Osteopontin/Secreted Phosphoprotein-1 Behaves as a Molecular Brake Regulating the Brain Translocator Protein-Dependent Neuroinflammatory Response to Chronic Viral Infection


 Background

Osteopontin (OPN) as a secreted signaling protein, is dramatically induced in response to cellular injury and neurodegeneration. Microglial inflammatory responses in the brain are tightly associated with the neuropathologic hallmarks of neurodegenerative disease, but understanding of the molecular mechanisms remains in several contexts, poorly understood.
Methods

Positron emission tomography (PET) neuroimaging using radioligands to detect increased expression of the translocator protein (TSPO) receptor in the brain, is a non-invasive tool used to track neuroinflammation in living mammals.
Results

In humanized, chronically HIV-infected mice in which OPN expression was knocked down with functional aptamers, uptake of TSPO radioligand, DPA-713 was markedly upregulated in the hippocampus, cortex, olfactory bulb, cerebellum and significantly increased in other key brain regions analyzed compared to controls. TSPO+ microglia were detected by immunolabeling of post-mortem brain tissue, thus validating the neuroimaging findings. Unexpectedly, two types of neurons also selectively stained positively for TSPO. The reactive cells were the specialized neurons of the cerebellum, Purkinje cells, and a subset of tyrosine hydroxylase positive neurons of the substantia nigra. Two-way ANOVA of immunoreactivity revealed that a well-validated marker of microglial activation in brain tissue, ionized calcium-binding adaptor molecule-1 (Iba-1) was significantly increased, in an interaction that depended on HIV replication. Interestingly, similar analyses of TSPO immunoreactivity showed a significant interaction with OPN expression.
Conclusions

Collectively, these findings using a model of chronic HIV-infection revealed for the first time two key findings: 1) two different pathways of neuroinflammation are activated, and 2) osteopontin acts as a molecular brake regulating in the brain, the inflammatory response to HIV infection.


brain.NO
.Cg-Prkdc scid IL2rγ tm1Wjl /SzJ (NSG) mice were selected for this study as several publications demonstrated that this strain could be reproducibly engrafted with human immune cells and productively infected with HIV.Moreover, in ammation of the meninges and a small number of HIV-infected immune cells could be detected in the brain parenchyma (15).With 90% power to detect a 30% difference in our experimental variables (HIV, OPN) using a twosided test with alpha = .05,we determined that eight mice in each group would be required.NSG mice are susceptible to graft-vs-host disease and thus additional animals were included in the event that ~ 5% of mice had to be euthanized.Animals (males and females were included equally in each grouping) were assigned to one of the four groups in a 2 × 2 treatment paradigm: two groups of TBS-injected animals served as uninfected controls and two groups were infected with HIV SF162 , strain originally isolated from the brain of an HIV-infected individual with toxoplasmosis of the brai

(as describe
below) (25).One group of TBS-injected, and one group of HIV-infected mice were intraperitoneally injected weekly with functional RNA aptamers ("OPN-R3").The OPN-R3 aptamers blocked the interaction of osteopontin with its receptors and stimulate a feedback loop that reduces gene an

protein exp
ession (26).The second groups of TBS-, or HIV-infected animals received mutated aptamers ("mutant OPN-R3") with sequence alterations that prevent engagement of OPN with aptamer receptors and thus, served as negative controls.The animal handlers were not blinded to each group due to the biosafety level 2 requirements of working with HIV-infected mice.

Generation of humanized mice.The animal protocol and all procedures were reviewed and approved by the Johns Hopkins University Animal Care and Use Committee and Institutional Review Board.Male and female NSG mice (5-8 weeks old) were purchased from Jackson Laboratory and maintained under pathogen-free conditions.Mice were rst acclimated to the suite with automatic watering system for two weeks before monogamous pairing.Mice had free access to food and water and were housed in automatically controlled light conditions (light 7 am-9 pm).To avoid secondary microbial infections of these immunocompromised mice, cages were changed in a vented biosafety cabinet.All work surfaces, as well as cages, were thoroughly sanitized with chlorohexidine solution.Two-day old pups were irradiated with 100 cGy using a Gamma Cell-40 Cesium Extractor (Theratronics) followed by sedation with iso urane in an induction chamber.The sedated pups were subjected to intrahepatic injection of 1-2 × 10 cells/mouse of human CD34 + HSCs in 50-60 µl of sterile PBS with a 30-gauge needle (HSCs from cord blood from STEM Cell, Inc. or fetal liver from Advanced Bioscience Resources using magnetic bead separation, Miltenyi Biotec, 113-117-043).The engrafted pups were weaned at 3-4 weeks of age and males and females housed as separate groups.Peripheral blood (~ 50 µl) taken by facial bleed was sampled at regular intervals to quantify human leukocytes by seven-color ow cytometry with the following antibodies in a protocol optimized for small mouse blood volumes: human pan-CD45-Viogreen or PerCP (Miltenyi Biotec, 130-110-638, ThermoFisher MHCD4531), CD3-PE (ThermoFisher, MHCD0304), CD4-Vioblue (Miltenyi Biotec,130-113-219), CD8-APC-Vio770 (Miltenyi Biotec, 130-113-155), CD14-APC (BioLegend, 301808), CD19-PerCP (eBiosciences, 11-0199-42; BioLegend, 302228) and mouse CD45-FITC (eBiosciences,11-0451-82). Three to four months after engraftment, mice were sedated with iso urane and intravenously injected with 1 × 10 5 tissue culture infectious doses/ml of HIV SF162 in ~ 50 µl or wi h TBS buffer using a 25-gauge needle.The HIV-infected and buffer-injected mice were housed in separate same-sex social groups (when possible).

