| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Review article |
Department of Medicine, University of Cambridge, Addenbrooke's Hospital, Hills Road, Cambridge CB2 2QQ, UK
Correspondence
John Sinclair
js{at}mole.bio.cam.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT REFERENCES |
|---|
Published online ahead of print on 15 March 2006 as DOI 10.1099/vir.0.81891-0.
Introduction
Human cytomegalovirus (HCMV) is a 
-herpesvirus (Human herpesvirus 5, genus Cytomegalovirus, subfamily Betaherpesvirinae, family Herpesviridae) that causes widespread, persistent human infection. Its seroprevalence can vary from 50 to 90 %, mainly depending on the socio-economic status of the population. HCMV does not usually cause clinically obvious disease upon primary infection of an immunocompetent individual, but clinical disease is much more likely when infection occurs in the immunosuppressed, particularly transplant recipients and patients with AIDS (Drew, 1988
; Rubin, 1990). Transplacental transmission during pregnancy or neonatal infection of premature newborns can lead to neurological damage, manifesting itself as deafness or learning disability in early life. Indeed, a study by the US Institute of Medicine that gave particular weighting to quality life years lost in consequence of an infection concluded that HCMV infection was one of the highest priorities for vaccine development (Arvin et al., 2004). During active HCMV infection and during recovery from illness, virus can be detected in many cell types and is excreted in the urine, saliva and breast milk, indicating the involvement of a number of tissues as sites of productive infection and viral transmission during active infection (Stagno et al., 1980; Sinzger et al., 1995). However, exactly how HCMV infection in specific cell types translates into specific pathology during active HCMV infection and disease is far from clear.
During productive infection in the primary human fibroblast, a cell type used extensively in vitro for culture of HCMV, the viral genes are expressed in a temporal cascade, the first viral genes expressed being termed the immediate-early (IE) genes (Fig. 1). The most abundantly expressed viral genes at this IE time are transcribed from the major IE locus, which encodes the major IE72 and IE86 gene products (Stenberg, 1996). These viral gene products are generated by differential splicing and polyadenylation of a primary transcript from a single transcription start site (Fig. 1) and play a pivotal role in controlling viral and cellular gene expression to optimize the cellular environment for the production of viral DNA (Colberg-Poley, 1996; Stenberg, 1996; Fortunato et al., 2000; Castillo & Kowalik, 2002; Goldmacher, 2004). After IE expression, transcription of viral early (E) genes and then late genes occurs. E genes generally encode functions associated with viral DNA replication and late genes encode virus structural proteins. Expression of both E and late genes is dependent on IE gene expression (Spector, 1996). Consequently, it has now become clear that expression of the viral major IE gene products is absolutely critical for this temporal cascade of viral gene expression and that the viral major IE proteins act to initiate the virus productive-infection programme.
|
HCMV can remain latent throughout the lifetime of the host and sporadic reactivation events, if they occur, are generally well-controlled by cell-mediated immunosurveillance. However, as with primary infection, when reactivation occurs in immunocompromised AIDS patients or immunosuppressed transplant patients, HCMV replication can become uncontrolled and often leads to high levels of morbidity and mortality. Reactivation of latent HCMV in these settings presents a very serious clinical problem.
Analyses of virus strains during HCMV infection of, for instance, organ-transplant patients has shown that the predominant cause of infection is reactivation of the transplant recipient's own HCMV rather than virus transfer from the donor (Smyth et al., 1991). Consequently, understanding where and how HCMV persists latently in the seropositive carrier, and particularly what regulates HCMV reactivation, is of major importance for the understanding of HCMV pathogenesis in the transplant scenario.
Consistent with the mechanism of maintenance of latency of other herpesviruses, such as Epstein–Barr virus (EBV) and herpes simplex virus type 1 (HSV1) (Speck et al., 1997; Everett, 2000; Halford & Schaffer, 2001; Sinclair, 2003; Amon & Farrell, 2005), it now appears that what determines reactivation from latency of HCMV is expression of viral IE genes; it is whether expression of these crucial viral lytic gene products occurs that essentially determines commitment to the latent- or lytic-infection programme.
HCMV latency before the era of PCR
Viral latency is defined operationally as the persistence of the viral genome in the absence of production of infectious virions, but with the ability of the viral genome to reactivate under specific stimuli. Initial attempts to define cell types carrying HCMV in healthy, seropositive individuals and to what extent viral gene expression occurred in any such cell types were, for a long time, confounded by the difficulty in detecting HCMV at all in normal, healthy individuals. The ability of healthy, HCMV-seropositive blood donors to transmit HCMV infection to recipients and the fact that transmission by blood could be reduced by using leukocyte-depleted blood products suggested strongly that one site of carriage of this virus was in the peripheral-blood compartment (Yeager et al., 1981; Adler, 1983; Tolpin et al., 1985; de Graan-Hentzen et al., 1989). However, HCMV is notoriously difficult to isolate directly from the blood of healthy donors (Jordan, 1983). Consequently, whilst the peripheral blood of healthy, seropositive individuals does not carry infectious virus, it must carry virus in a form that is reactivatable. Attempts to detect viral DNA, RNA or protein in such cell types by Southern, Northern and Western blot analyses, respectively, or from tissue biopsies of otherwise healthy individuals by in situ hybridization or immunohistochemistry produced data that were never reproduced when more sensitive and specific analyses became available (Schrier et al., 1985; Gnann et al., 1988; Hendrix et al., 1997). One immunohistochemical analysis from post-mortem tissue did suggest that viral IE expression could be detected in most, if not all, tissues analysed (Toorkey & Carrigan, 1989), but the effects of post-mortem stress on viral gene expression could never be ruled out and such levels of viral gene expression were never detected in live tissue biopsies (Gnann et al., 1988). It was not until the technological breakthrough of PCR and its use as a highly sensitive detection tool for low copy numbers of DNA sequences that the question of sites of carriage of HCMV in vivo was really able to be addressed.
