T(11;18)(q21;q21)-positive gastrointestinal MALT lymphomas are heterogeneous with respect to the VH gene mutation status

Abstract

AIM: To investigate how t(11;18)(q21;q21)-positive gastrointestinal MALT lymphomas relate to other marginal zone lymphomas with respect to the somatic mutation pattern of the V_H genes and the expression of the marker CD27.

METHODS: The V_H gene of 7 t(11;18)(q21;q21)-positive gastrointestinal MALT lymphomas was amplified by PCR using family specific V_H primers and a consensus J_H primer. PCR products were sequenced and mutation analysis of the CDR and the FR regions was performed. All cases were immunostained for CD27.

RESULTS: One case showed unmutated V_H genes while the others showed mutated V_H genes with mutation frequencies ranging from 1.3 to 14.7% and with evidence of antigen selection in 2 cases. These data suggest that the translocation t(11;18)(q21;q21) can target either B-cells at different stages of differentiation or naive B-cells that retain the capacity to differentiate upon antigen stimulation. All cases but one displayed weak to strong CD27 expression which did not correlate with the V_H gene mutation status.

CONCLUSION: t(11;18)(q21;q21)-positive gastrointestinal MALT lymphomas are heterogeneous with respect to the V_H mutation status and CD27 is not a marker of somatically mutated B-cells.

Key Words: Gastrointestinal MALT lymphoma; t(11; 18)(q21; q21); V_H mutation; Marginal zone; CD27

INTRODUCTION

Extranodal marginal zone lymphoma or MALT lymphoma is listed as a separate disease entity in the World Health Organisation (WHO) classification of lymphoid tumors, and comprises 8% of all non-Hodgkin lymphomas[1]. These tumors typically arise at sites normally devoid of lymphoid tissue, which have accumulated lymphoid tissue in the context of long-standing antigenic stimulation resulting from causes such as chronic bacterial infection [Helicobacter pylori (H. pylori), Campylobacter jejuni, Chlamydia psittaci, Borrelia Burgdorferi], chronic viral infection (Hepatitis C) or autoimmune disease (Sjögren’s syndrome, Hashimoto thyroiditis)[2]. Several genetic aberrations have been identified in MALT lymphomas, some of which appear to be specific for, or at least closely related to, this type of lymphoma. As such, the chromosomal translocations t(11;18)(q21;q21), t(1;14)(p22;q32), t(14;18)(q32;q21), and t(3;14)(p13;q32) are known to occur with variable frequencies in MALT lymphomas, resulting in IGH-BCL10, IGH-MALT1, API2-MALT1 and IGH-FOXP1 rearrangements respectively^3-5. The oncogenic activity of the 3 translocations t(1;14)(p22;q32), t(14;18)(q32;q21) and t(11;18)(q21;q21) is linked to the physiological role of the BCL10 and MALT1 proteins in the antigen receptor-mediated activation of NF-kB, which is the transcription factor of a number of survival- and proliferation-related genes in B cells[6,7]. The oncogenic role of the IGH-FOXP1 fusion transcript is unknown.

MALT lymphomas may be encountered at virtually any extranodal site of the human body, among which the stomach is the most frequently involved organ. In this particular location, the occurrence of MALT lymphomas is associated with H. pylori infection[8]. A high incidence (about 25%) of t(11;18)(q21;q21) is detected in gastric MALT lymphomas whereas this translocation is rarely found in MALT lymphomas outside the gastrointestinal tract (with the exception of pulmonary MALT lymphomas)[9-11]. Remarkably, the presence of t(11;18)(q21;q21) in MALT lymphomas correlates with the lack of any further genetic instability or chromosomal imbalance[12,13]. T(11;18)(q21;q21)-positive gastric MALT lymphomas do not differ from their t(11;18)(q21;q21)-negative counterparts, with respect to morphology and immunophenotype[3]. However, several studies have revealed that t(11;18)(q21;q21)-positive gastric MALT lymphomas are more often resistant to H. pylori eradication treatment[11,14-16]. Nevertheless, complete lymphoma regression can still be obtained in 20% of these cases after H. pylori eradication[17]. In addition, for a decade, researchers believed that t(11;18)(q21;q21)-positive gastric MALT lymphomas rarely if ever evolve to a more aggressive diffuse large B-cell lymphoma (DLBCL)[11,14-16], however, new data show that the t(11;18)(q21;q21) can be found in both gastric MALT lymphomas and gastric DLBCLs at approximately equivalent frequencies[18]. Although the presence of t(11;18)(q21;q21) may facilitate and/or confirm the diagnosis of a MALT lymphoma, current guidelines do not recommend a routine screening for the t(11;18)(q21;q21) once the diagnosis of a gastric MALT lymphoma has been established[19].

