PARP1 Inhibition Radiosensitizes Models of Inflammatory Breast Cancer to Ionizing Radiation Free

This article is featured in Highlights of This Issue, p. 1897

Inflammatory breast cancer (IBC) diagnoses represent well under 5% of new breast cancer cases but account for a disproportionate share of breast cancer mortality (1). Despite aggressive, multimodal therapy, patients have high rates of locoregional recurrence and distant metastases (1). Treatment strategies for many breast cancer subtypes are largely directed against the protein drivers of each molecular subtype, including targeted therapies against the estrogen receptor (ER) or the HER2. IBC, however, represents a heterogeneous population that includes tumors across all of the molecular subtypes (2). Current treatment guidelines for patients with IBC take into consideration the molecular subtype of the tumor and include anti-HER2 or anti-estrogen therapy when appropriate, but more effective and targeted therapeutic options for patients with IBC are extremely limited. Without more effective alternatives, patients with BC typically receive neoadjuvant chemotherapy followed by mastectomy and adjuvant radiation (RT) to the chest wall and regional lymphatics (1). The key molecular drivers of IBC are currently unknown, and this uncertainty manifests as ineffective clinical therapeutic strategies. In IBC, there is a critical need to identify more effective treatment strategies to decrease rates of locoregional recurrence.

In an attempt to understand the heterogeneity of IBC, a recent study of 53 IBC tumors demonstrated that over 90% of tumors studied contained actionable mutations in genes like PIK3CA and BRCA1/2 that could be targeted using therapies that are either FDA-approved or currently in clinical trial (3). In line with this finding, there are a number of phase I and phase II clinical trials seeking to repurpose other FDA-approved drugs for indication in IBC (1). Targeted therapies in these trials include agents against PD-1 (pembrolizumab), VEGF-A (bevacizumab; refs. 4, 5), JAK1/2 (ruxolitinib), and the viral agent T-VEC (talimogene laherparepvec; ref. 1). Many different chemotherapy and radiation therapy regimens have been explored in IBC, but rates of recurrence and overall survival have not significantly improved (6). However, the ability to sensitize IBC tumors to current treatments like radiation represents a promising treatment strategy for patients with IBC.

Inhibition of PARP1 has been explored in clinical trials for many cancer types. PARP1 inhibition (PARPi) does not demonstrate significant single-agent efficacy in the treatment of most breast cancers (7, 8); however, PARPi is an effective targeted therapy in subsets of patients harboring BRCA1/2 mutations (9). In addition to the use of PARP1 inhibitors as monotherapy, our group has shown previously that PARPi can effectively radiosensitize a large range of breast cancer cell lines, including those with functional BRCA1 and BRCA2 (10). PARP1, through the addition of poly-ADP ribose (PAR) moieties to sites of single-strand DNA (ssDNA) damage, plays a critical role in recognition and recruitment of DNA repair machinery for a variety of different DNA repair processes. If ssDNA lesions go unrepaired, double-strand DNA (dsDNA) breaks form. For cells with intact repair pathways, nonhomologous end joining (NHEJ) or homologous recombination (HR) allows the cell to repair DNA. In the case of cancers with BRCA1/2 mutations, where BRCA-mediated homologous recombination is already deficient, the use of PARP1 inhibitors alone can promote the lethal accumulation of dsDNA breaks, leading to selective death of tumor cells—a concept referred to as synthetic lethality. In cells with wild type BRCA, other deficiencies in DNA repair pathways—and the addition of PARPi—may predispose tumor cells to higher levels of DNA damage caused by therapeutic radiation (10). To that end, this study aimed to determine the effect and efficacy of combining PARPi and radiation in multiple preclinical models of IBC.

