Veliparib Alone or in Combination with Mitomycin C in Patients with Solid Tumors With Functional Deficiency in Homologous Recombination Repair

Abstract

Background:

BRCA germline mutations are being targeted for development of PARP inhibitors. BRCA genes collaborate with several others in the Fanconi Anemia (FA) pathway. We screened cancer patients’ tumors for FA functional defects then aimed to establish the safety/feasibility of administering PARP inhibitors as monotherapy and combined with a DNA-breaking agent.

Methods:

Patients underwent FA functional screening for the presence (or lack) of tumor FancD2 nuclear foci formation on their archival tumor material, utilizing a newly developed method (Fanconi Anemia triple-stain immunofluorescence [FATSI]), performed in a Clinical Laboratory Improvement Amendments–certified laboratory. FATSI-negative patients were selected for enrollment in a two-arm dose escalation trial of veliparib, or veliparib/mitomycin-C (MMC).

Results:

One hundred eighty-five of 643 (28.7%) screened patients were FATSI-negative. Sixty-one received veliparib or veliparib/MMC through 14 dose levels. Moderate/severe toxicities included fatigue (DLT at veliparib 400mg BID), diarrhea, and thrombocytopenia. Recommended doses are 300mg BID veliparib and veliparib 200mg BID for 21 days following 10mg/m ² MMC every 28 days. Six antitumor responses occurred, five in the combination arm (3 breast, 1 ovarian, 1 endometrial [uterine], and 1 non–small cell lung cancer). Two patients have received 36 and 60 cycles to date. BRCA germline analysis among 51 patients revealed five deleterious mutations while a targeted FA sequencing gene panel showed missense/nonsense mutations in 29 of 49 FATSI-negative tumor specimens.

Conclusions:

FATSI screening showed that a substantial number of patients’ tumors have FA functional deficiency, which led to germline alterations in several patients’ tumors. Veliparib alone or with MMC was safely administered to these patients and produced clinical benefit in some. However, a better understanding of resistance mechanisms in this setting is needed.

Mutations of the breast cancer susceptibility (BRCA) genes have been identified as potential predictors of antitumor response to PARP inhibitors ( 1–5 ). They collaborate with several others in the Fanconi Anemia (FA) repair pathway ( 6–19 ). FA patients have a high incidence of malignancies and their cells exhibit hypersensitivity to DNA cross-linking agents such as mitomycin C (MMC) and cisplatin ( 20–22 ). FA falls into 17 complementation group subtypes ( 12–19 ). Eight of these proteins and three associated factors are subunits of an FA core complex, a nuclear E3 ubiquitin ligase ( 12–14 , 22 ). Monoubiquitination of FancD2 and FancI by the FA core complex followed by nuclear colocalization with other DNA damage response proteins results in nuclear foci of repair ( Figure 1 ) ( 15 ). Any alteration that disrupts components of the core complex abrogates its E3 ligase function, leading to defective mono-ubiquitination and no repair foci formation ( 7 , 22 ).

Figure 1.

The Fanconi Anemia (FA) pathway and formation of repair foci. Following DNA interstrand crosslink damage, the FANCM-FAAP24-MHF1-MHF2 anchor complex recruits the FA core complex I, which functions to activate FANCD2 and FANCI by mono-ubiquitinating the proteins. The activated FANCD2 and FANCI heterodimers are subsequently transported to subnuclear foci, which in collaboration with additional genes result in homologous recombination DNA repair.

Disruptions of the FA/BRCA cascade have been noted in sporadic cancers, including epigenetic silencing of the FA core complex, mutations of FA/BRCA genes, or modification of encoded products ( 23–25 ). Cancers with a defective FA/BRCA pathway are likely to be more sensitive to cross-link-based therapy, and treatments in which an additional repair mechanism is targeted may have antitumor activity or provide therapy sensitization ( 26–31 , 57 ).

We hypothesized that given the number of modifications that could interfere with FA pathway functionality a substantial number of patients would be good candidates for PARP inhibitor or cross-link cytotoxic-based therapy. To identify these patients, we developed an assay, FancD2/DAPI/Ki67 (Fanconi Anemia triple-stain immunofluorescence [FATSI]), which permits the observation (or lack thereof) of FancD2 foci formation in proliferating cells ( 32 ). The FATSI assay demonstrated reliable performance in paraffin-embedded (FFPE) archival tumor material and underwent validation in a Clinical Laboratory Improvement Amendments (CLIA)–certified laboratory, thus it is suitable for large-scale screening.

