The interaction between intestinal bacterial metabolites and phosphatase and tensin homolog in autism spectrum disorder

Intestinal bacteria-associated para -cresyl sulfate ( p CS) and 4-ethylphenyl sulfate (4EPS) are elevated in autism spectrum disorder (ASD). Both metabolites can induce ASD-like behaviors in mice, but the molecular mechanisms are not known. Phosphatase and tensin homolog (PTEN) is a susceptibility gene for ASD. The present study investigated the relation between p CS and 4EPS and PTEN in ASD in a valproic acid (VPA)-induced murine ASD model and an in vitro LPS-activated microglial model. The VPA-induced intestinal inflammation and compro- mised permeability in the distal ileum was not associated with changes of PTEN expression and phosphorylation. In contrast, VPA reduced PTEN expression in the hippocampus of mice. In vitro results show that p CS and 4EPS reduced PTEN expression and derailed innate immune response of BV2 microglial cells. The PTEN inhibitor VO-OHpic did not affect innate immune response of microglial cells. In conclusion, PTEN does not play a role in intestinal inflammation and compromised permeability in VPA-induced murine model for ASD. Although p CS and 4EPS reduced PTEN expression in microglial cells, PTEN is not involved in the p CS and 4EPS-induced derailed innate immune response of microglial cells. Further studies are needed to investigate the possible involvement of reduced PTEN expression in the ASD brain regarding synapse function and neuronal connectivity.


Introduction
Autism spectrum disorder (ASD) is a neurodevelopmental disease which core symptoms are impairments in social interaction and communication, deficits in cognitive function as well as the presence of stereotyped behavior (Lord et al., 2018). Recent epidemiological surveys show the worldwide prevalence of ASD is about 1 % (Fombonne et al., 2021). At present, there is no effective treatment targeting ASD detrimental core symptoms (Genovese and Butler, 2020). Although the ASD pathogenesis is not clear, the abnormal changes along the microbiotagut-brain axis have been reported to be involved (Salim et al., 2021). The involvement of the gut in ASD is also supported by frequently reported gastrointestinal comorbidities, including intestinal inflammation and compromised intestinal barrier (Babinská et al., 2017;Dalton et al., 2014;Kang et al., 2017;Walker et al., 2019).
Children with ASD show changes in gut microbiota and microbiotaassociated metabolites (Kang et al., 2020;Kang et al., 2018;Zou et al., 2020). Para-cresyl sulfate (pCS) and 4-ethylphenyl sulfate (4EPS) are host metabolites derived from two bacterial metabolites, para-cresol (pcresol) and 4-ethylphenol, originating from intestinal bacterial fermentation of tyrosine and phenylalanine (Zheng et al., 2021). It has been reported that children diagnosed with ASD show a overgrowth of small intestinal bacteria, indicating that the p-cresol-and or 4EP-producing bacteria might be existing in distal ileum . p-Cresol and 4-ethylphenol can be absorbed from the gut and metabolized into pCS and 4EPS Zheng et al., 2021;Zheng et al., 2022). It has been suggested that 95 % of p-cresol and metabolized by the host into pCS via O-sulfonation, a process that occurs primarily in the liver, and to a smaller extent in colonic epithelial cells (Persico and Napolioni, 2013;Ramakrishna et al., 1991;Rong and Kiang, 2021). In addition, 4-ethylphenol has also been suggested to be sulfated into 4EPS via O-sulfonation in the host (Hsiao et al., 2013;Needham et al., 2022). Both pCS and 4EPS have been shown to be elevated in urine, feces or serum of children with ASD (Gabriele et al., 2014;Kang et al., 2020;Kang et al., 2018;Needham et al., 2021). We have demonstrated elevated pCS in the serum of the valproic acid (VPA)-induced mouse model for ASD (Zheng et al., 2022). Furthermore, p-cresol administration in drinking water causes social behavior deficits and repetitive behavior in mice through remodeling of the gut microbiota (Bermudez-Martin et al., 2021). In addition, intraperitoneal injection of 4EPS into mice induces anxiety-like behavior, a common co-morbidity that may contribute to core ASD symptoms (Hsiao et al., 2013;Needham et al., 2022). Moreover, a clinical trial has recently shown that an oral gastrointestinal (GI)-restricted adsorbent AB-2004 decreases urine pCS and 4EPS levels and ameliorates ASD-like behaviors of ASD children without serious side-effects (Stewart Campbell et al., 2022). However, it is unclear yet how these two bacteria-derived metabolites are mechanistically linked with ASD.
Phosphatase and tensin homolog (PTEN) is a well-recognized ASD susceptibility gene and a germline mutation in PTEN has been identified in up to 20 % of children diagnosed with ASD with macrocephaly (Busch et al., 2019;Yehia et al., 2020). Studies in mouse models demonstrate that deletion of PTEN in the cerebral cortex and hippocampus results in increased rates of macrocephaly and abnormal social interactions (Kwon et al., 2006). Moreover, PTEN expression in brain is decreased in in utero valproic acid (VPA)-exposed male mice (Mahmood et al., 2018). PTEN inhibition or knockdown attenuates neuroinflammation in mice (Pan et al., 2022;Shen et al., 2021). In contrast, loss-of-function mutation of PTEN in mice upregulates neuroinflammation and enhances microglial phagocytosis (Sarn et al., 2021a;Sarn et al., 2021b). Gut-derived pCS and 4EPS are regarded as an important protein-bound uremic toxins associated with chronic kidney disease in human and rodent models (Vanholder et al., 2007;Velenosi et al., 2016). Of interest is the recent finding that in uremic mice a downregulated PTEN expression is associated with peripheral inflammation, indicating the pCS can affect PTEN expression (Wei et al., 2022). Our previous study has demonstrated that pCS derailed the innate immune response as well as phagocytosis of BV2 microglial cells (Zheng et al., 2022). However, little is known about the interaction between bacteria-derived pCS and 4EPS and PTEN in ASDassociated intestinal inflammation and derailed neuroimmune responses in the brain.
The current study aims to investigate the potential relation between PTEN and pCS and 4EPS, using in vivo and in vitro models to unravel the molecular mechanisms involved of how pCS and 4EPS contribute to ASD. First, in tissues obtained from in utero VPA-exposed male mice the intestinal and brain PTEN phosphorylation and/or expression was assessed. Subsequently BV2 microglial cells were used to investigate whether pCS or 4EPS has a direct effect on PTEN expression. Next, the role of PTEN in microglial neuroimmune response and phagocytosis associated with pCS and 4EPS was investigated.

