Article Text

Extended report
Persistence of immunoglobulin-producing cells in parotid salivary glands of patients with primary Sjögren's syndrome after B cell depletion therapy
  1. Nishath Hamza1,
  2. Hendrika Bootsma1,
  3. Saravanan Yuvaraj2,
  4. Fred K L Spijkervet3,
  5. Erlin A Haacke4,
  6. Rodney P E Pollard3,
  7. Annie Visser1,2,
  8. Arjan Vissink3,
  9. Cees G M Kallenberg1,
  10. Frans G M Kroese1,2,
  11. Nicolaas A Bos1,2
  1. 1Department of Rheumatology and Clinical Immunology, University Medical Center Groningen, University of Groningen, Groningen, Netherlands
  2. 2Department of Cell Biology, Section Immunology, University of Groningen, University Medical Center Groningen, Groningen, Netherlands
  3. 3Department of Oral and Maxillofacial Surgery, University of Groningen, University Medical Center Groningen, Groningen, Netherlands
  4. 4Department of Pathology and Medical Biology, University of Groningen, University Medical Center Groningen, Groningen, Netherlands
  1. Correspondence to Nicolaas A Bos, University Medical Center Groningen, University of Groningen, Rheumatology and Clinical Immunology, Groningen 9700 RB, Netherlands; n.a.bos{at}umcg.nl

Abstract

Objectives To assess the persistence of immunoglobulin-producing cell populations in the parotid salivary glands of patients with primary Sjögren's syndrome (pSS) after B cell depletion therapy with rituximab.

Methods Thirteen patients with pSS and four control patients were included in this study. Patients with pSS were treated with rituximab or placebo. Sequence analysis was carried out on IgA- and IgG-encoding transcripts extracted from parotid salivary gland biopsy specimens taken before treatment and at 12–16 and 36–52 weeks after treatment.

Results At baseline, many clonally related sequences were seen in patients with pSS. The number of clonal expansions was significantly higher in patients with pSS than in control patients. Clonal expansions were composed of IgA- and/or IgG-expressing cells. Rituximab did not significantly alter the degree of clonal expansions. Groups of clonally related cells had members which were shared between biopsy specimens taken before and after treatment. Mutation frequencies of immunoglobulin sequences from clonally related cells in patients with pSS were higher after treatment.

Conclusions Rituximab treatment does not alter the characteristic features of increased clonal expansions seen in the parotid salivary glands of patients with pSS. The presence of clonally related immunoglobulin-producing cells before and after rituximab treatment strongly suggests that immunoglobulin-producing cells persist in salivary glands of patients with pSS despite B cell depletion. The presence of mixed isotype expression within groups of clonally related cells indicates local class switching in salivary glands of patients with pSS. Persistent immunoglobulin-producing cells may underlie disease relapse after treatment.

View Full Text

Statistics from Altmetric.com

Introduction

Primary Sjögren's syndrome (pSS) is an autoimmune disease characterised by chronic inflammation of the salivary and lachrymal glands1 and progressive dryness of mouth and eyes. The presence of B cell infiltrates in the salivary glands of patients with pSS, accompanied by disturbances in the relative proportions of peripheral B cell subsets and high titres of circulating autoantibodies such as anti-Ro (SS-A) and anti-La (SS-B) antibodies in the blood, strongly implicates B cells in the pathogenesis of pSS.2 This assumption was strengthened when therapeutic B cell depletion strategies targeting CD20, a B cell-specific transmembrane protein, resulted in relief from SS disease symptoms.3,,7

B cell targeting therapies with anti-CD20 monoclonal antibodies, such as the chimeric antibody rituximab (RTX) are now widely used to treat autoimmune disorders such as rheumatoid arthritis, systemic lupus erythematosus and vasculitis.8 We and others have previously shown that RTX can also be successfully used for the treatment of patients with pSS.3 ,6 ,9 ,10

