Aging suppresses sphingosine-1-phosphate (S1P) chaperone ApoM in circulation resulting in maladaptive organ repair

Summary

Here we show that the liver-derived apolipoprotein M (ApoM) protects the lung and kidney from pro-fibrotic insults and that this circulating factor is attenuated in aged mice. Aged mouse hepatocytes exhibit transcriptional suppression of ApoM. This leads to reduced sphingosine-1-phosphate (S1P) signaling via the S1P receptor 1 (S1PR1) in the vascular endothelial cells of lung and kidney. Suboptimal S1PR1 angiocrine signaling causes reduced resistance to injury-induced vascular leak and leads to organ fibrosis. Plasma transfusion from Apom transgenic mice but not Apom knockout mice blocked fibrosis in the lung. Similarly, infusion of recombinant therapeutics, ApoM-Fc fusion protein enhanced kidney and lung regeneration and attenuated fibrosis in aged mouse after injury. Furthermore, we identified that aging alters Sirtuin1-hepatic nuclear factor 4α circuit in hepatocytes to downregulate ApoM. These data reveal an integrative organ adaptation that involves circulating S1P chaperone ApoM⁺HDL which signals via endothelial S1PR1 to spur regeneration over fibrosis.

Introduction:

Organ regeneration in most vertebrates is limited by the aging process (Boon et al., 2013; Chen et al., 2016; Conboy et al., 2005; Gurtner et al., 2008; Han et al., 2018; Katsimpardi et al., 2014; Kusumbe et al., 2016; Loffredo et al., 2013; Martinod et al., 2017; Poss, 2010; Rajagopal and Stanger, 2016; Sinha et al., 2014; Sousounis et al., 2014; Wells and Watt, 2018; Wilhelm et al., 2016). The self-repair capacity of the mammalian lung and kidney is severely compromised in aging, leading to fibrosis and organ dysfunction upon injury (Armulik et al., 2011; Friedman et al., 2013; Hecker et al., 2014; Henderson et al., 2013; Jin et al., 2012; Kotton and Morrisey, 2014; Kramann et al., 2015; Lechner et al., 2017; Noble et al., 2012; Vaughan et al., 2015; Wynn and Ramalingam, 2012; Yang et al., 2010; Zeisberg et al., 2007). Although such regeneration to fibrotic switch in the lung or kidney maybe primarily governed by local mechanisms (Pardo-Saganta et al., 2015), systemic and circulating signals may also play an important role (Dyar et al., 2018; Han et al., 2018). Whether lung or kidney repair is modulated by other organs and the role of specific circulating factors are largely unknown.

Here, we found that a bioactive lipid sphingosine 1-phosphate (S1P) mediates an inter-organ communication network responsible for reparative responses in the lung and kidney. S1P is an extracellular mediator that signals via G protein-coupled receptors in vertebrates (Chun et al., 2010; Gaengel et al., 2012; Proia and Hla, 2015; Yanagida et al., 2020). Recent work suggests that S1P action is facilitated by its chaperones, presumably due to its poor water solubility (Deng et al., 2012; Ding et al., 2016; Galvani et al., 2015; Jung et al., 2012; Obinata et al., 2019). Two critical S1P chaperones, apolipoprotein M (ApoM) and albumin, are both produced by the liver (Christensen et al., 2016; Christoffersen and Nielsen, 2013; Christoffersen et al., 2011). As such, we hypothesized that the liver might regulate S1P signaling in the distant lung or kidney by circulating S1P chaperones.

The expression of S1P chaperones in the liver can be regulated by injury, and ApoM-bound S1P shows preferential signaling via the endothelial S1P receptor 1 (S1PR1). In this study, we found that suppression of ApoM production in the aged liver compromises S1PR1 signaling in the lung and kidney, which causes maladaptive repair and fibrosis. Therapeutically, transfusion of ApoM-rich plasma or injection of recombinant ApoM therapeutic promotes lung and kidney functional repair and lowers fibrosis in aged animals. As such, we describe a mechanism that is compromised in the crosstalk between aged liver, lung, and kidney, which promotes fibrosis in the lung and kidney. We also describe potential therapeutic strategies for aging-related fibrosis.

Results

Enhanced vascular leakage in the injured lung is associated with aging-related fibrosis

To define the mechanisms involved in aging-related lung fibrosis, lung injury was induced in 20 months old (old) mice and 3 months old (young) mice by intratracheal instillation of hydrochloric acid (acid) (Fig. 1A). Compared to young mice, old mice exhibited impaired expansion of type 2 alveolar epithelial cells (AEC2s), recovery of type 1 alveolar epithelial cells (AEC1s), and restoration of blood oxygenation (Fig. 1B–D, Fig. S1A). Thus, lung epithelial regenerative response was suppressed in old mice. In addition, there were enhanced hydroxyproline and collagen amounts (fibrosis) in the injured old lung (Fig. 1E). We then tested whether circulating cues might regulate lung repair. Lung endothelial cells (ECs) of the lung vasculature separate circulating components from perivascular cells, and increased permeability of lung ECs causes vascular leakage and extravasation of plasma and blood cells (Corada et al., 2010; Komarova and Malik, 2010; Murakami et al., 2008; Vila Ellis et al., 2020). Indeed, there was significantly augmented vascular leakage in the injured lung of old mice, compared to that of young mice (Fig. 1F–I). As such, we hypothesized that lung vascular leakage in old mice leads to perivascular extravasation of circulating pro-fibrotic mediators that activates perivascular fibroblasts, the major producer of matrix and contributor to lung fibrosis.

Figure 1. Aging suppresses sphingosine-1-phosphate receptor 1 (S1PR1) signaling in lung endothelial cells (ECs) to cause regeneration to fibrotic switch after hydrochloric acid injury.

(A) Schema describing strategy to test the self-repair capacity of the lung in 3-month-old (young) and 20-month-old (old) mice. Hydrochloric acid (Acid) was instilled into mouse via trachea to induce lung injury, and regenerative and fibrotic responses were measured.

(B-D) Restoration of alveolar gas exchanging function, epithelial architecture recovery, and epithelial proliferation were impaired in old mice after Acid injury. Type 1 alveolar epithelial cell (AEC1) markers aquaporin-5 and podoplanin were co-stained to examine the distribution of this alveolar epithelial cell that mediates gas exchange (C). Tissue architecture was also examined by H&E staining (C), and proliferation of type 2 alveolar epithelial progenitor cell (AEC2) was measured by staining of incorporated EDU and AEC2 marker surfactant protein C (SFTPC) (D). One way ANOVA followed by Tukey’s post hoc test (n = 6). *p < 0.05. Scale bars = 50 μm unless indicated.

