Cal‘MAM’ity at the Endoplasmic Reticulum-Mitochondrial Interface: A Potential Therapeutic Target for Neurodegeneration and Human Immunodeficiency Virus-Associated Neurocognitive Disorders

The endoplasmic reticulum (ER) is a multifunctional organelle and serves as the primary site for intracellular calcium storage, lipid biogenesis, protein synthesis, and quality control. Mitochondria are responsible for producing the majority of cellular energy required for cell survival and function and are integral for many metabolic and signaling processes. Mitochondria-associated ER membranes (MAMs) are direct contact sites between the ER and mitochondria that serve as platforms to coordinate fundamental cellular processes such as mitochondrial dynamics and bioenergetics, calcium and lipid homeostasis, autophagy, apoptosis, inflammation, and intracellular stress responses. Given the importance of MAM-mediated mechanisms in regulating cellular fate and function, MAMs are now known as key molecular and cellular hubs underlying disease pathology. Notably, neurons are uniquely susceptible to mitochondrial dysfunction and intracellular stress, which highlights the importance of MAMs as potential targets to manipulate MAM-associated mechanisms. However, whether altered MAM communication and connectivity are causative agents or compensatory mechanisms in disease development and progression remains elusive. Regardless, exploration is warranted to determine if MAMs are therapeutically targetable to combat neurodegeneration. Here, we review key MAM interactions and proteins both in vitro and in vivo models of Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis. We further discuss implications of MAMs in HIV-associated neurocognitive disorders (HAND), as MAMs have not yet been explored in this neuropathology. These perspectives specifically focus on mitochondrial dysfunction, calcium dysregulation and ER stress as notable MAM-mediated mechanisms underlying HAND pathology. Finally, we discuss potential targets to manipulate MAM function as a therapeutic intervention against neurodegeneration. Future investigations are warranted to better understand the interplay and therapeutic application of MAMs in glial dysfunction and neurotoxicity.

INTRODUCTION exploration before the concept of a physical ER-mitochondrial interaction gained acceptance and the term, mitochondriaassociated ER membranes (MAMs), was coined (Herrera-Cruz and Simmen, 2017). Following acceptance of this phenomenon, it was not until over 20 years later that we were able to produce the first comprehensive analysis of the MAM proteome (Herrera-Cruz and Simmen, 2017;Janikiewicz et al., 2018;Moltedo et al., 2019). Since these initial investigations, multiple mediators have been identified in regulating both the structure and function of the ER-mitochondrial interface, which are illustrated in Figure 1.
Strides have been made in our understanding of how MAMs are integral signaling platforms that regulate multiple cellular functions and maintain homeostasis. In addition to regulating the function and dynamics of both the ER and mitochondria, as discussed below, MAMs are considered central hubs for regulating key cellular processes including apoptosis, autophagy, Ca 2+ and lipid homeostasis, inflammation, and inflammasome formation Janikiewicz et al., 2018;Moltedo et al., 2019). Tether proteins between the ER and mitochondria serve as essential scaffolds in regulating MAM-mediated mechanisms whereas communication between these two organelles is primarily facilitated by Ca 2+ and redox signaling (Marchi et al., 2017;Yoboue et al., 2018;Moltedo et al., 2019).
Briefly, the transfer of Ca 2+ from the ER to mitochondria is facilitated by inositol 1,4,5-triphosphate receptors (IP 3 R) on the ER membrane, voltage-dependent anion-selective channel (VDAC) on the outer mitochondrial membrane (OMM), and cytosolic glucose-regulated protein 75 kDa (grp75), which stabilizes IP 3 R and VDAC association within the MAM interface. Notably, cytosolic DJ-1 is newly identified as a critical component in the IP 3 R-grp75-VDAC complex (Liu et al., 2019;Basso et al., 2020). On the inner mitochondrial membrane (IMM) mitochondrial Ca 2+ uniporter (MCU) enables the Ca 2+ transfer into the matrix, which increases the electrochemical potential and thus oxidative phosphorylation power van Vliet et al., 2014;Giorgi et al., 2018). On the ER membrane side of the interface, the sigma-1 receptor (σ1R) promotes of Ca 2+ transfer by associating with IP 3 R (Hayashi and Su, 2007).
Mitochondria-associated ER membranes are indispensable for ER physiology and health and are intricately involved in cellular responses to UPR/ER stress signaling. Interestingly, many studies have revealed distinct associations between the UPR sensors and MAM regulation both in response to and in the absence of ER stress, highlighting possible non-canonical functions for these proteins (Bravo et al., 2011;van Vliet et al., 2014;Saito and Imaizumi, 2018;van Vliet and Agostinis, 2018). Briefly, there are three arms to the UPR cascade: protein kinase RNAlike endoplasmic reticulum kinase (PERK), inositol-requiring protein 1α (IRE1α), and activating transcription factor 6 (ATF6). PERK is proposed as a key regulator of MAM tethering through direct interaction with MFN2 and is also linked to regulating mitochondrial dynamics and bioenergetics (Verfaillie et al., 2012;Munoz et al., 2013;Rainbolt et al., 2014;van Vliet and Agostinis, 2016;Lebeau et al., 2018;Balsa et al., 2019). IRE1α, which is commonly associated with cellular responses to infections or inflammation, is implicated in ER-mitochondrial Ca 2+ transfer, mitochondrial respiration and redox homeostasis through associations with IP 3 R (Son et al., 2014;Carreras-Sureda et al., 2019) and/or σ1R (Mori et al., 2013). ATF6 is shown to both interact with and be regulated by the key MAM tethering protein VAPB (Gkogkas et al., 2008). Moreover, ATF6 can regulate lipid biosynthesis and ER expansion suggesting a possible interplay in MAM-mediated lipid homeostasis and ER-mitochondrial physiology (Bommiasamy et al., 2009). Most recently, ATF6 has also been implicated in ER-mitochondrial Ca 2+ homeostasis (Burkewitz et al., 2020).
The ER and mitochondria contact sites are fundamental for regulating mitochondrial function, dynamics, and homeostasis. For example, Ca 2+ and redox signaling between the ER and mitochondria are essential for regulating mitochondrial integrity and bioenergetic activity. Moreover, mitochondrial dynamics are regulated by a balance between fission and fusion both of which are coordinated by the ER-mitochondrial interface. In fact, fission requires the physical maneuvering of the ER membrane to constrict around the mitochondrion fission site which then recruits the primary regulator of mitochondrial fission: dynaminrelated protein 1 (Drp1), along with fission adaptor proteins, mitochondrial fission factor and Fis1 (Friedman et al., 2011;Moltedo et al., 2019). Fusion of mitochondria require the involvement of MFN1 and MFN2 which, as aforementioned, are enriched at MAMs and regulate MAM tethering (Moltedo et al., 2019). Notably, MFN1/2 coordinated OMM fusion while optic atrophy protein 1 (OPA1) mediated IMM fusion.