In this first of its kind trial we set out to: 1) screen cancer patients to identify those with FA functional defects in their tumors, 2) establish the safety/feasibility of PARP inhibition as monotherapy and in combination with a DNA-breaking agent in these patients, and 3) recommend appropriate doses for subsequent studies.

Methods

Patients

The Institutional Review Boards of The Ohio State University (OSU) and the Georgetown University approved this study (clinicaltrials.gov; NCT01017640). The study was performed in two parts. Patients older than age 18 years with advanced solid malignancies consented to have their existing FFPE tumor tissue screened for FA deficiency by the FATSI assay. Those determined to be FA functionally deficient (FATSI-negative) were offered a place in the therapeutic portion of the trial. Separate written consents for the screening and therapeutic intervention were obtained ( Figure 2 ).

Figure 2.

Patient screening and flow diagram. CLIA = Clinical Laboratory Improvement Amendments.

Other eligibility requirements for the therapeutic portion included progressive disease, less than three previous cytotoxic chemotherapy regimens for metastatic disease, and a lapse of four weeks from chemotherapy or radiation therapy. Prior MMC restricted to topical applications or chemo-embolization was allowed. We required patients have an ECOG performance status of less than 2 and normal organ and marrow function (absolute neutrophil count ≥ 1.5 x 10 ⁹ ; platelets ≥ 100 x 10 ⁹ ; hemoglobin ≥ 9g/dL; serum creatinine and bilirubin ≤ 1.5 x the upper limit of normal; AST/ALT ≤2.5 x the upper limit of normal). Patients were excluded because of pregnancy, active brain metastases, recent history of seizures, uncontrolled concurrent illness, combination antiretroviral therapy, and previous treatment with PARP inhibitors.

Treatment Plan

Patients were allocated to one of two arms, and a 3+3 dose escalation design was followed. On arm 1, patients received the oral PARP inhibitor veliparib as monotherapy, and on arm 2 patients received veliparib combined with MMC. MMC was chosen because of its well-defined role in producing double-strand DNA breaks, its approval for use in a vast number of solid malignancies, demonstration in our laboratory of stimulation of the FA pathway, and because our previous experience has shown it induces sensitivity to veliparib ( 33 ). No pharmacokinetic interactions between veliparib or MMC were anticipated.

To allocate patients to the treatment arms, we used a “Ping-Pong” approach. As 3+3 phase I trials are designed to enroll three patients and evaluate toxicity before evaluating the next dose level or expanding, there is a gap of a few weeks for enrollment. We initiated with the monotherapy arm, continuing to the combination arm once the monotherapy arm at a particular dose level was full. The “Ping-Pong” approach discouraged bias from the referring physician (or patient) for a particular arm (with chemo vs without chemo).

The velaparib starting dose was 50mg twice daily (BID), which had demonstrated PARP catalytic activity inhibition in tumor in a previous report ( 34 ). As a safety precaution for the combination arm, an increase on the duration of administration of 50mg BID veliparib for the first three patients cohorts (7, 14, and 21 days, every 28 days) following MMC 10mg/m ² was evaluated prior to escalating veliparib. MMC was administered every 28 days, with a cumulative dose cap of 40mg/m ² . Thus, no patient in the combination arm would receive a higher dose than a dose previously cleared on the monotherapy arm. Patients experiencing clinical benefit continued on single-agent veliparib once the MMC dose cap was reached.

Dose-limiting toxicity (DLT) was defined as grade 4 neutropenia for more than seven days or with sepsis or fever, grade 4 thrombocytopenia, grade 3–4 nonhematologic toxicity that caused an interruption of veliparib dosing for seven or more days, or any grade 1–2 treatment-related toxicity requiring dose delays for more than four weeks. At least two cycles for DLT evaluation were required in the combination arm whereas one cycle (4 weeks) for the monotherapy arm was considered sufficient. The maximum tolerated dose (MTD) was the dose at which no more than one of six patients experienced DLT. Toxicities were graded according to National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events version 4. Measurable disease was not required. Imaging was repeated every two cycles.

