p62 promotes HIV-1 gp120 V3 loop-mediated microglial inammation by promoting noncanonical activation of Nrf2

Combined antiretroviral therapy (cART) has signicantly increased the life expectancy of AIDS patients; however, the prevalence of the neurocognitive impairment associated with HIV-1 continues to rise. HIV-1 gp120, an important subunit of the envelope spikes that decorate the surface of virions, is found to activate microglia in central nervous system (CNS) which leads to the cognitive and behavioral dysfunction known as HIV-1 associated neurocognitive disorder(HAND), and the V3 loop is the most important toxic domain of gp120. A study has shown that autophagy plays key role in the activation of microglia, p62 is an important autophagy substrate protein that is elevated in neuroinammation. In this study, we sought to explore the role of p62 in gp120 V3 loop-mediated microglial activation. Our results demonstrated that exposure of CHME-5 cells to the gp120 V3 loop resulted in elevated inammatory cytokines, accompanied by autophagy dysfunction and p62 upregulation. Subsequently, we found that the p62-dependent Nrf2 noncanonical signaling pathway was activated and that HO-1, the target protein of Nrf2, was also upregulated. Interestingly, the elevation of inammatory factors caused by the gp120 V3 loop was signicantly alleviated after knocking down p62, Nrf2 and HO-1. Further investigation revealed that in the microglial inammation induced by the gp120 V3 loop, up-regulated HO-1 promoted the expression of iNOS by interacting with iNOS, while enhanced autophagy by RAPA promoted the degradation of iNOS and alleviated inammation. These ndings provide a new perspective on the relationship between noncanonical Nrf2 activation and autophagy in microglial inammation and an experimental basis for HAND prevention and treatment. inammation by the gp120 V3 Silencing of p62 Nrf2 can prevent HIV-1 gp120 V3 loop mediated increased proinammatory cytokines


Introduction
HIV-1 not only destroys the peripheral immune system, but also infringes on the central nervous system and causes cognitive, motor or behavioral abnormalities [1,2], known as HIV-1 associated neurocognitive disorder (HAND) [3] . Dispite the extensive application of cART, the prevalence rate of HAND is still rising, due to the poor biological permeability and toxicity of drugs [4,5]. Therefore, it is urgent to explore and clarify the pathogenesis of HAND to provide a basis for its prevention and treatment. The main pathogenesis of HAND is the activation of microglial cells by HIV-1 associated proteins, such as gp120, and the release of proin ammatory cytokines, which results in the neuronal injury [6,7]. Our previous studies have shown that the gp120 peptide can promote autophagy in BV2 microglia [8], but the speci c mechanism of autophagy regulation in microglia is still unclear.
Macroautophagy, hereafter referred to as autophagy, is an important mechanism by which cells to maintain their homeostasis. Autophagy begins with the formation of phagocytic vesicles, and then the membrane of phagocytic vesicles stretches to form autophagosomes with a unique double-membrane structure. The whole autophagosome then fuses with the lysosome to form an autolysosome, containing a series of lysosomal degrading enzymes (such as cathepsin and other acidic hydrolases) that digested the tomont contents of the autolysome. In this study, we selected three autophagy-associated proteins such as LC3, Beclin-1 and p62, as indicators and assessed autophagic ux to determine the autophagy level and function. Autophagy ux is currently the most reliable method for measuring autophagy, which uses labeling of LC3 with vectors (such as plasmid, adenovirus, lentivirus, etc.) in tandem with GFP( green uorescent protein )and RFP(red uorescent protein) to re ect the process of autophagy.
