Synergistic killing effect of paclitaxel and honokiol in non-small cell lung cancer cells through paraptosis induction

Abstract

Purpose

Paclitaxel is an anticancer drug for the treatment of non-small cell lung cancer (NSCLC). However, drug-resistance remains a major problem. Honokiol is a natural component which has been found to exhibit anti-tumor activity. Paclitaxel and honokiol have been reported to be able to induce paraptosis. The aim of this study was to investigate whether honokiol can reverse paclitaxel resistance by inducing paraptosis in NSCLC cells.

Methods

NSCLC cell lines H1650 (paclitaxel-sensitive), H1299 and H1650/PTX (intrinsic and acquired paclitaxel-resistant, respectively) were used to assess the cytotoxic effects of paclitaxel and honokiol. Light and transmission electron microscopy were performed to detect cytoplasmic vacuolation. In vitro cell viability and clonogenic survival assays, as well as in vivo xenograft assays were conducted to test synergistic killing effects of paclitaxel and honokiol on NSCLC cells. Western blotting, flow cytometry and immunofluorescence were performed to evaluate paraptosis-regulating mechanisms.

Results

We found that combination treatment with paclitaxel and honokiol synergistically killed H1650, H1299 and H1650/PTX cells by inducing paraptosis, which is characterized by cytoplasmic vacuolation. Moreover, paclitaxel/honokiol treatment resulted in a significant growth delay in H1299 xenograft tumors that showed extensive cytoplasmic vacuolation. Mechanistically, proteasomal inhibition-mediated endoplasmic reticulum (ER) stress and unfolded protein responses leading to ER dilation, and the disruption of intracellular Ca²⁺ homeostasis and mitochondrial Ca²⁺ overload resulting in mitochondrial disfunction, were found to be involved in paclitaxel/honokiol-induced paraptosis. Cellular protein light chain 3 (LC3) may play an important role in paclitaxel/honokiol induced cytoplasmic vacuolation and NSCLC cell death.

Conclusions

Combination of honokiol and paclitaxel may represent a novel strategy for the treatment of paclitaxel-resistant NSCLC.

Electronic supplementary material

The online version of this article (10.1007/s13402-020-00557-x) contains supplementary material, which is available to authorized users.

Introduction

Non-small cell lung cancer (NSCLC) is the most common histological type of lung cancer, with high morbidity and mortality rates [1]. Although advances in the development of target therapies have resulted in improvements in the prognosis of NSCLC patients with EGFR and ALK activating mutations, chemotherapy continues to play an important role in the treatment of this disease [2]. Paclitaxel combined with platinum remains the first-line therapy regimen for advanced NSCLC wild-type EGFR and ALK patients [2–5]. However, the clinical response rate to paclitaxel is still modest, which may be attributed to intrinsic resistance to paclitaxel or acquired resistance developed after repeated exposure to this drug. Several mechanisms of paclitaxel resistance have been identified, including overexpression of the MDR1 efflux pump and β-tubulin mutations, alterations in drug efflux pump activity, deregulation of apoptotic signaling pathways and activation of hypoxia-induced factor1 (HIF-1) signaling [6–10]. During paclitaxel resistance development, cancer cells gradually escape paclitaxel-induced apoptosis leading to relapse and, finally, a poor prognosis [11–13].

Honokiol is a natural biphenolic compound initially isolated from bark of Magnolia plants with anti-oxidative, anti-inflammatory and anti-tumor activities [14, 15]. Honokiol has been found to induce survival inhibition and apoptosis in vitro in various cancer cell lines and to suppress in vivo tumor growth in xenograft models [16, 17]. Honokiol also enhances the sensitivity of different cancer cells to chemotherapeutic drugs [18, 19]. Some studies have shown that honokiol can induce cell death through paraptosis induction when treated as a single drug or combined with imatinib in leukemia cells [20, 21], indicating that paraptosis may be honokiol induced, another mode of programmed cell death. Also, previous studies have shown that high concentrations of paclitaxel may induce paraptotic death in lung cancer cells [22, 23]. Based on this information, we hypothesized that combination of paclitaxel with honokiol might have a synergistic inhibitory effect on NSCLC cells through enhanced paraptosis induction. We found that honokiol indeed potentiates the killing effect of paclitaxel on NSCLC cells in vitro and in vivo, and effectively reverses paclitaxel resistance in NSCLC cells by inducing cytoplasmic vacuolation and paraptosis-like cell death. This effect is independent of apoptosis and autophagy and may be attributed to perturbation of cellular proteostasis, endoplasmic reticulum (ER) stress and mitochondrial disfunction, mediated by proteasome inhibition and mitochondrial Ca²⁺ overload. Microtubule-associated protein 1 light chain 3 (LC3) is known to be involved in paraptosis induction. Therefore, the use of paclitaxel and honokiol combinations to induce paraptosis may provide an attractive strategy to efficiently treat paclitaxel resistant NSCLC.

Materials and methods

Cell culture

The human NSCLC cell lines H1299 and H1650, and a human normal bronchial epithelial cell line BEA5-2B, were purchased from the Shanghai Institute for Biological Science (China). Paclitaxel resistant H1650 cells (H1650/PTX) were established by sequential exposure of the cells to increasing concentrations of paclitaxel as described previously [24]. The cells were maintained in the presence of 400 ng/L docetaxel and grown in drug-free media for five days before the experiments. H1299 and H1650 cells were maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS), L-glulamine, and 100 units/ml penicillin-streptomycin in a humidified incubator containing 5% CO₂ at 37^oC. BEAS-2B cells were maintained in Dulbecco’s Modified Eagle’s medium (DMEM) containing 10% FBS and 100 units/ml penicillin-streptomycin. All cells were passed for 3 months or less before renewal from frozen stocks, and were authenticated based on growth rate, morphology and viability, and were tested for mycoplasma contamination using a MycoAlert Mycoplasma Detection kit (Lona Switzerland).

