Radiation dose, schedule, and novel systemic targets for radio-immunotherapy combinations

Abstract

Immunotherapy combinations are being investigated to expand the benefit of immune checkpoint blockade across many cancer types. Radiation combinations, in particular using stereotactic body radiotherapy, are of keen interest because of underlying mechanistic rationale, safety, and availability as a standard of care in certain cancers. In addition to direct tumor cytotoxicity, radiation therapy has immunomodulatory effects such as induction of immunogenic cell death, enhancement of antigen presentation, and expansion of the T-cell receptor repertoire as well as recruitment and increased activity of tumor-specific effector CD8⁺ cells. Combinations of radiation with cytokines and/or chemokines and anti-programmed death 1 and anticytotoxic T-lymphocyte antigen 4 therapies have demonstrated safety and feasibility, as well as the potential to improve long-term outcomes and possibly induce out of irradiated field or abscopal responses. Novel immunoradiotherapy combinations represent a promising therapeutic approach to overcome radioresistance and further enhance systemic immunotherapy. Potential benefits include reversing CD8+ T-cell exhaustion, inhibiting myeloid-derived suppressor cells, and reversing M2 macrophage polarization as well as decreasing levels of colony-stimulating factor-1 and transforming growth factor-β. Here, we discuss current data and mechanistic rationale for combining novel immunotherapy agents with radiation therapy.

Immune checkpoint blockade (ICB), with anti-programmed death 1 (PD1) receptor or anti-PD ligand 1 (PD-L1) and anti-cytotoxic T-lymphocyte antigen 4 (CTLA4) antibodies, have improved clinical outcomes across several cancer types (1). However, many patients still do not derive clinical benefit and eventually have disease progression. Thus, improving the efficacy of ICB is an ongoing unmet clinical need.

One promising strategy involves combining radiation therapy (RT) with immune-oncology (IO) agents to improve responses, potentially in a synergistic manner. This strategy is particularly attractive as RT is standard of care in many malignancies, allowing for a greater ease of study in clinical trials (2-5). Indeed, prior to the widespread use of ICB, clinical trials combining cytokines such as interleukin-2 (IL-2) and chemokines such as granulocyte-macrophage colony-stimulating factor (GMCSF) with high-dose RT suggested feasibility for IO-RT approaches, with preliminary evidence for improved treatment outcomes (6-9). In this article we discuss characteristics of RT such as dose, schedule, and delivery mode in relation to its immunomodulatory effects, review safety, and efficacy results of existing IO-RT clinical trials and propose novel IO-RT combinations based on mechanistic rationale of IO to overcome RT-induced immunosuppression and further enhance immunostimulant effects.

Radiation characteristics and potential impact on immunotherapy efficacy

In the majority of disease settings, the current standard, curative RT treatment involves small, fractionated doses of RT (generally 1.8-2.2 Gy) given daily to cumulative doses of 50-70 Gy depending on clinical context (Table 1). This schedule is used because of the necessity of treating both the primary tumor and commonly at-risk draining lymphatics while limiting normal tissues toxicity. This approach, although leading to cure in many patients, may be disadvantageous when used in immunotherapy combinations because of a potential immunosuppressive effect (10-12). Indeed, recent data suggest that elective nodal RT may be antagonistic to immunotherapy, at least in preclinical models (13).

Table 1.

Radiation characteristics and potential impact on immunotherapy efficacy^a

Radiation type	Description	Immunomodulatory effect
Conventional	Small, fractionated doses of RT (approximately 1.8-2.2 Gy) Cumulative doses of 50-70 Gy	Potential immunosuppressive Elective nodal irradiation→decreased immune activity
Hypofractionated SBRT SRS SABR	High dose, less fractions Eg, 30-50 Gy in 3-5 fractions	Potential immuno-enhancement Increase in IFN-γ–associated gene signature T-cell activation
Proton radiation treatment	High linear energy transfer Bragg peak phenomena	Increase in calreticulin Less lymphopenia
Carbon ion radiotherapy	High relative biologic effectiveness	Immunogenic cell death
SFRT-GRID SBRT-PATHY	Partial radiation Targeting tumor hypoxic segment	Induction of TNA-α and TRAIL Abscopal response in preclinical reports Targeting tumor hypoxia
FLASH-RT	RT delivery in milliseconds	Less lymphopenia Increase in CTL infiltration to tumor

^aCTL = cytotoxic T-lymphocyte infiltration; IFN = interferon; RT = radiation therapy; SABR = stereotactic ablative radiation therapy; SBRT = stereotactic body radiation therapy; SBRT-PATHY = stereotactic body radiation therapy for Partial tumor irradiation of unresectable bulky tumors targeting exclusively their hypoxic segment; SFRT = spatially fractionated radiation therapy (also known as GRID therapy); SRS = stereotactic radiosurgery; TNA = tumor necrosis factor; TRAIL = TNF-related apoptosis-inducing ligand.

Conversely, approaches treating small volumes to high doses over few treatments—referred to by several names such as hypofractionated RT, stereotactic body RT (SBRT), stereotactic ablative RT (SABR), or stereotactic radiosurgery—are likely associated with immune-stimulatory effects. For example, in murine preclinical models, high-dose RT has been shown to increase tumor-associated macrophages (TAM) and CD8+ T cells and enhance expression of T-cell activation and interferon (IFN)-γ–associated gene expression (14-16).

