IJC International Journal of Cancer

Chemotherapy of WAP-T mouse mammary carcinomas aggravates tumor phenotype and enhances tumor cell dissemination Katharina Jannasch1, Florian Wegwitz2,3, Eva Lenfert2, Claudia Maenz2, Wolfgang Deppert2 and Frauke Alves1,4,5 1

Department of Hematology and Medical Oncology, University Medical Center, 37075 Goettingen, Germany Department of Tumor Biology, University Medical Center Hamburg-Eppendorf, 20246 Hamburg, Germany 3 Department of General, Visceral and Pediatric Surgery, University Medical Center, 37075 Goettingen, Germany 4 Max-Planck Institute for Experimental Medicine, Molecular Biology of Neuronal Signals, 37075 Goettingen, Germany 5 Department of Diagnostic and Interventional Radiology, University Medical Center, 37075 Goettingen, Germany

In this study, the effects of the standard chemotherapy, cyclophosphamide/adriamycin/5-fluorouracil (CAF) on tumor growth, dissemination and recurrence after orthotopic implantation of murine G-2 cells were analyzed in the syngeneic immunocompetent whey acidic protein-T mouse model (Wegwitz et al., PLoS One 2010; 5:e12103; Schulze-Garg et al., Oncogene 2000; 19:1028–37). Single-dose CAF treatment reduced tumor size significantly, but was not able to eradicate all tumor cells, as recurrent tumor growth was observed 4 weeks after CAF treatment. Nine days after CAF treatment, residual tumors showed features of regressive alterations and were composed of mesenchymal-like tumor cells, infiltrating immune cells and some tumor-associated fibroblasts with an intense deposition of collagen. Recurrent tumors were characterized by coagulative necrosis and less tumor cell differentiation compared with untreated tumors, suggesting a more aggressive tumor phenotype. In support, tumor cell dissemination was strongly enhanced in mice that had developed recurrent tumors in comparison with untreated controls, although only few disseminated tumor cells could be detected in various organs 9 days after CAF application. In vitro experiments revealed that CAF treatment of G-2 cells eliminates the vast majority of epithelial tumor cells, whereas tumor cells with a mesenchymal phenotype survive. These results together with the in vivo findings suggest that tumor cells that underwent epithelial-mesenchymal transition and/or exhibit stem-cell-like properties are difficult to eliminate using one round of CAF chemotherapy. The model system described here provides a valuable tool for the characterization of the effects of chemotherapeutic regimens on recurrent tumor growth and on tumor cell dissemination, thereby enabling the development and preclinical evaluation of novel therapeutic strategies to target mammary carcinomas.

Mammary carcinoma is the most frequent type of cancer in women in the Western world.1 About 25% of the patients develop distant metastases and ultimately die of breast cancer.2 The prognosis of cancer patients is largely determined by the occurrence of distant metastases. Disseminated tumor cells (DTCs) in the bone marrow at diagnosis and during Key words: transgenic tumor mouse models, breast cancer, disseminated tumor cells, recurrent tumor growth, chemotherapy Additional Supporting Information may be found in the online version of this article. Grant sponsor: Deutsche Krebshilfe; Grant numbers: 109323 and 109315; Grant sponsor: Jung-Stiftung f€ ur Forschung, Hamburg, Germany DOI: 10.1002/ijc.29369 History: Received 1 Aug 2014; Accepted 13 Nov 2014; Online 2 Dec 2014 Correspondence to: Katharina Jannasch, Department of Hematology and Medical Oncology, University Medical Center, Robert-Koch-Str. 40, 37075 Goettingen, Germany, Tel.: 149-0-551/ 39–6991, Fax: 149-0-551/39–7853, E-mail: [email protected]

C 2014 UICC Int. J. Cancer: 137, 25–36 (2015) V

follow-up have been demonstrated in several early breast cancer studies to be of clinical relevance3,4 and are used in patient risk assessment. The high incidence of breast cancer and the high mortality of the disease ask for better treatment options. However, a major problem in evaluating the effects of cancer therapy in preclinical settings results from the fact that the main parameter analyzed is tumor growth, and treatment effects on DTCs are not monitored. Especially, for treating mammary carcinomas, strategies to target DTCs and metastases are absolutely essential, as they are known to disseminate at an early stage.5 Most preclinical studies are performed using xenograft models implanting human tumor cells into immunocompromised mice, which only inadequately mimic tumor development and progression in a natural setting.6 Although different tumor models based on immunodeficient NOD scid gamma mice have been developed and extended,7 to date preclinical models that would allow monitoring the effects of chemotherapy on tumor cell dissemination and metastasis in mammary carcinoma in an immunocompetent setting are extremely scarce.

Carcinogenesis

2

26

Phenotype changes in carcinoma after therapy

Carcinogenesis

What’s new? Despite their prognostic value in breast cancer, disseminated tumor cells (DTCs) aren’t usually monitored when treatments are evaluated. In this study, the authors developed a mouse model for analyzing the effects of chemotherapy on the growth, recurrence, and dissemination of mammary tumors in vivo. They found that tumors exposed to chemotherapy often develop a more aggressive phenotype, that tumor cells with a mesenchymal phenotype often survive exposure, and that recurrence is associated with increased DTCs. Their results also indicate that neoadjuvant chemotherapy should be carefully considered.

