Endogenous Retroviruses Transcriptional Modulation After Severe Infection, Trauma and Burn

Abstract

Although human endogenous retroviruses (HERVs) expression is a growing subject of interest, no study focused before on specific endogenous retroviruses loci activation in severely injured patients. Yet, HERV reactivation is observed in immunity compromised settings like some cancers and auto-immune diseases. Our objective was to assess the transcriptional modulation of HERVs in burn, trauma and septic shock patients. We analyzed HERV transcriptome with microarray data from whole blood samples of a burn cohort (n = 30), a trauma cohort (n = 105) and 2 septic shock cohorts (n = 28, n = 51), and healthy volunteers (HV, n = 60). We described expression of the 337 probesets targeting HERV from U133 plus 2.0 microarray in each dataset and then we compared HERVs transcriptional modulation of patients compared to healthy volunteers. Although all 4 cohorts contained critically ill patients, the majority of the 337 HERVs was not expressed (around 74% in mean). Each cohort had differentially expressed probesets in patients compared to HV (from 19 to 46). Strikingly, 5 HERVs were in common in all types of severely injured patients, with 4 being up-modulated in patients. We highlighted co-expressed profiles between HERV and nearby CD55 and CD300LF genes as well as autonomous HERV expression. We suggest an inflammatory-specific HERV transcriptional response, and importantly, we introduce that the HERVs close to immunity-related genes might have a role on its expression.

Keywords: burn; endogenous retroviruses; host response; septic shock; severe inflammatory injuries; transcriptome; trauma.

Figure 1

Heatmap representation of HERVs in three models of injury. Heatmap of the 25% most variant probesets targeting HERVs in the four datasets: burn, trauma, and 2 septic shock cohorts. Probesets are in rows and samples in columns. Samples are annotated (colored bars on the top) by type of samples (HV in pink, patients in cyan) and day after inclusion (blue scaled). Expression levels are color-coded from blue (low expression) to red (high expression). Similar patterns of expression are highlighted through hierarchical clustering of probesets (rows) and samples (columns) with Euclidean distance and complete clustering method. (A) Expression levels in burn patients. (B) Expression levels in trauma patients. (C) Expression levels in septic shock 1 patients. (D) Expression levels in septic shock 2 patients.

Figure 2

Differentially expressed HERVs in severely injured patients. (A) Venn diagram of differentially expressed HERVs for each dataset. (B) Expression profiles of commonly modulated probesets targeting HERVs in the 4 datasets, at D1. (C) Expression profiles of 2 selected probesets targeting HERVs. Boxes are color-coded by cohort. For each graphic, from top to bottom, title contains: probeset name, HERV name and closest gene.

Figure 3

LTR33 HERV and SLC8A1 gene expression. (A) SLC8A1 genomic region, with the position of HERV in green, probeset in dark blue, and PCR designs in purple. (B) Expression levels of specific transcripts by RT-qPCR, as described in A, in HV and patients at D1. Expression levels (copy number / μl) were normalized with reference gene (HPRT1). Boxes are color-coded by cohort. Statistically significant difference with HV is marked by * (Wilcoxon signed rank test, p < 0.05).

Figure 4

LTR101_Mam HERV and NFE4 gene expression. (A) NFE4 genomic region, with the position of HERV in green, of probeset in dark blue, of PCR designs in purple. (B) Expression levels of specific transcripts by RT-qPCR, as described in A, in HV and patients at D1. Expression levels (copy number / μl) were normalized with reference gene (HPRT1). Boxes are color-coded by cohort. Statistically significant difference with HV is marked by * (Wilcoxon signed rank test, p < 0.05).

Figure 5

CD55 associated HERV. (A) CD55 genomic region, with the positions of HERV in green, of probeset in dark blue, of PCR designs in purple. (B) Zoom in genomic region of HERV showing PCR designs in detail. (C) Expression levels of specific transcripts by RT-qPCR, as described in A and B, in HV and patients at D1. Expression levels (copy number/μl) were normalized with reference gene (HPRT1). Boxes are color-coded by cohort. (D) Protein expression levels (MFI), on monocytes (left) and neutrophils (right) from 8 burn patients (red), 11 septic shock patients (blue), and 9 HV (purple). Columns ISO B, ISO SS, and ISO HV correspond to isotypes for burn, septic shock, and HV, respectively. Statistically significant difference with HV is marked by * (Wilcoxon signed rank test, p < 0.05).

Figure 6

CD300LF associated HERV. (A) CD300LF genomic region, with the positions of HERV in green, of probeset in dark blue, of PCR designs in purple. (B) Zoom in genomic region of HERV showing PCR designs in detail. (C) Expression levels of specific transcripts by RT-qPCR, as described in A and B, in HV and patients at D1. Expression levels (copy number/μl) were normalized with reference gene (HPRT1). Boxes are color-coded by cohort. (D) Protein expression levels (MFI), on monocytes (left), and neutrophils (right) from 14 burn patients (red), 11 septic shock patients (blue), and 10 HV (purple). Columns ISO B, ISO SS, and ISO HV correspond to isotypes for burn, septic shock, and HV, respectively. Statistically significant difference with HV is marked by * (Wilcoxon signed rank test, p < 0.05).