Edinburgh Research Explorer Loss of integrin v8 in murine hepatocytes accelerates liver regeneration

Recent fate-mapping studies in mice have provided substantial evidence that mature adult hepatocytes are a major source of new hepatocytes following liver injury. In other systems, integrin αvβ8 has a major role in activating transforming growth factor beta (TGFβ), a potent inhibitor of hepatocyte proliferation. We hypothesized that depletion of hepatocyte integrin αvβ8 would increase hepatocyte proliferation and accelerate liver regeneration following injury. Using Itgb8 flox/flox ;Alb-Cre mice to deplete hepatocyte αvβ8, following partial hepatectomy, hepatocyte proliferation and liver-to-body weight ratio were significantly increased in Itgb8 flox/flox ;Alb-Cre mice compared to control. Antibody-mediated blockade of hepatocyte αvβ8 in vitro , with assessment of TGFβ signaling pathways by qPCR array, supported the hypothesis that integrin αvβ8 inhibition alters hepatocyte TGFβ signaling towards a pro-regenerative phenotype. A diethylnitrosamine-induced model of hepatocellular carcinoma, employed to examine the possibility that this pro-proliferative phenotype might be oncogenic, revealed no difference in either tumor number or size between Itgb8 flox/flox ;Alb-Cre and control mice. Immunohistochemistry for integrin αvβ8 in healthy and injured human liver demonstrated that human hepatocytes express integrin αvβ8. Depletion of hepatocyte integrin αvβ8 results in increased hepatocyte proliferation and accelerated liver regeneration following partial hepatectomy in mice. These data demonstrate that targeting integrin αvβ8 may represent a promising therapeutic strategy to drive liver regeneration in patients with a broad range of liver diseases. precise, and 13 The αv integrins, transmembrane heterodimeric proteins comprising an αv subunit and one of the five β subunits, bind to an arginine-glycine-aspartate (RGD) motif present on the tip of an exposed loop within the latency-associated peptide that maintains TGFβ in an inactive state. 14 All five αv integrins have been shown to interact with the RGD motif present in the latency-associated peptide. 15-19 This integrin-RGD interaction, in the presence of mechanical force supplied by the integrin-expressing cell, enables the release of the active TGFβ homodimer. 20 integrin αvβ8 the context of liver regeneration. We hypothesized that depletion of integrin αvβ8 from hepatocytes would reduce local activation of TGFβ and result in increased hepatocyte proliferation and accelerated liver regeneration following liver injury. Hematoxylin and eosin staining: sections were baked at 55 o C overnight, before de-waxing and rehydration. Slides were then placed in Harris Hematoxylin (Thermo Fisher Scientific, Paisley, UK) for five minutes. After washing, slides were placed in 1% acid alcohol for five seconds, followed by Scott’s tap water for two minutes. Slides were then transferred to Eosin Y solution (Thermo Fisher Scientific) for two minutes, followed by washing, dehydration, and mounting. For quantification of mitotic figures, a minimum of 1,000 hepatocytes were counted per sample. hepatectomy. However, inhibition of hepatocyte integrin αvβ8 in vitro , using a β8 integrin subunit blocking antibody, resulted in changes in expression of multiple TGFβ-responsive genes such as Plat and Hmox1 . Tissue plasminogen activator, encoded by Plat , can activate HGF 51 and has been shown to play a role in liver lobule reorganization following acute injury. 52 Knockout of tPA in mice worsens injury following bile duct ligation, but this phenotype is reversed by administration of HGF. 53 These data suggest that the regulatory role of integrin αvβ8 during hepatocyte proliferation is, at least in part, mediated via TGFβ signaling, and that integrin αvβ8 depletion or inhibition may drive hepatocyte proliferation through tPA-mediated activation of HGF. Inhibition of hepatocyte integrin αvβ8 did not alter the proliferative response to stimulation with EGF and HGF in vitro , suggesting that the accelerated liver regeneration observed following hepatocyte integrin αvβ8 depletion does not occur via modulation of HGF or EGF downstream signaling pathways. 54


