Main

Extranodal marginal zone lymphomas of mucosa-associated lymphoid tissue type (MALT lymphomas) are typically indolent B-cell neoplasms that develop out of acquired reactive infiltrates associated with localized long-standing autoimmune diseases or chronic infections.1, 2 The growth of early MALT lymphomas is thought to be dependent on antigen stimulation but may become antigen independent over time with the acquisition of additional genetic abnormalities.1, 3 Antigen stimulation of MALT lymphoma growth is strongly supported by studies of gastric MALT lymphomas that develop out of infiltrates associated with chronic Helicobacter pylori-associated infections and can often be cured with H. pylori-eradicating antibiotics.4, 5, 6 Studies by an Italian group have suggested Chlamydia psittaci may have a similar function in ocular adnexal MALT lymphoma and reported tumor regressions in patients receiving C. psittaci-eradicating antibiotic treatment.7, 8 However, studies of US-based cases have not found evidence for C. psittaci in ocular adnexal MALT lymphomas using sensitive PCR techniques.9, 10

Analysis of expressed immunoglobulin heavy chain variable region (VH) genes can provide strong support for surface immunoglobulin-mediated direct antigen stimulation of lymphoma growth without prior identification of the reactive MALT antigenic trigger.11, 12, 13, 14, 15 For example, studies of salivary gland MALT lymphoma VH genes have demonstrated that approximately 60% of cases use a single VH gene segment, V1-69, which is suggestive of antigen selection because there are approximately 40 different functional VH gene segments that can potentially be used.16 In addition, antigen selection of salivary gland MALT lymphoma growth is further supported from finding nonrandom distributions of point mutations typically present in the expressed VH genes and stretches of similar CDR3 amino-acid sequences in VH genes from different patients.12, 16 Evidence supporting antigen stimulation of low-grade B-cell neoplasm growth is not limited to MALT lymphomas as recent studies of chronic lymphocytic leukemias (CLLs) have documented nearly identical expressed heavy and light chain variable genes from different patients, suggesting that some of these cases recognize the same antigens or epitopes.14, 17, 18, 19

In this study, we identified and analyzed the VH genes used by 10 US-based ocular adnexal MALT lymphoma cases to look for evidence of antigen stimulation. Our findings provide support that direct antigen stimulation by a common antigen is important in the growth of these neoplasms that are not associated with C. psittaci.

Materials and methods

Cases

Cases of MALT lymphoma involving the ocular adnexa as defined by the WHO2 with available frozen tissue were identified by searching the pathology files of the University of Pittsburgh Medical Center and the Cleveland Clinic. Of the identified 13 cases, 10 were included in our previous study of ocular MALT lymphoma10 and 3 are new to this study. The hematoxylin-and-eosin-stained slides were reviewed on all cases. Immunohistochemistry and flow cytometry studies (done on all cases) preformed as described10 were also reviewed. Two independent real-time nested PCR assays for detection of C. psittaci were previously performed on 7 of the cases as described.10 The research use of these specimens was approved by the institutional review boards of the University of Pittsburgh (no. 0505102), the Cleveland Clinic (no. 81008), and the University of Utah (no. 13172).

VH Gene Analysis

Genomic DNA was prepared from frozen tissues using the Gentra Puregene kit (Qiagen Inc., Valencia, CA, USA) or the Genomic DNA Purification kit (Gentra Systems, Minneapolis, MN, USA) following the manufacturer's directions. Rearranged immunoglobulin heavy chain variable regions were amplified from the isolated DNA using a reverse 3′ JH primer and seven different 5′ VH leader region primers, six of which were previously described,20 with an additional primer that perfectly matches the leader sequence of the VH3-21 gene segment, 5′-CCATGGAACTGGGGCTCCGC. Genomic DNA (20 ng) was amplified in 1 × GoTaq flexi PCR buffer (Promega Corp., Madison, WI, USA), 3 mM MgCl2, 0.2 mM each dNTP, 0.2 μM leader prime, 0.2 μM J primer, and 1 U GoTaq Flexi DNA polymerase in a final volume of 20 μl. Cycling conditions were 2 min at 94°C followed by 35 cycles of 20 s at 94°C, 10 s at 55°C, 30 s at 72°C followed by a hold 2 min at 72°C and cool down. The PCR reactions were run on a 2% agarose gel and the DNA was visualized with ethidium bromide (0.5 μg/ml). Before sequencing, primers and dNTP were removed from the reactions with the most prominent bands by mixing 10 μl of reaction with 2 μl ExoSAP-IT (USB Corp., Cleveland, OH, USA) and incubating at 37°C for 45 min followed by 15 min at 85°C. The treated samples were appropriately diluted in water and 6 μl was mixed with 8 μl of the appropriate leader primer (0.8 μ M) and subjected to DNA sequence analysis using BigDye Terminator chemistry (Applied Biosystems Inc., Foster City, CA, USA). The readable DNA sequences were searched against the IMGT database (http://imgt.cines.fr/) to identify the most closely related VH, DH, and JH germ-line segments and to confirm a productive VDJ rearrangement.

