Immunologic heterogeneity of tumor cell subpopulations from a single mouse mammary tumor.

Five subpopulations (66, 67, 68H, 168, and 4.10LM) obtained from a single BALB/cfC3H mammary adenocarcinoma were used to assess intratumor immunologic heterogeneity. BALB/c and BALB/cfC3H mice were immunized with each of the subpopulations, and lymph node cells (LNC) from immunized animals were tested for cell-mediated immunity (CMI) to each subpopulation in vitro by chromium release and microcytotoxicity tests and in vivo by Winn assays. The immunogenic character of the subpopulations differed markedly. The pattern of cross-reactivity indicated that at least two determinants were involved, one of which was probably a viral antigen. The viral antigen was expressed on 4 subpopulations (66, 68H, 168, and 4.10LM). The other determinant was immunogenic in both BALB/c and BALB/cfC3H mice and was expressed on 4 subpopulations (66, 67, 168, and 4.10lm). Thus 1 subpopulation (68H) expressed only the viral antigen, 1 (67) expressed only the other antigen, and 3 (66, 168, and 4.10LM) expressed both. Expression of the determinants showed qualitative and quantitative variations. Quantitative differences were noted by the relative effectiveness of the subpopulations to induce CMI and by the relative sensitivities to LNC-mediated killing. Qualitative differences were indicated by the occurrence of unidirectional cross-reactivities between some pairs of subpopulations; a determinant could be expressed so that the subpopulation could induce cytotoxic cells but not be sensitive to them or vice versa.

Activity of lymphoid cells separated from mammary tumors in blastogenesis and Winn assays.

Lymphoid cells isolated from mouse mammary tumors by isokinetic gradients were not stimulated in vitro by either phytohemagglutinin or a soluble, tumor-associated antigen extract even though splenocytes from the tumor-bearing mice were responsive to both. In in vivo Winn assays, lymphoid cells isolated from tumors markedly stimulated tumor growth rate. The effect on growth rate was abrogated by exposure of the isolated lymphoid cells to antilymphocyte serum and complement.

Induction of antitumor immunity by interleukin-2 gene-transduced mouse mammary tumor cells versus transduced mammary stromal fibroblasts.

BACKGROUND: Tumor cell-targeted cytokine gene transfer has been used to generate tumor cell vaccines, but this approach is limited by the need to establish and implant live tumor cells. PURPOSE: The purpose of this study was to determine if stromal fibroblasts could be used as an alternative vehicle for delivery of the cytokine interleukin-2 (IL-2) into the tumor microenvironment. We attempted to establish the feasibility of (a) genetic immunotherapy in a mammary tumor system and (b) engineering stromal fibroblasts as well as tumor cells. We compared the effects of tumor cell-mediated and stromal fibroblast-mediated local IL-2 expression on the generation of antitumor immune responses. METHODS: Retroviral vectors containing a human IL-2 gene were used to transduce a mouse mammary tumor line, 4TO7, and an immortalized but nontumorigenic fibroblast line established from syngeneic mammary fatpads. Expression of the IL-2 gene in transduced cells was determined by measuring IL-2 secretion, by RNA-polymerase chain reaction, and by immunochemistry. Groups of 5-12 BALB/c mice were injected with either 4TO7 cells or various doses of IL-2-secreting 4TO7 cells (4TO7-IL-2); tumor growth was monitored. To test whether local IL-2 expression by transduced cells could influence the growth of unmodified tumor cells, we determined tumor development in groups of mice treated with 4TO7 cells co-injected with either 4TO7-IL-2 cells or IL-2-secreting fibroblasts. RESULTS: 4TO7-IL-2 cells induced active immunity able to reject the immunizing tumor and to resist challenge with parental 4TO7 cells on the contralateral side. Mice pretreated with 4TO7-IL-2 were significantly protected compared with untreated control animals or mice pretreated with irradiated 4TO7 cells. The immunity induced by 4TO7-IL-2 cells did not protect against challenge with another subline, 4T1, which was derived from the same spontaneously arising mammary tumor as 4TO7. Co-injection of 4TO7 cells with 4TO7-IL-2 cells reduced tumorigenicity, whereas co-injection of 4TO7 cells with IL-2 secreting fibroblasts did not. CONCLUSION: Our results suggest that induction of anti-tumor immune response by local IL-2 production is most effective when the helper cytokine is secreted by the tumor cell. IMPLICATION: Our studies caution against the use of IL-2 gene-transduced syngeneic stromal cells as an alternative strategy of gene therapy for cancer. However, they may allow study of the mechanisms of tumor antigen recognition and the possible involvement of co-stimulatory signals for effective tumor vaccination by gene-modified cells.

