Cytokine xxx (2013) xxx–xxx

Contents lists available at SciVerse ScienceDirect

Cytokine

journal homepage: www.journals .e lsev ier .com/cytokine

Increased CD38 expression in T cells and circulating anti-CD38 IgGautoantibodies differentially correlate with distinct cytokine profilesand disease activity in systemic lupus erythematosus patients

Esther J. Pavón a,1,4, Esther Zumaquero a,2,4, Antonio Rosal-Vela a,4, Keng-Meng Khoo b,3,Daniela Cerezo-Wallis a, Sonia García-Rodríguez a, Montserrat Carrascal c, Joaquin Abian c, Richard Graeff d,José-Luis Callejas-Rubio e, Norberto Ortego-Centeno e, Fabio Malavasi f, Mercedes Zubiaur a, Jaime Sancho a,⇑a Instituto de Parasitología y Biomedicina López-Neyra (IPBLN), Consejo Superior de Investigaciones Científicas (CSIC), Parque Tecnológico de Ciencias de la Salud (PTS), Avenida delConocimiento s/n, 18016 Armilla, Spainb Department of Rheumatology, Allergy and Immunology, Tan Tock Seng Hospital, 11 Jalan Tan Tock Seng, Singapore 308433, Singaporec Laboratory of Proteomics CSIC/UAB, Facultat de Medicina, Edifici M-UAB, Campus UAB, 08193 Bellaterra, Spaind Department of Physiology, University of Hong Kong, Hong Kong Special Administrative Regione Unidad de Enfermedades Autoinmunes Sistémicas, Hospital Clínico Universitario San Cecilio, Avenida Doctor Olóriz 16, 18012 Granada, Spainf Laboratory of Immunogenetics, Department of Medical Sciences, University of Torino, Turin, Italy

a r t i c l e i n f o

Article history:Received 20 June 2012Received in revised form 14 February 2013Accepted 16 February 2013Available online xxxx

Keywords:Anti-CD38 autoantibodiesCD38Cytokine profilesIL-10Systemic lupus erythematosus

1043-4666/$ - see front matter � 2013 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.cyto.2013.02.023

Abbreviations: dsDNA, double-stranded DNA; SLEsus; SLEDAI, SLE Disease Activity Index.⇑ Corresponding author. Tel.: +34 958181664; fax:

E-mail address: granada@ipb.csic.es (J. Sancho).1 Present address: Hospital Clínico Universitario San

ades Hepáticas y Digestivas (CIBEREHD), Avenida DoctSpain.

2 Present address: Department of Microbiology, 182531, University of Alabama at Birmingham, AL 35294-

3 Present address: Lonza, 10 Science Park Road, #Science Park II, Singapore 117684, Singapore.

4 These authors contributed equally to this work.

Please cite this article in press as: Pavón EJ et alwith distinct cytokine profiles and diseasej.cyto.2013.02.023

a b s t r a c t

CD38 is a multifunctional protein possessing ADP-ribosyl cyclase activity responsible for both the synthe-sis and the degradation of several Ca2+-mobilizing second messengers. In mammals, CD38 also functionsas a receptor. In this study CD38 expression in CD4+, CD8+, or CD25+ T cells was significantly higher insystemic lupus erythematosus (SLE) patients than in Normal controls. Increased CD38 expression inSLE T cells correlated with plasma levels of Th2 (IL-4, IL-10, IL-13) and Th1 (IL-1b, IL-12, IFN-c, TNF-a)cytokines, and was more prevalent in clinically active SLE patients than in Normal controls. In contrast,elevated anti-CD38 IgG autoantibodies were more frequent in clinically quiescent SLE patients (SLE-DAI = 0) than in Normal controls, and correlated with moderate increased plasma levels of IL-10 andIFN-c. However, clinically active SLE patients were mainly discriminated from quiescent SLE patientsby increased levels of IL-10 and anti-dsDNA antibodies, with odds ratios (ORs) of 3.7 and 4.8, respectively.Increased frequency of anti-CD38 autoantibodies showed an inverse relationship with clinical activity(OR = 0.43), and in particular with the frequency of anti-dsDNA autoantibodies (OR = 0.21). Increased celldeath occurred in CD38+ Jurkat T cells treated with anti-CD38+ SLE plasmas, and not in these cells treatedwith anti-CD38� SLE plasmas, or Normal plasmas. This effect did not occur in CD38-negative Jurkat Tcells, suggesting that it could be attributed to anti-CD38 autoantibodies. These results support thehypothesis that anti-CD38 IgG autoantibodies or their associated plasma factors may dampen immuneactivation by affecting the viability of CD38+ effector T cells and may provide protection from certain clin-ical SLE features.

� 2013 Elsevier Ltd. All rights reserved.

ll rights reserved.

, systemic lupus erythemato-

+34 958181632.

Cecilio, CIBER de Enfermed-or Olóriz 16, 18012 Granada,

5 University Boulevard, SHEL2182, USA.02-12, The Alpha, Singapore

. Increased CD38 expression in Tactivity in systemic lupus

1. Introduction

CD38 is a ubiquitous protein expressed in myriad cells and tis-sues [1–3]. It is a member of a gene family that has the ability toconvert NAD into cADPR [4]. In addition to its ADP-ribosyl cyclaseenzymatic function, CD38 also catalyzes a transglycosylation reac-tion exchanging the terminal nicotinamide group of the substrateNADP with nicotinic acid resulting in the production of NAADP[5]. Both cADPR and NAADP are Ca2+-mobilizing messengersthought to be important in various cellular signaling events [6,7].

Increased CD38 expression in different cell types has been asso-ciated with a number of human diseases [8,9]. Thus, CD38

cells and circulating anti-CD38 IgG autoantibodies differentially correlateerythematosus patients. Cytokine (2013), http://dx.doi.org/10.1016/

2 E.J. Pavón et al. / Cytokine xxx (2013) xxx–xxx

represents a reliable negative prognostic marker in chronic lym-phocytic leukemia (CLL) [10]. However, CD38 expression in CLL Bcells also affects expansion and proliferation of the neoplasticclones, and it is, therefore, considered as part of a network sustain-ing growth and survival of CLL cells [11].

Increased percentages of B cells positive for CD38 have beenconsistently shown in SLE patients by different groups [12–14]. Ac-tive-SLE patients have circulating CD38bright Ig-secreting cells thatare not found in Normal individuals [12]. This plasma cell subsetdisappears from circulation during treatment with humanizedanti-CD154 mAb, and it is associated with decreases in anti-dou-ble-stranded DNA (anti-dsDNA) Ab levels, proteinuria, and SLE dis-ease activity index [12]. In this sense, oxytocin at physiologicalconcentrations influences the distribution of surface CD38 and in-creases the growth rate of plasma cells [15]. Moreover, in SLE pa-tients with active disease, B cells expressing high levels of CD38spontaneously produce IgG class anti-dsDNA in vitro, whereas per-sistence of CD38+ B cells during periods with clinically quiescentdisease seems to underlie hypergammaglobulinaemia but notanti-dsDNA production [14]. Likewise, T cells from active SLE pa-tients over-express CD38 [16–18], which might be playing a rolein modulating TCR signaling [18].

It has been postulated that CD38 may play a role in the patho-genesis of diabetes mellitus, which is likely related to increasedanti-CD38 autoantibodies detected in Japanese type 2 [19], as wellas Caucasian type 1 and 2 diabetic patients [20]. Autoantibodies toCD38 are also detected in patients with chronic autoimmune thy-roiditis and Graves’ disease [21]. However, no similar studies havebeen carried out in other autoimmune diseases, such as SLE.

The aim of the present study is to define the prevalence of auto-antibodies to CD38 in SLE subjects, and to explore whether there isan association of these autoantibodies with cytokine plasma levels,clinical activity, and CD38 surface expression in T cells.

2. Materials and methods

2.1. Patients and controls

A total of 69 SLE consecutive patients (63 female, 6 male) attend-ing the outpatient clinic were selected. No exclusion criteria wereused for the selection of patients in order to obtain a heterogeneoussample representative of a broad spectrum of clinical and laboratoryphenotypes. All SLE patients fulfilled the American College of Rheu-matology criteria for SLE [22]. Disease activity was measured usingthe SLE Disease Activity Index (SLEDAI) [22]. The SLE patients hada median SLEDAI of 2 (range: 0–20). A total of 71 Normal control sub-jects were selected at the local blood bank (43 female, 28 male). Allpatients and Normal controls were Caucasians. The study protocolwas approved by the Hospital Clínico San Cecilio, and CSIC ReviewBoard and Ethics Committees. Written informed consent was ob-tained from all participating patients and volunteers according tothe Declaration of Helsinki.

2.2. Blood plasma samples

Blood was collected by the BD Vacutainer system into K2-EDTAtubes (BD Diagnostics, NJ, USA) and plasma was separated fromcells by density gradient centrifugation over HISTOPAQUE�-1077(Sigma–Aldrich, St. Louis, MO). The supernatant was collected,checked for the absence of cells by light microscopy, and fraction-ated in aliquots that were stored at �80 �C [18].

2.3. Recombinant CD38 proteins

The recombinant His-tagged GST-CD38 fusion protein used as thetarget antigen in the ELISA and Western blots experiments was

Please cite this article in press as: Pavón EJ et al. Increased CD38 expression in Twith distinct cytokine profiles and disease activity in systemic lupusj.cyto.2013.02.023

kindly obtained from Prof. C. F. Chang at the National University ofSingapore [23]. The other recombinant CD38 protein used in Wes-tern blot experiments (rCD38) lacks the four putative N-linked gly-cosylation sites and the intracellular and transmembrane spanningregions and was produced as a soluble, enzymatically active CD38in yeast [24].

