Registered report: Discovery and preclinical validation of drug indications using compendia of public gene expression data.

The Science Translational Medicine in 2011 (Sirota et al., 2011). The key experiments being replicated include Figure 4C and D and Supplemental Figure 1. In these figures, Sirota and colleagues. tested a proof of concept experiment validating their prediction that cimetidine, a histamine-2 (H2) receptor agonist commonly used to treat peptic ulcers (Kubecova et al., 2011), would be effective against lung adenocarcinoma (Figure 4C and D). As a control they also tested the effects of cimetidine against renal carcinoma, for which it was not predicted to be efficacious (Supplemental Figure 1). The Reproducibility Project: Cancer Biology is a collaboration between the eLife.

Introduction

In this paper, Sirota and colleagues tested their hypothesis that extant drugs could be repurposed to target alternative diseases; if so, this could improve efficiency in the search for new treatments. They compared data from the Gene Expression Omnibus (GEO)—which they used to determine gene expression signatures of diseases—to data from the Connectivity Map, which tracks the changes in mRNA expression caused by 164 drugs. By comparing these two mRNA expression sets, Sirota and colleagues created a similarity score to describe how similar the changes in mRNA expression were between each drug and each disease. They theorized that a similarity score close to −1 (exactly opposite signatures) might indicate that the drug could treat the disease.

In Figure 4C and D, Sirota and colleagues directly test their hypothesis by examining the effects of cimetidine, an H2 receptor blocker commonly used to treat gastric ulcers by reducing the production of stomach acid (Kubecova et al., 2011), on xenograft transplanted A549 lung adenocarcinoma cells. Mice treated with cimetidine showed a dose-dependent reduction in tumor size after 12 days of treatment. In Supplemental Figure 1, they also treated ACHN renal carcinoma cells with cimetidine, although cimetidine was not predicted to treat this cancer line. They observed no effect of cimetidine on the growth of this cancer cell line. These experiments will be replicated in Protocol 1. However, the conclusions that can be drawn from these experiments are limited by the fact that only a single cell line was tested with only a single drug.

To date, no direct replication of the experiments presented in Sirota and colleagues' Figure 4C and D or Supplemental Figure 1 has been reported. However, Stoyanov and colleagues did examine the effect of cimetidine on growth of A459 cells activated with histamine and reported that cimetidine did reduce proliferation in vitro (Stoyanov et al., 2012). An exploratory analysis of a cohort of diabetic patients demonstrated a decreased risk of developing lung cancer, specifically adenocarcinoma, in patients who took over-the-counter H2 receptor blockers, including cimetidine (Hsu et al., 2013).

Materials and methods

Unless otherwise noted, all protocol information was derived from the original paper, references from the original paper, or information obtained directly from the authors. An asterisk (*) indicates data or information provided by the Reproducibility Project: Cancer Biology core team. A hashtag (#) indicates information provided by the replicating lab.

Protocol 1: assessing the effect of cimetidine treatment on tumor growth in a xenograft model of lung carcinoma and a xenograft model of renal carcinoma

This protocol describes how to create xenograft tumors in severe combined immunodeficient (SCID) mice from A549 lung carcinoma cells (as seen in Figure 4C and D) or ACHN renal carcinoma cells (Supplemental Figure 1). Tumor growth is then assessed during 11 days of cimetidine treatment. Sirota and colleagues designed this experiment to test their predictions that A549 cells would be susceptible to cimetidine treatment while ACHN cells would not. Treatment with phosphate-buffered saline (PBS) alone will serve as the negative control, while treatment with the lung adenocarcinoma standard drug doxorubicin will serve as the positive control.

Sampling

This experiment will use at least 12 mice per group for a final power of 82.4%.

See ‘Power calculations’ for details.

The experiment contains five cohorts total:

A549 lung adenocarcinoma xenografts:

a. Cohort 1: mice treated with PBS (negative control).

i. N = 14.

■ To ensure at least 12 tumors develop.

b. Cohort 2: mice treated with 2 mg/kg doxorubicin (Dox) (positive control).

i. N = 5.

■ To ensure at least 3 tumors develop.

c. Cohort 3: mice treated with 100 mg/kg cimetidine.

i. N = 14.

■ To ensure at least 12 tumors develop.

ACHN renal carcinoma xenografts:

a. Cohort 1: mice treated with PBS (negative control).

i. N = 14.

■ To ensure at least 12 tumors develop.

b. Cohort 2: mice treated with 100 mg/kg cimetidine.

i. N = 14.

■ To ensure at least 12 tumors develop.

