Microregional Expression of Glucose Transporter-1 and Oxygenation Status: Lack of Correlation in Locally Advanced Cervical Cancers

Under physiologic conditions, glucose transporter-1 (GLUT-1) is expressed most strongly in erythrocytes and blood-brain or blood-nerve barriers (1–3). Overexpression of GLUT-1 is found in many types of solid malignancies (1), among them cancers of the uterine cervix (4). Overexpression of GLUT-1 significantly contributes to the phenomenon of increased glucose uptake (as measured by positron emission tomography imaging of ¹⁸F-fluorodeoxyglucose uptake) found in solid tumors which can also be exploited for diagnostic purposes (5, 6). Because expression of GLUT-1 is stimulated by hypoxia via activation of the hypoxia-inducible factor-1 (HIF-1; ref. 7), the question arises as to whether GLUT-1 expression in malignant tissue can be used as a surrogate measure of tissue oxygen (O₂) depletion and could thus serve as an endogenous marker of tumor hypoxia.

Hypoxia is a hallmark of solid tumors and has an effect both on the effectiveness of cancer treatment and on the aggressive traits of malignant cells. Reduced O₂ availability diminishes the efficacy of ionizing radiation (8). Significantly higher radiation doses are thus needed in hypoxic tissue to achieve equivalent cytotoxic effects compared with normoxic microenvironments. Additionally, severely hypoxic conditions (<0.1% O₂) can promote the accumulation of genomic mutations (9–11) and may thus expedite malignant progression (e.g., via activation of oncogenes or inactivation of tumor suppressor genes). Hypoxia also has a profound effect on gene expression (termed “hypoxic response”; e.g., ref. 12), which is mediated via oxygen-regulated transcription factors (13), among which HIF-1 occupies a pivotal position. GLUT-1 is one member of the broad range of target genes of HIF-1 which are believed to confer enhanced cell survival and proliferation under hypoxic conditions (for a recent review, see ref. 7) and are thus viewed as driving factors for tumor growth, at least when a net effect is considered. Consistent with such a concept, overexpression of GLUT-1 has been shown to be associated with an adverse prognosis in several tumor entities (14–23), among them cancers of the uterine cervix (4). The present study examined the possibility of assessing GLUT-1 expression to gain information on the extent of tumor hypoxia. Additionally, the role of GLUT-1 as a prognostic factor was assessed in this cohort of patients with long-term follow-up.

Patients. All patients in this study were enrolled in a prospective clinical trial for the evaluation of the significance of tumor oxygenation in primary, locally advanced carcinomas of the uterine cervix that commenced at the Department of Obstetrics and Gynecology, University of Mainz Medical School, in June 1989. The study design was approved by the local Medical Ethics Committee, with patients giving informed written consent before being enrolled. All 47 patients from the former study, for whom one or two tumor biopsy specimens of the oxygen measurement tracks were available, were included in the present study. Patients of this subgroup had been recruited between August 1991 and April 1997. Table 1 shows relevant patient and tumor characteristics at the time of pretreatment oxygen tension (oxygen partial pressure, pO₂) measurements.

For correlations involving survival, only patients treated with curative intent were included (n = 42). In 31 of these, the primary therapy was surgical. Abdominal radical hysterectomy and pelvic/periaortic lymph node dissection was the standard surgical procedure. In 5 of 31 patients, supralevator exenteration had to be done instead of radical hysterectomy. The remaining 11 cases were treated by radiation instead of surgery because their tumors were fixed to the pelvic wall or their comorbidity excluded extensive operations. Primary radiotherapy was given as combined teletherapy and brachytherapy. External beam irradiation was applied with 10 MV photons produced by a linear accelerator at the Division of Radiotherapy. For brachytherapy, a high–dose rate 192_Ir afterloading machine at the Department of Obstetrics and Gynecology was used (for details, see ref. 24). Adjuvant chemotherapy regimens are described in detail in Ref. 31. Median follow-up time was 28 months (SD 28), ranging from 4 to 95 months.