Mice
As previously described (de Theije et al., 2014), specific pathogenfree BALB/cByJ breeding pairs from Charles River laboratories (Maastricht, the Netherlands) were housed under a 12 h light/dark cycle with free access to water and standard food for laboratory rodents AIN-93 M. All animal procedures were conducted according to governmental guidelines and approved by the Ethical Committee of Animal Research of Utrecht University, Utrecht, the Netherlands (CCD number AVD108002017826). All females were mated until a vaginal plug was detected, indicated as gestational day 0 (G0). On G11, after neural tube closure, pregnant females were treated subcutaneously with 600 mg/kg VPA (Sigma, Zwijndrecht, the Netherlands, VPA: 100 mg/ml) or phosphate buffered saline (PBS). The offspring was weaned on postnatal day 21 (P21). On postnatal day 50, male mice were euthanized by decapitation to collect intestinal tissue and brain tissue (in utero PBS-exposed male mice: n = 3; In utero VPA-exposed male mice: n = 3).

BV2 microglial cell viability
To investigate the effect of 4EPS on cell viability, 5000 BV2 cells/ well were incubated with 4EPS (concentration range: 0.1-400 μM) for 24 h. After 24 h, 50 μl medium of each well was transferred into new 96well plate and the content of lactate dehydrogenase (LDH) in medium was measured by Cytotoxicity Detection KitPLUS (LDH) (4,744,926,001, Sigma) according to the manufacturer's instructions.
Meanwhile, the medium left was removed and 100 μl DMEM containing 0.5 mg/ml MTT (M2128, Sigma) was added to the cells for 4 h of incubation at 37 • C under 5 % CO 2 . 200 μl Dimethyl sulfoxide (DMSO) was added into each well after the DMEM was removed. Finally, optical density values were measured at wavelength of 570 nm. In all viability experiments, 1 μM and 10 μM Rotenone dissolved in DMSO was used as positive control (Avallone et al., 2020;Günaydin et al., 2021).

