Parasympathetic cholinergic innervation of the liver

In our previous study [16], we showed that a subset of DMV cholinergic neurons projected to the liver. To extend and further confirm the presence of DMV cholinergic neurons innervating the liver, we injected a Cre-dependent viral tracer (adeno-associated virus (AAV).PHP.eB-FLEX-tdTomato [17]) into the liver of choline acetyltransferase (ChAT)-IRES-Cre (ChAT^Cre) mice [18]. The ChAT^Cre transgenic strain has been used to selectively label parasympathetic cholinergic neurons [19]. At 3 weeks post-viral infection, we performed double immunocytochemistry with anti-ChAT and anti-tdTomato antibodies on the brain sections and found that the DMV displayed a small number of tdTomato-positive cells (Fig 1A). Importantly, all tdTomato-positive cells in the DMV were ChAT-expressing cells (n = 72 out of 72 neurons from 3 mice), and approximately 10% of DMV cholinergic neurons expressed tdTomato (range, 7% to 14%; n = 72 out of 660 cholinergic neurons from 3 mice).

We subsequently examined whether these DMV cholinergic neurons send axonal projections to the liver by employing a retrograde AAV that encodes a Cre-dependent placental alkaline phosphatase (PLAP), which is expressed on the plasma membrane (AAVrg-FLEX-PLAP). Injection of this AAVrg-FLEX-PLAP into the liver of ChAT^Cre mice resulted in the presence of PLAP-positive nerves in the parenchyma of the liver, particularly in the periportal and midlobular areas (Fig 1B). Immunostaining with an antibody against the presynaptic cholinergic marker vesicular acetylcholine transporter (vAChT) revealed the presence of vAChT-positive nerve terminals scattered throughout the liver parenchyma, but higher levels of vAChT-positive terminals were observed around the periportal zones (Fig 1C). In addition, the 3D projection view of the cleared liver tissue from the ChAT^Cre;channelrodopsin (ChR2)-tdTomato mice showed the presence of hepatocytes receiving tdTomato-positive puncta-like nerve terminals (S1 Movie). Consequently, our findings align with prior studies indicating that a subset of DMV cholinergic neurons innervate the liver [16,20,21].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Parasympathetic cholinergic innervation of the mouse liver.

(A) Schematic illustration of our experimental conditions (upper panel). Bottom panel: Images of confocal fluorescence microscopy showing co-expression of tdTomato and ChAT in a subset of DMV cholinergic neurons in ChAT^Cre mice receiving an injection of AAV.Php.eB-tdTomato into the liver, representing cholinergic neurons innervating the liver (white arrowheads). Scale bar, 30 μm. (B) Schematic illustration of our experimental conditions (upper panel). Bottom panel: Images showing PLAP-positive nerve fibers in the liver parenchyma. Right panel: Higher-magnification view of the black square box in the left panel. Scale bars, 100 μm (left), 20 μm (right). (C) Images of confocal fluorescence microscopy showing the expression of vAChT-positive nerve terminals (green) in the liver parenchyma (inset: higher-magnification view of vAChT-positive terminals). Scale bar, 70 μm. (D) Schematic illustration of our experimental conditions. (E) Images of confocal fluorescence microscopy showing tdTomato-positive cholinergic neurons in the DMV (upper panel; white arrowheads) and sypGFP-positive nerve terminals (bottom panel; white arrowheads) in the liver of ChAT^Cre mice injected with AAV1-FLEX-tdTomato-T2A-sypGFP into the DMV. A small number of hepatocytes and cholangiocytes received parasympathetic cholinergic innervation (inset). Blue: nucleus staining with DAPI. Upper panel: scale bar, 30 μm; lower panel: left, scale bar, 20 μm; right, 40 μm. (F) Images of confocal fluorescence microscopy showing that SOX9-positive cholangiocyte cells receive vAChT-positive nerve terminals (white arrowheads). Scale bar: 10 μm. (G) Image of the liver after tissue clearing. (H) Z-stack image showing parasympathetic GFP-positive cholinergic nerve terminals (green) around the periportal area in cleared liver tissue collected from ChAT^Cre mice injected with AAV1-FLEX-tdTomato-T2A-sypGFP (upper panel). Bottom panel: 2D image of GFP-positive cholinergic nerve terminals shown in the upper panel. White arrowheads indicate the cholinergic nerve terminals. Scale bar: 50 μm. AAV, adeno-associated virus; PLAP, placental alkaline phosphatase; vAChT, vesicular acetylcholine transporter.

https://doi.org/10.1371/journal.pbio.3002865.g001

Given that DMV cholinergic neurons projected to the liver, we injected an anterograde AAV1 encoding Cre-dependent tdTomato and a synaptophysin-GFP (sypGFP) fusion protein separated by the sequence of a T2A self-cleaving peptide (AAV1-FLEX-tdTomato-T2A-sypGFP [22]) into the DMV of ChAT^Cre mice (Fig 1C). Under these experimental conditions, we anticipated that tdTomato would be expressed in DMV cholinergic neurons and that sypGFP would move forward to the axon terminals of DMV cholinergic neurons. As shown in Fig 1E, immunostaining revealed that the DMV cholinergic neurons were positive for tdTomato, indicating that the DMV cholinergic neurons were successfully infected. Immunolabeling of liver sections further revealed numerous sypGFP-positive nerve terminals in a subset of hepatocytes and bile duct epithelial cells cholangiocytes (Fig 1E and 1F). SRY-Related HMG Box Gene 9 (SOX9)-positive cholangiocytes were surrounded by vAChT-positive nerve terminals (Fig 1F). 3D imaging analysis of cleared liver tissue collected from ChAT^Cre mice injected with AAV1-FLEX-tdTomato-T2A-sypGFP further confirmed the presence of cholinergic nerve terminals around the portal vein and in a subset of hepatocytes (Fig 1G and 1H and S2 Movie), consistent with a recent findings of parasympathetic cholinergic innervation of the bile duct in human liver tissues [15].

