Main

ALK-positive large B-cell lymphoma was initially described by Delsol et al1 in 1997 as an uncommon and aggressive subtype of diffuse large B-cell lymphoma with frequent immunoblastic morphology, a plasma cell phenotype and expression of the ALK protein. This protein is detected in most cases with a granular cytoplasmic pattern, which is commonly associated with the rearrangement of ALK with the clathrin gene CTCL, t(2;17).2 However, other genetic partners as well as other patterns of cell localization of the ALK protein have been described.3, 4, 5 In addition to the ALK rearrangements, conventional cytogenetic studies have reported the presence of frequent complex karyotypes with multiple structural and numerical alterations in other chromosomes.2, 6, 7, 8, 9, 10, 11, 12, 13, 14

The oncogenic mechanisms involved in the pathogenesis of human ALK-positive large B-cell lymphomas are not well known. ALK protein is a cytoplasmic receptor with tyrosine kinase activity that is not normally expressed in lymphoid cells. ALK rearrangements upregulate oncogenic fusion proteins in which the ALK fragment contains the catalytic domain and the fused partner provides a dimerization domain that activates the receptor without the need of the ligand.15, 16 ALK activation has the oncogenic potential in different types of cells including B and T lymphocytes, and in transgenic mice, it induces the development of plasma cell tumors.17, 18 The dimerization of ALK leads to the activation of different downstream signaling pathways, one of the most relevant being the STAT3 pathway.19, 20 STAT3 is an important signal transducer that triggers the program of plasma cell differentiation through BLIMP1 activation, XBP1 and IRF4 expression.21 ALK-positive large B-cell lymphomas have a phenotype similar to that of plasma cells, being negative for CD20 and positive for CD138. However, whether these tumors express a complete terminal B-cell differentiation program with the upregulation of BLIMP1 and XBP1 is not well known.22 STAT3 also regulates many proteins involved in tumor cell proliferation including MYC that may have an important role in the pathogenesis of these tumors.23

The aggressive behavior of some B-cell lymphomas has been correlated with the presence of MYC rearrangement. This genetic alteration has been observed frequently in lymphoid neoplasms with plasma cell differentiation, such as plasmablastic plasma cell myelomas and plasmablastic lymphomas.24, 25, 26 On the other hand, MYC protein expression has been recently described in some aggressive B-cell lymphomas, not always related to the presence of MYC gene alterations.27, 28, 29, 30 The presence of MYC alteration and protein expression, as well as the activation of STAT3 in ALK-positive large B-cell lymphomas are not well known.

The aims of this study were to determine whether ALK-positive large B-cell lymphomas express a complete terminal B-cell differentiation program with the upregulation of BLIMP1 and XBP1 and whether MYC genetic alterations and STAT3 activation may play a role in the pathogenesis of these tumors.

Materials and methods

Case Selection

Twelve ALK-positive large B-cell lymphoma cases were retrieved from the files of the Pathology Department of the Hospital Clinic of Barcelona; Netherlands Cancer Institute (Amsterdam, The Netherlands); Laboratorio de Hematopatología (Mendoza, Argentina) and Consultoria em Patologia (Botucatu, Brazil). The patients were predominantly males (92%) with a median age of 33 years (range 22–42 years). Ten patients (83%) presented with nodal disease and two had extranodal presentation: one in the right colon and the other one in the soft tissues of the thigh (Table 1).

Table 1 Clinical and pathological features of ALK+ LBCL
Full size table

For comparison purposes, we also studied 11 plasmablastic lymphomas and 16 diffuse large B-cell lymphomas that were selected because they were of the activated B-cell subtype by gene expression profiling (11 cases) or had been classified immunohistochemically as non-germinal center-like subtype using the Hans’ algorithm (5 cases).31

