In this novel study of HERV gene regulation and protein expression in brain tissue from ALS and non-ALS individuals, we show that there an ALS specific pattern of HERV-K subtypes expressed in neurons within the frontal lobe. HERV-K pol mRNA and RT demonstrate preferential regional expression within brain tissue, whereby prefrontal cortex exhibited the highest levels. For the first time, we demonstrate preferential expression of distinctive HERV-K pol sequences in patients with ALS compared to non-ALS individuals.
Multiple reports have previously demonstrated increased retroviral RT activity in the serum and CSF of patients with ALS compared to matched controls 2–4
. In most of these patients RT was functionally active however, a clinical trial with a HIV protease inhibitor, Indinavir was without clinical benefit 3
. Here we show that there is increased expression of HERV-K pol
transcripts and RT protein in patients with ALS, which may explain the prior observations of RT activity in these patients. Moreover, differential regional HERV-K expression was seen in brain tissue from patients with ALS. Increased amounts of HERV-K transcripts were detected in prefrontal and sensory cortex, areas adjacent to the motor cortex, while the motor cortex itself displayed minimal RNA expression of HERV-K. A limitation of using brain samples in patients with terminal disease was that the neuronal cells producing pol
mRNA were likely degenerated and replaced by glial cells, therefore the transcripts were detectable at lower levels in the motor cortex. Yet, HERV-K RT expression as detected by immunostaining remained robust in the surviving neurons of the motor cortex. Interestingly, RT was detected only in neurons and not in other cell types. This is in contrast to HERV-W which is expressed in glial cells of patients with Multiple Sclerosis 23
and has been associated with macrophage activation 24
. HERV-W gag protein has also been found in neurons of normal brain, but has striking accumulation in axons of patients with Multiple Sclerosis in demylinated regions 25
. While the role of HERV-K proteins in causing neuronal damage is unknown, previous studies have shown that HERV-W can cause inflammation in the brain leading to T cell-dependent hemorrhages in rodents 26
. The env
protein of HERV-W mediates its proinflammatory properties via the toll like receptor 427
. The env
protein of HIV, which is an exogenous human retrovirus, has been shown to be a potent neurotoxin 28
. Therefore, increased RT protein expression in neurons may serve as a marker of ALS-involved brain tissue and may be of pathophysiological significance. It remains to be determined if similar findings can be found in the motor neurons of the brainstem and spinal cord. One possibility is that HERV-K expression precedes tissue atrophy. There is extensive evidence of prefrontal involvement in ALS; upwards of one-third of patients with ALS demonstrate frontal executive deficits 29
. The observation that higher HERV-K expression occurs frequently in prefrontal cortex raises the possibility that RT over-expression may contribute to cognitive deficits attributed to the frontal lobe, although premortem cognitive assessments were not available on the patients studied. Moreover, HERV-K RT expression correlates with increased TDP-43 levels in neurons from frontal cortex tissue of ALS patients, suggesting that RT expression occurs in combination with other abberant cellular processes characteristic of ALS 19, 20
. Although it is possible that the increased levels of these proteins may be an epiphenomenon, it seems unlikely since there is a very strong correlation in the expression of boith genes; and the two proteins colocalize in the same neurons in the motor cortex. This pattern was absent from all the other control groups. TDP-43 has been shown to act as a retroviral restriction factor by repressing the transactivation of HIV 21
. Typical of most retroviral restriction factors, the TDP-43 promoter is likely responsive to interferon and inflammation associated transcription factors, as it contains binding sites for IRF1, IRF3 and NFκB. Hence the possibility that TDP-43 may also play a role in the innate immune response to HERV expression needs to be considered 30
While some expression of HERV-K pol
transcripts was noted in an age-matched group of patients who had prolonged systemic illnesses such as atherosclerotic disease and cancer, the frequency and pattern of protein expression was clearly different compared to patients with ALS. In these patients, the HERV-K transcripts were mostly defective and did not cluster in motor neurons, transcribe into protein or correlate with TDP-43 expression. In contrast, HERV-K pol
transcripts were undectable in patients with Parkinson’s disease who were much older or in a much younger group of individuals who had died acutely without preexisting systemic illnesses. It is possible that systemic diseases, by virtue of generalized immune dysfunction, may lead to non-specific activation of HERV transcripts, as has been shown following macrophage activation 24
and in patients with HIV infection 31
. Since the brain samples were obtained from different brain banks the possibility that some differences may occur to differences in handling of tissues cannot be entirely ruled out. However, this is unlikely since the postmortem intervals were similar in all groups (<24 hrs), they were all stored at −80°C until tested and other control genes could be amplified from all tissues.
