Multidrug Analyses in Patients Distinguish Efficacious Cancer Agents Based on Both Tumor Cell Killing and Immunomodulation

Abstract

The vision of a precision medicine-guided approach to novel cancer drug development is challenged by high intratumor heterogeneity and interpatient diversity. This complexity is rarely modeled accurately during preclinical drug development, hampering predictions of clinical drug efficacy. To address this issue, we developed Comparative In Vivo Oncology (CIVO) arrayed microinjection technology to test tumor responsiveness to simultaneous microdoses of multiple drugs directly in a patient's tumor. Here, in a study of 18 canine patients with soft tissue sarcoma (STS), CIVO captured complex, patient-specific tumor responses encompassing both cancer cells and multiple immune infiltrates following localized exposure to different chemotherapy agents. CIVO also classified patient-specific tumor resistance to the most effective agent, doxorubicin, and further enabled assessment of a preclinical autophagy inhibitor, PS-1001, to reverse doxorubicin resistance. In a CIVO-identified subset of doxorubicin-resistant tumors, PS-1001 resulted in enhanced antitumor activity, increased infiltration of macrophages, and skewed this infiltrate toward M1 polarization. The ability to evaluate and cross-compare multiple drugs and drug combinations simultaneously in living tumors and across a diverse immunocompetent patient population may provide a foundation from which to make informed drug development decisions. This method also represents a viable functional approach to complement current precision oncology strategies. Cancer Res; 77(11); 2869-80. ©2017 AACR.

Conflict of interest statement

Disclosure of Potential Conflicts of Interest

J. Frazier has ownership interest (including patents) in Presage Biosciences, Inc. A. Moreno-Gonzalez is a director of clinical technology development at Presage Biosciences. R.K. Amaravadi has ownership interest (including patents) in Presage Biosciences, has a patent licensed to Presage Biosciences, and is a consultant/advisory board member for Presage Biosciences. J.M. Olson is a director and has ownership interest (including patents) in Presage Biosciences. R. Klinghoffer has ownership interest (including patents) in Presage Biosciences. No potential conflicts of interest were disclosed.

©2017 American Association for Cancer Research.

Figures

Figure 1

CIVO microinjections result in easily observable spatially defined drug-specific responses in canine STS tumors. A, A representative image of a resected portion of a tumor subjected to CIVO microinjection with an eight-needle array. Resection occurred 48 hours following the CIVO procedure. A coinjected FTM denoting the position of each injection site is visualized with a SLR camera outfitted with a custom lens filter and LED light source. B, An H&E-stained 4-μm thick section of the same tumor. Microinjection sites 1 and 3 both contain doxorubicin. Scale bar, 2,000 μm. C–F, High magnification images of tumor cell responses captured at sites of microinjection with doxorubicin (C), docetaxel (D), gemcitabine (E), and vehicle (F). Scale bar, 25 μm. G–L, Images from a tumor removed from a different subject and subjected to analysis for drug responses by IHC staining with antibodies that detect DNA damage (phospho-γH2AX) and mitotic arrest (phospho-histone H3). G, Microinjection site of doxorubicin stained with an antibody to detect phospho-γH2AX as a marker of DNA damage response (yellow). Scale bar, 500 μm. H, High magnification image of the same site shown in G. Scale bar, 25 μm. I, The fraction of phospho-γH2AX cells plotted as a function of radial distance from the injection site 48 hours after microinjection of vehicle, doxorubicin (P < 0.001 vs. vehicle), gemcitabine (P = 0.97), or docetaxel (P = 0.77). Data are averages ± SEM (n = 18 tumors; J) Microinjection site of docetaxel stained with an antibody to detect pHH3 as a marker of mitotic arrest (white). Scale bar, 500 μm. K, High magnification image of the same site shown in K. Scale bar, 25 μm. L, The fraction of pHH3-positive cells plotted as a function of radial distance from the injection site 48 hours following microinjection of vehicle, doxorubicin (P = 0.36 vs. vehicle), gemcitabine (P = 0.94), or docetaxel (P < 0.001). Data are averages ± SEM (n = 18 tumors). DAPI-stained nuclei are shown in blue.

Figure 2

CIVO microinjection of doxorubicin results in tumor cell clearing and a localized immune cell infiltration. A, A representative image of an H&E-stained site 48 hours following localized exposure to doxorubicin. B, Anti-vimentin staining to detect loss of STS cells. Scale bar, 500 μm. C, Anti-cleaved caspase-3 staining to detect apoptotic cells. D, Anti-S100A9 staining (MAC387 antibody) to detect macrophage infiltration. E, Anti-CD3 staining to detect T lymphocytes. Scale bar for A–E, 500 μm. F, High magnification image of a CD3/granzyme B double-positive cell docking onto an adjacent STS cell. Scale bar, 10 μm. High magnification image of calreticulin IHC staining within a vehicle (G), doxorubicin (H), and docetaxel (I) injection site, respectively. Scale bars for G–I, 50 μm. DAPI-stained nuclei, blue.

