Immunogenicity of rat-neu+ mouse mammary tumours determines the T cell-dependent therapeutic efficacy of anti-neu monoclonal antibody treatment

The use of Trastuzumab (Herceptin), a monoclonal antibody (mAb) targeting HER2/neu, results in an increased median survival in Her2+ breast cancer patients. The tumour mutational burden and the presence of tumour infiltrating lymphocytes (TILs) clearly correlate with response to trastuzumab. Here, we investigated if the immunogenicity of the transplantable rat-neu+ tumour cell line (TUBO) derived from a BALB/c-NeuT primary tumour is associated with the response to anti-neu mAb therapy. We compared the TUBO tumour outgrowth and tumour infiltrating T cells in isogenic (BALB/c-NeuT) and non-isogenic (WT BALB/c) recipient mice. Furthermore, therapeutic efficacy of anti-neu mAb and the contribution of T cells were examined in both mouse strains. The outgrowth of untreated tumours was significantly better in BALB/c-NeuT than WT BALB/c mice. Moreover, tumour infiltrating T cells were more abundantly present in WT BALB/c than BALB/c-NeuT mice, showing that the TUBO tumour was more immunogenic in WT BALB/c mice. In TUBO tumour bearing WT BALB/c mice, anti-neu mAb therapy resulted in an increase of tumour infiltrating T cells and long-term survival. When T cells were depleted, this strong anti-tumour effect was reduced to an outgrowth delay. In contrast, in TUBO tumour bearing BALB/c-NeuT mice, treatment with anti-neu mAb resulted only in tumour outgrowth delay, both in the presence and absence of T cells. We concluded that in immunogenic tumours the response to anti-neu mAb therapy is enhanced by additional T cell involvement compared to the response to anti-neu mAb in non-immunogenic tumours.

tumour models and treatment. The TUBO tumour cell line 12  For the anti-neu mAb treatment efficacy studies using BALB/c-neuT female mice 12 , the mammary glands of BALB/c-neuT mice were inspected weekly for tumour appearance from the twelfth week of age. When a cumulative spontaneous tumour burden of 20 mm 2 was reached, the mice were treated with three intraperitoneal injections of 100 µg of anti-neu mAb (clone 7.16.4) at day 10, 15 and 20. The spontaneous tumours were measured 3 times per week with a calliper and their size was calculated by multiplying the tumour diameters in two dimensions. Cumulative tumour burden was calculated as the sum of all individual tumour sizes. Mice were sacrificed when established spontaneous tumours reached a cumulative tumour burden of 15 × 15 mm. flow cytometry. Flow cytometry analysis for tumour infiltrating CD45 + , CD4 + and CD8 + were performed as previously described 20 . In brief, tumours were harvested into 1 mL of non-supplemented IMDM media in 24-well plates and manually dissociated into small pieces with scalpels, incubated with 2.5 mg/mL Liberase TL (Roche) for 20 minutes at 37 °C and single-cell suspensions were made using 70-µm cell strainers (BD Biosciences). FcγRs were blocked with 10% normal mouse serum and anti-mouse CD16/CD32 antibody (2.4G2). Cell surface staining was performed using the following antibodies: anti-mouse CD8α (clone 53-6.7), CD4 (clone L3T4), CD3ε (clone 145-2c11)/ TCRβ chain (clone H57-597), CD45.2 (clone 104). Dead cells were excluded based on 7-AAD (Invitrogen). Analysis were performed using LSRII cytometer (BD) using FacsDIVA software (BD) and FlowJo Software (Tree Star).