All IBC cell lines were grown in HAMS F12 media (Gibco; 11765-054) in a 5% CO₂ incubator. Media for SUM-149 cells was supplemented with 5% FBS (Atlanta Biologicals), 10 mmol/L HEPES (Thermo Fisher Scientific; 15630080), 1× antibiotic-antimycotic (anti-anti, Thermo Fisher Scientific; 15240062), 1 μg/mL hydrocortisone (Sigma; H4001), and 5 μg/mL insulin (Sigma; I9278). SUM-190 media was supplemented with 1% FBS, 1 μg/mL hydrocortisone, 5 μg/mL insulin (Sigma; I0516), 50 nmol/L sodium selenite (Sigma; S9133), 5 μg/mL apo-Transferrin (Sigma; T-8158), 10 nmol/L triiodo thyronine (T3, Sigma; T5516), 10 mmol/L HEPES, and 0.03% ethanolamine (Sigma; 411000). MDA-IBC-3 cells were grown with 10% FBS, 1 μg/mL hydrocortisone, 1× anti-anti, and 5μg/mL insulin (Sigma; I0516). SUM cell lines were obtained from Stephen Ethier at the Medical University of South Carolina, and MDA-IBC-3 cells were obtained directly from Wendy Woodward at the University of Texas MD Anderson Cancer Center. All cell lines were routinely tested for mycoplasma contamination (Lonza; LT07-418) and were authenticated using fragment analysis at the University of Michigan DNA sequencing core. Olaparib (MedChem Express; HY-10162) and veliparib (MedChem Express; HY-10129) were reconstituted in 100% DMSO for cellular assays.

SUM-190 and SUM-149 cells were plated in 96-well plates overnight and treated the next morning with either olaparib or veliparib using a dose range of 1 pmol/L to 10 μmol/L. After 72 hours, AlamarBlue (Thermo Fisher Scientific; DAL1025) was added up to 10% of the final volume and read on a microplate reader after incubation at 37°C for 3 hours. MDA-IBC-3 cells were plated in 6-well plates and treated with a dose range of 1 nmol/L to 10 μmol/L of either olaparib or veliparib. After 72 hours, cells were trypsinized and counted with a hemocytometer.

SUM-149 and SUM-190 cells were plated at various densities from single cell suspension in 6-well plates and radiated the following day after a 1-hour pretreatment with olaparib. Cells were grown for up to 3 weeks, then fixed with methanol/acetic acid and stained with 1% crystal violet. Colonies with a minimum of 50 cells were counted for each treatment condition. Plating efficiency was determined and used to calculate toxicity. Cell survival curves were calculated as described previously (10). MDA-IBC-3 cells were grown in soft agar (Thermo Fisher Scientific; 214050) with a base layer of 0.5% agar solution and a top layer of 0.4% agar containing the cell suspension. Drug treatments in supernatant media were added fresh each week. Colonies were grown for up to 4 weeks before staining with 0.005% crystal violet.

Cells were plated on 18 mm coverslips in 12-well plates and allowed to adhere to coverslips overnight. The following day, cells were treated with media containing either olaparib or vehicle 1 hour before radiation (2 Gy), and coverslips were fixed at predetermined time points after radiation. γH2AX foci were detected using anti-phospho-histone H2AX (ser139) monoclonal antibody (Millipore; 05-636), with a goat anti-mouse fluorescent secondary antibody (Invitrogen; A11005). At least 100 cells were scored visually for γH2AX foci in 3 independent experiments. Cells containing ≥15 γH2AX foci were scored positive and were pooled for statistical analyses.

Cells were plated overnight and pretreated the next morning with olaparib. Plates were irradiated 1 hour after pretreatment, and cells were harvested at 6 and 24 hours after radiation. Lysates were extracted using RIPA buffer (Thermo Fisher Scientific; 89901) containing protease and phosphatase inhibitors (Sigma-Aldrich; PHOSS-RO, CO-RO). Proteins were detected using the anti-PAR antibody (LS-B12794; 1:5,000), the anti-PARP1 antibody (ab6079; 1:1,000), and anti-β-Actin (8H10D10; Cell Signaling Technology, 12262S, 1:50,000).

