Upregulation of two BH3-only proteins, Bmf and Bim, during TGFβ-induced apoptosis


Transforming growth factor-β (TGFβ)-activated signalling pathways can lead to apoptosis, growth arrest or promotion of malignant behaviour, dependent on cellular context. The molecular mechanisms involved in TGFβ-induced apoptosis remain controversial; although changes in gene expression are thought to be pivotal to the process, several different candidate apoptotic initiators and mediators have been proposed. Smad4, a critical component of the TGFβ-induced transcriptional machinery, is shown here to be essential for induction of apoptosis. Gene expression analysis identified the proapoptotic Bcl-2 family members, Bmf and Bim, as induced by TGFβ, dependent on both Smad4 and p38 function and the generation of reactive oxygen species. TGFβ-induced Bmf and Bim localize to cellular membranes implicated in apoptosis. Inhibition of the TGFβ-induced expression of both these proteins together provides significant protection of cells from apoptosis. The TGFβ-triggered cell death programme thus involves induction of multiple BH3-only proteins during the induction of apoptosis.


The transforming growth factor-β (TGFβ) family of cytokines activates an array of signalling pathways, including Smad2-, Smad3- and Smad4-mediated transcription and stress-activated kinases such as p38 and Jun-N-terminal kinase (JNK) (Wakefield and Roberts, 2002). These pathways can lead to reduction in cell growth and survival (Derynck and Zhang, 2003), with Smad4 expression being lost in certain tumours (Hahn et al., 1996). A key outcome of TGFβ signalling in some cell types is the induction of apoptosis; lymphocytes and hepatocytes are particularly sensitive. A striking characteristic of TGFβ-induced apoptosis is the slow onset, with requirement for changes in gene expression (Inayat-Hussain et al., 1997; Teramoto et al., 1998). This has precipitated a search for potential apoptotic regulators whose expression is controlled by TGFβ (Coyle et al., 2003; Yang et al., 2003).

To date, several proteins have been put forward as potential candidates for such a role, including IκB-α (Arsura et al., 1996), death-associated protein kinase (DAP-kinase) (Jang et al., 2002), the lipid phosphatase SHIP-1 (Valderrama-Carvajal et al., 2002), the stress and cytokine inducible GADD45b protein (Yoo et al., 2003) and the connective tissue growth factor (CTGF) (Hishikawa et al., 1999). Other proteins that play a facilitative role in TGFβ-induced death, but whose expression is not induced by it, include the adaptor proteins DAXX (Perlman et al., 2001) and CD2AP (Schiffer et al., 2004) and the septin ARTS-(Larisch et al., 2000). In addition, the proapoptotic Bcl-2 family protein Bim is a Smad3-dependent TGFβ-induced protein in B lymphocytes (Wildey et al., 2003), whose binding to Bcl-XL increases during TGFβ-induced apoptosis (Ohgushi et al., 2005). Furthermore, Bad (Kim et al., 2002) and Bid (Kim et al., 2004) undergo caspase-dependent cleavage to more potently apoptotic forms after TGFβ stimulation, while the antiapoptotic proteins Bcl-XL and Bcl-2 have been demonstrated to undergo TGFβ-mediated downregulation in several cell types (Saltzman et al., 1998; Francis et al., 2000; Chipuk et al., 2001; Kanamaru et al., 2002). Finally, TGFβ may antagonise survival signalling through the physical interaction of Smad3 with Akt (Conery et al., 2004; Remy et al., 2004).

Although many of these proteins may play some role in TGFβ-induced apoptosis, likely with variations between cell types, it has proved difficult to establish an unequivocal requirement for any of them as the key activator of the apoptotic machinery, leading to the idea that there might exist substantial redundancy in the transcriptional response to TGFβ that ensures efficient execution of the cell death programme. We report here the results of a microarray analysis of TGFβ-regulated transcription in a mouse hepatocyte cell line where apoptosis is the overwhelming response to TGFβ treatment. We find that TGFβ strongly induces expression of the proapoptotic Bcl-2 family proteins Bmf and Bim in these cells and also in a number of other cell systems. Suppression of the induction of both of these two BH3-only proteins together significantly inhibits TGFβ-driven apoptosis, suggesting that they might play at least partially redundant roles in the regulation of cell death. Bmf and Bim acting together may be significant inducers of TGFβ-regulated apoptosis in a number of cell types.


In order to investigate the mechanisms involved in TGFβ-induced cell death, we used the normal murine hepatocyte cell line, AML12, and the normal murine mammary epithelial cell line, NMuMG. Addition of TGFβ caused extensive cell death by 48 h in both cells with concomitant caspase-3 cleavage (Figure 1a and b). Analysis of the two cell types by Annexin V/4,6-diamidino-2-phenylindole (DAPI) or propidium iodide (PI) staining revealed differences in the kinetics of changes in their fluorescence-activated cell sorting (FACS) profiles after TGFβ stimulation (Figure 1c). AML12 cells rapidly became both Annexin V-positive and DAPI-positive, whereas NMuMG cells appeared to progress more slowly through the cell death programme, accumulating in the Annexin V-positive only state before membrane rupture and subsequent DAPI staining (Figure 1c). AML12 cells also appeared to undergo loss of mitochondrial membrane potential more rapidly than NMuMG cells (Figure 1c). As caspase-3 was activated in both cell types, it appeared likely that caspases were responsible for mediating cell death in both AML12 and NMuMG cells. This was confirmed using the broad-spectrum caspase inhibitor, Boc-D-FMK, which prevents TGFβ-induced apoptosis in both cell types (Figure 1d). However, AML12 cells are only partially protected by the more selective caspase inhibitor Z-VAD-FMK (Figure 1d), possibly suggesting the involvement of other effector caspases in this line.