The immunoglobulin heavy chain (IgH) locus on chromosome 14 contains an estimated 100 to 150 variable (V_H) genes, 30 diversity (D) genes, and six junctional (J_H) gene segments. During the maturation of a normal B cell, the unique combination of single germline V_H, D, and J_H gene fragments gives rise to a functional V_H-D-J_H unit[20]. Added diversity occurs through the junctional insertion of nucleotides at the boundaries of these fragments and through somatic hypermutation, where the nucleotide sequences of the germline fragments are altered. Because somatic hypermutation of Ig V_H genes appears to be restricted to B cells proliferating within the microenvironment of the germinal, somatically mutated V-regions are a hallmark of germinal B cells and their descendants[21]. The positive or negative selective pressure of somatic hypermutation in B cell malignancies can pinpoint the developmental stage of the neoplastic cells[22]. Therefore, through positive selective pressure, somatic hypermutations have been found in post-germinal centre B cell lymphomas such as follicular lymphoma[23], but not (or to a much lesser extent) in pregerminal B cell lymphomas, such as mantle cell lymphoma[24]. Lymphomas arising from the marginal zone (e.g. extranodal, nodal and splenic marginal zone lymphomas) seem to carry either unmutated, slightly mutated or highly mutated V_H genes, either with or without evidence of antigen selection[25-40]. These findings indicate that MALT lymphomas are heterogeneous with respect to the B cell differentiation stage at which clonal expansion occurs (being either a naive, early or late memory cell) and that direct antigen stimulation may be involved in the development or growth of the lymphoma.

The presence of somatic hypermutations in B cells has been associated, not only with post-germinal centre status, but also with CD27 expression. CD27 is a type I glycoprotein that is member of the tumor necrosis factor (TNF) receptor family, with unique cysteine-rich motifs. It is considered to be a marker for memory B cells based on the following observations: (a) circulating IgM⁺/IgD⁺/CD27⁺ B cells contain somatic V_H gene mutations in contrast to peripheral IgM⁺/IgD⁺/CD27^- B cells[41,42]; (b) immunoglobulin class switching is more frequent among CD27-positive than CD27-negative B cells[43]; (c) upon activation, CD27-positive B cells secrete antibodies more efficiently than CD27-negative B cells[44]; (d) cord blood B cells do not express CD27, while the percentage of CD27-positive B cells in blood increases with age[45]; and (e) in the presence of stimuli such as IL-10, SAC plus IL-2 and IL4, prompt differentiation into plasma cells occurs in CD27-positive B cells but not in CD27-negative B cells[42,46]. CD27 expression has not been investigated yet in lymphomas that are considered to arise from somatically mutated B cells.

Here, we investigated how the t(11;18)(q21;q21)-positive gastrointestinal MALT lymphoma subtype relates to other marginal zone lymphomas with respect to the somatic mutation pattern of the V_H genes and analyzed 7 t(11;18)(q21;q21)-positive gastrointestinal MALT lymphomas. In addition, these 7 cases were stained for CD27 to correlate expression of this protein with the somatic mutation pattern of the V_H genes.

MATERIALS AND METHODS

Cases

For this study, the number of representative cases documented with both surgical resection specimens and frozen tissue (as required for adequate mutation analysis) was limited. This is due to the current conservative treatment of gastric MALT lymphomas (the majority of gastric MALT lymphomas regress after H. pylori eradication therapy) and the rare occurrence of t(11;18)(q21;q21). As such, only 7 gastrointestinal MALT lymphomas, documented by the presence of the t(11;18)(q21;q21) on RT-PCR, were analysed. All these cases were described in detail in a previous study[3]. Freshly frozen tissue samples were obtained from surgical specimen: 6 gastrectomy specimens (case 1, 2, 3, 4, 5 and 7) and 1 enterectomy specimen (case 6). From case 6, in addition to the DNA derived from the intestinal biopsy, DNA derived from a gastric biopsy, taken after 54 months of follow-up, was analysed. In all 7 cases, the breakpoints within the API2 gene occurred consistently in intron 7 (now annotated as intron 6 according to Ensembl gene ENSG00000023445) whereas the breakpoints within the MALT1 coding DNA were variable and located in intron 4 (case 3 and 4), 7 (case 5, 6 and 7) and 8 (case 1 and 2), resulting respectively in API2/exon7-MALT1/exon5, API2/exon7-MALT1/exon8 and API2/exon7-MALT1/exon9 fusions.

