Publication of the lnternatlonal Union Against Cancer Publication de I'Union Internationale Contre le Cancer

Znt. J. Cancer: 65,658-663 (1996) 0 1996 Wiley-Liss, Inc.

SPATIAL HETEROGENEITY OF TUMOUR BLOOD FLOW MODIFICATION INDUCED BY ANGIOTENSIN 11: RELATIONSHIP TO RECEPTOR DISTRIBUTION Gillian M. TOZER~ , Katija M. SHAFFI, Vivien E. PRISE and Katrina M. BELL Turnour Microcirculation Group, Gray Laboratory Cancer Research Trust, Mount Vernon Hospital, Northwood, Middlesex HA6 Z R , United Kingdom.

Angiotensin II (ATll) has potential for improving delivery of blood-borne anti-cancer agents to tumours by increasing tu- mour blood flow. However, ATII-induced hypertension is not always accompanied by an increase in tumour blood flow due to significant constriction of the tumour vasculature. Such unpre- dictability in tumour response to ATll limits the clinical useful- ness of this approach. In this study, the potential of assessing numbers of binding sites for ATll as a predictor of tumour blood flow response to intravenous administration of ATll was investi- gated. The distribution of AT11 receptors in the rat P22 carcinosarcoma was related to tumour blood flow distribution and blood flow response to ATll using an autoradiographic approach. ATll (0.2 pg kg-I * min-I) increased mean arterial blood pressure of anaesthetized BD9 rats from 92.2 f 1.4 mmHg to 145.6 f 1.3 mmHg. Despite this increase in perfusion pressure, overall tumour blood flow to viable regions decreased by 2Ooh, indicating significant constriction of tumour blood vessels. Autoradiographic localisation of tumour blood flow showed that the decrease in flow was confined to the turnour periphery, with no change at the tumour centre. This pattern was consistent with 10% more binding sites for ATll at the tumour periphery than at the tumour centre. Maximum num- ber of binding sites (BhU) for the P22 turnour was 0.38 f 0.09 fmol - mg-I, which is approximately a factor of 10 lower than published values for various normal tissues. The dissociation constant K,, was I. 16 f 0. I8 nM. These results encourage the development of techniques for analysis of receptor binding characteristics for predicting the response of individual turnours to blood flow manipulation using vasoactive agents. 0 1996 Wiley-Liss, Znc.

Intravenous infusion of angiotensin I1 (ATII) induces hyper- tension and has been used clinically for improving blood flow to tumours relative to that in critical normal tissues. It therefore has potential for improving the relative delivery of chemotherapeutic agents to tumours (Suzuki et al., 1981; Takematsu et aZ., 1985; Noguchi et al., 1988; Kobayashi et aL, 1990, 1991; Anderson et al., 1991; Kerr et al., 1992; Mutoh et al., 1992).

We have previously found that overall blood flow to the P22 carcinosarcoma growing subcutaneously in BD9 rats is de- creased by about 20% during intravenous infusion of ATII, despite a redistribution of the cardiac output in favour of the tumour (Tozer and Shaffi, 1993). This decrease in absolute blood flow has been observed in other experimental tumours (Jirtle et al., 1978) but not in all of them (Tokuda et aZ., 1990; Hori et aZ., 1991; Tanda et al., 1991). The explanation for this discrepancy could be associated with different dose regimes and assay systems used by different research groups, but there are also likely to be real differences between tumours that modify the outcome. A decrease in absolute blood flow could compromise drug delivery to tumour microregions. Therefore, it is important to investigate the spatial heterogeneity of tumour blood flow modification induced by AT11 and also the factors determining response.

Blood flow to the periphery of P22 tumours is normally about 50% higher than that to the centre (Tozer et al., 1994a). We have found that AT11 abolishes this difference but, with the number of animals used, we could not determine whether an increase in blood flow to the turnour centre was contribut-

ing to this effect (Tozer et af., 1994a). Subsequently, we have found no evidence for ATII-induced re-opening of previously collapsed tumour blood vessels that could explain the spatial heterogeneity of the AT11 effect in these tumours (Tozer and Shaffi, 1995).