Xenograft model of progressive human proliferative breast disease.

BACKGROUND: Progression of proliferative breast disease has been associated with increased risk for development of invasive carcinoma. Cell lines have been developed to facilitate the study of this process. Human cell line MCF10A originated from spontaneous immortalization of breast epithelial cells obtained from a patient with fibrocystic disease, and cell lines MCF10AneoN and MCF10AneoT were created by stable transfection of these cells with the neomycin-resistance gene and either the HRAS gene or the mutated T-24 HRAS gene, respectively. PURPOSE: Our goal was to develop an experimental model of progressive human proliferative breast disease. METHODS: MCF10A, MCF10AneoN, and MCF10AneoT cells were injected subcutaneously into the dorsal flank of male nude/beige (C57/BALB/c nu/nu bg/bg) mice (12 mice for each cell type). These mice were examined periodically for formation and persistence or growth of palpable nodules. One mouse per group was killed 1 week after cell injection; thereafter, mice were observed as long as possible. Cells were recovered from palpable lesions by enzymatic dissociation of the excised lesions. Cells re-established in tissue culture from a week-14 tumor (MCF10AneoT.TG1) were injected into 12 male nude/beige mice. Southern blot hybridization analysis of the HRAS gene locus and cytogenetic analyses were performed. RESULTS: Transplanted MCF10A and MCF10AneoN cells formed transient, small palpable nodules that regressed and disappeared during the 4th and 5th weeks. In 10 of the 12 mice, T-24 HRAS gene-transfected MCF10A cells (MCF10AneoT) formed small, flat nodules that persisted for at least 1 year. Three of these xenografts became carcinomas. One (removed 7 weeks after transplantation) was an undifferentiated carcinoma composed of polygonal cells with large, vesicular nuclei and numerous mitoses. The second (removed after 14 weeks) was an invasive squamous cell carcinoma. The third (removed after 56 weeks) was a moderately differentiated adenocarcinoma. Initially, xenografts of MCF10AneoT.TG1 cells showed intraductal proliferative changes; after 23 weeks, the lesions showed histologic features resembling those seen in atypical hyperplasia of the human breast, and later lesions showed characteristics of carcinoma in situ. The MCF10 lineage of cells of three MCF10AneoT.TG1 xenografts was confirmed by DNA fingerprinting and karyotype analysis. CONCLUSIONS: MCF10AneoT and MCF10AneoT.TG1 comprise a transplantable xenograft model that produces a broad spectrum of human proliferative breast disease. IMPLICATIONS: The reproducible establishment of representative stages in early breast cancer progression from the MCF10 model offers a new opportunity to analyze critical events of carcinogenesis and progression in breast cancer.

Epithelial component of host-tumor interactions in the orthotopic site preference of a mouse mammary tumor.

Mouse mammary tumors grow preferentially upon transplantation into intact mammary glands compared to cleared mammary fat-pads. Both sites provide stroma of the orthotopic site, but the latter lacks epithelial elements. If epithelium from enzymatically dissociated normal mammary glands is added to the tumor cells prior to injection into cleared fat-pads, tumor growth is comparable to that seen in intact mammary fat-pads. The growth-enhancing effects of normal mammary cells are not duplicated by normal kidney or liver cells. These results demonstrate that epithelial-epithelial interactions, as well as stromal-epithelial interactions, are associated with the enhanced growth of mammary tumor cells transplanted into orthotopic sites. The results also suggest that enhancement of tumor growth does not require intact tissue architecture.