The authors of this study have previously developed and characterized a mouse model in which mammary carcinogenesis was initiated by whey acidic protein (WAP) promoterdriven expression of SV40 early proteins (WAP-T mice8–10). Initiated WAP-T mice develop mammary carcinomas, which could be cross-species validated with corresponding human tumors.11,12 From an undifferentiated WAP-T tumor (T1H229,13), the authors established a mammary cancer cell line (G-2).14 This clonal cell line exhibits stem-like properties and builds up a homeostatic, self-renewing “cancer cell system,” made up by subpopulations that, in a simplified manner, can be classified into “epithelial” and “mesenchymal” differentiation states. This cell line mimics the phenotypic heterogeneity observed in primary WAP-T tumors already during in vitro culture. Consequently, orthotopic transplantation of as low as 10 G-2 cells into immunocompetent syngeneic mice closely mimics endogenous mammary cancer development. Moreover, tumor cells from mammary carcinomas arising after orthotopic transplantation of G-2 cells disseminate into various organs (Maenz C, Lenfert E, Deppert W, Wegwitz F, manuscript in preparation). Transplanted G-2 cells thus provide a system that allows monitoring the effects of chemotherapy on tumor growth and on DTCs. In clinical treatment of mammary carcinomas, surgery, radiation and chemotherapy are used in different combinations or sequences to select an appropriate strategy for the type and stage of the disease. Because breast cancer often spreads early, today, surgical treatment as a combination of removal of the primary tumor and staging of the axillary lymph nodes is standard. Because of the risk of tumor recurrence in the remaining breast tissue, therapies in the form of radiation therapy and chemotherapy either in an adjuvant (after surgery) or a neoadjuvant (before surgery) way is used to eliminate microscopic disease and to prolong patient survival by reducing the chance of local recurrence. Anthracyclines as doxorubicin (adriamycin) have been proven to be very effective cytotoxic agents in the treatment of breast cancer.15 In combination with cyclophosphamide (cytoxan) and 5-fluorouracil (5FU), doxorubicin is regularly used in patients with early breast cancer. This standard regime, termed CAF (cyclophosphamide/adriamycin/5FU), is given in certain time intervals for 4 months and is well tolerated. This study, for the first time, analyzes the influence of CAF chemotherapy not only on tumor growth but also on tumor phenotype, recurrence and dissemination of tumor cells in an orthotopic immunocompetent syngeneic mam-

mary tumor mouse model. This model system thus enables a detailed characterization on the progression and regrowth of mammary carcinoma in response to chemotherapy, thereby representing a useful tool to investigate novel therapeutic approaches in mammary carcinomas.

Material and Methods In vivo experiments

All animals were handled according to German regulations for animal experimentations, and all animal protocols were approved by the administrations of Hamburg and Lower Saxony, Germany. G-2 cells14 were maintained in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% FCS and 2 mmol/l glutamine at 37 C in a humidified atmosphere with 5% CO2. For implantation, 1 3 106 cells were resuspended in 20 ml of a 1:1 mixture of serum-free DMEM and BD Matrigel Matrix High Concentration, Growth Factor Reduced (BD Bioscience, San Jose, CA). The 10 to 20 weeks old female virgin WAP-T-NP8 mice (Balb/c background) were anesthetized by intraperitoneal injection of ketamine/xylazine. After a 3–4 mm incision of the skin, the cell suspensions were injected into the right abdominal mammary gland (MG#6); carprofen (50 mg/ml) was applied as analgesic; and the skin was closed by interrupted sutures. The operation was performed under sterile conditions. Size of growing tumors was measured twice a week with a caliper, and volumes were calculated using the following formula: volume 5 0.5 3 (length 3 width 3 height). CAF chemotherapy [cyclophosphamide (Endoxan, Baxter, Deerfield, IL) 100 mg/kg body weight (BW), doxorubicin (Cell Pharm, Hannover, Germany) 5 mg/kg BW and 5FU (Medac, Wedel, Germany) 100 mg/kg BW in phosphate buffered-saline] was administered intraperitoneally at time points indicated. Dissection and preparation

Mice were killed using isoflurane, and blood was taken by cardiac puncture. Subsequently, mice were dissected, and the primary tumor and following organs were prepared: Mammary glands #2 (thoracal left), #3 (abdominal left) and #7 (thoracal right), liver, lung, spleen, brain and axial lymph nodes on the left and right side. Bone marrow was taken from the left and right femur. One half of the primary tumor, the lung, the liver and the brain were fixed using formalin, and all other samples were stored at 280 C. Blood and bone marrow were processed using erythrocyte lysis buffer of the C 2014 UICC Int. J. Cancer: 137, 25–36 (2015) V

27

Jannasch et al.

Name

Sequence

Zeb1-Ex3_for

CACCAGAAGCCAGCAGTCAT

Zeb1-Ex4_rev

CGTTCTTCTCATGGCGGTACT

Twist1-Q1

CGGACAAGCTGAGCAAGATTC

Twist1-Q2

TCCAGACGGAGAAGGCGTAG

Mmp3-Q1

TGTCCCGTTTCCATCTCTCTC

Mmp3-Q2

TGGTGATGTCTCAGGTTCCAG

Epcam-Q1

GAGTCCGAAGAACCGACAAGG

Epcam-Q2

CTGATGGTCGTAGGGGCTTTC

Vim-Q1

CGGCTGCGAGAGAAATTGC

Vim-Q2

CCACTTTCCGTTCAAGGTCAAG

Wisp1-Q1

ACAGAGGCTGCCATCTGTGAC

Wisp1-Q2

CTCGCCATTGGTGTAGCGTA

Tgfb1i1-Ex6–7_Q1

CCTCTGTGGCTCCTGCAATA

Tgfb1i1-Ex6–7_Q2

AGCGCTCAAAGTAGCACTCG

Thy1-Ex2-Qa1

GAACTCTTGGCACCATGAACC

Thy1-Ex3-Qa2

GTTATTCTCATGGCGGCAGTC

Snai1-Q1

CTGGTGAGAAGCCATTCTCCT

Snai1-Q2

CCTGGCACTGGTATCTCTTCA

SV40LTag-Q1

TCCTGGCTGTCTTCATCATC

Histology

Tissue specimens were fixed in formalin, embedded in paraffin, cut into sections of 2 mm and deparaffinized. Sections from all samples were stained with hematoxylin and eosin (H&E). In order to detect tumor cells in the lung, the lung was cut into serial sections (2 mm) and stained using antiSV40 T-antigen (T-Ag) antibody (homemade rabbit antibody, R15, 1:20,00014). Sections of all primary tumors (n 5 9–12 per group) were stained using antibodies against vimentin (rabbit monoclonal ab92547, Abcam, Cambridge, UK; 1:500), E-cadherin (rabbit monoclonal #3195, Cell Signaling Technology, Beverly, MA; 1:200), SV40 T-Ag (homemade), F4/80 (rat MCA497EL, AbD Serotec, Oxfordshire, UK; 1:100), CD45 (rat #103102, Biolegend, San Diego, CA; 1:600), Collagen-I (rabbit R1038, Acris; 1:200), alpha smooth muscle actin (rabbit ab5694, Abcam; 1:500) and ECF-L (M2 macrophages; goat #AF2446, R&D systems, Minneapolis, MN; 1:100) as well as the histochemical stainings with H&E, Masson–Goldner trichrome and Picrosirius red. Either Histofine Simple Stain MAX PO anti rabbit (Nichirei, Tokyo, Japan; undiluted) or Biotin anti-rat (BioLegend; 1:200) together with Avidin HRP (eBioscience, San Diego, CA; 1:400) were used as secondary antibodies. Slices for all antibody stainings were pretreated according to the manufacturers’ recommendations. Using H&E-stained tumor sections, proliferation of tumor cells was determined by counting mitoses in five to seven randomly taken pictures (magnification 340) from each tumor.

