Search tips
Search criteria 


Logo of neoplasiaGuide for AuthorsAbout this journalExplore this journal
Neoplasia. 2003 July; 5(4): 330–338.
PMCID: PMC1502420

Functional Response of Tumor Vasculature to PaCO2: Determination of Total and Microvascular Blood Volume by MRI


In order to identify differences in functional activity, we compared the reactivity of glioma vasculature and the native cerebral vasculature to both dilate and constrict in response to altered PaCO2. Gliomas were generated by unilateral implantation of U87MGdEGFR human glioma tumor cells into the striatum of adult female athymic rats. Relative changes in total and microvascular cerebral blood volume were determined by steady state contrast agent-enhanced magnetic resonance imaging for transitions from normocarbia to hypercarbia and hypocarbia. Although hypercarbia induced a significant increase in both total and microvascular blood volume in normal brain and glioma, reactivity of glioma vasculature was significantly blunted in comparison to normal striatum; glioma total CBV increased by 0.6±0.1% / mm Hg CO2 whereas normal striatum increased by 1.5±0.2%/ mm Hg CO2, (P < .0001, group t-test). Reactivity of microvascular blood volume was also significantly blunted. In contrast, hypocarbia decreased both total and microvascular blood volumes more in glioma than in normal striatum. These results indicate that cerebral blood vessels derived by tumor-directed angiogenesis do retain reactivity to CO2. Furthermore, reduced reactivity of tumor vessels to a single physiological perturbation, such as hypercarbia, should not be construed as a generalized reduction of functional activity of the tumor vascular bed.

Keywords: glioma, vasculature, CO2 reactivity, cerebral blood volume, MRI


In 1971, Folkman [1] proposed the necessity of angiogenesis for sustained tumor growth. This hypothesis sparked both academic as well as clinical interest in differentiating the newly developed, tumor-directed vascular bed from that of the tissue of origin, with the specific goal of better detection and treatment of neoplasia. Although it has been well established that tumor vasculature differs from the preexisting vasculature in morphological characteristics including vascular density and mean vessel radius [2–4], functional properties of these new vessels have not been as extensively investigated (for review, see Ref. [5]).

The ability of cerebral blood vessels to change diameter in response to alterations in PaCO2 is a well-characterized property of brain vasculature [6,7] by which small adjustments in ventilation result in significant changes in both cerebral blood flow (CBF) and cerebral blood volume (CBV) [8,9]. Recently, several studies have exploited this property of cerebral hemodynamics in order to assess the functional activity of blood vessels within tumors noninvasively with magnetic resonance imaging (MRI) [10–13]. In these reports, blood oxygenation level-dependent (BOLD) signal was used to detect the functional response of tumor vasculature to hypercarbia. BOLD signal is proportional to regional deoxyhemoglobin content [14,15]. In normal neural tissues with functionally active vasculature, BOLD signal intensity increases with hypercarbia because the increased influx of oxygenated blood with no significant change in oxygen utilization results in a decreased deoxyhemoglobin content. Indeed, BOLD contrast has been demonstrated to increase linearly in humans with arterial carbon dioxide tension [16].

Reports using BOLD signal contrast to assess tumor vascular reactivity to hypercarbia have produced conflicting results. Abramovitch et al. [10] detected no significant change in BOLD signal intensity in C6 gliomas with hypercarbia, but a robust increase in signal intensity with hypercarbic hyperoxia. Because hypercarbia appeared to have little effect on tumor blood flow, they concluded that the tumor vascular bed contained “immature” vessels, which lack the smooth muscle elements necessary for vasodilation in response to an elevation in carbon dioxide. Using the same technique, Gross et al. [12] similarly reported little signal change with hypercarbia in paraganglioma xenographs. Based on the lack of vascular reactivity, Neeman et al. [11] subsequently have proposed that the differential increase in tumor BOLD signal between induced hypercarbia and hypercarbic hyperoxia could be used as a noninvasive index of vascular maturity and angiogenesis. In contrast, however, Robinson et al. [13] reported a 10% to 30% increase in BOLD image intensity in GH3 prolactinomas with hypercarbia. In all of these studies, however, tumors had been implanted subcutaneously into the flank and the response of normal neural tissue was not assessed in comparison.

Although BOLD contrast is proportional to changes in local deoxyhemoglobin concentration, it is not possible to quantify alterations in either CBF or CBV based on BOLD signal intensity alone. This is largely due to the complex relationship that exists within each tissue type between these parameters and tissue oxygen extraction. Furthermore, BOLD signal changes are proportional to baseline tissue blood volume [17] that may be altered by angiogenesis. As such, BOLD signal changes are, at best, imprecise measures of local hemodynamic effects and, consequently, direct comparisons of physiological effects between tissues or across hemodynamic perturbations are not possible.

With the availability of superparamagnetic contrast agents that remain confined within the vascular space, it is now possible to accurately quantify relative changes in blood volume noninvasively by MRI with high spatial resolution. Furthermore, it is possible to preferentially sensitize MRI measurements of CBV toward the microvascular, or, alternatively, weight signal by total CBV by using steady state contrast agent-enhanced imaging [18]. Magnetic susceptibility contrast mechanisms predict that the change in relaxation rate observed with gradient echo (GE) images is proportional to the volume of blood contained in all vessels, whereas spin echo (SE) imaging is primarily sensitive to vessels less than 30 µm. This technique of interrogation of microvasculature was suggested by analytical simulations and verified experimentally [2]. Therefore, adjustments in both total and microvascular blood volumes can be assessed dynamically by rapidly imaging tissues during physiological perturbations.