Reagents and antibodies

Paclitaxel, honokiol, FBS, 3-mefhycaolenine (3-MA), chloroquine (CQ) and cycloheximide (CHX) were purchased from Sigma-Aldrich (St Louis, MO, USA). Z-VAD-fmk, Q-VD-Oph, staurosporine (STS), necrostatin-1, ruthenium red, SP600125, u0126, salubrinal and 4-phenylbutyrafe (4-PBA) were purchased from Calbiochem (San Diego, CA, USA). Antibodies directed against Nrf-1, Mcl-1, Noxa, ATF4, Bip, eIF2α, p-eIF2α, CHOP, XBP1s, Alix, JNK1/2, phospho-JNK1/2, ERK1/2, phospho-ERK1/2, p38, phospho-p38, LC3I/II, p62, β-actin, multi-ubiquitin and Lys 48-specific ubiquitin were purchased from Cell Signaling (Beverly, MA, USA).

Cell viability assay

Cell viability was determined using a cell counting kit-8 (CCK-8) assay before and after treatment with the indicated agents in accordance with the manufacturer’s instructions as described previously [25, 26]. Interactions between agent combinations were analyzed using the CalcuSyn software program (Biosoft, Gambride, UK), which allows combination index (CI) calculations based on the Chou-Talalay’s method. The CIs were determined from the CCK-8 assay as the fraction of cells killed by individual drugs, or combination of the drugs, compared to vehicle-treated cells (control). CI < 1, CI = 1 or CI > 1 indicate synergistic, additive or antagonistic effects, respectively.

Clonogenic assay

Colony formation assays were performed as reported previously [25]. Briefly, cells were seeded at a density of 1000 cells per well into 6-well culture plates and incubated for 24 h. After treatment with the indicated drugs, the cells were allowed to proliferate for 14 days in fresh growth medium. Next, the cells were fixed in 70% ethanol and stained with 0.05% crystal violet (Sigma) for the analysis of clonogenic cell survival. Colonies grown in each cell group were counted and measured using Image J software.

Light microscopy and TEM

To evaluated cell morphological changes caused by paclitaxel and/or honokiol, phase-contrast images were captured using a Nikon DS-File camera. Contrast and brightness were adjusted using Image J software. Transmission electron microscopy (TEM) was conducted as described previously [23]. Briefly, cells (10 × 10⁶ cells/ml) were treated with the indicated drugs during given time periods and subsequently washed in PBS, fixed with Karnovsky’s fixative for 1 h, washed twice with PBS and fixed with O_sO₄. Dehydration was performed using a graded ethanol series and propylene oxide, after which the cells were embedded in Araldit kit (Merck, Rahway, NT, USA). Ultra-thin sections were prepared on a LKB III Ultratome using a diamond knife, after which the sections were mounted on Formvar-coated 200-mesh nicked grids. The grids were double stained for 1 h with saturated uranyl acetate in 50% methanol and rinsed in 100% ethanol and for 17 min in 0.5% lead citrate solution, and rinsed in double deionized water (DDW). All staining procedures were conducted at room temperature. Finally, the sections were examined using a JEOL-100 CX transmission electron microscope, at 80 KV.

In vivo tumor growth inhibition assay

All animal experiments were performed in accordance with guidelines and protocols approved by the Institutional Animal Care and Use Committee of the Jiangsu University. Six-week-old male BALB/c nude mice (Shanghai SLAC Laboratory Animal Co., Ltd., Shanghai, China) were housed in a standard animal laboratory. H1299 cells were washed three times with cold PBS and suspended at a final concentration of 1 × 10⁷/ml in PBS. Next 100 µl cell suspensions were subcutaneously injected into the right flanks of the mice. When the tumor volumes reached approximately 0.1 cm³, the mice were randomly divided into four groups (n = 5 per group), after which treatment was initiated. Honokiol (20 mg/kg) was administrated by oral gavage daily, paclitaxel (10 mg/kg/week) was administrated by intravenous injection (i.v.) once weekly, honokiol (oral gavage) plus paclitaxel (i.v.) were administrated to the combination group, and the control group was treated with normal saline (NS) through oral gavage. The tumor volumes and body weights of the mice were measured every five days and tumor volumes were calculated using the formula: V = shortest diameter² × longest/2. After 30 days, the mice were sacrificed and the tumors were excised, after which H&E staining was carried out to evaluate histologic changes.

Immunofluorescence staining

For the detection of protein aggresome formation, cells were treated with the indicated drugs, fixed and permeabilized using 0.5% Triton-x-100/PBS for 15 min, and blocked in 1% bovine serum albumin for 1 h. Next, the cells were incubated for 2 h with the indicated primary antibodies followed by incubation with a secondary anti-rabbit Alexa Flur 488-conjugated antibody with Hoechst 33,342. Images were captured using an Olympus confocal laser scanning microscope.

For COX-II and PDI detection, cells were fixed and blocked with reagents as mentioned above, and incubated for 2 h with primary antibodies (mouse anti-COXII and rabbit anti-PDI) and secondary anti-rabbit or anti-mouse Alexa Flur 488 or 594 antibodies for 1 h at room temperature and mounted with Prolong Gold antifade mounting reagent. Finally, the slides were evaluated using an Olympus confocal laser scanning microscope.

Western blotting

Before and after treatment with the indicated drugs, cells were lysed and protein expression was determined using standard Western blot analysis as reported before [25, 26]. The antibodies used were as mentioned above.

Apoptosis assay

A Hoechst 33,258 staining kit (Beyotime, China) was used to detect apoptotic features. Cells were seeded into a six-well plate (1 × 10⁵ cells/well), cultured for 24 h and, subsequently, treated with the indicated drugs and stained with Hoechst 33,258. Apoptotic morphological features (chromatin condensation and nuclear fragmentation) were evaluated and imaged using fluorescence microscopy. Sub-G1 population fractions were analyzed using flow cytometry as described previously [26]. The percentage of apoptotic cells was determined using an Annexin V-PE/7-AAD kit (Multscience) according to the manufacturer’s protocol as reported previously [25].

