β III-Tubulin Is a Multifunctional Protein Involved in Drug Sensitivity and Tumorigenesis in Non – Small Cell Lung Cancer

Advanced non – small cell lung cancer (NSCLC) has a dismal prognosis. β III-Tubulin, a protein highly expressed in neuronal cells, is strongly associated with drug-refractory and aggressive NSCLC. To date, the role of this protein in in vivo drug resistance and tumorigenesis has not been determined. NSCLC cells stably expressing β III-tubulin short hairpin RNA displayed reduced growth and increased chemotherapy sensitivity when compared with control clones. In concordance with these results, stable suppression of β III-tubulin reduced the incidence and significantly delayed the growth of tumors in mice relative to controls. Our findings indicate that β III-tubulin mediates not only drug sensitivity but also the incidence and progression of lung cancer. β III-Tubulin is a cellular survival factor that, when suppressed, sensitizes cells to chemotherapy via enhanced apoptosis induction and decreased tumorigenesis. Findings establish that upregulation of a neuronal tubulin isotype is a key contributor to tumor progression and drug sensitivity in lung adenocarcinoma.


Introduction
Advanced non-small cell lung cancer (NSCLC) has a dismal prognosis and remains the most common cause of cancer-related death worldwide. Mechanisms mediating resistance and tumor aggressiveness are poorly defined. βIII-Tubulin, a microtubule protein highly expressed in neuronal cells, is strongly associated with drug-refractory and aggressive NSCLC (1). Microtubules are multifunctional cytoskeletal proteins involved in many essential cellular roles, including maintenance of cell shape, intracellular transport, and in mitosis, forming mitotic spindles to ensure proper chromosome segregation and cell division. The soluble α/β-tubulin heterodimers assemble to form the microtubule polymer, with the soluble and polymer forms coexisting in a state of dynamic equilibrium (2). A number of αand β-tubulin isotypes have been identified that display differential developmental and tissue expression and differ in their chromosomal localization (2). It is increasingly apparent that the cellular role of microtubules extends beyond structural support into key signaling and apoptotic roles.
β-Tubulin is the cellular target of clinically important tubulin-binding agents (TBA) used in cancer therapy. High expression of βIII-tubulin was associated with resistance in paclitaxel-selected NSCLC cells and clinical resistance to paclitaxel in ovarian cancer (3). Translational studies have now clearly established that expression of βIII-tubulin is associated with resistance to taxanes or vinorelbine in a range of tumor types, including lung, ovarian, breast, gastric, and cancers of unknown origin (reviewed in refs. 1,4). To date, resistance mediated by βIII-tubulin was thought to be restricted to TBAs. However, we recently identified a broader role for βIII-tubulin in NSCLC (5). βIII-Tubulin can mediate response in vitro not only to broad classes of TBAs but also to DNA-damaging agents. This broad chemosensitization is specific to βIII-tubulin, as silencing of βIIor βIVb-tubulin does not sensitize NSCLC cells to paclitaxel (6). In addition, βIIItubulin under stress conditions has been found to be a mediator of cell survival (7). Indeed, it has been known for some time that tumors expressing βIII-tubulin have a poorer clinical outcome than tumors that have low or no expression of this tubulin isotype (reviewed in refs. 1,8). In lung and a number of other cancers, increased expression of βIIItubulin is associated with poorly differentiated tumor tissue, high-grade malignancy, and metastatic potential (8,9). A recent study found that completely resectable NSCLC tumors, with high expression of βIII-tubulin, correlated with resistance to docetaxel. Collectively, the laboratory and clinical data strongly suggest that βIII-tubulin may have a broader role in the tumorigenesis of certain cancers, such as NSCLC (10). Major issues that have yet to be resolved include the role of βIII-tubulin in mediating in vivo drug sensitivity and whether βIII-tubulin is functionally involved in the tumorigenic phenotype of epithelial cancers.
Herein, we show that βIII-tubulin is mediating in vitro and in vivo drug sensitivity. Importantly, suppression of βIII-tubulin expression reduces anchorage-independent growth and leads to decreased incidence and progression of NSCLC cell xenografts, directly implicating this protein in the tumorigenic phenotype of NSCLC.

