Downregulation of CCR5 on brain perivascular macrophages in simian immunodeficiency virus‐infected rhesus macaques

Abstract Background C‐C chemokine receptor 5 (CCR5) is a major coreceptor for Human immunodeficiency virus (HIV) and simian immunodeficiency virus (SIV) cell entry; however, its role in brain pathogenesis is largely understudied. Thus, we sought to examine cell type‐specific protein expression of CCR5 during SIV infection of the brain. Methods We examined occipital cortical tissue from uninfected rhesus macaques and SIV‐infected animals with or without encephalitis using immunohistochemistry and immunofluorescence microscopy to determine the number and distribution of CCR5‐positive cells. Results An increase in the number of CCR5+ cells in the brain of SIV‐infected animals with encephalitis was accounted for by increased CD3+CD8+ cells expressing CCR5, but not by increased CCR5+ microglia or perivascular macrophages (PVMs), and a concurrent decrease in the percentage of CCR5+ PVMs was observed. Levels of CCR5 and SIV Gag p28 protein expression were examined on a per‐cell basis, and a significant, negative relationship was established indicating decreased CCR5 expression in productively infected cells. While investigating the endocytosis‐mediated CCR5 internalization as a mechanism for CCR5 downregulation, we found that phospho‐ERK1/2, an indicator of clathrin‐mediated endocytosis, was colocalized with infected PVMs and that macrophages from infected animals showed significantly increased expression of clathrin heavy chain 1. Conclusions These findings show a shift in CCR5‐positive cell types in the brain during SIV pathogenesis with an increase in the number of CCR5+ CD8 T cells, and downregulated CCR5 expression on infected PVMs, likely through ERK1/2‐driven, clathrin‐mediated endocytosis.


INTRODUCTION
Human immunodeficiency virus (HIV) infection spreads through the body at a very high rate, with diverse sites of infection established in the first 3 days, including immune privileged sites such as the brain.
Studies examining early antiretroviral therapy (ART) starting as early as 3 days post infection have shown that this rapid dissemination results in seeding of viral reservoirs that persist through long-term ART treatment (Whitney et al., 2014). With the inability to eliminate these reservoirs through current ART regimens, it has become vitally important to understand the cells that make up these reservoirs and their role in HIV pathogenesis (Hsu et al., 2018;Ko et al., 2019).
Previous studies by our group have shown brain perivascular macrophages (PVMs) and microglia to be the major viral reservoir in the brain (Ko et al., 2019). These cells are infected with macrophagetropic HIV/simian immunodeficiency virus (SIV) through CD4 receptor and a coreceptor C-C chemokine receptor 5 (CCR5) (Edinger et al., 1997). Viral envelope protein gp120 binds strongly with CD4 causing a conformational change in the gp120 protein which then allows for CCR5 interaction, leading to membrane fusion and viral entry into the cell (Cormier et al., 2000;Tamamis & Floudas, 2014). With low expression of CD4 on macrophages, including PVMs, levels of CCR5 surface expression likely contribute to the susceptibility of these cells to HIV infection. Interestingly, a mutation in CCR5, CCR5∆32, truncates the receptor leading to lower membrane expression and degrees of resistance to infection both in vitro and clinically in two successful ∆32/∆32 bone marrow transplantations in which the patients showed no viral rebound despite cessation of ART (Hütter et al., 2009). CCR5 is a seven-transmembrane G protein-coupled receptor expressed on immune cells, which facilitates the trafficking of cells during the inflammatory response. In addition, CCR5 has been shown to play a role in activation of CD8 T cells, with its expression highly correlated with activation status (Wang et al., 2019). Ligands for this receptor include CCL3, CCL4, CCL3L1, and CCL5 (RANTES), all of which, upon binding, cause rapid internalization of ligand and receptor (Venuti et al., 2017(Venuti et al., , 2015. In vitro work examining this internalization has shown the presence of CCR5 in invaginated clathrin pits following CCL5 treatment, as well as CCR5 internalization inhibited by depletion of clathrin (Mueller & Strange, 2004;Signoret et al., 2005). However, reports differ on the fate of CCR5 following internalization, some showing quick recycling to the cell surface, while others demonstrate sustained downregulation (Mueller et al., 2002;Signoret et al., 2000;Venuti et al., 2015).
This may suggest multiple, cell-fate specific, signaling pathways regulating CCR5 surface expression. ERK1/2 signaling has been shown to both colocalize with internalized CCR5 and prevent internalization of CCR5, when inhibited (Venuti et al., 2015). Further, the results of Venuti et al. (2015) indicate receptor degradation, showing cell surface re-expression dependent upon de novo synthesis.
The importance of CCR5 in the infection of myeloid cells is well established; however, its presence and role in HIV brain pathogenesis is still unknown, in part due to the paucity of specific anti-human CCR5 antibodies that are effective in routinely processed-that is, formalinfixed, paraffin-embedded-tissue. For this study, we tested multiple commercially available CCR5 antibodies, validating a single antibody, as well as a monoclonal human CCR5 antibody kindly gifted by Dr.
Mathias Mack at Regensburg University (Regensburg, Germany) for immunohistochemical reactivity to human and monkey formalin-fixed tissue. We showed an increase in CCR5-positive, activated CD8 infiltrates driving an increase in total CCR5-positive cells in the brains of encephalitic animals. Despite this progression to severe infection, there was a decline in percentage of CCR5-positive PVMs, which was surprising considering myeloid cell-driven neuropathogenesis in HIV/SIV infection. We showed strong co-localization of clathrin heavy chain 1 with PVMs, a significant inverse association between SIV Gag p28 protein and CCR5 levels, and increased levels of phospho-ERK1/2, indicating clathrin-mediated endocytosis, degradation, and sustained downregulation of CCR5 in SIV-infected PVMs.