The aims of the present study were (1) to characterise the spatial heterogeneity of blood flow response to AT11 in the P22 tumour and (2) to determine whether distribution of AT11 receptors could explain this response. These results should have significance for the potential use of receptor characterisa- tion for determining response to blood flow modification in individual tumours.

MATERIAL AND METHODS Turnours

Early generations (up to 12 away from the primary tumour) of the P22 transplanted rat carcinosarcoma were used for these experiments (Tozer and Shaffi, 1993). Tumours were grown subcutaneously in the left flank of 10-12 week old male BD9 rats. Tumours were used for experimentation when they reached 1-2 g (all 3 orthogonal diameters 10-15 mm, including skin thickness).

Blood flow Turnour blood flow was measured using uptake of 14C-

labelled iodo-antipyrine (l4C-IAP, Amersham, Aylesbury, UK). This method has been described previously (Tozer and Shaffi, 1993). Briefly, animals were anaesthetised with Hypnorm and Midazolam and 2 tail veins and 1 tail artery catheterised using polyethylene catheters (external diameter 0.96 mm; internal diameter 0.58 mm). Mean arterial blood pressure (MABP) was monitored via the tail artery catheter up to the point of blood flow measurement. Animals were heparinised with 0.2 ml of 1000 units . ml-l heparin (CP Pharmaceuticals, Wrexham, UK) via a catheterised tail vein. AT11 (0.2-0.3 pg . kg-l * min-I; Sigma, Poole, UK) or the vehicle for the drug (0.9% saline) was administered by constant infusion into the second tail vein at a rate of 0.2 ml * kg-l . min - l . Twenty minutes after the start of AT11 infusion, 0.44-0.56 MBq (12-15 pCi) of 14C-IAP in 0.8 ml saline was infused into the first tail vein over 30 sec. During the 30 sec period, free-flowing arterial blood from the tail artery was collected into pre-weighed vials at 1 sec intervals. At the end of the 30 sec, the rat was killed by intravenous injection of Euthatal. Tumours and overlying skin were rapidly excised. The skin was divided into small pieces ( < 0.2 g), weighed, solubilised and counted by liquid scintilla- tion with suitable quench correction. Blood samples were similarly weighed and counted. Immediately after excision, tumours were frozen in isopentane at approximately -40°C and stored at -70°C for subsequent autoradiography of 20 pm thick cryosections.

'To whom correspondence and reprint requests should be sent. Fax: (44) 1923-835210.

Received: August 2,1995 and in revised form October 20, 1995.

659 TUMOUR BLOOD FLOW AND AT11 RECEPTORS

Blood flow to tumour and skin was calculated from the tissue counts measured via liquid scintillation counting or autoradiog- raphy, the equilibrium partition coefficient of IAP in the different tissues (Tozer and Morris, 1990) and the arterial input function derived from the arterial blood counts. This method is based on principles derived by Kety (1960) and is described in detail elsewhere (Tozer et al., 19946).

A TII receptors Distribution of AT11 receptors in untreated P22 tumours

was assessed by measuring the binding characteristics to tumour sections of the AT11 analogue ATII(Sarl, Ile8)-Tyr4 labelled with 1251(1251-ATII SI) (DuPont NEN, Stevenage, UK), a selective antagonist of ATII. Autoradiography was used for localisation and quantitation. No attempt was made to discriminate between different ATII receptor sub-types.

Animals bearing P22 tumours (1-2 g) were sacrificed by cervical dislocation following concussion and the tumours excised and frozen as described for blood flow measurements. Within a week of storage at -7o"C, 20 pm thick cryosections were cut and mounted in serial pairs onto cold poly-L-lysine (Sigma) coated slides. Slides were stored at -20°C until completion of sectioning and then desiccated overnight at -70°C. In one group of 6 animals, receptor distribution was measured following the blood flow procedure described above. In this group, cryosections were cut in serial triplets. One section from each triplet was used immediately for autoradiog- raphy of blood flow. The remaining pair were processed for receptor studies.

