Drug Delivery ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/idrd20 GSDMD-dependent pyroptotic induction by a multivalent CXCR4-targeted nanotoxin blocks colorectal cancer metastases Rita Sala, Elisa Rioja-Blanco, Naroa Serna, Laura Sánchez-García, Patricia Álamo, Lorena Alba-Castellón, Isolda Casanova, Antonio López-Pousa, Ugutz Unzueta, María Virtudes Céspedes, Esther Vázquez, Antonio Villaverde & Ramon Mangues To cite this article: Rita Sala, Elisa Rioja-Blanco, Naroa Serna, Laura Sánchez-García, Patricia Álamo, Lorena Alba-Castellón, Isolda Casanova, Antonio López-Pousa, Ugutz Unzueta, María Virtudes Céspedes, Esther Vázquez, Antonio Villaverde & Ramon Mangues (2022) GSDMD- dependent pyroptotic induction by a multivalent CXCR4-targeted nanotoxin blocks colorectal cancer metastases, Drug Delivery, 29:1, 1384-1397, DOI: 10.1080/10717544.2022.2069302 To link to this article: https://doi.org/10.1080/10717544.2022.2069302 © 2022 The Author(s). Published by Informa Published online: 09 May 2022. UK Limited, trading as Taylor & Francis Group. Submit your article to this journal Article views: 1771 View related articles View Crossmark data Citing articles: 5 View citing articles Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=idrd20 DRUG DELIVERY 2022, VOL. 29, NO. 1, 1384–1397 https://doi.org/10.1080/10717544.2022.2069302 RESEARCH ARTICLE GSDMD-dependent pyroptotic induction by a multivalent CXCR4-targeted nanotoxin blocks colorectal cancer metastases  Rita Salaa,b,c, Elisa Rioja-Blancoa,c, Naroa Sernab,d,e, Laura Sanchez-Garcıab,d,e, Patricia Alamo a,b,c , Lorena Alba-Castello n , Isolda Casanova , Antonio Lo a,c a,b,c pez-Pousa , Ugutz Unzueta , b,f a,b,c Marıa Virtudes C espedesa, Esther Vazquezb,d,e, Antonio Villaverdeb,d,e and Ramon Manguesa,b,c a Biomedical Research Institute Sant Pau (IIB-Sant Pau), Barcelona, Spain; bCIBER en Bioingenierıa, Biomateriales Y Nanomedicina (CIBER- BBN), Madrid, Spain; cJosep Carreras Research Institute, Barcelona, Spain; dInstitut de Biotecnologia I de Biomedicina, Universitat Autonoma de Barcelona, Bellaterra, Spain; eDepartament de Genetica I de Microbiologia, Universitat Autonoma de Barcelona, Bellaterra, Spain; f Department of Medical Oncology, Hospital de la Santa Creu I Sant Pau, Barcelon, Spain ABSTRACT ARTICLE HISTORY Colorectal cancer (CRC) remains the third cause of cancer-related mortality in Western countries, meta- Received 1 March 2022 stases are the main cause of death. CRC treatment remains limited by systemic toxicity and chemo- Revised 3 April 2022 therapy resistance. Therefore, nanoparticle-mediated delivery of cytotoxic agents selectively to cancer Accepted 12 April 2022 cells represents an efficient strategy to increase the therapeutic index and overcome drug resistance. KEYWORDS We have developed the T22-PE24-H6 therapeutic protein-only nanoparticle that incorporates the exo- Metastasis; CXCR4; toxin A from Pseudomonas aeruginosa to selectively target CRC cells because of its multivalent ligand colorectal cancer; targeted display that triggers a high selectivity interaction with the CXCR4 receptor overexpressed on the sur- nanoparticle; PE24 exotoxin; face of CRC stem cells. We here observed a CXCR4-dependent cytotoxic effect for T22-PE24-H6, which GSDMD; pyroptosis was not mediated by apoptosis, but instead capable of inducing a time-dependent and sequential activation of pyroptotic markers in CRC cells in vitro. Next, we demonstrated that repeated doses of T22-PE24-H6 inhibit tumor growth in a subcutaneous CXCR4þ CRC model, also through pyroptotic activation. Most importantly, this nanoparticle also blocked the development of lymphatic and hema- togenous metastases, in a highly aggressive CXCR4þ SW1417 orthotopic CRC model, in the absence of systemic toxicity. This targeted drug delivery approach supports for the first time the clinical relevance of inducing GSDMD-dependent pyroptosis, a cell death mechanism alternative to apoptosis, in CRC models, leading to the selective elimination of CXCR4þ cancer stem cells, which are associated with resistance, metastases and anti-apoptotic upregulation. 1. Introduction CRC, and also in other solid tumors, involve proteins of the Bcl-2 family, such as upregulation of Bcl-xL or Mcl-1 anti- Metastatic dissemination remains the main cause of death in apoptotic proteins (Schulze-Bergkamen et al., 2008; Song colorectal cancer (CRC) as in other tumor types. Despite et al., 2020). Therefore, new anticancer therapies that trigger advances in systemic therapy, the five-year survival rate of mechanisms of action alternative to apoptosis induction, metastatic CRC patients is only 14% (Crooke et al., 2018), being the acquisition of resistance to therapy the major rea- could overcome drug resistance mediated by apoptosis eva- son for treatment failure. Nearly half of metastatic CRC sion, expecting to have a significant impact on cure rates patients are resistant to chemotherapy regimens, which are and patient survival. mainly based on 5-fluorouracil (5-FU) and oxaliplatin (Virag Furthermore, targeted drug delivery aims at increasing et al., 2013; Francipane et al., 2019). Cancer cells can display the effectiveness of anticancer agents while reducing sys- intrinsic resistance or acquired resistance developed along temic toxicity (Singh & Lillard, 2009). Here, we developed with treatment, which involves agent-specific mechanisms self-assembling protein-based nanoparticles with intrinsic that widely vary among genotoxic agents and molecularly cytotoxic activity, to explore their targeting selectivity and targeted drugs (Pan et al., 2016). Thus, cancer cells activate the possible activation of cell death pathways alternative to different mechanisms to escape from therapy-induced death, apoptosis. For that purpose, we chose to study the anti- and apoptosis evasion, as a result of genetic and epigenetic cancer effect of a novel protein-only nanoparticle that alterations, one of the most prevalent (Hanahan & Weinberg, actively targets the CXCR4 receptor, whose overexpression 2011). Accordingly, the most frequent alterations found in has been observed in many tumor types and correlates CONTACT Ramon Mangues rmangues@santpau.cat Biomedical Research Institute Sant Pau (IIB-Sant Pau), Sant Antoni Ma Claret 167, Barcelona, 08025, Spain; Esther Vazquez esther.vazquez@uab.es CIBER en Bioingenierıa, Biomateriales y Nanomedicina (CIBER-BBN), Monforte de Lemos 3-5, Madrid, 28029, Spain ß 2022 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. DRUG DELIVERY 1385 withpoor survival, dissemination, and recurrence in CRC (Zeiss) operating at 1KV as described in (Sanchez-Garcıa patients (Kim et al., 2005) and accommodates the de-immu- et al., 2018). nized 24 kDa catalytic domain of the Pseudomonas aerugi- nosa exotoxin A (PE24). This exotoxin is in the same position as the GFP protein in the T22-GFP-H6 nanocarrier (Falgas 2.2. Cell culture and CXCR4 expression analysis et al., 2020), forming self-assembled nanoparticles that The SW1417 CRC cell line was purchased from the American incorporate several monomers per nanotoxin (estimated to Type Culture Collection (ATCC, Manassas, VA, USA) and was be 10–12 in a similar construct) (Rueda et al., 2015). cultured at 37  C and 5% CO2 in DMEM supplemented with Interestingly, we previously reported the capacity of a dif- 10% fetal bovine serum, 100 units/ml penicillin, and 100 mg/ ferent nanotoxin that incorporates the diphtheria exotoxin ml streptomycin (Thermo Fisher, Waltham, MA, USA). (DITOX), to activate pyroptosis, an inflammatory cell death Moreover, SW1417 cells were engineered to express the luci- mechanism alternative to apoptosis, in subcutaneous