Low-Dose Maraviroc, an Antiretroviral Drug, Attenuates the Infiltration of T Cells into the Central Nervous System and Protects the Nigrostriatum in Hemiparkinsonian Monkeys

This article is featured in In This Issue, p.3337

Parkinson's disease (PD) is a neurodegenerative disorder that is characterized by tremor, bradykinesia, rigidity, and postural instability (1, 2). Pathologically, it is characterized by gliosis and progressive degeneration of the dopamine (DA) neurons associated with the presence of intracytoplasmic inclusions (Lewy bodies) in the substantia nigra (SN) pars compacta (SNpc) (3).

Although the cause of PD is poorly understood, recent studies indicate that PD is regulated by the adaptive arm of the immune system (4–11). According to Brochard et al. (8), both CD8⁺ and CD4⁺ T cells significantly invade the SNpc of PD patients and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)–intoxicated mice. We have also found the infiltration of both CD8⁺ and CD4⁺ T cells into the SN of hemiparkinsonian monkeys (12). Although the contribution of each of these cell types in nigrostriatal degeneration is not yet known, removal of CD4⁺ but not of CD8⁺ T cells in mice greatly reduces MPTP-induced nigrostriatal DA cell death (8). Whereas Th17 cells exacerbate nigrostriatal dopaminergic neurodegeneration, regulatory T cells have been shown to attenuate such neurodegeneration (9).

Mechanisms by which T cells infiltrate into the CNS under neurodegenerating conditions are poorly understood. Recently we have seen marked upregulation of RANTES and eotaxin, chemokines that are involved in the infiltration of T cells and other immune cells, in vivo in the SNpc and the serum of an MPTP-intoxicated monkey (10). In another study, we have also demonstrated the upregulation of RANTES and eotaxin in the SNpc of post mortem PD brains, as compared with age-matched controls (4). In an MPTP mouse model as well, we have found that RANTES and eotaxin are rapidly upregulated in the SN and serum and that functional blocking Abs against RANTES and eotaxin protect mice against MPTP-induced nigrostriatal degeneration (4), suggesting that these chemokines play a role in DA neuron death.

While investigating the function of these chemokines, we found that a common receptor, CCR5, is shared by both RANTES and eotaxin. Although different antagonists of CCR5 are available, maraviroc is the first member of the CCR5 antagonist class of antiretroviral medication that has been approved for the treatment of HIV-1 infection by the Food and Drug Administration for use in combination with other antiretroviral agents. In this study, we demonstrate that oral administration of low-dose maraviroc protects the integrity of the endothelial monolayer, reduces the infiltration of T cells into the SN, attenuates activation of glial cells, lowers α-synuclein (α-syn) pathology, protects dopaminergic neurons, normalizes striatal SNl neurotransmitters, and improves motor functions in hemiparkinsonian monkeys. Therefore, low-dose maraviroc may be of therapeutic benefit for PD patients.

Maraviroc was purchased from Selleck Chemicals (Houston, TX). Mouse anti–tyrosine hydroxylase (TH) Ab (ImmunoStar), mouse anti-human laminin α5 Ab clone number 4B12 (Millipore), rabbit anti–α-syn Ab (clone MJFR1) (Abcam), anti-CD4 Ab (Thermo Fisher Scientific), anti-CD8 Ab (Thermo Fisher Scientific), anti-Iba1 Ab (Abcam), anti-GFAP Ab (Agilent Technologies), and mouse anti-iNOS Ab (BD Biosciences) were purchased from different vendors. Cy2- and Cy5-conjugated secondary Abs were obtained from Jackson ImmunoResearch Laboratories (West Grove, PA).