Animals
Archival sections from a total of 23 rhesus macaque (Macaca mulatta), 14 adult and nine neonates listed in Table 1, were used in this study.
All animals except those marked with an asterisk (*) were housed at the Tulane National Primate Research Center (TNPRC), and all procedures were approved by the Tulane University Institutional Animal Care and Use Committee. Animals marked with an asterisk in Table 1 were housed at the New

Human samples
The Manhattan HIV Brain Bank (MHBB) graciously provided FFPE sections of frontal white matter for 3 HIVE cases and 3 seronegative controls (Table 2). Controls had minimal non-diagnostic abnormalities when autopsied. Postmortem intervals were all less than 48 h.

Immunohistochemistry
Immunohistochemistry was performed as previously described (Filipowicz et al., 2016). In brief, deparaffinization and rehydration TA B L E 1 Animals used in this study.

Immunofluorescence
Sections were deparaffinized, rehydrated, and treated with antigen retrieval as stated above. The sections were washed in phosphatebuffered saline with 0.2% fish skin gelatin (Sigma-Aldrich, St. Louis, MO, USA) permeabilized in wash buffer with 0.1% Triton X-100, washed, and then blocked with 5% normal horse or goat serum.
Without a wash, the sections were incubated with primary antibodies (Table 3)

Immunohistochemistry image analysis
Semi-quantitative analysis was performed; 15 (CD3) or 21 (CCR5) images were acquired per section at a magnification of 20× on a Nikon Coolscope digital microscope and manually counted. Counts were recorded as total positive cells per 15 or 20 images for each animal or human subject. Statistical analysis between groups was performed in GraphPad Prism 9.0.

Immunofluorescence image analysis
Images were acquired on a ZEISS Axio Observer.Z1 fluorescence microscope at a magnification of 20× or 40× using ZEISS AxioVision Statistical analysis between groups was performed in GraphPad Prism 9.0.

