Longitudinal and multimodal in vivo imaging of tumor hypoxia and its downstream molecular events

Abstract

Tumor hypoxia and the hypoxia-inducible factors (HIFs) play a central role in the development of cancer. To study the relationship between tumor growth, tumor hypoxia, the stabilization of HIF-1α, and HIF transcriptional activity, we have established an in vivo imaging tool that allows longitudinal and noninvasive monitoring of these processes in a mouse C51 allograft tumor model. We used positron emission tomography (PET) with the hypoxia-sensitive tracer [¹⁸F]-fluoromisonidazole (FMISO) to measure tumor hypoxia over 14 days. Stabilization of HIF-1α and HIF transcriptional activity were assessed by bioluminescence imaging using the reporter constructs HIF-1α-luciferase and hypoxia response element-luciferase, respectively, stably expressed in C51 cells. Interestingly, we did not observe any major change in the level of tumor hypoxia throughout the observation period whereas HIF-1α levels and HIF activity showed drastic temporal variations. When comparing the readouts as a function of time we found a good correlation between HIF-1α levels and HIF activity. In contrast, there was no significant correlation between the [¹⁸F]-FMISO PET and HIF readouts. The tool developed in this work allows for the longitudinal study of tumor hypoxia and HIF-1α in cancer in an individual animal and will be of value when monitoring the efficacy of therapeutical interventions targeting the HIF pathway.

As a result of their high proliferation rates and abnormal growth patterns, tumor cells outgrow their vascular supply territories, causing intratumoral areas of hypoxia, a hallmark of most solid tumors (1). Tumor hypoxia is accompanied by the stabilization of hypoxia-inducible factors (HIFs), oxygen-regulated transcription factors that mediate the adaptation of cells to decreased oxygen availability (2). Typically, accumulation of HIFs in solid tumors is associated with poorer patient prognosis, more aggressive tumor phenotypes, and an increased metastatic potential (3). Furthermore, HIF activity has been implied to decrease the effectiveness of chemotherapy and radiation therapy (4, 5). Inhibiting tumor progression by targeting the HIF signaling pathway in combination with other treatments thus appears to offer attractive options in the development of cancer therapeutics (6).

HIFs are heterodimeric transcription factors consisting of a β-subunit and 1 of 3 α-subunits (HIF-1α, HIF-2α, and HIF-3α). Whereas the HIF-β subunit is not affected by changes in oxygen availability, HIF-α subunits are subject to proteasomal degradation under normoxic conditions, but become stabilized during hypoxia. Details regarding the mechanism of HIF-α oxygen-dependent regulation have been described (7).

In view of the important role of tumor hypoxia and HIFs in cancer development, great efforts are being taken to further understand and elucidate their contribution to tumor growth. Noninvasive in vivo imaging techniques are of particular interest in this context because they allow longitudinal monitoring of cellular and molecular processes in the same subject. Here, we report on the development of a hypoxia imaging tool, which combines nuclear imaging strategies and optical imaging of reporter genes to study the relationship between tumor hypoxia and its molecular consequences in a mouse tumor allograft model. We stably transfected the murine colon cancer cell line C51 (8) with either a HIF-1α-luciferase fusion construct (pcDNA3.1-mHIF-1α-luciferase) or a hypoxia response element (HRE)-driven luciferase reporter gene [pGL(P2P)95bp]. After evaluation of these reporter constructs in vitro, tumor allografts were established by s.c. inoculation of nude mice with the stably transfected C51 reporter cells. In longitudinal in vivo experiments we daily assessed (i) tumor hypoxia by quantitatively measuring the uptake of [¹⁸F]-fluoromisonidazole (FMISO), (ii) the levels of HIF-1α (HIF-1 stability), and (iii) HIF transcriptional activity by using bioluminescence imaging over up to 14 days.

Results

Assessment of Tumor Hypoxia by Measuring Uptake of [¹⁸F]-FMISO.