CRC ferase reporter gene (pPK-CMV- F3, Promokine, TE Huissen, tumors derived from organoids that had acquired resistance The Netherlands) by transfection with Lipofectamine 2000 to 5-Fluorouracil by selection along successive in vitro pas- (Thermo Fisher). sages (Serna et al., 2020). Here, we assessed whether the de- Before conducting all in vitro and in vivo experiments, the immunized T22-PE24-H6 nanotoxin could induce the activa- levels of CXCR4 expression in the cell membrane were tion of the pyroptotic cell death pathway and provide new assessed by flow cytometry. For that purpose, we labeled 1 possibilities in oncotherapy. We also tested if this nanotoxin million cells in 100 ml of PBS with 5 ml of the PE-Cy5 Mouse was able to selectively eliminate CXCR4þ target cancer cells Anti-Human CD184 monoclonal antibody (560937, BD and whether this could lead to effective antitumor and anti- Biosciences, Becton Dickinson, Franklin Lakes, NY, USA) and 1 metastatic activity. million cells in 100 ml of PBS with 5 ml of the PE-Cy5 Mouse IgG2a isotype (BD Biosciences) as a negative control. Cells 2. Materials and methods were incubated in the dark for 20 minutes at 4  C with gentle agitation, washed, and analyzed in the FACSCalibur cytome- 2.1. Protein production, purification and ter (BD Biosciences). We measured both the percentage of characterization the population expressing CXCR4 and the intensity of the expression expressed as mean fluorescence intensity (MFI). Synthetic codon-optimized genes encoding for T22-DITOX-H6 and T22-PE24-H6 modular proteins were designed in-house (Figure 1B) and provided by Geneart (ThermoFisher) subcl- 2.3. Cytotoxic effect assays oned into a pET22b plasmid. Recombinant plasmids were then transformed into Escherichia coli Origami B (BL21, The in vitro cytotoxic effect of different nanoparticles was OmpT, Lon, TrxB, Gor, Novagen) and proteins were evaluated by measuring cell viability using the XTT colori- produced overnight at 20  C upon addition of 0.1 mM of metric cell proliferation kit II (Roche Diagnostics, Basel, IPTG (isopropyl-b-D-thiogalactopyranoside). Bacterial cells Switzerland). SW1417 cells were seeded in 96 well plates and were then harvested by centrifugation (5000 g, 5 min) and incubated O/N at 37  C. After 24 hours, the T22-GFP-H6 nano- the obtained pellets were resuspended in wash buffer particle, or its therapeutic derivatives were added at different (20 mM Tris-HCl pH 8.0, 500 mM NaCl, 10 mM imidazole) in concentration ranges [0–2 mM] for 48 hours. The competitive presence of complete EDTA-Free (Roche) protease inhibitors. assays were performed by pre-treating the cells with Cell disruption was performed in an EmulsiFlex C5 cell dis- AMD3100 (molar ratio 1 T22-PE24-H6: 10 AMD3100). Then, ruptor (Avestin) at 8000–10,000 psi, soluble fraction separated cells were incubated with the XTT labeling mixture for by centrifugation (15,000 g, 45 min), and the soluble protein 4 hours and the formed formazan dye was quantified using a purified by Immobilized Metal Affinity Chromatography scanning multi-well spectrophotometer at 490 nm (FLUOstar (IMAC) with a HisTrap HP 1 ml column (Cytiva) in an AKTA € OPTIMA, BMG Labtech, Ortenberg, Germany). The measured Pure system (Cytiva). Protein elution was achieved by a linear absorbance directly correlates with the number of viable gradient of elution buffer (20 mM Tris, 500 mM NaCl, 500 mM cells. Data were reported as a percentage of cell viability of Imidazole, pH ¼ 8) and pure protein fractions were then dia- nanoparticle-treated cells in relation to buffer-treated lyzed against a sodium carbonate solution (166 mM NaCO3H, cell viability. pH ¼ 8). Protein purity was determined by SDS-PAGE and Western blot using an anti-His monoclonal antibody (Santa 2.4. Antitumor activity assessment in a subcutaneous Cruz Biotechnology) and protein integrity was verified by CRC model MALDI-TOF mass spectrometry. Finally, protein concentration was determined by the Bradford assay. To assess the antitumor activity of the T22-PE24-H6 nano- The volume size distribution of generated nanoparticles toxin we used the SW1417 subcutaneous model in Swiss was analyzed by Dynamic Light Scattering (DLS) in a Nude mice. All procedures were conducted in accordance Zetasizer Pro Blue (Malvern) at 633 nm with a backscattering with the guidelines approved by the institutional animal detector (173 ). Additionally, the size and shape of nanopar- Ethics Committee of Hospital Sant Pau. Swiss nude female ticles were also evaluated at a nearly native state in a Field mice (4 weeks old) were obtained from Charles River emission scanning electron microscope (FESEM) Zeiss Merlin Laboratories and ten million CXCR4þ SW1417 cells were 1386 R. SALA ET AL. Figure 1. Antitumor activity of toxin-based nanoparticles in SW1417 CRC cells in vitro. (A) CXCR4 membrane expression of the SW1417 cell line was measured by flow cytometry or by ICC in SW1417 cell blocks. Scale bar: 50 mm. (B) Schematic representation of the polypeptidic building blocks of the T22-PE24-H6 and T22- DITOX-H6 constructs (box sizes are only indicative). Linker indicates a short peptidic spacer to ensure flexibility of the protein domains. (C) Cell viability of the CXCR4þ SW1417 cells upon exposure to a 0–20 nM range of T22-PE24-H6 and T22-DITOX-H6 toxin-based nanoparticles for 48 h. The selectivity of the cytotoxic effect was also studied by performing a 30 min treatment with the antagonist AMD3100 previous to nanotoxin exposure in CXCR4þ SW1417 cells. Data presented as mean ± s.e.m. (N ¼ 2 in triplicate experiments). injected in two subcutaneous flanks. Once tumors reached et al., 2007) and three days after, primary tumor biolumines- approximately 120 mm3 sizes, mice were randomized into cence was measured using the IVIS Spectrum equipment. the control or treated group. Treated mice received intraven- Mice were then randomized into the control (N ¼ 9) and ous doses of 10 lg of T22-PE24-H6 (N ¼ 6 mice), at a treated (N ¼ 9) group and received intravenous doses of T22- repeated dose regime of three times a week, per eight doses. PE24-H6 at a repeated-dose regime of 5 lg, three times a The control group received a buffer using the same adminis- week, for a total of eighteen doses. The control group tration schedule. Overall the experimental period, mouse received a buffer using the same administration schedule. body weight and tumor volume were registered three times Overall the experimental period, mouse body weight and a week. Tumor bioluminescence was also measured once a health status were controlled three times per week. Primary week using the IVIS Spectrum equipment (Perkin Elmer, tumor bioluminescence was registered once per week using Waltham, MA, USA). When tumors reached a volume of the IVIS Spectrum equipment and the experiment finished 600 mm3, mice were euthanized and tumor and non-tumor when the first mouse needed to be euthanized. Necropsy organs were collected for further analyses. and ex vivo imaging of the organs of interest were per- formed 24 hours after the last dose. 2.5. Antimetastatic effect assessment in an orthotopic CRC model 2.6. In vivo and ex vivo bioluminescence monitoring To assess the antimetastatic activity of the T22-PE24-H6 In vivo mice, tumor progression was registered once per nanotoxin we used the metastatic SW1417 orthotopic model week by injecting intraperitoneally 2.25 mg of D-Luciferin in 4 weeks old NSG mice (Charles River Laboratories, (150 ml of the prepared 15 mg/mL stock, diluted in sterile L’Arbresle, France). CXCR4þ SW1417 cells were injected into physiological solution; PerkinElmer). After 5 minutes and the cecum of NSG