Academic Editor:Nejla Fourati, Radiation Oncology Specialist, Habib Bourguiba Hospital. Member of the Faculty of Medicine, University of Sfax
Checked for plagiarism: Yes
Review by: Single-blind
Tumor Growth Dynamics: Dietary Fish Oil Induced Inhibition of Human Breast Carcinoma Growth, A Phenomenon of Reduced Cellular DNA Synthesis or Increased Cell Loss?
Diets high in unsaturated fatty acids, especially those containing high levels of linoleic acid, e.g., corn oil, enhance mammary gland tumorigenesis in experimental animals. In contrast, diets high in long-chain polyunsaturated fatty acids such as eicosapentaenoic (EPA) and docosahexaenoic (DHA), e.g. menhaden oil, appear to have a suppressive effect on this tumorigenic process. Many mechanisms have been proposed to explain the tumor inhibitory action exerted by menhaden oil and other fish oils, e.g., differences in prostaglandin metabolism, energy efficiency, alterations of the immune system, changes in lipid peroxidation, etc. Fundamental to a mechanistic understanding of this phenomenon, however, is an understanding as to whether or not the tumor inhibitory activities of dietary fish oil is mediated via an inhibition of tumor cell proliferation or mediated via an enhancement of tumor cell loss. Whether the amount of dietary fat or the type of fat effects mammary tumorigenic processes, via an effect on tumor cell proliferation or tumor cell loss, has not been clearly established. In the studies described in this communication, three methods were utilized to study tumor cell proliferation, i.e., H3-thymidine autoradiographic analysis, 5-bromo 2'-deoxyuridine (Brdu) flow cytometric analysis, and proliferative cell nuclear antigen (PCNA) flow cytometric analysis. Two methods were used to study tumor cell loss, i.e., a determination of the I125Urd tumor emission rate and a determination of a cell loss factor from the formulas of Steel and Begg. The tumor examined was the human breast carcinoma cell line MDA- MB231 maintained in athymic nude mouse. No significant difference in cell proliferation between carcinomas of mice fed a high corn oil diet (20% w/w) and a diet high in fish oil (19% menhaden oil, 1% corn oil). In contrast, a significant (p<0.05) increase in the rate of I125Urd emission rate and cell loss factor from the carcinomas in the fish oil fed mice compared to the corn oil fed mice was observed. In summary, the decreased tumor volume in the human breast carcinomas maintained in athymic nude mice fed a fish oil diet as compared to those fed a corn oil diet, appears to be due, at least in part, to an increased rate of carcinoma cell loss rather than a decreased rate of carcinoma cell proliferation.
The mechanism by which diets high in certain fats such as corn oil are capable of enhancing mammary tumorigenesis in rodents 1 and increase human breast carcinoma size in athymic nude mice 2 is unclear. Moreover, the mechanism by which diets high in long-chain polyunsaturated fatty acids (PUFA) such as fish oils, can effectively suppress mammary tumorigenesis in rodents 1 and inhibit human breast carcinoma growth in athymic nude mice 2, 3 also remains to be determined. This raises a fundamental question in tumor biology that has not been rigorously examined. How do dietary fats exert their enhancing or suppressive activity at a tumor growth kinetic (tumor cell loss vs tumor cell proliferation) level? Only a few research groups 4, 5, 6 have examined cell proliferation in tumors of animals fed high fat diets. One group 7 reports no difference in cell proliferation rates in transplantable mouse mammary tumors from animals fed diets composed of unsaturated vs saturated fats. In another study, Abraham et al. 8 hypothesized that the increase in tumor size induced by a corn oil diet compared to hydrogenated cottonseed oil or fish (menhaden) oil diet was due to a decrease in cell loss (cell death) as a result of the high corn oil diet impairing immune system activity.
In other studies 5, 6, 9, 10, an increase in cell proliferation was observed by an increase in H3-thymidine incorporation into DNA of carcinogen-induced rat mammary tumors from rats fed a high corn oil diet compared to those fed lower levels of corn oil. They concluded that this increase in mammary tumor growth was due to an increase in carcinoma cell proliferation by providing diets high in corn oil. Our study was designed to determine if the growth of a human breast carcinoma cell line (MDA-MB231) in vivo (athymic nude mice), as a function of feeding high levels of either corn oil or fish oil (menhaden), is due to changes in carcinoma cell proliferation and/or changes in carcinoma cell loss. The knowledge of how dietary fats can affect mammary carcinoma growth dynamics is critical to a mechanistic understanding of nutritional tumorigenesis.
