Regional Differences in Lymphomagenesis Identified through Quantification of AICDA-Mediated Subclonal Evolution in Follicular Lymphoma: Do Pesticides Play a Role?

Background:

Environmental risk factors for lymphoma are ill-defined and controversial. Epidemiologic studies have demonstrated an association between pesticide exposure and an increase in non-Hodgkin lymphomas (NHL), particularly those with a 14;18 translocation. However, in the absence of a mechanistic explanation for the effects of pesticide on lymphomagenesis, the data remain correlative and subject to alternative explanations.

Mutagenesis due to activation-induced cytidine deaminase (AICDA) is key to lymphomagenesis, particularly in follicular lymphoma (FL). AICDA exerts physiologic effects by creating mutations in the immunoglobulin heavy locus (IGH) for somatic hypermutation in order to produce highly specific antibodies. AICDA's off-target genomic damage creates "driver" and "passenger" mutations, which can be used to quantify the subclonal evolution of each tumor specimen. Using deep sequencing (10000-fold) with appropriate error-thresholding, our lab previously demonstrated that the number of AICDA-mediated mutations in the 5' UTR of BCL2 is linearly related to the sum of AICDA-mediated mutations at other off-target sites, implying that mutations within the 5'UTR of BCL2 can function as a surrogate marker for the genome-wide aberrant targeting (Spence, J Immunology, 2014). We have also demonstrated that the number of mutations within IGHV is unrelated to the sum of genome-wide aberrant mutations. The sequencing data can thus be used to determine the minimum number of subclonal populations within a specimen as defined by each region (BCL2 and IGHV).

We hypothesized that the number of AICDA-generated subclones would differ in patients with FL living in high and low pesticide use regions, elucidating whether pesticide-associated lymphomagenesis acts through AICDA or an alternative mechanism. We sought to assess the feasibility of demonstrating a difference in extent of AICDA-mediated subclonal evolution in patients with FL residing in regions with high and low pesticide usage by quantifying AICDA-mediated subclones in the 5' UTR of BCL2 and within IGHV.

Methods:

Using a state-wide database of pesticide application organized by ZIP code, we identified 35 patients with FL seen at our institution in Western New York in locales with high (N=19) and low (N=16) pesticide usage and quantified AICDA-mediated subclones in the 5' UTR of BCL2 and within IGHV for these patients' FL specimens. This sample size provided 75-92% power to detect a difference in the number of subclones by pesticide usage, using a 2-sided 0.05 level Wilcoxon test, assuming an effect size of 75-80% probability of more subclones in one pesticide usage group.

Results:

We identified 19 patients from high pesticide use (1000+ gigaliters (GL)) and 16 from low pesticide use ZIP codes (<1000 GL). Pesticide volume use by ZIP code in the low group ranged from 10-664 GL with a median of 263 GL, and in the high group ranged from 1636 to 67,140 GL with a median of 4787 GL.

Patients living in regions of high and low pesticide use were similar in terms of race and ethnicity. The proportion of women was also similar: 43% among high versus 50% among low. Mean age at diagnosis was 62 in the high versus 60 in the low group. Three patients were diagnosed at age < 45 in the high group while no one in the low group was diagnosed at age < 45.

BCL2 subclone count was lower (p = 0.046) among those in high pesticide ZIP codes (median 3.0), compared with those in low pesticide ZIP codes (median 6.5). Although the median IGH subclone count was 278 among those in the high group versus 91.5 in the low group, there was insufficient evidence of an association between IGH subclone count and the level of pesticide usage (p = 0.36).

Conclusions:

The extent of subclonal evolution was significantly different between lymphoma specimens of patients residing in ZIP codes with high versus low pesticide usage. Our data suggest that AICDA-mediated subclonal evolution is negatively correlated with residence in a ZIP code with high pesticide use, raising the question of whether pesticides could induce lymphoma through an alternate mechanism. This observation calls for both corroboration and examination of alternative hypotheses. We demonstrate a feasible approach to assessing these questions in a larger dataset.