Frozen tumour sections were brought to room temperature with a stream of cool air. Once dry, sections were pre- incubated for 30 min in ice-cold buffer containing 0.4% bovine serum albumin, 10 mM MgC12, 150 mM NaCI, 5 mM EGTA and 50 mM NaHP04 (pH 7.2) in order to remove any endogenous AT11 or related ligand. A low temperature and high sodium concentration for the buffer were specifically chosen because these factors have been shown to increase specific binding, most likely due to efficient dissociation of endogenous ligand (Simantov et al., 1976). Low temperature also helps to maintain tissue integrity. The pre-incubation also served to remove 14C-IAP from those tumours used for blood flow determination. Removal was confirmed by autoradiogra-

Preliminary experiments were carried out to determine suitable concentrations of lZ5I-ATII SI for incubation and suitable incubation and washing times to maximise specific binding. Most sections were incubated with 0.5 nM IZ5I-ATII SI, except for 1 experiment where doses of 0.1 nM to 5.6 nM were used for quantitating the maximum number of binding sites (BLmax) and the dissociation constant (Kd). Incubations were achieved by spotting directly onto each section. One section from each pair on a slide was incubated with l=I-ATII SI alone for determination of total binding and the other section was incubated with 1251-ATII SI in the presence of 3 pM unlabelled AT11 SI (Sigma) for determination of non- specific binding. Incubation was for 45 min in moist conditions at 21-22"C, terminated by a dip in ice-cold buffer. Further rinsing in buffer for 6 min and a final dip in distilled water were followed by fan-drying and overnight desiccation.

Quantitative autoradiography Slides were mounted onto thin perspex and overlaid with

autoradiographic film (Hyperfilm pmax for blood flow slides and Hyperfilm 3H for receptor slides, both from Amersham). 14C-labelled methylmethacrylate microscales (Amersham), of known isotope concentration were included with the blood flow slides for quantitation of autoradiograms. lZ5I microscales (Amersham) were included with receptor slides used for calculation of BLma and Kd. For the remaining receptor studies, optical density (O.D.) of autoradiograms was used to

PhY.

assess ligand binding. A linear relationship between O.D. and concentration of ligand (in fmol . g-l) pertains over the O.D. range used (data not shown). Following suitable exposure (several weeks), autoradiograms were developed and sections were fixed and stained with haematoxylin and eosin (H&E).

Autoradiograms were analysed using an Applied Imaging (Warrington, UK) image analysis system. Levels of radioactiv- ity in the microscales were converted into blood flow values (ml . g-' . min-') using arterial input functions specific to each rat or, where appropriate, into binding values (fmol . mg-I) using the specific activity of lZ5I-ATII SI supplied with the ligand. This information was used to generate standard curves from which O.D. images were transformed into blood flow or binding images. Mean values for blood flow, total binding or non-specific binding were obtained from regions of interest in each section. Specific binding (in fmol.mg-' or as O.D. values) for each pair of sections used for receptor studies was calculated by subtraction of the mean non-specific binding from the mean total binding using the whole sections as the regions of interest. Images of specific binding were obtained where it was possible to overlay total and non-specific binding images for subtraction. Histological images from H&E stained sections were also overlaid onto blood flow images in order to define viable or necrotic tumour areas from which blood flow values could be extracted. A peripheral region was defined, in sections cut from the tumour centre, as the rim of a section whose width was equal to 1/10 of the largest diameter of the section. A central region was defined as the remaining area of the section.

Analysis of blood fiow results Results from blood flow images were expressed as mean

blood flow in ml . g - l .min-l * standard deviation for individual regions in each section. This corresponds to 1000 to 12,000 pixel samples of blood flow per section. Results for groups of animals were expressed as means f 1 standard error of the mean. Vascular resistance in resistance units (1 resis- tance unit = 1 (mmHg) . (ml . g-' . min-l)-l was calculated from perfusion pressure + blood flow, where h4ABP was assumed to be equivalent to perfusion pressure. Coefficient of variation of blood flow for each section was calculated from standard deviation of pixel blood flow values +- mean blood flow.

