Age-related decrease in tyrosine hydroxylase immunoreactivity in the substantia nigra and region-specific changes in microglia morphology in HIV-1 Tg rats

Abstract

Animal models have been used to study cellular processes related to human immunodeficiency virus-1 (HIV)-associated neurocognitive disorders (HAND). The HIV-1 transgenic (Tg) rat expresses HIV viral genes except the gag-pol replication genes and exhibits neuropathological features similar to HIV patients receiving combined antiretroviral therapy (cART). Using this rat, alterations in dopaminergic function have been demonstrated however, the data for neuroinflammation and glial reactivity is conflicting. Differences in behavior, tyrosine hydroxylase (TH) immunoreactivity, neuroinflammation, and glia reactivity were assessed in HIV-1 Tg male rats. At 6 and 12 weeks-of-age, rota-rod performance was diminished, motor activity was not altered, and active avoidance latency performance and memory were diminished in HIV-1 Tg rats. TH+ immunoreactivity in the substantia nigra (SN) was decreased at 8 months but not at 2–5 months. At 5 months, astrocyte and microglia morphology was not altered in the cortex, hippocampus, or SN. In the striatum, astrocytes were unaltered, microglia displayed slightly thickened proximal processes, mRNA levels for Iba1 and Cd11b were elevated, and interleukin (Il)1αCxcr3, and cell adhesion molecule, Icam decreased. In the hippocampus, mRNA levels for Tnfa and Cd11b were slightly elevated. No changes were observed in the cortex or SN. The data support an age-related effect of HIV proteins upon the nigral-striatal dopaminergic system and suggest an early response of microglia in the terminal synaptic region with little evidence of an associated neuroinflammatory response across brain regions.

Introduction

Despite the introduction of combined antiretroviral therapy (cART), human immunodeficiency virus (HIV)-associated neurocognitive disorders (HAND) is present in the patient population (Letendre et al., 2010; McArthur et al., 2010; Heaton et al., 2010; Schouten et al., 2011). Compared to pre-cART, the differences reflect a shift from slowed motor function information processing to milder stages with effects on learning, memory, and executive function (Heaton et al., 2010; Schouten et al., 2011; Saylor et al., 2016; Sacktor and Robertson, 2014) and a concern for adverse effects on brain structure and function remains (Gelman, 2015; Chan et al., 2016; Boban et al., 2017; Clifford et al., 2017; Underwood et al., 2017; van den Dries et al., 2017; Sanford et al., 2018). In the pre- or naïve cART HIV patient, CNS-associated disorders have been linked to changes in the dopaminergic system including, dopamine (DA) metabolism (Berger et al., 1994), neuropathology in dopamine-rich brain regions (Kieburtz et al, 1991; Reyes et al., 1991; Lopez et al, 1999; Itoh et al., 2000; Gelman et al., 2006; Hu et al., 2009), reductions in substantia nigra (SN) tyrosine hydroxylase (TH), the enzyme responsible for catalyzing the conversion of the amino acid tyrosine to the dopamine precursor, L-DOPA (Silvers et al., 2006), and decreased dopamine transporter (DAT) levels in the putamen and ventral striatum (Wang et al., 2004). cART treatment in HIV patients appears to be insufficient in fully preventing such disruptions as reports of dopaminergic system dysfunction in this patient population continue (Kumar et al., 2009; Gaskill et al., 2017). The presence of viral proteins and possible exacerbation of dopaminergic system dysfunction has been suggested to contribute to neurocognitive disorders (Sheppard et al., 2015) and an association with Parkinsonism (DeVaughn et al., 2015). In addition, a role for neuroinflammation identified in pre-cART HAND appears to continue in cART-era HAND (Gannon et al. 2011; Zayyad and Spudich, 2015; Levine et al., 2016; Ginsberg et al., 2018) yet, the association with cognitive deficits is not as well defined (Chen et al., 2014; Rao et al., 2014; Hong and Banks, 2015). The observation that dopamine regulates a number of myeloid functions and that changes in DA concentration could influence myeloid function via DA receptor activation (Gaskill et al., 2012; Nolan et al., 2019) suggests a link between dopaminergic tone and neuroinflammation (Gaskill et al., 2013; 2014; Nolan et al., 2019). It is proposed that such chronic neuroinflammation may be tightly related to HAND co-morbidities (Saylor et al., 2016, Vera et al., 2016). Information on the temporal and spatial association between neuroinflammatory factors and the dopaminergic system will provide a better understanding of their role as contributing factors in the continue prevalence of HAND in the cART patient population.

While none of the available rodent models adequately replicate HIV patients under cART, they have been used to examine the potential association between HIV-related proteins and adverse nervous system outcomes such as, neuroinflammation, dopamine function, and learning and memory. Few have been used to examine the contribution of antiretroviral treatment which would more accurately reflect the cART patient population but may still have value as representative of a subgroup of the patient population. One model, the HIV-1 transgenic (Tg) rat, expresses all of the HIV viral genes, except the gag-pol replication genes, and proteins from birth, in the absence of actively replicating virus (Reid et al. 2001). This rat has been utilized to explore underlying biological processes associated with cART HAND (Reid et al., 2001; Vigorito et al, 2007; Lashomb et al., 2009; Lassiter et al., 2009; Peng et al., 2010; Royal et al., 2012; Roscoe et al., 2014) with the caveat that viral protein expression levels do not adequately reflect those in a cART patient. From the rodent studies, a number of processes have been implicated as important correlates including, synaptic simplification and injury (Atluri et al., 2013; Ru and Tang, 2017), mitochondrial alterations (Rozzi et al., 2017), axonal transport (Avdoshina et al., 2016), and neuroinflammation (Repunte-Canonigo et al., 2014; Nesil et al., 2015; Sanna et al., 2017). The presence of HIV viral proteins alone damages DA neurons (Nosheny et al., 2006; Theodore et al., 2012; Miller et al., 2018), alters DA transporter (Webb et al. 2010; McIntosh et al., 2015; Bertrand et al., 2018), mediates striatal dopaminergic synapses (Sinharay et al., 2017), facilitates synaptic loss (Kim et al., 2011; Shin et al., 2012), and alters dendritic spines of medium spiny neurons and DA uptake in the striatum (Javadi-Paydga et al., 2017). Diminished immunostaining in the striatum for TH suggested an enzymatic deficit in dopaminergic projections from the SN (Reid et al., 2016a). Such changes in dopamine function have been reflected in behavioral alterations in HIV-1 Tg rats (Liu et al., 2009; Moran et al., 2013; Zhu et al., 2016; Javadi-Paydar et al., 2017). While susceptibility of the dopaminergic system continues to gain support, questions remain as to the mechanism by which this occurs given that neurons are not infected by the virus. An alternative mechanism proposed is that infected microglia are a source of viral proteins or inflammatory factors that, upon release, can damage neurons. While neuroinflammation has been implicated in HIV-1 Tg rats, the findings are inconsistent (Table 1), making it difficult to identify a relationship with behavioral or dopaminergic alterations.

Table 1.

Summary of publications on neuroinflammation, glial response, behavior, in HIV-1 Tg rat brain regions.

