Rationale and evidence to combine radiation therapy and immunotherapy for cancer treatment

Abstract

Cancer immunotherapy exploits the immune system’s ability to differentiate between tumor target cells and host cells. Except for limited success against a few tumor types, most immunotherapies have not achieved the desired clinical efficacy until recently. The field of cancer immunotherapy has flourished with a variety of new agents for clinical use, and remarkable progress has been made in the design of effective immunotherapeutic regimens. Furthermore, the therapeutic outcome of these novel agents is enhanced when combined with conventional cancer treatment modalities including radiotherapy (RT). An increasing number of studies have demonstrated the abscopal effect, an immunologic response occurring in cancer sites distant from irradiated areas. The present work reviews studies on the combination between RT and immunotherapy to induce synergistic and abscopal effects involved in cancer immunomodulation. Further insight into the complex interactions between the immune system and cancer cells in the tumor microenvironment, and their modulation by RT, may reveal the abscopal effect as a clinically relevant and reproducible event leading to improved cancer outcome.

Introduction

Radiation therapy (RT) is an essential component in the treatment of various tumors. The advent of sophisticated imaging methods and irradiation devices has led to the development of three-dimensional (3D) conformal RT, which uses beams arranged in 3D to focus the doses. Intensity-modulated RT (IMRT) uses lead collimators to shape the beam to match the tumor. Stereotactic ablative radiation (SAbR or SBRT) is an emerging treatment paradigm that uses both technologies in combination with image-guided RT (IGRT) to further reduce target uncertainty. RT is indicated for the local control of primary or limited metastatic cancer, but is not applicable to widespread metastatic disease beyond palliative purposes. Immunotherapy can be an effective systemic treatment for metastatic cancer, but significant cancer subpopulations will either not respond or will develop resistance due to their immune evasive properties [1]. However, the combination of RT with immunotherapy is known to improve the therapeutic efficacy of either treatment alone.

Leukocytes are highly sensitive to radiation as shown by the immunosuppressive use of whole-body irradiation before bone marrow transplantation; their involvement in the synergy between RT and immunotherapy may be essential. The activation of immune cells due to irradiation during cancer therapy has been reported [2]. Reits et al. [3] showed that cancer cell surface expression of MHC class I molecules was increased after RT in a dose-dependent manner, leading to the recognition of irradiated cells by cytotoxic T lymphocytes (CTLs). The paradoxical relationship between the immunosuppressive and immune-activating effects of RT can be partly explained by the focused radiation field used in modern treatments. These radiation fields can release tumor-associated antigens (TAAs) from irradiation-induced dying tumor cells, alter the tumor microenvironment, and simultaneously spare adjacent lymph nodes needed for an adaptive immune response. The introduction of TAAs into the immune system and other radiation-induced changes in the tumor microenvironment can lead to the recruitment of radiation-naïve immune cells; these events increase immune cell infiltration at local and distant metastatic sites leading to increased tumor cell death (reviewed in [4]).

This article reviews the literature on the rational combination of RT and cancer immunotherapy, starting with RT-induced abscopal effects as evidence of the immunogenic effect of RT. We will also analyze preclinical studies to analyze optimal RT doses and fractionation in the rational combination with cancer immunotherapy.

The abscopal effect

The abscopal effect [from the Latin ab (outside) and scopus (target)] is the indirect, systemic anticancer effect of RT at distant, non-irradiated sites. Since its first description by Mole [5], the abscopal effect has been uncommonly observed in the clinic and its underlying mechanism has been poorly understood until recently. A limited number of cases have been reported in response to RT as monotherapy. In a retrospective analysis of 62 patients with stage II–IV breast cancer treated with preoperative RT, Konoeda [6] observed an abscopal effect in metastatic lymph nodes in 15 of 42 cases. Upon further examination of the lymph nodes, a histopathological abscopal effect was detected in 22 of 42 cases. Moreover, the abscopal effect was found in patients with infiltrating T cells around the degenerated cancer cells in irradiated primary tumor nests.

Demaria et al. [7] showed that the abscopal effect was abolished when RT was delivered to immunodeficient nude mice. Recent preclinical studies combining RT with immunotherapy have reliably reproduced distant tumor regression outside the irradiation field using tumor syngeneic mice models, validating these combinations as promising strategies. Wersall et al. [8] showed that non-irradiated renal cell carcinoma (RCC) lesions regressed in 4 of 28 patients (14 %) following treatment with hypofractionated RT. None of the 4 patients had received systemic treatment while one had previously been treated with systemic IL-2. These results suggested that the abscopal effect was due to RT and not to other treatments.

Camphousen et al. [9] reproduced the abscopal effect of RT using mice bearing tumor grafts. Tumor-bearing hind legs of mice received focal irradiation to determine whether an abscopal effect could be observed against lung or fibrosarcoma implanted at a distant site. In wild-type mice, tumors outside the RT field grew at a slower rate than sham-irradiated animals when the leg was exposed to hypofractionated RT. A deficiency in p53 or its pharmacological inhibition induced the loss of the abscopal effect even upon delivery of hypofractionated RT. These data suggested that p53 and pathways downstream of it may be involved in the RT-induced abscopal effect [9].

A recent prospective study reported an abscopal effect following different combination of chemo-irradiation and GM-CSF, a potent stimulator of dendritic cell (DC) maturation [10]. Forty-one patients with metastatic solid cancers (mostly lung or breast cancer) were treated with GM-CSF and concurrent fractionated RT delivered to one metastatic site. An abscopal response occurred in 11 of 41 accrued patients (27 %). Similarly, a study of 21 patients with melanoma who progressed after receiving ipilimumab and palliative RT reported an abscopal response in 52 % of the patients and an increase in the median survival [11]. A local response to RT was detected in 13 patients (62 %) with 11 (85 %) showing an abscopal response. An abscopal effect was primarily observed in those exhibiting a local response, strongly suggesting this as a prerequisite to attain this response. Although a larger prospective study is required to evaluate clinical relevance and applicability, these findings strongly suggest that a combination of RT with immunotherapy could improve abscopal response rates and increase survival. These studies have collectively reinvigorated the interest in combining radiation with immunomodulatory therapy.