2.4. Analysis of anti-CD38 autoantibodies by enzymatic immunoassay

Anti-CD38 IgG and IgM concentrations were measured by ELI-SA. Briefly, the His-tagged GST-CD38 fusion protein (Fr #20) wasadded to 96-well polystyrene plates (Nunc, Roskilde, Denmark)in coating buffer (0.05 M Na2CO3/NaHCO3, pH 9.6) and incubatedfor 16 h at 4 �C. Plates were then washed twice with phosphatebuffered-saline (PBS) containing 0.1% Tween 20 (PBS/Tween) andblocked for 2 h with 5% BSA in PBS. Plasma samples were diluted1:4 in PBS/Tween and 100 ll was added to each well and incubatedfor 2 h at 37 �C. Following four washings with PBS/Tween, horse-radish peroxidase-conjugated goat anti-human IgG (Sigma–Al-drich, St. Louis, MO) anti-serum was added and incubated for 1 h.Plates were then washed 6 times and developed with 100 llo-phenylenediamine dihydrochloride (OPD) (Sigma–Aldrich, St.Louis, MO). Substrate color development was stopped by 0.5 M sul-furic acid and absorbance was measured at 450 nm using a Versa-Max tunable microplate reader (Molecular Devices, Chicago, IL)with the results expressed in optical density units (ODU). ForIgM measurement, plasma samples were incubated as before anddetected with purified goat anti-human IgM (Sigma–Aldrich, St.Louis, MO) followed by HRP-conjugated rabbit anti-goat IgG (Sig-ma–Aldrich, St. Louis, MO). The development was as describedabove. Validation of both assays was done in-house. Control plas-ma derived from two Normal controls and two SLE patients wereadded to each assay to determine inter-assay variation, whichwas always less than 15%. The intra-assay variation of triplicateswas always less than 5%. A positive control was performed usingthe rabbit polyclonal IgG against human CD38 (H-170) (Santa CruzBiotechnology Inc., Santa Cruz, CA) raised against an epitope corre-sponding to amino acids 1–170 mapping at the amino terminus ofhuman CD38. This was followed with washing and incubation withthe secondary antibody (goat anti-rabbit IgG) conjugated withhorseradish peroxidase (Promega Co., Madison, WI). Other anti-bodies used were HB136, OKT10, and OKT3 mAbs.

2.5. Identification of GST-CD38 by mass spectrometry

Samples were analyzed by LC-MS/MS as described [25] (seeSupplemental material and methods)

2.6. Cytokine assay

The Bio-Plex Precision Pro Human Cytokine 10-Plex kit assay(Bio-Rad, Hercules, CA) was used to simultaneously test 10 cyto-kines: IL-1b, IL-2, IL-4, IL-5, IL-6, IL-10, IL-12 (p70), IL-13, IFN-cand TNF-a. Assays were performed according with manufacturers’protocol. Analyses of experimental data were performed usingfive-parameter logistic curve fitting to standard analyte values.

2.7. Ca2+ analysis

Changes in intracellular calcium concentrations in Jurkat T cellswere measured as described [26].

2.8. CD38 expression assay and assessment of plasma-induced cell death

The percentage of apoptotic cells was determined by flowcytometry analysis by evaluating the modification of scatter

cells and circulating anti-CD38 IgG autoantibodies differentially correlateerythematosus patients. Cytokine (2013), http://dx.doi.org/10.1016/

E.J. Pavón et al. / Cytokine xxx (2013) xxx–xxx 3

parameters in freshly isolated cultured cells incubated with orwithout the relevant plasmas or antibodies [27]. When cells aredying they decrease their size (forward characteristics: FSC) andincrease their complexity (scatter characteristics: SSC), and theycan be clearly distinguished from live cells. Briefly, Jurkat T cells(0.5–1 � 106) were resuspended in 100 ll of complete RPMI med-ium. Plasma from SLE patients, or Normal controls was then addedat several volume ratios (1:1; 1:5 and 1:10), and after incubationfor 1 h at 37 �C, these cells were washed once, and made to reactwith 0.1 lg of FITC-labeled anti-CD38 mAb HB136 (IgG1), or APC-labeled anti-CD38 mAb IB6 (IgG2b) for 20 min at room tempera-ture, and then analyzed by FACS. FITC-labeled mouse IgG1, orAPC-labeled mouse IgG2b were used as isotype-matched controls.Then, cells were first analyzed for their FSC and SSC as comparedwith cells non-incubated with plasma. Spontaneous or specific celldeath was defined as the percentage of FSClowSSChigh Jurkat cellsupon treatment for 1 h at 37 �C with medium alone or plasma,respectively. Then, CD38 expression was analyzed gating thosecells with Normal FSC and SSC.

2.9. Flow cytometry analysis

Peripheral blood mononuclear cells (PBMC) or purified T cellswere analyzed for surface expression of CD3, CD4, CD8, CD38,CD19 and CD56 by single or double-staining using FITC-, or PE-la-beled anti-CD3, anti-CD38, anti-CD19, and anti-CD56 mAbs in therelevant combinations. Compensation settings were adjusted usingsingle-stained PBMC, or T cell samples. Isotype-matching labeledantibodies were used to calculate the nonspecific staining. PBMCand lymphocytes were gated according with their forward andscatter characteristics [18]. Two color immunofluorescence analy-ses were performed on a FACScan, or a FACScalibur flow cytometer(BD Biosciences, San Jose, CA), using the CellQuest Pro (BD Biosci-ences), and FlowJo (Tree Star, Inc. San Carlos, CA) software.v

2.10. Statistical analysis

Values are given as the mean ± standard error (SE), unlessotherwise indicated. Comparisons between two groups were per-formed with the Student’s t test for normally distributed variablesand the Mann–Whitney U test for non-normal variables. Each indi-vidual marker was evaluated by a receiver-operating characteristic(ROC) curve, the area under the ROC curve (AUC), and sensitivity at95% specificity (ROC [0.005]), or higher. Fisher’s exact test was usedto compare two proportions. Sensitivity, specificity, ORs and P val-ues presented from Fisher’s exact test comparing cytokine levels,or autoantibody levels were dichotomized based on a previousROC analysis of the data to determine the cut-off values at thehighest likelihood ratio (LR) and specificity to discriminate SLEfrom Normal controls. The one-way ANOVA test was used to com-pare mean values of three or more groups. If the mean values werestatistically significant by the ANOVA test, then the Tukey–Kramertest was used to compare all pairs of groups. In some occasions theKruskal–Wallis nonparametric test was used to compare mediansinstead, followed by the Dunn’s multiple comparison test. Differ-ences were considered statistically significant for P values < 0.05.Analysis of the data was done using the GraphPad Prism version5.04 software (GraphPad Software, Inc., San Diego, CA).

3. Results

3.1. Increased levels of anti-CD38 IgG autoantibodies in blood plasmafrom SLE patients

A 56 kDa GST-CD38 recombinant protein comprised of theextracellular domain of human CD38, beginning with Pro 44, and

Please cite this article in press as: Pavón EJ et al. Increased CD38 expression in Twith distinct cytokine profiles and disease activity in systemic lupusj.cyto.2013.02.023

a 6 � His-containing GST fusion protein [23] was unambiguouslyidentified as GST-CD38 by tandem mass spectrometry (Supple-mental Table 1). Then, this protein was used to test for the pres-ence of anti-CD38 autoantibodies in SLE vs Normal plasmas byWestern blotting at 1:100 plasma dilution, and ELISA at 1:4 dilu-tion (Supplemental Figs. 1 and 2a). The ELISA data were comparedwith those obtained with plasmas at 1:100 dilution tested againstanother recombinant CD38 protein, rCD38 [24] that lacks irrele-vant epitopes such as GST or a poly-histidine tail, which may causenon-specific binding. The data obtained showed a linear correla-tion between them (Supplemental Fig. 2b and c), suggesting thatthe presence of GST and the poly-histidine tail in GST-CD38 didnot affect the anti-CD38 reactivity of plasmas.

Fig. 1a shows the distribution of anti-CD38 IgG levels among theNormal controls and SLE patients. Anti-CD38 IgG levels differenti-ate very well between SLE patients and Normal controls, as con-firmed by the ROC analysis. Thus, the area under the ROC curve(AUC) for anti-CD38 IgG was 0.60 ± 0.05 (mean ± SE; P = 0.0374).Setting the threshold value of 1.25 ODU by ROC analysis, 13% ofall SLE patients tested were positive for anti-CD38 IgG autoanti-bodies vs 2.8% of Normal controls (P = 0.0297, Fisher’s exact test).With this cut-off value the assay showed a sensitivity of 13% anda specificity of 97.2%, with a likelihood ratio (LR) of 4.63 to getanti-CD38+ in SLE patients and not in Normal controls. With acut-off of 1.163 the sensitivity increased up to 20.3%, with still areasonable specificity of 94.4% and a likelihood ratio of 3.6. Withthe cut-off of 1.163, 14/69 (20.3%) of SLE patients were positivefor anti-CD38 IgG antibodies vs 4/71 (5.6%) of Normal controlsand this difference was statistically significant (P = 0.0115, Fisher’sexact test).

