Although an immune response to CMV is generated during the course of an acute infection, the virus has evolved a number of mechanisms in order to limit the immune reaction and thus ensure its survival. However, in doing so, CMV may also provoke a pathogenic response from the host. As stated above (section 1.6), CMV can directly infect cells of leukocyte origin, and an acute infection with CMV is often associated with a sustained general dysfunction of the immune system. Infection of mononuclear cells with CMV has been shown to profoundly alter their function in both humans and mice. Virus-induced suppression of macrophage accessory activities can lead to decreased lymphocyte proliferation, diminished natural killer cell activity, and impaired CTL responses to CMV-infected cells (Schrier, Oldstone, 1986; Buchmeier, Cooper, 1989), thus dampening the immune response to the virus. In immunologically incompetent individuals, however, this may further perturb the ability of the host not only to control the dissemination o f CMV, but also to control other infectious agents as well (Chatterjee et al. 1978). Similarly, whilst the ability of CMV to establish an infection in haematopoietic progenitor and stem cells without
lytic effect may elicit a replicative environment which is not subject to immune surveillance, decreased regeneration of bone marrow precursors may promote graft loss in bone marrow recipients (Mayer et al. 1997; Mutter et al. 1988). Hence, infection with CMV may further complicate the condition of already immunosoppressed patients.
CMV induces the expression of an FcyR on the surface of infected fibroblasts, in common with other members of the herpesviridae (Litwin et al. 1990; McTaggart et al. 1978; Furukawa et al. 1975). These FcyR’s may protect the infected cells from antibody-mediated destruction by binding the Fc portion of nonvirus- specific antibody, thereby impeding the attachment of specific antiviral IgG to the cell surface (Adler et al. 1978). Moreover, the simultaneous binding of virus-specific antibody to the viral antigen and to the virus-induced FcyR could prevent the functional engagement of the Fc portion of the antibody, thus preventing its participation in antibody-dependent cell mediated cytotoxicity, or antibody plus complement-mediated lysis (Dubin et al. 1991). Alternatively, whilst the induction of FcyR’s may protect the virus from antibody-mediated host immune defences, aberrant expression of FcyRs might also result in the binding of immune complexes to the infected cell, a phenomenon that could contribute to tissue injury (Oldstone, 1975) or to the initial lesions involved in the development of atherosclerotic plaques (Griffith et al. 1988; Raines, Ross, 1993). Furthermore, there is evidence to suggest that immune complex formation may occur during, or be responsible for, the initiation of chronic rejection in kidney allograft recipients (Hayry et al. 1992a; Hancock et al.
1993a), thus CMV might initiate the deposition of immune complexes via the induction of a FcyR, which could lead to the destruction of host tissues.
Another potential immune escape mechanism which is associated with CMV involves the modulation of MHC Class I and Class II molecules. It has been shown that a down-regulation of Class I molecules occurs at the surface of CMV-infected fibroblasts in vitro (Yamashita et al. 1993; Bames, Grundy, 1992), and a similar phenomenon has been reported during murine CMV infection (Campbell, Slater,
1994). In both cases, the viral expression of IE and early genes was necessary to promote the Class I down-regulation (Campbell, Slater, 1994; Ahn et al. 1996). The
interaction of early viral glycoproteins with the assembly of Class Lpeptide complexes rendered them unstable, and consequently incapable of being transported to the cell surface (Davis-Poynter, Farrell, 1998). Furthermore, transfection of a CMV-encoded MHC Class I homologue (pULlS) led to an inhibition of natural killer cell function in both mice and man (Farrell et al. 1997; Reybum et al. 1997). However, the expression of pULlS during a natural infection, either in vitro or in vivo, has yet to be reported. Analogously, it has been suggested that Class II molecules were prevented from being induced by IFN-y at the surface of endothelial cells following infection with CMV in vitro (Scholz et al. 1992), and inducible expression of Class II at the surface of murine macrophages was also prevented following infection with murine CMV (Heise et al. 1998a). This modulation in Class II expression was thought to prevent the interaction of CD4+ effector lymphocytes with the CMV-infected target cell by blocking the antigen presenting functions of the latter cells.