Cell death assay

Cells were cultured in 96 well plates and treated with the indicated drugs during different time intervals. Next, cell death was detected by measuring lactate dehydrogenase (LDH) activity released from the dead cells using a cytotoxicity detection kit (Roch Diagnostics, Mannheim, Germany), according to manufacturer’s instructions.

Small interfering RNA transfection

Before and after treatment with the indicated drugs, transfections of siRNAs into the cells were carried out using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s protocol and as reported before [25, 26]. Western blotting was performed to determine protein expression downregulation.

Intracelluar and mitochodrial Ca²⁺ level measurements

For the measurement of intracellular Ca²⁺ levels, cells were washed with Hank’s balanced salt solution (HBSS without Ca²⁺ or Mg²⁺) after treatment with the indicated drugs, and incubated with Fluo-3 AM (2 µM) for 30 min at 37℃ in the dark. Next, the cells were incubated with HBSS at 37℃ for 15 min before flow cytometry analysis. For the measurement of mitochondrial Ca²⁺ levels, cell were washed with HBSS after treatment with the indicated drugs and incubated with Rhod-2 AM (2.5 µM) for 30 min at 4^oC in the dark. Following a wash with HBSS, the cells were incubated with HBSS at 37^oC for 15 min and collected for flow cytometric analysis.

Statistical analysis

The data are presented as mean ± SD. Student’s t-test and one-way ANOVA were used to calculate significance differences between groups. All values were considered statistically significant when p < 0.05. Statistical analyses were performed using SPSS version 20.0 software (Chicago, IL, USA). * p < 0.05, ** p < 0.005, and *** p < 0.001.

Results

Paclitaxel and honokiol synergistically induce paraptosis in NSCLC cells

Since paclitaxel and honokiol have been reported to be able to induce paraptosis in cancer cells [20–23], we hypothesized that paclitaxel and honokiol might synergistically induce paraptosis in NSCLC cells. Using phase-contrast microscopy, we found that 1 µM paclitaxel alone did not alter the morphology, while 50 µM paclitaxel caused cytoplasmic vacuolation in H1299, H1650 and H1650/PTX cells (Fig. 1a, b and Supplementary Fig. 1A, B). In addition, we found that 20 µM honokiol alone induced smaller cytoplasmic vacuoles, and 40 µM honokiol led to enlarged cytoplasmic vacuoles (Fig. 1c, d and Supplementary Fig. 1C, D). Transmission electron microscopy (TEM) revealed similar morphological features induced by lower doses of paclitaxel or honokiol alone. However, combination treatment with the two drugs led to extensive cytoplasmic vacuoles without distinguished double-layered membranes (Fig. 1e, f and Supplementary Fig. 1E, F). These drug combination-induced cytoplasmic vacuoles could not be prevented by co-treatment with the pan-caspase inhibitor Q-VD-Oph or the necroptosis inhibitor necrostatin-1 (Fig. 1f, g and Supplementary Fig. 1F, G), suggesting that apoptosis and necroptosis are not involved in paclitaxel/honokiol-induced cytoplasmic vacuolation. To determine whether the combination treatment-induced vacuoles result from dilated ER and mitochondria, the ER and mitochondrial structures in the cells were assessed for the presence of specific markers. We found that the ER membrane exhibited calnexin staining, a protein known to be located on the ER membrane (Fig. 1h and Supplementary Fig. 1H). We further performed immunostaining for PDI, an ER-resident protein, and COX-II, a protein known to be localized in the inner mitochondrial membrane. In H1299 and H165/PTX cells treated with paclitaxel and honokiol, PDI expression was detected as a large ring at the cellular periphery, while COX-II expression was observed as a small ring at the perinuclear area (Supplementary Fig. 1H, J). These results indicate that paclitaxel may act synergistically with honokiol to induce cytoplasmic vacuoles in NSCLC cells, which originate from the ER and mitochondria.

Fig. 1.

Combination of honokiol and paclitaxel induces extensive cytoplasmic vacuolation. (a-d) H1299 and H1650 cells were treated with the indicated doses of paclitaxel or honokiol for the given time intervals, and evaluated by phase-contrast microscopy for vacuolation. Bars, 20 µm. At least 300 cells were counted at different time intervals and the percentages of vacuolated cells were calculated. The data represent mean ± S.D. from three independent experiments. ***p < 0.001 vs. 1 µM paclitaxel; **p < 0.005 vs. 20 µM honokiol. (e) H1299 and H1650 cells were treated with the indicated doses of paclitaxel (P) and/or honokiol (H) for 24 h and evaluated by transmission electron microscopy. Bars, 5 µm. Massive cytoplasmic vacuolation is seen. (f, g) H1299 and H1650 cells were treated with the indicated doses of paclitaxel plus honokiol for 24 h in the absence or presence of Q-VD-Oph or necrostatin-1 and evaluated by phase-contrast microscopy. Bars, 20 µm. The percentages of vacuolated cells were calculated as described in (a, b). (h) H1299 cells were treated with the indicated doses of paclitaxel and/or honokiol for 24 h, followed by fixation in 4% formaldehyde and staining with calnexin (green). Nuclei were counterstained with DAPI (blue)

