Adenovirus-directed expression of Q227L-Gα_s inhibits growth of established tumors of later-stage human breast cancer cells in athymic mice

Abstract

Elevation of cAMP inhibits proliferation and expression of the transformed phenotype in several cell types. We studied the effects of elevation of cAMP and expression of mutant (Q227L) activated Gα_s on the proliferation and tumorigenic capability of the later stage, metastatic estrogen-independent human breast cancer cell lines MDA-231 and MDA-435. Our studies show that 8Br-cAMP inhibits proliferation of these cells in culture and their ability to form colonies in soft agar. This inhibition may occur by different mechanisms in the two cell types. In MDA-231 cells, cAMP elevation results in sustained expression of the cell cycle inhibitor p27kip1 and inhibition of CDK2 activity, whereas in MDA-435 cells inhibition of mitogen-activated protein (MAP) kinase 1,2 activity is observed. We tested whether these effects in culture could be translated into inhibition of tumor growth in vivo. Tumors were developed in athymic (Nu/Nu) mice by injections of MDA-231 or MDA-435 cells. Injection of Q227L-Gα_s expressing adenoviral vector into these established tumors inhibited further tumor growth under conditions where tumors injected with either saline or adenoviral vector containing β-galactosidase grew up to four to five times their original size. These results raise the possibility that sustained elevation of cAMP may have therapeutic value in the treatment of estrogen-resistant later stage breast cancers.

Complex biological processes such as proliferation result from a balance of interactions between signaling pathways. Unregulated cellular proliferation is a defining characteristic of cancer. In the case of cancer of mammary epithelia, progression from the estrogen sensitive to insensitive state and metastasis is accompanied by an increase in expression of several receptor tyrosine kinases, including the epidermal growth factor receptor, erb-B2 or erb-b4 and fibroblast growth factor receptor (1–4). Proliferative signals emanating from these tyrosine kinases are generally transmitted through the mitogen-activated protein kinase (MAPK) 1,2 pathway (5, 6). Signal transmission through the MAPK pathway is blocked by elevation of cellular cAMP levels and the consequent protein kinase A effects on c-Raf (7–9). Expression of activated Gα_s, which stimulates adenylyl cyclase, blocks Ras-stimulated MAPK activity and Ras-induced transformation of NIH 3T3 cells (4). Activated Gα_s, when expressed in the estrogen-dependent MCF-7 human breast cancer cells, inhibits their ability to form tumors in athymic mice (10). Elevation of cAMP also induces the cell cycle inhibitor p27kip1 (11). Decrease in p27kip1 levels is thought to be an important factor in breast tumor progression (12, 13). p27kip1 associates with active cyclin-dependent kinase (CDK) complexes to inhibit kinase activity and prevent cell cycle progression (14–16). Alternatively, the phosphorylation of p27kip1 by the active CDK-2/cyclin D complex leads to its degradation by proteasome (17, 18). Increases in cAMP lead to a decrease in CDK-2 activity, and levels of p27kip1 remain elevated. Thus, activated Gα_s may regulate proliferation at multiple sites, which may be advantageous in regulating aggressive proliferative programs in metastatic cells. The growth of these types of tumors cannot be inhibited by anti-estrogens, and generally only nonselective chemotherapeutic agents are therapeutically used. However, with our increasing understanding of how cell proliferation may be inhibited by interactions between signaling pathways, we wondered whether sustained elevation of cAMP in these tumor cells could inhibit tumor growth. Sustained elevation of cAMP may be obtained by expression of mutant (Q227L)-activated Gα_s (Gα*_s; refs. 4 and 10). Hence, we tested whether adenoviral vector-directed expression of Q227L-Gα*_s into solid tumors induced by either MDA-435 or MDA-231 cells inhibited tumor growth.

Methods

Cell Culture.

