Modulation of Tumor Angiogenesis by Stem Cell Factor¹

Angiogenesis, the growth of new blood vessels from a preexisting microvascular bed (1), is of crucial importance for the growth, maintenance, and metastasis of solid tumors(2, 3, 4). Therefore, a more complete understanding of the factors contributing to tumor angiogenesis is of paramount importance. Tumor angiogenesis depends on the interaction of different tumor components, e.g., tumor parenchymal cells, endothelial cells, infiltrating cells from the bloodstream, and perhaps mast cells. There is much evidence to suggest a link between mast cells and angiogenesis. For example, mast cells are distributed along blood vessels (5), giving them a perfect position to modulate vessel growth. Mast cell degranulation alone is sufficient to induce neovascularization in rat mesentery (6) and in the chick chorioallantoic membrane (7). Mast cells also accumulate within and around solid tumors (8). When tumor cells are injected into a chick embryo, there is a 40-fold increase in mast cell density around the tumor implantation site compared with normal tissue(9). Injection of mast cell suspensions into animals leads to acceleration of tumor growth (10), whereas decreasing the number of tissue mast cells leads to depression of tumor growth(11). Inhibiting mast cell degranulation with disodium cromoglycate also significantly depresses tumor growth (10, 12). In capillary hemangiomas, which are common benign vascular tumors that inflict young children, the mast cell concentration in the tumor is at least 5-fold higher than in normal tissue(13). When the hemangioma starts to shrink as the child gets older, a decrease in mast cell number precedes tumor shrinkage(13). On the basis of this circumstantial evidence, it has been suggested that mast cells in tumors modulate the neovascularization process. However, the role of mast cells in tumor angiogenesis has not been studied thoroughly.

Tumor-associated mast cells are often found to have degranulated and to have released their chemical mediators, especially in the late stages of tumor proliferation (14). Many components of mast cells are angiogenic or can modulate the angiogenesis process (15, 16). These components include basic fibroblast growth factor,vascular endothelial growth factor, heparin, heparinase, histamine,tumor necrosis factor-α, and various proteases. Therefore, mast cell degranulation may modulate angiogenesis.

The activities of mast cells are largely controlled by SCF,⁴a mast cell growth factor. For example, SCF is a chemoattractant for mast cells (17) and repeated injection of SCF into the skin of mice results in the appearance of large numbers of mast cells at the injection site (18, 19). SCF can drive the proliferation of mast cells as well as promote mast cell maturation in vitro (20, 21). Finally, SCF induces mediator release from mouse mast cells in vitro(22) and can trigger mast cell activation and a mast cell-dependent inflammatory response in vivo(23).

SCF is a product of the steel gene in mice and has two transmembrane isoforms, SCF-1 and SCF-2 (24, 25). SCF-1,encoded by full-length mRNA, is a 248-amino acid protein that can be hydrolyzed by proteases, resulting in a soluble form of SCF. SCF-2,derived from alternatively spliced SCF mRNA, gives rise to a smaller 220-amino acid protein that lacks the same proteolytic cleavage site. It is ineffectively cleaved at an alternative site and therefore remains almost exclusively as a cell membrane protein.

Several types of tumor cells exhibit an increased production of SCF(26) in addition to other growth factors. However, the direct effect of SCF on tumor angiogenesis has not been examined. We hypothesized that large amounts of SCF released from tumor cells may account for the increased number of mast cells in tumors and may lead to an accelerated angiogenic response. In this report, we provide direct evidence that SCF expressed by mammary tumor cells modulates tumor angiogenesis by regulating mast cell activity.

Ethylnitrosourea-induced mammary tumor cells were generated in Berlin Druckrey IV rats (27). This animal model for human breast cancer is characterized by a short latency period for tumor development, ovarian hormone dependency, high incidence of malignant tumors, and widespread metastases (27, 28, 29). In addition,these mammary tumors are well vascularized, and the supporting fibrovascular stroma is infiltrated with large numbers of mast cells. Numerous cell lines and clones with different tumorigenic and metastatic potentials have been isolated from these tumors(30). One cell line, designated Brc, was grown in a 37°C incubator with 10% CO₂ in complete medium consisting of DMEM with 10% fetal bovine serum, 2 mmglutamine, 1 mm sodium pyruvate, 100 units/ml penicillin,100 μg/ml streptomycin, and 0.25 μg/ml amphotericin B. Cells were given fresh medium every 3 days and subcultured once a week. For Western blot, immunoprecipitation, and growth curve studies, cells were switched to serum-free medium containing 6.25 ng/ml of sodium selenite,6.25 μg/ml of transferrin, 535 μg/ml of BSA-linoleic acid, 1.6μg/ml of putrescine, 10 μg/ml of high-density lipoprotein, 1 mg/ml BSA, 15 mm HEPES, 1 mm sodium pyruvate, 2 mm l-glutamine, 100 units/ml penicillin, 100μg/ml streptomycin, and 0.25 μg/ml amphotericin B.

