Skip to main content
NCBI home page
Scientific Reports logoLink to Scientific Reports
. 2024 Dec 1;14:29841. doi: 10.1038/s41598-024-80871-8

Effects of cholestasis and hyperammonemia on dendritic spine density and turnover in rat hippocampal neurons

Laurianne Giovannoni 1,2,✉,#, Katarzyna Pierzchala 3,4,#, Mathias De Roo 2,5, Olivier Braissant 6, Stephen Bruce 6, Valérie Anne McLin 1, Laszlo Vutskits 2,5
PMCID: PMC11609291  PMID: 39617839

Abstract

Adults and children with cholestatic liver disease are at risk for type C hepatic encephalopathy (HE) and may present lifelong neurocognitive impairment. While the underlying cellular and molecular mechanisms are still incompletely understood, ammonium and bile acids (BAs) seem to play a key role in this pathology, by crossing the blood-brain-barrier and modifying neuronal homeostasis and synaptic plasticity. This experimental study aimed to investigate the effects of ammonium and BAs on dendritic spines of rat hippocampal CA1 neurons. Taking advantage of the bile duct ligated (BDL) in vivo rat model and a hippocampal organotypic rat ex vivo slice model, we analyzed dendritic spine density in both models and spine turnover ex vivo. BDL rats showed decreased dendritic spine densities after 8 weeks, paralleled with increased concentrations of blood ammonium. In organotypic hippocampal slices, exposure to ammonium, tauro-α-muricholic and taurocholic acid induced a decrease in dendritic spine density during the first 3 days, followed by an increase in dendritic spinogenesis during days 4–5, resulting in an increased number of dendritic spines. These observations provide new insights into the effects of ammonium and BAs on dendritic spines and consequently synaptic plasticity in chronic cholestatic liver disease.

Subject terms: Hepatology, Neural circuits

Introduction

Children and adults with cholestatic liver disease are at risk for chronic hepatic encephalopathy and long-lasting neurocognitive sequelae1–4. The pathogenesis underlying these symptoms associated with cholestatic liver disease is widely considered as multifactorial, and the molecular and cellular mechanisms are incompletely understood. Since elevated plasma ammonium5 and bile acid6,7 levels are associated with impaired cognitive function, accumulation of these substrates in the central nervous system is hypothesized to play a key role in cognitive deficits8.

Bile duct ligation (BDL) in rats is a well-established experimental model recapitulating multiple pathophysiological features of human cholestatic liver disease. This surgical model is the most widely used model of type C HE9, the encephalopathy most commonly associated with chronic liver disease. Following BDL, adult rats develop motor and cognitive deficits10, reminiscent of what is seen in humans with type C HE. These behavioral deficits were associated with decreased dendritic spine density of pyramidal neurons both in the cerebral cortex and in the hippocampus11.

Dendritic spines are postsynaptic sites of excitatory inputs to pyramidal neurons. Changes in the density and size of these structures provide us with important morpho-functional read-out of synaptic plasticity12. Given the high complexity of concomitantly occurring pathophysiological changes following BDL, the specific effects of ammonium and bile acid on dendritic spines is difficult to explore in the in vivo BDL model. Organotypic slice cultures of the hippocampus offer an appealing experimental alternative since these preparations maintain three-dimensional architecture with intact synaptic circuitry13. Importantly, these slice cultures allow live, longitudinal imaging of dendritic spines over a period of several days14.

In this experimental study, we aimed to analyze the effects of ammonium and bile acids on dendritic spines. We hypothesized that ammonium and bile acids would lead to a decrease in dendritic spine density and turnover. To explore this hypothesis, we first focused on the in vivo model of BDL to evaluate if, in agreement with previous observations11, decreased dendritic spine densities in CA1 hippocampal neurons are indeed present in this experimental model. We then took advantage of our previously validated organotypic slice culture model of the hippocampus15 to test the hypothesis that ammonium and bile acids may influence dendritic spine density and turnover.

Results

Bile duct ligation leads to decreased dendritic spine density of hippocampal CA1 pyramidal neurons

To assess the impact of cholestasis on dendritic spines, we took advantage of our previously validated BDL model in adult male rats9,10. The presence of chronic liver disease (CLD) was confirmed by biochemical measurements in this experimental model. Two weeks following BDL surgery, a decrease in blood glucose, and increase in blood bilirubin, GOT and GPT were observed (Fig. 1a) as also shown previously10. An increase in plasma ammonium concentrations was observed by the end of the 2nd week post-BDL and further increase by the end of the 8th week (88.22 ± 40.78 µmol/L vs. 18.73 ± 5.38 µmol/L, p < 0.001).

Fig. 1.