(E) Higher hydroxyproline quantity (E) in young and old mouse lung after acid injury. Unpaired t test (n = 4). *p < 0.05;

(F-I) Enhanced vascular leakage (F, G, H) and myofibroblast activation (I) in young and old mouse lung after acid injury. Note that transmission electron microscopy image shows the thickened interstitial space in the lung of old mouse (G). Moreover, we found that red blood cell (RBC) leaks out of the capillary in the old injured lung (arrow). Myofibroblast activation was evaluated by staining of alpha-smooth muscle actin (SMA) and desmin (I). Scale bar = 1 μm in (G). To assess vascular permeability, fluorescent dye-conjugated high molecular weight Dextran or evans blue was intravenously injected to injured mice. Dextran distribution in the vasculature of perfused lung was determined after the section was co-stained with endothelial cell (EC)-specific marker VE-cadherin (F). Evans blue in perfused lung was quantified (H). One way ANOVA followed by Tukey’s post hoc test (n = 5). * p < 0.05.

(J) Heatmap of altered genes in lung endothelial cells (ECs) isolated from described mouse groups. Lung ECs were purified from 3-month-old (young) and 20-month-old (old) mice at day 3 after acid injury for transcriptomic analysis. Upregulated and downregulated genes selectively expressed by ECs are shown to define endothelial molecular alteration in aging-dependent regeneration to fibrotic switch. Sphingosine-1-phosphate receptor 1 (S1pr1) expression level in the lung ECs of old mice was lower than that of young mice.

(K) S1PR1 signaling in the lung vasculature was suppressed in old but not young mice after acid injury. To monitor S1PR1 activation in the injured lungs, S1PR1-GFP reporter mouse was tested. In this mouse line, S1PR1 activation stimulates histone 2B-GFP fusion protein expression in the cell nucleus. Nuclear GFP expression in the indicated mouse groups was examined in the lung section, and GFP was co-stained with EC-specific nuclear marker ETS-response gene (ERG) to assess the expression of Histone2B-GFP in lung ECs.

(L, M) Enhanced lung fibrosis (L) and vascular permeability (M) in young mice with inducible endothelial cell (EC)-specific deletion of S1pr1 (S1pr1^iΔEC/iΔEC) and control (S1pr1^iΔEC/iΔEC) mice after acid injury. Unpaired t test (n = 4). *p < 0.05;

(N) Proposed model showing that compromised lung endothelial S1PR1 signaling in old mice is associated with lung vascular leakage and fibrosis. Because S1PR1 signaling in the lung EC maintains vascular barrier and prevents injury, impaired endothelial S1PR1 signaling in the old lung triggers vascular leakage after acid injury, possibily augmenting activation of perivascular myofibroblast and fibrosis.

Compromised sphingosine-1-phosphate receptor 1 (S1PR1) signaling in lung ECs causes vascular leak and promotes lung fibrosis

To delineate the molecular mechanism involved in enhanced vascular permeability in the old injured lungs, we purified lung ECs from the lung of young and old mice at day 3 after acid injury, respectively. We chose this time point because epithelial proliferation was observed on day 3 after acid injury. RNA sequencing analysis was performed using isolated lung ECs, and gene differential analysis was used to compare the transcriptomic profiles between young and old lung ECs after injury (Fig. 1J). We then searched for the regulator of endothelial permeability in genes differentially expressed in young and old lung ECs. Among the altered genes, the receptor for lipid mediator S1P, S1PR1, was markedly lower in injured old lung ECs than that of young lung ECs. Since S1P modulates vascular permeability via its endothelial G protein-coupled receptor S1PR1 (Oo et al., 2011; Proia and Hla, 2015), we analyzed the kinetics of S1pr1 expression in the lung ECs of young and old mice after acid injury (Fig. S1B). Indeed, S1pr1 was induced in the lung ECs of young mice at day 3 after injury, which persisted for at least 30 days. By contrast, there was little upregulation of S1pr1 in the lung ECs of old mice after Acid injury. Hence, we hypothesized that suppression of endothelial S1PR1 signaling in the lung ECs of old mice might lead to endothelial permeability and vascular leak after injury.

S1PR1 signaling in lung ECs was compared in young and old S1PR1 reporter mice in which S1PR1 activation drives expression of green fluorescent protein (S1PR1-GFP) (Ding et al., 2016). After acid injury, S1PR1 activation (GFP expression) was localized in the lung ECs of young mice but not those of old mice (Fig. 1K, Fig. S1C). To verify the functional contribution of endothelial S1PR1 to endothelial permeability and lung fibrosis, we generated mice with inducible endothelial cell-specific deletion of S1pr1 (S1pr1^iΔEC/iΔEC). Compared to young control mice, Acid injury enhanced extravasation of Evans blue and hydroxyproline in the injured lung of young S1pr1^iΔEC/iΔEC mice, a phenotype that is reminiscent of old mouse lung (Fig. 1L, M). Therefore, it is likely that compromised S1PR1 signaling in lung ECs causes endothelial permeability and extravasation of pro-fibrotic mediators, resulting in lung fibrosis (Fig. 1N).

Suppressed ApoM production in the old liver blocks endothelial S1PR1 signaling in the injured lung

We then set out to seek systemic mediator(s) that might alter S1PR1 signaling in lung ECs. The liver produces the S1P chaperone apolipoprotein M (ApoM), a key constituent of high density lipoprotein (HDL) that promotes vascular homeostasis and suppresses inflammatory and pathologic processes (Blaho et al., 2015; Cartier and Hla, 2019; Swendeman et al., 2017). ApoM⁺HDL/S1P complex acts as a biased agonist to stimulate S1PR1 in lung ECs to preserve the vascular barrier and suppress inflammatory processes (Burg et al., 2018; Christensen et al., 2016; Christoffersen et al., 2011; Kitano et al., 2006). Therefore, we measured the hepatic expression and serum levels of ApoM in young and old mice after acid injury.

After injury, ApoM expression was lower in the old mouse liver than that of young mouse liver (Fig. 2A). Moreover, serum levels of ApoM were lower in old mice than that of young mice at all tested time points after acid injury (Fig. 2B). Reciprocally, lung endothelial S1PR1 signaling in old ApoM transgenic mice overexpressing Apom (Apom^TG) was enhanced compared to old control mice (Fig. 2C). The indispensible role of ApoM in promoting lung endothelial S1PR1 signaling was further evidenced by the diminished lung endothelial S1RP1 signaling in young S1PR1-GFP reporter mice lacking Apom (Apom^−/−) (Fig. 2D). These findings suggest that attenuated ApoM production in old mice might impede S1PR1 activation in lung ECs after Acid injury, and dysfunctional S1PR1 in lung ECs causes endothelial permeability. Indeed, old Apom^TG mice exhibited decreased lung vascular permeability after Acid injury (Fig. 2E, F), and young Apom^−/− mice showed higher extent of endothelial permeability after acid injury (Fig. 2G, H). As a result, lung fibrosis was enhanced in young Apom^−/− mice and blocked in old Apom^TG mice overexpressing Apom (Fig. 2I, J). These data imply that suppressed S1PR1 signaling in lung ECs promotes fibrosis in old injured lung, which might be due to impaired production of ApoM⁺HDL, S1P chaperone, in the aged liver (Fig. 2K).

