Small molecule modulation of the p75 neurotrophin receptor suppresses age- and genotype-associated neurodegeneration in HIV gp120 transgenic mice

Abstract

The persistence of HIV in the central nervous system leads to cognitive deficits in up to 50% of people living with HIV even with systemic suppression by antiretroviral treatment. The interaction of chronic inflammation with age-associated degeneration places these individuals at increased risk of accelerated aging and other neurodegenerative diseases and no treatments are available that effectively halt these processes. The adverse effects of aging and inflammation may be mediated, in part, by an increase in the expression of the p75 neurotrophin receptor (p75^NTR) which shifts the balance of neurotrophin signaling toward less protective pathways. To determine if modulation of p75^NTR could modify the disease process, we treated HIV gp120 transgenic mice with a small molecule ligand designed to engage p75^NTR and downregulate degenerative signaling. Daily treatment with 50 mg/kg LM11A-31 for 4 months suppressed age- and genotype-dependent activation of microglia, reduced expression of p75^NTR, increased microtubule associated protein-2 (MAP-2), reduced dendritic varicosities and slowed the loss of parvalbumin immunoreactive neurons in the hippocampus. An age related accumulation of microtubule associated protein Tau was identified in the hippocampus in extracellular clusters that co-expressed p75^NTR suggesting a link between Tau and p75^NTR. Although the significance of the relationship between p75^NTR and Tau is unclear, a decrease in Tau-1 immunoreactivity as gp120 mice entered old age (>16 months) suggests that the Tau may transition to more pathological modifications; a process blocked by LM11A-31. Overall, the effects of LM11A-31 are consistent with strong neuroprotective and anti-inflammatory actions that have significant therapeutic potential.

Introduction

HIV rapidly enters the nervous system resulting in a chronic infection, inflammatory responses and progressive neural damage. The neural damage has been linked, in part, to the release of soluble factors from microglia (MG) and macrophages (M) (Giulian et al., 1990; Pulliam et al., 1991; Bragg et al., 2002; Kolson, 2002; Meeker et al., 2012; Meeker et al., 2016b) as well as direct neural effects of HIV proteins. These factors lead to cognitive dysfunction in up to half of the individuals infected with HIV. Although antiretroviral therapy (ART) has reduced the severity of cognitive dysfunction through effective long-term control of systemic virus, HIV persists in the CNS and cognitive decline continues. As HIV infected individuals maintained on ART approach advanced ages, they are also displaying an accelerated aging-like phenotype that may synergize with other age-associated diseases such as Alzheimer Disease (AD)(Martin et al., 2013; Harper, 2015; Gianesin et al., 2016; Levine et al., 2016; Cohen and Torres, 2017; Pfefferbaum et al., 2018). Synergy between HIV infection and aging may be mediated in part through the effects of inflammation on neurons which are similar in both disease processes(Stokin et al., 2005; Kuchibhotla et al., 2008; Bittner et al., 2010; Meeker et al., 2016a). The potential for synergy is highlighted in postmortem studies that have identified aggregates of amyloid beta (Aβ) in HIV infected patients although the direct role of HIV is still debated(Fields et al., 2018; Stern et al., 2018; Fulop et al., 2019).

Efforts to reduce inflammation and neural damage have focused on ways to protect neurons or block the destructive inflammatory processes. Studies employing in vitro models of HIV induced inflammation and neuropathogenesis have shown that modulators that enhance neurotrophin signaling may offer strong neuroprotection. Alterations in nerve growth factor (NGF) signaling via TrkA and brain-derived neurotrophic factor (BDNF) signaling via TrkB are thought to contribute to both HIV- and age-associated neural dysfunction(Fahnestock et al., 2001; Peruzzi et al., 2002; Mocchetti and Bachis, 2004b, c; Costantini et al., 2005; Nosheny et al., 2005; Cuello et al., 2007; Mufson et al., 2007; Nosheny et al., 2007; Capsoni et al., 2011; Liu et al., 2018). Neurotrophins and related ligands are protective in various rodent models of neuropathogenesis (Ebadi et al., 1997; Mocchetti and Bachis, 2004a; Longo et al., 2007; Mocchetti et al., 2007; Yang et al., 2008; Aloe et al., 2012; Capsoni et al., 2012; Knowles et al., 2013; Longo and Massa, 2013; Shi et al., 2013; Nguyen et al., 2014a; Simmons et al., 2014) and in natural aging (Xie et al., 2019). Thus, modulation of neurotrophin signaling has substantial potential for the inhibition of both age- and disease-related cognitive decline.

The p75 neurotrophin receptor (p75^NTR) has proven to be an attractive target for modification of neurotrophin signaling(Longo and Massa, 2013; Meeker and Williams, 2014). In the context of aging and AD, an increase in the expression of the p75^NTR along with decreased levels of NGF and increased levels of proNGF are thought to contribute to the increased vulnerability of the aging nervous system by shifting neurotrophin signaling away from beneficial NGF/TrkA signaling toward deleterious proNGF/p75^NTR signaling(Fahnestock et al., 2001; Counts et al., 2004; Costantini et al., 2005; Counts and Mufson, 2005; Ginsberg et al., 2006; Mufson et al., 2007; Capsoni et al., 2010; Capsoni et al., 2011; Cuello et al., 2012; Mufson et al., 2019). A novel, ligand, LM11A-31, has been developed that binds p75^NTR to down regulate degenerative signaling, upregulate survival signaling and prevent deficits associated with pro-neurotrophin signaling (Massa et al., 2006). The compound is orally bioavailable, crosses the blood brain barrier and has no known side effects at therapeutic concentrations. Using in vitro models of HIV neuropathogenesis, we have shown that low nanomolar concentrations of LM11A-31 provide neuroprotection against the effects of virus and the viral envelope protein, gp120(Meeker et al., 2012; Meeker et al., 2016a). Similar protective effects have been seen in a variety of in vitro and in vivo disease models including AD (Yang et al., 2008; Knowles et al., 2013; Simmons et al., 2014) Huntington disease (Simmons et al., 2016), traumatic brain injury (Shi et al., 2013) and models involving pathological accumulations of proNGF including spinal cord injury (Tep et al., 2013), bladder degeneration (Ryu et al., 2018) and juvenile indiopathic arthritis (Minnone et al., 2017).

