Introduction

Head and neck cancers (HNC) rank among the ten most prevalent cancers globally, accounting for nearly 1 million new cases and approximately 500,000 deaths annually1. Head and Neck Squamous cell carcinomas (HNCC) are the most common subtype of HNC, comprising about 90% of cases2. HNSCCs originate from the mucosal epithelium of the oral cavity, sino-nasal cavity, pharynx, and larynx3. Most patients diagnosed at an early stage already exhibit loco-regional metastases at the initial diagnosis4. Current guidelines recommend a regimen of platinum-based chemotherapy combined with radiotherapy for these patients5. However, the severe toxicity of these treatments significantly diminishes patient quality of life, and not all patients respond to these therapies due to chemoresistance6. Over recent decades, efforts have been made to enhance the quality of life and reduce chemoresistance by modifying the therapeutic scheme7.

Over time, cancer research increasingly focused on natural molecules, particularly flavonoids—polyphenolic secondary metabolites derived from plants8. Xanthohumol (XN) (Fig. 1A), a prenylated chalcone present in the hop, the beer’s main ingredient, is one of the most studied flavonoids. It showed several activities against various diseases, such as diabetes9, aging10, inflammation11, infection12, and cancer13. Among the types of cancer, Xanthohumol has demonstrated significant effectivity against melanoma14, leukemia, OSCC14, colorectal15, pancreatic16, liver17, and breast cancer18. Indeed, XN showed the ability to inhibit cell growth, induce apoptosis, modulate cell signaling, and activate immune response19. However, detailed knowledge about XN’s activity remains limited, particularly its effects on HNSCC. Our previous research on HNSCC, along with studies on natural compounds and cellular processes such as apoptosis, stemness, autophagy, ER stress20,21,22 has prompted us further to delve into the effects of XN on this cancer cell model. In this work, we are studying the impact of XN and Hop Extract (that contains 2% Xanthohumol) on HNSCC in combination with standard chemotherapy molecules, demonstrating its potential in cancer treatment. We detail how XN influences cell viability, proliferation, stemness, and programmed cell death by activating autophagy and ER stress mechanisms.

Fig. 1
figure 1

XN and HOP EX affect HNSCC cell viability: (A) Xanthohumol chemical structure. (B) Viability assay measured by MTT (concentration range 0–25 µM XN, 0–1% HOP EX) at 48 h post-treatment. (C) Apoptosis analysis by Annexin V/PI assay 48 h post-treatment with IC50 (original plot Fig.S1C) (D) Synergyfinder + analysis shows significant synergism in combinations of XN + Cisplatin in UMSCC-103 and CAL-33. (E) Western blot results and quantification of BAX/BCL2 ratio in XN, Cisplatin, and combination treatment compared to the control. (F) Synergyfinder + analysis shows significant synergism in combinations of XN + Cisplatin in UMSCC-103 cis-R. (G) XN sensitivity (MTT assay) of UMSCC-103 cis-R cells.

Full size image

Results

Effects of Xanthohumol and hop extract on cell viability and synergy with cisplatin