Materials and Methods
Generation of βIII-tubulin stable short hairpin RNA-expressing cells The human NSCLC cell line H460 was maintained as previously described (5). For the generation of βIII-tubulin stable knockdown cells, H460 cells were transfected with the pRS vector containing the βIII-tubulin short hairpin RNA (shRNA) expression cassette (pRS/βIII SH ) and the pRS vector containing a noneffective shRNA cassette against green fluorescent protein that acts as a negative control (pRS/Ctrl SH ; OriGene Technologies, Inc.). The 29-mer shRNA sequence that targets βIII-tubulin is as follows: 5′-GTGTGAGCTGCTCCTGT-CTCTGTCTTATT-3′. Cells were selected in growth media containing puromycin (Sigma-Aldrich) to enrich the population of cells that had retained the expression plasmid. Approximately 6 to 60 individual clonal populations were isolated for each construct. After clonal expansion, each clone was examined for βIII-tubulin expression by Western blotting as described below.

Gene expression of βIII-tubulin by real-time PCR
The expression of βIII-tubulin in stably expressing shRNA cells was examined using real-time quantitative PCR. Total RNA was extracted and DNase treated using the Qiagen RNeasy Plus kit according to the manufacturer's instructions. Real-time PCR was done using the QuantiTect SYBR Green PCR kit (Qiagen). βIII-Tubulin mRNA sequences used were as follows: forward, 5′-GCGAGATGTACGAAGACGAC-3′; reverse, 5′-TTTAGACACTGCTGGCTTCG-3′. All data were normalized to the housekeeping gene β 2 -microglobulin (β 2 -Microglobulin QuantiTect Primer Assay, Qiagen).

Assessment of drug sensitivity in vitro
Drug-treated clonogenic assays using CDDP or paclitaxel were performed as described previously (5).

Assessment of drug sensitivity in vivo
BALB/c nude mice (6-8 wk old) were obtained from the Animal Resource Center at the University of New South Wales (Sydney, New South Wales, Australia) and maintained under specific pathogen-free conditions for the studies. All animal experiments were approved by the Animal Ethics Committee, University of New South Wales (ACEC#: 07/89B). Stably expressing βIII-tubulin shRNA (pRS/βIII SH4 or pRS/βIII SH59 ) or control (pRS/Ctrl SH1 and pRS/Ctrl SH2 ) cells (1 × 10 6 ) were inoculated s.c. Once tumors reached approximately 200 to 300 mm 3 , mice were randomized into treatment groups (five mice per treatment group) and treated with 3.33 mg/kg CDDP (one injection once a day for 3 d) or its vehicle control (PBS). Mice were then closely monitored and tumors were measured twice weekly. Tumor volume was calculated using the formula (a × b × c)/2, where a and b are the shorter and longer diameter, respectively, of the tumors and c is the depth. All mice were sacrificed once tumors reached 1,000 mm 3 or if they lost ≥20% body weight. At the time of sacrifice, tumor tissue was collected and placed into 4% paraformaldehyde for histologic analysis or snap frozen in liquid N 2 for biochemical analysis.

Immunohistochemistry
Immunohistochemistry was carried out to measure the expression of βIII-tubulin and Ki67 in paraffin-embedded subcutaneous tumor tissue using a βIII-tubulin monoclonal rabbit antibody (clone TUJ1 1-15-79; Covance) and a mouse monoclonal Ki67 antibody (clone MIB-1; Dako). 3,3′-Diaminobenzidine was used as a substrate for the peroxidase reaction and hematoxylin as the counterstain. The specificity of the primary antibodies was confirmed by including several negative controls: (a) omission of the primary antibody and (b) incubating sections with normal goat IgG (Vector Laboratories) at the same concentration as the primary antibody.

