IL15 Agonists Overcome the Immunosuppressive Effects of MEK Inhibitors

Small-molecule inhibitors targeting oncogenic signaling pathways have emerged as a promising new class of drugs in cancer therapy. While these molecules undergo rigorous testing to demonstrate their efficacy in tumor models, their effects on the interplay between leukocytes and tumors have been largely unstudied because of the use of preclinical xenograft models that lack a competent immune system. T cells, crucial for controlling the growth of immunogenic tumors (1), rely upon many of the same signaling pathways targeted by pharmaceutical inhibitors for activation of cytotoxicity against tumor cells. For instance, engagement of the T-cell receptor (TCR) and costimulatory receptors activates the RAS–MAPK and PI3K–AKT signaling cascades, which are necessary for proliferation and effector function in T cells (2).

The FDA-approved small-molecule inhibitor of MEK, trametinib, presents an example of seemingly paradoxical interactions with host antitumor immunity. Studies have shown that proper T-cell activation and proliferation is impaired by pharmacologic inhibition of MEK signaling with trametinib (3, 4) and other compounds (5). These data seem to imply that trametinib would impair antitumor T-cell function in tumor-bearing hosts. However, it was recently found that although trametinib impairs T-cell function in vitro, it does not limit the effectiveness of either adoptive cell therapy (6) or checkpoint blockade with antibodies against PD-1, PD-L1, and CTLA-4 (7) in mouse models. A potential explanation is that trametinib in these studies was coadministered with high doses of IL2 (6) and checkpoint inhibitors (7) that enhance the activation of T cells, thus allowing them to overcome small-molecule inhibition of MEK. The possibility that cytokines can rescue some deleterious effects associated with MEK inhibition on T cells has yet to be explored. In addition, MEK inhibition of tumor cells can lead to increased expression of tumor-specific antigens (5, 8), potentially enhancing recognition of tumors by CD8⁺ effector T cells, which offers another explanation for the synergy with immunotherapies. At the moment, the mechanisms explaining the paradoxical effects of MEK inhibitors on T-cell–mediated antitumor immunity remain elusive.

In a rapidly changing clinical scenario, first-line PD-1 inhibitors, alone or in combination with CTLA-4 inhibitors, may become a standard of care in the near future against tumors such as melanoma (9). However, the combined effects and optimal sequencing of targeted therapies and immunotherapy remains unknown. Clarifying the immunosuppressive effects of targeted therapies in vivo is crucial for the design of synergistic combinatorial interventions with emerging immunotherapies. Trametinib was the first MEK inhibitor to be approved for clinical use in 2013, and it has demonstrated to improve overall survival in combination with other targeted interventions (10). To elucidate the effects of multiple targeted therapies on the tumor immunoenvironment and, subsequently, antitumor immunity, we analyzed a panel of molecules for their inhibitory activity on T cells. Our results indicate that most small-molecule inhibitors, and in particular trametinib, exert direct suppressive effects on human T cells in vitro and antitumor mouse T cells in vivo in preclinical cancer models. However, the suppressive effects of MEK inhibitors can be overcome by various cytokines. We found that clinically available IL15 agonists, through a mechanism dependent on the activation of PI3K, were particularly effective at rescuing T-cell function.

WT C57BL/6 and congenic Ly5.1 female 6- to 8-week-old mice were procured from the NCI or Charles River Laboratories. OT-I C57BL/6-Tg (TcraTcrb)1100Mjb/J transgenic mice were obtained from The Jackson Laboratory. Transgenic Kras^tm4Tyj and Trp53^tm1Brn mice (11, 12) were obtained from NCI Mouse Models of Human Cancers Consortium, brought to a full C57BL/6 background (13, 14). All animals were maintained in specific pathogen-free barrier facilities and used in accordance with the Institutional Animal Care and Use Committee of the Wistar Institute.

