Virus-host interaction in the simian immunodeficiency virus-infected brain

Happy is he who gets to know the reasons for things (Virgil) Dedicated to the work, spirit and life of Opendra (Bill) Narayan

Introduction

With the onset of the AIDS pandemic, research on the pathobiology of lentivirus infections became paramount. Lessons learned from studies on other lentiviruses, such as Bill Narayan’s work in visna virus, revealing its persistent infection of the brain (Narayan et al, 1974) and development of viral variants (Narayan et al, 1977), were quite instructive for HIV research. His past research enabled other researchers not only to stand on the shoulders of the giant he was, but as he then set his great skills to work on HIV and its monkey counterpart, SIV, to gain needed insight from his studies on these viruses.

Bill Narayan’s accomplishments have had a profound influence on the studies performed by our lab and others on the effects of SIV on the brain of monkeys, modeling neuroAIDS in humans. One of the concepts we have focused on is best stated in his review on lentiviral diseases of ungulates: “Consistent low-grade viral replication sets the pace for disease by providing continuous antigenic stimulation for the inflammatory cellular immune response…” (Narayan and Cork, 1985). Indeed Bill’s work had nicely shown in rodents that the host response can be a major effector in central nervous system (CNS) dysfunction following viral infection (Narayan et al, 1983). Another concept concerned the nature of CNS infection itself, with the particular characteristics of both the virus and the target cell, studies he developed quite well in the SIV-infected monkey system.. Again, to directly quote from his writings, his lab found that “different macrophage populations in the body may select specific phenotypes of lentivirus from the quasispecies of virus … ” (Sharma et al, 1992).

There are many good reviews of the SIV model for HIV neuropathogenesis (Buch et al, 2004; Burudi and Fox, 2001; Kim et al, 2005; Sopper et al, 2002; Zink and Clements, 2002). Here we will review how our group has applied the lessons learned from the studies of Bill Narayan and others to uncover unique aspects of the virus-host interaction resulting from SIV infection of the nonhuman primate brain.

CNS Virus

The majority of viral strains and molecular clones utilized in monkey neuropathogenesis studies originate from the SIVmac251 virus stock, the first described pathogenic nonhuman primate model for AIDS (Letvin et al, 1985). This stock infects both CD4+ T cells and macrophages, and in a subset of animals results in a distinctive pathology, SIV encephalitis, resembling HIV encephalitis. Although a molecular clone (SIVmac239) derived from this stock was found not to infect macrophages, replacement of parts of its genome with sequences from other SIVs as well as HIV, followed in some cases by serial passage through bone marrow or brain, could restore macrophage and CNS infectivity (Anderson et al, 1993; Mankowski et al, 1997; Raghavan et al, 1997; Westmoreland et al, 1998; Zink et al, 1999).

Although some controversy exists over additional targets, myeloid-derived cells –macrophages and microglia – are the primary and predominant infected cells in the brains of HIV-infected people and SIV-infected monkeys. In order to enrich for a viral stock capable of infecting these cells in the brain, we isolated microglia from a chronically, SIVmac251-infected animal, and inoculated a naïve animal with these cells intravenously. This serial passage was repeated, and resulted in a high rate of encephalitis in the recipients (Watry et al, 1995).

Initial characterization of the changes in the viral genome that were induced by the microglia passage focused on the Env gene. Distinct variants were found, either through selection of subspecies present in the original SIVmac251 stock or through mutation, which were then maintained in subsequent passages (Lane et al, 1995). Further analysis from an additional microglial passage, including the Nef gene and the 3′ LTR in addition to Env, revealed the convergent evolution and uniqueness of the resulting virus when compared to other macrophage-tropic and/or neurovirulent SIVs (Gaskill et al, 2005). In this study, molecular clones derived from the microglia-passaged virus were also characterized for in vitro infectious phenotype. Differences were found when compared to the parental stock, in that the microglia-passaged virus showed a robust early virus production on macrophages and microglia, compared to a much lower level of infection for the parental SIVmac251 stock (Gaskill et al, 2005). Such an infectious phenotype may aid in both the establishment and spreading of viral infection in the brain.

