Introduction

Cancer immunotherapy is based on the recognition of tumor-associated mutant or non-mutant antigens (peptides) on cancer cells by the patient’s T cells. Cancer vaccines have successfully increased the number and activity of T cells that recognize tumor-associated antigens (TAAs) in many cases, but robust clinical responses remain anecdotal [1]. In contrast, US Food and Drug Administration-approved immunotherapies such as high-dose interleukin (IL)-2, interferon (IFN)-α, anti-CTLA-4 (cytotoxic T lymphocyte antigen-4) monoclonal antibodies and anti-PD-L1 (programmed death-ligand) therapies all appear to activate and/or expand tumor-specific T cells of largely unknown antigen specificity. This process allows the immune system to “decide” which antigen to target and may prove particularly fruitful as an approach to cancer immunotherapy. However, this approach requires the delivery of appropriate signals that convert immune responses from being ineffective to capable of rejecting established tumors. Using the tumor itself as a vaccine by introducing an immunomodulator or adjuvant can activate innate immunity and lead toward immunizing patients against shared (usually “self”) or patient-specific, mutated antigens. Besides generating tumor-specific immunity through T cell priming, intratumoral (i.t.) immunotherapy can also target the tumor by affecting innate immune cells like myeloid-derived suppressor cells (MDSC) and M2 macrophages, for example by polarizing them into tumor-suppressive phenotypes such as M1 macrophages [2, 3]. This modification of the tumor microenvironment and generation of systemic tumor-specific immunity by i.t. immunotherapy is not only useful to destroy the injected tumor but also suppresses distant metastasis. Intratumoral immunotherapy has been attempted for many years to treat solid tumors, especially melanoma due to its easily accessible cutaneous nature and high immunogenicity compared to most other malignancies. The majority of data on i.t. immunotherapy therefore involve melanoma. Here, we review various i.t. immunotherapy strategies to treat metastatic melanoma.

I.t. immunotherapy with bacteria

William B. Coley observed that patients with bacterial infections sometimes underwent spontaneous remissions of their cancers. Based on this observation, in 1891, he began treating bone sarcoma by injecting virus or bacteria into the tumors, resulting in some remarkable cures [4]. Since then, many bacteria and bacterial products have been used to activate the immune system to kill cancer cells. Activation of the toll-like receptor (TLR) pathway of immune cells by bacterial cell walls and their nucleic acids makes some bacteria useful in cancer treatment. I.t. bacillus Calmette–Guerin (BCG) alone and in combination with chemotherapy and other immunotherapies has been used to treat melanoma [5–8]. This bacterial therapy provides immunostimulatory DNA and activates the TLR9-MyD88 pathway [9], which leads to activation of macrophages and dendritic cells (DCs) and the production of various cytokines [9, 10]. These activated innate immune cells can also prime tumor-specific T cells against various TAAs and generate anti-tumor immunity. Recombinant BCG (rBCG 3A) has anti-tumor properties equivalent to those of wild-type BCG, but is a safer alternative for patients, because it does not contain infectious bacteria [11]. Udagawa et al. [12] reported that intratumoral administration of DCs stimulated with the BCG cell wall skeleton (BCG-CWS) suppressed the growth of not only the cryoablative tumors into which they were injected but also tumors at distant sites, which suggests that i.t. BCG can be an effective treatment for metastasis.

I.t. injection of the Salmonella enterica serovar typhimurium vaccine has also been shown to produce anti-tumor activity through transformation of MDSCs into tumor necrosis factor (TNF)-alpha-secreting neutrophils, reducing the generation of regulatory T (Treg) cells and increasing cytotoxic T cell infiltration [13]. This therapy is an example of i.t. therapy that not only generates tumor-specific T cells but also changes the tumor microenvironment from tumor promoting to tumor suppressive.

I.t. administration of attenuated Toxoplasma gondii parasites treated B16.F10 melanoma because it stimulated systemic anti-tumor immunity and tumor-specific memory CD8+ T cells [14, 15]. The results of these studies suggest that activation of tumor-associated innate immune cells by attenuated bacteria or parasites is a promising, safe and inexpensive approach to treat metastatic melanoma. However, efficacy is largely anecdotal at this point in time.