Tumor oxygen tension measurements. Tumor pO₂ was measured pretherapeutically with the computerized Eppendorf histography system (Eppendorf, Hamburg, Germany), using a protocol that has been described in detail previously (25). Briefly, pO₂ readings were done in the conscious patient along linear tracks, first in the s.c. fat of the mons pubis followed by cervical measurements at the 12 and 6 o'clock sites in macroscopically vital tumor tissue. Within the tumor tissue, up to 35 pO₂ measurements were made along each electrode track (70 readings in total) starting at a tissue depth of 5 mm. The individual pO₂ measurement points were situated 0.7 mm apart, resulting in an overall measurement track length of ∼25 mm. Immediately following pO₂ measurements, needle core biopsies (obtained using Biopty, Radioplast, Uppsala, Sweden) of ∼2 mm in diameter and 20 mm in length were taken from those tumor areas where pO₂ readings had been obtained. Both the pO₂ readings and the needle core biopsies were done without general anesthesia in all patients. Intravaginal temperature, arterial blood pressure, heart rate, hemoglobin concentration, hematocrit, and arterial oxyhemoglobin saturation were monitored at the time when pO₂ readings were taken. The pretherapeutic pO₂ measurements were usually done 1 to 5 days before oncological treatment. After histologic examination of the biopsy specimens, pO₂ measurements in necrotic tissue areas were excluded from analysis.

Immunohistochemistry. Expression of GLUT-1 was assessed in 80 biopsy specimens taken from the tumor pO₂ measurement tracks obtained directly after pO₂ measurement in 47 patients. Two biopsies, corresponding to the 6 and 12 o'clock positions of the tumor center were available for each of 33 patients and one biopsy for each of the remaining 14 cases. All material was fixed in formalin before being embedded in paraffin. Histologic slides were prepared from the paraffin blocks and dried overnight at 37°C. Subsequently, specimens were dewaxed in two changes of fresh xylene and rehydrated in a descending alcohol series. Retrieval of antigenic binding sites was done by heating specimens in 10 mmol/L citrate buffer (pH 6.0) in a microwave oven for 17 minutes. GLUT-1 polyclonal rabbit anti-human GLUT-1 clone MYM (DakoCytomation, Hamburg, Germany) was used as the primary antibody at a concentration of ∼1 μg/mL in antibody diluent (DakoCytomation, 1:200 dilution). Incubation took place overnight at 4°C. A biotinylated goat anti-mouse/rabbit secondary antibody was applied for 30 minutes at room temperature and further detection was carried out using a streptavidin-biotin-horseradish peroxidase system (Duet-Kit, DakoCytomation), in accordance with the manufacturer's instructions. Negative control specimens were treated with antibody diluent without the primary antibody under the same conditions. Slides were counterstained with Mayer's hematoxylin, dehydrated in an ascending alcohol series and covered with a coverslip using Eukitt mounting medium (Riedel-de Haen, Seelze, Germany).

Assessment of GLUT-1 expression. A semiquantitative scoring system was used to assess the degree of GLUT-1 expression in entire biopsy sections: score 0, no staining or only very few positive cells (“absent”); score 1, <10% positive (“weak”), score 2, 11% to 50% positive (“moderate”); score 3, >50% positive (“strong”). Each specimen was scored by two independent observers (A.M. and A.W.). Discordant cases were reevaluated and discussed using a conference microscope.

Statistical analysis. All statistical tests were done using the SPSS software package (version 11.5; SPSS, Inc., Chicago, IL). The significance level was set at α = 5% for all comparisons. Linear correlations between two variables were described by Spearman's rank correlation coefficient (ρ). Two-sided Mann Whitney U tests and Kruskal Wallis tests were used for comparison of categorized variables. Survival estimates were calculated using the Kaplan-Meier method and differences between groups were assessed with log-rank statistics. The Cox proportional hazards model was used for the multivariate analysis of the effect of individual factors on survival.