While many study the viral genome for the basis of changes in viral properties, changes in the virion, such as host modification of viral proteins or virion incorporation of host proteins, can also affect the phenotype of infection. Since HIV/SIV in the brain is largely derived from macrophages, we examined whether the infectious phenotype of virions produced in macrophages might differ from those produced in T cells. Such virions showed no difference in genomic RNA/Gag or Env/Gag ratios, and sequence analysis revealed no mutations resulting from production in the different cell types. Infectivity assays, using a number of in vitro and in vivo derived targets, revealed a significantly higher infectivity of macrophage derived virions (Gaskill et al, 2008). Since both Env and host proteins incorporated into virions can be glycosylated, and virion glycosylation can affect infectivity, we then assessed the effect of treating virions with glycosidases. Treatment of viral particles with mannosidase, but not neuraminidase, increased the infectivity of virions derived from T cells, but did not affect the infectivity of virions derived from macrophages (Gaskill et al, 2008). Thus the greater infectivity of macrophage derived virions may be linked to a lower number of mannose residues on proteins in the virions themselves. The 3 to 10 fold greater infectivity of macrophage derived virions, amplified over even a few rounds of viral replication, can result in an enormous increase in the spread of infection in the brain, where macrophages are the primary source and target for HIV infection.

In addition to enriching for a neuroinvasive, neurovirulent virus, we have also used a serial passage technique to address another problem in the field, the shortage of rhesus monkeys available for AIDS research. Investigators typically have limited their studies to Indian rhesus monkeys, as some data indicated that the course of SIV-induced disease might differ in monkeys from other provenances, such as China. However such Indian monkeys are in extremely short supply. Recognizing that, and the fact that the SIVmac251 virus itself was itself derived from a serial passage through Indian-derived monkeys (Mansfield et al, 1995), we performed a serial passage of our microglia-derived stock through Chinese rhesus monkeys (using plasma). Indeed, a virulent stock was derived, capable of inducing simian AIDS with CNS disease in China-derived monkeys (Burdo et al, 2005).

CNS Host Response

The innate and adaptive responses to HIV and SIV infection in the CNS play a large role in the course of CNS disease, but are less well characterized than the responses found in the blood and lymphoid organs. Virus enters the brain within two weeks of infection, and if an effective systemic immune response occurs, becomes relatively quiescent until end-stage disease, when in a fraction of people and animals pronounced neurological symptoms can occur accompanied by encephalitis. In monkeys, one can shorten the course of infection and obtain a rapid CNS disease through the use of combinations of viruses, which lead to pronounced CD4+ T cell depletion, high viral loads, and CNS disease (Mankowski et al, 2002), or by depletion of CD8 cells, which inhibits the immune response and also results in high viral loads and CNS disease (Madden et al, 2004; Roberts et al, 2003; Schmitz et al, 1999; Williams et al, 2005).

We have performed a molecular examination of the different stages of SIV infection of the brain by gene array transcriptional profiling. During the acute infection (2 weeks after inoculation) a large number of genes in the interferon and interleukin-6 responsive pathways were upregulated (Roberts et al, 2004a), likely representing the innate CNS response to viral invasion and infection. Indeed others have identified interferon-beta as a causative agent in the induction of viral transcriptional latency in the brain (Barber et al, 2006). During chronic infection in otherwise healthy, asymptomatic animals (9 months after inoculation), expression of only a few host genes was found to significantly differ in the brains of SIV infected animals. These upregulated genes represent discrete aspects of immune responsiveness (Roberts et al, 2006). Examination of the expression of these genes in the brain over the varied stages of disease revealed that CCL5 (RANTES) was upregulated throughout infection, and was produced by brain-infiltrating CD8 lymphocytes (Roberts et al, 2006). CCL5 can exert both protective and damaging effects in the infected brain, and exemplifies the duality of controlling HIV/SIV in the brain – activating processes that eliminate or restrain the virus and virally infected cells, while needing to minimize damage to the brain. Finally, during SIV encephalitis (induced in a rapid disease course by CD8 depletion) a wide range of genes are upregulated, again with a substantial proportion representing immune related processes (Roberts et al, 2003). Interestingly, the expression of these molecules localized to endogenous brain cells in addition to infiltrating immune cells (Roberts et al, 2003).

