ORIGINAL RESEARCH published: 23 February 2021 doi: 10.3389/fimmu.2021.622995 Natural Killer Anti-Tumor Activity Can Be Achieved by In Vitro Incubation With Heat-Killed BCG Gloria Esteso 1, Nacho Aguiló 2, Esther Julián 3, Omodele Ashiru 4†, Mei. M. Ho 4, Carlos Martı́n 2 and Mar Valés‐Gómez 1* 1 Department of Immunology and Oncology, Spanish National Centre for Biotechnology (CNB), Consejo Superior de Edited by: Investigaciones Científicas (CSIC), Madrid, Spain, 2 Mycobacteria Genetics Group, University of Zaragoza Medical School, IIS Carsten Watzl, Aragón, CIBERES, Zaragoza, Spain, 3 Departament de Genètica i Microbiologia, Facultat de Biociències, Universitat Autònoma Leibniz Research Centre for Working de Barcelona, Barcelona, Spain, 4 Custom Solutions Department, Bacteriology Division, Medicines and Healthcare Products Environment and Human Factors Regulatory Agency – National Institute for Biological Standards and Control (MHRA-NIBSC), Potters Bar, United Kingdom (IfADo), Germany Reviewed by: Francisco Borrego, Natural Killer cell receptors allow this heterogeneous immune population to efficiently fight Biocruces Bizkaia Health Research both tumors and infection, so their use as immunotherapy agents is an active field of Institute, Spain Martin R Goodier, research. Cytokine activation, particularly by myeloid cell-derived IL15, can induce potent University of London, United Kingdom NK anti-tumor responses. While studying the mechanism of action of intravesical *Correspondence: instillations of Bacille Calmette-Gueŕ in (BCG) as therapy for patients with high risk non- Mar Valés-Gómez mvales@cnb.csic.es muscle invasive bladder cancer, we showed that BCG can activate a cytotoxic CD56bright † Present address: NK cell population which efficiently recognized bladder cancer cells. This pioneer Omodele Ashiru, immunotherapy provides an invaluable model to understand the role of different Abcam plc, Cambridge, immune populations in tumor elimination. However, during the propagation of BCG United Kingdom worldwide a large number of genetically diverse BCG substrains developed. Here, we Specialty section: investigated the capacity of different BCG substrains to promote NK cell activation and This article was submitted to confirmed that they were able to activate lymphocytes. Tice, Connaught and Moreau were NK and Innate Lymphoid Cell Biology, a section of the journal the substrains with a stronger NK activation effect as measured by CD56 upregulation. Frontiers in Immunology Surprisingly, dead mycobacteria also stimulated PBMC cultures and we further Received: 29 October 2020 demonstrate here that subcellular fractions of BCG-Tice, in the absence of live Accepted: 04 January 2021 Published: 23 February 2021 mycobacteria, could also induce an NK cell response. Lipids from BCG-Tice, but not Citation: from Mycobacterium bovis, stimulated NK cell activation and degranulation, whereas the Esteso G, Aguiló N, Julián E, Ashiru O, aqueous fraction of either bacteria did not activate lymphocytes. However, delipidated Ho MM, Martı´n C and Valés‐Gómez M BCG-Tice bacteria were able to activate effector cells (CD3+CD56+ and NK, CD3-CD56+). (2021) Natural Killer Anti-Tumor Activity Can Be Achieved by In Vitro These data demonstrate that different components of mycobacteria can stimulate different Incubation With Heat-Killed BCG. immune subpopulations resulting in phenotypes suitable for cancer elimination. Front. Immunol. 12:622995. doi: 10.3389/fimmu.2021.622995 Keywords: cancer immunology, BCG strains, NK activation, CD56bright, bladder cancer, mycobacterial fractions Frontiers in Immunology | www.frontiersin.org 1 February 2021 | Volume 12 | Article 622995 Esteso et al. Non-Viable BGC and NK Cell Activation INTRODUCTION BCG has been used for treatment of bladder cancer as intravesical instillations for decades, after the first successful Natural Killer cells express a large number of receptors trials in the 1970s (14–16) and FDA approval in 1989 for CIS, recognizing both pathogen-infected and transformed cells followed by an extension in 1998 to treat Ta/T1 papillary tumors. which allow them to contribute to many arms of the immune However, because of the way in which BCG was developed, many response (1). The capacity of NK cells to recognize tumors has different regimes are used in the treatment of bladder cancer been long known, however, their role for cancer immunotherapy patients with BCG: different hospitals use different substrains, is currently the focus of active research (2). The contribution of mycobacterial concentrations and schedules (13, 17, 18). The NK cells in the elimination of human bladder cancer has been general regime of BCG treatment consists of an initial cycle of 6- extensively reported in the context of local instillations with the weekly intravesical instillations, called the induction phase. The tuberculosis vaccine Bacille Calmette-Guérin (BCG) (3, 4), which addition of a maintenance phase lasting one to three years, that is the treatment of choice for T1G3 non-muscle invasive bladder consists in successive cycles of 3 weekly instillations separated by cancer (NMIBC) appearing in the form of either papillary 3-month rest periods after the induction phase, has represented tumors or carcinoma in situ (CIS). Since the beginning of the the main improvement of this therapy in the multiple decades in use of this therapy several decades ago, the survival time of use. Importantly, clinical studies have directly addressed the bladder cancer patients increased notably. However, survival effect of the different protocols and there are contradicting data rates have not changed in the last 30 years and many questions on the influence of different substrains of BCG in patient about the mechanism of action of the BCG against bladder outcome. Clinical trials to study recurrence free survival (RFS) cancer and about the optimal dose and recall instillations to be using different substrains are difficult to compare since they also used in patients remain open. While studying the phenotypical differ in the schedules of induction vs induction + maintenance changes of NK cells mediating tumor elimination in the context (with different maintenance regimes), one-arm vs two to three-arms, of BCG, we have recently reported that, after exposure of 1–5 years follow-up, tens to thousands of patients (18). In this Peripheral Blood Mononuclear Cells (PBMCs) to BCG, NK context, a trial including a large number of patients (2,451 patients cells proliferate leading to a CD56bright phenotype while keeping with primary T1G3 tumors from 23 clinics) in a non-randomized functional characteristics of mature NK cells including cytotoxic retrospective comparison, found that BCG Connaught reduced the activity and a high