INTRODUCTION
Although the liver has a unique ability to regenerate, in many cases of liver disease this regenerative capacity is overwhelmed. A successful pro-regenerative therapy for the liver could have widespread application, reducing the need for transplantation in both acute and chronic liver failure, and potentially allowing more patients with primary or metastatic liver cancer to be treated successfully. Recent fate-mapping studies in mice have provided strong evidence that, in most murine models of liver injury and regeneration, restoration of liver mass occurs predominantly through self-duplication of hepatocytes. 1,2 Hence, identifying targets that promote proliferation and expansion of the pre-existent hepatocyte population represents an attractive therapeutic approach to drive liver regeneration.
Transforming growth factor beta (TGFβ) has pleiotropic roles in liver disease. In addition to its role as a major pro-inflammatory cytokine, 3 TGFβ is also a potent repressor of hepatocyte proliferation. [4][5][6][7] Therefore, in principle, TGFβ inhibition appears an attractive therapeutic strategy to promote hepatocyte proliferation and liver regeneration. An ideal therapy would target TGFβ with precision, allowing hepatocytes to escape the mitoinhibitory effects of TGFβ, while not abrogating the positive effects of TGFβ on extracellular matrix production and vascular remodeling during the regenerative process. 8,9 Furthermore, pan-TGFβ blockade may result in a number of unwanted, off-target effects, such as induction of autoimmunity and hepatocarcinogenesis. [10][11][12] Therefore, a more nuanced, selective approach targeting the TGFβ pathway to promote liver regeneration is required.

M A N U S C R I P T
A C C E P T E D ACCEPTED MANUSCRIPT 5 TGFβ is predominantly stored within the extracellular matrix in a latent state, and much of the regulation of TGFβ function results from precise, temporally and spatially restricted, extracellular activation of this latent complex. 13 The αv integrins, transmembrane heterodimeric proteins comprising an αv subunit and one of the five β subunits, bind to an arginine-glycine-aspartate (RGD) motif present on the tip of an exposed loop within the latency-associated peptide that maintains TGFβ in an inactive state. 14 All five αv integrins have been shown to interact with the RGD motif present in the latency-associated peptide. [15][16][17][18][19] This integrin-RGD interaction, in the presence of mechanical force supplied by the integrin-expressing cell, enables the release of the active TGFβ homodimer. 20 Inhibition of myofibroblast αv integrins in mice reduces fibrosis in multiple organs via a reduction in TGFβ activation. 21 Furthermore, combined global knockout of integrins αvβ6 and αvβ8 phenocopies the developmental effects of loss of TGFβ-1 and -3. 22 In the liver, expression of integrin αvβ6 appears restricted to activated cholangiocytes, transitional hepatocytes, and oval cells during biliary and portal fibrosis. 23,24 Conversely, αvβ8 expression by hepatic cell types has not been well-characterized. Integrin αvβ8 has been shown to play an important role in TGFβ activation in other systems, including dendritic cells, [25][26][27] regulatory T cells, 28 neuroepithelium, 29 and in fibroinflammatory airway disease. 30 Further, integrin αvβ8 inhibits proliferation of lung epithelium via TGFβ activation. 31 Therefore, given the important role of αvβ8 in mediating TGFβ activation in other organ systems and pathologies, we investigated the role of hepatocyte integrin αvβ8 in the context of liver regeneration. We hypothesized that depletion of integrin αvβ8 from hepatocytes would reduce local activation of TGFβ and result in increased hepatocyte proliferation and accelerated liver regeneration following liver injury.

M A N U S C R I P T
A C C E P T E D ACCEPTED MANUSCRIPT 6

Mice
Albumin-Cre (Alb-Cre) mice 32 were obtained from the Jackson Laboratory (Bar Harbor, ME), crossed with Itgb8 flox/flox mice 33 obtained from L. Reichardt, and the resulting Itgb8 flox/flox ;Alb-Cre mice were maintained on a C57BL/6 background. Pdgfrb-Cre mice (also on a C57BL/6 background) were obtained from R. Adams. 34 Mice used for all experiments were 8-to 16-week-old and were housed under specific pathogen-free conditions in the Animal Barrier Facility of the University of California, San Francisco, or the University of Edinburgh, UK.
Genotyping of all mice was performed by PCR. Sample size was determined statistically prior to experimentation. Age-and sex-matched littermate controls were used for all experiments. Investigators were blinded to mouse genotype and experimental order was decided randomly. All experimental animal procedures were approved by the Institutional Animal Care and Use Committee of the University of California, San Francisco, or performed in accordance with the UK Home Office regulations.