Immunohistochemistry

Immunohistochemical studies were carried out with a panel of monoclonal and polyclonal antibodies reactive in paraffin-embedded tissue sections using a peroxidase-labeled detection system, standard antigen retrieval protocols and an automated immunostainer (Dako Autostainer, Glostrup, Denmark and Bond-Max, Leica Microsystems, Wetzlar, Germany). The panel of antibodies used included CD20 (clone L26; Dako), CD79a (clone JCB117; Dako), PAX5 (clone 24; BD Biosciences), CD30 (clone Ber-H2; Dako) and EMA (clone E29; Dako). Polyclonal antibodies for IgA, IgG, kappa and lambda were applied (all from Dako). ALK expression was determined using the monoclonal ALK1 antibody (Dako).The terminal B-cell differentiation program was evaluated using antibodies anti-IRF4 (clone MUM1p; Dako) and CD138 (clone MI15; Dako), PRDM1/Blimp1 (clone Sc13206; Santa Cruz Biotechnology) and XBP1 (clone Sc7160; Santa Cruz Biotechnology). The conditions for all these antibodies were the same as described previously.32 A cutoff of 25% was used to interpret the results as positive.

MYC expression was assessed using the monoclonal Y69 antibody (Epitomics), at dilution 1/40, with incubation of the primary antibody for 1 h after antigen retrieval at pH 6 for 30 min. The activation of STAT3 was analyzed using the monoclonal phosphorylated-STAT3 (pSTAT3, phosphorylation on tyrosine 705) antibody (clone D3A7; Cell Signaling Technology). The samples were incubated with this antibody at a dilution of 1:100 for 2 h at 37 °C. MYC and pSTAT3 immunostaining were evaluated in a semiquantitative manner and a normal tonsil tissue was used as the control. A cutoff of 25% was also used to interpret the results as positive.

The presence of Epstein–Barr virus was examined by in situ hybridization to detect Epstein–Barr virus-encoded early nuclear RNAs. The latency-associated nuclear antigen 1 of human herpes virus type 8 was studied by immunohistochemistry using the clone LN53 (Advanced Biotechnologies), both as described earlier.25

Fluorescence In Situ Hybridization

Fluorescence in situ hybridization was performed on 3- to 4-μm-thick sections of formalin-fixed, paraffin-embedded tissues, using break-apart probes (Vysis, Abbott Molecular) specific for ALK (2p13) and MYC (8q24) as described previously.25

Tonsil sections were used as controls. For every tumor and tonsil sample, a minimum of 100 evaluable nuclei were scored. The cutoff values for the interphase FISH analyses were established following the criteria of Ventura et al.33 To detect rearrangements, the cutoff value was 3%. Moreover, the mean number of numerical genetic alterations was evaluated in each case to assess the incidence of the genetic events in the tumors. Gains were considered when three to four copies of the gene studied were identified, whereas more than four copies were considered as amplification. A deletion was considered when one of the alleles lost one of their extremes in >40% of the neoplastic cells.

Results

Pathological Features of ALK-Positive Large B-Cell Lymphomas

The clinical and immunophenotypic features of the 12 ALK-positive large B-cell lymphomas are summarized in Table 1. All cases had a granular cytoplasmic distribution of ALK expression. CD20 was negative in all of them. CD79a, PAX5 and CD30 were detected in isolated cells in 2/9, 1/10 and 2/12 cases, respectively, but none of the tumors had positivity for more than one of these B-cell markers. EMA was expressed in all cases. IgA was positive in 9/12 cases and two of them also expressed IgG. No clear expression of immunoglobulin heavy chains could be demonstrated in three cases. Lambda restriction was detected in 7/12 and kappa restriction in 2/12, and in 3 the study was not contributory. Epstein–Barr virus and/or human herpes virus type 8 were negative in all the cases investigated.

CD138 was positive in all cases examined. IRF4 was positive in 8/11 (73%) cases; two had weak expression in 10–25% of the neoplastic cells and one was negative. BLIMP1 was positive in all cases, most of them with strong expression in >50% of the cells. Similarly, XBP1 was also expressed in 11 of 12, although the number of positive cells was slightly lower than for BLIMP1.

The ALK break-apart probe showed different patterns of rearrangement and alterations in all cases (Table 1 and Figure 1). Seven of 12 (58%) cases showed a split ALK signal consistent with a chromosomal rearrangement (cases 1, 3, 4, 5, 7, 10 and 12). Two of these cases also had extra normal copies of the normal allele (cases 3 and 4). A split signal of both alleles was observed in two other cases (cases 2 and 8) and one of them also showed extra copy signals of normal ALK allele (case 8). Two other cases exhibited isolated copies of 3′ end (telomeric region) (cases 9 and 11), one of them with additional copies of the normal ALK (case 11). One additional case showed isolated extra copies of the 5′ end of ALK (centromeric region) and of the normal allele (case 6). These three cases with deletions of the 3′ or 5′ ends had ALK protein expression with the same granular cytoplasmic pattern as the other cases.