MB-MDA-231 cells were purchased from American Type Culture Collection (ATCC). MB-MDA-435 cells were a gift from J. Price (University of Texas, M.D. Anderson Laboratories, Houston, Texas). Low passage HEK 293 cells were purchased from Microbix (Toronto, ON, Canada). MDA-231, MDA-435, and HEK 293 cells were propagated at 37°C in an atmosphere containing 5% CO₂ in DMEM supplemented with 10% FCS containing 100 units/ml penicillin, 100 μg/ml streptomycin, and 2 mM l-glutamine (for HEK 293 cells).

Cell Proliferation Assay.

MDA-231 and MDA-435 cells were plated in six-well dishes (Costar) in triplicate at a density of 2.5 × 10⁴ cells per well. Cells were counted on day 1 to ensure accuracy of plating, then transduced by the adenoviral vectors ADV-βgal (adenoviral vector containing β-galactosidase) or ADV-Gα*_s [MDA-231, moi (multiplicity of infection) of 10; MDA-435, moi of 200] or treated with 1 mM 8Br-cAMP (Sigma). Media was replenished on day 3. Cells were counted on days 1, 2, 4, 5, 6, and 7.

Immunoblot Analysis.

For MAPK activity and for detection of B-Raf and c-Raf, cells were plated at a density of 10⁶ per 10-cm plate. Cells were serum starved 24 h before stimulation by 50 ng/ml epidermal growth factor (EGF; Upstate Biotechnology). Twelve hours before stimulation with EGF, 1 mM 8Br-cAMP was added to the plates. Control plates were not stimulated with EGF and did not have 8Br-cAMP added to the medium. Five minutes after stimulation, the plates were harvested, electrophoresed, and immunoblotted according to the manufacturer's protocol. For detection of phospho MAPK 1,2 and nonphosphorylated MAPK 1,2, antibodies were diluted 1:1000 (New England Biolabs). For Raf-B (C-19; Santa Cruz Biotechnology) and c-Raf (C-12; Santa Cruz Biotechnology), the antibodies were diluted at 1 μg/ml. Detection by chemiluminescence was done by using ECL (Amersham Pharmacia) and carried out as described by the manufacturer's protocol. For p27kip1 protein levels, cells were serum starved 24 h before stimulation with serum. Twelve hours before stimulation with serum, 1 mM 8Br-cAMP was added to the plates. Control plates did not have 8Br-cAMP added to the medium. Samples were collected at time 0, 4, 8, and 12 h after stimulation. Immunoblot analysis was performed p27kip1 antibody (1 μg/ml dilution; Upstate Biotechnology).

Soft Agar Assays.

Colony formation in soft agar was used as a measure of anchorage-independent growth, a defining characteristic of transformed cells. For this assay, bottom agar was laid in 60-mm/gridded dishes (Corning) and contained 0.5% Agar Noble (Difco), DMEM supplemented with 10% FCS containing 100 units/ml penicillin, 100 μg/ml streptomycin and 500 ng/ml Fungizone (GIBCO/BRL). Cells 10⁴ were seeded in DMEM supplemented with 10% FCS containing 100 units/ml penicillin, 100 μg/ml streptomycin, 500 ng/ml Fungizone, and 0.3% Agar Noble without or with 10 μM or 100 μM 8Br-cAMP. Plates were fed twice a week, and large colonies were counted after 6 weeks.

cAMP Accumulation Assays.

Cells were seeded in 12-well plates (Costar) in triplicate at a density of 5 × 10⁴ cells per well. The cells were labeled with [³H]adenine (2 μCi/ml; ICN) for 24 h. Cells were preincubated for 30 min with the phosphodiesterase inhibitor Ro-20174 (0.3 mM; Calbiochem), and accumulation of cAMP was measured. Accumulation of cAMP was measured as described in (9) and expressed as [³H]cAMP/([³H]ATP + [³H]cAMP) (×10⁻³). Cells were infected with ADV-Gα*_s or ADV-empty (Adenoviral vector) for 12 h before labeling.

Immune Complex CDK2 Kinase Assay.

