Organoid Cultures As Preclinical Models of Non–Small Cell Lung

Published OnlineFirst November 6, 2019; DOI: 10.1158/1078-0432.CCR-19-1376

CLINICAL CANCER RESEARCH | TRANSLATIONAL CANCER MECHANISMS AND THERAPY

Organoid Cultures as Preclinical Models of Non–Small Cell Lung Cancer Ruoshi Shi1,2, Nikolina Radulovich1, Christine Ng1, Ni Liu1, Hirotsugu Notsuda1, Michael Cabanero1, Sebastiao N. Martins-Filho1, Vibha Raghavan1, Quan Li1, Arvind Singh Mer1, Joshua C. Rosen1,3, Ming Li1, Yu-Hui Wang1, Laura Tamblyn1, Nhu-An Pham1, Benjamin Haibe-Kains1,2,4,5,6, Geoffrey Liu1,2,7, Nadeem Moghal1,2, and Ming-Sound Tsao1,2,3

ABSTRACT ◥ Purpose: Non–small cell lung cancer (NSCLC) is the most Results: We have identified cell culture conditions favoring the common cause of cancer-related deaths worldwide. There is an establishment of short-term and long-term expansion of NSCLC unmet need to develop novel clinically relevant models of NSCLC to organoids derived from primary lung patient and PDX tumor tissue. accelerate identification of drug targets and our understanding of The NSCLC organoids recapitulated the histology of the patient and the disease. PDX tumor. They also retained tumorigenicity, as evidenced by Experimental Design: Thirty surgically resected NSCLC cytologic features of malignancy, xenograft formation, preservation primary patient tissue and 35 previously established of mutations, copy number aberrations, and gene expression pro- patient-derived xenograft (PDX) models were processed files between the organoid and matched parental tumor tissue by for organoid culture establishment. Organoids were histo- whole-exome and RNA sequencing. NSCLC organoid models also logically and molecularly characterized by cytology and preserved the sensitivity of the matched parental tumor to targeted histology, exome sequencing, and RNA-sequencing analysis. therapeutics, and could be used to validate or discover biomarker– Tumorigenicity was assessed through subcutaneous injection drug combinations. of organoids in NOD/SCID mice. Organoids were subjected Conclusions: Our panel of NSCLC organoids closely recapitu- to drug testing using EGFR, FGFR, and MEK-targeted lates the genomics and biology of patient tumors, and is a potential therapies. platform for drug testing and biomarker validation.

Introduction tumors or drug sensitivity to targeted therapeutics of their patient tumors (5). In addition, although GEMMs and clinically relevant Non–small cell lung cancer (NSCLC) is the leading cause of cancer- PDXs may be closer to the ideal models to study drug response in related death worldwide with a 5-year overall survival rate of 15% (1). patients, studies using these models are labor intensive, costly, and Over the last decades, there has been tremendous effort in developing time consuming (6). Thus, research efforts are underway to develop preclinical models of NSCLC, including two-dimensional (2D) cell novel preclinical models derived from patient with NSCLC and PDX lines, air–liquid interface cultures, genetically engineered mouse mod- tissue that are economical, rapid to use, and accurately reflect the els (GEMM), and patient-derived xenografts (PDX; refs. 2–4). These biology of the disease. models have been used to accelerate our understanding of NSCLC Over the past few years, organoid cultures derived from primary biology and pathogenesis. Although cell lines are still widely used in patient tumors and PDXs of various cancers including the colon, preclinical studies, they often do not reflect the biology of their parental pancreas, prostate, liver, and breast have been described (7–16). These cancer organoids have been utilized for numerous applications, such as fi – 1University Health Network, Ontario Cancer Institute/Princess Margaret Cancer drug screening and biomarker identi cation (17 20). They have been Centre, Toronto, Ontario, Canada. 2Department of Medical Biophysics, Univer- proposed to be better in vitro models than 2D cell lines due to higher sity of Toronto, Toronto, Ontario, Canada. 3Department of Laboratory Medicine rates of preservation of key histologic and molecular traits of their 4 and Pathobiology, University of Toronto, Toronto, Ontario, Canada. Depart- parental tumors (14, 15). In addition, drug screening in patient- ment of Computer Science, University of Toronto, Toronto, Ontario, Canada. derived organoids (POD) has shown high concordance with that of 5Ontario Institute for Cancer Research, Toronto, Ontario, Canada. 6Vector Institute for Artificial Intelligence, Toronto, Ontario, Canada. 7Division of Medical the matched patient tumor (14, 18). Some reports have demonstrated Oncology and Hematology, Princess Margaret Cancer Centre, University of the ability to generate normal lung organoids composed of airway cell Toronto, Toronto, Ontario, Canada. lineages (21, 22). These models were primarily generated from normal Note: Supplementary data for this article are available at Clinical Cancer mouse and human airways to understand normal lung development Research Online (http://clincancerres.aacrjournals.org/). and function. In addition, methods to generate lung organoids from pluripotent stem cells have been reported to aid in the study of genetic R. Shi and N. Radulovich contributed equally to this article. pulmonary diseases such as cystic fibrosis (21, 23). A major advance Corresponding Author: Ming-Sound Tsao, University Health Network, 101 was outlined in recent reports describing protocols for the develop- College Street 11-314, Toronto, Ontario M5G 1L7, Canada. Phone: 416-634- – 8721; E-mail: [email protected] ment of NSCLC organoids (24 26). However, although many of the models reported in these studies were cultured short-term and were – Clin Cancer Res 2020;XX:XX XX useful for acute studies, lack of systematic documentation of doi: 10.1158/1078-0432.CCR-19-1376 organoid tumor cell purity was a significant issue and specific details 2019 American Association for Cancer Research. regarding long-term growth of the models were not provided (25, 26).

AACRJournals.org | OF1

Shi et al.

