Measurement of Mammary Tumor Blood Flow in Unanesthetized Rats 1,2 Randy Jirtle; Kelly H. Clifton, 3 and John H. G. Rankin 4, 5, 6 ABSTRACT-Blood flow to mammary tumor MT-W9B in minimally disturbed, unanesthetized inbred WF rats was measured with radioactively labeled 25-p. (diameter) microspheres. The distribution of the blood flow estimates was not described by a normal probability density function but could be normalized with a natural log transformation of the data. The relative blood flow to the tumor tissue was always greater than that to the surrounding normal tissue and was dependent on the tumor size, transplantation site (axillary vs. inguinal), and physiologic state of the animal. The growth of this tumor did not alter the blood flow to the adjacent normal tissues. However, the cardiac output and coronary blood flow increased with increasing tumor size. Two blood flow estimates could be performed on the same animal within a time-span of 15 minutes.-J Nail Cancer Inst 60: 881886,1978.
The effectiveness of tumor treatment with radiation, chemicals, and heat depends indirectly on tumor vasculature and blood flow. To more accurately predict the efficacy of such treatments, one should correctly deter-' mine the blood flow to the malignant and the surrounding normal tissue in conscious animals. Although the vasculature of tumors has been extensively studied (1-5) since Van der Kolk (6) first did injection experiments in neoplastic disease, it was not until 1961 that a direct estimate of tumor blood flow was reported by Gullino and Grantham (7). Indirect methods of determining tumor blood flow have since been primarily used. Of these, the 86Rb method described by Sapirstein (8) for assessment of the distribution of cardiac output is the most widely employed (7, 9-11 ). To estimate regional blood flow by this technique, one must simultaneously determine the cardiac output. Radioactively labeled antipyrine (12-14) and 133Xe (15) have also been used as reference substances. Most tumor blood flow estimates have been done in anesthetized animals. Kallman et al. (15) found that the tumor blood flow was lower in anesthetized than in unanesthetized animals. However, their estimate probably still did not reflect tumor blood flow in undisturbed conditions, inasmuch as the animals were handled and received intratumor injections of 133Xe. We applied the microsphere technique, previously used to measure blood flow to rabbit tumors (16), to measure the tumor blood flow in unanesthetized, minimally disturbed rats.
ABBREVIATIONS USED: cpm naturaL
counts per minute; In
logarithm,
Received August 10, 1977; accepted November 2, 1977. Supported by Public Health Service grant CA18756 from the National Cancer Institute, by grant PDT-46R from the American Cancer Society, and by a grant (to R. J.) from the Graduate School of the University of Wisconsin. a Radiobiology Laboratories, Departments of Human Oncology and Radiology, University of Wisconsin Medical School, Madison, Wis. 53706. 4 Departments of Physiology and Gynecology-Obstetrics, University of Wisconsin Medical School; and Wisconsin Perinatal Center, Madison General Hospital, Madison, Wis. 53715, 'Address reprint requests to Dr. John Rankin, Madison General Hospital, 202 South Park St., Madison, Wis. 53715. 6 We thank Terry M. Phernetton for excellent technical assistance. 1
2
MATERIALS AND METHODS Animals and tumor system.-Isogeneic female inbred WF rats (=240 g) were housed in hanging cages in a temperature-controlled room with 12 hours of light daily. Food and water were given ad libitum. Suspensions of MT-W9B mammary carcinomas (17) were preVOL. 60, NO.4, APRIL 1978
pared for transplantation with a Snell cytosieve as described in (18) and adjusted to a 33% volume of centrifugally packed cellular material. Inocula (0.10 ml) were injected into both axillary glands and the right inguinal mammary glands. Surgical technique.-When the tumors reached the desired size, the rats were surgically prepared under ether anesthesia. A PESO catheter (inside diameter, 0.023 inches; outside diameter, 0.038 inches) filled with heparinized isotonic saline (25 D/ml) was placed into the left ventricle of the heart via the right common carotid artery. A PElO catheter (inside diameter, 0.011 inches; outside diameter, 0.024 inches) that had previously been fused by heat with a PESO catheter was filled with heparinized isotonic saline (25 D/ml), and the PEIO segment was inserted 1.5 cm into the left femoral artery. Both catheters were tied off, placed subcutaneously, and externalized at the dorsal aspect of the neck. All animals were returned to cages and allowed to recover for at least 3 hours before the microspheres were injected. Microsphere technique.-Approximately 70,000 radioactively labeled microspheres (diameter, 25 /L) were sealed in PESO catheters (IS cm in length) and placed in wide-mouth gamma counting vials. The spheres were labeled with either 125 1, 46S C , 85S r , 141Ce (3M Co., St. Paul, Minn.), or 109Cd (New England Nuclear, Boston, Mass.). A microsphere standard for each isotope was prepared as described in (19), and the average cpm per sphere were measured with an NaI well counter equipped with a pulse height analyzer. Radioactivity in the sphere-containing catheters was measured with the same gamma counter, and since the cpm per sphere were known, the number of spheres contained in each catheter could be calculated. For spheres labeled with 125 1, I09Cd, and 141Ce, the error in this estimation was approximately 2%.