SLE patients were then segregated according with their SLEDAIscores, a validated measure of global disease activity [22], in twogroups: clinically quiescent (SLEDAI = 0, n = 26), and clinically ac-tive SLE (SLEDAI > 0, range: 2–20, median: 4, n = 43). Then, anti-CD38 IgG levels of the quiescent SLE group were compared withthose of the Normal control group. The AUC was 0.67 ± 0.07;P = 0.0099, and with a cut-off of 1.25 19% of the SLE patients withSLEDAI = 0 were positive for anti-CD38 IgG autoantibodies(P = 0.0140, Fisher’s exact test). With a cut-off of 1.163, 31% ofthe quiescent SLE patients were anti-CD38 IgG+ as compared with5.6% of the Normal controls, and the difference was statisticallysignificant (P = 0.0024, Fisher’s exact test). The LR for anti-CD38IgG+ in quiescent SLE relative to Normal controls ranged from6.83 to 5.5 for cut-offs of 1.25 and 1.163, respectively.

In the active SLE group (SLEDAI > 0), with the cut-off of 1.25, thenumber of SLE patients positive for anti-CD38 IgG autoantibodiesdropped to 9%, which was not statistically significant as comparedwith that of the Normal controls (P = 0.1966, Fisher’s exact test).With the cut-off of 1.163 the percentage of anti-CD38 IgG+ in-creased in the active SLE group up to 13.95%, which again it wasnot statistically significant compared with that in Normal controls.When anti-CD38 IgG autoantibodies levels were compared be-tween clinically active and quiescent SLE patients the odds ratio(OR) was of 0.43 (Table 4), which was indicative that anti-CD38IgG autoantibodies were inversely associated with the risk of ahigher disease activity (57% decrease). However, these differencesdid not reach statistical significance (P = 0.2816).

Therefore, anti-CD38 autoantibodies were more prevalent inSLE patients with inactive disease than in SLE patients with activedisease, when they were compared with Normal controls. In con-trast, the frequency of anti-dsDNA antibodies was significantlyhigher in SLE patients with clinically active disease than in quies-cent patients (32.5% vs 7.7%, respectively, P = 0.0201, Fisher’s exacttest). In comparison with the frequencies of anti-CD38 IgG anti-bodies, the frequency of anti-dsDNA was significantly higher inclinically active patients (32.5% vs 9.3%, P = 0.0155, Fisher’s exact

cells and circulating anti-CD38 IgG autoantibodies differentially correlateerythematosus patients. Cytokine (2013), http://dx.doi.org/10.1016/

Fig. 1. Increased levels of anti-CD38 autoantibodies of IgG class (anti-CD38 IgG) in SLE patients. (a) Dot diagram of anti-CD38 IgG autoantibody levels in SLE patients (closedblack circles), Normal controls (open circles), SLE patients with SLEDAI = 0 (clinically quiescent) (black squares), or with SLEDAI from 2 to 20 (clinically active) (open squares).(b) Dot diagram of anti-CD38 IgG autoantibodies in SLE patients of 20–40 years of age (closed circles), Normal controls of 20–40 years of age (open circles), SLE patients of>40 years-old (closed circles), or Normal controls of >40 years of age (open circles). (c) Dot diagram of anti-CD38 IgM autoantibody levels in SLE patients (closed circles),Normal controls (open circles), SLE patients with SLEDAI = 0 (black squares), or with SLEDAI from 2 to 20 (open squares). In the three panels, the horizontal line in each dotdiagram denotes the median value. Dashed line represents the cut-off value above of which samples were considered positive for anti-CD38 IgG (a and b), or IgMautoantibodies (c). Numbers on top of each dot diagram represent the number of individuals (%) that showed anti-CD38+ autoantibodies of IgG class (panels (a and b), or IgMclass (panel (c)).

4 E.J. Pavón et al. / Cytokine xxx (2013) xxx–xxx

test) but not in inactive patients, nor in the whole SLE population(data not shown). Other laboratory parameters, as serum comple-ment C3, C4 and cytokine plasma levels were, however, quite sim-ilar in anti-CD38+ and anti-dsDNA+ SLE patients (see SupplementalTable 3, and Supplemental Fig. 3).

Anti-CD38 IgG autoantibody profiles were also examined byage-matched groups. As a whole the SLE patients and Normal con-trols were not age-matched (40 vs 25 years of age, median values;P < 0.0001, Mann–Whitney test). However, they were age-matchedby separating them in two groups, group I comprised subjects from20 to 40 years-old, and group II comprised subjects > 40 years-old(see Supplemental Table 2). As shown in Fig. 1b, the anti-CD38IgG median levels in each SLE age group was significantly higherthan that in its respective age-matched Normal control group. Itis clear, however, that the titers of anti-CD38 antibodies dimin-

Please cite this article in press as: Pavón EJ et al. Increased CD38 expression in Twith distinct cytokine profiles and disease activity in systemic lupusj.cyto.2013.02.023

ished significantly with age in both SLE and Normal controls.Regarding the frequencies, with a cut-off value of 1.25 ODU, SLEpatients (20–40 years of age) were more frequently positive foranti-CD38 autoantibodies compared to age-matched Normal con-trols (17.5% vs 3.2%; P = 0.0270, Fisher’s exact test), and more fre-quently positive than SLE patients > 40 years of age, but thedifference was not statistically significant (17.5% vs 6.4%;P = 0.2818, Fisher’s exact test).

Fig. 1c shows the distribution of anti-CD38 IgM levels amongthe Normal donors and SLE patients. There was no significant dif-ference in anti-CD38 IgM median levels between SLE patientsand Normal donors (P = 0.5568). A higher proportion of clinicallyactive SLE patients showed increased levels of anti-CD38 IgM anti-bodies than quiescent SLE patients or than Normal controls (6.9%,3.8% and 1.4%, respectively), although these differences were not

cells and circulating anti-CD38 IgG autoantibodies differentially correlateerythematosus patients. Cytokine (2013), http://dx.doi.org/10.1016/

Table 2Analysis of the correlation of SLE anti-CD38 IgG autoantibodies with cytokine plasmalevels.

Cytokine Anti-CD38 IgG+a Anti-CD38 IgG�b

r P valuec r P valued

IL-1b 0.5061 nse 0.08037 nsIL-2 0.6341 ns �0.03848 nsIL-4 0.5644 ns 0.003930 nsIL-5 0.6225 ns �0.1387 nsIL-6 0.5667 ns �0.1821 nsIL-10 0.8167 �f �0.1953 nsIL-12 (p70) 0.5515 ns �0.05426 nsIL-13 0.5626 ns 0.1426 nsIFN-c 0.6795 �f �0.04527 nsTNF-a 0.5496 ns �0.02328 ns

Positive correlation coefficients, r, which are statistically significant are highlightedin bold.

a n = 9 Pairs.b n = 54 Pairs.c P values were obtained using Spearman’s (IL-10, IL-6) or Pearson’s (IL-1b, IL-2,

IL-4, IL-5, IL-12, IFN-c, TNF-a) correlation tests.d P values were obtained using Spearman’s correlation test.e ns: Not significant. P values > 0.05.f �P values from 0.01 to 0.05 were considered significant.

E.J. Pavón et al. / Cytokine xxx (2013) xxx–xxx 5

statistically significant. No correlation was found between anti-CD38 IgG and anti-CD38 IgM levels in any of the groups tested(data not shown).

3.2. Distinct blood plasma cytokine profile in SLE patients

In PBMCs, binding of agonistic anti-CD38 mAbs induces releaseof pro-inflammatory cytokines, i.e., IL-1, IL-6, and TNF-a [28], andsimilar effects may be triggered by anti-CD38 autoantibodies. Onthe other hand, increased levels of either proinflammatory, oranti-inflammatory cytokines in blood plasma or serum from SLEpatients have been reported earlier [29]. We used multiplex beadarray technology to simultaneously measure the levels of 10 differ-ent cytokines in blood plasma from SLE patients and Normal con-trols that had previously tested for anti-CD38 IgG autoantibodies.Cytokine plasma levels were similar between females and maleswithin the Normal control group (data not shown). As shown in Ta-ble 1, SLE patients had increased plasma levels of both Th1 (IL-12,IFN-c, and TNF-a) and Th2 (IL4, IL-6, IL-10, and IL-13) cytokines ascompared with Normal controls. Moreover, IL-2 plasma levels werealso elevated. IL-2 is considered a Th0 cytokine because it contrib-utes to the expansion of all T cell types including effector Th1, Th2,Th17 cells, memory CD8+ T cells, and regulatory T cells [30]. IL-2can also contribute to the expansion of CD25+ B cells, and indeedIL-2 is required for polyclonal Ig production by committed humanB cells [31]. Neither IL-1b, nor IL-5 plasma levels were significantlyelevated in SLE patients (Table 1).

To test whether anti-CD38 IgG and cytokine levels were corre-lated or not, SLE patients were first analyzed as a whole. Spear-man’s correlation analysis showed a weak, although statisticallysignificant, association of anti-CD38 IgG levels with IL-13 plasmalevels (r = 0.2499, P = 0.0483, n = 63). Then SLE patients were splitin two subgroups: anti-CD38 IgG positive (anti-CD38+, above thecut-off value of 1.25) and anti-CD38 IgG negative (anti-CD38-, be-low 1.25). Both subgroups showed a similar cytokine profile thanthat of the whole SLE group (data not shown). However, in theanti-CD38+ group, a significant positive correlation was found be-tween anti-CD38 IgG autoantibodies and IL-10, or IFN-c plasmalevels. In contrast, in the anti-CD38- group there was no correlationof anti-CD38 IgG levels with any of the cytokines tested (Table 2).

3.3. Increased CD38 expression in SLE T cells correlates with increasedplasma levels of Th1 and Th2 cytokines

PBMCs from patients with SLE (n = 67) and Normal controls(n = 25) were analyzed for expression of CD38 in CD3+ T cells. Sur-face CD38 expression was similar between females and males

Table 1Increased cytokine plasma levels in SLE patients.