Figure 1
figure 1

Different ALK patterns detected by fluorescence in situ hybridization (FISH). (a) Normal ALK pattern observed in a reactive lymph node; (b) case 1, simple rearrangement; (c) case 3, simple rearrangement and extra copies; (d) case 8, double rearrangement; (e) case 2, double rearrangement and extra copies; (f) case 6, three loci with deletion of the 3′ end and extra copies; (g) case 9, deletion of the 5′ end; (h) case 11, three loci with deletion of the 5′ end and extra copies. Yellow arrow: normal signal of ALK; red arrow: split signal of ALK, 3′ end; and green arrow: split signal of ALK, 5′ end.

Full size image

MYC Gene Alterations and Protein Expression

The genetic alterations and protein expression of MYC in ALK-positive large B-cell lymphomas, plasmablastic lymphomas and diffuse large B-cell lymphomas are summarized in Table 2 and Figure 2. MYC analysis by FISH was successful in the 11 ALK-positive large B-cell lymphomas cases studied. Six cases showed an increased copy number of the gene, 5 gains (cases 4, 5, 10, 11 and 12) and 1 amplification (case 3), whereas 5 cases did not show any genetic alteration. The mean copy number gains per cell was 3.26 (range 2.76–3.87) and the amplified copies 5.84. No evidence of MYC rearrangement was detected in any case.

Table 2 MYC: Genetic alterations by FISH and protein expression by IHC
Full size table
Figure 2
figure 2

Phosphorylated signal transducers and activators of transcription-3 (pSTAT3) and MYC in ALK+ large B-cell lymphoma (LBCL), plasmablastic lymphoma (PBL) and activated B-cell-like (ABC) diffuse large B-cell lymphoma (DLBCL). (a) Case 3, MYC amplified in an ALK+ LBCL; (b) case 3, MYC protein expression in >75% of the cells; (c) case 3, pSTAT3 expression in all tumor cells; (d) case 13, MYC rearranged in a PBL case; (e) case 13, MYC protein expression in 50–75% of the neoplastic cells; (f) case 13, pSTAT3 protein expression in 50–75% of the tumor cells. Immunohistochemistry (IHC) images (b, c, e and f) photographed at × 400. Yellow arrow: normal signal of MYC; red arrow: split signal of MYC, 5′ end; green signal: split signal of MYC, 3′ end.

Full size image

MYC protein was expressed in the 12 cases. Six tumors had MYC positivity in 25–50% of the cells and three of them had MYC gains (cases 10, 11 and 12). Three cases had MYC expression in 50–75% of the neoplastic cells, and two of them had MYC gains (cases 4 and 5). The other three tumors showed MYC expression in >75% of the tumor cells and one of them demonstrated amplification of the gene (case 3). No gene alterations were observed in six cases.

In plasmablastic lymphomas, MYC genetic alterations were observed in 9/11 cases, 4 rearrangements and 5 gains. MYC expression was detected in 7/11 cases. Four cases were positive in 25–50% of the neoplastic cells, two of them showed MYC gains (cases 16 and 19), one had a gene rearrangement and no gene alterations were observed in one. Three tumors were MYC positive in 50–75% and all of them had MYC rearrangement (cases 13, 14 and 18). No MYC expression was observed in three plasmablastic lymphomas with MYC gains (cases 17, 20 and 21) (Table 2).

In diffuse large B-cell lymphomas, MYC genetic alterations were observed in 2/16 cases, one rearrangement and one gain. MYC expression was detected in 7/16 cases. Two cases showed MYC expression in >75% of the neoplastic cells, both without any MYC gene alteration (cases 24 and 35). In two cases, MYC expression was observed in 50–75% of the tumor cells, but none of them showed MYC gene alterations by fluorescence in situ hybridization (cases 36 and 37). Three additional cases showed MYC expression in 25–50% of the neoplastic cells (cases 25, 26 and 27) and one of them presented rearrangement of the gene (case 25).

pSTAT3 Expression

pSTAT3 expression in ALK-positive large B-cell lymphomas, plasmablastic and diffuse large B-cell lymphoma-activated B-cell subtype is detailed in Table 2 and Figure 2. All ALK-positive large B-cell lymphomas were strongly positive for pSTAT3. This overexpression was observed in virtually all neoplastic cells, except in one case (case 12).

