The Reproducibility Project: Cancer Biology seeks to address growing concerns about reproducibility in scientific research by conducting replications of 50 papers in the field of cancer biology published between 2010 and 2012. This Registered report describes the proposed replication plan of key experiments from ‘The CD47-signal regulatory protein alpha (SIRPa) interaction is a therapeutic target for human solid tumors’ by Willingham et al., 2012, published in PNAS in 2012. The key experiments being replicated are those reported in Figure 6A–C and Table S4. In these experiments, Willingham et al., 2012 test the safety and efficacy of anti-CD47 antibody treatment in immune competent mice utilizing a syngeneic model of mammary tumor growth in FVB mice. The Reproducibility Project: Cancer Biology is a collaboration between the Center for Open Science and Science Exchange, and the results of the replications will be published in eLife.

DOI: http://dx.doi.org/10.7554/eLife.04586.001

Replication Study

Original article

  1. The CD47-signal regulatory protein alpha (SIRPa) interaction is a therapeutic target for human solid tumors

    1. SB Willingham
    2. JP Volkmer
    3. AJ Gentles
    4. D Sahoo
    5. P Dalerba
    6. SS Mitra
    7. J Wang
    8. H Contreras-Trujillo
    9. R Martin
    10. JD Cohen
    11. et al.
    Proceedings of the National Academy of Sciences of USA 2012;109:6662-6667

Main text


Phagocytosis is an essential process utilized by an organism for pathogen or apoptotic cell clearance (Poon et al., 2014). CD47 is a cell surface glycoprotein with a variety of functions including regulation of phagocytosis through binding to the macrophage and dendritic cell specific protein signal regulatory protein alpha (SIRPα) (Oldenborg, 2013). Binding of SIRPα to CD47 essentially sends a ‘don't eat me’ message to macrophages by initiating signaling to inhibit phagocytosis (Murata et al., 2014).

Increased expression of CD47 is proposed to be a mechanism through which cancer cells evade immune detection and phagocytosis. CD47 expression is increased in several cancer types including acute myeloid leukemia (AML), acute lymphoblastic leukemia (ALL), non-Hodgkin lymphoma (NHL), primary effusion lymphoma, multiple myeloma, leiomyosarcoma, and bladder cancer, and targeting of CD47 on cancer cells with an anti-CD47 blocking antibody can promote phagocytosis by macrophages in vitro (Chan et al., 2009; Jaiswal et al., 2009; Majeti et al., 2009; Chao et al., 2010a; Edris et al., 2012). Further, treatment with an anti-CD47 blocking antibody synergized with rituximab treatment to promote phagocytosis in vitro and to eliminate cancer cells in an in vivo xenograft model of non-Hodgkin lymphoma (Chao et al., 2010b). This is supported in two syngeneic murine tumor models, in melanoma and squamous cell carcinoma, where irradiation combined with antisense suppression of CD47 delayed tumor growth (Maxhimer et al., 2009). Willingham et al., 2012 further extend these results to demonstrate that CD47 expression increases in a variety of human solid tumor types and that blocking the SIRPα/CD47 interaction with an anti-CD47 antibody can promote phagocytosis of solid tumor cells in vitro and reduce growth of solid tumors in vivo. While it is not clear if SIRPα signaling is involved in the antitumor activity of an anti-CD47 antibody, these results indicate that anti-CD47 antibody therapy may be an effective treatment for a variety of solid tumor types (Soto-Pantoja et al., 2012).