Cells of the myeloid lineage are an important site of carriage of HCMV
By using highly sensitive PCR strategies, a number of laboratories have demonstrated the presence of HCMV DNA in the peripheral blood leukocytes of healthy, HCMV-seropositive individuals (Taylor-Wiedeman et al., 1991; Stanier et al., 1992; Larsson et al., 1998). These and other analyses, particularly using granulocyte–macrophage colony-stimulating factor (GM-CSF)-mobilized peripheral blood cells (Slobedman & Mocarski, 1999), suggest that the frequency of cells that carry the HCMV genome is extremely low (probably <1 in 10 000 peripheral blood mononuclear cells), requiring very sensitive PCR conditions. It is perhaps for this reason that the detection of HCMV DNA in the blood of healthy carriers by other detection techniques was so difficult and that some laboratories cannot detect HCMV DNA in donor blood by less-sensitive PCR protocols (Bitsch et al., 1992; Meyer-Konig et al., 1997). By using sorted peripheral blood-cell populations, however, it has now become clear that peripheral blood monocytes are a major site of carriage of HCMV DNA in healthy carriers (Taylor-Wiedeman et al., 1991; Larsson et al., 1998).
As with all lymphoid and myeloid cells, monocytes arise from pluripotent, CD34+ stem cells present in bone marrow. These CD34+ stem cells are believed to be the progenitors of all blood-cell lineages and capable of self-renewal (Metcalf, 1989). These CD34+ stem cells differentiate along the myeloid lineage to monoblasts, then promonocytes, in the bone marrow and then enter the bloodstream, where they lose CD34 cell-surface antigen and develop into monocytes (Katz et al., 1985). This lineage development depends upon responses to haematopoietic growth factors, particularly GM-CSF and interleukin 3 (IL-3), which stimulate cell division and differentiation (Adams & Hamilton, 1987).
Interestingly, HCMV DNA can also be detected in such CD34+ bone-marrow progenitors (Mendelson et al., 1996). Myeloid progenitors (CD34+ cells) give rise not only to monocytes, but also to a number of other cell types including B cells, T cells and polymorphonuclear leukocytes (PMNLs). Consequently, it might be predicted that these cell types could also carry HCMV in the blood of normal, healthy carriers. However, viral genome does not appear to be carried in the PMNL, T-cell or B-cell fractions of peripheral blood (Taylor-Wiedeman et al., 1991, 1993). How the HCMV genome is maintained selectively in only particular subsets of cells arising from common CD34+ stem cells that carry viral DNA is not yet understood.
Analysis of the conformation of HCMV DNA in peripheral blood mononuclear cells has also been assessed by using electrophoretic separation on native agarose gels in combination with PCR detection and has shown that the viral genome migrates as a circular plasmid (Bolovan-Fritts et al., 1999) – consistent with carriage of virus genome in these cells as an episome. Whether these viral genomes are replicating actively has been difficult to address. What is known is that little or no viral lytic gene expression can be detected during carriage of HCMV in CD34+ progenitors or monocytes, although this conclusion has been contentious. For instance, whilst some analyses using in situ hybridization have demonstrated the presence of RNA from the major IE region of HCMV in mononuclear cells of healthy carriers (Schrier et al., 1985), other analyses have failed to detect HCMV IE RNA in seropositive donor tissue, even though biopsies taken after transplant of these tissues into seronegative recipients showed high levels of HCMV IE RNA (Gnann et al., 1988). Such conflicting results may reflect limitations in the specificity or sensitivity of the detection method. However, more sensitive analyses using reverse transcription followed by PCR (RT-PCR) to detect HCMV lytic gene expression in monocytes or monocyte progenitors of healthy, seropositive carriers also failed routinely to detect viral IE gene expression, even though they clearly carried HCMV DNA (Taylor-Wiedeman et al., 1994; Mendelson et al., 1996), consistent with the inability to culture virus from cells. Consequently, at least in monocytic cells and their precursors, it would appear that HCMV is carried in a true latent state, with little or no accompanying viral IE gene expression.
Carriage of HCMV genomes in myeloid cells in healthy, seropositive individuals is, therefore, not associated with any substantial level of lytic infection, arguing against virus genome persistence in these cell types by, for instance, low-level productive infection. However, the fact that there is evidence suggesting that virus genomes are probably carried in such monocytic cells in an episomal form (Bolovan-Fritts et al., 1999) begs the question of whether HCMV DNA is maintained by some type of latent viral DNA replication akin to that observed for EBV (Adams, 1987; Leight & Sugden, 2000). There is a notable lack of direct evidence for a latent origin of replication in HCMV or for any HCMV genes homologous to latent genome-maintenance factors such as the EBV-encoded Epstein–Barr nuclear antigen 1 (EBNA1) and, unlike EBV, there are no established cell lines that are able to carry HCMV stably, which could help to identify such functions. However, Mocarski et al. (2006) observed that deletion of DNA sequences close to the viral major IE coding region appears to affect the maintenance of viral genomes in experimentally infected undifferentiated granulocyte–macrophage precursors (GMPs) maintained in long-term culture, an experimental latency system that maintains viral genomes in the absence of productive infection (see below). Whether this indicates a latent origin of replication awaits further investigation.
Alternatively, it is possible that carriage of HCMV genomes in differentiating myeloid cells, at least in the peripheral blood, does not involve replication of viral DNA at all. The differentiation and development of haematopoietic cells is quite rapid in vivo; a CD34+ bone-marrow progenitor may take only 2–3 days to exit the marrow and differentiate to terminally differentiated myeloid cell types, such as macrophages or dendritic cells (DCs) (Strobl, 2003). Therefore, it is conceivable that HCMV DNA is carried passively by these myeloid progenitors as they exit the bone marrow into the peripheral blood, where they differentiate, and the resulting macrophages and DCs then reactivate virus; new, latently infected myeloid cells are simply reseeded into the peripheral blood from the bone marrow by, perhaps, a small population of self-renewing, CD34+, pluripotent stem cells. However, if this is the case, then whilst there may be no active maintenance of viral genomes in peripheral blood monocytic cells, it is likely that carriage of virus genomes in any self-renewing stem-cell population would require some form of active viral genome maintenance and segregation to daughter cells or, alternatively, it might result from a continual low-level productive infection of, for instance, bone-marrow stromal cells. What actually happens in vivo is still unclear.
Attempts to shed light on this aspect of HCMV biology by trying to identify viral transcripts associated with latent carriage of the virus in naturally infected myeloid cells have been confounded by the extremely low frequency of HCMV genome-positive cells in seropositive carriers, and it is not possible to isolate CD34+ cells or monocytes enriched for the carriage of HCMV genomes from healthy, seropositive individuals to enable comprehensive analyses of viral transcription. This has led to extensive use of experimental infection of differentiating myeloid cell cultures, which takes advantage of the differentiation-dependent permissiveness of myeloid cells for productive infection in vitro with HCMV. Experimental infection of, for instance, pluripotent or totipotent bone-marrow progenitor cells results in the carriage of viral genomes for 2–3 weeks in the absence of any substantial production of infectious virus. Virus production can, however, be reactivated after differentiation of these long-term cultures (Kondo et al., 1994; Minton et al., 1994; Hahn et al., 1998; Maciejewski & St Jeor, 1999; Goodrum et al., 2002; Reeves et al., 2005a). Cell systems supporting this type of experimental latency of HCMV have been used to try to identify putative latent viral transcripts; expression of such transcripts can then be confirmed, or otherwise, in naturally infected individuals.