Female athymic nude mice (Harlan Sprague-Dawley Inc., Madison, WI) 4-5 weeks old were used in these experiments. The mice were maintained under aseptic conditions which included an enclosed overhead laminar flow hood and were housed in sterilized cages, with sterilized bedding and provided sterilized drinking water in a temperature-controlled (24*C) and light-controlled (14 h/day) room. Autoclaved laboratory mouse chow (Purina Mills Inc., St. Louis, MO) was fed ad libitum before and until 7-10 days after human breast carcinoma transplantation. Thereafter mice were fed ad libitum purified diets for 4 to 6 weeks (unless indicated otherwise) (Table 1). All dietary ingredients were obtained from U.S. Biochemicals Inc. (Cleveland, OH) except sucrose, which was obtained from ICN Biochemicals Inc. (Costa Mesa, CA), and fish oil (menhaden), which was obtained from Zapata Haynie Corp. (Reedville, VA). The percentages of predominant fatty acids (1% or greater, manufacturer's specifications) of the dietary oils are shown in Table 2. The diets were prepared weekly and stored at -20°C, individually packed in small plastic sealed bags of sufficient size for one day's feed. Mice were fed daily and non-consumed food discarded daily. Since purified diets were not sterilized, antibiotics (Bacitracin combined with Streptomycin or Neomycin, 1 g/L) were added to the distilled drinking water.Table 1. Diet composition
|g/100 g diet|
|AIN Mineral mix4||4.13|
|AIN Vitamin mix4||1.18|
|Fatty acidsa||Corn oil||Menhaden oil|
Palpable MDA-MB231 human breast carcinomas (American Type Culture Collection, Rockville, MD) were surgically excised from female athymic mice, cut into slices (2x4 mm, 0.1-0.3 mm thick) and implanted into recipient female athymic mice under aseptic conditions. Mice were anesthetized with sodium pentobarbital (60 µg/g,i.p.) prior to transplantation. An incision was made in the integument, the tumor slices were placed s.c. in the dorsum at distances from each other of at least 2 cm, 3 to 4 slices/mouse (autoradiograph and flow cytometer experiments). One slice per mouse was placed in the middle of the upper back between the shoulder blades in the animals used for the cell loss experiments. The carcinoma grafts were established in the host animals before the onset of experimental dietary treatments.
After being fed the diet for 5 to 6 weeks mice were sacrificed. The tumors were excised and cut into slices (1-2 mm). Tumor slices were incubated in 10x30 mm Falcon disposable Petri dishes (2 slices/dish) containing 2.5 ml of medium (10X Waysmouth MB 752/1 medium, GIBCO Labs, Grand Island, NY). Per 100 ml of media, the following constituents were added: 35 mg glutamine, 3.5 mg penicillin and 125 mg of sodium bicarbonate. Sterile H3-thymidine (45 Ci/nmol,
New England Nuclear, Boston, MA) was added at a concentration of 1 µCi/ml of medium. The Petri dishes were placed in a small gassing chamber, housed in an incubator at 37°C. The chamber was continuously infused with gas 95% 02: 5% C02 for a 4 hr incubation period. The slices were then fixed in Bouins Fluid, embedded in a paraffin preparation (Tissue-prep, Fisher Scientific Co., Fairlawn, NJ), sectioned at 5-7 /m and mounted on glass slides. Two series of tissue sections were prepared; one series was stained with hematoxylin and eosin (H & E) and the other series was used for the autoradiographs.
The slides for autoradiography were dipped in NTB2 nuclear tract emulsion (Eastman Kodak Co., Rochester, NY), dried and stored away from light in tight black boxes with a desiccant for 14 days at 4°C. After two weeks, the slides were developed and stained by H & E using a standard method 11. The slides were then coded (identity of treatment unknown) and the number of H3-thymidine labelled breast carcinoma cells per area was computed for each carcinoma of both dietary groups. Group mean differences between labelled cells were evaluated statistically by the students t-test.