In Figures 6B, 6C, and Table S4, the safety and efficacy of anti-CD47 antibody treatment are tested in immune competent mice using a syngeneic breast cancer model. MT1A2 mouse mammary cancer cells were implanted in the mammary fat pads of FVB mice and IgG control or anti-CD47 antibody treatment commenced upon detection of palpable tumors. Tumor growth was measured by gross weight and analyzed by immunohistochemistry. Willingham et al., 2012 showed that anti-CD47 antibody treatment reduced tumor growth and increased lymphocytic infiltration to the tumor site without unacceptable toxicity, thus demonstrating that anti-CD47 therapy is effective in reducing solid tumor growth in immune competent hosts. This key experiment demonstrates that CD47 is a therapeutic target for solid tumors and follows similar reports from the same laboratory that also demonstrated that anti-CD47 antibody treatment reduced growth of primary human cancer xenografts of several hematopoietic cancers and of solid leiomyosarcoma tumors (Jaiswal et al., 2009; Majeti et al., 2009; Chao et al., 2010a; Edris et al., 2012). Subsequent reports extended these results to multiple myeloma and primary effusion lymphoma models (Kim et al., 2012; Goto et al., 2014). These experiments will be replicated in Protocol 1.

Materials and methods

Protocol 1: Engraftment of mouse breast cancer cells and treatment with targeted antibodies

This experiment tests the safety and efficacy of anti-CD47 antibody treatment in immune competent mice using a syngeneic model of mammary cancer. This experiment replicates figures 6B, 6C, and table S4 of the original article, which assess tumor growth by weight of the tumor 30 days after implantation, lymphocytic infiltration by immunohistochemistry, and toxicity by blood analysis.


  1. Experiment has 2 cohorts:

    • A. MT1A2 allograft treated with IgG isotype control.

    • B. MT1A2 allograft treated with anti-CD47 clone MIAP410.

  2. Experiment will use seven mice per treatment group.

    • A. To account for unexpected deaths, seven mice will be used per group to ensure at least five will survive for a minimum power of 80%.

    • B. A separate, untreated cohort of three mice will be used to gather baseline readings for the blood analysis.

      • I. See ‘Power calculations’ section for details.

Materials and reagents

All known differences are indicated by an asterisk, with the originally used item listed in the comments section.

ReagentTypeManufacturerCatalog #Comments
MT1A2Cell lineOriginal labn/aFrom original lab
Dulbecco's Modified Eagle's Medium, high glucose with HEPES modificationCell cultureSigma–AldrichD6171Included during communication with original authors. Original lab used Gibco catalog # 12430-054
L-glutamineCell cultureSigma–AldrichG7513
Fetal bovine serumCell cultureSigma–AldrichF0392
Penicillin/Streptomycin (100×)Cell cultureSigma–AldrichP433
Trypsin-EDTACell cultureSigma–AldrichT3924
Hank's balanced salt solutionCell cultureSigma–AldrichH6648
T75 flaskLabwareSigma–AldrichZ707503
50-ml tubesLabwareSigma–AldrichCLS430290
1 M Hepes in normal salineBufferBiowhittaker17-737EIncluded during communication with original authors
BSA, IgG freeChemicalJackson Immunoresearch001-000-161Included during communication with original authors.
Kolliphor P188ChemicalSigma–AldrichK4894Included during communication with original authors. Original lab used Pluronic F-68 which has been discontinued
Leibovitz L15 media, no phenol redCell cultureLife Technologies21083027Included during communication with original authors
Matrigel Matrix High Concentration*Cell cultureCorning354248Original from Becton Dickinson
6–8 week old female FVB miceAnimal modelCharles RiverStrain Code: 207Original from Jackson Labs
27½G needleLabwareSigma–AldrichZ192384Included during communication with original authors
1-ml syringeLabwareSigma–AldrichZ192090
30½G needleLabwareSigma–AldrichZ192341Included during communication with original authors. Original lab used 31 gauge
1–5% isofluraneChemicalSpecific brand information will be left up to the discretion of the replicating lab and recorded later
Anti-CD47 clone MIAP410AntibodyOriginal labn/aFrom original lab
Mouse IgG isotype controlAntibodyInnovative ResearchIR-MS-GF
Hematology analyzer*InstrumentIdexx LaboratoriesProCyte DxOriginal lab used a Heska HemaTrue
Neutral buffered formalinBufferSpecific brand information will be left up to the discretion of the replicating lab and recorded later
Carazzi's HematoxylinStain



  • A. All cells will be sent for mycoplasma testing and STR profiling.