Beyond fission and fusion, biogenesis and degradation are also vital for ensuring healthy mitochondria turnover. Following mitochondrial fission events, damaged mitochondria are removed via mitophagy, in which phosphatase and tensin homolog-induced putative kinase (PINK1) and beclin 1 (BECN1) at the MAM interface promote autophagosome formation (Gelmetti et al., 2017). Notably, PINK1 phosphorylates MFN2 leading to Parkin recruitment and initiation of mitophagy machinery while proteins such as p62 and microtubule-associated protein 1A/1B-light chain 3 (LC3) coordinate cargo selection and autophagosome/mitophagosome maturation (Moltedo et al., 2019;Yang et al., 2020). Meanwhile, biogenesis is a self-renewal process in which mitochondrial machinery is 'replenished' by increased transcription and translation of mitochondrial DNA (mtDNA) as well as increased synthesis, import, and assembly of nuclear DNA-encoded mitochondrial proteins (Popov, 2020). While the role of MAMs in mitochondrial biogenesis is not clear, expression of proliferator-activated receptor γ coactivator 1α (PGC1α), a master regulator of mitochondrial biogenesis, controls ERmitochondrial contact. Notably, PGC1α knockout perturbs MAM contact while overexpression promotes increased interaction (Ciron et al., 2015). Finally, mitochondrial transport is essential to ensure mitochondria are distributed to meet the energetic demands of a cell. Trafficking of mitochondria is facilitated by mitochondrial Rho GTPases (Miro), which directly interact with MFN1/2 and regulate ER-mitochondrial contact (Misko et al., 2010(Misko et al., , 2012Modi et al., 2019). Moreover, functional MFN2 is required for mitochondrial mobility (Misko et al., 2010(Misko et al., , 2012.

MITOCHONDRIA-ASSOCIATED ENDOPLASMIC RETICULUM MEMBRANES IN NEUROPATHOLOGY
Given the importance of MAM-associated mechanisms in cellular homeostasis, MAMs are gaining attention as plausible pathological platforms underlying the development and/or progression of disease (Pinton, 2018). Briefly, the role of MAMs in neuropathology, which have been previously reviewed (Raeisossadati and Ferrari, 2020;Wilson and Metzakopian, 2020;Leal and Martins, 2021), have primarily been centered on neurons, highlighting MAMs as pivotal regulators of synaptic transmission and neuronal health (Bernard-Marissal et al., 2018;Gomez-Suaga et al., 2019;Shirokova et al., 2020;Leal and Martins, 2021). Astrocyte MAMs are largely a new exploration. Notably, enrichment of MAMs in astrocyte endfeet may be crucial for regulating the blood-brain interface and brain healing following injury (Bergami and Motori, 2020;Gbel et al., 2020). Moreover, knockdown of MAM-associated proteins, PACS2 or σ1R, induces degeneration of hippocampal neurons and astrocytes, supporting the importance of MAMs for neural cell survival (Hedskog et al., 2013). The presence and function of MAMs in other neural cells, microglia and oligodendrocytes, remain unconfirmed (Bernard-Marissal et al., 2018;Shirokova et al., 2020).
Many of the cellular processes that are implicated in the etiology of neurodegenerative diseases are coordinated at the ER-mitochondrial interface such as dysregulated lipid and Ca 2+ homeostasis, mitochondrial dysfunction, ER and oxidative stress, impaired autophagy, and inflammation. Neurons are particularly susceptible to mitochondrial dysfunction given the high energetic demand of brain tissue (Grimm and Eckert, 2017;Picca et al., 2020). Moreover, ER stress is induced by the aggregation of misfolded proteins, which is a physical hallmark of most neurodegenerative pathologies. Formation of protein aggregates can be a consequence of increased production of misfolded proteins (ER dysfunction), the impaired removal of dysfunctional proteins (impaired autophagy), or both (Sweeney et al., 2017;Monaco and Fraldi, 2020). Regardless, ER stress and mitochondrial dysfunction are increasingly being proposed as key therapeutic targets for combating neuropathology, with MAMs arising as the essential crossroad for this collaboration. It is worth considering whether these central hubs can be manipulated to reconcile cellular dysfunction and degeneration and to restore CNS homeostasis.
Beyond the classic hallmarks of neuropathology in the context of known MAM-associated functions, dysregulation of MAM tethering and activity are implicated in a number of neurodegenerative diseases such as Alzheimer's disease (AD), Parkinson's disease (PD), and amyotrophic lateral sclerosis (ALS). However, whether these MAM alterations arise as intended remedial responses or are the causative agents in these disease pathologies remains to be determined. The distinct interplay of MAMs in these neuropathological conditions is illustrated in Figure 2 and discussed below.

