SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

A novel theory in the field of tumor biology postulates that cancer growth is driven by a population of stem-like cells, called tumor-initiating cells (TICs). We previously identified a TIC population derived from hepatocellular carcinoma (HCC) that is characterized by membrane expression of CD133. Here, we describe a novel mechanism by which these cells mediate tumor growth and angiogenesis by systematic comparison of the gene expression profiles between sorted CD133 liver subpopulations through genome-wide microarray analysis. A significantly dysregulated interleukin-8 (IL-8) signaling network was identified in CD133+ liver TICs obtained from HCC clinical samples and cell lines. IL-8 was found to be overexpressed at both the genomic and proteomic levels in CD133+ cells isolated from HCC cell lines or clinical samples. Functional studies found enhanced IL-8 secretion in CD133+ liver TICs to exhibit a greater ability to self-renew, induce tumor angiogenesis, and initiate tumors. In further support of these observations, IL-8 repression in CD133+ liver TICs by knockdown or neutralizing antibody abolished these effects. Subsequent studies of the IL-8 functional network identified neurotensin (NTS) and CXCL1 to be preferentially expressed in CD133+ liver TICs. Addition of exogenous NTS resulted in concomitant up-regulation of IL-8 and CXCL1 with simultaneous activation of p-ERK1/2 and RAF-1, both key components of the mitogen-activated protein kinase (MAPK) pathway. Enhanced IL-8 secretion by CD133+ liver TICs can in turn activate an IL-8-dependent feedback loop that signals through the MAPK pathway. Further, in its role as a liver TIC marker CD133 also plays a functional part in regulating tumorigenesis of liver TICs by way of regulating NTS, IL-8, CXCL1, and MAPK signaling. Conclusion: CD133+ liver TICs promote angiogenesis, tumorigenesis, and self-renewal through NTS-induced activation of the IL-8 signaling cascade. (Hepatology 2012)

Hepatocellular carcinoma (HCC) is the most commonly diagnosed malignancy of the liver and is the fifth most frequently diagnosed cancer worldwide.1-3 This disease generally presents with a high degree of neovascularization, which plays an important role in tumor growth and progression. Surgical resection and liver transplantation are potentially curative in HCC patients diagnosed at an early stage of cancer progression. However, HCC is often detected only when it reaches an advanced stage, when these treatments are generally no longer an option. Additionally, this disease is particularly difficult to treat because of the high recurrence rate and its chemotherapy-resistant nature.4 Most nonsurgical therapies targeting HCC are directed against the bulk of the tumor mass. These therapies are generally able to initially reduce the size of the primary tumor; however, they ultimately fail to eradicate the lesion in full, resulting in disease relapse. Recent research efforts in the fields of stem cell and cancer biology have resulted in a “stem cell model of carcinogenesis” that postulates that the capability to maintain tumor formation and growth is found in a small population of cells called tumor-initiating cells (TICs) or cancer stem cells (CSCs). The stem cell- and cancer cell-like characteristics of these cells are believed to render these cells resistant to conventional therapies and allow them to drive tumorigenesis. Past studies by our research group have found that HCC is also hierarchically organized; it originates from a group of primitive stem cells distinguished by their CD133 surface phenotype.5, 6 CD133+ HCC TICs were also found to display elevated resistance to conventional chemotherapeutic drugs due to dysregulation of the AKT pathway.7 Recently, we delineated a novel pathway by which TICs sustain self-renewal through miR-130b and TP53INP1 dysregulation.6 Our findings are in agreement with past reports by other research groups that also identify CD133 as a marker of TICs in HCC.8-11 Despite our growing understanding concerning the existence of liver TICs, the underlying molecular mechanism by which this population of cells mediates tumor growth and maintenance remains unclear. In an attempt to better characterize these cells, a genome-wide messenger RNA (mRNA) expression profiling approach was employed to compare the gene expression profiles between CD133+ and CD133 cells isolated from the HCC cell lines Huh7 and PLC8024. Our results provide in vitro and in vivo evidence that interleukin-8 (IL-8) plays a vital role in driving HCC tumor growth, self-renewal, and angiogenesis in CD133+ liver TICs. These results demonstrate that the preferential expression of IL-8 is mediated through the neurotensin (NTS)-activated mitogen-activated protein kinase (MAPK) signaling cascade. Findings from this study should provide not only insight into the mechanism underlying the ability of TIC to drive HCC but also lead to new approaches for the development of more effective cancer therapeutics.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Fresh Clinical Tissue Collection and Processing.