To investigate the effect of pure Xanthohumol and hop extract on the viability of the HNSCC cell lines UMSCC-103 and CAL-33, we performed an MTT assay. We observed that both XN and HOP EX significantly affect cell viability after 48 h (Fig. 1B), in a dose-dependent manner (UMSCC-103 XN IC50 = 12.36µM, HOP EX IC50 = 0.96% - CAL33 XN IC50 = 11.97µM – HOP EX IC50 = 0,83%). Interestingly, we found that XN and HOP EX did not affect the cell viability of non-tumoral cells (HaCat) (Fig. S1B). We used Annexin V/PI assay to assess the mechanisms leading to cell death, observing a strong induction of both early and late apoptosis, with no necrosis evident (Fig. 1C-S1C). To explore potential synergistic effects, we combined XN and HOP EX with varying concentrations of Cisplatin. The combination showed a significant reduction in cell viability compared to individual treatments, with XN-Cisplatin displaying the most pronounced synergistic effect, as analyzed by SynergyFinder + software23. The analysis pointed out a strong synergism in the XN-Cisplatin combination (Fig. 1D). Reducing cisplatin dose is very desirable as it could reduce side effects and delay the onset of resistance7. In order to identify the mechanisms of apoptosis-mediated cell death, we monitored by Western blot analysis the expression level of the key markers of apoptotic cascade, such as PARP-1, involved in DNA damage repair, Bax, and Bcl-2. The results showed that XN induced an increase in the BAX/BCL-2 ratio and its combination with cisplatin was synergic. The same trend was observed for PARP-1 fragmentation (Fig. 1E). To further demonstrate the synergistic effect of XN with cisplatin, we created a cicisplatin-resistantine starting from UMSCC-103. This cell line (UMSCC-103-cisR) showed high resistance to cisplatin treatment at the concentration previously used. Treatment with XN has been shown to resensitize cells to treatment with cisplatin (Fig. 1F and G).

Impact of Xanthohumol and hop extract on HNSCC stemness and clonogenicity

Cancer stemness is linked to aggressive disease characteristics, therapy resistance, metastasis, and recurrence. Many stemness-reducing molecules are in pre-clinical and clinical development23. We examined the effects of XN and HOP EX on stemness markers in UMSCC-103 and CAL-33 cells using spheroid assays and flow cytometry. We used low toxicity concentrations (IC10) because we wanted the effects on the stemness were not linked to the effects on the viability, both treatments inhibited sphere formation (Fig. 2A-B). To investigate their effect on cancer stemness, we analyzed the expression of CD44, CD44v6, and CD133 on the cell membrane by Flow cytometry. Both treatments strongly reduced the expression of CD44v6 and CD133. Moreover, XN had the same effect on total CD44; however, for HOP EX, we did not observe any effect (Fig. 2C – S2A). We also analyzed the gene expression of HNSCC’s stem markers: CD44, CD44v6, CD133, NANOG, OCT3/4, and SOX2. XN and HOP EX strongly reduced gene expression of CD44 and CD44v6. XN also significantly reduces the expression of CD133, NANOG, OCT3/4, and SOX2. HOP EX had the same tendency but only significantly affected SOX2 (Fig. 2D– S2B). Colony formation assays confirmed that both XN and HOP EX markedly reduced the clonogenic growth of the UMSCC-103 cells, underscoring their potential to diminish cancer stemness24. We found that XN and HOP EX strongly reduced the colony formation ability of the UMSCC-103 (Fig. 2E – S2C). The stemness of the cells is regulated to some extent by a balanced regulation of ER stress and Autophagy. Excessive ER stress and uncontrolled Autophagy can lead to the depletion of the stem cell population and may cause cell death.

Fig. 2
figure 2

XN and HOP EX affect HNSCC stemness: (A) Sphere formation assays show how XN and HOP EX affect the ability to form spheroids by UMSCC103 AND CAL33. (B) Histograms represent the decrease in the size of spheres after treatments with XN and HOP EX (*** P≤0,001) (C) Surface stemness markers expression after treatments. (D) qRT PCR of “stemness” genes. XN and HOP EX treatments reduce the expression of all these genes in UMSCC103 (*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001). (E) Colony formation assay (images and ImageJ quantification) shows XN an HOP EX affects the clonogenicity ability of UMSCC103 (***P ≤ 0.001, ****P ≤ 0.0001).

Full size image

Induction of ER stress and autophagy by Xanthohumol and hop extract

ER stress and autophagy are essential for cellular homeostasis25. However, their deregulation can cause detrimental outcomes. For this reason, we investigated the effect of XN on ER stress and autophagy in HNSCC cells. Immunofluorescence and flow cytometry analyses showed that these treatments enlarged the endoplasmic reticulum, indicative of ER stress at IC10 concentration (Fig. 3A - S3A). Flow cytometry analysis confirmed the increased fluorescence in samples treated with XN and HOP EX (Fig. 3B – S3B).

