Long-term immunity elicited by antibody–cytokine fusion proteins protects against sequential challenge with murine mammary and colon malignancies

Abstract

We have previously reported that the antibody fusion proteins anti-HER2/neu IgG3 fused to IL-12 [(IL-12)-IgG3] or GM-CSF [IgG3-(GM-CSF)] independently or in combination are effective anti-tumor agents against D2F2/E2 murine mammary cancer cells expressing human HER2/neu in the peritoneum. Importantly, the long-term survivors were immune to the subcutaneous challenge with D2F2/E2 and the parental D2F2 not expressing HER2/neu. We now show that these long-term survivors also exhibit significant protection against subsequent subcutaneous challenge with the murine colon carcinoma CT26-HER2/neu, and later against subcutaneous challenge with the parental CT26. These results suggest that the long-term systemic protection against mammary cancer elicited by treatment with antibody–cytokine fusion proteins can be extended to prevent the growth of a tumor from different origin expressing HER2/neu, and that this protection is not limited to this antigen alone, since it also prevented the growth of the parental tumor cells.

Introduction

Recent improvements in the therapy of cancer have considerably extended the life expectancy of patients with primary breast and colon tumors. However, due to this higher number of survivors, there has been an increase in the frequency of patients with subsequent primary cancers [42]. Conditions in which patients exhibit different tumors that show a behavior, histology and date of diagnosis that are distinct from the initial tumor, characteristics that distinguish them from metastatic tumors or cases of relapse, are known as metachronous cancers [31]. Patients with metachronous tumors also reveal poorer survival rates compared to patients with only a single primary cancer diagnosis over time [31, 42]. Multiple primary cancers may arise due to several factors, including exposure to carcinogens in the environment, genetic predisposition, secondary effects of the therapy for the first cancer, immunological deficiency, or a combination of the above [11].

Breast cancer is one of the most frequently diagnosed cancers in the world and the occurrence of metachronous tumors away from the primary site has been widely documented [11, 14, 31, 42]. In breast cancer patients, the most frequent location of metachronous tumors includes the contra-lateral breast, skin, gynecologic region (ovary, uterus, cervix), gastrointestinal tract (colon, rectum, esophagus, pancreas) or lung tissues [11, 14, 31, 42]. Frequently, those tumors have the potential to turn into peritoneal carcinomatosis as metastatic disease [17, 24, 34], and in some cases exhibit HER2/neu overexpression [38]. HER2/neu, a 185 kDa type I transmembrane receptor tyrosine kinase and member of the epidermal growth factor receptor family, is a well-characterized tumor associated antigen (TAA) that activates signal transduction pathways involved in cell growth control [21]. The overexpression of HER2/neu in subsets of ovarian, breast and colon cancer is associated with poor prognosis and aggressive tumor attributes [22, 39]. Herceptin (Trastuzumab, Genentech Inc., San Francisco, CA, USA) is a humanized monoclonal antibody that targets the extracellular domain of HER2/neu. Treatment of metastatic breast cancer with Herceptin monotherapy has resulted in a response rate of 12–20% [2, 35], which can be further improved when used in combination with chemotherapeutics [44]. Unfortunately, the majority of breast cancer patients with advanced disease who initially respond to Herceptin-based treatments acquire resistance within 1 year [10, 26, 40]. Furthermore, Herceptin lacks considerable effectiveness against non-breast tumors expressing HER2/neu such as ovarian [3], pancreas [33], prostate [45], and lung cancer [4].

Antibodies fused to cytokines targeting TAAs have proven to be effective anti-tumor agents in murine models, retaining the tumor targeting capacity and eliciting an immune response against the disease [19, 37]. In our laboratory we have constructed a family of antibody fusion proteins (AbFPs) targeting HER2/neu composed of the variable region of Herceptin and the constant region of human IgG3 genetically fused to different cytokines: human interleukin (IL)-2 [IgG3-(IL-2)], murine single chain IL-12 [(IL-12)-IgG3] or murine granulocyte-macrophage colony-stimulating factor (GM-CSF) [IgG3-(GM-CSF)] [9, 27, 30]. For the IgG3-(IL-2) we used the human cytokine because it is also functional in mice [30]. In the cases of IL-12 and GM-CSF we used the murine cytokines due to the lack of activity of the human versions in mice [32, 36].