The Brpkp110 primary mammary tumor cell line was generated by culturing a mechanically dissociated C57BL/6 L-Stop-L-Kras^G12Dp53^flx/flxL-Stop-L-Myristoylated p110α−GFP^flx/+ primary breast tumor mass as described previously (15). Tumor cells were passaged a total of ten times and tested for mycoplasma before deriving the Brpkp110 cell line. Tumors were initiated by injecting 5 × 10⁵ cells into the axillary flanks. Tumor volume was calculated as: 0.5 × (L × W²), where L is the longer of the two measurements.

Peripheral blood lymphocytes were obtained by leukapheresis/elutriation and Miltenyi bead–purified. A2780 cells were obtained from AddexBio Technologies. ID8 cells (16) were provided by K. Roby (Department of Anatomy and Cell Biology, University of Kansas, Kansas City, KS) and retrovirally transduced to express Defb29 and Vegf-a (17) or OVA (18).

For human T-cell proliferation assays, K562 cells expressing human CD32, termed K32, were generated as described (19), γ-irradiated (100 Gy) and loaded with anti-CD3 (500 ng/mL, clone OKT3; eBioscience) plus anti-CD28 (500 ng/mL, clone 15E8; EMD Millipore) antibodies at room temperature for 10 minutes [artificial antigen-presenting cells (aAPCs)]. Peripheral blood mononuclear cells (PBMC) were labeled with Cell Trace Violet (Invitrogen) according to the manufacturer's instructions and cocultured with loaded aAPCs at a 10:1 PBMC:aAPC ratio or activated with Concanavalin A (ConA; 2 μg/mL, Sigma). Proliferation of T cells was determined 7 days later by FACS and division index was calculated using FlowJo software.

For mouse T-cell proliferation assays, pan-T cells were negatively purified from spleens with antibodies to B220 (RA3), Mac-1 (M170.13), and MHC-II (M5/114) using magnetic beads. T cells were labeled with Cell Trace Violet (Invitrogen) and stimulated with either agonistic CD3/CD28 beads (Dynabeads, Life Technologies) or tumor-pulsed bone marrow dendritic cells (BMDC) and analyzed for proliferation by FACS either 3 days (CD3/CD28 beads) or 7 days (BMDCs) later. Day 7 BMDCs were generated as described previously (20) and cultured overnight with double-irradiated (γ-irradiated, 100 Gy; and UV, 30 minutes) ID8-Defb29/Vegf-a cells. BMDCs were added to cultures of T cells at a 10:1 (T cell:BMDC) ratio. For recall ELISpot assays, mouse T cells were primed with tumor-pulsed BMDCs plus IL2 (30 U/mL) and IL7 (5 ng/mL), and restimulated 7 days later with fresh tumor-pulsed BMDCs at a 10:1 ratio in an IFN- ELISpot (eBioscience).

ALT-803 was generously provided by Altor BioScience Corporation and was diluted in sterile PBS for in vitro and in vivo studies. Recombinant human IL15 (Novoprotein), human IL2, human IL21, mouse IL7 (Peprotech), human IL27 (eBioscience), and ConA (Type VI, Sigma-Aldrich) were reconstituted in sterile PBS and stored at −20ºC. Trametinib (GSK-1120212) was purchased from LC Laboratories and suspended in vehicle solution of 10% PEG-300 (Sigma-Aldrich) and 10% Cremophor EL (EMD Millipore) in sterile dH₂0 for in vivo oral gavage experiments. For in vitro assays, all inhibitors were dissolved in sterile DMSO and diluted in the assays 1:1,000, so that the final concentration of DMSO was 0.1%.

Compound screening on A2780 cells was performed by adding compounds the morning after plating and measuring proliferation 72 hours later. Screening on human PBMCs was performed by adding compounds simultaneously with ConA stimulation (2 μg/mL) and measuring proliferation 7 days later. Normalized percent inhibition (NPI) was calculated by measuring resazurin fluorescence with respect to values obtained with DMSO-negative control and doxorubicin (5 μmol/L) positive control as NPI = 100% × (DMSO – compound)/(DMSO – doxorubicin).

