CCL2 is required for initiation but not persistence of HIV infection mediated neurocognitive disease in mice

HIV enters the brain within days of infection causing neurocognitive impairment (NCI) in up to half of infected people despite suppressive antiretroviral therapy. The virus is believed to enter the brain in infected monocytes through chemotaxis to the major monocyte chemokine, CCL2, but the roles of CCL2 in established NCI are not fully defined. We addressed this question during infection of conventional and CCL2 knockout mice with EcoHIV in which NCI can be verified in behavioral tests. EcoHIV enters mouse brain within 5 days of infection, but NCI develops gradually with established cognitive disease starting 25 days after infection. CCL2 knockout mice infected by intraperitoneal injection of virus failed to develop brain infection and NCI. However, when EcoHIV was directly injected into the brain, CCL2 knockout mice developed NCI. Knockout of CCL2 or its principal receptor, CCR2, slightly reduced macrophage infection in culture. Treatment of mice prior to and during EcoHIV infection with the CCL2 transcriptional inhibitor, bindarit, prevented brain infection and NCI and reduced macrophage infection. In contrast, bindarit treatment of mice 4 weeks after infection affected neither brain virus burden nor NCI. Based on these findings we propose that HIV enters the brain mainly through infected monocytes but that resident brain cells are sufficient to maintain NCI. These findings suggest that NCI therapy must act within the brain.


Introduction
Combination antiretroviral therapy (CART) in people living with HIV (PLWH) suppresses HIV replication, prevents immunode ciency, and extends lifetimes of people on CART.Nevertheless, other chronic ailments caused by HIV continue 1,2 .Among these, HIV-associated neurocognitive disorder (HAND) occurs in roughly half of CART compliant patients impairing day to day activities and reducing the quality of life 3 .With aging, PLWH are facing the fact that HAND tends to worsen with age 4 .To control or eliminate HAND, we must understand its viral and cellular bases.
In this effort, we have employed infection of conventional mice by EcoHIV, a chimeric HIV which encodes ecotropic murine leukemia virus gp80 in place of gp120 thereby switching viral tropism from human to rodent.We and others have shown that EcoHIV replicates in lymphocytes, macrophages, and the brain; establishes a lifelong chronic infection; and induces neuronal dysfunction and cognitive and depressionlike defects in all infected animals [5][6][7][8][9][10][11][12][13] .Using intracranial injection of EcoHIV we identi ed macrophage/microglial cells as its major target in the brain 6,8 , a result reproduced using RNAscope in infected rats 10 .In EcoHIV-infected mice, as in other animal models and PLWH themselves, macrophages and brain tissues overexpress in ammatory cytokines, a spectrum of host antiviral factors, and chemokines that induce cellular migration 6,8, 14−16 .Prominent in the latter category is CCL2, the primary chemokine for monocytes.CCL2 serves as an autocrine factor in HIV brain infection.It is elevated in cerebrospinal uid (CSF) of some individuals within weeks of HIV infection 15 .In macaque models of SIV encephalitis, elevated CCL2 in CSF compared to plasma predicts the development of disease 17 .Its synthesis is induced by viral gp120, Tat, and Nef proteins and by interaction of astrocytes with HIVinfected macrophages [18][19][20][21] .Conversely CCL2 promotes HIV infection in T cells and macrophages 22,23 .
Moreover, in culture models, CCL2 has been shown to prompt transit of monocytes, particularly HIVinfected monocytes, across the blood brain barrier 24,25 .PLWH with a naturally occurring promoter variant in which CCL2 expression is increased show increased risks of HIV-associated dementia 26 and their levels of CCL2 in the CSF correlate with the extent of impairment in cognitive tests 27 .Here we investigate the partnership of CCL2 and macrophages in neurocognitive impairment (NCI) during experimental EcoHIV infection of mice.Our studies show that CCL2 is required for HIV entry into the brain and the development of NCI during systemic HIV infection but is dispensable once brain infection and disease have been established.