Author	Sex	Age	mRNA		Proteins		Histochemistry		Imaging		Behavior
Avdoshina et al., 2018	M	5 mo	CTX (n=4)		CTX (n=4)		CTX (n=4)
F344/NHsd			Il1b	nc	Ac-tubulin	dec	iba1	nc
			Tnfa	nc
Blanchard et al 2015	M	9 mo			HEMIS (n=10-;12)
F344/NHsd					PGE2	in
					8-isoprostane	in
					15-Epi-LXA4	in
					TXB2	nc
					15-HETE	nc
					LTB4	nc
					LXA4	nc
Chivero et al., 2017	M/F	18 mo	STR (n=4)		STR (n=4)
			Il1b	in	IL-1b (17kDa)	in
			Il6	in	ASC	in
					caspase 1	in
Cho et al., 2017	ns	5 mo			HIPP(n=4)		HIPP(n=4)
F344/NHsd					AC3	in	H&E	cell death
					Bax	in	Cresyl violet	cell death
					NeuN	dec	AC3	in
					GFAP	in	GFAP	in
					C99	in	NeuN	in
					b-amyloid	in	Congo red	in
					p-Thr181	in	beta-Amyloid	in
					p-Thr231	in	CTX (n=4)
					p-Ser396	in	NeuN	dc
					TNFa	in	GFAP	dc
					MCP-1	in
Gou et al., 2012	ns	3–6 mo					CA1 Hipp
F344/NHsd							NeuN	nc
							Parvalbumin	nc
Lee et al 2015	M	3–16 mo			brain slice		Iba1	#	[18F]DPA-714 PET
F344/NHsd					ctx, hipp, st		1–10 mo (n=14,12)		3 mo (n=4; 4)	nc
					24 plex assay		Ctx	nc	9 mo (n=5; 5)	nc
					3 mo (n=5; 5)	nc	hipp	nc	16 mo (n=3; 6)	nc
					9 mo (n=4; 5)	nc	striatum	nc
Nemeth et al., 2014	F	48 days	HIPP (n=10,9)				HIPP(n=9,12)				Motor Function	nc
F344/NHsd			Ccl2	in			Ki-67	dec			Open field (10 min)	nc
			Tnf	nc							Startle	nc
			Il1b	nc							Sucrose preference	nc
			Nfxbia	nc							Forced swim	in
											Social behavior	dec
Pang and Panee, 2016	ns	10 mo	Hipp (n=5)		Hipp (n=5)
HIV-1 NL4–3			Gfap	nc	GFAP	in
gag/pol Tg			Iba1	nc	Iba1	dec
			Tnfa	nc	IL-1b	in
			Il1b	in	NFkB p65	nc
					c-Jun	in
Reid et al., 2016									18F-FDG Uptake		Rota Rod (n=4; 5)
F344/NHsd									4–31 weeks-old (n=4)	nc	4–7 wk	dec
									12 wk (n=5)	nc	8–20 wk	nc
									29 wk (n=4;5)	nc	Open-field (n=4–5)
									14C-DG		11–25 wk	dec
									7–8 mo (n=5)	nc
Reid et al., 2016	M	1–20 mo	1 mo (n=3)				1&3 mo (n=6; 8)
F344/NHsd			8–9 mo (n=5)				7&9 mo (n=7; 9)	NeuN/GFAP
			STR/HIPP				STR	dec/dec
			Gfap	nc			HIPP	nc/dec
			Iba1	nc			CTX	nc/dec
			Cd11b	nc				Iba-1#
							str/hipp/ctx	nc
Repunte-Canonigo et al., 2014	ns	4–5 mo	HIPP(n=6;5)				HIPP, CTX (n=5)	Intensity/#			T-maze (n=7)
SD			Gfap	in			GFAP	In/in			Spontaneous alternation	dec
			Iba1	in			Iba-1	in/in			latency	nc
			Cd11b	in
			INF signaling	in
			cell division	in
Rowson et al 2016	M/F	54 days	HIPP				HIPP (n=3)				Open-Field	LC/DC
F344/NHsd			Cfb	in			Iba-1				M(n=8;11)	dec/nc
			Lcn2	nc			# branches	in			F (n=11; 10)	dec/nc
			PFC				# junctions	in
			Cfb	dec			max branch length	in
			Lcn2	dec			avg branch length	nc			Novel Object
							RECA-1	nc			M (n=8;9)	dec
							PFC (n=3;5)	nc			F (n=9;7)	dec
							Amygdala (n=5;6)	nc				dec
							Hipp (n=5;6)	nc
Royal et al., 2012	ns	ns			HEMIS (n=2)		Brain (n=2)	area/intensity
F344/NHsd					IFNg	in	GFAP	nc/nc
					TNFa	in	MHC ii	in/in
					IL-1b	in	ED-1	in/in
							Iba-1	in/in
Yang et al. 2016	M	3 mo	mRNA		mRNA		mRNA
F344/NHsd			Nac		PFC		VTA
			Casp3	in	Casp1	dec	Casp3	in
			Cx3c11	in	Ccl5	dec	Cx3cr1	in
			LRF4	in	Tgfb	dec	Il1b	in
			Irf7	in	TLR4	dec	Il6	dec
			Ccl5	dec
			Cx3cr1	dec
			Il1a	dec
			Tgfb1	dec
			Tlr4	dec

Abbreviations: M-male; F-female; CTX-cortex; HEMIS-brain hemisphere; HIP-hippocampus; NAc-nucleus accumbens; PFC-prefrontal cortex; STR-striatum; VTA-ventral tegmental area

Overall, studies report an absence of neuronal death in HIV-1 Tg rats. However, one recent study reported hippocampal neuronal loss and an associated glia response to injury (Cho et al., 2017). With the exception of the Cho et al. (2017) study, only subtle changes in microglia and astrocytes have been observed, with no clear pattern of an association with elevations in pro-inflammatory related factors. For example, Repunte-Canonigo et al. (2014) reported an increase in microglia number in the hippocampus and cortex but changes in related genes occurred only in the hippocampus. Further observations of a morphological response of glia in the absence of an associated pro-inflammatory response were provided by Reid et al. (2016a) and Rowson et al. (2016) reported mild morphological differences in microglia with a slight elevation in complement factor b (Cfb) but no measurement of pro-inflammatory factors. Using translocator protein TSPO imaging and mRNA levels of pro-inflammatory cytokines, Lee et al (2015) reported no evidence of inflammation in the cortex, hippocampus, or striatum of 9 month-old HIV-1 Tg rats. The conflicting and less than robust reports raise doubts with regards to an early contribution of neuroinflammation and suggest alternative functions for the few morphological changes observed in microglia and astrocytes.

The current study was undertaken to evaluate alterations in HIV-1 Tg rats across a number of related factors. Rota-rod performance and motor activity were conducted to evaluate motor related deficits and compare to the literature and learning and memory was assessed using an active avoidance paradigm. TH immunoreactivity in the SN was examined across ages to examine alterations in dopaminergic neurons to compare to changes in TH immunoreactivity observed in the striatum of HIV-1 Tg rats (Reid et al., 2016a) and SN of GT-tg bigenic mice (Miller et al., 2018). Morphological alterations in astrocytes and microglia and mRNA levels for inflammatory factors were examined in specific brain regions previously reported. It is our expectation, that while descriptive, the collation of these endpoints across ages and in the same animal will help to set the framework to more fully describe the relationship. Such information will facilitate our ability to fully utilize this experimental model to address relevant biological and therapeutic questions as they may apply to the human patient population.

