Introduction
Patients with breast cancers of specific subtypes are at higher risk for recurrence. Human epidermal growth factor receptor 2 (Her2)
+ breast cancer is a higher risk subtype that constitutes 20–30 % of all breast tumors. Targeted therapies such as Herceptin and Lapatinib have improved recurrence-free survival and helped control metastatic or recurrent disease (as reviewed [
1]). However, response to these therapies is not uniform and resistance, either intrinsic or acquired, remains a significant clinical challenge. Strategies to treat breast cancers that are no longer sensitive to these targeted therapies could translate into improved outcomes for patients.
We initially identified protein tyrosine kinase 6 (PTK6) as a critical mediator of anoikis resistance of breast cancer cells in a functional genomic screen designed to identify regulators of anchorage-independent survival [
2]. PTK6, a member of a distinct family of non-receptor tyrosine kinases distantly related to Src kinases, is expressed in breast cancers and multiple other cancer types [
3‐
7]. We reported that PTK6 transcript expression has prognostic significance; higher levels of PTK6 are associated with adverse outcomes independently of other factors such as nodal status. Among the molecular subtypes of breast cancer, estrogen receptor (ER)
+ and Her2
+ cancers express the highest levels of PTK6 transcript [
2].
PTK6 is a non-receptor tyrosine kinase composed of an amino-terminal SH3 domain, SH2 domain, and carboxyl-terminal kinase domain (as reviewed [
6,
7]). PTK6 promotes oncogenic phenotypes including enhanced proliferation, enhanced anoikis resistance, regulation of autophagy, epithelial-mesenchymal transition, and migration/invasion, via kinase activity-dependent and possibly independent mechanisms [
2,
6‐
11]. There are increasing numbers of PTK6 kinase substrates, including Sam68, Stat3/5b, BKS, Fak, Cbl, and paxillin, many of which are known to play critical roles in oncogenic signaling [
12‐
19]. Unlike the distantly related src kinases, PTK6 lacks a myristylation sequence. Therefore, PTK6 exhibits a broader range of cellular localization that could impact its activities; PTK6 protein has been detected in the nucleus, cytosol, and membranes of cells [
4,
10,
20]. The preferential localization pattern of PTK6 appears to differ between normal vs tumor cells, which could account for differential access to substrates and differential activities in these contexts; while PTK6 is expressed in the nucleus of normal luminal prostate epithelial cells, PTK6 is largely cytosolic in more aggressive prostate cancer cells [
4,
12].
PTK6 impacts survival of both normal and cancer cells, and may seemingly play contradictory roles in these two contexts. In normal intestinal epithelial cells, PTK6 is required for apoptosis induced by DNA damage following UV irradiation [
21]. In contrast, in many tumor model systems PTK6 promotes survival. For example, enhanced PTK6 expression inhibits anoikis and autophagic death following matrix detachment and promotes soft agar colony growth [
2,
9,
17,
22]. Furthermore, downregulation of PTK6 enhances anoikis of breast, ovarian and prostate cancer cells [
2,
17]. PTK6 may also regulate sensitivity to targeted therapeutics. In the studies of Xiang et al., overexpression of PTK6 in ErbB2
+ MCF-10A cells suppressed the growth inhibitory effects of Lapatinib treatment [
23]. However, the precise molecular mechanisms by which PTK6 regulates survival and specifically the apoptotic machinery, of Her2-targeted therapy-resistant cells have not yet been elucidated.
In this study, we sought to determine the effects of PTK6 inhibition on growth and survival of Lapatinib-resistant Her2+ breast cancer cells. We demonstrate that PTK6 downregulation induces apoptosis of these cells by enhancing Bim protein expression. Induction of Bim is critical, as downregulation of Bim expression prevents PTK6 shRNA-induced apoptosis. We also present evidence for p38 activation as a mechanism for PTK6 shRNA-induced Bim induction, and provide the first link between PTK6 and the intrinsic apoptotic pathway.