Two-thirds partial hepatectomy
Two-thirds of the liver was surgically removed under isoflurane anesthesia as previously described. 35 All surgeries were performed in the first half of the day. To label proliferating hepatocytes, 5-bromo-2-deoxyuridine (BrdU (Roche), Sigma-Aldrich, Gillingham, UK) was injected two hours prior to liver harvest (100 mg/kg intraperitoneally). Mice and livers were weighed at harvest to calculate liver-to-body weight ratio.

Hepatocarcinogenesis model
Male mice were injected with diethylnitrosamine (DEN, Sigma-Aldrich) at 12 to 14 days (25 mg/kg intraperitoneally). Mice were sacrificed at 40 weeks, and macroscopic tumors counted and measured.

Liver biochemistry
Whole blood was collected immediately post mortem, allowed to clot, and serum obtained by centrifugation (9391g for 5 minutes twice). Samples were frozen at -20 °C pending analysis. Serum albumin, total bilirubin, alanine transaminase (ALT), and alkaline phosphatase (ALP) measurements were determined using commercial kits (Alpha  Hematoxylin and eosin staining: sections were baked at 55 o C overnight, before de-waxing and rehydration. Slides were then placed in Harris Hematoxylin (Thermo Fisher Scientific, Paisley, UK) for five minutes. After washing, slides were placed in 1% acid alcohol for five seconds, followed by Scott's tap water for two minutes. Slides were then transferred to Eosin Y solution (Thermo Fisher Scientific) for two minutes, followed by washing, dehydration, and mounting. For quantification of mitotic figures, a minimum of 1,000 hepatocytes were counted per sample.
No image processing was performed prior to quantitative analysis. Images presented in figures were contrast-enhanced by adjusting intensity minima and maxima. Images to be compared were processed identically and in a manner that preserved the visibility of dim and bright structures in the original image.

Primary mouse hepatocyte isolation
Primary mouse hepatocytes were isolated by retrograde perfusion of the liver with Liver Perfusion Medium (Thermo Fisher Scientific), followed by Liver Digest Medium (Thermo Fisher Scientific) at 37 °C. When hepatocytes were visually dispersed within the liver capsule, the liver was removed to a sterile dish and minced with scissors to release the crude cell isolate. The cells were then suspended in DMEM/F-12 (Thermo Fisher Scientific) and pelleted twice. Hepatocytes were purified from the washed pellets by resuspension in culture medium and centrifugation through 50% equilibrated Percoll (GE Healthcare Life Sciences, Little Chalfont, UK).

Standard primary hepatocyte culture
Primary hepatocytes were isolated as described above, resuspended in low-serum medium (DMEM (Thermo Fisher Scientific), 2.5% Fetal Bovine Serum (Thermo Fisher Scientific), 2% L-Glutamine (Thermo Fisher Scientific), 1% Penicillin Streptomycin (Thermo Fisher Scientific)), and plated onto collagen-coated wells (Collagen Type I, Millipore, Watford, UK) in a 6-well plate at a density of 500,000 cells per well. Either β8 integrin subunit blocking antibody 26 or non-binding control antibody were added at 20µg/mL and samples were incubated for 24 hours at 37 ºC in 5% CO 2 . Wells were then washed with PBS and cells lyzed as described below.

RT-qPCR
RNA was isolated from whole mouse liver, primary hepatocytes, or liver sinusoidal endothelial cells using an RNeasy Mini Kit (whole liver, hepatocytes) or RNeasy Plus Micro Kit (liver sinusoidal endothelial cells) (Qiagen, Manchester, UK). cDNA transcription and qPCR were performed using a SYBR-GreenER Two-Step qRT-PCR kit (Invitrogen, Thermo Fisher Scientific) or QuantiTect Reverse Transcription and SYBR Green PCR Kits (Qiagen).
Quantitect Primer Assays (Qiagen, 249990) were used for the following genes: Ccna2 To assess TGFβ signaling, a custom RT 2 Profiler PCR array (Qiagen, 330171) was designed containing primer sequences for the genes shown in Supplementary Table S1. RNA was isolated following primary hepatocyte culture as described above and reversed transcribed using the RT 2 First Strand Kit (Qiagen, 330401). qPCR was performed using RT 2 SYBR Green ROX qPCR Mastermix (Qiagen, 330522) on an ABI 7900HT thermocycler (Applied Biosystems), normalized to Actb and Gapdh expression.

