Background
Metastatic breast cancer (MBC), including locally advanced breast cancer (LABC) and distant relapse (DR) are, nowadays, major challenges to oncologists, and to researchers focused on the development of new treatments against these deadly disease situations. Despite advances in early diagnosis and multidisciplinary therapeutic approaches, MBC remains incurable and current main therapies goals range from symptoms palliation to extending survival [
1]. In severe cases of LABC, where an underlying invasive disease is often present at the time of diagnosis, neoadjuvant therapies (i.e., administration of therapeutic agents before a main treatment such as mastectomy and/or radiotherapy) are first-line choices [
2‐
4]. Presently, such neoadjuvant therapies relay mostly on cytotoxic chemotherapies which, unfortunately, have limited efficacy, as well as multiple toxic effects. It is, therefore, extremely necessary the development of alternative therapeutic strategies that are more effective against MBC and that can be applicable in neoadjuvancy to treat LABC. In addition, it is also equally important that the new therapeutic approaches are evaluated in clinically relevant animal models of breast cancer.
The murine 4T1 orthotopic model mimics aggressive types of breast cancer since it is rapidly progressive, highly angiogenic and angioinvasive and metastasizes spontaneously from little primary tumors to draining lymph nodes and distant organs, following a metastatic pattern that closely resembles the human counterpart [
5‐
7]. In addition, the carcinogenesis dynamics of 4T1 tumors established in immunocompetent BALB/c mice is comparable to human stage IV type of breast cancer [
8]. Moreover, and analogously to human breast cancer, 4T1-derived tumors are poorly immunogenic [
6]. In this regard, several mechanistic studies have described the cellular, molecular and soluble tumor-associated and immune-related factors that participate in inhibiting the host immunosurveillance effectors [
9‐
11], making the 4T1 model particularly challenging for evaluating novel strategies aimed at inducing an efficient antitumor immune response. For all these reasons, the investigation and discovery of effective neoadjuvant immunotherapies against the metastatic disease in the murine 4T1 breast cancer model could be of great clinical value.
Gene therapy is a relatively new paradigm in medicine with enormous therapeutic potential. Actually, almost 2000 gene therapy clinical trials (most of them for cancer applications) were registered worldwide, having 96 trials already reached phases II/III to IV, indicating the progress being made with respect to bringing gene therapy closer to the clinical setting [
12]. A number of different vectors and delivery systems have been applied; however, viral vectors remain by far the most versatile, popular and effective approach, been used in approximately two-thirds of such clinical trials [
13]. We have recently shown that a gene therapy strategy based on the short-term intratumoral (i.t) expression of the potent pro-inflamatory cytokine interleukin-12 (IL-12) expressed from a cytopathic Semliki Forest virus vector (SFV-IL-12) inhibits tumor growth and extends survival in a transgenic mice model of hepatocellular carcinoma (HCC) [
14]. Moreover, this vector was also efficient in reducing tumor volume and inducing T cell-mediated responses against HCC spontaneously developed in woodchucks chronically infected with an hepadnavirus very similar to hepatitis B virus, a situation that closely resembles HCC disease in humans [
15]. In addition, SFV-IL-12 was proven to have a stronger antitumor potential compared to other viral and non-viral-based IL-12 expression systems
in vivo due to the strong SFV-mediated induction of apoptosis and activation of type-I interferon responses specifically in the tumor [
14,
16,
17]. When searching whether the therapeutic spectrum of this vector was expanded to breast cancer, we found only one previous report where a similar SFV-IL-12 construct was tested in the 4T1 model. There, a significant reduction in primary tumor size as well as in the percentage and number of lung metastases was observed after several i.t. administrations of high amounts of vector [
18]. However, no information was provided about long-term survival (i.e, at least 6 months after treatment) and/or anti-metastatic effect of the therapy in tumor-excised mice, which would be a more clinically relevant condition, due to the fact that, in general, patients have their primary tumor surgically removed.
On the other hand, bacteria-based therapies are a modality of increasing interest in anticancer immunotherapeutic research due to the capability of many auxotrophic mutants to restrict their growth to the rich nutrient milieu tumor interior, inducing cell destruction and liberation of tumor antigens, as well as
in situ inflammation, triggering a versatile immune response upon bacterial infection [
19‐
21]. It has also been shown that facultative anaerobic bacteria, like attenuated
Salmonella enterica serovar Typhimurium (
S. Typhimurium) and genetically modified
Escherichia coli strains, elicit antitumor potential against breast cancer models, especially if combined with plasmids that express cytotoxic inducing factors or immunostimulatory cytokines [
22‐
26]. However, similarly as for IL-12-based gene therapies, these studies were mostly focused on evaluating the antitumor effect of bacteria against primary tumors; and so far there are no data reporting complete remission of 4T1 primary tumors or total inhibition of lung metastasis in operated mice.