Methods
Antibodies and reagents
GAPDH, cleaved PARP, phospho-ERK1/2, phospho-AKT (Ser473), phospho-p38 (Thr180/Tyr182), phospho-hsp27 (Ser82), phospho-JNK (Thr183/Tyr185), p-c-Jun (Ser73), p-ATF2 (Thr71), α-tubulin, phospho-Her2 (Tyr1289), phospho-Her2 (Tyr877), Hsp70, Bcl-2, Bcl-xL, Mcl-1, pro-apoptotic Bcl-2 family members (Puma, phospho-Bad (Ser112), Bid), and total p38 antibodies were purchased from Cell Signaling (Danvers, MA, USA). PTK6 (D7), PTK6 (C18), anti-rabbit-hrp, anti-mouse-hrp antibodies, and Protein A/G Plus-agarose (sc-2003) were purchased from Santa Cruz Biotechnology, Inc (Dallas, TX, USA). Bim antibody was purchased from Abcam (Cambridge, MA, USA). Growth factor-reduced MatrigelTM and Z-VAD-FMK were purchased from BD Bioscience (Franklin Lakes, NJ, USA). Lipofectamine 2000 and Plus reagent were purchased from Life Technologies (Grand Island, NY, USA). Lapatinib, SB203580, and SP600125 were purchased from Sellekchem (Houston, TX, USA).
RNAi
PTK6 Mission shRNAs (TRCN0000021549 (49), TRCN0000021552 (C9), TRCN0000196912 (12), TRCN0000199853 (53)) and Bim shRNAs (TRCN0000001051 (51), TRCN0000001054 (54)) were purchased from Sigma Aldrich (St. Louis, MO, USA).
Cell lines
MDA-MB-453, UACC893, SKBR3, and HCC1954 were purchased form ATCC (Manassas, VA, USA). MDA-MB-453 and UACC893 cells were maintained in complete DMEM medium supplemented with 10 % fetal bovine serum and penicillin/streptomycin. The Lapatinib-resistant cell line, UACC893R1, was generated by culturing parental UACC893 cells continuously over 6 months in the presence of increasing concentrations of Lapatinib (up to 5 μM). These cells were then maintained in DMEM complete medium in the presence of 1 μM Lapatinib. SKBR3 and HCC1954 cells were maintained in complete McCoy’s 5A and RPMI medium, respectively, supplemented with 10 % fetal bovine serum and penicillin/streptomycin.
Viral infections
Lentivirus was generated by co-transfecting 293T cells with lentiviral vector, Δ8.9, and pCMV-VSV-G using Lipofectamine 2000 and Plus reagent as described in Irie et al. [
2]. Supernatants were collected and frozen at −80 °C overnight. Retrovirus was generated by transfecting 100-mm plates of 293-GPG cells with retroviral vectors using Lipofectamine 2000 according to established protocols [
24]. Virus was collected, filtered and stored at −80 °C. Cells were infected with virus by spin-infection at 2,250 rpm for 30 minutes at room temperature followed by overnight incubation at 37 °C. Infected cells were selected in the presence of antibiotics (puromycin or G418) purchased from InvivoGen (San Diego, CA, USA).
Quantitative real-time PCR
GAPDH and β-Actin primers were purchased from Qiagen (Venlo, Netherlands). β-2M (Hs00984230_m1), PTK6 (Hs00963386_m1), and Bim (Hs00708019_s1) gene expression assays were purchased from Applied Biosystems (Grand Island, NY, USA). The Bim gene expression assay detects all isoforms of Bim transcript (BimEL, BimL, and BimS). For real-time PCR, RNA was extracted from cell lines using RNeasy kit (Qiagen) and cDNA was synthesized using Taqman cDNA synthesis kit and oligo-dt (16) primers (Life Technologies). Taqman PCR reactions were run using 2 × Taqman universal master mix II (Applied Biosystems, Cat. number 4440040), 5 μl of undiluted cDNA, and 1 μl of PTK6 gene expression assay (protocol used: UNG incubation (50 °C, 2 minutes), polymerase activation (95 °C, 10 minutes), 40 cycles of denaturation (95 °C, 15 sec) and annealing/extension (60 °C, 1 minute)). For other genes, the reactions were performed using 2 × Power SYBR green PCR master mix (Life Technologies), 5 μl of two-fold diluted cDNA, and 2.5 μl of 10 μM primer mix.
Three-dimensional (3-D) cell growth assays
Eight-well chamber slides (BD Biosciences) were coated with 50 μl of growth factor-reduced MatrigelTM; 400 μl of complete growth media containing 4,000 cells were added to each well coated with MatrigelTM. All samples were set up in triplicates. The chamber slides were incubated at 37 °C and re-fed every 3–4 days with complete growth media. Cells were imaged using the Axiovert 25 inverted microscope (Carl Zeiss AB).