Female rhesus monkeys (6–8 y old; 5–7 kg) were used in this study. All animals were singly housed with a 12-h light/dark cycle. Purine monkey chow and water were available ad libitum. Diets were supplemented with fruit during the testing sessions. The study was performed in accordance with federal guidelines of proper animal care and with the approval of the Institutional Animal Care and Use Committee. Monkeys were intoxicated with MPTP according to protocol described previously (12–16). Briefly, animals were tranquilized with ketamine (10 mg/kg, i.m.) and then maintained on an anesthetic plane with isoflurane (1–2%). The animals were put in the supine position. For each injection, a right-sided incision was made along the medial edge of the sternocleidomastoid muscle. The carotid sheath was opened and the common carotid artery, internal jugular vein, and vagus nerves were identified. The common carotid artery was exposed below the carotid bifurcation. The external carotid artery was then ligated. A 27-gauge butterfly needle was inserted into the common carotid artery in a direction retrograde to blood flow; for each injection, 20 ml of saline containing 3 mg of MPTP–HCl (Sigma-Aldrich) was infused at a rate of 1.33 ml/min (15 min). After the infusion was completed, 3 ml of saline was delivered and then the incision was closed.

Monkeys displaying hemiparkinsonian symptoms 7 d after the first intracarotid injection were included in the study. Therefore, hemiparkinsonian monkeys were randomized and one group of hemiparkinsonian monkeys (n = 4) was treated with maraviroc (1.0 mg/kg body wt/d) orally for 30 d through banana starting from 7 d after MPTP injection (Fig. 1). The control group of hemiparkinsonian monkeys (n = 4) received only banana as vehicle.

At the end point, monkeys were anesthetized with pentobarbital (25 mg/kg i.v.) and killed via perfusion with 0.9% saline. The brain was removed, immersed in ice-cold saline for 10 min, and slabbed on a monkey brain slicer (12, 13, 15, 16). Slabs through the head of the caudate and putamen were punched bilaterally with a 1-mm brain punch. These punches were processed for HPLC analysis (for quantification of neurotransmitters). Tissue slabs were immersed in Zamboni fixative followed by immersion in 30% sucrose in PBS. Slabs were sectioned frozen (40 μm) on a sliding knife microtome. Tissue sections were stored in a cryoprotectant solution at 4°C before use. For TH staining, midbrain sections were immunostained with mouse monoclonal anti-TH Ab (ImmunoStar) and visualized by using diaminobenzidine (DAB)/H₂O₂ (DAB Kit; Vector Laboratories) as described previously (13, 15, 17, 18).

Estimates of dopaminergic nigral cell number were performed bilaterally by an investigator blind to the treatment groups using an unbiased design-based counting method (Optical Fractionator, Stereo Investigator; MicroBrightField, Williston, VT). All counts were performed by a single investigator blinded to the experimental conditions. Using a random start, we outlined the SN under low magnification (1.25× objective) and sampled 20% of the treated nigra or 5% of the intact nigra in a random but systematic manner. Quantitation of striatal TH immunostaining was performed as previously described (15, 17, 18). OD measurements were obtained by digital image analysis (Scion, Frederick, MD).

For immunofluorescence staining, anti-Iba1 Ab (1:500), anti-iNOS Ab (1:1000), anti-human laminin α5 Ab (1:1000), and anti-GFAP Ab (1:1000) were used. The samples were mounted and observed under an Olympus IX81 fluorescence microscope as described before (17–19).

Samples were analyzed for neurotransmitters by an investigator blind to the treatment groups. On the day of the analysis, striatal tissues were sonicated in 0.2 M perchloric acid containing isoproterenol (internal standard), and the resulting homogenates were centrifuged at 20,000 × g for 15 min at 4C. After pH adjustment and filtration, 10 μl of supernatant was injected onto a Eicompak SC-3ODS column (HPLC-ECD System, Eicom HTEC-500; JM Science, Grand Island, NY) and levels of DA, 3,4-dihydroxyphenylacetic acid (DOPAC), and homovanillic acid (HVA) were analyzed as described earlier (13, 17, 18, 20).

The animals were trained to a consistent level of performance before intracarotid injection of MPTP. Monkeys were rated three times a week by an investigator blind to the treatment groups using a parkinsonian rating scale used to quantify the clinical status of the monkeys (13–16). The scale included ratings of 10 parkinsonian features (tremor, posture, locomotion, hypokinesia, bradykinesia, balance, fine and gross motor skills, startle response, and freezing). Each test consisted of 12 total trials alternating between arms, six per side. Delaying feeding time until after the task was completed ensured the animal compliance with the test. Data collected included the time taken for the animal to move its hand into the chamber where the fruit was located (reaction time), the time taken to pick up the fruit while the hand was in the chamber (reception time), and the total time taken to move the hand into the chamber, retrieve the fruit, and bring the hand back out of the panel and into the cage (total time).

