HCV Replication Found in CNS in HIV+
Pathogenesis of Hepatitis C Virus Coinfection in the Brains of Patients Infected with HIV
The Journal of Infectious Diseases July 15, 2007;196:361-370
Scott Letendre,1 Amy D. Paulino,2 Edward Rockenstein,2 Anthony Adame,2 Leslie Crews,3 Mariana Cherner,4 Robert Heaton,3,5 Ronald Ellis,2 Ian P. Everall,4 Igor Grant,4,5 Eliezer Masliah,2,3 and the HIV Neurobehavioral Research Center Groupa
Departments of 1Medicine, 2Neurosciences, 3Pathology, and 4Psychiatry, University of California, San Diego, and 5Veterans Affairs Healthcare System, La Jolla
â?oâ?¦.supports the possibility that productive HCV infection might be present in the brains of these personsâ?¦. replicating HCV capable of contributing to neurological damage is present in patients infected with HIVâ?¦. the present results demonstrate the presence of HCV antigens in the CNS of HIV/HCV-seropositive subjects, which supports the contention that HCV can promote neurological damage and is an important comorbidity factor in patients with AIDS.â?
Involvement of the nervous system by human immunodeficiency virus (HIV) continues to be a serious problem. Among individuals with HIV who have a history of illicit drug use, those coinfected with hepatitis C virus (HCV) are a fast-growing population. However, few studies have assessed the penetration of HCV into the central nervous system (CNS) and its clinical and neuropathological impacts on HIV-infected individuals. For this purpose, the distribution of HCV was investigated in the brains of patients infected with HIV. The presence of HCV RNA in the CNS as detected by nested polymerase chain reaction was associated with a history of methamphetamine use, considerable antemortem cognitive impairment and abundant astrogliosis, and less-severe HIV encephalitis. HCV antigens were detected by immunoblot analysis, using heparin-purified brain samples, and HCV immunoreactivity was detected in astrocytes and in macrophage-microglial cells. The results support the hypothesis that HCV traffics into the HIV-infected brain, where it might lead to a productive coinfection associated with cognitive impairment.
Nervous system involvement of HIV continues to be a serious problem, despite the development of more-potent combination antiretroviral therapy [1-3]. This may be related to the use of poorly penetrating antiretroviral therapy [4, 5], the emergence of resistant HIV species [6, 7], and an increase in the use of drugs such as methamphetamine . Drug users, one of the faster growing populations with HIV, often have comorbid conditions leading to neurological complications [9, 10].
Of the cofactors that might play an important role in the neuropathogenesis of HIV, hepatitis C virus (HCV) infection, deserves special attention because of its high prevalence in the drug-using population [11, 12]. HCV belongs to the Flaviviridae family, which includes several neurotropic viruses [13, 14]. There has been recent interest in the possibility of a link between chronic HCV infection and cognitive impairment independent of liver failure, because HCV sequences have been detected in cerebrospinal fluid (CSF) [15, 16] and brain tissue [17, 18] and because proton magnetic resonance spectroscopy has shown significant metabolite alterations [19, 20]. Plasma HCV levels are associated with the severity of memory impairment and immune activation [11, 21]. Furthermore, HIV infection can facilitate extrahepatic HCV replication, and both viruses are present in monocytes/macrophages , leading to considerable dysregulation of immune responses and interferon- production [23, 24]. HCV augments cognitive deficits associated with HIV infection and methamphetamine use [11, 21, 25], and neuropsychiatric dysfunction has been linked to HCV infection in HIV-positive patients . These studies have suggested that HIV and HCV might interact in the central nervous system (CNS) to promote greater neurological damage. However, it is unclear which cells HCV infects in the brain, whether HCV antigens can be detected in the CNS, and which brain regions are selectively affected.
For this reason, the distribution of HCV was investigated in the brains of well-characterized patients infected with HIV. Remarkably, we demonstrate, using heparin column preparations, the presence of HCV antigens in the CNS of patients infected with HIV. HCV immunoreactivity was present in astroglial cells and perivascular macrophages. These results support the view that HCV traffics into the CNS of patients infected with HIV and that the presence of HCV antigens is associated with cognitive impairment.
