Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 https://doi.org/10.1186/s13046-022-02267-8 RESEARCH Open Access CXCR4-targeted nanotoxins induce GSDME- dependent pyroptosis in head and neck squamous cell carcinoma Elisa Rioja‑Blanco1,2, Irene Arroyo‑Solera1,2,3, Patricia Álamo1,2,3, Isolda Casanova1,2,3, Alberto Gallardo1,4, Ugutz Unzueta1,2,3, Naroa Serna3,5,6, Laura Sánchez‑García3,5,6, Miquel Quer3,7,8, Antonio Villaverde3,5,6, Esther Vázquez3,5,6,9*, Xavier León3,7,8, Lorena Alba‑Castellón1,2,11* and Ramon Mangues1,2,3,10* Abstract Background: Therapy resistance, which leads to the development of loco-regional relapses and distant metasta‑ ses after treatment, constitutes one of the major problems that head and neck squamous cell carcinoma (HNSCC) patients currently face. Thus, novel therapeutic strategies are urgently needed. Targeted drug delivery to the chemokine receptor 4 (CXCR4) represents a promising approach for HNSCC management. In this context, we have developed the self-assembling protein nanotoxins T22-PE24-H6 and T22-DITOX-H6, which incorporate the de-immu‑ nized catalytic domain of Pseudomonas aeruginosa (PE24) exotoxin A and the diphtheria exotoxin (DITOX) domain, respectively. Both nanotoxins contain the T22 peptide ligand to specifically target CXCR4-overexpressing HNSCC cells. In this study, we evaluate the potential use of T22-PE24-H6 and T22-DITOX-H6 nanotoxins for the treatment of HNSCC. Methods: T22-PE24-H6 and T22-DITOX-H6 CXCR4-dependent cytotoxic effect was evaluated in vitro in two differ‑ ent HNSCC cell lines. Both nanotoxins cell death mechanisms were assessed in HNSCC cell lines by phase-contrast microscopy, AnnexinV/ propidium iodide (PI) staining, lactate dehydrogenase (LDH) release assays, and western blot‑ ting. Nanotoxins antitumor effect in vivo was studied in a C­ XCR4+ HNSCC subcutaneous mouse model. Immunohisto‑ chemistry, histopathology, and toxicity analyses were used to evaluate both nanotoxins antitumor effect and possible treatment toxicity. GSMDE and CXCR4 expression in HNSCC patient tumor samples was also assessed by immunohis‑ tochemical staining. Results: First, we found that both nanotoxins exhibit a potent CXCR4-dependent cytotoxic effect in vitro. Impor‑ tantly, nanotoxin treatment triggered caspase-3/Gasdermin E (GSDME)-mediated pyroptosis. The activation of this alternative cell death pathway that differs from traditional apoptosis, becomes a promising strategy to bypass therapy resistance. In addition, T22-PE24-H6 and T22-DITOX-H6 displayed a potent antitumor effect in the absence of systemic *Correspondence: esther.vazquez@uab.cat; lalba@santpau.cat; rmangues@santpau.cat 9 Institut de Biotecnologia i de Biomedicina and Departament de Genètica i de Microbiologia, Universitat Autònoma de Barcelona and CIBER, Bellaterra, Barcelona, Spain 10 Institut d’Investigacions Biomèdiques Sant Pau, Hospital de Sant Pau, CIBER and Josep Carreras Research Institute, 08041 Barcelona, Spain 11 Institut d’Investigacions Biomèdiques Sant Pau, Hospital de Sant Pau and Josep Carreras Research Institute, 08041 Barcelona, Spain Full list of author information is available at the end of the article © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 2 of 17 ­ XCR4+ subcutaneous HNSCC mouse model. Lastly, GSDME was found to be overexpressed in tumor tis‑ toxicity in a C sue from HNSCC patients, highlighting the relevance of this strategy. Conclusions: Altogether, our results show that T22-PE24-H6 and T22-DITOX-H6 represent a promising therapy for HNSCC patients. Remarkably, this is the first study showing that both nanotoxins are capable of activating caspase-3/ GSDME-dependent pyroptosis, opening a novel avenue for HNSCC treatment. Keywords: Targeted drug delivery, CXCR4, HNSCC, Pyroptosis, GSDME Background has been jeopardized by severe off-target toxicities, espe- One of the main problems of head and neck squamous cially because of their high immunogenicity and reduced cell carcinoma (HNSCC) therapy is that up to 60% of targeting capacity, which limits their long-term use in patients develop loco-regional relapses and up to 30% patients [12, 13]. On the other hand, different ADCs have distant metastases after treatment, dramatically affect- also been tested for the treatment of HNSCC, such as ing their survival. Currently, these HNSCC patients are ABBV-321, an EGFR-targeting antibody conjugated to a no longer candidates for a curative therapy and the main pyrrolobenzodiazepine (PBD) dimer cytotoxic molecule goals are palliation and prolongation of patient survival [14]. Nevertheless, ADCs present important drug leakage [1, 2]. Thus, the development of drug resistance, which during circulation, also limiting their clinical use [15–17]. leads to treatment failure and relapses, constitutes a Altogether, ADCs and immunotoxins present severe off- major issue in current HNSCC treatment, highlighting target toxicities, which dramatically narrow down their the urge for novel therapeutic strategies [3–6]. therapeutic window. In this context, targeting the chemokine receptor 4 In this context, we have developed self-assembling (CXCR4) has emerged as a promising approach in can- protein nanoparticles which incorporate the de-immu- cer treatment. CXCR4 overexpression in HNSCC pri- nized catalytic domain of Pseudomonas aeruginosa mary tumors associates with a higher risk of developing (PE24) exotoxin A or the diphtheria exotoxin (DITOX) loco-regional recurrences and distant metastases after domain from Corynebacterium diphtheriae, that spe- treatment and worse overall survival [7, 8]. Moreover, cifically target CXCR4-overexpressing ­(CXCR4+) cancer preclinical and clinical data suggest that this chemokine cells through the T22 peptide ligand. These nanotoxins, pathway contributes to a resistant phenotype [8]. Thus, named T22-PE24-H6 and T22-DITOX-H6 respectively, targeted drug delivery to CXCR4-overexpressing cells are recombinantly produced in Escherichia coli, where represents a promising antitumor strategy in HNSCC they self-assemble into multimeric nanoparticles. This treatment. fact enables a greater payload capacity while increasing In the last years, different protein toxins have gained the number of ligands per nanoparticle, which endows relevance as moieties of antitumor drugs because of their superselectivity [18]. Moreover, both nanotoxins are pro- interesting properties that can be exploited in clinical duced in a single step process, avoiding chemical conju- oncology [9–11]. Toxins display a great cytotoxicity in gation steps, which allows a straightforward production a wide range of cancer cells, presenting mechanisms of and purification while preventing drug leakage [19]. action capable of killing not only proliferating, but also Here, we describe the cytotoxic and antitumor effect quiescent cells [10]. In addition, toxins can be recom- of these two novel nanotoxins, T22-PE24-H6 and binantly produced enabling large scale production and T22-DITOX-H6 that specifically target ­CXCR4+ cells, for purification. All these facts make them ideal candidates the treatment of HNSCC. First, we evaluate the CXCR4- to replace current chemotherapeutic drugs. However, dependent cytotoxic effect of both nanotoxins in two to prevent undesired off-target toxicities, targeted drug different HNSCC cell lines. Moreover, we analyze the delivery specifically to tumor cells is key for the transla- mechanisms of cell death induced by both T22-PE24- tion of toxin-based drugs to the clinic. In this context, H6 and T22-DITOX-H6, finding that they are capable different antibody-drug conjugates (ADCs) and immu- of activating caspase-3/Gasdermin E (GSDME) medi- notoxins have exploited this strategy. An example is the ated pyroptosis. Since the activation of anti-apoptotic immunotoxin denileukin diftitox, that incorporates the pathways is a main mechanism of resistance to both interleukin-2 (IL-2) fused to the diphtheria exotoxin chemotherapy and radiotherapy in HNSCC [3, 5, 6, 20], domain to target T cell lymphoma cells that overexpress the development of drugs capable of triggering cell death the IL-2 receptor. However, it was withdrawn from