Primary Sjögren’s syndrome (pSS) is a systemic autoimmune disease, characterized by impaired secretion of exocrine glands. Activated B cells play a key role in the pathogenesis of pSS, through the production of autoantibodies and the development of ectopic germinal centers in the salivary glands and other affected sites (1, 2). The produced autoantibodies cover a wide spectrum, including antinuclear, anti-Ro/SS-A, anti-La/SS-B and rheumatoid factor (RF) antibodies (3). In addition, 5-10% of pSS patients develop B-cell lymphoma, typically extranodal marginal zone lymphomas (eMZL) of the mucosa-associated lymphoid tissue (MALT), of which around 70% in the salivary glands (4). Lymphoma is the main cause of a decreased survival in pSS (5). Strikingly, parotid eMZL in pSS frequently express antibodies with RF activity, suggesting autoreactivity is an early driver during lymphoma development (6, 7). RF clones were previously shown to be enriched in inflamed tissues, particularly during pSS-related eMZL development (6). Therefore, a putative model for parotid eMZL lymphomagenesis has been proposed where RF clones organize in ectopic germinal center-like structures in a salivary gland, stimulating somatic hypermutation, proliferation and accumulation of driver mutations (8–10).

During their care for patients with pSS clinicians are confronted with a number of dilemmas. Patients are monitored for development of a complicated disease course by rheumatologists or immunologists, with lymphoma considered to be one of the most severe complications (11). The time to development of lymphoma varies with the highest incidence occurring after >8 years follow-up (12). Nonetheless, 50-70% of patients do not develop a complicated disease course, resulting in unnecessary costs for specialist care and anxiety for patients. Another dilemma occurs in the event of a clinical suspicion of lymphoma in patients with persistent general swelling of salivary glands or lymph nodes, or an unexplained mass lesion in other organs. A confident histological diagnosis can be difficult to make as eMZL in pSS develops in a background of chronic inflammation. This can result in uncertain circumstances in which a clear diagnosis cannot be reached, especially in small biopsies. Mortality and morbidity is increased in older patients with co-morbidities, with stage III/IV disease and in those with progression into a more aggressive lymphoma (11). Additionally, lymphoma treatment depends on staging and varies from wait-and-see policy, immunotherapy alone or mild chemotherapy in early stages to intensive chemo-immunotherapy in late stages and progression to diffuse large B-cell lymphoma.

Previously reported risk factors for lymphoma development in pSS include a combination of epidemiological, clinical, laboratory and pathological features. These prognostic markers include male gender, permanent parotid enlargement, lymphadenopathy, mixed monoclonal cryoglobulinemia, leukopenia, RF autoantibodies, low complement levels and an extensive lymphocytic infiltrate in a salivary gland biopsy (termed a high focus score) (4, 11, 13, 14). Some of the other risk factors may emerge as a reflection of pre-lymphomatous conditions, particularly parotid enlargement and cryoglobulinemia. Nonetheless, the specificity and sensitivity of these risk factors for lymphoma development is limited, and these features are not restricted to pSS patients developing eMZL. A composite score of lymphoma biomarkers improved specificity at the cost of sensitivity (71.8% sensitivity and 79% specificity at a score of ≥2) (4, 15). Unfortunately, the lack of clarity resulting from the limited specificity and sensitivity of existing markers can result in a significant diagnostic delay for eMZL diagnosis in pSS patients in the order of months to years, alongside a monitoring burden for pSS patients who never progress to eMZL. Altogether, there is a clear need for novel methods for early detection in pSS patients with high risk features for lymphoma.

Previously, we have shown that chronic lymphocytic leukemia (CLL), an indolent subtype of lymphoma, can be detected over 16 years prior to clinical diagnosis by sequencing the B-cell receptor immunoglobulin (BCR IG) gene repertoire in the peripheral blood (16). Our findings showed promise for early detection of lymphoma and for the use of immunogenetic sequencing for precision oncology. We hypothesize that patient groups at increased risk of lymphoma (such as pSS patients) in particular would benefit from a sensitive and specific method of early detection through immunogenetic sequencing.

Hence, the aims of the current study on pSS patients are: 1) to investigate the presence of the eMZL clonotype in tissue biopsies taken prior to eMZL diagnosis. 2) to perform an unbiased study of IG heavy chain (IGH) gene repertoire dynamics in the peripheral blood prior to eMZL and 3) to validate the presence of a dominant clonotype using the eMZL clonotype in the eMZL diagnostic biopsy.