All values are expressed as means ± SEM. Differences among means were analyzed by one-way ANOVA or Kruskal–Wallis tests (comparison among all groups) and post hoc pairwise comparison. In other cases, two sample t tests were also used to compare control versus MPTP and MPTP versus MPTP plus maraviroc. For behavioral analysis, one-way repeated measure ANOVA models were used. Pairwise comparisons were conducted with Bonferroni adjustment.

Because, in contrast to AIDS, virus is not directly involved in the pathogenesis of PD and the degree of T cell infiltration is also less in the CNS of PD patients as compared with HIV-associated neurocognitive disorders, we used a low dose of maraviroc for treating hemiparkinsonian monkeys. Moreover, to make our primate work relevant to the clinical scenario, monkeys with established hemiparkinsonian symptoms were treated with maraviroc orally (Fig. 1). At first, we examined whether T cells egressed from capillaries and/or also passed the perivascular glia limitans in the CNS of hemiparkinsonian monkeys. Because laminin α5 serves as a marker of endothelial basement membrane and astroglial endfeet could be monitored by GFAP staining, to stain pial blood vessels, nigral sections were also double labeled for laminin α5 and GFAP. Fig. 2A shows the integrity of glia limitans in control SN that was lost after MPTP intoxication. However, after maraviroc treatment, the integrity of glia limitans was restored partially (Fig. 2A, 2B). Next, we monitored infiltration of T cells. Double labeling of laminin α5 and CD3 shows that T cells were almost absent around the intact endothelial layer of control SN (Fig. 2C, 2E). However, a significant increase in CD3⁺ T cells was seen around the ruptured endothelial layer in the SN of MPTP-treated monkey brain (Fig. 2C–E). In contrast, after maraviroc treatment, we found significant reduction in CD3⁺ T cells around the endothelial layer (Fig. 2C–E).

Next, we tried to characterize these T cells by immunostaining of nigral sections. CD4 immunostaining (Fig. 3A, 3B, DAB staining; Fig. 3C, TH & CD4 double labeling) clearly shows a typical CD4-immunoreactive inflammatory cuffing in the SN of hemiparkinsonian but not control monkeys. Similarly, we also observed infiltration of CD8⁺ T cells into the SN of hemiparkinsonian but not control monkeys (Fig. 3F, 3G). Although CD4⁺ T cells were found to be more abundant than CD8⁺ T cells near the blood vessel (Fig. 3D, 3H), more CD8⁺ T cells were detected in the deep parenchyma far from the lumen of blood vessels (Fig. 3E, 3I) in the SN of MPTP-treated monkey brain. However, oral treatment with maraviroc significantly attenuated the infiltration of both CD4⁺ and CD8⁺ T cells near the blood vessel as well as in deep parenchyma (Fig. 3). These results suggest maraviroc treatment is capable of suppressing the infiltration of T cells.

It is becoming clear that neuroinflammation driven by glial cells (astrocytes and microglia) plays an important role in the loss of dopaminergic neurons in PD and its animal model (17, 18, 21–25). As reported earlier (13), MPTP intoxication led to marked increase in nigral iNOS protein expression in hemiparkinsonian monkeys. Consistent with the involvement of glial cells, this iNOS was mainly expressed by either Iba1-positive microglia (Fig. 4A) or GFAP-positive astrocytes (Fig. 4B). However, oral maraviroc treatment reduced the expression of iNOS protein in the SN of hemiparkinsonian monkeys (Fig. 4A, 4B, 4E). Similar to the increase in iNOS, we also observed an increase in microglia and astrocytes in the SN of hemiparkinsonian monkeys (Fig. 4A–D). Interestingly, maraviroc treatment did not significantly reduce the number of astrocytes and microglia in the SN of hemiparkinsonian monkeys (Fig. 4A–D).