Fig. 1

Open in a new tab

BDL animals CA1 pyramidal neuron have a decreased dendritic spine density at w6 time point post surgery. Wistar male adult rats underwent bile-duct ligation (BDL) and sham surgery and Golgi Cox staining experiments were conducted at week 8 post surgery. (a) Longitudinal evolution of blood biochemical parameters induced by bile duct ligation (BDL: n = 1, sham: n = 3). (b) A representative photomicrograph of hippocampal histological sections of Golgi-Cox staining and neuronal morphology analysis of pyramidal CAl neurons of dendritic spines of Sham (left panels) and BDL rat (right panels), scale bar 10 µm. Graphical illustration of significant (c) apical and (d) basal spines density decrease in BDL rats (n = 100 neurons from 7 BDL animals and n= 50 neurons from 3 Sham animals). Data are presented individually with mean± SD and statistical significance (One-way ANOVA with post-hoc Tukey HSD, + within the group (pre BDL vs. post-op BDL), * between the groups (sham vs. post-op BDL)) •/+p < 0.05, ••/++p < 0.01, •••/+++p < 0.001.

Golgi-Cox staining of BDL rat brains (week 8 post BDL) showed a significant loss of dendritic spines density in CA1 hippocampal neurons: respectively 9.91 ± 0.24 vs. 5.61 ± 0.91 apical spines/10 µm (-43%, p < 0.003) and 9.19 ± 0.91 vs. 5.99 ± 1.13 basal spines/10 µm (-35%, p < 0.001) as compared to shams (Fig. 1b–d).

Exposure to bile acids and ammonium modifies spine turnover and transiently decreases CA1 pyramidal neuron dendritic spine density ex vivo

In order to mimic cholestastic conditions, we used a mixture of ammonium chloride (NH4Cl) and TαMCA and TCA (respectively 2,5 mM; 100 µM and 200 µM).

Using propidium iodide labeling, we first confirmed that NH4Cl, TαMCA and TCA, alone or in combination did not induce cell death in hippocampal organotypic slice cultures for up to 7-days of exposure (Fig. 2).

Fig. 2.

Fig. 2

Open in a new tab

Bile acids and ammonium do not change cell death level of CA1 pyramidal neurons. Bile acids (BAs) were added 4 days after hippocampal slice preparation (Day 0). For MIX condition: MEM + TCA 200 µmol/L + TaMCA l00 µmol/L + NH4CI 2.5 mM. Controls were exposed to MEM alone. Propidium iodide (Pl) was used to detect dead cells after 3 and 7 days of incubation (n = 7 hippocampal slices from 7 different animals for each condition). Propidium iodide positive red fluorescent cells were manually counted in high power field (×400) after 3 (a) and 7 (b) days of incubation. Ctrl = control MEM medium condition. Data are presented individually with means ± SD and statistical significance (One-way ANOVA with post-hoc Tukey HSD).

Next, we evaluated the effects of NH4Cl, TαMCA and TCA, alone or in combination, on dendritic spines of EGFP-transfected CA1 pyramidal neurons. Dendritic spines were imaged every 24 h for 5 consecutive days in the presence/absence of these substrates.

Spine density showed a bi-phasic response to combined NH4Cl, TαMCA and TCA or NH4Cl alone, TαMCA or TCA alone. A significant decrease was observed during the 1st 72 h of incubation followed by an increase between 72 h and 96 h. Indeed, spine density was significantly lower compared to control conditions when using a combination of NH4Cl, TαMCA and TCA from day 1 (after 24 h of incubation, 0.43 ± 0.21 vs. 1.11 ± 0.35 spines/µm, p < 0.01) to day 3 (after 72 h of incubation, 0.53 ± 0.24 vs. 1.02 ± 0.41 spines/µm/, p < 0.001) but not at days 4 (0.86 ± 0.33 vs. 0.95 ± 0.29 spines/µm) and 5 (0.93 ± 0.19 vs. 0.77 ± 0.28 spines/µm) (Fig. 3b).

Fig. 3.

Fig. 3

Open in a new tab

Exposure to bile acids and ammonium transiently decrease CA1 pyramidal neuron dendritic spine density. Slices were GFP-transfected after 7 days and BAs were added 3 days thereafter (i) MIX condition MEM + TCA 200µmol/L + TaMCA 100 µmol/L + NH4CI 2.5 mM; (ii) MEM + TCA 200 µmol/L; (iii) MEM + TaMCA l00 µmol/L; (iv) MEM + NH4CI 2.5 mM. Controls were exposed to MEM alone. CAl pyramidal neuron spines were then counted daily during 120h for Controls (n=15 neurons in 7 different animals). (a) Representative illustration for Ctrl and MIX condition after 48h and 96h of incubation. (b) MIX condition (n = 20 neurons in 7 different animals) (c) NH4CI condition (n = 11 neurons in 7 different animals) (d) TaMCA condition (n = 4 neurons in 4 different animals) (e) TCA condition (n = 14 neurons in 7 different animals). X axis for b, c, d and e represents hours of incubation time. Data are presented individually with mean ± SD and statistical significance (Two way ANOVA mixed model, Geisser-Greenhouse’s correction): *p < 0.05, **p < 0.01, ***p < 0.001, ••p < 0.0001.