Due to the inherent limitations of BOLD contrast as a reporter of vascular activity and the variability of the results from previous studies, the functional activity of tumor vasculature remains an open question. To address this issue, we quantified the difference in functional response to physiological challenge between tumor vasculature and normal tissue. We describe a direct quantitative measurement of the relative change in blood volume obtained simultaneously in both vascular beds with steady state, contrast agent-enhanced MRI. We assessed the functional response of the vasculatures to two challenges: 1) dilation in response to hypercarbia, and 2) constriction in response to hypocarbia. In order to test the proposal that newly derived or “immature” vessels lack functional activity, we compared the reactivity of blood volume between the entire vascular bed and the microvascular compartment selectively. Because the immature nonfunctional vessels are smaller in size and predominantly located in the microvascular compartment of the tumor bed, a difference in functional reactivity between compartments would indicate vessel size dependence in response.

We selected an experimental tumor model that permitted a direct comparison of functional activity of tumor vasculature with that of normal brain, specifically implantation of a human glioma-derived tumor cell line into normal brain. To minimize the potential for graft-versus-host reaction that could interfere with the functional behavior of tumor vasculature, tumor cells were implanted in athymic host animals that cannot mount such a response. The U87 cell line was selected because it was derived from a human glioma and is known to develop a highly vascular tumor similar to naturally occurring neuroaxial neoplasms. Our results indicated that although the reactivity of blood volume to carbon dioxide is maintained in glioma vasculature, it is significantly disrupted from normal.

Materials and Methods

Glioma Implantation

The U87MGdEGFR human glioblastoma cells were a generous gift of Dr. H. -J. Su Huang (University of California at San Diego). This cell line was established by retroviral transfer of a mutant epidermal growth factor receptor (de 2–7 EGFR) into the U87MG human glioblastoma cell line, enhancing its capacity to develop as a tumor in brains of nude rats [19]. Cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 100 U/ml penicillin, 100 Ag/ml streptomycin, and 50 µg/ml G418 at 37°C in 5% CO2.

Tumor Implantation

Animal studies were performed in accordance with guidelines issued by the Massachusetts General Hospital Subcommittee on Animal Care. Adult female nude rats (rnu/rnu; NCI, Frederick, MD) were anesthesized with 12.5 mg of ketamine and 2.5 mg of xylazine intraperitoneally. After immobilization in a stereotactic apparatus, a linear skin incision was made over the bregma and a 1-mm burr hole was drilled in the skull approximately 1 mm anterior and 2 mm lateral to bregma on the right side. A total of 200,000 tumor cells in 2 µl were injected at a depth of 3.5 mm from the dura using a 5-µl Hamilton syringe [19]. After awakening, rats were returned to their cages until imaging.

Animal Imaging

Ten to 12 days after tumor implantation, rats were anesthesized with 1.5% halothane in oxygen for surgery; a tracheostomy was placed and femoral arterial and venous cannulae were inserted. Following surgery, rats were mechanically ventilated with 0.7% halothane in a mixture of air and oxygen using a 3-ml tidal volume and a rate of 40 breaths/min with inspiratory time:expiratory time ratio of 1:1. After initiation of mechanical ventilation, rats were paralyzed with a 2-mg/kg bolus of pancuronium followed by continuous intravenous infusion at 2 mg/kg per hour. To minimize motion artifact, rats were placed into a custom-made plastic cradle attached to a head frame machined from plastic (David Kopf Instruments, Fremont, CA); heads were fixed with plastic screws inserted into the ear canals and a bar inserted under the front incisors. Rats were positioned prone in a 2-T, 11-cm horizontal bore SISCO NMR scanner (Varian Associates, Fremont, CA) with the head elevated approximately 1 cm with respect to the torso. A custom-built surface transmit and receive coil was applied firmly over the head. In order to reduce magnetic susceptibility artifacts arising from the air-tissue interface, rat heads were covered with gel toothpaste before the positioning of the coil; the ear canals and oropharynx were also filled with toothpaste. To maintain body temperature, the rat torso was wrapped in two heating blankets (Gaymar, Orchid Park, NY) that circulated water at 40°C.

Imaging Protocols

Relative CBV was assessed by injection of a superparamagnetic intravascular susceptibility contrast agent, monocrystalline iron oxide nanoparticle (MION), as described previously [2,20]. Briefly, after obtaining stable physiological condition and ventilation adjusted to attain normocarbia, a set of 60 precontrast agent SE and GE single-shot echo planar images (EPIs) was obtained through an interleaved double echo protocol (TR/TEGE/TESE = 2500/22/60 milliseconds, FOV 2.5 x 2.5 cm, in-plane resolution: 780 µm) with 11 slices of 1 mm thickness that covered the entire tumor. A total dose of 3 mg/kg MION was injected intravenously and the sequence was repeated to obtain postcontrast agent images. In images obtained with a GE pulse sequence, the change in relaxation rate resulting from injection of the intravascular contrast agent is proportional to total blood volume. Images obtained with an SE pulse sequence are preferentially sensitive to vessels less than 30 µm and, therefore, the change in relaxation reflects the microvascular blood volume compartment [18,21]. MION has a volumeweighted hydrodynamic diameter of 17 nm, and remains confined to the intravascular space in normal brain [22]. Previous studies have shown that there is no extravasation of MION into the interstitial space of tumors during the duration of these imaging experiments [2,23].

