7
Vaccine 30 (2012) 4859–4865 Contents lists available at SciVerse ScienceDirect Vaccine jou rn al h om epa ge: www.elsevier.com/locate/vaccine Use of human MonoMac6 cells for development of in vitro assay predictive of adjuvant safety in vivo Marina Zaitseva a,, Tatiana Romantseva a , Ksenia Blinova a , Joel Beren a , Lev Sirota a , Debbie Drane b , Hana Golding a a Divisions of Viral Products, Veterinary Services, and Biostatistics, Center for Biologics Evaluation and Research (CBER), Food and Drug Administration (FDA), Bethesda, MD 20892, USA b CSL Behring King of Prussia, PA 19406, USA a r t i c l e i n f o Article history: Received 8 December 2011 Received in revised form 9 April 2012 Accepted 1 May 2012 Available online 15 May 2012 Keywords: Adjuvants Human monocytes Cytokines TLR agonists Prostaglandin E2 a b s t r a c t Subunit vaccines composed of recombinant or purified antigens have a good safety record but are poorly immunogenic and require adjuvants to activate innate immunity and facilitate antigen specific immune response. Of the many adjuvant formulations that are under development, very few are licensed mainly due to concerns about adverse side effects. The goal of our study was to develop in vitro assays that could predict toxicity of adjuvants in vivo. Pro-inflammatory cytokines IL-, IL-6, TNF-, and IL-8 were mea- sured in human primary monocytes and the monocytoid cell line, MonoMac 6 (MM6), activated with a panel of TLR agonists or with adjuvants. A 0.5 EU/ml dose of Standard for endotoxin (previously shown to provide a margin between pyrogenic and non-pyrogenic substances in rabbits) was used as a comparator to establish a “safety threshold”. FSL-1, Pam3CSK4, flagellin, and R848 TLR agonists but not Alum, MF59, Poly I:C, or MPL adjuvants induced cytokines in MM6 cells above the safety threshold. To confirm the predictive value of the in vitro assays, FSL-1 and flagellin were injected intramuscularly into New Zealand White (NZW) rabbits. Both TLR agonists induced fever within 6–8 h post-injection followed 24–48 h later by increased C reactive protein (CRP). Importantly, an early peak in plasma prostaglandin E2 (PGE 2 ) lev- els preceded rise in body temperature. In vitro production of PGE 2 in monocytes and MM6 cells was found following treatments with various TLR agonists but not with alum, MF59, MPL, or Poly I:C adju- vants. Together, our studies demonstrated a strong correlation between production of pro-inflammatory cytokines above a “safety threshold” and production of PGE 2 in vitro and an increase in body temperature in rabbits. The developed human cell based assays could provide an important tool for early screening of new molecular moieties and adjuvant formulations and may assist in selection of safer products. Published by Elsevier Ltd. 1. Introduction The majority of vaccines under development contain highly purified recombinant molecules or subunits of pathogens, hence they lack inherent immunostimulatory properties of the original pathogens. The antigen specific immunity of these vaccines can be amplified with adjuvants to harness elements of the innate immune Disclaimer: The findings and conclusions in this article have not been formally disseminated by the Food and Drug Administration and should not be construed to represent any Agency determination or policy. Corresponding author at: Division of Viral Products, Center for Biologics Evalua- tion and Research, Food and Drug Administration, Building 29B, Room 4NN06, 8800 Rockville Pike, Bethesda, MD, 20892, USA. Tel.: +1 301 827 0788; fax: +1 301 496 1810. E-mail address: [email protected] (M. Zaitseva). system including pattern recognition receptors (PRR), evolutionarily developed to alert the host about the presence of infection [1–5]. Significant efforts are dedicated to identify new adjuvants including National Institute of Health-supported Global Adjuvant Develop- ment Initiative (GADI) for the discovery of new candidate adjuvants through screening of libraries of small molecules mimicking toll like receptor (TLR) agonists [6]. The use of immunostimulatory molecules as adjuvants poses safety concerns related to the potential for over production of inflammatory and pyrogenic molecules [7]. Preclinical studies of adjuvanted vaccines may not always identify an increased risk for vaccine associated adverse events (AEs) due to species-specific differences in receptors and tissue distribution of innate system components [8–11]. Febrile responses commonly observed in vaccines are activated by pyrogenic cytokines, IL-1, IL-6, and TNF- [12,13]. However, during infection with Gram-negative bacteria, Kupffer cells in the liver were shown to produce PGE 2 that was suggested to transmit a 0264-410X/$ see front matter. Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.vaccine.2012.05.002

Use of human MonoMac6 cells for development of in vitro assay predictive of adjuvant safety in vivo

Embed Size (px)

Citation preview

Page 1: Use of human MonoMac6 cells for development of in vitro assay predictive of adjuvant safety in vivo

Ua

MHa

Ub

a

ARRAA

KAHCTP

1

ptpa

dr

tRf

0h

Vaccine 30 (2012) 4859– 4865

Contents lists available at SciVerse ScienceDirect

Vaccine

jou rn al h om epa ge: www.elsev ier .com/ locate /vacc ine

se of human MonoMac6 cells for development of in vitro assay predictive ofdjuvant safety in vivo�

arina Zaitsevaa,∗, Tatiana Romantsevaa, Ksenia Blinovaa, Joel Berena, Lev Sirotaa, Debbie Draneb,ana Goldinga

