E U RO P E A N U R O L O GY 76 ( 2 019 ) 4 3 0 – 4 3 4 available at www.sciencedirect.com journal homepage: www.europeanurology.com Platinum Priority – Brief Correspondence – Editor’s Choice Editorial by Alberto Martini, John P. Sfakianos and Matthew D. Galsky on pp. 435–436 of this issue Personalized Drug Sensitivity Screening for Bladder Cancer Using Conditionally Reprogrammed Patient-derived Cells Kimmo Kettunen a, Peter J. Boström b, Tarja Lamminen a,b, Taija Heinosalo c, Gun West a, Irena Saarinen a, Katja Kaipio a,b, Juha Rantala d, Chris Albanese e, Matti Poutanen c, Pekka Taimen a,* a b Institute of Biomedicine, University of Turku, and Department of Pathology, Turku University Hospital, Turku, Finland; Department of Urology, Turku University Hospital and University of Turku, Turku, Finland; c Institute of Biomedicine, Research Center for Integrative Physiology and Pharmacology and Turku Center for Disease Modeling, University of Turku, Turku, Finland; d Misvik Biology, Turku, Finland; e Departments of Oncology and Pathology, Lombardi Comprehensive Cancer Center, Georgetown University, Washington, DC, USA Article info Abstract Article history: Many patients with muscle-invasive bladder cancer (BC) are either ineligible for or do Accepted June 12, 2019 not benefit from cisplatin-based chemotherapy, and there is an unmet need to estimate individuals’ drug sensitivities. We investigated the suitability of conditionally repro- Associate Editor: grammed (CR) cells for the characterization of BC properties and their feasibility for James Catto personalized drug sensitivity screening. The CR cultures were established from six BC tumors with varying histology and stage. Four cultures were successfully propagated for Statistical Editor: genomic, transcriptomic, and protein expression profiling and compared to the parental tumors. Two out of four CR cultures (urothelial carcinoma and small cell neuroendocrine Andrew Vickers carcinoma [SmCC]) corresponded well to their parental tumors and underwent drug sensitivity screening to identify novel drugs for the respective tumors. Both cultures Keywords: were sensitive to standard BC chemotherapy agents (eg cisplatin and gemcitabine) and Bladder cancer to conventional drugs such as taxanes and inhibitors of topoisomerase and proteasome. The SmCC cells were also sensitive to statins (eg, atorvastatin and pitavastatin). In Chemotherapy summary, after confirming their representativeness and origin, we conclude that CR Conditional reprogramming cells are a feasible platform for personalized drug sensitivity testing and might thus add Drug sensitivity testing to the approaches used to personalize BC treatment strategies. Patient-derived cultures Patient summary: We investigated the conditional reprogramming method for gener- Small cell carcinoma ating patient-derived bladder cancer cell cultures and studied their feasibility for planning personalized treatment strategies. © 2019 The Authors. Published by Elsevier B.V. on behalf of European Association of Urology. This is an open access article under the CC BY-NC-ND license (http://creati- vecommons.org/licenses/by-nc-nd/4.0/). Please visit www.eu-acme.org/europeanurology to answer questions on-line. The EU- ACME credits will then be attributed automatically. * Corresponding author. Institute of Biomedicine, University of Turku, Kiinamyllynkatu 10, 20520 Turku, Finland. Tel. +358 29 450 4524. E-mail address: pekka.taimen@utu.fi (P. Taimen). https://doi.org/10.1016/j.eururo.2019.06.016 0302-2838/© 2019 The Authors. Published by Elsevier B.V. on behalf of European Association of Urology. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). E U RO P E A N U RO L O GY 76 ( 2 019 ) 4 3 0 – 4 3 4 431 Muscle-invasive bladder cancer (BC) is a mutationally overgrowth of benign epithelial cells, highlighting the need heterogeneous malignancy with poor survival. Almost half for stringent sampling of tumor material and validation of of patients with this disease are ineligible for cisplatin- the origin of the cells via sequencing [5]. based adjuvant or neoadjuvant chemotherapy, and only a We applied the CR platform to generate patient-derived subset benefits from treatment [1]. Currently there are no cell cultures from six BC patients undergoing cystectomy or validated means in the clinical setting to predict transurethral resection (Fig. 1A). Four patients had pTaN0– chemotherapy responses and/or clinical outcomes. T4N1 high-grade urothelial carcinoma (HG-UC), one patient As BC is characterized by a lack of common druggable had pT4aN1 small-cell neuroendocrine carcinoma (SmCC), driver mutations and high intertumor heterogeneity, and one patient had pT2bN1 primary bladder adenocarci- recapitulation of tumor characteristics ex vivo would offer noma (Fig. 1A and B, Supplementary Table 1). The cultures tools for studying drug sensitivities for individual patients were established according to a published CR procedure [6] upfront and potential novel treatment approaches. Various with some modifications (Supplementary material). The patient-derived organoid and xenograft BC models have cultures were considered stably established if they could be recently been established and used to identify personalized cultured for five passages, and after cryopreservation drug sensitivities [2]. Although these three-dimensional repropagated for follow-up analysis. (3D) models replicate the tumor environment, it may not be Four (3 HG-UCs and 1 SmCC) out of six CR cultures were feasible to use them for routine high-throughput diagnos- successfully repropagated after cryopreservation (success tics because of limitations associated with the culture rate 67%; Supplementary Table 1) and were denoted HG-Ta- complexity and the inability to accurately assess responses CR, HG-T1-CR, HG-T4-CR, and SmCC-T4-CR. All the urothelial at the individual cell level. CR cultures shared typical CR culture morphology, while the Conditional reprogramming (CR) is a method that allows SmCC-T4-CR cells had the expected smaller cellular size rapid expansion of malignant and normal epithelial cells (Fig. 1C and E). HG-T1-CR and SmCC-T4-CR cells showed without genetic manipulation. The CR approach has been exponential growth during 30-d follow-up (Fig. 1D). Exome used to generate patient-derived cultures from various sequencing analysis showed that two of the cultures (HG-T1- neoplasms, including prostate cancer and respiratory CR and SmCC-T4-CR) retained the majority of mutations (eg, papillomatosis [3,4]; however, the method has not been in RB1) found in the corresponding tumors, whereas the HG- applied in human BC. The main challenge has been the Ta-CR and HG-T4-CR cultures did not, suggesting that normal [(Fig._1)TD$IG] Fig. 1 – Establishment of conditionally reprogrammed (CR) cultures. (A) Overview of the study design. CR cultures were established from fresh tumor samples obtained from radical cystectomy (RC) or transurethral resection of bladder tumor (TUR-BT). The cultures and the corresponding tumor samples were characterized by immunohistochemical (IHC) analyses, Western blotting, and whole-exome DNA sequencing and RNA sequencing. Finally, high-throughput drug sensitivity screening was performed on the cancerous CR cultures. (B) Histology of parental tumors (low- and high-power magnifications). Three patients had high-grade (HG) urothelial carcinoma staged as pTa, pT1, and pT4aN1 (HG-Ta, HG-T1 and HG-T4, respectively; hematoxylin and eosin stain), and one patient had small cell neuroendocrine carcinoma (SmCC; Van Gieson stain) staged as pT4aN1 (SmCC-T4). Scale bars, 100 mm. Only tumors with stably established CR cultures are shown. (C) Representative phase-contrast microscopy image showing the typical cell morphology of a CR culture (example from HG-T1 CR). Scale bars, 50 mm. (D) Proliferation assays of HG-T1-CR and SmCC-T4-CR cultures. Both cultures showed exponential growth over the 30-d follow-up period. (E) May-Grünwald Giemsa-stained cytospin samples from CR cultures. SmCC-T4-CR cells showed an overall smaller cellular size compared to cells from urothelial carcinomas. Scale bars, 50 mm. 432 E U R O P E A N U R O L O GY 76 ( 2 019 ) 4 3 0 – 4 3 4 [(Fig._2)TD$IG] Fig. 2 – Characterization of conditionally reprogrammed (CR) cultures and their