Bilateral subcutaneous flank injections were performed on 4- to 6-week-old CB17-SCID female mice with 1 × 10⁶ SUM-190 cells resuspended in 100 μL PBS with 50% Matrigel (Thermo Fisher Scientific; CB-40234). Tumors were allowed to grow until reaching approximately 80 mm³. Olaparib treatment was given by intraperitoneal injection 24 hours prior to the first radiation treatment. For long-term studies, mice were treated with vehicle (10% 2-hydroxylpropyl-beta-cyclodextrin in phosphate buffered saline; Thermo Fisher Scientific; 10010-023), olaparib (50 mg/kg) alone, radiation alone (2 Gy × 8 fractions) or the combination of olaparib + RT, with 16 to 20 tumors per treatment group. Tumor growth was measured 3 times a week using digital calipers, and mice were weighed on the same days. Tumor volume was calculated using the equation V = (L × W²) × π/6. For short-term studies, mice were treated with vehicle control, olaparib, or radiation for 48 hours before the tumors were harvested. Mice treated with both olaparib and radiation received olaparib treatment 24 hours before radiation treatment. The tumors were then harvested 48 hours after radiation. IHC staining was performed on tumors for all 4 conditions. All procedures involving mice were approved by the Institutional Animal Care & Use Committee (IACUC) at the University of Michigan and conform to their relevant regulatory standards.

Irradiation was carried out using a Philips RT250 (Kimtron Medical) at a dose rate of approximately 2 Gy/min in the University of Michigan Experimental Irradiation Core as described previously (10). Irradiation of mouse tumors was carried out as described previously (11).

IHC staining was performed on the DAKO Autostainer (Agilent) using Envision+ or liquid streptavidin–biotin and diaminobenzadine (DAB) as the chromogen. De-paraffinized sections were labeled with the antibodies listed in Supplementary Table S1 for 30 minutes at ambient temperature. Microwave epitope retrieval, as specified in Supplementary Table S1, was used prior to staining for all antibodies. Appropriate negative (no primary antibody) and positive controls (as listed in Supplementary Table S1) were stained in parallel with each set of slides studied. Whole-slide digital images were generated using an Aperio AT2 scanner (Leica Biosystems Imaging; Vista) at ×20 magnification, with a resolution of 0.5 μm per pixel. The scanner uses a ×20/0.75 NA objective and an LED light source. The same instrument and settings were used throughout the study for all whole-slide images generated. The images were checked for quality before use, and scans were repeated as necessary. Digital slides were analyzed using the Visopharm image analysis software suite (DK-2970 Hoersholm; v2019.2) to count stained and unstained nuclei.

Cells were plated in 6-well plates and allowed to adhere overnight. Cells were pretreated with olaparib for 1 hour before radiation and collected at designated time points after radiation. Cells were mixed with low melting point agarose (Thermo Fisher Scientific; 15-455-200) and spread on CometSlides (Trevigen; 4250-050-03). The cells were lysed with lysis solution (Trevigen; 4250-050-01), and DNA was separated by electrophoresis. Propidium iodide (Thermo Fisher Scientific; P3566) was used to stain DNA. A fluorescent microscope was used to take images of at least 50 cells/treatment. Images were analyzed using Comet Assay IV Software Version 4.3 to calculate the Olive tail moment. Results were pooled for statistical analyses.

GraphPad Prism 7.0 was used to perform statistical tests. In vitro statistical analyses were performed using the 2-tailed Student's t-test or a 1-way ANOVA in the case of multiple comparisons. For in vivo studies, a 2-way ANOVA was used to compare tumor growth, and the fractional tumor volume (FTV) method for assessing synergy in vivo was used as described previously (12, 13).