881
J
NATL CANCER INST
882
JIRTLE, CLIFTON, AND RANKIN
Both ends of a sphere-containing catheter were cut off prior to use and placed in the counting vial. One end was connected to the externalized portion of the PESO catheter residing in the left ventricle of the heart. The other end was connected to a syringe containing 0.6 ml heparinized isotonic saline (10 U/ml). The'microspheres were dispersed by gentle movement of the syringe plunger. The femoral catheter was attached to a stainless steel "T"-connector that was in turn joined to a Harvard withdrawal pump and a Statham P23Db pressure transducer. With this arrangement, the average systemic pressure could be continuously monitored on a chart recorder whenever blood was not being withdrawn from the femoral artery. Because the connection procedure usually disturbed the animal, the rat was allowed to rest before the microspheres were injected (=S-lO min). During this period, the left ventricle pressure changes maintained the dispersal of the microspheres. When the monitored pressure demonstrated that the animal was in a resting state (i.e., constant at 100 mmHg), the connector to the pressure transducer was damped, the connector to the withdrawal pump undamped, and the withdrawal pump started. Once it was confirmed that the blood was being withdrawn at a constant rate, the 25-p.. microspheres were gradually flushed into the left ventricle with 0.6 ml heparinized isotonic saline. The pump was allowed to withdraw blood for exactly 1 minute after all the spheres were injected, and then the withdrawn blood sample was flushed into a counting vial with 4 ml water. The time required for the procedure was approximately I.S minutes, and the total quantity of blood removed was 0.7S ml [=S% of the total blood volume (20)]. Tissue blood flow could be measured a second time by the same procedure with microspheres with a different label. The animals were killed by injection of 0.1 ml euthanasia solution (Veterinary Laboratories Inc., Lenexa, Kans.) into the left ventricle of the heart. Normal and malignant tissues were removed, dissected free of extraneous surrounding tissue, weighed, and placed into wide-mouth counting vials. The normal tissues that were removed included the soleus and adjacent muscles, the right and left kidneys, the uterine horns, a IO-cm section of the duodenum, the heart, the whole brain, and the mammary gland and skin from the axillary and inguinal regions. The IS-cm catheter that previously contained the spheres was also removed and placed into the counting vial which contained its end pieces. Sample radioactivity was determined with an NaI well counter and converted into the number of spheres by the use of microsphere standards. The blood flow to various tissues and cardiac output were calculated by the formulas: Fr=(WR/NB)·N r and CO=(WR/NB)'NJ, where: FT=tissue blood flow (ml/ min), CO=cardiac output (ml/min), WR=arterial blood withdrawal rate (ml/min), NB=number of microspheres contained in the withdrawn blood sample, NT=number of micro spheres contained in the tissue, andN 1 =number of microspheres initially injected into the left ventricle of the heart. The value of N I was calculated by subtractI NATL CANCER INST
ing the number of spheres remammg in the catheter pieces after injection from the number of spheres initially contained in the sealed catheter. A nesthetized animals . -The surgical preparation of the rats to be anesthetized prior to blood flow estimation was identical to that of unanesthetized animals. However, after each animal recovered from surgery, 0.15 ml Nembutal (SO mg/ml) was flushed into the left ventricle of the heart with 0.5 ml isotonic saline. The blood flows were measured 20 minutes thereafter by the microsphere technique as described for unanesthetized animals. Statistical analysis.-A X2 test was used to determine whether the un transformed or In-transformed blood flow data were normally distributed. The nonparametric Spearman's p test for correlation was used to determine the extent of correlation between two variables (21). A two-sample t-test assuming unequal variances was used to compare the average In-transformed tissue blood flows in unanesthetized animals to those of the respective tissues in anesthetized animals. The mammary gland and the skin blood flows in the axillary regions were compared to those in the inguinal region by a paired t-test. Analysis of variance (ANOV A) was used for the comparison of more than two independent sample means.