Figure 1

Characterization of TGFβ-mediated apoptosis in AML12 and NMuMG cells. (a) Photographs of AML12 and NMuMG cells undergoing TGFβ-induced apoptosis taken 48 h after TGFβ stimulation. Scale bars: 25 μm. (b) A Western blot time course analysis of caspase-3 cleavage during TGFβ-induced apoptosis. The processed and cleaved form of caspase-3 are indicated. (c) Kinetics of TGFβ-mediated cell death. Alexa Fluor 647 Annexin V/DAPI staining showing cell death in AML12 and NMuMG cells at the indicated times after addition of TGFβ. Mitochondrial membrane potential was also monitored with TMRE. (d) Different caspase inhibitors have different effects on cell death in AML12 and NMuMG cells. Cells preincubated with the broad-spectrum caspase inhibitors, Z-VAD-FMK (Z-VAD) or Boc-D-FMK (Boc-D) were scored for the percentage of Annexin V and PI-positive cells after 48 h of TGFβ exposure. The bar graphs are a representative example of the experiment. This experiment was repeated at least three independent times. The data show means+s.d. The numbers shown represent the fold changes over unstimulated cells.

To investigate the role of the various TGFβ-induced signalling pathways in the induction of apoptosis, we knocked down Smad4 expression in both cell types using RNA interference (RNAi). Both stable cell clones expressing a short hairpin RNA vector and cell populations transiently transfected with synthetic small interfering RNA oligonucleotides showed downregulation of Smad4 (Figure 2a). The transient and stable RNAi systems targeted different sequences in the Smad4 mRNA, reducing the likelihood of artifactual off-target effects. Smad4 downregulation dramatically inhibited the induction of cell death by TGFβ in both cell types (Figure 2a), confirming the importance of the Smad pathway in apoptosis.

Figure 2

Smad4 and p38 play distinct roles during TGFβ-mediated cell death. (a) Western blots showing downregulation of Smad4 in a clone of AML12 cells stably expressing smad4 RNAi oligonucleotides, and a pool of NMuMG cells transfected with smad4 siRNA for 48 h (left panel). Smad4 downregulation inhibits TGFβ-induced cell death. Stable AML12 cells or transiently transfected NMuMG cells scored for cell death using Annexin V/PI staining following TGFβ-stimulation for 48 and 24 h, respectively. The numbers shown represent the fold changes over unstimulated conditions (right panel). (b) The p38 MAPK pathway plays a role during the later phase of apoptotic cell death in AML12 cells, but not in NMuMG cells. Western blotting analysis of the p38-substrate MAPKAPK-2 after 24 h of TGFβ stimulation with or without the p38 inhibitor, SB202190 (top panel). Cell death analysis after inhibition of the p38 MAPK pathway during early (0–12 h) and late (12–48 h) phases of TGFβ-induced apoptotic cell death (bottom panel). The bar graphs are a representative example of the experiment. This experiment was repeated at least three independent times. The data show means+s.d. The numbers shown represent the fold changes over unstimulated cells.

The p38 pathway has also been implicated in TGFβ-induced cell death (Liao et al., 2001; Schrantz et al., 2001; Park et al., 2002; Yu et al., 2002), with use of an MKK3 dominant-negative construct or p38 inhibitors providing protection against TGFβ-induced killing in some systems (Yu et al., 2002; Edlund et al., 2003; Yoo et al., 2003). However, there are cell type differences as to whether TGFβ induces p38 activation and whether this is Smad-dependent or Smad-independent (Takekawa et al., 2002; Yu et al., 2002; Yoo et al., 2003), with early and late phases of p38 activation also likely employing different mechanisms. In the cell lines used here, TGFβ induced only minor activation of p38 as shown by analysis of MAPKAPK-2 phosphorylation (Figure 2b) and p38 phosphorylation at various time points (data not shown). The p38 kinase inhibitors SB202190 (Figure 2b) and SB203580 (data not shown) markedly inhibited the basal level of p38 activity in AML12 and NMuMG cells. In order to address whether, and over what time period, p38 activity might be essential for TGFβ-mediated cell death, the p38 inhibitors were either used during the first 12 h or during the final 36 h of TGFβ stimulation. Inhibition of the p38 pathway in AML12 cells robustly abrogated TGFβ-induced cell death, but only when the inhibitor was applied at the later time interval (12–48 h). By contrast, NMuMG cells demonstrated little alteration in their response to TGFβ on treatment with p38 inhibitors (Figure 2b). The importance of the p38 pathway in the induction of apoptosis by TGFβ therefore differs between these two cell types, while Smad function is essential in both cases. As TGFβ only causes relatively small increases in p38 activity, basal p38 signalling may play an important role in the induction of apoptosis by TGFβ in AML12 cells. In these cells, it has been proposed that p38 activation is downstream of Smad-mediated induction of GADD45b (Yoo et al., 2003).

As changes in gene expression are required for TGFβ-induced apoptosis, we sought to identify apoptotic genes regulated by TGFβ in a Smad4- and/or p38-dependent manner in AML12 cells using Affymetrix microarrays. From three independent replicates, a number of cell death-associated proteins were identified as being differentially up- or downregulated by TGFβ between wild-type and the Smad4-deficient cells, or between wild-type cells in the presence or absence of the p38 inhibitor. In addition, well-characterized TGFβ-regulated genes such as PAI-1 (induced 4.1-fold) and c-Myc (suppressed 2.8-fold) were identified, acting as positive controls for microarray function. The Bcl-2 protein family was significantly represented in our data set, with Bmf, Bim and Bcl-XL detected in the TGFβ-regulated Smad4-dependent gene list, as well as Bmf and Bim in the p38-dependent gene list. We chose to focus on cell death-associated genes whose expression was most differentially regulated by TGFβ (greater than 1.5-fold; Table 1). Significantly, the proapoptotic BH3-only protein, Bmf, emerged as the most strongly upregulated cell death-associated protein in both the Smad4-dependent and p38-dependent gene lists. None of the other previously reported TGFβ-induced apoptosis regulators were found in this analysis to be modulated in the direction expected, a result also confirmed by Western blotting for IκB-α, DAP-kinase, Bid and Bad (Figure 3b) and SHIP-1 (data not shown).