In this study, we sequenced the IG gene repertoire of tissue biopsies and/or the peripheral blood of eight pSS patients who developed eMZL and nine matched pSS controls. Our results indicate eMZL clonotypes are already present in both tissue biopsies and the peripheral blood prior to eMZL diagnosis. In pSS patients, dendritic cells, T-cells and B-cells react strongly to auto-antigens such as the ribonucleoprotein particles Ro/SS-A and La/SS-B expressed by the epithelium of the exocrine glands. They produce high levels of cytokines and chemokines, resulting in chronic inflammation of the exocrine glands and loss of physiological function (25, 26). The salivary gland epithelium plays a central role in this local autoimmune process by actively stimulating immune cells to accumulate, activate and differentiate (27). The inflammatory microenvironment and activated immune cells then create a vicious cycle by activating the epithelial cells and promoting epithelial cell survival, resulting in a perpetual maintenance and escalation of pSS-associated auto-immune responses (27). A particularly important factor in the context of eMZL development is that epithelial cells may be directly involved in B-cell activation. Salivary gland epithelial cells of pSS patients have been shown to produce BAFF mRNA upon Type 1 IFN stimulation (28). BAFF is essential in promoting B lymphocyte activation and survival and has also been associated with autoimmune B-cell activation (27, 29). B-cells in pSS can be found in focal mononuclear infiltrates and between the ductal epithelial cells in the salivary glands (30, 31). Interestingly, the B-cells found in the focal mononuclear infiltrates express markers associated with memory B-cells and plasma cells, while the B-cells in the lympho-epithelial lesions between the ductal epithelial cells express markers of chronic activation and proliferation in absence of classical memory and plasma cell markers (30, 31). This population of chronically expanding B-cells has been described as the putative source of the eMZLs associated with pSS (32, 33).

One of the most impactful and most studied pathways in pSS is the Type 1 interferon (IFN) pathway (34). The IFN receptor IFNAR is expressed on nearly all cells in the body and its downstream mediators affect the transcription of up to 10 percent of all human genes (35, 36). A remarkable and consistent upregulation of IFN stimulated genes (commonly referred to as the IFN signature) has been reported in pSS (37, 38). A higher expression of IFN stimulated genes is associated with a more severe disease phenotype characterized by B-cell hyperactivity, manifesting through a higher focus score, autoantibody seropositivity, hypergammaglobulinemia, complement consumption, reduced saliva secretion and an increased risk of lymphoma development (39).

The susceptibility of RF clones for lymphomagenesis compared to anti-Ro/La clones has been attributed to differences in affinity maturation. The increased incidence of lymphomagenesis among RF clones likely results from differences in the structure of the recognized autoantigens. Compared to anti-Ro/La clones, RF clones express a restricted public set of immunoglobulin variable regions and an IgM constant region, resulting in extensive intraclonal diversification, which may over time result in formation of cryoglobulins and monoclonal lymphoproliferation (6, 8). In the current study, we observed evidence for this intraclonal diversification in the diagnostic tissue biopsies, but not in the eMZL clonotype in the peripheral blood. Additionally, the degree of intraclonal diversification in the eMZL clonotype in the peripheral blood was stable over time to diagnosis (data not shown). This observation would suggest that intraclonal diversification may hold value as a marker for eMZL development in the context of tissue biopsies but not in the peripheral blood, although this finding should be validated in a larger study.

In the context of this chronic auto-immune inflammatory process, it is perhaps not surprising that malignant B-cell clonotypes are already detectable prior to eMZL diagnosis. It remains unclear at what point after the emergence of these clonotypes progression to eMZL occurs and what drivers are involved in progression. To answer this question it will be essential to ascertain which (genetic) driver events lead to the transformation from auto-immune RF-stereotyped B cells to eMZL and which role the local environment of the parotid gland epithelium plays in this process. Further study of the features of the prediagnostic IG gene repertoire may be an effective method of providing clarification on the dynamics of the expansion during eMZL development, but paired experiments evaluating the local environment will likely be required to achieve new insights. Additionally, in the event of the identification of a clonal population in the initial salivary gland biopsy during pSS diagnosis it may be beneficial to monitor the putative eMZL clonotype in the peripheral blood instead, reducing the burden of additional biopsies required for monitoring. Another factor to consider is whether there is a difference in the sensitivity of eMZL detection through immunogenetics when sequencing DNA from labial salivary gland biopsies vs. parotid gland biopsies.

The detection of a Kde clonotype alone without an IGH rearrangement in the tissue biopsy of one pSS patient (pSS3) highlights the benefit of immunogenetic sequencing beyond the IGH gene repertoire, as in this instance the malignant clonotype would have been missed. Potentially, the extensive degree of somatic hypermutation observed in eMZL has impeded amplification of the IGH rearrangement in this patient (24). The IGK-Kde is a IGK rearrangement in which recombination occurs of a recombination signal sequence 24 kb downstream of IGKC gene with a IGKV gene or with the intron between the IGKJ and IGKC genes. In lymphoma, this rearrangement is well recognized to be used as a clonal marker (23, 24).