The mammalian expression vector pCEP4 (Invitrogen, San Diego, CA) was used in all studies. The rat SCF cDNA (770 bp, −184 to +586 bp) was cloned into pCEP4 in the sense direction. The design of antisense oligomers was based on the sequence of the leading strand of SCF cDNA. The 37-bp sequence downstream of the initiation codon but including the first ATG codon was chosen as a target site. To ensure that the oligomers would be inserted into the pCEP4 vector in an antisense orientation, specific cohesive ends were engineered for the two oligomers. The oligomers were synthesized by Sigma-Genosys (The Woodlands, TX). The sequences of these oligomers were the following:5′-TCGAG[ATG]AAGAAGACACAAACTTTGGATTATCACTTGCATTA-3′and 5′-AGCTTAATGCAAGTGATAATCCAAAGTTTGTGTCTTCTTCATC-3′. Underlined bases indicate the restriction enzyme sites, and the ATG in brackets is the initiation codon in the open reading frame for SCF. The two oligomers were annealed and inserted into the multiple cloning site of the pCEP4 plasmid. Both sense and antisense constructs were checked for fidelity by restriction enzyme mapping.

PanVera TransIT polyamine transfection reagent (PanVera Corp., Madison,WI) was used to transfect the Brc cell line with either the control vector (C-P, no insert) or the pCEP4 vector containing SCF cDNA fragments in either the sense (S-P) or the antisense (AS-P) direction. Approximately 24 h prior to transfection, 1 × 10⁵ Brc cells in complete medium were plated into 35-mm dishes so that the cells were 40–70% confluent the next day. Before transfection, 18 μl of TransIT-LTI transfection reagent were added dropwise into tubes containing 200 μl of Opti-MEM I reduced serum medium (Life Technologies, Grand Island, NY) and incubated at room temperature for 5 min. Three μg of control plasmid (C-P),plasmid with SCF cDNA in the sense orientation (S-P), or plasmid with SCF oligomers in the antisense orientation (AS-P) were then added to designated tubes, mixed gently, and incubated at room temperature for another 5 min. The medium was removed from the dishes, cells were washed once with PBS, and 2 ml of fresh Opti-MEM were added. The cells were then treated with the TransIT reagent/DNA complex mixture for 6 h. After 6 h, medium containing the TransIT/DNA complex mixture was removed and replaced with growth medium. The cells were incubated for 24 h after transfection and then subcultured at a 1:5 ratio in 60-mm dishes. Hygromycin B (Life Technologies), a selection agent, was added at its optimal concentration (predetermined to be 300 μg/ml for the Brc cell line). This concentration of hygromycin was maintained until discrete colonies appeared (usually 10–14 days).

All of the transfected colonies were carefully screened for SCF expression by Northern blot, Western blot, and immunoprecipitation. Some colonies were further characterized for growth rate, VEGF, and bFGF expression. One colony from each transfection group was used to induce tumors for the in vivo studies.

Transfected cell lines were incubated in serum-free medium for 72 h. The cells were washed extensively, and whole-cell lysates were prepared using 1 ml of lysis buffer [10 mm Tris (pH 8.0),1 mm EDTA, 0.1% SDS, 1% deoxycholate, 1% NP40, 0.14 m NaCl, 0.5 mg/ml Pefabloc, 1 μg/ml leupeptin, 1 μg/ml aprotinin, and 1 μg/ml pepstatin]. Total protein was determined by bicinchoninic acid assay (Pierce, Rockford, IL). Ten μg of protein were loaded onto 9–16% gradient acrylamide gels, subjected to electrophoresis, and transferred to nitrocellulose. The blot was blocked with nonfat milk and reacted with 1 μg/ml rabbit-antimouse SCF antibody (Genzyme, Cambridge, MA). After washing, the blot was incubated for 1 h with a horseradish peroxidase-conjugated goat-antirabbit IgG (Jackson ImmunoResearch Laboratories, West Grove,PA) at a 1:100,000 dilution. The blot was washed again and exposed to Super Signal West Dura chemiluminescence reagent (Pierce) for 5 min and exposed to Biomax ML film (Eastman Kodak, Rochester, NY).