We have previously studied the use of (IL-12)-IgG3, IgG3-(IL-2) and IgG3-(GM-CSF) AbFPs alone or in combination as effective therapeutics for the treatment of murine models of different peritoneal tumors of epithelial origin expressing human HER2/neu [20]. In those studies, mice were challenged intraperitoneally (i.p.) with the syngeneic tumor cell lines D2F2/E2, CT26-HER2/neu, and MC38-HER2/neu, and subsequently treated i.p. for 3 days with the AbFPs. Among the models tested, mice challenged with the syngeneic mammary tumor cell line D2F2/E2 expressing human HER2/neu and later treated with (IL-12)-IgG3 in combination with IgG3-(GM-CSF), (IL-12)-IgG3 alone, and to a lesser extent IgG3-(GM-CSF), resulted in the largest number of survivors [20]. Moreover, those survivors demonstrated systemic protection and immunological memory that prevented tumor growth after later subcutaneous (s.c.) challenge with D2F2/E2 tumor cells [20]. We also reported that the long-term survivors treated with (IL-12)-IgG3 alone or combined with IgG3-(GM-CSF) achieved significant protection after subsequent s.c. challenge with the parental cell line D2F2 not expressing human HER2/neu, compared to control time matched naïve mice [20]. This observation suggests that the immune response generated by the treatment against the original challenge with mammary cancer cells expressing HER2/neu elicited epitope spreading against antigens different of the originally targeted [20]. This response was strong enough to protect 100% of the mice treated with (IL-12)-IgG3 in combination with IgG3-(GM-CSF) against the D2F2 challenge, averting one mechanism of tumor evasion to immunotherapies that is the down-regulation of the TAA [23]. In the present study we explored the quality of the immune memory elicited by the treatment with the combination of AbFPs in a model of metachronous tumor, challenging the long-term survivors with a tumor from different origin, the murine colon cell line CT26-HER2/neu, and afterward with the parental CT26.

Materials and methods

Cell lines

The murine colon carcinoma cell lines CT26 (syngeneic for BALB/c) (Generously provided by Dr. Young Chul Sung, Pohang University of Science and Technology, South Korea) and its derivative expressing human HER2/neu (CT26-HER2/neu [29]) were cultured in Iscove’s modified Dulbecco Medium (IMDM) (Invitrogen, Carlsbad, CA, USA) supplemented with 2 mM l-glutamine, 10 U/ml penicillin, 10 μg/ml streptomycin (Sigma Chemical, St. Louis, MO, USA) and 5% calf serum (Atlanta Biologicals, Norcross, GA, USA) at 37°C and 5% CO₂. Dr. Wei-Zen Wei (Wayne State University, Detroit, MI, USA) kindly provided the mammary tumor cell lines syngeneic for BALB/c mice D2F2 and D2F2/E2, the latter expressing human HER2/neu [43]. Both cell lines were cultured in IMDM supplemented with 2 mM l-glutamine, 10 U/ml penicillin, 10 μg/ml streptomycin and 10% calf serum at 37°C and 5% CO₂. The construction, purification and characterization of the anti-HER2/neu IgG3, IgG3-(GM-CSF), and (IL-12)-IgG3, cytokine fusion proteins expressed in murine myeloma cells have been described previously [9, 27].

Animal tumor challenges

Groups of 8 female BALB/c mice 6–8 weeks old (Taconic Farms, Germantown, NY, USA) were challenged i.p. at day 0 with 2 × 10⁵ D2F2/E2 cells in 150 μl Hank’s buffered saline solution (HBSS) (Invitrogen, Carlsbad, CA, USA). Afterwards, animals were treated with one daily i.p. injection for 3 days of PBS, 2.3 μg of (IL-12)-IgG3, the equivalent molar amount of IgG3, IgG3-(GM-CSF), and a mixture of (IL-12)-IgG3 and IgG3-(GM-CSF). On day 134, long-term survivors and age-matched naïve mice were challenged s.c. on the lower right flank with 2 × 10⁵ D2F2/E2 cells. As control for all the following challenges we used naïve mice, since the PBS and IgG3 treated mice died within the first 6 weeks after i.p. tumor challenge. We repeated the procedure on day 183, challenging long-term survivors and naïve mice (n = 7) s.c. on the left lower flank with the parental cell line D2F2 (2 × 10⁵ cells in 150 μl HBSS). Later (day 300) we challenged s.c. the long-term survivors and control naïve mice (n = 8) in the upper right flank with CT26-HER2/neu (10⁶ cells in 150 μl HBSS). Finally, at day 448 we challenged the remaining long-term survivors and control naïve mice (n = 8) on the opposite flank (s.c. upper left flank) with the parental CT26 tumor cell line (10⁶ cells in 150 μl HBSS) and monitored for 52 days. Tumor progression was followed with caliper on a daily basis, and once the tumor reached 15 mm in diameter the animals were euthanized. Differences in survival between experimental groups were determined by the non-parametric Breslow–Gehan–Wilcoxon test.