Results

Of the 13 ocular adnexal MALT lymphoma cases initially identified with available frozen tissue, the expressed VH genes were successfully obtained from the 10 cases further described in Table 1. The patients mean age was 60 years, range 43–82 years, and included 6 men and 4 women. Evaluated biopsies were from orbital soft tissue in nine cases, the lacrimal gland in two cases, with one case (no. 9) having both orbit and lacrimal gland biopsies. All showed morphological features typical of orbital MALT lymphoma, namely a predominance of small lymphocytes admixed with variable numbers of monocytoid appearing cells and sometimes smaller numbers of plasma cells (Figure 1). Flow cytometry that was performed in all cases revealed populations of CD5-negative, CD10-negative B cells with monotypic expression of κ (8 of 10 cases) or λ (2 of 10 cases) light chains. Sensitive PCR testing for C. psittaci was negative in all seven cases evaluated. Only one patient (no. 4) had a prior history of lymphoma, reportedly a diffuse large-cell lymphoma of the small bowel treated with R-CHOP 2 years before the orbit biopsy, and 9 months later a cutaneous low-grade lymphoma with a follicular growth pattern treated with radiation. Slides were not available for review from either biopsy, but the patient was clinically thought to represent a stage 1 orbital MALT lymphoma and received only radiation treatment.

Table 1 Ocular MALT lymphoma case information
Full size table
Figure 1
figure 1

Hematoxylin-and-eosin-stained section of a representative case (no. 4) showing morphological features typical of ocular adnexal MALT lymphoma.

Full size image

A single rearranged functional VH gene without stop codons was identified in the 10 cases by directly sequencing PCR products generated by amplifying lymphoma DNA with VH leader and JH primers. Comparing the lymphoma VH genes to germ-line genes in the IMGT database revealed that two used VH gene segments from the VH1 family, four used VH4 family gene segments, three used VH3 family gene segments, and one used a VH5 family segment (Table 2). Interestingly, the two VH1 family cases used the same V1-2 gene segment, whereas three of the four VH4 family cases used the same V4-34 gene segment. The three VH3 family cases all used different V3 gene segments.

Table 2 Ocular MALT lymphomas use of VH gene segments
Full size table

All of the lymphoma VH gene segments showed significant numbers of point mutations from the most closely related germ-line counterparts, with percent homologies ranging between 96.9 and 89.0%, mean percent homology 93.4±2.4%. The mutation locations, complementarity determining regions (CDRs) or framework regions (FWRs), and type, those causing an amino-acid replacement (R) or not, ie silent (S), are listed in Table 3. All of the lymphoma VH genes segments had fewer replacement mutations in the FWRs than would be expected by chance alone,21 ie random mutation, consistent with the maintenance of immunoglobulin functionality.22 Five cases with higher numbers of mutations (no. 2, 4, 5, 8, and 10) also had fewer replacement mutations than would be expected by chance alone in the CDRs, suggestive of negative selection or selection against replacement mutations in these areas. In only one case (no. 1) were mutations above what would be expected by chance alone in the CDRs that could be suggestive of positive selection.

Table 3 Distribution of mutations in lymphoma VH segments
Full size table

The CDR3 regions of the lymphoma VH genes appeared to be encoded by a variety of different D segments and to have variable sizes (Table 4). Most used either J4 or J6 joining segments, which are also used in the majority of normal B-cells.23 No conserved amino-acid motifs in the CDR3 regions were appreciated among the different cases.

Table 4 Ocular lymphoma VH gene CDR3 sequences
Full size table

Discussion

The development and early growth of MALT lymphomas appears to be dependent on antigen stimulation, which can be direct, mediated by surface immunoglobulin, and/or indirect and mediated by T cells. The antigens having key functions in MALT lymphomagenesis are mostly unknown, but are likely dependent on the type of tissue the lymphomas arise from and could also vary with different geographical areas. We analyzed the VH genes expressed by 10 cases of ocular adnexal MALT lymphoma to look for evidence of direct antigen stimulation. An important finding along this line was observing preferential use of certain VH gene segments with two cases using the same V1-2 gene segment and three other cases using the same V4-34 VH gene segment. Because there are approximately 40 different functional VH gene segments that can potentially be used,23, 24 the probability that 2 or 3 of the 10 cases evaluated would use the same VH gene segments by chance alone is 0.023 and 0.002, respectively. By contrast, approximately 3% of normal peripheral blood B cells use V1-2 and V4-34, which is the expected frequency if use is random.23 Of the other 10 cases, 1 expressed a VH4 family gene segment different from V4-34, V4-31, which further suggests orbital MALT lymphomas may also display preferential use of VH4 family gene segments.