These distinct and opposed outcomes following the combination of RT and immunotherapy based on the type of RT given (eg, conventional vs hypofractionated RT) complicate the inclusion of immunotherapy into the curative management of many solid tumors. Indeed, at the time of this publication, the only randomized clinical trial of combined conventionally fractionated chemoRT and immunotherapy that improved survival is the PACIFIC trial in non-small cell lung cancer (NSCLC), which held immunotherapy until the completion of RT (17-19). This approach would seem at least to partially avoid the immunosuppressive effects of conventionally fractionated RT while reaping some of the immune-stimulatory effects of tumor cell death due to RT. Similar trials that gave immunotherapy concurrent with RT have had no beneficial effects on outcome (20-22). Ongoing trials will determine the short- and intermediate-term sequencing of current definitive conventionally fractionated RT and immunotherapy for clinical use.

In the metastatic setting, where RT may be an effective immune adjunct as opposed to a curative therapy, utilizing hypofractionated regimens is likely a better approach. In addition to its immune-stimulatory effects, hypofractionated RT has the added benefit of potentially ablating 1 or more larger tumors, which are typically more resistant to immune-checkpoint inhibitors (ICI) with a minimal added toxicity (23). Increased tumor bulk has been associated with poor response to anti-PD1 (24-26), whereas the tumor microenvironment in large tumors is associated with increased infiltration of myeloid lineage cells, systemic immunosuppressive cytokines, and regulatory T-cell (Treg) infiltration (27-29).

There are several aspects of RT delivery that impact the interaction between RT and IO beyond dose, schedule, and combination with specific IO agents. These include heavy particle therapy such as proton, carbon ion, and π meson and spatially fractionated radiotherapy (SFRT) such as SFRT (GRID therapy) and SBRT PArtial Tumor irRT of unresectable bulky tumors targeting exclusively their HYpoxic segment (SBRT-PATHY), as well as an emerging paradigm of understanding the effect of dose rates with FLASH radiotherapy.

Photon-based RT treatment utilizes x-ray and γ-rays that deposit energy to tissue that causes single- and double-stranded DNA breaks to damage tumor cell DNA. Should enough DNA strand breaks occur in key regions of the genome, RT damage ultimately results in mitotic catastrophe and tumor cell death. The deposition of this energy is characterized by linear energy transfer (LET) or the energy transfer per unit length of the track of ionizing RT. Heavier ionizing particles have a greater chance to interact and transfer energy with the surrounding tissue. These heavier ionizing particles such as protons, carbon ion, and π meson are considered densely ionizing RT with a high LET. Lighter particles such x-ray and γ-rays classically are low LET and are considered sparsely ionizing RT. To properly compare the biological effects of each form of ionizing RT, we use the relative biologic effectiveness (RBE), defined as the ratio of a normalized dose of x-ray RT to a test dose (proton, carbon ion, etc) thus allowing one-to-one comparisons of the effects of different forms of RT treatment. Larger ion particles tend to have a higher RBE compared with traditional RT treatment. RBE is affected by LET, dose, fraction, dose rate, and biological system (30,31).

Proton RT treatment has unique properties that render it intriguing for pairing with IO. It has a higher LET and slightly higher RBE of 1.1 compared to photon treatment. It has unique properties in which the majority of its ionizing RT is given off at a specific location known as the Bragg peak, after which energy deposition promptly drops off and RT dose to deeper tissue is largely spared. One of the main reasons for excitement for proton therapy was the utilization of this Bragg peak phenomena to limit dose deposition to normal tissue. There are some reports of surprising results of increased tissue toxicity.

Some preclinical evidence suggests a difference in affecting tumors and tumor microenvironment with proton therapy. In vitro experiments suggested higher cell surface levels for the damage-associated molecular pattern (DAMP) molecule calreticulin, which might lead to increased immunogenicity and recognition by CTLs (32,33). Additionally, the lower dose bath of RT from proton therapy decreases the time and/or level of exposure of lymphocytes to RT ultimately decreasing the level of lymphopenia (34,35). Other studies simply suggest in vitro that proton therapy can produce unique antitumor responses, albeit this study did not directly compare proton with photon effects (36).

Carbon ion radiotherapy is another form of particle RT treatment that is not widely used today given the current expense and extremely large size facility needed to operate this. Carbon ions operate under the similar principles for proton therapy, however, the RBE is even higher than protons likely because of more dose deposition due to particle fragmentation into the surrounding tissue (37). Interestingly, some preclinical models with carbon ion radiotherapy suggest an enhancement of immune responsiveness to combinatorial RT and immunotherapy via enhanced immunologic cell death and increased tumor immunogenicity (38).

Pion therapy (negatively charged π mesons) is another modality that was previously explored because of its Bragg peak and increased effectiveness in hypoxic environments (39). Pions are heavy ions that are produced in cyclotrons or linear accelerators using protons to hit a beryllium target. Pion therapy Bragg peak is higher than the other heavy particles because of its negative charge being attracted to nucleus in materials resulting ultimately into nuclear fragmentation, alpha particle production, and protons and neutrons, which all interact with tissue. Pion therapy is not widely used because of cost and availability as well as the loss of conformality to tumor targets because of increased lateral scatter and worse penumbra due to the other charged particles it produces. Two smaller clinical trials compared overall survival (OS) in prostate and high-grade astrocytoma, however, pion therapy improves OS compared with conventional photon therapy (40,41),

