Potential conflict of interest: ASC-J9® was patented by the University of Rochester, the University of North Carolina, and AndroScience and then licensed to AndroScience. Both the University of Rochester and C.C. own royalties and equity in AndroScience.
This work was supported by the National Institutes of Health (grant no.: CA122295) and the Taiwan Department of Health Clinical Trial and Research Center of Excellence (grant no.: DOH99-TD-B-111-004; China Medical University, Taichung, Taiwan). HSC-T6 and LX2 cells were kindly provided by Dr. Scott L. Friedman (The Mount Sinai Hospital, New York, NY).
Transplantation of bone marrow mesenchymal stem cells (BM-MSCs) has been considered as an alternative therapy, replacing liver transplantation in clinical trials, to treat liver cirrhosis, an irreversible disease that may eventually lead to liver cancer development. However, low survival rate of the BM-MSCs leading to unsatisfactory efficacy remains a major concern. Gender differences have been suggested in BM-MSCs therapeutic application, but the effect of the androgen receptor (AR), a key factor in male sexual phenotype, in this application is not clear. Using two liver cirrhosis mouse models induced by CCl4 or thioacetamide, we showed that targeting AR in the BM-MSCs improved their self-renewal and migration potentials and increased paracrine effects to exert anti-inflammatory and anti-fibrotic actions to enhance liver repair. Mechanism dissection studies suggested that knocking out AR in BM-MSCs led to improved self-renewal and migration by alteration of the signaling of epidermal growth factor receptor and matrix metalloproteinase 9 and resulted in suppression of infiltrating macrophages and hepatic stellate cell activation through modulation of interleukin (IL)1R/IL1Ra signaling. Therapeutic approaches using either AR/small interfering RNA or the AR degradation enhancer, ASC-J9®, to target AR in BM-MSCs all led to increased efficacy for liver repair. Conclusion: Targeting AR, a key factor in male sexual phenotype, in BM-MSCs improves transplantation therapeutic efficacy for treating liver fibrosis. (HEPATOLOGY 2013;57:1550–1563)
Chronic liver disease (CLD) with cirrhosis is the twelfth leading cause of death, with 27,555 deaths each year in the United States,1 and the 10-year mortality rate is 32%-66%,2 depending on the cause of the cirrhosis. Many factors may contribute to liver cirrhosis, including hepatitis B, hepatitis C, alcoholic liver disease, hepatotoxic drugs, and toxins. Among those factors, chronic alcoholism and hepatitis C are the most common causes to induce cirrhosis in the western world. Because patients with liver cirrhosis may also develop hepatocellular carcinomas (HCCs),3 early treatment of liver cirrhosis with proper therapy will not only improve cirrhotic symptoms, but also prevent HCC incidence.
Current treatment for liver cirrhosis is to prevent further damage of functional hepatocytes, and liver transplantation (LT) remains the standard treatment for advanced liver cirrhosis. However, the efficacy of LT is limited by the availability of donor organs and several adverse effects, such as graft-versus-host disease, mental changes, and complications resulting from perioperation,4, 5 which all may complicate this therapy. Therefore, alternative therapeutic approaches, such as transplantation of hepatocytes, hematopoietic stem cells, endothelial progenitor cells, and bone marrow mesenchymal stem cells (BM-MSCs),6-10 may become other options.
BM-MSC transplantation has been extensively studied in clinical trials. Clinical outcomes displayed short-term relief in liver function. But, unfortunately, long-term observations showed failure with recurring symptoms,11 and only low numbers of BM-MSCs finally migrated to the target liver, which could be the result of high apoptotic rates of BM-MSCs from microischemia.12 Therefore, improving self-renewal and survival of BM-MSCs may become a key step to improve the efficacy of using BM-MSCs to treat cirrhotic livers.
Gender differences have been suggested in BM-MSC transplantation therapy of myocardial infarction, because female BM-MSCs have better therapeutic potential than male BM-MSCs.13 It also has been shown that expression of the androgen receptor (AR), a key factor in male sexual phenotype, was increased as embryonic stem cells proceeded to differentiation, a process initiated after cell proliferation stopped,14 implicating that interruption of androgen/AR signaling may lead to increased stem cell proliferation. Here, we found that targeting AR in BM-MSCs significantly enhanced the self-renewal of BM-MSCs. BM-MSCs treatment using either AR/small interfering RNA (siRNA) or ASC-J9®, an AR degradation enhancer, both lead to better therapeutic efficacy to treat liver cirrhosis.
