Abstract
Introduction: Human endogenous retroviruses (HERVs) are genomic sequences that result from ancestral germ-line infections by exogenous RVs and therefore are transmitted in a Mendelian fashion. Several lines of genetic evidence suggest the involvement of HERVs in the etiology of autoimmune diseases and cancer. Here, by means of genetic epidemiology, we describe a case-control association study of SNPs located in the proximity of HERVs elements in patients with myeloproliferative neoplasms (MPNs).
Methods: Genomic DNA from two cohorts recruited from 4 university hospitals, in total 376 patients with MPN and 527 control subjects, was isolated from whole blood. 148 SNPs related to 49 HERV loci were tested using the Sequenom platform. To quantify genomic DNA copy number variation, real-time PCR was performed using the LightCycler system. Data was analyzed using the χ2-test. For synergy, we recoded controls and cases as 0 and 1, respectively, and used ANOVA with type III sum of squares. Adjustment for multiple hypothesis testing was done using the Bonferroni correction (pB <0.05).
Results: In cohort 1, analysis of 148 SNPs in 188 MPN patients and 453 controls demonstrated a significant association in MPNs of the two SNPs rs219077 and rs5993426 belonging to HERV-K and HERV-9, respectively. The two SNPs were validated in an additional cohort of 188 MPNs and were again highly significant, rs219077 (OR 1.97 (1.43 - 2.7), combined pB = 0.005) and rs5993426 (OR 3.94 (2.72 - 5.71), combined pB = 9x10-11). Additionally, we report that in patients carrying the SNP risk allele (G), rs219077, qPCR analysis results in significantly higher copy numbers compared to patients not carrying the risk allele.
Discussion and conclusion: About 30 years ago HERV-K particles were reported in megakaryocytes cultured from patients with ET. Surprisingly, since then no studies have focused upon a potential link between virus infection and the development of MPNs. In the present study, we revive the HERV-story in MPNs. Thus, we show that two SNPs, rs5993426 and rs219077, close to HERVs correlate with risk of MPNs supporting a role of these genetics elements in MPN pathogenesis. In our study, rs219077, located on the chromosome 2, associates with the disease and maps in the proximity of the HERV-9 sequence. Additionally, the SNP risk allele (G) in rs219077 results in significantly higher copy numbers. There are two possible explanations for the amplification of rs219077 in MPNs. First, it might be an amplification of a part of chromosome 2, as genomic features of MPNs often include clonal, non-random chromosomal amplifications. Second, it might be due to the formation of a transducing virus from the right LTR of HERV-9, possibly now (after a few rounds of replication) with HERV-K LTRs. The observed amplification of risk allele (G) in MPN may indicate that the recombinant virus pathogenic copy can retrotranspose in PBMCs with an active viral reverse transcriptase and, eventually, integrate through recombination/integration events. rs219077 is also located in genetic proximity to MIR3 SINE elements. Thus, one fascinating possibility is that the putative transducing ERV-9 virus could encode or regulate microRNA. The second marker showing significant association with MPN risk was rs5993426. This marker in the proximity of HERV-K encoding Gag, Pol has been reported to be associated with rheumatoid arthritis. Our genetic data are currently exclusively based on genetic association and cannot as such predict any functional relationship beyond presence or absence of SNPs. Nevertheless, statistical associations are most easily explained by functional relations and it is therefore pertinent to pursue the potential molecular mechanisms mediated by HERVs in MPNs. As an example, the induction and progression of rodent erythroleukemias by Friend spleen-forming virus are mainly due to the ability of proviruses to activate cellular oncogenes or inactivate tumor suppressor genes. In this regard, we have most recently shown downregulation of TP53 in MPNs to be significantly upregulated during treatment with interferon-alpha2. In conclusion, we have for the first time demonstrated a genetic association of MPNs with HERV elements. The association supports a pathogenetic role of these loci in MPNs. Our observations stimulate to further studies on HERVs in the pathogenesis of MPNs and our results need to be confirmed in larger cohorts of patients.
Hasselbalch:Novartis: Research Funding.
