Failure to process chromatin on apoptotic microparticles in the absence of deoxyribonuclease 1 like 3 drives the development of systemic lupus

erythematosus

Benjamin Sally

Submitted in partial fulfillment of the requirements for the degree of

Doctor of Philosophy in the Graduate School of Arts and Sciences

COLUMBIA UNIVERSITY

2017

© 2017 Benjamin Sally

All rights reserved

ABSTRACT

Failure to process chromatin on apoptotic microparticles in the absence of deoxyribonuclease 1 like 3 drives the development of systemic lupus

erythematosus

Benjamin Sally

Systemic lupus erythematosus is an autoinflammatory disorder driven by

the development of autoantibodies to self-nucleic acids, in particular to DNA and

to chromatin. Loss-of-function mutations of the secreted deoxyribonuclease

DNASE1L3 have been implicated in the development of aggressive familial lupus.

In addition, recent genome-wide association studies have linked a hypomorphic

variant of DNASE1L3 to sporadic lupus. Studies in the lab determined that

Dnase1l3-deficient mice develop rapid autoantibody responses against dsDNA

and chromatin, and at older ages this leads to a lupus-like inflammatory disease.

These disease manifestations were completely independent of the intracellular

DNA sensor STING, which has been implicated in other examples of self-DNA

driven autoinflammatory diseases. My project focused on developing assays to

track the activity of DNASE1L3, as well as identifying the endogenous source of

self-DNA normally processed by DNASE1L3. Using mouse models that allow the

depletion of specific cell populations, we found that circulating DNASE1L3 is

produced by hematopoietic cells, in particular by CD11c+ dendritic cells and by

tissue macrophages. Taking into account the unique properties of DNASE1L3,

we discovered that this enzyme is uniquely able to digest chromatin contained

within and on the surface of apoptotic microparticles. Loss of DNASE1L3 activity

in circulation results in elevated levels of DNA in plasma, in particular within

microparticles. Microparticles are extensively bound by anti-chromatin

autoantibodies isolated from both murine models of lupus as well as prototypical

human clones. In addition, Dnase1l3-deficient mice have high levels of circulating

IgG that bind to microparticles from young ages, and these titers increased as

disease progressed in aged animals. Pretreatment of microparticles with

DNASE1L3 largely abrogated this binding, demonstrating that DNASE1L3

directly reduces the immunogenicity of microparticles. We also studied two

human patients with null mutations in DNASE1L3, and observed increased DNA

circulating in plasma and, in particular, in their microparticles, demonstrating a

conserved role for DNASE1L3 in mice and humans. Finally, we obtained plasma

samples from a cohort of patients with sporadic SLE, and found that roughly 80%

had circulating IgG that avidly bound microparticles. Roughly half of this group

failed to bind to microparticles that had been pretreated with DNASE1L3, and this

DNASE1L3-sensitive group also presented with lower levels of DNASE1L3

activity. We conclude that extracellular chromatin associated with microparticles

acts as a potential self-antigen capable of causing loss of tolerance to self-DNA

and inflammatory disease in both mice and humans. The secretion of a DNA-

processing enzyme thus represents a novel, conserved tolerogenic mechanism

by which dendritic cells restrict autoimmunity.

i

Table of Contents

List of figures ii Introduction Systemic lupus erythematosus in humans and mice 1 Processing of self-DNA as a mechanism to prevent autoimmunity 14 Mice lacking Dnase1l3 develop autoimmunity 21 Chapter 1: Development of a functional assay to identify sources of DNASE1L3 Background 26 Results 29 Conclusions 44 Chapter 2: Microparticles as the physiological target of DNASE1L3 Background 46 Results 48 Conclusions 60 Chapter 3: Microparticles as autoantigens Background 62 Results 63 Conclusions 74 Chapter 4: DNASE1L3 and microparticles in human lupus patients Background 76 Results 77 Conclusions 89 Discussion and future directions 91 Materials and methods 107 References 121

ii

List of figures

Figure 1 26 Phylogenetic tree of human and mouse deoxyribonucleases Figure 2 28 Representation of the key features of DNASE1L3 Figure 3 29 Standard curve for detection of DNASE1L3 activity by qPCR Figure 4 31 The C terminal domain of DNASE1L3 allows its activity to be specifically tracked as it confers the ability to digest coated DNA Figure 5 33 The catalytic domain of DNASE1 and DNASE1L3 share extensive homology Figure 6 35 Perturbations of the C terminal domain of DNASE1L3 dramatically affect its function Figure 7 36 Dnase1l3-deficient animals do not process liposome-coated DNA ex vivo or in vivo Figure 8 37 Active DNASE1L3 in circulation is produced by hematopoietic cells Figure 9 39 The production of autoantibodies in Dnase1l3-deficient mice is driven by hematopoietic cells Figure 10 40 Restoration of circulating DNASE1L3 into Dnase1l3-deficient animals delays the development of autoimmunity Figure 11 42 DNASE1L3 is expressed primarily by dendritic cells and macrophages Figure 12 44 DNASE1L3 in circulation is produced primarily by dendritic cells and macrophages

iii

Figure 13 49 Treatment with staurosporine induces rapid cell death and microparticle release Figure 14 52 DNASE1L3 is uniquely able to process microparticle DNA Figure 15 54 A membrane permeable DNA dye extensively stains microparticles only in the absence of DNASE1L3 Figure 16 55 DNASE1L3 treatment of microparticles alters their surface composition Figure 17 57 Dnase1l3-deficient mice have higher levels of DNA circulating in plasma Figure 18 58 Microparticles from Dnase1l3-deficient mice carry increased amounts of DNA compared to those of WT mice Figure 19 59 Dnase1l3-deficient mice fail to process exogenous microparticles in vivo Figure 20 65 DNASE1L3-sensitive chromatin on the surface of microparticles is antigenic in Dnase1l3-deficient mice Figure 21 66 Microparticles from Dnase1l3-deficient mice at young ages are more coated by IgG and have higher levels of exposed self-antigens Figure 22 68 Immunization with microparticles drives anti-chromatin antibody responses in inflammatory contexts Figure 23 69 Treatment of microparticles with DNASE1L3 but not DNASE1 prevents binding by DNA-specific autoantibodies Figure 24 71 The C terminal domain of DNASE1L3 is required for the processing of microparticle chromatin

iv

Figure 25 73 A subset of human 9G4+ antibodies bind to microparticles only in the absence of DNASE1L3 Figure 26 78 Analysis of human microparticles Figure 27 80 DNASE1L3 null mutations in human patients result in defective DNA processing that parallels Dnase1l3-deficient mice Figure 28 82 DNASE1L3 null mutations in humans result in defective processing of endogenous microparticles in circulation Figure 29 83 DNASE1L3-sensitive binding of patient IgG to exogenous microparticles Figure 30 85 Extensive binding of microparticles by the plasma of patients with sporadic SLE Figure 31 86 Classification of sporadic SLE patients based on DNASE1L3 sensitivity of IgG binding Figure 32 88 The activity of DNASE1L3 in the circulation of sporadic SLE patients differs according to the observed pattern of staining Figure 33 89 Changes in DNASE1L3 activity in sporadic patients impacts the DNA cargo of microparticles Figure 34 97 Proposed mechanism by which DNASE1L3 prevents the development of autoimmunity

v

Acknowledgements

I would be remiss to not begin by acknowledging my PI, Dr. Boris Reizis.

He is not only a brilliant scientist and educator, but a phenomenal mentor as well.

Special mention must also go to Dr. Vanja Sisirak, a postdoctoral researcher in

the lab with whom I worked very closely on this project. Indeed, Vanja was

instrumental in my training after I joined the lab, and his input and guidance have

proved vital many times over. I have been fortunate enough to have worked with

two wonderful departments; the Department of Microbiology and Immunology of

Columbia and the Department of Pathology of NYU. In addition, I would like to

thank all of my fellow lab members. The environment is truly wonderful; everyone

is always willing to assist with troublesome experiments or give advice.

Furthermore, I would like to acknowledge all of our collaborators, who

have provided invaluable reagents, materials, and discussions. In particular, I’d

like to acknowledge Dr. Jill Buyon and Dr. Robert Clancy, who have graciously

allowed us access to all of the human samples analyzed. In addition, they have

been tremendously generous with their time and expertise. I would also like to

thank all of the patients for providing material for this study; without such

contributions our work would not be possible.

Finally, my profound thanks to my advisory committee at Columbia: Dr.

Donna Farber, Dr. Uttiya Basu, Dr. Kang Liu, and the non-affiliated evaluator Dr.

Robert Clancy. Your advice and support has been paramount during my time as

a PhD student, and I appreciate your willingness to be available for this whole

process.

vi

To my family, friends, and Mariana: Your love and support has been the cornerstone

for anything and everything I’ve accomplished

- 1 -

Introduction

Systemic lupus erythematosus: prevalence and therapeutic options

Systemic lupus erythematosus (SLE or lupus) is an autoinflammatory

disease wherein pathogenesis is driven by the production of antibodies (Abs) to

self-nuclear antigens. Targets may include ribonucleoproteins and/or nucleic

acids. Following the initial loss of tolerance, auto-Ab production amplifies further

inflammatory responses via the innate immune system, resulting in the activation

of myeloid pathways and the production of type 1 interferon (interferon /, IFN)

(1). In turn, the production of auto-Abs is dramatically enhanced by these

inflammatory molecules, resulting in a feed forward pathogenic loop (1). Immune

complexes are then formed by these auto-Abs and nucleic acids, and upon their

deposition in various tissues cause chronic inflammation, including arthritis,

vasculitis, and glomerulonephritis (1). An additional aspect of SLE pathogenesis

is its unpredictable flaring, the causes of which vary considerably between

patients (1, 2). Known triggers include but are not limited to pregnancy (3), viral

or bacterial infection (4), ultraviolet radiation (5), stress (6), and exposure to

various environmental agents including pharmaceuticals, chemicals, or food

products (7).

Current estimates as to the prevalence of SLE range from 20-150 cases

per 100,000 persons (2). Like most complex inflammatory disorders, there are

numerous genetic and environmental factors that have been linked to SLE

development. Indeed, studies have shown that the concordance for SLE among

- 2 -

monozygotic twins is approximately 24% (8). In addition, there is strong evidence

for involvement of the endocrine system in SLE, as roughly 90% of patients are

female (9). Furthermore, rates of SLE dramatically differ depending on ethnicity,

with African Americans being a particularly vulnerable population for reasons

unknown. Consequently, among African American women the rate is over 400

cases per 100,000 persons (10).

Because the presentation of SLE can vary wildly between patients, or,

indeed, between flares of a single individual, its identification and assessment

can be challenging. Currently, there are two main scoring systems for SLE

patients: the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI)

and the American College of Rheumatology (ACR) criteria (11, 12). Both

ultimately evaluate patients based on the presence or absence of a wide variety

of symptoms, ranging from neurological (psychosis, seizures, sensory or motor

neuropathy, etc.) to hematological (anemia, leukopenia, thrombocytopenia) to

immunological (presence of auto-antibodies). Importantly, not every symptom

has to be present to result in a diagnosis of SLE; of the 11 ACR criteria the

presence of at 4 yields a sensitivity of 85% and a specificity of 95% for SLE (12).

SLEDAI scores, alternatively, are used to monitor patients longitudinally, as

fluctuations in score play a major role in determining changes to the course of

treatment.

Despite the high (and increasing) prevalence of SLE, therapeutic options

remain inadequate. Because little is known regarding the mechanisms driving

loss of tolerance in SLE, current treatments are limited to the treatment of

- 3 -

symptoms rather than the underlying pathology (13). Widely used strategies

include over the counter or prescription nonsteroidal anti-inflammatory drugs

(NSAIDs), the anti-malarial drug hydroxychloroquine, prednisone or other

corticosteroids, and, in severe cases, immunosuppression with

chemotherapeutics such as methotrexate (13, 14). These therapies all represent

broad anti-inflammatory interventions that fail to address the specific dysfunction

of normal immune pathways driving SLE. Furthermore, there are few effective

therapies in development; indeed, there has been only one new drug

successfully developed and approved by the FDA for the treatment of SLE in the

past 60 years (15).

Pathogenesis of SLE: cells and cytokines

At the macro level, both branches of the immune system play essential

roles in the development of SLE. The innate immune system contributes both in

the early and late stages of the disease. Initially, myeloid populations such as

dendritic cells activate T cells and produce B-cell activating factor (BAFF), which

drives the adaptive response (16). Activated T lymphocytes, specifically TH1 and

TH17 helper populations then drive the systemic and renal activation of B cells

(17-19). B cell activation results in the production of anti-nuclear antigen (ANA)

autoantibodies, which form immune complexes and are deposited throughout the

body in various tissues, causing inflammation and pathogenesis via the activation

of localized innate pathways (19, 20). Deposition of these autoantibody

complexes at different sites in the body account for the variation in symptoms

- 4 -

between individuals and/or flares (21). Thus, the activation of autoreactive B cells

and their subsequent release of autoantibodies represents the central hallmark of

SLE.

While the primary cause of SLE is the deposition of immune complexes,

deregulation of inflammatory cytokines is a major contributor to immune

dysfunction and ultimately tissue injury. Inflammatory cytokines, such as type I

and type II interferons (IFNs), interleukin (IL)-6, IL-1, and tumor necrosis factor-

alpha (TNF), have all been implicated in the pathogenesis of SLE (22-27).

Furthermore, immunomodulatory cytokines have also shown to play roles in

preventing SLE, including IL-10 (28) and transforming growth factor-beta (TGF)

(29). Finally and more broadly, regulatory T cells play an important role in limiting

deleterious autoinflammatory conditions such as lupus, as shown by recent

studies linking IL-21, IL-17, and IL-2 to SLE (30).

Work from the group of Virginia Pascual has shown that SLE patients

frequently present with an enhanced type I IFN signature (31). Type I IFNs can

have extensive systemic effects at multiple levels that contribute to SLE

pathogenesis. Firstly, IFN promotes positive feedback loops, enhancing the

synthesis of additional type I IFN via the induction of Toll-like receptor (TLR) 7

(32). Secondly, IFNs drive the maturation of dendritic cells (DCs), shifting the

balance of immature versus mature DCs (33). Crucially, immature DCs are

important in maintaining regulatory T cells in the periphery, and also help

promote deletion of autoreactive T cells by presenting self-antigens in the

absence of costimulatory molecules (33). Thirdly, mature DCs produce BAFF,

- 5 -

which directly enhances the selection and survival of autoreactive B cells (34).

Finally, IFN both enhances the cytotoxic capabilities of CD8 T cells (35) and

also increases the number of autoreactive CD4 T cells by driving the upregulated

expression of CD80 and CD86 on APCs (33).

IL-6 is a proinflammatory cytokine that, in SLE, acts primarily at the level

of B cells and T cells. It has been shown to play a key role in driving the

hyperactivation of B cells (36) and in promoting autoreactive T cell responses

(37). In mouse models of rheumatoid arthritis, IL-6 has been shown to be vital for

the differentiation of TH17 helper cells (38), a proinflammatory population that has

also been implicated in the pathogenesis of SLE (18). Finally, IL-6 has been

shown to play important roles in mediating local inflammatory responses in lupus

patients. Elevated IL-6 is found in the kidneys of patients presenting with lupus

nephritis (39), as well as in the cerebrospinal fluid of patients presenting with

neuropsychiatric complications (40).

IFN-gamma (IFN) is the primary cytokine that drives TH1 responses. In

other contexts, TH1 responses have been shown to counteract aberrant

autoinflammatory reactions, largely thought to be because they inhibit the

development of autoreactive TH17 cells (41). However, in the context of lupus it

has been shown to induce production of IgG2a isotype antibodies, which can

contribute to autoimmunity by the activation of the complement system (42).

Lupus-prone mice deficient for IFN present with an ameliorated form of the

disease (43), and in human patients it has been suggested that an imbalance in

favor of TH1 responses can drive disease (44). However, additional clinical

- 6 -

studies have proved inconclusive, with some suggesting that levels of IFN are

tied closely to the severity of lupus nephritis (45), while others observe no

correlation (46). Although these data are not clear cut, what is apparent is that if

IFN plays a role in SLE pathogenesis, it is likely in exacerbating disease after

the initial loss of tolerance.

In addition to the overproduction of proinflammatory cytokines, SLE can

be exacerbated by disruption of regulatory cytokine production. Lupus-prone

mice that also lack IL-10 develop a more severe form of the disease (28). IL-10

acts to dampen proinflammatory responses, including but not limited to TH1

helper cells, TH17 helper cells, and macrophages, making it a key molecule in

preventing autoimmunity (47). Along these lines, TGF- is strikingly reduced in

lupus patients undergoing active disease (29). TGF- also plays an

immunosuppressive role, particularly via modulation of lymphocyte activation and

differentiation (48). Reductions in TGF- lead to a loss of regulatory T cells in the

periphery and an overall increase in inflammatory profile, providing a permissive

environment for autoreactive responses (48).

IL-23 and IL-17 act synergistically to drive the differentiation of TH17

helper cells, which as previously discussed are a potent proinflammatory subtype.

These cells are generally balanced by the presence of regulatory T cells, which

produce anti-inflammatory cytokines and prevent autoimmunity. IL-2, produced

by other T cells, is a critical growth factor for regulatory T cells in particular.

Lower levels of IL-2 lead to a striking reduction in their population (49), and a

corresponding increase in TH17 cells (50). The T cells of active SLE patients

- 7 -

produce less IL-2 compared to controls (51), and patients also present with a

reduced population of regulatory T cells (52). Thus, perturbing the balance of

proinflammatory versus regulatory T cells correlates with SLE severity, indicating

complex and opposing roles for these populations.

Mouse models of SLE and their applicability to human disease

One major factor that has drastically hampered understanding of human

SLE pathogenesis and consequently drug development is the lack of

representative mouse models. The oldest and most widely used model is the

New Zealand Black/New Zealand White (NZB/NZW) F1 hybrid, which

spontaneously develops lupus-like symptoms such as lymphadenopathy,

splenomegaly, and ANA autoantibodies whose repertoire includes anti-DNA (53,

54). These anti-double stranded DNA (dsDNA) antibodies typically lead to

glomerulonephritis and eventually fatal kidney failure (53, 54). Additional study of

these mice led to the identification of three genetic loci that, when all were

present in combination, drove the development of lupus-like disease in these

mice, dubbed Sle1-3 (55). Unfortunately, due to the requirement for all three loci

to be present, extensive genetic studies using this model are extremely

challenging and, generally, unfeasible. Furthermore, in stark contrast to many

human patients where anti-DNA antibodies are detectable prior to the

development of overt symptoms (56), NZB/NZW mice develop autoantibodies

against DNA slowly, and typically only after extensive immune activation (57).

Ultimately, this suggests that while anti-dsDNA antibodies may play a role in the

- 8 -

ongoing pathogenesis of these animals, loss of tolerance to DNA is unlikely to be

the initial insult driving SLE development.

An alternative model was subsequently discovered. A substrain of the

inbred MRL line was characterized that developed a sex-independent SLE-like

disease characterized by lymphadenopathy driven by double negative immature

T cells (58). These mice also presented with high levels of circulating immune

complexes driven by ANA IgG, including high anti-dsDNA (58). Ultimately, this

phenotype was attributed to an autosomal recessive mutation in the Fas receptor,

dubbed lymphoproliferation (lpr). Mice harboring this mutation are deficient for

Fas activity, which dramatically impairs apoptosis in lymphocytes (59). This

defect in negative selection of autoreactive cells is sufficient to drive the

development of lupus-like disease. Crucially however, subsequent studies

demonstrated that mice positive for the lpr mutation from other strains were

asymptomatic, indicating vital roles for genes specific for the MRL background

(60). Thus, the requirement for a specific background represents a significant

obstacle for in depth genetic analysis. In addition, the lpr model suffers from the

same issue of specificity as the NZB/NZW mice, namely that loss of tolerance to

DNA is a consequence of inappropriate immune activation and is not driving

pathogenesis.

Further efforts to pinpoint a monogenic model of disease led to the

identification of the BXSB mouse strain, derived from the F1 intercross between

C57BL6 and SB/Le strains and subsequent backcrossing onto SB/Le (54). These

mice developed lupus-like disease that, curiously, was much more severe in

- 9 -

males, and further study identified the presence of the Y-linked autoimmune

accelerator (Yaa) genetic element (61, 62). It was subsequently discovered that

this element is due to a translocation of a portion of the X chromosome to the Y

chromosome, resulting in the duplication of several genes and a 2-fold increase

in expression (63, 64). Crucially, one of these duplicated genes is TLR7, which

encodes an innate pattern recognition receptor specific for viral ssRNA (63-66).

Activation of TLR7 results in signaling through myeloid differentiation primary

response gene 88 (MyD88), leading to the activation of proinflammatory

pathways in DCs and B cells (67, 68). Importantly, signaling through TLR7 has

been shown to act as a second signal in B cells, allowing for activation of

autoreactive cells upon interactions with endogenous RNA ligands (69, 70).

Consequently, TLR7-transgenic mice have become a widely-used model to study

SLE (32, 63, 66, 67, 71, 72) and further analysis has revealed that

polymorphisms in TLR7 are linked to human disease in certain populations (65,

73).

However, this model does have drawbacks. Disease in TLR7-transgenic

animals is primarily driven by loss of tolerance to various forms of self-RNA (32,

63, 67), but in human disease anti-DNA antibodies are more pathogenic and

predictive of disease outcomes (56, 74). An additional complication stems from

the mechanism by which TLR7 is regulated. The endosomal TLRs (TLR7, TLR9,

and TLR3) are produced in the endoplasmic reticulum and are dependent on the

chaperone protein Unc-93 homolog B1 (UNC93B1) for trafficking to endosomes

(75). Consequently, changes in the expression level of one endosomal TLR also

- 10 -

affect the availability of UNC93B1, meaning that TLR7 transgenic mice also

present with reduced levels of endosomal TLR9 (71). Importantly, loss of TLR9

has been linked to exacerbation of autoimmunity (76), and it is unclear if this is

due to increases in TLR7 or if TLR9 has additional immunomodulatory functions.

A final downside of this model is that TLR7 is sufficient to drive disease only at

very high doses, as the two-fold increase caused by Yaa translocation requires

the presence of additional susceptibility loci for overt autoimmune disease (77).

Thus, the requirement for high levels of TLR7 complicates genetic analysis and

the extent to which it mimics human disease broadly is unclear.

In addition to genetic models of SLE, exposure to environmental triggers

has been used to induce the disease independently of mouse background. Chief

among these is the pristane-induced lupus model, which is driven by

intraperitoneal injections of pristane (2,6,10,14-tetramethylpentadecane) into

mice on a BALB/c background (78, 79). These mice develop autoantibodies

characteristic of lupus, including anti-histone, anti-ribonucleoprotein, and anti-

DNA variants, generally presenting with similar repertoires and disease

progression as MRL/lpr mice (78, 79). Additional mouse backgrounds have

subsequently been tested, and develop autoantibodies in response to pristane

injection to varying extents (80). Although this model is again not driven by loss

of tolerance to DNA specifically, it has proved useful in determining whether

various inflammatory pathways contribute to lupus pathogenesis. For instance,

injection of pristane into mice deficient for IL-6 demonstrated the importance of

this cytokine for the development of anti-DNA autoantibodies (23). Additionally,

- 11 -

pristane injection into mice deficient for IFN-I receptor (IFNAR) led to extensive

but not total ablation of the disease (81, 82), which corresponds with studies

linking overexpression of interferon with disease severity in human patients (83-

85). One key difference, however, is that whereas in humans plasmacytoid

dendritic cells (pDCs) are the key producers of interferon (86), in the pristane-

induced mouse model it is produced by Ly6Chi monocytes that aggregate in the

peritoneal cavity (87). Furthermore, interferon production by these cells requires

the presence of TLR7, again suggesting that there is no loss of tolerance to DNA

specifically in this model (88).

Thus, there is a critical need for the development of a monogenic mouse

model that presents with extensive anti-DNA responses before proceeding to

disease. While some of the aforementioned mouse models feature these

autoantibodies to varying extents, none of these models recapitulate loss of

tolerance to self-DNA. Modeling anti-DNA responses is important because high

affinity IgG specific for dsDNA are especially pathogenic and have been

demonstrated to correlate with the severity of symptoms in human disease, in

particular glomerulonephritis and kidney injury (56, 74, 83, 89). Furthermore, loss

of tolerance to self-DNA seems to be a key event leading to the development of

sporadic disease, as many patients present with high levels of anti-DNA

autoantibodies despite being otherwise asymptomatic (56).

- 12 -

Genetics of SLE

SLE in humans typically falls into one of two categories: sporadic disease

and familial (or monogenic) disease. Inherited SLE is very rare, and in humans

has been definitively linked to only a small number of genes, all of which are

involved in cell death pathways or in the clearance of apoptotic debris (90). In the

study of sporadic disease, numerous genome wide association studies (GWAS)

have identified more than 50 common risk loci for SLE susceptibility.

Unsurprisingly, the strongest association signal among common SLE variants is

obtained from the HLA region (90). Other genes implicated generally fall into four

different categories: activation of the adaptive immune system, innate immune

signaling, renal-specific factors, and clearance of self-nucleic acids (91). Genes

affecting activation of the adaptive system include those that modulate B cell

receptor signaling (92), T and B cell crosstalk via MHC (93), and the

differentiation of T cells into various effector phenotypes (94). Innate signaling

pathways known to be dysregulated include the type I IFNs (95), immunoglobulin

Fc receptors (96), and various molecules upstream of nuclear factor kappa-light-

chain-enhancer of activated B cells (NF-B) (97). Renal-specific factors include

localized activation of myeloid cells or lymphocytes (98, 99), clearance of

immune complexes from glomeruli (100), and potential disease pathways active

in resident renal cells, such as angiotensin-converting enzyme (ACE) (101). The

last category of molecules implicated by GWAS are predominantly involved in the

processing of self-antigens, and includes the complement system (102) and the

clearance of apoptotic cells (103).

- 13 -

Identifying the antigenic form of self-DNA driving pathogenesis in SLE

In addition, a major outstanding question in the field concerns the nature

of antigenic self-DNA that drives loss of tolerance. Despite extensive efforts, it

remains unclear which specific form of self-DNA is recognized by autoreactive B

cells, although different types of self-DNA have been implicated in SLE

pathogenesis such as oxidized mitochondrial DNA and neutrophil extracellular

traps (104-107). Mitochondria are organelles present in all nucleated cells, and

contain their own DNA, packaged into structures called nucleoids (108). Because

mitochondria are the site of oxidative phosphorylation in the body, they

accumulate large quantities of reactive oxygen species, leading to a high rate of

mutations due to oxidation of their DNA and the lack of proofreading DNA repair

machinery (108). In lupus patients, neutrophils accumulate this oxidized

mitochondrial DNA as a consequence of impaired nucleoid digestion, and this

DNA is instead extruded into the extracellular space (104). Oxidized

mitochondrial DNA is both a potent stimulator of interferon and a target of

autoantibodies in these individuals, as it is recognized by TLR9 in pDCs (104,

107).

In addition, oxidized mitochondrial DNA can be incorporated into

neutrophil extracellular traps (NETs) (107). NETs are composed of processed

self-DNA bound to various cytoplasmic proteins that are then organized into large

filamentous structures (109). NET generation occurs via a specific form of cell

death called NETosis, wherein fragmentation of the nucleus results in the

availability of chromatin to be integrated and released in NETs (109). These

- 14 -

structures play key antimicrobial roles, both via direct binding to pathogens (110)

and by delivery of high local concentrations of antimicrobial peptides (111).

Importantly, they also play a role in the inactivation of virulence factors via the

delivery of intracellular enzymes, such as elastase (109). Furthermore, they act

to contain microbes at the initial site of infection, preventing dissemination (109).

However, because NETs contain abundant self-DNA, they have been linked to

the development of autoimmunity, in particular through activation of TLR9 (105,

106). Thus, although both oxidized mitochondrial DNA and NETs may act as

second signals, it is unclear whether they are capable of functioning as antigens

for DNA-specific autoreactive B cells.

Vital roles for deoxyribonucleases in the prevention of autoimmunity

The existence of nucleic acid-sensing innate immune receptors presents a

fundamental problem for the body, given the abundance of self-DNA and the

limited capacity for structural differences been self and non-self nucleic acids.

This paradigm is addressed by the vitally important processing of self-DNA so as

to render it non-immunogenic. This function is performed by a class of enzymes

called deoxyribonucleases (DNASEs), which catalyze the hydrolytic cleavage of

DNA phosphodiester linkages. These fall into three broad categories, DNASE1

and the DNASE1-like family members, DNASE2, and DNASE3, which is also

known as three-prime repair exonuclease 1 (TREX1).

DNASE2 is expressed at high levels in phagocytic cells, in particular

macrophages, and is responsible for the digestion of DNA accumulated in

- 15 -

lysosomes (112). Its potent activity is regulated by the requirement for activation

by low pH and cleavage by lysosomal proteases (113). DNASE2 plays important

roles in the digestion of DNA from apoptotic cells and also nuclei extruded from

erythrocytes, both of which are normally engulfed by phagocytes (114). This

function is absolutely vital, as mice deficient for DNASE2 die during embryonic

development due to massive production of interferon triggered by undigested

self-DNA (115, 116). Mice doubly deficient for both DNASE2 and IFNAR were

viable, though did develop chronic polyarthritis that resembled human

rheumatoid arthritis, driven in part by the leaking of DNA from lysosomes into the

cytoplasm (116). It was subsequently discovered that recognition of accumulated

self-DNA by the cytosolic DNA sensor and adaptor protein Stimulator of

interferon genes (STING) was responsible for driving autoimmunity in this model,

as mice lacking both DNASE2 and STING were both viable and completely failed

to develop arthritis (117). STING activates the transcription factor IFN regulatory

factor 3 (IRF3), leading to the production of IFN and other inflammatory

chemokines (118). It remains unclear what upstream mechanisms are

responsible for activation of STING in this context. STING is commonly triggered

by the presence of cyclic 2’-5’ 3’-5’ GMP-AMP dinucleotides (cGAMP) produced

by cyclic GMP-AMP synthase (cGAS) (119). However, double deficiency of both

cGAS and DNASE2 failed to rescue the inflammatory phenotype despite the

accumulation of cGAMP in DNASE2 knockout mice (120), suggesting that there

are additional cytosolic DNA sensors upstream of STING activated by self-DNA.

