|Home | About | Journals | Submit | Contact Us | Français|
Here the authors discuss evidence in human and animal models supporting two opposing views regarding the pathogenesis of human immunodeficiency virus (HIV) in the central nervous system (CNS): (1) HIV infection in the CNS is a compartmentalized infection, with the virus-infected macrophages entering the CNS early, infecting resident microglia and astrocytes, and achieving a state of latency with evolution toward a fulminant CNS infection late in the course of disease; or alternatively, (2) events in the periphery lead to altered monocyte/macrophage (MΦ) homeostasis, with increased CNS invasion of infected and/or uninfected MΦs. Here the authors have reevaluated evidence presented in the favor of the latter model, with a discussion of phenotypic characteristics distinguishing normal resident microglia with those accumulating in HIV encephalitis (HIVE). CD163 is normally expressed by perivascular MΦs but not resident microglia in normal CNS of humans and rhesus macaques. In agreement with other studies, the authors demonstrate expression of CD163 by brain MΦs in HIVE and simian immunodeficiency virus encephalitis (SIVE). CNS tissues from HIV-sero positive individuals with HIVE or HIV-associated progressive multifocal leukoencephalopathy (PML) were also examined. In HIVE, the authors further demonstrate colocalization of CD163 and CD16 (FcγIII recptor) gene expression, the latter marker associated with HIV infection of monocyte in vivo and permissivity of infection. Indeed, CD163+ MΦs and microglia are often productively infected in HIVE CNS. In SIV infected rhesus macaques, CD163+ cells accumulate perivascularly, within nodular lesions and the parenchyma in animals with encephalitis. Likewise, parenchymal microglia and perivascular MΦs are CD163+ in HIVE. In contrast to HIVE, CD163+perivascular and parenchymal MΦs in HIV-associated PML were only associated with areas of demyelinating lesions. Interestingly, SIV-infected rhesus macaques whose viral burden was predominantly at 1 × 106 copies/ml or greater developed encephalitis. To further investigate the relationship between CD163+/CD16+ MΦs/microglia in the CNS and altered homeostasis in the periphery, the authors performed flow-cytometric analyses of peripheral blood mononuclear cells (PBMCs) from SIV-infected rhesus macaques. The results demonstrate an increase in the percent frequency of CD163+/CD16+ monocytes in animals with detectable virus that correlated significantly with increased viral burden and CD4+ T-cell decline. These results suggest the importance of this monocyte subset in HIV/SIV CNS disease, and also in the immune pathogenesis of lentiviral infection. The authors further discuss the potential role of CD163+/CD16+ monocyte/MΦ subset expansion, altered myeloid homeostasis, and potential consequences for immune polarization and suppression. The results and discussion here suggest new avenues for the development of acquired immunodeficiency syndrome (AIDS) therapeutics and vaccine design.
Human immunodeficiency virus type 1 (HIV-1)–associated dementia (HIV-D) is a syndrome of motor and cognitive dysfunction observed in approximately 5% to 10% of patients infected with HIV-1 with acquired immune deficiency syndrome (AIDS) (McArthur et al, 1993; Sacktor et al, 2001). The neuropathogenesis of HIV-D is not completely understood. HIV enters the central nervous system (CNS) early, during the acute phase of infection; however, the action of cytotoxic T cells eliminates productively infected cells. Later during the course of disease, productive infection of the CNS ensues, with the concomitant development of CNS disease. The source and mechanism of this latter infection of the CNS has been a matter of considerable debate. It has been unclear if virus recrudesces from a latent reservoir or, alternatively, there is a new invasion of virus-infected cells. In a previous review (Fischer-Smith and Rappaport, 2005), we discussed two models: the Trojan Horse Model and the Late Invasion Model. In the Trojan Horse Model, the virus enters the CNS early, and replicates at low levels as a reservoir separated from the periphery. A viral phenotype that is more virulent in the context of the CNS emerges, leading to the development of disease. In the Late Invasion Model, uncontrolled virus replication and resulting immune deficiency lead to alterations in the myeloid differentiation pathway, promoting the expansion of an activated monocyte subset that is capable of tissue invasion.