One hour prior to sacrifice, mice were injected i.p. with 5-bromo 2'-deoxyuridine (Brdu) at a concentration of 50 mg/kg body weight (Sigma Chemical Co., St. Louis, MO). After 1 hr, animals were terminated by an overdose of C02 and the tumors excised. Necrotic tissue was trimmed from the tumorand a 7 mm biopsy punch was used to obtain a tumor sample. The sample was minced with a single-edged razor blade and placed in a 12x75 mm glass tube containing 2 ml of ice cold 70% ETOH. The tubes were then sealed with parafilm and stored at -20°C until the dissociation step. Tissue samples were removed from -20°C storage and approximately 40 mg of tissue was finely minced with a scalpel or single-edged razor blade. The tissue was then simultaneously dissociated and denatured by placing in a 25 ml Erlenmeyer flask containing 2 ml of 0.4 mg/ml pepsin in 2 N HCl. The flasks were placed in a shaking water bath at room temperature for 1.5 to 2 hours or until the cells could be easily dispersed by gentle up and down pipetting with a pasteur pipette. The cell suspensions were then filtered through a 50 µg mesh and washed twice with 2 ml of PBS (pH 7.4) containing 0.1% BSA and 0.05% Tween 20 (PBT buffer).
Approximately 5x10s to 1xl06 cells were resuspended in 100 µl of PBT buffer containing 2 µg/ml of anti-Brdu antibody (Boehringer- Mannheim Co., Indianapolis, IN). The tubes were incubated for 30 min at room temperature, washed with 2 ml of PBT buffer and resuspended in 100 µg/ml of PBS containing 10 ^g/ml goat antimouse IgG-FITC. The tubes were incubated for 30 min at room temperature, washed twice with PBS and resuspended in 1 ml of PBS containing 10µg/ml propidium iodide (Sigma Chemical Co., St. Louis, MO). A control for non-specific binding was run for each sample by preparing a duplicate tube with no anti-Brdu. The tubes were incubated overnight at 4°C and analyzed using a flow cytometer (Ortho 50H, Ortho Diagnostics, Westward, MA).
The amount of Brdu uptake was reported as the percent of cells with green fluorescence intensity above that of the non-specific binding control. Mice that were not injected with Brdu were used as a control to eliminate background fluorescence.
The cold 70% ethanol-fixed tumors prepared for the Brdu assay were also used for the PCNA assay. The tissue samples were removed from the 70% ethanol, rinsed and placed in an Erlenmeyer flask containing 2 ml of pepsin (0.4 mg/ml in 0.1N HC1). Dissociation was carried out on a shaking water bath at room temperature for 30-60 min or until the cells were easily dispersed by gentle up and down pipetting with a Pasteur pipette. Cells were washed twice in PBS (pH 7.4) containing 0.1% Triton-X 100. Cells were then suspended in 100 µl PBS containing 25 µg/ml of PCNA (Boehringer-Mannheim Co., Indianapolis, IN) and 1% BSA. Cells were then incubated at room temperature for 30 min. Cells were washed and resuspended in 100 µl of PBS containing 10 ng/ml goat anti-mouse FITC. Tubes were incubated for 30 min at room temperature, washed twice with PBS and resuspended in 1 ml of PBS containing 10 of propidium iodide. The cold 70% ethanol-fixed tumors prepared for the Brdu assay were also used for the PCNA assay.
The tissue samples were removed from the 70% ethanol, minced and placed in an Erlenmeyer flask containing 2 ml of pepsin (0.4 mg/ml in 1N HC1). Dissociation was carried out on a shaking water bath at room temperature for 30-60 min or until the cells were easily dispersed by gentle up and down pipetting with a pasteur pipette. An IgG¹ antibody of irrelevant specificity was used as control to monitor nonspecific binding. The cells were analyzed using a flow cytometer (Ortho 50H, Ortho Diagnostics, Westward, MA). The amount of PCNA was reported as percentage of green fluorescent cells.
Cell loss is defined as the rate of loss of cells as a fraction of the rate at which cells are being added to the tumor volume by cell proliferation. Cell loss is an important factor in estimating the growth potential of a tumor 12. In order to facilitate the study of this phenomenon in our experimental model, we proceeded as follows.