Recent studies demonstrate alteration in mitochondrial homeostasis in the CNS of patients with different neurodegenerative diseases, including PD (26–30). We have also described decreased mitochondrial biogenesis in the SN of MPTP mouse model of PD (31). Therefore, in this study, we examined the effect of maraviroc on mitochondrial biogenesis in the SN of hemiparkinsonian monkeys. Whereas PPARγ coactivator 1α (PGC1α) is considered as the master regulator of mitochondrial biogenesis (32), mitochondrial transcription factor A (TFAM) is known to control the transcription of mitochondrial genes (33). Mitochondrial biogenesis was monitored in nigral sections by double labeling of TH & PGC1α (Fig. 5A, 5C) and TH & TFAM (Fig. 5B, 5D). In control SN, both PGC1α and TFAM were present in TH-positive neurons as well as TH-negative cells (Fig. 5A, 5B). MPTP challenge resulted in a marked decrease in mitochondrial content in the SN as compared with control, which is evident from decrease in PGC1α (Fig. 5A, 5C) and TFAM (Fig. 5B, 5D). However, low-dose maraviroc treatment led to significant protection of both PGC1α (Fig. 5A, 5C) and TFAM (Fig. 5B, 5D) in the SN of hemiparkinsonian monkeys.

Neuropathological hallmarks of PD are the presence of intracytoplasmic inclusions containing α-syn and the demise of dopaminergic neurons in the SNpc. In addition to PD, accumulation of α-syn is also an important pathological hallmark of dementia with Lewy bodies and multiple system atrophy (34–36). Many in vitro and in vivo studies have shown that posttranslational modifications of α-syn and associated protein aggregation are intimately coupled with neurotoxicity. Therefore, decreasing α-synucleinopathy may have therapeutic importance in PD and other Lewy body diseases. Accordingly, we observed widespread α-syn pathology in the SN of monkeys after 37 d of MPTP intoxication (Fig. 6A–C). However, α-syn pathology was almost missing in the SN of hemiparkinsonian monkeys after oral maraviroc treatment (Fig. 6A–C).

Because low-dose maraviroc treatment suppressed T cell infiltration into the SN and inhibited glial activation as well as α-synucleinopathy in the SN of MPTP-intoxicated mice, we examined the effect of maraviroc on nigrostriatal pathology. Therefore, after 30 d of treatment, animals were processed for quantification of dopaminergic cell bodies in the SNpc and of projecting dopaminergic fibers in the striatum using TH immunostaining. As expected, MPTP intoxication led to marked loss of nigral TH-positive neurons on day 37 of MPTP intoxication (Fig. 7A, 7B) compared with the intact side (control). Stereological counting of TH-immunoreactive neurons in nigral sections showed that MPTP intoxication led to almost 80% loss of SNpc TH-positive neurons (Fig. 7D) in the lesion side compared with the control intact side. Similarly, quantitative measurements of OD of the TH staining in caudate nucleus and putamen exhibited marked loss of striatal TH fibers (Fig. 7C, 7E) in the lesion side compared with the intact side. However, oral maraviroc treatment significantly restored SNpc TH-positive neurons (Fig. 7A, 7B, 7D) and striatal TH fibers (Fig. 7C, 7E) in hemiparkinsonian monkeys.

Because maraviroc protected dopaminergic neurons and fibers, we next determined if maraviroc also protected against biochemical deficits caused in hemiparkinsonian monkeys. Therefore, we quantified levels of DA, DOPAC, and HVA in different parts of striatum. MPTP intoxication led to marked loss of DA in the dorsomedial (Fig. 8A) and ventromedial caudate (Fig. 8B), the dorsal (Fig. 8C) and ventral putamen (Fig. 8D), and the nucleus accumbens (Fig. 8E) of the lesion side compared with the intact control side. In contrast, hemiparkinsonian monkeys that were treated with low-dose maraviroc exhibited pronounced preservation of DA in different parts of striatum (Fig. 8A–E). Similarly, maraviroc treatment also restored the levels of DOPAC and HVA in different parts of striatum of hemiparkinsonian monkeys (Fig. 9A–E).