The role of partial-tumor RT is being currently explored and may hold specific promise for RT-IO combinations. Two techniques are currently being used: SFRT, also known as GRID therapy, and SBRT-PATHY. The SFRT principle is to deliver large, heterogenous doses to large, bulky tumors utilizing a grid of pencil beams dose distribution to the tumor. The premise of GRID therapy relies on sparing some portions of the overlying skin to reduce the risk of skin necrosis. Originally, a metallic block with small beam holes was used to create this treatment pattern, but with modern techniques, this can be recreated with standard multileaf collimator blocks in contemporary radiotherapy machines. GRID traditionally is used for palliative purposes typically with photons, although proton and carbon ion therapy have been used as well. The excitement with SFRT for RT-IO combinations is that it has been shown to induce bystander effect to kill tumor cells in the lower dose region adjacent to the high-dose pencil beam areas, which is not typically seen in conventional fractionated treatment. This bystander effect might be mediated by induction of cytokines such as tumor necrosis factor alpha (TNA-α) and TNF-related apoptosis-inducing ligand (TRAIL), destruction of tumor microvasculature endothelium by induction of cermamide, or induction of immune response leading to abscopal effects, seen mainly in preclinical animal models in nonirradiated tumors (42). SBRT-PATHY takes the inhomogeneity concept and applies it to certain areas of the tumor microenvironment. SBRT-PATHY specifically treats only part of a large, bulky unresectable tumor and targets hypoxic areas of the tumor microenvironment. These hypoxic cores of tumors tend to be more radioresistant to conventional RT treatment, thus SBRT-PATHY is an attractive treatment that may trigger bystander and abscopal effects on tumors (42).

Another exciting technology or technique that may lend itself useful to the combinatorial IO-RT treatment paradigm is FLASH radiotherapy (FLASH-RT). FLASH-RT delivers a standard dose of RT treatment in a matter of milliseconds, a dramatic reduction from the typical length measured in minutes. In the 1960s, FLASH-RT showed increased normal cell survival in vivo when compared with more typical lower dose rates (43). A few interesting theories exist as to why normal tissue seems to be spared. The first is that FLASH-RT might cause the immediate consumption of oxygen in the environment, and thus normal tissue becomes hypoxic, and RT damage is most optimal in an oxygenated environment, therefore normal tissue is not as easily damaged. Tumors that are already hypoxic are not affected by this sudden change of oxygenation. The second theory that is of greater interest here is that by delivering dose so rapidly, circulating lymphocytes are no longer irradiated, and participants do not become lymphopenic. Additionally, FLASH may cause improved CTL infiltration into the tumor microenvironment, less chronic neuroinflammation, and differential expression of cytokines (43). Most of our knowledge with FLASH comes from preclinical animal models; however, in the past few years, the first patient, who had multiply recurrent CTCLs, was treated with FLASH therapy (44). Results were excellent with limited skin toxicity. It was noted that there was an increase in edema, which is speculated to be due to a bigger immune-mediated response.

In summary, the impact of these techniques or modalities on RT-IO interactions beyond conventional photon-based RT approaches is the subject of ongoing investigations, and despite great enthusiasm, definitive conclusions on the contributions of specific RT parameters remain largely uncertain. Only systematic and rigorous analysis of the different aspects of RT will help investigators identify the optimal RT approaches to pair with IO.

Immunomodulatory effects of RT

Immunostimulant effects of RT

Beyond direct cytocidal effect in large—and potentially ICI-resistant—tumors, hypofractionated RT has immunomodulatory effects including but not limited to 1) induction of immunogenic cell death (ICD), 2) enhancement of antigen presentation by dendritic cells (DCs), 3) expansion of the T-cell receptor (TCR) repertoire, and 4) recruitment and increased activity of tumor-specific effector CD8⁺ cells (45-48). The process of ICD includes a series of steps that can induce antitumor immunity and collectively shape RT-induced responses (49,50). After RT, ICD includes the release of adenosine triphosphate and high mobility group box 1 protein, which binds to toll-like receptor (TLR) 4 on DCs, and translocation of calreticulin to the cell surface. Calreticulin then serves as a potent prophagocytic signal (51,52). Potential immunostimulant effects of RT are summarized in Figure 1.

Figure 1.

Immunostimulant and immunosuppressant effects of radiation therapy (RT). The figure illustrates the mechanisms underlying immunostimulant effects of radiation therapy, which include enhancement of tumor-associated antigen (TAA) shedding, immunogenic cell death, increase in TAA presentation to antigen-presenting cells, major histocompatibility complex (MHC) class I expression, increase in expression of NKG2D ligand, upregulation of adhesion molecules. The mechanisms underlying immunosuppressant effects include TGF-beta signaling activation, M2 macrophage polarization, increase in infiltration of myeloid-derived suppressor cell infiltration, activation of colony-stimulating factor-1 (CSF1) pathway, hypoxia inducible factor 1 (HIF1), T-regulatory cell (Treg) recruitment, programmed-death ligand-1 (PD-L1) upregulation, and vascular endothelial growth factor (VEGF) upregulation.

Because of its immunomodulatory effects, hypofractionated RT may induce responses in nonirradiated lesions, a phenomenon known as the abscopal effect. Initially observed preclinically as early as 1953, isolated case reports of this abscopal effect pepper the RT literature but are rarely observed in clinical practice as their existence has been questioned (53).

Favorable initial data for abscopal effects were observed in initial studies combining RT with GM-CSF (8), however, follow-up trials have shown mixed results for abscopal or distant effects following the combination of hypofractionated RT with ICI, underscoring the lack of understanding of this phenomenon (54-57). Elucidating any distant immunomodulatory effects of hypofractionated RT is even more critical in light of several trials showing the survival benefit of consolidation hypofractionated RT to multiple oligometastatic sites (48,58-60).