SV40LTag-Q2

TACAGACCTGTGGCTGAGTT

Klf4-Q1

CGACTAACCGTTGGCGTGAG

Klf4-Q2

GTCGTTGAACTCCTCGGTCTC

Nanog_Q1

CCTGATTCTTCTACCAGTCCCA

Chemotherapy and proliferation in vitro

Nanog_Q2

GGCCTGAGAGAACACAGTCC

Sox2_Q1

GACAGCTACGCGCACATGA

Sox2_Q2

GTGCATCGGTTGCATCTG

qRT-Hspa8-3

CCGATGAAGCTGTTGCCTAT

qRT-Hspa8-4

GTGACATCCAAGAGCAGCAA

mutp53_HA-1a

GACCGCCGTACAGAAGAAGAA

mutp53_HA-Q2

TCAGATCTTCAGGCGTAGTCG

Notch4-Ex27–28_Q1

CTGCACCTAGCTGCCAGATTC

Notch4-Ex27–28_Q2

CTGTCTGCTGGCCAATAGGAG

CAF was applied on G-2 cells in the following concentrations: cyclophosphamide 1 mg, doxorubicin 0.05 mg and 5FU 1 mg per 100 ml DMEM (modified after Ref. [16]). Cell proliferation was determined using the CellTiter 96V AQueous One Solution Cell Proliferation Assay (Promega Corp., Mannheim, Germany) according to the manual. 50,000 G-2 cells were seeded in 96-well microtiter plates, and 24 hr later extinction values of cells and control wells with DMEM were determined. Chemotherapy was administered to half of the wells, and viable cells were measured 1, 2, 3, 8 and 11 days afterward. The experiment was repeated twice. To investigate cell recovery after chemotherapy, G-2 cells were seeded and treated with CAF as described earlier. After 24 or 48 hr, respectively, CAF was washed away, and cells were cultured in DMEM 1 10% FCS for up to 22 days until viable cells were determined. R

QIAamp RNA Blood Mini Kit (Qiagen, Hilden, Germany) for 30 min on ice followed by washing twice. Polymerase chain reaction

Genomic DNA was isolated from 50 mg of each sample and additionally from serum of tumor-bearing WAP-T mice by using phenol–chloroform extraction. G-2 cells were detected via the G-2-specific HA-sequence (primer mutp53_HA-1a and mutp53_HA-Q2, Table 1) by polymerase chain reaction (PCR) in at least two independent assays with 200 ng of DNA each. DNA quality was verified by performing a PCR of the Notch4-Gene (Table 1). DNA of the corresponding primary tumor was used as positive control, and DNA from Balb/c mice was used as negative control. C 2014 UICC Int. J. Cancer: 137, 25–36 (2015) V

Quantitative real-time PCR

RNA was purified using the innuPrep RNA Mini Kit (Analytik Jena, Jena, Germany) and reverse transcribed with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA). Routinely, 1 mg RNA was used for cDNA synthesis. Quantitative real-time PCR was performed using the Power SYBR Green PCR Mastermix (Applied Biosystems) in an ABI 7500 Fast thermal cycler

Carcinogenesis

Table 1. Primers used for PCR and RT-qPCR

28

(Applied Biosystems). Five nanograms cDNA per 10 ml mix and primer pairs in concentrations of 100 nM were used (Table 1).

Carcinogenesis

Western blot analysis

Cells were lysed with Laemmli lysis buffer. Protein concentration was estimated using the BCA Protein Assay (Thermo Scientific, Waltham, MA). Equal amounts of protein were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis and transferred onto nitrocellulose (Hybond, Amersham, GE Healthcare Life Sciences, Pittsburgh, PA). The membrane was blocked with 5% milk in Tris-buffered saline and Tween 20 for at least 60 min. Following primary antibodies were used: E-cadherin (rabbit monoclonal #3195, Cell Signaling; 1:2,000), vimentin (goat sc-7557, Santa Cruz Biotechnology, Santa Cruz; 1:500) and a-tubulin (mouse CP06, Millipore, Billerica, MA; 1:5,000). All secondary antibodies conjugated with peroxidase (Dianova, Hamburg, Germany) were diluted 1:5,000 to 1:10,000 in 5% milk (Trisbuffered saline and Tween 20).

Results Effects of chemotherapy on tumor growth

To gain insight into the effect of chemotherapy on mammary carcinoma growth, the standard chemotherapy regimen CAF (Cytoxan (cyclophosphamide), Adriamycin (doxorubicin) and 5FU) was applied to WAP-T mice after orthotopic implantation of G-2 cells. Mice were divided into four experimental groups, each containing 9–12 animals as shown in Figure 1a. As a control, untreated tumor-bearing mice in Group 1 were dissected when the tumor volume reached 0.5 cm3. Mice in Group 2 were treated with CAF at a tumor volume of 0.5 cm3 and dissected 9 days later. In this group, the primary tumor volume was reduced by 30–50% already 24 hr after treatment. Within 7–9 days, the tumor volumes decreased to 0.02–0.05 cm3. All mice well tolerated one round of CAF treatment, none of the tumors disappeared completely and, until dissection 9 days after CAF treatment, none of the tumors had started to grow again (Fig. 1b). Unfortunately, a second round of CAF treatment, as applied in human patients, was not possible, as mice died within 1 week after repeated chemotherapeutic treatment. As CAF treatment was not able to eradicate tumors completely, a further treatment group was set up to investigate potential recurrent tumor growth (Group 3). Mice were treated with CAF like mice in group 2 (Fig. 1a). After 20 6 3 days of CAF treatment, the tumors started to grow again, and mice were dissected when the tumors reached their initial volume of 0.5 cm3 31 6 3 days after CAF treatment. Growth rates of recurrent tumors were similar to those of the initial tumors (Fig. 1b). Although not related to any clinical situation, the effect of chemotherapy on tumor onset was examined by treating mice with CAF 3 days after tumor cell implantation (Group 4). Like in Groups 1 and 3, mice of Group 4 were dissected