A comparable pattern and level of significance were observed for NH4Cl (day 1: 0.41 ± 0.10 vs. 1.11 ± 0.35 spines/µm, p < 0.001 to day 3: 0.66 ± 0.17 vs. 1.02 ± 0.41 spines/µm, p < 0.05, Fig. 3c) and TαMCA (day 1: 0.44 ± 0.20 vs. 1.11 ± 0.35 spines/µm, p < 0.001 to day 3: 0.45 ± 0.14 vs. 1.02 ± 0.41 spines/µm, p < 0.05, Fig. 3d).

When TCA alone was added to the culture medium (Fig. 3e), dendritic spine density was not significantly different from control conditions at day 1 (after 24 h of incubation) (0.79 ± 0.34 vs. 1.11 ± 0.35 spines/µm, ns). At days 2 and 3, dendritic spine densities were significantly lower compared to controls (day 2: 0.50 ± 0.16 vs. 0.97 ± 0.30 spines/µm, p < 0.01; and day 3: 0.59 ± 0.27 vs. 1.02 ± 0.41 spines/µm, p = 0.016) and increased back to reach control values at day 4 (after 96 h of incubation, 0.97 ± 0.18 vs. 1.00 ± 0.29 spines/µm-).

Dendritic spines are dynamic structures characterized by ongoing formation and elimination. We therefore focused on dendritic spine dynamics to better understand the bi-phasic response in spine densities observed ex vivo when NH4Cl and the bile acids TαMCA and TCA were combined. Spine stability, defined as the proportion of spines observed at the first timepoint still present on consecutive days, was significantly decreased at 48 h of combined exposure to NH4Cl, TαMCA and TCA, compared to controls (Figs. 4a and 80.50 ± 1.63% vs. 93.33 ± 4.95%, p = 0.013). Spine stability was however not significantly different between control conditions at 48 h of exposure to TCA alone, TαMCA alone, or NH4Cl alone (Fig. 4b). Spine stability under combined conditions was comparable to control conditions at 72 h and 96 h of exposure.

Fig. 4.

Fig. 4

Open in a new tab

Exposure to bile acids and ammonium modify CA1 pyramidal neuron dendritic spine turnover. Slices were GFP-transfected 7 days after slide preparation and bile acids (BAs) added 3 days later: MIX = MEM + TCA 200 µmol/L + TaMCA l00 µmol/L + NH4CI 2.5mM; (ii) MEM + TCA 200µmol/L; (iii) MEM + TaMCA 100 µmol/L; (iv) MEM + NH4CI 2.5 mM. Controls were exposed to MEM alone. CAl pyramidal neuron spines were then counted daily during 5 days. (a,b) Spine stability was defined as the proportion of spines observed at the first timepoint still present on consecutive days , x axis represent incubation time points, for example day 2 after 48 hours of incubation. We counted (c) missing (lost) spines, that could no longer be identified on the next observation (d) all new spines appearing between two observations, for example between day 1 and day 2, between 24 to 48 hours of incubation. n = 3 CAl pyramidal neurons from 3 different animals were analyzed for each condition. Data are presented as mean ± SD and statistical significance (Two way ANOVA mixed model, Geisser-Greenhouse’s correction): *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001..

Hippocampal slices were also followed longitudinally for daily spine elimination and spinogenesis (respectively Fig. 4c and d). At 48 h and 72 h, we observed significantly more spine elimination in TCA exposure compared to controls (respectively 0.12 ± 0.02 vs. 0.02 ± 0.03 spines/µm, p = 0.013 and 0.08 ± 0.01 vs. 0.02 ± 0.03 spines/µm, p = 0.008) and significantly more spinogenesis, compared to controls after 48 h of TαMCA exposure (0.12 ± 0.01 vs. 0.04 ± 0.03 spines/µm, p = 0.009). Next, at 72 h, we observed significantly more spinogenesis, compared to control for NH4Cl alone and TCA alone exposure (respectively 0.09 ± 0.01 vs. 0.03 ± 0.01 spines/µm, p < 0.001 and 0.08 ± 0.03 vs. 0.03 ± 0.01 spines/µm, p = 0.02).

Finally, at 96 h, we observed significantly more spine elimination in combined NH4Cl + BAs or NH4Cl alone exposure as compared to controls (respectively 0.15 ± 0.08 or 0.07 ± 0.02 vs. 0.02 ± 0.01 spines/µm, p = 0.04 or 0.03). We perceived also significantly more spinogenesis, compared to controls of NH4Cl + BAs (0.28 ± 0.23 vs. 0.02 ± 0.01 spines/µm, p = 0.01).