The principle clinical features of the MALT lymphoma case are shown in Table 1. It is important to notice that all diagnoses were made in the late eighties before gastric MALT lymphoma was known to be related to chronic H. pylori infection. However, immunostaining for H. pylori at a later time revealed the presence of this bacterium in 5 of the 6 gastrectomy specimens as well as in the enterectomy specimen. As H. pylori eradication therapy was not known yet at time of diagnosis, surgical resection of the lymphoma site was the first choice of treatment in all patients (total gastrectomy in 6 patients and partial enterectomy in one patient). All cases were morphologically reviewed according to the WHO classification criteria for lymphoid neoplasms[1]. Histological examination was performed on paraffin-embedded tissue using HE staining. Paraffin-embedded and frozen sections were subjected to immunohistochemical stainings using the following markers: CD20, CD3, CD5, IgM, IgD, kappa, lambda, CD10, CD23 and CD27 (Table 2).

Table 1 Clinical details of cases with t(11;18)(q21;q21)-positive gastrointestinal MALT lymphomas.

Case	Age (yr)	Sex	Hp	Site	Stage	Therapy	Follow-up data
1	60	F	+	Stomach	IE	Surg	CR
2	40	M	+	Stomach	II	Surg + RT	CR
3	50	M	+	Stomach	II	Surg + RT	CR
4	52	M	+	Stomach	IE	Surg	CR
5	42	M	+	Stomach	IE	Surg	CR
6	46	M	+	Small intestine	IE	Surg	Relapse in stomach 1 year after initial diagnosis, with PR and SD after (recurrent) therapy (surg, RT, CT, Hp eradication)
7	52	M	-	Stomach	IE	Surg	Relapse in lung 8 years after initial diagnosis with CR after secondary treatment (CT)

Surg: Surgerical resection; RT: Radiotherapy; CT: Chemotherapy; CR: Complete response; PR: Partial response; SD: Stable disease; Hp: Helicobacter pylori.

Table 2 Panel of primary antibodies used in immunohistochemical techniques.

Antibody	Clone	Company	Dilution	Specificity
CD20	L26	Dakocytomation, Denmark	1/100	pan-B cell marker
CD3	-	Dakocytomation, Denmark	1/50	pan-T cell marker
CD5	Leu1	Becton, Dickinson and Company, US	1/5	pan-T cell marker, some B cell lymphomas
CD10	56C6	Novocastra, UK	1/20	GC B cells, GC B cell lymphomas
CD23	1B12	Novocastra, UK	1/10	Mature B cells, dendritic cells
CD27	137B4	Novocastra, UK	1/50	T cells, activated B cells
IgM	R1/69	Dakocytomation, Denmark	1/5000	B cells
IgD	IgD26	Dakocytomation, Denmark	1/10	B cells
Kappa	163-42	Becton, Dickinson and Company, US	1/25	B cells
Lambda	1-155-2	Becton, Dickinson and Company, US	1/100	B cells

PCR amplification of the rearranged IgH gene

For all cases, DNA was extracted from whole frozen tissue sections. PCR for the IgH gene was performed in 6 separate reactions with 6 V_H family specific primers (primers V_H1-FR1, V_H2-FR1, V_H3-FR1, V_H4-FR1, V_H5-FR1, V_H6-FR1) directed to the FR1 region and a common consensus primer directed to the J_H region (primer J_H) (for primer sequences see Table 3). One µL of DNA was used in a final volume of 20 μL containing primers (2 mmol/L), dNTPs (200 μmol/L), commercially prepared PCR buffer (Promega), MgCl₂ (2.5 mmol/L) and Taq Polymerase (2.5 units). Thermocycler conditions were: denaturation at 94°C, annealing at 59°C and extension at 72°C, each for 40 s with a total number of 50 cycles. All PCR products were electrophoresed in a 1.5% agarose gel containing ethidium-bromide. All amplifications were performed in duplicate. Positive and negative controls were included, consisting respectively of Namalwa cell line DNA and dH₂O. In cases where two or more similarly dominant bands were obtained, a second PCR amplification was performed on the respective PCR products (diluted 1/100), as well as on the original DNA, using primer IgH FR3 region (which is internal to the FR1 primers) (Table 3) and primer J_H under the same conditions, allowing amplification of the CDR3 region. The obtained products were electrophoresed on an 8% polyacrylamide gel and visualised with ethidium-bromide. By comparing the lengths of the PCR products, it was established which product of the initial PCR was amplified from the monoclonally rearranged IgH gene.