In the present study, we show, using heparin columns, that, in addition to HCV RNA, HCV antigens can be detected in the brains of patients infected with HIV. Recent evidence has indicated that HCV replicates in extrahepatic organs such as the brain and might contribute to neurological and psychiatric manifestations in HCV-infected patients [35, 36]. These studies have also suggested that low-level HCV replication in the brain might be more likely in the setting of immunosuppression associated with HIV infection than in immunocompetent patients [17, 36-38].
Supporting this possibility, HCV RNA has been detected in the CSF of HIV-positive patients [15, 39], HCV sequences in serum differ from those in the CNS , negative-strand RNA has been identified in the CNS of HIV-positive patients [16, 17], and viral sequences with evidence of tissue compartmentalization have been amplified in postmortem brain samples . However, it has been difficult to demonstrate the presence of HCV in the CNS, probably because of low levels of replication, the presence of cross-reacting proteins, and the possibility that a large proportion of HCV antigens might form complexes with IgGs.
Consistent with the findings of previous studies [17, 40], we have shown by nested restriction site-specific PCR and qRT-PCR that all the HIV/HCV-positive subjects had HCV RNA in the CNS, whereas no HIV-positive/HCV-negative subjects had HCV RNA in the brain. Remarkably, in 85% of the HIV/HCV-positive subjects, the HCV NS5A and NS3 proteins and the core antigen were detected by immunoblot analysis after brain homogenates were semipurified with heparin columns, which supports the possibility that productive HCV infection might be present in the brains of these persons. The heparin chromatography method was recently developed to purify HBV and HCV from serum samples  and represents an important step in the development of new approaches for the detection of viral antigens.
Of the antibodies against HCV, NS5A-specific antibodies rendered the most consistent results. This might suggest that this HCV protein could be more abundantly expressed in the CNS, that antibodies against NS5A might have higher specificity or affinity for the protein, or that it is more difficult to retrieve or unmask HCV antigens other than NS5A. More detailed studies will be necessary to better understand the significance of this finding.
The NS molecules (2, 3, 4A, 4B, 5A, and 5B) have been considered to function in the replication of HCV subgenomic RNA . NS5A is an integral part of the replication complex of the virus [42, 43]. The detection of NS5A, NS3, and core antigen in brain samples of our HCV-positive subjects, in combination with previous studies detecting negative-strand RNA in the brains of HCV-positive patients , supports the contention that replicating HCV capable of contributing to neurological damage is present in patients infected with HIV.
Several studies have shown that nonhepatic cells such as peripheral blood mononuclear cells (PBMCs) and lymphocytes are capable of supporting HCV replication [40, 42, 44], which suggests that HCV might traffic into the CNS in macrophages and might infect microglia [43, 45]. In this regard, and consistent with the immunoblot studies, immunocytochemical analysis showed that HCV immunoreactivity was present in glial cells and macrophages only in HCV-positive subjects. We found that most of the HCV-immunoreactive cells were astroglia in white matter and, on occasion, in the neocortex and basal ganglia. However, we also detected HCV-immunoreactive perivascular macrophages and microglia. This is in agreement with the results of a previous study that showed by laser microcapture that CD68+ cells in the CNS display HCV positive and negative RNA strands as well as NS3 immunoreactivity [43, 45].
Compared with the report by Adair et al. , we found more-abundant HCV immunoreactivity associated with astroglia than with macrophages/microglia. This difference might be related to protocols for tissue processing, strategies for exposing HCV antigens, antibody specificity, and population used. For the present study, all included subjects were HIV infected, and previous reports have suggested that HCV and HIV influence each other's replication [22, 46].
