Vorinostat Reverses Paclitaxel Resistance Induced by SIRT6 Upregulation Through Enhancing Autophagy in Lung Cancer

Background Major obstacle to improve the lung squamous cell carcinoma (L-SCC) patients’ prognosis is the acquired chemoresistance. Materials and methods we investigated the clinical signicance of SIRT6, which plays a key role in the resistance to paclitaxel-based chemotherapy. Firstly, we detected the expression and localization of SIRT6, then explored the inhibitory effect of histone deacetylase inhibitor (vorinostat) on paclitaxel-resistant squamous cell carcinoma cell lines, and determined the role of autophagy in reversing the paclitaxel-resistance induced by SIRT6 upregulation.


Introduction
The chemotherapeutic regimens are less for lung squamous cell carcinoma (L-SCC) patients compared with lung adenocarcinoma(L-AD) ones, there are only several options, of which cisplatin plus paclitaxel (TP) regimen is the most common. 1,2 However, taxol-resistance occurred frequently in clinic, and how to carry subsequent treatments after taxol-resistance still is one challenge.
Sirtuins (SIRTs) is one highly conserved family of NAD + -dependent deacetylases. 3 SIRT family members are believed to play distinct roles in tumorigenesis. SIRT6, locating on a chromosomal locus (19p13. 3) that is a frequent breakage site in human acute myeloid leukemia, 4,5 is over-expressed in several cancers, including prostate, 6 endometrioid carcinomas, and keratinocyte-derived skin squamous cell carcinomas. 7 Furthermore, it reported that SIRT6 was down-regulated in pancreatic cancer, 8 hepatocellular carcinoma (HCC) 9,10 and colorectal carcinoma. 11 However, in HCC,over-expressed SIRT6 could inhibit HCC cell growth. In breast cancer, high nuclear expression of SIRT6 might predict poor prognosis. 12,13 SIRT6 also enhanced sensitivity to radiation damage and reduced cell viability in lung cancer. 14 Therefore, the role of SIRT6 in cancers was complicated. In previous study, we found that SIRT6 was highly expressed in L-SCC than L-AD, indicating its association with taxol-resistance.
HDAC inhibitors (HDACi) have been identi ed as potential anticancer agents. Histones comprise the protein core of nucleosomes and their modi cation by acetylation and/or deacetylation affects chromatin structure and regulates gene expression. 15 Drugs that function as HDAC inhibitors are hypothesized to induce differentiation, cause growth arrest and promote apoptosis. It is reported that vorinostat (SAHA) increases carboplatin and paclitaxel activity in non-small-cell lung cancer (NSCLC) cells; 16 and SAHA potentiates paclitaxel-induced apoptosis in ovarian cancer cell lines. 17 In addition, it has demonstrated that SAHA and paclitaxel could cause synergistic effects on apoptosis and microtubule stabilization in papillary serous endometrial cancer cells. 18 In present study, we intended to demonstrated: a) the relationship between SIRT6 and taxol-resistance in L-SCC; b) the molecular mechanism of taxol-resistance induced by SIRT6 upregulation in L-SCC; c) illuminated the anticancer effect of SAHA on taxol-resistance in L-SCC; d) analyzed the role of autophagy in reversing taxol-resistance in the lung cancer.

Materials And Methods
All experimental protocols were approved by our local quality committee, and were carried out in accordance with our institutional guidelines.

Human Samples
Human lung cancer and corresponding non-cancer tissues were collected at the time of resection from 64 patients with NSCLC from May 2017 to June 2018 at the department of thoracic surgery in HuaDong hospital a liated to FuDan University. Tissues were frozen in liquid nitrogen and stored at -80℃ refrigerator immediately. Patients who relapsed within 3 months after the rst course of chemotherapy were de ned as chemoresistance; ≥3 months was sensitive.
The signed informed consent was obtained from all patients and the clinical research ethics committee of FuDan university approved the present study.