Viral gene expression associated with experimental infection in latent model systems
Kondo and colleagues (Kondo et al., 1994; Kondo & Mocarski, 1995), using experimentally infected GMPs derived from fetal liver cells, originally identified several HCMV transcripts that were expressed in the absence of virus production in these long-term cultures. These cytomegalovirus latency-specific transcripts (CLTs) included novel spliced and unspliced RNA transcripts mapping to both strands of the HCMV major IE region and included open reading frames (ORFs) that were, in some cases, recognized at the protein level by antisera from infected individuals (Kondo et al., 1996). Whilst some of these transcripts were also identified in healthy, seropositive carriers (Kondo & Mocarski, 1995; Kondo et al., 1996; Hahn et al., 1998) and antibodies to some putative CLT ORFs were detected in healthy carriers (Landini et al., 2000), many of the transcripts were also detected in infected cells in culture (Lunetta & Wiedeman, 2000). Hence, the true role of these CLTs, if any, in HCMV latency is yet to be established fully. Initial analysis of one major ORF (ORF94) using an ORF94-deletion virus failed to observe any effect on latency or reactivation after experimental infection of GMPs (White et al., 2000). So far, no viral gene products from these CLTs have been shown to play a role in latency.
The same experimental model of latency has also identified a novel transcript from the unique long (UL) region of the viral genome (UL111.5A), which is predicted to encode a viral homologue of IL-10 (vIL-10). IL-10 is a cytokine that inhibits immune responses and the authors suggested that expression of a homologue may help to avoid host immunosurveillance during latency. However, the role of this transcript in latency is not clear and is somewhat confounded by the observation that detection of this putative viral transcript is not correlated with HCMV serostatus; the transcript was also detected in monocytes from HCMV-seronegative individuals (Jenkins et al., 2004).
Other types of experimentally infected primary haematopoetic cells have also been used to try to identify viral transcripts associated with latent carriage of viral genomes. Long-term growth of CD34+ progenitors from healthy, seropositive individuals on murine stromal cells has been used for a much more comprehensive analysis of viral gene expression using HCMV gene arrays, although the exact phenotype of these cells after their long-term growth was not established (Goodrum et al., 2002). Such experiments originally identified multiple viral RNAs associated with carriage of the virus in the absence of production of infectious virions; many, if not all, of which are also detected during productive infection (Goodrum et al., 2002). However, it was not determined whether any were also detectable in mononuclear cells of normal, healthy carriers. A more detailed analysis of subpopulations of these CD34+ bone-marrow progenitors suggested that different subpopulations of CD34+ progenitor cells are able to support different levels of HCMV infection, ranging from productive to abortive infection to latent carriage of the virus, depending on the phenotype, but, again, this study did not define any latent-specific transcripts (Goodrum et al., 2004).
Based on transcripts identified by Goodrum et al. (2002), a more recent analysis of RNA isolated from monocytes of healthy, seropositive carriers has identified viral RNAs antisense to the UL81–82 region of the viral genome (UL81–82ast). Detectable in bone-marrow mononuclear cells of seropositive but not seronegative individuals (Bego et al., 2005), this transcript contains an ORF of 133 aa that can be identified by Western blot analysis during lytic infection of fibroblasts. As this transcript is antisense to the viral UL82 gene, which encodes the viral pp71 tegument protein, a known transcriptional activator of the viral major IE promoter/enhancer (MIEP), it has been suggested that the transcript or its protein product may be involved in restricting IE gene expression, perhaps helping to maintain latency (Bego et al., 2005). However, confirmation of its role in HCMV latency awaits further study.
One other study using abortive infection of a myelomonocytic cell line (THP1) as a model of latency has also identified the viral US28 gene product as a putative latent transcript (Beisser et al., 2001). However, virus was never reactivated from these cells and the authors did not attempt to detect US28 RNAs in naturally latently infected individuals.
Attempts to identify viral transcription associated with carriage of HCMV in myeloid cells from naturally latently infected individuals has clearly been frustrated by the low frequency of cells carrying latent HCMV in vivo. The model systems described above, based on experimental infection of undifferentiated myeloid cells, have succeeded in identifying a number of putative latent viral RNAs. However, a definitive role for any of these RNAs in HCMV latency has yet to be shown.
If latent viral RNAs do exist, it is likely that they are either associated with maintenance of the viral genome directly or that they confer on the cell a gene-expression profile more conducive to carriage of viral genomes. On this basis, experimental infection of undifferentiated myeloid cells, resulting in long-term carriage of virus genomes in the absence of overt productive infection, has also been used to try to identify changes in cellular gene expression that may be associated with HCMV latency.
Cellular gene expression associated with infection in latent model systems
Changes in cellular gene expression associated with productive infection of numerous cell types by HCMV have been studied extensively, either at the single-gene level or at a global level by microarray analyses (Zhu et al., 1998; Challacombe et al., 2004). However, studies to determine changes in the cellular transcriptome of latently infected cells have, once again, been frustrated by the low level of cells carrying viral genomes in vivo. Consequently, long-term carriage of experimentally infected, undifferentiated cells has also been used as a model to identify these types of changes in cellular gene expression that may be associated with ‘latent’ infection.