Figure 2. Downregulation of S1P chaperone, Apolipoprotein M (ApoM) in old liver impedes S1PR1 signaling in lung EC to cause vascular leakage and fibrosis.

(A) Transcriptional level of Apolipoprotein M (ApoM), a major component of HDL and chaperone of S1P selectively stimulating S1PR1 signaling, in the liver of young and old mice after acid injury. Unpaired t test (n = 5). *p < 0.05.

(B) Serum level of ApoM in old and young mice at indicated time points after acid injury (n = 5).

(C) After acid injury, S1PR1 signaling in lung EC was enhanced in old transgenic mice overexpressing ApoM (Apom^TG), an S1PR1-specific biased agonist. To monitor S1PR1 activation in injured lung, Apom^TG mice were bred with S1PR1-GFP reporter mouse and studied in acid injury model. Induced Histone 2B-GFP fusion protein was co-stained with endothelial nuclear antigen ERG.

(D) Deletion of Apom in young mice (Apom^−/−) suppresses endothelial S1PR1 signaling in acid-injured lungs. Apom^−/− mice were crossed with S1PR1-GFP mice and resulting 3-month-old young S1PR1-GFP mice lacking Apom (Apom^−/−) were tested after acid injury. GFP expression was co-stained with endothelial ERG in the lung section. Young S1PR1-GFP mice were used as control.

(E, F) Old ApoM^TG mice exhibited lower extent of vascular leakage in the acid-injured lungs than old control group. Evans blue amounts in the perfused lungs were quantified and compared in indicated mouse groups. One way ANOVA followed by Tukey’s post hoc test (n = 5). * p < 0.05.

(G-I) Young Apom^−/− mice manifested higher extents of vascular leakage (G, H) and fibrosis (I) in injured lung than that of control young mice. Unpaired t test (n = 4). *p < 0.05;

(J) Lower degree of lung fibrosis in old ApoM^TG mice after acid injury than old wild type control mice. Unpaired t test (n = 4). *p < 0.05;

(K) Suppressed expression of ApoM in old liver results in lower amounts of circulating ApoM⁺HDL-S1P. Reduction of this S1PR1-selective biased agonist impairs protective S1PR1 signaling in the lung EC, causing endothelial permeability and vascular leak. Perivascular distribution of S1P and other pro-fibrotic mediators stimulate myofibroblast activation and enhance lung fibrosis.

Transplantation of ApoM-enriched plasma and injection of recombinant ApoM-Fc reduce lung fibrosis in old mice via stimulating endothelial S1PR1 signaling

We further explored the mechanism whereby systemically derived ApoM modulates lung fibrosis. First, plasma was isolated from Apom^TG and Apom^−/− mice and intravenously (i.v.) transplanted to old wild type mice (Christensen et al., 2016) after acid injury, respectively. Influence of transplanted plasma on lung fibrosis in recipient old mice was evaluated (Fig. 3A). Transplantation of Apom^TG plasma but not Apom^−/− plasma enhanced lung endothelial S1PR1 signaling and blocked lung fibrotic responses in the injured recipient mice (Fig. 3B–E). In addition, plasma also enhanced the recovery of AEC1s and blood oxygenation in recipient mice (Fig. 3F–G). These data suggest that increases in circulating ApoM enhanced functional lung repair in old mice after acid injury.

Figure 3. Plasma from Apom^TG mice but not Apom^−/− mice blocks lung fibrosis in old mice after acid injury.

(A) Plasma transplantation experiment to test the contribution of circulating ApoM protein to blocking aging-related lung fibrosis. Hydrochloric acid was instilled into the trachea of old mice. 100 microliters of plasma isolated from Apom^TG or Apom^−/− mice were i.v. injected into injured mice after acid injury, respectively. S1PR1 signaling activation (S1PR1-GFP), blood oxygenation, lung collagen deposition, and hydroxyproline level in the lung were analyzed in recipient mice.

(B) S1PR1 signaling in S1PR1-GFP reporter mice transplanted with Apom^TG or Apom^−/− plasma. Old reporter mice without plasma transplantation were used as control. GFP signal was co-stained with endothelial nuclear marker ERG in the lung section to test endothelial S1PR1 signaling after transplantation of Apom^TG or Apom^−/− plasma.

(C-E) Lung hydroxyproline quantity (C), collagen I deposition (D), and activation of perivascular fibroblasts (E) in injured old mice transplanted with Apom^TG or Apom^−/− plasma. Myofibroblast activation was tested by staining of SMA and desmin.

(F, G) Recovery of pulmonary AEC1s and restoration of blood oxygenation in indicated mouse groups after transplantation of Apom^TG or Apom^−/− plasma.

One way ANOVA followed by Tukey’s post hoc test (n = 5). *, p < 0.05.

We went on to test the therapeutic effects of an ApoM-Fc fusion protein on the promotion of lung regeneration and blockage of fibrosis in old mice (Fig. 4A). This ApoM-Fc fusion recombinant protein was engineered by fusing the ligand binding lipocalin domain of ApoM-Fc to the N-terminal IL-2 signal peptide which allows efficient secretion, and C-terminal Fc domain of IgG to allow in vivo stability (Swendeman et al., 2017). ApoM-Fc retains the capacity to bind to S1P and activates S1P receptors with a preference for endothelial S1PR1 (Ding et al., 2016; Galvani et al., 2015). Importantly, ApoM-Fc strongly induces sustained barrier function because of sustained signaling in the lung ECs. Therefore, ApoM-Fc-S1P can uniquely trigger endothelial S1PR1-dependent protective effects and bypass side effects associated with small molecule S1P-related therapeutics which are designed to induce endocytosis of S1PR1 in lymphocytes.

Figure 4. S1P chaperone-based therapeutic, recombinant ApoM-Fc stimulates lung endothelial S1PR1 signaling to promote regeneration over lung fibrosis in old mice.

(A) Therapeutic effect of recombinant ApoM-Fc fusion protein was tested in old mice after acid injury. 1 mg/kg recombinant ApoM-Fc or vehicle (PBS) was intravenously (i.v.) injected into mice every three days after acid administration. Lung repair was assessed at indicated time points.

(B, C) Proliferation of AEC2s in mice treated with ApoM-Fc or vehicle after acid injury. AEC2 proliferation was determined by co-staining of incorporated Edu with AEC2 marker SFTPC. Representative immunostaining image is shown in (B), and quantification of proliferating AEC2 in treated mice is shown in the bar graph (C).

