Cancer Research AACR Membership Protein Translation and Cancer
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
Cancer Research 66, 9929, October 15, 2006. doi: 10.1158/0008-5472.CAN-06-0886
© 2006 American Association for Cancer Research
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Raman, V.
Right arrow Articles by Bhujwalla, Z. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Raman, V.
Right arrow Articles by Bhujwalla, Z. M.

Cell, Tumor, and Stem Cell Biology

Characterizing Vascular Parameters in Hypoxic Regions: A Combined Magnetic Resonance and Optical Imaging Study of a Human Prostate Cancer Model

Venu Raman1, Dmitri Artemov1, Arvind P. Pathak1, Paul T. Winnard, Jr.1, Stephen McNutt1, Anna Yudina2, Alexei Bogdanov, Jr.2 and Zaver M. Bhujwalla1

1 The Johns Hopkins University In vivo Cellular Molecular Imaging Center Program, The Russell H. Morgan Department of Radiology and Radiological Science, The Johns Hopkins University School of Medicine, Baltimore, Maryland and 2 Department of Radiology, University of Massachusetts Medical School, Worcester, Massachusetts

Requests for reprints: Zaver M. Bhujwalla, Department of Radiology, The Johns Hopkins University School of Medicine, Room 208C Traylor Building, 720 Rutland Avenue, Baltimore, MD 21205. Phone: 410-955-9698; Fax: 410-614-1948; E-mail: zaver{at}mri.jhu.edu.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
The integration of imaging technologies with the capabilities of genetic engineering has created novel opportunities for understanding and imaging cancer. Here, we have combined vascular magnetic resonance imaging (MRI) and optical imaging to understand the relationship between hypoxia and vascularization in a human prostate cancer model engineered to express enhanced green fluorescent protein (EGFP) under hypoxia. Characterization and validation of EGFP expression under hypoxic conditions was done in culture and in solid tumors in vivo. MRI measurements showed that vascular volume was significantly lower in fluorescing regions. These regions also frequently exhibited high permeability. These data were further supported by the detection of low vessel density in EGFP-positive regions, as determined by the distribution of intravascularly administered, fluorescence-labeled Lycopersicon esculentum lectin in frozen tumor sections. These observations are consistent with the possibility that regions of low vascular volumes are hypoxic, which induces increased expression of functionally active vascular endothelial growth factor, a potent vascular permeability factor. (Cancer Res 2006; 12(20): 9929-36)


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Hypoxic environments are frequently observed in solid tumors and have posed a major problem for successful radiation therapy as well as chemotherapy of cancer for several decades (1, 2). Although hypoxic environments occur in other ischemic diseases, cells in these environments do not contribute to the recurrence of a disease and are generally doomed to die if salvaging efforts fail. In contrast, cancer cells in hypoxic environments adapt to this environment, are resistant to radiation and chemotherapy, and are likely to lead to recurrence of the disease (3).

The combination of molecular imaging strategies together with multimodality imaging has provided novel opportunities to further characterize the tumor microenvironment. Characterization of the vasculature in hypoxic regions is important as it provides further understanding of the tumor microenvironment. Vascular patterns typical of hypoxic regions may be exploited as surrogate markers of hypoxia using noninvasive contrast-enhanced magnetic resonance imaging (MRI), with the associated advantages for radiation and chemotherapy planning. It is also important to know the vascular characteristics of hypoxic regions for drug delivery, especially for therapeutic strategies targeting hypoxic cells (4).

A critical response to hypoxia, in both normal and cancer cells, is the induction of the hypoxia-inducible transcription factor (HIF-1). HIF-1 can act as a transcriptional activator for, among others, vascular endothelial growth factor (VEGF), inducible nitric oxide synthase, heme oxygenase 1, glucose transporter (GLUT1), and the glycolytic enzymes aldolase A, enolase 1, lactate dehydrogenase A, and phosphoglycerate kinase 1 (5). Each of these target genes contains hypoxia response elements (HRE), which include one or more HIF-1–binding sites. Thus, HIF-1 mediates the adaptive response of cells to hypoxia.

In this study, we have exploited the HRE to study the relationship between hypoxia and vascularization using MRI and optical imaging of a human tumor xenograft model derived from tumor cells that stably express an inducible enhanced green fluorescent protein (EGFP), under the regulation of an HRE promoter (6). Validation of the relationship between fluorescence and hypoxia was done in culture as well as in vivo using an Oxylite fiber optic oxygen probe inserted in fluorescing and non-fluorescing regions of the tumor under image guidance. MRI of the macromolecular contrast agent albumin-Gd-diethylenetriaminepentaacetate (albumin-GdDTPA) was done to obtain colocalized maps of vascular volume and vascular permeability (7), after which the EGFP distribution in fresh tissue slices obtained from the imaged slices was determined. In nearly all the tumors examined, a coarse colocalization was observed between regions of high EGFP expression in the fluorescent images and low vascular volume. In several of the tumors, these regions also exhibited high permeability in the MRI maps. The results showed that areas of hypoxia are characterized by low vascular volume, and that high permeability occurred in or around the high EGFP expressing regions. Further validation of the relationship between hypoxia and low vessel density was done using fluorescence-labeled tomato lectin as a marker of perfused blood vessels.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Generation of 5xHRE-EGFP clones in PC-3 cell line. The 5xHRE-EGFP construct (schematic shown in Fig. 1A ) is similar to the construct described by Shibata et al. (6). The promoter construct contains five copies of the HRE of the VEGF gene attached to a minimal cytomegalovirus (CMV) promoter in the pd2EGFP vector (Clontech, Mountain View, CA). To generate a stable cell line with the integrated vector, 5 x 105 PC-3 prostate cancer cells were plated into a 60-mm dish 24 hours before transfection. Duplicate plates were transfected using 1 µg of CMVpd2EGFP (control vector) and 5xHRE-EGFP with LipofectAMINE (Invitrogen, Carlsbad, CA). Approximately 12 hours later, each plate was split into three 100-mm plates with medium containing 500 µg/mL of G418 (Calbiochem, San Diego, CA). Following selection for 2 weeks, large healthy colonies were picked and expanded individually into cell lines.