A recent article reviewed preclinical and clinical studies on RT leading to an abscopal effect with targeted cancer immunotherapy but no concurrent cytotoxic treatment [12]. The authors described 23 case reports and 13 preclinical studies on the abscopal effect. They observed that 11 of the 13 preclinical studies used immune-modulating treatment to achieve an abscopal response. With radiation dose fractionation sizes ranging from 1.2 to 26 Gy per fraction, the median time to induce an abscopal response was 5 months [12]. The review also reported a progression-free median time of 13 months following an abscopal response, suggesting strong synergism between RT and immune treatments. Another article recently discussed the abscopal effect generated by the combination of local irradiation with immunotherapy, focusing on the combination of radiation and chemotherapy, which was also immunostimulatory [13]. These studies demonstrated that combining RT with immunotherapeutic agents may increase the likelihood of inducing an abscopal effect (Fig. 1).

Fig. 1.

Proposed synergistic effect showing irradiation of one site of the tumor in combination with immune therapy can lead to enhanced immunity against tumor cells outside of the radiation field

Radiation and immune checkpoint inhibitors

Studies on the regulation of T cell responses have elucidated immune checkpoint mechanisms and have examined the synergistic effect of RT combined with various immune checkpoint inhibitors (Table 1). Ipilimumab is a mAb against cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) that ordinarily transmits an inhibitory signal for T cells, limiting the induction of immunity by acting in the priming phase of CTLs. Therefore, the administration of anti-CTLA-4 mAb leads to an increase in T cell activation. Anti-CTLA-4 mAb improved the survival of patients with metastatic melanoma in two randomized phase III trials, when administered as either a second-line [14] or a first-line therapy [15]. A landmark preclinical study showed that combining anti-CTLA-4 mAbs with radiation leads to an immune-mediated abscopal effect [16]. Breast or colon carcinoma cells were injected into mice in two separate locations with the second site outside the irradiation field. The combination of anti-CTLA-4 mAb and fractionated RT achieved complete regression at the primary site and delayed tumor growth at the secondary site, while treatment with either therapy alone could only delay tumor growth at the primary site. Furthermore, a significant increase in tumor-infiltrating lymphocytes (TILs) was found in secondary tumors of mice treated with a combination of anti-CTLA-4 mAb and fractionated RT [16]. These results suggested that the combination of targeted immune checkpoint therapy and radiation may achieve tumor regression via enhanced immune cell activation, recruitment of immune cells, and tumor cell killing. Kwon et al. conducted a randomized phase III trial on 799 patients with metastatic castration-resistant prostate cancer (mCRPC) who progressed after docetaxel chemotherapy. They compared the efficacy of anti-CTLA-4 antibody versus placebo after RT to bone metastasis [17]. Anti-CTLA-4 with RT improved progression-free survival and reduced prostate-specific antigen levels; yet, no significant difference was found in the overall survival (OS) (hazard ratio (HR) 0.85, p 0.053). The authors examined the piecewise hazard model that showed a change in HR over time: HR was reported as 1.46 for 0–5 months, 0.65 for 5–12 months, and 0.60 for 12 months and over. While these results did not meet the improved OS primary endpoint, maybe because of insufficient power, this was one of the first studies to provide evidence of enhanced antitumor activity due to the combination of RT with immune checkpoint inhibitors in a clinical setting. These data indicated that combination therapy could be further optimized to improve clinical outcome.

Table 1.

Preclinical studies that combine RT with immune checkpoint therapies

Tumor/cell line	Immune therapy	Radiation therapy	Highlights	References
Breast (4T1)	Anti-CTLA-4	12 Gy×1 or ×2	Control or anti-CTLA-4Ab alone did not have an effect on primary tumor growth or survival RT + Anti-CTLA-4Ab treatment had a significant survival advantage, which correlated with inhibition of lung mets; required CD8+ but not CD4+ T cells	[51]
Breast cancer (AT-3)	Anti-CD137, anti-PD1	12 Gy×1 subQ tumor 12 Gy×1 or 4–5 Gy×4 orthotopic tumor	Alpha-CD137 and alpha-PD-1 mAbs combined with single- or low-dose fractionated radiotherapy cured all mice with established orthotopic AT-3 mammary tumors CD8+ T cells were essential for curative responses to this combinatorial regime	[21]
Breast (TUBO) and colon (MC38)	Anti-PD1	12 Gy for TUBO, 20 Gy for MC38	Anti-PD-L1 Ab combined with RT enhanced the efficacy of treatment in a CD8+ T-cell-dependent manner Anti-PD-L1 Ab decreased MDSCs accumulation T cell-derived TNF induced MDSC death TNF blockade counteracted the efficacy of combination treatment	[52]
Colon (MC38)	Anti-CTLA-4Ab	20–30 Gy (20 Gy×1, 8 Gy×3 or 6 Gy×5 in consecutive days)	Each RT regimen caused comparable growth delay of primary tumors but had no effect on secondary tumors outside the RT field Anti-CTLA-4Ab alone had no detectable effect Combination of anti-CTLA Ab and RT enhanced the tumor response at the primary site An abscopal effect only occurred in mice treated with a combination of anti-CTLA Ab and RT The frequency of CD8+ T cells, producing tumor-specific IFN-γ, was proportional to the inhibition of the secondary tumor	[16]
RCC (RENCA) melanoma (B16-OVA) and breast (4T1)	Anti-PD-1	SABR 15 Gy×1	PD-1 expression correlated with a decrease in the survival of tumor-bearing mice Combination therapy induced near-complete response of irradiated tumors as compared to either therapy alone and a 66 % reduction in non-irradiated secondary tumors outside of RT field CD11aCD8(+) T cells were identified as a tumor-reactive population associated with SABR-induced antitumor response in frequency and function	[23]
Colon (CT26) breast (4T1) melanoma (BRAF-4434)	Anti-PD-1 and anti-PD-L1	2 Gy×5	Fractionated RT increased PD-L1 of tumor cells via IFN-gamma produced by CD8+ T cells Concurrent but not sequential administration of anti-PD-L1 mAb with fractionated RT was required to improve the survival of tumor-bearing mice	[25]
GBM (GL261)	Anti-PD-1	10 Gy×1	Mice bearing GL261 tumors stratified into groups: (1) control; (2) RT only; (3) anti-PD-1 Ab only; and (4) RT plus anti-PD-1 Ab Improved survival with combination therapy compared with either modality alone Increased tumor infiltration by cytotoxic T cells and decreased Tregs in mice that received combined therapy	[53]