In addition to direct neuroprotection, LM11A-31 may restrict inflammation. Studies of cultured human monocyte-derived macrophages (hMDM) have shown expression of TrkA and the p75^NTR which respond to their respective ligands, mature NGF and its precursor, proNGF, to control the phenotype of these cells. NGF was shown to induce a morphological and functional phenotype the opposite of proNGF suggesting a prominent role for neurotrophin signaling in the control of macrophage activity (Williams, et al. 2015). Macrophages stimulated with proNGF displayed a phenotype that enhanced damage to neurons whereas NGF was partially protective. This dichotomy in the actions of NGF versus proNGF is similar to the proposed actions on neurons where a shift in the neurotrophin balance to higher proNGF:NGF ratios predisposes neurons to degeneration during aging and Alzheimer disease(Fahnestock et al., 2001; Peng et al., 2004; Al-Shawi et al., 2008; Zeng et al., 2011; Mufson et al., 2019). These studies suggest that neurotrophin ligands may have beneficial anti-inflammatory effects on macrophages and microglia. This may be particularly true for a p75^NTR ligand given the role of proNGF/p75^NTR signaling in the activation of neurotoxic responses in macrophages.

The robust neuroprotective effects combined with potential anti-inflammatory effects suggest that LM11A-31 may have substantial therapeutic benefit. In the current study, we used the gp120 transgenic mouse engineered to express the HIV envelope protein gp120_IIIB to examine the ability of LM11A-31 to modify the development of inflammation and neural damage. The model recapitulates many of the features of HIV infection including the gradual development of inflammation making it a good model to investigate the interactions between HIV-associated inflammation and aging. LM11A-31 was found to suppress inflammation and neural damage in support of a potential therapeutic effect in HIV patients.

Materials and Methods

gp120 transgenic (Tg) mice

Breeder mice expressing the HIV gp120_IIIB envelope protein driven by the GFAP promoter(Toggas et al., 1994) were kindly provided by Dr. Marcus Kaul who has characterized the development of neuropathology in the mice(Kaul et al., 2007). Mice were bred at the University of North Carolina and experiments were conducted under an approved IACUC protocol in an ALAC certified facility. All mice were genotyped at weaning and randomly separated into treatment groups. Wild type littermates were used as controls. Since small numbers of mice homozygous for gp120 were obtained, heterozygous male and female mice were used in the experiments. Three major treatment groups were designated to assess the interaction of the gp120 phenotype with age: young mice (3–8 months), middle aged mice (12–16 months) and old mice (19–23 months). Males and females negative for the gp120 transgene (Wild Type, WT) were included as controls for each experiment.

Mice were dosed once daily with 50 mg/kg LM11A-31 bisulfate or vehicle (water) by oral gavage. The compound was prepared at a concentration of 5 mg/ml in sterile deionized water. Mice in the middle aged and old groups were treated for 3 months ending at the indicated times with some variation in the start times depending on mouse availability. After the last dose, each mouse was euthanized by the isoflurane drop method in accordance with the NIH Guide for the Care and Use of Laboratory Animals and perfused with 0.01 M PBS. The brain was removed and split in the sagittal plane. One half was fixed in phosphate buffered 4% paraformaldehyde and the other half frozen.

Tissue preparation and staining

The fixed brain was embedded in paraffin and sagittal sections were cut at 5 μm for subsequent staining. Sections were deparaffinized, incubated for 25 min in 1% H₂O₂ and gently boiled for 20 min in 0.1 M citrate buffer, pH 6.0 to unmask antigen. Sections were then washed 3x in PBS and incubated in 3% normal goat serum for 1 h at room temperature. Immunostaining was performed for MAP-2 (EMD Millipore AB5622, MAB 3418, Aves, MAP), Iba1 (Wako, 019–1941), GFAP (DAKO, Z0334), Tau-1 (EMD Millipore, MAB 3420), Tau T22 (EMD Millipore, ABN 454), p75^NTR (EMD Millipore, AB 1554, 07–476) and parvalbumin (Swant PV25, Sigma P3088). In addition, staining was done with antibodies to amyloid beta (EMD Millipore, MAB 348), phospho-Tau (AT8, Biolegend,137401), and C1q (Abcam, ab182451) but no positive staining was found (not shown). A thioflavin T stain to identify plaque-like material was also negative (not shown). For the primary comparisons, changes in immunostaining were compared between the gp120 Tg mice treated with LM11A-31 versus placebo.

Isolation and culture of human monocyte-derived macrophages

Human monocyte-derived macrophages were used to generate conditioned medium that would mimic the effects of inflammation in vitro. De-identified human buffy coat leukocytes were purchased and shipped within 24 hours after blood draw from healthy donors at the New York Blood Center (http://nybloodcenter.org/), a non-profit organization specifically for the collection and distribution of blood for clinical and research purposes. The proposed research was screened and approved by the New York Blood Center prior to initiation of the studies. Blood was diluted 1:1 with phosphate buffered saline (PBS) centrifuged at 500 X g for 20 min over Ficoll-Paque (GE Healthcare 17–1440-03) and the peripheral blood mononuclear cells (PBMCs) were collected from the interface. PBMCs were washed for 10 min with 10 ml of red blood cell lysis buffer to remove any red blood cell contamination, diluted to 40 ml with PBS and centrifuged at 100 X g for 15 min. After a second wash, the supernatant was aspirated and the pellet re-suspended in Dulbecco’s modified eagle medium (DMEM) with high glucose, 10% fetal bovine serum (Gibco 160000–044) and 20 μg/ml gentamicin (Gibco 15750–60). Cells were seeded into ultralow adhesion 6 well plates (Corning 3471) at a density of 10⁷ cells/well. PBMCs were cultured for 5–7 days to allow monocyte attachment. The remaining white blood cells were washed, from the plate yielding a pure monocyte/macrophage culture. The adherent cells were differentiated into monocyte-derived macrophages using human GM-CSF (15 ng/ml) in complete DMEM for 5 days. A previous comparison of GM-CSF with M-CSF indicated that the former yielded cultures with slightly greater neurotoxic secretory activity in response to HIV gp120. Following differentiation, the cells were maintained in complete DMEM.

Primary cultures of mouse microglia

Mouse microglia were obtained from E16 fetal mice and were cultured from explants of brain tissue. Mice were euthanized in isoflurane and the uterus was removed, washed briefly, and placed in sterile ice cold HEPES buffered Hank’s Balance Salt Solution (HBSS). The brain was removed from each fetus, washed sequentially in sterile HBSS followed by complete medium (DMEM+glutamine+10% FBS+20 μg/ml gentamicin). The brain was stripped of membranes and vessels and minced with microdissecting scissors. The tissue pieces were then gently drawn into a syringe fitted with a 21 ga needle engineered to cut smoothly through the tissue (blunt tip with sharpened edges). The needle was removed and the small cores of tissue that were collected in the syringe were then transferred to a 15 ml tube containing 8–10 ml of medium. The tissue “punches” were washed twice with gentle agitation and then re-suspended in 12 ml fresh medium. The punches were then evenly distributed into 6 well ultra low adhesion plates such that each well had the tissue from ~1–2 brains. The tissue punches formed small organoid-like balls of tissue that were maintained as a suspension culture. Microglia are the only cells in the culture capable of attaching to the low adhesion surface. After about 4–5 days, microglia growing within the punches begin to migrate out and attach to the plate. Over the next week, the numbers of microglia increased until the well was confluent, at approximately 2 weeks (depending on the density of tissue punches). This procedure allows the development of the microglia in a neural environment. The neural tissue (suspended punches) is easily removed by washing leaving a pure microglial culture.