Hepatocyte adhesion assay
Adhesion was assessed using a colorimetric ECM Cell Adhesion Array Kit (Millipore, ECM540) according to the manufacturer's instructions. Primary mouse hepatocytes were isolated as described above, plated in triplicate at 50,000 cells per well, and incubated for two hours at 37 °C in 5% CO 2 . Absorbance was measured at 570nm using a Synergy HT microplate reader (BioTek, Swindon, UK). Relative absorbance was calculated by standardizing to absorbance in the Collagen I well, prior to calculation of mean relative absorbance for each extracellular matrix protein for each sample.

Hepatocyte proliferation assay
Primary mouse hepatocytes were isolated as above and plated at 10,000 cells per well in 24- Imaging was performed using an LSM780 confocal microscope system (Carl Zeiss Ltd, Cambridge, UK). Tiled images were acquired, with three non-overlapping areas of 18µm 2 imaged per well. Imaris (version 8.4.1, Bitplane AG, Zurich, Switzerland) was used to identify the total (Hoechst-positive) nuclei number and the number of EdU-positive nuclei, and the percentage of proliferating nuclei was calculated.

Whole liver microarray
Sample preparation, labeling, and array hybridizations were performed using the Agilent GE The statistical significance of differences between groups was calculated with a 2-tailed Student's t test or Mann Whitney test as appropriate. Differences with a P-value of less than 0.05 were considered statistically significant. PCR data obtained for individual genes were log-transformed prior to analysis and a Bonferroni correction was applied to account for multiple testing. PCR array data were standardized as previously reported, 37 to identify genes in test samples with a 95% confidence interval for standardized relative fold change that did not overlap 1 (the value assigned to the fold change for the same gene in control samples).
For microarray analysis, differential gene expression was examined with the R package limma (version 3.32.7). 38 Quality control was performed by identifying outliers in the log2 intensity between arrays and comparison of multidimensional scaling of distances between microarray expression profiles. Background correction was conducted according to the normexp method and the data were normalized using the quantile normalization method. 39
Assessment of hepatocyte proliferation following two-thirds partial hepatectomy showed significantly increased proliferation in Itgb8 flox/flox ;Alb-Cre mice at 36, 48, and 72 hours following liver injury compared to controls (Fig 1B,C). This increased hepatocyte proliferation was not followed by an increase in hepatocyte apoptosis at day five post partial hepatectomy, when liver regeneration was nearing completion in the Itgb8 flox/flox ;Alb-Cre mouse (Supplemental Fig S1). Interestingly, the proportion of hepatocyte mitoses (identified morphologically) was decreased in Itgb8 flox/flox ;Alb-Cre mice at 72 hours following liver injury compared to controls (Fig 1D). However, liver-to-body weight ratio was significantly increased in Itgb8 flox/flox ;Alb-Cre mice at 72 and 96 hours after partial hepatectomy, demonstrating that the increase in hepatocyte proliferation in Itgb8 flox/flox ;Alb-Cre mice detected by BrdU immunohistochemistry resulted in accelerated restoration of liver mass compared to control (Fig 1E).

Depletion of hepatocyte integrin αvβ8 does not alter baseline hepatocyte proliferation or subsequent inflammatory phenotype
As integrin αvβ8 is able to activate TGFβ, a well-characterized suppressor of epithelial proliferation, it was assessed whether genetic depletion of hepatocyte αvβ8 alters baseline hepatocyte proliferation or liver-to-body weight ratio. Hepatocyte BrdU incorporation and M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT 16 mitoses, and liver-to-body weight ratio was measured in uninjured Itgb8 flox/flox ;Alb-Cre mice and controls (Fig 1B, D, E), and no difference was found in any of these variables between groups. Furthermore, there was no difference in baseline liver biochemistry, hepatic morphology, or resident non-parenchymal cell populations (Kupffer cells and HSCs) between uninjured Itgb8 flox/flox ;Alb-Cre mice and controls (Fig 1F-H, Supplemental Fig S2). Following partial hepatectomy, there was also no difference in hepatic inflammation (Kupffer cells or neutrophils), or HSC immunostaining (Fig 2A-C). This suggests that the increased liver regeneration observed following partial hepatectomy in Itgb8 flox/flox ;Alb-Cre mice was not due to differences in degree of initial injury or the subsequent inflammatory response.