Cells were plated into 10-cm dishes to reach a density of 1 × 10⁶ cells per dish the day of the assay. The cells were serum starved 24 h before the assay, and 100 μM 8Br-cAMP was added 12–16 h before the start. The cells were then stimulated with 10% FBS and collected at the following time points: 0, 4, 8, and 12 h after stimulation. The assay was carried out as previously described by Matsushime et al. (19).

Construction, Preparation and Titering of Adenovirus.

Adeno-β-Gal, Adeno-empty vector, and pSKAC were generous gifts from K. Peppel of the Lefkowitz Laboratory (Duke University, Durham, NC). The construction, preparation, and titering was carried out as previously described (20).

Optimal moi Determined by Transduction by Adeno-β-Galactosidase.

To determine an optimal moi for transduction by the adenoviral vectors, MDA-231 and MDA-435 were transduced by ADV-βgal and stained with a 5-bromo-4-chloro-3-indolyl β-d-galactoside (X-Gal) solution. MDA-231 and MDA-435 cells were transduced with dilutions of ADV-βgal ranging from 10⁵–10⁻⁴ plaque-forming units (pfu) to determine the optimal moi for transduction of the two cell lines. Thirty-six hours after transduction, the cells were washed twice with PBS and fixed in 0.25% glutaraldehyde for 15 min. The plates were then rinsed three times with PBS and incubated with 0.2% X-Gal solution for 1 h at 37°C. The X-Gal solution was then removed, and the plates were covered with 70% glycerol.

Tumorigenesis Experiments.

Female athymic mice were purchased from Taconic Laboratories at 2–3 weeks old and were placed in a pathogen-free environment. The mice were allowed to acclimate for a week before being used in studies. Exponentially growing MDA-435 cells or MDA-231 cells were injected s.c. into the Nu/Nu mice. At the time of injection, the cells were trypsinized, counted, and washed twice in PBS. MDA-231 cells were resuspended in PBS at 2.5 × 10⁷ cells per ml, and 5 × 10⁶ cells (200 μl) were injected per mouse. MDA-435 cells were resuspended at 5 × 10⁶ cells per ml, and 1 × 10⁶ cells (200 μl) were injected per mouse s.c. Mice were monitored for 2–4 weeks for tumor formation. When the tumor volume reached 100–200 mm³, the animals were randomly divided into groups and were injected with vehicle or vehicle containing adenoviral vector. For this procedure, 100 μl of PBS, or PBS containing ADV-Gα*_s or ADV-βgal (≈10⁹ pfu), was used. This procedure was done every week for 4 weeks. In the second type of experiments, when tumor volume reached 100–200 mm³, the animals were randomly divided and were injected with only a single dose of 100 μl of PBS, or PBS containing ADV-Gα*_s or ADV-βgal (≈10⁹ pfu) at the beginning of the observation period. Tumors size was determined by use of calipers. Growth of tumors was monitored weekly for 5 weeks.

Immunohistochemistry.

Tumors were removed 7 days after the last injection and placed in 10%/formalin/PBS. The solid tumors were embedded in paraffin and were cut into 5-μm sections. Sections were processed for immunocytochemistry with anti-βgal antibody (ICN). Vectastain Elite ABC Kit was used for detection according to the manufacturer's protocol (Vector Laboratories). Tissue sections were counterstained with hematoxylin and eosin.

All experiments were repeated at least three times, and qualitatively similar results were obtained. For animal experiments, the number of animals that was used is shown for each experiment.

Results

We selected two human breast cancer cell lines that represent some of the features found in tumors with advanced stages of the disease. Both MDA-MB-231 and MDA-MB-435 are cell lines that represent later estrogen-independent stages of metastatic breast cancer (21). MDA-MB-231 cells have high levels of EGF receptors and low levels of erb-B2, whereas MDA-MB-435 have been shown to overexpress erb-B2 (21–24). Both receptors have been previously characterized as markers for later stage estrogen-independent mammary tumors (25, 26). We tested both cell lines for the effects of cAMP on proliferation. For both cell types, when 8Br-cAMP was present, cell proliferation in serum containing medium was substantially inhibited (Fig. 1).

