Clinical studies have shown that patients with highly hypoxic primary tumors may have poor disease-free and overall survival rates. Studies of experimental tumors have revealed that acutely hypoxic cells may be more metastatic than normoxic or chronically hypoxic cells. In the present work, causal relations between acute cyclic hypoxia and metastasis were studied by periodically exposing BALB/c nu/nu mice bearing A-07 human melanoma xenografts to a low oxygen atmosphere. The hypoxia treatment consisted of 12 cycles of 10 min of 8% O2 in N2 followed by 10 min of air for a total of 4 hr, began on the first day after tumor cell inoculation and was given daily until the tumors reached a volume of 100 mm3. Twenty-four hours after the last hypoxia exposure, the primary tumors were subjected to dynamic contrast-enhanced magnetic resonance imaging for assessment of blood perfusion before being resected and processed for immunohistochemical examinations of microvascular density and expression of proangiogenic factors. Mice exposed to acute cyclic hypoxia showed increased incidence of pulmonary metastases, and the primary tumors of these mice showed increased blood perfusion, microvascular density and vascular endothelial growth factor-A (VEGF-A) expression; whereas, the expression of interleukin-8, platelet-derived endothelial cell growth factor and basic fibroblast growth factor was unchanged. The increased pulmonary metastasis was most likely a consequence of hypoxia-induced VEGF-A upregulation, which resulted in increased angiogenic activity and blood perfusion in the primary tumor and thus facilitated tumor cell intravasation and hematogenous transport into the general circulation.
Most tumors show severe microvascular abnormalities, are heterogeneous in oxygen tension (pO2) and develop regions with hypoxic (pO2 < 10 mmHg) tissue during growth.1 Experimental studies have indicated that tumor hypoxia may cause resistance to treatment and promote metastatic dissemination and growth.2 Clinical studies have shown that patients with highly hypoxic primary tumors may have increased frequency of locoregional treatment failure, increased incidence of distant metastases and poor disease-free and overall survival rates following radiation therapy alone or radiation therapy combined with surgery and/or chemotherapy.3 It has been suggested that hypoxia may promote metastasis by increasing the genetic instability of tumors, by selecting for particularly aggressive tumor cell phenotypes with a diminished apoptotic potential or by upregulating the expression of metastasis-enhancing genes.4 However, the mechanisms linking tumor hypoxia to increased metastasis may be multiple and are far from fully understood.
Two main categories of hypoxia have been recognized in tumors: chronic hypoxia, also known as diffusion-limited or permanent hypoxia, and acute hypoxia, also known as perfusion-limited or transient hypoxia.5 Chronic hypoxia is a consequence of increased intervessel distances or a low number of tumor-supplying arterioles, and is typically found in regions far from blood vessels or in regions with low oxyhemoglobin saturation.6 In untreated tumors, chronically hypoxic cells are believed to remain hypoxic until they die because of lack of oxygen or nutrients. These cells have been shown to have a lifetime within the range of 4–10 days in experimental tumors.7 Acute hypoxia is a consequence of transient limitations in blood perfusion caused by vasomotor activity in tumor-supplying arterioles, mechanical compression of microvessel walls caused by proliferating tumor cells, intravessel stasis or vessel obstruction caused by circulating white blood cells.8 Acutely hypoxic cells are found downstream of perfusion-impairing vessel abnormalities and are believed to experience several short-term periods of hypoxia during their lifetime.8, 9 In experimental tumors, the duration of these periods has been shown to range from less than a minute to several hours.10
It has been hypothesized that acutely hypoxic cells may have a higher metastatic potential than chronically hypoxic cells because they, in general, are located closer to blood vessels.2, 11 In addition to having easier access to the circulation, acutely hypoxic cells may have higher energy status than chronically hypoxic cells and thus be in better condition to accomplish the different steps of the metastatic process.2–5 Studies of experimental tumors have supported the hypothesis that associations between tumor hypoxia and metastasis observed in clinical investigations may be attributed primarily to acutely rather than chronically hypoxic cells.12, 13