About 4 ml of conditioned medium was collected from culture dishes of C-P, S-P, or AS-P transfected tumor cells, normalized for cell number,and filtered through a 0.2 μm filter to remove any cell debris. The medium was subsequently incubated with 1 μg/ml of rabbit-antimouse SCF antibody (Genzyme) for 2 h at room temperature. A 100-μl volume of 50/50 (vol/vol) protein A-agarose (Amersham-Pharmacia,Piscataway, NJ) in TBS was added, and the tubes were mixed gently for 2 h. The agarose beads were washed three times with TBS. Immune complexes were then eluted by addition of 2× sample buffer [0.126 m Tris (pH 6.8), 12.6% glycerol, 10% β-mercaptoethanol,0.004% bromphenol blue, and 5% SDS] and boiling for 5 min. The relative SCF amounts in the supernatant were determined by SDS-PAGE,Western blotting, and densitometry.

Female Berlin Druckrey IV rats, 40–50 days of age, were randomly allocated into three groups. A suspension (1 × 10⁶ cells/0.5 ml PBS) of transfected cells(S-P-Brc and AS-P-Brc) or transfection controls cells (C-P-Brc) was injected into the mammary fat pads of corresponding animals. After 2 weeks, the tumors reached a diameter of about 1–3 cm. The animals were anesthetized, and the tumors were excised. The connective tissue surrounding the tumor mass was carefully and thoroughly dissected away,and the tumor weight was measured.

Half of each tumor was put into liquid nitrogen for protein extraction. Frozen tumor tissue was finely ground using a mortar and pestle. The resulting powder was solubilized in lysis buffer, and protein concentration was determined. To evaluate the expression of SCF by the tumors, 10 μg of protein from each tumor were loaded onto 9–16%gradient gels and subjected to Western blot analysis (as described above for transfected cells). After the incubation of the blots with chemiluminescence reagent and exposure to Biomax film, the films were analyzed by densitometry.

The expression of vWF by tumors was used as an index of vascular density. Because of the large molecular weight of the vWF monomer, 10μg of protein from each tumor were separated on 8–20% NOVEX NuPAGE Tris-acetate gels (Novex, San Diego, CA) and subjected to Western blot analysis. Rabbit-antihuman vWF antibody (1:300 in blocking solution;Dako, Santa Barbara, CA) was used as the primary antibody, and horseradish peroxidase-conjugated donkey-antirabbit IgG was used as the secondary antibody (1:100,000 dilution in blocking solution; Jackson ImmunoResearch Laboratories). Protein expression was analyzed by densitometry.

Half of each tumor was fixed in 10% buffered formalin for at least 24 h, progressively dehydrated in a graded series of ethanol,cleared in Histoclear, embedded in paraffin, sectioned at 5-μm thickness, and placed on poly-l-lysine-coated slides. To detect mast cells, tumor sections were stained with acidic toluidine blue (Sigma Chemical Co., St. Louis, MO) for 5 min. Toluidine blue is dissolved in 60% ethanol to a final concentration of 0.4 mg/ml and acidified with hydrochloric acid (pH 2.0). Toluidine blue binds to sulfated glycosaminoglycans in mast cell granules (31, 32), staining them purple while the tumor tissue stains blue. The number of toluidine blue-positive mast cells was counted using a 1-mm square counting grid and recorded as mast cell number/mm² of tumor surface area. The entire tumor section surface area was measured.

