New Insights into Endogenous Retrovirus-K Transcripts in Amyotrophic Lateral Sclerosis

Abstract

Retroviral reverse transcriptase activity and the increased expression of human endogenous retroviruses (HERVs) are associated with amyotrophic lateral sclerosis (ALS). We were interested in confirming HERVK overexpression in the ALS brain, its use as an accessory diagnostic marker for ALS, and its potential interplay with neuroinflammation. Using qPCR to analyze HERVK expression in peripheral blood mononuclear cells (PBMCs) and in postmortem brain samples from ALS patients, no significant differences were observed between patients and control subjects. By contrast, we report alterations in the expression patterns of specific HERVK copies, especially in the brainstem. Out of 27 HERVK copies sampled, the relative expression of 17 loci was >1.2-fold changed in samples from ALS patients. In particular, the relative expression of two HERVK copies (Chr3-3 and Chr3-5) was significantly different in brainstem samples from ALS patients compared with controls. Further qPCR analysis of inflammation markers in brain samples revealed a significant increase in NLRP3 levels, while TNFA, IL6, and GZMB showed slight decreases. We cannot confirm global HERVK overexpression in ALS, but we can report the ALS-specific overexpression of selected HERVK copies in the ALS brain. Our data are compatible with the requirement for better patient stratification and support the potential importance of particular HERVK copies in ALS.

Keywords: NLRP3; amyotrophic lateral sclerosis; copy-specific expression; endogenous retroviruses; quantitative PCR.

Figure 1

HERVK expression in ALS. Expression analysis of human endogenous retrovirus ENV and POL was carried out via qPCR in cDNA samples from ALS patients (ALS), patients with other neuropathies (ONPs), and controls (NDIs). Results were normalized using RPL19 and GAPDH as reference genes, calculated using the 2^−ΔΔCt method, and represented as the fold expression compared with the mean expression level in the controls. ENV (A) and POL (B) relative expression levels in PBMC samples from 10 ALS patients were compared with NDI samples (n = 10) and ONP samples (n = 10) (Student’s t-test; p = 0.982). Correlations (C,D) between mRNA expression of HERVK ENV and POL genes in the PBMC samples analyzed in (A,B) are represented, and Pearson’s correlation coefficients are indicated. ENV (E,G) and POL (F,H) levels were analyzed in the cortex and brainstem from ALS patients and control individuals (NDI). Cortex samples: ALS, n = 10; NDI, n = 6; brainstem samples: ALS, n = 8; NDI, n = 8.

Figure 2

HERVK copies expressed in ALS PBMCs and the brain. A qPCR that specifically amplifies HERVK cDNA (described in Figure 1) was applied to PBMCs and brain samples taken from ALS patients and healthy controls (Table 1 and Table 2). Products were sequenced via next-generation sequencing (NGS), and the reads were mapped to the HERVK copies in the human genome. The relative frequency (%) of individual HERVK copies was calculated as the number of reads mapping to a particular HERVK locus relative to the total number of mapped reads. This relative frequency is a measure of the relative transcription of each individual HERVK copy. The 27 HERVK copies to which mapping was counted are listed on the right; their genomic location can be found in Table S2. Relative frequencies are indicated in pie charts; color codes for each copy are indicated on the right. Analysis was performed on either the PBMCs of ALS patients (n = 10) and healthy controls (n = 10) (A) or on postmortem samples of the brainstems of ALS patients or non-diseased individuals (NDIs) (n = 4 in each group). Both the overall similarity between copies expressed in ALS vs. controls and the differentially expressed individual copies can be discerned. (A) Pie charts indicate the relative frequency (%) of reads mapping to each of the 27 HERVK copies in PBMC. (B) Same for brainstems.

Figure 3

Altered expression of HERVK copies in ALS brains. (A) Fold change or fold difference between all 27 HERVK loci. The change in each locus (Δfold) was calculated as the mean/average relative frequency (%) of reads mapped to an individual HERVK locus in ALS patients versus NDIs (or the inverse for results < 1). Subsequently, the fold change (ΔΔfold) was calculated for each locus as Δfold-1. The median fold changes over all loci were calculated for PBMCs, the cortex (Cx), and the brainstem (Stem) and are represented in the figure. The corresponding p-values are indicated above. (B) Same analysis as in C for 17 loci showing differences in brains between ALS patients and NDIs. Loci that showed 0.8 < Δfold < 1.2 between ALS patients and controls for both the cortex (Cx) and brainstem were excluded from the analysis. (C) The relative frequency (%) of reads mapped to the individual HERVK copies is indicated. The bottom and top of the box plot are, respectively, the first and third quartiles, and the band inside the box is the second quartile (the median). Lines extending vertically from the boxes indicate variability outside the upper and lower quartiles. Outliers are indicated as individual points. IR: interquartile range. The statistical significance for the difference between ALS and NDI (calculated using the DESeq2 pipeline) is shown in the picture (* adjusted p-value (p-adj.) < 0.05).

Figure 4

Inflammation-related gene expression in ALS brain. Expression analysis of the genes indicated was carried out via qPCR on cDNA samples prepared from the brain samples of ALS patients and controls (NDIs). Results were normalized using GAPDH and RPL19 as reference genes, calculated using the 2^−ΔΔCt method, and are represented as a percentage of the fold expression compared with the median expression level in the controls. NLRP3 expression was analyzed in both cortex and brainstem (A) and in both tissues together (B). ALS patient samples and control (NDI) samples were prepared from either the cerebral cortex or brainstem. Cortex ALS, n = 10; cortex NDI, n = 3; brainstem ALS, n = 9; brainstem NDI, n = 4. Student’s t-test: * p < 0.05, *** p < 0.001.