Abstract
Hypoxia is a common feature of cancer and a prognostic factor for many types of cancer. 18F-fluoromisonidazole (18F-FMISO) PET can detect tumor hypoxia noninvasively. Hypoxia-inducible factor-1 (HIF-1) is a key player in the transcriptional response to low oxygen tension in many types of cancer. Its activity is mainly dependent on the stability and modification of HIF-1α, which is a composition of HIF-1. However, it is unclear whether 18F-FMISO PET can identify HIF-1α expression in oral squamous cell carcinoma (OSCC). The present study was performed to elucidate the correlation between 18F-FMISO PET findings and HIF-1α expression in OSCC. Methods: Twenty-three patients (age range, 42–84 y; 15 men, 8 women) with OSCC were enrolled in this study. The T-stages of cancer were T1 in 1 patient, T2 in 9, T3 in 2, and T4a in 11. The N-stages were N0 in 13 patients, N1 in 5, and N2 in 5. Each patient underwent 18F-FMISO and 18F-FDG PET before surgery, and the maximum standardized uptake value (SUVmax) of both PET studies was measured. HIF-1α expression in the operation materials was evaluated by immunohistochemical staining. The SUVmax of both PET studies and HIF-1α findings were compared statistically. Results: 18F-FMISO PET detected uptake in the primary site in 14 of the 23 patients (61%). The median SUVmax of 18F-FMISO and 18F-FDG PET in the primary site was 1.83 (range, 0.8–2.7) and 16.5 (range, 1.0–32.3), respectively. There was a weak significant correlation between 18F-FMISO and 18F-FDG PET SUVmax (P = 0.02, r = 0.48). HIF-1α expression was clearly detected in 11 of the 23 patients (48%). The 18F-FMISO PET SUVmax was significantly higher in the HIF-1α–positive cases than in the HIF-1α–negative cases (median, 2.1; range, 1.5–2.4, vs. median, 1.6; range, 0.8–2.0, respectively) (P = 0.002). However, there were no significant correlations between 18F-FDG PET SUVmax and HIF-1α expression (median, 21.8; range, 7.7–29.1 vs. 1.0–32.2) (P = 0.06). Conclusion: 18F-FMISO uptake in the primary site of OSCC indicates a hypoxic environment with HIF-1α expression.
Hypoxia is a common feature of cancer and thus is a prognostic factor for many types of cancer (1,2). Clinically, the prognosis of cancer with low oxygenation levels is poor, and there is strong evidence that this is due to the effects of hypoxia on therapy resistance and malignant progression (2). In particular, it is clear that hypoxia is a negative factor in the treatment of head and neck cancers, reducing the chance of cure (1). However, it is difficult to determine the conditions of hypoxia in solid tumors.
18F-FDG PET is frequently used for the diagnosis and evaluation of treatment outcomes for several tumors. Multiple radiotracers have been developed for hypoxia imaging (1). 18F-fluoromisonidazole (18F-FMISO) PET is a promising noninvasive method of measuring hypoxia (3–6). This method is sensitive to the presence of hypoxia in viable cells and can cover the entire region of interest (7,8). Our recent study demonstrated high reproducibility of tumor hypoxia evaluated by 18F-FMISO PET for head and neck cancer (9).
Hypoxia achieves many of its effects through the activation of the transcription factor hypoxia-inducible factor-1 (HIF-1) (10). HIF-1 is recognized as a key player in the transcriptional response to hypoxia (11–13). HIF-1 is a heterodimer composed of an α-subunit (HIF-1α) and a β-subunit (HIF-1β), and its activity is mainly dependent on the stability and modification of HIF-1α, the expression of which is highly regulated (10,12). In the case of HIF-1α, its synthesis is regulated via O2-independent mechanisms, whereas its degradation is regulated primarily via O2-dependent mechanisms (10). The heterodimer HIF-1 binds to its cognate enhancer sequence, the hypoxia-responsive element, and induces the expression of various genes responsible for the adaptation of cellular metabolism to hypoxia (the switch from oxidative respiration to anoxic respiration) (14), escape from hypoxia (invasion and metastasis of cancer cells) (15), and improvement of hypoxia (angiogenesis) (10,16). Because HIF-1 activity depends on HIF-1α expression level (17), the visualization of HIF-1α expression provides useful information on the tumor microenvironment (17).