The impact of acute hypoxia on metastatic growth has been studied by exposing tumor cells to acute hypoxia in vitro and then inoculating the cells intravenously in recipient mice for formation of lung colonies.14–16 These studies showed that the lung colonization potential of tumor cells may be transiently enhanced by acute hypoxia, possibly because of an altered apoptotic response or hypoxia-induced upregulation of metastasis-promoting gene products. However, this type of experiment considers only the last steps of the metastatic process. To study causal relations between acute hypoxia and spontaneous metastasis, Hill and coworkers have examined the development of metastases in tumor-bearing mice exposed to acute cyclic hypoxic stress in vivo.17, 18 They showed that experimentally imposed acute hypoxia increased the incidence of metastases in mice bearing KHT murine fibrosarcomas17 or ME-180 human cervical carcinomas.18 In contrast, similar experiments have suggested that tumor progression and metastasis in transgenic mouse breast cancer models are not driven primarily by acute hypoxic stress.19, 20
Associations between tumor hypoxia and metastasis have been studied in our laboratory by using human melanoma xenografts as preclinical models of human cancer.21, 22 Hypoxia was found to promote pulmonary metastasis in D-12 tumors by upregulating the proangiogenic factor interleukin-8 (IL-8),21 and lymph node metastasis in R-18 tumors by upregulating the invasive growth-promoting receptor urokinase-type plasminogen activator receptor.22 The purpose of the study reported here was to search for causal relations between acute hypoxia and spontaneous metastasis by using an experimental procedure similar to that used by Hill and coworkers.17 The A-07 melanoma was selected for the study because naturally occurring hypoxia cannot be detected in small tumors of this line.23 Tumor-bearing mice exposed to acute cyclic hypoxic stress were found to show increased incidence of lung metastases, most likely because of increased angiogenesis in the primary tumor as a consequence of hypoxia-induced upregulation of vascular endothelial growth factor-A (VEGF-A).
Adult (8–10 weeks of age) female BALB/c nu/nu mice, bred and maintained as described elsewhere,12 were used as host animals for xenografted tumors. The animal experiments were approved by the institutional committee on research animal care and were done according to the US Public Health Service Policy on Humane Care and Use of Laboratory Animals.
The A-07 human melanoma cell line was established and characterized in our laboratory as described previously.24 The cells used in the present experiments were obtained from our frozen stock and maintained in monolayer culture in RPMI 1640 (25 mM HEPES and L-glutamine) supplemented with 13% bovine calf serum, 250 mg/l penicillin and 50 mg/l streptomycin.12
Intradermal A-07 tumors were initiated from monolayer cultures by inoculating aliquots of 3.5 × 105 cells into the left mouse flank.12 Tumor volume (V) was calculated as V = π/6 × ab2, where a is the longer and b is the shorter of two orthogonal diameters.
Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and pO2 measurements were carried out with anesthetized mice in the prone position. Fentanyl citrate (Janssen Pharmaceutica, Beerse, Belgium), fluanisone (Janssen Pharmaceutica) and midazolam (Hoffmann-La Roche, Basel, Switzerland) were administered in doses of 0.63, 20 and 10 mg/kg, respectively. By using the 86Rb method, it has been verified that this anesthesia does not alter the blood perfusion of A-07 tumors in mice having a normal body core temperature.25 A heating pad was used to keep the body core temperature of the mice at 37–38°C during experiments.
Hypoxia treatment in vivo
Unanesthetized mice were placed in an in-house-made incubation chamber and exposed to a continuous flow of a humidified gas mixture at room temperature to induce hypoxia. The hypoxia treatment consisted of 12 cycles of 10 min of 8% O2 in N2 followed by 10 min of air for a total of 4 hr. Control mice were exposed to a continuous flow of humidified air for 4 hr. The hypoxia treatment began on the first day after tumor cell inoculation and was given once per day, 7 days per week, until the tumors reached a volume of 100 mm3.