Quantitation of receptor binding Specific binding (BL) in fmol . mg-l versus free l=I-ATII SI

concentration [L] in nM was fitted to the Langmuir binding isotherm (Matthews, 1993):

Non-linear regression analysis was used to obtain estimates of Bhax and Kd. Statistics

All data was tested for normality of distribution and equal variances between groups using a W-test and an F-test, respectively (JMP Statistics for the Apple Macintosh). In the majority of cases, these conditions were fulfilled and the Student's t-test for unpaired data was used for comparison of means unless otherwise stated.

RESULTS

A dose of 0.2 pg . kg-I . min-I AT11 increased MABP from 92.2 f 1.4 mmHg to 145.6 5 1.3 mmHg. Figure 1 shows that AT11 also produced a significant reduction in blood flow to both skin and viable regions within the P22 tumour. Skin blood flow was reduced by 65%, whereas tumour blood flow was only reduced by 20%. These results are consistent with our earlier

TOZER ETAL. 660

a) b)

0’5 f p = 0.05 0 4 0 4 -

z c 0 3 P E O 3

_o - 02

- x -m 02 = L

0 1 0 1

0 0 control ATll control AT1 I viable turnour regions skin overlying tumour

FIGURE 1 - ATII-induced changes in blood flow to (a) viable tumour regions and (b) skin overlying tumour. Results in (a) were derived from autoradiography of 5 to 7 sections per tumour spaced through the whole tumour mass. Results in (b) were derived from liquid scintillation of skin samples. Results are means t 1 SEM for n = 11 animals (control group) and n = 10 animals (ATII group); p values below 0.05 indicate a significant difference between means at the 5% level. *t-test for unequal variances.

findings (Tozer and Shaffi, 1993; Tozer et al., 1994~) . Necrotic fractions were low in these tumours but slightly higher in the control group than in the ATII-treated group (7.3 f 1.8% compared to 3.3 f 1.1%, respectively). This difference prob- ably accounts for the fact that the tendency for AT11 to reduce blood flow in whole tumours (viable plus necrotic regions) did not quite reach significance ( p = 0.07, results not shown).

Figure 2a and b shows that blood flow to the tumour periphery (which under normal conditions tends to be higher than that to the centre) is significantly reduced by AT11 ( p = 0.035). There was no significant change in flow to the tumour centre ( p = 0.328). This point was unresolved in our previous publications (Tozer and Shaffi, 1993; Tozer et al., 1994~) . Microregional heterogeneity in tumour blood flow tended to decrease with ATII, as measured by a reduction in the coefficient of variation of blood flow from 0.98 to 0.81. However, this change was not significant ( p > 0.05).

Tendency for blood flow to decrease is partly offset by the increase in perfusion pressure (MABP). This is demonstrated in Figure 2c and d, which shows that ATII-induced increases in vascular resistance were larger than the decreases in blood flow (e.g., there was a 130% increase in flow resistance in the tumour periphery compared to a 32% decrease in blood flow). Although blood flow to the tumour centre was unchanged by ATII, there was still a significant increase in vascular resis- tance ( p = 0.003, Fig. 26, d) , indicating that there was ATII- induced vasoconstriction throughout the tumour mass.

Specific binding of 1251-ATII SI to its binding site was usually around 30% of the total binding. Variation between tumours from sequential transplants was significant but small, and there was no indication of a progressive increase or decrease with transplant generation (data not shown). A larger variation in percentage specific binding was found between tumours aris- ing from different frozen stocks. Numbers of binding sites were also quite variable in this situation. Figure 3 shows binding data for 2 tumours from different frozen stocks that have very different numbers of binding sites (BLmax = 0.38 f 0.02 fmol . mg-’ for tumour 1 and 0.043 fmol . mg-l for tumour 2). The dissociation constant (&) is similar for the 2 tumours shown in Figure 3 despite the difference in BLmax (Kd = 0.90 -t 0.15 nM for tumour 1 and 0.57 nM for tumour 2). The mean Kd for all tumours analysed was 1.16 -I- 0.18 nM (n = 7). Mean BLma for all tumours analysed was 0.38 f 0.09 fmol . mg-’ (n = 7).