the pathways alternative to apoptosis is an important ave- market in 2014 due to life-threatening toxicity in patients nue of research that could increase cure rate in HNSCC [12, 13]. In fact, immunotoxin translation to the clinic patients. Importantly, T22-PE24-H6 and T22-DITOX-H6 Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 3 of 17 also present a potent antitumor effect in the absence of Life Technologies) supplemented with 10% Fetal Bovine systemic toxicity in a ­CXCR4+ subcutaneous HNSCC Serum (FBS), 100 U/mL penicillin/ streptomycin, 2 mM mouse model. Lastly, we incorporate clinical data show- glutamine, and 400 ng/ml hydrocortisone (Life Technolo- ing that GSDME (also named DFNA5) is overexpressed gies) at 37 °C and 5% ­CO2 in a humidified atmosphere. in tumor tissue of HNSCC patients, highlighting the relevance of this strategy. Thus, activating caspase-3/ Patient samples GSDME-dependent pyroptosis specifically in ­ CXCR4+ HNSCC patient samples were obtained by the Otorhi- HNSCC tumor cells may represent a novel and enticing nolaryngology Department of the Hospital de la Santa approach for the treatment of HNSCC patients. Remark- Creu i Sant Pau (Barcelona, Spain) in accordance with ably, this is the first study showing that T22-PE24-H6 and the Institutional Review Board of the institution. Writ- T22-DITOX-H6 activate caspase-3/GSDME mediated ten informed consent was acquired from all the patients pyroptosis in oncotherapy. involved in this study. Methods Cell viability assays Nanoparticles production, purification, Cell viability upon T22-PE24-H6 and T22-DITOX-H6 and characterization exposure was assessed with the Cell Proliferation Kit II T22-PE24-H6 and T22-DITOX-H6 production, purifica- (XTT) (Roche) according to the manufacturer’s instruc- tion, and characterization have been previously described tions. Cells were seeded in 96-well plates (5000 cells/ [19]. T22-PE24-H6 nanoparticles self-assemble into well for 22A and 2500 cells/well for 74B) and treated ~ 60 nm nanoparticles, whereas T22-DITOX-H6 form 38 with either buffer (166 mM ­ NaCO3H pH 8) or differ- and 90 nm nanoparticles. ent concentrations of T22-PE24-H6 or T22-DITOX-H6 (0–50 nM) for 48 h. For AMD3100 blocking experiments, Cell lines and cell culture 1 μM AMD3100 was added 1 h prior to the addition of the UM-SCC-22A (22A) and UM-SCC-74B (74B) human nanotoxins. For the zVAD pre-treatment, 100 μM zVAD papillomavirus negative (­HPV−) HNSCC cell lines [21] (Calbiochem) was added to the plates and incubated at were kindly provided by Dr. R. H. Brakenhoff and Dr. 37 °C for 2 h before nanotoxin treatment. After 48 h, XTT Gregory Oakley respectively. 22A mock, 74B mock, reagent was added to the plate and further incubated 22A-CXCR4+, and 74B-CXCR4+ were obtained by lenti- at 37 °C for 4 h, then absorbance, which directly corre- viral transduction with the plasmids pLenti-III-UbC-luc lates to the number of viable cells, was measured using a and pLenti-III-UbC-CXCR4-2A-luc (abm, Vancouver, multi-well spectrophotometer (FLUOstar Optima, BMG Canada) respectively, as already described in previous Labtech). All experiments were performed in triplicate. work [22]. HNSCC cell lines were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) (Gibco, Life Technol- Flow cytometry ogies) supplemented with 10% Fetal Bovine Serum (FBS), Cell death induced by T22-PE24-H6 and T22-DITOX-H6 100 U/mL penicillin/ streptomycin, and 2 mM glutamine was further assessed using the Annexin V-FITC / propid- (Life Technologies) and incubated at 37 °C and 5% ­CO2 ium iodide (PI) detection kit (Merck Millipore) following in a humidified atmosphere. CXCR4 expression in 22A manufacturer’s instructions. 74B-CXCR4+ cells were cul- mock, 74B mock, 22A-CXCR4+, and 74B-CXCR4+ has tured in 6-well plates (25,000 cells/well) and exposed to been already evaluated in previous work [22]. 50 nM of either of the two nanotoxins for different times 575 and 909 patient-derived cell cultures were obtained (15 h, 24 h, and 48 h). Cells were analyzed by MACS- from two HNSCC patient tumor samples with high Quant analyzer flow cytometry with the MACS Quantify CXCR4 tumor expression. Tumor samples were disaggre- version 2.3 software (Miltenyi Biotech). The experiment gated by incubation in Trypsin-EDTA 0.25% (Gibco, Life was performed in triplicate. Technologies) for 2 h at 4 °C, followed by further incuba- tion with collagenase type II (200 mg/mL, Life Technolo- LDH release assay gies) and DNase (20 mg/mL, Sigma-Aldrich) for 2 h at LDH release from 22A-CXCR4+ and 74B-CXCR4+ cells 37 °C. After some mechanical disruption, samples were upon nanotoxin exposure was studied using the CytoTox filtered through a 40 μm mesh, and cultured. Epithelial 96 Non-Radioactive Cytotoxicity Assay (Promega). Cells cell enrichment was performed by differential trypsiniza- were seeded in 96-well plates (5000 cells/well for 22A tion and maintaining the cells in Defined Keratinocyte- and 2500 cells/well for 74B) and exposed to the nano- SFM medium (Gibco, Life Technologies). 575 and 909 toxins (5 nM for T22-PE24-H6 in 22A-CXCR4+ cell line, cell cultures were maintained Dulbecco’s Modified Eagle 50 nM in the rest of conditions). zVAD (Calbiochem) Medium: Nutrient Mixture F-12 (DMEM/F-12) (Gibco, pre-treatment was performed at a 100 μM concentration Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 4 of 17 and incubated at 37 °C for 2 h prior to nanotoxin addi- The subcutaneous tumor model was generated by tion. After 48 h of treatment, cytotoxicity assay reagents subcutaneous injection of 10 million 74B-CXCR4+ cells were added according to the manufacturer’s instructions, in both flanks of the animal. To assess T22-PE24-H6 and the absorbance at 492 nm was measured using a and T22-DITOX-H6 antitumor effect, animals bearing multi-well spectrophotometer (FLUOstar Optima, BMG tumors around 60–100 ­mm3 were randomized into three Labtech). All experiments were performed in triplicate. groups (n = 10 per group). Animals were intravenously administered buffer (166 mM ­NaCO3H pH 8), or 10 μg of Western blotting either T22-PE24-H6 or T22-DITOX-H6 every day up to T22-PE24-H6 and T22-DITOX-H6 cell death mecha- 8 doses. Animal body weight and tumor size were meas- nisms were further studied by western blotting (WB). For ured with a caliper (tumor volume = ­width2 x length/2) that, 22A-CXCR4+ and 74B-CXCR4+ cells were treated through the time course of the treatment. Animals were with either of the nanotoxins (5 nM for T22-PE24-H6 in euthanized 48 h after the last dose, when tumors were 22A-CXCR4+ cell line, 50 nM in the rest of conditions) weighted and collected together with different organs for for different times (15 h, 24 h, and 48 h). In the zVAD later analysis. Plasma was also obtained by centrifugation conditions, the pan-caspase inhibitor (Calbiochem) was of total blood, extracted from the animals by intracardiac added at 100 μM and incubated for 2 h before nanotoxin puncture. treatment. WB was also used to evaluate the GSDME To evaluate the possible long-term toxicity of the expression in two patient-derived cultures (575 and 909). nanotoxins treatment, four-week-old female non-tumor Cells were lysed in RIPA buffer (Sigma) containing pro- bearing Swiss nude mice (NU (Ico)-Foxn1nu) were teinase inhibitors (Roche) and phosphatase inhibitors intravenously administered buffer or 10 μg of either (Roche). The protein extracts (50 μg) were subjected to T22-PE24-H6 or T22-DITOX-H6 daily up to 8 doses, SDS-PAGE and transferred to a nitrocellulose blotting similarly to the antitumor effect experiment. Animals membrane (GE Healthcare life sciences). After blockage were weighted twice a week during the study. After the with 5% skim milk in TBS-T for 1 h at room temperature, last dose, blood samples were collected from the tail membranes were incubated overnight at 4 °C with the every week to assess cell blood count (CBC). Animals primary antibodies: anti-human caspase-3 (1:1000, BD were euthanized 4 weeks after the end of the treatment Biosciences), cleaved caspase-3 (1:1000, Cell Signaling), and different organs were collected for histopathological PARP (1:2000, BD