Although overexpression of the epidermal growth factor receptor (EGFR; ErbB1) has been correlated with poor prognosis in breast and other cancers, clinical trials of ErbB1 inhibitors have shown limited efficacy in inhibiting tumor proliferation. To evaluate other possible roles of ErbB1 in tumor malignancy besides proliferation, we have developed a series of tools for analysis of intravasation. Overexpression of ErbB1 in MTLn3 mammary adenocarcinoma cells results in increased intravasation and lung metastasis from tumors formed by injection of cells in the mammary fat pad. However, increased ErbB1 expression has no effect on primary tumor growth and lung seeding efficiency of cells injected i.v. Chemotactic responses to low concentrations of EGF in vitro and cell motility in vivo in the primary tumor measured using intravital imaging are significantly increased by ErbB1 overexpression. The increased cell motility is restricted to ErbB1-overexpressing cells in tumors containing mixtures of cells expressing different ErbB1 levels, arguing for a cell-autonomous effect of increased ErbB1 expression rather than alteration of the tumor microenvironment. In summary, we propose that ErbB1 overexpression makes more significant contributions to intravasation than growth in some tumors and present a novel model for studying ErbB1 contributions to tumor metastasis via chemotaxis and intravasation. (Cancer Res 2006; 66(1): 192-7)
Metastasis involves multiple steps, including growth of a primary tumor, intravasation, and arrest and growth in secondary sites. The steps of growth of the primary tumor, as well as arrest and growth in secondary sites have been extensively studied, with genes affecting these steps being identified and studied (1–5). Studies of intravasation have lagged due to limitations in techniques for analyzing this step. The requirements for study of intravasation include quantitation of the cell number in the circulation, the ability to observe cells in the primary tumor, and identification of cell lines in which intravasation is selectively altered (6).
The epidermal growth factor receptor (EGFR; also referred to as ErbB1) is often overexpressed in breast and other cancers (7–9). It has been correlated with poor prognosis (10, 11), and these observations have stimulated the development of ErbB1 inhibitors (12). Both antibody inhibitors, such as cetuximab, and small molecule inhibitors, such as gefitinib and Tarceva, have been developed to inhibit ErbB1 (13–15). Clinical trials of these ErbB1 inhibitors have shown some but much more limited efficacy in blocking tumor growth than would be expected based on the proportions of tumors with overexpressed ErbB1 (16, 17). These results raise the possibility that ErbB1 might contribute to the malignant potential of tumors through affecting other characteristics besides just proliferation. ErbB1 can increase in vitro tumor cell motility and invasion (18–23). However, the in vivo contribution of the increased motility and invasion, which is critical to understanding metastasis of tumors, is still not clear.
We have developed methods for evaluating cell properties at the primary tumor, including multiphoton imaging and quantitation of tumor cell number in the blood (6, 24–28). Multiphoton imaging of green fluorescent protein (GFP)–labeled primary tumors enables direct visualization of tumor cell behavior in vivo. Quantitation of tumor cell number in the blood is a direct evaluation of the efficiency of intravasation. The combination of these two measurements allows identification of changes in cell movement that can contribute to intravasation efficiency. Using MTLn3 and MTC rat mammary adenocarcinoma cell lines, we showed that there can be significant differences in intravasation between metastatic and nonmetastatic cells. More specifically, the metastatic MTLn3 cells showed greater orientation toward blood vessels within the primary tumor, whereas nonmetastatic MTC cells fragmented when interacting with vessels (6, 24). The findings illustrate the value of a direct visualization of cell properties in vivo for dissection of the metastatic process.
In this article, we provide direct in vivo evidence for contributions of ErbB1 to metastasis independent of effects on proliferation. Overexpression of ErbB1 in MTLn3 rat mammary tumor cells results in significantly increased intravasation and metastasis from the primary tumor, without any changes in primary tumor growth. Chemotactic efficiency in vitro and tumor cell motility in vivo were also significantly increased. Increased ErbB1 expression did not affect lung seeding efficiency of cells injected i.v. Our studies are consistent with in vitro and in vivo studies showing that ErbB1 expression can enhance invasiveness (18–23).