Depletion of HSC integrin αvβ8 does not lead to increased hepatocyte proliferation
Integrin αvβ8 is also expressed on HSCs. 21 As HSCs have been shown to play an important regulatory role in liver regeneration, 43,44 mice in which integrin αvβ8 had been depleted from HSCs (Itgb8 flox/flox ;Pdgfrb-Cre) were used to examine the role of HSC integrin αvβ8 during liver regeneration. Following two-thirds partial hepatectomy, there was no significant difference in hepatocyte proliferation between Itgb8 flox/flox ;Pdgfrb-Cre mice and controls ( Fig 2D). Liver sinusoidal endothelial cells have also been shown to play a key role in liver regeneration. [45][46][47] However, integrin αvβ8 expression was not observed in liver sinusoidal endothelial cells by qPCR.

Assessment of hepatic cell cycle genes following depletion of hepatocyte integrin αvβ8 and partial hepatectomy
To examine whether depletion of hepatocyte integrin αvβ8 might have a direct effect on the cell cycle, the expression of genes with key roles in cell cycle regulation was measured at M A N U S C R I P T

A C C E P T E D ACCEPTED MANUSCRIPT
17 multiple time points following partial hepatectomy. Overall, partial hepatectomy resulted in expected changes in gene expression in whole liver from both Itgb8 flox/flox ;Alb-Cre mice and controls (Fig 3A, B, Supplemental Fig S3). There was a trend towards increased expression of Ccna2 and Ccnb1 in Itgb8 flox/flox ;Alb-Cre mice compared to controls; however, this did not reach statistical significance at any time point (Fig 3A, B). Analysis of other cell cycle-related genes (Ccnd1, Ccne1, Cdkn1a, Cdkn1b) showed no difference between Itgb8 flox/flox ;Alb-Cre mice and controls (Supplemental Fig S3).

Depletion of integrin αvβ8 on hepatocytes does not alter adhesion to multiple matrix proteins present in normal and regenerating liver
Integrin αvβ8 binds extracellular matrix ligands such as vitronectin, fibronectin, collagen IV, and fibronectin. 48,49 To test whether depletion of integrin αvβ8 on hepatocytes alters adhesion to cell matrix proteins present in normal and regenerating liver, an in vitro cell adhesion assay with multiple different matrix substrates was used. No difference was found in adhesion between Itgb8 flox/flox ;Alb-Cre and control hepatocytes across all seven matrix proteins tested (Fig 3C), suggesting that altered hepatocyte adhesion is not responsible for the pro-regenerative phenotype observed in Itgb8 flox/flox ;Alb-Cre mice.

Inhibition of integrin αvβ8 modulates TGFβ-responsive genes in hepatocytes
Integrin αvβ8 has previously been shown to play a key role in the activation of latent TGFβ, 25-31 a potent inhibitor of hepatocyte proliferation. 5-7 Therefore, we hypothesized that depletion of hepatocyte integrin αvβ8 might promote hepatocyte proliferation through modulation of TGFβ signaling pathways. The time course of hepatic Itgb8 expression following partial hepatectomy supports a role for integrin αvβ8 as a suppressor of M A N U S C R I P T

A C C E P T E D ACCEPTED MANUSCRIPT
18 hepatocyte proliferation during liver regeneration (Fig 3D). Hepatic Itgb8 expression falls markedly in the 24 hours immediately following partial hepatectomy, and this downregulation appears to be permissive for hepatocyte proliferation. As the liver approaches full restoration of its functional mass at five days post partial hepatectomy, hepatocyte Itgb8 expression peaks at 10 times baseline expression, consistent with a role for integrin αvβ8 as a brake on hepatocyte proliferation (Fig 3D).
Detecting the modulation of TGFβ activation in the hepatocyte regenerative niche is very challenging, as it is not possible to measure the levels of active TGFβ in tissue directly.
Therefore, an experiment was designed to examine how inhibition of integrin αvβ8 might modulate TGFβ-responsive genes in primary mouse hepatocytes (Fig 3E). Firstly, a custom qPCR array was designed, containing 87 genes either shown to be responsive to TGFβ signaling in hepatocytes 50 or comprising components of the TGFβ pathway (Supplemental Table S1). Primary murine hepatocytes were isolated from wild-type mice and plated onto collagen in the presence of either a β8 integrin subunit blocking antibody or a non-binding control antibody. 26 After incubation for 24 hours, hepatocytes were lyzed, RNA was extracted, and gene expression quantified using the custom qPCR array. Data were then three-fold in wild-type hepatocytes treated with β8 integrin subunit blocking antibody.
Increased expression of Plat was not observed when hepatocytes from Itgb8 flox/flox ;Alb-Cre mice were treated with β8 integrin subunit blocking antibody, suggesting the observed response is specific to β8 integrin subunit inhibition (Supplemental Fig S4). Conversely, hepatocyte expression of the TGFβ-responsive gene Hmox1 (heme oxygenase 1) was downregulated in the presence of β8 integrin subunit blocking antibody. These data demonstrate that inhibition of integrin αvβ8 modulates TGFβ-responsive genes in hepatocytes, suggesting a possible mechanism through which integrin αvβ8 depletion promotes hepatocyte proliferation.