Treatment with inhibitory aptamers.Two weeks after infection, mice were injected (intraperitoneal) with 1 mg/kg of HPLC-puri ed RNA aptamers against OPN (IBA Lifesciences).The sequences for the aptamers were: mutant OPN-R3 (CGG-CCA-CAG-AAU-GAA-UCA-UCG-AUG-UUG-CAU-AGU-UG) and OPN-R3 (CGG-CCA-CAG-AAU-GAA-AAA-CCU-CAU-CGA-UGU-UGC-AUA-GUU-G).Both aptamers were modi ed at the 2'-O position with a methyl group, and the OPN-R3 aptamer included phosphorothioate modi ed bases (PTO) to help protect the nucleic acids from rapid attack by endo-and exonucleases (as given in reference (26).HPLC-puri ed lyophilized aptamers were refolded by resuspension in Folding Buffer (1 x PBS 8 mM sodium phosphate dibasic; 2 mM potassium phosphate monobasic;137 mM NaCl; pH 7.4,) with 1 mM MgCl 2 , heated at 95 o C f r 5 minutes (min) and allowed to cool to room temperature for 15 min before use.For the PTO modi ed OPN-R3 aptamer, an additional reduction step was required and performed using Tris(2-carboxyethyl)phosphine (TCEP) disul de reducing gel slurry (ThermoScienti c).A volume of TCEP slurry equal to the volume of aptamer solution was added to a microfuge tube with 3 volumes of Folding Buffer, vortexed then centrifuged for 1 min at 1000 x g and the supernatant discarded.This step was repeated.The cooled aptamer from above was added to the washed slurry, vortexed brie y and the tube allowed to incubate at room temperature for 10 min with mixing after the rst 5 min.The tube was centrifuged as above and the supernatant containing the reduced aptamer collected into a clean microfuge tube for use as described above.

HIV Viral Load Quanti cation.RNA was isolated from 56 mice using QIAamp Viral RNA mini kit (Qiagen) and concentrations determined using NanoDrop 2000 (Thermo Scienti c).Samples of insu cient quality were treated with DNase to remove contaminating DNA ( ife Technologies, Carlsbad, CA), followed by ethanol precipitation to concentrate any samples with less than 10 ng/µl of RNA.Samples were mixed with 1/10th volume of 3M sodium acetate, 2.5 volumes of ice-cold 100% ethanol, and 1µ l of 20 mg/ml glycogen, and kept at -80 o C for one hour.Samples were then centrifuged at 4 o C for 30 minutes, followed by two 70% ice-cold ethanol washes of the pellet.Pellets were resuspended in the appropriate volume of sterile, deionized water.RNA of insu cient concentration and quality were not used in the nal results (7 out of 47 total samples).RNA was reverse transcribed in a 20 µl reaction using 50 ng/µl of random hexamers, 10 mM dNTP mix, 5X First-Strand Buffer, 0.1M DTT, RNaseOUT recombinant RNase Inhibitor, 10 pg-5 µg of RNA, sterile water, and SuperScript III reverse transcriptase (Invitrogen/ThermoFisher). For Taqman qPCR, a custom design was optimized using forward primer Seq162F: 5'CGA-ACC-CAG-ATT-GTA-AGA-CT, reverse primer Seq162R: 5'ACA-TGC-TGT-CAT-CAT-TTC-TTC and probe: 5'-FAM/AG-CAT-TAG-G/ZEN/A-CCA-G

-GCT-ACA-CT/3IABKFQ (I
tegrated DNA Technologies).Ampli cation was performed with TaqMan Fast Advanced Master Mix, with conditions as follows: UNG incubation ) were used to create a 10-fold dilution standard curve performed on the cDNA with a lower detection limit of 1.0 copies/ml.The slope was − 3.759 and the R 2 value was 0.995, which both fell within the acceptable range for real-time PCR assays.