resuspended in 100% growth factor–reduced Matrigel (VWR), plated Translational Relevance in 24-well tissue culture plates as Matrigel domes and maintained in Currently, there is an urgent need for clinically relevant pre- 37 C5%CO2 with media overlaying the Matrigel dome. Refer to clinical models of non–small cell lung cancer (NSCLC) for bio- Supplementary Materials and Methods for a list of media components. marker and drug discovery due to the lack of preclinical models Organoid growth was monitored weekly for the detection of initiated that recapitulate the biology of the patient tumor. Three- organoids, and organoids were kept in the same passage for no longer dimensional (3D) organoids have become valuable preclinical than four weeks. The identity of PDX and organoids were authenti- models to study disease pathogenesis and identify novel drug cated by short tandem repeat (STR) analysis and matched to patient targets. We have established a protocol for the development tissue. Organoid cultures were tested routinely for Mycoplasma. of NSCLC organoids from patient tumor and patient-derived Additional methods can be found in Supplementary Materials and xenograft models. This protocol allowed for the efficient generation Methods. of organoids for multiple potential applications. Importantly, we showed that these organoids retained the histologic and molecular IHC features of their parental tumors and demonstrated their utility for Fresh tumor tissue was fixed in 10% formalin for 24 to 48 hours, drug testing. Our organoid platform provides additional preclinical followed by fixation in 70% ethanol prior to paraffin embedding. models of NSCLC and may be useful for future drug screening Organoids were fixed with 10% formalin for 24 to 48 hours and 70% biomarker identification. ethanol with eosin and embedded in Histolgel (Thermo Fisher Sci- entific) before processing for H&E and IHC. Formalin-fixed paraffin- embedded tumor tissues and organoids were cut into 4-mm-thick slices and allowed to dry overnight at 60C. Prepared tissue sections were Furthermore, there still remains a great need to develop a NSCLC stained with appropriate antibodies using BenchMark XT autostainer organoid platform suitable for drug screening and biomarker identi- (Ventana Medical Systems). Primary antibody specific to CK5/6 fication in lung cancer. (Ventana), TP63, TTF-1, and CK7 (Dako) were used for IHC analysis. Here, we describe a culturing protocol that enables generation of The slides were scanned and imaged using Aperio Scanscope XT short-term (1–3 months, 1–9 passages) and long-term (>3 months, (Leica). >10 passages) NSCLC organoids from most and a subset of primary lung patient tumors and PDXs, respectively. These models were able to DNA extraction and WES analysis initiate from tumor tissues with 88% (57/65) success rate. Specifically, Snap-frozen tumor tissues and fresh organoid pellets were lysed in 72% (47/65) of the organoids were maintained in culture short-term, tris-buffered saline solution with 10% SDS and proteinase K (1 mg/mL) whereas 15% (10/65) were maintained in culture long-term. We overnight at 55C. DNA was isolated and eluted on spin columns using demonstrated that both short-term and long-term established NSCLC proprietary solutions provided by a DNA Extraction Kit (Norgen organoids grown in vitro and as xenografts recapitulated the histologic Biotek). DNA quality was assessed using Bioanalyzer, Tapestation, and features and tumorigenicity of their matched tumor tissue. Whole- qPCR. One-hundred to 200 ng of genomic DNA was used for library exome sequencing (WES) and RNA-sequencing revealed that the preparation (Agilent SureSelect Human All Exon v5 Capture Kit). long-term NSCLC organoids, despite having been grown in in vitro DNA was sequenced using 125-cycle paired-end protocol and multi- environments with multiple passaging, preserved the mutation, copy plexing to obtain 150 coverage on Illumina Hiseq2500 sequencer. number, and gene expression profiles of their parental tumors. Finally, Xenome (29) was used to eliminate reads pertaining to mouse stroma. we demonstrate that these biologically relevant models of NSCLC can Sequence reads were subsequently aligned to the human reference be used for drug testing, supporting their application for both disease genome (GRCh37) using Burrows-Wheeler Aligner v0.7.12 (30). The modeling and therapeutic testing. mapped data were further processed for quality control using the standard GATK pipeline, including Picard v1.140 (31), Mutect v1.1.5 (32), and Varscan v2.3.8 (33) were used for mutation calling, Materials and Methods whereas dbSNP (34), ExAC (35), and ESP (36) were used as filters for Tumor tissue processing and organoid establishment samples without matched normal tissue. Annovar (37), vcf2maf The collection of surgically resected primary tumors from patients v1.6.14, and Variant Effect Predictor v87 (38) were used to annotate with early-stage NSCLC and the development of PDXs were approved final mutation calls, following which, the R package “ComplexHeat- by the University Health Network Research Ethics Board (REB: 17- map” (39) was used to generate oncoprints and visualize the data. CNV 558) and Animal Care Committee (AUP: 5555). Informed written kit (40) was used to infer copy number from exome-sequenced samples consent was received from all patients. All studies were performed in by applying circular binary segmentation (CBS; ref. 41) to make calls in accordance with TRI-Council Policy Statement: Ethical Conduct for both targeted regions and nontargeted regions. The targeted regions Research Involving Humans. Clinical diagnosis of NSCLC subtypes used for this algorithm were a combination of SureSelect Human All was validated by pathologic review. The protocol for establishing Exon V4 and V5 regions. A panel of normal lung tissue was used for NSCLC PDXs was previously described (4, 27, 28). For organoid samples lacking a matched normal. Subsequently, GISTIC2.0 (42) was cultures, tumor tissues were processed into 4-mm-diameter pieces and run to identify genes affected by copy number alterations, while also washed with ice-cold PBS. Tumor pieces were dissociated into single taking into account the frequency and amplitude of the events. WES cells in Advanced DMEMF12 (GIBCO) with Liberase TM (Sigma) for data were deposited in the sequence read archive (SRA; accession no.: 1 hour followed by 10-minute incubation with TrypLE Express SRP158596). (Invitrogen) in 37C with gentle shaking. Mouse cell depletion in PDX samples was performed after tissue dissociation using H-2Kb/ RNA extraction and RNA-sequencing analysis H-2Db antibody (#MA5-17998; Invitrogen) labeling and Streptavidin Organoids were extracted from Matrigel using Cell Recovery (BD Biosciences) bead magnetic separation. Cells were counted and Solution (Corning) on ice for 1 hour. Total RNA from homogenized

OF2 Clin Cancer Res; 2020 CLINICAL CANCER RESEARCH

Non–Small Cell Lung Cancer Organoids

tumor tissue and pelleted organoids were extracted using TRIzol this study: pFGFR (Y653/654; #3471), FGFR1 (#9740), pErk (Invitrogen) method, followed by isolation and precipitation in (T202/Y204; #9101), Erk (#9102), pAkt (S473; #9271), and Akt chloroform and 70% ethanol. DNA cleanup was performed using (#9272) were obtained from Cell Signaling Technology. b-Actin DNA Cleanup Kit (Invitrogen). Total RNA was quality checked via antibody (#A1978) was obtained from Sigma. BioAnalyzer (Agilent), Tapestation, and qPCR. Library preparation was performed using Illumina TruSeq Stranded mRNA Sample In vivo organoid implantations Preparation Kit (Illumina). RNA was sequenced using HiSeq 2000 Dissociated organoids were isolated from growth factor–reduced sequencer with 75-cycle paired end protocol and multiplexing to Matrigel using Cell Recovery Solution (Corning) for 1 hour on ice. obtain 40–80 million reads/sample. Xenome (version 1.0.1 with Organoids were resuspended with 500,000 cells in 200 mLM26 standard parameters; ref. 29) was used to filter mouse reads from media prior to injection in the subcutaneous flank of 4- to 6-week- human reads. For transcript quantification, Salmon (version 0.8.2 old NOD/SCID mice. Tumor growth was monitored once or twice with default parameters; ref. 43) with quasi-mapping was applied to weekly by caliper measurement. Tumors were harvested, formalin- assign reads directly to transcripts to obtain transcripts per million fixed paraffin embedded for histologic analysis, and snap frozen for þ (TPM) values. The log2(TPM 1) were used for all statistical DNA/RNA/protein isolation. analysis. ComBat (44) was applied to adjust for batch effects. For correlation analysis, genes that are differentially expressed between In vivo therapeutic studies LUAD and LUSC at a 2-fold or greater cutoff were identified from Cryopreserved PDX tissue (below passage 10) was thawed and profiling of PDX models (4) or primary patient tumors (TCGA). implanted into the subcutaneous flank of NOD/SCID mice. The tumor These gene sets consisted of 893 and 1,492 differentially expressed was harvested and cut into 4-mm–diameter pieces at endpoint and genes, respectively, and were used to calculate correlation coeffi- expanded into experimental arms for drug testing when the average cients between patient, PDX, and organoids. RNA-sequencing data size reached 150 to 200 mm3. Trametinib (1 mg/kg) and BGJ398 were deposited in the Gene Expression Omnibus (accession no. (25 mg/kg) were dissolved in 0.5% hydroxyethyl cellulose with 0.2% GSE119004). Tween80 in sterile H2O and 10% Tween80, respectively. Compounds were delivered once daily via oral gavage for 21 to 28 days. Tumor size In vitro drug studies was monitored twice weekly by caliper measurement. For in vitro drug testing, compounds were purchased from UHN Shanghai and dissolved in DMSO. Organoids were dissociated into Statistical analysis single cells, counted, and plated in Matrigel-coated 384 well plates All exome sequencing and RNA-sequencing analysis were per- (3,000 cells per well) in triplicate for 24 hours prior to drug treatment. formed in the open-source R Statistical Computing software (http:// – Organoids were treated with a range of drug concentrations (0.01 www.r-project.org/). All statistical analysis for obtaining IC50s for drug 10 mmol/L) for 96 hours and cell viability was determined by CellTiter screening were performed in GraphPad prism 6.0. Combination Glo 3D viability assay (protocol mentioned above). Drug–response indices for drug combination studies were performed in CompuSyn P in vivo curves were graphed and IC50 values were calculated using Graphpad (http://www.combosyn.com/). values in the drug studies were Prism 6.0. CompuSyn software (45) was used to calculate combination obtained using Student t test at specific time points. indices for combination drug studies. qRT-PCR Results Total RNA was extracted according to the methods mentioned Organoid establishment from NSCLC patient tumor and PDXs above. RNA was reverse transcribed to cDNA using a Reverse Tran- From December 2015 to 2017, 19 surgically resected lung scription Kit (Thermo Fisher Scientific). Primers used for qPCR adenocarcinomas (LUAD), 15 lung squamous cell carcinomas included FGFR1 F50-GCATCAACCACACATACCAGC-30, FGFR1 (LUSC), 16 LUAD PDXs, and 26 LUSC PDXs were processed for R50-CACGTTGCTACCCAGGGC-30, ACTB F50-TCCTAAAAGC- organoid establishment (Fig. 1A and B; Supplementary Table S1). CACCCCACTTCT-30, ACTB R50-GGGAGAGGACTGGGCCATT- Note that the nomenclature LPTO and PDXO were used to denote 30, B2M F50-GAGTGCTGTCTCCATGTTTGATGT-30, B2M R50- organoid models derived from lungprimarypatienttumorand AAGTTGCCAGCCCTCCTAGAG-30. The following conditions were PDX, respectively. Of the 76 tissues processed, 11 models were used for qPCR: 94C for 1 minute, 60C for 30 seconds, and 72C for 1 excluded from the final count due to lack of starting tumor cell minute for 35 cycles. material, mouse/normal epithelial cell contamination, and sites of metastasis (Supplementary Table S1). We attempted to grow orga- Western blotting noids in advanced DMEMF12 base media with additional supple- Matrigel/organoid suspension was dissociated with TrypLE ments that we termed M26 (Supplementary Materials and Methods, Express and organoid pellets were lysed with RIPA buffer (Sigma) page 4). Our M26 media was modified from the media used to with phenylmethylsulfonylfluoride, sodium vanadate, and protease derive normal lung organoids from induced pluripotent stem inhibitor cocktail (Roche). Protein was quantified via Bradford cells (21). In terms of our success rates, 88% (57/65) of our assay (Bio-Rad), denatured in sample buffer (Bio-Rad), and loaded dissociated NSCLC tissue successfully initiated organoid cultures, for SDS-PAGE. Proteins were transferred onto nitrocellulose which is defined as organoid formation upon plating in passage membranes (Bio-Rad) and blocked in 5% skim milk for 1 hour zero (Supplementary Table S2). Seventy-two percent (47/65) of the and probed overnight with appropriate primary antibodies. models exhibited a range of short-term growth (passage 1–9, 1– The membrane was probed with secondary anti-rabbit/mouse IgG, 3 months), providing opportunities for most tissues to be used in HRP-linked antibodies (#7074 and #7076; Cell Signaling Technol- acute studies (Fig. 1B; Supplementary Tables S1 and S2). In ogy) for 1 hour prior to imaging. ECL reagent (GE Healthcare) was addition, 15% (10/65) of the models achieved long-term growth used to detect proteins of interest. The primary antibodies used in (Fig. 1B; Supplementary Tables S1 and S2).

