High frequency of inactivating tetraspanin CD37 mutations in diffuse large B-cell lymphoma at immune-privileged sites

Diffuse large B-cell lymphoma (DLBCL) is the most common type of B-cell non-Hodgkin lymphoma. Several genetic aberrations are known to underlie DLBCL development and progression. Recently, we identified that Cd37-knockout mice spontaneously develop B-cell lymphoma.¹  Moreover, CD37 protein expression has been reported to be an independent prognostic factor for patient outcome in R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone)-treated DLBCL patients.² 

CD37 (Tspan-26), a member of the tetraspanin superfamily, is highly expressed on mature B lymphocytes³  and is currently studied as novel target for chimeric antigen receptor (CAR)-T cell therapy⁴  and antibody-based therapies for patients with relapsed/refractory non-Hodgkin lymphoma.^5,6  Tetraspanins control the spatial distribution of membrane proteins through specific interactions with (immune-)receptors and signaling molecules, thereby facilitating intracellular signaling pathways.^7,8  Tetraspanins are associated with different malignancies,⁹  but the role of altered tetraspanin function in DLBCL pathogenesis is unclear.

This study evaluated the mutational status of the human CD37 gene in DLBCL patients, particularly in DLBCL at immune-privileged sites (IP-DLBCL), including primary testis lymphoma (PTL) and primary central nervous system lymphoma (PCNSL). With functional analysis, the effect of identified CD37 mutations on CD37 protein expression at the cell membrane was studied.

The study was conducted with 137 archived primary DLBCL samples, including 44 IP-DLBCL. The discovery cohort included 31 DLBCL samples and 106 DLBCL samples were analyzed in the validation cohort. Samples were collected from multiple Dutch university hospitals (supplemental Table 1, available on the Blood Web site).

CD37 mutation analysis of the discovery cohort was performed by amplicon-based sequencing with Ion Torrent and Sanger sequencing. For the validation cohort, single-molecule molecular inversion probe mutation analysis and sequencing with Ion Torrent included the CD37 gene and hotspot loci of CARD11, CD79A, CD79B, and MYD88.

The CD37-Gly88Asp-green fluorescent protein (GFP) and CD37-Gly65Glu-GFP mutant constructs were generated by introducing point mutations c.263G>A or c.194G>A in hCD37-wild-type (WT)-GFP¹  using site-directed mutagenesis. Human BJAB and OCI-Ly8 lymphoma B cells were transfected with the AMAXA Nucleofector or the Neon transfection system and analyzed for CD37 protein expression and subcellular localization using western blot and confocal fluorescence microscopy.

Additional experimental details are provided in supplemental Materials and methods.

In a discovery cohort of 31 DLBCL cases, including 5 PTL cases, CD37 mutations were analyzed. CD37 mutations were identified in 2 PTL cases, 1 harboring a missense mutation (c.263G>A (p.Gly88Asp)) and the other a splice-site mutation (c.343-1G>C), but none were identified in the non-IP-DLBCL samples (n = 26). To confirm this finding, a validation cohort consisting of 39 IP-DLBCL cases (18 PTL and 21 PCNSL) and 67 non-IP-DLBCL cases was analyzed for CD37 mutations along with CARD11, CD79A, CD79B, and MYD88 hotspot mutations (Figure 1; supplemental Table 2). In this analysis, we also included the 5 PTL cases from the discovery cohort. Besides confirmation of the 2 initially identified CD37 mutations, we uncovered 12 additional mutations (variant allele frequency 5.7% to 49%; Figure 1B-C). Intriguingly, all CD37 mutations were exclusively present in IP-DLBCL cases (n = 10/44, 23%); and none were present in the non-IP-DLBCL cases (n = 0/93; Fisher’s exact test, P = .0001; Figure 1A). Mutation rates of frequently affected genes CARD11 (5%), CD79A (5%), CD79B (50%), and MYD88 (64%) in our cohort (Figure 1B; supplemental Table 2) were in line with other DLBCL studies.^10,,-13  Whole-genome and exome sequencing studies of DLBCL report low mutation frequencies of CD37, ranging from no mutations to 8%.^11,14,,-17 

The CD37 mutations identified in the validation cohort were detected in 8 IP-DLBCL cases and included 9 missense mutations (c.109C>T (p.Leu37Phe), c.194G>A (p.Gly65Glu) [twice], c.279G>A (p.Met93Ile), c.279G>T (p.Met93Ile), c.460G>A, (p.Gly154Ser), c.656T>G (p.Val219Gly), c.337G>A (p.Ala113Thr), and c.827G>A (p.Arg276Gln)), 1 in-frame deletion (c.364_399del (p.Val122_Asn133del)), and 2 splice-site mutations (c.266_267+10del, c.268-2A>C) (Figure 1C; pathogenicity scores are listed in supplemental Table 2). Within 1 patient, missense mutation c.194G>A (p.Gly65Glu) was detected in 2 primary lymphoma specimens as well as healthy control tissue, but not described as a single-nucleotide polymorphism, indicating the presence of a CD37 germline mutation. To the best of our knowledge, this is the first reported CD37 germline mutation in humans.