Fresh human liver tumor and adjacent nontumor tissues were obtained from 12 patients undergoing hepatectomy for HCC in accordance with the ethical standards established by the Institutional Committee on Human Experimentation. Specimens were collected between 2008-2009 at the Queen Mary Hospital or the Prince of Wales Hospital in Hong Kong. The patients had received no previous local or systemic treatment prior to operation. Histological examination was conducted by staff pathologists of the hospitals, and the diagnosis of HCC was established by microscopic examination in every case. Surgical specimens were obtained at the time of resection from all patients. All samples reached the laboratory within 20 minutes, where they were immediately mechanically disaggregated and digested with type IV collagenase (Sigma-Aldrich, St. Louis, MO) and then resuspended in Dulbecco's modified Eagle's medium (DMEM)/F12 medium containing penicillin (500 U/mL) and streptomycin (500 μg/mL). Single-cell suspensions were obtained by filtration through a 100-μm filter (BD Biosciences, Franklin Lakes, NJ). Dead cells and red blood cells were removed using the Ficoll gradient centrifugation method. The remaining red blood cells were lysed with ACK buffer (Invitrogen, Carlsbad, CA). The viable cells were counted using Trypan blue staining.

Cell Culture.

Hep3B was obtained from the American Type Culture Collection (Manassas, VA). Huh7 was provided by Dr. H. Nakabayashi, Hokkaido University School of Medicine, Japan.12 HepG2, QSG-7701, QGY-7703, BEL7402, and PLC8024 were obtained from the Institute of Virology, Chinese Academy of Medical Sciences (Beijing, China). MiHA was provided by Dr. J.R. Chowdhury, Albert Einstein College of Medicine (New York).13 All cell lines were cultured in DMEM medium (Invitrogen) supplemented with 10% fetal bovine serum (Invitrogen), penicillin (500 U/mL), and streptomycin (500 μg/mL) in a 5% CO2 incubator at 37°C.

Lentiviral Transduction.

IL-8 short hairpin RNA (shRNA) lentiviral knockdown vectors (Sigma-Aldrich, NM_000584) and CD133 shRNA lentiviral knockdown vectors (Sigma-Aldrich, NM_006017) were packaged using MISSION Lentiviral Packaging Mix (Sigma-Aldrich). Stable clones were selected using puromycin. For both systems cells were infected with lentiviral media at multiples of infection of 10 in the presence of 8 mg/mL polybrene (Sigma-Aldrich) overnight in a 37°C incubator.

Reagents.

Recombinant IL-8 and neurotensin were purchased from R&D Systems and Sigma-Aldrich, respectively.

Statistical Methods.

All statistical analyses were performed using PASW Statistics 18.0 (SPSS, Chicago, IL), with the exception of the results depicted in bar graphs. These bar graph analyses were performed by applying independent t tests using Microsoft Office Excel software (Redmond, WA). P < 0.05 was considered significant.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

CD133+ Liver TICs Display Unique Gene Signatures and Functional Networks.

In an attempt to characterize the molecular mechanisms by which CD133+ liver TICs mediate tumor formation and growth, a genome-wide mRNA expression profiling screen was employed to compare gene expression profiles between CD133+ liver TICs and their CD133 counterparts isolated from the HCC cell lines Huh7 and PLC8024. CD133 sorted cells were isolated by flow cytometry, resulting in a considerable enrichment in the CD133+ cell population (purity >94%) and, more important, a very pure negative selection (purity >97%) of CD133 cells (Supporting Information Fig. S1). By integrating both Robust Multichip Average (RMA) and MicroArray Suite 5 (MAS5) analysis using Genespring GX software, 149 common differentially expressed genes that displayed a fold change of greater than two were identified (Fig. 1A,B). Of these 149 genes, 113 were up-regulated and 36 were down-regulated (Supporting Information Table S2). Included in these genes were the known hepatic stem/progenitor markers CD44 and EpCAM, which have previously been shown to mark liver TICs and overlap with CD133 expression.10, 14 Additional analysis by ingenuity pathway analysis (IPA) software identified significantly altered biological and toxicity profiles with regard to cancer, liver proliferation, liver cirrhosis, liver hepatitis, and liver degradation activities (Fig. 1C,D).

thumbnail image

Figure 1. CD133+ liver TICs display unique gene signatures. (A) Unique gene signatures of CD133+ TICs and their CD133 counterparts isolated from the HCC cell lines Huh7 and PLC8024, as shown by hierarchical cluster analysis (2-fold cutoff). Left, analysis by Genespring GX, MicroArray Suite 5 (MAS5) and right, analysis by Robust Multichip Average (RMA). Each cell in the matrix represents a particular expression level of genes, where red and green cells indicate high and low gene expression, respectively. (B) Venn analyses of the up- and down-regulated genes in the HCC cell lines PLC8024 and Huh7 identified by MAS5 and/or RMA analyses. Genes that were significantly and commonly dysregulated are shown in the overlapping area. Only genes whose expression levels displayed greater than a 0.5-fold decrease or a 2-fold increase were selected. (C,D) Functional analysis of CD133-regulated genes by IPA.

Download figure to PowerPoint

IL-8 Is Significantly Overexpressed in CD133+ Liver TICs.