mice following the orthotopic cell micro- under isoflurane anesthesia (XGI-8 system, Xenogen, injection (OCMI) procedure described elsewhere (Cespedes Alameda, CA, USA), bioluminescent signals emitted by tumor DRUG DELIVERY 1387 cells were detected and measured using the IVIS Spectrum CA, USA), cleaved caspase-1 (1:400, PA5-99390, Thermo equipment. Registration of bioluminescence emission was Fisher) and cleaved GSDMD (1:500, 36425, Cell Signaling, performed in the mouse ventral or dorsal position depend- Danvers, MA, USA), were assessed. The percentage of stained ing on the cell injection site. surface and its intensity were calculated using the ImageJ At the end of each experiment, a complete necropsy pro- software and the Color Deconvolution Plugin with the H cedure was performed, and all relevant organs (liver, kidneys, DAB vector to split the brown staining adjusting the thresh- lungs, diaphragm, spleen, pancreas, colon, intestinal tract, old to each marker. Then, the Analyze particles plugin was and mesenteric lymph nodes) were collected and covered in used to detect all stained areas and the mean gray value a 0.75 mg/kg D-Luciferin solution in 12 or 6 wells plates and was obtained by combining all selected black areas. The imaged as well. Exposure time was adjusted automatically intensity value was calculated by subtracting to 255 the according to the bioluminescence intensity of each mouse mean gray value obtained in the analysis. Up to 5 high- and cell line. Then, the obtained images were analyzed with power fields (400x) of each sample were analyzed and the the Living Image Software v4.5.5 (PerkinElmer) and the pho- results were expressed as mean ± s.e.m. tons/sec emitted by cells were quantified for each mouse We have also used IHC in the tissue samples obtained or organ. from in vivo experiments to assess the expression of CXCR4 (1:200, ab124824, Abcam, Cambridge, UK) in subcutaneous tumors, CRC primary tumors, and metastases located in dis- 2.7. Histopathological analysis and metastatic foci tal organs. quantitation All relevant organs described above were paraffin-embedded and stained with H&E to perform a complete histopatho- 2.9. Statistical analysis logical analysis and a clinical pathologist supervised all sam- In vitro cell, viability assays were performed in biological trip- ples for toxicity evaluation. Tumor sections stained with licates and the data were expressed as mean ± standard error either H&E or DAPI were also used to assess the cytotoxic of the mean (s.e.m.). IC50 was calculated with SigmaPlot effect of the administered T22-PE24-H6 nanotoxin in vivo, by (Systat Software, Inc) using a non-linear regression test with counting the number of cell death bodies in 10 high-power Hill-3 parameter adjustment. In the in vivo experiments, we fields (magnification 400x). DAPI staining was performed in analyzed differences in the number of metastatic foci and Triton X-100 (0.5%) permeabilized sections mounted with foci size between buffer-treated vs nanoparticle-treated mice, DAPI mounting media (ProLong Gold antifade reagent, using the Mann-Whitney U test, since data was obtained for Thermo Fisher). Samples were evaluated under a fluores- all measured variables did not follow a normal distribution. cence microscope at a wavelength of kex ¼ 334 nm/kem Results on the number of mitotic figures, number of cell ¼ 465 nm. death bodies, and active caspase 3, proteolyzed PARP, In the orthotopic in vivo experiments, the spreading of cleaved caspase-1, and cleaved GSDMD stained cells cells from the primary tumor (colon) to other distant organs between groups were analyzed using both the Student’s t- was analyzed in H&E-stained samples. We studied those test and the Mann-Whitney U test. Data were reported as organs in which metastatic dissemination is expected in CRC: mean ± s.e.m. and differences between groups were consid- liver, lung, mesenteric lymph nodes, and peritoneum. An ered significant at p < .05. The differences between relevant Olympus microscope with the Cell^D Olympus software was data were indicated as  for statistical significance among used to count the number and measure the size (expressed the designated groups. Statistical calculations were per- as mm2) of all observed metastatic foci in three different sli- formed using GraphPad Prism 8 software. ces of each organ. 3. Results 2.8. Immunocytochemistry and immunohistochemistry 3.1. Highly potent and CXCR4-dependent T22-PE24-H6 The study of cell death pathways induced by the T22-PE24- cytotoxic effect in CXCR41 CRC cells H6 was performed both in in vitro treated cells aggregated in cell blocks and in the subcutaneous CRC tumors of treated We first confirmed the high expression of the CXCR4 recep- mice. SW1417 cells were treated with 6 nM of T22-PE24-H6 tor in SW1417 CRC cells by flow cytometry and immunocyto- for 2, 5, 24, and 48 hours and then trypsinized and centri- chemistry (Figure 1A). Following, we tested the activity of fuged. The obtained pellet was mixed with plasma and two different previously generated toxin-based nanoparticles thrombin (Sigma Aldrich, San Luis, MO, USA) and quickly agi- (Sanchez-Garcıa et al., 2018), which structure is graphically tated to form a clot. The clot was placed in a tissue cassette, described in Figure 1(B), in CXCR4þ SW1417 cells in vitro. In fixed with formaldehyde, and then paraffin-embedded for short, T22-PE24-H6 is a modular protein that contains the further immunocytochemistry (ICC) analysis. Markers of apop- de-immunized catalytic domain of the Pseudomonas aerugi- tosis such as active caspase-3 protein (early apoptosis, 1:300, nosa exotoxin A fused to the CXCR4-specific ligand T22 at C92-605, BD Bioscience) or proteolyzed PARP (late apoptosis, the amino-terminal and a poly-histidine tag at the carboxy- 1:300, g7341, Promega, Madison, WI, USA) or pyroptosis such terminal. It also contains a KDEL sequence to induce its cyto- as NLRP3 (1:300, AG-20B-0014-C100, Adipogen, San Diego, solic release via retrograde transport through the secretory 1388 R. SALA ET AL. pathway. On the other hand, T22-DITOX-H6 is an equivalent 3.3. Lack of apoptosis activation after T22-PE24-H6 construction that contains the catalytic domain of the diph- nanotoxin treatment theria toxin from Corynebacterium diphtheriae together with We previously reported that the diverse therapeutic nanopar- its natural translocation domain for efficient cytosolic release ticles, either protein-only nanoparticles or nanoconjugates (Sanchez-Garcıa et al., 2018). Both modular proteins contain developed so far in our group, containing different cytotoxic a furin-cleavable site between T22 and the catalytic domain targets (DNA, microtubules, or mitochondrial proteins), kill to release the amino-terminal peptide once internalized. cancer cells mainly by apoptosis or mitotic catastrophe. Thus, Finally, as described before, both proteins self-assemble into we next determined whether the potent antitumor effect of pseudo-spherical nanoscale structures with average sizes the T22-PE24-H6 nanotoxin was mediated by activation of around 40 nm as determined by DLS and FESEM (Sanchez- the classical apoptotic cell death pathway. For that purpose, Garcıa et al., 2018). We surprisingly found that CXCR4þ CXCR4þ SW1417 cells were cultured and exposed to the IC50 SW1417 cells were not sensitive to T22-DITOX-H6 nanopar- of T22-PE24-H6 (6 nM) at different exposure times (2, 5, 24, ticles (Figure 1C). In contrast, CXCR4þ SW1417 cell exposure or 48 hours). Control CXCR4þ SW1417 cells were also cul- to a concentration range of 0–20 nM of T22-PE24-H6 nano- tured without being exposed to the nanotoxin. The immuno- toxin, showed