Biomarker and Correlative Studies

Tumor Tissue Screening

Archival tumor tissue was retrieved and sent to the Department of Pathology. Tissue sections were cut to 4 microns, and FATSI staining and interpretation was performed in the CLIA-certified OSUCCC Molecular Pathology Core Laboratory (MPCL) by the same pathologist, as previously described ( 32 ).

BRCA Germline Mutational Analysis

Blood samples (8mL) from patients in the therapeutic portion (receiving veliparib or veliparib/MMC) were collected prior to treatment using blood collection specimen kits obtained from Myriad Genetics (Salt Lake City, Utah), where sequencing for detection of BRCA1 and 2 mutations and the five most common BRCA1 large rearrangements was performed.

Gamma-H2AX and PARP Activity

Blood samples prior to dosing and at two, four, six, and 24 hours on day 1 of cycles 1 and 2 were obtained. PBMC were isolated with BD Vacutainer Cell Preparation Tubes (BD Diagnostics, Franklin Lakes, NJ) and fixed by immersing the cells in 2% formaldehyde solution at 37°C for 20 minutes and then stored in 70% ethanol until batch-analyzed. Quantitative staining for gammaH2AX and PARP activity using enzyme-linked immunosorbent assay was performed at the National Clinical Target Validation Laboratory, NCI Frederick, as previously described ( 35 ).

Repair Sequencing Panel

To evaluate potential genetic alterations that led to the FA functional deficiency, we developed a targeted FA sequencing panel (25 genes) using Agilent’s Haloplex Custom Library system to test a group of random FATSI-negative samples from the screening portion of the trial ( Supplementary Table 1 , available online). The designed panel includes the entire coding sequence of each gene. One cell line and a patient sample with known mutations in FANCC and FANCD2 were positive controls. A library was prepared for each patient specimen as previously described ( 36 ) and sequenced on an Illumina HighSeq2000 to an average coverage of 1000X. Sequencing data was analyzed using the Agilent SureCall software. Variant results were filtered based on read depth and quality, nonsynonymous amino acid changes or splice site base changes, and allelic frequency in the general population.

Results

Screening

From October 2009 to June 2014, 724 patients with solid malignancies consented to FATSI screening. Of these, 71 had insufficient archived tissue, four failed screening because of technical issues, and six either withdrew or died before testing was conducted. FATSI was performed in 643 patients ( Table 1 ). For patients with archived material at OSU, the retrieval of and specimen selection, cutting, and processing took an average of two days. Staining and interpretation at the OSU-MPCL were conducted within 48 hours.

Table 1.

Tissue screening results

Primary tumor type	FATSI negative	FATSI positive	Patients tested	% FATSI negative
Colon	63	104	167	37.7
Rectal	5	9	14	35.7
Breast	41	101	142	28.8
Lung non–small cell	19	78	97	19.6
Lung small cell	6	25	31	19.4
Ovarian	6	20	26	23.1
Biliary track	8	13	21	38.1
Pancreatic	3	17	20	15.0
Bladder	5	13	18	27.8
Head & neck	6	12	18	33.3
Endometrium	1	12	13	7.7
Prostate	1	12	13	7.7
Adenoca unknown primary	2	8	10	20.0
Neuroendocrine	3	6	9	33.3
Mesothelioma	2	5	7	28.6
Cervical	0	5	5	0
Thymic	2	3	5	40.0
Sarcoma	1	2	3	33.3
Melanoma	2	1	3	66.7
Gastro-intestinal stromal tumor	1	2	3	33.3
Stomach	1	2	3	33.3
Small bowell	2	1	3	66.7
Appendix	1	1	2	50.0
Renal	1	1	2	50.0
Adrenal	1	1	2	50.0
Testicular	1	0	1	100.0
Vulva	0	1	1	0
Penile	0	1	1	0
Fallopian	0	1	1	0
Peritoneal	1	0	1	100.0
Hepatic	0	1	1	0
Total	185	458	643	28.7

Colorectal, breast, and lung cancers were the most common malignancies tested. Overall, 28.7% of patients demonstrated no FancD2 foci formation in their tumor specimens. FA functional deficiency was observed throughout most histological types tested.

Dose Escalation and Toxicities

When appropriate, FATSI-negative patients consented to the therapeutic portion of the study. Sixty-one patients were enrolled through one of 14 dose levels in the two arms. Table 2 depicts the demographics and tumor types of these patients.