parental tumors. (A) For each tumor and cell-line pair, Venn diagrams depict the total number of mutations per sample with the number of co-occurring mutations shown in the overlapping circles. Tumor samples are depicted in green and cell cultures in blue. HG-T1-CR and SmCC-T4-CR cells retained the majority of mutations found in the corresponding tumors, whereas HG-Ta-CR and HG-T4-CR did not. (B) Top panel: Tumor mutation burden shown as synonymous (green) and nonsynonymous (blue) mutation rates per Mb. Bottom panel: The significantly mutated genes in the CR cultures (C) and in the corresponding tumors (T), as reported in The Cancer Genome Atlas BC cohort [11]. Only the exonic or splicing mutations found in either HG-T1 or SmCC-T4 tumors/cells are shown. Supplementary Fig. 1 and Supplementary Tables 2 and 3 show individual mutations also found in HG-Ta and HG-T4 tumors/cells. (C) Immunohistochemistry analysis of CR cultures and the HG-T1 and SmCC-T4 parental tumors shown as a heat map. The HG-T1 tumor shows characteristics of the luminal type, while the CR culture shows transition towards a more basaloid/squamous phenotype as observed by the expression of keratin 5/6 (CK5/6) and reduced expressions of keratin 20 (CK20) and GATA3 in the corresponding CR culture. The SmCC-T4 tumor and CR culture correspond to the neuronal type. The number indicates the percentage of positively stained cells. CGA = chromogranin A. (D) Gene expression clustering analysis shows distinct sample groups E U RO P E A N U RO L O GY 76 ( 2 019 ) 4 3 0 – 4 3 4 433 epithelial cells had overgrown in these two cultures (Fig. 2A failed to retain the specific driver mutations found in their and B, Supplementary Fig. 1, Supplementary Tables 2 and 3). parental tumors, suggesting contamination by nonmalig- On immunohistochemical analysis, HG-T1-CR cells nant cells. Therefore, detailed selection of original tumor showed strong keratin 5/6 expression, indicating a shift material and genomic analysis are crucial to confirm the towards a more basaloid phenotype (Fig. 2C, Supplementary origin of the established culture. Fig. 2). In SmCC-T4-CR cells, strong expression of neuroen- The results of personalized drug screening are of docrine markers and loss of keratins were detected, similar particular interest and might point to novel therapy targets to the parental tumor (Fig. 2C, Supplementary Fig. 2). Both for individual patients or subgroups. While clinically used cancer-originating cultures were strongly p53-positive and platin-based compounds were effective in the CR cultures in showed a higher percentage of Ki-67-positive cells when vitro, we also observed high sensitivity of SmCC-T4-CR cells compared to noncancerous cultures (Fig. 2C, Supplementa- to statins, which could be a promising, well-tolerated, and ry Figs. 2 and 3). low-cost candidate for further studies. The specific mecha- In RNA sequencing and gene set enrichment analysis, the nism of statins depends on the type of statin used and the expression of cell adhesion-related transcripts clearly type of cancer cells. In case of small cell lung cancer, the differentiated the primary tumors from CR cells cultured mechanism seems to be HMG-CoA reductase inhibition of in the absence of stromal cells (Fig. 2D and E). Genes cholesterol biosynthesis and subsequent impairment of Ras representative of basal cell fate (eg, KRT5, KRT6, KRT17, and signaling [7]. In addition, for some urothelial BC cell KRT19) were enriched in CR cultures of urothelial origin, cultures, statins appear to induce cell cycle arrest and while SmCC-T4-CR and the parental tumor were distinct inhibit proliferation via the PPARg signaling pathway [8]. from the other cultures and tumors on the basis of strong While the results from the current study are encourag- expression of neuroendocrine genes. The CR cultures showed ing, the data are limited by the small number of the cultures very little gain of functionality compared to the primary produced and potential intratumor heterogeneity. Culture tumors. However, there was some increase in translation- conditions, including the culture media used and the related