However, symptoms usually return 6–9 months after treatment.6 ,10 Although RTX treatment results in the almost complete absence of B cells in the peripheral blood of patients with pSS for 3–6 months after the initiation of the treatment,10,,12 this is not mirrored by their parotid salivary glands, where B cells remain, albeit in decreased numbers.7 Also, disease relapse seemingly coincides with the reappearance of peripheral B cell subpopulations in the blood.9,,11

B cells can be distinguished from each other by the unique rearrangements of their immunoglobulin variable heavy (IGHV) and light chain genes. B cell clones (and plasma cells derived from them) share the same IGHV gene rearrangements. Previous studies have reported the presence of B cell clones within salivary gland biopsy specimens from patients with pSS using immunoglobulin sequence analysis.13,,15 To assess whether such B cell clones might persist in salivary glands of patients with pSS after RTX treatment, we analysed immunoglobulin sequences obtained from repeated biopsy specimens taken from the same parotid gland before and after RTX treatment. If clonally related immunoglobulin (Ig)-producing cells are seen before and after RTX treatment in patients with pSS, this may suggest a potential role for these persistent cells in the disease relapse occurring after B cell depletion therapy with RTX.

Patients and methods

Patients

Thirteen patients with pSS fulfilling the American–European criteria for pSS16 (all women; mean age 40.5 years; range 18–65 years), with a disease duration of <5 years, were enrolled in this study after providing written informed consent. The study protocol was approved by the institutional review board of the University Medical Center Groningen. Patients included in this study had participated in previous pilot studies, such as the open-label rituximab study (n=4)6 or the placebo-controlled rituximab study (n=9).9

Inclusion criteria for patients with pSS to participate in this study were stimulated whole saliva secretion flow >0.15 ml/min, presence of autoantibodies (IgM-rheumatoid factor ≥10 klU/l in combination with anti-SS-A and/or anti-SS-B autoantibodies), and a salivary parotid gland biopsy specimen (obtained ≤12 months before inclusion) showing characteristic features for SS.17 Patients with a history of any malignancy (such as MALT lymphoma) or underlying cardiac, pulmonary, metabolic, renal or gastrointestinal conditions or with chronic or latent infectious diseases or immune deficiency were excluded. Furthermore, patients who had been treated previously with monoclonal antibodies were excluded. Treatment with prednisone and hydroxychloroquine had to be discontinued at least 1 month before baseline, and treatment with other disease-modifying antirheumatic drugs at least 6 months before baseline.

Eleven patients who had received RTX (patients with pSS-RTX) had been treated according to the following two schemes: a RTX infusion of 375 mg/m2/week for 4 weeks (in the open-label study; n=4)6 or two intravenous infusions of 1000 mg RTX (in the placebo-controlled trial; n=7)9 on days 1 and 15. Two patients had been treated with placebo infusions (patients with pSS-placebo).9 To minimise side effects (infusion reactions, serum sickness), all patients, both RTX- and placebo-treated had been premedicated with methylprednisolone (100 mg/intravenously), paracetamol (1000 mg/by mouth) and clemastine (2 mg/intravenously) and received 60 mg of oral prednisone on days 1 and 2, 30 mg on days 3 and 4 and 15 mg on day 5 after each infusion. No corticosteroids or other disease-modifying antirheumatic drugs were allowed during follow-up.

As controls (all women; mean age 51 years; range 45–57 years), two patients with sicca complaints not fulfilling the American–European criteria for pSS and two patients with malignancies (squamous cell carcinoma of the oral cavity) without involvement of the parotid salivary glands were included.

An incisional biopsy specimen of the parotid gland was obtained from the same gland before treatment (RTX or placebo) and 12–16 weeks later.17 Moreover, in five pSS-RTX patients, a third sample was taken at 36–52 weeks from baseline, at the time of recurrence of clinical symptoms. In control patients, a biopsy specimen of the parotid gland was obtained either as part of the clinical diagnostic investigation for SS or as part of a neck dissection procedure in patients with malignancy during surgical intervention. Histopathological examination of the parotid glands of all four control patients showed a normal histology of the glandular tissue.