Figure 1.

Effects of 8Br-cAMP on the proliferation of MDA-231 and MDA-435 cells. Cells were cultured in DMEM and 10% FCS with or without (control) the addition of 1 mM 8Br-cAMP, MDA-231 (Left) and MDA-435 (bottom) cells. The cells were counted over a period of 7 days. Medium was replenished on day 3. A two-way ANOVA showed that, over the growth period, there was a significant interaction between the cell treatment and cell number for both MDA-231 cells (F_10,36 = 73.74; P < 0.0001) and MDA-435 cells (F_5,24 = 1739; P < 0.0001). cAMP treatment caused significantly lower cell numbers on days 2–7 for MDA-231 (P < 0.01 by Student's t test) and MDA-435 cells (P < 0.005 by Student's t test). Each experiment was carried out in triplicate and repeated at least three times.

We analyzed the probable cause(s) for the antiproliferative effects of cAMP on these cell lines. First, we determined the effect of 8Br-cAMP on MAPK 1,2 activation. Serum-starved cells were treated as indicated, and the phosphorylation state of MAPK was measured with phosphoMAPK-1,2-specific antibodies. In the case of MDA-231 cells, 8Br-cAMP, as well as EGF, strongly stimulated MAPK activity. 8Br-cAMP did not inhibit EGF stimulation of MAPK (Fig. 2A). The data from the MDA-435 cells was different. Here, 8Br-cAMP by itself stimulated MAPK activity, whereas strongly inhibiting EGF stimulated MAPK activity (Fig. 2B). These seemingly contradictory results can be explained by the presence of both c-Raf and B-Raf in these cells because c-Raf is inhibited by protein kinase A (PKA) whereas B-Raf is stimulated by the cAMP pathway (27). Hence, we resolved cellular extracts on SDS/PAGE and blotted for c-Raf and B-Raf with selective antibodies. Both isoforms were found to be present, and no consistent differences in expression were observed under the varying conditions for both cell lines (Fig. 2AII and 2BII). Because 8Br-cAMP substantially inhibited proliferation of both cell types while having differing effects on receptor-tyrosine kinase-stimulated MAPK activity, we looked for other mechanisms by which elevation of cellular cAMP could inhibit proliferation, especially in 231 cells. cAMP is known to trigger the expression of p27kip1, a cyclin-dependent kinase inhibitor, resulting in the inhibition of proliferation (11, 15). In 231 cells, 8Br-cAMP treatment resulted in elevation of p27kip1 in sustained fashion for 12 h (Fig. 3A). In contrast, in the absence of 8Br-cAMP, p27kip1 levels are transiently increased on serum stimulation. The sustained elevation p27Kip1 could contribute to the inhibition of CDK2 (28). Hence, we determined whether CDK2 activity is chronically inhibited in cells treated with 8Br-cAMP. 8Br-cAMP completely suppresses serum-stimulated CDK2 activity in serum-starved cells, although levels of the CDK2 protein do not change (Fig. 3B). Similar results were obtained for MDA-435 cells as well (Fig. 7, which is published as supporting information on the PNAS web site, www.pnas.org). These biochemical experiments suggest that the mechanism by which cAMP inhibits proliferation of MDA-231 cells is different from that of MDA-435 cells. In the case of MDA-231 cells, the major locus of action would have to be at the level of p27kip1, which is elevated in a sustained manner because EGF-stimulated MAPK activity is not inhibited by 8Br-cAMP. In contrast, in MDA-435 cells, it appears that the major effect of cAMP may be at the level of the signaling cascade leading to the activation of MAPK 1,2, but induction of p27kip1 could also play a role. On the basis of the biochemical and cell proliferation data, we tested whether 8Br-cAMP could block expression of the transformed phenotype in both cell lines. By using the colony formation in soft agar, we found that the ability of both cell types [MDA-231, Fig. 8A (which is published as supporting information on the PNAS web site); and MDA-435, Fig. 8B)] to grow in soft agar was substantially inhibited by cAMP. The experiments in Figs. 1–3 and Figs. 7 and 8 suggested that targeted expression of genes that results in constitutive elevation of cAMP may be useful in inhibiting the growth of these cells representing advanced stages of cancer.

