Inoculated mammary carcinoma-associated fibroblasts: contribution to hormone independent tumor growth

Two-month-old virgin female BALB/c mice (Animal Facility, Instituto de Biología y Medicina Experimental) were housed in groups of four per cage. Transgenic mice expressing enhanced GFP under the direction of the human ubiquitin C promoter were purchased from the Jackson Laboratories, Bar Harbor, Maine. These mice express Green Fluorescent Protein (GFP) in all tissues examined. The animals were in an air-conditioned room at 20 ± 2°C with a 12-hour light/dark cycle and had access to food and water ad libitum. Animal care and manipulation were in agreement with institutional guidelines and the Guide for the Care and Use of Laboratory Animals [14]. Protocols are approved by the Institutional Bioethical Committee.

C4-HD, a transplantable ductal mammary tumor, was originally induced by the continuous administration of MPA to a BALB/c female mouse [9, 15] and was subsequently maintained by serial subcutaneous (sc) transplantations into syngeneic MPA-treated female mice. This HD tumor expresses high levels of estrogen (ER) and progesterone receptors (PR) [16] and shows a nearly diploid karyotype [17]. The C4-HI variant arose from the C4-HD tumors when it started to grow in the absence of MPA. C4-HI cells also exhibit high ER and PR levels but have acquired an aneuploid karyotype [17, 18].

Primary cultures were prepared from C4-HI and C4-HD tumors as described previously [16, 19]. Briefly, tumors were aseptically removed, minced, washed with DMEM/F12 medium (Dulbecco's modified Eagle's medium: Ham's F12, 1:1, without phenol red, Sigma Chem. Co.), suspended in 5 ml of enzymatic solution [2.5 mg/ml trypsin (Life Technologies Inc.), 5 mg/ml albumin (Life Technologies Inc.), and 850 U/ml collagenase type II (Life Technologies Inc.) in phosphate buffered saline], and incubated at 37°C for 20 minutes with continuous stirring. The liquid phase of the suspension was then removed, and the undigested tissue was incubated for an additional 20 minutes with fresh enzymatic solution. Enzyme action was interrupted by adding a washing medium with 10% fetal calf serum (FCS, BioSer, Buenos Aires, Argentina). Epithelial (EPI) and fibroblastic cells (CAF) were isolated by resuspending the digested material in 20 ml of medium plus 10% FCS and allowing it to settle for 20 minutes. The upper 7 ml, containing single cells, were seeded into plates and the medium was changed after two hours. This time period is enough for the CAF to settle and not enough for the EPI to attach to the plastic plates. The fraction corresponding to the lower 5 ml, containing the sedimented cells, was resuspended in another 20 ml of washing medium (plus 5% FCS) and allowed to settle for 20 minutes. The upper 15 ml were discarded and the procedure was repeated 10 times, or until no single cells were observed in the upper fraction. Instead of plating the purified EPI, they were directly used for in vivo experiments. CAF, on the other hand, were obtained from primary cultures that were grown for one week prior to inoculation.

Purified EPI-HI, CAF-HI or EPI-HI+CAF-HI were seeded onto chamber slides. Slides were fixed with 100% ethanol for 30 minutes at -20°C, washed and treated with 0.25% Triton X-100 for 20 minutes. Slides were rinsed and blocked with 5% FCS for one hour at room temperature. Incubation with cytokeratin (CK; polyclonal rabbit antibody Z0622, 1:250 dilution, DakoCytomation, Carpinteria, CA) or smooth muscle actin (SMA; monoclonal mouse antibody Sigma Aldrich) was performed overnight at 4°C. After rinsing in PBS, slides were incubated with fluorescein isothiocyanate (FITC)-conjugated anti-rabbit (for CK) or anti-mouse (for SMA) IgG at 1:100 dilution (Vector Laboratories, Burlingame, CA). Nuclei were counterstained with propidium iodide (PI). The signal was viewed in a Nikon Eclipse E800 confocal microscope.

In the first set of experiments, untreated female BALB/c mice were sc inoculated with 1.5 × 10⁴ EPI-HD or EPI-HI alone or together with the same number of CAF-HI (n = 6 mice/group). Isolated CAF-HI were also inoculated to rule out the possibility that contaminating EPI-HI cells could account for the observed effects. In the second set of experiments, 1.5 × 10⁴ EPI-HI or EPI-HI+CAF-HI were sc inoculated in untreated female BALB/c mice and the animals were euthanized at 2, 7, 12 and 17 days post-inocula (n = 3/group). These cell injections were mixed with trypan blue prior to inoculation to localize the area of inoculation. After removal, isografts were immediately fixed in 10% formalin, embedded in paraffin and prepared for histological evaluation.