In addition to cholinergic input, RT-qPCR of liver tissue homogenates revealed the presence of muscarinic ACh receptor types 1, 2, 3, 4, and 5 (Chrm1, 2, 3, 4, and 5) (S1A Fig) consistent with our prior study [16]. Western blot analysis of liver tissue homogenates showed the expression of CHRM1, 2, 4, and 5, but not CHRM3 (S1B and S1C Fig). As CHRM4 protein expression was highly expressed compared to other CHRM subtypes, we immunolabeled the liver tissues with an anti-CHRM4 antibody. CHRM4 was detected in areas adjacent to vAChT-positive terminals (S1D Fig). Similarly, CHRM2 was also found in hepatocytes (S1E Fig). Taken together, our analyses of cleared liver tissues using 3D imaging techniques, as well as antero- and retrograde neural tracing robustly support the interpretation that the liver is innervated by parasympathetic cholinergic neurons.

Liver-innervating parasympathetic cholinergic neurons are indispensable for the development of hepatic steatosis during an HFD feeding

The role of DMV cholinergic neurons innervating the liver in controlling hepatic lipid metabolism remains unexplored. Consequently, we examined the physiological impact of hepatic cholinergic loss-of-function on hepatic fat accumulation. First, to selectively target liver-projecting cholinergic neurons, we used a retrograde AAV (AAVrg) encoding a Cre-dependent diphtheria toxin subunit A fragment (dtA) that induces apoptosis of target cells [23]. AAVrg-FLEX-dtA and AAVrg-FLEX-tdTomato as control were injected into the livers of the ChAT^Cre mice (Fig 2A). To validate cholinergic cell death in the DMV, we sacrificed mice at 10 weeks post-viral inoculation and stained the DMV cholinergic neurons with an anti-ChAT antibody. The number of DMV cholinergic neurons was significantly lower in ChAT^Cre mice that received retroAAV-FLEX-dtA than in the controls (Figs 2B, 2C, S2A, and S2B). This reduced number of DMV cholinergic neurons was strongly associated with a lack of vAChT-positive nerve terminals in the livers of ChAT^Cre mice that received AAVrg-FLEX-dtA (Fig 2D). Importantly, measurement of ACh levels revealed a significant reduction in hepatic ACh content in the experimental mice relative to controls, whereas there was no difference in plasma ACh levels between the experimental and control groups (Fig 2E and 2F). These results support the interpretation that our experimental approach effectively ablates parasympathetic cholinergic neurons innervating the liver.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Loss of parasympathetic cholinergic neurons innervating the liver.

(A) Schematic illustration of the experimental configurations, where AAVrg-FLEX-tdTomato (control, ο) and AAVrg-FLEX-dTA (•) viral vectors were injected directly into the livers of ChAT^Cre mice. (B) Images of confocal fluorescence microscopy showing reduced numbers of cholinergic neurons in the DMV of ChAT^Cre mice receiving an injection of AAVrg-FLEX-dtA compared with the controls. Scale bar, 80 μm, cc: central canal, 12N: hypoglossal nucleus. (C) Summary graph showing the number of cholinergic neurons on the left and right sides of the DMV in the control (n = 4 mice) and experimental mice (n = 4 mice). Unpaired t test, **p < 0.01. (D) Images showing a lack of vAChT-positive nerve terminals in the liver parenchyma of ChAT^{Cre + AAVrg-FLEX-dtA} mice (right panel). Left panel: control. Scale bar, 50 μm. (E, F) Summary plots showing a significant decrease in hepatic (control, n = 7 mice; experimental mice, n = 6 mice), but not plasma (control, n = 11 mice; experimental mice, n = 8 mice), ACh contents in ChAT^{Cre + AAVrg-FLEX-dtA} mice compared to the controls. Unpaired t test, *p < 0.05. The data supporting the graphs shown in the figure (Fig 2C, 2E, and 2F) are available in the S1 Data file. AAV, adeno-associated virus; vAChT, vesicular acetylcholine transporter.

https://doi.org/10.1371/journal.pbio.3002865.g002

We then examined whether deleting liver-innervating cholinergic neurons alters hepatic lipid metabolism during HFD feeding. To address this question, we first investigated if there are differences in the liver histology between the groups. Liver tissues were labeled with liver zonation markers such as the pericentral marker glutamine synthetase (GS) and the periportal marker E-cadherin (E-cad) [24]. The expression pattern of these zonation markers in the livers of the control group fed a standard chow diet demonstrated that GS- and E-cad-positive areas were clearly separated in the liver (S3A Fig). Interestingly, high-fat feeding robustly reduced E-cad expression in periportal hepatocytes in the livers of control mice compared to control mice fed a standard chow diet (Fig 3A), consistent with disrupted E-cad expression in chronic liver disease in mice [25]. Surprisingly, deleting the parasympathetic cholinergic input to the liver in ChAT^Cre mice fed HFD restored the expression of E-cad in periportal hepatocytes (Fig 3B). In contrast, GS was similarly expressed in pericentral hepatocytes in both control and experimental mice fed HFD (Fig 3A and 3B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Deleting liver-innervating cholinergic neurons prevents hepatic steatosis during HFD feeding.

(A, B) Macroscopic appearance of livers of ChAT^Cre mice receiving AAVrg-FLEX-tdTomato and AAVrg-FLEX-dtA. E-Cad expression was lower in the control group than in the experimental group. Scale bar, 100 μm. (C, D) HE staining of liver tissues revealed increased numbers of lipid droplets and ballooned hepatocytes in the controls compared to those in the experimental mice (top panel). Scale bar, 100 mm. Bottom panel: Higher magnification view of liver tissue in the black dotted square. Scale bar, 30 μm. (E, F) Oil red O staining (red) of liver tissues from ChAT^Cre mice with and without hepatic cholinergic input. The experimental group did not exhibit fat accumulation in the liver parenchyma during high fat feeding. Nuclear staining (blue): hematoxylin. Scale bar, 100 μm. (G) Graphs showing hepatic (n = 9 vs. 8 mice) and plasma (n = 11 vs. 12 mice) TG levels in the controls and experimental mice. There was a significant reduction in both hepatic and plasma TG levels in the experimental group relative to those in the control group (unpaired t test, **p < 0.01, ***p < 0.001). (H) Summary graphs showing changes in hepatic mRNA expression of enzymes involved in lipid metabolism. Srebp1c mRNA expression was significantly down-regulated, while Ppara mRNA expression was increased in the livers of ChAT^Cre mice receiving AAVrg-FLEX-dtA relative to controls (control, n = 6–8 mice; experimental, n = 6–8 mice, unpaired t test, *p < 0.05, **p < 0.01). The data supporting the graphs shown in the figure (Fig 3G and 3H) are available in the S1 Data file. AAV, adeno-associated virus; HE, hematoxylin and eosin; HFD, high-fat diet.

https://doi.org/10.1371/journal.pbio.3002865.g003

Hematoxylin and eosin (HE) staining revealed a clear difference in the macroscopic appearance of the livers of ChAT^Cre mice receiving AAVrg-FLEX-tdTomato and AAVrg-FLEX-dtA. Control mice exhibited an increased number of large and small lipid droplets and ballooned hepatocytes compared to experimental mice (Fig 3C and 3D). Oil Red O staining further showed that ChAT^Cre mice without DMV cholinergic neurons innervating the liver contained fewer lipid droplets in the liver than controls (Fig 3E and 3F). These results suggest that parasympathetic cholinergic input to the liver limits hepatic steatosis during HFD feeding.

