mmun

Cas

A

a

b

mcTtaaseacIvcitetblmmaaitomcDhpoe

d

AM

PJC

a

b

c

hlMcp(aww

Abstracts / Molecular I

hicken class II DMA and DMB are similar to mammals, butsecond DMB gene differs in expression level, regulation and

tructural features

imée Parker a,∗, Colin Butter b, Karen Staines b, Jim Kaufman a

Department of Pathology, University of Cambridge, UKInstitute for Animal Health, Compton, UK

Though it is smaller, simpler and rearranged, the chickenajor histocompatibility complex (MHC) contains most of the

ore antigen presentation genes found in the mammalian MHC.he well-documented strong associations of chicken MHC haplo-ypes with resistance and susceptibility to particular pathogensnd inactivated vaccines have been attributed to the presence ofsingle dominantly expressed class I and class II molecule. The

ingle dominant class I has been shown to result from of co-volution with the closely linked antigen processing genes TAPnd tapasin, but the situation for class II is less clear. In mammals,lass II peptide presentation is assisted by the non-classical classI molecule DM, but little is known about DM in non-mammalianertebrates. We therefore began by characterising the chickenlass II DM region and show that chickens have three DM genesn their MHC, a single alpha chain gene chDMA, and, unusually,wo beta chain genes chDMB1 and chDMB2. The chDM genesncode proteins with high structural and sequence homology tohe DM molecules described in other species. However, the twoeta chains differ substantially from each other at the amino acid

evel and the lack of an endosomal sorting signal for chDMB1ay suggest functional differences. We identify proximal pro-oter modules in keeping with mammalian class II promoters,

nd describe splice variants of the alpha chain gene involving anlternative upstream first exon. All three genes are expressed inmmune related tissues, though levels of chDMB1 are much lowerhan chDMA and chDMB2. In pursuing the functional propertiesf the chDMs and their interaction with the classical class II BLBolecules, we aim to establish whether the dominantly expressed

lass II could result from co-evolution with a dominantly expressedM molecule, and whether the more poorly expressed moleculesave a different or context-dependent function, questions directlyertinent to any species where multiple DMB genes are found andf wider importance for deepening our understanding of class IIvolution.

oi:10.1016/j.molimm.2012.02.072

multi-dimensional RNAi screen reveals pathways controllingHC Class II antigen presentation

etra Paul a,∗, Tineke van den Hoorn a, Marlieke Jongsma a, Lennertanssen a, Peter Cresswell b, Marieke van Ham c, Anja ten Brinke c,oenraad Kuijl a, Jacques Neefjes a

The Netherlands Cancer Institute AmsterdamHoward Hughes Medical Institute New Haven CTSanquin Research and Landsteiner Laboratory Amsterdam

MHC class II molecules (MHC-II) present peptides to CD4 + Telper cells to facilitate immune responses and are strongly

inked to autoimmune diseases. To unravel processes controllingHC-II antigen presentation, we performed a genome-wide flow

ytometry-based RNAi screen detecting MHC-II expression andeptide loading followed by additional high-throughput assaysPaul et al., 2011). We identified 276 genes controlling expression

nd peptide loading of MHC class II molecules. Among them, 69ere selectively expressed in immune tissues and 21 correlatedith autoimmune diseases.

ology 51 (2012) 5–41 27

All data sets were integrated to answer two fundamental ques-tions: what regulates tissue specific MHC-II transcription, and whatcontrols MHC-II transport in dendritic cells?

MHC-II transcription was controlled by nine regulators actingin feedback networks with higher-order control by signaling path-ways, including TGF�. MHC-II transport was controlled by theGTPase ARL14/ARF7, which recruits the motor myosin 1E via aneffector protein ARF7EP. This complex controls movement of MHC-II vesicles along the actin cytoskeleton in human dendritic cells.

Furthermore, we have identified molecules playing a crucial rolein the cross-talk between antigen presenting cell (APC) and T cell,such as PD-L1. This protein and others were shown to control themovement of MHC-II vesicles in human dendritic cells. Currently,ways to manipulate the activation of T cells are being explored, bothfrom the APC- and the T cell side.

These genome-wide systems analyses have identified factorsand pathways controlling MHC-II transcription and transport,defining targets for manipulation of MHC-II antigen presentation,and hence immune activation, in infection and autoimmunity.