Materials And Methods
Mice, cells, and tissue culture.The protocols for all experiments involving animal use were executed with the approval of the Icahn School of Medicine at Mount Sinai Animal Care and Use Committee under the protocol IACUC-2014-0124 (DJV), renewed on August 11, 2022.Any and all experiments using mice were designed and conducted with the approval of Icahn School of Medicine at Mount Sinai Animal Care and Use Committee and according to all relevant guidelines and regulations including all relevant ARRIVE guidelines including reporting.All mice were purchased from Jackson Laboratory.C57BL/6 (stock# 000664), B6.129S4-CCL2 (CCL2 KO mice, strain stock# 004434), and B6.129S4-Ccr2 tm1Ifc /J (CCR2 KO) mice, strain stock# 004999).After virus infection, mice were euthanized under methods approved by the Icahn School of Medicine at Mount Sinai Animal Care and Use Committee and peritoneal microphages, spleens, and brain tissues were removed and prepared for measurement of HIV burden, cellular gene expression and microscopy as described 6,28,29 .For bone marrow macrophage (BMM) culture, 8-10 weeks old mice were sacri ced by carbon dioxide asphyxiation.Marrow was harvested from separate hind legs, erythrocytes were lysed using ACK lysing buffer (Lonza, Walkersville, MD) for 5 min at room temperature, and nucleated cells were cultured in RPMI1640 with 10% horse serum, 5% fetal bovine serum, 100 units/ml penicillin/streptomycin (ThermoFisher Scienti c, Waltham, MA), and 20% L929 conditioned medium 30 .For infection, cells were cultured in complete culture medium without antibiotics 24 h before infection.Cells were then infected with virus at 1 pg p24 per cell in culture medium without serum and antibiotics for 3 h then replaced the medium to complete culture medium and incubated in 5% CO 2 at 37°C for 7 days.
Virus construction and preparation.EcoHIV/NDK-V5 and EcoHIV/NDK-EGFP were constructed as described 6,9 respectively and virus stocks were prepared as described 14 .
Preparation of bindarit and treatment.Bindarit, 2-((1-benzyl-1H-indazol-3-yl)methoxy)-2-methylpropanoic acid (MW 324.37), was synthesized by Biorbyt (Biorbyt Limited, UK) and was prepared as a suspension in 0.5% methylcellulose (MTC) at a concentration of 20 mg/ml as previously described 31 .Animals were injected IP with bindarit, suspended in 0.5% MTC aqueous solution, at the dose of 100 mg/kg daily 32 either for prevention from 1 day before EcoHIV infection until euthanasia or for treatment starting four weeks after infection until euthanasia.
Virus or vehicle injection.IP injections were in a volume of 0.5 ml.Infection was conducted at a dose of 2.0x10 6 pg p24 unless otherwise stated.Intracranial infection was conducted as described 6,28,29 .
Behavioral tests.Learning and memory were evaluated in groups of 8 to 10 mice using the radial arm water maze (RAWM) and auditory-cued fear conditioning (FC) tests as the authors have described previously 9,28,29 .Brie y, RAWM testing consisted of four training trials (T), followed by a retention trial (RT) administered after a 30-min rest, repeated daily until control mice reached asymptotic performance of one error or fewer on trials T4 and RT.Errors for the last 3 days of testing were averaged for statistical analysis.The hidden platform tests were followed by measuring the latency for nding a visible platform as a control for possible effects of treatment on animal vision, motivation, and motor ability.FC testing was conducted using an SDI Freeze Monitor (San Diego Instruments, San Diego, CA).Conditioning sessions included three consecutive pairings of 10-Hz sound signals and 0.7-mA electric shock signals; cued associative fear memory was measured the following day in a novel context by presenting sound signal alone.Results are shown as the mean total percentage of time spent freezing pre-and post-cue on both the conditioning and the cued memory days.
Quantitative real-time RT-PCR (QPCR).For QPCR, RNA and DNA were isolated from cells and tissues as described 14 and ampli cation was conducted as described 28,34 .

Results
Kinetics and dose response of EcoHIV brain disease development.
Reasoning that the development of NCI requires the presence of EcoHIV in the brain, we infected mice by intraperitoneal (IP) injection of virus stock and then at graded times after infection performed RAWM assays of learning and euthanized cohorts and collected tissues for measurement of viral DNA and RNA (Fig. 1).Mice that began the RAWM 1 or 10 days after infection displayed learning behavior equivalent to that of uninfected mice (Fig. 1a-c).However, when infection was allowed to proceed 25 or 55 days, mice were signi cantly cognitively impaired.All mice tested were able to nd and swim to a visible platform demonstrating their motor and visual competence (Fig. 1c).Virus DNA and RNA burden in spleen reached an early peak and declined 28 (Fig. 1d-f).Viral DNA in brain was also detected within days of infection and was relatively stable over the rst three weeks of infection (Fig. 1g).PLWH and SIV-infected macaques show similar evidence of central nervous system infection within days of exposure 15,39 .With the persistence of viral DNA in the brain and behavioral defects, we investigated whether NCI development requires a minimum EcoHIV burden in the brain and performed virus dose response studies (Sup.Figure 1).The splenic virus burden tended to increase with dose.At EcoHIV infections from 4 µg p24 to 60 ng p24 each, mice were signi cantly impaired in RAWM (Sup. Figure 1).Together, these results outline a time-period when EcoHIV establishes neuropathogenic infection in the brain and virus dose thresholds for cognitive impairment.We next investigated the mechanism for this staged virus spread from the periphery to the brain.

