Neoadjuvant administration of Semliki Forest virus expressing interleukin-12 combined with attenuated Salmonella eradicates breast cancer metastasis and achieves long-term survival in immunocompetent mice

Breast cancer aggressiveness and poor prognosis is commonly associated with metastasis occurrence to distant organs, such as lungs, liver and bones. In this regard, it is well known that early resection of primary tumors contributes to control metastasis; however clinical outcome depends on many factors, including growth dynamics and metastatic characteristics of these tumors. In order to correlate the 4T1 breast cancer model with clinical cases of LABC in respect to its metastatic progression, we evaluated the therapeutic impact of primary tumor resection at different time points. Small tumors (4-5 mm diameter) could be homogeneously detected 10 days after orthotopic implantation of 7×10⁴ 4T1 cells and by day 16 tumors reached 6–7 mm diameter, still a feasible surgically size. Thus, we defined days 10 and 16 post-4T1 implantation as early and middle-early time points for tumor resection, respectively. Control tumor-bearing mice have a life span of 44 ± 5 days after 4T1 cells implantation (Figs. 1 and 8b). Animals whose tumors were surgically resected at day 10 showed an improved overall survival (with 40 % mice surviving more than 6 months), while 100 % animals operated at day 16 died at the same rate as non-operated mice (Fig. 1a). In agreement with these data, lung micro-metastatic cells were present at day 10 in 40 % of the animals, while the rest were metastasis-free; however, 6 days later (day 16), 100 % animals presented a variable number (10-1300) of lung micro-metastasis (Fig. 1b). Middle-early excised tumor mice and non-operated controls showed comparable life span and a similar amount of established lung metastasis (in the order of 10⁵ 4T1 tumor cells/lungs) 35 days after tumor implantation (Fig. 1b) suggesting that both facts are associated. The autopsy of moribund mice revealed in all cases the presence of numerous macroscopic metastatic foci and a massive distortion of both lungs (Fig. 1c), accompanied by severe breathing difficulties at the time of death. These data confirmed the determinant role of the metastatic disease in the survival outcome of the 4T1 model of breast cancer [6]. In addition, non-survivor animals occasionally presented visible metastatic tumor nodules in liver, lymph nodes, spleen and/or heart (not shown). It is important to note, that just a small number of 4T1 metastatic cells (around 10–20, Fig. 1b) are sufficient to progress to a severe lung metastatic disease in the absence of primary tumors, indicating that these tumor cells may be particularly resistant or indifferent to the host immune system in order to persist. Due to the early metastatic dissemination and their growing characteristics, we conclude that the 4T1-tumor model could reflect most aggressive clinical cases of LABC. Moreover, middle-early tumor-excised mice would represent an excellent model of minimal residual disease (MRD) for therapeutic intervention against metastatic disease, since 100 % of animals develop lung metastasis and do not survive in the long-term. Our data also point out to a temporal window between days 10 and 16 (initiation - establishment of the metastatic process) for evaluation of novel neoadjuvant strategies (prior to surgical removal of primary tumors) aimed at preventing and treating lethal MRD.

We and other groups have shown that the amount of IL-12 expressed at the tumor site using different viral vectors is critical for tumor regression in a number of cancer models [14‐16, 18, 28, 32]. In order to evaluate the therapeutic potential of SFV-IL-12 against breast cancer, we took into consideration all these previous experiences, and injected a relatively high dose of SFV-IL-12 (2×10⁸ vp) at two early stages of the disease: the first dose was given at day 10 and the second dose was given at day 13 after 4T1 implantation (Fig. 2). Tumor growth was periodically measured and animals were monitored for survival outcome. Similarly, control groups received two i.t. doses of an SFV vector expressing β-galactosidase (SFV-LacZ) or PBS. As shown in Fig. 2a, treatment with SFV-IL-12 caused a significant inhibition of tumor growth prolonged until days 20-25. At this stage, control animals showed large ulcerated and irrigated tumors, whereas SFV-IL-12 treated tumors were small and whitish, due to a visible reduction of vasculature (Fig. 2c). These data are consistent with the described antiangiogenic effect of IL-12 in 4T1 tumors and in other IL-12-treated tumor models [18, 32‐35] indicating the in vivo activity of this cytokine. The survival of SFV-IL-12 injected animals was significantly extended compared to the SFV-LacZ and PBS control groups, although no animals showed long-term survival (Fig. 2c). Moreover, a significant lower number of lung metastasis was observed at day 35 in SFV-IL-12 treated mice (Fig. 2d), suggesting that their extended survival rate may be a consequence of the reduced metastatic burden at a time point when untreated animals start to die. Together, these results show that SFV-IL-12 induce a marked local antiangiogenic effect in implanted 4T1 tumors, which may contribute to inhibit primary tumor growth for a limited period of time, as well as to reduce dissemination of metastasis in the organism, prolonging consequently animal survival. However, SFV-IL-12 monotherapy, at least at the employed dose regimen, was insufficient to achieve complete tumor remission and to prevent the spread of 4T1 lung metastasis.