In support of the Trojan Horse Model, several studies suggest compartmentalization of HIV in the CNS through comparisons of viral quasispecies inside and outside the CNS compartment (Clements et al, 2002, 2005). Many of these studies have compared plasma virus with virus found in cerebrospinal fluid (CSF) (Cunningham et al, 2000; Stingele et al, 2001; Strain et al, 2005; Tashima et al, 2002). Such comparisons may reflect the differences in the cellular source of virus, rather than viral evolution per se, because plasma virus is derived from lymphoid tissues, involving both infected T cells and macrophages (MΦs), and CSF virus is derived from infected perivascular MΦs and microglia. Studies in simian immunodeficiency virus (SIV)-infected rhesus macaques model also provide some evidence in support of a compartmentalized infection. These studies demonstrate suppression of virus replication in the CNS as early as 21 days, without loss of CNS viral DNA, with virus replication recrudescing later during the course of disease in a rapid model of SIV encephalitis (SIVE) (Barber et al, 2004). MΦ invasion does indeed occur in this model, however, suggesting that microglial activation is not the sole contributor to CNS disease. MΦ invasion is not limited to the CNS in this model, and peripheral neuropathy is also a consequence of global MΦ activation and trafficking (Laast et al, 2007). Although these studies do not preclude the contribution of invading MΦs, the constant level of viral DNA in CNS suggested a major contribution of a latent CNS reservoir.
There are several lines of evidence, however, that support an alternative hypothesis, as depicted in the Late Invasion Model. It has been proposed that virus entry into the CNS is due largely to the trafficking of HIV-1–infected monocytes/MΦs from circulation (Meltzer et al, 1990). Accordingly, the number of total brain MΦs is dramatically increased in HIV encephalopathy (HIVE), the pathology of HIV-D, without additional evidence of local proliferation by these cells (Fischer-Smith et al, 2004; Glass et al, 1995). Furthermore, MΦs/microglia represent the principle productive reservoir of HIV1 infection in the CNS (Kure et al, 1991; Porwit et al, 1989; Pumarola-Sune et al, 1987; Rostad et al, 1987). The contribution of infected MΦs and microglial cells to neuronal injury through secretion of viral and host factors has been the subject of numerous reviews (Fischer-Smith and Rappaport, 2005; Gonzalez-Scarano and Martin-Garcia, 2005; Kaul et al, 2001).
The Late Invasion Model assumes that there is a great influence of HIV/SIV infection status within the peripheral compartment that contributes to the development of CNS disease. Previous studies comparing HIV-1 Gp120 sequences have demonstrated the greatest similarity between envelope sequences derived from brain with those derived from bone marrow and blood (Liu et al, 2000). The role of monocyte/MΦ trafficking from the periphery into the CNS is further supported by the beneficial effects of highly active antiretroviral therapy (HAART) despite poor CNS penetrance of most antiretroviral compounds (Vehmas et al, 2004).
Determining the origin of the significant number of brain MΦs in CNS with or without disease has been difficult because no single cluster of differentiation (CD) can conclusively discriminate between resident microglia and perivascular MΦs. As such, combined CD markers have been used to make this distinction. Perivascular MΦs are positive for CD14 (lipopolysaccharide [LPS] receptor) and CD45 (leukocyte common antigen [LCA]); however, microglia do not express detectable levels of these antigens by standard immunohistochemical methods (Ford et al, 1995; Sedgwick et al, 1993; Ulvestad et al, 1994; Williams et al, 1992). Previously, we identified two populations of activated MΦs in the CNS of patients with HIVE (Fischer-Smith et al, 2001). MΦs accumulating perivascularly are CD14+/CD45(LCA)+/CD16+ (FcγIII receptor), and appear to be the principal reservoir of productive HIV-1 infection in the CNS. Similar observations were also reported in SIVE (Williams et al, 2001). These cells are phenotypically similar to a subpopulation of monocytes reported to be expanded in patients with HIV-D, as compared to patients with HIV-1 infection without dementia and seronegative controls (Pulliam et al, 1997). Importantly, CD16+ monocytes preferentially harbor HIV in vivo, are more permissive to HIV infection than CD16− monocytes, and are likely important as reservoirs of infection and tissue dissemination (Ellery et al, 2007; Joworoski et al, 2007). Additionally, ramified cells with microglial morphology located within the brain parenchyma were also found to harbor productive HIV-1 infection, although to a lesser degree than MΦs located perivascularly (Fischer-Smith et al, 2001). These ramified cells are CD14−/CD45(LCA)−/CD16+ (Fischer-Smith et al, 2001).We previously suggested that based on their increased numbers and immunophenotype, these cells may represent activated resident microglia and/or MΦs that have recently migrated from the peripheral blood into the brain parenchyma, with apparent loss of CD14 and CD45(LCA) expression (Fischer-Smith et al, 2001). In support of the latter hypothesis, the increase in total brain MΦs associated with HIVE appears to be due to trafficking of monocytes/MΦs into the CNS from the periphery, rather than local microglial proliferation (Fischer-Smith et al, 2004).
