Defined Blocks in Terminal Differentiation of Common Variable Immunodeficiency Patients

This information is current as Nadine Taubenheim, Marcus von Hornung, Anne Durandy, of September 24, 2021. Klaus Warnatz, Lynn Corcoran, Hans-Hartmut Peter and Hermann Eibel J Immunol 2005; 175:5498-5503; ; doi: 10.4049/jimmunol.175.8.5498

http://www.jimmunol.org/content/175/8/5498 Downloaded from

References This article cites 35 articles, 13 of which you can access for free at: http://www.jimmunol.org/content/175/8/5498.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 24, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2005 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Defined Blocks in Terminal Plasma Cell Differentiation of Common Variable Immunodeficiency Patients1

Nadine Taubenheim,* Marcus von Hornung,* Anne Durandy,† Klaus Warnatz,‡ Lynn Corcoran,§ Hans-Hartmut Peter,‡ and Hermann Eibel2*

Common variable immunodeficiency (CVID) is a heterogeneous disorder characterized by defective Ab production and recurrent bacterial infections. The largely unknown causes are likely to comprise a diverse set of genetic or acquired defects. In this study, we investigated terminal differentiation in lymph nodes from CVID patients. Up to the B cell stage, B cell differentiation was normal but terminal plasma cell development was found to be impaired. Using differential Blimp-1 and Syndecan-1 expression in controls, we defined three different plasma cell subsets that correspond to progressive developmental stages locating to different sites in the lymph node. In the CVID patients, we could only detect one or two of these subsets indicating a defective differentiation. Thus, terminal plasma cell differentiation was found to be impaired despite normal expression of Downloaded from Blimp-1. B cells reaching only the first stage of plasma cell differentiation were further unable to undergo isotype switching and to up-regulate activation markers on B cells stimulated in vitro. The Journal of Immunology, 2005, 175: 5498–5503.

ommon variable immunodeficiency (CVID)3 is the sec- hypermutation (SHM) (9), as well as defective formation of mem- ond most frequent primary immunodeficiency in humans ory B cells (10–13). C with a prevalence of ϳ1 in 30,000 (1). The major hall- Class switch recombination (CSR) as well as SHM of the V http://www.jimmunol.org/ mark of CVID is a significant reduction or absence of Abs in the regions take place in germinal centers (GC) of secondary lymphoid serum of the patients leading to recurrent bacterial infections pre- organs and lead to the generation of Ab-secreting plasma cells and dominantly of the respiratory and gastrointestinal tracts. Further- Ag-specific memory cells. The GC reaction is initiated by the ac- more, an increased incidence of granulomatous inflammation, au- tivation of B cells, which up-regulate BCL-6 and migrate into B toimmune disorders, and gastrointestinal malignancies has been cell follicles where they undergo a phase of strong proliferation. At described (2). The syndrome can occur in both sporadic and fa- the same time, the specificity of the BCR is modified by SHM, milial forms and covers a heterogeneous group of disorders of allowing for affinity maturation due to positive selection of B cells which the underlying molecular bases are largely unknown. Re- for BCRs with the highest affinity. cently, a homozygous deletion in the ICOS gene has been identi- BCL-6 is a zinc finger protein that acts as a transcriptional re- by guest on September 24, 2021 fied in a small group of CVID patients (3). However, in the ma- pressor. BCL-6 mRNA can be found in a variety of tissues (14), jority of the patients, the characterization of genetic defects but on the protein level, the expression is mainly restricted to lym- remains elusive. Several functional defects have been described phocytes (14–16), with the highest expression in GC affecting proliferation, activation, and cytokine production in T and centrocytes. PRDM1, encoding the transcriptional repressor cells (4, 5). In B cells, early and late differentiation stages may be Blimp-1, represents a central target gene of BCL-6 (17). Blimp-1 affected resulting in defective up-regulation of surface molecules was proposed to be a key regulator of terminal plasma cell differ- such as CD86 and CD70 (6, 7), impaired signaling (8), and somatic entiation because ectopic expression of the protein was sufficient to drive the differentiation of mature B cells to plasma cells (18). The defective expression of proteins essential for the differen- *Clinical Research Unit for Rheumatology, University Hospital of Freiburg, Freiburg, tiation of B cells to GC B cells or plasma cells, such as BCL-6 or Germany; †Institut National de la Sante´et de la Recherche Me´dicale (INSERM) Unite´ Blimp-1, could be responsible for the phenotypic features observed 429, Hoˆpital Necker-Enfants Malades, Paris, France; ‡Division of Rheumatology and Clinical Immunology, Department of Medicine, University Hospital of Freiburg, in CVID. In this report, we describe for the first time defects in the Freiburg, Germany; §The Walter and Eliza Hall Institute of Medical Research, GC reaction and plasma cell development in lymph nodes (LN) Parkville, Victoria, Australia from CVID patients. We revealed three different stages by which Received for publication April 28, 2005. Accepted for publication July 22, 2005. plasma cell differentiation takes place in the control LN and could The costs of publication of this article were defrayed in part by the payment of page show distinct blocks in this process in the LN from three CVID charges. This article must therefore be hereby marked advertisement in accordance patients. with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by the European Community through Grants QLCT-2001- 01536 and LSHM-CT-2004-005264 to H.E. and the Grants EURO-POLICY-PID PL Materials and Methods 006411, INSERM, the French Rare Disease Program and the Assistance-Publique, Patients Hopitaux de Paris (to A.D.). 2 In this study patients were only included with established diagnosis of Address correspondence and reprint requests to Dr. Hermann Eibel, Clinical Re- ͗ ͘ search Unit for Rheumatology, University Hospital Freiburg, Zentrum Klinische CVID (ESID criteria, www.esid.org ). The patients were regularly fol- Forschung, Breisacher Strasse 66, 79106 Freiburg, Germany. E-mail address: lowed in our outpatient clinic and received monthly i.v. or weekly s.c. [email protected] replacement therapy, respectively. LN biopsies were from patients who developed during their follow-up regional mediastinal and/or LN swellings 3 Abbreviations used in this paper: CVID, common variable immunodeficiency; CSR, class switch recombination; CT, circle transcript; AICDA, activation-induced cytidine to exclude malignant . None of the biopsies of the three CVID deaminase; EGFP, enhanced GFP; GC, germinal center; GLT, germline transcript; patients described here showed malignant . Control tissues HD, healthy donor; LN, lymph node; MFI, mean fluorescence intensity; SHM, so- were obtained from patients undergoing LN dissection within the scope of matic hypermutation. lung cancer treatment.