Therefore, and in order to go a step forward towards a possible clinical application of a gene- and bacteria-based neoadjuvant therapy for MBC, we have evaluated the anti-metastatic potential of the i.t. administration of SFV-IL-12 and/or
aroC
−
attenuated
S. Typhimurium strain LVR01 [
27] into 4T1 primary tumors followed by their surgical resection. Our data showed a clear synergic antitumor effect of the combined therapy compared to SFV-IL-12 and LVR01 monotherapies, achieving metastasis-free and long-term survival in 90 % of treated animals. Moreover, we observed that the efficacy of the combined therapy radically depends on the order in which both factors were administered; suggesting the existence of specific mechanisms underlying the observed synergic therapeutic effect.
Discussion
Neoadjuvant therapies aim to reduce the size or extent of primary tumors before radical intervention, as well as to eliminate early disseminated micro-metastasis in locally advanced malignancies, such as LABC [
2]. In the present study, we have evaluated the efficacy of a combined neoadjuvant gene therapy approach against the highly metastatic 4T1 mouse model, which we have correlated here with severe LABC type of cancers (Fig.
1). Our neoadjuvant approach is based on the use of IL-12, a potent immunostimulatory cytokine with strong antiangiogenic activity [
32,
35,
43] in combination with LVR01, an attenuated auxotrophic mutant of
S. Typhimurium [
27,
42] as co-adjuvant factor. To allow IL-12 expression into the tumor mass we employed a SFV-based vector encoding IL-12 genes (SFV-IL-12), which was previously shown to express IL-12 (locally and systemically) upon i.t. administration [
14,
28]. In this construct, each IL-12 subunit (p35 and p40) is driven by an independent SFV subgenomic promoter fused to the SFV capsid translation enhancer achieving higher expression levels and stronger antitumor activity compared to adenoviral vectors engineered to express IL-12 [
16,
28]. SFV-based vectors have also the capability of inducing apoptosis of infected cancer cells, allowing the release of tumor-associated antigens that could be taken up and cross-presented by antigen-presenting cells [
14,
36,
44,
45]. In addition, SFV-IL-12 achieves transgene expression after multiple i.t. administrations and triggers efficient antitumor immune responses in different cancer models [
14‐
16,
18,
28,
46].
Salmonella LVR01 was also demonstrated to induce death of infected tumor cells and to trigger innate and adaptive immune responses after multiple i.t. administrations [
42]. These responses are associated to Toll-like receptors recognition of pathogen associated molecular patterns (PAMPs) in the bacteria, such as lipopolysaccharide (LPS), teichoic acid, peptidoglycan, and bacterial CpG DNA [
41,
47]. Although the LVR01 bacterial strain has been extensively tested as carrier for heterologous antigens for human and veterinary vaccination applications [
48‐
50], its use as anticancer agent is relatively recent [
42,
51] and therefore is still an open area for investigation. Still, the use of
Salmonella for anticancer therapies, either as monotherapy or combined, has received extensive consideration in the last decade, showing different degree of effectivity in a number of animal models of major human cancer types including breast cancer [
52‐
58]. Particularly, a
Salmonella A1-R-based monotherapy was recently reported with promising results in a 4T1 brain metastatic model [
59], although these experiments were conducted in nude mice and long-term protection was not evaluated. In any case, to date only a few approaches using attenuated strains of
Salmonella have moved into initial clinical trials and none of them have progressed into large phase II clinical trials, paving the way for the search of alternative combined therapies.
Our data showed a clear reduction of tumor blood supply, delay of tumor growth, lower number of lung metastasis and a better survival in SFV-IL-12 treated mice compared to PBS and SFV-LacZ controls. Even though, only 20 % of animals survived long-term after receiving one or two i.t. doses of SFV-IL-12 in the neoadjuvant setting, indicating that an increased dose regimen of SFV-IL-12 monotherapy did not seem to augment its antitumor effectiveness (Figs.