I.t. immunotherapy with oncolytic viruses

Tumor cell killing by genetically engineered and naturally occurring oncolytic viruses is another approach for cancer treatment that has been studied for decades [16]. These viruses selectively replicate in tumor cells because viral receptors are overexpressed on the surface of those cells or because specific anti-viral pathways, such as the type I IFN pathway, are disrupted in tumor cells or tumor stromal cells. Viruses can also be genetically modified for selective homing to tumor cells, for example, by genetic alterations in viral capsids or envelops, generating viruses that specifically recognize only tumor-associated surface markers and infect tumor cells [17]. Many DNA and RNA viruses have been used in melanoma treatment, including Newcastle disease virus (NDV) [18], adenovirus [19], herpes simplex virus [20], influenza virus [21], coxsackievirus [22], reovirus [23], vesicular stomatitis virus (VSV) [24], parvovirus [25], vaccinia virus [26, 27], measles virus [28] and myxoma virus [29].

Oncolytic viruses typically infect only a fraction of all tumor cells in the injected tumor mass. However, these viruses induce anti-tumor immunity through antigen release from lysed tumor cells, in combination with virus-induced local innate immune activation. As a result, much of the eventual cancer cell death is mediated by the host’s adaptive immune system. It has been reported that most oncolytic viruses trigger innate immune responses because they produce factors that can specifically bind to toll-like receptors (TLR) on host cells resulting in signaling through the MyD88 pathway [24, 30, 31]. The resulting activation of the IRF and NF-kB pathways causes DCs to produce large amounts of cytokines including type I IFNs [32, 33], which promotes the activation and tumor infiltration of tumor-specific T cells [34, 35]. Wild-type coxsackievirus A21 is a common cold virus that selectively infects and kills melanoma cells through its interaction with intercellular adhesion molecule (ICAM-1) and decay accelerating factor (DAF). These molecules are overexpressed on the surface of malignant melanoma cells compared to the surrounding benign tissue [36]. ICAM-1 is also recognized as a viral attachment receptor for many enteroviruses including CVA13, CVA15 and CVA18 [22]. Phase I/II trials using oncolytic A21 coxsackievirus (Cavatak) have been conducted and found effective and safe [37].

To further enhance the activation of innate, and subsequently adaptive immunity, many oncolytic viruses are modified to express immunostimulatory molecules. A recent study used i.t. therapy with a genetically modified adenovirus, encoding CD40 ligand (Ad-CD40L) in murine melanoma, resulting in tumor regression that was associated with specific T cell responses against TAAs, including the melanocyte differentiation antigen, TRP-2 and the model antigen, chicken Ovalbumin [38]. In another study, the novel oncolytic adenovirus Ad5/3-hTERT-E1A-hCD40L, which includes chimeric Ad5/3 capsid for enhanced tumor transduction, a human telomerase reverse transcriptase (hTERT) promoter for tumor selectivity, and human CD40L for increased efficacy, caused tumor destruction and generated a significantly stronger tumor-specific CD8 T cell response than wild-type adenovirus [39]. Similar to Ad-CD40L, i.t. immunotherapy with an adenovirus vector expressing OX40L (AdOX40L) generated anti-tumor immunity mediated by cytotoxic CD8 T lymphocytes (CTLs) [40]. Furthermore, an adenoviral vector expressing siRNA against the mouse IL-17A gene (Ad-si-IL-17) significantly inhibited both MC38 and B16 tumor growth and induced a Th1-dominant environment, which selectively eliminated MDSCs and Tregs at tumor sites but not in the spleen [41].

A recent study compared i.t. lipopolysaccharide (LPS) plus i.t. VSV to i.t. LPS plus i.v. VSV. Administration of both drugs i.t. synergized well and generated a profound anti-tumor immune response, whereas the combination of i.t. LPS with systemically delivered VSV resulted in rapid morbidity and mortality in the majority of mice [42]. This study highlights the benefits of i.t. over systemic treatment in terms of reduced toxicity. Another recent study showed that combination therapy with i.t. NDV and systemic CTLA-4 blockade regressed distant tumors and developed long-term CD8 T cells memory against poorly immunogenic B16 tumors. The therapeutic effect was dependent on CD8+ cells, natural killer cells and type I IFN [43].

