Characterization of in vivo chemoresistant human hepatocellular carcinoma cells with transendothelial differentiation capacities
© Marfels et al.; licensee BioMed Central Ltd. 2013
Received: 4 December 2012
Accepted: 19 March 2013
Published: 2 April 2013
Chemotherapeutic treatment of hepatocellular carcinoma often leads to chemoresistance during therapy or upon relapse of tumors. For the development of better treatments a better understanding of biochemical changes in the resistant tumors is needed. In this study, we focus on the characterization of in vivo chemoresistant human hepatocellular carcinoma HUH-REISO established from a metronomically cyclophosphamide (CPA) treated HUH7 xenograft model.
SCID mice bearing subcutaneous HUH7 tumors were treated i.p. with 75 mg/kg CPA every six days. Tumors were evaluated by immunohistochemistry, a functional blood-flow Hoechst dye assay, and qRT-PCR for ALDH-1, Notch-1, Notch-3, HES-1, Thy-1, Oct-4, Sox-2 and Nanog mRNA levels. Cell lines of these tumors were analyzed by qRT-PCR and in endothelial transdifferentiation studies on matrigel.
HUH-REISO cells, although slightly more sensitive against activated CPA in vitro than parental HUH-7 cells, fully retained their in vivo CPA chemoresistance upon xenografting into SCID mice. Histochemical analysis of HUH-REISO tumors in comparison to parental HUH-7 cells and passaged HUH-PAS cells (in vivo passaged without chemotherapeutic pressure) revealed significant changes in host vascularization of tumors and especially in expression of the tumor-derived human endothelial marker gene PECAM-1/CD31 in HUH-REISO. In transdifferentiation studies with limited oxygen and metabolite diffusion, followed by a matrigel assay, only the chemoresistant HUH-REISO cells exhibited tube formation potential and expression of human endothelial markers ICAM-2 and PECAM-1/CD31. A comparative study on stemness and plasticity markers revealed upregulation of Thy-1, Oct-4, Sox-2 and Nanog in resistant xenografts. Under therapeutic pressure by CPA, tumors of HUH-PAS and HUH-REISO displayed regulations in Notch-1 and Notch-3 expression.
Chemoresistance of HUH-REISO was not manifested under standard in vitro but under in vivo conditions. HUH-REISO cells showed increased pluripotent capacities and the ability of transdifferentiation to endothelial like cells in vitro and in vivo. These cells expressed typical endothelial surface marker and functionality. Although the mechanism behind chemoresistance of HUH-REISO and involvement of plasticity remains to be clarified, we hypothesize that the observed Notch regulations and upregulation of stemness genes in resistant xenografts are involved in the observed cell plasticity.
KeywordsCyclophosphamide Chemoresistance Tumor plasticity Cancer stemness
Hepatocellular carcinoma (HCC) is the fifth most common malignancy worldwide . Moreover, its incidence increases due to hepatitis B and C viral infections. Therefore, HCC is in the focus of several treatment studies. Compared to other solid tumors, HCC is characterized by high levels of vascularization. The status of angiogenesis correlates with cancer progression and prognosis. Therefore, antiangiogenic strategies are suggested for treatment of HCC  due to survival advantages, as revealed in recent studies [3, 4]. In addition, the usage of preclinical antiangiogenic, metronomic regimen of cyclophosphamide (CPA) revealed encouraging results in terms of tumor growth suppression and survival in an in vivo rat model of hepatocellular carcinoma . The metronomic treatment regimen is characterized by significantly reduced side effects, compared to conventional maximum tolerated chemotherapy administration and by antitumoral activity in respect to its antiangiogenic properties. The metronomic treatment regimes target preferentially genetic stable tumor vessel endothelial cells and thus, the development of resistance against the therapy should be avoided [6, 7]. However, several studies point towards the induction of in vivo chemoresistance mechanisms that let tumors escape from metronomic CPA therapy [8–10].
In this study, we investigated changes in transcription factors, controlling plasticity and stemness of tumor cells in an in vivo chemoresistance HCC xenograft mouse model. Resistant HCC xenografts were generated by metronomically scheduled CPA treatment in SCID mice, resulting in resistant tumor outgrowth after an initial chemoresponsive phase of 10 weeks. Histological analysis revealed significant changes in tissue organization and blood flow. Re-xenografted tumors from HUH-REISO cell culture manifested immediate chemoresistance, lacking an initial response phase. In order to detect gene expression associated with the chemoresistance and its development, expression levels of Notch-1 and downstream HES-1, Notch-3, Thy-1, Oct-4, Sox-2 and Nanog were determined in in vivo passaged control xenografts and in their resistant counterparts with and without therapeutic pressure. Furthermore, several aspects of cell differentiation were traceable in specialized in vitro models, mimicking features of environmental properties of solid tumors.