3 and
6, respectively). In this regard, a previous report showed inhibition of primary tumor growth after 6 consecutive i.t. injections of a similar SFV-IL-12 vector in 4T1 tumor-bearing mice, inferring an effective antitumor response using SFV-IL-12 alone [
18]. However, these studies were followed up only until day 22 (when mice were sacrificed in order to measure the presence of lung metastasis), a time period where we also observed tumor growth inhibition. Nevertheless, in our long-term studies, all tumors recovered their growth dynamics after day 25-30 and animals died between days 50 and 65 due to metastatic disease, indicating that IL-12 alone was efficient in delaying tumor growth and death, but insufficient to cure these animals. In accordance with our data, the authors also observed metastasis reduction and a potent antiangiogenic effect of SFV-IL-12 in 4T1 injected tumors, a fact that may delay nutrient access to the tumor mass and metastasis spread [
18]. The antiangiogenic contribution of IL-12 together with its immunomediated antitumor properties were consistently studied and reported previously [
32,
34,
35,
43,
60], as well as its limitations as a single antitumor agent [
36]. For that reason, in this work we focused on enhancing the SFV-IL-12–based therapy by combining it with an appropriate “therapeutic companion” in order to achieve long-term cure, which was the main goal of the present study.
A number of different immunomodulatory factors and tumor-specific antigens have shown synergistic effects with IL-12-based therapies in different cancer models [
36,
60‐
62]. However, most of these studies were aimed at evaluating antitumor efficacy against primary tumors, without assessing their effects in neoadjuvancy. In this regard, a relevant study performed in a lung alveolar metastatic model, showed that i.t. administration of IL-12-encapsulated biodegradable microspheres combined with granulocyte-macrophage colony-stimulating factor (GM-CSF), followed by surgical removal of tumors, achieved significant post-operatory long-term survival and reduction of metastasis in animals receiving the combined treatment, compared to each monotherapy [
37]. These data indicate that
in situ IL-12-based neoadjuvant treatments could help to prevent post-operatory dissemination of the metastatic disease, and, together with other studies, confirm that antitumor IL-12-based therapies could benefit from combinations with other cytokines, antibodies or chemotherapy [
36‐
40]. We hypothesized that live attenuated
Salmonella could be a proper IL-12 co-adjuvant due to the bacterial intrinsic immune-stimulatory and antitumor properties, its easy and cheap high-scale production, and its enhanced therapeutic activity observed when transformed with plasmids expressing immune-stimulatory cytokines ([
41] and references therein). Recently, Grille et al, showed an improved survival outcome in mice receiving multiple i.t. doses of
Salmonella LVR01 in the A20 lymphoma model [
42]. However, in the 4T1 breast cancer model, we did not see any therapeutic response in tumor-bearing animals treated solely with LVR01, confirming the previously inferred resistance of 4T1 tumors to immunotherapies [
13], although a significant increase in the overall survival was achieved in neoadjuvancy. Anyhow, we detected the infiltration of granulocytes to the tumor site, a relevant observation, since it was suggested that this cell population (specially neutrophyls) is involved in
Salmonella-mediated therapeutic effect [
42,
63]. In addition, LVR01-treated tumors showed microscopic distortions, probably due to the bacterial replication and destruction of infected cells. It is important to remark that all 4T1 tumors treated with LVR01 alone (either with 1 or 2 doses) experienced a significant increase in size, which is consistent with a strong inflammatory reaction and observed vascularization. Consequently, these tumors were more difficult to resect than their SFV-IL-12 counterparts, because they got strongly attached to the peritoneum and skin borders. This fact made the evaluation of a multiple dose regimen protocol surgically impracticable.
We discovered an impressive synergistic effect when a single i.t. dose of SFV-IL-12 (2×10
8 vp given at day 10 after 4T1 cells implantation) was followed by a single i.t. dose of LVR01 (2×10
7 cfu given at day 13) prior to surgical removal of the treated tumor (day 16). Here, 90-100 % treated animals survived for a time period that could be considered life-long in mice (total 350 days; Figs.