CCl4 (Sigma-Aldrich, diluted 1:1 in olive oil; Sigma-Aldrich, St. Louis, MO) or vehicle (olive oil) was administered by intraperitoneal (IP) injection at a dose of 1 mL/kg of body weight twice per week to induce liver cirrhosis. Transplantation of BM-MSCs of wild type (WT) and AR knockout (ARKO) was performed by tail vein injection 1 day after the eighth CCl4 treatment. For knocking down BM-MSCs AR with siRNA, BM-MSCs were infected with lentivirus, which contained either scramble control or AR-siRNA, and green fluorescence protein (GFP). BM-MSCs containing more than 90% GFP-positive signals were prepared to perform transplantation. For ASC-J9®-treated BM-MSCs, BM-MSCs were cultured and treated with either dimethyl sulfoxide (DMSO) or 10 μM of ASC-J9® for 1 week before transplantation. After 8 weeks of treatments (total 16 treatments) with CCl4, mice were sacrificed with pentobarbital, mouse livers were removed to examine for fibrosis, and mouse sera were isolated to assay for liver functions. Thioacetamide (TAA) was administered at a concentration of 300 mg/L in drinking water for 8 weeks to induce liver cirrhosis. Similarly, BM-MSCs transplantation was performed after 4-week treatment, and mice were sacrificed after an additional 4 weeks of TAA challenge. Other materials and methods, including cells, animals, reagents, isolation of BM-MSCs, histologic analysis of liver tissues, liver functional assay, colony-forming unit-fibroblast (CFU-f) assay, cell-growth assay, flow cytometry, western blotting analysis, quantitative reverse-transcriptase polymerase chain reaction (qRT-PCR), lentivirus production, migration, and statistical analysis, are described in the Supporting Materials.
Knockout of AR in BM-MSCs Improves Their Therapeutic Potential in Liver Cirrhosis.
We first characterized the purity of primary isolated BM-MSCs from mice by checking their surface marker expressions and found they were highly positive for CD44 (99.0%) and CD29 (98.5%) and moderately positive for CD117 (29.4%) and CD106 (32.5%), but only had little expression of CD34 (0.59%) and CD45 (1.28%) (as shown in Supporting Fig. 1). These results were consistent with previous reports.15 We also compared surface marker expressions between WT BM-MSCs and ARKO BM-MSCs to see whether knockout (KO) of AR changes BM-MSCs subtypes. The results showed no significant difference between WT and ARKO BM-MSCs (Supporting Fig. 2). We then compared ARKO BM-MSCs versus WT BM-MSCs for their transplantation therapeutic efficacy in CCl4-induced liver cirrhotic mice. Because BM-MSCs express Cre recombinase (Cre) and GFP, we therefore used these two markers to monitor transplantation efficacy. First, we checked whether BM-MSCs were able to migrate to the liver and found both Cre expressions and GFP signals were detected in BM-MSCs-transplanted mouse liver tissues, but not in liver tissues from healthy control or untransplanted CCl4-treated mice (Supporting Fig. 3), indicating that transplanted BM-MSCs indeed migrated to the cirrhotic liver.
Histological analysis revealed that WT BM-MSCs-transplanted livers showed lower levels of collagen (shown by Picro Sirus Red staining; Fig. 1A-a), fibronectin (Fig. 1A-b), and alpha smooth muscle actin (α-SMA; Fig. 1A-c), compared to untransplanted mice. Importantly, we found that ARKO BM-MSCs-transplanted liver showed even much lower fibrotic marker expressions than WT BM-MSC-transplanted liver (Fig. 1A-a-c), suggesting that WT BM-MSCs transplantation improved liver cirrhosis and that ARKO BM-MSCs further enhanced this transplantation therapeutic efficiency. Liver functional assays, performed by measuring aspartate aminotransferase (AST), alanine aminotransferase (ALT), and albumin (Fig. 1B-d-f), all showed consistent results in that BM-MSCs-transplanted mice had improvement in liver cirrhosis and that KO of AR in BM-MSCs increased transplantation therapeutic efficacy.