Thus, HIF-1α overexpression may affect the prognosis of cancer patients. HIF-1α expression in oral squamous cell carcinoma (OSCC) has been shown in several recent studies (11,13,18–21). HIF-1α expression has been shown to correlate with poor prognosis and is an aggressive index of OSCC (19,22,23). Although both 18F-FMISO PET and HIF-1α are important factors for hypoxia evaluation in cancer, there have been no previous reports regarding the correlation between 18F-FMISO PET findings and HIF-1α expression in patients with OSCC. Therefore, the present study was performed to elucidate the correlation between 18F-FMISO PET uptake and HIF-1α expression in patients with OSCC. Imaging and targeting of hypoxic cells using 18F-FMISO PET and HIF-1α expression are becoming more important (17).
MATERIALS AND METHODS
Patients
Twenty-three consecutive patients (15 men, 8 women; median age, 67 y; age range, 42–84 y) with untreated primary OSCC undergoing medical examination and radical surgery for OSCC between October 2009 and October 2011 in our department at Hokkaido University Hospital were enrolled in this study (Table 1). None of the patients received palliative treatment. The primary tumor sites were the tongue (n = 5), upper gingiva (n = 6), lower gingiva (n = 9), buccal mucosa (n = 1), and oral floor (n = 2). One tumor (4%) was classified as T1, 9 (39%) as T2, 2 (9%) as T3, and 11 (48%) as T4a. The N-classifications were N0 in 13 patients (57%), N1 in 5 (21%), and N2 in 5 (22%) (24).
Patients in This Study
Intraoperative materials were evaluated histopathologically by hematoxylin and eosin staining and examined by 2 specialists in oral pathology masked to the specimen origin. The degree of histologic differentiation was determined in accordance with the criteria of the World Health Organization published in 1997. Of the 23 tumors, 9 (39%) were classified as grade 1, 6 (26%) as grade 2, 4 (17%) as grade 3, and 4 (17%) as unclear (25). The histologic mode of invasion of cancer was classified according to the Yamamoto and Kohama (YK) classification system (26), with YK-1 tumors having well-defined borders and YK-4 tumors having diffuse growth or invasion. Of the 23 tumors, 1 (4%) was classified as YK-1, 6 (26%) as YK-2, 11 (48%) as YK-3, 1 (4%) as YK-4, and 4 as unclear.
The protocol of this study was approved by the Institutional Review Board of Hokkaido University Hospital (2009). All the patients signed a written informed consent.
18F-FMISO and 18F-FDG PET Imaging
After giving their written informed consent, all the patients underwent 18F-FMISO and 18F-FDG PET imaging before surgery. For 18F-FMISO PET, 10-min static PET images were acquired in 3-dimensional mode using an ET/CT scanner (True Point Biograph 64 with true V option; Asahi-Siemens) 4 h after the injection of 400 MBq of 18F-FMISO. For 18F-FDG PET, the same scanning protocol as for 18F-FMISO PET was applied. At 1 h after the injection of 18F-FDG (4.5 MBq/kg), the energy window was 425–650 keV, the transaxial field of view was 216 mm, and the reconstruction matrix was 168 × 168. Images were reconstructed using the iterative TrueX reconstruction method, which included partial-volume correction. The spatial resolution was 6.7 mm after reconstruction (9). The detailed methods of 18F-FMISO PET and high reproducibility of tumor hypoxia evaluated by 18F-FMISO PET for head and neck cancer in our institution have been described in our recent study (9).
For the semiquantitative evaluation of both 18F-FMISO and 18F-FDG uptake in the primary tumor, the highest uptake in the tumor was estimated using a maximum standardized uptake value (SUVmax).
PET images were evaluated by specialists in nuclear medicine masked to the clinical information for each case. The patients underwent 18F-FMISO and 18F-FDG PET on different days. Therefore, the nuclear medicine specialists also evaluated each image independently on different days. When necessary, they referred to enhanced CT images to confirm the tumor region. The average and median of the interval of both PET studies were 3.1 ± 3.7 d and 1 d (range, 1–16 d), respectively.