Ab antibody, EBRT external beam radiation therapy, IV intravenous, hr hour, LLC Lewis Lung Carcinoma, mets metastases, SABR Stereotactic Ablative Radiotherapy, TH1 T helper type 1 cell, Tregs: regulatory T cells, RT radiotherapy

Programmed death 1 (PD-1) is a cell surface receptor that signals T cell programmed death, thereby limiting the overall extent and duration of an immune response by acting on the CTLs' effector phase. Increased expression of PD ligand 1 (PD-L1, which induces T cell death via PD-1 activation) by tumor cells has been documented in various malignancies, including pancreatic, esophageal, and ovarian cancers, and melanoma (reviewed in [18]). The inhibitory stimulus from these receptors that ordinarily lead to immune tolerance is exploited by the tumor cells; immune checkpoint inhibitors attempt to inhibit this immune evasion process [19]. Reversing immune suppression using anti-CTLA-4 or anti-PD-1 mAbs inhibits T cell suppression in the tumor microenvironment as shown by promising completed and ongoing clinical trials (reviewed in [20]). Some of the underlying mechanisms used by RT to enhance the therapeutic effects of anti-PD-1 and PD-L1 mAbs have been investigated. Verbrugge et al. [21] found that the combination of RT with anti-PD-1 and anti-CD137 antibodies, acting as co-stimulatory molecules, was a curative treatment for both established and secondary tumors. Since the expression of CD137 on tumor-specific CD8+ T cells was mostly restricted to a subset that highly expressed PD-1, the curative effect of the combined regimen may be mediated by the activation and/or enhancement of CD8+ T cell response against tumor cells [21]. Deng et al. [22] reported that PD-L1 was upregulated in the tumor microenvironment after high-dose RT (12–20 Gy×1), while the administration of anti-PD-L1 mAb enhanced the efficacy of RT in a CD8+ T cell-dependent manner. In addition, they noticed that anti-PD-L1 mAb in combination with RT decreased the local accumulation of MDSCs suppressing T cells through an enhanced production of T cell-derived TNF. Park et al. [23] showed how the combination of SAbR with PD-1 blockade induced the near-complete regression of irradiated primary tumors, as opposed to either therapy alone. Furthermore, PD-1-deficient tumor-bearing mice showed an enhanced SAbR-induced abscopal effect and survived longer than control mice. These data suggested that SAbR and checkpoint inhibitor combination therapy can be particularly effective in the metastatic setting.

The therapeutic efficacy of conventionally fractionated RT is enhanced by the blockade of PD-1 [24] or PD-L1 [25]. Dovedi et al. reported that low doses of fractionated RT induced PD-L1 upregulation in tumor cells via IFN production by CD8+ T cells. The investigators found that concurrent, non-sequential administration of anti-PD-L1 mAb with fractionated RT was required to improve the survival of tumor-bearing mice. Concurrent administration may be required because the upregulation of PD-L1 peaked 72 h after RT. Deng et al. [22] also reported that high-dose RT induced PD-L1 upregulation of DC and tumor cells. Another recent report showed that the addition of anti-PD-1 mAb extended the survival of tumor-bearing mice receiving RT [24]. These studies strongly suggested that combining RT with immune checkpoint inhibitors can enhance the efficacy of each therapy in the clinical setting. The development of agents functioning as immune checkpoint inhibitors is being translated into clinical success with promising results in several tumor types. A recent breakthrough study by Twyman-Saint Victor et al. [26] reported the activation of non-redundant immune mechanisms in melanoma by dual checkpoint blockade with radiation. They found that anti-CTLA-4 predominantly inhibited Tregs, thereby increasing the ratio of tumor-infiltrating CD8+ T cells to Tregs. PD-L1 was upregulated in melanoma cells and was associated with CD8+ T cell exhaustion in patients treated with anti-CTLA-4 and RT who exhibited resistance [26]. Furthermore, they showed that RT increased the T cell repertoire in the host. Because RT increases PD-L1 expression of tumor cells [22, 25], these results provide the mechanistic basis of resistance to RT, and indicate the use of immune checkpoint inhibitors to circumvent treatment failure.

