|Home | About | Journals | Submit | Contact Us | Français|
Protein Inhibitor of Activated Signal Transducer and Activators of Transcription 3 (PIAS3) is an endogenous inhibitor of STAT3 transcriptional activity. We have previously demonstrated the concentration‐dependent negative regulatory effect of PIAS3 on STAT3 signaling and its capacity to decrease lung cancer proliferation and synergize with epidermal growth factor inhibition. We now investigate PIAS3 expression in both non‐small cell lung cancer (NSCLC) cell lines and human resected NSCLC specimens. We also investigated the mechanism by which some lung cancers have significantly decreased PIAS3 expression. Expression of PIAS3 is variable in lung cancer cells lines with 2 of 3 squamous cell carcinoma (SCC) cell lines having no or little PIAS3 protein expression. Similarly, the majority of human SCCs of the lung lack PIAS3 expression by immunohistochemistry; this despite the finding that SCCs have significantly higher levels of PIAS3 mRNA compared to adenocarcinomas. High PIAS3 expression generally correlates with decreased phosphorylated STAT3 in both SCC cell lines and human specimens compatible with the negative regulatory effect of this protein on STAT3 signaling. To investigate this variable expression of PIAS3 we first performed sequencing of the PIAS3 gene that demonstrated single nucleotide polymorphisms but no mutations. Exposure of lung cancer cells to 5‐azacytidine and trichostatin A results in a significant increase in PIAS3 mRNA and protein expression. However, methylation‐specific PCR demonstrates a lack of CpG island methylation in the promoter region of PIAS3. Exposure of cells to an agent blocking proteosomal degradation results in a significant increase in PIAS3. Our data thus shows that SCC of the lung commonly lacks PIAS3 protein expression and that post‐translational modifications may explain this finding in some cases. PIAS3 is a potential therapeutic molecule to target STAT3 pathway in lung cancer.
Lung cancer remains the most common cause of cancer death within the United States (Jemal et al., 2009) and the overall prognosis for this disease, even surgically resected cases, remains dismal (Arriagada et al., 2004). Although recent therapeutic advances in non‐small cell lung cancer (NSCLC) have identified several oncogenic signaling pathways, including the epidermal growth factor receptor (EGFR) (Kumar et al., 2008) and the EML4‐ALK fusion oncogene (Horn and Pao, 2009), the majority of NSCLCs do not carry these well described mutations and little therapeutic advance has been made.
Signal transducers and activators of transcription (STAT) are latent cytoplasmic transcription factors, which become activated and phosphorylated by upstream receptor and non‐receptor tyrosine kinases. With activation, STAT3 molecules dimerize with subsequent nuclear translocation and modulation of gene transcription of its target genes. This pathway is of importance because of its functions in a number of important biological processes including hematopoiesis, immune regulation, and tumorigenesis (Buettner et al., 2002). Out of the seven STATs (1, 2, 3, 4, 5a, 5b and 6), STAT3 is of particular importance due to its involvement in important biological functions including tumor growth and progression. It is activated and phosphorylated by a variety of cytokines and growth factors, such as platelet derived growth factor (PGDF) and epidermal growth factor (EGF) in a large number of human malignancies (Levy and Darnell, 2002). STAT proteins have three families of natural inhibitors: the protein inhibitors of activated STAT (PIAS) (Chung et al., 1997), the suppressors of cytokine signaling (SOCS) (Endo et al., 1997; Naka et al., 1997; Starr et al., 1997), and the Src homology 2 containing phosphatase (SHP) (Shen et al., 1991). All three are known to participate in the negative regulation of the STAT signal transduction pathway (Rakesh and Agrawal, 2005).
PIAS3 belongs to a multi‐gene family which was first identified as a transcriptional repressor of activated STAT3 that inhibits transactivation of a STAT3‐responsive reporter gene and inhibition of the DNA‐binding activity of STAT3 (Chung et al., 1997; Liu et al., 1998). Limited information exists on its relevance and function in solid tumors. PIAS3 is expressed in prostate cancer cells and functions as a transcriptional cofactor for the androgen receptor and its overexpression can induce apoptosis in prostate cancer cells (Ogata et al., 2006).