Preparation of macrophage/microglial conditioned medium (MCM)

Cultures of macrophages or microglia were stimulated with 200 pg/ml HIV gp120_ADA virions for a period of 1 hour to generate conditioned medium. The MCM was centrifuged at 1000 x g for 10 minutes to remove any cells, ultrafiltered at 300 kDa to remove virions and then frozen in aliquots at −80° C. Prior to use, the medium was tested to verify the ability to provoke a dysregulation of neuronal intracellular calcium homeostasis and cytoskeletal damage as previously described(Meeker et al., 2005; Meeker et al., 2012; Meeker et al., 2016a). The medium was used in the current studies to provoke inflammation-associated damage and determine if the p75^NTR was expressed during the earliest stages of damage (calcium dysregulation and beading).

Primary cultures of mouse cortex and hippocampus

All culture work was done in accordance with NIH animal welfare guidelines and was approved by the University of North Carolina-Chapel Hill Institutional Animal Care and Use Committee. Timed gestational embryonic day 16 (E16) pregnant female CD1 mice (Charles Rivers) were sacrificed by anesthetizing with isoflurane drop method until breathing and heart stopped. A thoracotomy was then performed, the uterus removed, briefly rinsed in ice cold 70% ethanol and rinsed twice in ice cold, sterile HEPES-buffered Hank’s balanced salt solution (HBSS). The brain was dissected from each fetus, extensively washed and cleaned of dura-arachnoid membrane and visible vessels. The cortex/hippocampus was dissected from each brain and transferred to a 15 ml tube containing 5 ml calcium-magnesium free-HBSS + 2.4 U/ml dispase + 2 U/ml DNase I and incubated for 25–30 min at 36° C. Tissue was triturated and allowed to settle for 2 min. The suspended cells were transferred to a 50 ml culture tube containing 25 ml of Neurobasal medium with added B27 supplement, glutamine, 5% fetal bovine serum and 20 μg/ml gentamicin (Note: more recently we have determined that Neurobasal Plus with B27 Plus provides even better results). After several rounds of trituration in 2–3 ml calcium-magnesium free HBSS, the dissociated cells were seeded at a density of 12,000 – 20,000 cells/cm² on poly-D-lysine-treated coverslips. After 24 hours, cultures were transferred to Neurobasal medium with added B27 supplement and glutamine. The resulting cultures were >95% neurons at day 4 after seeding.

Statistical analysis

Graphpad Prism software was used for data summaries and graphics. Parametric statistics were used to evaluate most changes induced by LM11A-31 relative to untreated controls. In cases where the data were not normally distributed non-parametric statistics were used. Means and the standard error of the mean were calculated for at least three replicate experiments. T-tests were used for paired comparisons, analysis of variance for multiple groups, analysis of variance with repeated measures for temporal data and Chi-square for the analysis of cell populations. A probability of <0.05 for rejection of the null hypothesis was considered significant.

Results

Inflammatory stimuli increase expression of the p75^NTR

The p75^NTR is often upregulated in the nervous system in response to injury or stress (Ibanez and Simi, 2012). To determine if neural damage associated with inflammatory conditions was associated with changes in neuronal p75^NTR expression, we first challenged cultured mouse neurons with macrophage or microglial conditioned medium (MCM) that was prepared from human monocyte-derived macrophages or mouse microglia treated with HIV gp120. MCM from macrophages or microglia produced the same results.

When challenged with MCM, foci of p75^NTR staining appeared in the dendritic varicosities that form in response to calcium overload (Figure 1A, solid arrows). Neurons with the highest expression of p75^NTR were often deficient in MAP-2 staining (Figure 1A, open arrows) suggesting that the expression of p75^NTR preceded the development of more extensive damage. To quantify the relationship between p75^NTR and MAP-2 loss, we compared the expression of p75^NTR to MAP-2 in 789 neurons sampled from cultures treated with MCM. A frequency plot of p75^NTR expression in cells with low (lower half) versus high (upper half) MAP-2 expression (Figure 1B) illustrated a distinct sub-population of neurons expressing high p75^NTR in the low MAP-2 expression group. Since only a small number of cells show severe depletion of MAP-2 expression within the limited time frame of these experiments, which reflects the earliest stages of pathogenesis, the analysis in Figure 1B is likely to be a conservative representation of the p75^NTR/MAP-2 relationship. The anatomical relationship of p75^NTR expression to MAP-2 in the swellings was quantified by mapping the intensity of the stain with a resolution of 0.32 μm along the length of dendrites from 10 representative cells. The location of the MAP-2 immunoreactive varicosities matched closely to the p75^NTR stain across all runs with an average correlation of r=0.961 ± 0.005 (n=20, not shown). On average, p75^NTR expression in the swellings was more than 5-fold greater than the surrounding regions of the dendrite (Figure 1C) verifying the enrichment of p75^NTR within the swellings. In addition to the increased expression of p75^NTR, dendritic and axonal varicosities were also immunoreactive for Tau-1 (Figure 1D) in agreement with previous reports (Tseng et al., 2017). The variant of Tau in the swellings is hypophosphorylated and acetylated as well as immunoreactive for the T22 antibody that recognizes oligomeric species (Tseng et al., 2017). The swellings also accumulated aggregates of F-actin as previously described (Meeker et al., 2016a), mitochondria (Figure 1E) and endoplasmic reticulum (Figure 1F) providing an environment conducive to the production of toxic species. It is notable that our previous studies showed that mitochondria in the swellings identified by staining for TOMM20 are not readily visible using probes based on the mitochondrial membrane potential, such as tetramethylrhodamine methyl ester (Meeker et al., 2016a) indicating that the mitochondria are depolarized. Depolarized mitochondria trapped within the swellings would likely generate reactive oxygen species that enhance damage. Thus, p75^NTR accumulated in the swellings in concert with Tau, actin, mitochondria and endoplasmic reticulum under conditions that could promote pathology. These observations demonstrated for the first time the rapid appearance of p75^NTR in response to an inflammatory stimulus and guided the pathological studies in the gp120 transgenic mouse. Since treatment with LM11A-31 prevents the development of the swellings and preserves MAP-2 expression (Meeker et al., 2016a), we hypothesized that p75^NTR expression may underlie the early development of inflammation-associated pathology and may be a viable therapeutic target in vivo.