Table 1 Smad4/p38-dependent TGFβ-regulated genes associated with cell death
Figure 3

TGFβ induces expression of the apoptotic proteins, Bmf and Bim, which localize to mitochondria/ER membranes. (a) TGFβ induces transcription of Bmf and Bim in a Smad4- and p38-dependent manner. RT–PCR performed from wild-type or Smad4-deficient AML12 cell total RNA after 0, 4 and 16 h of TGFβ incubation in the presence or absence of the p38 inhibitor (SB202190). (b) TGFβ induces proapoptotic Bmf and Bim protein expression and downregulates the prosurvival Bcl-2 family member, Bcl-XL. Western blot analysis of Bmf, Bim and Bcl-xL after 24 h of TGFβ stimulation, in wild-type and Smad4-deficient AML12 cells, as well as in the presence of the p38 inhibitor (SB202190). The Bmf and Bim antibodies recognize multiple isoforms of the proteins. Other Bcl-2 family members, Bid and Bad, as well as the apoptosis-associated proteins DAP kinase and IκB-α were also assessed. A p38 antibody was used to show equal protein loading. (c) TGFβ-mediated upregulation of Bmf and Bim is an early event. Western blot of time course analysis of Bmf and Bim upregulation after TGFβ addition. (d) Preincubation of cycloheximide inhibits Bmf and Bim upregulation (upper panel) and partially abrogates TGFβ-induced apoptosis (lower panel). The bar graphs are a representative example of the experiment. This experiment was repeated at least three independent times. The data show means+s.d. The numbers shown represent the fold changes over unstimulated cells. (e) Upregulated Bmf and Bim localize to mitochondrial/endoplasmic reticulum (ER) membranes in Western blot analysis of subcellular fractionation of AML12 cells. AML12 cells stimulated for 24 h with TGFβ, before performing subcellular fractionation of cellular membranes. The HDM fraction consists mainly of mitochondrial and ER membranes. Other fractions include the PNS from which nuclei and cell debris have been removed, LDMenriched for small vesicular membranes and the cytosol (Cyt). Equal protein was loaded from each fraction.

To confirm that Bmf and Bim are both TGFβ-induced Smad4- and p38-dependent genes in AML12 cells, we used reverse transcription–polymerase chain reaction (RT–PCR) (Figure 3a) and Western blotting (Figure 3b). TGFβ-mediated upregulation of Bmf and Bim protein and mRNA is greatly reduced in the absence of Smad4. In addition, Bmf induction is very sensitive to inhibition of p38, while Bim induction is partially inhibited. Reduction of both Smad4 and p38 function results in complete loss of Bmf and Bim expression in response to TGFβ. Bmf and Bim proteins were induced rapidly, between 1 and 4 h after addition of TGFβ (Figure 3c), suggesting that they are relatively direct targets of these pathways.

Bmf and Bim are reported to be additionally regulated through their interaction with specific components of the cytoskeleton (Puthalakath et al., 1999, 2001). Thus, we sought to determine the localization of the TGFβ upregulated Bmf and Bim through subcellular fractionation. Purified components of the cytoskeleton revealed enrichment for a large portion of the induced BH3-only proteins (data not shown). However, we also detected significant noncytoskeletal-associated pools of Bmf and Bim in cells. Subcellular fractionation revealed that upregulated Bmf and Bim were also enriched on high-density microsomes (HDM), which consist of endoplasmic reticulum and mitochondrial membranes (Figure 3d). The presence of Bmf and Bim here is consistent with their ability to induce cell death through their sequestration of prosurvival members of the Bcl-2 family away from the apoptotic regulators, Bax and Bak. Indeed, Bmf was able to interact with Bcl-XL in vitro (data not shown). In addition, similar to previous data demonstrating downregulation of Bcl-XL in TGFβ-stimulated cells (Nass et al., 1996; Saltzman et al., 1998; Chipuk et al., 2001; Herrera et al., 2001b), we also detected Bcl-XL as one of the apoptotic genes downregulated by TGFβ in the microarray. Furthermore, Bcl-XL protein levels in both AML12 (Figure 3b) and NMuMG (Figure 3d) cells were correspondingly lower in cells exposed to TGFβ and were also dependent upon both Smad signalling and p38 activity (Figure 3b).

To demonstrate that TGFβ-mediated regulation of the Bcl-2 family was not unique to AML12 cells, a survey of different cell lines demonstrated that similar regulation occurs in other cell systems. Bmf was induced by TGFβ in rat FAO hepatoma cells, the lung CMT170 cell line and the B lymphoma cell line, A20 (Figure S1A), as well as the myeloid cell, M1 (data not shown). Bim was also induced in FAO cells. In addition, examination of NMuMG cells revealed a very similar pattern of regulation of Bmf and Bim to that in AML12 cells, with major inhibition by Smad4 RNAi and at least partial sensitivity to p38 inhibition (Figure 4a). In all cases where the BH3 proteins underwent TGFβ-induction, TGFβ-induced cell death was observed (Figure S1B). To ensure the phenomenon was not specific to cell lines only, TGFβ-sensitive primary rat hepatocytes were also examined. Similar TGFβ-mediated regulation of Bmf and Bim was also observed in primary cells to that seen in the cell lines (Figure S1C).

Figure 4

Bmf and Bim are regulated by the same pathways in NMuMG cells as in AML12 cells, and are capable of causing apoptosis through overexpression. (a) NMuMG cells upregulate Bmf and Bim in response to TGFβ. NMuMG cell lysates from cells exposed to TGFβ and/or the p38 inhibitor (SB202190) for 24 h, probed with antibodies to Bmf, Bim, Bcl-XL and tubulin (left panel). Western blot analysis of Bmf, Bim and Bcl-xL after 24 h of TGFβ stimulation, in wild-type and Smad4-deficient AML12 cells, as well as in the presence of the p38 inhibitor (SB202190; right panel). (b) ROS scavengers prevent Bim and Bmf upregulation. Western blot analysis of AML12 and NMuMG cells pretreated with PDTC and ascorbic acid (Asc) before addition of TGFβ. (c) GFP/Bim or GFP/Bmf co-transfected AML12 cells were subjected to cell death analysis using Annexin V/PI FACS.