In vitro tumorigenesis assay
Cells stably expressing βIII-tubulin (pRS/βIII SH60 , pRS/ βIII SH59 , and pRS/βIII SH4 ) or control (pRS/Ctrl SH1 and pRS/Ctrl SH2 ) shRNA were resuspended in 0.33% agar in growth media and plated on 0.5% solidified agar (bottom supportive layer). Triplicate plates were set up for each sample. After 12 days of culture, individual colonies were counted and photographed using the Zeiss Axiovert S100 inverted microscope and SPOT digital camera. The results were expressed as the percentage of colony formation, according to the formula: (number of colonies formed/number of cells seeded) × 100%.
In vivo tumorigenesis assay BALB/c nude mice (6-8 wk old) were inoculated s.c. with cells stably expressing βIII-tubulin (pRS/βIII SH4 or pRS/βIII SH59 ) or control (pRS/Ctrl SH1 and pRS/Ctrl SH2 ) shRNA (1 × 10 6 ) cells (10 mice per group). Tumors were measured twice weekly using digital calipers, and tumor volume was calculated using the formula (a × b × c)/2, where a and b are the shorter and longer diameter, respectively, of the tumors and c is the depth. All mice were sacrificed once tumors reached 1,000 mm 3 or if mice had lost ≥20% body weight. At the time of sacrifice, tumors were divided into sections and placed into 4% paraformaldehyde for histologic analysis or snap frozen in liquid N 2 for biochemical analysis.

Statistical analyses
Data are expressed as the mean ± SE and analyzed using ANOVA or Student's t test followed by the nonparametric Dunnett or Mann-Whitney tests using the GraphPad Prism program. Survival curves were plotted by the Kaplan-Meier method and tested for differences with the log-rank statistic. A P value of <0.05 was considered statistically significant.

Generation of stable βIII-tubulin shRNA-expressing NSCLC cells
To allow a long-term study of the possible effects of silencing βIII-tubulin expression on in vivo drug sensitivity and the tumorigenic potential, NSCLC cells were generated, which stably express the pRS/βIII SH construct (refer to Materials and Methods). Control cells were transfected with a nonfunctional 29-mer shRNA. Three independent βIII-tubulinexpressing shRNA clones, designated pRS/βIII SH60 , pRS/ βIII SH59 , and pRS/βIII SH4 , were identified as cells that stably express significantly reduced levels (>80% knockdown) of βIII-tubulin at both the gene transcript and protein level when compared with the control clones pRS/Ctrl SH1 and pRS/Ctrl SH2 (Supplementary Fig. S1A and B). Specific stable knockdown of βIII-tubulin was achieved without compensatory upregulation of other β-tubulin isotypes ( Supplementary  Fig. S2). This is not surprising given that βIII-tubulin makes up ∼7% of the total β-tubulin pool in H460 cells (11).

Inhibition of βIII-tubulin increases sensitivity to cisplatin in NSCLC cells both in vitro and in vivo
Previously, we have shown that transient knockdown of βIII-tubulin using conventional small interfering RNA (siRNA) significantly sensitized NSCLC cells to both TBAs and DNA-damaging agents such as cisplatin (CDDP; ref. 5). To confirm that stable βIII-tubulin shRNA-expressing NSCLC cells behaved in a similar fashion, we assessed their sensitivity to CDDP (a platinum-based agent that is often used as the backbone for systemic chemotherapy in the treatment of NSCLC). All three βIII-tubulin knockdown clones (pRS/βIII SH60 , pRS/βIII SH59 , and pRS/βIII SH4 ) were found to exhibit increased sensitivity to CDDP (Fig. 1A). pRS/βIII SH4 , which has the greatest amount of knockdown, showed the greatest sensitivity to CDDP (mean ID 50 value, 0.17 μmol/L). Similarly, pRS/βIII SH60 and pRS/βIII SH59 were significantly more sensitive to CDDP (mean ID 50 values of 0.22 and 0.24 μmol/L, respectively) compared with control clones (pRS/Ctrl SH1 and pRS/Ctrl SH2 ; mean ID 50 values of 0.41 and 0.38 μmol/L, respectively). Similar results were observed when cells were treated with the TBA paclitaxel or the cisplatin analogue carboplatin (Supplementary Fig.  S3A and B).