DCE-MRI and image processing and analysis were carried out as described previously.26 Briefly, a 24-G neoflon connected to syringe by a polyethylene tubing was inserted in the tail vein, and gadolinium diethylene-triamine penta-acetic acid (Gd-DTPA; Schering, Berlin, Germany) was administered in a bolus dose of 0.3 mmol/kg after the mice had been placed in the magnet. T1-weighted DCE-MRI was performed at a spatial resolution of 0.23 × 0.47 × 2.0 mm3 and a time resolution of 14 sec by using a 1.5-T whole-body scanner (Signa; General Electric, Milwaukee, WI) and a slotted tube resonator transceiver coil constructed for mice. Two calibration tubes, 1 with 0.5 mM Gd-DTPA in 0.9% saline and the other with 0.9% saline only, were placed adjacent to the mice in the coil. The tumors were imaged axially in a single section through the tumor center by using a scan thickness of 2 mm, a number of excitations of 1, an image matrix of 256 × 64 and a field of view of 6 × 3 cm2. Two types of spoiled gradient recalled images were recorded: proton density images with repetition time TR = 900 ms, echo time TE = 3.2 ms and flip angle αPD = 20° and T1-weighted images with TR = 200 ms, TE = 3.2 ms and αT1 = 80°. Two proton density images and 3 T1-weighted images were acquired before Gd-DTPA was administered, and T1-weighted images were recorded for 15 min after the administration of Gd-DTPA. Finally, 1 proton density image was acquired. Tumor images were analyzed on a voxel-by-voxel basis by using software developed in IDL (Interactive Data Language, Boulder, CO). Gd-DTPA concentrations were calculated from signal intensities as described by Hittmair et al.,27 and the DCE-MRI series were analyzed by using the arterial input function established by Benjaminsen et al.28 and the modified Kety pharmacokinetic model.29 Parametric images of E × F [blood perfusion in units of ml/(g min)] and λ (parameter proportional to extravascular extracellular volume fraction) were generated by using the SigmaPlot software (SPSS Science, Chicago, IL). E × F and λ are related to the parameters of the commonly used Tofts pharmacokinetic model (Ktrans, the volume transfer constant of Gd-DTPA, and ve, the extravascular extracellular volume fraction) by the following expressions: E × F = Ktrans / [ρ × (1 − Hct)] and λ = ve / (1 − Hct), where ρ is the density of the tumor tissue (1 g/ml) and Hct is the hematocrit.29 A previous study in our laboratory has shown that the experimental procedure described above provides numerical values of E × F and ve of human melanoma xenografts that are closely related to the absolute values of blood perfusion and extravascular extracellular volume fraction.30
Immunohistochemical detection of microvessels, tumor hypoxia and proangiogenic factors
CD31 was used as a marker of tumor endothelial cells, and pimonidazole [1-[(2-hydroxy-3-piperidinyl)-propyl]-2-nitroimidazole], administered as described previously,31 was used as a marker of tumor hypoxia. Tumor slices were fixed in phosphate-buffered 4% paraformaldehyde or liquid nitrogen. Immunohistochemistry was done by using a peroxidase-based indirect staining method.12 An anti-pimonidazole rabbit polyclonal antibody (gift from Prof. J.A. Raleigh, Department of Radiation Oncology, University of North Carolina School of Medicine, Chapel Hill, NC), an anti-human VEGF-A rabbit polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA), an anti-human IL-8 rabbit polyclonal antibody (Endogen, Woburn, MA), an anti-human platelet-derived endothelial cell growth factor (PD-ECGF) goat polyclonal antibody (R&D Systems, Abingdon, UK), an anti-human basic fibroblast growth factor (bFGF) rabbit polyclonal antibody (Oncogene Science, Cambridge, MA) or an anti-mouse CD31 rat monoclonal antibody (Research Diagnostics, Flanders, NJ) was used as primary antibody. Controls included omission of the primary antibody, incubation with normal rabbit and goat immunoglobulin or normal rabbit and goat serum, and incubation with blocking peptides before staining. Diaminobenzidine was used as chromogen, and hematoxylin was used for counterstaining. Quantitative studies of microvascular density and the expression of proangiogenic factors were based on 4 cross-sections of each tumor. Staining intensities and positive area fractions were determined by image analysis.22, 31 Microvessels were defined and scored manually as described previously.21
The number of microscopic lung metastases was determined by histological examination. The lungs were resected from the host mice, fixed in phosphate-buffered 4% paraformaldehyde and embedded in paraffin. Histological sections were cut through each lobe at 100-μm intervals and stained with hematoxylin and eosin. Groups of 5 or more clearly identifiable tumor cells were scored as a metastasis. The analysis was carried out by a trained scientist blinded to the treatment group. The results were confirmed by an experienced pathologist.