Clinical trials of i.t. immunotherapy with oncolytic viruses

Mastrangelo et al. [44] used i.t. vaccinia virus encoding recombinant granulocyte–macrophage colony-stimulating factor (GM-CSF) and found it to be safe and effective in patients with metastatic melanoma. Similarly, in a phase II trial of talimogene laherparepvec (T-VEC), an oncolytic herpes simplex virus encoding GM-CSF, the virus was administered i.t. to patients with metastatic melanoma and caused complete regression of injected and uninjected lesions in eight of 50 patients [45] (Table 1). T-VEC also showed improvement in durable response rate (DRR) and improvement in overall survival approached statistical significance in a Phase III trial when compared to systemic GM-CSF control treatment [46]. Phase I/II trial based on intratumoral injections of adenovirus-IL-2 (TG1024) have been conducted on patients with melanoma. The virus induced pronounced inflammation in the treated lesions with predominant CD8+, TIA+ (a granule-associated protein of cytotoxic T cells) lymphocytic infiltrates [47]. In a phase I clinical trial, i.t. injections of adenovirus-delivered mda-7/IL-24 (Ad.mda-7) proved safe, elicited tumor-suppressive and immune-enhancing processes including Th1 cytokines production and CD8 T cells activation and provided clinically significant anti-tumor activity [48]. Herpes simplex virus (HSV) have also been used to treat mouse and human glioblastoma and found to be a promising treatment for patients with glioblastoma [49].

Table 1 Clinical trials based on intratumoral immunotherapy
Full size table

Naturally occurring or genetically modified oncolytic viruses are thus promising agents for the treatment of melanoma. The host-induced primary immune response against the virally infected tumor cells causes limited tumor destruction and cytokine production, and APC are activated, take up tumor antigen from killed tumor cells and prime tumor-specific CD8 T cells that may ultimately cause systemic anti-tumor immunity.

I.t. immunotherapy with synthetic immunoagonists

Synthetic TLR agonists can stimulate the TLR-MyD88 pathway, activate innate immune cells and lead to tumor-specific adaptive immunity. Combination therapy with i.t. poly I:C (a TLR-3 agonist) and CpG (a TLR-9 agonist) with adoptive T cell transfer eradicated established melanoma through an IFN-gamma-dependent mechanism [50]. We and others have shown that i.t. and peritumoral CpG administration routes are superior to intravenous and subcutaneous routes for the activation of innate immune cells leading to induction of tumor-specific CTLs and long-lasting tumor protection [51–53]. It has also been reported that TLR-9 expression on plasmacytoid DC is critical for the therapeutic effect of CpG [53]. A recent preclinical study showed that combination therapy with anti-CTLA-4 and anti-OX40 antibodies together with CpG was able to modify the tumor microenvironment by depleting tumor-infiltrating Tregs; this increased the therapeutic efficacy of CpG and generated a systemic anti-tumor immune response that eradicated distant tumors, including in the brain [54]. Shirota et al. reported that i.t. administration of CpG oligodeoxynucleotides (ODNs) reduced the immunosuppressive activity of monocytic MDSCs (CD11b+Ly6G−Ly6Chi) which differentiated into tumoricidal macrophages (CD11b+F4/80+Ly6Chi). Monocytic MDSCs are present in high numbers in many tumors and suppress anti-tumor T cell function; therefore, switching the phenotypes of these cells from tumor promoting to tumor suppressing may enhance anti-tumor immunity. These studies provide insight into a novel mechanism by which CpG ODNs contribute to tumor regression and also provide an example of how local treatment by a TLR agonist can convert a tumor-promoting microenvironment to a tumor-suppressing one [55]. I.t. injection of polyguanosine ODNs boosted anti-tumor immunity mediated through direct phosphorylation of Lck in CD8 T cells, resulting in expansion of CD8 T cells and IL-2 production [56].