PTEN expression and inflammatory response in BV2 microglial cells
To study the effect of pCS and 4EPS on the release of TNF-α and IL-6 or on the expression of PTEN, COX-2 and iNOS in cells, 5000 BV2 cells/ well were seeded in a 96-well plate (3599, Corning, NY, USA) or 50,000 BV2 cells/well in a 12-well plate (3512, Corning, NY, USA), respectively. On the following day, the cells were incubated with pCS and 4EPS concentrations in the presence or absence of 1000 ng/ml LPS stimulation (L3024, Sigma). After 24 h of incubation, the medium in 96-well plate was collected for measurements of TNF-α and IL-6 by enzymelinked immunosorbent assay (ELISA). The BV2 cells in 12-well plate were lysed in RIPA lysis buffer containing Proteinase inhibitor (1:200), 5 μM GI254023X and 10 mM 1,10-Phenanthroline for WB analysis. For experiments to study the effect of PTEN inhibition on the inflammatory response of microglial cells, 5000 BV2 cells/well were seeded into 96well plate overnight. On the following day, BV2 cells were incubated with a potent PTEN inhibitor VO-OHpic trihydrate (V8639, Sigma) (Mak et al., 2010) for 24 h in the absence or presence of 1000 ng/ml LPS stimulation. Finally, the medium was used for measurements of inflammatory cytokines TNF-α and IL-6 using ELISA.

Phagocytosis activity assay
50,000 BV2 cells/well were seeded into 96-well plate (3599, Corning, NY, USA). To assess the effect of 4EPS on phagocytosis, BV2 cells were immediately incubated with 4EPS concentrations (0.1, 1, 10 and 100 (μM)) for 24 h in the presence or absence of 1000 ng/ml LPS stimulation. To study the effect of PTEN inhibition on constitutive and LPS-stimulated microglial phagocytosis, BV2 cells were immediately incubated with VO-OHpic trihydrate concentrations (50, 100, 200 and 400 (nM)) for 24 h with or without 1000 ng/ml LPS. Next the phagocytic effect was measured with Vybrant™ Phagocytosis Assay Kit (V6694, Thermo Scientific) according to manufacturer's instructions. Briefly, the medium was discarded completely after 24 h incubation, and cells were incubated with K-12 strain bioParticles for 2 h at 37 • C, followed by Trypan Blue solution incubation for 1 min. Finally, the fluoresce intensity of BV2 cells was measured using ~480 nm excitation, ~520 nm emission.

Statistics
All data analyses and statistics were performed using GraphPad Prism (version 9.1.1; GraphPad software, La Jolla, CA, USA). The in vivo results were analyzed by two-tailed Student's test. Two-tailed Pearson correlation was used for correlation analysis of E-cadherin and iNOS expression in the distal ileum of mice. For multiple comparisons of the in vitro data, one-way ANOVA was used, followed by Dunnett's multiple comparisons as post-hoc test. All data are shown as mean ± SEM or mean ± SD; P < 0.05 is considered to be statistically significant.