6 and
8). However, to our surprise, if LVR01 was administered before SFV-IL-12, the combined antitumor synergy was completely lost in our neoadjuvant experimental setting. The reason by which this synergy operates or can be abolished may involve multiple mechanisms, since the type of vector, cytokine and bacteria employed in this study are able to trigger a diverse spectrum of effects and pathways [
17,
18,
28,
42,
45] and therefore, their relative contribution will require further investigations. So far, and based on our data, we could mainly speculate that i.t administration of SFV-IL-12 allow sufficient IL-12 expression to activate antiangiogenic responses contributing to inhibit primary tumor growth, and in this favorable context (which might converge with an adequate amount and type of SFV-IL-12-induced immune mediators [
17,
34]), the addition of a potent adjuvant, like LVR01, can probably promote a systemic response able to eliminate disseminated micro-metastases in the organism, something that is not achieved by SFV-IL-12 or LVR01 alone. In accordance with the higher expression levels of IL-12 observed in the SFV-IL12 + LVR01
vs LVR01 + SFV-IL12 treated animals, also a significantly higher production of IFN-γ was achieved in the former group at early time points, promoting the idea that the better therapeutic outcome in this group could be associated to a T helper 1 primed response timely induced for which an early source of IFN-γ is strictly required [
64]. Indeed, previous studies demonstrated that CD8
+ cytotoxic T cells are major players in antitumor responses triggered by SFV-IL-12 either alone or combined with a proper agonist [
28,
36,
46]. Consistently, we obtained initial data showing a higher amount of CD8
+ T cells in draining lymph nodes isolated from SFV-IL-12 + LVR01 compared to LVR01 + SFV-IL-12 treated animals. The fact that synergy is not observed when LVR01 is given before SFV-IL-12 suggests that the first agent is not as efficient as SFV-IL-12 to prime an adequate microenvironment and efficient immune responses, or alternatively, that LVR01 infection could affect posterior SFV-IL-12 tumor cells transduction. In any case, we found that tumor pre-infection with LVR01 resulted in a significant decrease of SFV-IL-12-mediated expression of IL-12, which would explain the lack of IL-12-mediated antiangiogenic effect in treated tumors, as well as the reduced IFN-γ levels and CD8
+ T cells in these animals. Overall, our data suggest the importance of reducing angiogenesis and inducing an efficient immune response at early stages of the disease in order to achieve a potent therapeutic effect.
The lack of protection against a secondary challenge may indicate that our therapeutic procedure does not stimulate an efficient long-lasting antitumor immunity. Nevertheless, we found that the application of same treatment (combined adjuvant therapy followed by surgical removal of primary tumors) is equally effective to cure animals when used into re-challenged mice, a finding that we believe could be clinically relevant to treat patients who suffer from relapse. Thus, this combined therapy may be worth of considering for clinical trials.
In summary, we have described a promising neoadjuvant therapeutic strategy to treat a highly aggressive type of breast cancer. To our knowledge, this is the first experimental protocol reported to date that combine: gene therapy + bacteria-based therapy + surgical removal of primary tumors, for the prevention and treatment of the metastatic disease, and that achieved long-term curative results in the clinically relevant 4T1 model. In addition, we demonstrated that all re-challenged mice were able to respond to a second neoadjuvant SFV-IL-12 + LVR01 treatment, rising further optimism about this experimental approach, since in the clinic, relapsed mammary tumors are often more difficult to be therapeutically solved.
This strategy would have at least three main theoretical advantages compared to other immuno- or chemical- based therapies. First, the
in situ administration of each therapeutic agent would avoid, in theory, systemic toxicity. Second, an expected locally induced immunity would specifically recognize and destroy cancer cells, thus diminishing unspecific effects. Third, by combining SFV-IL-12 with
Salmonella and surgery we could reduce the doses of the viral vector, more complex to produce than bacteria, lowering the cost of therapy. In addition, we believe that such a strategy could be clinically feasible because: (i) administration of alphavirus-based vectors similar to SFV, has been already proven to be safe in phase I/II clinical trials [
65] (ii) live attenuated
Salmonella strains have also been assayed in clinical trials showing a good safety profile [
66] with a bacteria-based therapy successfully being used for the treatment of patients with bladder cancer [
67,
68], (iii) neoadjuvant i.t. administration of a gene therapy vector combined with chemotherapy has been already clinically tested in LABC patients, showing to be feasible and safe [
69]. Finally, combinatorial treatments are perceived as a major pathway for progress in cancer therapy with a number of pre-clinical and clinical events revealing synergistic antitumor activities. It is therefore encouraging to anticipate that combined biological approaches could become interesting options for the treatment of malignancies that lack cure in the present.
Authors’ contribution
MGK conceived and supervised the project, carried out all surgeries and in vivo experiments, analyzed the data, prepared the final illustrations and wrote the manuscript. MMa performed the ELISAs, clonogenic and immunohistochemical assays and helped with statistical and in vivo studies. EC produced and titrated the SFV-IL-12 and SFV-LacZ vectors used in this work. MMo provided orientation for MMa to perform the initial immunological analysis. CS and JACh supplied the viral vectors and attenuated bacteria, respectively, and critically revised the manuscript. All authors contributed with enriching discussions. All authors read and approved the final manuscript.