In order to determine the relative change in total and microvascular blood volume during transitions to hypercarbia and hypocarbia, an imaging sequence utilizing the same parameters was used to record GE and SE images, respectively, every 20 seconds. Following a baseline period of approximately 6 minutes, in one set of rats (n = 8, Hypercarbia Group), hypercarbia was induced by addition of 5% carbon dioxide to the inspired gas mixture and imaging continued for approximately 20 minutes. In another set of rats (n = 6, Hypocarbia Group), hypocarbia was induced by increasing ventilation rate. The ratio of inspiratory to expiratory time was maintained at 1:1. Arterial blood samples were obtained just prior to, and immediately after, imaging to quantify pH, PaO2, and PaCO2; measurements were performed with a Corning Blood Gas Analysis Machine (model 1304) on samples withdrawn from the femoral artery. Rat core temperature and mean arterial blood pressure (MABP) were monitored throughout the imaging session.

Image Data Analysis

Regions of tumor and contralateral striatum were selected for analysis by examination of postcontrast T2-weighted MR images. Mean tumor diameter was determined by counting the number of voxels in each slice representing the tumor and converting voxels to millimeters. Regional CBV was calculated relative to the mean blood volume determined for the contralateral normal striatum. The change in T2 relaxation rate produced by the injection of MION was used to determine relative total and microvascular CBV from GE (ΔR2*) and SE (ΔR2) images, respectively, by use of the formula: ΔR2 = -ln(Spre/Spost)/Te as previously described [2,20], where Spre and Spost represent signal intensities preinjection and postinjection of MION and Te is the echo time. Percent changes in CBV in tumor and contralateral striatum during hypercarbia and hypocarbia were calculated by comparisons of the average signal intensity at normocarbic baseline with the average signal intensity obtained after the perturbation in the respiratory status by the formula: CBV(t)/CBV(0) = ΔR2(t)/ΔR2(0) as previously described [20], where CBV(t) and ΔR2(t) are CBV and ΔR2 at a specific time, respectively, and CBV(0) and ΔR2(0) are the initial values for CBV and ΔR2, respectively. Transition to a new stable signal intensity was determined from review of the time course of signal intensity detected for the region examined. Reactivities of total and microvascular CBV were calculated for each rat by dividing the percentage change in CBV by the change in PaCO2 in millimeters Hg. Data are presented as mean±SEM; two-tailed t-tests were used to quantify the significance of difference.

Histology and Analysis

Brains from six rats bearing U87 gliomas were fixed in situ after imaging by intracardiac perfusion with 4% neutral paraformaldehyde in phosphate-buffered saline (pH = 7), removed, and then stored for 24 hours in perfusion fixative followed by storage for 48 hours in 30% sucrose before freezing. Frozen tissue sections, 10 µm thick, were cut on a cryostat (Cryocut 2800 E; Leica, Deerfield, IL), thawmounted onto glass microscope slides, and air-dried at room temperature. Alternating tumor sections were stained with NBT/BCIP (nitro blue tetrazolium/5-bromo-4-cloro-3-indolyl-phosphate, toluidine salt) in order to identify vessels as described; NBT/BCIP is enzymatically converted to a blue precipitate by alkaline phosphatase present in blood vessel endothelium [24]. Intervening sections were stained with hematoxylin and eosin to aid in identification of anatomy.

Based on anatomical landmarks and tumor size, tissue sections through tumors that corresponded to MR images were selected for microscopic analysis. Four separate regions for analysis were identified on each slide: contralateral striatum and three subsections of tumor as edge, middle, and center. Tumors were subdivided arbitrarily into geometric areas such that the diameter of the center region was approximately the same as the widths of the two ring-like regions surrounding the core for each tumor. Four separate photo images were obtained for each region with a digitizing microscope (Olympus BX60, Melville, NY) and processed with the ImageProPlus software package (MediaCybernetics, Silver Spring, MD). The program identified vessels based on the vascular stain; vessels appeared darker than the counterstained tissue background. A semiautomated method was used to select vessels in the field based on setting a threshold of overall light intensity and breakdown of color into red-green-blue components. Vessels were counted in each field; vascular area was determined as the fraction of total area contained within stained vessels. Vessel radius was determined as the minimum distance in microns from the center of each object to its edge. The results from the four images were averaged for each of the individual regions for all six animal brains.


Tumor size and baseline physiological parameters were similar between the Hypercarbia Group and the Hypocarbia Group of rats. There was no significant difference in mean tumor diameter between the two groups: 5.1 ± 0.5 mm in the Hypercarbia Group and 5.0 ± 0.2 mm in the Hypocarbia Group. Furthermore, there were no significant differences in normocarbic baseline values of pH, PaCO2, PaO2, and MABP (Table 1). Hematocrit was slightly higher in rats subjected to hypercarbia: 44% vs 37% (P < .01, pooled t-test). After the addition of 5% carbon dioxide to the breathing mixture, pH decreased and PaCO2 increased significantly (P < .0001, paired t-test); with hyperventilation, pH increased and PaCO2 decreased significantly (P < .003, paired t-test). There was no significant change in MABP, PaO2, or hematocrit as a result of either transition.