Immunosuppressant effects of RT

In addition to the potentially immunosuppressant effects of RT targeting lymphocyte-producing organs, and circulating lymphocytes, intratumoral immunosuppression by RT has been described as well, particularly following conventionally fractionated RT. These effects can include 1) promoting MDSCs infiltration (61), 2) enhancing CSF1 gene expression and CSF1 levels in tumor (62), 3) increasing tissue growth factor (TGF)-β (63) and Tregs (64), 4) upregulating hypoxia-inducible factors (HIF) and vascular endothelial growth factor (VEGF) (65), upregulating PD-L1 expression (66,67), and M2 macrophage polarization (Figure 1) (68). Targeting each of those pathways can be promising to improve response to RT and expand the role of RT as a therapeutic approach in patients with metastatic cancer. Here we will discuss current data and mechanistic rationale of combining novel immunotherapy agents with RT.

Optimal approach to combine immunotherapy and RT

Many unanswered questions remain regarding the optimal approach to combine IO with RT (Table 2). Firstly, the optimal timing and sequencing of RT and immunotherapy is unknown. It is established that responses following the combination of PD-L1 monoclonal antibody (mAb) and hypofractionated RT vary dramatically in preclinical models based on the timing of either treatment in relation to the other (66,69,70). This importance is underscored, albeit in the conventionally fractionated setting, based on the aforementioned positive clinical trial for ICI consolidation after chemoRT vs several negative concurrent RT-ICI trials (17-20). With appropriate caveats, these data provide tantalizing clinical insight highlighting the potential importance of RT and ICI sequencing.

Table 2.

Optimal approach to combine immunotherapy and radiation therapy^a

Key question	Approach and considerations	Supportive findings
Sequencing immunotherapy with RT	Concurrent vs sequential	Sequential Positive PACIFIC trial CRT→durvalumab Negative JAVELIN HN100 CRT with avelumab
Dose and fractionation of RT	Single vs multiple fractions	Fractionated RT Antitumor effect with 8 Gyx3 combined with anti-CTLA4 Ab in MCA38 model Elevated IFN-γ with fractionated RT Anatomically driven fractionated RT→increase in IFN-linked gene expression despite induction of TREX1
Number of lesions for RT	Single vs multisite RT	Multisite RT Irradiated tumor response is associated with OS To alter local immunosuppressive TME of each lesion
Irradiated tumor location	Clinically relevant and/or large tumor vs targeting and/or not targeting specific location	None to TDLN RT to TDLN has immunosuppressive effects through decrease in tumor-specific T cells Clinical data are lacking to prove this approach is beneficial

^aAb = antibody; CRT = chemoradiation therapy; CTLA4 = cytotoxic T-lymphocyte antigen 4; HN = head and neck; IFN = interferon; OS = overall survival; RT = radiation therapy; TDLN = tumor-draining lymph node; TME = tumor microenvironment; TREX1 = Three prime repair exonuclease 1.

Secondly, the optimal dose and fractionation of RT as an ICI adjunct is worthy of further study. At least one study showed that a fractionated regimen (8 Gy x 3 fractions [fx]) vs a single fraction of 30 Gy resulted in upregulation of type I IFN genes mediated by stimulator of IFN genes (STING) pathway activation and induced abscopal responses in TSA breast and MCA38 colon mice models (71,72). Single RT dose above 12 Gy induced activation of 3 prime repair exonuclease 1 (TREX1), which decreased cytosolic DNA level aborting activation of STING pathway, nominating TREX1 as a key playing factor in RT effect (73). However, regimens such as that pursued in NRG-BR001 recommended anatomically driven doses of 30-50 Gy in 3-5 fractions resulted in induction of IFN-linked gene expression, exemplified by Deoxyribonuclease I (DNASE1), despite induction of DNASEIII (TREX1) (57,74).

Thirdly, the optimal number of tumor lesions to target via RT is unclear. As mentioned above, targeting multiple oligometastatic sites with RT has been associated with improved survival in several small trials. However, the majority of IO-RT trials performed to date have pursued single-site irradiation approaches with the goal of inducing abscopal responses (75-77). Moreover, in at least 1 trial, OS of patients receiving RT-IO was correlated with irradiated tumor response (74). This highlights the potential importance of multisite irradiation, as targeting several tumor lesions may further alter the local tumor immunosuppressive microenvironment.

Fourthly, the unique role of irradiated tumor location in improving IO-RT combination efficacy is an area of investigation. In most RT-IO trials, when choosing metastatic sites to irradiate, preference has been given to the most clinically relevant and/or larger lesions, although there are reports about the differential immunogenic potential of irradiated sites in enhancing peripheral CD8+ T-cell response (76). Another consideration for metastatic site targeting is the sparing of lymphocyte-producing organs to prevent RT-induced lymphopenia (78). The irradiation to tumor-draining lymph nodes (TDLN) may create local immunosuppressive microenvironment with the effect being dependent on RT dose (13,79,80,81). In the MC38 murine colon carcinoma model, elective irradiation to TDLN downregulated CXCR3-CXCR5–associated T-cell chemoattractant chemokines resulting in decreased effector CD8 T cells in lymph node and reduced efficacy of IO-RT (80). RT to TDLN reduced abscopal effect and decreased count of stem-like and total CD8 T cells in B16 murine melanoma model (81). Elective nodal irradiation led to a decrease in antigen-specific T cells in tumor microenvironment of Head and Neck (HN) cancer model and resulted in low numbers of circulating T lymphocytes supporting further the strategy to spear lymph node (LN) irradiation (82). However, prospective clinical data regarding beneficial effect of TDLN-sparing RT are lacking.