Initial observations have shown that, whilst a number of cell-surface markers appear not to be altered on experimentally infected GMP cultures, GMPs that carry viral genomes downregulate major histocompatibility class II (MHC II) expression at the protein level (Slobedman et al., 2002). What mediates this MHC class II downregulation is unknown, but it appears to be independent of the previously characterized MHC I and MHC II immunomodulatory genes in the US2–11 region of HCMV, which are known to function during productive infection. Similarly, why MHC class II expression should be targeted in this way is also unknown, but it may help latently infected cells to avoid immunosurveillance (Slobedman et al., 2002). The same latent model system has also recently been used to carry out a global survey of changes in cellular gene expression associated with carriage of experimentally infected virus in GMPs by microarray analysis (Slobedman et al., 2004). This analysis failed to recapitulate the downregulation of MHC class II expression at the RNA level observed previously, but identified an assortment of other cellular transcripts that were up- or downregulated in experimentally infected GMPs carrying HCMV in the absence of productive infection. Many of these changes involved transcripts from cellular genes associated with the immune response, cell growth, signalling and transcriptional regulation, but the viral gene products that mediate these changes and the mechanisms by which they are altered await further analysis, as does the biological impact of such cellular changes on virus–host-cell interactions in vivo. Underlying these studies is the crucial question of whether experimental infection of undifferentiated myeloid progenitors and their long-term culture with concomitant carriage of viral genomes in the absence of virus production can really ever recapitulate natural latency in vivo. It is very difficult to rule out the possibility that changes in cellular gene expression observed in such cultures result from the initial insult of the experimental infection (normally at very high m.o.i.s) or that cells of a specific phenotype are the only ones that actually support the initial infection in vitro. The ability to directly address such cell-specific changes during natural latency awaits technological breakthroughs (perhaps as innovative as PCR was 15 years ago) that will allow the identification and enrichment of bone-marrow progenitor cells or other myeloid cells that specifically carry latent viral genomes from healthy, seropositive individuals.
Similarly, the question as to what changes in cellular gene expression accompany true reactivation of HCMV in the natural setting is of equal importance, yet will only be addressed comprehensively when techniques to identify naturally infected latent cells in vivo have been developed.
Are there other sites of latency or long-term carriage of HCMV besides myeloid cells?
Although there is now a real consensus that myeloid cells are an important site of true latency of HCMV in vivo, the possibility that other sites of latency occur in normal, healthy carriers requires serious consideration. A vast number of cell types quickly become infected productively upon clinical reactivation, begging the question of whether or not virus in these cells results from reactivation of HCMV per se or reflects de novo infection after rapid dissemination from reactivating mononuclear cells.
Similarly, CD34+ bone marrow-derived cells may also give rise to endothelial cells (ECs) (Quirici et al., 2001) and, hence, the possibility that ECs may be one other reservoir of latent virus has been suggested (Jarvis & Nelson, 2002), perhaps lending support to the circumstantial evidence that has linked HCMV and atherosclerosis (Epstein et al., 1996).
However, we have been unable to detect latent HCMV genomes in ECs or vascular smooth-muscle cells from saphenous vein of healthy, seropositive individuals, even though HCMV DNA can be detected readily in monocytes at the same time (Reeves et al., 2004), suggesting that cells of the microvasculature are unlikely to be important sites of HCMV latency in vivo. Whether or not cells of the macrovasculature, such as aortic ECs, are sites of carriage of the virus in vivo is more difficult to address, not least because of the obvious difficulty in obtaining such cells ex vivo from healthy donors. For the same reason, it has not been possible to analyse many other cell types for latent carriage of HCMV in vivo.
The possibility that HCMV may persist in certain cells of healthy carriers as a low-level productive infection with little or no pathology or cell lysis has also been raised. Work using experimental infection of cultured ECs from different sources has suggested that HCMV might persist as a low-level productive infection in aortic ECs in the absence of cytopathic effects (Fish et al., 1996; Streblow & Nelson, 2003) and would go some way to explaining the speed with which a wide variety of cell types become infected in vivo during reactivation resulting from, for instance, immunosuppression. However, other workers have observed that ECs, regardless of tissue of origin, show full cytopathic changes upon HCMV infection that always progress to complete lysis (Kahl et al., 2000). Whether other sites in vivo (such as bone-marrow stromal cells) might, similarly, act as a virus reservoir maintained by low-level productive infection in the absence of cell lysis has not been addressed fully. An argument against this would be the observation that HCMV appears to undergo a fully lytic, productive infection in bone-marrow fibroblasts and myofibroblasts in vitro (Apperley et al., 1989; Michelson et al., 2001).
Nevertheless, these types of considerations do suggest that carriage of the virus in vivo is likely to be a complicated affair, with latency (no productive infection) and persistence (low-level productive infection) linked intimately and occurring simultaneously in different cell types. Clearly, myeloid cells represent one true site of latency in normal, healthy carriers, with little or no lytic gene expression. However, terminally differentiated macrophages and DCs in the same individual are likely sites of continual, subclinical reactivation. Consequently it is likely that, in vivo, HCMV reactivation occurs routinely in normal, healthy virus carriers, but that this is unlikely to present a problem in the immunocompetent, due to a robust CD8+ cytotoxic T-lymphocyte (CTL) response to the virus. Consistent with this is the observation that the T-cell repertoire of healthy, seropositive individuals contains a strikingly high frequency of CTLs that recognize HCMV epitopes (Borysiewicz et al., 1988; Riddell et al., 1991; McLaughlin-Taylor et al., 1994; Wills et al., 1996; Sylwester et al., 2005). This recognition is dominated by a relatively small number of major viral peptide epitopes, several of them derived from viral structural proteins, suggesting that the immune response to HCMV is subject to regular restimulation, with consequent focusing of the CTL response on a limited number of epitopes (McLaughlin-Taylor et al., 1994; Wills et al., 1996, 2002; Kern et al., 2002; Sylwester et al., 2005) – an event possibly mediated by sporadic levels of subclinical reactivation.
It is also still unclear whether any increased frequency of cells reactivating HCMV from latency results from immunosuppression per se; reactivation itself could be stimulated greatly by the numerous cytokines elicited by other infections, allogeneic stimulation, transplant rejection or graft-versus-host disease – all of which often result in, or are treated by, immunosuppression.
The question of what cell types in vivo carry HCMV DNA and whether these cells are infected lytically or truly latently is crucial to our understanding of HCMV reactivation and pathogenesis. However, addressing such questions is problematic, mainly because it is difficult to obtain tissue samples for analysis from healthy individuals. Consequently, our knowledge of the sites of carriage of HCMV in the healthy seropositive, other than in cells of the myeloid lineage, and the extent of virus gene expression in those sites is far from comprehensive.
Reactivation of HCMV from naturally latently infected myeloid cells requires differentiation
Nevertheless, the myeloid lineage has provided a wealth of information about the possible mechanisms of latency and reactivation of HCMV in vivo.