Design of in vivo cyclic hypoxia experiment
Eight hypoxia-treated mice and 8 control mice were included in the main experiment. The mice were given the last hypoxia treatment on the same day as their tumor reached a volume of 100 mm3, and ∼24 hr later, they were subjected to DCE-MRI and pimonidazole administration. The mice were euthanized ∼4 hr after the pimonidazole administration, and tumors and lungs were resected and processed for histological examinations. In a separate experiment, the primary tumors of 3 hypoxia-treated and 3 control mice were resected immediately after the last hypoxia exposure and processed for histological examinations of the expression of proangiogenic factors.
Assessment of the oxygenation status of untreated tumors
The oxygenation status of untreated ∼100-mm3 tumors was assessed by pimonidazole immunohistochemistry as described earlier, using ∼1,000-mm3 tumors as a positive control. In addition, a radiobiological assay was used in attempts to detect hypoxic cells in untreated 100-mm3 tumors.31 Briefly, a Siemens Stabilipan X-ray unit, operated at 220 kV, 19–20 mA and with 0.5-mm Cu filtration, was used for irradiation. Monolayer cultures in exponential growth were irradiated under aerobic conditions at a dose rate of 3.4 Gy/min. Tumors were irradiated by using a radiation field of 15 × 15 mm2 and a dose rate of 5.1 Gy/min. Cell survival was measured by using a plastic surface colony assay.31 Single-cell suspensions were obtained by treating resected tumors with an enzyme solution (0.2% collagenase, 0.05% pronase and 0.02% DNase) at 37°C for 2 hr. Trypan blue-negative cells were plated in 25-cm2 tissue culture flasks and incubated at 37°C for 14 days. Cells giving rise to colonies > 50 cells were scored as clonogenic. The cell surviving fraction of an irradiated tumor was calculated from the plating efficiency of the cells of the tumor and the mean plating efficiency of the cells of 6 untreated control tumors. Plots of cell surviving fraction versus radiation dose were generated, and cellular radiation sensitivies (i.e., D0 values) were determined by regression analysis.
Measurement of tumor oxygen tension in mice exposed to cyclic hypoxia
In separate experiments, tissue pO2 was measured in ∼100-mm3 tumors in mice given hypoxia treatment, using a fiberoptic oxygen-sensing device (OxyLite 2000, Oxford Optronix, Oxford, UK) as described elsewhere.32 Briefly, a pO2 probe was inserted into the tumor center, and pO2 readings were recorded every 5 or 10 sec. The experiments began 20 min after the probe insertion to avoid artifacts caused by the insertion procedure.32
Treatment with blocking antibody in vivo
The specific roles of VEGF-A, IL-8, PD-ECGF and bFGF in tumor angiogenesis and metastasis in mice exposed to cyclic hypoxic stress were investigated by using neutralizing antibodies against these proangiogenic factors. The antibodies used for treatment were anti-human VEGF-A mouse monoclonal antibody (R&D Systems), anti-human IL-8 mouse monoclonal antibody (R&D Systems), anti-human PD-ECGF goat polyclonal antibody (R&D Systems) and anti-human bFGF goat polyclonal antibody (R&D Systems). The antibody solutions were diluted in PBS and administered to the mice in volumes of 0.25 ml by intraperitoneal injection. The treatments started on the same day as the tumors reached a volume of 75 mm3 and consisted of 8 doses of 25 μg (VEGF-A and bFGF) or 100 μg (IL-8 and PD-ECGF) of antibody given in intervals of 24 hr. Control mice were treated at 24-hr intervals with 8 doses of an irrelevant antihuman monoclonal antibody of the same isotype as the blocking antibody. The number of lung metastases and the microvascular density of the primary tumor were assessed as described previously.21
Hypoxia treatment in vitro
Monolayer cultures were exposed to acute cyclic or acute continuous hypoxia by using the steel chamber method as described elsewhere.15 Briefly, the steel chambers were flushed with humidified, highly purified gas mixtures consisting of 5% CO2 in 95% N2 or 5% CO2 in air. The cyclic hypoxia treatment consisted of 6 cycles of 30 min under hypoxic conditions followed by 30 min under aerobic conditions for a total of 6 hr. The continuous hypoxia treatment consisted of a single 6-hr incubation under hypoxic conditions. Control cells were exposed to a continuous flow of 5% CO2 in air for 6 hr. During hypoxia treatment, the concentration of O2 in the medium was 10–100 ppm. The cell cultures were kept at 37°C during treatment.