Immune checkpoint inhibitors provide non-specific immune stimulation that relies on preexisting tumor-specific immunity for optimal therapeutic efficacy. RT increases PD-L1 expression of tumor cells [25], and SAbR is well known for its ability to initiate tumor antigen presentation and increase tumor-specific CTLs [27], suggesting that RT can enhance the efficacy of mAb to PD-1/L1 by increasing their binding when administered concurrently. Increased PD-L1 expression by tumor tissue clinically correlated with an increased response of mAb to PD-L1 [28]. Hence, the PD-L1 expression level in tumors is being investigated as a predictive marker. Considering that immune checkpoint inhibition can counter RT immunosuppression and that RT antigen-presenting properties can provide tumor specificity to immune checkpoint inhibition, the concurrent administration of the two modalities is likely synergistic. Despite multiple preclinical studies that demonstrated the synergy between SAbR and immune checkpoint inhibitors [16, 21–24, 26, 29, 30], including an immune-mediated abscopal effect [31], clinical application of this combination has not been reported. Multiple clinical trials are currently being conducted to apply this synergy to the clinic.

Radiation with cancer vaccines, vaccine adjuvants, and adaptive T cell transfer

Immunotherapy targets antitumor responses by directly administering tumor antigens, effector cells, or growth factors to promote the maturation of APCs. A TAA response generated following RT induces antitumor-specific immunity, impacting our understanding of vaccine-immune therapy. Multiple preclinical studies have investigated the efficacy of RT combined with vaccination and its underlying mechanisms, as summarized in Table 2. As mentioned earlier, Nesslinger et al. [32] observed that RT in prostate cancer induced de novo antibodies to prostate antigens in both clinical and preclinical settings. In addition to humoral responses, an increase in T cell receptor (TCR) repertoire diversity of tumor-infiltrating T cells recognizing tumor antigens is expected following RT-mediated cell death or apoptosis. Hannan et al. [33] showed significant tumor growth delay and complete regression of established tumors in 60 % of mice treated with RT and immunotherapy using a Listeria monocytogenes-based vaccine in a prostate-specific antigen (PSA)-transfected TRAMP-C1 syngeneic mouse model. An increase in PSA-specific CD8+ cytotoxic T cells was observed in both groups receiving the vaccine alone and the combination therapy. However, an increase in tumor-specific CD8+ CTL and complete regression were only observed in mice receiving the combination therapy. Tumor challenge delivered 6 months or more following combination therapy resulted in tumor rejection. These results clearly showed a synergistic effect of RT and immunomodulatory therapy in enhancing the PSA-specific immune response to prostate cancer.

Table 2.

Prelinical studies that combine RT with vaccine/vaccine adjuvants/adaptive T cell transfer

