Aging reduces liver resiliency by dysregulating Hedgehog signaling

Abstract Older age is a major risk factor for damage to many tissues, including liver. Aging undermines resiliency and impairs liver regeneration. The mechanisms whereby aging reduces resiliency are poorly understood. Hedgehog is a signaling pathway with critical mitogenic and morphogenic functions during development. Recent studies indicate that Hedgehog regulates metabolic homeostasis in adult liver. The present study evaluates the hypothesis that Hedgehog signaling becomes dysregulated in hepatocytes during aging, resulting in decreased resiliency and therefore, impaired regeneration and enhanced vulnerability to damage. Partial hepatectomy (PH) was performed on young and old wild‐type mice and Smoothened (Smo)‐floxed mice treated with viral vectors to conditionally delete Smo and disrupt Hedgehog signaling specifically in hepatocytes. Changes in signaling were correlated with changes in regenerative responses and compared among groups. Old livers had fewer hepatocytes proliferating after PH. RNA sequencing identified Hedgehog as a top downregulated pathway in old hepatocytes before and after the regenerative challenge. Deleting Smo in young hepatocytes before PH prevented Hedgehog pathway activation after PH and inhibited regeneration. Gene Ontogeny analysis demonstrated that both old and Smo‐deleted young hepatocytes had activation of pathways involved in innate immune responses and suppression of several signaling pathways that control liver growth and metabolism. Hedgehog inhibition promoted telomere shortening and mitochondrial dysfunction in hepatocytes, consequences of aging that promote inflammation and impair tissue growth and metabolic homeostasis. Hedgehog signaling is dysregulated in old hepatocytes. This accelerates aging, resulting in decreased resiliency and therefore, impaired liver regeneration and enhanced vulnerability to damage.

evaluates the hypothesis that Hedgehog signaling becomes dysregulated in hepatocytes during aging, resulting in decreased resiliency and therefore, impaired regeneration and enhanced vulnerability to damage. Partial hepatectomy (

| INTRODUC TI ON
Older age is a major risk factor for damage to many tissues, including liver (Kennedy et al., 2014). Aging undermines resiliency, i.e., the ability to withstand, and recover from, stressors (López-Otín et al., 2013). Age-related deterioration of liver resiliency is important because this both enhances vulnerability to acute and chronic liver injury and causes defective regenerative responses that promote progressive replacement of functional hepatic parenchyma with scar (i.e., cirrhosis; Frith et al., 2009;Poynard et al., 2001;Stine et al., 2013;Thabut et al., 2006). Indeed, similar to other age-related degenerative diseases, the incidence and prevalence of cirrhosis and potentially lethal complications of cirrhosis, such as liver failure and liver cancer, are increasing as populations age (Maeso-Díaz & Gracia-Sancho, 2020). The underlying mechanisms that suppress resiliency during aging are not well understood and thus, therapies that restore tissue resiliency to a more robust, youthful state remain elusive (Hunt et al., 2019;Pibiri, 2018). Preclinical models that challenge resiliency by imposing a regenerative stress are useful for revealing age-related differences in signaling that associate with reduced resiliency and resultant tissue degeneration. Prior work by us and others demonstrated that aging inhibits liver regeneration after 70% partial hepatectomy (PH) (Nevzorova et al., 2015;Oh et al., 2018;Timchenko, 2009;Timchenko et al., 1998). Therefore, in the present study we used the mouse 70% partial hepatectomy (PH) model to identify age-related differences in signaling pathways that control liver resiliency.
In the present study, deep sequencing and transcriptomic analysis of RNA isolated from hepatocytes before and after this regenerative challenge identified Hedgehog as one of the signaling pathways that was most differentially activated after PH in young versus old hepatocytes. This discovery has intriguing implications for aging because enforcing Hedgehog signaling was recently reported to extend healthy life span in adult flies (Rallis et al., 2020). Previously, we had shown that inhibiting Hedgehog pathway activity systemically, or selectively in liver stromal cells, of young adult mice blocks liver regeneration after PH (Ochoa et al., 2010;Swiderska-Syn et al., 2014. However, more research is necessary to determine if and how activation of this pathway in adult hepatocytes promotes hepatic resiliency because other work by ourselves and others clearly indicates that excessive Hedgehog activity in hepatic stromal cells promotes the evolution of cirrhosis, and shows that uncontrolled Hedgehog signaling in hepatocytes themselves drives hepatocarcinogenesis (Chan et al., 2014;Jung et al., 2007;Verdelho Machado & Diehl, 2018). Therefore, the present study evaluates the provocative hypothesis that Hedgehog signaling becomes deficient in hepatocytes during aging, resulting in decreased hepatocyte resiliency, impaired liver regeneration and therefore, enhanced vulnerability to liver damage. Herein we report definitive evidence that Hedgehog signaling in adult hepatocytes is required to maintain liver resiliency, identify Hedgehog-sensitive mechanisms that promote resiliency, and show that disrupting Hedgehog signaling in young hepatocytes rapidly phenocopies old hepatocytes, thereby identifying a novel mechanism that helps to explain why aging decreases resiliency and enhances vulnerability to liver damage.