As there was a lack of lipid droplets in the livers of the experimental mice, we measured the hepatic and plasma triglyceride (TG) levels. Following HFD feeding, ChAT^Cre mice without parasympathetic cholinergic neurons innervating the liver on high-fat feeding exhibited reduced hepatic and plasma TG levels in the liver relative to controls (Fig 3G). This was associated with decreased plasma TG levels (Fig 3G). Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) analysis of liver tissues revealed a significant decrease in the expression of the lipogenic transcription factor sterol regulatory element-binding protein 1-c (Srebp-1) in the experimental mice relative to the controls. The experimental mice also exhibited down-regulation of its target genes, including acetyl-coenzyme A carboxylase 1 (Acaca), fatty acid synthase (Fasn), stearoyl-CoA desaturase-1 (Scd1) (Fig 3H). In addition, the ChAT^Cre mice without parasympathetic cholinergic input to the liver had increased expression of peroxisome proliferator-activated receptor alpha (PPAR-α), the central transcriptional regulator of fatty acid oxidation relative the control group. As a result, there was increased carnitine palmitoyltransferase 1a (Cpt1a) and acyl-CoA dehydrogenase medium chain (Acadm) mRNA expression in the experimental mice relative to controls (Fig 3H). In contrast to male mice, high-fat feeding did not cause any changes in liver histology or fat accumulation in both control and experimental nhigh-fat feeding [26–28].

Hepatic vagal denervation results in an increase in hepatic interleukin 6 (Il6) gene expression in mice, which subsequently modulates the expression of hepatic gluconeogenic enzymes [29]. In our experiments, where we selectively deleted liver-projecting cholinergic neurons, we observed a significant decrease in Il6 gene expression (S5A Fig). Notably, there was no significant difference in gluconeogenic gene expression between the experimental and control groups (S5B Fig). Furthermore, we investigated whether the loss of parasympathetic cholinergic innervation affects sympathetic input to the liver. We used immunostaining with an anti-tyrosine hydroxylase (Th) antibody and found no detectable difference in Th-positive nerve fibers (S6A and S6B Fig). Additionally, hepatic norepinephrine levels were similar between the groups (S6C Fig). Taken together, these findings suggest that the loss of parasympathetic cholinergic neurons innervating the liver prevents the development of hepatic steatosis during HFD feeding in male mice.

The male control mice displayed dense lipid accumulation in the pericentral areas compared to the experimental mice. Therefore, we further examined if there are distinct gene expression patterns throughout the hepatic zonation in both groups by performing spatial transcriptomics (S7 Fig). Pericentral, midlobular, and periportal clusters were identified based on known zonation markers (S7A Fig). We found that the midlobular areas exhibited 391 differentially expressed genes (DEGs) (S7B Fig and S1 Table) and that 271 genes in the PC and 550 genes in the PP areas were significantly different between the groups (S7B Fig and S2 and S3 Tables). Importantly, the loss of liver-projecting cholinergic neurons down-regulated hepatic lipogenesis-related genes, including Scd1, Fabp4, Cd36, Lpin1, Nampt, Nr1l2, and Me1 and up-regulated the Trib1 gene across the liver zonation (S7C Fig).

A selective gene ontology (GO) analysis of the DEGs in each zonation using Enrichr-KG [30] identified the top 5 enriched GO terms in biological process (S7D Fig). These included “cotranslational protein targeting to membrane” (GO:0006613), “SRP-dependent cotranslational protein targeting to membrane” (GO:0006614), “protein targeting to ER” (GO:0045047), “cytoplasmic translation” (GO:0002181), and “cellular protein metabolic process” (GO:0044267). According to the mammalian phenotype ontology analysis [31], each zonation shared 3 terms, including “abnormal lipid homeostasis” (MP:0002118), “abnormal liver physiology” (MP:0000609), and “hepatic steatosis” (MP:0002628). These results further support that liver-projecting cholinergic neurons play an important role on developing hepatic steatosis in mice fed HFD.

Loss of liver-innervating cholinergic neurons causes beiging of ingWAT

One of the hallmarks of diet-induced obesity is a set of increases in inguinal WAT depot size, accompanied by increased adipocyte cell size [32]. Importantly, we found that ChAT^Cre mice without liver-innervating cholinergic neurons had reduced ingWAT size relative to control mice (Fig 4A). Histological analysis of ingWAT revealed smaller adipocytes and decrease adipocyte area in experimental mice compared to the control mice (Fig 4B–4D). These results suggest that a loss of DMV cholinergic input to the liver prevents ingWAT hypertrophy.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Deleting liver-innervating cholinergic neurons induces beiging of ingWAT.