pathways and ribogenesis. Furthermore, the cancer- absence of stromal and inflammatory cells, may also affect originating cultures exhibited significantly higher expression the tumor microenvironment and drug sensitivities of various cell cycle-related genes and lower expression of [9,10]. Since conditioned media appear to be a prerequisite CCND1/CCND2 (due to loss of RB1) compared to the for successful culture, a more detailed characterization of noncancerous cultures (Supplementary Table 4). the impact on drug sensitivities during the development of Drug sensitivity screening performed on cancer-origi- rapid 3D culture models needs to be tested in future studies. nating cultures showed sensitivity to platinum-based drugs, taxanes, topoisomerase inhibitors, and proteasome Author contributions: Pekka Taimen had full access to all the data in the inhibitors, independent of the difference in the proliferation study and takes responsibility for the integrity of the data and the rates of the cell cultures (Fig. 2F, Supplementary Fig. 4). The accuracy of the data analysis. HG-T1-CR cells were resistant to the EGFR inhibitor Study concept and design: Kettunen, Boström, Taimen. erlotinib (in line with the activating E322K mutation Acquisition of data: Kettunen, Boström, Lamminen, Heinosalo, West, detected in MAPK1) while the SmCC-T4-CR cells were Saarinen, Kaipio, Rantala, Taimen. highly responsive to statins at low concentrations. The Analysis and interpretation of data: Kettunen, Boström, Lamminen, effects of cisplatin (standard treatment) and atorvastatin Heinosalo, West, Saarinen, Kaipio, Rantala, Taimen. (because of a high statin response) were further validated Drafting of the manuscript: Kettunen, Boström, Poutanen, Taimen. with more detailed sensitivity measurements, with highly Critical revision of the manuscript for important intellectual content: similar results compared to screening obtained (Fig. 2G). Boström, Albanese, Poutanen, Taimen. In the current study we demonstrated the feasibility of Statistical analysis: Kettunen, Rantala. establishing patient-derived BC CR cultures that retain the Obtaining funding: Boström, Poutanen, Taimen. hallmark mutations of the primary tumor without signifi- Administrative, technical, or material support: Boström, Albanese, Poutanen, Taimen. cant phenotypic drift. Furthermore, the CR cultures were Supervision: Boström, Poutanen, Taimen. suitable for drug sensitivity screening. Exome sequencing Other: None. analysis demonstrated that two of the four CR cultures characterized matched the corresponding tumor, as seen in Financial disclosures: Pekka Taimen certifies that all conflicts of interest, the previous studies [3]. However, the other two CR cultures including specific financial interests and relationships and affiliations relating to the tumor lineage (epithelial or neuroendocrine) and sample type (cell culture or primary tumor). Transcripts were selected on the basis of high variation across samples (n = 825): (1) genes representative of a basal cell fate, indicating enrichment of basal/tumor cells (eg, KRT5, KRT6, KRT17 and KRT19 positive) in the CR cultures compared to tumors; (2) cell adhesion molecules upregulated in tumors compared to the cell cultures (presumably due to lack of stroma in cell cultures); and (3) neuroendocrine genes upregulated in SmCC compared to UCs. (E) Gene set enrichment analysis further details the expression profiles that differentiate the sample classes. The top panel identifies cell adhesion as one of the main transcriptional themes differentiating primary tumors from the cell cultures (cluster 2 in D). The bottom panel clearly shows that both the SmCC tumor and SmCC-T4-CR retained strong neuroendocrine characteristics compared to the other samples (cluster 3 in D). (F) Heatmap of responses of the cancerous CR cultures to various drugs, including OSI-420 (desmethyl erlotinib), atorvastatin, the MEK1/2 inhibitor trametinib, a large group of commonly used chemotherapy agents (eg, cisplatin/carboplatin, taxanes), pitavastatin, and carboplatin. The intensity of the blue color correlates with sensitivity to a drug and reduced viability. Supplementary Fig. 4 shows the complete screening results. (G) Dose–response curves for HG-T1-CR and SmCC-T4-CR cells treated with cisplatin and atorvastatin. Both cultures were sensitive to cisplatin, while SmCC-T4-CR showed significantly reduced viability after treatment with atorvastatin compared to the HG-T1-CR culture. Data points denote the mean  standard deviation (n = 3). 