NanoString
Total RNA was extracted from FFPE tissue samples from uninfected

Statistical analysis
Statistical analysis was performed using GraphPad

Cells expressing HIV coreceptor CCR5 increase with progression to AIDS and encephalitis
Due to the critical role played by CCR5 in HIV infection of macrophages, we hypothesized that the presence of CCR5 on brain PVMs, the major cellular target of HIV/SIV in the brain, would increase throughout disease progression. To begin investigating CCR5 in the brain, we examined occipital cortical tissue from adult rhesus  (Bohannon et al., 2020;Delery et al., 2019;Dickson et al., 1993;Kure et al., 1991). As such we compared UI and SIV-infected infant rhesus macaques for CCR5 expression to elucidate whether differential expression of the HIV/SIV coreceptor may play a role in lesion formation. No change was found between UI and infected neonates in the number of CCR5-positive cells ( Figure 1f); however, numbers were significantly lower in infants than adults as reported by Delery et al. (2019). Due to extremely low CCR5 expression in uninfected and infected neonates/infants and the complexity of macrophage subsets in neonatal SIV infection, neonates were not investigated further in this study (Bohannon et al., 2020).

Microglia, T cells, and PVMs express CCR5
CCR5 expression on various cell types throughout the body has been reported, and as such we set out to examine the phenotype of cells

Activated, brain-infiltrating, CD8 T cells, not PVMs or microglia, account for the observed increase in CCR5+ cell count
Because PVMs are the major cell type infected with HIV and SIV in the brain, we assumed that the increase in CCR5-positive cell counts would closely correlate with an increase in CCR5+ PVMs. We there- and SIVnoE animals were found both in association with vessels and scattered in the brain parenchyma. To confirm these cells were positive for CCR5, contributing to the increase in total CCR5-positive cells, triple-label immunofluorescence was performed, and the percentage of CCR5+ CD3 cells was established with approximately 80% of CD3 T cells co-expressing CCR5 in animals with encephalitis. Of note, both SIVnoE and SIVE animals showed significantly higher percentage of CCR5+ CD3 T cells than UI, with no significant variation in percentage between the two groups ( Figure S2). Taken together, these results show that infiltrating, activated CCR5+ CD8 T cells and not PVMs account for the increase in total number of CCR5-positive cells in the brains of encephalitic animals. CCR5+ microglia were rare and did not contribute to changes in subpopulation distribution of total CCR5-positive cells.

With increased severity of infection and increased numbers of infected
PVMs in SIVE animals, (Filipowicz et al., 2016) we expected to see an increase in the percentage of CCR5+ PVMs but found the opposite.
Given these observations, we updated our initial hypothesis and sought to investigate whether spreading infection may cause an internalization and downregulation, or complete degradation, of CCR5 following PVM infection, leading to a decreased percentage. To test this, we

Activated ERK1/2 MAP kinases in association with downregulated CCR5 in SIV-p28+ macrophages
Multiple studies have examined the endocytosis of CCR5 in vitro, presenting differing mechanisms that lead to either recycling of the receptor to the cell surface or sustained downregulation. Activated ERK1/2 signaling leading to clathrin-mediated endocytosis has been shown to lead to sustained downregulation, with membrane expression following internalization dependent at least in part on de novo CCR5 synthesis (Venuti et al., 2015). To see if this mechanism persists in vivo, we first checked for the colocalization of clathrin heavy chain 1 with PVMs in SIVE tissue and observed strong colocalization (Figure 5a).
The percentage of clathrin-positive macrophages and MPI of F I G U R E 3 T cells and not perivascular macrophages (PVMs) or microglia account for the increase in the number of C-C chemokine receptor 5 (CCR5)-positive cells in the brain. Triple-label immunofluorescence staining was performed for CCR5, CD163, and DAPI showing a significant decrease in the percentage of CD163+ PVMs that are positive for CCR5 in animals with encephalitis compared with both uninfected (UI) and SIVnoE, one-way analysis of variance (ANOVA) with Tukey's multiple comparison test, p-values = .0113 and .0297 (a). Immunohistochemistry (IHC) staining for CD3 was performed on UI, SIVnoE, and SIVE animals showing a significant increase in infiltrating CD3 T cells in animals with encephalitis, one-way ANOVA with Tukey's multiple comparison test, p-values = .0004 and .0026 (b and c).