Cell viability assay
The viability of cells from control and hypoxia-treated cultures was assessed by measuring plating efficiency in vitro using the plastic surface colony assay described above.15 Aliquots of 200 cells were plated in 25-cm2 culture flasks, and after an incubation period of 14 days, the cultures were fixed and colonies >50 cells were counted.31
VEGF-A secretion assay
Rate of VEGF-A secretion was determined as described previously by using a commercially available human VEGF165 enzyme-linked immunosorbent assay kit (R&D Systems) according to manufacturer's instructions.15
Lung colony assay
Aliquots of 1.0 × 106 cells suspended in 0.2 ml of HBSS were inoculated into the lateral tail vein of mice. The mice were euthanized and autopsied 5 weeks after the cell inoculation. The lungs were removed and fixed in Bouin's solution for 24 hr, and the number of surface colonies was recorded by using a stereomicroscope. Lung colonization efficiency was determined by correcting the number of colonies for the plating efficiency measured in vitro.15
Results are presented as arithmetic mean ± SE unless otherwise stated. Correlations between 2 parameters were searched for by linear regression analysis. Statistical comparisons of data were carried out by using the Student t test when the data complied with the conditions of normality and equal variance. Under other conditions, comparisons were carried out by nonparametric analysis using the Mann-Whitney rank sum test. The Kolmogorov-Smirnov method was used to test for normality. Probability values of p < 0.05, determined from 2-sided tests, were considered significant. The statistical analysis was performed by using the SigmaStat statistical software (SPSS Science).
Tumors in untreated mice did not show hypoxic tissue regions
To investigate whether small A-07 tumors showed naturally occurring hypoxia, cell survival curves were established for ∼100-mm3 tumors irradiated in vivo and cell cultures irradiated under aerobic conditions in vitro (Fig. 1a). The D0 values were determined to be 0.88 ± 0.09 Gy (tumors) and 0.90 ± 0.07 Gy (cultures). The cellular radiation sensitivity did not differ between the tumors and the cultures (p > 0.05), consistent with the assumption that ∼100-mm3 tumors do not show radiation-resistant hypoxic cells. Moreover, ∼100-mm3 tumors did not stain positive for the hypoxia marker pimonidazole, whereas ∼1,000-mm3 control tumors showed small positive foci scattered throughout the tissue and thin bands of positive staining adjacent to central necroses (Fig. 1b).
Tumors in mice undergoing hypoxia treatment showed acute cyclic hypoxia
The pO2 in tumors of mice undergoing hypoxia treatment varied cyclically with time. This is illustrated in Figure 2, which shows pO2 profiles for 3 different tumors, recorded in regions differing substantially in pO2 under air-breathing conditions. The pO2 decreased rapidly to 0 mmHg after flushing with 8% O2 in N2 was initiated (red lines) and increased rapidly to the pretreatment level after flushing with air was initiated (blue lines). These data demonstrate clearly that the hypoxia treatment used in the present work resulted in acute cyclic hypoxia in A-07 tumors.
Tumors exposed to acute cyclic hypoxia showed increased metastasis and angiogenesis
The time from tumor cell inoculation to the tumors reached a volume of 100 mm3 varied from 13 to 16 days, and was not significantly different for the hypoxia-treated and the control mice (p > 0.05). Moreover, the volume of the tumors on the last day of treatment was not significantly different for the hypoxia-treated (110 ± 7 mm3) and the control (112 ± 5 mm3) tumors (p > 0.05). The number of lung micrometastases varied among the mice from 0 to 5, and was significantly higher for the hypoxia-treated mice than for the control mice (Fig. 3a; p = 0.021). An example of a lung micrometastasis is illustrated in the right panel of Figure 3a. Furthermore, the microvascular density of the primary tumor (#/mm2) varied from 80 to 175, and was significantly higher for the hypoxia-treated tumors than for the control tumors (Fig. 3b; p = 0.022). The general quality of the CD31 staining is shown in the right panel of Figure 3b. Positive pimonidazole staining could not be detected in any of the tumors. The area fraction staining positive for VEGF-A did not differ between the hypoxia-treated and the control tumors (p > 0.05), whereas the staining intensity was significantly higher in the hypoxia-treated than in the control tumors (Fig. 3c; p < 0.0001). The staining for IL-8, PD-ECGF and bFGF was weak in all 16 tumors, and neither the positive area fraction nor the staining intensity differed between the hypoxia-treated and the control tumors (p > 0.05 for both parameters and all 3 proangiogenic factors). Furthermore, hypoxia-treated tumors resected immediately after the last hypoxia exposure did not differ from untreated control tumors in IL-8 (Fig. 4a), PD-ECGF (Fig. 4b), or bFGF (Fig. 4c) staining intensity or positive area fraction (p > 0.05 for both parameters and all 3 proangiogenic factors).