GLUT-1 expression. GLUT-1 expression exhibited a characteristic pattern, with staining intensity increasing as a function of distance from the vascularized tumor stroma (Fig. 1), being particularly strong in the viable cell layers immediately adjacent to necrotic areas. Erythrocytes and perineural tissue invariably stained positive. Variation of erythrocyte staining intensity was very low, indicating a neglectable batch to batch variation in overall GLUT-1 immunoreactivity. No positive staining was seen in the vascular tumor stroma. GLUT-1 expression was present in 59 of 80 biopsies (∼74%). Of these, GLUT-1 expression was weak in 37 cases (∼63%), moderate in 18 cases (∼30%), and strong in four cases (∼7%).

GLUT-1 expression, clinical, and pathohistologic data. GLUT-1 expression increased linearly with Fédération Internationale de Gynécologie et d' Obstétrique (FIGO) stage (r = 0.35, P = 0.002; see Fig. 2). Significantly higher expression of GLUT-1 was also found in larger tumors (r = 0.42, P = 0.0001) and in tumors with a higher pT stage (r = 0.34, P = 0.015). GLUT-1 expression showed no correlation with histologic grading, pN stage, patient age, parity, and pretherapeutic hemoglobin level.

GLUT-1 expression and oxygenation data. The Kruskal-Wallis test showed no differences in median pO₂, hypoxic fraction ≤2.5 mm Hg (HF 2.5) and hypoxic fraction ≤5 mm Hg (HF 5) between the four GLUT-1 expression scores. There were also no statistically significant differences in GLUT-1 expression between individual categories (e.g., absent expression versus strong expression; Fig. 3). A subgroup analysis of squamous cell carcinomas only (n = 60) also showed no differences in any of the oxygenation variables between the four classes of intensity of GLUT-1 expression. A weak trend (r = 0.34; P = 0.14, n = 20) was seen for higher values of HF 5 in non–squamous cell histology cases with higher GLUT-1 expression. As Fig. 3 shows, some severely hypoxic tumors had weak to absent expression of GLUT-1 and normoxic tumors repeatedly showed moderate to strong GLUT-1 expression.

GLUT-1 expression and survival. Univariate Kaplan-Meier survival analysis showed significantly improved overall (P = 0.004) and recurrence-free (P = 0.007) survival for patients whose tumors entirely lacked (both biopsies negative in cases where two biopsies were available) expression of GLUT-1 (see Fig. 4). In addition, correlations with poorer overall and recurrence-free survival, respectively, were found for the presence of lymph node metastasis (P = 0.0002 and P = 0.0001), higher pT stage (P = 0.02 and P = 0.0367) and higher FIGO stage (P = 0.0067 and P = 0.0248). When either pT stage or pN stage were included in a multivariate Cox proportional hazards analysis (only applicable in cases treated with surgery), there was no significant independent influence of GLUT-1 expression on prognosis. Only pN stage remained a significant prognostic factor for overall (P = 0.015) and recurrence-free (P = 0.007) survival.

The primary aim of this study was to evaluate the suitability of GLUT-1 as an endogenous marker of tumor hypoxia. The expression of GLUT-1 was analyzed using immunohistochemistry in biopsy specimens taken from oxygenation measurement tracks done with the Eppendorf microsensor system. The fact that both measurements originate from identical tissue microareas is a novel feature of this study. Using this methodology, no correlation of GLUT-1 expression and oxygenation variables (median pO₂, HF 2.5 and HF 5) were found. Several severely hypoxic tissue biopsies showed weak or no expression of GLUT-1, whereas moderate to strong expression was repeatedly found in normoxic specimens. In a recent study, Airley et al. (4) described a weak, albeit statistically significant, correlation of higher GLUT-1 expression in cases with higher values of HF 2.5. This finding may in the first instance be interpreted as being contradictory to our results. On closer examination however, the suitability of GLUT-1 as an endogenous hypoxia marker seems highly questionable from the data of both studies. In agreement with our own results, Airley et al. (4) found no correlation of GLUT-1 expression with HF 5 and the study also does not mention a correlation with the median pO₂. Both studies show that GLUT-1 expression may be absent in a significant amount of severely hypoxic tumors and that well-oxygenated tumors may exhibit very strong expression of GLUT-1. Two recent studies compared GLUT-1 (and carbonic anhydrase IX) expression with the accumulation of the “hypoxia-marker” pimonidazole and found a strong spatial colocalization and correlation between the two variables, concluding that both proteins may be regarded as endogenous hypoxia markers (17, 26). This interpretation is problematic since pimonidazole, as well as another “extrinsic” hypoxia marker, EF5, have been shown not to correlate with the oxygenation status, as directly measured with microelectrodes (27–30). It also has to be kept in mind that prognostic correlations have thus far only been shown for the “snapshot” picture of hypoxia acquired with the Eppendorf microsensor technique, assessing all pathophysiologically relevant types of hypoxia (acute and chronic), provided measurements in necrotic tissue areas can be excluded as done in our study. Similar to our findings, Hedley et al. (31) in a recent study found no correlation between a further HIF-1α target gene, carbonic anhydrase IX, and the oxygenation status measured by microelectrodes in cervical cancers. Applying the same experimental setup as that used in this study, we recently showed that there is also no direct correlation between HIF-1α expression and oxygenation status in squamous cell carcinomas of the uterine cervix (32). HIF-1α data originating from this former study were available for most of the cases reported in the present study and a preliminary analysis revealed no correlation between the two proteins. However, these data are not shown, because the methods for assessment were different and the two markers were not always analyzed in consecutive sections.