Micro-Positron Emission Tomography (PET) imaging.Animals to be imaged were moved to the small animal PET imaging room at least 1 hour before the experiments commenced.A dedicated small animal PET scanner (eXplore VISTA; GE Healthcare) and small animal CT scanner (X-SPECT/CT; Gamma Medica) were used.In all experiments, the PET scan was conducted for two mice simultaneously with the same batch of [ 11 C]DPA-713.Batches of [11C]DPA-713 were synthesized at high speci c activity and radiochemical purity at the Johns Hopkins PET Radiotracer Center according to the literature (27).One female mouse from a pair represented a control group (Buffer-OPN +/− ) whereas the other female mouse represented an experimental group (HIV-infected-OPN +/− ).Mice were anesthetized with iso urane and catheterized and PET-imaging began immediately after an intravenous bolus injection of the tracer.For the quantitative analyses of uptake intensity, the following tracer parameters were recorded: mouse weight, tracer speci c activity, amount of tracer injected, amount of tracer remaining in the syringe, time at which residual was measured, and start time of scan.Dynamic PET scans were acquired for 60 min (20 sec x 3, 30 sec x 2, 1 min x 2, 2 min x 3, 5 min x 10) and a CT scan was acquired immediately after for the purpose of localizing brain regions as described below.The individual performing the analyses were blinded to the identity of the experimental groups.PET images were re onstructed using an iterative 2D ordered-subject expectation-maximization method, using a trans-axial pixel size of 0.39 mm and axial slice thickness of 0.78 mm [1].No attenuation and scatter corrections were applied, as they have relatively small impact on mouse brain imaging.Image analysis was performed using the PMOD software package (v3.7,PMOD Technologies Ltd, Zürich, Switzerland).For each study, the reconstructed PET and CT brain images were rst co-registered; the existing Mouse Brain Template Volumes of Interest (VOIs) [2,3] were then "morphed" to match the brain image of the fused PET-CT; nally, the pre-de ned VOIs were then applied to the dynamic PET data to generate time-activity curves (TACs) in the unit of percent injected dose per gram of tissue (%ID/g).Tissue harvest.Mice were deeply anesthetized with iso urane, then subjected to intracardiac puncture for terminal blood collection followed by perfusion with cold PBS buffer.The brain was isolated from the skull intact, cut along midline in the sagittal plane into two halves, with one xed in 4% paraformaldehyde for ~ 24 hrs and later transferred to 70% ethanol before further processing for para n embedding and sectioning into 5 µm slices (Johns Hopkins Path Services).The remaining half of brain was placed in RNAlater (ThermoFisher) overnight and frozen within 18-24 hrs at -80 o C before RNA extraction.RNA from the brain was isolated using Monarch Total RNA Miniprep Kit (New England Biolabs) and concentrations were determined using NanoDrop 2000 (Thermo Scienti c).Each brain sample was thawed from − 80 °C and 30 mg slices were homogenized with 600 uL of 1X DNA/RNA protection reagent.60uL of Protease K reaction Buffer and 30 uL of Protease K were added, and then the tube was vortexed.Each sample was incubated at 55 °C for 7 minutes, followed by the addition of one volume of lysis buffer.Samples were vortexed well, and then transferred to a gDNA Removal Column.After spinning down for 30 seconds at 16,000 x g, the elute was combined with one volume of 100% ethanol and mixed by pipette.The mixture was transferred to an RNA puri cation column and spun for 30 seconds.The speed remained at 16,000 x g for all centrifugation steps.After the rst spin in the RNA puri cation column, the ow through was discarded and extra RNA enhancement steps were performed as described in the Monarch Total RNA Miniprep Kit.Following these extra RNA enhancement steps, 500uL of RNA Priming Buffer was added to the column and spun for 30 seconds.Two washes with RNA Wash Buffer were preformed, the second lasting for 2 minutes instead of 30 seconds.Finally, the column was transferred to a sterile 1.5 mL RNase free microfuge tube and eluted with 60uL of nuclease free water by a 30 second spin.

Immunohistochemistry. After warming for 10 min at 60 o C, brain sections were depara nized in a 100%, 95%, 70% alcohol series with nal rehydration in 1X tris-buffered saline-TBS (0.02 M Tris-HCl, 0.15 M NaCl, pH 7.4 #351-086-101).Sections were incubated with proteinase K (Electron Microscope Sciences) at 37 o C for 20 min then immers d in citrate buffer pH 6.0 at 99 o C for 10 min then allowed to cool at room temperature for 20 min.Slides were placed in 1X TBS then treated for 10 min with Bloxall (VectorLabs) and 3% hydrogen peroxide to inactivate tissue peroxidases and alkaline phosphatases (AP).Slides were then incubated with 2% Donkey sera/0.3%Triton X-100, 0.1% Tween-20 for one hour at room temperature, followed by blocking with Donkey anti-mouse F ab (1:25; #715-007-003, Jackson ImmunoResearch Laboratories, Inc) for 1 hour.Sections were incubated with antibodies at 4 o C for at least 16 hrs.For double labeling, staining for TSPO was performed rst followed by detection of Iba-1.Antibodies used: goat anti-mouse OPN (R&D Systems #AF808; tissue blocked in 10% rabbit serum/0.3%Triton X-100, 0.1% Tween-20), mouse anti-TSPO (1:100, MA5-24844, Invitrogen/ThermoFisher), mouse anti-Iba-1 (MA5-27726, Invitrogen/ThermoFisher), anti-tyrosine hydroxylase (1:500, #Ab112, Abcam).Slides were rinsed in 1X TBS then incubated with 1:500 dilution of donkey anti-mouse secondary antibodies conjugated to AP (#AP15999-anti-goat, MilliporeSigma, #715-056-151-anti-mouse, Jackson ImmunoResearch Laboratories, Inc) or horseradish peroxidase (HRP) (#715-036-150, Jackson ImmunoResearch Laboratories, Inc).Slides were developed with the Vector Immpact Red and DAB kits (#SK-5105, #SK-4105, Vector Labs) followed by dehydration and mounting in Cytoseal 60 (#8310-4 Richard-Allan Scienti c, ThermoFisher).Images were captured on a Zeiss Axio Observer A1 inverted microscope (20x objective), and a digital copy of the raw image was adjusted in the same manner for each, for optimal brightness and contrast, using Adobe Photoshop CS5.1.Staining intensity was quanti ed using ImageJ/Figi 2.0.A binary image was created and the threshold for the maximum and minimum intensity values in each channel applied.The pixel area was calculated using the measure function.