Deparaffinized tissue sections were exposed to 0.1% trypsin in 0.1%CaCl₂ in PBS at a temperature of 37°C for 10 min. To block endogenous peroxidase activity, the tissue sections were subsequently incubated in 0.3%H₂O₂ for 30 min at room temperature, followed by incubation with a 1:500 dilution of rabbit-antihuman vWF antibody (Dako) at room temperature for 30 min in a moist chamber. After washing in PBS, the sections were treated further using a commercial kit (Vectastain Elite ABC; Vector Laboratories, Burlingame, CA), which included a biotinylated goat-antirabbit IgG secondary antibody and peroxidase-labeled avidin. Peroxidase substrate solution (Vector NovaRED substrate) was applied to each section to produce a brick red colorimetric signal, and tissue sections were counterstained with methyl green (Vector Laboratories). Areas of the tumor containing the most capillaries and small venules(i.e., areas of most intense neovascularization), so-called hot spots, were identified (33). Three hot spots for each tumor were counted, and each count was expressed as the highest number of microvessels identified within any ×400 field. The numbers of microvessels were then averaged.

Results are expressed as mean ± SE unless otherwise indicated. Statistical significance was determined by one-way ANOVA, followed by the Student-Newman-Keuls multiple comparison method. P < 0.05 was used to indicate statistically significant differences.

More than 10 clones from each transfection were carefully prescreened for SCF expression. Fig. 1,A shows representative data from nontransfected Brc cells as well as transfected C-P-Brc, S-P-Brc, and AS-P-Brc clones analyzed by Western blotting. Two proteins, M_r33,000 and M_r 30,000,corresponding to the SCF-1 and SCF-2 isoforms, respectively, were detected in cells. Transfection of control plasmid (original pCEP4) did not affect SCF expression in control cells (Fig. 1,A, compare Lanes 1 and 2). On the other hand, S-P Brc cells(Fig. 1,A, Lane 3) expressed more SCF-1 and SCF-2 than C-P Brc cells. In AS-P Brc cells (Fig. 1 A, Lane 4), only SCF-2 was detectable, and the amount was lower than that of C-P-Brc. A similar pattern for SCF mRNA was observed by Northern blot analysis (data not shown).

To further evaluate the effects of transfection, the production of soluble SCF was also examined. Conditioned media were collected and subjected to immunoprecipitation, SDS-PAGE, and Western blotting. Both the soluble form of SCF (M_r 23,000,derived primarily from SCF-1 by proteolysis) and the membrane-bound SCF-2 (M_r 30,000) were found in media. Fig. 1 B shows that S-P Brc cells (Lane 3)released more soluble SCF into the medium, whereas AS-P Brc cells (Lane 4) released less soluble SCF into the medium,compared with C-P Brc cells (Lane 2) or nontransfected cells(Lane 1). The M_r 30,000 SCF found in the medium was presumably attributable to cell death/lysis with release of membrane-associated proteins into the medium.

After induction of tumors by transfected cells, 35 tumor samples were collected (11 C-P Brc, 12 S-P Brc, and 12 AS-P Brc tumors). Both SCF-1 and SCF-2 were found in control and sense transfected rat mammary tumors, with SCF-1 being predominant (data not shown). On the other hand, antisense SCF cDNA-transfected Brc cells generated tumors expressing mainly SCF 1. Fig. 2 A shows densitometric analysis of SCF-1 expression on Western blots. Expression of SCF-1 in tumors was elevated by sense transfection of the cells and decreased by antisense transfection of the cells.

Tumor sections were stained with toluidine blue to identify mast cells. The number of mast cells in each tumor section was counted and reported as mast cell number/mm². The densities of mast cells in C-P, S-P, and AS-P tumors were 0.43 ± 0.04,0.73 ± 0.07, and 0.30 ± 0.08/mm², respectively (Fig. 2 B).

Polyclonal vWF antibody was used to stain microvessels in tumor sections. No cross-reactivity with other cell types was noted. Blood vessels stained brick red or dark brown, and the background stained blue. The mean microvessel counts, per ×400 field, were 20.8 ± 3, 29.9 ± 4, and 16.4 ± 3, for control, sense, and antisense transfected tumors (Fig. 2,C),respectively. A similar pattern was observed with vWF Western blotting. As shown in Fig. 3,A, tumors derived from S-P Brc cells (Lanes 1 and 2) expressed more vWF than those derived form C-P Brc cells(Lanes 3 and 4). On the other hand, antisense tumors (Lanes 5 and 6) expressed less vWF compared with control. Fig. 3 B shows the densitometric analysis of all of the vWF Western blots. The vWF expressed by sense SCF cDNA-transfected tumors was 96% higher than that of controls,whereas antisense SCF cDNA-transfected tumors expressed only 55% of vWF found in control tumors.