capacity to mediate Antibody Dependent Cellular recurrence rate, overall progression and death compared to BCG- Cytotoxicity (ADCC) (5). This unconventional cytotoxic Tice when no maintenance was used, but the opposite was true when subpopulation of CD56bright NK cells is reminiscent of the potent maintenance was given. However, for time to progression and overall anti-tumor NK cells described after blood cell IL15 priming that survival, Connaught and Tice had a similar efficacy (19, 20). These result in enhanced elimination of multiple myeloma (6). The anti- data agree with a prospective randomized single-institution trial with tumor BCG-stimulated CD56bright NK cell population that we treatment of 142 patients with BCG Connaught or Tice study in previously described (5) can be distinguished from “classical” which Connaught had better RFS rate. However, the low number of CD56bright NK cells normally found in a small percentage in cases did not allow statistical conclusions on progression-free peripheral blood, because they have markers of mature NK cells. survival, disease-specific survival or overall-survival. In the same Most express high levels of CD94 and are CD16+, and a subset is report, Rentsch et al. (20) treated mice with the same BCG KIR2D+. Further, this population is able to mediate both substrains and showed, in the animal model, that BCG Connaught degranulation and ADCC. The role of BCG in CD56 upregulation induced a stronger Th1 response, greater priming of BCG-specific was consistent when using large numbers of different donors, CD8+ T cells, and more robust T-cell recruitment to the bladder than however, the bacterial components involved were not studied. BCG Tice. Other retrospective studies in BCG-treated bladder cancer BCG was generated in 1921, after 13 years of passage of patients conclude, however, that there is not much difference between Mycobacterium bovis (M. bovis) in vitro, during which this BCG substrains in the treatment of bladder cancer on progression- tuberculosis-causing agent acquired a number of mutations and free survival or disease-specific survival (21). gene modifications, with the loss of the RD1 region being the main Since data in the literature point towards a high degree of cause for attenuation (7). Through this process, a new strain named heterogeneity between the different BCG preparations available BCG was generated. The new cultures of attenuated mycobacteria, for bladder cancer patients, we decided to compare the effect of perfectly able to reproduce, were distributed worldwide (8, 9). Since different BCG substrains on the activation of PBMCs, with the no freeze-drying methods were available until the 1960s, the different hypothesis that both the intensity and the quality of the immune host laboratories kept different subcultures of the new bacilli growing activation might vary between different BCG preparations. There and BCG acquired new genetic modifications thereby generating is a strong Natural Killer (NK) cell component in the mechanism independent substrains which received different names, e.g. BCG- of action of BCG in bladder cancer (3, 22) evidenced by the Pasteur, -Connaught, -Tokyo, -Danish, -Tice (10–12). Thus, a whole proliferation and activation of BCG-stimulated CD56bright NK family of BCG substrains are included under the BCG denomination. cells, hereafter, referred to as CD56bright NK cells (5). Here, we Commercial preparations of lyophilized BCG of different substrains specifically characterized the generation of the anti-tumoral are used for intravesical instillations of bladder cancer patients and CD56bright NK cell subpopulations in vitro in response to the viable bacteria content is highly variable ranging from 106–109 different substrains of BCG. During these studies we discovered (10, 13). that, in addition to different numbers of viable bacilli, the different Frontiers in Immunology | www.frontiersin.org 2 February 2021 | Volume 12 | Article 622995 Esteso et al. Non-Viable BGC and NK Cell Activation commercially available presentations of BCG can contain high ratios formaldehyde for 10 min at RT, permeabilized with 0.2% saponin of dead mycobacteria accompanying the colony-forming units for 10 min at RT and stained with Perforin-PE, Granzyme B-PE (CFU), information that cannot be inferred from the supplier (Biolegend) for 30 min. Cells were washed in PBA and analyzed leaflet. This finding led us to demonstrate that dead BCG also using a Gallios Flow Cytometer (Beckman Coulter). Analysis of the contribute to the activation of certain pathways of the immune experiments was performed using Kaluza software. response, in particular, NK cell anti-tumor activity. These data are consistent with, and may explain, previous in vivo findings in which PBMC Stimulation With BCG autoclaved BCG inhibited tumor growth in mice with transplanted Aliquots of reconstituted BCG were prepared in RPMI with 10% bladder cancer cells (23). Interestingly, in other models, NK and T DMSO and stored at −20°C. For BCG stimulation, experiments cell recruitment to tumors was dependent on BCG viability (24, 25), were performed as previously described (5). Briefly, 0.5 x 106 suggesting that other immune cell types need to be activated for a PBMCs/ml were incubated in 48-well plates with or without complete response. Evaluating subcellular mycobacterial components BCG at a 1:10 ratio (viable bacteria:PBMC). One week later, cells from M. bovis and BCG-Tice, we determined that fragments from M. in suspension were recovered from the co-culture, centrifuged, bovis could strongly provoke lymphocyte proliferation, but less analyzed by flow cytometry. skewed towards an NK cells response when compared with BCG- Tice fragments. Delipidated BCG-Tice was very efficient in Live Versus Dead Mycobacteria Counting stimulating CD56 upregulation, suggesting that non-covalently Viable bacteria were determined by serial dilutions in bound mycobacterial lipids and glycolipids are not strongly Middlebrook 7H10 medium supplemented with Albumin involved in NK activation. Dextrose Catalase (ADC). To obtain heat-killed BCG, bacteria were washed in PBS and treated at 100°C for 20–30 min. The total number of BCG, including live and dead bacilli, was MATERIALS AND METHODS calculated by flow cytometry using a protocol optimized for this study as follows. A calibration curve was prepared, using a Cells, BCG Substrains, and Reagents log-phase growth culture of BCG, in Middlebrook 7H9 medium Bladder cancer cell lines, T24 and RT112, and the erythroleukemia supplemented with ADC, assuming that the whole culture was K562 cell line were previously described (5). viable. Then, serial dilutions (1:2) were prepared and analyzed by PBMCs from buffy coats of healthy donors were obtained from