The human breast carcinomas, maintained in athymic nude mice, were measured weekly with a Vernier caliper. The weekly increase in volume (cm3) was determined for each carcinoma. After the mice had been fed diet for 6 weeks, the carcinoma-bearing athymic nude mice were injected p. with 5 µCi of I125-iodo 2' -deoxyuridine (I125Urd, 6 mCi/mg, Sigma Chemical Co., St. Louis, MO). In order to prevent excess concentration of I125 in the thyroid, each mouse was given 0.1% KI in the drinking water commencing 3 days prior to I125Urd administration. Twenty-four hours after I125Urd injection, mice were lightly anesthetized with ether and secured in a holding apparatusto allow for gamma emission readings. Emissions were read using a Geiger counter with a NaI crystal, 2 inch diameter and 0.04 inches thick, Model leg-1, low energy gamma probe, 61% efficiency, Eberline Inc., Santa Fe, NM. Care was taken to place the probe in an identical position on top of the carcinoma in contact with the integument overlying the outer surface. Duplicate 1 minute emission readings (cpm) were recorded for each carcinoma for seven consecutive days, subtracting background emissions. Mean rate of I125Urd loss from each tumor was calculated as follows:
y=mx+b y=natural log of the daily mean I125Urd emissions (from duplicate measurements) (cpm)-background emissions (cpm)
m=slope (Kl, rate constant)
b-y intercept=activity at time zero Using the above equation, a graph was generated for each carcinoma as follows:
lncpm(l vs time (days), ln=natural log
cpm(l-7)=counts per minute (emissions) from day 1 to day 7
cpm(0)=counts per minute (emissions) at day 0
The resulting slope, or rate constant KL, was utilized to compute carcinoma cell loss factor using the following formulas:
Ø = (Ti/2+Td)Begg’s formula 13
Ø =cell loss factor=cell loss rate expressed as percent of the cell birth rate
TD=tumor doubling time in days (calculated by determining the number of days for tumor to double in size)
Ti/2=time (days) for I125Urd emission from the tumor to reach 1/2 of initial (time 0) emission rate
Ø =1-Tp Steel's Formula12
Ø =cell loss factor=cell loss rate expressed as percent of the cell birth rate Td=tumor doubling time in days (calculated by determining the number of days for tumor to double in size)
Principles for assessing cell loss from growing tumors in situ using these formulas have been validated by Kallman et al 14.
In Table 3 after mice were fed corn oil (CO) and fish oil (FO) diets for a period of only one week (Study 1), the difference in mean tumor volumes did not reach a level of 5% significance. Also no significant difference between mice fed CO and FO diets was obtained in DNA synthesis parameters (H3-thymidine autoradiograph analysis and Brdu flow cytometry analyses). This trend was also observed in tumors of animals fed CO and FO diets for a period of two weeks (Study 2), in which mean tumor volumes and mean tumor DNA synthesis parameters were not significantly different. The animals fed the CO diet for four weeks (Study 3) had a significantly larger (p<0.05) tumor volume than those fed a FO diet; nevertheless no significant difference was detected in mean H3-thymidine autoradiograph analysis. In another study in which animals were fed diets for 4 weeks (Study 4) we also observed a significantly larger tumor volume (p<0.05) in the CO fed animals compared to those fed FO. However, when FO was supplemented with excess antioxidants, mean tumor volume of animals fed the supplemented FO was comparable to the mean tumor volume of the CO fed animals. Again no significant difference was observed in tumor DNA synthesis parameters. In addition, animals fed FO supplemented with iron, mean tumor volume were significantly less compared to the other three experimental groups (CO, FO, FO+antioxidants) but no significant difference in mean tumor H3-thymidine analysis was detected. After feeding CO and FO diets for six weeks (Study 5), CO fed animals had a significantly larger (p<0.05) mean tumor volume than those fed FO. However, no significant difference in tumor DNA synthesis parameters was observed. After feeding diets for 10 weeks (Study 6), a significant difference in mean tumor volume was not reached, neither was a significant difference obtained in tumor DNA synthesis parameters.Table 3. Effect of dietary fat (corn oil and fish oil) on DNA synthesis (H -thymidine autoradiography and Brdu) of human breast carcinoma MDA-MB231 maintained in athymic nude mice.
