BMP4 signaling induces senescence and modulates the oncogenic phenotype of A549 lung adenocarcinoma cells

S. Buckley, W. Shi, B. Driscoll, A. Ferrario, K. Anderson, D. Warburton


Lung cancer is the most common visceral malignancy in males, with rapidly increasing incidence in females, and a devastatingly poor prognosis. Transforming growth factor (TGF)-β has been shown to induce senescence in A549 lung cancer cells, and both TGF-β and bone morphogenetic protein (BMP) 2 can suppress the transformed phenotype of A549 cells in vitro. We examined the effects of BMP4, another member of the TGF-β superfamily, on specific oncogenic properties of A549 cancer cells. When A549 cancer cells were treated continuously with 100 ng/ml of BMP4, a senescent phenotype was observed after 2 wk of treatment. The BMP-treated cells appeared larger than untreated cells, grew more slowly, had more senescence-associated β-galactosidase activity, and had less telomerase activity, as measured by the telomeric repeat amplification protocol assay. Invasion through Engelbreth Holm-Swarm matrix was inhibited in the senescent cell population. Senescent BMP4-treated cells had lower ERK activation, VEGF expression, and Bcl2 expression than wild-type cells, consistent with a less proliferative, less angiogenic phenotype with increased susceptibility to death by apoptosis. BMP4 treatment also resulted in sustained elevation of Smad1. In vivo xenograft studies in the flanks of nude mice confirmed that the BMP-treated cells were significantly less tumorigenic than untreated cells. Direct overexpression of Smad1 using adenoviral constructs resulted in cell death within 5 days. These studies suggest that BMP4 pathway signaling can induce senescence and thus negatively regulate the growth of A549 lung cancer cells.

  • bone morphogenetic protein 4
  • senescence-associated β-galactosidase activity
  • telomerase
  • Engelbreth Holm-Swarm matrix invasion
  • xenograft tumor

lung cancer is the primary cause of cancer deaths in the United States, with a devastatingly poor prognosis (22). There has been little improvement in 5-year survival rates over the last 20 years, and, for the most common treatments, toxicity exceeds efficacy.

Bone morphogenetic proteins (BMPs) are one of the subgroups of the transforming growth factor-β (TGF-β) superfamily. TGF-β and BMP transduce signals through downstream proteins, Smads, controlling cell proliferation, differentiation, migration, and apoptosis (1, 14). Frequent activating mutations of Smad4 (10) or less common somatic mutations of Smad2 (19) are seen in solid tumors and suggest that these genes may have a suppressor function. A ubiquitously expressed truncated Smad5, Smad5-β, has been implicated in stem cell homeostasis as a mechanism to protect pluripotent stem cells and malignant cells from the growth inhibitory and differentiation signals of BMPs (9). Functionally impaired Smad mutations may also play a role in lung tumorigenesis (24).

The concept of lung cancer growth suppression through TGF-β signaling is supported by several studies. TGF-β has been shown to drive A549 cancer cells into replicative senescence in vitro (11) by mechanisms unknown, and TGF-β or BMP2 treatment can suppress the growth of A549 cells in vitro (17, 18). However, the effects of BMP4 signaling on nonsmall cell lung cancers have not been explored. Herein, we investigate the effects of BMP4 on cell senescence and tumorigenicity in A549 lung cancer cells. We present data demonstrating that BMP4 at a dose of 100 ng/ml is effective in inducing senescence in A549 lung cancer cells in vitro. More importantly, we show that BMP4 pretreatment can significantly reduce tumorigenicity of A549 xenograft tumors generated by these cells in nude mice. We demonstrate that BMP4-mediated overexpression of Smad1 in A549 cells correlates with senescence, whereas direct overexpression of Smad1 using adenoviral vectors results in cell death. These data suggest that BMP4 or downstream targets may be effective in the control of certain lung cancers.


Parental cell line. A549 lung carcinoma cells (13) were obtained from ATCC (CCL-185), and parental cultures were maintained in HEPES-buffered RPMI with 10% FCS and antibiotics. The medium was replaced every other day, and the cells were passaged before confluence. As the A549 cell line consists of four distinct clones (2), each BMP4 treatment was done on cells fresh from the ATCC to ensure a reproducible response.