  • B. Cells maintained in DMEM supplemented with 4 mM L-glutamine, 10% FBS, 100 U/ml penicillin and 100 µg/ml streptomycin at 37°C in a humidified atmosphere at 5% CO2.

  1. Culture MT1A2 cells, count cells, gently spin down, and resuspend in FACS buffer so that 50,000 cells can be injected per mouse.

    • A. Prepare cells in FACS buffer at 50,000 cells/75 µl.

    • B. FACS buffer (500 ml):

      • i. 5 ml 1 M Hepes, pH = 7.4.

      • ii. 500 mg BSA, IgG free.

      • iii. 680 mg Kolliphor P188.

      • iv. 5 ml 100X Pen/strep.

      • v. Bring up to 500 ml with Leibovitz L15 media, no phenol red.

  2. Add high protein matrigel to obtain 25% vol/vol solution.

    • A. Total volume/injection is 100 µl and 50,000 cells.

  3. Inject 50,000 cells into the left abdominal mammary fat pad (#4) of 6 to 8-week-old female FVB mice using a 27½ G needle.

    • A. Use 1–5% isoflurane at 1–2 l/min to anesthetize the mice.

    • B. Total number of mice injected is 14.

  4. Check mice until palpable tumors form in at least 12 animals.

    • A. Approximately 7–10 days after injection palpable tumors will arise.

  5. Randomize mice with palpable tumors to two treatment groups using the following method.

    • A. On the day the mice are randomized, measure tumors. Animals with no detectable tumors are excluded from the study.

    • B. Animals are ranked according to tumor size, to balance groups for baseline tumor characteristics, and assigned to group 1 or group 2 using an alternating serpentine method. (rank 1 = group 1, rank 2 = group 2, rank 3 = group 1, rank 4 = group 2, etc).

      • i. Designation of IgG or CD47 antibody treatment as group 1 or group 2 determined by randomly assigning the two treatments into one block using www.randomization.com. Record seed number.

  6. Inject 400 µg of antibody, administered in 100 µl PBS, into mammary fat pad proximal to tumor with an approximate distance of 2 mm to the tumor (do not inject directly into tumor), every other day for 30 days using a 30 G needle.

    • C. anti-CD47 clone MIAP410 (mouse IgG1).

    • D. mouse IgG isotype control.

  7. 5 days after the beginning of antibody injections perform complete blood cell counts to assess treatment toxicity with hematology analyzer.

    • A. Collect 0.2 ml blood by retro-orbital bleeding.

    • B. Collect 0.2 ml blood from three untreated female FVB mice to gather a baseline reading.

  8. After 30 days of antibody treatment, sacrifice mice, dissect, and weigh tumor.

  9. Dissected tumors are processed for further analysis.

    • A. Immediately, place tissues into 10% neutral buffered formalin overnight at room temperature.

    • B. Dehydrate tissues through graded alcohols (50%, 70%, 95% ethanol) and clear xylene.

    • C. Infiltrate with paraffin, and then embed tissues in a paraffin block.

    • D. Cut paraffin blocks on a microtome with a section thickness of 5 µm.

  10. Stain tumor sections with H&E (total: 2 stained sections per tumor).

    • A. Deparaffinize sections 2 times in xylene, then rehydrate through graded alcohols (95%, 70%, 50% ethanol) to water.

    • B. Stain sections with Carazzi's hematoxylin, then rinse slides in water.

    • C. Stain sections with eosin.

    • D. Dehydrate sections through graded alcohols (50%, 70%, 95% ethanol) and then place in xylene.

    • E. Apply coverslips to slides with Permount and store slides at room temperature.

  11. Blindly image stained sections and have images blindly analyzed by a Board Certified Pathologist to analyze lymphocytic infiltration of the tissue sections.

    • A. Assess absence or presence of tumor infiltrating lymphocytes in at least 10 random fields at high power magnification (×400) and score lymphocytic infiltration using the following system (Demaria et al., 2001):

      • i. 0 = absent.

