Breast cancer cell CD200 expression regulates immune response to EMT6 tumor cells in mice
- First Online:
- Cite this article as:
- Gorczynski, R.M., Chen, Z., Diao, J. et al. Breast Cancer Res Treat (2010) 123: 405. doi:10.1007/s10549-009-0667-8
- 325 Views
CD200 has been characterized as an important immunoregulatory molecule, increased expression of which can lead to decreased transplant rejection, autoimmunity, and allergic disease. Elevated CD200 expression has been reported to be associated with poor prognosis in a number of human malignancies. We have found that cells of the transplantable EMT6 mouse breast cancer line growing in vitro express low levels of CD200, but levels increase markedly during growth in immunocompetent mice. Similar increased in vivo expression does not occur in NOD-SCID.IL-2γr−/− mice or mice with generalized over-expression of a CD200 transgene. In both mice, tumor growth occurs faster. Altered CD200 expression in control versus transgenic mice is accompanied by reproducible changes in tumor-infiltrating host cells, and altered cell composition in lymph nodes draining the tumor (DLN). Neutralization of expressed CD200 by anti-CD200mAbs leads to decreased tumor growth in immunocompetent mice, with improved detection of cytotoxic anti-tumor immune cells in DLN. Finally, we report that tumor growth in vivo can be monitored by levels of soluble CD200 (sCD200) in serum of tumor-bearing animals.
KeywordsBreast cancerTregCD200 transgeneImmunotherapy
Aberrant signaling from the microenvironment and cell autonomous mechanisms, such as microRNA or aberrant chromosomes, contributing to malignant transformation through introduction of defects in mitotic control and apoptosis  are independent routes whereby cell proliferation and malignancy are fostered . Many of the signals within the external environment which are implicated in regulating tumor growth, including integrin-binding extracellular matrix molecules, and other soluble mediators (tumor necrosis factor (TNF) family members such as 4-1BBL , cytokines, and Toll-like receptor molecules ), also regulate host defense mechanisms. We have characterized immunoregulation and anti-inflammatory activity following interaction between the two cell surface proteins CD200:CD200R. There is a growing body of information implicating a role for CD200 expression in cancer progression in several human solid tumors [5–8] and hematological tumors, including CLL , AML , and multiple myeloma (MM) . The mechanism(s) whereby CD200 expression by tumors is related to disease status and regulates the response to therapy remains essentially unexplored. However, increased inflammation and activation of myeloid-derived cells occurs in CD200KO mice, with increased numbers of CD200R+ cells , and the increased expression of cell surface CD200 (mCD200) on human CLL tumors is reported to cause T cell suppression in vitro .
One hypothesis is that tumor cells exploit the CD200–CD200R pathway as a means of immune evasion. This suggests that CD200 blockade could represent a novel strategy for cancer immunotherapy . It has recently been hypothesized that CD200 expressed on cancer stem cells (CSCs)  may be relevant to control of disease, and help explain why targeting differentiated cells has generally proven unsuccessful . In the studies described below, we asked whether, in immunocompetent mice, growth of EMT6 mouse breast cancer cells was associated with increased expression of membrane CD200 (mCD200) and altered tumor-infiltrating cells and/or tumor-draining lymph node populations, and whether tumor growth in vivo in normal mice is modulated by anti-CD200 mAb therapy. We also investigated mCD200 expression on EMT6 cells in an immunocompromised environment, using NOD-SCID.IL-2γr−/− mice or CD200 transgenic mice (CD200tg) with systemic over-expression of CD200 after doxycycline induction .
Materials and methods
Stock male and female BALB/c and C57BL/6 mice were purchased from the Jackson laboratories, Bar Harbour, Maine. (BALB/cxC57BL/6)F1 (BBF1) mice were bred at the Toronto Hospital. All mice were housed 5/cage and allowed food and water ad libitum. All mice were used at 8–12 weeks of age. All animal experimentation was performed following guidelines of an accredited animal care committee (protocol No. AUP.1.5). NOD-SCID.IL-2γr−/− mice (BALB/c background) were bred and maintained at the Toronto Hospital.