BMP4 treatment. A549 cells were grown in RPMI with 1% FCS ± 100 ng/ml of human recombinant BMP4 (R & D Systems) continuously present. Control cells were grown in parallel in 1% FCS. The medium ± BMP4 was replaced every other day, and the cells were passaged before confluence. At various times during the BMP4 treatment, flasks of A549 cells ± BMP4 were grown to near confluence, trypsinized, and counted by hemocytometer to compare growth rates.

Senescence-associated β-galactosidase assay. Cells were lysed in reporter lysis buffer (Promega, Madison, WI) and centrifuged at 20,800 g at 4°C for 1 min, and the supernatants were removed. Proteins were measured with Bio-Rad protein dye (Bio-Rad, Hercules, CA). Cell lysates containing equal amounts of protein were diluted in equal volumes of 2× assay buffer containing 1.33 mg/ml o-nitrophenyl-β-d-galactopyranoside, 2 mM MgCl2, and 100 μl 2-mercaptoethanol in 200 mM phosphate buffer, pH 6.0, and incubated at 37°C for 4 h. The absorbance at 420 nm was measured after the addition of an equal volume of 1 M Na2CO3.

Telomeric repeat amplification protocol assay. Sample preparation and telomeric repeat amplification protocol (TRAP) assays were performed according to the TRAP-EZE protocol provided by Serologicals (Norcross, GA) (3). Briefly, at least 106 cells for each sample were lysed in 1× 3-([3-cholamidopropyl]dimethylammonio)-1-propanesulfonate lysis buffer. The lysate was clarified by centrifugation, and protein content was measured with Bio-Rad dye. To assay telomerase activity, we incubated samples with a [γ-32P]dATP end-labeled telomerase specific primer at 30°C for 30 min for telomere primer extension. The telomerase products were then amplified by 30 rounds of two-step PCR (94°C/30 s, 60°C/30 s). The samples were subjected to 12.5% nondenaturing polyacrylamide gel electrophoresis in 0.5× TBE buffer (45 mM Tris-borate, 1 mM EDTA) for 1 h at 500 V. Gels were dried and exposed to X-ray film to visualize the telomerase products. Each assay included telomerase-positive A549 cells that had not been treated with BMP4, as a cellular control, as well as a PCR internal amplification control, provided by Intergen, and a PCR contamination control lane consisting of all sample elements with the exception of cell lysate.

Western blotting of proteins. Cells were washed in PBS, lysed in iced radioimmunoprecipitation assay buffer (1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS in PBS, with 10 mg/ml PMSF, 30 U/ml aprotinin, and 100 mM sodium orthovanadate), and centrifuged at 14,000 g for 10 min at 4°C. Protein (20 μg) was electrophoresed through commercially prepared polyacrylamide gels (Novex, San Diego, CA) and transferred to Immobilon membrane (Millipore, Bedford, MA). Proteins of interest were detected with horseradish peroxidase-linked secondary antibodies (Sigma, St. Louis, MO) followed by ECL detection (Amersham, Little Chalfont, UK). We confirmed equal loading by blotting the lower half of the blots with an antibody to actin. Antibodies to VEGF, phosphorylated ERK (p-ERK), Bcl2, interleukin-1β-converting enzyme (ICE), and p-Smad1 were from Santa Cruz Biotechnology (Santa Cruz, CA). Antibody to actin was from ICN (Irvine, CA). Antibody to FLAG, a tagging epitope, was from Sigma.

Cell invasion assay. A commercial cell invasion assay kit (Chemicon, Temecula, CA) was used according to the manufacturer's instructions. The inserts, which contain 8-μm pores, are coated with a thin layer of biological matrix derived from the Engelbreth Holm-Swarm (EHS) mouse tumor. Cells (3.5 × 105) in 300 μl of serum-free RPMI + 0.1% BSA were added to triplicate upper wells, and migration through the matrix to the lower wells, containing 500 μl of RPMI + 10% FCS, was measured after 24-h culture at 37°C. The noninvading cells and matrix were removed with a cotton-tipped swab, and the cells were stained. Cell-free filters were processed as blanks. The stain was eluted with 10% acetic acid, and the absorbance at 560 nm was measured.