Cytokine SLE (n = 63) Controls (n = 67)

Mediana (IQR)b Mediana (IQR)b P valuec

IL-1b 0.25 (0.2–0.89) 0.21 (0.21–0.21) 0.3694IL-2 16.98 (1.86–35.04) 1.05 (1.05–7.56) <0.0001d

IL-4 0.59 (0.3–1.57) 0.12 (0.12–0.25) <0.0001IL-5 1.42 (0.51–4.25) 2.02 (1.42–2.02) 0.8492IL-6 9.1 (1.5–33.79) 0.82 (0.82–1.5) <0.0001d

IL-10 5.35 (1.33–11.21) 0.9 (0.9–4.29) <0.0001d

IL-12 (p70) 0.93 (0.46–2.92) 0.18 (0.16–1.1) 0.0004d

IL-13 2.2 (0.55–9.16) 0.28 (0.28–2-2) 0.0005d

IFN-c 1.51 (0.4–4.07) 0.33 (0.33–0.78) 0.0010d

TNF-a 0.82 (0.31–2.37) 0.14 (0.14–0.63) <0.0001d

Values that are statistically significant are highlighted in bold.a pg/ml.b IQR, interquartile range.c Mann–Whitney test.d Statistically significant, P < 0.05.

Please cite this article in press as: Pavón EJ et al. Increased CD38 expression in Twith distinct cytokine profiles and disease activity in systemic lupusj.cyto.2013.02.023

within the Normal control group (data not shown). In contrast,the proportion of SLE T cells expressing CD38 varied from 2.5% to78%, and a threshold was set at 23.68% (95% percentile), abovewhich samples were classified as CD38+ T cells. On the basis of thiscut-off value, 26.9% of the SLE patients analyzed were consideredCD38+ vs 3.8% of Normal controls (OR = 9.2, 95% CI: 1.2–72.8,P = 0.0193, Fisher’s exact test) (Fig. 2b). The differences with Nor-mal controls were even higher in clinically active SLE patients(34.1%, OR = 12.96, 95% CI: 1.6 to 106.0, P = 0.0054), while therewere not statistically significant differences in quiescent SLE pa-tients vs Normal controls (15.4%, P = 0.3497), nor between bothSLE subgroups (P = 0.1562).

As shown in Fig. 2c, when SLE patients were stratified by age(Group I: 20–40 years-old (n = 35) vs Group II: >40 years-old(n = 30)) no major differences were observed in the frequency ofCD38+ T cells between the two SLE age groups (28.6% vs 25%,P = 0.7885). However, the difference in percentages between theage-matched 20–40 SLE group and 20–40 Normal group was statis-tically significant (28.6% vs 4.8%, P = 0.0389, Fisher’s exact test). ForNormal controls it seemed that there was a decreased CD38expression in T cells in the group > 40 years as compared withthe group 20–40, although the low number of subjects in the firstgroup precluded any strong conclusion about it. Likewise, the lackof statistical significance between the percentages of CD38+ T cellsin SLE > 40 vs Normal > 40 (25% vs 0%) was likely due to the lownumber of Normal subjects of that age.

CD38 expression in T cells showed a positive correlation withplasma levels of a number of cytokines, including both Th2 cyto-kines such as IL-4, IL-10, and IL-13 and Th1 cytokines such as IL-1b, IL-12, IFN-c, and TNF-a (Table 3). Moreover, plasma levels of9 out of 10 cytokines tested were significantly increased in SLE pa-tients with CD38+ T cells compared to SLE patients with CD38- Tcells (data not shown).

In B cells (CD19+) and NK cells (CD56+) the differences in CD38expression between SLE and Normal controls were statistically sig-nificant (data not shown). The ROC curve for CD38 expression ineach cell type showed significant ability to differentiate SLE pa-tients from Normal controls, as AUCs were all significantly above0.5. Similar percentages of individuals were CD38+ in B, or NK cellsfrom clinically active and quiescent SLE patients. The data alsoshowed that CD38 expression in B cells correlated exclusively withthe plasma levels of IL-10 (r = 0.3481, P = 0.0471, Spearman’s

cells and circulating anti-CD38 IgG autoantibodies differentially correlateerythematosus patients. Cytokine (2013), http://dx.doi.org/10.1016/

Fig. 2. Increased CD38 expression in lymphoid cells from SLE patients. (a) PBMCs from Normal controls and SLE patients were double-stained with monoclonal Abs againstCD3 (X-axis) and CD38 (Y-axis) and analyzed by FACS for CD38+CD3+ cells. Left panels were gated on lymphocytes in FSC vs SSC dot plots of PBMCs, and the percentages arerelative to total lymphocyte population. Percentages of CD38+CD3+ cells in the dot plots are from a representative experiment of one Normal control (upper panel) and oneSLE patient (lower panel). Right panel histograms were obtained by gating on CD3+ cells in left panels. It represents an overlay of CD38 expression within CD3+ T cells from aSLE patient (green), or a Normal control (red), with a gate set on CD38+ cells. The percentages of CD38+CD3+ cells were in relation to the total CD3+ T cell population. (b) Dotdiagrams of CD38 expression in CD3+ T cells from SLE patients (left closed circles), Normal controls (open circles), clinically quiescent SLE patients (closed circles), andclinically active SLE patients (closed circles). The dashed line represents the cut-off value above which T cells were considered CD38+ (95% percentile). Numbers on top of eachdot diagram represent the number of individuals (%) that showed CD38+ T cells. (c) Dot diagrams of CD38+ T cells from SLE patients (closed circles) and Normal controls (opencircles) segregated in two age-matched groups: 20–40 years-old and >40 years-old. (For interpretation of the references to colour in this figure legend, the reader is referredto the web version of this article.)

6 E.J. Pavón et al. / Cytokine xxx (2013) xxx–xxx

correlation), whereas CD38 expression in NK cells did not correlatewith the plasma levels of any of the cytokines tested (data notshown).

The SLE cohort studied in our laboratory comprised many pa-tients being successfully maintained on low doses of corticoste-roids (S), and/or hydroxychloroquine (HCQ), and some requiringimmunosuppressive (IS) therapy. The three major groups were:1) patients treated with HCQ alone (29.4%), 2) treated with HCQplus S (25%), or 3) treated with HCQ + S + IS (22.05%). Very fewwere treated with S alone (7.3%), and none of them were treatedwith IS alone. There were not statistically significant differencesbetween these groups regarding anti-CD38 autoantibodies, CD38expression, and cytokine median levels, although patients treatedwith HCQ + PD + IS showed a trend toward a decrease in anti-CD38 IgG, and an increase in both IL-6 and CD38+ T cells medianlevels (Supplemental Fig. 4). Since these therapies are adminis-tered to patients with an active disease, these data may reflect highdisease activity, or may reflect a relative lack of effect of the agentson the relevant immunological pathway.

3.4. Predictive value of IL-10 plasma levels to discriminate betweenactive and quiescent SLE patients

The relative importance of individual cytokines in classifyingbetween the different disease states (clinically active and quies-cent) and Normal controls was determined by calculating theirrespective ORs from analyses of 2 � 2 contingency tables. IL-2,IL-6, IL-10 and IL-13 were chosen because they showed the highestmedian values above those in Normal controls and they repre-sented both anti-inflammatory and pro-inflammatory responses.As shown in Table 4, all cytokines tested showed higher OR scoresthan CD38+ T cells, or anti-CD38 IgG autoantibodies to discriminate

Please cite this article in press as: Pavón EJ et al. Increased CD38 expression in Twith distinct cytokine profiles and disease activity in systemic lupusj.cyto.2013.02.023

between SLE patients and Normal controls, being IL-10 the onewith the highest OR score. Likewise, the highest sensitivity andOR to discriminate active SLE patients from Normal controls wasfor high levels of IL-10, followed by IL-2, IL-6, and IL-13. Increasedexpression of CD38 in T cells had also predictive value, with thesame sensitivity as IL-13 (33.3%), although with a lower OR (12.5vs 36.5). Higher levels of anti-CD38 autoantibodies had not signif-icant predictive value to distinguish between active SLE patientsand Normal controls.

In contrast, IL-13 showed the highest sensitivity and OR for dis-criminating quiescent SLE patients from Normal controls, and IL-10the lowest. Increased levels of anti-CD38 autoantibodies showedthe same sensitivity as IL-10 to discriminate quiescent SLE patientsfrom Normal controls (19.2%) and a lower OR (8.2 vs 16.7). CD38+ Tcells had not significant predictive value.

Of note is that IL-10 was the only cytokine with enough sensi-tivity and specificity to discriminate between active and quiescentSLE patients, with an OR of 3.7 that was close to the OR of 5.8 foranti-dsDNA autoantibodies, which both were statistically signifi-cant. Higher levels of anti-CD38 autoantibodies showed an inverserelationship with clinical activity (OR = 0.43), although this valuewas not statistically significant (P = 0.2816). However these differ-ences were statistically significant when the frequency ofanti-CD38 autoantibodies was compared with the frequency ofanti-dsDNA autoantibodies within the active SLE group(OR = 0.21 (CI = 0.06–0.7), P = 0.0155, Fisher’s exact test), which isindicative of negative association.

3.5. Altered phenotype of SLE T cells

In a previous paper, we had evidence that SLE T cells show analtered phenotype similar to that of activated/effector T cells,

cells and circulating anti-CD38 IgG autoantibodies differentially correlateerythematosus patients. Cytokine (2013), http://dx.doi.org/10.1016/

Table 4Predictive value of cytokines.