Copyright © 2005 by The American Association of Immunologists, Inc. 0022-1767/05/$02.00 The Journal of Immunology 5499

Table I summarizes the clinical and immunological features of the pa- endogenous peroxidases. Immunohistochemical stainings were then per- tients. Informed consent was obtained from all patients and controls using formed according to the procedures specified in Elite Vectastain kits, using biopsy material and peripheral blood for scientific purposes according to a the Nova Red, Vector SG, and Vector AP substrates (Vector Laboratories). local ethics committee approved research protocol (no. 239/1999). Reagents Flow cytometry Primary Abs used for immunohistochemistry: anti-IgD-AP (Southern Bio- PBMC and single cell suspensions of the LN were stained with fluores- technology Associates), anti-Syndecan-1 (Serotec), anti-Ki67 (BD Pharm- cence-conjugated Abs. Cells were acquired using a FACSCalibur analyzer, ingen), anti-BCL 6 and anti-CD23 (DakoCytomation), anti-BOB-1 and anti- and data were analyzed using CellQuest software (BD Biosciences). Dead OCT-2 (Santa Cruz Biotechnology) and anti-Blimp-1 (19). Abs and cells were excluded by gating on propidium iodide-negative cells and on reagents used for flow cytometry: anti-CD19 allophycocyanin and anti- viable lymphocytes according to their forward and side scatters. CD38 FITC (Caltag Laboratories), anti-IgM-bio (Sigma-Aldrich), anti- CD19 PC5 (IO-Test), SA-allophycocyanin, anti-CD27 PE, anti-CD69 FITC and anti-CD86 PE (BD Pharmingen). Results Defective B cell differentiation in GC from CVID patients Cell preparation and stimulations In immunohistochemical stainings of LN from CVID patients, we PBMC were isolated from EDTA-blood using Ficoll-Hypaque density gra- dient centrifugation (PAN Biotech) and cultivated in IMDM (Invitrogen analyzed the expression of different GC and plasma cell markers to Life Technologies) supplemented with 10% FCS (Linaris). Cells were detect possible defects during late B cell differentiation. stimulated with CD40L using irradiated NIH 3T3 fibroblasts expressing Large granulomas were detected in the lymph node of one pa- ␣ human CD40L under the control of the human EF1 promoter. For BCR tient whereas in another patient a lymphoma had developed. In the ϫ 5

stimulation, 6 10 PBMC were cultivated in 48-well plates in the pres- Downloaded from ␮ lymph nodes of three patients, we found the GCs to be highly ence of 1.5 g/ml anti-IgM-F(ab)2 for 24 h. For CSR, 500 U/ml IL-4 (R&D Systems) was added and supernatants were analyzed after 13 days of cul- enlarged compared with those found in control tissues. Analysis of tivation by ELISA as described elsewhere (20). Expression of activation- Ki67 illustrated that these GC consisted of a high number of pro- induced cytidine deaminase (AICDA), IgE germline transcript (GLT), and liferating cells, correlating well with high expression levels of the circle transcript (CT) was examined as described in the report of Imai et al. transcriptional repressor BCL-6. Therefore, B cell differentiation (21). in the LN of these CVID patients seems not to be disturbed until