Inhibition of βIII-tubulin increases sensitivity to apoptosis in the presence of cisplatin
To understand the basis for the increased drug sensitivity following suppression of βIII-tubulin expression, cell death pathways were investigated. The increase in sensitivity to CDDP was associated with an enhanced induction of apoptosis in the βIII-tubulin shRNA-expressing cells as evidenced by a significant increase in Annexin V (a measure of phosphatidylserine externalization) staining ( Supplementary Fig. S4A). In addition, to explore whether the intrinsic or extrinsic apoptotic pathways were involved, we measured the activity of the initiator caspase-8 and caspase-9 after exposure to CDDP. βIII-Tubulin shRNA-expressing cells and their controls displayed an increase in caspase-9 activity when exposed to 2 and 4 μmol/L CDDP (Supplementary Fig. S4B). However, there was no significant difference in activity between the βIII-tubulin shRNA-expressing and control shRNA-expressing cells (Supplementary Fig. S4B). In contrast, a significant increase in caspase-8 activity was observed in the βIII-tubulin shRNA-expressing cells when compared with controls after exposure to CDDP (Supplementary Fig.  S4C). This increase in activity also correlated with a significant increase in the activation of the effector caspase-3/ caspase-7 and cleavage of its substrate poly(ADP-ribose) polymerase (PARP) in the βIII-tubulin shRNA-expressing cells (Supplementary Fig. S5). Collectively, these data suggest that inhibiting βIII-tubulin enhances drug-induced apoptosis via a caspase-mediated cascade in lung cancer cells.