Divisions of Viral Products, Veterinary Services, and Biostatistics, Center for Biologics Evaluation and Research (CBER), Food and Drug Administration (FDA), Bethesda, MD 20892,SACSL Behring King of Prussia, PA 19406, USA

r t i c l e i n f o

rticle history:eceived 8 December 2011eceived in revised form 9 April 2012ccepted 1 May 2012vailable online 15 May 2012

eywords:djuvantsuman monocytesytokinesLR agonistsrostaglandin E2

a b s t r a c t

Subunit vaccines composed of recombinant or purified antigens have a good safety record but are poorlyimmunogenic and require adjuvants to activate innate immunity and facilitate antigen specific immuneresponse. Of the many adjuvant formulations that are under development, very few are licensed mainlydue to concerns about adverse side effects. The goal of our study was to develop in vitro assays that couldpredict toxicity of adjuvants in vivo. Pro-inflammatory cytokines IL-�, IL-6, TNF-�, and IL-8 were mea-sured in human primary monocytes and the monocytoid cell line, MonoMac 6 (MM6), activated with apanel of TLR agonists or with adjuvants. A 0.5 EU/ml dose of Standard for endotoxin (previously shown toprovide a margin between pyrogenic and non-pyrogenic substances in rabbits) was used as a comparatorto establish a “safety threshold”. FSL-1, Pam3CSK4, flagellin, and R848 TLR agonists but not Alum, MF59,Poly I:C, or MPL adjuvants induced cytokines in MM6 cells above the safety threshold. To confirm thepredictive value of the in vitro assays, FSL-1 and flagellin were injected intramuscularly into New ZealandWhite (NZW) rabbits. Both TLR agonists induced fever within 6–8 h post-injection followed 24–48 h laterby increased C reactive protein (CRP). Importantly, an early peak in plasma prostaglandin E2 (PGE2) lev-

els preceded rise in body temperature. In vitro production of PGE2 in monocytes and MM6 cells wasfound following treatments with various TLR agonists but not with alum, MF59, MPL, or Poly I:C adju-vants. Together, our studies demonstrated a strong correlation between production of pro-inflammatorycytokines above a “safety threshold” and production of PGE2 in vitro and an increase in body temperaturein rabbits. The developed human cell based assays could provide an important tool for early screening ofnew molecular moieties and adjuvant formulations and may assist in selection of safer products.

. Introduction

The majority of vaccines under development contain highly

urified recombinant molecules or subunits of pathogens, hencehey lack inherent immunostimulatory properties of the originalathogens. The antigen specific immunity of these vaccines can bemplified with adjuvants to harness elements of the innate immune

� Disclaimer: The findings and conclusions in this article have not been formallyisseminated by the Food and Drug Administration and should not be construed toepresent any Agency determination or policy.∗ Corresponding author at: Division of Viral Products, Center for Biologics Evalua-

ion and Research, Food and Drug Administration, Building 29B, Room 4NN06, 8800ockville Pike, Bethesda, MD, 20892, USA. Tel.: +1 301 827 0788;

ax: +1 301 496 1810.E-mail address: [email protected] (M. Zaitseva).

264-410X/$ – see front matter. Published by Elsevier Ltd.ttp://dx.doi.org/10.1016/j.vaccine.2012.05.002

Published by Elsevier Ltd.

system including pattern recognition receptors (PRR), evolutionarilydeveloped to alert the host about the presence of infection [1–5].Significant efforts are dedicated to identify new adjuvants includingNational Institute of Health-supported Global Adjuvant Develop-ment Initiative (GADI) for the discovery of new candidate adjuvantsthrough screening of libraries of small molecules mimicking toll likereceptor (TLR) agonists [6].

The use of immunostimulatory molecules as adjuvants posessafety concerns related to the potential for over production ofinflammatory and pyrogenic molecules [7]. Preclinical studies ofadjuvanted vaccines may not always identify an increased riskfor vaccine associated adverse events (AEs) due to species-specificdifferences in receptors and tissue distribution of innate systemcomponents [8–11].

Febrile responses commonly observed in vaccines are activatedby pyrogenic cytokines, IL-1�, IL-6, and TNF-� [12,13]. However,during infection with Gram-negative bacteria, Kupffer cells in theliver were shown to produce PGE2 that was suggested to transmit a

Page 2: Use of human MonoMac6 cells for development of in vitro assay predictive of adjuvant safety in vivo

4 ccine 3

flcimp

ucacT

2

2

SW((laSa(Labr

2

D(t1pTwnC

tiAr

cfaaahEVcpautf

860 M. Zaitseva et al. / Va

ebrile message to the ventromedial preoptic area in the hypotha-amus via a peripheral neural route [14]. Thus, proinflammatoryytokines and PGE2 may both trigger febrile responses follow-ng innate immune activation in vivo, and measurements of these

olecules in vitro or in vivo could help identify adjuvants with theotential to elicit unacceptable reactions in humans.

In the current study, human monocytes and MM6 cell line weresed to develop an in vitro assay differentiating between adjuvantandidates based on relative levels of pro-inflammatory cytokinesnd PGE2. Additional proof-of-concept studies in rabbits providedorrelation between the in vitro findings with the in vivo toxicity ofLR agonists.