None of the myeloid cells identified in peripheral blood of healthy individuals that do carry viral DNA produce infectious virus or express viral lytic genes to any appreciable extent (Taylor-Wiedeman et al., 1994; Mendelson et al., 1996); they thus appear to be sites of classical viral latency. However, what has now become apparent is that, once these myeloid cells differentiate to macrophages and DCs, a fundamental change in their ability to support viral IE gene expression occurs. Differentiation of monocytes from healthy, seropositive donors in vitro to macrophages is known to result in reactivation of viral IE gene expression (Taylor-Wiedeman et al., 1994), but there was no evidence of reactivation of infectious virus from these cells. However, complete reactivation of infectious virus from peripheral blood mononuclear cells has been reported after their long-term culture and differentiation in the presence of supernatant medium containing cytokines produced by the allogeneic stimulation of T cells (Söderberg-Nauclér et al., 1997). Despite attempts to characterize the specific factors in the culture media that might be responsible for causing this reactivation, their nature remained unresolved (Söderberg-Nauclér et al., 2001). More detailed analyses have been complicated somewhat by the fact that this approach to reactivating HCMV from naturally latent peripheral blood mononuclear cells has not yet been reproduced independently by other investigators. Nevertheless, cells from which HCMV reactivated in this study were seen to carry macrophage and some DC markers. DCs are highly specialized cells that present antigen to cells of the immune system and thus play a key role in the induction of the immune response. There are at least two types of myeloid lineage-derived DCs: the Langerhans DCs that are located in the epithelium and the interstitial DCs, which reside in deeper-lying tissues (Banchereau et al., 2000).
Consistent with the observation that the cells within naturally infected peripheral blood mononuclear cells from which HCMV reactivates after allogeneic T-cell stimulation carried some DC markers, our laboratory has observed full reactivation of infectious HCMV after ex vivo differentiation of CD34+ bone-marrow progenitors to a mature DC phenotype consistent with that of mature Langerhans DCs (Reeves et al., 2005b). These data were based entirely on direct analysis of material isolated from healthy virus carriers and show how the virus selectively utilizes the differentiation pathway of key antigen-presenting cells to persist in its human host. However, only three out of five HCMV-seropositive samples tested reactivated endogenous latent virus, suggesting that some, but not all, donors respond to differentiation-dependent reactivation for reasons that are not yet clear. It was, however, noticed that donors from whom infectious virus could be routinely reactivated tended to carry higher levels of latent viral genome, suggesting that higher HCMV ‘latent loads' in a donor may make it easier either to reactivate virus per se, as has been suggested for HSV1 (Sawtell, 1998), or to detect reactivation (Reeves et al., 2005b). Consequently, the robustness of this as a system for the routine reactivation of HCMV from naturally infected cells also awaits independent confirmation from other laboratories.
These caveats notwithstanding, there appears to be a fundamental correlation between cellular differentiation and productive infection and the implications of this with respect to the characteristics of HCMV persistence are discussed extensively below.
Permissiveness of myeloid cells for experimental HCMV infection requires cellular differentiation
The differentiation-dependent reactivation of viral lytic gene expression from myeloid cells infected naturally with HCMV, perhaps not surprisingly, also mimics well the relative permissiveness of these different myeloid cell types for experimental infection in vitro with virus.
It has been known for some time that peripheral blood monocytic cells are generally non-permissive for experimental infection with HCMV (Rice et al., 1984; Einhorn et al., 1985), whereas their differentiation to a macrophage-like phenotype results in cells that are fully permissive for infection (Ibanez et al., 1991; Lathey & Spector, 1991). More recently, other work (Riegler et al., 2000; Hertel et al., 2003) has shown that, unlike their monocytic or CD34+ progenitors, mature DCs are also fully permissive for experimental HCMV infection and this is consistent with the observations that cells that reactivated naturally acquired latent virus after allogeneic T-cell stimulation of peripheral blood monocytes expressed some DC markers (Söderberg-Nauclér et al., 1997), as well as with the demonstration of HCMV reactivation after specific differentiation of CD34+ cells to mature DCs (Reeves et al., 2005b); clearly, cells that reactivate naturally acquired latent virus would be predicted to be fully permissive for HCMV after experimental infection.
A number of analyses of this differentiation-dependent permissiveness for HCMV infection in vitro have shown that this block in permissiveness for HCMV infection in undifferentiated cells is due to their inability to support viral IE gene expression (Riegler et al., 2000; Hertel et al., 2003) and is not due to the inability of cells to bind or internalize virus. Similarly, this block is also observed when using myelotropic and EC-tropic strains of HCMV, which carry about 20 kb of additional DNA sequence that is known to be lost upon long-term carriage of clinical isolates of HCMV in fibroblasts in vitro (Cha et al., 1996); loss of specific genes in this region generally leads to the inability to passage fibroblast-adapted strains of HCMV in ECs or DCs (Hahn et al., 2004; Gerna et al., 2005; Wang & Shenk, 2005).
The difficulty in obtaining primary cells from blood for these types of analyses has led investigators to try to identify immortalized or transformed human cell lines that could recapitulate differentiation-dependent permissiveness for HCMV infection, and thus permit a more thorough analysis of what actually may be restricting viral IE gene expression in such a differentiation-dependent manner. Two such cell lines identified were the myelomonocytic THP1 cell line (Weinshenker et al., 1988) and the teratocarcinoma cell line NT2D1 (Gonczol et al., 1984). Both of these lines showed a clear restriction in their ability to support HCMV infection, but their differentiation resulted in a cell type that was fully permissive for full, productive infection.
Clearly, such blocks in virus productive infection could occur at many stages of the viral life cycle and, indeed, there is evidence that some cultured cells are able to express the virus major IE72 protein after infection, but are unable to replicate viral DNA (LaFemina & Hayward, 1988). However, many experiments have concluded that the lack of viral lytic gene expression in undifferentiated NT2D1 or THP1 cells is also not due to a lack of binding or internalization of virus, nor an inability to transport virus genome to the nucleus in these cells, but rather is due to a specific block in expression of viral IE genes, which then prevents later classes of viral gene expression. Differentiation of NT2D1 cells to a neuronal phenotype with retinoic acid (RA) or of THP1 to a macrophage phenotype with phorbol esters (PMA) lifts this block in IE expression, resulting in productive infection (Gonczol et al., 1984; LaFemina & Hayward, 1986; Nelson & Groudine, 1986; Weinshenker et al., 1988; Shelbourn et al., 1989).