Selective events in the metastatic process defined by analysis of the sequential dissemination of subpopulations of a mouse mammary tumor.

To identify selective steps in metastasis, those that eliminate nonmetastatic tumor cells more efficiently than metastatic cells, we have evaluated the sequential dissemination of tumor cells from a mammary fatpad, using both metastatic (4T1 and 66cl4) and nonmetastatic (67NR, 168FARN, and 4TO7) subpopulations of a single mouse mammary tumor. Each of these variant subpopulations is resistant to one or more selective drugs so they could be quantitatively identified by colony formation in selective media. We found that the 2 metastatic cell lines metastasized by different routes and that the nonmetastatic tumor cell lines failed at different points in dissemination. Line 67NR did not leave the primary site; clonogenic tumor cells were not detected in the nodes, blood, or lungs during the experiment (7 weeks). Tumor line 168FARN disseminated from the primary tumor because clonogenic cells were cultured from the draining lymph nodes throughout the experiment. However, dissemination essentially stopped in the node as cells were rarely isolated from blood, lungs, or lives. Whether 168FARN cells failed to reach these tissues or were killed very rapidly after traversing the lymph node is unknown. Line 4TO7 cells disseminated via the blood and were consistently recovered from lungs by day 19 but failed to proliferate. This panel of 5 subpopulations thus identifies different points of selective failure in tumor cell dissemination and should be valuable in the assessment of antimetastatic therapies.

Production of a more aggressive tumor cell variant by spontaneous fusion of two mouse tumor subpopulations.

A hybrid cell formed by the spontaneous fusion of two sister subpopulations was isolated and was found to combine clinically aggressive features of both parental cells. Subpopulation 168FAR is highly tumorigenic but does not metastasize spontaneously from a s.c. site. Subpopulation 44FTO is a variant selected for resistance to 6-thioguanine and to ouabain. It is poorly tumorigenic but spontaneously metastasizes to the lungs and livers of mice in which primary tumors do form. The two parental subpopulations are differentially sensitive to chemotherapeutic drugs; 168FAR is relatively more resistant to methotrexate, whereas 44FTO is more resistant to 5-fluorouracil. Both are about equally sensitive to melphalan. The hybrid 168FAR x 44FTO clone was isolated by growth in hypoxanthine-aminopterin-thymidine medium which also contained 3 mM ouabain; neither 168FAR nor 44FTO survive in this medium. The hybrid nature of the isolated clone was confirmed by cytogenetic analysis and determination of DNA content. The hybrid clone was highly tumorigenic, spontaneously metastasized from the subcutis, was nearly as resistant to 5-fluorouracil as the most resistant parental line, and was more resistant to both melphalan and methotrexate than either parental subpopulation. These results illustrate the role that cell fusion could play in progression by allowing the rapid assimilation of aggressive phenotypes from distinct coexisting subpopulations.

Factors affecting growth and drug sensitivity of mouse mammary tumor lines in collagen gel cultures.

A series of mouse mammary tumor subpopulation lines were compared for growth properties and sensitivity to chemotherapeutic drugs when grown as boluses in a collagen gel matrix versus in monolayer culture. Although the cell lines exhibited characteristic rates of bolus expansion in collagen, this growth was not paralleled by an exponential increase in cell number with time. Cell boluses contained a higher proportion of cells in G0-G1 phases of the cell cycle than did the same cell lines in monolayer cultures. Histological examination revealed areas of necrosis in boluses. Thus cells growing in collagen cultures resembled cells growing as solid tumors and cells from other three-dimensional culture systems. The growth of cell boluses in collagen gel cultures was reduced nonexponentially by melphalan, methotrexate, and 5-fluorouracil in contrast to the exponential decrease in growth measured in cloning assays. The lowest concentration to which cells first responded to drug was in general similar for collagen gel assays and for cloning assays. The rank order of sensitivity of different cell lines in the two assays was identical for methotrexate (four cell lines), similar for melphalan (four of five lines), but quite different for 5-fluorouracil. In contrast to cloning assays cell boluses continued to grow, albeit at a reduced rate, in the presence of high drug concentrations. This was not due to either diminished drug availability in collagen gel or drug penetration into the bolus.