Fig. 3
figure 3

XN and HOP EX induce ER stress and autophagy: (A) Live staining fluorescence of ER, XN, and HOP EX increase the size and the fluorescence of UMSCC103 ER. (B) Flow cytometry histograms of ER tracker staining, median fluorescence increases in treated UMSCC103 (C) Immunoblotting and quantification of GRP78 expression (*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001). (D) qRT PCR of CHOP and sXBP1. XN, HOP EX, and Tunicamycin (positive control) treatments increase their expression (*P ≤ 0.05, **P ≤ 0.01). (E) Live staining fluorescence with lysotracker, XN, and HOP EX increases the number and the fluorescence of UMSCC103 lysosomes. (F) Flow cytometry histograms of lysotracker staining median fluorescence increases in treated UMSCC103. (G) Immunoblotting and quantification of ATG7, P62 AND LC3B expression (*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001).

Full size image

The enlarged endoplasmic reticulum suggested that these molecules most likely induce stress on this organelle. The Glucose-Regulated Protein 78 (GRP78) is an ER chaperone protein and it is the master regulator of the unfolded protein response (UPR) activated during ER stress26. We compared GRP78 expression in XN and HOP EX-treated cells with cells treated with Tunicamycin, a known inducer of ER stress, used as a positive control. We found that both treatments increased the expression of this chaperone, confirming the induction of ER stress (Fig. 3C). We also tested the gene expression of two ER stress response mediators: X-box binding protein 1 (sXBP1) mRNA, unconventional alternative splicing that activates XBP1 transcription factor to upregulate ER chaperones, and CCAAT-enhancer-binding protein homologous protein (CHOP), which drives the activation of apoptosis27. sXBP1 and CHOP transcriptions also increased in response to XN and HOP EX treatments (Fig. 3D). Autophagy is a cellular process frequently associated with ER stress. The increase in the autophagic flux is a recycling mechanism of cellular materials (organelles, proteins, debris) that get incorporated into vesicles (autophagosomes) and acquire a degrading function after fusing with lysosomes (autophagolysosomes). Autophagosome formation involves the recruitment of LC3B (Microtubule-associated protein 1 A/1B-light chain 3) and p62/sequestosome proteins. To investigate the activation of autophagy after XN and HOP EX treatments, we performed an IF and a quantitative analysis by flow cytometry using LysoTracker (acidic organelles tracker for lysosomes and autophagolysosomes in live cells) and using as positive control the Chloroquine (CLORO), an autophagosome degradation blocker. This staining showed an increased number of acidic vesicles in treated cells, with median fluorescence increased in XN and HOP EX samples (Fig. 3E-F, S3C-D). To understand if there was an autophagic activation, we performed an immunoblotting analysis to evaluate the level of the autophagy-related markers LC3B, ATG7, and p62/SQSTSM1. LC3B is a significant marker of autophagy present in the cytoplasm (LC3B-I), and when autophagy is activated, it comes lipidated and cleaved to form LC3B-II. ATG7 acts as an E1-like activating enzyme facilitating autophagic vesicle expansion. The immunoblotting results in Fig. 3G show that the LC3B II/I ratio and ATG7 increased after XN and HOP EX treatments. p62/SQSTSM1 is a cargo protein indicator of autolysosomal degradation. Interestingly, we have found that its expression decreased in treated cells28 (Fig. 3G).

ER Stress-Mediated autophagy activation

Autophagy is a process activated in physiological conditions that counteracts stress. ER stress, mediated by UPR factors, is one of the many processes activating autophagy. Therefore, the UPR primary transducers Inositol-requiring enzyme type 1 (IRE1) and protein R-like endoplasmic reticulum kinase (PERK), are proven inducers of autophagy.