(D-G) Restoration of AEC1 coverage (D), deposition of collagen (E, F), and activation of myofibroblast (G) in the lung of indicated mouse groups.

(H, I) Therapeutic effect of ApoM-Fc relies on endothelial S1PR1 signaling in injured lungs. In mice lacking endothelial S1pr1 (S1pr1^EC), treatment of ApoM-Fc did not reduce hydroxyproline quantity or collagen deposition level after Acid injury, compared to vehicle treatment.

(J, K) ApoM-Fc did not enhance the proliferation of AEC2s in S1pr1^EC mice after Acid injury, relative to vehicle treatment. Representative immunostaining image is shown in (J). N.S., not statistically different.

Unpaired t test (n = 5). *p < 0.05.

Recombinant ApoM-Fc was injected i.v. to old wild type (WT) mice after acid injury, and lung repair was assessed. Compared with vehicle (PBS) control, administration of ApoM-Fc enhanced proliferation of AEC2s (Fig. 4B, C) and lowered myofibroblast activation and collagen deposition in the lung of old injured mice (Fig. 4D–G). Notably, relative to vehicle, ApoM-Fc did not show higher efficacy in blocking fibrosis or promoting regeneration in the lungs of S1pr1^iΔEC/iΔEC mice lacking endothelial S1pr1 (Fig. 4H–K). Hence, i.v. injection of ApoM-Fc therapeutic reduces fibrosis and increases epithelial proliferative responses in injured lungs at least partially via stimulating endothelial S1PR1 signaling.

Compromised Sirtuin-1 (SIRT1) signaling in old hepatocytes causes suppression of ApoM expression

To establish the influence of hepatocyte-derived ApoM on lung repair, we isolated hepatocytes from Apom^TG mice and transplanted to old mice after partial hepatectomy (Fig. 5A). As shown before, intrasplenic injection of hepatocytes integrated into aged mouse liver after partial hepatectomy (Ding et al., 2010) (Fig. S2A). Transplantation of ApoM-overexpressing hepatocytes reduced vascular permeability (Fig. 5B, C) in injured lungs. These findings imply that ApoM from hepatocyte modulates lung repair, and suppression of ApoM in old liver leads to leaky lung vasculature that augments lung fibrosis.

Figure 5. Reprogramming of SIRT1-Hepatic nuclear factor 4α (HNF4A)-ApoM circuit in old liver modulates S1PR1 signaling in lung repair.

(A) Schema showing approach to test the influence of ApoM-overexpressing (Apom^TG) hepatocytes on lung fibrosis in old mice. Hepatocytes were isolated from Apom^TG mice and transplanted into old mice via partial hepatectomy (PH) and intrasplenic injection, and lung repair was determined after Acid injury.

(B, C) Lung vascular permeability in old mice transplanted with wild type (WT) or Apom^TG hepatocytes (Hepato) after Acid injury. Amounts of Evans blue (B) and high molecular weight Dextran (red fluorescence) (C) in the injured lung were determined after perfusion.

(D, E) Expression of Sirtuin 1 (SIRT1), hepatic nuclear factor 1α (HNF1A), hepatic nuclear factor 4α (HNF4A) in the liver of young and old mice after Acid lung injury. HNF4A was co-stained with hepatocyte marker cholesterol 7 alpha-hydroxylase (CYP7A1) in the liver section.

(F) Generation of hepatocyte-specific Sirt1 knockout of mice. Mice expressing hepatocyte-specific Cre driven by Albumin (Alb-cre) were bred with Sirt1 floxed (Sirt1^flox/flox) mice. Generated young offspring mice with hepatocyte-specific deletion of Sirt1 (Sirt1^ΔAlb/ΔAlb) were tested after Acid injury.

(G, H) Expression of HNF4A and ApoM is suppressed in the liver of Sirt1^ΔAlb/ΔAlb young mice than that of control Sirt1^+/+ mice, recapitulating the phenotype of old mouse liver.

(I) To test whether SIRT1 regulates ApoM expression in hepatocyte and in vivo contribution of SIRT1-ApoM pathway, young Sirt1^ΔAlb/ΔAlb mice were transplanted with plasma from Apom^TG or Apom^−/− mice, respectively.

(J-L) Compared to young control mice and young mice transplanted with Apom^TG plasma, young Sirt1^ΔAlb/ΔAlb mice transplanted with Apom^−/− plasma exhibited enhanced lung vascular leakage (J), augmented perivascular fibroblast activation, impeded alveolar epithelial reconstitution (K), and enhanced hydroxyproline quantity (L) after Acid lung injury. To test perivascular fibroblast activation, desmin was co-stained with SMA and endothelial marker VE-cadherin in the lung section. Distribution of AEC1s was examined by staining of podoplanin and aquaporin-5.

(M) Old mouse liver and hepatocyte (hepato) showed suppressed expression of Gankyrin, a regulatory subunit of proteasome that degrades HNF4A (Jiang et al., 2013; Sun et al., 2011).

(N, O) HNF4A and ApoM are induced by SIRT1 and inhibited by Gankyrin in hepatocytes. Hepatocytes were isolated from old and young mouse liver. Gankyrin was silenced by shRNA (shGank) in the old hepatocytes and Sirt1 was knocked down (shSirt1) in the young hepatocytes, respectively. Protein level of HNF4A was tested by immunoblot (N), and mRNA levels of ApoM and HNF4A were determined (O). Scrambled (Srb) sequence was used as control.

(P) Proposed model of how SIRT1 and Gankyrin differentially regulate HNF4A and ApoM expression in hepatocytes. In young hepatocytes, SIRT1 maintains HNF4A-dependent production of ApoM. In old hepatocytes, upregulation of Gankyrin and SIRT1 downregulation lead to degradation of HNF4A and suppression of ApoM production.

One way ANOVA followed by Tukey’s post hoc test (n = 5). * p < 0.05.

We then explored mechanisms that regulate ApoM production in aged hepatocytes. Hepatic nuclear factor 4α (HNF4A) was shown to initiate ApoM production in the liver, and SIRT1 can promote HNF4A expression (Mosialou et al., 2010; Palu and Thummel, 2016). Indeed, expression of both Sirt1 and Hnf4a were suppressed in the old mouse liver (Fig. 5D, E). To test whether Sirt1 downregulation causes suppression of HNF4A and ApoM, floxed Sirt1 mice were bred with mice expressing hepatocyte-specific albumin-driven Cre (Alb-Cre) to generate mice lacking Sirt1 specifically in hepatocytes (Sirt1^ΔAlb/ΔAlb) (Fig. 5F). Young Sirt1^ΔAlb/ΔAlb mice at 3-month age showed markedly suppressed expression of HNF4A in the liver (Fig. 5G, H). These data suggest that ApoM expression is regulated by SIRT1 in hepatocytes.