Table 3 Primer sequences.

Primer	Sequence
VH1-FR1	5’CCTCAGTGAAGTCTCCTGCAAGG 3’
VH2-FR1	5’TCCTGCGCTGGTGAAAGCCACACA 3’
VH3-FR1	5’GGTCCCTGAGACTCTCCTGTGCA 3’
VH4-FR1	5’TCGGAGACCCTGTCCCTCACCTGCA 3’
VH5-FR1	5’GAAAAAGCCCGGGGAGTCTCTGGA 3’
VH6-FR1	5’CCTGTGCCATCTCCGGGGACAGTG 3’
JH	5’ACCTGAGGAGACGGTGACC 3’
IgH FR3	5’CTGTCGACACGGCCGTGTATTACTG 3’

Sequencing of PCR products and sequencing analysis

Sequencing was performed in both directions with the same oligonucleotides used in the PCR amplifications. The nucleotide sequences of the amplification products representative of the IgH rearrangements were compared with the VBASE directory (VBASE Sequence Directory, I.M.; Tomlinson, MRC Centre for Protein Engineering, Cambridge, UK; URL: http://www.mrc-cpe.cam.ac.uk/imt-doc)[47] using the DNAPLOT analysis software. A diversity (D) germline segment was assigned to the longest stretches with the highest nucleotide homology with a minimum of six successive matches or seven matches interrupted by one mismatch.

Somatic mutation analysis

Mutations in the variable region were identified by comparing the nucleotide sequence of each case with the closest germline V_H sequence. Two nucleotide exchanges in one codon resulting in a replacement (R) mutation, was considered as a single mutation. Mutations at the FR1 primer site and at the joining site of the V_H segment were not regarded as mutations. The number of expected R mutations in the complementary determining region (CDR) or framework region (FR) was calculated using the formula: Exp R (CDR or FR) = n× (CDR R_f or Fr R_f) × (CDRrel or FRrel); where n is the total number of observed mutations, CDRrel or FRrel is the relative size of CDRs or FRs and R_f is the inherent replacement frequency to CDRs or FRs sequences. The R_f value was calculated for the amplified region (excluding primer site) of each V_H germline gene used by our cases, according to Chang and Casali[48]. According to the binomial distribution model, the probability that excess or scarcity of R mutations in CDRs or FRs occurred on the basis of chance alone was calculated using the formula: p = {n!/[k! (n - k)!]} × q^k× (1 - q)^n-k; where k is the number of observed mutations in the CDRs or FRs and q is the probability that an R mutation will localise to CDRs or FRs (q = CDRrel × CDR R_f or FR rel × FR R_f)⁴⁸.

RESULTS

Morphology and immunohistochemistry

All 7 gastrointestinal MALT lymphomas demonstrated a similar monomorphic morphology, being a proliferation predominantly composed of centrocyte-like cells mixed with a small number of lymphocyte-like cells and large activated B-cells. The results of the immunohistostainings are summarized in Table 4. Of interest, in all but one cases, weak to strong CD27 expression was present.

Table 4 Immunohistochemical staining data.

Case	CD20	CD3	CD5	CD10	CD23	CD27	IgM	IgD	Igκ/Igλ
1	+	-	-	-	-	-	+	-/+	Igκ
2	+	-	-	-	-	+	+	-	Igλ
3	+	-	-	-	-	++	+	-	Igκ
4	+	-	-	-	-	+	+	-	Igκ
5	+	-	-	-	-	++	+	-	Igλ
6	+	-	-	-	-	++	+	-	Igκ
7	+	-	-	-	-	++	+	-/+	Igκ

PCR amplification of the rearranged IgH gene

A predominant band was obtained in cases 2, 3, 6 and 7. Cases 1, 4 and 5 revealed two similarly dominant bands but PCR of the CDR3 region allowed identification the V_H family involved (Figure 1). PCR analysis thus revealed the use of V_H3 family, V_H4 family and V_H1 family genes in case 4, 2 and 1 respectively (Table 5). The corresponding PCR products were used for sequencing.