Further analysis using the IPA software identified that a functional network involving IL-8 was significantly altered in the CD133+ liver TIC subpopulation (Fig. 2A). To validate our profiling findings, we performed quantitative polymerase chain reaction (qPCR) on selected commonly deregulated mRNAs involved in the IL-8 pathway, as well as known hepatic stem cells and cancer stem cell markers, using sorted CD133 cells isolated from the HCC cell lines PLC8024 and Huh7. Preferential dysregulation of the IL-8 functional network genes endothelin-1 (EDN1), CXCL1, CXCL3, CXCL5, CXCL6, NTS, IL-8 (Fig. 2A); and hepatic stem cell and cancer stem cell markers cyto-keratin 19 (KRT19), CD44, and EpCAM was verified (Supporting Information Table S3). As a number of members in the suggested network function in IL-8 signaling, and as it is known that IL-8 signaling plays an important role in carcinogenesis, we focused our studies on IL-8 (up-regulated in the CD133+ subpopulation: 5.11-fold in Huh7 and 2.95-fold in PLC8024, as detected by complementary DNA [cDNA] microarray analysis). The IL-8 receptors CXCR1 and CXCR2 were also consistently expressed in cell lines that express CD133 and IL-8, including Huh7, PLC8024, and Hep3B (Supporting Information Fig. S2). Screening a panel of liver cell lines revealed that IL-8 expression was strongly correlated with the expression of CD133. Liver cell lines with low CD133 expression also displayed low or no expression of IL-8 (MiHA, QSG-7701, QGY-7703, BEL7402, HepG2), whereas liver cell lines exhibiting high CD133 expression showed a more abundant expression of IL-8 (PLC8024, Huh7, Hep3B). As IL-8 is a secreted protein, we next extended our studies to the proteomic level by using both immunoblot and enzyme-linked immunosorbent assay (ELISA). We also found a positive correlation between CD133 and endogenous and secreted IL-8 across the panel of liver cell lines (Fig. 2B,C). Additionally, preferential expression of both endogenous and secreted IL-8 was identified in CD133+ subpopulations isolated from Huh7 and PLC8024 cell lines (Fig. 2D,E), compared to the CD133 subpopulations. Similar observations were also consistently found in CD133+ (CD45 and CD31, depleted of both hematopoietic and endothelial progenitors, respectively) cells isolated from HCC clinical samples (Fig. 2F; n = 12; P = 0.006).

thumbnail image

Figure 2. IL-8 is significantly overexpressed in CD133+ liver TICs. (A) Interactions and functional network identified by IPA. Genes in red and green indicate under- and over-expressed genes found in CD133+ TICs compared to their CD133 counterparts, and genes enclosed in blue are members of the IL-8 signaling pathway. (B) IL-8 and CD133 gene expression coincide across a panel of liver cell lines, as detected by qPCR (top) and ELISA (bottom) (Pearson correlation: P = 0.002). (C) Endogenous (End.) and secreted (Sec.) IL-8 expression coincides with CD133 expression across the same panel of liver cell lines as detected by immunoblot analyses (Pearson correlation: P = 0.003). (D,E) IL-8 is preferentially expressed in CD133+ TICs compared with their CD133 counterparts isolated from Huh7 and PLC8024, as detected by immunoblot and ELISA assays (*P = 0.017 and **P = 0.002). (F) A dot-plot graph shows significant up-regulation of IL-8 in CD133+ TICs as compared to CD133 cells in 12 HCC clinical specimens, as detected by qPCR (paired sample t test; P = 0.006).

Download figure to PowerPoint

IL-8 Promotes Tumorigenicity and Angiogenesis of CD133+ Liver TICs.

To determine whether IL-8 is essential for the maintenance of CD133+ liver TICs, we further investigated the effect of IL-8 on the ability of CD133+ liver TICs to promote tumorigenicity and angiogenesis. As IL-8 is a well-known angiogenic factor, we first investigated the effect of the tumor cell supernatant on the proliferation and tube formation ability of HUVEC endothelial cells. Bromodeoxyuridine (BrdU) proliferation and capillary tube formation assays revealed that the supernatants of CD133+ liver TICs isolated from the HCC cell lines Huh7 and PLC8024, which are known to contain elevated levels of IL-8, were preferentially capable of promoting growth and stimulating capillary tube structure formation in HUVECs when compared to supernatants collected from CD133 cells (Fig. 3A,B). To confirm that this effect was indeed dependent on the preferential overexpression of IL-8 in CD133+ liver TICs, an IL-8 neutralizing antibody and IL-8 shRNA knockdown were used to abolish IL-8 expression/activity. When supernatants from CD133+ liver TICs were pretreated with an IL-8 neutralizing antibody (1 μg/mL), the observed angiogenic and pro-proliferation activity was eliminated (Fig. 3A,B). Similarly, stable IL-8 knockdown in CD133-expressing HCC cells (Fig. 3A, right) resulted in a reduced ability of these cells to proliferate and form capillary tubes when compared to cells transduced with an NTC (Fig. 3A, left, and B). The effect of IL-8 on the growth of CD133+ liver TICs was also investigated using an animal model. CD133-expressing Huh7 cells were injected subcutaneously in the right flanks of nude mice to initiate xenograft tumors. Treatment of the xenografts with an IL-8 neutralizing antibody (1 μg/μL) began 4 weeks after the initial tumor inoculation, when tumor size reached approximately 1 mm3. A 2-week treatment of Huh7 cells that expressed a high percentage of CD133+ liver TICs with an IL-8 neutralizing antibody resulted in a significant reduction in tumorigenic potential (Fig. 3C; n = 5 per group). Immunohistochemical staining of resected xenografts revealed a significant reduction in microvessel density (MVD), as indicated by the reduced expression of the endothelial marker CD34. A reduction in tumor proliferation was indicated by a decreased expression of the proliferation marker proliferation cell nuclear antigen (PCNA), and an increase in tumor apoptosis was demonstrated by increased Transferase-Mediated dUTP Nick-End Labeling (TUNEL) staining following IL-8 neutralizing antibody treatment (Fig. 3D). The increased tumorigenicity mediated by IL-8 was further validated by a gene knockdown approach, where IL-8-repressed cells (clones 30 and 31) showed a diminished ability to initiate tumor growth as indicated by a reduction in both tumor incidence and size (Supporting Information Fig. S3; n = 5 per group).