The presence of HCV immunoreactivity in astroglial cells raises several questions regarding the potential mechanisms involved, because HCV might reach the brain via PBMCs and might not directly infect neural cells. One possibility is that HCV-infected macrophages might shed viral particles or antigens that, in turn, might be taken up by astroglia; another is that viral elements might be exchanged through cell junctions between microglia and astroglia. Similar mechanisms have been suggested for HIV [47, 48]-recent studies have shown the presence of HIV antigens in astroglial cells. Alternatively, the presence of HCV immunoreactivity in the astroglia of HIV/HCV-positive subjects might represent cross-reactivity with an unknown host antigen. However, this possibility is less likely in view of the specificity of the immunoblot studies, the reproducibility of the results with antibodies against NS5A and core antigen from different sources, and the negative results in HCV-negative subjects. Additional in vitro and in situ hybridization studies to detect negative-strand HCV RNA will be necessary to clarify the mechanisms at play.
The mechanisms of HCV toxicity in the CNS of patients infected with HIV are not completely clear; however, if HCV indeed affects astroglia, this is of significance, because these cells regulate glutamate levels at the synapse via glutamate transporters. In support of this possibility, previous studies have shown that, in patients with HCV infection, glutamate transport is defective . Similarly, previous studies have shown that HIV proteins such as gp120 interfere with the astroglial EAAT2 glutamate transporter . This suggests that HCV and HIV might cooperate to block astroglial glutamate clearance, potentially leading to excitotoxicity.
In conclusion, the present results demonstrate the presence of HCV antigens in the CNS of HIV/HCV-seropositive subjects, which supports the contention that HCV can promote neurological damage and is an important comorbidity factor in patients with AIDS.
SUBJECTS, MATERIALS, AND METHODS
Subjects and postmortem examination. Subjects had a complete neuromedical assessment and detailed laboratory antemortem data. Neurobehavioral assessment methods have been described in detail elsewhere [9, 26]. A total of 25 HIV-seropositive autopsy cases (12 HCV seropositive and 13 HCV seronegative) were selected for the present study on the basis of their enrollment in a longitudinal study at the University of California San Diego HIV Neurobehavioral Research Center, confirmed HIV and HCV serostatus, and availability of postmortem tissues (table 1). Exclusion criteria included a history of non-HIV-related neurological or medical disorders affecting the CNS (e.g., head trauma, neurosyphilis, or schizophrenia). In all cases, the complete postmortem examination included a detailed histopathological analysis of the brain and liver and of other tissues [9, 27]. A database including the patient demographic information, final anatomic diagnosis, and pathological cause of death was constructed (table 1).
RNA preparation, nested polymerase chain reaction (PCR), and quantitative reverse-transcription (qRT)-PCR for HCV detection. As described elsewhere , total RNA was extracted from frozen brain and liver samples from the 25 subjects by use of TriReagent (Molecular Research Center). RNA from blood and liver of an HIV-seronegative subject with confirmed HCV infection and from an uninfected subject was used as positive and negative controls. RNA quality was verified by inspection of the 18S and 28S ribosomal RNA bands. Nested restriction site-specific PCR was performed as described by Krekulova et al.  with some modifications. The first step amplified a 661-bp segment spanning nucleotide positions -285 in the 5 region to +376 in the NS5A region of HCV. The second round amplified a 451-bp segment spanning nucleotide positions -121 to +330 inside the first-round PCR product. To estimate the levels of HCV RNA in the brain samples, a modified version of the qRT-PCR method of Takeuchi et al.  was used.
Serial dilutions of glyceraldehyde 3-phosphate dehydrogenase (Invitrogen) and JFH-1 (HCV genotype 2) that contained 102-107 copies were used to generate a standard curve. PCR was performed on the iQ5 iCycler PCR Detection System (Bio-Rad). The number of HCV copies was determined using the standard 10-fold dilutions of HCV JFH-1.
Tissue preparation and HCV antigen purification with heparin columns. Because viral particles are usually complexed with immunoglobulins and HCV antigens are difficult to detect and might be present in small amounts in the CNS, an alternative strategy developed by Zahn and Allain  was used that uses a heparin column to partially purify HCV.