Cell Culture And Regents
The NSCLC cell lines (SW900, SW900-Taxol) were purchased from the Cell Resource Center, Shanghai Institute of Biochemistry and Cell Biology at the Chinese Academy of Sciences. SW900-Taxol cell line refers to paclitaxel-resistant SW900 cell line. Both cells were maintained at 37°C in humidi ed air atmosphere containing 5% carbon dioxide in RPMI 1640 supplemented with 10% FBS. Rapamycin (R0395) and chloroquine (C6628) were purchased from Sigma; vorinostat (SAHA) was purchased from Selleck (Houston, TX, USA) and dissolved in dimethylsulfoxide (DMSO) as 10 mmol/L stock solutions.
RNA extraction and quantitative real-time PCR Total RNA was extracted from cultured cells using the Isolation Kit (Ambion; Life Technologies), RNA was extracted from formalin-xed/para n-embed normal and lung cancer specimens using the Recover All Total Nucleic Acid Isolation Kit (Ambion; Life Technologies). cDNA was synthesized from total RNA with speci c stem-loop primers and the TaqMan Reverse Transcription Kit (Applied Biosystems; Life Technologies).
The sequences of the primers were as follows:

Migration And Invasion Assay
For transwell migration assays, 2.5×10 4 to 5×10 4 cells were plated in the top chamber with the noncoated membrane (24-well insert; pore size, 8 um; BD Biosciences). For invasion assays, 1.25×10 5 cells were plated in the top chamber with matrigel coated membrane (24-well insert; pore size, 8 um; BD Biosciences). In both assays, cells were plated in medium without serum or growth factors, and medium which supplemented with growth factors or serum was used as a chemo-attractant in the lower chamber. The cells were incubated for 24 h and cells that did not migrate or invade through the pores were removed by a cotton swab. Cells on the lower surface of the membrane were stained with the Crystal Violet (GuangZhou Xilong Chemical industry co.) and counted.

Immunohistochemistry
The resected tissues were xed with 10% formaldehyde, embedded in para n blocks. All sections were incubated at 63°C for 60 mins, depara nized in xylene, rehydrated, and incubated with fresh 0.3% hydrogen peroxide in 100% methanol for 30 mins at 37°C to block endogenous peroxidase activity. After rehydration through a graded ethanol series, antigen retrieval was carried out in 10 mM citrate buffer (pH6.4) at 98°C-100°C for 20 mins.The sections were passively cooled to 30°C. After rinsing the sections in 0.1M phosphate -buffered saline (pH7.4), non-speci c binding sites were blocked by incubating with 10% normal goat serum for 30 mins. The sections were incubated overnight at 4°C or 37°C for 30 mins with anti-SIRT6 antibodies (CST) at a dilution of 1:100 in PBST. The sections were washed using PBST, incubated with biotinylated anti-mouse IgG, A, M solution (NichireiCo., Tokyo, Japan) for 30 mins at 37°C, and nally incubated in S-ABC solution (Nichirei Co.) for 30 mins. The chromogen, 3,3'-diaminobenzidine tetrahydrochloride, was applied as a 0.02% solution containing 0.005% hydrogen peroxide in a 50 mM ammonium acetate-citrate acid buffer (pH6.0).
The sections were counterstained in Mayer's hematoxylin and mounted on glass slides. The expression levels were de ned as follows:(1) low expression: no staining, weak staining or strong complete cytoplasm/ nuclear staining in <20% of tumor cells; (2) high expression/strong complete cytoplasm/nuclear staining in ≥20% of tumor cells. The expression levels were evaluated by two independent investigators who reached a consensus for all samples.

Microscopy
Cells were plated at low con uence in 12-well plates (5*10 4 cells/well). On the 2 day, cells were exposed to serum starvation (0% FBS), normal medium (10% FBS), Rapamycin (30 umol/L), chloroquine (50 mmol/L) and SAHA for 24 hours. Medium was removed, cells were washed with PBS and treated with 4% paraformalde hyde/PBS for 20 minutes at room temperature, washed, and permeabilized with 0.1% Triton X-100 for 10 mins. Cells were blocked with 5% normal goat serum (CST) containing 0.4% Triton X-100 in PBS for 60 mins. Diluted primary antibody, anti-mouse LC3 A/B (CST), was applied in blocking buffer overnight at 4℃. Alexa Fluor-555 secondary antibodies diluted in 1% normal goat serum in PBS were added for 1 hour at ambient temperature. Cells were xed using Vectashield hard set mounting medium containing DAPI dye (Vector Laboratories). Images were acquired using confocal microscopy (Olympus FV-1000) and overlaid using ImageJ.

Cell Viability Assays
Cells were seeded in 96-well plates (1*10 3 -3*10 3 cells per well depending on the cell type) and incubated overnight before treatment. Cell viability was measured using the CellTiter 96® AQueous MTS Reagent (Promega, Madison, WI). For Giemsa staining, cells were washed with PBS and xed with formaldehyde and washed again before incubation with Giemsa staining solution (Sigma, St Louis, MO). After 30 mins, cells were washed and allowed to dry. Each subgroup has three holes.