thumbnail image

Figure 3. IL-8 promotes tumorigenicity and angiogenesis of CD133+ liver TICs. (A) Significant induction of endothelial cell proliferation by supernatants collected from CD133+ TICs as compared to CD133 cells (Huh7), as shown by BrdU proliferation assay (*P = 0.002; **P = 0.003). This enhancement was abolished upon addition of an IL-8-neutralizing antibody (IL-8 N. Ab). (B, left) Decreased induction of endothelial cell proliferation was also found following silencing of IL-8 by shRNA knockdown in Huh7 (clones 30 and 31), compared to nontransduced cells (*P = 0.003; **P = 0.004). (A, right) Immunoblot analysis showing decreased endogenous and secretory IL-8 expression in Huh7 following stable shRNA-mediated IL-8 knockdown, as compared to nontransduced cells. (B) Increased endothelial cell tube formation by supernatant collected from CD133+ TICs (Huh7) as shown by capillary tube formation assay. This effect was abolished upon addition of IL-8 N. Ab. A similar loss of tube formation was observed for supernatants from Huh7 cells when IL-8 was repressed as compared to nontransduced cells. (C) Xenogen images showing significant inhibition of tumor growth in xenograft tumors treated with IL-8 N. Ab (n = 5 per group). Plot of tumor volume from 3 to 13 days following IL-8 N. Ab treatment (General Linear Model, repeated measures: P < 0.001). (D) Hematoxylin and eosin (H&E) staining of harvested tumors confirmed a primary HCC phenotype. Immunostaining revealed a reduction of MVD (CD34), a reduction in tumor proliferation (PCNA), and an increase in tumor apoptosis (TUNEL) of xenograft tumors treated with IL-8 N. Ab (800× magnification).

Download figure to PowerPoint

IL-8 Inhibition Attenuates the Self-Renewal Ability of CD133+ Liver TICs.

We next chose to determine if IL-8 can regulate the self-renewal ability of CD133+ TICs. Repression of IL-8 expression (clones 30 and 31) in CD133+ TICs significantly suppressed the ability of these cells to initiate spheroid formation when grown in nonadherent serum-free conditions in vitro (Fig. 4A). These cells also exhibited a reduced ability to expand in subsequent in vitro serial propagations (Supporting Information Fig. S4). Furthermore, CD133+ TICs with IL-8 repression displayed a significantly lower expression of stem cell-associated genes, including Nanog, Notch-1, and ATP-binding cassette half-transporter ABCC1 (Fig. 4B).

thumbnail image

Figure 4. IL-8 is critical for growth and self-renewal of CD133+ liver TICs. (A) Compared to nontransduced cells, Huh7 cells with IL-8 suppression (clones 30 and 31) generated smaller and fewer spheroids. (B) qPCR analysis of stem cell-associated and multiple-drug-resistant transporter genes in Huh7 cells with IL-8 suppression and in nontransduced cells. Cells transduced with NTC were defined as 1 (calibrator) and the relative expression of stem-cell associated genes in cells transduced with IL-8 shRNA clones 30 and 31 were expressed as the fold difference relative to this baseline. (C) Representative xenogen images of tumors (red arrows) formed in NOD/SCID mice when injected subcutaneously with CD133+IL-8+, CD133IL-8+, CD133+IL-8, and CD133IL-8 cells isolated from Huh7 cells (picture shown representative of the 50,000 cells injected group; see also Table 1) (n = 5 per group). (D) H&E staining of harvested tumors confirmed a primary HCC phenotype. Immunostaining revealed a reduction of MVD (CD34), a reduction in tumor proliferation (PCNA), and an increase in tumor apoptosis (TUNEL) of the xenograft tumors (200× magnifications).

Download figure to PowerPoint

CD133+ IL-8 Expressing Subpopulation Is Critical for Growth and Maintenance of Liver TICs.