RESULTS Distribution of Blood Flow Data A X2 test was used to determine whether the distributions for the tissue blood flow estimates were adequately described by a normal probability density function. To perform these tests, we generated normal distributions with the same means and standard deviations as the tissue blood flow frequency distributions. A X2 test was used to compare the observed frequency to the frequency that would be expected if the distribution was described by a normal probability density function. For example, when this test was performed on the blood flow data for the mammary gland (text-fig. 1), the X2 value was 16.2 with 6 degrees of freedom (P=O.Ol). We therefore concluded that the frequency distribution of the mammary gland blood flow data was not normally distributed; this conclusion was true for all tissues, whether normal or malignant. When a similar X2 test for normality was done on the In-transformed data for the mammary gland (text-fig. 1), the X2 value was 3.3 with S degrees of freedom. The P-value of 0.6 demonstrated that the In-transformation adequately normalized the mammary gland blood flow data. The In transformation similarly normalized the blood flow data for the other tissues studied.
Effect of Blood Withdrawal Rate and Time Femoral arterial blood was removed at a rate of 1.03 ml/minute for O.S minutes, O.SI ml/minute for 1 minute, or O.SI ml/minute for 2 minutes. The average estimated VOL. 60, NO.4, APRIL 1978
883
MEASUREMENT OF TUMOR BLOOD FLOW
-
I.-Ratio of blood flows estimated with 15-, 25-, and 50-/k (diameter) microspheres in MT-W9B tumor-bearing rats
TABLE
f--I
,,
r-
The effectiveness of tumor treatment with radiation, chemicals, and heat depends indirectly on tumor vasculature and blood flow. To more accurately predict the efficacy of such treatments, one should correctly deter-' mine the blood flow to the malignant and the surrounding normal tissue in conscious animals. Although the vasculature of tumors has been extensively studied (1-5) since Van der Kolk (6) first did injection experiments in neoplastic disease, it was not until 1961 that a direct estimate of tumor blood flow was reported by Gullino and Grantham (7). Indirect methods of determining tumor blood flow have since been primarily used. Of these, the 86Rb method described by Sapirstein (8) for assessment of the distribution of cardiac output is the most widely employed (7, 9-11 ). To estimate regional blood flow by this technique, one must simultaneously determine the cardiac output. Radioactively labeled antipyrine (12-14) and 133Xe (15) have also been used as reference substances. Most tumor blood flow estimates have been done in anesthetized animals. Kallman et al. (15) found that the tumor blood flow was lower in anesthetized than in unanesthetized animals. However, their estimate probably still did not reflect tumor blood flow in undisturbed conditions, inasmuch as the animals were handled and received intratumor injections of 133Xe. We applied the microsphere technique, previously used to measure blood flow to rabbit tumors (16), to measure the tumor blood flow in unanesthetized, minimally disturbed rats.
