Liver regeneration therapy through the hepatic artery-infusion of cultured bone marrow cells in a canine liver fibrosis model

Background We previously reported regenerative therapies for decompensated cirrhosis based on peripheral venous drip infusion using non-cultured whole bone marrow (BM) cells, or the less invasive cultured BM-derived mesenchymal stem cells (BMSCs). Here, we assessed the efficacy and safety of hepatic arterial infusion using cultured autologous BMSCs, comparing it with peripheral infusion, using our established canine liver fibrosis model. Methods Canine BM cells were harvested and cultured, and the resultant BMSCs were returned to carbon tetrachloride (CCl4)-induced liver cirrhosis model canines via either a peripheral vein (Vein group) or hepatic artery (Artery group). A variety of assays were performed before and 4, 8, and 12 weeks after BMSC infusion, and liver fibrosis and indocyanine green (ICG) half-life (t1/2) were compared to those in a control group that received CCl4 but not BMSCs. The safety of this approach was evaluated by contrast-enhanced computed tomography (CT) and serial blood examinations after infusion. Results Four weeks after infusing BMSCs, a significant improvement was observed in the Vein group (n = 8) compared to outcome in the Control group (n = 10), along with a decrease in ICG t1/2. In the Artery group (n = 4), ICG t1/2 was significantly shorter than that in the Vein group at 8 weeks (Δt1/2: −3.8 ± 1.7 min vs. +0.4 ± 2.4 min; p < 0.01) and 12 weeks (Δt1/2: −4.2 ± 1.7 min vs. +0.4 ± 2.7 min; p < 0.01) after BMSC administration. Post-infusion contrast-enhanced CT showed no liver infarction, and blood tests showed no elevations in either serum lactate dehydrogenase concentrations or hypercoagulability. Conclusions We confirmed the efficacy and safety of the hepatic arterial infusion of cultured autologous BMSCs using a canine model, thereby providing non-clinical proof-of-concept.

Introduction Liver cirrhosis represents the advanced stage of liver fibrosis, presenting as regenerative nodules surrounded by fibrotic bands that develop in response to chronic liver injury [1][2][3]. Currently, the most effective therapy for advanced cirrhosis is liver transplantation. However, the procedure has several limitations including the lack of donors, surgical complications, immunological suppression, and high medical costs [4]. Therefore, novel alternative therapies for decompensated liver cirrhosis are urgently needed.
In recent years, several studies using animal models of liver diseases have demonstrated that bone marrow (BM) cell transplantation might accelerate liver regeneration and reduce hepatic fibrosis [5][6][7][8].
Accordingly, we developed autologous BM cell infusion (ABMi) therapy for liver cirrhosis and reported its efficacy and safety for liver cirrhosis [9,10]. However, since ABMi therapy requires BM aspiration under general anesthesia, some patients are excluded due to poor liver or cardiopulmonary functions. In an effort to expand the applicability of ABMi therapy, we developed a less invasive method for liver regeneration therapy using autologous BM-derived mesenchymal stem cells (BMSCs) cultured from small amounts of BM fluid aspirated under local anesthesia. However, before human clinical trials can be considered, the safety and efficacy of cultured autologous BMSC infusion must be confirmed in medium-to-large animals. To this end, we previously established a useful canine liver fibrosis model generated through the repeated administration of carbon tetrachloride (CCl 4 ) through a catheter for 10 weeks. Using this model, we found that peripheral venous infusion using cultured autologous BMSCs could improve liver fibrosis without adverse effects, suggesting that this could represent a less invasive therapeutic option [11].
Here, to develop liver regeneration therapy with enhanced therapeutic effects, we evaluated the efficacy and safety of the hepatic arterial infusion of cultured autologous BMSCs, comparing it with peripheral infusion, using our established canine model of liver fibrosis.