Phenotype changes in carcinoma after therapy

at a tumor volume of 0.5 cm3. All mice in Group 4 developed tumors, which started to become palpable 18–30 days later than the tumors in the control group. Thus, tumor onset was delayed in comparison with tumors in untreated control mice. However, growth rates were not affected by early CAF treatment (Fig. 1b). Comparable numbers of mitotic figures were determined in all tumor sections from Groups 1, 3 and 4, demonstrating no differences in proliferation in the groups with recurrent or delayed growth in comparison with controls (data not shown).

Effects of chemotherapy on tumor cell dissemination

To detect DTCs in response to CAF treatment in different mouse organs, a PCR recognizing the HA-tag of the mutp53 transgene present in G-2 cells13 but not in host NP8 mice was designed. Using spiking experiments, it was possible to detect tumor cells with a sensitivity of at least 100 cells in 50 mg liver tissue (Fig. 2a), corresponding to about 25 tumor cells in 1 3 106 tissue cells. To exclude that this PCR detected free circulating DNA, PCRs were performed in blood serum of G-2-tumor-bearing animals. In these samples, HA-signals were never detected (data not shown). Furthermore, several lung tissue samples that had scored positive for tumor cells by PCR analysis were stained for T-Ag by immunohistochemistry. In these lung samples, single T-Ag positive tumor cells were able to be detected (Fig. 2b). To assess the dissemination of tumor cells and their organ distribution at the end of the experiments, samples from nine different organs (lung, kidney, spleen, liver, brain, left and right thoracal and left abdominal mammary glands and left and right axial lymph nodes), as well as bone marrow from the right and left femur and blood from each animal were taken for further PCR analysis. On average, in mice of the control group 3.1 out of 12 organ samples per mouse were positive for DTCs at a tumor volume of 0.5 cm3 (Fig. 2c). In each animal, DTCs could be detected in at least one organ. In mice dissected directly after chemotherapy (Group 2), no DTCs were found in five of nine animals. In total, only 1.6 organ samples per animal contained DTCs. In contrast, mice with recurrent tumor growth, on average, contained 5.75 organ samples per mouse positive for DTCs at the end of the experiment; only in one animal no DTCs could be detected. In mice with delayed tumor growth after CAF treatment (Group 4) even 6.1 positive organ samples were found per mouse, and all animals had at least one organ positive for DTCs (Figs. 2c and 2d). No CTCs could be found in the blood of mice from the control and chemotherapy groups, whereas blood from mice with recurrent or delayed tumor growth was positive for CTCs at a low incidence (in 8% and 10% of all mice, respectively). In contrast, bone marrow and lymph node samples were positive for DTCs in mice from all groups, with a higher incidence in lymph nodes in Groups 3 and 4 (Fig. 2e). Moreover, in the control and the chemotherapy group some organs were always negative for DTCs, C 2014 UICC Int. J. Cancer: 137, 25–36 (2015) V

29

Carcinogenesis

Jannasch et al.

Figure 1. (a) Schematic representation of experimental group design. Group 1: control group; no chemotherapy, dissection at tumor volume of 0.5 cm3; Group 2: treatment with CAF at tumor volume of 0.5 cm3, dissection 9 days later; Group 3: treatment with CAF at tumor volume of 0.5 cm3, dissection after recurrent tumor growth at tumor volume of 0.5 cm3; Group 4: treatment with CAF 3 days after implantation of tumor cells, dissection at tumor volume of 0.5 cm3. (b) Growth rates of primary tumors in different groups. In graphs of Groups 1 and 2, the mean tumor volume of all mice from the groups is shown in mm3 over time. Graphs for Groups 3 and 4 show the tumor volume of every single animal in mm3 over time. Application of chemotherapy is indicated. Significant reduction of tumor volume in Group 2 was observed between days 22 and 26, p < 0.0001.

whereas every organ tested was positive for DTCs at least in one animal in the groups with recurrent or delayed growth. Chemotherapy is associated with an altered tumor morphology and phenotype

The influence of chemotherapy on tumor phenotype was then investigated by immunohistochemistry of primary tumor sections (n 5 9–12 per group) from each experimental group. Staining with H&E revealed clear differences in tumor morphology in the individual groups (Fig. 3a). Untreated primary tumors from control Group 1 are compact, composed of proliferation-active undifferentiated cells, with sparse residual glandular tissue and often a large central necrosis (Fig. 3a, yelC 2014 UICC Int. J. Cancer: 137, 25–36 (2015) V

low arrow). Nine days after CAF treatment, residual tumor areas of Group 2 were compact, including tumor cells with heterochromatic nuclei. Multiple regressive alterations such as apoptosis, fat deposition, multinucleated giant tumor cells, interstitial edema or dystrophic calcifications could be observed (Fig. 3a, red arrow on fat deposition). Necrotic areas were small, focal and distributed all over the tumor (Fig. 3a, yellow arrow). Moreover, spindle-shaped nuclei could be found in tumor cells, indicating epithelial-mesenchymal transition (EMT) (Supporting Information Fig. 1d, arrows). Tumors with recurrent or delayed growth (Groups 3 and 4) are less dense than tumors from Group 1 or 2, reflected by a higher degree of intercellular spaces and a nodular