Table 2.

Patient characteristics*

Characteristic	No. of patients
	Arm 1 (veliparib)	Arm 2 (veliparib +MMC)	All
No. of enrolled patients	32	29	61
Men/women, No.	15/17	15/14	30/31
Race/ethnicity, No.
White	29	28	47
African American	3	1	4
Hispanic/Latino	0	0	0
Other	0	0	0
Median age (range), y	57 (27–79)	59 (30–77)	58 (27–79)
ECOG Performance Status, No.
0	10	13	23
1	20	15	35
2	2	1	3
Primary tumor type, No.
Breast	6	6	12
Colorectal	11	9	20
Non–small cell lung	3	3	6
Small cell lung	2	0	2
Ovarian	3	1	4
Small bowel	0	2	2
Pancreatic	0	2	2
Cholangiocarcinoma	1	1	2
Mesothelioma	1	1	2
Thymic, bladder, ampullary, anal, testicular	5	0	5
Nasopharyngeal, prostate, melanoma, endometrial	0	4	4
Previous chemotherapy
None	0	0	0
1 line	3	6	9
2 lines	18	14	32
≥3 lines	11	9	20

* MMC = mitomycin C.

Dose escalation, numbers of patients with dose reductions, DLT, and the MTD for each arm of the trial are depicted in Table 3 . Overall toxicities are depicted on Table 4 . For veliparib monotherapy, two of three patients developed DLT (grade 3 fatigue) at 400mg BID during their first cycle. The first subject at this level tolerated well one cycle but discontinued trial participation for surgical resection of his mesothelioma. However, a subject age 77 years with metastatic breast cancer developed grade 2 nausea/vomiting and diarrhea 24 hours after veliparib initiation, which improved by the following day with drug interruption. Reintroduction of veliparib resulted in reappearance of symptoms, development of grade 3 fatigue and consent withdrawal (received 8 doses). The other subject with DLT (a woman age 63 years with colon cancer) developed lightheadedness and worsening fatigue over the first cycle, which she completed, spending more than 50% of her day resting or in bed. Interruption of veliparib for two weeks resulted in improvement of fatigue to grade 2. Treatment was discontinued because of clinical disease progression.

Table 3.

Dose escalation and toxicities of veliparib alone and with MMC

Dose level	Veliparib mg	MMC mg/m 2	No. patients treated	No. cycles	Pts with DLT/ evaluable	Pts with dose delays/ reductions
Arm 1
1	50 BID	-	3	64	0/3	0
2	80 BID	-	3	5	0/3	0
3	100 BID	-	3	17	0/3	0
4	150AM/100PM	-	3	6	0/3	0
5	150 BID	-	3	8	0/3	0
6	200AM/150PM	-	4	11	0/3	0
7	200 BID	-	3	7	0/3	0
8	300 BID	-	7	15	0/6	0
9	400 BID	-	3	4	2 /3	0
Total	-	-	32	137	-	-
Arm 2*
1A	50 BID x 7d	10 q 28d	4	11	0/3	0
2A	50 BID x 14d	10 q 28d	7	23	0/6†	0
3A	50 BID x 21d	10 q 28d	3	11	0/3	0
4A	100 BID x21d	10 q 28d	4	45	0/3	1/1
5A	200 BID x 21d	10 q 28d	11	22	1/6	0/1
Total	-	-	29	112	-	-

* Two evaluable cycles required for dose-limiting toxicity evaluation. BID = twice daily; DLT = dose-limiting toxicity; MMC = mitomycin C; Pts = patients.

† Because of a G3 dyspnea episode, which was unclear if related to MMC or to tumor progression after a third cycle, the dose level was expanded to six patients.

Table 4.