PTEN expression and PTEN phosphorylation did not change in the distal ileum of male mice after in utero exposure to VPA
Previous studies have shown intestinal inflammation in VPA-induced murine model of ASD (de Theije et al., 2014). COX-2 expression was significantly increased to six folds in the distal ileum of in utero VPAexposed mice compared to control mice, indicating a VPA-induced intestinal inflammatory response ( Fig. 1A & B). Furthermore, membranebound E-cadherin was significantly decreased in the distal ileum of in utero VPA-exposed mice compared to control mice ( Fig. 1C & D). The decreased membrane-bound E-cadherin leads to increased C-terminal fragment of E-cadherin levels (CT E-cadherin). Indeed, we found a trend of increased CT E-cadherin ( Fig. 1E P = 0.083) and significantly increased ratio of CT E-cadherin to membrane-bound E-cadherin (Fig. 1F). It has been shown that iNOS co-localizes with adhesive protein E-cadherin, indicating E-cadherin might regulate iNOS expression or vice versa (Frank and Hostetter, 2007;Glynne et al., 2002;Terciolo et al., 2017). The decreased E-cadherin expression was associated with a decreased iNOS expression in the distal ileum of in utero VPA-exposed mice compared to control mice ( Fig. 1G & H (P = 0.056). Furthermore, the Pearson correlation analysis shows a strong correlation between Ecadherin expression and iNOS expression (r = 0.87, P = 0.024, Fig. 1I). Next, we assessed whether PTEN expression or PTEN phosphorylation is changed in the inflamed distal ileum of in utero VPA-exposed male mice. As shown in Fig. 1K & M, both PTEN and phosphorylated PTEN expression show no differences between in utero VPA-exposed and control mice.

In utero exposure to VPA attenuated PTEN expression in hippocampus, without effects in other brain regions in male mice
PTEN expression significantly decreased in the hippocampus of in utero VPA-exposed male mice compared to control mice ( Fig. 2A & B). In contrast, PTEN expression of in utero VPA-exposed mice was not significantly changed compared to control mice in the prefrontal cortex (supplementary fig. 1A & B), cerebellum (supplementary fig. 1C & D), olfactory bulb (supplementary fig. 1E & F), and in the rest of other mice brain regions (supplementary fig. 1G & H).
These in vivo results indicate that in utero exposure to VPA down regulates PTEN expression in the hippocampus but not affect PTEN expression in the distal ileum in male mice. Elevated levels of the intestinal bacteria-derived metabolite pCS in serum of in utero VPAexposed male mice has been shown in previous studies (Zheng et al., 2022). In the present study, we hypothesize that the intestinal bacteriaderived pCS and 4EPS induce ASD-like behaviors via affecting PTEN expression of microglial cells in the brain in this murine model for ASD. Therefore, the direct effect of pCS and 4EPS on PTEN expression in BV2 microglial cells and the possible association with neuro-immune responses and microglial phagocytosis function are investigated.

pCS and 4EPS attenuated constitutive and LPS-activated PTEN expression in BV2 microglial cells
Previously we demonstrated that 24-h exposure of BV2 microglial cells up to a concentration of 500 μM pCS did not affect their viability (Zheng et al., 2022). In the present study, we show that 4EPS up to a concentration of 400 μM did also not have toxic effects on BV2 microglial cells after 24 h incubation using two cell viability assays (MTT and LDH) (supplementary fig. 2). The direct effects of pCS and 4EPS on PTEN expression in BV2 microglial cells were assessed. As shown in Fig. 3A-D, 5 to 150 μM pCS,  The quantification of hippocampal PTEN expression. β-ACTIN was used as loading control. n = 3 for in utero PBS-or VPA-exposed male mice. Results are expressed as mean ± SD. *P < 0.05. The quantification of PTEN expression of microglial cells exposed to pCS (n = 4 from 4 independent experiments). (C) The representative immunoblots of PTEN with 4EPS treatment. (D) The quantification of PTEN expression of microglial cells exposed to 4EPS treatment (n = 6 from 6 independent experiments). Calnexin was used as loading control. Black: control; red: control LPS-activated microglial cells; blue: constitutive active microglial cells; green: LPS-activated microglial cells. Results are shown as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.
Next, the effect of a PTEN inhibitor VO-OHpic (50, 100, 200 and 400 nM) on the constitutive and LPS-induced release of TNF-α and IL-6 by microglial cells was investigated (Mak et al., 2010). As Fig. 4E & F show, VO-OHpic neither affected the constitutive nor the LPS-induced releases of TNF-α and IL-6 by microglial cells, indicating that PTEN is not involved in the constitutive or LPS-induced release of TNF-α and IL-6 by microglial cells. In summary, these results indicate that the pCS and 4EPS-induced inhibition of PTEN expression is not involved in the pCS and 4EPS-induced effects on constitutive and LPS-induced release of TNF-α and IL-6 of microglial cells.
In addition to the inflammatory cytokines TNF-α and IL-6, iNOS and COX-2 play important role in neuroinflammatory response in brain (Aïd and Bosetti, 2011;Sonar and Lal, 2019). It has been shown previously that LPS stimulation upregulates the protein expressions of these two inflammatory targets in BV2 microglial cells (Nam et al., 2018;Yang Fig. 4. The effects of pCS, 4EPS and PTEN inhibitor VO-OHpic on TNF-α and IL-6 release from BV2 microglial cells. BV2 microglial cells were seeded with a density of 5000 each well into 96-well plate, then exposed to pCS, 4EPS, and VO-OHpic for 24 h in the absence or presence of 1000 ng/ml LPS, respectively, culture medium was collected for ELISA measurements. (A, B) The TNF-α and IL-6 concentration in medium after pCS treatment. (n = 6 and 4 from 6 and 4 independent experiments respectively). (C, D) The TNF-α and IL-6 concentration in medium after 4EPS treatment. (n = 3 and 4 from 3 and 4 independent experiments respectively). (E, F) The TNF-α and IL-6 concentration in medium after VO-OHpic treatment. (n = 4 and 3 from 4 and 3 independent experiments respectively). Black: control; red: control LPS-activated microglial cells; blue: constitutive active microglial cells; green: LPS-activated microglial cells. Results are shown as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. et al., 2020), but there is little known about the effect of pCS and 4EPS on the expression of iNOS and COX-2. Therefore, the same in vitro model of BV2 microglial cells was further used to investigate the effect of pCS and 4EPS on iNOS and COX-2 expression in BV2 microglia.