Cells were lysed in RIPA buffer (Thermo Fisher Scientific) with Complete Protease Inhibitor Cocktail Tablets (Roche) and phosphatase inhibitors (Halt Phosphatase Inhibitor, Thermo Fisher Scientific, and Na₃VO₄, 1 mmol/L) and cleared by centrifugation. Proteins were quantified by BCA assay (Thermo Fisher Scientific), diluted in reducing Lamelli buffer, denatured at 95°C, run on mini Protean TGX Ready Gels (Bio-Rad Laboratories), transferred to a polyvinylidene difluoride membrane, blocked, and incubated with primary antibodies for p-ERK1/2 (D13.14.4E), p-AKT (D9E), and β-tubulin (9F3), all from Cell Signaling Technology; plus β-actin (Sigma; AC-15). Immunoreactive bands were developed using horseradish peroxidase–conjugated secondary antibodies (Bio-Rad) and ECL substrate (GE Healthcare).

CD8⁺ T cells were sorted from PBMCs and rested overnight in R10. T cells (0.5 × 10⁶ per condition) were stained with OKT3-biotin (BioLegend, 10 μg/mL) for 15 minutes on ice, and washed in cold PBS. TCR ligation was performed by adding streptavidin (Promega, 25 μg/mL) and anti-CD28 antibody (Millipore, clone 15E8, 1 μg/mL) in the presence of indicated inhibitors for 10 minutes at 37°C.

Congenic Ly5.1 mice were injected with 1.5 × 10⁶ ID8-OVA cells intraperitoneally (21). Mice were oral gavaged with trametinib or vehicle on days 9–13. On day 10, mice were injected intraperitoneally with 1.5 × 10⁶ Cell-Trace Violet labeled, unstimulated OT-I T cells. Mice were administered ALT-803 (0.2 mg/kg on day 10) or IL2 (50,000 IU/mouse on days 10–12) intraperitoneally On day 14, peritoneal washes were analyzed for proliferating OT-I T cells.

To determine the sensitivity of human T cells to inhibition of signaling pathways commonly targeted by small molecules in cancer therapy (22), we first designed a high-throughput assay to test a diverse panel of 41 inhibitors over a 6 log concentration range on ConA-induced activation and expansion of human T cells from PBMCs. At doses equivalent to or below those required to limit proliferation of A2780 ovarian cancer cells, known to be sensitive to PI3K and MEK inhibitors (23, 24), a variety of inhibitor classes prevented ConA-driven T-cell expansion (Fig. 1A). Small molecules targeting PI3K, mTOR, MAPK, and CDK signaling, as well as transcriptional regulators [histone deacetylase (HDAC)] and survival molecules (Bcl-2) were deleterious for T-cell expansion. Among these, trametinib, the MEK1/2 inhibitor approved by the FDA for BRAF-mutant melanoma, was particularly potent at inhibiting the in vitro proliferation of human T cells. Overall, the observed EC₅₀ of every tested molecule with some activity on A2780 cells was lower for human T cells than for A2780 cells (Table 1; Supplementary Fig. S1 and S2), highlighting the immunosuppressive effects of most small-molecule targeted therapies.

We validated our screening approach by focusing on inhibitors of the PI3K and MEK signaling pathways. Abrogation of T-cell activation elicited by small-molecule inhibitors was not restricted to ConA stimulation, because pan-PI3K (BKM120) and MEK (GDC0973) inhibitors also restricted the proliferation of human T cells in response to aAPC coated with agonistic CD3 and CD28 antibodies (Fig. 1B; ref. 19). Importantly, these effects were consistent among three different donors (Fig. 1C). Comparable results were obtained with the MEK inhibitor trametinib (Fig. 1D). As expected, T cells were more sensitive to kinase inhibitors when they were activated in the absence of costimulation (aAPCs lacking anti-CD28), as could occur within the immunosuppressive microenvironment of tumor-bearing hosts (Fig. 2A and B). We noticed only minor and nonsignificant differences in the repertoire of memory and effector T-cell subsets when cultures were activated without anti-CD28 and with lower amounts of anti-CD3, indicating that differences in sensitivity to the kinase inhibitors cannot be attributed to altered T-cell differentiation (Supplementary Fig. S3A–SD).

