Bone Marrow as a Potential Source of Hepatic Oval Cells

See allHide authors and affiliations

Science  14 May 1999:
Vol. 284, Issue 5417, pp. 1168-1170
DOI: 10.1126/science.284.5417.1168


Bone marrow stem cells develop into hematopoietic and mesenchymal lineages but have not been known to participate in production of hepatocytes, biliary cells, or oval cells during liver regeneration. Cross-sex or cross-strain bone marrow and whole liver transplantation were used to trace the origin of the repopulating liver cells. Transplanted rats were treated with 2-acetylaminofluorene, to block hepatocyte proliferation, and then hepatic injury, to induce oval cell proliferation. Markers for Y chromosome, dipeptidyl peptidase IV enzyme, and L21-6 antigen were used to identify liver cells of bone marrow origin. From these cells, a proportion of the regenerated hepatic cells were shown to be donor-derived. Thus, a stem cell associated with the bone marrow has epithelial cell lineage capability.

Hepatic oval cells proliferate under certain conditions, mainly when hepatocytes are prevented from proliferating in response to liver damage, and may be stem cells for hepatocytes and bile duct cells or the intermediate progeny of a hepatic stem cell. Oval cells may originate either from cells present in the canals of Herring (1) or from blastlike cells located next to bile ducts (2). Oval cells have been shown to proliferate in certain pathological conditions, in which hepatocyte proliferation is inhibited before severe hepatic injury. In experimental models, hepatocyte proliferation is suppressed by exposure of the animal to 2-acetylaminofluorene (2-AAF), and hepatic injury can be induced by partial hepatectomy or by administration of carbon tetrachloride (CCl4) (3,4). Oval cells express CD34, Thy-1, and c-kit mRNAs and proteins (5–7) and flt-3 receptor mRNA (8), all of which are also found in hematopoietic stem cells (HSC).

We tested the hypothesis that oval cells and other liver cells may arise from a cell population originating in, or associated with, the bone marrow (BM). This hypothesis was tested by three approaches: (i) bone marrow transplantation (BMTx) from male rats into lethally irradiated syngeneic females and detection of donor cells in the recipients by means of DNA probes to the Y chromosome sryregion; (ii) BMTx from dipeptidyl peptidase IV–positive (DPPIV+) male rats into DPPIV syngeneic females and detection of DPPIV-expressing cells in the recipient animals; and (iii) whole liver transplantation (WLTx) with Lewis rats that express the L21-6 antigen as recipients and Brown-Norway rats that do not express this antigen as allogeneic donors (9) to confirm that an extrahepatic source (L21-6+ cells) could repopulate the transplanted (L21-6 cells) liver. In conjunction with these approaches, the 2-AAF–CCl4 protocol (4, 5) was used to induce oval cell activation and proliferation. In situ hybridization, polymerase chain reaction (PCR), and immunocytochemistry were used to distinguish donor cells from recipient cells.

Female rats were lethally irradiated and rescued with a BM transplant from a male animal. Nucleated blood cells of the transplanted animals were tested by PCR to establish that the BMTx was successful (10). Engrafted females were then placed on the oval cell protocol (4, 5). Both Thy-1+and Thy-1 cell populations of nonparenchymal cells (NPC) from days 9 and 13 after hepatic injury were positive for the Y chromosome PCR product (Fig. 1). The strong signal from the Thy-1 fraction was probably due to donor hematopoietic cells that are in the NPC population and Thy-1 but are positive for the Y chromosome. The day 9 hepatocyte fraction had no Y chromosome signal, but by day 13 a signal from the hepatocyte fraction was detected. At this time, the oval cells are beginning to differentiate into hepatocytes (1). If all oval cells that differentiate into hepatocytes were derived from the liver, then one would expect that none of the hepatocytes tested would be positive for the Y chromosome. The finding that some hepatocytes were Y chromosome–positive suggests that they were derived from the BM donor cells. The combined data suggest that at day 9 the oval cells (Thy-1+) (5) in the recipient female are derived from the donor male and that they continue to differentiate into mature hepatocytes by day 13.

Figure 1

PCR analyses of DNA from female rats transplanted with BM from male donor rats. Primers for thesry gene were used (10). sryexpression is seen in hepatocytes at day 13 after the 2-AAF–CCl4 protocol from donor-derived cells. No expression is seen in females. The data presented are representative of four independent experiments. MW STD, molecular weight standard; D-9, day 9; D-13, day 13. β-actin was used as control.

To confirm the PCR results, we performed in situ hybridization for the Y chromosome sry gene on frozen sections (11). Hepatocytes carrying a positive reaction product (blue staining) in the nuclei were seen in untreated control male rats (Fig. 2A). In agreement with the results obtained by PCR analyses of the isolated hepatocyte fraction, cells with positive signal (Y chromosome) were detected in oval cells of females subjected to BMTx and 2-AAF–CCl4 at both days 9 (Fig. 2B) and 13 and hepatocytes at day 13 (Fig. 2C). No signal was detected in the liver of an untreated control female (Fig. 2D). In 100 random fields, about 0.14% of the total hepatocytes were positive for the Y chromosome. Therefore, about 9.9 × 105hepatocytes originated from transplanted BM cells on day 13 after hepatic injury. In addition, the proportion of Thy-1+ oval cells expressing the Y marker was about 0.1% (12).

Figure 2

In situ hybridizations of the Y chromosomesry gene on frozen liver sections. Arrows indicate positive signals in the nuclei of cells. (A) Untreated normal male rat control. (B and C) Female rats treated with BMTx and the 2-AFF–CCl4 protocol and killed at day 9 (B) and day 13 (C) after hepatic injury. (D) Untreated female (negative control). Scale bar, 15 μm.