HIV-1 RNA in situ hybridization with RNAscope.Chromogenic RNA in situ hybridization was performed using kits and protocol purchased from ACD Biosciences.HIV SF162 target probes were custom designed to target Env-Nef sequences obtained from the accession number M65024.1 (2621-3701).PFA-xed para n embedded brains and spleens were sectioned (10 µM) and placed on SuperFrost Plus slides (Johns Hopkins Path services).Slides were baked in a slide oven for 1 hour at 60 °C prior to depara nization.Slides were incubated twice in Histoclear for 5 minutes and washed in 100% ethanol twice for 1 minute.Sections were incubated in Bloxall (VectorLabs) for 10 minutes at room temperature to quench endogenous peroxidases.Slides were then washed with distilled water and moved to a steamer containing 2 slide incubation chambers.One contained distilled wate and the other contained 1X RNAscope target retrieval buffer heated to at least 99 °C.Slides were temperature adjusted by gently agitating them for 10 seconds in the distilled water chamber and then moved to target retrieval buffer and allowed to boil for 15 minutes.Slides were then washed in distilled water and transferred to 100% ethanol for 3 minutes and then allowed to dry.Hydrophobic barrier pen was used to make a boundary around the tissue.Proteinase K treatment was applied to the tissues and incubated in a HybEZ oven at 40 °C for 15 minutes.Slides were washed with distilled water and incubated with pre-warmed target probes for 2 hours in the HybEZ oven at 40 °C.Slides were saved overnight in in 5X SSC buffer.Further amplication of target probe signal was performed the following day according to the manufacturer's instructions (RNAscope 2.5 HD detection protocol).Amp 1 was applied and incubated in HybEZ at 40 °C for 30 minutes.Slides were washed in RNAscope wash buffer and treated with Amp 2 and incubated in HybEZ at 40 °C for 15 minutes.Following another wash Amp 3 was applied and incubated in HybEZ at 40 °C for 30 minutes.Amp 4 was applied after another wash and incubated in HybEZ at 40 °C for 15 minutes.Amp 5 was applied following another wash and incubated at room temperature for 30 minutes.Amp 6 was applied before signal detection and incubated at 15 minutes at room temperature.Fast red was detected by combining Red-A and Red-B (1:60) and added to the sections which developed in 10 minutes at room temperature.The slides were counterstained with 30% Gill's Hematoxylin I (MilliporeSigma GHS132) and allowed to dry on a 60 °C slide warmer for 15 minutes.Slides were mounted with Cytoseal 60 (#8310-4 Richard-Allan Scienti c, ThermoFisher) and imaged on ECHO Revolve microscope at 40x magni cation.

Statistical analyses.Data were analyzed using ordinary one-way ANOVA and subsequent Tukey's test for multiple comparisons or Kruskal-Wallis test, followed by Dunn's multiple omparison test (non-parametric data).To determine whether there was any interaction between knockdown of OPN or not and HIV infection, data were also analyzed by two-way ANOVA with Spearman's test for heteroscedasticity, and normality of residuals (Anderson-Darling, D'Agostino-Pearson omnibus as ± standard deviation (SD), and a minimum p value of 0.05 was estimated as the signi cance level for all tests (GraphPad Prism 8).ANOVA two-tailed Student's t test and signi cance (P < .05).


Results


A. A model of chronic low-level HIV persistence in CD34 + humanized mice

Despite the lack of robust HIV replication in the brain of NSG-hCD34 mice, which lack human microglia, productive replication in human target T-cells and myeloid cells in the CNS and periphery induces microgliosis, astrocytosis and alterations in neuronal metabolites (15).We adapted this model to test the hypothesis that osteopontin (OPN) also known as secreted phosphoprotein-1 (SPP1), is required for HIVmediated in ammation in the central nervous system (CNS).The experimental design included four groups composed of male and female mice engrafted with human CD34 + hematopoietic stem cells (hCD34, HSCs), which differentiate into mature human T-cells and macrophages that permit infection with HIV.Two groups were intraveneously inoculated with buffer, and the other two groups with HIV SF162 .