Table 1
Physiologic Data for U87- and GLI36-Bearing Rats During Perturbations of PaCO2.

Both total and microvascular blood volumes contained within U87 tumor vasculature were significantly elevated relative to the contralateral striatum at the time of experimentation. Figure 1 shows maps of relative total and microvascular CBV generated from the change in R2* and R2, respectively, after injection of MION contrast agent. At normocarbic baseline, tumor total blood volume was elevated in the Hypercarbia Group and the Hypocarbia Group by factors of 3.5 ± 0.3 and 4.1 ± 0.2, respectively, relative to the contralateral normal striatum (P < .0001, pooled t-test, comparison between tumor and normal striatum). Similarly, microvascular CBV was increased by factors of 2.2 ± 0.2 and 2.9 ± 0.2, respectively (P < .0001, pooled t-test, comparison between tumor and normal striatum). There was no significant difference in elevation of either total or microvascular tumor blood volume between the two treatment groups. In both treatment groups, the elevation of total blood volume within tumors was significantly greater than the elevation of microvascular blood volume (P < .008, pooled t-test).

Figure 1
Representative relative CBV maps generated from ΔR2 and ΔR2* for SE (a) and GE (b) images, respectively, produced by intravenous administration of MION (3.3 mg/kg iron) contrast agent. Each panel consists of five consecutive 1-mm-thick ...

Microscopic analysis of tissue sections stained for vascular endothelium revealed that gliomas were significantly more vascular than contralateral striatum (Figure 2), in good agreement with the increased blood volumes determined by MRI. Vessel density was increased significantly in gliomas relative to striatum; the number of vessels identified per high power field increased from 86 ± 15 in the striatum to 191 ± 32 in the glioma (P < .02, pooled t-test tumor versus striatum, n = 6). Similarly, vascular area was increased significantly in gliomas; the percentage of area contained within blood vessels per field increased from 2.2 ± 0.2% in the striatum to 5.3 ± 1.2% in tumors (P < .003, pooled t-test). The diameters of 3138 vessels in glioma and 2113 vessels in normal striatum were determined from histological sections; a histogram of the distribution of blood vessel sizes as a percentage of the total population of vessels examined is shown in Figure 3. Average vessel size was significantly elevated in gliomas; mean blood vessel diameter increased from 10.3 ± 0.4 µm in the striatum to 16.8 ± 1.3 ±m in the glioma (P < .004, pooled t-test). The vessel size distribution demonstrated a higher proportion of large vessels contained within gliomas; whereas only 20% of striatum vessels were larger than 14 µm, 52% of tumor vessels fell into this range. No necrotic areas were identified within the gliomas examined, and there were no significant differences detected in either vascular area or mean vessel diameter between tumor center, middle, or periphery.

Figure 2
Photomicrographs of normal striatum (a) and U87 glioma (b) stained with NBT/BCIP to identify vascular endothelium; tissue counterstained with eosin.
Figure 3
Histograms of blood vessel diameters for normal striatum (left) and glioma (right). Data are presented as percentage of total population of vessels examined in 1 µm size bins for 2113 vessels measured in normal striatum and 3138 vessels measured ...

A representative time course for the increase in total CBV in tumor (closed circles) and contralateral normal striatum (open circles) during transition from normocarbia to hypercarbia is shown in Figure 4. There was no significant drift with time in either GE or SE magnetic resonance signal intensity in images obtained either at baseline or after transition to hypercarbia or hypocarbia. Because progressive loss of MR signal would indicate leakage of the intravascular contrast agent into the tumor interstitial space, these results confirm the finding of Dennie et al. [2] that MION does not pass into the interstitial space within tumors within the time frame of these experiments.

Figure 4
Time course generated for the change in total blood volume in a region of U87 tumor (closed circles) and contralateral normal striatum (open circles) produced by transition to hypercarbia. A complete set of single-shot GE planar images covering eleven ...

Both hypercarbia and hypocarbia produced significant changes in tumor total and microvascular CBV. Although hypercarbia resulted in a significant increase in blood volume in both tumor and normal striatum (P < .003, pretransition versus posttransition comparison, paired t-test), the reactivity of tumor CBV to carbon dioxide was significantly blunted as compared to the contralateral normal striatum (Table 2). Total CBV increased by only 38% as much as the normal striatum whereas microvascular CBV increased by 58% as much as the normal striatum (P < .001, group t-test). In contrast, whereas hyperventilation resulted in significant hypocarbia (Table 1), blood volume decreased significantly only in tumor vasculature with no significant change in either total or microvascular CBV in normal striatum. In tumors, as in normal striatum, there was no significant difference in reactivity to carbon dioxide between total and microvascular blood volume compartment either for transition to hypercarbia or hypocarbia.

Table 2
Reactivity of Total and Microvascular CBV to PaCO2.


We interrogated the physiology of cerebral vasculature in order to identify differences in functional activity between tumor-derived vasculature and the native vasculature from which it sprouted. Specifically, we quantified the reactivity of both total and microvascular CBV to alterations in PaCO2. Although vasodilation was blunted significantly in glioma vascular bed as compared to normal striatum, tumor vessels did constrict effectively in response to hypocarbia and to a greater extent than the response of normal subcortex.