Adenoviral transduction of Smad1. FLAG-tagged Smad1 recombinant adenovirus (5) was a kind gift from Dr. M. Fujii (Japan Society for the Promotion of Science, Tokyo). High titer stocks of recombinant virus were grown in 293 cells and purified with a Viraprep kit (Virapur, Carlsbad, CA) according to the manufacturer's instructions. Purified adenovirus containing green fluorescence protein (GFP) or Smad1 was added to nonconfluent cultures of A549 cells at a concentration of 20 live particles/cell in RPMI with 5% FCS. The virus-conditioned medium was removed after overnight incubation, and the cells were further incubated in RPMI and 5% FCS. The GFP-transduced cells were observed daily under fluorescent microscopy for GFP expression, which was seen maximally after 48 h (estimated to be 80-90% of the culture), and persisted until the death of the Smad1-transduced cells around day 5. We confirmed overexpression of Smad1 by lysing the Smad1 and control GFP-transduced cells and then Western blotting for FLAG expression.

Tumorigenicity in nude mice. Viable single cell suspensions (4 × 106) from BMP4-treated and untreated A549 cells were injected subcutaneously into the flanks of nude mice obtained from Harlan Sprague-Dawley. Tumor size was measured in three dimensions weekly, and the tumor volume was calculated by the formula (length × width × height × π/6). The animals were killed when necrosis occurred in the larger tumors arising from untreated cells (45 days postinjection). Animal experiments were approved by the Institutional Animal Care and Use Committee and conducted in accordance with Institutional guidelines.


BMP4 treatment induces a senescent morphology in A549 lung cancer cells. Cultures of A549 cells treated continuously with BMP4 showed reduced growth compared with untreated cells (Fig. 1A). A change in morphology was apparent after 2 wk of treatment. The slower-growing cells were larger, flattened, and more granular than untreated cells (Fig. 1B)

Fig. 1.

A: bone morphogenetic protein (BMP) 4-treated cells grow more slowly than untreated cells cultured under identical low-serum conditions. At various times during the BMP4 treatment, flasks of A549 cells ± BMP4 were grown to near confluence, trypsinized, and counted by hemocytometer to compare growth rates. B: BMP4-treated senescent A549 cells have a larger, more flattened phenotype than quiescent, untreated cells grown in parallel in low serum. Left: untreated cells after 21 days (d) of culture. Right: BMP4-treated, senescent cells after 21-d culture.

BMP4 treatment induces senescence-activated β-galactosidase activity in A549 cancer cells. A549 cells were treated with BMP4 until the growth rate slowed and the cells assumed an enlarged, flattened appearance compared with untreated cells grown under parallel low-serum conditions. BMP4-treated and untreated cultures were lysed and assayed for senescence-activated β-galactosidase (SA-β-gal; pH 6.0) activity, a biomarker associated with senescence in human cells (23). A fourfold increase in SA-β-gal activity was seen after 17 days of BMP4 treatment (Fig. 2). Increased SA-β-gal activity was sustained through prolonged culture in the presence of BMP4 (60 days) and for at least 1 wk after BMP4 was withdrawn (data not shown).

Fig. 2.

A549 cells treated with BMP4 for 17 d have significantly higher senescence-activated β-galactosidase (SA-β-gal) activity than untreated cells cultured under identical low-serum conditions. Data are presented ± SD, n = 4, ***P < 0.05. A549 cells cultured ± 100 ng/ml BMP4 for 17 d were lysed, and the lysates tested for SA-β-gal activity, using o-nitrophenyl-β-d-galactopyranoside as substrate at pH 6.0.

BMP4 treatment downregulates telomerase activity in A549 cells. A549 cells, in common with many cancer cells, have abundant telomerase activity that correlates with a highly proliferative phenotype (3). A549 cells treated with TGF-β have reduced telomerase activity that is associated with senescence (11). We analyzed lysates of senescent BMP4-treated and quiescent untreated A549 cells for telomerase activity, using the PCR-based TRAP. Autoradiography showed characteristic laddering, representing radiolabeled telomere primer extension, in the wild-type A549 cells (Fig. 3, lane 5). Laddering was reduced in untreated A549 cells grown for 17 days in low-serum conditions making them essentially quiescent (lane 3) and dramatically reduced in cells treated with BMP4 for 17 days and expressing the senescence biomarker SA-β-gal (lane 4).