Cell culture media, antibiotics, fetal bovine serum (FBS) and trypsin/EDTA solution were purchased from Invitrogen GmbH (Karlsruhe, Germany). Human hepatoma cells (HUH-7) (JCRB0403) were cultured in a mixture of Ham’s-F12 and Dulbecco’s modified Eagle’s medium (DMEM) in a ratio of 1:1 supplemented with 10% FBS. Cells were grown at 37°C in 5% CO2 in a humidified atmosphere. HUH-7 cells were cultured without antibiotics for at least 3–4 passages before tumor cell implantation and were harvested just as reaching approx. 70% confluency.
In vivoanimal model
Male SCID mice (CB17/lcr-PrkdcSCID/Crl) (8–10 weeks) were housed in individually vented cages under specific pathogen free conditions with a 12 h day/night cycle and with food and water ad libitum. HUH-7 cells were cultured as described above. The number of 106 HUH-7 cells in 100 μl PBS was injected subcutaneously with a 25 G needle (Braun, Melsungen, Germany) into the flank of SCID mice. The animals were checked regularly for tumor progression. The moment that tumor volume reached the size of at least 10 mm3, tumor progression was monitored using a digital measuring slide (Digi-Met, Preisser, Gammertingen). Each measurement consisted of three diameters, length (a), width (b), and height (c). Tumor volume was calculated by the formula a × b × c × π/6 (with a, b and c indicating the three diameters and π/6 as correction factor for tumor shape). Tumor volume doubling time was calculated with TVDT = ln2x(t2 − t1)/ln[V(t2)/V(t1)]. All animal experiments were performed with 6 animals per group. All animal procedures were approved and controlled by animal experiments ethical committee of Regierung von Oberbayern, District Government of Upper Bavaria, Germany and carried out according to the guidelines of the German law of protection of animal life.
Isolation of tumor cells
Cyclophosphamide (CPA) (Sigma, Taufkirchen, Germany) was solved in PBS (10 mg/ml) and applied intraperitoneally. 75 mg/kg CPA solution was administered with a 25 G needle (Braun, Melsungen, Germany). The application of the CPA solution was carried out every 6 days. A single dose of each application was based on animal body weight. Toleration of CPA treatment was monitored by regular measurement of body weight. The vehicle group (PBS) and the drug treatment group (CPA dissolved in PBS) were housed separately.
HE stain of tumors
Cryosections were fixed with 4% paraformaldehyde and stained with haematoxilin (Sigma, St. Louis, USA) for 30 min. After washing with PBS and aqua dest., sections were incubated with eosin (1:100 in aqua dest.) (Sigma, St. Louis, USA) for 4 min. Afterwards, sections were washed with aqua dest., embedded with PBS and analyzed by transmission light microscopy at the Axiovert 200 microscope (Zeiss, Jena, Germany).
For immunohistology, tumors were embedded in tissue freezing medium (Jung, Nussloch, Germany). They were cut into sections of 5–10 μm thickness with a cryomicrotome (Leica CM 3050 s, Wetzlar, Germany) at − 20°C. Sections were transferred to a microscope slide, tissue freezing medium was removed and tissue was fixed with 4% paraformaldehyde (in PBS). Afterwards, sections were rehydrated and washed with blocking solution (PBS containing 5% FBS) prior to antibody incubation. Antibodies, which were used for the stains, are listed in Additional file 2: Table S2. All primary antibodies were diluted 1:200 in blocking solution. After incubation for 12 h at 4°C in humidified atmosphere, sections were washed repeatedly with blocking solution followed by secondary antibody staining. Secondary antibodies were diluted 1:400 in blocking solution and sections were incubated for 2 h at room temperature in humidified atmosphere. Sections were washed with blocking solution repeatedly, before fluorescence analysis at the Axiovert 200 microscope (Zeiss, Jena, Germany) using appropriate filter sets.
Agarose overlay method
Cells were seeded in 6-well plates 24 h before addition of the agarose overlay. Culture medium was removed and replaced with 1 ml medium containing 0.6% (w/v) agarose (Invitrogen, Carlsbad, USA). The agarose-containing medium was obtained by stepwise dilution of complete medium with melted agarose (5% agarose in medium without supplementations, w/v). Before applying the agarose-containing medium to the seeded cells, the medium was allowed to cool to 37°C. After solidification, 2 ml of complete culture medium without agarose was added to the cells.