Accumulation of [¹⁸F]-FMISO in the C51 allograft tumors was analyzed from days 6 to 14 after tumor inoculation in 2 groups of 6 animals by using an interleaved observation scheme. Two groups of mice had to be used for the longitudinal study because of experimental restrictions (number of anesthesia episodes per animal). Uptake of [¹⁸F]-FMISO tended to increase with increasing tumor volumes, implying a higher degree of hypoxia, as illustrated by transversal positron emission tomography (PET) images of a representative animal recorded longitudinally on days 7, 10, 12, and 14 (Fig. 1A). After inoculation, tumors grew rapidly, with fastest growth rates between days 11 and 13 (Fig. 1B). In [¹⁸F]-FMISO experiments, tissue is considered hypoxic when the activity displays a tissue-to-background ratio ≥1.4. Taking muscle as reference tissue, we calculated the tumor-to-muscle retention ratio (TMRR) for all hypoxic tumor voxels, i.e., all tumor voxels displaying a TMRR ≥1.4 (9). In both groups of animals, the TMRR of the hypoxic tumor fraction (hypoxic TMRR) showed no major change in signal intensity over the entire observation period. Only a slight increase was observed between days 9 and 14 (Fig. 1C). On average, at least 50% of all tumor voxels displayed TMRR values indicative of hypoxia, clearly showing that the C51 tumor allografts were highly hypoxic irrespective of the absolute tumor mass.

Fig. 1.

Assessment of tumor hypoxia by measuring the uptake of [¹⁸F]-FMISO. (A) Four consecutive PET images representing cross-sections of the same tumor-bearing mouse on days 7, 10, 12, and 14 are shown. Tumor regions are indicated by the black dotted line and show increased uptake of [¹⁸F]-FMISO. (B) Tumor volume as function of time relative to the values measured on day 6; mean ± SEM (n = 6). (C) Average tumor-to-muscle retention signal (hypoxic TMRR) for the hypoxic fraction in the tumor defined by a TMRR ≥1.4 (10); mean ± SEM (black), individual animals (gray).

In Vivo Monitoring of the Stabilization of HIF-1α Using Bioluminescence Imaging.

To gain further insight into the complex and dynamic events driving the stabilization of HIF in cancer we generated a reporter construct to monitor HIF-1α stability by means of bioluminescence imaging. HIF-1α was C-terminally fused to firefly luciferase via a short linker sequence (pcDNA3.1-mHIF1α-luciferase; Fig. 2A). To analyze the functionality and localization of the fusion protein mouse embryonic fibroblasts (MEFs) were transiently transfected with the reporter construct, followed by treatment with dimethyloxalylglycine (DMOG), a chemical substance that mimicks hypoxia by inhibiting the degradation of HIF-α subunits (10). Immunofluorescence staining of these cells using an antibody against firefly luciferase revealed that the fusion protein was localized in the cell nucleus as expected (Fig. 2B). To examine the oxygen-dependent regulation of the fusion construct, MEF wild-type or HIF-1α-deficient cells (MEF-Hif1a^+/+ and MEF-Hif1a^−/−, respectively) were cotransfected with both the HIF-1α-luciferase and a reporter plasmid (pH3SVB) that drives the expression of β-galactosidase from a HIF-responsive promoter. Whereas β-galactosidase activity was low in the absence of DMOG for HIF-1α-luciferase-transfected MEF-Hif1a^+/+ and MEF-Hif1a^−/−, the reporter activity increased in both cell lines after DMOG treatment (Fig. 2C). In cells transfected only with the β-galactosidase reporter plasmid, an induction of reporter activity after DMOG treatment could be observed exclusively in the MEF-Hif1a^+/+, reflecting endogenous HIF-1α activity. Next, we stably expressed the reporter construct in C51 colon cancer cells. Single cell clones were screened for the highest induction in response to hypoxia and high luciferase counts. To study HIF-1α stabilization in vivo, a highly expressing clone was injected s.c. into the necks of nude mice. C51 cells stably expressing firefly luciferase from a CMV promoter (pcDNA3.1-luciferase) were used as a reference; these cells should yield an almost constant light output per viable tumor cell and could therefore be used for normalizing the HIF-1α-luciferase-induced bioluminescent signals. Bioluminescence imaging after the administration of d-luciferin revealed an increase in photon output in animals carrying HIF-1α-luciferase-expressing tumors as the tumors grew (Fig. 2D). Two groups of 5 animals were analyzed by using an interleaved scheme starting on days 5 or 6. Tumor growth was similar in both the control luciferase and the HIF-1α-luciferase tumors (Fig. S1). Luciferase activity was estimated by normalizing (total photon counts) measured in both HIF-1α-luciferase and luciferase control tumors to the theoretical counts calculated for luciferase control tumors of the same volume (for further details on the normalization procedure see Fig. S2 and SI Text). Normalized intensities are expressed relative to the values measured on day 5 for both HIF-1α-luciferase and luciferase tumors (Fig. 2E). The normalization procedure compensates to some extent for variations in signal intensity associated with changes in tumor volume, blood absorption, or necrosis. Intensities in Fig. 2D represent total photon counts measured in the tumors and hence differ from the normalized values shown in Fig. 2E. The signal detected in the luciferase control tumors deviates from the value 1 mainly for reasons of biological variability: (i) clonal differences may lead to varying levels of luciferase expression and (ii) tumor shapes may differ, both factors influencing the signal intensity measured on the surface.

