PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Cancer Cell. Author manuscript; available in PMC Apr 17, 2013.
Published in final edited form as:
PMCID: PMC3332002
NIHMSID: NIHMS361151
Imaging tumor-stroma interactions during chemotherapy reveals contributions of the microenvironment to resistance
Elizabeth S. Nakasone,1,2,10 Hanne A. Askautrud,1,3,4,5,10 Tim Kees,1 Jae-Hyun Park,1 Vicki Plaks,3 Andrew J. Ewald,3,6 Miriam Fein,1,7 Morten G. Rasch,1,8 Ying-Xim Tan,3 Jing Qiu,1 Juwon Park,1 Pranay Sinha,1 Mina J. Bissell,9 Eirik Frengen,4,5 Zena Werb,3,11 and Mikala Egeblad1,3,11
1Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 11724, USA
2Watson School of Biological Sciences, Cold Spring Harbor, NY 11724, USA
3Department of Anatomy, University of California, San Francisco, San Francisco, CA 94143, USA
4Department of Medical Genetics, Institute of Clinical Medicine, University of Oslo, N-0315 Oslo, Norway
5Department of Medical Genetics, Oslo University Hospital, N-0424 Oslo, Norway
6Departments of Cell Biology and Oncology, Center for Cell Dynamics, Johns Hopkins School of Medicine, Baltimore, MD, 21205, USA
7Graduate Program in Genetics, Stony Brook University, Stony Brook, New York 11794, USA
8Finsen Laboratory, Copenhagen University Hospital, Denmark
9Life Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA
Contact: Mikala Egeblad, Cold Spring Harbor Laboratory, 1 Bungtown Road, Cold Spring Harbor, NY 11724, USA, egeblad/at/cshl.edu
10These authors contributed equally
11These authors contributed equally
Little is known about the dynamics of cancer cell death in response to therapy in the tumor microenvironment. Intravital microscopy of chemotherapy-treated mouse mammary carcinomas allowed us to follow drug distribution, cell death and tumor-stroma interactions. We observed associations between vascular leakage and response to doxorubicin, including improved response in matrix metalloproteinase-9 null mice that had increased vascular leakage. Furthermore, we observed CCR2-dependent infiltration of myeloid cells after treatment and that Ccr2 null host mice responded better to treatment with doxorubicin or cisplatin. These data show that the microenvironment contributes critically to drug response via regulation of vascular permeability and innate immune cell infiltration. Thus, live imaging can be used to gain insights into drug responses in situ.
Keywords: Chemoresistance, breast cancer, chemokine receptors, doxorubicin, intravital imaging, drug distribution, matrix metalloproteinase, myeloid cell infiltrate, tumor microenvironment, vascular permeability
One of the major challenges in treating cancer is resistance to therapy. It is well appreciated that cancer cell intrinsic factors, such as genetic or epigenetic changes, can cause development of therapy resistance (Dean et al., 2005). Extrinsic factors in the microenvironment of certain organs, such as the bone marrow and thymus, can also confer resistance (Gilbert and Hemann, 2010; Meads et al., 2009). In these cases, resistance is mediated by factors secreted from stromal cells such as IL-6. In addition, impaired drug penetration through the extracellular matrix (ECM) influences drug response in primary solid tumors (Loeffler et al., 2006; Olive et al., 2009).
Surprisingly little is known about how cells in intact tumors respond to classical chemotherapies (Minchinton and Tannock, 2006; Rottenberg et al., 2010). Most of the knowledge about these responses has been obtained from cell culture or xenograft experiments where cancer cells grow under conditions very different from the microenvironments of human tumors. Indeed, such experiments are often not predictive of drug responses in patients (Johnson et al., 2001; Minchinton and Tannock, 2006).
To investigate the role of the microenvironment in chemoresponsiveness, we chose a well documented mouse model, the mammary tumor virus (MMTV) promoter-driven polyoma middle T oncogene (PyMT) model, that displays progressive stages of tumorigenesis similar to human luminal type B breast cancer (Herschkowitz et al., 2007; Lin et al., 2003). As MMTV-PyMT tumors progress, cancer cells undergo molecular changes including increased expression of ErbB receptor family members (Lin et al., 2003). In parallel, the stromal microenvironment undergoes changes in blood vessel architecture, ECM composition, and immune cell infiltration (Egeblad et al., 2008; Egeblad et al., 2010).
Doxorubicin induces necrotic cell death in vivo
Doxorubicin (Adriamycin) is a cytotoxic drug used to treat advanced breast cancer (Rouzier et al., 2005). To investigate the acute, cellular response to doxorubicin treatment in the context of the tumor microenvironment, we monitored mammary carcinomas in live mice using spinning disk confocal microscopy (Egeblad et al., 2008). MMTV-PyMT mice were crossbred with the ACTB-ECFP and c-fms-EGFP reporter mice to enable visualization of different tumor stages and tracking of the most abundant stromal cell type, myeloid cells (Egeblad et al., 2008). Dead cells were labeled with propidium iodide (PI) administered intraperitoneally (i.p.). Cell death became apparent 24-30 hr after doxorubicin administration and increased as imaging continued, while there was limited cell death in untreated control mice (Figures 1A, S1A, see Figure 2 for quantification). The death of individual cells was readily observed as the appearance of PI staining (Figure 1B, Movie S1). In addition to cell death in the tumor mass, we observed stromal cell death (Figure 1A).
Figure 1
Figure 1
Visualizing cellular responses to chemotherapy in vivo
Figure 2
Figure 2
Doxorubicin sensitivity changes with tumor stage in vivo but not in vitro
Necrosis is a major pathway of cell death after doxorubicin treatment in vitro (Obeid et al., 2007). Cell death by apoptosis and necrosis are characterized by specific changes in nuclear morphology and plasma membrane integrity (Dive et al., 1992). We next determined the nature of doxorubicin-induced cell death in vivo by imaging changes in nuclear morphology in MMTV-PyMT;ACTB-ECFP mice crossbred with mice expressing histone H2B conjugated to EGFP (Hadjantonakis and Papaioannou, 2004). The majority of the nuclei became positive for PI before their morphology changed, indicating early loss of plasma membrane integrity and necrosis-like cell death. A small percentage of cells died with breakdown of chromatin into condensed bodies followed by acquisition of PI labeling, indicating late loss of plasma membrane integrity and apoptosis-like cell death (Figures 1C, 1D, Movie S2). By histology, we observed only a minor increase in the number of cells with the condensed chromatin that is typical of apoptosis after doxorubicin treatment (Figure S1B). Thus, doxorubicin predominantly induces necrosis in vivo.