. Materials and methods

.1. Reagents

The following reagents were used: the USP Reference Endotoxintandard (from Escherichia coli strain O55:B5, TLR4 ligand, Lonza,alkersville, MD); synthetic bacterial lipoprotein, Pam3CSK4

TLR1/2 ligand), a purified Flagellin from Salmonella typhimuriumTLR5 ligand), and a synthetic diacylated lipoprotein FSL-1 (TLR2/6igand), all from InvivoGen (San Diego, CA); Poly I:C (TLR3 lig-nd) and R848 (TLR7/8 ligand in human) were from Enzo Lifeciences Inc., Farmingdale, NY); Aluminum Hydroxide Gel (Alum)nd Monophosphoryl Lipid A isolated from Salmonella MinnesotaMPL), both from Sigma–Aldrich (St. Louis, MO) or from Axxora,LC (San Diego, CA). ISCOMATRIX adjuvant and ISCOPREP Saponin,nd MF59 were generously provided for this study by CSL andy Novartis Vaccines, respectively. ISCOMATRIX and ISCOPREP areegistered trademarks of ISCOTEC Ab, a CSL Limited company.

.2. Protocol for in vitro treatments

Human elutriated monocytes (MO) were obtained from theepartment of Transfusion at the National Institute of Health

Bethesda, MD). MonoMac6 (MM6) cells (American Type Cul-ure Collection, Manassas, VA) [15] were maintained in RPMI640 medium, supplemented with 2 mM l-glutamine, 200 U/mlenicillin/200 �g/ml streptomycin and 10 �g/ml ITS-X (Insulin-ransferrin-Selenium), all from Mediatech Inc. (Manassas, VA);ith 1 mM OAA (Oxaloactic acid, Sigma), 15% FBS (Sigma), and 1×onessential amino acids and 1 mM Sodium Pyruvate (Invitrogenorporation, Carlsbad, CA).

Cytokines and PGE2 were measured in MO and MM6 cells cul-ured in triplicate wells in 96-well plates at 1–1.5 × 106 cells/mln RPMI medium supplemented with 2% human heat-inactivatedB serum (Mediatech) or with 15% FBS for cytokines and PGE2,espectively.

Endotoxin standard, TLR ligands, and adjuvants were added toells at indicated concentrations in triplicates and cultured for 5 hor endotoxin, and for 18 h for TLR agonists and adjuvants at 37 ◦Cnd 5% CO2. At the end of culture period, plates were centrifugedt 1400 rpm for 15 min, cell culture supernatants were collectednd were assayed for IL-1�, TNF-�, IL-6, and IL-8 cytokines usinguman Quantikine ELISA kits (R&D Systems, Minneapolis, MN) andLx808 Ultra Microplate reader (BioTek Instruments, Winooski,T) and for PGE2 using Homogeneous Time-Resolved Fluores-ence assay (HTRF®) (Cisbio Bioassays, Bedford, MA) and Novostarlate reader (BMG Labtech, Offenburg, Germany) equipped with

time-resolved module. PGE2 concentrations were calculatedsing four-parameter logistic fit using Origin software applica-ion (OriginLab, Northampton, MA). The detectable PGE2 range wasrom 10 to 5000 pg/ml.

0 (2012) 4859– 4865

2.3. Rabbits and protocol for in vivo treatments

Female NZW rabbits weighting 4 kg were purchased fromCharles River Laboratories. One week before experiment, an IVHickman Catheter was surgically implanted, under general anes-thesia, into the right jugular vein in the ventro-lateral portion of theright side of the neck. Animals were injected intramuscularly with1 ml solution of flagellin or FSL-1 in PBS, or with 1 ml of PBS. Rectalbody temperature was measured at time 0, then hourly for 8 h, andat 24, 48, and 72 h time points using SureTemp® Plus thermome-ter (Welch Allyn, Skaneateles Falls, NY). Immediately prior to eachtemperature measurement, 4 ml of whole blood was collected viathe Hickman Catheter. All experimental procedures were approvedby the CBER Institutional Animal Care and Use Committee.

2.4. CRP and PGE2 measurements in rabbits

PGE2 and CRP were detected in rabbit plasma and sera, respec-tively, using PGE2 EIA (Assay Designs, Ann Arbor, MI) and C ReactiveProtein ELISA kits (Alpco Immunoassays, Salem, NH) in triplicates.

2.5. Statistical analysis

All in vitro experiments were performed using triplicate cell cul-tures and testing of rabbit sera and plasma was performed usingtriplicate assay wells. Mean ± SD values from triplicate wells werecalculated using Excel software. JMP 9 Software (SAS Institute Inc.,Cary, NC) was used for the statistical analysis of the dose-responseslopes in double logarithmic coordinates.

3. Results

3.1. Production of IL-1ˇ, IL-8, IL-6, and TNF ̨ proinflammatorycytokines in human monocytes following in vitro treatments withTLR ligands or with adjuvants

Excessive production of proinflammatory cytokines followingactivation of monocytes in vitro could be indicative of increasedreactogenicity in vivo [16]. To explore this possibility, monocyteswere incubated with TLR agonists LPS, FSL-1, Pam3CSK4, Flagellin,and R848 (1 pg to 10 �g/ml) (Fig. 1). At the end of cell culture,supernatants were assayed for IL-1�, IL-8, IL-6, and TNF� by ELISA.Secreted cytokines were detected in monocytes treated with LPS,Pam3CSK4, or FSL-1 with a maximum of 10,000 pg/ml of IL-1� andTNF�, 100 ng/ml of IL-8, and 100,000 pg/ml of IL-6 (Fig. 1A–D).All four cytokines reached plateau levels at 1 ng/ml of LPS or10–100 ng/ml of Pam3CSK4 and FSL-1 (Fig. 1A–D). Other TLR ago-nists exhibited different dose responses. Cytokines were induced byR848 and by flagellin at doses starting from 1 or 30 ng/ml, respec-tively (Fig. 1E–H). These data showed that the dynamic range ofproinflammatory cytokines production in primary human mono-cytes differed for different TLR agonists.