F I G U R E 4
Simian immunodeficiency virus (SIV) p28 protein is associated with decreased CCR5 in SIV p28-positive cells. Triple-label immunofluorescence microscopy was performed for SIV p28 (red), CCR5 (green), and DAPI (blue) in the SIVE rhesus macaques, and images were acquired on a ZEISS AxioVision Release 4.8.2 (a). Background subtraction using the AxioVision and Adobe Photoshop software was applied. Normalized mean pixel intensity (MPI) data were sent to a statistician (Jiangtao Luo) who modeled the relationship between SIV p28 and CCR5 using a general additive model with a linear function followed by a smooth function estimated using a scatterplot smoother with both parts reaching statistical significance with p-value < .0001 (b). clathrin-positive macrophages was determined for UI, SIVnoE, and SIVE groups, with a significant difference in proportion between groups and significantly higher normalized MPI in SIVE animals (Figure 5b,c). We then examined phospho-ERK1/2 as a measure of activated ERK1/2 signaling, and phospho-ERK1/2 staining was present in cells with cytoplasmic CCR5. To further examine if phospho-ERK1/2 contributed to the SIV-associated CCR5 downregulation, MPI of phospho-ERK1/2 was examined on SIV-p28-positive and SIV-p28-negative cells, and it was found that SIV-p28-positive cells express significantly more phospho-ERK1/2 than uninfected cells (Figure 5d,e).
Indeed, in SIVE animals, there was a significant increase in the ratio of phospho-ERK1/2 to CCR5 in SIV p28-positive cells (Figure 5g), indicating a conserved ERK1/2-dependent mechanism of internalization that is accelerated in SIV-infected cells. A study conducted by Venuti et al. (2015) found that specific ERK1 inhibition led to significantly less cytoplasmic CCR5 and β-arrestin 1/2 accumulation, indicating F I G U R E 5 Phospho-ERK1/2 is upregulated in infected cells. Triple-label immunofluorescence staining for clathrin heavy chain 1 (green), CD163 (red), and DAPI (blue) was performed on SIVE tissue (a) showing colocalization of clathrin heavy chain 1 with perivascular macrophages (PVMs). Chi Squared test was performed comparing proportion of clathrin-positive CD163+ macrophages between groups, with 50 macrophage cells analyzed per group (b). Mean pixel intensity (MPI) was measured and normalized for each clathrin-positive macrophage, and a one-way analysis of variance (ANOVA) with Tukey's post hoc test was performed to compare between groups showing significantly higher clathrin MPI in SIVE animals, p-value < .0001 and < .0001 (c). Triple-label immunofluorescence staining for SIV p28 (green), phospho-ERK1/2 (yellow), and CCR5 (red) was performed on SIVE tissue (d). Unpaired t tests were performed for all statistical analysis. MPI analysis on a per-cell basis was performed on 61 SIV p28− cells and 91 SIV p28+ cells showing a significant increase in phospho-ERK1/2 MPI, p-value < .0001 (e) and a significant decrease in CCR5 MPI, p-value = .0013 (f). Only cells with above background CCR5 levels, > five adjusted MPI, were counted, while cells with a T-cell morphology, small and very round, were excluded from this analysis. The ratio of phospho-ERK1/2 to CCR5 was calculated for each cell showing a significant increase in SIV p28+ cells, p-value < .0001 (g). Analysis of NanoString nCounter data for ERK1 (h) and ERK2 (i) showed a significant increase in ERK1 mRNA count, p-value = .0256, but no change in ERK2 in animals with encephalitis.
inhibited internalization. To investigate potential differences in ERK1 versus ERK2, NanoString data of four UI and four SIVE frontal cortices were probed with the NHP Immunology V2 code set and examined for ERK1 and ERK2 mRNA expression levels in the brain. No difference was seen in ERK2 between uninfected and infected animals; however, there was a significant increase in ERK1 mRNA in SIVE-infected animals supporting Venuti et al.'s findings (Figure 5h,i). This suggests that the increased internalization of CCR5 may be driven by increased ERK1 signaling.