Figure 3

CIVO identifies doxorubicin sensitive and resistant tumors. A, Plotting region of extent of tumor cell kill (loss of vimentin, orange) versus extent of doxorubicin exposure (γH2AX-positive cells, yellow), both as a function of radial distance from the initial site of doxorubicin microinjection. Drug-resistant tumors were identified as those that demonstrated little or no cell death following doxorubicin exposure (blue circles). Other tumors demonstrated poor distribution of doxorubicin leading to a reduction in overall exposure and tumor cell kill (red open circles). Groups were statistically segregated by K-means clustering. B, Representative examples of patient tumors classified as drug sensitive and drug resistant. Scale bars, 500 μm. Note that one of the 18 tumors included for analysis in this study did not have a visible doxorubicin injection site and was excluded from this doxorubicin CIVO responder analysis.

Figure 4

A potent lysosomal autophagy inhibitor, PS-1001, combines with doxorubicin to convert drug-resistant tumors into doxorubicin-sensitive tumors. A and B, Anti-LC3A/B staining to detect autophagy inhibition in vehicle (A) and PS-1001 (B) exposed tumor tissue 48 hours following CIVO microinjection. Scale bar, 20 μm. C, Quantification of localized sarcoma cell killing activity induced by doxorubicin alone, or by the combination of doxorubicin plus PS-1001 in the five subjects classified as doxorubicin resistant. Plotting extent of the loss of vimentin-positive cells (mm) as a function of radial distance from initial site of microinjection. D, Images of doxorubicin distribution as tracked by γH2AX (yellow) and tumor cell killing activity (loss of vimentin, orange) comparing doxorubicin (top row) with the combination of PS-1001 and doxorubicin (bottom row) in the three tumors exhibiting enhanced antitumor responses to microinjection of the drug combination. DAPI-stained nuclei, blue; FTM, green. Scale bars, 500 μm.

Figure 5

PS-1001 increases infiltration of macrophages in a subset of tumors coinjected with doxorubicin. Tumor sections classified as poor distribution (left; n = 5), doxorubicin-resistant (middle; n = 5), and doxorubicin-sensitive (right; n = 6) were stained for macrophages with the MAC387 antibody (yellow). DAPI-stained nuclei and FTM are shown in blue and green, respectively. Scale bars, 500 μm. Bottom row, quantification of macrophage infiltration by plotting the density of MAC387-positive cells (cells/mm2) as a function of radial distance from the site of CIVO microinjection (data are averages ± SEM). The overall difference across all tumors was statistically significant (P = 0.04). Significance macrophage infiltration was also observed in the three doxorubicin-resistant tumors converted to sensitive by the addition of PS-1001 (P = 0.028). Strong trends were noted within the doxorubicin-resistant (P = 0.051) and doxorubicin-sensitive (P =0.12) subgroups, but not the subgroup with poor distribution (P = 0.65).

Figure 6

PS-1001 skews macrophage polarization toward the M1 state. A, CIVO analysis to examine the impact of PS-1001 on macrophage polarization in canine STS patients. Representative images from localized regions of macrophage infiltration at sites of CIVO microinjection of doxorubicin + PS-1001 into tumors classified as doxorubicin resistant. Tumor sections were costained with MAC387 (red) and a marker of M1 polarization (pSTAT1, green, top row), or a marker of M2 polarization (c-Maf, green, bottom row) to assess dual positivity, visualized in yellow. Scale bars, 500 μm. B,In vitro cell–based biomarker analysis to examine the effect of PS-1001 on stimuli known to directly influence macrophage polarization status. qPCR analysis for transcriptional biomarkers of macrophage polarization status was performed following 24-hour exposure of RAW 264.7 cells to M1 stimulus (LPS + IFNγ) or M2 stimulus (IL4), −/+ PS-1001. Transcripts assessed included M1 markers (light gray bars) IL6, iNOS, and Socs3, and M2 markers (dark gray bars) Arg1, Ym2, Mrc1, and Socs1. All data were normalized to signal observed in the absence of PS-1001 exposure. Error bars, SD of triplicate samples. C, Cytokine profiling of RAW 264.7 cell supernatant by ELISA assay following 24-hour exposure to PS-1001. Samples were run in triplicate and averages values were plotted. Error bars, SD of the triplicate samples.

Figure 7

CIVO captures the complexity of tumor response to drug exposure and highlights interplay between diverse components of the tumor microenvironment. A, Site of doxorubicin + PS-1001 microinjection showing intense punctate LC3A/B staining (white) as a surrogate for PS-1001 distribution. Scale bar, 500 μm. Inset shows a high magnification image from a representative LC3A/B–positive cell. B, Staining with MAC387 (red) and pSTAT1 (green) reveals polarized macrophages (dual positive, yellow) within the region of dying tumor cells, as well as a periphery of MAC387(−)/pSTAT1(+) cells. Scale bar, 500 μm. Inset shows a high magnification image of the dual positive cells. C, Infiltrating CD3-positive T cells (yellow) surround the periphery of the drug exposure zone and the infiltrating macrophage population. The inset shows staining for CD3 (red) and pSTAT1 (green) as a marker of active T cells showing that the pSTAT1-positive population observed at the periphery of the drug exposure zone contains active infiltrating T cells.