Ex Vivo Explant Cultures of Non–Small Cell Lung Carcinoma Enable Evaluation of Primary Tumor Responses to Anticancer Therapy

Published OnlineFirst February 15, 2017; DOI: 10.1158/0008-5472.CAN-16-1121

Cancer Therapeutics, Targets, and Chemical Biology Research

Ex Vivo Explant Cultures of Non–Small Cell Lung Carcinoma Enable Evaluation of Primary Tumor Responses to Anticancer Therapy Ellie Karekla1, Wen-Jing Liao1, Barry Sharp2, John Pugh2, Helen Reid2, John Le Quesne1,3, David Moore1, Catrin Pritchard1, Marion MacFarlane3, and James Howard Pringle1

Abstract

To improve treatment outcomes in non–small cell lung cancer stage. In explant tissue, cisplatin-resistant tumors excluded plat- (NSCLC), preclinical models that can better predict individual inum ions from tumor areas in contrast to cisplatin-sensitive patient response to novel therapies are urgently needed. Using tumors. Intact TP53 did not predict cisplatin sensitivity, but a freshly resected tumor tissue, we describe an optimized ex vivo positive correlation was observed between cisplatin sensitivity explant culture model that enables concurrent evaluation of and TP53 mutation status (P ¼ 0.003). Treatment of NSCLC NSCLC response to therapy while maintaining the tumor micro- explants with the targeted agent TRAIL revealed differential sen- environment. We found that approximately 70% of primary sitivity with the majority of tumors resistant to single-agent or NSCLC specimens were amenable to explant culture with tissue cisplatin combination therapy. Overall, our results validated a integrity intact for up to 72 hours. Variations in cisplatin sensi- rapid, reproducible, and low-cost platform for assessing drug tivity were noted with approximately 50% of cases responding ex responses in patient tumors ex vivo, thereby enabling preclinical vivo. Notably, explant responses to cisplatin correlated signiﬁ- testing of novel drugs and helping stratify patients using bio- cantly with patient survival (P ¼ 0.006) irrespective of tumor marker evaluation. Cancer Res; 77(8); 2029–39. 2017 AACR.

Introduction The era of personalized medicine has heralded the development of targeted therapies for NSCLC, some of which rely on Non–small cell lung cancer (NSCLC) is a leading cause of preselection of cancers according to genetic mutation. For cancer death worldwide. Patients with stage I–III tumors are example, selective EGFR inhibitors gefitinib and erlotinib pro- surgically resected and given adjuvant chemotherapy or radio- vide clinical benefit over standard chemotherapy for NSCLC therapy. Patients with stage IV disease receive palliative chemo- tumors bearing EGFR mutations (3, 4), whereas the ALK inhib- therapy only unless they can be stratified for targeted therapy. itor crizotinib benefits ALK-mutated cases (5). A global industry Most patients receive combination chemotherapy based on clin- is centered on assessing additional mono- or combinatorial ical parameters of cisplatin or carboplatin with at least one other treatments in NSCLC clinical trials. Despite this momentum, drug such as vinorelbine, gemcitabine, or paclitaxel. Unfortunate- late-stage failures are a reality and there is less than 11% success ly, only approximately 5% of patients receiving adjuvant therapy in bringing a drug to market (6), attributable in part to non- show 5-year average survival benefit (1, 2). Therefore, more predictive preclinical drug platforms (7, 8). The incorporation accurate methods for predicting chemotherapeutic benefit are of patient-derived xenograft (PDX) mouse models (9, 10) into urgently required to improve clinical outcomes. preclinical studies has improved predictive accuracy somewhat (11, 12). However, PDX efficacy studies are expensive, requiring large numbers of mice. Furthermore, not all primary human 1Department of Cancer Studies, University of Leicester, Leicester, United King- tumors generate PDXs and, of those that do, serial propagation dom. 2Centre for Analytical Science, Department of Chemistry, Loughborough can select tumors that adapt to grow in an immunocompro- 3 University, Loughborough, Leicestershire, United Kingdom. MRC Toxicology mised environment. Unit, Leicester, United Kingdom. An alternative approach is to use 3-dimensional ex vivo cul- Note: Supplementary data for this article are available at Cancer Research ture of fresh human tumors. Methods for ex vivo culture of Online (http://cancerres.aacrjournals.org/). human tumors have been available for many years, and evi- C. Pritchard, M. MacFarlane, and J.H. Pringle contributed equally to this article. dence shows that they can reliably reflect tumor growth in vivo Corresponding Authors: Catrin Pritchard, Department of Cancer Studies, Uni- (13–19). Here, we have developed and perfected an ex vivo versity of Leicester, Lancaster Road, Leicester LE1 9HN, United Kingdom. Phone: culture method for NSCLC tumor samples that is both simple 4411-6229-7061; Fax: 4411-6229-7018; E-mail: [email protected]; and Marion Mac- and reproducible. We have optimized culture conditions and Farlane, MRC Toxicology Unit, Hodgkin Building, Lancaster Road, Leicester LE1 show that tumor and stroma are retained intact and are viable. 9HN, United Kingdom. E-mail: [email protected] As proof-of-concept, NSCLC explant response to the standard- doi: 10.1158/0008-5472.CAN-16-1121 of-care chemotherapy drug cisplatin was examined, as well as 2017 American Association for Cancer Research. response to the targeted agent TRAIL. We also illustrate how

www.aacrjournals.org 2029

Karekla et al.