BALB/c mice (n = 3-10/group) were sc inoculated with saline, CAF-HI (1.5 × 10⁴), EPI-HI (1.5 × 10⁴), or EPI-HI+CAF-HI (1.5 × 10⁴ + 1.5 × 10⁴). Trypan blue was used to visualize the site of injection. The animals were euthanized at 7 and 12 days post inocula; tumors were then fixed with 10% formalin and embedded in paraffin.

Frozen tissue sections of the tumors were fixed in a cold solution of methanol:acetic acid (3:1) for one hour. The tissue sections were denatured in 70% formamide/2× standard saline-citrate. Mouse paint biotinylated DNA probes (Cambio, Cambridge, UK) for chromosomes X and Y were denatured and preannealed. The hybridization was performed overnight at 37°C. After hybridization, the slides were washed and the probes were detected with Avidin-TexasRed (Vector Laboratories, Burlingame, CA). Immunofluorescence using a CK antibody (polyclonal rabbit antibody Z0622, 1:250 dilution, DakoCytomation, Carpinteria, CA) was performed on the same slides. A secondary anti-rabbit-FITC coupled antibody was used. Nuclei were counterstained with DAPI. The signal was viewed in a Nikon Eclipse E800 confocal microscope.

Each BALB/c mouse was implanted intradermally on the ventral surface with the cells. Saline, CAF-HI (1.5 × 10⁴), EPI-HI (1.5 × 10⁴) or EPI-HI+CAF-HI (1.5 × 10⁴ + 1.5 × 10⁴) were injected in a volume of 50 μl together with one drop of 0.4% trypan blue to visualize the sites of injection. After 5 days, the mice were sacrificed, the skin carefully separated, and photographs captured under a Trinocular Zoom Stereomicroscope (Leica). The number of vessels was scored using a grid, subdivided into squares of 0.16 mm², and superimposed onto the photograph of the injection site.

Primary cultures from HD and HI tumors were generated, and the epithelial cells (EPI) and CAF were separated as described in Materials and Methods. We verified the purity of both cell populations by immunostaining. Isolated EPI strongly expressed CK whereas CAF expressed the fibroblastic marker SMA (Figure 1A) and vimentin (data not shown). This method proved to be very effective for the isolation of EPI. As confirmed by CK staining, a purity of 99 ± 0.5% could be attained. CAF enriched cultures were 93 ± 9.8% pure. To assess the contribution of CAF-HI to HD tumor growth, we inoculated these cell populations either alone or in combination. In other words, we mixed CAF-HI with EPI-HD or EPI-HI cells in a 1:1 ratio and inoculated these mixtures sc in BALB/c mice. We also inoculated EPI-HD, EPI-HI and CAF-HI alone. EPI-HI co-mixed with CAF-HI induced a significant increase in tumor size at the end of the experiment (Figure 1B), a decrease in tumor take (p < 0.05; EPI-HI vs. EPI-HI+CAF-HI) and an increase in tumor growth rate during the first period of tumor development (Slope EPI-HI: 5.2 ± 0.5 mm²/day; EPI-HI + CAF-HI: 8.3 ± 0.5 mm²/day p < 0.01). To ensure that contaminating EPI cells could not account for the enhancement of EPI-HI tumor growth, 2 × 10⁵ and 2.2 × 10⁵ cells (plus 10% EPI-HI) were inoculated. No differences were observed between either group (data not shown), indicating that the inoculated CAF-HI were responsible for promoting tumor growth. Moreover, no tumor growth was observed in EPI-HD co-inoculated with CAF-HI. These results raised two possibilities: CAF-HI do not provide sufficient stimuli to induce EPI-HD tumor growth, or inoculated CAF-HI do not remain in the tumor for initial signaling to induce EPI-HD tumor growth and thus, to bypass the hormone requirement.

To address the possibility that inoculated CAF-HI do not persist in the tumor and that the host stroma forms part of the tumor mass, we developed a murine isograft model that enabled us to distinguish the cells that come from the host from those that come from the donor. Thus, we isolated EPI-HI and CAF-HI from tumors growing in BALB/c mice and inoculated the neoplastic cells alone or admixed with CAF-HI in BALB/c-GFP mice. These mice express the GFP in all tissues examined (Jackson lab UBC-GFP). As shown in Figure 2A by fluorescent analysis of frozen sections, as early as 13 days after cell injection, there is already a strong invasion of host stroma within the tumor mass even in the admixed group. Primary cultures of these tumors and using an anti-smooth muscle actin antibody confirmed that all the SMA positive cells were positive for GFP too (Figure 2B), suggesting that even at early time points after cell inoculation the inoculated CAF-HI are undetected. These data were supported by additional assays in which neoplastic female cells were inoculated alone or admixed with CAF-HI in male BALB/c mice and the results were analyzed by FISH (data not shown).