We have previously demonstrated that EGFR activation leads to PIAS3‐STAT3 binding (Kluge et al., 2009). This PIAS3‐STAT3 complex forms within 5 minutes, then translocates to the nucleus and binds STAT3 consensus sequences (Dabir et al., 2009). PIAS3 regulates STAT3 transcriptional activity in lung cancer cells by affecting its DNA transcriptional properties but also by affecting STAT3 phosphorylation (Dabir et al., 2009; Kluge et al., 2009). The effect of PIAS3 on STAT3 transcriptional activity and phosphorylation is PIAS3 concentration dependent. We have also demonstrated that increasing intracellular PIAS3 levels has growth inhibitory effects on NSCLC cell lines (Kluge et al., 2009). A recent study showed the absence of PIAS3 protein expression in glioblastoma multiforme and restoring PIAS3 results in growth inhibition in U251 cells (Brantley et al., 2008). We thus set to evaluate PIAS3 expression in NSCLC cell lines and resected human lung cancer tissue specimens to determine if mutations or epigenetic changes may explain variability in expression of this protein in lung cancer.
Non‐small cell lung cancer cell lines (adenocarcinoma: A549, H1650, H522, H441, H1975, H1666, H23; squamous cell carcinoma: Calu1, H520, SW900, H1869) were purchased through American Type Culture Collection (Manassas, VA) and maintained in DMEM/Ham's F12 media containing 1% glutamine, 10% fetal bovine serum, and 1% penicillin/ streptomycin in a humidified 5% CO2 environment. NuLi (normal lung epithelial) cells were maintained in Bronchial Epithelial Cell Growth Media (BEGM; Cambrex Corporation, East Rutherford, NJ). BEGM media was supplemented with Hydrocortisone, Bovine Pituitary Extract, Epidermal Growth Factor, Transferrin, Bovine Insulin, Triiodotronine, Epinephrine and Retinoic Acid (all supplied with media), as well as with 50 μg/ml penicillin‐streptomycin, 50 μg/ml gentamycin, and 1.25 μg/ml amphotericin B. To obtain protein lysates, cells were lysed in buffer containing 1% Triton X‐100, 0.15 M sodium chloride, 50 mM Tris, pH 7.4, and protease inhibitors (aprotinin 50 μg/mL, pepstatin 50 μg/mL, leupeptin 10 μg/mL, EDTA 0.4 mM, sodium orthovanadate 0.4 mM, sodium fluoride 10 mM, sodium pyrophosphate 10 mM, phenylmethylsulfonyl fluoride 5 mM). Protein concentrations were determined by the Bradford method (BioRad Protein Assay, Invitrogen, Carlsbad, CA). For electrophoresis, fifteen micrograms of protein from each sample was separated on 10% SDS‐PAGE, transferred to PVDF membrane (Immobilon, Millipore, Bedford, MA), and blotted with specific antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA).
Following electrophoresis, the membranes were blocked with 5% bovine serum albumin in PBS containing 0.1% Tween 20 (PBS‐Tween) for 1 h and then incubated at room temperature with primary antibody (1:200 dilution) for 90 min for the STAT3 antibody and at 4 °C overnight with the other primary antibodies at 1:1000 dilution. After washing in PBS‐Tween, membranes were incubated for 1 h in horseradish peroxidase‐conjugated anti‐immunoglobulin (secondary antibody) (1:5000). Following three washes in PBS‐Tween (5 min each wash), bands were visualized with an enhanced chemiluminescent substrate and subsequent exposure to hyper‐film. The antibody to STAT3 (Santa Cruz Biotechnology, Santa Cruz, CA) is a rabbit polyclonal antibody raised against the recombinant protein mapping to amino acids 50–240. The antibody against phosphorylated STAT3 (p STAT3) (Cell Signaling, Boston, MA) is polyclonal rabbit antibody raised against a peptide corresponding to amino acid sequence containing phosphorylated Tyr‐705 of STAT3 human origin. The antibody to PIAS3 is polyclonal rabbit raised against amino acids 451–619 mapping at the C terminus of human origin. For proteosome inhibition, A549 and SW900 cells were incubated with MG‐132 (10 μmol/l) for different time points, followed by protein collection for PIAS3 western blotting.