Tumor/cell line	Immune therapy	Radiation therapy	Highlights	References
Prostate (TRAMP-C1)	Listeria monocytogene-based PSA vaccine (ADXS31-142)	10 Gy×1	Significant tumor growth delay in mice that received combined ADXS31-142 and RT treatment as compared with mice of other cohorts Combined treatment caused complete regression of established tumors in 60 % of the mice Significant increase in IFN-γ production by CD8+ CTL in mice with combined treatment PSA-specific CTLs were also increased in mice receiving ADXS31-142 alone and combination treatment, respectively	[33]
Prostate	Adoptively transferred CD8 + T cells	12 Gy×1	The combination of radiation and immunotherapy resulted in a significant decrease in pathologic tumor grade and gross tumor bulk that was not evident with either modality therapy alone Combination therapy improved overall survival of tumor-bearing mice with an increase in effector-to-regulatory T cell ratio for both CD4 and CD8 tumor-infiltrating lymphocytes	[54]
Prostate, lung, and breast cancer cell lines	CEA-TRICOM recombinant pox virus-based vaccine with CEA	Samarium-153 (0–25 Gy×4; total 0–100 Gy by Samarium-153	Tumor cells upregulated the expression of surface molecules Fas, CEA, mucin-1, MHC class I, and ICAM-1 in response to S153Sm-EDTMP Exposure to 153Sm-EDTMP rendered LNCaP cells more susceptible to killing by CTLs specific for prostate-specific antigen, CEA and mucin-1 RT with vaccination (CEA-TRICOM) altered the phenotype of tumor cells, suggesting that RT may synergistically increase tumor cell susceptibility to CTL killing	[55]
Lung cancer (LLC)	Th1 Cells with Ovalbumin antigen (OVA)	20 Gy×1 or 18 Gy×1	Mice bearing established LLC expressing model tumor antigen, LLC-OVA marginally responded to local RT, but mice were not cured Treatment of tumor-bearing mice with intratumoral injection of tumor-specific TH1 cells and OVA after RT prolonged survival and induced complete cure in 80 % of the mice	[56]
Lung (LLC)	Flt3-ligand	60 Gy×1	RT + Flt3L reduced pulmonary mets in mice with LLC and significantly improved survival in C57Bl/6 mice with established footpad tumors Mice treated with Flt3L alone showed delayed tumor growth but eventually succumbed to tumor progression The combination therapy of RT + Flt3L did not impact the survival of immunodeficient athymic mice	[57]
Breast (67NR)	Flt3-Ligand	2–6 Gy×1	Flt3-L alone had no effect without RT An abscopal effect was shown to be tumor-specific as it did not affect the growth of a non-irradiated A20 lymphoma in the same mice containing a treated 67NR tumor No abscopal effect (growth delay of non-irradiated 67NR tumors) was observed when T-cell-deficient (nude) mice were treated with RT plus Flt3-L	[7]
Colon (MC38 and MC38 cells expressing CEA)	Vaccinia, Avipox-CEA; anti-CEA mAb	2–30 Gy in vivo, 8 Gy in vitro	Enhanced tumor-infiltrating T cells were found after combination of vaccine and local radiation, which was not seen with either modality alone CD4+ and CD8+ T cell responses specific for CEA were detected in mice cured of tumors; high levels of T cell responses to two other antigens (p53 and gp70) were overexpressed in the tumors	[58]
Colon (MC38 and MC38 cells expressing CEA)	Yttrium-90 labeled anti-CEA mAb and CEA/TRICOM	10 Gy×1	Mice receiving the combination therapy survived longer than those receiving the vaccine or the mAb alone via the Fas/Fas ligand pathway Mice receiving the combination therapy also showed a significant increase in viable tumor-infiltrating CEA-specific CD8+ T cells percentage as compared to the vaccine alone Mice cured of tumors showed CD4+ and CD8+ T cell responses not only specific for the administered CEA, but also toward p53 and gp70	[59]
Colon (MC38); Lung (LL2-CEA+ tumors)	Vaccine against CEA; CD8 + T cells	8 Gy×1 or brachytherapy using 125I-brachytherapy seed	The combination of vaccine and EBRT in primary MC38 tumor resulted in regression of the distant tumor The combination of the vaccine and RT (EBRT or brachy) in LL2-CEA+ primary tumor mediated significant regression of distant or LL2-CEA+ pulmonary mets	[30]
Colorectal (HC 116, SW620)	TSA: histone deacetylase inhibitor & 5-Aza-2′-deoxycytidine: DNA methyltransferases	10 Gy×1	Inhibition of histone deacetylases and of DNA methyltransferases resulted in increased OX40L and 41BBL mRNA and protein expression Co-incubation of T cells with TSA-treated tumor enhanced T cell survival and activation Chromatin immunoprecipitation experiments revealed significantly increased histone H3 acetylation of 41BBL promoters specifically following irradiation	[50]
Fibrosarcoma (FSAR, FSAN, FSAN-JmIL-3 cells)	IL-3 gene-transduced tumor cell vaccines	25 Gy for FSAR (immunogenic cells) and 35 Gy for FSAN (non-immunogenic cells)	Systemic IL-3 vaccine treatment increased intratumoral levels of ICAM-1, Mac-1, EB22/5.3, TNF-alpha, and IL-1 mRNA in irradiated tumors IL-3 vaccines were less effective in TNF-alpha-deficient C3H/HeJ mice as compared to control mice Local RT decreased the tumor burden, thereby enhancing the efficacy of genetically altered vaccine-based immunotherapy for cancer	[60]
Gliosarcoma (9L)	GM-CSF (via adenoviral vector) + DC	30 Gy = 10 Gy×3	RT and GM-CSF vector administration cured 60 % of rats with tumors The same efficacy was observed with a second generation of vaccines combining DC, local tumor irradiation and the controlled supply of recombinant GM-CSF	[61]
Fibrosarcoma (MCA102), lymphoma (EL4), Colon (CT-26)	DC ± TNA-alpha ± LPS	15 Gy×1 or 7 Gy×2	Intratumoral injection of DCs into irradiated tumor induced cytotoxicity of splenocytes against tumor cells as compared to either therapy alone Combined therapy induced antitumor immunity at a distant site Addition of TNF-α or LPS to DC suspension before intratumoral injection potentiated antitumor immunity	[62]
Melanoma (B16), lymphoma (EL4, EG7), Lung (LLC)	Ovalbumin-specific Th1 cells	2 or 15 Gy	Therapeutic effect of RT was almost completely abolished in tumor-bearing mice by depleting CD8+ T cells through anti-CD8 mAb administration Surgical ablation of draining lymph nodes or genetic alteration (Aly/Aly mice) reduced the number of RT-induced tumor-specific CTL	[63]
Melanoma (D5)/sarcoma (MCA205)	Intratumoral injection of DCs	42.5 Gy = 8.5 Gy×5	RT intensified the antitumor efficacy of DC administration independent of apoptosis or necrosis within the tumor mass Combination treatment of DCs and RT was better than subcutaneous injections of tumor lysate-pulsed DCs with IL-2 in inhibiting melanoma tumor growth and prolonged survival Splenocytes from mice treated with DCs and RT contained more tumor-specific, IFN-γ-secreting T cells as compared to control mice Adoptive transfer of these splenocytes mediated significant tumor regression in mice with established pulmonary metastases	[64]

CEA carcinoembryonic antigen, draining lymph node (DLN), ECOG Eastern Cooperative Oncology Group, EBRT external beam radiation therapy, IFN interferon, IL-2 interleukin-2, LLC Lewis Lung cancer, mets metastases, OVA ovalbumin, oligodeoxynucleotides in the CpG motif (CpG-ODN), ORR objective response rate, patient pt, SRS stereotactic radiosurgery, WBRT whole-brain radiation therapy, yo year old

Mason et al. [34] showed that a combination of RT with oligodeoxynucleotides containing unmethylated CpG motifs (that stimulate TLR-9 from APCs) was effective against murine immunogenic fibrosarcoma tumor in the leg of mice. A total dose of 83.1 Gy achieved tumor regression in 50 % of the mice treated with RT alone, while 23 Gy achieved tumor regression in 50 % of mice treated with RT and vaccination. Following combination therapy, tumor-free mice were highly resistant to tumor inoculation re-challenge [34]. These data suggest that RT combined with cancer vaccination amplifies the cancer-specific immune response and induces long-lasting T cell memory preventing recurrence and metastasis.

Radiation with cytokine therapy or cytokine inducers

Cytokines can indirectly affect tumor growth by inducing CTLs or directly act on tumor cells. IFN-α and IL-2 are widely studied examples of passive immunotherapy. IL-2 is a cytokine with pleiotropic effects on T cell function [35] and is approved for use in the US as monotherapy for metastatic renal cell cancer (mRCC) and metastatic melanoma. Since IL-2 was believed to have no direct impact on cancer cells, its effect on mRCC and melanoma is likely due to its ability to expand preexisting T cell populations leading to antitumor activity [36]. Evidence from IL-2-deficient mice also suggests a regulatory role in homeostasis and Treg function [37]. The reported response rates for treatment with IL-2 in multiple phase II and III trials range from 20 to 23.2 % and complete response rates range from 7 to 9 % [38–40]. Although the overall response rates to IL-2 are low, the complete responses induced are durable and potentially curative.