These results demonstrate several important features regarding angiogenesis in experimentally derived tumors. Firstly, although tumor vascularity may differ significantly from the tissue of origin, blood vessels developed as a result of tumor angiogenesis can exhibit functional activity typical of the tissue from which the vessels sprouted. At normocarbia, blood volume in the tumors examined here was significantly elevated relative to surrounding brain and contralateral normal striatum. In addition, the elevation in the total blood volume relative to normal brain was significantly larger than the elevation in the microvascular compartment alone; total blood volume was elevated by 51% more than microvascular blood volume in gliomas. This indicates a disproportionate increase in blood contained within larger vessels in the tumor vascular bed. This derangement of the tumor vascular bed observed with imaging agreed well with the histological analysis that confirmed the significantly more vascular nature of these gliomas relative to normal striatum. Fractional vascular area and vessel number in gliomas were increased 2.4- and 2.2-fold, respectively, relative to normal striatum. These values are comparable to the 2.5-fold increase in microvascular blood volume determined by MRI. Moreover, the distribution of tumor vessel diameters revealed a pronounced shift in the population to larger vessels; mean vessel diameter was increased by 63% in tumors relative to striatum, confirming the disproportionate increase in blood volume contained within larger vessels in tumors. These results also agree well with previous reports of elevated blood volumes in experimentally implanted gliomas and the dilated and tortuous blood vessels [4] that characterize the three-dimensional vascular network [2,25]. Nonetheless, despite these distortions of normal vessel anatomy and vascular architecture, we demonstrated an appropriate functional response of the U87 glioma vascular bed to both dilate and constrict with an acute change in PaCO2.

Secondly, although the response of the tumor vascular bed to hypercarbia was significantly blunted as compared to normal tissue, this did not translate into a reduction in functional ability to constrict in response to hypocarbia. Previous reports have described little or no increase in tumor BOLD signal intensity in experimentally implanted tumors challenged with hypercarbia [10–12]. Neeman et al. [11] proposed that the differential response in BOLD contrast between tumors subjected to hypercarbia alone and hypercarbic hyperoxia could be used as a noninvasive technique to identify immature vessels that lack the functional activity to dilate. Although we, as did they, observed impaired vasodilation, our demonstration of highly effective vasoconstriction within the tumor bed contradicts the notion that the blunted response to hypercarbia is due to vessels within the tumor that lack functional activity and, therefore, can be classified as “immature.” Although these results do not rule out the presence of a subpopulation of nonfunctional vessels present in the tumor, they do point out that the gross response of the vasculature to a single physiological challenge should be used with caution to comment on the microanatomy of the vascular bed.

Further evidence arguing against a preponderance of small nonfunctional blood vessels within the vascular bed is the good agreement in reactivity between the total blood volume and microvascular blood volume compartments. Because relative CBV determined by GE imaging is sensitive to all blood vessels, whereas SE imaging is primarily sensitive to vessels less than approximately 30 µm, it is possible to distinguish blood volume compartments with MRI [18]. If, as has been proposed, the tumor vascular bed contained a large population of “immature” blood vessels that lack structural elements necessary for vasoreactivity, then the microvascular compartment that contains predominantly these smaller “immature” vessels would be expected to exhibit less reactivity than the total blood volume compartment compartment that contains larger, more reactive vessels. Similar to the behavior of normal brain, however, there was no significant difference in reactivity between blood volume compartments. Tumors derived from the C6 glioma cell line also have a comparably disproportionate increase in blood volume contained within large vessels [2] as demonstrated for the U87 cell line reported here. Therefore, vessel size distribution cannot account entirely for the difference in results reported here and those of Abramovitch et al. [10] and Neeman et al. [11]. It is also important to note, however, that previous studies have indicated only a minor contribution of capillaries to adjustments in blood volume produced by moderate hypercarbia [27–29].

As has been reported previously [2,20,30] and confirmed here, we found no significant difference in reactivity to carbon dioxide between total and microvascular blood volume compartments in normal brain. Therefore, the mechanical activity responsible for alterations in vascular tone within tissues affected both blood compartments equally. Because the reactivity to carbon dioxide was also similar for both blood volume compartments in gliomas, we conclude that this physiological function was also preserved during angiogenesis. Despite the disproportionate increase in blood volume contained within large vessels and distortions in normal vessel anatomy, the mechanisms necessary for normal maintenance of vascular resistance appear intact in these tumors. Therefore, the tumor vascular bed appears to function more like normal brain than would be expected based on the gross differences in vasculature.

There are two previous reports of tumor vasculature that retain reactivity to changes in PaCO2. Robinson et al. [13] attributed a 10–30% increase in T2*-weighted MRI signal intensity during hypercarbia to be consistent with an increase in tumor blood flow resulting from vasodilation in GH3 prolactinomas. Cenic et al. [31] reported significant decreases in both blood flow and blood volume in VX2 tumors implanted intracranially in isoflurane-anesthesized rabbits during hypocarbia. Interestingly, neither normal brain nor tumor demonstrated any significant decrease in CBF or CBV with hyperventilation when rabbits were anesthesized with propofol. Although the specific characteristics of the C6 glioma cell line and the angiogenesis-derived tumor vasculature may have been responsible for the lack of response to hypercarbia in tumor xenografts implanted subcutaneously in nude mouse flank reported by Abramovitch et al. [10] and Neeman et al. [11], the limitations of BOLD imaging to detect small changes in physiological parameters may also have played a role. Alternatively, anesthetic conditions, baseline blood volume, or location of implantation may have contributed to the lack of response.