Honokiol enhances NSCLC cell killing effect of paclitaxel in vitro and in vivo

As paclitaxel plus honokiol induce paraptosis, we inferred that honokiol may enhance the killing effect of paclitaxel on NSCLC cells by inducing paraptotic cell death. To test this notion, firstly the anti-cancer activity of paclitaxel and honokiol was tested by determining the viability of three NSCLC cell lines. Using a CCK-8 assay, we found that the 50% inhibitory concentration of paclitaxel (IC₅₀) for H1299 and H1650/PTX cells were 448.6 nM and 751.4 nM, respectively, being nine and fifteen times higher than that for the H1650 cells (49.6 nM) (Supplementary Fig. 2A), whereas honokiol caused similar and dose-dependent inhibitory effects on the three cell lines after 48 h treatment (Supplementary Fig. 2B). Importantly, honokiol dose-dependently increased NSCLC cell death when combined with paclitaxel (Fig. 2a, b and Supplementary Fig. 2C). Particularly in H1299 and H1650/PTX cells that exhibit intrinsic and acquired resistance to paclitaxel, respectively, 1 µM paclitaxel was sufficient to kill 80% of these cells when combined with 20 µM honokiol (Fig. 2a and Supplementary Fig. 2C). Moreover, most combination index (CI) values at different combinations were < 7.0, indicating a strong synergism of paclitaxel and honokiol in the NSCLC cells tested (Supplementary Fig. 2D-F). Likewise, co-treatment with paclitaxel and honokiol overtly resulted in decreased colony formation in the NSCLC cells compared to either drug alone (Supplementary Fig. 2G, H). However, paclitaxel/honokiol failed to synergistically kill BEAS-2B cells and failed to induce cytoplasmic vacuolation in the these cells (Supplementary Fig. S3A, B), suggesting a selective effect of paclitaxel/honokiol on the NSCLC cells. Next, we evaluated the synergistic anti-tumor effect of paclitaxel and honokiol in vivo using a mouse H1299 cell xenograft model. The mice received honokiol alone (20 mg/kg, oral gavage daily), paclitaxel alone (10 mg/kg, i.v. weekly), or honokiol plus paclitaxel until 30 days after the start of the treatment. We found that paclitaxel alone hardly affected the growth of H1299 cell tumors. Honokiol alone had a limited inhibitory effect on tumor growth and paclitaxel/honokiol markedly suppressed tumor growth (Fig. 2c-e) without causing any significant loss of body weight (Fig. 2g). H&E staining revealed that intracellular vacuolation was present in tumor sections obtained from the paclitaxel/honokiol treated mice (Fig. 2f). Together, these findings indicate that co-treatment with honokiol can efficaciously overcome paclitaxel resistance in NSCLC cells via the induction of paraptotic cell death.

Fig. 2.

(a, b) H1299 and H1650 cells were treated with the indicated doses of paclitaxel and/or honokiol for 48 h after which their viability was assessed using a CCK-8 assay. The data represent mean ± S.D. from three independent experiments. ***p < 0.001. (c, d) Six-week-old male nude mice were xenografted with H1299 cells and treated with vehicle (control), paclitaxel (10 mg/kg/week), honokiol (20 mg/kg/day) and honokiol plus paclitaxel as described in Materials and methods. Tumor sizes were measured every 5 days after the start of the treatment, and tumor weights were measured at the end of the experiment. The data represent mean ± S.D. ***p < 0.001 vs. honokiol. (e) Representative pictures of tumors from the mice receiving different treatments as indicated. (f) H&E staining of tumor tissue sections of the mice treated with different agents as indicated. Bars, 20 µm. (g) Body weights of the mice measured every five days during the treatment period

The enhanced killing effect of paclitaxel/honokiol on NSCLC cells is apoptosis independent

Next, we set out to determine whether the enhanced killing effect of honokiol/paclitaxel on NSCLC cells may be attributed to apoptosis induction. Using Western blotting, we found that paclitaxel and/or honokiol only induced minor increases in expression of cleaved-caspase 3 and cleaved-PARP, whereas treatment with STS, a potent apoptosis inducer, resulted in dramatic activation of caspase 3 and PARP (Fig. 3a and Supplementary Fig. 4A). Also, treatment with STS led to significant chromatin condensation and nuclear fragmentation with increased bright blue fluorescence, which are typical features of apoptosis. In contrast, these features were not seen in the NSCLC cells treated with paclitaxel and/or honokiol (Fig. 3b and Supplementary Fig. 4B). Subsequent cell cycle analysis revealed that STS treatment led to striking increases in sub G1 NSCLC cell populations, while paclitaxel and/or honokiol only induced slight or minimal increases in sub G1 populations in these cells (Fig. 3c, d and Supplementary Fig. 4C, D). Likewise, the percentages of apoptotic cells induced by STS were markedly higher than those induced by paclitaxel and/or honokiol alone (Supplementary Fig. 4E and F). We noted that the cleaved-caspase 3 and cleaved-PARP expression levels, the sub G1 cell populations and apoptotic rates in H1650 cells treated with paclitaxel or paclitaxel/honokiol were higher than those in H1299 and H1650/PTX cells receiving the same treatment, suggesting that H1299 and H1650/PTX cells are more resistant to apoptosis induced by paclitaxel. In addition, we found that paclitaxel/honokiol induced cell death was not blocked by Q-VD-Oph, Z-VAD-fmk and necrostatin-1 (Fig. 3e and Supplementary 4g). Collectively, these results indicate that the synergistic inhibitory effects of paclitaxel/honokiol on NSCLC cells may not be attributed to apoptosis.

Fig. 3.

Synergistically paclitaxel and honokiol-induced cell death is not primarily attributed to apoptosis. (a) H1299 and H1650 cells were treated with the indicated doses of honokiol and/or paclitaxel, or staurosporin (STS) for 24 h, after which extracts were prepared and subjected to Western blotting to detect caspase 3, cleaved-caspase 3, PARP and cleaved-PARP expression. (b) Cells treated with drugs as described in (a) and stained with Hoechst 33,258. (c, d) Cells treated with drugs as described in (a) for 48 h subjected to cytometry to determine sub G1 cell fractions as apoptotic cells. Representative cell cycle analysis images are shown (c); quantitative analysis of the sub G1 population of cells treated with the indicated drugs (d). **p < 0.005 vs. paclitaxel plus honokiol. (e) Cells were treated with paclitaxel and/or honokiol for 48 h in the absence or presence of Q-VD-Oph, Z-VAD-fmk or necroptatin-1, after which the percentages of cell death were determined using a LDH activity assay. The data represent mean ± S.D. from three independent experiments

Proteasomal dysfunction and ER stress contribute to paclitaxel/honokiol-induced paraptosis