To determine the effects of MEK and pan-PI3K inhibitors on different T-cell subsets, we FACS purified human CD8 T cells into naïve, memory, and effector populations based upon CD45RA and CD27 expression and activated them with aAPCs. We found that trametinib equally inhibited proliferation of naïve (CD45RA⁺CD27⁺), memory (CD45RA⁻CD27⁺), and effector/effector memory (CD45RA⁻CD27⁻) cells, although one donor showed a trend that memory and effector/effector memory cells were less sensitive than naïve cells to trametinib (Fig. 3A and B). Interestingly, memory and effector/effector memory cells were more sensitive to PI3K inhibition with BKM120 than naïve cells, a result consistent among three donors (Fig. 3C and D). Differentiated effector cells (CD45RA⁺CD27⁻) did not proliferate in response to aAPCs, so we could not conclude their sensitivity to kinase inhibition from this analysis (Supplementary Fig. S4A).

We also explored the result of MEK and PI3K inhibition on physiologic activation of T cells with tumor antigens. MEK and PI3K inhibitors completely abrogated the initial priming response of murine T cells activated with tumor lysate-pulsed DCs (13, 20; Fig. 3E and F). More importantly, the direct suppressive effects of these inhibitors were not restricted to proliferative responses, because the frequency of tumor-primed T cells secreting IFNγ in response to restimulation with fresh tumor lysate–pulsed DCs was also significantly reduced when either PI3K or MEK were inhibited (Fig. 3G and H).

Overall, these data underscore the T-cell–suppressive effects of most small-molecule targeted therapies clinically approved or in the pipeline of clinical development, at a time when combinatorial immunotherapeutic interventions are being tested against multiple tumors.

Despite the strong T-cell inhibitory activity of FDA-approved trametinib on both initial priming and recall responses, recent reports suggest that trametinib does not limit the effectiveness of adoptive cell therapy in preclinical tumor models (6). Interestingly, these studies included the administration of high doses of IL2. We reasoned that cytokines signaling on immune (but not tumor) cells could rescue the deleterious effects of trametinib on T-cell activity. We therefore tested a panel of cytokines known to play a role in T-cell survival and proliferation for their ability to recover T-cell expansion from MEK inhibition. Supporting our hypothesis, IL2, IL7, and IL15 were able to individually rescue the proliferation of human T cells in the presence of trametinib (Fig. 4A and Supplementary Fig. S4B). In contrast, IL21 and IL27 had no significant rescuing effect (Fig. 4A and Supplementary Fig. S4B). Among all cytokines tested, we focused on IL15 because it provides a strong stimulating signaling to both effector and memory CD8⁺ T cells without inducing the expansion of Tregs, as compared with IL2 (25, 26). Indeed, IL15 was able to dramatically rescue the proliferation of purified CD8 naïve, memory, and effector/effector memory T cells (Fig. 4B).