Western blots
Protein lysates were prepared using 1 % NP40 Lysis Buffer (Boston Bioproducts, Ashland, MA, USA) and quantified by bicinchoninic acid (BCA) assay. Lysates were resolved using 4–12 % Bis-Tris gradient gels (Life Technologies), transferred at 100 V onto polyvinylidene fluoride (PVDF) membranes using transfer buffer solution (Boston Bioproducts) containing 10 % methanol. The membranes were blocked in 5 % BSA/TBS + 0.05 % Tween solution at room temperature and incubated in primary antibody overnight at 4 °C. Membranes were incubated with secondary antibody (1:2,000) for 1 h at room temperature and were washed in 1 × TBS+ 0.05 % Tween. Blots were developed using ECL (Pierce).
Growth curve analysis
We plated 5 × 104 MDA-MB-453 or UACC893R1 cells infected with either control or PTK6 shRNAs onto 12-well plates in triplicate. The number of live cells was counted every 3–4 days to generate the growth curves. Experiments were performed two or three times with UACC893R1 and MDA-MB-453 cells, respectively. For growth curve analysis in the presence of Lapatinib, 1 × 105 UACC893 and UACC893R cells were plated in triplicate in 24-well plates. Cells were treated with either DMSO or Lapatinib (0.5–5 μM). The number of live cells was counted to generate the growth curves.
Fluorescence-activated cell sorting (FACS)
We detached 5 × 105 cells from the plates using 3mM EDTA/PBS and resuspended in 500 μl of ice-cold PBS. Annexin V-fluorescein isothiocyanate (FITC) staining was performed according to the manufacturer’s protocol (BD Pharmingen, San Diego, CA, USA). For cell cycle profile analysis, cells were fixed using 80 % ethanol, stored overnight at 4 °C, and then spun at 1,500 rpm for 5 minutes at 4 °C using a centrifuge (Eppendorf 5810R). Pellets were washed with cold PBS + 1 % serum, mixed, spun for 5 minutes at 1,200 rpm, and stained with propium iodide (PI)/RNase solution (BD Pharmingen). All analysis was performed using FACS Diva software on a BD FACSCanto II flow cytometer.
Soft agar assays
Base agar was prepared with 0.8 % agarose (Lonza, Basel, Switzerland) at 42 °C. Equal volumes of agarose and 2 × complete growth media were mixed and plated. Top agar (final concentration of 0.4 %) was prepared using a 1:1 mixture of 0.8 % agarose and 2 × growth medium at 37 °C. Cells resuspended in top agar were plated and cultured for 30 days. Cells were re-fed one to two times a week with fresh media.
Tumor xenografts
We resuspended 5 × 105 UACC398R1 cells infected with either control or PTK6 shRNA lentivirus in 100 μl of growth-factor-reduced MatrigelTM on ice. Cell suspensions were injected subcutaneously into the flanks of 6-week-old female nude (nu/nu) mice (Charles River Laboratories). Tumor measurements were performed twice weekly and tumor volume was calculated using the formula: V =1/2 (L × W2). All procedures and studies with mice were performed in accordance with protocols pre-approved by the Institutional Animal Care and Use Committee of Mount Sinai.
Immunoprecipitation (IP)
We lysed 5 × 105 UACC893R1 cells infected with either control or PTK6 shRNA lentivirus in IP lysis buffer (NP-40 lysis buffer (Boston Bioproducts: BP-431), PMSF, leupeptin, aprotinin, NaF, Na3VO4, and phosSTOP (Roche, Basel, Switzerland)) and incubated at 4 °C for 20 minutes. Cell extract was prepared by spinning at full speed for 20 minutes at 4 °C using a centrifuge (Eppendorf Centrifuge 5424R). Supernatant was pre-cleared with 50 % slurry beads (pre-equilibrated with IP lysis buffer) for 30 minutes, incubated with PTK6 antibody (C18) for 2 h, and incubated with 35 μl of 50 % slurry beads for 1 h in a rotating wheel at 4 °C. Beads were washed three times with IP wash buffer (IP lysis buffer without protease inhibitor) and boiled with 2 × SDS sample buffer for 3 minutes at 95 °C. The samples were analyzed by western blotting following the above protocol.