AACRJournals.org Clin Cancer Res; 2020 OF3

Shi et al.

AB Patient primary Lung adenocarcinoma Lung squamous cell lung tumor carcinoma LPTO131 LPTO77 LPTO130 LPTO76 LPTO75 LPTO129 LPTO74 PDX LPTO128 t

n LPTO73 LPTO127 e LPTO72 Implant LPTO126 ti LPTO71

LPTO94 Pa LPTO65 tumor LPTO92 LPTO59 LPTO91 LPTO55 Patient LPTO90 LPTO40 LPTO39 LPTO86 LPTO20 LPTO85 LPTO7 Dissociate tumor, LPTO83 422 LPTO80 377 H-2K mouse LPTO57 369 LPTO54 365 Dissociate tumor cell depletion 4056 356 4009 353 Organoid model 426 322 321 402 296 Organoids Organoids 344 295 325 274 314 271 PDX

309 PDX 268 256 267 148 252 137 225 200 110 188 83 162 152 0 20406080100 149 Days 86 Short-term (<3 months, P1−9) and Short-term Long-term 020406080100 Long-term (>3 months, >P10) Days Organoid models CD100

LUAD primary LUAD PDX LUSC PDX 80 LPTO126 PDXO137 PDXO321 60 PDXO344 40

Wells, n LPTO126 Tumor 20 LPTO131 PDXO137 0 0255075 Days Bright-field E PDXO344 EGFR wt LPTO126 EGFR wt PDXO137 EGFR mut LPTO131 EGFR wt 100 Organoid 80

60 TP63

% Viability 40

20 TTF-1 −3 −2 −10 1 2 log Erlotinib (μmol/L)

Figure 1. Establishment of NSCLC-derived organoids and characterization of short-term organoid cultures. A, Schematic of NSCLC organoid development from surgically resected tumors or PDX. Models propagated below 10 passages and under 3 months were considered to be short-term cultures, whereas models propagated beyond 10 passages and over 3 months were considered to be long-term cultures. B, Maximum number of days in culture of all models attempted visualized on a swimmer's plot. Models contaminated with mouse or normal cells or derived from metastasis were excluded. C, Selected short-term NSCLC organoid histology and IHC staining. Note that LPTO126 patient tumor was both TTF-1 and TP63 negative, whereas PDXO137 PDX was TTF-1 positive and TP63 negative. The organoids reﬂected the TTF-1 and TP63 staining of their parental tumors. Scale bar, 100 mm. D, Organoid cell growth of short-term organoid cultures. Each point on the graph represents a passage. Growth was calculated by plotting the time to passaging and the cumulative sum of the number of wells plated. E, Erlotinib testing in short-term organoid models.

OF4 Clin Cancer Res; 2020 CLINICAL CANCER RESEARCH

Non–Small Cell Lung Cancer Organoids

Tumor purity of short-term organoid cultures inhibitor erlotinib in three models with wild-type EGFR and one A recurrent issue highlighted by previous work is the outgrowth of model with EGFR exon 19 deletion. The EGFR exon 19 deleted normal epithelial cells of organoid cultures derived from primary organoid model PDXO137 was the most sensitive to erlotinib, whereas patient tumor (24, 26). Consistent with previous reports of NSCLC the EGFR wild-type models were less sensitive (Fig. 1E). The parental organoids, we observed the outgrowth of normal epithelial cells in 58% PDX of PDXO137 has also been previously shown to respond to (7/12) of our short-term organoid cultures derived from patient tumor erlotinib (27), demonstrating that organoid drug responses reflect and PDX (Supplementary Tables S1 and S2). Because surgically those of its parental tumor. Therefore, we demonstrated as a proof-of- resected lung tumors or biopsies may contain entrapped normal lung principle that short-term organoids contain sufficient cell numbers for airway/alveolar epithelial cells, we speculate that these normal orga- drug testing and may be used as preclinical models for biomarker noids derived from patient tissue likely arose from this region. In validation. contrast, normal organoids arising from subcutaneously implanted PDX could have arisen from entrapped murine breast/sweat gland Characterization of long-term NSCLC organoid cultures tissues at the implantation site. To determine tumor purity in PDOs, (growth, purity, histologic/lineage marker) we performed cytologic evaluation by H&E and IHC for the lung Fifteen percent of NSCLC organoid models became long-term markers TTF-1 and TP63 of the cultured cells and the original patient cultures, as defined by continuous cell growth that maintained the tumor. We observed that the normal-like organoid models do not same split ratios in late passages (beyond 10 passages, over 3 months) reflect the IHC results of their parental tumor (Supplementary and retained a high percentage of tumor cells (Fig. 1B; Supplementary Table S1; Supplementary Fig. S1). For example, the LPTO124 patient Figs. S1 and S3). These cultures could be propagated beyond 10 tumor is an adenocarcinoma that stains negative for both TTF-1 and passages for over 1 year in culture with a splitting ratio of at least TP63, but the matched organoid stains positive for TP63 and negative 1:3, and without a decline in proliferation as the passage number for TTF-1. Because TP63 is a marker for lung basal cells, we speculate increased (Fig. 2A). They were also recoverable from >1 year of that the organoids derived from the LPTO124 patient tumor reflects a cryopreservation and could be expanded in culture after thawing. cell population growing from normal cells of basal cell origin. To assess Using the same method to assess tumor purity as described for the percentage of tumor cells versus mouse cells in the PDX-derived short-term, in long-term cultures, none were contaminated with þ þ organoids (XDO), EpCAM (human epithelial cells) and H2K normal or nonhuman cells. PDOs consisted of over 85% of tumor þ (mouse cells) cell populations were characterized by flow cytometry cells and the majority of the XDOs contained over 65% of EpCAM analysis. Overall, for short-term cultures, 75% (3/4) of the evaluable cells (Fig. 2B; Supplementary Table S2), with <8% of H2K-positive PDO models were contaminated with this normal cell population, cells in all of the long-term organoid models. whereas 50% (4/8) of the evaluable XDOs were contaminated with Long-term established NSCLC organoids also retained the histo- mouse cells (<2% human EpCAM, >60% H2K; Supplementary Tables logic features of their parental tumors. LUAD tumors can be classified S1 and S2). PDO and XDO models that were deemed to not be largely into multiple histologic subtypes, which include acinar, lepidic, solid, contaminated with normal cell populations exhibited 75%–97% and papillary, and mixed histology. Four patient with LUAD and PDX 50%–90% tumor cell populations, respectively (Supplementary tumors (LPTO54 tumor, LPTO85 tumor, PDXO426 PDX, and Table S1). Finally, we were not able to detect the presence of fibroblasts PDXO4056 PDX) collectively represented three histologic subtypes and immune cells in the short-term organoid cultures by histologic of LUAD: acinar predominant, mucinous, and solid predominant assessment. (poorly differentiated; Fig. 2C; Supplementary Fig. S2A). These histologic subtype patterns were reflected in the matched organoids of the Recapitulation of histologic and cell lineage features of parental tumor samples. In addition, expression of LUAD lineage markers such tumors by short-term NSCLC organoid cultures as TTF-1 was preserved. LPTO54 tumor and PDXO4056 PDX, as well To assess the quality of our short-term NSCLC organoids for as their matched organoids were positive for TTF-1, whereas LPTO85 downstream applications, we assessed the organoid models by cytol- tumor and PDXO426 PDX, along with their respective organoid ogy/histology. Note that histologic and tumor purity assessment were models, were TTF-1 negative (Fig. 2C; Supplementary Fig. S2A). performed in the short-term cultured LPTO126 and PDXO137 orga- All three LUSC PDX models exhibited features of moderately noids before they were later established as long-term models. Histo- differentiated LUSC. Likewise, our long-term LUSC PDXs and logic evaluation of the short-term models revealed LUAD and LUSC matched organoids were moderately differentiated and nonkeratiniz- representing various histologic subtypes such as mucinous (LPTO126) ing squamous cell carcinomas (PDXO274 and PDXO377), except for and acinar (PDXO137) morphology in LUAD, and moderate differ- PDXO149 which was a keratinizing LUSC (Fig. 2D; Supplementary entiation (PDXO321) in LUSC (Fig. 1C). The organoids also reflected Fig. S3A and S3B). The LUSC organoids preserved the histology of the TTF-1 and TP63 staining pattern of their parental tumors, their matched PDX models. LUSC organoids were positive for TP63 suggesting that they recapitulate the histology of their parental tumors and CK5, and negative for TTF-1 and CK7, which is characteristic of (Fig. 1C). LUSC (Supplementary Fig. S3C). To demonstrate the utility of short-term organoid models for drug To determine whether the organoid culture conditions preserved testing, we first determined whether there was sufficient number and the tumorigenic properties of the cancer cells, organoids derived growth of cells for these experiments. The four short-term models used from patient tissue and PDX were implanted into immunocompro- in our drug test were propagable in the first few passages and contained mised NOD/SCID mice. The NSCLC organoid models formed tumor enough cells for plating (Fig. 1D). We evaluated the efficacy of xenografts that histologically recapitulated their parental tumors clinically approved EGFR-targeted therapy in NSCLC in four short- (Supplementary Fig. S4A–S4D). Among LUAD models, the term organoid models. Although some of the organoids later on PDXO4056 xenograft formed a solid LUAD positive for TTF-1 while became long-term models, the drug test was performed in early the LPTO85 xenograft exhibited features of a mucinous LUAD passages (P1) of those organoids for the purpose of assessing the negative for TTF-1, which is typical of mucinous LUAD (46). Among ability of short-term models for drug testing. We evaluated the EGFR LUSC models, both PDXO274 and PDXO149 organoid xenografts

