SEARCH

SEARCH BY CITATION

Keywords:

  • CD133;
  • hepatocellular carcinoma;
  • tumorigenicity

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Recently increasing reported data have suggested that only a small subset of cancer cells possess capability to initiate malignancies including leukemia and solid tumors, which was based on investigation in these cells displaying a distinct surface marker pattern within the primary cancers. CD133 is a putative hematopoietic and neuronal stem-cell marker, which was also considered as a tumorigenic marker in brain and prostate cancer. We hypothesized that CD133 was a marker closely correlated with tumorigenicity, since it was reported that CD133 expressed in human fetal liver and repairing liver tissues, which tightly associated with hepatocarcinogenesis. Our findings showed that a small population of CD133 positive cells indeed exists in human hepatocellular carcinoma (HCC) cell lines and primary HCC tissues. From SMMC-7721 cell line, CD133+ cells isolated by MACS manifested high tumorigenecity and clonogenicity as compared with CD133 HCC cells. The implication that CD133 might be one of the markers for HCC cancer stem-like cells needed further investigation. © 2006 Wiley-Liss, Inc.

Hepatocellular carcinoma (HCC) is one of most prevalent cancers in Asia and Africa, and the incidence is rising in western world due to increased incidence of hepatitis virus C infection.1 Although advance of conventional clinical treatment for HCC has been achieved, the mortality had not been improved significantly over the past several decades. Current progress in the targeted therapy has showed promising prospect in selective killing of cancer cells2; however, what may be the most needed is a better understanding for the key subpopulation of cancer cells responsible for the tumor growth. It has been proposed that in malignancy there is a hierarchy in which only a small subset of cells drive cancer propagation and metastasis, while the rest of the bulk consists of cells with relatively weak tumorigenicity.3, 4, 5, 6 Recent data in leukemia and several solid tumors supported the existence of such a subpopulation, which was successfully isolated and manifested marked tumorigenic capacity assessed by NOD/SCID mice xenograft assay.7, 8, 9, 10, 11 These clonogenic cells, which is termed as “tumor initiating cells” or “cancer stem (like) cells,” displayed distinguishable surface marker patterns such as CD34+CD38 in leukemia, CD44+CD24low/lin− in breast cancer and CD133+ in brain and prostate tumors.9, 10, 11, 12

CD133 (AC133) is a highly conserved antigen as the human homologue of mouse Prominin-1, which was originally identified as a 5 transmembrane cell surface glycoprotein expressed in a subpopulation of the CD34+ hematopoietic stem and progenitor cells derived from human fetal liver and bone marrow.13, 14 From reports about the expression distribution of the antigens, CD133 was also detected on neuroepithelial cells, embryonic epithelia and adult immature epithelia of several normal tissues.15, 16, 17, 18 Notably, CD133+ was expressed in some types of tumor tissues, especially in some cancer stem cells within AML, brain tumor, ependymoma and prostate cancer.10, 11, 12, 19, 20 In glioblastoma, as few as 100 CD133+ cells were described to be able to produce tumors in immunodeficient mice, whereas 1 × 105 cells from the same tumor without this surface molecule failed.11 CD133 expression was observed in liver by Northern blot but not affirmed by immunostaining in human normal liver or HCC tissues14; however, Craig et al. identified regenerative CD117+/CD133+ hepatic precursor cells in fresh frozen liver samples from patient suffered from massive liver necrosis.21 Whether there is a subset of CD133+ cells playing a role in tumorigenicity in HCC needed to be further determined.

In the present study, we isolated and characterized a small population of CD133+ cells, which existed in human HCC cell line SMMC-7721, and CD133 expressions were also identified in human primary HCC tissues by immunohistochemistry. These isolated CD133+ HCC cells, though representing an extremely small subpopulation, were distinctive for their high clonogenicity in vitro and tumorigenicity in immunodeficiency mice xenograft model. Our findings might provide some insights into isolation and characterization of cancer initiating cells in HCC development.