Spheroid growth in agarose gel
To avoid cell growth in a two dimensional way, cells were seeded into an agarose gel. Therefore, cells were harvested as soon as reaching approx. 70% confluency. The gel was prepared by mixing low melting agarose from Cambrex BioScience (Rockland, USA) with medium in a percentage of 10% (w/v). After autoclaving, the gel was diluted with medium to 1.2%. Afterwards gel and single cell suspension were mixed and 5,000 cells in 1 ml of 0.6% gel were seeded in a 24 well plate (TPP, Trasadingen, Switzerland). Grown spheroids were counted after 42 days. Pictures were taken with an Infinity 2 camera and Infinity capture software (both: Lumenera corporation, Ottawa, Canada).
In vitromatrigel angiogenesis assay
Matrigel was purchased from BD (Franklin Lakes, USA). The coating procedure was done as described in the BD guidelines for thin gel layers. Matrigel was thawed over night at 4°C. After short homogenization of the gel by pipetting with cooled tips, 10 μl of matrigel were added to each well of a μ-slide angiogenesis uncoated chamber (ibidi, München, Germany). The slide was incubated for 30 minutes at 37°C and afterwards cells were seeded on the gel layer in an amount of 50,000 cells in 50 μl of cell culture medium per well. Cells were grown at 37°C in 5% CO2 in a humidified atmosphere and observed with an axiovert 200 microscope (Zeiss, Jena, Germany). Pictures were taken after 24 hours with an Infinity 2 camera and Infinity capture software (both: Lumenera corporation, Ottawa, Canada). Image analysis was performed with the analysis software from S.Core (Hoehenkirchen, Germany).
Marker gene analysis was done by quantitative polymerase chain reaction. Therefore, pairs of primer were designed with the universal probe library of Roche and purchased from Matabion (Martinsried, Germany). All pairs of primer, which were used for qRT-PCR, are listed in Additional file 3: Table S3. Total RNA of in vivo and in vitro samples were purified by using two different methods. For the in vitro angiogenesis assay samples, 106 cells were seeded on a thin layer of matrigel (BD, Franklin Lakes, USA) in a six well plate. Total RNA of cells from one well was purified by phenol-chloroform extraction using peqGOLD TriFast kit (Peqlab, Erlangen, Germany). Preparation of in vivo RNA-Samples was done with a NucleoSpin RNA II kit (Macherey-Nagel, Düren, Germany) using 30 mg of tumor tissue. 5 μg of purified RNA were applied for cDNA production. For each qRT-PCR 25 ng cDNA were used for amplification and all samples were measured in duplicates. After an activation cycle with 90°C for 10 Minutes, 45 cycles were performed with a 10 seconds denaturation step at 95°C, 30 seconds of annealing at 60°C and polymerase extension for 1 second at 72°C. The PCR runs were performed with Light Cycler 480 (Roche, Mannheim, Germany). For detection, the corresponding probe out of the universal probe library (Roche, Mannheim, Germany) was applied. Fold changes were calculated with the ΔΔCt- method. As GAPDH and Act-B showed stable unregulated expression status over all tested tumor samples in initial experiments, GAPDH was used as reference gene for measurements of Oct-4, Thy-1, and ALDH-1 and Act-B for measurements of HES-1, Notch-3, Notch-1, Nanog, Sox-2, PECAM-1/CD31, and ICAM-2. The ready to use Universal Probe Library (UPL) reference gene assays for GAPDH and ACT-B (Roche Diagnostics, Mannheim, Germany) were applied.
In vitroCPA sensitivity
In vitro sensitivity was assayed by measuring the DNA content of a cell population composed of 25% of CPA activating X39 cells mixed with either HUH-wt, or HUH-REISO cells, followed by CPA treatment. Generation of CYP2B1 transgene expressing X39 cells is described previously . In total, 1500 cells per well were plated in 48-well plates. Twenty-four hours after seeding, the culture medium was removed and replaced by either 200 μl of fresh medium or by fresh cell culture medium containing CPA at the indicated concentrations. Treated cells and controls were incubated for 5 days in a humidified atmosphere containing 5% CO2 at 37°C. DNA contents were assayed after Hoechst 33258 incorporation, as previously described . Briefly, cells were lysed with Millipore water followed by a freeze–thaw cycle. Cell lysis buffer (1 mM Tris-EDTA, pH 7.4, 200 mM NaCl) containing 0.2 ng/ml Hoechst 33258 was applied to each well, followed by another freeze–thaw cycle. The DNA content was measured by quantifying fluorescence with a plate reader (Tecan, Grödig, Austria) equipped with filters for excitation at 360 nm and emission at 465 nm. Relative DNA content was calculated using the ratio of DNA content treated /DNA content untreated cell culture.
U-Test (Mann–Whitney) analysis was performed with WinStat for Exel to proof statistical significance in all cases. * stands for p≤0.05, ** for p≤0.01.