- 16 -

TREX1 is the most abundant 3’ to 5’ exonuclease, and is localized to the

endoplasmic reticulum (112, 121). It functions to digest DNA that has been

reverse transcribed from endogenous retroelements, which are present

throughout the genome and are capable of driving the production of IFN if

unprocessed (122). In TREX1 deficient cells, ssDNA accumulates and is

recognized by cGAS, leading to activation of STING and aberrant production of

IFN via IRF3 (122, 123). Mice deficient for TREX1 are viable, but succumb to

IFN-driven inflammatory disease postweaning, with a median survival of less

than 6 months (122, 124). Cause of death is typically cardiomyopathy leading to

circulatory failure (124), but knockout mice also present with increased

macrophage activation and neural inflammation (125, 126). Furthermore,

mutations in TREX1 have been implicated in human disease. Discovered in 1984,

Aicardi-Goutières syndrome (AGS) is a rare inflammatory disorder that most

commonly affects the brain and skin, causing encephalopathy and ultimately

brain atrophy (127-129). The disease is driven by extremely high levels of IFN in

the cerebrospinal fluid in the absence of infection (130, 131). Subsequent work

revealed that mutations in any of seven genes causes AGS: the three genes

encoding the RNase H2 endonuclease complex (132), the deoxynucleoside

triphosphate triphosphohydrolase SAMHD1 (133), the dsRNA deaminase

ADAR1 (134), the cytosolic dsRNA receptor MDA5 (135), and TREX1 (136).

Notably, these enzymes are all involved in the processing or sensing of nucleic

acids, implicating the accumulation of self-nucleic acids in the IFN-driven

pathogenesis of AGS.

- 17 -

Additionally, a separate mutation in TREX1 has been shown to drive the

development of familial chilblain lupus erythematosus (CHLE) (137). CHLE is a

poorly understood form of lupus wherein lesions are triggered by exposure to

cold and/or damp climates, typically distributed symmetrically in the extremities

(138). These patients also often develop antibodies against Sjögren’s-syndrome-

related antigen A (SSA), commonly referred to as anti-SSA/Ro (139). Anti-

SSA/Ro antibodies recognize nuclear antigens and are one of the most

prominent specificities of autoantibodies in sporadic SLE (140), and have been

strongly implicated in the development of neonatal lupus and/or congenital heart

defects in utero (141). More broadly, genome-wide association studies have

revealed strong associations between TREX1 and sporadic disease severity,

suggesting a potential contribution to the exacerbation of symptoms (142).

However, although DNASE2 and TREX1 deficiencies lead to autoimmunity,

these responses are driven by aberrant interferon production rather than loss of

tolerance to DNA and are thus distinct from classical SLE. Nevertheless, these

findings emphasize the importance of self-DNA processing in the prevention of

autoimmunity.

While DNASE2 and TREX1 both digest intracellular forms of self-DNA,

DNASE1 is instead secreted extracellularly. DNASE1 was originally

characterized in the context of pancreatic exocrine function, as it is released

extensively into the digestive tract in addition to being present in circulation (143).

Given the importance of TLR signaling in the pathogenesis of lupus which is

suggestive of an extracellular form of DNA being key, there was considerable

- 18 -

interest in investigating potential roles for DNASE1 in SLE. Initial reports

suggested that DNASE1 knockout animals developed mild anti-DNA and ANA

(144). However, the autoimmunity displayed by these mice was underwhelming

compared to other models, and these results were never replicated. There were

also tenuous links between DNASE1 and human disease; reports of minor

reductions in DNASE1 activity (144) or DNASE1 mutations in small cohorts of

SLE patients (145). However, these studies were subsequently expanded upon

and it is now accepted that if a role for DNASE1 in promoting autoimmunity exists,

it is at best a minor contributor (146, 147). In addition, clinical trials wherein

DNASE1 was administered to lupus patients failed to demonstrate any

therapeutic benefit whatsoever (148).

Strong links between DNASE1L3 and multiple forms of SLE

However, DNASE1 is just one member of the DNASE1 family, which also

includes DNASE1-like 1 (DNASE1L1), DNASE1L2, and DNASE1L3. All family

members share very similar catalytic domains and depend on the presence of

Ca2+ and Mg2+ (149). DNASE1L3 came to prominence when it was identified in a

small cohort of consanguineous Saudi Arabian patients with an aggressive form

of inherited SLE affecting children at extremely young ages characterized by very

high titers of anti-DNA antibodies (150). Further study of these patients revealed

that they had a frameshift mutation in DNASE1L3 that led to total loss of the

enzyme’s function (150). A second study of similarly consanguineous patients in

Turkey with hypocomplementemic urticarial vasculitis syndrome (HUVS) also

- 19 -

identified two separate inactivating mutations (one frameshift, one insertion

leading to exon skipping) in DNASE1L3 (151). HUVS is a recurring urticaria

characterized by glomerulonephritis and loss of complement components, in

particular C1q (151). This seems to be driven by the development of high titers of

anti-C1q autoantibodies, leading to C1q depletion (151). Indeed, anti-C1q

antibodies have been extensively linked to the development of SLE, suggesting

that complement plays a key role in the clearance of potentially immunogenic

molecules from circulation (152-155). Consequently, HUVS is very strongly

associated with SLE, as the majority of patients with HUVS progress to develop

SLE at some point in life (151). Finally and most recently, an additional family of

patients was identified that also presented with extremely aggressive, early onset

disease in Italy (156). Featuring the same mutation as the Turkish family, one

extensively-studied Italian patient first presented with HUVS before proceeding to

extensive autoinflammatory disease, including SLE, polyarthritis, and intestinal

vasculitis, all of which were largely resistant to treatment with anti-inflammatory

therapeutics (156). Thus, these studies demonstrated that very rare cases where

DNASE1L3 function is completely lost lead to extremely aggressive anti-DNA

and anti-C1q responses, leading to rapid onset of SLE.

However, to date fewer than 10 patients lacking DNASE1L3 entirely have

been identified. It was unclear whether DNASE1L3 was relevant to sporadic

disease, as it had never been identified as having potential involvement in

sporadic SLE by GWAS. Curiously, one locus linked to SLE by GWAS

extensively is PXK (157, 158). PXK encodes for a kinase of unknown function,

- 20 -

and efforts to link PXK to SLE have yet to be conclusive. However, PXK is in very

close proximity to DNASE1L3 on chromosome 3, and closer examination of this

genetic region led to the association between SLE and PXK to be reassigned to

a single nucleotide polymorphism (SNP) in DNASE1L3 (159, 160). This particular

SNP results in an amino acid substitution wherein arginine is replaced by

cysteine at position 206 of DNASE1L3, and studies have shown that this R206C

DNASE1L3 is hypomorphic (161, 162). Thus, reinterpretation of GWAS have

now linked the function of DNASE1L3 with the development of sporadic SLE.

Unique features of DNASE1L3

Given these results, we decided to look more closely at DNASE1L3, also

known as DNASE. Like the other DNASE1 family members, it has the same

critical active site residues, an essential disulfide bridge, a calcium-binding

domain, and an N terminal signaling peptide (163, 164). Unlike DNASE1,

however, which is bound and inhibited by glomerular actin (165), DNASE1L3

lacks the residues that facilitate this interaction and is thus actin independent

(163). In addition, DNASE1L3 contains a greater number of positively charged

residues, resulting in a much more basic isoelectric point compared to DNASE1

(9.5 vs 4.8) (164). Notably, efforts to enhance the activity of DNASE1 to improve

therapeutic applications led researchers to replace residues at six positions with

basic substitutes, resulting in a hyperactive DNASE1 (166). At four of these six

positions DNASE1L3 already features basic residues, suggesting that it likely

more potent than DNASE1.

- 21 -

Another important difference is the C terminus of these enzymes.

DNASE1L3 has a unique and highly basic domain at its C terminus that is very

strongly conserved across species (167). One of the earliest described

applications for DNASE1L3 was its ability to block liposomal transfection, a

function not possessed by DNASE1 (167). This activity was dependent on the C

terminal domain, as a truncated form of DNASE1L3 was able to process

uncoated but not coated DNA (167). In addition, this domain is sufficient to confer

this activity, as fusing it to the C terminus of DNASE1 allows for the resulting

chimeric protein to block transfection (167). In addition to digesting liposome-

coated DNA, DNASE1L3 is also able to digest chromatin even in the absence of

a helper protease, unlike DNASE1 (168, 169). This finding has led to suggestions

that DNASE1L3 plays a role in nuclear fragmentation during apoptosis (170, 171)

and necrosis (172). However, DNASE1L3 is secreted into the extracellular space

and requires cleavage of its N-terminal signal peptide by the secretory pathway

(164), pointing to a likely function for DNASE1L3 in processing extracellular DNA.

In addition, DNASE1L3 is not upregulated in apoptotic cells (164), so any role

during cell death may be a secondary function.

Generation and characterization of Dnase1l3-deficient mice

Given the strong links between DNASE1L3 and both inherited and

sporadic SLE, our laboratory decided to investigate whether mice lacking

DNASE1L3 would develop the disease. Essential coding exons of Dnase1l3 was

targeted via gene-trapping, resulting in the introduction of a LacZ cassette in its

- 22 -

place (173). Dnase1l3LacZ/LacZ knockout (KO) mice were viable, fertile, and were

born at Mendelian frequencies. Given that a major shortcoming of other lupus

models was the reliance on a particular background, Dnase1l3 KO mice were

crossed onto two common inbred strains, C57BL/6 (B6) and 129SvEv (129). In

addition, mice on mixed and F1 backgrounds were analyzed.

All KO mice on both pure backgrounds and on mixed backgrounds

developed ANA that stained HEp-2 cells perinuclearly, a pattern consistent with

severe human SLE (173). Both male and female mice as young as 5 weeks of

age presented with elevated levels of autoantibodies against both dsDNA and

chromatin as measured by enzyme linked immunosorbent assay (ELISA) (173).

Notably, there was no increase in total IgG in anti-RNA IgG at any time point

tested (173), demonstrating that loss of DNASE1L3 causes loss of tolerance

against DNA specifically rather than broad immune activation. The rapid and

specific responses to dsDNA and chromatin in these animals thus strongly

suggest endogenous genomic DNA represents the primary autoantigen.

Having observed loss of tolerance to self-DNA at very early ages in these

knockout mice, our laboratory next wanted to determine whether they went on to

develop other symptoms of SLE. As these mice aged, there was an expansion of

the CD11c+ MHC class II- CD11b+ Ly-6C- inflammatory monocyte population

(173), which have been described in other models as being key for clearance of

immune complexes (174). KO mice also developed splenomegaly at 50 weeks of

age, and within these spleens presented with spontaneous germinal center (GC)

formation and a corresponding increase in GC B cells (173). All KO mice also

- 23 -

displayed overt kidney pathology by 50 weeks, with significantly higher IgG

deposition in kidney glomeruli compared to WT mice (173). In addition, KO mice

on the 129 background had additional kidney irregularities, presenting with

extensive glomerulonephritis (173). These studies in the lab thus conclude that

the initial loss of tolerance to self-DNA in KO mice precedes extensive immune

activation, deposition of immune complexes, and glomerulonephritis, all

hallmarks of sporadic SLE in humans.

Acceleration of disease in Dnase1l3-deficient mice after treatment with IFN

Our laboratory was intrigued by the observation that despite extremely

early anti-DNA responses, extensive immune activation was not present in KO

mice until roughly 30 weeks of age. Previous studies have shown that artificially

increasing the production of IFN can dramatically accelerate the development of

SLE (175). To test whether this also applied to Dnase1l3-deficient animals, WT

and KO mice at young ages were injected with an adenoviral vector encoding

IFN-5. The efficacy of the vector was determined by both direct measurement of

serum IFN and of the IFN-inducible marker Sca-1 (173). Notably, KO mice

treated with IFN rapidly developed anti-dsDNA IgG, and in addition presented

with novel anti-RNA IgG responses (173). In contrast to previous experiments

using KO mice, animals that received IFN had expanded compartments of

inflammatory monocytes and T cells one-week post-injection (173). Furthermore,

roughly two thirds of KO mice who received IFN perished after 35 weeks (173).

These studies from the lab thus conclude that increases in IFN expression

- 24 -

dramatically accelerate the immune activation observed in Dnase1l3-deficient

animals, which is reflective of severe SLE in DNASE1L3-deficient human

patients and sporadic patients where IFN is upregulated.

Disease in Dnase1l3 KO animals is independent of STING but not MyD88

Previous studies of DNASE1L3 have led to suggestions that it processes

intracellular DNA (170-172), in a manner similar to DNASE2 or TREX1 (90, 176).

If this were the case, one would predict that autoreactivity in Dnase1l3-deficient

animals would be dependent on STING, as is the case in DNASE2 and TREX1

knockout animals. To test the potential involvement of STING in driving anti-DNA

responses in our model, mice doubly deficient in STING and DNASE1L3 or

MyD88 and DNASE1L3 were generated. Characterization of these double

knockouts revealed that ablation of STING did not reduce the levels of ANA, anti-

dsDNA IgG, anti-dsDNA-specific antibody secreting cells (ASCs) as measured by

ELISPOT, kidney deposition of IgG, or splenomegaly (173). In stark contrast,

when MyD88 was deleted in Dnase1l3-deficient animals every parameter tested

returned to WT levels (173). Because MyD88 is responsible for transducing

signals through TLRs and the IL-1 receptor, these data strongly suggest that

DNASE1L3 targets extracellular DNA, which is in keeping with its secreted

nature. In addition, these results clearly distinguish loss of DNASE1L3 from other

examples of DNASE deficiency, as the autoinflammation caused by loss of

DNASE2 and TREX1 are both rescued by the additional loss of STING.

- 25 -

Thus, these studies from the lab strongly suggest that Dnase1l3-deficient

animals represent a novel model of SLE, wherein the initial loss of tolerance is

against an extracellular form of genomic DNA. These mice then go on to develop

overt disease much later in life, presenting with classical hallmarks of SLE:

expansion of inflammatory monocytes, splenomegaly, increased germinal center

reactions, and kidney IgG deposition resulting in overt glomerulonephritis.

Furthermore, the dispensability of STING for disease demonstrates that

DNASE1L3 functions differently than TREX1 or DNASE2 in that it is not

processing intracellular DNA. My project has focused on addressing outstanding

questions concerning the mechanism by which DNASE1L3 prevents

autoimmunity. Chiefly, we have attempted to identify the form or forms of

extracellular DNA processed by DNASE1L3, and also sought to understand the

cell type or types responsible for the production of DNASE1L3 in the steady state.

- 26 -

Chapter 1

Development of a functional assay to identify sources of DNASE1L3

Evolutionary relationships between the DNASE1 family members

Because DNASE1L3 is more potent, actin-independent, and has the

unique C terminal domain, we wanted to determine the evolutionary relationship

between the various deoxyribonucleases. To wit, using software from

www.phylogeny.fr (177, 178), we generated a phylogenetic tree using human

and mouse DNASE1, DNASE1L1, DNASE1L2, DNASE1L3, DNASE2, and

TREX1. Our analysis revealed that DNASE1L3 appears to have been the first

DNASE1 family member that evolved from DNASE2 (Fig. 1). Indeed, DNASE1

appears to have evolved significantly later in comparison, even after DNASE1L1,

which potentially explains why it lacks some of the features of DNASE1L3. The

very high degree of homology between human and mouse DNASEs is also

notable, demonstrating that these enzymes are very strongly conserved between

species.

Figure 1: Phylogenetic tree of human and mouse deoxyribonucleases. Branch length is proportional to the number of amino acid substitutions per site. Generated with one-click analysis from www.phylogeny.fr: alignment by MUSCLE, curation by Gblocks, tree construction by PhyML, visualization by TreeDyn.

- 27 -

Technical challenges presented by DNASE1L3

Having observed the development of a lupus-like disease in mice deficient

for Dnase1l3, we sought to explore the mechanism by which DNASE1L3

contributes to the enforcement of tolerance against self-DNA. The first priority

was to establish a means of tracking DNASE1L3. Given the secreted nature of

DNASE1L3 (168), we anticipated assessment of DNASE1L3 to consist of

conventional methods, such western blotting, intracellular staining and flow

cytometry, and ELISA. Unfortunately, due to the extensive homology between

DNASE1L3 and DNASE1 (163), we were unable to identify a commercially

available antibody that functioned adequately in any of our experiments (data not

shown).

We then decided to investigate the possibility of generating a custom

antibody within the lab by immunizing our knockout animals. This approach

presented additional problems surrounding protein purification, however.

DNASE1L3 has a signal peptide at its N terminus (169), rendering a

hexahistidine (His) tag ineffective (Fig. 2). Furthermore and as we later confirmed,

the critical nature of the C terminal domain for the function of DNASE1L3 (167)

suggested that a His tag at the end of this region would dramatically affect its

stability and/or function. Lacking the tools to purify DNASE1L3 to the degree

required for antibody generation, an alternative approach was required.

- 28 -

Instead, we sought to develop a functional assay that would allow us to

track the activity of DNASE1L3. We opted to take advantage of the unique C

terminal domain of the enzyme. It had previously been noted that unlike DNASE1,

DNASE1L3 has the unique ability to block liposomally-mediated transfection

(167). This activity depended on the presence of the C terminal domain, as

truncated forms of DNASE1L3 lost the ability to digest liposomally coated DNA

(167). Thus, since coated DNA is only able to be processed by DNASE1L3, we

can use its digestion as a specific readout.

qPCR as a functional readout for the activity of DNASE1L3

We initially attempted to assess digestion of coated DNA by gel

electrophoresis, but were unable to do so due to the nullification of the negative

charge of DNA upon coating with positively-charged liposomal transfection

reagents such as N-[1-(2,3-Dioleoyloxy)propyl]-N,N,N-trimethylammonium

methyl-sulfate (DOTAP) or TransIT. Instead, we decided to use quantitative

polymerase chain reaction (qPCR) to determine the extent to which DNA had

been digested after incubation with enzyme. For each qPCR reaction, a standard

curve was generated using serial dilutions of the target plasmid. Importantly, 2

Figure 2: Representation of the key features of DNASE1L3. Asterisks denote vital catalytic residues; bracket denotes a disulfide bridge that is required for function. The N terminal signal pepdide is cleaved during secretion.

- 29 -

M of Mg2+ and Ca2+ were added to the reaction, since DNASE1L3 is dependent

on these ions (169). Notably, this technique was extremely sensitive, as plasmid

could be consistently detected at the femtogram level. Using the standard curve,

the amount of DNA remaining after each digestion was extrapolated and the

percent of input (i.e. 1 ng) was calculated (Fig. 3).

DNASE1L3 is uniquely able to digest liposome-coated and nucleosomal

DNA via its C terminus

To test our assay, we generated recombinant DNASE1L3 or DNASE1 via

transfection of HEK293 cells, followed by harvest of the supernatant. Importantly,

we used KnockOut serum replacement from Thermo Fisher in lieu of using fetal

bovine serum (FBS) in our cell cultures, as the latter contains abundant DNASE1

and DNASE1L3. We noted that recombinantly-generated DNASE1L3 was able to

digest coated plasmid, whereas DNASE1 or an empty vector control did not (Fig.

4A). In contrast, both DNASE1 and DNASE1L3 digested naked DNA with similar

efficiency. Furthermore, we generated a truncated form of DNASE1L3 lacking the

Figure 3: Standard curve for detection of DNASE1L3 activity by qPCR. Serial dilutions of template DNA were used to generate a standard curve. DNA remaining after digestion with various DNASE-containing supernatants or sera was assessed by qPCR and the standard used to calculate the percentage of DNA remaining. A unique standard curve was generated for each individual experiment.

- 30 -

positively charged C terminal domain, and noted that it also completely failed to

process coated DNA, but was still able to digest uncoated DNA (Fig. 4A).

Encouraged by our ability to replicate previously published results with

our assay, we next sought to test the mutated form of DNASE1L3 linked to

sporadic lupus. The SNP rs35677470 is a nonsynonymous mutation of C686 to

T686, and results in the substitution of a cysteine residue in place of an arginine

at position 206 of DNASE1L3 (159, 160). Notably, this substitution results in the

loss of a potentially crucial hydrogen bond (159). The R206C variant had

previously been reported to be inactive (179), so we decided to generate this

variant using site-directed mutagenesis to test its activity in our assay. We noted

that although the R206C variant is not completely bereft of activity, it was roughly

10 fold less active than wild type DNASE1L3 (Fig. 4A). Crucially, this reduced

activity was observed in the digestion of both coated and uncoated DNA,

suggesting that this variant is hypomorphic but can still target liposome-coated

DNA.

- 31 -

Finally, we were intrigued by previous studies of DNASE1L3 where it was

shown to digest nucleosomal DNA from isolated nuclei much more efficiently

than DNASE1 (169). Furthermore, DNASE1L3 was demonstrated to digest

chromatin independently of helper proteases, whereas DNASE1 required

supplementary proteolysis (168). Additional findings hypothesized a role for

DNASE1L3 in the processing of genomic DNA during cell death via both

apoptosis (170) and necrosis (172). To test the efficacy of our DNASE1L3

against chromatin, we designed qPCR primers specific for the endogenous

Figure 4: The C terminal domain of DNASE1L3 allows its activity to be specifically tracked as it confers the ability to digest coated DNA. A. Digestion of native or liposome-coated plasmid DNA by recombinant DNASEs as assessed by qPCR. Supernatants containing the indicated DNASE were incubated with uncoated or coated plasmid DNA for one hour at 37oC, and the DNA remaining after digestion assessed by qPCR using GFP-specific primers. Data are expressed as a percentage of input DNA (means +/- SD of three independent experiments). B. Digestion of purified human genomic DNA (gDNA) or purified human nucleosomes (nDNA). DNASEs were incubated with the DNA substrates for 10 minutes at 37oC, and the DNA remaining was assessed by qPCR using primers specific for the human Alu genomic repeat. Data are expressed as a percentage of input DNA (means +/- SD of three independent experiments).

- 32 -

retroelement Alu, of which roughly 106 copies are present in all primate genomes.

We then incubated recombinant DNASEs with commercially purchased human

nucleosomes (EpiCypher) and again performed qPCR to assess the extent to

which the DNA had been digested. We allowed the incubation to proceed for only

10 minutes, since at longer timepoints DNASE1 was able to digest nucleosomal

DNA (data not shown). Overall though, we noted that DNASE1L3 was much

more efficient than DNASE1 at digesting nucleosomal DNA, and that this activity

was dependent on the presence of the basic C terminal domain (Fig. 4B). In

addition, the hypomorphic R206C variant was again roughly 10 fold less efficient

in processing nucleosomal DNA (Fig. 4B). Notably, all recombinant DNASEs had

comparable activity against control genomic DNA (Fig. 4B), suggesting that the

activity of DNASE1L3 against nucleosomal DNA is a unique property and is not

due to differences in its production or release as opposed to other DNASEs.

Minor perturbations of the C terminus caused by mutations dramatically

affect the activity of DNASE1L3

Having established a means of testing DNASE1L3 activity specifically, we

once again decided to pursue the generation of a tagged form of the enzyme.

Initially, we began by placing a His tag immediately behind the signal peptide at

the N terminus (DNASE1L3 His-N). Although this variant could digest uncoated

DNA normally, it was unable to process coated DNA (data not shown). As there

is no crystal structure for DNASE1L3, we examined that of the well-characterized

DNASE1 (149, 165). The tertiary structure of DNASE1 is such that the N

- 33 -

terminus and C terminus are placed in close proximity, and the high degree of

homology between DNASE1L3 and DNASE1 (Fig. 5A) suggests that this is likely

the case for DNASE1L3 as well. Additionally, the key residues forming the

interface between DNASE1 and DNA are very closely replicated in DNASE1L3

(Fig. 5B). Thus, it seems likely that our His tag disrupts the C terminal domain,

preventing this form of DNASE1L3 from digesting coated DNA despite it being

active catalytically.

Given the challenges DNASE1L3 presented, we decided to collaborate

with structural biologists in order to more thoroughly understand the enzyme and

to understand the importance of the C terminus. Molecular modeling revealed

that the C terminus of DNASE1L3 forms an extremely rigid alpha helical structure

Figure 5: The catalytic domains of DNASE1 and DNASE1L3 share extensive homology. A. Global sequence alignment of bovine DNASE1 (PDB:2DNJ.a; top) and human DNASE1L3 (UniprotQ:Q13609). Red boxes represent amino acids of DNASE1 that contact DNA and orange lines represent selected residues contacting DNA. B. 3D structure of the complex of DNASE1 with DNA. Selected residues contacting DNA are colored orange.

- 34 -

when isolated (Fig. 6A) and also in the context of the entire protein (Fig. 6B). To

test the importance of positioning of the C terminus, we created versions of

DNASE1L3 with a His tag immediately preceding the C terminal domain (His-

preCT) and at the C terminus itself (His-CT). Modeling of these variants revealed

that although the alpha helical nature of the basic domain was preserved, the

tertiary structure was affected in the case of the His-preCT variant and the

domain was masked in the His-CT form. (Fig 6C). Indeed, analysis of the activity

of the His-CT DNASE1L3 revealed that while digestion of naked DNA was

preserved, it was unable to process coated DNA or nucleosomal DNA,

suggesting that the C terminus must be exposed to function (Fig. 6D). In contrast,

the His-preCT DNASE1L3 retained its activity against coated DNA but was

deficient in digestion of nucleosomal DNA (Fig. 6D). Overall, this suggests two

distinct roles for the C terminal domain: it both allows for the penetration of

DNASE1L3 through liposomes and, likely, lipid bilayers, and also allows it to

digest nucleosomal DNA.

- 35 -

Figure 6: Perturbations of the C terminal domain of DNASE1L3 dramatically affect its function. A. ab initio structure prediction of the C terminus of DNASE1L3 (aa 282-305). Shown is the 3D structure of the lowest energy conformation (left) and the energy spectrum of the folding simulation (right). Importantly, there is a roughly 5 kcal gap between the lowest energy conformation and the first non-helical conformation, strongly suggesting an extremely rigid helical structure. B. Homology model of DNASE1L3 based on PDB:2NDJ, with ab initio

structure prediction of the C terminus. The rigid -helical conformation of the C-terminal

domain is unperturbed in the context of the whole protein. C. Homology modeling of

hexahistidine-tagged versions of DNASE1L3. Although the -helical conformation of the C

terminus is not affected, it is shielded in the His-CT variant and its packing against the protein is altered in the His-preCT mutant. D. Digestion by His-tagged DNASE1L3 variants, as in Figures 7A and 7B. Data are presented as percentage of input digestion, and are means +/- SD of three independent experiments.

- 36 -

Dnase1l3-deficient mice are unable to process liposome-coated DNA

In addition to testing recombinant DNASE1L3, we wanted to validate our

assay using our Dnase1l3-deficient animals. We began by isolating serum from

wild type or knockout animals and assessing its capacity to digest coated or

uncoated DNA. Indeed, although knockout mouse serum could readily process

uncoated DNA, we observed a defect in the digestion of coated DNA (Fig. 7A).

To further confirm this phenotype, either uncoated or coated plasmid DNA was

injected intravenously into wild type or knockout mice, with serum being collected

3 hours post-injection and assessed by qPCR using GFP-specific primers. Again,

while both wild type and knockout animals cleared uncoated DNA, the knockout

animals were unable to process DNA after liposome coating (Fig. 7B).

Figure 7: Dnase1l3-deficient animals do not process liposome-coated DNA ex vivo or in vivo. A. Sera from Dnase1l3-deficient animals does not process coated plasmid DNA as assessed by qPCR. Although robust digestion of naked DNA was observed by sera of both WT and KO mice, loss of DNASE1L3 resulted in an inability to digest liposome-coated DNA. Data are presented as % of input DNA, displaying individual mice as well as the median value. B. Dnase1l3-deficient animals fail to process circulating liposome-coated plasmid DNA. WT or KO

mice were injected intravenously with 1 g of native or liposome-coated plasmid DNA, then bled

3 hours later. After isolation of serum, the presence of plasmid DNA was determined by qPCR for GFP. Individual mice are shown, with bars representing the median.

- 37 -

Circulating DNASE1L3 produced by hematopoietic cells prevents

autoimmunity

In order to test whether circulating DNASE1L3 was produced by

hematopoietic cells or a radioresistant cell type, we performed reciprocal bone

marrow (BM) transfers between WT and KO mice. We then tracked DNASE1L3

activity over time in these mice by making a slight modification to our qPCR

assay in order to compare small changes in

circulating DNASE1L3. Whereas previously the

standard curve was generated using serial

dilutions of DNA in order to assess the

percentage of DNA digested, we now diluted

wild type serum to different percentages and

incubated it with the same amount of coated

DNA as our test reactions for a shorter amount

of time (10 minutes vs. 1 hour). This allows for

smaller changes in the amount of DNASE1L3 to

be resolved. Using this technique, we observed

progressive loss of systemic DNASE1L3 in WT

mice that received KO bone marrow, and progressive gain in KO mice that

received WT bone marrow (Fig. 8). Intrigued at these results, we decided to set

up an additional set of long-term chimeras, wherein WT or KO bone marrow was

injected into lethally irradiated WT mice. Consistent with our previous results, we

observed progressive loss of DNASE1L3 activity in mice receiving KO bone

Figure 8: Active DNASE1L3 in circulation is produced by hematopoietic cells. Reciprocal bone marrow chimeras were set up and circulating DNASE1L3 activity tracked over time by qPCR. Data are expressed as percentage of WT activity and are shown as means +/- SD of 6-8 mice per group.

- 38 -

marrow (Fig. 9A). In the group receiving KO BM, circulating DNASE1L3 activity

was lost by 40 weeks post-transfer, which coincided exactly with the

development of anti-dsDNA auto-Ab (Fig. 9A). At the experimental endpoint, we

observed that these mice had eventually developed the hallmark symptoms of

SLE: ANA, kidney deposition of IgG, and glomerulonephritis (Fig. 9B-D). We

conclude that autoreactivity to self-DNA inversely correlates with circulating

DNASE1L3 produced by hematopoietic cells.

- 39 -

Figure 9: The production of autoantibodies in Dnase1l3-deficient mice is driven by hematopoietic cells. A-D: The development of autoantibodies in WT mice lethally irradiated and reconstituted with bone marrow from WT or KO mice. A. Recipient mice were simultaneously assessed for circulating DNASE1L3 activity (left panel) and for the presence of anti-dsDNA IgG by ELISA (right panel) post reconstitution. The left panel is presented as percentage of WT activity (means +/- SD of six animals per group). The right panel shows individual mice, with bars representing the median. B. ANA in the sera of recipient mice at the 55 week experimental endpoint, as in Figure 3A. Images are representative of six animals per group. C. Deposition of IgG in the kidneys of recipient mice, as in Figure 4G. Kidney sections stained for the presence of IgG by immunofluorescence are shown (left panels, representative of six animals per group), as well as the percentage of glomeruli with IgG deposition (right panel, out of 40 glomeruli per kidney, individual mice plus bars representing median). D. Assessment of kidney pathology in recipient mice, as in Figure 4I. Kidney sections were stained by H&E and histopathology scores were generated by a blinded pathologist. Shown is the percentage of the kidney cortex affected by inflammation (left panel, individual mice and median) and the median cumulative score of glomerulonephritis (right panel, individual mice and median).