Homozygous rtTA2s-M2 CD200tg mice (on a C57BL/6 background), in which CD200 expression was induced by Doxycycline (Dox) in the drinking supply (1 μg/ml), are described elsewhere . Both transgenes (rtTA2s-M2 and CD200) were backcrossed onto the BALB/c background and the heterozygote transgenic mice used below (BBtgF1) were all at the 5th backcross generation, and showed that ubiquitous over-expression of CD200 following Dox for 7 days .
Real-time PCR for genotyping and quantitating mRNA expression (e.g., of CD200) was performed using methodology/primers discussed elsewhere .
FITC-conjugated rat anti-mouse CD200 was obtained from Serotec. PE- and FITC-labeled monoclonal antibodies (mAbs) to mouse CD4, CD8, CD11b, CD11c, GR1, CD25, MAC-3, and Foxp3 were obtained from Pharmingen (San Diego, CA, USA) and used as per the supplier’s instructions permeabilization for intracellular staining with anti-Foxp3 and other antibodies (see below) used a CytoFix/CytoPerm kit from BD Pharmingen. 105 cells were incubated with PE- (or FITC-) coupled mAbs, or the respective isotype controls, at 4°C for 60 min in PBS with 2% mouse serum to block Fcr binding. Cells were washed thrice with PBS and analysed in a Cytomics cytometer using Cytomics™ software (Beckman Coulter, Miami, FL, USA) as previously described .
Custom large scale preparation of a previously described rat anti-mouse CD200 (10A5) was performed by Cedarlane Labs (Hornby, Ontario, Canada). Paired mAbs for cytokine ELISAs (anti-IL-2, -IFNγ, -TNFα, and -IL-10), and recombinant cytokines were obtained from Pharmingen . Two independent antibodies were used interchangeably (after permeabilization of cells) to stain for tumor cell content in cell populations. A goat anti-human GRP78 antibody Ab39, recognizing a determinant at the cell surface of a variety of cancers  and cross-reactive with EMT6 (see below) was purchased from Santa Cruz Biotechnologies (Santa Cruz, CA, USA). In addition, a mAb (clone 35C1) for a breast cancer-amplified kinase (BTAK-1) was obtained from Hycult Biotechnology (Netherlands) .
An ELISA assay to detect a soluble form of CD200 (sCD200) in serum of normal and tumor-bearing mice  used as capture a rat anti-mCD200 mAb (10A5), with a newly developed heterologous rabbit anti-mouse CD200 as developing reagent (along with HRP-conjugated anti-rabbit Ig and appropriate substrate; see text and figure for details). This assay detected a molecularly cloned soluble CD200, in which the extracellular domains of CD200 were linked to a murine Fc region, CD200Fc , at a sensitivity of 20 pg/ml, with no reactivity with purified mouse Fc or mouse IgG (Wong et al. 2009, in press).
EMT6 breast tumor cells and induction of tumor growth in BBF1 mice
The breast tumor cell line (EMT6 growing in female BALB/c mice), used to model hormonal, radiation, and chemotherapeutic effects on primary and metastatic breast cancer growth , was obtained from ATCC and passaged in vitro in α-minimal essential medium (α-MEM) medium with 10% fetal calf serum (αF10). BBF1 (or BBtgF1) mice received 5 × 105 tumor cells subcutaneously in 100 μl PBS, with tumor growth monitored daily thereafter. In some cases, mice received Dox drinking water. Mouse tumors were harvested and cells digested with mixture of collagenase, trypsin, and DNase. Cells were centrifuged over mouse lymphopaque (Cedarlane Labs, Hornby, Ontario, Canada), and resuspended in αF10. Tumor-draining lymph nodes (DLN) were harvested from the same mice.