These cell types, therefore, represent tractable systems in which permissiveness for viral IE expression and productive infection is differentiation-dependent, similar to that observed for peripheral blood myeloid cells in vitro. A number of laboratories have used these cell systems to try to identify differentiation-specific regulators of IE gene expression on the basis that they may help to define factors that play a role in viral latency and reactivation.
Fortunately, the restriction in viral major IE gene expression upon virus infection of undifferentiated cell types in vitro also extends to the ability of the viral MIEP to express in transient-transfection analyses in these cell types (LaFemina & Hayward, 1986; Nelson & Groudine, 1986; Shelbourn et al., 1989; Sinclair et al., 1992). This has permitted analyses of the regions of the viral MIEP responsible for such differentiation-dependent repression.
Cellular transcription factors that may be involved in mediating latency and reactivation of HCMV
HCMV major IE gene expression is under the control of the MIEP (Fig. 2). This region is extremely complex and contains one of the strongest-known transcriptional enhancers (made up of an array of 17, 18, 19 and 21 bp-repeat elements) and a far-upstream element, which has been termed the modulator and which includes an imperfect dyad symmetry (Boshart et al., 1985; Lubon et al., 1989; Meier & Stinski, 1996). The modulator region has been shown to be highly sensitive to DNase I and the extent of hypersensitivity changes upon differentiation of cells to a phenotype permissive for HCMV infection (Nelson et al., 1987; Lubon et al., 1989). It has, therefore, long been suggested that changes in binding of cellular transcription factors to the MIEP may play an important role in determining the permissiveness of cells for HCMV IE expression (Lubon et al., 1989; Ghazal et al., 1990; Nelson et al., 1990) and, hence, may help to identify cellular factors controlling virus latency and reactivation in vivo.
|
). For example, nuclear factor 
B (NF-B) (Sambucetti et al., 1989), cAMP response element-binding protein (CREB) (Hunninghake et al., 1989) and Sp1 (Lang et al., 1992) are known to activate the MIEP by binding to the 18 (NF-B) and 19 (CREB and SP1) bp-repeat sequences.
However, sequence analysis of the MIEP regions has also identified putative binding sites for a number of other cellular transcription factors that are known to act as transcriptional repressors. These include proteins such as modulator-binding factors (Shelbourn et al., 1989; Kothari et al., 1991), Yin Yang 1 (YY1) (Liu et al., 1994), methylated DNA-binding protein (Zhang et al., 1991), modulator-recognition factor (Huang et al., 1996), growth factor independence-1 (Zweidler-McKay et al., 1996) and Ets-2 repressor factor (ERF) (Bain et al., 2003). Consequently, it was suggested that a balance of these positively and negatively regulating factors determines whether a specific cell type is permissive or non-permissive for viral major IE gene expression and that a change in this balance accompanies differentiation of cells to a permissive phenotype (Sinclair & Sissons, 1996).
Consistent with this, we have found that differentiation-specific repression of the MIEP in non-permissive NT2D1 and THP1 cells is due, at least in part, to DNA sequences within the modulator region and the 21 bp-repeat elements of the enhancer (Shelbourn et al., 1989; Kothari et al., 1991; Liu et al., 1994). These DNA sequences within the MIEP bind YY1 and ERF specifically in vitro (Fig. 2
) and both YY1 and ERF can mediate repression of the MIEP in transient co-transfection experiments (Liu et al., 1994; Bain et al., 2003). Whilst many groups have shown consistently that the modulator, 21 bp repeats and other upstream elements of the MIEP play a clear role in inhibiting IE gene expression in transfection assays (Shelbourn et al., 1989; Kothari et al., 1991; Zhang et al., 1991; Liu et al., 1994; Huang et al., 1996; Zweidler-McKay et al., 1996; Bain et al., 2003), the question of whether these repressor sites play any role in the context of the virus genome is a valid one.
Recent studies using short interfering RNAs to specifically inhibit expression of some of these putative repressors in normally non-permissive cells have asked whether knock-down of their expression permits expression from the viral MIEP. Unfortunately YY1, as well as negatively regulating the viral MIEP, is a cellular housekeeping gene whose knock-down leads to profoundly detrimental effects on the cell (Bain & Sinclair, 2005).
Other studies using deletion mutants of HCMV lacking the binding sites for these factors in the MIEP have argued against these sites playing a role in the differentiation-dependent repression of the MIEP (Meier & Stinski, 1997). However, the large deletions in the viruses used in such studies have also been shown by the same workers to have perturbed IE gene expression even in permissive cells (Meier & Pruessner, 2000) and this may cause difficulties in such comparisons of IE gene expression in non-permissive and permissive cell types. Consequently, confirmation or otherwise of the role of the modulator and 21 bp repeats in repression of the viral MIEP in the context of the virus awaits more defined, specific virus deletions or mutations.
This notwithstanding, transfection analyses of MIEP reporter constructs in such differentiation-dependent permissive cell lines do suggest that the control of the MIEP by differentiation-specific cellular transcription factors could be a mechanism by which viral lytic gene expression is restricted in vitro and that these factors may also be involved in the regulation of viral latency and reactivation in vivo.
Until recently, the mechanism by which cellular transcription factors such as YY1 or ERF might act to repress the MIEP was unclear. YY1 is expressed differentially in NT2D1 and THP1 cells and appears to be controlled post-translationally, in that steady-state levels of YY1 RNA are constant throughout differentiation, whereas YY1 protein levels are reduced substantially in differentiated NT2D1 (Pizzorno, 2001) and THP1 cells. In contrast, differentiation of both NT2D1 and THP1 cells is not associated with decreases in steady-state levels of ERF protein (Bain et al., 2003). Consequently, it cannot simply be that differentiation results in a global decrease in protein expression of these MIEP-binding factors. It has now become clear that the mechanism by which YY1 and ERF repress the viral MIEP involves the ability of these proteins to interact physically with and recruit chromatin-modification enzymes to the viral MIEP (Wright et al., 2005); specifically, histone deacetylases (HDACs) that remove acetyl groups from core histone tails. It is the changes in post-translational modification of histones around the viral MIEP resulting from such HDAC recruitment that appear to play a pivotal role in the differentiation-dependent regulation of the MIEP during virus infection of undifferentiated and differentiated cell types.