Figure 1
View larger version (43K):
[in this window]
[in a new window]
  		Figure 1. A, schematic of the construct used to create stably transfected HRE-EGFP PC-3 cells. B, immunoblot evaluation of the expression of EGFP in PC-3-pd2EGFP–transformed cells. Cell lysates were prepared from PC-3-pd2EGFP cells cultured for 24 hours in the presence of 0 to 500 µmol/L CoCl2. Induction of EGFP was highest in protein extracts from cells incubated with 200 to 300 µmol/L CoCl2. Treatment with 500 µmol/L CoCl2 nearly abolished EGFP induction. Approximately 7 µg of total protein (bicinchoninic acid assay) was loaded onto each lane. The blot was probed with monoclonal antibodies to EGFP and an internal loading control, ß-actin, which was detected with a horseradish peroxidase–conjugated rabbit anti-mouse antibody and its chemiluminescent substrate. C, stably transfected PC-3 cells expressing HRE-EGFP under normoxia and following 24 hours treatment with 0 to 500 µmol/L cobalt chloride. D, characterization of EGFP in cells following incubation under hypoxic conditions.

 
SDS-PAGE and immunoblot of PC-3 cell lysates following CoCl2 treatments. After 24 hours of treatment with CoCl2, PC-3 cells were lysed in radioimmunoprecipitation assay buffer + 1.0 mmol/L EDTA solution with 1:200 dilution of protease inhibitor cocktail (P8340, Sigma, St. Louis, MO). Protein concentration of lysates was determined using Bicinchoninic Acid Protein Assay kit (Bio-Rad, Hercules, CA). SDS-PAGE was run on a 12% gel loaded with 7 µg of total protein from each CoCl2 concentration tested. The proteins were blotted onto Hybond-N nitrocellulose (Amersham, Piscataway, NJ) membranes, blocked in 5% skim milk-TBST (TBS Tween) overnight at 4°C, and developed at room temperature with 1:1,000 dilution of monoclonal anti-EGFP (BD Biosciences, San Jose, CA) as well as with a 1:10,000 dilution of monoclonal anti-actin antibody (Sigma).

Characterization of the hypoxic response of cells in culture. Stably transfected 5HRE-pd2EGFP clones of PC-3 cells (PC-3-5HRE-EGFP) were grown in 35-mm plates and incubated with varying concentrations of CoCl2, which mimics hypoxia by stabilizing HIF-1{alpha} (8), for 24 hours. In separate studies, the effect of continuous hypoxia on the fluorescence of the cells was determined using a commercially available controlled atmosphere-humidified culture chamber, placed within a bench top incubator to maintain temperature, and connected to a PROOX model 110 oxygen controller (BioSpherix Ltd., Redfield, NY). Oxygen tension within the chamber was maintained under 1%. For these studies, 106 cells were plated in 100-mm plates and allowed to reach 50% confluency. At the end of each hypoxia time point, regions within the plate were randomly identified and imaged at x20 magnification, using a Nikon inverted microscope equipped with a Nikon Coolpix digital camera (Nikon Instruments, Inc., Melville, NY). After imaging, cells were lysed, and immunoblot assays for cells exposed to hypoxia were done as described earlier, except 2.5 µg of total protein was used per lane with 1:2,000 dilution of monoclonal anti-EGFP (BD Biosciences) and 1:2,500 dilution of monoclonal anti-actin antibody (Sigma).

Tumor studies. Solid tumors were derived from HRE-EGFP PC-3 cells by inoculating 106 cells in 0.05 mL of Hanks Balanced Salt solution (Sigma) s.c. in the right flank of male severe combined immunodeficient mice. Tumor volumes used in the study were 100 to 200 mm3 and developed within 5 weeks of inoculation. No differences in growth rate were observed between wild-type PC-3 cells and HRE-EGFP PC-3 cells or tumors.

Oxygen tension measurements of tumors in vivo. A home-built epifluorescence device was used to detect oxygen tensions in fluorescing regions of the tumor under image guidance. Oxygen tensions in muscle and fluorescing and non-fluorescing regions of the tumor (n = 3) were measured using an OxyLab system (Oxford Optronix, Oxford, United Kingdom) with a fiber optic probe of 250 µm diameter. Mice were anesthetized using a mixture of ketamine (25 mg/kg; Phoenix Scientific, Inc., St. Joseph, MO) and acepromazine (250 mg/kg; Aveco, Phoenix Scientific) diluted in saline and injected i.p. in a volume of 0.05 mL. Anesthetized mice were immobilized by gently taping the animal to a platform to minimize motion, and the probe was inserted at a depth of 2 to 3 mm within the tumor through a 20-G detachable trochar (FastBreak Cannulas, Harvard Apparatus, Holliston, MA). The probe was allowed to stabilize after the trochar was removed, and pO2 measurements were recorded over a period of 20 minutes at each location once the probe readings had stabilized. Probe readings stabilized with 15 to 20 minutes of insertion.