It was important to determine if a monocyte-based assaycould distinguish between strong TLR agonists and adjuvants withproven safety in humans: Alum (licensed world wide), MF59(licensed in Europe), MPL (TLR4 ligand; a component of the licensedAS04-adjuvanted human papillomavirus vaccine, Cervarix, Glaxo-SmithKline), and Poly I:C (TLR3 ligand), a prototype of several PolyI:C-based adjuvants for vaccines against cancer, HIV, and influenza[17]. MF59 and MPL induced pro-inflammatory cytokines very closeto background levels (Fig. 1I–L). Alum induced IL-1� and TNF� (butnot IL-8 or IL-6) only at doses ≥ 1 �g/ml (Fig. 1 panels I and K), and

Poly I:C induced up to 400 pg/ml of IL-6 when added at ≥ 1 �g/ml(Fig. 1 panel L). These data suggested that adjuvants that have beenshown to be safe in clinical settings induce no or very low levelsof proinflammatory cytokines in human monocytes, while strong
Page 3: Use of human MonoMac6 cells for development of in vitro assay predictive of adjuvant safety in vivo

M. Zaitseva et al. / Vaccine 30 (2012) 4859– 4865 4861

Fig. 1. TLR ligands, but not adjuvants induced IL-1�, IL-8, TNF�, and IL-6 in primary human monocytes. Elutriated human monocytes were incubated with TLR agonists: LPS,Pam3CSK4, FSL-1 (A–D), flagellin and R848 (E–H), or with MF59, ALOH, Poly I:C, and MPL adjuvants (I–L). After 5 h of incubation for LPS and 18 h for all other TLR agonistsa ted ILM rimen

Tr

lpmrlodM

Foal

nd adjuvants, the cell culture supernatants were collected and the levels of secreean ± SD of triplicate cell cultures are shown. Results are representative of 5 expe

LR agonists induce proinflammatory cytokines across a broad doseange.

Use of MM6 cells and Standard for endotoxin to estab-ish a threshold between “safe” and “non-safe” levels ofro-inflammatory cytokines. Experiments using primary humanonocytes confirmed their potential to produce cytokines in

esponse to stimulation with TLR agonists. However, data from our

aboratory and from others have shown that the expression levelf TLRs in primary monocytes vary significantly among normalonors [18,19]. Therefore, we selected the human promonocyticM6 cell line, previously suggested as an alternative model to

ig. 2. Establishing the range of cytokine production in response to 0.5 EU of endotoxin sf endotoxin and 5 h later the supernatants were tested for the presence of IL-1� (A), IL-8

re shown. The average levels of cytokine production in MM6 cells in response to 0.5 EU aines above and below the solid lines.

-1� (A, E, I) IL-8 (B, F, J) TNF� (C, G, K) and IL-6 (D, H, L) were measured by ELISA.ts.

rabbit pyrogenic assay for Gram-negative organisms [20]. Standardfor endotoxin (LPS derived from Gram-negative bacteria) was eval-uated as a “comparator” since the 0.5 EU/ml dose was shown toinduce a 0.5 ◦C increase in body temperature in rabbits followingintramuscular injection and was defined as the threshold betweenpyrogenic and non-pyrogenic samples [21]. In repeat experiments,the levels of cytokines secreted by MM6 cells in response to LPS

stimulation were highly reproducible (Fig. 2 and Table 1). Basedon the concentration of cytokines produced by MM6 cells in thepresence of 0.5 EU of Standard for endotoxin (80–100 pg/ml ofLPS), the mean pyrogenic thresholds were calculated as follows:

tandard in MM6 cell line. MM6 cells were activated with indicated concentrations(B), TNF� (C), and IL-6 (D) cytokines by ELISA. The results of individual experimentsre shown as solid lines and the 95% upper and lower CI limits are shown as dashed

Page 4: Use of human MonoMac6 cells for development of in vitro assay predictive of adjuvant safety in vivo

4862 M. Zaitseva et al. / Vaccine 30 (2012) 4859– 4865

Table 1Statistical analysis of the dose-response of the cytokine production in MM6 cells incubated with endotoxin.

Cytokinea Meanb StDev StError 95% CI

IL-1� 0.779 0.239 0.090 (0.558, 0.999)IL-6 0.578 0.197 0.074 (0.396, 0.760)IL-8 0.648 0.179 0.063 (0.499, 0.798)TNF-� 0.904 0.224 0.091 (0.669, 1.139)

tanda 95% c

u

I7poh

3T

ccAdeiw(eiT

3u

dfiMp8i

FoFM

a Each cytokine concentration produced by MM6 cells in presence of endotoxin sb For each cytokine mean, standard deviation (StDev), standard error (StError), and

sing JMP 9 software.

L-1�, 27.35 pg/ml; IL-8, 24.8 ng/ml; IL-6, 3186.0 pg/ml; TNF-�,53 pg/ml; (Fig. 2 panels A, B, C, and D, respectively). Statisticsresented in Table 1 indicate that for each cytokine the slopesf dose-response curves in double logarithmic coordinates wereighly reproducible.