To explore the epistatic relationship between ApoM and hepatic SIRT1, we transfused Apom^TG or Apom^−/− plasma to young Sirt1^ΔAlb/ΔAlb mice, respectively (Fig. 5I). While young control mice showed little lung fibrosis after acid injury, young Sirt1^ΔAlb/ΔAlb mice transplanted with Apom^−/− plasma displayed increased endothelial permeability and fibrosis in the injured lungs that is reminiscent of old mouse lung (Fig. 5J). Of importance, the phenotype of young Sirt1^ΔAlb/ΔAlb mice was rescued by Apom^TG plasma transplantation, as evidenced by lower degree of vascular permeability and myofibroblast activation (Fig. 5J–K). Suppressed recovery of AEC1s and enhanced hydroxyproline amounts after lung injury were also reversed in young Sirt1^ΔAlb/ΔAlb mice transplanted with Apom^TG plasma, compared with those injected with Apom^−/− plasma (Fig. 5K, L). Data from these plasma transplantation experiments imply that disruption of SIRT1-HNF4-ApoM circuit in the old liver interferes with lung regeneration and stimulates fibrosis.

In vitro and in vivo silencing of Gankyrin restores ApoM production in old hepatocytes

Mechanisms underlying HNF4A suppression in old hepatocytes were further examined. Gankyrin, a component of the 19S regulatory cap of the proteasome, was shown to be upregulated by aging and contributes to HNF4A degradation (Iakova et al., 2003; Jiang et al., 2013; Sun et al., 2011; Wang et al., 2010a). Indeed, Gankyrin was upregulated in old mouse liver and hepatocytes relative of young liver and hepatocytes (Fig. 5M). Based on this finding, we examined how SIRT1 and Gankyrin differentially regulate HNF4A in aged hepatocytes. Silencing of Gankyrin by shRNA in old hepatocytes elevated the expression of both HNF4A and ApoM (Fig. 5N, O). Meanwhile, knocking down Sirt1 in the young hepatocytes suppressed expression of both HNF4A and ApoM, and knocking down Gankyrin in the old hepatocytes reversed ApoM and HNF4 suppression (Fig. 5N, O). These data suggest that inhibition of the SIRT1-HNF4A pathway by Gankyrin in old liver suppresses production of ApoM, which alters lung vascular function (Fig. 5P).

Because silencing of Gankyrin in cultured hepatocytes restored HNF4A-dependent ApoM production, we tested whether in vivo blocking Gankyrin in the old liver recovers lung regeneration over fibrosis. Adeno-associated virus (AAV) was used to express Gankyrin shRNA (AAV-shGank) and injected to old mice via intraperitoneal injection (Fig. 6A). Injection of shGank restored expression of HNF4A and ApoM in the old liver (Fig. 6B–C, Fig. S2B–C). Furthermore, restoring ApoM in the old liver displayed a series of anti-fibrotic and pro-regenerative effects in the injured lungs, including reduced activation of perivascular fibroblasts, lower hydroxyproline amounts, enhanced restoration of AEC1 architecture and alveolar function, and augmented vascular barrier (Fig. 6D–H). Of note, in vivo silencing Gankyrin in the liver of old ApoM^−/− mice showed little beneficial effects, suggesting that the protective effect of AAV-shGank relies on ApoM production (Fig. 6I–K). As such, suppression of Gankyrin expression promotes pro-regenerative HNF4A-ApoM in the old liver to enhance lung repair and ameliorate fibrosis.

Figure 6. In vivo silencing of HNF4A degrader Gankyrin in the old liver promotes lung repair and abrogates fibrosis.

(A) Schema describing approach to suppress Gankyrin expression in the old liver by intraperitoneal injection of Adeno-associated virus (AAV) expressing Gankyrin shRNA (AAV-shGank). AAV expressing scrambled sequence (Srb) was injected as control.

(B, C) Expression of ApoM and HNF4A were enhanced in the liver of old mice by injection of AAV-shGank. HNF4A was co-stained with CYP7A1 in the liver section (B), and ApoM and HNF4A expression levels in the old liver were tested (C) (n = 5).

(D-H) Injection of AAV-shGank inhibited activation of perivascular fibroblasts, attenuated lung hydroxyproline amounts, reduced pulmonary vascular permeability, promoted alveolar epithelial cell recovery, restored blood oxygenation in the injured lung of old mice. Myofibroblast activation (D), epithelial recovery (D), and hydroxyproline quantity (E) were tested at day 60 after injury, and vascular permeability (F, G) and oxygenation (H) were measured at day 30 after injury. (I-K) In ApoM knockout (ApoM^−/−) mice, AAV-shGank injection did not alter lung hydroxyproline quantity, vascular permeability, or gas exchanging function. These results suggest that the effect of AAV-shGank relies on ApoM expression.

Unpaired t test (E, n = 5), (H, n = 6), (I, n = 5), (J, n = 5), (K, n = 6). One way ANOVA followed by Tukey’s post hoc test (F) (n = 5). * p < 0.05; N.S., not statistically different.

Recombinant ApoM-Fc decreases aging-related kidney fibrosis

Subsequently, we tested whether the reduced SIRT1-ApoM axis in old liver is associated fibrosis in organs other than the lung, such as the kidney. Acute kidney injury was induced in young, old, and young Sirt1^ΔAlb/ΔAlb mice by ischemia and reperfusion procedure (Kidney I/R) (Yang et al., 2010) (Fig. 7A). Kidney function was compromised with enhanced fibrosis in old mice after I/R, in contrast to young mouse kidney without renal injury and fibrotic responses (Fig. 7B–C). I/R similarly increased collagen deposition and myofibroblast activation in the kidney of young Sirt1^ΔAlb/ΔAlb mice, resembling the kidney phenotype of old injured mice (Fig. 7B, C). Moreover, i.v. injection of recombinant ApoM-Fc fusion protein preserved renal function and attenuated kidney fibrosis in young Sirt1^ΔAlb/ΔAlb mice (Fig. 7B, C). Therefore, dysfunctional SIRT1-ApoM axis in old liver augments maladaptive repair and fibrosis in injured kidney.

Figure 7. Suppression of SIRT1-ApoM pathway in aged liver promotes kidney fibrosis after ischemia reperfusion (I/R).

(A) Kidney repair after kidney ischemia reperfusion (I/R) was tested in young and old mice.

(B) Serum creatinine was measured in the indicated mouse groups to assess kidney injury. Old and young wild type mice, and young mice with hepatocyte-specific deletion of Sirt1 (Sirt1^ΔAlb) were subjected to kidney I/R. Young Sirt1^ΔAlb mice were also treated with ApoM-Fc every three days to assess the protective efficacy.