HUH7 tumors under metronomic CPA therapy in vivo
Male and female SCID mice bearing subcutaneously implanted human HUH7 tumors were treated with metronomically scheduled CPA (75 mg/kg every 6 days). CPA treatment was started on day 12 after tumor cell implantation, just as tumors reached an average volume of 32 mm3. Metronomically scheduled CPA treatment resulted in a significant tumor growth delay. The tumor volume of treated mice was constant at around 100 mm3 up to day 75 after tumor cell implantation, whereas tumors in the control group exhibited a tumor volume doubling time of 2.5 days. Around day 75 after tumor cell implantation, tumor volume began to increase in the CPA treated group, with a tumor doubling time of 3.5 days, despite ongoing treatment (Figure 1A). Metronomically scheduled CPA therapy was well tolerated, indicated by a constant animal body weight up to day 85 after tumor cell implantation (data not shown). Further CPA treatment resulted in significant loss of body weight, observed in all CPA treated animals (data not shown). Treatment was stopped and the mice were sacrificed as soon as the average body weight loss reached 20%. At this therapy endpoint, tumors were collected and subjected to macroscopical and histological analyses. Furthermore, tumor cells were extracted from viable tumor tissue for characterization and cell culture experiments. To establish appropriate control cells (HUH-PAS), HUH7 were grown in male and female SCID mice, without exposure to therapy. As soon as tumors reached about 300 to 400 mm3, viable cells were extracted from tumor tissue. Reisolated cells (HUH-REISO and HUH-PAS) exhibited the same morphology in comparison to the parental HUH7 cells (HUH-wt) and were identified by human EGF-Receptor staining for their human origin (Additional file 4: Figure S1).
Influence of CPA therapy on tumor macroscopic appearance, tumor histology and functional blood flow
Immunohistochemical analysis of vascular structures in xenografts
Regarding plasticity aspects, tumor tissue was stained for human, besides mouse, endothelial specific marker PECAM-1/CD31 (hPECAM-1 and mPECAM-1). Counter stain of cell nuclei with DAPI can be seen in Figure 3C (HUH-wt), 3G (HUH-PAS) and 3K (HUH-REISO). Control stains revealed no hPECAM-1 signal for the parental HUH7 xenografts (Figure 3E and F), whereas mPECAM-1 (mCD31) positive cells showed a large network of vessel (Figure 3D, highlighted with black arrows). Most interestingly, Figure 3M and merged 3N showed hPECAM-1 (hCD31) positive structures (highlighted with white arrows) in reimplanted resistant xenografts (HUH-REISO) in close neighbourhood to murine vascular structures (Figure 3L), indicating HCC plasticity towards the endothelial lineage. Control tumors of HUH-PAS, which were treated twice with CPA, showed rare positive signals for mPECAM-1 (Figure 3H) and no positive signal for hPECAM-1 (Figure 3I). Those HUH-PAS tumors had to be treated late, when they had reached an average volume of 254 mm3. At an earlier starting point of therapy of these chemoresponsive cells there would not be enough tumor material left for reliable analysis (Figure 1B).
No evidence of acquired resistance in vitro
Resistance of reimplanted tumors towards metronomic CPA therapy in vivo
Reisolated HUH-REISO tumor cells and reisolated in vivo passaged (HUH-PAS) cells were implanted in the flank of SCID mice. On day 10 after tumor cell implantation, just as average tumor volume reached 14 mm3, mice were subjected to CPA treatment (75mg/kg, every 6 days). Tumor volume and body weight were measured regularly during the treatment. No growth retardation effect was detectable for xenografts established from HUH-REISO cells, in contrast to blocked growth of xenografts established from in vivo passaged cells (HUH-PAS) (Figure 1B). Resistant xenografts exhibited an average tumor volume doubling time of approximately 4.5 days. Metronomically scheduled CPA was again well tolerated, indicated by no significant loss in body weight (data not shown).
Regulation of ALDH-1 expression in response to CPA therapy in vivo
Expression profiles of Thy-1, Oct-4, Sox-2 and Nanog in vivo
Expression profiles of Notch-1, Notch-3 and HES-1
Anchorage independent growth of HUH-wt, HUH-PAS and HUH-REISO spheroids
Endothelial transdifferentiation in vitro
Quantification of tube formation in matrigel was performed via software based analysis (Additional file 6: Figure S2 and Figure 9G-J). Comparison of HUH-wt, HUH-PAS, and HUH-REISO (all pre-cultured by agarose overlay) revealed a far higher number of branching points (Figure 9G) and tubes (Figure 9H), a very extended length of skeleton (Figure 9I), and a decreased amount of confluent areas without tube formation (Figure 9J) for HUH-REISO cells.