|Diet||Number of tumors||Mean tumor volume (cm iS.E.)||Mean H3-thymidine autoradiographs (# labeled tumors cells/mm of tissueiS.E.)e||Brdu labeling- index (% tumor cells showing Brdu uptake 1S.E.)e|
|1 week on diet||(Study 1)|
|2 weeks on diet||(Study 2)|
|4 weeks on diet||(Study 3)|
|4 weeks on diet||plus diet .||supplementation (Study 4)|
|Fish oil +||17||1.48l0.29f||45.4614.63(17)||13.8811.25(17)|
|Fish oil +||10||0.2410.07h||43.8615.83(10)||n.d.|
|6 weeks on diet||(Study 5)|
|10 weeks on diet||(Study 6)|
In Table 4 (Study 7), after mice were fed a CO and FO diet for six weeks, animals fed a CO diet had a significantly larger (p<0.05) tumor volume compared to those fed FO but not compared to those fed the antioxidant supplemented FO diet. No significant difference in tumor DNA synthesis parameters (Brdu analysis and PCNA analysis) was detected between these three dietary groups.Table 4. Effect of dietary fat (corn oil and fish oil) on DNA synthesis (Brdu and PCNA) of human breast carcinoma MDA-MB231 maintained in athymic nude mice
|Diet 6 weeks on diet (Study 7)||Number tumors||of Mean tumor volume (cm 1 S.E.)||Brdu labeling- index (% tumor cells showing Brdu uptakeiS.E.)d||PCNA labeling- index (% tumor cells PCNA positive lS.E.)d|
|Corn oila||24||1.40±0.21e||4.0210.52 (24)||14.1710.85 (24)|
|Fish oilb||18||0.43±0.llf||2.6310.38 (18)||12.7610.80 (18)|
|Fish oil +||7||1.03±0.18e||2.1410.47 (7)||17.4712.86 (7)|
In Table 5 (Study 8), after feeding different ratios of CO and FO for a period of six weeks the animals fed 15% CO/5% FO had a significantly larger (p<0.05) tumor volume than the ones fed 10% C0/10% FO and 5% CO/15% FO. The animals fed 10% C0/10% FO had a higher tumor volume than those fed 5% C0/15% FO but this difference did not reach a level of 5% significance. Also no significant difference was detected in tumor mean Brdu analysis between these three groups.Table 5. Effect of different ratios of dietary fats (corn oil and fish oil) on DNA synthesis (Brdu) of human breast carcinoma MDA-MB231 maintained in athymic nude mice.1738630603123000
|Diet 6 weeks on diet (Study 8)||Number of tumors||Mean tumor volume (cm 1 S.E.)||Brdu labeling- index (% tumor cells showing Brdu uptakeiS.E.)b|
|Corn oil 15%/ fish oila 5%||20||1.24±0.23c||10.1111.55(16)|
|Corn oil 10%/ fish oila 10%||18||0.65±0.15d||7.5011.09(17)|
|Corn oil 5%/ fish oi1a 15%||17||0.34±0.08d||6.2111.30(9)|
In Table 6 (Study 9), after feeding tumor bearing mice a CO and FO diet for two weeks no significant difference was detected in mean tumor volume, mean rate of I125Urd loss from tumors nor mean tumor cell loss factors betweenthe two diet groups. In Study 10, after feeding a CO and FO diet for four weeks, no significant difference was detected in mean tumor volume, mean rate of I125Urd loss from tumors, nor mean tumor cell loss factors. In Study 11, after feeding a CO and FO diet for six weeks the CO fed animals had a significantly larger (p<0.05) mean tumor volume compared to the FO fed animals; the tumor volume in the CO fed animals, however, was not significantly different from that observed in the antioxidant supplemented FO fed animals. Mean rate of I125Urd loss from tumors was significantly (p<0.05) lower in the CO fed animals compared to the FO fed animals and the antioxidant supplemented FO group. Mean tumor cell loss factor as determined by Steel 12 and Begg 13 was also significantly lower in the CO fed group compared to the FO fed group and to the antioxidant supplemented FO group. Figure 1 compares the mean slopes of the rate of I125Urd loss from the tumors of the three dietary groups (CO, FO+A and FO). A significant difference (p<0.05) in slopes between CO and FO and CO and FO+A was observed; the slopes of the FO and FO+A dietary groups were virtually identical.
Figure 1. Rate of I125Urd loss from human breast carcinomas in athymic nude mice fed corn oil (CO), fish oil (FO) and fish oil supplemented with antioxidants (FO+A) for a period of 6 weeks. Rates were 0.174±-0.01 (N=52), 0.224±0.01 (N=45) and 0.223±0.02 (N=14) for CO, FO and FO+A, respectively, p<0.05.