We investigated whether ER stress induced by XN is responsible for autophagy activation. Using specific inhibitors for IRE1 (4µ8C) and PERK (GSK2606414), we demonstrated that inhibition of these UPR transducers mitigated ER enlargement (Fig. 4A-B and Fig. S4A-B) and reduced GRP78 expression, confirming the linkage between ER stress and autophagy induction(Fig. 4C). The immunofluorescence of Fig. 4D-S4C demonstrates how the treatments with UPR inhibitors reduced the Lysotracker staining, highlighting the decrease in the number and size of acidic organelles. The results were confirmed by cytofluorimetric quantitative analysis (Fig. 4E-S4D), where the fluorescence decreased by about 40% for cells treated with XN in combination with 4µ8C or GSK2606414. The LC3 lipidation was also completely reverted by inhibitors in XN-treated cells (Fig. 4F).

Fig. 4
figure 4

ER stress drives to autophagy: (A) Live staining fluorescence of ER and (B) Flow cytometry analysis shows how 4µ8C and GSK2606414 revert XN and HOP EX effect on UMSCC103 ER. (C) Immunoblotting and quantification of GRP78 expression (**P ≤ 0.01) (D) Live staining fluorescence with lysotracker and (E) Flow cytometry histograms of lysotracker staining, IRE1, and PERK inhibitors revert the XN effect on acidic organelles, decreasing their number and median fluorescence in UMSCC103. (F) Immunoblotting and quantification of LC3B II/I ratio. Inhibitors reduced lipidation of LC3B.

Full size image

ER stress and autophagy drive programmed cell death

XN strongly affects the viability of UMSCC-103 and CAL-33. As demonstrated, this effect is due to the activation of apoptosis, which can be induced by both ER stress and autophagy. To better understand the role of these two biological processes, we first tested if the inhibition of PERK (0.25 µM) or IRE1 (2.5 µM) phosphorylation affected viability in XN-treated cells. The results revealed both UPR inhibitors could revert the effect of XN 4 µM (IC25) treatment for 48 h (Fig. 5A). Furthermore, the treatment with the inhibitors provoked a substantial reduction of apoptosis compared to the cells treated only with XN (Fig. 5B- S5A original plots). This suggested a causative link between ER stress and programmed cell death. Considering that we demonstrated that ER stress activates autophagy, a cellular process that can also lead to apoptosis, we verified whether its inhibition reduced the effect on XN-induced cell death. 3-methyladenine (3-MA) inhibits class III PI3K activity, which is essential for the induction of autophagy. Interestingly, the MTT assay showed that combined treatment with 3-MA (1mM) increased the viability in XN-treated cells (Fig. 5C). We observed the same effect in the Annexin V/Pi assay, where 3-MA treatment was able to reduce apoptosis (Fig. 5D- S5B original plot). Our findings indicate that ER stress and autophagy collectively contribute to the reduced viability and increased apoptosis observed in UMSCC-103 and CAL-33 cells treated with XN. Inhibition of PERK and IRE1 significantly reversed these effects, reducing apoptosis and improving cell survival. Treatment with 3-methyladenine (3-MA), which blocks early stages of autophagy, further validated that autophagy plays a pro-apoptotic role in the context of XN-induced cell stress.

Fig. 5
figure 5

ER stress causes programmed cell death through autophagy: (A) Viability assay measured by MTT 4µ8C and GSK2606414 revert XN cytotoxicity of XN 48 h post-treatment (*P ≤ 0.05, **P ≤ 0.01). (B) Apoptosis analysis by Annexin V/PI assay with IRE1 and PERK inhibitors 48 h post-treatment (original plot Fig.S5A) (C) Viability assay shows that XN cytotoxicity is reverted with 3-MA (*P ≤ 0.05). (D) 3-MA reduces also the apoptosis induced by XN in UMSCC103 and CAL33 cells (original plot Fig.S5B).