Biosciences), GSDME (1:1000, abm), analysis. and α/β-tubulin (1:1000, Cell Signaling). After washing with TBS-T to remove nonspecific antibody binding, Histopathology, DAPI staining, and immunohistochemical membranes were incubated with the corresponding sec- analysis ondary antibodies (1:10,000, Jackson Immune Research) 4 μm paraffin-embedded sections obtained from tumor for 1 h at room temperature. Finally, membranes were patient samples collected at the Hospital de la Santa Creu SuperSignal™ West Pico Chemiluminescent Substrate further washed with TBS-T and visualized with the i Sant Pau, as well as tumors and organs extracted from and SuperSignal™ West Femto Maximum Sensitivity the animals were used to performed different histopatho- logical analysis. Organ sections were stained with H&E Substrate (Thermo Scientific) using the ChemiDoc XRS+ and analyzed by two independent observers (one section imaging system (Biorad). Pro-caspase-3, cleaved cas- of the whole organ/tumor). Cell death in tumor tissues was pase-3, cleaved PARP, and cleaved GSDME levels were assessed by DAPI staining, paraffin-embedded sections quantified using Fiji, ImageJ software. The densitomet- were dewaxed, rehydrated, and permeabilized with 0.5% ing medium (ProLong™ Gold Antifade Mountant, Thermo ric analysis was performed by dividing the value of each Triton X-100. Then, slides were stained with DAPI mount- nanotoxin-treated sample by the buffer-treated sample and normalized by the loading control (α/β-tubulin). All Scientific) and visualized by fluorescence microscopy. Rep- experiments were performed at least in triplicate. resentative pictures were taken using an Olympus DP73 digital camera and the number of dead cells was quantified In vivo experiments by counting the number of condensed nuclei per 10 high- Four-week-old female Swiss nude mice (NU (Ico)- power fields (magnification 400x). Immunohistochemical Foxn1nu) weighing 18–25 g were purchased from Charles (IHC) staining of animal tumors was performed to study River (France). Animals were housed in a specific patho- the cytotoxic effect of the nanotoxin treatment. CXCR4 gen-free (SPF) environment with sterile food and water (1:200, Abcam. Retrieval pH high, Dako) and F4/80 (ready ad libitum. All animal experiments were approved by to use, Dako) IHC were performed in a DAKO Autostainer the Hospital de la Santa Creu i Sant Pau Animal Ethics Link48 following the manufacturer’s instructions. Similarly, Committee. GSDME IHC staining (1:300, Abcam. Retrieval pH high, Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 5 of 17 Dako) was assessed in patient tumor samples. Representa- both nanotoxins, ranging between 1 and 5 nM, further sup- tive images were captured using an Olympus DP73 digital porting T22-PE24-H6 and T22-DITOX-H6 potent cyto- camera and processed with the Olympus CellD Imaging toxicity (Fig. 1A-D). On the other hand, neither 22A mock 3.3 software. CXCR4 and F4/80 expression levels in tumors nor 74B mock (CXCR4 negative cells) cell viability was were quantified as mean gray values using Fiji, ImageJ altered upon nanoparticle exposure, suggesting a CXCR4- software. dependent cytotoxic effect (Fig. 1A-D). T22-PE24-H6 and T22-DITOX-H6 CXCR4-dependent Toxicity analyses in plasma and total blood samples cytotoxicity was further corroborated by pre-incubat- To further evaluate the toxicity of T22-PE24-H6 and ing both 22A-CXCR4+ and 74B-CXCR4+ cells with the T22-DITOX-H6 treatment in mice, plasma glutamic CXCR4 antagonist AMD3100, 1 h prior to the addition oxaloacetic transaminase (GOT) and glutamic pyruvic of the nanotoxins. Treatment with AMD3100 blocked transaminase (GPT) enzyme activities, as well as creatinine nanotoxin binding to CXCR4, leading to a practically and uric acid levels were assessed in plasma samples using complete remission of their cytotoxic effect in both commercial kits (Roche) in a COBAS 6000 autoanalyzer ­CXCR4+ cell lines (Fig. 1E and F). (Roche). Thus, both T22-PE24-H6 and T22-DITOX-H6 display Cell blood count (CBC) from buffer and nanotoxin- a potent CXCR4-dependent cytotoxic effect in vitro in treated animals was analyzed using the Mindray BC-5000 both HNSCC cell lines. Vet hematology analyzer. T22‑PE24‑H6 and T22‑DITOX‑H6 nanotoxins induce Statistical analysis pyroptosis in HNSCC cell lines Data was represented as mean ± Standard error (SEM). The aforementioned potent CXCR4-dependent cytotoxic Statistical analyses were performed using the GraphPad effect prompted us to investigate the mechanism of cell Prism 5 software (GraphPad Software, San Diego, Cali- death induced by both T22-PE24-H6 and T22-DITOX- fornia USA). Results were analyzed by Student t-test. Dif- H6 nanotoxins. Interestingly, upon nanoparticle expo- ferences were considered statistically significant when sure, we observed a balloon-like morphology of the p-values < 0.05. cells, which clearly differ from classic apoptotic blebbing (Fig. 2A). These swelling cells were especially noticeable Results in the 22A-CXCR4+ cultures treated with either of the T22‑PE24‑H6 and T22‑DITOX‑H6 exhibit a potent two nanotoxins, although they could also be observed CXCR4‑dependent cytotoxicity in HNSCC cell lines in the 74B-CXCR4+ cell line (Fig. 2A). Remarkably, this T22-PE24-H6 and T22-DITOX-H6 cytotoxic effect was balloon-like shape is a characteristic of pyroptotic cell evaluated in vitro in two HNSCC cell lines, 22A mock and morphology. 22A-CXCR4+; and 74B mock and 74B-CXCR4+. Both 22A In addition to these observations, we performed an mock and 74B mock cell lines were negative for the recep- Annexin V/ Propidium iodide (PI) assay by flow cytom- tor, whereas 22A-CXCR4+ and 74B-CXCR4+ displayed etry to further study the type of cell death induced by a strong CXCR4 membrane expression, as it has been T22-PE24-H6 and T22-DITOX-H6. In classic apoptosis, already studied by flow cytometry and immunocytochem- cells first undergo an early apoptosis phase (Annexin ­V+/ istry in previous work [22]. Cells were exposed to different PI−), followed by a late apoptosis phase (Annexin ­V+/ concentrations of either of the two nanotoxins (0–50 nM) PI+), characterized by plasma membrane rupture and for 48 h before assessing their viability. Both T22-PE24- leakage. However, in this case we did not observe these H6 and T22-DITOX-H6 were able to induce cell death in phases, revealing a lytic type of cell death, as Annexin V/ 22A-CXCR4+ and 74B-CXCR4+ cells, which express high PI double-positive stained cells increased upon treatment levels of CXCR4 in their membranes (Fig. 1A-D). Remark- with both nanotoxins (Fig. 2B). ably, nanotoxins present a potent cytotoxic effect, as they To further elucidate the mechanism of cell death were capable of inducing cell death at low concentrations induced by both nanotoxins, we evaluated the LDH (nM range) (Fig. 1A-D). ­IC50 values were calculated for released from cells 48 h after nanoparticle treatment, (See figure on next page.) Fig. 1 T22-PE24-H6 and T22-DITOX-H6 CXCR4-dependent cytotoxic effect in HNSCC cell lines. A and B) T22-PE24-H6 (A) and T22-DITOX-H6 (B) cytotoxic effect (0-50 nM) after 48 h of exposure in 22A mock and 22A-CXCR4+ cell lines represented as percentage of cell viability and ­IC50 values. C and D) T22-PE24-H6 (C) and T22-DITOX-H6 (D) cytotoxic effect (0–50 nM) after 48 h of exposure in 74B mock and 74B-CXCR4+ cell lines represented as percentage of cell viability and ­IC50 values. E and F) AMD3100 blocking assay (1 μM) in 22A-CXCR4+ and 74B-CXCR4+ cell lines treated with T22-PE24-H6 (E) (5 nM for 22A-CXCR4+ and 50 nM for 74B-CXCR4+) and T22-DITOX-H6 (F) (50 nM) for 48 h. ** p < 0.01; *** p < 0.001. Each column represents the mean value of three biological replicates. Statistical analysis performed by Student t-test. Error bars indicate SEM Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 6 of 17 Fig. 1 (See legend on previous page.) Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 7 of 17 which correlates with cell membrane disruption and to an inhibition of cell death, further corroborating the leakage. In agreement with the previous findings, nano- involvement of caspase-3/GSDME pathway in nano- toxin treated cells showed an increase in LDH release, toxin cytotoxicity. Remarkably, this is the first time that further corroborating a lytic form of cell death (Fig. 2C). we describe the activation of this pathway by T22-PE24- Finally, we studied different cell death markers by west- H6 and T22-DITOX-H6, which opens a novel avenue for ern blotting to determine the exact mechanisms of cell HNSCC treatment. death activated by T22-PE24-H6 and T22-DITOX-H6. We found an activation of caspase-3, PARP, and GSDME Nanotoxins repeated dosage potently inhibits tumor in both HNSCC cell lines treated with either of the two growth in a ­CXCR4+ subcutaneous HNSCC mouse model nanotoxins (Fig. 2D). 22A-CXCR4+, which has shown a in the absence of systemic toxicity greater sensitivity to both nanotoxins, presented cleaved The potent cytotoxic effect induced by both T22-PE24- caspase-3 at 15 h and 24 h, leading to the activation of H6 and T22-DITOX-H6 nanotoxins in the HNSCC cell PARP and GSDME also at 15 h and 24 h. On the other lines encouraged us to evaluate their antineoplastic effect hand, activation of caspase-3, PARP, and GSDME in the in vivo. For that, we generated a CXCR4-overexpress- 74B-CXCR4+ cell line was only observed at 48 h (Fig. 2D, ing subcutaneous mouse model. One week after the Supplementary Fig. 1). In the last years, different stud- implantation, animals were intravenously administered ies have determined that pyroptosis, a lytic form of cell buffer or 10 μg of either T22-PE24-H6 or T22-DITOX- death, can be triggered by GSDME N-terminal domain H6 daily up to 8 doses. Tumor volume and body weight which is activated by cleaved caspase-3, also responsible were measured on alternate days. Animals were eutha- for the proteolytic activation of PARP during apoptosis nized 48 h after the last dose, and tumors were weighed [23, 24]. Interestingly, we observed the simultaneous acti- and collected for later analysis, as well as different organs vation of both PARP and GSDME upon nanotoxin treat- (Fig. 4A). ment, which has also been described in other molecular Treatment with T22-PE24-H6 and T22-DITOX- therapies [25–28]. H6 nanotoxins clearly impaired tumor growth in the CXCR4-overexpressing subcutaneous tumors, as tumors T22‑PE24‑H6 and T22‑DITOX‑H6 activate caspase‑3/ from the buffer-treated animals reached bigger vol- GSDME‑mediated pyroptosis in HNSCC cell lines umes compared to their nanotoxin-treated counterparts In order to validate the previous findings which suggested (Fig. 4B). Remarkably, treatment with T22-DITOX-H6 that the caspase-3/GSDME pathway would be responsi- practically inhibited tumor growth, as tumor volumes ble for nanotoxin cytotoxic effect, we exposed the cells to did not significantly vary throughout the treatment. Con- the pan-caspase inhibitor zVAD prior to the nanotoxin sequently, tumor weight at the endpoint of the experi- treatment. Pre-treatment with zVAD clearly abrogated ment was significantly higher in the tumors derived from balloon-like morphology of the nanotoxin treated cells, buffer-treated mice compared to the nanotoxin-treated showing a decrease in cell swelling (Fig. 3A). Conse- animals, especially the ones treated with T22-DITOX-H6 quently, cell viability was also protected by pre-treating (Fig. 4C). In addition, nanotoxin treatment did not affect the cell with the pan-caspase inhibitor, further indicat- animal body weight in the time course of the experiment, ing a caspase-dependent cell death mechanism (Fig. 3B). suggesting a lack of systemic toxicity for the treatment In agreement, zVAD pre-treatment led to a reduction of (Fig. 4 D). LDH release in nanotoxin treated cells (Fig. 3C). Moreo- Tumor histology was also studied to further evaluate ver, western blotting analysis revealed that as expected, nanotoxin antineoplastic effect. Tumors after nanotoxin caspase-3 proteolytic activation was inhibited by zVAD repeated treatment maintained their undifferentiated pre-treatment. Hence, PARP and GSDME activa- phenotype (Supplementary Fig. 3A). CXCR4 expression tion were also abolished in the zVAD pre-treated cells in tumor tissue was also maintained upon nanotoxin (Fig. 3D, Supplementary Fig. 2). Thus, pre-treatment treatment, as no differences in percentage of positive of the cells with zVAD prior to nanotoxin exposure led cells were observed between groups (Supplementary (See figure on next page.) Fig. 2 T22-PE24-H6 and T22-DITOX-H6 nanotoxins induce tumor cell pyroptosis. A Phase-contrast imaging of 22A-CXCR4+ and 74B-CXCR4+ cells treated with T22-PE24-H6 or T22-DITOX-H6 for 48 h exhibiting pyroptotic cell morphology (magnification 200x). B Flow cytometry analysis of 74B-CXCR4+ after 15 h, 24 h, and 48 h of exposure to T22-PE24-H6 or T22-DITOX-H6 stained with Annexin V-FITC and propidium iodide (PI). Percentage of stained cells is represented in the column graph. C LDH release from 22A-CXCR4+ and 74B-CXCR4+ exposed to either T22-PE24-H6 or T22-DITOX-H6 for 48 h. D Representative images of pro-caspase-3, cleaved caspase-3, PARP, GSDME, and tubulin immunoblotting in protein extracts from 22A-CXCR4+ and 74B-CXCR4+ cell lines treated with T22-PE24-H6 and T22-DITOX-H6 for 15 h, 24 h, and 48 h. * p < 0.05. Each column represents the mean value of three biological replicates. Statistical analysis performed by Student t-test. Error bars indicate SEM Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 8 of 17 Fig. 2 (See legend on previous page.) Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 9 of 17 Fig. 3B). Besides, nanotoxin treatment induced cell death (Supplementary Fig. 4A). Importantly, non-tumor bear- in the tumor cells as detected by condensate DNA (DAPI ing mice were utilized, as the lack of tumor and conse- staining) (Fig. 5A). The number of DAPI stained cells was quently the absence of nanotoxin uptake by tumor tissue, found to be higher in tumor treated with either of the will potentially increase nanotoxin concentration in the two nanotoxins as compared to the control group, sup- bloodstream, thus providing a better evaluation of their porting nanotoxin antitumor effect (Fig. 5A). Last but potential off-target toxicity. Remarkably, none of the not least, we wanted to assess immune cell recruitment nanotoxins, neither T22-PE24-H6 nor T22-DITOX- to the tumor site, as it is described that tumor leakage H6, shortly after their administration at repeated doses from pyroptotic cells enhances the number and activity or 4 weeks after the end of the treatment, induced any of tumor-associated macrophages (TAMs) that phago- changes in animal body weight (Supplementary Fig. 4B). cyte tumor cells [23]. For that, we performed a F4/80 In addition, cell blood count (CBC) analyses were per- immunohistochemistry, a well-known macrophage cell formed weekly to assess treatment derived toxicity. No marker, finding an increase in the percentage of tumor- differences between buffer and nanotoxin-treated ani- infiltrating macrophages in nanotoxin-treated tumors as mals were observed in terms of white blood cells (white compared to the buffer-treated ones (Fig. 5B). Altogether, blood cell count (WBC), neutrophils, lymphocytes, these results corroborate nanotoxin potent antitumor monocytes, eosinophils, and basophils), red blood cells effect, suggesting that the activation of pyroptosis in (red blood cell count (RCB), hemoglobin (HGB), hema- tumor cells also enhances immune cell recruitment to the tocrit (HCT), mean cell volume (MCV), mean corpuscu- tumor site, further contributing to the antitumor effect. lar hemoglobin (MCH), mean corpuscular hemoglobin Importantly, no off-target toxicity was observed upon concentration (MCHC), and red blood cell distribution nanotoxin treatment, neither in the T22-PE24-H6 nor in width (RDW)), or platelets (platelet count, mean plate- the T22-DITOX-H6 group. We assessed nanotoxin toxic- let volume (MPV), platelet distribution width (PDW), ity by histopathology (H&E staining) in liver and kidneys and plateletcrit (PCT)) (Supplementary Fig. 5). Moreo- tissue, organs involved in the metabolism and elimina- ver, these values were between the normal range for a tion of drugs. As it could be observed by two