Figure 1 Gel electrophoresis of PCR products from case 1, obtained with 6 primer sets amplifying the FR1-J_H region (A) and the FR3-J_H region (B). Experiments were performed in duplicate. For all amplifications, the reverse primer was J_H. A: The forward primers used are indicated on the gel. The agarose gel shows predominant bands with primer sets V_H3-J_H and V_H5-J_H; B: Polyacrylamide gel showing FR3-J_H amplicons of the same length starting from the V_H3-J_H PCR product (lanes 1, 2) and from the initial DNA sample (lanes 5, 6). No monoclonal FR3-J_H product was obtained starting from the V_H5-J_H PCR product (lanes 3, 4). These results indicate that the lymphoma cells use a V_H3 family gene and that the product obtained with the V_H5-J_H set is an unspecific product.

Table 5 Usage of the rearranged V_H, D_H and J_H gene segments and V_H mutation analysis of gastric MALT-type lymphoma, characterised by the presence or absence of the API2-MALT1 fusion abnormality.

Case	API2-MALT1	Germline VH		Identity (%)	Obs R/S CDR	Obs R/S FR	Exp R/S CDR	Exp R/S FR	pCDR	pFR	DH	JH
		Locus	Name
1	+	VH3-30	DP-49	92.4	4/2	5/4	3/1	8/3	0.217	0.062	DH3-10	JH3
2	+	VH3	HHG4	97.7	1/2	0/1	1/0	2/1	0.421	-	NI	JH4
3	+	VH1-69	DP-10	85.3	11/3	11/4	7/2	15/5	0.030	0.040	DH1-1	JH5
4	+	VH4-34	DP-63	98.7	0/0	1/2	1/0	2/1	-	0.352	NI	JH6
5	+	VH4-39	DP-79	98.6	0/0	0/3	1/0	2/1	-	-	NI	JH4
61	+	VH3-33	DP-50	100	0/0	0/0	-	-	-	-	DH1-20	JH6
7	+	VH3-30	DP-49	95.1	5/0	2/2	2/1	5/2	0.030	0.048	DH3-10	JH4

¹Results obtained on the diagnostic intestinal biopsy as well as on a gastric biopsy after 54 mo follow-up; V_H genes: Immunoglobulin heavy chain variable gene; R: Replacement mutation; S: Silent mutation; CDR: Complementarity determining regions; FR: Framework regions; Obs R/S CDR and Obs R/S FR: Number of the observed R and S mutations in the CDR and FR, respectively; Exp R/S CDR and Exp R/S FR: Number of expected R and S mutations in the CDR and FR, respectively; pCDR and pFR: Probability that excess or scarcity of the R mutations in the V_H gene CDR or FR resulted from chance only; P values < 0.05 were considered statistical significant; NI: Not identified. V_H sequences were compared with germline V_H sequences included in the VBASE sequence directory.

Sequence analysis

All VDJ rearrangements were in frame and none contained a stop codon, indicating that all were productive. Results are summarised in Table 5.

Of the 4 cases using a V_H3 family gene, 2 (cases 1 and 7) used a V_H gene most closely related to the V_H3-30 (DP-49) germline gene with a nucleotide similarity of 92.4 and 95.1%, respectively. Other V_H3 family germline genes were the V_H3-33 (DP-50) gene (case 6) and the HHG4 gene (case 2) with identities of 100 and 97.7% respectively. In the one case using a V_H1 family gene (case 3), the V_H segment was closely related to the same germline gene, being the V_H1-69 (DP-10) gene with 85.3% homology. The 2 V_H4 family genes identified showed the highest homology to the V_H4-34 (DP-63) (case 4, 98.7% homology) and the V_H4-39 (DP-79) (case 5, 98.6% homology) germline genes.

D_H segments and J_H segments were assigned in respectively 4 and 7 of 7 cases. A preferential use of the J_H4 segment was observed (3 times). Both cases with involvement of the V_H3-30 germline gene (cases 1 and 7), used the same D_H segment (D_H3-10) but different J_H segments (J_H3 and J_H4 respectively).

The follow-up gastric sample (taken at 54 mo follow-up) of case 6 revealed the same VDJ rearrangement as obtained from the diagnostic sample.

Mutation analysis

In 6 of 7 cases, the V_H segment showed single nucleotide substitutions with respect to the closest germline gene with identities ranging from 85.3% to 98.8% (Table 5). Some of these nucleotide differences may represent polymorphisms. However, sequence polymorphism among V_H genes is generally low and the detected differences probably represent somatic mutations[7]. The V_H segment used by the remaining case (case 6) was 100% identical to the assigned V_H gene. Furthermore, in case 6, the V_H gene amplified from the gastric follow-up sample did not show somatic mutations. The nucleotide sequences of the V_H genes and comparison to the germline sequences for cases 1 and 7, both using the DP-49 germline, are illustrated in Figure 2.