Knockdown of βIII-tubulin decreases anchorageindependent growth in NSCLC cells
The ability of transformed cells to grow under anchorageindependent conditions is one of the hallmark properties that are associated with the tumorigenic potential of a cancer cell (12). Given that increased levels of βIII-tubulin are associated with more aggressive disease in NSCLC, we then proceeded to assess whether βIII-tubulin plays a role in tumorigenicity. βIII-Tubulin knockdown clones (pRS/βIII SH4 , pRS/βIII SH59 , and pRS/βIII SH60 ) were grown in soft agar, and the number of colonies formed was assessed. Stable knockdown of βIII-tubulin significantly reduced the number of colonies formed in all three βIII-tubulin knockdown clones (pRS/βIII SH4 , mean % colony formation = 15.89; pRS/βIII SH59 , mean % colony formation = 33.39; pRS/βIII SH60 , mean % colony formation = 32.88) when compared with control (pRS/Ctrl SH1 , mean % colony formation = 58.12; pRS/ Ctrl SH2 , mean colony % formation = 59.82; Fig. 3A). Similar results were also observed in cells treated with two different 27-mer dicer siRNA targeting different regions of the βIIItubulin gene as well as conventional siRNA targeting βIIItubulin ( Supplementary Fig. S6A). Notably, the effect of reduced βIII-tubulin on tumorigenicity seemed to be tubulin isotype specific, as no significant effect was observed when cells were treated with siRNA targeting another β-tubulin isotype, βIVb-tubulin ( Supplementary Fig. S6B). This suggests that βIII-tubulin may play a novel and specific role in the tumorigenic potential of NSCLC cells.
Importantly, βIII-tubulin mRNA and protein levels were significantly reduced in βIII-tubulin shRNA-expressing tumors (pRS/βIII SH59 and pRS/βIII SH4 ) when compared with control (pRS/Ctrl SH1 and pRS/Ctrl SH2 ; Fig. 4B and C). The decrease in βIII-tubulin expression in tumor tissue was also confirmed by immunohistochemistry (Fig. 4D and E). Moreover, tumors expressing βIII-tubulin shRNA showed a significant delay in tumor growth (βIII-tubulin shRNA, 67.8 d versus control shRNA, 36.15 d) and size when compared with control tumors ( Fig. 5A and B). Taken together, these results confirm that the activity of βIII-tubulin shRNA is highly potent and active in an in vivo setting. More importantly, these results clearly show for the first time that βIII-tubulin has an important functional role in tumor development and progression in NSCLC.
Discussion βIII-tubulin is predominantly a neuronal-expressed cytoskeletal protein that is associated with drug resistance and aggressive tumors in a range of cancer types, including NSCLC, ovarian, and breast cancers (1,4). Despite strong correlative preclinical and clinical evidence implicating a role for βIII-tubulin in tumorigenesis, its role in tumor formation and aggression has not been addressed (1,13,14). Herein, we have identified and validated a new mechanism of action for βIII-tubulin as a cellular survival factor that, when suppressed, sensitizes cells to chemotherapy via enhanced apoptosis induction and decreased tumorigenesis.
Previous studies have attempted to address the role of βIII-tubulin in drug resistance by overexpressing βIII-tubulin (15,16). However, the results have been difficult to interpret due to increased cell toxicity and significant compensatory changes in the levels of other β-tubulin isotypes. In this study, we overcame a number of these challenges by using a RNA interference approach and examined the effects of stable knockdown of βIII-tubulin on drug sensitivity. Consistent with transient knockdown of βIII-tubulin in two independent NSCLC cell lines (5), stable knockdown of βIII-tubulin resulted in increased in vitro sensitivity to cisplatin, carboplatin, and paclitaxel. Enhanced sensitivity to cisplatin in the βIII-tubulin knockdown cells was correlated with a significant induction of drug-induced apoptosis, as evidenced by an increase in Annexin V staining as well as increased activity of the initiator caspase-8 and the downstream effector caspase-3/caspase-7 and its substrate PARP, suggesting that suppressing βIII-tubulin levels in NSCLC may increase cell death on exposure to chemotherapy by modulating caspase activity. Importantly, confirmation that suppression of βIIItubulin was directly responsible for the enhanced sensitivity to CDDP and not off-target effects of shRNA was obtained when βIII-tubulin levels were "rescued" back into the cells.
Given the strong clinical evidence linking high βIII-tubulin levels and drug resistance, our finding that stable suppression of βIII-tubulin expression increases in vivo sensitivity to CDDP has direct clinical relevance. Taken together with the in vitro data showing increased susceptibility to druginduced apoptosis in βIII-tubulin knockdown cells, βIIItubulin seems to be mechanistically involved as a survival factor, which helps protect cancer cells from cell death by chemotherapy drugs. In support, a recent study in ovarian cancer cells exposed to the stress condition hypoxia showed that βIII-tubulin expression was significantly increased (17) and hypoxia-induced βIII-tubulin expression correlated with paclitaxel resistance (18). In our model, we cannot exclude the possibility that βIII-tubulin may be induced under hypoxic conditions, as the potent stable knockdown would mask any effects. Of note is that a number of signaling proteins involved in regulating drug resistance coimmunoprecipitate or colocalize with βIII-tubulin (17,19). Therefore, it is feasible that silencing βIII-tubulin expression in cancer cells disrupts important protein interactions and signaling processes, which are vital for providing cancer cells with a survival advantage when exposed to cytotoxic stressors.