Figure 2.

Regulation of MAPK 1,2 activation in MDA-231 and -435 cells. Immunoblot of extracts from MDA-231 (AI) and MDA-435 (BI) cells with anti-phospho-MAPK-1,2 and MAPK-1,2 antibody (total). Cells were treated with or without 1 mM 8Br-cAMP (for 12 h before stimulation) and stimulated with or without EGF for 5 min before harvesting protein. Relative p42 phospho-MAPK levels were determined by densitometric analysis. Normalized values were calculated as ratios of phospho-MAPK to that of total MAPK. (AII) Immunoblot of MDA-231 and (BII) MDA-435 cell extracts with B-Raf antibody (Upper bands) or c-Raf (Lower bands). Cells were treated with or without 8Br-cAMP or EGF and stimulated with or without EGF.

Figure 3.

8Br-cAMP effects on p27kip1 levels and CDK2 activity in MDA-231 cells. (A) Immunoblot of extracts from MDA-231 cells treated with or without 8Br-cAMP, (Upper) with p27 antibody. Relative amounts of p27 (Lower) are shown in arbitrary units determined by densitometric analysis. (B) Immunoblot and activities of CDK2. Extracts from MDA-231 cells, treated with or without 8Br-cAMP for 12 h before stimulation by serum, (Upper bands) were blotted with CDK-2 antibody to determine total CDK-2 protein. Cyclin-dependent kinase-2 activity in immune complexes obtained from MDA-231 cells, treated with or without 8Br-cAMP. Anti-CDK-2 antibodies were used to immunoprecipitate the enzyme. The immune complexes were assayed for kinase activity by their ability to phosphorylate histone H1. The phosphorylated histones were resolved by SDS/PAGE and visualized by autoradiography (Lower band). Quantitation of the phospho-histone bands is shown in arbitrary units as determined by densitometric analysis.

Mutations in the GTPase domains of Gα_s result in constitutive activation of this protein. Stimulation of adenylyl cyclase by Gα*_s results in constitutive elevation of cAMP when expressed in cells, including MCF-7 mammary epithelia cells (10). To introduce the mutant-activated Gα_s into established tumors in mice, we constructed replication-deficient adenoviral vector containing Q227L-Gα_s (29). MDA-231 and MDA-435 cells were transduced by ADV-βgal to determine the optimal moi. An moi of 10 and 200, respectively, were chosen for transduction by ADV-Gα*_s or ADV-βgal as control. These mois were chosen because, at an moi of 10, all MDA-231 cells appeared to be transduced (Fig. 9A, which is published as supporting information on the PNAS web site). At a moi of 100, 30–50% of the MDA-435 cells appeared to be transduced (Fig. 9B), and, at a moi of 200, 85–95% of the MDA-435 cells were transduced (data not shown). Transduction at these mois significantly raised cAMP levels in both cell types (Fig. 9 C and D) when compared with nontransduced cells and cells transduced by the empty adenoviral vector (ADV-EMP). We determined the effect of transduction by ADV-Gα*_s on the proliferation of MDA-231 and MDA-435 cells. When compared with the nontransduced or cells transduced by ADV-β-gal, the cells transduced by ADV-Gα*_s proliferated very slowly (Fig. 4, filled circles). Transduction by ADV-βgal had some inhibitory effects although the cells were able to maintain their proliferative status (Fig. 4, squares).

Figure 4.