Diets high in polyunsaturated fatty acids (e.g.,corn oil),when fed to rodents, causes an increase in size and number of mammary tumors when compared to rodents fed low levels of the same fat or high levels of other types of fat (e.g., beef tallow and certain fish oils) 15. To examine the tumor growth kinetics of this phenomenon, only a few laboratories 4, 5, 6, 7, 8 have investigated the differential effects of dietary fat on tumor cell proliferation or on tumor cell loss (cytolysis). Abraham et al. (8) reported a significantly smaller tumor size of transplantable mouse mammary tumors in mice fed a high FO diet compared to tumors of mice fed a high CO diet. They accounted for this result by providing data of an increase cell loss in tumors of rodents fed a high FO diet compared to the cell loss obtained from those tumors of mice fed a high CO diet. Previously they reported 4, 7 no significant difference in tumor cell proliferation parameters when feeding diets high in unsaturated and saturated fatty acids to mammary tumor bearing rodents in spite of obtaining a significant difference in tumor size. When examining different levels of dietary fat, Oyaizu et al. 5 and Noguchi and colleagues 6, 9, 10 reported a smaller mammary tumor size in rats fed a low level of CO compared to rats fed a high level of CO; a decrease in carcinoma cell proliferation parameters in the tumors of rats fed the low level CO diet compared to those fed the high level of CO was observed. Clearly, more studies are required to have a more definitive understanding as to the effect of the type and amount of dietary fat on tumor cell proliferation dynamics in mammary tumors in order to have a better understanding of the nutritional influence on tumor growth processes.
Thus, an important question remains unanswered. Is the decreased size of tumors of animals fed high FO diets due to a decrease in DNA synthesis, or because of an increase in cell loss (cytolysis)? The H3-thymidine autoradiographic methodology is a precise means to assess DNA synthesis 16. Furthermore, this technique is an effective means of providing a quantitative differentiation between carcinoma cell and stromal cell proliferation processes 17. One drawback of the H3-thymidine autoradiographic methodology is that it is extremely time consuming. On the other hand, the flow cytometry method of measuring DNA synthesis provides a fast (labeling and detection can be performed the same day), sensitive and quantitative way to measure DNA synthesis in suspended cells. Brdu is an analog of thymidine that is concentrated only in cells in active DNA synthesis 18. Quantitation of Brdu concentration in DNA is made possible by the development of a monoclonal antibody against Brdu 19. The H3-thymidine autoradiographic technique and the Brdu flow cytometric technique, as methods of estimating cell proliferation have been reported to be in close agreement with each other 20. The PCNA flow cytometric technique has also been used to study cell proliferation processes. PCNA possesses a temporal specificity which makes it a suitable marker for cell proliferation. PCNA begins to accumulate during the G1 phase of the cell cycle, is most abundant during the S phase and declines during G2/M phase 21. PCNA has been successfully used to selectively identify proliferating cells in solid tumors 22.
The method for the assessment of cell loss from growing tumors was originally described by Steel 12 and validated by Begg 13 and Kallman et al. 14. Steel's cell loss factor measures the rate of loss of cells as a fraction of the rate at which cells are being added to the tumor volume by cell proliferation 23. This factor, therefore, expresses the growth potential of a tumor (ratio of cell loss rate to the cell birth rate). The extent to which processes of cell loss are competing with the process of cell proliferation can be obtained utilizing the formula:
Ø =1-Tp, where Ø =the cell loss factor.