(A) Representative images of the ingWAT of control and experimental mice. (B) Representative images of adipocytes in the ingWAT of control and experimental groups. Scale bar, 100 μm. (C, D) Plots showing the distribution of the average adipocyte size (left panel). Right panel: The mean cell area was significantly smaller in the experimental groups than in controls (control, n = 4 mice; experimental, n = 5 mice, unpaired t test, *p < 0.05). (E) Summary graph showing changes in thermogenic gene expression in ingWAT of ChAT^Cre mice with and without DMV cholinergic neurons innervating the liver. Ablation of liver-projecting cholinergic neurons significantly up-regulated the thermogenic gene expression (unpaired t test, *p < 0.05, **p < 0.01). (F, G) Western blot images of ingWAT showing an increase in UCP1 expression in experimental mice. Images of confocal fluorescence microscopy showing increased UCP1 expression in the ingWAT of experimental mice (G). Scale bar, 50 μm. (H) Summary plots showing ingWAT temperature in the light and dark phases (control, n = 5 mice; experimental, n = 4 mice, unpaired t test, **p < 0.01). (I) Western blot images showing a decrease in pS660-HSL expression in ingWAT of experimental mice relative to controls. (J) Summary plot showing plasma glycerol levels in the control (n = 8 mice) and experimental mice (n = 8 mice, unpaired t test, *p < 0.05). The data supporting the graphs shown in the figure (Fig 4C, 4D, 4E, 4H, and 4J) are available in the S1 Data file. AAV, adeno-associated virus.

https://doi.org/10.1371/journal.pbio.3002865.g004

RT-qPCR analysis revealed that deleting DMV cholinergic neurons innervating the liver significantly up-regulated the expression of Peroxisome proliferator-activated receptor-gamma coactivator (PGC)-1alpha (PGC1a) in ingWAT, which is required for the expression of mitochondrial biogenesis and thermogenic genes in brown fat cells [33,34] (Fig 4E). Likewise, the expression of PR (PRD1-BF1-RIZ1 homologous)-domain containing 16 (PRDM16), a transcription factor that plays a key role in browning and thermogenesis of subcutaneous WAT [35], was significantly increased in experimental mice relative to controls (Fig 4E). The up-regulation of these genes was associated with a robust increase in the mRNA expression of uncoupling protein 1 (Ucp1), iodothyronine deiodinase 2 (dio2), and cell death-inducing DFFA-like effector A (Cidea) (Fig 4E). Western blotting and immunostaining further showed an increase in UCP1 protein levels in ingWAT of ChAT^Cre mice receiving AAVrg-FLEX-dtA to the liver (Fig 4F and 4G). In addition, daily measurement of ingWAT temperature revealed a significant elevation in the ingWAT temperature of the experimental mice in the dark phase (Fig 4H). These data support that a loss of parasympathetic cholinergic neurons innervating the liver induces beiging of ingWAT, resulting in an increase in thermogenesis.

WAT lipolysis in obesity can result in an excess of fatty acids and glycerol in the liver [36]. This lipid influx can cause hepatic TG synthesis, which results in MAFLD [37]. We investigated whether eliminating cholinergic input to the liver from the DMV would decrease HSL-mediated lipolysis in ingWAT. The western blot analysis of ingWAT demonstrated a decrease in the levels of phospho-HSL in the experimental mice (Fig 4I). Moreover, we observed a decrease in plasma glycerol levels in the experimental mice compared to the control mice (Fig 4J). Taken together, these findings demonstrate that the loss of hepatic cholinergic innervation led to an increase in the expression of thermogenic genes while simultaneously reducing lipolysis in ingWAT.

Increased bile acid secretion induces ingWAT beiging via TGR5

How would removing hepatic cholinergic innervation drive these reductions in adiposity in the ingWAT, which does not receive parasympathetic cholinergic innervation in rodents [38,39]. Our results suggest that removing hepatic parasympathetic cholinergic input enhances liver-adipose crosstalk, limiting ingWAT lipolysis and thereby limiting hepatic steatosis during HFD feeding. Previous studies have indicated that bile acid and its corresponding receptor TGR5 are crucial in the browning of WAT [12,40]. We thus investigated whether the loss of function of liver-innervating cholinergic neurons affects the expression of genes related to bile acid synthesis, such as Cyp7a1 and Cyp7b1. We found that hepatic cholinergic loss of function up-regulated expression of the Cyp7b1 gene involved in an alternative BA synthesis pathway [41] (Fig 5A). Furthermore, our spatial transcriptomic analysis of the liver revealed that ChAT^Cre mice without DMV cholinergic neurons innervating the liver displayed a trend toward an increase in Cyp7b1 gene expression in all 3 zones (Fig 5B), whereas Cyp7a1 expression was not altered, as observed in qPCR experiments (Fig 5A and 5B). Elevated Cpy7b1 expression led to a substantial increase in fasting serum BA levels in ChAT^Cre mice that received AAVrg-FLEX-dtA (Fig 5C). In our further analysis of the fasting serum BA profiles in both control and experimental mice, we observed a significant increase in primary BAs such as cholate and chenodeoxycholate, which was attributed to the loss of hepatic cholinergic input (Fig 5D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Ablation of liver-innervating cholinergic neurons elevates serum BA levels.

(A) Graph showing mRNA expression of the enzymes involved in BA synthesis in liver. While there was no difference in Cyp7a1 gene expression, the loss of hepatic cholinergic input significantly up-regulated Cyp7b1 mRNA expression (control (open circle), n = 7–8 mice; experimental (closed black circle), n = 7 mice). (B) Graphs showing Cyp7a1 and Cyp7b1 gene expression in the periportal (PP), midlobular (Mid), and pericentral (PC) zones in the livers of the controls and the experimental groups. (C) Graph showing the fasting serum BA levels in the control group (n = 6 mice) and the experimental group (n = 5 mice). Unpaired t test, *p < 0.05. (D) Graph showing fasting serum BA profile in the control (n = 4 mice) and experimental (n = 4 mice) mice. The Mann–Whitney test revealed a statistically significant difference between the 2 groups (*p < 0.05). The data supporting the graphs shown in the figure (Fig 5A–5D) are available in the S1 Data file. AAV, adeno-associated virus; BA, bile acid.

https://doi.org/10.1371/journal.pbio.3002865.g005

We then examined whether the expression of its corresponding receptor, TGR5, is altered in the ingWAT of the experimental mice. Our observations revealed elevated levels of both TGR5 gene and protein expression in the ingWAT of the experimental mice compared to the control group (Fig 6A and 6B). This suggests that the increase in serum BAs and Tgr5 expression in the ingWAT leads to beiging of the ingWAT in the experimental mice. To investigate this possibility, we administered Tgr5 siRNA or control siRNA into the ingWAT of ChAT^Cre mice without liver-innervating cholinergic neurons to exclusively knock down Tgr5 in the ingWAT (Fig 6C). Ten weeks after a viral infection, we conducted RT-qPCR and western blot analyses on samples and found that the Tgr5 gene and protein levels in ingWAT were significantly reduced following the injection of Tgr5 siRNA (Fig 6D), demonstrating the effectiveness of our approach in knocking down Tgr5 expression. Our experiments also revealed that the number of large-sized adipocytes and crown-like structures in the ingWAT of mice that received Tgr5 siRNA was increased (Fig 6E), accompanied by a notable increase in adipocyte area (Fig 6F and 6G).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Liver-ingWAT communication contributes to beiging of ingWAT via BAs.