I.t. injection of plasmid DNA encoding CD40L (pSP-D-CD40L) with TLR-9 and TLR-3 agonists (CpG and poly I:C) changed the tumor microenvironment by increasing the number of cytotoxic CD8+ T cells and decreasing the number of DCs (it may due to the activation and migration of DCs to tumor-draining lymph nodes): These changes slowed tumor growth and prolonged mouse survival [57]. It has been shown that i.t. treatment with GVAX (GM-CSF secreting whole-cell tumor cell vaccine) plus LPS (TLR-4 agonist) was more efficient at generating anti-tumor responses than GVAX alone [58]. An i.t. TLR-2 agonist switched mast cell phenotypes from tumor promoting to tumor inhibiting and secreted cytokines IL-6 from activated mast cells causing B16 melanoma regression [59]. These studies suggest that direct activation of the tumor-infiltrating TLR+ innate immune cells by TLR agonists induces effective innate immune response and also prime tumor-specific CD8 T cell responses that could be long lasting and systemic if combined with T cells activating drug.

The synthetic TLR-7 agonist imiquimod showed synergy with live recombinant listeria vaccine and significantly enhanced its anti-tumor effects against murine melanoma [60]; similarly, this TLR-7 agonist also enhanced the anti-melanoma effects of IL-2 [61]. Recently, two studies [62, 63] showed that topical treatment with imiquimod suppressed tumor growth by converting pDCs into granzyme B-expressing, tumor-killing effector cells. In addition, we found that i.t. treatment with a tissue-retained, injectable form of a TLR-7/8 agonist, 3M-052, suppressed melanoma growth through T and B cell-dependent mechanisms. It also stimulated tumor-associated macrophages and polarized them from M2 to M1 phenotypes, and these macrophages contributed to the anti-tumor activity [3].

CD137 or 4-1BB is a member of the TNF receptor family that is expressed on activated T cells and crosslinking of CD137 with its ligand enhances T cell proliferation and cytokine production. It has been shown that CD137 agonist antibodies enhance the cytolytic function and proliferation of CD8 T cells and leads to regression of established tumors [64] I.t. delivery of liposome-coupled anti-CD137 plus IL-2-Fc fusion protein appeared very effective against B16 melanoma and did not cause toxicity [65]. I.t. SFV-IL-12 (Semliki Forest virus encoding IL-12) and systemic anti-CD137 agonist antibodies therapy generated robust anti-tumor immunity against B16 melanomas (B16-OVA and B16.F10) and TC-1 lung carcinomas. I.t. injection of SFV-IL-12 induced strong expression of CD137 on CD8+ T lymphocytes, providing targets for the action of the CD137 agonist antibody [66]. Recently, Marabelle et al. described that pattern recognition receptor agonists including TLR agonists and immunostimulatory monoclonal antibodies like anti-CD137 and anti-CTLA-4 showed synergistic effect when delivered intratumorally and stimulated the tumor-infiltrating leukocytes. This suggests that i.t. immunotherapy would be a better option to generate systemic anti-tumor immune response with lower toxicity than after systemic administration [67, 68]. Another report also compared local delivery of slow release formulation of anti-CTLA-4 or anti-CD40 near to tumor lesion with systemic therapy and found that controlled local delivery of immunomodulating antibodies generates systemic anti-tumor CD8 T cells immunity but minimal toxicity compared to systemic treatment [69]. Interestingly, peritumoral route of treatment initiated different types of immune response than intratumoral treatment, likely because of differences in the immune environment of tumors versus normal tissues such as skin. Therefore, it is important to understand which immune cells initiate and modulate effective adaptive anti-tumor immune response.

I.t. immunotherapy by synthetic STING agonist

STING (stimulator of interferon genes) is a transmembrane containing protein that is localized in the endoplasmic reticulum (ER) of numerous cell types such as macrophages, dendritic cells, endothelial and epithelial cells [70]. It induces type I interferon production through recognition of pathogen or tumor-derived cytosolic DNA metabolites [71, 72]. It has been shown to activate downstream transcription factors STAT6 and IRF3 through TBK1, resulting in anti-viral and innate immune responses against intracellular pathogens [73]. CD8α+DCs and type I interferons are required for spontaneous T cell priming in growing tumors [74], and a major defect in both type I IFN induction and T cell priming was observed in STING−/− mice, as well in mice lacking the downstream transcription factor IRF3 [75]. Presence of tumor cell DNA in host APCs correlated with STING pathway activation and IFN-β production [76].