Based on previous reports that cytoplasmic vacuolation may result from dilated ER, and that proteasome inhibition and unfolded protein response (UPR) may be involved in paraptosis [27, 28], we next set out to test the expressions of ER stress markers and ubiquitinated proteins in the NSCLC cells. We found that, compared to paclitaxel or honokiol alone, the combination treatment notably upregulated the expression of polyubiquitinated proteins and the proteasome-substrate proteins Nrf 1, Mcl-1 and Noxa in a time-dependent manner (Fig. 4a and Supplementary Fig. 5A, B). In addition, we noted increased phosphorylated eIF2α levels and increased expression levels of CHOP, ATF4 and Bip, as well as increased splicing of X-box binding protein1s (XBP1s) and decreased expression of Alix in the NSCLC cells (Fig. 4b and Supplementary Fig. 5C). The effect of paclitaxel/honokiol on these proteins was, however, not observed in BEAS-2B cells (Supplementary Fig. 3C). Given that protein synthesis is required for paraptosis [27], we assessed the impact of protein synthesis inhibition on this process, using CHX as an inhibitor. We found that pretreatment with CHX prevented the paclitaxel/honokiol-induced upregulation of polyubiquitinated proteins, proteasome-substrate proteins and ER stress marker proteins, and blocked the reduction of the Alix protein (Fig. 4a, b and Supplementary 5A and C). In addition, we found that pretreatment with the ER stress inhibitors salubrinal and 4-PBA impeded CHOP and ATF4 upregulation induced by paclitaxel/honokiol (Fig. 4c). Furthermore, paclitaxel/honokiol-induced cytoplasmic vacuolation and growth inhibition were found to be retarded by pretreatment with CHX or salubrinal (Fig. 4d-f and Supplementary Fig. 5D, E), underscoring that protein synthesis and ER stress are necessary for paraptotic cell death induced by paclitaxel/honokiol.

Fig. 4.

Combination of paclitaxel and honokiol induces biochemical features of paraptosis in NSCLC cells. (a, b) H1299 and H1650 cells were pretreated with or without CHX for 2 h, and next treated with the indicated doses of paclitaxel and/or honokiol for 24 h. Subsequently, cell extracts were prepared for Western blotting of the indicated proteins; β-actin was used as a loading control. (c) H1299 and H1650 cells were pretreated with the indicated doses of salubrinal or 4-PBA for 4 h, and next treated with paclitaxel (P) plus honokiol (H) for 24 h. Subsequently, cell extracts were prepared for Western blotting of ATF4 and CHOP; β-actin was used as a loading control. (d) Cells were pretreated with or without CHX (2 µM, 2 h), or salubrinal (20 µM, 4 h), and next treated with paclitaxel plus honokiol for 24 h, after which phase-contrast images were obtained. Bars, 20 µm. (e, f) Cells were treated with the indicated doses of paclitaxel and/or honokiol for 48 h in the absence or presence of CHX (2 µM) or salubrinal (20 µM), after which cell viability was determined using a CCK-8 assay. ***p < 0.001. The data represent mean ± S.D. from three independent-experiments. (g) Cells were treated with the indicated doses of paclitaxel and/or honokiol for 24 h, after which cell extracts were prepared and subjected to Western blotting to detect proteins as indicated. (h) Cells were pretreated with U0126 (10 µM) or SP600125 (15 µM) for 1 h, and next treated with paclitaxel (1 µM) plus honokiol (20 µM) for 48 h, after which cell death was determined using a LDH activity assay. **p < 0.005 vs. paclitaxel plus honokiol.

In addition, we assessed the role of paclitaxel/honokiol on the MAPK pathway, which is considered to be involved in the modulation of paraptosis [27, 28]. We found that paclitaxel/honokiol clearly increased the expression of phosphorylated JNK1/2 and ERK1/2 in a time-dependent manner, but did not augment the expression of p38 (Fig. 4g, h and Supplementary Fig. 5F, G). The JNK and ERK inhibitors U0126 and SP600125 blocked paclitaxe/honokiol-induced cell death, indicating that JNK and ERK may positively regulate paclitaxel/honokiol-induced cell death.

UBE1 is essential for ubiquitinated protein accumulation and paraptosis induced by paclitaxel/honokiol

Since ubiquitinated protein aggregates often cause cytoplasmic vacuolation [29–31], we next evaluated the role of ubiquitin-activating enzyme 1 (UBE1) on paclitaxel/honokiol-induced cytoplasmic vacuolation and cell death. In the three NSCLC cell lines treated with paclitaxel/honokiol and next immunostained with antibodies specific for multi-ubiquitin and K48-linked ubiquitin, fluorescence microscopy revealed distinct punctate formations of ubiquitinated protein aggregates (Fig. 5a, b and Supplementary Fig. 6A). Subsequent knockdown of UBE1 (Fig. 5f) attenuated the paclitaxel/honokiol-induced accumulation of ubiquitinated protein aggregates (Fig. 5c and Supplementary Fig. 6B), which is consistent with the repression of cytoplasmic vacuolation by UBE1 silencing (Fig. 5d and Supplementary Fig. 6C). Given that the upregulation of CHOP induced by paclitaxel/honokiol may contribute to cytoplasmic vacuolation and paraptotic cell death, we next examined whether CHOP knockdown also influences paclitaxel/honokiol-induced protein aggregate formation, cytoplasmic vacuolation and cell death. As expected, we found that siRNA-mediated CHOP silencing (Fig. 5f) led to results similar to those obtained by UBE1 knockdown (Fig. 5c, d and Supplementary Fig. 6B, C). Similarly, we found that pretreatment with CHX, salubrinal and 4-PBA suppressed paclitaxel/honokiol-induced protein aggregate formation in H1299 and H1650 cells (Supplementary Fig. 6D). Interestingly, we found that UBE1 knockdown led to retarded upregulation of CHOP, ATF4 and XBP1s protein expression levels induced by paclitaxel/honokiol (Fig. 5e and Supplementary Fig. 6E), suggesting that UBE1-induced ubiquitinated protein aggregate formation may act upstream of CHOP. Accordingly, silencing of UBEI and CHOP impeded paclitaxel/honokiol-induced cell death (Fig. 5g and Supplementary Fig. 6F). Together, these results indicate that UBEI and CHOP play critical roles in mediating paclitaxel/honokiol-induced paraptosis.