Figure 2 Mutation analysis of the DP-49 gene used in cases 1 and 7, compared to the germline DP-49 gene sequence. CDR regions are indicated. Dashes indicate sequence similarity. Replacement mutations are shown in bold.

The distribution of replacement (R) and silent (S) somatic mutations in each V_H sequence was analysed according to Chang and Casali (Table 5). The P values indicating whether excess or scarcity of the R mutations in the CDR or FR resulted from chance alone were considered statistically significant if below 0.05. For cases 3 and 7, a preferential clustering of R mutations in the CDR regions was observed (P values respectively 0.03 and 0.03) as well as a significantly lower number of R mutations than expected in the FR regions (P values respectively 0.040 and 0.048). This somatic mutation pattern is indicative for positive antigen selection as well as negative selection against R mutations within the FR regions to preserve the structure of the immunoglobulin. No significant clustering of R mutations was observed for the remaining cases with somatic V_H mutations (cases 1, 2 and 4).

DISCUSSION

This is the first study analysing the V_H mutation status in a series of gastrointestinal MALT lymphomas characterised by t(11;18)(q21;q21). One of the investigated cases (case 6) showed unmutated V_H genes while the others showed mutated V_H genes with mutation frequencies ranging from 1.3 to 14.7%. Two of the cases with mutated V_H gene (cases 3 and 7) showed evidence of both positive and negative antigen-driven selection with clustering of replacement mutations within the CDR regions and a lower incidence of replacement mutations than expected in the FR regions. The latter was also observed in a further mutated case (case 1: statistical significance not reached), without evidence for positive selective pressure, and is consistent with a selection to preserve immunoglobulin structure. Three V_H mutated cases (case 2, 4, 5) showed only few differences from the corresponding germline genes, not indicative of antigen mediated selection. The presence of a low number of somatic V_H mutations without evidence of antigen selection may indicate an origin from an early memory B cell which had not yet undergone multiple rounds of antigen selection in the germinal centre[49].

These data show that t(11;18)(q21;q21)-positive gastrointestinal MALT lymphomas may harbour unmutated, slightly or highly mutated V_H genes, indicating that they are composed either of naive B cells or of early or late memory B cells. The fact that at least part of the cases were composed of memory B cells, shows evidence of antigen selection that may have taken place at the onset of the lymphomatous process. This heterogeneity within t(11;18)(q21;q21)-positive gastrointestinal MALT lymphomas is consistent with that reported in similar studies on both gastric and non-gastric MALT lymphomas in which data on the presence of t(11;18)(q21;q21) was not available[26,27,30,32-36,38-40]. In addition, analysis of the V_H mutation status in 2 t(11;18)(q21;q21)-negative MALT lymphomas (data not shown) revealed a mutation frequency of 0% and 4.5% respectively, indicating that, similar to what is observed in t(11;18)(q21;q21)-positive MALT lymphomas, t(11;18)(q21;q21)-negative MALT lymphomas can be derived from both naïve and memory B cells. Thus, the presence or absence of the API2-MALT1 fusion transcript is not related to a particular B cell stage from which the MALT lymphoma originates. Our results are also in line with studies analysing the composition of the reactive marginal zone, which is thought to represent the normal counterpart of marginal zone cell lymphomas. Indeed, mutation analysis of the rearranged V_H genes of reactive marginal zone B cells indicated that the marginal zone is composed of a heterogeneous B cell population, with naive as well as memory B cells[50,51]. A subpopulation of memory B cells present in the normal marginal zone shows a characteristic pattern of somatic mutations indicative of antigen selection and therefore probably participates in the T cell dependent immune reaction[52,53]. It is of interest that 3 distinct pathways for recruitment of B cells in the marginal zone have been demonstrated in animal models: (a) maturation of virgin recirculating B cells, a process that occurs in the absence of T cells and does not require B cell proliferation; (b) colonisation by memory B cells generated in the germinal centres; (c) recruitment of both virgin and memory marginal zone B cells into antibody responses[54].