p62 is an important autophagy adaptor protein that can recruit ubiquitinized substrates to autophagosomes for degradation and is eventually degraded in autolysosomes [9]. Therefore, the cellular level of p62 has been considered a marker for autophagy activity, and the amount of p62 is considered to be inversely correlated with autophagic ux [10]. It is also worth noting that p62 is also a multifunctional signaling center that is involved in a variety of cellular activities such as apoptosis, in ammation, and the oxidative stress response [9]. p62 directly binds to Keap-1 and thus leads to the noncanonical activation of Nrf2. In general, Nrf2 is stored in the cytoplasm in an inactive state and is mainly negatively regulated by Keap-1, which can induce cullin3 E3 ubiquitin ligase to promote the ubiquitination of Nrf2 and its degradation by the proteasome. Haefen et al found that in an LPS-induced model of central nervous in ammation in rats, toxic lentils can reduce neuroin ammation by reducing p62 levels and improving autophagy dysfunction [11]. Additionally, Jin et al reported that koumine can reduce the level of p62 and further inhibit the activation of astrocytes [12]. Moreover, it has been reported that in a rat model of stress hypertension, glial cells are largely activated and release in ammatory cytokines such as IL-1β and TNF-α , while minocycline can alleviate in ammation by reducing p62 levels [13]. These studies have shown that the high-expression of p62 is common in a neuroin ammation model, but the exact pathogenesis is not clear, nor is its role in HAND.
Nrf2 is a transcription factor belonging to cap 'n' collar family of zippers, and there are two ways to activate Nrf2 that are known. Under oxidative stress, the canonical activation of ROS and other electrophiles modi es the speci c cysteine of Keap1, thereby blocking the ubiquitination degradation pathway of Nrf2, and increasing the level of Nrf2 further promotes its downstream protein expression [14]. The other way is the noncanonical way, in which various reasons cause autophagy dysfunction leading to p62 accumulation, after which p62 binds to Keap-1 competitively so that the level of Nrf2 increases [15], while p62 is also one of the target protein of Nrf2, with the p62-dependent upregulation of NRF2 increasing the transcription of p62 itself, creating a positive feedback loop that promotes the prolonged activity of NRF2, leading to intracellular redox imbalance [16,17]. Based on these data outlined above, this study investigated the role and mechanism of p62 mediating crosstalk between autophagy and the Nrf2 noncanonical signaling pathway in HIV-1 gp120 V3 loop-mediated microglial in ammation.

Result
HIV-1 gp120 V3 loop-mediated defective autophagy increases microglial activation and elevates proin ammatory cytokines in CHME-5 cells First, we observed the level of in ammatory cytokines released by microglia induced by thebHIV-1 gp120 V3 loop. TNF-α, MCP-1 and IL-1β were chosen as markers of in ammation. As seen in gure 1A, the gp120 V3 loop can induced the upregulation of these in ammatory factors, while the autophagy inducer rapamycin can inhibited this effect.
Based on the changes in in ammatory cytokines, we then observed the effect of the HIV-1 gp120V3 loop on autophagy in CHME-5 cells. Since autophagy is a dynamic process, we used the GFP-RFP-LC3 adenovirus to infect cells to study the changes in autophagic ux in CHME-5 cells treated with or without HIV-1 gp120 V3 loop. Following the formation of the autophagosomes, LC3 is labeled with both red and green uorescence that appeared as merged yellow puncta due to the colocalization of GFP and RFP-LC3B. However, under a pH<4 condition, the GFP signal of the adenovirus-expressed uorescently tagged LC3B fusion protein is quenched, whereas the red uorescence remains stable, showing red after the overlay. The increased accumulation of the yellow puncta is thus an indicator of impaired autophagy, while the increased accumulation of the red puncta indicates increased autophagic ux ( gure 1B). As shown in gure 1(B, C), CHME-5 cells transfected with the GFP-RFP-LC3 adenovirus followed by exposure to the HIV-1 gp120 V3 loop (50 ng/mL) for 24 h, exhibited a signi cant increase in the yellow puncta compared with the random peptide group (the control group), with signi cantly more yellow puncta than red puncta in the group, indicating incomplete autophagosome maturation. However, decreased numbers of yellow puncta were observed in CHME-5 cells exposed to rapamycin (20 nM), an autophagy ux inducer, along with an increased formation of red puncta ( gure 1B, C)), indicating increased autophagy ux. Next, we assessed changes in the levels of the autophagy-related proteins Beclin-1, LC3 and p62. As shown in gure 1C, western blot analysis indicated that the gp120 V3 loop increased the expression of LC3, Beclin-1 and p62, while pretreatment with RAPA reversed these effects in CHME-5 cells( gure 1D).