Fig. 5.

UBE1-mediated ubiquitinated protein aggregates contribute to paclitaxel plus honokiol-induced paraptotic cell death. (a, b) H1299 and H1650 cells were treated with 1 µM paclitaxel (P) plus 20 µM honokiol (H) for 12 h, fixed and incubated with the indicated anti-ubiquitin antibodies and Hoechest 33,342. Images were obtained by fluorescence microscopy. Bars, 20 µm. (c) H1299 and H1650 cells were transfected with or without siRNAs as indicated for 24 h, treated with 1 µM paclitaxel plus 20 µM honokiol for 12 h, and next fixed and incubated with the indicated anti-ubiquitin antibody and Hoechst 33,342. Bars, 20 µm. (d) Cells were treated as described in (c) and evaluated by phase-contrast microscopy. Bars, 20 µm. (e) Cells were transfected with siRNA against UBE1 for 24 h, and treated with the indicated doses of paclitaxel plus honokiol for 24 h. Next, cell extracts were prepared and subjected to Western blotting of CHOP, ATF4 and XBP1s; β-actin was used as a loading control. (f) siRNA-mediated knockdown of UBE1 and CHOP verified by Western blotting. (g) Cells were treated as described in (e), after which cell death was determined using a LDH activity assay. **p < 0.005 vs. paclitaxel plus honokiol

LC3 is involved in paclitaxel/honokiol-induced paraptosis and cell death

Since LC3 expression can be upregulated by paraptosis inducers [30, 31], we next tested changes in LC3 expression and its role in paclitaxel/honokiol-induced paraptosis in NSCLC cells. Although we found that paclitaxel or honokiol alone induced minor increases in LC3I/II expression, paclitaxel/honokiol clearly increased LC3I/II and p62 protein levels, which were not inhibited by 3-MA and CQ (Fig. 6a, b and Supplementary Fig. 7A, B), implying that this process is autophagy independent. Intriguingly, we found that pretreatment with CHX not only retarded paclitaxel/honokiol-induced increases in CHOP and Bip expression, but also attenuated upregulation of the LC3-I/II and P62 proteins (Fig. 6b and Supplementary Fig. 7B). Concurrently, p62 aggresome formation was observed following treatment with paclitaxel/honokiol, which was diminished by pretreatment with CHX, but not affected by 3-MA (Fig. 6c and Supplementary Fig. 7C). Again, 3-MA and CQ failed to impede paclitaxel/honokiol-induced cytoplasmic vacuolation and cell death (Fig. 6d, e and Supplementary Fig. 7D, E). However, LC3 knockdown blocked paclitaxel/honokiol-induced upregulation of ubiquitinated proteins, p62, CHOP and Bip proteins (Fig. 6f and Supplementary Fig. 7F), and rescued paclitaxel/honokiol-induced cell death (Fig. 6g and Supplementary Fig. 7g) and repression of cell colony formation (Supplementary Fig. 7H, I). Additionally, knockdown of LC3 further increased the paclitaxel/honokiol-induced expression levels of PTEN, p27 and p21, and reduced phosphorylated AKT levels in the NSCLC cells (Supplementary Fig. 6J, K). These data indicate that LC3 may play a critical role in paclitaxel/honokiol-induced cell death through proteasomal inhibition, ER stress and cytoplasmic vacuolation.

Fig. 6.

LC3 is involved in modulation of paclitaxel plus honokiol-induced paraptotic cell death. (a) H1299 and H1650 cells were treated with the indicated doses of paclitaxel (P) and/or honokiol (H) for 24 h, after which Western blotting of LC3-I/II and p62 was performed; β-actin was used as a loading control. (b) H1299 and H1650 cells were treated with 1 µM paclitaxel plus 20 µM honokiol for 24 h in the absence or presence of CQ (10 µM), 3-MA (10 µM) or CHX (2 µ M), after which Western blotting of the indicated proteins was performed; β-actin was used as a loading control. (c) Cells were treated as described in (b), fixed and incubated with a primary anti-p62 antibody and secondary antibodies. Images were obtained by fluorescence microscopy. Bars, 20 µm. (d) Cells were treated as described in (b) and evaluated by phase-contrast microscopy. Bars, 20 µm. (e) Cells were treated as described above, after which cell death was determined using a LDH activity assay. (f) Cells were transfected with LC3-specific and control siRNAs for 24 h, and next treated with 1 µM paclitaxel plus 20 µM honokiol for 24 h. Subsequently, Western blotting of the indicated proteins was performed; β-actin was used as a loading control. (g) Cells were treated as described in (f), after which cell death was determined using a LDH activity assay. **p < 0.005 vs. siControl plus paclitaxel and honokiol

Mitochondrial Ca²⁺ overload is involved in paclitaxel/honokiol-induced paraptosis in NSCLC cells