Our results also show that t(11;18)(q21;q21)-positive gastrointestinal MALT lymphomas predominantly rearrange V_H3 family and to a lesser extent V_H4 and V_H1 family genes, in cases 4, 2 and 1 respectively. This is similar to what was previously observed in several studies on MALT lymphoma[27,30,32-35,39] and one study on nodal marginal zone lymphoma[28]. In addition, it was found that the two t(11;18)(q21;q21)-negative gastric MALT lymphomas rearranged V_H3 and V_H1 family genes respectively (data not shown). The restricted use of V_H genes further suggests that antigen stimulation may play a role in MALT lymphomagenesis. Interestingly, one particular V_H gene was represented twice, i.e.the V_H3-30 (DP-49) gene in cases 1 and 7. In both cases, the V_H3-30 gene was joined with the D3-10 gene and the mutation pattern was indicative for antigen selection. The V_H3-30 gene seems to be frequently involved in auto-antibody production[38,55]. Therefore, our results support the idea that at least some t(11;18)(q21;q21)-positive gastric MALT lymphomas may result from the transformation of auto-reactive B cell clones. This is in line with the observation that gastric MALT lymphoma cells produce immunoglobulins that do recognise various autoantigens[55]. Also, antibodies to gastric epithelial cells are commonly present in serum samples from patients with H. pylori gastritis, and an anti-idiotype antibody to immunoglobulins of a gastric MALT lymphoma cross-reacts specifically with reactive B cells in H. pylori gastritis[56,57]. Another V_H gene, V_H1-69 (DP-10), was also found in two gastric MALT lymphoma cases, one a t(11;18)(q21;q21)-positive case (case 3) and the other a t(11;18)(q21;q21)-negative case (data not shown). The V_H1-69 gene was reported to be used frequently in nodal marginal zone lymphomas, with a similar mutation pattern as observed in our study, and also in salivary gland MALT lymphomas[26,36]. This indicates that nodal marginal zone lymphomas and MALT lymphomas, occurring at different anatomic sites and either with or without the t(11;18)(q21;q21), may be derived from B cells that are driven by similar antigen epitopes. Moreover, Marasca et al. observed an association between the use of the V_H1-69 gene in a series of nodal marginal zone lymphomas and the occurrence of Hepatitis C virus infection. Unfortunately, the Hepatitis C virus infection status of our 2 cases using V_H1-69 gene was not known.

In view of the reported association between CD27 expression in B cells and the presence of somatic hypermutations, we stained all 7 MALT lymphomas for CD27 and correlated these immunohistochemical data with the V_H gene mutation status. No specific correlation between CD27 expression and the mutation status of the V_H gene was found, as strong CD27 expression did occur in the one case (case 6) with an intact germline V_H gene, while there was a lack of CD27 expression in the one case (case 1) that displayed the second most diversified V_H gene. This observation is in line with data on mice models that questioned whether CD27 is a marker of memory B cells, since CD27 expression did occur in somatically unmutated B cells and CD27 deficiency did not affect somatic diversification in a CD27 knockout mice model[58]. As such, our study does not support the hypothesis that CD27 is a general marker of somatically mutated B cells and further emphasises that the only definitive marker of memory B cells thus far identified is the presence of somatically mutated immunoglobulins. However, our data on CD27 expression should be interpreted with caution as the precise role of (loss of) CD27 expression in (marginal zone) lymphomas remains to be investigated.

Apart from the evidence for antigen selection of the lymphomatous cells, others have found that ongoing mutation (i.e. intraclonal variation) of the V_H genes indicating that continuing antigen stimulation may play a role in the clonal expansion of the MALT lymphoma B cells[30,32,34,35,38]. Our approach using direct sequencing of DNA extracted from whole tissue sections does not allow the identification of subclones within a lymphomatous proliferation. Nevertheless, from case 6, we were able to analyse DNA from the diagnostic sample as well as DNA from a gastric biopsy taken 54 months later. Case 6 was a t(11;18)(q21;q21)-positive MALT lymphoma diagnosed with large masses in the upper intestine, extending into the stomach. This patient showed evidence of minimal residual disease at all time points of his follow-up, as shown by the detection of the API2-MALT1 fusion transcripts and of the reciprocal genomic fusion. The latter is confirmed by the results of the present study, demonstrating the presence of lymphoma cells in the stomach harbouring the same VDJ rearrangement as used by the initial intestinal lymphoma cells, 4.5 years after the original diagnosis. Moreover, the V_H gene was unmutated in both samples. The lack of V_H somatic mutations in the diagnostic as well as in the follow-up sample, indicates that this t(11;18)(q21;q21)-positive MALT lymphoma is composed of naive B cells that have not yet gone through the germinal centre reaction and, therefore, tumor growth seems not to be stimulated by a T cell dependent antigen response. As shown in Table 1, this case was characterised by relapse in the stomach after one year, with stable disease thereafter (despite intensive therapy). Remarkably, in some B cell non-Hodgkin lymphomas, the lack of somatic mutations has been linked to an unfavourable clinical outcome. As such, patients with chronic lymphocytic leukaemia or mantle cell lymphoma whose tumors have mutated V_H genes have a better prognosis than those with intact germline V_H genes[59-61].