Materials and Methods
Cell culture and establishing EGFR transfectants in MTLn3 rat mammary adenocarcinoma cells. Rat mammary adenocarcinoma cell line MTLn3 was obtained from Dr. Garth Nicolson (Institute for Molecular Medicine, Huntington Beach, CA; ref. 29) and was maintained in αMEM (Life Technologies, Inc., Gaithersburg, MD) supplemented with 5% fetal bovine serum and penicillin-streptomycin (Life Technologies). For transfection assays, MTLn3 cells were seeded in six-well plates and grown until 70% confluent. These cells were subjected to retroviral transduction of constructs containing the retroviral expression vector pLXSN alone or pLXSN driving expression of ErbB1 (ref. 30; provided by Dr. David Stern, Yale University). The vectors were packaged in the Phoenix cell line following standard protocols (31), and geneticin-resistant clones were isolated. Cell growth rate and morphology of transductants was determined to be the same as parental cells, and the expression of ErbB1 was confirmed by flow cytometry and Western blot.
Microchemotaxis chamber assay. A 48-well microchemotaxis chamber (Neuroprobe, Cabin John, MD) was used as previously described (32), except that L15 containing 0.35% bovine serum albumin was used instead of αMEMH. For measurement of migration in response to EGF (Life Technologies Bethesda Research Laboratories, Gaithersburg, MD), filters were coated with 27 μg/mL rat tail collagen (BD Biosciences, Palo Alto, CA). After inserting the filters in the chamber, 20,000 cells detached by trypsin/EDTA were plated into the wells of the upper chamber. The chambers were incubated for 3 hours at 37°C before counting the number of cells crossing the filter.
To evaluate the in vitro cell-autonomous chemotactic efficiency of MTLn3-B1 cells compared with MTLn3-PL cells, mixtures of equal numbers of fluorescent MTLn3-B1-GFP and nonfluorescent MTLn3-PL cells were used in the 48-well chemotaxis assay as described above. After fixation, the filters were stained with.5 mg/mL 4′,6-diamidino-2-phenylindole (DAPI) for 2 minutes, and then images of DAPI and GFP fluorescence were taken using a ×10 objective. In each experiment, images of unscraped wells were also taken to determine the actual relative percentage of cells plated per experiment, and this information was used in calculating the relative chemotactic efficiency. ImageJ (developed by Wayne Rasband at the NIH, Bethesda, MD; http://rsb.info.nih.gov/ij/, 1997-2005) macros were written to identify all cells using the DAPI image and MTLn3-B1 cells by GFP fluorescence. For each condition, the relative proportion of MTLn3-B1 cells was determined and used to calculate the relative chemotactic efficiency of MTLn3-B1 cells compared with MTLn3-PL cells. Iressa (AstraZeneca, Manchester, England) and AG879 (SigmaT2067) were used to inhibit ErbB1 and ErbB2, respectively.
Immunoblotting. Cells were grown to 70% confluency in a 10-cm cell-culturing dish and then incubated with serum-free medium for 4 hours. The medium was changed to fresh serum-free medium with or without 5 nmol/L EGF, and cells were incubated for 30 seconds. Cells were then washed with cold PBS and lysed in 0.5 mL lysis buffer [50 mmol/L Tris-HCl (pH 7.5), 1% Triton X-100, 5 mmol/L EGTA, 1 mmol/L EDTA, 150 mmol/L NaCl, 10% glycerol, 1 mmol/L phenylmethylsulfonyl fluoride, 50 mmol/L sodium fluoride, Complete Protease Inhibitor Cocktail (Roche Diagnostics Corp., Indianapolis, IN)]. The plates were scraped with a rubber policeman and incubated on ice for 20 minutes. The lysate was precleared by centrifugation at 15,000 rpm for 15 minutes. Lysate protein concentration was measured using bicinchoninic acid protein assay reagent (Pierce, Rockford, IL). Protein (20 μg) of each sample was loaded and separated by SDS-PAGE and transferred electrophoretically to nitrocellulose membranes, which were immunoblotted by appropriate antibodies followed by incubation with horseradish peroxidase–conjugated secondary antibodies. The following antibodies were used: anti-β-actin (Sigma, St. Louis, MO), anti-phosphotyrosine (PY20, BD Biosciences, San Diego, CA), anti-phospho-ErbB1 (Tyr1173, Cell Signaling Technology, Beverly, MA), anti-phospho-ErbB2 (Tyr877, Cell Signaling Technology), and anti-ErbB2 monoclonal (NeoMarkers, Fremont, CA). The blots were developed by using SuperSignal West Pico Chemiluminescent Substrate (Pierce).