Detection of SHM in the IgH V regions the centrocyte stage. At later stages of B cell differentiation, http://www.jimmunol.org/ Frequency and characteristics of SHMs in the variable region of the IgM H Blimp-1 and Syndecan-1 are expressed in plasma cells and/or in a ϩ ϩ chain (VH region) was studied in purified CD19 CD27 B cells as de- subset of GC B cells, which are committed to plasma cell differ- scribed previously (9). PBMCs were labeled with anti-CD19 mAb and entiation (23). The expression of Blimp-1 and Syndecan-1 in con- anti-CD27 mAb (Immunotech) and then purified by FACS sorting. The purity of sorted CD19ϩCD27ϩ B cells was Ͼ99%. SHMs were identified trol LN revealed three different subsets of plasma cells that could be further distinguished by their localization in the LN (Fig. 1): in the VH3-23 (GenBank accession no. AB019439) region after PCR am- ␮ ϩ Ϫ plification using VH3-23 and C primers and cloning as previously de- Blimp-1 Syndecan-1 cells represented a relatively small number scribed (9). of late GC B cells and were found only in the light zone, whereas ϩ ϩ Engraftment of immunodeficient mice with purified T and B cells Blimp-1 Syndecan-1 were found both in and around the GC (arrows in Fig. 1). Finally, Syndecan-1ϩ cells with low expression by guest on September 24, 2021 Human T and B cells from PBMC were isolated with CD4-positive selec- low high tion kit and anti-CD19 PanB magnetic beads, respectively (Dynal). of Blimp-1 (Blimp-1 Syndecan-1 ) were situated in large Purified cells were transferred into Rag-2/common-␥-chain-deficient numbers only outside of GC. Because plasma cell differentiation mice (originally described by Colucci et al. (22)). Within 24 h of their requires exit from the cell cycle, all of these cells expressed only 7 isolation, 1–3 ϫ 10 cells in a 50-␮l suspension were directly injected into very low levels of Ki67. the spleens of the mice (6:1 T/B cell ratio). Blood samples were taken from The LN from all patients contained cells expressing Blimp-1 the tail vein and human IgM and IgG concentrations were measured in the serum by ELISA. Commercial coating Abs, alkaline phosphatase-conju- and low levels of Ki67. Expression of Syndecan-1 could be de- gated detection Abs and standards for IgM and IgG (Jackson ImmunoRe- tected only in the LN from patient P3. All Blimp-1-positive cells search Laboratories) were used. The assays were developed with AP sub- in the tissues from the patients P1 and P2 therefore were Blimp- strate p-nitrophenyl-phosphate (Sigma-Aldrich). After 3 wk, mice were 1ϩSyndecan-1Ϫ, whereas the LN from patient P3 additionally con- sacrificed and spleens were taken for immunohistochemical analyses. ϩ ϩ ϩ tained Blimp-1 Syndecan-1 cells. Thus, Syndecan-1 cells ex- Immunohistochemistry pressing low Blimp-1 levels (Blimp-1lowSyndecan-1high), which Methanol/acetone (1:1) fixed 10-␮m cryosections of LN were treated with were present in the control sections, could not be found in the LN 0.1% phenylhydrazine in PBS for 30 min at room temperature to inactivate from any of these three CVID patients (Fig. 1A).

Table I. Characteristics of CVID patients

Classification

Bryant Warnatz Serum IgM Serum IgG SHM In Vivo et al. et al. (g/L) (g/L) % CD19 % Mutations/bp Patient Age/Sex (Ref. 35) (Ref. 10) (0.4–2.3) (7–16)a (6–19%) (2.6–6.3) Clinic Infections

P1 38/F B Ia Ͻ0.2 3.2 4.2 NAb Splenomegaly, enlarged Chronic bronchitis, mediastinal LN chronic sinusitis P2 45/F B Ia 0.4 2.1 7.0 0.4 Splenomegaly, enlarged Chronic sinusitis mediastinal LN, sarcoid-like lesions P3 29/F B Ia 0.3 4.0 13.7 2.9 Splenomegaly, enlarged Chronic bronchitis, mediastinal and chronic sinusitis abdominal LN, bronchiectasia

aPrior to monthly i.v. IgG substitution. b NA, not available. 5500 PLASMA CELL DIFFERENTIATION IN CVID PATIENTS

FIGURE 1. Immunohistochemical and flow cyto- metric analysis of LN and peripheral blood cells. A, Immunohistochemical analysis of LN sections from patients P1, P2, and P3 and controls (C). Arrowheads point toward different plasma cell subsets. Magnifica- tions are ϫ100 at Ki67 stainings except for C (ϫ200), ϫ200 at BCL-6 (P3), ϫ100 at C, P1 and P2, ϫ100 at Syndecan-1 except for P3 (ϫ1000) and ϫ1000 at all Blimp-1 stainings. For the controls, four independent samples were analyzed. B, Numbers of Blimp-1 and/or Syndecan-1 expressing cells determined in multiple (n ϭ 4) distinct areas of equal size in LN from controls and the CVID patients. C, Percentages of CD19ϩCD27ϩ B cells in LN from CVID patients P1 (᭜), P2 (F), and P3 (Œ) and from controls in a flow cytometric analysis. D, Dot plot of the cells from the Downloaded from CVID LN and one representative control. E, Percent- age of IgMϩCD27ϩ and IgMϪCD27ϩ cells of B cells in the peripheral blood of patients P1, P2, and P3 and of controls. F, IgM and CD27 expression on CD19ϩ B cells from the peripheral blood of the CVID patients P1, P2, and P3 and one representative of four controls. http://www.jimmunol.org/

Quantitative analysis of cell numbers within each Blimp-1ϩ not of CD69 (7). To examine defects in the response to Ag en- subset was performed by counting cells expressing Blimp-1 and counter and to correlate those data with the immunohistochemical