Phenotype changes in carcinoma after therapy

Carcinogenesis

30

Figure 2. Detection of tumor cell dissemination in G-2-cell transplanted mice. (a) Sensitivity of HA-PCR in spiking experiments. 101 to 104 G-2 cells in 50 mg liver tissue, G-2 cells from two different passages, liver tissue from native WAP-T-NP8 and Balb/c mice as well as H2O as negative control were tested via HA-PCR. For each PCR reaction, 200 ng of DNA were used. It was possible to detect as few as 100 G-2 cells in 50 mg tissue. (b) Staining of disseminated tumor cells. Above: Staining of DTCs in blood vessels of a tumorbearing mouse using H&E (left) and T-Ag antibody (right). Below: Staining of DTCs in a lung section using T-Ag antibody. Red box shows area of magnification. Arrows indicate single DTCs. (c) Treatment with chemotherapy influences dissemination of tumor cells. Mean value of DTC-positive organs. Application of chemotherapy leads to a reduction of organs positive for DTCs, whereas animals with recurrent as well as delayed tumor growth show a higher amount of organs positive for DTCs in comparison with the control group. (d) Number of animals without DTCs. Nine days after application of chemotherapy, in half of the animals, not a single DTC can be detected any more. In all other groups, in most of the animals, at least one organ was positive for DTCs. (e) Distribution of DTCs in organs (in percentage of all animals from one group).

growth pattern (Fig. 3a, blue arrows). Moreover, tumor cells in these groups were entirely undifferentiated. Tumors from both groups exhibited extensive coagulative necroses, distributed all over the tumor (Fig. 3a, yellow arrows).

To test whether CAF treatment might alter the phenotypic differentiation state of tumor cells, all tumors were stained using antibodies against the mesenchymal marker vimentin and the epithelial marker E-cadherin. As shown in Figure 3b, C 2014 UICC Int. J. Cancer: 137, 25–36 (2015) V

31

Carcinogenesis

Jannasch et al.

Figure 3. (a) Morphological changes in primary tumors. H&E-stained sections of primary tumors from all groups show differences in compactness and distribution of necrosis (yellow arrows). Higher magnification reveals that application of chemotherapy (Group 2) leads to primary tumors that are small and compact but still consist of vital tumor cells. Moreover, signs of regression as, e.g., fat deposition (red arrow), can be found. Independently of whether the tumor growth is recurrent (Group 3) or delayed (Group 5), tumors show a much looser structure than tumors from Group 1 (control) or 2 (chemotherapy). This is characterized by a higher degree of intracellular spaces and a nodular growth pattern (blue arrows). (b) Phenotypic alterations in primary tumors. Representative pictures of sections stained for vimentin and E-cadherin show complementary staining patterns. Application of chemotherapy (Group 2) leads to a massive reduction of cells positive for the epithelial marker E-cadherin in favor of cells that show the mesenchymal marker vimentin. The tumors per mouse group was n 5 9–12.

the staining patterns for vimentin and E-cadherin are complementary. In all groups, cells were positive for either vimentin or E-cadherin, except for necrotic areas. However, the control group showed a balanced ratio of cells expressing vimentin and cells expressing E-cadherin, whereas in tumors of Group 2, only sparse cells were positive for E-cadherin and almost all cells were positive for vimentin 9 days after CAF treatment. Interestingly, tumors after recurrent or delayed tumor growth in Groups 3 and 4 again displayed a balanced distribution of vimentin- and E-cadherin-positive C 2014 UICC Int. J. Cancer: 137, 25–36 (2015) V

cells, comparable with tumors of the control group. As the mutp53 transgene is not expressed in G-2 cells,17 only T-Ag expression could be used as marker for tumor cells. Although the majority of cells in tumors of Group 1 (control) and in tumors with recurrent or delayed growth (Groups 3 and 4) were T-Ag positive, thus representing tumor cells, tumors analyzed directly after chemotherapy treatment (Group 2) contained only a few T-Ag positive cells (Fig. 4a; Supporting Information Figs. 1c and 1d). Residual tumor areas consisted mainly of vimentin-positive cells of other origin, e.g., immune

Carcinogenesis

32

cells or fibroblasts, and of few vimentin- and T-Agpositive mesenchymal-like tumor cells (Supporting Information Fig. 1). To unravel the origin of the T-Ag-negative and vimentinpositive cells present in tumors of Group 2, immunohistochemistry was performed with antibodies directed against immune cells, including macrophages. Staining against macrophagespecific F4/80 (Fig. 4b and Supporting Information Fig. 1e) demonstrated that all G-2 tumors contain a cellular subpopulation of macrophages. Massive macrophage infiltration was observed in residual tumor areas of Group 2. Interestingly, considerably less infiltrating macrophages than in the control group were observed after recurrent or delayed tumor growth. Staining against CD45, detecting all leucocytes, including macrophages, confirmed these results (Fig. 4b lower panel; Supporting Information Figs. 1e and 1f). Staining for the M2 macrophage subtype revealed only few positive cells in all tumors, but showed no differences in the staining patterns among the individual groups (data not shown). The extracellular matrix (ECM) is an important regulator of tumor growth, tumor cell phenotype, tumor cell invasion and EMT processes.18–20 To gain more insight regarding ECM composition in this experimental system, staining for collagen was performed using Masson–Goldner trichrome and Picrosirius red. Primary tumors of mice from all groups contain collagen, which is partly found as fibroblast collagen in stromal streaks (Fig. 4c, blue arrows), and also as tumor cell surrounding collagen (Fig. 4c, red arrow). Tumors of control mice are poor in stromal fibers, and only little tumor cell collagen surrounds the cells. After CAF treatment, an altered ECM composition is seen in the tumors (Group 2), where tumor tissues contain considerably more collagen. This includes an increase in stromal fibers as well as more deposition of tumor cell collagen (Fig. 4c; Supporting Information Fig. 1h). Some of the vimentin-positive cells were identified as collagen-producing fibroblasts (Supporting Information Fig. 1g). Tumors after recurrent or delayed tumor growth (Groups 3 and 4) also contain a higher amount of collagen, predominantly consisting of stromal collagen, compared with tumors of the control group. Collagen fiber staining with Picrosirius red (Fig. 4c, lower panel) not only shows a strong increase in collagen deposition in the residual tumor areas after CAF treatment, but also an altered distribution of collagen in tumors after recurrent or delayed growth. Collagen in tumors of the control group is found mainly in the periphery of the tumor, whereas collagen additionally pervades the tumor in all CAF-treated groups (Fig. 4c). Staining for Type I collagen also revealed a strong increase in collagen Type I in tumors 9 days after chemotherapy (data not shown). Taken together, shortly after CAF treatment, the remaining tumor areas consist mainly of mesenchymal-like tumor cells with a massive leukocyte infiltration, including macrophages, and deposition of collagen. A balanced epithelialmesenchymal tumor cell composition is re-established during recurrent or delayed tumor growth.