Full size image

Discussion

In recent decades, numerous studies have aimed to find an effective anti-cancer drug with lower side effects. Head and neck squamous cell carcinoma (HNSCC) is particularly challenging due to its poor prognosis and the often ineffective treatments that significantly diminish patients’ quality of life. Furthermore, Mundi et al. highlighted gender differences in disease frequency and survival rates i.e. high in males and low in females, prompting our use of both male (CAL-33) and female (UMSCC-103) derived cell lines to explore these variances further29Many researchers have used natural molecules to counter tumors and even enhance the effects of standard medications. Among these compounds, our study focuses on Xanthohumol (XN), a naturally occurring molecule from the female inflorescence of Humulus lupulus, and its potential in cancer therapeutics. Xanthohumol and a commercially viable hop extract (HOP EX) containing a known concentration of XN were evaluated for their anticancer effects. The affordability and dietary compatibility of hop extract make it an attractive alternative to pure Xanthohumol.

Our results have demonstrated that treatments with XN and HOP EX (XN 2%) have a significant antiproliferative in vitro effect on HNSCC cells, activating apoptosis. Cisplatin is the most common antineoplastic drug used against solid cancers including HNSCC patients. However, treatment with this molecule often causes many side effects30. For this reason, the central goal of many studies is lowering Cisplatin concentrations without decreasing its efficacy. Interestingly, we discovered that the XN and HOP EX combined with Cisplatin had a synergistic effect on UMSCC-103 viability, particularly the XN combination. This synergy suggests potential dietary applications for Xanthohumol to boost Cisplatin’s effectiveness or allow for lower dosages, reducing its toxicity.

Despite the initial success with the HNCC therapies, the tumor recurrence rate is high. These features are correlated with cancer stem cells (CSCs) which are directly linked to resistance, relapse, tumorigenesis, and poor clinical outcomes31. The sphere formation assay provides a valuable tool to assess the stem cell population residing in cancer cell lines and screen for drugs. Our investigations into the sphere-forming capabilities of CSCs showed that XN and HOP EX could significantly reduce these properties. These compounds also downregulated critical stemness markers such as CD44, CD44v6, and CD133, and impacted the gene expressions of key transcription factors involved in CSC maintenance32,33, including HNSCC34,35; moreover, HOP EX had the same tendency but did not affect CD44 protein expression. The gene expressions for the activated transcription factors in CSCs (Nanog, OCT3/4, SOX2)36 were strongly influenced by XN treatment, while the HOP EX effect was significant only on SOX2 expression. These results suggested that XN impaired the stemness of UMSCC-103, reducing the CSC population. HOP EX had a weaker but significant effect that almost blocked the ability of CCs to form spheres. Finally, we evaluated the clonogenic activity of the cells after treatment with our compounds, showing that both had a significant ability to reduce clonal growth in vitro.

Our previous studies have demonstrated the link between ER stress, autophagy, and apoptosis21,37. Treatment with XN and HOP EX led to ER stress, evidenced by the upregulation of UPR mediators such as GRP78, CHOP, and sXBP1, and an increase in autophagic activity We also found that the Lysotracker activated the acidic compartment, suggesting further activation of other cellular processes, such as autophagy. During autophagy induction, LC3-I becomes in LC3-II with a concurrent decrease in p62. We observed that the LC3II/I ratio increased and p62 decreased by immunoblotting after treatments. To confirm autophagy’s activation, we also detected ATG7, a protein essential for the formation of autophagosomes, which also increased in cells treated with XN and HOP EX. To understand whether observed cellular processes such as ER stress, autophagy, and cell death were linked, we inhibited the UPR. It is known that UPR, activated by ER stress, can mediate several events such as the activation of autophagy and apoptosis. As shown in our previous research, IRE1 and PERK are the two primary mediators of this process37. We inhibited UPR mediators with 4µ8C (IRE1) and GSK2606414 (PERK), finding that treatment of cells with these molecules inhibited XN-induced ER expansion, suggesting that UPR plays a central role. The reduction of GRP78 in XN-treated cells confirmed that the use of inhibitors blocked the UPR. Many authors have demonstrated that using these molecules causes potent and selective inhibition of ER stress responses38,39. Considering the decrease of the LC3II/I ratio induced by the combination with 4µ8C and GSK2606414, we assume that UPR blockage inhibited the activation of autophagy.