Variable Sensitivity Specificity ORa 95% CIb P valuec

SLE vs NormalIL-10 37 98.6 41.09 5.4–313.1 <0.0001IL-13 32.4 98.6 35.04 4.6–267.6 <0.0001IL-6 32.4 98.6 33.6 4.4–256.7 <0.0001IL-2 31.08 98.6 31.6 4.1–241.5 <0.0001CD38 + T cells 26.1 96.1 8.8 1.1–70.5 0.0183anti-CD38 13.04 97.2 5.2 1.1–24.9 0.0297

Active SLE vs NIL-10 46.8 98.6 61.6 7.9–481.3 <0.0001IL-2 37.5 98.6 42 5.4–329.2 <0.0001IL-6 35.4 98.6 38.4 4.9–301.5 <0.0001IL-13 33.3 98.6 36.5 4.6–287.3 <0.0001CD38+ T cells 33.3 96.1 12.5 1.5–102.8 0.005anti-CD38 9.3 97.2 3.5 0.6–20.2 0.1966

Quiescent vs NIL-13 30.8 98.6 32.4 3.8–276.4 <0.0001IL-6 26.9 98.6 25.8 3.0–222.8 0.0003IL-2 19.2 98.6 16.7 1.8–150.8 0.005IL-10 19.2 98.6 16.7 1.8–150.8 0.005anti-CD38 19.2 97.2 8.2 1.5–45.5 0.014CD38+ T cells 15.4 96.1 4.5 0.5–43.8 0.3497

Active vs quiescentanti-dsDNA 32.6 92.3 5.8 1.2–28.1 0.0201IL-10 46.8 80.8 3.7 1.2–11.5 0.0238CD38+ T cells 33.3 84.6 2.7 0.8–9.7 0.1517IL-2 37.5 80.8 2.5 0.8–7.9 0.1225IL-13 25 84.6 1.8 0.5–6.4 0.3911IL-6 35.4 73.08 1.5 0.5–4.2 0.6042anti-CD38 9.8 80.7 0.43 0.1–1.8 0.2816

ORs that are statistically significant are highlighted in bold.a OR = odds ratio.b 95% CI: 95% Confidence interval.c Sensitivity, specificity, ORs and P values presented from Fisher’s exact test

comparing cytokine levels, or autoantibody levels were dichotomized based on thecut-off values at the highest likelihood ratio and specificity to discriminate SLE fromNormal controls (IL-2: 24.81; IL-6: 25.06; IL-10: 8.5; IL-13: 7.9; anti-CD38 IgG:1.25). The cut-off value for CD38+ T cell expression was the 95% percentile of theNormal control population (23.6%). Quiescent SLE (SLEDAI = 0), and clinically activeSLE (SLEDAI > 0, range: 2–20).

Table 3Analysis of the correlation of the percentage of SLE CD38+CD3+ T cells with thecytokine plasma levels.

Cytokinea CD38+CD3+ T cellsa

Spearman r P valueb P value summaryc

IL-1b 0.5012 0.0030 ��d

IL-2 0.2717 0.1262 nse

IL-4 0.4889 0.0039 ��IL-5 0.2928 0.0982 nsIL-6 0.3339 0.0575 nsIL-10 0.5173 0.0021 ��IL-12 (p70) 0.5029 0.0029 ��IL-13 0.4675 0.0061 ��IFN-c 0.4967 0.0033 ��TNF-a 0.4908 0.0037 ��

Positive correlation coefficients, r, which are statistically significant are highlightedin bold.

a n = 33 Pairs.b P values were obtained using Spearman’s correlation test.c Gaussian approximation.d ��P values from 0.001 to 0.01 were considered very significant.e ns: Not significant. P values > 0.05.

E.J. Pavón et al. / Cytokine xxx (2013) xxx–xxx 7

although the number of patients analyzed were relatively low (18).To further confirm this, CD38 expression in CD4+, CD8+, and CD25+

T cells was analyzed by flow cytometry in a larger number of SLEpatients. Indeed, the proportion of CD38+ cells in the three T cellsubsets studied was significantly increased in SLE patients as com-

Please cite this article in press as: Pavón EJ et al. Increased CD38 expression in Twith distinct cytokine profiles and disease activity in systemic lupusj.cyto.2013.02.023

pared with that in Normal controls (Fig. 3a–c, respectively). More-over, most SLE patients had an altered CD4:CD8 ratio comparedwith that of healthy controls (17.1% CD4:CD8high + 48.6%CD4:CD8low = 65.7%). We also tested for the expression of the earlyactivation marker CD69 in the lymphocyte cell fraction of PBMCs.56% of the SLE lymphocytes were CD69+ vs only 12% of the Normalcontrols (P = 0.0461, Fisher’s exact test, data not shown). The in-creased CD69 expression may correspond mainly to T cells, sinceabout 70% of the lymphocytes were T cells, and similar results wereobtained in highly purified untouched SLE T cells isolated by neg-ative selection from eight different SLE PBMCs (data not shown).Overall, these data further confirm the altered phenotype of SLET cells, which resembles that of activated/effector T cells.

3.6. Effect of SLE plasmas on CD38 expression and cell number inCD38+ Jurkat T cells

The same SLE patients analyzed for CD38 expression in T cells inFig. 2 were tested for anti-CD38 IgG levels (n = 67), and there wasno correlation between both parameters (r = -0.07404, P = 0.5515,Spearman’s correlation). In fact, in a dot-plot representation thesetwo parameters seemed to be mutually exclusive (Fig. 4a). Thus,those SLE patients positive for CD38 expression were negative foranti-CD38 IgG autoantibodies (lower right quadrant), and thoseSLE patients positive for anti-CD38 IgG were negative for CD38expression in T cells (upper left quadrant). Essentially, there werenot double-positives (upper right quadrant). One possible interpre-tation is that the antibodies against CD38 in SLE plasma can inducedown-modulation of CD38 surface expression in CD38+ T cellsleading to a drastic decrease in the number of circulating CD38+

T cells. Alternatively, these antibodies could be cytotoxic toCD38+ T cells upon binding to the T cell surface by activating thecomplement system or other means. To assess whether anti-CD38 autoantibodies could alter CD38 surface expression in T cells,we used the subclone D8 of Jurkat T cells, which is highly positivefor CD38 on the cell surface (Fig. 4c, upper right panel) [32]. Thesecells were incubated at 37 �C with either plasma from SLE patientswith high levels of anti-CD38 IgG antibodies, or with plasma fromSLE patients negative for anti-CD38 IgG autoantibodies, and com-pared with the effect of incubating with plasmas from Normal con-trols. Upon incubation with plasmas at several volume to volumeratios (1:1; 1:5 and 1:10), cells were washed, marked with ananti-CD38-FITC antibody and analyzed by FACS. Gating those cellswith Normal FSC and SSC (region R2 in Fig. 4c) give you an estimateof the number of alive cells. Plasmas from anti-CD38+ SLE patients(n = 5) did not induce a significant reduction on CD38 surfaceexpression as compared with plasmas from Normal controls(n = 5), or with plasmas from anti-CD38- SLE patients (n = 4) (fora representative experiment see Supplemental Fig. 5). FACS analy-sis was also used to monitor how many cells were dying upon plas-ma incubation. When cells are dying they decrease their size (FSC)and increase their complexity (SSC) (region R1, Fig. 4c), and theycan be clearly distinguished from live cells in R2 [27]. Therefore,cells were analyzed for their FSC and SSC characteristics as com-pared with cells non-incubated with plasma. As shown in a repre-sentative experiment in Fig. 4c, there was a significant increase inthe percentage of dead cells upon incubation of CD38+ Jurkat Tcells with an SLE plasma positive for anti-CD38 autoantibodies(third upper panel, R1 region) as compared with that upon incuba-tion with a Normal plasma negative for anti-CD38 autoantibodies(second upper panel), or with cells kept at 37 �C in medium alone(spontaneous cell-death, first upper panel). The percentage of spe-cific cell death was calculated after subtracting the percentage ofspontaneous cell death yielding 22.5% of dead cells in SLE plas-ma-treated cells vs 5.3% in Normal plasma-treated cells. This effectwas dose- and temperature-dependent (data not shown), and the

cells and circulating anti-CD38 IgG autoantibodies differentially correlateerythematosus patients. Cytokine (2013), http://dx.doi.org/10.1016/

Fig. 3. Altered phenotype of SLE T cells. (a) PBMCs from Normal controls and SLE patients were double-stained with monoclonal Abs against CD38 (X-axis) and CD4 (Y-axis)and analyzed by FACS for CD4+CD38+ cells. Left panels were gated on lymphocytes in FSC vs SSC dot plots of PBMCs, and the percentages are relative to total lymphocytepopulation. Percentages of CD38+CD4+ cells in the dot plots are from a representative experiment of one Normal control (upper panel) and one SLE patient (lower panel).Middle panel histograms were obtained by gating on CD4+ cells in left panels. It represents an overlay of CD38 expression within CD4+ T cells from a SLE patient (green), or aNormal control (red), with a gate set on CD38+ cells. The percentages of CD38+CD4+ cells were in relation to the total CD4+ T cell population. Right panel represents dotdiagrams of CD38 expression in CD4+ T cells from SLE patients (closed circles), and Normal controls (open circles). (b) Dot plots, histograms and dot diagrams of CD38expression in CD8+ T cells from the same groups as in (a). (c) Dot plots, histograms and dot diagram of CD38 expression in CD25+ T cells from the same groups as in (a). Thehorizontal line in each dot diagram denotes the median. The P values were obtained using the Mann–Whitney U test. (For interpretation of the references to color in thisfigure legend, the reader is referred to the web version of this article.)