| Liver regeneration is impaired with aging
Comparison of young and old wild-type male mice before and at various time points after PH ( Figure S1A) confirmed earlier work (Timchenko, 2009) which showed that aging inhibits liver regeneration. Recovery of pre-PH liver mass was suppressed in old mice relative to young mice ( Figure 1a). In young mice, liver weight started to increase 48h after PH, following the peak in hepatocyte mitosis at that time point (Figure 1b, c, Figure S1B and C). The number of hepatocytes entering mitosis after PH was significantly decreased in old mice, suggesting that the reduced proliferative activity of old hepatocytes slowed liver mass recovery after PH (Figure 1b, c, Figure S1B and C).

| Hedgehog signaling pathway is downregulated in old and old regenerating hepatocytes
To identify mechanisms for the reduced proliferative response in old livers, we used genome set enrichment analysis (GSEA) and Deseq to compare the transcriptomes of young and old hepatocytes before PH and young and old regenerating hepatocytes at 48 h post-PH.
We included only the Hallmarks database, defined a very restrictive FDR cutoff (<0.05) for the exploratory GSEA, and discovered that the Hedgehog signaling pathway was one of the most down- aging, Hedgehog, hepatocytes, liver regeneration, resiliency, smoothened following a regenerative challenge. Together, these findings provide novel evidence that Hedgehog signaling is suppressed in hepatocytes of old livers and cannot be induced appropriately after liver injury. In a second GSEA, we expanded the number of databases included and assessed other developmental and metabolic pathways that are known to be regulated by Hedgehog. We found that several of these pathways, including Notch, Wntβ-Catenin, insulin like growth factor (IGF), and mammalian target of rapamycin (mTOR)), were significantly downregulated in old and old regenerating hepatocytes ( Figure 2b).
To confirm the results derived from the RNA analyses we per- bile duct cells in young livers before PH. 48h after PH, there was a significant increase in the number of hepatocytes expressing Gli2 in the nuclei of young hepatocytes, many of which were undergoing mitosis. However, Gli2 protein could not be detected in the nuclei of old hepatocytes before or after the PH, and the number of mitotic hepatocytes was also reduced significantly in old livers. Therefore, the immunostaining data confirm the mRNA findings and together, demonstrate that hepatocytes in old livers with reduced regenerative capability cannot upregulate the Hedgehog pathway appropriately after PH, suggesting that Hedgehog induction is required to increase hepatocyte proliferation after PH.

| Deletion of Smo in young hepatocytes prevents Hedgehog pathway activation after PH
Previously, we reported that induction of hepatocyte proliferation after PH required activation Hedgehog signaling in myofibroblastic cells , consistent with evidence that these stromal cells are critical sources of hepatocyte mitogens and other F I G U R E 1 Liver regeneration is impaired with aging. Partial hepatectomy (70%) was performed in young (3-monthold) and old (24-month-old) mice. Liver tissue and hepatocytes samples were collected at 0 h and 24, 48, 72 and 96 post-PH. Liver/body weight ratio (a), number of mitotic figures (P-Histone-3 positive Hepatocytes) (b) and representative micrographs of P-Histone 3 immunohistochemistry (c) in liver sections from 0h and 48h following PH in young and old mice (100X magnification). Results shown as MEAN +/− SEM (n = 6 mice/ group/time, *p < 0.05,***p < 0.  Figure   S4D).  Figure S5B, F), a cell cycle protein that controls transition through G1-S and is necessary to increase hepatocyte proliferation after PH (Leong et al., 1964).

| Deletion of Smo in hepatocytes inhibits hepatocyte proliferation after partial hepatectomy
Hedgehog signaling also induces cellular proliferation by upregulating FOXM1, another cell cycle regulator that is required for hepatocyte proliferation after PH (Xiang et al., 2014). Transcript levels of FOXM1 and Ki67, a protein widely used as a proliferation marker, were also significantly downregulated in the Smo-deficient mice ( Figure S5A, 5C and 5D). Together, these results confirm that deleting Smo disrupts Hedgehog signaling in adult hepatocytes and prove that simply abrogating Smo activity in these cells is sufficient to prevent them from progressing through the cell cycle to regenerate the liver after PH.