flow cytometry. Bacterial viable units in the culture were the Regional Transfusion Centre (Madrid) with ethical permission calculated by plating in solid medium and counting colonies and experimental protocols approved by the institutional for each standard curve point. For these analyses, FSC and SSC committees: Regional Transfusion Centre (PO-DIS-09) and parameters were set in the flow cytometer to visualize bacteria assessed by the bioethics committee of CSIC. Informed consent (Supplementary Figure 1). Samples were acquired, using a was obtained at the Transfusion Centre from all participants. All Beckman Coulter Gallios flow cytometer, at constant speed methods were carried out in accordance with biosafety guidelines during 30 s and the number of events was defined as the total and regulations authorized by CNB-CSIC. amount of bacteria. For analysis, weasel software was used. Data PBMCs were isolated by centrifugation on Ficoll-HyPaque and of the different dilutions were used to build the calibration curve cultured for one week in RPMI-1640 (Biowest) supplemented with and, the latter to interpolate and calculate the total amount of 5% FBS, 5% human male AB serum, 2 mM glutamine, 1 mM BCG present in different formulations. sodium pyruvate, 0.1 mM non-essential amino acids, 10 mM Hepes, 100 U/ml penicillin and 100 U/ml streptomycin (Biowest) Degranulation Experiments at a concentration of 1 x 106 cells/ml in 96-well plates. Untreated or BCG-treated PBMCs were used as effector cells; World Health Organization Reference Reagents of BCG- bladder cancer cell lines or the erythroleukemia K562 (as positive Russian (code 07/274) and BCG-Tokyo 172 (code 07/272) control for NK degranulation) were used as target cells, always substrains were obtained from the National Institute for pretreated with HP1F7 antibody to block MHC-I. The Biological Standards and Control (NIBSC, UK); BCG-Pasteur percentage of NK cells was determined for each donor, so that strain 1173P2 was originally obtained from Institut Pasteur Paris, equivalent effector NK cell ratios were set up for the different BCG-Moreau was a gift from Dr Sylvia Leao, Federal University donors and conditions. NK cells were allowed to degranulate in of Sao Paulo, Brazil); BCG-Tice (OncoTICE) and BCG- co-cultures of PBMCs and targets for 2 h at an E:T ratio of 5:1 Connaught (ImmuCyst) were both commercially available. (which corresponds usually to 1:2 NK:target ratio), following the method published by Bryceson et al. (26). Surface expression of Flow Cytometry LAMP1 (CD107a) was analyzed by flow cytometry as shown in For flow cytometry, cells were washed with PBA [PBS Supplementary Figure 3. Statistical analyses were performed supplemented with 0.5% bovine serum albumin (BSA), 1% FBS using the Prism 6 software. and 0.1% sodium azide] and incubated with antibodies against surface markers [CD3-CF (Immunostep); CD3-FITC, CD56-PC5, Mycobacterial Subcellular Fractions CD16-PE-Cy7, CD4-PC5.5, CD8-PC7 and LAMP1-APC The following reagents were obtained through the NIH (Biolegend)] at 4°C for 30 min in the dark. For intracellular Biodefense and Emerging Infections Research Resources staining, after surface labelling, cells were fixed with 1% p- Repository, NIAID, NIH: g-irradiated whole cells M. bovis, Frontiers in Immunology | www.frontiersin.org 3 February 2021 | Volume 12 | Article 622995 Esteso et al. Non-Viable BGC and NK Cell Activation Strain AF 2122/97 (ATCC® BAA-935™) Catalog No. NR-31210; described previously that this population of “BCG-stimulated Total Lipids, Catalog No. NR-44100, Cell Membrane Fraction, CD56bright NK cells” differ both in phenotype and function Catalog No. NR-31214 and Cytosol Fraction, Catalog No. NR- capacities from “peripheral blood CD56bright NK cells”, for 31215 from M. bovis; Trehalose-6,6-dimycolate (TDM), Catalog simplicity, NK cells expressing high vs low amounts of CD56 No. NR-14844 and Mycolic Acid, Catalog No. NR-14854 from are called in this paper CD56bright and dim. However, this only M. tuberculosis, strain H37Rv. refers to the level of expression of the receptor. Thus, to confirm To obtain BCG-Tice fragments, BCG-Tice was grown at 37°C on the activation of an anti-tumor NK population, in this paper, we Middlebrook 7H10 medium supplemented with Oleic-Albumin- check the expression of high levels of CD56 together with CD16 Dextrose-Catalase (OADC). After 4 weeks, cells were scraped from and the degranulation capacity of these cells after incubation the surface and non-covalently attached cell wall lipids were with bladder cancer cells. extracted, first with chloroform/methanol (2:1, v/v), and then Firstly, we compared the effect of two reference BCG vaccines with chloroform/methanol (1:2, v/v) mixtures. Pooled organic with OncoTICE® on the upregulation of the surface marker extracts, containing the lipids, were combined, dried and CD56 on NK cells using the same ratio as previously. Incubation partitioned with chloroform/methanol/water (8:4:2 v/v). Aqueous of PBMCs with both Tokyo and Russian substrains (lyophilized phase (“aqueous” antigen) and chloroform phase (“lipid” antigen) preparations obtained from NIBSC) also resulted in an increase were then separated and evaporated to dryness. Delipidated bacteria in the percentage of CD56bright NK cells (Figure 1A). Since different were collected and stored in aliquots in sterile glass tubes commercial formulations contain different concentrations of BCG, (“delipidated” antigen). laboratory grown cultures of the mentioned different substrains were subsequently used to plan new experiments using equal numbers of mycobacteria. When using laboratory grown substrains, upregulation RESULTS of CD56 also could be observed, although surprisingly, the percentages of CD56bright cells obtained were much lower with the Differential Activation of CD56bright NK laboratory preparations than with the commercial BCG (Figure 1B). Cells by BCG Substrains Depends on the Since the pellet obtained after centrifugation of the suspensions from Number of Total Mycobacteria commercially available preparations was substantial, we hypothesized To test whether the different BCG substrains used in the clinics that a larger number of bacteria, presumably dead, were present in could differently activate the immune response, experiments to addition to the live BCG indicated in the leaflet. Thus, the numbers of study the variability of mycobacteria in the context of NK live vs dead BCG were analyzed in each lyophilized sample. Live BCG activation were initiated. We recently reported that stimulation were measured by counting CFU in Middlebrook 7H10 agar plates with BCG, i.e., culture with the clinical grade preparation and the total number of BCG was estimated by flow cytometry, using OncoTICE® at a