Protein (0.5 mg) was diluted in 1.0 mL of wash buffer (0.02 mol/L Tris/HCl and 0.15 mol/L NaCl [pH 7.4]) and loaded onto a 1.0-mL HiTrap Heparin HP column (GE Healthcare). A 0.3 mol/L NaCl elution buffer (0.3 mol/L NaCl and 0.02 mol/L Tris/HCl [pH 7.4]) was used to wash the column and to remove any proteins that did not bind to the column. This was followed with a 0.4 mol/L elution buffer (0.4 mol/L NaCl and 0.02 mol/L Tris/HCl [pH 7.4]). It has been previously determined that HCV is in this elution fraction . Blood, liver, and brain control samples from verified HCV-seronegative and -seropositive subjects were used to validate the assay. However, because of the different content of heparin-binding proteins in different organs, comparisons of immunoreactivity levels between purified brain and liver samples is difficult.
Antibodies. The following antibodies were used for immunoblot and/or immunocytochemical analysis: mouse monoclonal antibody (MAb) against HCV nonstructural protein (NS) 5A (Abcam), rabbit polyclonal antibody against HCV NS5A (GeneTex), mouse MAb against HCV NS3 (aa 1378-1458; GeneTex), mouse MAb against HCV core (aa 21-40; GeneTex), mouse MAb against HCV core antigen (aa 1-80; Chemicon), rabbit polyclonal antibody against HCV core antigen (Virogen), mouse MAb against HIV-1 gp41 (Virogen), and mouse MAb against HIV p24 (DakoCytomation).
Western-blot analysis for the detection of HCV antigens in the CNS. Heparin-purified brain fractions from each subject and control samples were analyzed by immunoblot assay as described elsewhere . Blots were probed with antibodies against HCV NS5A, NS3, and core antigen and with an antibody against HIV gp41. All antibodies were used at a dilution of 1 : 1000. After the addition of primary antibody, blots were then incubated with a horseradish peroxidase (HRP)-conjugated anti-rabbit or anti-mouse IgG (1 : 5000; American Qualex) and a Western Lighting Chemiluminescence Reagent (PerkinElmer) and were analyzed using the Versadoc gel imaging system (BioRad).
Neuropathological analysis and immunocytochemical detection of HCV. Paraffin sections of frontal neocortex, basal ganglia gray matter, and subcortical white matter (centrum semiovale) were used for routine neuropathological and immunocytochemical analyses with a mouse MAb against HIV p24 (DakoCytomation), as described elsewhere [27, 33]. The diagnosis of HIV encephalitis was based on the presence of HIV in the brain by HIV p24 and/or PCR , activated microglia, microglial nodules, multinucleated giant cells, astrogliosis, and myelin pallor.
For immunocytochemical analysis, paraformaldehyde-fixed tissue blocks from midfrontal cortex and basal ganglia were serially sectioned at 40 umol/L with the Vibratome 2000 (Leica) and incubated with antibodies against HCV (NS5A, NS3, and core antigen). After overnight incubation with the primary antibodies, sections were incubated with biotinylated anti-rabbit or anti-mouse IgG (1 : 100; Vector Laboratories) and Avidin D-HRP (1 : 200; ABC Elite) and reacted with diaminobenzidine tetrahydrochloride that contained 0.001% H2O2.
Double-immunocytochemical analysis was performed, as described elsewhere , to determine the cellular localization of HCV antigens in the CNS. For this purpose, vibratome sections were double-immunolabeled with a polyclonal antibody against NS5A protein (1 : 10,000) detected with the Tyramide Signal Amplification-Direct (Red) system (1 : 100; NEN Life Sciences) and mouse MAbs against either glial fibrillary acidic protein (1 : 500; Chemicon), CD68 (1 : 100; Wako), p24 (DakoCytomation), CD45 (1 : 40; Chemicon), or neurofilament (1 : 1000; Sternberger Monoclonals) detected with fluorescein isothiocyanate-conjugated secondary antibodies (1 : 75; Vector). All sections were processed simultaneously under the same conditions, and experiments were performed twice to assess reproducibility. Sections were imaged with a Zeiss 63Ã- (numerical aperture, 1.4) objective on an Axiovert 35 microscope (Zeiss) with an attached MRC1024 laser scanning confocal microscope system (BioRad). To confirm the specificity of primary antibodies, control experiments were performed with sections incubated overnight in the absence of primary antibody, preimmune serum, and primary antibody only.