As shown in Fig. 4, tumors derived from AS-P Brc cells were significantly smaller than control tumors. Sense-transfected tumors, on the other hand, were not statistically different from the control group.

To rule out the possibility of a change in tumor cell proliferation rate attributable to sense and antisense SCF gene transfection, the proliferative rates of C-P, S-P, and AS-P Brc cells were determined. No significant difference in proliferation rate was found among the three cell lines when grown in serum-free or complete medium containing 300 μg/ml hygromycin B (data not shown).

We also examined the expression of VEGF and bFGF by transfected cells. As determined by Western blot, the expression of VEGF by C-P, S-P, and AS-P transfected cell lines was similar (data not shown). We were unable to find bFGF in any cell line using Western blotting or immunoprecipitation techniques (data not shown).

Breast cancer is one of the most frequent causes of death among women (34). The growth of solid tumors in the breast depends upon the induction of new blood vessel growth into the tumor. There is mounting evidence that this neovascularization, or angiogenesis, also plays a relevant role in the biological aggressiveness of breast cancer. Several studies have shown a worse prognosis for those patients exhibiting tumors with high angiogenic activity (33, 35, 36). Unfortunately, the cellular mechanisms controlling tumor angiogenesis are poorly understood. Although endothelial cells and pericytes are key players, many other cells may also be important in angiogenesis. A more complete understanding of the factors contributing to intercellular cross-talk in tumors may provide clues for preventing tumor angiogenesis.

Coussens et al. (37) demonstrated a role for mast cells in initiation of premalignant neovascularization in a transgenic mouse model of epithelial carcinogenesis. Infiltration of mast cells accompanied activation of angiogenesis in this model. Using mast cell-deficient mice, these investigators demonstrated an attenuation of neoplasia with a quiescent vasculature similar to normal tissue. This finding agrees with earlier studies showing decreased tumor angiogenesis in mast cell-deficient mice (38) or in rats treated with inhibitors of mast cell degranulation(39).

Despite this knowledge of a connection between mast cells and tumor angiogenesis, very little information is available to explain how mast cell activity is controlled by tumors. SCF, a mast cell growth factor,is released by tumor cells and represents a suitable target for studying the cross-talk between mast cells and parenchymal cells in tumor angiogenesis. At the same time, gene transfection techniques provide a powerful tool for modulating cell behavior. In this pilot study, the ethylnitrosourea-induced rat mammary tumor cell line, Brc,was used as a biological model to directly evaluate the relationship between SCF and tumor angiogenesis. Brc cells were stably transfected with a rat sense SCF cDNA or with a rat antisense SCF cDNA fragment; a vector-only transfected cell line was used as a control. The rationale for using sense and antisense SCF cDNA-transfected cells is that this approach will allow for the selection of tumor cell clones with defined, consistent SCF expression.

Transfected tumor cells were inoculated into rat mammary fat pads, and the role of SCF in mammary tumor angiogenesis was directly studied using this in vivo approach. We provide evidence to show that SCF is a key regulator of the mast cell population within a rat mammary tumor. As shown in Fig. 5 A, the antisense SCF-transfected tumors expressed a lower level of SCF (17% of the SCF in transfection control tumors), and this resulted in a mean mast cell density of 0.31/mm²compared with 0.43/mm² in the control tumors (SCF expression in controls normalized to 100%). This was a 28% decrease in mast cell density. On the other hand, sense SCF-transfected tumors had a 37% increase in SCF expression, resulting in a 69% increase in mean mast cell density (0.73/mm²) compared with control. As the primary regulating factor for mast cell growth and function, SCF is known to induce mast cell maturation, migration,proliferation, and degranulation (40, 41). It also enhances mature mast cell survival by inhibiting cell loss by apoptosis(42, 43). Our data show that SCF is an important regulator of rat mammary tumor mast cell density. By expressing SCF, the tumor itself may establish a microenvironment capable of regulating mast cell presence and activity.