a very potent and dose-dependent antineo- cytochemical assessment of the expression of active caspase- plastic effect, with an IC50 of 6.56 nM (Figure 1C). Moreover, 3 yielded a lack of differences in staining between control competition assays using CXCR4þ SW1417 cells pretreated and nanotoxin-treated cells (Figure 3A). with the CXCR4 inhibitor AMD3100, showed a dramatic We also studied apoptotic activation in the CXCR4þ reduction in T22-PE24-H6 anticancer activity, preventing the SW1417 subcutaneous tumors after treatment with eight nanotoxin-induced cell death and maintaining cell viability 10 mg T22-PE24-H6 doses, in which we had found a 3-fold close to 100% (Figure 1C). Thus, the T22-PE24-H6 nanotoxin increase in the number of cell death bodies and a decrease displays a potent and completely CXCR4-dependent cyto- in the expression of CXCR4. By immunohistochemistry toxic effect in CXCR4þ SW1417 cells in vitro. against apoptotic markers in tumor tissue sections, we found no detectable or very low signal of active caspase-3 or pro- teolyzed-PARP staining in nanotoxin-treated tumors (Figure 3B). These results suggested that the mechanism of cell 3.2. T22-PE24-H6 repeated dose administration death was not mediated by apoptosis induction. reduces tumor volume without toxicity in non- target organs 3.4. T22-PE24-H6 induction of the pyroptotic cell We next evaluated the antitumor effect of the T22-PE24-H6 death pathway and whether the induction of cell death observed in vitro, Since we could not detect activation of the apoptotic path- could be replicated in in vivo models. Thus, CXCR4þ SW1417 way after T22-PE24-H6 therapy, both in vitro and in vivo, we cells were injected into Swiss nude mice to generate a sub- explored possible alternative mechanisms of cell death. cutaneous tumor model. When tumor size reached an Pyroptosis has been recently described to be triggered by approximately 120 mm3 of size, mice were randomized into epithelial cells upon infection by intracellular bacterial patho- two groups (N ¼ 6) that were administered with a repeated gens (Jorgensen & Miao, 2015). Although this mechanism of dosage of 10 mg T22-PE24-H6 or buffer, three-time a week, action has been barely exploited in cancer therapy, we have per 8 doses (which we had previously observed not to pro- recently reported the T22-DITOX-H6 induction of pyroptosis duce any toxic effect in non-target organs in the same in tumors unable to induce apoptosis (Serna et al., 2020). mouse strain). Therefore, we analyzed whether the protein-only nanopar- After only four T22-PE24-H6 doses, significant differences ticle T22-PE24-H6, incorporating the PE24 exotoxin, was cap- (p ¼ 0.02) in tumor volume were already detected between able to induce pyroptosis in our CRC models, by determining control (buffer-treated) and T22-PE24-H6-treated mice (Figure the activation of different molecular components of this 2A). Moreover, at the end of the experiment, after the eighth pathway after exposure to this nanotoxin in vitro or after its dose, we observed a 1.6-fold reduction in tumor volume, as intravenous injection in mice in vivo (Figure 4). compared to buffer-treated mice (p ¼ 0.005), which was asso- After exposure of CXCR4þ SW1417 cells to the T22-PE24- ciated with a significant reduction in final tumor weight H6 nanotoxin for a short time, at 2 or 5 hours, we found in (p ¼ 0.04) (Figure 2B). This antitumor effect was a likely con- almost all cancer cells a high overexpression of the NLRP3 sequence of the 3.0-fold increase in the number of cell death protein, an inflammatory marker that initiates the signaling bodies observed in tumor tissue (p < 0.05) (Figure 2C). cascade that leads to pyroptosis. This expression was main- Moreover, at the end of the experiment, nanotoxin-treated tained in a reduced percent of cancer cells during the and buffer-treated mice did not show significant differences 24–48 hours period, which however continued being in mouse body weight (Figure 2D). The histology of non-tar- expressed at a high intensity (Figure 4A and B). We next get organs, such as the liver and kidneys, was evaluated by observed that the activation of the NLRP3 inflammasome H&E-stained sections, observing a complete lack of histo- was subsequently followed by the generation of active cas- logical alterations in these tissues. pase-1 at longer exposure times (5, 24, and 48 hours), which DRUG DELIVERY 1389 Figure 2. T22-PE24-H6 antitumor effect in a cell-derived subcutaneous SW1417 CRC model. (A) Antitumor effect of T22-PE24-H6 measured by the reduction of tumor volume (mm3). (B) Antitumor effect of T22-PE24-H6 measured by tumor weight (g) at the end of the experiment. (C) Increase in the number of cell death bodies in the SC SW1417 tumors counted in 10 high-power fields of H&E-stained tissue samples collected at the end of the experiment. (D) Follow-up of mouse body weight (g) during the repeated dose administration of 10 mg T22-PE24-H6 (three times a week, 8 total doses). All data are presented as mean ± s.e.m., N ¼ 6. p < 0.05; p < 0.01. Figure 3. In vitro and in vivo assessment of apoptosis activation after T22-PE24-H6 treatment. (A) Representative images of ICC staining of active caspase-3 in cell blocks of CXCR4þ SW1417 cells exposed to 6 nM T22-PE24-H6 for 5, 24 or 48 h. Stained sections were compared to untreated control CXCR4þ SW1417 cells. (B) Representative images of H&E and IHC staining of CXCR4, active caspase-3 and proteolyzed PARP in SW1417 subcutaneous tumors from Swiss nude mice treated with buffer or 10 mg of T22-PE24-H6, three times a week and 8 total doses. Scale bars: 50 mm. 1390 R. SALA ET AL. Figure 4. Activation of the pyroptotic pathway in CXCR4þ SW1417 cultured cells and in subcutaneous tumors after T22-PE24-H6 treatment. (A) Detection of pyrop- totic markers, including NLRP3, cleaved caspase-1 and cleaved GSDMD, by ICC in cell blocks of CXCR4þ SW1417 cells exposed for 2, 5, 24 or 48 h to 6 nM of T22- PE24-H6. Stained sections were compared to untreated control CXCR4þ SW1417 cells. Scale bars: 50 mm. (B) Quantitation values of each marker staining in SW1417 control cells or in cells treated with T22-PE24-H6. In ICC of NLRP3 and caspase-1, up to 5 high-power fields (400x) were imaged and the percentage of stained sur- face was measured with ImageJ. In GSDMD staining, up to 5 high-power fields (400x) were analyzed by counting the number of GSDMD positive cells. (C) Representative images of IHC staining of NLRP3, cleaved caspase-1 and cleaved GSDMD in SW1417 SC tumors from Swiss nude mice treated with 10 mg of T22- PE24-H6, three times a week and 8 total doses. Scale bars: 50 mm. (D) Quantitation of pyroptotic markers in tumors from buffer-treated and T22-PE24-H6-treated mice. NLRP3 staining (5 high-power fields (400x) analyzed with ImageJ) was expressed as a percentage of the stained surface. To assess caspase-1 and GSDMD stained sections, we instead counted the number of positive cells in 5 high-power fields (400x). Measurements in tissue sections were performed 24 h after the last administered dose, being compared to buffer-treated control mice. All the data are expressed as mean ± s.e.m.  p < 0.05;  p < 0.001. DRUG DELIVERY 1391 in turn, cleaved the pro-Gasdermin D (pro-GSDMD) protein. nanotechnological anticancer approach as a powerful alter- It is known that the active GSDMD N-terminal fragment native for the treatment of CRC tumors once they develop released after cleavage translocates to the plasma mem- relapse and metastasize, especially, when recurrence is due brane, where it oligomerizes to form pores in the cell surface to the upregulation of anti-apoptotic proteins (Hanahan & (Weldon & Pastan, 2011; Mulvihill et al., 2018). Consistently, Weinberg, 2011) (Figure 5). we observed a highly evident GSDMD translocation to the membrane of CXCR4þ SW1417 cells with very high intensity in 10–20% of exposed cancer cells, mainly after 48 hours of 3.6. T22-PE24-H6 prevents the development of exposure to T22-PE24-H6 (Figure 4A). lymphatic and hematogenous metastases in the The sequential activation of highly expressed pyroptotic CXCR41 SW1417 cell-derived CRC model markers was also observed in T22-PE24-H6-treated CXCR4þ Once we demonstrated that the T22-PE24-H6 nanotoxin has SW1417 tumors, as compared to their negligible expression a potent antitumor effect, mediated by pyroptotic cell death in buffer-treated tumors. Thus, NLRP3, cleaved caspase-1, induction, in CXCR4þ SW147 subcutaneous tumors, we per- and cleaved Gasdermin D (GSDMD) showed a higher expres- formed an experiment to determine its capacity to prevent sion 24 hours after the last dose (Figure 4C), in the repeated cancer cell dissemination and metastatic foci growth at the T22-PE24-H6 nanotoxin dosage administration described clinically relevant organs affected by metastases in an ortho- above, which previously showed significant antitumor activity topic CXCR4þ SW1417 CRC model. This model metastasizes (Figure 2). Again, the first marker to be activated in tumor to lymph nodes, liver, lung, and peritoneum. We expected tissues was NLRP3 with a cytoplasmic expression pattern in that treatment with a T22-PE24-H6 repeated dosage could most of the tumor cells. After that, caspase-1 is activated reduce the number and size of metastatic foci at the end of showing very specific staining in the subset of cancer cells the experiment. For that purpose, NSG mice were implanted undergoing cell death, which increased 6-fold the number of in the cecum with two million CXCR4þ SW1417 CRC cells. stained cells as compared to buffer-treated tumors. Finally, After three days, we started administering the nanotoxin to GSDMD was also activated in the pyroptotic bodies causing the mice in the experimental group, following the same pore formation and cell death, according to its role as a repeated-dose schedule used for antitumor treatment, or pyroptotic cell effector (Figure 4D). with buffer in the control group. Treatment continued until the first mouse, belonging to either the control or treated group, achieved the criteria for euthanasia. 3.5. Selective pyroptotic effect on CXCR41 cancer cells After 18 doses, control mice started to lose weight due to based on T22-PE24-H6 multivalency metastatic dissemination, a time point at which all mice The integration of the previously described structural infor- were euthanized (Figure 6A). At this point, we could find dif- mation on the T22-PE24-H6 self-assembling nanoparticle and ferences close to significance in total body bioluminescence, the antitumor effect observed in this work, which is medi- emitted by luciferase-expressing cancer cells, between mice ated by the activation of the pyroptotic cell death pathway, treated with T22-PE24-H6 and buffer-treated mice (Figure leads us to propose a mechanistic view of this novel thera- 6B). The ex vivo bioluminescence analysis of each cancer- peutic approach (Figure 5) that could overcome, at least in affected organ showed differences in signal between both part, the relapse and metastatic dissemination that occurs groups (Figure 6C and E). The bioluminescence emitted by when resistance to classical chemotherapy develops, in CRC the tested organs of the nanotoxin-treated mice, which and in other tumor types. Thus, the fact that the homo-oligo- included metastatic foci disseminated to the peritoneum, meric T22-PE24-H6 nanoparticle displays on its surface sev- liver, and lung, which maintained the expression of the eral copies of the CXCR4 ligand T22 (estimated to be 10–12 CXCR4 receptor (Figure 6D), was in all cases lower than that in a very similar construct) (Rueda et al., 2015), and that emitted by the same organs of buffer-treated mice, which CXCR4þ tumors in CRC patients show an overexpression of correlates with a reduced metastatic load. the CXCR4 receptor (as compared to normal tissues, includ- Following, a histological evaluation of the metastatic load ing the CXCR4þ bone marrow cells) (Serna et al., 2020), as it after T22-PE24-H6-treatment was performed by quantifying happens with the CXCR4þ SW1417 model that we here used the developed metastatic foci and size for each affected (Figure 1A), makes the delivery of exotoxin cytotoxic domain organ in H&E-stained sections. Mice treated with T22-PE24- in the target cell cytosol of high selectivity compared with H6 presented a reduction in metastatic load (total metastatic the negligible delivery of this domain in normal cells (Falgas area) in all affected organs, especially evident in lymph et al., 2020). This would lead to a high antitumor effect in nodes (p < 0.05) (data not shown). Compared to the findings tumor tissue and a lack of toxicity in non-tumor organs in the buffer-treated group, T22-PE24-H6 treatment potently (Figure 5). prevented lymphatic (Lymph node Mets) and hematogenous On the other hand, the lack of induction of apoptosis by (Liver Mets and Lung Mets) metastasis development, whereas the T22-PE24-H6 in CXCR4þ CRC cells, both in vitro and its capacity to prevent transcelomic metastases (Peritoneal in vivo, and the capacity of this nanotoxin to sequentially Mets) were lower. Consistent with the ex vivo biolumines- activate the pyroptotic markers NLRP3, caspase-1, and cence results, mice treated with T22-PE24-H6 showed a 2- GSDMD, in CXCR4þ CRC cells, leading to its selective and fold reduction in mean metastatic foci number in lymph CXCR4-dependent cell death, would establish this novel node (p ¼ 0.008), a 2.1-fold reduction in the liver (p ¼ 0.007) 1392 R. SALA ET AL. Figure 5. Graphical summary showing the T22-PE24-H6 highly selective antitumor effect in CXCR4 overexpressing CRC cells, through the activation of pyroptosis. The image describes the advantage of this novel approach in front of other anticancer therapies, due to the super selectivity of the polytpeptidic nano toxins that allows its specific internalization in CXCR4þ target cancer cells and the induction of pyroptosis, a non-apoptotic cell death mechanism. T22-PE24-H6 interacts with the CXCR4 receptor in the cell surface of cancer cells, being subsequently internalized by endocytosis and trafficked through the Golgi apparatus and the endoplas- mic reticulum (ER). Once in the cytoplasm, the Pseudomonas aeruginosa exotoxin is able to induce target cancer cell death by activation of the pyroptotic effectors NLRP3 (part of the inflammasome complex), caspase-1 and Gasdermin-D (GSDMD). The activation of this novel mechanism will overcome resistance to classical anticancer drugs mediated by cancer cell upregulation of antiapoptotic pathways (e.g. overexpression of Bcl-2 protein family). and a 2.5-fold reduction in the lung (p ¼ 0.01), as compared examined for possible histological alterations in the kidneys to metastatic foci number of the corresponding sites in buf- and liver, the two organs with higher nanoparticle biodistri- fer-treated mice (Table 1). There was also a trend of a reduc- bution, but we also analyzed all described relevant normal tion of the mean number of peritoneal metastatic foci in organs. No differences in histology were observed between T22-PE24-H6-treated mice (similar to the reduction in bio- control buffer-treated mice and T22-PE24-H6-treated mice in luminescence emission by peritoneal metastases described these organs. This lack of toxicity was previously found in