Prior clinical results of RT with anti-PD1 and/or anti-CTLA4 therapy

There is a strong scientific rationale to combine RT with ICI therapy not only to enhance the therapeutic effect of irradiation and immunotherapy alone but also to overcome the resistance to each therapy individually. Increased PD-L1 expression and T-cell exhaustion are established resistance mechanisms to RT, which might be overcome by combining RT with anti-PD1 or anti–PD-L1 therapies (83). In human papilloma virus (HPV)–negative head and neck squamous cell cancer (HNSCC), upregulation of PD-L1 was associated with Axl-PI3K/Akt expression and led to locoregional failure for patients after RT (84).

Safety results of RT with anti-PD1 and/or anti-CTLA4 therapy

Despite conceptual promise, there were early clinical concerns surrounding the potential safety of RT combined with ICI given the possibility of overlapping toxicities. The safety profile may vary between using combination of anti-PD1 and/or L1 vs anti-CTLA4 and/or anti-PD1/L1 plus anti-CTLA4 with PD1/-L1 targeting appearing to be associated with fewer adverse events. The central nervous system and the possibility of treatment-associated brain necrosis (TABN) and/or radionecrosis is a particularly important anatomic consideration for IO-RT combination toxicity. The incidence of TABN is reported to be approximately 15%-30% with studies reporting association of IO with hypofractionated RT with both radiological and clinical TABN (85-88). The association was especially strong in patients with melanoma (85). The incidence of TABN observed in 2 single-institutional studies was 13.5%-15% in melanoma patients. In contrast, another group reported the association between immunotherapy and TABN when combined with stereotactic radiosurgery in patients with melanoma (hazard ratio [HR] = 4.02, 95% confidence interval [CI]= 1.17 to 13.82; P = .03) (85).

Anti-CTLA4

In a retrospective, single-center study of patients with metastatic melanoma, the combination of ipilimumab administered before fractionated RT resulted in 18% immune-related adverse events, which required holding or delaying ipilimumab administration (89). In patients with metastatic castrate-resistant prostate cancer, bone-directed RT at the dose of 8 Gy in 1 fraction followed by ipilimumab resulted in high-grade AEs in 26% of patients (75).

Anti-PD1 and/or L1

In a safety analysis of combining anti-PD1 therapy with RT, one study reported a drug limiting toxicity (DLT) rate of less than 10% with using a multisite hypofractionated RT approach (57), whereas combination of single-site hypofractionated RT and pembrolizumab resulted in no DLTs (90). In the former study, all patients with toxicities had at least 2 irradiated lesions. The use of concurrent avelumab with chemoRT resulted in G3 and higher treatment-related AEs in approximately one-third of patients, which was similar to chemoRT alone (20). Finally, the most robust analysis to date to address this issue was the recent pooled analysis of patients enrolled in prospective clinical trials of ICI (anti-PD1, anti–PD-L1, CTLA4) with RT within 90 days of initiation of ICI. The study revealed no increased rate of G3 and higher treatment-related AEs in comparison with those who received immunotherapy alone (91).

Concurrent vs sequential

Another important consideration for safety evaluation is whether IO was given concurrently with hypofractionated RT or sequentially after RT. The randomized phase I trial of Concurrent or Sequential Ipilimumab, Nivolumab and Stereotactic Body Radiotherapy in Patients with Stage IV NSCLC Study (COSINR) study combined anti-PD1 with anti-CTLA4 therapy with hypofractionated RT randomized to concurrent and sequential multisite SBRT. In the sequential arm, the combination of all 3 anti-CTLA4 with anti-PD1 and RT resulted in increased risk of pneumonitis and prompted dose de-escalation of hypofractionated RT in patients with NSCLC during central lung radiation (92). There were no DLTs observed in the concurrent arm supporting the concept of safe initiation of IO during RT. The US Food and Drug Administration database pooled analysis of clinical trials combining ICB with RT revealed that ICB therapy within 90 days of RT was not associated with an increased risk of serious adverse events.

In summary, anti-PD1 and/or anti-CTLA4 and RT combination approach was found to be generally safe in terms of immune-related adverse events with only minor concern for increased incidence of pneumonitis with only anti-PD1 with anti-CTLA4 and RT combination and radionecrosis especially among patients with metastatic melanoma (86,92).

Efficacy results of RT with anti-PD1 and/or anti-CTLA4 therapy

The anti-CTLA4 and/or anti-PD1 or L1 plus RT trials provided informative data about sequencing of IO and RT, preferential anatomic site of RT, and IO-RT–induced immune profile changes.

Sequencing IO and RT

The efficacy of sequential vs concurrent IO-RT has not been formally evaluated and compared in large clinical trials, and it remains unknown whether one is more beneficial over another. It is noteworthy to mention that individual clinical trial data may support sequential over concurrent IO-RT in terms of efficacy endpoint. In patients with metastatic liver and lung (excluding melanoma) advanced tumors, therapy with ipilimumab and sequential or concurrent SBRT (total 50 Gy in 4 fx or total 60 Gy in 10 fx) resulted in a clinical benefit rate of 26% and reached to 42% in patients who received sequential RT and irradiation to lung lesions. There was a trend toward improved responses while using sequential as opposed to concurrent RT as well as radiating lung as opposed to liver lesions (55). ECOG-ACRIN is an ongoing randomized phase III trial that aims to evaluate concomitant vs sequential durvalumab with chemoRT for unresectable stage III NSCLC (93). Concurrent avelumab with chemoRT vs chemoRT alone failed to improve PFS in patients with locally advanced HNSCC cancers. In fact, the hazard ratio was equal to 1.21, raising the question about proper combination strategy of IO and chemoRT and the hypothesis that concurrent design may have impacted the beneficial effect of adding avelumab (20). In the cohort of patients with intermediate risk HPV-positive and locally advanced HPV-negative HNSCC, sequential pembrolizumab with cisplatin and RT resulted in numerically higher 1- and 2-year progression-free survival (PFS) in comparison with the concurrent approach (94). Pooled analysis of patients with NSCLC treated with pembrolizumab and RT revealed statistically significant improvement in abscopal response rate (odds ratio [OR] = 2.96; P = .0039), median PFS (HR = 0.67; P = .045), and OS (HR = 0.67; P = .0004) in combination of pembrolizumab with RT vs RT alone (95).