There were 10% more 1251-ATII SI specific binding sites in peripheral tumour regions than in central regions (Fig. 4). This

a)

p = 0.035 o’6 T

control AT11 peripheral tumour

C)

800

700

control ATll peripheral turnour

t

p = 0.328 0.4

control ATll central tumour

d) F ‘p = 0.003 I

700

600

500

400

300

200

100

n control AT11

central turnour

FIGURE 2 - ATII-induced changes in (a) absolute blood flow to peripheral tumour regions, (b) absolute blood flow to central tumour regions, (c) vascular resistance in peripheral tumour regions and (d) vascular resistance in central tumour regions. All results were derived from autoradiography of 2 sections per tumour cut from the tumour centre. Results are means t 1 SEM for n = 11 animals (control group) and n = 10 animals (ATII group); p values below 0.05 indicate a significant difference between means at the 5% level. *Wilcoxon test requiring no assumption of normality of distribution.

difference was significant using Student’s t-test for paired data ( p = 0.004) and is consistent with the greater vasoconstrictive effect of AT11 in the tumour periphery (Fig. 2 ) . Figure 5 shows a typical example of specific binding of 1251-ATII SI in a section cut from the centre of a P22 tumour (Fig. 5a) together with the blood flow pattern from an adjacent section of the same tumour (Fig. 5b). This pair of images graphically demonstrates that both binding and blood flow are higher at the tumour periphery than at its centre. However, the binding image has a much sharper, high intensity rim than the blood flow image, and within the general mass of the tumour, specific binding is relatively uniform, whereas blood flow is highly heteroge- neous. The only exception to uniformity of specific binding in the tumour centre is that there are areas of low binding in areas of very low blood flow. These regions usually contained small necrotic regions, although reduced specific binding extended well beyond the necrotic border. This pattern was typical of all 6 tumours analysed in this way.

DISCUSSION

In this study we have shown that vasoconstriction of the P22 tumour in response to AT11 occurs throughout the tumour mass. However, there is more vasoconstriction at the periphery than at the centre. There is no evidence to support the suggestion that AT11 “opens up” previously collapsed tumour blood vessels (Trotter et al., 1991; Hemingway et al., 1992). The net result of vasoconstriction combined with the increase in systemic blood pressure is a decrease in blood flow at the periphery and no change in the centre.

This pattern of blood flow changes is consistent with the binding pattern of lZ51-ATII SI to its receptors. That is, binding

TUMOUR BLOOD FLOW AND AT11 RECEPTORS

- c

2 0.4:

661

- - 0.04-

- 2 0.26- E ._

v

C

0.24- ._

0.22{

% 0.2- u ._

FIGURE 3 - Bound ligand versus free ligand for lZ5I-ATII SI in 2 tumours (tumour 1 and turnour 2). Each point represents mean 2 1 SEM for 3 pairs of sections. Non-linear regression analysis was used to fit the data to the Langmuir binding isotherm and extract values for BLmax and Kd (see text). BLmax was 0.38 t 0.02 fmol . mg~-’ for tumour 1 and 0.043 ? 0.01 fmol . mg-I for tumour 2. Kd was 0.90 t 0.15 nM for tumour 1 and 0.57 t 0.11 nM for tumour 2.

P P

FIGURE 4 - Specific binding of 12sI-ATII SI to peripheral turnour regions (open symbols) and central tumour regions (closed sym- bols). Each pair of symbols represents a single tumour.