Inhibition of hepatocyte integrin αvβ8 does not alter the proliferative response to mitogenic growth factors
Hepatocyte growth factor (HGF) and epidermal growth factor (EGF) are key drivers of liver regeneration. 8 To investigate whether integrin αvβ8 might have a role in regulating the hepatocyte response to these direct mitogens, the effect of β8 integrin subunit inhibition during in vitro proliferation of primary hepatocytes in response to EGF and HGF was examined. A robust increase in hepatocyte proliferation was achieved with addition of either, or both, EGF and HGF, compared to standard culture medium (Fig 3G). However, inhibition of integrin αvβ8 had no effect on the degree of in vitro hepatocyte proliferation.
This suggests that the accelerated liver regeneration observed following hepatocyte integrin αvβ8 depletion does not occur via modulation of HGF or EGF downstream signaling pathways.  Figure 4C. These terms were relatively non-specific, relating to a range of intracellular metabolic processes.

Microarray analysis of whole liver from control and
Given the increased hepatocyte proliferation observed following partial hepatectomy in  Table S1).

Depletion of hepatocyte integrin αvβ8 does not increase tumor formation in a mouse model of HCC
In addition to promoting hepatocyte proliferation, disruption of TGFβ signaling can accelerate the development of HCC in mice following DEN administration. 12 As depletion of integrin αvβ8 on hepatocytes increased hepatocyte proliferation and accelerated liver regeneration following injury, and blockade of hepatocyte integrin αvβ8 in vitro modulated TGFβ-responsive genes, the possibility that this pro-proliferative phenotype might increase the risk of HCC development was also assessed. Itgb8 flox/flox ;Alb-Cre and control mice were injected with DEN at 12 to 14 days of age to induce HCC (Fig 5A). Following sacrifice at forty weeks, the number and size of tumors in each liver was quantified (Fig 5B). There was no difference in either tumor number or median tumor size between Itgb8 flox/flox ;Alb-Cre and control mice (Fig 5C, D). This demonstrates that depletion of hepatocyte integrin αvβ8 does not predispose to increased tumor formation in this mouse model of HCC.

Human hepatocytes express integrin αvβ8 and represent a viable therapeutic target to promote liver regeneration in patients with liver disease
To assess the potential utility of integrin αvβ8 as a therapeutic target to promote hepatocyte proliferation and liver regeneration in patients with liver disease, the expression of integrin αvβ8 was assessed in samples of human liver. Uninjured liver tissue and tissue obtained from patients with acute liver failure secondary to acetaminophen overdose or from patients with cirrhosis was stained for the β8 integrin subunit. Widespread expression in hepatocytes was detected in all samples, demonstrating that integrin αvβ8 is a viable potential therapeutic target in patients with a broad range of liver diseases (Fig 5E).