The effect of transduction by ADV-βgal and ADV-Gα*_s on proliferation of MDA-231 (A) and MDA-435 cells in culture (B). Cells were plated into six-well dishes 12 h before vector transduction. Cells were cultured in supplemented DMEM with ADV-βgal and ADV-Gα*_s or without (control). The cells were counted over a period of 7 days. Medium was replenished on day 3. A two-way ANOVA showed that, over the growth period, there was a significant interaction between treatment and cell number for MDA-231 (F_10,36 = 881.14; P < 0.0001). An ANOVA with Newman-Keuls post hoc test showed that ADV-Gα*_s-treated cells grew to levels significantly different from either ADV-βgal-treated or control cells on day 4 (F_2,6 = 98.07; P < 0.001), day 5 (F_2,6 = 750.2; P < 0.001), day 6 (F_2,6 = 735.4; P < 0.001), and day 7 (F_2,6 = 4510; P < 0.001). A two-way ANOVA showed that, over the growth period, there was a significant interaction between the cell treatment and cell number for MDA-435 (F_10,36 = 73.74; P < 0.0001). An ANOVA with Newman-Keuls post hoc test ADV-Gα*_s-treated cells grew to levels significantly different from either ADV-βgal-treated or control cells on day 2 (F_2,6 =10.88; P < 0.05), day 4 [F_2,6 = 11.72; P < 0.01 (control), P < 0.05 (βgal)], day 5 (F_2,6 = 7.121; P < 0.05), day 6 (F_2,6 = 14.91; P < 0.01), and day 7 (F_2,6 = 282.2; P < 0.001).

We tested whether expression of Gα*_s inhibited the growth of established tumors in vivo. For this test, 10⁶ MDA-435 cells or 5–10 × 10⁶ MDA-231 cells were injected s.c. into athymic mice, and the tumors were allowed to develop until they reached a tumor volume of 100–200 mm³. At this time, ADV-Gα*_s, ADV-βgal, or PBS was directly injected into the tumor in a volume of 100 μl (≈10⁹ infectious particles). Every seventh day, tumors were measured, then the animals were injected with adenoviral vectors or PBS and followed for up to 30 days after the first injection.

We assessed the extent to which the adenoviral vector transduction resulted in expression of the exogenous proteins 7 days after infection. Although we had tagged Gα*_s with the FLAG epitope, we were unable to conclusively and reproducibly detect the expression of Gα*_s in tissue using the commercially available antibodies by immunohistochemical methods. Hence, we monitored expression by detection of β-galactosidase by using an antibody against the enzyme. For this procedure, tumors were excised 7 days after injection of ADV-βgal, fixed, and embedded in paraffin. Five micrometer cross sections were obtained and then stained with anti-βgal antibody. Staining with anti-βgal antibody showed that 55–65% of the tumor sections expressed the β-galactosidase protein (Fig. 10, which is published as supporting information on the PNAS web site). It appeared that, in the periphery, most cells appeared transduced and that the interior did not show any staining for β-gal (Fig. 10 Upper). Control sections showed no background staining with the anti-βgal antibody (Fig. 10 Lower).

After injections with the adenoviral vectors, we measured tumor growth. For the MDA-231 cells, ADV-Gα*_s-injected cells stopped growing completely and showed a small regression (Fig. 5A). For MDA-231 cells, the ADV-βgal injections also appeared to slow the growth of tumors, as compared with the PBS-injected tumors (Fig. 5A, middle curve), although there was substantial growth (doubling of tumor volume in 30 days). We next determined whether multiple injections of ADV-Gα*_s are necessary. We injected a single bolus of adenoviral vectors or saline into established tumors and followed tumor growth for the next 4 weeks. A single injection was sufficient to halt tumor growth (Fig. 5B). We conducted similar experiments with MDA-435 cells. Subcutaneous tumors were formed by injection of cells. PBS or PBS containing ADV-βgal or ADV-Gα*_s was injected when the tumors reached a volume of 100–200 mm³. Tumors in animals injected with ADV-Gα*_s on a weekly basis stopped growing whereas the tumors in animals injected with PBS or ADV-βgal nearly quadrupled in sized (Fig. 6A). For MDA-435 cells as well, a single injection of ADV-Gα*_s into tumors blocked further growth (Fig. 6B). A representative mouse bearing a tumor formed by MDA-435 cells before injections is shown in Fig. 11A (which is published as supporting information on the PNAS web site). Fig. 11 B and C shows representative mice bearing tumors after treatment with ADV-Gα*_s or ADV-βgal, respectively.