Materials and reagents

Procedure

Notes:

A549 cells are maintained in F-12 Ham's medium supplemented with 10% fetal bovine serum (FBS), 2 mM sodium pyruvate and 1 μM lipoic acid, based on ATCC recommendations.

Lipoic acid is maintained as a 50 mg/ml stock in ethanol.

ACHN cells are maintained in EMEM supplemented with 10% FBS, 2 mM glutamine and 1 mM sodium pyruvate, based on ATCC recommendations.

All cells are grown at 37°C/5% CO2.

All cell lines will be sent for STR profiling and mycoplasma testing.

Culture A549 cells and ACHN cells.

Resuspend 5 × 106 cells in 100# μl PBS per injection.

Inject 5 × 106 cells (i.e., 100 µl of cell suspension) into the upper flank of 4–6-week-old* female SCID mice.

a. Mice will be randomly assigned to receive injections with A549 cells or ACHN cells.

i. Injections will be balanced so the total number of mice receiving A549 injections will be 33 and ACHN will be 28.

Measure tumor volume with calipers daily.

a. Record daily tumor volume.

b. Volume is defined as mm3 = 0.52 × [width (cm)]2 × height (cm).

i. Mice shall be euthanized if they appear in undue distress according to the replicating lab's guidelines; if the animal has lost >20% body weight.

When tumor reaches a minimum of 100 mm3 in volume (estimated time 2–3 weeks#), initiate treatment. Continue treatment for 11 days past this point.

a. As each mouse reaches the injection criteria (i.e., 100 mm3 tumor volume), randomly assign to a treatment group using the adaptive randomization approach with the time from injection of cells to when tumors reach at least 100 mm3 and tumor volume at time of assignment as the covariates that are assessed as mice are sequentially assigned to a particular treatment group.

i. Assignment will also take into account the pre-determined size of each treatment group.

b. Treat mice by intraperitoneal injection according to cohort:

i. A459 lung adenocarcinoma xenografts:

Cohort 1: PBS (daily).

Cohort 2: 2 mg/kg Doxorubicin (biweekly).

Cohort 3: 100 mg/kg cimetidine (daily).

ii. ACHN renal carcinoma xenograft injections:

Cohort 1: PBS (daily).

Cohort 2: 100 mg/kg cimetidine (daily).

c. Continue daily tumor volume measurements.

Euthanize mice.

a. Euthanize mice by CO2 inhalation followed by cervical dislocation.

Harvest tumors and record weight (additional parameter).

a. Image tumors alongside a ruler.

Deliverables

Data to be collected:

Mouse health records, including age and tumor volume at start of injections, time of tumor detection, any excluded mice (including reason for exclusion).

Raw data of tumor dimensions by day.

Final weight of tumors.

Graph of relative mean tumor weight in each cohort starting on Day 1 post-100 mm3 (as seen in Figure 4C and Supplemental Figure 1).

a. Normalize Day 2 onwards to the weight at Day 1.

Image of all tumors alongside ruler (as seen in Figure 4D) for both A459 xenografts and ACHN xenografts.

Confirmatory analysis plan

Statistical analysis of replication data:

At the time of analysis, we will perform the Shapiro–Wilk test and generate a quantile–quantile (q–q) plot to attempt to assess the normality of the data and also perform Levene's test to assess homoscedasiticity. If the data appear skewed, we will attempt a transformation in order to proceed with the proposed statistical analysis listed below and possibly perform the appropriate non-parametric test.

a. Comparison of the mean relative tumor weight of 100 mg/kg cimetidine treatment at day 11 as compared to PBS treatment at day 11 for both A549 and ACHN xenograft tumors.

i. Two-way analysis of variance (ANOVA) (2 × 2 factorial) followed by Bonferroni corrected Welch's t-tests for the following comparisons:

■ PBS-treated A549 tumors vs cimetidine-treated A459 tumors.

■ PBS-treated ACHN tumors vs cimetidine-treated ACHN tumors.

ii. Additional comparison of PBS-treated A459 tumors to doxorubicin-treated tumors.

■Bonferroni corrected Welch's t-test outside the framework of the ANOVA.

Meta-analysis of original and replication attempt effect sizes:

This replication attempt will perform the statistical analysis listed above, compute the effects sizes, compare them against the reported effect size in the original paper and use a meta-analytic approach to combine the original and replication effects, which will be presented as a forest plot.

Known differences from the original study

The replication attempt will encompass the PBS control, the doxorubicin control and the highest dose of cimetidine (100 mg/kg). It will not include the 25 mg/ml or 50 mg/ml cimetidine treatment groups.