After two weeks, mice were injected weekly with either functional (HIV OPN − ) or mutant aptamers (HIV OPN + ) that do not block OPN expression (26).We make note that due to 1) the capability to conduct microPET-imaging on two mice at a time and 2) the need to synthesize the radiotracer fresh each time because of its short-half-life, we had to limit the number of mice that could be imaged and given these limitations, decided to test only females based on a recent publication suggesting sex differences in HIV CNS disorders (28).We also note that fewer data points at earliest time periods after infection.We discovered weeks later that the e ciency of viral RNA recovery after low temperature freeze/thaw of mouse blood was ine cient and needed to be performed within 24 hrs after collection.Quantitative PCR analyses of virus release into the plasma at 10 and 16-weeks after aptamer treatmen revealed no signi cant differences between experimental groups (Fig. 1SA-B).However, analyses by two-way ANOVA demonstrated that time accounted for 20% of the variance between the viral loads at the 10-and 16-week time points (n = 3 HIV-OPN+, n = 8 HIV-OPN-, P = .0369).Similarly, analyses of virus release into the plasma over a period of 16 weeks (112 days) after infection revealed a wide variation in replication levels among individual female mice and overall, no signi cant differences among the two groups were detected (Fig. 1A).However, in HIV OPN + mice, replication increased from the initiation of aptamer treatment until the last sampling point at 98 days after infection, a trend that was less apparent in the HIV OPN − group (Fig. 1B).During this same time period, ow cytometric analyses of circulating white blood cells in the plasma revealed a gradual and stark decline of hCD45 cells in the HIV OPN + , which contrasted with the higher levels detected in the HIV OPN − group, however the differences in slopes were not quite signi cant (Fig. 1C, F = 3.800.DFn = 1, DFd = 17, P = 0.0680).While a trend in the decrease in the absolute percentage of hCD4 cells was similar in both groups, over time the decline in hCD8 cells in the plasma of HIV OPN − was less compared to the HIV OPN + group, but neither trends were signi cant (Fig. 1D-E, CD4, F = 0.3051.DFn = 1, DFd = 17, P = 0.5879; CD8, F = 1.489.DFn = 1, DFd = 17, P = 0.2390).Therefore, we surmised that the detection of virus replication in the plasma at the 16-week study endpoint was likely due to a combination of virus release from productively infected circulating immune cells and tissue reservoirs.To detect HIV RNA (vRNA) expression in the brain and other tissues, we used in-situ hybridization or RNAscope.Large foci of vRNA positive cells were readily detected in spleen, a tissue rich in HIV target T-cells and macrophage populations (Fig. 1F).In mice with the highest viral loads, vRNA near the brain vasculature and in the parenchyma in both the infected aptamer treated groups were detected, but infrequently (Fig. 1F).

To ascertain the effectiveness of the aptamers in reducing OPN gene expression, the entire gastrointestinal (GI) tract collected at the study endpoint was immunostained for OPN protein.Because the GI tract is the largest tissue system in the body, it was used as a proxy for assessing the e ciency of the aptamer treatment.While there were no differences in OPN reactivity between the buffer treated groups, signi cantly elevated staining was detected in the small intestines of HIV OPN + mice in contrast to infected mice injected with functional aptamers (HIV OPN − group) (Fig. 1G & H).This d fference in reactivity was in uenced by highly signi cant interactions with HIV (two-way ANOVA, P < .0001)and OPN levels (two-way ANOVA, P = .0003)(Fig. 1H, right graph).Quantitative RT-PCR analyses on RNA puri ed from brain tissue showed a trend of decreased message in both the buffer OPN − and HIV OPN − mice treated with functional aptamers, but the differences did not reach statistical signi cance (Fig. 1I).

B. Increased uptake of TSPO ligand in mice with decreased osteopontin (OPN) expression To determine the impact of chronic HIV infection in the presence of normal or lowered OPN levels on microglial activation in live mice, female animals in each gro

thetized and subjected to PET-imaging using the ligand [11C]DPA-713 at 12
weeks post infection (27).Females were chosen because, over the course of the study, anecdotal differences in their anxiety compared to male mice was observed.DPA-713 ligand binds to the translocator receptor (TSPO) expressed prominently on activated microglia (29), and to a lesser extent, on the endothelium, astrocytes, and neurons of the dorsal root ganglion (30).Signi cant differences in DPA-713 uptake between the two HIV-infected groups treated with functional (HIV OPN − ) or mutated aptamers (HIV OPN + ) and both buffer control groups in the 20-50 min time window after bolus injection were observed across the entire brain (Fig. 2A-C).Interestingly, the highest levels of uptake of DPA-713 ligand amongst the four experimental groups were seen in the HIV OPN − group (Fig. 2B-C).Analyses of s eci c brain regions showed signi cantly increased ligand uptake in the hippocampus, striatum, olfactory bulb, cortex, thalamus, cerebellum, amygdala, central grey matter, midbrain, basal forebrain septum, hypothalamus, and in the inferior and superior colliculi of HIV OPN − mice compared to the other experimental groups (Fig. 2D-R).In the right striatum, olfactory bulb, cortex, and basal forebrain in particular, mice in the HIV OPN − group that had the highest viral loads also showed the greatest uptake of DPA-713 ligand, suggesting a direct relationship between the extent of virus replication, and the amount of TSPO binding (Fig. 2E-G, 2N).Interestingly, in these regions, the uptake of label was the highest and sustained over time (Fig. 2 Supplemental, F, G, H, Q).In two regions, the right striatum and right amygdala, the HIV OPN + group had more ligand uptake than the buffer OPN − group (Fig. 2E, 2J).Analyses of time-activity curves and the early 8-10 min window after bolus injection did not show any signi cant differences in the amygdala among the experimental groups in DPA-713 uptake (Fig. 2Supplemental, K), which was in contrast to the 20-50 min window (Fig. 2J).It was notable that ligand uptake in the cortex and olfactory bulb was relatively fast and sustained over the 1hr imaging period (Fig. 2 Supplemental, G-H).