Cloning and sequencing of immunoglobulin transcripts

From every biopsy specimen, three batches of four serial tissue sections (5–7 µm thick) were obtained from different parts of the specimen. Total RNA was extracted from each batch. RNA was converted to cDNA and amplified using primers specific for the variable region of IGHV genes18 in combination with constant region primers specific for the CH1 domains of the Cα (5′-GAATTCGAGTGGCTCCTGGGGGAAGA-3′) or Cγ (5′-GAGTTCCACGACACCGTCAC-3′) constant regions.19 For four patients with pSS-RTX and four control patients, we performed a multiplex PCR of all IGHV families. For the remaining nine patients with pSS-RTX, we focused our analysis on the largest IGHV family—that is, IGHV3. PCR was performed (see online supplementary data) and the PCR products were gel-extracted and cloned into plasmid vectors using the Fermentas GeneJet Kit (Fermentas, St Leon-Rot, Germany) and Xli Blue competent bacteria. A total of 72 (36 IgA+36 IgG) plasmid-PCR product constructs were picked separately for each biopsy and submitted for sequencing. Only plasmids containing IGHV sequence inserts were considered for further analysis.

The independent sampling of different parts of each biopsy ensured that 100% identical immunoglobulin sequences derived from separate tissue batches could be attributed to different cells. Identical immunoglobulin sequences within a single tissue batch could be derived from multiple transcripts of the same cell or from different cells. For this reason, 100% identical sequences obtained from one tissue batch were counted as one.

Analysis of rearranged immunoglobulin genes

The nucleotide sequences of the variable region were compared with the international ImMunoGeneTics information system (IMGT) databases of human immunoglobulin germline sequences.20 ,21 Imunoglobulin sequences obtained from independent PCRs were assigned to clonally related Ig-producing cells based on their similarity at the complementarity determining region (CDR) 3 and their shared IGHV gene usage.22 Shared mutations within the IGHV regions were also considered to be indicative of clonal relationships. Mutation frequencies of immunoglobulin sequences were computed with an in-house software tool.23

Immunohistochemistry

Sections of fixed and paraffin-embedded parotid gland biopsy specimens from patients with pSS included in this study were immunohistochemically stained with anti-CD79, anti-IgA and anti-IgG antibodies (see supplementary data for protocol).

Statistical analysis

All statistical analyses were performed using GraphPad Prism software (version 3.0). Statistical comparisons of data from patients with pSS and non-pSS controls were carried out using unpaired t test. Data from baseline and at different times after treatment were analysed using paired t test.

Results

A total of 1314 immunoglobulin (803 IgA and 511 IgG) sequences were collected from patients with pSS (at baseline and after RTX treatment) and from control patients. Of these sequences, 109 (8%) were designated as unproductive (non-protein coding) by the IMGT Quest bioinformatic tool. After filtering out unproductive sequences and 100% identical sequences from the same PCRs, the number of productive (coding for a functional protein) immunoglobulin sequences from different Ig-producing cells was 1172 (1084 from patients with pSS and 88 from control patients) and only these sequences were used for subsequent analysis.

Increased clonal expansions in parotid salivary glands of patients with pSS

Analysis of immunoglobulin sequences obtained from parotid salivary glands, from both patients with pSS and non-pSS controls, clearly showed the existence of sequences derived from clonally related cells. In patients with pSS, at baseline, on average 20±1.85% of the sequences collected from a parotid gland biopsy specimen belonged to clonally related cells, whereas in controls, the average percentage (7±2.37%) was significantly lower (p=0.0027, unpaired t test; figure 1A). The number of groups of clonally related cells per biopsy (normalised to 100 sequences/biopsy) from patients with pSS at baseline was also significantly higher than those from controls (p=0.0022, unpaired t test figure 1B); and the sizes (ie, the number of clonally related sequences) of these clonal groups were significantly larger in pSS patient biopsy specimens at baseline than in controls (p=0.0457, unpaired t test; figure 1C).