Spontaneous and experimental metastasis assays. All animal studies described here were done according to protocols approved by the Institutional Animal Care and Use Committee of Albert Einstein College of Medicine. To measure spontaneous metastasis, tumor cells were grown to 70% to 85% confluence before being harvested for cell counting. Cells (5 × 105) were injected into the right thoracic mammary fat pads of 5- to 7-week-old female BALB/c severe combined immunodeficient (SCID) mice (National Cancer Institute, Bethesda, MD). The cells were injected in a volume of 100 μL of PBS with calcium and magnesium through a 25-gauge needle. Tumor growth rate was monitored at weekly intervals after inoculation of tumor cells. The tumors were measured in two dimensions, and tumor volumes were calculated using the formula: (length × width2) / 2. Five weeks after inoculation, the animals were anesthetized, blood burden was determined as described below followed by sacrifice of the animals and removal of primary tumors and lungs for histologic analysis as described below.
For the experimental lung metastasis assay, 2.5 × 105 cells were injected into the lateral tail veins of 5- to 7-week-old female BALB/c SCID mice (National Cancer Institute). Two weeks after injection, the mice were euthanized, and the lungs were removed and subjected to histologic examination as described below.
Measurement of tumor cell blood burden. At the end point of the spontaneous metastasis assay, mice were anesthetized with Aerrane (Isoflurane, Baxter Pharmaceutical Products, Inc., Deerfield, IL). The right chest was exposed by a simple skin flap surgery. Blood was taken from the right atrium via heart puncture with a 25-gauge needle and 1-mL syringe coated with heparin. Blood (0.2-1.05 mL) was harvested from each animal. The blood was immediately plated into 150-mm-diameter dishes filled with 5% fetal bovine serum/αMEM growth medium. The next day, the plates were rinsed with fresh medium containing 0.8 mg/mL geneticin to selectively grow the tumor cells. After 3 to 7 days, all tumor cell clones in the dish were counted. Tumor blood burden was calculated as total colonies in the dish divided by the volume of blood taken. To test tumor cell viability in the blood colony assay process, cultured MTLn3-PL cells were removed from dishes by 0.25% trypsin/EDTA, and 1,000 tumor cells were added to fresh mouse blood. The mixtures were then processed as described above.
Tumor histology and quantitative assessment of the efficiency of metastasis. The primary tumors and lungs from each mouse were used for histologic analysis. Samples were fixed in formalin and embedded in paraffin, and 5-μm sections were stained with H&E. For each lung sample, all micrometastases were counted under a light microscope at ×10 magnification, and the total lung area was measured using a UMAX PowerLook III color scanner (UMAX Technologies, Inc., Dallas, TX) and Adobe Photoshop 5.5 software. Briefly, after scanning lung sections, the cross-sectional area in pixels was measured using PhotoShop and converted to mm2 using a calibration factor of 1 pixel = 0.00179 mm2. The efficiency of lung metastasis then was calculated as metastases per mm2 of lung area.
Intravital imaging by multiphoton microscopy. Tumor imaging was done as described previously (6, 24, 27, 33). Cells (5 × 105) were injected into the mammary fat pads of SCID mice as described above and allowed to grow for 4 weeks. The animal was placed under isoflurane anesthesia, and the tumor was exposed using a simple skin flap surgery, with as little disruption of the surrounding vasculature as possible. The animal was then placed in a 30°C temperature chamber on an inverted Olympus IX70 microscope and imaged using a ×20 objective. Briefly, a 10-W Millennium Xs laser (Spectra Physics, Mountain View, CA) was used to run a Radiance 2000 multiphoton system (Bio-Rad, Hercules, CA) at 880 nm using a 450/480-nm filter to image matrix and CFP and a 515/530-nm filter to image CFP and GFP. Time lapse images of GFP-labeled, MTLn3-PL- and MTLn3-B1-generated tumors were taken at 60-second intervals for 30 minutes. The images were collected using Bio-Rad's Lasersharp 2000 software. During each 1-minute interval, a z series of 9 to 12 images was taken at a spacing of 10 μm between images, beginning at the periphery of the tumor and moving into the tumor. For each tumor, this image acquisition process was repeated for 30 minutes, resulting in a time lapse three-dimensional z series for analysis of tumor cell motility.