C. TSPO-Iba-1 immunoreactivity is abundant in cortical and hippocampal regions PET imaging with ligands that detect microglial cell activation as an indirect measure of neuroin ammation has been an active area of research as it allows insight into neuropathologic processes in living persons.Recent ndings have noted that cells other than microglia can upregulate TSPO expression (31).To determine whether increases in TSPO uptake detected by PET-imaging in these mice correlated with measures in brain tissue at the 16-week study endpoint, double-label IHC was performed to detect cells labeling for TSPO + (red color) and/or Iba-1 + microglia (brown color).There are relatively few published studies that have analyzed TSPO expression in brain tissue in conjunction with neuroimaging (31)(32)(33)(34).We rst conducted a qualitative survey of brain sections from each group of mice.Abundant TSPO reactive neurons and microglia of variable intensities were seen in the cortex (Fig. 3A, representative shown, Buffer OPN + ).TSPO + Iba-I + cells having rami ed or dystrophic phenotypes were readily detected in the midbrain (Fig. 3B, representative shown, HIV OPN+).In both HIV-infected groups (Fig. 3C iii and iv) the organization of the dentate gyrus appeared altered compared to the buffer control groups (Fig. 3C i and ii).Additionally, there were more cells within hippocampal ssures in the HIV-OPN + group compared to the other three groups (Fig. 3Ciii, red circles).TSPO labeling was strong within the hippocampal neurons and ssures and in a small number of mice, reactivity in the meningeal layer was also found (Fig. 3D-G).In contrast to the ordered array of Purkinje neur ns of the cerebellum in HIV-OPN + mice (Fig. 4A-C), in the HIV-OPN − group, the intensely stained TSPO + Purkinje neurons displayed a disordered pattern (Fig. 4E) that was not seen in the Buffer-OPN − group (Fig. 4D).Moreover, a select subset of neurons, which we suspected were located in the substantia nigra, reacted very strongly with TSPO antisera (Fig. 5A-F).The identity of this brain region as the substantia nigra region was con rmed by the speci c reactivity of these neurons individually with antisera against either TSPO or tyrosine hydroxylase (Fig. 5H-I).E. Iba-1 but not TSPO immunoreactivity is increased with osteopontin (OPN) suppression For postmortem immunohistochemistry and quantitative analyses, brains from mice that were subjected to neuroimaging were included in the analyses (Buffer-OPN + , n = 3; Buffer-OPN + , n = 5; HIV-OPN + , n = 3; HIV-OPN + , n = 3).In the HIV-OPN − group, one mouse died before necropsy and another from that group which was not imaged, was included to have n = 3.
Quantitative analyses of TSPO antibody reactivity (one-way ANOVA with Krusal-wallis test, Dunn's test) showed signi cantly reduced intensity in the buffer-OPN − (8.62 +/-5.24)compared to the buffer (13.73 +/-8.24) and HIV OPN + (13.7 +/-7.8)groups (Fig. 6A, P = .0087and P = .003,respectively).Interestingly, while no signi cant differences in TSPO immunoreactivity between the HIV-infected groups were detected, the presence or absence of OPN was a signi cant interacting factor (P = .0136,F (1,190), Fig. 6B).The HIV-OPN − group showed signi cantly elevated Iba-1 immunoreactivity compared to uninfected mice with normal expression of OPN (Fig. 6C, p = .0074,one-way ANOVA, Kruskal-Wallis, Dunn's test).Perhaps to the lower number of mice available for the analyses, the increase in Iba-I immunoreactivity did not reach statistical signi cance, between the HIV OPN + and HIV OPN − groups (Fig. 6C).Nevertheless, viral infection signi cantly drove the increased level of Iba-I immunoreactivity in the brain (Fig. 6D, two-way ANOVA p = .0051).Collectively, these results suggest that HIV-infection increases the number of activated microglia in the brain, while the level of OPN in this compartment had a greater impact on TSPO expression.


Discussion

Knowledge that there is continuous monitoring of proper neuronal function and integrity by microglia, the resident immune cells in the CNS, fundamentally altered the way in which we think about the role these cells play in health and disease (35,36).Signi cant disruptions of the homeostatic brain microenvironment, whether secondary to injury, neurodegeneration or viral infection, dramatically activate microglia.The result is upregulation of gene expression, morphologic conversion to a functional phagocytic phenotype, and the elaboration of pro-and anti-in ammatory mediators (37).Interestingly, recent detailed transcriptomic analyses of microglia from humans with neurodegenerative disorders, and the corresponding mouse disease models, revealed a common gene signature which includes osteopontin (OPN/SPP1) as a signi cantly enriched expressed gene in so-named "neurodegenerative" microglia (38)(39)(40).Evidence that OPN is signi cantly elevated in the plasma and/or CSF of people with neurodegenerative disorders and in HIV-associated cognitive impairment or HAND, has been reported (24).Prior studies have examined a role for OPN in protecting neurons after injury and stroke, but few have examined its potential functional impac on microglia (41).Our prior studies implicated OPN as an upregulated protein in HIV-infected macrophages that also stimulates viral replication through a NF-κbdependent mechanism (22).We went on to demonstrate, using post-mortem brain tissue from individuals with HAND, that neurons as well as microglia, show elevated OPN expression (19).OPN knockout mice are viable and display no gross physical or behavioral abnormalities but do show disorganized wound remodeling and defective macrophage in ltration after injury or infection (42,43).This led us to our current study to determine whether or not OPN is required for HIV-mediated neuroin ammation in the brain.