Figure 5.

(A) The effect of PBS, ADV-βgal, and ADV-Gα*_s on established tumors formed by MDA-231 cells in athymic mice. Tumors were measured and injected with ADV-Gα*_s, ADV-βgal, or saline every seventh day for 30 days (indicated by arrows). Calculated volumes were plotted as percentage change of volume from the initial volume measurements (before injections) were normalized to 100%. The data were analyzed by using a two-way ANOVA, with treatment condition as a between subject factor and time after injection as a repeated factor. There was a significant effect of treatment (F_2,104 = 21.60; P < 0.0001) and time (F_3,104 = 7.71; P < 0.0001). There was a significant interaction between the treatment and time (F_6,104 = 2.47; P < 0.03). The data were further analyzed by using Tukey post hoc tests. Mice treated with ADV-Gα*_s differed significantly from controls on days 22 (P < 0.004) and 29 (P < 0.0001), and on no day was there any significant difference to βgal controls. (B) The effect of a single vector injection (arrow) of PBS, ADV-βgal, and ADV-Gα*_s on established tumors of MDA-231. Tumors were measured and injected with a single dose of ADV-Gα*_s, ADV-βgal, or saline and then measured every seventh day for 30 days (indicated by the arrow). Calculated volumes were plotted as percentage change of volume from the initial volume measurements (before injection) were normalized to 100%. The data were analyzed by using a two-way ANOVA, with treatment condition as a between subject factor and time after injection as a repeated factor. There was a significant effect of treatment (F_2,141 = 38.52; P < 0.0001) and time (F_3,141 = 9.99; P < 0.0001). There was a significant interaction between the treatment and time (F_6,141 = 2.63; P < 0.02). The data were further analyzed by using Tukey post hoc tests. Mice treated with ADV-Gα*_s differed significantly from βgal controls on days 15 (P < 0.03), 22 (P < 0.0001), and 29 (P < 0.0001) and differed significantly from PBS controls only on day 29 (P < 0.001).

Figure 6.

(A) The effect of PBS, ADV-βgal, and ADV-Gα*_s on established tumors of MDA-435 origin in athymic nude mice. Tumors were measured and injected with ADV-Gα*_s, ADV-βgal, or saline every seventh day for 30 days (indicated by arrows). Calculated volumes were plotted as percentage change of volume from the initial volume measurements (before injections) were normalized to 100%. The data were analyzed by using a two-way ANOVA, with treatment condition as a between subject factor and time after injection as a repeated factor. There was a significant effect of treatment (F_2,108 = 17.71; P < 0.0001) and time (F_3,108 = 12.09; P < 0.0001). There was a significant interaction between the treatment and time (F_6,108 = 307; P < 0.008). The data were further analyzed by using Tukey post hoc tests. Mice treated with ADV-Gα*_s differed significantly from PBS controls on days 22 (P < 0.05) and 29 (P < 0.0001) and differed significantly from βgal controls only on day 29 (P < 0.0001). (B) The effect of a single injection (arrow) of PBS, ADV- βgal, and ADV-Gα*_s on preformed tumors of MDA-435. Tumors were measured and injected with a single dose of ADV-Gα*_s, ADV-βgal, or saline and then measured every seventh day for 30 days (indicated by the arrow). Calculated volumes were plotted as percentage change of volume from the initial volume measurements (before injection) were normalized to 100%. The data were analyzed by using a two-way ANOVA, with treatment condition as a between subject factor and time after injection as a repeated factor. There was a significant effect of treatment (F_2,62 = 33.39; P < 0.0001) and time (F_3,62 = 13.92; P < 0.0001). There was a significant interaction between the treatment and time (F_6,62 = 2.44; P < 0.035). The data were further analyzed by using Tukey post hoc tests. Mice treated with ADV-Gα*_s differed significantly from βgal controls on days 22 (P < 0.005) and 29 (P < 0.0001) and differed significantly from PBS controls only on day 29 (P < 0.0035).