While the original study performed injections of 5 × 106 cells per microliter of PBS, on the advice of the replicating lab we will inject the same number of cells but suspended in 100 µl PBS.

Provisions for quality control

Mice will be randomly assigned to xenograft model and treatment type. All data obtained from the experiment—raw data, data analysis, control data and quality control data—will be made publicly available, either in the published manuscript or as an open access dataset available on the Open Science Framework (https://osf.io/hxrmm/).

Power calculations

For details on power calculations, please see analysis files on the Open Science Framework:

https://osf.io/uazfe/wiki/home/

Protocol 1

Note: data values estimated from published figures. Error bars assumed to represent SEM.

Summary of original data

Test family

Two way ANOVA (2 × 2 factorial, PBS cohort and cimetidine cohorts only) followed by Bonferroni corrected Welch's t-tests for the following comparisons:

PBS-treated A549 tumors vs cimetidine-treated A459 tumors.

PBS-treated ACHN tumors vs cimetidine-treated ACHN tumors.

Comparison of PBS-treated A459 tumors to doxorubicin-treated tumors.

Bonferroni corrected Welch's t-test outside the framework of the ANOVA.

Power calculations

Power calculations were performed with R software 3.1.2 (R Core team, 2014) and G*Power (Faul et al., 2007).

eLife posts the editorial decision letter and author response on a selection of the published articles (subject to the approval of the authors). An edited version of the letter sent to the authors after peer review is shown, indicating the substantive concerns or comments; minor concerns are not usually shown. Reviewers have the opportunity to discuss the decision before the letter is sent (see review process). Similarly, the author response typically shows only responses to the major concerns raised by the reviewers.

Thank you for sending your work entitled “Registered report: Discovery and preclinical validation of drug indications using compendia of public gene expression data” for consideration at eLife. Your article has been favorably evaluated by Stylianos Antonarakis (Senior editor), Chi Dang (Reviewing editor), and 3 reviewers, one of whom is a biostatistician.

The Reviewing editor and the reviewers discussed their comments before we reached this decision, and the Reviewing editor has assembled the following comments to help you prepare a revised submission.

In this study, the authors propose a study to reproduce the findings reported in Figure 4C/D and Supplementary Figure 1 from a previously published manuscript (Sirota et al. Sci Trans Med, 2010), which aimed at assessing the ability to predict drug repurposing opportunities based on connectivity map data analysis. Specifically, the previous Sci Trans Med paper reports that cimetidine, a histamine-2 (H2) receptor agonist commonly used to treat peptic ulcers, can diminish lung cancer tumorigenesis in vivo. There are several key concerns about the design of the study. The first concern is about the duration of the experiment and statistical analysis, and the second about conclusions drawn from using only one lung cancer cell line.

1) At the beginning of the Materials and methods section: The authors plan to follow the mice for 11 days instead of 12 days. Is there a good reason to follow the mice one day short? In addition, the experiment contains five cohorts. Among the five cohorts, cohort 2 only has 5 mice while the other 4 cohorts have 14 mice. Please justify.

2) Power calculation was based on t-test. It is suggested that the authors use two-tailed unequal variance t-test if normality is not violated or the use of Wilcoxon rank-sum test if normality is violated. The authors propose the use of two-way ANOVA followed by t-test for analyzing tumor weight data (in the subsection headed “Confirmatory analysis plan”). Please make sure that the data do not violate the assumptions of ANOVA: normality and homoscedasiticity. If the data do not fit the assumptions well enough, please try to find a data transformation that makes them fit. If this doesn't work, please apply a nonparametric counterpart of ANOVA such as Kruskal–Wallis test. In addition, I suggest the use of contrast within the ANOVA framework instead of t-test if the assumptions of ANOVA are met.

3) To compare growth curves of tumors, the authors propose ANCOVA followed by Bonferroni corrected t-test. Please make sure that the data do not violate the assumptions of ANCOVA and perform transformation or use non-parametric ANCOVA if needed.

4) For the additional comparison of PBS-treated A459 tumors to Doxorubicin treated tumors (in the subsection headed “Confirmatory analysis plan” and in the subsection headed “Test family”), I suggest the use of two-tailed unequal variance t-test instead of t-test if normality is not violated or the use of Wilcoxon rank-sum test if normality is violated.

5) Although the reproducibility project is aimed toward reproducing previously published results, the reviewers would like for the authors to address the limitation of drawing conclusions for the use of only one cell line, A549. Specifically, activity of drugs in cell lines and xenografts is generally highly idiosyncratic. As a result, most journals require that any in vitro and in vivo experiments are replicated in multiple cell lines and in vivo models.