ratio of 1:10 (105 live bacteria to 106 PBMC), a protocol optimized for the present study (see Materials and led to the generation of an anti-tumoral NK cell population Methods and Supplementary Figure 1). These numbers were characterized by upregulation of the surface marker CD56 on compared with the concentration range provided by the NK cells together with expression of CD16 (5). Although we have manufacturer. While the CFU range provided in the product insert A B FIGURE 1 | NK cell activation, based on the increase of CD56bright cells, occurs after incubation with several substrains of BCG. PBMCs from healthy donors were incubated with or without the indicated substrains of BCG at a 1:10 ratio (viable bacteria to PBMC). Commercial OncoTice was used for comparison. At day 7, cells in suspension were recovered from the co-culture, centrifuged and analyzed by flow cytometry. The percentage of the CD56bright and CD56dim populations within NK cells was obtained for each donor and plotted in a bar graph. Statistical analysis was performed using T-student test to compare each substrain with OncoTice (*p < 0.05; **p < 0.01). (A) Comparison of reference vaccines with Oncotice. (B) Comparison of laboratory-grown substrains with Oncotice. Frontiers in Immunology | www.frontiersin.org 4 February 2021 | Volume 12 | Article 622995 Esteso et al. Non-Viable BGC and NK Cell Activation agreed with the numbers that were obtained by the culture method, Although the final amount of different subtypes of lymphocytes the number of total BCG obtained by flow cytometry was, in general, varied slightly, all the BCG substrains tested gave rise to the much higher. Table 1 summarizes the numbers obtained for the upregulation of CD56 on NK cells (Figure 3A, right panel) and different BCG preparations. For example, OncoTICE® contains increased degranulation against bladder cancer cells (Figure 3B). 2–8 x 108 viable BCG per vial, according to the manufacturer’s Indeed, upregulation of CD56 was significant for all seven donors information but the total number of mycobacteria measured by flow tested, compared to untreated PBMCs, and the percentage of BCG- cytometry was 5.95x 109, thus, only around 10% of mycobacteria stimulated CD56bright cells was very consistent when comparing present in the vial were alive after reconstitution in buffer. Similarly, between the different substrains. In parallel, the capacity of these only 38% of mycobacteria were alive in the BCG Tokyo lyophilized activated NK cells to degranulate after recognition of two bladder ampoule (NIBSC, WHO reference reagent, Japan BCG laboratory) cancer cell lines [T24, a grade 3 cell line recognized by activated NK (Table 1). cells from most donors, and RT112, a low grade cell line usually Thus, the differences observed in BCG-NK activation between poorly recognized by activated NK cells (22)], was tested. the commercial lyophilized product (OncoTice) and the Supplementary Figure 3 shows a gating strategy for this type of laboratory cultures might be due to the presence of large experiments. These experiments demonstrate an increase in numbers of dead bacteria in the former. degranulation of BCG-stimulated CD56bright NK cells when using different BCG substrains containing approximately 10% of viable Dead Mycobacteria Can Activate Anti- bacteria (Figure 3B). In these experiments, the regions Tumor CD56bright NK Cells to Degranulate corresponding to T cells could also be analyzed, confirming that Against Bladder Cancer Cell Lines after BCG exposure the main granule mobilization activity was To test whether dead BCG could contribute to the activation of observed for CD56bright cells. CD3 degranulation against bladder CD56bright NK cells, the capacities of clinical grade Tice and cancer cells did not occur and it was not different to that seen for several laboratory-cultured BCG preparations to stimulate CD3 cells in the absence of target. To sum up, these results suggest PBMCs were compared using either viable or heat-inactivated that dead mycobacteria in co-culture with PBMCs contribute to the BCG. In order to avoid heterogeneity due to culture/production activation of NK cells for an enhanced anti-tumoral activity. conditions, the different substrains were grown in laboratory using the same culture medium and method. BCG added to PBMCs Different Mycobacterial Fragments were either mostly viable (at log phase), completely inactivated or Lead to the Activation of Different mixed 50% dead/50% alive. In all the cases, PBMC activation was Immune Populations observed as shown by the emergence of lymphoblast cells in the Since dead BCG contributed to the activation of BCG-stimulated FCShigh/SSChigh region (Figure 2A, Supplementary Figure 2). CD56bright NK cells, experiments were designed to study their effect The percentage of activated lymphocytes was plotted for the on different PBMC subpopulations. Different subcellular fractions different donors showing very similar percentages with viable of mycobacterial extracts from M. bovis and BCG-Tice were also and heat-killed bacteria. The commercial product, OncoTICE, used in co-culture experiments with PBMCs to explore the idea that was used as a positive control for comparison, since it was the one the use of fragments of bacteria or different strains (including non- used for our initial description of NK cell activation. CD56bright pathogenic bacteria) might avoid toxicity in bladder cancer patients, NK cells were present in all the cases (Figure 2B). The percentage as has been suggested previously (28–30). Whole irradiated M. of CD56bright NK cells increased significantly with respect to bovis, and different fractions, membrane, cytosolic, total lipids and untreated cells even when PBMCs were incubated with 100% individual lipids [Trehalose-6,6-dimycolate (TDM) and mycolic heat-inactivated BCG, suggesting that the activation of NK cells in acid] were incubated at different concentrations in co-culture with these cultures does not require interaction with live mycobacteria. PBMCs (7 days) and the percentage of activated lymphocytes and These experiments show that NK cells can be activated when NK subpopulations was analyzed by flow cytometry non-viable BCG is co-cultured with PBMCs. To further confirm (Supplementary Figure 4A). Since most of the compounds tested this hypothesis, the capacity of different laboratory-grown BCG could activate lymphocytes, experiments to define the relative substrains using approximately 10% of live bacteria (proportion activation of each lymphocyte subpopulation were performed found in OncoTICE) to activate PBMCs was tested showing (Figure 4A). Whole irradiated M. bovis and membrane fractions lymphocyte activation once again (Figure 3A, left panel). led to an increase in lymphocyte