Figure 1

Immunoglobulin sequences from clonally related cells detected in parotid gland biopsy specimens from patients with primary Sjögren's syndrome (pSS) and control patients. (A) Percentage of immunoglobulin sequences belonging to clonally related cells in biopsy specimens from patients with pSS at baseline compared with those from control patient specimens. (B) Number of groups of clonally related cells per biopsy specimen (normalised to 100 sequences/biopsy) from patients with pSS at baseline compared with those from control patient specimens. (C) Size of clonal groups from patients with pSS at baseline compared with those from control patient biopsy specimens.

Clonally related cells in salivary glands before and after RTX treatment

Comparison of immunoglobulin sequences obtained at different time points showed that clonally related cells were not only present within biopsy specimens taken at baseline, but that members of these groups of clonally related cells were also present in specimens taken after RTX treatment. We did not observe any identical IGHV sequences between different patients.

In total, 42 groups of immunoglobulin sequences from clonally related cells were obtained both from samples taken at baseline and after RTX. In 35 of these groups, members were found both in samples taken at baseline and at 12–16 weeks after RTX (occurring in eight out of 11 patients with pSS-RTX). In seven groups, members were found both in samples at baseline and in samples taken at 36–52 weeks after RTX. In one pSS-RTX patient, members of two groups of clonally related Ig-producing cells were detected at all three time points. A graphical representation of the immunoglobulin sequences from clonally related cells detected at different times in five patients with pSS-RTX is shown in figure 2. In the two patients with pSS-placebo, we observed five groups of immunoglobulin sequences from clonally related cells before and 12–16 weeks after treatment with placebo.

Figure 2

Immunoglobulin (Ig) sequences from clonally related cells in parotid salivary glands at baseline and after rituximab (RTX) treatment. Each circle represents the sequence from an Ig-producing cell. This figure shows Ig sequences detected at three different time points in five patients with primary Sjögren's syndrome treated with RTX. The black and white shaded circles are clonally related cells found only at a single time point. Coloured circles represent clonally related cells seen at different time points within a particular patient. Note that identical colour codes indicate the B cell clones within each patient and do not imply any B cell clonality between patients.

The percentage of sequences belonging to clonally related cells did not change significantly at 12–16 weeks and 36–52 weeks after treatment with RTX, compared with those at baseline. Both the number of groups and the clone size at all time points showed the same distribution as seen in patients with pSS at baseline (figure 3).

Figure 3

Immunoglobulin sequences from clonally related cells detected at baseline and after rituximab (RTX) treatment in patients with primary Sjögren's syndrome (pSS). (A) Percentage of immunoglobulin sequences belonging to clonally related cells in biopsy specimens from patients with pSS at baseline compared with those at 12–16 weeks and 36–52 weeks. (B) Number of groups of clonally related cells per biopsy (normalised to 100 sequences/biopsy) in biopsy specimens from patients with pSS at baseline compared with those at 12–16 weeks and 36–52 weeks. (C) Size of clonal groups in biopsy specimens from patients with pSS at baseline compared with those at 12–16 weeks and 36–52 weeks.

B cells and plasma cells are present in parotid glands before and after RTX treatment

In parotid sections from patients with pSS, we observed the presence of B cells as well as IgA- and IgG-producing cells at all three time points. Both B cell and plasma cell numbers apparently decreased at 12 weeks after RTX treatment compared with baseline (a representative sample is shown in figure 4).

Figure 4

B cells, IgA and IgG-expressing cells in a parotid gland from a patient with primary Sjögren's syndrome before and after rituximab (RTX) treatment. Sections of a parotid gland were stained with anti-CD79, anti-IgA and anti-IgG antibodies before RTX and at 16 weeks after RTX. B cells as well as IgA- and IgG-producing cells were detected at all time points (data not shown for 36–52 weeks after RTX).