Sites of moving tumor cells in each plane of the z series were identified by playing the time lapse movie of that plane in ImageJ and marking sites of movement using a specially written plugin. The total number of movement sites for each plane over the entire 30 minutes was then calculated. The top four planes (corresponding to the outermost 40 μm of the tumor) reproducibly provided good enough images for determination of movement sites and were used to calculate the average number of motility sites for each tumor type.
Increasing ErbB1 expression in MTLn3 cells significantly enhances lung metastasis but not tumor size. MTLn3 mammary adenocarcinoma cells were transfected with the pLXSN retroviral expression vector containing full-length ErbB1 or with the empty vector as a control and were designated as MTLn3-B1 and MTLn3-PL, respectively. Increased expression of ErbB1 protein in MTLn3-B1 was confirmed by Western blot (Fig. 1A,, left). As shown in Fig. 1A, expression of other ErbB family members, such as ErbB2 and ErbB3, was not affected (ErbB4 was not detectable). Fluorescence-activated cell sorting (FACS) analysis confirmed higher cell surface expression levels of ErbB1 for MTLn3-B1 (data not shown). No significant changes in morphology or growth rate in vitro were observed for MTLn3-B1 cells compared with MTLn3-PL cells. Stimulation of MTLn3-B1 and MTLn3-PL cells with EGF showed increased tyrosine phosphorylation of ErbB1 and ErbB2 in MTLn3-B1 cells, indicating that ErbB1/ErbB2 heterodimer signaling is increased in these cells (Fig. 1B).
To determine the effects of increased ErbB1 expression on tumor growth and metastasis, the ErbB1 transductants were injected into the mammary fat pads of SCID mice, and tumor growth was monitored. All animals implanted with tumor cells formed tumors at the site of injection, and the tumors were palpable about 2 weeks after injection. The dynamics of in vivo tumor growth rates of MTLn3-B1 and MTLn3-PL were similar. At 5 weeks, the animals were sacrificed for analysis of spontaneous metastasis efficiency. The mean volumes of MTLn3-pL and MTLn3-B1 tumors were 1,825 ± 215 and 1,666 ± 250 mm3, respectively (Fig. 1C). There are no statistically significant differences in tumor volume (P > 0.05), indicating that increased ErbB1 expression does not affect tumor growth rate of MTLn3 cells.
To evaluate the effect of increased ErbB1 on the efficiency of metastasis, lung samples from the mice were examined physically at autopsy and by light microscopy on H&E-stained paraffin sections. As shown in Fig. 1D, mice bearing MTLn3-B1 tumors had significantly more lung metastases than mice carrying MTLn3-PL tumors. On average, MTLn3-B1 tumors generated 6.00 ± 2.59 metastases/mm2 compared with 0.49 ± 0.26 metastases/mm2 generated by MTLn3-PL tumors. The difference is highly significant (P = 0.0001) and shows that ErbB1 increases spontaneous metastatic efficiency of MTLn3 cells without affecting primary tumor growth.
Increasing tumor cell motility and intravasation by ErbB1 expression. Intravasation is an important metastatic property and can be evaluated by determining the number of tumor cells present in blood collected from the right atrium of the heart, before filtration by the lungs. The number of viable tumor cells present is then determined by culturing the blood and counting the number of tumor cell colonies that form. Using blood added to cultured tumor cells, this assay has an ∼87% recovery rate. FACS analysis of blood taken from animals bearing tumors formed by GFP-labeled cells shows a good correlation between the number of colonies present on the plate and the number of cells in the blood (correlation coefficient, 0.86). On average, 0.7 mL of blood was taken from the right atrium via heart puncture from each animal bearing tumors generated by MTLn3-PL and MTLn3-B1. Animals bearing MTLn3-PL tumors had 37 ± 17 tumor cells/mL of blood, whereas animals bearing MTLn3-B1 tumors had 441 ± 172 tumor cells/mL blood, a highly significant difference (Fig. 2A; P = 0.0055). Thus, increased ErbB1 expression can significantly increase intravasation rate without affecting primary tumor growth.