In some studies, we used EMT 6 tumor cells transduced with a retrovirus encoding murine CD200 (plmCD200sn) or a blank vector control (plxsn), and selected with G418 (1 mg/ml) for 2 weeks . CD200 expression was monitored by flow cytometry, and highly expressing cells cloned following FACS sorting. Stably transduced tumor cells were maintained in G418 (0.5 mg/ml) until used in vivo.
Unsorted and CD200+/CD200− cells were enriched from tumor digests by FACS sorting (FITC-anti-mouse CD200), and permeabilized for staining with anti-GRP78 or -BTAK to determine the percentage of tumor cells in the population. Throughout these studies, approximately 30–40% of unsorted cells in the tumor digest stained for BTAK and GRP78.
Immunohistochemistry of tumor tissue sections were performed using formalin fixed, paraffin embedded sections. Samples were immunostained with a double-labeled indirect immunofluorescence procedure using mixtures of first antibodies (goat polyclonal anti-GRP78, or rabbit anti-mCD200) followed by FITC- (anti-goat) or rhodamine- (anti-rabbit) conjugated second antibodies.
Preparation of cells and cytotoxicity, proliferation, and cytokine assays
Single cell suspensions were prepared aseptically from DLN of mice and cells resuspended in α-MEM supplemented with 2-mercaptoethanol and 10% fetal calf serum (αF10). In tumor cell, stimulation assays used to assess induction of cytotoxic T cells and/or cytokine production, 5×106 BBF1 (or BBtgF1) responder DLN from control or tumor-bearing mice were stimulated with 2 × 104 irradiated (2500 Rads) cultured EMT6 tumor stimulator cells in duplicate in αF10. Supernatants were pooled at 40 h from replicate wells and assayed in triplicate in ELISA assay, with capture and biotinylated detection mAbs as reported . Varying volumes of supernatant were bound in triplicate at 4°C to plates pre-coated with 100 ng/ml mAb, washed thrice, and biotinylated detection antibody added. After washing, plates were incubated with streptavidin–horseradish peroxidase (Cedarlane Labs), developed with appropriate substrate and OD405 determined using an ELISA plate reader. Recombinant cytokines for standardization were obtained from Pharmingen (USA), with assay sensitivity in the range ~40 pg/ml. When cytotoxicity was assayed, cultures were harvested at 6 days, and cells were titrated at various effector:target ratios for killing (18 h at 37°C) of 51Cr-labeled EMT6 targets.
In cultures, used to assess the presence of regulatory cell populations, CD4+ T cells isolated from tumor-bearing mice were used as a regulatory pool source. CD4+ cells were purified from spleen cell preparations by negative selection using a CD4+ T Cell Isolation Kit according to the manufacturer’s instructions (StemCell Technologies, Vancouver, Canada). 5×106 responder lymph node cells were used from a pool of three BBF1 mice challenged 14 days earlier with 10 × 106 irradiated cultured (CD200−) EMT6 tumor cells subcutaneously, and 1 × 106 regulatory DLN were added as described. Mixed cultures received stimulation with 2 × 104 irradiated (2500 Rads) cultured (CD200−) EMT6 tumor stimulator cells in duplicate as described above.
Within experiments comparison between groups used ANOVA, with subsequent paired t-tests as indicated.
Increased expression of membrane CD200 on EMT6 cells in immunocompetent mice
Tumor cell digests from immunocompetent mice expressed more mCD200 than tumors grown in immunocompromised NOD-SCID.IL-2γr−/− mice (panel a). When tumor cells were stained, after permeabilization, for either GRP78 or breast cancer antigen BTAK-1 [17, 18], in immunocompetent mice, a similar number of mCD200+ and mCD200− BTAK-1+ cells was detected (~60% of CD200+ cells and ~15% of CD200− cells stained for BTAK/GRP78). However, the majority of BTAK-1+ cells in NOD-SCID.IL-2γr−/− mice were mCD200− (panel b), supporting the hypothesis that mCD200+ BTAK-1+ cells have a growth advantage only in immunocompetent mice. Tumors in NOD-SCID.IL-2γr−/− mice were approximately 2-fold larger than tumors in control immunocompetent mice, implying a host resistance mechanism in the latter which was absent in NOD-SCID.IL-2γr−/−.