Reference

Paul, et al., 2011. A Genome-wide multi-dimensional RNAi screen reveals pathwayscontrolling MHC Class II antigen presentation. Cell 145 (April), 268–283.

doi:10.1016/j.molimm.2012.02.073

Microbial detection controls defective ribosomal proteinsdegradation by autophagy and subsequent endogenous MHCII-restricted presentation in dendritic cells

Philippe Pierre

Centre d’Immunologie de Marseille-Luminy CNRS-INSERM-AMU.Campus de Luminy, Case 906,13288 Marseille, Cedex 09, France

Dendritic cells (DCs) are antigen presenting cells with theunique capacity to initiate primary immune responses. DCs havea remarkable pattern of differentiation (maturation) that exhibitshighly specific mechanisms to control antigen presentation inresponse to microbial stimuli. In particular, expression, transportand loading of MHC class I and II molecules are strongly acti-vated during DC maturation. MHC class I molecules present toCD8 + cytotoxic T cells peptides mostly derived from defectivecytosolic proteins (DRiPS), which are ubiquitinated and degradedby the proteasome. Here, we show that newly synthesized defec-tive proteins are also targeted to autophagosomes and degraded.Upon DC maturation, autophagy is inhibited leading to the accu-mulation of these DRiPs in large cytosolic structures. Interestingly,these Dendritic cell Aggresome Like Induced Structures (DALIS)are transient, and require both translation inhibition and protea-some activity for their disapearance. Thus, autophagy affects theefficiency of the classical proteasome-mediated antigen process-ing by competing for a fraction of the same DRiPs substrates.Autophagy was estimated to reduce the level MHC I presentationby 20%. We also demonstrate, that this pathway is used by DCsto present endogenous proteins on the MHC II molecules and thatimmature cells have the capacity of presenting peptides derivedfrom the same endogenous antigens both on the MHC I and MHCII. Upon microbial activation MHC II presentation of the endoge-nous antigens is lost, while the MHC I antigens source changesfrom newly synthesized proteins to other undefined sources. Mor-ever, we demonstrate that several cytokines can revert this process

and induce alternative pathways for the autophagic processing ofantigens as well as different immunological functions. This phe-nomenon is likely to be important for regulating MHC class I andII presentation and the establishment of tolerogenic or immunos-

2 mmun

tm

d

Wa

B

mvpmoCbeo(opimgpihi

eerggmirustpmaplgmbfao

apcFdsamIW

cross-presentation by human and mouse DC that need to be con-sidered when designing human vaccines.

doi:10.1016/j.molimm.2012.02.076

8 Abstracts / Molecular I

imulatory conditions during the necessary functional switch of DCaturation and CD4 + T cells licensing.

oi:10.1016/j.molimm.2012.02.074

hat can listeria teach us about MHC class I antigen processingnd presentation?

enjamin Wolf, Michael Princiotta ∗

SUNY Upstate Medical University, Syracuse, NY

Most, though not all, peptides presented by MHC class Iolecules (pMHC) are generated by the degradation of proteins

ia the ubiquitin proteasome system. A general feature of thisathway is that, for a given protein, increasing the number ofolecules degraded will result in a proportionally greater supply

f peptides available for presentation on MHC class I molecules.onsequently, the supply of peptides available for presentationy MHC class I molecules can be increased by radically short-ning the cellular half-life of a protein through the introductionf amino acid sequences that target the protein for degradatione.g. N-degrons, PEST sequences, degradation domains). We previ-usly demonstrated that decreasing the cellular t1/2 of a substraterotein expressed from recombinant vaccinia virus (rVV) resulted

n a concomitant increase in pMHC generation. We also deter-ined that the average efficiency of surface pMHC expression for a

iven peptide generated from a recombinant protein was ∼1 pMHCer 3000 protein molecules degraded. One of the more intrigu-

ng observations made during the course of these studies was thatighly similar proteins can be processed with different efficiencies

n the same cell.Why highly similar proteins would be processed with different

fficiencies is not clear. To date, the only other study to quantify thefficiency of pMHC generation is that of Pamer and colleagues, whoeported an efficiency of pMHC generation 2–3 orders of magnitudereater for native bacterial proteins secreted by the intracellularram positive bacterium, Listeria monocytogenes, relative to thoseeasured using our rVV system. It is possible that the differences