Results

The effectiveness of (IL-12)-IgG3 treatment alone and in combination with IgG3-(GM-CSF) prompted us to test a condition of metachronous tumors in this model. For this purpose, we challenged D2F2 long-term survivors and naïve controls with a different syngeneic tumor model, the colon carcinoma cell line CT26 expressing human HER2/neu (CT26-HER2/neu) (Fig. 1). (IL-12)-IgG3 and IgG3-(GM-CSF) long-term survivors showed significant protection (median 148 days) compared to control naïve mice (median 20 days) (Breslow-Gehan-Wilcoxon test P < 0.001), with 62% survival versus 0% of control mice (Table1; Fig. 1). We further tested the long-term survivors tumor protection by challenging with the parental colon carcinoma CT26. Figure 1 shows that only two out of the five mice treated with (IL-12)-IgG3 and IgG3-(GM-CSF) survived challenge with CT26. Table 1 shows that the therapy with combination of (IL-12)-IgG3 and IgG3-(GM-CSF) exhibited a median survival of 40 days, which was significantly higher than the 20 days survival of naïve mice controls challenged with CT26 (Breslow–Gehan–Wilcoxon test P < 0.002).

Fig. 1.

Survival plot of mice treated with GM-CSF and IL-12 anti-HER2/neu AbFPs and in combination against challenge with different tumor models. On the ordinates are indicated the treatment applied and on the abscissa the days after the first tumor challenge challenge. Cohorts of 8 female BALB/c mice were challenged i.p. at day 0 with D2F2/E2. The next 3 days animals were treated with one daily i.p. injection of PBS, IgG3, IgG3-(GM-CSF), (IL-12)-IgG3, and a mixture of (IL-12)-IgG3 and IgG3-(GM-CSF). Open circles refer to animals that succumbed to tumor challenge, and black circles to long-term survivors. Afterwards, long-term survivors and control-matched naïve mice were challenged s.c. with the tumor cell lines D2F2/E2, D2F2, CT26-HER2/neu and CT26. On the top of the figure are indicated the time points of treatment, challenges and the cell lines used. The result of the challenge with D2F2/E2 and with D2F2 cells has been previously reported [20]

Table 1.

Statistical analysis of long-term survival of mice treated with AbFPs and sequentially challenged with different syngeneic tumor models

Treatment	Re-challenge s.c. D2F2/E2		Challenge s.c. D2F2		Challenge s.c. CT26-HER2/neu		Challenge s.c. CT26
	Median (days)	P*	Median (days)	P	Median (days)	P	Median (days)	P
Control naïve mice	24	NA‡	16	NA	20	NA	12	NA
IgG3-(GM-CSF)	45	0.056	117	0.043	91	0.042	18	0.095
(IL-12)-IgG3	45	0.003	117	0.003	23	0.112	29	0.030
IgG3-(GM-CSF) + (IL-12)-IgG3	45	<0.001	117	<0.001	148	0.001	40	0.002

* The statistical significance in survival for each tumor challenge between the different treatments versus control naïve mice was determined by the non-parametric Breslow–Gehan–Wilcoxon test. In bold are indicated P values < 0.05