Fig. 3.

BMP4-induced senescence in A549 cells is associated with decreased telomerase activity. The representative autoradiograph shows that A549 cells treated with BMP4 for 17 d (lane 4) have greatly decreased telomerase activity compared with untreated cells (lane 3), as measured by the laddering, which represents radioactive telomere extension products. Each assay included telomerase-positive parental A549 cells grown in 10% serum as a cellular control (lane 5), as well as a PCR internal amplification control (lane 1), and a PCR contamination control lane, consisting of all sample elements with the exception of cell lysate (lane 2). TRAP, telomeric repeat amplification protocol; WT, wild type.

BMP4-induced senescence in A549 cells is associated with downregulation of p-ERK, VEGF, and Bcl2. Western analysis of lysates of BMP4-treated A549 cells (Fig. 4) suggests a decrease in proliferative and angiogenic phenotypic markers compared with untreated cells cultured under parallel low-serum conditions, as shown by decreased p-ERK and VEGF expression, respectively. The expression of survival factor Bcl2 is lower in BMP4-treated cells, suggesting an increased potential for apoptosis in the senescent population. Despite the decreased Bcl2 levels, a population of cells can survive culture in the presence of BMP4 for 8 wk or more. This survival may reflect the polyclonal nature of the A549 cell line (2), with diverse effects elicited from different subclones.

Fig. 4.

BMP4 treatment of A549 cells results in a sustained less proliferative, less angiogenic phenotype, as reflected by decreased expression of phosphorylated ERK (p-ERK) and VEGF, as seen in this representative Western blot. The expression of Bcl2 is also decreased in BMP4-treated cells, suggesting that the senescent cells may be more prone to death by apoptosis. Actin was used as loading control.

BMP4-treated A549 cells are less invasive through EHS matrix. We compared the invasive properties of senescent BMP4-treated cells with untreated cells by measuring invasion through EHS matrix (Fig. 5). Migration through EHS matrix was significantly retarded by ∼50% in the BMP4-treated cells compared with untreated cells (P < 0.05). The nonmigrating population was viable and attached readily to the matrix. The decreased invasion of BMP4-treated A549 cells through a basement membrane model suggests that BMP4 treatment may result in a less metastatic phenotype by downregulating the proteolytic enzymes required for matrix penetration. Wild-type A549 cells are highly metastatic and secrete high levels of matrix metalloproteinase-2 (25).

Fig. 5.

Cells treated with BMP4 for 17 d are significantly less invasive through Engelbreth Holm-Swarm (EHS) matrix-coated porous inserts than untreated cells grown in identical low-serum conditions, **P < 0.01. Data are presented ± SD, n = 4. BMP4-treated and untreated A549 cells, cultured under identical low-serum conditions, were plated in serum-free medium on a layer of EHS matrix over a porous filter, atop serum-containing medium. Invasion through matrix was measured after 24 h.

BMP4-treated A549 cells generate slower-growing xenograft tumors in nude mice. Xenograft tumors were established by subcutaneous injection of 4 × 106 viable untreated or BMP4-treated A549 cells into the flanks of nude mice. The tumor size was recorded weekly, and the tumors were excised when necrosis was observed in the larger tumors generated from untreated cells (45 days postinjection). The tumors from BMP4-treated A549 cells grew more slowly than tumors from untreated cells (Fig. 6A). The weights at excision of tumors arising from BMP4-treated cells were significantly (50%) less than tumors generated by untreated cells (Fig. 6B). These data support the in vitro findings, which suggest that the BMP4-treated cells are less proliferative and angiogenic than the quiescent untreated cells.

Fig. 6.