Furthermore, expression levels of endothelial markers (CD31/PECAM-1, ICAM-2, VEGFR2, VE-cadherin and vWF) of agarose overlay HUH-REISO cells, showing positive tube formation, were compared to the corresponding standard culture HUH-REISO (Figure 9K-L). Expression levels were determined after the matrigel assay. In agarose overlay HUH-REISO cells, ICAM-2 (p=0.009) (Figure 9K), as well as CD31/PECAM-1 (p=0.028) (Figure 9L) were significantly upregulated. Expression of the other endothelial markers was not detectable in HUH-REISO under any culture condition (data not shown).
In the present study, acquired in vivo chemoresistance against metronomic cyclophosphamide (CPA) treatment was studied in a human hepatocellular carcinoma HUH7 xenograft mouse model. During treatment, a two phase development of tumor progression was observable: In the beginning of treatment (response phase), tumor progression was significantly decreased, indicated by constant tumor volume for about 75 days. In the following, second phase (escape phase), tumor volume increased with a tumor volume doubling time of 3.5 days, despite ongoing therapeutic intervention (Figure 1A). Viable tumor cells were extracted from resistant tumors (HUH-REISO), whereas control cells (HUH-PAS) were obtained from in vivo passaging HUH7 tumor cells without CPA treatment. Subsequently, HUH-PAS and HUH-REISO were characterized and identified in terms of cell morphology and representative human epidermal growth factor (EGF receptor) expression for their human origin (Additional file 4: Figure S1).
Interestingly, in vivo chemoresistant HUH-REISO did not manifest their drug resistant phenotype in a two-dimensional monolayer culture in presence of in situ activated CPA (Figure 4). In addition, significant changes in macroscopic appearance and tumor tissue organization of chemoresistant tumors indicate resistance mechanisms, which were only gainful in the in vivo situation. However, an endogenous imprinted component for in vivo chemoresistance was obvious, as the chemoresistant phenotype of isolated tumor cells was immediately manifested again after reimplantation and reapplied chemotherapy. In the reimplantation experiment, chemoresistance was manifested lacking the response phase. In contrast, HUH-PAS, which were only adjusted to the in vivo environment but not to the treatment, remained sensitive (Figure 1B) in a response phase. The qRT-PCR assay on in vivo samples revealed no significant difference in basal expression of the detoxification enzyme ALDH-1 (Figure 5), which converts aldophosphamide into carboxyphosphamide. ALDH-1 expression was detected to be significant upregulated in vivo during therapeutic pressure. However, the extent of upregulation was not significantly different in chemosensitive and chemoresistant tumors. The detected chemosensitivity of HUH-PAS exclude resistance by unspecific selection processes in vivo, which might change cellular properties independently of therapeutic pressure [12, 13].
As metronomic CPA treatment is known to suppress tumor angiogenesis , functional blood flow analysis was performed, revealing blood flow in a part of the newly formed cavities of chemoresistant tumors (Figure 2F). In contrast to untreated animals, blood flow was not obligatory colocalized with immunohistochemical detected laminin and mouse PECAM-1/CD31 signal (Figure 3B). Further analysis on human vessel markers revealed within the tumor the presence of cells expressing human PECAM-1/CD31 (Figure 3M/N), indicating plasticity of the tumor cells and initiation of differentiation towards the endothelial lineage. This differentiation was observed only in HUH-REISO. In HUH-PAS the CPA therapy led to a deletion of murine vessels (Figure 3H), while the capacity of building new human vessels was not observed. As HUH-REISO showed no diminished blood-flow (Additional file 5: Figure S3) and also a supplementation of destroyed murine vessels by human endothelial cells, resistance is partly caused by a better blood-supplementation due to the higher plasticity of HUH-REISO. Better blood-supplementation leading to higher drug delivery into the tumor presumably requires also other detoxification mechanisms for the resistance. The higher levels of ALDH-1 in HUH-REISO (6×CPA, Figure 5) are consistent with this hypothesis. Therefore, further changes in the microenvironment and proteome of the resistant tumor cells remain to be examined in future work.
In line with literature , the process of reprogramming was connected to induction of Notch-1 expression, as a first response to chemotherapeutic pressure (Figure 7A). Induction of Notch-1 leads to the expression of GIMAP5 , which is antiapoptotic and has potential to save cells. ALDH-1 was simultaneously induced (Figure 5) and led to cell survival under therapeutic pressure. Furthermore, induction of HES-1, which is a downstream target of Notch-1, was observable (Figure 7B). HES-1 belongs to the basic helix loop helix family of transcription factors and is described by Kageyama et al. as a crucial factor in many tissues to maintain the status of pluripotency . Notch-1 and HES-1 were only upregulated in the response phase of therapy, indicating an initial key factor in this process. In the escape phase of already chemoresistant tumors, Notch-1 was steadily regulated under chemotherapeutic pressure, indicating that Notch-1 was essential for the preservation of the reprogrammed status, once Thy-1, Oct-4, Sox-2 and Nanog provides pluripotency and self-renewal.