Therefore, these results indicate that the gp120 V3 loop can increase proin ammatory cytokines and mediate autophagy dysfunction in CHME-5 cells, while promoting autophagy with RAPA can improve microglial in ammation caused by the gp120 V3 loop Silencing of p62 or Nrf2 can prevent HIV-1 gp120 V3 loop mediated increased proin ammatory cytokines in CHME-5 cells Next, we wanted to explore whether silencing the autophagy-associated protein p62 or its related signaling molecule Nrf2 genes could alleviate in ammation in CHME-5 cells treated with the gp120 V3 loop. Therefore, in ammatory cytokines such as TNF-α, MCP-1 and IL-1β were assessed in gp120 V3 loop-exposed CHME-5 cells transfected with either p62 siRNA or Nrf2 siRNA ( gure 2A, C). As shown in gure 2B and D, silencing of both p62 and Nrf2 reversed the upregulation of in ammatory cytokines induced by the gp120 V3 loop in CHME-5 cells. These results suggest that silencing of p62 and Nrf2 can prevent HIV-1 gp120 V3 loop-mediated increases in proin ammatory cytokines in CHME-5 cells.
HIV-1 gp120 V3 loop leads to noncanonical activation of Nrf2 in CHME-5 cells To further study the mechanism by which silencing p62 or Nrf2 can prevent HIV-1 gp120 V3 loopmediated increases in proin ammatory cytokines in CHME-5 cells, we observed the exact activation way of Nrf2 in CHME-5 cells treated with the gp120 V3 loop. As mentioned above, Nrf2 can be activated in two ways, among which noncanonical activation is marked by the competitive binding of p62 to Keap-1, which is the negative regulator of Nrf2. First, we observed the effect of the knockdown of p62 on Nrf2 signaling pathway-related protein levels, including Nrf2 cytoplasmic and nuclear levels, Keap-1, and the Nrf2 target protein HO-1 ( gure 3 A, B). The results show that in the gp120 V3 loop treatment group, Nrf2 was increased in both the cytoplasm and the nucleus, as well as Keap-1 and HO-1and that knockdown p62 could reverse these changes caused by the gp120 V3 loop. Moreover, immunoprecipitation assays revealed an interaction between p62 and Keap-1 in CHME-5 cells treated with the gp120 V3 loop ( gure 3C), which is a prominent sign of noncanonical activation of Nrf2. To further explore the relationship between p62 and Keap-Nrf2 signaling, we observed changes in p62 and Keap-1 protein levels after silencing Nrf2. Figure 3D shows that silenting Nrf2 signi cantly reduced the levels of p62 and Keap-1, which indicates that p62 and Keap-1 are also targets of Nrf2 in CHME-5 cells treated with the gp120 V3 loop. In summary, these results suggest that the gp120V3 loop induces Nrf2 noncanonical activation of in CHME-5 cells.
We next explored the mechanism of in ammation in CHME-5 cells induced by noncanonical activation of Nrf2 induced by the gp120 V3 loop. In a previous study, we found that HO-1, the target protein of Nrf2, was increased in the gp120 V3 loop treatment group. Therefore, we performed a knockdown of HO-1 with RNA interference ( gure 4A), and we found that the TNF-α, MCP-1 and IL-1β levels were signi cantly decreased compared with the gp120 V3 loop group after knockdown of HO-1( gure 4C). At the same time, we found that the knockeddown of HO-1 can signi cantly reduced the protein level of iNOS( gure 4B), Immuno uorescence staining results showed that iNOS and HO-1 levels in the gp120 V3 loop group were upregulated and colocalized ( gure 4D) ; however, iNOS protein levels were signi cantly decreased in the HO-1 knockdown group compared to the gp120 V3 loop group. These results suggest that the HIV-1 gp120 V3 loop can up-regulate HO-1, which interacts with iNOS to promote in ammation.
Rapamycin promotes the degradation of iNOS through autophagy in HIV-1 gp120 V3 loop treated CHME-5 cells To explore whether the upregulation of iNOS was related to impaired autophagy in addition to its interaction with HO-1, we used the autophagy inducer rapamycin to observe the relationship between autophagy and iNOS levels in HIV-1 gp120 V3 loop treated CHME-5 cells. As seen in gure 5, compared with the gp120 V3 loop group, the autophagy inducer rapamycin plus gp120 V3 loop group showed signi cantly decreased iNOS levels in both the immuno uorescence assay ( gure 5A) and the western blot ( gure 5B, C). In addition, we further found by an immuno uorescence assa that the autophagy inducer rapamycin could downregulate the expression of iNOS. These results suggest that rapamycin promotes the degradation of iNOS through autophagy in HIV-1 gp120 V3 loop-treated CHME-5 cells.