We also performed the opposite experiment. We separated CAF-HI from a 28-day HI tumor growing in BALB/c-GFP mice. As expected, these CAF-HI expressed high levels of GFP (Figure 3). We then inoculated them with EPI-HI into BALB/c mice. Inoculated CAF-HI-GFP were not detected 13 days after transplantation (Figure 4A). However, the non-fluorescent tumor stroma was positively stained with SMA in immunofluorescence studies, demonstrating that the inoculated CAF-HI do not persist within the tumor and that neoplastic epithelial cells recruit new CAF from the host (Figure 4B). Furthermore, primary cultures of these tumors followed by immunofluorescence showed that 100% of SMA positive cells were negative for GFP (data not shown). In a different set of experiments, we transplanted EPI-HI into BALB/c-GFP mice and a small piece of this tumor was transplanted into BALB/c mice. Fluorescence analysis of frozen sections displayed similar results: 13 days after tumor transplantation, no GFP stromal cells were detected within the tumor mass (Figure 5A and 5B), and the non-fluorescent stroma was stained with SMA (Figure 5D). In addition, isolated CAF-HI, which were positive for SMA, were 100% negative for GFP (Figure 5C). These results indicate that even CAF within tumor transplants do not persist in the tumors after trocar transplantation and neoplastic cells recruit the stroma from the host.

We demonstrated that inoculated CAF-HI do not participate in HD tumor growth; however, they enhanced HI tumor growth (Figure 1B). Thus, our next goal was to investigate the growth-enhancing effects of CAF-HI on HI tumor growth during the first days of tumor formation. EPI-HI alone or mixed with CAF-HI were inoculated into female BALB/c mice. At 2, 7, 12 and 17 days after inoculation, the area of tumor inoculum was excised and histologically evaluated. At day 2 after cell injection, inflammatory infiltrates such as monocytes, lymphocytes and polymorphonuclear leukocytes (PMN) were present in both groups. There were no evident signs of vascularization and it was clear that the inoculated isografts had not yet developed into solid tumors (Figure 6). At day 7, significant differences were observed between both groups. The mixed isografts generated tumors with ductal and cribiform differentiation, which are the morphological features of these tumors, that were still absent in EPI-HI isografts (Figure 6). Several lymphocytes and polymorphonuclear cells were observed surrounding the tumors (Figure 6). Despite the clear differences on tumor histology observed at day 7, on day 12, the EPI-HI cells developed tumors that were very similar to the EPI-HI+CAF-HI carcinomas observed at day 7 (Figure 6). At subsequent time points, the tumors continued to grow, although the EPI-HI lesions were always smaller than the admixed tumors (p < 0.001). Even though EPI-HI isografts began growing later than EPI-HI+CAF-HI, the tumor growth rate was similar (Figure 1B). It is evident, that there are no histological differences in either tumor except for the lag in the development of the tumor structure. These results suggested that the inoculated CAF-HI influence the microenvironment during the first stages of tumor formation, contributing to early tumor development.

We previously demonstrated that FGF-2, a well known angiogenic factor [20], exerts a direct proliferative effect on epithelial HD and HI cells [13] and that CAF-HI express higher levels of FGF-2 than CAF-HD. To investigate if co-inoculated CAF-HI could influence angiogenesis during the first stages in tumor development, we performed intradermal inoculations of saline, CAF-HI, EPI-HI or EPI-HI+CAF-HI in BALB/c mice. Trypan blue stain was co-administered with the cells to localize the site of the inoculum. Mice were sacrificed 5 days after inoculation and tumor vessels were counted. A significant increase in tumor vessels (p < 0.05) was observed in co-inoculated animals compared to the other groups (Figure 7). Similarly, endothelial cells were positive for Von Willebrand factor expression in the vascular structures at the outside edge of EPI-HI+CAF-HI tumors, whereas no vessels were evident in EPI-HI isografts at this time point (data not shown). Collectively, these results indicate that inoculated CAF-HI induce angiogenesis at the site of injection contributing to rapid tumor formation.

To investigate whether the inoculated CAF-HI could be modulating the innate host's immune system, we evaluated at 5 days post-inoculum the number of PMN and mast cells using standard morphological criteria and toluidine blue staining, respectively. Both cell types were detected but we found no significant differences between both groups (Figure 8A). To rule out the role of the immune system we inoculated EPI-HI alone or commingled with CAF-HI into NOD/SCID mice. The tumor growth pattern in NOD/SCID mice was similar to the one obtained in BALB/c mice: CAF-HI significantly enhanced tumor growth throughout the experiment and the tumor weight at the end of the experiment was significantly higher in the admixed group (Figure 8B). Because GFP expressed by inoculated CAF-HI can potentially elicit an immune response, NOD/SCID mice were used as recipients. CAF-HI-GFP enhanced tumor growth even in NOD/SCID mice (data not shown). Collectively, these data indicate that the host's immune system is not responsible for the enhanced tumor growth induced by CAF.