In vivo vascular imaging and imaging EGFP in fresh tissue slices. The mice were anesthetized as described earlier, and the tail vein was catheterized before placing the animal in the spectrometer. Nine tumors were studied with combined MRI and EGFP. All imaging studies were done on a 4.7 T Bruker Avance (Bruker, Billerica, MA) spectrometer, as described in Bhujwalla et al. (7), using a home-built solenoid coil placed around the tumor. Briefly, multislice relaxation rate (1/T1) maps were obtained by a saturation recovery method combined with fast T1 SNAPSHOT-FLASH imaging (flip angle of 10 degrees, echo time of 2 milliseconds). First, an Mo map with a recovery delay of 7 seconds was acquired once. Then, images of four slices (1 mm thick), acquired with an in-plane spatial resolution of 125 µm (128 x 128 matrix, 16 mm field of view, NS = 8), were obtained for three relaxation delays (100 and 500 milliseconds and 1 second). These T1 recovery maps were obtained before i.v. administration of albumin-GdDTPA, which was synthesized by us. A volume of 0.2 mL of 60 mg/mL of albumin-GdDTPA in saline was given at a dose of 500 mg/kg. Then, starting 3 minutes after i.v. injection of albumin-GdDTPA, imaging was repeated every 3 minutes for a total of six times. Quantitative T1 relaxation maps were reconstructed from data sets for three different relaxation times and the Mo data set on a pixel-by-pixel basis. A small glass capillary containing water doped with GdDTPA was used as a spatial reference, together with the orientation of the head and tail of the animal.

At the end of the imaging studies, the animals were sacrificed and 0.5 mL of blood was withdrawn from the inferior vena cava or tail vein, and the T1 of blood was measured. Each tumor was marked for spatial referencing and sectioned to obtain 1-mm-thick slices coarsely matching the MR-imaged slices. Optical imaging of EGFP expression in freshly cut tumor slices was used to visualize hypoxia in relation to the noninvasively obtained MR vascular maps. After recording the EGFP images, slices were fixed in 10% formalin, sectioned, and stained with H&E.

Generation of functional vascular MR maps. Vascular volume and vascular permeability surface area product (PSP) maps were generated from the ratio of {Delta}(1/T1) values in the images to that of blood. The slope of {Delta}(1/T1) ratios versus time in each pixel was used to compute PSP, and the intercept of the line at zero time was used to compute vascular volume. Thus, vascular volumes were corrected for vessel permeability (7). Three-dimensional volume data were processed with an operator-independent computer program that enabled selection, mapping, and display of the regions. Values of vascular volume and PSP were computed for every voxel in the tumor. The routine was written with Interactive Data Language (IDL, Research Systems, Boulder, CO) for UNIX workstations.

Processing strategy for registration and colocalization of MR and EGFP images. All image processing was done using ImageJ v1.34s (freeware for Windows developed by Wayne Rasband at the NIH).3 Before image registration, a binary mask of the entire tumor was created from the magnetization or Mo map. The next step was to register the EGFP image obtained from microscopy, which was considered as the source image, to the MR image that was the target image. To eliminate any bias in subsequent selection and analysis of regions of interest (ROIs), all EGFP fluorescence images were registered to the magnetization or Mo MRI maps rather than the vascular volume (VV) or PSP maps. This was achieved by subjecting the source image to an affine transformation as defined in (9). Briefly, during an affine transformation, a straight line in the source image is mapped to a straight line in the target image while conserving flat angles between lines (i.e., parallel or coincident lines remain parallel or coincident). In two-dimensions, three user-defined landmarks are required in each image to give a complete description of the affine transformation. This mapping is of the form x = {(a11, a12), (a21, a22)}·u + {Delta}u, where one of the three points helps define the translation {Delta}u of the source (u), and the other two points, the projections. After completion of the registration process, ImageJ uses the final position of the source and target landmarks to create a warped image that has the size of the target and contains a distorted version of the source. The warping is such that the landmarks of the source are mapped to those of the target. ImageJ refines the landmarks of the source automatically to minimize the mean-square difference between the target and the warped image using cubic spline interpolation. Warping of the EGFP maps for three of the nine tumors was unsuccessful because of major alterations in the shape of the slice following excision of the tumor. This precluded the identification of anatomic landmarks for warping the images. Hypoxic ROIs were manually identified on the registered version of the EGFP image by picking regions that exhibited fluorescence. The 8-bit signal intensity (mean ± 1 SD) for the EGFP ROIs (n = 6) was 108.69 ± 21.39, and for the background was 71.53 ± 20.06 (n = 6). A two-tailed Aspin-Welch unequal-variance t test showed that this difference was significant (P = 0.011). These regions were, in turn, employed to generate a binary mask of the hypoxic ROIs that we called the "hypoxic" mask. This hypoxic mask was subtracted from the mask of the whole tumor to yield another binary mask representing the remainder of the tumor that we called the "normoxic" mask. The VV and PSP maps were then masked using either the "hypoxic" or "normoxic" masks, and pixels that passed each mask were saved to a spread sheet for further analyses. Because we had no a priori knowledge of the distributions of PSP and VV, the median PSP and VV were computed for the hypoxic and normoxic regions, respectively.

