Abstract

Abstract 1597

Ocular adnexal mucosa associated lymphoid tissue lymphomas (OAMALTL) are the most common lymphomas of the eye. The potential roles for specific antigens in these lymphomas are still controversial. Previously we examined the usage and mutations of the IGVH in Chlamydophila (C) psittaci-negative OAMALTL, demonstrating biased use of the IGHV4 family and IGHV4–34 gene and evidence for antigen selection (PLoS One. 2011;6(12):e29114). However, there have been no previous reports characterizing the Ig light chain (IGKV/IGLV) usage and properties in OAMALTL. Herein, we examined IGKV/IGLV gene usage and mutations in 34 C. psittaci-negative OAMALTL originating from the orbit (15), conjunctivae (14) and lacrimal gland (5).

Clonal potentially functional IGKV/IGLV gene sequences were identified in 30 tumors (18 kappa and 12 lambda). Among the 7 kappa light chain families, IGKV1 (n=5, 28%), IGKV2 (n=1, 6%), IGKV3 (n=8, 44%) and IGKV4 (n=4, 22%) were observed. Among the 8 lambda chain families, IGLV2 (n=5, 42%), IGLV3 (n=1, 8%), IGLV6 (n=1, 8%), IGLV7 (n=3, 25%) and IGLV8 (n=2, 17%) were detected. In comparison with the IGKV and IGLV repertoire employed in normal peripheral blood B-lymphocytes, an overrepresentation of the IGKV4 family (P<0.01) was observed in the analyzed cohort of tumors. No statistically significant difference in the use of other kappa and lambda light chain families was observed. In the most commonly used IGKV3 family, only 2 (IGKV3–20*01 (n=6) and IGKV3–15*01 (n=2)) of the 7 family members were used by the OAMALTL. The IGKV3–20*01 allele was used at a greater frequency than in normal peripheral blood B-lymphocytes (p=0.02). The observed frequency of the allele in OAMALTL (20%) is higher than the previously reported frequency in other non-MALT lymphoma types. Of the six cases which utilized the IGKV3–20*01 allele, corresponding heavy chains were available for five cases. Three of the IGKV3–20*01 alleles paired with the IGHV4–34 allele, one paired with the IGHV3–23 allele and one paired with the IGHV1–69 allele. Interestingly, each of these three alleles has been previously implicated in autoreactivity. The IGKV4-1*01 allele was observed to pair with IGHV3–74 (n=2), IGHV1–24 (n=1) and IGHV2–5 (n=1), respectively. Of the 30 clonal tumor sequences, 27 displayed mutations from germline. The average percent homology to germline was 96.35%. Most mutated cases (14 out of 27, 52%) exhibited less than a 2% difference from their original germline sequence. Four of the six tumors utilizing the IGKV3–20*01 allele exhibited more than 2% mutation away from germline, with the remaining two sequences unmutated. Three of the four tumors employing IGKV4-1*01 exhibited greater than 2% mutation away from germline. No acquired N-glycosylation sites were observed in the OAMALTL. To analyze for potential selective pressure exerted by antigens, we utilized the BASELINe algorithm, which assesses for both presence and strength of antigen selection. 23 sequences exhibited evidence of selection in the CDR regions, while 26 sequences exhibited evidence of selection in the FR regions. Of the 26 sequences for which paired light and heavy chains were available, 9 heavy chains exhibited strong evidence of selection (8 in FR, 1 in CDR) and 5 light chain exhibited strong evidence of selection (all in FR). Two patients' tumors exhibited strong evidence of selection in both heavy and light chains. A recurrent mutation was observed in the FR2 region of sequences derived from the IGKV3–20*01, with three of six cases showing an acquisition of histidine in place of glutamine at position 39. An analysis of the tumor derived CDR3 regions did not reveal stereotyped sequence determinants and showed no homology to other known bacterial binding antibodies. Our findings demonstrate that C. psittaci negative OAMALTL exhibit biased usage of IGKV families and genes with evidence of antigen selection. Combined with our previous findings, we demonstrate the frequent use and common pairing of IGVH and IGKV genes, which are usually used by autoantibodies, suggesting that the C. psittaci negative OAMALTL encoded immunoglobulins may be directed to autoantigens. Indeed, generated recombinant immunoglobulins derived from multiple OAMALTLs react to self and not to bacterial antigens. These findings and the identity of the autoantigens will be presented at the meeting.

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.