Brusatol alleviates neurocognitive dysfunction in mice induced by the HIV-1 gp120 V3 loop Because our in vitro data suggested that silencing Nrf2 can reduce in ammation caused by the gp120 V3 loop, we used brusatol, an inhibitor of Nrf2, to verify the mechanism by which noncanonical activation of Nrf2 promotes microglial in ammation in vivo. We rst observed the alleviation effect of brusatol on learning and memory dysfunction induced by the gp120 V3 loop in mice. The Water maze and Stepthrough passive avoidance test were used to evaluate learning and memory function in our research. Mice were randomly divided into four groups, and the details of the groups were the same as those described previously (Materials and Methods). As seen in gure 6A and B, the gp120V3 loop prolonged escape latency signi cantly, while brusatol reversed the effects on the 5 th and 6 th days of testing. Representative search strategy trials also showed a difference between the gp120 V3 loop group and the gp120 V3 loop plus brusatol group ( gure 6A). Similarly, in the probe trial, the group injected with the gp120V3 loop demonstrated fewer platform crosses and spent less time in the target quadrant than that of the random peptide group, while the group administered with the gp120V3 loop and brusatol demonstrated more platform crosses and spent more time in the target quadrant than did the group administered with the gp120V3 loop alone. The representative search strategy trials in the probe test were also showed that mice in the gp120 V3 loop plus brusatol group performed better in nding the platform than those in the gp120 group ( gure 6B). Furthermore, the results of the Step-through passive avoidance test showed that the brusatol group showed longer latency and fewer errors than the gp120V3 loop group ( gure 6C and 6D). These data suggested that brusatol alleviates cognitive impairments caused by the HIV-1gp120 V3 loop.
Brusatol inhibits HIV-1 gp120 V3 loop mediated increased autophagy marker expression, microglial activation and proin ammatory cytokines in mice To further con rm the the role of Nrf2 in microglial in ammation induced by the gp120 V3 loop, Western blotting was used to observe the levels of autophagy-related proteins, Nrf2 signaling related proteins and iNOS levels in the brain tissues of mice in each group. As shown in gure 7 A and 7 B, we found that compared with the control group, Beclin-1, LC3 and p62 in the gp120 group were signi cantly upregulated, suggesting autophagy dysfunction, that Nrf2, Keap-1 and HO-1 were upregulated, indicating that Nrf2 was noncanonical activated; and that iNOS signi cantly upregulated in ammation. However, brusatol can further promoted the upregulation of Beclin-1 and LC3, and downregulated p62, indicating that brusatol alleviated autophagy dysfunction caused by the gp120 V3 loop, and downregulated Nrf2, Keap-1, HO-1 and iNOS, indicating that brusatol can inhibited the noncanonical activation of Nrf2.
Furthermore, the ELISA results showed that the gp120 V3 loop promoted the upregulation of in ammatory factors in brain tissues (including TNF-α, MCP-1 and IL-1β), while brusatol alleviated in ammation ( gure 7C). To further con rm the role of autophagy in microglial activation, we next performed immuno uorescence staining for autophagy-associated protein p62 and microglial marker Iba-1 in the hippocampus ( gure 8A) and the cortex ( gure 8B) of the mice from each group. As shown in gure 8, the intensity of the p62 (green) and Iba-1 (yellow) puncta formation was signi cantly increased in the hippocampus and cortex of gp120V3 loop mice compared to the random peptide group. Furthermore, the red uorescence (Iba-1) was signi cantly increased and colocalized with the autophagyassociated protein p62 (green) in the gp120 V3 loop group. These ndings thus underscore the effects of brusatol on HIV-1 gp120 V3 loop-mediated increased autophagy marker expression, microglial activation and proin ammatory cytokines in mice.