(A, B) Graphs showing mRNA and protein expression levels of Tgr5 in ingWAT. There was a significant increase in TGR5 expression in the ingWAT of ChAT^Cre mice that received AAVrg-FLEX-dtA (RT-qPCR, control (open circle), n = 7 mice; experimental (closed circle), n = 7 mice, unpaired t test, **p < 0.01; western blotting, control (open circle), n = 3 mice; experimental (closed circle), n = 3 mice, unpaired t test, ***p < 0.001). (C) Schematic illustration showing our experimental configurations. (D) Reduction in Tgr5 gene and protein levels in ChAT^Cre mice following administration of AAVrg-FLEX-dtA to the liver (black circle) and another injection of Tgr5 siRNA into both sides of the ingWAT (blue circle) (unpaired t test, *** P < 0.001). (E) Representative images of ingWAT sections stained with HE. ChAT^Cre mice without DMV cholinergic neurons innervating the liver receiving an injection of Tgr5 siRNA into ingWAT exhibited larger adipocytes compared to the control mice (upper panel). Scale bar, 50 μm. Bottom panel: higher magnification view of the black dotted square. Arrow represents crown-like structure. Scale bar, 100 μm. (F, G) Plot showing the distribution of average adipocyte size. The mean cell area was significantly larger in the experimental group receiving Tgr5 siRNA than in the experimental group receiving control siRNA (I) (control, n = 5 mice; experimental, n = 5 mice, unpaired t test, *p < 0.05). The data supporting the graphs shown in the figure (Fig 6A, 6B, 6D, 6F, and 6G) are available in the S1 Data file. AAV, adeno-associated virus; BA, bile acid; HE, hematoxylin and eosin; RT-qPCR, reverse transcription-quantitative polymerase chain reaction.

https://doi.org/10.1371/journal.pbio.3002865.g006

We examined whether Tgr5 knockdown impacts HSL-mediated lipolysis in ingWAT. Our findings showed an increase in phospho-HSL levels in the ingWAT (Fig 7A), which subsequently led to higher levels of plasma glycerol (Fig 7B). Considering the well-established connection between elevated plasma glycerol levels and lipid accumulation in the liver, we sought to determine if Tgr5 knockdown results in hepatic steatosis. Our results revealed the existence of large lipid droplets and ballooned hepatocytes in the liver sections of mice that received Tgr5 siRNA (Fig 7C). Furthermore, Oil Red O staining confirmed the accumulation of lipids in the livers of mice treated with Tgr5 siRNA in the ingWAT (Fig 7D). Additionally, we found that the knockdown of the Tgr5 gene in the ingWAT of ChAT^Cre mice without DMV cholinergic neurons to the liver resulted in a significant increase in hepatic TG content, which is indicative of enhanced lipid accumulation in the liver (Fig 7E). Collectively, our findings support the idea that cross-talk between the liver and ingWAT through elevated bile acid signaling plays a vital role in maintaining the lipid metabolism balance in both the liver and ingWAT.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Tgr5 knockdown in ingWAT causes hepatic steatosis in ChAT^Cre mice receiving AAVrg-FLEX-dtA during HFD feeding.

(A) Gel images showing the western blot analysis of ingWAT from experimental mice revealed an increase in the expression of pS660-HSL when compared to the control group. (B) Graph showing plasma glycerol levels in the controls (black circle, n = 6 mice) and the experimental groups (blue circle, n = 5 mice). Unpaired t test, **p < 0.01. (C, D) Representative images of liver sections stained with HE (C) and Oil Red O (D). Upper panel: scale bar, 100 μm; bottom panel: higher magnification view of the black dotted square scale bar, 50 μm. (E) Graph showing hepatic TG levels in the controls (n = 5 mice) and the experimental groups (n = 6 mice). Unpaired t test, *p < 0.05. The data supporting the graphs shown in the figure (Fig 7B and 7E) are available in the S1 Data file. AAV, adeno-associated virus; HE, hematoxylin and eosin; HFD, high-fat diet; TG, triglyceride.

https://doi.org/10.1371/journal.pbio.3002865.g007

Deleting the brain–liver axis has a slight but significant effect on body weight

We also investigated whether the disruption of the brain–liver axis could impact body weight in mice fed HFD. After 10 weeks of HFD feeding, we observed a small but significant difference in body weight gain between the control and experimental groups (S8A Fig). The experimental mice gained less weight than the control mice during the HFD period. This decrease in weight gain was closely related to a significant reduction in fat mass in the experimental group, while lean mass was similar between the 2 groups (S8B Fig). When we compared adipose tissue mass collected from controls with that of adipose tissue from the experimental group, we found a significant decrease in the mass of ingWAT (S8C Fig). It is important to note that both groups consumed an equal amount of food during the light and dark phases (S8D Fig).

The reduction in body weight, despite no change in energy intake, may be due to an increase in energy expenditure. To assess metabolic rate, we placed animals in metabolic cages and observed that the experimental mice consumed more oxygen and produced more carbon dioxide during the dark phase than the controls (S8E and S8F Fig). These findings indicate an increase in total energy expenditure in the experimental mice (S8G Fig). However, the respiratory exchange ratio was similar between the 2 groups (S8H Fig). Notably, the experimental mice exhibited more locomotor activity during the dark phase (S8I and S8J Fig). Collectively, these results suggest that deleting liver-projecting cholinergic neurons plays a role in increasing energy expenditure.

Loss of the brain–liver axis improves insulin tolerance in male mice fed HFD

We evaluated whether deleting the brain–liver axis affects glucose homeostasis in mice fed an HFD, as the liver plays a crucial role in regulating systemic glucose levels [42]. To assess glucose intolerance and insulin sensitivity, we measured blood glucose and insulin levels. Our results showed that basal glucose levels were lower in ChAT^Cre mice receiving AAVrg-FLEX-dtA injection than in the controls (S9A Fig). Additionally, the loss of DMV cholinergic neurons innervating the liver led to a significant reduction in fasting glucose levels (S9B Fig).