explants can be used to inform mechanisms of drug action by Statistical analysis evaluating biomarkers of drug response. Together, our data Significance of proliferation/death indices was determined show that the explant platform can effectively predict patient by Wilcoxon matched pairs test and Jonckheere–Terpstra trend response to therapy and can be used for monitoring clinically test, respectively. Unpaired data were compared by the Mann– relevant biomarkers. Whitney U test or Kruskal–Wallis one-way ANOVA. Paired data were analyzed by the Page L test (Unistat Statistical Package, Materials and Methods version 5.0, Unistat), and interrelationships were investigated by Spearman rank correlation (SPSS, version 22, IBM). The Ex vivo explant culture optimal cutoff point to determine the relationship between Fresh NSCLC tumors were collected from consented patients explant response and patient survival was examined using a undergoing lung surgery (Ethical approval: LREC: 07/MRE08/ 07). plot of sensitivity against 1 specificity as a receiver operator Patients had no prior exposure to chemotherapy. Viable tumor characteristic (ROC) curve (SPSS). Survival was investigated by fi areas were identi ed by frozen tissue sectioning and hematoxylin Kaplan–Meier analysis (SPSS) of cell indices, which were com- and eosin (H&E) staining. Tissue was placed in Hank balanced salt pared by the log-rank Mantel–Haenszel (Peto) test, and by 3 solution and cut into fragments of 2 to 3 mm using 2 skin graft univariate and multivariate Cox regression (SPSS). P < 0.05 was blades on a dental wax surface. These were placed in fresh culture considered statistically significant. media (DMEM with 4.5 g/L glucose þ 0%–5% FCS and 1% pen/ strep); 9 fragments were randomly selected and placed on a 0.4-mm culture insert disc (Millipore) floated on 1.5 mL of media in a Results 6-well dish. Explants were incubated at 37 C and 5% CO2 for 16 to Histopathology of NSCLC tumors used for explants 20 hours. Discs were then transferred to new wells containing Table 1 provides a summary of patient demographics, tumor 1.5-mL fresh media, and drugs or carrier control were added to each type, and stage for all 45 samples utilized for this study. The well in a volume of 1.5 mL for 24 hours. Cisplatin (Sigma) was histologic types and stages were broadly consistent with the utilized over a dose range of 0 to 50 mmol/L (dissolved in known distribution of NSCLC cases in the United Kingdom dimethylformamide). TRAIL (20, 21) was utilized at 1 mg/mL, (26). A proportion of NSCLC tumors is known to be necrotic diluted in DMEM media from a stock of 1 mg/mL. After treatment, (27), and an important first step was to exclude such tumors from explants were washed with PBS and transferred to new wells analysis using H&E assessment. This led to the identification of containing 1 mL of 4% (w/v) paraformaldehyde for 20 hours. 13 tumors (29%) that were excluded from explant generation. Explants were transferred onto sponges, pre-soaked in 70% (v/v) Supplementary Fig. S2 indicates the histologic type (Supplemen- ethanol, and placed in histology cassettes. They were embedded tary Fig. S2A) and stage (Supplementary Fig. S2B) of viable and into paraffin blocks from which 4-mm sections were generated. nonviable tumors. The highest proportion of nonviable tumors was within the adenocarcinoma subtype [36% compared with Histologic analysis 25% of squamous cell carcinoma (SCC) cases]. However, there H&E staining of formalin-fixed, paraffin-embedded (FFPE) was no correlation between tumor stage and viability. material sections was generated by standard approaches and, for Thirty-two viable tumors were processed for explant culture. immunohistochemistry (IHC), sections were processed as Intrinsic levels of cell proliferation and cell death were first described (22). The Novolink Polymer Detection System Kit assessed in uncultured samples by Ki67 or cleaved PARP (cPARP) (Leica Microsystems) was used according to the manufacturer's instructions. Primary antibodies were: cleaved PARP [E51]: Abcam 1:6,000, Ki67 Clone MIB-1: DAKO 1:2,000, p53 DO1: Table 1. Summary of patients' characteristics and tumors used for this study gift from David Lane 1:1,000, cytokeratin clone MNF116: DAKO Characteristic Number (%) of patients/tumors 1:5,000. Antibodies were diluted in blocking solution made with Sex 3% (w/v) BSA, 0.1% (v/v) Triton X-100 (Fisons) in TBS. Staining Male 25 (55.6) was visualized under a LEICA DM 2500 microscope and photo- Female 20 (44.4) graphed with a LEICA DFC 420 camera. Age Median 70 Range 54–85 Quantitation of IHC staining Histology Images of the tumor explants were taken at 10 magnification Adenocarcinoma 22 (48.9) and merged using Adobe Photoshop CS5.1, generating a single SCC 20 (44.4) image of one explant. Tumor area was determined using ImageJ Large cell carcinoma 0 (0) analysis (23), excluding areas of necrosis and stroma. The labeling Atypical carcinoid 3 (6.7) Stage index was determined using ImmunoRatio (24), and a single IA 7 (15.6) value was obtained for all 9 explants derived from one treatment IB 8 (17.8) that was expressed as a percentage of the total tumor area. IIA 8 (17.8) IIB 9 (20) Laser ablation inductively coupled plasma mass spectrometry IIIA 11 (24.4) IIIB 1 (2.2) Sections of explants treated with cisplatin were subjected to IV 1 (2.2) Laser ablation inductively coupled plasma mass spectrometry NOTE: Tumor samples were collected from consented patients undergoing (LA-ICP-MS) to produce elemental maps showing the spatial lung surgery at Glenfield Hospital, Leicester. Clinical data, histology, and distribution of platinum in tissue sections (25). The method is stagewereprovidedbyofficial histopathologic reports submitted by con- described in Supplementary Fig. S1. sultant pathologists at University Hospitals of Leicester, Leicester, UK.

2030 Cancer Res; 77(8) April 15, 2017 Cancer Research

Assessing Tumor Drug Response with a Novel Ex Vivo Platform

A 100 100 100

80 80 P = 0.003 80

60 60 60 H 40 40 40 % Ki67 +ve M % cPARP +ve % Labeling index 20 20 20 L 0 0 0 cPARP +ve Ki67 +ve ADC SCC AC ADC SCC AC

LT103

LT98

LT104

Ki67 cPARP

Figure 1. Intrinsic levels of proliferation and apoptosis in the tumors used for explants. A, Proliferation was assessed by quantitating Ki67 IHC staining and cell death by quantitating cPARP staining. Left, percentage of intrinsic proliferation and cell death in tumors. A single dot represents a single tumor sample. For the Ki67 staining, the samples were grouped into 1 of 3 groups: H, high (>40%); M, medium (20%–40%); L, low (0%–20%). The samples all had low intrinsic levels of cPARP staining. Middle and right, percentage of Ki67 and percentage of cPARP staining for tumors with adenocarcinoma (ADC), SCC, and atypical carcinoid (AC) histologies. The SCC samples had signiﬁcantly higher levels of intrinsic proliferation compared with the adenocarcinoma samples whereas, the atypical carcinoid samples were indolent. Cell death levels were consistent across the histologies. B, Representative images of Ki67 (left) and cPARP (right) IHC staining of high (LT103; top), medium (LT98; middle), and low (LT104; bottom) proliferative tumors. IHC stains were counterstained with hematoxylin. LT103 (high proliferative tumor) and LT98 (medium proliferative tumor) represent SCC samples, whereas LT104 (low proliferative tumor) is an adenocarcinoma. Scale bars, 100 mm.