Discussion

In an established model of type C HE in rats, we have shown decreased dendritic spine density of CA1 pyramidal neurons at 8 weeks from disease onset, and that these morpho-functional changes are concurrent with a rise in blood ammonium and CSF concentrations of bile acids. Ex vivo longitudinal imaging of dendritic spines of CA1 neurons in organotypic slice cultures of the hippocampus suggest that the underlying mechanism may be more complex and could involve a dynamic process of dendritic spine loss and de novo spinogenesis.

At the molecular level, ammonium is an accepted culprit in the CNS manifestations of liver disease both in humans and in the BDL model. It exerts its effect via different mechanisms, among which the Glutamine load is one of the first to manifest16. However, multiple other molecular events have been described in hepatic encephalopathy, including fluctuations in organic osmolytes, antioxidants and neurotransmitters10,17,18 and more recently bile acids7,8. Further, evidence suggests that blood-brain barrier permeability is increased in BDL rats19–21.

Whether molecular changes described in type C HE are primary or secondary to mild edema22 and increased blood-brain barrier permeability is difficult to determine. Nevertheless, they are known to be associated with in vivo changes in astrocyte morphology and neuronal growth in vitro 3D aggregates cell culture prepared from mechanically dissociated brain of 16 d rat embryos10,23,24.

To better understand the neuronal changes in cholestasis, we combined the advantages of in vivo and ex vivo approaches. First, we analyzed the rat BDL model which is associated with both molecular and structural changes in the central nervous system17, and is accepted to recapitulate some of the molecular and cellular events in humans with chronic liver disease4,25. Next, we used an ex vivo model of CA1 pyramidal neurons to analyze the direct effects of NH4Cl and bile acids on dendritic spines.

In line with previous observations, in vivo dendritic spine density of CA1 hippocampal pyramidal neurons was significantly decreased 8 weeks after BDL11. Chen et al. showed that the decrease in spine densities was associated with increased CSF concentrations of ammonium and bile acids26, something confirmed in the present study.

Using longitudinal ex vivo studies, we also showed that BAs and NH4Cl modify the spine turnover of CA1 pyramidal neurons. When BAs and NH4Cl were combined, an acute destabilization of the pre-existing synaptic network was observed, with significant decrease in spine stability and dendritic spine loss over the first 3 days of exposure. This was followed by an increase in dendritic spinogenesis, resulting in spine densities reaching control values at experimental day 5, suggestive of a compensatory neuronal reaction.

The seeming discrepancy between results obtained from our in vivo and ex vivo models most probably stems from the distinct specificities of these models. While the in vivo approach allows us to obtain a post-hoc static analysis of overall dendritic spine densities over a longer period of time (8 weeks following BDL), the ex vivo experiments allow us to gain insights into the dynamics underlying these changes over a shorter time frame (up to one week). Based on the in vivo data showing a decrease in spine densities after 8 weeks, a plausible interpretation of our ex vivo data is that the compensatory increase in dendritic spine formation may only be temporary and cannot compensate for the loss of dendritic spines.

While the functional relevance of our observations remains to be established, these results suggest that increased ammonium and bile acid concentrations may affect dendritic spine dynamics and, therefore, synaptic plasticity in animals and humans with chronic cholestatic liver disease. Indeed, dendritic spines undergo a constant turnover in physiological conditions27,28. This turnover is increased upon learning-related patterns of activity in an input specific manner29 and is positively correlated with learning performance30. Thus, the transient pathological alteration of spine turnover observed in this study, as any pathological alteration of spine turnover, may lead to an impairment, at least transient, of learning and memory processes with possible consequences on cognitive function31 and behavior32.

The relative role and contribution of ammonium and bile acids to spine dynamics and the underlying mechanisms remains to be determined.

One possible, albeit unexplored, hypothesis is illustrated in Fig. 5.

Fig. 5.

Fig. 5

Open in a new tab

Hypothetical cellular mechanisms involved in CA1 neuronal dendritic spine bi-phasic answer to cholestasis. It is known that BAs can affect blood-brain-barrier permeability and enter neurons using specific transporters, like ASBT or MRP6. BAs have been shown to link directly to purified muscle actin, slowing the rate of polymerization, thus modulating actin dynamics in HeLa cells or cholangiocytes. Therefore, we speculate that TCA and TαMCA link to actin cytoskeleton in our in vitro CA1 pyramidal neurons organotypic slice model (a) and modifying actin polymerization, in turn leading to spine loss (b). This mechanism could arguable be further enhanced by astrocyte swelling observed in hyperammoniemic animals as astrocytes are known to be a major determinant of actin shape and stability in CA1 neurons. Finally, FXR-mediated Bas efflux from cell could contribute to restoring actin stability ans thus, spine integrity (c). BSEP: Bile Salt Export Pump; FXR: Farnesoid X Receptor; MRP6: Multi-Resistance Protein 6; ASBT: Apical Bila Acid Salt Transporter; BBB: Blood Brain Barrier.