Doxorubicin sensitivity changes with tumor stage
At the macroscopic level, doxorubicin treatment reduced tumor volume in MMTV-PyMT mice bearing multiple tumors (Figures 2A, S2A). However, the smallest or largest tumors tended to be resistant (Figure S2B), suggesting that resistance was associated with tumor progression.
MMTV-PyMT tumors have been classified into four pathological stages, based on cellular morphology, lack of basement membrane, and infiltration of immune cells (Lin et al., 2003). We adapted this classification to allow a simplified assessment of tumor stage in live animals: “hyperplasia”, small lesions showing increased accumulation of cells as compared to normal epithelium; “early carcinoma”, lesions with evidence of myeloid cell infiltration and/or cancer cell invasion; and “late carcinoma”, large lesions with densely packed cancer cells (Egeblad et al., 2008). MMTV-PyMT mammary glands have multiple tumors, which are often at different stages of progression. This allowed us to image doxorubicin response at different stages in the same mouse (Figure 2B, S2C). Cell death increased after doxorubicin-treatment in early carcinomas (Figure 2C); on average, 10% of the Fields Of View (FOV) were positive for PI at 31 hr after doxorubicin-treatment compared to 2% of the FOV in the controls (Figure S2D). In contrast, cell death did not increase in hyperplasias and late carcinomas after treatment (Figures 2B, 2C, S2C, S2D, Movie S3), suggesting marked differences in drug sensitivity between tumor stages. Cancer cells in early carcinomas also accumulated more DNA damage after doxorubicin treatment than those in hyperplasias or late carcinomas, as determined by immunostaining for histone γ-H2AX (Figure 2D). Thus, doxorubicin sensitivity changes with tumor stage, but not in a linear manner.
Cancer cell proliferation does not correspond to doxorubicin sensitivity in vivo
Doxorubicin intercalates into DNA, inhibiting topoisomerase II. Hence, proliferating cells are predicted to be more sensitive than resting cells (Campiglio et al., 2003). Yet, cancer cell proliferation differed insignificantly between tumor stages (Figure 2E). Furthermore, in pulse-chase experiments, the percentage of dead cells that were BrdU+ was similar in doxorubicin- and PBS-treated tumors (Figure 2F). These data suggest that differences in cellular proliferation between tumor stages are not the main cause of differences in doxorubicin sensitivity in vivo.
Primary cancer cells from different tumor stages do not exhibit intrinsic differences in doxorubicin sensitivity in culture
As tumors progress, cancer cells acquire mutations and silence or activate genes. These changes can give rise to differences in sensitivity between tumor stages. To distinguish the contribution of cancer cell intrinsic changes from those of the tissue microenvironment, we isolated cancer cells from MMTV-PyMT;ACTB-ECFP mice using a fluorescent dissection microscope. This allowed us to visualize the cancer epithelium during tissue removal, ensuring that tumors were representative of the different stages (Figure S2E). Indeed, in vitro sensitivity to the ErbB1/ErbB2 inhibitor lapatinib was highest for cells isolated from late carcinomas (Figure 2G), as predicted from the increased expression of the ErbB2/Neu oncoprotein in late stage tumors (Lin et al., 2003). In contrast, cells from hyperplasia, early carcinoma, and late carcinoma stages had similar sensitivity to doxorubicin in both 2D and 3D cultures (Figures 2H, 2I, S2F). Primary macrophages and the macrophage cell line RAW 264.7 were also sensitive to doxorubicin (Figure S2G), in accordance with the stromal cell death observed by imaging. Taken together, the in vitro data indicated that cancer cell intrinsic mechanisms are unlikely to be responsible for the differences in doxorubicin sensitivity that were observed between tumor stages and that extrinsic mechanisms are operative in vivo.
Visualization of drug distribution in real-time reveals its association with drug response
Interstitial fluid pressure is one extrinsic factor that increases in late stage tumors as compared to early stages and it hinders drug penetration into tumors (Hagendoorn et al., 2006; Netti et al., 1999). Therefore, we hypothesized that differences in drug response between tumor stages might be related to differences in drug accessibility. Doxorubicin is naturally fluorescent, allowing for imaging of its distribution. However, its broad emission and excitation spectra overlap with those of our fluorescent reporters. It was therefore not possible to image doxorubicin distribution in MMTV-PyMT;ACTB-ECFP mice where tumor stage could be determined. Instead, we imaged unlabeled MMTV-PyMT mice and identified tumor areas by palpation and abnormal tumor vasculature, visualized using intravenously (i.v.) injected fluorescent nanoparticles (AngioSPARK 680).
We imaged small (just palpable), medium (~160 mm3), and large (~845 mm3) tumors (Figure 3A, Movies S4 and S5). Doxorubicin rapidly leaked out into the tissue and accumulated in the nuclei of cells near blood vessels (Figure 3A). This resulted in an intravascular half-life of 9.5 min (range 7.9-11.8 min, Figure 3B), consistent with the initial plasma half-life of 8.5 min (range 7.2-9.8 min) reported in human patients (Greene et al., 1983). Doxorubicin distribution varied among tumors of different sizes (Figures 3A, 3C), with the most leakage in larger tumors and the least leakage in those that were just palpable. Furthermore, significantly more doxorubicin was extractable from small tumors (13-81 mm3) than from hyperplastic MMTV-PyMT tissues or normal mammary glands (Figure 3D). Solid tumors tend to have progressed further than softer ones, and when tumors were analyzed based on consistency, as well as size, the highest levels of doxorubicin were found in soft tumors, although the values were not significantly higher than in solid tumors of similar size (Figure 3E).
Figure 3
Figure 3
Distribution and uptake of doxorubicin in different tumor sizes
Tumor size and consistency is not an accurate indicator of tumor stage. Thus, to quantify vascular leakage in tumors of different stages, we co-injected i.v. fluorescent 10 kD (low) and 2 MD (high) molecular mass dextrans into MMTV-PyMT;ACTB-ECFP mice (Figures 4A, S3A, Movie S6). The intravascular half-lives of dextrans were similar in tumor microenvironments at different stages (Figures S3B, S3C), with a higher intravascular half-life for 2 MD dextran (41 [range 32-58] min) than for 10 kD dextran (15 [range 14-16] min, Figure S3D).