Several possible explanations could account for reduced vasodilation but normal vasoconstriction in glioma vasculature observed here. In their native state, tumor capacitance vessels could be already near maximally dilated and have little capacity to increase the diameter further in response to an increase in CBF. Alternatively, local glioma CBF may be near maximal with little capacity to increase. CBF has been shown to be elevated in this tumor model [32]. Because the majority of tissue blood volume is located in the capillary and venous compartments and is under relatively low pressure, physical constraints of the immediate tumor environment could influence venous compliance as well. The presence of a space-occupying lesion within the closed confines of the cranium can increase intracranial pressure—a condition that impacts on cerebral hemodynamics. Elevation of tumor intrinsic tissue pressure could exert a similar effect [33,34]. Finally, vascular reactivity could be a characteristic of individual tumor cell lines. Using the same experimental design described here, we quantified the vascular reactivity to carbon dioxide for tumors derived from the human glioma cell line GLI36. In that case, the increase in total and microvascular blood volume induced by hypercarbia was not different from the contralateral normal striatum (unpublished observation). Simultaneous determination of changes in CBF and CBV during physiological challenge could be helpful to distinguish between these possibilities.

The reactivity of CBV to hypercarbia for normal striatum reported here (1.5 ± 0.2% CBV/mm Hg PaCO2) compares favorably to results obtained previously by others. Payen et al. [35] reported a 1.7 ± 0.3% increase in CBV per millimeters mercury PaCO2 in the striatum for transition to hypercarbia over a similar range of PaCO2 using GE planar MRI in halothane/nitrous oxide-anesthesized normal rats. Similarly, Mandeville et al. [20] reported a 1.3% increase and Zaharchuk et al. [30] a 1.8% increase in CBV per millimeters Hg PaCO2 in the whole brain of halothane-anesthesized rats by GE MRI. For transition to hypocarbia, results are more variable in quantifying the reduction in total CBV. Using dynamic contrast-enhanced CT to measure CBV, Cenic et al. [31,36] reported no significant difference in CBV during transition to hypocarbia in propofol-anesthesized rabbits, but a significant decrease of 1.1 ± 0.5% CBV/mm Hg PaCO2 under isoflurane anesthesia. This difference in reactivity was attributed to the vasoactive properties of the anesthetic to effect baseline values of CBV. Similarly, Weeks et al. [37] found that reactivity to carbon dioxide varied widely depending on anesthetic conditions; although halothane and pentobarbital resulted in no significant change in CBV with hypocarbia, CBV decreased by 1.4% mm Hg PaCO2 under isoflurane anesthesia. In good agreement with these results, we also found no significant decrease in CBV in normal brain during hypocarbia under halothane anesthesia.

Among the functional properties that have been described for normal cerebral circulation, MRI has been reported to assess reactivity to carbon dioxide, autoregulation of blood flow, blood flow metabolism coupling, and integrity of the blood-brain barrier. Although penetration of the blood-brain barrier by small-molecular-weight contrast agents is used routinely to image neoplasms, the noninvasive determination of disruption of other physiological properties of tumor vasculature has not become common in clinical practice. Recent evidence has suggested, though, that neuroaxis neoplasms may be graded for malignancy noninvasively based on metabolic rate [38,39], blood volume, and blood flow [40–42], and the kinetics of permeability to contrast agents [43,44]. Further studies are necessary to determine the value of vascular reactivity as a useful clinical marker either of malignancy or of degree of neovascularity. The availability of superparamagnetic intravascular susceptibility contrast agents that do not penetrate a disrupted blood-brain barrier will greatly facilitate the assessment of tumor vascular function in the clinical setting. Although it is unlikely that imaging will replace surgical biopsies in the near future, imaging of the functional properties of tumors may aid in the selection of biopsy sites to improve the accuracy of diagnosis and selection of therapy. Finally, MRI assessment of vascular function may help in elucidating the mechanism of angiogenesis and prove useful in the development and evaluation of antiangiogenic pharmaceuticals.


Although vasodilation in response to hypercarbia was significantly blunted within U87 gliomas relative to normal brain, vasoconstriction during hypocarbia was enhanced. This pattern of reactivity was similar in response to the total as well as the microvascular blood volume compartments. Simple assessment of vascular response to hypercarbia with indirect measures of hemodynamics may be inadequate to determine functional activity and classify tumor-derived blood vessels as immature.