- 40 -

In order to directly test the effects of circulating DNASE1L3 on the

development of autoreactivity, we injected young Dnase1l3 deficient mice with an

adenoviral vector encoding human DNASE1L3 (Ad-DNASE1L3). Adenoviral

injection in vivo results in the transduction of hepatocytes, allowing for the

production of protein for several weeks until cleared. Indeed, after four weeks we

observed total restoration of WT levels of DNASE1L3 in the knockout mice

injected with Ad-DNASE1L3 compared to a control adenovirus (Fig. 10A).

Crucially, we observed a significant delay in the development of anti-dsDNA IgG

in these mice compared to controls (Fig. 10B). Overall, these results suggest that

circulating DNASE1L3 produced by hematopoietic cells prevents the

development of autoimmunity by digesting an extracellular source of self-DNA.

Figure 10: Restoration of circulating DNASE1L3 into Dnase1l3-deficient animals delays the development of autoimmunity. Young 4-week-old Dnase1l3-deficient mice were injected with adenoviruses encoding human DNASE1L3 (Ad-DNASE1L3) or GFP (Ad-GFP). A. Serum DNASE1L3 activity after adenoviral administration at the indicated time points, along with age-matched WT controls. Data are presented as percentage of WT activity, and individual animals are shown with bars as the median. B. Serum titers of anti-dsDNA IgG in the same mice as measured by ELISA. Data are shown as the median +/- the range of four (Ad-GFP or WT) and nine (Ad-DNASE1L3) animals per group.

- 41 -

Identifying the cellular sources of circulating DNASE1L3

We then wanted to understand which cell type was responsible for the

production of DNASE1L3 in systemic circulation. Because our Dnase1l3-deficient

animals were generated through the use of a gene trap insertion of a LacZ

cassette, we were able to use the fluorogenic substrate fluorescein di-V-

galactoside (FDG) to quantify -galactosidase activity in cells expressing

Dnase1l3 by flow cytometry. We found that Dnase1l3 expression is largely

restricted to myeloid cells, in particular CD11c+ classical dendritic cells (cDCs)

and, to a lesser extent, macrophages (Fig 11A). Additionally, while it seems to

not be expressed broadly in lymphocytes, innate-like B cell populations such as

marginal zone (MZB) and CD5+ B-1a cells do have low levels of Dnase1l3 (Fig.

11A). Furthermore, this limited expression we observed was confirmed by

previously published microarray data in both humans (Fig. 11B) (180) and mice

(Fig. 11C) (181). In addition, we performed qPCR on sorted cell populations from

primary mouse splenocytes, and consistent with these previous results we noted

that classical DCs had the highest levels of Dnase1l3 expression (Fig. 11D).

- 42 -

Figure 11: DNASE1L3 is expressed primarily by dendritic cells and macrophages. A. Flow cytometry analysis of Dnase1l3 expression in immune cells. Splenocytes from Dnase1l3LacZ/LacZ KO or WT mice were stained for the presence of LacZ as a proxy for DNASE1L3 using the fluorogenic substrate FDG. Histograms of FDG in the indicated gated cell populations are shown. Data are representative of three independent experiments. B. The expression of DNASE1L3 in human tissues and cell types. The expression profile of DNASE1L3 (probe 205554_s_at) in the Primary Cell Atlas microarray expression database as visualized in the BioGPS browser (biogps.org) is shown, with relevant cell types indicated. C. The expression of Dnase1l3 in murine immune cell populations. The expression profile of Dnase1l3 in the Immgen microarray expression database of key populations (top panel) and monocytes/macrophages (bottom panel) are shown. D. The expression of Dnase1l3 in sorted murine splenocyte populations, as determined by qRT-PCR. Relative expression levels in the indicated cell types are shown, normalized to Actb (mean +/- SD of triplicate PCR reactions). Data are representative of five experiments.

- 43 -

We went on to utilize the alternative protocol to test DNASE1L3 activity in

our mice lacking specific populations to understand the extent to which each

population contributes to normal WT activity in circulation. In Rag1-deficient

animals that lack all lymphocytes, circulating DNASE1L3 was not reduced

compared to wild type controls, ruling out potential contributions from innate-like

B cells (Fig. 12A). However, upon transient depletion of cDCs and intestinal

macrophages via injection of diphtheria toxin (DTX) into CD11c-diphtheria toxin

receptor (DTR) mice, systemic DNASE1L3 levels were reduced by more than

75% (Fig. 12B). To rule out any effects of the transgene, we injected WT mice

with chlodronate liposomes, depleting tissue macrophages and reducing cDC

numbers. These mice lost roughly 50% of DNASE1L3 activity compared to mice

injected with control liposomes (Fig. 12C). And finally, depletion of intestinal

macrophages via a single injection of anti-macrophage colony stimulating factor

1 receptor (anti-Csf1r) blocking antibody resulted in a loss of roughly 15% of

systemic DNASE1L3 (Fig. 12D).

- 44 -

Conclusions

Overall, we have designed an assay allowing the tracking of systemic

DNASE1L3 activity based on its unique C terminal domain and resultant ability to

digest liposome-coated DNA. Our data emphasize the extent to which this C

terminal domain is crucial for the enzyme’s full activity, as the function of

DNASE1L3 is majorly impaired by small adjustments to the stability or angle of

the alpha helix. Importantly, this C terminal domain appears to have two roles for

DNASE1L3; it both allows for the penetration of lipid bilayers and also for the

digestion of nucleosomal DNA. We hypothesize that the former activity may be

Figure 12: DNASE1L3 in circulation is produced primarily by dendritic cells and macrophages. A. DNASE1L3 activity in the sera of Rag1-deficient mice (individual animals and median). B. DNASE1L3 activity in the sera of DC-depleted mice. Animals with Cre-inducible diphtheria toxin receptor (DTR) with or without DC-specific Cre deletion (Cd11c-Cre) were injected with diphtheria toxin (DTX) intraperitoneally every other day for 2 weeks, and their sera was analyzed for DNASE1L3 activity (individual animals and median). C. DNASE1L3 activity in the sera of WT animals treated with liposomes containing either PBS or chlodronate to facilitate macrophage depletion on the indicated days after treatment (individual animals and median). D. DNASE1L3 activity in the sera of WT animals 12 days after intraperitoneal injection of control IgG or anti-Csf1r blocking antibody (individual animals and median).

- 45 -

due to its extremely basic properties while the latter may be due to the very

stable and rigid structure this domain adopts, though further experiments are

necessary to confirm.

Furthermore, while dendritic cells and macrophages have been previously

described to play key roles in the maintenance of tolerance and prevention of

autoimmunity (182, 183), our data suggest a novel mechanism by which this

takes place: the secretion of a systemic enzyme that processes extracellular self-

DNA. Its affinity for both membrane-encapsulated DNA and for nucleosomal DNA

suggest that it processes some form of extracellular chromatin, preventing the

activation of autoreactive B cells. The inverse correlation between DNASE1L3

levels and anti-dsDNA autoAb indicates that DNASE1L3 is constantly required

for the prevention of autoimmunity against a ubiquitous source of potentially

antigenic self-DNA.

- 46 -

Chapter 2

Microparticles as the physiological target of DNASE1L3

Microparticles as potential targets of DNASE1L3

We next wanted to determine the physiological form of self-DNA targeted

by DNASE1L3. Taking into account the known properties of DNASE1L3 and our

initial results, we inferred several likely characteristics of its target. Firstly, the

self-DNA would probably be extracellular in nature, given the secreted nature of

DNASE1L3 and its detectable presence in systemic circulation. Also pointing to

an extracellular antigen is the dispensability of the intracellular DNA sensor

STING in driving autoimmunity in knockout animals and the subsequent lack of a

pronounced interferon (IFN) signature. Secondly, this DNA must be more or less

ubiquitous, due to the constant requirement for the presence of DNASE1L3 to

prevent autoimmunity. Thirdly, given the efficiency with which DNASE1L3

processes coated DNA and nucleosomal DNA, it is likely membrane associated

and chromatin based. Given these parameters, we hypothesized that apoptotic

microparticles might be a potential target of DNASE1L3. Although microparticles

have been primarily investigated in the context of thrombosis (184, 185) and

cancer (186, 187), recent work has linked them to autoimmune disorders such as

systemic sclerosis (188), rheumatoid arthritis (189), and other rheumatic

diseases such as vasculitis (190).

Also known as microvesicles, microparticles are released from cells during

both cell death and activation (191). They are small, membrane-bound vesicles

- 47 -

that are known to contain a variety of cytoplasmic and nuclear cargo, including

proteins, messenger RNAs, micro RNAs, and other signaling molecules (192).

Consequently, microparticles have been implicated in intercellular crosstalk and

regulation, both locally and systemically (190, 193, 194). Their biological

functions include modulation of the immune system, regulation of coagulation,

and transfer of signaling molecules (191, 195-198). Finally, on top of potential

roles in disease pathogenesis, microparticles can be used as biomarkers, since

they reflect the surface markers and intracellular contents of their parent cells

(191, 199-202). In addition, microparticles arising from apoptotic cells display

phosphatidylserine on their surfaces, allowing for the binding of annexin V (191).

The extent to which microparticle DNA is reflective of or represents total DNA in

human plasma, however, remains unclear (203, 204).

Links between microparticles and disease

Several recent studies have linked microparticles to systemic lupus

erythematosus (189, 205-210). A particularly interesting study showed that

although lupus patients did not have an increase in the numbers of microparticles,

microparticles isolated from their bloodstreams had higher levels of autoantibody

coating compared to controls (205, 207). In addition, it was shown that

microparticles can be bound by DNA-specific autoantibodies (210). Furthermore,

it has been established that apoptotic cells incorporate genomic DNA into

released microparticles (211). Finally, these microparticles have been shown to

- 48 -

expose chromatin on their surfaces (210, 212-214), thus representing a

potentially novel source of self-DNA for DNA-reactive B cells.

Previous therapeutic efforts focused on microparticles were conducted

using DNASE1, which upon its discovery was tenuously linked to SLE (144, 145,

215). Although no role for DNASE1 could ever be conclusively demonstrated, we

hypothesized that microparticles may instead represent a potentially novel target

of DNASE1L3 for several reasons. They are released extracellularly into

circulation and are stable enough to have systemic effects (193), are present

normally in the plasma of both healthy subjects and patients with SLE (205, 207,

216), and are loaded with chromatin incorporated into cell membranes (190, 210,

211). Given that our previous results suggested that the natural substrate of

DNASE1L3 was likely to be ubiquitously produced, present in the blood, and

chromatin based and/or membrane encapsulated (see Chapter 1), microparticles

thus met all criteria as a physiological target. To test the potential effects of

DNASE1L3 on microparticle DNA, we began by inducing the release of

microparticles from the Jurkat T cell leukemia line.

Induction and isolation of microparticles

Staurosporine (STS) is a molecule originally isolated from the bacterium

Streptomyces staurosporeus (217). STS functions as a competitive inhibitor of

ATP, as it binds with very high affinity to many protein kinases (218). Previous

studies have shown that Jurkat cells readily release microparticles upon STS

treatment (191). We sought to confirm these findings, and indeed observed rapid

- 49 -

release of microparticles by Jurkat cells treated with STS (Fig. 13A)

accompanied by comprehensive apoptosis (Fig. 13B). To quantify microparticles,

we employed a combination of rapid centrifugation and flow cytometry to

determine purity (Fig. 13C).

Figure 13: Treatment with staurosporine induces rapid cell death and microparticle release. A. Extensive production of microparticles from Jurkat cells after treatment with staurosporine. Jurkat cells were treated with 1 mM staurosporine (STS) for the indicated amount of time. Cultures were harvested and stained with Annexin V and propidium iodide (PI) to assess cell death. Cell death was accompanied by a concordant increase in microparticle release. B. Near total induction of apoptosis in Jurkat cells treated with staurosporine. Jurkat cells were stained with Annexin V and PI after treatment with DMSO control (top panels) or STS (bottom panels) for 24 hours. C. Preparation of microparticles yields significant enrichment. STS-treated Jurkat cell cultures were first centrifuged to pellet cells, and supernatants were then spun at high speeds to pellet microparticles. After resuspension in sterile PBS, microparticles are assessed for purity by flow cytometry.

- 50 -

DNASE1L3 is uniquely able to digest microparticle DNA

In order to test the digestion of human microparticle DNA, we designed

qPCR primers specific for the Alu retroelement. The Alu element is a primate-

specific short interspersed element (SINE) that comprises roughly 11% of the

human genetic code (219). Because more than 106 copies are dispersed

throughout the genome, PCR typing between adjacent copies has been widely

used both in forensic science and in the study of population genetics (220). In

lieu of using PCR to assess the distances between individual elements, we

designed primers to amplify Alu itself, reasoning that its abundance and wide

distribution meant it was likely to be incorporated into microparticles at a high

level.

We began by assessing the capacity of recombinantly generated DNASEs

to digest microparticle DNA by qPCR. As hypothesized, DNASE1L3 proved to be

extremely efficient at eliminating microparticle DNA (Fig. 14A). This was in stark

contrast to empty vector controls as well as DNASE1 (Fig. 14A), in concordance

with previous studies that demonstrated an inability of DNASE1 to digest

microparticle DNA (221). Wanting to understand whether this activity required the

C terminal basic domain of DNASE1L3, we also assessed the capacity of our

truncated DNASE1L3 to digest microparticle DNA. Confirming the importance of

the C terminus, this modified DNASE1L3 was completely deficient in its ability to

process microparticle DNA (Fig. 14A). Finally, we also assessed the variant of

DNASE1L3 associated with sporadic lupus, R206C. As with uncoated and

- 51 -

coated plasmid DNA, DNASE1L3-R206C was hypomorphic, processing

microparticle DNA with roughly 10-fold reduced efficiency (Fig. 14A).

One potential caveat was our reliance on microparticles produced from the

Jurkat cell line, which after extensive culture may not be reflective of cellular

behavior in physiological settings. We decided to validate our results using

microparticles produced ex vivo from mouse splenocytes. Because STS induces

apoptosis and microparticle release much more extensively in proliferating cells,

we stimulated the splenocytes with phorbol 12-myristate 13-acetate (PMA) and

ionomycin. These two compounds are widely used in combination to drive

proliferation, cytokine production, and cell activation via activation of protein

kinase C (PKC) (222, 223). After 72 hours of stimulation with PMA/ionomycin,

cells were treated with STS to induce microparticle release and microparticles

were harvested as previously described for Jurkat cells.

In order to measure digestion of mouse DNA by qPCR, we designed

primers specific for the B1 family of retroelements. B1 repetitive elements are

SINEs analogous to Alu in humans, and are present in similar abundance in the

mouse genome (224). Consequently, we could perform qPCR using these

primers and previously isolated mouse DNA to generate a standard curve. After

digestion with recombinant DNASEs, we observed that again only DNASE1L3

was capable of digesting DNA from splenocyte microparticles (Fig. 14B and data

not shown), suggesting that the activity of DNASE1L3 is not limited to

microparticles produced from a particular type of cell.

- 52 -

Finally, we also wanted to assess whether the sera of Dnase1l3-deficient

animals would be able to digest microparticle DNA, or if loss of DNASE1L3

resulted in a comparable phenotype to that observed for liposome-coated DNA.

Jurkat microparticles were isolated and incubated with sera from WT or

Dnase1l3-deficient mice. qPCR for the remaining human DNA revealed that sera

from KO animals was largely unable to process microparticle DNA, whereas sera

from WT animals efficiently digested this substrate extensively (Fig. 14C). These

data emphasize the importance of DNASE1L3 in the digestion of circulating

microparticle DNA, as other serum DNASEs had minimal effect on microparticle

chromatin.

In addition to our qPCR-based assay, we wanted to assess digestion of

microparticle DNA by an alternative readout. We ultimately opted to utilize flow

Figure 14: DNASE1L3 is uniquely able to process microparticle DNA. A. Digestion of Jurkat microparticle DNA by recombinant DNASEs, as measured by qPCR. Data presented as means +/- SD of three independent experiments. B. Digestion of mouse splenocyte MPs by recombinant DNASE1L3. Four independent experiments are shown. C. Digestion of Jurkat MP DNA by sera from WT or KO mice. Individual animals and median are shown.

- 53 -

cytometry in conjunction with a DNA-specific fluorescent dye to assess DNA

integrity. Previous studies had shown limited effectiveness of propidium iodide

(PI) in staining microparticles (221), likely due to their intact membranes. As an

alternative, we decided to use a membrane-permeable dye used primarily for

cell-cycle analysis, Vybrant DyeCycle Green (VG; ThermoFisher). We found that

this dye extensively stained microparticles produced from Jurkat cells at very low

concentrations (1:200,000) (Fig. 15). When microparticles were treated with

recombinant DNASE1L3, there was dramatically reduced VG staining compared

to microparticles treated with empty vector controls (Fig. 15). In stark contrast,

treatment with DNASE1 had minimal effect on the VG staining of microparticles

(Fig. 15). We also assessed the digestion of microparticles produced from

splenocytes by flow cytometry, and confirmed that DNASE1L3 but not DNASE1

dramatically reduces VG staining (data not shown).

- 54 -

The surface composition of microparticles is altered by DNASE1L3

We also wanted to test for the presence of high mobility group box protein

1 (HMGB1). HMGB1 is a non-structural nuclear protein that is known to

associate with nucleosomes, helping to facilitate transcription (225). However,

HMGB1 also functions as a key proinflammatory cytokine extracellularly. Upon

stimulation of macrophages with lipopolysaccharide (LPS), HMGB1 is released,

which activates NF-B via TLR4 (226). In addition to being produced by activated

cells, HMGB1 functions as an alarmin, as it is also commonly released during cell

death (227). HMGB1 has been described to be released concomitantly with

microparticles, and may indeed associate with them due to its affinity for

Figure 15: A membrane permeable DNA dye extensively stains microparticles only in the absence of DNASE1L3. Jurkat microparticles were isolated and treated with recombinant DNASEs, then stained with Vybrant Green membrane permeable DNA dye at a dilution of 1 to 200,000 for 20 minutes at room temperature. MPs were then analyzed on an Attune NxT flow cytometer. Data are representative of six independent experiments.

- 55 -

nucleosomal DNA (226-228). Furthermore, HMGB1 has been linked to the

pathogenesis of SLE as both a proinflammatory signal that contributes to cell

activation and as a potential target of autoantibody development (229-231).

Consequently, we wanted to investigate whether DNASE1L3 was capable of

reducing HMGB1 levels on microparticles through the processing of nucleosomal

DNA using flow cytometry. Upon digestion of microparticle DNA by DNASE1L3,

there was comprehensive loss of HMGB1 binding to their surfaces (Fig. 16A).

Furthermore, we noted that HMGB1 staining was restricted to DNA-positive MPs

as labeled by Vybrant green, and we further confirmed that DNASE1L3 has

dramatic effects on microparticle DNA (Fig. 16B). We thus conclude that

DNASE1L3 processes microparticle DNA in vitro, likely reducing their

immunogenicity and surface composition.

Figure 16: DNASE1L3 treatment of microparticles alters their surface composition. A. Jurkat MPs were treated with rDNASEs and then stained for the presence of HMGB1. Mean fluorescence intensity is indicated. B. Costaining of pretreated Jurkat MPs with anti-HMGB1 and Vybrant green. Notably, there are very few HMGB1 single positive MPs, indicated strong associations between HMGB1 and MP DNA.

- 56 -

Loss of DNASE1L3 results in a failure to process endogenous circulating

microparticles in vivo

We next sought to assess potential changes to endogenous microparticles

present in DNASE1L3 knockout animals. Previous work in the lab had examined

circulating DNA in the sera of knockout versus wild type mice using the

fluorescent DNA-intercalating dye PicoGreen. We had observed no differences in

DNA in the sera between wild type and knockout animals (Fig. 17A). In light of

our findings regarding microparticles, however, we decided to return to these

experiments. Importantly, the process by which sera is collected from mice

involves extended centrifugation at 22,000 x g, meaning that our serum samples

were likely free of microparticles. Furthermore, during serum collection the

coagulation process is allowed to proceed, which could result in the release of

DNA not normally present in steady state circulation.

Thus, in lieu of sera, we decided to isolate plasma using less rapid

centrifugation. In order to collect plasma, we required an anticoagulating

substance. Because we intended to perform qPCR on our isolates, we could not

use the most common anticoagulant, ethylenediaminetetraacetic acid (EDTA).

Instead, we used commercially purchased heparin to isolate plasma from WT

and Dnase1l3-deficient animals. When whole plasma was assessed by qPCR,

we could now observe a significant difference between DNA present in the

plasma of wild type and knockout animals (Fig. 17B).

- 57 -

We then wanted to assess the endogenous microparticles of these mice

more specifically. To that end, we isolated plasma as above, and after isolation

spun the plasma at 22,000 x g for 30 minutes to pellet microparticles. These

were then resuspended and stained with anti-CD41 and anti-Ter119, which are

specific for platelets and red blood cells, respectively. This was necessary

because platelets and debris from erythrocytes are approximately the same size

as microparticles and are also pelleted by these centrifugation steps. The

negative fraction was then counted on an Accuri tabletop flow cytometer in order

to precisely determine the numbers of relevant apoptotic microparticles (Fig.

18A). Notably, although microparticle count varied considerably from mouse to

mouse, we observed no difference between wild type and knockout animals (Fig.

18B). We then performed qPCR using B1-specific primers on this material and in

conjunction with the flow cytometry data calculated the amount of DNA per

microparticle. We noted that microparticles isolated from knockout animals

contained on average four orders of magnitude more DNA compared to those

from wild type animals (Fig. 18C). To further show the specificity of the defect to

Figure 17: Dnase1l3-deficient mice have higher levels of DNA circulating in plasma. A. No difference in the amount of DNA in the sera of WT and KO mice, as determined by staining with the DNA intercalator PicoGreen. Individual mice and median are shown. B. qPCR on whole murine plasma from WT or KO animals using primers specific for the B1 genomic repeat. Individual mice and median are shown.

- 58 -

microparticle DNA, we also performed qPCR on the non-microparticle fraction of

plasma (i.e. the supernatant after the rapid spin to pellet microparticles). We

noted that there was virtually no DNA present in plasma outside of microparticles,

and indeed were unable to detect DNA reliably despite our qPCR amplifying

standards at the femtogram level (data not shown).

Finally, we also decided to inject exogenous microparticles generated

from Jurkat cells into wild type and knockout animals to assess their ability to

cope with a sudden burden of microparticle DNA. 6 x 106 Jurkat microparticles

were injected intravenously into 6 wild type and 6 knockout animals, and the

mice were then bled 3 hours later. After isolation of plasma, qPCR using Alu-

specific primers was then performed. We again observed a defect in knockout

Figure 18: Microparticles from Dnase1l3-deficient mice carry increased amounts of DNA compared to those of WT mice. A. Identification of the relevant MP fraction from the plasma of mice. Only events within the MP gate that were negative for CD41 and Ter119 were counted as bona fide MPs. No differences between WT and KO mice were observed. Representative of nine mice per group. B. No difference in the total numbers of MPs isolated from WT or KO mice was observed. Individual animals and median are shown. C. Increased DNA in MPs isolated from KO mice compared to WTs. qPCR was performed on isolated MPs using B1 primers, and the total amount of DNA in the reaction was then divided by the number of MPs included. Individual mice and median are shown.

- 59 -

animals in the processing of exogenous microparticles, noting a significant

increase in microparticle DNA equivalents per L of plasma (Fig. 19). Knockout

animals were hugely deficient in clearing microparticle DNA, as they made little

impact on overall amounts for 24 hours (data not shown). We thus conclude that

DNASE1L3 processes DNA in circulating microparticles in vivo.

Conclusions

These data suggest that DNASE1L3 is responsible for the processing of

nucleosomal DNA contained within circulating microparticles. Our initial results in

Dnase1l3-deficient animals suggested that its endogenous substrate was likely to

be nucleosomal and/or membrane associated in nature. Microparticles represent

an underappreciated source of extracellular genomic DNA (192, 202, 211, 221),

and by definition have intact plasma membranes (221). Because chromatin is

potentially immunogenic, its persistence in circulation on microparticles may

compromise tolerance to self-DNA. DNASE1L3 rapidly processes the DNA load

of microparticles, as we have demonstrated by two independent readouts.

Figure 19: Dnase1l3-deficient mice fail to process exogenous microparticles in vivo. 6 x 106 Jurkat microparticles were injected intravenously into WT or KO mice, and plasma samples taken 3 hours post administration. qPCR for human Alu repeats was then performed. Individual mice and median are shown.

- 60 -

Furthermore, treatment of microparticles with DNASE1L3 reduces the binding of

HMGB1 to their surfaces, which represents a potent proinflammatory molecule

that has been extensively linked to the pathogenesis of SLE (225, 229-232). This

finding thus demonstrates that DNASE1L3 can fundamentally affect the profile of

circulating microparticles even beyond their DNA content.

The sheer scale of granulocyte turnover alone (>109 per kg per day) (233)

ensures that microparticles are ubiquitously present throughout circulation at the

steady state. Thus, the previously observed requirement for the presence of

DNASE1L3 in order to prevent the development of anti-dsDNA autoreactivity is in

keeping with the context of continuous microparticle release. Indeed, a recent

paper that focused on deep sequencing of cell-free DNA from human plasma

found that the majority is hematopoietic in origin, with an especially large amount

coming from granulocytes (203). We believe that the vast majority of DNA

analyzed by these authors originated from microparticles, since they used whole

plasma and because we found virtually no DNA in the microparticle-free plasma

fraction in mice.

Overall our results indicate that microparticles represent a novel source of

antigenic self-DNA, in particular in the absence of DNASE1L3 where it is allowed

to persist in circulation for extended periods of time. DNASE1L3 is uniquely

capable of processing microparticle DNA due to its basic C terminal domain.

DNASE1 in contrast had minimal impact on microparticle DNA, confirming

previous studies (191) and again emphasizing the importance of DNASE1L3 as

opposed to DNASE1 in circulation. Importantly, our data demonstrate that there

- 61 -

is no defect in microparticle release or production, but rather that produced

microparticles are not rendered immunologically benign in circulation.

Additionally, the production of microparticles as a consequence of cell death may

be relevant for human disease, which typically features flares that can be

triggered by a huge variety of stimuli, including but not limited to exposure to UV

radiation, viral infection, pregnancy, and physical injury (1, 12, 234-236). Even in

individuals with no defect in DNASE1L3, overwhelming the system with a large

quantity of DNA-loaded microparticles may allow for increased persistence and

the opportunity for microparticles to trigger anti-DNA autoimmunity.

- 62 -

Chapter 3

Microparticles as autoantigens

Roles for microparticles in human disease

Having established that microparticles are not processed fully in the

absence of DNASE1L3, we next wanted to assess their potential as antigen-

loaded targets of autoimmune responses. While the rapid development of

autoantibodies against dsDNA and, in particular, chromatin in our knockout

animals suggest a loss of tolerance to these molecules that may be caused by

their increased presence in microparticles, we sought to determine whether

microparticles themselves represent suitable B cell antigens. Understanding

whether microparticle-driven autoimmunity is based on direct interaction with B

cells is potentially key for further elucidation of the mechanisms underlying

pathogenesis. Importantly, DNA may behave as an unconventional antigen due

to its ability to stimulate the production of cytokines via TLR activation in addition

to its intrinsic immunological properties.

Numerous studies have offered potential explanations for the systemic

effects attributed to microparticles, including but not limited to priming of dendritic

cell and neutrophil subsets (216), metalloprotease and cytokine production by

fibroblasts (196), induction of soluble selectin release (188), crosslinking of

receptors on endothelial cells (198), interactions with galectin-3 binding protein

(206), activation of interferon production by plasmacytoid DCs (237), and direct

delivery of cytokines or other signaling molecules (185, 193, 208). Alternatively,

- 63 -

microparticles have been proposed to be targets of autoantibodies, directly

contributing to the formation of immune complexes in rheumatic diseases such

as SLE (192, 205, 207, 209, 210, 214, 238). Furthermore, other mouse models of

SLE, including the MRL-lpr/lpr spontaneous lymphoproliferative mutants as well

as the NZB/NZW F1 generation present with anti-microparticle autoantibodies

(209, 214). Given the rapid development of anti-dsDNA and anti-chromatin

autoantibodies in our mice and the strong association between anti-DNA

autoantibodies and disease severity in lupus patients (56, 74, 83, 89), we wanted

to investigate the potential binding of autoantibodies to microparticles and the

effects of DNASE1L3 on the repertoire of antigens present on microparticle

surfaces.

Microparticles are bound by circulating IgG from Dnase1l3-deficint animals

We began by investigating the extent to which microparticles were bound

by autoreactive antibodies from Dnase1l3-deficient mice compared to controls

using flow cytometry. Microparticles isolated from Jurkat cells were incubated

with sera from wild type or knockout mice, followed by a fluorescent secondary

antibody. We observed dramatically increased binding of the sera of knockout

animals compared to that of controls to microparticles (Fig. 20A), suggesting that

microparticles represent viable B cell autoantigens. In addition, we tested sera

from mice at different ages to establish whether binding to microparticles

worsened during lupus pathogenesis. Strikingly, we observed clear increases in

microparticle binding over time in knockout mice but not in wild type animals (Fig.

- 64 -

20B), indicating that microparticles are indeed bound by pathogenic

autoantibodies that increase in prevalence over time and correlate with disease

severity in Dnase1l3-deficient mice.

We next wanted to investigate the potential effects of treatment with

DNASE1L3 on autoantibody binding to microparticles. To this end, we treated

Jurkat microparticles with an empty vector control, recombinant DNASE1, or

recombinant DNASE1L3. From there, we stained each group of microparticles

with the sera from five different knockout mice at 20 weeks of age. While

DNASE1 had minimal effects on the binding of mouse sera to autoantibodies

compared to the empty vector control, treatment with DNASE1L3 completely

ablated all binding (Fig. 20C). Overall these results demonstrate that

microparticles represent viable B cell antigens, and that treatment of

microparticles with DNASE1L3 dramatically reduces their immunogenicity.