immunostaining (Fig. 1A and B). SCC tumors displayed signif- Explants were routinely allowed to recover for a period of icantly higher levels of proliferation than adenocarcinoma, 16 to 20 hours after their initial generation; viability was whereas atypical carcinoid tumors were essentially indolent (Fig. assessed over a time range of 24 to 72 hours after recovery for 1A). These observations are consistent with several previous 5 tumors. As shown in Fig. 2A, a trend of decreasing cell reports (28–30). With regard to cPARP staining, the majority of proliferation and increasing cell death with increasing time of samples showed less than 20% of staining, indicating low levels of culture was observed, suggesting ex vivo explant cultures are intrinsic cell death (Fig. 1A and B). more viable in short-term culture. Varying FCS concentration, from 0% to 5%, at 24 hours of culture after the initial 16 to Optimization of explant culture 20 hours of culture recovery showed no statistically significant Our approach for the NSCLC explant culture system was based difference in levels of proliferation or cell death (Fig. 2B). on previous experience with breast cancer samples (MacFarlane, The data from above suggest that the 24-hour time point gives unpublished observations; ref. 31; ). As a first step in implement- the greatest viability but that FCS concentration is not a signi- ing protocols for NSCLC, we first investigated the effects of culture ficant factor. Subsequent analyses of drug responses were there- time and FCS concentration. fore performed for 24 hours, using 1% as the standard FCS

www.aacrjournals.org Cancer Res; 77(8) April 15, 2017 2031

Karekla et al.

A 100 100 Uncultured Uncultured 80 24 h 24 h Figure 2. 80 48 h 48 h Establishment of optimal explant culture 72 h 72 h 60 60 conditions. A, Evaluation of proliferation and cell death levels with increasing time 40 40 of explant culture. The percentage of Ki67 % Ki67 +ve (left graphs) and the percentage cPARP % cPARP +ve % cPARP 20 20 (right graphs) IHC staining of 5 NSCLC ex vivo explants over time are shown. The 0 0 LT31 LT33 LT34 LT36 LT38 LT31 LT33 LT34 LT36 LT38 staining values were generated from the uncultured tumor in its native state and 100 100 from explants from the same tumor at 24, 80 80 48, and 72 hours of culture after an initial recovery of 16 to 20 hours. The graphs at 60 60 the top show individual values for each of the 5 tumors and the graphs on the 40 40 bottom indicate pooled mean values for % Ki67 +ve

20 % cPARP +ve 20 the 5 samples 95% conﬁdence interval (CI). Page L nonparametric trend test 0 0 showed a negative trend for Ki67 staining Uncultured 24 h 48 h 72 h Uncultured 24 h 48 h 72 h with increasing culture time (P ¼ 0.01; L statistic ¼ 143) and a positive trend for 100 B 0 100 0 cPARP staining with increasing culture 0.5 0.5 time (P ¼ 0.05; L statistic ¼ 140). B, 80 80 1 1 Evaluation of percentages of Ki67 and %FCS

2.5 %FCS 2.5 60 5 60 5 cPARP staining with varying FCS concentration. The percentage of Ki67 40 40 (left graphs) and percentage of cPARP

% Ki67 +ve (right graphs) IHC staining of 5 NSCLC 20 +ve % cPARP 20 ex vivo explants over time are shown. The staining values were generated from the 0 0 uncultured tumor in its native state and LT31 LT32 LT33 LT36 LT38 LT31 LT32 LT33 LT36 LT38 from explants from the same tumor 100 100 cultured in varying FCS concentrations for 24 hours after an initial recovery of 16 to 80 80 20 hours. The graphs at the top show individual values for each of the 5 tumors, 60 60 and the graphs on the bottom indicate pooled mean values for the 5 samples 40 40 95% CI. Page L nonparametric trend test % Ki67 +ve

% cPARP +ve % cPARP showed no signiﬁcant difference for Ki67 20 20 or cPARP staining with varying FCS 0 0 concentration. C, Summary of the effect 0 0.5 1 2.5 5 0 0.5 1 2.5 5 of cultivation. The percentage of Ki67 (left %FCS %FCS graph) and percentage of cPARP (right 100 graph) staining were determined for 21 C 100 explant cultures. The box and whiskers P = 0.004 80 80 plots show the data for uncultured tumors P = 0.0001 compared with tumors cultured for 60 60 24 hours þ 16 to 20 hours of recovery in 1% FCS. The boxes extend from the 25th to 40 40 75th percentiles, and the lines in the % Ki67 +ve middle of the boxes represent the median. 20 % cPARP +ve 20 The whiskers extend from the smallest 0 0 to the largest values. Uncultured 24 h culture Uncultured 24 h culture

concentration. Pooled data for 21 explants under these conditions inant drug, with the drug given in combination with cisplatin are shown in Fig. 2C. Overall, there is a small but signiﬁcant effect not modifying the effect of chemotherapy on overall survival of cultivation on explant viability, with there being approximately (1, 32). Given this evidence, we focused on examining explant 10% decrease in proliferation and almost 10% increase in cell responses to cisplatin alone. death compared with the uncultured but freshly ﬁxed native Twenty-one explants were treated with a dose range of cisplatin tumor. (0–50 mmol/L) for 24 hours following the initial recovery of 16 to 20 hours. Data for individual cases are shown in Supplemen- Explant responses to cisplatin tary Fig. S3. Levels of cell proliferation were only marginally Adjuvant chemotherapy for NSCLC usually consists of cis- affected by the drug (Supplementary Fig. S3B), and therefore the platin often in combination with another chemotherapy emphasis was placed on assessing cell death responses (Supple- drug. Clinical trial data have shown that cisplatin is the dom- mentary Fig. S3A). Cell death response for each tumor was