Tumors exposed to acute cyclic hypoxia showed increased blood perfusion
The blood perfusion was higher in the tumors exposed to hypoxia than in the control tumors. This is illustrated qualitatively in Figure 5a, which shows the E × F image and frequency distribution of a representative hypoxia-treated tumor and a representative control tumor. Representative single-voxel Kety curves are shown in the lower panels of Figure 5a, illustrating that accurate curve fits were achieved, also in the early phase of the curves, probably because the tumors studied here have a large extravascular extracellular volume fraction.33 Quantitative analysis showed that median E × F differed among the tumors from 0.18 to 1.19 ml/(g min), and was significantly higher for the hypoxia-treated tumors than for the control tumors (Fig. 5b; p = 0.010). Furthermore, there was a strong linear correlation between median E × F and microvascular density (Fig. 5c; R2 = 0.83; p < 0.00001). On the other hand, median λ did not differ significantly between the hypoxia-treated and the control tumors (data not shown; p > 0.05).
Anti-VEGF-A treatment resulted in decreased metastasis and decreased microvascular density
Treatment with blocking antibody had significant effects on the development of lung metastases (Fig. 6a) and the microvascular density of the primary tumor (Fig. 6b). Anti-VEGF-A treatment resulted in significantly decreased incidence of metastases (p = 0.019) and significantly reduced microvascular density (p = 0.0098). In contrast, the primary tumor microvascular density was not reduced after treatment with blocking antibody against IL-8 (p > 0.05), PD-ECGF (p > 0.05), or bFGF (p > 0.05). The incidence of lung metastases tended to decrease after anti-IL-8 treatment (p = 0.075), anti-PD-ECGF treatment (p = 0.14) and anti-bFGF treatment (p = 0.24), however, the number of metastases was not significantly reduced by any of these 3 treatments.
Cells exposed to hypoxia in vitro showed increased VEGF-A secretion
Monolayer cell cultures showed a high rate of VEGF-A secretion varying from 500 to 1,500 pg/(hr 106 cells), depending on the experimental conditions. The secretion rate was higher for hypoxia-treated cells than for aerobic control cells, regardless of whether the cells were exposed to the continuous (p = 0.033) or the cyclic (p = 0.038) hypoxia treatment (Fig. 7a).
Cells exposed to hypoxia in vitro did not show increased lung colonization potential
The lung colonization potential was not significantly different for hypoxia-treated and aerobic control cells, regardless of whether the cells were exposed to the continuous (p > 0.05) or the cyclic (p > 0.05) hypoxia treatment (Fig. 7b). Cell viability was not reduced by the hypoxia treatments. The plating efficiency of the cells was measured to be 86 ± 6% (aerobic controls), 82 ± 5% (continuous hypoxia), or 88 ± 7% (cyclic hypoxia).
Causal relations between acute cyclic hypoxia and spontaneous metastasis were studied by exposing mice bearing A-07 human melanoma xenografts to 12 cycles of 10 min of 8% O2 in N2 followed by 10 min of air daily for 13–16 days until the tumors reached a volume of 100 mm3. This treatment was verified to cause acute cyclic hypoxia in the tumor tissue and resulted in increased incidence of microscopic lung metastases. The increased metastasis was most likely a consequence of hypoxia-induced changes in the primary tumor rather than increased cell survival in the circulation or increased cell extravasation and growth at the secondary site. This suggestion follows from the observation that A-07 cells did not show increased incidence of lung colonies when exposed to acute continuous or acute cyclic hypoxia in vitro and inoculated intravenously. The lack of increased lung colonization after hypoxia treatment in vitro cannot be attributed to the experimental method or to the relatively short durations of the hypoxia exposures. This conclusion follows from earlier experiments with this method, which showed that the lung colonization potential of D-12 melanoma cells was increased after a 4-hr continuous hypoxia treatment in vitro, whereas hypoxia treatments with durations up to 24 hr did not increase the lung colonization potential of A-07 cells.15
Primary tumors exposed to acute cyclic hypoxia showed increased microvascular density and blood perfusion, probably as a consequence of increased angiogenesis induced by the hypoxia treatment. High angiogenic activity in tumors results in microvascular networks composed of immature vessels and vessel sprouts that may facilitate tumor cell intravasation and metastasis by several mechanisms.34 Immature vessels are leaky and have fragmented basement membranes, making them more accessible to tumor cells than mature vessels. Vessel sprouts are characterized by extensive endothelial cell proliferation and migration, and proliferating and migrating endothelial cells secrete paracrine growth factors and proteolytic enzymes that may facilitate the escape of tumor cells into the circulation. Finally, growing vessels may actively promote intravasation by engulfing tumor cells.