ABBREVIATIONS USED: cpm naturaL
counts per minute; In
logarithm,
Received August 10, 1977; accepted November 2, 1977. Supported by Public Health Service grant CA18756 from the National Cancer Institute, by grant PDT-46R from the American Cancer Society, and by a grant (to R. J.) from the Graduate School of the University of Wisconsin. a Radiobiology Laboratories, Departments of Human Oncology and Radiology, University of Wisconsin Medical School, Madison, Wis. 53706. 4 Departments of Physiology and Gynecology-Obstetrics, University of Wisconsin Medical School; and Wisconsin Perinatal Center, Madison General Hospital, Madison, Wis. 53715, 'Address reprint requests to Dr. John Rankin, Madison General Hospital, 202 South Park St., Madison, Wis. 53715. 6 We thank Terry M. Phernetton for excellent technical assistance. 1
2
MATERIALS AND METHODS Animals and tumor system.-Isogeneic female inbred WF rats (=240 g) were housed in hanging cages in a temperature-controlled room with 12 hours of light daily. Food and water were given ad libitum. Suspensions of MT-W9B mammary carcinomas (17) were preVOL. 60, NO.4, APRIL 1978
pared for transplantation with a Snell cytosieve as described in (18) and adjusted to a 33% volume of centrifugally packed cellular material. Inocula (0.10 ml) were injected into both axillary glands and the right inguinal mammary glands. Surgical technique.-When the tumors reached the desired size, the rats were surgically prepared under ether anesthesia. A PESO catheter (inside diameter, 0.023 inches; outside diameter, 0.038 inches) filled with heparinized isotonic saline (25 D/ml) was placed into the left ventricle of the heart via the right common carotid artery. A PElO catheter (inside diameter, 0.011 inches; outside diameter, 0.024 inches) that had previously been fused by heat with a PESO catheter was filled with heparinized isotonic saline (25 D/ml), and the PEIO segment was inserted 1.5 cm into the left femoral artery. Both catheters were tied off, placed subcutaneously, and externalized at the dorsal aspect of the neck. All animals were returned to cages and allowed to recover for at least 3 hours before the microspheres were injected. Microsphere technique.-Approximately 70,000 radioactively labeled microspheres (diameter, 25 /L) were sealed in PESO catheters (IS cm in length) and placed in wide-mouth gamma counting vials. The spheres were labeled with either 125 1, 46S C , 85S r , 141Ce (3M Co., St. Paul, Minn.), or 109Cd (New England Nuclear, Boston, Mass.). A microsphere standard for each isotope was prepared as described in (19), and the average cpm per sphere were measured with an NaI well counter equipped with a pulse height analyzer. Radioactivity in the sphere-containing catheters was measured with the same gamma counter, and since the cpm per sphere were known, the number of spheres contained in each catheter could be calculated. For spheres labeled with 125 1, I09Cd, and 141Ce, the error in this estimation was approximately 2%.
881
J
NATL CANCER INST
882
JIRTLE, CLIFTON, AND RANKIN
Both ends of a sphere-containing catheter were cut off prior to use and placed in the counting vial. One end was connected to the externalized portion of the PESO catheter residing in the left ventricle of the heart. The other end was connected to a syringe containing 0.6 ml heparinized isotonic saline (10 U/ml). The'microspheres were dispersed by gentle movement of the syringe plunger. The femoral catheter was attached to a stainless steel "T"-connector that was in turn joined to a Harvard withdrawal pump and a Statham P23Db pressure transducer. With this arrangement, the average systemic pressure could be continuously monitored on a chart recorder whenever blood was not being withdrawn from the femoral artery. Because the connection procedure usually disturbed the animal, the rat was allowed to rest before the microspheres were injected (=S-lO min). During this period, the left ventricle pressure changes maintained the dispersal of the microspheres. When the monitored pressure demonstrated that the animal was in a resting state (i.e., constant at 100 mmHg), the connector to the pressure transducer was damped, the connector to the withdrawal pump undamped, and the withdrawal pump started. Once it was confirmed that the blood was being withdrawn at a constant rate, the 25-p.. microspheres were gradually flushed into the left ventricle with 0.6 ml heparinized isotonic saline. The pump was allowed to withdraw blood for exactly 1 minute after all the spheres were injected, and then the withdrawn blood