RNA was extracted using either the Trizol method (Invitrogen, Carlsbad, CA) or with Qiagen's RNeasy Mini Kit (Valencia, CA) according to their respective protocols. DNA was extracted using the UltraClean Tissue & Cells DNA Isolation Kit (MO BIO Laboratories, Inc., Carlsbad, CA) or Qiagen's DNeasy Kit (Valencia, CA) according to each manufacturer's protocol. DNA and RNA were quantified by spectrophotometry using the Nanodrop 2000.
Surgically resected human NSCLC specimens were obtained from University Hospitals Department of Pathology with IRB approval. Formalin‐fixed paraffin‐embedded tissue was freshly cut (3–4 μm), mounted on pretreated glass slides, dried at 37 °C overnight followed by 60 °C for 30 min. Specimens were rehydrated in xylenes for 5 min (twice), then washed with absolute alcohol and distilled water. Heat Induced Epitope Retrieval(HIER) was performed in a 900 W microwave for 2 × 5 min increments. Antigens were retrieved in a citrate‐based Antigen Unmasking Solution (Vector Laboratories, Inc.). Slides were incubated with anti‐PIAS3 antibody (1:50 dilution, Santa Cruz Biotechnology Inc.) overnight at room temperature in a humidity chamber and detected using R.T.U. VectaStain ABC Kit and DAB Substrate Kit (Vector Laboratories, Inc.). After staining and evaluation of the entire section under low power (4×), a representative area was evaluated with high power (400×). Scoring of the results was performed by an experienced pathologist. The IHC staining was scored as 0 (<5% of tumor cells staining), 1+ (5–19% of tumor cells staining), 2+ (20–50% of tumor cells staining), and 3+ (>50% of tumor cells staining). No staining of stromal cells was seen using anti‐PIAS3 antibodies.
Following extraction, RNA was treated with DNase I according to manufacturer's protocol (Promega). Primers were designed to span the 14 exons of PIAS3 (Table 1). Reverse transcriptase‐PCR was performed using the SuperScript One‐Step RT‐PCR with Platinum Taq system (Invitrogen). PCR products were verified with electrophoresis on a 2% agarose gel. Samples were then purified using Exo‐SAP‐it enzyme according to manufacturer's instructions (USB Corporation, Solon, OH) and submitted to a Genetics Core for sequencing.
Sense and antisense primers for PIAS3 gene sequencing. Seven sets of primers were used. Exon 14, which is the largest exon, required 4 sets of primers.
H1650 cells were grown in DMEM/HF12 media supplemented with 10% fetal bovine serum and 1% penicillin‐streptomycin. Cells were seeded into three T25 flasks at very low density 48 h before treatment. The methodology used is described by (Yamashita et al., 2002) Two flasks were treated with 10 μM 5‐aza‐2′‐deoxycytidine (5AZA, a demethylating agent) (Sigma), diluted from 100 mM stock (dissolved in 50% acetic acid/PBS). The third flask, a mock control, was treated with a like concentration of acetic acid/PBS solution. After 48 h treatment of 5AZA, Trichostatin A (TSA, a histone deacetylating agent) (Sigma) was added to one of the 2 treated flasks at a final concentration of 600 nM. Following a 48 h incubation, cells were collected and protein extracted from all three flasks. Protein was quantified, SDS‐PAGE and transfer to PVDF membrane was performed, and the membrane was stained with anti‐PIAS3 antibody (Santa Cruz Biotechnology, Inc.). Membrane was stripped and reblotted with β‐actin as a loading control. Densitometry was done using Scion Image to quantitate the difference in expression levels of the protein samples.
H1650 cells were grown and treated with 5AZA and TSA as described above, but in triplicate. Following the final 48 h incubation, RNA was extracted from cell slurries according to the Trizol method or with the RNeasy Mini Kit. After spectrophotometric quantification, RNA was purified using DNase I (Promega) according to protocol. Following DNase I treatment, reverse transcription was performed using Invitrogen's Superscript III First Strand for RT‐PCR (Carlsbad, CA) to synthesize cDNA. Real‐time PCR was performed using TaqMan Fast Universal Master Mix and TaqMan probe for PIAS3 (Hs00966025_g1) on the Applied Biosystems 7500 Fast Sequence Detection System according to the manufacturer's instructions (Applied Biosystems, Foster City, CA). Samples were run in triplicate and normalized against β‐actin (Human ACTB Endogenous Control, Applied Biosystems). Relative PIAS3 mRNA amounts in each sample were standardized against the amount of β‐actin RNA and expressed as CT = (CTβ‐actin RNA–CTPIAS3 RNA).