Syndecan-1 in multiple areas of equal size inside and around GCs. results, we stimulated PBMC from CVID patients P2 and P3 for by guest on September 24, 2021 The mean number of Blimp-1ϩSyndecan-1Ϫ cells was increased in 24 h with CD40L ϩ anti-IgM and analyzed the surface expression the GC of CVID patients (P1, 15.5 Ϯ 3.6; P2, 9.0 Ϯ 0.9; P3, of CD86 and CD69. 10.3 Ϯ 4.2; control, 5.7 Ϯ 0.6; Fig. 1B), whereas the number of The B cells from patient P2 exhibited a lower CD86 expression Blimp-1ϩSyndecan-1ϩ cells was lower in LN from P3 and absent after stimulation (Fig. 2A, mean fluorescence intensity (MFI) 94.1) in P1 and P2 (P3, 6.3 Ϯ 1.2; control, 15.2 Ϯ 5.1). Blimp-1lowSyn- whereas on B cells from patient P3 the expression was within the decan-1high cells were only found in control LN (14.7 Ϯ 3.2) at normal range (MFI 284.7; HD range MFI 195–370). The up-reg- about equal numbers as Blimp-1ϩSyndecan-1ϩ cells. ulation of CD69 was found to be much higher on B cells from patient P3 (Fig. 2B; MFI 67) than on B cells from the controls Defective formation of memory B cells in CVID patients (MFI 17.4–38) or from patient P2 (MFI 22). In terms of CD69 Flow cytometric characterization of the LN from the patients showed up-regulation, the cells of patient P3 therefore seemed to be more a significant reduction in the percentage of CD19ϩCD27ϩ memory B prone to stimulation than the cells of patient P2 or the controls. cells (28.4 Ϯ 4.1% of LN cells compared with a mean of 59.7 Ϯ After 13 days of stimulation of PBMC via CD40 and the IL-4 10.9% of LN cells from control tissues; Fig. 1, C and D; p Ͻ 0.001). receptor, only 19% of the B cells from patient P2 expressed CD38, Thus, besides a defective generation of Blimp-1ϩSyndecan-1ϩ and/or compared with 58% in patient P3 and 50.9 Ϯ 15.2% in HD (Fig. Blimp-1lowSyndecan-1high cells, impaired generation of memory B 2C). The analysis of other activation markers, such as CD27, cells in the LN from all three CVID patients was found. CD130, or CD180 revealed no difference between the stimulated The defective development of memory B cells was also detect- cells of patients and controls (data not shown). able in the peripheral blood of the patients (Fig. 1, E and F). Anal- ysis of the IgM and CD27 expression on CD19ϩ B cells of the Generation of SHM patients revealed a strong reduction in the percentage of Because the generation of SHM is closely linked to the selection of B IgMϩCD27ϩ B cells (mean value 9.0 Ϯ 2.7% of the CD19ϩ cells cells during memory and plasma B cell formation, we analyzed the (CVID) vs 18.2 Ϯ 5.7% (healthy donor; HD); p ϭ 0.004; Fig. 1E). frequency and the nucleotide substitution of SHM in CD19ϩCD27ϩ The percentage of isotype-switched memory B cells (IgMϪCD27ϩ) B cells from patients P2 and P3. As shown in Table I, the SHM was even more drastically reduced with a mean value of 1.0 Ϯ frequency was found to be very low in B cells from patient P2, but in 0.4% of the CD19ϩ B cells (CVID) vs 17.8 Ϯ 5.6% (HD) ( p Ͻ normal range (with a normal pattern) in patient P3. 0.001; Fig. 1E). CSR in vitro and in transfer experiments into immunodeficient mice Altered responses of patients’ B cells to in vitro stimulation As a strongly impaired formation of class-switched Abs is a We could previously show for a subgroup of CVID patients that common feature of CVID, we analyzed CSR in vitro and in cell BCR stimulation leads to an impaired up-regulation of CD86, but transfer experiments to elucidate possible defects. The Journal of Immunology 5501