      • ii. 1 = minimal.

      • iii. 2 = moderate.

      • iv. 3 = brisk.


  1. Data to be collected:

    • A. Mouse health records: general health, weight, and age at time of transplant, survival.

    • B. Lab records on time course of tumor formation (noting when tumors become palpable), transplants, and antibody injections (including seed number of randomization).

    • C. Image of each tumor at harvest.

    • D. Raw numbers and graph of tumor weight in control and anti-CD47-treated mice (compare to Figure 6B).

    • E. H&E stained sections from each tumor analyzed (compare to Figure 6C).

    • F. Pathology report for each section and tumor analyzed.

    • G. Complete blood cell counts in control and anti-CD47-treated mice.

Confirmatory analysis plan

This experiment assesses if treatment with an anti-CD47 therapeutic antibody alters breast tumor growth in immune competent hosts and examines the safety of the treatment by looking at blood toxicity. The histological analysis of lymphocytic infiltration of the tumors will be reported for each tumor generated during the study, along with the H&E stained sections. This replication attempt will perform the following statistical analysis.

  • A. Statistical analysis:

Note: At the time of analysis we will perform the Shapiro–Wilk test and generate a quantile–quantile plot to assess the normality of the data. We will also perform Levene's test to assess homoscedasticity. If the data appear skewed, we will perform an appropriate transformation in order to proceed with the proposed statistical analysis. If this is not possible, we will perform the equivalent non-parametric test.

  1. Tumor weight from isotype control treated mice to anti-CD47 clone MIAP410 treated mice.

    • i. Unpaired two-tailed Welch's t-test.

  • B. Hematological parameters (13 parameters) tested in untreated mice, IgG isotype control treated mice, and anti-CD47 clone MIAP410 treated mice.

    1. Two-way ANOVA (3 × 13 design) with the following planned comparisons with the Bonferroni correction:

      • i. One-way ANOVA of untreated, IgG, and anti-CD47-treated mice for each hematological parameter.

      • ii. Compare the effect size of the original data to the replication data 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 not include the anti-CD47 clone MIAP301. The replication attempt will analyze lymphocytic infiltration of the tumors using a scoring system of the H&E stained sections, which was not implemented by the original study, and is included as exploratory analysis. Toxicity will be assessed during the course of the efficacy experiment instead of on a different strain and cohort of mice as the original study, which was determined in BALB/c mice. Additionally, the replication study will analyze blood 5 days after the beginning of antibody treatment, which precedes a total of three antibody treatments at a dose of 400 µg/injection. This is similar to the original study, which analyzed blood 5 days after two successive daily antibody injections of 500 µg/injection. To determine the baseline reading, untreated female FVB mice will be also be analyzed. The Idexx Laboratories ProCyte Dx hematology analyzer will assess the same parameters as the Heska HemaTrue used in the original study. All known differences of materials and reagents are listed in the ‘Materials and reagents’ section above, indicated by an asterisk, with the originally used item listed in the comments section. All differences have the same capabilities as the original and are not expected to alter the experimental design.

The original study analyzed the data using a Student's t-test, however since the original data variance is not homogenous the Welch's t-test for comparing the samples will be used instead.

Provisions for quality control

The cell line used in this experiment will undergo STR profiling to confirm its identity and will be sent for mycoplasma testing to ensure there is no contamination, as well as rodent pathogen screening to ensure there are no detectable pathogens. The anti-CD47 clone MIAP410 will be checked to verify the specificity by ELISA, which is being conducted by the original lab. The retro-orbital bleeds will be performed by a Science Exchange lab with expertise in this technique to minimize stress. Subjective data collection will be performed blinded to the experimental conditions and treatment groups will be assigned in a random manner with the seed number recorded to reproduce the plan. All of the raw data, including the H&E stained sections, will be uploaded to the project page on the OSF (https://osf.io/9pbos) and made publically available.