Under normoxic conditions, the angiogenesis of A-07 tumors has been shown to be mediated by several proangiogenic factors, including VEGF-A, IL-8, PD-ECGF and bFGF.35 By counting the number of capillaries in the dermis oriented toward small 7-days-old tumors in untreated mice and mice treated with blocking antibody, it was shown that the initial tumor vascularization was inhibited by anti-VEGF-A, anti-IL-8, anti-PD-ECGF and anti-bFGF treatment. However, in contrast to the effect of anti-VEGF-A treatment, the effects of anti-IL-8, anti-PD-ECGF and anti-bFGF treatment were small and did not translate into decreased intratumor microvascular density, possibly because the volume of the tumors was reduced by the antibody treatments.35
The hypoxia-induced increase in angiogenic activity observed here was also most likely caused primarily by VEGF-A. This suggestion follows from the immunohistochemical examinations, which demonstrated unequivocally that the hypoxia-treated tumors showed higher expression of VEGF-A than the control tumors, whereas the expression of IL-8, PD-ECGF and bFGF was weak and did not differ between the 2 tumor groups, regardless of whether the primary tumors were resected immediately after or ∼24 hr after the last hypoxia exposure. The importance of VEGF-A in the angiogenesis of A-07 primary tumors was underscored further by the angiogenesis inhibition experiments, which showed that anti-VEGF-A treatment, in contrast to anti-IL-8, anti-PD-ECGF and anti-bFGF treatment, resulted in reduced microvascular density.
The angiogenesis inhibition experiments also showed that VEGF-A may play an important role in the metastasis of A-07 tumors as the incidence of lung metastases decreased significantly after anti-VEGF-A treatment. Since the incidence of lung metastasis tended to decrease after anti-IL-8, anti-PD-ECGF and anti-bFGF treatment, though without being significantly different from that in untreated mice, the possibility that also IL-8, PD-ECGF and/or bFGF may promote lung metastasis of A-07 tumors cannot be excluded. However, if these factors are involved, this is probably a concequence of other effects than their proangiogenic effects. This suggestion follows from the observation that the microvascular density of the primary tumors did not decrease after anti-IL-8, anti-PD-ECGF and anti-bFGF treatment.
Taken together, the present observations suggest that exposure to acute cyclic hypoxic stress promoted lung metastasis in A-07 tumors by the following mechanism: acute cyclic hypoxia caused upregulation of VEGF-A, increased VEGF-A expression resulted in increased microvascular density and blood perfusion in the primary tumor, and these vascular changes facilitated hematogenous metastasis.