We further confirmed these findings in the second liver cirrhosis mouse model in TAA-induced liver cirrhosis mice with consistent results (Fig. 1C,D-g,h), showing better therapeutic outcome in ARKO BM-MSCs-transplanted mice than WT BM-MSCs-transplanted mice, compared to untransplanted CCl4-treated mice.
ARKO in BM-MSCs Demonstrates Improved Therapeutic Effects by Antifibrotic Action in Cirrhotic Livers.
BM-MSCs transplantation has been shown to improve autoimmune disease, sepsis, and myocardial infarction through anti-inflammatory effects. Pro-inflammatory and pro-fibrogenic signals have been linked to liver fibrosis16 and showed that BM-MSCs improved liver cirrhosis through antifibrosis by down-regulating transforming growth factor beta 1 (TGFβ1).17 We examined the expression of the fibrotic marker, TGFβ1, in mice liver tissues and found that ARKO BM-MSCs-transplanted livers showed lower expressions of TGFβ1 and TGFβ receptor 2, compared with WT BM-MSCs-transplanted mice (Fig. 2A-a-c and Supporting Fig. 4A).
We then examined the proliferation of myofibroblasts with double immunofluorescence (IF) staining in liver tissues using antibodies (Abs) of α-SMA and proliferating cell nuclear antigen (PCNA) and found decreased numbers of double stained cells (indicating less proliferating myofibroblasts) in liver tissue of ARKO BM-MSCs-transplanted mice, compared with those transplanted with WT BM-MSCs (Fig. 2B-d,e), suggesting that ARKO BM-MSC transplantation in mice inhibited fibrosis more significantly.
Tissue inhibitor of metalloproteinase 2 (TIMP-2) has been shown to possess antiapoptotic effects on hepatic stellate cells (HSCs) and plays an important role in promoting liver cirrhosis.18 We found that BM-MSCs-transplanted livers have decreased TIMP-2 expression, compared to livers without transplantation. More important, it was shown that ARKO BM-MSCs-transplanted livers showed even lower TIMP-2 expression level, compared with WT BM-MSCs-transplanted mice (Fig. 2B-f), suggesting that HSCs in BM-MSCs-transplanted mice have higher apoptotic potential than untransplanted mice and that knockout in BM-MSCs enhanced this potential.
Depletion of AR in BM-MSCs Improves Its Therapeutic Effects by Anti-inflammatory Action.
Clinically, it has been shown that patients with liver cirrhosis have higher circulating cytokines, including interleukin (IL)-1β, IL-6, and tumor necrosis factor alpha, than healthy patients.19 The increased circulating cytokines could then elevate the circulating monocytes that lead to enhance monocyte/macrophage infiltration in damaged livers.19, 20 We observed lower numbers of F4/80 positively stained cells (indicating infiltrating macrophages) in BM-MSCs-transplanted mice livers, compared to untransplanted mice, and even lower numbers of infiltrated macrophages were detected in ARKO BM-MSCs-treated mice (Fig. 2C-g,h). We also found that ARKO BM-MSCs-treated livers have significantly reduced expression of monocyte chemotactic protein-1 (MCP-1; an indicator of anti-inflammatory action in liver tissues) (Fig. 2C-i), suggesting that transplantation of ARKO BM-MSCs does exert potent anti-inflammation effects in fibrotic livers.