2032 Cancer Res; 77(8) April 15, 2017 Cancer Research

Assessing Tumor Drug Response with a Novel Ex Vivo Platform

calculated as fold induction relative to the control over the dose export, drug deactivation, increased repair of DNA damage, or range (Fig. 3A). Of 21 tumors, 9 (43%) showed less than 2-fold alterations in apoptosis (33, 34). To examine drug uptake/export, induction in cell death in response to the drug, whereas the we investigated Pt ion distribution across explant tissue using LA- remaining 12 of 21 (57%) showed cell death induction ranging ICP-MS (see Supplementary Fig. S1) imaging (Fig. 4 and Supple- from 2- to 25-fold. The majority of these tumors only showed a mentary Fig. S5). For resistant cases, Pt ions were depleted from response at high levels of cisplatin (50 mmol/L), with only 2 areas corresponding to tumor cells but were present in the stroma tumors responding at the lower dose of 10 mmol/L. (Fig. 4A and Supplementary Fig. S5). In contrast, for sensitive In addition to the 21 explants treated with a dose range of cases, Pt ions were present throughout the tumor and stromal cisplatin, a further 9 were treated with a single dose of 50 mmol/L areas of the explant, indicating widespread cisplatin uptake (Fig. 4B cisplatin. We obtained clinical and histopathologic information on and Supplementary Fig. S5). Thus, while cisplatin is available to all 30 patients and their tumors (Supplementary Table S1). Cell the resistant explants, there is decreased intracellular drug concen- death difference compared with control in response to cisplatin is tration in tumor cells. included alongside this information. One tumor was excluded from the analysis due to complex histopathology and 3 atypical TP53 expression in the explants carcinoids were excluded because of their different biologic behav- The TP53 gene is frequently mutated in NSCLC (35). Wild- ior compared to adenocarcinomas and SCCs. For the remaining 26, type TP53 protein is induced by DNA-damaging agents such as a ROC curve was used to determine the threshold for resistance/ cisplatin, whereas mutated TP53 is either not expressed or is sensitivity to cisplatin and this analysis gave an area under the curve constitutively expressed. We utilized IHC to gain an indication of 0.6485 0.1122 SE, a likelihood ratio of 3.30 and identified of TP53 function in the 30 tumors, identifying 3 categories: 28.45% as the optimal cutoff (Supplementary Fig. S4A). (i) WTTP53 tumors (12 tumors), (ii) MUTTP53 tumors with We then categorized each explant into being either sensitive or constitutively high TP53 levels (16 tumors), and (iii) MUTTP53 resistant to cisplatin (Supplementary Fig. S4B and Supplementary tumors expressing undetectable TP53 (2 tumors). IHC TP53 Table S1). Using clinical information on corresponding patients staining of positive and negative tumors is shown in Fig. 5A, (Supplementary Table S1), the relationship of cisplatin sensitiv- whereas Fig. 5B indicates induction of TP53 following treat- ity/resistance in explant culture to patient survival post-surgery ment of a WTTP53 tumor with a dose range of cisplatin and was determined (Fig. 3B). The data show a statistically significant quantitation of the staining. Overall, 40% of tumors were relationship (P ¼ 0.006) with sensitive cases demonstrating a WTTP53 and 60% MUTTP53 based on IHC criteria (Supplemen- mean survival time (MST) of 28 months and resistant cases an tary Table S1). This is approximately consistent with the known MST of 14 months. To rule out an effect of tumor stage, we mutation rate of TP53 in human NSCLC (34). separated stage I/IIA and IIB/III cases (Fig. 3C). There is a statis- As expected, TP53MUT tumors had significantly higher intrin- tically significant relationship between cisplatin sensitivity and sic levels of proliferation compared with TP53WT tumors patient survival for both stage I/IIA cases (P ¼ 0.02) and for stage (Fig. 5C), and the majority of TP53MUT cancers were of the SCC IIB/III cases (P ¼ 0.05), indicating the correlation is independent subtype (Fig. 5D). In terms of response to cisplatin, TP53MUT of stage; importantly, this relationship was also shown to be samples had significantly higher levels of cell death induction independent of stage in a multivariate Cox survival analysis. Of compared with TP53WT samples (Fig. 5D, left) and significantly the 26 patients, 12 received adjuvant therapy (Supplementary higher levels of suppression of cell proliferation (Fig. 5D, right). Table S1), 8 received platinum-based chemotherapy, 3 received These data counteract the view that TP53MUT tumors are defec- radiotherapy, and 1 received taxane-based chemotherapy. In tive in their apoptotic response to DNA damage induced by those cases receiving any form of adjuvant therapy, there is a cisplatin. significant correlation with survival of patients and response to cisplatin in explants (P ¼ 0.01; Fig. 3D, left). However, there was Explant responses to TRAIL no relationship between cisplatin sensitivity in explants and TRAIL is a death receptor ligand that has been developed for survival for patients reported to have lymph node involvement therapy, although clinical trials have been disappointing (36). (P ¼ 0.13; Fig. 3D, right). Overall, the data show a strong It is thought that preclinical in vitro studies using cell lines have relationship between patient survival and explant sensitivity to not faithfully represented the clinical situation. This is sup- cisplatin, indicating that the explant platform is predictive of ported by data demonstrating that the majority of primary disease recurrence and response to adjuvant therapy. human tumor cells are resistant to TRAIL receptor agonists Cisplatin sensitivity in explants was also correlated with tumor (36–38). To investigate TRAIL sensitivity in NSCLC, 12 explants stage and histologic type (Fig. 3E). There was a significant negative were treated with TRAIL either as a single agent or in combi- trend between difference in percentage of cPARP staining com- nation with cisplatin (Fig. 6A). TRAIL alone did not elicit a pared with control in response to cisplatin and increasing tumor strong response, except for one case (LT22) that demonstrated stage (P ¼ 0.007), suggesting that more advanced tumors are more approximately 4-fold induction of cell death. Similarly, TRAIL resistant to the drug. There was also a significant correlation did not enhance the effects of cisplatin in the majority of cases, between cisplatin sensitivity and tumor type (P ¼ 0.0004), with except for one tumor (LT83) for which slightly greater cell death SCC cases demonstrating greater cisplatin sensitivity than adeno- induction (6-fold) than cisplatin alone (4-fold) was detected carcinoma subtypes (Fig. 3E). (Fig. 6A and B).