To evaluate whether ChAT^Cre mice without parasympathetic cholinergic neurons innervating the liver exhibit any changes in their ability to dispose a glucose load, we conducted intraperitoneal glucose tolerance tests (i.p.GTT) and found no significant difference in blood glucose levels in response to a bolus of glucose (i.p., 2 g/kg) between the groups (S9C and S9D Fig). However, the experimental mice showed increased insulin sensitivity (S9E Fig). During the insulin tolerance tests (ITTs) (i.p., 1 U/kg), blood glucose levels were significantly lower in ChAT^Cre mice injected with AAVrg-FLEX-dtA than in controls (S9E Fig). Furthermore, the improved insulin tolerance in the experimental group was associated with reduced plasma insulin levels relative to those in the controls (S9F Fig). These results indicate that the loss of DMV cholinergic input to the liver enhanced insulin sensitivity in ChAT^Cre male mice fed HFD.

Intriguingly, a growing number of studies have revealed sex-specific differences in metabolic homeostasis, diabetes, and obesity in both rodents and humans [43–49]. For instance, premenopausal women were relatively protected from diseases associated with metabolic syndrome [50]. Female rodents were less prone to developing diet-induced obesity (DIO), insulin resistance, hyperglycemia, and hypertriglyceridemia [43,44,47,48]. In particular, C57BL/6 female mice were less likely to develop DIO and hyperglycemia than C57BL/6 male mice during high-fat feeding [47–49,51,52]. Our study also found that control female mice did not develop DIO or hyperglycemia (S10A–S10F Fig), and hepatic cholinergic loss of function had no effect on body weight or glucose homeostasis (S10A–S10F Fig). In other words, both groups did not show any notable differences in body weight, food intake, body composition, blood glucose levels, GTT, and ITT compared with the controls (S10A–S10H Fig).

Ethics statement

All mouse care and experimental procedures were approved by the Institutional Animal Care Research Advisory Committee of the Albert Einstein College of Medicine (Approval number: 00001377) and performed in accordance with the guidelines described in the NIH Guide for the Care and Use of Laboratory Animals. Viral injections were performed under isoflurane anesthesia.

Animals

Eight- to nine-week-old ChAT^Cre mice were purchased from Jackson Laboratory (JAX stock # 006410) [75]. We backcrossed them to C57BL/6J mice for several generations, and the mutant mice were bred together to make this line homozygous. Mice were housed in cages under conditions of controlled temperature (22°C) with a 12:12 h light-dark cycle and fed an HFD (Research Diets, D12492; 20% calories by carbohydrate, 20% by protein, and 60% by fat) for 10 to 12 weeks, with water provided ad libitum. Mice were euthanized with an overdose of isoflurane at the end of the experiment.

Western blotting

Proteins were isolated from the liver using CelLytic MT (Sigma, C3228) in the presence of a protease inhibitor/phosphatase inhibitor cocktail (Thermo Fisher Scientific, 78443). Proteins from the ingWAT were extracted using a protein extraction kit for adipocyte tissue (Invent, AT-022).

Protein concentrations were determined using the BCA Protein Assay Kit (Thermo Fisher Scientific, 23225). Proteins (30 to 50 μg) were prepared by adding laemmli sample buffer (Bio-Rad, 1610747) and heated at 95°C for 5 min. The samples were then separated using 10% SDS-PAGE and transferred to a polyvinylidene difluoride (PVDF) membrane. The PVDF membrane was incubated with 5% w/v non-fat dry milk or 5% BSA for 1 h at room temperature and immunoblotted with anti-HSL (phospho-serine 660) (1:1,000, Cell Signaling Technology, 45804S), anti-HSL (1:1,000, Cell Signaling Technology, 4107), rabbit anti-UCP1 (1:1,000, Abcam, ab234430), rabbit anti-CHRM1 (1:200, Alomone Labs, AMR-001), rabbit anti-CHRM2 (Alomone Labs, AMR-002), rabbit anti-CHRM3 (Alomone labs, AMR-006), rabbit anti-CHRM4 (Alomone labs, AMR-004), rabbit anti-CHRM5 (Alomone Labs, AMR-005), and rabbit anti-β-actin (Cell Signaling, 4970S) antibodies. The specificity of anti-CHRM1-5 antibodies was tested by preincubation with the blocking peptide (CHRM1 with BLP-MR001, CHRM2 with BLP-MR002, CHRM3 with BLP-MR006, CHRM4 with BLP-MR004, and CHRM5 with BLP-MR005). Following incubation with primary antibodies, the membrane was washed 3 times in TBS-T and then incubated with an anti-rabbit IgG, HRP-linked antibody (1:10,000, Cell Signaling Technology, 7074) for 2 h at room temperature. ECL reagents were applied to the membrane and protein bands were detected using an Odyssey Fc imaging system (Li-COR) and ChemiSOLO (Azure biosystems).

Immunostaining to examine the hepatic cholinergic system

To examine DMV cholinergic neurons innervating the liver, ChAT^Cre mice received an AAV.Php.eB-flex-tdTomato (titer; 1.9 × 10¹³ pfu/ml, Addgene 28306-PHP.eB) injection into the liver. We also injected AAV-phSyn1(s)-FLEX-tdTomato-T2A-SypGFP-WPRE (Addgene 51509-AAV1) into the DMV (500 nl of 4 × 10¹² pfu/ml per site; coordinate; AP: −7.5 mm, ML: ±0.1 mm, DV: −3.5 mm) to label cholinergic nerve terminals in the liver. Mice were killed 5 and 10 weeks post-viral inoculation.