(C) Collagen deposition, expression of SMA, desmin, and VE-cadherin in the kidney of described mouse groups 45 days after I/R.

(D) Activation of S1PR1 signaling in the kidney ECs of indicated mouse groups. Young and old S1PR1-GFP mice were subjected to kidney I/R and nuclear histone2-GFP expression was co-stained with endothelial marker ERG to evaluate S1PR1 signaling in kidney ECs.

(E-G) Old WT mice (E, F) and mice with EC-specific knockout of S1pr1 (S1pr1^ΔEC) (G) were treated with vehicle and ApoM-Fc after kidney I/R, respectively. Levels of serum creatinine and kidney collagen deposition were measured to assess kidney injury and fibrosis.

One way ANOVA followed by Tukey’s post hoc test (B, n = 7 young and old WT, n = 5 young Sirt1^ΔAlb, n = 6 young Sirt1^ΔAlb treated with ApoM-Fc). Unpaired t test (F, n = 7), (G, n = 6), * p < 0.05.N.S., not statistically different.

The molecular mechanism underlying the anti-fibrotic effects of ApoM-Fc in the kidney was defined using S1PR1-GFP reporter mice (Fig. 7D). Young and old S1PR1-GFP reporter mice were subjected to kidney I/R. S1PR1 signaling was activated in kidney ECs of young but not old mice after I/R. Endothelial S1PR1 activation in injured kidney was also absent in S1PR1-GFP mice with hepatocyte-specific knockout of Sirt1 (Sirt1^ΔAlb/ΔAlb), which was augmented by injection of recombinant ApoM-Fc (Fig. 7D). Therapeutic effects of recombinant ApoM-Fc were also determined in old WT mice after kidney I/R. Compared to vehicle, ApoM-Fc attenuated kidney fibrosis and injury in old WT mice after I/R (Fig. 7E, F). Of note, the protective effect of ApoM-Fc was absent in mice lacking endothelial S1pr1 (S1pr1^iΔEC/iΔEC) (Fig. 7G). These experiments suggest that recombinant ApoM-Fc blocks aging-related kidney fibrosis via stimulating endothelial S1PR1 signaling.

Discussion:

Our data show that the liver supplies circulating ApoM⁺-HDL, an S1P chaperone, to act on vascular S1PR1 to regulate lung and kidney repair. We previously demonstrated lung ECs can release pro-regenerative or anti-fibrotic factors (Cao et al., 2016; Ding et al., 2011; Rafii et al., 2016). Here we show that the endothelium serves as gatekeeper for circulatory S1P. Our data also suggest that a mechanism promoting fibrosis in the aged organ might be extravasation of pro-fibrotic mediators activating perivascular fibroblasts (upregulation of SMA). Attenuated S1PR1 signaling in lung or kidney ECs is probably due to low level of circulating ApoM⁺HDL-S1P in aged animals. As such, liver-derived ApoM is a possible regulator that maintains the pro-regenerative function of S1P signaling in different organs.

When ApoM is suppressed in the old liver, compromised endothelial S1PR1 signaling in injured lung or kidney causes endothelial permeability and possible perivascular distribution of pro- fibrotic mediators, stimulating a transition from regeneration to fibrosis. This function of S1P signaling complements previous work which suggests that this bioactive lipid can regulate embryonic stem cell self renewal (Pebay et al., 2005), vascular development (Proia and Hla, 2015), and early embryonic development (Osborne et al., 2008). Previous literature implicates a dramatic transcriptional alteration in the aged liver (Hong et al., 2014; Iakova et al., 2003; Sun et al., 2011). But how this transcriptional reprogramming affects crosstalk between the liver and other organs remains to be defined. Our data suggest that suppression of transcription factor HNF4A and subsequent reduced production of S1P chaperone, ApoM, in old hepatocytes might be a “setpoint” overturning regeneration to fibrosis in injured lung or kidney.

The influence of liver-derived ApoM on lung S1P signaling may be explored in human lung diseases in the future. For example, suppressed production of ApoM in the old liver is likely to cause prolonged lung vascular leakage in human patients with pneumonia or other lung injurious conditions. It also remains to be determined whether lung S1P signaling is more defective in old patients with idiopathic pulmonary fibrosis (IPF), compared to the younger counterparts. In addition, our data show that endothelial S1PR1 signaling is attenuated in conditions where the liver produces lower ApoM production, thus reducing circulating ApoM⁺HDL and its cargo, S1P. Suppression of ApoM might also interfere with S1P signaling in other cell types such as epithelial cells. Indeed, we observed defective lung epithelial restoration in multiple mouse models with dysfunctional ApoM production. This result suggests that there may be link between fibrosis and epithelial function via S1PR1 or other S1P receptors.

HNF4A-ApoM axis is suppressed in the old hepatocytes by two distinct mechanisms: Gankyrin upregulation and SIRT1 downregulation. As such, aging-induced reprogramming of old liver influences S1P signaling in the lung or kidney vasculature to promote maladaptive repair in a systemic manner. Moreover, therapeutic infusion of recombinant ApoM-Fc reinstates regenerative responses over fibrosis in the lung and kidney. Therefore, normalization of suppressed ApoM production might provide a therapeutic approach for age-related fibrosis.

STAR Methods:

RESOURCE AVAILABILITY

Lead Contact,

Further information and requests should be directed to and will be fulfilled by the Lead Contact, Zhongwei Cao (zhongwei.cao@outlook.com).

Materials Availability,

This study did not generate new unique reagents.

Data and Code Availability

RNA sequencing data that support the findings of this study have been deposited in NCBI Gene Expression Omnibus (accession number GSE148893).

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Animal Care and Mouse lines used

Mice were housed in a specific pathogen-free, temperature-controlled facility with a 12-h light/dark cycle in individual ventilated cages. Animals were provided food and water ad libitum. Male and female old animals were randomly utilized in our study because no apparent differences were found between male and female animals in the tested parameters. Animal experiments were carried out by protocols approved by the Institutional Animal Care and Use Committee at Weill Cornell Medicine and Icahn School of Medicine at Mount Sinai. Floxed Sirt1 (Sirt1^flox/flox) and hepatocyte-specific Alb-Cre mice were obtained from Jackson lab. Floxed S1pr1 mice were crossed with endothelial cell-specific Cdh5-Cre^ERT2 to generate EC-specific S1pr1 knockout mice as previously described (Ding et al., 2016). In brief, mice were injected with three consecutive injections of 200 mg/kg tamoxifen, rested for three days, followed by another three consecutive injections. Mice lacking ApoM and ApoM transgenic mice were previously described (Christoffersen et al., 2008) . C57BL/6J mice were obtained from Jackson Laboratories. Sirt1^flox/flox mice were crossed with Alb-Cre mice to establish hepatocyte-specific Sirt1 knockout (Alb-Cre⁺Sirt1^flox/flox) mice and control Sirt1^flox/flox mice. ApoM transgenic mouse overexpressing ApoM (ApoM^TG) and S1pr1-Gfp mouse whereby S1PR1 activation induces GFP expression were previously described (Kono et al., 2014). Deletion of target genes was corroborated by quantitative PCR (Cao et al., 2017).