Discussion
Studies have shown that neuroin ammation of caused by HIV-1 activated microglia is the key factor in the occurrence and development of HAND [18,19]. Microglia are the most important immune cells in the brain tissue and are widely distributed in the brain [20]. There is also evidences showning that even in HAND patients treated with HAART(highly active antiretroviral therapy), the severity of the disease is related to the activation of microglia cells [19]. gp120, which is an envelope protein of the HIV-1 virus, can can induce a series of in ammatory factors, such as TNF-α, MCP-1, IL-1β, IL-6 and iNOS by binding the corresponding receptors on the microglial cell surface, and these in ammatory factors can act on other glial cells to amplify the neuroin ammation. In this study, the HIV-1 gp120 V3 loop was used, which is the third variable region of gp120 [21] and the most important structural domain of gp120-induced toxicity [22]. In addition, random peptide segments were selected as the control, ensuring the scienti c nature of the control.
Autophagy is associated with in ammation, tumors, neurodegenerative diseases and other diseases. Recent studies, including our preliminary study, have demonstrated the role of impaired autophagy in neurons as a correlate of neurodegeneration in HAND [23][24][25]. However, the speci c mechanism by which autophagy regulating microglial activation remains unclear. In this study, we demonstrated that the persistent stimulation of the HIV-1 gp120 V3 loop impaired the integration of autophagosomes and lysosomes in CHME-5 cells. Additionally, in ammatory factors were increased in CHME-5 cells treated with the gp120 V3 loop and reduced when treated with rapamycin simultaneously, suggesting that autophagy dysfunction promotes in ammation.
We further observed the changes in autophagy-related proteins and found that p62 was upregulated.
Intracellular proteins or organelles can be degraded by two pathways: the proteasome pathway and the autophagy lysosome pathway. p62, as an autophagy substrates of chitosan protein combines with LC3 through polymerization , and is degraded in the targeted autophagosome [26], which is the main negative regulatory mechanism by which p62 maintain its basic level in cells, an thus it is possible that autophagy dysfunction causes abnormal accumulation of p62. Furthermore, abnormal accumulation of p62 affects many signaling pathways due to its special structure, and studies have shown that a lack of p62 in mice results in lower energy consumption and fever, as well as lower transcription [27]. This may be related to p62 participating in nutritional mTORC1 activation in sensing to adjust the adipocyte differentiation and energy controlling [27,28]. Interestingly, in ammatory factors were reduced after p62 knockdown in our in vitro studies, suggesting that p62 accumulation may promote microglial in ammation, which is consistent with the results of many studies. For example, excessive accumulation of p62 causes the in ammatory environment to promote the occurrence of hepatocellular carcinoma [29], and there is a large amount of p62 accumulation in microglial toxic reactions caused by trimethyl tin chloride [30]. The canonical activation of Nrf2 is caused by the conformational changes in Keap-1 caused by various oxidative stimuli, and the dissociation of Keap-1 and Nrf2 causes the activation of Nrf2, in which p62 is not involved. p62 is involved in the noncanonical activation of Nrf2, which is characterized by the competitive binding of p62 to Keap-1, the main negative regulator of Nrf2 [15]. After Nrf2 activation and incorporation into the nucleus, the antioxidant damage effects of target genes such as HO-1, NQO1 and GSH have been con rmed, and it cannot be ignored that some studies have noticed the negative effects of Nrf2, such as liver brosis, liver injury and Kaposi's sarcoma as mentioned above. In nervous system diseases, neuro brillary tangling of brain tissue of AD brain tissue was found to polymerize the phosphorylated p62-Keap-1 complex, and the expression of the Nrf2 target gene was increased simultaneously [31][32][33]. However, the relationship between the noncanonical activation of Nrf2 and HAND was still unknown. In this study, we found that autophagy dysfunction in CHME-5 cells caused by HIV-1 gp120 V3 loop led to the upregulation of p62 and the release of in ammatory factors. In addition, Keap-1 expression was up-regulated, and the knockdown of p62 could reduce Nrf2 levels. Combined with the Co-IP results, this indicated that Nrf2 was noncanonically activated by p62. Furthermore, in ammatory factors were downregulated after the knockdown of Nrf2, suggesting that noncanonically activated Nrf2 was related to in ammation.Nrf2 was related to in ammation. In addition, Keap-1 and p62 have also found to be the target proteins of Nrf2 [34,35]. This indicates that when Nrf2 is activated, Keap-1 and p62 levels are upregulated, and thus p62-Nrf2 forms a positive feedback loop, which was also con rmed in our study, which also showed that in ammatory cytokines were downregulated after blocking this positive feedback in gp120 V3 loop-treated CHME-5 cells. Therefore, the continuous upregulation of p62, Keap-1 and Nrf2 levels leads to the imbalance of intracellular redox.