To determine the effect of splice-site mutation c.343-1G>C, complementary DNA of this DLBCL sample was sequenced, which revealed an 11-bp deletion (supplemental Figure 1). This resulted in a frameshift in the CD37 mRNA sequence, leading to a completely altered protein after transmembrane domain 3. Based on previous studies on tetraspanin CD81 in B cells,¹⁸  this mutation is predicted to result in a nonfunctional CD37 protein, since the essential extracellular domain 2 is truncated.¹⁹  Next, the CD37 missense mutations and in-frame deletion were mapped onto the reported crystal structure of tetraspanin CD81²⁰  (supplemental Figure 2). Deletion c.364_399del (p.Val122_Asn133del) disrupts an important structural element of extracellular domain 2, namely the A helix, and is therefore also predicted to result in a nonfunctional CD37 protein.¹⁹ 

To further assess whether the observed missense mutations resulted in CD37 loss of function, we introduced 2 of the identified CD37 mutations, c.263G>A (p.Gly88Asp) and c.194G>A (p.Gly65Glu), in human lymphoma B cells using a GFP expression vector. WT CD37 protein appeared as 50- to 75-kDa bands on western blot (Figure 2A-B; supplemental Figure 3A) due to heavy glycosylation. Protein isoforms with the highest glycosylation state (±70 kDa) were less pronounced in lymphoma B cells expressing mutant CD37-Gly88Asp-GFP (Figure 2A), and CD37-Gly65Glu-GFP showed a lower-molecular-weight band on western blot indicating a glycosylation defect (Figure 2B). Total CD37-GFP protein levels of both mutants were similar to CD37-WT-GFP (Figure 2A-B).

Next, the subcellular localization of the CD37 mutants was determined using confocal microscopy. Lymphoma B cells expressing the CD37-GFP mutants had significantly lower cell membrane expression and more retention of intracellular CD37 protein than cells expressing CD37-WT-GFP (Figure 2C; supplemental Figure 3B). Moreover, CD37-Gly88Asp had a dominant-negative effect on the membrane expression of CD37-WT protein (supplemental Figure 4). These data indicate that these missense mutations result in defective glycosylation and trafficking of CD37 in human lymphoma B cells, which corresponds to the reported role of glycosylation as a sorting signal for protein transport to the plasma membrane.²¹  Moreover, modification of cell-surface proteins by defective glycosylation frequently contributes to cancer development by enhancing invasive growth and tumor cell migration.²² 

We previously showed that loss of CD37 results in increased interleukin-6 signaling and STAT3 activation,¹  which are both known to be involved in the pathogenesis of hematological malignancies.²³ MYD88 mutations also stimulate cell survival by promoting JAK/STAT3 signaling and interleukin-6 secretion, suggesting a similar phenotype.²⁴  As proposed for MYD88 and CD79B mutations in IP-DLBCL,¹²  mutations in CD37 could therefore provide a survival advantage in the otherwise stimulus-poor environment of immune-privileged sites.

Seven out of 9 evaluable cases with a CD37 mutation did not express CD37 protein (supplemental Figure 5 and supplemental Table 1). Since immunohistochemistry does not distinguish between cell-surface and cytosolic expression, the importance of CD37 in DLBCL biology^1,2  is likely underestimated. Moreover, the recent identification of loss of CD19 expression as a mechanism underlying disease relapse upon CD19 CAR-T cell therapy²⁵  emphasizes the need to investigate CD37 surface expression in hematologic malignancies prior to the use of novel CD37 CAR-T cell therapies.

In conclusion, inactivating CD37 mutations occur predominantly in IP-DLBCL and lead to aberrant CD37 cell-surface expression. This study provides novel molecular insight into DLBCL at immune-privileged sites and indicates that CD37-based therapies are more likely to be beneficial for DLBCL patients that harbor no CD37 mutations.

The authors thank Piet Gros (Department of Chemistry, Utrecht University) for his valuable help with the modeling studies and Shannon van Lent-van Vliet and Sanne Sweegers (Department of Pathology, Radboudumc) for their valuable help with the single-molecule molecular inversion probe analysis. The authors apologize to all authors whose work could not be cited due to space limitations.

S.E. is supported by a Radboudumc grant. C.M.d.W. is supported by a Rubicon Fellowship from The Netherlands Organization for Scientific Research (019.162LW.004). J.V., A.C., and R.G. are supported by research funding from Stichting Fonds Oncologie Holland. A.B.v.S. is recipient of a Netherlands Organization for Scientific Research grant (NWO-ALW VIDI grant 864.11.006), a Netherlands Organization for Scientific Research Gravitation Programme 2013 grant (ICI-024.002.009), and a Dutch Cancer Society grant (KUN2014-6845) and was awarded an European Research Council Consolidator grant (Secret Surface, 724281).

Contribution: S.E., C.M.d.W., M.v.d.B., S.v.D., B.S., and A.B.v.S. designed the study; S.E., C.M.d.W., M.v.d.B., S.v.D., and A.B.v.S. wrote the manuscript; M.v.d.B., C.J.H., W.S., M.G.M.R., D.d.J., J.S.P.V., A.H.G.C., and B.S. collected DLBCL tissue and corresponding pathological information; S.E., C.M.d.W., M.v.d.B., E.J., A.v.d.S., S.v.D., P.J.T.A.G., A.E., R.A.L.d.G., and M.B. performed mutation analysis; S.E., F.A., L.J., and S.v.D. performed functional in vitro experiments; V.N. performed CD37 modeling studies; and J.H.v.K., S.v.D., B.S., and A.B.v.S. supervised the work.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Annemiek B. van Spriel, Department of Tumor Immunology, Radboud Institute for Molecular Life Sciences, Radboud University Medical Center, Geert Grooteplein-Zuid 26-28, 6525 GA Nijmegen, The Netherlands; e-mail: annemiek.vanspriel@radboudumc.nl.