Previous reports have suggested a role for reactive oxygen species (ROS), which can be rapidly generated after TGFβ addition, in the induction of apoptosis (Thannickal and Fanburg, 1995; Sanchez et al., 1996; Herrera et al., 2001a). In order to determine if ROS played a role in the regulation of the Bcl-2 family proteins, we used the ROS scavengers pyrrolidinedithiocarbamic acid (PDTC) and ascorbic acid to inhibit ROS generation. Similar to previous reports, we found that TGFβ-mediated Bcl-XL down-regulation was attenuated in the presence of the ROS scavengers in both AML12 and NMuMG cells. More strikingly, TGFβ-mediated Bmf and Bim induction were almost completely prevented in the presence of the ROS scavengers (Figure 4b). This suggested a role for ROS in the TGFβ-mediated regulation of changes in apoptotic gene transcription. To determine whether oxidative stress alone was sufficient to induce changes in the expression of the apoptotic proteins, we treated the cells with exogenous hydrogen peroxide. In both AML12 and NMuMG cells, hydrogen peroxide failed to induce either Bmf or Bim protein levels (data not shown). This suggested that TGFβ-mediated ROS production alone is insufficient to induce changes to Bcl-2 family protein levels, but only in conjunction with other TGFβ-mediated signalling pathways, most likely involving Smad-mediated transcription, can such changes occur.

To address whether overexpression of Bim or Bmf by themselves is sufficient to induce cell death in AML12 cells, a GFP marker plasmid was co-transfected along with Bim or Bmf expression constructs. The level of cell death in the transfected cells was assessed 48 h later using FACS analysis of Annexin V and PI staining (Figure 4c). As expected, both Bim and Bmf, but not an inactive mutant of Bmf, can induce cell death, with Bmf being particularly potent.

In order to determine the importance of the induction of Bmf and Bim in the TGFβ-mediated cell death process, we sought to inhibit their upregulation using siRNA oligonucleotides. In both AML12 and NMuMG cells, Bmf levels were undetectable in the absence of TGFβ and their induction by the cytokine was almost totally blocked by transient transfection of siRNA oligos targeting Bmf (Figure 5a). Bim levels were barely detectable in NMuMG cells in the absence of TGFβ and their induction was completely blocked by siRNA oligos targeting Bim. In AML12 cells, a significant basal level of Bim expression could be detected: 50 nM oligo against Bim reduced the level of Bim following TGFβ treatment to about the basal level in control cells (Figure 5a, right hand lane). Combinations of the two oligos effectively inhibited the induction of the two targets together. It was thus possible to use RNAi to maintain the amounts of both Bmf and Bim in each cell line at the basal level even following prolonged exposure to TGFβ. A second pair of siRNA oligos against different sequences on Bmf and Bim could cause similar inhibition of protein expression (Figure S2A).

Figure 5

Inhibition of Bmf and Bim expression protects cells from TGFβ-mediated apoptosis. (a) siRNA oligonucleotides to Bmf and Bim reduce their TGFβ-induced expression in AML12 and NMuMG cells. Western blot analysis of cells transfected with either GFP (control), Bmf #1, and/or Bim #1 siRNA oligonucleotides 18 h before addition of TGFβ. After 24 h, lysates were prepared and probed with antibodies to Bmf, Bim, tubulin and caspase-3. The 17 kDa fragment of cleaved caspase-3 is shown. (b) Inhibition of Bmf and Bim expression protects cells from TGFβ-induced apoptosis. Cell death analysis of cells transfected with either GFP (control) or Bmf #1 and/or Bim #1 siRNA oligonucleotides before addition of TGFβ. After 48 h cells were subjected to Annexin-V/PI FACS analysis to determine percentage of cell death. The bar graphs are a representative example of the experiment. This experiment was repeated at least three independent times. The data show means+s.d. Numbers shown represent the fold changes over unstimulated cells.

Individually preventing Bmf or Bim upregulation provided only partial protection for AML12 and NMuMG cells, as determined by FACS analysis (Figure 5b, Figure S2B) or by the appearance of the activated, cleaved fragment of caspase-3 (Figure 5a, bottom panel). However, simultaneous use of RNAi to reduce Bmf and Bim to basal levels was sufficient to provide significantly more protection in both cell types (Figure 5b, Figure S2B), with protection being close to complete in NMuMG cells, indicating that Bmf and Bim induction function together during TGFβ-mediated apoptosis. As loss of neither one alone is sufficient to protect cells, increasing levels of Bmf and Bim are likely to have at least partially redundant function in the induction of cell death by TGFβ.


To date it has been unclear exactly which proteins induced on TGFβ treatment of cells lead to activation of the apoptotic programme, in particular the identity of the entry points into the apoptosis signalling pathways. While several proapoptotic proteins have been shown to be induced by TGFβ, intervention to change their levels within the range controlled by TGFβ has generally proved ineffective in providing major protection in different cell types. In some cases where the expression or activity of TGFβ-induced proteins has been targeted, appropriate controls to check for the specificity of any protective effects have not been carried out (Hishikawa et al., 1999; Jang et al., 2002). In other cases, the importance of signalling intermediates varies between cell systems, such as GADD45b-mediated activation of p38 in hepatocytes (Yoo et al., 2003), while p38 signalling is not essential in mammary cells (Figure 1).

From the data provided here, it is likely that the difficulty in finding single mediators of TGFβ-induced apoptosis is due to redundancy in the proapoptotic transcriptional programme. The combination of induction of both Bmf and Bim increases the robustness of the apoptotic response to TGFβ and makes it less likely that susceptible cells can escape TGFβ-induced killing by inactivation of single genes. Downregulation of Bcl-XL may also be important in contributing to cell death in certain systems and may explain the incomplete protection provided by Bmf and Bim inhibition in AML12 cells. Bcl-XL downregulation at the same time that Bmf and Bim are upregulated means that not only do the higher levels of these BH3-only proteins sequester more of the Bcl-XL and other similar prosurvival proteins, but also the overall reduced level of Bcl-XL limits that available for sequestration of Bax and Bak.