Our study presents several limitations. First, we used only male animals, limiting the generalizability of our findings, since in humans, females may be more prone to cholestastic liver disease than males33. Also, due to commercial limitations in bile acids availability and stability, we also limited our studies to TCA, TαMCA, while in fact other species such as cholic or β- and Ω-muricholic acids34 have also been identified in the CSF of BDL rats. Finally, we focused on changes in dendritic spines numbers of CA1 pyramidal neuron, but we have yet to correlate the anatomical changes with functional measures. Indeed, the significance of the delayed increase in dendritic spine density in culture remains to be determined, and further investigations are needed to establish if these observations reflect a biphasic response to ammonium / bile acids on synaptic plasticity or merely represent a compensatory mechanism intrinsically related to culture conditions.

In conclusion, our results confirm previous experimental observations suggesting that chronic cholestatic liver disease may induce synaptic loss in the central nervous system. We also provide some new insights suggesting that elevated levels of NH4Cl and bile acids may be mediating these effects. Altogether, these exploratory data suggest that increased central nervous system concentrations of NH4Cl and bile acids as seen in chronic cholestatic liver disease, may affect synaptic plasticity.

Methods

Animal welfare

All experimental protocols were approved by the Committee on Animal Experimentation for the Canton de Vaud for BDL and Golgi Cox staining experiments (VD3022/VD2439) and the Geneva Cantonal Veterinary Office and the University of Geneva for organotypic hippocampal slice experiments (authorization GE_65_18). All methods were performed in accordance with the relevant guidelines and regulations and all methods are reported in accordance with ARRIVE guidelines.

Due to the complexity of the study different animals have been used for each experiment. The detailed number of used animals is provided under each procedure.

Chronic liver disease model

Wistar male adult rats (175–200 g at surgery, Charles River Laboratories, L’Arbresle, France) underwent bile duct ligation (BDL) and Sham surgery (BDL n = 9, Sham n = 5) as previously described9,10,35. Rats were group-housed in the animal facilities of EPFL / CIBM-AIT in Lausanne.

Liver parameters (plasma bilirubin, aspartate aminotransferase (AST/GOT) and Alanine Aminotransferase (ALT/GPT) (Reflotron® Plus system, F. Hoffmann-La Roche Ltd, Switzerland), glucose (CONTOUR®XT) and blood NH4+ (PocketChemTM BA PA-4140) were monitored longitudinally (week 0, 2, 4, 6 and 8, BDL n = 7 and Sham n = 3).

Histology assessments

Golgi Cox staining based on the principle of metallic impregnation of neurons was applied to reveal the cytoarchitecture of the hippocampus and its detailed neuronal morphology36,37. Before brain extraction, rats (BDL n = 7, Sham n = 3) were deeply anesthetized with 4% isoflurane for 5 min and then injected with intraperitoneal Temgesic (buprenorphine, ESSEX) (analgesic dose: 250 µl of 0.03 mg/ml in 0.9% NaCl), perfused with cell culture medium RPMI 1640 (pH 7.4, Sigma), supplemented with 10% of Fetal Bovine Serum (FBS, Sigma-Aldrich® Buchs, Switzerland) and 1% of antibiotic mix (50.5 units/ml penicillin, 50.5 µg/ml streptomycin and 101 µg/ml neomycin, Sigma-Aldrich® Buchs, Switzerland) to wash out blood and keep the brain cells alive. Extracted half-brains were immediately immersed in Golgi Cox staining solution and kept in the dark at room temperature for 25 days followed by PBS wash and 48 h cryopreservation in 30% sucrose in PBS at 4 °C. Brains were frozen in liquid nitrogen and preserved at -80oC until the developing procedure. Brains were sliced at 110 μm thick sagittal sections using Leica VT1200 S vibratome. Sections were mounted on Superfrost Plus microscope slides (Thermo Scientific®, Waltham, USA). After staining procedure and dehydratation, the slides were mounted with Neo-Mount (EMD Millipore®, Schaffausen, Switzerland) and covered with a coverslip. Golgi Cox-stained tissue sections of hippocampus were examined with light microscopy. Only tissues uniformly stained where dendritic segments and spines were clearly visible were used for quantitative analysis. Measurements were obtained from the CA1 neurons of hippocampus using manual counting of dendritic spines. Apical and basal dendrite branches were analyzed. Dendrite fragments were chosen with the following criteria: (i) staining without breaks, (ii) distance from soma: ~120 μm (apical) or 30 μm (basal), (iii) dendrite branch had to be in the same focus plane having a length ~ 20–70 μm, and (iv) to minimize errors connected with length measuring a relatively straight branch fragments were chosen. An average of 15 branches per brain were analyzed. Twenty-five (25) slides / hemisphere, n = 100 neurons from 7 BDL animals and n = 50 neurons from 3 Sham animals were analyzed.