Purified DNA samples were sent to OncoMethylome Science SA (ONCO, Liège, Belgium). ONCO performed real‐time quantitative methylation‐specific PCR. Briefly, DNA was sodium bisulfite‐modified using the EZ DNA Methylation Kit (Zymo Research, Orange, CA) according to the manufacturer's instructions. Analyte (PIAS and ACTB) quantification was performed by real‐time MSP assays. These consisted of parallel amplification/quantification processes using specific primer and primer/detector pairs for each analyte using the Amplifluor® assay format on an ABI Prism® 7900HT instrument (Applied Biosystems). The four PIAS analytes were promoter sequences and detected the fully methylated version. ACTB was used as a reference gene in the assay, using primers which are outside any CpG islands. The following thermal profile was used: Stage1: 50 °C for 2 min; Stage2: 95 °C for 10 min; Stage3: 95 °C for 15 sec; 62 °C for 1 min (= plateau data collection) for 45 cycles. The results were generated using the SDS 2.2 software (Applied Biosystems), exported as Ct values (cycle number at which the amplification curves cross the threshold value, set automatically by the software), and then used to calculate copy numbers based on a linear regression of the values plotted on a standard curve, using plasmid DNA containing the bisulfite modified sequence of interest.
We first evaluated the expression of PIAS3 in 6 different NSCLC cell lines‐ 3 adenocarcinomas and 3 squamous cell carcinomas. As can be seen in Figure 1, expression of PIAS3 is variable. Two of three squamous cell cancers had very low to undetectable PIAS3 levels. These two cell lines (SW900, Calu1) had higher levels of phosphorylated STAT3 (pSTAT3) expression while H520 had high PIAS3 levels and low pSTAT3. In adenocarcinomas this relation was not seen with two of three cell lines showing high PIAS3 and high pSTAT3 levels.
Western blot demonstrates variable expression of PIAS3 in NSCLC cell lines. Phosphorylated STAT3 expression and STAT3 expression were also evaluated. β‐actin was used as a protein loading control.
We evaluated for PIAS3 expression by immunohistochemistry in forty‐four resected NSCLC specimens. Twenty eight specimens were adenocarcinomas and 16 squamous cell carcinomas (Figure 2). In adenocarcinomas, 25/28 (89%) stained positive for PIAS3. By contrast 6/16 (38%) squamous cell carcinomas showed evidence of staining. This difference between adenocarcinoma and squamous cell carcinoma is statistically significant (p = 0.0005). Nine specimens stained 3+ based on the scoring system. Eight of these nine were adenocarcinomas. One 3+ specimen was a large cell carcinoma. All squamous cell carcinomas that did stain for PIAS3 were 1+ except for 3 cases that were 2+.
Immunohistochemistry for PIAS3 in resected lung cancer. Panel A demonstrates a squamous cell carcinoma with no PIAS3 staining and panel B demonstrates an adenocarcinoma with 3+ staining.
We stained the same 44 specimens for pSTAT3 expression. We had previously demonstrated from a larger cohort of resected NSCLC specimens that 38% of NSCLC tumors stain positive for pSTAT3 (Cortas et al., 2007) with a higher degree of staining in adenocarcinomas versus squamous cell cancers (48% vs 27% respectively, p = 0.02). We thus searched our 44 specimen cohort to see if any correlation existed between PIAS3 staining and pSTAT3 staining. All cases that stained for 3+ for PIAS3 were pSTAT3 negative. There was no statistical difference between stages of cancer (Stage I vs Stage II) in terms of positivity for PIAS3 (p = 0.3).