Materials and Methods

Animals

Male HIV-1 transgenic rats (Fischer 344/NHsd; Tg) and Fischer 344 (F344) rats were obtained from Harlan Laboratories (Madison, WI). All HIV-1 Tg rats displayed cataracts. Male rats were examined as based upon the predominance of data from male rats, or unspecified sex distributions, in the existing literature on neuroinflammation in HIV-1 Tg rats. Rats were assessed across multiple cohorts and randomly assigned to endpoints and age of analysis. Animals were housed within an Association for Assessment and Accreditation of Laboratory Animal Care accredited facility (40–60% humidity; 12-h light/dark cycle: 6:00–18:00 EST; 20–24^oC) under isolator housing conditions to minimize any possibility of non-specific immune challenge. Rats were allowed ad libitum access to reverse osmosis deionized drinking water and maintained on their Harland Labs original diet (Teklad global 18% protein 2018S diet; Teklad Harlan, Madison, WI; sterilized for controls and gamma-irradiated for HIV-1 Tg rats to minimize risk of infection). The diet contained (as % of total fatty acid) 16.7% saturated, 21.8% monounsaturated, 54.8% linoleic acid, 6.2% α-linolenic acid, 0.03% AA, 0.02% eicosapentaenoic acid (EPA, 20:5 n-3) and 0.06% docosahexaenoic acid (DHA, 22:6). Animal studies were done in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals following approved animal protocols from the Animal Care and Use Committee of the National Institute of Environmental Health Sciences.

Behavioral Assessments

Behavioral assessments of motor function used test systems similar to those in the literature and learning and memory was evaluated within each of these tests and with a non-visually dependent task of active avoidance. Assessments were conducted at ages prior to overt evidence of immunohistological changes in the SN. Handling and husbandry of rats for behavioral testing followed the National Toxicology Program (NTP) guidelines for neurobehavioral testing (NTP, 2015). Rats were transported to animal facility holding area 1 h prior to testing to allow for acclimation. All testing was conducted between 10:00–15:00 h. Assignment to testing chamber and to time of testing was counterbalanced and animals were assessed under experimenter-blinded conditions.

Rota-rod

One of the more consistent effects reported in the HIV-1 Tg rat is a deficit in rota-rod performance (June et al., 2009; Reid et al., 2016b) and thus, was included with assessment at 6 weeks (n=18), 12 weeks (n: F344 = 14; HIV-1 Tg = 17), and 24 weeks (n=7) of-age to compare our findings to those reported in the literature. Performance on an accelerating rota-rod was evaluated using a Rotamex-5 (Columbus Instruments, Columbus, OH) equipped with the rat spindle (7 cm x 9.5 cm). Rats were placed on the stationary spindle and rotations initiated at acceleration increments of 1 rpm/10 sec. Rats were allowed two initial training trials followed by a sequence of 6 trials with an inter-trial interval (ITI) of 30 min. Latency for remaining on the rotating rod, as automatically photocell-detected, was recorded with a maximum time of 5 min allowed for each trial. At 8 months-of-age, the weight and size of the rats were such that the rat spindle was not of sufficient size to allow the rats to perform the task and thus, all rats fell within the first few rotations and the task did not provide a valid assessment. This data is therefore not provided.

Motor Activity

Motor activity has been previously reported to be diminished in adult male HIV-1 Tg rats 11–25 weeks of age (Reid et al., 2016b). Adult rats (24 weeks-of-age; n=10) were assessed for exploratory ambulatory activity in a novel environment. Activity was recorded in 5-min epochs over 30 min in an OptoMax Activity photocell device (42cm x 42cm x 20cm; Columbus Instruments, Columbus, OH) under dim-light conditions. Ambulatory activity, time spent within the margin of the arena, and distance travelled in the margin zone of the arena were recorded.

Active Avoidance

As a measure of learning and memory, rats were sequentially examined at 6 (n=8), 12 (n=8), and 24 weeks (n=7) for performance in an active avoidance procedure. Animals were placed into one compartment of the shuttle box (Gemini II shuttle box, San Diego Instruments, San Diego, CA), the guillotine door was opened, and a 2-min general exploratory period was initiated. The session was initiated with the delivery of a cue light and tone (conditioned stimulus, CS) on the side of the apparatus containing the rat. The CS was delivered 10 sec prior to and continued throughout the delivery of a 10-sec 0.6 mA scrambled foot shock (unconditioned stimulus; US), resulting in a 20-sec response period. Movement of the rat to the “safe” side terminated the CS and US. A total of 60 CS/US pairings were delivered within a session on a 15-sec variable ITI schedule. Avoidance responses and latency were recorded for the animal moving to the “safe” side during the 10-sec interval between CS and US. Escape responses were recorded for shifting to the “safe” side during the US delivery. Mean responses were calculated for each 10-trial unit for a total of 6 units representing 60 trials.

Tissue Collection for Histology

Rats were euthanized with isofluorane, whole body cardiac perfused with saline, decapitated and the brain rapidly excised, cut in the midsagittal plane, and one hemisphere immersion-fixed in 4% paraformaldehyde/phosphate buffer (pH 7.2) for 18h, rinsed, and placed in FD Tissue Cyroprotection Solutions (FD NeuroTechnologies, Inc, Columbia, MD). Samples were maintained at −20^oC until sectioning. Cyroprotected brain hemispheres were shipped to FD NeuroTechnologies (Ellicott City, MD) for sectioning and immunostaining. Forty-μm free-floating coronal serial cryo-sections (every 1–10^th section of each series of 10 sections with an interval of 400 μm between 3.00 and −7.04 mm Bregma) were collected and stored in cryoprotection solution (FD Neurotechnologies, Baltimore, MD) at −20°C. For comparison to the existing literature, morphology of microglia and astrocytes was examined in the hippocampus, cortex, striatum, and SN.

Unbiased Stereology for TH⁺ Neurons in SN

To determine if TH immunoreactivity was diminished in neurons of the SN/VTA, unbiased stereological analysis was conducted. Systematic-random samples of 40 μm serial sections through the SN_PC-VTA were collected between Bregma −4.36 mm to −7.04 mm with an interval of 120 μm. Sections were washed with 1xTRIS-buffered saline (TBS; Quality Biological, Inc, Gaithersburg, MD) and endogenous peroxidases quenched by 1% hydrogen peroxide for 30 min at RT, followed by rinses with 1xTBS and incubation with 0.3% Triton-X. Sections were blocked with 5% normal goat serum in 1×TBS and then incubated with 1:2000 rabbit anti-tyrosine hydroxylase (Santa Cruz Biotechnology, Dallas, TX) overnight at 4°C. Following primary incubation, sections were rinsed with 1xTBS and incubated in biotinylated anti-rabbit secondary antibody (1:400; Vector Laboratories, Burlingame, CA) with normal goat serum in 1xTBS for 90 min at RT. Rinsed sections were incubated using Vectastain Elite ABC kit for 90 min at RT, washed in 1xTBS and treated with 3-diaminobenzidine (DAB; Agilent Technologies, Santa Clara, CA) [10 mg DAB in 40 ml 1xTBS] and chromogen with nickel chloride amplification. Sections were mounted on poly-L lysine coated slides and coverslipped.