In order to mimic frequent management of patients with LABC, we have implemented a therapeutic protocol that included surgical removal of primary tumors after a preliminary treatment. Unlike conventional neadjuvant systemic approaches, here we performed local administration of the therapeutic agent (Fig. 3a). Two doses of SFV-IL-12 were injected inside the tumor mass at days 10 and 13 after 4T1 cells inoculation. Tumors were surgically removed 3 days later (day 16) and survival outcome was compared to SFV-LacZ or PBS treated mice. As shown in Fig. 3b, the overall survival of SFV-IL-12 treated animals was significantly higher than that of controls. In addition, 20 % of long-term survivors were observed in the IL-12-based neoadjuvant post-operatory setting. All of the non-survivor animals experienced breathing difficulties before death and showed numerous macroscopic lung metastases as was previously observed in untreated mice (see Fig. 1c). In contrast, mice that survived at the end of this study did not present any visible metastasis. However, the fact that long-term survival was only achieved in a small percentage of treated mice indicated that our strategy to treat this type of cancer needed to be improved. Based on several studies that report the benefit obtained when combining IL-12-based gene therapy with complementary strategies or drugs [36‐40], we next decided to combine SFV-IL-12 with another biological agent that could collaborate in inducing a more efficient response against the metastatic disease.

We hypothesized that live attenuated Salmonella could be a good complement for IL-12-based antitumor therapy due to the intrinsic immune-stimulatory properties and the better antitumor responses observed when administering bacteria carrying plasmids that express pro-inflammatory cytokines (reviewed in [41]). Additionally, a recent related study of our group has shown that i.t. administrations of Salmonella LVR01 generate a considerable dose-dependent antitumor effect in a mouse model of B-cell lymphoma [42]. Despite this result, when two doses of LVR01 were inoculated into primary 4T1 tumors, no survival improvement was seen in tumor-bearing animals (Fig. 4). Surgical removal of the tumor after LVR01 inoculations led to a significant death delay, although no animals survived for the long-term (Fig. 4). In this case, all mice died with large macroscopic lung and disseminated metastasis, indicating that LVR01 alone was insufficient to control the 4T1 metastatic process. Nevertheless, the inoculated bacteria were able to invade tumor cells and induce local microscopic tissue distortion and granulocyte infiltration (Fig. 5), persisting inside the tumor mass for several days (see Additional file 1).

On the other hand, when SFV-IL-12 (day 10) was combined with LVR01 administration (day 13) followed by surgical removal of the treated tumor (day 16, Fig. 6a), a clear synergy in the antitumor action was observed, evidenced by 90 % long-term survival without metastasis. However, if treated tumors were not resected, all mice died, indicating that this combined therapy only works in neoadjuvancy (Fig. 6b). Interestingly, we found that the order of the injected factors dramatically affected the antitumor response, since the administration of LVR01 prior to SFV-IL-12 did not show any synergy and all animals died at similar rate than control groups (Fig. 6a and b). It is worth to note, that at the time of surgery, tumors receiving only LVR01 or the ineffective LVR01 + SFV-IL-12 combination showed a bigger size compared to SFV-IL-12 alone or SFV-IL-12 + LVR01 treated animals (Fig. 6c). In addition, inhibition of vasculature was observed with effective SFV-IL-12 + LVR01, but not with the LVR01 + SFV-IL-12 combination (Fig. 6d). Moreover, we found that IL-12 expression was significantly reduced in mice receiving LVR01 prior to SFV-IL-12 (day 14, Fig. 7a), while a higher amount of IFN-γ (a main mediator of IL-12 activity) was induced earlier in the effective combination (day 11, Fig. 7b), indicating that the right levels and dynamics of both cytokines expression may be associated with the curative effect of the combined therapy. In this regard, we also observed a significant increase of CD8⁺ and CD4⁺ T cells in draining lymph nodes isolated from the effective LVR01 + SFV-IL-12 combined group (see Additional file 2). Overall, these data demonstrate that the synergic neoadjuvant action of SFV-IL-12 and Salmonella LVR01 prevent and eradicate breast cancer metastasis that could be disseminated in the organism before surgical removal of the tumors. Our observations point out to the importance of reducing angiogenesis and inducing an efficient immune response at early stages of the disease.

As shown above, surgery of primary tumors performed 16 days after 4T1 cells inoculation did not prevent lung metastasis formation and animal death, whereas an efficient long-term antitumor effect was observed after receiving the right neoadjuvant combined therapy. This data suggests that the mice cured from metastatic disease may have developed a systemic antitumor immunity, which, in theory, could protect them from a second 4T1 cells challenge. To test this hypothesis, we inoculated 10⁴ 4T1 cells in the left abdominal mammary gland of cured animals 6 months after the first 4T1 tumor implantation (Fig. 8a). Contrary to our expectations, all of these animals developed mammary tumors and died, although with a significant slower rate than control naïve mice that received an equal dose of 4T1 cells (Fig. 8b), indicating that specific antitumor immunity might be elicited. However, removal of the primary tumor at day 16 post-inoculation (performed in order to reduce tumor cells burden in the organism) did not improve survival, suggesting that this elicited immune response was too weak to protect these mice from a secondary challenge. On the other hand, treatment with SFV-IL-12 and Salmonella LVR01, when followed by excision of the tumor, again showed a very potent antitumoral effect, achieving 100 % of long-term survival (Fig. 8b). These data indicate that although the combined therapy does not induce an efficient long-lasing protection against 4T1 tumors, it allows the use of the same treatment to cure a secondary mammary tumor in the same animal, a situation that could be clinically relevant to treat patients who suffer from relapse.