In an independent approach, BM cells from DPPIV+ F-344 male rats were injected into lethally irradiated DPPIV F-344 females, and the recipients were treated with the 2-AAF–CCl4 protocol. This constituted a system in which the presence of cells originating from donor cells in the recipient liver could be easily detected, by revealing cytochemically the activity of the enzyme DPPIV (Fig. 3). A diffuse red to red-orange staining of the bile canalicular site between hepatocytes (13) was observed in the DPPIV+F-344 male rats (Fig. 3A). The control untreated DPPIVfemales showed no staining (Fig. 3B). DPPIV expression was observed in a number of bile canalicular sites between hepatocytes from several different transplanted animals (Fig. 3, C to F). DPPIV expression was also observed on oval cells and transitional (small hepatocyte) cells in the liver from these rats (Fig. 3, D and E). Out of 250 random fields, about 0.16% of the hepatocytes were positive for DPPIV. Therefore, about 1.0 × 106 hepatocytes originated from transplanted BM cells by day 13 after hepatic injury (12). In control DDPIV females, no DPPIV+ hepatocytes were observed after BMTx but before 2-AAF–CCl4 (Fig. 3G). To demonstrate that the donor-derived hepatocytes were mature, we performed three-color immunofluorescence. DPPIV+ hepatocytes also stained positive for the mature hepatocyte-specific markers, H4 and C-CAM (Fig. 3H) (14). This result shows definitively that the cells are mature hepatocytes (H4- and C-CAM–positive) that have been integrated into the parenchyma and are carrying the transplant marker (DPPIV+).

Figure 3

DPPIV activity in frozen liver sections. (A) Untreated DPPIV+ rat (positive control). DPPIV staining is evident (red-orange color). (B) Untreated DPPIV rat showing absence of DPPIV (negative control). (C to F) Four different BMTx. Male DPPIV+ (donor) or female DPPIV (recipient) rats were exposed to the 2-AAF–CCl4 protocol for oval cell induction and killed at day 11 or 13 after hepatic injury. A positive reaction is evident between hepatocytes from all four animals and on a few oval or transitional cells (E). DPPIV staining appears as a line (open arrowheads) or dot (closed arrowheads) depending on the plane of the section through the bile canalicular region. In all cases, there are clusters of cells expressing DPPIV. The asterisks indicate donor origin hepatocytes. (G) Control DPPIV female after BMTx but before 2-AAF–CCL4 protocol. No DPPIV+ hepatocytes were observed; an occasional DPPIV+ hematopoietic cell was seen (arrowhead). (H) Immunofluorescent staining of mature hepatocytes, after transplantation and 2-AAF–CCL4protocol. Three markers were used: DPPIV–Texas Red, H4-AMCA (stains the cytoplasm of mature hepatocytes), and C-CAM-FITC (a hepatocyte-specific cell adhesion molecule). In the bile canalicular region, the blue cells (H4 positive), which are also are positive for both DPPIV and C-CAM, appear yellow. Scale bars, 60 μm (F) and 30 μm (H).

Our final approach to examine extrahepatic cells repopulating the liver was WLTx. Lewis rats that express the major histocompatibility complex class II L21-6 isozyme were recipients of liver from Brown-Norway rats that do not express L21-6 (15). A monoclonal antibody to L21-6 differentiated donor from recipient cells. Oval cells that originated from an extrahepatic source would be L21-6–positive, whereas those originating in situ would be negative. Brown-Norway rat livers transplanted into Lewis rats after the 2-AAF–CCl4 protocol show a widespread staining of L21-6 in the transplanted Brown-Norway livers, a result of the influx of Lewis immune cells reacting to the allogeneic Brown-Norway liver. Ductal structures also contained L21-6–positive cells in a pattern often seen in the organization and differentiation of actively proliferating oval cells (16, 17), indicating that some oval cells were derived from an extrahepatic source and others, L21-6–negative, were derived in situ from the donor liver (18).

Under specific physio-pathological conditions, oval cells can differentiate into the two types of epithelial cells present in the liver: ductular cells and hepatocytes (1, 16,17). Oval cells may have a role in the genesis of chemically induced hepatocellular carcinomas, but they also have a potential use in the therapy of acute liver failure syndromes. Oval cell precursors are thought to be located either in the canals of Herring or next to the bile ducts (1, 2). Ductal epithelium is required for oval cell proliferation (19), indicating that either it is the source of the precursors or it acts in a supportive or inductive role. Oval cells have some phenotypic traits that are typical of BM stem cells (5–8). The liver harbors many hematopoietic cell types, including HSC (20), and if oval cells could originate from either BM-derived cells or self-replication, one would expect to see a subpopulation of negative oval cells from resident cells in the liver of Brown-Norway rats, as well as a population of positive cells originating from cells that migrated into the liver from the L21-6–positive host. Our Brown-Norway into Lewis liver transplant data support this hypothesis (18).

Our data lend credence to the possibility that a cell associated with the BM may act, under certain physio-pathological conditions, as the progenitor of several types of liver cells. The limits and biological importance of these BM-derived cells need to be assessed; the BM-derived cells add to the growing body of evidence suggesting that cells in the adult organism have a remarkable degree of plasticity (21–23).

  • Present address: Department of Oncology, King Faisal Specialist Hospital and Research Centre, Post Office Box 3354, Riyadh, Kingdom of Saudi Arabia.


View Abstract

Stay Connected to Science

Navigate This Article