Discussion

Tumor progression is a complex process. In the case of mammary epithelia, there are several characteristics that stand out. One is expression of receptor tyrosine kinases. For some time now, expression of such receptor tyrosine kinases has been viewed as progression markers (25, 26, 30–32). That such expression is not merely coincidental but in part causal in the later-stage phenotype is best exemplified by the effects of the erb-b2 antibody in the treatment of the advanced metastatic breast cancer (32). Another noteworthy characteristic that co-relates with tumor progression is the decrease in the cell cycle inhibitor p27kip1 levels in primary breast tumor tissue (12). Elevation of cellular cAMP inhibits signal flow from receptor tyrosine kinases to MAPK and stimulates the expression of the cell cycle inhibitor p27kip1 and subsequently decreasing the activity of CDKs. Thus, elevation of cellular cAMP levels in mammary tumors in a sustained manner could be a rational mechanistic approach to block the growth of later stage tumors. Expression of Q227L-Gα_s in the target tissue is a reliable method of constitutively elevating cAMP levels. Our experiments with the adenovirus-directed expression of Q227L-Gα_s in established tumors indicate that this rational approach is successful in blocking tumor growth. This result was not predictable because in vivo other compensatory mechanisms may have easily negated the effects of sustained cAMP elevation. However, it appears that the cAMP-mediated cross-regulation does function in vivo. Hence, induction of sustained interactions between upstream components of different signaling pathways may be a useful approach in regulating the growth of later stage mammary tumors.

Our data and the multiplicity of the sites of action of cAMP suggest that, in different types of mammary tumors, expression of Gα_s could use different mechanisms to inhibit tumor progression. We suggest that, in MDA-231 cells, the biologically relevant cAMP effects are likely to be at the level of p27kip. In contrast, in the MDA-435 cells, the effects are more upstream because elevation of cAMP blocks receptor activation of MAPK. It has been long noted that there several distinct subclasses of later stage breast cancers, as characterized by progression markers such as erb-B2 or EGF receptors (21–26). Our functional analysis suggests that there are differences in the organization of the signaling pathways as well. In the MDA-435 cells, the action of c-Raf seems to predominate because cAMP inhibits EGF receptor activation of MAPK. In contrast, in MDA-231 cells, the action of B-Raf appears to predominate because cAMP by itself extensively stimulates MAPK and cAMP does not inhibit tyrosine kinase-stimulated MAPK stimulation. Here, inhibition of proliferation appears to be achieved by induction of the cell cycle inhibitor p27kip1 and a decrease in CDK2 activity. Sustained elevation of cAMP acting at multiple sites may be of general value because it is capable of inhibiting the growth of different subclasses of late stage mammary tumors.

Although we could block tumor growth completely, we were unable to obtain tumor regression by expression of Gα_s. This result could be due to two reasons. First, because we only infect the outer layer of the tumor cells, the bystander effect is not sufficiently robust to kill all of the tumor cells. Alternatively, the cells in the tumor are quiescent but not committed to apoptosis and an agent that promotes apoptosis when used in conjunction with expression of Gα*_s will cause full tumor regression. The observation of a lack of gross morphological differences between the Gα*_s vs. the β-gal and vehicle-injected tumors (data not shown) suggests that the later explanation might be correct. These alternatives will have to be definitively resolved by further experiments.

In conclusion these studies show that the interactions between signaling pathways could be a basis for therapy of later stage breast cancers. The development of appropriate small molecule agents targeted to such interaction sites may be therapeutically valuable.

Abbreviations

MAPK

mitogen-activated protein kinase

CDK

cyclin-dependent kinase

moi

multiplicity of infection

EGF

epidermal growth factor

pfu

plaque-forming unit

X-gal

5-bromo-4-chloro-3-indolyl β-d-galactoside

ADV-βgal

adenoviral vector containing β-galactosidase