DMXAA is a known strong agonist of the mouse STING pathway, and i.t. injection of DMXAA-induced tumor antigen-specific CD8+ T cells had robust anti-tumor activity via a mechanism that was dependent on the host STING [77]. Intratumoral administration of a different STING agonist (cyclic diguanylate monophosphate; c-di-GMP) improved the survival of glioma-bearing mice associated with enhanced type I IFN signaling and T cell migration into the brain [78].

Thus, activation of STING pathway by naturally occurring tumor DNA is the main source of type I interferon production in tumor, which initiates tumor-specific immunity at some extent, and that immune response can be enhanced multifold by direct i.t. delivery of synthetic STING agonists and can have profound therapeutic efficacy. However, all known murine STING agonists do not bind to human STING which may limit their use clinically. New generations of STING agonist are under development that binds murine as well as all known human STING variants.

Clinical trials of i.t. immunotherapy with synthetic immunoagonists

Food and Drug Administration-approved synthetic TLR7 agonist imiquimod is a cream formulation for the treatment of cutaneous basal cell carcinoma, actinic keratosis and genital warts, and has limited activity against cutaneous melanoma and breast tumors [79–82]. The cream formulation of imiquimod limits its application for deep, non-cutaneous tumors and systemic administration of TLR agonists is limited by severe toxicity, including cytokine storm [83]. Thus, the newly developed injectable, lipid-modified TLR7/8 dual agonist 3M-052 that is shown to have therapeutic efficacy against mice melanoma, could be a better option for future clinical trials of i.t. therapy of melanoma [3, 84].

A phase I trial of i.t. treatment with a TLR-9 agonist, PF-3512676, showed local tumor regression in patients with basal cell carcinoma (one complete regression and four partial regressions out of five treated patients) and metastatic melanoma (one complete regression out of five treated patients). All patient’s post-treatment biopsies showed moderate to abundant cellular infiltrates of lymphocytes in injected and uninjected lesions [85]. Carpentier et al. [86] conducted a phase II trial to test the efficacy of i.t. CpG ODNs in patients with recurrent glioblastoma after radiotherapy and chemotherapy; the median overall survival was 28 weeks; however, this trial did not meet the targeted progression-free survival benefit in patients with recurrent GBM. Recently, Salazar et al. [87] treated a patient with facial embryonal rhabdomyosarcoma by i.t. and intramuscular injections of a stabilized dsRNA viral mimic, polyinosinic-polycytidylic acid-polylysine-carboxymethyl cellulose (poly-ICLC, Hiltonol) and reported tumor regression with extended survival. Thus, there is some evidence of clinical success with i.t. TLR agonists in specific settings, particularly superficial cutaneous disease.

TLR agonists are very promising adjuvants for i.t. immunotherapy because most tumor-associated innate immune cells express TLRs and can easily be activated in response to these agonists. Some of these cells are able to switch their phenotypes from immunosuppressive to immune enhancing and kill tumor cells directly and/or prime tumor-specific T cells. If these tumor-specific T cells receive further stimulation through TNF receptor super family (TNFRSF) members such as 4-1BB or OX40, which keeps them in the activated stage and enhances their functions, the result can be a strong, long-lasting, systemic anti-tumor immunity. Therefore, combination therapy with TLR agonists and 4-1BB/OX40 agonists may be a promising strategy to treat established tumors.