Flow cytometry after annexin-V/propidium iodide (PI) staining
The collected cells were centrifuged at 2000 rpm/min for 5 min, and the centrifuged cells were washed twice with PBS. The washed cells were centrifuged and collected, and the treated cells (1*10 6 ) were transferred into an Ependorff (EP) tube and 500 µL of bidding buffer was added into EP tube. PI (5 µL) and Annexin V (5 µL) (KeyGEN BioTECH, Nanjing, China) were added into each EP tube. After the dyes were added, the cells were incubated in the dark for 10-15 mins. The incubated cell suspension was ltered through 400 mesh cell screen to lter off the cell clumps. After ltration, the cells were placed on ice and analyzed within 30 mins using ow cytometer. Annexin V + PI − indicated early apoptosis, Annexin V + PI + indicated late apoptosis. Apoptotic rate = early + late apoptosis rates.

Statistical analysis
Data were shown as mean±SD, the student-t test was used for statistical analysis; all statistical analyses were performed with the SPSS 22.0 software. The Cox regression method was adopted to estimate overall survival (OS) and progression free survival (PFS), and the log-rank test was used to compare survival time. The independent prognostic factors were performed in multivariate analysis using the cox hazard model. P value <0.05 were considered to be signi cant. Graphical displays were prepared using Graph Pad Software (Graph Pad Software, Inc, La Jolla, CA) to show the distributions of expression.

Page 7/23
The upregulation of SIRT6 in lung squamous cell carcinoma The characteristics of patients were shown in Table 1; there was no difference in terms of TNM stage, histology, age, gender and smoking history. Firstly, we detected the expression of SIRT6 in 64 pairs of primary lung cancer and corresponding non-cancer tissues, the results revealed that the SIRT6 was upregulated by more than 2.81-fold in lung cancer compared with non-cancer tissues ( Figure 1A; P<0.001).
Whereas, there was no signi cant relationship between the SIRT6 expression and age, tumor size, location, gender, smoking, and differentiation. Moreover, SIRT6 increased in L-SCC than L-AD ( Figure 1B, and 1C; P<0.01); and advanced stage patients (stage III) had higher SIRT6 expression than early stages (stage I and II) ( Figure 1D, and 1E). In order to explore the relationship between the SIRT6 expression and paclitaxel response, we compared the SIRT6 expression between chemo-sensitive and chemo-resistant patients, and found that the SIRT6 increased in patients who were resistant to chemotherapy (Table1, P<0.001). Analysis also showed that SIRT6 along with TNM stage was independent prognostic factor ( Figure 1F-a, P<0.001) when Cox regression analysis with OS (overall survival) as the end point to identify independent prognostic factors among covariates including SIRT6, histological type, gender, and smoking history. Multivariate analysis with PFS (progression free survival) as the end point revealed that only SIRT6 was an independent prognostic factor ( Figure 1F-b, P<0.001).

SIRT6 promotes paclitaxel resistance via up-regulating TGFβ
We evaluated SIRT6 expression in the cytoplasm and nucleus of cells within the set of NSCLC tissues, and analyzed the correlation between SIRT6 cellular localization with clinicopathological factors. The associations between cytoplasmic SIRT6 expression in the NSCLC specimens and clinicopathological characteristics were shown in Table 2. IHC ( Figure 2) and Immuno uorescence ( Figure 3) showed that there was a correlation between SIRT6 location and TNM stage (P=0.024,). However, no difference was observed for other variables. We also explored the molecular mechanism of SIRT6-induced chemoresistance, it has reported that "drug resistance originating from a TGF-β/FGF-2-driven EMT and its reversion in human lung adenocarcinoma cell lines harboring an EGFR mutation"; therefore, we determined whether TGF-β was the resistant protein that induced paclitaxel resistance through the SIRT6 upregulation? We detected the SIRT6 and TGF-β expression in the SIRT6 transduced and knocked out SW900 cell lines, and found that the TGF-β was consistent with SIRT6 (Supplemental Figure 1), qRT-PCR and western blot also showed that SIRT6 and TGF-β improved in paclitaxel-resistant patients, whereas SIRT6 and TGF-β decreased in non-resistant ones ( Figure 4A and 4C). Simultaneously, we detected the expression of SIRT6 and TGF-β in SW900-Taxol (paclitaxel resistant cell line) compared with SW900 cells. The results showed that SIRT6 and TGF-β were higher in SW900-Taxol than that in SW900 cell line ( Figure 4B and 4D). In summary, we speculated that the paclitaxel resistance of L-SCC induced by up-regulated SIRT6 through increasing TGF-β, further molecular mechanism is still under study.