Hematologic toxicities of veliparib alone and in combination with mitomycin C in FA pathway deficient patients

Toxicity*	No. of patients’ cycles with toxicity by toxicity grade
	Veliparip, all doses			Veliparib, MTD			Veliparib-MMC, all doses			Veliparib-MMC, MTD
	2	3	4	2	3	4	2	3	4	2	3	4
Hematologic
Neutropenia	1	0	0	0	0	0	4	1	0	0	0	0
Anemia	11	2	1	5	0	0	16	4	0	2	3	0
Thrombocytopenia	1	0	0	1	0	0	9	7†	1	5	3	0
Lymphocytopenia	13	3	0	1	1	0	16	11	0	8	7	0
Febrile neutropenia	0	0	0	0	0	0	0	0	0	0	0	0
Nonhematologic
Nausea	6	1	0	3	0	0	5	1	0	4	0	0
Vomiting	9	2	0	3	0	0	2	1	0	1	1	0
Anorexia	2	1	0	3	0	0	7	0	0	5	0	0
Diarrhea	7	2	0	4	1	0	3	1	0	0	0	0
Dysgeusia	0	0	0	0	0	0	8	0	0	8	0	0
Headaches	0	1	0	0	0	0	1	1	0	0	1	0
Transaminitis	2	0	0	0	0	0	1	2	0	1	0	0
Alkaline phosphatase	2	1	0	0	0	0	2	1	1	2	1	0
Fatigue/asthenia	3	7	0	2	2	0	23	9	0	12	4	0
Arthralgia/myalgia	1	2	0	1	0	0	2	0	0	2	0	0
Peripheral neuropathy	10	0	0	2	0	0	7	0	0	0	0	0
Hyperglycemia	3	1	0	0	0	0	3	0	1	2	0	1
Peripheral edema	0	0	0	0	0	0	3	0	0	3	0	0

* Toxicities are reported as per National Cancer Institute common toxicity criteria 4. Two hundred forty-nine cycles were administered: 137 for veliparib, 15 for veliparib at maximum tolerated dose (MTD), 112 for mitomycin C (MMC)/veliparib, 22 for MMC/veliparib at MTD. FA = Fanconi Anemia; MMC = mitomycin C; MTD = maximum tolerated dose.

† One episode of thrombotic microangiopathy.

The 300mg BID dose was expanded to six evaluable patients. At this dose, toxicities included one patient with grade 3 fatigue lasting fewer than seven days, a grade 3 pneumonia during cycle 2, and a grade 3 diarrhea during cycle 5. The 300mg BID veliparib dose was therefore declared the MTD for veliparib as monotherapy for patients with FA-deficient tumors.

For the combination arm, one patient receiving 14 days of veliparib dosing (2A) developed grade 3 fatigue, dyspnea, hypoxia, and a pleural effusion after completing two cycles. Because of the possibility of MMC-induced hypoxia, the dose level was expanded, with no additional moderate or severe toxicities. Two cycles were required subsequently for the combination arm as a safety precaution for evaluation of DLT. For patients receiving 21 days of veliparib (3A), no moderate/severe toxicities occurred during the first two cycles. However, one patient developed a thrombotic thrombocytopenic purpura-like syndrome after cycle 3. Veliparib dose in this schedule was subsequently escalated, up to 200mg BID. At this last dose level, the enrollment of 11 patients was necessary to assure that at least six patients completed two cycles. Although only one DLT (grade 3 fatigue) occurred, the inability to deliver MMC consistently on schedule after cycle 3 made additional dose escalation impractical. Thus MMC 10mg/m ² followed by 200mg BID veliparib is recommended for initial dosing in subsequent trials.

Antitumor Activity

Supplementary Table 4 (available online) depicts tumor assessments results. Six antitumor responses were confirmed (5-RECIST/1-PET criteria) ( 37 , 38 ). Three patients withdrew consent during the first cycle without progression. Two withdrew because of ill effects, and the third to undergo an extrapleural pneumonectomy. Histologic sections demonstrated less than 10% residual viable tumor cells in the resected pleura. Two patients in the combination arm withdrew for side effects without progression after one cycle.

A patient with metastatic ovarian cancer who had pathological documentation of PET avid metastasis to liver after treatment with cisplatin/gemcitabine had resolution of PET uptake after two cycles of therapy at dose level 5A. Previous treatment included paclitaxel/carboplatin for ovarian cancer and adjuvant doxorubicin for a previous bilateral breast cancer. Partial responses occurred in three patients with breast cancer, two of whom are still receiving therapy (dose level 1, 60 cycles; and dose level 4A, 36 cycles), in a non–small cell lung cancer patient (1A), and in a patient with endometrial carcinoma metastatic to peritoneum (5A). Stable disease was the best response in 18 patients (median = 6 months, range = 3–15 months).