Figure 4
Figure 4
Leakage and distribution of intravenously injected dextran varies with tumor stage
In contrast to intravascular half-life, leakage of dextrans into the extravascular space and their retention in tissues were clearly influenced by tumor stage (Figures 4A, 4B). Regions with hyperplasias showed limited leakage into the extravascular space, while early carcinomas and the tumor-stroma borders of late carcinomas exhibited substantially more leakage. Extravasated 10 kD dextran reached a larger area of the tissue than did extravasated 2 MD dextran (Figures 4A, 4B). Following extravasation, the dextrans accumulated in stromal cells, which we have previously shown are c-fms+CD68+CD206+ macrophages (Egeblad et al., 2008). Leakage in normal mammary glands was minimal and below detection limits when the settings used for tumor-bearing mice were applied (not shown).
Finally, we co-injected i.v. doxorubicin-treated mice with 10 kD dextran and Ricinus communis agglutinin I, a lectin that binds to basement membrane exposed to the vascular lumen, marking leaky vasculature (Thurston et al., 1999). We observed that only tumor areas in close proximity to these two markers contained cells with nuclear uptake of doxorubicin (Figure S3E). Interestingly, some of the cells with doxorubicin-positive nuclei had also taken up dextran, suggesting that they were macrophages (Figure S3E).
Absence of MMP9 results in increased vascular leakage and sensitizes tumors to doxorubicin treatment
Tumor-associated macrophages promote vascular changes (Egeblad et al., 2010). We found that the extent of local c-fms-EGFP+ myeloid cell infiltration correlated significantly with the degree of local 10 kD dextran leakage, regardless of tumor stage (Figures 4C, 4D). Myeloid cells can regulate vessel stability via secretion of vascular endothelial growth factor (VEGF) and transforming growth factor (TGF)-β, both of which are sequestered in the ECM and released by matrix metalloproteinases (MMPs) (Yu and Stamenkovic, 2000; Ebrahem et al., 2010). Nevertheless, MMPs may also negatively regulate vascular permeability (Sounni et al., 2010). Several MMPs, including MMP9, are expressed at high levels by tumor-infiltrating myeloid cells (Figure 5A). To determine if host-derived MMP9 affects vascular stability and doxorubicin response, we transplanted MMTV-PyMT cancer cells to wild type FVB/n and Mmp9−/− hosts and treated them with doxorubicin. Although MMP9 does not influence primary tumor growth in the transgenic MMTV-PyMT model (Martin et al., 2008), we found that the growth of grafted tumors was reduced in Mmp9−/− hosts compared to FVB/n hosts (Figure S4A). Interestingly, tumors in Mmp9−/− hosts responded better to doxorubicin than those in FVB/n hosts (Figure S4B).
Figure 5
Figure 5
Matrix metalloproteinase-9 is expressed by myeloid cells and influences vascular leakage and response to doxorubicin
Since transplanted MMTV-PyMT tumors grew slower in Mmp9−/− hosts, we could not rule out that the influence of MMP9 on doxorubicin response was related to its effects on cancer cell proliferation. Therefore, we tested if MMP9 influenced the response in the MMTV-Neu model, another doxorubicin-sensitive, luminal breast cancer model, in which the absence of MMP9 does not affect cancer cell proliferation or tumor growth (Figures S4C, S4D, and data not shown). Vascular volume, determined by i.v. tomato lectin staining, was not affected by MMP9 status, but tumor vessels were leakier in the absence of MMP9, as determined by i.v. staining with Ricinus communis agglutinin I (Figures 5B-D). In the absence of MMP9, the endothelial cell adhesion molecule VE-cadherin showed increased phosphorylation (Figures 5E, 5F), which results in loose adherence junctions (Gavard, 2009). Vascular coverage with pericytes, which support endothelial cells, was also decreased (Figures 5G, S4E). In keeping with the increased permeability of the vasculature, MMTV-Neu;Mmp9−/− tumors treated with doxorubicin responded better than MMTV-Neu;Mmp9+/+ tumors (Figure 5H). In contrast, C3(1)-Tag tumors, a basal-like mammary carcinoma model that is largely doxorubicin-resistant (Figure S4F), did not respond better in the absence of MMP9 (Figure S4G).
Doxorubicin treatment leads to recruitment of myeloid cells to tumors
We observed that vascular leakage was associated with higher infiltration of macrophages (Figure 4D), so we sought to determine the effect of MMP9 on macrophage infiltration. There were increased numbers of cells expressing the F4/80 and CD206 macrophage markers in untreated tumors of Mmp9−/− hosts compared to those of FVB/n hosts (Figures 6A, 6B). Interestingly, doxorubicin treatment significantly reduced the number of macrophages in Mmp9−/− hosts, suggesting that macrophages were killed.
Figure 6
Figure 6
Myeloid cells are recruited to areas of tumor necrosis
Despite the reduction in macrophage numbers observed after doxorubicin-treatment of Mmp9−/− hosts, imaging consistently showed recruitment of myeloid cells to doxorubicin-treated tumors in wild type mice (Figures 6C, 6D, Movie S7). These myeloid cells often formed granuloma-like structures in areas with cell death (Movie S7). Rarely, we observed myeloid cell recruitment prior to cancer cell death (Figure S5A).
Myeloid cells are a diverse family of innate immune cells, including neutrophils, monocytes and macrophages, that are all recruited to areas of cell death (Murdoch et al., 2004). To determine a basis for myeloid cell infiltration, we injected necrotic cell debris into mammary glands of non-tumor-bearing mice and imaged the response. In this model, the results are independent of any direct effects of doxorubicin on myeloid cells, such as induction of cell death. Myeloid cell infiltration increased in areas injected with necrotic debris as compared to areas injected with only saline and dextran (Figure S5B). Once a myeloid cell recognized the cell debris, other myeloid cells were rapidly recruited and a granuloma-like structure was formed (Figure S5B, Movie S8). Myeloid cell recruitment is often mediated by Gi-protein coupled chemokine receptors (Sadik et al., 2011). Mice pretreated with pertussis toxin, a Gi-protein inhibitor, showed reduced myeloid cell recruitment to cell debris, suggesting the involvement of chemokine receptors in this process (Figure S5C, Movie S9).