We have demonstrated that ER stress-induced autophagy leads to cell death in HNSCC. We found that treatment with ER stress inhibitors reduced the XN effect on cell viability and apoptosis. Autophagy can be a cytoprotective process, stimulating the recycling of damaged cytoplasmic components and maintaining cellular homeostasis. Many studies have shown that it can also be a proapoptotic mechanism40. To understand if the activation of autophagy was pro-survival or proapoptotic, we combined XN treatment with 3-MA. Different from chloroquine, 3-MA inhibits autophagosome formation at early stages, inactivating autophagy functional effects41. We found that 3-MA significantly reduced the XN effect on viability and apoptosis, suggesting that autophagy was a proapoptotic event in this case.

Conclusion

Ultimately, our study illustrates that XN and HOP EX not only reduce HNSCC cell viability and stemness but also trigger cellular mechanisms like ER stress and autophagy that lead to apoptosis. This complex interplay suggests that enhancing autophagic and ER stress responses could be a novel approach to improve cancer treatment outcomes. The potential for XN to act both as a therapeutic and a dietary supplement opens new avenues for integrating natural compounds into cancer treatment regimens, aiming to enhance efficacy while minimizing side effects.

Materials and methods

Chemicals, cell culture, and in vitro treatment

Xanthohumol pure extract (XN) and Xanthohumol Hop extract 2% purity (HOP-EX) were obtained from Greensky. Cisplatin (purity ≥ 98%) was purchased from Sigma-Aldrich (St. Louis, MO, USA). Selective UPR inhibitors 4µ8C and GSK2606414 were obtained from Tocris. 3-MA or 3-methyladenine was purchased from MERK (Darmstadt, Germany).

The human tongue cancer cell line UMSCC-103 was obtained from the University of Michigan under a protocol approved by the Institutional Review Board Office under the university’s regulations and described here42. The human tongue squamous cell carcinoma cell line CAL-33 (DSMZ no. ACC447) was obtained from DSMZ- German Collection of Microorganisms and Cell Cultures GmbH. Human keratinocytes HaCaT cells were from Cell Line Service (Eppelheim, Germany). Cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum, 100 U/mL penicillin, 100 µg/mL streptomycin and 1% l-glutamine and cultured at 37 °C in a 5% CO2 humidified atmosphere. The cells were tested Mycoplasma negative.

To establish cisplatin-resistant cell lines we treated UMSCC-103 with an initial dose of 1/10 IC50 (50 ng/ml, ), and the dose was increased by adding 100 ng/ml to each previous dose. The next dose was administered until the cells entered a stable growth period. After 8 weeks the IC50 was increased at 50 µM. This cell line was renamed UMSCC-103-cisR.

MTT assay

Cell viability was measured by the colorimetric 3-(4,5-dimethyl-2-thiazolyl)-2,5- diphenyltetrazolium bromide (MTT) assay. Cells were seeded in 96-well plates at 104 cells per well density. Then they were treated with 100µL of 1 mg/mL MTT (Sigma) in DMEM medium containing 10% fetal bovine serum for 3–4 h at 37 °C. The medium was then replaced with 100µL of DMSO and then absorbance at 570 nm was measured using a microplate ELISA reader (Multiskan FC Microplate, Thermo Scientific, Rockford, USA) Thermo with DMSO was used as the blank. Data are presented as means standard deviations (SD); error bars represent triplicate samples.