Fig. 2.

In vivo monitoring of the stabilization of HIF-1α using bioluminescence imaging. (A) pcDNA3.1-mHIF-1α-luciferase (HIF-1α-luc) and pcDNA3.1-luciferase (luciferase) reporter constructs. (B) Immunofluorescence staining of firefly luciferase in MEF cells transiently transfected with pcDNA3.1-mHIF-1α-luciferase. Nuclei were stained with DAPI. (Magnification: × 300.) (C) Assessment of oxygen-dependent regulation and transcriptional activity of the HIF-1α-luciferase fusion construct. MEF-Hif1a^+/+ or MEF-Hif1a^−/− were cotransfected with pcDNA3.1 mHIF-1α-luciferase, and the reporter plasmid pH3SVB, which drives the expression of β-galactosidase from a HIF responsive promoter. β-Galactosidase activity was assessed. Data are representative of 2 independent experiments. (D) Bioluminescent images of a mouse carrying a HIF-1α-luciferase C51 tumor in the neck. Four consecutive images of the same animal are shown. Color bar indicates total photon counts. (E) Normalized bioluminescence photon counts (mean ± SEM) relative to day 5 values for both the HIF1α-luciferase and the luciferase control tumors. Values indicated by * significantly (P ≤ 0.05) differ from values measured in control groups.

Measuring HIF Transcriptional Activity Using in Vivo Bioluminescence Imaging.

A reporter construct for in vivo monitoring of HIF transcriptional activity was generated to investigate the relationship between HIF-1α stability and HIF transcriptional activity. A DNA element comprising the minimal HRE from the human PHD2 promoter was used to drive expression of the firefly luciferase gene [pGL(P2P)95bp; Fig. 3A]. Hypoxic induction of luciferase activity was assessed in MEF-Hif1a^+/+ and MEF-Hif1a^−/− cells transiently expressing the HRE reporter construct. Normalized photon counts were increased under hypoxic conditions only in cells expressing HIF-1α, confirming the key role of this factor in transcriptional activation of the reporter gene (Fig. 3C). C51 cells were stably transfected with the reporter construct, and single cell clones were screened for high luciferase counts and high sensitivity to oxygen levels. Cellular extracts of a selected clone were analyzed for HIF-1α and luciferase expression by immunoblotting (Fig. 3B). In the same clone, bioluminescence increased with decreasing oxygen levels, revealing the graduated induction of HRE-luciferase (Fig. 3D). Allograft tumors were established in nude mice with the selected HRE-luciferase clone and C51 control cells stably expressing firefly luciferase from an SV40 promoter (pGL3prom). Two groups consisting of 5 animals each were used for the measurements between days 5 and 15. Tumor growth rates were comparable in the HRE-luciferase and the luciferase control tumors (see Fig. S1). Normalized in vivo luciferase activity was calculated as described above by dividing the actual photon counts measured for both the HRE-luciferase and the luciferase control tumors by the estimation value for the photon counts of a luciferase control tumor of the same volume. Normalized photon counts in control C51 tumors did not significantly change during the observation period. However, tumors expressing the HRE reporter construct displayed an up to 14-fold signal increase ≈day 10 when compared with the initial measurement on day 5 (Fig. 3E). For a more detailed description of the normalization procedure please see SI Text.