above); however, it did not reach statistical significance. the same organs, including normal colon, kidney, liver, and Thus, the nanotoxin reduces mainly the number of meta- spleen or bone marrow, in a lymphoma model (Falgas et al., static foci rather than their size, since in T22-PE24-H6 treated 2020) and in a head and neck cancer model (Rioja-Blanco mice, primary tumor, lymph node, and lung metastases des- et al., 2022) treated with the same T22-PE26-H6 dosage. pite of showing a trend toward a smaller mean size, they did Consequently, the glomeruli and surrounding renal tubules not achieve statistically significant differences. were clearly visible and presented no cytoplasmic vacuol- ation or eosinophilic protein accumulation. In liver tissue, the hepatocytes did not lose their architecture and did not pre- 3.7. Lack of toxicity of T22-PE24-H6 repeated sent steatosis or any other histological alteration (Figure 7). administration in normal and metastasis- Therefore, consistently with the negligible nanoparticle bio- affected organs distribution to normal tissues, the lack of histological altera- We also evaluated the effect of the T22-PE24-H6 nanotoxin tions in all analyzed tissues, including also the lung, heart, in normal and metastasis affected organs in which we had pancreas and spleen, and the lack of mouse body weight previously described a transient accumulation in biodistribu- loss during the repeated administration of T22-PE24-H6 tion assays with the T22-GFP-H6 nanoparticle by measuring (Figure 2D and 6A), indicate a wide therapeutic index for the fluorescence emission, because of the presence of fenes- T22-PE24-H6 nanotoxin at a dosage that achieves a potent trated vessels (Cespedes et al., 2016). We principally antimetastatic effect. DRUG DELIVERY 1393 Figure 6. T22-PE24-H6 antimetastatic effect in the cell-derived orthotopic CXCR4þ SW1417 CRC model. (A) Evolution of mouse body weight (g) during the repeated dose administration of 5 mg of T22-PE24-H6 (three times a week, 18 total doses). (B) Total body bioluminescence (BLI; Total Flux [p/s]) of buffer and T22- PE24-H6 treated mice, was measured once per week throughout the experiment. (C) Comparison of bioluminescence emission between the buffer-treated and the 5 mg T22-PE24-H6 repeated-dose schedule groups in the primary tumor, mesenteric lymph nodes, liver, lung, and peritoneum. Results are presented as mean- ± s.e.m. bioluminescence values in photons per second (total flux [p/s]). (D) CXCR4 IHQ images show that the receptor is highly expressed in cancer cells from the primary tumor, liver, and lung metastasis. Scale bar: 50 mm. (E) Representative bioluminescence images comparing primary tumor and metastatic dissemination in liver and lungs of mice treated with the nanotoxin (5 mg T22-PE24-H6) or buffer (control). 4. Discussion & Pastan, 2011). On this basis, we previously developed pro- Among all cytotoxic domains found in organisms, bacterial tein-only nanoparticles, incorporating the de-immunized ver- toxins are already being used in new strategies for cancer sion of the catalytic domain of the exotoxin A from therapy, exhibiting a highly potent anticancer effect (Weldon Pseudomonas aeruginosa together with a polypeptide 1394 R. SALA ET AL. Table 1. T22-PE24-H6 antimetastatic effect measured by prevention of metastases development in the CXCR4þ SW1417 cell-derived CRC metastatic model. T22-PE24-H6 prevention of metastasis SW1417 cell-derived orthotopic model Primary tumor Lymph node Mets Liver Mets Lung Mets Peritoneal Mets Group Mice % Mice % # foci Mice % # foci Mice % # foci Mice % # foci Buffer 9/9 100% 100% 100% 100% 89% 7.1 ± 1.1a 15 ± 1.9b 49.2 ± 9.6c 4.4 ± 0.7 T22-PE24-H6 9/9 100% 100% 100% 89% 89% 3.4 ± 0.5a 7.1 ± 1.6b 19.7 ± 4.1c 3.2 ± 0.9 Metastatic foci size (mm2  103) Group Primary tumor Lymph nods Mets Liver Mets Lung Mets Peritoneal Mets Buffer 9428.8 ± 1081.1 739.8 ± 94.6 15.7 ± 3.5 44.1 ± 9.1 2423.4 ± 1141.2 T22-PE24-H6 7331.5 ± 978.8 513.8 ± 99.3 16.5 ± 4.2 36.1 ± 8.9 2579.2 ± 1345.9 Mean þ s.e.m. metastatic foci number or area (mm2) per mouse, counted in three entire histology sections. p ¼ 0.008; bp ¼ 0.007; cp ¼ .001. a Figure 7. Lack of histological alterations in normal and colonized organs of mice treated with T22-PE24-H6. Representative images of H&E staining in normal organs (kidney, heart, pancreas, and spleen) and in metastasis-affected organs (liver and lung). These organs were collected 24 hours after the last buffer or T22- PE24-H6 administered dose of the repeated administration schedule (5 mg of T22-PE24-H6, 18 total doses, three times per week) in NSG mice orthotopically injected with the CXCR4þ SW1417 CRC cell line. Scale bars: 100 mm. containing the T22 peptide and a His-tag; thus, generating bearing subcutaneous tumors underwent a 1.6-fold reduction the self-assembling multivalent T22-PE24-H6 nanotoxin in tumor volume and a 3-fold increase in tumor cell death (Sanchez-Garcıa et al., 2018). This novel protein-only nano- body induction at the end of the experiment, compared to particle approach allowed us to achieve a faster production buffer-treated mice. Moreover, when repeatedly administer- process by skipping the conjugation step needed for the ing low doses of T22-PE24-H6 in the highly metastatic cell- therapeutic nanoconjugates synthesis (Sharma et al., 2012) derived orthotopic model, we observed a significant 2-fold while reducing its immunogenicity due to the replacement reduction in the number of lymphatic and hematogenous of the exogenous GFP protein, found in our previously devel- metastatic foci and a decreasing trend in the primary tumor oped nanoconjugates (Cespedes et al., 2018; Pallares and metastatic foci size. Consistently with these results, our et al., 2020). group also proved the T22-PE24-H6 capacity to block dissem- Our results demonstrate that the T22-PE24-H6 nanotoxin ination in a diffuse large B-cell lymphoma (DLBCL) model displays a potent CXCR4-dependent cytotoxic effect in CRC without associated toxicity, by the elimination of CXCR4þ cells and also that its use at low doses in a repeated treat- cancer cells through apoptosis induction (Falgas et al., 2020), ment regime is capable of inhibiting tumor growth and a result consistent with the pharmacokinetics of T22-P26-H6 metastasis development without associated systemic toxicity. that show a fast biodistribution in the bloodstream followed Thus, exposure of the CXCR4þ SW1417 cell line to the toxin, by a slow elimination phase having a half-life of 30 hours showed a reduction in cell viability and a lack of antitumor (Cespedes et al., 2007). activity after competition with the AMD3100 antagonist, The success of this new approach relies again on achiev- showing a highly selective CXCR4 receptor-dependent cell ing highly selective targeted drug delivery by exploiting the death induction. Furthermore, T22-PE24-H6-treated mice CXCR4 membrane overexpression of CRC cells as compared DRUG DELIVERY 1395 to normal tissues but also due to T22-PE24-H6 multivalency been demonstrated that the downregulation of Bcl-2 and since it leads to super selectivity in target cell internalization Bcl-XL using antisense techniques was able to sensitize cells (Liu et al., 2020). Also importantly, because of the capacity of to chemotherapy (Hayward et al., 2004), whereas a loss of the PE24 toxin to trigger the mechanism of pyroptosis, a cell Bax (pro-apoptotic protein) expression resulted in increased death alternative to apoptosis. Thus, we found that target resistance (Zhang et al., 2000). Therefore, treatment with the cell death induction by the T22-PE24-H6 nanotoxin in in vitro T22-PE24-H6 nanotoxin could represent a