activation and, although an increase in CD56bright cells could be observed, the population of NK cells did not proliferate as much as CD3 lymphocytes. In TABLE 1 | Live vs dead BCG in commercial lyophilized preparations. particular, the percentage of CD4 cells was higher in cultures SUBSTRAIN BCG CFU* TOTAL BCG# where the membrane fraction was added (Figure 4B). Data in this figure show the relative percentage of CD4 compared to CD8 TICE 2-8 x 108 (3mg) 6.0 x 109 within the CD3 lymphocyte region, indicating that in general the CONNAUGHT 1.8-15.9 x 108 (81mg) 5.2 x 109 RUSSIAN 3.4 x 106 (0.5mg) 1.5 x 107 two subpopulations of T lymphocytes maintain the same ratio in the TOKYO 5 x 107 (1mg) 1.3 x 108 co-cultures, except when the membrane fractions of M. bovis were used. In this case, the proportion of CD4 cells increased *CFU provided by the manufacturer or NIBSC (27) (expressed as dry weight) and confirmed in the laboratory. significantly. In agreement with a moderate activation of NK # Number of BCG cells counted using flow cytometry analysis (see Materials and Methods). cells, low degranulation capacity of NK cells against the two Frontiers in Immunology | www.frontiersin.org 5 February 2021 | Volume 12 | Article 622995 Esteso et al. Non-Viable BGC and NK Cell Activation A B FIGURE 2 | Lymphocyte activation occurs in co-cultures using either live or dead BCG. PBMCs from healthy donors were incubated with the indicated BCG substrains. The ratio was calculated to mimic the amount of total bacteria in the commercial vaccine, 6:1 ratio (total bacteria to PBMC) from cultures either 100% viable, 50% viable or completely inactivated by 20 min at 100°C. At day 7, cells in suspension were recovered from the co-culture, centrifuged and analyzed by flow cytometry. (A) The percentage of activation was determined by identifying the resting and activating lymphocyte regions by FSC vs SSC. The percentage of activated lymphocytes was plotted for each donor (different colors) and viability condition. Statistical analysis was performed using one-way ANOVA comparing each condition with the untreated culture (*p < 0.05; **p < 0.01; ns, non-significant). (B) The percentage of the CD56bright population was obtained within the NK cell region for each donor and analyzed as in (A). bladder cancer cell lines, T24 and RT112, was found in these The same type of experiments were also performed to evaluate cultures (Figure 4C). These experiments demonstrated that, NK activation using BCG-Tice fragments (Supplementary although OncoTICE led to a lower percentage of activated Figure 4B; Figure 5). In this case, delipidated-BCG, as well as lymphocytes than M. bovis, the composition of activated the lipidic and the aqueous fractions were used. In contrast to M. subpopulations was more skewed towards NK cells in the former bovis fragments, Tice lipids, mildly activated PBMC cultures while leading to higher percentage of CD56bright upregulation and NK the aqueous phase did not activate lymphocytes at all (Figure 5A). degranulation capacity. On the other hand, CD4 cells were activated The more activated cultures were obtained upon incubation with more efficiently by M. bovis fragments or irradiated cells. The delipidated BCG. No fragment was as efficient as intact OncoTICE cytosolic fraction of M. bovis only stimulated significant NK cell at stimulating NK CD56 upregulation, however the percentage of degranulation against the T24 cell line although an upward trend, activated lymphocytes was higher in cultures with both that did not become significant, could be observed for RT112. The delipidated-BCG and lipid fragments than in untreated PBMCs lipidic fractions of M. bovis, in the concentrations used for these (Figure 5B). Both the lipid and the delipidated, but not the experiments, did not stimulate the PBMC culture. aqueous fraction, led to only low level of CD56bright activation, Frontiers in Immunology | www.frontiersin.org 6 February 2021 | Volume 12 | Article 622995 Esteso et al. Non-Viable BGC and NK Cell Activation A B FIGURE 3 | Lymphocyte populations and NK degranulation after co-cultures with 10% viable BCG. PBMCs from healthy donors were incubated with the indicated substrains of BCG at a 1:10 ratio (live bacteria to PBMC), using cultures containing approximately 10% viable bacteria plus 90% inactivated bacteria by 20-30 min at 100°C. OncoTice resuspended from the commercial vial was used as a control. At day 7, cells in suspension were recovered from the co-culture, centrifuged and analyzed by flow cytometry. (A) The percentage of activated lymphocytes was determined by identifying the resting and activating lymphocyte regions by FSC vs SSC and further analyzed to distinguish T cells from NK cells by staining with antibodies against population markers (CD3, CD56) (left). The percentage of activated lymphocytes was plotted for each donor (different populations are displayed in different colors). The percentage of the CD56bright population within the NK cell region was obtained for each donor (different colors) and analyzed by one-way ANOVA (**p < 0.01; ***p < 0.001; ****< 0.0001) (right). (B) Degranulation of CD56bright NK cells against bladder cancer cells (T24, RT- 112) was measured by analyzing surface LAMP-1 (CD107a) within the CD3-CD56+ region. Different donors are represented by symbols and compared with the untreated culture. Statistical analysis was performed using one-way ANOVA (**p < 0.01; ***p < 0.001; ****p < 0.0001; ns, non-significant). but when present, these cells were able to recognize bladder cancer presented here confirm that the different mycobacterial strains cells in degranulation assays (Figure 5C). In fact, the amounts of have different capacities to stimulate leukocytes and that the intracellular granzyme B and perforin in BCG-stimulated subpopulations of lymphocytes stimulated by different bacterial CD56bright NK cells were clearly increased when PBMCs were compounds diverge. treated with delipidated-BCG in a comparable manner to cultures A key discovery reported here is that, although all BCG stimulated with BCG-Tice (Figure 6). Although NK cells became substrains can stimulate NK cell proliferation and response activated by BCG lipid, CD3 lymphocytes did not show against high grade bladder cancer, clearly some substrains degranulation when exposed to bladder cancer cells after provoke NK cell degranulation more strongly than others. stimulation with BCG components (not shown). These data Different BCG substrains were generated from 1921 when the suggest that different mycobacteria fractions can activate immune original culture was continuously grown separately in different populations differently and that, depending on the intended use of countries (13, 17). Several genetic groups of BCG appeared