Biliary fibrosis drives liver repopulation and phenotype transition of transplanted hepatocytes

Abstract

Background & aims: Current research focuses on developing alternative strategies to restore decreased liver mass prior to the onset of end-stage liver disease. Cell engraftment/repopulation requires regeneration in normal liver, but we have shown that severe liver injury stimulates repopulation without partial hepatectomy (PH). We have now investigated whether a less severe injury, secondary biliary fibrosis, would drive engraftment/repopulation of ectopically transplanted mature hepatocytes.

Methods: Ductular proliferation and progressive fibrosis in dipeptidyl-peptidase IV (DPPIV)(-) F344 rats was induced by common bile duct ligation (BDL). Purified DPPIV(+)/green fluorescent protein (GFP)(+) hepatocytes were infused without PH into the spleen of BDL rats and compared to rats without BDL.

Results: Within one week, transplanted hepatocytes were detected in hepatic portal areas and at the periphery of expanding portal regions. DPPIV(+)/GFP(+) repopulating cell clusters of different sizes were observed in BDL rats but not untreated normal recipients. Surprisingly, some engrafted hepatocytes formed CK-19/claudin-7 expressing epithelial cells resembling cholangiocytes within repopulating clusters. In addition, substantial numbers of hepatocytes engrafted at the intrasplenic injection site assembled into multicellular groups. These also showed biliary "transdifferentiation" in the majority of intrasplenic injection sites of rats that received BDL but not in untreated recipients. PCR array analysis showed upregulation of osteopontin (SPP1). Cell culture studies demonstrated increased Itgβ4, HNF1β, HNF6, Sox-9, and CK-19 mRNA expression in hepatocytes incubated with osteopontin, suggesting that this secreted protein promotes dedifferentiation of hepatocytes.

Conclusions: Our studies show that biliary fibrosis stimulates liver repopulation by ectopically transplanted hepatocytes and also stimulates hepatocyte transition towards a biliary epithelial phenotype.

Keywords: Bile duct ligation; Cell transplantation; Hepatocyte dedifferentiation; Hepatocyte transdifferentiation; Osteopontin.

Fig. 1. Liver fibrosis in bile duct-ligated rats

Biliary fibrosis was induced by BDL in DPPIV⁻ F344 rats. Pathologic changes at 2/4 weeks after BDL were determined using histochemical and immunohistochemical techniques (A,B). Original magnification, ×50. (C) RNA extracts from laser-captured liver regions (fibrotic septa, surrounding parenchyma) at 1 month after BDL were analyzed, using pooled RNA samples. LCM-derived RNA had high integrity without DNA contamination (one sample is shown). Mean ± SEM values of two replicate PCR analyses are expressed as fold differences in gene expression compared to age-matched normal control rats, set at a value of 1.

Fig. 2. Engraftment of mature GFP⁺ hepatocytes transplanted into the spleen of DPPIV⁻ F344 rats at 1 month after BDL

Immunohistochemistry for GFP showing engraftment of highly purified GFP⁺ hepatocytes (~1 × 10⁷ cells) in the injection site of the spleen (A, B) and after migration into the liver (C–F). Rats were sacrificed at 1, 3 or 7 days after cell infusion. Original magnification, ×200 (A–E), ×400 (F).

Fig. 3. Tissue replacement in the recipient liver by donor hepatocytes

Highly purified DPPIV⁺ hepatocytes (~1×10⁷) were transplanted into the spleen of DPPIV⁻ F344 rats w/o (A,B) or with BDL (C–F). After migration and engraftment, only single DPPIV⁺ cells or very small cell groups were found in normal recipient livers (see arrows in A,B). In contrast, repopulation of transplanted hepatocytes was observed in host livers of BDL recipient rats. Donor-derived hepatocytes incorporated in the host parenchyma and generated DPPIV⁺ cell clusters (C–F). Arrowheads highlight donor-derived bile ducts. Examples at 2 months after transplantation are shown (A–F). Original magnification, ×50 (A,C,E), ×100 (B,D,F).

Fig. 4. Evidence of biliary epithelial differentiation of transplanted hepatocytes

DPPIV⁺ hepatocytes were transplanted into the spleen of DPPIV⁻ F344 rats w/o (A) or with BDL (B–H). DPPIV⁺ cells are shown in the injection site of recipient spleens of normal (A) and BDL rats (B). Transplanted hepatocytes expressed CK-19 (C,D). DPPIV expression pattern of engrafted hepatocytes in livers of BDL rats (E–G). Note the presence of hepatocyte-like cells expressing DPPIV in the cytoplasm (see arrow heads; F). DPPIV⁺ hepatocytes co-expressed claudin-7 (H). Panels show transplanted cells at 1 month (A–F) and 2 months (G,H) after infusion. Original magnification, ×100 (A,B), ×400 (C–H).

Fig. 5. γ–Glutamyl transpeptidase (GGT) expression in donor-derived cells in recipient rats

DPPIV⁺ hepatocytes were transplanted into the spleen of DPPIV⁻ F344 rats with BDL. The panels compare DPPIV and GGT staining in serial sections of recipient spleen at 1 month and liver at 2 months after cell infusion. Note the presence of bile duct-like cells expressing DPPIV and GGT (see arrow heads). Host DPPIV⁻ hepatocytes can also express GGT with a pattern of weak canalicular staining (F), a phenomenon observed after hepatocyte transplantation (see ref. 15). Original magnification, ×200 (A–D), ×400 (E,F).

Fig. 6. Timecource of phenotype conversion of donor-derived hepatocytes in recipient rats

DPPIV⁺ hepatocytes were intrasplenically transplanted into DPPIV⁻ F344 rats with BDL. (A–D) Simultaneous immunohistochemistry for DPPIV and CK-7. Note the presence of DPPIV⁺ donor cells co-expressing CK-7 (see arrow heads) in close proximity to expanding host bile ducts. (E–H) Double-immunohistochemistry for DPPIV/Sox-9. Original magnification, ×400 (A,C), ×200 (B,D), ×1000 (E,F), ×640 (G,H).

Fig. 7. Direct phenotype transition of transplanted hepatocytes into bile duct-like cells in recipient spleen

(A) DPPIV staining compares morphology of transplanted hepatocytes engrafted at the splenic injection. (B) RNA extracts derived from the injection site at 1 day and 1 month after hepatocyte infusion into BDL rats (n=3/3) were pooled and analyzed by qRT-PCR. Mean ± SEM values of two replicate analyses for selected genes are shown. (C–F) At 1 month, mesenchymal markers were apparent in the stroma but not in donor-derived cells. Panels in E and F compare stainings in consecutive sections. Original magnification, ×200 (A;D,left panel), ×400 (C, left;D,right;E;F), ×640 (C; right).

Fig. 8. Detection of osteopontin (SPP1) in fibrotic liver and its effect on hepatocytes

(A) RNA extracts from laser-captured liver regions (Sep, fibrotic septa; Par, surrounding parenchyma) at 1 month after BDL compared to normal livers (NL) were analyzed, using pooled RNA samples (3 rats, each group). One representative analysis of 2 independent RT-PCR analyses is shown. (B) Immunohistochemistry for osteopontin 4 weeks after BDL. Image contains 36 merged adjacent microscopic fields (×200). Normal liver is shown in Supplemental Figure 3. (C,D) Hepatocytes were cultured in triplicates with osteopontin or bile acids for 4 days (n=2). Two independent RT-PCR analyses were performed for each experiment. Mean ± SEM values are expressed as fold changes of RNA expression with respect to cells cultured without osteopontin/bile acids. (*P < 0.05; Student’s t test)