Increased mutation frequencies in immunoglobulin sequences from clonally related cells observed after RTX treatment in patients with pSS

Virtually all productive immunoglobulin sequences from patients with pSS and control patients carried mutations in comparison with known germline IGHV genes (1170/1172 sequences). Overall, the combined average mutation frequencies of all IGHV sequences (from clonally related and unrelated cells) were not significantly different before (11.1±0.49%) and after RTX treatment (10.9±0.38% at 12–16 weeks and 11.0±0.90% at 36–52 weeks after RTX treatment).

However, when only the groups of clonally related cells were analysed separately, we observed significantly higher (p=0.0007, paired t test) mutation frequencies from cells at 36–52 weeks after RTX treatment than its closest clone (homologous sequence) seen at baseline. An example of a group of sequences from clonally related cells in patients with pSS-RTX at different time points and their differences in mutation patterns is shown in figure 5A,B.

Figure 5

Immunoglobulin sequence alignment of clonally related cells at baseline and after rituximab (RTX) treatment. An example of six immunoglobulin sequences belonging to persistent clonally related cells from patient 2 indicated in figure 2 (depicted as pink circles) is shown here. (A) Sequences are aligned with the germline gene and their predicted CDR3-junction amino acid (aa) sequences. Dashes indicate identical nucleotides and gaps resulting from ImMunoGeneTics information system (IMGT) subdivision are marked by dots. (B) A genealogical tree is deduced from the mutations observed in the immunoglobulin variable heavy genes and each cell (sequence) is shown with the observed isotype. P1, P2.1, P2.2 and P3 indicate consecutive hypothetical precursor B cells that share mutations with the sequences from clonally related cells.

Evidence for localised immunoglobulin class switching in parotid salivary glands of patients with pSS

One hundred and one groups of clonally related sequences were obtained from biopsy specimens from patients with pSS. Thirty-nine per cent (39/101) of sequences in these groups included only IgA sequences and another 39% had only IgG sequences. Twenty-three per cent (23/101) included members of both isotypes (IgA and IgG), of which 15 groups included immunoglobulin sequences seen at different time points and the remaining eight groups were detected within the same specimens. These cells probably switched after the somatic hypermutation process occurred as suggested by shared mutations in the IGHV genes.

Discussion

In this study, using IGHV sequence analyses, we provide molecular evidence for the local persistence of certain B cell populations within the salivary glands of patients with pSS after RTX treatment.

Increased clonal expansions of Ig-producing cells are a characteristic feature of parotid glands from patients with pSS

Our comparison of the immunoglobulin sequences from biopsies of pSS and control patients indicates that patients with pSS can be characterised by the presence of increased clonal expansions within the IgA- or IgG-expressing cell populations present in parotid glands. Presence of clonally related cells has been previously seen in non-inflamed parotid salivary glands in humans24 and in rats.25 Clonal expansions in salivary glands of patients with pSS have also been previously indicated.13,,15 ,26 Our study adds to these previous reports by including comparisons with non-pSS control patients, thus enabling us to attribute increased clonal expansions as a feature observed within salivary glands of patients with pSS. Moreover, we show that the intrinsic property of an increased degree of clonal expansions in patients with pSS at baseline was largely unaltered by RTX treatment.

As we analysed mRNA of entire tissue sections, we did not select for specific immunoglobulin-expressing cell subsets. For this reason it is conceivable that the majority of the immunoglobulin sequence data obtained reflects the immunoglobulin expression of class-switched plasma blasts or plasma cells, which express approximately 100–1000-fold more immunoglobulin transcripts than B cells.27 On the other hand, this approach has the advantage of providing a greater representation of plasma cell populations, which may include those that actively produce autoantibodies.

It is possible that the biopsy sections from control patients may have much fewer Ig-expressing cells than patients with pSS. We attempted to control for any quantitative bias by picking 36 plasmids for IgA and 36 plasmids for IgG analysis from each biopsy specimen. Even then, owing to the amplification biases that may be introduced by PCR, the number of IGHV sequences recovered from these plasmids can only be a rough estimate of the abundance of Ig-producing cells in the biopsy sections.

