Comprehensive analysis of human endogenous retrovirus transcriptional activity in human tissues with a retrovirus-specific microarray

Abstract

Retrovirus-like sequences account for 8 to 9% of the human genome. Among these sequences, about 8,000 pol-containing proviral elements have been identified to date. As part of our ongoing search for active and possibly disease-relevant human endogenous retroviruses (HERVs), we have recently developed an oligonucleotide-based microarray. The assay allows for both the detection and the identification of most known retroviral reverse transcriptase (RT)-related nucleic acids in biological samples. In the present study, we have investigated the transcriptional activity of representative members of 20 HERV families in 19 different normal human tissues. Qualitative evaluation of chip hybridization signals and quantitative analysis by real-time RT-PCR revealed distinct HERV activity in the human tissues under investigation, suggesting that HERV elements are active in human cells in a tissue-specific manner. Most active members of HERV families were found in mRNA prepared from skin, thyroid gland, placenta, and tissues of reproductive organs. In contrast, only few active HERVs were detectable in muscle cells. Human tissues that lack HERV transcription could not be found, confirming that human endogenous retroviruses are permanent components of the human transcriptome. Distinct activity patterns may reflect the characteristics of the regulatory machinery in these cells, e.g., cell type-dependent occurrence of transcriptional regulatory factors.

FIG. 1.

Alignment of false-color chip data sets corresponding to HERV classes I, II, and III transcriptional activity observed in 19 normal human tissues by microarray hybridization. A housekeeping gene panel served as an internal control for mRNA integrity. DNA-free mRNA samples were analyzed by microarray hybridization after Cy3-labeled DNA hybridization probes had been generated by RT-PCR according to the standardized protocol. For data reliability, two replicas of the capture probe set were present on each DNA chip. Assays were carried out at least two to four times. For origins and identities of dots, see Table 2. Transcripts of exogenous retroviruses listed in Table 2 were not detected in any tissue. For validation and quantification, real-time PCR has been performed for a subset of eight tissues, denoted in red letters, and a subset of eight HERV elements, marked by red asterisks.

FIG. 2.

Relative quantification of HERV transcriptional activity by real-time PCR. Transcriptional activity of eight HERV elements [ERV-FRD, HERV-E, HERV-F, HERV-W, HERV-K(HML-2), HERV-K(HML-3), HERV-K(HML-5), and HERV-L, all with dot localization codes shown in Fig. 1] was analyzed in eight different human tissues (brain, skin, liver, mammary gland, ovary, cervix, testes, and thyroid gland). The relative abundance of HERV transcripts in each tissue normalized by HPRT levels is depicted by bars and standard deviations (triple experiments). The results of the corresponding chip hybridization experiments are symbolized by small boxes allowing discrimination between no (white boxes), weak (grey boxes), and strong (black boxes) signals.