Monocyte chemotaxis to CCL2 contributes to EcoHIV NCI.
Many previous studies implicate infected monocytes as vectors of HIV entry to the brain [39][40][41] .Indeed, we previously showed that leukocytes that in ltrate the brains of infected mice carry integrated EcoHIV DNA 28 .The likely signal for this migration is CCL2; it is the major monocyte chemokine 42 and is highly induced by EcoHIV infection in mouse macrophages 43 and in astrocytes and monocytes/microglia in the HIV or EcoHIV-infected brain 8, 44 .We exploited a mouse strain with knockout of CCL2 and compared the course of its EcoHIV infection to that of wildtype C57BL/6 mice (Fig. 2).
To con rm that CCL2 is essential for monocyte migration to the brain, we isolated mononuclear cells from brains of both mouse strains 25 days after EcoHIV infection and performed ow cytometry to detect in ammatory monocytes de ned as CD11b + Ly6G neg CD45 high Ly6C + as described 28,45 .This critical cell type was signi cantly reduced in brains of CCL2 KO mice indicating a requirement for CCL2 for entry into the brain (Fig. 2a).Testing fear associated learning and memory function in fear conditioning tests revealed that wildtype mice were highly impaired by IP EcoHIV infection, however, infected CCL2 KO mice were not different than uninfected mice in cognitive function (Fig. 2b-c).The same mice were tested in RAWM function with similar results, only infected C57BL/6 mice were impaired (Fig. 2d-f).While both mouse strains were susceptible to EcoHIV infection of spleen cells and peritoneal macrophages, viral DNA was not detected in the brain over 5 weeks of infection in CCL2 KO mice (Fig. 2g-i).Taken together these ndings indicate that the absence of cognitive disease and profound reduction of virus in the brain of CCL2 KO mice resulted from the loss of CCL2 driven EcoHIV-infected monocyte tra cking to the brain.To test whether CCL2 has additional roles in HIV NCI, we performed direct EcoHIV infection of the brain by intracranial (IC) injection and measured cognitive function in wildtype and CCL2 KO mice.IC infection signi cantly impaired performance in RAWM by wildtype and CCL2 KO mice but affected neither vision nor motor function (Fig. 2j-l).The ability of brain resident but not systemic EcoHIV to cause NCI in the absence of the chemokine CCL2 strongly suggests that during systemic HIV infection, initiation of cognitive disease is driven by CCL2 directed migration of monocytes from the periphery to the brain.These ndings also indicate that CCL2 has no other roles in HIV NCI.

CCL2 promotes EcoHIV infection of macrophages in culture.
Previous studies indicate that CCL2 promotes HIV infection in culture 22,23 .To assess the role of CCL2 and its major receptor, CCR2, on cultured macrophage susceptibility to EcoHIV, we employed one week infection in tissue culture of fully differentiated bone marrow macrophages (BMM) from mouse strains C57BL/6, CCL2 KO, and CCR2 KO, the latter lacking the principal CCL2 receptor (Fig. 3).We performed uorescence microscopy and image analysis to count cells expressing enhanced green uorescent protein (EGFP) from EcoHIV carrying the marker downstream of nef 6 and co-stained cells for macrophage marker Iba-1 (red) and cellular DNA (blue) (Fig. 3a).Knockout of CCL2 or CCR2 signi cantly impaired BMM infection measured by EGFP expression (Fig 3b).We also conducted QPCR to measure newly synthesized, circular 2-LTR DNA and found a signi cant reduction of infection in CCL2 KO but not CCR2 KO macrophage compared to wildtype cells (Fig. 3c).Finally, we repeated the experiment and using immunohistochemistry, we counted cells expressing HIV p24 and found reduced virus expression in both CCL2 and CCR2 KO cells (Fig. 3d).However, the reduction in virus expression is roughly 25-60% and is transient during infection of macrophages in mice (Fig. 2), suggesting that in vivo long-term susceptibility to or expression of EcoHIV is affected by interactions among diverse cell types and their production of chemokines and cytokines.