Immunohistochemical Study for HIF-1α
The immunohistochemical detection of HIF-1α was conducted using formalin-fixed paraffin-embedded tissue sections. The sections were deparaffinized, rehydrated, and immersed in 3% hydrogen peroxide in distilled water for 5 min to block endogenous peroxidase activity. They were incubated with a primary mouse monoclonal antibody to HIF-1α (sc-13515, 1:100 dilution; Santa Cruz Biotechnology) overnight at 4°C. The epitope of this antibody is mapped within amino acids 329–530 of HIF-1α of human origin, and the antibody has no cross-reactivity to HIF-2α or HIF-3α. After being washed with phosphate-buffered saline, the sections were stained with Simple Stain MAX–PO (Nichirei Biosciences) for 30 min and visualized with Envision Plus Kits/HRP (Dako) at room temperature. The peroxidase reaction products were developed with 3,3′-diaminobenzidine, and the sections were counterstained with hematoxylin. Negative controls in which the primary antibody was replaced with normal rat IgG were run with each specimen. HIF-1α positivity was evaluated by counting positive cells among 500–1,000 tumor cells at a magnification of ×200 in 3 different areas. We set the cutoff value of HIF-1α–positive cells at 5% of the positively stained cells (21). This work was performed by 2 of the authors who were masked to the identities of the patients from whom the specimens had been obtained.
Statistical Analysis
A Spearman correlation coefficient was used to compare the relationship between the SUVmax of 18F-FMISO and of 18F-FDG PET. The Mann–Whitney U test was used to compare the SUVmax of 18F-FMISO and of 18F-FDG PET in patients with and without HIF-1α expression. In the same way, the Mann–Whitney U test, χ2 test, and logistic regression analysis were used to compare SUVmax or HIF-1α expression and other factors, including patient age, T- and N-classifications, degree of histologic differentiation, and histologic mode of invasion. The factors assessed included sex, T-classification (T1 + 2 vs. T3 + 4), N-classification (N0 vs. N1 + 2), degree of histologic differentiation (grade 1 vs. grade 2 + 3), and mode of invasion (YK-1 + 2 vs. YK-3 + 4). All statistical analyses were performed using Stat View J-5.0 statistical software (Abacus Concepts). In all analyses, a P value of less than 0.05 was taken to indicate statistical significance.
RESULTS
18F-FMISO and 18F-FDG PET Imaging
18F-FMISO and 18F-FDG PET detected uptake in primary sites of OSCC in 14 (61%) and 22 (96%) of the 23 patients, respectively (Fig. 1). Only 1 patient (patient 21) showed no 18F-FDG uptake. The median SUVmax of 18F-FMISO and 18F-FDG PET in the primary site was 1.83 (range, 0.8–2.7) and 16.5 (range, 1.0–32.3), respectively (Table 1; Figs. 1–3⇓⇓).
Relationship of SUVmax between 18F-FMISO and 18F-FDG PET. There was weak significant correlation in SUVmax between 18F-FMISO and 18F-FDG PET by Spearman correlation coefficient (P = 0.02, r = 0.48). Median SUVmax of 18F-FMISO and 18F-FDG PET was 1.83 and 16.5, respectively. ○ = results of cases without 18F-FMISO uptake.
Clinical findings, 18F-FMISO and 18F-FDG PET images, and immunohistochemical staining of HIF-1α for patient 2. (A) Case of left upper gingival squamous cell carcinoma. Tumor can be seen in left upper maxillary region. (B) For 18F-FMISO PET, definite uptake was detected in left maxillary primary site of cancer (SUVmax, 1.57). (C) For 18F-FDG PET, definite uptake was also seen in left maxillary primary site of cancer (SUVmax, 12.2). (D) HIF-1α expression was detected in nucleus and cytoplasm of cancer cells by immunohistochemical analysis.
Clinical findings, 18F-FMISO and 18F-FDG PET images, and immunohistochemical staining of HIF-1α for patient 1. (A) Case of left upper gingival SCC. Tumor can be seen in left upper maxillary region. (B) For 18F-FMISO PET, definite uptake was not detected in primary site of cancer (SUVmax, 1.39). (C) For 18F-FDG PET, definite uptake was detected in primary site of cancer (SUVmax, 30.1). (D) Weak HIF-1α expression was observed in cancer cells by immunohistochemical analysis.
The weak correlation between 18F-FMISO and 18F-FDG SUVmax was statistically significant (P = 0.02, r = 0.48) (Fig. 1). However, if we delete 2 extreme data points of 18F-FMISO SUVmax (patients 18 and 19) from the analysis, the correlation is not significant (P = 0.06, r = 0.41; data not shown).