SAHA inhibits invasion and metastasis of lung squamous carcinoma cells
To study the effect of SAHA on cell proliferation, SW900 and SW900-Taxol were cultured in the presence of various SAHA concentrations. The results revealed that it could inhibit cell growth in a dose-dependent manner. The effective dose of SAHA that inhibited cell growth by 50% at 24 h was 8.6 µM for SW900, 12.0 µM for SW900-Taxol. The effective dose of paclitaxel that inhibited cell growth by 50% at 24 h was 28.5 µM for SW900, 37.2 µM for SW900-Taxol.
We investigated the effect of SAHA on the migration and invasion of lung squamous carcinoma cells. As a result, the inhibitory rates of migration and invasion on SW900 cells and SW900-Taxol cells were shown in Figure 5-I and Table 3, 4. Next, the viabilities of SW900 and SW900-Taxol treated with SAHA or paclitaxel were compared, as a result, combined treatments with SAHA and paclitaxel resulted in more reduction in cell viability compared with SAHA or paclitaxel treatment alone ( Figure 5-II, P<0.001), which indicated that SAHA has re-sensitizing effect on paclitaxel-resistant lung squamous carcinoma cells.  Table 4 The inhibitory rates of migration and invasion on SW900- Taxol  After treating with SAHA, western blot has detected the SIRT6 expression in these cell lines. Treatment with SAHA in SW900-Taxol cells ( Figure 5-III) reduced SIRT6 expression more signi cantly than in SW900 cells ( Figure 5-IV, P<0.001).

SAHA reverses paclitaxel resistance in lung squamous cell carcinoma through enhancing autophagy
In order to investigate the mechanism of re-sensitizing on paclitaxel-resistant cells, we examined the autophagy in SW900-Taxol cells. Compared with control and rapamycin (autophagic inducer) ( Figure 6-I), SAHA induced more autophagosomes than paclitaxel (Figure 6-II, P<0.010). Combined treatments with SAHA and paclitaxel further enhanced the number of autophagosomes compared with SAHA alone, which chloroquine (autophagy inhibitor) could relieve it (Figure 6-III, P<0.010); so it suggested that SAHA could reverse the chemoresistance through enhancing autophagy.
Furthermore, the apoptotic changes in each group detected by the Annexin V-FITC/PI assay were consistent with the changes in apoptosis observed by ow cytometry. As shown in Figure 7-I, higher rates of apoptosis were induced in SW900-Taxol cells by SAHA combined with paclitaxel than SAHA alone (P<0.01).
To further validate the autophagic ux, different autophagy markers were analyzed. As a result, LC3 was the highest when treated with SAHA combined with paclitaxel (P<0.001), whereas, the degradation protein of p62, an adaptor protein which served as an autophagy receptor targeting ubiquitin proteins to autophagosomes for degradation; decreased signi cantly (P<0.001), indicating an enhanced autophagic ux (Figure 7-II).