TSPO, formerly known as the peripheral benzodiazepine receptor, is a small 18 kDa protein found in the outer membrane of mitochondria whose expression is signi cantly increased in the brain after injury and neuroin ammation (reviewed in (18)), however there is no consensus about its exact physiologic roles (18).The availability of effective tools to clinically evaluate HIV-infected persons for neuropathologic evidence of cognitive impairment by non-invasive methods has been a critical focus of the eld (29).TSPO PET-imaging with rst-generation ligands showed promise in both the rhesus macaque models of HIV brain disease (44) and in humans (45).Imaging with the second-generation ligand DPA-713 revealed signi cant abnormalities in the frontal cortex of HIV-infected persons with severe cognitive impairment (29).A more recent association study found positive correlation between increased DPA-713 uptake and neurocognitive performance in treated-HIV-infected persons in line with prior human imaging studies (28,(46)(47)(48).

The microPET-imaging results from this study strongly suggests that, rather than potentiating HIV injury, OPN acts as a molecular brake helping to dampen the microglial in ammatory response.In this regard, TSPO levels quanti ed by IHC were higher in OPN + compared to OPN − mice, with signi cant increases seen in the buffer-injected group.Moreover, additional analyses revealed an interdependence between TSPO and OPN levels that warrants further investigation to understand the disconnect between our neuroimaging and IHC ndings.The detection of TSPO expressio in cortical, Purkinje and striatal neurons by IHC, suggests that these cells also contribute to the overall in ammatory state in the brain, but the underlying neurobiological basis for this nding is unknown (18).In this regard, a recent report describes low-level basal expression of TSPO in several regions of the brain including vascular endothelial cells and in Purkinje neurons in normal mouse brain compared to tissues from a TSPO knockout mouse (33).Moreover, TSPO can also be expressed in astrocytes, and these cells were not assessed in this study.Early intraneuronal accumulation of toxic amyloid in a mouse model of Alzheimer's disease was recently shown to initiate in ammatory gene expression of chemokines CCL2 and CCL3 in pyramidal neurons of the hippocampus, thus demonstrating that neurons have the capacity to initiate such signaling (49).Clearly, additional studies are needed to have a more complete understanding of the neuroin ammatory response.

In contrast, for Iba-1, a well-established marker for microglial activation, elevation was dependent on the extent of HIV replication.This nding is in alignment with early neuropathological ndings of microgliosis in post-mortem brains of HIV-infected individuals (50).The association of activated microglia and increased secretion of proin ammatory cytokines in plasma and CSF are neuropathologic hallmarks in HAND (50).However, with antiviral treatment, cellular activation related to ongoing virus replication is greatly diminished, though not completely eliminated (51).Moreover, our ndings demonstrate that there may be two or more different signaling pathways targeting microglia, TSPO and/or Iba-1, which are activated by HIV infec

on and coul
potentially explain why, in TSPO knockout mice, microglia remained, able to respond appropriately to neuronal injury (32).

Ruling out underlying genetic susceptibility to dementia, there are competing ideas, supported by data suggesting that cognitive de cits could be related to a legacy effect (52) or low-level chronic HIV replication in the brain (53), both of which would in ict accumulating damage to neural networks over the span of time.Our mice model in which HIV RNA expression was detected in the brain in only a few cells per ve-micron section represents the latter, low-level chronic infection modeling approximately 48-55 human years.Nevertheless, only by the disruption of OPN function could we detect by microPET-imaging the profound impact HIV has on promoting neuroin ammation.


Conclusions

Although OPN has diverse functions in different physiological and pathological processes, much evidence supports it having a central role in in ammatory signaling (24).It was identi ed as a highly phosphorylated protein puri ed from osteoclasts, which are the macrophages of the bone (54,55), but it can also be made by T-cells and broblasts (24).Mice express a single OPN allele.In contrast, human cells can express one or more of three splice variants of OPN, leading to an intracellular or extracellular form of protein which is further subject to proteolytic cleavage that can generate peptides possessing signaling activity (reviewed in (24)).However, the signi cance of these variants in regulating neuroin ammation has not, to our knowledge, been studied in the CNS.Our ndings suggest that a deeper understanding of OPN-regulation of innate immune signaling of neuroin ammation would be extremely valuable in the search for novel approaches which are urgently needed to promote the return to homeostasis in the CNS microenvironment in HAND and other neurocognitive disorders.Iba-1, but not TSPO mmunoreactivity is increased in the brains of mice suppressed for osteopontin.(A-D) Quantitative analyses of TSPO or Iba-1 antibody reactivity, Buffer-OPN+ n=3; Buffer-OPN-n=5, HIV-OPN+, n=3; HIV-OPN-, n=3; one-way ANOVA with Krusal-wallis test, Dunn's test or with two-way ANOVA to determine any interaction between osteopontin levels or HIV infection.The mean and standard deviation is shown.(A) While there was no difference in TSPO immunoreactivity in amongst the HIV groups (B) a signi cant interaction between TSPO levels and OPN (P=.0002) was observed.(D) A signi cant interaction between Iba-1 protein levels and HIV infection was found (P=.0051).