AACRJournals.org Clin Cancer Res; 2020 OF5

Shi et al.

AB Continuous organoid growth Histology through passaging (%Tumor Flow analysis cells) (%EpCAM) 1,000 100 LPTO54 800 LPTO85 PDXO4056 nt 50 600 PDXO426 90 85 92 68 90 96 87 PDXO149 P erce 400 Wells, n PDXO377 200 PDXO274 0

4056 0 LPTO54LPTO85 050100 PDXO426PDXO274PDXO149PDXO377 PDXO Days Organoid model C D Lung adenocarcinoma Lung squamous cell carcinoma LPTO54 PDXO426 PDXO4056 PDXO274 PDXO149 Tumor PDX tumor Patient/PDX Bright-field Bright-field Organoid Organoid TTF-1 CK5/6 p63

Figure 2. Histologic and growth characterization of long-term organoid cultures. A, Growth curves of seven long-term established organoid models. Each point on the graph represents a passage. Growth was calculated by plotting the time to passaging and the cumulative sum of the number of wells plated. B, Tumor cell purity in seven long-term established organoid models assessed by histologic examination or ﬂow cytometry analysis. H&E and IHC of representative LUAD (C) and LUSC (D) models demonstrating histologic recapitulation of the patient tumor or PDX to the matched organoid. Scale bar, 200 mm.

formed LUSC expressing the LUSC markers CK5 and TP63 (Supple- their parental tumors by WES. These samples included three patient– mentary Fig. S4A–S4D). Overall, our data indicate that even over long- organoid pair (LPTO54, LPTO85, LPTO126), one PDX–organoid pair term, our organoid culture conditions allow the cancer cells to retain (PDXO4056), and five patient–PDX–organoid groupings (PDXO426, key biological properties observed in the patient tumors, including PDXO344, PDXO137, PDXO149, PDXO274). The spectrum of muta- histologic differentiation and tumorigenicity. tions was highly concordant between the organoid and their matched patient tumor and/or PDX tissue (Fig. 3A; Supplementary Table S3). NSCLC organoids preserve the mutation and copy number Furthermore, the mutation burden in the five long-term established landscape of their parental tumors organoids was also similar to that of their parental patient/PDX tumors To evaluate the genome profile concordance of organoid cultures to (Fig. 3A), indicating that the culture conditions do not destabilize the their source, we compared the spectrum of somatic mutation and copy cancer genomes. The WES data further revealed that our organoid number aberrations between nine long-term organoid cultures and models harbored common mutations that were previously identified in

OF6 Clin Cancer Res; 2020 CLINICAL CANCER RESEARCH

Non–Small Cell Lung Cancer Organoids

A Color Key Fraction of concordant mutations 0 0.2 0.6 1 Value P 54 248 O 258 X 4056 315 O 316 P 248 426 X 279 O 249 P 126 336 O 362 P 309 85 O 460 P 183 137 X 314 O 177 P 254 344 X 739 O 802 P 1040 274 X 2679 O 3188 P 210 149 X 183 O 200 POXO PXOPOPPOXPO O XOPOX XP 4056 426 126 85 137 344 274 54 149 0 1,000 2,000 3,000 4,000 Number of somatic mutations Copy number concordance B C P 1 54 O Gene expression correlation among X 0.8 patient tumor, PDX and organoids 4056 O 1 P 0.6 0.8 426 X Patient O 0.6 0.4 P 126 0.4 O P 0.2 0.2 85 0.66 O PDX 0 P 0 [0.57,0.75] 137 X −0.2 O −0.2 −0.4 P −0.6 344 X 0.59 0.8 −0.4 Organoid O [0.54,0.64] [0.75,0.85] −0.8 P −1 274 X −0.6 O P −0.8 149 X O −1 PPXOPOOOXO PPOXPOXOPOX XP 54 4056 426 126 85 137 344 274 149

Figure 3. Mutation, copy number, and transcriptomic landscape of organoids and matched patient tumor/PDX. A, Mutational concordance and mutation burden between patient tumor and respective PDX and organoids. Heatmap represents the fraction of concordant mutations between corresponding samples. B, Copy number concordance heatmap on the global gene level. Pearson correlation of gene copy number was computed per sample. A panel of normal tissues was used for copy number calling for samples without matched normal tissue (model 274, 54, 4056, 426, 126, 85, 344, 137). C, Gene expression Pearson correlation heatmap (95% conﬁdence interval) showing gene expression clustering using 893 genes differentially expressed between LUAD and LUSC PDX models in 9 patient/PDX-organoid models (total of 23 samples). P, patient; X, patient-derived xenograft; O, organoid.