Total number of TH⁺ neurons in the SN reference space was quantified using the optical fractionator method and the Stereologer system (Stereology Resource Center, Chester, MD), as previously reported (Mouton et al., 2002; Marcario et al., 2004; Mouton and Gordon, 2010). The reference space was outlined under low magnification (5×) for each section and TH⁺ neurons counted at high magnification (63× oil immersion), with a guard volume of 2 μm. For counts of total neuron number, the following sampling fractions were used: section sampling fraction (ssf, number of sections sampled divided by the total number of sections); area sampling fraction (asf, area of the sampling frame divided by the area of the xy sampling step); and thickness sampling fraction (tsf, height of the dissector divided by the section thickness). The counting criteria were TH immunoreactivity in the cytoplasm of cells with a neuronal phenotype, including a clear nuclear membrane and distinct nucleolus. Sampling of x-y locations was continued to a high level of sampling stringency, i.e., 0.10 to 0.15 mean coefficient of error (CE, Gundersen et al., 1999) per group.

Immunostaining for Glial Fibrillary Acidic Protein (GFAP) and Ionized Calcium-binding Adapter 1 (Iba-1)

Sections containing the SN (−4.0 to −7.0 from Bregma), striatum (−2.6 to −1.mm from Bregma), and hippocampus (−3.0 to −4.5 from Bregma) were washed with phosphate buffered saline (PBS), equilibrated to room temperature (RT), and incubated for 2 h in a blocking solution containing 2% goat serum, 1% bovine serum albumin, and 0.1% Triton X-100 in automation buffer (Biomedia, Foster City, CA). Sections were incubated with anti-GFAP (1:200, Dako Agilent Technologies, Carpinteria, CA) or anti-Iba-1 (1:500, Wako Chemicals, Richmond, VA) in blocking solution (18 h; 4°C), re-equilibrated to RT. Rinsed sections were incubated using Vectastain Elite ABC kit, and visualized by DAB. Sections used for immunofluorescence were incubated with Alexa Fluor antibody conjugates (594 nm or 488 nm; 1:250, Invitrogen ThermoFisher, Carlsbad, CA) in blocking solution without Triton X-100 (2 h; RT), mounted on charged slides in Prolong with DAPI (Invitrogen, ThermoFisher), and coverslipped.

Digital images of immunostaining were collected using a LSM 410 inverted confocal laser-scanning microscope (Carl Zeiss, Oberkochen, Germany). For acquisition, the pinhole diameter, scan speed, and amplifier gain were kept constant. The detector gain and amplifier offset were determined from the range indicator function in Fast XY mode. Image stacks were collected through the full depth of the tissue. At 20x magnification 1.5μm steps were collected with a 2048 × 2048 frame size; at 63x magnification, 1.0μm steps were collected with a frame size of 1728 × 1728. Stacks were displayed as a single image using 3D maximum projection. DAB sections were scanned under 20x magnification (Aperio ScanScope T2 scanner, Aperio Technologies, Inc., Vista, CA), image stitched, and viewed using Aperio ImageScope v.6.25.0.1117. Stained slides were assigned random numbers and blinded for evaluation. Defined regions of interest (ROI) in the hippocampal dentate gyrus, motor cortex, striatum, and SN were evaluated for overt changes in astrocyte or microglia morphology and any evidence of neuronal loss. The scoring system for morphology (20 cells/2 sections/ rat) was based on previously published work and reflected the different morphological features observed within each region (Heppner et al., 1998; Kanaan et al., 2008; Harry et al., 2014). Observations of a thickened proximal process of microglia in the striatum were confirmed using Image J 1.48v (National Institutes of Health). with the width of the cell body and the proximal process calculated for a minimum of 20 cells within the ROI for each section. For the hippocampus and cortex, images of Iba-1 immunohistochemical (DAB) sections were analyzed using Contrast was optimized within the green channel grayscale images and the threshold was set to a lower level of 0 and an upper level between 175–185. This range allowed for inclusion of all Iba-1⁺ microglia within the image plane. The area fraction occupied by Iba-1⁺ cells was calculated using the measurement tool and expressed as percent total area.

qRT-PCR for Inflammatory Factors

From the contralateral hemisphere to that used for histology, the striatum, motor cortex, and hippocampus were dissected. Under a dissection microscopy, an area enriched for the SN/VTA was micro-punched from the relevant brain slice obtained using a brain matrix. Samples were stored at −80^oC. Total RNA was extracted using TRIzol Reagent (Invitrogen, ThermoFisher). RNA quantity and purity were assessed using NanoDrop (Thermos Scientific, Wilmington, DE) and complementary deoxyribonucleic acid (cDNA) was synthesized from 2.5 μg total RNA using SuperScript™ II reverse transcriptase with random hexamers (Invitrogen, ThermoFisher). qPCR was performed in duplicate using 2.5 μl cDNA as a template in combination with Power SYBR Green Master Mix (Applied Biosystems, Foster City, CA) and optimized forward and reverse primers (Supplementary Table 1). The reaction mixtures were held at 50° C for 2 min, 95° C for 10 min, followed by 40 cycles at 95° C for 15 s and 1 min at 60°C. Amplification curves were generated with sequence detection system 1.9.1 software (Applied Biosystems). Threshold cycle values were determined and the mean fold changes over age-matched WT controls were calculated according to the 2^ΔΔCT method and normalized to housekeeping gene. For all primers, a melting curve analysis was performed by denaturation at 95°C for 15 sec, annealing at 60°C for 1 min at a melting rate of 0.3°C/sec to 95°C. Each qRT-PCR was carried out in sample duplicates and replicated with two different housekeeping genes, glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and cyclophilin, by different investigators. Data were obtained from samples that met inclusion criteria of transcript detection at ≤ 32 PCR cycles.

Statistical Analysis

Statistical analyses were performed using GraphPad Prism software (GraphPad Software, Inc., San Diego, CA). Levene’s test was used to test for homogeneity of variance. Rota-rod, motor activity, and active avoidance at each age were analyzed by repeated measures ANOVA (RM ANOVA) with genetic background and epoch/trial as main factors. Sidak’s multiple comparison test was used to examine differences at individual trial blocks. Unbiased stereology data and imaging density data were analyzed by Student’s t-test as each age assessment was conducted in independent groups of rats. Rating scale for microglia morphology in striatum was calculated as percentage of total microglia within the ROI and analyzed by Mann Whitney U test. Thickness of microglia processes was analyzed by Student’s t-test. A maximum of 5 endpoints for qPCR were assessed in any one independent group of RT samples. Student’s t-tests were conducted with brain region considered as an independent factor given that, while from the same rat, the samples were experimentally handled as independent. To address issues of experiment-wise error rate, statistical analysis of mRNA was conducted on transcripts that showed a minimum of a 20% difference. False discovery rate was determined using the Benjamini Hochberg procedure (BH). The order of analysis for transcripts was maintained across regions and any BH corrections are noted in the text. Statistical significance was set at p<0.05. Results are expressed as mean ± standard deviation or standard error as indicated.

RESULTS

Behavioral Testing

Behavioral assessments of motor coordination were conducted at 6, 12, and 24 weeks of age, allowing for a comparison to a clear reproducible effect reported in the literature. Learning and memory were assessed using a shock based active avoidance procedure at 6 and 12 weeks of age. Motor activity was assessed at 24 weeks of age for comparison to findings in the literature. As noted in the methods section, the size and weight of rats the older ages presented issues with assessing behavior in these tasks.