Myeloid cells are recruited to doxorubicin-treated tumors through a stromal CCL2/CCR2 chemokine/chemokine receptor axis
To screen for candidate chemokines involved in the recruitment of myeloid cells, we used a protein array. Tumor lysates isolated 48 hr after doxorubicin treatment showed increased protein levels for CCL2 and CCL12 (Figure 7A). Both of these chemokines are ligands for the CCR2 receptor, which is expressed on monocytes and is responsible for their recruitment to sites of inflammation (Tsou et al., 2007). We also observed small increases in macrophage colony stimulating factor (M-CSF/CSF1) and tissue inhibitor of metalloproteinase 1 (TIMP1) (Figure 7A). We confirmed the increases in CCL2 and CCL12 by ELISA and immunostaining (Figures 7B, S6A, S6B). CCL2-expressing cells were large cells located in the stroma that did not express markers of fibroblasts and pericytes (Figure 7C) or endothelial cells (Figure 7D).
Figure 7
Figure 7
Myeloid cells are recruited to doxorubicin-treated tumors through a stromal CCL2/CCR2 chemokine/chemokine receptor axis
Next, we investigated the types of myeloid cells recruited after doxorubicin treatment. Infiltration of cells expressing a neutrophil/monocyte marker, the 7/4 antigen (Ly6B.2), was significantly increased (Figure 7E). This increase was exclusively found among cells that also expressed the CCL2 receptor, CCR2, and had a monocytic nuclear morphology (Figures 7E, S6C). The number of 7/4+CCR2- cells did not increase, but rather decreased, after doxorubicin treatment (Figure 7E). A small number of CCR2+ cells did not express the 7/4 antigen, but their numbers did not change after doxorubicin treatment (Figure 7E). In contrast to the acute increase in the infiltration of 7/4+CCR2+ cells, overall macrophage infiltration, as determined by the F4/80 marker, was not changed 48 hr after doxorubicin treatment (Figure S6D). However, a small increase was seen in the number of cells within the tumor mass that expressed CD206, a marker of alternatively activated macrophages (Figure S6E).
We next tested whether CCR2 mediates the myeloid cell infiltration seen after doxorubicin treatment. We transplanted primary cancer cells isolated from MMTV-PyMT mice to Ccr2−/− or C57BL/6 hosts and characterized the myeloid cell population after doxorubicin treatment. In C57BL/6 hosts, the fraction of cells co-expressing the myeloid cell marker CD11b with the monocyte/neutrophil markers Gr1 and 7/4 increased significantly after doxorubicin treatment (Figures 7F, 7G). The total proportion of cells in the tumors that expressed the myeloid CD11b marker did not increase (Figure S6F). Tumors of Ccr2−/− hosts had a higher fraction of CD11b+7/4+Gr1+ cells than did C57BL/6 hosts (Figures 7F, 7G), consistent with previous reports (Pahler et al., 2008). However, there was no significant increase in this cell population after doxorubicin treatment in Ccr2−/− hosts (Figures 7F, 7G). Doxorubicin treatment did not affect the subpopulations of CD11b+ cells that expressed F4/80 or CXCR4 in either genotype (Figures S6G, S6H, S6I). Together, our results suggest that doxorubicin treatment leads to a specific, acute recruitment of CCR2-expressing myeloid cells of the monocytic lineage through stromally expressed CCL2.
Host CCR2 influences tumor response to chemotherapy
To determine if CCR2-mediated recruitment of myeloid cells influenced the response to doxorubicin, we generated cohorts of MMTV-PyMT;Ccr2+/− and MMTV-PyMT;Ccr2−/− mice. For doxorubicin-sensitive tumors (250-750 mm3), the absence of CCR2 was associated with a significantly better response, although tumors in both cohorts ultimately relapsed (Figure 8A). In contrast, tumors below 250 mm3, which respond poorly to doxorubicin, did not show a better response to treatment when mice were deficient for CCR2 (Figure S7A).
Figure 8
Figure 8
Host CCR2 regulates response to doxorubicin
CCL2 has been proposed to promote cancer cell survival through CCR2 receptors on cancer cells (Zhang et al., 2010). To rule out the involvement of CCR2 signaling in cancer cells, we transplanted primary cancer cells from MMTV-PyMT mice to Ccr2−/− or C57BL/6 hosts. When treated with doxorubicin or cisplatin, a different class of chemotherapeutic drug with efficacy against MMTV-PyMT tumors (Figure S7B), tumors in Ccr2−/− hosts relapsed later than those in C57BL/6 hosts (Figures 8B, 8C). Tumor growth before treatment was not significantly different between Ccr2−/− and C57BL/6 hosts (Figure 8B). Myeloid cells can secrete factors that increase cancer cell survival (Shree et al., 2011). However, we detected no lasting survival benefit on primary cancer cells when they were cultured with primary macrophages (Figures S7C, S7D).
Although tumors in Ccr2−/− hosts relapsed, they were of a strikingly lower histological grade than the high-grade, relapsed tumors of C57BL/6 hosts (Figure 8D). Histological differences were already evident 48 hr after doxorubicin treatment as tumors in Ccr2−/− hosts were more cystic than those in C57BL/6 hosts (Figure 8D). Taken together, our data suggest that infiltration of CCR2-expressing myeloid cells into chemotherapy-treated tumors contributes to tumor regrowth after treatment.
Like CCR2, MMP9 can also influence myeloid cell recruitment (Gong et al., 2008). We therefore tested whether the effects of MMP9 on doxorubicin response could be explained by effects on the acute recruitment of myeloid cells to treated tumors. However, adaptively transferred bone marrow cells from c-fms-EGFP;Mmp9+/+ and c-fms-EGFP;Mmp9−/− mice were recruited equally well to tumors of MMTV-PyMT mice after treatment with doxorubicin (Figure S7E). Next, we tested whether the changed myeloid cell environment in Ccr2−/− mice correlated with vascular changes, akin to the effects of loss of Mmp9 (Figure 8E). Whereas the vascular volume was higher in tumors of Ccr2−/− mice (Figure 8F), the percentage that was leaky was significantly lower, with the net result being no change in the total volume of leaky vasculature (Figures 8G, 8H). Furthermore, pericyte coverage of the vasculature was increased in Ccr2−/− hosts, while VE-cadherin phosphorylation levels were unaffected (Figures S7F-I). Thus, although the response to doxorubicin is improved in both Mmp9−/− and Ccr2−/− microenvironments, the effects on vascular structure and myeloid cell recruitment after treatment are different.