DISCUSSION
We show that depletion of hepatocyte integrin αvβ8 leads to increased hepatocyte proliferation and accelerated liver regeneration following partial hepatectomy in mice. The time course of hepatic Itgb8 expression following partial hepatectomy, namely a rapid down-regulation followed by rebound up-regulation as the liver returns to its normal size, is consistent with a role for integrin αvβ8 as a brake on hepatocyte proliferation. This antiproliferative role for integrin αvβ8 appears to be mediated via TGFβ, rather than altered hepatocyte adhesion, since blocking integrin αvβ8 on hepatocytes alters TGFβ-responsive gene expression. Importantly, the augmentation in hepatocyte proliferation in Itgb8 flox/flox ;Alb-Cre mice was not accompanied by increased susceptibility to hepatocellular tumor formation. Finally, human hepatocytes also express integrin αvβ8 in both acute and chronic liver disease, and therefore integrin αvβ8 represents a viable therapeutic target to promote liver regeneration in patients with a broad range of liver diseases.
Integrin αvβ8 has previously been shown to have a key regulatory role in the activation of latent TGFβ. [25][26][27][28][29][30][31] The inhibitory effect of active TGFβ on hepatocyte proliferation is wellestablished, including evidence demonstrating tonic inhibition of hepatocyte proliferation in the uninjured liver. [5][6][7] The rapid down-regulation of hepatic Itgb8 expression observed following partial hepatectomy is in line with the hypothesis that a reduction in integrin αvβ8-mediated activation of TGFβ is permissive for a pro-regenerative environment in the liver. Demonstrating subtle changes in activation status of TGFβ within the hepatic regenerative niche is very challenging, given the magnitude and localized nature of these changes, and also the small amount of remnant tissue present following two-thirds partial M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT 23 hepatectomy. However, inhibition of hepatocyte integrin αvβ8 in vitro, using a β8 integrin subunit blocking antibody, resulted in changes in expression of multiple TGFβ-responsive genes such as Plat and Hmox1. Tissue plasminogen activator, encoded by Plat, can activate HGF 51 and has been shown to play a role in liver lobule reorganization following acute injury. 52 Knockout of tPA in mice worsens injury following bile duct ligation, but this phenotype is reversed by administration of HGF. 53 These data suggest that the regulatory role of integrin αvβ8 during hepatocyte proliferation is, at least in part, mediated via TGFβ signaling, and that integrin αvβ8 depletion or inhibition may drive hepatocyte proliferation through tPA-mediated activation of HGF. Inhibition of hepatocyte integrin αvβ8 did not alter the proliferative response to stimulation with EGF and HGF in vitro, suggesting that the accelerated liver regeneration observed following hepatocyte integrin αvβ8 depletion does not occur via modulation of HGF or EGF downstream signaling pathways.
Detectable changes in the expression of genes regulating the cell cycle were identified following partial hepatectomy, similar to those previously reported, 54 but did not differ between Itgb8 flox/flox ;Alb-Cre mice and controls. Inhibiting integrin αvβ8 in vitro also had no Even following partial hepatectomy, only a minority of hepatocytes will be proliferating at any one time and the presence of non-parenchymal cell mRNA in the whole liver lysates that were analyzed will further reduce the signal-to-noise ratio. their cytoplasmic domains can bind the cytoskeleton. 14 However, it has previously been suggested that the cytoplasmic domain of the β8 subunit does not bind the cytoskeleton. 48 Furthermore, no difference was found in the ability of hepatocytes isolated from Itgb8 flox/flox ;Alb-Cre mice and controls to adhere to multiple extracellular matrix proteins found in both normal and regenerating liver.
Targeting of TGFβ pathways has been a major focus of research across several fields, particularly in the context of inflammation, wound healing, and oncogenesis. Unfortunately, global inhibition of TGFβ signaling can be associated with serious, undesirable effects, including excessive inflammation and development of neoplasia. [10][11][12] This is highly likely to be due to the pleiotropic, context-dependent functions of TGFβ. Selective targeting of TGFβ activation by inhibition of integrin αvβ8 in the hepatic regenerative niche may potentially avoid many of the adverse effects noted with pan-TGFβ blockade, while still promoting the desired effects on hepatocyte proliferation and liver regeneration. Importantly, these results did not demonstrate an increase in either hepatic inflammation or carcinogenesis in mice following depletion of hepatocyte integrin αvβ8.
Human hepatocytes express integrin αvβ8 in uninjured liver, following acute hepatic injury secondary to acetaminophen overdose, and also in cirrhosis. Therefore, hepatocyte integrin αvβ8 appears to be a viable translational target. There are potentially multiple clinical scenarios to which integrin αvβ8 inhibition could be applied. For example, using αvβ8 inhibition as a pro-regenerative therapy in the setting of acute liver failure may obviate the requirement for, or buy more time prior to, liver transplantation. Furthermore, combination with anti-fibrotic therapies could permit the restoration of functional, parenchymal liver M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT 26 mass in tandem with a reduction in fibrosis in patients with chronic liver disease. It might also allow more patients with primary or metastatic liver cancer to be treated successfully.
In summary, depletion of integrin αvβ8 on murine hepatocytes leads to increased hepatocyte proliferation and accelerated liver regeneration. Targeting integrin αvβ8 may therefore represent a promising therapeutic strategy to drive liver regeneration in patients with a broad range of liver diseases.    error bars -SEM. *P < 0.05, **P < 0.01, ****P < 0.0001. Scale bars 100µm.