.2. In vitro cytokine dose-response in MM6 cells cultured withLR ligands

To determine whether TLR ligands induce proinflammatoryytokines above the endotoxin-based pyrogenic threshold, MM6ells were cultured with flagellin, FSL-1, or with Pam3CSK4 (Fig. 3).ll three TLR agonists induced IL-1�, IL-6, IL-8, and TNF� in aose-dependent manner (Fig. 3A–D). Cytokine production at lev-ls above the mean pyrogenic thresholds (horizontal broken linesn each panel) were detected in MM6 cell supernatants cultured

ith FSL-1 (≥11.1 ng/ml), flagellin (≥33.3 ng/ml), and Pam3CSK4≥33.3 ng/ml) (Fig. 3). Only FSL-1 induced above-threshold lev-ls of IL-1�, IL-6, IL-8, and TNF�. Pam3CSK4 and flagellinnduced above-threshold levels of IL-1�, IL-6, and IL-8 but notNF�.

.3. Evaluation of adjuvants with known clinical safety recordsing secretion of cytokines by MM6 cells

To verify whether the pyrogenicity threshold in MM6 cells canifferentiate between adjuvants with an established safety pro-le, we tested adjuvants that have been used in clinical trials [22].

F59, Alum, and MPL induced IL-1�, IL-8, TNF-�, and IL-6 below the

yrogenicity threshold in MM6 cells; R848 induced IL-1� and IL-, IL-6, and TNF� at above or near the threshold levels (Fig. 3E–H)

n agreement with reported dose-limiting toxicity in human trials

ig. 3. Pro-inflammatory cytokines are secreted by MM6 cells in presence of TLR ligandf Pam3CSK4 (Pam3), flagellin, and FSL-1 (A–D) or with MF59, ALOH, R848 and MPL (E–H), IL-6 (C and G), and TNF� (D and H). Horizontal dashed lines indicate the average leveean ± SD of triplicate cell cultures are shown. The experiment shown is representative

rd and doses of endotoxin used in cell culture were log-transformed.onfidence interval (CI) were calculated for slopes in double logarithmic coordinates

[23]. In separate experiments, we tested the new investigationalISCOMATRIX adjuvant along with ISCOPREP saponin which is thesaponin fraction used in the manufacture of the ISCOMATRIX adju-vant [24]. The saponin induced IL-1�, IL-8, and TNF�, but not IL-6at equal or higher than threshold levels, (Fig. 4A–D). In contrast,ISCOMATRIX adjuvant did not induce significant levels of any ofthe four cytokines, which supports the improved safety profile inhuman trials compared with the free saponin such as QS21 [24,25].While not completely identical, QS21 and ISCOPREP saponin con-tain the same or similar molecules and would be expected to exhibitthe same activities.

3.4. Induction of pyrogenic responses in rabbits by TLR agonists

In order to demonstrate that the results in MM6 cells were pre-dictive of pyrogenic activity in vivo, FSL-1 and flagellin TLR agoniststhat were most and least active in the MM6 assay were admin-istered in groups of 3 NZW rabbits at low (10 �g/kg) and high(50 �g/kg) doses; 1 ml of PBS was administered to two control rab-bits (Fig. 5A–F). Body temperature was recorded, and blood sampleswere collected in all rabbits prior to treatment (time 0), hourlyfor the first 8 h, and at 24 h (flagellin), or at 48 and 72 h (FSL-1).No increase in body temperature was detected in control rabbits(Fig. 5A). One rabbit that received a low dose of flagellin did notdevelop fever (R4), while in rabbits R2 and R3, a 1.0 ◦C rise in bodytemperature was observed 5 h post injection and resolved 3 h later(Fig. 5B; blue symbols). All animals that received 50 �g/kg of flag-ellin (R5, R6, R8) showed a 1.1–1.3 ◦C increase in body temperature

4–6 h post-treatment; fever sustained for longer periods of timecompared with the low dose group (Fig. 5B; red vs. blue symbols).In the low dose FSL-1-treated rabbits, the maximal increases inbody temperatures were 0.7 ◦C, 0.9 ◦C, and 1.3 ◦C for R9, R11, and

s but not adjuvants. MM6 cells were incubated with the indicated concentrations) for 24 h. Supernatants were assayed for the levels of IL-1� (A and E), IL-8 (B andls of cytokines generated by 0.5 EU endotoxin as described in the legend to Fig. 2.

of 4 separate experiments.

Page 5: Use of human MonoMac6 cells for development of in vitro assay predictive of adjuvant safety in vivo

M. Zaitseva et al. / Vaccine 30 (2012) 4859– 4865 4863

Fig. 4. ISCOPREP saponin alone, but not ISCOMATRIX adjuvant (containing saponin) induced IL-1�, IL-8, and TNF� above safety threshold in MM6 cells. ISCOMATRIX adjuvant( icatedf icate

l

RaerIa

FEli

IMX) and ISCOPREP saponin (ISP) were added to the cultures of MM6 cells at the indor IL-1� (A), IL-8 (B), IL-6 (C), and TNF� (D). Horizontal dashed lines in panels indegend to Fig. 2. Mean ± SD of triplicate cell cultures are shown. See Fig. 3.