To further confirm anti-inflammation effects, we also examined expressions of IL-1 receptor antagonist (IL1Ra), hepatocyte growth factor (HGF), vascular endothelial growth factor (VEGF), and matrix metalloproteinases (MMPs) that have been linked to BM-MSC-mediated anti-inflammation/anti-fibrosis therapies in myocardial infarction and liver cirrhosis.7, 21-23 We found significantly increased IL1Ra expressions, at both messenger RNA (mRNA) and protein levels, in ARKO BM-MSCs-transplanted livers, as compared with those transplanted with WT BM-MSCs (Fig. 2D-j-l), and transplanted BM-MSCs are the major cells secreting IL1Ra (Supporting Fig. 5A,B). However, we detected no significant difference in HGF, VEGFB, and VEGFC expressions (Supporting Fig. 4C-E). Surprisingly, we observed significantly reduced MMP-2 and -9 expressions upon BM-MSC transplantation, and ARKO BM-MSCs showed better reduction than WT BM-MSCs (Supporting Fig. 4F,G), implying that BM-MSCs transplantation inhibited inflammatory response. These results are consistent with the clinical observation showing MMP-2 and -9 were elevated in chronic and inflammatory liver disease patients,24, 25 but opposite to the report proposing BM-MSCs therapeutic effects through elevating MMPs.23
AR Suppresses Self-Renewal and Migration Potentials of BM-MSCs.
To dissect the potential mechanisms by which knockout of AR in BM-MSCs could lead to better transplantation efficacy through anti-inflammation/anti-fibrosis signals, we investigated the self-renewal and migration potentials of BM-MSCs that have been shown to improve therapeutic outcomes on myocardial infarction and liver cirrhosis through anti-inflammatory and anti-fibrotic actions.26-28 We found higher self-renewal potential in ARKO BM-MSCs than WT BM-MSCs using the CUF-f assay29 (Fig. 3A-a). Western blotting analysis also showed higher PCNA expression in ARKO BM-MSCs than WT BM-MSCs (Fig. 3A-b).
We then dissected the mechanisms by which ARKO BM-MSCs have higher self-renewal ability, and found that knocking out AR in BM-MSCs led to activation of extracellular signal-related kinase 1 and 2 (Erk1/2) and protein kinase B (Akt) signals and their upstream signal, endothelial growth factor/endothelial growth factor receptor (EGF/EGFR; Fig. 3A-c,d), suggesting that AR in BM-MSCs might be able to promote the self-renewal potential through modulation of EGF-Erk1/2 and EGF-Akt signals.
It is interesting to know whether human MSCs (hMSCs) also express AR and whether knockdown of AR in hMSCs results in the similar mechanistic regulation as observed in mouse models. We demonstrated that hMSCs have detectable AR expression (Supporting Fig. 6A). Knockdown of AR in hMSCs enhanced EGFR expression to result in activation of Akt and Erk1/2 (Supporting Fig. 6B,E,F).
We then examined AR knockout effect in BM-MSCs on cell migration using Boyden chamber assays, and found that ARKO BM-MSCs have higher migration ability than WT BM-MSCs, as demonstrated by positively stained migrated cells (Fig. 3B-e,f).
We then dissected the mechanisms by which the ARKO BM-MSCs have higher migration ability, and found that ARKO BM-MSCs have higher MMP-9 expression than WT BM-MSCs (Fig. 3B-g). Similar results were also obtained in the established MSCs cell line (C3H10T1/2 cells) manipulated with AR-siRNA (Fig. 3B-h).
To further dissect how AR regulates MMP-9 at the transcriptional level, we constructed an MMP-9 promoter (ranged from +2 to −2629) hooked with a luciferase vector to test whether AR could negatively regulate MMP-9 promoter transactivation activity, and found that AR could suppress MMP-9 expression in promoter regulation (Supporting Fig. 7A-C). We also applied the zymography assay to detect MMP-9 activity, and found higher MMP-9 proteolytic activity in ARKO BM-MSCs, compared with WT BM-MSCs (Fig. 3B-i). This was also confirmed in studies using hMSCs manipulated with AR-siRNA (Supporting Fig. 6C,F).
To test whether ARKO-mediated enhanced migration is MMP-9 dependent, we pretreated ARKO BM-MSCs with an MMP-9 inhibitor and performed the migration assay, and results showed that the addition of the MMP-9 inhibitor indeed masked ARKO-mediated enhanced migration ability (Fig. 3B-j), suggesting that AR needs to go through MMP-9 to exert its influence on BM-MSC migration.
Together, results from three different types of assays all proved that MMP-9 is a critical molecule to mediate the enhanced migration ability of ARKO BM-MSCs.