Results
Anti-neu mAb therapy improved survival in transgenic BALB/c-NeuT spontaneous tumour model. We first analysed the anti-tumour effect of anti-neu mAb monotherapy in the BABL/c-NeuT female mice, which develop spontaneously invasive mammary tumours around 4 months of age 13,15,21 . As shown in Fig. 1A, administration of anti-neu mAb to spontaneous mammary tumours bearing BALB/c-NeuT mice resulted in significant delay in tumour outgrowth (p < 0.001), leading to significant improvement in survival when compared with untreated mice (Fig. 1B) (p < 0.01). Our results support previous studies showing the therapeutic effect of anti-neu antibodies in vivo 8,14 . Sub-optimal TUBO tumour take in WT BALB/c recipient mice. Similar to the majority of the oncogene-driven models of cancer in genetically engineered mice (GEMMs) 22,23 , spontaneous mammary tumours driven by activated rat-neu in the MMTV-NeuT mouse model may not harbour high neo-antigen load 24 . To evaluate whether the therapeutic efficacy of anti-neu mAb therapy would be stronger when the tumour is more immunogenic, we used a transplantable tumour model, the oncogenic rat-neu (NeuT) expressing tumour cell line (TUBO) established from a spontaneously developed mammary gland tumour of a female BALB/c-NeuT mouse 12 . As recipients of the tumour cells, we used either non-syngeneic WT BALB/c mice in which, due to the fact that the rat-neu molecule is a non-self-antigen, the tumour might be highly immunogenic, or syngeneic transgenic BALB/c-NeuT mice for which the rat-neu is a self-antigen and therefore the tumour might be poorly immunogenic 25 .
WT BALB/c and transgenic BALB/c-NeuT female mice were injected subcutaneously with TUBO tumour cells and tumour outgrowth was monitored. All WT BALB/c mice had palpable tumours at around day 8 upon transplantation, however, the growth of established tumours was inconsistent, and some underwent spontaneous regression without therapeutic intervention ( Fig. 2A). In contrast, spontaneous regression of established TUBO tumours did not occur in syngeneic BALB/c-NeuT mice (Fig. 2B). These observations suggest that TUBO cells are more immunogenic in WT BALB/c than in BALB/c-NeuT recipient mice. Flow cytometric analysis revealed that there was an increase in percentage of total CD3 + T cells and in CD8 + /CD4 + T cell ratio in TUBO tumours from WT BALB/c mice compared to TUBO tumours from BALB/c-NeuT mice (Fig. 2C,D), suggesting that T cells may contribute to the tumour regression in WT BALB/c mice.

Stronger therapeutic efficacy of Anti-neu mAb in TUBO bearing WT than BALB/c-NeuT mice.
Based on the difference in immunogenicity of TUBO in WT and BALB/c-NeuT mice, we anticipated different therapeutic responses to anti-neu mAb therapy in WT and BALB/c-NeuT mice. There was a significant increase in survival of untreated TUBO bearing WT mice compared to untreated TUBO bearing BALB/c-NeuT mice, in keeping with the earlier suggestion that an effective endogenous anti-tumour immune responses against the TUBO tumour is induced in WT but not BALB/c-NeuT mice (Fig. 3A). Short-term treatment with anti-neu mAb www.nature.com/scientificreports www.nature.com/scientificreports/ significantly increased the survival of both TUBO bearing WT and BALB/c-NeuT mice. However, the survival of WT BALB/c was significantly higher compared to the survival of BALB/c-NeuT mice. Taken together, our data strongly suggest that the response of established TUBO tumours to anti-neu mAb treatment correlates with the immunogenicity of tumours.

Anti-neu mAb monotherapy enhances the anti-tumour t cell responses against immunogenic tumours.
To investigate the role of T cells in the anti-tumour activity of anti-neu mAb in both WT and BALB/c-NeuT mice, this therapy was tested in mice depleted for CD4 + and CD8 + T cells. Efficiency of T cell depletion was validated during the experiment (Supplemental Fig. 1A,B). As expected, depletion of T cells clearly increased the growth of established TUBO tumours in WT BALB/c mice (Fig. 4A). Whereas treatment with anti-neu mAb resulted in strong tumour regression in WT BALB/c mice, tumours relapsed progressively in the www.nature.com/scientificreports www.nature.com/scientificreports/ absence of T cells, suggesting that T cells are essential for the maximal therapeutic efficacy of anti-neu mAb. In a follow-up experiment, we observed increased CD4 + and CD8 + T cells in the TUBO tumour of WT BALB/c mice treated with anti-neu mAb (Fig. 4B). This increase was not observed in BALB/c-NeuT mice (data not shown). Together, these data indicate that in immunogenic tumours, anti-neu mAb therapy contributes to strong tumour growth inhibition and eradication by improving the anti-tumour T cell responses.