A nonparametric two-tailed Mann-Whitney test was done on these data to determine if there were any significant ({alpha} = 0.05) differences between the VV and PSP of hypoxic and normoxic ROIs, respectively. It should be reiterated that only the coregistered EGFP images were employed to determine which ROIs were hypoxic, without consulting the VV or PSP maps. In essence, all subsequent analyses were blinded to the individual spatial distributions of VV and PSP for each animal.

Tomato lectin staining of blood vessels and apoptosis assay. Carbocyanine near-IR fluorescence dye–labeled Cy5.5-tomato (Lycopersicon esculentum) lectin was prepared by using Cy5.5-mono N-hydroxysuccinimide ester (Amersham Biosciences, Piscataway, NJ) as described by the manufacturer. Fluorescent conjugates were purified using Biospin P30 minicolumns (Bio-Rad). A subset of animals (n = 4) were i.v. injected with 0.2 mL of Cy5.5-lectin (1.25 mg/kg) in saline, at the end of the MRI study, and mice were sacrificed 10 minutes later. These tumors were frozen and thin-sectioned (8-µm sections) around the center (equatorial) part and the edges (10 sections per tumor sample) for vessel density analyses using fluorescence microscopy. Terminal deoxynucleotidyl transferase–mediated nick-end labeling assay (ApopTag kit, Chemicon, Temecula, CA) was done by using Cy3-labeled anti-digoxigenin F(ab')2 fragment (Roche Diagnostics, Indianapolis, IN) for detecting fragmented genomic DNA. Fluorescence was observed and recorded in two channels using a Zeiss Axiovert TV100 microscope equipped with a CoolSnap HQ CCD camera (Photometrics, Tucson, AZ). Color coding and fusion of two fluorescence channels was achieved by using IPLab Spectrum software (Scanalytics, Inc., Fairfax, VA).

Correlative VEGF and EGFP immunofluorescent microscopy. Tumors (n = 3) were excised and placed in cold formalin. Following 24 hours of fixation, each tumor was transferred to a precooled (4°C) solution of 30% sucrose in PBS and placed in 4°C until it sank to the bottom of the flask, indicating adequate dehydration. Tumors were then frozen in liquid N2 after embedding in ornithine carbamyl transferase (Tissue-Tek Optimal Cutting Temperature, Sakura Finetech, Torrance, CA) and immediately cut into 10-µm-thick sections on a cryostat. After rinsing in PBS, endogenous peroxidase activity was blocked by incubating slides with Peroxoblock (Invitrogen/Zymed, Carlsbad, CA) for 2 minutes. After a PBS rinse, slides were blocked with normal donkey serum blocking solution for 45 minutes. After another PBS rinse, slides were incubated overnight at 4°C with 5 µg/mL of mouse anti-EGFP antibody (Clontech). After a PBS wash, slides were incubated with donkey anti-mouse FITC (Invitrogen) for 2 hours. Following a PBS wash, slides were incubated with 15 µg/mL of goat anti-human VEGF polyclonal antibody (R&D Systems, Minneapolis, MN) overnight at 4°C. Slides were then washed again in PBS before being incubated with goat anti-rat Cy3-conjugated secondary antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) for 3 hours at room temperature. Following a final PBS wash, slides were counterstained with 4',6-diamidino-2-phenylindole (DAPI), mounted in an aqueous mounting medium Faramount (DAKO Corp., Carpinteria, CA), and coverslipped. VEGF was identified by red fluorescence, enhanced EGFP (EGFP) by green fluorescence, and cell nuclei by blue fluorescence, respectively. All slides were viewed at x20 using a Nikon ECLIPSE-TS100 microscope (Nikon Instruments) equipped with Plan-Fluor lenses and filters for detecting EGFP/FITC, RFP/TRITC, and DAPI fluorescence.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Correlating hypoxia with EGFP synthesis in stably transfected 5HRE-pd2EGFP clones. Several clones from PC-3 tumor cell lines stably expressed EGFP under the control of five copies of the HRE (5'-CCACAGTGCATACGTGGGCTCCAACAGGTCCTCTT-3') derived from the 5'-flanking region of the human VEGF gene (Fig. 1A).

The clone used in this study was selected based on its robust response to CoCl2 and hypoxia as shown in Fig. 1C and D. CoCl2 concentrations ranging from 0 to 500 µmol/L were tested to determine the optimal concentration for induction. Figure 1B shows immunoblots obtained from cells incubated with CoCl2 for 24 hours probed with anti-EGFP antibody and anti-actin as a loading control. As evident in Fig. 1B, HIF-1 activity was nondetectable under normoxic growth conditions showing the absence of aberrant signaling pathways resulting in stable expression of HIF-1 (10). The immunoblot data show that these cells produced large amounts of EGFP following treatment with CoCl2, with optimal concentrations in the range of 200 to 300 µmol/L. Concentrations of 400 to 500 µmol/L CoCl2 diminished EGFP expression most likely due to toxicity. The stability of 5HRE-pd2EGFP and the response to CoCl2 was studied after passaging cells for 20 generations. The results obtained were identical to the original isolate (data not shown) showing stable expression of the transfected gene that is not lost with the passage of cells in culture.

Representative fluorescent images obtained before and 24 hours after the addition of 0 to 500 µmol/L CoCl2 are shown in Fig. 1C and show the robust expression of EGFP in these cells following treatment. Representative fluorescent images from cells obtained before and time points after hypoxia are shown in Fig. 1D and illustrate the increase in fluorescence within 6 hours of hypoxia and the robust expression by 20 hours of hypoxia. Immunoblots of EGFP corresponding to these images are also shown in Fig. 1D.