Discussion
Lung adenocarcinoma patients have more therapeutic choices, such as chemotherapy and epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs); not only the overall survival is appreciable, but also the side effect is less. However, the treatment options for L-SCC patients are limited, some L-SCC patients are intrinsically resistant to chemotherapy. 19 In fact, all patients, including initial responders, rapidly develop chemoresistance, which, in the absence of alternative therapies, leads to 5-year survival rate less than 15%. 20 Therefore, better understanding the molecular determinants is critically important to overcome the chemoresistance responsible for the dismal survival.
Tumor metastasis is characterized by rapidly proliferating and drug resistance. An increasing evidence suggested that the survival of a small portion of cells with stem-like properties may be responsible for tumor recurrences after initial response to chemotherapy. 21  revealed that TGF-β inhibition enhanced chemotherapy action against triple-negative breast cancer; 24 Yhun Yhong Sheen et al. also reported that combinatorial TGF-β attenuation with paclitaxel inhibited the EMT and breast cancer stem-like cells. 25 Based on above studies, we speculated that: 1) the EMT induced by TGF-β was closely associated with tumor metastasis and acquired chemoresistance; 2) blocking TGFβ pathway would enhance the chemotherapeutic effect; 3) combined paclitaxel with TGF-β pathway inhibitors have a synergistic effect. These ndings have provided guidance for the therapy of chemoresistance induced by TGF-β improvement.
However, which gene to regulate the TGF-β pathway function? It is reported that SIRT6 inhibited EMT during idiopathic pulmonary brosis (IPF) via inactivating TGF-β1/Smad3 signaling, 26 and others have discovered that SIRT6 suppressed cell proliferation through inhibiting Twist1 in the NSCLC. 27 Whereas, in present study, we reported that the upregulated SIRT6 induced paclitaxel-resistance via TGF-β pathway, and the improvement of SIRT6 predicted poor prognosis in L-SCC patients. Therefore, we speculated that: 1) SIRT6 might regulate the TGF-β expression; 2) the regulatory role of SIRT6 was in contradiction with previous study. The possible reasons are as follows: 1) the microenvironments in IFP and lung cancer were distinct; resulting in that SIRT6 might play different function; just as has revealed that SIRT6 suppresses pancreatic cancer through controling Lin28b, and SIRT6 was a tumor suppressor that controled cancer metabolism. 8 Although SIRT6 has been shown to be a tumor suppressor in pancreatic cancer; some researchers have demonstrated that the upregulated SIRT6 predicted poor prognosis and promoted the NSCLC metastasis via ERK1/2/MMP9 pathway. 28 It also has reported that SIRT6 was associated with poor prognosis and chemosensitivity in the NSCLC patients, 29 all these ndings indicated that SIRT6 was one oncogene in lung cancer; which were consistent with us, and SIRT6 played diverse epigenetic role on different pathways and metabolic processes, and the obvious divergence has existed in EMT-mediated tumor process and drug resistance.
Next, we intended to determine the therapeutic effect of SAHA alone or combined with paclitaxel on lung cancer. As a result, we found that SAHA has signi cant effect on paclitaxel-resistant L-SCC and resensitized the resistant cells when combined with paclitaxel. However, it has been demonstrated that in vitro, SAHA alone decreased viability and increased apoptosis similarly to paclitaxel. In vivo, paclitaxel followed by SAHA and paclitaxel alone increased survival compared with SAHA alone or SAHA followed by paclitaxel. 30 This suggested adding SAHA to ovarian cancer chemotherapy could increase the e cacy and the agents sequencing was important. In addition, they found that the ef cacy of combination and sequence of SAHA and paclitaxel administration were cell line dependent, suggesting that the responses varied based on tumor speci c characteristics.
In several cancers, it has shown that HDACi lead to autophagy induction, whereas, whether it was protective or toxic for cancer cells has been depending on the types of cancer. 31 Therefore, we examined whether the combined treatments enhanced the autophagy and what was the role of autophagy in the synergistic cytotoxic effect. We demonstrated that the combined treatment depended on autophagy and that inhibition of autophagy resulted in reduced apoptosis and cell death. The interplay between autophagy and apoptosis was complicated, 32 however, it was clear that the autophagy and apoptosis were cross-talk. Under certain conditions, the activation of autophagy inhibited apoptosis, in other condition, autophagy activated it. That was, apoptosis could be induced directly by the treatments or indirectly via autophagy induced by the treatments. Our results showed that the autophagy acted as a cell death mechanism of combined treatment, because the autophagy inhibition brought about a decrease in cell death.
In present study, there were several limitations requiring further exploring: 1) absence of SIRT6 lentivirus interference in cell experiments, so the regulatory mechanism of SIRT6 needed to be validated in-depth; 2) there may be many other mechanism in reversing drug resistance except for autophagy; 3) the number of patients needed to expanded and the follow-up time to be extended.  Figure 1 A. SIRT6 was up-regulated in lung cancer compared with non-tumorous tissues (P<0.001); B, C. showed that SIRT6 increased signi cantly in L-SC than L-AD (P<0.01); D, E. showed that SIRT6 expression increased with TNM stage.
F. showed that SIRT6 was an independent prognostic factor (P<0.001); 1F-a: OS as the end point; 1F-b: PFS as the end point.

Figure 2
IHC showed the SIRT6 location; A. Para-cancer; B. IA stage, High Nuclear expression; C. IIB stage, Cytoplasm and nucleus co-expression; D. III stage, Cytoplasm expression. A and C. qPCR and Western blot showed that SIRT6 and TGF-β improved in paclitaxel-resistant patients, decreased in non-resistance. B and D. SIRT6 and TGF-β were higher in SW900-Taxol cell lines than that in SW900.