Alzheimer's Disease
As one of the most common forms of dementia, Alzheimer's disease is characterized by the aggregation of extracellular amyloid β (Aβ) peptides and intracellular hyperphosphorylated tau proteins, termed plaques and tangles, respectively. Treatment with Aβ in primary hippocampal neurons increased expression of MAM-associated PACS2 and σ1R proteins, as well as direct ER-mitochondrial contact via the IP 3 R-VDAC bridging complex. Expression of PACS2 and σ1R was also increased in the hippocampus, cortex, and cerebellum brain regions of a potent amyloid precursor protein (APP) mutant AD mouse model. However, in human AD postmortem cortical tissues, PACS2 increased, but σ1R expression decreased (Hedskog et al., 2013). Interestingly, MAMs serve as a key production site for Aβ peptides (Schreiner et al., 2015). The formation of Aβ plaques occurs following the catalytic processing of APP by the γ-secretase complex. Two major components of this complex, presenilin (PS) 1 and PS2, localize to the ER-mitochondria interface and modulate MAM functions, specifically lipid and Ca 2+ homeostasis (Area-Gomez et al., 2009;Zampese et al., 2011;Area-Gomez et al., 2012;Galla et al., 2020). Notably, overexpressing PS2 increases both ER-mitochondria contact and Ca 2+ transfer from the ER to mitochondria (Zampese et al., 2011;Galla et al., 2020). This phenomenon appears to be mediated through interactions between PS2 and MFN2 . Furthermore, silencing MFN2 in human embryonic kidney (HEK) 293 cells stably expressing an APP mutant to overproduce Aβ, increased MAM contact and Ca 2+ transfer, which impairs γ-secretase maturation and activity, ultimately decreasing Aβ production (Leal et al., 2016). Leal et al. also recently reported decreased expression of MFN1/2 in postmortem human AD brain tissues. They further confirmed a positive correlation between Aβ and MAM contact in multiple AD mouse models and in vitro neuron cultures exposed to Aβ. The increased connectivity between ER and mitochondria in response to Aβ FIGURE 2 | MAMs in AD, PD, and ALS. While dysregulated MAM-associated mechanisms are hallmarks of neurodegenerative pathology, the unique pathological features of AD, PD, and ALS alter MAM contact and communication emphasizing MAMs as pivotal players in neuropathology. In AD, MAM contact, Ca 2+ transfer, and MAM-mediated mechanisms (i.e., autophagy) are increased among in vitro and mouse models, yet decreased MAM-mediator expression and interaction was found in human AD cortical tissues. In PD, both increased and decreased MAM contact and communication were reported. In ALS, MAM tethering and Ca 2+ transfer were significantly impaired. MAMs, mitochondria-associated ER membranes; Mito, mitochondria; ER, endoplasmic reticulum; AD, Alzheimer's disease; Aβ, amyloid β; APP, amyloid precursor protein; ApoE4, ε4 allele of apolipoprotein E; PS, presenilin; MFN, mitofusin; PACS2, phosphofurin acidic cluster sorting 2; Ca 2+ , calcium; IP 3 R, inositol 1,4,5-triphosphate receptors; VDAC, voltage-dependent anion-selective channel; grp75, glucose-regulated protein 75 kDa; MCU, mitochondrial Ca 2+ uniporter; σ1R, sigma-1 receptor; ETC, electron transport chain; PD, Parkinson's disease; α-syn, α-synuclein; VAPB, vesicle-associated membrane protein-associated protein B; PTPIP51, protein tyrosine phosphatase-interacting protein 51; PINK1, phosphatase and tensin homolog-induced putative kinase; ALS, amyotrophic lateral sclerosis; TDP-43, transactive response (TAR) DNA-binding protein 43; GSK-3β, glycogen synthase kinase-3β; and FUS, fused in sarcoma. Image created with BioRender.com.
subsequently increased mitochondrial metabolic function and autophagosome formation, likely to promote removal of Aβ aggregates (Leal et al., 2020). Of note, another recent examination of human AD cortical tissues showed a decreased expression and interaction of MAM tether proteins, VAPB and PTPIP51, in addition to decreased IP 3 R expression. The decreased expression of these proteins also correlated with increased disease stage severity (Lau et al., 2020). Thus, decreased expression of MAMmediators in human AD tissues was reported across three separate studies (Hedskog et al., 2013;Lau et al., 2020;Leal et al., 2020) suggesting impaired ER-mitochondrial tethering, highlighting unique differences between animal models and human tissues. Regardless, these findings strongly implicate MAMs as possible targets to combat AD/Aβ pathology.
Disease-associated tau protein alters mitochondrial transport, dynamics, bioenergetics and degradation (reviewed by Szabo et al., 2020). These findings further discuss the implications of MAMs in tau pathology as only two studies have so far explored this phenomenon; both of these models identified increased ERmitochondrial contact during tau pathology (Perreault et al., 2009;Cieri et al., 2018). The ε4 allele of apolipoprotein E (ApoE4) significantly associates with an increased risk for sporadic AD. Treatment of astrocyte conditioned media cultured from ApoE4 knock-in mice, upregulated MAM activity in human fibroblasts and mice neurons, as measured by the synthesis of phospholipids and cholesteryl esters (Tambini et al., 2016). This outcome was averted when repeated in mouse embryonic fibroblasts (MEF) harboring MFN2 knockout, which in this model, decreased MAM tethering (de Brito and Scorrano, 2008;Tambini et al., 2016).

Parkinson's Disease
Parkinson's disease is a neurodegenerative disease manifesting with primarily motor deficits due to damaged and degenerative dopaminergic neurons in the substantia nigra. A key pathophysiology underlying PD is the presence of 'Lewy bodies' comprised of α-synuclein (α-syn) aggregates. However, tau pathology is also present in some cases. PD pathology is characterized by a number of MAM-associated cellular processes including impaired autophagy, Ca 2+ homeostasis, lipid metabolism, ER stress, and mitochondrial dynamics (reviewed by Rodriguez-Arribas et al., 2017;Gomez-Suaga et al., 2018). Notably, overexpression or silencing of α-syn confirms regulation on mitochondrial dynamics, ER-mitochondrial coupling and Ca 2+ transfer (Cali et al., 2012;Guardia-Laguarta et al., 2014). However, these findings are also inconsistent as one demonstrates increased organelle contact while the other reports disruption in tethering. From a more recent report, α-syn binds to VAPB on the ER membrane, disrupting interaction with PTPIP51 on the OMM and subsequently, MAM tethering, Ca 2+ transfer, and ATP production . Thus, the effects of α-syn on MAM function are linked to impaired bioenergetic activity.
Another MAM connection to the neuropathology of PD involves PINK1 and Parkin. Mutations in PINK1 and Parkin are key risk factors for the development of PD. The functions of these proteins are critical in maintaining mitochondrial health by regulating mitochondrial biogenesis, degradation, dynamics, function, and transport. PINK1 and Parkin localize to MAMs, which is not surprising given their prominent roles in mitophagy and MAM's being the site for mitochondrial fission (Gomez-Suaga et al., 2018). Moreover, Parkin is implicated as a possible regulator for MAM tethering although controverting evidence has been reported as to whether it is a positive or negative regulator Gautier et al., 2016). Briefly, overexpression of Parkin in both HeLa and a neuroblastoma cell line enhanced MAM coupling and Ca 2+ transfer to increase ATP production . Expression of Parkin is encoded by the PARK2 gene. Increased mitochondrial-ER contact was found in fibroblasts from a PARK2 knockout mouse and PD patients with PARK2 mutation (Gautier et al., 2016). Moreover, mutations in the gene for DJ-1, which as mentioned above has been identified as a novel component in the IP 3 R-grp75-VDAC complex, are causative for autosomal recessive familial PD. The DJ-1 protein has been identified to have a protective role in oxidative stress and modulates mitochondrial morphology. Similar to a report on α-syn, overexpression of DJ-1 enhances ER-mitochondrial coupling and Ca 2+ transfer while silencing of DJ-1 impairs mitochondrial Ca 2+ flux and induces fragmentation . All these findings highlight the importance of MAMs as regulators of mitochondrial function and physiology in the context of PD pathology.