A list of preclinical studies that examined the effect of RT combined with cytokines is presented in Table 3. Yasuda et al. reported the complete regression of established primary colorectal carcinoma and liver metastasis formation in mice receiving RT and intratumoral IL-2 [41]. Analysis of splenocytes from these mice showed a higher percentage of CD4(+) T cells and a lower percentage of Tregs and MDSCs as compared to splenocytes from controls, suggesting that combination therapy augments T-cell-mediated antitumor immunity. Using mouse models of melanoma and breast cancer, Lee et al. showed that ablative RT (20 Gy×1) dramatically increased T cell priming in draining lymphoid tissues, leading to the regression of the primary tumor and distant metastases in a CD8+ T-cell-dependent fashion. Fractionated RT (5 Gy×4) induced initial tumor regression followed by relapse. Lee et al. also found that adjuvant chemotherapy suppressed the ablative RT-initiated antitumor response. However, the sub-optimal fractionated RT in combination with local immunotherapy using ad-LIGHT initiated an antitumor response and controlled metastasis development [42]. Ad-LIGHT is a TNF superfamily member, a stromal cell-expressed lymphotoxin-beta receptor ligand, and a T cell-expressed herpes viral entry mediator. These results suggested that while cytokine immunotherapy in combination with RT is a viable option, current systemic cancer treatment strategies may attenuate the antitumor effects of the immune system.

Table 3.

Prelinical studies that combine RT with cytokine/cytokine inducers

Tumor/cell line	Immune therapy	Radiation therapy	Highlights	References
B16 melanoma, 4T1 Breast	ad-LIGHT: TNF superfamily member	12 Gy×1 (with ad-LIGHT), 15 Gy×1, 20 Gy×1 or 20 Gy = 5 Gy×4, 25 Gy×1	Ablative RT (20 Gy×1)-initiated immune responses. Ad-LIGHT amplified RT-initiated immunity to eradicate disseminated metastases Tumor reduction was abrogated by fractionated RT (5 Gy×4) or adjuvant chemotherapy but amplified with ad-LIGHT Ablative RT dramatically increased T cell priming in DLT, leading to reduction/eradication of the primary tumor or distant metastasis in a CD8+ T-cell-dependent fashion	[42]
4T1 breast	Anti-TGF-beta, anti-PD-1 antibody	6 Gy×5	TGF-beta inhibition during RT is effective in generating CD8+ T cell responses to multiple tumor Ags in poorly immunogenic mice tumor Addition of anti-PD-1 antibodies extended survival achieved with RT and anti-TGF-beta Ab	[24]
B16 melanoma	IFN-gamma	15 Gy×1	Irradiation upregulated the expression of VCAM-1 on the vasculature of tumors grown in control but not in IFN-gamma-/-mice In addition to inducing molecular cues necessary for T cell infiltration, surface MHC class I expression was also upregulated in response to IFN-gamma produced after irradiation	[65]
B16 melanoma	OVA or gp51 peptide emulsified in CFA and injected into left flank and nape of neck of mice	15 Gy×1 or 5 Gy×3	Irradiated mice had cells with greater capability to present tumor antigens and specific T cells that secreted IFN-gamma upon peptide stimulation within tumor-draining lymph nodes than non-irradiated mice Immune activation in tumor-draining lymph nodes correlated with an increase in the number of CD45(+) cells infiltrating single dose irradiated tumors compared with non-irradiated mice Localized radiation can increase both the generation of antitumor immune effector cells and their trafficking to the tumor site	[27]
Colon 26	IL-2	2 Gy×10 (5 consecutive days, 2 cycles)	RT with intratumoral injection of IL-2 in mice bearing subcutaneous xenografts of colon adenocarcinoma cells abrogated primary tumors Liver metastasis formation was also completely inhibited by the combination of IL-2 and RT	[41]
Lung cancer (LLC; 3LL)	TLR9 agonist	20 Gy×1	RT with TLR9 agonist inhibited 3LL tumor growth in both control and B-cell-deficient mice Mice receiving combination therapy had fewer lung mets and higher survival than the single treatment cohort Tumor-specific humoral immune responses were found in wild-type mice, whereas tumor-infiltrating natural killer DCs were increased in B-cell-deficient mice following combined RT with TLR9 agonist treatment	[66]
Lung cancer LLC (OVA-expressing LLC-OVA cells)	TLR9 agonist/CpG-ODN	14 Gy×2 (24 h apart)	CpG-ODN as a ligand of TLR9 The tumor growth of mice treated with radiation and CpG+ OVA liposome was greatly inhibited, and ~60 % of mice treated were completely cured Combined therapy with radiation and CpG+ OVA liposome allowed the augmented induction of OVA tetramer(+) LLC-OVA-specific cytotoxic T lymphocyte (CTL) in DLN of tumor-bearing mice	[67]
Murine immune-genic fibrosarcoma	TLR9 agonist/CpG-ODN	2 Gy×10 (BID) or 20 Gy×1 for growth delay expts, 2–9 Gy×10 for TCD50 assay	A total dose of 83.1 Gy achieved tumor cure in 50 % of mice treated with RT alone Only 23 Gy achieved tumor cures in 50 % of mice treated with CpG-ODN Mice cured of their tumors by combined therapy were highly resistant to subcutaneous tumor take or development of tumor nodules when re-challenged with fibrosarcoma cells after treatment	[34]