1. Folkman J. Tumor angiogenesis: therapeutic implications. N Engl J Med. 1971;285:1182–1186. [PubMed]
2. Dennie J, Mandeville JB, Boxerman JL, Packard SD, Rosen BR, Weisskoff RM. NMR imaging of changes in vascular morphology due to tumor angiogenesis. Magn Reson Med. 1998;40:793–799. [PubMed]
3. Deane BR, Lantos PL. The vasculature of experimental brain tumours: Part 2. A quantitative assessment of morphological abnormalities. J Neurol Sci. 1981;49:67–77.
4. Scatliff JH, Radcliffe WB, Pittman HH, Park CH. Vascular structure of glioblastomas. Am J Roentgenol Radium Ther Nucl Med. 1969;105:795–805.
5. Vajkoczy P, Menger MD. Vascular microenvironment in gliomas. J Neuro-Oncol. 2000;50:99–108.
6. Auerr L. Pial arterial reactions to hyper- and hypocapnia: a dynamic experimental study in cats. Eur Neurol. 1978;17:351–362. [PubMed]
7. Atkinson JL, Anderson RE, Sundt TM., Jr The effect of carbon dioxide on the diameter of brain capillaries. Brain Res. 1990;517:333–340. [PubMed]
8. Reivich M. Regulation of the cerebral circulation. Clin Neurosurg. 1969;16:378–418. [PubMed]
9. Reich T, Rusinek H. Cerebral cortical and white matter reactivity to carbon dioxide. Stroke. 1989;20:453–457. [PubMed]
10. Abramovitch R, Dafni H, Smouha E, Benjamin LE, Neeman M. In vivo prediction of vascular susceptibility to vascular susceptibility endothelial growth factor withdrawal: magnetic resonance imaging of C6 rat glioma in nude mice. Cancer Res. 1999;59:5012–5016. [PubMed]
11. Neeman M, Dafni H, Bukhari O, Braun RD, Dewhirst MW. In vivo bold contrast MRI mapping of subcutaneous vascular function and maturation: validation by intravital microscopy. Magn Reson Med. 2001;45:887–898. [PubMed]
12. Gross DJ, Reibstein I, Weiss L, Slavin S, Stein I, Neeman M, Abramovitch R, Benjamin LE. The antiangiogenic agent linomide inhibits the growth rate of Von Hippel - Lindau paraganglioma xenografts to mice. Clin Cancer Res. 1999;5:3669–3675. [PubMed]
13. Robinson SP, Collingridge DR, Howe FA, Rodrigues LM, Chaplin JR, Griffiths JR. Tumour response to hypercapnia and hyperoxia monitored by flood magnetic resonance imaging. NMR Biomed. 1999;12:98–106. [PubMed]
14. Ogawa S, Lee TM. Magnetic resonance imaging of blood vessels at high fields: in vivo and in vitro measurements and image simulation. Magn Reson Med. 1990;16:9–18. [PubMed]
15. Ogawa S, Lee TM, Barrere B. The sensitivity of magnetic resonance image signals of a rat brain to changes in the cerebral venous blood oxygenation. Magn Reson Med. 1993;29:205–210. [PubMed]
16. Rostrup E, Larsson HB, Toft PB, Garde K, Thomsen C, Ring P, Sondergaard L, Henriksen O. Functional MRI of CO2 induced increase in cerebral perfusion. NMR Biomed. 1994;7:29–34. [PubMed]
17. Mandeville JB, Marota JJ. Vascular filters of functional MRI: spatial localization using bold and CBV contrast. Magn Reson Med. 1999;42:591–598. [PubMed]
18. Boxerman JL, Hamberg LM, Rosen BR, Weisskoff RM. MR contrast due to intravascular magnetic susceptibility perturbations. Magn Reson Med. 1995;34:555–566. [PubMed]
19. Nishikawa R, Ji XD, Harmon RC, Lazar CS, Gill GN, Cavenee WK, Huang HJ. A mutant epidermal growth factor receptor common in human glioma confers enhanced tumorigenicity. Proc Natl Acad Sci USA. 1994;91:7727–7731. [PubMed]
20. Mandeville JB, Marota JJ, Kosofsky BE, Keltner JR, Weissleder R, Rosen BR, Weisskoff RM. Dynamic functional imaging of relative cerebral blood volume during rat forepaw stimulation. Magn Reson Med. 1998;39:615–624. [PubMed]
21. Boxerman JL, Bandettini PA, Kwong KK, Baker JR, Davis TL, Rosen BR, Weisskoff RM. The intravascular contribution to FMRI signal change: Monte Carlo modeling and diffusion-weighted studies in vivo. Magn Reson Med. 1995;34:4–10. [PubMed]
22. Shen T, Weissleder R, Papisov M, Bogdanov A, Jr, Brady TJ. Monocrystalline iron oxide nanocompounds (MION): physicochemical properties. Magn Reson Med. 1993;29:599–604. [PubMed]
23. Zimmer C, Weissleder R, Poss K, Bogdanova A, Wright SC, Jr, Enochs WS. MR imaging of phagocytosis in experimental gliomas. Radiology. 1995;197:533–538. [PubMed]
24. Weiss HR, Buchweitz E, Murtha TJ, Auletta M. Quantitative regional determination of morphometric indices of the total and perfused capillary network in the rat brain. Circ Res. 1982;51:494–503. [PubMed]
25. Zama A, Tamura M, Inoue HK. Three-dimensional observations on microvascular growth in rat glioma using a vascular casting method. J Cancer Res Clin Oncol. 1991;117:396–402. [PubMed]