Given the differences in PIAS3 protein expression amongst NSCLC cell lines we then looked to see if any mutations can be found in the PIAS3 gene in NSCLC. Primers were designed to span all fourteen exons of the PIAS3 gene (Table 1), some of which covered exon–exon gaps to eliminate non‐specific amplification. Reverse transcription PCR was performed on RNA extracted from all cell lines, the PCR products were purified and samples were submitted for sequencing. After scanning electropherograms for any heterozygous mutations, alignments were performed with published sequence for the PIAS3 gene. Alignments indicated no mutations within the exons but there were 2 single nucleotide polymorphisms (SNPs) in the untranslated region following exon 14 (Figure 3).
Single nucleotide polymorphisms found in A549 sense sequence, change from C/T or G/T. Verified the SNP changes by sequencing the A549 antisense and confirmed the G/A and C/A changes. These C/T and G/T SNPs were found in other sequenced cell lines, both ...
Given the variable expression of PIAS3 in NSCLC cell lines and the absence of expression in a percentage of surgically resected NSCLCs, we looked at potential epigenetic regulation as a mechanism of PIAS3 down regulation. We exposed H1650 cells to 10 μM 5AZA (demethylating agent) and 600 nM TSA (an HDAC inhibitor). A western blot of proteins treated at this level and stained with anti‐PIAS3 antibody showed a significant increase in protein expression (Figure 4A). For more accurate comparison of PIAS3 expression of the samples, densitometry was done to normalize against β‐actin expression for the differences in protein loading. No significant increase in PIAS3 protein expression was seen with 5AZA; however, a significant increase in PIAS3 protein expression was seen when 5AZA and TSA were combined (8‐fold increase compared to control) (Figure 4B). We also observed an increase in mRNA levels with a combination of 5AZA and TSA suggesting a potential epigenetic mechanism for the regulation of PIAS3 (Figure 4C).
A‐ PIAS3 expression and β‐actin expression in H1650 cells. Cells were treated with 5AZA for 48 h followed by a 24 h treatment with TSA. B‐ Relative expression of PIAS3 standardized to β‐actin ...
We initially looked at DNA from 12 cell lines and constructed 4 MSP assays encompassing a region −1000 to +500 around the transcription start site of PIAS3 (Table 2, Figure 5). Operating efficiencies were deemed acceptable (from 96.5% to 78.5%) and there was good discrimination between in vitro methylated DNA and unmethylated DNA. Beta‐actin was run as an internal control. No positive PIAS3‐MSP signal was obtained from any of the 12 cell lines tested. Copy numbers for β‐actin indicate that bisulfite treatment went correctly. The assays suggest that perhaps we are not targeting the appropriate location of the promoter or that there are modified sites of methylation outside this region. We then focused on regions further upstream and failed to see any sign of methylation (Figure 5).
Localization of PIAS3 qMSP assays on Chromosome 1: Transcription Start Site (TSS) is shown as pink vertical lines, the green box represents the first exon, and the open box represents the PIAS3 promoter region. Open red boxes represent location of PIAS3 ...
DNA from twelve NSCLC cell lines and normal lung epithelial cells (NuLi) was used in four real‐time quantitative MSP assays in the area surrounding the transcription start site of PIAS3 and compared with β‐actin as the internal ...
We evaluated the mRNA level across our panel of 6 lung cancer cell lines. Squamous cell carcinomas in general had significantly higher PIAS3 mRNA levels, confirming our data above that epigenetic changes are unlikely to explain this difference (Figure 6). We therefore exposed SW900 and A549 cells to MG‐132 which blocks protein proteosomal degradation. This exposure led to a significant increase in PIAS3 levels in SW900 cells but not in A549 cells, suggesting that in certain cell lines post‐translational protein degradation may be responsible for the decreased PIAS3 levels (Figure 7).
Relative mRNA expression of adenocarcinoma v. squamous cell carcinoma cell lines.
Post‐translational proteosomal degradation by MG‐132 in adenocarcinoma v. squamous cell carcinoma cell lines.
