Prominin-1 (PROM1), also known as CD133, is known to be expressed in hepatic progenitor cells (HPCs) and cholangiocytes of fibrotic liver. In this study, we show that PROM1 was upregulated in the plasma membranes of fibrotic hepatocytes. The hepatocellular expression of PROM1 was also demonstrated in the mice (Prom1^CreER; R26^TdTom) where cells express TdTom under the control of Prom1 promoter. To understand the role of hepatocellular PROM1 on liver fibrosis, global and liver- and cholangiocyte-specific Prom1-deficient mice were analyzed after bile duct ligation (BDL) and carbon tetrachloride (CCl₄) treatment. BDL- and CCl₄-induced liver fibrosis was aggravated with increased phosphorylation of SMAD2/3 and decreased level of SMAD7 by global or liver-specific Prom1 deficiency but not by cholangiocyte-specific Prom1 deficiency. Indeed, PROM1 prevented SMURF2-induced SMAD7 ubiquitination and degradation by interfering with the molecular association of SMAD7 with SMURF2. We also demonstrated that hepatocyte-specific ovxpression of SMAD7 ameliorated BDL-induced liver fibrosis in liver-specific Prom1-deficient mice. Thus, we concluded that PROM1 is necessary for the negative regulation of TGFb signaling during liver fibrosis.

Patient samples

Samples from human patients with mild or severe fibrosis were obtained from Dongguk University in Gyeongju, South Korea (Approval number: 110757-201909-HR-06-02). The patients had been diagnosed with liver fibrosis or cirrhosis by histological examination at hospitals in South Korea.

Animal models and experiments

Global Prom1 knockout mice (The Jackson Laboratory, Bar Harbor, ME, USA), in which the ATG start codon of the CD133 gene was replaced by a CreERT2 fusion protein and an IRES-b-galactosidase (lacZ) gene, were utilized. Prom1 KO mice were backcrossed onto a C57BL/6N background for at least five generations. To generate liver-specific Prom1 deficient mice (Prom1^f/f; Alb-Cre), Prom1^f/f mice (f/f) mice were created by ToolGen (Seoul, Korea) and crossed with Alb-Cre (Alb-Cre) mice containing a Cre recombinase driven by albumin promoter (The Jackson Laboratory, Bar Harbor, ME, USA).

We generated f/f; Krt19-Cre, and f/f mice in which Krt19-Cre (Stock No: 026925, Jackson Laboratory) mice were crossed with f/f mice. For Cre-loxP recombination, tamoxifen (T5648; Sigma) at a concentration of 20mg/ml corn oil was injected intraperitoneally at 100 mg/kg of body weight for 5 consecutive days on 6 to 8 weeks old mice. To generate Krt19^CreER; Prom1^f/f; R26^TdTom mice, Krt19^CreER (Stock No: 026925, Jackson Laboratory), R26^TdTom (Stock No: 007914, Jackson Laboratory), Prom1^f/f mice were mated with each other. For Cre-loxP recombination, tamoxifen (T5648; Sigma) at a concentration of 20mg/ml corn oil was injected intraperitoneally at 100 mg/kg of bodyweight for 5 consecutive days on 6 to 8 weeks old mice.

Generation of Prom1 cell lineage tracing mice (Prom1^CreER; R26^TdTom) were crossed with Prom1^CreER (Stock No: 017743, Jackson Laboratory) and R26^TdTom (Stock No: 007914, Jackson Laboratory) mice. To achieve Cre-loxP recombination, tamoxifen (T5648; Sigma) at a concentration of 20mg/ml corn oil was injected once intraperitoneally at 150 mg/kg of body weight after a day of bile duct ligation on 8-10 weeks old mice.

Mice, housed in plastic cages under a 12:12 light-dark photoperiod with free access to water and food, were bred, maintained, and cared for in a manner consistent with criteria outlined in the Principles of Laboratory Animal Care (NIH publication no. 85-23, revised 1985), and protocols were approved by the Institutional Animal Care and Use Committee of Korea University. Notably, Prom1 KO mice are viable and live to adulthood without exhibiting phenotypic differences from their WT littermates. All animal studies were conducted with the approval of the Korea University Institutional Animal Care and Use Committee and the Korean Animal Protection Law (KUIACUC-2018-06 and -2019-0111).

Mouse model of liver fibrosis

To generate different mouse fibrosis models, CCl₄ and BDL treatment were performed. Toxicant-induced models were generated by subcutaneous injections with either CCl₄ (1 ml/kg body weight) in corn oil (1:3 dilution), as a control, in six-week old male WT and KO C57BL/6 mice every third day for 6 weeks. The mice were killed 2 days after the final CCl₄ injection.

For the BDL model, the mice were randomly divided into two groups. Animals that underwent BDL and sham operated animals were used as healthy controls. Briefly, mice were anesthetized using isoflurane. The extra-hepatic bile duct was isolated and doubly ligated. The peritoneal cavity was filled with saline and the incisions were closed. Mouse body weight were measured daily. Seven days after the surgery, BDL and sham control animals were sacrificed.