Evaluation of Pharmacodynamic Responses to Cancer Therapeutic Agents Using DNA Damage Markers

Abstract

Purpose: We sought to examine the pharmacodynamic activation of the DNA damage response (DDR) pathway in tumors following anticancer treatment for confirmation of target engagement.

Experimental design: We evaluated the time course and spatial activation of 3 protein biomarkers of DNA damage recognition and repair (γH2AX, pS343-Nbs1, and Rad51) simultaneously in a quantitative multiplex immunofluorescence assay (IFA) to assess DDR pathway activation in tumor tissues following exposure to DNA-damaging agents.

Results: Because of inherent biological variability, baseline DDR biomarker levels were evaluated in a colorectal cancer microarray to establish clinically relevant thresholds for pharmacodynamic activation. Xenograft-bearing mice and clinical colorectal tumor biopsies obtained from subjects exposed to DNA-damaging therapeutic regimens demonstrated marked intratumor heterogeneity in the timing and extent of DDR biomarker activation due, in part, to the cell-cycle dependency of DNA damage biomarker expression.

Conclusions: We have demonstrated the clinical utility of this DDR multiplex IFA in preclinical models and clinical specimens following exposure to multiple classes of cytotoxic agents, DNA repair protein inhibitors, and molecularly targeted agents, in both homologous recombination-proficient and -deficient contexts. Levels exceeding 4% nuclear area positive (NAP) γH2AX, 4% NAP pS343-Nbs1, and 5% cells with ≥5 Rad51 nuclear foci indicate a DDR activation response to treatment in human colorectal cancer tissue. Determination of effect-level cutoffs allows for robust interpretation of biomarkers with significant interpatient and intratumor heterogeneity; simultaneous assessment of biomarkers induced at different phases of the DDR guards against the risk of false negatives due to an ill-timed biopsy.

Conflict of interest statement

Conflict of Interest: none

©2019 American Association for Cancer Research.

Figures

Figure 1.. Timing of the activation of pS343-Nbs1, and γH2AX in response to gemcitabine in vivo.

(A) Analysis of gemcitabine-treated A673 Ewing sarcoma xenograft samples from 2, 4, 7, 12, 24 and 52 h after one dose gemcitabine (240 mg/kg). % Nuclear area positive (NAP) was measured for pS343-Nbs1 and γH2AX on co-stained slides. Marker expression per field is plotted. (B) Representative images showing pS343-Nbs1 (green), γH2AX (red), and DAPI (blue).

Figure 2.. Baseline expression of DDR markers and establishment of response cutoff values.

(A) Baseline expression of selected DDR markers was demonstrated and quantified in a colorectal tissue array with 32 tumors (2 cores per case are included) and 15 normal colon tissues. At least 500 individual tumor nuclei were quantified in 95% of the tissue microarray cores. (B-C) Baseline marker quantitation in advanced stage colorectal cancers from patients enrolled in Phase 1 clinical trials at NCI. Median expression and inter-quantile range is indicated for each marker. A minimum of 4,000 individual tumor nuclei were quantitated across at least two nonadjacent slides per biopsy specimen. (D) Baseline expression in human colon adenocarcinoma patient-derived xenografts. At least 5,000 nuclei quantified per model. (E) Baseline expression of selected DDR markers was demonstrated and quantified in 8 colorectal cancer cell lines. Over 1000 individual nuclei were quantified for each cell line. (F) Inter-lesion baseline Rad51 quantitation from patients with advanced stage cancers with two biopsies each collected from the same lesion. *p< 0.05. The dashed line represents our empirically determined baseline value cutoff of 5% of cells ≥ 5 Rad51 foci per nucleus.

Figure 3.. DDR multiplex is broadly applicable across classes of different genotoxic therapeutic agents in vivo.

Representative immunofluorescence images showing the induction of DDR markers in a variety of xenograft models treated with different genotoxic agents: (A) 60 mg/kg clofarabine Q2Dx5, (C) 2 mg/kg 5-aza-T-dCyd (AzaTdC) QDx5 with 2-day rest, and (E) 0.75 mg/kg decitabine (DAC) QDx5 with 2-day rest in HCT-116 colon xenografts; (G) single 6 mg/kg dose of cisplatin in A2780 ovarian; and (I) 4.7 mg/kg topotecan QDx5 in A375 melanoma xenografts. Resolution: 20X confocal, 0.24 μm/pixel. Corresponding changes in DDR marker expression are shown in (B, D, F, H, J), and the timing of xenograft sampling is indicated on the x-axes. Although all 3 markers are significantly activated by treatment with the 5 agents, at the cellular level, DDR expression varies widely. Statistical significance for the difference in marker expression levels between vehicle-treated and drug-treated xenografts corresponds to *p< 0.05, **p< 0.01, ***p< 0.001.

Figure 4.. Heterogeneous DDR marker activation in clinical colorectal cancer specimens.

Representative H&E and immunofluorescence (IFA) images from 3 patients with advanced colorectal cancer enrolled in NCI trial NCT01851369 before and 5 days after start of treatment with a DNA damaging therapeutic regimen consisting of TCR102 plus temozolomide administered orally once daily. Red scale bar represents 10 μm.

Figure 5.. DNA damage marker expression in a cell cycle–dependent manner contributes to heterogeneous DDR response.

(A) Representative immunofluorescence images showing expression of DAPI, DDR markers (Rad51, pS343-Nbs1, γH2AX), cell cycle markers (p21, cyclin B1, pHH3, geminin), and apoptosis markers (cleaved caspase 3 and γH2AX) in vehicle- and topotecan-treated A375 xenografts (4.7 mg/kg). (B) Quantitated expression of cell cycle markers p21, cyclin B1, pHH3, and geminin pre- and post-topotecan treatment (4.7 mg/kg). (C) Co-expression of cell cycle markers and DDR markers. White arrows point out yellow co-localization of pS343-Nbs1/geminin cells and Rad51/geminin-positive cells. The percentage of cells expressing the combination of markers indicated is provided on each slide. There was no colocalization observed of any DDR marker with pHH3 so numbers in middle panel reflect p21 co-localization with each DNA repair biomarker. The white bar represents 90 μm. Red bar 50 μm. (D) Schematic illustration of the effects of topotecan on the activation of DDR and cell cycle markers derived from the A375 xenograft experiments 4 h after a single 4.7 mg/kg dose of topotecan.