Markedly delayed outgrowth of TUBO tumours was also observed in BALB/c-NeuT mice (Fig. 4C). However, the therapeutic efficacy of anti-neu mAb was not impaired in the absence of CD4 + and CD8 + T cells (Fig. 4C), indicating that the T cells in BALB/c-NeuT mice did not contribute to the therapy in contrast to the T cells in WT BALB/c mice. Together, our results suggest that T cells are responsible for the poor TUBO tumour outgrowth in WT BALB/c mice and that anti-neu mAb monotherapy can enhance the anti-tumour T cell responses to further delay the tumour outgrowth. In contrast, the much weaker anti-neu mAb-mediated anti-tumour effect in TUBO bearing BALB/c-NeuT mice, in which the tumour is less immunogenic, is T cells independent. When T cells were absent in WT BALB/c mice, the anti-tumour response of anti-neu mAb resembled strongly the response in BALB/c-NeuT mice, indicating a T cell-independent tumour-outgrowth inhibition by anti-neu mAb.

Discussion
Here we report that the immunogenicity of mouse mammary tumours can have a significant impact on the response to anti-neu mAb immunotherapy. When rat-neu is a non-self-antigen, making the rat-neu + tumour highly immunogenic as in WT BALB/c recipient mice, T cells contributed to a more pronounced anti-neu mAb-mediated tumour growth inhibition and eradication of the transplanted tumours than when rat-neu on the tumour is a self-antigen, as in BALB/c-NeuT recipient mice. In the latter mice, the weaker anti-tumour response induced by anti-neu mAb therapy was T cell independent. www.nature.com/scientificreports www.nature.com/scientificreports/ Our observation that rat-neu expressing TUBO tumours are highly immunogenic in WT BALB/c mice is in agreement with the finding of Reilly et al. 26 who demonstrated that the minimum tumour cell dose required for tumour outgrowth in 100% of the transplanted animals was 100-fold lower for the neu transgenic mice compared with non-transgenic mice. Our study also confirms previous studies showing that CD4 + and CD8 + T cells can contribute to anti-neu mAb monotherapy to suppress the growth of the TUBO tumour in WT BALB/c immunocompetent mice 8,14 . However, we did not observe a contribution of T cells to the therapeutic efficacy in anti-neu mAb treated TUBO bearing BALB/c-NeuT transgenic mice. An important question raising from these experiments is: why T cells contributed to the therapeutic efficacy of anti-neu mAb in WT BALB/c but not BALB/c-NeuT tumour bearing mice? A simple explanation is that rat-neu is a foreign antigen in normal BALB/c mice, and that the rejection of the TUBO tumour likely involved high avidity rat-neu specific T cells that are strongly, and rapidly activated and expanded after transplantation. This hypothesis is in agreement with our data showing the poor TUBO tumour growth and development in WT BALB/c mice. In contrast, BALB/c-NeuT mice which express rat-neu as self-antigen, harbour lower numbers of self-reactive low-avidity T cells which are more difficult to activate and do not expand effectively [27][28][29] . It has been demonstrated that spontaneous Her2-driven mammary tumorigenesis in BALB/c-NeuT animals is not suppressed by the adaptive immune system 15 , suggesting the lack of neo-antigens that can be recognised by T cells.