12, respectively (Fig. 5C; blue symbols). In the high dose group,ll rabbits developed febrile responses with maximal temperaturelevations of 1.6 ◦C and 2.2 ◦C in rabbits R15 and R18, and R13,

espectively, between 5–8 h post-treatment (Fig 5C; red symbols).n all animals, body temperatures returned to pre-injection levelst 24 h post-treatment.

ig. 5. Administration of flagellin and of FSL-1 induced fever and increased CRP and PGE), or with FSL-1 (C and F) at 10 �g/kg (blue curves) or at 50 �g/kg (red curves). Body tempevels of CRP were measured in blood samples collected at the indicated time points (D–noculated PBS (G) or with FSL-1 at 10 �g/kg (blue curves) or at 50 �g/kg (red curves) by

concentrations. Twenty-four hours later, the cell culture supernatant was assayedthe average levels of cytokines generated by 0.5 EU endotoxin, as described in the

No increase in CRP was detected in control rabbits at any timepoint (Fig. 5D). CRP modestly increased in the flagellin low dosegroup at 8 h and returned to baseline levels at 24 h (Fig 5E, blue

symbols). In contrast, high levels of CRP were found in the flagellinhigh dose group at 24 h post-treatment (Fig. 5E; red symbols). CRPelevations were much more pronounced in the FSL-1 treated groups

2 in rabbits. NZW rabbits were inoculated with PBS (A and D), with flagellin (B anderatures were recorded at indicated times up to 72 h post treatment (A–C) and the

F). PGE2 was measured in the samples of plasma from individual rabbits that wereEIA. Mean ± SD of triplicate assay wells are shown.

Page 6: Use of human MonoMac6 cells for development of in vitro assay predictive of adjuvant safety in vivo

4864 M. Zaitseva et al. / Vaccine 30 (2012) 4859– 4865

F and int or ALOo ll cultu

cCtIbaa

3

wimFrnrvseT

3m

bmcPatcMA

4

a

ig. 6. TLR agonists but not adjuvants induced PGE2 in human primary monocyteshe indicated concentrations of LPS, FSL-1, Pam3CSK4, flagellin, and R848 (A and C)

f cell culture and assayed for PGE2 using the HTRF assay. Mean ± SD of triplicate ce

ompared with the flagellin-treated animals (Fig. 5F vs. Fig. 5E).RP concentrations in all experimental rabbits peaked at 24 h postreatment and returned to pre-challenge levels at 72 h (Fig. 5F).n summary, flagellin and FSL-1 induced dose-dependent rises inody temperature that peaked between 3 and 8 h post-treatmentnd were followed by elevations of CRP that reached a maximumt 24 h post-treatment.

.5. Assessment of PGE2 in the plasma of FSL-1-treated rabbits

To identify earlier biomarkers predictive of pyrogenic reaction,e referred to studies that showed increased PGE2 in the blood and

n the brain of rats within 30 min post-inoculation of LPS [26]. Weeasured PGE2 in the plasma collected from rabbits that received

SL-1 or PBS. No changes in PGE2 were detected in both controlabbits (Fig. 5G). In FSL-1-treated animals, an increase in PGE2 wasoted at 2 h post-treatment and was followed by a second PGE2ise at 5 h in rabbits R22 and R24 (Fig. 5H). Importantly, PGE2 ele-ation preceded the increase in body temperature. In spite of themall group size, these data showed that PGE2 may be a usefularly predictor of pyrogenic response following treatments withLR agonists.

.6. TLR agonists but not adjuvants induce PGE2 in humanonocytes and MM6 cells

Increased levels of PGE2 in the plasma of FSL-1-treated rab-its that developed fever suggested that peripheral circulatingonocytes might produce PGE2 in response to TLR agonists. Mono-

ytes and MM6 cells were incubated with TLR agonists LPS, FSL-1,am3CSK4, flagellin, and R848, and cell culture supernatants weressayed for PGE2 by the HTRF assay (Fig. 6A and C). In both cellypes, TLR agonists induced PGE2 in a dose-dependent manner. Inontrast, no PGE2 was produced in monocytes cultured with Alum,F59, MPL, or Poly I:C adjuvants (Fig. 6B) or in MM6 cultured with

lum or MF59 (Fig. 6D).

. Discussion

We investigated whether in vivo toxicity of TLR agonists anddjuvants can be predicted based on the levels of proinflammatory

MM6 cell line. Monocytes (A and B) or MM6 cells (C and D) were incubated withH, MF59, MPL, Poly I:C (B and D) overnight. Supernatants were collected at the endres are shown. The experiment shown is representative of 4 separate experiments.

cytokines secreted by human MM6 cell line in vitro. A 0.5 EU/mldose of Standard for endotoxin was used as a “safety thresh-old” where the levels of cytokines above the threshold shouldbe indicative of toxicity. At sub-microgram doses, TLR agonistsFSL-1, Pam3CSK4, flagellin and R848 induced cytokines in MM6cells above the safety threshold. In contrast, cytokines secretedin the presence of adjuvants shown to be safe in humans (Alum,MF-59, MPL) were below the safety threshold. ISCOPREP saponinbut not ISCOMATRIX adjuvant elicited IL-1�, IL-8, and TNF� atlevels equal or above the safety threshold, in good agreementwith their safety in humans. The validity of the LPS-based thresh-old was confirmed in rabbit studies where animals developedfebrile responses and elevated CRP following intramuscular injec-tions with FSL-1 and with flagellin. For example, 10–100 ng/ml ofFSL-1 induced cytokines at above safety threshold in MM6 cellsin vitro and 10–50 �g/kg of FSL-1 induced pyrogenic response andincreased CRP in rabbits in vivo. Thus although it is difficult to accu-rately compare the in vitro assay performed in one particular celltype with in vivo animal data, a clear correlation was observedbetween the range of doses that induced pyrogenic response inrabbits and cytokine elevation in MM6 cells. The cell based assaysallow testing of broad dose ranges, which could help to select theappropriate dose for in vivo testing prior to moving into humantrials.

Additionally, we have identified PGE2 as an early marker offebrile responses in vivo. To date our studies demonstrated goodinternal consistency. However, both primary monocytes and MM6cells do not express endosomal TLRs (TLR7 and 9) and may not beappropriate for evaluating adjuvants that target these receptors.