Power calculations

Protocol 1
Summary of original data (provided by original authors).

Dataset being analyzedNMeanSD
Tumor weight of IgG-treated mice50.14450.05203
Tumor weight of anti-CD47 clone MIAP410 treated mice50.012240.002258
Test family

  • A. 2 tailed Welch's t test, difference between two independent means, alpha error = 0.05.

Power calculations (performed with R software, version 3.1.2) (R Core Team, 2014).

Group 1Group 2Effect size (Glass' Δ)*A priori powerGroup 1 sample sizeGroup 2 sample size
  • The IgG control group SD was used as the divisor.

Summary analysis of original data presented in Table S4 (Willingham et al., 2012):

Test family

  • A. 2-way ANOVA (3 treatments x 13 hematology parameters), Fixed effects, special, main effects, and interactions, alpha error of 0.05.

F(Dfn, Dfd)Partial η2Original effect size fReplication total sample sizeDetectable effect size f
F(24,39) = 0.8678 (interaction)0.3481200.7307699169*0.3895070
F(2,39) = 0.8075 (treatments)0.0397660.2035014169*0.2415459
F(12,39) = 187.6811 (hematology parameters)0.9829787.599178169*0.3331365
  • The replication sample size includes 13 parameters from 3 untreated, 5 IgG-treated, and 5 CD47-treated mice.

  • The original data did not detect a statistically significant interaction or treatment main effect, making these the detectable effect size with 80.0% power.

  • The original data reported a statistically significant effect for the hematology parameters, which the replication is powered to 99.9% to detect. This is the detectable effect size with 80% power.

Test family

  • A. ANOVA, Fixed effects, omnibus, one-way, Bonferroni's correction alpha error of 0.05/13 = 0.00385.

    • i. Power calculations performed with G*Power software, version 3.1.7 (Faul et al., 2007).

Hematology parameterF(Dfn, Dfd)Partial η2Original effect size fReplication total sample sizeDetectable effect size f
WBCF(2,3) = 0.49750.2490710.5759213*1.5234072
LymF(2,3) = 0.32970.1802090.46885313*1.5234072
MonoF(2,3) = 0.97810.3947090.807525813*1.5234072
GranF(2,3) = 1.07060.4164760.844822813*1.5234072
HCTF(2,3) = 3.76730.7152221.58477413*1.5234072
MCVF(2,3) = 58.27100.9749046.23273513*1.5234072
RDWaF(2,3) = 96.10000.9846318.00412713*1.5234072
HGBF(2,3) = 2.00360.5718641.15572813*1.5234072
MCHCF(2,3) = 83.14500.9822797.44514813*1.5234072
RBCF(2,3) = 2.97970.6651531.40941113*1.5234072
MCHF(2,3) = 1.37140.4776120.95618313*1.5234072
PLTF(2,3) = 0.85360.3626820.754370913*1.5234072
MPVF(2,3) = 1.92310.5617981.13227813*1.5234072
  • The replication sample size includes three untreated, five IgG-treated, and five CD47-treated mice.



The Reproducibility Project: Cancer Biology core team would like to thank the original authors, in particular Stephen Willingham and Jens-Peter Volkmer, for generously sharing critical information as well as reagents to ensure the fidelity and quality of this replication attempt, as well as Frank Graham and McMaster University for facilitating the transfer of MT1A2 cells. We thank Courtney Soderberg at the Center for Open Science for assistance with statistical analyses. We would also like to thank the following companies for generously donating reagents to the Reproducibility Project: Cancer Biology; American Type Culture Collection (ATCC), BioLegend, Cell Signaling Technology, Charles River Laboratories, Corning Incorporated, DDC Medical, EMD Millipore, Harlan Laboratories, LI-COR Biosciences, Mirus Bio, Novus Biologicals, Sigma–Aldrich, and System Biosciences (SBI).