In solid tumors in vivo, the image-guided placement of a fiber optic probe in fluorescing and non-fluorescing regions confirmed the results observed in cells. As shown in Fig. 2 , oxygen tensions in two of the fluorescing regions were below the sensitivity of detection of the probe (i.e., at 0 mm Hg). The third measurement, located within a fluorescing region, reported a value of 0.12 mm Hg. Studies done in three animals showed that oxygen tensions in fluorescing regions of the tumor ranged from 0 to 0.6 mm Hg, in non-fluorescing regions from 5 to 40 mm Hg, and in normal muscle tissue from 23 to 50 mm Hg. The mean ± SD of oxygen tensions in fluorescing and non-fluorescing tumor regions were 0.34 ± 0.31 and 18.7 ± 16 mm Hg, respectively. Representative images of colocalized vascular volume and permeability with the corresponding EGFP image and histologic section from the imaged slice of a tumor are shown in Fig. 3A to D . Distinct vascular and fluorescence distributions were detected in the tumors. Low vascular volume regions (see Fig. 3A, arrow) measured by MRI were associated with high fluorescence (Fig. 3C, arrow). High permeability regions (Fig. 3B, arrow) were coincident with these low vascular volume regions and high fluorescence. Histologic evaluation of the H&E-stained section obtained from this imaged slice revealed only a small focus of necrosis (Fig. 3D, arrow). Upon histologic examination, high EGFP expressing regions mostly contained viable cells with only small foci of necrosis. Each EGFP image was used as a mask to obtain vascular volume and permeability values from fluorescing and non-fluorescing regions of each tumor. A representative example of this image analysis for the tumor slice presented in Fig. 3 is shown in Fig. 4A to G . The relationship between the MRI and EGFP data obtained using this analysis is summarized for six tumors in Fig. 5A and B . High EGFP expressing regions were consistently characterized by a significantly lower vascular volume. With this analysis, higher permeability was detected in these regions for three of the six tumors.


Figure 2
View larger version (49K):
[in this window]
[in a new window]
  		Figure 2. Image-guided oxygen tension measurements in an HRE-EGFP PC-3 tumor. Measurements were made in normal tissue and fluorescing and non-fluorescing regions using an Oxylite probe.

 

Figure 3
View larger version (37K):
[in this window]
[in a new window]
  		Figure 3. Maps of (A) VV and (B) PSP obtained from a central slice of an HRE-EGFP PC-3 tumor (180 mm3). VV ranged from 0 to 344 µL/g and PSP from 0 to 24 µL/g min. C, fluorescent microscopy of a fresh tissue slice obtained from the imaged slice, using a Nikon TS100-F microscope (x1 objective) with a wavelength of 512 nm. D, H&E-stained, 5-µm-thick section from the central MR-imaged slice. The region exhibiting EGFP consisted of viable cells. The less dense staining in the upper part of the section is due to uneven sectioning. The only area of dying cells was in a small necrotic focus (black arrow).

 

Figure 4
View larger version (31K):
[in this window]
[in a new window]
  		Figure 4. Functional colormaps of (A) the PSP, (B) the VV, and (C) EGFP fluorescence image of a freshly excised tumor section from optical microscopy, warped to match the magnetization (or Mo) map obtained from MR images. Also shown on each colormap are hypoxic (cyan squares) and normoxic (red squares) ROIs. D, histogram comparing the distribution of PSP from hypoxic and normoxic ROIs in (A), showing that hypoxic regions tended to exhibit elevated PSP values compared with normoxic regions. E, histogram of the distribution of VV from the hypoxic and normoxic ROIs in (B), showing that hypoxic regions exhibited lower VV values compared with normoxic regions. F, histogram comparing the distribution of EGFP fluorescence intensities from the hypoxic and normoxic ROIs in (C), showing that hypoxic regions fluoresced more intensely than normoxic regions. G, a composite RGB image composed from the channels shown in (A) to (C), showing clearly distinguishable hypoxic regions characterized by low VV and elevated PSP (yellowish) and normoxic regions characterized by lower PSP and higher VV (cobalt blue). ROIs were directly drawn on the warped EGFP maps (C) and the exact same ROIs analyzed on the MR maps without any a priori inspection of the latter.

 

Figure 5
View larger version (21K):
[in this window]
[in a new window]
  		Figure 5. Summary of VV and PSP values obtained from fluorescing and non-fluorescing regions of tumors. Box and whisker plots were created to graphically display the relationship between the VV and PSP for normoxic and hypoxic ROIs. In these plots, a box is drawn with its top at the third quartile and bottom at the first quartile of the respective distribution. The length of the box is a visual representation of the interquartile range; that is, it represents 50% of the data. The location of the midpoint or median of the distribution is indicated by a horizontal line in the box, and the straight lines or "whiskers" above and below the 25th and 75th percentiles extend 1.5 times the interquartile range. A, box and whisker plots comparing the VV from six of the nine animals (warping of the EGFP maps for three animals was unsuccessful) for normoxic and hypoxic ROIs, respectively. Inset, median VV of hypoxic ROIs was significantly lower than that of the normoxic ROIs for all animals imaged. *, P < 0.05, with the two-sided Mann-Whitney U test. B, box and whisker plots comparing the PSP from the same six animals for normoxic and hypoxic ROIs, respectively. Whereas there was no significant difference in the median PSP of hypoxic and normoxic ROIs (see inset), three of the six animals did exhibit a trend of elevated PSP in hypoxic ROIs compared with normoxic ROIs (brackets).