Fig. 3.

Measurement of HIF transcriptional activity reporter. (A) Reporter construct pGL(P2P)95bp (HRE-luciferase). (B) Immunoblot analysis of stably transfected C51 cells cultured at different oxygen concentrations or treated with DMOG. Unspec. refers to an unspecific band on the Western blots used as loading control. (C) In vitro luciferase activity measurement. MEF-Hif1a^+/+ and MEF-Hif1a^−/− cells were transiently transfected with the HRE-luciferase reporter construct and exposed to normoxic or hypoxic conditions. (D) Luciferase activity (photon counts) normalized to total amount of protein in HRE-luciferase C51 reporter cells exposed to varying oxygen concentrations. (E) Normalized bioluminescence photon counts (mean ± SEM) relative to day 5 values for both the HRE-luciferase and the luciferase control tumors. * indicate significant differences (P ≤ 0.05).

In Vitro Evaluation of in Vivo Reporter Activity.

For confirmation of the reporter activities measured in vivo, sections from tumors isolated on days 7, 8, 9, 11 and 14 were analyzed by immunohistochemistry. The sections were stained for the nitroimidazole derivative pimonidazole (11), HIF-1α, GLUT1 [a typical HIF-1α target gene (12)], and CD31 as a marker for endothelial cells, i.e., angiogenesis (13) (Fig. 4A). Higher-magnification images obtained from tumor sections harvested on day 14 demonstrate the cellular localization of the antigens detected (Fig. 4B). Whereas pimonidazole adducts were observed in both the cytoplasm and nucleus, HIF-1α expression was predominantly nuclear. As expected, GLUT1 and CD31 were localized to the plasma membrane. These in vitro stainings supported the results of the in vivo measurements: HIF-1α and GLUT1 protein levels and the hypoxia marker pimonidazole showed a decrease from days 7 to 8 and then recovered to increase throughout the remaining observation period. In contrast, there was only little CD31 detected on day 7. CD31 expression started to increase from day 8, concomitant with the onset of tumor growth. Interestingly, the staining patterns of pimonidazole, HIF-1α, and GLUT1 were often comparable but not identical. In particular, large tumors displayed significant discrepancies in the spatial correlation of hypoxia markers: regions with high HIF-1α and GLUT1 expression but hardly any pimonidazole staining were observed, implying that the HIF cascade in tumors may not be exclusively activated by low oxygen tensions. Furthermore, there were well-vascularized tumor areas that nevertheless showed intensive pimonidazole staining, indicating that the vascularization in those regions may not (yet) be fully functional (days 11 and 14). This interpretation was confirmed by Hoechst perfusion experiments and immunofluorescence stainings: regions that accumulated pimonidazole and were also positive for CD31 tended to be weakly perfused (Fig. 4C). However, we also found regions positive for CD31, pimonidazole, and Hoechst, indicating that in these areas cells were hypoxic because of increased metabolic activity or there was plasma flow with only low intravascular oxygen content.

Fig. 4.

Immunohistochemical analyses of tumor sections. (A) Immunohistochemical stainings of tumor sections extracted on days 7, 8, 9, 11, and 14. Images of whole tissue sections are shown. (Magnification: × 0.8.) (B) Tumor regions of a section from a tumor isolated on day 14 are shown with high magnification to confirm the subcellular localization of the detected antigens. (Scale bars: 20 μm for H&E, pimonidazole, HIF-1α, and GLUT1 stainings; 50 μm for the CD31 image.) (C) Immunofluorescence stainings. (i) Pimonidazole staining. (ii) CD31 staining. (iii) Hoechst 33342 (perfusion marker). (iv) Overlay of i–iii. (Scale bar: 100 μm.)

Comparison of in Vivo Tumor Hypoxia, HIF Stability, and HIF Activity Measurements.