pCS and 4EPS inhibited LPS-activated iNOS expression in BV2 microglial cells
iNOS is an inflammation-induced protein and is barely constitutively expressed in microglial cells (Nakazawa et al., 2017;Zamora et al., 2000). As shown in Fig. 5A-D, LPS stimulation strongly increased iNOS expression in microglial cells compared to control, which can be attenuated by co-incubation with pCS (50 μM) and 4EPS (10 and 100 μM), respectively. These results further support that pCS and 4EPS can attenuate LPS-induced neuroinflammation in microglial cells.

pCS and 4EPS were unable to affect LPS-activated COX-2 expression. 4EPS alone increased constitutive COX-2 expression in BV2 microglial cells
LPS stimulation significantly increased COX-2 expression in microglial cells (Fig. 6A-D). Neither pCS nor 4EPS treatment affected LPSinduced COX-2 expression. To better detect COX-2 bands without LPS stimulation, microglial cell lysates after pCS or 4EPS exposures were exposed longer separately. As shown in Fig. 6E & F, no clear difference was observed in relation to the constitutive COX-2 expression in microglial cells exposed to 5, 10, 50 or 150 μM pCS compared to control. In contrast, 1, 10 and 100 μM 4EPS significantly increased COX-2 expression in microglial cells compared to control without LPS stimulation ( Fig. 6G & H). These results indicate that the bacterial metabolites pCS and 4EPS do not affect LPS-activated COX-2 expression, but 4EPS can increase constitutive COX-2 expression of microglial cells.

4EPS or PTEN inhibitor VO-OHpic did not affect constitutive and LPS-activated microglial phagocytosis activity
Previous results have shown that pCS treatment attenuated constitutive phagocytosis activity in microglial cells without affecting LPSinduced microglial enhanced phagocytosis activity (Zheng et al., 2022). In the present study, the effects of 4EPS or PTEN inhibitor VO-OHpic on microglial phagocytosis were studied. Confirming the previous finding, LPS stimulation significantly increases microglial phagocytosis activity (Fig. 7A & B). As shown in Fig. 7A, 4EPS treatments from 0.1 μM to 100 μM did not affect constitutive or LPS-induced microglial phagocytosis. In addition, 50 nM to 400 nM VO-OHpic also did not affect microglial phagocytic activity with or without LPS stimulation (Fig. 7B), indicating that 4EPS and PTEN do not play a role in microglial phagocytosis.