Animals and ethics
Twenty-two beagles (1-2 years of age; 10 males and 12 females) were used in this study. These animals included six canines in the control group and four in the peripheral venous infusion group, from which 0-and 4-week data were used from our previous study [11]. These canines were housed in our animal facility and treated in accordance with the university's Animal Care Guidelines. In particular, they were housed in individual metal cages (100 cm in width and 120 cm in height) and given standard solid food (Oriental Yeast Co., Tokyo, Japan), according to dietary recommendations that were based on body weight, age, and activity level of each canine. They had free access to water at all times and were housed in a climate-controlled building at 22-23˚C, with a 16-h light/8-h dark cycle. They were taken for walks daily in the evening (15:00 to 17:00) and were inspected at least three times per day to monitor their health and well-being. Facility staff members including our group checked several health conditions as follows: appetite, thirst, defecation, urination, behavior and lethargy, etc. After surgery, we also checked these health conditions at least three times per day. This study was approved by the Institutional Animal Care and Use Committee of Yamaguchi University, "Ube" area. (Approval number: 21-033).

Catheter implantation
All catheter implantations were performed as previously described [11]. Briefly, this was performed under general anesthesia for all canines. The stomach was directly punctured with an 18-gauge Teflon IV catheter (6-French P-U catheter; Toray Medical Co. Ltd., Tokyo, Japan). An infusion port (P-U Celsite port; Toray) was then placed in a subcutaneous pocket created on the back. Carbon tetrachloride (CCl 4 , Wako, Osaka, Japan) was administered from the port, which essentially represented oral administration. Intramuscular buprenorphine at 10 μg/kg body weight (BW) and subcutaneous cefovecin (Convenia; Zoetis Japan, Tokyo, Japan) at 8 mg/kg BW were given at the end of the procedure for post-operative analgesia and to prevent infections, respectively. As stated, we checked several health conditions after surgery at least three times per day.

Aspiration of bone marrow fluid, culture, and infusion of BMSCs
After catheter implantation, a 16-gauge biopsy needle (Angiotech, Gainesville, FL) was used to puncture the proximal humerus of each canine and BM was aspirated using a 5-mL syringe filled with 1 mL of heparin. To culture BMSCs, the aspirated BM fluid was seeded in T-75 flasks (Life Technologies Corp., Grand Island, NY) and cultured with 10% fetal bovine serum (Life Technologies) and 100 μg/mL gentamicin (Life Technologies) in Dulbecco's modified Eagle Medium (DMEM, Life Technologies) in a 5% CO 2 incubator at 37˚C. After 2 days of incubation, non-adherent cells were removed during medium replacement. The culture medium was changed every 2 days for 2-3 weeks. Expanded BMSCs were detached using cell detachment solution and then collected.

Experimental model (induction of liver fibrosis)
CCl 4 was diluted 1:1 in corn oil before injecting it intra-gastrically via the catheter. Injections ware repeated for 10 weeks using the implanted catheter (high-dose period: 1.0 mL/kg BW once per week and 0.5 mL/kg BW once per week for 6 weeks; low-dose period: 0.25 mL/kg BW twice per week for 4 weeks) to induce liver fibrosis (Fig 1). After the infusion of BMSCs, low-dose CCl 4 was further administrated for 12 weeks. CCl 4 injections were performed on the first and fourth days of each week.

Experimental groups
After 10 weeks of CCl 4 injections, the 22 canines were divided into three groups including the Control, Vein, and Artery groups. This point was defined as "0W". Subsequently, the 10 animals in the Control group received low-dose CCl 4 injection for a further 12 weeks. The eight canines in the Vein group were infused with cultured autologous BMSCs (4 × 10 5 /kg) through a peripheral vein and low-dose CCl 4 injection was continued for 12 more weeks. The four canines in the Artery group were infused with cultured autologous BMSCs (4 × 10 5 /kg) through a hepatic artery using angiography, and low-dose CCl 4 injection was continued for 12 more weeks.

Safety evaluations with an increased BMSC dose
BMSCs were increased to 1.2 × 10 6 cells/kg and infused through a hepatic artery into the canine model of liver fibrosis to test the effects of a higher dose (n = 1). The general condition and vital signs were monitored before and after BMSC infusion. Blood parameters were measured before BMSC infusion and at 1, 3, 5, and 7 days after infusion. Seven days after BMSC The canines were administered repeated carbon tetrachloride (CCl 4 ) injections through the implanted catheter for 10 weeks (high-dose period: 1.0 mL/kg body weight once per week and 0.5 mL/kg body weight once per week for 6 weeks; low-dose period: 0.25 mL/kg body weight twice per week for 4 weeks). CCl 4 was injected on the first and fourth days of each week. At 0 weeks (0W; i.e. after 10 weeks of CCl 4 injections), the canines were divided into three groups. In the control group, low-dose CCl 4 injection was continued for a further 12 weeks without bone marrowderived mesenchymal stem cell (BMSC) infusion. In the Vein group, cultured autologous canine BMSCs (4 × 10 5 /kg) were infused through a peripheral vein, and low-dose CCl 4 injections were continued for 12 more weeks. In the Artery group, cultured autologous BMSCs (4 × 10 5 /kg) were infused through a hepatic artery using angiography, and low-dose CCl 4 injections were continued for 12 more weeks. In all three groups, blood assays, indocyanine green (ICG) testing, and liver biopsy based on ultrasonography were performed at −10W, 0W, 4W, 8W, and 12W, as indicated by the arrows. infusion, after collecting blood samples, we conducted a liver biopsy to evaluate the presence of liver infarction based on histological analysis.