It has been reported that disturbance of intracellular Ca²⁺ homeostasis can cause ER stress and mitochondrial disfunction, resulting in paraptosis [32, 33]. Thus, we next set out to evaluate the effect of paclitaxel/honokiol treatment on changes in intracellular Ca²⁺ levels in the three NSCLC cell lines. Detection of the intracellular Ca²⁺ concentrations was performed using Fluo-3AM staining and flow cytometry. Our results showed that paclitaxel/honokiol prominently increased the intracellular Ca²⁺ levels compared to honokiol alone, in a time-dependent manner, up to a peak at 6 h and declined slightly thereafter, whereas paclitaxel alone did not cause any increases in intracellular Ca²⁺ levels (Fig. 7a, b and Supplementary Fig. 8A, B). Likewise, flow cytometry using Rhod-2 AM staining revealed that the mitochondrial Ca²⁺ levels were markedly increased after paclitaxel/honokiol treatment compared to honokiol alone, whereas paclitaxel alone did not alter the mitochondrial Ca²⁺ levels (Fig. 7c, d and Supplementary Fig. 8C, D). As Ca²⁺ can influx into mitochondria through uniporters when intracellular Ca²⁺ levels are high [34], and ruthenium red (RR) can inhibit this mitochondrial uptake of Ca²⁺ [34, 35], we next examined the effect of RR on mitochondrial Ca²⁺ levels, cell death and cytoplasmic vacuolation induced by paclitaxel/honokiol. We found that RR dose-dependently repressed paclitaxel/honokiol-induced increases in mitochondrial Ca²⁺ levels (Supplementary Fig. 8E), cell death and cytoplasmic vacuoles (Fig. 7e, f, g and Supplementary Fig. 8F, G), and increases in polyubiquitinated proteins, CHOP and ATF4 expression (Fig. 7h and Supplementary Fig. 8H). In addition, we found that RR suppressed paclitaxel/honokiol-induced phosphorylation of JNK1/2 and ERK1/2 (Fig. 7i). These results indicate that paclitaxel/honokiol-induced disruption of intracellular Ca²⁺ homeostasis and mitochondrial Ca²⁺ overload are responsible for mediating paraptosis events in mitochondria.

Fig. 7.

Mitochondrial Ca²⁺ overload contributes to paclitaxel plus honokiol-induced paraptosis. (a, b) H1299 and H1650 cells treated with 1 µM paclitaxel plus 20 µM honokiol for the indicated time periods were stained with 2.5 µM Fluo-3AM and processed for FACS analysis. Fluo-3AM fluorescence intensities (FI) in cells treated with paclitaxel plus honokiol were compared with those of untreated cells. Histograms of cells treated with paclitaxel plus honokiol for 6 h are shown. X axis, fluorescence intensity, Y axis, relative number of cells. (c, d) H1299 and H1650 cells treated with or without 1 µM paclitaxel plus 20 µM honokiol for the indicated time periods were stained with 2.5 µM Rhod-2 AM and processed for FACS analysis. Rho-2 fluorescence intensities (FI) were compared with those of untreated cells and denoted in the graph. Histograms of the cells treated with paclitaxel plus honokiol for 8 h are shown. (e, f) Cells were treated with the indicated doses of paclitaxel and/or honokiol for 24 h in the absence and presence of 2 µM or 4 µM ruthenium red (RR), after which cell viability was determined using a CCK-8 assay. ***p < 0.001. (g) Cells were pretreated with the indicated doses of RR for 4 h, and next treated with 1 µM paclitaxel (P) plus 20 µM honokiol (H) for 24 h, after which phase-contrast images were obtained. Bar, 20 µm. (h, i) Cells treated as described in (g) were subjected to Western blotting to detect the proteins as indicated; β-actin was used as a loading control

Discussion

Paclitaxel is an anti-mitotic agent currently used as standard first-line therapy with platinum agents for advanced NSCLC with wild-type EGFR and ALK [2, 3]. Paclitaxel is thought to act through binding tubulin polymers that stabilize spindle microtubules and prevent their disassembly, thereby arresting tumor cells in the G2/M phase of the cell cycle, ultimately leading to apoptosis [36]. Although some patients benefit from paclitaxel treatment, most patients show resistance to the drug. Therefore, novel agents that exhibit improved efficacies against NSCLC cells with less toxicity toward normal cells are urgently needed. It has been reported that honokiol, a small-molecule polyphenol, may induce paraptosis and apoptosis in leukemia cells [20, 21], and that honokiol is effective in overcoming drug resistance in colon, melanoma and breast cancer [18, 37, 38]. In addition, some studies revealed that paclitaxel can induce concentration-dependent and caspase-independent cytoplasmic vacuolation and paraptosis-like cell death [22, 23]. These observations prompted us to assess whether honokiol can overcome paclitaxel resistance in NSCLC cells through inducing paratosis.

We found that, although low-dose paclitaxel (1 µM) did not induce cytoplasmic vaculation and 20 µM honokiol induced relatively small vacuoles, combination treatment of paclitaxel and honokiol at relative low-doses led to extensive cytoplasmic vacuolation in H1299, H1650 and H1650/PTX cells. Interestingly, co-treatment with the two drugs elicited synergistic killing effects in the NSCLC cells, including paclitaxel-sensitive and -resistant NSCLC cells. Moreover, the enhanced anti-NSCLC effect and cytoplasmic vacuolation by paclitaxel/honokiol was confirmed in mice bearing H1299 cell-derived xenograft tumors. Therefore, these results signify that honokiol can sensitize drug-resistant NSCLC cells to paclitaxel, and increase the cytotoxicity of paclitaxel to drug-sensitive NSCLC cells by inducing paraptotic cell death. We did, however, not observe cytoplasmic vacuolation and cell death in paclitaxel/honokiol-treated BEAS-2B cells, indicating that paraptosis and cytotoxicity elicited by paclitaxel/honokiol is selective to NSCLC cells. Also paclitaxel/honokiol-induced cytoplasmic vacuolation and cell death were found to be independent of apoptosis and autophagy, as indicated by the finding that inhibitors of apoptosis and autophagy could not prevent the effects induced by paclitaxel/honokiol. Additionally, H1650 cells treated with paclitaxel alone or paclitaxel/honokiol showed more obvious apoptotic features and more apoptotic cells compared to H1299 and H1650/PTX cells treated with the same drugs, suggesting that H1299 and H1650/PTX cells are more resistant to apoptosis induced by paclitaxel. Apoptosis induced by chemotherapeutic drugs is the mechanism most commonly used in cancer therapy. However, cancer cells may develop adaptive and/or intrinsic mechanisms, such as modification of drug targets, alteration of drug transport, potentiation of DNA damage repair and deregulation of apoptotic signaling, to facilitate the formation of apoptosis-resistant cells leading to reduced chemotherapeutic drug efficacies [39, 40]. Thus strategies to induce non-apoptotic cell death in apoptosis-resistant cancers may be useful to overcome chemotherapy resistance [40]. Paraptosis is an non-apoptotic form of programmed cell death that is characterized by a process of cytoplasmic vacuolation arising from dilation of the ER and mitochondria. This mode of cell death lacks characteristic apoptotic features and the cells do not respond to typical apoptosis and autophagy inhibitors [41, 42]. De novo protein synthesis, proteasomal disfunction and MAP kinase activation have been reported to be involved in paraptosis induction [42, 43], and AIP-1/Alix has been identified as an inhibitor of paraptosis [43]. Here, immunostaining with ER- and mitochondrial membrane-specific markers indicated that paclitaxel/honokiol-induced vacuoles originate from the ER and mitochondria. Moreover, we found that paclitaxel/honokiol inhibited proteasome activity and triggered ER stress responses, as indicated by the accumulation of polyubiquitinated proteins and increases in proteasome-substrate proteins and ER stress markers, which were impeded by CHX. Furthermore, CHX and salubrinal, an ER stress inhibitor, suppressed paclitaxel/honokiol-induced CHOP and ATF4 expression and cytoplasmic vacuolation, further demonstrating that paclitaxel/honokiol-induced cytoplasmic vacuoles are derived from the ER and mitochondria. It has been reported that proteasome inhibition-induced accumulation of misfolded proteins in the ER lumen may magnify osmotic pressure that draws fluid from the cytoplasm to the ER, thereby dilating their lumen [44]. Vacuoles derived from dilated ER in NSCLC cells are preceded by a significant increase in expression of an ER-chaperone, Bip, which is anti-apoptotic [45]. A persistent presence of ER stress with higher levels of Bip may shift the fate of cells away from apoptosis, possibly towards non-apoptotic cell death. An alternative caspase-independent non-apoptotic death pathway may become active under conditions of persisting ER stress when apoptosis gets inhibited. Accumulation of misfolded proteins within the mitochondria has also been found to lead to mitochondrial stress and expansion in a manner similar to ER dilation [46].