AACRJournals.org Clin Cancer Res; 2020 OF7

Shi et al.

independent NSCLC patient profiling studies (47–49). The affected type cell lines (50, 51). To determine whether the KRAS mutation genes included TP53, DDR2, KRAS, KEAP1, CUL3, NOTCH, etc. and amplification in the PDXO426 organoid confers sensitivity to (Supplementary Table S3). the MEK inhibitor trametinib, we compared its response with the Copy number variation (CNV) analysis also supported the tumor drug to that of three other organoid models without KRAS alterations. origin of the organoids and indicated that CNV profiles of the parental Consistent with previous studies, we found that the KRAS mutant tumors were largely preserved during organoid culture (Fig. 3B). PDXO426 was much more sensitive to the MEK inhibitor trametinib < m KRAS Notably, major chromosomal copy number changes associated with (IC50 0.05 mol/L) than three other organoids with wild-type LUAD and LUSC including, chr.1q and 3q amplifications, respectively, (IC50 > 0.5 mmol/L; Fig. 4A). Similar results were also generated were detected in our patient/PDX/organoid cohorts (Supplementary with the MEK inhibitor selumetinib (Fig. 4B). To confirm that the Table S4). LPTO85 and LPTO344 patient tumors appeared to be KRAS alterations in PDXO426 act specifically to confer sensitivity to highly correlated with one another by CNV. We confirmed the distinct targeted therapy, we also examined the responses of these four identities of these samples by STR profiling. However, more detailed organoid models to the EGFR inhibitor afatinib (Fig. 4C). As expected, analysis of their genomes revealed that they were both close to copy none of the four organoid models responded to afatinib relative to the number neutral, suggesting that this similarity largely accounted for HCC827 cell line, which has an EGFR mutation that sensitizes cells to their close correlation by CNV analysis. At the gene level, we also EGFR inhibitors. detected amplification in KRAS and deletion of CDKN2A in To determine whether the specific MEK inhibitor sensitivity dis- PDXO426, which are frequent occurrences in LUAD (47), as well as played by the PDXO426 organoid model reflects a biological property FGFR1 amplification in PDXO274, which is commonly observed in of the PDX tumor from which it was derived, we evaluated trametinib LUSC (48). Overall, the cancer mutations and CNV detected in our sensitivity in the parental PDXO426 PDX model. PDXO426 study support the tumor origin of the organoids and indicate that they PDX exhibited trametinib sensitivity, whereas the KRAS wild-type retain the genomic aberrations and potentially other key cancer PDXO274 PDX was resistant, supporting ex vivo organoid drug properties of their parental tumors. responses being reflective of in vivo responses of nonculture adapted tumor cells (Fig. 4D). Overall, these results support organoids being Gene expression profiles are similar between NSCLC organoids clinically relevant surrogates for patient tumors for drug testing. and parental tumors To determine whether gene expression profiles are preserved in the Combination therapy in NSCLC organoids organoids, we used RNA-seq to analyze gene expression of nine We next explored whether NSCLC organoids can also be used as matched organoid–patient and/or PDX tumor pairs described in the discovery tools for novel biomarker and combination therapy genomic analysis. Because of the confounding situation of human approaches. CNV analysis revealed chromosome 8p amplification stromal cells uniquely contributing to gene expression in the patient in the patient, PDX, and organoid model of PDXO274. FGFR1 samples, we sought to identify a gene set that reduced the number of amplification in this region is a common occurrence in LUSC, stromal-specific genes and enriched for genes expressed in tumor which occurs in 20% of LUSC cases (48). However, FGFR1 ampli- epithelial cells. To identify such a gene set, we used gene expression fication by itself is not a good biomarker for FGFR inhibitor profiles for primary patient LUAD and LUSC growing as PDXs that monotherapy in LUSC, as only 7%–11% of preselected patients were obtained used human-specific microarray chips (4). From these demonstrated durable response in clinical trials (52, 53). Thus, we gene expression profiles, we obtained a list of 893 genes that are utilized PDXO274 to model potential combination therapies in differentially expressed between LUAD and LUSC at a level of 2-fold or FGFR1-amplified LUSC. more. Using this list, we determined that the overall gene expression FGFR1 mRNA and protein quantification by RT-qPCR and West- correlation between patient tumor and organoids is 0.59, whereas PDX ern blot analysis revealed that PDXO274 exhibited more than a 10-fold and organoids is 0.8 (Fig. 3C). Furthermore, gene expression between increase in FGFR1 mRNA expression and higher phospho-FGFR1 five of six XODs and one of three PODs was more highly correlated (pFGFR1) and total FGFR1 protein expression relative to PDXO149 with their matched tumor tissue from which the organoid was derived, (FGFR1 wild-type; Fig. 5A and B). These results indicated that in the than with any other organoid model (Supplementary Table S5). In PDXO274 organoid model, FGFR1 amplification correlated with addition, gene expression correlations were also calculated with 1,492 increased FGFR1 mRNA levels, protein expression, and pathway differentially expressed genes (a cutoff of 2-fold change) between activation. However, reflective of the low response rates to FGFR patient with LUAD and LUSC samples from the TCGA. Using this inhibitors in patients, in vitro drug testing of the FGFR inhibitor gene set, the correlation coefficient between patient organoids is 0.66, BGJ398 revealed that PDXO274 was largely insensitive to FGFR whereas PDX organoids is 0.85 (Supplementary Fig. S5), similar to inhibition (Fig. 5C). On the basis of previous cell line studies showing what we observed using the tumor epithelial-enriched gene list. efficacy of the combination of MEK and PI3K inhibitors with FGFR Overall, the molecular data indicate that our in vitro growth conditions inhibitors in FGFR-aberrant cancers (54, 55), we tested trametinib and allow organoid tumor cells to largely maintain key molecular prop- the PI3K inhibitor BKM120 with BGJ398 in our FGFR1-amplified erties of their parental tumors. organoid model. Strong synergy (combination index < 0.5) was observed in the BGJ398þtrametinib combination, whereas weaker Utility of long-term NSCLC organoids for drug testing synergy (combination index >0.5) was observed in the To explore the utility of our long-term NSCLC organoids for drug BGJ398þBKM120 combination (Fig. 5D). In addition, although testing, we first surveyed genomic data of the five well-characterized single-agent BGJ398 inhibited pFGFR and pAkt, and single-agent long-term models for potential sensitizing biomarker alterations. The trametinib inhibited only pErk, targeted inhibition of all three phos- KRAS G13C mutation and amplification was detected in the patient, phoproteins was achieved with the combination of the two compounds PDX, and organoid of PDXO426. Preclinical studies with KRAS (Fig. 5E). The efficacy of the trametinib and BGJ398 combination was mutant cell lines have suggested that such mutant tumor cells may further verified in vivo, in the parental PDXO274 PDX (Fig. 5F), which be more sensitive to MEK inhibitors, as compared with KRAS wild- further supports our earlier contention that organoid models can

OF8 Clin Cancer Res; 2020 CLINICAL CANCER RESEARCH

Non–Small Cell Lung Cancer Organoids

PDXO274 PDXO4056 PDXO274 PDXO4056 AB LPTO54 PDXO426 LPTO54 PDXO426

100 100 KRAS wt KRAS wt KRAS wt 50 50

KRAS wt % Survival % Survival

KRAS mut KRAS mut 0 0 -2 -1 0 1 -2 -1 0 1 log[Trametinib] μmol/L log[Selumetinib] μmol/L

C PDXO274 PDXO4056 LPTO54 PDXO426 HCC827

100 EGFR wt EGFR wt 50 EGFR wt % Survival

EGFR mut 0 -4 -2 0 log[Afatinib], μmol/L D PDXO426 PDX PDXO274 PDX 1,200 1,000 Vehicle Vehicle

) 1,000 Trametinib Trametinib 3 800 800 * 600 Treat 600 Treat 400 400 200 Tumor volume (mm 200 **

0 0 0 102030405060010203040 Days Days

Figure 4. Drug testing in long-term NSCLC organoids. Trametinib (A), selumetinib (B), and afatinib (C) drug testing in three LUAD models and one LUSC organoid model performed in technical and biological triplicates. HCC827 cell line was used as an EGFR-positive control. Error bars were determined as the SEM. Final drug curves were calculated as an average of three independent experiments. D, In vivo trametinib (1 mg/kg) sensitivity curves at experimental endpoint in PDXO426 and PDXO274 PDX. Error bars were determined as the SEM. N ¼ 5 mice (PDXO426 PDX) and N ¼ 6 mice (PDXO274 PDX) were used for each arm (, P > 0.05; , P < 0.05). retain the targeted therapy sensitivity of its source tumor tissue. Our Discussion data thus support combining FGFR and MEK inhibitors in FGFR1- Organoid methodology has gained widespread popularity in the ampliﬁed LUSC. Collectively, our drug studies generally support the past few years for its utility in disease modeling and drug addition of organoid and PDX models to a validation/discovery screening (56–59). We aimed to establish a protocol of culturing pipeline from cell lines to the patient, which may vastly improve the NSCLC organoids from patient tumors and PDXs, with the eventual clinical response rate.

AACRJournals.org Clin Cancer Res; 2020 OF9

Shi et al.