sample was flushed into a counting vial with 4 ml water. The time required for the procedure was approximately I.S minutes, and the total quantity of blood removed was 0.7S ml [=S% of the total blood volume (20)]. Tissue blood flow could be measured a second time by the same procedure with microspheres with a different label. The animals were killed by injection of 0.1 ml euthanasia solution (Veterinary Laboratories Inc., Lenexa, Kans.) into the left ventricle of the heart. Normal and malignant tissues were removed, dissected free of extraneous surrounding tissue, weighed, and placed into wide-mouth counting vials. The normal tissues that were removed included the soleus and adjacent muscles, the right and left kidneys, the uterine horns, a IO-cm section of the duodenum, the heart, the whole brain, and the mammary gland and skin from the axillary and inguinal regions. The IS-cm catheter that previously contained the spheres was also removed and placed into the counting vial which contained its end pieces. Sample radioactivity was determined with an NaI well counter and converted into the number of spheres by the use of microsphere standards. The blood flow to various tissues and cardiac output were calculated by the formulas: Fr=(WR/NB)·N r and CO=(WR/NB)'NJ, where: FT=tissue blood flow (ml/ min), CO=cardiac output (ml/min), WR=arterial blood withdrawal rate (ml/min), NB=number of microspheres contained in the withdrawn blood sample, NT=number of micro spheres contained in the tissue, andN 1 =number of microspheres initially injected into the left ventricle of the heart. The value of N I was calculated by subtractI NATL CANCER INST
ing the number of spheres remammg in the catheter pieces after injection from the number of spheres initially contained in the sealed catheter. A nesthetized animals . -The surgical preparation of the rats to be anesthetized prior to blood flow estimation was identical to that of unanesthetized animals. However, after each animal recovered from surgery, 0.15 ml Nembutal (SO mg/ml) was flushed into the left ventricle of the heart with 0.5 ml isotonic saline. The blood flows were measured 20 minutes thereafter by the microsphere technique as described for unanesthetized animals. Statistical analysis.-A X2 test was used to determine whether the un transformed or In-transformed blood flow data were normally distributed. The nonparametric Spearman's p test for correlation was used to determine the extent of correlation between two variables (21). A two-sample t-test assuming unequal variances was used to compare the average In-transformed tissue blood flows in unanesthetized animals to those of the respective tissues in anesthetized animals. The mammary gland and the skin blood flows in the axillary regions were compared to those in the inguinal region by a paired t-test. Analysis of variance (ANOV A) was used for the comparison of more than two independent sample means.
RESULTS Distribution of Blood Flow Data A X2 test was used to determine whether the distributions for the tissue blood flow estimates were adequately described by a normal probability density function. To perform these tests, we generated normal distributions with the same means and standard deviations as the tissue blood flow frequency distributions. A X2 test was used to compare the observed frequency to the frequency that would be expected if the distribution was described by a normal probability density function. For example, when this test was performed on the blood flow data for the mammary gland (text-fig. 1), the X2 value was 16.2 with 6 degrees of freedom (P=O.Ol). We therefore concluded that the frequency distribution of the mammary gland blood flow data was not normally distributed; this conclusion was true for all tissues, whether normal or malignant. When a similar X2 test for normality was done on the In-transformed data for the mammary gland (text-fig. 1), the X2 value was 3.3 with S degrees of freedom. The P-value of 0.6 demonstrated that the In-transformation adequately normalized the mammary gland blood flow data. The In transformation similarly normalized the blood flow data for the other tissues studied.
Effect of Blood Withdrawal Rate and Time Femoral arterial blood was removed at a rate of 1.03 ml/minute for O.S minutes, O.SI ml/minute for 1 minute, or O.SI ml/minute for 2 minutes. The average estimated VOL. 60, NO.4, APRIL 1978
883
MEASUREMENT OF TUMOR BLOOD FLOW
-
I.-Ratio of blood flows estimated with 15-, 25-, and 50-/k (diameter) microspheres in MT-W9B tumor-bearing rats
TABLE
f--I
,,
r-