As all cases analysed in this study showed the presence of the t(11;18)(q21;q21), we may speculate on the stage of B cell differentiation at which this genetic abnormality is acquired. Several possibilities can be considered. Firstly, our results may indicate that the t(11;18)(q21;q21) can be acquired at different stages of B cell differentiation.However, the hypothesis that the t(11;18)(q21;q21) may confer a transforming potential to the preceding antigen mediated lymphoid proliferation, is supported only by some of the cases, i.e. those with mutated V_H genes and evidence of antigen selection. Secondly, it may be hypothesised that the acquisition of the t(11;18)(q21;q21) occurs at a naive B cell stage in all cases, representing the primary proliferation stimulus, and that somatic V_H mutations are acquired after the malignant transformation induced by the API2-MALT1 fusion transcript. An additional pathogenic role of antigen stimulation at the origin and/or expansion of the lymphoma, acting synergistic with the API2-MLT fusion protein cannot be excluded. The evidence for (auto-)antigen selection that we found in some of the cases may be supportive of this latter suggestion. However, this hypothesis seems to conflict with the fact that H. pylori reactive T-cells have been shown to drive gastric MALT lymphoma cells, regardless of the presence or absence of the t(11;18)(q21;q21)[62]. Alternatively, the acquisition of somatic V_H mutations in an t(11;18)(q21;q21)-positive MALT lymphoma might represent an epiphenomenon without important significance.

COMMENTS

Background

MALT lymphomas are tumors that arise from the marginal zone compartment of the B-follicle at extranodal sites. Gastric MALT lymphoma accounts for 5% of all gastric neoplasia and at least 50% of all gastric lymphomas, making it the most frequent lymphoma of the gastrointestinal tract. Its pathogenesis is clearly linked with Helicobacter pylori (H. pylori) gastritis and 25% of all gastric MALT lymphomas display the translocation t(11;18)(q21;q21).

Research frontiers

The t(11;18)(q21;q21) results in the expression of the aberrant API2-MALT1 fusion protein. Presence of API2-MALT1 in gastric MALT lymphomas is associated with resistance to H. pylori eradication therapy and with the absence of any further genetic instability or chromosomal imbalances. In this study, the authors investigate how t(11;18)(q21;q21)-positive gastrointestinal MALT lymphomas relate to other marginal zone lymphomas with respect to the somatic mutation pattern of the V_H genes and the expression of the marker CD27.

Innovations and breakthroughs

Reports show that both gastric and non-gastric MALT lymphomas harbour unmutated, slightly or highly mutated V_H genes, indicating that they are composed either of naive B cells or of early or late memory B cells. This is the first study analysing the V_H mutation status in a series of gastrointestinal MALT lymphomas in correlation with the t(11;18)(q21;q21). It was found that the presence or absence of the API2-MALT1 fusion transcript is not related to a particular B cell stage from which the MALT lymphoma originates. In addition, we provide evidence that CD27 is not a marker of somatically mutated B-cells, contrary to what is generally believed.

Applications

By understanding at what B-cell stage the t(11;18)(q21;q21) is acquired, this study may contribute to a better understanding of the pathogenesis of gastrointestinal MALT lymphomas in general and may be of importance to therapeutic strategies that target the API2-MALT1 fusion transcript and/or interfere with the NF-kB pathway.

Terminology

API2 and MALT1 are proteins that play a key role in the antigen receptor-mediated activation of NF-kB, which is the transcription factor of a number of survival- and proliferation-related genes in B cells. Not surprisingly, the formation of an aberrant fusion protein API2-MALT1 will deregulate the NF-kB pathway and result in uncontrolled B-cell proliferation and thus (gastrointestinal MALT) lymphomagenesis.

Peer review

In this work, the author reported the association between the VH gene mutation and T(11;18)(q21;q21)-positive gastric MALT lymphomas. And they suggest that this translocation will target different stage of B-cells. Their observation is interesting, however, the clinical sample size is somewhat small.