promising tool for and in vivo CRC models is activated through the following targeting CXCR4þ cancer stem cells and rendering sensitive ordered steps (Figure 5): 1) the nanotoxin, through its mul- the CRC tumors with acquired resistance to chemotherapy- tiple T22 ligands, binds to CXCR4 receptors exposed in the induced apoptosis after their treatment, by inducing pyrop- membrane of target cells and internalizes by endocytosis, 2) tosis as an alternative cell death mechanism (Figure 5). furin-cleavage sites inserted between the T22 peptide and The exotoxin of Pseudomonas aeruginosa is being used in the cytotoxic PE24 domain allow its cleavage and the intra- the clinical setting in the form of immunotoxins for cancer cellular release of the toxin active fragment, 3) the carboxyl- therapy. Many toxin-based therapies are under clinical trials, terminal KDEL sequence in PE24 domain increases the effi- but Moxetumomab pasudotox, composed of an anti-CD22 ciency of retrograde transport through the endoplasmic antibody fused to a 38 kDa portion of the Pseudomonas exo- reticulum and its subsequent release into the cytosol and 4) toxin A, is one of the few approved immunotoxins by the the released catalytic domain will block protein synthesis by FDA, in this case for the treatment of hairy-cell leukemia inhibition of the elongation factor 2 (EF2) to induce the acti- (Kreitman et al., 2012). However, immunotoxin therapy has vation of pyroptosis, a scarcely explored non-apoptotic cell been successfully applied against hematological malignan- death mechanism in cancer therapy (Tait et al., 2014; cies, despite of several issues that still represent significant Nagarajan et al., 2019). barriers to their effective use for solid cancer treatment Importantly, the finding of T22-PE24-H6 pyroptotic induc- (Shan et al., 2013). These hurdles include dose-limiting toxic- tion, mediated by activation of the inflammosome marker ities, immunogenicity, and reduced efficacy in cytosolic deliv- NLRP3, followed by caspase-1 cleavage and GSDMD cleavage ery (Kim et al., 2020). We believe that the T22-PE24-H6 that translocate to the membrane in CRC models, in the nanotoxin, displays a wider therapeutic window than immu- absence of apoptosis activation (in sharp contrast with the notoxins, based on the fact that this nanotoxin is a self- observation of potent apoptosis induced by the same nano- assembled oligomeric construction expectedly formed by toxin in a DLBCL lymphoma model (Falgas et al., 2020) rein- around 10–12 protein monomers, that therefore displays forces its use to induce anticancer activity in tumors with 10–12 T22 ligands and that consequently shows super select- inactivated apoptotic pathways, at least in CRC and other ivity (Martinez-Veracoechea & Frenkel, 2011; Sharma et al., solid tumors, especially after recurrence or active metastatic 2012). Thus, its multivalency makes this construct internalize dissemination. It is also highly relevant to point out that T22- faster and more intensively in target cancer cells that overex- PE24-H6 induces a mechanism of pyroptosis not previously press the CXCR4 receptor than immunotoxins, which only described in CRC and different from that induced by cur- display two Fab fragments per one exotoxin domain in each rently used chemotherapeutic anticancer drugs, which is immunotoxin molecule, to interact and internalize in target mediated by activation of caspase-3 and GSDME (Wang cancer cells. On the other hand, most of the developed fused et al., 2017; Aizawa et al., 2020). Instead, the nanotoxin acti- immunotoxins used the PE38 fragment of the exotoxin. vates the inflammasome to activate caspase-1 followed by Recent studies demonstrated that de-immunization of GSDMD. In this regard, resistance to the classical chemother- recombinant toxins by identifying and removing B-cell epito- apy regimens is an important problem in different types of pes, decreased the immune response when administered in cancer treatment, including CRC. Exposure to 5-fluorouracil mice. Investigations have focused, for instance, on the elim- can cause tumor relapse by the emergence of a population ination of the B-cell epitopes of the 25 kDa portion (HA22- of cancer cells with stem-like properties, resistant to this LR-8M) of the PE toxin to produce a fully cytotoxic protein drug (Xu et al., 2015; Francipane et al., 2019). Cancer cells effective against leukemia cell lines (Onda et al., 2011). acquire resistance to chemotherapy through the emergence Therefore, we incorporated the PE24 toxin into our T22- of genetic mutations or epigenetic changes, favoring the based nanoparticles, which has been demonstrated to be activation of signaling pathways related to chemotaxis, cell less immunogenic and better, tolerated than previous ver- survival or proliferation, including the CXCL12/CXCR4 axis sions of this toxin, as we observed in this study. (Touil et al., 2014). In human cancers, the anti-apoptotic pro- teins (e.g. Bcl-2, Bcl-XL, and Mcl1) are often upregulated in 5. Conclusion cancer cells, enabling them to evade apoptotic cell death and losing the capacity to undergo apoptosis in response to We have developed an intrinsically cytotoxic and de-immu- chemotherapeutic drugs (Wilson et al., 2006; Pan et al., nized nanotoxin that offers a novel therapeutic approach, 2016). The Bcl-2 gene has been shown to be overexpressed which reaches a wider therapeutic index, because of its in many solid tumor cell lines (Yang et al., 2004; Yip & Reed, multivalency, than the previously developed nanoconjugates 2008; Goldsmith et al., 2012) and clinically, a high Bcl-2 used to treat metastasis in CRC models. Most importantly, expression in patient samples correlates with a poor we here demonstrate for the first time that the T22-PE24-H6 response to therapy (Geng et al., 2013). Furthermore, it has nanotoxin induces GSDMD-mediated pyroptotic cell death in 1396 R. SALA ET AL. CRC cells but is unable to trigger apoptosis. Thus, this nano- Cespedes MV, Unzueta U, Avin ~o  A, et al. (2018). Selective depletion of toxin could be a promising tool for effectively overcoming metastatic stem cells as therapy for human colorectal cancer. EMBO Mol Med 10:e8772 (1-22). the apoptosis blockade associated with relapsed chemother- Crooke H, Kobayashi M, Mitchell B, et al. (2018). Estimating 1- and 5-year apeutic-resistant and metastatic CRC tumors, by triggering relative survival trends in colorectal cancer (CRC) in the United States: pyroptosis, an alternative cell death pathway, mechanistically 2004 to 2014. JCO 36:587. different from apoptosis, and poorly studied in oncotherapy. Falgas A, Pallares V, Serna N, et al. (2020). Selective delivery of T22-PE24- H6 to CXCR4þ diffuse large B-cell lymphoma cells leads to wide therapeutic index in a disseminated mouse model. Theranostics 10: Acknowledgements 5169–80. Falgas A, Pallares V, Unzueta U, et al. (2020). A CXCR4-targeted nanocar- The authors would like to thank the free access to Servier Medical Art rier achieves highly selective tumor uptake in diffuse large B-cell images (http://smart.servier.com/) that allowed us to create a new com- lymphoma mouse models. Haematologica 105:741–53. posed image. Francipane MG, Bulanin D, Lagasse E. (2019). Establishment and charac- terization of 5-fluorouracil-resistant human colorectal cancer stem-like cells: tumor dynamics under selection pressure. Int J Mol Sci 20:1817 Ethical approval (1-23). Geng M, Wang L, Li P. (2013 Oct 15). Correlation between chemosensi- All the animal experiments were approved by the institutional animal tivity to anticancer drugs and Bcl-2 expression in gastric cancer. Int J Ethics Committee of Hospital de Sant Pau. Clin Exp Pathol 6:2554–9. Goldsmith KC, Gross M, Peirce S, et al. (2012 May 15). Mitochondrial Bcl- 2 family dynamics define therapy response and resistance in neuro- Disclosure statement blastoma. Cancer Res 72:2565–77. AV, EV, UU, RM, IC are cited as inventors in three patents: PCT/EP2018/ Hanahan D, Weinberg RA. (2011). Hallmarks of cancer: the next gener- 061732, covering Nanostructured Proteins and uses thereof; PCT/EP2018/ ation. Cell 144:646–74. 