In plasmablastic lymphomas, pSTAT3 expression was observed in 4/6 cases. Three of them (cases 13, 15 and 19) showed pSTAT3 expression in 50–75% of the neoplastic cells. MYC expression correlated with pSTAT3 expression in these cases. The fourth case (case 21) showed pSTAT3 expression in virtually 100% of the neoplastic cells, but MYC expression was negative. Two cases were negative for pSTAT3, one expressed MYC in 25–50% and the other was negative.

pSTAT3 expression was detected in 4/11 diffuse large B-cell lymphoma cases in 50–75% of the neoplastic cells. Two of these cases (cases 24 and 37) also showed MYC expression in a similar percentage, whereas the other two cases were negative for MYC (cases 28 and 38). None of them had MYC genetic alterations. Three cases without pSTAT3 expression (cases 27, 35 and 36) expressed MYC, but none of them showed genetic alterations of the gene.

Discussion

ALK-positive large B-cell lymphoma is an uncommon subtype of diffuse large B-cell lymphoma carrying ALK rearrangement and protein expression. The most common gene partner of ALK in these tumors is CLTC at 17q232, 8, 14, 34, 35, 36, 37, 38, 39, 40, 41 followed by NPM1 at 5q356, 10 and other genes identified in occasional cases such as SQSTM142 and SEC31A.13 Genetic studies have recognized other translocations between ALK and chromosomal regions at 4q22, 12q24 or Xq21, but the corresponding gene partner has not been identified.11, 12 All these translocations lead to the overexpression and oncogenic activation of the ALK protein. The CLTC-ALK fusion generates a protein recognizable by immunohistochemistry with a cytoplasmic granular pattern. However, this pattern is also detected in other translocations, with the exception of the nuclear and cytoplasmic pattern of the NPM1-ALK fusion. All our cases had ALK expression with a cytoplasmic granular pattern, although the partner involved could not be determined since we used a break-apart ALK probe. A split signal consistent with an ALK translocation was detected in 7 of the 12 cases. Intriguingly, two cases showed only a deletion of the 5′ end of the probe. This FISH pattern has been observed occasionally in ALK-positive large B-cell lymphomas.4, 11, 12, 13 The combination of FISH and molecular studies have confirmed that this pattern is due to cryptic translocations of the 3′ end of ALK fused with different partners and associated with the deletion of the ALK 5′ end. Similar to the reported cases, our two tumors with the 5′ end deletion expressed ALK with the same granular cytoplasmic pattern as the other cases with the typical split signal. In addition to the rearranged gene, our FISH study detected double splits in two cases and extra copies of the normal ALK gene in five cases. These findings are consistent with the multiple chromosomal alterations and karyotypes close to the tetraploidy detected in around 50% of these tumors by conventional cytogenetics.2, 6, 7, 8, 9, 10, 11, 12, 13, 14 However, as we have not used centromer probe in this study, we cannot distinguish between true polysomies or partial chromosomal gains. We also found a case with deletion of 3′ end, which has not been reported before in the literature. This case also expressed ALK protein with the same granular and cytoplasmic pattern of expression, suggesting that ALK may have a cryptic rearrangement with a similar functional effect upregulating ALK protein.

The oncogenic pathways involved in the pathogenesis of ALK-positive large B-cell lymphomas are not well defined. ALK rearrangements produce fusion proteins with constitutive tyrosine kinase activity, which activates several downstream pathways. One of them is the activation of STAT3 signaling that seems to be required for ALK-mediated lymphomagenesis in different cell models including murine and human B and T cells.19, 20, 43 In this study, we have demonstrated that activated pSTAT3 is strongly expressed in all ALK-positive large B-cell lymphomas. These observations expand the previous identification of pSTAT3 in two ALK-positive large B-cell lymphomas44 and suggest that this pathway may be important in the pathogenesis of these lymphomas. The relevance of STAT3 in ALK-positive large B-cell lymphomas is highlighted by its downregulation in cell lines derived from these tumors treated with an ALK inhibitor.7, 13