BCG, different fragments could be used to produce specific effects. subsequently differing in regions of difference (RD) that have particular deletions, insertions and mutations. RD1 deletion was the first main genetic change that occurred between M. bovis and DISCUSSION BCG. Interestingly, in our degranulation experiments, the stronger increase in response against T24 bladder cells BCG has been used both as TB vaccine and to treat bladder occurred using as stimulation of PBMCs Tice, Connaught and cancer for many decades, however, the heterogeneity of action Moreau. The genetic branch of the Moreau substrain separated and the existence of many different substrains still poses many early, in 1924, before the loss of IS6110, and belongs to the same unresolved questions. This study evaluates the effect of BCG group as Tokyo, 1925, which stimulated a more discrete NK heterogeneity in the immune response exerted by NK cells response. Moreau and Tokyo differ in the regions, RD16 and against bladder cancer cells, as a model of the successful DRv3405c, while Tice (1934) and Connaught (1948) belong to immunotherapy used routinely in NMIBC patients. The data the same genetic group [tandem duplication 2 (DU2) group 4] Frontiers in Immunology | www.frontiersin.org 7 February 2021 | Volume 12 | Article 622995 Esteso et al. Non-Viable BGC and NK Cell Activation A B C FIGURE 4 | Lymphocyte activation using M. bovis and its subcellular fractions. PBMCs from healthy donors were incubated with the indicated fractions of M. bovis obtained through BEI Resources, NIAID, NIH. At day 7, cells in suspension were recovered from the co-culture, centrifuged and analyzed by flow cytometry. (A) The percentage of activation was determined by identifying the resting and activating lymphocyte regions by FSC vs SSC and further analyzed to distinguish T cells from NK cells by staining with antibodies against population markers (CD3, CD56) (top). The percentage of activated lymphocytes was plotted for each donor (different populations are displayed in different shades of grey/black) (bottom). (B) The percentage of the CD56bright population was obtained within the NK cell region and the percentage of CD4 vs CD8 was obtained within the CD3+ region (labelled as CD4/CD8). Data are displayed as different colors for each donor. Statistical analysis was performed using one-way ANOVA comparing each condition with the untreated culture (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, non-significant). (C) Degranulation of CD56bright NK cells against bladder cancer cells (T24, RT-112), was measured by analyzing surface LAMP-1 (CD107a) within the CD3-CD56+ region. Different donors are represented by symbols and compared with the untreated culture. Statistical analysis was performed using one-way ANOVA (*p < 0.05; **p < 0.01; ***p < 0.001; ns, non-significant). and their main difference is the loss of RD15 that occurred in lymphocyte proliferation and NK cell activation in co-cultures Connaught. Whether these genes are involved in the production of PBMCs was similar to the stimulation caused by live bacteria. of a stronger or milder NK cytotoxic response was not explored In certain strains, dead bacteria were even more potent. Early here, but could contribute to the functional differences found animal model experiments already concluded that dead BCG among the different BCG substrains. could mediate an anti-tumor response when bacteria were A second finding from this work is that vaccine preparations applied by intradermal injection (31) and, since then, several contain, in addition to the CFU specified, a large number of dead approaches have been investigated in order to eliminate the use bacteria and this prompted a series of experiments to evaluate the of live bacteria in bladder cancer treatment to avoid toxicity, for effect of dead mycobacteria on stimulation of leukocytes. Here, example using bacterial cell wall and nucleic acids that could we confirm that the capacity of dead bacteria to stimulate benefit patients failing other therapies (32). This study also Frontiers in Immunology | www.frontiersin.org 8 February 2021 | Volume 12 | Article 622995 Esteso et al. Non-Viable BGC and NK Cell Activation A B C FIGURE 5 | Lymphocyte activation using Tice BCG subcellular fractions. PBMCs from healthy donors were incubated with the indicated fractions of Tice BCG (see Materials and Methods). At day 7, cells in suspension were recovered from the co-culture, centrifuged and analyzed by flow cytometry. (A) The percentage of activation was determined by identifying the resting and activating lymphocyte regions by FSC vs SSC and further analyzed to distinguish T cells from NK cells by staining with antibodies against population markers (CD3, CD56). The percentage of activated lymphocytes was plotted for each donor (different populations are displayed in different shades of grey/black) (bottom). (B) The percentage of the CD56bright population was obtained within the NK cell region and percentage of CD4 vs CD8 was obtained within the CD3+ region (labelled as CD4/CD8). Data are displayed as different colors for each donor. Statistical analysis was performed using one-way ANOVA comparing each condition with the untreated culture (*p < 0.05; **p < 0.01; ***p < 0.001; ns, non-significant). (C) Degranulation of CD56bright NK cells against bladder cancer cells (T24, RT-112), was measured by analyzing surface LAMP-1 (CD107a) within the CD3-CD56+ region. Different donors are represented by symbols and compared with the untreated culture. Statistical analysis was performed using one-way ANOVA (*p < 0.05; **p < 0.01; ns, non-significant). provides a first evaluation of the impact of different mycobacterial an increase in effector cells including both CD3+CD56+ and NK components in the activation of NK cells to recognize bladder cells. All these data support the idea that new inactivated BCG- cancer cells. The experiments using M. bovis and BCG-Tice based preparations could be possible both for cancer treatment and subcellular components confirmed that CD56bright NK cells other situations in which NK cell stimulation could be beneficial. In respond strongly against bladder cancer and that this activation fact, new approaches using intravenous BCG to stimulate certain might be modulated by different mycobacterial compounds, but immune populations, such as Th1/Th17 CD4 or CD8 T cells, again, experiments showed differences between the two proposed to increase TB protection in youngster or adults could mycobacteria strains. While the lipidic fraction of M. bovis did benefit from the use of fractions and not viable bacteria (33). Data not stimulate any immune population, the lipids from Tice mainly presented here show that different substrains induce