In order to confirm a role for CD200 expression in regulation of growth of EMT6 in vivo, we also compared growth of 5 × 105 EMT6 cells in BBF1 (or BBtgF1) mice, with all mice receiving Dox (to induce ubiquitous over-expression of CD200 in BBtgF1). Tumors harvested at day 16 were measured (volume) and CD200+/CD200− cells assessed as above for BTAK-1 expression. These data are also shown in Fig. 1b, and confirm that in an immunosuppressed environment following systemic over-expression of mCD200, EMT6 BTAK+ tumor cells have low mCD200 expression, while tumor growth is enhanced. Note that in BBtgF1 (~50% CD200+ cells in the tumor digest), some 15% of CD200+ cells and 50% of CD200− cells stained for BTAK/GRP78, in contrast (see above) to the staining pattern in BBF1 or BALB/c.
In an alternate approach, we measured tumor volume at day 16, CD200 mRNA and tumor cell CD200 expression (FACS sorting followed by BTAK staining) using tumor digests obtained from groups of three control BALB/c or NOD-SCID.IL-2γr−/− mice receiving 5 × 105 EMT6 cells stably transfected with an “empty” retrovirus vector or with a vector expressing CD200. These data, in Fig. 1c, confirm that superior growth occurred for the CD200-transfected EMT6 cells only in normal (but not immunoincompetent) BALB/c mice, and that both increased CD200mRNA and cell surface expression was seen in control BALB/c versus NOD-SCID.IL-2γr−/− mice receiving control EMT6 cells.
Immunohistology sections (panel d) of EMT6 growing in control or CD200tg mice (first two columns of panel d), or of CD200-transfected EMT6 in control mice (last column in panel d), lend further support to the hypothesis that tumors in control (but not CD200tg) mice showed that high levels of dual staining with both anti-CD200 and GRP78 antibodies (bottom row).
Comparison of DLN cells and tumor-infiltrating cells in BBF1 and BBtgF1
Data in Fig. 2 support the hypothesis that decreased tumor growth in normal versus NOD-SCID.IL-2γr−/− (or BBtgF1 mice; see Fig. 1) reflects immune resistance in control animals. While CD4+, CD4+CD25+ Foxp3−, and MAC-3+ cells in DLN were increased in control mice at early (CD4+: days 9–12) and later times (MAC-3+: days 15–18) following tumor induction, GR1+, and CD4+CD25+Foxp3+ cells increased from approximately days 12–15. Within the tumor mass (Fig. 3), we confirmed the increase in CD200+BTAK+ tumor cells in control mice compared with BBtgF1 mice. Consistent with data from the DLN cells, the tumor infiltrate contained approximately 2-fold less CD4+Foxp3− and CD4+Foxp3+ cells in BBtgF1 mice compared with cells in BBF1 animals. TILs in both sets of mice had a high frequency of CD11b+, MAC-3+ cells, and GR1+ cells (see legend to Fig. 3).
Preferential cloning of tumor cells in vitro from CD200+ not CD200− cells
Whether tumor cells were cloned from BBF1 or BBtgF1 mice, CD200+ cells showed that approximately 10-fold greater cloning efficiency that CD200− cells. Although, there were increased numbers of BTAK+CD200+ cells in BBF1 mice (Fig. 4), tumors in immunocompromised BBtgF1 mice grew faster than in BBF1 as shown in Fig. 1. These data support the hypothesis that in BBF1 mice an ongoing immune response exists which attenuates tumor growth.