 
A representative example of a fluorescent image obtained from a frozen section of a tumor following i.v. administration of Cy5.5-lectin is shown in Fig. 6A . Most of the lectin-stained vessels (color coded in blue) can be observed in EGFP-negative regions that constituted 24 ± 2% of total section area (n = 30 sections processed, values represent mean ± SD). Consistent with the MRI results, green fluorescing regions (EGFP positive) contained significantly lower numbers of blood vessels. The average microvascular densities was 16 ± 2 vessels/mm2 in EGFP-negative, normoxic regions versus 2 ± 1 vessels/mm2 in EGFP-positive, hypoxic areas. No significant population of apoptotic cells was detected in either of the regions.


Figure 6
View larger version (23K):
[in this window]
[in a new window]
  		Figure 6. A, representative section showing Cy5.5-tomato lectin–mediated visualization of blood vessels (blue) and hypoxic EGFP-positive areas (green). Minimal colocalization of the lectin stain was observed with HRE-driven EGFP expression. Some apoptotic cells are also visible within the hypoxic regions (terminal deoxynucleotidyl transferase–mediated nick-end labeling positive, color coded in red). B to D, patterns of spatial distribution of anti-human VEGF (red) and anti-EGFP (green) in tumor sections. B, example of region showing spatial overlap of VEGF and EGFP expression. C, example of region with low EGFP and low VEGF expression. D, example of region exhibiting complementary spatial localization of VEGF and EGFP.

 
Immunofluorescent staining of tumor sections for VEGF and EGFP showed three spatial patterns as shown in Fig. 6B to D. Increased VEGF expression colocalized with regions expressing high EGFP (Fig. 6B), and low VEGF expression was observed in regions with low EGFP expression (Fig. 6C). In addition, there were some regions where high VEGF expression was observed adjacent to regions with high EGFP expression (Fig. 6D).


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Most studies characterizing the vasculature of hypoxic regions of tumors have been done using excised specimens (1). Thus, the in vivo functional characterization of these vessels is either unknown or can only be assessed for a terminal time point. The combined in vivo functional MRI and EGFP images in this study show that hypoxic regions in the prostate cancer xenograft studied were consistently characterized by low vascular volume. This low vascular volume detected in hypoxic regions by MRI was independently corroborated by microscopy of cryosections of lectin-perfused tumors. Significantly lower vessel densities were measured in EGFP-expressing regions following i.v. administration of fluorescently labeled lectin. Clusters of dead cells were also observed in the center of larger hypoxic tumors surrounded by EGFP-positive cells. These data also support observations made in previous studies relating vascular volume, derived from dynamic contrast-enhanced (DCE) MRI of GdDTPA, to oxygen tensions, derived from 19F MRI of the relaxation rate of perfluoro-15-crown-ether emulsion injected in the tumor (11). In these studies, pO2 values predicted using Krogh's cylinder model and the DCE-measured vascular volume values correlated well with pO2 values measured with 19F MRI. Because intravascular administration of perfluorocarbon emulsions results in significant reticuloendothelial uptake, 19F MRI following direct i.t. injection of perfluorocabons is being used successfully to measure tumor oxygen tension in vivo (12). Egeland et al. (13) have also found a strong correlation between number of voxels with low extraction fraction of GdDTPA measured by DCE MRI and the fraction of hypoxic cells.

Several of the high EGFP regions also exhibited high permeability. This elevated permeability in and around hypoxic areas can be explained by the stabilization of the HIF-1{alpha} protein by hypoxia, which in turn activates downstream genes like the potent permeability factor VEGF (14). Immunofluorescently stained tumor sections did show colocalization of VEGF and EGFP expression. However, unlike the significant decrease of vascular volume observed in high EGFP expressing regions, the increase of vascular permeability was not consistently observed in all six tumors. Based on the time course of the evolution of EGFP expression of cells under hypoxia, one can infer that the hypoxia detected in vivo was extant for at least 6 hours or longer. It is possible that for some of these tumors the MRI studies may have been done either before or after the response of VEGF to hypoxia. It is also possible that the action of VEGF may be most intense at the periphery of hypoxic regions, as observed by us in this study. The pattern of high VEGF expression adjacent to high EGFP expression in some regions and colocalized with high EGFP expression in other regions supports this possibility. Because the entire area of high EGFP expression was used as a mask for analysis, this may have resulted in volume averaging in these areas. In addition, as recent studies have shown, temporal variations in oxygen concentrations as well as the effects of extracellular pH may play a role in the variability of VEGF expression (15, 16).

The relationship between hypoxia and vascular variables described here was for studies done in a human prostate cancer xenograft transplanted s.c. We are currently evaluating this relationship in the orthotopic site using the same prostate cancer model.

A limitation of using a promoter such as HRE in combination with reporter genes to detect oxygen concentration is that it is currently not possible to quantify the oxygen tension in vivo. However, characterization of the cell line in culture, as well as the image-guided insertion of the Oxylite probe into fluorescing regions of solid tumors in vivo showed a robust expression of EGFP under hypoxic conditions. Using a similar construct, Shibata et al. (6) and Cao et al. (17) have also shown strong reporter expression under hypoxic conditions in culture. Similarly, Serganova et al. (18) have used the HRE enhancer region of the Erythropoietin gene to detect hypoxia with positron emission tomography (PET) and EGFP expression. The use of PET allowed longitudinal imaging of HIF-1–specific reporter gene expression under hypoxic conditions. Further validation of the hypoxia reporting ability of such a biological reporter system was done by Cao et al. (17) in tumors stained with pimonidazole, a validated hypoxic marker (19). In the studies by Cao et al., 98% and 94%, respectively, of the EGFP expressing areas in HCT116 and 4T1 xenograft tumors colocalized with pimonidazole staining. These results support the visualization of EGFP under the control of HRE as a functional biomarker to detect tumor hypoxia. Although there have been reports that EGFP protein levels decrease under anoxic conditions (20), such studies were done using a minimal promoter devoid of any HIF-1{alpha}–binding sites. The use of an inducible promoter such as HRE will produce significantly larger quantities of EGFP under hypoxic conditions that may override the detection limits seen by Coralli et al. (20).