A mechanism by which increased ErbB1 expression could enhance intravasation is via motility. Because we have shown that MTLn3 cells exhibit in vivo invasion in response to gradients of EGF (6), gradients of EGF from blood vessels or stromal cells in tumors could enhance intravasation. To test whether increasing ErbB1 expression enhanced the chemotactic response of MTLn3 cells to EGF, we used a microchemotaxis chamber migration assay to measure chemotactic sensitivity. MTLn3-B1 cells showed stronger responses at all concentrations of EGF (Fig. 2B), with 5-fold more cells migrating in response to 0.05 and 0.5 nmol/L EGF. Thus, MTLn3-B1 cells show an increased chemotactic response to EGF in vitro that is consistent with the enhanced intravasation rate seen in vivo.
Intravital imaging reveals that ErbB1 enhances tumor cell motility in vivo. To determine whether increased ErbB1 expression enhances tumor cell motility in vivo, we imaged tumor cells directly in primary tumors in mice. MTLn3-PL and MTLn3-B1 cells were stably transduced with a GFP expression vector and then injected into the mammary fat pads of SCID mice to form a primary tumor as in the spontaneous metastasis assay. Four weeks after injection, animals were anesthetized, a partial skin flap dissection done to expose the primary tumor, and the tumors were imaged using a multiphoton microscope. Tumor cell motility was identified in time lapse movies (Fig. 3A and accompanying movie) and the number of moving tumor cells per time lapse movie determined. There was significantly enhanced motility of tumor cells in MTLn3-B1 tumors compared with MTLn3-PL tumors (Fig. 3B).
To test whether the motility changes observed in MTLn3-B1 tumors reflect a difference in the tumor microenvironment of MTLn3-B1 tumors compared with MTLn3-PL tumors, we coinjected CFP-expressing MTLn3-B1 tumor cells with GFP-expressing MTLn3 tumor cells. On average, the same number of GFP- and CFP-labeled tumor cells were present in each imaging plane. ErbB1-overexpressing cells showed significantly greater movement and invasiveness (Fig. 3C and D), compared with neighboring nonoverexpressors. These results are consistent with increased ErbB1 expression stimulating increased tumor cell motility and invasiveness in vivo through enhanced ErbB1 signaling and argue against general modifications of the tumor microenvironment being the major cause of the increased motility.
This observation was confirmed in vitro using the microchemotaxis chamber. Equal numbers of GFP-labeled ErbB1 overexpressors were mixed with unlabeled normal expressors and exposed to 0.05 and 0.5 nmol/L EGF gradients. The relative number of GFP-labeled to non–GFP-labeled migrating cells was found to be 3 to 4 (Fig. 4A), consistent with the increased migratory capability of ErbB1 cells in response to EGF being due to an increased sensitivity afforded by the higher receptor number. The chemotactic response to EGF of both cell lines was inhibited with dose dependencies consistent with the published IC50's of 0.033 μmol/L for inhibition of ErbB1 by Iressa (ref. 34; Fig. 4B) and 0.35 μmol/L for AG825 (ref. 35; Fig. 4C), indicating that signaling from both ErbB1 and ErbB2 is required and consistent with the results of Fig. 1B showing that both ErbB1 and ErbB2 are phosphorylated in response to EGF.
ErbB1-driven metastasis is not caused by enhanced growth of metastases in the lungs. To determine whether increased ErbB1 expression enhances lung metastasis by increasing the arrest, survival, or growth of tumor cells in the lungs, we compared MTLn3-PL and MTLn3-B1 properties in the experimental metastasis assay. Cells were injected into the lateral tail veins of 5- to 7-week-old female BALB/c SCID mice, and 2 weeks later, the mice were sacrificed, and the lungs were removed and examined for metastases. All animals receiving MTLn3-PL or MTLn3-B1 cells formed metastatic lesions in the lungs. Animals injected with MTLn3-PL cells had 3.2 ± 0.6 metastases/mm2, whereas animals injected with MTLn3-B1 cells had 3.3 ± 0.5 metastases/mm2. There is no statistically significant difference between the two groups (P > 0.05). These results indicate that that ErbB1 overexpression in MTLn3 cells does not affect later steps of metastasis and are consistent with the major effect being an enhancement of intravasation.