was more apparent at the tumour periphery than at the tumour centre. It is evident that a pixel to pixel comparison of blood flow and specific binding of AT11 to its receptors would not yield a good correlation. This is not surprising, because ATII-induced vasoconstriction, primarily at the arteriolar level, would induce reduction in blood flow “downstream” from these sites. Quantitation of binding showed that the P22 tumour possesses relatively few binding sites for ATII. BLma was 0.38 fmol . mg-I compared to previously published values of 8-36 fmol . mg-I for normal human kidney (Grone et al., 1992), 14.5-17.5 fmol . mg-’ for normal rat kidney (Gehlert et al., 1984) and 2.5-10.7 fmol. mg-I for normal rat brain (Gehlert et al., 1986) using autoradiography. Other authors have also reported a loss of AT11 receptors in malignant tissue (Bryson et al., 1992; Sitzmann et al., 1994). This is consistent with our previous finding that vasoconstriction in the P22 tumour tends to be less than that in most normal tissues (Tozer and Shaffi, 1993). The Kd for the P22 tumour was of the same order as for various normal tissues, indicating a similar affinity of ligand for its receptors: 1.2 nM for tumour compared with 0.14-0.89 nM for normal human kidney (Grone et al., 1992), 0.20-0.85 nM for normal rat kidney (Gehlert et al., 1984) and 0.5-2.0 nM for normal rat brain (Gehlert et al., 1986).

receptor binding characteristics are predictive of physiological response of the P22 tumour to exogenously administered ATII. EIOW- ever, the rather uniform distribution of specific binding of lz5I-ATII SI in central tumour regions suggests that binding is

Our results support the hypothesis that FIGURE 5 - Computed images of specific binding of 12SI-ATII SI (a) and blood flow (b) in adjacent sections from the turnour. Note the very narrow, incomplete rim of high specific binding in (a) and the region of negligible specific binding in (a) that corresponds to a region of very low blood flow in (b).

TOZER ETAL. 662

REFERENCES

associated with tumour parenchymal cells as well as the vasculature. This has been amply demonstrated in tumour cells growing in vitro (Chen et al., 1993) and is likely to be associated with ATII's growth-promoting characteristics. Low specific binding in areas of low blood flow suggests that stability of AT11 binding sites is dependent on in vivo microenvironmental factors such as oxygenation or glucose availability. In order to determine the location of specific binding in our tumours it would be necessary to use photographic emulsion applied directly to fixed tumour sections for higher resolution autoradi- ography. Alternatively, immunohistochemical techniques for localisation of AT11 receptors are now becoming available (Reagan et al., 1994). Use of a non-peptide antagonist specific to the AT1 receptor sub-type, which is associated with vasocon- striction (Timmermans et al., 1992), may also help to clarify the situation.

The response of tumours to vasoactive agents is highly unpredictable, and this currently limits the usefulness of this approach for improving delivery of blood-borne anti-cancer agents to tumours. Absolute blood flow to some rodent tumours has been found to increase in response to exogenously administered AT11 (Tokuda et at., 1990; Hori et a/., 1991; Tanda et al., 1991). The results from the present study suggest that this may be due to a deficiency of AT11 receptors even compared to the P22 tumour. The tumours used in studies showing improved blood flow were very large, such that the

contribution of the periphery, where receptor density should be high, would be relatively small. A differential response to AT11 between small and large tumours has, indeed, been reported (Thews et al., 1995), although it has also been reported (Hemingway et al., 1992) that AT11 has very little effect on blood flow to very large (> 5 cm diameter) colorectal liver metastases in human patients despite a significant rise in blood flow to smaller deposits, as measured by laser Doppler flowmetry. In these very large confluent tumours, it may be that a high proportion of the tissue is suffering from temporary or permanent ischaemia, which would be unaffected by ATII- induced hypertension. A systematic study of the effects of tumour size on blood flow response to AT11 would be worthwhile.

In conclusion, a spatially heterogeneous blood flow response to AT11 has been found for the P22 tumour, which is consistent with the spatial heterogeneity of receptor binding. This encour- ages the development of techniques for analysis of receptor binding characteristics for predicting the response of indi- vidual tumours to blood flow manipulation using vasoactive agents.

ACKNOWLEDGEMENTS

We thank Gray Laboratory staff for care of the animals and the Cancer Research Campaign for funding this work.