8 E.J. Pavón et al. / Cytokine xxx (2013) xxx–xxx

Please cite this article in press as: Pavón EJ et al. Increased CD38 expression in T cells and circulating anti-CD38 IgG autoantibodies differentially correlatewith distinct cytokine profiles and disease activity in systemic lupus erythematosus patients. Cytokine (2013), http://dx.doi.org/10.1016/j.cyto.2013.02.023

E.J. Pavón et al. / Cytokine xxx (2013) xxx–xxx 9

results of cumulative experiments shown in Fig. 4b strongly sug-gests the presence of plasma factors in anti-CD38+ SLE patientsthat contribute to increased cell death of CD38+ T cells. We consid-ered that the anti-CD38 autoantibodies present in the SLE plasmasmay bind to CD38 in Jurkat T cells and induce directly or indirectlythe observed events. To test this possibility, we treated the CD38-negative 5A1 Jurkat T cell subclone (Fig. 4c, lower right panel) withan anti-CD38+ SLE plasma for 1 h at 37 �C and subjected to FACSanalysis as above. As shown in Fig. 4c, lower panels, cell-deathdid not increased significantly in CD38-negative Jurkat cells trea-ted with an anti-CD38+ SLE plasma as compared with that obtainedupon incubation with Normal control plasma. The specific celldeath upon subtracting spontaneous cell death was 0.6% and0.2%, respectively. A second experiment with another CD38-dim

Fig. 4. (a) Dot-plot of CD38 expression in SLE CD3+ T cells (X axis) vs anti-CD38 IgG plasmCD38 expression in T cells (upper and lower right quadrants 2), or anti-CD38 IgG autoanNote the absence of double-positive samples in upper right quadrant. (b) Box–and–whiskplasmas from SLE patients (anti-CD38+, n = 5; anti-CD38-, n = 4), or from Normal controlsthe interquartile ranges, and whiskers the 5–95% percentiles. Horizontal lines within thevaried significantly (P = 0.0445, Kruskal–Wallis test). Shown are the P values obtained bforward parameters of CD38+ Jurkat T cell clone D8 (upper panels), or CD38-negative Jurkcontrol plasma (second panels), or SLE plasma (third panels). X axis: forward scatter; Y axiof cells with the scatter parameters of R1 (dead cells) closely correlates with the perceexpression levels of CD38 in CD38+ Jurkat D8 (upper panel) and CD38-negative Jurkat 5events were acquired and analyzed for each figure.

Please cite this article in press as: Pavón EJ et al. Increased CD38 expression in Twith distinct cytokine profiles and disease activity in systemic lupusj.cyto.2013.02.023

5C6 Jurkat T cell subclone yielded 1.7% of specific cell death uponincubation with the same anti-CD38+ SLE plasma vs 2.7% with Nor-mal plasma (data not shown).

3.7. Effect of SLE plasmas on intracellular calcium levels in CD38+

Jurkat T cells

Next we tested whether the plasma from SLE patients inducedany effect on Ca2+ mobilization in Jurkat D8 cells. The patterns ofCa2+ triggering by a plasma from a SLE patient, or a healthy controlis presented in Fig. 5a. The SLE plasma displayed agonistic activity,documented by a consistent increase in [Ca2+]i in Jurkat T cells. TheCa2+ mobilization pattern highlighted by the SLE plasma showedslower kinetics that that elicited by the agonistic anti-CD3 mAb

a levels (Y axis). The dashed lines represent the cut-off values above of which eithertibodies in plasma (upper left and upper right quadrants) were considered positive.ers plot of the number of live cells recovered upon incubation of Jurkat T cells with(n = 5) as described in panel (c) and in Material and Methods. The boxes representboxes represent the median value for each group. The median of the three groups

y the post test Dunn’s Multiple Comparison test. (c) Dot plots show the scatter andat T cell clone 5A1 (lower panels) after incubation with medium (left panels), Normals: side scatter. R1 and R2 represent dead and alive cells, respectively. The percentagentage of apoptotic cells (see Ref. [33]). Histograms on the right show the surfaceA1 (lower panel). Y axis: number of events; X axis: fluorescence intensity. 10,000

cells and circulating anti-CD38 IgG autoantibodies differentially correlateerythematosus patients. Cytokine (2013), http://dx.doi.org/10.1016/

Fig. 5. Mobilization of intracellular calcium in Jurkat T cells upon incubation withplasma from SLE patients. (a) Kinetics of intracellular calcium release (expressed asthe 340/380 Ratio (Y axis) in Jurkat T cells incubated with the indicated reagents forvarious periods of time (X axis, in minutes). (b) Dot diagram of intracellular Ca2+

levels in Jurkat T cells incubated with plasma from SLE patients (closed blackcircles), Normal controls (open circles), SLE patients positive for anti-CD38 IgGautoantibodies (closed blue circles), or negative for anti-CD38 IgG autoantibodies(closed red circles). The dashed line represents the cut-off value above of whichincreases in [Ca2+]i were considered statistically significant (ROC analysis). The Pvalues were obtained using the Fisher’s exact test. (For interpretation of thereferences to color in this figure legend, the reader is referred to the web version ofthis article.)

10 E.J. Pavón et al. / Cytokine xxx (2013) xxx–xxx

OKT3, and resembled the Ca2+ profile induced by an anti-CD38mAb, IB4, in these cells [27]. In contrast, no Ca2+ mobilizationwas observed by incubating Jurkat T cells with plasma from ahealthy control.

[Ca2+]i was then tested in Jurkat T cells incubated with a num-ber of plasmas from Normal controls or SLE patients. As shownin Fig. 5b, most SLE plasmas (13/16) induced increased [Ca2+]i inJurkat T cells as compared with Normal plasmas (81% vs 14%,P = 0.0049, Fisher’s exact test). When SLE patients were segregatedin two groups (anti-CD38 IgG+ or anti-CD38 IgG�), there was astrong induction of Ca2+ mobilization by most of the samples ana-lyzed, independently of the presence or not of anti-CD38autoantibodies.

4. Discussions

SLE patients show increased levels of anti-CD38 IgG autoanti-bodies and higher percentages of CD38+ lymphoid cells (T, B andNK cells) as compared with Normal controls. In contrast, in SLE pa-tients increased levels of anti-CD38 IgM autoantibodies are not de-tected, suggesting that the anti-CD38 IgG autoantibodies are notrelated to low avidity auto- and poly-reactive antibodies (natural

Please cite this article in press as: Pavón EJ et al. Increased CD38 expression in Twith distinct cytokine profiles and disease activity in systemic lupusj.cyto.2013.02.023

antibodies), which usually are of the IgM class, and are producedby CD5+ B cells [33].

The key relevance of autoantibodies in clinical assessment ofSLE is clearly established. Thus, autoantibodies of IgG class aremore frequent in overt and active disease [34,35]. However, someautoantibodies are found in a relatively small proportion of pa-tients, and in clinically quiescent SLE patients. In this sense, thereis an inverse relationship between disease activity and the pres-ence of anti-IFN-a autoantibodies [36,37]. Moreover, a subgroupof SLE defined as having clinically quiescent symptoms despitepersistent serological activity (increased anti-dsDNA and/or lowserum complement levels) have a significantly higher frequencyof anti-IFN-a autoantibodies than SLE patients without SACQ[38]. IgM autoantibodies to phosphorylcoline (PC) were signifi-cantly higher in SLE patients with low disease activity and less or-gan damage [39]. Moreover, the expression patterns of IgG anti-PCwere also different from all other types of IgG autoantibodiestested (MDA, b2-GPI, Cardiolipin, dsDNA), as anti-PC was the onlyIgG autoantibody that neither showed a direct correlation with dis-ease activity measurements nor with standard clinical inflamma-tory markers or disease manifestations [39]. In our study,clinically quiescent SLE patients show increased levels of anti-CD38 IgG autoantibodies as compared with healthy controls withan OR of 8.2, whereas clinically active SLE patients do not show asignificant increase as compared with Normal controls. Rather,when active and quiescent SLE patients are compared a negativeassociation seems to occur regarding the frequency of anti-CD38IgG autoantibodies (OR = 0.43), although it does not reach statisti-cal significance. Focusing in active SLE patients with positive anti-dsDNA autoantibodies, it is clearly shown the inverse relationshipof anti-dsDNA and anti-CD38 IgG autoantibody frequencies(OR = 0.21, P = 0.0155).

CD38 expression in SLE B cells correlates exclusively with IL-10plasma levels, which may also be linked to the maintenance of highlevels of anti-CD38 autoantibodies. Thus, IL-10 is produced at ahigh level by B lymphocytes and monocytes of SLE patients, andit contributes to the abnormal production of immunoglobulinsand of autoantibodies [34,35]. In this sense, in this study increasedIL-10 and anti-dsDNA autoantibody plasma levels are the best pre-dictors of increased clinical activity. On the other hand, regulatoryB cells that preferentially produce IL-10 have a role in controllingautoimmune responses. Thus, it has recently been shown that inhuman healthy controls a CD19+CD24highCD38high B cell subsetinhibits proinflammatory IFN-c and TNF-a production by CD4+ Tcells, which is mediated mostly by IL-10 [40]. This subset is notfunctional in SLE patients, at least upon CD40 engagement [40].Therefore, moderate increased levels of IL-10 may be beneficialfor the individual because stimulates regulatory events, whereasprolonged and higher levels may stimulate autoantibody produc-tion. In this sense, SLE patients positive for anti-CD38 IgG autoan-tibodies correlate with increased plasma levels of IL-10, which is inapparent contradiction with the positive association of IL-10 plas-ma levels and clinical activity. Looking closely to the results it isclear that the anti-CD38� SLE patients represent an heterogeneouspopulation with a significant subgroup of patients showing higherIL-10 plasma levels than any of the anti-CD38+ SLE patients, andanother subgroup with Normal IL-10 levels (data not shown). Insummary, anti-CD38+ SLE patients show moderate, although sig-nificant increases of IL-10 plasma levels suggesting that these pa-tients have a relatively well controlled disease, which iscorroborated by their relatively low SLEDAI score, and low fre-quency of anti-dsDNA autoantibodies.