434 E U R O P E A N U R O L O GY 76 ( 2 019 ) 4 3 0 – 4 3 4 relevant to the subject matter or materials discussed in the manuscript [3] Saeed K, Rahkama V, Eldfors S, et al. Comprehensive drug testing of (eg, employment/affiliation, grants or funding, consultancies, honoraria, patient-derived conditionally reprogrammed cells from castration- stock ownership or options, expert testimony, royalties, or patents filed, resistant prostate cancer. Eur Urol 2017;71:319–27. http://dx.doi. received, or pending), are the following: Juha Rantala is founder of org/10.1016/j.eururo.2016.04.019. Misvik Biology. The remaining authors have nothing to disclose. [4] Yuan H, Myers S, Wang J, et al. Use of reprogrammed cells to identify therapy for respiratory papillomatosis. N Engl J Med Funding/Support and role of the sponsor: This study was supported by 2012;367:1220–7. http://dx.doi.org/10.1056/NEJMoa1203055. ERVA funding from the Hospital District of Southwest Finland (P.T.), [5] Sette G, Salvati V, Giordani I, et al. Conditionally reprogrammed cells Sigrid Juselius Foundation (P.J.B.) and the Finnish Cancer Foundation (P.T., (CRC) methodology does not allow the in vitro expansion of patient- P.J.B.). The sponsors played no direct role in the study. derived primary and metastatic lung cancer cells. Int J Cancer Acknowledgments: We thank Sinikka Collanus and Anu Salminen for 2018;143:88–99. http://dx.doi.org/10.1002/ijc.31260. their valuable help with immunohistochemistry and cell culture, [6] Liu X, Krawczyk E, Suprynowicz FA, et al. Conditional reprogram- respectively; the Finnish Functional Genomics Centre (University of ming and long-term expansion of normal and tumor cells from Turku and Åbo Akademi University) and Biocenter Finland for their human biospecimens. Nat Protoc 2017;12:439–51. http://dx.doi. expertise in sequencing and imaging techniques; Erik Fredlund org/10.1038/nprot.2016.174. (DoubleStrand Bioinformatics, Stockholm, Sweden) for sequencing data [7] Seckl MJ, Ottensmeier CH, Cullen M, et al. Multicenter, phase III, analysis; and Auria Biobank for help in tissue collection. randomized, double-blind, placebo-controlled trial of pravastatin added to first-line standard chemotherapy in small-cell lung cancer (LUNGSTAR). J Clin Oncol 2017;35:1506–14. http://dx.doi.org/10. 1200/JCO.2016.69.7391. Appendix A. Supplementary data [8] Wang G, Cao R, Wang Y, et al. Simvastatin induces cell cycle arrest and inhibits proliferation of bladder cancer cells via PPARg signal- Supplementary data associated with this article can be ling pathway. Sci Rep 2016;6:35783. http://dx.doi.org/10.1038/ found, in the online version, at https://doi.org/10.1016/j. srep35783. eururo.2019.06.016. [9] Kodack DP, Farago AF, Dastur A, et al. Primary patient-derived cancer cells and their potential for personalized cancer patient care. Cell Rep 2017;21:3298–309. http://dx.doi.org/10.1016/j. References celrep.2017.11.051. [1] Witjes JA, Lebret T, Compérat EM, et al. Updated 2016 EAU guide- [10] Suprynowicz FA, Kamonjoh CM, Krawczyk E, et al. Conditional cell lines on muscle-invasive and metastatic bladder cancer. Eur Urol reprogramming involves non-canonical b-catenin activation and 2017;71:462–75. http://dx.doi.org/10.1016/j.eururo.2016.06.020. mTOR-mediated inactivation of Akt. PLoS One 2017;12:e0180897, [2] Lee SH, Hu W, Matulay JT, et al. Tumor evolution and drug https://doi.org/10.1371/journal.pone.0180897. response in patient-derived organoid models of bladder [11] Robertson AG, Kim J, Al-Ahmadie H, et al. Comprehensive molecular cancer. Cell 2018;173:. http://dx.doi.org/10.1016/j.cell.2018.03.017, characterization of muscle-invasive bladder cancer. Cell 2017;171:. 515–28.e17. http://dx.doi.org/10.1016/j.cell.2017.09.007, 540–56.e25.