Although maraviroc treatment displayed beneficial effects in hemiparkinsonian monkeys, it was important to know if this antiretroviral drug might evoke adverse effects on dopaminergic neurons in normal monkeys. However, in our unilateral model, we did not see any significant change in nigral dopaminergic neurons, striatal TH fiber density, and striatal neurotransmitters in the intact side even after 30 d of maraviroc treatment, suggesting that maraviroc does not have a detrimental effect on the normal nigrostriatum.

The eventual objective of any neuroprotective therapy is to decrease functional impairments. Therefore, monkeys were examined performance on a hand-reaching task, general activity, and clinical dysfunction (tremor, posture, locomotion, bradykinesia, balance, fine and gross motor skills, startle response, and freezing), based on a clinical rating scale modeled after the Unified PD Rating Scale. As reported earlier (13–16), MPTP injection caused a marked decrease in performance in the left hand (contralateral to the MPTP infusion) while maintaining relatively normal function with the right hand (Table I). However, oral maraviroc treatment significantly improved functional ability, as observed on both a subjective clinical rating scale and an objective operant motor test (Table I). Importantly, during maraviroc treatment, we did not observe any drug-related side effect (e.g., hair loss, untoward infection, hyperkinesia, psychological disturbance, vomiting, diarrhea, etc.) in any of these monkeys.

At present, no effective therapy is available to halt the progression of PD. Carbidopa/Levodopa (Sinemet) and/or a DA agonist has been the standard treatment for PD. However, it is often associated with a number of side effects and unsatisfactory outcomes. For example, after 2 years of therapy with Carbidopa/Levodopa (Sinemet), 30–50% of patients begin to experience dopaminergic-related complications, including dyskinesis, wearing off, or on–off motor fluctuations. These motor complications eventually influence the clinical picture. Therefore, developing an effective therapeutic approach to halt the progression of PD is of paramount importance. Numerous studies have been carried out in rodents on the role of innate and adaptive immune responses in PD (4, 37–42). However, there is a wide immunological gap between inbred rodent strains and heterogeneous human population. As a result, sometimes rodent results are not translated to humans. In contrast, rhesus monkeys are old world monkeys, and both rhesus monkeys and humans are thought to have diverged from a common ancestor, a primitive anthropoid,∼25 million years ago. Interestingly, despite 25 million years of evolutionary separation, humans and macaques apparently share∼93% of their DNA sequence (43). It is also important to note that the immune systems of rhesus monkeys and humans are very similar, suggesting that, in situations in which it is desirable but impossible to study human neuroimmune responses in vivo, the rhesus monkey provides the best available alternative.

Maraviroc (Selzentry) is a Food and Drug Administration–approved drug for treating HIV infection. This drug is also currently being explored for treating different cancers (44, 45). Recent studies have shown that maraviroc suppresses the metastasis of breast cancer (46) in different models. One clinical trial is also underway to study the efficacy of maraviroc in metastatic colorectal cancer. Several lines of evidence presented in this manuscript clearly demonstrate that oral maraviroc treatment results in protection of hemiparkinsonian monkeys, both clinically and pathologically. Our conclusions are based on the following: first, only the monkeys that exhibited parkinsonian posture, tremor, and left hand freezing after intracarotid injection of MPTP were selected for maraviroc treatment. Second, chronic inflammation is a hallmark of the PD pathologic condition. Similarly, even after 37 days of a single MPTP injection, levels of GFAP, Iba1, and iNOS remained markedly higher in the lesioned SN than nonlesioned SN. However, oral maraviroc markedly reduced nigral expression of iNOS. This result was specific, as we did not notice any significant inhibition in the number of GFAP-positive astroglia and Iba1-positive microglia after maraviroc treatment. Third, mitochondrial dysfunction is known to contribute to nigrostriatal pathologic condition in PD patients and in animal models of PD (27, 28). We also observed a decrease in mitochondrial biogenesis in the SN of hemiparkinsonian monkeys. However, marked restoration of mitochondrial biogenesis was observed in the SN after maraviroc treatment. Fourth, α-syn accumulation is one of the hallmarks of both sporadic and familial PD (34–36), which is usually observed in hemiparkinsonian monkeys. However, maraviroc treatment markedly decreased α-syn pathology in the SN of hemiparkinsonian monkeys. Fifth, as observed in PD, nigral dopaminergic neurons disappeared and the level of DA decreased in hemiparkinsonian monkeys, but treatment with maraviroc protected TH-positive dopaminergic neurons and restored the level of DA. Sixth, oral maraviroc also reversed motor deficits in a hand-reach task. Seventh, daily Food and Drug Administration–approved doses of maraviroc for adult HIV patients are 300 and 600 mg twice a day per patient. It is prescribed at a dosage of 150 mg/d for baby HIV patients. However, in hemiparkinsonian monkeys, maraviroc at a dosage of 1 mg/kg body wt/d markedly protected the nigrostriatum and improved locomotor activities at a much lower dose. Interestingly, a dosage of 1 mg/kg body wt/d is equivalent to 70–80 mg daily per patient that is lower than even the baby dose of maraviroc. These results suggest that maraviroc may be effective in PD patients at very low doses. Moreover, at regular doses (300 and 600 mg twice a day per patient), maraviroc exhibits side effects related to some kind of infection (e.g., upper respiratory infections, herpes infection, etc.) in <10% HIV patients. Although maraviroc itself does not cause hyperkinesia in HIV patients, concomitant use of maraviroc and ritonavir may lead to hyperkinesia. However, at very low dosages (70–80 mg daily per patient), we should not expect such side effects from maraviroc.