were analyzed. We could not detect IgE CTs, indicating that CSR had not taken place. However, the first steps of CSR initiation were not impaired because AICDA transcript and eps-GLT expression were shown to be normal (Fig. 3B). The low but present amounts of IgE secreted by stimulated B cells from patient P2 may result from low numbers of memory B cells present in the peripheral blood (Fig. 1, E and F), suggesting that the patient might have had B cells able to undergo CSR but lost them over the time and during the time course of the disease. As the cells of patient P3, but not of patient P2 showed a normal CSR after stimulation in vitro, we further examined the capacity of the cells from these patients to undergo CSR in transfer experi- ments using immunodeficient mice as recipients. Mixtures of pu- rified peripheral CD4ϩ T cells and CD19ϩ B cells (6:1) from pa- tients P2 and P3 and from HD were transferred into the spleens of Rag-2/common-␥-chain-deficient mice. At time points of 7, 14, and 21 days after engraftment, serum concentration of human IgM and IgG were determined. After 3 wk, mice were sacrificed, and ϩ spleens were analyzed immunohistochemically for human CD4 Downloaded from T and CD19ϩ B cells. Consistent with the in vitro results, trans- FIGURE 2. In vitro stimulation of PBMC. A and B, PBMC were stim- ferred cells from patient P3 produced normal amounts of IgG (557 ulated for 24 h with CD40L ϩ anti-IgM and expression of CD86 (A) and ␮g/ml vs 503 ␮g/ml in the control), whereas cells from patient P2 CD69 (B) was analyzed on the surface of stimulated B cells from patients failed to secrete IgG (3 ␮g/ml) (Fig. 3D). The serum levels of IgM P2 and P3 and a HD representative for four independent controls. C, were ϳ10-fold higher in the mice engrafted with the cells from PBMC from patients P2 and P3 and one representative HD were stimulated patient P3 (107 ␮g/ml) than in control mice (11.2 ␮g/ml) or those http://www.jimmunol.org/ ϩ ϩ with CD40L IL4 for 13 days. Percentages of CD38 B cells are shown with cells from patient P2 (4.6 ␮g/ml) (Fig. 3C). In the spleens of for each dot plot. All stimulation experiments were reproduced at least all mice, human CD4ϩ T cells were detected, reflecting a success- twice depending on the availability of patient samples. ful transfer in all cases and the presence of the T cells 4 wk after engraftment (data not shown). In contrast, human CD19ϩ B cells We determined IgE levels in supernatants of PBMC stimulated were detectable neither in mice engrafted with cells from HD nor with CD40L ϩ IL-4 for 13 days. The PBMC from patient P3 from CVID patients. This might be due to the differentiation of the Ϫ produced normal amounts of IgE (3744 pg/ml), whereas the stim- B cells into Ab-secreting CD19 plasma cells, which remained ϩ ulated cells of patient P2 secreted diminished levels of IgE (994 undetected, or to a shorter half-life of CD19 B cells. pg/ml; Fig. 3A). The transfer experiments underscore the in vitro switch exper- by guest on September 24, 2021 To further characterize CSR at a molecular level in cells from iments confirming the inability of B cells from patient P2 to un- patient P2, PBMC were stimulated for 5 days with CD40L ϩ IL-4 dergo CSR. Conversely, the cells of patient P3 were able to switch and expression of AICDA as well as IgE eps-GLT and IgE eps-CT in both experiments. Discussion CVID is characterized by low concentrations or the lack of Abs. We were interested in analyzing whether any of the Ab deficien- cies result from blocks in plasma cell differentiation. To this end, we investigated late B cell differentiation steps in LNs from a subgroup of CVID patients who underwent LN biopsies. Analyz- ing samples from five patients, we found that the LN from three CVID patients contained large hyperplastic GC consisting of highly proliferating, Ki67ϩ B lymphocytes. A major fraction of cells in these GC also expressed the transcription factor BCL-6, which is essential for the proliferation of GC B cells. BCL-6 in- hibits genes involved in cell cycle control, B cell activation, and B cell differentiation, such as p27kip1, Id2, CD69, or Blimp-1, thereby maintaining the rapid proliferation of GC B cells whereas terminal differentiation into plasma cells is delayed (24). BCL-6 is essential for the GC reaction, because the formation of GC in LN and spleen was absent, and Ab affinity maturation was impaired in BCL-6-deficient mice (25, 26). Both GC histology and expression FIGURE 3. CSR in B cells from CVID patients. A, IgE secretion in of BCL-6 highly suggest that B cell differentiation is not disturbed ϩ supernatants of PBMC stimulated for 13 days with CD40L IL-4. B, CSR up to the centroblast or centrocyte stage in the LN of these patients. in PBMC after 5 days of stimulation with CD40L ϩ IL-4 (Ϫ, unstimulated; This was further underscored by the detection of OCT-2 and ϩ, CD40L ϩ IL-4). C and D, Transfer of purified B and T cells into the spleens of RAG-2/common-␥-chain-deficient mice. Human IgM (C) and BOB-1 expression (data not shown). IgG (D) titers in the serum of mice after 7, 14, and 21 days of engraftment Blimp-1 is a “master” regulator of plasma cell formation as it is with T and B cells from HD (open squares), P2 (circles), and P3 (triangles). both sufficient and required for plasmacytic differentiation (18, All experiments were reproduced at least once depending on the availabil- 27). Genes repressed by Blimp-1 are associated with proliferation, ity of the patient samples. BCR signaling, CSR or cell cycle, such as BCL-6, c-myc, btk,or 5502 PLASMA CELL DIFFERENTIATION IN CVID PATIENTS