Laboratory tests
The blood samples collected at the stipulated time points were used to assess serum albumin, alanine aminotransferase, aspartate aminotransferase, bilirubin, prothrombin time PT, anti-thrombin3 (AT3) activity, and fibrin degradation products.

ICG tests to assess liver functions
A 22-G indwelling venous cannula was inserted into the internal jugular vein and blood was collected. Following the collection of blood for the 0 time-point sample, 0.1 mg/kg of ICG (Diagnogreen Inj.; Daiichi Sankyo Co. Ltd., Tokyo, Japan) was administered via the peripheral vein and blood was collected 5, 10, 15, 20, and 30 min later from the maintained venous route. Blood samples were then centrifuged and sera were harvested from blood samples and analyzed for ICG content using an ICG Meter (Fuchu Giken, Inc., Tokyo, Japan) The difference in the ICG half-life (t 1/2 ), or ΔICG (min), was calculated using the formula, ΔICG (min) = ICG t 1/2 (min) at 4W, 8W, or 12W -ICG t 1/2 (min) at 0W.

CE-CT imaging
CE-CT examinations were performed using an eight-detector-row CT system (ECLOS 8; Hitachi Medical Corp., Tokyo, Japan). All canines were first placed in ventral recumbency and all CT scans were performed during apnea under anesthesia. Canines were administered 2 mL/kg iopamidol with an iodine concentration of 370 mg/mL (Oiparomin 370; Fuji Pharmaceutical Co., Toyama, Japan) as intravenous contrast medium.

Liver biopsy
After the intramuscular injection of 20 μg/kg medetomidine (Dorbene; Kyoritsu Seiyaku Corp., Tokyo, Japan), the canines were placed in the left lateral decubitus position. The skin over the biopsy site was shaved and then disinfected with povidone-iodine. The local anesthetic (0.5% lidocaine; Pfizer, Tokyo, Japan) was injected into a small area of the skin and into tissues over part of the liver, and ultrasonography-guided liver biopsies were collected using a 16-gauge biopsy needle (Aragon Medical Devices, Plano, TX). After removing the biopsy needle, the puncture site was manually compressed and atipamezole hydrochloride was injected intramuscularly.
Some samples were fixed in 4% paraformaldehyde overnight and used for the evaluation of liver fibrosis by Sirius-red staining, whereas others were stored at −80˚C for real-time quantitative polymerase chain reaction (RT-PCR). These experiments including ultrasonographyguided liver biopsies were performed over time. We analyzed gene expression related to liver fibrosis by RT-PCR using ultrasonography-guided biopsied specimens; therefore, we have been taking care of these canines after this study, and we did not sacrifice them.

Sirius-red staining
Paraffin-embedded liver samples were sectioned (3-μm) and stained with Sirius-red staining. For this, paraffin sections were de-waxed, hydrated, and stained with Sirius-red solution (1.3% aqueous picric acid/0.1% Direct Red 80) for 15-20 min to detect liver fibrosis. Samples were then washed with 0.5% acetic acid for 1 min and dehydrated using 100% ethanol for 5 min. Samples were cleared in xylene for 10 min and then mounted in a resinous medium.