We imaged the dynamics of doxorubicin responses in a progressive cancer model and revealed that chemosensitivity in vivo is influenced by the tumor microenvironment. The highest sensitivity to doxorubicin was not observed in the earliest or latest tumor stages, but rather in the intermediate stage. These differences in drug response between stages were associated with parallel differences in vascular leakage. We therefore tested doxorubicin response in Mmp9 null tumor microenvironments with increased vascular permeability and found an improved response. Doxorubicin treatment led to the recruitment of CCR2+ monocytic cells to tumors. In Ccr2 null microenvironments, recruitment was inhibited and the response to chemotherapy better. These results have clinical implications, as myeloid cell infiltration is increased in human breast tumors after chemotherapy and the composition of the immune infiltrate is a predictor of survival (Denardo et al., 2011).
Stromal regulation of vascular leakage
Primary cancer cells from different stages responded similarly to doxorubicin in vitro, whereas doxorubicin sensitivity, tumor stage, and vascular leakage were related in vivo. This suggests that stage-specific changes in vascular leakage contributed to doxorubicin response, although we cannot exclude contributions from stromal cells. Strikingly, increased leakage of the tumor vasculature in Mmp9−/− mice corresponded to a better response to doxorubicin. The increased vascular permeability in Mmp9−/− mice was associated with decreased pericyte-coverage of the vasculature and increased phosphorylation of VE-cadherin, which affect endothelial cell-cell adhesions (Gavard, 2009; Goel et al., 2011). Although MMP inhibitors have failed in clinical trials (Coussens et al., 2002), our data suggest that these, or other drugs that affect vascular permeability, could be used to achieve better responses to chemotherapies.
MMP9 may regulate vascular structure by acting on a substrate in the tumor microenvironment. Indeed, a reduction of infiltrating MMP9+ myeloid cells through inhibition of the CSF-1 receptor results in a better response to anti-VEGFR2 treatment (Priceman et al., 2010). However, MMP9 may also regulate vascular permeability indirectly through its effects on macrophage infiltration. Macrophages can secrete VEGF (also known as vascular permeability factor), which induces VE-cadherin phosphorylation in endothelial cells (Gavard, 2009).
Myeloid cells are not the only stromal cells capable of regulating drug delivery in tumors. By decreasing the fibroblast pool, and thus the ECM, drugs enter the tissue better (Loeffler et al., 2006; Olive et al., 2009). Interestingly, the improved drug delivery in these cases is achieved by increased vascular density and “vascular normalization”, which involves better pericyte coverage and reduced permeability (Goel et al., 2011). Since vascular volume and pericyte coverage increased in tumors of Ccr2−/− compared to C57BL/6 hosts, such vascular changes could play a role in the MMTV-PyMT model. Untangling the roles of MMP9 and CCR2 on vasculature, myeloid cell infiltration, and drug response will require the ability to temporally and conditionally manipulate these genes.
The role of myeloid cell recruitment in chemotherapy responses
Recruitment of CCR2+ monocytes after doxorubicin treatment correlated with tumor relapse. Similarly, an increased number of macrophages is found after treatment of tumors with chemotherapy and is associated with poor drug response (Denardo et al., 2011; Shree et al., 2011). Such macrophages can promote cancer cell survival through the secretion of cysteine cathepsins (Shree et al., 2011). CCR2+ monocytes could be the source of tumor-associated macrophages in the post-treatment microenvironment. Indeed, we did not observe increased numbers of macrophages 48 hr after treatment, while they have been observed 7-12 days after the first dose of chemotherapy in the MMTV-PyMT model (Denardo et al., 2011; Shree et al., 2011). However, macrophage infiltration depends on CSF-1 (Denardo et al., 2011). Furthermore, macrophages do not promote survival of cancer cells in a cathepsin-dependent manner after treatment with the cisplatin-related drug carboplatin (Shree et al., 2011), whereas the response to cisplatin was better in Ccr2−/− hosts. Thus, it is possible that CCR2+ monocytes and tumor-associated macrophages are recruited through independent pathways and influence drug responses through independent mechanisms. Several such mechanisms likely exist. For example, Tie-2-expressing macrophages are recruited after hypoxic tissue injury through a CXCL12/CXCR4 chemokine axis (Welford et al., 2011).
Our results, as well as those of others (Ahn et al., 2010; Denardo et al., 2011; Shree et al., 2011; Welford et al., 2011), indicate that the myeloid cell infiltration that occurs after chemotherapy, radiation or tissue injury impedes the response to therapy. However, myeloid cell recruitment can also lead to the direct killing of cancer cells (Guerriero et al., 2011), thereby increasing the response to chemotherapy. These differences in the effects of recruited myeloid cells are likely due to the recruitment of different subpopulations of myeloid cells. Indeed, infiltration of CD206+ macrophages was associated with increased vascular leakage and better doxorubicin response, while reduced CCR2-dependent recruitment of monocytic cells was associated with delayed tumor relapse.
Interestingly, stromal cells expressed CCL2 after doxorubicin treatment. This offers a possible explanation for the observation that stromally derived, but not cancer cell-derived, CCL2 is associated with decreased relapse-free survival in breast cancer patients (Fujimoto et al., 2009). Furthermore, administration of antibodies specific for mouse CCL2 enhances the response to docetaxol in a xenograft model of prostate cancer (Loberg et al., 2007).
Conclusions
In vivo imaging of tumors shows that different components of the microenvironment participate in the development of chemoresistance. Disruption of these microenvironments is beneficial for the response to doxorubicin and cisplatin. Our data suggest that existing drugs that inhibit MMPs or chemokine signaling may be effective when combined with traditional chemotherapies. The order and timing of administration of such combination therapies could, however, be critical, because of the complexity of the interactions between myeloid cells and vasculature in chemotherapy responses. Future studies combining imaging with molecular approaches hold promise for gaining further insights into the targeting of tumors in the context of its microenvironment.