Finally, we confirmed the above-described findings showing KO of AR in BM-MSCs increased self-renewal potential and migration capacity in CCl4-induced liver cirrhotic mice. Consistently, ARKO BM-MSCs-transplanted liver showed higher Ki67/GFP double-positive stained cells (representing proliferating transplanted BM-MSCs) than WT BM-MSCs (Fig. 3C-k-m).
Enhanced Self-Renewal and Migration Potentials of BM-MSCs Leads to Better Anti-fibrotic and Anti-inflammatory Actions.
To correlate the increased self-renewal and migration potentials of ARKO BM-MSCs improvement in anti-fibrosis and anti-inflammatory actions, we used conditioned medium (CM) of BM-MSCs to test their effects on macrophage migration and HSCs proliferation. Results showed that BM-MSCs-inhibited macrophage migration (anti-inflammatory effects) and HSCs proliferation (anti-fibrotic actions) were BM-MSCs-number dependent (Fig. 3D,E), suggesting that KO of AR-increased self-renewal and migration of BM-MSCs resulted in more BM-MSCs to exert better anti-inflammation and anti-fibrotic actions. Together, results (from Fig. 3A-E) concluded that KO of AR in BM-MSCs led to increased self-renewal and migration potentials of BM-MSCs and these resulted in better transplantation therapeutic efficacy to treat liver cirrhosis by exerting better anti-fibrotic and anti-inflammatory effects. These phenotypes were involved in the modulation of EGF-Erk1/2/Akt signals, as well as MMP-9 signals.
IL1Ra, Secreted From BM-MSCs, Is the Key Molecule That Mediates ARKO BM-MSCs-Exerted Anti-inflammatory and Anti-fibrotic Effects.
All above-described results demonstrated that higher numbers of BM-MSCs migrating into the cirrhotic liver led to better transplantation therapeutic efficacy with higher anti-inflammatory and anti-fibrotic effects (Fig. 3D,E). We were interested to know whether there are any secreted paracrine factors influenced by knockout of AR in BM-MSCs to contribute to anti-inflammatory and -fibrotic actions. To test this possibility, we investigated expressions of HGF, VEGFB, VEGFC, TIMP-2, bone morphogenetic protein, activin membrane-bound inhibitor (BAMBI), and IL1Ra in ARKO and WT BM-MSCs, and found that ARKO BM-MSCs expressed much higher levels of IL1Ra (but not HGF, VEGFB, VEGFC, TIMP-2, and BAMBI) than WT BM-MSCs (Fig. 4A-a and Supporting Fig. 8A-E). Similar results were observed in hMSCs manipulated with AR-siRNA (Supporting Fig. 6D,F). Molecular mechanism dissection revealed that AR acts on the promoter region to inhibit IL1Ra transcription (Supporting Fig. 9A-C).
Clinical evidence has indicated that IL-1β, the ligand of IL-1 receptor, is elevated in patients with liver cirrhosis and that this elevation is correlated with increased circulating monocytes.19, 20 In addition, IL-1 signals have been suggested to play critical roles in HSCs activation and proliferation and are linked to macrophage infiltration.30, 31 We demonstrated that the addition of one of the most powerful inflammatory stimuli, lipopolysaccharide (LPS), and IL-1β both led to increased macrophage migration into HSCs. We also observed that IL-1β stimulated HSCs growth. When we pretreated HSCs with WT or ARKO BM-MSCs CM, the increased macrophage migration and HSCs growth were suppressed probably the result of the effects of secreted IL1Ra (macrophage migration result is shown in Fig. 4A-b,c and Supporting Fig. 10A; HSCs growth result is shown in Fig. 4A-d and Supporting Fig. 10B,C). ARKO BM-MSC CM showed better suppression than WT BM-MSC CM, suggesting that higher levels of IL1Ra molecule were secreted in ARKO BM-MSCs.