List Of Abbreviations


Supplementary Files

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

Fig1S.pdf FIGURE2SUPPLEMENTAL.pdf



50 o C, 2 min, one cycle; enzyme activation, 95 o C, 20 sec, one cycle; 40 cycles of denature 95 o C, one sec, anneal, 48 o C, 4 sec; extension, 60 o C, 16 sec QuantStudio 3 (Applied Biosystems).RNA isolated from viral partic es (HIV-1 SF162


1
1
Figures


Figure 2 Increased
2
Figure 2


Figure 3 TSPO
3
Figure 3


Figure 4 Robust
4
Figure 4


Figure 5 Strong
5
Figure 5


Figure 6 Page 27 / 27
62727
Figure 6






Acknowledgements.A.M.B. acknowledges Gilbert Green for training received in the early stages of this project.Availability of data and materials.All data generated and analyzed during this study are included in this article and its supplementary information les.Funding.This study was supported by funding from the USA, National Institutes of Neurological Disorders and Stroke (NINDS), R01NS102006 to A.M.B.DeclarationsEthics approval for research involving animals.Johns Hopkins University is fully accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care that shares foundational principles articulated in the Basel Declaration.The research protocol was approved by the Johns Hopkins Institutional Review Board and the Animal Use and Care Committee and conducted using biosafety levelSupplementary MaterialsReferences
Molecular Imaging of Neuroin ammation in HIV. A Boerwinkle, B M Ances, Journal of neuroimmune pharmacology: the o cial journal of the Society on NeuroImmune Pharmacology. 1412019

. J Ghosn, B Taiwo, S Seedat, B Autran, C Katlama, Hiv, Lancet. 3922018. 10148

A Review of Management of In ammation in the HIV Population. J Slim, C F Saling, BioMed research international. 201634206382016

HIV-associated neurocognitive disorder before and during the era of combination antiretroviral therapy differences in rates, nature and predictors. R K Heaton, D R Franklin, R J Ellis, J A Mccutchan, S L Letendre, S Leblanc, J Neurovirol. 172011

HIV-associated neurocognitive disorders persist in the era of potent antiretroviral therapy: CHARTER Study. R K Heaton, Clifford Db , Franklin DrJr, S P Woods, C Ake, F Vaida, Neurology. 752010

et al Differences in Cognitive Function Between Women and Men With HIV. P M Maki, L H Rubin, G Springer, E C Seaberg, N Sacktor, E N Miller, Journal of Acquired Immune De ciency Syndromes. 7911999. 2018

Human immunode ciency virus-associated neurocognitive disorders: Mind the gap. J Mcarthur, J Steiner, N Sacktor, A Nath, Ann Neurol. 672010

Toward a biological de nition of Alzheimer's disease. Jack CrJr, D A Bennett, K Blennow, M C Carrillo, B Dunn, S B Haeberlein, Journal of the Alzheimer's Association. 1442018Alzheimer's dementia the

HIV-associated neurocognitive disorder -pathogenesis and prospects for treatment. D Saylor, A M Dickens, N Sacktor, N Haughey, B Slusher, M Pletnikov, Nat Rev Neurol. 122016

HIV's double strike at the brain: neuronal toxicity and compromised neurogenesis. M Kaul, Front Biosci. 132008

One Century of Evolving Concepts. M Prinz, Jung S Priller, J Microglia Biology, Cell. 17922019

The development and maintenance of resident macrophages. E G Perdiguero, F Geissmann, Nat Immunol. 1712016

Microglia emerge as central players in brain disease. M W Salter, B Stevens, Nat Med. 2392017

New tools for studying microglia in the mouse and human CNS. M L Bennett, F C Benne

, S A Liddel
w, B Ajami, J L Zamanian, N B Fernhoff, Proc Natl Acad Sci. 113122016

Associations between brain microstructures, metabolites, and cognitive de cits during chronic HIV-1 infection of humanized mice. M D Boska, P K Dash, J Knibbe, A A Epstein, S P Akhter, N Fields, Mol Neurodegener. 9582014

Loss of neuronal integrity during progressive HIV-1 infection of humanized mice. P Dash, S Gorantla, H Gendelman, J Knibbe, G Casale, E Makarov, J Neurosci. 312011

Links between progressive HIV-1 infection of humanized mice and viral neuropathogenesis. S Gorantla, E Makarov, J Finke-Dwyer, A Castanedo, A Holguin, C Gebhart, Am J Pathol. 1772010

TSPO in diverse CNS pathologies and psychiatric disease: A critical review and a way forward. T R Guilarte, Pharmacol Ther. 1942019

Cortical neurons are a prominent source of the proin ammatory cytokine osteopontin in HIV-associated neurocognitive disorders. K Silva, C Hope-Lucas, T White, T Hairston, T Rameau, A Brown, J Neurovirol. 212015

Osteopontin is increased in HIV-associated dementia. T H Burdo, R J Ellis, H S Fox, J Infect Dis. 1982008

Increased expression of monocyte CD44v6 correlates with the development of encephalitis in rhesus macaques infected with simian immunode ciency viru

lication and is incr
ased in the brain and cerebrospinal uid of HIV-infected individuals. A Brown, T Islam, R Adams, S Nerle, M Kamara, C Eger