We then compared the readout for tumor hypoxia with the stabilization of HIF-1α and HIF transcriptional activity in the C51 allograft tumor model over 14 days (Fig. 5). Tumor hypoxia as assessed by measuring the average TMRR in the hypoxic tumor fraction, showed only a slight increase in tracer uptake between days 6 and 14. In contrast, the HIF-1α stability reporter signal (HIF-1α-luc) peaked around days 9 and 10 for both groups, decreased until day 12, and then reached a plateau. For the HIF activity readout, maximum photon counts were observed around day 10 in both groups of animals, i.e., slightly delayed when compared with the HIF-1α stability reporter signal; toward the end of the observation period (day 12 and later) the activity readout significantly decreased similar to the HIF-1α signal (Fig. 5A). Quantitative correlations of the various hypoxia-related signals over time (Fig. 5 B–D) revealed a reasonable correlation between HIF-1α stability and HIF activity signals with Spearman r = 0.6364 and P = 0.05: the higher the HIF-1α stability, the more HIF transcriptional activity was observed. However, only weak or no correlations were observed between the luciferase reporters readouts and the degree of hypoxia as measured by [¹⁸F]-FMISO PET (Spearman r = −0.5357, P = 0.24 for HIF-1α-luciferase and Spearman r = 0.2143, P = 0.62 for HRE-luciferase, respectively).

Fig. 5.

Comparisons of in vivo tumor hypoxia, HIF-1α stability, and HIF activity measurements. (A) Tumor hypoxia as assessed by [¹⁸F]-FMISO PET, HIF-1α stability, and HIF activity readouts as a function of time. HIF-1α stability and HIF activity reporter values were normalized to the values measured on day 5 (left y axis). Tumor hypoxia is given by the hypoxic TMRR (right y axis). For each readout, mean ± SEM values of the 2 groups measured are displayed. (B–D) Quantitative correlations of the different hypoxia readouts. (B) Spearman r = 0.6364, P = 0.05. (C) Spearman r = −0.5357, P = 0.24. (D) Spearman r = 0.2143, P = 0.62. For each readout, mean ± SEM values of the 2 groups measured are displayed.

Discussion

By combining nuclear and bioluminescence imaging approaches we have generated a tool for longitudinal assessment of tumor hypoxia, HIF-1α stability, and HIF activity in a mouse allograft model. This multimodal imaging approach allows investigating the relationship between tumor hypoxia and HIF signaling, which up to now had been only poorly understood. It may also be used for monitoring therapeutic interventions targeting the HIF pathway in a semiquantitative fashion. Remarkably, the level of overall tumor hypoxia only slightly increased with tumor growth over 14 days. However, the HIF-related readouts indicated dramatic changes in HIF-1α levels and HIF activity in the early phase of tumor development, before the onset of massive tumor growth. Toward the end of the observation period, when tumors have reached a volume of ≈1 cm³, these readouts displayed a drastic decrease in signal intensity. Decreased HIF activity, mediated by a HIF-induced negative feedback mechanism, has also been observed in in vitro experiments when exposing cells to chronic hypoxia conditions (14). To which extent, if at all, the decrease in HIF stability and HIF activity observed in our in vivo model is regulated by such a negative feedback mechanism remains to be investigated. Even though direct comparison of an allograft model and spontaneously arising tumors in cancer patients is difficult, variability in HIF activity is likely to also occur in the clinical situation. Because HIFs play an important role in mediating resistance to chemotherapy and radiation therapy, our results, in combination with the study of HIF activity in other tumor models, may allow for the future identification of a time window for most effective treatment (15).