In many cell types CD38 expression is regulated by a variety ofcytokines including IL-1b, TNF-a, IL-13, IFN-c, and IFN-a [41–44].Thus, CD38 appears to be a relevant marker for activation of den-dritic cells by TLR3 or IFN-a [45]. Moreover, IFN-a/b, produced

cells and circulating anti-CD38 IgG autoantibodies differentially correlateerythematosus patients. Cytokine (2013), http://dx.doi.org/10.1016/

E.J. Pavón et al. / Cytokine xxx (2013) xxx–xxx 11

by HIV-activated plasmacytoid dendritic cells, is necessary and suf-ficient for upregulation of CD38 (and CD69) in CD4+ and CD8+ Tcells [46]. Interestingly, TNF-a and IFNs (b or c) synergistically reg-ulate CD38 expression in human airway smooth muscle cells, mak-ing these cells refractory to the anti-inflammatory action ofsteroids [47]. In our study, increased CD38 expression in T cellsstrongly correlates with increased plasma levels of several cyto-kines, including TNF-a, IFN-c and IL-13, which may contribute toan autocrine or paracrine induction and stabilization of CD38expression. In SLE patients neutralizing IFN-a activity with anti-IFN-a autoantibodies may mediate a reduction in CD38 expression,at least in B cells.

CD38 overexpression in SLE T cells may also contribute to per-petuate the altered cytokine profile observed in SLE plasmas. Thus,up-regulation of CD38 expression in Normal T cells enhances TCR-mediated calcium mobilization, and IL-2 production in an antigen-dependent manner [26], while opposite results are obtained bydown-regulating surface CD38 expression by means of CD38 siRNA[26]. Increased expression of CD38 in SLE T cells, and its alteredpresence in membrane lipid rafts [18], may contribute, in additionwith other signaling molecules that show altered T cell-surface dis-tribution, to the observed T cell signaling abnormalities in SLE pa-tients [48].

Increased CD38 expression and increased anti-CD38 IgG levelsseem to be mutually exclusive. In anti-CD38+ SLE patients CD38could be down-modulated from the cell surface of T cells by highlevels of circulating anti-CD38 IgG antibodies. This hypothesis isnot supported by our observation that plasmas from anti-CD38+

SLE patients cannot effectively reduce the CD38 surface expressionin Jurkat T cells. However, the increased cell death induced by plas-mas from anti-CD38+ SLE patients in CD38+ but not in CD38-nega-tive Jurkat T cells suggests either the anti-CD38 may induceapoptosis directly as it occurs in Jurkat T cells incubated with ago-nistic anti-CD38 mAbs [27], or it is required the intervention ofadditional plasma factors not present or nor activated in anti-CD38- SLE and Normal plasmas. In contrast, it is observed that bothanti-CD38+ and anti-CD38- SLE plasmas induce increased [Ca2+]i

mobilization in CD38+ Jurkat T cells. Therefore, it is unlikely thatthe anti-CD38 IgG autoantibodies can account for the observed in-crease in [Ca2+]i. Additional functional experiments with purifiedanti-CD38 autoantibodies from SLE patients will be required to fur-ther prove that anti-CD38 autoantibodies play a direct role in theseor another biological processes.

5. Conclusions

Prospective studies are needed to evaluate whether changes inCD38 expression in lymphoid cells and anti-CD38 levels in plasmaor serum can be useful to predict disease flares and/or organ dam-age. Increased CD38 expression in SLE T cells could be the conse-quence of the action of proinflammatory cytokines such as TNF-aand IFN-c, and is indicative of SLE patients with a more active dis-ease, and with an overt abnormal Th2 and Th1 cytokine profile (seeSupplemental Fig. 6 for a tentative model). In contrast, the pres-ence of anti-CD38 autoantibodies could be indicative of SLE pa-tients with a relatively well-controlled disease, which iscorroborated by their relatively low or negative SLEDAI score,and low frequency of anti-dsDNA autoantibodies. However, the in-creased cytokine profile was still significantly higher than that ofNormal controls, which is consistent with the hypothesis of persis-tent B cell activation in SLE, even in clinically quiescent patients[14]. Anti-CD38 autoantibodies or their associated plasma factorsmay dampen immune activation by affecting the viability ofCD38+ effector T cells and be correlated with a favorablephenotype.

Please cite this article in press as: Pavón EJ et al. Increased CD38 expression in Twith distinct cytokine profiles and disease activity in systemic lupusj.cyto.2013.02.023

Disclosure

None of the authors has any potential financial conflict of inter-est related to this manuscript.

Funding

This work was supported in part by the European Commissionin collaboration with the following Funding Agencies:

Ministerio de Ciencia e Innovación (MICINN) del Gobierno deEspaña [Grant Nos. SAF2005-06056-C02-01, SAF2008-03685, andSAF2011-27261] (to J.S.); Consejería de Innovación, Ciencia yEmpresa de la Junta de Andalucía [Grant Nos. P05-CVI-00908,and P08-CTS-04046] (to J.S.); Ministerio de Sanidad y Consumo-IS-CIII-FIS [Grant Nos. FIS03/0389, UIPY-1465/07] (to MZ); MICINN-ISCIII-FIS [Grant No. FIS06/1502] (to M.Z.); and MICINN-CSIC[Grant No. PI-200820I216] (to M.Z.).

F.M. The work done here is supported by PRIN and FIRB Projects(Ministry of University and Research, Italy) and by the FondazioneInternazionale Ricerca in Medicina Sperimentale (FIRMS).

E.J.P. and D.C.W. were supported by grant-contracts fromMICINN.

E.Z. and A.R.V. were supported by fellowship-contracts fromConsejería de Innovación, Ciencia y Empresa de la Junta deAndalucía.

S.G.R. was supported by a JAEDoc contract from CSIC.

Note added in proof

While this work was in revision García-Rodríguez et al. (Medi-ators Inflamm. 2012;2012:495934. http://dx.doi.org/10.1155/2012/495934) published findings in which they showed, in a com-pletely different set of SLE patients, that increased IL-10 geneexpression in SLE PBMCs positively correlated with both CD38and MAPK1 gene expression in these cells.

Acknowledgements

We thank Pilar Navarro-Cuesta, and Salvador J. Guerrero-Ferná-ndez from the IPBLN, and Blanca Martínez-López from the HospitalClínico for their technical assistance. The proteomic protein identi-fication by LC-MS/MS analysis was carried out in the LP-CSIC/UAB,a member of ProteoRed network.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.cyto.2013.02.023.

References

[1] Malavasi F, Deaglio S, Funaro A, Ferrero E, Horenstein AL, Ortolan E, et al.Evolution and function of the ADP ribosyl cyclase/CD38 gene family inphysiology and pathology. Physiol Rev 2008;88:841–86.

[2] Khoo KM, Chang CF. Characterization and localization of CD38 in thevertebrate eye. Brain Res. 1999;821:17–25.

[3] Khoo KM, Chang CF. Localization of plasma membrane CD38 is domain specificin rat hepatocyte. Arch Biochem Biophys. 2000;373:35–43.

[4] Graeff R, Munshi C, Aarhus R, Johns M, Lee HC. A single residue at the activesite of CD38 determines its NAD cyclizing and hydrolyzing activities. J BiolChem 2001;276:12169–73.

[5] Lee HC. Physiological functions of cyclic ADP-ribose and NAADP as calciummessengers. Annu Rev Pharmacol Toxicol 2001;41:317–45.

[6] Guse AH. Second messenger function and the structure-activity relationship ofcyclic adenosine diphosphoribose (cADPR). Febs J. 2005;272:4590–7.

[7] Dammermann W, Zhang B, Nebel M, Cordiglieri C, Odoardi F, Kirchberger T,et al. NAADP-mediated Ca2+ signaling via type 1 ryanodine receptor in T cellsrevealed by a synthetic NAADP antagonist. PNAS 2009;106:10678–83.

cells and circulating anti-CD38 IgG autoantibodies differentially correlateerythematosus patients. Cytokine (2013), http://dx.doi.org/10.1016/

12 E.J. Pavón et al. / Cytokine xxx (2013) xxx–xxx

[8] Giorgi JV, Lyles RH, Matud JL, Yamashita TE, Mellors JW, Hultin LE, et al.Predictive value of immunologic and virologic markers after long or shortduration of HIV-1 infection. J Acquir Immune Defic Syndr 2002;29:346–55.

[9] Keyhani A, Huh YO, Jendiroba D, Pagliaro L, Cortez J, Pierce S, et al. IncreasedCD38 expression is associated with favorable prognosis in adult acuteleukemia. Leuk Res 2000;24:153–9.

[10] Deaglio S, Vaisitti T, Aydin S, Ferrero E, Malavasi F. In-tandem insight frombasic science combined with clinical research: CD38 as both marker and keycomponent of the pathogenetic network underlying chronic lymphocyticleukemia. Blood 2006;108:1135–44.

[11] Deaglio S, Aydin S, Vaisitti T, Bergui L, Malavasi F. CD38 at the junctionbetween prognostic marker and therapeutic target. Trends Mol Med2008;14:210–8.