I.t. immunotherapy with cytokines

I.t. administration of cytokines such as IL-2, IL-21, IFN-alpha/beta and IL-12 can also generate anti-tumor immunity and suppress tumor growth. Combination therapy with i.t. IL-12 and systemic anti-CTLA-4 led to eradication of murine glioblastoma in mouse [88]. I.t. administration of IL-21 showed better treatment effect than subcutaneous injection and caused superior CD8 T cell proliferation [89]. I.t. delivery of plasmid DNA encoding IL-12 by in vivo electroporation induced systemic immunity that was able to kill both injected and uninjected, distant tumors without the systemic toxicity commonly observed after systemic administration of cytokine protein [90]. The combination of i.t. IL-12 with T cells redirected against vascular endothelial growth factor receptor-2 had therapeutic efficacy in mice with a variety of solid tumor types, including melanoma [91]. I.t. hu14.18-IL-2 showed better anti-tumor activity in mouse models compared with i.v. hu14.18-IL-2 [92]. I.t. DC-IFN-gamma efficiently induced cross-presentation of tumor antigens to specific CD8+ T cells and generated anti-tumor immunity against pre-established B16 melanoma [93]. Van der Jeught et al. [94] showed that i.t. administration of mRNA encoding a fusion protein consisting of interferon-β and the ectodomain of the transforming growth factor-β receptor II induced anti-tumor immunity in mice by enhancing the antigen-presenting capacity of dendritic cells and reducing the suppressive activity of myeloid-derived suppressor cells.

Clinical trials of i.t. immunotherapy with cytokines

In a clinical trial, i.t. administration of plasmid encoding IL-12 was found to be effective in melanoma. Two of nine patients showed stable disease, and one had a complete response. Patients, especially responders, generated antigen-specific immunity against MAGE-1 and MART-1 antigens [95]. L19-IL-2, an immunocytokine made up of the recombinant human antibody fragment L19 (specific to the alternatively spliced EDB domain of fibronectin, a well-characterized marker of tumor neo-vasculature) and of human IL-2, has shown therapeutic efficacy in animal cancer models. Twenty-five patients with stage IIIB/IIIC melanoma and cutaneous/subcutaneous injectable metastases were treated i.t. with L19-IL-2, resulting in complete response (CR) in 25 % of patients by modified immune-related response criteria (irRC) [96] (Table 1).

Overall, i.t. cytokine therapy is a rapid and effective method to generate tumor-specific adaptive immunity in several mouse models, with some evidence of efficacy in patients with melanoma.

I.t. immunotherapy with activated immune cells

Activated antigen-presenting cells are needed to prime tumor-specific T cells and develop long-lasting immunity. In this regard, many studies have been performed to determine the effectiveness of in vitro activated DCs in in vivo tumor killing. While most of these studies used antigen-loaded DCs as vaccines for subcutaneous, intravenous or intranodal delivery, some studies directly introduced DCs i.t. to promote uptake of tumor antigen and T cell priming. Okano et al. reported that i.t. delivery of Fas-inhibited allogeneic DCs had anti-tumor effects similar to those of autologous DCs, and this approach offers an alternative in patients where autologous DCs cannot be used. This study suggests that blocking the Fas–FasL interaction between allogeneic DC and host T cells may be an useful strategy to overcome the rejection response against alloantigens on the DCs [97]. We have shown that i.t. administration of TLR9-triggered pDCs induced robust anti-tumor immunity that resulted in regression of the treated tumor as well as distant tumors by natural killer cells and CD8 T cell-mediated mechanisms [98]. I.t. injection of immature DCs and IFN-gamma into malignant tumors in dogs produced anti-tumor immunity including four complete responses and two partial responses out of seven treated dogs [99]. In B16 melanoma, i.t. injection of poly I:C-treated DCs generated anti-tumor immunity and led to infiltration of TRP-2-specific IFN-gamma-producing CD8+ T cells [100].

Other i.t. immunotherapy

Recently, i.t. injections of recombinant heat-shock protein (Hsp)70 were used to treat malignant brain tumors in children; the therapy was safe but not highly effective (one of 12 children had a complete response, and one had a partial response). However, an increased number of Th1 T cells and decreased number of B and Treg cells were seen in the blood of all children in response to Hsp70 treatment [101], suggesting that locally injected Hsp70 may generate a systemic anti-tumor immune response that could be further enhanced with additional immunomodulators such as the T cell checkpoint blockade agents approved for the treatment of metastatic melanoma, anti-CTLA-4 (ipilimumab) and anti-PD-1 (pembrolizumab).

In a study of i.t. therapy in mice, injection of alpha-gal glycolipids into experimental melanomas induced CD8 T cell-mediated protective immunity that was not only effective against the treated tumor but also suppressed distant metastasis [102].