Rota-rod

Performance on the accelerating rota-rod represents primarily motor strength and coordination but can also reflect task-specific learning as indicated by an increase in performance with training. Consistent with previous reported deficits in motor strength and coordination in HIV-1 Tg rats (June et al., 2009; Reid et al., 2016b), an overall deficit in rota-rod performance was observed in HIV-1 Tg rats (Fig. 1A). Over age, performance in all rats declined (Reid et al., 2016b). At 6 weeks-of-age, latency-to-fall over trials, was significantly shorter in HIV-1 Tg rats as compared to F344 rats (F_(1,34)=11.12; P=0.0021). However, both F344 and HIV-1 Tg rats showed a similar increase in performance over successive trials (F_(5,170)= 9.275, p<0.0001). At 12 weeks-of-age, all rats showed slightly lower latencies than that observed at 6 weeks of age (Fig. 1A). Over the session, HIV-1 Tg rats displayed significantly shorter latencies to fall as compared to F344 rats (F_(1,29)=5.69, p=0.0238). Yet, an increase in performance over successive trials was observed in both groups (F_(5,145)=6.718, p<0.0001), with no evidence of a significant interaction between trial and genotype. At 24 weeks-of-age, all rats showed shorter latencies to fall as compared to younger ages but also failed to show a betterment in performance over trials (Fig. 1A). No difference was observed between groups; however, this absence of differences was likely related more to the overall age and animal size-related decrease in performance rather than genetic differences. The overall age-related decrease in performance seen between 6 and 24 weeks-of-age across all animals and the performance deficit observed in HIV-1 Tg rats are consistent with findings reported by Reid et al. (2016b). They do however, raise questions with any interpretation of data obtained from the older, larger rats.

Figure 1.

A. Accelerating Rota-Rod. Latency to fall from a rota-rod was determined across 6 trials (3–8), following 2 training trial, in F344 and HIV-1 Tg male rats at 6 (n=18), 12 (F344 n=14; HIV-1 Tg n=17), and 24 weeks-of-age (n=7). Performance increased over successive trials at 6 and 12 weeks-of-age (p<0.0001). HIV-1 Tg showed significantly overall shorter latencies at 6 (P<0.01) and 12 weeks (p<0.05) but not at 24 weeks. B. Open-field Activity in 5 month-old rats (n=10). Activity significantly decreased over the session (p<0.0001) in both groups. Time spent within the margin was significantly less in the HIV-1 Tg rats (p<0.05). No difference was observed for distance travelled in the margin. C. Shuttle-box Active Avoidance at 6 and 12 weeks of age. Avoidance latency (sec) and # of avoidance responses occurring over 6 blocks of 10 trials each. At 6 and 12 weeks-of-age (n=8), rats showed acquisition of the task in latency (p<0.0001) and # avoidance responses (p<0.0001). At both ages, HIV-1 Tg rats displayed longer latencies (p<0.05) and lower # of avoidance responses (p<0.05). At 12 weeks, the latency response pattern for HIV-1 Tg rats was similar to that observed on training at 6 weeks while F344 reached performance plateau by trial block 3. Data represents mean +/− SEM * denotes p<0.05.

Motor Activity

Ambulatory activity within the open field provides an assessment of exploratory activity and is often considered to reflect a dopaminergic-related motor behavior. In 20 week-old rats, total ambulatory activity was similar between groups (Fig. 1B). A normal pattern of activity was observed characterized by a progressive decrease in activity over the session across 5-min epochs (F_(5,90)=85.22, p<0.0001) with no significant differences observed between F344 and HIV-1 Tg rats. The absence of any difference in ambulatory activity at this age is consistent with findings of Reid et al. (2016b). The test paradigm however is somewhat different in that motor activity was assessed over 2-week intervals in the Reid et al. (2016b) study thus, acclimating the rats with the open-field environment no longer having characteristics of a novel environment towards the end of assessment. As compared to the analysis of total session activity levels in the Reid et al. (2016b) study, the current study examined the response of rats naïve to the novel open-field environment and examined activity over time epochs to examine habituation. A normal pattern of activity was observed in both groups, characterized by a progressive decrease in activity over 5-min epochs (F_(5,90)=85.22, p<0.0001), suggestive of normal response and habituation to a novel environment. Based on a previous report that HIV-1 Tg rats display a greater “anxiety-like” response (Reid et al., 2016b), ambulatory activity localized to the margin, thigmotaxis, was recorded (Fig. 1B). Consistent with the decrease in overall activity levels across epochs, activity in the margin decreased over the session in both groups (time spent: F_(5,90)=3.37, p=0.0191; distance travelled: F_(5,90)=8.483, p<0.0001). In contrast to Reid et al. (2016b), total time spent within the margin zone (F_(1,18)=6.621, p<0.05) was less in HIV-1 rats while no difference was observed for total distance traveled (p<0.1). F344 rats showed an increase in time in the margin during the 5^th and 6^th epoch that was not observed in the HIV-1 Tg rats (p<0.05). The difference in margin time was not related to differences in overall activity levels of the rats. Thus, at this age, we observed no evidence of “anxiety-like” behavior with regards to preference for the margin region.

Active Avoidance

In the active avoidance task, rats were required to learn an association between the conditioned stimulus (CS: i.e., light/tone) and the delivery of the unconditional stimulus (US: i.e., foot shock) to respond in a manner to avoid receiving the shock. As training proceeds, animals learn to avoid the US by shuttling across a divided chamber during the CS. This crossing terminates the tone and prevents receipt of the shock. Signaling for active avoidance requires an intact flow of information between the basal amygdala and the shell region of the nucleus accumbens (Ramirez et al., 2015) and possibly requires suppression of the amygdala-mediated defensive reactions by the prefrontal cortex (Moscarello and LeDoux, 2013). At 6 weeks-of-age, all rats showed a significant decrease in avoidance latency to respond to the CS over session trials (F_(5,70)=13.25, p<0.0001), indicative of learning (Fig. 1C). However, significant overall longer avoidance latencies across trials were observed in HIV-1 Tg rats as compared to F344 rats (F_(1,14)=8.322, p=0.012) with no significant interaction observed between trials and genetic background. Avoidance performance (#avoidance responses) significantly improved over the session (F_(5,70)=7.983, p<0.0001) for both groups, again suggestive of learning however, a deficit was suggested with the significantly less # avoidance responses in the HIV-1 Tg rats as compared to F344 rats (F_(1,14)=7.853, p<0.05)(Fig. 1B). By 12 weeks-of-age, a progressive decrease in avoidance latency over trials was observed for both groups (F_(5,70)=16.68, p<0.0001; Fig. 1C). HIV-1 Tg rats showed a higher overall avoidance latency, similar to that seen at 6 weeks and, while the overall differences across trials failed to reach statistical significance (p=0.08). F344 rats show improved performance by the 2 and 3^rd trial as compared the HIV-1 Tg rats, suggestive of a memory deficit. The # avoidance responses significantly increased over trials for both groups (F_(5,70)=15.24, p<0.0001; Fig. 1B). HIV-1 Tg rats showed overall significantly lower # avoidance responses (F_(1,14)=5.242, p<0.05) reaching a maximum level by trial block 5 as compared to trial block 3 in the F344 suggestive slight effect on memory. At 24 weeks of age, rats had progressed to a body size and weight that significantly limited their ability to transgress through the door connecting the two chambers. This resulted in significantly lower avoidance responses (2–3) and longer latencies (8–10 sec) in both groups and by the 3^rd trial block of the session all responses were “escape losses” with the rats adopting a freezing behavior. Sessions were terminated at the end of the 3^rd trial block. Data is not reported.