069303, covering Therapeutic Nanoconjugates and uses thereof; and Hayward RL, Macpherson JS, Cummings J, et al. (2004) Enhanced oxali- PCT/EP2012/050513 (WO2012/095527), covering Methods and reagents platin-induced apoptosis following antisense Bcl-xl down-regulation is for efficient and targeted delivery of therapeutic molecules to CXCR4 p53 and Bax dependent: Genetic evidence for specificity of the anti- cells, all licensed to Nanoligent. AV, EV, RM are scientific advisors and sense effect. Mol Cancer Ther 3(2):169–78. cofounders of Nanoligent. The authors report no other conflicts of inter- Jorgensen I, Miao EA. (2015 May). Pyroptotic cell death defends against est in this work. intracellular pathogens. Immunol Rev 265:130–42. Kim J, Takeuchi H, Lam ST, et al. (2005). Chemokine receptor CXCR4 expression in colorectal cancer patients increases the risk for recur- Funding rence and for poor survival. JCO 23:2744–53. Kim J-S, Jun S-Y, Kim Y-S. (2020). Critical issues in the development of This work was supported by Instituto de Salud Carlos III (ISCIII, Co-fund- immunotoxins for anticancer therapy. J Pharm Sci 109:104–15. ing from FEDER) [FI16/00017 to R.S., PI18/00650, PIE15/00028, PI21/ Kreitman RJ, Tallman MS, Robak T, et al. (2012 May 20). Phase I trial of 00150 and EU COST Action CA 17140 to R.M.]; Agencia Estatal de anti-CD22 recombinant immunotoxin moxetumomab pasudotox n (AEI) and Fondo Europeo de Desarrollo Regional (FEDER) Investigacio (CAT-8015 or HA22) in patients with hairy cell leukemia. J Clin Oncol [grant BIO2016-76063-R, AEI/FEDER, UE to A.V. and PID2019-105416RB- 30:1822–8. I00/AEI/10.13039/501100011033 to E.V]; CIBER-BBN [CB06/01/1031 and Liu M, Apriceno A, Sipin M, et al. (2020). Combinatorial entropy behav- 4NanoMets to R.M., VENOM4CANCER to A.V. NANOSCAPE to U.U. and iour leads to range selective binding in ligand-receptor interactions. NANOREMOTE to E.V.]; AGAUR [2017- SGR-865 to R.M., 2017SGR-229 to Nat Commun 11:4836. A.V., and 2018FI_B2_00051 to L.S.G.] and a predoctoral fellowship from Martinez-Veracoechea FJ, Frenkel D. (2011). Designing super selectivity the Government of Navarra to N.S.; Josep Carreras Leukemia Research in multivalent nano-particle binding. Proc Natl Acad Sci USA 108: Institute [P/AG to R.M.] and the Generalitat de Catalunya CERCA 10963–8. Programme. A.V. received an Icrea Academia Award. UU was supported Mulvihill E, Sborgi L, Mari SA, et al. (2018). Mechanism of membrane by Miguel Servet contract (CP19/00028) from ISCIII co-funded by pore formation by human gasdermin-D. Embo J 37(14):e98321 (1-11). European Social Fund (ESF investing in your future). The authors would Nagarajan K, Soundarapandian K, Thorne RF, et al. (2019). Activation of like to thank the ICTS-141007 Nanbiosis Platform, especifically the pyroptotic cell death pathways in cancer: an alternative therapeutic Platform of Protein Production and the ICTS Nanbiosis Nantoxicology approach. Transl Oncol 12:925–31. Unit of CIBER-BBN where the bioluminescent follow-up of cancer cells Onda M, Beers R, Xiang L, et al. (2011). Recombinant immunotoxin and toxicity studies were performed. against B-cell malignancies with no immunogenicity in mice by removal of B-cell epitopes. Proc Natl Acad Sci USA 108:5742–7. Pallar es V, Unzueta U, Falgas A, et al. (2020). An Auristatin nanoconju- gate targeting CXCR4þ leukemic cells blocks acute myeloid leukemia References dissemination. J Hematol Oncol 13:36 (1-19). Pan S-T, Li Z-L, He Z-X, et al. (2016). Molecular mechanisms for tumour Aizawa E, Karasawa T, Watanabe S, et al. (2020 May 22). GSDME-depend- resistance to chemotherapy. Clin Exp Pharmacol Physiol 43:723–37. ent incomplete pyroptosis permits selective IL-1a release under cas-  Rioja-Blanco E, Arroyo-Solera I, Alamo P, et al. (2022). CXCR4-targeted pase-1 inhibition. iScience 23:101070. nanotoxins induce GSDME-dependent pyroptosis in head and neck Cespedes MV, Espina C, Garcıa-Cabezas MA, et al. (2007). Orthotopic squamous cell carcinoma. J Exp Clin Cancer Res 41:49. microinjection of human colon cancer cells in nude mice induces Rueda F, Cespedes MV, Conchillo-Sol e O, et al. (2015). Bottom-up tumor foci in all clinically relevant metastatic sites. Am J Pathol 170: instructive quality control in the biofabrication of smart protein mate- 1077–85. rials. Adv Mater 27:7816–22.  Cespedes MV, Unzueta U, Alamo P, et al. (2016). Cancer-specific uptake  Sanchez-Garcıa L, Serna N, Alamo P, et al. (2018). Self-assembling toxin- of a liganded protein nanocarrier targeting aggressive CXCR4þ colo- based nanoparticles as self-delivered antitumoral drugs. J Control rectal cancer models. Nanomedicine 12:1987–96. Release 274:81–92. DRUG DELIVERY 1397 Schulze-Bergkamen H, Ehrenberg R, Hickmann L, et al. (2008). Bcl-x(L) Virag P, Fischer-Fodor E, Perde-Schrepler M, et al. (2013). Oxaliplatin and myeloid cell leukaemia-1 contribute to apoptosis resistance of induces different cellular and molecular chemoresistance patterns colorectal cancer cells. World J Gastroenterol 14:3829–40. in colorectal cancer cell lines of identical origins. BMC Genomics  Serna N, Alamo P, Ramesh P, et al. (2020). Nanostructured toxins for the 14:480. selective destruction of drug-resistant human CXCR4þ colorectal can- Wang Y, Gao W, Shi X, et al. (2017). Chemotherapy drugs induce pyrop- cer stem cells. J Control Release 320:96–104. tosis through caspase-3 cleavage of a gasdermin. Nature 547:99–103. Shan L, Liu Y, Wang P. (2013). Recombinant immunotoxin therapy of Weldon JE, Pastan I. (2011). A guide to taming a toxin-recombinant solid tumors: challenges and strategies. J Basic Clin Med 2:1–6. immunotoxins constructed from Pseudomonas exotoxin A for the Sharma A, Madhunapantula SV, Robertson GP. (2012). Toxicological con- treatment of cancer. Febs J 278:4683–700. Wilson TR, Longley DB, Johnston PG. (2006). Chemoresistance in solid siderations when creating nanoparticle-based drugs and drug delivery tumours. Ann Oncol 17(Suppl 10):x315–324. systems Expert Opin Drug Metab Toxicol 8:47–69. Xu Z-Y, Tang J-N, Xie H-X, et al. (2015). 5-Fluorouracil chemotherapy of Singh R, Lillard JW. (2009). Nanoparticle-based targeted drug delivery. gastric cancer generates residual cells with properties of cancer stem Exp Mol Pathol 86:215–23. cells. Int J Biol Sci 11:284–94. Song X, Shen L, Tong J, et al. (2020). Mcl-1 inhibition overcomes intrinsic Yang X, Zheng F, Xing H, et al. (2004). Resistance to chemotherapy- and acquired regorafenib resistance in colorectal cancer. Theranostics induced apoptosis via decreased caspase-3 activity and overexpres- 10:8098–110. sion of antiapoptotic proteins in ovarian cancer. J Cancer Res Clin Tait SWG, Ichim G, Green DR. (2014). Die another way-non-apoptotic Oncol 130:423–8. mechanisms of cell death . J Cell Sci 127:2135–44. Yip KW, Reed JC. (2008). Bcl-2 family proteins and cancer. Oncogene 27: Touil Y, Igoudjil W, Corvaisier M, et al. (2014). Colon cancer cells escape 6398–406. 5FU chemotherapy-induced cell death by entering stemness and quies- Zhang L, Yu J, Park BH, et al. (2000). Role of BAX in the apoptotic cence associated with the c-Yes/YAP axis. Clin Cancer Res 20:837–46. response to anticancer agents. Science 290:989–92.