Allovectin (velimogene aliplasmid) is an immunotherapeutic drug for direct i.t. administration. It is a plasmid that encodes both major histocompatibility complex (MHC) class I heavy (HLA-B7) and light chains (β2-microglobulin) and i.t. administration of this plasmid stimulates both local and systemic anti-tumor immune responses [103]. Phase II and III trials were conducted using allovectin in 127 and 375 patients, respectively. However, in the phase III trial, allovectin failed to improve overall survival [37].

Sandin et al. [104] have suggested that local administration of anti-CTLA-4 monoclonal antibodies is a better option to treat pancreatic adenocarcinoma than systemic treatment because local treatment had similar treatment efficacy to systemic treatment and did not cause accumulation of Treg cells in secondary lymphoid organs. However, i.t. therapy of tumors at visceral sites such as the pancreas is more challenging than systemic therapy due to the need for image-guided i.t. administration, requiring specific equipment and expertise.

Perspectives

The induction of anti-tumor immunity by immune activation within tumors is an effective method to generate immune responses against multiple (self and non-self) tumor antigens and modify the tumor microenvironment. A great advantage is the possibility of producing “off-the-shelf” reagents that can be applied across patient populations and tumor types, since antigens do not need to be identified. This approach could therefore be a more rapid, broadly applied and cost-effective therapeutic option than personalized immunotherapy. However, care should be taken in choosing the specific i.t. immunotherapy method because strong immune activation results in the up-regulation of chemokines within a tumor and may cause an influx of immune cells only in treated tumors and a lack of immunity or cell killing in untreated tumors. At this time, a wealth of approaches is being tried, and it is as of yet unclear which of these are most potent in their ability to induce systemic anti-tumor immunity and therapeutic benefit. Therefore, well-designed preclinical studies should be conducted to observe the effect of a particular drug and its mechanisms of action in metastatic disease models and especially in clinical trials. Administration of a reproducible amount of therapeutic agent into tumors and immune-suppressive tumor environments can be challenging; however, i.t. immunotherapy could be safer than systemic treatment, with its local nature giving rise to fewer side-effects such as systemic tissue inflammation or cytokine storm. As an exciting prospect, i.t. therapy can enhance the effect of the FDA-approved checkpoint blockade therapeutics, anti-CTLA-4 and anti-PD-1, by promoting the generation of activated, tumor-specific T cells that can then become the targets for checkpoint blockade.

Conclusion

Many local tumor immunomodulation therapies to generate systemic anti-tumor immunity are emerging. The benefits of i.t. immunotherapy depend not only on the generation of tumor-specific immunity but also on changing the tumor microenvironment from immunosuppressive to immunostimulatory (Fig. 1). This approach holds great promise in combination with chemotherapy and systemic immunotherapies. Opportunities include (1) i.t. immunotherapy with adoptive T cell transfer, (2) i.t. oncolytic viruses encoding immunomodulatory molecules and (3) combination of i.t. immunoagonists with i.t. or systemic anti-CTLA-4 or anti-PD1/PD-L1 blockade. These synergistic combinations may provide a promising approach to generate systemic anti-tumor immunity for the treatment of metastatic melanoma with superior efficacy over single-agent approaches.

Fig. 1
figure 1

Activation of tumor-associated immune cells by i.t. immunotherapy generates strong local and systemic immunity: (i) i.t. delivery of immunoagonists, cytokines and other immune activating agents, (ii) activation of tumor-associated innate immune cells through toll-like receptor, CD40 or cytokine pathway leads to conversion of M2 macrophages, MDSC, DCs/pDCs and B cell to M1 macrophages, killer pDC/DC and B1 cells, respectively. (iii) These activated cells produce many cytokines or/and lytic molecules and kill tumor cells directly. (iv) They can also act as an antigen-presenting cells and (v) migrate to TDLN where they process and present tumor antigens to CD8 T cells to (vi) expand tumor-specific CD8 T cells and generate long-lasting tumor-specific immunity. (vii, viii, ix) Secondary co-stimulation or blocking of inhibitory pathway further activates and facilitates migration of these tumor-specific CD8 T cells to injected and uninjected tumor and cause tumor lysis

Full size image