Interestingly, a correlation (P = 0.05) was found only between the HIF-1α stability and the HIF activity readout, whereas this was not the case when comparing HIF-1α stability or HIF activity to the [¹⁸F]-FMISO PET readout (P = 0.24 and 0.62, respectively). This observation was supported by in vitro immunohistochemical stainings for pimonidazole, HIF-1α, and the HIF target GLUT1 in tumor sections: even though it was possible to identify regions with comparable distribution patterns, the overlay between pimonidazole and HIF-1α or HIF target proteins was generally poor. This finding is in line with earlier studies demonstrating that there is no significant correlation in the degree of hypoxia and expression of HIF and its target proteins in human or xenograft tumor sections (16–18). Notably, oncogenic signaling pathways such as PI3K or MAPK signaling in response to activation by oncogenic Ras have been shown to activate the HIF pathway independently of tumor oxygenation (3). At least in part this may account for the discrepancies observed between the hypoxia and HIF readouts. Alternatively, transient changes in oxygen concentration, which cause the accumulation of [¹⁸F]-FMISO, but may be detected with the HIF readouts only if of sufficient duration, might account for the missing correlation observed in our study. To investigate whether and to what extent C51 tumors would undergo such acute, transient changes in tumor oxygenation, we performed dynamic PET scans to assess the uptake of [¹⁸F]-FMISO over 4 h (Fig. S3 a and b). We did not observe significant temporo-spatial fluctuations in [¹⁸F]-FMISO activity pattern during this time window, indicating that there were no major changes in tumor hypoxia over 4 h in the C51 allograft model. These results were in agreement with experiments involving sequential injection (1-h delay) of 2 hypoxia markers, CCl-103F and pimonidazole (19), in C51 tumor-bearing mice (Fig. S3c). Immunofluorescence analysis of tumor sections from these animals did not reveal any profound changes in the overall distribution of the 2 hypoxia markers. We concluded that transient changes in tumor hypoxia occurring at pO₂ ≤ 10 mm Hg are unlikely to account for the lack in correlation between the hypoxia and HIF-related imaging readouts in our study. However, we cannot exclude that tumor oxygenation transiently changes at pO₂ levels ≥10 mm Hg. These fluctuations would not be detected with bioreductive marker molecules, but could nevertheless lead to the activation of the HIF system (20).

Even though HIF-1α stability and HIF activity readouts were clearly correlated, discrepancies were observed also between these readouts, which is not surprising considering the probable temporal delay between the activation of the 2 reporters. Moreover, the HIF activity construct used is driven by the HRE isolated from the human PHD2 promoter. Although it has been shown that this enzyme is induced predominantly by HIF-1 and not HIF-2 (21) and our in vitro results indicate a strong regulation of the reporter construct by HIF-1, we cannot rule out some activation by HIF-2 or other unknown transcription factors, although the use of a minimized regulatory element should have alleviated such interplay.

In summary, we have developed a multimodal imaging strategy for studying the molecular events induced by tumor hypoxia in a time-resolved manner. Our findings revealed significant discrepancies between [¹⁸F]-FMISO PET and the HIF-1α stability or HIF transcriptional activity readout, which has to be further investigated in view of the importance of the PET approach in clinical tumor diagnostics. The HIF signaling pathway has, besides other pathways, emerged as an attractive target in the development of anticancer drugs. The imaging tool proposed in this work might support these developments by enabling visualization of mechanistic aspects of drug action.

Materials and Methods

Plasmid Constructions.

A short linker sequence was used to C-terminally fuse HIF-1α to firefly luciferase, generating pcDNA3.1 HIF-1α-luciferase. The firefly luciferase reporter vector pGL(P2P)95bp (HRE-luciferase) is driven by a truncation of the previously described human PHD2 promoter (22) and was constructed by placing synthetic oligonucleotides encompassing 90 nt of the core region of the HIF binding site into pGL3 basic (Promega). To generate the pcDNA3.1-luciferase control plasmid, HIF-1α was excised from the pcDNA3.1 HIF-1α-luciferase construct and the vector was subsequently recircularized. The β-galactosidase reporter vector pH3SVB is a subversion of pH3SVL (23) where a lacZ ORF replaces the firefly luciferase gene. β-Galactosidase is expressed under the control of a minimal SV40 promoter flanked by 3 HIF-responsive elements from the human transferrin gene (24).

Generation of Stably Transfected Cell Lines.