IO-RT–induced immune profile changes

A growing body of evidence supports IO-RT combinations as inducing improved immune responses, and several of those changes may be linked to clinical benefit. In patients with metastatic castrate-resistant prostate cancer, ipilimumab with RT (8 Gy x 1) resulted in increased peripheral blood CD8+ T cells, and high ratio of CD8+ to CD4+ T-cell ratio was associated with improved clinical benefit (75). In patients with chemotherapy-resistant NSCLC, concurrent RT (total 30 Gy in 5 fx or total 27 Gy in 3 fx) with ipilimumab resulted in expansion of TCR clones. In addition, increases in IFN-β were associated with response to therapy, whereas baseline tumor PD-L1 expression, CD8+ T-cell infiltration, and increased proliferation of peripheral blood CD4⁺CD8⁺ cells were not (96). Pembrolizumab combined with multisite SBRT (total 30-50 Gy in 3-5 fx) resulted in increased expression of IFN-γ–linked genes as well as a decrease in those associated with the cell cycle and DNA repair (57). Single-site RT regimens (total 24 Gy in 3 fx or 17 Gyx1 fx) with pembrolizumab resulted in increase in Ki67⁺ PD1⁺ CTLA⁺CD8 T cells in peripheral blood mononuclear cells (PBMCs) (90,97). Concurrent RT (total 30 Gy in 10 fx or total 27 Gy in 3 fx) resulted in an increase in peripheral blood TCR diversity in patients with advanced melanoma (98).

Does IO-RT improve response rates and/or PFS/OS over IO or RT alone?

Even though above-mentioned trials demonstrated evidence of enhanced immune responses after IO-RT, these changes have not translated into consistent improvement in efficacy. In a phase III trial of patients with metastatic castrate-resistant prostate cancer who progressed on docetaxel, the addition of ipilimumab to RT (8 Gyx1) did not improve OS in comparison to placebo (75). In a phase II randomized trial of patients with advanced adenoid cystic carcinoma, the addition of RT (total 30 Gy in 5 fx) to pembrolizumab did not improve PFS in comparison with pembrolizumab alone (99). In a phase II trial of patients with metastatic renal cell carcinoma (RCC) who progressed on prior anti-angiogenic therapies combination of nivolumab with RT (total 30 Gy in 3 fx) resulted in an objective response rate (ORR) of 17% and did not meet the prespecified ORR of 40% (100,101). Concurrent RT (total 30 Gy in 10 fx or total 27 Gy in 3 fx) with ipilimumab and nivolumab did not demonstrate improved efficacy over immunotherapy alone in patients with advanced melanoma who received prior immunotherapy and/or targeted therapy (98). An ongoing large, randomized phase II trial aims to evaluate the efficacy of ipilimumab, nivolumab, and RT in melanoma patients with brain metastases (102). The combination of ipilimumab-nivolumab with SBRT (total 45-50 Gy in 3-5 fx) demonstrated ORR of 45% in treatment-naïve NSCLC patients, which is numerically higher than ipilimumab-nivolumab ORR of 33% in metastatic NSCLC patients unselected for PD-L1 status (92,103). Noteworthy, in patients with microsatellite stable colorectal and pancreatic cancers who progressed on first-line therapy, the combination of ipilimumab-nivolumab and RT (total 24 Gy in 3 fx) resulted in an encouraging disease control rate of 25% and 20% in colorectal and pancreatic groups, respectively (104). Table 3 further summarizes randomized trials exploring these combinations.

Table 3.