Surprisingly, after two treatments with CPA, expression levels of Thy-1, Oct-4, Sox-2 and Nanog were decreased in HUH-REISO (Figure 6A-D), indicating a process of differentiation, which obviously antagonize further enrichment of this subpopulation at this point of treatment. This correlated with the induction of Notch-3 expression. The Notch-3 pathway is described as an important signalling pathway in the development of vascularization. N. Lawson et al. [21, 22] showed the important role of Notch-3 in arterial cell fate during blood vessel development. However, after prolonged CPA treatment, expression levels of all pluripotency markers recovered. The initial expression profile of Notch-3 and the profile of Thy-1, Oct-4, Sox-2 and Nanog pointed to an adaptation process, as a response to acute chemotherapeutic pressure. Obviously, differentiation was repressed again after adaptation and the balance returned to enrichment of pluripotent cells. An important step to the in vivo chemoresistance in our model was therefore the development of a pool of cells, which reveal increased levels of pluripotency markers and obviously help to overcome chemotherapeutic pressure by initiation of differentiation (Figure 10). In contrast to chemoresistant tumors, the CPA therapy in chemosensitive tumors led to the induction of differentiation processes and cell death, which in the beginning of treatment nearly flushed out the whole pool of pluripotent tumor cells. Recovering of this pool took long time (growth delay phase Figure 1B) and led, beside other mechanisms, in the end to resistance (Figure 1A).
This differentiation process, resulting in the reorganization of tumor tissue, helped tumors to escape the metronomic treatment. To investigate the trigger, which caused differentiation, HUH-wt, HUH-PAS, and HUH-REISO cells, were pre-cultured in an in vitro cell culture system, which mimics several features of solid tumors under therapeutic pressure. In a following matrigel tube formation assay, usually performed for angiogenesis studies , HUH-REISO cells showed tube formation capacity (Figure 9F), whereas cells pre-cultured under conventional conditions showed no functionality (Figure 9C). Importantly, HUH-wt and HUH-PAS revealed no tube formation potential, independently of pre-culture conditions. Obviously, initiation of differentiation only took place in an environment, which was characterized by limitation of oxygen supply and simultaneous diffusion limitation of paracrine and autocrine factors. In recent studies, the phenomena of tumor cell tube formation on matrigel was detected e.g. for glioblastoma, breast cancer  and multiple myeloma . In several studies the feature of tube formation was independent from the expression of endothelial markers (vascular mimicry) . However, in the case of HUH-REISO cells, tube formation was associated with induction of the endothelial genes PECAM-1/CD31 and ICAM-2 (Figure 9K/L). In contrast, other endothelial marker genes as vWF, VEGFR2 and VE-Cadherin (data not shown) were neither expressed, nor regulated. Expression of endothelial genes was published by Bussolati et al. and Bruno et al., who detected induction of endothelial marker expression, derived from tumor initiating cells (tumor stem cells) by treatment via VEGF  or after xenografting . Apparently, HUH-REISO cells, or at least a subpopulation acquired a reprogrammed status characterized by enormous plasticity. This pool of cells reacted on environmental requirements by initiation of differentiation in specialized cells to maintain a balanced tumor microenvironment, ensured sufficient oxygen and nutrient supply and finally counteracted metronomic therapy.
In conclusion, studying the escape mechanism towards metronomically applied CPA therapy in a HCC xenograft model revealed a multistep process, going beyond unspecific enrichment of a certain subpopulation by the in vivo tumor environment. First step of this process was the enrichment of detoxicating enzyme ALDH-1 in combination with upregulation of the Notch1-pathway and its protective downstream effector proteins. Second step was a selection process of pluripotent, stem cell marker expressing cells. These cells increased endothelial transdifferentiation in vivo and in vitro. Functionality of such endothelial-like cells helped resistant tumors to overcome anti-angiogenic therapy and was the most important finding of our study. Further examinations how modification of Notch signalling may impact cell plasticity and differentiation remain to be explored in future.
These studies were supported by the DFG Cluster of Excellence Nanosystems Initiative Munich (NIM). We are very grateful to Monika Fuchs and Christina Troiber for carefully reading and correcting the manuscript text.
