A low-grade B-cell lymphoma with prolymphocytic/paraimmunoblastic proliferation and IRF4 rearrangement

Translocations between IRF4 and immunoglobulin genes, present in myeloma1 and high-grade B-cell lymphomas2 have not been reported for low-grade B-cell neoplasms. We report 3 low-grade B-cell lymphoma cases with IRF4 rearrangement showing characteristic morphological features and immunophenotypes.

While checking our laboratory-developed FISH probes for IRF4 split assay on paraffin sections, we incidentally found IRF4 rearrangement in case 1 (the sample at recurrence; Figure 1A). The original pathological diagnosis was low-grade B-cell lymphoma, unclassified. The cytogenetic analysis record indicated that this tumor harbored a balanced translocation between the IGK and IRF4 genes, t(2;6)(p11.2;p25). We then performed fusion FISH assays and confirmed IRF4-IGK present and IRF4-IGH absent (Figure 1B-E). The primary lesion (biopsied eight years prior to the index sample) also displayed IRF4 rearrangement. A case that had been diagnosed as low-grade B-cell lymphoma, unclassified in the original institute, was presented in a regional pathology conference, ten days after the identification of case 1. The cytomorphological feature of the case was reminiscent of that of case 1 although paraimmunoblasts were more predominant (Figure 1P). The morphology prompted us to perform FISH assays for IRF4, obtaining positive results (Figure 1L-O) (case 2). Lymphoma samples (784 cases including 232 low-grade B-cell lymphomas) were screened by split FISH assay for IRF4. Case 3 was identified in this process, and the original pathology diagnosis was low-grade B-cell lymphoma, unclassified. The cytogenetic analysis record, t(2;6)(p12;p25), supported the presence of IRF4-IGK. Case 1 was also in this cohort and, therefore, 2 of 232 low-grade B-cell lymphomas (0.86%) harbored IRF4 rearrangement in the cohort.

Figure 1.Morphological and histomolecular examination results. Panels A to J and K to T illustrate cases 1 (at recurrence) and 2, respectively. All of the present lesions had positive split FISH assay results for IRF4. The lymphoma cells harbored a yellow signal (wild-type IRF4) and individual red (5′-side of IRF4) and green signals (3′-side of IRF4; A). The split FISH assays for IGK and IGL were positive in cases 1 (B, C, and E) and 2 (L, M, and O), respectively. On the same sections, IRF4 stained blue (D and N) and was observed fusing to the 3′-side of IGK (E, arrowhead) or 3′-side of IGL (O, arrowhead). The infiltration pattern was diffuse, but it was divided by sclerotic bands in case 2 (K, 4× objective). The infiltrate comprised prolymphocytes and occasional paraimmunoblasts and small lymphocytes (F, 60× objective). In case 2, paraimmunoblasts were prominent and neoplastic small lymphocytes were rare (P, 60× objective). Most of the lym-phoma cells expressed MUM1/IRF4 (G and Q, 40× objective) and BCL6 (H and R, 40× objective) simultaneously. IgM was strongly expressed (I and S, 40× objective). The Ki67-labeling index was low (J and T, 40× objective).

Histologically, the infiltrate was basically diffuse, and the lymph node architecture was totally effaced. In case 2, however, broad fibrotic bands divided the lymphoma infiltrate into nodules (Figure 1K). The lymphoma cells in the 4 lesions from the 3 patients comprised prolymphocytes, paraimmunoblasts, and small lymphocytes, varying in proportion between lesions. No proliferation centers or lymph follicles were seen, except for the primary lesion in case 1, in which areas of prolymphocytic and paraimmunoblastic infiltration were confluent with a background of neoplastic small lymphocytes and a few regressed primary follicles without follicular colonization were highlighted by the presence of follicular dendritic cell (FDC) network as determined by anti-CD23 immunohistochemistry. In case 1, the recurring lesion showed a diffuse infiltrate of prolymphocytes and occasional paraimmunoblasts (Figure 1F). In case 2, most tumor cells were paraimmunoblasts admixed with occasional prolymphocytes (Figure 1P). On morphological examination only, this cytomorphological feature, together with fibrotic nodularity (Figure 1K), might lead to a differential diagnosis of grade 3 follicular lymphoma (FL). However, these nodules had no FDC networks. Case 3 comprised prolymphocytes and small-to-medium cells with a small central nucleolus in an irregular nucleus (giving an impression of mantle cell lymphoma, MCL), but it was somewhat indefinite because of admixed T cells and tissue scarcity. The immunohisto-chemical results for the 3 cases indicated CD5 (dim in case 3 by flow cytometry), CD10, CD20, CD23, CD43, CD138, MUM1/IRF4 (Figure 1G and Q), BCL2, BCL6 (Figure 1H and R), IgM (Figure 1I and S), IgD, TdT, and Cyclin D1. LMO2 was negative in cases 1 and 2 (occasional lymphoma cells were positive in the latter) (Online Supplementary Figure S1). The Ki67-labeling index was universally low (approximately 10%; Figure 1J and T). The clinical courses and the results of FISH analyses are summarized in Table 1.