Cisplatin sensitivity is linked to drug accumulation in tumor areas Discussion A number of mechanisms have been reported to render cells Predicting drug response in patients with cancer is a major resistant to cisplatin including reduced drug uptake, enhanced challenge in the clinic. Cell line xenograft mouse models have

www.aacrjournals.org Cancer Res; 77(8) April 15, 2017 2033

Karekla et al.

AB30 25 All cases 100 Cis Sens Cis Res 20 P = 0.006 15 50 10 % Survival relative to control

Fold cPARP increase 5 0 0 0 80604020 0 1 10 50 Months elapsed Cisplatin dose (mmol/L)

Stage I+IIA Stage IIB+III 100 100 C Cis Sens Cis Sens Cis Res Cis Res P = 0.02 P = 0.05

50 50 % Survival % Survival

0 0 0 20 40 60 80 0 20 40 60 80 Months elapsed Months elapsed

Adjuvant therapy Lymph node involvement D 100 100 Cis Sens Cis Sens Cis Res Cis Res P = 0.01 P = 0.13

50 50 % Survival % Survival

0 0 0 80604020 0 40302010 Months elapsed Months elapsed

E P = 0.0004 100 100

50 50 from control

0 from control 0 Difference of % cPARP IIIIII ADC SCC Difference of % cPARP Stage Tumor type

Figure 3. NSCLC explant response to cisplatin. A, Fold cell death relative to control of 21 NSCLC explant cultures treated with a dose range (0–50 mmol/L) of cisplatin. The percentage of cPARP staining of each sample was quantitated within explants from the same tumor cultured in carrier control (DMF) or increasing cisplatin concentrations (1, 10, and 50 mmol/L) for 24 hours after an initial recovery of 16 to 20 hours. The value for each treatment was divided by the carrier control to obtain a fold change. B, Kaplan–Meier patient survival for all cases correlated with sensitivity of explants to cisplatin at 50 mmol/L. Data were evaluated for 26 patients/explants (Supplementary Table S1). The threshold of sensitivity/resistance to the drug was determined using an ROC curve (Supplementary Fig. S4A). The Mantel–Cox log-rank test identified a statistically significant relationship (P ¼ 0.006) between patient survival and cisplatin sensitivity. C, Kaplan–Meier patient survival according to stage correlated with sensitivity of explants to cisplatin at 50 mmol/L. Left graph, stage I/IIA samples (n ¼ 14, P ¼ 0.02). Right graph, stage IIB/III samples (n ¼ 12, P ¼ 0.05). D, Kaplan–Meier patient survival according to adjuvant chemotherapy or lymph node involvement correlated with sensitivity of explants to cisplatin at 50 mmol/L. Left graph, subgroup of patients receiving adjuvant therapy (n ¼ 12), including platinum-based chemotherapy (n ¼ 8), radiotherapy (n ¼ 3), and taxane-based chemotherapy (n ¼ 1). The Mantel–Cox log-rank test identified a statistically significant relationship between survival of patients and response to cisplatin in explants (P ¼ 0.01) Right graph, subgroup of patients with lymph node involvement (n ¼ 10) for which the Mantel–Cox log-rank test did not identify a statistically significant relationship (P ¼ 0.13) between patient survival and cisplatin sensitivity. E, Correlation of response to cisplatin in explant culture with tumor stage (left) and histology (right). The box and whiskers plots show data for difference in percentage of cPARP staining in response to 50 mmol/L cisplatin compared with control treatment for each explant relative to tumor stage/histology. The box extends from the 25th to 75th percentiles, and the lines in the middle of the boxes represent the median. The whiskers extend from the smallest to the largest values. The Jonckheere–Terpstra test for ordered alternatives showed a significant negative trend (P ¼ 0.007) between increasing stage and cisplatin response. Correlation of tumor histology with cisplatin response demonstrated statistically significant differences between SCC samples and adenocarcinoma (ADC) types with a Mann–Whitney test of P ¼ 0.0004.

2034 Cancer Res; 77(8) April 15, 2017 Cancer Research

Assessing Tumor Drug Response with a Novel Ex Vivo Platform

A H&E MNF116 cPARP

Ki67 LA-ICP-MS ≥ 7,500 6,625 5,750 4,875 4,000 3,125 2,250 1,375 ≤ 500

B H&E MNF116 cPARP

Ki67 LA-ICP-MS

≥ 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0

Figure 4. Pt ion distribution in cisplatin-sensitive and -resistant explants. A and B, Cisplatin-resistant explant (LT31; A) and cisplatin-sensitive explant (LT88; B). Both samples were treated with 10 mmol/L cisplatin. Serial sections of explants stained with H&E staining are shown in the top left, and IHC staining with antibodies for MNF116 (top middle), cPARP (top right), and Ki67 (bottom left) are also shown. LA-ICP-MS samplings to indicate the distribution of Pt ions within each explant are shown in the bottom right. The white lines indicate tumor areas as determined by H&E and MNF116 staining. Scale bars, 100 mm.

been extensively used for preclinical drug testing, but while genetic lesion (11, 40), in practice, these models are expensive these models can provide an initial indication of in vivo drug and lose the characteristics of the original human tumor micro- efﬁcacy, data are often not predictive of patient outcome environment over time. Here, we have perfected a rapid and (7, 39). Although the advent of mouse PDX models has opened low-cost platform that relies on the in situ assessment of drug up the possibility of tailoring drugs to a tumor with a speciﬁc responses within real human tumors. We validate this platform

www.aacrjournals.org Cancer Res; 77(8) April 15, 2017 2035

Karekla et al.