Consent statement
We confirm that this study did not involve human patients and no consent was necessary.
Discussion
The treatment of breast cancers resistant to current standard therapies poses significant clinical challenges. Cancers intrinsically possess or develop mechanisms to evade the death-inducing effects of cytotoxic agents, as well as targeted therapies. Treatment resistance contributes to the development of recurrent or metastatic breast cancers, which are responsible for the majority of deaths due to breast cancer. Therefore, strategies to effectively inhibit the growth and induce death of breast cancer cells resistant to currently available targeted therapies could lead to novel therapeutic options for patients with breast cancer. In this study, we report for the first time that inhibition of PTK6 induces apoptotic cell death of Her2+ breast cancer cells that are relatively resistant to Lapatinib at baseline or after continuous treatment in the presence of this Her2 tyrosine kinase inhibitor (TKI). Apoptosis is induced via enhanced expression of Bim, a BH3-only member of the Bcl2 family via a p38-dependent mechanism.
Our studies show for the first time a link between PTK6, pro-apoptotic Bim and apoptosis of Her2
+ breast cancer cells. Bim, which is expressed as three major isoforms (Bim
EL, Bim
L, and Bim
S), is a regulator of the mitochondrial (intrinsic) apoptotic pathway ([
29] and also reviewed in [
30]). Bim is emerging as a biomarker of sensitivity to targeted therapies, including those that target EGFR family members such as Lapatinib. Bim is frequently downregulated in cancers and lower levels of Bim expression are associated with poorer response to targeted therapy treatment [
31]. Tumors with relatively lower levels of Bim due to a common deletion polymorphism are also more resistant to EGFR tyrosine kinase inhibitors [
32].
Induction of Bim expression tips the functional balance of interacting Bcl2 family members in favor of apoptosis. Bim induction is required for targeted therapy-induced apoptosis of colon, lung, and breast cancers; for example, siRNA-mediated Bim downregulation impaired apoptosis of Her2
+ BT474 cells in response to Lapatinib treatment [
26,
28,
33,
34]. In studies presented in this report, we found that Bim expression is not significantly induced in Her2
+ breast cancer cells that are resistant to Lapatinib, either at baseline or acquired through continuous exposure to Lapatinib, and this lack of induction correlates with lack of apoptosis in response to Lapatinib treatment. Interestingly, we did not observe induction of PUMA, another BH3 only protein implicated in Lapatinib-induced apoptosis, in either the Lapatinib-sensitive or resistant cell lines evaluated in this report (Additional file
1: Figure S1C and [
26]).
Strategies that enhance Bim expression, such as PTK6 inhibition, could therefore be an effective strategy to induce death of TKI-resistant Her2
+ breast cancer cells. Bim is regulated on multiple levels via transcriptional and post-transcriptional mechanisms. Transcription factors such as Foxo3a, NF-κB, c-Myc, CHOP, and AP-1 are known to regulate Bim transcription [
35‐
39]; these are in turn regulated by major survival signaling molecules, such as Erk/MAPK, Akt, and p38 [
39‐
42]. Bim transcript levels are also influenced by epigenetic modifications of the BIM locus and by microRNA-dependent suppression [
43‐
46]. Post-transcriptionally, the stability of Bim protein is regulated by Erk-dependent ubiquitination and proteasome-dependent degradation [
47]. PTK6 inhibition is one approach to induce death by enhancing the expression of Bim protein to levels sufficient to induce apoptosis in Lapatinib-resistant cells. The increased expression of Bim protein is at least partially accounted for by increased levels of Bim transcript in PTK6 shRNA-expressing cells. Future studies will elucidate the specific Bim transcriptional programs regulated by PTK6.