- 65 -

Circulating microparticles from knockout animals are coated with IgG

Next, we investigated the coating of endogenous circulating microparticles

by antibodies in wild type or knockout mice. As previously, we exsanguinated

young mice at 10 weeks of age into heparin in order to isolate plasma. We found

Figure 20: DNASE1L3-sensitive chromatin on the surface of microparticles is antigenic in Dnase1l3-deficient mice. A-C: Binding of circulating murine IgG to the surface of human apoptotic microparticles. Jurkat MPs were incubated with sera from Dnase1l3 KO or control WT animals, followed by secondary anti-mouse IgG fluorescent antibody. A. Representative histograms of mouse IgG fluorescence on the surface of Jurkat MPs. Percentage of positive MPs is indicated. B. Percentage of mouse IgG positive Jurkat MPs bound by WT or KO mice at the indicated time points. Data are presented as median +/- range of five animals per group. C. Jurkat microparticles were incubated with supernatants containing recombinant human DNASE1L3, DNASE1, or an empty vector control prior to staining with serum from Dnase1l3-deficient animals. Representative histograms are shown in the left panel, with the percentage of IgG positive MPs indicated. The right panel shows individual mice, with the bars representing the median.

- 66 -

that even at 10 weeks of age, knockout animals had more extensive binding of

IgG to endogenous microparticle surfaces (Fig. 21A). Additionally, we stained

these endogenous microparticles with PR1-3, a monoclonal antibody specific for

anti-DNA/histone 2a/2b complexes (210). We noted significantly higher levels of

PR1-3 binding to microparticles isolated from knockout mice compared to those

from wild type mice (Fig. 21B), suggesting that in addition to being more coated,

endogenous microparticles in young Dnase1l3-deficient animals have higher

levels of exposed chromatin and are thus more immunogenic.

Figure 21: Microparticles from Dnase1l3-deficient mice at young ages are more coated by IgG and have higher levels of exposed self-antigens. A. Staining of MPs isolated from mouse plasma from WT and KO mice for mouse IgG. The left panel shows representative histograms with the MFI indicated, and the right shows MFIs of individual mice with bars representing the median. B. MPs isolated from young WT or KO mice were stained using the anti-DNA/histone antibody PR1-3. Representative histograms are shown in the left panel with the percentage of PR1-3 positive MPs indicated, and the right shows percentage of PR1-3 positive MPs from individual mice with bars representing the median.

- 67 -

Microparticles drive loss of tolerance to self-DNA in inflamed conditions

To further examine the ability of microparticles to drive anti-nucleic acid

autoimmunity, we decided to inject them into mice and assess the subsequent

autoantibody levels. Our initial experiments used microparticles isolated from

Jurkat cells, but unfortunately these elicited massive germinal center and

inflammatory responses in all settings. In retrospect, this is unsurprising due to

the fact that Jurkat cells constitutively release murine gammaretroviruses picked

up during extensive passaging in culture (239) in addition to the presence of

potent cross-species and anti-tumor host responses. Thus, we instead opted to

produce syngeneic microparticles from mouse splenocytes treated with PMA and

ionomycin, as previously (see Chapter 2). In order to assess de novo anti-DNA

autoimmunity, we could not use Dnase1l3-deficient animals, as they have

autoantibodies from extremely early ages. Instead, we injected wild type mice

with an adenovirus encoding IFN-5, delivery of which has been shown to

accelerate the development of SLE in experimental models (175) as well as in

Dnase1l3-deficient mice (173). These mice rapidly presented elevated levels of

both IFN and the IFN-inducible gene Sca-1, confirming adenoviral efficacy (data

not shown). Microparticle injections into these mice but not naïve mice produced

very high titers of anti-nucleosome IgG (Fig. 22). These data demonstrate that

MPs can function as antigens that elicit chromatin-specific B cell autoimmunity in

the context of elevated interferon.

- 68 -

DNASE1L3-sensitive chromatin on microparticles is accessible to

autoantibodies

We next wanted to establish whether microparticles were bound by a

variety of anti-chromatin monoclonal antibodies, and whether this binding was

affected by treatment of microparticles with DNASE1L3. To this end, we obtained

four hybridomas from our collaborator Dr. Mark Shlomchik that produce

antibodies specific for different DNA/histone complexes implicated in SLE: PR1-3

(DNA/histone 2a/2b), PR2-8 (DNA/histone 2a/2b, DNA/histone 3/4), PR9-11

(naked DNA), and 29A-3H9 (isolated from 3H9 transgenic mouse that presents

with very high anti-dsDNA and anti-ssDNA).

To test binding of these anti-chromatin antibodies and the potential effects

of DNASE1L3, Jurkat microparticles were treated with recombinant DNASEs

Figure 22: Immunization with microparticles drives anti-chromatin antibody responses in inflammatory contexts. WT

mice were administered IFN5

adenovirus (IFN), MPs from syngeneic apoptotic splenocytes (MP), or both. Serum titers of anti-nucleosome IgG was measured one week post final injection by ELISA. Titers from individual animals and the median are shown.

- 69 -

prior to staining with our hybridoma antibodies. We noted that all antibodies

bound to microparticles treated with the empty vector control extensively, further

confirming that microparticles represent viable antigens for anti-chromatin B cell

responses (Fig. 23). Furthermore, treatment with DNASE1 had a minimal impact

on autoantibody binding (Fig. 23). In contrast, treatment with DNASE1L3

significantly reducing binding of all autoantibodies, providing further evidence that

DNASE1L3 is key for modulating the immunogenicity of microparticles, limiting

the avidity of anti-chromatin immune responses (Fig. 23).

Figure 23: Treatment of microparticles with DNASE1L3 but not DNASE1 prevents binding by DNA-specific autoantibodies. Microparticles isolated from Jurkat cells were treated with supernatants containing recombinant DNASE1, DNASE1L3, or an empty vector control. They were then stained with murine antibodies against DNA and/or chromatin originally isolated from mouse SLE models, followed by a fluorescent anti-mouse secondary. Shown are representative histograms from four experiments, with the percent of IgG positive MPs indicated.

- 70 -

The C terminus of DNASE1L3 is required to prevent autoantibody binding

to microparticles

We also wanted to test the extent to which the C terminus of DNASE1L3

is required for this activity, as well as the extent to which the DNASE1L3-R206C

hypomorphic form retains its function. To that end, we again isolated Jurkat

microparticles, and treated them with a vector control, DNASE1, DNASE1L3,

truncated DNASE1L3, or DNASE1L3-R206C. From there, we stained these

treated microparticles with purified PR1-3, which was provided by our

collaborator Dr. Keith Elkon. Washing and staining with our secondary antibody

was unchanged from previous experiments. Strikingly, we observed that

truncated DNASE1L3 had a minimal effect on PR1-3 binding to microparticles, in

stark contrast to wild type DNASE1L3 (Fig. 24A). In contrast, DNASE1L3-R206C

was capable of preventing PR1-3 binding to similar levels, but given that our

previous qPCR experiments suggested that it still retained considerable activity

this was not unanticipated (Fig. 24A). We also went on to test the activity of our

self-generated His-tagged DNASE1L3 variants. Although we had previously seen

that the his-preCT variant, which has a his tag immediately preceding the C

terminal -helix retained its activity against coated DNA. When tested for its

ability to process microparticle DNA, however, we found that this variant was

unable to reduce binding of PR1-3 (Fig. 24B). The his-CT variant of DNASE1L3,

which has a His tag at the end of the C terminus, was previously unable to

process coated DNA and was similarly deficient here (Fig. 24B). These data

demonstrate that in the absence of DNASE1L3, microparticles are extensively

- 71 -

bound by anti-chromatin autoantibodies and that incubation of microparticles with

DNASE1L3 is sufficient to render them non-immunogenic. In addition, the spatial

integrity of the C terminal domain is vital for this function of DNASE1L3, as even

slight perturbations of the enzyme’s tertiary structure can dramatically impact its

ability to process microparticles.

Microparticles are bound by prototypical autoantibodies from human SLE

Finally, we wanted to assess the relevance of microparticles in human

disease. Human antibodies that are positive for the 9G4 idiotype are greatly

Figure 24: The C terminal domain of DNASE1L3 is required for the processing of microparticle chromatin. A. Microparticles from Jurkat cells were incubated with recombinant DNASEs and subsequently stained with the purified anti-nucleosome antibody PR1-3 followed by a fluorescent secondary antibody. Histograms are shown with the percentage of IgG positive MPs indicated. Data are representative of three experiments. B. As in A, Jurkat MPs were treated with the indicated mutant forms of DNASE1L3 and then stained with PR1-3. Histograms are shown with the percentage of IgG positive MPs indicated. Data are representative of three experiments.

- 72 -

enriched among SLE-associated autoantibody repertoires (240, 241). They have

been shown to bind to a variety of antigens, including DNA, nuclear components,

and apoptotic cell nuclei (240, 241). We obtained a panel of 22 chimeric

antibodies consisting of Fab fragments isolated from patients fused with the

mouse IgG Fc region from our collaborator, Dr. Iñaki Sanz. We wanted to assess

binding of these antibodies to Jurkat microparticles that had been treated with an

empty vector or with DNASE1L3 by flow cytometry, as we had done previously.

Given the minimal effects we had observed of DNASE1 on antibody binding and

the limited quantities of primary antibody available, we elected to omit it from our

analysis. We observed three different binding patterns: some antibodies did not

bind to MPs at all, some bound to all MPs, and some bound to MPs treated with

the vector control only (Fig. 25A). Of our panel of 22 antibodies, 9 bound to

microparticles treated with the vector control, and of these, 6 then failed to bind

to microparticles after DNASE1L3 treatment (Fig. 25B). These DNASE1L3-

sensitive antibodies were reactive against apoptotic cell membranes, and were

positive for anti-nuclear antigen staining (241). These findings suggest that

microparticles may represent novel sources of antigen in sporadic SLE in

addition to rare cases where there is DNASE1L3 deficiency.

- 73 -

Conclusions

Overall these results demonstrate that microparticles represent sources of

self-antigen that are accessible to B cells. Furthermore, data from our knockout

mice suggest that antibody responses to microparticles are enhanced over time,

indicating that they play a role in disease pathogenesis. Additional evidence for

this stems from the observation of increased coating of IgG of endogenous

microparticles from knockout animals compared to wild type controls. On top of

that, our experiments injecting microparticles into permissive mice show that they

also likely play a role in the loss of tolerance observed in SLE. We hypothesize

that failure to process microparticle DNA represents the primary insult driving the

development of anti-dsDNA autoimmunity in Dnase1l3-deficient mice.

Identifying the specific form or forms of DNA recognized by B cells during

the pathogenesis of SLE remains a major challenge of the field. While multiple

Figure 25: A subset of human 9G4+ antibodies bind to microparticles only in the absence of DNASE1L3. A. Jurkat MPs were treated with an empty vector control or with DNASE1L3, and were then incubated with 9G4+ mAbs. Three patterns were observed: no binding to any MPs, binding to all MPs, and binding to MPs treated with vector control only. Representative histograms of each group are shown from left to right, respectively. Data are reflective of three independent experiments. MFIs of IgG bound to MPs are indicated. B. Classification of the 22 different 9G4+ clones tested based on binding pattern.

- 74 -

types of DNA have been linked to the pathogenesis of SLE, including neutrophil

extracellular traps, self-DNA complexed with anti-microbial peptides, and

oxidized mitochondrial DNA (104-107, 242), their potential as B cell antigens

remains unresolved. In contrast, our results demonstrate that microparticles

enriched in genomic DNA in the absence of DNASE1L3 are extensively bound by

autoreactive antibodies, many of which have been implicated in the pathogenesis

of SLE (210, 214, 240, 241). Furthermore, the reductions in autoantibody binding

observed specifically after treatment of microparticles with DNASE1L3 but not

DNASE1 indicate a fundamental role for DNASE1L3 in preventing the

development of autoimmunity.

Importantly, our data emphasize that the defect in DNASE1L3-deficient

animals stems from a failure to process microparticles rather than an issue in

their production. After digestion with DNASE1L3, microparticles appear to be

rendered inert with respect to anti-DNA and anti-chromatin immune responses,

though the presence of additional autoantigens on their surface is also likely as

evidenced by the binding of three 9G4+ antibodies to DNASE1L3-treated

microparticles. Furthermore, our results using 9G4+ antibodies indicate a

potential role for microparticles in sporadic lupus. Although it seems unlikely that

there may be overt defects in microparticle DNA processing at the steady state,

microparticles may represent novel sources of self-antigen in sporadic patients.

Finally, given the widespread links between cell death and/or stress and the

pathogenesis of SLE (235, 236), the resulting temporary increase in microparticle

burden may contribute to disease and/or flare onset.

- 75 -

Overall we believe our Dnase1l3-deficient mouse represents a novel,

comprehensive, monogenic model of SLE. This lupus is driven by a failure to

process the DNA contained within apoptotic microparticles in circulation, which

are normally produced primarily from granulocytes. Chromatin-containing

microparticles function as potent self-antigens for autoreactive B cells, and the

unique properties of DNASE1L3 allow it to process this chromatin and render

microparticles to be nonimmunogenic. Finally, we believe that our model

describes a novel cell-extrinsic mechanism of self-tolerance to DNA mediated by

DCs and macrophages that may allow for therapeutic intervention.

- 76 -

Chapter 4

DNASE1L3 and microparticles in human lupus patients

Circulating cell-free DNA in humans is contained within microparticles

Given our comprehensive studies of Dnase1l3-deficient mice, our findings

that a subset of 9G4+ antibodies are reactive against microparticles in a

DNASE1L3-sensitive manner, and the DNASE1L3-sensitive nature of

microparticle chromatin, we believed that microparticles may represent

unappreciated sources of antigen in human disease. In particular, recent work

has suggested that loss of tolerance in human SLE is initially to chromatin, with

subsequent epitope spreading to DNA specifically and other nuclear

autoantigens (152, 243). Furthermore, it is increasingly accepted among the field

that DNA-specific autoantibodies are among the most pathogenic variants, and

their prevalence has been strongly linked to poorer outcomes for patients (56, 74,

83, 89). Because DNASE1L3 is able to dramatically reduce the efficiency with

which DNA and chromatin-specific antibodies bind to microparticles in other

contexts, we wanted to directly investigate human patients.

Thanks to a collaboration with the Department of Rheumatology at NYU

and Dr. Jill P. Buyon and Dr. Robert M. Clancy, we were able to obtain samples

from a large cohort of patients with SLE. We were initially interested in assessing

the surface markers present on microparticles to determine the cells from which

they are primarily derived. These studies were far more feasible to perform in

human settings compared to mice simply due to the major differences in

- 77 -

obtainable plasma volume. A recent study of cell-free DNA in human circulation

focused on deep sequencing as a means of determining nucleosome positioning

and, hence, cell type of origin (203). They concluded that the vast majority of cell-

free DNA comes from hematopoietic cells, with a particularly large amount

stemming from granulocytes (203). This was in line with a previous study where

the authors examined a cohort of patients that received sex-mismatched bone

marrow transplants, then examined cell-free DNA (244). They noted that the

DNA present in plasma was predominantly of donor origin, demonstrating that

hematopoietic cells contribute the majority of cell-free DNA (244).

Circulating microparticles in humans are primarily of hematopoietic origin

We performed flow cytometry on microparticles isolated from human

plasma, and observed positive staining of CD45 and CD66 on Annexin V positive

microparticles, confirming that circulating microparticles originate largely from

granulocytes (Fig. 26A). In addition, we wanted to determine the extent to which

microparticle DNA comprises total plasma cell-free DNA. To assess this, after

rapid plasma centrifugation of human samples, qPCR was run in parallel

between the pelleted microparticles, the remaining soluble fraction, and

unfractionated plasma. We found that there was virtually no DNA present in the

non-microparticle fraction of plasma, and correspondingly microparticle DNA

made up nearly all detectable DNA in total plasma (Fig. 26B). We thus conclude

that microparticles released from granulocytes and other hematopoietic cells

- 78 -

make up the bulk of circulating DNA, and therefore are likely an important source

of self-antigen in humans.

Mutations in DNASE1L3 result in defective digestion of coated and

nucleosomal DNA in humans

Via our collaboration with Drs. Buyon and Clancy, we were extremely

fortunate to obtain plasma samples from DNASE1L3-deficient human patients

previously characterized with hypocomplementemic urticarial vasculitis syndrome

(HUVS) (151). Patient 1 had presented with SLE in addition to HUVS but is now

in remission, while patient 2 presented with HUVS only. Thus, neither patient was

affected by active SLE at the time of blood sampling, ruling out potential

Figure 26: Analysis of human microparticles. A. Phenotype of circulating microparticles in human plasma. MPs were isolated and stained for CD42b and CD235a to eliminate platelets and RBCs. We defined MPs as being FSClo SSClo CD42bneg CD235aneg. Representative stainings are shown comparing Annexin V versus the endothelial marker CD31, the leukocyte marker CD45, and the granulocyte marker CD66b. B. Distribution of human genomic DNA between the microparticle fraction and the soluble fraction of human plasma. DNA was measured by qPCR in unfractioned plasma, MPs, and the soluble

fraction. DNA amounts per L

of the original unfractioned volume are shown in samples from SLE patients (n=8) or healthy controls (n=2).

- 79 -

secondary effects of the disease. We also obtained blood samples from their

haplodeficient parents, neither of whom ever manifested with effects of SLE nor

HUVS (151).

We began by assessing the activity of DNASE1L3 in the plasma of these

patients. Using the digestion of liposome-coated plasmid DNA, we confirmed the

total lack of DNASE1L3 activity in the plasma of these patients, while their

haplodeficient parents had roughly 50% activity (Fig. 27A). In addition, we

analyzed 3 patients who were heterozygous for the hypomorphic R206C variant

of DNASE1L3, and observed that these patients had roughly 60% of DNASE1L3

activity (Fig. 27A), indicating that this variant is approximately 5 fold less active

than the wild-type form in vivo.

In addition, we assessed the extent to which patient plasma could digest

commercially-purchased nucleosomal DNA over short incubations, as we

previously showed in mice (see Chapter 1). As in Dnase1l3-deficient mice,

DNASE1L3-null human patients were totally unable to process human

nucleosomal DNA compared to normal controls (Fig. 27B). Additionally, while the

haplodeficient parents could digest nucleosomes to a greater extent than the

patients, they were also broadly deficient compared to controls (Fig. 27B). Finally,

we determined whether null and haplodeficient patients were capable of

processing microparticle DNA. We induced and isolated microparticles from

Jurkat cells, then incubated them with plasma from these null patients. As with

digestion of nucleosomal DNA, null patients were greatly deficient in their

- 80 -

capacity to digest microparticle DNA, while the haplodeficient parents were

partially deficient (Fig. 27C).

Patients lacking DNASE1L3 present with increased microparticle DNA

Having confirmed that these patients were indeed deficient for DNASE1L3

and that our assay developed in mice could be applied to these rare human

patients, we wanted to analyze the DNA content of circulating microparticles ex

vivo. As we were working with human plasma, we utilized the same protocol

optimized in mice in order to pellet microparticles from plasma. These pellets

Figure 27: DNASE1L3 null mutations in human patients result in defective DNA processing that parallels Dnase1l3-deficient mice. A-C. Plasma from 2 DNASE1L3-deficient patients, their haplodeficient parents, individuals with the hypomorphic R206C polymorphism (n=3), and normal control subjects (n=2-5) were analyzed. A. DNASE1L3 activity in patient plasma, measured by qPCR of DOTAP-coated plasmid DNA after incubation. Data are expressed as a percentage of the activity in reference control plasma. Individual human subjects and median are shown. B. Digestion of human nucleosomal DNA by human plasma. Data are expressed as a percentage of the input DNA. Individual human subjects and median are shown. C. Digestion of chromatin in microparticles by the soluble fraction of patient plasma. Individual subjects and median are shown.

- 81 -

were resuspended and stained with anti-CD42b and anti-CD235a in order to

eliminate platelets and erythrocyte debris respectively, which are commonly

included after microparticle isolation due to their similar size. The stained

microparticles were then assessed by flow cytometry, and the negative fraction

counted precisely. In parallel, qPCR on these isolated microparticles was

performed along with a human DNA standard curve, and, in conjunction with the

precise quantification of starting material as determined by flow cytometry, the

mass of DNA contained in each microparticle was calculated. Because platelets

and erythrocytes do not contain DNA, we were unconcerned about potential false

reads stemming from their inclusion in the qPCR reaction. Strikingly, we found

that DNASE1L3-null patients had hugely elevated levels of DNA per microparticle

compared to controls, with haplodeficient parents again at intermediate levels

(Fig. 28A). Additionally, we assessed three patients heterozygous for the R206C

hypomorphic variant and found that they also had elevated levels of DNA in

circulating microparticles (Fig. 28A).

As we previously showed that microparticle DNA represents the vast

majority of circulating DNA in human plasma, we also assessed the levels of

DNA in unfractionated plasma from these patients by qPCR. Indeed, we

observed increased DNA in total plasma of DNASE1L3-null patients compared to

controls, with haplodeficient parents and R206C heterozygous patients

intermediate (Fig. 28B). Collectively, we believe that these genetic data in both

mice and humans demonstrate that DNASE1L3 is required for normal digestion

of DNA in plasma, in particular in microparticle DNA.

- 82 -

DNASE1L3-sensitive binding of patient IgG to microparticles

We next wanted to assess the extent of autoreactivity against

microparticles in these null patients. To do this, we employed the same protocol

used previously in Dnase1l3-deficient mice. Briefly, microparticles were isolated

from Jurkat cells, then treated with an empty vector control or with recombinant

DNASE1L3. After incubation these microparticles were stained with plasma from

patients, then washed and stained again with an anti-human IgG secondary

antibody. Flow cytometry was employed to determine the extent of binding. We

found that only plasma from the patient who had previously developed SLE

contained microparticle-specific autoantibodies (Fig 29). Notably, this binding

was completely absent when microparticles were treated with DNASE1L3 (Fig.

29). The patient who did not develop full blown SLE and the healthy

haplodeficient parents did not have microparticle-specific autoantibodies (Fig. 29).

It is also notable that the first patient retains detectable levels of these

Figure 28: DNASE1L3 null mutations in humans result in defective processing of endogenous microparticles in circulation. A. The amount of genomic DNA in MPs isolated from patient plasma. Results are shown as the amount of DNA per MP as determined by qPCR. Individual subjects and median values are shown. B. The amount of genomic DNA per volume of total unfractionated patient plasma, as determined by qPCR. Individual subjects and median values are shown.

- 83 -

autoantibodies despite being in disease remission after treatment with rituximab,

a potent therapeutic antibody that blocks B cell activation. Given the very high

levels of DNA in the microparticles of both patients and the lack of an immune

response in the second patient, these data strongly suggest that accumulation of

DNA in microparticles is the primary insult that precedes loss of tolerance.

Binding of circulating IgG from sporadic SLE patients to microparticles

While studies of these DNASE1L3-null patients provide compelling genetic

evidence validating many of our experiments using mice, the applicability of our

findings to the much more prevalent sporadic disease remain unclear.

Furthermore, the rarity of patients that fully lack DNASE1L3 make detailed

analysis unfeasible. Consequently, we sought to determine whether these

findings were of relevance to SLE more generally. To accomplish this, we again

Figure 29: DNASE1L3-sensitive binding of patient IgG to exogenous microparticles. Jurkat microparticles were treated with an empty vector control or with DNASE1L3, then incubated with plasma from human subjects followed by a secondary anti-human IgG fluorescent antibody. Histograms from DNASE1L3 haplodeficient parents are shown as well as the fully deficient patients. Percentage of IgG positive MPs is indicated for each condition.

- 84 -

took advantage of our collaboration with Drs. Buyon and Clancy of the NYU

Department of Rheumatology, obtaining plasma samples from 136 patients with

SLE and 12 controls. We began by assessing the prevalence of anti-

microparticle autoantibodies in all patients and controls, again using

microparticles isolated from Jurkat cells that were treated with a vector control or

with DNASE1L3.

We observed major variability in the extent to which both treated and

untreated microparticles were bound by patient plasma. We began by examining

binding to untreated microparticles. As expected, there was a very strong

correlation between the percentage of microparticles bound by IgG and the mean

fluorescence intensity of stained microparticles (R2=0.82) (Fig. 30). We decided

to define patients as being positive for anti-microparticle IgG if we observed

binding to at least 5% of microparticles, calling them “Binders”. The roughly 20%

of patients that failed to meet this threshold (28/136) we deemed to be “Non-

binders” (Fig. 30). Notably, none of the control plasma tested bound more than

5% of microparticles (Fig. 30). Patients with anti-microparticle antibodies bound

to 30.59 +/- 16.50% of control treated microparticles (Fig. 30).

- 85 -

We then compared the binding of each patient to microparticles treated

with DNASE1L3, and noted that patients fell into three classifications: one group

bound to microparticles only in the absence of DNASE1L3 (DNASE1L3-

sensitive); one group bound to microparticles regardless of DNASE1L3 treatment

(DNASE1L3-insensitive); and one group failed to bind to microparticles

regardless of treatment (non-binder) (Fig. 31A). To mathematically categorize

patients as sensitive or insensitive, we compared the differences in both

0

10

20

30

40

50

60

70

80

90

100 1000 10000 100000 1000000

% o

f Ig

G p

ositiv

e M

Ps

IgG total MFI

Binders Non binders Controls

Figure 30: Extensive binding of microparticles by the plasma of patients with sporadic SLE. Microparticles from Jurkat cells were stained with the soluble fraction of patient plasma, followed by a fluorescent anti-human secondary antibody. Shown is the MFI of IgG fluorescence versus the percentage of positively stained microparticles for each human subject. Positive binding was defined as > 5% IgG positive MPs (blue circles). Patients who bound to fewer than 5% were classified as non-binders (red circles). Healthy control subjects all failed to bind (green circles).

- 86 -

percentages of microparticles bound and the total mean fluorescence intensity

(MFI) between incubation with native or DNASE1L3-treated microparticles.

Patients were considered insensitive to DNASE1L3 treatment if there were no

changes of at least 2 fold in either percent of microparticles bound or in MFI. Our

calculations showed that 35.3% of patients bound to microparticles in a

DNASE1L3-sensitive manner (48/136), while 44.1% (60/136) bound regardless

of treatment with DNASE1L3 (Fig. 31B). These data demonstrate that

microparticles represent an important source of B cell self-antigens in sporadic

SLE and that treatment of microparticles with DNASE1L3 is sufficient to

modulate their immunogenicity in some patients.

Figure 31: Classification of sporadic SLE patients based on DNASE1L3 sensitivity of IgG binding. A. Patient plasma was used to stain MPs treated with an empty vector control or with DNASE1L3. Representative histograms of IgG staining for each reactivity pattern are shown. Percentages of IgG positive MPs are indicated. B. The distribution of sporadic SLE patients into categories based on their staining pattern (n=136).

- 87 -

DNASE1L3 activity varies between sporadic patients

We also wanted to assess the activity of circulating DNASE1L3 in these

patients. To do this, we employed the same assay we had developed to look at

depletion mouse models (see Chapter 1). In brief, different amounts of wild type

serum were titrated and used to digest 1 ng of coated plasmid DNA in order to

determine a standard. Digestion with patient plasma could thus be calculated as

a percentage of normal activity. We began by assessing the activity of controls

versus SLE patients, noting a much greater range of activity levels across

patients (Fig. 32A). However, when we separated patients by their categorized

binding to microparticles, we noted that patients that bound to microparticles in a

DNASE1L3-sensitive manner had significantly lower levels of DNASE1L3 activity

compared to controls, non-binding patients, and DNASE1L3-insensitive binding

patients (Fig. 32B). Because DNASE1L3 is an extracellular protein whose activity

in circulation is key to protect against the development of autoimmunity, this

cohort may be the most amenable to therapeutic intervention via the delivery of

DNASE1L3. Further investigation is required to determine whether the decline in

DNASE1L3 activity in these patients is potentially contributing to loss of tolerance

or if it is a secondary effect due to ongoing disease. In addition, understanding

exactly what is causing these differences could yield valuable insights into the

regulation of DNASE1L3 both in inflamed conditions and at the steady state.

- 88 -

Finally, we wanted to assess whether the variability we observed in

DNASE1L3 activity was reflected in differences in endogenous microparticle

DNA in patients. We isolated microparticles from human plasma by centrifugation,

then counted microparticles by flow cytometry as previously, again only

considering the fraction negative for the platelet and erythrocyte markers CD41b

and CD235a. Due to limitations of detection, we only considered patients from

whom we were able to isolate at least 1000 microparticles, leaving us with 36

Figure 32: The activity of DNASE1L3 in the circulation of sporadic SLE patients differs according to the observed pattern of staining. A. DNASE1L3 activity assays of all sporadic patients and healthy controls, as measured by the digestion of DOTAP-coated plasmid DNA, expressed as a percentage of the activity in reference control plasma. Individual subjects and median are shown. B. DNASE1L3 activity assays of SLE patients only, classified based on their previously determined pattern of binding to native and DNASE1L3-treated MPs. Individual subjects and median are shown.

- 89 -

patients. We performed qPCR using Alu-specific primers to quantify the amount

of DNA present in each sample, then divided by the number of microparticles

used in the test to yield DNA per microparticle. Notably, there was a strong

inverse correlation (R2=0.492) between circulating DNASE1L3 activity and DNA

per microparticle in sporadic patients (Fig. 33). These data suggest that even

small changes in DNASE1L3 activity can impact the amount of DNA present in

circulating microparticles in sporadic patients.

Conclusions

Overall these data demonstrate that chromatin on the surface of

circulating apoptotic microparticles represents an antigen for autoreactive B cells

and immune complexes in sporadic and inherited human SLE. In patients lacking

active DNASE1L3, we observed strikingly higher levels of DNA in circulating

microparticles, confirming our studies using Dnase1l3-deficient mice.

Furthermore, the presence of high levels of microparticle DNA in the absence of

Figure 33: Changes in DNASE1L3 activity in sporadic patients impacts the DNA cargo of microparticles. DNA load of MPs isolated from the plasma of sporadic patients was determined by qPCR and flow cytometry. The soluble fraction was used to assess DNASE1L3 activity. Individual patients and best fit line are shown.

- 90 -

active SLE in the second patient indicate that this elevation may be the primary

defect that leads to loss of tolerance upon interaction of autoreactive B cells with

DNA-laden microparticles. Also intriguing is the consistent observation of anti-

C1q autoimmunity in DNASE1L3-deficient patients; further studies are required

to determine the potential links between microparticles and complement. One

possibility is that increased chromatin in microparticles due to loss of DNASE1L3

results in C1q binding with greater avidity, as previous studies have

demonstrated that DNA is a substrate for C1q (153, 245, 246). This increased

binding could lead to enhanced clearance of C1q and also potential activation of

autoreactive clones, resulting in the high levels of anti-C1q autoantibodies

observed in these patients and ultimately HUVS (151).