independ- healthy Swiss nude mouse. Lastly, no histopathological ent observers, no histological alterations were detected alterations were observed in liver, kidneys, spleen, or in either of the two organs, which presented their nor- bone marrow, implying that nanotoxin treatment did not mal architecture and morphology (Fig. 6A). Moreover, to induce any long-term toxic effects in the animals (Sup- further corroborate treatment’s lack of toxicity, hepatic plementary Fig. 4C). Thus, the lack of change in mouse and renal function were studied by hepatic transami- body weight, together with unaltered CBCs, and the nases activity, as well as creatinine and uric acid levels in absence of histological alterations in non-tumor tissues, plasma. Results showed no statistically significant differ- indicate a lack of long-term toxicity by the nanotoxins ences in transaminases activity, nor in creatinine or uric when administered at a dosage that achieves a highly sig- acid concentrations in plasma between control and nano- nificant antitumor effect. toxin treated animals (Fig. 6B-E). In addition, these val- ues were between the normal range of a healthy animal GSDME and CXCR4 expression in HNSCC and clinical [29]. Thus, T22-PE24-H6 and T22-DITOX-H6 nanotoxin implications treatment at the chosen administration conditions, does Taking into consideration these findings, we wanted to not induce off-target toxicity in the ­CXCR4+ HNSCC evaluate the clinical relevance of GSDME activation in subcutaneous mouse model. HNSCC patients. Given that GSDME presents a tumor Moreover, to evaluate the long-term toxicity induced suppressive role, it has been reported that GSDME inac- by nanotoxin treatment, animals received the same dos- tivation is a strategy developed by cancer cells to avoid age as in the antitumor effect experiment, but were fur- cell death. However, both H­ PV− HNSCC cells lines used ther maintained for 4 weeks after the end of the treatment in this study expressed GSDME, and this protein was able (See figure on next page.) Fig. 3 zVAD pre-treatment of 22A-CXCR4+ and 74B-CXCR4+ cells shows T22-PE24-H6 and T22-DITOX-H6 activation of caspase-3/GSDME-mediated pyroptosis. A Phase-contrast imaging of 22A-CXCR4+ and 74B-CXCR4+ cell lines with and without zVAD pre-treatment (100 μM) 1 h prior to the addition of T22-PE24-H6 or T22-DITOX-H6 (magnification 200x). zVAD clearly inhibits pyroptotic cell morphology in both cell lines. B Cell viability of 22A-CXCR4+ and 74B-CXCR4+ cells either pre-treated or not with zVAD before the addition of T22-PE24-H6 and T22-DITOX-H6 nanotoxins. C LDH release from 22A-CXCR4+ and 74B-CXCR4+ treated with T22-PE24-H6 or T22-DITOX-H6 for 48 h, with and without zVAD pre-treatment. D Representative images of pro-caspase-3, cleaved caspase-3, PARP, GSDME, and tubulin western blots of samples from 22A-CXCR4+ and 74B-CXCR4+ cell lines exposed to the inhibitor zVAD before nanotoxin treatment for 48 h. * p < 0.05; ** p < 0.01. Each column represents the mean value of three biological replicates. Statistical analysis performed by Student t-test. Error bars indicate SEM Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 10 of 17 Fig. 3 (See legend on previous page.) Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 11 of 17 Fig. 4 T22-PE24-H6 and T22-DITOX-H6 antitumor effect in a ­CXCR4+ HNSCC subcutaneous mouse model. A) Schematic representation of the experimental design followed in this study. B Variation of the tumor volume in each group (buffer, T22-PE24-H6, and T22-DITOX-H6) in the time course of the experiment. C Tumor weight registered at the end point of the experiment for the three experimental groups. D Body weight of buffer, T22-PE24-H6, and T22-DITOX-H6 treated animals along the study. * p < 0.05; ** p < 0.01; *** p < 0.001; n = 10 per group (total animal number 30). Statistical analysis performed by Student t-test. Error bars indicate SEM to exert its activity upon nanotoxin treatment. Moreo- Interestingly, high DFNA5 expression in tumor tissue ­ XCR4+ ver, two patient-derived samples obtained from C correlated with worse overall survival (OS) (Supple- tumors maintained in vitro also expressed the protein mentary Fig. 7C). These results prompted us to perform (Fig. 7A). Thus, to further investigate GSDME expres- a IHC analysis of GSMDE expression in a cohort of 17 sion in HNSCC, we conducted an analysis using data HNSCC patients (Fig. 7B). First, we assessed the CXCR4 from The Cancer Genome Atlas (TCGA) with the UAL- expression in the HNSCC samples finding that 88.2% CAN analysis software [30]. First, we found that CXCR4 were positive for the receptor (Fig. 7C). Moreover, 94.1% is overexpressed in tumoral tissue compared to healthy of these patient samples presented positive GSDME tissue (Supplementary Fig. 6). Similarly, DFNA5 (gene staining, independently of their prognosis (Fig. 7C). encoding GSDME) was also overexpressed in tumors Importantly, 87.5% of the patient samples expressing compared to non-tumor tissue (Supplementary Fig. 7A). GSDME were also positive for CXCR4 immunostaining Moreover, DFNA5 was more expressed in H ­ PV− tumors (Fig. 7C). This subset of patients could potentially benefit + compared to the ­HPV ones (Supplementary Fig. 7B). from a CXCR4-targeted treatment capable of activating Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 12 of 17 Fig. 5 Cytotoxic effect of T22-PE24-H6 and T22-DITOX-H6 in tumor tissue. A Detection of dead cells by condensate DNA staining (DAPI) in the buffer, T22-PE24-H6 or T22-DITOX-H6 treated. Quantification of the number of DAPI positive stained cells in tumor tissue represented as fold-change respect to the buffer. B IHC analysis of tumor infiltrated macrophages in tumors from buffer, T22-PE24-H6 or T22-DITOX-H6 treated animals detected by F4/80 immunostaining. Quantification of the percentage of F4/80 positive stained cells in tumor samples from each group. Scale bars = 50 μm. F4/80 expression was quantified as mean gray value and represented as mean ± SEM. ** p < 0.01; *** p < 0.001. Statistical analysis performed by Student t-test GSDME, such as the nanotoxins T22-PE24-H6 and Importantly, pre-treatment with the pan-caspase inhib- T22-DITOX-H6. Thus, we found that GSDME is widely itor zVAD blocked all the above described pyroptotic expressed in HNSCC patients highlighting the relevance features. Thus, the selective bacterial exotoxin release of GSDME-dependent pyroptosis in disease course and in ­CXCR4+ HNSCC cells activate caspase-3/GSDME opening a novel avenue for future treatments. switch from apoptosis to pyroptosis, as described for different untargeted and targeted drugs in other cancer Discussion types [24, 25, 32–34]. In this study (which rationale is summarized in Sup- In addition, and according to the potent cytotoxic- plementary Fig. 8), we report for the first time that tar- ity (low nanomolar I­C50) observed in vitro, repeated geted toxin delivery to CXCR4-overexpresing ­(CXCR4+) intravenous administration of each nanotoxin in human HNSCC cells, achieved by the T22-PE24-H6 a subcutaneous ­ C XCR4+ HNSCC mouse model, and T22-DITOX-H6 nanotoxins, activates caspase-3/ induced a potent blockade of tumor growth, flatten- GSDME-dependent pyroptosis in ­CXCR4+ cancer cells. ing the growth curves, especially for the T22-DITOX- Consequently, nanotoxin treatment leads to a potent H6 nanotoxin, in the absence of systemic toxicity or blockade of tumor growth in a HNSCC model, without adverse effects. This high therapeutic window is most inducing systemic toxicity. Importantly, activation of this likely due to the strict CXCR4-dependent killing cell death mechanism alternative to apoptosis is expected achieved by the nanostructured toxins. T22-PE24- to overcome the low response rate to chemotherapy and H6 and T22-DITOX-H6 internalize exclusively in radiotherapy in HNSCC patients [5, 31]. ­C XCR4+ cells, releasing their toxin domains by We meticulously demonstrate that both nanotox- furin cleavage upon cell internalization, leading to ins activate caspase-3/GSDME-dependent pyroptosis EF-2 inactivation and protein synthesis inhibition in ­CXCR4+ human HNSCC cell lines. First, we