Expression of GLUT-1, much like HIF-1α, is induced by a variety of stimuli besides hypoxia. For GLUT-1, established inducing factors are glucose deprivation (33), oncogenic transformation (e.g., overexpression of c-MYC; ref. 34), inhibition of oxidative phosphorylation (35), angiotensin II (in mesangial cells; ref. 36), and osmotic stress (37, 38). According to our interpretation of the data, induction of HIF-1α and subsequent transactivation of GLUT-1 by hypoxia, although undoubtedly present, cannot be selectively identified due to the heterogeneous occurrence of the other abovementioned factors in human cancer specimens.

Another important issue is the prognostic effect of GLUT-1 as a marker of the hypoxic response. In univariate analysis, an improved overall and recurrence-free survival in patients with completely absent GLUT-1 expression (i.e., two negative biopsies) was found. Multivariate Cox regression analysis revealed that both correlations were independent of FIGO stage, clinical tumor size, histologic grading, patient age, and pretherapeutic hemoglobin concentration. Inclusion of pT stage or pN stage into the model, however, abrogated the independent prognostic effect of GLUT-1 expression. The dependency of the prognostic relevance of GLUT-1 expression on T stage and N stage has been described for other tumor entities (e.g., breast carcinoma; ref. 23) and colorectal cancer (16, 22). The only study that evaluated the prognostic effect of GLUT-1 expression in cancers of the uterine cervix found a significant correlation with prognosis only for metastasis-free survival. A possible dependency of this correlation on nodal status could not be analyzed, as only patients treated with radiotherapy were examined (4).

A remarkable finding in the present study is the correlation of GLUT-1 expression with FIGO stage, T stage, and maximum clinical tumor size. Correlations between GLUT-1 expression and tumor size have also been described by others (14, 39–41) and may have important pathophysiologic implications. Because it is known (24) that tumor oxygenation is independent of tumor size, the finding is consistent with a partial GLUT-1 activation by factors other than hypoxia. Because the expression pattern of GLUT-1 shows increasing intensity towards areas of necrosis and with increasing distance from the stroma (containing the microvessels), it is probable that environmental factors (e.g., glucose deprivation) are important for the expression of the protein. Activated oncogenes may have an effect on the degree of activation by these factors. This influence is likely to become more relevant in higher stages and larger tumors, because oncogenic mutations, accumulated during malignant progression, may become more prevalent.

In conclusion, from the data of the present study as well as from recent findings by other groups, the role of GLUT-1 as an endogenous marker of tumor hypoxia is questionable, at least for cancers of the uterine cervix. There is an association of GLUT-1 expression with prognosis, although correlations were not independent, but instead were due to the association of GLUT-1 expression with established factors of dominant prognostic relevance.

We thank Prof. Dr. M.A. Konerding (Department of Anatomy, University of Mainz) for providing the DISKUS image acquisition system, Beate Köhler for excellent technical assistance, and Dr. Debra K. Kelleher for her valuable assistance in preparing this article.