Owing to its effects on tumor angiogenesis and normal tissue microvascular permeability, VEGF-A may be involved in hematogenous metastasis both by facilitating tumor cell intravasation at the primary site and by facilitating tumor cell extravasation at the secondary site.4, 34 The present work showed that hypoxia-induced up-regulation of VEGF-A at the primary site enhanced the metastatic propensity of A-07 tumors, whereas the increased VEGF-A expression induced by hypoxia in vitro did not enhance the lung colonization potential of A-07 cells. As A-07 cells have a particularly high constitutive expression of VEGF-A,35 it is possible that the constitutive VEGF-A expression was sufficiently high for efficient lung colonization but not sufficiently high for optimal cell intravasation in the primary tumor. Earlier studies in our laboratory have shown that the incidence of spontaneous lung metastases in D-12 melanoma xenografts is associated with the extent of hypoxia in the primary tumor and that D-12 cells exposed to hypoxia in vitro show increased lung colonization when inoculated into the tail vein of recipient mice.15, 21 As the constitutive VEGF-A expression is low in D-12 cells,35 efficient lung colonization may require hypoxia-induced VEGF-A up-regulation. It should also be noticed that although VEGF-A is involved in the angiogenesis of D-12 tumors,15 the hypoxia-induced angiogenesis is driven primarily by IL-8 as immunohistochemical studies have shown that D-12 tumors stain homogeneously for VEGF-A, whereas the staining for IL-8 is heterogeneous and localized mainly to tissue regions staining positive for the hypoxia marker pimonidazole.21
An experimental set-up similar to that reported here was used by Hill and colleagues to study effects of acute cyclic hypoxic stress on the development of spontaneous metastases in mice bearing KHT murine fibrosarcomas17 or ME-180 human cervical carcinomas.18 Exposure to acute cyclic hypoxia increased the incidence of lung micrometastases in mice with KHT tumors,17 possibly because the hypoxia treatment caused increased resistance to apoptosis by upregulating Mdm2 (murine double minute-2, an inhibitor of p53 transcriptional activation) and hence caused increased survival of tumor cells arrested in the lungs.36 Mice with ME-180 tumors exposed to acute cyclic hypoxia showed unchanged incidence of lung metastases, whereas the incidence of lymph node metastases was increased.18 The molecular mechanism leading to the increased lymph node metastasis was not identified. However, it has been shown that exposure to cyclic hypoxic stress may cause up-regulation of several metastasis-related genes in ME-180 tumors, including chemokine receptor-4, urokinase-type plasminogen activator receptor, VEGF-C, transformed 3T3 double minute-2 and osteopontin.37 Thus, the mechanism of acute cyclic hypoxia-induced metastasis suggested here for A-07 tumors is fundamentally different from those suggested for KHT and ME-180 tumors. Interestingly, the A-07 primary tumors showed unchanged growth rate and increased microvascular density when exposed to acute cyclic hypoxic stress, in sharp contrast to the ME-180 tumors, which showed reduced growth and reduced CD31-positive vascular area fraction.18
It should also be noticed that small A-07 melanoma xenografts are particularly suitable models for studying biological consequences of acute cyclic hypoxia in tumor tissues, because these tumors do not show naturally occurring hypoxia.23 This was confirmed in the present study by comparing the cellular radiation sensitivity of ∼100-mm3 tumors with that of aerobic cell cultures. Moreover, positive pimonidazole staining could not be detected in any hypoxia-treated or control tumor 24 hr after the last hypoxia exposure. In contrast, the KHT and ME-180 tumors studied by the Hill group contained significant fractions of naturally occurring hypoxic cells, up to ∼50% in the ME-180 tumors, as detected by staining for the hypoxia marker EF5.17, 18
Hill and coworkers have also demonstrated that exposure to acute cyclic hypoxic stress may result in increased plasma levels of mutagenic lipid peroxidation products and increased tumor levels of 8-oxo-7,8-dihydro-2′-deoxyguanosine DNA lesions in transgenic mouse breast cancer models.19, 20 However, these increases did not translate into enhanced tumor progesssion and elevated incidence of metastases. Thus, the experiments with the transgenic tumor models provided results that differ from those obtained with the transplantable KHT, ME-180 and A-07 tumor models. The biological causes underlying these disparate observations have not been identified. However, it is tempting to speculate that the transplantable tumor models may be more aggressive than the transgenic spontaneous tumor models and, hence, may be more prone to develop metastases after hypoxic stress.
The treatment of many cancer patients involves strategies that may cause transient changes in the oxygenation status of the tumor tissue. For example, patients with chronic hypoxia (e.g., some lung cancer patients and old patients with emphysema) may undergo oxygen treatment, and the treatment of patients with unresectable hepatocellular carcinoma may involve hepatic artery occlusion. The possibility that treatment strategies of this type may promote malignant progression and metastatic dissemination merits thorough preclinical and clinical investigations.
In summary, acute cyclic hypoxic stress promoted spontaneous pulmonary metastasis in A-07 human melanoma xenografts, most likely by up-regulating the expression of the proangiogenic factor VEGF-A. The increased VEGF-A expression caused increased angiogenic activity in the primary tumor that facilitated tumor cell intravasation and metastatic dissemination.