Analysis of the promoter regions of Bmf and Bim reveals consensus binding sequences for Smad3/4 complexes, as well as, in the case of Bmf, for the known p38-regulated transcription factor, MEF2C, (data not shown) suggesting that the transcriptional regulation of these genes by TGFβ may be relatively direct. In the case of Bim, ROS upregulation of FoxO Forkhead transcription factors (Liu et al., 2005) could also play a role in Bim induction by TGFβ (Dijkers et al., 2000), possibly with FoxO and Smad3/Smad4 acting coordinately (Seoane et al., 2004). In the case of Bcl-XL, upregulation of the transcriptional repressor Bcl-6 (Table 1), a known suppressor of Bcl-XL transcription, could explain its suppression by TGFβ (Tang et al., 2002).

ROS play an important role in the apoptotic response to TGFβ, being necessary, but not sufficient, for upregulation of Bmf and Bim expression and downregulation of Bcl-XL expression (Figure 3f). TGFβ treatment of hepatocytes has been reported to suppress the expression of a number of key proteins involved in protecting cells from oxidative stress, including manganese-superoxide dismutase, copper/zinc-superoxide dismutase and catalase (Kayanoki et al., 1994). An earlier microarray analysis of TGFβ-regulated transcription in a hepatoma cell line reported downregulation of some nine antioxidant defence proteins, which could lead to elevation of the levels of ROS (Coyle et al., 2003). Many of these genes are components of the Phase II detoxification response, which are regulated by Smad3-ATF3-mediated repression of the ARE antioxidant response element (Kang et al., 2003; Bakin et al., 2005).

The data presented here suggest that no single Bcl-2 family member is the sole critical mediator of the TGFβ-induced cell death response, but instead apoptosis is a result of activation of a transcriptional programme in which a small subset of regulatory proteins, including Bcl-XL, Bmf and Bim, undergo significant expression changes, the combined effect of which is to trigger the cell death programme. In certain cell types, other proapoptotic proteins could also contribute, such as SHIP in haematopoietic cells, although this is not induced in the cell systems used here (Valderrama-Carvajal et al., 2002). The redundancy of the apoptotic programme of gene expression induced by TGFβ is likely to be critical in ensuring that death of sensitive cell types occurs on exposure to TGFβ and minimizing the possibility of escape from this fate. The full significance of the combined induction of Bmf and Bim by TGFβ will require study of the phenotype of mice in which both genes have been deleted; unfortunately the close proximity of these two genes, less than 10 Mb apart on mouse chromosome 2, will hamper the generation of these animals.

Materials and methods


The AML12, NMuMG, CMT170 (clone E9) and A20 cell lines were from ATCC. The FAO cells were a kind gift from Kelvin Cain/Marion MacFarlane (Leicester, UK), and the primary rat hepatocytes were a gift from the Sharon Tooze laboratory (Cancer Research UK). The p38 mitogen-activated protein kinase (MAPK) inhibitor (SB202190) and the caspase inhibitors (Z-VAD-FMK and Boc-D-FMK) were purchased from Calbiochem Inc. (San Diego, CA, USA). Alexa Fluor 647 Annexin V and tetramethyl rhodamine ethyl ester (TMRE) were purchased from Molecular Probes Inc. (Eugene, OR, USA), ascorbic acid, pyrrolidine carbodithioic acid (PDTC), DAPI, PI and cycloheximide were obtained from Sigma Inc. (St Louis, MO, USA). Cycloheximide was used at a final concentration of 0.5 μg/ml. Recombinant human TGFβ was from R&D Systems Inc. (Minneapolis, MN, USA) and was used at a final concentration of 10ng/ml in all experiments.


The Smad4 (B-8), IκB-α (C-21) and Bax (P-19) antibodies were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, Ca, USA). The phospho-MAPKAPK-2, MAPKAPK-2 and p38 polyclonal antibodies were all obtained from Cell Signalling Technology Inc. (Danvers, MA, USA). The Bcl-XL, cytochrome c, Caspase-3 and Bad antibodies were obtained from BD Biosciences (Bedford, MA, USA). Anti-Bim antibodies was obtained from Calbiochem. The antibody recognizing DAP-kinase was from Sigma, and the Bid antibody from R&D systems. The monoclonal hybridoma tubulin (TAT-1) antibody was generated in-house. Finally, the mouse-specific Bmf antibody was a kind gift from David Huang.


To generate stable Smad4 downregulated AML12 cells, the following oligonucleotides were synthesized and high-performance liquid chromatography purified: forward primer 5-IndexTermGATCCCCGGTGGGGAAAGTGAAACCTTTCAAGAG AAGGTTTCACTTTCCCCACCTTTTTGGAAA, and reverse primer: 5-IndexTermAGCTTTTCCAAAAAGGTGGGGAAAGTGA AACCTTCTCTTGAAAGGTTTCACTTTCCCCACCGGG. The forward and reverse oligos were annealed, phosphorylated and cloned into the BglII–HindIII sites of pSuper, using standard techniques. AML12 cells were co-transfected with both the Smad4 RNAi vector and pBabe-puro (Morgenstern and Land, 1990) using Effectene (Qiagen, Valencia, CA, USA) at a ratio of 10:1. After 48 h, cells were selected with puromycin, and colonies were available for picking several weeks later.

For synthetic siRNA oligonucleotides against Smad4, Bim and Bmf, predesigned mouse-specific oligos were obtained from Ambion Inc. (Austin, TX, USA). The oligo sequences were as follows: (control/GFP) 5′-IndexTermGACCCGCGCCGAGGU GAAGtt, (Smad4) 5′-IndexTermGGAUUUCCUCAUGUGAUCUtt, (Bim #1) 5′-IndexTermGGAGGAACCUGAAGAUCUGtt, (Bim #2) 5-IndexTermGACGAGUUCAACGAAACUUtt, (Bmf #1) 5′-IndexTermGGAUU AUUCAAGGACUUUGtt, (Bmf #2) 5-IndexTermGGUCUUCCUUU UCCUUCAAtt. These were transfected into cells using Oligofectamine (Invitrogen, Carlsbad, CA, USA) at a concentration of – 2–100 nM according to the manufacturer's recommendations. For transient Bmf and Bim RNAi, cells were transfected 18 h before addition of TGFβ and cells were analysed for cell death after 48 h. For Smad4 oligonucleotides, cells were transfected 48 h before TGFβ was added, and cell death was quantified 24 h later.