A B

100 µm µ µ 100 m 100 µm 100 m LT18 LT27 Carrier control 1 µmol/L cisplatin

µ 100 µm 100 m 100 µm 100 µm LT23 LT116 10 µmol/L cisplatin 50 µmol/L cisplatin

C 100 100

P = 0.002 80 80 60 60 40

40 20 % Ki67 +ve % Labeling index of p53 0 20 0 1 10 50 Cisplatin dose (mmol/L)

0 TP53 -ve TP53 +ve D P = 0.003 P = 0.006 100 10

ADC 0 ADC AC AC SCC SCC

–20 50

–40 %Decrease of Ki67 compared with control 0 –60 % Increase of cPARP compared with control Inducible p53 Noninducible p53 Inducible p53 Noninducible p53

Figure 5. TP53 expression. A, IHC staining of explants treated with cisplatin demonstrating tumors with constitutively high levels of nuclear TP53 in tumor cells (LT18, LT27, and LT23) or low levels of TP53 (LT116). Scale bars, 100 mm. B, IHC staining and quantitation of nuclear p53 staining in a TP53WT explant sample demonstrating dose-dependent TP53 expression following cisplatin treatment. The graph shows the percentage of mean labeling index of TP53 SD following treatment of the tumor sample with increasing doses of cisplatin (P ¼ 0.0001). Scale bars, 100 mm. C, Correlation of intrinsic proliferation index with TP53 IHC staining. The graph shows the percentage of Ki67 staining of tumors classified according to TP53 IHC staining. Each circle represents one sample. D, Induction of cell death, as assessed by cPARP staining (left graph), and reduction of proliferation (right graph), as assessed by Ki67 staining, upon 50 mmol/L cisplatin treatment of 30 NSCLC ex vivo explants stratified according to their ability to induce TP53 expression upon treatment with the drug. The data show that TP53-inducible tumors have a significantly reduced (P ¼ 0.003) ability to undergo cell death in response to cisplatin compared with TP53-noninducible tumors. The majority of these tumors are of the SCC subtype.

2036 Cancer Res; 77(8) April 15, 2017 Cancer Research

Assessing Tumor Drug Response with a Novel Ex Vivo Platform

A 18

16 Carrier control

14 TRAIL 1 µg/mL

50 µmol/L Cisplatin 12 TRAIL 1 µg/mL + 50 µmol/L Cisplatin 10

4 Figure 6. Response to TRAIL. A, Fold cell death 2 Fold cPARP increase relative to control induction in response to 1 mg/mL of TRAIL, 50 mmol/L cisplatin, or a 0 combination of the two relative to LT15 LT16 LT18 LT20 LT22 LT23 LT27 LT83 LT84 LT88 LT89 LT92 the carrier control is shown. The percentage of cell death of 12 ex vivo Carrier 1 mg/µL TRAIL 50 µmol/L Cisplatin TRAIL + Cisplatin explant cultures as determined by B cPARP staining was determined after each treatment for 24 hours after an initial recovery of 16 to 20 hours. These values were divided by the value for LT83 each carrier control to calculate the fold difference. B, Representative images of cPARP staining of LT83, LT18, and LT22 treated with 1 mg/mL of TRAIL, 50 mmol/L cisplatin, or a combination of the two. Scale bars, 100 mm. LT18

LT22

by showing that explant response to the standard-of-care culture. Correlation of organotypic culture data with patient chemotherapy drug cisplatin is related to survival of patients outcomes has been previously reported for the Histoculture Drug (P ¼ 0.006), indicating that explant response to cisplatin is Response Assay system (15–17). However, a disadvantage of this predictive of disease progression. Responses to the targeted technique is that the endpoint requires enzymatic digestion of agent TRAIL are also more consistent with clinical outcomes tissue, thus preventing assessment of the speciﬁc cell type affected than standard cell line model systems (36–38). We demon- by the drug. This disadvantage can be overcome by using our in strate how the explant platform can be used to inform mechan- situ FFPE/IHC approach. isms of drug action by biomarker monitoring. Our data show that the majority of the cisplatin-resistant A number of organotypic culture systems have been previously tumors are of a higher stage (Fig. 3 and Supplementary Table developed for human tumors (13–18). In most of these previous S1), but the ability to induce cell death in response to cisplatin systems, viability of tumors has been demonstrated for up to 7 does not correlate with intact TP53 (Fig. 5). In fact, we have days (13–19). Here, we identiﬁed a mild effect of cultivation after found that TP53-mutated NSCLC cases are more sensitive to 24 hours of culture (Fig. 2C), but tissue architecture was main- cisplatin in explants thanWTTP53 cases (Fig. 5D). Previous tained intact for up to 72 hours. Our preference is to examine drug studies have investigated whether TP53 mutations are of prog- responses immediately after cultivation to minimize any effects of nostic value in predicting response to chemotherapy in NSCLC;

www.aacrjournals.org Cancer Res; 77(8) April 15, 2017 2037

Karekla et al.

the results are controversial (41). In a 35-patient study, the Authors' Contributions presence of mutant TP53 was highly indicative of resistance to Conception and design: B. Sharp, C. Pritchard, M. MacFarlane, J.H. Pringle cisplatin (P ¼ 0.002) (42), while in a study involving 253 Development of methodology: E. Karekla, B. Sharp, J. Le Quesne, C. Pritchard, patients, TP53-positive patients had a significantly greater sur- M. MacFarlane, J.H. Pringle Acquisition of data (provided animals, acquired and managed patients, vival benefit from adjuvant chemotherapy compared with provided facilities, etc.): E. Karekla, W.-J. Liao, J. Pugh, M. MacFarlane TP53-negative patients (43). Another report in the Internation- Analysis and interpretation of data (e.g., statistical analysis, biostatistics, al Adjuvant Lung Cancer Trial (IALT), a randomized trial of computational analysis): E. Karekla, B. Sharp, J. Pugh, H. Reid, J. Le Quesne, adjuvant cisplatin-based chemotherapy, found no correlation D. Moore, C. Pritchard, M. MacFarlane, J.H. Pringle between TP53 mutation and outcome in 524 patients (44). Writing, review, and/or revision of the manuscript: E. Karekla, B. Sharp, Overall, it will be important to extend analysis to a greater J. Pugh, H. Reid, C. Pritchard, M. MacFarlane, J.H. Pringle Administrative, technical, or material support (i.e., reporting or organizing number of explants/patients to robustly determine the prog- TP53 data, constructing databases): E. Karekla, H. Reid, J.H. Pringle nostic value of mutation. Lack of response to cisplatin Study supervision: C. Pritchard, M. MacFarlane, J.H. Pringle does, however, correlate with exclusion of the drug from tumor Other (histopathologicinterpretation): J. Le Quesne areas (Fig. 4). Cisplatin import is mediated by the copper transporter CTR1, whereas the copper transporters ATP7A and Acknowledgments ATP7B regulate the efflux of cisplatin (45). Resistance to cis- We thank Andrew Wardlaw for providing the ethical framework for this platin has been associated with alterations in the expression project; Hilary Marshall and Will Monteiro for support in tissue collection; fi status of these transporters (46) and so it will also be important thoracic surgeons at Glen eld Hospital, Leicester, for providing clinical samples; and Chris Baines for providing clinical data. We also thank Angie Gillies for to evaluate these transporters in the explant system used here. assistance with histology. In summary, the explant platform provides a patient-relevant model system for the preclinical evaluation of novel anticancer Grant Support fi agents. When combined with tumor strati cation approaches, This work was supported by a Medical Research Council (MRC) Doctoral the platform has the potential for personalizing drug treatment. Training Grant to E. Karekla, the MRC Toxicology Unit (MC A/600), and the The technology is low-cost, rapid, and achievable within an Leicester Experimental Cancer Medicine Centre (C325/A15575 Cancer Research integrated cancer translational research setting. An important UK/UK Department of Health). C. Pritchard was supported by a Royal Society- next step will be to conduct a clinical study aimed at deter- Wolfson merit award. mining which patients would best respond to chemotherapy The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked prospectively; such a study is currently being developed within advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate our center. this fact.