| Smo (-) hepatocytes resemble aged hepatocytes
Having discovered that old hepatocytes (which are known to have reduced proliferative activity) have impaired Hedgehog signaling and proven that simply disrupting Hedgehog signaling in young hepatocytes is sufficient to inhibit their proliferation, we next sought to identify differentially regulated pathways that could account for this shared 'reduced regenerative' phenotype. Bulk RNAseq was hepatocytes and old hepatocytes also exhibited decreased hepatocyte mitochondrial DNA copy number, as assessed with two important mitochondrially encoded genes; NADH dehydrogenase 1 (MT-ND1) and ribosomal protein S16 (MT-S16), suggesting that the number of mitochondria in these livers was reduced ( Figure 6b). Mitochondrial loss is predicted to compromise energy homeostasis and reducing energy stores is known to restrict growth (Khiati et al., 2015). Therefore, we treated young primary hepatocytes with cyclopamine, a direct Smo antagonist that was shown to inhibit liver regeneration after PH (Ochoa et al., 2010)

| DISCUSS ION
Loss of resiliency occurs during aging and limits the ability of tissues to tolerate and recover from various stresses, leading to gradual degeneration of tissue structure and function. Consistent with this, aging enhances vulnerability to acute liver injury and compromises repair, thereby promoting chronic liver damage. Thus, it is not surprising that older age is a major risk factor for decompensated cirrhosis, the fatal end stage of all types of chronic liver disease. It follows that age-related increases in the incidence and prevalence of cirrhosis and consequent hepatic decompensation could be reduced by restoring hepatic resiliency. However, more research to identify tractable mechanisms that undermine liver resiliency is required in order to achieve this objective.
Herein we report novel evidence that Hedgehog signaling is suppressed in old hepatocytes and show that this helps to explain how aging impairs hepatocyte resiliency by demonstrating that Accumulation of hepatocytes with short telomeres associates with liver atrophy and progressive hepatic dysfunction in young humans, and aging enriches the liver with hepatocytes that have shortened telomeres (Wiemann et al., 2002). In mice, defective telomere maintenance in hepatocytes has been linked to an increased risk for chronic inflammatory conditions, and exacerbates cirrhosis induced by carbon tetrachloride (Sato et al., 2004). Interestingly, simply provoking telomere shortening in the intestinal epithelial lineage was shown to cause intestinal inflammation in mice, inducing an inflammasome gene signature in intestinal epithelial cells and increasing their production of IL1, IL-18, TNFα, and interferon-gamma proteins, leading to intestinal accumulation of cytotoxic T cells. These responses were accompanied by mild intestinal degeneration in young mice that gradually worsened with age, and was rescued by reactivating telomerase (Chakravarti et al., 2020).
Telomerase restores telomere length after DNA damage, and cell cycle progression is generally blocked until the defective telomeres are repaired (Barnes et al., 2019). When challenged by DNA damaging agents, blood cell lineages that activate Smo are able to escape the checkpoint inhibition that is typically induced by DNA damage and thus, can progress through the cell cycle despite having DNA adducts; blocking Smo activation abrogates DNA damage-resistant proliferation in those cells (Scheffold et al., 2020). The mechanisms whereby Smo enables resistance to typical sequelae of DNA damage remain poorly understood but Hedgehog-regulated, Gli-family transcription factors are known to promote telomerase transcription in some cells (Mazumdar et al., 2013). Our findings indicate that Gli-activity is reduced in old hepatocytes and show that telomerase mRNA expression is also reduced in these cells, suggesting that repair of shortened telomeres is reduced by inhibiting Hedgehog signaling and renders old hepatocytes susceptible to checkpoint inhibition that arrests their progression through the cell cycle. On the other hand, telomerase mRNA levels were not decreased in young hepatocytes a week or so after Smo was deleted acutely, although these young Smo (−) cells had shortened telomeres and reduced proliferative activity.
Cell cycle arrest after DNA damage has been linked with reduced expression of circadian clock genes in blood cell lineages with low Hedgehog pathway activity (Scheffold et al., 2020). Interestingly, our GSEA demonstrated decreased 'circadian regulation of gene expression' in Smo-deficient hepatocytes, even in young mice.