have that results in cell death [19]. This effect was dem- reported that, upon nanotoxin treatment, HNSCC cell onstrated in vitro in two different HNSCC cell lines, lines displayed distinctive pyroptotic features, includ- showing that only ­ C XCR4+ cells are killed by the ing balloon-like morphology, increase in LDH release, nanotoxins and also by the demonstration of a com- and lytic cell death. These observations associated plete cell death blockade by AMD3100 (CXCR4 with caspase-3 activation followed by GSDME cleav- antagonist) exposure prior to nanotoxins treatment. age to generate the pyroptotic effector pore inductor Consistently, our previous work supports the selec- GSDME N-terminus, as detected by immunoblotting. tive biodistribution of the nanotoxins to the high Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 13 of 17 Fig. 6 T22-PE24-H6 and T22-DITOX-H6 toxicity assessment. A Histopathological analysis by H&E staining in liver and kidneys samples from buffer, T22-PE24-H6 or T22-DITOX-H6 treated animals. B, C, D, and E Oxaloacetic transaminase (GOT) (B), and glutamic pyruvic transaminase (GPT) (C) enzyme activities, as well as creatinine (D) and uric acid (E) levels in plasma samples from buffer, T22-PE24-H6 or T22-DITOX-H6 treated animals. Scale bars = 100 μm and 50 μm (zoom in). Statistical analysis performed by Student t-test. Error bars indicate SEM CXCR4-overexpressing tumor tissue, avoiding the accu- ­ XCR4+ target cell internalization, which [18] regarding C mulation and off-target toxicity in normal organs with exploits the CXCR4 overexpression in tumor compared low or negligible CXCR4 expression. In this context, we to normal tissue. have recently reported this high tumor selectivity using Moreover, innate antitumor immunity may have con- a CXCR4-targeted fluorescent nanocarrier, that dis- tributed to the anticancer activity based on the dramatic plays the exact self-assembling nanoparticle structure of increase in tumor-infiltrating macrophages (TAM) found the nanotoxins, in HNSCC, colorectal cancer [35], and in nanotoxin-treated animals. It has been reported that lymphoma [36] models. Their size above the 7 nm renal GSDME activation enhances antitumor effect by stimu- filtration cut-off allows a high recirculation in the blood- lating TAM phagocytosis, as well as NKs and C ­ D8+ T stream, while the multivalency derived from the display lymphocytes activation [6, 23, 37]. Although preliminary, of multiple T22 peptide domains enables superselectivity these findings may suggest an activation of the pyroptotic Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 14 of 17 Fig. 7 GSDME expression levels in HNSCC patient samples. A GSDME expression analysis by western blotting in two patient derived tumor samples (909 and 575) maintained in vitro. Phase-contrast imaging of 909 and 575 cultures showing their morphology (magnification 400x). B Representative IHC images of GSDME expression in HNSCC patient tumor samples presenting different levels of expression. C Representative IHC images of CXCR4 receptor expression in patient tumor samples. D Percentage of CXCR4 and GSDME positive and negative stained samples in the HNSCC patient cohort. Percentage of ­GSDME+ samples that were also positive for CXCR4 IHC. Scale bars = 200 μm pathway in vivo upon treatment, leading to an enhanced toxicities are induced by the activation of GSDME immune cell recruitment to the tumor site. mediated pyroptosis in non-tumor cells, as these Importantly, different studies have described that agents lack tumor targeting capacity [24, 38, 39]. conventional chemotherapy threatening off-target In contrast, our CXCR4-targeted nanotoxins, Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 15 of 17 administered at a repeated dosage schedule, did not So far, most of the targeted drugs to treat resist- induce acute systemic toxicity (at the end of treat- ant HNSCC have failed at early stages. Immunotoxins ment) nor long-term toxicity (4 weeks after treat- based on Pseudomonas exotoxin that target mesothelin ment), in the HNSCC mouse model, displaying in HNSCC stopped development in early clinical trials undetectable markers of toxicity in plasma, unaltered [50]. Similarly, ADCs targeting different surface mark- cell blood count and lack of histological alterations ers (EGFR, c-MET, HER2, etc.) conjugated to the PBD in non-tumor organs, at a dosage that achieves a toxin or the microtubule inhibitor Auristatin, did not potent antitumor effect. Thus, a strict targeted drug progress due to toxicity concerns [14]. Only the EGFR delivery to cancer cells is crucial to prevent GSDME inhibitor cetuximab and the PD-1/PD-L1 immune activation in healthy tissues. checkpoint inhibitors pembrolizumab, nivolumab, and Regarding the clinical relevance of our results, it is durvalumab, improve survival in recurrent or meta- important to highlight that apoptosis avoidance is the static HNSCC [51]. We believe that T22-PE24-H6 and main mechanism of resistance to radiotherapy, cispl- T22-DITOX-H6 nanotoxins present important features atin, 5-flourouracil or docetaxel/paclitaxel, leading to that will exceed the performance of current immuno- treatment failure and patient death [3, 5, 6, 20]. More- toxins and ADCs. These nanotoxins are formed by self- over, CXCR4 overexpression, in HNSCC and other assembly of multiple monomers, displaying numerous neoplasias, also associates with resistance, relapse targeting ligands that confer superselectivity, and the and metastatic potential [7, 8, 40]. Thus, our nano- ability to incorporate multiple cytotoxic domains into toxin approach aims to overcome tumor resistance a single nanotoxin (Supplementary Fig. 8). These facts and increase cure rates in HNSCC, by switching cell contrast with immunotoxins or ADCs, that display only death induction from apoptosis to caspase-3/GSDME- one ligand per molecule, a lower cytotoxic payload, and dependent pyroptosis, as pursued in other cancer types show drug leakage during circulation inducing off-tar- [34, 37, 41]. get effects [12, 13, 15–17]. Interestingly, a large percent of HNSCC patients could be candidates to targeted nanotoxin therapy since in our 17-patient cohort, 88.2% were positive for the CXCR4 Conclusions receptor, 94% of them expressed GSDME in tumor tis- In summary, the activation of caspase-3/GSDME- sue, whereas 87% of G ­ SDME+ tumors also co-expressed dependent pyroptosis by nanostructured toxins tar- CXCR4. Moreover, we have also studied CXCR4 expres- geting CXCR4 opens a novel and virtually unexplored sion in two different cell cultures derived from HNSCC therapeutic approach for HNSCC or other cancer patient tumor samples, named as 575 and 909. Interest- types. Thus, T22-PE24-H6 and T22-DITOX-H6 ingly, although both patient samples presented a strong nanotoxin treatments may turn sensitive the recur- CXCR4 expression, when cultured in vitro, 575 and 909 rent or metastatic HNSCC because of their ability to cell cultures lost their CXCR4 expression with succes- trigger a cell death mechanism alternative to apopto- sive culture passages. Remarkably, when re-inoculated sis, since apoptosis blockade is the main mechanism in vivo in immunodeficient mice, the generated tumors of resistance to currently used chemotherapy and expressed again CXCR4 (Supplementary Fig. 9). These radiotherapy in HNSCC patients [5, 6]. Addition- results demonstrate that CXCR4 expression is downreg- ally, the nanotoxin switch from non-inflammatory ulated in vitro, suggesting that CXCR4 regulation plays apoptosis to the inflammatory pyroptosis may also an important role in tumor progression in vivo. In addi- engage immune cells that could enhance anti-tumor tion, our TGCA analysis of HNSCC samples showed immunity. that GSDME overexpression in tumor compared to normal tissue, correlates with worse OS, which may be Abbreviations explained by the reported lack of tumor infiltrated lym- HNSCC: Head and neck squamous cell carcinoma; CXCR4: Chemokine recep‑ phocytes in ­GSDME+ HNSCC tumors by other authors tor 4; GSDME: Gasdermin E; LDH: Lactate dehydrogenase; ADC: Antibody- [42]. Consistently with our results, GSDME overexpres- drug conjugate; IL-2: Interleukin-2; PBD: Pyrrolobenzodiazepine; HPV: Human papilloma virus; DMEM: Dulbecco’s Modified Eagle’s Medium; FBS: Fetal Bovine sion in tumor tissue has also been reported in different Serum; PI: Propidium iodide; WB: Western Blotting; SPF: Specific pathogen- cancer types [25, 32, 43, 44], which contrasts with the free; IHC: Immunohistochemistry; GOT: Glutamic oxaloacetic transaminase; assumption that GSDME is silenced in tumors given its GPT: Glutamic pyruvic transaminase; SD: Standard deviation; SEM: Standard error; TAM: Tumor-associated macrophage; H&E: Hematoxylin-eosin; WBC: tumor suppressor role [23, 24, 45–49]. In any case, it is White blood cell count; RCB: Red blood cell count; HGB: Hemoglobin; HCT: clear that CXCR4 and GSDME markers could be used Hematocrit; MCV: Mean cell volume; MCH: Mean corpuscular hemoglobin; to select HNSCC patients that might benefit from our MCHC: Mean corpuscular hemoglobin concentration; RDW: Red blood cell distribution width; MPV: Mean platelet volume; PDW: Platelet distribution nanotoxin therapy. width; PCT: Plateletcrit. Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 16 of 17 Supplementary Information Targeted delivery of therapeutic molecules to CXCR4 cells, and in PCT/ EP2018/061732, covering Therapeutic Nanostructured Proteins. All other The online version contains supplementary material available at https://​doi.​ authors report no conflicts of interest in this work. org/​10.​1186/​s13046-​022-​02267-8. Author details 1 Additional file 1. Institut d’Investigació Biomèdica Sant Pau (IIB-Sant Pau), Sant Quintí, 77, 08041 Barcelona, Spain. 2 Institut de Recerca contra la Leucèmia Josep Carreras, 08025 Barcelona, Spain. 3 CIBER de Bioingeniería, Biomateriales y Nanome‑ Acknowledgements dicina (CIBER-BBN), Monforte de Lemos 3‑5, 28029 Madrid, Spain. 4 Depart‑ The authors would like to thank Clara Seira (IIB-Sant Pau, Barcelona) for her ment of Pathology, Hospital de la Santa Creu i Sant Pau, Sant Quintí, 89, technical support, and Dr. Gemma Vilahur (IIB-Sant Pau, Barcelona) for letting 08041 Barcelona, Spain. 5 Institut de Biotecnologia i de Biomedicina, Universitat us use the Mindray BC-5000 Vet hematology analyzer. Autònoma de Barcelona, 08193 Bellaterra, Spain. 6 Departament de Genètica i de Microbiologia, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain. 7 Authors’ contributions Department of Otorhinolaryngology, Hospital de la Santa Creu i Sant Pau, We declare that all authors made fundamental contributions to the manuscript. Universitat Autònoma de Barcelona, Sant Quintí, 89, 08041 Barcelona, Spain. 8 ERB, EV, RM, LAC, and XL designed the study. ERB, IAS, UU, and IC contributed to the Department of Surgery, Hospital de la Santa Creu i Sant Pau, Universitat development of the methodology. ERB, IAS, PA, NS, and LSG acquired the data and Autònoma de Barcelona, Sant Quintí, 89, 08041 Barcelona, Spain. 9 Institut de carried out the experiments. ERB and AG carried out the histopathological analysis. Biotecnologia i de Biomedicina and Departament de Genètica i de Microbio‑ The analysis and interpretation of data was performed by ERB, MQ, XL, AV, EV, RM, logia, Universitat Autònoma de Barcelona and CIBER, Bellaterra, Barcelona, and LAC. ERB prepared the manuscript. UU, IC, AV, EV, RM, and LAC performed the Spain. 10 Institut d’Investigacions Biomèdiques Sant Pau, Hospital de Sant Pau, manuscript reviews and revisions. The study was supervised by EV, RM, and LAC. All CIBER and Josep Carreras Research Institute, 08041 Barcelona, Spain. 11 Institut authors revised the manuscript and approved the final version. d’Investigacions Biomèdiques Sant Pau, Hospital de Sant Pau and Josep Car‑ reras Research Institute, 08041 Barcelona, Spain. Funding This work was supported by grants PI21/00150, PI18/ 00650, PIE15/00028, Received: 23 September 2021 Accepted: 19 January 2022 PI15/00378 and EU COST Action CA 17140 to Ramon Mangues, PI19/01661 to Xavier León, and PI17/00584 to Miquel Quer, founded by Instituto de Salud Carlos III (ISCIII, Co-funding from FEDER, Spain). Grants BIO2016–76063-R, AEI/ FEDER, UE to Antonio Villaverde and PID2019-105416RB-I00/AEI/https://​doi.​ org/​10.​13039/​50110​00110​33 to Esther Vázquez, founded by Agencia Estatal References de Investigación (AEI) and Fondo Europeo de Desarrollo Regional (FEDER) 1. Sacco AG, Cohen EE. Current treatment options for recurrent or (Spain). CIBER-BBN (Spain) grants CB06/01/ 1031 and 4NanoMets to Ramon metastatic head and neck squamous cell carcinoma. J Clin Oncol. Mangues, VENOM4CANCER to Antonio Villaverde, NANOREMOTE to Esther 2015;33:3305–15. Vázquez, and NANOSCAPE to Ugutz Unzueta. AGAUR 2017- SGR-865 (Spain) to 2. Johnson DE, Burtness B, Leemans CR, Lui VWY, Bauman JE, Grandis JR. Ramon Mangues, and 2017SGR-229 (Spain) to Antonio Villaverde. Josep Car‑ Head and neck squamous cell carcinoma. Nat Rev Dis Prim. 2020;6:92. reras Leukemia Research Institute (Spain) P/AG to Ramon Mangues. Elisa Rioja- 3. Picon H, Guddati AK. Mechanisms of resistance in head and neck cancer. Blanco and Laura Sánchez-García were supported by a predoctoral fellowship Am J Cancer Res. 2020;10:2742–51. from AGAUR (2020FI_B2 00168 and 2018FI_B2_00051 (Spain)) co-funded by 4. López-Verdín S, Lavalle-Carrasco J, Carreón-Burciaga RG, Serafín-Higuera European Social Fund (ESF investing in your future). Lorena Alba-Castellón N, Molina-Frechero N, González-González R, et al. Molecular markers of was supported by a postdoctoral fellowship from AECC (Spanish Associa‑ anticancer drug resistance in head and neck squamous cell carcinoma: a tion of Cancer Research, Spain). Antonio Villaverde was granted an Icrea literature review. Cancers (Basel). 2018;10(10):376. Academia Award (Spain). Ugutz Unzueta was also supported by Grant PERIS 5. Kanno Y, Chen C-Y, Lee H-L, Chiou J-F, Chen Y-J. Molecular mechanisms SLT006/17/00093 from la Generalitat de Catalunya (Spain) and Miguel Servet of chemotherapy resistance in head and neck cancers. Front Oncol. fellowship (CP19/00028) from Instituto de Salud Carlos III (Spain) co-funded by 2021;11:640392. European Social Fund (ESF investing in your future). The toxicity studies have 6. Raudenská M, Balvan J, Masařík M. Cell death in head and neck cancer been performed in the ICTS-141007 Nanbiosis Platform, using its CIBER-BBN pathogenesis and treatment. Cell Death Dis. 2021;12:192. Nanotoxicology Unit (http://​www.​nanbi​osis.​es/​portf​olio/​u18-​nanot​oxico​logy-​ 7. León X, Diez S, García J, Lop J, Sumarroca A, Quer M, et al. Expression unit/). Protein production has been partially performed by the ICTS “NANBIO‑ of the CXCL12/CXCR4 chemokine axis predicts regional control in SIS”, more specifically by the Protein Production Platform of CIBER-BBN/ IBB head and neck squamous cell carcinoma. Eur Arch Otorhinolaryngol. (http://​www.​nanbi​osis.​es/​unit/​u1-​prote​in-​produ​ction-​platf​orm-​ppp/). 2016;273:4525–33. 8. De-Colle C, Menegakis A, Mönnich D, Welz S, Boeke S, Sipos B, et al. Availability of data and materials SDF-1/CXCR4 expression is an independent negative prognostic bio‑ The datasets used and/or analysed during the current study are available from marker in patients with head and neck cancer after primary radiochemo‑ the corresponding author on reasonable request. therapy. Radiother Oncol. 2018;126:125–31. 9. Pastan I, Hassan R, Fitzgerald DJ, Kreitman RJ. Immunotoxin therapy of cancer. Nat Rev Cancer England. 2006;6:559–65. Declarations 10. Alewine C, Hassan R, Pastan I. Advances in anticancer immunotoxin therapy. Oncologist. 2015;20:176–85. Ethics approval and consent to participate 11. Wu T, Zhu J. Recent development and optimization of pseudomonas HNSCC patient samples were obtained in accordance with the Hospital de la aeruginosa exotoxin immunotoxins in cancer therapeutic applications. Int Santa Creu i Sant Pau Institutional Review Board. Written informed consent Immunopharmacol. 2021;96:107759. was acquired from all the patients involved in this study. All animal experi‑ 12. Vallera DA, Kreitman RJ. Immunotoxins targeting B cell malignancy- ments were approved by the Hospital de la Santa Creu i Sant Pau Animal Progress and problems with immunogenicity. Biomedicines. 