Given the importance and role of STAT3 in multiple oncogenic pathways, this latent cytoplasmic transcription factor has become a prime target for therapeutic anticancer inhibition (Buettner et al., 2002). We have shown that PIAS3, an endogenous inhibitor of STAT3, can be modulated to decrease STAT3 transcriptional activity and decrease growth of lung cancer cells (Kluge et al., 2009). Furthermore, the effect of PIAS3 on STAT3‐driven growth may be synergistic with EGFR inhibitors. This therapeutic approach may be of substantial value in tumors that lack PIAS3 expression. In a first report of PIAS3 expression, Wang et al. evaluated the expression of PIAS3 by immunohistochemistry in 103 commercially available tumor specimens including 13 NSCLC specimens (1 small cell, 8 squamous cell, 4 adenocarcinoma, 1 large cell carcinoma) and noted increased expression in all 13 specimens (Wang and Banerjee, 2004). This study contains a number of technical issues including the absence of definition of staining intensity, absence of validation of antibody in vitro and the use of commercial specimens. Furthermore, the results of the study of Wang et al. are also in stark contrast with the study of Brantley et al. that convincingly showed a significant loss of PIAS3 expression in glioblastomas (Brantley et al., 2008). In our study, 2/3 of SCC cell lines and human resected SCCs expressed low to no PIAS3. This suggests that restoration of PIAS3 negative regulatory effects on STAT3 may be important as a therapeutic modality for this disease. In addition, all tumors that stained 3+ were adenocarcinomas.
We also observed that adenocarcinomas with 3+ PIAS3 expression lacked phosphorylated STAT3 expression. This is consistent with data from GBM where increase in pSTAT3 was observed in tumor specimens that lacked PIAS3 expression (Brantley et al., 2008). It is also consistent with our data where PIAS3 concentration dependent dephosphorylation of STAT3 has been observed (Dabir et al., 2009).
In order to determine the mechanism by which certain lung cancers lack PIAS3 expression, we first turned our attention to potential mutations within the PIAS3 gene. No mutations were observed. We therefore set to determine if epigenetic silencing of PIAS3 gene may play a role. An extensive search for CpG island methylation did not reveal any significant methylation patterns upstream of the gene. Functional studies showed that 5AZA alone did not increase PIAS3 level; however, the combination of 5AZA and TSA drastically increased PIAS3 concentrations. Although TSA is utilized in the laboratory as a histone deacetylase inhibitor it may have another mechanism by which it can modify protein expression. Similar to glioblastomas we believe that post‐translational modifications, including degradation by the ubiquitin‐proteosome pathway, may play a role in the variable expression of PIAS3 in at least some cell lines. Indeed, it has been shown in glioblastoma cells that treatment with MG‐132, a proteosome inhibitor, leads to increased intracellular PIAS3 levels (Brantley et al., 2008). In our cell lines, A549, which has high PIAS3 protein expression, saw no change in response to MG‐132 while the SCC cell line SW900 showed a significant increase in protein levels within 1 hour of exposure to this agent. This is further confirmed by the fact that SCC has significant PIAS3 mRNA levels. However in a number of cells lines post‐translational modification may not be the underlying pathogenesis for decreased PIAS3. When we exposed H1650 (adenocarcinoma) to 5AZA and TSA a substantial increase in both mRNA transcript and protein was seen. Since no sites of CpG island methylation were seen upstream from the PIAS3 gene in H1650 cells, it is possible that TSA resulted in expression of another gene product which in turn regulated PIAS3.
In conclusion, we demonstrate a lack of PIAS3 expression in a minority of adenocarcinomas of the lung and a majority of squamous cell lung cancers. PIAS3 high expressing tumors have an absence of phosphorylated STAT3 consistent with the negative role of PIAS3 on STAT3 signaling. The lack of PIAS3 in some tumors is not a result of PIAS3 gene mutation or methylation of CpG islands in the promoter region of PIAS3, but most likely a result of increased protein degradation in some cell lines.
We would like to thank OncoMethylome Science SA (ONCO, Liège, Belgium) for performing the real‐time quantitative methylation‐specific PCR.
Kluge Amy, Dabir Snehal, Vlassenbroeck Ilse, Eisenberg Rosana and Dowlati Afshin, (2011), Protein inhibitor of activated STAT3 expression in lung cancer, Molecular Oncology, 5, doi: 10.1016/j.molonc.2011.03.004.
†Supported by grant K23 CA109348 from the National Institutes of Health.