A PDX Organoid B P 15 15 = 0.01 Organoid PDXO 149 274 ion ion pFGFR 10 10 P = 0.004 (Y653/654) FGFR1 5 5

β-Actin FGFR1 Relative expre ss FGFR1 Relative expre ss 0 0 PDXO149 PDXO274 PDXO149 PDXO274

CDPDXO274 PDXO274 150 PDXO274 BGJ398 Trametinib BGJ398 BKM120 PDXO149 Combination Combination 100 100 100

50 50 50 % Survival 0 -2 -1 0 1 0 0 log[BGJ398], μmol/L -2 -1 0 1 -2 -1 0 1 log[Drug], μmol/L log[Drug], μmol/L E PDXO274 1.5 ED50 BGJ398 TrametinibCombo

x ED75 0 1 3 5 1 3 5 1 3 5 μmol/L

pFGFR i nde 1.0 (Y653/654) ti on Synergy FGFR1 0.5 pErk1/2

(T202/Y204) Combi na Strong synergy Erk1/2 0.0 pAkt (S473) BGJ398 BGJ398 Akt + + Trametinib BKM120 β-Actin

F 1,000 Vehicle ) 3 BGJ398 800 Trametinib 600 Combination Treat 400 ** 200 mor vol umeTu (mm 0 0 1020304050 Time (days)

Figure 5. Combination of FGFR1 and MEK inhibitors in LUSC organoid models. A, qRT-PCR of FGFR1 in PDXO274 PDX and organoid. FGFR1 expression was normalized to PDXO149 PDX. Error bars were determined as the SEM. B, FGFR1 protein expression in PDXO274 validated by Western blot analysis. C, In vitro screen of BGJ398 in PDXO274 and PDXO149 performed in technical and biological triplicates. Error bars were determined as the SEM. Final drug curves were calculated as an average of

three independent experiments. D, Combination drug screen of BGJ398 with trametinib and BKM120 performed in technical and biological triplicates. ED50 is the drug synergy at 50% inhibition of cell viability, and ED75 is the drug synergy at 75% inhibition of cell viability. Error bars were determined as the SEM. Final drug curves were calculated as an average of three independent experiments. Combination indices were determined in CompuSyn software. E, Targeted inhibition of FGFR1 downstream proteins with single agents and combination treatment at 1, 3, and 5 mmol/L for 24 hours. F, In vivo conﬁrmation of the trametinib (1 mg/kg) and BGJ398 (25 mg/kg) combination in PDXO274 PDX. N ¼ 4–6 mice were used per arm. Error bars were determined as the SEM. , P < 0.05 comparing between single agent and combination therapy using Student t test at the 38-day timepoint.

OF10 Clin Cancer Res; 2020 CLINICAL CANCER RESEARCH

Non–Small Cell Lung Cancer Organoids

goal of establishing an improved platform for drug testing and of gene expression profiles between organoids and their parental biomarker discovery in NSCLC. Recent reports described two distinct tumors (15). We showed that majority of our organoid models methods for generating NSCLC organoid cultures, which harbored a (>10 passages) exhibited a strong correlation in gene expression with mixture of normal and tumor cell populations (24, 26). In one method, their parental tumors. However, it did appear that the gene expression the authors found it necessary to treat the organoid cultures with the correlations were better for organoids derived from PDXs than patient MDM2 inhibitor nutlin-3a to enrich for tumor cells harboring TP53 tumors. This may reflect the patient tumors having the most hetero- mutation, due to the large amount of contamination by nontumor geneity among cell populations, with the PDX and organoid models cells (26). However, long-term exposure to such chemicals may have potentially selecting for fewer clonal populations. unexpected consequences on tumor cell biology and although TP53 One of the attractive applications of our short-term cultures is mutations are common in NSCLC, they do not occur in all that there is usually enough cell material for drug testing, which we cases (47, 48, 60), precluding this method for such tumor samples. have demonstrated for some targeted therapeutics. This may be In addition, there should be continued efforts to improve media to particularly useful not only for biomarker validation studies, better enrich for tumor cells from a greater number of samples. Future but also for quick assessment of which therapies would be suitable development of organoid models and media formulations would for patient use when they fail on first or second-line therapies. benefit from a standardized set of parameters to calculate “success Recent reports have demonstrated the utility of PDXs for drug rates” of model establishment. Currently, different studies use different screening and personalized medicine (64). However, PDXs are criteria for deeming their cultures to be initiated or established. The limited by slow growth rate and are economically challenging to definition of established models, for example, is defined by continuous maintain. Thus, our protocol for establishing NSCLC organoids propagation for 6 months in one study (61) whereas others use from patient tumor and PDX could potentially provide additional 1 month (25) or ambiguous cutoffs (26) to define the longevity of models of NSCLC that are biologically relevant for future drug their organoid cultures. Using our methods, which includes histologic screening studies. Indeed, we find that we are usually able to and flow cytometric characterization of cultures, we were able to obtain enough cells even in short-term culture for drug testing, achieve a collective overall establishment rate of 88% for both which would be particularly useful for quick assessment of short-term and long-term NSCLC organoids. One question that would alternative therapies when patients fail on first or second-line be interesting to address in future studies is what are the differences treatments. With short-term cultures, we verified a previous between tumor samples that determine whether tumor cells can adapt suggestion that MEK could be a potential clinically relevant target at all versus short-term versus long-term to organoid culture. Reasons in some KRAS mutant LUADs (50, 51). Using our long-term for these differences are currently unknown, although mitotic insta- organoid cultures, we also found evidence to support combining bility of cancer cells, lack of physical environmental support and FGFR and MEK inhibitors in FGFR1-amplified LUSC, the basis essential media components, and the quiescent nature of lung stem for which was also initially suggested from cell line work (54). cells have been speculated to account for some failures (22). We Importantly, we found that the organoid drug response is similar attempted to compare tumor features (histologic subtype, size) and to that of the matched PDX, which was not adapted to ex vivo mutation status between long-term and short-term organoid models culture. Thus, organoids developed with our protocol appear to (Supplementary Tables S6 and S7). However, these association studies be good surrogates for clinical tissue for drug screening and are currently limited by our small sample size, which may be overcome biological studies. as more organoid models become available. Interestingly, we did In conclusion, our study provides a methodology of developing observe that organoids were more easily formed from PDXs than short-term and long-term organoid cultures from NSCLC patient and primary patient tumors. This finding suggests that a prior selection PDX tumor tissue, and demonstrates the utility of the established pressure in the nonorthotopic in vivo PDX environment might enrich organoids for drug testing and biomarker validation. Our collection of for tumor populations that are more likely to survive in vitro. We also NSCLC organoids is also a novel addition to existing preclinical observed that it was more difficult to establish organoids from LUSC models of NSCLC that may be useful for identifying viable therapeutic patient tumors than LUAD tumors. These suggest that the ex vivo options for this disease. conditions are still not optimal for LUSC, and consequently, LUSC exhibits a marked preference for an in vivo environment, even non- Disclosure of Potential Conflicts of Interest orthotopic. Indeed, it has generally been more difficult to establish 2D G. Liu is a paid consultant for AstraZeneca, Takeda, Pfizer, Novartis, Roche, cell lines from LUSC, as compared with LUAD, whereas LUSC patient Bristol-Myers Squibb, Bayer, and EMD Serono, and reports receiving commercial tumors engraft much better as PDXs (60% vs. 25%, respectively; research grants from Takeda, AstraZeneca, and Boehringer Ingelheim. No potential conflicts of interest were disclosed by the other authors. refs. 4, 28). In addition to recapitulating the biology that drives histologic ’ appearance of their parental tumors growing in vivo, another attractive Authors Contributions feature of our organoid models is that they have not yet drifted on the Conception and design: R. Shi, N. Radulovich, N. Moghal, M.-S. Tsao Development of methodology: R. Shi, N. Radulovich, C. Ng, N. Liu, H. Notsuda, molecular level, as many cell lines have. Previous studies have shown fi S.N. Martins-Filho, G. Liu that organoids from other cancers can retain the molecular pro les of Acquisition of data (provided animals, acquired and managed patients, provided their parental tumors, even with subsequent passaging (15, 61–63). We facilities, etc.): R. Shi, N. Radulovich, C. Ng, H. Notsuda, S.N. Martins-Filho, examined the mutation and CNV profiles of our NSCLC organoids J.C. Rosen, G. Liu, M.-S. Tsao and identified major somatic alterations in lung cancer that were Analysis and interpretation of data (e.g., statistical analysis, biostatistics, preserved in the parental tumors. We also observed that organoids of computational analysis): R. Shi, N. Radulovich, C. Ng, H. Notsuda, M.Cabanero,S.N.Martins-Filho,V.Raghavan,Q.Li,A.S.Mer,B.Haibe- late passage (even after more than 10 passages) still retained the Kains, G. Liu, M.-S. Tsao molecular features of their parental tumors. This suggests that muta- Writing, review, and/or revision of the manuscript: R. Shi, N. Radulovich, tion and CNV profiles are largely stable in organoid cultures. In M. Cabanero, S.N. Martins-Filho, V. Raghavan, Q. Li, N.-A. Pham, B. Haibe- addition, there have been reports of both concordance and discordance Kains, G. Liu, N. Moghal, M.-S. Tsao

AACRJournals.org Clin Cancer Res; 2020 OF11

Shi et al.