Disclosure of Potential Conﬂicts of Interest Received April 18, 2016; revised January 10, 2017; accepted January 30, 2017; No potential conﬂicts of interest were disclosed. published OnlineFirst February 15, 2017.

References 1. Pignon JP, Tribodet H, Scagliotti GV, Douillard JY, Shepherd FA, Stephens personalized tumorgrafts in patients with advanced cancer. Mol Cancer RJ, et al. Lung adjuvant cisplatin evaluation: a pooled analysis by the Ther 2011;10:1311–6. LACE Collaborative Group. J Clin Oncol 2008;26:3552–9. 12. Bertotti A, Migliardi G, Galimi F, Sassi F, Torti D, Isella C, et al. A 2. Douillard JY, Tribodet H, Aubert D, Shepherd FA, Rosell R, Ding K, et al. molecularly annotated platform of patient-derived xenografts ("xenopa- Adjuvant cisplatin and vinorelbine for completely resected non-small cell tients") identifies HER2 as an effective therapeutic target in cetuximab- lung cancer: subgroup analysis of the Lung Adjuvant Cisplatin Evaluation. resistant colorectal cancer. Cancer Discov 2011;1:508–23. J Thorac Oncol 2010;5:220–8. 13. Pampaloni F, Reynaud EG, Stelzer EH. The third dimension bridges the gap 3. Mok TS, Wu YL, Thongprasert S, Yang CH, Chu DT, Saijo N, et al. Gefitinib between cell culture and live tissue. Nat Rev Mol Cell Biol 2007;8:839–45. or carboplatin-paclitaxel in pulmonary adenocarcinoma. N Engl J Med 14. Freeman AE, Hoffman RM. In vivo-like growth of human tumors in vitro. 2009;361:947–57. Proc Natl Acad Sci U S A 1986;83:2694–8. 4. Rosell R, Molina MA, Serrano MJ. EGFR mutations in circulating tumour 15. Yoshimasu T, Oura S, Hirai I, Tamaki T, Kokawa Y, Hata K, et al. Data DNA. Lancet Oncol 2012;13:971–3. acquisition for the histoculture drug response assay in lung cancer. J Thorac 5. Solomon BJ, Mok T, Kim DW, Wu YL, Nakagawa K, Mekhail T, et al. First- Cardiovasc Surg 2007;133:303–8. line crizotinib versus chemotherapy in ALK-positive lung cancer. N Engl J 16. Kubota T, Sasano N, Abe O, Nakao I, Kawamura E, Saito T, et al. Potential of Med 2014;371:2167–77. the histoculture drug-response assay to contribute to cancer patient sur- 6. Kola I, Landis J. Can the pharmaceutical industry reduce attrition rates? vival. Clin Cancer Res 1995;1:1537–43. Nat Rev Drug Discov 2004;3:711–5. 17. Furukawa T, Kubota T, Hoffman RM. Clinical applications of the histo- 7. Johnson JI, Decker S, Zaharevitz D, Rubinstein LV, Venditti JM, Sche- culture drug response assay. Clin Cancer Res 1995;1:305–11. partz S, et al. Relationships between drug activity in NCI preclinical 18. Vaira V, Fedele G, Pyne S, Fasoli E, Zadra G, Bailey D, et al. Preclinical model in vitro and in vivo models and early clinical trials. Br J Cancer 2001;84: of organotypic culture for pharmacodynamic profiling of human tumors. 1424–31. Proc Natl Acad Sci U S A 2010;107:8352–6. 8. Fricker J.Time for reform in the drug-development process. Lancet Oncol 19. Majumder B, Baraneedharan U, Thiyagarajan S, Radhakrishnan P, Nara- 2008;9:1125–6. simhan H, Dhandapani M, et al. Predicting clinical response to anti- 9. Tentler JJ, Tan AC, Weekes CD, Jimeno A, Leong S, Pitts TM, et al. Patient- cancer drugs using an ex vivo platform that captures tumour heterogeneity. derived tumour xenografts as models for oncology drug development. Nat Nat Commun 2015;6:6169. Rev Clin Oncol 2012;9:338–50. 20. Harper N, MacFarlane M. Recombinant TRAIL and TRAIL receptor analysis. 10. Siolas D, Hannon GJ. Patient-derived tumor xenografts: transforming Methods Enzymol 2008;446:293–313. clinical samples into mouse models. Cancer Res 2013;73:5315–9. 21. MacFarlane M, Ahmad M, Srinivasula SM, Fernandes-Alnemri T, Cohen 11. Hidalgo M, Bruckheimer E, Rajeshkumar NV, Garrido-Laguna I, De Oli- GM, Alnemri ES. Identification and molecular cloning of two novel veira E, Rubio-Viqueira B, et al. A pilot clinical study of treatment guided by receptors for the cytotoxic ligand TRAIL. J Biol Chem 1997;272:25417–20.