By depleting CD8 + T cells, Park et al. 8 and Stagg et al. 11 also noted the importance of CD8 + T cells in anti-neu mAb therapy in TUBO tumor bearing NeuT transgenic animals resulting in complete tumor regression in 20% of the mice. This is not in agreement with our results and it might be due to differences in experimental conditions. Park et al. 8 used F1 rat-neu transgenic mice (BALB/c x FVB/N MMTV-neu) as the recipient of TUBO cells. Although these mice are 'tolerized' for rat neu, most likely there are still genetic differences between the TUBO cell line and the F1 BALB/c x FVB/N MMTV-neu recipient mouse, which can explain the stronger, partially CD8 + T cell dependent, anti-TUBO immune response they report, compared to our studies in the fully syngenic BALB/c-NeuT mice. Besides, all other experiments of Park et al. 8 showing dependence on adaptive and innate immunity of anti-neu mAb therapy were performed in WT BALB/c or KO BALB/c mice not tolerized for the rat-neu antigen 11 11 used a much more intensive treatment regime (8 times 100 µg anti-neu Ab within a period of 14 days) than we did (3 times 100 µg anti-neu Ab within 10 days). It could be that sustained tumour cell killing by a longer period of treatment leads to induction of a strong inflammatory response activating low avidity anti-neu CD8 + T cells. This might also explain the syngergistic effect of anti-neu mAb and anti-PD-1 mAb 11 in this setting. Notably, CD8 + T cell dependency has also been reported for the effective combination therapy of anti-death receptor 5 (DR5) and anti-neu mAb in BALB/c-NeuT mice 30 . Since anti-DR-5 mAb triggers mainly apoptosis in tumour cells, this study raises the possibility that the sustained release of danger signals or damage-associated molecular patterns (DAMPs) from tumour cell death could be key at breaking immune tolerance and inducing meaningful CD8 + T cell antitumor immunity in neu-transgenic mice. This might as well explain why neu-specific whole tumour vaccination 26 but not DNA vaccination 12,31,32 therapies elicit antitumor CD8 + T cell responses in neu transgenic mice.
There is an increasing amount of literature supporting that anti-tumour immunity of anti-neu mAb can be enhanced with immunomodulatory agents such as CD73 33 , PolyI:C and CpG 34 in the highly immunogenic TUBO transplanted WT BALB/c mice. Our study suggests that the potential and underlying mechanisms of action of these combination therapies have to be elucidated in low immunogenic setting using transgenic BALB/c-NeuT recipient mice in order to complement the results obtained in WT BALB/c mice. Our result in BALB/c-NeuT mice suggest that direct tumour growth inhibition by blockade of rat-neu signalling is one of the major mechanisms of anti-neu mAb in transgenic NeuT mice. Improved anti-tumour effects could be achieved by the continuation of treatment of these mice with anti-neu mAb, similar to the therapy in breast cancer patients who often receive anti-Her2 mAb over long periods of time 35 or combining it with a tumour targeting agent such as anti-DR5 mAb that induces tumour cell death directly 11 . We do not rule out, however, that Fc-mediated effector functions might be involved also in anti-neu mAb therapy. Further investigations are needed to reveal the contribution of FcγR and FcγR expressing immune cells to the anti-tumour efficacy of anti-neu mAb and whether FcγR expression can be modulated in the tumour microenvironment to augment mAb-mediated 36 effector functions for both immunogenic and non-immunogenic tumours. Our study supports the notion that intense immune phenotyping of the various syngeneic tumour models is critical for both rational model selection and data interpretation for clinical translation [37][38][39] . For example, the high immunogenicity of a fully non-self-antigen such as rat-neu in the TUBO tumour bearing WT BALB/c model, will most likely result in overestimation of the potency of combining anti-neu mAb therapy with immunotherapy. On the other hand, using a completely syngeneic model such as MMTV BALB/c-NeuT mice that develop spontaneous mammary tumours or TUBO transplanted in fully syngeneic BALB/c-NeuT mice might result in underestimation of the potential anti-tumour immune response in patients, in which mutations may have led to neo-epitope formation and a certain level of immunogenicity.
Antibody targeting Her2 (trastuzumab, pertuzumab) is among the first of the approved antibodies directed against Her2 + cancer cells. Although these antibodies have become standard of care and improved overall survival of patients with Her2 + breast cancer, heterogeneity exists within Her2-positive tumours, and the overall response rate to anti-Her2 mAb-based therapies remains modest, approximately 26% when used as a single therapy and 40-60% when used in combination with chemotherapy 2,40,41 . To date, Her2 expression level remains the only suitable marker for patient selection for anti-Her2 based regimen 42,43 . Nonetheless, the relationship between