In SIVE and HIVE, MΦs/microglia represent the only cells in the CNS to demonstrate productive SIV or HIV-1 infection, respectively. Microglia, the resident MΦ of the brain, populate the CNS during fetal development and for a brief period after birth, after which there is little or very slow turnover (Unger et al, 1993). In contrast, perivascular MΦs are more rapidly and continuously replaced by monocytes from circulation (Hickey and Kimura, 1988). CD163, a monocyte/MΦ-specific scavenger receptor for hemoglobin-haptoglobin complex (Kristiansen et al, 2001), is reported to be expressed by perivascular MΦs, but not resident microglia, in normal human CNS (Fabriek et al, 2005b; Rezaie and Male, 2003). CD163 expression has been identified previously in HIVE and SIVE CNS, but not in AIDS patients without CNS disease with little, and no expression was observed in variant Creutzfeldt-Jakob disease or Alzheimer’s dementia (Roberts et al, 2004). From these studies, the authors suggested that the CD163 expression seen in HIVE and SIVE represents a specific type of microglial activation seen only in some pathogen-induced inflammatory CNS conditions (Roberts et al, 2004).
In view of the disparity in the expression pattern of CD163 by perivascular MΦs versus microglia that has been previously reported (Fabriek et al, 2005a), together with the increases in total MΦs and microglia we previously reported in HIVE (Fischer-Smith et al), we sought to revisit this issue with the notion that the accumulation of CD163+ “microglia” in HIVE and SIVE occur as a consequence of monocyte/MΦ trafficking and the adaptation and engraftment of peripheral derived MΦs within the brain parenchyma.
In agreement with previous reports (Kim et al, 2006; Roberts et al, 2004), we observed significant CD163 accumulation within the perivascular cuff and nodular lesions in HIVE CNS (Figure 1A, panels F and G). Additionally, abundant CD163 expression was seen in cells in the brain parenchyma, many having ramified microglial morphology (Figure 1A, panel C). In consideration of other reports that demonstrate CD163 expression by perivascular MΦs but not resident microglia in normal CNS (Fabriek et al, 2005b; Rezaie and Male, 2003), CD163 expression by parenchymal microglia is in agreement with our previously reported findings suggesting that the large numbers of ramified microglia accumulating in HIVE are peripheral blood derived (Fischer-Smith et al, 2004; Fischer-Smith et al, 2001). Furthermore, the proposed “activation” of CD163 expression in microglia in HIVE would not explain the increase in the numbers of MΦs/microglia we observed in HIVE CNS (Fischer-Smith et al, 2004), without evidence for proliferation. In contrast to HIVE, brain tissues derived from seronegative and HIV-infected individuals exhibit few CD163+ perivascular cells and the absence of detectable CD163+ parenchymal cells (Figure 1A, panels A, B, D, and E).
As we had observed and reported previously, CD16+ cells are seen in the parenchyma and perivascular cuffs of HIVE CNS (Figure 1B, panels A and D). Colocalization studies revealed that the majority of these CD16+ cells also expressed the perivascular MΦ marker, CD163 (Figure 1B, panels C and F). CD16+ monocytes may be important to the development of HIV-D and HIVE where the expansion of this subset was demonstrated in patients with AIDS/dementia over AIDS patients without dementia and seronegative individuals (Pulliam et al, 1997). CD16+ monocytes exhibit features of tissue MΦs and are more phagocytic and express high levels of inflammatory cytokines (Scherberich and Nockher, 2000). It was suggested that these cells are more invasive and thus able to enter the CNS compartment in HIV-D (Pulliam et al, 1997). In support of this hypothesis, we reported a significant accumulation of CD16+ cells in the CNS of patients with HIVE (Fischer-Smith et al, 2001). In our current study, we find CD163 colocalizes with CD16 to a significant degree throughout the CNS in HIVE. Enhanced CD163 expression on CD16+ monocytes may contribute to the invasiveness of these cells, as CD163 has been demonstrated to augment monocyte adherence to LPS or cytokine-stimulated endothelial cells (Wenzel et al, 1996). Interestingly, the CD16+ monocyte subset reportedly shows the highest CD163 expression of all human monocyte subsets (Buechler et al, 2000).