AID (28, 29). The mutual inhibition of BCL-6 and Blimp-1 rep- can-1Ϫ subset as a less differentiated population than cells express- resents a double-negative feedback loop that ensures the exclusive ing both factors (Blimp-1ϩSyndecan-1ϩ). At first glance, the expression of genes necessary at either the GC or plasma cell model put forward by Kallies et al. suggesting a correlation be- stage. In control LN, we could observe three different plasma cell tween increasing expression of Blimp-1 and progressive matura- subsets by analyzing their Blimp-1 and Syndecan-1 expression. tion of plasma cells seems to be inconsistent with our results. How- According to these results, we propose a model in which late ever, it should be kept in mind that Kallies et al. used as a tracer plasma cell differentiation can be divided into at least three steps for Blimp-1 expression the fluorescence of enhanced green fluo- defined by their Blimp-1 and Syndecan-1 expression (Fig. 4). rescent protein (EGFP) integrated by homologous recombination Plasma cell precursors up-regulating Blimp-1 further become into the Blimp-1 encoding Prdm1 gene. Because EGFP is a highly Blimp-1/Syndecan-1 double-positive plasma cells and can differ- stable protein it tends to accumulate within cells resulting in in- entiate into Blimp-1lowSyndecan-1high cells. creased light emission (30). Therefore, changes in gene expression Blimp-1-expressing cells were observed in the LN from all three are reliably detected only by destabilized EGFP mutants (31). It is CVID patients. This finding was unexpected because the serum of therefore not unlikely that Blimp-1 and EGFP may differ in half- these patients contains no Abs, suggesting a defect in the differ- lives and turnover rates during plasma cell development, and entiation of plasma cells. However, only the LN from patient P3 increasing EGFP fluorescence intensities may not reflect propor- contained Syndecan-1ϩ cells, although at lower numbers com- tionally increasing amounts of Blimp-1 protein. Thus, the EGFP- pared with the control sections; these cells all coexpressed positive cells described by Kallies et al. may correspond to the Blimp-1. Thus, only subsets representing the first or second stages Blimp-1lowSyndecan-1high plasma cells detected in the control LN of our model of plasma cell differentiation were detected in the LN sections. Further, it is also conceivable that the observed differ- Downloaded from of these CVID patients. Further, the numbers of Blimp-1ϩSynde- ences are due to species differences and/or might reflect the distinct can-1Ϫ cells in the LN of the patients were higher than in the tissues examined. control LN, which is consistent with our hypothesis of a blocked A previous report ascribed the appearance of Syndecan-1ϩ but plasma cell differentiation. Blimp-1Ϫ B cells to technical problems arising when the plane of In addition to the arrested plasma cell differentiation, peripheral the section does not pass through the center of the nucleus, thereby

B cells from patient P2 also showed an impaired activation in impeding a detection of Blimp-1 (23). Based on the facts that such http://www.jimmunol.org/ terms of CD86 and CD38 up-regulation after in vitro stimulation cells abundantly surround the GC of control LN and tonsils but via the BCR or CD40L ϩ IL-4, respectively. Defective isotype completely lack in LN from patient P3, we are putting forward that switch was shown in vitro and by cell transfer experiments into Blimp-1lowSyndecan-1high B cells represent a discrete B lympho- Rag-2/common-␥-chain-deficient mice, where the cells did not se- cyte subset rather than a technical artifact. Because Blimp-1 is crete IgG but normal levels of IgM. Further, analysis of SHM in known to be a transcription factor required for the transition from CD19ϩCD27ϩ cells from the peripheral blood of patient P2 GC B cells to plasma cells, it is conceivable that Blimp-1 expres- showed no mutated V regions. In contrast, the defect of patient P3 sion is only high at this point and that the Blimp-1lowSyndecan- seems not to interfere with these processes because the cells re- 1high1low cells represent plasma cells of a later stage, in which vealed intact CSR both in vitro and in transfer experiments and Blimp-1 is reduced to a minimum sufficient to maintain the plasma by guest on September 24, 2021 exhibited no defective up-regulation of CD86 and CD38 and V cell stage. regions in CD19ϩCD27ϩ cells were found to be mutated both with Terminal plasma cell formation therefore may be impaired even normal frequency and pattern. though Blimp-1 is expressed. The blocked differentiation step var- The finding of normal proliferation and production of cytokines ied between patients P1 and P2 vs patient P3 suggesting different such as IFN-␥, IL-2, and IL-10 by T cells (data not shown) cor- underlying mechanisms. It will be of interest to examine the func- roborates the B cell specificity of the defect in patient P2, whereas tional integrity of Blimp-1, especially seen in patients P1 and P2, in P3 the defect seems to affect cells other than B or T cells, which where no Syndecan-1ϩ cells were detectable. However, a severe might provide factors necessary for B cell differentiation in LN. defect of Blimp-1 is improbable because Blimp-1 also has impor- In mice, distinct subsets of plasma cells differentially expressing tant functions in cells other than plasma cells rendering a defi- Blimp-1 and Syndecan-1 have also been reported by Kallies et al. ciency for Blimp-1 embryonic lethal (32). Besides Blimp-1, the (19). The authors describe plasma cells of the Blimp-1ϩSynde- expression of XBP-1 is also pivotal for terminal plasma cell dif- ferentiation (33). Because XBP-1 acts downstream of Blimp-1, it is conceivable, that the induction of this protein might be impaired in our patients. However, deletion of XBP-1 is embryonic lethal too, because it is an important factor in the unfolded protein re- sponse and is essential for the development of fetal hepatocytes (34). Keeping this in mind, it is unlikely that a genetic defect abrogating the function of XBP-1 accounts for the defective ter- minal plasma cell differentiation in our CVID patients. In conclusion, our results define three subsets of plasma cells, according to their differential Blimp-1 and Syndecan-1 expression, in control LN. In LN from the three CVID patients that we ana- lyzed here, we were able to unravel distinct blocks during these stages of plasma cell differentiation. Our data point toward differ- ent B cell defects downstream of Blimp-1 that seem to represent essential factors in plasma cell development. Although different subgroups of CVID patients may suffer from other defects or mu- tations (3, 6, 7) defective plasma cell differentiation is a key find- FIGURE 4. A proposed simple model of late plasma cell differentiation ing in CVID and it will be of major interest to further investigate defined by Ki67, BCL-6, Blimp-1, and Syndecan-1 expression. target genes downstream of Blimp-1. The Journal of Immunology 5503