We found that increased expression of ErbB1 in MTLn3 mammary adenocarcinoma cells enhanced metastasis through significantly increasing intravasation. There was no effect on primary tumor growth or the efficiency of lung colonization by injection of tumor cells via the tail vein. In parallel with intravasation, tumor cell motility in the primary tumor in vivo was increased. This increase in motility was not due to alterations in the tumor microenvironment, because the motility of MTLn3-B1 cells was also significantly enhanced over cells expressing normal levels of ErbB1 in tumors composed of mixtures of the two cell types. In vitro, increased expression of ErbB1 significantly enhanced chemotaxis to EGF in a cell-autonomous fashion. The enhanced chemotactic response was dependent on both ErbB1 and ErbB2 signaling, as shown by inhibition with Iressa or AG825.
To our knowledge, this is the first direct demonstration that increased expression of ErbB1 can enhance tumor cell motility in the primary tumor and intravasation without affecting tumor growth. Our studies are consistent with other in vitro and in vivo studies showing that ErbB1 expression can enhance invasiveness (18, 19). In addition to its growth-promoting effects, ErbB1 mediates chemotaxis to EGF (36–39) and transforming growth factor-α (40, 41) in vitro, and increased levels of ErbB1 can enhance chemotactic responses to these growth factors (39, 42–44). We have previously shown that tumor cells chemotax towards EGF in vivo, using an in vivo invasion assay in which needles containing EGF are inserted into tumors (45), and MTLn3-B1 cells show enhanced responses in this in vivo invasion assay (46). Thus, increased ErbB1 can enhance invasiveness both in vitro and in vivo in response to ErbB1 ligands. This likely reflects increased sensitivity provided by higher levels of ErbB1 as a heterodimer with ErbB2, which also shows increased tyrosine phosphorylation in response to EGF in MTLn3-B1 cells.
Potential sources of EGF in the primary tumor that could direct tumor cells toward blood vessels include both blood vessels themselves and macrophages (47–49). EGF is reported as present in serum at around 0.1 nmol/L (50), concentrations at which we see a significant difference in chemotactic response between MTLn3-B1 and MTLn3-PL cells. Thus, gradients of EGF formed from leaky blood vessels could stimulate tumor cell motility directly. In addition, macrophages also produce EGF and are present around blood vessels (1). We have shown that macrophages can form a paracrine loop with tumor cells to enhance invasiveness in response to gradients of EGF (46). Thus, there are multiple sources of EGF in vivo that can direct tumor cells towards blood vessels.
In summary, metastasis is an inefficient process, and it is important to dissect each step to identify the mechanisms that contribute to metastasis (1, 2). A major obstacle to examining the intravasation step has been the lack of reliable tools to track tumor cell movement in vivo. Critical elements required for examination of the process of intravasation in detail include (a) a method for evaluation of intravasation by measuring the number of viable tumor cells present in the blood; (b) development of multiphoton imaging methods, which allow direct visualization of tumor behavior in vivo; and (c) identification of tumor models, which show a strong correlation between intravasation rate and metastatic rate. In this article, we provide an example of an appropriate tumor model for studying intravasation efficiency and apply measurements of intravasation and intravital imaging to show its use. These tools provide access to studies of intravasation as a new target for anticancer therapy.
Note: E. Sahai is currently at the Tumour Cell Biology Laboratory, Cancer Research United Kingdom, London Research Institute, 44 Lincoln's Inn Fields, London WC2A 3PX, United Kingdom.
Grant support: National Cancer Institute grants R01 CA77522 (J.E. Segall), P01 CA100324 (J.E. Segall and J. Condeelis), and R01 CA69202 (Z-Y. Zhang).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank the members of the Cox, Segall, and Condeelis labs for comments and suggestions and David Stern, AstraZeneca, Bristol Myers Squibb (Princeton, NJ), and Gary Nolan (Stanford University, Stanford, CA) for the reagents.