In the present study, we attempt to examine the CD133 expression in human primary HCC. Furthermore, we try to isolate the CD133+ cells from human HCC cell lines and investigate whether the CD133+ cells possess the potential for clonogenicity and tumorigenicity.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Antibodies and other reagents

Monoclonal CD133/1 antibody (IgG1), FcR Blocking Reagent and Rat anti-mouse IgG1 MicroBeads were from Miltenyi Biotec (Auburn, CA). Goat polyclonal anti-CD133 was from Santa Cruz (Santa Cruz, CA). Monoclonal anti β-actin was from Sigma (St. Louis, MO). FITC and Alexa 594 conjugated goat anti-mouse IgGs were from Sigma-Aldrich or Molecular Probes (Eugene, OR). EnVision HRP (mouse) kits, DAB Liquid and Mayer's Hematoxylin were from DAKO (Glostrup, Denmark). Other chemicals and reagent without special mention were from Sigma-Aldrich (St. Louis, MO).

HCC cell lines and culture

HCC cell line SMMC-7721 was obtained from the Department of Pathology of The Second Military Medical University (Shanghai). SMMC7721-GFP HCC cell line that stably expresses green fluorescent protein was kindly provided by Dr. Qian Huang' laboratory at the First People's Hospital of Shanghai Jiaotong University. All cell lines were grown in DEM medium/10% FBS, supplemented with 100 I.U./ml penicillin G and 100 μg/ml streptomycin (Sigma). In all experiments, these cells were maintained at 37°C in a humidified 5% CO2 incubator.

Human tissue samples

Eighteen liver cirrhosis (LC) and 85 tumor specimens along with matched noncancerous tissues were obtained from patients with hepatocelluar carcinoma after hepatectomy or liver transplantation in Qidong Liver Cancer Institute, Shanghai Eastern Hepatobiliary Surgery Hospital, Guangxi Medical University and the First Affiliated Hospital, Zhejiang University. The 85 HCC patients comprised 68 males and 17 females with mean age of 47.3 years (ranging from 27 to 86). HCC grade were histologically defined according to Edmondson-Steiner criteria. Ten normal liver samples were from persons of accident death. All human sample collection procedures were approved by the China Ethical Review Committee.

Magnetic sorting and culture of CD133+ HCC cells

Cells were labeled with primary CD133/1 antibody (mouse IgG1, Miltenyi Biotec, 1 μl per million cells), subsequently magnetically labeled with rat anti-mouse IgG1 Micro beads (Miltenyi Biotec, 20 μl per 10 million cells) and separated on MACS LS column (Miltenyi Biotec). All the procedures were carried out according to manufacturer's instructions. The purity of sorted cells was evaluated by flow cytometry and Western blot. The flow cytometry was carried out with a Epics Altra machine (Beckman Coulter), using CD133/1 primary antibody (Miltenyi Biotec) and FITC-conjugated secondary antibody (Sigma).

The isolated CD133+ cells were cultured before assay in DMEM/17% FBS (Invitrogen, Carlsbad, CA), supplemented with 200 mM L-glutamine (Invitrogen), 25 mM HEPES (Sigma), 10 mM nonessential amino acids (Invitrogen), 10 ng/ml leukemia inhibitory factor (Chemicon, Temecula, CA), 2%(v/v) 2-Mercaptoethanol (Sigma) and antibiotics.

Immunostaining of cultured cells and tissue sections

The CD133 sorted cells were plated and grown on glass slides for 48 hr, fixed with 100% ice-cold acetone, and blocked with SuperBlock solution (Pierce, Rockford, IL). Slides were incubated using goat anti-CD133 polyclonal antibody (Santa Cruz) with optimal dilution and time, labeled by EnVision HRP (mouse) kits, stained by DAB Liquid and counterstained by Mayer's hematoxylin.

All of tissue samples for immunohistochemistry were fixed in phosphate-buffered neutral formalin, embedded in paraffin, and cut into 5-μm-thick sections. The tissues for cryostat sections were freshly collected, snap-frozen in liquid nitrogen, OCT-embedded and cut into 10-μm-thick sections.