Amyotrophic Lateral Sclerosis
A key component of ALS pathology is deposits of transactive response (TAR) DNA-binding protein 43 (TDP-43). Like α-syn, TDP-43 alters MAM tethering and Ca 2+ homeostasis by interrupting the relationship between VAPB and PTPIP51. However, the mechanism of disruption is instead mediated through activation of glycogen synthase kinase-3β (GSK-3β) (Stoica et al., 2014). Interestingly, accumulation of fused in sarcoma (FUS), a key pathological feature characterizing ALS, also interrupts VAPB and PTPIP51 association and MAM tethering via GSK-3β activation . This interruption was accompanied by compromised mitochondrial Ca 2+ uptake and ATP production. Thus, ALS pathology is characterized by two factors that activate GSK-3β to disentangle the MAM interface; however, the target of GSK-3β to facilitate this interference remains unknown. The significance of MAM tethering in ALS pathology is further supported as a P56S mutation in VAPB is causative of ALS (Nishimura et al., 2004). Moreover, mutation or perturbed function of another MAM protein, σ1R, which is essential for Ca 2+ transfer from the ER to mitochondria, is a causal agent to ALS pathology and motor neuron degeneration (Al-Saif et al., 2011;Bernard-Marissal et al., 2015). Additional implications and associations of MAMs in ALS pathology are previously reviewed (Manfredi and Kawamata, 2016;Lau et al., 2018).

HUMAN IMMUNODEFICIENCY VIRUS-ASSOCIATED NEUROCOGNITIVE DISORDERS
Human immunodeficiency virus type 1 (HIV-1) can invade the CNS early during infection and infect residential glial cells (astroglia, microglia, and oligodendrocytes) where infection can persist for life. Even with the medical advancement of antiretroviral therapy (ART), low viral replication, chronic neuroinflammation, glial dysfunction, and HIV-1 protein toxicity contribute to the development of a spectrum of HIV-associated neurocognitive disorders (HAND). In fact, HAND continue to afflict approximately 30-70% of people living with HIV, depending on cohort demographics (Simioni et al., 2010;Heaton et al., 2011). HAND are characterized by different levels of cognitive impairments and interference with one's daily functioning. At the extreme end of the spectrum, symptoms clinically manifest as dementia (Mackiewicz et al., 2019).
The pathology of HIV-1 infection often includes "accelerated aging." Thus, HIV-1 patients are more prone to developing early onset symptoms for a number of age-related diseases including the neurodegenerative pathologies discussed above (Robertson et al., 2007;Cody and Vance, 2016;Wang et al., 2017;Mackiewicz et al., 2019). In fact, Aβ plaques as well as tauopathy can both associate with and be exacerbated by HAND pathology (Kim et al., 2013;Brown et al., 2014;Hategan et al., 2019). Of note, the primary HIV-1 proteins implicated in neurotoxicity are transactivator of transcription (Tat), glycoprotein (gp)120, viral protein R (Vpr), and negative factor (Nef). Moreover, ART drugs are now identified as key contributors to the cellular senescence and accelerated aging underlying HIV/HAND pathology, which was reviewed previously with an emphasis on mitochondrial dysfunction (Schank et al., 2021). Khan et al. (2019) recently published insights into potential inter-organelle collaborations in HIV/HAND pathogenesis. Despite obvious disruptions in ER and mitochondrial homeostasis in HAND pathology, there remains limited investigations and/or considerations of MAMs (Ma et al., 2016;Nooka and Ghorpade, 2017;Khan et al., 2019). Below we discuss the implications of MAMs in HAND pathology by reviewing the effects of HIV-1 on ER stress, Ca 2+ dysregulation and mitochondrial dysfunction, which is also illustrated in Figure 3. Together, these findings strongly support altered MAM signaling as a prominent contributor to HAND pathology.

Human Immunodeficiency Virus-1 and Endoplasmic Reticulum Stress
As aforementioned, ER stress is a key characteristic in neurodegenerative pathologies, and MAMs are intricately involved in cellular responses to ER stress. Indeed, the three UPR arms are now increasingly considered integral mediators within the MAM proteome (van Vliet andAgostinis, 2016, 2018;Saito and Imaizumi, 2018;Malli and Graier, 2019). In the context of HAND, significant increases in ER stress markers are detected in the frontal cortex of brain tissues from HIV-1 positive individuals. In fact, the levels of ER stress positively correlate with cognitive decline, wherein the most severe cases of HAND have the highest expression of ER stress markers (Lindl et al., 2007;Akay et al., 2012). Interestingly, these studies show that while ER stress is evident in both neurons and astrocytes, astrocytes appear to have higher expression of ER stress markers. These findings do not necessarily mean neurons are less likely to endure ER stress during HIV-1 infection. Rather, astrocytes may be more resilient to prolonged ER stress, and neurons could be more susceptible to succumb to apoptosis.