Figure 1.

Neuronal varicosities induced by an inflammatory challenge accumulate p75 ^NTR, Tau and organelles. A. Primary mouse neurons treated with microglial conditioned medium developed extensive beading of MAP-2 immunoreactive dendrites (solid arrows), an early index of cytoskeletal damage. Expression of p75^NTR was robust within the varicosities (solid arrows). Neurons expressing high levels of p75^NTR had dendritic pruning and greatly reduced expression of MAP-2 (open arrows). B. Frequency analysis of the distribution of p75^NTR stain intensities in 789 individual neurons from 30 coverslips treated with MCM. A subset of cells with low MAP-2 expression (arrow) showed significantly higher p75^NTR expression (triangles, X², p<0.001). C. Quantification of the intensity of the p75^NTR stain illustrating significant accumulation within varicosities relative to the surrounding dendrite (paired t test, p<0.001). D. Dual stain for MAP-2 (green) and Tau-1 (red) illustrating the accumulation of Tau in axons (small arrows) and dendrites (large arrow) of the same cell. MAP-2 positive varicosities also stained for TOMM20 (mitochondria, E) and calnexin (endoplasmic reticulum, F).

Plaque-like clusters of p75^NTR in the hippocampus.

To determine if expression of p75^NTR is increased in the nervous system of the gp120 Tg mice, we stained sections of gp120 and WT mice for p75^NTR at ages ranging from 4–23 months. Within neurons, p75^NTR was relatively diffuse with a subset of neurons showing increased intensity. In addition to the diffuse neuronal stain for p75^NTR, robust extracellular plaque-like clusters of p75^NTR were seen localized predominantly to the hippocampus. The clusters appeared as numerous small deposits of p75^NTR immunoreactive material within clusters (Figure 2A, arrow). The greatest accumulation was associated with the pyramidal cell dendrites (stratum radiatum) although occasional clusters were seen in stratum oriens and cortex. The area occupied by the clusters was measured using Metamorph analysis software and the average density at ages ranging from 6 to 23 months is illustrated in Figure 2B. Wild type mice showed age-related increases in p75^NTR cluster expression after approximately 12 months (Figure 2B). The gp120 Tg mice showed an earlier and more robust increase in p75^NTR expression by 12 months of age. A sub-analysis of males versus females indicated that the accumulation of p75^NTR was similar for both sexes (not shown). Treatment with LM11A-31 failed to decrease the expression of these clusters.

Figure 2.

Clusters of p75^NTR immunoreactivity appeared in the hippocampus in an age-related fashion in both wild type (WT) and gp120 transgenic mice (gp120). A. Extensive p75^NTR immunoreactive clusters (arrow) in the hippocampus of a 23 month old gp120 Tg mouse. B. The clusters appeared earlier in the gp120 transgenic mice and occupied a greater area at all ages (two way ANOVA, *Age effect: F=6.60, p=0.0006; **genotype effect: F=5.11, p=0.027). The gp120 transgenic mice treated with 50 mg/kg LM11A-31 were not different from untreated mice.

Tau-1 immunoreactive clusters accumulate with age and overlap with p75^NTR

Mice do not normally develop classical AD-like Tau pathology. Indeed, these mice failed to show significant staining with an antibody specific for pathologically phosphorylated Tau (AT8) and stains for amyloid accumulation (thioflavin T). However, when we examined non-phosphorylated Tau epitopes, shown to accumulate in inflammation-induced dendritic swellings in vitro (Tseng et al., 2017), aging mice showed an accumulation of distinct plaque-like clusters mainly localized to the hippocampus (Figure 3A). All mice showed an accumulation of these clusters although the density was variable between mice. Since the plaque-like clusters of Tau-1 immunoreactive material appeared almost identical to the p75^NTR clusters, we compared the localization in double staining experiments. Tau-1 immunoreactive clusters overlapped extensively with the p75^NTR immunoreactive clusters described above. Figure 3B shows double staining for p75^NTR (green) and Tau-1 (red) demonstrating partial overlap (yellow). An analysis of 12 mice with the best double staining showed that on average, 67.4 ± 4.8% of the p75^NTR(+) aggregates overlapped Tau-1 with a range of 47.4 to 91.7%. The data suggested a close but non-essential or perhaps sequential relationship between p75^NTR and Tau in the formation of these deposits. To determine if the deposits contained oligomeric protein, we stained the brains with the oligomer specific T22 antibody (Lasagna-Reeves et al., 2012). As illustrated in Figure 3C, the T22 stain overlapped partially with the Tau-1 stain indicating that many deposits contained an oligomeric, form of a Tau-like protein. Some foci immunoreactive for Tau-1 showed very weak staining for T22 (Fig 3C, arrows) and vice versa (Fig 3C, arrowheads) suggesting that these foci may contain a combination of oligomeric and non-oligomeric Tau. These clusters were identical to those recently described in the hippocampus of normal aged and P301S transgenic mice(Tseng et al., 2017). They also match the description of “age-related accumulation of reelin in amyloid-like deposits” (Knuesel et al., 2009) including a propensity to decrease in a subset of mice at older ages. Thus, these clusters appear to be a normal product of aging that contain several different proteins. The Tau-1 immunoreactive clusters, on average, first appeared after approximately 8 months of age and continued to increase with age (Figure 3D). The age-associated accumulation of Tau-1 immunoreactive clusters was identical in wild type and gp120 Tg mice up to 16 months of age and was unaffected by treatment with LM11A-31 up to 16 months. As the WT mice entered old age (20–23 months) the Tau-1 immunoreactivity continued to increase. In contrast, the gp120 Tg mice showed a 54.2% decrease in Tau-1 immunoreactive clusters at old ages (19–23 months) relative to WT mice. Old gp120 Tg mice treated with LM11A-31 contrasted sharply with the vehicle treated mice showing continued accumulation of the Tau-1 immunoreactive clusters, exceeding vehicle treated gp120 Tg mice by 3.8-fold (Figure 3D). An additional group of 12 month old p75^NTR null mice stained at the same time showed very low immunoreactivity for Tau-1 (Figure 3D, yellow diamond) indicating that p75^NTR expression may be necessary for the development of the Tau clusters. It is notable that substantial variation in the development of Tau-1 clusters was seen across individual mice. The variation increased with age suggesting important individual differences in either the development of the clusters or the decrease seen at older ages. A sub-comparison of male versus female mice failed to show a significant difference in the development of the clusters with age.