Cell death assays

Cells were stimulated with TGFβ and cell death was quantified 48 h later. At 10 min before harvesting, TMRE was added to the media to label active mitochondria. Cells were processed for levels of cell death using Alexa Fluor 647 Annexin V conjugate and PI or DAPI staining. Dead and dying cells were scored as Annexin V-positive and/or PI-positive on a FACSCalibur or LSR (BD Biosciences). For the caspase inhibitor experiments, 50 μ M Z-VAD-FMK or Boc-D-FMK were preincubated with cells 2 h before addition of TGFβ. The cells were replaced with media containing fresh inhibitor (and TGFβ) after 24 h. In the case of the p38 inhibitor experiments, inhibition of early p38 activity (0–12 h) was achieved by preincubation of the inhibitor 30 min before TGFβ addition, before being washed out and replaced with fresh media containing TGFβ 12 h later. At this 12 h time point, the p38 inhibitor was also added to cells for inhibition of late p38 activity (12–48 h). All cell death experiments were performed in triplicate, and at least three independent times.

Protein expression analysis

For analysis of protein levels, crude cell homogenates were lysed in sample buffer (without bromophenol blue or β-mercaptoethanol) following 24 h of TGFβ stimulation in the presence of dexamethasone. Lysed samples were sonicated briefly, and total protein was quantified using BCA reagent (Pierce, Rockford, IL, USA). Samples were routinely separated on 12% sodium dodecyl sulphate–polyacrylamide gel electrophoresis gels. In the case of ROS scavenger experiments, cells were pretreated with PDTC (50 μ M for AML12; 100 μ M for NMuMG cells) and Ascorbic acid (1 mM for AML12; 2 mM for NMuMG) for 1 h before addition of TGFβ. Cell lysates were prepared 24 h later. All protein expression analysis was carried out at least twice.

Microarray acquisition and analysis

Experiments using Affymetrix GeneChip Mouse Genome MOE 430A Array oligonucleotide arrays were performed following the manufacturers recommendation (http://www.affymetrix.com/support/technical/manual/expression_manual.affx). A complete description of all procedures and statistical analysis is available in the MIAME checklist file in the Supplementary information. Annotation and classification of the genes was performed using the Gene Ontology browser on the Affymetrix website (http://www.affymetrix.com/analysis/index.affx).


Total RNA was extracted from cells using the RNeasy kit (Qiagen), and subjected to RT and PCR using standard protocols. Primers used for PCR were as follows: Bim forward primer 5′-IndexTermATGGCCAAGCAACCTTCTGATGTAAG, Bim reverse primer 5′-IndexTermTCAATGCCTTCTCCATACCAGACGG. For Mouse Bmf, the forward primer was 5′-IndexTermATGCCCG GAGCGGGCGTATTTTG, and the reverse primer was 5′-IndexTermTCACCAGGGCCCCACCCCTTC. For Rat Bmf, the sequences were as follows: the forward primer was 5′-IndexTermGAG ACATGGAGCCACCTCAGTG, and the reverse primer was 5′-IndexTermGTAGCCAGCGTTGCCGTAAAAGAGTC. For Bim PCR, the parameters used were Tm=56°C for 25 cycles, and for Bmf PCR the parameters Tm=61°C, 35 cycles were used. Control glyceraldehyde-3-phosphate dehydrogenase primers and parameters were described previously (Yoo et al., 2003). All RT–PCR experiments were carried out at least two independent times.

Subcellular fractionation of membranes

Membrane fractions were prepared as described previously (Kaufmann et al., 2003). Essentially, cells were grown in 15 cm plates to 80% confluency. TGFβ was incubated with cells for 24 h, before harvesting in lysis buffer (210 mM Mannitol, 70 mM Sucrose, 20 mM Hepes, pH 7.4, 1 mM ethylenediamine-N, N, N′, N′-tetraacetic acid and protease inhibitors). Cells were homogenized in a Dounce homogenizer (20 strokes) and nuclei and cellular debris were removed by centrifugation at 500 g for 5 min to obtain a postnuclear supernatant (PNS). The PNS was centrifuged at 5100 g for 10 min to obtain the mitochondrial/endoplasmic reticulum pellet (HDM). Centrifugation of the post-HDM supernatant at 100 000 g for 60 min in a TLA100.1 yielded the low-density microsomal pellet (LDM) and the cytosolic supernatant (Cyt). 100 μg protein/lane of each fraction were separated on 12% polyacrylamide gels. Protein localization analysis was carried out at least twice.

Accession codes




  1. Arsura M, Wu M, Sonenshein GE . (1996). TGF beta 1 inhibits NF-kappa B/Rel activity inducing apoptosis of B cells: transcriptional activation of I kappa B alpha. Immunity 5: 31–40.

  2. Bakin AV, Stourman NV, Sekhar KR, Rinehart C, Yan X, Meredith MJ et al. (2005). Smad3-ATF3 signaling mediates TGF-beta suppression of genes encoding Phase II detoxifying proteins. Free Radic Biol Med 38: 375–387.

  3. Chipuk JE, Bhat M, Hsing AY, Ma J, Danielpour D . (2001). Bcl-xL blocks transforming growth factor-beta 1-induced apoptosis by inhibiting cytochrome c release and not by directly antagonizing Apaf-1-dependent caspase activation in prostate epithelial cells. J Biol Chem 276: 26614–26621.

  4. Conery AR, Cao Y, Thompson EA, Townsend Jr CM, Ko TC, Luo K . (2004). Akt interacts directly with Smad3 to regulate the sensitivity to TGF-beta induced apoptosis. Nat Cell Biol 6: 366–372.