Correlative Studies

PBMC BRCA analysis (Myriad) among 51 patients receiving veliparib or veliparib/MMC showed that five patients (2 breast, 1 ampullary, 2 ovarian) carried BRCA-deleterious mutations ( Supplementary Table 5 , available online). Of note is the uncovering of four cases of breast cancer in the family of an ampullary carcinoma patient with a previously unsuspected BRCA2 deleterious mutation and the discovery of this mutation in two of his three daughters ( Figure 3 ).

Figure 3.

Patient with ampullary carcinoma and BRCA2 mutation. Four cases of breast cancer in his family were identified. The testing of his three daughters identified the deleterious mutation in the germline of two of them (subjects 29 and 35).

Seventeen patients receiving veliparib had PAR assessment on PBMC ( Supplementary Figure 1 , available online). Thirteen of 15 patients with evaluable data had reductions during cycle 1. PAR level was lower at cycle 2 predose timepoint than that of cycle 1 predose in six of seven patients on veliparib alone, suggesting long-term modulation of PARP activity. Following MMC, four of five patients had higher PAR levels at cycle 2 predose than that of C1, suggesting mitomycin-C activated PARP. However, PAR levels went down in all patients in subsequent samples.

Forty-seven patients were evaluated for gamma-H2AX on PBMC ( Supplementary Tables 2 and 3 , available online). Limited data points from informative patients of gama-H2AX cytospin made it difficult to draw any conclusion. Significant induction (>2%) only occurred in two patients in arm 1 and five from arm 2.

We sequenced 49 random patient tumor DNA samples, FATSI-negative per screening, with the FA sequencing panel. Thirty-four unique alterations were identified in 29 of the 49 patient specimens ( Supplementary Table 6 , available online). Seventeen patients who received veliparib were among those sequenced ( Supplementary Table 7 , available online). Tumor DNA from the two patients with germline BRCA mutations analyzed showed the same germline mutations at a high variant allele frequency (VAF). Loss of heterozigosity (LOH) was confirmed in the tumor. A RAD51c mutation (c.223_224insA p.Y75*) with high VAF was detected in the tumor from a breast cancer patient with a PR who was on maintenance veliparib for two years after initial induction with veliparib/MMC. The same RAD51c mutation was demonstrated in her germline tissue (Invitae Panel, San Francisco, CA), as well as family history consistent for a cancer predisposition syndrome ( Figure 4 ).

Figure 4.

Family tree of patient with breast cancer and a Rad51c mutation in tumor and germline. Three cases of breast cancer, one of ovarian cancer, and one of leukemia in the family were detected.

One additional patient with metastatic lung adenocarcinoma and tracheal infiltration had a truncation mutation in the ataxia telangiectasia mutated gene (ATM c.6976-1 G>T), which was not present in his germline, in addition to an ERCC4 missense mutation (P379S) in both germline and tumor with LOH. An episode of massive hemoptysis occurred during his first cycle of veliparib monotherapy, requiring intubation. Continuation of life support was declined by the family.

Tumors and adjacent tissue from 10 patients FATSI-positive per screening were analyzed as controls with the FA sequencing panel. A deleterious mutation (ERCC4), along with a germline potentially damaging mutation in FAN1, was found in only one patient. One other patient had a somatic missense mutation in ATM ( Supplementary Table 6 , available online).

Discussion

The FATSI screening results confirm that a substantial number of patients throughout a variety of primary organ sites have FA functional deficiency in their tumors. Colorectal, breast, and lung cancers had the most representation. The 29% rate is consistent with our previously reported data in tumors obtained from the Cooperative Human Tissue Network ( 32 ). Although the Ki67 (a commonly used tumor proliferation index marker) expression requirement for the test (>10%) to be evaluable helps to eliminate false negatives because of necrosis, fibrosis, etc., low-proliferating tumors may be affected in a similar way, thus they are not suitable for this test.

The targeted FA sequencing panel was able to show FA-associated repair alterations at the genomic level in a substantial fraction of FATSI-negative patients, as compared with a control group of FATSI positives. Other mechanisms for foci formation loss not examined could include large deletions not readily captured by the panel, epigenetic silencing of the FA genes, or mutations in genes not included. To test this notion, an expansion cohort of FATSI-negative cancer patients is being accrued who will provide fresh tumor biopsies prior to treatment. Our prediction is that by adding RNAseq, we will be able to identify the cause of the functional defect detected by FATSI in most patients with no mutations on the repair sequencing panel. Therefore, it is our contention that the test could not only provide an inexpensive and practical tool for population screening for tumor-associated DNA-repair dysfunction, which then could be targeted for sequencing, but could also provide the phenotypic dysfunction readout of a yet-uncharacterized repair gene alteration identified by genetic screening.