Histomorphometry
Histomorphometry was performed using an imaging system coupled to a fluorescence microscope (Biorevo BZ9000; Keyence, Osaka, Japan). The fibrotic area was calculated as the percent Sirius red-stained area relative to the total sample using a BZ Analyzer II (Keyence). Vessels stained with Sirius red were excluded from the calculation. We defined the percent fibrotic area as the fibrosis level. The difference in the fibrotic area, or Δfibrosis level (%), was calculated using the formula Δfibrosis level (%) = fibrosis level (%) at 4W, 8W, or 12W − fibrosis level (%) at 0W.
Next, we inserted a catheter into the left hepatic artery, right medial lobe branch, and right lateral lobe branch and administered BMSCs via each of these arteries at doses of 2 × 10 5 cells/ kg, 1 × 10 5 cells/kg, and 1 × 10 5 cells/kg, respectively. After removing the sheath, the puncture site was manually compressed. Cefazolin sodium (25 mg/kg, intravenously) and buprenorphine (20 mg/kg, intramuscularly) were administered after treatment.

Statistical analysis
Data were analyzed using a Student's t-test and paired t-test. Values of p < 0.05 were considered statistically significant. Data are presented as the mean ± standard deviation.

Biochemical results
In the Control group, albumin and AT3 gradually tended to become exacerbated. In the Vein group, although improvement was observed at 4 weeks, subsequent gradual exacerbation was observed. However, in the Artery group, this improvement at 4 and 8 weeks was maintained at 12 weeks (Table 1 and S3 Table).

Expression of liver fibrosis-related genes
The expression of COL1A2, COL3, and TIMP-1 mRNA, which encode proteins involved in hepatic fibrosis, was analyzed. Expression levels of all genes tended to decrease in the Vein

Safety evaluations
Post-infusion contrast-enhanced CT showed no sign of liver infarction and the absence of major PSS (Fig 7a). Further, we did not observe oxygen desaturation, a remarkable change in pulse rate, or a decline in general condition after BMSC infusion in any of the canines including the canine used to test an increased dose. Moreover, in the canine infused with a higher number of BMSCs (1.2 × 10 6 cells/kg), blood tests showed no elevation in either serum lactate dehydrogenase concentrations or hypercoagulability (Table 2). Moreover, arterial embolization and necrotic biliary areas were not observed based on hematoxylin and eosin staining of the liver tissue (Fig 7b).

Discussion
The present study showed the efficacy and safety of hepatic artery infusion using cultured autologous BMSCs based on a canine liver fibrosis model.
In a previous study, we placed a port and a catheter in beagles, and using this system, administered CCl 4 over a 10-week period to induce liver fibrosis. We then administered cultured autologous BMSCs via peripheral veins and confirmed the safety and efficacy of this approach for the treatment of fibrosis [11].
In an attempt to increase the efficacy of this treatment, we considered the selective administration of BMSCs into the liver via the hepatic artery to facilitate more effective liver regeneration. Hence, in this study, we administered cultured autologous BMSCs into the hepatic artery, and compared treatment efficacy and safety between the two administration routes.
Using canines, we established a technique based on abdominal angiography for the administration of BMSCs via the hepatic artery, which is used routinely in medical practice. Canine abdominal blood vessels differ from those of humans in that after an artery branches from the common hepatic artery toward the outer right liver lobe, the gastroduodenal artery branches into the artery leading to the inner right lobe and the artery leading to the left lobe. In a joint study carried out at the Veterinary Surgery Laboratory of Yamaguchi University Joint Faculty of Veterinary Medicine, the abdominal vascular course of 30 beagles was studied using 3-D CT, and in 80% of these animals, the blood vessel to the outer right lobe branched before the gastroduodenal artery [12]. Based on this vascular anatomy, which is different from that of humans, we introduced a catheter into the left hepatic artery, the right medial lobe branch, and the right lateral lobe branch, and administered BMSCs at 2 × 10 5 cells/kg 1 × 10 5 cells/kg, and 1 × 10 5 cells/kg, respectively. We also assessed the ICG t 1/2 to evaluate liver function. Boothe et al. proved the diagnostic benefits of using ICG disposition kinetics as a method to evaluate hepatic function in canines with progressive liver disease [13]. However, in cases with PSS, the results of ICG testing might exaggerate the diminished functional reserve [14,15]. Since we confirmed the absence of PSS in our experimental model by CE-CT (Fig 7A), the prolonged ICG t 1/2 observed in our canine model is a true reflection of liver function.   Liver regeneration via hepatic artery BM cell infusion At 4 weeks after the administration of BMSCs, the ICG t 1/2 increased in the Control group but decreased in the Vein and Artery groups. Moreover, whereas the ICG t 1/2 increased in the Vein group at 8 weeks post-BMSC administration, the effects of treatment were maintained in the Artery group at 8 and 12 weeks post-administration. Consequently, the effects of BMSCs injected via a peripheral vein in this canine liver fibrosis model were suggested to last for approximately 4 weeks. In contrast, with the administration of BMSCs via the hepatic artery, it is anticipated that these effects can be maintained for at least 12 weeks after administration.  Boothe et al. induced hepatic disease in canines with dimethylnitrosamine, and reported that Alb, T-bil, and prothrombin time were only slightly different between mild and severe hepatic disease groups, and that AT3 activity correlated with hepatic disease severity, similar to that observed in ICG experiments [13]. In our study, AT3 decreased in the Vein group 8 weeks after BMSC administration, similar to results obtained for the ICG experiment, although the decrease in the Artery group was maintained at 8 and 12 weeks postadministration.