Tissue sections were deparaffinized, rehydrated and microwave-heated in sodium citrate buffer (10 mM, pH 6.0) for antigen retrieval. Then, the sections were incubated with 3% hydrogen peroxide/PBS for 5 min, and blocked with SuperBlock solution (Pierce). Immunodetection was conducted as mentioned earlier. In addition, fluorescent immunostaining of CD133 antigen in human HCC tissues were carried out on frozen sections. All the slides were observed and photographed with a Axioskop 2 microscope (Carl Zeiss, Oberkochen, Germany).

Western blotting

Cell lysis, sample preparation, SDS-PAGE separation and electrotransferation to nitrocellulose membrane were performed with common method. Immunoblotting was carried out with mouse anti-CD133/1 IgG1 (Miltenyi Biotec) and visualized using SuperSignal West Femto Maximum Sensitivity Substrate (Pierce). β-actin was reprobed as a loading control.

Colony formation assay

Colony formation assay in soft agar was carried out as described previously.22 In brief, base layer was made by mixing 1% soft agar and equivalent 2× medium and prepared in 96-well plate at first, then CD133-sorted SMMC-7721-GFP cells were harvested, suspended in medium containing 0.3% soft agar (Invitrogen), and seeded upon the base layer at a density of 20 cells per well. All experiments were conducted in triplicates. Plates were maintained at 37°C in humidified incubator and were fed every 5 days with 0.1 ml of medium. As for colony formation assay in Matrigel, CD133 sorted cells were plated in DMEM/10% FBS on 96-well plates precoated with 100 μl Matrigel (Becton Dickinson, Franklin Lakes, NJ). After 3 weeks, the number of the colony formation was assessed by counting under microscope. Representative views were photographed.

For the second round of colony formation assay, some colonies from the first round were picked up and their cells were reseeded in 96-well plates with soft agar and Matrigel with the same procedures as the first cycle. The results of colony formation were assessed same as earlier.

Animal preparation and xenograft tumorigenicity assay

Six- to eight-week-old female congenitally immune-deficient nonobese diabetic/severe combined immune-deficiency (NOD/SCID) mice and bg/nu/xid (BNX) mice were randomly divided into groups and maintained under standard conditions according to the institution's guidelines.

For peritoneal inoculation, various numbers of CD133+ and CD133 SMMC-7721 cells ranging from 100 to 500 cells per mouse was injected i.p. in 200 μl serum-free DMEM / Matrigel (BD) (1:1) to the female BNX mice.

For orthotopic inoculation, an 8-mm transverse incision was made in mouse's upper abdomen under anesthesia. Two-thousand sorted CD133+ and CD133 SMMC-7721 cells, suspended in 50 μl serum-free DMEM/Matrigel (1:1), were injected into the left hepatic lobe of the mice with a microsyringe. Tumor formation was monitored from 1 week after inoculation. After 10 weeks, all of the mice were killed, and tumor masses and inoculated murine liver tissue samples were dissected, weighed grossly and microscopically examined.

Statistical analysis

Data were presented as mean ± SD and evaluated with Student's t-test. p < 0.05 was accepted as statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

CD133+ HCC cells exist in human hepatoma cell line, human primary HCC and cirrhotic liver tissues

After screening in HCC cell lines, we found that CD133-positive cells were presented in SMMC-7721 HCC cell line (Fig. 1a), which was established from a male HCC patient in 197723 and well used in our laboratory. Although the CD133+ cells represented only a very small subpopulation (0.1–1%), they could be sorted out by utilizing MACS, which enriched the CD133+ cells as revealed by immunocytochemical examination (Fig. 1b) and Western blot analysis (Fig. 1e). CD133+ SMMC-7721 cells were also analyzed by flow cytometry before and after MACS sorting for purity, which ranged from 60.2% to 91.2% compared to unsorted SMMC-7721 cells (0.1–2%) (Figs. 1c and 1d).