Discussion
The present study investigated the relation between two bacteriaderived metabolites, pCS and 4EPS, and PTEN in the pathogenesis of ASD along the gut-brain axis using an in utero VPA-induced murine model of ASD and an in vitro LPS-activated microglial neuroinflammation model. In this current study, it was demonstrated that PTEN expression was not changed in the distal ileum of in utero VPAexposed male mice, which is associated with enhanced levels of serum pCS. The observation that PTEN expression is reduced in the brain of in utero VPA-exposed male mice directed us to investigate the role of PTEN regarding the effects of pCS and 4EPS on microglial cell function in vitro. The decreased PTEN expression in the brain in the murine model for ASD is mirrored by pCS and 4EPS-induced decrease in PTEN in microglial cells. However, this pCS and 4EPS-induced reduction in PTEN expression does not appear to play a role in the bacteria-derived metabolite-induced derailed innate immune and phagocytotic response of microglial cells.
Increased intestinal permeability and intestinal inflammation frequently occur in children with ASD as well as in rodent models for ASD (Babinská et al., 2017;Kang et al., 2017;Ristori et al., 2019;Walker et al., 2019). Seike et al. have previously shown that Clostridioides perfringens Delta-toxin causes increased intestinal permeability or epithilial permeability through decreaseing E-cadherin expression in mouse ileal loop or epithelial Caco-2 cells, respectively (Seike et al., 2019;Seike et al., 2018). In the present study, it was demonstrated that in utero exposure to VPA decreased E-cadherin expression in distal ileum, indicative of increased intestinal permeability, which is also supported   by the enhanced intestinal permeability reported in utero VPA-exposed rats . Glynne et al. have shown that E-cadherin colocalizes with epithelial iNOS (Glynne et al., 2002). In addition, decreased iNOS activity or expression increases intestinal permeability through attenuating the expression of tight junctional proteins in mouse ileal tissue (Han et al., 2004). These previous studies support the decreased iNOS expression that was observed in the distal ileum of in utero VPA-exposed male mice in the current study. In addition, it was demonstrated that the distal ileum of in utero VPA-exposed male mice was inflamed as indicated by enhanced COX-2 protein expression that is an indicator of intestinal and epithelial inflammation (Golden et al., 2020;Wang and Dubois, 2010;Wu et al., 2019), which is supported by previous study showing the presence of inflammation in small intestine of in in utero VPA exposed mice (de Theije et al., 2014). Next, PTEN and p-PTEN expression in the same ileal tissue were examined. We found that in utero exposure to VPA did not to affect (p-)PTEN expression in ileum. These data suggest that PTEN is not relevant for the compromised barrier and intestinal inflammation in the ileum of in utero VPA-exposed male mice.
Both pCS and 4EPS are found to be elevated in urine, feces or serum of children diagnosed with ASD and can induce ASD-like behaviors in mice (Bermudez-Martin et al., 2021;Gabriele et al., 2014;Hsiao et al., 2013;Kang et al., 2020;Kang et al., 2018;Needham et al., 2021;Needham et al., 2022;Pascucci et al., 2020). Furthermore, our previous study has shown an increase in serum pCS levels of in in utero VPAexposed male mice (Zheng et al., 2022). pCS has been detected in mouse brain tissue (Sun et al., 2020;Zgoda-Pols et al., 2011) and cerebrospinal fluid of parients with Parkinson's disease (Sankowski et al., 2020). In addition, 4EPS has also been demonstrated to be able to reach the brain in mice (Hsiao et al., 2013). In in utero VPA-exposed mice, the present study shows that hippocampal PTEN expression is decreased, which is in line with the previous finding that PTEN expression is decreased in hippocampus and cortex of in in utero VPA-exposed male mice (Mahmood et al., 2018). Decreased PTEN expression in brain might induce ASD-like behaviors, including impaired social interaction and increased repetitive behavior, which has also been demonstrated previously with PTEN-deficient mice (Lugo et al., 2014;Mahmood et al., 2018). However, it is not clear whether pCS or 4EPS derailes the neuroimmune responses of microglial cells via the induction of decreased PTEN in the hippocampus of in utero VPA-exposed male mice. This study showed that both pCS and 4EPS attenuated PTEN expression in BV2 micorglial cells constitutively and during LPS stimulation, indicating that these two ASD-associated bacterial derivates might be able to decrease PTEN expression directly in brain.