TH Immunohistochemistry and Unbiased Stereology in the SN

TH expression in the SN is considered to reflect dopaminergic neuronal integrity. Thus, we examined TH staining to reveal whether HIV-1 Tg rats exhibited alterations of the dopaminergic system. By 8 months of age, a significant decrease in TH+ immunoreactive neurons was observed in HIV-1 Tg rats as compared to age-matched F344 rats (t=5.198, p=0.002; Fig. 2A). As HIV-1 Tg rats constitutively express HIV viral proteins throughout development, TH+ immunoreactivity was examined at earlier ages to determine if the changes were developmentally present or if they represented a loss over time. Between 2 and 6 months of age, the number of TH+ immunoreactive neurons was similar between HIV-1 Tg and F344 rats (Fig. 2A). Representative images of TH immunohistochemistry in the SN showed a normal dense staining in neuronal cell bodies and processes of the SN at both 8 and 5 months (Fig. 2B,C). In HIV-1 Tg rats, TH immunoreactivity within neuronal cell bodies and processes was diminished at 8 months (Fig. 2B) but not at 5 months (Fig. 2C).

Figure 2.

Tyrosine Hydroxylase Immunoreactivity and Unbiased Stereology. A. Estimates of TH+ neuronal number between 2 and 8 months of age. Scatter graph of individual values (% of F344 control) at each age. Bar graphs of mean +/− SEM for 2–5 months (n=9) and 8 months (n=4–5). No difference was observed between 2–5 months of age. At 8 months-of-age, HIV-1 Tg rats showed significantly fewer TH+ neurons as compared to age-matched controls (p=0.002; n=4–5). B. Representative images of TH immunoreactivity (3,3-diaminobenzidine staining: brown, black) in the SN at 8 months-of-age. C. At 5 months of age, TH immunoreactivity in the SN showed no difference between F344 and HIV-1 rats with DAB staining or fluorescence (green) staining. Iba-1+ microglia showed similar morphology across groups as shown by immunostaining for Iba-1 (red, DAB). GFAP staining (DAB) showed no difference between groups.

GFAP and Iba-1 Immunohistochemistry in SN and Striatum

Microglia rapidly respond to changes that occur in neurons in close proximity and with synaptic damage and remodeling. Given the question of whether changes in glia might serve as an early indicator of neuronal alterations in the SN, we examined morphological changes in the SN and the striatum of rats at 5 months-of-age as representative of a time prior to pronounced changes in TH immunoreactivity in the SN. In the SN, Iba-1+ microglia and GFAP+ astrocytes showed a normal distribution and morphology (Fig. 2C; Supplementary Fig. 1). In the striatum, GFAP+ astrocytes showed similar morphology across groups (Fig. 3). In F344 rats, Iba-1+ cells (Fig. 3) were characterized by small cell bodies and long, thin processes. As compared to F344 rats, immunostaining for Iba-1 revealed a slight change in microglia in HIV-1 Tg rats. The visual appearance of a slightly more pronounced microglia morphology was supported by morphometric measurements in 70%+/−10% of the microglia showing slightly larger (10% +/− 5%) cell bodies and thickened proximal processes (30%+/−8% (t= 5.8; p<0.01) within a defined ROI.

Figure 3.

Immunoreactivity for GFAP and Iba-1. Representative images of GFAP⁺ astrocytes and Iba-1⁺ microglia in the striatum of F344 and HIV-1 Tg 5 month-old male rats. Immunofluorescent staining for GFAP (red) with DAPI (blue) counterstain showed astrocytes with elongated thin processes in F344 and HIV-1 Tg rats with no evidence of astrocyte hypertrophy. Iba-1⁺ microglia (red) were characterized as thin-process bearing cells in F344 rats. In HIV-1 Tg rats, microglia were similar but displayed slightly thickened proximal processes.

Astrocyte and Microglia Morphology in Hippocampus and Cortex

Previous studies have reported conflicting findings with regards to measures of astrocyte or microglia in the hippocampus or cortex of the HIV-1 Tg rat (Table 1). GFAP staining of astrocytes has been reported as either increased (Repunte-Canonigo et al., 2014) or decreased (Reid et al., 2016a). In the current study, a relatively uniform staining intensity and morphological distribution was observed for GFAP⁺ astrocytes in both regions with no differences observed between F344 and HIV-1 Tg rats at 5 months (Fig. 4). Astrocytes maintained a normal cell body with thin processes and showed no evidence of hypertrophy. A similar pattern of conflicting results has been reported for microglia. Previous reports indicated that HIV-1 Tg rats showed no difference in microglia number in the cortex or hippocampus but an increase in microglia arborization (Repunte-Canonigo et al., 2014), number (Rowson et al., 2016), or no changes (Lee et al., 2015). In the current study, DAB stained sections were examined for comparison to the literature and revealed a robust staining of microglia with no morphological differences observed in the HIV-1 Tg rats at 5 months of age (Fig. 4) Fluorescent staining showed similar results (Supplementary Fig. 2; Supplementary Fig. 3). Iba-1⁺ microglia retained normal morphology characterized by a small cell body and fine ramified processes. No morphological evidence of microglia hypertrophy or of differentiation into an amoeboid phagocytic phenotype was observed. The total surface area occupied by Iba-1 immunoreactivity as calculated within multiple defined ROIs in the cortex and hippocampus showed no significant differences in HIV-1 Tg rats, as compared to F344 (Fig. 4).

Figure 4.

Immunoreactivity for GFAP and Iba-1. Representative images of GFAP⁺ astrocytes and Iba-1⁺ microglia in the cortex and hippocampus of 5 month-old male F344 and HIV-1 Tg rats. 3,3-diaminobenzidine (brown) stained cells displayed normal process-bearing morphologies with no evidence of microglia activation or astrocyte hypertrophy. Quantitation of the % area occupied by Iba-1⁺ immunoreactive cells (mean +/− SD) in the cortex (n=6) and hippocampus (n=10) is provided within each representative image.

mRNA Levels for Inflammatory Factors

The published literature is inconclusive with regards to evidence of an inflammatory response occurring in the brain of HIV1-Tg rats. Much of this appears to be related to the method of analysis, the age of analysis, the region examined (Table 1) and could possibly be associated with a phenotypic drift in the rats over time. We focused our analysis on changes in inflammatory markers that occurred at 5 months-of-age, prior to the occurrence of diminished TH immunoreactivity in the SN and within an age range for an association to neurobehavioral alterations. As a comparison to previous published reports, multiple regions were examined including the striatum, SN, cortex, and hippocampus.