thumbnail image

Figure 1. CD133 expression and sorting purity analysis in hepatoma cells. Human hepatocellular carcinoma SMMC7721 cell line was immunocytochemically stained using an anti-CD133 polyclonal antibody. CD133+ cells were indicated (arrows). CD133+-SMMC7721-GFP cells were enriched by magnetic beads sorting, (a) before- (arrows) and (b) after-sorting (brown) cells were showed (×400). Flow cytometry histogram in representative HCC SMMC-7721 cells, with the first peak (gate M1) representing cells negative for CD133 expression, and the second peak (gate M2) representing CD133 positive cells. (c) CD133+ cells range from 0.1% to 2% before sorting for CD133 expression by magnetic bead cell sorting (MACS). (d) MACS sorted CD133+ cells were collected, checked for purity by flow cytometry; the purity was found to range from 60.2% to 91.2%. (e) Western blotting: CD133 expression of sorted positive (+) and negative (−) SMMC-7721 cells were analyzed by using a monoclonal anti-CD133 IgG. β-actin was used as an internal control.

Download figure to PowerPoint

In addition to HCC cell lines, we further studied the CD133 expression in 85 human HCC tissue samples by immunohistochemistry examination. Among 85 HCC samples, CD133+ cancer cells were found in most of these tissue samples, though they only represented only a subpopulation of 0.1–1% of the total cancer cells, varying from case to case. In most instances, these CD133+ cells were distributed in a scattered pattern. But in some cases, they may also present in a clustered form. The representative features were illustrated in Figures 2a2d.

thumbnail image

Figure 2. Detection of CD133+ cells in human primary HCC and relevant liver tissues. Paraffin-embedded sections of HCC (a, b), noncancerous liver (e, f) and cirrhotic patient liver (g, h) were prepared and detected by immunohistochemical staining with a polyclonal anti-CD133 IgG. HCC frozen sections were directly stained witha monoclonal anti-CD133-PE IgG1 (c, d). [(e, g) ×100; (ad, f, h) ×400; inserts (c, d) ×1,000].

Download figure to PowerPoint

As cirrhosis usually occurs in most HCC noncancerous liver, we examined the CD133 expression pattern in these noncancerous liver tissues in HCC patients as well as the liver tissues from liver cirrhosis patients. In fact, the CD133 expression pattern was quite similar in these 2 types of cirrhotic tissues (Figs. 2e2h). Notably, we found that CD133 was expressed in a small subset of hepatocytes, particularly at the peripheral area of cirrhotic pseudolobules. These CD133+ periportal hepatocytes were in oval or cubical shape, with a size much smaller than the mature hepatocytes. Moreover, CD133 was only occasionally expressed in epithelia of biliary tubules (Figs. 2g2h), and in some structures at the portal region, which displayed as lumen-less double-layered epithelial clusters or stretches (arrow at Figs. 2g2h), suggesting as transitional structures from biliary ductule toward small hepatocytes. In some case, we also found that CD133 was expressed in a proportion of endothelial cells of vascular capillaries at portal area. For comparisons, we also examined liver tissues from normal individuals. In contrast to liver tissues in HCC or cirrhosis, no CD133 expression was observed in normal hepatocytes or biliary epithelial structures in portal area.

CD133+ HCC cells showed high clonogenicity

CD133+ SMMC-7721-GFP HCC cells enriched by MACS efficiently formed colonies in 3 weeks after growing both in soft agar and Matrigel (Figs. 3a3d). Based on calculation of numbers of colony formation per 100 seeded cells, the value was around 37.7 ± 1.0/13.7 ± 0.9 for CD133+/CD133 cells in soft agar, and 41.2 ± 1.1/12.1 ± 0.4 in Matrigel, respectively (Table I). We further tested the colony formation from cells isolated from primary colonies; the colony formation efficiency in soft agar in the secondary round of plating was about 24.8 ± 0.8/4.1 ± 1.1 for CD133+/CD133 cells in soft agar, and 28.7 ± 0.7/3.2 ± 0.4 in Matrigel, respectively (Table I). These results strongly indicated that the CD133+ cells possessed the capability for maintenance of their high clonogenicity.

thumbnail image

Figure 3. Colony formation assays of magnetically sorted SMMC-7721-GFP cells in soft agar and Matrigel. CD133+ HCC cells (a)/Matrigel (c) and CD133 HCC cells in soft agar (b)/Matrigel (d) were taken under inverted microscope [(a, b) ×400; (c, d) ×100]. The difference of colony formation efficiency in soft agar between CD133+ and CD133 cells are shown in Table I.