Mast cells have been implicated in the angiogenic process because of the temporal relationship between their appearance in the tumor and the in-growth of vessels as well as the observation that tumor angiogenesis is retarded in mast cell-deficient animals (37, 38). As discussed earlier, many components of mast granules are potent angiogenesis mediators. Mast cell degranulation may result in the release of these angiogenesis factors into the interstitium and stimulation of the angiogenesis process. Furthermore, under the influence of SCF, mast cells not only release angiogenic factors from a preformed pool but also up-regulate expression of some potent angiogenic factors. For example, treatment of mouse tissue mast cells or mast cell lines with SCF results in an increase in the expression of bFGF at the mRNA level (44). Boesiger et al.(45) showed that, in response to SCF, mouse or human mast cells can rapidly release vascular permeability factor/VEGF by degranulation and can then sustain release by secreting newly synthesized protein. Recent evidence suggests that the mast cell itself is a cellular source of SCF (46), as well as being a target cell for this growth factor. Thus, SCF may regulate mast cell growth and function via both paracrine and autocrine mechanisms. These findings further extend the repertoire of cytokines synthesized by mast cells and widen their biological potential.

One aim of this study was to explain the relationship between mast cell density and angiogenesis of tumors in the context of changing SCF levels. As revealed by our in vivo study, the mast cell density of solid tumors directly correlated with tumor angiogenesis(Fig. 5 B). As determined by immunohistochemistry, the average microvascular densities of control, sense, and antisense transfected tumors were 20.8 ± 1.8, 29.9 ± 2.3, and 16.4 ± 1.9 vessels/high power field,respectively. Thus, sense transfection resulted in a 43% increase and antisense transfection resulted in a 21% decrease in mean tumor microvascular density, respectively. In a histochemical study, the changes in mean mast cell density were 69% and −28% for corresponding samples. The mast cells were found to accumulate around blood vessels (data not shown), in agreement with other published studies (5, 8, 9, 37).

Despite the availability of information on the actions of SCF at the molecular level and the regulation of mast cell activity by SCF, there have been no studies showing that SCF modulates angiogenesis. We constructed mammalian expression plasmids carrying either sense or antisense SCF cDNA fragments and transfected these plasmids into rat mammary tumor cells to modulate their expression of SCF. We found that an 83% decrease in SCF expression resulted in a 21% decrease in microvascular density, and a 37% increase in SCF expression resulted in a 43% increase in mean microvascular density (shown in Fig. 5 C). This is the first study to show that SCF expression and tumor angiogenesis have a “cause and effect” relationship. It provides direct evidence that this cytokine, expressed by tumor cells,has profound effects on tumor angiogenesis via modulation of mast cell activities and provides a novel experimental model for further investigation.

Importantly, we found that antisense transfection reduced tumor mast cell density and inhibited tumor angiogenesis so that tumor growth was limited. However, although sense transfection successfully increased tumor mast cell density and tumor angiogenesis, tumor growth did not change proportionally. The reason for this absence of increased tumor growth in response to elevated tumor angiogenesis is not clear. We harvested tumors 2 weeks after tumor cell inoculation to generate tumors of a sufficient size for all of the histochemical and biochemical analyses to be performed. We speculate that sense tumors may have had a growth advantage earlier. By 2 weeks, the control tumors were able to “catch up” in size, whereas the antisense tumors continued to lag behind, creating significantly smaller tumors. In addition, microscopic evaluation of tumors indicated more severe necrosis in the larger tumors (i.e., control and sense tumors). Therefore, it may be that increased growth of the sense tumors is offset by an increase in central necrosis, thus limiting the overall growth rate of the sense tumors.

In conclusion, transfection of the rat mammary tumor cells with sense or antisense SCF cDNA successfully up- or down-regulated SCF expression, respectively. The sense-transfected tumor cell clones produced tumors that were infiltrated with more mast cells and were more intensively vascularized. The stimulus that promoted angiogenesis in these tumors was provided by mast cells attracted to the tumor by SCF. SCF stimulates degranulation of mast cells, releasing several mediators that can amplify the angiogenic process. On the other hand,mast cell density in the antisense-transfected tumors was lower,consistent with the absence of an effective stimulator for mast cell migration, maturation, proliferation, and activation. Presumably, as a result of decreased mast cell number and degranulation, tumor angiogenesis was inhibited, as illustrated by decreased vascular density and lower total vWF expression. Finally, antisense-transfected tumors were significantly smaller than those induced by control or sense-transfected cells. These findings indicate that SCF should be considered as a possible target for antiangiogenic therapy of malignant breast tumors.