Chromatin structure of the MIEP upon infection determines the permissiveness of cell lines for productive infection in vitro
How do such changes in the post-translational modification of histones regulate the transcriptional activity of specific promoters? It is now very clear that the chromatinization of eukaryotic DNA, mediated by its association with histone proteins, has the dual role of both tightly packaging the DNA into the nucleus and regulating the transcription of cellular genes (Weintraub & Groudine, 1976; Fig. 3). This regulation is imparted by post-translational modification of specific histones (Strahl & Allis, 2000). Generally, histone acetylation resulting from activity of specific histone acetylases is a marker of transcriptionally active chromatin (Lusser, 2002; Kuo & Allis, 1998; Eberharter & Becker, 2002). In contrast, deacetylated histones, resulting from the activity of HDACs, become targets for methylation and the subsequent recruitment of transcription-silencing proteins such as heterochromatin protein 1 (HP1) – methylated histones and HP1 binding both being markers of transcriptionally silenced chromatin (Bannister et al., 2001; Khochbin et al., 2001; Schotta et al., 2002; Rice et al., 2003; Khorasanizadeh, 2004). Recent work has suggested that expression of key genes from a number of viruses is also regulated by their chromatin structure. Indeed, changes in chromatinization of both EBV and HSV1 have been implicated strongly in the control of reactivation of these herpesviruses (Jenkins et al., 2000; Arthur et al., 2001; Kubat et al., 2004; Amon & Farrell, 2005).
|
; Wright et al., 2005), but to be due, at least in part, to a reduction in co-factor repressors, such as HDACs, that YY1 and ERF recruit to the MIEP (Murphy et al., 2002; Wright et al., 2005). A prediction of this would be that, in undifferentiated NT2D1 and monocytic cells, which are refractory to viral IE gene expression, HCMV infection results in association of the viral MIEP with methylated histones and silencing proteins such as HP1. In contrast, infection of differentiated cells, in which IE gene expression is not repressed, should result in association of the viral MIEP with markers of transcriptional activation.
Consistent with this prediction is the observation that treatment of undifferentiated NT2D1 cells with HDAC inhibitors, such as trichostatin A, leads to an increase in IE gene expression in these normally non-permissive cells after experimental infection in vitro (Meier, 2001; Murphy et al., 2002). Analysis of the viral MIEP by chromatin immunoprecipitation assays also shows very clearly that, in undifferentiated NT2D1 or monocytic cells, experimental infection results in the association of the viral MIEP with methylated histones and HP1 protein. In contrast, experimental infection of differentiated NT2D1 cells or monocytic cells results in the association of the MIEP with acetylated histones (Murphy et al., 2002), all entirely consistent with the inability of undifferentiated, but not differentiated, cells to support viral IE expression.
Similar experiments using an in vitro latency system based on experimental infection of CD34+ cells, their subsequent long-term culture in the absence of lytic gene expression and differentiation to mature DCs resulting in virus reactivation (Reeves et al., 2005a) have shown identical changes in MIEP chromatin acetylation/deacetylation and recruitment of HP1; again, this was correlated specifically to reactivation of viral IE expression and production of infectious virus (Reeves et al., 2005a).
Chromatin structure of the MIEP regulates latency and reactivation of HCMV in natural infection in vivo
Clearly, the use of cell lines or myeloid cell-culture systems, which are conditionally permissive for experimental HCMV infection due to their inability to support viral IE gene expression whilst undifferentiated, but not when differentiated, has given enormous insight into the mechanisms by which cell differentiation regulates HCMV lytic gene expression and productive infection. However, whether such mechanisms play a role in the control of latency and reactivation in vivo, even though there is a clear link between differentiation of myeloid cells and reactivation of both viral IE expression (Taylor-Wiedeman et al., 1994) and infectious virus (Söderberg-Nauclér et al., 1997; Reeves et al., 2005b) in naturally latently infected cells as well as in experimental latent systems, is a key question.
Recent work from our laboratory has now shown that such differentiation-dependent chromatin remodelling of the viral MIEP by post-translational modification of histones is also involved in the control of latency and reactivation of HCMV in vivo (Reeves et al., 2005b). Analysis of the chromatin state of the HCMV MIEP carried out directly on cells isolated from naturally infected, healthy, seropositive donors has shown that, in latent HCMV genomes in CD34+ cells and monocytes of naturally infected, healthy carriers, the MIEP is also associated with HP1 and not with acetylated histones; this is consistent with transcriptional repression. However, the differentiation of both CD34+ cells and monocytes to mature DCs results in chromatin remodelling of the viral MIEP. Specifically, HP1 is lost from the MIEP and the histones bound to the MIEP become acetylated, consistent with transcriptional activation of viral gene expression. These changes in chromatinization of the viral MIEP correlate precisely with reactivation of infectious virus from the terminally differentiated, mature DCs (Reeves et al., 2005b). Consequently, reactivation of HCMV from true latency appears to be linked intrinsically to the differentiation status of the myeloid cell and, hence, to the cellular mechanisms that normally control host gene expression (Fig. 4).
|
) and ERF (Wright et al., 2005).
Clearly, the differentiation of CD34+ cells or monocytes to DCs does not result in global acetylation of histones bound to all cellular promoters and, hence, their universal transcriptional activation. So what determines this differentiation-specific acetylation of histones associated with the viral MIEP? It is known that a number of cellular promoters are activated specifically upon myeloid differentiation; this includes promoters driving expression of a number of inflammatory cytokines and chemokines required for the function of DCs (Saccani & Natoli, 2002). The activation of such cellular genes is promoted by the activation of the p38–mitogen-activated protein kinase (MAPK) pathway by extracellular signals, such as mitogens and stress signals. The overall result of the activation of the p38–MAPK pathway is to promote the phosphorylation of a number of cellular proteins, including histone H3. This phosphorylation leads to the recruitment of histone acetyltransferases, subsequent acetylation of the histones and ultimately the activation of the target promoter. This response is restricted to a subset of cellular genes that are NF-B-responsive (Saccani et al., 2002) and it is possible that the activation of the MIEP, an NF-B-responsive promoter itself, may be due to the activation of the p38–MAPK pathway. Indeed, the stimuli for the activation of the p38–MAPK pathway in mature DCs can be delivered either by signals resulting from engagement of a cell-surface molecule on DCs, called CD40, or by lipopolysaccharide-induced maturation of DCs – an event that also appears to be crucial for reactivation of HCMV in DCs (Reeves et al., 2005b). Thus, it is likely that differentiation-dependent activation of the viral MIEP requires the coordinated expression of positive regulators of the MIEP that are not found in undifferentiated cells (Meier & Stinski, 2006), as well as changes in expression of specific negative regulators of IE gene expression, and may well be mimicking the positive regulation of inflammatory genes normally associated with DC maturation.