Decision letter

Joan Massagué, Reviewing editor, Memorial Sloan-Kettering Cancer Center, United States

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: The CD47-signal regulated protein alpha interaction is a therapeutic target for human solid tumors” for consideration at eLife. Your article has been favorably evaluated by Tadatsugu Taniguchi (Senior editor), a Reviewing editor, and four 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. There was considerable discussion about the proposed replication. One of the reviewers asked for studies with human patient xenograft tumors to be included for a comprehensive and representative reproduction of the original study, but the other reviewers and the editors favor strict reproduction. The key experiment from the prior publication that suggests therapeutic potential for an anti-CD47-based strategy is the syngeneic experiment in an immune-competent host, where the potential toxicity of the antibody against the host could be assessed. This experiment seems to be the most critical to reproduce as it most closely mimics a therapeutic scenario. Therefore we support the original proposed design, with comments on the proposal and revision requests following.


Chroscinski et al. propose to replicate the experiments from Figure 6B, 6C, and Table S4 of Willingham et al, to verify the safety and efficacy of anti-CD47 antibody treatment in immune competent mice utilizing a syngeneic breast cancer model, and to assess the effects on tumor growth and lymphocytic infiltration. The tumor model consists of MT1A2 tumor cells injected into the mammary fat pad of FVB mice. The proposed protocol is detailed and carefully considered. Tumors will be weighed after 30 days, and embedded in paraffin and processed for histological analysis of lymphocytic infiltration. Toxicity will be determined five days after tumor cell injection by complete blood cell counts.

The authors appropriately discuss the relevant background for the study, and have also thoroughly considered all parameters of the proposed experiments, for which they have also consulted with the original authors. Author consultation allowed for the inclusion of key parameters including multiple details associated with proper media conditions for the tumor cells.

Notable differences between this proposed work and the previous study include the exclusion of anti-CD47 clone MIAP301, which is appropriate given the lack of a statistically significant effect of this clone in the previous study, and toxicity in this replicate study will be conducted on the same mice for which efficacy is also being considered, which appears to be preferable to the previous work that used a different mouse strain (BALB/c). Blood analysis will also be performed with the same dose of antibody used to determine efficacy, which also seems preferable. The number of mice is appropriate, as the authors seek to derive data from the same number of animals as in the original study (n=5 per group final). Finally the tumor cell line will also be subjected to STR profiling to verify its identity.

Overall the authors propose to replicate what amounts to the key findings of Willingham et al, in a carefully considered and complete proposal.

Revisions to the proposal and questions to consider:

1) Methods: In the original report two antibodies reactive against CD47, with different isotypes were used; one isotype control [not specified which) was used as a control. Why is only one antibody used in replicate here?

2) Methods: In an immunocompetent mouse, tumor can be cleared via CMC, ADCC, opsonization, phagocytosis, etc, mediated by a number of different effectors. In the proposed experiment, the test antibody might mediate these mechanisms in addition to blocking the CD47-SIRPa interaction, with a similar or partial therapeutic result. Using an isotype control, therefore, does not yield an interpretable result unless the isotype antibody binds to the same target cell (optimally CD47) with similar affinity, but does not block the interaction with SIRPa. Such a control antibody was used in some of the original experiments in vitro, (PNAS, Figure 3A-C), but apparently the antibody later became “unavailable” and hence was not included in subsequent in vitro or any animal experiments. Alternatively, use of the same CD47 antibody with a D265A Heavy Chain mutation to eliminate FcR binding would be a useful contriol. Without such controls, at the completion of this replicate, a conclusion that the mechanism is via blockade of the interaction of CD47 with SIRPa, (as in title) is not possible.

3) Methods: Injection of such large and frequent antibody doses, and directly into the tumor bed, is an unusual administration plan. Use of an IV or IP administration route would be advised in addition as a control group.

4) How will injections into the orthotropic site, but not the tumor itself be achieved? It is not clear what Willingham did from reading the paper.

5) Methods: Step 5, How are mice randomized? Please state method.

DOI: http://dx.doi.org/10.7554/eLife.04586.002

Author response