- 91 -

Discussion and Future Directions

Dnase1l3-deficient mice represent a novel, monogenic model for the study

of inherited SLE

Loss of DNASE1L3 in humans results in extremely aggressive, early onset

SLE driven by anti-dsDNA IgG that lacks a sex bias (150, 151). Upon knockout of

Dnase1l3 in mice, we observed the same features on two independent pure

strains as well as on mixed backgrounds: loss of tolerance to self-DNA leading to

elevated titers of anti-DNA and anti-chromatin antibodies at very early ages,

followed by late-onset immune activation driven by immune complex deposition.

Importantly, this model provides the means to study inherited SLE in mice;

previous efforts have been unsuccessful. As previously discussed, defects in the

complement component C1Q have been tied to severe SLE, and in fact null

mutations in C1Q cause SLE in human patients (90, 176). In mice, however,

C1Q is dispensable, as its knockdown fails to cause any discernable pathology

(247). Consequently, to our knowledge Dnase1l3-deficient mice represent the

only mouse model of hereditary SLE reflective of human disease. In addition,

these data provide an explanation for the extensive conservation observed

between mouse and human DNASE1L3.

An outstanding question regarding this model concerns the delayed

immune activation observed, which is in contrast to other mouse models and to

DNASE1L3-deficient patients. The rapid development of autoreactivity against

dsDNA and chromatin only led to comprehensive immune activation in old

- 92 -

animals, and furthermore these mice failed to lose tolerance to other self-

antigens (e.g. RNA). One possible explanation for this is the lack of

supplementary mutations prominent in other mouse models. For instance, the

Yaa mouse features enhanced sensing of RNA due to duplication of TLR7, while

the Faslpr model is driven by aberrant lymphoproliferation and defective negative

selection. Another important consideration is the lack of baseline immune

stimulation in mice housed under specific pathogen-free conditions compared to

human patients, especially considering the extensive literature linking viral

infections and the resulting antiviral immune responses to SLE flaring in patients

(4, 24, 84). Strong evidence for this hypothesis stems from our observation that

artificially introducing IFN into Dnase1l3-deficient animals dramatically

accelerated autoimmunity and induced novel anti-RNA responses, and

furthermore led to substantial mortality. Further studies could similarly take

advantage of this model to examine the effects of other triggers linked to flares in

human disease. In particular, induction of apoptosis in vivo by pharmaceutical

means may yield insights into environmental effects on pathogenesis in humans.

We thus conclude that anti-DNA reactivity directed against chromatin in these

mice causes the observed pathologies, and furthermore primes for severe SLE-

like disease upon inflammatory stimulus.

- 93 -

Autoimmunity in Dnase1l3-deficient mice is distinct from that seen in

Dnase2- or Trex1-deficient animals

Previous studies of DNASE1L3 pointed towards a potential role in the

processing of intracellular DNA molecules, in particular during apoptosis or

necrosis (170-172). However, cytoplasmic sensing of DNA requires the

intracellular adaptor molecule STING, as evidenced by the ablation of disease

when STING is deleted in conjunction with either DNASE2 or TREX1 (126, 248).

Disease in Dnase1l3-deficient animals, in contrast, was totally independent of

STING, suggesting that DNASE1L3 processes an extracellular form of DNA. In

addition, our lab observed that disease was completely dependent on MyD88,

which may point to a role for TLR or IL-1 signaling. Importantly, aberrant

activation or any developmental defects in either DCs or B cells in very young

Dnase1l3-deficient animals was not detected. We therefore hypothesize that the

observed primary response to DNA may be driven by direct recognition by and

subsequent expansion of DNA-reactive B cells, in all probability by extrafollicular

activation. These primary signals are then amplified by a MyD88-dependent

pathway or pathways in B cells or possibly other cell types. Thus, a critical future

direction remains identifying the specific MyD88-dependent signals involved,

which would provide additional insight into the mechanisms underlying loss of

tolerance to genomic DNA which drives autoimmunity in Dnase1l3-deficient

animals. More broadly, the extrafollicular production of autoantibodies may be a

unique property of self-DNA as an antigen, as it has recently been shown that

- 94 -

activation of TLR7 or TLR9 in conjunction with BCR signaling results in robust

affinity maturation in the absence of germinal centers (249).

The presence of DNASE1L3 in circulation is required constitutively to

prevent the development of anti-dsDNA autoimmunity

We observed that the levels of DNASE1L3 inversely correlated with the

development of anti-dsDNA Abs. This suggests that the presence of DNASE1L3

is constitutively required in circulation in order to prevent interactions between

extracellular DNA and autoreactive B cells. We also demonstrate that the

majority of serum DNASE1L3 is produced by DCs and subsets of tissue

macrophages, in concordance with previously published microarray data. In

particular, CD11c+ DCs produced roughly 70-75% of circulating DNASE1L3.

While DCs and macrophages have been previously shown to play key roles in

mediating self-tolerance and thus preventing autoimmunity (182, 183), we

believe our data present a novel mechanism by which this occurs. While it is

established that phagocytosis of apoptotic cells, induction of tolerogenic T cell

programs, and the production of anti-inflammatory molecules and cytokines are

all means by which self-tolerance is maintained by DCs and macrophages, here

we demonstrate that the release of a DNA-digesting enzyme into circulation is

required to prevent the activation of autoreactive B cells.

One confounding observation we made concerned the kinetics of

DNASE1L3 activity after hematopoietic reconstitution. In wild-type mice receiving

Dnase1l3-deficient bone marrow, activity persisted in these animals until roughly

- 95 -

30 weeks, despite typical splenic dendritic cell turnover being on the order of

days (250). This may be due to the production of DNASE1L3 into circulation in

response to lethal irradiation by a non-hematopoietic radio-resistant cell

population such as hepatocytes, which normally express DNASE1L3 although

they seemingly do not contribute to circulating levels at the steady state.

Alternatively, long-lived resident macrophage populations may persist and

upregulate DNASE1L3 in response to the inflammation induced by total body

irradiation, allowing for some degree of compensation. Further studies examining

potential contributions of hepatocytes or tissue resident cells to systemic

DNASE1L3 under duress may provide context for these findings.

Microparticles are an endogenous substrate of DNASE1L3

The unique features of DNASE1L3 described here and by others led us to

hypothesize that its endogenous substrate is chromatin-loaded circulating

microparticles, for several reasons. Firstly, the requirement for DNASE1L3

systemically in order to prevent autoimmunity indicates that its substrate must be

present in circulation ubiquitously. Microparticles containing DNA are present

normally in circulation (205, 216) as a consequence of the constant and rapid

turnover of various blood cell types, in particular granulocytes. Secondly,

DNASE1L3 is uniquely able to digest membrane-encapsulated DNA. This activity

is dependent on its highly basic C terminus, the importance of which is made

clear by the unusual degree to which it is conserved evolutionarily. Microparticles

are intact structures that sequester DNA, as evidenced from the requirement to

- 96 -

disrupt microparticle membrane integrity with detergents in order to robustly stain

with propidium iodide. Thirdly, the efficiency with which DNASE1L3 digests DNA

packaged into nucleosomes suggests its target is likely chromatin-based.

Microparticles contain the vast majority of detectable genomic DNA in human

plasma, which is comprised of primarily nucleosomal chromatin fragments

derived from leukocytes (203, 204, 251). Consequently, mice deficient for

Dnase1l3 had significantly higher levels of DNA within their circulating

microparticles as well as in their total plasma, which was corroborated in both

DNASE1L3-deficient human patients and their haplodeficient parents. Thus,

although it is possible that DNASE1L3 may process other types of self-DNA, our

data demonstrate that genomic DNA contained within microparticles is a relevant

endogenous target (Fig. 34).

- 97 -

Microparticles are relevant sources of B cell antigens

Previous studies have also shown that chromatin on circulating

microparticles is exposed and can be bound by autoantibodies (210, 212, 213),

demonstrating that microparticle DNA represents a viable antigen for B cells.

DNASE1L3 digests this exposed chromatin on microparticles, preventing the

binding of autoantibodies from Dnase1l3-deficient animals and from the

DNASE1L3-deficient patient who had previously developed SLE. In addition, our

Figure 34: Proposed mechanism by which DNASE1L3 prevents the development of autoimmunity. As microparticles are released from cells undergoing constant turnover, DNASE1L3 released into circulation by dendritic cells and macrophages digests their chromatin contents. This prevents interactions between DNA-specific B cells and this antigenic self-DNA, ultimately resulting in tolerance to self-DNA. In the absence of DNASE1L3, chromatin-loaded microparticles persist in circulation and are able to be presented to DNA-specific T cells, ultimately resulting in the activation of autoreactive B cells and the release of anti-dsDNA antibodies.

- 98 -

experiments using murine hybridomas and 9G4+ monoclonals demonstrate that

this chromatin is recognized by numerous archetypal autoantibodies found in

both mouse and human SLE in a DNASE1L3-sensitive manner. Finally and most

broadly, we observed binding by IgG from roughly 40% of sporadic patients to

DNASE1L3-sensitive chromatin on microparticles. Endogenous microparticles

from lupus patients have been previously described to be extensively bound by

complement and IgG (207, 210), demonstrating their relevance as SLE

autoantigens in vivo. Furthermore, the ability of microparticles to drive

autoimmunity in sporadic patients may explain the flaring aspect of SLE, as

temporal increases in microparticle production and/or release could overwhelm

the available pool of DNASE1L3, allowing for persistence of unprocessed

chromatin in circulation.

In addition to the patients whose antibodies bound to microparticles only in

the absence of DNASE1L3, a further 40% of patients had IgG that bound to

microparticles both before and after DNASE1L3 treatment. Because they contain

a myriad of potentially immunogenic proteins, lipids, and nucleic acids, extensive

characterization of the exposed surface antigens available to be bound by patient

antibodies is required. One such molecule that is present on microparticle

surfaces and is unaffected by DNASE1L3 is Sjögren’s-syndrome-related antigen

A (SSA), antibodies against which are very strongly associated with both SLE

and with disease-related complications (139-141). Furthermore, the apoptotic

mechanisms that drive the export of chromatin from nuclei to microparticle

surfaces remain uncharacterized, and additional in depth analysis of their

- 99 -

contents may yield insights to this somewhat enigmatic pathway. Notably, the

R206C hypomorphic variant of DNASE1L3 has also been linked by GWAS to the

development of scleroderma (160), which is often characterized by the presence

of anti-centromere or anti-kinetochore autoantibodies (252). As evidenced by its

ability to reduce the levels of the DNA-associated protein HMGB1 on

microparticle surfaces, DNASE1L3 may also be capable of processing other

potentially immunogenic molecules to prevent loss of tolerance in systemic

scleroderma.

Beyond SLE: DNASE1L3 and other diseases

Another outstanding question concerns the strong associations between

human DNASE1L3 deficiency and the development of anti-C1q antibodies, which

eventually results in hypocomplemetemia and likely contributes to the severity of

SLE. C1q is normally involved in the clearance of apoptotic or other cellular

debris, and strikingly has been shown to bind with increased avidity to such

structures when they contain chromatin (153, 245, 246). Thus, we would

hypothesize that in the absence of DNASE1L3, microparticles are more

extensively bound by C1q. This may both deplete the pool of free C1q in

circulation and may also contribute to activation of C1q-specific B cells, since

C1q is now presented in conjunction with immunogenic molecules contained

within microparticles. As these cells become activated and produce anti-C1q, the

pool of C1q is further reduced, leading to impaired clearance of chromatin-loaded,

unprocessed microparticles. Persistence of these microparticles would allow for

- 100 -

presentation to DNA-specific T and B cells, resulting in loss of tolerance and

progression to SLE. Future experiments will investigate the potential impact of

treatment with DNASE1L3 on the binding of complement to microparticles. In

addition, determining whether anti-C1q antibodies are elevated in Dnase1l3

knockout mice is key, albeit with the caveat that these antibodies do not lead to

SLE development in mice as in humans (247). Finally, assessment of potential

anti-C1q antibodies in sporadic patients will also be an important parameter,

especially in the context of anti-microparticle immunity.

In addition, we are keen to investigate potential roles for DNASE1L3 in the

regulation of cancer microparticles. As previously mentioned, microparticles have

been implicated in the pathogenesis of cancer, both by enhancing metastasis

and transformation through the delivery of various types of nucleic acids (184,

186, 187, 200). Thus, DNASE1L3 may be an important host factor in limiting

tumorigenesis, which will be tested in both spontaneous and injected mouse

models. An alternative possibility is that the presence of DNASE1L3 prevents

host interactions with oncogenic DNA or oncoproteins, which may impair the

development of specific immune responses. More broadly, we would like to

explore the potential usage of microparticles as a diagnostic surveillance tool.

Because they contain mRNA in addition to DNA, we would like to perform RNA-

seq on microparticles isolated from various disease models, in particular cancer.

These data can then be used to determine the cells of origin, so deviations from

the steady state may be indicative of transformation. Such a liquid biopsy could

- 101 -

be easily applied to humans, and may be a way to improve early detection of and

screening for cancer.

DNASE1L3 activity as a marker for sporadic SLE

An intriguing finding in our sporadic cohort was the observation that

diminished levels of circulating DNASE1L3 correlated with larger reductions in

binding to microparticles after treatment with exogenous DNASE1L3. Although

statistically significant, the biological impact is unclear, since individuals carrying

the hypomorphic R206C variant have much greater reductions in activity without

necessarily progressing to overt disease. In addition, the mechanisms driving

differences in the activity of DNASE1L3 remain undiscovered. One possibility is

the development of autoantibodies against DNASE1L3 itself, which will be tested

in sporadic patients. Alternatively, preliminary data in the lab has suggested that

DNASE1L3 expression is induced in macrophages in response to inflammatory

stimuli, which may also explain variability between not only patients but also

samples from the same patient. With the help of a collaboration with experts in

the field of biostatistics, we are in the process of closely examining our data in

conjunction with all measured parameters available. We are keen to determine

whether DNASE1L3-sensitive chromatin binding correlates with data obtained in

the clinic, which may lead to translational opportunities. In parallel, we are

beginning to investigate patients longitudinally, which may allow for in depth

assessment of the impact of disease flares, therapeutic interventions, and/or

- 102 -

inflammatory status on IgG binding to microparticles as well as the activity of

DNASE1L3.

Our initial studies of sporadic patients were deliberately limited to those of

European descent, with the aim being to specifically examine the impact of the

R206C variant. We subsequently expanded our studies to include African

American and Hispanic patients. Among the first group, we observed a greater

fraction of patients who had no circulating anti-microparticle IgG compared to the

cohort overall (~35% vs 20%). Given that disease in the latter populations tends

to be more severe, anti-microparticle responses may represent an important and

predictive disease parameter. Indeed, such a diagnostic test would be relatively

straightforward to implement, and may lead to substantial quality of life

improvements for patients who may be able to receive therapeutics

prophylactically to prevent flaring as opposed to reactively.

Therapeutic applications of DNASE1L3

One therapeutic option that we are eager to explore further concerns the

administration of exogenous DNASE1L3, since we were able to restore normal

levels of DNASE1L3 activity and consequently significantly delay the onset of

anti-dsDNA autoantibodies by the use of an adenoviral vector in Dnase1l3-

deficient mice. A major drawback of this system, however, is that adenoviral

vectors persist in hepatocytes for only 6-8 weeks, after which they are cleared by

the immune system and subsequent administrations are ineffective due to

memory responses. One solution is to take advantage of adeno-associated

- 103 -

viruses (AAVs), which do not trigger immune responses and can thus be used for

prolonged expression in hosts (253, 254). Our lab has obtained two such AAVs

encoding DNASE1L3, and both have proved effective at restoring DNASE1L3

levels to those comparable to wild-type animals. It remains to be seen whether

these AAVs are effective at ameliorating disease development, but if so such a

treatment could be beneficial in sporadic SLE patients, especially since the

cohort with DNASE1L3-sensitive antibodies also presented with lower

DNASE1L3 activity.

In addition, our laboratory is working on the development of hyperactive

DNASE1L3. As previously discussed, DNASE1 was modified to be actin-

independent and more potent catalytically by the substitution of six amino acids,

four of which are already present in native DNASE1L3 (166). Using site-directed

mutagenesis, the additional two substitutions have been introduced, and we are

in the process of testing its efficacy. Although it may prove challenging due to the

importance of the tertiary structure of the C terminus, we are also investigating

potential modifications of DNASE1L3 that may allow it to persist for longer in

circulation. Ultimately, we envision such a molecule as being a valuable

therapeutic tool for the treatment of both inherited and sporadic SLE: a potent,

long-acting injectable enzyme, the dosage of which can be modified based on

well-defined longitudinal parameters.

We are also intrigued at potential roles for DNASE1L3 outside of

circulation. As mentioned, it is expressed by hepatocytes, despite our

observations that hematopoietic cells are responsible for the pool of DNASE1L3

- 104 -

in circulation. Consequently, it is possible that DNASE1L3 is also released into

the digestive tract, as is the case for DNASE1. Given its potent activity and ability

to process coated DNA, DNASE1L3 may play an important role in regulating the

intestinal microbiome, potentially by limiting bacterial horizontal gene transfer. In

addition, we have observed much more efficient adeno-associated viral

transduction in Dnase1l3-deficient mice compared to wild-type animals, which

may suggest that DNASE1L3 also plays a role in defense against DNA viruses.

Characterization of the microbiome and viriome of Dnase1l3-deficient animals

may prove illuminating. If DNASE1L3 does indeed have a role as an anti-viral

molecule in particular, its administration may prove effective for the treatment of

enveloped DNA viruses, such as herpesviruses or poxviruses.

Conclusion

SLE is a very challenging disease to study for numerous reasons, chief

among them being the heterogeneity of presentation between individual patients

and/or flares. Indeed, it seems likely that what is collectively referred to as SLE in

fact represents a collection of closely related disorders, wherein loss of tolerance

occurs to a self-nuclear antigen. This can be followed by epitope spreading and

the development of broader autoimmunity. Our studies have focused on

DNASE1L3, since it is strongly tied to the very early and robust development of

anti-dsDNA autoantibodies. These antibodies are predictive of both disease

flaring and of disease severity between patients, underlying the urgent need to

- 105 -

understand more about how and why they are generated (255). Our data provide

an example of one potential mechanism by which tolerance to self-DNA is lost.

In addition, it is possible that microparticles may be fundamentally

involved in loss of tolerance to self-antigens that are not DNA-based. Given the

large quantities of RNA and nuclear proteins they contain, their persistence in the

bloodstream may also allow for delivery of additional autoantigens to reactive T

or B cells. Further work is required to understand the extent of their contents and

in what contexts they can contribute to the initial steps of autoimmunity. The

complement system seems to be involved; preliminary work in the lab has found

that microparticles are actively bound by multiple forms of complement.

Furthermore, C1q deficiency represents an additional monogenic form of SLE in

humans (155), suggesting that microparticle clearance is fundamentally required

to prevent aberrant self-reactivity.

In conclusion, our data demonstrate that chromatin in microparticles is a

potential self-antigen for autoreactive B cells, and that it is processed by

circulating DNASE1L3 in both mice and humans. This digestion is both restricted

to DNASE1L3 and is absolutely required to prevent the development of rapid

anti-DNA responses that progress over time to full blown SLE. In addition, these

results provide mechanistic insight to the observed associations between null

and hypomorphic mutations in DNASE1L3 with inherited and sporadic SLE,

respectively. More broadly, we have discovered a novel, cell-extrinsic means by

which tolerance to a self-nucleic acid is enforced by DCs and, to a lesser extent,

macrophages. The secretion of an enzyme that protects against autoimmune

- 106 -

development allows for unique opportunities for therapeutic intervention, as

evidenced by our observation of delayed anti-DNA reactivity after re-expression

of DNASE1L3 using an adenoviral vector. In particular, such a therapy may

prove efficacious in the roughly 40% of patients who presented with extensive

IgG binding to DNASE1L3-sensitive chromatin, especially as this same cohort

also featured lower levels of DNASE1L3 activity. In addition, mutations in

DNASE1L3 have been linked to other autoimmune diseases such as systemic

scleroderma and HUVS, so delivery of exogenous DNASE1L3 may not be limited

as a therapeutic tool to SLE alone.

- 107 -

Materials and Methods

Dnase1l3-deficient animals

All experiments were performed according to the investigator’s protocol,

which was approved by the Institutional Animal Care and Use Committees of

Columbia University and New York University. Mice with a targeted germline

replacement of Dnase1l3 (Dnase1l3LacZ) were purchased from Taconic Knockout

Repository (model TF2732). Backcrossing onto 129SvEvTac and C57BL/6

backgrounds was performed for >10 generations, and pure strains were then

intercrossed to yield Dnase1l3LacZ/LacZ knockout animals. Age-matched wild-type

mice of the same backgrounds were bred in the same colony for use as controls.

Deletion of STING or MyD88 in Dnase1l3-deficient mice

STING-deficient mice (Tmem173gt/J) (256) and MyD88-deficient mice

(Myd88tm1.1Defr/J) (257) were obtained from Jackson Laboratories, both on pure

B6 backgrounds. Both strains were crossed to Dnase1l3LacZ B6 mice to produce

double-heterozygous F1 litters, which were then backcrossed to Dnase1l3LacZ

mice, thus generating Dnase1l3LacZ/LacZ mice that were either wild-type or

homozygous null for Tmem173 or Myd88 (double knockout, dKO). To allow for

infection-free survival of MyD88-deficient animals, breeder cages were

administered trimethoprim and sulfadiazine-supplemented animal chow

(Uniprim®, Harlan Teklad). Upon weaning, mice were switched to a normal diet.

- 108 -

Importantly, this diet had no effect on the development of autoimmunity in the

littermate controls of dKO mice who lacked only DNASE1L3.

Human subjects

DNASE1L3-deficient HUVS patients 1 and 2 correspond to patients IV-4

and IV-5 from family 1 described in (151). Patient 1 (14-year-old female)

developed SLE with ANA and high anti-dsDNA 6 years ago. She was treated

with rituximab (most recent dose in May 2012) and has been in clinical remission

since (limited ANA, normal anti-dsDNA), although patient still presents with active

uveitis. Patient 2 (12-year-old female) has not developed SLE (negative ANA, no

anti-dsDNA) but presents with ongoing HUVS with moderate to severe vasculitis

of the kidney. Their study was approved by the Ethics Committees of Ankara

University (Turkey) and the Institutional Review Board of the University of Miami.

Informed consents were obtained from the parents.

Blood from patients with sporadic SLE and healthy controls was obtained

from the NYU IRB-approved Rheumatology SAMPLE (Specimen and Matched

Phenotype Linked Evaluation) Biorepository. All patients signed IRB-approved

informed consent forms. All patients with sporadic SLE met at least 4 revised

American College of Rheumatology criteria (12). Subjects were genotyped for

DNASE1L3 rs35677470 using the TaqMan SNP genotyping kit (ThermoFisher).

Three subjects (one healthy control and two SLE patients) were found to be

heterozygous for the rare R206C hypomorphic allele.

- 109 -

Adenoviruses

The adenoviral vector encoding IFN-5 has previously been described

(175). Human DNASE1L3 was cloned into an adenoviral vector and constructed

by Welgen. Viral particles were produced and purified at Welgen, and were

injected into Dnase1l3LacZ/LacZ mice intravenously at 0.5 or 1 x 1010 particles per

mouse. No discernable differences were observed between the two doses. As a

control, Dnase1l3LacZ/LacZ mice were injected with equivalent doses of adenoviral

particles encoding GFP.

ELISA

Anti-dsDNA and anti-RNA IgG titers were determined as previously

described (72, 258). Plates were pre-coated with poly-L-lysine (0.05 mg/mL) for 2

hours at room temperature, followed by coating with 0.1 mg/mL calf thymus DNA

or yeast RNA (Sigma-Aldrich) as antigens. Anti-nucleosome IgG titers were

assessed using plates coated with 1 g/mL of purified HeLa polynucleosomes

(Epicypher). After coating, sera were added and incubated overnight. The

amount of bound IgG was measured with an alkaline phosphatase (AP)-

conjugated goat anti-mouse IgG antibody (1:5000, Jackson ImmunoResearch).

Antigen-specific IgG titers were determined using serial dilution of serum from a

positive animal as a standard, and thus expressed as arbitrary units per volume.

Anti-dsDNA IgG subtypes were measured as above using AP-conjugated

antibodies to the indicated IgG subtypes (Southern Biotech). Total

- 110 -

immunoglobulin isotype levels were also determined as above, using AP-

conjugated antibodies to IgM, IgA, and IgG (Southern Biotech).

ELISPOT

To detect antibody-secreting cells (ASC) specific to dsDNA, 96 well

multiscreen plates (Millipore) were pretreated with 35% ethanol for 1 minute,

coated with 100 g/mL calf thymus dsDNA (Sigma-Aldrich), then washed and

blocked with a solution of 3% fetal calf serum (FCS) and 3% bovine serum

albumin (BSA) in PBS. Single-cell suspensions of total splenocytes were plated

in duplicate in six two-fold serial dilutions starting at 106 cells per well. Cells were

incubated on the plate for 5 hours at 37oC. Plates were then washed and

incubated with AP-conjugated goat anti-mouse IgG (1:5000, Jackson

ImmunoResearch) overnight at 4oC. Spots were developed using the NBT/BCIT

(Sigma-Aldrich) system and counted on an ImmunoSpot Series 1 ELISPOT

analyzer (Cellular Technology Ltd).

Anti-nuclear antibody staining (ANA)

Prefixed HEp-2 cells on glass slides were purchased from MBL Bion, and

were incubated with mouse serum (1:100 dilution). This was followed by PE-

labeled goat anti-mouse IgG and DAPI. Images were captured on a confocal

fluorescent microscope (LSM 710 NLO) and were processed by Zen software

(Carl Zeiss).

- 111 -

Kidney IgG deposition and histopathology

Both kidneys were collected and fixed in 4% paraformaldehyde,

dehydrated in 30% sucrose, then frozen in OCT (TissueTek). Frozen sections (5

m) were stained with DAPI and PE-labeled goat anti-mouse IgG (eBioscience),

then visualized by confocal microscopy as above.

Sections of formalin-fixed kidneys (2 m) were stained with hematoxylin

and eosin, then evaluated by a pathologist (V. D’Agati) blinded to sample identity.

Mesangial and endocapillary proliferation, leukocyte infiltration, glomerular

deposition, and apoptosis were scored individually on a scale from 0 (none) to 4

(severe), then added to produce a cumulative score. The percentage of cortical

parenchyma with interstitial inflammation was also determined. Images were

captured on a Zeiss AxioImager upright microscope and processed by AxioVision

software (Carl Zeiss).

Germinal center immunohistochemistry

To visualize germinal center reactions, frozen spleen sections (5 m) were

stained with PE-labeled anti-mouse B220 (eBioscience) and biotin-conjugated

peanut agglutinin (PNA; Vector Laboratories) followed by FITC-conjugated

Streptavidin (eBioscience). Sections were then visualized by confocal

microscopy as above.

- 112 -

Flow cytometry

Suspensions of peripheral blood leukocytes or splenocytes were

subjected to red blood cell lysis and were subsequently washed and stained with

directly conjugated fluorescent antibodies to the relevant surface markers

(eBioscience). LacZ expression was analyzed as in (259). Single cell

suspensions were washed and resuspended in HBSS containing 1 mM HEPES

and 2% FCS. These cells were incubated for 20 minutes at 37oC, then were

mixed with pre-warmed 2 mM fluorescein di--D-galactopyranoside (FDG,

Sigma-Aldrich) in a hypotonic solution for 1 minute. Immediately thereafter, cells

were placed on ice and washed with ice-cold HBSS buffer. Endogenous

lysosomal -gal activity was blocked by 0.3 mM chloroquine diphosphate. After

loading with FDG, cells were stained for surface markers as above, then

analyzed by flow cytometry. Data was collected on either LSR II (BD) or Attune

NxT (ThermoFisher) flow cytometers and was analyzed using FlowJo software

(Tree Star).

Recombinant deoxyribonucleases

Cloned open reading frames (ORFs) of human DNASE1 (NCBI:

NP_005214.2) and DNASE1L3 (NCBI: NP_004935.1) were subcloned into the

pMSCV-IRES-GFP (pMIG) retroviral expression vector. The constructs for

DNASE1L3 variants were generated using the Q5 site-directed mutagenesis kit

(NEB), including the R206 to C substitution, the C-terminal truncation (amino

acids 282-305), and the insertions of hexahistidine tags between aa 282-283

- 113 -

(His-preCT) or between aa 305 and the stop codon (His-CT). The resulting

constructs or the empty pMIG vector were used as plasmids for the transient

transfection of HEK293 cells using the TransIT-293 transfection reagent (Mirrus).

Equal transfection efficiency was determined by GFP expression, as measured

on an Accuri C6 tabletop flow cytometer (BD Biosciences). To avoid

contamination with bovine DNASEs, transfection was performed in medium

containing 15% KnockOUT serum supplement (Thermo Fisher) in lieu of FCS.

Transfected cells were cultured for 48 hours, and the supernatants were

collected, filtered, supplemented with 4 mM CaCl2 and 4 mM MgCl2, then frozen

in aliquots.

Analysis of DNASE1L3 activity

To measure the digestion of liposome-coated DNA, pMIG plasmid DNA

was pre-incubated with N-[1-(2,3-Dioleoyloxy)propyl]-N,N,N-trimethylammonium

methyl-sulfate (DOTAP) (Roche) in HBSS according to the manufacturer’s

protocol. 1 ng of naked or DOTAP-coated plasmid per reaction was incubated

with an equal volume of DNASE-containing supernatants or sera for 60 minutes

at 37oC, for a final reaction volume of 2 L. The amount of DNA remaining after

digestion was measured by qPCR with GFP-specific primers and expressed as a

percentage of input DNA using a calibration curve generated from serial plasmid

dilutions.

For the measurement of relative DNASE1L3 activity in vivo, the digestion

was performed for 10 minutes using 1 L of mouse or human serum or plasma,

- 114 -

again in a final reaction volume of 2 L. After qPCR, the amount of DNA

remaining post-digestion was converted to a percentage of DNASE1L3 activity

using a calibration curve generated from serial dilutions of multiple wild-type

serum or plasma samples.

Purified human polynucleosomes from HeLa cells (Epicypher) or purified

Jurkat cell DNA (2 ng/reaction) were incubated with an equal volume of DNASE-

containing supernatants for 15 minutes at 37oC in a total volume of 2 L. The

amount of remaining DNA was measured by qPCR with primers specific for

human genomic Alu repeats and expressed as a percentage of input DNA using

a calibration curve generated from serial DNA dilutions.