Administrative, technical, or material support (i.e., reporting or organizing data, Terry Fox Research Institute and Princess Margaret Cancer Foundation. G. Liu is the constructing databases): R. Shi, N. Radulovich, C. Ng, H. Notsuda, M. Li, Alan B. Brown Chair in Molecular Genomics. M.-S. Tsao is the M. Qasim Choksi Y.-H. Wang, L. Tamblyn, N.-A. Pham, G. Liu, M.-S. Tsao Chair in Lung Cancer Translational Research. The authors thank Jing Xu, Wendy So, Study supervision: N. Radulovich, M.-S. Tsao and Jian Zhou for all IHC staining. They acknowledge the Princess Margaret Biobank for providing patient tissue samples. They thank the Princess Margaret Genomics Acknowledgments Centre (PMGC), especially Julissa Tsao, for exome sequencing RNA sequencing. They This work was supported by Canadian Institute of Health Research (CIHR) thank Drs. Trevor Pugh and Vuk Stambolic for advice and guidance of the project. Foundation grant FDN-148395 (to M.-S. Tsao), Canadian Cancer Society Research Institute IMPACT grants 701595 (to M.-S. Tsao) and 703206 (to G. Liu), and the The costs of publication of this article were defrayed in part by the payment of page Princess Margaret Cancer Foundation (for PM Living Biobank core). R. Shi is funded charges. This article must therefore be hereby marked advertisement in accordance by a University of Toronto Ontario Student Opportunity Trust Fund (OSOTF) and with 18 U.S.C. Section 1734 solely to indicate this fact. Ontario Graduate Scholarship (OGS). M. Cabanero was supported by the Terry Fox Foundation Training Program in Molecular Pathology of Cancer at CIHR (STP Received May 5, 2019; revised September 19, 2019; accepted October 30, 2019; 53912). S.N. Martins-Filho was supported by a Training Program grant from the published ﬁrst November 6, 2019.

References 1. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, et al. Cancer 19. Aurora M, Spence JR. hPSC-derived lung and intestinal organoids as models of incidence and mortality worldwide: sources, methods and major patterns in human fetal tissue. Dev Biol 2016;420:230–8. GLOBOCAN 2012. Int J Cancer 2015;136:E359–86. 20. Nadkarni RR, Abed S, Draper JS. Organoids as a model system for studying 2. GazdarAF,GirardL,LockwoodWW,Lam WL, Minna JD. Lung cancer cell human lung development and disease. Biochem Biophys Res Commun 2016; lines as tools for biomedical discovery and research. J Natl Cancer Inst 2010; 473:675–82. 102:1310–21. 21. Dye BR, Hill DR, Ferguson MAH, Tsai Y, Nagy MS, Dyal R, et al. In vitro 3. Azzopardi D, Haswell LE, Foss-smith G, Hewitt K, Asquith N, Corke S, et al. generation of human pluripotent stem cell derived lung organoids. eLife Evaluation of an air-liquid interface cell culture model for studies on the 2015;4:1–25. inflammatory and cytotoxic responses to tobacco smoke aerosols. Toxicol In 22. Rock JR, Onaitis MW, Rawlins EL, Lu Y, Clark CP, Xue Y, et al. Basal cells as stem Vitro 2015;29:1720–8. cells of the mouse trachea and human airway epithelium. Proc Natl Acad Sci U S 4. Wang D, Pham NA, Tong J, Sakashita S, Allo G, Kim L, et al. Molecular A 2009;106:12771–5. heterogeneity of non-small cell lung carcinoma patient-derived xenografts 23. Wong AP, Chin S, Xia S, Garner J, Bear CE, Rossant J. Efficient generation of closely reflect their primary tumors. Int J Cancer 2017;140:662–73. functional CFTR-expressing airway epithelial cells from human pluripotent stem 5. Gillet JP, Varma S, Gottesman MM. The clinical relevance of cancer cell lines. cells. Nat Protol 2015;10:363–81. J Natl Cancer Inst 2013;105:452–8. 24. Dijkstra KK, Cattaneo CM, Weeber F, Chalabi M, van de Haar J, Fanchi LF, et al. 6. Jung J, Seol HS, Chang S. The generation and application of patient-derived Generation of tumor-reactive T cells by co-culture of peripheral blood lympho- xenograft model for cancer research. Cancer Res Treat 2018;50:1–10. cytes and tumor organoids. Cell 2018;174:1586–98. 7. Fatehullah A, Tan SH, Barker N. Organoids as an in vitro model of human 25. Neal JT, Li X, Zhu J, Giangarra V, Grzeskowiak CL, Ju J, et al. Organoid modeling development and disease. Nat Cell Biol Rev 2016;18:246–54. of the tumor microenvironment. Cell 2018;175:1972–88. 8. Sato T, Stange DE, Ferrante M, Viries RG, Van Es JH, Van den Brink S, et al. 26. Sachs N, Papaspyropoulos A, Ommen DDZ, Heo I, Bottinger L, Klay D, et al. Long-term expansion of epithelial organoids from human colon, adenoma, Long-term expanding human airway organoids for disease modeling. EMBO J adenocarcinoma, and barrett's epithelium. Gastroenterology 2011;141:1762–72. 2019;38:pii: e100300. 9. Boj SF, Hwang CI, Baker LA, Chio II, Engle DD, Corbo V, et al. Organoid models 27. Stewart EL, Mascaux C, Pham NA, Sakashita S, Skyes J, Kim L, et al. Clinical of human and mouse ductal pancreatic cancer. Cell 2015;160:324–38. utility of patient-derived xenografts to determine biomarkers of prognosis and 10. Chua CW, Shibata M, Lei M, Toivanen R, Barlow LJ, Bergen SK, et al. Single map resistance pathways in EGFR-mutant lung adenocarcinoma. J Clin Oncol luminal epithelial progenitors can generate prostate organoids in culture. 2015;33:2472–80. Nat Cell Biol 2014;16:951–61. 28. John T, Kohler D, Pintilie M, Yanagawa N, Pham N, Li M, et al. The ability to 11.DrostJ,KarthausWR,GaoD,DriehuisE,SawyersCL,ChenY,etal. form primary tumor xenografts is predictive of increased risk of disease Organoid culture systems for prostate epithelial and cancer tissue. Nat Protoc recurrence in early-stage non–small cell lung cancer. Clin Cancer Res 2011; 2016;11:347–58. 17:134–41. 12. Gao D, Vela I, Sboner A, Iaquinta PJ, Karthaus WR, Gopalan A, et al. Organoid 29. Conway T, Wazny J, Bromage A, Tymms M, Sooraj D, Willams ED, et al. cultures derived from patients with advanced prostate cancer. Cell 2014;159: Xenome—a tool for classifying reads from xenograft samples. Bioinformatics 176–87. 2012;28:172–8. 13. Huch M, Gehart H, van Boxtel R, Hamer K, Blokzijl F, Verstegen MM, et al. 30. Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler Long-term culture of genome-stable bipotent stem cells from adult human liver. transform. Bioinformatics 2009;25:1754–60. Cell 2015;160:299–312. 31. DePristo MA, Banks E, Poplin R, Garimella KV, Maguire JR, Hartl C, et al. A 14. Sachs N, de Ligt J, Kopper O, Gogola E, Bounova G, Weeber F, et al. A living framework for variation discovery and genotyping using next-generation DNA biobank of breast cancer organoids captures disease heterogeneity. Cell 2018; sequencing data. Nat Genet 2011;43:491–8. 172:373–86. 32. Cibulskis K, Lawrence MS, Carter SL, Sivachenko A, Jaffe D, Sougnez C, et al. 15. Broutier L, Mastrogiovanni G, Verstegen MM, Francies HE, Gavarro LM, Sensitive detection of somatic point mutations in impure and heterogeneous Bradshaw CR, et al. Human primary liver cancer-derived organoid cultures for cancer samples. Nat Biotechnol 2013;31:213–9. disease modeling and drug screening. Nat Med 2017;23:1424–35. 33. Koboldt DC, Zhang Q, Larson DE, Shen D, McLellan MD, Lin L, et al. VarScan 2: 16. Huang L, Holtzinger A, Jagan I, BeGora M, Lohse I, Ngai N, et al. Ductal somatic mutation and copy number alteration discovery in cancer by exome pancreatic cancer modeling and drug screening using human pluripotent stem sequencing. Genome Res 2012;22:568–76. cell– and patient-derived tumor organoids. Nat Med 2015;21:1364–71. 34. Sherry ST, Ward MH, Kholodov M, Baker J, Phan L, Smigielski EM, et al. dbSNP: 17. Pauli C, Hopkins BD, Prandi D, Shaw R, Fedrizzi T, Sboner A, et al. Personalized the NCBI database of genetic variation. Nucleic Acids Res 2001;29:308–11. in vitro and in vivo cancer models to guide precision medicine. Cancer Discov 35. Lek M, Karczewski KJ, Minikel EV, Samocha KE, Banks E, Fennell T, et al. 2017;5:462–77. Analysis of protein-coding genetic variation in 60,706 humans. Nature 2016;536: 18. Schutte M, Risch T, Abdavi-Azar N, Boehnke K, Schumacher D, Keil M, et al. 285–91. Molecular dissection of colorectal cancer in pre-clinical models identifies 36. Fu W, O’Connor TD, Jun G, Kang HM, Abecasis G, Lean SM, et al. Analysis of biomarkers predicting sensitivity to EGFR inhibitors. Nat Commun 2017;8: 6,515 exomes reveals the recent origin of most human protein-coding variants. 14262. Nature 2013;10:216–20.