Statistical analysis. All the analyses were conducted in triplicate on blind-coded samples. After results were obtained, the code was broken and data were analyzed with the StatView program (version 3; SAS Institute).
Clinicopathological characteristics of the subjects with HIV and HCV. A total of 25 HIV-positive subjects from a random sample were included (13 HCV negative and 12 HCV positive) by serological criteria. Both groups were matched for age and had a similar ethnic distribution. Although 75% of the subjects with HCV infection had history of illicit drug use that, in most cases, included methamphetamine and injection drugs, only 30% of the HCV-negative subjects had a history of illicit drug use. In both groups, the most common causes of death were sepsis and bronchopneumonia (table 1).
As summarized in table 2, most subjects had abnormal transaminase levels, with platelets counts >100,000 cells/mm3 and serum albumin levels >2.0 mg/dL. Sixty percent of the HCV-positive subjects showed evidence of portal fibrosis or liver cirrhosis, and some presented with multiple alterations, whereas, in the HCV-negative group, only 1 subject had evidence of fibrosis (table 1). This subject had serological evidence of hepatitis B virus (HBV) infection but was HCV negative. Additional liver findings in the HCV-positive subjects included necrosis, steatosis, and abundant mononuclear inflammation in the portal spaces (table 1). None of the subjects had evidence of transformation to hepatocellular carcinoma.
Although 33% of the HCV-positive subjects developed HIV-associated dementia and 67% developed either minor cognitive-motor disorder or some form of neuropsychological impairment, 23% of HCV-negative subjects presented with HIV-associated dementia, and 62% presented with either minor cognitive-motor disorder or a type of neuropsychological impairment. Of the 13 HCV-negative subjects, 11 had evidence of HIV encephalitis, whereas, of the HCV-positive subjects, 7 had HIV encephalitis (table 3). Overall, the frequency of opportunistic infections and neoplasms in these subjects was low, and type II Alzheimer disease was found in 1 HCV-negative subject and 4 HCV-positive subjects (table 3).
Frequent presence of HCV RNA and antigens in the CNS of HIV-positive subjects. To investigate the frequency of HCV RNA in the brains of HIV-positive subjects with serological evidence of HCV infection, nested PCR analysis was performed essentially as described by Krekulova et al. . RNA quality was verified in all cases by analysis of ribosomal RNA. Control experiments with a known HCV-positive, HIV-negative subject confirmed that only in the second round of PCR were HCV RNA sequences amplified from blood, liver, and brain (figure 1A). By contrast, tissue samples from an HCV-negative control subject (figure 1A) as well as from a subject with HBV (but no HCV) infection were negative by nested PCR for HCV RNA. Next, we analyzed the samples from the HIV-positive subjects; remarkably, all subjects who were HCV seropositive had HCV RNA sequences in the brain (figure 1B). In contrast, none of the 13 subjects who were HCV seronegative had evidence of HCV RNA in the brain (figure 1B). HCV RNA was present in the frontal cortex, basal ganglia, and white matter, but no HCV RNA was amplified from the cerebellum, brainstem, occipital cortex, and thalamus; in 60% of these subjects, HCV RNA was identified in the temporal cortex and hippocampus (data not shown). To verify the specificity of the band detected by nested PCR, gel extraction, purification, and sequencing were performed. This analysis confirmed that the segment being amplified corresponded to the NS5A region of HCV. Analysis of the HCV RNA in the CNS was performed by qRT-PCR. This study showed that 8 of the 12 subjects who were HCV positive by nested PCR had significant amounts of HCV RNA. qRT-PCR experiments in the frontal cortex and plasma showed substantial amounts of RNA only in the HCV-positive subjects and not in the HCV-negative group (figure 1C and 1D).