Figure 3.

A. Immunostaining of a 16 month old normal mouse brain with antibodies to MAP-2 (green) and Tau-1 (red). The Tau-1 stain appears as clusters of immunoreactive material predominantly associated with the pyramidal cell dendrites. While some overlap is seen with the MAP-2 stain (yellow) no such clusters were seen for MAP-2. B. Combined staining for p75^NTR and Tau-1 in the hippocampus illustrated partial overlap of the two stains (yellow, open arrow). Examples of p75^NTR stain in the absence of Tau (arrow) can be seen. C. Double staining with the T22 antibody to oligomeric protein and Tau-1 showed substantial overlap of the two signals indicating that much of the Tau was present in oligomeric form. T22 stain could occasionally be seen that was negative for Tau-1 (arrowhead) and Tau-1 stain was occasionally negative for T22 (arrows) indicating that both oligomeric and non-oligomeric forms of Tau were present in the same region. D. Quantification of the Tau cluster density in the hippocampus at young (4–8 months) middle (11–16 months) and old ages (20–23 months) showed that the Tau-1 immunoreactive clusters increased progressively with age in wild type mice. The gp120 transgenic mice showed a similar increase up to approximately 16 months but then declined at older ages. Treatment of gp120 Tg mice with LM11A-31 preserved the increase in cluster density in old age relative to vehicle treated mice (p=0.045, gp120 vs gp120+LM11A-31). Mice null for p75 (yellow diamond, n=5, mean ± sem) did not express Tau clusters (p=0.0119 vs WT mice). E. No significant difference between males and females was seen in the development of Tau clusters with age although a slight lag was present in males at 12 months.

LM11A-31 had no effect on GFAP expression

An increase in GFAP expression was evident in the gp120 Tg mice at the earliest ages, most likely due to the expression of gp120 in the astrocytes. GFAP increased with age in the WT mice such that by 12 months of age, there were no differences in expression (Figure 4). Older mice displayed numerous large astrocytes throughout the hippocampus. In some cases, small varicose GFAP+ structures were seen which suggested some degenerating cells (arrow). Treatment with LM11A-31 had no effect on the expression of GFAP.

Figure 4.

No differences in GFAP immunoreactive astrocytes were seen in treated and untreated older mice. A. Quantification of the area of GFAP immunoreactive cells in the hippocampus showed increased expression at young ages but no significant differences between WT, gp120 or gp120+LM11A-31 mice from 12 to 23 months. B. Example of GFAP stained astrocytes in a 23 month old gp120 mouse illustrating large, strongly positive cells. In addition, small stained elements resembling varicosities (arrow) were occasionally seen suggestive of degenerating cells.

LM11A-31 reduced the numbers of activated Iba1+ microglia across age and genotype

The numbers of Iba1 immunoreactive microglia increased progressively in both WT and gp120 Tg mice after approximately 8 months of age. A genotype effect in the appearance of Iba1 immunoreactive microglia was apparent at all ages with the gp120 mice showing the highest numbers of Iba1+ microglia (Figure 5A). Treatment of the gp120 Tg mice with LM11A-31 significantly reduced the expression of Iba1+ cells as illustrated in Figure 5B–C. The level of Iba1+ microglia was slightly below that of age-matched WT mice indicating some protection against age-associated microglial activation. A sub-comparison of the age-related accumulation of Iba 1 immunoreactive microglia in WT males versus females using an automated threshold analysis to showed slightly enhanced expression of Iba 1 microglia in females relative to males (Figure 5E). After treatment with LM11A-31, males and females both returned to levels (Figure 5F) that were at or below the matched WT controls (Figure 5D).

Figure 5.

Changes in the density of Iba1 immunoreactive microglia in the hippocampus in response to age, gp120 expression and treatment with LM11A-31. A. The density of Iba1 stained cells increased progressively with age in approximately a linear fashion for both WT mice (r=0.987, p=0.013) and gp120 mice (r=0.977, p=0.023). The accumulation of Iba1 microglia for gp120 transgenic mice treated with LM11A-31 was significantly lower than untreated mice (*F=9.45, p=0.004). B. Large robustly stained microglia were seen in the aged gp120 mice. C. Matched gp120 mice treated with LM11A-31 showed fewer and less activated microglia. D. Iba1 immunoreactive microglia in young, middle aged and old males and females showed similar age-associated increases (2-way ANOVA, F=5.76, p=0.0068). No significant difference was seen in Iba1+ microglia in males versus females. E. Both male and female gp120 transgenic mice showed greater expression of Iba1+ microglia relative to WT mice (D) and females were consistently higher than males (2 way ANOVA, F=4.077, p=0.054). F. Treatment with LM11A-31 prevented the accumulation of Iba1+ microglia in both females (F=9.56, p=0.006) and males (F=7.30, p=0.015).

Microglia were occasionally seen in the same general area as Tau-1 immunoreactive deposits (Figure 6A). Regions with higher microglial densities were often not associated with Tau clusters and vice versa. To determine if microglia actively phagocytosed Tau, we treated cultured primary WT mouse microglia with normal and mutant (P301L) Myc tagged Tau. Within 1 hour, the microglia accumulated 13.6- (WT) and 60.4- (P301L mutant) times background levels of Tau fibrils indicating a robust phagocytic response. Significantly higher uptake of the aggregation prone P301L Tau indicated that it may be a preferred substrate for phagocytosis by microglia.

Figure 6.

A. Tau-1 (green) and Iba1 (red) immunoreactivity in the hippocampus. B. Phagocytosis of Myc tagged Tau fibrils by mouse microglia in vitro over a period of 1 hour. Both WT Tau and P301L mutant Tau fibrils were phagocytosed by microglia. Significantly greater accumulation was seen for the P301L mutant Tau (*t=3.73, p=0.0004). Close inspection of Iba1 stained microglia in vivo (C) revealed small particles of Tau1 immunoreactive material (D) in 35.3 ± 6.5% of the microglia in CA1-CA3 (n=14). E. Merged image.

Quantification of the overlap of Tau-1 particles with Iba 1 indicated that most Tau-1 (98.3%) was not associated with microglia. However, close examination of individual Iba 1 immunoreactive microglia (Figure 6C) in brains with robust Tau1 clusters indicated that 35.3 ± 6.5% of the microglia in CA1-CA3 contained small particles of Tau (Figure 6D–E) indicating that many microglia actively phagocytose Tau in vivo.

Comparison of the relative rates of p75^NTR, Tau-1 and Iba-1 accumulation.

