Human Endogenous Retrovirus-K HML-2 integration within RASGRF2 is associated with intravenous drug abuse and modulates transcription in a cell-line model

Abstract

HERV-K HML-2 (HK2) has been proliferating in the germ line of humans at least as recently as 250,000 years ago, with some integrations that remain polymorphic in the modern human population. One of the solitary HK2 LTR polymorphic integrations lies between exons 17 and 18 of RASGRF2, a gene that affects dopaminergic activity and is thus related to addiction. Here we show that this antisense HK2 integration (namely RASGRF2-int) is found more frequently in persons who inject drugs compared with the general population. In a Greek HIV-1-positive population (n = 202), we found RASGRF2-int 2.5 times (14 versus 6%) more frequently in patients infected through i.v. drug use compared with other transmission route controls (P = 0.03). Independently, in a United Kingdom-based hepatitis C virus-positive population (n = 184), we found RASGRF2-int 3.6 times (34 versus 9.5%) more frequently in patients infected during chronic drug abuse compared with controls (P < 0.001). We then tested whether RASGRF2-int could be mechanistically responsible for this association by modulating transcription of RASGRF2 We show that the CRISPR/Cas9-mediated insertion of HK2 in HEK293 cells in the exact RASGRF2 intronic position found in the population resulted in significant transcriptional and phenotypic changes. We also explored mechanistic features of other intronic HK2 integrations and show that HK2 LTRs can be responsible for generation of cis-natural antisense transcripts, which could interfere with the transcription of nearby genes. Our findings suggest that RASGRF2-int is a strong candidate for dopaminergic manipulation, and emphasize the importance of accurate mapping of neglected HERV polymorphisms in human genomic studies.

Keywords: HERV-K HML-2; RASGRF2; addiction; endogenous retrovirus; persons who inject drugs.

Fig. 1.

(A) LTR integration screening—PCR design. Primer mapping and product sizes relative to the wild type (Top) and the HK2-LTR (red) integrated allele (Bottom). Exonic and intronic regions are in blue and gray, respectively. (B) Integration screening results of eight (p1 to p8) random patients: Primers LTR_splc_F and dn_intg_R were used in “p,” while primers up_intg_F and dn_intg_R were used in “s” mastermix. Patients 98523 and 99583 are positive (heterozygous) for the integration. (C) Primer sequences used for the LTR integration screening and the confirmation of the editing in HEK293 cells and for the expression assessment of the RASGRF2 exonic junctions. (D) Tabular index for the genotypic interpretation of the PCR products in B. (E) CRISPR/Cas9 editing of the HEK293 cell line to incorporate RASGRF2-int. pcDNA3.1(+) plasmid containing the LTR sequence flanked by 1-kb preintegration site homologous arms (light blue) was used for the homology directed repair (HDR) insertion mechanism. Primers intg_conf_F (mastermix “c”), which avoid the false-positive amplification of the HDR plasmid, and ingr_conf_R (mapping the 5′ LTR splice site) were used in conjunction with primers up_intg_F and dn_intg_R to confirm integration/genotyping of the cell line. (F) Post CRISPR/Cas9 clonal selection and screening/genotyping for the RASGRF2-int HEK293 cells. The gel is annotated according to the scale of the cells in culture over the days posttransfection: 96w/18d (96-well plate, 18 d), 24w/34d (24-well plate, 34 d), and T75f/48d (T75 flask, 48 d). The aneuploidy of the cell line selects for the wild-type alleles versus the edited ones. (G) Differential expression of exons (triplicates of the same clone) upon editing of HEK293 cells: SYBR Green qPCR was used to evaluate the expression levels of exon–exon junctions 2–3, 16–17, 17–18, and 18–19. Error bars represent 95% confidence intervals. The expression of the exons around the integration is reduced by more than >70% (P = 0.01, P = 0.047, and P = 0.002 for 16–17, 17–18, and 18–19, respectively, t test), while the expression of exons 2–3 is increased by more than threefold (P = 0.01, t test) compared with the wild-type HEK293 cells, 18 d posttransfection. The expression levels of the exons revert to normal after 48 d posttransfection. (H) RASGRF2 gene and alternative transcripts. Positioning of HK2 solo-LTR integration, between exons 17 and 18 of the main, 201, transcript.

Fig. 2.

(A) Schematic representation of the pipeline used for the estimation of the expression at the edges of the HK2 integrations in NCCIT cells. LTR (in red) and full-length HK2 elements were used to build a sequence database. RNA-seq HK2-filtered reads were locally realigned against an HK2–host junction database. (B) A scatterplot of counts spanning LTR-start vs. LTR-end edges. (C–E) (Log) Boxplots of mean counts vs. type of element, over LTR-start and LTR-end edges. (F) Active (>5 reads) vs. not active (0 to 5 reads) elements, in combination with their age. (G) The magnitude of activity (more/less than 500 reads) in combination with the age of the elements, over LTR-start and LTR-end edges. NS, not statistically significant. Error bars indicate 95% confidence intervals.