n efficiency observed between the Listeria and vaccinia studieseflect differences inherent to the specific proteins and/or peptidessed for these studies. To address this issue, we designed a novelystem for expressing and secreting recombinant proteins from Lis-eria, which would allow us to determine the efficiency of pMHCroduction from a stable and rapidly degraded fusion protein in aanner analogous to our previously findings using rVV. Remark-

bly, we found that the rate of pMHC generation for recombinantrotein expressed by Listeria did not change, regardless of the cel-

ular half-life of the protein. Furthermore, the efficiency of pMHCeneration from recombinant Listeria proteins was at least 19-foldore efficient than for the same recombinant proteins expressed

y vaccinia. Together, our results suggest that proteins secretedrom Listeria that are processed for MHC class I presentation follow,lmost exclusively, a high efficiency processing pathway that cannly be accessed by a subset of endogenously synthesized proteins.

Our current studies are focused on deducing the cellular mech-nism(s) responsible for the high-efficiency processing of Listeriaroteins. To date, we have determined that differences in effi-iency occur prior to the liberation of peptide from the proteasome.urthermore, this high efficiency pathway is proteasome and TAPependent and does not require expression of immunoproteasomeubunits. Listerial infection of antigen presenting cells does notlter the processing efficiency of non-listerial proteins. Further-

ore, processing of Listeria proteins is independent of MyD88 and

L1R signaling and does not require formation of autophagosomes.e are currently examining the role of several molecules asso-

ology 51 (2012) 5–41

ciated with cross-presentation in professional antigen processingcells.

doi:10.1016/j.molimm.2012.02.075

Cross-presentation by human dendritic subsets depends on thetype of antigen and exogenous stimuli

Kylie McDonald a, Andrew Kassianos a, Varinder Jeet a, XinshengJu a, Yitian Ding a, Derek Hart b, Kristen Radford a,∗

a Mater Medical Research Institute, Brisbane, Australiab Anzac Research Institute, Sydney, Australia

The dendritic cell (DC) network in mice and humans consistsof multiple subsets with specialised functions. In mice, the CD8+conventional DC subset is specialised in cross-presentation andknown to play a crucial role in the induction of anti-viral and anti-tumor CTL responses. We and others have recently identified thehuman CD141+ DC subset as the equivalents of the mouse CD8+DC subset. Like mouse CD8+ DC, CD141+ DC are effective at cross-presentation. However, human CD1c+ DC and monocyte-derivedDC (MoDC) can also cross-present in some circumstances and theefficiency of cross-presentation by these three subsets has notbeen directly compared. We examined cross-presentation of theHLA-A*0201-restricted human cytomegalovirus (CMV) pp65495-503 epitope following uptake of recombinant pp65 protein byhuman DC subsets. In the absence of activation, CD1c+ DC andMoDC cross-presented the pp65495-503 epitope, but this was notfurther enhanced by the addition of exogenous activators such asLPS or poly I:C. Cross-presentation by CD1c+ DC and MoDC wasenhanced in the presence of bafilomycin, a specific inhibitor ofvacuolar-type H(+)-ATPase that inhibits endosomal acidification.The proteasome inhibitor, lactacystin, inhibited cross-presentationby CD1c+ DC but not MoDC. In contrast, cross-presentation ofrecombinant pp65 protein by CD141+ DC was weak in the absenceof activation but was significantly enhanced in the presence of polyI:C. Cross-presentation by CD141+ DC was proteasome dependentbut not further enhanced by inhibiting endosomal acidification.We next examined the capacity of human DC subsets to cross-present the pp65495-503 epitope following uptake of necroticcytomegalovirus-infected fibroblasts. Although all DC subsets wereefficient in uptake of necrotic cells, CD141+ DC were superiorto CD1c+ DC and MoDC in their capacity to cross-present thepp65495-503 epitope. Cross-presentation of the virus-infectedfibroblasts appeared to use similar mechanisms to recombinantprotein. These data show that, like mouse CD8+ DC, human CD141+DC are specialised in cross-presenting Ag from dead cells. However,CD1c+ DC, CD141+ DC and MoDC are all capable of cross-presentingrecombinant protein but process the pp65495-503 epitope viadifferent mechanisms. These data suggest subtle differences in