A low rate of AEs occurs in recipients of vaccines: 11.4 AEs per100,000 doses of vaccines (adjuvanted and non-adjuvanted) werereported to the Vaccine Adverse Event Reporting System between1999 and 2001 with fever being the most frequent event, 25.8%[13]. Febrile signals are transmitted to the brain by pyrogeniccytokines with IL-1� being able to induce fever at doses 10–100lower than TNF� and IL-6 in humans and rabbits [27,28] [29–32].In addition, IL-8 was suggested to serve as an early biomarker ofsepsis [33,34]. In our experiments, the levels of cytokines induced

by TLR agonists were higher in monocytes than in MM6 cellswhich may be attributed in part to the absence of constitutivelyactivated casapase-1 in macrophages and in cell lines which
Page 7: Use of human MonoMac6 cells for development of in vitro assay predictive of adjuvant safety in vivo

ccine 3

uadcw

PrcbTdtmsw

pitao

A

i

C

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

M. Zaitseva et al. / Va

nlike monocytes require two signals for IL-1� gene transcriptionnd caspase 1 activation [35]. Thus although MO and MM6 cellsiffered in the levels of cytokine production, all three pyrogenicytokines and IL-8 were produced by primary cells and cell lineith similar pattern.

Detection of cytokines in vivo was substituted by assaying ofGE2 in rabbit’s plasma. FSL-1 induced upregulation of PGE2 inabbits in agreement with previous reports showing increase in cir-ulating PGE2 in rats after injections of LPS [26]. PGE2 is synthesizedy many cell types; however, studies in mice selectively lackingLR4 receptor in hematopoietic cells showed that fever criticallyepends on TLR4-positive hepatic and pulmonary macrophageshat quickly upregulate PGE2 after LPS injection [36,37]. In agree-

ent with these studies, in our system, monocytes and MM6 cellsecreted significant quantities of PGE2 in response to activationith TLR agonists but not adjuvants with good safety records.

Our findings demonstrate the feasibility of measurements ofroinflammatory cytokines and PGE2 in human MM6 cells as an

nitial screening to select against adjuvants with a potential for sys-emic toxicity. This approach can be utilized to select TLR basedttenuated adjuvants and during the manufacturing process toptimize adjuvant formulation.

cknowledgements

We thank Basil Golding and Surender Khurana for critical read-ng of the manuscript.

This project has been funded in part with Federal funds from theritical Path Initiative at CBER, FDA.

eferences

[1] Medzhitov R, Preston-Hurlburt P, Janeway Jr CA. A human homologue ofthe Drosophila Toll protein signals activation of adaptive immunity. Nature1997;388(July (6640)):394–7.

[2] O‘Hagan DT, Valiante NM. Recent advances in the discovery and deliv-ery of vaccine adjuvants. Nature Reviews Drug Discovery 2003;2(September(9)):727–35.

[3] Reis e Sousa C. Toll-like receptors and dendritic cells: for whom the bug tolls.Seminars in Immunology 2004;16(1):27–34.

[4] Akira S, Uematsu S, Takeuchi O. Pathogen recognition and innate immunity.Cell 2006;124(February (4)):783–801.

[5] McKee AS, Munks MW, Marrack P. How do adjuvants work? Importantconsiderations for new generation adjuvants. Immunity 2007;27(November(5)):687–90.

[6] Harandi AM, Davies G, Olesen OF. Vaccine adjuvants: scientific challenges andstrategic initiatives. Expert Review of Vaccines 2009;8(3):293–8, 2011/10/17.

[7] Gupta RK, Relyveld EH, Lindblad EB, Bizzini B, Ben-Efraim S, GuptaCK. Adjuvants—a balance between toxicity and adjuvanticity. Vaccine1993;11(3):293–306.

[8] Waite DC, Jacobson EW, Ennis FA, Edelman R, White B, Kammer R, et al.Three double-blind, randomized trials evaluating the safety and tolerance ofdifferent formulations of the saponin adjuvant QS-21. Vaccine 2001;19(July(28–29)):3957–67.

[9] Alving CR. Design and selection of vaccine adjuvants: animal models andhuman trials. Vaccine 2002;20(May (Suppl. 3)):S56–64.

10] Mestas J, Hughes CC. Of mice and not men: differences between mouse andhuman immunology. Journal of Immunology 2004;172(March (5)):2731–8.

11] Sesardic D, Rijpkema S, Patel BP. New adjuvants: EU regulatory developments.Expert Review of Vaccines 2007;6(October (5)):849–61.

12] Luheshi GN. Cytokines and fever. Mechanisms and sites of action. Annals of theNew York Academy of Sciences 1998;856(September):83–9.

13] Zhou W, Pool V, Iskander JK, English-Bullard R, Ball R, Wise RP, et al.

Surveillance for safety after immunization: Vaccine Adverse Event ReportingSystem (VAERS)—United States, 1991–2001. MMWR Surveillance Summaries2003;52(January (1)):1–24.

14] Blatteis CM. The onset of fever: new insights into its mechanism. Progress inBrain Research 2007;162:3–14.

[

0 (2012) 4859– 4865 4865

15] Ziegler-Heitbrock HW, Thiel E, Futterer A, Herzog V, Wirtz A, RiethmullerG. Establishment of a human cell line (Mono Mac 6) with characteristicsof mature monocytes. International Journal of Cancer 1988;41(March (3)):456–61.