[12] Grammer AC, Slota R, Fischer R, Gur H, Girschick H, Yarboro C, et al. Abnormalgerminal center reactions in systemic lupus erythematosus demonstrated byblockade of CD154-CD40 interactions. J Clin Invest 2003;112:1506–20.

[13] Lee J, Kuchen S, Fischer R, Chang S, Lipsky PE. Identification andcharacterization of a human CD5+ pre-naive B cell population. J Immunol2009;182:4116–26.

[14] Spronk PE, Horst G, Van Der Gun BT, Limburg PC, Kallenberg CG. Anti-dsDNAproduction coincides with concurrent B and T cell activation duringdevelopment of active disease in systemic lupus erythematosus (SLE). ClinExp Immunol 1996;104:446–53.

[15] Malavasi F, Deaglio S, Zaccarello G, Horenstein AL, Chillemi A, Audrito V, et al.The hidden life of NAD+-consuming ectoenzymes in the endocrine system. JMol Endocrinol 2010;45:183–91.

[16] Alcocer-Varela J, Alarcon-Riquelme M, Laffon A, Sanchez-Madrid F, Alarcon-Segovia D. Activation markers on peripheral blood T cells from patients withactive or inactive systemic lupus erythematosus. Correlation with proliferativeresponses and production of IL-2. J Autoimmun 1991;4:935–45.

[17] al-Janadi M, Raziuddin S. B cell hyperactivity is a function of T cell derivedcytokines in systemic lupus erythematosus. J Rheumatol 1993;20:1885–91.

[18] Pavon EJ, Munoz P, Navarro MD, Raya-Alvarez E, Callejas-Rubio JL, Navarro-Pelayo F, et al. Increased association of CD38 with lipid rafts in T cells frompatients with systemic lupus erythematosus and in activated normal T cells.Mol Immunol 2006;43:1029–39.

[19] Ikehata F, Satoh J, Nata K, Tohgo A, Nakazawa T, Kato I, et al. Autoantibodiesagainst CD38 (ADP-ribosyl cyclase/cyclic ADP-ribose hydrolase) that impairglucose-induced insulin secretion in noninsulin-dependent diabetes patients. JClin Invest 1998;102:395–401.

[20] Mallone R, Ortolan E, Baj G, Funaro A, Giunti S, Lillaz E, et al. Autoantibodyresponse to CD38 in Caucasian patients with type 1 and type 2 diabetes:immunological and genetic characterization. Diabetes 2001;50:752–62.

[21] Antonelli A, Fallahi P, Nesti C, Pupilli C, Marchetti P, Takasawa S, et al. Anti-CD38 autoimmunity in patients with chronic autoimmune thyroiditis orGraves’ disease. Clin Exp Immunol 2001;126:426–31.

[22] Bombardier C, Gladman DD, Urowitz MB, Caron D, Chang CH. Derivation of theSLEDAI. A disease activity index for lupus patients. The committee onprognosis studies in SLE. Arthritis Rheum 1992;35:630–40.

[23] Khoo KM, Chang CF, Schubert J, Wondrak E, Chng HH. Expression andpurification of the recombinant His-tagged GST-CD38 fusion protein using thebaculovirus/insect cell expression system. Protein Expr Purif2005;40:396–403.

[24] Fryxell KB, O’Donoghue K, Graeff RM, Lee HC, Branton WD. Functionalexpression of soluble forms of human CD38 in Escherichia coli and Pichiapastoris. Protein Expr Purif 1995;6:329–36.

[25] Carrascal M, Ovelleiro D, Casas V, Gay M, Abian J. Phosphorylation analysis ofprimary human T lymphocytes using sequential IMAC and titanium oxideenrichment. J Proteome Res 2008;7:5167–76.

[26] Munoz P, Mittelbrunn M, de la Fuente H, Perez-Martinez M, Garcia-Perez A,Ariza-Veguillas A, et al. Antigen-induced clustering of surface CD38 andrecruitment of intracellular CD38 to the immunologic synapse. Blood2008;111:3653–64.

[27] Morra M, Zubiaur M, Terhorst C, Sancho J, Malavasi F. CD38 is functionallydependent on the TCR/CD3 complex in human T cells. FASEB J1998;12:581–92.

Please cite this article in press as: Pavón EJ et al. Increased CD38 expression in Twith distinct cytokine profiles and disease activity in systemic lupusj.cyto.2013.02.023

[28] Ausiello CM, Urbani F, la Sala A, Funaro A, Malavasi F. CD38 ligation inducesdiscrete cytokine mRNA expression in human cultured lymphocytes. Eur JImmunol 1995;25:1477–80.

[29] Illei GG, Tackey E, Lapteva L, Lipsky PE. Biomarkers in systemic lupuserythematosus: II. Markers of disease activity. Arthritis Rheum2004;50:2048–65.

[30] Letourneau S, Krieg C, Pantaleo G, Boyman O. IL-2- and CD25-dependentimmunoregulatory mechanisms in the homeostasis of T-cell subsets. J AllergyClin Immunol 2009;123:758–62.

[31] van Vlasselaer P, Gascan H. De Waal Malefyt R, de Vries JE. IL-2 and a contact-mediated signal provided by TCR alpha beta+ or TCR gamma delta+ CD4+ T cellsinduce polyclonal Ig production by committed human B cells. Enhancement byIL-5, specific inhibition of IgA synthesis by IL-4. J Immunol 1992;148:1674–84.

[32] Zubiaur M, Izquierdo M, Terhorst C, Malavasi F, Sancho J. CD38 ligation resultsin activation of the Raf-1/mitogen-activated protein kinase and the CD3-zeta/zeta-associated protein-70 signaling pathways in Jurkat T lymphocytes. JImmunol 1997;159:193–205.

[33] Lydyard PM, Lamour A, MacKenzie LE, Jamin C, Mageed RA, Youinou P. CD5+ Bcells and the immune system. Immunol Lett 1993;38:159–66.

[34] Rekvig OP, Putterman C, Casu C, Gao HX, Ghirardello A, Mortensen ES, et al.Autoantibodies in lupus: culprits or passive bystanders. Autoimmun Rev 2011.

[35] Meroni PL, Shoenfeld Y. Systemic lupus erythematosus and the SLE galaxy.Autoimmun Rev 2010;10:1–2.

[36] Morimoto AM, Flesher DT, Yang J, Wolslegel K, Wang X, Brady A, et al.Association of endogenous anti-interferon-alpha autoantibodies withdecreased interferon-pathway and disease activity in patients with systemiclupus erythematosus. Arthritis Rheum 2011;63:2407–15.

[37] von Wussow P, Jakschies D, Hartung K, Deicher H. Presence of interferon andanti-interferon in patients with systemic lupus erythematosus. Rheumatol Int1988;8:225–30.

[38] Ching KH, Burbelo PD, Tipton C, Wei C, Petri M, Sanz I, et al. Two majorautoantibody clusters in systemic lupus erythematosus. PLoS One2012;7:e32001.

[39] Gronwall C, Akhter E, Oh C, Burlingame RW, Petri M, Silverman GJ. IgMautoantibodies to distinct apoptosis-associated antigens correlate withprotection from cardiovascular events and renal disease in patients with SLE.Clin Immunol 2012;142:390–8.

[40] Blair PA, Norena LY, Flores-Borja F, Rawlings DJ, Isenberg DA, Ehrenstein MR,et al. CD19(+)CD24(hi)CD38(hi) B cells exhibit regulatory capacity in healthyindividuals but are functionally impaired in systemic Lupus Erythematosuspatients. Immunity 2010;32:129–40.

[41] Tirumurugaan KG, Kang BN, Panettieri RA, Foster DN, Walseth TF, Kannan MS.Regulation of the cd38 promoter in human airway smooth muscle cells byTNF-alpha and dexamethasone. Respir Res 2008;9:26.

[42] Musso T, Deaglio S, Franco L, Calosso L, Badolato R, Garbarino G, et al. CD38expression and functional activities are up-regulated by IFN-gamma on humanmonocytes and monocytic cell lines. J Leukoc Biol 2001;69:605–12.

[43] Deshpande DA, Dogan S, Walseth TF, Miller SM, Amrani Y, Panettieri RA, et al.Modulation of calcium signaling by interleukin-13 in human airway smoothmuscle: role of CD38/cyclic adenosine diphosphate ribose pathway. Am JRespir Cell Mol Biol 2004;31:36–42.

[44] Bauvois B, Durant L, Laboureau J, Barthelemy E, Rouillard D, Boulla G, et al.Upregulation of CD38 gene expression in leukemic B cells by interferon types Iand II. J Interferon Cytokine Res 1999;19:1059–66.

[45] Trepiakas R, Pedersen AE, Met O, Svane IM. Addition of interferon-alpha to astandard maturation cocktail induces CD38 up-regulation and increasesdendritic cell function. Vaccine 2009;27:2213–9.

[46] Boasso A, Hardy AW, Anderson SA, Dolan MJ, Shearer GM. HIV-induced type Iinterferon and tryptophan catabolism drive T cell dysfunction despitephenotypic activation. PLoS One 2008;3:e2961.

[47] Tliba O, Cidlowski JA, Amrani Y. CD38 expression is insensitive to steroidaction in cells treated with tumor necrosis factor-alpha and interferon-gammaby a mechanism involving the up-regulation of the glucocorticoid receptorbeta isoform. Mol Pharmacol 2006;69:588–96.

[48] Tenbrock K, Juang YT, Kyttaris VC, Tsokos GC. Altered signal transduction inSLE T cells. Rheumatology (Oxford, England) 2007;46:1525–30.

cells and circulating anti-CD38 IgG autoantibodies differentially correlateerythematosus patients. Cytokine (2013), http://dx.doi.org/10.1016/