Sections were digitized using a MEIJI TECHNO TC5600 Inverted Biological Microscope equipped with the INFINITYX-32 pixel shifting camera with the picture size 6464 × 4864 pixels. Images ware collected using the x5, x20 and x50 objectives and processed with INFINITY ANALYZE 7 software (Lumenera®, Nepean, Canada).

Organotypic hippocampal slice cultures

Wistar rats of local breeding were group-housed and bred in the University of Geneva Medical School animal facility for organotypic hippocampal slice culture experiments. All animals were under a 12 h light/dark cycle and 22 °C temperature. Food and water were available ad libitum.

Six to seven-day old male Wistar rats (n = 4 to 7 per group, see figure legend) were used to prepare transverse hippocampal organotypic slice cultures (400 μm thick) as described previously15.

Slices were maintained for 11 days for cytotoxicity experiments and 18 days for spine observation in a CO2 incubator at 33 °C, as previously published38. Transfection was performed with a pc-DNA3.1-EGFP plasmid using a biolistic method (Helios Gene Gun, Bio-Rad®, Cressier, Switzerland) 7 days after slice preparation. Fluorescence was detectable after 24–48 h and then remained stable for at least 15 days.

Culture medium

0,14% NaCl (Sigma-Aldrich® Buchs, Switzerland) and with 1,2% Hepes (Sigma-Aldrich® Buchs, Switzerland) were added to Minimum Essential Media (MEM medium 2x: MEM 10x diluted 1:5 (Sigma-Aldrich® Buchs, Switzerland), which was then stored at -20 °C. We used this medium as control. Bile acids or ammonium chloride (NH4Cl) were added as follows to mimic CLD conditions (MIX):

Bile acid concentration

In vivo, tauro-α-muricholic acid (TαMCA) (Steraloids® Newport, USA) and taurocholic acid (TCA) (Sigma-Aldrich® Buchs, Switzerland) are the 2 available bile acids that seem to be increased in the CSF of BDL rats 6 weeks after BDL (preliminary data, not shown). Therefore, we added 100 µM tauro-α-muricholic acid (TαMCA) (Steraloids® Newport, USA) and 200 µM taurocholic acid (TCA) (Sigma-Aldrich® Buchs, Switzerland) to the MEM to mimic cholestasis condition. We chose these concentrations in a comparable fold change between BDL and sham operated CSF (preliminary data not shown), and also comparable to BA concentrations in the brain of a mouse model of acute liver failure7, and to BA concentrations used, without cytotoxicity, in the literature in vitro on primary neurone culture39or murine hypothalamic neuron cell line40.

Ammonium concentration

Ammonium chloride (Sigma-Aldrich® Buchs, Switzerland) was added to the culture medium according to previous reports in organotypic brain cell cultures in aggregates23 using a 2.5 mM concentration.

pH

The medium was changed every 3 days and pH was maintained between 7.40 and 7.42.

Cell death

Bile acids and / or ammonium chloride were added to the culture medium starting from day 4 after slice preparation. Three (3) or seven (7) days after exposure, hippocampal slices were incubated 20 min in MEM medium with 5 µg/ml propidium iodide (PI, Sigma-Aldrich® Buchs, Switzerland). Dead CA1 cells (red) were counted manually in 600 μm2 fields using a Zeiss Axioskop 2 Plus system coupled to Axiovision Software.

Confocal imaging

Short imaging sessions (10–15 min) of transfected hippocampal slices were carried out, from 4 (day 1) to 8 (day 5) days after transfection, with an Olympus Fluoview 300 system coupled to a single photon laser as described38,41. We focused on dendritic basal segments of about 100 μm in length and located between 100 and 350 μm from the soma on secondary dendrites of CA1 pyramidal neurons using a 40x objective and a 10x additional zoom (final resolution: 25 pixels per micron; steps between scans: 0.4 μm, Fig. 3a). We used confocal microscopic analysis of dendritic spines as proxies for excitatory synapses in transfected cells.

Analysis of spine morphology and turnover

Spine density was analyzed in CA1 pyramidal neurons. Dendritic spine analysis was conducted as described previously38,41. Dendritic spines were defined as protrusions from the dendritic shaft exhibiting an enlargement at the tip of dendrites. New spines were defined as spines appearing between two observations (day 1 to 5). Lost spines were defined as spines that could no longer be identified on the next observation. Spine stability was defined as percentage of initial spines (day 1) that could be still observed at each time point (day 2 to 5).

All turnover and stability analyses were carried out manually by scrolling across single z-stacks of raw images using ImageJ software. We counted spines located behind each other on z-stacks whenever the distinction was possible. Dubious situations due to possible changes in spine shape, size or orientation were discarded, but overall accounted for only a small number of cases (less than 1% of observed spines).

Statistics

Statistical analysis was conducted using PRISM software for multiple t-test, with Holm-Sidak correction method, One-way Anova with post-hoc Tukey HSD or mixed model ANOVA. P < 0,05 was considered for significance. All statistics are given with the standard deviation (s.d.).