ANDERSON, N.H., WILLMOTT, N., BESSENT, R., ANGERSON, W.J., KERR, D.J. and MCARDLE, C.S., Regional chemotherapy for inoper- able renal carcinoma: a method of targeting therapeutic microspheres to tumour.Brit. J. Cancer, 64,365-368 (1991). BRYSON, S.E., WARBURTON, P., WINTERSGILL, H.P., DREW, G.M., MICHEL, A.D., BALL, S.G. and BALMFORTH, A.J., Induction of the angiotensin AT2 receptor subtype expression by differentiation of the neuroblastoma x glioma hybrid, NG-1080-15. Europ. J. Pharmacol.,

CHEN, L., PRAKASH, 0. and RE, R.N., The interaction of insulin and angiotensin I1 on the regulation of human neuroblastoma cell growth. Mol. chem. Neuropathol., 18,189-196 (1993). GEHLERT, D.R., SPETH, R.C. and WAMSLEY, J.K., Autoradiographic localisation of angiotensin I1 receptors in the rat brain and kidney. Europ. J. Pharmacol., 98,145-146 (1984). GEHLERT, D.R., SPETH, R.C. and WAMSLEY, J.R., Distribution of ['2SI]angiotensin I1 binding in the rat brain: a quantitative autoradio- graphic study. Neuroscience, 18,837-856 (1986). GRONE, H.-J., SIMON, M. and FUCHS, E., Autoradiographic character- ization of angiotensin receptor sub-types in fetal and adult human kidney. Amer. J. Physiol., 262 (Renal Fluid Electrolyte Physiol., 31),

HEMINGWAY, D.M., ANGERSON, W.J., ANDERSON, J.H., GOLDBERG, J.A., MCARDLE, C.S. and COOKE, T.G., Monitoring blood flow to colorectal liver metastases using laser Do ler flowmetry: the effect of angiotensin 11. Brit. J. Cancer, 66,958-9688992). HORI, K., SUZUKI, M., TANDA, S., SAITO, S., SHINOZAKI, M. and ZHANG, Q.H., Fluctuations in tumor blood flow under normotension and the effect of angiotensin 11-induced hypertension. Jpn. J. Cancer Res., 82,1309-1316 (1991). JIRTLE, R., CLIFTON, K.H. and RANKIN, J.H.G., Effects of several vasoactive drugs on the vascular resistance of MT-W9B tumours in WiFu rats. Cancer Res., 38,2385-2390 (1978). KERR, D.J., GOLDBERG, J.A., ANDERSON, J.R., WILMOTT, N., WHAT- ELEY, A.T., MCARDLE, C.S. and MCKILLOP, J., The effect of angioten- sin I1 on tumor blood flow and the delivery of microparticulate cytotoxic drugs. EXS, 61,340-345 (1992). KETY, S.S., Theory of blood tissue exchange and its application to measurements of blood flow. Meth. med. Res., 8,223-227 (1960). KOBAYASHI, H., HASUDA, K., AOKI, K., TANIGUCHI, S. and BABA, T., Systemic chemotherapy in tumour-bearing rats using high-dose cis- diamminedichloroplatinum (11) with low nephrotoxicity in combina- tion with angiotensin I1 and sodium thiosulfate. Inf. J. Cancer, 45,

225,119-127 (1992).

F326-F331 (1992).

940-944 ( 1990).

KOBAYASHI, H., HASUDA, K., TANIGUCHI, S. and BABA, T., Therapeutic efficacy of two-route chemotherapy using cis-diamminedichloroplati- num (11) and its antidote, sodium thiosulfate, combined with the angiotensin-11-induced hypertension method in a rat uterine tumor. Int. J. Cancer, 47,893-898 (1991). MATTHEWS, J.C., Fundamentals of receptor, enzyme, and transpori kinetics, CRC Press, Boca Raton, Florida (1993). MUTOH, S., AIKOU, I., SOEJIMA, K., UEDA, S., FUKUSHIMA, S., KISHI- MOTO, S. and TAKAGI, Y., Local control of prostate cancer by intraarterial infusion chemotherapy facilitated by the use of angioten- sin 11. Urol. Int., 48,175-180 (1992). NOGUCHI, S., MIYAUCHI, K., NISHIZAWA, Y., SASAKI, Y., IMAOKO, S., IWANAGA, T., KOYAMA, H. and TERASAWA, T., Augmentation of anti-cancer effect with angiotensin I1 in intra-arterial infusion chemo- therapy for breast carcinoma. Cancer, 62,467-473 (1988). REAGAN, L.P., FLANAGAN-CATO, L.M., YEE, D.K., MA, L.Y., SAKAI, R.R. and FLUHARTY, S.J., Immunohistochemical mapping of angioten- sin type 2 (AT2) receptors in rat brain. Brain Res., 662,45-59 (1994). SIMANTOV, R., SNOWMAN, A.M. and SNYDER, S.H., Temperature and ionic influences on opiate receptor binding. Mol. Pharmacol., 12,