To assess the degree by which enhanced IL-8 secretion supports growth and maintenance of CD133+ liver TICs, IL-8 repressed cells (clone 30) and IL-8 expressing cells (nontarget control) were sorted into CD133+ and CD133 populations, and then injected subcutaneously into NOD/SCID mice to assess their tumorigenicity. The results showed that the tumorigenicity of the IL-8 repressed CD133+ liver TICs (CD133+IL-8) was greatly reduced compared to the IL-8-expressing CD133+ cells (CD133+IL-8+) (Fig. 4C, Table 1; n = 5 per group). Xenograft tumors were harvested from both IL-8 repressed or expressing CD133+ liver TICs groups (CD133+IL-8 and CD133+IL-8+) and single cells were isolated from these tumors. Cells isolated from the two groups were again sorted into CD133+ and CD133 populations, which were injected subcutaneously into NOD/SCID mice for secondary transplantation to assess their ability to self-renew and to maintain tumorigenicity. We found that only CD133+ TIC isolated from the IL-8-expressing tumors (CD133+IL-8+) retained the ability of tumor initiation. After serial propagation, none of the cells isolated from an IL-8 repressed tumor were able to form secondary tumors, indicating that CD133+ liver TICs mediate self-renewal through IL-8. In all cases, CD133+ liver TICs were much more efficient when compared to their CD133 counterparts in their ability to initiate tumors (Table 1). Immunohistochemical staining of resected primary xenografts from IL-8-repressed CD133+ liver TICs showed a significant reduction in MVD and proliferation, as well as an increased apoptotic rate, when compared with IL-8-expressing CD133+ liver TICs and/or their IL-8-expressing CD133 HCC counterparts (Fig. 4D).

Table 1. Engraftment Rates of CD133+ and CD133- Subpopulations With or Without IL-8 Silenced (Huh7)
 Cell Number and Tumor Incidence*
Cell Type50,00010,0005,000500
  • *

    Five animals per group.

  • Significance calculation: chi-square test.

Primary Implantation
 CD133+IL-8+4433
 CD133IL-8+2011
 CD133+IL-81000
 CD133IL-80000
 P = 0.058P = 0.01P = 0.038P = 0.038
Secondary Implantation
 CD133+IL-8+5534
 CD133IL-8+1101
 CD133+IL-80000
 CD133IL-80000
 P = 0.002P = 0.002P = 0.038P = 0.01

NTS Activates an IL-8-Mediated Feedback Loop Through MAPK Signaling in CD133+ Liver TICs.

Similar to IL-8, NTS was identified by the initial cDNA microarray analysis to be preferentially expressed in the CD133+ liver TIC subpopulation derived from the HCC cell lines Huh7 (5.25-fold increase) and PLC8024 (2.99-fold increase). Additionally, IPA analysis identified NTS as likely to be involved in the IL-8 functional network (Fig. 2A). Expression of NTS was previously reported to occur in the liver only during fetal development or cancer formation.15 This expression pattern is very similar to that of CD133, as previously found by our group, where CD133 was found to be expressed during the early stages of liver regeneration and in liver cancer.5, 16 Because NTS is known to induce IL-8 production through MAPK pathway activation in colon cancer cell lines,17 we decided to investigate if the preferential activation of IL-8 signaling in CD133+ liver TICs is induced by NTS up-regulation. NTS was preferentially expressed at both the genomic and proteomic (endogenous and secretory) levels in CD133+ liver TICs (Fig. 5A), as validated by subsequent qPCR and immunoblot assays. Its expression also coincided with the expression of CD133 across a panel of HCC cell lines (Fig. 5B). Enhanced NTS expression was also further validated in the CD133+ subpopulation of two additional HCC clinical samples: Patients 3 and 72 (Supporting Information Fig. S5A). The two most common receptors of NTS, NTSR1 and NTSR2, were also expressed in these HCC cell types (Supporting Information Fig. S2). To examine if enhanced NTS secretion in CD133+ liver TICs can activate the MAPK pathway and subsequent IL-8 signaling, exogenous NTS was added to the liver cell line Huh7. Upon NTS treatment (1 μg/mL), phosphorylated ERK1/2 and RAF-1 were both significantly elevated. Expression of IL-8 was also similarly increased, suggesting that increased IL-8 expression in CD133+ liver TICs is at least partially a result of increased NTS secretion due to MAPK signaling (Fig. 5C). Interestingly, the cytokine CXCL1, another member of the IL-8 family, was also found to be preferentially expressed in the CD133+ liver TIC subpopulation isolated from HCC cell lines Huh7 (2.46-fold) and PLC8024 (5.06-fold) (Fig. 5A), as well as in two HCC clinical samples (Supporting Information Fig. S5A). Further, expression of CXCL1 was also significantly enhanced following NTS induction (Fig. 5C). IL-8 and CXCL1 are closely related cytokines with important roles in inflammation, angiogenesis, and tumorigenesis. Additionally, they share common receptors (CXCR1 and CXCR2).18-20 Because IL-8 was previously found to activate MAPK signaling through the CXCR1 and CXCR2 receptors,21, 22 we evaluated whether NTS-stimulated IL-8 production could in turn initiate a positive feedback loop in CD133+ HCC cells. Treatment with exogenous recombinant IL-8 (100 ng/mL) led to a dramatic increase in phosphorylated ERK1/2 expression and the expression of IL-8 and CXCL1 (Fig. 5D). Conversely, IL-8 repression by shRNA led to a concomitant down-regulation of IL-8 and CXCL1 (Supporting Information Fig. S5D). To further confirm that only CD133+ liver TICs are responsive to the stimulatory effects of NTS and IL-8, we sorted CD133+ and CD133 cells from Huh7 and treated the two subpopulations with exogenous NTS or recombinant IL-8. Increased CD133, NTS, and CXCL1 were observed only in the CD133+ subpopulation, suggesting that NTS-induced IL-8 expression in CD133+ liver TICs aids in the maintenance of this specific subpopulation of TICs (Supporting Information Fig. S5B). In addition, treatment of another CD133-expressing HCC cell line Hep3B with exogenous NTS or recombinant IL-8 similarly resulted in an increase in IL-8 and CXCL1 expression (Supporting Information Fig. S5C). Taken together, these results indicate that preferential IL-8 signaling in CD133+ liver TICs is at least in part a result of a positive feedback loop initiated by increased NTS expression.

