Metagenome analyses identify human endogenous retrovirus–K113 (HML-2) subtype in glioblastoma

We read with great interest the article by Shah et al., demonstrating human endogenous retrovirus K (HERV-K; HML-2) overexpression in glioblastoma multiforme (GBM) and its role in maintaining stemness and tumorigenesis (1).

This article coincided with our findings from metagenomic analyses on the potential involvement of viruses in GBM etiology. Using and comparing two different bioinformatic approaches, we discovered HERV-K113 in the GBM tissues tested and validated these combined results by PCR and RT-PCR. In analyzing GBM-related samples (15 samples) downloaded from three Sequence Read Archive (15 samples) studies with the centrifuge metagenomics classification tool (2), many reads aligned to HERV-K (Figure 1A, the flow of reads from the left side of the graph through different viral taxonomic units eventually identifies a viral species on the right). Furthermore, compared with our analysis of normal brain samples (3 samples), HERV-K113 viral abundance was only detected in GBM samples. The results of this analysis were compared to our long-read sequences of GBM samples (2 samples), and we analyzed the subsequent metagenomic classification using NCBI BLAST (Figure 1B, BLAST summary of a GBM sample; the graph displays the total read lengths and number of reads for different viruses on the y axis, while the x axis shows the log scale for these variables). To confirm the outcomes of the metagenomic analyses and to evaluate which region of the HERV-K113 might be of importance, PCR primers were designed to target specific areas of HERV-K113, as identified in the alignment results of the BLAST analysis (Figure 1C).

Figure 1

Identification of HERV-K113 in GBM. (A) Centrifuge metagenomics classification analysis. D, dogma; K, kingdom; P, phylum; F, family; G, genus; S, species. (B) BLAST summary. (C) Map of locations of HERV-K113 LR primers. (D) Gel electrophoresis. (E) Conserved domain. The bold letters in the sequence, represent the part of the query sequence that is identical (hit) to the subject (RVT_1).

RT-PCR was performed on the isolated RNA from these samples using long range 1 and 2 (LR1 and LR2) primers, revealing 50% and 60% positivity, respectively. (Figure 1D, results of RT-PCR using LR1 [expected product size, 2,719 bp] and using LR2 primer [expected product size, 2,750 bp]). Sanger sequencing of all PCR products confirmed HERV-K113. PCR products of LR1 revealed a sequence identity of 97% to HERV-K113 — the query sequence clustered with an integrated HERV-K (HML-2) sequence (HML-2_19p12b). Using LR1 primers, the HERV-K113 polyprotein was identified as the longest ORF on the first frame of the positive strand, with which it shared 93.5% sequence identity. Our data suggest that the HERV-K113 sequences are integrated into the GBM genomes and transcribed. Specific hits for the conserved domain of the reverse transcriptase (RVT_1) protein were identified with ORF predictions by long-range PCR products (Figure 1E).