METHOD DETAILS

Mouse lung injury model

To test mouse lung alveolar regeneration and repair, mouse lung injury model (Cao et al., 2016; Paris et al., 2016) was adapted. Mice were anesthetized with a cocktail of ketamine (100 mg/kg) and xylazine (10 mg/kg). Intratracheal hydrochloric acid (acid) injection model was employed to induce lung injury (Paris et al., 2016). Orotracheal instillation was performed in anesthetized mice, and 20 μl of an iso-osmolar solution of 0.1 M hydrochloric acid (pH 1.0) was instilled. After injection, mice were observed to ensure the full recovery from anesthesia, and body temperature was maintained using external heat sources. After recovery, mice were transferred to ventilated cages with access to food and water. Lung ECs were purified from 20 months (old) and 3 months (young) old mice after injury, and total RNA was isolated and subjected to transcriptomic analysis (Ding et al., 2010; Nolan et al., 2013). To measure cell proliferation, 100 mg/kg 5-ethynyl-2’-deoxyuridine (EDU) was i.p. injected into mice one hour before sacrificing, and EDU incorporation was measured by EDU cell proliferation kit. Mice were i.p. supplemented with 1 ml PBS after all surgery procedures. At indicated days, mice were killed and oxygen tension in the arterial blood were measured with I-Stat per vender’s instruction (Rafii et al., 2015).

Mouse kidney ischemia reperfusion (I/R) model

Mouse kidney I/R was generated as previously described (Yang et al., 2010). In brief, kidney ischemia was induced in anesthetized mice with a retroperitoneal approach on two kidneys for 30 min. After removal of the clamps, the kidneys were monitored for restoration of blood flow. Then the abdomen was closed. Mice were treated 1 mg/kg ApoM-Fc (Swendeman et al., 2017) every three days after kidney I/R. At indicated time, serum creatinine concentration was measured by the picric acid method, and tissue morphology and fibrosis were analyzed by Sirius red and hematoxylin co-stainings.

Analysis of lung vascular permeability

To measure the pulmonary vascular permeability, 150 μl of 5 mg/ml evans blue (Sigma) in PBS was injected through jugular vein 1 h before sacrificing the animals. Residual blood in the pulmonary vasculature was flushed out by PBS perfusion via right ventricle. Lungs were removed, rinsed, and homogenized in formamide. Lung Evans Blue dye was extracted by incubation at 55 ^oC for 24 h and centrifugation at 5000 g for 15 min. Supernatants were aliquoted and 100 μl was used to measure optical density at 620 nm. Evans blue in the lung was quantified by comparing with standard curve derived from serial dilutions of evans blue standards in formamide.

Isolation and transplantation of hepatocytes

Hepatocytes were isolated from mouse liver and cultured (Ding et al., 2010). For hepatocyte transplantation, intrasplenic injection was performed as described (Ding et al., 2010). Ten million fluorescently labeled mouse hepatocytes were transplanted into recipient mice after partial hepatectomy (PH). Mice were anaesthetized by 100 mg/kg intraperitoneal ketamine and 10 mg/kg xylazine. Midline laparotomy was performed in the anaesthetized mice, and three most anterior lobes (right medial, left medial and left lateral lobes) containing 70% of the liver weight were surgically removed. Briefly, after opening the upper abdomen and the exposure of the liver, the left lobe to be removed was lifted. A 5-0 silk suture tie (Roboz) was placed under the lobe and positioned to the origin of the lobe. After three knots were tied, the tied lobe distal to the suture was resected by a microdissecting scissor. This surgical procedure was then repeated for the other median lobes to complete PH procedure. Following surgical removal of 70% of liver mass, the peritoneum was re-approximated, and the skin was closed. Sham-operated mice underwent laparotomy without lobe resection. Recipient mice were then subjected to acid-mediated lung injury and sacrificed at day 30 after acid injection to assess lung repair and fibrosis. Incorporation of transplanted GFP-labeled hepatocytes in the recipient liver was determined by fluorescent microscopy (Ding et al., 2010).

Immunostaining and histological analysis of lung and liver cryosections

Lung and liver tissues were harvested for histological analysis (Ding et al., 2014; Ding et al., 2010). Mouse tissues were fixed with 4% paraformaldehyde (PFA) and cryopreserved in OCT. For immunofluorescent (IF) microscopy, the sections (10 μm) were blocked (5% donkey serum/0.1% Tween80) and incubated in primary antibodies. After incubation in fluorophore-conjugated secondary antibodies, nuclear staining was carried out with DAPI (Invitrogen, CA) using Prolong Gold mounting medium (Invitrogen). To determine the immunofluorescent staining signal in the prepared tissue sections, fluorescent cells in each slide were independently evaluated on five different high-power fields and quantified, representing the results for individual specimen.

Tissue fibrosis and serum ApoM determination

Lung tissues were harvested for fibrosis analysis at the indicated time after injury. Sirius red and hematoxylin stainings were performed on paraffin-embedded tissue sections to determine the tissue morphology and Collagen deposition and distribution (Cao et al., 2016). Sirius red-positive fibrotic parenchyma was determined from five random fields in each section and quantified using Image J (Ding et al., 2014). Hydroxyproline amount was quantified in the lung to determine the extent of fibrosis. To measure ApoM concentration, blood was collected from mice via vena cava, and serum was prepared from coagulated blood. Serum ApoM concentration was measured by ELISA kit (MolBio, Shanghai).

Transplantation of plasma from Apom^−/− and Apom^TG mice and administration of recombinant ApoM-Fc fusion protein

Blood was collected from Apom^−/− or Apom^TG mice via vena cava, respectively. Plasma was prepared from anti-coagulated blood by centrifugation (150g for 5 minutes). 100 microliters of plasma from Apom^−/− or Apom^TG mice were injected intravenously into old recipient mice via jugular vein or tail vein every six days after acid injury (Christensen et al., 2016). Recipient mice were sacrificed at indicated time points, and tissues were isolated for analyses of fibrosis, vascular leakage, or endothelial signaling.