Download figure to PowerPoint

Table I. Colony Formation Efficiency (Colony Numbers per 100 Seeded Cells)1
 Soft agarMatrigel
CD133+CD133CD133+CD133
  • 1

    CD133+ and CD133 cells were seeded in soft agar and matrigel to analyze their colony formation capacity (see Materials and Methods). The colony formation efficiency was presented as the number of colonies formed per 100 seeded cell.

  • *

    p < 0.05.

First round (primary cycle)37.7 ± 1.0*13.7 ± 0.941.2 ± 1.1*12.1 ± 0.4
Second round (secondary cycle)24.8 ± 0.8*4.1 ± 1.128.7 ± 0.7*3.2 ± 0.4

Tumorigenicity of CD133+ HCC cells in immunodefficient mice by intraperitoneal and intrahepatic inoculation

To validate the capacity of tumorigenicity of CD133+versus CD133 HCC cells in immunodeficient mice, we inoculated intraperitoneally 100, 500 and 2,000 CD133+ SMMC-7721 cells and the same amount of CD133 cells into BNX mice, respectively. In our pilot studies, only more than 2 × 106 unsorted SMMC-7721 cells could generate tumors by intraperitoneal inoculation in 1 month. In the present experiments, 3/6 mice received 100 CD133+ cells, 3/5 with 500 CD133+ cells and 5/5 with 2,000 CD133+ cells developed tumors in abdominal cavity, while no tumors were found in mice treated with same amount of CD133 HCC cells (Table II). The tumors generated by CD133+ cells grew at the parietal peritoneum and the mesentery adipose tissue. The lesions were confirmed by histopathological examination.

Table II. Tumorigenicity of CD133+ SMMC-7721 Cells in BNX Mice Xenograft Assay* by I.P. Injection
Cell numbers injected per mouseCD133+ SMMC-7721 cellsCD133 SMMC-7721 cells
  • 1

    Number of cancer cell inoculated in BNX mice. CD133+ cells were sorted with MACS. The result of tumor formation was observed for 10 weeks after injection.

10013/60/6
50013/50/5
2,00015/50/5

We also implanted 2,000 CD133+ SMMC7721 cells into NOD/SCID mice by intrahepatic inoculation. Tumor nodules were detected in the liver of 4 out of 6 mice, among which two were microscopic lesions; while no tumors were found in all 6 mice with intrahepatic injection of 2,000 CD133 HCC cells (Table III). The gross and microscopic features were illustrated in Figures 4c4f. The results represented above strongly indicated that the CD133 marker has sorted out a subpopulation of HCC cells with high capability for tumor formation.

thumbnail image

Figure 4. Tumorigenecity of CD133+ HCC cells. CD133+ and CD133 SMMC-7721-GFP cells ranging from 100 to 2,000 cells per mouse were injected intraperitoneally in 200 μl serum-free DMEM/Matrigel (BD) (1:1) to the female BNX mice. (a, b) shows the representative gross and microscopic pictures summarized in Table II. Two-thousand CD133+ and CD133 SMMC-7721-GFP cells per mouse were inoculated intrahepatically into NOD/SCID mice. cf shows the microscopic features of tumor nodules formed in livers (margins were shown by black arrows). [(b) ×50; (c, d) ×200; (e, f) ×400]. (Data refer to Table III).

Download figure to PowerPoint

Table III. Tumorigenicity of CD133+ SMMC-7721 Cells in NOD/SCID Mice Xenograft Assay by Intrahepatic Injection
 Number of mice with tumor formation in liver/total number of mice inoculated with 2,000 cancer cells
  1. Two-thousand cancer cells were inoculated into liver of NOD/SCID. CD133+ cells were sorted with MACS. The result of tumor formation was observed for 10 weeks after injection.