Immunohistochemical Staining for HIF-1α and PET Imaging
The immunohistochemical staining of HIF-1α was performed on 23 cases of squamous cell carcinoma in the oral region. Positive signals were observed in 11 of the 23 cases (48%) (Table 1; Figs. 2 and 3). HIF-1α was detected in the cytoplasm and nucleus of cancer cells (Fig. 2). No obvious signals of HIF-1α were observed in normal stromal cells around the cancer parenchyma and negative-control specimens (data not shown).
The prevalence of HIF-1α positivity was significantly higher in patients with 18F-FMISO uptake in primary sites than in those without uptake (10/14 vs. 1/9; P < 0.005) (Table 2). The SUVmax of 18F-FMISO PET was significantly higher in patients positive for HIF-1α than in those negative for HIF-1α (median, 2.1; range, 1.5–2.4, vs. median, 1.6; range, 0.8–2.0 [P = 0.002]) (Fig. 4). If we delete 2 extreme data points of 18F-FMISO SUVmax (patients 18 and 19) from the analysis, the correlation is still significant (P = 0.006; data not shown). However, there were no significant correlations between the SUVmax of 18F-FDG PET and HIF-1α (median, 21.8; range, 7.7–29.1, vs. median, 8.4; range, 1.0–32.2 [P = 0.06]) (Fig. 5).
Relationship Between 18F-FMISO Uptake and HIF-1α Expression
Relationship between 18F-FMISO SUVmax and HIF-1α expression. SUVmax of 18F-FMISO was significantly higher in HIF-1α–positive cases than in HIF-1α–negative cases (P = 0.002, Mann–Whitney U test). Bars indicate median of SUVmax.
Relationship between 18F-FDG SUVmax and HIF-1α expression. There was no significant correlation between 18F-FDG PET SUVmax and HIF-1α expression (P = 0.06, Mann–Whitney U test). Bars indicate median of SUVmax.
As for the correlations of SUVmax with clinical findings, there was no significant correlation between 18F-FMISO PET SUVmax and age of the patients (P = 0.29) (data not shown), T-classification (P > 0.99) (Fig. 6), N-classification (P = 0.14) (data not shown), degree of histologic differentiation (P = 0.51) (data not shown), or histologic mode of invasion (P = 0.52) (data not shown).
Relationship between 18F-FMISO SUV max and T-classification. There was no significant correlation between 18F-FMISO PET SUVmax and T-classification (P > 0.99, Mann–Whitney U test). Bars indicate median of SUVmax.
DISCUSSION
The results of the present study indicated a significant correlation between 18F-FMISO uptake and HIF-1α expression in patients with OSCC. Our results confirmed that 18F-FMISO uptake indicates the presence of hypoxic areas in a tumor with a high degree of confidence.
Hypoxia is characteristic of solid tumors arising from a less ordered vasculature and necrosis (20,27,28). Hypoxia can also induce major changes in gene expression, thereby enhancing the metastatic ability of tumor cells (1). Therefore, intratumoral hypoxia is one of the most important mechanisms promoting tumor aggressiveness, metastasis, and poor prognosis (11). Consequently, the noninvasive imaging of HIF-1α expression and activity is of great interest because it may be useful for predicting cancer prognosis (29). Thus, the significant association observed between 18F-FMISO uptake and HIF-1α expression in patients with OSCC is an important finding of the present study.
18F-FMISO PET, showing specific uptake in hypoxic tissue, is a promising noninvasive method of measuring hypoxia (4–6). Eschmann et al. (3) performed 18F-FMISO PET for 14 cases of advanced head and neck cancers and reported that before radiotherapy all the patients showed enhanced 18F-FMISO uptake throughout the entire tumor area. Moreover, they found a correlation between the SUVmax of 18F-FMISO PET and tumor recurrence, because both are linked to hypoxia. In the present study, 14 of the 23 patients (61%) showed 18F-FMISO uptake in primary sites of OSCC. However, 9 patients showed no definite 18F-FMISO uptake in such primary sites. This observation may be explained by differences between the characteristics of the patients; in the study by Eschmann et al. (3), only 4 of the 14 patients had OSCC (29%). In oral lesions, metal artifacts due to dental restoration may also decrease image quality. 18F-FMISO PET images typically have low signal-to-noise characteristics, and metal artifacts may easily decrease image quality (7).