CEA carcinoembryonic antigen, DLT draining lymphoid tissue, IFN interferon, IL-2 interleukin-2, Expts experiments, MHC major histocompatibility complex, OVA ovalbumin, TLR Toll-like receptor, VCAM vascular cell adhesion molecule

A recent study combined RT with immunotherapy to modulate cytokines and immune checkpoint inhibitors [24]. Vanpouille-Box et al. hypothesized that TGF-β activity hinders RT-induced TAA presentation; inhibition of TGF-β activity may lead to an in situ tumor vaccination effect. TGF-β inhibition with RT-associated fresolimumab was effective in generating CD8+ T cell responses to multiple TAAs even in poorly immunogenic mouse tumors (4T1 breast tumor cells) [24]. Both irradiated primary tumors and non-irradiated lung metastases regressed with enhanced CD8+ TILs cells, strongly suggesting that these cells directly caused tumor regression (aka abscopal effects). Furthermore, they observed that the addition of anti-PD-1 mAbs prolonged the survival achieved with RT and TGF-β blockade. This remarkable study employed all aforementioned strategies with RT to achieve the desired abscopal response. Investigators then designed a clinical trial to evaluate the application of this concept in metastatic breast cancer patients (NCT01401062) [43]. Another recent study investigated the antitumor effect of RT combined with IL-2 along with L19, which targets the extra domain of fibronectin, a marker of tumor neoangiogenesis [44]. Zegers et al. [44] found that L19-IL-2 and RT combination therapy had a synergistic effect in colon and lung cancer cells but not in breast cancer cells following injection into mice. A phase I clinical study is currently recruiting patients with oligometastatic solid tumors (NCT02086721) [43].

In a phase I study, patients with metastatic melanoma or RCC were treated with SAbR followed by IL-2 [29]. Eligible patients included those who had not previously received medical therapy for metastatic disease; they received one to three fractions of high-dose SAbR with the last dose administered 3 days before IL-2 treatment [29]. Eight of 12 patients (66.6 %) achieved either complete or partial responses to non-irradiated sites. While the number of subjects was too small to indicate the efficacy of this regimen, this response rate is still highly promising as compared to the reported response rate of 20 % with IL-2 monotherapy in metastatic RCC [38]. This regimen is currently being evaluated in two phase II trials in metastatic melanoma (NCT01884961) and mRCC (NCT01896271).

The influence of radiation dose and fractionation

External beam radiation therapy (EBRT) is a commonly used method to deliver conformal RT to target areas. Examples of EBRT include 2D and 3D conformal RT, IMRT, and SAbR. While definitions vary, conventional fractionation generally refers to radiation doses lower than 3 Gy per fraction. EBRT can be delivered as a standard/conventional fractionation of 1.8–2 Gy daily treatments, as a large dose with a lower number of fractions (hypofractionation), or as a smaller dose with a higher number of fractions (hyperfractionation). The most focused form of EBRT is SAbR which is usually reserved for extreme hypofractionation, often exceeding 6 Gy per fraction. Different doses are used because of the varying radiosensitivity of different tumor sites and surrounding organs at risk. For example, small cell carcinoma responds dramatically to small doses of radiation per fraction, although it is hard to deplete because of early metastasis formation. In contrast, melanoma and RCC are well known for their resistance to conventionally fractionated doses, but respond well to hypofractionation. A dose of 6 Gy×5 is a common and effective treatment modality for localized melanoma [45]. Therefore, treatment doses and fractionation are additional variables that need to be optimized for effective RT and immunotherapy combination regimens.

Schaue et al. [46] reported that a single dose of 7.5 Gy or above induced an increase in IFN-γ+ tumor-specific splenocytes in a mouse model of melanoma. A dose of 5 Gy did not exert the same effect. This observation is consistent with the report by Lee et al. [42] who found that ablative RT (20 Gy×1), but not the same dose delivered in 4 fractions (5 Gy×4) over 2 weeks, increased T cell priming in lymphoid tissues, leading to the eradication of the primary tumor or the formation of distant metastases in a CD8+ T-cell-dependent fashion. Furthermore, Camphausen et al. [9] reported tumor growth inhibition at distant sites following radiation with 10 Gy×5 fractions as compared to 2 Gy×12 fractions in mouse models of lung cancer and fibrosarcoma. These studies suggest a radiation dose dependency of the abscopal effect. The radiobiological mechanisms underlying the enhanced immune response elicited by higher RT doses (and possibly to partial tumor volumes) remain an active area of investigation.

Studies conducted by Wattenberg et al. [47] advocated that ablative RT doses greater than 8 Gy in a single fraction lead to changes in cytokine, chemokine, and chemoattractant profiles within the tumor microenvironment and induce an immunogenic tumor cell death. Filatenkov et al. [48] demonstrated a significant enhancement in response rate and survival following dose-escalation studies using single-fraction irradiation in mice bearing advanced CT26 tumors. Local irradiation with 15 Gy resulted in complete remission in only one of 14 mice. Despite an overall decrease in tumor growth, only two of 14 mice survived beyond 100 days after tumor irradiation. In contrast, three of five mice in the 20 Gy arm and 13 of 15 mice in the 30 Gy arm achieved complete remission and survival beyond 100 days. [48]. High RT doses transformed the immunosuppressive tumor microenvironment as shown by an increase in CD8+ TILs and a decrease in MDSCs. Similarly, Kulzer et al. [2] showed that the in vitro co-incubation of immature DCs with supernatants of irradiated human SW480 colorectal tumor cells resulted in DC maturation and activation. Both high doses (15 Gy×1 or 5 Gy×3) and conventionally fractionated RT (2 Gy×5) induced tumor cell death and generated supernatants that could activate immature DCs. Hence, the underlying mechanisms regulating the effects of high-dose RT, as opposed to fractionated RT, are still unclear. As discussed earlier, Dewan et al. [16] showed that fractionated (8 Gy×3 or 6 Gy×5), but not single-dose RT (20 Gy), combined with ipilimumab induced an immune-mediated abscopal effect in mice with either breast or colon cancer. Differences in tumor cell radiosensitivity may explain this difference, as indicated by different thresholds for efficient TAA production for each tumor type, leading to efficient immune cell activation.