The mode of action of maraviroc is becoming clear. Recently, our laboratory has demonstrated that RANTES and eotaxin, chemokines that are involved in the infiltration of T cells and other immune cells, are upregulated in the SNpc and the serum of MPTP-intoxicated monkey (10). These two chemokines are also upregulated in vivo in the SNpc of post mortem PD brains as compared with age-matched controls (4). According to Tang et al. (47), serum RANTES levels strongly correlated with Hoehn–Yahr score and disease duration in PD patients. Interestingly, both RANTES and eotaxin share the same receptor, CCR5, which is inhibited by maraviroc. CCR5-mediated T cell infiltration into the target organ occurs during many pathological conditions, including HIV infection and cancer (48, 49). By antagonizing CCR5, maraviroc inhibits the infiltration of T cells in HIV infection and cancer. T cells enter into brain parenchyma through a bilayer of endothelial basement membrane and glial limitans made of astrocytic end feet. Accordingly, we found loss of glia limitans and infiltration of both CD8⁺ and CD4⁺ T cells into ventral midbrain of hemiparkinsonian monkeys. Although we did not characterize circulating T cell subsets in the blood on hemiparkinsonian monkeys before and after maraviroc treatment, CD3 immunostaining showed a decrease in overall T cell infiltration into the SN by maraviroc. There are several direct and indirect pathways by which T cell infiltration could influence dopaminergic neurodegeneration. For example, it has been reported that the migration of Ag-specific CD4⁺ T cells from the periphery to the CNS generates immunocyte-microglial activities that perpetuate neuroinflammation and affect neuronal survival (50). Earlier we have shown that effector T cells are capable of activating microglia for the production of various proinflammatory molecules via cell-to-cell contact (51, 52). This contact process involves VLA4 (α4β1) integrin of T cells and VCAM1 of microglia (52, 53). Furthermore, activated T cells may also activate microglia via CD40–CD40 ligation (7, 9, 11). According to Nitsch et al. (10), cytotoxic T cell–mediated lethal increase in neuronal calcium could be prevented by blocking both perforin and glutamate receptors. Interestingly, oral administration of maraviroc restored the integrity of endothelial monolayer and strikingly suppressed the infiltration of CD8⁺ and CD4⁺ T cells into the SN. Adaptive immune response can be driven by a number of mechanisms (54, 55). In this study, the beauty of our finding is that antagonism of a particular component of the adaptive immune arm (CCR5) is sufficient to reinstate glia limitans, protect the loss of dopaminergic neurons, improve striatal DA, and reverse motor deficits in hemiparkinsonian monkeys.

The authors have no financial conflicts of interest.