First, we sought to characterize the effect of two PARP1 inhibitors, olaparib (AZD2281) and veliparib (ABT-888) (14), on the proliferation of IBC cell lines. In SUM-190 and MDA-IBC-3 cells, single agent PARPi with olaparib or veliparb does not cause a significant decrease in proliferation at concentrations up to 10 μmol/L (Fig. 1A–D). While veliparib does not appear to impact proliferation of SUM-149 cells (IC₅₀ > 10 μmol/L, Fig. 1E), olaparib does have a modest effect as a single agent in SUM-149 cells (IC₅₀ = 2.2 μmol/L, Fig. 1F). All models tested were isolated from patients with IBC; however, SUM-149 cells are unique as they harbor a BRCA1 2288delT mutation as well as allelic loss of the wild type BRCA1 gene, rendering them BRCA1 deficient (15). Thus, SUM-149 cells may be especially sensitive to additional inhibition of DNA repair pathways (15).

While single agent PARPi with either olaparib or veliparib did not inhibit cell proliferation, we sought to determine the effect of PARPi on the radiosensitivity of IBC cell lines. Clonogenic survival assays were performed with olaparib in each of the 3 IBC cell lines; compared to veliparib, olaparib is a more potent PARP1inhibitor, with both PARP1 enzymatic inhibition efficacy and PARP trapping function. All IBC cell lines displayed significant radiosensitization as a result of pretreatment with olaparib. In SUM-190 cells, a dose-dependent radiosensitization was observed, with average radiation enhancement ratios (rER) of 1.45 ± 0.03 and 1.64 ± 0.21 at concentrations of 1 μmol/L and 2 μmol/L olaparib, respectively (Fig. 2A; Supplementary Fig. S1A). A similar trend was observed in MDA-IBC-3 cells, with enhancement ratios of 1.12 ± 0.08 and 1.28 ± 0.06 under the same treatment conditions (Fig. 2C, Supplementary Fig. S1B). Because SUM-149 cells express a truncated form of the BRCA1 protein, treatment with olaparib leads to marked radiosensitization at much lower doses. At 10 nmol/L and 20 nmol/L, the average enhancement ratios for SUM-149 cells were approximately 1.42 ± 0.01 and 1.76 ± 0.11 (Fig. 2E, Supplementary Fig. S1C). The enhancement ratios observed here are similar to or greater than that of cisplatin (rER = 1.2–1.3), a compound well-characterized for its ability to act as a radiosensitizing agent (16, 17). Furthermore, the surviving fraction of cells at 6 Gy (Fig. 2B, D and F) was significantly lower across all 3 inflammatory cell lines with the addition of olaparib. The radiation enhancement ratios demonstrated a marked dose-dependent increase, while toxicity from each treatment was minimal (Supplementary Fig. S1).

In cancer cells, ionizing radiation induces both single strand and double strand DNA breaks. In situations where DNA repair is inhibited and single strand breaks go unrepaired, the collapse of replication forks can propagate chromosomal damage and lead to the accumulation of lethal dsDNA breaks. Because PARP1 is involved in the recruitment of DNA repair proteins to DNA strand breaks, we sought to understand the effect of PARPi and radiation on the accumulation of DNA damage in IBC cell lines. In SUM-190 and SUM-149 cells, radiation treatment alone (2 Gy) induces γH2AX foci in greater than 75% of cells (Fig. 3A and B). In both cell lines, dsDNA breaks are retained at significantly higher levels at 12 and 16 hours after treatment with olaparib and radiation compared with treatment with radiation alone (Fig. 3C and D). Furthermore, a similar difference in dsDNA breaks was observed between RT alone and combination treatment in SUM-190 cells at 4 hours after radiation. In short, the presence of olaparib leads to the accumulation and persistence of dsDNA breaks in the combination treatment compared with the radiation treatment alone in both SUM-190 and SUM-149 cells. In order to independently confirm these findings, we performed the neutral comet assay to assess for dsDNA breaks (Fig. 4A). In SUM-190 cells, the combination of PARPi and RT in SUM-190 cells lead to a significantly longer tail moment, indicating increased dsDNA breaks compared with treatment with RT alone (P = 0.029). The tail moment was also significantly higher in combination treated cells compared with cells treated with vehicle or olaparib as a single agent (Fig. 4A). Representative images for each treatment condition are shown (Fig. 4A).