Therapeutic perturbation of the tumor ecosystem in reconstructed heterogeneous mouse mammary tumors.

We have measured the response to methotrexate in vivo of paired mixtures of sister subpopulation lines from a mouse mammary tumor, as a model of drug response of a heterogeneous tumor. The subpopulation lines differed in intrinsic sensitivity to methotrexate. Response was measured both as growth delay and as a shift in tumor cell population distribution toward the more resistant cell line. We found differences between two pairs of cell lines in growth delay: line 66 plus 4T07 mixtures tended to be as responsive as was line 4T07 (the more sensitive line) alone, whereas line 168 plus 4T07 mixtures tended to be less responsive than line 4T07 alone. With both paired mixtures, the tumors arising after treatment tended to contain more line 66 or line 168 than did untreated tumors, but this shift was extremely variable among individual tumors. Within most treatment groups, there was no correlation between the growth rate of individual mixed tumors and the final tumor cell distribution. Likewise, between experiments, there was no correlation between the amount of growth delay in mixed tumors and the final tumor cell distribution. Thus, the cellular composition of treated tumors did not directly reflect the response to therapy.

Growth regulation of mouse mammary tumor cells in collagen gel cultures by diffusible factors produced by normal mammary gland epithelium and stromal fibroblasts.

We have reported previously that both normal mammary epithelium and stroma stimulate the growth of mouse mammary tumors in vivo. We have devised a method to investigate the role of diffusible factors in these growth interactions in vitro. Because an appropriate matrix, a particular cell shape, or multicellular organization may be required for the production of factors and/or for the response to such factors, the method assesses the expansion of boluses of cells in collagen gel matrix. Under these culture conditions, paracrine effects were detected which were not observed when cells were growing in monolayer on plastic. The growth of mammary tumor lines 66, 410.4, and D2A1 was stimulated by both mammary epithelial cells and fibroblasts prepared from midpregnant mouse mammary glands. In contrast, these mammary tumor lines were inhibited by coculturing with normal mammary cells in monolayer on plastic. Stimulation in collagen and inhibition in monolayer cultures were both dose dependent; increasing numbers of regulator mammary cells increased the effects on tumor growth. Additional tumor cells did not stimulate growth of target tumor cell boluses in collagen gel cultures. This attempt to model the in situ situation may be a more appropriate model for the detection and characterization of relevant diffusible growth regulatory factors than monolayers on plastic and/or colonies growing in agar.

Preferential growth of mammary tumors in intact mammary fatpads.

Four transplantable tumors, three (66, 410, and 168cl) isolated from a spontaneously occurring strain BALB/cfC3H mammary tumor and one (D2) arising from a BALB/c hyperplastic alveolar nodule were found to grow better in mammary fatpads than at s.c. sites. Furthermore, tumor growth was better (p less than 0.05) in intact mammary glands than in cleared mammary fatpads for the D2, 410, and 66 tumors (168cl was not tested). The role of immunity in these differences was investigated using the highly immunogenic 410 tumors. Tumor 410 induced equally effective immunity to subsequent challenge whether it was implanted s.c. or in intact fatpads. Furthermore, in immunized animals, Tumor 410 was rejected equally well when the challenge site was intact fatpad as when s.c. Similarly, Tumor 410 induced immunity after implantation into cleared fatpads and, in immunized animals, was rejected when the challenge site was the cleared fatpad. We thus found no evidence that the mammary fatpad is immunologically privileged, as compared to the s.c. site, with respect to tumor transplantation antigens.

Quantitative selectivity of contact-mediated intercellular communication in a metastatic mouse mammary tumor line.