Conclusion
Our understanding of HCMV latency and reactivation has clearly been hampered by the lack of a good model system. However, sensitive detection techniques that have allowed the analysis of sites of carriage of virus, as well as studies of specific cell types that can be infected with HCMV in vitro, have begun to give insights into the mechanisms by which HCMV is carried in the healthy host. The difficulty in obtaining samples for analysis from healthy carriers has generally restricted these analyses to the peripheral blood. It cannot be overstated that a comprehensive analysis of other sites, besides the peripheral blood, that may act as true sites of latency/persistence of HCMV in the healthy host is an important goal in HCMV research. For instance, a number of observations hint at the possibility of other, non-haematopoietic sites of latency: the speed with which HCMV reactivation occurs in ductal epithelial cells, the much higher incidence of HCMV transmission in solid-organ compared with bone-marrow transplants and the known ability to detect locally restricted virus reactivation in breast milk during lactation suggest that epithelial cells, as well as other cell types, may harbour latent or persistent virus in vivo.
Indeed, analyses of latency of murine cytomegalovirus (MCMV), which we are unable to address in any detail in this review due to space constraints, suggest that latent infection in mice is associated with high loads of latent viral genomes in many tissues (Simon, 2006). At first glance, this seems incompatible with what we know, to date, about HCMV latency. However, it will be crucial to analyse other tissues systematically for the carriage of latent HCMV and to determine whether there really are such fundamental differences between HCMV and MCMV latency in vivo.
Still, a substantial body of work to date points clearly to the important role that myeloid cells play in the maintenance of HCMV in the normal host and to the role of cellular differentiation in the control of viral gene expression.
A speculative model of HCMV latency/persistence in the healthy, seropositive carrier would be that virus is maintained in monocytes and their progenitors and probably particularly in the DC precursors within this population, in a truly latent state with the absence of viral lytic gene expression – which in itself may help to evade the host immune response (Fig. 5). This results from the transcriptional milieu of the monocytes/DCs preventing viral IE expression and would correlate well with the known inability to productively infect monocytes in vitro or to culture virus from peripheral blood monocytes of healthy, seropositive carriers. However, as monocytes/DCs become terminally differentiated to macrophages/mature DCs, the block in viral IE expression is lifted and the viral genome becomes fully reactivated such that these sites in the host may be looked upon as being persistently infected. In this situation, however, the host's cytotoxic T-cell response acts efficiently to clear these productively infected cells unless they are immunocompromised, at which time virus disseminates, resulting in disease.
|
), is not known. It is clear that during experimental infection in vitro, HCMV genomes, as with other herpesviruses, are deposited at or close to specific nuclear structures termed nuclear domain 10 (ND10), which have been suggested to be sites of suppression of viral IE gene expression (Tang & Maul, 2006). Whether such sites play any role in repression of IE gene expression during latency will need to be investigated actively.
Similarly, our understanding of the carriage of HCMV in bone marrow is also incomplete. Virus can clearly enter bone-marrow progenitors after experimental infection in vitro and persist without lytic gene expression during long-term culture of these cells; HCMV DNA can be detected by PCR in the absence of IE RNA in bone-marrow progenitors of naturally infected carriers. Taken together, these observations would be consistent with the bone marrow acting as a reservoir of latent HCMV after primary infection, which then seeds latent virus into the peripheral blood via the monocytes/DCs. Differentiation of these cells to tissue macrophages and DCs then results in local reactivation. This model of HCMV carriage and latency/persistence of HCMV in the healthy carrier, although still speculative, is consistent with many of the observations that have been made to date (Fig. 5).
The recent in vivo studies that we have outlined above now suggest a particularly strong relationship between HCMV and DCs and show that persistence of HCMV is associated intimately with the normal programme of myeloid-cell differentiation; it is the changes in the internal cellular environment that accompany differentiation that promote virus reactivation. HCMV DNA remains latent in myeloid DC progenitors of naturally infected carriers and latent carriage of HCMV is maintained until specific differentiation of these cells to a mature DC phenotype. This results in reactivation of infectious virus from the DCs of some healthy, seropositive individuals, consistent with DC being a biologically significant site of viral reactivation in vivo.
As with other herpesviruses, the induction of IE lytic gene expression from latent virus represents the critical event required for the switch from latency to reactivation. Analyses in vitro have shown that transcriptional activation of viral IE gene expression from the viral MIEP requires the action of cellular transcription factors and changes in chromatinization of the MIEP. Consistent with this, chromatin remodelling of the viral MIEP also plays a crucial role in reactivation of HCMV from latency upon DC differentiation in naturally infected individuals. Latent HCMV in DC progenitors in vivo is in a closed, transcriptionally silent chromatin conformation. In contrast, differentiation of these progenitors to mature DCs results in specific chromatin remodelling of the MIEP to an open, transcriptionally active chromatin structure, resulting in reactivation of viral lytic IE gene expression and production of infectious virus.
Understanding the relationships between HCMV and such professional antigen-presenting cells in detail at the cellular level is important for our understanding of how this virus persists in the host and the complexities of its interaction with the immune system. For example, it is well-recognized that HCMV possesses an unusually large number of genes encoding functions designed to help it evade the cellular immune response; it now seems plausible that these may reflect the particular problems facing the virus in residing in, and in particular reactivating from, cells whose specific function is to initiate immune responses. This new knowledge may also aid the development of more effective treatments for the prevention of HCMV-mediated disease due to reactivation. Clearly, future studies will be needed to define precisely the biochemical triggers responsible for myeloid DC differentiation, as these also appear to promote the switch from viral latency to reactivation. These may help to generate specific strategies to more effectively control virus reactivation, an important aspect of HCMV biology that also leads to a substantial amount of HCMV-mediated disease – particularly that occurring in the context of solid-organ and bone-marrow transplantation.
|
ACKNOWLEDGEMENTS |
|---|
) such that the MIEP remains in a transcriptionally repressed state (e). However, if the CD34+ cells are differentiated to a mature DC phenotype, the reactivation of lytic gene expression occurs from the previously quiescent genome, which ultimately results in the release of infectious virus progeny (d). The reactivation of lytic gene expression is concomitant with chromatin remodelling of the MIEP into a transcriptionally active state (f). Such chromatin remodelling of the MIEP allows the reactivation of viral lytic gene expression to occur, following specific differentiation of the CD34+ cells to mature DCs. Reproduced with permission from Reeves et al. (2005c)