Hematopoietic reconstitution

To reconstitute the hematopoietic system of recipient mice, 2 x 106 total

BM cells were isolated from WT or KO mice on the 129 background. Cells were

transferred into lethally irradiated WT and KO recipients on the B6 background,

which has the same MHC haplotype. Reconstitution was tracked by staining for

the 129 donor strain specific CD229.1 marker. An additional set of experiments

was performed using WT recipients only. BM cells from WT or KO mice on the

(129xB6) F1 background were transferred into lethally irradiated congenic

(129xB6.SJL) F1 recipients. Reconstitution was tracked by staining for the

B6.SJL recipient strain specific CD45.1 leukocyte marker. In all analyzed

chimeras, the donor contribution was greater than 95%.

- 115 -

Specific depletion of cell populations

Rag1-deficient mice were purchased from Jackson Laboratories. To

deplete dendritic cells (DCs), animals with a Cre-inducible diphtheria toxin

receptor (iDTR) allele (260) were crossed with the DC-specific Cre deletion strain

Cd11c-Cre (261). The resulting Cd11c-Cre/iDTR mice were administered

diphtheria toxin (DTX) at a dose of 20 ng per gram of body weight every other

day for two weeks as described (262). Efficient DC depletion was confirmed by

flow cytometry at the endpoint, and minimal effects of DTX were observed on DC

populations in mice lacking iDTR. To deplete macrophages, B6 mice were

injected intraperitoneally with 150 g per gram of body weight of macrophage

colony-stimulating factor 1 receptor (Csf1r) blocking antibody (clone AFS98,

(263)) purified from a hybridoma as previously described (264). Depletion was

confirmed by flow cytometry. For macrophage depletion by clodronate, mice

were injected intravenously with a single dose (200 L) of clodronate liposomes

(5 mg/mL) or control liposomes containing PBS (obtained from

ClondronateLiposomes.com). Depletion was again confirmed by flow cytometry.

Molecular modeling

ZEGA global sequence alignment of DNASE1 (PDB:2DNJ, chain a) and

DNASE1L3 (Uniprot:Q13609) was performed as previously described (265). 3D

structures were visualized and analyzed with Internal Coordinate Mechanics

Software (ICM-Pro Molsoft LLC, La Jolla, CA). Amino acids in the DNASE1

crystal structure 2DNJ, chain a contacting DNA were identified by selecting

- 116 -

amino acids exhibiting any atom center within a 5 Å radius of any atom center in

the DNA. Prediction of the structure of the isolated DNASE1L3 C-terminus (282-

SSRAFTNSKKSVTLRKKTKSKRS-305) and the C-terminus and His mutants in

the context of the whole DNASE1L3 protein was performed using the Biased-

Probability Monte Carlo algorithm, which was shown to be as accurate as

experimental structure determination for short peptides (266). Homology

modeling of DNASE1L3 was performed as previously described (267).

Generation, isolation, and analysis of microparticles

Microparticles from Jurkat cells were generated as previously described

(210). In brief, 1 mM staurosporine (STS, Sigma-Aldrich) was added overnight to

Jurkat cell cultures. Cells were harvested and collected by centrifugation for 5

minutes at 300 x g. The supernatants were collected and centrifuged at 22,000 x

g for 30 minutes to pellet microparticles. Pellets were resuspended in sterile PBS

and then analyzed on an Accuri C6 tabletop flow cytometer (BD Biosciences) to

determine absolute numbers and ensure >95% enrichment. Where indicated,

microparticles (105 per L) were incubated with an equal volume of DNASE-

containing transfection supernatants for 1 hour at 37oC.

To generate mouse microparticles, primary splenocytes were cultured for

2-3 days with 50 ng/mL phorbol 12-myristate 13-acetate (PMA) and 2 g/mL

ionomycin (Sigma-Aldrich). The resulting cultures of activated, proliferating

splenocytes were treated with 1 mM STS overnight, and microparticles were

collected and analyzed as for Jurkat cells. Weekly, 107 of these microparticles

- 117 -

were injected intravenously into wild-type female 129 mice who had previously

been administered either PBS or 5 x 109 IFN5 adenoviral particles. After four

instances of microparticle injection, the mice were analyzed one week later.

To isolate microparticles from human plasma, blood was collected in tubes

containing ethylenediaminetetraacetic acid (EDTA), and blood cells were

removed by centrifugation at 2,000 x g for 10 minutes at 4oC. In some

experiments, a second centrifugation step at 3,000 x g for 10 minutes at 4oC was

used to deplete platelets as previously described (205). The resulting plasma

(either fresh or stored at -80oC) was centrifuged at 22,000 x g for 30-60 minutes

to pellet microparticles, and the supernatant was used to measure DNASE1L3

activity and antibody binding specificities. Microparticles from murine plasma

were isolated as well. Mice were euthanized and immediately exsanguinated by

cardiac puncture into tubes containing heparin. Plasma was then isolated by

centrifugation at 2,000 x g for 10 minutes at 4oC, and then centrifuged at 22,000

x g for 30 minutes to pellet microparticles. Plasma microparticles were

resuspended in sterile PBS and stained for CD42b and CD235a (human) or

CD41 and Ter119 (mouse) to exclude platelets and erythrocyte debris,

respectively. Stained microparticles were assessed on an Accuri C6 tabletop flow

cytometer (BD Biosciences) to count absolute numbers of the unstained fraction.

To test the ability of mouse serum to digest microparticle DNA, Jurkat

microparticles (5 x 105 per L) were incubated with an equal volume of serum for

1 hour at 37oC in a final reaction volume of 10 L. Human plasma was tested

using the same protocol, with the exception that microparticles (5 x 105 per L)

- 118 -

derived from mouse splenocytes were used instead. The DNA content of Jurkat-

derived microparticles, human plasma microparticles, and total human plasma

was determined by qPCR for human genomic Alu repeats. The DNA content of

mouse splenocyte-derived microparticles, murine plasma microparticles, and

total murine plasma was determined by qPCR for mouse genomic B1 repeats. In

both cases, raw qPCR data was converted into the amount of genomic DNA

remaining after digestion using calibration curves derived from serial dilutions of

the respective genomic DNA. Data were then expressed either as the percent of

input DNA or, in conjunction with our counting of absolute microparticle numbers

through the use of the Accuri, as amount of DNA per microparticle.

Persistence of microparticles in vivo

6 x 106 Jurkat cell-derived microparticles were injected intravenously per

mouse into wild-type mice. Recipients were euthanized at the indicated time

points, and their tissues were analyzed by qPCR for the presence of human DNA

using primers specific for the Alu human genomic repeat.

Staining of microparticles

For surface staining, 2.5 x 105 native or DNASE-treated Jurkat cell-derived

microparticles were incubated with either purified anti-DNA/histone 2a/2b mAb

PR1-3 (10 g/mL), mouse sera (1:10 dilution), purified 9G4-positive Fab

fragments derived from human patients fused to the Fc region of mouse IgG (240,

241, 268), or hybridoma supernatants (PR1-3, PL2-8 (anti-DNA/histone), PL9-11

- 119 -

(anti-DNA) (269, 270), and 3H9 (anti-DNA) (271)) for 30 minutes at 4oC.

Hybridomas were kindly provided by M. Shlomchik (University of Pittsburgh),

while 9G4-positive antibodies were generously provided by G. Silverman (New

York University) and I. Sanz (Emory University). Stained microparticles were

washed by centrifugation at 22,000 x g for 30 minutes, and were then incubated

with PD-labeled goat anti-mouse IgG secondary antibody (eBioscience) at 1:200

for 30 minutes at 4oC. Microparticles were then analyzed by flow cytometry

without further washing on an Attune NxT flow cytometer (ThermoFisher).

Staining with human plasma or sera was performed as above at a 1:20 dilution,

with a PE-labeled goat anti-human IgG secondary antibody (eBioscience).

Statistical analysis

In the interest of robust statistical analysis, normal distribution of values

was not assumed (272). Unless otherwise indicated in figure legends, data were

displayed as medians with a range of indicated values. Statistical significance

was calculated by the nonparametric Mann-Whitney test. Significance is

indicated throughout as follows: *=p ≤ 0.05; **=p ≤ 0.01; ***=p ≤ 0.001.

PCR primer sequences

GFP:

FW: AAGTTCATCTGCACCACCGG; RV: GCGCTCCTGGACGTAGCCTT

Human Alu repeats:

FW: TCACGCCTGTAATCCCAGCA; RV: AGCTGGGACTACAGGCGCCC

- 120 -

Mouse B1 repeats:

FW: GGGCATGGTGGCGCACGCCT; RV: GAGACAGGGTTTCTCTGTGT

Disclosures

Neither myself nor anyone in my laboratory have any financial conflicts of interest

to disclose.

This project was supported by NIH grants AR064460 and AI072571, the Lupus

Research Institute, and the Judith and Stewart Colton Center for Autoimmunity.

The majority of this work was published on June 9th, 2016 (173).

- 121 -

References

1. Choi, J., S. T. Kim, and J. Craft. 2012. The pathogenesis of systemic lupus erythematosus-an update. Curr Opin Immunol 24: 651-657.

2. Pons-Estel, G. J., G. S. Alarcon, L. Scofield, L. Reinlib, and G. S. Cooper. 2010. Understanding the epidemiology and progression of systemic lupus erythematosus. Semin Arthritis Rheum 39: 257-268.

3. Khamashta, M. A. 2006. Systemic lupus erythematosus and pregnancy. Best Pract Res Clin Rheumatol 20: 685-694.

4. Doria, A., M. Canova, M. Tonon, M. Zen, E. Rampudda, N. Bassi, F. Atzeni, S. Zampieri, and A. Ghirardello. 2008. Infections as triggers and complications of systemic lupus erythematosus. Autoimmun Rev 8: 24-28.

5. Casciola-Rosen, L., and A. Rosen. 1997. Ultraviolet light-induced keratinocyte apoptosis: a potential mechanism for the induction of skin lesions and autoantibody production in LE. Lupus 6: 175-180.

6. Bricou, O., O. Taieb, T. Baubet, B. Gal, L. Guillevin, and M. R. Moro. 2006. Stress and coping strategies in systemic lupus erythematosus: a review. Neuroimmunomodulation 13: 283-293.

7. Vasoo, S. 2006. Drug-induced lupus: an update. Lupus 15: 757-761. 8. Deapen, D., A. Escalante, L. Weinrib, D. Horwitz, B. Bachman, P. Roy-

Burman, A. Walker, and T. M. Mack. 1992. A revised estimate of twin concordance in systemic lupus erythematosus. Arthritis Rheum 35: 311-318.

9. Lahita, R. G. 1999. The role of sex hormones in systemic lupus erythematosus. Curr Opin Rheumatol 11: 352-356.

10. Chakravarty, E. F., T. M. Bush, S. Manzi, A. E. Clarke, and M. M. Ward. 2007. Prevalence of adult systemic lupus erythematosus in California and Pennsylvania in 2000: estimates obtained using hospitalization data. Arthritis Rheum 56: 2092-2094.

11. Gladman, D. D., D. Ibanez, and M. B. Urowitz. 2002. Systemic lupus erythematosus disease activity index 2000. J Rheumatol 29: 288-291.

12. Hochberg, M. C. 1997. Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 40: 1725.

13. Hahn, B. H., M. A. McMahon, A. Wilkinson, W. D. Wallace, D. I. Daikh, J. D. Fitzgerald, G. A. Karpouzas, J. T. Merrill, D. J. Wallace, J. Yazdany, R. Ramsey-Goldman, K. Singh, M. Khalighi, S. I. Choi, M. Gogia, S. Kafaja, M. Kamgar, C. Lau, W. J. Martin, S. Parikh, J. Peng, A. Rastogi, W. Chen, J. M. Grossman, and R. American College of. 2012. American College of Rheumatology guidelines for screening, treatment, and management of lupus nephritis. Arthritis Care Res (Hoboken) 64: 797-808.

14. Lee, S. J., E. Silverman, and J. M. Bargman. 2011. The role of antimalarial agents in the treatment of SLE and lupus nephritis. Nat Rev Nephrol 7: 718-729.

- 122 -

15. Furie, R., M. Petri, O. Zamani, R. Cervera, D. J. Wallace, D. Tegzova, J. Sanchez-Guerrero, A. Schwarting, J. T. Merrill, W. W. Chatham, W. Stohl, E. M. Ginzler, D. R. Hough, Z. J. Zhong, W. Freimuth, R. F. van Vollenhoven, and B.-S. Group. 2011. A phase III, randomized, placebo-controlled study of belimumab, a monoclonal antibody that inhibits B lymphocyte stimulator, in patients with systemic lupus erythematosus. Arthritis Rheum 63: 3918-3930.

16. Aringer, M., C. Günther, and M. A. Lee-Kirsch. 2013. Innate immune processes in lupus erythematosus. Clinical Immunology 147: 216-222.

17. Okamoto, A., K. Fujio, N. H. Tsuno, K. Takahashi, and K. Yamamoto. 2012. Kidney-infiltrating CD4+ T-cell clones promote nephritis in lupus-prone mice. Kidney Int 82: 969-979.

18. Zhang, Z., V. C. Kyttaris, and G. C. Tsokos. 2009. The role of IL-23/IL-17 axis in lupus nephritis. J Immunol 183: 3160-3169.

19. Rahman, A., and D. A. Isenberg. 2008. Systemic lupus erythematosus. N Engl J Med 358: 929-939.

20. Nashi, E., Y. Wang, and B. Diamond. 2010. The role of B cells in lupus pathogenesis. Int J Biochem Cell Biol 42: 543-550.

21. Rekvig, O. P., C. Putterman, C. Casu, H. X. Gao, A. Ghirardello, E. S. Mortensen, A. Tincani, and A. Doria. 2012. Autoantibodies in lupus: culprits or passive bystanders? Autoimmun Rev 11: 596-603.

22. Kuida, K., J. A. Lippke, G. Ku, M. W. Harding, D. J. Livingston, M. S. Su, and R. A. Flavell. 1995. Altered cytokine export and apoptosis in mice deficient in interleukin-1 beta converting enzyme. Science 267: 2000-2003.

23. Richards, H. B., M. Satoh, M. Shaw, C. Libert, V. Poli, and W. H. Reeves. 1998. Interleukin 6 dependence of anti-DNA antibody production: evidence for two pathways of autoantibody formation in pristane-induced lupus. J Exp Med 188: 985-990.

24. Lopez de Padilla, C. M., and T. B. Niewold. 2016. The type I interferons: Basic concepts and clinical relevance in immune-mediated inflammatory diseases. Gene 576: 14-21.

25. Sisirak, V., D. Ganguly, K. L. Lewis, C. Couillault, L. Tanaka, S. Bolland, V. D'Agati, K. B. Elkon, and B. Reizis. 2014. Genetic evidence for the role of plasmacytoid dendritic cells in systemic lupus erythematosus. J Exp Med 211: 1969-1976.

26. Domeier, P. P., S. B. Chodisetti, C. Soni, S. L. Schell, M. J. Elias, E. B. Wong, T. K. Cooper, D. Kitamura, and Z. S. Rahman. 2016. IFN-gamma receptor and STAT1 signaling in B cells are central to spontaneous germinal center formation and autoimmunity. J Exp Med 213: 715-732.

27. Aringer, M., and J. S. Smolen. 2008. The role of tumor necrosis factor-alpha in systemic lupus erythematosus. Arthritis Res Ther 10: 202.

28. Yin, Z., G. Bahtiyar, N. Zhang, L. Liu, P. Zhu, M. E. Robert, J. McNiff, M. P. Madaio, and J. Craft. 2002. IL-10 regulates murine lupus. J Immunol 169: 2148-2155.

- 123 -

29. Becker-Merok, A., G. O. Eilertsen, and J. C. Nossent. 2010. Levels of transforming growth factor-beta are low in systemic lupus erythematosus patients with active disease. J Rheumatol 37: 2039-2045.

30. Ohl, K., and K. Tenbrock. 2011. Inflammatory cytokines in systemic lupus erythematosus. J Biomed Biotechnol 2011: 432595.

31. Obermoser, G., and V. Pascual. 2010. The interferon-alpha signature of systemic lupus erythematosus. Lupus 19: 1012-1019.

32. Ganguly, D., G. Chamilos, R. Lande, J. Gregorio, S. Meller, V. Facchinetti, B. Homey, F. J. Barrat, T. Zal, and M. Gilliet. 2009. Self-RNA-antimicrobial peptide complexes activate human dendritic cells through TLR7 and TLR8. J Exp Med 206: 1983-1994.

33. Banchereau, J., F. Briere, C. Caux, J. Davoust, S. Lebecque, Y. J. Liu, B. Pulendran, and K. Palucka. 2000. Immunobiology of dendritic cells. Annu Rev Immunol 18: 767-811.

34. Boule, M. W., C. Broughton, F. Mackay, S. Akira, A. Marshak-Rothstein, and I. R. Rifkin. 2004. Toll-like receptor 9-dependent and -independent dendritic cell activation by chromatin-immunoglobulin G complexes. J Exp Med 199: 1631-1640.

35. Blanco, P., V. Pitard, J. F. Viallard, J. L. Taupin, J. L. Pellegrin, and J. F. Moreau. 2005. Increase in activated CD8+ T lymphocytes expressing perforin and granzyme B correlates with disease activity in patients with systemic lupus erythematosus. Arthritis Rheum 52: 201-211.

36. Cash, H., M. Relle, J. Menke, C. Brochhausen, S. A. Jones, N. Topley, P. R. Galle, and A. Schwarting. 2010. Interleukin 6 (IL-6) deficiency delays lupus nephritis in MRL-Faslpr mice: the IL-6 pathway as a new therapeutic target in treatment of autoimmune kidney disease in systemic lupus erythematosus. J Rheumatol 37: 60-70.

37. Liang, B., D. B. Gardner, D. E. Griswold, P. J. Bugelski, and X. Y. Song. 2006. Anti-interleukin-6 monoclonal antibody inhibits autoimmune responses in a murine model of systemic lupus erythematosus. Immunology 119: 296-305.

38. Fujimoto, M., S. Serada, M. Mihara, Y. Uchiyama, H. Yoshida, N. Koike, Y. Ohsugi, T. Nishikawa, B. Ripley, A. Kimura, T. Kishimoto, and T. Naka. 2008. Interleukin-6 blockade suppresses autoimmune arthritis in mice by the inhibition of inflammatory Th17 responses. Arthritis Rheum 58: 3710-3719.

39. Fukatsu, A., S. Matsuo, H. Tamai, N. Sakamoto, T. Matsuda, and T. Hirano. 1991. Distribution of interleukin-6 in normal and diseased human kidney. Lab Invest 65: 61-66.

40. Alcocer-Varela, J., D. Aleman-Hoey, and D. Alarcon-Segovia. 1992. Interleukin-1 and interleukin-6 activities are increased in the cerebrospinal fluid of patients with CNS lupus erythematosus and correlate with local late T-cell activation markers. Lupus 1: 111-117.

41. Kelchtermans, H., A. Billiau, and P. Matthys. 2008. How interferon-gamma keeps autoimmune diseases in check. Trends Immunol 29: 479-486.

- 124 -

42. Bossie, A., and E. S. Vitetta. 1991. IFN-gamma enhances secretion of IgG2a from IgG2a-committed LPS-stimulated murine B cells: implications for the role of IFN-gamma in class switching. Cell Immunol 135: 95-104.

43. Balomenos, D., R. Rumold, and A. N. Theofilopoulos. 1998. Interferon-gamma is required for lupus-like disease and lymphoaccumulation in MRL-lpr mice. J Clin Invest 101: 364-371.

44. Tokano, Y., S. Morimoto, H. Kaneko, H. Amano, K. Nozawa, Y. Takasaki, and H. Hashimoto. 1999. Levels of IL-12 in the sera of patients with systemic lupus erythematosus (SLE)--relation to Th1- and Th2-derived cytokines. Clin Exp Immunol 116: 169-173.

45. Uhm, W. S., K. Na, G. W. Song, S. S. Jung, T. Lee, M. H. Park, and D. H. Yoo. 2003. Cytokine balance in kidney tissue from lupus nephritis patients. Rheumatology (Oxford) 42: 935-938.

46. Min, D. J., M. L. Cho, C. S. Cho, S. Y. Min, W. U. Kim, S. Y. Yang, J. K. Min, Y. S. Hong, S. H. Lee, S. H. Park, and H. Y. Kim. 2001. Decreased production of interleukin-12 and interferon-gamma is associated with renal involvement in systemic lupus erythematosus. Scand J Rheumatol 30: 159-163.

47. Couper, K. N., D. G. Blount, and E. M. Riley. 2008. IL-10: the master regulator of immunity to infection. J Immunol 180: 5771-5777.

48. Aoki, C. A., A. T. Borchers, M. Li, R. A. Flavell, C. L. Bowlus, A. A. Ansari, and M. E. Gershwin. 2005. Transforming growth factor beta (TGF-beta) and autoimmunity. Autoimmun Rev 4: 450-459.

49. Setoguchi, R., S. Hori, T. Takahashi, and S. Sakaguchi. 2005. Homeostatic maintenance of natural Foxp3(+) CD25(+) CD4(+) regulatory T cells by interleukin (IL)-2 and induction of autoimmune disease by IL-2 neutralization. J Exp Med 201: 723-735.

50. Yang, X. P., K. Ghoreschi, S. M. Steward-Tharp, J. Rodriguez-Canales, J. Zhu, J. R. Grainger, K. Hirahara, H. W. Sun, L. Wei, G. Vahedi, Y. Kanno, J. J. O'Shea, and A. Laurence. 2011. Opposing regulation of the locus encoding IL-17 through direct, reciprocal actions of STAT3 and STAT5. Nat Immunol 12: 247-254.

51. Alcocer-Varela, J., and D. Alarcon-Segovia. 1982. Decreased production of and response to interleukin-2 by cultured lymphocytes from patients with systemic lupus erythematosus. J Clin Invest 69: 1388-1392.

52. Miyara, M., Z. Amoura, C. Parizot, C. Badoual, K. Dorgham, S. Trad, D. Nochy, P. Debre, J. C. Piette, and G. Gorochov. 2005. Global natural regulatory T cell depletion in active systemic lupus erythematosus. J Immunol 175: 8392-8400.

53. Kono, D. H., and A. N. Theofilopoulos. 2006. Genetics of SLE in mice. Springer Semin Immunopathol 28: 83-96.

54. Perry, D., A. Sang, Y. Yin, Y. Y. Zheng, and L. Morel. 2010. Murine models of systemic lupus erythematosus. J Biomed Biotechnol 2011: 271694.

- 125 -

55. Morel, L., U. H. Rudofsky, J. A. Longmate, J. Schiffenbauer, and E. K. Wakeland. 1994. Polygenic control of susceptibility to murine systemic lupus erythematosus. Immunity 1: 219-229.

56. Pisetsky, D. S. 2016. Anti-DNA antibodies - quintessential biomarkers of SLE. Nat Rev Rheumatol 12: 102-110.

57. Gilkeson, G. S., A. M. Pippen, and D. S. Pisetsky. 1995. Induction of cross-reactive anti-dsDNA antibodies in preautoimmune NZB/NZW mice by immunization with bacterial DNA. J Clin Invest 95: 1398-1402.

58. Andrews, B. S., R. A. Eisenberg, A. N. Theofilopoulos, S. Izui, C. B. Wilson, P. J. McConahey, E. D. Murphy, J. B. Roths, and F. J. Dixon. 1978. Spontaneous murine lupus-like syndromes. Clinical and immunopathological manifestations in several strains. J Exp Med 148: 1198-1215.

59. Reap, E. A., D. Leslie, M. Abrahams, R. A. Eisenberg, and P. L. Cohen. 1995. Apoptosis abnormalities of splenic lymphocytes in autoimmune lpr and gld mice. J Immunol 154: 936-943.

60. Kelley, V. E., and J. B. Roths. 1985. Interaction of mutant lpr gene with background strain influences renal disease. Clin Immunol Immunopathol 37: 220-229.

61. Hudgins, C. C., R. T. Steinberg, D. M. Klinman, M. J. Reeves, and A. D. Steinberg. 1985. Studies of consomic mice bearing the Y chromosome of the BXSB mouse. J Immunol 134: 3849-3854.

62. Merino, R., T. Shibata, S. De Kossodo, and S. Izui. 1989. Differential effect of the autoimmune Yaa and lpr genes on the acceleration of lupus-like syndrome in MRL/MpJ mice. Eur J Immunol 19: 2131-2137.

63. Pisitkun, P., J. A. Deane, M. J. Difilippantonio, T. Tarasenko, A. B. Satterthwaite, and S. Bolland. 2006. Autoreactive B cell responses to RNA-related antigens due to TLR7 gene duplication. Science 312: 1669-1672.

64. Subramanian, S., K. Tus, Q. Z. Li, A. Wang, X. H. Tian, J. Zhou, C. Liang, G. Bartov, L. D. McDaniel, X. J. Zhou, R. A. Schultz, and E. K. Wakeland. 2006. A Tlr7 translocation accelerates systemic autoimmunity in murine lupus. Proc Natl Acad Sci U S A 103: 9970-9975.

65. Shen, N., Q. Fu, Y. Deng, X. Qian, J. Zhao, K. M. Kaufman, Y. L. Wu, C. Y. Yu, Y. Tang, J. Y. Chen, W. Yang, M. Wong, A. Kawasaki, N. Tsuchiya, T. Sumida, Y. Kawaguchi, H. S. Howe, M. Y. Mok, S. Y. Bang, F. L. Liu, D. M. Chang, Y. Takasaki, H. Hashimoto, J. B. Harley, J. M. Guthridge, J. M. Grossman, R. M. Cantor, Y. W. Song, S. C. Bae, S. Chen, B. H. Hahn, Y. L. Lau, and B. P. Tsao. 2010. Sex-specific association of X-linked Toll-like receptor 7 (TLR7) with male systemic lupus erythematosus. Proc Natl Acad Sci U S A 107: 15838-15843.

66. Soni, C., E. B. Wong, P. P. Domeier, T. N. Khan, T. Satoh, S. Akira, and Z. S. Rahman. 2014. B cell-intrinsic TLR7 signaling is essential for the development of spontaneous germinal centers. J Immunol 193: 4400-4414.

67. Celhar, T., R. Magalhaes, and A. M. Fairhurst. 2012. TLR7 and TLR9 in SLE: when sensing self goes wrong. Immunol Res 53: 58-77.

- 126 -

68. Nickerson, K. M., S. R. Christensen, J. Shupe, M. Kashgarian, D. Kim, K. Elkon, and M. J. Shlomchik. 2010. TLR9 regulates TLR7- and MyD88-dependent autoantibody production and disease in a murine model of lupus. J Immunol 184: 1840-1848.

69. Lau, C. M., C. Broughton, A. S. Tabor, S. Akira, R. A. Flavell, M. J. Mamula, S. R. Christensen, M. J. Shlomchik, G. A. Viglianti, I. R. Rifkin, and A. Marshak-Rothstein. 2005. RNA-associated autoantigens activate B cells by combined B cell antigen receptor/Toll-like receptor 7 engagement. J Exp Med 202: 1171-1177.

70. Leadbetter, E. A., I. R. Rifkin, A. M. Hohlbaum, B. C. Beaudette, M. J. Shlomchik, and A. Marshak-Rothstein. 2002. Chromatin-IgG complexes activate B cells by dual engagement of IgM and Toll-like receptors. Nature 416: 603-607.

71. Desnues, B., A. B. Macedo, A. Roussel-Queval, J. Bonnardel, S. Henri, O. Demaria, and L. Alexopoulou. 2014. TLR8 on dendritic cells and TLR9 on B cells restrain TLR7-mediated spontaneous autoimmunity in C57BL/6 mice. Proc Natl Acad Sci U S A 111: 1497-1502.

72. Giltiay, N. V., C. P. Chappell, X. Sun, N. Kolhatkar, T. H. Teal, A. E. Wiedeman, J. Kim, L. Tanaka, M. B. Buechler, J. A. Hamerman, T. Imanishi-Kari, E. A. Clark, and K. B. Elkon. 2013. Overexpression of TLR7 promotes cell-intrinsic expansion and autoantibody production by transitional T1 B cells. J Exp Med 210: 2773-2789.

73. Lee, Y. H., S. J. Choi, J. D. Ji, and G. G. Song. 2016. Association between toll-like receptor polymorphisms and systemic lupus erythematosus: a meta-analysis update. Lupus 25: 593-601.

74. Rekvig, O. P. 2015. The anti-DNA antibody: origin and impact, dogmas and controversies. Nat Rev Rheumatol 11: 530-540.

75. Tabeta, K., K. Hoebe, E. M. Janssen, X. Du, P. Georgel, K. Crozat, S. Mudd, N. Mann, S. Sovath, J. Goode, L. Shamel, A. A. Herskovits, D. A. Portnoy, M. Cooke, L. M. Tarantino, T. Wiltshire, B. E. Steinberg, S. Grinstein, and B. Beutler. 2006. The Unc93b1 mutation 3d disrupts exogenous antigen presentation and signaling via Toll-like receptors 3, 7 and 9. Nat Immunol 7: 156-164.

76. Christensen, S. R., J. Shupe, K. Nickerson, M. Kashgarian, R. A. Flavell, and M. J. Shlomchik. 2006. Toll-like receptor 7 and TLR9 dictate autoantibody specificity and have opposing inflammatory and regulatory roles in a murine model of lupus. Immunity 25: 417-428.

77. Deane, J. A., P. Pisitkun, R. S. Barrett, L. Feigenbaum, T. Town, J. M. Ward, R. A. Flavell, and S. Bolland. 2007. Control of toll-like receptor 7 expression is essential to restrict autoimmunity and dendritic cell proliferation. Immunity 27: 801-810.

78. Satoh, M., A. Kumar, Y. S. Kanwar, and W. H. Reeves. 1995. Anti-nuclear antibody production and immune-complex glomerulonephritis in BALB/c mice treated with pristane. Proc Natl Acad Sci U S A 92: 10934-10938.

- 127 -

79. Satoh, M., and W. H. Reeves. 1994. Induction of lupus-associated autoantibodies in BALB/c mice by intraperitoneal injection of pristane. J Exp Med 180: 2341-2346.

80. Satoh, M., H. B. Richards, V. M. Shaheen, H. Yoshida, M. Shaw, J. O. Naim, P. H. Wooley, and W. H. Reeves. 2000. Widespread susceptibility among inbred mouse strains to the induction of lupus autoantibodies by pristane. Clin Exp Immunol 121: 399-405.

81. Nacionales, D. C., K. M. Kelly-Scumpia, P. Y. Lee, J. S. Weinstein, R. Lyons, E. Sobel, M. Satoh, and W. H. Reeves. 2007. Deficiency of the type I interferon receptor protects mice from experimental lupus. Arthritis Rheum 56: 3770-3783.