We have examined contact-mediated intercellular communication by measuring the transfer of thioguanine sensitivity to a hypoxanthine phosphoribosyltransferase (EC 2.4.2.8)-negative clone (66cl-4) selected from one subline isolated previously from a spontaneously arising mammary tumor of a BALB/cfC3H mouse. We tested other sublines from the same tumor and unrelated cell types for their ability to serve as 6-thioguanine nucleotide donors to 66cl-4 cells. The degree of communication, measured by the number of donor cells required to reduce the number of thioguanine-resistant colonies, varied with the donor cell type. The 66cl-4 line communicated with the parent cell line from which the thioguanine-resistant cell was selected and with other sublines from the parent tumor, with some unrelated tumor cells, and with some nonneoplastic cells (3T3, hamster kidney and lung fibroblasts, and mouse mammary epithelial cells). There was a quantitative difference in the amount of communication which took place with the various cells tested, but no pattern of difference could be discerned. Line 66cl-4 did not preferentially communicate with cells of epithelial versus fibroblast morphology, nor with tumor versus nontumor cells. The 66cl-4 cells retained the ability of their parent line to form metastatic tumors when injected s.c. into BALB/c mice. A quantitative selectivity of communication is thus expressed in these malignant metastatic cells, but it is apparently unrelated to either the morphological or malignant phenotype of the donor. Contact-mediated communication between tumor subpopulations may differentially affect growth and drug sensitivity within a tumor.

Tumor suppression by p21WAF1.

The p21WAF1 gene encodes a cyclin-dependent kinase inhibitor and mediates tumor suppressor gene p53-induced cell cycle arrest. To directly test whether p21WAF1 can act as a tumor suppressor, we have expressed the p21WAF1 cDNA in several human tumor cell lines using a tetracycline-inducible system. Overexpression of p21WAF1 suppresses proliferation and soft agar growth of tumor cells in vitro, as well as tumorigenicity in vivo. Our data provide direct evidence for the tumor-suppressive activity of p21WAF1.

Dominance of a tumor subpopulation line in mixed heterogeneous mouse mammary tumors.

When mixtures of cell lines 168 and 4T07, both derived from the same mouse mammary tumor, were injected into syngeneic mice, the resulting tumors, analyzed over a large size range by colony-forming assays in selective media, consisted primarily of line 4T07, even when the ratio injected was 100:1 or greater in favor of line 168. This result indicated a suppression of growth of line 168, since the volume-doubling time of line 168 tumors in the absence of line 4T07 was one-half that of line 4T07 tumors. That growth suppression was not due to inhibition of line 168 by immunity induced to line 4T07 was shown in two ways: (a) line 168 tumors grew almost as well in mice preimmunized with line 4T07 as in controls, whereas line 4T07 tumor growth was strongly inhibited in preimmunized mice; and (b) the final composition (favoring line 4T07) in mixed tumors was similar in tumors grown in mice immunosuppressed by irradiation to that in nonirradiated controls. The strong suppression of line 168 did not occur when the two cell lines were injected simultaneously at different s.c. sites, nor did it occur when line 168 cells were injected in mixtures with lethally irradiated line 4T07 cells. Line 4T07 cells also suppressed the growth of line 168 cells in monolayer cultures. It was not likely that suppression was due to competition for growth factors, since the effect required cell contact. Suppression probably was not mediated through junctional communication, since these cells do not engage in metabolic cooperation. We suggest that a growth-inhibitory factor produced by line 4T07 mediates the suppression of 168 cells.

Growth interaction in vivo between tumor subpopulations derived from a single mouse mammary tumor.

Our laboratory has previously isolated several tumor cell populations from a single, spontaneously arising mammary tumor of a BALB/cfC3H mouse and established them in tissue culture as independent sublines. These subpopulations differ according to many criteria including growth parameters and expression of tumor-associated antigens. We have tested the interaction in vivo of several of these subpopulations by injecting cell suspensions of the same or different sublines into opposite flanks of BALB/cfC3H or BALB/c mice. The growth characteristics of certain subpopulations were altered by the presence of a different subpopulation on the opposite side. In order to understand the mechanism of interaction, we chose two subpopulations (410 and 168) for further study. In BALB/cfC3H mice, the presence of line 410 tumors on one flank inhibited both 410 and 168 tumors on the other flank. Line 168 tumors did not inhibit either 410 or 168 tumors. The inhibitory effect of line 410 appeared to be immunological, since (a) it was increased by injecting line 410 several weeks before line 168, (b) it was abrogated in mice subjected to 400-rad X-irradiation 2 days prior to tumor cell injection, (c) mice could be made resistant to both line 410 and line 168 tumors by implantation followed by surgical removal of line 410 but not of line 168, and (d) resistance could be adaptively transferred with lymph node cells from line 410-sensitized mice in Winn assays. Thus, immunity to tumor-associated antigens may be one way by which cells of a heterogeneous tumor can interact.