OF12 Clin Cancer Res; 2020 CLINICAL CANCER RESEARCH

Non–Small Cell Lung Cancer Organoids

37. Wang K, Li M, Hakonarson H. ANNOVAR: functional annotation of genetic patients with advanced solid tumors harboring genetic alterations in fibroblast variants from high-throughput sequencing data. Nucleic Acids Res 2010;38: growth factor receptors: results of a global phase I, dose-escalation and dose- e164. expansion study. J Clin Oncol 2017;35:157–65. 38. McLauren W, Gil L, Hunt SE, Riat HS, Ritchie GR, Thormann A, et al. The 53. Chae YK, Ranganath K, Hammerman PS, Vaklavas C, Mohindra N, Kalyan ensembl variant effect predictor. Genome Biol 2016;17:122. A, et al. Inhibition of the fibroblast growth factor receptor (FGFR) pathway: 39. Gu Z, Eils R, Schlesner M. Complex heatmaps reveal patterns and correlations in the current landscape and barriers to clinical application. Oncotarget 2017;8: multidimensional genomic data. Bioinformatics 2016;32:2847–9. 16052–74. 40. Talevich E, Shain AH, Botton T, Bastian BC. CNVkit: genome-wide copy 54. Bockorny B, Rusan M, Chen W, Liao RG, Li Y, Piccioni F, et al. RAS-MAPK number detection and visualization from targeted DNA sequencing. reactivation facilitates acquired resistance in FGFR1-amplified lung cancer and PLoS Comput Biol 2016;12:e1004873. underlies a rationale for upfront FGFR-MEK blockade. Mol Cancer Ther 2018; 41. Olshen AB, Venkatraman ES, Lucito R, Wigler M. Circular binary segmen- 17:1526–39. tation for the analysis of array-based DNA copy number data. Biostatistics 55. Wang L, Sustic T, Leite de Oliveira R, Lieftink C, Halonen P, van de Ven M, et al. 2004;5:557–72. A functional genetic screen identifies the phosphoinositide 3-kinase pathway as a 42. Mermel CH, Schumacher SE, Hill B, Meyerson ML, Beroukhim R, Getz G. determinant of resistance to fibroblast growth factor receptor inhibitors in FGFR GISTIC2.0 facilitates sensitive and confident localization of the targets of focal mutant urothelial cell carcinoma. Eur Urol 2017;71:858–62. somatic copy-number alteration in human cancers. Genome Biol 2011;12:R41. 56. Watanabe M, Buth JE, Vishlaghi N, de la Torre-Ubieta L, Taxidis J, Khakh BS, 43. Patro R, Duggal G, Love MI, Irizarry RA, Kingsford C. Salmon provides fast and et al. Self-organized cerebral organoids with human-specific features predict bias-aware quantification of transcript expression. Nat Methods 2017;14:417–9. effective drugs to combat zika virus infection. Cell Rep 2017;2:517–32. 44. Johnson WE, Li C, Rabinovic A. Adjusting batch effects in microarray expression 57. Dekkers JF, Berkers G, Kruisselbrink E, Vonk A, de Jonge HR, Janssens HM, et al. data using empirical Bayes methods. Biostatistics 2007;8:118–27. Characterizing responses to CFTR-modulating drugs using rectal organoids 45. Chou TC, Martin N. CompuSyn for drug combinations: PC software and User's derived from subjects with cystic fibrosis. Sci Transl Med 2016;8:344ra84. guide: a computer program for quantitation of synergism and antagonism in 58. Tiriac H, Belleau P, Engle DD, Pienker D, Deschenes A, Somerville TDD, et al. fi fi drug combinations, and the determination of IC50 and ED50 values. Paramus, NJ: Organoid pro ling identi es common responders to chemotherapy in pancre- Combo-Syn; 2005. Available from: http://www.combosyn.com. atic cancer. Cancer Discov 2018;8:1112–29. 46. Zhang Y, Wang R, Li Y, Hu H, Zhang Y, Li H, et al. Negative thyroid transcription 59. Vlachogiannis G, Hedayat S, Vatsiou A, Jamin Y, Fernandez-Mateos J, Khan K, factor 1 expression defines an unfavorable subgroup of lung adenocarcinomas. et al. Patient-derived organoids model treatment response of metastatic gastro- J Thorac Oncol 2015;10:1444–50. intestinal cancers. Science 2018;359:920–6. 47. The Cancer Genome Atlas Research Network. Comprehensive molecular pro- 60. Hollstein M, Sidransky D, Vogelstein B, Harris CC. p53 mutations in human filing of lung adenocarcinoma. Nature 2014;511:543–50. cancers. Science 1991;253:49–53. 48. The Cancer Genome Atlas Research Network. Comprehensive genomic char- 61. Lee SH, Hu W, Matulay JT, Silva MV, Owczarek TB, Kim K, et al. Tumor acterization of squamous cell lung cancers. Nature 2012;489:519–25. evolution and drug response in patient- derived organoid models of bladder 49. Shtivelman E, Hensing T, Simon GR, Dennis PA, Otterson GA, Bueno R, et al. cancer. Cell 2018;173:515–28. Molecular pathways and therapeutic targets in lung cancer. Oncotarget 2014;5: 62. Weeber F, van de Wetering M, Hoogstraat M, Dijkstra KK, Krijgsman O, 1392–433. Kuilman T, et al. Preserved genetic diversity in organoids cultured from biopsies 50. Wee S, Jagani Z, Xiang KX, Loo A, Dorsch M, Yao YM, et al. PI3K pathway of human colorectal cancer metastases. Proc Natl Acad Sci U S A 2015;112: activation mediates resistance to MEK inhibitors in KRAS mutant cancers. 13308–11. Clin Cancer Res 2009;69:4286–93. 63. Gendoo DMA, Denroche RE, Zhang A, Radulovich N, Jang GH, Lemire M, et al. 51. Yeh JJ, Routh ED, Rubinas T, Peacock J, Martin TD, Shen XJ, et al. KRAS/BRAF Whole genomes define concordance of matched primary, xenograft, and orga- mutation status and ERK1/2 activation as biomarkers from MEK1/2 inhibitor noid models of pancreas cancer. PLoS Comput Biol 2019;15:e1006596. therapy in colorectal cancer. Mol Cancer Ther 2009;8:834–43. 64. Gao H, Korn JM, Ferretti S, Monahan JE, Wang Y, Singh M, et al. High- 52. Nogova L, Sequist LV, Perez Garcia JM, Andre F, Delord JP, Hidalgo M, et al. throughput screening using patient-derived tumor xenografts to predict clinical Evaluation of BGJ398, a fibroblast growth factor receptor 1-3 kinase inhibitor, in trial drug response. Nat Med 2015;21:1318–25.

AACRJournals.org Clin Cancer Res; 2020 OF13

Organoid Cultures as Preclinical Models of Non−Small Cell Lung Cancer

Ruoshi Shi, Nikolina Radulovich, Christine Ng, et al.

Clin Cancer Res Published OnlineFirst November 6, 2019.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-19-1376

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2019/11/06/1078-0432.CCR-19-1376.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/early/2020/01/08/1078-0432.CCR-19-1376. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.