Block of CCL2 synthesis prevents development of HIV-NCI.
To test the effects of CCL2 upon HIV brain disease by an alternate approach, we employed bindarit that inhibits transcription of CCL2 through interference with NFkB activation at the CCL2 promoter 46 .In mice, bindarit blocks expression of CCL2 RNA and protein and prevents in ammatory monocyte entry into the brain 47 as well as preventing pathological microglia activation during impaired neonatal brain development 48 .C57BL/6 mice were pretreated with bindarit one day before EcoHIV infection by IP injection and then daily through RAWM testing at Day 25 until euthanasia (Fig. 4).As shown in Fig. 1-2, EcoHIV-infected mice were unable to learn in the RAWM but bindarit treated, infected mice learned as well as uninfected mice (Fig. 4a-b).Bindarit administration had no effect upon performance of uninfected mice and in no case was motor or visual capacity affected (Fig 4c).To evaluate the effects of bindarit upon monocyte tra cking we measured cells in the peritoneal cavity, staining for monocyte marker F4/80.Bindarit increased the number of F4/80 positive peritoneal cells in both infected and uninfected mice indicating some restriction in exit from this compartment (Fig. 4d).Measurement of virus burden revealed that after bindarit treatment, EcoHIV was undetectable in brain, signi cantly increased in spleen and signi cantly reduced in peritoneal macrophages (Fig 4e -g).Overall, these ndings are consistent with those from CCL2 KO mice and indicate that CCL2 is required for e cient monocyte migration and HIV entry into the brain, functions essential for the development of HIV NCI.

Block of CCL2 synthesis does not affect established EcoHIV brain infection or disease.
The ability to conditionally eliminate CCL2 with bindarit permits investigation of the role of CCL2 in NCI and the persistence of EcoHIV in the brain.Mice were infected by IP injection and allowed four weeks for development of NCI; then bindarit was administered daily for one week prior to and during the RAWM for a total of two weeks of drug treatment.Mice were then euthanized for tissue collection and measurement of virus burden (Fig. 5).EcoHIV-infected mice showed learning defects regardless of bindarit treatment, the drug itself did not affect performance of uninfected mice in the behavioral assay.Virus burden in the brain was not altered by bindarit, however virus burden in macrophages was signi cantly increased, likely due to inhibition of exit from the peritoneal cavity as shown in Fig. 4g.The maintenance of brain virus and disease during treatment with an effective preventative of HIV brain infection is consistent with the development of HAND in PLWH despite suppressive ART 49,50 and suggests that interventions to treat HAND must act in the brain.