Animals
MMTV-PyMT (FVB/n), MMTV-Neu (FVB/n), ACTB-ECFP (FVB/n), ACTB-H2B-EGFP (obtained on mixed background and backcrossed to FVB/n for 6 generations) and Ccr2−/− (C57BL/6) mice were from Jackson Laboratory. MMTV-PyMT (C57BL/6) mice were provided by Dr. Kasper Almholt and c-fms-EGFP mice were provided by Dr. Jeffrey Pollard and backcrossed to FVB/n mice for 6 generations. Mmp9−/− mice (FVB/n) were previously described (Vu et al., 1998). All animal experiments were conducted in accordance with procedures approved by the IACUC at Cold Spring Harbor Laboratory or the University of California, San Francisco.
Tumor transplantation experiments
Virgin females of 6-16 weeks of age were used as hosts for transplantation. Cancer cells were isolated from 2-3 tumors at 8-10 mm diameter from MMTV-PyMT mice. Tumors were mechanically dissociated and digested with collagenase (0.2% w/v), trypsin (0.2% w/v) and DNase I (8 U/ml) in RPMI-1640 medium. Single cells and debris were removed from the resulting carcinoma organoid preparation by differential centrifugation. Purified carcinoma organoids were dissociated into single cell suspension in 0.25% trypsin with 0.1% EDTA and washed in PBS. Cells (4×105 in 20 μl PBS) were injected into the inguinal mammary glands of host mice.
Tumor response to doxorubicin and cisplatin
Mice received 8 mg/kg doxorubicin hydrochloride (in PBS, Sigma-Aldrich) or 10 mg/kg cisplatin (in 10% DMSO in PBS, MBL International) i.p. on days 0, 7 and 14. Control mice were injected with sterile PBS. Tumors were measured 2-3 times a week by caliper, and tumor volumes were calculated as length × width2/2.
Spinning disk confocal imaging of live mice
Details of the microscope design and imaging procedure were previously described (Egeblad et al., 2008). To track cell death, mice were injected i.p. with 50 μl/h propidium iodide (PI, 0.05-0.1 mg/ml, Invitrogen) in sterile PBS. To determine doxorubicin distribution, MMTV-PyMT mice were injected i.v. with AngioSPARK 680 (VisEn Medical, 100 μl of stock solution) and doxorubicin (8 mg/kg body weight in 200 μl PBS). To determine vascular leakage, mice were injected i.v. with 100 μl sterile PBS containing 1 mg/ml 10 kD Alexa-Fluor-647-conjugated dextran and 1 mg/ml 2 MD rhodamine-conjugated dextran (Invitrogen).
Histology and immunostaining
Paraformaldehyde-fixed, paraffin-embedded sections were stained with Mayer’s hematoxylin and eosin. Immunostaining was done with primary antibodies against 7/4 (Cedarlane), BrdU and MECA-32 (Developmental Studies Hybridoma Bank, University of Iowa), CCL2, CD206, and F4/80 (AbD Serotec), CCR2 (Novus Biologicals), phospho-histone H2AX and phospho-histone H3 (Cell Signaling), αSMA (Sigma-Aldrich), CD31 and phospho-VE-cadherin (Abcam), and MMP9 (Rasch et al., 2010). Immunostained slides were quantified by counting (for 7/4, BrdU, CCL2, CCR2, CD206, phospho-histone H2AX, and phospho-histone H3) or by fluorescence intensity with Volocity software (PerkinElmer; for F4/80, phospho-VE-cadherin, αSMA, and CD31).
In vitro drug sensitivity
Tumor lesions were isolated from MMTV-PyMT;ACTB-ECFP mice using a fluorescent dissection microscope, primary mammary organoids isolated from different tumor stages and cultured in Growth Factor Reduced Matrigel (BD Biosciences). To determine doxorubicin sensitivity, treatment of tumor organoids was started 72 hr after isolation. To determine lapatinib sensitivity, single cell suspensions were generated from tumor organoids, isolated as described above, grown to subconfluence, re-seeded and treated with lapatinib (Selleck Chemicals, LLC). Cell viability was measured 48 hr after addition of drugs using the CellTiter 96 Aqueous One Solution Cell Proliferation Assay (Promega).
For further experimental details, see Supplementary Experimental Procedures.
SIGNIFICANCE
It is well appreciated that intrinsic factors regulate chemoresponsiveness. Using in vivo microscopy of tumors, we show that extrinsic factors within the microenvironment support the development of chemoresistance by regulating drug distribution and the inflammatory response. These results have clinical implications, as myeloid cell infiltration is increased in human breast tumors after chemotherapy and the cellular composition of the immune infiltrate is a strong predictor of overall survival. Our data suggest that the response to classical chemotherapeutic drugs can be improved by changing the tumor microenvironment with agents that modify matrix metalloproteinase activity and chemokine signaling. Our study shows that observing the tumor response in real-time at cellular resolution can yield insights into the biology of chemoresistance.
Supplementary Material
01
10
02
03
04
07
08
09
ACKNOWLEDGEMENTS
We thank E. Atamaniuc, Y. Yu, H. Capili, J. Cappellani, G. DiMino, M.B. Ebert, J. Waage, J. Paterek and the Shared Resources at CSHL for technical support. Dr. Claire Lewis is thanked for helpful comments on the manuscript. Antibodies against BrdU and MECA-32 were obtained from the Developmental Studies Hybridoma Bank maintained by the University of Iowa. This work was supported by funds from the National Cancer Institute (U01 CA141451 to M.E., R01 CA057621 to Z.W. and M.J.B., and P50 CA088843 to A.J.E.), the Starr Cancer Consortium, the Breast Cancer Alliance, Susan G. Komen for the Cure, Long Island 2 Day Walk to Fight Breast Cancer and Manhasset Women’s Coalition Against Breast Cancer to M.E., the Stand Up to Cancer-American Association for Cancer Research Dream Team Translational Cancer Research Grant (SU2C-AACR-DT0409 to Z.W.), a predoctoral fellowship from the Congressionally Directed Breast Cancer Research Program, U.S. (E.S.N.), the Research Council of Norway (160698/V40 and 151882 [FUGE] to E.F.), and Southeastern Regional Health Authorities (2007060 to E.F.). E.S.N. is the recipient of the Leslie C. Quick and William Randolph Hearst Foundation Fellowships from the Watson School of Biological Sciences. E.F. and H.A.A. were supported by “University of Oslo Research Fund (UNIFOR)” and Ullevål University Hospital Research Fund (VIRUUS).