- Farazi PA, DePinho RA: Hepatocellular carcinoma pathogenesis: from genes to environment. Nat Rev Cancer. 2006, 6: 674-687. 10.1038/nrc1934.View ArticlePubMedGoogle Scholar
- Pang R, Poon RT: Angiogenesis and antiangiogenic therapy in hepatocellular carcinoma. Cancer Lett. 2006, 242: 151-167. 10.1016/j.canlet.2006.01.008.View ArticlePubMedGoogle Scholar
- Finn RS, Zhu AX: Targeting angiogenesis in hepatocellular carcinoma: focus on VEGF and bevacizumab. Expert Rev Anticancer Ther. 2009, 9: 503-509. 10.1586/era.09.6.View ArticlePubMedGoogle Scholar
- Llovet JM, Ricci S, Mazzaferro V, Hilgard P, Gane E, Blanc JF, de Oliveira AC, Santoro A, Raoul JL, Forner A, Schwartz M, Porta C, Zeuzem S, Bolondi L, Greten TF, Galle PR, Seitz JF, Borbath I, Haussinger D, Giannaris T, Shan M, Moscovici M, Voliotis D, Bruix J: Sorafenib in advanced hepatocellular carcinoma. N Engl J Med. 2008, 359: 378-390. 10.1056/NEJMoa0708857.View ArticlePubMedGoogle Scholar
- Park ST, Jang JW, Kim GD, Park JA, Hur W, Woo HY, Kim JD, Kwon JH, Yoo CR, Bae SH, Choi JY, Yoon SK: Beneficial effect of metronomic chemotherapy on tumor suppression and survival in a rat model of hepatocellular carcinoma with liver cirrhosis. Cancer Chemother Pharmacol. 2009, 65: 1029-1037.View ArticlePubMedGoogle Scholar
- Browder T, Butterfield CE, Kraling BM, Shi B, Marshall B, O’Reilly MS, Folkman J: Antiangiogenic scheduling of chemotherapy improves efficacy against experimental drug-resistant cancer. Cancer Res. 2000, 60: 1878-1886.PubMedGoogle Scholar
- Folkman J: Angiogenesis. Annu Rev Med. 2006, 57: 1-18. 10.1146/annurev.med.57.121304.131306.View ArticlePubMedGoogle Scholar
- Broxterman HJ, Lankelma J, Hoekman K: Resistance to cytotoxic and anti-angiogenic anticancer agents: similarities and differences. Drug Resist Updat. 2003, 6: 111-127. 10.1016/S1368-7646(03)00026-8.View ArticlePubMedGoogle Scholar
- Emmenegger U, Shaked Y, Man S, Bocci G, Spasojevic I, Francia G, Kouri A, Coke R, Cruz-Munoz W, Ludeman SM, Colvin OM, Kerbel RS: Pharmacodynamic and pharmacokinetic study of chronic low-dose metronomic cyclophosphamide therapy in mice. Mol Cancer Ther. 2007, 6: 2280-2289. 10.1158/1535-7163.MCT-07-0181.View ArticlePubMedGoogle Scholar
- Thoenes L, Hoehn M, Kashirin R, Ogris M, Arnold GJ, Wagner E, Guenther M: In vivo chemoresistance of prostate cancer in metronomic cyclophosphamide therapy. J Proteomics. 2010, 73: 1342-1354. 10.1016/j.jprot.2010.02.019.View ArticlePubMedGoogle Scholar
- Gunther M, Waxman DJ, Wagner E, Ogris M: Effects of hypoxia and limited diffusion in tumor cell microenvironment on bystander effect of P450 prodrug therapy. Cancer Gene Ther. 2006, 13: 771-779. 10.1038/sj.cgt.7700955.View ArticlePubMedGoogle Scholar
- Abe T, Tada M, Shinohara N, Okada F, Itoh T, Hamada J, Harabayashi T, Chen Q, Moriuchi T, Nonomura K: Establishment and characterization of human urothelial cancer xenografts in severe combined immunodeficient mice. Int J Urol. 2006, 13: 47-57. 10.1111/j.1442-2042.2006.01220.x.View ArticlePubMedGoogle Scholar
- Mueller MM, Peter W, Mappes M, Huelsen A, Steinbauer H, Boukamp P, Vaccariello M, Garlick J, Fusenig NE: Tumor progression of skin carcinoma cells in vivo promoted by clonal selection, mutagenesis, and autocrine growth regulation by granulocyte colony-stimulating factor and granulocyte-macrophage colony-stimulating factor. Am J Pathol. 2001, 159: 1567-1579. 10.1016/S0002-9440(10)62541-2.View ArticlePubMedPubMed CentralGoogle Scholar
- Gu G, Yuan J, Wills M, Kasper S: Prostate cancer cells with stem cell characteristics reconstitute the original human tumor in vivo. Cancer Res. 2007, 67: 4807-4815. 10.1158/0008-5472.CAN-06-4608.View ArticlePubMedGoogle Scholar