Discussion
We report that through its promotion of infected monocyte migration, CCL2 is required for the development of HAND during systemic EcoHIV infection of mice, however once HAND is established CCL2 is dispensable.We mapped the time-course of EcoHIV brain infection and cognitive impairment to outline its brain entry, a possible CCL2 function.Kinetics of EcoHIV replication after IP inoculation indicate that viral DNA burdens in lymphocytes decline moderately but are stable in the brain.The timecourse of HAND development indicates a progressive disease: 10 days of infection are insu cient to impair learning but mice infected for 25 days or 55 are highly impaired, we previously reported that EcoHIV NCI persists for at least 12 weeks 28 .Virus dose response studies also show a threshold of productive HIV brain infection required for HAND development.Because CCL2 promotes monocyte migration in blood brain barrier models 24,25 , and its levels in the CNS correlate with cognitive dysfunction in HIV infection in PLWH 27 we investigated the roles of CCL2 in HAND in EcoHIV-infected mice.
Using existing mouse knockouts of CCL2, we found that they fail to develop HAND during systemic EcoHIV infection but succumb to cognitive disease when the virus is injected into the brain.These results show that EcoHIV must enter the brain to cause disease, either by direct injection or by CCL2 driven entry of infected monocytes from the periphery into the brain, the latter route previously indicated [39][40][41] .The absence of disease and brain EcoHIV in CCL2 KO mice after peripheral infection indicates that other chemotactic factors are insu cient to summon virally infected cells to the brain.Consistent with these ndings we reported previously that brain in ltrating leukocytes carry integrated HIV DNA 28 .Indeed here we show that during systemic infection of CCL2 KO mice in ammatory monocytes are reduced in the brain and EcoHIV DNA was undetectable there.The necessity of monocyte migration to the brain during EcoHIV infection and NCI was recently reported through treatment of mice prior to infection with buprenorphine, an opiate agonist 45 .Brain in ammation and induction of CCL2 are prominent after IC EcoHIV infection of wildtype mice 6,8 .However, cognitive impairment observed here after IC infection of CCL2 KO mice demonstrates that responses to CCL2 in the brain 44,51,52 are dispensable for HAND development.
During HIV/EcoHIV infection of macrophages in culture CCL2 levels or signaling and virus replication are positively correlated.Exogenous CCL2 increases HIV replication 23 and CCL2 neutralization decreases HIV replication 22 .We found that knockout of either CCL2 or CCR2 slightly reduced EcoHIV infection of BMM in culture depending upon the assay; however, CCL2 KO did not affect infection of macrophage in mice suggesting that the cellular and factor environment for EcoHIV infection in vivo is less dependent upon CCL2 than in macrophage culture.
In mice, bindarit treatment prior to infection impeded macrophage exit from the peritoneal cavity, increased EcoHIV replication in spleen, fully blocked EcoHIV infection of the brain, reduced infection of peritoneal macrophages, and prevented development of HAND.Bindarit treatment of uninfected mice showed no effects upon cognitive function in RAWM.This effect is presumed to re ect bindarit inhibition of CCL2 transcription 46 , however in other disease models it can also affect expression of other cytokines 53 .Administration of bindarit following virus infection enabled us to determine whether CCL2 functions are required to maintain NCI.We found that bindarit failed to affect NCI in mice with established disease, also it failed to affect virus burden in spleen or the brain.Its apparent increase of virus burden in macrophages may arise from its inhibition of macrophage exit from the peritoneal cavity.These ndings indicate that cells resident in the brain, presumably perivascular macrophages and microglia, maintain EcoHIV infection and its pathogenic effects upon cognitive function regardless of the presumed reduction in CCL2 driven monocyte migration to the brain.These ndings, consistent with other reports 24 , indicate that CCL2 is the major chemokine responsible for monocyte migration to the brain and its absence signi cantly impairs HIV brain infection and disease.Buprenorphine has recently been shown to reverse EcoHIV NCI in mice 54 ; possibly through its action on CCL2 expression 38 .This suggests that buprenorphine reverses NCI through inhibition of monocyte migration, however, buprenorphine also has effects within the brain upon various aspects of behavior 55,56 , so the mechanism of buprenorphine action upon EcoHIV NCI remains under study.
If, as shown here, once EcoHIV infects the brain, resident cells maintain NCI, then treatments that target EcoHIV pathogenesis in the brain should reverse NCI.This prediction was con rmed by three interventions in chronically EcoHIV-infected mice with NCI.Dysregulated brain gene and protein expression as well as NCI during EcoHIV infection were normalized by intranasal insulin treatment three months after infection 9 .Amelioration of cognitive defects by insulin have also been observed in Alzheimer's disease and amnestic mild cognitive impairment 57 .Likewise, administration of a glutamine antagonist, 6-diazo-5-oxo-l-norleucine, to mice to block neuropathogenic effects of excess glutaminase in the central nervous system as observed during EcoHIV infection reversed NCI and restored normal levels of glutamate in CSF and glutaminase in microglia 11 .This intervention targets dysregulation of metabolism in the brain as observed post-mortem in PLWH.In another approach to address HIV brain disease, it was shown that avoidance behavior and depression in EcoHIV-infected mice is accompanied by dysregulation of synthesis of extracellular vesicles (EV) and ceramide in the brain and that EV synthesis and behavior in infected mice can be normalized three weeks after infection by a novel neutral sphingomyelinase inhibitor, (R)- Of particular relevance to the work presented here are reports that CCL2 overexpression by a promoter variant increases the risk of HIV associated dementia in human beings 26 and the extent of this overexpression correlates with the extent of cognitive impairment observed 27 .Although direct evidence of the impact of CCL2 overexpression upon macrophage entry into the brain in PLWH is not available, a similar phenomenon of increased macrophage entry into tissue correlated with the CCL2 promoter variant during cell migration to the kidney during SLE nephritis 26 .The results presented here place the role of CCL2 in NCI as the signal to HIV-infected monocytes to enter the brain where infection and disease ensue.
We nd that continued cognitive impairment is independent of CCL2 and further monocyte migration or response.These ndings suggest that reversal of HAND cannot rely upon treatments that act only in the periphery but must target neuropathogenic events within the brain itself.

Supplementary Files
This is a list of supplementary les associated with this preprint.Click to download.

Figures
Figures

Figure 2 HAND
Figure 2