Next, we studied whether PTEN plays a role in the bacteria-derived metabolite-induced dysfunction of microglial cells in vitro. Both pCS and 4EPS attenuate constitutive and LPS-induced TNF-α and IL-6 release by BV2 microglial cells, which is supported by the finding that pCS attenuates LPS-induced inflammation in macrophages (Murakami et al., 2014;Shiba et al., 2016). However, PTEN inhibition did not affect the release of these cytokines in microglial cells, indicating that the pCS-and 4EPS-induced decreased PTEN expression does not play a role in the inhibitory effects of pCS and 4EPS on TNF-α and IL-6 release of by microglial cells. Interestingly, the present study showed that 4EPS, but not pCS, significantly increased constitutive COX-2 expression in microglial cells indicative for prostaglandin-associated inflammatory response. In contrast, COX-2 deficient mice show ASD pathogenesis and ASD-related behaviors (Kissoondoyal et al., 2021;Rai-Bhogal et al., 2018;Wong et al., 2019). The possible role of increased COX-2 expression triggered by 4EPS in microglial cells in the context of ASD remains to be investigated. Microglial phagocytosis is necessary to control synaptic function and brain development via regulating synaptic pruning and neuro-immune response (Irfan et al., 2022;Lenz and Nelson, 2018). In the present study, we found that neither 4EPS nor PTEN inhibition affected BV2 microglial phagocytosis activity. PTEN deficient mice show upregulated neuroinflammation and microglial phagocytosis (Sarn et al., 2021a;Sarn et al., 2021b). In contrast, it has also been reported that PTEN inhibition or knockdown attenuates neuroinflammation in mice (Pan et al., 2022;Shen et al., 2021). This study reports that PTEN did not play a role in ASD-associated bacterial metabolite-induced derailed immune and phagocytosis response of microglial cells. Additionally, PTEN is essential to maintain neuronal growth, synaptic function and brain connectivity (Skelton et al., 2019;Spina Nagy et al., 2021). Therefore, although beyond the scope of the present study, the role of the reduced PTEN expression observed in the brain of in utero VPA-exposed male mice, in the derailed neuronal network development and function remains to be further investigated.
In summary, this study demonstrated that both pCS and 4EPS decreased PTEN expression directly in microglial cells, which mirrors the decreased PTEN expression observed in the hippocampus of in utero VPA-exposed male mice. The current findings indicate that pCS and 4EPS-induced down-regulated PTEN does not play a role in ASDassociated neuroinflammation in the brain. Future studies are warranted in order to investigate how these two bacteria-derived metabolites affect neuronal synaptic function in vitro primary neurons and in in vivo ASD murine models with a focus on PTEN. Besides, It would be interesting that future experiments investigate whether p-cresol and 4ethylphenol treatment can affect intestinal permeability in colonic epithelial cells and mice with a focus on PTEN. In addition, future studies colonizing mice with p-cresol and 4-ethylphenol producing bacteria would be promising approach to investigate their role in ASDassociated detrimental symptoms.

CRediT authorship contribution statement
Conceptualization-YZ, PPP, ADK; supervision-PPP, ADK, JG; investigation and data collection-YZ, NP, CVH; data analysis-YZ; writing-original draft preparation: YZ; writing-review and editing: YZ, NP, CVH, PPP, JG, ADK. All authors have read and agreed to the published version of the manuscript.

Declaration of competing interest
All authors have seen and approved the final version of the manuscript being submitted. Johan Garssen is a part time employee at Danone Nutricia Research, Utrecht, The Netherlands. All other authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Data availability
Data will be made available on request.

Acknowledgement
We would like to thank Prof. Dr. Ulrich L.M. Eisel for kindly donating mouse BV2 microglial cells.

Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi. org/10.1016/j.mcn.2022.103805.