2018;7(1):1. Ethics Committee. 13. Kim J-S, Jun S-Y, Kim Y-S. Critical issues in the development of immuno‑ toxins for anticancer therapy. J Pharm Sci. 2020;109:104–15. Consent for publication 14. Perrotti V, Caponio VCA, Mascitti M, Lo Muzio L, Piattelli A, Rubini C, et al. Not applicable. Therapeutic potential of antibody-drug conjugate-based therapy in head and neck cancer: a systematic review. Cancers (Basel). 2021;13(13):3126. Competing interests 15. Beck A, Goetsch L, Dumontet C, Corvaïa N. Strategies and challenges for Antonio Villaverde, Esther Vázquez, Ugutz Unzueta, Ramon Mangues, and the next generation of antibody-drug conjugates. Nat Rev Drug Discov. Isolda Casanova are cited as inventors in PCT/EP2012/050513 covering 2017;16:315–37. Rioja‑Blanco et al. J Exp Clin Cancer Res (2022) 41:49 Page 17 of 17 16. Wolska-Washer A, Robak T. Safety and tolerability of antibody-drug conju‑ 39. Huang Z, Zhang Q, Wang Y, Chen R, Wang Y, Huang Z, et al. Inhibition of gates in Cancer. Drug Saf. 2019;42:295–314. caspase-3-mediated GSDME-derived pyroptosis aids in noncancerous 17. Masters JC, Nickens DJ, Xuan D, Shazer RL, Amantea M. Clinical toxicity of tissue protection of squamous cell carcinoma patients during cisplatin- antibody drug conjugates: a meta-analysis of payloads. Invest New Drugs. based chemotherapy. Am J Cancer Res. 2020;10:4287–307. 2018;36:121–35. 40. Chatterjee S, Behnam Azad B, Nimmagadda S. The intricate role of CXCR4 18. Liu M, Apriceno A, Sipin M, Scarpa E, Rodriguez-Arco L, Poma A, et al. in cancer. Adv Cancer Res. 2014;124:31–82. Combinatorial entropy behaviour leads to range selective binding in 41. Wu D, Wang S, Yu G, Chen X. Cell death mediated by the Pyroptosis ligand-receptor interactions. Nat Commun. 2020;11:4836. pathway with the aid of nanotechnology: prospects for Cancer therapy. 19. Sánchez-García L, Serna N, Álamo P, Sala R, Céspedes MV, Roldan M, et al. Angew Chem Int Ed Engl. 2021;60:8018–34. Self-assembling toxin-based nanoparticles as self-delivered antitumoral 42. Liu Z, Liu H, Dong Q, Li H, Zhang B, Liu Y, et al. Prognostic role of DFNA5 drugs. J Control Release. 2018;274:81–92. in head and neck squamous cell carcinoma revealed by systematic 20. Holohan C, Van Schaeybroeck S, Longley DB, Johnston PG. Cancer drug expression analysis: Research Square; 2020. https://​doi.​org/​10.​21203/​rs.3.​ resistance: an evolving paradigm. Nat Rev Cancer. 2013;13:714–26. rs-​114898/​v1. 21. Brenner JC, Graham MP, Kumar B, Saunders LM, Kupfer R, Lyons RH, et al. 43. Wu M, Wang Y, Yang D, Gong Y, Rao F, Liu R, et al. A PLK1 kinase inhibitor Genotyping of 73 UM-SCC head and neck squamous cell carcinoma cell enhances the chemosensitivity of cisplatin by inducing pyroptosis in lines. Head Neck. 2010;32:417–26. oesophageal squamous cell carcinoma. EBioMedicine. 2019;41:244–55. 22. Rioja-Blanco E, Arroyo-Solera I, Álamo P, Casanova I, Gallardo A, Unzueta 44. Zhou B, Zhang J-Y, Liu X-S, Chen H-Z, Ai Y-L, Cheng K, et al. Tom20 senses U, et al. Self-assembling protein nanocarrier for selective delivery of cyto‑ iron-activated ROS signaling to promote melanoma cell pyroptosis. Cell toxic polypeptides to CXCR4+ head and neck squamous cell carcinoma Res. 2018;28:1171–85. tumors. Acta Pharm Sin B. 2021; Available from: https://​www.​scien​cedir​ 45. Xia X, Wang X, Cheng Z, Qin W, Lei L, Jiang J, et al. The role of pyroptosis in ect.​com/​scien​ce/​artic​le/​pii/​S2211​38352​10039​19. cancer: pro-cancer or pro-“host”? Cell Death Dis. 2019;10. https://​doi.​org/​ 23. Zhang Z, Zhang Y, Xia S, Kong Q, Li S, Liu X, et al. Gasdermin E sup‑ 10.​1038/​s41419-​019-​1883-8. presses tumour growth by activating anti-tumour immunity. Nature. 46. Lage H, Helmbach H, Grottke C, Dietel M, Schadendorf D. DFNA5 (ICERE- 2020;579:415–20. 1) contributes to acquired etoposide resistance in melanoma cells. FEBS 24. Wang Y, Gao W, Shi X, Ding J, Liu W, He H, et al. Chemotherapy drugs Lett. 2001;494:54–9. induce pyroptosis through caspase-3 cleavage of a gasdermin. Nature. 47. Kim MS, Chang X, Yamashita K, Nagpal JK, Baek JH, Wu G, et al. Aberrant 2017;547:99–103. promoter methylation and tumor suppressive activity of the DFNA5 gene 25. Lu H, Zhang S, Wu J, Chen M, Cai MC, Fu Y, et al. Molecular targeted in colorectal carcinoma. Oncogene. 2008;27:3624–34. therapies elicit concurrent apoptotic and GSDME-dependent pyroptotic 48. Akino K, Toyota M, Suzuki H, Imai T, Maruyama R, Kusano M, et al. Identi‑ tumor cell death. Clin Cancer Res. 2018;24:6066–77. fication of DFNA5 as a target of epigenetic inactivation in gastric cancer. 26. Zhang X, Zhang P, An L, Sun N, Peng L, Tang W, et al. Miltirone induces Cancer Sci. 2007;98:88–95. cell death in hepatocellular carcinoma cell through GSDME-dependent 49. Kim MS, Lebron C, Nagpal JK, Chae YK, Chang X, Huang Y, et al. Methyla‑ pyroptosis. Acta Pharm Sin B. 2020;10:1397–413. tion of the DFNA5 increases risk of lymph node metastasis in human 27. Zhang Y, Yang J, Wen Z, Chen X, Yu J, Yuan D, et al. A novel breast cancer. Biochem Biophys Res Commun. 2008;370:38–43. 3′,5′-diprenylated chalcone induces concurrent apoptosis and GSDME- 50. Hassan R, Ho M. Mesothelin targeted cancer immunotherapy. Eur J dependent pyroptosis through activating PKCδ/JNK signal in prostate Cancer. 2008;44:46–53. cancer. Aging (Albany NY). 2020;12:9103–24. 51. Cramer JD, Burtness B, Le QT, Ferris RL. The changing therapeutic land‑ 28. Kong Y, Feng Z, Chen A, Qi Q, Han M, Wang S, et al. The natural flavonoid scape of head and neck cancer. Nat Rev Clin Oncol. 2019;16:669–83. Galangin elicits apoptosis, Pyroptosis, and autophagy in glioblastoma. Front Oncol. 2019;9:942. 29. Cañes L, Martí-Pàmies I, Ballester-Servera C, Alonso J, Serrano E, Briones Publisher’s Note AM, et al. High NOR-1 (neuron-derived orphan receptor 1) expression Springer Nature remains neutral with regard to jurisdictional claims in pub‑ strengthens the Vascular Wall response to angiotensin II leading to aneu‑ lished maps and institutional affiliations. rysm formation in mice. Hypertens (Dallas, Tex 1979). 2021;77:557–70. 30. Chandrashekar DS, Bashel B, Balasubramanya SAH, Creighton CJ, Ponce- Rodriguez I, Chakravarthi BVSK, et al. UALCAN: a portal for facilitating tumor subgroup gene expression and survival analyses. Neoplasia. 2017;19:649–58. 31. Wang L, Qin X, Liang J, Ge P. Induction of Pyroptosis: a promising strategy for Cancer treatment. Front Oncol. 2021;11:635774. 32. Zhang C-C, Li C-G, Wang Y-F, Xu L-H, He X-H, Zeng Q-Z, et al. Chemo‑ therapeutic paclitaxel and cisplatin differentially induce pyroptosis in A549 lung cancer cells via caspase-3/GSDME activation. Apoptosis. 2019;24:312–25. 33. Tsuchiya K. Switching from apoptosis to Pyroptosis: Gasdermin-elicited inflammation and antitumor immunity. Int J Mol Sci. 2021;22(1):426. 34. Jiang M, Qi L, Li L, Li Y. The caspase-3/GSDME signal pathway as a Ready to submit your research ? Choose BMC and benefit from: switch between apoptosis and pyroptosis in cancer. Cell Death Discov. 2020;6:112. • fast, convenient online submission 35. Céspedes MV, Unzueta U, Tatkiewicz W, Sánchez-Chardi A, Conchillo-Solé O, Álamo P, et al. In vivo architectonic stability of fully de novo designed • thorough peer review by experienced researchers in your field protein-only nanoparticles. ACS Nano. 2014;8:4166–76. • rapid publication on acceptance 36. Falgàs A, Pallarès V, Unzueta U, Céspedes MV, Arroyo-Solera I, Moreno • support for research data, including large and complex data types MJ, et al. A CXCR4-targeted nanocarrier achieves highly selective tumor uptake in diffuse large B-cell lymphoma mouse models. Haematologica. • gold Open Access which fosters wider collaboration and increased citations 2020;105:741–53. • maximum visibility for your research: over 100M website views per year 37. Yu P, Zhang X, Liu N, Tang L, Peng C, Chen X. Pyroptosis: mechanisms and diseases. Signal Transduct Target Ther. 2021;6:128. At BMC, research is always in progress. 38. Wang Y-Y, Liu X-L, Zhao R. Induction of Pyroptosis and its implications in Cancer management. Front Oncol. 2019;9:971. Learn more biomedcentral.com/submissions