Nonrandom or preferential use of certain VH gene segments suggests that these encode important determinants that have been selected by an antigen binding to surface immunoglobulin leading to stimulation of B-cell growth. Besides our study, preferential use of V4-34 and V1-2 has been reported for other types of low-grade B-cell neoplasms. The V4-34 gene segment is used by approximately 20% of CLL cases with mutated VH genes,25 and by approximately 60% of cases of primary central nervous system lymphoma26, 27 (Table 5). Preferential use of V1-2 is seen in splenic marginal zone lymphomas and used by approximately 40% of cases.28, 29 It is interesting that the expression of V1-2 and V4-34 does not appear to be increased above random use levels in MALT lymphomas that occur at other sites such as the stomach, lung, or salivary gland.12, 30, 31 This is consistent with the concept that antigens stimulating MALT lymphoma growth in the ocular adnexal will be different than those at other MALT lymphoma sites. It may be significant regarding the antigen specificity of ocular adnexal MALT lymphomas that both V1-2 and V4-34 are used by autoantibodies, and almost all cold agglutinin autoantibodies use V4-34.32

Table 5 Differential use of V1-2 and V4-34 segments
Full size table

Our findings of preferential use of V1-2 and V4-34 differ from the two earlier studies of ocular adnexal MALT lymphoma VH genes. In the study by Coupland et al33 that involved 26 cases, 2 (8%) used V4-34 and only 1 (4%) used V1-2, whereas none of the 12 cases analyzed by Hara et al34 used V1-2 or V4-34. Because the cases analyzed in Coupland's study were European and those in Hara's study were Japanese, it is possible that antigens driving ocular adnexal MALT lymphomagenesis vary depending on geographical location. This possibility is also supported by reports showing an association of C. psittaci with mostly Italian orbital MALT lymphoma cases,7, 8, 35 whereas US-based orbital MALT cases using sensitive PCR techniques appear to be C. psittaci negative.9, 10, 35 Technical considerations could also have complicated positive lymphoma VH gene identification in the study by Coupland et al33 because paraffin-extracted DNA was used in nested PCR reactions involving consensus FW1 and FW2 primers and 65 total cycles of amplification.

All of the lymphoma VH genes identified in this study were mutated relative to their most closely related germ-line counterparts consistent with the proposed post-germinal center B-cell origin. In addition, the level of mutation we observed in the orbital MALT lymphoma VH gene segments (mean percent homology to germ line of 93±2.4%) is also close to the mutation load values reported for MALT lymphomas that occur at other locations.12, 16 Analysis of the point mutation types (replacement or silent) and distribution (FWRs or CDRs) within many of the VH gene segments suggested there is selection against mutations occurring in CDR1 and CDR2, or negative selection. Finding evidence of negative selection in these CDRs was not an artifact related to analyzing small numbers of mutations because the five cases showing clear negative selection also had some of the highest mutation loads. Negative selection observed in the CDRs of these cases is similar to the negative selection that was observed in the FWRs of these cases, which is an indicator of functionality in that many FWR residues cannot withstand replacement mutations if immunoglobulin function is to be maintained.22 Selection against replacement mutations in the CDRs is also a feature of VH genes expressed by MALT lymphomas that occur at other sites such as the salivary gland,36, 37 and could function in preventing the disruption of antigen-mediated signaling that may be detrimental to lymphoma cell survival.38, 39 Negative selection is also seen with antibodies that display reactivity towards autoantigens and may have a function in preventing the emergence of high-affinity variants.40

The CDR3 nucleotide sequence, made up of randomly templated nucleotides and diversity and joining segments that have been shortened by nuclease trimming, is the most variable part of a VH gene and serves as a clonal marker.41 Because of the tremendous variability, antibodies that recognize the same antigens typically do not have similar CDR3 sequences.42, 43, 44 As such, not finding similarities of CDR3 sequences among our MALT lymphoma cases does not alter the possibility that some may recognize similar or common antigens. Because the size of the CDR3 sequence also has an important function in antigen binding,45 it may be significant that the CDR3 of two cases using V4-34 are the same size, whereas the CDR3 sequences for the two V1-2 cases differ by only two amino acids.

In summary, our study suggests that US-based ocular adnexal MALT lymphomas that are not associated with C. psittaci preferentially express a limited set of VH gene segments not frequently used by MALT lymphomas that occur at other sites. This finding is consistent with these cases recognizing a common antigen that is responsible for selecting and promoting the transformation of B cells with the frequently expressed VH gene segments through immunoglobulin-mediated growth stimulation. Although the nature of the antigen is unknown, frequent use of the V4-34 and V1-2 gene segments as well as selection against replacement point mutations occurring in the CDRs, favors these cases recognizing a self-antigen or autoantigen.