Acknowledgments nuclear phosphoprotein, is highly expressed in germinal center B cells and their neoplastic counterparts. Blood 86: 28–37. We gratefully acknowledge the help, suggestions, and advice of Uli Salzer 17. Keller, A. D., and T. Maniatis. 1992. Only two of the five zinc fingers of the and Michael Schlesier. eukaryotic transcriptional repressor PRDI-BF1 are required for sequence-specific DNA binding. Mol. Cell Biol. 12: 1940–1949. Disclosures 18. Turner, C. A., Jr., D. H. Mack, and M. M. Davis. 1994. Blimp-1, a novel zinc finger-containing protein that can drive the maturation of B lymphocytes into The authors have no financial conflict of interest. immunoglobulin-secreting cells. Cell 77: 297–306. 19. Kallies, A., J. Hasbold, D. M. Tarlinton, W. Dietrich, L. M. Corcoran, References P. D. Hodgkin, and S. L. Nutt. 2004. Plasma cell ontogeny defined by quantitative changes in blimp-1 expression. J. Exp. Med. 200: 967–977. 1. International Union of Immunological Societies. 1999. Primary immunodefi- 20. Durandy, A., C. Schiff, J. Y. Bonnefoy, M. Forveille, F. Rousset, G. Mazzei, ciency diseases. Report of an IUIS Scientific Committee. Clin. Exp. Immunol. M. Milili, and A. Fischer. 1993. Induction by anti-CD40 antibody or soluble 118(Suppl. 1): 1–28. CD40 ligand and cytokines of IgG, IgA and IgE production by B cells from 2. Cunningham-Rundles, C., and C. Bodian. 1999. Common variable immunodefi- patients with X-linked hyper IgM syndrome. Eur. J. Immunol. 23: 2294–2299. ciency: clinical and immunological features of 248 patients. Clin. Immunol. 92: 34–48. 21. Imai, K., N. Catalan, A. Plebani, L. Marodi, O. Sanal, S. Kumaki, V. Nagendran, 3. Grimbacher, B., A. Hutloff, M. Schlesier, E. Glocker, K. Warnatz, R. Drager, P. Wood, C. Glastre, F. Sarrot-Reynauld, et al. 2003. Hyper-IgM syndrome type H. Eibel, B. Fischer, A. A. Schaffer, H. W. Mages, et al. 2003. Homozygous loss 4 with a B lymphocyte-intrinsic selective deficiency in Ig class-switch recombi- of ICOS is associated with adult-onset common variable immunodeficiency. Nat. nation. J. Clin. Invest. 112: 136–142. Immunol. 4: 261–268. 22. Colucci, F., C. Soudais, E. Rosmaraki, L. Vanes, V. L. Tybulewicz, and 4. North, M. E., A. D. Webster, and J. Farrant. 1991. Defects in proliferative re- J. P. Di Santo. 1999. Dissecting NK cell development using a novel alymphoid sponses of T cells from patients with common variable immunodeficiency on mouse model: investigating the role of the c-abl proto-oncogene in murine NK direct activation of protein kinase C. Clin. Exp. Immunol. 85: 198–201. cell differentiation. J. Immunol. 162: 2761–2765. 5. Rump, J. A., A. Jahreis, M. Schlesier, R. Drager, I. Melchers, and H. H. Peter. 23. Angelin-Duclos, C., G. Cattoretti, K. I. Lin, and K. Calame. 2000. Commitment of B lymphocytes to a plasma cell fate is associated with Blimp-1 expression in 1992. Possible role of IL-2 deficiency for hypogammaglobulinaemia in patients Downloaded from with common variable immunodeficiency. Clin. Exp. Immunol. 89: 204–210. vivo. J. Immunol. 165: 5462–5471. 6. Groth, C., R. Drager, K. Warnatz, G. Wolff-Vorbeck, S. Schmidt, H. Eibel, 24. Shaffer, A. L., X. Yu, Y. He, J. Boldrick, E. P. Chan, and L. M. Staudt. 2000. M. Schlesier, and H. H. Peter. 2002. Impaired up-regulation of CD70 and CD86 BCL-6 represses genes that function in lymphocyte differentiation, inflammation, in naive (CD27Ϫ) B cells from patients with common variable immunodeficiency and cell cycle control. Immunity 13: 199–212. (CVID). Clin. Exp. Immunol. 129: 133–139. 25. Dent, A. L., A. L. Shaffer, X. Yu, D. Allman, and L. M. Staudt. 1997. Control of 7. Denz, A., H. Eibel, H. Illges, G. Kienzle, M. Schlesier, and H. H. Peter. 2000. inflammation, cytokine expression, and germinal center formation by BCL-6. Impaired up-regulation of CD86 in B cells of “type A” common variable immu- Science 276: 589–592. nodeficiency patients. Eur. J. Immunol. 30: 1069–1077. 26. Ye, B. H., G. Cattoretti, Q. Shen, J. Zhang, N. Hawe, R. de Waard, C. Leung, 8. Eisenstein, E. M., and W. Strober. 