Metabolic cooperation between mouse mammary tumor subpopulations in three-dimensional collagen gel cultures.

A series of subpopulation lines derived from a single mouse mammary tumor were tested for their ability to interact with each other in metabolic cooperation assays in three-dimensional collagen gel cultures. Inhibition of growth in the presence of selective drug (6-thioguanine or 2-fluoroadenine) of two 6-thioguanine-resistant cell lines, 66cl4 and 44FTO, as well as a 2-fluoroadenine-resistant line, 168FAR, occurred in the presence of other sensitive cells, demonstrating the transfer of drug sensitivity to drug-resistant cells. Transfer of drug resistance was shown by using selective medium containing hypoxanthine, aminopterin, and thymidine plus ouabain in which mutual metabolic cooperation between line 66cl4 (also ouabain resistant) and "wild-type" lines is required for growth. Effective communication leading to the transfer of drug resistance and ultimate survival of both cell types occurred in mixtures containing as little as 10% of either cell type. The requirement for mutual metabolic cooperation in hypoxanthine, aminopterin, and thymidine plus ouabain medium prevented line 66 from overgrowing line 66cl4 to the extent that it did in control medium. The ranked abilities of lines 66, 410.4, and 168 to promote growth when mixed with line 66cl4 in collagen cultures in hypoxanthine, aminopterin, and thymidine plus ouabain medium were correlated with the ranked cell densities required for these same cell lines to inhibit growth of line 66cl4 by metabolic cooperation in the presence of 6-thioguanine in monolayer cultures. These ranked abilities may be related to the effectiveness with which each of these cell lines is able to form communicating junctions with line 66cl4.

Heterogeneity in drug sensitivity among tumor cell subpopulations of a single mammary tumor.

Three distinct subpopulations of tumor cells derived from a single parent strain BALB/cfC3H mammary adenocarcinoma were tested in vivo for sensitivity to cyclophosphamide, methotrexate, and 5-fluorouracil. Treatment was begun either 2 days after s.c. tumor cell injection or at the time when the tumors became palpable. It was given on a weekly basis for 4 weeks. The mice were observed for growth of the primary implant and for development of spontaneous metastases. The three subpopulations differed markedly in their sensitivity to the drugs. The effects of the drugs ranged from induction of regression of the "primary" to enhancement of metastases. The effect on primary growth was independent of that on metastasis. The effect of the time of administration of the drugs also varied among the subpopulations. The sublines were also tested in vitro with methotrexate and 5-fluorouracil. Again there were marked differences in sensitivity to inhibition of cell division by the drugs. The relative sensitivities in vitro did not correlate with observations in vivo. The existence of subpopulations of tumor cells, differing in sensitivity to therapeutic agents, within a single neoplasm, presents a challenge to development of assays capable of predicting drug response and to the selection of combination therapies.

Expression of epithelial-like markers and class I major histocompatibility antigens by a murine carcinoma growing in the mammary gland and in metastases: orthotopic site effects.