Because macrophages have been shown to be one source of the IL-β in CLDs32 and macrophage CM could enhance HSCs activation and growth,33 we then tested BM-MSCs CM effect on macrophage-induced HSCs growth and activation (Fig. 4B-e), and found that ARKO BM-MSCs showed better suppression than WT BM-MSCs on HSCs growth (Figure 4B-f), indicating that ARKO BM-MSCs could suppress macrophage-induced HSCs growth more significantly through secreting more IL1Ra to block IL1R signaling. Importantly, the addition of the IL1Ra neutralizing Ab did abolish the inhibitory effect significantly (Fig. 4B-g,h), confirming that the effect of ARKO BM-MSCs on anti-inflammatory and -fibrotic actions might need to go through the IL1Ra molecule.
To further explore whether BM-MSCs-secreted IL1Ra is the major source to mediate anti-inflammatory and -fibrotic effects, IL-1β was used to stimulate macrophage (chemoattractant) protein expression and WT and ARKO BM-MSCs CM were used to test whether they could block this enhancement. Because IL-1β showed induction in MCP-1 (chemokine [C-C motif] ligand 2; CCL2), chemokine (C-X-C motif) ligand (CXCL)1, and CXCL2, consistently in two different types of HSCs (Supporting Fig. 11), expressions of CCL2, CXCL1, and CXCL2 were further examined upon BM-MSC CM treatment. Results showed that ARKO BM-MSC CM had higher suppression on CCL2, CXCL1, and CXCL2 expression than did WT BM-MSCs CM (Fig. 4C-i-k). Pretreating BM-MSCs CM with IL1Ra neutralization Ab significantly masked BM-MSCs CM inhibition on CCL2, CXCL1, and CXCL2, suggesting that BM-MSCs exerted anti-inflammatory actions through IL1Ra inhibiting IL1 signaling to abolish the elevation of CCL2, CXCL1, and CXCL2 blocking macrophage infiltration (Fig. 4D-l-n).
Together, results (from Fig. 4) concluded that KO of AR in BM-MSCs led to more secreted IL1Ra that resulted in suppression of macrophage infiltration (anti-inflammation) and HSCs activation (anti-fibrosis) and then yielded better transplantation therapeutic efficacy to treat liver cirrhosis.
Targeting AR in BM-MSCs With AR-siRNA or ASC-J9® Leads to Better Self-Renewal and Migration Potential.
To apply these findings in clinical application by targeting AR in BM-MSCs (mimicking genetic ARKO BM-MSC effects in treating liver cirrhosis), we applied the currently available agents, such as AR-siRNA and ASC-J9®, that could degrade AR in selective cells with little side effects.34 We found ARKO with lentiviral AR-siRNA infection in primary WT BM-MSCs (or C3H10T1/2 and D1 cells) led to increased cell migration and proliferation (Fig. 5A-a-c,B). Similar results were also obtained when we replaced AR-siRNA with ASC-J9®. Results showed that ASC-J9® treatment in WT BM-MSCs caused elevated migration into regular media or to hepatocytes (Fig. 5C-d-f). ASC-J9® was also applied to WT BM-MSCs to determine its effect on WT BM-MSC self-renewal and proliferation potential, and results from the bromodeoxyuridine (BrdU) assay proved that ASC-J9® treatment led to enhanced self-renewal and proliferation in WT BM-MSCs (Fig. 5D). Zymographic analysis also showed that AR-siRNA or ASC-J9® treatment increased MMP-9 activity (Fig. 5E,F). Together, results (from Fig. 5A-F) conclude that targeting AR in BM-MSCs with either AR-siRNA or ASC-J9® yielded similar effects, when compared with BM-MSC effects isolated from ARKO mice, showing better anti-inflammation and anti-fibrosis effects.
New Therapy of Targeting AR in BM-MSCs Using AR-siRNA and/or ASC-J9® to Treat Liver Cirrhosis.
With consistent in vitro results obtained (Fig. 5), it was essential to test whether concordant outcomes could be reached in the in vivo mouse liver cirrhosis model. As expected, we found that lentiviral AR-siRNA infected BM-MSCs have better transplantation therapeutic effects in treating liver cirrhotic mice induced with CCl4 or TAA than scramble control BM-MSCs, when demonstrated using collagen deposition staining and expressions of fibronectin and α-SMA (Fig. 6A-a-c and Supporting Fig. 12A,B). This conclusion was further supported in liver functional assays in CCl4-induced liver cirrhosis mice (Fig. 6B-d-f). Similar results were also obtained in the TAA-induced liver cirrhosis mouse model (Supporting Fig. 12C).