A modified version of Steel's cell loss measurement concept was first utilized in a dietary study of tumor growth by Abraham and colleagues 18. The method of measuring cell loss from tumors in situ by using the I125 deoxyuridine (I125Urd) technique was first described by Begg 13. Begg's derived tumor kinetic parameters originated from Steel's formula for cell loss factor 11. Begg equates the I125Urd emission rate (KL, loss of radioactivity) to the rate of cell loss. The slope of the I125Urd emission rate in a semi logarithmic plot is defined as T1/2 (time to halve the radioactivity). The modified adaptation of Steel's formula is:
The pragmatic difference between these two formulas is that Begg's derived formula takes into account only one point (point of 1/2 radioactivity) on the curve generated by the I125Urd emission data, whereas Steel's formula takes into account the total curve generated by the I125 emission data. Therefore, by utilizing the whole curve, Steel's formula provides a more precise assessment of cell loss factor. In general, both formulas were in close agreement in these studies. These techniques to measure cell loss are very attractive since they provide a direct determination, and therefore, are superior to methods which depend solely on calculated and measured doubling times of tumor growth. Another advantage of these methods is that they require fewer animals than methods requiring the excision of tumors, since each animal contributes several time points. In addition, variation is reduced since each tumor acts as its own control. Moreover, it allows for cell loss rates of individual tumors to be determined. Currently, the I125Urd technique offers the only direct non-invasive method of assessing cell loss in individual tumors. Although errors of the in situ technique are smaller and less frequent than those occurring with other methods, the problems of reutilization of the isotope are still present. The isotope could also be trapped in necrotic areas inside or surrounding the tumor. For this reason, in our experiments, tumors with overt necrotic areas were not used. Since our cell lines do not elicit any substantial immune response, in athymic nude mice, we did not have the problem of additional necrosis induced by immune cell infiltration of the tumor tissue. Thus, our experimental model utilizing athymic nude mice bearing human breast carcinoma cell lines is suitable for determination of the cell loss factor by the I125Urd in situ technique. In addition to reporting Steel's and Begg's cell loss factor, we report the tumor I125Urd emission rate which, indirectly can be equated to cell loss rate.
No significant differences in DNA synthesis parameters between the diet groups (CO and FO) were observed (Studies 1-6, in Table 3, Study 7 in Table 4 and Study 8 in Table 5), despite the significant differences in tumor size that were observed in a number of these studies (Studies 3,4,5,7,8). The small numerical decrease in DNA synthesis parameters in tumors of FO fed mice (3-5% decrease) may have relevant biological significance. This very small decrease in DNA synthesis may prove of importance, if this difference is real and can be extended throughout the entire dietary feeding period.
In Study 4 (Table 3), two additional FO groups were added, an antioxidant supplemented FO group (FO+A), and an iron-supplemented FO group (FO+I). In this study, supplementation with antioxidants significantly enhanced tumor volume of the FO fed animals, while in contrast, supplementation with iron significantly decreased tumor volume of the FO fed animals. This appears to be due to differences in cytostatic/cytolytic lipid peroxidation product accumulation in the tumors as we reported previously 2, 3. Nevertheless, in spite of this tumor volume difference, no significant difference in cell proliferation parameters in the tumors was observed. Study 7 (Table 4) followed a similar trend as Study 4 in which a significant difference was observed in mean tumor volume between the CO and the FO fed groups, with the antioxidant supplemented FO having a comparable tumor volume to the CO fed group; no significant difference in cell proliferation parameters was once again observed. In Study 8 (Table 5), different ratios of CO and FO were fed which resulted in an inverse relationship in which a decreasing tumor volume was evident as the FO content of the diet increased; once again no significant difference in cell proliferation parameters was observed. These results are in accordance with those of Abraham et al. 4, 7, 8 in which the difference in mammary tumor size between CO and FO fed animals cannot be accounted for by the fraction of tumor cells that were actively proliferating.
Our results suggest that parameters other than cell proliferation may be the primary mechanism by which differences in tumor volume between dietary CO and FO fed animals is achieved. Studies 9, 10, 11 (Table 6) furnish a possible answer to the mechanistic question of how dietary FO affects tumor growth-related kinetic parameters by providing Also in Study 11, no difference in tumor cell loss parameters was obtained between FO+A and the FO group. The reason for this is not known. It is conceivable that the FO used in Study 11 could have been substantially oxidized prior to diet preparation. The already excessively oxidized FO could prevent any substantial antioxidative effect by the addition of antioxidants. This would result in similar tumor volumes and cell loss parameters in animals fed FO and FO+A diets. These results, albeit preliminary, suggest that differences in mammary tumor cell loss parameters in CO and FO fed animals are very important, perhaps more important than tumor cell proliferation, in determining the extent of volume of these tumors. More studies are needed to confirm these preliminary results in order to provide a conclusive unifying concept to explain how dietary fat affects tumor growth. Nevertheless, these experiments and those reported earlier 2, 3 support the concept that FO suppresses human breast carcinoma growth in athymic nude mice by increasing the concentration of secondary products of lipid peroxidation in the tumor; such products (cytostatic/cytolytic) significantly increase tumor cell loss.