  5. Coyle B, Freathy C, Gant TW, Roberts RA, Cain K . (2003). Characterization of the transforming growth factor-beta 1-induced apoptotic transcriptome in FaO hepatoma cells. J Biol Chem 278: 5920–5928.

  6. Derynck R, Zhang YE . (2003). Smad-dependent and Smad-independent pathways in TGF-beta family signalling. Nature 425: 577–584.

  7. Dijkers PF, Medema RH, Lammers JW, Koenderman L, Coffer PJ . (2000). Expression of the pro-apoptotic Bcl-2 family member Bim is regulated by the forkhead transcription factor FKHR-L1. Curr Biol 10: 1201–1204.

  8. Edlund S, Bu S, Schuster N, Aspenstrom P, Heuchel R, Heldin NE et al. (2003). Transforming growth factor-beta1 (TGF-beta)-induced apoptosis of prostate cancer cells involves Smad7-dependent activation of p38 by TGF-beta-activated kinase 1 and mitogen-activated protein kinase kinase 3. Mol Biol Cell 14: 529–544.

  9. Francis JM, Heyworth CM, Spooncer E, Pierce A, Dexter TM, Whetton AD . (2000). Transforming growth factor-beta 1 induces apoptosis independently of p53 and selectively reduces expression of Bcl-2 in multipotent hematopoietic cells. J Biol Chem 275: 39137–39145.

  10. Hahn SA, Schutte M, Hoque AT, Moskaluk CA, da Costa LT, Rozenblum E et al. (1996). DPC4, a candidate tumor suppressor gene at human chromosome 18q21.1. Science 271: 350–353.

  11. Herrera B, Alvarez AM, Sanchez A, Fernandez M, Roncero C, Benito M et al. (2001a). Reactive oxygen species (ROS) mediates the mitochondrial-dependent apoptosis induced by transforming growth factor (beta) in fetal hepatocytes. FASEB J 15: 741–751.

  12. Herrera B, Fernandez M, Alvarez AM, Roncero C, Benito M, Gil J et al. (2001b). Activation of caspases occurs downstream from radical oxygen species production, Bcl-xL down-regulation, and early cytochrome C release in apoptosis induced by transforming growth factor beta in rat fetal hepatocytes. Hepatology 34: 548–556.

  13. Hishikawa K, Oemar BS, Tanner FC, Nakaki T, Luscher TF, Fujii T . (1999). Connective tissue growth factor induces apoptosis in human breast cancer cell line MCF-7. J Biol Chem 274: 37461–37466.

  14. Inayat-Hussain SH, Couet C, Cohen GM, Cain K . (1997). Processing/activation of CPP32-like proteases is involved in transforming growth factor beta1-induced apoptosis in rat hepatocytes. Hepatology 25: 1516–1526.

  15. Jang CW, Chen CH, Chen CC, Chen JY, Su YH, Chen RH . (2002). TGF-beta induces apoptosis through Smad-mediated expression of DAP-kinase. Nat Cell Biol 4: 51–58.

  16. Kanamaru C, Yasuda H, Fujita T . (2002). Involvement of Smad proteins in TGF-beta and activin A-induced apoptosis and growth inhibition of liver cells. Hepatol Res 23: 211–219.

  17. Kang Y, Chen CR, Massague J . (2003). A self-enabling TGFbeta response coupled to stress signaling: Smad engages stress response factor ATF3 for Id1 repression in epithelial cells. Mol Cell 11: 915–926.

  18. Kaufmann T, Schlipf S, Sanz J, Neubert K, Stein R, Borner C . (2003). Characterization of the signal that directs Bcl-x(L), but not Bcl-2, to the mitochondrial outer membrane. J Cell Biol 160: 53–64.

  19. Kayanoki Y, Fujii J, Suzuki K, Kawata S, Matsuzawa Y, Taniguchi N . (1994). Suppression of antioxidative enzyme expression by transforming growth factor-beta 1 in rat hepatocytes. J Biol Chem 269: 15488–15492.

  20. Kim BC, Mamura M, Choi KS, Calabretta B, Kim SJ . (2002). Transforming growth factor beta 1 induces apoptosis through cleavage of BAD in a Smad3-dependent mechanism in FaO hepatoma cells. Mol Cell Biol 22: 1369–1378.

  21. Kim SG, Jong HS, Kim TY, Lee JW, Kim NK, Hong SH et al. (2004). Transforming growth factor-beta 1 induces apoptosis through Fas ligand-independent activation of the Fas death pathway in human gastric SNU-620 carcinoma cells. Mol Biol Cell 15: 420–434.

  22. Larisch S, Yi Y, Lotan R, Kerner H, Eimerl S, Tony Parks W et al. (2000). A novel mitochondrial septin-like protein, ARTS, mediates apoptosis dependent on its P-loop motif. Nat Cell Biol 2: 915–921.

  23. Liao JH, Chen JS, Chai MQ, Zhao S, Song JG . (2001). The involvement of p38 MAPK in transforming growth factor beta1-induced apoptosis in murine hepatocytes. Cell Res 11: 89–94.

  24. Liu JW, Chandra D, Rudd MD, Butler AP, Pallotta V, Brown D et al. (2005). Induction of prosurvival molecules by apoptotic stimuli: involvement of FOXO3a and ROS. Oncogene 24: 2020–2031.

  25. Morgenstern JP, Land H . (1990). Advanced mammalian gene transfer: high titre retroviral vectors with multiple drug selection markers and a complementary helper-free packaging cell line. Nucleic Acids Res 18: 3587–3596.

  26. Nass SJ, Li M, Amundadottir LT, Furth PA, Dickson RB . (1996). Role for Bcl-xL in the regulation of apoptosis by EGF and TGF beta 1 in c-myc overexpressing mammary epithelial cells. Biochem Biophys Res Commun 227: 248–256.