Clinical trial results of RT with anti-PD1 (PD-L1) and/or anti-CTLA4 therapy^a

Study	Immunotherapy	RT	Study population	Key finding(s)
Qin et al. (89)	Ipilimumab (Ipi)	Median Ablative-total 20 Gy in 1 fraction Non-ablative-total 35 Gy in 11 fractions	Advanced melanoma	Ipi before RT vs Ipi after RT resulted in longer duration of irradiated tumor response.
Kwon et al. (75)	Ipilimumab	Total 8 Gy in 1 fraction	mCRPC with bone Mets	Ipi vs placebo after bone directed RT did not improve OS.
Welsh et al. (55), Tang et al. (76)	Ipilimumab	Total 50 Gy divided in 4 fractions or total 60 Gy in 10 fractions	Advanced solid tumors with liver and/or lung Mets	26% of clinical benefit rate of nonirradiated lesions Sequential vs concurrent (28% vs 20%; P = .25) Lung vs liver (31% vs 14%; P = .06) Liver>lung RT→T cell activation
Formenti et al. (96)	Ipilimumab	Total 30 Gy in 5 fractions or total 27 Gy in 3 fractions	Chemo refractory mNSCLC	Disease control rate of 31% Increased peripheral blood IFN-β and expansion of T cell clones were predictors for response
Luke et al. (57,74)	Pembrolizumab	Total 30-50 Gy in 3-5 fractions	Advanced solid tumors	Multisite SBRT→DLT < 10% ORR 13% SBRT→ increased expression of IFN-g–linked genes Irradiated tumor response was a/w improved OS
Mahmood et al. (99)	Pembrolizumab	Total 30 Gy in 5 fractions	mACC	No objective response outside of radiation field Local irradiated tumor response was a/w PD-L1 expression and MYB/NFIB translocation
Maity et al. (90)	Pembrolizumab	Total 24 Gy in 3 fractions Total 17 Gy in 1 fraction	Cohort 1-anti-PD1 refractory melanoma and NSCLC Cohort 2-advanced solid tumors	No DLTs Cohort 1 ORR 16.7% Cohort 2 ORR 8% RT→ increase in Ki67+ PD1+CTLA+CD8 T cells in PBMCs
Lee et al. (20)	Avelumab	Total 70 Gy in 35 fractions	Locally advanced SCHNC	No improvement in mPFS HR = 1.21 for PFS
Masini et al. (100,101)	Nivolumab	Total 50 Gy in 5 fractions	mRCC	ORR 17% DCR 55% mPFS 5.6 months NO benefit with Nivo with RT Disease control rate of irradiated tumor lesions was 82%
Postow et al. (98)	Ipilimumab Nivolumab	Total 30 Gy in 10 fractions or total 27 Gy in 3 fractions	Advanced melanoma	G ≥ 3 AEs 40% RT→increase in peripheral blood T cell receptor diversity
Hammers et al. (97)	Ipilimumab Nivolumab	50 Gy in 5 fractions	mRCC	40% required systemic steroids ORR 56%
Patel et al. (105)	Ipilimumab Nivolumab	Total 40-50 Gy in 3-5 fractions	Treatment naïve mNSCLC	DLTs 17% DCR 74% PD-L1 did not affect PFS and OS
Parikh et al. (104)	Ipilimumab Nivolumab	Total 24 Gy in 3 fractions	MSS CRC and PDAC	DCR 25% and 20%
Kim et al. (106)	Ipilimumab Nivolumab	Total 24 Gy in 2 fractions	Merkel cell carcinoma	ORR 100% in ICB-naïve patients ORR 31% with ICB-treated patients No significant difference in ORR between IO vs IO with RT group ≥G3 TRAEs 32% in IO with RT group
Clump et al. (94)	Pembrolizumab	Total 70 Gy in 35 fractions	SCHNC	1-year PFS 89% in sequential arm vs 82% in concurrent arm 2-year PFS 89% in sequential arm vs 78% in concurrent arm G3 TRAEs 58.9% in sequential arm vs 76.7% in concurrent arm

^aAE = adverse effect; a/w = associated with; CRC = colorectal cancer; DCR = disease control rate; DLT = dose-limiting toxicity; G3 = grade 3; HR = hazard ratio; ICB = immune checkpoint blockade; IFN = interferon; IO = immune-oncology; mACC = metastatic adenoid cystic carcinoma; mCRPC = metastatic castrate resistant prostate cancer; Mets = metastases; mPFS = median progression free survival; mRCC = metastatic renal cell carcinoma; MSS = microsatellite stable; Nivo = nivolumab; NSCLC = non-small cell lung cancer; ORR = objective response rate; OS = overall survival; PBMC = peripheral blood mononuclear cell; PDAC = pancreatic ductal adenocarcinoma; PD-L1 = programmed death ligand 1; PFS = progression-free survival; RT = radiation therapy; SBRT = stereotactic body radiation therapy; SCHNC = squamous cell head and neck cancer; TRAE = treatment-related adverse event.

In summary, clinical studies combining anti-PD1 or PD-L1 and/or anti-CTLA4 therapies with RT demonstrate immunomodulation triggered by RT but efficacy is yet required to be investigated in further large-scale studies.

RT in combination with novel immunotherapy agents

The introduction of novel immunotherapy agents may be a promising approach to overcome RT-induced immunosuppression and synergize with the immunostimulatory effects of RT. Although many approaches might be considered, near-term possibilities may include but are not necessarily limited to innate immune targets to enhance tumor-associated antigen shedding and presentation, cytokines and/or chemokines, T-cell checkpoints, and anti-CSF1 to overcome immunosuppressive macrophages.

Below, we review current data on combination therapy of RT with innate immune targets such as TLR, STING agonists, cytokines/chemokines, and IFN-γ–linked mechanisms.

Targeting innate immunity

RT with TLR agonist therapy

TLRs are a class of pattern recognition receptor, which recognize damage-associated molecular patterns and play an essential role in innate immunity (107-109). Targeting TLR pathway along with RT may have synergistic activity in increasing tumor antigenicity, activation of DCs, and induction of tumor-specific cytotoxic T lymphocytes (110,111). The TLR/MyD88 pathway has been described to play an important role in RT-induced antitumor activity (51).

The activity of TLR agonists, such as TLR3, TLR7, and TLR9 combined with RT, has been tested via various preclinical models. In the Lewis lung carcinoma model, RT with TLR3 agonist polyinosinic‐polycytidylic acid (poly [I:C])–induced tumor growth inhibition was dependent on Batf3-positive DCs, CD8 T-cell activation, and TNF-α production. Pretreatment with poly(I:C) as opposed to after RT showed more robust antitumor efficacy mainly driven by decreased infiltration of MDSCs (112). The efficacy of intratumorally as well as intravenously administered TLR7 agonists combined with RT was demonstrated in murine colorectal, lymphoma, and melanoma models. The combination resulted in autophagic cell death, increase in CD8+ T cells, and decrease in Tregs and MDSCs with subsequent clearance of tumor cells and reduction of metastatic potential of tumor cells in melanoma and pancreatic models (113,114). This antitumor effect was more potent with fractionated RT (2 Gyx2) as opposed to single 10 Gy RT (113,115,116). TLR9 agonists and RT have been found to have a synergistic activity in increasing tumor antigenicity. They activated nuclear factor-kB and mitogen-activated protein kinases and enhanced both innate and adaptive immune system and the immune system response against tumor-associated antigens (117-119).