HO-1, one of the target proteins of Nrf2, was found to be correlated with in ammation or oxidative stress in lung ischemia reperfusion [36]. In our study, HO-1 and iNOS were both upregulated and co-localized, as observed by immuno uorescence detection. Interestingly, iNOS was downregulated after knocking down HO-1, it is suggesting that HO-1 can promote iNOS expression by interacting with it, but the speci c sites of its interaction need to be further studied.
In vivo, a HAND mouse model was constructed by 8-week-old SPF mice injected with the HIV-1 gp120 V3 loop in the lateral ventricle. This model is copied based on the papers previously published by our research group [37] , but it is less di cult and more reproducible than the previous rat model. In the preexperiment, the brain was harvested for observation 10 min after the injection of trypan blue for 10 min, so as to ensure the accuracy of the injection mode and the injection point. Brusatol is a speci c blocker of Nrf2 that can rapidly reduce the expression of Nrf2 independent of on Keap-1. It has been reported that brusatol may reduce the protein level of Nrf2 through demethylation [38] and has a strong lipid solubility to allow it to cross the blood-brain barrier. Brusatol can inhibit cytokine dependent induction of iNOS [39] in β cells, furthermore, brusatol can inhibit the c-myc /ROS signaling pathway in rectal cancer [40]. Our study found that brusatol alone had no signi cant effect on behavior or in ammatory factor expression in mice, while it can signi cantly improved the performance of the water maze and step-through passive avoidance in mice lateral in the lateral ventricle with gp120 V3 loop. Further experiments showed that brusatol reduced the release of in ammatory cytokines in mouse brain tissue induced by the HIV-1 gp120 V3 loop and reduced the expression of the mouse hippocampus and cortex microglia marker protein Iba-1. Molecular biology experiments found that Nrf2, Keap-1, HO-1 and iNOS expression was upregulated in the gp120 V3 loop group, autophagy-related proteins Belin-1, LC3 and p62 was upregulated, and the microglial marker protein iba-1 was upregulated and colocalized with p62 and LC3. Brusatol inhibited the noncanonical activation of p62-dependent Nrf2 induced by the gp120 V3 loop, and the downregulation of p62 and Keap-1 suggested that inhibition of Nrf2 blocked the positive feedback loop of p62-Nrf2. Therefore, the results of the in vitro experiments were veri ed well by the reduction in the level of in ammatory cytokines in the brain tissue and in reducing the impairment of learning and memory function in mice.

Conclusions
This study investigated the role of p62 in CHME-5 cells in ammation induced by the HIV-1 gp120 V3 loop. We demonstrated that the gp120 V3 loop can cause autophagic ow blockade, cause p62 accumulation, promote noncanonical activation of Nrf2, upregulate the level of HO-1, interact with iNOS to upregulate in ammation, and degrade iNOS through the autophagic lysosome pathway. Thus, targeting the noncanonical activation of Nrf2 in microglia may provide a novel therapeutic approach to treat HAND.