Some recent studies have shown that HIF-1–active regions are heterogeneous (30,31). Kudo et al. revealed that HIF-1α expression is not ubiquitous but heterogeneous and is small in tumors (30). On the other hand, Fillies et al. (21) found that in most cases, HIF-1α expression did not vary much among the tumor cores of each individual patient, indicating a rather homogeneous HIF-1α expression throughout the tumor in patients with oral floor cancer. A recent study has indicated that HIF-1 expression supports the adaptation of hypoxic cancer cells by inducing the expression of genes related to glucose metabolism and glucose transport, producing angiogenic and growth factors that help improve the nutritional environment, and preventing death by apoptosis (17). These chain-of-survival actions are linked to the malignancy change of the entire cancer (17). These findings suggest that the HIF-1 action affects the malignancy of the entire tumor, regardless of whether or not the HIF-1α expression is heterogeneous.
Although there have been some immunohistochemical studies of HIF-1α (21,32,33), the literature gives no uniform recommendation for a cutoff point of HIF-1α expression. Beasley et al. (34) used a cutoff value of greater than 1% in oral cancer. Other authors used cutoff values between 1% and 5% (21,33). In this study, we set the cutoff value at 5% of the HIF-1α–positive cells in accordance with the study of Fillies et al. (21), who demonstrated that the 5% HIF-1α expression threshold can be used to distinguish 2 different populations with significant statistical differences in survival prognosis.
The direct measurement of tumor partial oxygen pressure (pO2) can be made using an Eppendorf electrode. Tumor hypoxia is usually identified by pO2 values less than or equal to 10 mm Hg; whereas normal tissues have pO2 values of 24–66 mm Hg (1,2). 18F-FMISO uptake increases only when the pO2 value falls below 2–3 mm Hg (29). A previous immunohistochemical study showed that HIF-1α is more frequently observed adjacent to blood vessels than in pimonidazole-positive regions (29) and that the pO2 values in HIF-1–active regions are approximately 10–15 mm Hg (29,35). This result suggests that 18F-FMISO uptake reflects more severe hypoxic conditions than the region of HIF-1α expression. Moreover, this finding suggests that there are some discrepancies between the regions of 18F-FMISO uptake and HIF-1α expression. However, in the present study, we could not compare definite locations between 18F-FMISO uptake and HIF-1α expression, because the specialists in oral pathology and nuclear medicine evaluated HIF-1α expression and PET images independently.
In the present study, there was no significant correlation between 18F-FDG uptake and HIF-1α expression. This finding suggests that 18F-FMISO uptake and HIF-1α expression indicate not only increased glucose metabolism, but also hypoxia.
The major limitation of this study was the small patient population used. Moreover, the distributions of both 18F-FMISO uptake and HIF-1α expression in cancer could not be accurately determined. Further studies using a larger number of cases are required to address these issues.
In addition, in the present study, we could demonstrate HIF-1α expression in the hypoxic region of a tumor evaluated by 18F-FMISO PET. However, we could not elucidate the function and activity of HIF-1, a heterodimer composed of HIF-1α and HIF-1β, in OSCC in the present study. The correlation between HIF-1 activity and 18F-FMISO uptake is important for understanding the microenvironment of OSCC and for planning a strategy for OSCC treatment. Examining this correlation is the next target of our research.
HIF-1 was initially identified because of its response to low O2 concentrations, but it is now apparent that HIF-1 can be regulated by other factors such as oncogene activation or the loss of tumor suppression (36). The expression of oncogenes such as HRAS-V12 leads to the accumulation of HIF-1α under both normoxic and hypoxic conditions (36). To our regret, we could not provide information on the possibility of HIF-1α expression under normoxic conditions in OSCC. This is another limitation of this study.
CONCLUSION
We demonstrated that there is a significant relationship between definite 18F-FMISO uptake and HIF-1α expression, which are both key features of hypoxia, in primary sites of cancer in patients with OSCC. In the future, 18F-FMISO PET may be used in the decision-making process regarding treatment strategies.
DISCLOSURE
The costs of publication of this article were defrayed in part by the payment of page charges. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734. This study was partially supported by a grant-in-aid for scientific research (2010–2011: 22592203). No other potential conflict of interest relevant to this article was reported.
Footnotes
Published online May 22, 2013.
- © 2013 by the Society of Nuclear Medicine and Molecular Imaging, Inc.
REFERENCES
- Received for publication September 28, 2012.
- Accepted for publication January 31, 2013.