Control group
Whereas the expression of mRNA involved in hepatic fibrosis tended to be lower in the Vein group than the Control group, it was significantly lower at 4, 8, and 12 weeks in the Artery group, compared to that in the Control group. Kanemoto et al. reported that the expression of these genes correlates well with the histologic degree of fibrosis in canines [16]. Hence, assuming that the reduced expression of these genes in the Artery group is consistent with the degree of fibrosis (as suggested by the results of Kanemoto et al.), it appears that a stronger anti-fibrotic effect occurred in the Artery group than in the Vein group.
The mechanisms underlying BMSC-mediated improvements in fibrosis remain to be clarified. Several studies have suggested that mesenchymal stem cells (MSCs) exert their effects on different aspects of fibrosis to facilitate this improvement. The first of these is the effect of MSCs on reducing liver cell damage. MSCs are believed to prevent liver cell damage by suppressing oxidative stress induced at the time of injury and by downregulating inflammatory cytokines such as TNF-α and IFN-γ [17,18]. Therefore, they are thought to have an antiinflammatory effect. At the same time, some reports have suggested that MSCs have immunosuppressive and immunoregulatory effects, and that they inhibit T cell proliferation during liver inflammation and promote regulatory T cell proliferation [19,20]. It has also been reported that MSCs tend to suppress inflammation through their effects on macrophages and dendritic cells, which indirectly inhibits the action of stellate cells and is involved in the inhibition of fibrosis [21]. In addition, there are many aspects related to the direct effects of MSCs that are unclear, although numerous reports have reported improved fibrosis following the administration of MSCs. Thus, based on these effects on the abatement of liver disease and the suppression of inflammation, it appears likely that MSCs exert a secondary effect by indirectly inducing endogenous mechanisms to improve fibrosis. Further clarification of such mechanisms will be an area for future work.
The optimal route of MSC delivery to the liver is an important concern. Most studies have used an intravenous route, although a large proportion of MSCs that are injected via this route become trapped in the lungs upon first passage [22][23][24][25]. Moreover, several studies have revealed that intra-arterial injection can significantly decrease the number of MSCs that are trapped within the lungs, thus increasing their uptake by other organs, especially the liver [26][27][28]. We, therefore hypothesized that the administration of MSCs via the hepatic artery, rather than via a peripheral vein, would prevent MSCs from becoming trapped in the lungs, ensuring greater accumulation in the liver, and leading to improved liver function and the suppression of liver fibrosis. However, even in our canine liver fibrosis model, the engraftment of administered cells needs to be analyzed and studied in the future. Suk et al. also reported that BMSC administration into the hepatic artery could augment improvements in fibrosis and Child-Pugh scores, an index of liver function, in the group administered cells via both routes compared to those in the single administration group [29].
BMSCs can be cryopreserved once they have been cultured, allowing for frequent administration while reducing burden to the patient; this will likely lead to a more lasting therapeutic effect. Since multiple administrations of cryopreserved autologous BMSCs will likely bring about greater therapeutic efficacy for hepatic fibrosis, it will be important to perform comparative studies to test this in the future.
In conclusion, we confirmed the efficacy and safety of the hepatic artery infusion of cultured autologous BMSCs for the treatment of liver fibrosis using a canine model, therefore providing non-clinical proof-of-concept for this approach.
Supporting information S1 Table.