16] O‘Mahony DS, Pham U, Iyer R, Hawn TR, Liles WC. Differential constitutiveand cytokine-modulated expression of human Toll-like receptors in primaryneutrophils, monocytes, and macrophages. International Journal of MedicalSciences 2008;5(1):1–8.

17] Hedayat M, Netea MG, Rezaei N. Targeting of toll-like receptors: a decadeof progress in combating infectious diseases. The Lancet Infectious Diseases2011;11(9):702–12.

18] Visintin A, Mazzoni A, Spitzer JH, Wyllie DH, Dower SK, Segal DM. Regula-tion of toll-like receptors in human monocytes and dendritic cells. Journal ofImmunology 2001;166(January (1)):249–55.

19] Jack CS, Arbour N, Manusow J, Montgrain V, Blain M, McCrea E, et al. TLR sig-naling tailors innate immune responses in human microglia and astrocytes.Journal of Immunology 2005;175(October (7)):4320–30.

20] Eperon S, De Groote D, Werner-Felmayer G, Jungi TW. Human monocy-toid cell lines as indicators of endotoxin: comparison with rabbit pyrogenand Limulus amoebocyte lysate assay. Journal of Immunological Methods1997;207(September (2)):135–45.

21] Hoffmann S, Peterbauer A, Schindler S, Fennrich S, Poole S, Mistry Y, et al. Inter-national validation of novel pyrogen tests based on human monocytoid cells.Journal of Immunological Methods 2005;298(March (1–2)):161–73.

22] Reed SG, Bertholet S, Coler RN, Friede M. New horizons in adjuvants for vaccinedevelopment. Trends in Immunology 2009;30(January (1)):23–32.

23] Tomai MA, Miller RL, Lipson KE, Kieper WC, Zarraga IE, Vasilakos JP. Resiquimodand other immune response modifiers as vaccine adjuvants. Expert Review ofVaccines 2007;6(5):835–47, 2011/10/18.

24] Drane D, Gittleson C, Boyle J, Maraskovsky E. ISCOMATRIX adjuvant for pro-phylactic and therapeutic vaccines. Expert Review of Vaccines 2007;6(October(5)):761–72.

25] Mbawuike I, Zang Y, Couch RB. Humoral and cell-mediated immune responsesof humans to inactivated influenza vaccine with or without QS21 adjuvant.Vaccine 2007;25(April (17)):3263–9.

26] Davidson J, Abul HT, Milton AS, Rotondo D. Cytokines and cytokine inducersstimulate prostaglandin E2 entry into the brain. Pflugers Archiv: Europeanjournal of physiology 2001;442(July (4)):526–33.

27] Banks WA, Kastin AJ, Durham DA. Bidirectional transport of interleukin-1 alphaacross the blood–brain barrier. Brain Research Bulletin 1989;23(December(6)):433–7.

28] Watkins LR, Goehler LE, Relton JK, Tartaglia N, Silbert L, Martin D, et al. Block-ade of interleukin-1 induced hyperthermia by subdiaphragmatic vagotomy:evidence for vagal mediation of immune-brain communication. NeuroscienceLetters 1995;183(January (1–2)):27–31.

29] Ogilvie AC, Hack CE, Wagstaff J, van Mierlo GJ, Erenberg AJ, Thomsen LL, et al.IL-1 beta does not cause neutrophil degranulation but does lead to IL-6, IL-8, and nitrite/nitrate release when used in patients with cancer. Journal ofImmunology 1996;156(January (1)):389–94.

30] Creagan ET, Kovach JS, Moertel CG, Frytak S, Kvols LK. A phase I clinicaltrial of recombinant human tumor necrosis factor. Cancer 1988;62(December(12)):2467–71.

31] Feinberg B, Kurzrock R, Talpaz M, Blick M, Saks S, Gutterman JU. A phaseI trial of intravenously-administered recombinant tumor necrosis factor-alpha in cancer patients. Journal of Clinical Oncology 1988;6(August (8)):1328–34.

32] Dinarello CA, Cannon JG, Mancilla J, Bishai I, Lees J, Coceani F. Interleukin-6 as an endogenous pyrogen: induction of prostaglandin E2 in brain but notin peripheral blood mononuclear cells. Brain Research 1991;562(October (2)):199–206.

33] Santolaya ME, Alvarez AM, Aviles CL, Becker A, King A, Mosso C, et al. Predic-tors of severe sepsis not clinically apparent during the first twenty-four hoursof hospitalization in children with cancer, neutropenia, and fever: a prospec-tive, multicenter trial. Pediatric Infectious Disease Journal 2008;27(June(6)):538–43.

34] Tromp YH, Daenen SMGJ, Sluiter WJ, Vellenga E. The predictive value ofinterleukin-8 (IL-8) in hospitalised patients with fever and chemotherapy-induced neutropenia. European Journal of Cancer 2009;45(4):596–600.

35] Netea MG, Nold-Petry CA, Nold MF, Joosten LA, Opitz B, van der Meer JH, et al.Differential requirement for the activation of the inflammasome for processingand release of IL-1beta in monocytes and macrophages. Blood 2009;113(March(10)):2324–35.

36] Steiner AA, Chakravarty S, Rudaya AY, Herkenham M, Romanovsky AA. Bacterial

lipopolysaccharide fever is initiated via toll-like receptor 4 on hematopoieticcells. Blood 2006;107(May (10)):4000–2.

37] Steiner AA, Ivanov AI, Serrats J, Hosokawa H, Phayre AN, Robbins JR,et al. Cellular and molecular bases of the initiation of fever. PLoS Biology2006;4(September (9)):e284.