In vivo, however, the data to support reduced MHC distribution at the surface of CMV-infected cells are scant. In contrast, there are several reports suggesting that CMV infection augments the expression of Class I molecules at the surface of epithelial (van Dorp et al. 1993), smooth muscle (Hosenpud et al. 1991), and endothelial cells in vitro (Tuder et al. 1994; van Dorp et al. 1989). Similarly, CMV infection of thyroid follicular cells in vitro resulted in the increased expression of Class II molecules (Khoury et al. 1991), as did the in vitro infection of rat endothelial cells with rat CMV (Ustinov et al. 1993). Furthermore, adult rats undergoing an acute CMV infection in vivo displayed increased Class II antigens on endothelial cells in the kidney, liver and heart (Ustinov et al. 1994b), and the expression of these molecules was similarly augmented on tubular and endothelial cells of human renal allograft recipients undergoing an acute CMV infection in vivo (von Willebrand et al. 1986). However, it is unclear whether these in vivo effects were a result of direct viral infection, or mediated through the secretion of soluble inflammatory mediators.
Regardless of the mechanism involved, the increased MHC expression observed during CMV infection is interesting because MHC antigens are thought to be an important target in transplant rejection (Shoskes, Wood, 1994). Increased expression of Class I and Class II molecules has been with associated with the
initiation of rejection episodes in rat cardiac allografts (Milton, Fabre, 1985). In humans, the increase in Class II molecules is similarly associated with acute rejection episodes in kidney transplant recipients (von Willebrand et al. 1993; Hall et al. 1984), and with the development of accelerated atherosclerosis in cardiac transplant recipients (Labarrere et al. 1995). Similarly, increased Class I expression was observed on hepatocytes, bile duct epithelium and sinusoidal endothelium, and increased Class II expression on Kuppfer cells and sinusoidal endothelium during liver allograft failure (Gouw et al. 1988). The induced MHC expression on the allograft was thought to promote immune activation, by providing a greater concentration of target molecules on the donor cells (Shoskes, Wood, 1994). Thus, the increased expression of Class I and Class II molecules on various cell types during CMV infection may elicit CD8+ and CD4+ lymphocyte activation, leading to the generalised destruction of the allograft.
Graft rejection and accelerated atherosclerosis are also characterised by the infiltration of T-lymphocytes, monocytes (Hayry et al. 1992b) and neutrophils (Adams et al. 1990) into the blood vessels and tissues of the donor organ, and alterations to adhesion molecules have been associated with the extravasation of these leukocytes during both acute and chronic rejection. In cardiac transplant recipients, there was an increase of ICAM-1 and VCAM-1 expression on capillary venular endothelial cells associated with a T-lymphocyte infiltrate during acute allograft rejection (Briscoe et al. 1991), and an increase ICAM-1 induction on tubular cells during kidney rejection has been demonstrated in biopsy specimens (Brockmeyer et al. 1993). Interestingly, up-regulation of these molecules was also associated with CMV infection in solid organ transplant recipients (Koskinen, 1993; Lautenschlager et al. 1996). Increased expression of ICAM-1 and LFA-3 on hepatocytes, bile duct epithelium and endothelial cells of liver allografts undergoing severe rejection episodes has been demonstrated (Adams et al. 1989a; Steinhoff et al. 1993a), and these two adhesion molecules have also been shown to be up-regulated at the surface of fibroblasts following infection with CMV in vitro (Grundy, Downes, 1993).
The expression of E-selectin has been implicated, albeit to a lesser extent, as a predictor of rejection in both heart and renal allograft recipients (Ferran et al. 1993; Briscoe et al. 1995). This molecule has also been shown to be transiently up-regulated
during CMV infection in heart transplant recipients (Koskinen, 1993). The expression of E-selectin was also increased at the surface of endothelial cells following infection with CMV in vitro, and this has been implicated in promoting increased polymorphonuclear cell binding to the infected endothelium (Span et al. 1991a). Increased expression of ICAM-1 and VCAM-1 on CMV-infected endothelium was also implicated in promoting increased adhesiveness for T-lymphocytes and monocytes (Shahgasempour et al. 1997a). Thus, increased expression of adhesion molecules may result in a prolonged inflammatory reaction in the allograft following infection with CMV. However, in vitro, other authors have either failed to detect alterations in the expression of these molecules at the endothelial cell surface (Scholz et al. 1992), but rather have shown that infection with CMV prevents the cytokine inducible expression of VCAM-1 and E-selectin (Sedmak et al. 1994b). Thus the effect of CMV infection on the expression of both adhesion and MHC molecules at the surface of the endothelial cell is still controversial, and further studies are necessary to resolve these issues.