These studies not only afforded a comprehensive assessment of CNS gene expression, but have also have provided clues to mechanistic insights in HIV/SIV neuropathogenesis. For example, we have recently focused on osteopontin, which was identified in our gene array studies as being upregulated in the brain in SIVE (Roberts et al, 2003). Although osteopontin has been suggested to function as a chemokine, we found its ability to induce cell accumulation is due to other properties. Using an in vitro model system where monocytes cross an endothelial layer, osteopontin inhibited monocytes from reverse transmigrating back across the endothelial cells, the equivalent of leaving a tissue such as the brain (Burdo et al, 2007). In addition, we found that osteopontin is also an anti-apoptotic factor for monocytes (Burdo et al, 2007). Thus osteopontin expression in the brain can lead to accumulation of macrophages, which was found to be the best pathological correlate of HIV-induced CNS dysfunction (Glass et al, 1995). Interestingly, our recent work in monkeys reveals that both plasma osteopontin (Burdo et al, 2008), and one of its receptors on monocytes, CD44v6 (Marcondes et al, 2008), are increased in animals developing SIV encephalitis, and that in HIV-infected people, osteopontin levels increase proportionally to the severity of CNS dysfunction (Burdo et al, 2008).

A second molecular clue to HIV/SIV neuropathogenesis was provided by our finding of increased expression of CD163, the monocyte/macrophage receptor for the haptoglobin-hemoglobin complex. In SIV encephalitis, we found increased expression of CD16e, which localized not only to CNS macrophages but to microglia cells (Roberts et al, 2003). We also examined brains with HIV encephalitis, Alzheimer’s disease, and variant Creutzfeldt-Jakob disease, and found CD163 expressing ramified microglia only in SIV and HIV encephalitis (Roberts et al, 2004b). Recent work by others has revealed that expression of CD163 could be induced on microglia by the haptoglobin-hemoglobin complex, and such a complex could be found in the brain during SIV encephalitis, suggesting that such expression marks breakdown of the blood-brain barrier in this condition (Borda et al, 2008).

The adaptive immune response to the chronic infection of the brain has been a long-standing interest of our lab. During the chronic, asymptomatic phase of infection, through behavioral and neurophysiological testing, we find functional abnormalities of the CNS in the absence of observable CNS histopathology. However, in keeping with the above results of upregulated immune molecules in the chronically infected brain, we find infiltrating lymphocytes in the infected CNS, specifically CD8+ T cells (Marcondes et al, 2007; Marcondes et al, 2001; Marcondes et al, 2003; Roberts et al, 2006; von Herrath et al, 1995). These represent, in large part, SIV-specific cytotoxic T lymphocytes (CTL), demonstrated through SIV protein specific in vitro killing assays (von Herrath et al, 1995) and by the use of SIV epitope specific tetramer reagents (Marcondes et al, 2007).

These CTLs have enabled us to identify unique aspects of adaptive immunity in the CNS. An enrichment of SIV-specific CTLs is found in the brain, compared to the blood, lymphoid and other organs (Marcondes et al, 2007). In addition, certain CTL specificities, both in terms of antigen recognition and T cell receptor clonal repertoire, could be found in the brain, but not in the rest of the body (Marcondes et al, 2007; Marcondes et al, 2003). This accumulation and persistence is due to the specialized nature of immune interactions in the brain, in particular the cytokine environment. Following SIV infection, astrocyte expression of IL-15 is increased in the brain. This increase in IL-15, in the absence of IL-2, creates an environment conducive to CTL persistence, confirmed by in vitro studies (Marcondes et al, 2007). As with the molecules we find increased in infected brains, these CTL represent double-edged swords – necessary to control the virus, but with a high capacity to induce bystander damage to the surrounding brain tissue.