1995. Evidence for a generalized signaling M. Nouri-Shirazi, A. Orazi, R. S. Chaganti, et al. 1997. The BCL-6 proto-onco- http://www.jimmunol.org/ abnormality in B cells from patients with common variable immunodeficiency. gene controls germinal-centre formation and Th2-type inflammation. Nat. Genet. Adv. Exp. Med. Biol. 371B: 699–704. 16: 161–170. 9. Levy, Y., N. Gupta, F. Le Deist, C. Garcia, A. Fischer, J. C. Weill, and 27. Shapiro-Shelef, M., K. I. Lin, L. J. McHeyzer-Williams, J. Liao, C. A. Reynaud. 1998. Defect in IgV gene in common M. G. McHeyzer-Williams, and K. Calame. 2003. Blimp-1 is required for the variable immuno-deficiency syndrome. Proc. Natl. Acad. Sci. USA 95: formation of immunoglobulin secreting plasma cells and pre-plasma memory B 13135–13140. cells. Immunity 19: 607–620. 10. Warnatz, K., A. Denz, R. Drager, M. Braun, C. Groth, G. Wolff-Vorbeck, 28. Shaffer, A. L., A. Rosenwald, E. M. Hurt, J. M. Giltnane, L. T. Lam, H. Eibel, M. Schlesier, and H. H. Peter. 2002. Severe deficiency of switched O. K. Pickeral, and L. M. Staudt. 2001. Signatures of the immune response. memory B cells (CD27ϩIgMϪIgDϪ) in subgroups of patients with common vari- Immunity 15: 375–385. able immunodeficiency: a new approach to classify a heterogeneous disease. 29. Shaffer, A. L., K. I. Lin, T. C. Kuo, X. Yu, E. M. Hurt, A. Rosenwald, Blood 99: 1544–1551. J. M. Giltnane, L. Yang, H. Zhao, K. Calame, and L. M. Staudt. 2002. Blimp-1 11. Brouet, J. C., A. Chedeville, J. P. Fermand, and B. Royer. 2000. Study of the B orchestrates plasma cell differentiation by extinguishing the mature B cell gene by guest on September 24, 2021 cell memory compartment in common variable immunodeficiency. Eur. J. Im- expression program. Immunity 17: 51–62. munol. 30: 2516–2520. 30. Chalfie, M., Y. Tu, G. Euskirchen, W. W. Ward, and D. C. Prasher. 1994. Green 12. Agematsu, K., T. Futatani, S. Hokibara, N. Kobayashi, M. Takamoto, S. Tsukada, fluorescent protein as a marker for gene expression. Science 263: 802–805. H. Suzuki, S. Koyasu, T. Miyawaki, K. Sugane, et al. 2002. Absence of memory 31. Li, X., X. Zhao, Y. Fang, X. Jiang, T. Duong, C. Fan, C. C. Huang, and B cells in patients with common variable immunodeficiency. Clin. Immunol. 103: S. R. Kain. 1998. Generation of destabilized green fluorescent protein as a tran- 34–42. scription reporter. J. Biol. Chem. 273: 34970–34975. 13. Piqueras, B., C. Lavenu-Bombled, L. Galicier, F. Bergeron-van der Cruyssen, 32. Chang, D. H., C. Angelin-Duclos, and K. Calame. 2000. BLIMP-1: trigger for L. Mouthon, S. Chevret, P. Debre, C. Schmitt, and E. Oksenhendler. 2003. Com- differentiation of myeloid lineage. Nat. Immunol. 1: 169–176. mon variable immunodeficiency patient classification based on impaired B cell 33. Reimold, A. M., N. N. Iwakoshi, J. Manis, P. Vallabhajosyula, memory differentiation correlates with clinical aspects. J. Clin. Immunol. 23: E. Szomolanyi-Tsuda, E. M. Gravallese, D. Friend, M. J. Grusby, F. Alt, and 385–400. L. H. Glimcher. 2001. Plasma cell differentiation requires the transcription factor 14. Allman, D., A. Jain, A. Dent, R. R. Maile, T. Selvaggi, M. R. Kehry, and XBP-1. Nature 412: 300–307. L. M. Staudt. 1996. BCL-6 expression during B-cell activation. Blood 87: 34. Reimold, A. M., A. Etkin, I. Clauss, A. Perkins, D. S. Friend, J. Zhang, 5257–5268. H. F. Horton, A. Scott, S. H. Orkin, M. C. Byrne, et al. 2000. An essential role 15. Cattoretti, G., C. C. Chang, K. Cechova, J. Zhang, B. H. Ye, B. Falini, in liver development for transcription factor XBP-1. Genes Dev. 14: 152–157. D. C. Louie, K. Offit, R. S. Chaganti, and R. Dalla-Favera. 1995. BCL-6 protein 35. Bryant, A., N. C. Calver, E. Toubi, A. D. Webster, and J. Farrant. 1990. Clas- is expressed in germinal-center B cells. Blood 86: 45–53. sification of patients with common variable immunodeficiency by B cell secretion 16. Onizuka, T., M. Moriyama, T. Yamochi, T. Kuroda, A. Kazama, N. Kanazawa, of IgM and IgG in response to anti-IgM and interleukin-2. Clin. Immunol. Im- K. Sato, T. Kato, H. Ota, and S. Mori. 1995. BCL-6 gene product, a 92- to 98-kD munopathol. 56: 239–248.