Additionally in HIVE CNS, many CD163+ cells were found to harbor productive infection, as indicated by colocalization of CD163 and HIV-1 p24. The majority of these cells were found within the perivascular space and in nodular lesions (Figure 1C, panels F and I). HIV-1 p24 positivity was also observed in a number of CD163+ cells located in the brain parenchyma; however, some CD163+ cells do not appear to harbor productive infection (Figure 1C, panel C).
To gain additional understanding into the presence of the considerable number of CD163+ cells in the CNS of patients with HIVE, we investigated brain tissue from patients with HIV-1 infection/AIDS with progressive multifocal leukoencephalopathy (PML) for CD163 expression. PML is a demyelinating disorder of the CNS associated with the polyomavirus, JC virus. In the CNS, JC virus infects and replicates in oligodendrocytes, ultimately destroying the infected cell, resulting in regions of myelin loss (lesions) in multiple areas of the brain (for review, see Wyen et al, 2005). In our studies, we found significant CD163 positivity by perivascular MΦs, as well as MΦs/microglia located in the brain parenchyma; however, these CD163+ cells were primarily localized to within the area of the PML lesion(s) (Figure 1D, panels A and C). Additionally, parenchymal CD163+ cells do not demonstrate the microglial morphology we see in HIVE (Figure 1A, panel C), but had a round and foamy appearance (Figure 1D, panel A). In areas outside of PML lesions, CD163+ cells were not seen in the parenchyma (Figure 1D, panel B) and only few cells were observed perivascularly (Figure 1D, panel D). This is in stark contrast to what was observed in HIVE CNS tissues, which demonstrated a generalized CD163 positivity throughout the white matter in tissues associated with frontal cortex and basal ganglia (Figure 1A, panels C, F, and G). Presumably, the CD163+ foamy MΦs seen in PML originated as monocytes, which were recruited into the CNS to areas of involvement. They become lipid-laden by clearing up myelin released from the infected oligodendrocytes, resulting in their foamy appearance. In support of this hypothesis, CSF from PML patients has demonstrated a significantly higher concentration of monocyte chemoattractant protein-1 (MCP-1), a chemokine involved in chemotactic migration of monocytes, when compared to HIV-1–infected individuals without PML and seronegative controls (Marzocchetti et al, 2005). Interestingly, CD163 expression by foamy MΦs has also been observed in multiple sclerosis (MS) (Fabriek et al, 2005b), a demyelinating CNS disease also involving monocyte recruitment. Also in MS, MCP-1 expression has been observed in astrocytes and MΦs within acute lesions (Simpson et al, 1998).
In addition to our studies in human tissues, we expanded our studies to include examination of uninfected and SIVmac251-infected rhesus macaques with and without CNS complications for alterations in CD163+ peripheral blood monocyte subsets by flow cytometry and assessed how these changes might correlate with the development of CNS disease. These studies were performed in eight SIVmac251-infected animals during the natural course of SIV disease. An additional two uninfected animals and eight SIVmac251-infected animals treated with antiretroviral therapy (ART) (PMPA and FTC, with varying degrees of success) were included in the blood monocyte studies.
Of the eight SIV-infected animals that were followed for up to 1 year without treatment, four developed SIVE as determined by immunohistochemical examination. In situ hybridization using an RNA probe against SIVmac239 identified productively infected cells in animals with encephalopathy (Figure 2A, panels C, F, and G) but not in SIV-infected animals without CNS disease (Figure 2A, panels B and E). Similar to that seen in HIVE, the majority of SIV+ cells form perivascular cuffs and nodules (Figure 2A, panels F and G) with few positive cells seen in the parenchyma (Figure 2A, panel C). Our immunohistochemical studies would suggest that these cells are also largely CD163+ (Figure 2B, panels F and G) and may support the hypothesis that the virus seen in the CNS in SIVE has recently entered the CNS compartment from the peripheral blood.