Mice were anesthetized with isoflurane (3%) and transcardially perfused with pre-perfusion solution (9 g NaCl, 5 g sodium nitrate, 10,000 U heparin in 1L distilled water) followed by 4% paraformaldehyde (PFA) solution. Liver sections were stained with a rabbit anti-GFP antibody (Rockland, 600-401-215, 1:500) to visualize cholinergic nerve terminals. Brainstem sections were labeled with rabbit anti-RFP (Rockland, 600-401-379, 1:500) and goat anti-ChAT antibodies (Millipore, AB144P, 1:200). The expression of muscarinic ACh receptors CHRM2 and 4 was examined using rabbit anti-CHRM2 (Alomone, AMR-002, 1:200) and rabbit anti-CHRM4 (Alomone, AMR-004, 1:200) antibodies. Cholinergic nerve terminals were stained with goat anti-vAChT (Millipore, ABN100) and goat anti-ChAT (Millipore, AB144P) antibodies. The tissue samples were incubated with AlexaFluor 568 donkey anti-goat IgG (Invitrogen, A11057, 1:200), AlexaFluor 568 donkey anti-rabbit IgG (Invitrogen, A11011, 1:200), and AlexaFluor 488 donkey anti-rabbit IgG (Invitrogen, A21206, 1:200). The tissues were washed, dried, and mounted with VECTASHIELD medium containing DAPI. Images were acquired using a Leica SP8 confocal microscope.

To visualize cholinergic nerve fibers in the liver, ChAT^Cre mice were given an injection of AAVrg-CMV-FLEX-PLAP (with a titer of 1.0 × 10¹² GC/ml from ABM) into the liver. After 2 months following viral infections, the mice were perfused with PBS and fixed with a 4% PFA solution. The liver tissues were then sliced to a thickness of 30 μm using a vibratome and incubated in alkaline phosphatase (AP) buffer (0.1 M Tris HCl (pH 9.5), 0.1 M NaCl, 50 mM MgCl2, 0.1% Tween20, and 5 mM levamisole) for 1 h and 30 min at room temperature. The tissue was then visualized using the NBT/BCIP substrate solution (Thermo Fisher Scientific, 34042) until the desired stain developed. The tissues were washed with distilled water and mounted with aqueous mounting media (Vector, H-5501). The stained samples were captured in brightfield scanning mode using the Panoramic 250 FLASH III F2 and analyzed using slideViewer 2.7 (3DHISTECH).

Clear, unobstructed brain imaging cocktails and computational analysis (CUBIC) tissue clearing of the brainstem and liver tissues

The CUBIC method [76] was used to clear liver and brainstem tissues. Mice were perfused with 4% (w/v) PFA in PBS. Liver and brain tissues were collected and post-fixed in 4% PFA at 4°C for 24 h. Tissue samples were sectioned using a vibratome at a thickness of 300 μm. The sections were immersed in CUBIC reagent 1(25wt% Urea, 25wt% Quadrol, 15wt% Triton X-100) with gentle shaking at 37°C for 3 days. After decolorization (transparent tissues), the sections were washed with PBS 3 times at RT for 2 h and then incubated in a blocking solution composed of 1% donkey serum, 0.5% BSA, 0.3% Triton X-100, and 0.05% sodium azide for 3 h at RT with rotation. The tissue samples were incubated with goat anti-ChAT (Millipore, AB144P, 1:200), rabbit anti-RFP (Rockland, 200-301-379 1:200), and rabbit anti-GFP (Rockland, 600-401-215, 1:200) antibodies for 5 days at RT with mild shaking, and washed 3 times with PBS (each for at least 2 h). Following wash-out, the tissue samples were incubated with AlexaFluor 568 donkey anti-goat IgG (1:200, Invitrogen, A11057), AlexaFluor 568 donkey anti-rabbit IgG (1:200, Invitrogen, A11011), and AlexaFluor 488 donkey anti-rabbit IgG (1:200, Invitrogen, A21206) for 3 days at room temperature. Tissues were washed and immersed in CUBIC-reagent 2 (50wt% Sucrose, 25wt% Urea, 10wt% Triethanolamine, and 0.1%(v/v) Triton X-100) for 1 d and mounted with a mixture of mineral oil and silicone oil. A z-stack consisting of 120 images was acquired with 0.33 μm steps with a Leica SP8 confocal microscope, resulting in a total optical thickness of approximately 40 μm and 3D construction was performed with AIVIA (version 13.1).

Viral injection to delete DMV cholinergic neurons innervating the liver

pAAV-mCherry-FLEX-dtA (Addgene, 58536) was subcloned and packaged in an AAV serotype retrograde (Applied Biological Materials). A total volume of 20 μl of AAVrg-mCherry-FLEX-dtA (titer, >1 × 10¹² pfu/ml) was injected into the medial and left lobes of the livers of the ChAT^Cre mice. AAVrg-FLEX-tdTomato viruses (titer, >1 × 10¹³ pfu/ml) were used as the controls.

To quantify cell death over time, the mice were killed 10 to 12 weeks after viral inoculation. Mice were anesthetized with isoflurane (3%) and transcardially perfused with PBS and heparin, followed by 10% neutral buffered formalin solution (Sigma, HT501128). Brainstem and liver tissues were embedded in paraffin to create paraffin blocks. Paraffin blocks were cut into 10 μm sections for the brainstem and 5 μm sections for liver tissues at our histology and comparative pathology core facility.

Coronal sections of the brainstem from approximately −7.2 to −7.6 mm posterior to bregma were stained with an anti-ChAT antibody (Millipore, AB144P). Cell counting of DMV cholinergic neurons was performed using the ImageJ software (version FIJI). The adipocyte cell area was measured using the Adiposoft plugin (version 1.16) in the Image J software.

We administered Tgr5 siRNA or control siRNA (Titer, 1 × 10⁹ IU/ml, volume 20 μm) into the ingWAT of ChAT^Cre mice without liver-innervating cholinergic neurons to exclusively knock down Tgr5 in the ingWAT.

Measurement of body weight, body composition, and blood glucose

Body weight was measured weekly at 9 AM. Body composition for fat mass and fat-free mass was assessed by ECHO MRI at our animal physiology core. Blood samples were collected from the mouse tail, and a small drop of blood was placed on the test strip of a glucose meter. Non-fasting basal glucose levels were measured at 9:00 AM. Fasted blood glucose levels were measured after overnight fasting at 10 weeks post-viral injection.

Assessment of energy expenditure and locomotor activity

To examine if the loss of function of hepatic cholinergic input regulates energy expenditure and locomotor activity, we performed indirect calorimetry on mice fed HFD for 10 weeks. Mice were individually housed in the calorimeter cages and acclimated to the respiratory chambers for at least 2 days prior to gas exchange measurements. Indirect calorimetry was performed for 3 days at the end of 10 weeks using an open-circuit calorimetry system. O₂ consumption and CO₂ production were measured for each mouse at 10-min intervals over a 24-h period. The respiratory exchange ratio was calculated as the ratio of CO₂ production over O₂ consumption. Since there was no difference in food intake, VO₂ was normalized for body lean mass. Locomotor activity in X-Y and Z planes was measured by infrared beam breaks in the calorimetry cages.