Recombinant ApoM-Fc fusion protein was engineered and produced as previously described (Swendeman et al., 2017). In brief, the ligand binding lipocalin domain of ApoM-Fc was fused to the N-terminal IL-2 signal peptide which allows efficient secretion, and C-terminal Fc domain of IgG to allow in vivo stability. Recombinant ApoM-Fc was homogenously purified by a two-column procedure. Then ApoM-Fc was allowed to bind to S1P and repurified, which resulted in S1P chaperone that is occupied >50% mol/mol. Quality control procedures described before (Swendeman et al 2017) were done on every lot of ApoM-Fc-S1P complex. 1 mg/kg ApoM-Fc in sterile PBS intravenously injected into recipient mice after lung or kidney injury. Same volume of vehicle PBS was used as control.

Image acquisition and analysis

Histology analysis and Sirius red staining of lung slides were captured with Olympus BX51 microscope (Olympus America, NY), and fluorescent images were recorded on AxioVert LSM710 confocal microscope (Zeiss). Digital images were analyzed using Image J (NIH, MD). Investigators that performed experiments and determined the extent and pattern of staining were randomly assigned with animal samples from different experimental groups and were blinded to the genotype of samples.

Generation of AAV to knockdown Gankyrin expression in the liver

AAV was generated using HEK293T cells. Designed Gankyrin shRNA was cloned into BamH1 and HindIII sites of double stranded AAV vector AAV-CMV-EGFP, and three plasmid transfection system was used to generate AAV. AAV vector with Gankyrin shRNA was mixed with pHelper and pAAV9 at the ratio of 1:1:1 and used to transfect HEK293T cells using calcium phosphate method. Twenty-four hours post-transfection, the medium was removed from the transfected cells and replaced with fresh medium. Forty-eight to seventy-two hours post-transfection, virus was harvested from the 293T cells. To isolate AAV from the 293T cells, the transfected 293T cells were scraped into tubes and centrifuged at 200g for 5 min at 4°C. Cell pellets were gently resuspended in PBS, frozen completely at −80°C for 15min and incubate 37°C for 5 min in water bath. This process was then repeated for three times. To knock down Gankyrin in mouse liver in vivo, a titer of 3 × 10¹¹ viral genome copies AAV expressing EGFP or shGank (AAV-shGank) were i.p. injected in 200 μL PBS. Mice were subjected to acid injury after 1 month, and expression of GFP and Gankyrin in mouse organs were compared by immunostaining and qPCR.

QUANTIFICATION AND STATISTICAL ANALYSIS

All calculations or analyses were performed using Prism 8 software package (GraphPad). For datasets containing two groups, an unpaired two-tailed Student’s t test was employed to determine significant differences. For datasets containing more than two groups, one-way ANOVA followed by Tukey’s post hoc test was used to assess significant differences. The statistical details of experiments can be found in the figure legends, and n represents the number of individual animals. All data are presented as mean ± standard error of mean (S.E.M). Center represents mean and error bar represents S.E.M. P values of < 0.05 were considered statistically significant.

Key Resources Table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Experimental Models: Organisms/Strains
Mouse: Floxed S1pr1	(Allende et al., 2003)	RRID: MGI:2681963
Mouse: Apom transgenic	(Christoffersen et al., 2008)	PMID: 18006500
Mouse: Apom knockout	(Christoffersen et al., 2008)	PMID: 18006500
Mouse:Cdh5-CreERT2	(Wang et al., 2010b)	RRID: MGI:3848984
Mouse: Albumin-Cre	The Jackson Laboratory	RRID: IMSR_JAX: 003574
Mouse: S1pr1-Gfp	(Kono et al., 2014)	PMID: 24667638
Mouse: C57BL6/J	The Jackson Laboratory	N/A
Bacterial and Viral Strains
AAV9	(Gao et al., 2004)	PMID:15163731
Antibodies
Rabbit polyclonal anti-Collagen I	Abcam	Cat#ab34710
Rabbit polyclonal anti-SMA	Abcam	Cat#ab5694
Rabbit polyclonal anti-desmin	Abcam	Cat#ab15200
Goat polyclonal anti-VE-Cadherin	R&D	Cat#AF1002
Rabbit polyclonal anti-SFTPC	Abcam	Cat#ab28744
Rabbit polyclonal anti-Aquaporin	Abcam	Cat#ab78486
Goat polyclonal anti-Podoplanin	R&D	Cat#AF3244
AF647 donkey anti-rabbit IgG	Jackson ImmunoResearch Labs	Cat#711-605-152
AF488 donkey anti-rabbit IgG	Jackson ImmunoResearch Labs	Cat#711-545-152
AF647 donkey anti-goat IgG	Jackson ImmunoResearch Labs	Cat#705-605-147
AF488 donkey anti-goat IgG	Jackson ImmunoResearch Labs	Cat#705-545-147
Deposited data
RNA Seq raw data	This paper	GEO
Software
Image J 1.51 s	NIH	https://imagej.nih.gov/ij/
Graphpad Prism 8.0	Graphpad Software	https://www.graphpad.com/scientificsoftware/prism/
Photoshop	Adobe	https://www.adobe.com/products/photoshop.html
Recombinant DNA
ApoM-Fc fusion construct	(Swendeman et al., 2017)	PMID: 28811382
Chemicals, Peptides, and Recombinant Proteins
Tamoxifen	Sigma-Aldrich	Cat#T5648
Corn Oil	Sigma-Aldrich	Cat#C8267
Collagenase I	Roche	Cat#11088793001
Dispase II	Roche	Cat#04942078001
RIPA Lysis Buffer	Santa Cruz	Cat#sc-364162
Phosphatase Inhibitor Cocktail	Roche	Cat#04906837001
Protease Inhibitor Cocktail	Roche	Cat#04693159001
Lipofectamine™ RNAiMAX Transfection Reagent	Invitrogen	Cat#13778150
EndoGRO-VEGF Complete Culture Media Kit	Millipore	Cat#SCME002
DMEM	Gibco	Cat#11965092
DAPI	Invitrogen	Cat#D1306
Ketamine	Akorn	NDC:59399-114-10
Xylazine	Akorn	NDC:59399-110-20
PFA	Electron Microscopy Sciences	Cat# 50-980-487
Hydrochloric acid	Fisher Scientific	Cat# A481-212
Tween80	Sigma-Aldrich	Cat#P1754
ApoM-Fc	(Swendeman et al., 2017)	PMID: 28811382
Critical Commercial Assays
Mouse ApoM Elisa Kit	Mlbio Biotechnology	Cat# ML722506
Hydroxyproline (Hyp) Kit	Abcam	Cat# ab222941
RNeasy Mini Kit	QIAGEN	Cat# 74104
EDU Proliferation Kit	BD Pharmingen	Cat# 565456
Oligonucleotides
shRNA targeting sequence: Gankyrin 5’- CCGGG CAGCT TCGAA GAATA GGCAT CTCGA GATGC CTATT CTTCG AAGCT GCTTT TTG -3’	This paper	N/A