CNS Function

Neurological abnormalities are often apparent in animals that rapidly progress to end-stage simian AIDS with the development of SIV encephalitis, and can be documented with behavioral and neurophysiological testing (Gray et al, 2006; Marcario et al, 1999; Raymond et al, 1998; Raymond et al, 2000; Weed et al, 2003). However an early study in SIV-infected monkeys documented behavioral/cognitive abnormalities during the asymptomatic phase of infection showing a normal time course (Murray et al, 1992). We have utilized a number of testing modalities, including brainstem and cortical sensory evoked potentials, motor and cognitive skill tasks, and body movement assessment. These indeed revealed that CNS abnormalities are common in otherwise asymptomatic SIV-infected rhesus monkeys (Gold et al, 1998; Horn et al, 1998; Marcondes et al, 2007; Marcondes et al, 2001; Prospero-Garcia et al, 1996; Roberts et al, 2006; Weed et al, 2004). Although often the specific functional deficits can differ between the animals, we have found two of these tests: a bimanual motor task, used to measure manual dexterity, procedural learning and motivation to work for a preferred food reinforcer (raisins); and electrophysiological measurement of brainstem auditory evoked potentials, are quite sensitive and reproducible to assess the CNS deficits induced by SIV infection in the chronic, asymptomatic, as well as he later symptomatic stages of disease.

CNS functional abnormalities are more profound in animals with SIV encephalitis. Interestingly, analysis of the circadian rhythms of body temperature and movement in monkeys that developed SIV encephalitis revealed impairments in circadian rhythms that preceded clinical symptoms, and which became more severe with the progression of disease (Huitron-Resendiz et al, 2007). Interestingly circadian abnormalities have also been documented in HIV infection (Rondanelli et al, 1997; Swoyer et al, 1990; White et al, 1995), and circadian alterations can lead to a number of cognitive and behavioral abnormalities. Examination of the hypothalamus in animals with SIV encephalitis and circadian abnormalities reveals increased macrophage accumulation, microglia activation, and the presence of viral gene expression (Huitron-Resendiz et al, 2007). Our previous studies revealed the dominance of interferon-induced genes in the brains of animals with SIV encephalitis (Roberts et al, 2003). Furthermore, others have found that interferon treatment of mice induced altered circadian rhythms and increased expression of interferon-induced genes in the hypothalamus (Ohdo et al, 2001). Therefore, interferon or other immune mediator effectors are prime candidates in causing circadian dysfunction.

The link between the CNS immune findings and the CNS functional abnormalities is not clear, the presence of both increased immune response related gene expression and accumulation of SIV-specific CTL are certainly correlated with the functional abnormalities in the chronic, asymptomatic stage, and greatly increased inflammatory gene expression and macrophage accumulation in encephalitis at end-stage CNS disease. Yet other contributing mechanisms, such as the actions of viral proteins, remain potential mechanisms leading to functional deficits.

Summary

HIV, and its nonhuman primate counterpart, SIV, provides a unique opportunity to study the effect of a chronic viral infection of the CNS. Aspects of both the virus and the host response are noteworthy in the brain. Viruses in the brain have specific genotypic and phenotypic properties that enable efficient infection of macrophage/microglia, enabling the maintenance of a chronic infection. The host response within the brain to infection is also distinctive. Following the acute phase, where the brain’s innate responses occur, a chronic virus-host interaction occurs, in which the adaptive immune response is active in the brain. With the development of immunodeficiency, unchecked viral replication, accompanied again by an innate-response picture, can occur in the CNS of a proportion of subjects. Although this end-stage disorder can lead to severe CNS symptoms, the finding of CNS disorders during the chronic stage, both in the SIV model as well as in HIV-infected humans (Antinori et al, 2007), point to the need to better understand this aspect of the pathogen-host relation in the brain, which, over the now prolonged course of HIV infection, is resulting in increased morbidity in infected individuals.