We found that IL15 can rescue early (within 10 minutes postactivation) TCR-induced MAPK signaling from MEK inhibition, as shown by ERK1/2 phosphorylation (Fig. 4C). Mechanistically, this effect depends on activation of PI3K by IL15 (27), because IL15 was not able to rescue the defect in ERK1/2 phosphorylation in the presence of a pan-PI3K inhibitor (Fig. 4C). Furthermore, activation of protein kinase C (PKC) with phorbol-12-myristate-13-acetate (PMA) completely overcomes the suppressive effect of PI3K inhibition on ERK1/2 phosphorylation without fully restoring PI3K activity as assayed by AKT phosphorylation (Fig. 4C). This is consistent with a mechanism of ERK phosphorylation (28) mediated by the activation of PKC isoforms upon stimulation of the PI3K pathway, which is known to result in the production of phosphatidylinositol (3,4,5)-trisphosphate (PIP3) and, subsequently, activation of PDK1 kinase (29). Together, these results show that IL15 can augment early signaling events downstream of TCR activation to enhance the amplitude of MAPK signaling to overcome MEK1/2 inhibition by trametinib.

To define whether IL15 signaling can overcome MEK inhibition-induced suppression of T-cell activation in the tumor microenvironment in vivo, we investigated the IL15 superagonist ALT-803 (30) for its ability to rescue T-cell functions from MEK inhibition. ALT-803 is a new IL15 superagonist complex with IL15N72D bound to the IL15RαSu/Fc (30). This IL15N72D:IL15RαSu/Fc has a significantly longer serum half-life and increased biologic activity compared with native IL15 (30) and is undergoing extensive clinical testing (Supplementary Table S1). In this study, we found that ALT-803 was more effective than IL7 (Fig. 4D) and IL15 (Fig. 5A and B) in rescuing human T-cell proliferation in vitro. Human T-cell proliferation in the presence of ALT-803 remained unaffected by trametinib, even at the relatively high concentration of 5 μmol/L (Fig. 5B). As was the case for IL15, we found that ALT-803 activity was also dependent on PI3K signaling because T-cell expansion could not be fully rescued when pan-PI3K and MEK inhibitors were combined (Fig. 5B). We determined that the reduction in proliferation when pan-PI3K and MEK inhibitors were combined was due to a block in cell division rather than an increase in cell death (Supplementary Fig. S5A–S5E).

ALT-803 was also able to rescue the proliferation of purified human CD8⁺ naïve, memory, and effector/effector memory T cells from trametinib, indicating broad activity on a range of T-cell subsets (Supplementary Fig. S6A). Intriguingly, ALT-803 (and IL15 for one donor) induced the proliferation of purified, differentiated effector (CD45RA⁺CD27⁻) CD8⁺ T cells that were otherwise unable to proliferate in response to CD3/CD28 activation (Supplementary Fig. S4A). ALT-803 showed activity on mouse T cells as well, demonstrated by its ability to restore proliferation of bead-activated T cells (Fig. 5C) and IFNγ recall responses of tumor-primed T cells in the presence of trametinib (Fig. 5D).

To test the activity of ALT-803 in the tumor microenvironment, we transferred Cell Trace Violet-labeled [Ovalbumin (OVA)-specific] OT-I T cells into mice growing OVA-transduced syngeneic ID8 ovarian tumors, a system that allows the recovery of tumor microenvironment lymphocytes through peritoneal wash (17, 18, 31). After 4 days, we found that in mice treated with trametinib, the OT-I T cells proliferated significantly less than in mice gavaged with vehicle (Fig. 5E and F). Importantly, when a single dose of ALT-803 was coadministered with OT-I T cells, proliferation was dramatically enhanced and was not restricted by trametinib (Fig. 5E and F). The observation that ALT-803 induces proliferation of some CD44^lo cells in addition to CD44^hi (antigen-experienced) OT-I T cells suggests that ALT-803 may also result in homeostatic proliferation, as has been reported previously (32). We also found that treatment of mice with a high-dose IL2 regimen was able to rescue OT-I T-cell proliferation (Supplementary Fig. S6B and S6C). These results indicate, first, that trametinib impairs antigen-specific T-cell responses in vivo, although to a lesser degree than in vitro. And second, that therapeutic activation of IL15 or IL2 signaling can completely overcome trametinib-induced CD8⁺ T-cell suppression in the tumor microenvironment.