thumbnail image

Figure 5. NTS activates an IL-8-mediated feedback loop through MAPK signaling in CD133+ liver TICs. (A) Increased expression of NTS and CXCL1 in CD133+ liver TICs as compared to CD133 cells at the genomic and endogenous and secreted proteomic levels in PLC8024 cells (*P = 0.001; **P = 0.002). (B) Genomic and endogenous and secretory NTS and CXCL1 expression coincided with CD133 expression across a panel of liver cell lines (Pearson correlation: P = 0.02). (C,D) Enhanced IL-8 and CXCL1 production, through MAPK signaling activation (phosphorylated ERK1/2 and RAF-1), was detected upon exogenous NTS or IL-8 addition (PLC8024) (D) t test: *P = 0.001; **P = 0.029. (E) t test: *P = 0.002, **P = 0.004).

Download figure to PowerPoint

Knockdown of CD133 Inhibits the NTS/IL-8/CXCL1 Signaling Cascade and Abolishes TIC Properties.

We observed a strong relationship between the expression of CD133, IL-8, NTS, and CXCL1 in CD133+ liver TICs and a panel of HCC cell lines with varying CD133 expression (Figs. 2, 5B). We next determined whether the transmembrane protein CD133 serves only as a marker that is coexpressed with these genes or if it possesses a functional role. Using a lentiviral-based shRNA approach, CD133 expression was stably suppressed (clones 44 and 47) in CD133+ cells isolated from the HCC cell line PLC8024 (Supporting Information Fig. S6). In these CD133 knockdown cells, down-regulation of IL-8, NTS, CXCL1, and RAF-1 was observed (Fig. 6A). Suppression of IL-8 signaling by CD133 knockdown resulted in a significant reduction in the ability of these cells to self-renew, as demonstrated by spheroid formation in vitro (Fig. 6B). Furthermore, an in vivo study revealed that these CD133-repressed clones possessed reduced tumorigenic potential (Fig. 6C; n = 5 per group), as indicated by reduced tumor incidence and size, reduced tumor angiogenesis and proliferation, and increased apoptosis, as indicated by immunostaining for CD34, PCNA, and TUNEL (Fig. 6D). These results suggest that CD133, apart from serving as a marker for liver TICs, also possesses a functional role in TIC properties. The function appears to be mediated, at least in part, by the NTS-induced IL-8 signaling cascade.

thumbnail image

Figure 6. Knockdown of CD133 inhibits the NTS/IL-8/CXCL1 signaling cascade and abolishes TIC properties. (A) Knockdown of CD133 expression led to concomitant down-regulation of RAF-1, NTS, IL-8, and CXCL1 in PLC8024 cells. (B) Compared to nontransduced cells, PLC8024 cells with repressed CD133 generated smaller and fewer spheroids that could not be serially propagated (ANOVA, P = 0.002; multiple comparison with Bonferroni correction, *P = 0.004; **P = 0.003). (C) Suppression of CD133 leaded to reduced engraftment ability, as demonstrated by significantly decreased tumor size (Mann-Whitney test: P = 0.018) and decreased tumor incidence (n = 5 per group). (D) H&E staining of harvested tumors confirmed a primary HCC phenotype. Immunostaining revealed a reduction of IL-8 expression, a reduction of MVD (CD34), a reduction in tumor proliferation (PCNA), and an increase in tumor apoptosis (TUNEL) in the xenograft tumors (200× magnification).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