A spontaneous murine mammary carcinoma, designated SP1, grew more aggressively in the mammary gland than in the subcutis exhibiting a 10-fold lower 50% lethal tumor dose and the ability to metastasize spontaneously from the orthotopic mammary gland site. The appearance of metastasis could be abrogated by resection of the primary tumor up to 21 days postinjection, arguing against the possibility that metastasis occurred due to trauma of the injection and/or healing processes. In addition, tumor cells recovered from lung metastases exhibited an increased ability to metastasize when reinjected into either the s.c. or mammary sites. Tumor cells from lung metastases showed low levels of Class I major histocompatibility (MHC) antigens, like the parental SP1 cells, but were found to express differentiation markers typical of normal basal and luminal mammary epithelium. SP1 tumors expressed increased Class I MHC antigens, as well as high levels of basal and luminal breast epithelial markers, within 7 days of implantation into the mammary gland. On the other hand, SP1 tumors growing in the subcutis never expressed increased Class I MHC levels and expressed the epithelial marker antigens at lower levels and not until at least 21 days of growth. Removal of host epithelium by cauterization of the mammary bud at 3 weeks had no effect on the increased growth, metastasis and acquired heterogeneity of MHC and epithelial associated antigens, suggesting that the mammary gland stroma was responsible for the observed phenomenon. These findings suggest that the mammary gland either selects distinct tumor subpopulations, or induces a phenotypic change leading to tumor progression and the generation of metastatic subpopulations.

The cellular basis of tumor progression.

Variability in disease presentation and course is a hallmark of cancer. Variability is seen among similarly diagnosed cancers in different patients or animal hosts and in the same cancer at different periods of time. This latter type of variability, termed "tumor progression," was defined by Foulds in a series of six rules that describe the independent behavior of individual cancers and the independent evolution of different cancer characteristics. Tumor progression is believed to result from variability among subpopulations of tumor cells within individual cancers and from selection of these subpopulations by conditions within the cancer environment, such that different subpopulations come to prominence over the course of cancer development and growth. Interactions among subpopulations, however, modulate tumor behavior as well as tumor evolution. The leading hypothesis for the origin of tumor subpopulations is the genetic instability of cancer cells. There are a number of possible mechanisms of genetic instability, some internal to cancer cells (mutation, amplification, mutator phenotypes, DNA repair deficiencies) and some present in the tumor microenvironment (endogenous mutagens). There are also potential epigenetic mechanisms of variability, including alterations in gene regulation, differentiation, adaptation, and cell fusion. Regardless of mechanism, the heterogeneity of tumor subpopulations poses a number of challenges to the practice of cancer research, including the design of reproducible and meaningful experiments. Tumor heterogeneity also has significant consequences for the clinical assessment of tumor prognosis and the development of effective treatment regimens.

Genetic Regulation of Development in Sorghum bicolor (VIII. Shoot Growth, Tillering, Flowering, Gibberellin Biosynthesis, and Phytochrome Levels Are Differentially Affected by Dosage of the ma3R Allele.

Sorghum [Sorghum bicolor (L.) Moench] homozygous for ma3R lacks a type II, light-stable phytochrome of 123 kD and has a number of phenotypic characteristics consistent with the absence of functional phytochrome B. We have used plants heterozygous at Ma3 (Ma3/ma3R and ma3/ma3R) to determine the effect of dosage of ma3R on plant growth, flowering, gibberellin (GA) levels, and content of the 123-kD phytochrome. Both Ma3/ma3R and ma3/ma3R produced the same number of tillers per plant as their respective homozygous non-ma3R parents. Height of the heterozygotes was intermediate between the homozygous parents, although it was more similar to the non-ma3R genotypes. In both field and growth-chamber environments, the timing of floral initiation and anthesis in the heterozygotes also was intermediate, again more similar to non-ma3R plants. In Ma3/ma3R, levels of GA53, GA19, GA20, and GA1 were almost exactly intermediate between levels detected in Ma3/Ma3 and ma3R/ma3R plants. Immunoblot analysis indicated that there was less of the 123-kD phytochrome in Ma3/ma3R than in homozygous Ma3, whereas none was detected in ma3R/ma3R. The degree of dominance of Ma3 and ma3 over ma3R varies with phenotypic trait, indicating that mechanisms of activity of the 123-kD phytochrome vary among the biochemical processes involved in each phenotypic character. Although the heterozygotes were similar to homozygous Ma3 and ma3 plants in growth and flowering behavior, Ma3/ma3R contained 50% less of the bioactive GA (GA1) than non-ma3R genotypes. Thus, sensitivity to endogenous GAs also may be regulated by the 123-kD phytochrome. To fully regulate plant growth and development, two copies of Ma3 or ma3 are required to produce sufficient quantities of the light-stable, 123-kD phytochrome.