To investigate the therapeutic potential of combining trametinib and ALT-803 against established KRas-mutated tumors, we utilized a syngeneic tumor model derived from an autochthonous breast cancer initiated in triple transgenic (L-Stop-L-KRas^G12Dp53^flx/flxL-Stop-L-Myristoylated p110α) mice with adenovirus-Cre (11, 12, 15). We chose this cell line, termed breast-p53-KRas-p110alpha (Brpkp110), to model treatment against tumors that evade single-molecule targeting of the MAPK pathway through PI3K activation, as has been commonly reported in human cancer cells (33–35). Brpkp110 cells have detectable signaling through MEK that can be inhibited with trametinib (Fig. 6A) and generate aggressive tumors when grown subcutaneously in mice.

Oral gavage with trametinib significantly reduced the growth of Brpkp110 tumors as a single intervention (Fig. 6B), although all mice eventually progressed to terminal disease (Fig. 6B–D). In contrast, when ALT-803 was combined with trametinib treatment, Brpkp110 tumors progressed even slower, with 19% of mice remaining tumor free at 50 days, and 15% exhibiting complete regression (Fig. 6D). No mice in either single treatment group remained tumor free after 50 days. Most importantly, mice that recovered from tumor challenge with trametinib/ALT-803 combination treatment developed immunologic memory against the tumor because these mice were resistant to subsequent rechallenge with Brpkp110 cells in the opposite flank over 30 days after initial tumor rejection, whereas all naïve control mice developed tumors (Fig. 6E). These results indicate that a targeted-immunotherapy combination with ALT-803 and trametinib could provide potent antitumor activity against some established and aggressive tumors and elicit protective immunity for tumor recurrences.

Immunotherapies are revolutionizing cancer treatments (36). In addition, the effects of existing chemotherapy on immune cells and the tumor microenvironment have recently become better appreciated (37). Here we show that many small molecules used as targeted therapies have significant immunosuppressive effects in vitro by interfering with signaling pathways that are critical for priming the responses of T cells. Nevertheless, MEK inhibitor-induced suppression can be rescued by signaling from some common gamma chain family cytokines, of which the most effective is the IL15 agonist, ALT-803, as shown in this study.

One of the most immunosuppressive drugs ex vivo in our hands was the FDA-approved MEK inhibitor trametinib. These data are in agreement with previous reports identifying a strong inhibitory activity for this drug in vitro (3, 4). However, recent studies indicate that trametinib does not limit the effectiveness of adoptive cell therapy in vivo (6) and synergizes with PD-1 inhibitors (7). Our study provides a mechanistic rationale to reconcile this paradox by demonstrating that direct T-cell inhibition by trametinib, while still detectable in vivo in tumor-bearing hosts, can be effectively overcome by cytokines such as IL2, IL7, and IL15. Therefore, a combination of these cytokines endogenously in tumor-bearing hosts may explain our observed lower suppressive effects of trametinib in vivo as compared with in vitro. In addition, the study showing that trametinib synergized with an adoptive transfer immunotherapy (ACT) in a mouse BRAF-driven tumor model included high-dose IL2 treatment (6), a regimen that was capable of rescuing T-cell proliferation from trametinib in our in vivo experiments. Our results suggest that potential trametinib-driven T-cell inhibition in ACT models may be offset by IL2 administration. Nevertheless, as MEK inhibitors are being clinically tested against multiple KRas-driven malignancies (38–40), our data provide an actionable approach to effectively overcome any direct T-cell inhibitory effects in future combinations of trametinib and emerging immune therapies, through the use of T-cell–rescuing agonists.