Another dimension added by FATSI screening was that it led to identification of unsuspected germline mutations, which impacted screening and prevention measures in the families of some of our patients. These included an uncommon Rad51c insertion mutation in a breast cancer patient predicted to be pathogenic ( 39–40 ). Both ovarian and breast cancer were present in her genealogical tree. The same mutation was identified in both germline and tumor material from a sister recently diagnosed with ovarian cancer.

Our study also showed that it is safe to administer veliparib at doses of up to 300mg BID to FA-deficient patients and that veliparib can be safely combined with the DNA-breaking agent mitomycin-C. However, veliparib as monotherapy did not produce a substantial number of tumor regressions. Potential reasons would include: 1) veliparib spectrum of doses below MTD utilized (it should also be noted that doses of up to 400mg BID have been tolerated in a different population of patients); 2) veliparib relatively low PARP trapping activity (a newly described mechanism of action for PARPi) ( 41 , 42 ); 3) restoration of FA functionality as a result of systemic treatment in the interval between the material in archives and enrollment in the trial ( 43 , 48 ); or 4) the presence of an additional antiapoptotic stimulus to which the repair deficient cell has become addicted (nononcogenic addiction/induced-essentiality) ( 49 , 50 ). The latter mechanism is expected to operate more commonly in non-BRCA-mutated, repair-deficient tumors, as BRCA mutation is in itself a potent anti-apoptotic stimulus.

A few genomic signatures have come forward as suggestive of “BRCAness” in tumors, and evaluations of their potential for enriching patient populations most sensitive to PARPi are actively being pursued ( 51–54 ). Given that FATSI is a functional, rather than a genomic, analysis, it would be of great interest to explore correlations or overlap between these signatures and FATSI. Properly powered prospective therapeutic trials to evaluate concurrently FATSI and other potential biomarkers as predictors of PARPi tumor sensitivity at proper doses across diverse malignancies would help to clarify their comparative clinical value.

Limitations of our study include the lack of fresh biopsies to cross-validate and assess potential changes in the pathway activation induced by chemotherapy given in the interval between the available archival material and the initiation of the pharmacological intervention and the inability to assess well the functionality in low-proliferating tumors.

Future plans include an ongoing expansion cohort of the combination of veliparib/MMC at recommended doses in FATSI-negative colorectal cancer patients, in which fresh biopsies are being obtained for patient-derived tumor xenografts creation, in order to address and overcome resistance. A follow-up phase II clinical trial proposal has been endorsed by the NCI to test the ability of a potent PARP trapping agent ( 55 , 56 ) to achieve synthetic lethality in nonbreast/nonovarian FA-deficient patients. Patients are selected by either FATSI (1 cohort) or genomically defined cohorts harboring mutations pertinent to homologous recombination ( 58 ), De novo and acquired resistance will be evaluated. Fresh biopsies, cross validation between functional and genomic abnormalities, and whole-exome tumor sequencing will be performed in all patients.

Funding

This work was supported by National Institutes of Health (NIH) grants R01-CA152101, N01-CM-2011-00070 (HHSN261201100070C) to MAV, and OSUCCC P30 CA016058-38.

Notes

The study funders had no role in design of the study; the collection, analysis, or interpretation of the data; the writing of the manuscript; or the decision to submit the manuscript for publication.

The authors revealed no conflict of interests related to the performance of the work reported or the writing of this manuscript.

Special thanks at the OSU Molecular Core Facility to Kristin Kovach (FATSI) and Nehad Mohamed (KRAS and BRAF); at the OSUCCC Clinical Trial Office to Kirsten Keinsenmayer (outside tissue coordination and reporting) and Andrea Lively (regulatory); at Tissue Archives to Mariya Kravets; at Cancer Genetics to Robert Pilarski; and Yiping Zhang (NCI), Li Gao, and Britanny Aguila (MAV’s lab, preparation, and performance of the correlative assays) for their diligent work.