To identify HCV antigens in the brains of HIV-positive subjects, Western-blot analysis was performed with samples purified using heparin columns. An HCV immunoreactive band at about 50 kDa was detected with the rabbit polyclonal antibody against NS5A antigen in blood, liver, and brain samples from HCV-positive subjects, compared with those from HCV-negative control subjects (figure 2A). These results were confirmed with a mouse MAb against NS5A from an independent source (data not shown).
Further analysis was performed in the brain samples from the 25 HIV-positive subjects by use of frontal cortex samples purified with heparin columns. HCV immunoreactivity was detected by immunoblot analysis in 8 of 12 HCV-positive subjects, whereas all 13 HCV-negative subjects were negative with the polyclonal antibody against NS5A (figure 2B). In fact, the same 8 subjects who were HCV positive by Western-blot analysis were positive by qRT-PCR (figure 1C). These results were confirmed with mouse MAbs against NS5A and HCV core (figure 2B). Similarly to the PCR analysis, HCV immunoreactivity was detected by immunoblot in the frontal cortex, basal ganglia, and white matter but not in the cerebellum of HCV/HIV-positive subjects (data not shown). An MAb against NS3 also recognized HCV antigens in the heparin columns (data not shown). By contrast, the polyclonal antibodies against HCV core that were tested displayed a nonspecific labeling pattern (data not shown). The specificity of the antibodies was further tested in liver sections from HCV-positive/HCV-negative control subjects. Results similar to those in brain samples were obtained with liver samples (figure 2C-2F). Consistent with the neuropathological analysis, subjects with HIV encephalitis displayed HIV gp41 immunoreactivity in the heparin-purified preparation (figure 2B).
HCV immunoreactivity in astroglial and macrophage/microglial cells in HIV-positive subjects. To investigate the cellular localization of HCV immunoreactivity in the nervous system of HIV-positive subjects, immunocytochemical analysis was performed with antibodies against HCV antigens. In contrast to results in HIV-positive/HCV-negative subjects (figure 3A), in 8 HCV-positive subjects, the polyclonal antibody against NS5A immunostained glial cells (figure 3B and 3C). The HCV immunoreactive glial cells were more abundant in white-matter sections of the frontal cortex (figure 3B and 3C) and basal ganglia and were not detected in the cerebellum (data not shown). In addition, the polyclonal antibody against NS5A yielded immunoreactivity in cells with a macrophage-like appearance (figure 3D). Similar results were obtained when the mouse MAb against NS5A was used (figure 3E-3H). The MAb against HCV core also labeled glial cells (figure 3I-3L); however, background levels were higher than those for NS5A antibodies. All 13 HCV-seronegative subjects were negative by immunocytochemical analysis when antibodies against HCV were used (figure 3A, 3E, and 3I). Double-immunolabeling studies showed that most HCV-immunoreactive cells were astroglial because they colocalized with glial fibrillary acidic protein (figure 4A-4I). HCV NS5A immunoreactivity was also identified in microglial/macrophage cells displaying CD68 (figure 4J-4L) and CD45 (figure 4M-4O) immunostaining. HCV-immunoreactive macrophages/microglia were usually distributed around blood vessels (figure 4J-4L). No HCV immunoreactivity was detected in neurons or oligodendroglial cells. Approximately 80% of the HCV-immunoreactive cells colocalized with the astroglial marker glial fibrillary acidic protein, whereas 15% colocalized with the macrophage marker CD68, and in 5% none of these glial markers was present. Approximately 10% of the astrocytes and 10% of the macrophages were immunoreactive with HCV. Double-immunocytochemical analysis with antibodies against HCV NS5A and HIV p24 showed that, in about 18% of the perivascular macrophages and microglia, HIV and HCV colocalized (figure 4P-4R).
[Non-text portions of this message have been removed]