Four stable cell lines were generated by cotransfecting mouse colon carcinoma C51 cells with either pcDNA3.1-mHIF-1α-luciferase, pcDNA3.1-luciferase, pGL3prom (Promega), or the pGL(P2P)95bp reporter plasmid, the latter two in combination with neomycin resistance gene-containing pcDNA3.1 in a ratio of 10:1. Stable transfectants were selected by adding G418 (400 μg/mL). Resistant clones were isolated by limited dilution and in the case of pcDNA3.1-mHIF-1α-luciferase and pGL(P2P)95bp they were screened for (i) good oxygen-dependent regulation and (ii) high absolute luciferase photon counts by using luciferase assays. pcDNA3.1-luciferase and pGL3prom luciferase clones were screened only for high luciferase activity.

In Vivo Allograft Tumor Models.

All animal protocols were approved by the Cantonal Veterinary Office in Zurich (129/2007 XIMO_Y2). To establish allograft tumors, we injected 1 × 10⁶ reporter cells into the neck of 8-week-old BALB/c nude mice (Charles River) that were maintained under optimized hygienic conditions. The termination criteria were reached and animals had to be euthanized when tumors showed a volume of 2 cm³. Caliper measurements allowed determination of tumor length and width. From these parameters tumor volumes were calculated by using the formula: tumor volume = (length × width²)/2 .

[¹⁸F]-FMISO PET Experiment.

The radiosynthesis of [¹⁸F]-FMISO was carried out according to the 2-step procedure reported by Lim and Berridge (25). The total synthesis time was ≈120 min, and radiochemical purity was >99% as assayed by HPLC. Specific radioactivities obtained immediately after the synthesis were always >100 GBq/μmol. PET experiments were performed on the 16-module variant of the quad-HIDAC tomograph (Oxford Positron Systems) with performance characteristics as described (26). Animals were lightly restrained and injected with 5–20 MBq of the radiotracer (100–120 μL per injection) via a lateral tail vein. Animals were anesthetized with isoflurane (Abbott) in an air/oxygen mixture at 80 min after injection and monitored as described (27). PET data were acquired in list-mode from 90 to 120 min after injection and reconstructed in a single time frame with a voxel size of 1 mm and a matrix size of 120 × 120 × 200 mm. Regions of interest (ROIs) were manually defined by using the dedicated software PMOD (PMOD Technologies). ROIs were drawn for the whole tumor on all coronal planes containing tumor tissue yielding a volume of interest. Reference tissue ROIs were drawn on 5–10 subsequent coronal planes containing muscle tissue at the contralateral forelimb. The quantification of [¹⁸F]-FMISO uptake was based on the TMRR. This ratio was calculated for all hypoxic tumor voxels that were determined in analogy to the method described by Koh et al. (9): according to their definition a tumor voxel with a TMRR ≥1.4 defines the presence of hypoxia. Additionally, the fractional hypoxic volume of the tumors was computed that is represented by the percentage of hypoxic voxels (with a TMRR ≥1.4) of all voxels within a tumor volume of interest (VOI). For visual inspection and comparison of [¹⁸F]-FMISO tumor uptake PET images were normalized to the injected dose per g of body weight.

Bioluminescence Imaging.

Mice were gas-anesthetized by using 3% isoflurane (MINRAD) and oxygen as a carrier gas. Each mouse was given an i.p. injection of 100 μL of luciferin in PBS (15 mg/mL; Caliper Life Sciences). Ten minutes later, the animals were placed in a light-tight chamber equipped with a charge-coupled device imaging camera (IVIS 100; Xenogen). Photons were collected between 5 and 300 s depending on the reporter line that was analyzed: (i) HIF-1α-luciferase, 300 s, (ii) HRE-luciferase, 60 s, (iii) pcDNA3.1-luciferase, 120 s, and (iv) pGL3prom-luciferase, 10 s. Images were analyzed with Living Image software (Xenogen) and IGOR image analysis software (Xenogen). Total photon counts were determined by drawing an ROI around the peak of photon emission. The border of a ROI was formed by those pixels whose signal intensity was 5% of the maximal signal in the ROI. To correct for the loss of signal associated with bigger tumor volumes in the oxygen-regulated HIF-1α-luciferase and the HRE-luciferase tumors, we normalized total counts from those tumors to pcDNA3.1-luciferase or pGL3prom-luciferase control tumor counts, respectively.

For more detailed information regarding standard methods used in this study please refer to SI Text.