^‡NA Not applicable

Discussion

Our data suggest that the immune response induced by the treatment with the AbFPs generated systemic long-term memory, which was extended to antigens different from the originally targeted antigen, as evidenced by the protection against the parental cell line D2F2. Importantly, the protection was also significant against later challenge with a tumor from different origin expressing the same antigen, CT26-HER2/neu. In addition, although in smaller number, long-term survivors of the cross-challenge with CT26-HER2/neu showed significant protection against the parental cell line CT26 at a dose that killed all the control mice. It is possible that the protection against D2F2/E2 may have elicited a strong immune response against the antigen HER2/neu, which prevented the growth of the CT26-HER2/neu tumor. This protection may have primed the immune system against different antigens expressed on the surface of the CT26-HER2/neu cells, which protected against the challenge with the parental CT26. Alternatively, the immunity generated against D2F2/E2 and the priming with D2F2 may have generated a long-term immune response against one or more previously unidentified shared antigens expressed in both parental cell lines, D2F2 and CT26. These circumstances may explain the protection against challenge with CT26-HER2/neu and CT26 independently of the generation of HER2/neu immunity. However, these mechanisms are not mutually exclusive, and we cannot discard the possibility that a combination of both may be responsible for the protection against CT26. It is important to note that the protection exhibited after the challenge with CT26-HER2/neu at day 300 was inferior to that of previous tumor challenges, and was even lower for the challenge with CT26 at day 448. This is somewhat expected since, in addition to the aging of the immune system, they are different tumor models that are sensitive to different mechanisms of tumor protection [25, 28]. In agreement with what we presented here, we ran a parallel experiment using CT26-HER2/neu for the initial i.p. challenge, followed by the same AbFPs treatments, s.c. challenge with CT26-HER2/neu, and later s.c. injection with CT26. In this case, although the number of long-term survivors was reduced, we observed significant protection with (IL-12)-IgG3 and the combination of GM-CSF and IL-12 AbFPs, but not with IgG3-(GM-CSF) alone [20]. When the long-term survivors were subsequently cross-challenged s.c. with D2F2/E2, we observed significant protection in all cohorts compared to the controls. Although after s.c. challenge with D2F2 only the long term survivors of the combination of (IL-12)-IgG3 and IgG3-(GM-CSF) exhibited a significant protection compared to controls, we still were able to achieve significant protection in an alternative metachronous tumor model (unpublished data). We do not discard the possibility that further injections of AbFPs after the different tumor challenges would have resulted in even better survival, even though the aged mice may still have difficulties eliciting an immune response. A murine model of tumor spontaneous regression/complete remission (SR/CR) has been described that exhibits resistance to multiple transplantable solid cancers such as sarcoma S180. However, aging precludes this capacity and is completely lost in animals older than 56 weeks [5]. The fact that we could observe significant immunity at a later age is intriguing and at the same time encouraging, given that the immunization was performed in 2-month-old animals and the last tumor challenge was done approximately 64 weeks after the first tumor injection. To the best of our knowledge, the present study is the first report of an experimental therapy that is able to generate protection against mammary cancer and subsequent cross-challenge with other tumor models such as the colon carcinoma CT26-HER2/neu and its parental CT26. Several immune mechanisms are expected to be involved in the protection elicited by our AbFPs [6, 8, 12]. D2F2/E2 cells exhibit sensitivity to cytotoxic T lymphocyte cells (CTLs) and IFN-γ based immunity [26], while CT26-HER2/neu tumors NK and T cells are both important for conferring tumor protection [30]. IL-12 and GM-CSF are known to be potent immunostimulators that can activate innate and acquired immunity. IL-12 promotes cell-mediated immunity [16] by enhancing the cytotoxicity of NK and CTLs [43], and inducing the secretion of IFN-γ by T and NK [15, 41]. GM-CSF promotes the augmentation of antigen presentation [13], and favor mechanisms of immunological memory [1]. Our previous studies using AbFPs as adjuvants in a vaccination setting against the extracellular domain of HER2/neu demonstrated that IgG3-(GM-CSF) alone or combined with (IL-12)-IgG3 could enhance Th1 and Th2 responses, while (IL-12)-IgG3 alone switched the immune response to Th1 [6, 7, 18]. Both CTL (Th1) and/or antibodies (Th2) may be able to confer long-term protection against tumors expressing HER2/neu [6]. The studies reported herein suggest the feasibility of targeted immunotherapy using antibodies fused to GM-CSF and IL-12 in a therapeutic setting of disseminated peritoneal malignancies expressing HER2/neu to generate broad long-term memory. However, additional studies are needed to decipher the specificity of the immune response and the basic immune mechanisms involved in the observed protection. Changing the variable regions to target a combination of other shared TAAs such as CEA, MAGE3, or mucin1 [12] may further increase the versatility of AbFPs as direct anti-tumor therapeutics and help broaden the range of effective treatments against peritoneal carcinomatosis. Our results suggest that this approach may be relevant to treat patients affected with primary or metastatic peritoneal tumors as those observed in ovarian, colon, stomach, bladder, lung, and breast cancers. In addition, it may also be effective in reducing the risk of relapse of the primary tumor whether or not they express the targeted antigen, as well as preventing the development of metachronous cancers that share at least one antigen with the primary tumor.