We previously reported that the transmembrane cell-surface glycoprotein CD133 serves as a marker for liver TICs.5, 6 Our findings are consistent with reports from other research groups who have also found that liver TICs are identified by the CD133 surface phenotype.8-11 In our current study, we compared the gene expression profiles of CD133 subpopulations sorted from HCC cells using genome-wide mRNA expression profiling. Functional analysis of significantly dysregulated genes revealed that most of them were involved in cancer, liver cell proliferation, and liver diseases; these findings suggest that CD133+ liver TICs play a vital role in hepatocarcinogenesis. Interestingly, hepatic stem/progenitor cell markers including CD44, EpCAM, and CK19 were also found to be preferentially up-regulated in the CD133+ TIC subpopulation. This observation is consistent with previous reports demonstrating that CD44 and EpCAM are coexpressed with CD133 and can mark liver TICs.10, 14, 23, 24 In addition to EpCAM and CD44,10, 14 two recent studies by Yang et al.23, 24 also utilized CD90 for the identification of liver TICs. To examine whether our CD133+ liver TIC subpopulation display overlapping expression with other known liver TIC markers, we performed additional flow cytometry analysis in CD133 cells using CD90, EpCAM, and CD44 using one HCC clinical specimen and three HCC cell lines (Huh7, PLC8024, and Hep3B). CD133 expression extensively overlapped with EpCAM in three of the four samples, and to a much smaller extent, also overlapped with some expression of CD44; suggesting that subpopulations marked by CD133/EpCAM and possibly also CD133/CD44 may share similar phenotypes and be controlled under similar regulatory mechanisms. However, we did not find our CD133 liver TICs to express CD90. In fact, CD90 expression was only detected at very low levels (<1%), if any at all, in all the HCC cell lines and clinical specimen examined (Supporting Information Fig. S7).

Past studies have shown that elevated expression of hepatic stem/progenitor cell markers is related to increased angiogenesis, poor prognosis of patients with HCC, and increased vascular endothelial growth factor expression levels.25 HCC is a highly neovascularized tumor, and angiogenesis is an early event in tumorigenesis; thus it is logical to hypothesize that CD133+ liver TICs may have angiogenic properties. In this study, we identified a significantly dysregulated functional network involving IL-8 (along with a number of its members, including NTS, CXCL1, CXCL3, CXCL5, CXCL6, endothelin-1, CD44, and EpCAM) as being preferentially activated in CD133+ liver TICs. Subsequently, we demonstrated that IL-8 plays a prominent role in promoting tumor angiogenesis, growth, and self-renewal in CD133+ liver TICs. These observations were further validated when silencing of IL-8 in CD133+ liver TICs by shRNA-mediated knockdown or a neutralizing antibody resulted in the abolition of these effects.

IL-8 was first identified as a proinflammatory cytokine that promotes neutrophil chemotaxis and activation.26 It functions as a chemoattractant and is also a potent angiogenic factor. Recently, there has been increasing evidence to support a role for IL-8 in regulating oncologic processes in various types of human cancer. In HCC, IL-8 levels have been reported to be elevated during disease progression and to contribute to the development of distant metastasis. Serum IL-8 was also found to be highly elevated in HCC patients and correlated with poor prognosis and survival rates of less than 1 year.27-30 Furthermore, IL-8 was recently shown to play a role in the self-renewal, chemoresistance, and tumorigenesis in TIC populations of lung, breast, colon, and endometrial tumors.31-35

IL-8 is known to act through two high-affinity receptors: CXCR1 and CXCR2. Past studies have suggested that IL-8 promotes tumorigenesis and self-renewal in a paracrine or autocrine manner.18, 31, 35 Here, we found that both CXCR1 and CXCR2 are expressed in CD133+ liver TICs and that exogenous addition of recombinant IL-8 activates the MAPK signaling pathway, as evidenced by increased phosphorylated ERK1/2 and RAF-1 expression. IL-8 treatment also further induced the expression of IL-8 and a closely related family member, CXCL1, which also acts through CXCR1 and CXCR2. These data suggest that IL-8 promotes tumorigenesis in both paracrine and autocrine manners. Most important, we provide evidence that activation of the IL-8 positive feedback loop through MAPK signaling in CD133+ liver TICs is induced, at least in part, by enhanced NTS expression (Fig. 7).

thumbnail image

Figure 7. Cartoon diagram of a potential IL-8/NTS cell signaling in CD133+ liver TICs. CXCR1/2 and NTSR1/2 activation upon IL-8 and NTS binding, respectively, induces MAPK phosphorylation. Activation of MAPK leads to increased production of IL-8 and CXCL1, which in turn can promote angiogenesis by acting on neighboring endothelial cells and also promote tumorigenicity and self-renewal in CD133+ liver TICs in a paracrine and autocrine manner. Transmembrane glycoprotein CD133, apart from serving as a marker for liver TICs, is also functionally critical in conferring tumorigenic and self-renewal abilities through NTS-induced IL-8 signaling.