26. Abramovitch R, Frenkiel D, Neeman M. Analysis of subcutaneous angiogenesis by gradient echo magnetic resonance imaging. Mag Reson Med. 1998;39:813–824.
27. Duelli R, Kuschinsky W. Changes in brain capillary diameter during hypocapnia and hypercapnia. J Cereb Blood Flow Metab. 1993;13:1025–1028. [PubMed]
28. Chen JL, Wei L, Acuff V, Bereczki D, Hans FJ, Otsuka T, Finnegan W, Patlak C, Fenstermacher J. Slightly altered permeability - surface area products imply some cerebral capillary recruitment during hypercapnia. Microvasc Res. 1994;48:190–211. [PubMed]
29. Keyeux A, Ochrymowicz-Bemelmans D, Charlier AA. Induced response to hypercapnia in the two-compartment total cerebral blood volume: influence on brain vascular reserve and flow efficiency. J Cereb Blood Flow Metab. 1995;15:1121–1131. [PubMed]
30. Zaharchuk G, Bogdanov AA, Jr, Marota JJ, Shimizu-Sasamata M, Weisskoff RM, Kwong KK, Jenkins BG, Weissleder R, Rosen BR. Continuous Assessment of Perfusion by Tagging Including Volume and Water Extraction (Captive): a steady-state contrast agent technique for measuring blood flow, relative blood volume fraction, and the water extraction fraction. Magn Reson Med. 1998;40:666–678. [PubMed]
31. Cenic A, Craen RA, Lee TY, Gelb AW. Cerebral blood volume and blood flow responses to hyperventilation in brain tumors during isoflurane or propofol anesthesia. Anesth Analg. 2002;94:661–666. (table of contents) [PubMed]
32. Packard SD, Ichikawa T, Mandeville JB, Chiocca EA, Rosen BR, Marota JJA. CBV and CBF are mismatched in experimental model of human glioma: measuring tumor perfusion efficiency. Int Soc Magn Reson Med. 2000;8:1064.
33. Boucher Y, Leunig M, Jain RK. Tumor angiogenesis and interstitial hypertension. Cancer Res. 1996;56:4261–4266.
34. Luthert PJ, Greenwood J. Selective closure of the vascular bed of an experimental glioma model during in situ saline perfusion. Neuropathol Appl Neurobiol. 1994;20:448–453. [PubMed]
35. Payen JF, Vath A, Koenigsberg B, Bourlier V, Decorps M. Regional cerebral plasma volume response to carbon dioxide using magnetic resonance imaging. Anesthesiology. 1998;88:984–992. [PubMed]
36. Cenic A, Craen RA, Howard-Lech VL, Lee TY, Gelb AW. Cerebral blood volume and blood flow at varying arterial carbon dioxide tension levels in rabbits during propofol anesthesia. Anesth Analg. 2000;90:1376–1383. [PubMed]
37. Weeks JB, Todd MM, Warner DS, Katz J. The influence of halothane, isoflurane, and pentobarbital on cerebral plasma volume in hypocapnic and normocapnic rats. Anesthesiology. 1990;73:461–466. [PubMed]
38. Aronen HJ, Gazit IE, Louis DN, Buchbinder BR, Pardo FS, Weisskoff RM, Harsh GR, Cosgrove GR, Halpern EF, Hochberg FH, et al. Cerebral blood volume maps of gliomas: comparison with tumor grade and histologic findings. Radiology. 1994;191:41–51. [PubMed]
39. Aronen HJ, Glass J, Pardo FS, Belliveau JW, Gruber ML, Buchbinder BR, Gazit IE, Linggood RM, Fischman AJ, Rosen BR, et al. Echoplanar MR cerebral blood volume mapping of gliomas. Clin Utility Acta Radiol. 1995;36:520–528.
40. Sugahara T, Korogi Y, Kochi M, Ikushima I, Hirai T, Okuda T, Shigematsu Y, Liang L, Ge Y, Ushio Y, Takahashi M. Correlation of MR imaging-determined cerebral blood volume maps with histologic and angiographic determination of vascularity of gliomas. AJR, Am J Roentgenol. 1998;171:1479–1486. [PubMed]
41. Sugahara T, Korogi Y, Kochi M, Ushio Y, Takahashi M. Perfusion-sensitive MR imaging of gliomas: comparison between gradient-echo and spin-echo echo-planar imaging techniques. AJNR, Am J Neuroradiol. 2001;22:1306–1315. [PubMed]
42. Shin JH, Lee HK, Kwun BD, Kim JS, Kang W, Choi CG, Suh DC. Using relative cerebral blood flow and volume to evaluate the histopathologic grade of cerebral gliomas: preliminary results. AJR, Am J Roentgenol. 2002;179:783–789. [PubMed]
43. Tofts PS, Brix G, Buckley DL, Evelhoch JL, Henderson E, Knopp MV, Larsson HB, Lee TY, Mayr NA, Parker GJ, Port RE, Taylor J, Weisskoff RM. Estimating kinetic parameters from dynamic contrast-enhanced T(1)-weighted MRI of a diffusable tracer: standardized quantities and symbols. J Magn Reson Imaging. 1999;10:223–232. [PubMed]
44. Provenzale JM, Wang GR, Brenner T, Petrella JR, Sorensen AG. Comparison of permeability in high-grade and low-grade brain tumors using dynamic susceptibility contrast MR imaging. AJR, Am J Roentgenol. 2002;178:711–716. [PubMed]

Articles from Neoplasia (New York, N.Y.) are provided here courtesy of Neoplasia Press