SITZMANN, J.V., Wu, Y. and CAMERON, J.L., Altered angiotensin-I1 receptors in human hepatocellular and hepatic metastatic colon cancers.Ann. Surg., 219,500-507 (1994). SUZUKI, M., HORI, K., ABE, I., SAITO, S. and SATO, H., A new approach to cancer chemotherapy: selective enhancement of tumor blood flow with angiotensin 1I.J. nut. Cancer Inst., 67,663-669 (1981). TAKEMATSU, H., TOMITA, Y. and KATo, T., Angiotensin-induced hypertension and chemotherapy for multiple lesions of malignant melanoma. Brit. J. Dermatol., 113,463-465 (1985). TANDA, S., HORI, K., SAITO, S., SHINOZAKI, M., ZHANG, Q.H. and SUZUKI, M., Comparison of the effects of intravenously bolus- administered endothelin-1 and infused angiotensin I1 on the subcuta- neous tumor blood flow in anaesthetized rats. Jpn. J. Cancer Res., 82,

THEWS, O., KELLEHER, D.K. and VAUPEL, P.W., Modulation of spatial 0 2 tension distribution in experimental tumors by increasing arterial 0 2 supply. Acta Oncol., 34,291-295 (1995). TIMMERMANS, P.M.W.M., CHIU, A.T., HERBLIN, W.F., WONG, P.C. and SMITH, R.D., Angiotensin I1 receptor sub-types. Amer. J. Hyperten- sion, 5,406-410 (1992). TOKUDA, K., ABE, H., AIDA, T., SUGIMOTO, S. and KANEKO, S., Modification of tumor blood flow and enhancement of therapeutic

977-986 (1976).

958-963 (1991).

TUMOUR BLOOD FLOW AND AT11 RECEPTORS 663

effect of ACNU on experimental rat gliomas with angiotensin 11. J. Neurooncol., 8,205-212 (1990). TOZER, G.M. and MORRIS, C., Blood flow and blood volume in a transplanted rat fibrosarcoma: corn arison with various normal tis- sues. Radiother. Oncol., 17,153-166 8990). TOZER, G.M. and SHAFFI, K.M., Modification of tumour blood flow using the hypertensive agent, angiotensin 11. Brit. J. Cancer, 67, 981-988 (1993). TOZER, G.M. and SHAFFI, K.M., The response of tumour vasculature to angiotensin I1 revealed by its systemic and local administration to “tissue-isolated’ tumours. Brit. J. Cancer 72,595-600 (1995).

TOZER, G.M., SHAFFI, K.M. and HIRST, D.G., Use of the hypertensive agent angiotensin I1 for modifying oxygen delivery to tumours. In: P. Vaupel (ed.), oxygen transport to tissues pp. 423429, Plenum Press, New

TOZER, G.M., SHAFFI, K.M., PRISE, V.E. and CUNNINGHAM, V.J., Characterisation of turnour blood flow using a ‘yissue-isolated” preparation, ~ k t , J, cancer, 70,104@1046 (19946).

TROTTER, M.J., CHAPLIN, D.J. and OLIVE, P.L., Effect of angiotensin I1 on intermittent tumour blood flow and acute hypoxia in the murine SCCVII carcinoma. Europ. J. Cancer, 27,887-893 (1991).

(1994a).