Phenotype changes in carcinoma after therapy

Mesenchymal cells and cells with stem cell properties survive chemotherapy in vitro

Because analysis of chemotherapy effects on tumors in vivo provides only limited information on underlying molecular mechanisms, the effects of chemotherapy on G-2 tumor cells were examined in vitro. Application of chemotherapy in vitro led to a slight decrease in vital tumor cells within 24 hr, as 70–90% of the tumor cells were still vital (Fig. 5a). After 48 hr of CAF application, the percentage of vital cells decreased further to about 20–30%. After additional 24 hr, i.e., 72 hr after CAF treatment, the fraction of vital cells reduced to 10– 20% and remained at this level for the remaining observation time of 11 days. Only re-application of chemotherapy after 2 weeks followed by continuous cultivation for a total of 24 days was able to reduce the number of vital tumor cells below the detection limit (data not shown). In recovery experiments, G-2 cells were treated with CAF for 24 or 48 hr, respectively, washed and afterward cultured in standard medium for defined periods of time. Independent of how long G-2 cells were incubated with CAF or how long they had been subsequently cultured in fresh medium, the fraction of viable cells remained constant, indicating that after in vitro treatment with CAF these cells were not able to proliferate again (Fig. 5b). To test whether a mouse (in vivo) surrounding would enable such cells to proliferate again, G-2 cells were treated with CAF in vitro for 48 hr followed by recovery for 6 days in DMEM and implanted into NP8 mice. Even 5 months after implantation, no tumor formation could be observed (data not shown). As after CAF treatment residual tumor areas displayed a more mesenchymal morphology, the phenotype of G-2 cells that had survived CAF application for 48 hr in vitro was examined. As the mesenchymal phenotype is characterized by expression of EMT-inducing transcriptions factors, the expression of selected transcription factors was analyzed (by q-PCR, Fig. 5c). It was found out that expression of Zeb1, Twist1, Twist2 and Snai1 was enhanced in G-2 cells after chemotherapy. Recovery in DMEM for 6 days still revealed elevated expression levels, although levels were not as high as directly after treatment. Similarly, expression levels of the EMT markers Mmp3 and Tgfb1i1 were highly elevated after chemotherapy of G-2 cells and remained high after recovery. This supports the assumption that chemotherapy preferably affects epithelial differentiated G-2 tumor cells, whereas G-2 tumor cells in a mesenchymal differentiation state survive. This assumption was further supported by analyzing the expression of vimentin, EpCAM and E-cadherin as markers for mesenchymal and epithelial phenotypes. After 48 hr of chemotherapy, a decrease in cells with an epithelial phenotype and an increase in cells with a mesenchymal phenotype, characterized by a reduction in E-cadherin/EpCAM expression and an increase in vimentin expression, were observed. Cultivation of G-2 cells in DMEM for 6 days did not lead to recovery but, on the contrary, to further reduction of epithelial markers. Expression of vimentin was still strongly

C 2014 UICC Int. J. Cancer: 137, 25–36 (2015) V

33

Carcinogenesis

Jannasch et al.

Figure 4. (a) Number of cells positive for T-Ag is strongly reduced 9 days after application of chemotherapy (Group 2). (b) Macrophages are depicted by using an anti-F4/80 antibody (upper panel). A high number of infiltrating macrophages can be observed 9 days after chemotherapy whereas only a few macrophages that infiltrate the tumors can be observed after recurrent (Group 3) or delayed (Group 4) tumor growth. These results were confirmed using an antibody against CD45 that stains all leucocytes, including macrophages (lower panel). (c) Collagen staining using Masson–Goldner trichrome (upper panel) and Picrosirius red (lower panel) shows variable distribution and hints to different origin of collagen in the diverse groups. CAF-treated Group 2 contains increased amounts of collagen, mainly surrounding mesenchymal tumor cells (red arrow). In Groups 3 (recurrent growth) and 4 (delayed growth), mainly stromal fibroblast collagen can be found (blue arrows), which is also slightly elevated in comparison with the control Group 1. In control Group 1, most collagen can be found at the rim of the tumor as shown by Picrosirius red staining (lower panel), whereas in all other groups collagen is mainly distributed all over the tumor.

C 2014 UICC Int. J. Cancer: 137, 25–36 (2015) V

34

Carcinogenesis

Phenotype changes in carcinoma after therapy

Figure 5. Impact of chemotherapy on G-2 cells in vitro. (a) Proliferation after application of chemotherapy in vitro. The amount of vital tumor cells in percentage of initially seeded G-2 cells 1, 2, 3 and 11 days after treatment with chemotherapy is shown (n 5 3 with repeated determination). (b) Recovery experiment. G-2 cells were treated with CAF for 24 or 48 hr, respectively, and afterwards cultivated in DMEM for the periods indicated. The amount of vital tumor cells in percentage of initially seeded cells is shown (n 5 2 with repeated determination). (c) RNAexpression analysis of CAF-treated G-2 cells. Relative quantification of the expression of phenotypic markers, EMT network markers, EMT transcription factors and stem cell markers in vitro are shown after treatment with CAF for 48 hr and after recovery for 6 days in comparison with untreated G-2 cells (*p < 0.05). (d) Protein expression levels of vimentin and E-cadherin were verified by Western blot analyses.

enhanced after recovery. These findings were confirmed by Western blot analyses of protein expression levels of vimentin and E-cadherin (Fig. 5d). A relationship between a mesenchymal phenotype induced by EMT and stem cell properties has been described.21 Therefore, the expression of three transcription factors (Klf4, Nanog and Sox2), which are necessary to maintain stem cell properties, was analyzed. Treatment of G-2 cells with chemotherapy for 48 hr led to a strongly increased expression of

these stem-cell-associated transcription factors. Recovery of tumor cells in all cases led to a slight reduction in the expression levels compared with expression levels in chemotherapytreated cells, but expression levels were still elevated (Fig. 5c).