Cytotoxicity and cytokine production by EMT6 stimulated DLN from tumor-bearing mice
As analysed by EMT6-stimulated cytokine production, BBF1 mice developed a greater immunity to EMT6 cells than did BBtgF1 mice (Fig. 5a), with only IL-10 produced at greater levels from the DLN of the latter. Assaying cytotoxicity for EMT6, DLN from BBtgF1 mice showed that no increased immunity over control (non-tumor injected mice) at any time post-tumor injection, while BBF1 mice showed that a peak response by approximately days 12–15 following tumor injection, with a decline in cytotoxicity at later times (day 18). The decline in cytotoxic responses from BBF1 with time, and the absence of a significant cytotoxic response in BBtgF1 mice (Fig. 5b), was associated with CD4+ cells able to suppress the cytotoxic response from EMT6-immune BBF1 mice, as shown in Fig. 5c.
Anti-CD200mAb augments immune response in BBF1 mice and decreases tumor growth
Whether judged by tumor growth or development of tumor directed cytotoxic responses, ongoing infusion of anti-CD200mAb into BBF1 mice augmented tumor resistance. No similar effects were observed after mAb treatment of BBtgF1 mice, despite evidence for detectable anti-CD200 mAb in the serum of transgenic recipients (i.e., not all of the anti-CD200mAb had been neutralized by the over-expressed CD200 transgene).
Evidence for a soluble form of CD200 in BBF1 mice whose levels parallel tumor growth
In human chronic lymphocytic leukemia patients, we reported a correlation between patient tumor burden and the presence of a soluble form of CD200, sCD200, in serum . This sCD200 may be derived from metalloprotease cleavage of the cell surface CD200, as much as, has been described for a number of other membrane-bound proteins, including EGF, heparin-binding growth factor and the TNFα family member, 4-1BBL . In order to explore whether a soluble form of CD200 was present in mice with EMT6 tumor cells, we sampled serum from BBF1 mice at various times post-tumor inoculation, analyzing sCD200 using an ELISA assay as described in the “Materials and methods”. Figure 6c shows a comparison of tumor volume and serum sCD200 levels for six mice/group, indicating that sCD200 levels increase greater than 4-fold during tumor growth.
Tumor immunotherapy aims to provide active and/or passive immunity against malignancies by harnessing the immune system to target tumors. We reported that a soluble form of CD200, linking the extracellular domain to an IgGFc region (CD200Fc), suppressed tumor immunity . Human lymphoma cells have been shown to express mCD200 at high levels in >70% of tumor isolates . Cells transfected to over-express CD200 suppressed anti-tumor immunity, an effect ameliorated by anti-CD200 antibody . mCD200 expression on melanoma, multiple myeloma, acute myeloid leukemia, and follicular lymphoma tumors has been correlated with poor prognosis [5, 10]. Most studies have not investigated how CD200 modulates tumor resistance, and to the best of our knowledge, no studies to date have addressed such issues in vivo in a breast cancer model.
The mechanism(s) responsible for tumor immunity are themselves not always well-characterized. There is evidence for both suppression of antibody and T cell immune reactivity in loss of control of tumor growth, e.g., suppressed type-1 (IFNγ/IL-2) and elevated type-2 (IL-4/IL-13) cytokine production, and of macrophage (Mph) activation . All of these responses can be regulated by CD200:CD200R . The role of macrophages in tumor immunity is, however, highly controversial, and in several human cancers, an abundance of macrophages in the tumor microenvironment is associated with poor prognosis . Induction of populations of regulatory cells (Tr) has also been implicated in diminished anti-tumor reactivity [28, 29]. Many Tr populations exist including: (i) CD4+CD25brightFoxp3+ cells ; (ii) Tr-releasing cytokines (IL-10 and/or TGFβ) for function ; and (iii) Foxp3− Tr, including NKT , γδTCR+ , and CD4−CD8− (DNT) cells . Importantly, CD200:CD200R1 interactions can induce both IL-10/TGFβ-producing Tr cells  and Foxp3+Treg . Myeloid-derived suppressor cells (CD200R+) are implicated in suppression of tumor immunity  and in a BALB-neuT breast cancer model a decline in (GR1+) cells correlated with decreased tumor growth . Circulating myeloid-derived suppressor cells levels are reported to be correlated with clinical cancer stage and metastatic tumor burden .