More research is necessary to determine how inhibiting Smo
promotes telomere loss and impacts DNA damage responses in hepatocytes. One possibility is that progressive DNA damage and telomere shortening are both by-products of an inflammatory response in liver tissue that is triggered by dysfunctional telomeres in Smodeficient hepatocytes, similar to the tissue inflammation and degeneration that occurred in intestinal mucosa when telomere function was directly disrupted in intestinal epithelial cells (Chakravarti et al., 2020(Chakravarti et al., , 2021Jurk et al., 2014). In this regard, it is important to note that inflammatory cytokines increase mitochondrial membrane permeability and thereby, exacerbate cellular exposure to mitochondria-derived reactive oxygen species (ROS) that cause oxidative DNA damage (Horssen et al., 2019). Oxidative damage to the nuclear genome, in turn, is known to promote cell cycle arrest and telomere loss (Barnes et al., 2019), responses that perpetuate inflammation and drive further tissue degeneration. We discovered It is important to emphasize that our study has limitations and raises questions that will require further research to address.
First, we did not study female mice in order to minimize potentially confounding effects of aging on sex hormones. However, as Hedgehog signaling and many other aspects of liver biology are sexually di-morphic, future studies are needed to determine if reduced Hedgehog signaling also contributes to aging-related decreases in hepatocyte resiliency and liver regenerative capacity in female mice. Second, the hepatocyte RNAseq data presented in this paper were generated by deep sequencing bulk RNA from hepatocyte isolates derived from a relatively small number of Smo-floxed mice. More studies are needed to confirm the reproducibility of our findings but the bulk RNAseq approach is not ideal because it may obscure zonal differences in Hedgehog pathway activity. This is an important consideration since other morphogenic signaling pathways are zonally localized in healthy adult livers, and both hepatocyte metabolism and proliferative activity are known to vary along the hepatic perfusion gradient (Halpern et al., 2017;MacParland et al., 2018;Ramachandran et al., 2019).
To rectify the limitations of our bulk RNAseq analysis, we are currently performing single cell RNA sequencing (scRNA seq) stud- Finally, more research is needed to determine if the aging-related decline in hepatocyte Hedgehog signaling helps to explain why solid epidemiologic evidence demonstrates a decreased incidence of HCC in the very elderly when the incidence of hepatocellular carcinomas (HCC) increases progressively until about 70 years of age (Sheedfar et al., 2013). The authors of a recent review speculate that this paradox might reflect the cumulative effects of age-related declines in insulin growth factor (IGF)-1 signaling (Xu et al., 2013). Loss of IGF1 activity inhibits liver regeneration (Desbois-Mouthon et al., 2006) and thus, is expected to promote tissue degeneration, but it might also curtail the outgrowth of malignant hepatocytes (Adamek & Kasprzak, 2018)

| Hepatocyte isolation
To obtain primary hepatocytes, livers of the remaining mice were perfused with collagenase as described (Oh et al., 2018). Hepatocyte preparations were evaluated by light microscopy to assure that viability and purity were at least 95%. Freshly isolated hepatocytes were immediately processed to obtain RNA and protein.

| RNA-Seq and analysis
Global transcriptome profiling was performed by RNAseq using freshly isolated hepatocytes from young and old wild-type mice (n = 3 mice/group/time point) or Smo-floxed mice (n = 1-2 mice/ group/time point) before and at 48 h after PH. Total RNA was extracted using RNeasy mini kit (Qiagen, Hilden, Germany) following the manufacturer's instructions. mRNA library preparation (poly A enrichment) and mRNA-Seq was performed by Novogene using Illumina NovaSeq PE150 platform. Resulting data was pre-processed: Trim_galore was used to trim off adapter and low-quality reads, STAR for reads alignment to the reference genome (mm10_STAR_genome_ idx), Samtools index sorted bam files, Picard removed duplicates and HTSeq for quantification of the gene expression data. Differential expression analysis and gene set enrichment analysis were performed using Deseq and GSEA, respectively. The dataset is available at the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus database, accession number GSE18176.

| Statistics
Data are expressed as mean ± SEM, unless otherwise specified.
Statistical significance between two groups was analyzed by twotailed Student's t test, whereas comparisons of multiple groups were evaluated by two-way analysis of variance (ANOVA) as specified followed by a post-hoc Tukey's test. p-values <0.05 were considered statistically significant. Drawing graphs and statistical analyses were performed using GraphPad Prism 8 (GraphPad Software, Inc. La Jolla, CA, USA).

Authors recognize the Duke Claude D. Pepper Older Americans
Independence Center (OAIC) and the National Institute on Aging (NIA) for providing the initial funding to start this project and the valuable resource of aged rodent colonies. Authors are grateful to Xiyou Zhou for his mouse colony management.

CO N FLI C T O F I NTE R E S T
The authors have declared that no conflict of interest exists.

AUTH O R CO NTR I B UTI O N S
RMD designed the research studies, conducted experiments, acquired data, analyzed data, provided reagents, wrote the manuscript; GDD, SHO, KD, LT, TC, RKD conducted experiments and acquired data. JHH and JNM designed seahorse experiments and analyzed data. AMD designed the research studies, analyzed data, wrote the manuscript, and secured funding for the research.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data supporting the findings of this study are available within the article and/or supplementary materials.