High levels of βIII-tubulin in clinical samples are correlated strongly with a poorly differentiated and invasive tumor phenotype in a number of epithelial-derived cancers, including NSCLC. βIII-Tubulin expression was positively correlated Figure 5. Knockdown of βIII-tubulin delays tumor growth. A, tumor volume (mm 3 ) of mice injected with stably expressing βIII-tubulin knockdown cells (pRS/βIII SH4 or pRS/βIII SH59 ) when compared with control (pRS/Ctrl SH1 and pRS/Ctrl SH2 ) 7 wk after inoculation. Points, mean (n = 20 animals per group); bars, SE. *, P < 0.05. B, graph showing a significant delay in the time for tumors to reach 1,000 mm 3 in mice injected with stably expressing βIII-tubulin knockdown cells (pRS/βIII SH4 or pRS/βIII SH59 ) when compared with control (pRS/Ctrl SH1 and pRS/Ctrl SH2 ). Columns, mean (n = 20 animals per group); bars, SE. **, P < 0.01. with decreased overall survival in ovarian and NSCLC patients regardless of response to chemotherapy, suggesting that βIII-tubulin may contribute to the aggressive behavior of a tumor rather than only a marker of chemotherapy resistance (1,20). However, despite strong correlative evidence implicating βIII-tubulin in tumorigenesis, its role in tumor incidence and development has been lacking. This study provides the first evidence that βIII-tubulin levels directly influence anchorage-independent growth (a measure of tumorigenic potential). This raises an important question: Is this phenotype specific to suppression of βIII-tubulin or would another β-tubulin isotype mediate a similar effect? The phenotype seems to be specific to βIII-tubulin, as no effect was observed in cells treated with siRNA targeting another β-tubulin isotype, βIVb-tubulin ( Supplementary Fig.  S6B). However, it cannot be ruled out that other β-tubulin isotypes may also affect tumorigenicity. We also addressed the possibility that the βIII-tubulin shRNA was mediating "off-target" effects. However, rescue of βIII-tubulin restored anchorage-independent growth to control levels, again confirming that the observed effect was a direct result of βIII-tubulin knockdown. Strikingly, stable knockdown of βIII-tubulin significantly delayed tumor growth and reduced tumor incidence of subcutaneous xenografted tumors. The reduced growth and incidence of the tumors was not due to reduced cell proliferation, as stable knockdown of βIIItubulin did not affect cell proliferation in vitro (Supplementary Fig. S7), suggesting that other factors associated with βIII-tubulin were at play. In the tumor environment, cancer cells survive under stressful conditions, and our finding that βIII-tubulin is a cellular survival factor for drug-induced cell death may extend to other known cellular stressors, including hypoxia and cytokine exposure (21). Consequently, there may be a balance between cell proliferation and cell death during tumor growth in our model. Alternatively, in the tumor microenvironment, the βIII-tubulin knockdown cells are proliferating at a slower rate. We support the former possibility as, once the tumors reached 1 cm 3 , staining with a proliferation marker (Ki67) did not reveal any difference in tumor cell proliferation in βIII-tubulin knockdown cells ( Supplementary Fig. S8). Further investigation of cell proliferation and cell death parameters is under way to identify the cellular events mediated by βIII-tubulin during the early stages of in vivo tumor growth. The ability of βIII-tubulin knockdown to markedly suppress tumor progression strongly suggests an important role for this protein in lung cancer growth.
Our finding that βIII-tubulin is associated with tumorigenesis is not without precedent for a microtubule-related protein. For example, the microtubule-destabilizing protein stathmin has been linked with tumorigenesis in breast cancer and hepatocellular carcinoma (22)(23)(24). Moreover, a mutation in stathmin identified in esophageal adenocarcinoma was found to play a role in tumorigenesis (25). Despite the fact that βIII-tubulin is present in relatively small amounts in the NSCLC H460 cells (7.8% of total β-tubulin isotypes; ref. 11), knockdown of this minor isotype produces marked effects in drug response and tumorigenesis, suggesting that βIII-tubulin is functionally important in NSCLC cells. Our data strongly support a mechanistic role for βIII-tubulin in tumor cell behavior.
This study has shown that βIII-tubulin is a multifunctional protein that has a key role in the pathobiology and aggressiveness of human lung cancer by influencing drug sensitivity, tumor incidence, and progression. Our findings have direct clinical relevance and raise the possibility that future therapeutic strategies aimed at specifically blocking βIIItubulin activity may have the dual advantage of suppressing lung cancer growth while enhancing the chemosensitivity of the tumor cells.