3.4.1. Substantia Nigra and Striatum

The absence of overt morphological changes in astrocytes in the SN and striatum was supported by absence of differences in GFAP (Fig. 5a,b). In the SN, Iba1, Cdllb, Tnfa, Il1a, Il1b, C-X-C motif chemokine receptor 3 (Cxcr3), and intracellular adhesion molecule (Icam) mRNA levels were not significantly different in HIV-1 Tg rats as compared to F344 rats (Fig. 5). In the striatum, the subtle morphological change in microglia morphology in the HIV-1 Tg rat was accompanied by a significant elevation in mRNA levels for Iba1 (t=2.719, p<0.05). HIV-1 Tg rats showed significantly elevated mRNA levels for Cdllb (t=2.378, df=6, p<0.05) and significantly lower mRNA levels for Il1a (t=3.022, df=6, p<0.05) and Icam (t=4.322, df=6, p<0.01 (p<0.02 corrected). The elevation in Cxcr3 was no longer statistically significant following multiple comparisons correction (t=2.07, df=6, p<0.05; p<1.0 corrected). No differences were observed in mRNA levels for Tnfa, Il1b, or or the inflammasome related NLR family pyrin domain containing 3 (Nlrp3) (Fig. 5b ). Differences were not observed at 2 months-of-age (Supplementary Fig. 4).

Figure 5.

mRNA Levels for Inflammatory Genes in 5 month-old F344 and HIV-1 Tg male rats. A. Substantia nigra (SN); B. Striatum; C. Cortex, D. Hippocampus. mRNA levels were determined by qRT-PCR and normalized to cyclophillin in each sample. Data was calculated by 2”^ΔΔCT presented as mean fold change in 5 month-old HIV-1 Tg male rats as compared to age-matched F344 rats +/− SEM (n=4). * denotes p<0.05; ** p<0.01.

3.4.2. Cortex and hippocampus

Previous reports of immunohistochemical changes in GFAP astrocytes were accompanied by an increase in GFAP mRNA levels in the hippocampus but not the cortex (Repunte-Canonigo et al., 2014) or with no changes in GFAP mRNA levels in the hippocampus even with substantially diminished GFAP staining intensity (Reid et al., 2016a). In the current study, consistent with the absence of immunohistochemical changes, mRNA levels for Gfap and Iba-1 in the hippocampus and cortex were not significantly different in the HIV-1 Tg rat as compared to the F344 controls (Fig. 5c,d). In the cortex, mRNA levels for Tnfa, Il1a Nlrp3 were not significantly altered in HIV-1 Tg rats as compared to F344 rats (Fig. 5c). In the hippocampus, mRNA levels for Il1a, Il1b, or Nlrp3 were not altered however, increases were observed for Cd11b (t=3.504, df=6, p<0.01(p<0.02 corrected)) and Tnfa (t=3.77, df=6, p<0.01 (p<0.02 corrected); Fig. 5d).

Discussion

In the human population, the association of neuroinflammation with the progression of HAND has strong support both in the pre and post-cART eras (review, Hong and Banks, 2015). An age-dependent scenario has been proposed that involves alterations in neurotrophic factors (Bachis et al., 2016) or normal microglia function (DeVaughn et al., 2015) that can influence synaptic strength and remodeling. An involvement of dopaminergic system disruption appears to have strong supporting data but at what stage a contribution of inflammatory factors might come into play cannot be adequately evaluated by the current conflicting literature. Using the HIV-1 Tg rat as a model, we report an age-related loss of TH⁺ immunoreactivity in the SN that supports the various reports of an alteration in dopaminergic integrity. Microglia and astrocytes rapidly respond to changes in their environment and we observed subtle changes in microglia morphology within the terminal dopaminergic synaptic field in the striatum. The thickening of the proximal process is consistent with an responsive action of microglia and given the location as well as absence of mRNA elevations in various pro-inflammatory cytokines, may be primarily indicative of early TH neuronal stress or dysfunction rather than cell death. No glial-related changes were noted in the cortex and only subtle elevations in Cd11b and Tnfa mRNA levels in the hippocampus, somewhat diminishing the hypothesis that behavioral changes were related to a neuroinflammatory process.

Even in the cART era, HIV-1 positive individuals display symptoms of cognitive impairment, some of which associate subcortical and frontal-striatal pathways (Becker et al., 2011). The utility of the HIV-1 Tg rat in assessing neurobehavioral functions has been well established with a large number of the studies conducted in animals between the ages of 2 and 6 months. One primary endpoint reported for HIV-1 Tg rats is a deficit in rota-rod performance (June et al., 2009; Reid et al. 2016b). Our data is consistent with these findings in that HIV-1 Tg rats showed poorer performance at young ages however, over a session performance improved in both groups suggesting effective learning of the task. This conclusion is somewhat in disagreement with that of Reid et al. (2016b) where performance was compared across weeks and the authors concluded that the animals failed to habituate to the task given that the 60% decrease in latency did not reach statistical significance. The Reid et al. (2016b) study and the current study, demonstrated that rota-rod performance diminished with age such that, by the adult ages it became difficult to discriminate between poorer motor performance and learning. Motor activity data have also been somewhat conflicting in HIV-1 Tg rats. Previous work reported no difference in motor activity over a 10-min session in 2 month-old HIV-1 Tg rats (Nemeth et al., 2014) while, adult rats showed lower ambulatory and rearing activity over a 20-min test session (June et al., 2009). In comparison, in adult rats we found no differences in ambulatory activity over a 30-min test session or at shorter intervals. Acclimation to a novel environment, such as an activity arena, can represent learning as well as response to stress. In adult male HIV-1 Tg rats we did not observe activity patterns that would indicate an anxiety response in a novel environment. Additional assessments of learning and memory have utilized a modified Morris Water Maze (MWM) adjusted for cataract-related visual deficits in the HIV-1 Tg rat. Vigorito et al. (2007; 2013) reported acquisition of the task and thus, learning, in 5-month-old HIV-1 Tg rats; however, longer latencies were observed during the acquisition phase without a clear deficit in swim speed. In the probe test for memory, all rats displayed a similar preference for the escape quadrant suggesting equivalent learning and memory. Lashomb et al. (2009) expanded on this work and confirmed the ability of HIV-1 rats to learn the MWM however, impairments were noted in search strategies that may involve the striatum circuitry (Packard and White, 1991). The authors speculated that these outcomes were mediated by the D2 dopamine pathway of the basal ganglia, hippocampus, and cortex given the prolonged escape latencies without decreased swim speed (Stuchlik et al., 2007). Using tasks not as dependent on latency, Nesil et al. (2015) reported deficits in working memory in a spontaneous alternation maze paradigm, exploration and memory in a novel object recognition task, in a step-through passive avoidance task in 2-month-old HIV-1 Tg rats. In the current study, we relied on an active avoidance procedure and found that HIV-1 Tg rats displayed a longer latency for the avoidance response and less number of avoidance responses however, performance improved across the session. It is possible that such deficits in latency were related to changes in the dopaminergic pathway. The increased performance observed with training suggested basic learning capabilities in the HIV-1 Tg rats but in the retest 6 weeks later, performance suggested memory of the task may not be as well formed. While this behavior was evident prior to changes in TH immunoreactivity in the SN, it may still be related to an alteration in the dopaminergic pathway similar to that proposed for the MWM. It is possible that aging or a pharmacological challenge may unmask early alterations in dopaminergic-dependent behaviors and identify a greater level of compromised function in the HIV-1 Tg rat.