1) At the beginning of the Materials and methods section: The authors plan to follow the mice for 11 days instead of 12 days. Is there a good reason to follow the mice one day short? In addition, the experiment contains five cohorts. Among the five cohorts, cohort 2 only has 5 mice while the other 4 cohorts have 14 mice. Please justify.

Figure 4C and Supplemental Figure 1 show data from Days 1 through 11. Although Day 12 is displayed on the graphs, no data is present on that timepoint. We interpreted this to mean that the original authors counted Day 0 as one of the days of monitoring; including Day 0 accounts for the 12 days of monitoring mentioned by the authors and reconciles that statement with the 11 days of data displayed in Figure 4D and Supplemental Figure 1.

For budgetary and ethical reasons, we wished to minimize the number of animals required for these experiments. Thus, we did not use equal sample sizes for the additional positive control (doxorubicin) treatment group when performing our power calculations. These calculations demonstrated that we could use 3 mice in the control group and still achieve 80% power to detect the original data’s effect size. The main aim of the experiment is to test the effect of vehicle (PBS) treatment compared to experimental (cimetidine) treatment, which includes an equal number of mice in each cohort. Considering the unbalanced nature of this design we will be performing a planned comparison between the PBS and doxorubicin outside the framework of the ANOVA, however, the four cohorts that are the main aim of the experiment are analyzed within the balanced ANOVA framework.

2) Power calculation was based on t-test. It is suggested that the authors use two-tailed unequal variance t-test if normality is not violated or the use of Wilcoxon rank-sum test if normality is violated. The authors propose the use of two-way ANOVA followed by t-test for analyzing tumor weight data (in the subsection headed “Confirmatory analysis plan”). Please make sure that the data do not violate the assumptions of ANOVA: normality and homoscedasiticity. If the data do not fit the assumptions well enough, please try to find a data transformation that makes them fit. If this doesn't work, please apply a nonparametric counterpart of ANOVA such as Kruskal–Wallis test. In addition, I suggest the use of contrast within the ANOVA framework instead of t-test if the assumptions of ANOVA are met.

We have added language to the manuscript to clarify that we will perform the normality and homoscedasticity tests. The original data was not shared, so instead summary statistics estimated from the graph presented in Figure 4C and Supplemental Figure 1 were used. This limits what we can ascertain from the original data. We recalculated the samples sizes using a Welch’s t-test instead of a Student’s. The sample size we have planned is still sufficient. This is also true for the non-parametric Wilcoxon rank-sum test. We plan to use the contrast within the ANOVA framework if the assumptions are met, with the exception of the doxorubicin cohort, which will be performed outside the framework because of the unbalanced design.

3) To compare growth curves of tumors, the authors propose ANCOVA followed by Bonferroni corrected t-tests. Please make sure that the data do not violate the assumptions of ANCOVA and perform transformation or use non-parametric ANCOVA if needed.

Pursuant to one of the later comments, we have removed the exploratory analysis by area under the curve from the manuscript.

4) For the additional comparison of PBS-treated A459 tumors to Doxorubicin treated tumors (in the subsection headed “Confirmatory analysis plan” and in the subsection headed “Test family”), I suggest the use of two-tailed unequal variance t-test instead of t-test if normality is not violated or the use of Wilcoxon rank-sum test if normality is violated.

We have added language to the manuscript to clarify that we will perform the normality and homoscedasticity tests. As described earlier, we used summary statistics estimated from the published data to perform power calculations. This limits what we can ascertain from the original data. We recalculated the samples sizes using a Welch’s t-test instead of a Student’s. The sample size we have planned is still sufficient. This is also true for the non-parametric Wilcoxon rank-sum test.

5) Although the reproducibility project is aimed toward reproducing previously published results, the reviewers would like for the authors to address the limitation of drawing conclusions for the use of only one cell line, A549. Specifically, activity of drugs in cell lines and xenografts is generally highly idiosyncratic. As a result, most journals require that any in vitro and in vivo experiments are replicated in multiple cell lines and in vivo models.

Thank you for providing this insight. We have added a sentence addressing this point to the Introduction. We will also include this point in the Discussion section of the Replication Study that will be published once the replication data has been generated. We agree the use of one model limits the overall conclusions that can be drawn. This project focuses on understanding if the effects drawn from a single model can be reproduced. While this does not speak to the robustness of the effect, such as can be inferred through multiple models/approaches, it does provide a mechanism to examine the extent to which an effect with a given model can be observed again. We will also limit the conclusions that can be drawn to only this model.