To allow a qualitative comparison of the relative rates of accumulation for p75^NTR, Tau-1 and Iba-1 immunoreactivity during normal aging, we normalized the expression of each protein as a proportion of the level in aged mice (Figure 7). Basal Iba 1 expression in young mice was arbitrarily set to zero and data was broken down into smaller groups to better illustrate the rates of accumulation. Tau immunoreactive deposits appeared before a significant increase in Iba 1 immunoreactive microglia and was paralleled by the expression of 75^NTR. As the mice entered old age, a decrease in Tau-1 contrasted with large increases in Iba 1 and p75^NTR.

Figure 7.

Relative rate of appearance of Tau-1, p75^NTR and Iba-1 immunoreactivity in the hippocampus of WT mice as they age. Data from the preceding figures were normalized to expression in old mice to facilitate the comparison. Tau-1 and p75^NTR were increased as early as 8 months whereas Iba-1 immunoreactive microglia began to increase from baseline expression at approximately 12 months. Two way ANOVA of the relative changes showed a significant main effect of age (F=19.38, p>0.001) and protein (F=3.316, p=0.0385) with post hoc significant differences between Iba-1 and Tau1 indicating different relative rates of progression.

MAP-2 Expression is restored by LM11A-31

Expression of MAP-2 in neurons often decreases as an early response to degenerative stimuli(Kaul et al., 2007; Maung et al., 2012). WT and gp20 Tg mice both showed a progressive decrease in hippocampal CA1 MAP-2 expression with age (Figure 8A). However, the gp120 Tg mice had less MAP-2 at each age indicating an overall suppression of MAP-2 expression. Less MAP-2 loss was seen in the CA3 region of the hippocampus (Figure 8B). Treatment of the middle aged and old gp120 Tg mice with 50 mg/kg LM11A-31 for three months resulted in a restoration of hippocampal MAP-2 expression to the level of young WT mice in both CA1 and CA3 (Figure 8 A–B).

Figure 8.

Age associated changes in hippocampal MAP-2 expression in WT and gp120 transgenic mice. A. MAP-2 immunostaining in the CA1 region of the hippocampus decreased with age in both the WT and the gp120 transgenic mice. However, the gp120 transgenic mice showed less MAP-2 expression relative to WT mice at all ages (F=20.17, p<0.0001). Treatment with LM11A-31 prevented the loss of MAP-2 in the gp120 transgenic mice (F=74.54, p<0.0001). B. A smaller genotype effect was seen in the CA3 region of the hippocampus although similar protection was seen against the age associated loss of MAP-2 by LM11A-31 (F=14.50, p<0.001).

An example of the MAP-2 stain in the hippocampus of untreated and treated gp120 Tg mice at 20 and 23 months respectively, is illustrated in Figure 9. The gp120 Tg mice (Figure 9A–C) showed a decrease in the apparent density of the dendrites (MAP-2 intensity) in the stratum radiatum under the pyramidal cell layer. Occasional damaged and beaded (Figure 9B, arrow) dendrites were seen in CA1 and CA2 of gp120 Tg mice although the overall structure of the hippocampus was largely intact. In CA3 there was generally less damage and loss of MAP-2. In contrast, the gp120 mice treated with LM11A-31 (Figure 9D–F) had a dense network of stained dendrites throughout the hippocampus that were indistinguishable from WT mice.

Figure 9.

MAP-2 staining in the stratum radiatum of the hippocampus of aged gp120 transgenic mice. A-C. Modest staining of dendrites is seen in gp120 Tg mice. Signs of early damage in the form of beaded processes was occasionally seen (arrow). D-F. Mice treated with LM11A-31 had more robust MAP-2 stain and fewer examples of damage such as beaded processes. G. MAP-2 stained varicosities were often seen associated with the hippocampal dendrites of aged WT and gp120 Tg mice appearing as vesicular-like structures with an unstained core (G) or as small solid varicosities (H). I. Quantification of the varicosities showed that they were seen in older WT mice (F=9.04, p<0.0001). In gp120 transgenic mice, the varicosities first appeared in middle age and expression was partially reduced by treatment with LM11A-31 (F=5.35, p=0.026). Since the varicosities increased exponentially with age values were normalized by log transformation for the comparison.

An additional pathology was seen in the gp120 mice as well as aged WT mice in the form of MAP-2 stained varicosities within the hippocampal dendritic field (Figure 9G–H). The varicosities had either a donut-like appearance with an unstained core (Figure 9G, arrows) or a solid core (Figure 9H, arrows). Quantification of the number of hippocampal varicosities in WT mice, illustrated in Figure 9I, indicated that they only increased in the very old mice. In contrast, the gp120 Tg mice showed increases in the number of varicosities starting at middle ages. Significantly fewer varicosities were seen in the middle aged and old gp120 Tg mice treated with LM11A-31 (Figure 9I).

Loss of parvalbumin immunoreactive GABAergic neurons is reversed by LM11A-31.

Previous studies have shown that p75^NTR regulates the maturation of parvalbumin (PVA) cell connectivity (Baho, et al., 2019) and promotes a selective, albeit indirect, neurotrophin-dependent expansion and maintenance of GABAergic cells in the basal forebrain in development and aging (Lin et al., 2007). To determine if the increase in p75^NTR with age and inflammation was associated with a change in the PVA+ GABAergic cells in the hippocampus, we quantified PVA immunoreactive neurons. The gp120 Tg mice consistently showed lower average expression with age. Treatment of the gp120 Tg mice with LM11A-31 resulted in preservation of the density of PVA+ neurons relative to placebo. To allow an analysis of the effects of sex on the expression of PVA+ cells, we combined the data for middle aged and old mice. Comparison of young male and female mice showed that PVA+ cells in the hippocampus varied by less than 1.3% across groups indicating no sex differences at early ages. The data in Figure 10C, representing middle aged and old mice, illustrate that as the mice age males developed a more pronounced age-related loss of PVA immunoreactive GABAergic neurons. The response to LM11A-31 was similar in both sexes.

Figure 10.

Age and genotype associated changes in the density of parvalbumin (PVA) immunoreactive neurons in the hippocampus of wild type (WT), gp120 transgenic (gp120) and gp120 Tg mice treated with LM11A-31. A. PVA+ neurons (arrows) in the hippocampus of a 19 month old gp120 Tg mouse illustrating the decreased numbers of PVA+ cells. Many cells positive for PVA appear unhealthy (inset) B. Even very old gp120 Tg mice (23 months) treated with LM11A-31 have greater numbers of PVA+ cells with overall healthier profiles (inset). C. Scatterplot of PVA cell density in the hippocampus of young, middle aged and old mice. Both WT and gp120 Tg mice showed a significant age-associated decline in PVA immunoreactive cells (F=10.13, p=0.0129) with gp120 mice consistently lower than WT (F=13.01, p=0.0194). Treatment of middle aged and old gp120 mice with 50 mg/kg LM11A-31 for 3 months preserved the density of PVA immunoreactive neurons in the hippocampus at a level significantly higher than placebo mice (F=32.56, p=0.0003). C. Comparison of males and females showed consistently lower PVA+ cell numbers for males across all groups (F=8.14, p=0.033). Treatment with LM11A-31 increased PVA+ cells in both males and females (F=19.57, p=0.0043).