Consistent therapeutic outcomes in cirrhotic liver mice were obtained from ASC-J9®-treated WT BM-MSCs. Expression of fibrosis markers confirmed that WT BM-MSCs treated with the ASC-J9® suppressed liver cirrhosis better than vehicle-treated WT BM-MSCs (Fig. 6C-g-i and Supporting Fig. 12D,E). Liver functional assays showed similar results in the CCl4 (Fig. 6D-j-l) and TAA models (Supporting Fig. 12F).
Collectively, targeting AR in BM-MSCs with either AR-siRNA or ASC-J9® may represent a new potential therapeutic approach to battle liver cirrhosis.
Androgen/AR signaling, as a critical determining factor in sexual phenotypes and maturation, is well documented. It is generally agreed that androgen/AR signaling stimulates differentiation of prostate, bone, and muscle.35-37 During the embryonic stem cell differentiation process, AR amount gradually increases.38 In natural development, androgen/AR signaling seems to enhance the differentiation process, except for adipocyte differentiation. This observation triggered us to think about whether this fact exists in disease treatment when BM-MSCs transplantation therapy is performed. Indeed, it has been suggested that gender difference does exist using BM-MSCs transplantation therapy in heart diseases.13
Although gender differences have been suggested in BM-MSC transplantation therapy with the observation that female BM-MSCs have better therapeutic potential than male BM-MSCs,13 effects of AR, a key factor in male sexual phenotype, in BM-MSC therapeutic application remain unclear. In this study, we used a genetic deletion approach to prove that deletion of AR in BM-MSCs increases their self-renewal potential (consistent results were observed in female ARKO mice; data not shown), migration capacity, anti-inflammatory actions, and anti-fibrotic effects to better treat liver cirrhosis. Because loss of BM-MSCs after transplantation cannot be prevented as a result of microischemia, our finding that knockout of AR in BM-MSCs increases functions and numbers of BM-MSCs after transplantation can provide the future basis for improving BM-MSC transplantation in liver cirrhosis as well as other diseases that have currently adopted BM-MSC transplantation therapy in clinical trials.
In therapeutic approaches, we also clearly demonstrated that targeting AR using either AR-siRNA or ASC-J9® in BM-MSCs both led to better therapeutic outcomes in treating liver cirrhosis, suggesting that targeting BM-MSC AR could exert better BM-MSC transplantation therapeutic efficacy in treating cirrhotic liver. The advantage of targeting AR with ASC-J9® is that we can handle AR degradation in in vitro cultured BM-MSCs before transplanting into liver, and several early in vivo mice studies all proved that ASC-J9® has little side effects and mice have normal sexual activity with healthy serum testosterone level.
Our mechanism studies pointed out that targeting AR in BM-MSCs leads to better therapeutic efficiency to treat liver cirrhosis through four major pathways (as illustrated in Fig. 7). Targeting AR in BM-MSCs could (1) increase EGFR to enhance their self-renewal potential (Fig. 3A), (2) elevate the MMP-9 to stimulate their migration (Fig. 3B), and (3) promote IL1Ra production to inhibit macrophage infiltration and HSCs proliferation (Fig. 4A,B). Increased IL1Ra production, in turn, suppress IL-1/IL1R signaling involved in macrophage infiltration and HSCs growth (Fig. 4). The combinational effects from those four pathways all led to improve therapeutic potential in liver cirrhosis.
In summary, we describe a process by which targeting AR, a key factor in male sexual phenotype, in BM-MSCs improves transplantation therapeutic efficacy for treating liver fibrosis. This finding might also be helpful in other diseases that have recently adopted BM-MSCs transplantation therapy in clinical trials.40
The authors thank Karen Wolf (University of Rochester Medical Center, Rochester, NY) for help in editing the manuscript for this article. The authors also thank Dr. Haiyan Pang's (University of Rochester Medical Center) help in BM-MSCs transplantation.