82. Thibault, D. L., K. L. Graham, L. Y. Lee, I. Balboni, P. J. Hertzog, and P. J. Utz. 2009. Type I interferon receptor controls B-cell expression of nucleic acid-sensing Toll-like receptors and autoantibody production in a murine model of lupus. Arthritis Res Ther 11: R112.

83. Banchereau, R., S. Hong, B. Cantarel, N. Baldwin, J. Baisch, M. Edens, A. M. Cepika, P. Acs, J. Turner, E. Anguiano, P. Vinod, S. Kahn, G. Obermoser, D. Blankenship, E. Wakeland, L. Nassi, A. Gotte, M. Punaro, Y. J. Liu, J. Banchereau, J. Rossello-Urgell, T. Wright, and V. Pascual. 2016. Personalized Immunomonitoring Uncovers Molecular Networks that Stratify Lupus Patients. Cell.

84. Elkon, K. B., and A. Wiedeman. 2012. Type I IFN system in the development and manifestations of SLE. Curr Opin Rheumatol 24: 499-505.

85. Pascual, V., L. Farkas, and J. Banchereau. 2006. Systemic lupus erythematosus: all roads lead to type I interferons. Curr Opin Immunol 18: 676-682.

86. Ronnblom, L., and V. Pascual. 2008. The innate immune system in SLE: type I interferons and dendritic cells. Lupus 17: 394-399.

87. Lee, P. Y., J. S. Weinstein, D. C. Nacionales, P. O. Scumpia, Y. Li, E. Butfiloski, N. van Rooijen, L. Moldawer, M. Satoh, and W. H. Reeves. 2008. A novel type I IFN-producing cell subset in murine lupus. J Immunol 180: 5101-5108.

88. Lee, P. Y., Y. Kumagai, Y. Li, O. Takeuchi, H. Yoshida, J. Weinstein, E. S. Kellner, D. Nacionales, T. Barker, K. Kelly-Scumpia, N. van Rooijen, H. Kumar, T. Kawai, M. Satoh, S. Akira, and W. H. Reeves. 2008. TLR7-dependent and FcgammaR-independent production of type I interferon in experimental mouse lupus. J Exp Med 205: 2995-3006.

89. Yung, S., and T. M. Chan. 2015. Mechanisms of kidney injury in lupus nephritis – the role of anti-dsDNA antibodies. Front Immunol 6: 475.

90. Ghodke-Puranik, Y., and T. B. Niewold. 2015. Immunogenetics of systemic lupus erythematosus: A comprehensive review. J Autoimmun 64: 125-136.

91. Mohan, C., and C. Putterman. 2015. Genetics and pathogenesis of systemic lupus erythematosus and lupus nephritis. Nat Rev Nephrol 11: 329-341.

- 128 -

92. Avalos, A. M., F. Meyer-Wentrup, and H. L. Ploegh. 2014. B-cell receptor signaling in lymphoid malignancies and autoimmunity. Adv Immunol 123: 1-49.

93. Graham, R. R., W. Ortmann, P. Rodine, K. Espe, C. Langefeld, E. Lange, A. Williams, S. Beck, C. Kyogoku, K. Moser, P. Gaffney, P. K. Gregersen, L. A. Criswell, J. B. Harley, and T. W. Behrens. 2007. Specific combinations of HLA-DR2 and DR3 class II haplotypes contribute graded risk for disease susceptibility and autoantibodies in human SLE. Eur J Hum Genet 15: 823-830.

94. Tan, W., K. Sunahori, J. Zhao, Y. Deng, K. M. Kaufman, J. A. Kelly, C. D. Langefeld, A. H. Williams, M. E. Comeau, J. T. Ziegler, M. C. Marion, S. C. Bae, J. H. Lee, J. S. Lee, D. M. Chang, Y. W. Song, C. Y. Yu, R. P. Kimberly, J. C. Edberg, E. E. Brown, M. A. Petri, R. Ramsey-Goldman, L. M. Vila, J. D. Reveille, M. E. Alarcon-Riquelme, J. B. Harley, S. A. Boackle, A. M. Stevens, R. H. Scofield, J. T. Merrill, B. I. Freedman, J. M. Anaya, L. A. Criswell, C. O. Jacob, T. J. Vyse, T. B. Niewold, P. M. Gaffney, K. L. Moser, G. S. Gilkeson, D. L. Kamen, J. A. James, J. M. Grossman, B. H. Hahn, G. C. Tsokos, B. P. Tsao, G. S. Alarcon, B. Network, and G. Network. 2011. Association of PPP2CA polymorphisms with systemic lupus erythematosus susceptibility in multiple ethnic groups. Arthritis Rheum 63: 2755-2763.

95. Dang, J., S. Shan, J. Li, H. Zhao, Q. Xin, Y. Liu, X. Bian, and Q. Liu. 2014. Gene-gene interactions of IRF5, STAT4, IKZF1 and ETS1 in systemic lupus erythematosus. Tissue Antigens 83: 401-408.

96. Karassa, F. B., T. A. Trikalinos, J. P. Ioannidis, and R.-S. L. E. m.-a. i. Fc gamma. 2003. The Fc gamma RIIIA-F158 allele is a risk factor for the development of lupus nephritis: a meta-analysis. Kidney Int 63: 1475-1482.

97. Kim-Howard, X., A. K. Maiti, J. M. Anaya, G. R. Bruner, E. Brown, J. T. Merrill, J. C. Edberg, M. A. Petri, J. D. Reveille, R. Ramsey-Goldman, G. S. Alarcon, T. J. Vyse, G. Gilkeson, R. P. Kimberly, J. A. James, J. M. Guthridge, J. B. Harley, and S. K. Nath. 2010. ITGAM coding variant (rs1143679) influences the risk of renal disease, discoid rash and immunological manifestations in patients with systemic lupus erythematosus with European ancestry. Ann Rheum Dis 69: 1329-1332.

98. Chang, A., S. G. Henderson, D. Brandt, N. Liu, R. Guttikonda, C. Hsieh, N. Kaverina, T. O. Utset, S. M. Meehan, R. J. Quigg, E. Meffre, and M. R. Clark. 2011. In situ B cell-mediated immune responses and tubulointerstitial inflammation in human lupus nephritis. J Immunol 186: 1849-1860.

99. Orme, J., and C. Mohan. 2012. Macrophages and neutrophils in SLE-An online molecular catalog. Autoimmun Rev 11: 365-372.

100. Brown, E. E., J. C. Edberg, and R. P. Kimberly. 2007. Fc receptor genes and the systemic lupus erythematosus diathesis. Autoimmunity 40: 567-581.

101. Zhou, T. B., Y. G. Liu, N. Lin, Y. H. Qin, K. Huang, M. B. Shao, and D. D. Peng. 2012. Relationship between angiotensin-converting enzyme

- 129 -

insertion/deletion gene polymorphism and systemic lupus erythematosus/lupus nephritis: a systematic review and metaanalysis. J Rheumatol 39: 686-693.

102. Manderson, A. P., M. Botto, and M. J. Walport. 2004. The role of complement in the development of systemic lupus erythematosus. Annu Rev Immunol 22: 431-456.

103. Munoz, L. E., K. Lauber, M. Schiller, A. A. Manfredi, and M. Herrmann. 2010. The role of defective clearance of apoptotic cells in systemic autoimmunity. Nat Rev Rheumatol 6: 280-289.

104. Caielli, S., S. Athale, B. Domic, E. Murat, M. Chandra, R. Banchereau, J. Baisch, K. Phelps, S. Clayton, M. Gong, T. Wright, M. Punaro, K. Palucka, C. Guiducci, J. Banchereau, and V. Pascual. 2016. Oxidized mitochondrial nucleoids released by neutrophils drive type I interferon production in human lupus. J Exp Med.

105. Garcia-Romo, G. S., S. Caielli, B. Vega, J. Connolly, F. Allantaz, Z. Xu, M. Punaro, J. Baisch, C. Guiducci, R. L. Coffman, F. J. Barrat, J. Banchereau, and V. Pascual. 2011. Netting neutrophils are major inducers of type I IFN production in pediatric systemic lupus erythematosus. Sci Transl Med 3: 73ra20.

106. Lande, R., D. Ganguly, V. Facchinetti, L. Frasca, C. Conrad, J. Gregorio, S. Meller, G. Chamilos, R. Sebasigari, V. Riccieri, R. Bassett, H. Amuro, S. Fukuhara, T. Ito, Y. J. Liu, and M. Gilliet. 2011. Neutrophils activate plasmacytoid dendritic cells by releasing self-DNA-peptide complexes in systemic lupus erythematosus. Sci Transl Med 3: 73ra19.

107. Lood, C., L. P. Blanco, M. M. Purmalek, C. Carmona-Rivera, S. S. De Ravin, C. K. Smith, H. L. Malech, J. A. Ledbetter, K. B. Elkon, and M. J. Kaplan. 2016. Neutrophil extracellular traps enriched in oxidized mitochondrial DNA are interferogenic and contribute to lupus-like disease. Nat Med 22: 146-153.

108. Linnane, A. W., S. Marzuki, T. Ozawa, and M. Tanaka. 1989. Mitochondrial DNA mutations as an important contributor to ageing and degenerative diseases. Lancet 1: 642-645.

109. Brinkmann, V., and A. Zychlinsky. 2012. Neutrophil extracellular traps: is immunity the second function of chromatin? J Cell Biol 198: 773-783.

110. Brinkmann, V., U. Reichard, C. Goosmann, B. Fauler, Y. Uhlemann, D. S. Weiss, Y. Weinrauch, and A. Zychlinsky. 2004. Neutrophil extracellular traps kill bacteria. Science 303: 1532-1535.

111. Thomas, M. P., J. Whangbo, G. McCrossan, A. J. Deutsch, K. Martinod, M. Walch, and J. Lieberman. 2014. Leukocyte protease binding to nucleic acids promotes nuclear localization and cleavage of nucleic acid binding proteins. J Immunol 192: 5390-5397.

112. Atianand, M. K., and K. A. Fitzgerald. 2013. Molecular basis of DNA recognition in the immune system. J Immunol 190: 1911-1918.

113. Ohkouchi, S., M. Shibata, M. Sasaki, M. Koike, P. Safig, C. Peters, S. Nagata, and Y. Uchiyama. 2013. Biogenesis and proteolytic processing of lysosomal DNase II. PLoS One 8: e59148.

- 130 -

114. Crow, Y. J. 2010. The story of DNase II: a stifled death-wish leads to self-harm. Eur J Immunol 40: 2376-2378.

115. Evans, C. J., and R. J. Aguilera. 2003. DNase II: genes, enzymes and function. Gene 322: 1-15.

116. Kawane, K., M. Ohtani, K. Miwa, T. Kizawa, Y. Kanbara, Y. Yoshioka, H. Yoshikawa, and S. Nagata. 2006. Chronic polyarthritis caused by mammalian DNA that escapes from degradation in macrophages. Nature 443: 998-1002.

117. Ahn, J., D. Gutman, S. Saijo, and G. N. Barber. 2012. STING manifests self DNA-dependent inflammatory disease. Proc Natl Acad Sci U S A 109: 19386-19391.

118. Tanaka, Y., and Z. J. Chen. 2012. STING specifies IRF3 phosphorylation by TBK1 in the cytosolic DNA signaling pathway. Sci Signal 5: ra20.

119. Ablasser, A., M. Goldeck, T. Cavlar, T. Deimling, G. Witte, I. Rohl, K. P. Hopfner, J. Ludwig, and V. Hornung. 2013. cGAS produces a 2'-5'-linked cyclic dinucleotide second messenger that activates STING. Nature 498: 380-384.

120. Motani, K., S. Ito, and S. Nagata. 2015. DNA-Mediated Cyclic GMP-AMP Synthase-Dependent and -Independent Regulation of Innate Immune Responses. J Immunol 194: 4914-4923.

121. Mazur, D. J., and F. W. Perrino. 2001. Structure and expression of the TREX1 and TREX2 3' --> 5' exonuclease genes. J Biol Chem 276: 14718-14727.

122. Stetson, D. B., J. S. Ko, T. Heidmann, and R. Medzhitov. 2008. Trex1 prevents cell-intrinsic initiation of autoimmunity. Cell 134: 587-598.

123. Gray, E. E., P. M. Treuting, J. J. Woodward, and D. B. Stetson. 2015. Cutting Edge: cGAS Is Required for Lethal Autoimmune Disease in the Trex1-Deficient Mouse Model of Aicardi-Goutieres Syndrome. J Immunol 195: 1939-1943.

124. Morita, M., G. Stamp, P. Robins, A. Dulic, I. Rosewell, G. Hrivnak, G. Daly, T. Lindahl, and D. E. Barnes. 2004. Gene-targeted mice lacking the Trex1 (DNase III) 3'-->5' DNA exonuclease develop inflammatory myocarditis. Mol Cell Biol 24: 6719-6727.

125. Pereira-Lopes, S., T. Celhar, G. Sans-Fons, M. Serra, A. M. Fairhurst, J. Lloberas, and A. Celada. 2013. The exonuclease Trex1 restrains macrophage proinflammatory activation. J Immunol 191: 6128-6135.

126. Ahn, J., P. Ruiz, and G. N. Barber. 2014. Intrinsic self-DNA triggers inflammatory disease dependent on STING. J Immunol 193: 4634-4642.

127. Aicardi, J., and F. Goutieres. 1984. A progressive familial encephalopathy in infancy with calcifications of the basal ganglia and chronic cerebrospinal fluid lymphocytosis. Ann Neurol 15: 49-54.

128. Crow, Y. J., and J. Rehwinkel. 2009. Aicardi-Goutieres syndrome and related phenotypes: linking nucleic acid metabolism with autoimmunity. Hum Mol Genet 18: R130-136.

129. Tolmie, J. L., P. Shillito, R. Hughes-Benzie, and J. B. Stephenson. 1995. The Aicardi-Goutieres syndrome (familial, early onset encephalopathy with

- 131 -

calcifications of the basal ganglia and chronic cerebrospinal fluid lymphocytosis). J Med Genet 32: 881-884.

130. Lebon, P., J. Badoual, G. Ponsot, F. Goutieres, F. Hemeury-Cukier, and J. Aicardi. 1988. Intrathecal synthesis of interferon-alpha in infants with progressive familial encephalopathy. J Neurol Sci 84: 201-208.

131. Rice, G. I., G. M. Forte, M. Szynkiewicz, D. S. Chase, A. Aeby, M. S. Abdel-Hamid, S. Ackroyd, R. Allcock, K. M. Bailey, U. Balottin, C. Barnerias, G. Bernard, C. Bodemer, M. P. Botella, C. Cereda, K. E. Chandler, L. Dabydeen, R. C. Dale, C. De Laet, C. G. De Goede, M. Del Toro, L. Effat, N. N. Enamorado, E. Fazzi, B. Gener, M. Haldre, J. P. Lin, J. H. Livingston, C. M. Lourenco, W. Marques, Jr., P. Oades, P. Peterson, M. Rasmussen, A. Roubertie, J. L. Schmidt, S. A. Shalev, R. Simon, R. Spiegel, K. J. Swoboda, S. A. Temtamy, G. Vassallo, C. N. Vilain, J. Vogt, V. Wermenbol, W. P. Whitehouse, D. Soler, I. Olivieri, S. Orcesi, M. S. Aglan, M. S. Zaki, G. M. Abdel-Salam, A. Vanderver, K. Kisand, F. Rozenberg, P. Lebon, and Y. J. Crow. 2013. Assessment of interferon-related biomarkers in Aicardi-Goutieres syndrome associated with mutations in TREX1, RNASEH2A, RNASEH2B, RNASEH2C, SAMHD1, and ADAR: a case-control study. Lancet Neurol 12: 1159-1169.

132. Crow, Y. J., A. Leitch, B. E. Hayward, A. Garner, R. Parmar, E. Griffith, M. Ali, C. Semple, J. Aicardi, R. Babul-Hirji, C. Baumann, P. Baxter, E. Bertini, K. E. Chandler, D. Chitayat, D. Cau, C. Dery, E. Fazzi, C. Goizet, M. D. King, J. Klepper, D. Lacombe, G. Lanzi, H. Lyall, M. L. Martinez-Frias, M. Mathieu, C. McKeown, A. Monier, Y. Oade, O. W. Quarrell, C. D. Rittey, R. C. Rogers, A. Sanchis, J. B. Stephenson, U. Tacke, M. Till, J. L. Tolmie, P. Tomlin, T. Voit, B. Weschke, C. G. Woods, P. Lebon, D. T. Bonthron, C. P. Ponting, and A. P. Jackson. 2006. Mutations in genes encoding ribonuclease H2 subunits cause Aicardi-Goutieres syndrome and mimic congenital viral brain infection. Nat Genet 38: 910-916.

133. Rice, G. I., J. Bond, A. Asipu, R. L. Brunette, I. W. Manfield, I. M. Carr, J. C. Fuller, R. M. Jackson, T. Lamb, T. A. Briggs, M. Ali, H. Gornall, L. R. Couthard, A. Aeby, S. P. Attard-Montalto, E. Bertini, C. Bodemer, K. Brockmann, L. A. Brueton, P. C. Corry, I. Desguerre, E. Fazzi, A. G. Cazorla, B. Gener, B. C. Hamel, A. Heiberg, M. Hunter, M. S. van der Knaap, R. Kumar, L. Lagae, P. G. Landrieu, C. M. Lourenco, D. Marom, M. F. McDermott, W. van der Merwe, S. Orcesi, J. S. Prendiville, M. Rasmussen, S. A. Shalev, D. M. Soler, M. Shinawi, R. Spiegel, T. Y. Tan, A. Vanderver, E. L. Wakeling, E. Wassmer, E. Whittaker, P. Lebon, D. B. Stetson, D. T. Bonthron, and Y. J. Crow. 2009. Mutations involved in Aicardi-Goutieres syndrome implicate SAMHD1 as regulator of the innate immune response. Nat Genet 41: 829-832.

134. Rice, G. I., P. R. Kasher, G. M. Forte, N. M. Mannion, S. M. Greenwood, M. Szynkiewicz, J. E. Dickerson, S. S. Bhaskar, M. Zampini, T. A. Briggs, E. M. Jenkinson, C. A. Bacino, R. Battini, E. Bertini, P. A. Brogan, L. A. Brueton, M. Carpanelli, C. De Laet, P. de Lonlay, M. del Toro, I. Desguerre, E. Fazzi, A. Garcia-Cazorla, A. Heiberg, M. Kawaguchi, R.

- 132 -

Kumar, J. P. Lin, C. M. Lourenco, A. M. Male, W. Marques, Jr., C. Mignot, I. Olivieri, S. Orcesi, P. Prabhakar, M. Rasmussen, R. A. Robinson, F. Rozenberg, J. L. Schmidt, K. Steindl, T. Y. Tan, W. G. van der Merwe, A. Vanderver, G. Vassallo, E. L. Wakeling, E. Wassmer, E. Whittaker, J. H. Livingston, P. Lebon, T. Suzuki, P. J. McLaughlin, L. P. Keegan, M. A. O'Connell, S. C. Lovell, and Y. J. Crow. 2012. Mutations in ADAR1 cause Aicardi-Goutieres syndrome associated with a type I interferon signature. Nat Genet 44: 1243-1248.

135. Oda, H., K. Nakagawa, J. Abe, T. Awaya, M. Funabiki, A. Hijikata, R. Nishikomori, M. Funatsuka, Y. Ohshima, Y. Sugawara, T. Yasumi, H. Kato, T. Shirai, O. Ohara, T. Fujita, and T. Heike. 2014. Aicardi-Goutieres syndrome is caused by IFIH1 mutations. Am J Hum Genet 95: 121-125.

136. Crow, Y. J., B. E. Hayward, R. Parmar, P. Robins, A. Leitch, M. Ali, D. N. Black, H. van Bokhoven, H. G. Brunner, B. C. Hamel, P. C. Corry, F. M. Cowan, S. G. Frints, J. Klepper, J. H. Livingston, S. A. Lynch, R. F. Massey, J. F. Meritet, J. L. Michaud, G. Ponsot, T. Voit, P. Lebon, D. T. Bonthron, A. P. Jackson, D. E. Barnes, and T. Lindahl. 2006. Mutations in the gene encoding the 3'-5' DNA exonuclease TREX1 cause Aicardi-Goutieres syndrome at the AGS1 locus. Nat Genet 38: 917-920.

137. Grieves, J. L., J. M. Fye, S. Harvey, J. M. Grayson, T. Hollis, and F. W. Perrino. 2015. Exonuclease TREX1 degrades double-stranded DNA to prevent spontaneous lupus-like inflammatory disease. Proc Natl Acad Sci U S A 112: 5117-5122.

138. Su, W. P., C. Perniciaro, R. S. Rogers, 3rd, and J. W. White, Jr. 1994. Chilblain lupus erythematosus (lupus pernio): clinical review of the Mayo Clinic experience and proposal of diagnostic criteria. Cutis 54: 395-399.

139. Franceschini, F., P. Calzavara-Pinton, L. Valsecchi, M. Quinzanini, C. Zane, F. Facchetti, P. Airo, and R. Cattaneo. 1999. Chilblain lupus erythematosus is associated with antibodies to SSA/Ro. Adv Exp Med Biol 455: 167-171.

140. Franceschini, F., and I. Cavazzana. 2005. Anti-Ro/SSA and La/SSB antibodies. Autoimmunity 38: 55-63.

141. Brucato, A., R. Cimaz, R. Caporali, V. Ramoni, and J. Buyon. 2011. Pregnancy outcomes in patients with autoimmune diseases and anti-Ro/SSA antibodies. Clin Rev Allergy Immunol 40: 27-41.

142. Harley, I. T., K. M. Kaufman, C. D. Langefeld, J. B. Harley, and J. A. Kelly. 2009. Genetic susceptibility to SLE: new insights from fine mapping and genome-wide association studies. Nat Rev Genet 10: 285-290.

143. Love, J. D., and R. R. Hewitt. 1979. The relationship between human serum and human pancreatic DNase I. J Biol Chem 254: 12588-12594.

144. Napirei, M., H. Karsunky, B. Zevnik, H. Stephan, H. G. Mannherz, and T. Moroy. 2000. Features of systemic lupus erythematosus in Dnase1-deficient mice. Nat Genet 25: 177-181.

145. Yasutomo, K., T. Horiuchi, S. Kagami, H. Tsukamoto, C. Hashimura, M. Urushihara, and Y. Kuroda. 2001. Mutation of DNASE1 in people with systemic lupus erythematosus. Nat Genet 28: 313-314.

- 133 -

146. Martinez-Valle, F., E. Balada, J. Ordi-Ros, S. Bujan-Rivas, A. Sellas-Fernandez, and M. Vilardell-Tarres. 2010. DNase1 activity in systemic lupus erythematosus patients with and without nephropathy. Rheumatol Int 30: 1601-1604.

147. Fismen, S., E. S. Mortensen, and O. P. Rekvig. 2011. Nuclease deficiencies promote end-stage lupus nephritis but not nephritogenic autoimmunity in (NZB x NZW) F1 mice. Immunol Cell Biol 89: 90-99.

148. Davis, J. C., Jr., S. Manzi, C. Yarboro, J. Rairie, I. McInnes, D. Averthelyi, D. Sinicropi, V. G. Hale, J. Balow, H. Austin, D. T. Boumpas, and J. H. Klippel. 1999. Recombinant human Dnase I (rhDNase) in patients with lupus nephritis. Lupus 8: 68-76.

149. Parsiegla, G., C. Noguere, L. Santell, R. A. Lazarus, and Y. Bourne. 2012. The structure of human DNase I bound to magnesium and phosphate ions points to a catalytic mechanism common to members of the DNase I-like superfamily. Biochemistry 51: 10250-10258.

150. Al-Mayouf, S. M., A. Sunker, R. Abdwani, S. A. Abrawi, F. Almurshedi, N. Alhashmi, A. Al Sonbul, W. Sewairi, A. Qari, E. Abdallah, M. Al-Owain, S. Al Motywee, H. Al-Rayes, M. Hashem, H. Khalak, L. Al-Jebali, and F. S. Alkuraya. 2011. Loss-of-function variant in DNASE1L3 causes a familial form of systemic lupus erythematosus. Nat Genet 43: 1186-1188.

151. Ozcakar, Z. B., J. Foster, 2nd, O. Diaz-Horta, O. Kasapcopur, Y. S. Fan, F. Yalcinkaya, and M. Tekin. 2013. DNASE1L3 mutations in hypocomplementemic urticarial vasculitis syndrome. Arthritis Rheum 65: 2183-2189.

152. Braun, A., J. Sis, R. Max, K. Mueller, C. Fiehn, M. Zeier, and K. Andrassy. 2007. Anti-chromatin and anti-C1q antibodies in systemic lupus erythematosus compared to other systemic autoimmune diseases. Scand J Rheumatol 36: 291-298.

153. Nauta, A. J., L. A. Trouw, M. R. Daha, O. Tijsma, R. Nieuwland, W. J. Schwaeble, A. R. Gingras, A. Mantovani, E. C. Hack, and A. Roos. 2002. Direct binding of C1q to apoptotic cells and cell blebs induces complement activation. Eur J Immunol 32: 1726-1736.

154. Orbai, A. M., L. Truedsson, G. Sturfelt, O. Nived, H. Fang, G. S. Alarcon, C. Gordon, J. Merrill, P. R. Fortin, I. N. Bruce, D. A. Isenberg, D. J. Wallace, R. Ramsey-Goldman, S. C. Bae, J. G. Hanly, J. Sanchez-Guerrero, A. E. Clarke, C. B. Aranow, S. Manzi, M. B. Urowitz, D. D. Gladman, K. C. Kalunian, M. I. Costner, V. P. Werth, A. Zoma, S. Bernatsky, G. Ruiz-Irastorza, M. A. Khamashta, S. Jacobsen, J. P. Buyon, P. Maddison, M. A. Dooley, R. F. Van Vollenhoven, E. Ginzler, T. Stoll, C. Peschken, J. L. Jorizzo, J. P. Callen, S. S. Lim, B. J. Fessler, M. Inanc, D. L. Kamen, A. Rahman, K. Steinsson, A. G. Franks, Jr., L. Sigler, S. Hameed, N. Pham, R. Brey, M. H. Weisman, G. McGwin, Jr., L. S. Magder, and M. Petri. 2015. Anti-C1q antibodies in systemic lupus erythematosus. Lupus 24: 42-49.

155. Walport, M. J., K. A. Davies, and M. Botto. 1998. C1q and systemic lupus erythematosus. Immunobiology 199: 265-285.

- 134 -

156. Carbonella, A., G. Mancano, E. Gremese, F. S. Alkuraya, N. Patel, F. Gurrieri, and G. Ferraccioli. 2016. An autosomal recessive DNASE1L3-related autoimmune disease with unusual clinical presentation mimicking systemic lupus erythematosus. Lupus.

157. Harley, J. B., M. E. Alarcon-Riquelme, L. A. Criswell, C. O. Jacob, R. P. Kimberly, K. L. Moser, B. P. Tsao, T. J. Vyse, C. D. Langefeld, S. K. Nath, J. M. Guthridge, B. L. Cobb, D. B. Mirel, M. C. Marion, A. H. Williams, J. Divers, W. Wang, S. G. Frank, B. Namjou, S. B. Gabriel, A. T. Lee, P. K. Gregersen, T. W. Behrens, K. E. Taylor, M. Fernando, R. Zidovetzki, P. M. Gaffney, J. C. Edberg, J. D. Rioux, J. O. Ojwang, J. A. James, J. T. Merrill, G. S. Gilkeson, M. F. Seldin, H. Yin, E. C. Baechler, Q. Z. Li, E. K. Wakeland, G. R. Bruner, K. M. Kaufman, and J. A. Kelly. 2008. Genome-wide association scan in women with systemic lupus erythematosus identifies susceptibility variants in ITGAM, PXK, KIAA1542 and other loci. Nat Genet 40: 204-210.

158. Kim, E. M., S. Y. Bang, I. Kim, H. D. Shin, B. L. Park, H. S. Lee, and S. C. Bae. 2012. Different genetic effect of PXK on systemic lupus erythematosus in the Korean population. Rheumatol Int 32: 277-280.

159. Mayes, M. D., L. Bossini-Castillo, O. Gorlova, J. E. Martin, X. Zhou, W. V. Chen, S. Assassi, J. Ying, F. K. Tan, F. C. Arnett, J. D. Reveille, S. Guerra, M. Teruel, F. D. Carmona, P. K. Gregersen, A. T. Lee, E. Lopez-Isac, E. Ochoa, P. Carreira, C. P. Simeon, I. Castellvi, M. A. Gonzalez-Gay, A. Zhernakova, L. Padyukov, M. Alarcon-Riquelme, C. Wijmenga, M. Brown, L. Beretta, G. Riemekasten, T. Witte, N. Hunzelmann, A. Kreuter, J. H. Distler, A. E. Voskuyl, A. J. Schuerwegh, R. Hesselstrand, A. Nordin, P. Airo, C. Lunardi, P. Shiels, J. M. van Laar, A. Herrick, J. Worthington, C. Denton, F. M. Wigley, L. K. Hummers, J. Varga, M. E. Hinchcliff, M. Baron, M. Hudson, J. E. Pope, D. E. Furst, D. Khanna, K. Phillips, E. Schiopu, B. M. Segal, J. A. Molitor, R. M. Silver, V. D. Steen, R. W. Simms, R. A. Lafyatis, B. J. Fessler, T. M. Frech, F. Alkassab, P. Docherty, E. Kaminska, N. Khalidi, H. N. Jones, J. Markland, D. Robinson, J. Broen, T. R. Radstake, C. Fonseca, B. P. Koeleman, and J. Martin. 2014. Immunochip analysis identifies multiple susceptibility loci for systemic sclerosis. Am J Hum Genet 94: 47-61.

160. Zochling, J., F. Newell, J. C. Charlesworth, P. Leo, J. Stankovich, A. Cortes, Y. Zhou, W. Stevens, J. Sahhar, J. Roddy, P. Nash, K. Tymms, M. Rischmueller, S. Lester, S. Proudman, and M. A. Brown. 2014. An Immunochip-based interrogation of scleroderma susceptibility variants identifies a novel association at DNASE1L3. Arthritis Res Ther 16: 438.

161. Ueki, M., J. Fujihara, H. Takeshita, K. Kimura-Kataoka, R. Iida, I. Yuasa, H. Kato, and T. Yasuda. 2011. Global genetic analysis of all single nucleotide polymorphisms in exons of the human deoxyribonuclease I-like 3 gene and their effect on its catalytic activity. Electrophoresis 32: 1465-1472.