Discussion

p75^NTR and Tau-1 are early markers of neural damage

Neurons exposed to medium from activated macrophages or microglia, conditions that mimic inflammation, go through a series of pathological changes beginning with a prolonged increase in intracellular calcium that is followed by cytoskeletal damage and focal swelling (beading) of dendrites and axons. Organelles, actin aggregates, p75^NTR,Tau-immunoreactive oligomers and amyloid precursor protein (APP) accumulate within the swellings (Stokin et al., 2005; Yoon et al., 2006; Sanchez-Varo et al., 2012; Meeker et al., 2016a; Tseng et al., 2017). This pattern of beading is observed in vivo in response to many different neurodegenerative conditions including HIV infection and Alzheimer disease and has been shown to be an early feature of neural damage in mouse models of Alzheimer disease(Stokin et al., 2005; Kuchibhotla et al., 2009; Bittner et al., 2010; Dawson et al., 2010). The accumulation of p75^NTR in the varicosities shown in this study indicates that expression of p75^NTR is an early marker of pathological activity at a time when the calcium dysregulation and structural changes are still reversible. Co-localization of 75^NTR with Tau-1 immunoreactivity suggests a relationship between the two proteins. The relationship could simply reflect inflammation associated changes in Tau that disrupt microtubule transport leading to passive accumulation of p75^NTR as well as other proteins and organelles at sites of dysfunction. On the other hand, neurotrophin signaling may help to maintain Tau function through cytoskeletal support or suppression of enzymes that modify Tau. Stressful conditions where p75^NTR is increased may reduce this support and allow more pathological modifications of Tau. The lack of Tau clusters in p75^NTR deficient mice would support the latter possibility.

While it is reasonable to hypothesize that the swellings give rise to the p75^NTR and Tau aggregates seen in vivo, the unique processes that generate the swellings and protein deposits remain poorly understood. Breakdown of the cytoskeleton at specific locations appears to play an important early role via deficits in transport that promote the accumulation of depolarized mitochondria (Stokin et al., 2005; Meeker et al., 2016a). The presence of dysfunctional mitochondria in close proximity to Tau, APP and other proteins provides an environment conducive to pathological interactions. A study by Takeuchi, et al. using microglial conditioned medium (Takeuchi et al., 2005) suggested that swelling was secondary to mitochondrial dysfunction. Under these conditions, the generation of reactive oxygen species and local ER stress would also be expected, setting into motion a cascade of neuropathological events. Thus, prevention of the underlying calcium accumulation and cytoskeletal modifications should facilitate healthy aging and suppress neurodegeneration.

Aging and gp120

Expression of gp120 in the mice enhanced the aging phenotype by promoting earlier and more robust appearance of p75^NTR, Iba-1 immunoreactive microglia and dendritic varicosities as well as early and more robust loss of MAP-2. Young gp120 Tg mice showed enhanced GFAP immunoreactivity, as expected. However, by 12 months, WT mice rose to the level of the gp120 Tg mice due to age-associated increases in GFAP. Tau-1 immunoreactive deposits were present in both WT and gp120 Tg mice and accumulated at the same rate up to about 16 months. Thus, the appearance of Tau-1 immunoreactivity appears to be an early marker of aging independent of gp120 expression. The decrease in Tau-1 immunoreactivity in old gp120 Tg mice (>16 months) suggested that the Tau-1 may be a transient phenotype that appears in response to early perturbations of neural function but then disappears with disease progression and aging.

Similar clusters of protein have been described in varying contexts. The morphology of the clusters is similar to corpora amylacea(Auge et al., 2018; Navarro et al., 2018; Auge et al., 2019) and matches previous descriptions of reelin plaques(Knuesel et al., 2009; Doehner et al., 2012; Notter and Knuesel, 2013). The clusters also have features similar to astrocytic plaques(Yoshida, 2014) which have similar accumulation of Tau deposits in fine distal astrocyte processes (Feany and Dickson, 1995; Kovacs et al., 2011). These deposits have been identified in elderly subjects with or without dementia and are thought to reflect age-dependent Tau aggregation in astrocytes that is highly localized to the hippocampus and independent of neuronal loss and reactive gliosis(Kovacs et al., 2011). Whether the generation of these aggregates is a protective or pathological response to stress is unclear. Reelin deposits, for example, are thought to reflect protective functions(Pujadas et al., 2014) including suppression of pathological Tau phosphorylation. An age-dependent decrease in reelin deposits, similar to the decrease in Tau-1, was interpreted as a loss of protection. This hypothesis is consistent with the non-phosphorylated status of Tau within swellings in vitro(Tseng et al., 2017) indicating that Tau modifications have not progressed to hyperphosphorylated forms.

Effects of LM11A-31 on age and gp120 associated pathology in vivo.

Treatment of aging gp120 transgenic mice with 50 mg/kg LM11A-31 suppressed microglial activation (Iba-1), preserved the neuronal cytoskeleton (MAP-2) and prevented the appearance of varicosities in the hippocampus. Decreases in the formation of hippocampal varicosities and stabilization of MAP-2 in vivo are consistent with in vitro studies showing that the ligand suppresses calcium accumulation and cytoskeletal dysfunction(Meeker et al., 2012; Meeker et al., 2016a). These effects indicate that LM11A-31 preserves normal neuronal structure and function while also reducing inflammation.