Table 1.Clinicopathological summary

The overall histological impression of the present cases is reminiscent of chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL) with prominent prolymphocytic/paraimmunoblastic infiltration (the tumor-forming subtype of CLL/SLL43 or paraimmunoblastic variant of CLL/SLL85) or tissue involvement of B-cell prolymphocytic leukemia (B-PLL). We examined cases 1 and 2 for 4 chromosomal aberrations, which are commonly observed in CLL/SLL [del(11)(q22.3), +12, del(13)(q14.3), del(17)(p13.1)] by FISH, and LEF1, which is characteristically expressed in CLL/SLL,9 by immunohistochemistry. All were negative in case 2, but 13q14.3 was deleted and LEF1 was weakly expressed in case 1 (Online Supplementary Figure S2), implying that case 1 might be more closely related to CLL/SLL.

In CLL/SLL, small lymphocytes are negative or only faintly positive for MUM1/IRF4 and BCL6 expression, whereas prolymphocytes and paraimmunoblasts are moderately positive for MUM1/IRF4 but negative for BCL6.1110 In contrast, most lymphoma cells in the present cases expressed MUM1/IRF4 and BCL6, an uncommon characteristic of well-known low-grade B-cell lym-phomas. The cases reported as paraimmunoblastic variants of CLL/SLL were positive for CD5, followed an aggressive clinical course, and harbored IGH-CCND1 (5 of 8 examined cases).85 These features suggest that most of the reported cases represented MCL (especially pleomorphic variant). None of the present cases showed any aggressive clinical behavior and they showed IRF4 rearrangement without other translocations commonly observed in B-cell neoplasms including CCND1 rearrangement. Variable expression of CD5, strong expression of IgM, and absence of IgD in the present cases are not typical in CLL/SLL and may be suggestive of the immunophenotype of B-PLL. However, none of our patients showed leukemic involvement of the peripheral blood and IRF4 rearrangement has not been reported in B-PLL. Nodal involvement by B-PLL is rare, and previous histological studies probably included MCL.12 The morphological features, immunophenotype, genetics, and clinical presentation of the present cases seemed incompatible with those of existing subtypes, prompting us to distinguish the present cases as “pro-lymphocytic/ paraimmunoblastic lymphoma (PPL)”. The relationship with other low-grade B-cell neoplasms remains to be clarified.

The question of whether the present cases are ‘precursors’ of the IRF4-rearranged high-grade B-cell lymphomas2 remains unanswered. Cytogenetic analyses were successful in case 1, 46XX,t(2;6)(p11.2;p25) [10]/46XX,t(1;11)(q21;q23)[1]/46XX[9], and case 3, 46XY,t(2;6)(p12;p 25)[1]/46,sl,-Y,-4,-8,-9,-9,add(11)(q23),+5mar[1]/46XY[11]. No evidence of rearrangement in BCL2, BCL6, MYC, and CCND1 was obtained in the present cases, whereas in previous cases,2 7 and 1 of the 19 IRF4-rearranged high-grade B-cell lymphomas showed BCL6 and MYC rearrangements, respectively. PCR for IGH gene rearrangement was successful in cases 1 and 2. The IGHV usage was V3-11*01 and V1-8*01, and identity to germline sequences from CDR1 to FWR3 regions was 100% and 94.7%, respectively, with intraclonal diversity in the latter. These mutation statuses may contrast to those in the IRF4-rearranged high-grade B-cell lymphomas (93.9%-86.1%).2 Light-chain genes were the translocation partners of IRF4 in the present cases and IGH in 17 of the 19 IRF4-rearranged high-grade B-cell lymphoma cases studied by Salaverria et al.2 (P=0.0065, Fisher’s exact test). Most patients with high-grade cases were young,2 indicating that IRF4-rearranged high-grade B-cell lymphomas are likely to be de novo. However, it is interesting that case 15 in the series of Salaverria et al.2 harbored IGK as its IRF4 partner and was a transformed lymphoma.2 The present and previous findings indicate that IRF4 rearrangement may be related to the development of low-grade and primary high-grade lymphomas and that some IRF4-rearranged low-grade lymphomas may progress to high-grade lymphomas.

To clarify the questions presented here, more “PPL” cases should be examined. The prolymphocytic/paraim-munoblastic morphological features and co-expression of BCL6 and MUM1/IRF4 in most lymphoma cells are key indications for confirmatory FISH for IRF4 rearrangement.