162. Ueki, M., K. Kimura-Kataoka, H. Takeshita, J. Fujihara, R. Iida, R. Sano, T. Nakajima, Y. Kominato, Y. Kawai, and T. Yasuda. Evaluation of all non-synonymous single nucleotide polymorphisms (SNPs) in the genes

- 135 -

encoding human deoxyribonuclease I and I-like 3 as a functional SNP potentially implicated in autoimmunity. FEBS J 281: 376-390.

163. Baron, W. F., C. Q. Pan, S. A. Spencer, A. M. Ryan, R. A. Lazarus, and K. P. Baker. 1998. Cloning and characterization of an actin-resistant DNase I-like endonuclease secreted by macrophages. Gene 215: 291-301.

164. Liu, Q. Y., S. Pandey, R. K. Singh, W. Lin, M. Ribecco, H. Borowy-Borowski, B. Smith, J. LeBlanc, P. R. Walker, and M. Sikorska. 1998. DNaseY: a rat DNaseI-like gene coding for a constitutively expressed chromatin-bound endonuclease. Biochemistry 37: 10134-10143.

165. Kabsch, W., H. G. Mannherz, D. Suck, E. F. Pai, and K. C. Holmes. 1990. Atomic structure of the actin:DNase I complex. Nature 347: 37-44.

166. Pan, C. Q., T. H. Dodge, D. L. Baker, W. S. Prince, D. V. Sinicropi, and R. A. Lazarus. 1998. Improved potency of hyperactive and actin-resistant human DNase I variants for treatment of cystic fibrosis and systemic lupus erythematosus. J Biol Chem 273: 18374-18381.

167. Wilber, A., M. Lu, and M. C. Schneider. 2002. Deoxyribonuclease I-like III is an inducible macrophage barrier to liposomal transfection. Mol Ther 6: 35-42.

168. Napirei, M., S. Ludwig, J. Mezrhab, T. Klockl, and H. G. Mannherz. 2009. Murine serum nucleases--contrasting effects of plasmin and heparin on the activities of DNase1 and DNase1-like 3 (DNase1l3). FEBS J 276: 1059-1073.

169. Napirei, M., S. Wulf, D. Eulitz, H. G. Mannherz, and T. Kloeckl. 2005. Comparative characterization of rat deoxyribonuclease 1 (Dnase1) and murine deoxyribonuclease 1-like 3 (Dnase1l3). Biochem J 389: 355-364.

170. Liu, Q. Y., J. X. Lei, J. LeBlanc, C. Sodja, D. Ly, C. Charlebois, P. R. Walker, T. Yamada, S. Hirohashi, and M. Sikorska. 2004. Regulation of DNaseY activity by actinin-alpha4 during apoptosis. Cell Death Differ 11: 645-654.

171. Liu, Q. Y., M. Ribecco, S. Pandey, P. R. Walker, and M. Sikorska. 1999. Apoptosis-related functional features of the DNaseI-like family of nucleases. Ann N Y Acad Sci 887: 60-76.

172. Mizuta, R., S. Araki, M. Furukawa, Y. Furukawa, S. Ebara, D. Shiokawa, K. Hayashi, S. Tanuma, and D. Kitamura. 2013. DNase gamma is the effector endonuclease for internucleosomal DNA fragmentation in necrosis. PLoS One 8: e80223.

173. Sisirak, V., B. Sally, V. D'Agati, W. Martinez-Ortiz, Z. B. Ozcakar, J. David, A. Rashidfarrokhi, A. Yeste, C. Panea, A. S. Chida, M. Bogunovic, Ivanov, II, F. J. Quintana, I. Sanz, K. B. Elkon, M. Tekin, F. Yalcinkaya, T. J. Cardozo, R. M. Clancy, J. P. Buyon, and B. Reizis. 2016. Digestion of Chromatin in Apoptotic Cell Microparticles Prevents Autoimmunity. Cell 166: 88-101.

174. Santiago-Raber, M. L., H. Amano, E. Amano, L. Baudino, M. Otani, Q. Lin, F. Nimmerjahn, J. S. Verbeek, J. V. Ravetch, Y. Takasaki, S. Hirose, and S. Izui. 2009. Fcgamma receptor-dependent expansion of a hyperactive monocyte subset in lupus-prone mice. Arthritis Rheum 60: 2408-2417.

- 136 -

175. Mathian, A., A. Weinberg, M. Gallegos, J. Banchereau, and S. Koutouzov. 2005. IFN-alpha induces early lethal lupus in preautoimmune (New Zealand Black x New Zealand White) F1 but not in BALB/c mice. J Immunol 174: 2499-2506.

176. Picard, C., A. L. Mathieu, U. Hasan, T. Henry, Y. Jamilloux, T. Walzer, and A. Belot. 2015. Inherited anomalies of innate immune receptors in pediatric-onset inflammatory diseases. Autoimmun Rev 14: 1147-1153.

177. Dereeper, A., S. Audic, J. M. Claverie, and G. Blanc. 2010. BLAST-EXPLORER helps you building datasets for phylogenetic analysis. BMC Evol Biol 10: 8.

178. Dereeper, A., V. Guignon, G. Blanc, S. Audic, S. Buffet, F. Chevenet, J. F. Dufayard, S. Guindon, V. Lefort, M. Lescot, J. M. Claverie, and O. Gascuel. 2008. Phylogeny.fr: robust phylogenetic analysis for the non-specialist. Nucleic Acids Res 36: W465-469.

179. Ueki, M., H. Takeshita, J. Fujihara, R. Iida, I. Yuasa, H. Kato, A. Panduro, T. Nakajima, Y. Kominato, and T. Yasuda. 2009. Caucasian-specific allele in non-synonymous single nucleotide polymorphisms of the gene encoding deoxyribonuclease I-like 3, potentially relevant to autoimmunity, produces an inactive enzyme. Clin Chim Acta 407: 20-24.

180. Mabbott, N. A., J. K. Baillie, H. Brown, T. C. Freeman, and D. A. Hume. 2013. An expression atlas of human primary cells: inference of gene function from coexpression networks. BMC Genomics 14: 632.

181. Heng, T. S., and M. W. Painter. 2008. The Immunological Genome Project: networks of gene expression in immune cells. Nat Immunol 9: 1091-1094.

182. Ganguly, D., S. Haak, V. Sisirak, and B. Reizis. 2013. The role of dendritic cells in autoimmunity. Nat Rev Immunol 13: 566-577.

183. Lavin, Y., A. Mortha, A. Rahman, and M. Merad. 2015. Regulation of macrophage development and function in peripheral tissues. Nat Rev Immunol 15: 731-744.

184. Date, K., J. Hall, J. Greenman, A. Maraveyas, and L. A. Madden. 2013. Tumour and microparticle tissue factor expression and cancer thrombosis. Thromb Res 131: 109-115.

185. Owens, A. P., 3rd, and N. Mackman. 2011. Microparticles in hemostasis and thrombosis. Circ Res 108: 1284-1297.

186. Gong, J., R. Jaiswal, P. Dalla, F. Luk, and M. Bebawy. 2015. Microparticles in cancer: A review of recent developments and the potential for clinical application. Semin Cell Dev Biol 40: 35-40.

187. Rak, J. 2010. Microparticles in cancer. Semin Thromb Hemost 36: 888-906.

188. Iversen, L. V., O. Ostergaard, S. Ullman, C. T. Nielsen, P. Halberg, T. Karlsmark, N. H. Heegaard, and S. Jacobsen. 2013. Circulating microparticles and plasma levels of soluble E- and P-selectins in patients with systemic sclerosis. Scand J Rheumatol 42: 473-482.

- 137 -

189. Dye, J. R., A. J. Ullal, and D. S. Pisetsky. The role of microparticles in the pathogenesis of rheumatoid arthritis and systemic lupus erythematosus. Scand J Immunol 78: 140-148.

190. Beyer, C., and D. S. Pisetsky. 2010. The role of microparticles in the pathogenesis of rheumatic diseases. Nat Rev Rheumatol 6: 21-29.

191. Ullal, A. J., and D. S. Pisetsky. 2010. The release of microparticles by Jurkat leukemia T cells treated with staurosporine and related kinase inhibitors to induce apoptosis. Apoptosis 15: 586-596.

192. Spencer, D. M., J. Gauley, and D. S. Pisetsky. 2014. The properties of microparticles from RAW 264.7 macrophage cells undergoing in vitro activation or apoptosis. Innate Immun 20: 239-248.

193. Piccin, A., W. G. Murphy, and O. P. Smith. 2007. Circulating microparticles: pathophysiology and clinical implications. Blood Rev 21: 157-171.

194. Ratajczak, J., M. Wysoczynski, F. Hayek, A. Janowska-Wieczorek, and M. Z. Ratajczak. 2006. Membrane-derived microvesicles: important and underappreciated mediators of cell-to-cell communication. Leukemia 20: 1487-1495.

195. Del Conde, I., C. N. Shrimpton, P. Thiagarajan, and J. A. Lopez. 2005. Tissue-factor-bearing microvesicles arise from lipid rafts and fuse with activated platelets to initiate coagulation. Blood 106: 1604-1611.

196. Distler, J. H., A. Jungel, L. C. Huber, C. A. Seemayer, C. F. Reich, 3rd, R. E. Gay, B. A. Michel, A. Fontana, S. Gay, D. S. Pisetsky, and O. Distler. 2005. The induction of matrix metalloproteinase and cytokine expression in synovial fibroblasts stimulated with immune cell microparticles. Proc Natl Acad Sci U S A 102: 2892-2897.

197. Kalinkovich, A., S. Tavor, A. Avigdor, J. Kahn, A. Brill, I. Petit, P. Goichberg, M. Tesio, N. Netzer, E. Naparstek, I. Hardan, A. Nagler, I. Resnick, A. Tsimanis, and T. Lapidot. 2006. Functional CXCR4-expressing microparticles and SDF-1 correlate with circulating acute myelogenous leukemia cells. Cancer Res 66: 11013-11020.

198. Mesri, M., and D. C. Altieri. 1998. Endothelial cell activation by leukocyte microparticles. J Immunol 161: 4382-4387.

199. Hron, G., M. Kollars, H. Weber, V. Sagaster, P. Quehenberger, S. Eichinger, P. A. Kyrle, and A. Weltermann. 2007. Tissue factor-positive microparticles: cellular origin and association with coagulation activation in patients with colorectal cancer. Thromb Haemost 97: 119-123.

200. Janowska-Wieczorek, A., M. Wysoczynski, J. Kijowski, L. Marquez-Curtis, B. Machalinski, J. Ratajczak, and M. Z. Ratajczak. 2005. Microvesicles derived from activated platelets induce metastasis and angiogenesis in lung cancer. Int J Cancer 113: 752-760.

201. Mostefai, H. A., R. Andriantsitohaina, and M. C. Martinez. 2008. Plasma membrane microparticles in angiogenesis: role in ischemic diseases and in cancer. Physiol Res 57: 311-320.

- 138 -

202. Shah, M. D., A. L. Bergeron, J. F. Dong, and J. A. Lopez. 2008. Flow cytometric measurement of microparticles: pitfalls and protocol modifications. Platelets 19: 365-372.

203. Snyder, M. W., M. Kircher, A. J. Hill, R. M. Daza, and J. Shendure. 2016. Cell-free DNA Comprises an In Vivo Nucleosome Footprint that Informs Its Tissues-Of-Origin. Cell 164: 57-68.

204. Sun, K., P. Jiang, K. C. Chan, J. Wong, Y. K. Cheng, R. H. Liang, W. K. Chan, E. S. Ma, S. L. Chan, S. H. Cheng, R. W. Chan, Y. K. Tong, S. S. Ng, R. S. Wong, D. S. Hui, T. N. Leung, T. Y. Leung, P. B. Lai, R. W. Chiu, and Y. M. Lo. 2015. Plasma DNA tissue mapping by genome-wide methylation sequencing for noninvasive prenatal, cancer, and transplantation assessments. Proc Natl Acad Sci U S A 112: E5503-5512.

205. Nielsen, C. T., O. Ostergaard, C. Johnsen, S. Jacobsen, and N. H. Heegaard. 2011. Distinct features of circulating microparticles and their relationship to clinical manifestations in systemic lupus erythematosus. Arthritis Rheum 63: 3067-3077.

206. Nielsen, C. T., O. Ostergaard, O. P. Rekvig, G. Sturfelt, S. Jacobsen, and N. H. Heegaard. 2015. Galectin-3 binding protein links circulating microparticles with electron dense glomerular deposits in lupus nephritis. Lupus 24: 1150-1160.

207. Nielsen, C. T., O. Ostergaard, L. Stener, L. V. Iversen, L. Truedsson, B. Gullstrand, S. Jacobsen, and N. H. Heegaard. 2012. Increased IgG on cell-derived plasma microparticles in systemic lupus erythematosus is associated with autoantibodies and complement activation. Arthritis Rheum 64: 1227-1236.

208. Ostergaard, O., C. T. Nielsen, L. V. Iversen, J. T. Tanassi, S. Knudsen, S. Jacobsen, and N. H. Heegaard. 2013. Unique protein signature of circulating microparticles in systemic lupus erythematosus. Arthritis Rheum 65: 2680-2690.

209. Ullal, A. J., and D. S. Pisetsky. The role of microparticles in the generation of immune complexes in murine lupus. Clin Immunol 146: 1-9.

210. Ullal, A. J., C. F. Reich, 3rd, M. Clowse, L. G. Criscione-Schreiber, M. Tochacek, M. Monestier, and D. S. Pisetsky. 2011. Microparticles as antigenic targets of antibodies to DNA and nucleosomes in systemic lupus erythematosus. J Autoimmun 36: 173-180.

211. Pisetsky, D. S., J. Gauley, and A. J. Ullal. 2010. Microparticles as a source of extracellular DNA. Immunol Res 49: 227-234.

212. Casciola-Rosen, L. A., G. Anhalt, and A. Rosen. 1994. Autoantigens targeted in systemic lupus erythematosus are clustered in two populations of surface structures on apoptotic keratinocytes. J Exp Med 179: 1317-1330.

213. Radic, M., T. Marion, and M. Monestier. 2004. Nucleosomes are exposed at the cell surface in apoptosis. J Immunol 172: 6692-6700.

214. Ullal, A. J., T. N. Marion, and D. S. Pisetsky. 2014. The role of antigen specificity in the binding of murine monoclonal anti-DNA antibodies to microparticles from apoptotic cells. Clin Immunol 154: 178-187.

- 139 -

215. Martinez Valle, F., E. Balada, J. Ordi-Ros, and M. Vilardell-Tarres. 2008. DNase 1 and systemic lupus erythematosus. Autoimmun Rev 7: 359-363.

216. Dieker, J., J. Tel, E. Pieterse, A. Thielen, N. Rother, M. Bakker, J. Fransen, H. B. Dijkman, J. H. Berden, J. M. de Vries, L. B. Hilbrands, and J. van der Vlag. 2016. Circulating Apoptotic Microparticles in Systemic Lupus Erythematosus Patients Drive the Activation of Dendritic Cell Subsets and Prime Neutrophils for NETosis. Arthritis Rheumatol 68: 462-472.

217. Omura, S., Y. Iwai, A. Hirano, A. Nakagawa, J. Awaya, H. Tsuchya, Y. Takahashi, and R. Masuma. 1977. A new alkaloid AM-2282 OF Streptomyces origin. Taxonomy, fermentation, isolation and preliminary characterization. J Antibiot (Tokyo) 30: 275-282.

218. Karaman, M. W., S. Herrgard, D. K. Treiber, P. Gallant, C. E. Atteridge, B. T. Campbell, K. W. Chan, P. Ciceri, M. I. Davis, P. T. Edeen, R. Faraoni, M. Floyd, J. P. Hunt, D. J. Lockhart, Z. V. Milanov, M. J. Morrison, G. Pallares, H. K. Patel, S. Pritchard, L. M. Wodicka, and P. P. Zarrinkar. 2008. A quantitative analysis of kinase inhibitor selectivity. Nat Biotechnol 26: 127-132.

219. Deininger, P. 2011. Alu elements: know the SINEs. Genome Biol 12: 236. 220. Zhao, G. S., L. Chang, and Y. N. Mo. 2010. [Applications of Alu family in

forensic DNA analysis]. Fa Yi Xue Za Zhi 26: 47-50. 221. Reich, C. F., 3rd, and D. S. Pisetsky. 2009. The content of DNA and RNA

in microparticles released by Jurkat and HL-60 cells undergoing in vitro apoptosis. Exp Cell Res 315: 760-768.

222. Castagna, M., Y. Takai, K. Kaibuchi, K. Sano, U. Kikkawa, and Y. Nishizuka. 1982. Direct activation of calcium-activated, phospholipid-dependent protein kinase by tumor-promoting phorbol esters. J Biol Chem 257: 7847-7851.

223. Chatila, T., L. Silverman, R. Miller, and R. Geha. 1989. Mechanisms of T cell activation by the calcium ionophore ionomycin. J Immunol 143: 1283-1289.

224. Roman, A. C., F. J. Gonzalez-Rico, and P. M. Fernandez-Salguero. 2011. B1-SINE retrotransposons: Establishing genomic insulatory networks. Mob Genet Elements 1: 66-70.

225. Andersson, U., H. Erlandsson-Harris, H. Yang, and K. J. Tracey. 2002. HMGB1 as a DNA-binding cytokine. J Leukoc Biol 72: 1084-1091.

226. Pisetsky, D. S. 2014. The expression of HMGB1 on microparticles released during cell activation and cell death in vitro and in vivo. Mol Med 20: 158-163.

227. Pisetsky, D. S., J. Gauley, and A. J. Ullal. HMGB1 and microparticles as mediators of the immune response to cell death. Antioxid Redox Signal 15: 2209-2219.

228. Spencer, D. M., F. Mobarrez, H. Wallen, and D. S. Pisetsky. 2014. The expression of HMGB1 on microparticles from Jurkat and HL-60 cells undergoing apoptosis in vitro. Scand J Immunol 80: 101-110.

- 140 -

229. Abdulahad, D. A., J. Westra, P. C. Limburg, C. G. Kallenberg, and M. Bijl. 2010. HMGB1 in systemic lupus Erythematosus: Its role in cutaneous lesions development. Autoimmun Rev 9: 661-665.

230. Lu, M., S. Yu, W. Xu, B. Gao, and S. Xiong. 2015. HMGB1 Promotes Systemic Lupus Erythematosus by Enhancing Macrophage Inflammatory Response. J Immunol Res 2015: 946748.

231. Wirestam, L., H. Schierbeck, T. Skogh, I. Gunnarsson, L. Ottosson, H. Erlandsson-Harris, J. Wettero, and C. Sjowall. 2015. Antibodies against High Mobility Group Box protein-1 (HMGB1) versus other anti-nuclear antibody fine-specificities and disease activity in systemic lupus erythematosus. Arthritis Res Ther 17: 338.

232. Urbonaviciute, V., B. G. Furnrohr, S. Meister, L. Munoz, P. Heyder, F. De Marchis, M. E. Bianchi, C. Kirschning, H. Wagner, A. A. Manfredi, J. R. Kalden, G. Schett, P. Rovere-Querini, M. Herrmann, and R. E. Voll. 2008. Induction of inflammatory and immune responses by HMGB1-nucleosome complexes: implications for the pathogenesis of SLE. J Exp Med 205: 3007-3018.

233. Athens, J. W., O. P. Haab, S. O. Raab, A. M. Mauer, H. Ashenbrucker, G. E. Cartwright, and M. M. Wintrobe. 1961. Leukokinetic studies. IV. The total blood, circulating and marginal granulocyte pools and the granulocyte turnover rate in normal subjects. J Clin Invest 40: 989-995.

234. Buyon, J. P., M. Y. Kim, M. M. Guerra, C. A. Laskin, M. Petri, M. D. Lockshin, L. Sammaritano, D. W. Branch, T. F. Porter, A. Sawitzke, J. T. Merrill, M. D. Stephenson, E. Cohn, L. Garabet, and J. E. Salmon. 2015. Predictors of Pregnancy Outcomes in Patients With Lupus: A Cohort Study. Ann Intern Med 163: 153-163.

235. Crispin, J. C., S. N. Liossis, K. Kis-Toth, L. A. Lieberman, V. C. Kyttaris, Y. T. Juang, and G. C. Tsokos. 2010. Pathogenesis of human systemic lupus erythematosus: recent advances. Trends Mol Med 16: 47-57.

236. Fairhurst, A. M., A. E. Wandstrat, and E. K. Wakeland. 2006. Systemic lupus erythematosus: multiple immunological phenotypes in a complex genetic disease. Adv Immunol 92: 1-69.

237. Schiller, M., M. Parcina, P. Heyder, S. Foermer, J. Ostrop, A. Leo, K. Heeg, M. Herrmann, H. M. Lorenz, and I. Bekeredjian-Ding. 2012. Induction of type I IFN is a physiological immune reaction to apoptotic cell-derived membrane microparticles. J Immunol 189: 1747-1756.

238. Pisetsky, D. S., A. J. Ullal, J. Gauley, and T. C. Ning. Microparticles as mediators and biomarkers of rheumatic disease. Rheumatology (Oxford) 51: 1737-1746.

239. Takeuchi, Y., M. O. McClure, and M. Pizzato. 2008. Identification of gammaretroviruses constitutively released from cell lines used for human immunodeficiency virus research. J Virol 82: 12585-12588.

240. Jenks, S. A., E. M. Palmer, E. Y. Marin, L. Hartson, A. S. Chida, C. Richardson, and I. Sanz. 2013. 9G4+ autoantibodies are an important source of apoptotic cell reactivity associated with high levels of disease activity in systemic lupus erythematosus. Arthritis Rheum 65: 3165-3175.

- 141 -

241. Richardson, C., A. S. Chida, D. Adlowitz, L. Silver, E. Fox, S. A. Jenks, E. Palmer, Y. Wang, J. Heimburg-Molinaro, Q. Z. Li, C. Mohan, R. Cummings, C. Tipton, and I. Sanz. 2013. Molecular basis of 9G4 B cell autoreactivity in human systemic lupus erythematosus. J Immunol 191: 4926-4939.

242. Lande, R., J. Gregorio, V. Facchinetti, B. Chatterjee, Y. H. Wang, B. Homey, W. Cao, B. Su, F. O. Nestle, T. Zal, I. Mellman, J. M. Schroder, Y. J. Liu, and M. Gilliet. 2007. Plasmacytoid dendritic cells sense self-DNA coupled with antimicrobial peptide. Nature 449: 564-569.

243. Rekvig, O. P., and E. S. Mortensen. 2012. Immunity and autoimmunity to dsDNA and chromatin--the role of immunogenic DNA-binding proteins and nuclease deficiencies. Autoimmunity 45: 588-592.

244. Lui, Y. Y., K. W. Chik, R. W. Chiu, C. Y. Ho, C. W. Lam, and Y. M. Lo. 2002. Predominant hematopoietic origin of cell-free DNA in plasma and serum after sex-mismatched bone marrow transplantation. Clin Chem 48: 421-427.

245. Garlatti, V., A. Chouquet, T. Lunardi, R. Vives, H. Paidassi, H. Lortat-Jacob, N. M. Thielens, G. J. Arlaud, and C. Gaboriaud. 2010. Cutting edge: C1q binds deoxyribose and heparan sulfate through neighboring sites of its recognition domain. J Immunol 185: 808-812.

246. Paidassi, H., P. Tacnet-Delorme, T. Lunardi, G. J. Arlaud, N. M. Thielens, and P. Frachet. 2008. The lectin-like activity of human C1q and its implication in DNA and apoptotic cell recognition. FEBS Lett 582: 3111-3116.

247. Heidari, Y., A. E. Bygrave, R. J. Rigby, K. L. Rose, M. J. Walport, H. T. Cook, T. J. Vyse, and M. Botto. 2006. Identification of chromosome intervals from 129 and C57BL/6 mouse strains linked to the development of systemic lupus erythematosus. Genes Immun 7: 592-599.

248. Gall, A., P. Treuting, K. B. Elkon, Y. M. Loo, M. Gale, Jr., G. N. Barber, and D. B. Stetson. 2012. Autoimmunity initiates in nonhematopoietic cells and progresses via lymphocytes in an interferon-dependent autoimmune disease. Immunity 36: 120-131.

249. Chappell, C. P., K. E. Draves, N. V. Giltiay, and E. A. Clark. 2012. Extrafollicular B cell activation by marginal zone dendritic cells drives T cell–dependent antibody responses. The Journal of Experimental Medicine 209: 1825-1840.

250. Kamath, A. T., J. Pooley, M. A. O’Keeffe, D. Vremec, Y. Zhan, A. M. Lew, A. D’Amico, L. Wu, D. F. Tough, and K. Shortman. 2000. The Development, Maturation, and Turnover Rate of Mouse Spleen Dendritic Cell Populations. The Journal of Immunology 165: 6762-6770.

251. Holdenrieder, S., P. Stieber, L. Y. Chan, S. Geiger, A. Kremer, D. Nagel, and Y. M. Lo. 2005. Cell-free DNA in serum and plasma: comparison of ELISA and quantitative PCR. Clin Chem 51: 1544-1546.

252. Ho, K. T., and J. D. Reveille. 2003. The clinical relevance of autoantibodies in scleroderma. Arthritis Res Ther 5: 80.

- 142 -

253. Kotterman, M. A., and D. V. Schaffer. 2014. Engineering adeno-associated viruses for clinical gene therapy. Nat Rev Genet 15: 445-451.

254. Lisowski, L., S. S. Tay, and I. E. Alexander. 2015. Adeno-associated virus serotypes for gene therapeutics. Curr Opin Pharmacol 24: 59-67.

255. Bentow, C., R. Rosenblum, P. Correia, E. Karayev, D. Karayev, D. Williams, J. Kulczycka, M. J. Fritzler, and M. Mahler. 2016. Development and multi-center evaluation of a novel immunoadsorption method for anti-DFS70 antibodies. Lupus 25: 897-904.

256. Sauer, J. D., K. Sotelo-Troha, J. von Moltke, K. M. Monroe, C. S. Rae, S. W. Brubaker, M. Hyodo, Y. Hayakawa, J. J. Woodward, D. A. Portnoy, and R. E. Vance. 2010. The N-ethyl-N-nitrosourea-induced Goldenticket mouse mutant reveals an essential function of Sting in the in vivo interferon response to Listeria monocytogenes and cyclic dinucleotides. Infect Immun 79: 688-694.

257. Hou, B., B. Reizis, and A. L. DeFranco. 2008. Toll-like receptors activate innate and adaptive immunity by using dendritic cell-intrinsic and -extrinsic mechanisms. Immunity 29: 272-282.

258. Blanco, F., J. Kalsi, and D. A. Isenberg. 1991. Analysis of antibodies to RNA in patients with systemic lupus erythematosus and other autoimmune rheumatic diseases. Clin Exp Immunol 86: 66-70.

259. Guo, W., and H. Wu. 2008. Detection of LacZ expression by FACS-Gal analysis. Nature Protocol Exchange.

260. Buch, T., F. L. Heppner, C. Tertilt, T. J. Heinen, M. Kremer, F. T. Wunderlich, S. Jung, and A. Waisman. 2005. A Cre-inducible diphtheria toxin receptor mediates cell lineage ablation after toxin administration. Nat Methods 2: 419-426.

261. Caton, M. L., M. R. Smith-Raska, and B. Reizis. 2007. Notch-RBP-J signaling controls the homeostasis of CD8- dendritic cells in the spleen. J Exp Med 204: 1653-1664.

262. Yogev, N., F. Frommer, D. Lukas, K. Kautz-Neu, K. Karram, D. Ielo, E. von Stebut, H. C. Probst, M. van den Broek, D. Riethmacher, T. Birnberg, T. Blank, B. Reizis, T. Korn, H. Wiendl, S. Jung, M. Prinz, F. C. Kurschus, and A. Waisman. 2012. Dendritic cells ameliorate autoimmunity in the CNS by controlling the homeostasis of PD-1 receptor(+) regulatory T cells. Immunity 37: 264-275.

263. Sudo, T., S. Nishikawa, M. Ogawa, H. Kataoka, N. Ohno, A. Izawa, S. Hayashi, and S. Nishikawa. 1995. Functional hierarchy of c-kit and c-fms in intramarrow production of CFU-M. Oncogene 11: 2469-2476.

264. Hashimoto, D., A. Chow, M. Greter, Y. Saenger, W. H. Kwan, M. Leboeuf, F. Ginhoux, J. C. Ochando, Y. Kunisaki, N. van Rooijen, C. Liu, T. Teshima, P. S. Heeger, E. R. Stanley, P. S. Frenette, and M. Merad. 2011. Pretransplant CSF-1 therapy expands recipient macrophages and ameliorates GVHD after allogeneic hematopoietic cell transplantation. J Exp Med 208: 1069-1082.

265. Abagyan, R. A., and S. Batalov. 1997. Do aligned sequences share the same fold? J Mol Biol 273: 355-368.

- 143 -

266. Abagyan, R., and M. Totrov. 1994. Biased probability Monte Carlo conformational searches and electrostatic calculations for peptides and proteins. J Mol Biol 235: 983-1002.

267. Cardozo, T., M. Totrov, and R. Abagyan. 1995. Homology modeling by the ICM method. Proteins 23: 403-414.

268. Tipton, C. M., C. F. Fucile, J. Darce, A. Chida, T. Ichikawa, I. Gregoretti, S. Schieferl, J. Hom, S. Jenks, R. J. Feldman, R. Mehr, C. Wei, F. E. Lee, W. C. Cheung, A. F. Rosenberg, and I. Sanz. 2015. Diversity, cellular origin and autoreactivity of antibody-secreting cell population expansions in acute systemic lupus erythematosus. Nat Immunol 16: 755-765.

269. Losman, M. J., T. M. Fasy, K. E. Novick, and M. Monestier. 1993. Relationships among antinuclear antibodies from autoimmune MRL mice reacting with histone H2A-H2B dimers and DNA. Int Immunol 5: 513-523.

270. Monestier, M., and K. E. Novick. 1996. Specificities and genetic characteristics of nucleosome-reactive antibodies from autoimmune mice. Mol Immunol 33: 89-99.

271. Shlomchik, M. J., A. H. Aucoin, D. S. Pisetsky, and M. G. Weigert. 1987. Structure and function of anti-DNA autoantibodies derived from a single autoimmune mouse. Proc Natl Acad Sci U S A 84: 9150-9154.

272. Weissgerber, T. L., N. M. Milic, S. J. Winham, and V. D. Garovic. 2015. Beyond bar and line graphs: time for a new data presentation paradigm. PLoS Biol 13: e1002128.