Footnotes
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
  • Ahn GO, Tseng D, Liao CH, Dorie MJ, Czechowicz A, Brown JM. Inhibition of Mac-1 (CD11b/CD18) enhances tumor response to radiation by reducing myeloid cell recruitment. Proc Natl Acad Sci U S A. 2010;107:8363–8368. [PubMed]
  • Campiglio M, Somenzi G, Olgiati C, Beretta G, Balsari A, Zaffaroni N, Valagussa P, Menard S. Role of proliferation in HER2 status predicted response to doxorubicin. Int J Cancer. 2003;105:568–573. [PubMed]
  • Coussens LM, Fingleton B, Matrisian LM. Matrix metalloproteinase inhibitors and cancer: trials and tribulations. Science. 2002;295:2387–2392. [PubMed]
  • Dean M, Fojo T, Bates S. Tumour stem cells and drug resistance. Nat Rev Cancer. 2005;5:275–284. [PubMed]
  • Denardo DG, Brennan DJ, Rexhepaj E, Ruffell B, Shiao SL, Madden SF, Gallagher WM, Wadhwani N, Keil SD, Junaid SA, et al. Leukocyte Complexity Predicts Breast Cancer Survival and Functionally Regulates Response to Chemotherapy. Cancer Discov. 2011;1:54–67. [PMC free article] [PubMed]
  • Dive C, Gregory CD, Phipps DJ, Evans DL, Milner AE, Wyllie AH. Analysis and discrimination of necrosis and apoptosis (programmed cell death) by multiparameter flow cytometry. Biochim Biophys Acta. 1992;1133:275–285. [PubMed]
  • Ebrahem Q, Chaurasia SS, Vasanji A, Qi JH, Klenotic PA, Cutler A, Asosingh K, Erzurum S, Anand-Apte B. Cross-talk between vascular endothelial growth factor and matrix metalloproteinases in the induction of neovascularization in vivo. Am J Pathol. 2010;176:496–503. [PubMed]
  • Egeblad M, Ewald AJ, Askautrud HA, Truitt ML, Welm BE, Bainbridge E, Peeters G, Krummel MF, Werb Z. Visualizing stromal cell dynamics in different tumor microenvironments by spinning disk confocal microscopy. Dis Model Mech. 2008;1:155–167. [PMC free article] [PubMed]
  • Egeblad M, Nakasone ES, Werb Z. Tumors as organs: complex tissues that interface with the entire organism. Dev Cell. 2010;18:884–901. [PMC free article] [PubMed]
  • Fujimoto H, Sangai T, Ishii G, Ikehara A, Nagashima T, Miyazaki M, Ochiai A. Stromal MCP-1 in mammary tumors induces tumor-associated macrophage infiltration and contributes to tumor progression. Int J Cancer. 2009;125:1276–1284. [PubMed]
  • Gavard J. Breaking the VE-cadherin bonds. FEBS Lett. 2009;583:1–6. [PubMed]
  • Gilbert LA, Hemann MT. DNA damage-mediated induction of a chemoresistant niche. Cell. 2010;143:355–366. [PMC free article] [PubMed]
  • Goel S, Duda DG, Xu L, Munn LL, Boucher Y, Fukumura D, Jain RK. Normalization of the vasculature for treatment of cancer and other diseases. Physiol Rev. 2011;91:1071–1121. [PMC free article] [PubMed]
  • Gong Y, Hart E, Shchurin A, Hoover-Plow J. Inflammatory macrophage migration requires MMP-9 activation by plasminogen in mice. J Clin Invest. 2008;118:3012–3024. [PMC free article] [PubMed]
  • Greene RF, Collins JM, Jenkins JF, Speyer JL, Myers CE. Plasma pharmacokinetics of adriamycin and adriamycinol: implications for the design of in vitro experiments and treatment protocols. Cancer Res. 1983;43:3417–3421. [PubMed]
  • Guerriero JL, Ditsworth D, Catanzaro JM, Sabino G, Furie MB, Kew RR, Crawford HC, Zong WX. DNA alkylating therapy induces tumor regression through an HMGB1-mediated activation of innate immunity. J Immunol. 2011;186:3517–3526. [PMC free article] [PubMed]
  • Hadjantonakis AK, Papaioannou VE. Dynamic in vivo imaging and cell tracking using a histone fluorescent protein fusion in mice. BMC Biotechnol. 2004;4:33. [PMC free article] [PubMed]
  • Hagendoorn J, Tong R, Fukumura D, Lin Q, Lobo J, Padera TP, Xu L, Kucherlapati R, Jain RK. Onset of abnormal blood and lymphatic vessel function and interstitial hypertension in early stages of carcinogenesis. Cancer Res. 2006;66:3360–3364. [PubMed]
  • Herschkowitz JI, Simin K, Weigman VJ, Mikaelian I, Usary J, Hu Z, Rasmussen KE, Jones LP, Assefnia S, Chandrasekharan S, et al. Identification of conserved gene expression features between murine mammary carcinoma models and human breast tumors. Genome Biol. 2007;8:R76. [PMC free article] [PubMed]
  • Johnson JI, Decker S, Zaharevitz D, Rubinstein LV, Venditti JM, Schepartz S, Kalyandrug S, Christian M, Arbuck S, Hollingshead M, Sausville EA. Relationships between drug activity in NCI preclinical in vitro and in vivo models and early clinical trials. Br J Cancer. 2001;84:1424–1431. [PMC free article] [PubMed]
  • Lin EY, Jones JG, Li P, Zhu L, Whitney KD, Muller WJ, Pollard JW. Progression to malignancy in the polyoma middle T oncoprotein mouse breast cancer model provides a reliable model for human diseases. Am J Pathol. 2003;163:2113–2126. [PubMed]
  • Loberg RD, Ying C, Craig M, Day LL, Sargent E, Neeley C, Wojno K, Snyder LA, Yan L, Pienta KJ. Targeting CCL2 with systemic delivery of neutralizing antibodies induces prostate cancer tumor regression in vivo. Cancer Res. 2007;67:9417–9424. [PubMed]