Reagent	Type	Manufacturer	Catalog #	Comments
PBS	Reagent	Invitrogen	10010023
Fetal bovine serum	Reagent	Invitrogen	16000-044
A549 cells	Cells	ATCC	#CCL-185	Original unspecified
ACHN cells	Cells	ATCC	#CRL-1611	Original unspecified
Hydrochloric acid (HCl)	Chemical	Sigma–Aldrich	320331
Sodium hydroxide (NaOH)	Chemical	Sigma–Aldrich	221465
Cimetidine	Drug	Sigma–Aldrich	C4522
Doxorubicin	Drug	Sigma–Aldrich	D1515
4–6-week-old female SCID mice	Mice	Charles River	Strain code 236
EMEM	Media	Sigma	M2279
F-12 Ham's	Media	Sigma	N3520
Sodium pyruvate	Reagent	Sigma	S8636
Lipoic acid	Reagent	Sigma	T1395
Glutamine	Reagent	Sigma	59202

Figure 4C: A549 xenograft tumor size		Normalized mean weight	SEM	SD	N
PBS	Day 1	1	0.25	0.61	6
	Day 2	1.28	0.25	0.61	6
	Day 3	0.98	0.34	0.83	6
	Day 4	1.35	0.24	0.59	6
	Day 5	1.28	0.26	0.64	6
	Day 6	1.39	0.24	0.59	6
	Day 7	1.63	0.25	0.61	6
	Day 8	1.98	0.24	0.59	6
	Day 9	2.98	0.25	0.61	6
	Day 10	2.83	0.31	0.76	6
	Day 11	3.3	0.23	0.56	6
Dox	Day 1	1	0.25	0.61	6
	Day 2	0.96	0.12	0.29	6
	Day 3	0.9	0.17	0.42	6
	Day 4	0.87	0.12	0.29	6
	Day 5	0.8	0.13	0.32	6
	Day 6	0.94	0.14	0.34	6
	Day 7	1.2	0.14	0.34	6
	Day 8	1.63	0.21	0.51	6
	Day 9	1.55	0.13	0.32	6
	Day 10	1.84	0.14	0.34	6
	Day 11	1.96	0.12	0.29	6
100 mg/kg cimetidine	Day 1	1	0.25	0.61	6
	Day 2	1.05	0.3	0.73	6
	Day 3	1.05	0.28	0.69	6
	Day 4	1.25	0.3	0.73	6
	Day 5	1.17	0.23	0.56	6
	Day 6	1.37	0.26	0.64	6
	Day 7	1.47	0.21	0.51	6
	Day 8	1.73	0.16	0.39	6
	Day 9	1.88	0.16	0.39	6
	Day 10	2.4	0.15	0.37	6
	Day 11	2.34	0.34	0.83	6

Stdev was calculated using formula SD = SEM*(SQRT n).

Supplemental Figure 1: ACHN xenograft tumor size		Normalized mean weight	SEM	SD	N
PBS	Day 1	1	0.09	0.22	6
	Day 2	1.37	0.09	0.22	6
	Day 3	1.39	0.09	0.22	6
	Day 4	1.45	0.09	0.22	6
	Day 5	1.39	0.09	0.22	6
	Day 6	1.52	0.08	0.20	6
	Day 7	1.64	0.09	0.22	6
	Day 8	1.84	0.09	0.22	6
	Day 9	1.67	0.13	0.32	6
	Day 10	1.92	0.08	0.20	6
	Day 11	2.14	0.09	0.22	6
100 mg/kg cimetidine	Day 1	1	0.2	0.49	6
	Day 2	1.26	0.14	0.34	6
	Day 3	1.23	0.11	0.27	6
	Day 4	1.1	0.1	0.24	6
	Day 5	1.23	0.11	0.27	6
	Day 6	1.34	0.09	0.22	6
	Day 7	1.18	0.07	0.17	6
	Day 8	1.34	0.09	0.22	6
	Day 9	1.7	0.1	0.24	6
	Day 10	1.7	0.08	0.20	6
	Day 11	2	0.1	0.24	6

Stdev was calculated using formula SD = SEM*(SQRT n).

ANOVA; all groups at day 11 time point
F (1,20) (interaction)	ηP2	Effect size f	Power	Total sample size across all groups
3.639500	0.153959	0.4265862	82.39%	48

The PBS control group SD was used as the divisor.

With a sample size of 12 per group derived from the ANOVA, achieved power will be at least 84.36%.