It is well-known that during neurodegenerative or neuroinflammatory conditions, astrocytes become 'activated' , shifting their function from neurotrophic to a more neurotoxic phenotype. Astrocyte-mediated neurotoxicity has arisen FIGURE 3 | MAMs in HAND. HIV-1 in the CNS leads to glial activation and neurotoxicity, which are key mechanisms mediating HAND pathology. HIV-induced UPR/ER stress, Ca 2+ dysregulation, and mitochondrial dysfunction strongly implicate MAMs as potential regulators of glial activation and neurotoxicity. In microglia, the initiation of mitophagy is essential to buffer HIV-induced mitochondrial dysfunction but accumulation of damaged mitochondria induces a heightened inflammatory response which is associated with impaired mitochondrial integrity and a significant decrease in bioenergetic capacity. Conversely, astrocytes increase their Ca 2+ signaling and metabolic capacity. Astrocyte shift in function leads to increased release of neurotoxic factors and well as impaired provision of essential nutrients to neurons. Similar to microglia, a balance between mitophagy and inflammasome activation appears to be a critical determinate to astrocyte fate. Notably, ER-mitochondrial Ca 2+ transfer and UPR/ER stress also arise as potential pivotal players in astrocyte-mediated neurotoxicity. While astrocytes and microglia teeter between toxic and tropic, neurons are highly vulnerable to HIV-induced mitochondrial dysfunction caused by both direct and indirect insults. Targeting MAMs to combat mitochondrial dysfunction is of heightened importance to enhance CNS fitness against neuropathologic challenges, including HIV/HAND. MAMs, mitochondria-associated ER membranes; HIV-1, human immunodeficiency virus type 1; HAND, HIV-associated neurocognitive disorders; ER, endoplasmic reticulum; UPR, unfolded protein response; PERK, protein kinase RNA-like endoplasmic reticulum kinase; IRE1, inositol-requiring protein 1α; ATF6, activating transcription factor 6; Ca 2+ , calcium; IP 3 R, inositol 1,4,5-triphosphate receptors; VDAC, voltage-dependent anion-selective channel; grp75, glucose-regulated protein 75 kDa; MCU, mitochondrial Ca 2+ uniporter; RyR, ryanodine receptors; MFN, mitofusin; Drp1, dynamin-related protein 1; mPTP, mitochondrial permeability transitional pore; ROS, reactive oxygen species; ETC, electron transport chain; PGC1α, proliferator-activated receptor γ coactivator 1α; TFAM, transcription factor A. Image created with BioRender.com. as a key feature in HAND pathology with UPR/ER stress as a potential regulator (Fan and He, 2016;Shah et al., 2016;Ghorpade, 2017, 2018;Natarajaseenivasan et al., 2018). In fact, while HIV-1-induced ER stress has not yet been extensively investigated, most investigations are primarily centered around astrocytes. Notably, expression of HIV-1 Tat in astrocytes induces aggregation of glial fibrillary acid protein (GFAP). Both Tat expression and GFAP aggregation activate the three UPR pathways (Fan and He, 2016). HIV-1 gp120 specifically activates the IRE1α branch of UPR signaling in an astrocyte cell line, primary human astrocytes, and astrocyte-restricted gp120 transgenic mice, which is linked to the initiation of apoptotic signaling (Shah et al., 2016). Inhibition of ER stress or UPR signaling in both of these studies reverses astrocyte-mediated neurotoxicity and apoptotic signaling (Fan and He, 2016;Shah et al., 2016). Moreover, studies from our lab demonstrate induction of the three UPR pathways in primary human astrocytes in response to HAND-relevant stimuli (whole HIV-1, inflammation, and ART drugs), which associate with mitochondrial depolarization (Nooka and Ghorpade, 2017). HIV-1 Tat-induced ER stress in a CD4+ T cell line (Campestrini et al., 2018) neurons (Norman et al., 2008) and human brain microvascular endothelial cells (HBMEC) (Ma et al., 2016) are also accompanied by changes in mitochondrial function and apoptotic signaling supporting cooperation of the ER and mitochondria during HIV/HAND pathology. Indeed, inhibition of ER stress reversed HIV-1-induced mitochondrial dysfunction and increased cell viability in HBMEC (Ma et al., 2016). These findings not only implicate MAMs in HAND pathology but also emphasize UPR/ER stress signaling as an important regulator of mitochondrial function, cellular fate and astrocyte-mediated neurotoxicity.

Human Immunodeficiency Virus-1 and Ca 2+ Dysregulation
As discussed above, the ER is the primary site for intracellular Ca 2+ storage, and MAMs are hubs for regulating Ca 2+ homeostasis. HIV-1 modulates Ca 2+ signaling and perturbs Ca 2+ homeostasis. Many of these studies have been previously reviewed or discussed (Haughey and Mattson, 2002;Hu, 2016;Fields and Ellis, 2019). Altered Ca 2+ dynamics during HIV-1 exposure is directly linked to apoptotic signaling, UPR induction, and impaired mitochondrial integrity, energetics, and/or quality control thus implicating Ca 2+ dysregulation as a key component in HIV-1-induced neurotoxicity and astrogliosis. One of the earliest studies identifying this relationship was in rat hippocampal neurons exposed to HIV-1 Tat, which induced apoptosis through the elevation of cytoplasmic Ca 2+ and increased mitochondrial Ca 2+ uptake. Chelating cytosolic Ca 2+ or pharmacological inhibiting mitochondrial Ca 2+ uptake via MCU protected neurons from HIV-1 Tat-mediated neurotoxicity (Kruman et al., 1998). Interestingly, a later study in rat cortical neurons linked ER and mitochondrial Ca 2+ loss via ryanodine receptors (RyR) as an upstream mechanism for HIV-1 Tat-mediated UPR induction and mitochondrial hyperpolarization (Norman et al., 2008). While less prominent, Ca 2+ mobilization is also implicated in HIV-1 gp120- (Meeker et al., 2016) and Vpr- (Jones et al., 2007) mediated neurotoxicity, which associated with impaired axonal mitochondria transport and apoptotic signaling, respectively.
In the context of astrocytes, our previous studies indicate a pivotal role of Ca 2+ signaling in regulating ER stress and mitochondrial dysfunction induced by HAND-relevant stimuli (whole HIV-1, inflammation, and ART drugs). Notably, chelation of cytoplasmic Ca 2+ was able to reverse HIV-1-induced ER stress and mitochondrial depolarization (Nooka and Ghorpade, 2017). Moreover, knockdown of MCU to reduce mitochondrial Ca 2+ uptake during exposure to HIV-1 Tat and/or cocaine reverses astrocyte mitochondrial dysfunction and metabolic switching to restore a neuroprotective phenotype (Natarajaseenivasan et al., 2018). Manipulation of VDAC1 (the OMM Ca 2+ channel within the MAM interface) or "mortalin" (aka grp75, the scaffolding protein between IP 3 R on the ER membrane and VDAC on the mitochondria) is also able to rescue neurons from HIV-1 Tat-induced astrocyte-mediated neurotoxicity (Fatima et al., 2017;Priyanka et al., 2020). Specifically, HIV-1 Tat expressing primary human astrocytes trigger neuronal death by excessive ATP release, a mechanism that was counteracted by repression of VDAC1 (Fatima et al., 2017). Neuroinflammation and glutamate excitotoxicity are additional mechanisms for which astrocytes inflict neuronal damage during HAND, as previously reviewed by our research team (Cisneros and Ghorpade, 2012;Borgmann and Ghorpade, 2015). Overexpression of mortalin/grp75 protects neurons from astrocyte-mediated neurotoxicity by reversing HIV-1 Tat-induced astrocyte mitochondrial dysfunction and fragmentation while also reducing the release of excess ATP, inflammatory cytokines, and extracellular glutamate (Priyanka et al., 2020). Altogether, these findings strongly support MAMmediated Ca 2+ transfer as a pivotal regulator of astrocytemediated neurotoxicity during HAND pathogenesis.
It is also noteworthy to mention, that while research with HIV-1 Nef is less investigated in HAND pathogenesis, it is known to play a prominent role for Ca 2+ dysregulation in T cells (Manninen and Saksela, 2002;Shelton et al., 2012;Silva et al., 2016). In fact, HIV-1 Nef directly interacts with both mortalin/grp75 (Shelton et al., 2012) and IP 3 R (Manninen and Saksela, 2002), which are key mediators in ER to mitochondria Ca 2+ transfer. Thus, there is likely an interplay between HIV-1 Nef and the MAM interface. More studies are needed to determine the role of HIV-1 Nef and Ca 2+ dysregulation in neural cells.