  • Loeffler M, Kruger JA, Niethammer AG, Reisfeld RA. Targeting tumor-associated fibroblasts improves cancer chemotherapy by increasing intratumoral drug uptake. J Clin Invest. 2006;116:1955–1962. [PMC free article] [PubMed]
  • Martin MD, Carter KJ, Jean-Philippe SR, Chang M, Mobashery S, Thiolloy S, Lynch CC, Matrisian LM, Fingleton B. Effect of ablation or inhibition of stromal matrix metalloproteinase-9 on lung metastasis in a breast cancer model is dependent on genetic background. Cancer Res. 2008;68:6251–6259. [PMC free article] [PubMed]
  • Meads MB, Gatenby RA, Dalton WS. Environment-mediated drug resistance: a major contributor to minimal residual disease. Nat Rev Cancer. 2009;9:665–674. [PubMed]
  • Minchinton AI, Tannock IF. Drug penetration in solid tumours. Nat Rev Cancer. 2006;6:583–592. [PubMed]
  • Murdoch C, Giannoudis A, Lewis CE. Mechanisms regulating the recruitment of macrophages into hypoxic areas of tumors and other ischemic tissues. Blood. 2004;104:2224–2234. [PubMed]
  • Netti PA, Hamberg LM, Babich JW, Kierstead D, Graham W, Hunter GJ, Wolf GL, Fischman A, Boucher Y, Jain RK. Enhancement of fluid filtration across tumor vessels: implication for delivery of macromolecules. Proc Natl Acad Sci U S A. 1999;96:3137–3142. [PubMed]
  • Obeid M, Tesniere A, Ghiringhelli F, Fimia GM, Apetoh L, Perfettini JL, Castedo M, Mignot G, Panaretakis T, Casares N, et al. Calreticulin exposure dictates the immunogenicity of cancer cell death. Nat Med. 2007;13:54–61. [PubMed]
  • Olive KP, Jacobetz MA, Davidson CJ, Gopinathan A, McIntyre D, Honess D, Madhu B, Goldgraben MA, Caldwell ME, Allard D, et al. Inhibition of Hedgehog signaling enhances delivery of chemotherapy in a mouse model of pancreatic cancer. Science. 2009;324:1457–1461. [PMC free article] [PubMed]
  • Pahler JC, Tazzyman S, Erez N, Chen YY, Murdoch C, Nozawa H, Lewis CE, Hanahan D. Plasticity in tumor-promoting inflammation: impairment of macrophage recruitment evokes a compensatory neutrophil response. Neoplasia. 2008;10:329–340. [PMC free article] [PubMed]
  • Priceman SJ, Sung JL, Shaposhnik Z, Burton JB, Torres-Collado AX, Moughon DL, Johnson M, Lusis AJ, Cohen DA, Iruela-Arispe ML, Wu L. Targeting distinct tumor-infiltrating myeloid cells by inhibiting CSF-1 receptor: combating tumor evasion of antiangiogenic therapy. Blood. 2010;115:1461–1471. [PubMed]
  • Rasch MG, Lund IK, Illemann M, Hoyer-Hansen G, Gardsvoll H. Purification and characterization of recombinant full-length and protease domain of murine MMP-9 expressed in Drosophila S2 cells. Protein Expr Purif. 2010;72:87–94. [PubMed]
  • Rottenberg S, Pajic M, Jonkers J. Studying drug resistance using genetically engineered mouse models for breast cancer. Methods Mol Biol. 2010;596:33–45. [PubMed]
  • Rouzier R, Perou CM, Symmans WF, Ibrahim N, Cristofanilli M, Anderson K, Hess KR, Stec J, Ayers M, Wagner P, et al. Breast cancer molecular subtypes respond differently to preoperative chemotherapy. Clin Cancer Res. 2005;11:5678–5685. [PubMed]
  • Sadik CD, Kim ND, Luster AD. Neutrophils cascading their way to inflammation. Trends Immunol. 2011;32:452–460. [PMC free article] [PubMed]
  • Shree T, Olson OC, Elie BT, Kester JC, Garfall AL, Simpson K, Bell-McGuinn KM, Zabor EC, Brogi E, Joyce JA. Macrophages and cathepsin proteases blunt chemotherapeutic response in breast cancer. Genes Dev. 2011;25:2465–2479. [PubMed]
  • Sounni NE, Dehne K, van Kempen L, Egeblad M, Affara NI, Cuevas I, Wiesen J, Junankar S, Korets L, Lee J, et al. Stromal regulation of vessel stability by MMP14 and TGFbeta. Dis Model Mech. 2010;3:317–332. [PubMed]
  • Thurston G, Suri C, Smith K, McClain J, Sato TN, Yancopoulos GD, McDonald DM. Leakage-resistant blood vessels in mice transgenically overexpressing angiopoietin-1. Science. 1999;286:2511–2514. [PubMed]
  • Tsou CL, Peters W, Si Y, Slaymaker S, Aslanian AM, Weisberg SP, Mack M, Charo IF. Critical roles for CCR2 and MCP-3 in monocyte mobilization from bone marrow and recruitment to inflammatory sites. J Clin Invest. 2007;117:902–909. [PMC free article] [PubMed]
  • Vu TH, Shipley JM, Bergers G, Berger JE, Helms JA, Hanahan D, Shapiro SD, Senior RM, Werb Z. MMP-9/gelatinase B is a key regulator of growth plate angiogenesis and apoptosis of hypertrophic chondrocytes. Cell. 1998;93:411–422. [PMC free article] [PubMed]
  • Welford AF, Biziato D, Coffelt SB, Nucera S, Fisher M, Pucci F, Di Serio C, Naldini L, De Palma M, Tozer GM, Lewis CE. TIE2-expressing macrophages limit the therapeutic efficacy of the vascular-disrupting agent combretastatin A4 phosphate in mice. J Clin Invest. 2011;121:1969–1973. [PMC free article] [PubMed]
  • Yu Q, Stamenkovic I. Cell surface-localized matrix metalloproteinase-9 proteolytically activates TGF-beta and promotes tumor invasion and angiogenesis. Genes Dev. 2000;14:163–176. [PubMed]
  • Zhang J, Lu Y, Pienta KJ. Multiple roles of chemokine (C-C motif) ligand 2 in promoting prostate cancer growth. J Natl Cancer Inst. 2010;102:522–528. [PMC free article] [PubMed]