2038 Cancer Res; 77(8) April 15, 2017 Cancer Research

Assessing Tumor Drug Response with a Novel Ex Vivo Platform

22. Caramel J, Papadogeorgakis E, Hill L, Browne GJ, Richard G, Wierinckx A, 33. Kelland LR.Preclinical perspectives on platinum resistance. Drugs 2000;59 et al. A switch in the expression of embryonic EMT-inducers drives the Suppl 4:1–8; discussion 37–8. development of malignant melanoma. Cancer Cell 2013;24:466–80. 34. Cosaert J, Quoix E. Platinum drugs in the treatment of non-small-cell lung 23. Schneider CA, Rasband WS, Eliceiri KW. NIH Image to ImageJ: 25 years of cancer. Br J Cancer 2002;87:825–33. image analysis. Nat Methods 2012;9:671–5. 35. Le Calvez F, Mukeria A, Hunt JD, Kelm O, Hung RJ, Taniere P, et al. TP53 24. Tuominen VJ, Ruotoistenmaki S, Viitanen A, Jumppanen M, Isola J. and KRAS mutation load and types in lung cancers in relation to tobacco ImmunoRatio: a publicly available web application for quantitative image smoke: distinct patterns in never, former, and current smokers. Cancer Res analysis of estrogen receptor (ER), progesterone receptor (PR), and Ki-67. 2005;65:5076–83. Breast Cancer Res 2010;12:R56. 36. Amarante-Mendes GP, Griffith TS. Therapeutic applications of TRAIL 25. Koppen C, Reifschneider O, Castanheira I, Sperling M, Karst U, Ciarimboli receptor agonists in cancer and beyond. Pharmacol Ther 2015;155 G. Quantitative imaging of platinum based on laser ablation-inductively :117–31. coupled plasma-mass spectrometry to investigate toxic side effects of 37. Dyer MJ, MacFarlane M, Cohen GM. Barriers to effective TRAIL-targeted cisplatin. Metallomics 2015;7:1595–603. therapy of malignancy. J Clin Oncol 2007;25:4505–6. 26. Baldwin DR, White B, Schmidt-Hansen M, Champion AR, Melder AM; 38. MacFarlane M, Kohlhaas SL, Sutcliffe MJ, Dyer MJ, Cohen GM. TRAIL Guideline Development Group. Diagnosis and treatment of lung cancer: receptor-selective mutants signal to apoptosis via TRAIL-R1 in primary summary of updated NICE guidance. BMJ 2011;342:d2110. lymphoid malignancies. Cancer Res 2005;65:11265–70. 27. Swinson DE, Jones JL, Richardson D, Cox G, Edwards JG, O'Byrne KJ. 39. Kung AL.Practices and pitfalls of mouse cancer models in drug discovery. Tumour necrosis is an independent prognostic marker in non-small cell Adv Cancer Res 2007;96:191–212. lung cancer: correlation with biological variables. Lung Cancer 2002;37: 40. Aparicio S, Hidalgo M, Kung AL. Examining the utility of patient-derived 235–40. xenograft mouse models. Nat Rev Cancer 2015;15:311–6. 28. Mehdi SA, Etzell JE, Newman NB, Weidner N, Kohman LJ, Graziano SL. 41. Toyooka S, Tsuda T, Gazdar AF. The TP53 gene, tobacco exposure, and lung Prognostic significance of Ki-67 immunostaining and symptoms in cancer. Hum Mutat 2003;21:229–39. resected stage I and II non-small cell lung cancer. Lung Cancer 1998; 42. Kandioler D, Stamatis G, Eberhardt W, Kappel S, Zochbauer-Muller S, 20:99–108. Kuhrer I, et al. Growing clinical evidence for the interaction of the p53 29. Nguyen VN, Mirejovsky P, Mirejovsky T, Melinova L, Mandys V. Expression genotype and response to induction chemotherapy in advanced non-small of cyclin D1, Ki-67 and PCNA in non-small cell lung cancer: prognostic cell lung cancer. J Thorac Cardiovasc Surg 2008;135:1036–41. significance and comparison with p53 and bcl-2. Acta Histochem 2000; 43. Tsao MS, Aviel-Ronen S, Ding K, Lau D, Liu N, Sakurada A, et al. Prognostic 102:323–38. and predictive importance of p53 and RAS for adjuvant chemotherapy in 30. Warth A, Cortis J, Soltermann A, Meister M, Budczies J, Stenzinger A, et al. non small-cell lung cancer. J Clin Oncol 2007;25:5240–7. Tumour cell proliferation (Ki-67) in non-small cell lung cancer: a critical 44. Ma X, Rousseau V, Sun H, Lantuejoul S, Filipits M, Pirker R, et al. reappraisal of its prognostic role. Br J Cancer 2014;111:1222–9. Significance of TP53 mutations as predictive markers of adjuvant cisplat- 31. Twiddy D, Edwards J, Walker RA, Cohen GM, MacFarlane M. A TRAIL- in-based chemotherapy in completely resected non-small-cell lung cancer. R1-specific ligand in combination with doxorubicin selectively targets Mol Oncol 2014;8:555–64. primary breast tumour cells for apoptosis. Breast Cancer Res 2010; 45. Safaei R, Howell SB. Copper transporters regulate the cellular phar- 12Suppl 1:17–8. macology and sensitivity to Pt drugs. Crit Rev Oncol Hematol 2005; 32. Cortes J, Rodriguez J, Calvo E, Gurpide A, Garcia-Foncillas J, Salgado E, 53:13–23. et al. Paclitaxel, cisplatin, and vinorelbine combination chemotherapy in 46. Kuo MT, Chen HH, Song IS, Savaraj N, Ishikawa T. The roles of copper metastatic non-small-cell lung cancer. Am J Clin Oncol 2004;27:299–303. transporters in cisplatin resistance. Cancer Metastasis Rev 2007;26:71–83.

www.aacrjournals.org Cancer Res; 77(8) April 15, 2017 2039

Ex Vivo Explant Cultures of Non−Small Cell Lung Carcinoma Enable Evaluation of Primary Tumor Responses to Anticancer Therapy

Ellie Karekla, Wen-Jing Liao, Barry Sharp, et al.

Cancer Res 2017;77:2029-2039. Published OnlineFirst February 15, 2017.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-16-1121

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2017/02/15/0008-5472.CAN-16-1121.DC1

Cited articles This article cites 46 articles, 14 of which you can access for free at: http://cancerres.aacrjournals.org/content/77/8/2029.full#ref-list-1

Citing articles This article has been cited by 5 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/77/8/2029.full#related-urls

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 Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/77/8/2029. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.