A: BMP4-treated cells resulted in insignificantly smaller xenograft tumors in nude mice compared with untreated cells grown under identical low-serum conditions. Viable BMP4-treated and untreated cells (4 × 106) cultured under identical low-serum conditions were injected into the flanks of nude mice, and the tumors were measured weekly in 3 dimensions. Tumor volume was calculated assuming spheroid shape, using the formula length × width × height × π/6. The experiment was stopped when necrosis was observed in the larger tumors. Data are expressed as means ± SD (untreated, n = 6; BMP4-treated, n = 8). B: weights of tumors resulting from BMP4-treated A549 cells were significantly less at excision (45 days postinjection) than tumors arising from untreated cells, *P < 0.05. Data are presented as means ± SD.

Smad1 phosphorylation is induced by BMP4 treatment of A549 cells. Western analysis shows that Smad1 is phosphorylated in A549 cells within 1 h after BMP4 treatment (Fig. 7A). Interestingly, p-Smad1 is not downregulated with time and remains elevated during culture with BMP4 (up to 40 days), suggesting that Smad1 activation correlates with slowed growth rate and senescence in A549 cells. To determine whether BMP4-induced senescence in A549 cells is receptor mediated, we bypassed the receptors by overexpressing Smad1. Purified recombinant adenoviral vectors bearing FLAG-tagged Smad1 or GFP were transduced into A549 cells. Maximal expression was achieved at 48 h posttransduction, as detected by fluorescent GFP expression in the control group. The cells were then lysed and analyzed to confirm Smad1 overexpression by Western blotting for FLAG (Fig. 7B). Interestingly, overexpression of Smad1 resulted in cell death, rather than senescence. Smad1-transduced cells began lifting off the plate soon after Smad1 overexpression was established, whereas GFP-transduced controls remained viable. Elevated levels of bax and ICE observed in the BMP4-treated cells before lifting from the plate, compared with the GFP-transduced controls, suggested that cell death may be due to apoptosis. Death of the total cell population occurred over several days (Fig. 7C).

Fig. 7.

A: representative Western analysis of lysates of A549 cells treated continuously with 100 ng/ml of BMP4 for various periods of time shows sustained activation of Smad1 in the BMP4-treated group compared with cells cultured under identical low-serum conditions. B: A549 cells were transduced with adenovirus bearing FLAG-tagged Smad1 or green fluorescence protein (GFP) at the same multiplicity of infection (MOI). Western analysis of FLAG expression in lysates of A549 cells at 48 h posttransduction confirms successful overexpression of Smad1 and shows increased bax and interleukin-1β-converting enzyme (ICE) expression in the cells overexpressing SMAD1. These cells would subsequently die. C: flasks of A549 cells were transduced with either adenovirus bearing FLAG-tagged Smad1 or GFP at the same MOI and harvested at various times after GFP expression was detected. Cell counts show that overexpression of Smad1 in A549 cells results in cell death within 5 days.


Novel therapeutic approaches to lung cancer are crucial because of the disease's dismal prognosis, which has persisted for more than 20 years (22). Although a cure is the primary goal of such therapies, control of tumor growth may be more achievable and would still afford significant benefits to the patient, such as longer survival and improved quality of life. Several reports demonstrate that members of the BMP family (or downstream targets) can inhibit growth of various cancers in vitro, including breast, skin, thyroid, and multiple myeloma (4, 6, 7, 16, 21). We examined the effects of in vitro BMP4 treatment on human lung cancer cells using the A549 lung cancer cell line derived from type II alveolar epithelial cells (13). These cells are highly proliferative, tumorigenic, and invasive (25) and have abundant telomerase activity (3). We show that BMP4 induces senescence in A549 lung cancer cells in vitro, as characterized by slowed growth rate, increased β-galactosidase expression, and reduced telomerase activity compared with quiescent cells grown in parallel in low serum. Although the BMP4-treated cells express decreased levels of Bcl2, the reduced cell numbers in the treated cultures probably reflect slowed growth rather than apoptosis, since there was no observable cell detachment from the plates and we could detect neither bax nor ICE in these cells (data not shown). In addition, the cultures survived in the presence of BMP4 for up to 8 wk.