The role of HIF-1{alpha} as the primary regulator of gene expression under hypoxic conditions is well documented (5, 14). However, another caveat of using the HRE system in tumors is the variation of normal HIF-1{alpha} expression levels in cancer cell lines where the expression can be independent of oxygen tension (10). This dysregulation is brought about by several pathways. For example, insulin growth factor-1 can stimulate HIF-1{alpha} expression through a posttranscriptional mechanism (21). In addition, cell lines that express mutated VHL can prevent HIF-1{alpha} from degrading and promote HIF-1{alpha} subunit accumulation (10). More recently, it was shown that the proangiogenic factor interleukin-8 can stimulate VEGF expression in HIF-1{alpha}–deficient colon cancer cells (22). However, the tight regulation of HIF-1{alpha} observed in our PC-3 cancer cells showed the absence of any aberrant signaling pathways in these cells.

In conclusion, the most likely explanation for our observations is that low vascular volume resulted in hypoxia, which in turn increased VEGF expression in these regions, leading to increased vascular permeability. These data are also consistent with the data obtained by Cao et al. (17) where hypoxia stimulated intensive tumor angiogenesis. Similarly, recent observations that VEGF independently predicts the efficacy of postoperative radiotherapy in node-negative breast cancer patients may be related to hypoxia-driven up-regulation of VEGF (23). These results also have implications for cancer therapy because, in contrast to hypoxic regions with leaky vessels, viable oxygenated and well-vascularized regions of solid tumors will not be readily accessible to macromolecular therapeutic agents such as monoclonal antibodies. However, if the circulation time of the agent in the blood is long, macromolecular therapeutic agents may accumulate in hypoxic regions despite the low vascular volume. Hypoxic cancer cells that have survived radiation and chemotherapy reoxygenate following debulking from radiation or chemotherapy to reestablish the tumor. To avoid this vicious cycle of hypoxia, drug and radiation resistance, and survival of cancer cells, hypoxic and well-oxygenated cells should be targeted and eliminated simultaneously. The differential in vessel permeability of hypoxic and non-hypoxic regions may be exploited to design agents to selectively target these regions within the tumor.


   Acknowledgments
 
Grant support: NIH grants R01 CA73850 and P50 CA103175.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Gary Cromwell for transplanting the tumors and maintaining the cell lines, Yelena Mironchik for assistance with cloning and evaluating the cells, and Dr. Ellen Ackerstaff for assistance with the measurement of oxygen tension of cells in culture.