The strong and constant expression of pSTAT3 in ALK-positive large B-cell lymphomas contrasted with the lower expression observed in non-germinal center B-cell-like diffuse large B-cell lymphomas (36%) and slightly lower in plasmablastic lymphomas (67%). STAT3 has not been previously investigated in plasmablastic lymphoma and our findings suggest that this pathway may be also relevant in these tumors. The expression of pSTAT3 observed in our subset of non-germinal center B-cell-like diffuse large B-cell lymphomas is similar to the frequency observed in previous studies.45, 46 Although the prognostic significance of this finding in diffuse large B-cell lymphoma is controversial,47, 48 studies in cell lines and animal models suggest that STAT3 may be a relevant therapeutic target for this subgroup of diffuse large B-cell lymphomas.45, 46, 49

The high expression of pSTAT3 in ALK-positive large B-cell lymphomas may be relevant for the plasmablastic phenotype of these tumors since STAT3 promotes the upregulation of BLIMP1, a master regulator of the plasma cell differentiation program.21 Previous studies have characterized the phenotype of ALK-positive large B-cell lymphomas with CD138 positivity and lack of mature B-cell markers such as CD20 and CD79a, but the expression of BLIMP1 and other markers of the plasma cell differentiation program was not known. Consistent with the role of pSTAT3, we have demonstrated that ALK-positive large B-cell lymphomas constantly express BLIMP1 and it seems to be functional because the tumors lack PAX5 and express XBP1, two genes, respectively, repressed and upregulated by BLIMP1. PAX5 is essential to maintain the mature B-cell identity and its repression is required for the development of the plasma cell differentiation program, whereas XBP1 is an important transcription factor that modulates the terminal B-cell differentiation program.50, 51 This complete plasmablastic phenotype observed in ALK-positive large B-cell lymphomas is less common in plasmablastic lymphomas. These tumors constantly express BLIMP1 in virtually all cases, but XBP1 is only expressed in approximately 50% of them.52, 53

MYC rearrangements have been observed in 50% of plasmablastic lymphomas and plasmablastic plasma cell myelomas. BLIMP1, constantly expressed in these tumors, represses genes related to cell proliferation and growth including MYC.54 The high frequency of MYC rearrangements in these tumors may be an oncogenic mechanism to overcome the repressor effect of BLIMP1 on MYC expression.25 In this study, we have observed that, contrary to plasmablastic lymphomas, ALK-positive large B-cell lymphomas do not have MYC rearrangements, although one case had MYC amplifications. In spite of the low incidence of genetic alterations, MYC protein was expressed in all tumors, suggesting that it may be important for the pathogenesis of the tumor. Previous studies have shown that MYC is downstream of ALK55 and may also be a target of activated STAT3.56, 57 Therefore, MYC expression in ALK-positive large B-cell lymphomas may be driven by these alternative mechanisms that, as postulated for plasmablastic lymphomas, may lead to overcome the repressor effect of BLIMP1.

In conclusion, our study shows that ALK-positive large B-cell lymphomas express markers associated with a complete plasma cell differentiation program and active STAT3, but contrary to other aggressive neoplasms with plasmablastic phenotype lack MYC rearrangements. These findings are consistent with a model in which ALK rearrangements activate STAT3 that in turn promotes the plasma cell differentiation program through BLIMP1, and upregulates the expression of MYC (Figure 3). The relevance of the ALK/STAT3 pathway in the pathogenesis of these lymphomas may offer an attractive target for new therapies.

Figure 3
figure 3

Model of signal transducers and activators of transcription-3 (STAT3) activation and BLIMP1 and MYC expression in ALK+ large B-cell lymphoma (LBCL). ALK rearranged receptor bound to a non-rearranged ALK receptor (ALK wild type, ALK WT) triggers a STAT3 homodimer formation without the presence of an external ligand. STAT3 dimerization results in the phosphorylation of its tyrosine residues. This activation allows the entrance of the STAT3 dimer into the nucleus of the cell, where it promotes the activation of BLIMP1, which jointly with interferon regulatory factor-4 (IRF4) and X-box binding protein 1 (XBP1) trigger the plasma cell differentiation, and the activation of MYC in the cell proliferation.

Full size image