Since Iba-1 immunoreactivity increased in parallel with the Tau-1 in aging mice, it is possible that the inflammatory activity is either a reaction to the same processes that generate Tau-1 immunoreactive deposits or a direct response to the Tau oligomers. The expression of p75^NTR not only paralleled the increase in Tau-1 but spatially overlapped, consistent with the idea that both are a response to neural insults and are poised to interact. It is not yet clear why p75^NTR accumulates within swellings in response to inflammation but its upregulation provides a therapeutic target unique to early pathological changes. Studies of p75^NTR function have indicated that it may be either a positive or negative force in recovery from injury and regulation of neurodegeneration depending on the conditions (Ramer and Bisby, 1997; Yeo et al., 1997; Kramer et al., 1999; Ward et al., 2000; Culmsee et al., 2002; van der Zee and Hagg, 2002; Gschwendtner et al., 2003; Murray et al., 2003; Copray et al., 2004; Jiang et al., 2005; Rosch et al., 2005; Scott et al., 2005; Teng et al., 2005; Zagrebelsky et al., 2005; Song et al., 2006; Chu et al., 2007; Lebrun-Julien et al., 2009; Jiang and Jakobsen, 2010; Tan et al., 2010; Ibanez and Simi, 2012; Hu et al., 2013; La Rosa et al., 2013; Longo and Massa, 2013; Matusica et al., 2013; Mysona et al., 2013; Tep et al., 2013; Dokter et al., 2015; Meeker and Williams, 2015; Poser et al., 2015). Although the molecular mechanisms that support survival versus degradation are still under investigation, in vitro studies clearly support the idea that modulation of the p75^NTR can play a major role in cell survival and function. The beneficial functions of LM11A-31 are due in part to enhanced Akt signaling and suppression of proteins kinases that support pathology (Longo and Massa, 2013). The protective actions in the gp120 Tg mice are paralleled by effects seen in mouse models of other neurodegenerative diseases such as Alzheimer (Yang et al., 2008; Knowles et al., 2013; Simmons et al., 2014) and Huntington disease (Simmons et al., 2016), indicating that p75^NTR has a general role in the development of pathology. Regulation of inflammation may be a feature common to each.

The absence of Tau clusters in p75^NTR (−/−) mice suggests that p75^NTR is important for the accumulation of Tau but the relationship between these proteins is not clear. Since Tau-1 immunoreactive clusters decrease at older ages in the gp120 mice, much like reelin in aged mice, it is possible that they reflect a similar protective function. Sustained accumulation of the clusters in gp120 transgenic mice treated with LM11A-31 in parallel with improved MAP-2 stained dendrites and reduced inflammation would be consistent with the idea that the ligand suppresses p75^NTR dependent progression of Tau pathology. Evidence for a deleterious functional relationship between Tau and p75^NTR has been provided by Nguyen, et al.(Nguyen et al., 2014b) in studies examining the effects of LM11A-31 in the AβPP^L/S mouse model of AD. Treatment of the AβPP^L/S mice with LM11A-31 decreased the formation of phosphorylated Tau clusters identified with the AT8 antibody. In addition, LM11A-31 decreased MC-1 immunoreactive clusters in AβPP^L/S mice indicating suppression of pathological Tau folding. No significant change in Thioflavin S stained amyloid deposits was seen indicating that the effects of p75^NTR are relatively specific to Tau. The reduction in Tau pathology was also paralleled by a decrease in CD68 immunoreactive microglia, again suggesting a close relationship between inflammation and Tau pathogenesis. LM11A-31 may afford protection, in part, by encouraging proteolytic cleavage of p75^NTR and/or a decrease in microglial activation. The p75^NTR ectodomain has been reported to have neuroprotective properties in APP/PS1 and P301L transgenic mice(Yao et al., 2015; Shen et al., 2018). Other actions of the ligand, such as inhibition of GSK3β, cdk5 or JNK, may also protect Tau from pathological modifications. Knockout of p75^NTR has been shown to prevent kinase activation and hyperphosphorylation of Tau in mice expressing P301L mutant human Tau (Manucat-Tan et al., 2019). Additional studies support the likelihood that modulation of p75^NTR may have beneficial effects via direct interactions with p75^NTR on mononuclear phagocytes(Williams et al., 2015; Williams et al., 2016). Activation of p75^NTR by its natural ligand proNGF is associated with secretion of neurotoxic factors that disrupt calcium homeostasis and induced beading. Suppression of this interaction by LM11A-31 may promote a more protective, repair prone environment as seen following stimulation with the TrkA ligand, NGF(Williams et al., 2016). More work will be needed to clarify the nature of these deposits and the potential interactions between p75^NTR, inflammation and Tau modifications. Nevertheless, p75^NTR appears to play an important role in the progression of neural damage and modulation by LM11A-31 can afford significant protection.

Protection of GABAergic neurons.

The loss of GABAergic interneurons in aging, Alzheimer disease and other neurodegenerative conditions is thought to contribute to loss of neural function by promoting a hyperexcitable phenotype and impaired network communication(Palop and Mucke, 2016; Hijazi et al., 2019). Parvalbumin (PVA) containing neurons which control the generation of hippocampal theta activity (Bartos et al., 2007; Amilhon et al., 2015; Ognjanovski et al., 2017) are dependent on p75^NTR (Fasulo et al., 2017; Baho et al., 2019) and may be highly vulnerable to HIV proteins(Marks et al., 2016). The ability of LM11A-31 treatment to restore PVA neurons to levels slightly above age matched controls suggests that, in addition to protection, there may be some restoration of the age associated decline in PVA neurons similar to the restoration of aging cholinergic neurons in previous studies(Xie et al., 2019). This preservation of PVA containing interneurons in the mice treated with LM11A-31 would be expected to facilitate retention of normal network activity and plasticity in the hippocampus. The ability to decrease the degeneration of PVA interneurons also has potential relevance to therapeutic approaches for increased seizure incidence in HIV infection (Mateen et al., 2012).

Conclusions

Overall, gp120 expression synergizes in an additive fashion with the effects of aging resulting in earlier and more robust p75^NTR expression, microglial activation, development of varicosities and loss of MAP-2 and PVA staining. The ability of LM11A-31 to prevent many of these effects is consistent with strong neuroprotective and anti-inflammatory actions of the compound. The accumulation of p75^NTR immunoreactive clusters in the hippocampus and the overlap with Tau-1 immunoreactive deposits was an unexpected observation but suggests a close relationship between p75^NTR expression and Tau, solidified by the lack of Tau-1 stained clusters in p75^NTR(−/−) mice. These clusters appear at relatively early ages and Tau-1 clusters begin to decline as the mice reach an age where cholinergic degeneration and decreased cognitive function first begin to appear. If these clusters are the products of cellular efforts to maintain homeostasis, as suggested in some studies, their preservation by LM11A-31 as the mice age may reflect the stabilization of poorly understood neuroprotective processes. Together, these neuroprotective and anti-inflammatory effects of LM11A-31 in the gp120 transgenic mice indicate that the compound has excellent potential to provide a disease modifying intervention that protects against HIV-associated neurodegeneration.

Highlights.

• HIV gp120 and aging synergize to accelerate inflammation and neurodegeneration

• Extracellular Tau and p75 receptor deposits co-accumulate in the aging hippocampus

• Inflammation and neurodegeneration are suppressed by a novel p75 receptor ligand

• Hippocampal inhibitory interneurons are preserved by the p75 receptor ligand

• Modulation of p75 has significant therapeutic potential in people living with HIV