   Footnotes
 
3 http://rsb.info.nih.gov/ij/. Back

Received 3/13/06. Revised 7/22/06. Accepted 8/22/06.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Harrison L, Blackwell K. Hypoxia and anemia: factors in decreased sensitivity to radiation therapy and chemotherapy? Oncologist 2004;9:31–40.[Abstract/Free Full Text]
  2. Thomlinson RH, Gray LH. The histological structure of some human lung cancers and the possible implications for radiotherapy. Br J Cancer 1955;9:539–49.[Medline]
  3. Brizel DM, Scully SP, Harrelson JM, et al. Tumor oxygenation predicts for the likelihood of distant metastases in human soft tissue sarcoma. Cancer Res 1996;56:941–3.[Abstract/Free Full Text]
  4. Jordan BF, Runquist M, Raghunand N, et al. Dynamic contrast-enhanced and diffusion MRI show rapid and dramatic changes in tumor microenvironment in response to inhibition of HIF-1{alpha} using PX-478. Neoplasia 2005;7:475–85.[CrossRef][Medline]
  5. Greijer AE, van der Groep P, Kemming D, et al. Up-regulation of gene expression by hypoxia is mediated predominantly by hypoxia-inducible factor 1 (HIF-1). J Pathol 2005;206:291–304.[CrossRef][Medline]
  6. Shibata T, Giaccia AJ, Brown JM. Development of a hypoxia-responsive vector for tumor-specific gene therapy. Gene Ther 2000;7:493–8.[CrossRef][Medline]
  7. Bhujwalla ZM, Artemov D, Natarajan K, Ackerstaff E, Solaiyappan M. Vascular differences detected by MRI for metastatic versus nonmetastatic breast and prostate cancer xenografts. Neoplasia 2001;3:143–53.[CrossRef][Medline]
  8. Yuan Y, Hilliard G, Ferguson T, Millhorn DE. Cobalt inhibits the interaction between hypoxia-inducible factor-{alpha} and von Hippel-Lindau protein by direct binding to hypoxia-inducible factor-{alpha}. J Biol Chem 2003;278:15911–6. Epub 2003 Feb 26.[Abstract/Free Full Text]
  9. Thevenaz P, Ruttiman UE, Unser M. A pyramid approach to subpixel registration based on intensity. IEEE Trans Image Process 1998;7:27–41.[Medline]
  10. Atkins DJ, Gingert C, Justenhoven C, et al. Concomitant deregulation of HIF1{alpha} and cell cycle proteins in VHL-mutated renal cell carcinomas. Virchows Arch 2005;447:634–42. Epub 2005 Jul 1.[CrossRef][Medline]
  11. Wang Z, Su MY, Nalcioglu O. Applications of dynamic contrast enhanced MRI in oncology: measurement of tumor oxygen tension. Technol Cancer Res Treat 2002;1:29–38.[Medline]
  12. Zhao D, Jiang L, Mason RP. Measuring changes in tumor oxygenation. Methods Enzymol 2004;386:378–418.[Medline]
  13. Egeland TA, Gaustad JV, Vestvik IK, Benjaminsen IC, Mathiesen B, Rofstad EK. Assessment of fraction of radiobiologically hypoxic cells in human melanoma xenografts by dynamic contrast-enhanced MRI. Magn Reson Med 2006;55:874–82.[CrossRef][Medline]
  14. Brahimi-Horn MC, Pouyssegur J. The hypoxia-inducible factor and tumor progression along the angiogenic pathway. Int Rev Cytol 2005;242:157–213.[Medline]
  15. Lanzen J, Braun RD, Klitzman B, Brizel D, Secomb TW, Dewhirst MW. Direct demonstration of instabilities in oxygen concentrations within the extravascular compartment of an experimental tumor. Cancer Res 2006;66:2219–23.[Abstract/Free Full Text]
  16. Goerges AL, Nugent MA. Regulation of vascular endothelial growth factor binding and activity by extracellular pH. J Biol Chem 2003;278:19518–25. Epub 2003 Mar 11.[Abstract/Free Full Text]
  17. Cao Y, Li CY, Moeller BJ, et al. Observation of incipient tumor angiogenesis that is independent of hypoxia and hypoxia inducible factor-1 activation. Cancer Res 2005;65:5498–505.[Abstract/Free Full Text]
  18. Serganova I, Doubrovin M, Vider J, et al. Molecular imaging of temporal dynamics and spatial heterogeneity of hypoxia-inducible factor-1 signal transduction activity in tumors in living mice. Cancer Res 2004;64:6101–8.[Abstract/Free Full Text]
  19. Raleigh JA, Chou SC, Arteel GE, Horsman MR. Comparisons among pimonidazole binding, oxygen electrode measurements, and radiation response in C3H mouse tumors. Radiat Res 1999;151:580–9.[Medline]
  20. Coralli C, Cemazar M, Kanthou C, Tozer GM, Dachs GU. Limitations of the reporter green fluorescent protein under simulated tumor conditions. Cancer Res 2001;61:4784–90.[Abstract/Free Full Text]
  21. Treins C, Giorgetti-Peraldi S, Murdaca J, Monthouel-Kartmann MN, Van Obberghen E. Regulation of hypoxia-inducible factor (HIF)-1 activity and expression of HIF hydroxylases in response to insulin-like growth factor I. Mol Endocrinol 2005;19:1304–17. Epub 2005 Feb 3.[Abstract/Free Full Text]
  22. Mizukami Y, Jo WS, Duerr EM, et al. Induction of interleukin-8 preserves the angiogenic response in HIF-1{alpha}-deficient colon cancer cells. Nat Med 2005;11:992–7. Epub 2005 Aug 28.[Medline]
  23. Manders P, Sweep FC, Tjan-Heijnen VC, et al. Vascular endothelial growth factor independently predicts the efficacy of postoperative radiotherapy in node-negative breast cancer patients. Clin Cancer Res 2003;9:6363–70.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 Cancer Res.Home page
M.-F. Penet, A. P. Pathak, V. Raman, P. Ballesteros, D. Artemov, and Z. M. Bhujwalla
Noninvasive Multiparametric Imaging of Metastasis-Permissive Microenvironments in a Human Prostate Cancer Xenograft
Cancer Res., November 15, 2009; 69(22): 8822 - 8829.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Res.Home page
B. Krishnamachary, K. Glunde, F. Wildes, N. Mori, T. Takagi, V. Raman, and Z. M. Bhujwalla
Noninvasive Detection of Lentiviral-Mediated Choline Kinase Targeting in a Human Breast Cancer Xenograft
Cancer Res., April 15, 2009; 69(8): 3464 - 3471.
[Abstract] [Full Text] [PDF]

Home page
 JNMHome page
M.-F. Penet, K. Glunde, M. A. Jacobs, A. P. Pathak, and Z. M. Bhujwalla
Molecular and Functional MRI of the Tumor Microenvironment
J. Nucl. Med., May 1, 2008; 49(5): 687 - 690.
[Full Text] [PDF]

Home page
 aacredbookHome page
Z. M Bhujwalla
Molecular Imaging of the Tumor Microenvironment
Am. Assoc. Cancer Res. Educ. Book, April 12, 2008; 2008(1): 519 - 522.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Res.Home page
K. Glunde, T. Shah, P. T. Winnard Jr., V. Raman, T. Takagi, F. Vesuna, D. Artemov, and Z. M. Bhujwalla
Hypoxia Regulates Choline Kinase Expression through Hypoxia-Inducible Factor-1{alpha} Signaling in a Human Prostate Cancer Model
Cancer Res., January 1, 2008; 68(1): 172 - 180.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Raman, V.
Right arrow Articles by Bhujwalla, Z. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Raman, V.
Right arrow Articles by Bhujwalla, Z. M.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online