CD133+ SMMC-7721 cells4/6
CD133 SMMC-7721 cells0/6

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

In past several years, some putative tissue-restrict stem-cell markers were also found to be present in their tumor counterparts, and could serve to identify a minor subpopulation of cells possessing high tumorigenicity.7, 10, 11 CD133 is widely used as a molecular marker for characterizing normal neuronal, hematopoietic and other stem cells, and some types of cancer stem or stem-like cells in tumors originating from neuroepithelia, prostate epithelia and hematopoietic tissues.10, 11, 12 It was reported that in regenerative liver after massive necrosis, CD133 was coexpressed with c-Kit on the surface of candidate hepatic progenitor cells.21 By immunostaining, a group of stem-like cells, with surface marker pattern similar to normal liver progenitors, had been revealed in hepatoblastoma.24

In the present study, we demonstrated that CD133 was expressed in SMMC-7721 cells, and also consistently present in other human HCC cell lines we examined: BEL-7402, Hep3B, Huh-7, MHCC-LM3, MHCC-97L (data not shown). Although long-term cultured cell lines are thought to lose some properties of parental tumor and consist of less divergent population, the cell heterogeneity still exists. For example, side population, which features the efflux of Hoechst 33342 dye and some cytotoxic drugs, was reported to be present in HCC cell line and highly tumorigenic.25 Our data support the existence of a highly tumorigenic subpopulation in HCC cell line.

CD133+ cells were also found here in some of the primary human HCC tissue samples, their noncancerous counterparts and cirrhotic liver tissues, but not in normal liver tissues. CD133 expression was detected in HCC-adjacent lesions as well as liver cirrhotic tissues, but not in normal human liver. In cirrhotic tissues, CD133 was expressed only in some small hepatocytes at periphery of pseudolobules, occasionally in trabeculae-like hepatocytes extending inside pseudolobules. Of particular interest, CD133 was also expressed in some cells at portal areas in a form of ductule-like structures, extending into periportal hepatocytes of pseudolobules (Figs. 2g and 2h). Further studies are needed to define if these cells revealed by CD133 correspond to cells of Hering's canal or other putative hepatic stem cells.26, 27 As oval cells or hepatic progenitor cells are supposed to proliferate only under certain conditions such as liver regeneration, it could explain why CD133 was negative in normal liver.

Early from 1977, Hamburger et al. introduced the primary bioassay of human tumor stem cells,28 and found that most tumor types are heterogeneous, comprising cells with different phenotypes and different proliferative and malignant potentials. Since then a number of reports have described that only a small subset of cells is clonogenic in culture and in vivo.7, 8, 10, 11 Our data showed that CD133+ HCC SMMC-7721 cells have high clonogenicity in vitro and high tumorigenicity in vivo compared with CD133 cells, which suggests that these cells have at least some of the properties of cancer stem or stem-like cells. We had noticed the difference in potential of tumorigenicity by peritoneal and orthotopic injection: the intraperitoneal inoculation developed multifocal tumor quickly, but intrahepatic implantation showed a slower pattern even with increased amount of cells injection. These data implicated that the crucial role of the microenviroment for cancer cell propagation should be taken into consideration for assessment of tumorigenicity in vivo.

As to the origin of HCC, hepatic progenitor cells are considered to be prone to accumulating gene mutations and are targets of malignant transformation. HCC shares some immature differentiation markers with fetal liver such as AFP and CK19, which indicates the “stemness” of the cancer. It is also widely reported that hepatic progenitor cells were activated in the context of consistent tissue repair, such as alcoholic liver damage and chronic viral hepatitis.27 Whether the CD 133 expression on HCC cells was a consequence of dedifferentiation or a phenotype of maturation-arrest remained to be determined. Moreover, whether CD133 could be one of the markers for characterizing HCC stem-like cells needed to be further confirmed by direct cell sorting from primary HCC tumor tissues. Many technical problems in isolating and sorting sufficient amount specified types of intact cells from a particularly fibrous tissues enriched cancer like HCC remained a challenge for identification of HCC stem cells.

Taken together, our data demonstrated that a small subpopulation of CD133+ HCC cells existed in HCC cell lines, and harbored higher clonogenicity in vitro and potent tumorigenicity in immunodeficient mice, these findings could provide some insight into isolation and characterization of cancer initiating cells, which might be a potential target for drug design or biotherapy, such as tumor vaccination.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

We thank Dr. Yuhong Xu for critically reading the manuscript.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References