In view of the observation that CD163 is not normally expressed by resident microglia in normal brain tissue (Fabriek et al, 2005a), the numerous ramified CD163+cells we observe in SIVE and HIVE suggest a mechanism involving emigration of MΦs into the CNS compartment from the peripheral blood and adaptation to microglial morphology. We have not excluded, however, the possibility that CD163 expression in microglia represents an unusual state of microglia activation. Indeed, CD163 expression can be induced by the anti-inflammatory mediators, glucocorticoids and interleukin (IL)-10, as well as by the proinflammatory cytokine, IL-6 (Buechler et al, 2000; Hogger et al, 1998; Sulahian et al, 2000). Additionally, peripheral blood monocytes differentiated to MΦs in the presence of macrophage colony-stimulating factor (M-CSF) have shown increased CD163 mRNA and protein expression (Buechler et al, 2000). M-CSF activation by HIV infection in MΦs (Gruber et al, 1995) may promote alterations in monocyte/MΦ homeostasis. Indeed, M-CSF levels in plasma and CSF correlate inversely with time to death in patients with advanced HIV disease (Sevigny et al, 2007).
To begin to investigate the relationship between alterations in circulating monocyte subsets and the development of CNS disease, we performed flow cytometric studies on isolated peripheral blood mononuclear cells (PBMCs) in SIVmac251-infected rhesus macaques during the chronic phase of SIV infection. These studies were performed to investigate potential alterations in CD163+ monocyte subsets and associated changes in viral burden. This analysis included PBMCs from untreated animals, as well as PBMCs from additional SIVmac251-infected animals treated with antiretroviral therapy (ART). Data were collected at days 154, 168, and 196 post challenge Monocytes were identified first by forward and side scatter parameters, then by CD14 expression. In our studies we found that, like humans, rhesus macaques express CD163 by the majority of monocytes with a subset of these expressing CD16. This CD163+/CD16+ monocyte (CD14+) subset was expanded in all chronically infected animals with detectable virus as compared to those with undetectable viral loads and seronegative animals, and this expansion was found to correlate with a greater viral load (Figure 3, panel A). Interestingly, animals with CNS disease at necropsy generally maintained very high viral loads (>1×106 copies/ml) throughout infection (Graph 1). Additionally, the expansion of this monocyte subset was also found to correlate inversely with the number of CD4+ T cells in animals with counts below 1100 cells/µl (Figure 3, panel B). These data suggest that CD14+/CD163+/CD16+ monocytes may play a role in virus production and disease progression.
Monocyte/MΦs exhibit functional and phenotypic heterogeneity, which is influenced largely by the surrounding cytokine environment (Gratchev et al, 2006; Park-Min et al, 2005; Porcheray et al, 2005). MΦs can exhibit immune polarization representing MΦs involved in promoting inflammation or ‘proinflammatory MΦs’ (type 1 MΦ or classically activated MΦ) and those involved in resolving inflammation or ‘anti-inflammatory MΦs’ (type 2 MΦ or alternatively activated MΦ) (Gratchev et al, 2006; Porcheray et al, 2005; Van Ginderachter et al, 2006). The CD163 positivity of the MΦs accumulating in the perivascular and parenchymal region in the CNS in HIVE and SIVE might suggest that these cells represent alternatively activated MΦs (Komohara et al, 2006). It is possible that these MΦs in the context of HIV and SIV infection exhibit a unique gene expression program. It is tempting, however, to speculate that alterations in monocyte/MΦ homeostasis (as illustrated by our immunohistochemical and flow cytometric studies) contribute not only to the pathogenesis of HIV in CNS, but also to T-cell immune dysfunction in HIV infection leading to AIDS. It is interesting to note that in previous studies using a recombinant SIV/HIV virus designated SHIV ku-2, increased MΦ tropism (despite CXCR4 utilization) was associated with increased virulence, CD4 depletion, and CNS and renal diseases (Liu et al, 1999). Additional studies will be required to characterize the phenotypic and functional characteristics of monocyte/MΦs in the context of HIV and SIV infection. Such studies should open additional avenues for therapeutic intervention.
This work was supported by NIH/NINDS grants R01 NS047031 to J.R. and P01 NS30916-09 to J.R. and K.K. T.F.S. was supported by training grant (T32-DA07237) from NIH/NIDA
Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf
This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden.
Publisher's Disclaimer: The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.