Controlled extracellular matrix degradation in breast cancer tumors improves therapy by trastuzumab

Abstract

Extracellular matrix (ECM) in solid tumors affects the effectiveness of therapeutics through blocking of intratumoral diffusion and/or physical masking of target receptors on malignant cells. In immunohistochemical studies of tumor sections from breast cancer patients and xenografts, we observed colocalization of ECM proteins and Her2/neu, a tumor-associated antigen that is the target for the widely used monoclonal antibody trastuzumab (Herceptin). We tested whether intratumoral expression of the peptide hormone relaxin (Rlx) would result in ECM degradation and the improvement of trastuzumab therapy. As viral gene delivery into epithelial tumors with extensive tumor ECM is inefficient, we used a hematopoietic stem cell (HSC)-based approach to deliver the Rlx gene to the tumor. In mouse models with syngeneic breast cancer tumors, HSC-mediated intratumoral Rlx expression resulted in a decrease of ECM proteins and enabled control of tumor growth. Moreover, in a model with Her2/neu-positive BT474-M1 tumors and more treatment-refractory tumors derived from HCC1954 cells, we observed a significant delay of tumor growth when trastuzumab therapy was combined with Rlx expression. Our results have implications for antibody therapy of cancer as well as for other anticancer treatment approaches that are based on T-cells or encapsulated chemotherapy drugs.

Figure 1

Immunohistochemical colocalization of Her2/neu and ECM proteins in breast cancer. (a) Representative sections of a tumor biopsy from a patient with stage III ductal mammary carcinoma. (b) Sections of a biopsy from a patient with stage IV clear cell ovarian cancer. Bar = 40 µm. (c) Confocal microscopy of BT474-M1 tumor cells in vitro. Shown are representative images of stacked XZ and XY sections. Bar = 20 µm. (d,e) Sections of xenograft tumors derived from (d) BT474-M1 and (e) HCC1954 cells. Bar = 40 µm. ECM, extracellular matrix.

Figure 2

Insulated SIN-lentivirus vector for regulated Rlx expression (Ins-SIN-LV-Rlx). (a) Scheme of integrated provirus. The Rlx gene is under the control of a tTR-KRAB system. tRT-KRAB bound to tet-operator sequences represses promoters in the vicinity of 3–4 kb. Addition of Dox releases this repression. The vector also contains a central polypurine tract (cPPT) and a woodchuck hepatitis virus post-transcriptional regulatory element (WPRC). A 0.4-kb cHS4 insulator element is inserted into the 398-bp U3 promoter/enhancer deletion (U3δ). Upon proviral integration into host genome, the U3 region containing the cHS4 is copied over to the 5′ LTR. (b) BT474-M1 cells were transduced with VSV-G-pseudotyped Ins-SIN-LV-Rlx at an MOI of 1. Twenty-four hours after transduction, cells were subjected to limited dilution in 96-well plates. Individual colonies were expanded and treated with Dox for 24 hours. Then, mRNA was isolated and subjected to qRT-PCR for GAPDH and Rlx mRNA. The powered δCt values represent Rlx mRNA levels compared to GAPDH mRNA levels. The right column shows the induction factor upon Dox addition. Dox, doxycycline; LTR, long-terminal repeat; MOI, multiplicity of infection; qRT-PCR, quantitative reverse transcription; Rlx, relaxin; SIN, self-inactivating.

Figure 3

Effect of Dox-induced Rlx expression on trastuzumab killing in vitro. These studies used BT474-M1-Rlx clone #4 (see Figure 2b) and a MDA-MB-231-Rlx clone with a comparable Rlx mRNA expression level. (a) Immunofluorescent analysis of BT474-M1-Rlx clone #4 for laminin and Her2/neu 2 days after Dox addition. (b) In vitro killing of (Her2/neu-positive) BT474-M1-Rlx and (Her2/neu-negative) MDA-MB-231-Rlx cells by trastuzumab. BT474-M1-Rlx cells were grown to confluence, incubated with Dox (1 µg/ml) or PBS for 2 days and then treated with trastuzumab (15 µg/ml) for 30 minutes. Cell viability was measured 2 hours later by WST-1 assay. Viability of PBS-treated cells was taken as 100%. N = 5, *P < 0.05 for BT474-M1-Rlx+trastuzumab w/ Dox versus w/o Dox. Three independent experiments were performed and the data used for statistical analysis. Dox, doxycycline; PBS, phosphate-buffered saline; Rlx, relaxin.