Interestingly, we found that UBE1 knockdown led to repressed paclitaxel/honokiol-induced accumulation of ubiquitinated protein aggregates and cytoplasmic vacuoles, and decreased paclitaxel/honokiol-induced expression of CHOP, ATF4 and XBP1s as well as NSCLC cell death, signifying an indispensable role of UBE1 in the process of cytoplasmic vacuolation. Silencing of CHOP led to similar effects, indicating that CHOP also plays a critical role in paraptosis induction by triggering ER stress and UPR, and subsequent accumulation of ubiquitinated proteins. Nevertheless, the effects of CHOP on ubiquitinated protein aggregation and cytoplasmic vacuolation may act downstream of UBE1.

In this study, we show that LC3, an important component of the autophagy pathway, may play a vital role in mediating paclitaxel/honokiol-induced vacuolation and cell death. Autophagy inhibitors did not block paclitaxel/honokiol-induced upregulation of LC3 and cytoplasmic vacuolation/cell death, which is in line with a previous report by Karetal et al. [30]. siRNA-mediated LC3 depletion abated paclitaxel/honokiol-induced accumulation of ubiquitinated proteins, upregulation of ER stress marker proteins as well as cell death, which appear to be related to an increased expression of the growth inhibitory proteins PTEN, p21 and p27 and a decreased formation of growth promoting p-AKT in the cells. p62 protein, an autophagy substrate and ubiquitin receptor [47], was found to be increased upon paclitaxel/honokiol treatment, suggesting inhibition of autophagy due to autophagy-mediated promotion of the degradation of p62 protein, and the notion that there is a correlation between autophagy induction and downregulation of p62 expression [48]. Although the role of p62 in regulating paraptotic cell death is unclear, Boridy et al. proposed that it is likely that the inhibition of autophagy leading to accumulation of autophagy substrates and receptors may lie upstream of proteasomal dysfunction and programmed cell death in certain cases [49]. p62 may act by sequestering and delaying the delivery of substrates destined for proteasomal degradation [49], thereby causing the buildup of aggregated substrates leading to cell death by paraptosis.

While inhibition of proteasome activity can increase the accumulation of ubiquitinated proteins in the ER, resulting in ER stress [33, 43], proteasomal disfunction by paclitaxel/honokiol may contribute to paraptotic changes in the ER, with features distinct from ER stress-induced apoptosis. Therefore, proteasomal impairment may be necessary, but not sufficient, for paclitaxel/honokiol-induced paratosis, implying the existence of other signals that are involved in mitochondrial paraptotic changes. Our data show that uniporter-mediated mitochondrial Ca²⁺ overload may be an initial and key signal in paclitaxel/honokiol-induced paraptotic signaling in NSCLC cells, since RR pretreatment blocked paclitaxel/honokiol-induced paraptotic signals, including proteasomal inhibition, CHOP and ATF4 expression and JNK and ERK activation, and prevented paclitaxel/honokiol-induced cytoplasmic vacuolation and cell death. Ample evidence indicates that disturbance of intracellular Ca²⁺ compartmentalization, such as Ca²⁺ overload in mitochondria and the ER, can cause mitochondrial disfunction and ER stress and subsequent cell death [50, 51]. Here, we found that mitochondrial Ca²⁺ overload resulting from elevated cytoplasmic Ca²⁺ levels following paclitaxel/honokiol treatment may contribute to mitochondrial dilation and subsequent paraptotic cell death.

In summary, the data reported here show that combination treatment of paclitaxel and honokiol potently suppresses the viability of NSCLC cells in vitro and represses the growth of paclitaxel-resistant NSCLC cell xenografted tumors in vivo. This suppression appears to be achieved through the induction of paraptotic cell death. These observations may point at strategies to surmount paclitaxel resistance in NSCLC cells.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Funding

This work was partially supported by the National Youth Science Foundation of China (Grant No. 81402485).

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Ethical approval

All experimental procedures were conducted in accordance with the Chinese legislation regarding experimental animals and were approved by the Institutional Animal Care and Use Committee of the Jiangsu University.