Differing frequency of HERV-K pol
mRNA transcripts in ALS versus non-ALS brain tissue may reflect either insertional polymorphism of select HERV-K sequences in the genome 11
or a global epigenetic change leading to cellular expression of a selection of HERV-K sequences not expressed in non-ALS patients 32
. Examination of expressed HERV-K sequences suggests that a proportion is non-coding RNA which is unable to contribute to RT levels. Therefore, the degree of RT protein expression in patients with ALS is not due to non-specific increase in mRNA levels in affected tissue. Select HERV-K sequences must contribute to RT expression, and may exhibit differential enzymatic activity depending on amino acid fidelity to RT active sites. This was further confirmed by detection of RT protein by immunostaining in neurons of patients with ALS. Clearly, the ALS-specific HERV-K mRNA derived from 7q36.1 encodes an RT with amino acid insertions prior to the LPQG active site which may result in altered RT activity. The overall relevance of RT activity within cortical neurons is unknown; however, there is evidence that endogenous RT activity (and retrotransposition) can affect cell proliferation, differentiation and gene expression patterns 33, 34
. Thus, RT expression in cortical neurons of patients with ALS may contribute to neurodegeneration.
The analysis of HERV expression in ALS brain tissue shows that there are specific loci of gene activation. Analysis of these loci suggest that there were ORFs for several HERV-K proteins besides RT. The expression of transcripts from the HERV-K loci in 7q34 and 7q36.1 in patients with ALS was specific for this population and may thus serve as markers of disease. HERV long terminal repeats (LTRs) can act as alternate promoters and enhancers for nearby gene expression 35, 36
, regulating activation of adjacent genes. This may have pathophysiological significance for a gene such as DPP6 in 7q36.2 which has been associated with ALS susceptibility 37
. Moreover, HERV-K 7q34 and 7q36.1 reside within a candidate interval for MND in which the susceptibility genes were not identified 14
. This suggests that cytogenetic identification of HERV transcripts can prove to be a useful tool to identify disease-associated genes and sites of altered transcriptional activity.
The prototypic HERV-K family HML-2 has been previously studied in association with disease, however here we report substantial expression of HML-3 transcripts in cortical brain tissue from patients with ALS. Interestingly, HML-3 is believed to have amplified its presence in the human genome using retrotransposition in cis
, a process that requires functional gag
but not env 38
. It is estimated that there are 150 HML-3 loci within the human genome compared to only 60 HML-2 copies 39, 40
. Yet, HML-2 loci are transcriptional more active than their HML-3 counterparts, largely due to HML-3’s longer period of integration. Phylogenetic analysis of HERV-K transcripts from patients with ALS confirms that at least some of HML-3 loci coding for RT, such as 7q36.1, remain actively transcribed during disease. It remains unclear if recombination of various HERV-K proteins originating from multiple loci may activate cycles of retrotransposition (or reinfection) and result in DNA damage precipitating cell death.
Evidence of RT activity has led to the search for novel retroviruses in ALS with no success 4
. Here, for the first time, we show that HERV-K expression may account for the RT activity in ALS. In contrast to previous studies, we evaluated the expression of all HERV-K families at the mRNA and sequence level. Expression of the HERV-K RT protein, derived from the specific genomic loci 7q34-7q36.1 and HERV-K HML2 and HML3 subtypes, was detected in cortical and motor neurons of patients with ALS. Viral RNA and protein expression strongly correlated with expression of TDP-43, defining a specific phenotype for ALS. This pattern was not seen in patients with chronic systemic illness, nor in patients with Parkinson’s disease or younger patients with accidental death without evidence of disease. , These findings identify novel HERV-K and genomic markers of ALS, which may have important implications for defining the pathophysiology of sporadic forms of this disease. However, the mechanisms by which these viruses cause or contribute to pathological changes requires further study.