Synergy assay

To analyze MTT assay data from XN-Cisplatin HOP EX-Cisplatin combinations treatment in the UMSCC-103 cell line, we used Synergyfinder software43. Data from the combination viability assay were processed using traditional Synergy models (Loewe, Bliss, and HSA) as single experiments for each treatment. We chose the HSA model based on the results because The HSA model predominates in experimental combination screens and is the product of chemical combinations of inhibitors with unrelated targets.

Flow cytometry analysis

The UMSCC-103 cell line was characterized with flow cytometry for three evaluations: apoptosis assay, stemness, autophagy, and ER size. Cells were incubated in 12 well plates at a density of 0.06 × 106 cells/well. After allowing them to attach for 1 day, they were treated for 48 h with the optimum concentrations of the drug combinations. Cells were analyzed with a FACSAria III (BD Biosciences, San Jose, CA), and data were analyzed by FlowJo V10 software (FlowJo LLC, USA). Apoptosis was evaluated with an Annexin V apoptosis detection kit (BD Biosciences). For this analysis, cells were treated with IC50 concentrations of the drug combinations. For stemness characterization, we evaluated the following markers: CD44-PE conjugated (BD Bioscience #550989), CD44v6 (Invitrogen #33-6700), CD133-PE conjugated (BD Bioscience #566594). For the staining, the UMSCC-103 cells were collected in FACS tubes at approximately 0.2 × 106 cells/tube and then incubated with the antibodies for 30 min at 4 °C in the dark. After incubation, cells were resuspended in 200µL of PBS for analysis at the Flow cytometer. To evaluate the effects of the treatment on the autophagy process, we used LysoTracker-red staining (LTR) (Thermo Fisher Scientific, USA), and for the ER size evaluation, the ER Tracker-blue staining (ERTR) (Invitrogen). After 48 h, LTR and ERTR were added to the cells for 20 min at 37 °C at a final concentration of 0.1 µM in DMEM. After, cells were washed with PBS and observed at the fluorescence microscope EVOS FL Cell Imaging System (Thermo Scientific, Rockford, USA). The fluorescence intensity was then analyzed by flow cytometry as previously described.

Sphere formation assay

For sphere formation, cells were plated in ultra-low-attachment flasks (Corning, NY, USA) in DMEM/F12 supplemented with B27 and N2 supplement (Gibco, NY, USA), 100 IU/ml penicillin, and 100 µg/ml streptomycin (Invitrogen, Carlsbad, CA). Human fibroblast growth factor (FGF, 20ng/ml) and epidermal growth factor (EGF, 20ng/ml) (both from Sigma Aldrich, MO, USA) were added every other day. After 48–72 h, spheres were disaggregated using trypsin/EDTA for 5 min and re-plated in the same condition for 3 passages, the spheres formed in the last passage were used for experiments.

Colony forming assay

Cells were washed twice with PBS. Cells were then plated. Untreated cells were used as controls and to assess plating efficiency. Cultures were observed for 10 days (depending on growth rate differences between cell lines). Colonies were then fixed and stained with crystal violet in 20% methanol. The analysis was performed with ImageJ with a specific plugin. All experiments were done in triplicate.

Immunofluorescence

UMSCC-103 cells were seeded at 0.03 × 106 cells/mL 24 well plates. They were treated with XN and HOP-EX for 48 h. After 48 h, ERTB was added to the adherent cells for 20 min at 37 °C at a final concentration of 0.1 µM in DMEM. Then cells were washed twice with PBS and fixed with paraformaldehyde 4% for 20 min at 4 °C. After washing each well twice again with PBS, the last washing was kept in the wells, and the plate was taken to image on a fluorescence microscope EVOS FL Cell Imaging System (Thermo Scientific, Rockford, USA).

qRT-PCR

Total RNA was extracted from UMSCC-103 after treatment with drug combinations (XN-Cisplatin and HOP EX-Cisplatin) with a High Pure RNA Isolation Kit (Lifescience Roche) following the manufacturer’s instructions. cDNA synthesis was done from total RNA (1 µg) using a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems™). PCR analyses were conducted using an Insta Q96® − 6.0 Real Time thermal cycler (Himedia). All samples underwent a 2-min UDG activation at 50 °C and a Dual-Lock DNA polymerase activation 2-min at 95ºC. Then, primers with annealing temperatures below 60 °C did 40 cycles of denaturation at 95 °C for 15 s and annealing/extension at 60 °C for 1 min. Instead, primers with annealing temperatures over 60 °C did 40 cycles of denaturation at 95 °C for 15 s, annealing at 55–60 °C for 15 s, and extension at 72 °C for 1 min.