We detected immunoglobulin sequences within groups of clonally related cells expressing different isotypes in patients with pSS. Similarly, Dunn-Walters et al24 had also observed sequences of both IgA and IgM isotypes among members of three groups of clonally related cells in non-inflamed human parotid salivary glands. In our study, such clonally related sequences consisted either of combinations of IgG1 and IgA transcripts or of IgA1 and IgA2 transcripts. These findings provide evidence for the existence of localised class switching within the salivary glands of patients with pSS. Since class switching is associated with proliferation, these observations are in line with the presence of clonal expansions, reflecting the hyperactive state of B cells in affected salivary glands of patients with pSS.

Persistence of clonally related cells before and after treatment with RTX

Our detection of immunoglobulin sequences from clonally related cells in nearly all (nine out of 11) patients with pSS before and after RTX indicates that Ig-producing cells can persist within the parotid salivary glands even at time points after RTX administration when the peripheral blood is almost completely devoid of B cells.11 This was also confirmed by immunohistochemical staining of parotid gland tissue sections. All sequences belonging to groups of clonally related cells obtained at baseline and after RTX treatment were heavily mutated. Furthermore, at 36–52 weeks after RTX, mutation frequencies of all sequences from clonally related cells were higher than at baseline, indicating continuing proliferation of these Ig-producing cells. Although we see mutations in both the framework and CDR regions, many of the mutations in the framework region are silent and there is no clear mutation pattern that is consistent with positive selection within CDR regions (data not shown). All of the above observations give credence to a model of disease relapse after RTX treatment that is seeded by persisting Ig-producing cells.

Several non-mutually exclusive mechanisms may explain the apparent survival of the persistent Ig-producing cells after RTX. One possibility is that these might be long-lived plasma cells that do not express CD20 and are therefore not targeted for depletion by RTX.28 Another explanation might be that these are persistent cells derived from memory B cells that possess a survival advantage to resist RTX depletion.29 Alternatively, the surviving cells may be situated in restricted niches that enable them to evade depletion and proliferate with time. This seems plausible, given that higher mutation patterns were seen in immunoglobulin sequence members from clonally related cells in biopsy specimens taken after RTX compared with baseline. The existence of niches with apparently restricted access to RTX was previously reported in murine systems, where B cells in certain tissue sites, such as the splenic marginal zone, germinal centre and peritoneal cavities, exhibit significant resistance to anti-CD20 depletion.30 ,31 This phenomenon was also seen in germinal centre B cells of the lymph nodes in non-human primates32 and in tonsils33 and other lymphoid organs in humans.34

The transient clinical relief from SS symptoms after RTX is probably due to ablation of B cell and CD20+ Ig-producing cell numbers.35 Although no significant decrease in serum Ig levels was seen after RTX,6 lower levels of certain (auto) antibodies may contribute to clinical relief. Reduction in other effector B cell functions, such as antigen presentation and cytokine production, are alternative explanations.36 ,37 At the same time, the underlying autoimmune mechanisms are probably maintained by long-lived plasma cells, as has been indicated in studies in systemic lupus erythematosus where patients who expressed autoantibodies secreted by long-lived plasma cells, had a higher chance of experiencing early flares or disease relapse after B cell depletion therapy than patients who did not express these autoantibodies.36 ,38,,40 Long-lived plasma cells were also reported in labial salivary glands of patients with pSS.41 In this context, our study provides a robust argument for the development and inclusion of therapeutic strategies that deplete plasma blast and/or plasma cell populations in patients with pSS.36

To conclude, we provide evidence for the existence of certain Ig-producing cell populations in the parotid salivary glands of patients with pSS that may have escaped depletion. The cells that survive after RTX may ultimately contribute to the disease relapse seen in patients with pSS who have undergone RTX treatment.

References

View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Funding Dutch Arthritis Foundation.

  • Competing interests None.

  • Ethics approval University Medical Center Groningen.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.