Human Immunodeficiency Virus-1 and Mitochondrial Dysfunction
The MAMs are essential for regulating mitochondrial function and homeostasis. Defects in mitochondrial bioenergetics, biogenesis, degradation, dynamics, integrity and transport are all present in the pathology of HAND and are discussed below. The role of mitochondrial dysfunction in HIV/HAND pathology is rapidly gaining attention with two recent reviews by Fields and Ellis (2019) and Schank et al. (2021); however, neither discuss the potential interplay of MAMs. It is well established that HIV-1 can induce neuronal apoptosis, in which mitochondria [and MAMs] are known upstream regulators through Ca 2+ and redox signaling (Kruman et al., 1998;van Vliet et al., 2014;Marchi et al., 2017). In addition to regulating apoptosis, Ca 2+ and redox signaling between mitochondria and ER are also essential for regulating mitochondrial integrity and bioenergetic activity, autophagy and inflammasome activation (van Vliet et al., 2014), again highlighting the importance of ER-mitochondrial inter-organelle collaboration in negotiating cellular fate.
In human neurons, exposure to HIV-1 Tat and Vpr decreased VDAC (OMM Ca 2+ transporter) protein expression, dysregulated RNA expression of several genes regulating mitochondrial metabolism and decreased ATP levels (Darbinian et al., 2020). These findings were also associated with mtDNA damage, reactive oxygen species (ROS) accumulation, increased expression apoptotic proteins, and release of cytochrome C, emphasizing the relationship of mitochondrial bioenergetics and integrity. Less clear are the effects of HIV-1 on neuronal mitochondrial dynamics, which vary greatly across investigations. Two separate studies looking at the brain tissues from HIV + donors with or without HIV encephalitis, specifically centered on neurons, report enlarged mitochondria as a unique pathological feature of HAND (Avdoshina et al., 2016;Fields et al., 2016).
Changes in mitochondrial morphology associated with increased expression of fusion proteins (MFN1 and OPA1) and decreased expression and/or activation of mitochondrial fission proteins (Fis1 and Drp1). Interestingly, as previously discussed with ER stress markers, decreased expression of mitochondrial Drp1 also coincided with increased HIVassociated neurocognitive decline . Similar disruptions on mitochondrial morphology were confirmed in gp120 transgenic mice and by gp120 treatment to neuronal cell cultures (Avdoshina et al., 2016;Fields et al., 2016). These reports also found increased extracellular acidification rate indicating increased glycolysis  and reduced oxygen consumption rate (OCR) (Avdoshina et al., 2016) suggesting a deviation in mitochondrial bioenergetics. Noteworthy, overexpressing Drp1 reverses gp120meditated neuronal mitochondrial dysfunction reducing both neuroinflammation and neurodegeneration . These findings suggest mitochondrial fission as a potential therapeutic mechanism to combat HIV-mediated mitochondrial dysfunction in neurons.
On the contrary, more recent studies by Teodorof-Diedrich and Spector (2018) looking at the effects of both HIV-1 gp120 and Tat on mitochondrial dynamics in human neurons and Rozzi et al. (2018) looking at the effects of HIV-1 Tat in rat cortical neurons, instead identified a proclivity toward mitochondrial fission and fragmentation. In fact, increased Drp1 expression, activity, and/or translocation were critical for these outcomes. Both studies also reported impaired mitochondrial integrity as demonstrated by decreased membrane potential ( m). Noteworthy, Teodorof-Diedrich and Spector (2018) also found an increased recruitment of mitophagy proteins (Parkin, p62, and LC3) and mitophagosome formation, yet an impaired mitophagic flux leading to accumulation of damaged mitochondria due to incomplete mitophagy. The discrepancy regarding HIV-1 gp120 on neuronal mitochondrial dynamics may be attributed to different in vitro models as the prior two studies were in rodents (Avdoshina et al., 2016;Fields et al., 2016) and the later in primary human neurons (Teodorof-Diedrich and Spector, 2018). Moreover, there may be general discrepancies between the effects of HIV-1 gp120 versus HIV-1 Tat. Regardless, the identical findings of elongated mitochondria in neurons of HIV+ brain tissues remain a significant find.
However, there is one similarity across these four studies in regard to mitochondrial distribution/localization. There appears to be an overall decrease in mitochondrial trafficking throughout the processes with increased aggregation of mitochondria near the soma. Impaired mitochondria axonal transport is also evident in neurons exposed to HIV-1 Vpr, which was further associated with decreased ATP production and increased expression of senescent markers (Wang et al., 2017). As mitochondria are essential for synaptic maintenance and to meet the energetic demands for neurotransmission, disrupted axonal mitochondria transport arises as a potential key pathological feature in HAND and age-related axonal degeneration. Interestingly, although HIV-1 Vpr is primarily known as a nuclear protein, it also directly interacts with adenine nucleotide translocase (ANT) on the OMM, which is a key regulator of mitochondrial integrity via formation of the mitochondrial permeability transitional pore Wang et al., 2017;Cowan et al., 2019). In fact, inhibiting the interaction between HIV-1 Vpr and ANT was able to reverse deficits in mitochondrial trafficking (Wang et al., 2017). Finally, this study also suggests HIV-1 Vpr may disrupt mitochondrial biogenesis in neurons via reduction of PGC1α expression. Swinton et al. (2019) confirmed decreased expression of PGC1α in HIV+ brain cortical tissues as well as decreased transcription factor A (TFAM), another regulator of mitochondrial biogenesis (Swinton et al., 2019).