The inhibitory action of BMP4 on certain cancer cell lines in vitro has been shown to be related to upregulation of the cyclin-dependent kinase (cdk) inhibitors p21 and p27, with subsequent decreased activities of cdk2 and cdk6 and hypophosphorylation of the retinoblastoma protein pRb (6). In BMP2-mediated cell cycle arrest of breast cancer cells, p21 promoter activity required both Smad1 and -4 and was induced by either BMP2 or constitutively active type 1 BMP receptors (BMPR) (16). The response of A549 lung cancer cells to BMP4 may not be simple cell cycle arrest, as seen with BMP2-treated breast cancer cells, since we found that BMP4-induced senescence was associated with only a modest increase in p21 compared with the untreated cells and no significant changes in the phosphorylation status of Rb (data not shown).

It has been reported that A5DC7 cells, a subline of A549 cells, senesce with long-term culture in low serum but can be restored to normal phenotype within days by their return to 10% serum culture conditions (12). In our experiments, A549 cells grown in low serum for several weeks in parallel with the BMP4-treated group did not show the hallmarks of senescence seen in the BMP4-treated group, i.e., increased SA-β-gal activity and reduced telomerase activity. In addition, we found that the senescence induced by BMP4 in A549 cells was not immediately reversible, since increased levels of p-Smad1, decreased growth rate, and increased SA-β-gal activity persisted for several weeks after the removal of BMP4 and restoration of 10% serum culture conditions (data not shown).

The persistence of the senescent phenotype after removal of the ligand may explain why BMP4-treated A549 cells generate significantly smaller xenograft tumors in nude mice than untreated cells. This is the first report to document the antitumorigenic potential of a BMP family member using A549 cancer cells in an in vivo model. Although in vivo metastasis was not examined in this project, in vitro matrix invasion assays suggest that the BMP4-treated cells are less invasive. Tada et al. (18) have reported that BMP2 at a dose of 100 ng/ml suppresses anchorage-dependent growth in cultured A549 cells.

Sustained overexpression of Smad1 was associated with a senescent, less invasive, and less tumorigenic phenotype in BMP4-treated A549 cells. In contrast, sustained expression of Smad5, which also transduces the BMP4 signal, was not detected (data not shown). Smads are pivotal intracellular nuclear effectors of TGF family members (8). BMP-mediated activation of downstream Smads involves ligand binding to the high-affinity receptor BMPR-1 and subsequent recruitment of BMPR-2 into the complex. This is followed by phosphorylation of Smads 1, 5, and 8, association with Smad4, and translocation to the nucleus. However, before ligand binding, low but measurable preformed receptor complexes have been detected. Receptor mutant studies show that signals induced by binding of BMP2 to preformed receptor complexes activates the Smad pathway, whereas BMP2-induced recruitment of receptors activates a different, Smad-independent pathway via p38 MAPK (15). Additionally, members of the BMP-Smad pathway can also physically interact with components of other signaling pathways to establish cross talk (20). Thus BMP4 ligand may trigger multiple downstream pathways. To determine whether sustained Smad1 activation itself is sufficient to result in senescence, we overexpressed Smad1 in A549 cells. Adenovirus-mediated Smad1 overexpression resulted in the death of A549 cells, rather than senescence, as seen with BMP4 treatment. The Smad1-overexpressing cells started to lift after 48-72 h posttreatment, and all were dead within 1 wk. Whether this response is unique to A549 cells or can be induced in other cancer cell lines remains to be determined. In the other lung cancer cell lines we tested, overexpression of Smad1 in CALU1 (squamous cell carcinoma) cells resulted in senescence yet had no deleterious effects on SW900 (epidermoid carcinoma) cells, indicating cell-specific responses (data not shown).

In A549 cells, activation of BMP4 downstream Smad1 signaling causes cell death, whereas the ligand itself induces a senescent phenotype. Although the specific downstream targets of BMP4 and/or Smad1 that mediate the phenotypic changes described herein remain to be determined, our data suggest that BMP4 and downstream mediators may have potential to control or retard the growth of certain lung cancers.



This work was funded by National Heart, Lung, and Blood Institute Grants HL-44060, HL-44977, HL-61286, HL-60231 and HL-68597.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract