Diffuse large B cell lymphoma derived from nodular lymphocyte predominant Hodgkin lymphoma presents with variable histopathology

Nodular lymphocyte predominant Hodgkin lymphoma (NLPHL) is a rare subtype of Hodgkin lymphoma (HL), which accounts for approximately 5% of all cases [1] and generally presents with early clinical stages and an indolent clinical behavior [2, 3]. Middle aged males account for 75% of cases [3]. Histologically, NLPHL typically consists of large nodules of reactive small B cells with a few intermingled tumor cells, the so called lymphocyte predominant or LP cells. Morphologic variants with a diffuse pattern or an abundant reactive T cell infiltrate mimicking T cell/histiocyte rich B cell lymphoma (THRLBCL) have been described [4, 5]. Gene expression profiling studies have identified only marginal differences between LP cells and tumor cells of THRLBCL [6, 7]. In contrast to Hodgkin and Reed-Sternberg cells of classical HL, LP cells usually show a preserved or only partially downregulated B cell phenotype [6, 8]. They typically exhibit constitutive NF-kappaB activity [6] and active JAK-STAT-signaling [9]. Recently, germline mutations of the NPAT gene have been identified as risk factor for NLPHL in the Finnish population [10] and there are reports about a high familial risk in first degree relatives of NLPHL patients [11, 12].

In the past 20 years, simultaneous or secondary transformation of NLPHL into diffuse large B cell lymphoma (DLBCL) has been described in up to 30% of cases [13‐18]. Several studies have suggested that these patients have a favorable outcome [13, 14, 16, 17, 19]. However, one study did not confirm this observation [15]. The male: female ratio for cases of NLPHL transformed into DLBCL was slightly higher than for primary NLPHL (77 - 81%) [14‐17]. The aim of the present study was to characterize morphology and immunophenotype of DLBCL simultaneously or sequentially derived from NLPHL in 33 cases. A second aim was the comparison of the immunophenotype observed in these particular cases with conventional DLBCL, in order to identify a marker panel, which allows to detect DLBCL, representing a putative transformation from NLPHL, in an independent series of DLBCL.

In the present study we investigated 33 patients with NLPHL and transformation into DLBCL. Although 26 of 33 patients in the present study showed a simultaneous presentation of NLPHL and DLBCL, the clinical data match well with previous studies of secondary transformations of NLPHL into DLBCL [16, 17]. Histopathologic variant patterns (“Non-A/B”) as described by Fan et al. [4] were seen in the NLPHL component in 70% of composite cases investigated in the present study. Histopathologic variants have been associated with advanced stage and an increased risk of relapse [4, 5, 22] and the high frequency in the present study likely points to an increased risk of transformation into DLBCL in these patients.

In the present study, we observed in most composite NLPHL-DLBCL cases, a particular pattern of infiltration with a sharply demarcated sheet-like blast expansion or with a relatively high number of histiocytes admixed with the DLBCL component. The morphology of the nuclei was variable, as previously observed [14]. The most common immunophenotype, differing from conventional DLBCL, was coexpression of EMA and J-chain. However, this immunophenotype was not specific. EMA expression in DLBCL derived from NLPHL was also observed in one previous study [14]. Another remarkable observation was the upregulation of B cell markers as well as BCL2 in the DLBCL component compared to the NLPHL component. Possibly, the slight downregulation of B cell receptor signaling observed in the LP cells of NLPHL [6, 8] is substituted by stimulation from rosetting follicular T helper cells. Since T cell rosetting was never observed in the DLBCL component, these tumor cells must have acquired additional properties to ensure their survival. One of these features may represent enhanced tonic active B cell receptor signaling [23‐25] or the upregulation of BCL2 anti-apoptotic protein [26, 27]. The increased number of tumor cells in “LP type” DLBCL compared to the NLPHL component may result from enhanced anti-apoptotic properties resulting from BCL2 expression. In the present series, p-STAT6 and JAK2 were expressed in both components less frequently than previously observed [9]. Possibly, SOCS1-mutated NLPHL cases with active JAK-STAT-signaling are less likely to transform into DLBCL and may, therefore, have a better prognosis similar to DLBCL with truncating SOCS1 mutations [28]. We also observed upregulation of CD15 and CD10 in some cases of “LP type” DLBCL. CD15 expression has been described in a subset of NLPHL-DLBCL composite lymphomas [14], whereas CD15 expression in NLPHL generally is a rare phenomenon [3, 29, 30]. Upregulation of these markers may be associated with activation of the tumor cells as observed for CD30 expression [31‐33].

Since we also identified four cases with a typical “LP type” infiltration pattern and immunophenotype in a series of conventional DLBCL, we hypothesize that a series of conventional DLBCL cases will include cases with transformation from NLPHL, particularly if diagnoses are made on small core biopsies. Since some of the composite cases showed a downregulation of J-chain and EMA as well as an upregulation of CD10 and BCL2 in the DLBCL component, these cases may be difficult to identify applying current histopathologic methods if only the DLBCL component is sampled. Furthermore, the immunohistochemical stainings, observed to be frequently positive in the present study, were not specific, as was demonstrated in the case of transformed marginal zone lymphoma, which was positive for both EMA and J-chain.

In the present study we observed a heterogeneous immunophenotype of DLBCL derived from NLPHL, which may reflect different mechanisms of transformation. This fact makes these cases difficult to recognize, if the NLPHL component is not sampled. Men with sharply demarcated blast infiltrates with EMA- and/or J-chain expression and abdominal, splenic, or axillary localization are more likely to have “LP type” DLBCL. However, it is important to state, that by applying current immunohistochemical markers, there is no way to specifically recognize these cases unless the coexisting NLPHL component can be identified. Therefore, further characterization of the particular clinical and molecular features of “LP type” DLBCL among conventional DLBCL is warranted, and may lead to identification of better diagnostic markers.