The data reported in this manuscript support the hypothesis that in an immunocompetent environment breast cancer cells expressing mCD200 are selected for growth. Selection does not take place in an immunodeficient environment (NOD-SCID.IL-2γr−/−), or in an environment in which attenuation of immunity occurs itself through systemic over-expression of CD200 (BBtgF1 mice); see Fig. 1b. CD4+CD25+ cells are increased in DLN of tumor-bearing BBF1 mice relative to BBtgF1 with an increased in the presence of Foxp3+ and GR1+ cells in both the DLN and TIL population, consistent with the hypothesis that immunity generated in BBF1 mice was regulated by Foxp3+ cells and/or GR1+ cells. Elevated cytokine production and induction of cytotoxicity in DLN from BBF1 mice was seen compared to the “immunosuppressed” BBtgF1 (Fig. 5a, b), and blockade of CD200 expression by anti-CD200mAb further attenuated tumor growth and improved cytotoxic immune responses in BBF1 recipients (Fig. 6a, b). These data support the hypothesis that over-expression of CD200, whether on tumor cells themselves or in the surrounding environment, inhibits tumor immunity, and augments tumor growth in this breast cancer model, the first evidence that CD200 expression may be an important modulator of host resistance in solid(breast) tumors.
A cancer stem cell (CSC) hypothesis was developed to help explain the failure of tumor therapy and the dormant behavior exhibited by many solid tumors, including breast cancer . This hypothesis posits that solid tumors are sustained by subpopulations of CSCs and that therapies targeting mature (differentiated) tumor cells in the tumor burden are doomed to failure . In support of this hypothesis, the “gene signatures” of CSCs are often independent prognostic indicators of tumor growth/outcome . Like mesenchymal stem cells, CSCs have been reported to express surface CD200 , and thus mCD200 expressed by CSCs may be an important target for future exploration.
EMT6 tumor cells obtained from tumor digests stained (after permeabilization; see Fig. 1d) with both a goat antibody to the glucose-regulated protein of 78 kDa (GRP78), a marker for breast and other tumor cells  and an antibody to a mitotic serine/threonine kinase Aurora-A (breast cancer-amplified kinase [BTAK-1]) which plays a critical role in regulating centrosome segregation and spindle assembly, and whose overexpression leads to tumor progression . Separation of tumor digests from EMT6 injected mice into CD200+/CD200− populations before cloning in vitro revealed that CD200+ cells giving rise to clonable tumors at a higher frequency (approximately 8-fold greater) than CD200− cells; see Fig. 5. Despite the relative increase in clonable CD200+ cells in immunocompetent BBF1 versus the BBtgF1 counterpart (see Figs. 1b, 2, 3, and 4), tumors grew slower BBF1, consistent with the hypothesis that an immune response in those animals attenuated tumor growth.
It is acknowledged that the tumor micro- and macro-environment plays a role in cancer growth, progression, invasion, and metastasis. Delivery and receipt of signals for tumors can occur from both membrane-bound forms of ligands and their soluble counterparts, released by ectodomain shedding or by other means (e.g., exocytosis) . Interestingly, pre-operative serum levels of a tissue inhibitor of matrix metalloproteinase-1 (TIMP-1) enzymes implicated in shedding membrane molecules were recently associated with outcome in breast cancer patients . We recently reported on a soluble form of CD200 (sCD200) in serum of leukemia patients and speculated that sCD200 may have a role in cancer progression . In the EMT6 breast cancer model described tumor growth in vivo is also correlated with serum sCD200 levels (Fig. 6c), and in further studies (not shown), a similar correlation was observed for a mouse thymoma (EL4) in C57BL/6 mice. This is the first evidence of a potential role for a soluble form of CD200 (sCD200) in monitoring tumor growth in vivo. We speculate that measurement of sCD200 may provide useful information in monitoring cancer growth, and that regulation by CD200:CD200R is important in controlling solid tumor (as well as hematological tumor) growth in vivo.
This study was supported by a POP II grant from the Canadian Institutes of Health Research (CIHR) to RMG.