Assessment of glucose tolerance and insulin tolerance

For GTT, the experimental and control mice were fasted for 15 h (6:00 PM to 9:00 AM) at 10 weeks post-viral inoculation. Sterile glucose solution was intraperitoneally administered at a concentration of 2 g/kg (glucose/body weight) at time 0. Blood glucose levels were measured at 15, 30, 60, 90, and 120 min after glucose injection. Blood glucose levels versus time after glucose injection were plotted, and the area under the curve was calculated and compared between the experimental and control groups.

For ITT, the mice were fasted for 4 h (9:00 AM to 1:00 PM). Blood glucose levels were measured at 0, 15, 30, 60, 90, and 120 min following i.p. injection of insulin (1 U/kg). We immediately injected glucose (2 g/kg) if the mice appeared ill owing to insulin-induced hypoglycemia.

RT-qPCR analysis

Liver tissues were collected from the experimental and control groups 10 to 12 weeks post-viral inoculation. Liver tissues were homogenized in Trizol reagent (Thermo Fisher Scientific, 15596–018), and total RNAs were isolated according to the manufacturer’s instructions. First-strand cDNAs were synthesized using the SuperScript III First-Strand synthesis kit (Thermo Fisher Scientific, 18080–051). qPCR was performed in sealed 96-well plates with SYBR Green I master Mix (Applied Biosystems, A25742) using the Quant Studio 3 system (Applied Biosystems). qPCR reactions were prepared in a final volume of 20 μl containing 2 μl cDNAs, and 10 μl of SYBR Green master mix in the presence of primers at 0.5 μm. Glyceraldehyde 3-phosphate dehydrogenase (Gapdh) for liver and 18S ribosomal RNA (18s rRNA) for white adipocyte tissue were used as an internal control for quantification of each sample. Amplification was performed under the following conditions: denaturation at 95  °C for 30 s, followed by 40 cycles of denaturation at 95 °C for 30 s, and annealing/extension at 60  °C for 1 min.

We examined the following genes; Srebp1c (NM_001358315.1), Chrebp (NM_021455.5), Acly (NM_134037.3), Acaca (NM_133360.3), Fasn (NM_007988.3), Scd1 (NM_009127.4), Ppara (NM_011144.6), Prdm16 (NM_027504), Cpt1a (NM_013495.2), Cpt2 (NM_009949.2), Acadm (NM_007382.5), Acadl (NM_007381.4), G6pase (NM_008061.4), Pck1 (NM_011044.3), Slc2a2 (NM_031197.2), Pgc1a (NM_008904.3), Ucp1 (NM_009463.3), Dio2 (NM_010050.4), Cidea (NM_007702.2), Gpbar1 (NM_174985.1), Cyp7a1 (NM_007824), Cyp7b1 (NM_007825). All the primer sequences (Table 1) will be validated with “Primer BLAST” primer design program to ensure specificity for the target gene. Melt curves and gel electrophoresis were analyzed to confirm the specificity of the PCR products. Relative gene expression was determined using a comparative method (2^-ΔΔCT).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. List of primer sets for qPCR.

https://doi.org/10.1371/journal.pbio.3002865.t001

Spatial transcriptomics

Livers were snap-frozen in liquid nitrogen and thawed briefly on ice, followed by embedding in a cryomold (Sakura, 4565) filled with Neg-50 (Epredia, 6502). The cryomold was then frozen by placing into a container with pre-chilled 2-methylbutane on dry-ice. The samples were cryosectioned (Cryostar NX70 Cryostat, −18°C for blade, −18°C for specimen head, set at 10 mm thickness) and placed on a pre-chilled Visium Spatial Gene Expression Slide (10x Genomics, PN-2000233). The tissue sections were processed following 10x Genomics protocols (PN-1000187, PN-1000215). The adaptors and sequencing primers were obtained from Illumina. The libraries were sequenced by Novogene using the NovaSeq 6000 (PE150) instrument targeting 80,000 reads per spot. Fastq files were aligned against mouse mm10 reference and histology images were processed with Spaceranger (2.0). Pericentral, mid-zone, and periportal clusters were identified based on Graph-Based clusters and known zonation markers (Cyp2e1, Glul, Gulo for pericentral, Cyp2f2, Hsd17b13, Sds for periportal). Violin plots were generated using Loupe Browser (6.4.1).

Measurement of hepatic and plasma TG and ACh levels

Liver samples were harvested from mice with or without liver-innervating cholinergic neurons. Tissue samples were weighed and homogenized in 1 ml of diluted NP-40 solution to solubilize triglycerides. The samples were centrifuged at 16,000 × g to remove insoluble material. The TG levels were measured using a TG assay kit (Cayman Chemical, 10010303).

For the acetylcholine assay, liver tissues were homogenized in PBS. The homogenates were then centrifuged at 5,000 g for 5 min. The supernatants were collected, and ACh levels were determined using an acetylcholine ELISA kit (Novus, NBP2-66389).

Measurement of plasma insulin levels and fasting serum bile acid levels

Blood samples from the experimental and control groups were collected from the retroorbital plexus using heparinized capillary tubes. Whole blood was centrifuged at 12,000 × g for 10 min, and plasma was separated and stored at −20°C until use. Plasma insulin concentrations were determined using two-site sandwich ELISA kits (Mercodia, cat# 10-1247-0). To analyze fasting total bile acid levels in the serum, fasting blood samples were collected without anticoagulant and determined using mouse total bile acid (Crystal Biochem, Cat#80471). BA profiling was performed at our metabolomic core facility.

Statistics

All statistical results are presented as the mean ± SEM. Statistical analyses were performed using GraphPad Prism, version 10.0. Two-tailed t tests were used to determine the p-values for each pairwise comparison. The Mann–Whitney U test was performed to evaluate the differences in BA composition between the groups. Time-course comparisons between groups were analyzed using a two-way repeated-measures analysis of variance (RM ANOVA) with Sidak’s correction for multiple comparisons. Data were considered significantly different when the probability value was less than 0.05.