- Zhang S, Balch C, Chan MW, Lai HC, Matei D, Schilder JM, Yan PS, Huang TH, Nephew KP: Identification and characterization of ovarian cancer-initiating cells from primary human tumors. Cancer Res. 2008, 68: 4311-4320. 10.1158/0008-5472.CAN-08-0364.View ArticlePubMedPubMed CentralGoogle Scholar
- Hay DC, Sutherland L, Clark J, Burdon T: Oct-4 knockdown induces similar patterns of endoderm and trophoblast differentiation markers in human and mouse embryonic stem cells. Stem Cells. 2004, 22: 225-235. 10.1634/stemcells.22-2-225.View ArticlePubMedGoogle Scholar
- Boyer LA, Lee TI, Cole MF, Johnstone SE, Levine SS, Zucker JP, Guenther MG, Kumar RM, Murray HL, Jenner RG, Gifford DK, Melton DA, Jaenisch R, Young RA: Core transcriptional regulatory circuitry in human embryonic stem cells. Cell. 2005, 122: 947-956. 10.1016/j.cell.2005.08.020.View ArticlePubMedPubMed CentralGoogle Scholar
- Lu JW, Chang JG, Yeh KT, Chen RM, Tsai JJ, Hu RM: Overexpression of Thy1/CD90 in human hepatocellular carcinoma is associated with HBV infection and poor prognosis. Acta Histochem. 2011, 113: 833-838. 10.1016/j.acthis.2011.01.001.View ArticlePubMedGoogle Scholar
- Chadwick N, Zeef L, Portillo V, Fennessy C, Warrander F, Hoyle S, Buckle AM: Identification of novel Notch target genes in T cell leukaemia. Mol Cancer. 2009, 8: 35-10.1186/1476-4598-8-35.View ArticlePubMedPubMed CentralGoogle Scholar
- Kageyama R, Ohtsuka T, Tomita K: The bHLH gene Hes1 regulates differentiation of multiple cell types. Mol Cells. 2000, 10: 1-7. 10.1007/s10059-000-0001-0.View ArticlePubMedGoogle Scholar
- Lawson ND, Scheer N, Pham VN, Kim CH, Chitnis AB, Campos-Ortega JA, Weinstein BM: Notch signaling is required for arterial-venous differentiation during embryonic vascular development. Development. 2001, 128: 3675-3683.PubMedGoogle Scholar
- Lawson ND, Vogel AM, Weinstein BM: Sonic hedgehog and vascular endothelial growth factor act upstream of the Notch pathway during arterial endothelial differentiation. Dev Cell. 2002, 3: 127-136. 10.1016/S1534-5807(02)00198-3.View ArticlePubMedGoogle Scholar
- Grant DS, Tashiro K, Segui-Real B, Yamada Y, Martin GR, Kleinman HK: Two different laminin domains mediate the differentiation of human endothelial cells into capillary-like structures in vitro. Cell. 1989, 58: 933-943. 10.1016/0092-8674(89)90945-8.View ArticlePubMedGoogle Scholar
- Basu GD, Liang WS, Stephan DA, Wegener LT, Conley CR, Pockaj BA, Mukherjee P: A novel role for cyclooxygenase-2 in regulating vascular channel formation by human breast cancer cells. Breast Cancer Res. 2006, 8: R69-10.1186/bcr1626.View ArticlePubMedPubMed CentralGoogle Scholar
- Scavelli C, Nico B, Cirulli T, Ria R, Di Pietro G, Mangieri D, Bacigalupo A, Mangialardi G, Coluccia AM, Caravita T, Molica S, Ribatti D, Dammacco F, Vacca A: Vasculogenic mimicry by bone marrow macrophages in patients with multiple myeloma. Oncogene. 2008, 27: 663-674. 10.1038/sj.onc.1210691.View ArticlePubMedGoogle Scholar
- El Hallani S, Boisselier B, Peglion F, Rousseau A, Colin C, Idbaih A, Marie Y, Mokhtari K, Thomas JL, Eichmann A, Delattre JY, Maniotis AJ, Sanson M: A new alternative mechanism in glioblastoma vascularization: tubular vasculogenic mimicry. Brain. 2010, 133: 973-982. 10.1093/brain/awq044.View ArticlePubMedPubMed CentralGoogle Scholar
- Bussolati B, Grange C, Sapino A, Camussi G: Endothelial cell differentiation of human breast tumour stem/progenitor cells. J Cell Mol Med. 2009, 13: 309-319. 10.1111/j.1582-4934.2008.00338.x.View ArticlePubMedGoogle Scholar
- Bruno S, Bussolati B, Grange C, Collino F, Graziano ME, Ferrando U, Camussi G: CD133+ renal progenitor cells contribute to tumor angiogenesis. Am J Pathol. 2006, 169: 2223-2235. 10.2353/ajpath.2006.060498.View ArticlePubMedPubMed CentralGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2407/13/176/prepub
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.