Each PCR reaction was performed in a total volume of 10 µl containing PowerUp SYBR Green master Mix (Thermo Fisher, #438577) 5 µl, 2 µl of Primer Mix, 2 µl of Nuclease-Free Water, and 1 µl of the sample.

PCR was performed using the following primer sequences and PCR product annealing:

GAPDH, fw: GGAGTCAACGGATTTGGTCG,

rev: CTTCCCGTTCTCAGCCTTGA,

NANOG, fw: TTCAGTCTGGACACTGGCTG,

rev: CTCGGTGATTAGGGTCCAAC,

SOX-2, fw: CGATGCCGACAAGAAAACTT,

rev: CAAACTTCCTGCAAAGCTCC,

OCT3/4, fw: ACATGTGTAAGCTGCGGCC,

rev: GTTGTGCATAGTCGCTGCTTG,

CD133, fw: ATGACAAGCCCATCACAACA,

rev: CCTGAGTCACTACGTTGCCA.

CD44, fw: TCCAAAGGTTTTCCATCCTG,

rev: AGGGCCAGCCTCTATGAAAT.

CD44v6, fw: GACACATATTGCTTCAATGCTTCAGC,

rev: TACTAGGAGTTGCCTGGATGGTAG.

The transcript amount of each gene was normalized to Glyceraldehyde 3- phosphate dehydrogenase (GAPDH).

Western blotting

The UMSCC-103 cell lysates were prepared by resuspending cell pellets in complete RIPA buffer (1 ml of RIPA, 10 µl of PMSF 200mM, 10 µl of ORTOVANADATO 100mM, and 10–20 µl of protease inhibition cocktail, PIC). The lysate was incubated for 30 min on ice and centrifuged at 14,000 rpm for 15 min at 4ºC to obtain the proteins’ supernatant. The protein concentration was established using the Bradford protein assay using Protein Assay Dye Reagent Concentrate (Bio-Rad). The samples containing 50 µg of protein were run on 12% polyacrylamide gel by SDS-PAGE and transferred to Trans-Blot Turbo Mini 0.2 μm Nitrocellulose membranes (Bio-Rad) using Trans-Blot Turbo Transfer System (Bio-Rad). The transfer efficiency and equal protein loading were determined using the Rosso Ponceau (Sigma) membrane staining. The membranes were blocked in 5% bovine serum albumin (BSA) in Tris-buffered saline with 0.1% Tween 20 (TBS-T). Membranes were washed 2–3 times for 5 min each with TBS-T and incubated with Primary antibodies LC3B (Cell signaling #2775, rabbit), ATG7 (Cell signaling #8558, rabbit) and p62/SQSTSM1 (Cell signaling # 5114, rabbit), BAX (cell signaling #5023, rabbit), Bcl-2 (cell signaling 15071, mouse), PARP-1 (cell signaling #9532, rabbit) diluted in 3% BSE overnight, in the dark and at 4ºC. The next day, the membrane was washed 6 times, 5 min each with TBS-T, and then incubated with Secondary Peroxidase-conjugate anti-rabbit IgG (Santa Cruz) was used for enhanced chemiluminescence (ECL) detection diluted in 3%, for 1 h in the dark and at 4ºC. Finally, the membrane was washed 6 times, 5 min each with TBS-T, incubated with Clarity Western ECL Substrate (Bio-Rad), and after reading with the ChemiDoc XRS + System (Bio-Rad).