DISCUSSION
The role of CCR5 in HIV pathogenesis as the major coreceptor of viral entry has been known in the field for over 25 years; however, its specific expression and importance in brain pathogenesis are as of yet unclear.
In this study, we show an increase in total CCR5-positive cells in the brain with progression to AIDS and encephalitis with infiltrating CD8 T cells representing the majority of these CCR5-positive cells. Previous studies by our group have shown PVMs to represent the major cellular target of HIV and SIV and proliferation of PVMs during the development of SIVE (Filipowicz et al., 2016;Ko et al., 2019). In this study, we demonstrate for the first time, decreased expression of CCR5 in virally infected macrophages, which strongly suggests a virally driven CCR5 downregulation and a decrease in CCR5+ macrophages. Overall, we argue that increased activated ERK1/2 MAP kinase signaling seen here leads to an increase in clathrin-mediated endocytosis of CCR5 and subsequent proteasomal degradation.
The significantly higher CD3 cell count in the brain of SIVE animals is in agreement with current literature on CD8 infiltrate in HIV/SIV infection, and the novel finding of high CCR5 expression on this infiltrate shown here indicates a substantial activated population (Subra & Trautmann, 2019). Studies by Fukada et al. (2002) and Wang et al. (2019) have shown all SIV/HIV specific tetramer-positive CD8 T cells to be positive for CCR5 in PBMC's, lymph node, and jejunum in infected macaques and humans. These findings are in concurrence with previous studies of T cell populations in the brain of SIV-infected macaques, which show an accumulation of antigen (virus)-specific, activated CD8+ T cells in the brains of these animals (von Herrath et al., 1995;Marcondes et al., 2003). CCR5 expression on CD8 T cells is seen mainly on effector memory cells defined as CD45RA − CD28 + CCR7 − , with different groups reporting either negative (Wang et al., 2019) or lower expression of CCR5 on other subsets (Fukada et al., 2002), with no expression on naive cells. Despite this heightened specific CD8 response, the body is unable to clear infection or halt pathogenesis.
T-cell exhaustion in chronic infections due to over exposure to antigen has been shown in patients with HIV and may play a role in this inability to combat the infection (Li et al., 2021;Trautmann et al., 2006;Zhang et al., 2007). Further studies confirming phenotype of CCR5 + T cell subsets and examining exhaustion status on these cells are warranted.
In vitro CCR5 downregulation in response to receptor binding has been shown to be transient, with eventual re-expression at the cell surface unless de novo synthesis of CCR5 is also inhibited (Mueller & Strange, 2004;Venuti et al., 2017Venuti et al., , 2015. We demonstrate increased activated MAP kinase signaling in vivo, which leads to CCR5 internalization and degradation, in cells with high SIV p28 viral protein levels. HIV/SIV viral accessory protein nef is produced following the infection, and the expression of nef has been shown to downregulate multiple cellular proteins in a transcriptionally independent manner, including CCR5 (Dubey et al., 2008;Haller et al., 2014;Michel et al., 2005;Pawlak et al., 2018 Gupta et al., 2019). While these results are encouraging, the treatment and procedure is both dangerous and damaging to the recipients. Attempts to pharmacologically block CCR5 through antagonist treatment using maraviroc have shown reduction in viral replication and myeloid activation when applied soon after infection (Kelly et al., 2013). Maraviroc showed some success clinically, meeting non-inferiority conditions in phase III clinical trials compared to thencurrent ART and gaining approval by the FDA (Woollard & Kanmogne, 2015); however, clinical results have not fully lived up to promising preclinical investigations. Maraviroc has shown success halting the spread of infection; however, it likely has little to no effect on already infected cells with low to no levels of CCR5 expression (Tilton et al., 2010;Tiraboschi et al., 2017). This implies that maraviroc application will be most successful soon after infection, with reduced effectiveness in long standing infections. Combination therapy targeting CCR5-low cells while simultaneously blocking spread to CCR5-positive cells should be investigated.

CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.