Western Blot Analysis
The cells were treated according to the study design and harvested, The brain samples were suspended in buffer solution and then centrifuged at 12,000 rpm for 30 min at 4°C. The supernatant was separated and collected for protein analysis. Protein was extracted using a protein extraction kit (Beyotime, China), and the concentration was determined using a Bio-Rad Protein Assay. The cytosolic and nuclear fractions were separated using a nuclear and cytoplasmic protein extraction kit, according to the manufacturer's instructions (WLA020a, Wanleibio Co., Ltd. Shenyang China). SDS-polyacrylamide gels were used for separation via electrophoresis and then transferred to polyvinylidene uoride membranes at 120 V for 30 min and 80 V for 2 h. The transferred membranes were blocked using fresh 5% nonfat dry milk dissolved in Tris-buffered saline Tween-20 (TBST) at room temperature for 1 h. The immunoblots were then probed with the appropriate antibodies. Primary antibodies used were: anti-GAPDH (5174s), anti-p62 (23214s), beclin-1 (3495s), anti-LC3 (3868s) anti-Keap1 (4678s) were purchased from Cell Signaling Technology, and anti-Nrf2 (WL02135, 1:1000, Wanleibio, Shenyang, China), anti-Lamin B (WL01775,1:1000, Wanleibio, Shenyang, China), anti-HO-1 (ab223349, 1:1000, Abcam), anti-iNOS (ab178945, 1:1000, Abcam). Each primary antibody was individually incubated with the membrane at 4°C for 12 h. Following incubation with primary antibodies and washing three times with TBST for 10 min each, the membranes were incubated for 1 h with the appropriate secondary antibody (diluted 1:5000 in TBST, 5% skim milk) and subsequently washed an additional three times with TBST for 10 min each. The detection of speci c proteins was performed with an enhanced chemiluminescent detection system. The bands were quanti ed by densitometry using Image-ProPlus 6.0. ELISA TNF-a, MCP-1 and IL-1β kits (USCN Business Co., Ltd, China) were used to determine the cytokine levels in mouse serum according to the manufacturer's instructions.

RFP-GFP-LC3 Adenovirus Transfection
To analyse autophagic ux, CHME-5 cells were transfected with mRFP-GFP-LC3-expressing adenovirus (Genechem Co.,Ltd. China), according to the manufacturer's instructions. The procedure was performed as follows: the CHME-5 cells were plated on a confocal dish and treated with 250 μL of infection enhancement medium containing 1×10 5 PFU RFP-GFP-LC3-expressing adenovirus. After 2 h of incubation, the medium was replaced by the adenovirus-free medium. Later, at 24 h, the treatment solutions were then applied to the plate, followed by incubation for 24 h. The images were acquired with an Leica uorescence microscope with the cells xed.
SiRNA transfection CHME-5 cells were transfected either with p62, Nrf2 or HO-1 siRNAs. Brie y, mPMs were seeded into 6well plate at a density of 3×10 5 cells/well and cultured in a humidi ed, 5% CO2 incubator at 37°C. On the next day, the culture medium was replaced with Opti-MEM® I Reduced Serum Medium. Meanwhile, the individual targeted siRNA and scrambled siRNA (500 pmol/ml) and Lipofectamine® 2000 (2 μl/ml) were incubated separately with Opti-MEM® I Reduced Serum Medium for 5 min at room temperature. After incubation, both individually targeted siRNAs and Lipofectamine® 2000 mix were mixed and subsequently, incubated at room temperature for another 15 min, mixed liquids were gently added to the culture medium. After 6 to 8 h the culture medium was replaced with DMEM supplemented with 10% heat-inactivated FBS. The transfected mouse primary microglia cells were then exposed to HIV-1 gp120V3 loop for 24 h, and transfection e ciency was analyzed by western blotting. All experiments were repeated at least 6 times.
Co-Immunoprecipitation CHME-5 cells were lysed and analyzed using previously established protocols in Western Blot Analysis. Lysates were centrifuged at 12,000 g for 15 min at 4 °C. Cleared lysates were incubated overnight with the indicated antibodies and for 1 h with protein A/G beads (Med Chem express, USA). The pellets were then washed four times with ice-cold lysis buffer. The immunoprecipitation samples were eluted from the beads by heating with SDS sample buffer. Eluted immunoprecipitates or whole cell lysates were separated by SDS-PAGE and analyzed by western blot.

Immuno uorescence Analysis
In vitro: After incubation according to the study design, the CHME-5 cells were xed with 4% PFA in PBS for 30 min, washed with PBS three times and then permeabilized with Triton-X-100 for 15 min. After washing with PBS three times, the cells were blocked with 5% BSA for 1 h at room temperature and then incubated with anti-HO-1 (ab223349, 1:100, Abcam), anti-iNOS (ab178945, 1:100, Abcam) in 5% BSA overnight at 4 °C. Subsequently, the cells were washed in PBS three times and incubated with uorescein isothiocyanate-labeled secondary antibody for 1 h at 37 °C. The cells were then washed and incubated with DAPI for 15 min and then washed again and mounted. The cells were then imaged using Leica uorescence microscope.