In order to determine if inhibition of PARP1 enzymatic activity occurs at concentrations of olaparib that are sufficient to induce radiosensitization, we treated cells with olaparib ± 4 Gy radiation and measured the total PAR and PARP1 levels in IBC cell lines. In SUM-190 and MDA-IBC-3 cells, PAR formation is significantly inhibited with 1 μmol/L of olaparib (Fig. 4B and C). The same effect is seen in SUM-149 cells after treatment with 1 μmol/L olaparib (Supplementary Fig. S2). Inhibition of PAR formation, however, can also be achieved at the same level in SUM-149 cells with 20 nmol/L of olaparib (Fig. 4D). Therefore, inhibition of PAR formation with olaparib occurs at low concentrations (20 nmol/L) that are sufficient to confer radiosensitization in SUM-149 cells. Though olaparib effectively inhibits PARylation at these concentrations, the amount of PARP1 in the cell lines remains relatively constant in all models (Fig. 4B–D).

Having demonstrated that PARPi can effectively radiosensitize IBC cell lines in vitro, we next sought to validate these findings in an in vivo xenograft model. For in vivo studies, subcutaneous tumors were allowed to reach ∼80 mm³ in CB-17 SCID mice whereupon treatment was initiated with one of the following: vehicle, 50 mg/kg olaparib alone daily, radiation alone (8 fractions of 2 Gy), or the combination (olaparib 50 mg/kg + 2 Gy RT daily for 8 fractions; Fig. 5A). To truly assess the radiosensitizing effects of PARPi, olaparib treatment was started 1 day before initiation of radiation and discontinued after the last fraction of radiation.

Consistent with the in vitro proliferation assays, treatment with olaparib alone did not significantly delay tumor growth or doubling time of xenograft tumors. As expected, radiation alone did lead to a decrease in tumor size initially, but tumors continued to grow after the completion of fractionated radiation (Fig. 5B). Mice receiving both radiation and olaparib treatment had significantly smaller tumors after completion of the study compared with those receiving radiation alone (P < 0.0001). There was a significant delay in the time to tumor doubling (P < 0.0001, Fig. 5C) and tripling (P < 0.0001, Fig. 5D) in the animals treated with combination olaparib and RT. In addition, time to tumor doubling and tripling was not reached in the combination treated group after 35 days. Weights of the mice (Fig. 5E) remained relatively constant throughout the experiment, indicating there was limited toxicity observed with combination treatment. Interestingly, the effects of the combination treatment with olaparib and radiation were found to be synergistic using the FTV method as described previously (Fig. 5F; ref. 12). IHC studies in tumors harvested from the mice at the end of the experiment demonstrated that levels of Ki67, a marker of cell proliferation, were significantly decreased in all treatment groups compared with control mice, with the most significant decrease in the combination treated animals (P = 0.0004; Supplementary Fig. S3A and S3B). There was also a decrease in p16 staining levels in the mice treated with radiation alone (P = 0.0072) and the combination treated group (P = 0.039) in the long-term experiments (Supplementary Fig. S3C and S3D), suggesting a decrease in cellular senescence in these tumors (18). The on-target effects of olaparib were confirmed in the short-term studies (48 hours of PARPi treatment alone or 24 hours of PARPi pretreatment before radiation). As expected, total levels of PARP1 were unaffected by treatment with olaparib, radiation, or the combination treatment (Supplementary Fig. S4A and S4B), whereas PAR levels were significantly lower in the PARPi treated animals (Supplementary Fig. S4C and S4D).

In this study, we demonstrate that PARPi alone is insufficient in delaying IBC cell line growth and proliferation (Fig. 1). Combination treatment with PARPi and ionizing radiation, however, results in significant radiosensitization of IBC models in vitro (Fig. 2), and the combination treatment results in delayed tumor growth in vivo (Fig. 5). Additionally, we demonstrate that PARPi in combination with radiation significantly delays resolution of dsDNA breaks using in vitro models of IBC (Fig. 3 and 4). Taken together, these results suggest that PARPi with radiation therapy may be a promising strategy for the treatment of IBC.