Acknowledgements

LG benefited from consumable support from the Ernst and Lucie Schmidheiny Foundation, and from a 2-year .5FTE grant from the University of Geneva Faculty of Medicine. The in vivo experiments were supported by the CIBM Center for Biomedical Imaging of the UNIL, UNIGE, HUG, CHUV, EPFL, the Leenaards and Jeantet. The authors thank Dr Vladimir Cousin and Dre Anaïs Schneider for their clinical support during LG lab time, Michèle Brunet (HUG, CMU) for her assistance in rat hippocampal culture and transfection and Dr Cristina Cudalbu for her support with the in vivo BDL rat model.

Author contributions

L.G performed ex vivo experiments and K.P. the in vivo experiments. L.G. and K.P. wrote the main manuscript text ans prepared Figs. 1, 2, 3 and 4. O.B. prepared supper Fig. 1. All authors reviewed the manuscript.

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

These authors contributed equally: Laurianne Giovannoni, Katarzyna Pierzchala, Valérie Anne McLin and Laszlo Vutskits.

References

  • 1.Stewart, S. M., Campbell, R. A., McCallon, D., Waller, D. A. & Andrews, W. S. Cognitive patterns in school-age children with end-stage liver disease. J. Dev. Behav. Pediatr.13, 331–338 (1992). [PubMed] [Google Scholar]
  • 2.Caudle, S. E., Katzenstein, J. M., Karpen, S. J. & McLin, V. A. Language and Motor skills are impaired in infants with biliary atresia before transplantation. J. Pediatr.156, 936–940e1 (2010). [DOI] [PubMed] [Google Scholar]
  • 3.Gold, A. et al. Assessment of school readiness in chronic cholestatic liver disease: a pilot study examining children with and without liver transplantation. Can. J. Gastroenterol. Hepatol.2017, 1–8 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Cudalbu, C. & Taylor-Robinson, S. D. Brain edema in chronic hepatic encephalopathy. J. Clin. Experimental Hepatol.9, 362–382 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Riordan, S. M. & Williams, R. Treatment of hepatic encephalopathy. N Engl. J. Med.337, 473–479 (1997). [DOI] [PubMed] [Google Scholar]
  • 6.Wang, X. et al. Serum bile acids are associated with pathological progression of hepatitis B-induced cirrhosis. J. Proteome Res.15, 1126–1134 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.McMillin, M. et al. Bile acid signaling is involved in the neurological decline in a murine model of acute liver failure. Am. J. Pathol.186, 312–323 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.DeMorrow, S. Bile acids in hepatic encephalopathy. J. Clin. Experimental Hepatol.9, 117–124 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.DeMorrow, S., Cudalbu, C., Davies, N., Jayakumar, A. R. & Rose, C. F. 2021 ISHEN guidelines on animal models of hepatic encephalopathy. Liver Int.41, 1474–1488 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Braissant, O. et al. Longitudinal neurometabolic changes in the hippocampus of a rat model of chronic hepatic encephalopathy. J. Hepatol.71, 505–515 (2019). [DOI] [PubMed] [Google Scholar]
  • 11.Chen, J. R. et al. Morphological changes of cortical pyramidal neurons in hepatic encephalopathy. BMC Neurosci.15, 15 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Arellano, J. I. Ultrastructure of dendritic spines: correlation between synaptic and spine morphologies. Front. Neurosci.1, 131–143 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Stoppini, L., Buchs, P. A. & Muller, D. A simple method for organotypic cultures of nervous tissue. J. Neurosci. Methods. 37, 173–182 (1991). [DOI] [PubMed] [Google Scholar]
  • 14.De Roo, M., Klauser, P., Mendez, P., Poglia, L. & Muller, D. Activity-dependent PSD formation and stabilization of newly formed spines in hippocampal slice cultures. Cereb. Cortex. 18, 151–161 (2008). [DOI] [PubMed] [Google Scholar]
  • 15.De Roo, M. & Ribic, A. Analyzing structural plasticity of dendritic spines in organotypic slice culture. In Synapse Development (ed Poulopoulos, A.) Vol. 1538, 277–289 (Springer New York, 2017). [DOI] [PubMed] [Google Scholar]
  • 16.Mosso, J. et al. Diffusion of brain metabolites highlights altered brain microstructure in type C hepatic encephalopathy: a 9.4 T preliminary study. Front. Neurosci.18, 1344076 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Rackayova, V. et al. 1H and 31P magnetic resonance spectroscopy in a rat model of chronic hepatic encephalopathy: in vivo longitudinal measurements of brain energy metabolism. Metab. Brain Dis.31, 1303–1314 (2016). [DOI] [PubMed] [Google Scholar]