In vivo: After irrigation with 4% PFA, the brain tissue was removed and xed in PFA for 24h. After washing with PBS for three times, 10% sucrose →20% sucrose →30% sucrose was used for gradient dehydration, and then oct-embedded and frozen at -80°C for about 40min before sectioning. The brain was cut into 20 m thick coronal sections with a frozen slicer, and baked for 2h at 38-40°C on a sheet baking machine. After drying to room temperature, the tissue slices could be stained with immuno uorescence. Sections (10 μm thick) were cut and mounted on microscope slides. Staining of p62 and Iba-1 using anti-p62 (1:100, 23214s, Cell Signaling Technology), anti-Iba-1 (1:100, ab5076, Abcam) primary antibodies were performed in mouse hippocampus and cortex.

Animals and intracerebroventricular injection
Male C57BL6J mice (8 weeks old) were procured from Shandong Experimental Animal Center (37009200014313). All animals were kept under a 12-hour light/12-hour dark cycle and allowed free access to water and food. mice were housed in wire cages at 20-22℃ and 50 ± 5% humidity. All experiments were approved by the Laboratory Animal Ethics Committee of Jinan University. Intracerebroventricular injection: mice were anesthetized with 1.2% tribromoethanol and placed in a stereotaxic apparatus (RWD Life Science, Shenzhen China). A burr hole was drilled 0.2-0.5 mm posterior to bregma and 1.0 mm lateral to the midline, and the lateral ventricle was injected with a micro syringe at 2.5-3.0mm below the skull. The sham group was given random peptides and gp120V3 loop group was given 100ng/5μL/d, which was administered continuously for 3 days, once a day. After lateral ventricle injection, the brusatol group and gp120V3 loop+brusatol group were intraperitoneally injected with 2 mg/kg of brusatol, once every other day, until sampling. The control group and gp120 group were intraperitoneally injected with the same amount of normal saline. The random peptide and gp120 were

Morris water maze
The Morris water maze test is used to assess memory. We performed this test as described by Morris et al [41]. A circular plastic pool (height: 60 cm, diameter: 120 cm) was lled with milk kept at 22-24°C. An escape platform (height: 30 cm, diameter: 10 cm) was submerged 1-1.5 cm below the surface of the water in position. A camera was xed to the ceiling above the water maze and connected to a computerbased program (Ethovision XT, Noldus Netherlands). mice were tested over 5 days (four trials/day) to nd the hidden escape platform in the bottom left quadrant. They had a maximum of 60 s to nd the platform. If animals were unable to locate the platform, they were gently guided to the platform and required to stay on the platform for 30 s before being removed. Mice were subjected to a probe trial by removing the platform during 60s. The number of rats crossing the target quadrant and time spent in the target quadrant were used as measures of spatial memory.
Step-Through Passive Avoidance test Step-Through Passive Avoidance test is a behavioral experiment designed to observe the short-term memory reproduction ability of mice. The experimental device (Chengdu tmin life technology co., LTD) is half dark room and half-light room, connected with adjustable small holes, and the bottom of the dark room is equipped with an electric shock device. If mice escape from the dark room to the dark room, they will be shocked immediately. The experiment was divided into two days of Training and Test at an Test: clean the experimental equipment and remove the smell to ensure that the environment is basically consistent with the training. Test was performed on the 24th h after the training of mice, the hole was closed before startup, the hole was opened immediately after the mice were put into the open chamber and the electric power was switched on. Latency and error times of the corresponding mice were observed and recorded.

Statistical Analysis
Data are presented as mean ± SEM. The data are from at least three different experiments, with each data point within an individual experiment representing triplicate measurements. Statistical analyses were performed using SPSS 13.0. Values were compared using the Nonparametric Kruskal -Wallis Oneway ANOVA followed by the Dunn post hoc test was used to determine the statistical signi cance Page 14/18 between multiple groups and the Wilcoxon test was used to compare between 2 groups .The results were considered statistically signi cant when P < 0.05.