RFs present with stereotypic combinations of IGHV and IGHJ genes with shared CDR3 sequence motifs (32, 40, 41). Interestingly, the stereotypy seen in RFs is shared with parotid eMZL developed by pSS patients (7, 32, 40). These eMZLs express one of a specific set of stereotyped combinations: IGHV4-59/IGHJ2, IGHV4-59/IGHJ5, IGHV1-69/IGHJ4, IGHV3-7/IGHJ3 (40). These IGH rearrangements are usually accompanied by a IGK light chain rearrangement involving the IGKV3-15 or IGKV3-20 gene (40). In the current study, five out of eight pSS patients developed eMZL expressing IGHV1-69/IGHJ4 or IGHV4-59/IGHJ5 rearrangements, supporting their putative value as prediagnostic markers for eMZL. Not all eMZL in pSS patients present with RF-like features, as described previously. Additionally, we observed a clonotype with RF-like features (IGHV3-7/IGHJ3) in one control (pSS13) at a ratio of 2.8 ( Supplementary Table 3 ). This clonotype could still be detected 1.6 years later, though it was no longer the clonotype with the highest abundance in the repertoire. Another control (pSS17) presented with a clonotype with RF-like features (IGHV1-69/IGHJ4) at a ratio of 3.5. This clonotype had significantly diminished 7 months later (6-fold reduction in abundance). Therefore, a combination of factors should be considered before characterizing a clonotype as potentially (pre)malignant. These include RF-like stereotypic features, consistent elevation of a clonotype over time and the magnitude of the increase in abundance and/or ratio. Characterization of a clonotype as potentially malignant based on measurement in the circulation alone appears prone to false positives. An interesting potential application of our findings would be to perform routine immunogenetic sequencing of the parotid or labial gland biopsy taken at pSS diagnosis. In the event that a (RF-like) clonotype is detected at an abnormal abundance, this clonotype could then be monitored through the peripheral blood, rather than through repeated biopsies, significantly reducing the monitoring burden for the patient, while putatively detecting eMZL at an earlier stage. Integration of immunogenetic sequencing based risk factors with previously reported risk factors for lymphoma in pSS such as parotid enlargement, cryoglobulinemia, leukopenia, detection of rheumatoid factor autoantibodies and low complement levels may further improve risk stratification (4). In our cohort, pSS patients developing eMZL were characterized by a combination of intermittent parotid enlargement, cryoglobulinemic vasculitis, high focal score (≥4) (13), anti-Ro, anti-La and low complement at time of sampling (seven years to three months before eMZL diagnosis), while most of these risk factors (except anti-Ro) were not reported in the controls ( Supplementary Table 4 ).

The main limitation of the current pilot study is its sample size, preventing further investigation into heterogeneous features of both the lymphomas and included pSS patients. The distinct prediagnostic dynamics of the eMZL clonotypes observed in both the tissue biopsies and the peripheral blood indicate a larger cohort is needed to fully understand the implications of the observed prediagnostic eMZL clonotypes, both for our understanding of lymphomagenesis in pSS patients and to evaluate the potential of these clonotypes for the early detection of eMZL.

In conclusion, eMZL clonotypes can be detected prior to eMZL diagnosis in both tissue biopsies and in the peripheral blood of pSS patients. The context-dependent abnormalities we observed in the IGH gene repertoire of blood and tissues in pSS patients have potential implications for early detection of lymphoma risk. While the absolute abundance of the observed eMZL clonotypes in tissues and peripheral blood was not sufficient to readily distinguish them from clonotypic expansions in controls, features such as RF-stereotypy, ongoing SHM and a persistently increased abundance relative to the other clonotypes in the IGH gene repertoire have the potential to aid in risk stratification and warrant further investigation. Furthermore, our findings highlight the potential of pSS-associated RF clones to transform into eMZL, confirming previous reports. In the age of precision oncology, there is a need for high resolution data and insights on targeted patient groups. While our results suggest potential merit for early detection of eMZL development through immunogenetic sequencing, expansion of the cohort size will be an essential step in validating the intriguing findings from the current pilot study.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ncbi.nlm.nih.gov/, GSE221412.

The studies involving human participants were reviewed and approved by Medical Ethics Review Committee Erasmus MC and Central Committee on Research Involving Human Subjects, Netherlands. The patients/participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

PMK and JR performed the experiments. PMK and PJTAG analyzed the data. PMK, EH, MJW, CGH-M, PLAD, ZB, JR, KMH, PJTAG, MAV, RMT and AWL interpreted results. PMK and AQL wrote the manuscript. EH, MJW, CGH-M, PLAD, ZB, JR, KMH, PJTAG, MAV, RMT critically reviewed and edited the manuscript. MAV, RMT and AWL designed and supervised the study. All authors contributed to the article and approved the submitted version.