Our study indicates that ALT-803, an IL15 superagonist complex (30), could be the agent of choice for such a combination therapy. ALT-803 induces memory CD8⁺ T cells to proliferate, upregulate receptors for innate immunity, secrete IFNγ, and acquire the ability to kill tumor cells (30, 41, 42). Stimulation of NK cells by ALT-803 can also contribute to enhanced antitumor immunity. ALT-803 exhibits more potent antitumor activities against tumors than recombinant IL15 in various animal models (30), likely due to its much stronger binding capability to IL2Rβγ displayed on T and NK cells, longer serum half-life, and better biologic distribution to and detainment in lymphoid tissues. Currently, ALT-803 is in multiple clinical trials against solid and hematologic tumors either as a single agent or in combination with other FDA-approved immunostimulators or therapeutic antibodies (Supplementary Table S1).

As most tumor-reactive T cells in cancer patients are unlikely to be naïve T cells, it is important to understand the outcome of kinase inhibition on memory and effector T cells. Sensitivity to MEK inhibition has been reported to correlate negatively with T-cell differentiation, suggesting that naïve T cells require more MEK signaling than memory and effector cells (43). We observed this trend in one of three healthy donors, whereas two others showed similar sensitivities between memory, effector memory, and naïve cells. IL15 agonists were remarkably successful at reversing the block in proliferation by trametinib on naïve, memory, and effector memory T cells. In addition, murine tumor-reactive effector T cells were inhibited by trametinib, but ALT-803 was also able to rescue this activity.

Importantly, we found that IL15 signaling rescues MEK inhibition-induced T-cell suppression through the activation of the PI3K pathway. The synergy between PI3K and MEK inhibitors on tumor cells, (44) therefore, also exists for human T cells. These data suggest that combinatorial therapies in patients may compromise antitumor immunity, especially considering that combination treatment prevents the rescue by IL15 agonists in vitro. Because treatments combining PI3K and MEK inhibitors are being investigated clinically (39), it is important that future studies determine whether this combination negatively impacts the antitumor activity of T cells in tumor-bearing hosts and thus represents a poor option for targeted-immunotherapy combination.

Overall, our findings illustrate that a greater understanding of how targeted small molecules impact the host's immune system and tumor microenvironment could lead to more effective therapies against cancer and provide a novel intervention that could pave the way for combining targeted anticancer and immune therapies.

E.K. Jeng has ownership interest (including patents) in Altor BioScience. H.C. Wong has ownership interest (including patents) in Altor BioScience Corporation. No potential conflicts of interest were disclosed by the other authors.

Conception and design: M.J. Allegrezza, T.L. Stephen, N. Svoronos, H.C. Wong, S.N. Fiering, J.R. Conejo-Garcia

Development of methodology: M.J. Allegrezza, M.R. Rutkowski, T.L. Stephen

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M.J. Allegrezza, M.R. Rutkowski, N. Svoronos, A.J. Tesone, A. Perales-Puchalt, J.M. Nguyen

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.J. Allegrezza, M.R. Rutkowski, T.L. Stephen, A.J. Tesone, A. Perales-Puchalt, J. Tchou, H.C. Wong, J.R. Conejo-Garcia

Writing, review, and/or revision of the manuscript: M.J. Allegrezza, M.R. Rutkowski, N. Svoronos, A.J. Tesone, A. Perales-Puchalt, M.R. Sheen, E.K. Jeng, H.C. Wong, S.N. Fiering, J.R. Conejo-Garcia

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M.J. Allegrezza, J.M. Nguyen, F. Sarmin

Study supervision: J.R. Conejo-Garcia

The authors thank David Schultz for support with molecular screening and the Wistar Flow Cytometry and Imaging facilities for technical assistance.

This study was supported by R01CA157664, R01CA124515, R01CA178687, U54CA151662, P30CA10815, The Jayne Koskinas & Ted Giovanis Breast Cancer Research Consortium at Wistar, and Ovarian Cancer Research Fund Program Project Development awards (J.R. Conejo-Garcia). M.J. Allegrezza and N. Svoronos were supported by T32CA009171. A. Perales-Puchalt was supported by the Ann Schreiber Mentored Investigator Award (OCRF).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.