Figure 4

Studies with transplanted mouse HSCs in neu-tg mice carrying MMC tumors. (a) Scheme of experiment. A total of 5 × 10⁵ of mock- or LV-transduced HSCs were transplanted into lethally irradiated neu-tg mice via tail vein injection. Six weeks after HSCs engraftment, MMC tumors were established via injection of 5 × 10⁵ MMC cells subcutaneously. Mice received intraperitoneal injection of PBS or Dox (0.5 mg/mouse in 500 µl PBS) starting at day 7 after MMC cell transplantation and then every other day. (b) Tumor homing of gene modified cells. Left panel: GFP expression in HSCs before transplantation, middle panel: representative MMC tumor section from mice that received LV-GFP transduced HSCs. Right panel: F4/80 and GFP flow cytometry analyses of MMC tumors from mice that received LV-GFP transduced HSCs. Tumors were digested with collagenase to generate single cell suspensions. The gated sections P4 and P5 represent GFP⁺/F4/80⁺ and GFP⁺/F4/40⁻ cells, respectively. Shown is a representative sample. (c) Therapy study with mice that received mock-transduced (upper panels) and Ins-SIN-LV-Rlx-transduced (lower panels) mouse HSCs. Dox or PBS was injected intraperitoneally at day 7 after MMC cell implantation and then every other day. Rlx/Dox− versus Rlx/Dox+: P = 0.0021 for day 19. The P value has been calculated based data from three independent experiments with different numbers of inoculated tumor cells in each experiment. The figure shows the data from animals injected with 5 × 10⁵ MMC cells. (d) Representative sections stained for basement membrane using Jones' periodic acid silver staining method. Basement membrane appears in dark brown/black. Dox, doxycycline; GFP, green fluorescent protein; HSC, hematopoietic stem cell; LV, lentivirus; MMC, mammary carcinoma cell; PBS, phosphate-buffered saline; Rlx, relaxin.

Figure 5

Infiltration of human breast cancer tumors by mouse TAMs. (a) BT474-M1 and (b) HCC1954 tumors were established in CB17-SCID-beige mice. Tumors were harvested at 4 weeks after tumor cell injection and sections analyzed for vascularization using antibodies against mouse CD31 (endothelial marker); for tumor ECM, using antibodies against mouse laminin; for leukocyte infiltration using antibodies against mouse CD45; and for TAM infiltration using antibodies against mouse F4/80. Bar = 40 µm. TAM, tumor-associated macrophage.

Figure 6

Therapy studies in the BT474-M1 xenograft model. CB17-SCID-beige mice were transplanted with either mock- or Ins-SIN-LV-Rlx-transduced HSCs cells [Tx (Mock) and Tx (Rlx), respectively]. After engraftment of HSCs, BT474-M1 cells were implanted into the mammary fat pad. Intraperitoneal Dox or PBS injections started 7 days after tumor cell implantation. Trastuzumab injections started 21 days after BT474-M1 cell transplantation. (a) Tumor growth in Tx (Mock) and Tx (Rlx) mice without trastuzumab treatment. Shown is the relative increase of tumor volume. Tumor volumes at the day on which Dox/PBS injections were started (“b.t.”) were taken as 100%. b.t., before treatment; p.t., post-treatment. N = 7. Shown are the average tumor volumes and SD. (b) Tumor growth in Tx (Mock) and Tx (Rlx) mice with trastuzumab treatment. Note that the scale of the y-axis is different. (c,d) Representative tumor sections of mice without Rlx expression [Tx (Mock)] and with Dox-induced Rlx expression [Tx (Rlx)+Dox] stained with (c) H&E and for (d) Her2/neu. Tumors from trastuzumab-treated mice were either absent or too small and could therefore not be evaluated. (e,f) Representative sections stained for basement membrane using (e) Jones' periodic acid silver staining method and (f) collagen IV. Collagen staining in f appears in brown. Collagen IV stained sections were used for morphometry. Bar = 40 µm. Dox, doxycycline; GFP, green fluorescent protein; HSC, hematopoietic stem cell; LV, lentivirus; PBS, phosphate-buffered saline; Rlx, relaxin; SIN, self-inactivating.

Figure 7

Therapy studies in the HCC1954 model. Mice were treated as described in Figure 6. (a) Tumor growth in mice. Shown is the relative increase of tumor volume. Tumor volumes at the day Dox/PBS injection were started (“b.t.”) were taken as 100%. b.t., before treatment; p.t., post-treatment. N = 5. Shown are the average tumor volumes and SD. In the Tx(Rlx)Dox group, the decrease in tumor volume between days 13 and 15 is not significant. (b) Tumor lysates were subjected to immunoprecipitation with polyclonal anti-Rlx antibodies and analyzed by western blot with monoclonal anti-Rlx antibodies. Representative samples are shown. In agreement with the manufacturer, the monoclonal anti-Rlx antibody reacts with 25 kd and ~50 kd Rlx forms. (c) Representative paraffin sections of tumors from mice without Rlx expression [Tx (Mock)] stained with H&E (left panel) and for Her2/neu (right panel). Tumor sections for the other experimental groups were similar. (d) Immunohistochemistry for collagen IV. Positive signals appear in brown. Bar = 40 µm. Dox, doxycycline; GFP, green fluorescent protein; H&E, hematoxylin and eosin; HSC, hematopoietic stem cell; LV, lentivirus; PBS, phosphate-buffered saline; Rlx, relaxin; SIN, self-inactivating.