Although these studies suggest that PARPi may be an effective radiosensitization strategy for the treatment of IBC, other potential targets for treatment have also been identified. Several groups have identified molecular alterations in IBC tumors and in vitro models that may help to describe the aggressive phenotype associated with IBC (1). Owing to the inflammatory nature of these cancers, the use of lipid lowering agents like statins has been met with some success (19, 20). Preclinical data using statins in IBC show statin treatment can lead to increased apoptosis and radiosensitivity, inhibition of proliferation and invasion, and decreased metastatic dissemination of tumors (21). In a population-based cohort study in patients with IBC, statin use was associated with improved progression-free survival in patients with IBC (21). Recent studies have sought to better define this inflammatory microenvironment, and many have noted that macrophages may be important in mediating the radiosensitivity and metastatic potential of IBC tumors (22–25). Immune-regulating agents have also been implicated in the aggressiveness of IBC. In addition to the role that cytokines like INFα and TNFα may play in pathogenesis (26), many studies have reported that PD-L1 is consistently overexpressed in IBC tumors (27, 28). Upregulation of downstream signaling proteins like mTOR and JAK2/STAT3 have also been observed (28, 29). The role of RhoC in IBC has also been reported, but recent evidence suggests that downstream signaling may lead to unique metabolic regulation (30) and changes in lipid raft formation (31). Transcriptional reprogramming of IBC cells is also common, including C/EBPδ-mediated upregulation of VEGF-A (32) and upregulation of the redox-sensitive transcription factor NF-κB and the E3 ubiquitin ligase XIAP (33–36). These promising studies suggest that more effective treatment strategies are on the horizon.

Although we have demonstrated the radiosensitizing effects of olaparib in our models, these studies highlight the challenges of studying IBC. This study uses most of the available preclinical models of IBC but also highlights that there are a limited number of available models in which to study IBC. Thus, the need for additional model systems is critical to gaining a better understanding of the heterogeneity and pathogenesis of IBCs. Although our studies were conducted in IBC cell lines, an important future direction of this work will involve the use of patient-derived xenograft (PDX) models of IBC. In addition, this study primarily utilized the more potent PARPi olaparib, although our previous studies also evaluated the efficacy of radiosensitization using veliparib (10). Olaparib may be more potent given its dual functionality as a PARP enzymatic inhibitor and PARP trapper, whereas veliparib only has functions as an enzymatic inhibitor of PARP1 at the doses used for these studies (37). Although more potent, toxicity in clinical trials to date does not appear worse with olaparib and clinical data suggests that olaparib is well tolerated in vivo (8).

The dual functionality of some PARP inhibitors (like olaparib) to both inhibit enzymatic activity of the PARP1 protein as well as induce PARP trapping has been well documented (37–40). Recent literature also suggests that PARPi may cause an increase in replication fork acceleration, resulting in replicative stress that ultimately leads to cell death (41). Although the study reported here does not directly address the relative contributions of enzymatic PARPi verses PARP trapping on radiosensitization, studies are underway to determine how these functions may differentially contribute to the compounds' radiosensitizing effects. In addition to olaparib, PARP inhibitors such as talazoparib and rucaparib are used to treat other types of breast cancer (42, 43). These inhibitors may also be valuable in the treatment of IBC in combination with radiation and are currently being investigated.

Although we have shown that PARPi can be used for the radiosensitization of IBC, olaparib and other PARP1 inhibitors are currently being investigated as radiosensitization agents for the treatment of triple negative breast cancer (RadioPARP/NCT03109080), head and neck cancer (44), pancreatic cancer (45), prostate cancer (46), and ovarian cancer (47). More recent trials are testing whether PARPi is effective in combination with radiation in squamous cell carcinoma (48) (NCT02229656), locally advanced rectal cancer (NCT02921256 and NCT01589419), high grade gliomas (NCT03212742), non–small cell lung cancer (NCT01386385 and NCT02412371), and soft tissue sarcoma (NCT02787642; ref. 49). Thus, although this study is the first to report that PARP inhibition may be an effective strategy in patients with IBC, the concept of PARP inhibitor-mediated radiosensitization is being explored in many other cancer contexts.