In experimental models of HIV, pathology and dysfunction of the dopaminergic system have been reported (Bansal et al., 2000; Gelman et al, 2006; Silvers et al., 2006; Webb et al., 2010; Li et al., 2013; Reid et al., 2016a,b; Gaskill et al., 2017). In HIV-1 Tg rats, a dysregulation of the dopaminergic system has been suggested by molecular profiles related to alteration of DA transmission and Parkinson’s disease (Repunte-Canonigo et al., 2014), a slight elevation in [¹⁸ F]DPA-714] uptake (Lee et al., 2015) and decreased TH staining intensity in the striatum (Reid et al., 2016b). PET imaging of [18F]-Fallypride striatal binding in aged HIV-1 Tg rats suggested a progressive degeneration of the DA system and decreased dopaminergic synaptic function in the striatum (Sinharay et al., 2017). Similar to the current findings, Miller et al., (2018) reported diminished TH immunostaining in the SN in GT-tg bigenic mice expressing HIV-1 Tat under GFAP. This occurred with reduced firing activity of DA neurons in the absence of neuronal death. While these deficits are observed in adult animals, it is also likely that subtle differences in the circuitry are present at younger ages. The work or Casas et al. (2017) suggested an effect of diminished volume and growth rate of the striatum between 5 and 9 weeks-of-age in HIV-1 Tg rats. This maturation pattern was functionally associated with rota-rod performance. In the current study, the early performance deficits observed on the rota-rod may reflect early alterations in the nigra-striatal network as impairment of the striatum can lead to motor coordination issues (Massaquoi and Hallett, 1998).

It is thought that HIV viral products such as gp120, Tat, nef, and vif may promote ongoing inflammation and possible degeneration (Gannnon et al. 2011; Zayyad and Spudich, 2015). A subset of mononuclear phagocytes and astrocytes in HAND patients have been implicated as the source of pro-inflammatory cytokines and other neurotoxic molecules that may exacerbate damage to surrounding cells (Yadav and Collman, 2009). How this translates to the rodent models is not as clear given the conflicting findings of glial response and cytokine elevation (Table 1). For example, astrocyte and microglia number was increased in the hippocampus and cortex of 5-month old HIV-1 Tg rats yet, mRNA levels for Gfap, Iba1, and Cd11b were elevated only in the hippocampus (Repunte-Canonigo et al., 2014). With a detailed evaluation of microglia morphology in the hippocampus of young female rats, Rowson et al. (2016) reported no difference in number or average branch length but an increase in branch number, maximum length, and junctions. This was accompanied by a slight elevation in mRNA levels for complement factor b with no change in Lcn2. There was no indication of hypertrophy or cells differentiating to an ameoboid phenotype. Thus, a relationship of microglia morphology to an inflammatory response is not established. Given the maturation pattern of microglia morphology, the enhanced arborization in the HIV-1 Tg rats may be reflective of differences manifesting during development. Using microPET imaging for TSPO, Lee et al. (2015) reported no differences in Iba-1⁺ microglia cell density in HIV-1 Tg rats up to 9-months-of-age with no elevations in mRNA levels for pro-inflammatory cytokines in samples containing the striatum, cortex, and hippocampus. Reid et al. (2016a) reported diminished GFAP staining intensity in the striatum with no changes in microglia, similar to findings of a recent study by Sinharay et al. (2017) showing no differences in Iba-1⁺ microglia but significantly lower GFAP staining of astrocytes as early as 1 month-of-age, extending until 18 months-of-age. Thus, overall the literature does not provide a solid representation of a neuroinflammatory state in the HIV-1 Tg rat brain that would account for the neurobehavioral differences reported. In general, these studies have relied on commonly analyzed components of neuroinflammation and do not exclude the possible involvement of other inflammatory processes such as lipid mediators. For example, Blanchard et al., (2015) reported elevations in whole brain prostaglandin E2, 15-epi-lipoxinA4, and 8-isoprostane in 9-month-old HIV-1 Tg rats while Repunte-Canonigo et al., (2014) reported an elevation in hippocampal mRNA levels for prostaglandin D2 at 5 months of age. It is possible that changes associated with neuroinflammation follow an age-dependent path with greater differences observed with increasing age. When older (18 months) HIV-1 Tg rats were examined, IL-1β and IL-6 levels were elevated in the striatum (Chivero et al. 2017). The elevation in Asc (apoptosis-associated speck-like protein containing a caspase recruitment domain) protein suggested an inflammatory component and possible activation of the inflammasome as a regulatory mechanism (Chivero et al. 2017), possibly as a direct activation of microglia (Walsh et al., 2014; Mamik et al., 2017). With the advanced age it was not clear if the source of the inflammatory cytokines was related to resident microglia or the result of peripheral monocytes infiltration through a permeable blood-brain-barrier as reported in Tat-expressing transgenic mice (Leibrand et al., 2017). Thus, differences observed for neuroinflammatory factors as a function of age may reflect the progressive damage to barrier integrity and raise a concern for a contribution of peripherally-derived inflammatory factors in disease progression.

The current study provides a structural correlate to previous reports of alterations in the dopaminergic system. The data suggests that age-related loss of TH immunoreactivity in the SN contributed to earlier changes in the striatum. It remains to be determined if the subtle microglia hypertrophy in the striatum represented a response to changes in the terminal endings of the projecting TH⁺ neurons (Schier et al., 2017). The current work was conducted in male rats and, given the differential sensitivity of the female HIV-1 Tg rat to DA changes (McLaurin et al., 2017), the relationship between TH immunoreactivity in the SN and microglia changes in the striatum may occur earlier or be more pronounced in females. However, gender was not found to correlate with a decrease in DA levels observed in HIV-1 patients (Kumar et al., 2009) and the recent work by McLaurin et al. (2018) suggested that sex-related differences in motor activity and acoustic startle response were not observed across the adult ages examined in the current study. There was a difference in pre-pulse startle inhibition. Thus, we feel that examining only one sex does not diminish the characteristics described but does open the possibility for examining sex-related differences in future studies. Alternatively, responses may be associated with an accumulation of viral-related proteins that occurs in the striatum between 2 and 10 months-of-age (Peng et al., 2010). A direct cortical injection of HIV-1 Tat1–72 was associated with microglial contacts with multiple non-degenerating neuronal components within as early as 24 hrs (Marker et al., 2013). It was suggested that cell-cell contacts mediated by microglial filopodia may function in a preliminary step in the elimination of synaptic structures. While it is often expected that microglia activation is associated with an upregulation of pro-inflammatory factors, this may not necessarily occur with microglial phagocytosis that occurs in response to apoptotic cellular material. In this case, activation of such receptors that recognize phosphatidylserine (PS) stimulate an anti-inflammatory response in phagocytes (Ravichandran, 2003). The observation of a subtle microglia morphological response and lower Il1a level with no change in Tnfa in the striatum may be related synaptic phagocytic activity of microglia. Cxcr3 functions to recruit microglia to a site of injury (Rappert et al., 2004) and thus, lower levels of Cxcr3 and Icam may serve to minimize recruitment as a regulatory mechanism in this early stage of cellular insult. These findings raise questions for future studies to determine if this represents a tightly regulated response to minimize cell activation to the neuronal insult or a dysregulated microglia response. Given the findings of an age-related change in TH immunoreactivity in the SN from the current study and in the striatum (Reid et al., 2016a), additional behavioral tests and pharmacological challenges can now be incorporated into future studies to examine the long-term implications of these anatomical findings.