Discussion The findings demonstrate that a single round of CAF treatment of mammary carcinoma-bearing mice leads to a significant reduction in tumor volume within a few days, but is not C 2014 UICC Int. J. Cancer: 137, 25–36 (2015) V

able to eliminate all tumor cells in the syngeneic immunocompetent G-2-WAP-T mouse model. Moreover, recurrent tumor growth with increased tumor cell dissemination can be observed. Both incomplete response to chemotherapeutic treatment and recurrent tumor growth are crucial factors in patients and have been described earlier.22,23 Therefore, the tumor model closely mimics the clinical situation. In vivo application of chemotherapy in the model resulted in strong regression of tumors with remaining tumor cells and a massive infiltration of leucocytes, including macrophages. Small residual tumor areas consisted mainly of vimentinpositive mesenchymal-like tumor cells; only a few epithelial-like E-cadherin-positive T-Ag-positive tumor cells remained. Moreover, spindle-shaped nuclei of tumor cells hint to the occurrence of EMT. In support, expression analyses of selected genes in G-2 cells after in vitro CAF treatment showed a decreased E-cadherin, coupled with an increased vimentin expression. The mesenchymal phenotype of cells in CAF-treated tumors, thus, is consistent with the EMT phenotype of CAF-treated G2 cells in vitro, where the expression of EMT-related phenotypic markers as well as of EMT-related transcription factors was considerably increased after CAF treatment. Additionally, the expression of stem cell factors increased after CAF treatment in vitro, which relates to poor prognosis and increased risk of metastases.24,25 Therefore, application of chemotherapy leads to a more mesenchymal tumor phenotype in this model. A mesenchymal phenotype of epithelial tumors corresponds to higher aggressiveness and poor prognosis.26 Although in tumors of CAF-treated mice (Group 2) the strong infiltration of macrophages most likely reflects their phagocytic activity on apoptotic tumor cell remnants, the putative role of tumor-associated macrophages in tumors of the other groups is not yet clear, and is currently under investigation. An increased deposition of collagen was also found in G-2 tumors 9 days after CAF treatment. After recurrent tumor growth, the amount of collagen is still increased and displayed an altered distribution in comparison with control tumors. Enhanced expression and deposition of ECM is a hallmark of tumor progression.18,19 Antibody staining for Type I collagen revealed an increase in this collagen type in tumors after chemotherapy compared with that in the control group. Type I collagen is the prevalent component of the stromal ECM and has been identified as a prognostic marker associated with cancer recurrence in human breast cancer patients.20 In accordance with the aggravated tumor cell phenotype after CAF treatment in vivo and in vitro, it is assumed that in this model the intense deposition of collagen after CAF treatment changes the homeostasis of ECM remodeling in favor of tumor growth, enabling the development of recurrent tumors. CAF application not only had an impact on growth and phenotype of the primary tumor, but also strongly influenced tumor cell dissemination. The number of DTCs decreased considerably 9 days after CAF application. However, recurrent as well as delayed tumor growth was associated with a C 2014 UICC Int. J. Cancer: 137, 25–36 (2015) V

significant increase in DTCs in comparison with that in untreated tumor-bearing mice. Furthermore, G-2 cells that had survived CAF treatment in vitro exhibited tumor stem cell properties and had undergone EMT. It has been suggested that DTCs able to initiate metastasis exhibit exactly these properties.27,28 Therefore, it can be hypothesized that recurrent tumor growth in this model is associated with an increased risk of formation of metastases. An important result of the chemotherapy studies was that in vivo tumor cells surviving CAF treatment in primary tumors were able to proliferate again, whereas in vitro treatment of G-2 cells led to nonproliferating cells. Although G-2 cells after CAF treatment remained vital, they were not able to overcome their growth arrest and to start to proliferate again, even after implantation into mice. It is assumed that the tumor microenvironment plays a decisive role in maintaining the capacity of treated tumor cells to re-establish a vital tumor cell system. A possibly important side-aspect of these results by the authors addresses the possible impact on neoadjuvant therapy. It has been shown that adjuvant chemotherapy reduces the risk of death in breast cancer patients.23 However, 70– 80% of patients receiving this treatment would have survived without it.20 Currently, neoadjuvant chemotherapy is also offered to patients with resectable tumors, as no difference between neoadjuvant and adjuvant treatment in terms of overall survival and the probability of disease relapse is evident.29 However, it is not clear whether therapy response of the primary tumor is representative for entire tumor burden.29 Although the experiments do not allow any conclusion about adjuvant therapy, the results suggest that application of neoadjuvant chemotherapy should be carefully considered. In a new setting, surgical removal of the tumor and subsequent CAF treatment would provide further information about the implications of adjuvant and neoadjuvant therapy. However, although it is believed that this model is appropriate for these kind of experiments, it should be taken into account that a mouse model cannot completely mimic the human situation, e.g., because of differences between mouse and human immune systems.30 Nevertheless, investigations using the G-2-WAP-T model should be well suited for the characterization of chemotherapeutic effects not only on tumor progression and cell dissemination but also on recurrent tumor growth, which is necessary for developing novel therapeutic strategies to target mammary carcinomas in patients.

Acknowledgements The authors acknowledge the excellent technical assistance of Hanna Widera, Bettina Jeep, Julia Schirmer, Sarah Greco, Roswitha Streich, Annette Preuß, Renke Brixel and Gundula Pilnitz-Stolze. Special thanks go to Philipp Str€ obel for pathological consultation. This work was supported by grants from the Deutsche Krebshilfe (109323 to F. Alves and 109315 to W. Deppert). The senior professorship of W. Deppert is supported by the JungStiftung f€ ur Forschung, Hamburg. The authors do not have any duality of interest (both financial and personal).

Carcinogenesis

35

Jannasch et al.

36

Phenotype changes in carcinoma after therapy

References

Carcinogenesis

1.

Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA Cancer J Clin 2013;63:11–30. 2. Mathiesen RR, Fjelldal R, Liestol K, et al. Highresolution analyses of copy number changes in disseminated tumor cells of patients with breast cancer. Int J Cancer 2012;131:E405–15. 3. Braun S, Vogl FD, Naume B, et al. A pooled analysis of bone marrow micrometastasis in breast cancer. N Engl J Med 2005;353:793–802. 4. Janni W, Vogl FD, Wiedswang G, et al. Persistence of disseminated tumor cells in the bone marrow of breast cancer patients predicts increased risk for relapse—a European pooled analysis. Clin Cancer Res 2011;17:2967–76. 5. Husemann Y, Geigl JB, Schubert F, et al. Systemic spread is an early step in breast cancer. Cancer Cell 2008;13:58–68. 6. Fantozzi A, Christofori G. Mouse models of breast cancer metastasis. Breast Cancer Res 2006;8:212. 7. Shultz LD, Ishikawa F, Greiner DL. Humanized mice in translational biomedical research. Nat Rev Immunol 2007;7:118–30. 8. Schulze-Garg C, Lohler J, Gocht A, et al. A transgenic mouse model for the ductal carcinoma in situ (DCIS) of the mammary gland. Oncogene 2000;19:1028–37. 9. Heinlein C, Krepulat F, Lohler J, et al. Mutant p53(R270H) gain of function phenotype in a mouse model for oncogene-induced mammary carcinogenesis. Int J Cancer 2008;122:1701–9. 10. Jannasch K, Dullin C, Heinlein C, et al. Detection of different tumor growth kinetics in single transgenic mice with oncogene-induced mammary carcinomas by flat-panel volume computed tomography. Int J Cancer 2009;125:62–70. 11. Otto B, Gruner K, Heinlein C, et al. Low-grade and high-grade mammary carcinomas in WAP-T transgenic mice are independent entities distinguished by Met expression. Int J Cancer 2013; 132:1300–10.

12. Otto B, Streichert T, Wegwitz F, et al. Transcription factors link mouse WAP-T mammary tumors with human breast cancer. Int J Cancer 2013;132:1311–22. 13. Krepulat F, Lohler J, Heinlein C, et al. Epigenetic mechanisms affect mutant p53 transgene expression in WAP-mutp53 transgenic mice. Oncogene 2005;24:4645–59. 14. Wegwitz F, Kluth MA, Manz C, et al. Tumorigenic WAP-T mouse mammary carcinoma cells: a model for a self-reproducing homeostatic cancer cell system. PLoS One 2010;5:e12103. 15. Levine MN, Pritchard KI, Bramwell VH, et al. Randomized trial comparing cyclophosphamide, epirubicin, and fluorouracil with cyclophosphamide, methotrexate, and fluorouracil in premenopausal women with node-positive breast cancer: update of National Cancer Institute of Canada Clinical Trials Group Trial MA5. J Clin Oncol 2005;23:5166–70. 16. Shrivastav S, Stone KR, Paulson DF, et al. Activation of cyclophosphamide for in vitro testing of cell sensitivity. Cancer Res 1980;40:4443–5. 17. Lenfert E, Maenz C, Heinlein C, et al. Mutant p53 promotes epithelial-mesenchymal plasticity and enhances metastasis in mammary carcinomas of WAP-T mice. Int J Cancer 2014; doi: 10.1002/ ijc.29186. 18. Provenzano PP, Eliceiri KW, Campbell JM, et al. Collagen reorganization at the tumor-stromal interface facilitates local invasion. BMC Med 2006;4:38. 19. Provenzano PP, Inman DR, Eliceiri KW, et al. Collagen density promotes mammary tumor initiation and progression. BMC Med 2008;6:11. 20. van’t Veer LJ, Dai H, van de Vijver MJ, et al. Gene expression profiling predicts clinical outcome of breast cancer. Nature 2002;415:530–6. 21. Tam WL, Weinberg RA. The epigenetics of epithelial-mesenchymal plasticity in cancer. Nat Med 2013;19:1438–49.

22. Rottenberg S, Nygren AO, Pajic M, et al. Selective induction of chemotherapy resistance of mammary tumors in a conditional mouse model for hereditary breast cancer. Proc Natl Acad Sci U S A 2007;104:12117–22. 23. Chang JC, Wooten EC, Tsimelzon A, et al. Patterns of resistance and incomplete response to docetaxel by gene expression profiling in breast cancer patients. J Clin Oncol 2005;23:1169–77. 24. Idowu MO, Kmieciak M, Dumur C, et al. CD44(1)/CD24(-/low) cancer stem/progenitor cells are more abundant in triple-negative invasive breast carcinoma phenotype and are associated with poor outcome. Hum Pathol 2012;43:364–73. 25. Martin TA, Jiang WG. Evaluation of the expression of stem cell markers in human breast cancer reveals a correlation with clinical progression and metastatic disease in ductal carcinoma. Oncol Rep 2014;31:262–72. 26. Taube JH, Herschkowitz JI, Komurov K, et al. Core epithelial-to-mesenchymal transition interactome gene-expression signature is associated with claudin-low and metaplastic breast cancer subtypes. Proc Natl Acad Sci U S A 2010;107: 15449–54. 27. Clarke MF, Dick JE, Dirks PB, et al. Cancer stem cells—perspectives on current status and future directions: AACR Workshop on cancer stem cells. Cancer Res 2006;66:9339–44. 28. Gupta PB, Chaffer CL, Weinberg RA. Cancer stem cells: mirage or reality? Nat Med 2009;15:1010–2. 29. Hartkopf AD, Taran FA, Wallwiener M, et al. The presence and prognostic impact of apoptotic and nonapoptotic disseminated tumor cells in the bone marrow of primary breast cancer patients after neoadjuvant chemotherapy. Breast Cancer Res 2013;15:R94. 30. Mestas J, Hughes CC. Of mice and not men: differences between mouse and human immunology. J Immunol 2004;172:2731–8.

C 2014 UICC Int. J. Cancer: 137, 25–36 (2015) V

Chemotherapy of WAP-T mouse mammary carcinomas aggravates tumor phenotype and enhances tumor cell dissemination.

In this study, the effects of the standard chemotherapy, cyclophosphamide/adriamycin/5-fluorouracil (CAF) on tumor growth, dissemination and recurrenc...
1MB Sizes 0 Downloads 8 Views