  • 18.Rackayova, V., Braissant, O., Rougemont, A. L., Cudalbu, C. & McLin, V. A. Longitudinal osmotic and neurometabolic changes in young rats with chronic cholestatic liver disease. Sci. Rep.10, 7536 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Faropoulos, K. et al. Altered occludin expression in brain capillaries induced by obstructive jaundice in rats. Brain Res.1325, 121–127 (2010). [DOI] [PubMed] [Google Scholar]
  • 20.Quinn, M. et al. Suppression of the HPA axis during extrahepatic biliary obstruction induces cholangiocyte proliferation in the rat. Am. J. Physiology-Gastrointestinal Liver Physiol.302, G182–G193 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Sandhir, R., Kaur, S. & Dhanda, S. N-acetyl-l-cysteine prevents bile duct ligation induced renal injury by modulating oxidative stress. Ind. J. Clin. Biochem.32, 411–419 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Pierzchala, K. et al. Lessons on brain edema in HE: from cellular to animal models and clinical studies. Metab. Brain Dis.39, 403–437 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Braissant, O. et al. Ammonium-induced impairment of axonal growth is prevented through glial creatine. J. Neurosci.22, 9810–9820 (2002). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Braissant, O., McLin, V. A. & Cudalbu, C. Ammonia toxicity to the brain. J. Inher Metab. Disea. 36, 595–612 (2013). [DOI] [PubMed] [Google Scholar]
  • 25.Córdoba, J. et al. T2 hyperintensity along the cortico-spinal tract in cirrhosis relates to functional abnormalities. Hepatology38, 1026–1033 (2003). [DOI] [PubMed] [Google Scholar]
  • 26.Chen, M. et al. Enriched environment ameliorates learning and memory deficits in hepatic encephalopathy mice by restoration of the structure of dendrites and dendritic spines. Brain Res.1804, 148264 (2023). [DOI] [PubMed] [Google Scholar]
  • 27.Trachtenberg, J. T. et al. Long-term in vivo imaging of experience-dependent synaptic plasticity in adult cortex. Nature420, 788–794 (2002). [DOI] [PubMed] [Google Scholar]
  • 28.Holtmaat, A. J. G. D. et al. Transient and persistent dendritic spines in the neocortex in vivo. Neuron45, 279–291 (2005). [DOI] [PubMed] [Google Scholar]
  • 29.De Roo, M., Klauser, P. & Muller, D. LTP promotes a selective long-term stabilization and clustering of dendritic spines. PLoS Biol.6, e219 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Roberts, T. F., Tschida, K. A., Klein, M. E. & Mooney, R. Rapid spine stabilization and synaptic enhancement at the onset of behavioural learning. Nature463, 948–952 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Caroni, P., Donato, F. & Muller, D. Structural plasticity upon learning: regulation and functions. Nat. Rev. Neurosci.13, 478–490 (2012). [DOI] [PubMed] [Google Scholar]
  • 32.Forrest, M. P., Parnell, E. & Penzes, P. Dendritic structural plasticity and neuropsychiatric disease. Nat. Rev. Neurosci.19, 215–234 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Bhandari, B. M., Bayat, H. & Rothstein, K. D. Primary biliary cirrhosis. Gastroenterol. Clin. N. Am.40, 373–386 (2011). [DOI] [PubMed] [Google Scholar]
  • 34.Pan, X. et al. Metabolomic profiling of bile acids in clinical and experimental samples of Alzheimer’s disease. Metabolites7, 28 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Cudalbu, C. BDL surgery protocol. 10.5281/ZENODO.10652104 (2024).
  • 36.Zaqout, S. & Kaindl, A. M. Golgi-Cox Staining Step by Step. Front. Neuroanat.10 (2016). [DOI] [PMC free article] [PubMed]
  • 37.Louth, E. L., Sutton, C. D., Mendell, A. L., MacLusky, N. J. & Bailey, C. D. C. Imaging neurons within thick brain sections using the golgi-Cox Method. JoVE5535810.3791/55358 (2017). [DOI] [PMC free article] [PubMed]
  • 38.De Roo, M., Klauser, P., Garcia, P. M., Poglia, L. & Muller, D. Chapter 11 spine dynamics and synapse remodeling during LTP and memory processes. in Prog. Brain Res.169 199–207 (Elsevier, 2008). [DOI] [PubMed]
  • 39.Morales, C. et al. Ursodeoxycholic acid binds PERK and ameliorates neurite atrophy in a cellular model of GM2 gangliosidosis. Int. J. Mol. Sci.24, 7209 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.McMillin, M. et al. Suppression of the HPA axis during cholestasis can be attributed to hypothalamic bile acid signaling. Mol. Endocrinol.29, 1720–1730 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.De Roo, M. et al. Anesthetics rapidly promote synaptogenesis during a critical period of brain development. PLoS ONE. 4, e7043 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Articles from Scientific Reports are provided here courtesy of Nature Publishing Group

RESOURCES