In astrocytes, both HIV-1 infection and external HIV-1 exposure perturb mitochondrial integrity (Nooka and Ghorpade, 2017;Natarajaseenivasan et al., 2018;Ojeda et al., 2018;Priyanka et al., 2020). Unlike neurons, astrocytes increase their metabolic activity and ATP production (Natarajaseenivasan et al., 2018;Swinton et al., 2019). Moreover, while neurons have decreased TFAM expression in HIV+ brain tissues, astrocytes have increased TFAM expression (Swinton et al., 2019). The increased metabolic and mitochondrial biosynthetic profile may be a key feature underlying astrocyte activation and astrocytemediated neurotoxicity during HAND. For example, HIV-1 Tat provokes astrocytes to undergo a distinct metabolic shift from glucose to fatty acid oxidation, which restricts astrocyte provision of lactate to neurons (Natarajaseenivasan et al., 2018). In addition to decreasing release of neurotrophic factors when 'activated' by HIV-1 Tat, astrocytes increase their release of neurotoxic factors including excessive ATP (Fatima et al., 2017;Priyanka et al., 2020), ROS (Natarajaseenivasan et al., 2018;Priyanka et al., 2020), and inflammatory cytokines (Natarajaseenivasan et al., 2018). As briefly discussed in the previous section, blocking ER-mitochondrial Ca 2+ transfer via VDAC (on the OMM) (Fatima et al., 2017), MCU (on the IMM) (Natarajaseenivasan et al., 2018), or mortalin/grp75 (scaffold between IP 3 R and VDAC) (Priyanka et al., 2020) restored an astrocyte neurotrophic phenotype, highlighting a probable MAM interplay in astrocyte-mediated neurotoxicity during HAND. Moreover, HIV-1 infection (Ojeda et al., 2018) or HIV-1 Tat expression (Priyanka et al., 2020) in astrocytes induced mitochondrial fragmentation, which may be followed by defective mitophagy, similar to what was seen by Teodorof-Diedrich and Spector (2018) in neurons. Notably, accumulation of damaged mitochondria promoted inflammasome activation, which was subsequently followed by cell death (Ojeda et al., 2018). However, astrocytes that had successful mitophagy, were able to attenuate mitochondrial dysfunction and resist cell death. The divergence between these fates appeared dependent on the mode of infection, where productively infected astrocytes favored survival, and non-productive infection succumbed to inflammasome-mediated cell death.
Mitochondrial dysfunction in microglia has been less studied; however, there is a recent review discussing the important role of microglia in HAND (Borrajo et al., 2021). Similar to astrocytes, there appears to be a delicate balance between mitophagy and inflammasome formation underlying microglia activation during HAND (Thangaraj et al., 2018;Rawat et al., 2019). In response to HIV-1 Tat (Thangaraj et al., 2018) or ssRNA (Rawat et al., 2019), there is increased expression of autophagy/mitophagy proteins (PINK1, Parkin, p62, LC3, and BECN1) with a subsequent blockade of mitophagy flux, leading to the accumulation of mitophagosomes and damaged mitochondria. These changes are associated with increased ROS generation and impaired mitochondrial integrity (Rawat et al., 2019). However, unlike astrocytes, there is a significant decrease in mitochondrial bioenergetic activity (Thangaraj et al., 2018). Importantly, defects in mitophagy appear to be central to microglial activation and inflammasome formation (Thangaraj et al., 2018;Rawat et al., 2019). Indeed, HIV-1 gp120 induces inflammasome activation in microglia in vitro and in vivo, whereas inhibiting inflammasome activation, alleviates microglia-mediated neurotoxicity, promotes neuronal regeneration, and improves cognitive function (He et al., 2020).
In summary, HIV-induced changes in mitochondrial bioenergetics, dynamics, degradation, integrity, and transport are MAM-regulated processes, implicating MAMs in HIV/HAND pathology. Noteworthy, mitochondrial fission events are key consequences of Tat-mediated toxicity in neurons (Rozzi et al., 2018;Teodorof-Diedrich and Spector, 2018), astrocytes (Ojeda et al., 2018;Priyanka et al., 2020), and microglia (Thangaraj et al., 2018;Rawat et al., 2019). Further, defective mitophagy is a crucial mechanism underlying HAND pathology. The accumulation of damaged mitochondria promotes inflammasome activation in astrocytes and microglia. Thus, the balance between mitophagy and inflammasome activation is a critical determinant of glial fate during HIV-1 toxicity (Ojeda et al., 2018;Rawat et al., 2019). As MAMs are the site for both mitophagy and inflammasome initiation, ER-mitochondrial cooperation is likely essential for negotiating these cellular outcomes. Finally, evidence of MAMs in HAND is further supported by a previous study in T cells showing that HIV-1 Vpr localizes to both the ER and mitochondria, and MAMs serve as a possible route for intracellular trafficking of Vpr . Notably, exposure to HIV-1 Vpr decreased MFN2 and Drp1 expression, impaired ER-mitochondrial interaction and morphology, and induced mitochondrial depolarization and deformation. Overexpressing MFN2 or Drp1 was able to prevent T cell mitochondrial depolarization and deformation. Additional studies in neural cells are needed to determine the role of MAMs in HAND. Moreover, HIV-1 Tat and Vpr are historically considered as nuclear proteins; however, Tat and Vpr toxicity on ER and mitochondrial homeostasis emphasizes the need to expand our classical understanding of host-viral interactions during HIV-1 infection.

POTENTIAL THERAPEUTIC TARGETS
Dysfunctional MAM-mediated mechanisms are hallmarks of neurodegenerative pathologies including AD, PD, ALS, and HAND. Notably, ER-mitochondrial contact and communication are critical in regulating mitochondrial function and health.
Throughout this review, we have specifically highlighted Ca 2+ , UPR/ER stress, mitochondrial fission/fusion, mitophagy and inflammasome signaling pathways as potential targets to combat mitochondrial dysfunction in neural cells during HIV/HAND, whereas MAMs serve as the central therapeutic platform. Table 1 summarizes the potential therapeutic targets discussed throughout this manuscript.
Coupling between the ER and mitochondria can increase in response to stress to tailor to the functional demands of the cell (Bravo-Sagua et al., 2016). However, whether increased/decreased contact contributes to cellular dysfunction or improves cell outcomes, remains largely unknown. Notably, forcing increased ER-mitochondrial contact in Drosophila was able to extend lifespan and improve overall motor function (Garrido-Maraver et al., 2020). As a MAM tether protein, MFN2 may serve as a potential target to manipulate ERmitochondria contact and communication. Overexpression of MFN2 in CD4+ T cells restored mitochondrial integrity and increased cell viability against HIV-1 Vpr toxicity . However, the effects of MFN2 manipulation vary across cell types. For example, knockdown of MFN2 decreases MAM contact and communication in MEF (de Brito and Scorrano, 2008) but increases MAM contact and communication in an APP mutant in HEK 293 cells (Leal et al., 2016). It should be noted that increased MAM tethering following MFN2 knockdown decreased Aβ production and improved cellular outcome.
Manipulating MAM tethering can subsequently affect MAMmediated mechanisms, such as ER to mitochondrial Ca 2+ transfer. However, with these studies, it becomes difficult to delineate the primary therapeutic mechanism, tethering versus Ca 2+ transfer. Conversely, studies that directly target MAM-mediated Ca 2+ transfer provide a clear demonstration of the therapeutic potential of targeting these pathways. For example, in astrocytes exposed to HIV-1, blocking mitochondrial Ca 2+ uptake by targeting MCU (Natarajaseenivasan et al., 2018), VDAC (Fatima et al., 2017), or cytosolic Ca 2+ (Nooka and Ghorpade, 2017) can prevent astrocyte mitochondrial dysfunction and reverse astrocyte-mediated neurotoxicity. The same outcome was achieved by overexpression of mortalin/grp75 (Priyanka et al., 2020). However, this was due to a direct interaction of mortalin/grp75 with HIV-1 Tat leading to Tat degradation rather than modulating ER-mitochondrial Ca 2+ transfer. This is an important discrepancy as mortalin/grp75 manipulation in neurons can protect neurons from oxidative cell death when repressed or increase susceptibility when overexpressed (Honrath et al., 2017). Similarly, blocking MCU or cytosolic Ca 2+ in neurons improved neuronal survival against HIV-1 Tat toxicity (Kruman et al., 1998), while blocking ER Ca 2+ release via RyR can also attenuate Tat-induced UPR induction and mitochondrial dysfunction (Norman et al., 2008). The potential therapeutic applications of targeting MCU and σ1R to regulate ER-mitochondrial Ca 2+ transfer to combat neurodegenerative pathologies have also been previously reviewed (Liao et al., 2017;Weng et al., 2017).
Other potential upstream regulators of mitochondrial dysfunction underlying neuropathology are the three UPR/ER stress pathways. In fact, targeting UPR pathways as potential neurodegenerative therapies are previously reviewed (Halliday and Mallucci, 2014;Remondelli and Renna, 2017;Martinez et al., 2019); however, these reports do not focus on UPR/ER stress in the context of mitochondrial dysfunction, regardless of the obvious crosstalk between these two organelles. In the context of HIV/HAND, blocking UPR/ER stress in astrocytes reverses astrocyte-mediated neurotoxicity and apoptotic signaling (Fan and He, 2016;Shah et al., 2016). Similarly, inhibiting UPR/ER stress in HBMECs restores mitochondrial integrity and increases cell viability during HIV-1 Tat toxicity (Ma et al., 2016). It is important to note that there are unique signaling pathways amongst the three arms in addition to their 'non-canonical' functions that remain ill-defined. However, given the potential regulation of the UPR/ER stress arms on MAM tethering [PERK/MFN2 (Munoz et al., 2013) and ATF6/VAPB (Gkogkas et al., 2008)], mitochondrial dynamics [PERK (Lebeau et al., 2018) and ATF6 (Bommiasamy et al., 2009)], ER-mitochondrial Ca 2+ transfer [IRE1α (Carreras-Sureda et al., 2019) and ATF6 (Burkewitz et al., 2020)], and mitochondrial bioenergetics [IRE1α (Carreras-Sureda et al., 2019) and PERK (Balsa et al., 2019)], more research is needed to delineate their therapeutic applications for regulating MAM/mitochondrial dysfunction in neurodegenerative pathologies. Moreover, IRE1α may also regulate mitochondrial dysfunction through inflammasome activation to promote inflammation (Bronner et al., 2015). Indeed, blocking inflammasome signaling in astrocyterestricted HIV-1 gp120 transgenic mice not only reduced neuroinflammation and neurodegeneration but also promoted neuronal regeneration and restored neurocognitive function (He et al., 2020).
The therapeutic potential of targeting inflammasome signaling gains support when considering the delicate balance between inflammasome activation and mitophagy that is evident in astrocytes and microglia during HIV/HAND. Proceeding mitophagy, mitochondrial fission via Drp1 is required for removal of damaged mitochondria, which is essential for cellular survival. Overexpression of Drp1 in astrocyte-restricted HIV-1 gp120 transgenic mice was able to reverse HIV-1 gp120induced mitochondrial elongation in neurons and reduce neuroinflammation (astrogliosis) and neurodegeneration . Similarly, Drp1 overexpression in CD4+ T cells expressing HIV-1 Vpr protected mitochondrial integrity and increased cell survival . Interestingly, cells exposed to HIV-1 Tat increase Drp1 expression and activity favoring mitochondrial fragmentation. Rozzi et al. (2018) was able to prevent fragmentation by indirectly inhibiting Drp1 activation. Thus, while altered mitochondrial dynamics is a notable characteristic of HIV/HAND pathology, the therapeutic application may largely depend on the model/stimuli.

CONCLUDING REMARKS
Cooperation and communication between the ER and mitochondria are essential to ensure maintenance of cellular and organelle homeostasis. The significance of MAMs in different pathologies is indisputable; however, whether MAM dysfunction is a cause or consequence in these pathologies is not yet known. This review presents MAMs as pivotal platforms in neuropathology highlighting specific interplay of MAMs in AD, PD, and ALS. We then discuss the implications of MAMs in HAND for the first time. These reports emphasize Ca 2+ , UPR and inflammasome signaling as potential targets to regulate MAM/mitochondrial dysfunction during neuropathological challenges.
Importantly, the presence and function of MAMs can differ across cell types and pathological conditions. A better understanding of the mechanisms regulating changes in MAMs in a respective cell type and/or pathology is critical to illuminate possible targets for therapeutic manipulation. For example, neurons are not directly infected by HIV-1; thus, indirect HIV-induced toxicity via infected glia is a key mechanism in HAND pathology. While glial cells (astrocytes, microglia, and oligodendrocytes) are essential for neural and CNS homeostasis, studies focusing on the presence and function of MAMs in glia are either severely deficient (astroglia) or non-existent (microglia and oligodendrocytes) highlighting the need to investigate ER-mitochondria contact and communication beyond the scope of neurons. Particularly, astrocytes provide essential metabolic and antioxidant support to neurons. Aberrant astrocyte mitochondrial function is a prominent threat to neuronal health and function. Targeting MAMs to manipulate astrocyte mitochondrial function could be a promising avenue to optimize the metabolic and antioxidant coupling between astrocytes and neurons and promote neuronal fitness against CNS pathologies. Moreover, microglia and astrocytes are the residential CNS immune cells participating in neuroinflammation, and MAMs may serve as key hinges for their activated/inflammatory status.

AUTHOR CONTRIBUTIONS
JP drafted the manuscript text and figures with guidance and supervision from I-WP and KB. All authors participated in editing or revising the article and agreed to be accountable for the content of the work.

FUNDING
This work was supported by R01 DA039789 and F31 DA053151 from the National Institute on Drug Abuse (NIDA), T32 AG020494 from National Institute on Aging (NIA) and the HSC Institute for Healthy Aging.