Download figure to PowerPoint

NTS is a short peptide that was first discovered to act as a neurotransmitter/neuromodulator or an endocrine molecule.36 NTS was later found to be overexpressed in a number of cancer types and to play a role in tumorigenesis.37 To our knowledge, this is the first report to provide data showing that NTS can activate an IL-8 positive feedback loop through MAPK signaling in CD133+ liver TICs and that this cascade is critical in conferring angiogenic, tumorigenic, and stem cell-like properties to these cells. In addition, we also show for the first time that CD133, apart from serving as a liver TIC marker, also plays a functional role in conferring tumorigenic potential to liver TICs, as stable repression of CD133 in CD133+ liver TICs by a lentiviral-based approach resulted in inactivation of the NTS/IL-8/CXCL1/MAPK signaling cascade and attenuation of TIC self-renewal and tumorigenic properties. The characterization of molecular targets downstream of CD133+ in liver TICs in this study should provide not only a better understanding of the mechanisms regulating this specific population of cells but also novel insights that could be used for the development of more effective cancer therapies for the treatment of this disease.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Transcript profiling: Microarray data are available publicly at Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) under the accession numbers of GSE23450 and GSE23451. Author contributions: K.H. Tang: study concept and design, acquisition of data, analysis and interpretation of data, statistical analysis, drafting of the article; S. Ma: study concept and design, acquisition of data, analysis and interpretation of data, critical revision of the article for important intellectual content, study supervision; Y.P. Chan: statistical analysis, technical support; T.K. Lee, P.K. Kwan, and C.M. Tong: technical support; I.O. Ng, K. Man, K.F. To, P.B. Lai, and C.M. Lo: material support; X.Y. Guan and K.W. Chan: critical revision of the article for important intellectual content, obtained funding, study supervision. The authors thank Dr. Davy Lee of Department of Paediatrics and Adolescent Medicine at the University of Hong Kong for expert assistance with the cell sorting facility, and the University of Hong Kong Li Ka Shing Faculty of Medicine Faculty Core Facility.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
HEP_24739_sm_SuppFig1.tif4523KSupporting Information Fig. S1. Representative flow cytometry histogram plot showing the distribution of CD133-expressing cells in unsorted and sorted HCC cells (Huh7 and PLC8024). The red histogram represents the isotype control. CD133 expression of unsorted (black) and sorted CD133+ (green) and CD133- (blue) cells is relative to this isotype control. The purity for the positive fraction ranged from 90-98%, while the negative fraction achieved greater than 97% purity.
HEP_24739_sm_SuppFig2.tif1127KSupporting Information Fig. S2. Immunoblot analyses of the expression of IL-8 receptors (CXCR1 and CXCR2) and NTS receptors (NTSR1 and NTSR2) in a panel of liver cell lines.
HEP_24739_sm_SuppFig3.tif4012KSupporting Information Fig. S3. Compared to non-transduced cells, Huh7 cells with IL-8 suppression displayed reduced engraftment ability, as demonstrated by significantly decreased tumor size (Mann-Whitney test: p = 0.05) and decreased tumor incidence (n = 5 per group).
HEP_24739_sm_SuppFig4.tif4556KSupporting Information Fig. S4. Compared to non-transduced cells, Huh7 cells with IL-8 suppression (clones 30 and 31) generated fewer spheroids that could not be serially propagated (ANOVA: p < 0.001; Multiple comparison with Bonferroni correction: *p < 0.001; **p < 0.001).
HEP_24739_sm_SuppFig5.tif3626KSupporting Information Fig. S5. (A) Increased expression of NTS and CXCL1 in CD133+ liver TICs as compared to CD133- cells in freshly resected clinical HCC specimens (t-test: *p = 0.002; **p = 0.003; ***p = 0.005; ****p = 0.03). (B) qPCR analyses of CD133, IL-8 and CXCL1 expression in sorted CD133 subpopulations (Huh7) in the presence or absence of exogenous NTS (1 µg/mL) or IL-8 (100 ng/mL) for 0-, 1- or 2-hr time points. (C) qPCR analyses of IL-8 and CXCL1 expression in CD133-expressing HCC cell line Hep3B in the presence or absence of exogenous NTS (1 µg/mL) or IL-8 (100 ng/mL) for 0-, 1- or 2-hr time points. (D) A reduction in IL-8 and CXCL1 expression was detected in IL-8-repressed PLC8024 cells (clones 30 and 31) when compared to non-transduced cells (ANOVA, p = 0.002; Multiple comparison with Bonferroni correction, p = 0.009).
HEP_24739_sm_SuppFig6.tif7375KSupporting Information Fig. S6. Significant reduction of CD133 expression in CD133 shRNA knockdown clones 44 and 47, compared with non-transduced cells, and in PLC8024 CD133+ liver TICs, as detected by flow cytometry.
HEP_24739_sm_SuppFig7.tif4213KSupporting Information Fig. S7. Dual-color flow cytometry analyses of CD133 with CD90, EpCAM and CD44 in HCC cell lines (Huh7, PLC8024, Hep3B) and HCC clinical specimen (Patient 72).
HEP_24739_sm_SuppInfo.doc38KSupporting Information Table S1. Primer sequences for qPCR analysis.
HEP_24739_sm_SuppTab1.doc37KSupporting Information Table S2. List of commonly dysregulated genes with a relative fold change greater than 2 in CD133+ vs. CD133- cells in the HCC cell lines Huh7 and PLC8024 (MAS5 and RMA).
HEP_24739_sm_SuppTab2.doc217KSupporting Information Table S3. Validation of commonly dysregulated known hepatic stem cells / TIC markers and genes involved in the IL-8 pathway in sorted CD133 subpopulations (Huh7 and PLC8024) by qPCR analysis.
HEP_24739_sm_SuppTab3.doc39KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.