Tsai et al. [49] reported significant differences in gene expression profiles in breast, prostate, and glioma tumor cells exposed to either single-dose (10 Gy) or fractionated (2 Gy×5) radiation. IFN-related genes such as signal transducers and activators of transcription factor 1 were induced by fractionated radiation in all three tumor cell lines. While the role of these genes in promoting T cell activation and radioresistance requires further investigation, Kumari et al. [50] demonstrated the upregulation of genes involved in T cell antitumor effector activity following 10 Gy radiation of tumors. Examples of these genes, ordinarily downregulated by tumor cells to evade immune recognition and response, include Fas (CD95), intracellular adhesion molecule-1 (ICAM-1), TAA, and MHC-I. Furthermore, investigators found that a single 10 Gy RT to colorectal cancer cells enhanced the expression of OX40 and 41BB ligands, previously identified as important co-stimulators of effector CTLs on tumor cells. Because T cells become anergic without proper co-stimulatory signals, this study showed that optimal radiation doses can reduce T cell anergy.

Preclinical studies have reported that high dose per fraction irradiation has more immunogenic effects than conventionally fractionated treatments, as discussed above. The majority of the reported preclinical studies showing the synergy of RT and immunotherapy used either single or a few fractions of high dose RT, although this is likely an effect of the technical difficulties of repeated animal irradiation. It stands to reason that the required steps to generate an adaptive immune response must occur after tumor irradiation. These steps include the recruitment of DCs to the irradiated tumor, phagocytosis of dying tumor cells by DCs and DC migration to the draining lymph nodes for antigen presentation and T cell priming. In this scenario, RT delivered in multiple fractions likely eradicates the recruited DCs and CTLs, and may be less immunogenic than either a single fraction or a few fractions given in close succession. Dewan et al. [16] reported that fractionated RT induced an abscopal effect more effectively when combined with ipilimumab anti-CTLA-4 mAb, although the fractionation regimen (8 Gy×3 or 6 Gy×5) was delivered in consecutive days consistent with a short course of RT. The depletion of tumor-specific CD8 T cell repertoire by long and fractionated RT may even lead to tumor tolerance. After reviewing the preclinical literature and the limited clinical data, the abscopal effects of RT and RT-IT synergy were most commonly observed at doses higher than 8 Gy per fraction. Despite this, the most immunogenic irradiation dose is still unknown because each cancer has distinct histological features, origins, and genetics. Hence, RT immunogenicity likely depends on cancer type, irradiated metastatic sites and species, radiosensitivity of cancer cells and the type of immunotherapy used in the combination regimen. Preclinical dose fractionation may not be readily applied to the clinical setting. Rigorous translational studies are needed to identify optimal RT doses and fractionation for each cancer and immunotherapy type.

Future directions

Conformal RT is an important modality for local cancer treatment and may induce immune stimulation leading to systemic antitumor responses. With the remarkable responses obtained using various immunotherapies combined with RT in preclinical settings, clinical applications of these strategies are being studied in a variety of cancers. The results from these trials may lead to the application of RT in cancer therapy solely for its immunogenic and antigen-presenting effects in conjunction with immunotherapy. While numerous preclinical studies have been reported on RT-IT combination regimens, additional clinical and translational studies are clearly needed to investigate optimal RT dosing, fractionation, timing, sequencing, and the interactions with various immunotherapies. For effective clinical application, the immunogenic properties of RT will have to be carefully balanced with countering its immunosuppressive effects (i.e., increased PD-L1 expression). The strategic combination of RT doses, fractionation, and sequencing and the careful selection of immunotherapy are required for therapeutic success. Hypofractionation may be the ideal choice to obtain optimal immunogenic effects of RT with either single or limited fractions administered closely. SAbR can deliver high doses focally and safely at any metastatic site and spares the draining lymph nodes which are an integral part of an immune response. The combination of immunotherapy and SAbR, often referred to as the i-SAbR regimen, is a promising and exciting new paradigm to further improve anticancer immunotherapy.

Abbreviations

2D

Two dimensional

3D

Three dimensional

CD

Cluster of differentiation

CTL

Cytotoxic T lymphocyte

CTLA-4

Cytotoxic T-lymphocyte-associated protein 4

DC

Dendritic cell

EBRT

External beam radiation therapy

Gy

Gray

HR

Hazard ratio

ICAM

Intracellular adhesion molecule

IFN

Interferon

IGRT

Image-guided radiation therapy

IL-2

Interleukin-2

IMRT

Intensity-modulated radiation therapy

mRCC

Metastatic renal cell carcinoma

OS

Overall survival

PD-1

Programmed death 1

PD-L1

Programmed death-ligand 1

PSA

Prostate-specific antigen

RCC

Renal cell carcinoma

RT

Radiation therapy

SAbR

Stereotactic ablative radiation

TAA

Tumor-associated antigen

TCR

T cell receptor

TGF

Transforming growth factor

TIL

Tumor-infiltrating lymphocyte

Treg

Regulatory T cell

Compliance with ethical standards

Conflict of interest

All authors declare that they have no conflict of interest.

Ethical statements

This article does not contain any studies with human participants or animals performed by any of the authors.