  27. Ohgushi M, Kuroki S, Fukamachi H, O'Reilly LA, Kuida K, Strasser A et al. (2005). Transforming growth factor beta-dependent sequential activation of Smad, Bim, and caspase-9 mediates physiological apoptosis in gastric epithelial cells. Mol Cell Biol 25: 10017–10028.

  28. Park HJ, Kim BC, Kim SJ, Choi KS . (2002). Role of MAP kinases and their cross-talk in TGF-beta1-induced apoptosis in FaO rat hepatoma cell line. Hepatology 35: 1360–1371.

  29. Perlman R, Schiemann WP, Brooks MW, Lodish HF, Weinberg RA . (2001). TGF-beta-induced apoptosis is mediated by the adapter protein Daxx that facilitates JNK activation. Nat Cell Biol 3: 708–714.

  30. Puthalakath H, Huang DC, O'Reilly LA, King SM, Strasser A . (1999). The proapoptotic activity of the Bcl-2 family member Bim is regulated by interaction with the dynein motor complex. Mol Cell 3: 287–296.

  31. Puthalakath H, Villunger A, O'Reilly LA, Beaumont JG, Coultas L, Cheney RE et al. (2001). Bmf: a proapoptotic BH3-only protein regulated by interaction with the myosin V actin motor complex, activated by anoikis. Science 293: 1829–1832.

  32. Remy I, Montmarquette A, Michnick SW . (2004). PKB/Akt modulates TGF-beta signaling through a direct interaction with Smad3. Nat Cell Biol 6: 358–365.

  33. Saltzman A, Munro R, Searfoss G, Franks C, Jaye M, Ivashchenko Y . (1998). Transforming growth factor-beta-mediated apoptosis in the Ramos B-lymphoma cell line is accompanied by caspase activation and Bcl-XL downregulation. Exp Cell Res 242: 244–254.

  34. Sanchez A, Alvarez AM, Benito M, Fabregat I . (1996). Apoptosis induced by transforming growth factor-beta in fetal hepatocyte primary cultures: involvement of reactive oxygen intermediates. J Biol Chem 271: 7416–7422.

  35. Schiffer M, Mundel P, Shaw AS, Bottinger EP . (2004). A novel role for the adaptor molecule CD2-associated protein in transforming growth factor-beta-induced apoptosis. J Biol Chem 279: 37004–37012.

  36. Schrantz N, Bourgeade MF, Mouhamad S, Leca G, Sharma S, Vazquez A . (2001). p38-mediated regulation of an Fas-associated death domain protein-independent pathway leading to caspase-8 activation during TGFbeta-induced apoptosis in human Burkitt lymphoma B cells BL41. Mol Biol Cell 12: 3139–3151.

  37. Seoane J, Le HV, Shen L, Anderson SA, Massague J . (2004). Integration of Smad and forkhead pathways in the control of neuroepithelial and glioblastoma cell proliferation. Cell 117: 211–223.

  38. Takekawa M, Tatebayashi K, Itoh F, Adachi M, Imai K, Saito H . (2002). Smad-dependent GADD45beta expression mediates delayed activation of p38 MAP kinase by TGF-beta. EMBO J 21: 6473–6482.

  39. Tang TT, Dowbenko D, Jackson A, Toney L, Lewin DA, Dent AL et al. (2002). The forkhead transcription factor AFX activates apoptosis by induction of the BCL-6 transcriptional repressor. J Biol Chem 277: 14255–14265.

  40. Teramoto T, Kiss A, Thorgeirsson SS . (1998). Induction of p53 and Bax during TGF-beta 1 initiated apoptosis in rat liver epithelial cells. Biochem Biophys Res Commun 251: 56–60.

  41. Thannickal VJ, Fanburg BL . (1995). Activation of an H2O2-generating NADH oxidase in human lung fibroblasts by transforming growth factor beta 1. J Biol Chem 270: 30334–30338.

  42. Valderrama-Carvajal H, Cocolakis E, Lacerte A, Lee EH, Krystal G, Ali S et al. (2002). Activin/TGF-beta induce apoptosis through Smad-dependent expression of the lipid phosphatase SHIP. Nat Cell Biol 4: 963–969.

  43. Wakefield LM, Roberts AB . (2002). TGF-beta signaling: positive and negative effects on tumorigenesis. Curr Opin Genet Dev 12: 22–29.

  44. Wildey GM, Patil S, Howe PH . (2003). Smad3 potentiates transforming growth factor beta (TGFbeta)-induced apoptosis and expression of the BH3-only protein Bim in WEHI 231 B lymphocytes. J Biol Chem 278: 18069–18077.

  45. Yang YC, Piek E, Zavadil J, Liang D, Xie D, Heyer J et al. (2003). Hierarchical model of gene regulation by transforming growth factor beta. Proc Natl Acad Sci USA 100: 10269–10274.

  46. Yoo J, Ghiassi M, Jirmanova L, Balliet AG, Hoffman B, Fornace Jr AJ et al. (2003). Transforming growth factor-beta-induced apoptosis is mediated by Smad-dependent expression of GADD45b through p38 activation. J Biol Chem 278: 43001–43007.

  47. Yu L, Hebert MC, Zhang YE . (2002). TGF-beta receptor-activated p38 MAP kinase mediates Smad-independent TGF-beta responses. EMBO J 21: 3749–3759.

Download references


We would like to thank Yvonne Hey at the Paterson Institute for performing the Affymetrix Microarray analysis. We would also like to thank David Huang, Lorraine O'Reilly and Andreas Strasser for Bmf antibody, as well as members of the Downward laboratory for helpful discussions and advice. This work was funded by Cancer Research UK.

Author information

Correspondence to J Downward.

Additional information

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Ramjaun, A., Tomlinson, S., Eddaoudi, A. et al. Upregulation of two BH3-only proteins, Bmf and Bim, during TGFβ-induced apoptosis. Oncogene 26, 970–981 (2007). https://doi.org/10.1038/sj.onc.1209852

Download citation


  • Smad4
  • TGFβ
  • apoptosis
  • p38
  • Bmf
  • Bim

Further reading