Our studies also show for the first time a link between p38 signaling and PTK6-dependent Bim regulation in Her2
+ breast cancer cells. Following PTK6 downregulation, we did not consistently observe changes in Erk/MAPK or Akt signaling. Rather, p38 MAPK was robustly activated and contributes to the induction of Bim protein expression and apoptosis. The rescue of cells from apoptosis and inhibition of Bim expression by p38 inhibitor treatment is not due to generalized rescue of stress signaling as inhibition of JNK, which is also activated by PTK6 shRNA expression, does not block apoptosis or Bim induction. The pro-apoptotic roles of p38 in the setting of cellular stress are well documented. P38 induces Bim activity and apoptosis via direct phosphorylation at Serine 65 following Sodium arsenite treatment [
48]. In addition, p38 activity led to increased Bim transcription following glucocorticoid treatment of lymphoblastic leukemia cells [
49]. However, in other studies, p38 promotes cellular survival, for example, in response to DNA damage or activation of growth factor receptors (e.g., IGF-R1) due to ionizing radiation [
50,
51]. Our results are also interesting in light of a previous study showing that PTK6 promotes p38 MAPK activation, subsequent Cyclin D1 expression and migration in the context of heregulin- or EGF-stimulated breast cancer cells [
52]. These seemingly conflicting roles of p38 in normal and cancer cell phenotypes have been repeatedly observed. P38 may play a role in pro- or anti-proliferative functions, as well as pro- or anti-apoptotic signaling depending on the cell-type-specific context, the specific stimuli that are used to activate p38, and the intensity or duration of its activation. Future studies are aimed at further dissecting the mechanism by which PTK6 inhibition activates p38 signaling, as well as the mechanisms responsible for p38-mediated regulation of Bim and apoptosis downstream of PTK6.
Recently Ludyga et al. showed that downregulation of PTK6 expression, alone or in combination with Her2 downregulation in Lapatinib and Tratuzumab-resistant, JIMT-1 breast cancer cells inhibited their proliferation without causing cell death [
53,
54]. The lack of apoptosis following PTK6 downregulation contrasts with our findings in two independent Her2
+ cell lines that are resistant to Lapatinib treatment. This may potentially be due to several factors: 1) the relatively high levels of autophagy reported in JIMT-1 cells which may protect cells from apoptosis-inducing stimuli [
55]; 2) the differential expression of Bcl2 family members and other proteins (e.g., high MUC4 expression in JIMT-1 cells) that could modify the threshold for apoptosis induction [
56]; and 3) differences in the genetic background of these cells (e.g., PTEN status) that could modify apoptotic responses [
57,
58]. Nevertheless, it is encouraging that our studies are complementary in demonstrating the efficacy of PTK6 inhibition in inhibiting the growth of Her2 targeted therapy-resistant breast cancer cells, and future studies are aimed at identifying biomarkers associated with cytostatic vs. cytocidal responses to PTK6 downregulation.
The studies presented in our current report support the clinical translation of PTK6 inhibition. There are already several small molecule inhibitors of PTK6 that have been developed and they potently inhibit kinase activity in vitro [
59‐
61]. As these become available, it will be critical to assess whether inhibition of kinase activity phenocopies our results with shRNA expression vectors. The kinase dependency of PTK6-induced oncogenic phenotypes has previously been reported by us and others; in our previous report, we showed that the ability of PTK6 to enhance anoikis resistance when overexpressed in immortalized breast epithelial cells was dependent on PTK6 kinase activity [
2]. Kinase activity of overexpressed PTK6 is also responsible for enhanced cell migration and invasion of MDA-MB-231 triple-negative breast cancer cells [
19]. These kinase-dependent functions are likely due to phosphorylation and/or activation of an increasing number of PTK6 substrate molecules, including Sam68, Stat3/5b, paxillin, BKS/STAP2, p130CAS, AKT, β-catenin, and p190RhoGAP [
12‐
17,
19,
20,
62‐
64]. However, Harvey et al. have also reported that overexpression of a kinase-inactive PTK6 is able to enhance proliferation of T47D breast cancer cells relative to vector control-expressing cells [
9]. As these studies did not include simultaneous evaluation of a kinase-active PTK6, it is difficult to specifically assess the relative contribution of kinase activity to enhanced proliferation. Nevertheless, it is possible that PTK6 is able to regulate proliferation via protein-protein interactions independently of kinase activity. Small molecule inhibitors of PTK6 should prove useful in determining the role of PTK6 kinase activity in Bim and apoptosis regulation of Her2
+ breast cancer cells.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
SHP, KI, WO, GHN, IK and HYI conceived of the design of the studies and carried out the experiments. SHP, KI, WO, GHN, IK and HYI wrote and revised the manuscript. IK carried out bioinformatics analysis of the TCGA data set to identify PTK6 correlated genes. All authors read and approved the final manuscript.