IBC is a subset of breast cancer with limited treatment options and the lowest 5-year survival rates of any breast cancer type (1). Despite the limitations of the model systems, these data have provided the preclinical rationale for further clinical investigation. In a phase I trial, our group previously demonstrated that PARPi in combination with radiation may be a safe and effective strategy for women with IBC (and in women with locoregionally recurrent breast cancer; ref. 50). To that end, a randomized phase II trial (SWOG 1706, NCT03598257) comparing the effects of olaparib and radiation therapy to radiation therapy alone in patients with IBC is now underway. Patients in the combination arm begin treatment with olaparib one day prior to the initiation of radiation therapy, and olaparib is administered until the final day of radiation treatment. Invasive disease-free survival of women receiving treatment with olaparib and radiation will be compared with that of the group receiving radiation alone. Secondary endpoints, like local disease control, distant relapse-free survival and overall survival will also be assessed. In addition, correlative studies from this trial will be used to see if biomarkers of treatment response and efficacy can be identified. These correlative studies will also define the genomic and transcriptomic landscape of IBC in a large patient population and will assess how circulating tumor DNA (ctDNA) levels are affected by combination and single agent treatment.

Although it is evident from our study that PARPi with olaparib leads to radiosensitization of IBC cell lines, further studies are needed to determine the exact mechanism of olaparib-induced radiosensitization in IBC. Future transcriptomic and proteomic analyses of current model systems across multiple platforms may provide some insight as to the mechanism of this radiosensitization, and such studies are currently underway. Finally, correlative studies from SWOG 1706 will help inform future mechanistic studies and will provide a platform in which to evaluate potential predictive or prognostic biomarkers that may be able to help more effectively guide selection of patients with IBC to this approach for treatment intensification.

R. Jagsi reports of receiving other commercial research support from AbbVie provided drug only for a phase I trial of a PARP inhibitor on which she was principal investigator. No potential conflicts of interest were disclosed by the other authors.

Conception and design: A.R. Michmerhuizen, A.M. Pesch, L. Moubadder, B.C. Chandler, R. Jagsi, C. Speers

Development of methodology: A.R. Michmerhuizen, A.M. Pesch, L. Moubadder, B.C. Chandler, C. Speers

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.R. Michmerhuizen, A.M. Pesch, L. Moubadder, K. Wilder-Romans, M. Cameron, E. Olsen, D.G. Thomas, A. Zhang, N. Hirsh, C.L. Ritter, M. Liu, C. Speers

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.R. Michmerhuizen, A.M. Pesch, L. Moubadder, B.C. Chandler, D.G. Thomas, C.L. Ritter, M. Liu, S. Nyati, R. Jagsi, C. Speers

Writing, review, and/or revision of the manuscript: A.R. Michmerhuizen, A.M. Pesch, L. Moubadder, B.C. Chandler, M. Cameron, E. Olsen, D.G. Thomas, A. Zhang, N. Hirsh, S. Nyati, L.J. Pierce, R. Jagsi, C. Speers

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A.R. Michmerhuizen, A.M. Pesch

Study supervision: C. Speers

The Breast Cancer Research Foundation (N026000 to L. Pierce), The Komen for the Cure Foundation (N053349 to R. Jagsi), the University of Michigan Rogel Cancer Center (P30CA046592 to C. Speers). A. Michmerhuizen, A. Pesch, and B. Chandler are all supported by training grants through the NIH. A. Michmerhuizen is supported by T32-GM007315 (NIGMS), A. Pesch is supported by T32-GM007767 (NIGMS), and B. Chandler is supported by T32-CA140044 (NCI). Wendy Woodward at MD Anderson Cancer Center for providing MDA-IBC-3 cells, and Stephen Ethier at the Medical University of South Carolina for providing SUM-149 and SUM-190 cells.

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.