differential stimulated T cell proliferation, increasing the proportion of CD4 vs proliferation of T vs NK cells. Connaught and Pasteur provoke a CD8. However, these lymphocytes did not degranulate in the strong lymphocyte proliferation and a greater percentage of CD3+ presence of bladder cancer cells. On the other hand, delipidated cells than after co-culture of PBMCs with Tice. Culture with these BCG-Tice did not change the proportion of CD4/CD8, but caused substrains also led to an increase in the proportion of effector Frontiers in Immunology | www.frontiersin.org 9 February 2021 | Volume 12 | Article 622995 Esteso et al. Non-Viable BGC and NK Cell Activation FIGURE 6 | Granzyme B and perforin within NK cells activated using BCG-Tice subcellular fractions. PBMCs from healthy donors were incubated with the indicated fractions of BCG-Tice (see Materials and Methods). At day 7, cells in suspension were recovered from the co-culture, centrifuged and analyzed by flow cytometry. The percentage of activation was determined by identifying the resting and activating lymphocyte regions by FSC vs SSC and further analyzed to distinguish T cells from NK cells by staining with antibodies against population markers (CD3, CD56). The percentage of activated lymphocytes was plotted for each donor (different colors) (left). The percentage of the CD56bright population was obtained within the NK cell region and the content of Granzyme B (middle) and perforin (right) were determined by intracellular staining. Statistical analysis was performed using one-way ANOVA (*p < 0.05; **p < 0.01; ***p < 0.001; ns, non-significant) and compared with the untreated culture. lymphocytes CD3+CD56+. Thus, it is tempting to propose that a FUNDING particular immune population could be stimulated selectively by choosing the right BCG substrain or component. In this line, more This work was supported by grants from the Spanish Ministry of research should be conducted to comprehensively identify the Science and Innovation (RTC-2017-6379-1, RTI2018-093569-B- immune cell subpopulations responding to specific mycobacterial I00 (MCIU/AEI/FEDER, EU) (MV-G) and the regional products. The use of NK cells as a therapeutic agent in cancer has government of Madrid (S2017/BMD-3733-2); Generalitat of been long discussed and nowadays, a large body of research is Catalunya (2017SGR-229) (EJ). The group of MV-G belongs to available on the use of NK cells for treatment of cancer, both the research network TENTACLES (RED2018-102411-T) autologous or allogeneic after ex vivo stimulation. Recent research funded by the Spanish Ministry of Science. and clinical trials are starting to generate promising results (34–37). The work presented here could open a new manner for NK stimulation with therapeutic purposes. ACKNOWLEDGMENTS In conclusion, here we demonstrate that BCG stimulates the activation of NK cells through the interaction with mycobacteria The authors would like to thank Dr Sylvia Leao, for the gift of subcellular components and this type of stimulation could be of BCG. The following reagents were obtained through BEI interest beyond bladder cancer treatment. The presence of dead Resources, NIAID, NIH: g-irradiated whole cells mycobacteria in lyophilized BCG preparations suggest that live Mycobacterium bovis, Strain AF 2122/97 (ATCC® BAA-935™) bacilli could be less important than initially thought for cancer Catalog No. NR-31210; Total Lipids, Catalog No. NR-44100, Cell elimination and this opens new possibilities for a desirable Membrane Fraction, Catalog No. NR-31214 and Cytosol decrease of unwanted effects caused by live BCG. Fraction, Catalog No. NR-31215 from M. bovis; Trehalose-6,6- dimycolate (TDM), Catalog No. NR-14844 and Mycolic Acid, Catalog No. NR-14854 from Mycobacterium tuberculosis, Strain DATA AVAILABILITY STATEMENT H37Rv. Data appeared online as a preprint in BioRxiv (BIORXIV/2020/129361). We acknowledge support of the The raw data supporting the conclusions of this article will be publication fee by the CSIC Open Access Publication Support made available by the authors, without undue reservation. Initiative through its Unit of Information Resources for Research (URICI). AUTHOR CONTRIBUTIONS SUPPLEMENTARY MATERIAL GE and NA: acquired, analyzed, and interpreted data. EJ, OA, and MH: generated reagents. NA, CM, and MV-G: provided The Supplementary Material for this article can be found online material support and supervised the study. All authors at: https://www.frontiersin.org/articles/10.3389/fimmu.2021. contributed to the article and approved the submitted version. 622995/full#supplementary-material Frontiers in Immunology | www.frontiersin.org 10 February 2021 | Volume 12 | Article 622995 Esteso et al. Non-Viable BGC and NK Cell Activation REFERENCES 22. Garcia-Cuesta EM, Lopez-Cobo S, Alvarez-Maestro M, Esteso G, Romera- Cardenas G, Rey M, et al. NKG2D is a Key Receptor for Recognition of 1. Shimasaki N, Jain A, Campana D. NK cells for cancer immunotherapy. Nat Bladder Cancer Cells by IL-2-Activated NK Cells and BCG Promotes NK Cell Rev Drug Discovery (2020) 19:200–18. doi: 10.1038/s41573-019-0052-1 Activation. Front Immunol (2015) 6:284. doi: 10.3389/fimmu.2015.00284 2. Morvan MG, Lanier LL. NK cells and cancer: you can teach innate cells new 23. Akaza H, Iwasaki A, Ohtani M, Ikeda N, Niijima K, Toida I, et al. Expression tricks. Nat Rev Cancer (2016) 16:7–19. doi: 10.1038/nrc.2015.5 of antitumor response. Role of attachment and viability of bacillus Calmette- 3. Brandau S, Riemensberger J, Jacobsen M, Kemp D, Zhao W, Zhao X, et al. NK Guerin to bladder cancer cells. Cancer (1993) 72:558–63. doi: 10.1002/1097- cells are essential for effective BCG immunotherapy. Int J Cancer (2001) 0142(19930715)72:2<558::AID-CNCR2820720237>3.0.CO;2-H 92:697–702. doi: 10.1002/1097-0215(20010601)92:5<697::AID- 24. Gunther JH, Frambach M, Deinert I, Brandau S, Jocham D, Bohle A. Effects of IJC1245>3.0.CO;2-Z acetylic salicylic acid and pentoxifylline on the efficacy of intravesical BCG 4. Suttmann H, Jacobsen M, Reiss K, Jocham D, Bohle A, Brandau S. therapy in orthotopic murine bladder cancer (MB49). J Urol (1999) 161:1702– Mechanisms of bacillus Calmette-Guerin mediated natural killer cell 6. doi: 10.1016/S0022-5347(05)69008-8 activation. J Urol (2004) 172:1490–5. doi: 10.1097/01.ju.0000131944.52354.63 25. Shapiro A, Ratliff TL, Oakley DM, Catalona WJ. Reduction of bladder tumor 5. Garcia-Cuesta EM, Esteso G, Ashiru O, Lopez-Cobo S, Alvarez-Maestro M, growth in mice treated with intravesical Bacillus Calmette-Guerin and its Linares A, et al. Characterization of a human anti-tumoral NK cell population correlation with Bacillus Calmette-Guerin viability and natural killer cell expanded after BCG treatment of leukocytes. Oncoimmunology (2017) 6: activity. Cancer Res (1983) 43:1611–5. e1293212. doi: 10.1080/2162402X.2017.1293212 26. Bryceson YT, Fauriat C, Nunes JM, Wood SM, Bjorkstrom NK, Long EO, 6. Wagner JA, Rosario M, Romee R, Berrien-Elliott MM, Schneider SE, Leong et al. Functional analysis of human NK cells by flow cytometry. Methods Mol JW, et al. CD56bright NK cells exhibit potent antitumor responses Biol (2010) 612:335–52. doi: 10.1007/978-1-60761-362-6_23 following IL-15 priming. J Clin Invest (2017) 127:4042–58. doi: 10.1172/ 27. Ho MM, Markey K, Rigsby P, Hockley J, Corbel MJ. Report of an JCI90387 International collaborative study to establish the first WHO reference 7. Pym AS, Brodin P, Brosch R, Huerre M, Cole ST. Loss of RD1 contributed to reagents for BCG vaccines of three different sub-strains. Vaccine (2011) the attenuation of the live tuberculosis vaccines Mycobacterium bovis BCG 29:512–8. doi: 10.1016/j.vaccine.2010.10.066 and Mycobacterium microti. Mol Microbiol (2002) 46:709–17. doi: 10.1046/ 28. Morales A, Johnston B, Emerson L, Heaton JW. Intralesional administration of j.1365-2958.2002.03237.x biological response modifiers in the treatment of localized cancer of the prostate: a 8. Calmette A, Bocquet A, Negre L. Contribution a l’etude du bacilli tuberculeux feasibility study. Urology (1997) 50:495–502. doi: 10.1016/S0090-4295(97)00402-0 bilie. Ann Inst Pasteur (1921) 9:561–70. 29. Morales A, Chin JL, Ramsey EW. Mycobacterial cell wall extract for treatment 9. Gandhi NM, Morales A, Lamm DL. Bacillus Calmette-Guerin of carcinoma in situ of the bladder. J Urol (2001) 166:1633–7discussion 1637– immunotherapy for genitourinary cancer. BJU Int (2013) 112:288–97. doi: 8. doi: 10.1016/S0022-5347(05)65642-X 10.1111/j.1464-410X.2012.11754.x 30. Juliá n E, Noguera-Ortega E. Bacteria-derived alternatives to live Mycobacterium 10. Brosch R, Gordon SV, Garnier T, Eiglmeier K, Frigui W, Valenti P, et al. bovis Bacillus Calmette-Guerin for non-muscle invasive bladder cancer Genome plasticity of BCG and impact on vaccine efficacy. Proc Natl Acad Sci treatment. In: AM Fialho, AM Chakrabarty, editors. Microbial Infections and U.S.A. (2007) 104:5596–601. doi: 10.1073/pnas.0700869104 Cancer Therapy. Singapore: Pan Stanford Publishing Pte. Ltd (2019). p. 123–88. 11. Ritz N, Curtis N. Mapping the global use of different BCG vaccine strains. 31. Zbar B, Rapp HJ. Immunotherapy of guinea pig cancer with BCG. Cancer Tuberculosis (Edinb) (2009) 89:248–51. doi: 10.1016/j.tube.2009.03.002 (1974) 34(suppl):1532–40. doi: 10.1002/1097-0142(197410)34:8+<1532::AID- 12. Zwerling A, Behr MA, Verma A, Brewer TF, Menzies D, Pai M. The BCG CNCR2820340827>3.0.CO;2-H World Atlas: a database of global BCG vaccination policies and practices. PloS 32. Packiam VT, Pearce SM, Steinberg GD. The role of mycobacterial cell wall Med (2011) 8:e1001012. doi: 10.1371/journal.pmed.1001012 nucleic acid complex in the treatment of bacillus Calmette-Guerin failures for 13. Gan C, Mostafid H, Khan MS, Lewis DJ. BCG immunotherapy for bladder non-muscle-invasive bladder cancer. Ther Adv Urol (2016) 8:29–37. doi: cancer–the effects of substrain differences. Nat Rev Urol (2013) 10:580–8. doi: 10.1177/1756287215607818 10.1038/nrurol.2013.194 33. Darrah PA, Zeppa JJ, Maiello P, Hackney JA, Wadsworth MH,2, Hughes TK, 14. Morales A, Eidinger D, Bruce AW. Intracavitary Bacillus Calmette-Guerin in et al. Prevention of tuberculosis in macaques after intravenous BCG the treatment of superficial bladder tumors. J Urol (1976) 116:180–3. doi: immunization. Nature (2020) 577:95–102. doi: 10.1038/s41586-019-1817-8 10.1016/S0022-5347(17)58737-6 34. Suck G, Odendahl M, Nowakowska P, Seidl C, Wels WS, Klingemann HG, 15. Martinez-Pineiro JA, Muntanola P. Nonspecific immunotherapy with BCG et al. NK-92: an ‘off-the-shelf therapeutic’ for adoptive natural killer cell-based vaccine in bladder tumors: a preliminary report. Eur Urol (1977) 3:11–22. doi: cancer immunotherapy. Cancer Immunol Immunother (2016) 65:485–92. doi: 10.1159/000472047 10.1007/s00262-015-1761-x 16. Lamm DL, Thor DE, Harris SC, Reyna JA, Stogdill VD, Radwin HM. Bacillus 35. Guillerey C, Huntington ND, Smyth MJ. Targeting natural killer cells in cancer Calmette-Guerin immunotherapy of superficial bladder cancer. J Urol (1980) immunotherapy. Nat Immunol (2016) 17:1025–36. doi: 10.1038/ni.3518 124:38–40. doi: 10.1016/S0022-5347(17)55282-9 36. Krasnova Y, Putz EM, Smyth MJ, Souza-Fonseca-Guimaraes F. Bench to 17. Behr MA, Small PM. A historical and molecular phylogeny of BCG strains. bedside: NK cells and control of metastasis. Clin Immunol (2017) 177:50–9. Vaccine (1999) 17:915–22. doi: 10.1016/S0264-410X(98)00277-1 doi: 10.1016/j.clim.2015.10.001 18. D’Andrea D, Gontero P, Shariat SF, Soria F. Intravesical bacillus Calmette- 37. Sanchez-Martinez D, Lanuza PM, Gomez N, Muntasell A, Cisneros E, Moraru M, Guerin for bladder cancer: are all the strains equal? Transl Androl Urol (2019) et al. Activated Allogeneic NK Cells Preferentially Kill Poor Prognosis B-Cell 8:85–93. doi: 10.21037/tau.2018.08.19 Chronic Lymphocytic Leukemia Cells. Front Immunol (2016) 7:454. doi: 10.3389/ 19. Witjes JA, Dalbagni G, Karnes RJ, Shariat S, Joniau S, Palou J, et al. The fimmu.2016.00454 efficacy of BCG TICE and BCG Connaught in a cohort of 2,099 patients with T1G3 non-muscle-invasive bladder cancer. Urol Oncol (2016) 34:484 e19–484 Conflict of Interest: The authors declare that the research was conducted in the e25. doi: 10.1016/j.urolonc.2016.05.033 absence of any commercial or financial relationships that could be construed as a 20. Rentsch CA, Birkhauser FD, Biot C, Gsponer JR, Bisiaux A, Wetterauer C, potential conflict of interest. et al. Bacillus Calmette-Guerin strain differences have an impact on clinical outcome in bladder cancer immunotherapy. Eur Urol (2014) 66:677–88. doi: Copyright © 2021 Esteso, Aguiló, Julián, Ashiru, Ho, Martı ́n and Valés‐Gómez. This 10.1016/j.eururo.2014.02.061 is an open-access article distributed under the terms of the Creative Commons 21. Unda-Urzaiz M, Cozar-Olmos JM, Minana-Lopez B, Camarero-Jimenez J, Attribution License (CC BY). The use, distribution or reproduction in other forums is Brugarolas-Rossello X, Zubiaur-Libano C, et al. Safety and efficacy of various permitted, provided the original author(s) and the copyright owner(s) are credited and strains of bacille Calmette-Guerin in the treatment of bladder tumours in standard that the original publication in this journal is cited, in accordance with accepted clinical practice. Actas Urol Esp (2018) 42:238–48. doi: 10.1016/j.acuroe. academic practice. No use, distribution or reproduction is permitted which does not 2018.03.004 comply with these terms. Frontiers in Immunology | www.frontiersin.org 11 February 2021 | Volume 12 | Article 622995