Preclinical data support combining TLR agonists with RT and has led to multiple early phase trials evaluating the role of TLR agonists to enhance responsiveness to RT (115,117,120-124). Two phase II clinical trials have studied RT with TLR3 agonist combinations in patients with glioblastoma multiforme (123,125). Polyinosinic-Polycytidylic (Poly-ICLC) with RT (2 Gyx30) was well tolerated and resulted in median OS of approximately 15 months, which is comparable with standard-of-care RT plus temozolomide. Adjuvant poly-ICLC and chemoRT was well tolerated but provided only mild improvement in median OS (17.2 months). The combination of intratumorally injected TLR9 agonists and RT (2 Gyx2) was safe in patients with low grade B-cell lymphoma patients and resulted in an overall response rate of 26.7%. This combination was later evaluated in patients with mycosis fungoides and was well tolerated resulting in clinically meaningful results in approximately 30% of patients (126-128). The injection of class C CpG (TLR9 agonist) SD-101 with RT was safe and resulted in tumor burden reduction in approximately 90% of patients with indolent lymphoma. The treatment was associated with an increase in the number of CD4⁺ effector and CD8⁺ cells and a decrease in Treg cells in the injected lesions (126,129,130,131). These studies were done in a small cohort of patients, and despite demonstrating safety and potentially clinically meaningful profiles along with promising immune reinvigoration, they did not reach to large phase II-III trials to officially evaluate the clinical benefit.

RT with STING agonist therapy

STING is another pattern recognition receptor that drives immune responses via activation of IFN-1, vascular normalization, and infiltration of CD8 T cells, as well as induction of nonclassical tertiary lymphoid structures (132-134). cyclic GMP-AMP synthase (cGAS)-STING pathway plays an important role in RT-induced antitumor immunity. Radiation-induced IFN-1 signaling and DC sensing of irradiated tumor cells are dependent on STING protein (135). In STING-deficient mice, RT fails to induce increases in IFN-β and CXCL10. STING is also a mediator for RT-induced adaptive immunity. RT induction of STING increases in tumor-associated antigen-specific CD8⁺ T cells in wild-type mice, whereas in STING-deficient Tmem173^-/^-, mice failed to demonstrate increases in CD8⁺ T cells (135). In HNSCC cell lines, STING was shown to regulate genes involved in reactive oxygen homeostasis with STING-deficient cells demonstrating decreased reactive oxygen species. Intravenously administered STING agonist SB11285 demonstrated synergistic activity with RT resulting in prolonged tumor growth inhibition in HNSCC (136). Although there are no results from human clinical studies that evaluated the combination of RT with STING agonists, this approach is promising to investigate (137).

RT with cytokine therapy

The combination of RT with cytokine therapy has been demonstrated to induce immunostimulatory changes and antitumor activity in preclinical models. The interaction of TNF-α and RT has been described since the 1990s, reporting radiosensitizing effects of TNF-α while using it before irradiation (138). Low-dose total body irradiation with IL-2 resulted in reduction of tumor burden in comparison to IL-2 alone and increased intratumoral natural killer cells and macrophages in a murine melanoma model (139). Hypofractionated RT at the dose of 6 Gyx4 combined with IL-12 microsphere resulted in greater tumor reduction in orthotopic pancreatic ductal carcinoma model in comparison to hypofractionated RT alone. The treatment resulted in intratumoral increase in IFN-γ and CXCL10, upregulation of M1, and downregulation of immunosuppressive M2 genes (140). RT at the dose of 10 Gyx1 combined with intratumorally injected IL-12 resulted in tumor regression in orthotopic hepatocellular carcinoma model. The therapy resulted in a decrease in MDSCs and increase in CD40⁺, CD80⁺, MHC II expressed DCs, CD8+, and natural killer cells (141). The RT at the dose of 8 Gyx3 in contrast to 20 Gyx1 combined with subcutaneous IL-15 vs IL-15 or RT alone resulted in long-term irradiated and nonirradiated tumor growth control in TSA mammary cancer cell lines. The RT with IL-15 resulted in an increase in the number of CD103 dendritic cells (cDCs), expression of CD40, CD80, and CD86 costimulatory molecules, and intratumoral infiltration of CD8+ T cells (142,143).

While thinking about combining cytokines with RT in the clinical setting, careful evaluation of safety data must be considered. Cytokine therapies have historically been associated with adverse events limiting their broad use in clinical settings. Small case series of melanoma patients raised concerns regarding safety of IFN with RT combination because of observed RT necrosis in brain and subcutaneous tissue (144). In stage III melanoma, IFN with RT (6 Gyx5) combination showed acceptable safety profile with the most common dermatological severe toxicity being witnessed in 9% of patients (145). Other combinations including IL-2, GM-CSF, and TNF-α demonstrated no dose-limiting toxicities and were overall well tolerated in patients with melanoma, renal cell carcinoma, and pancreatic cancer (6,8,146,147,148). In summary, the studies combining cytokines with RT have not demonstrated significant increases in toxicities in comparison with each modality alone.

With the introduction of IL-2 as a therapeutic modality for patients with melanoma and RCC, several studies then investigated combining IL-2 with RT aiming to increase antitumor responses. In a small study of 12 patients with melanoma and RCC, the combination of SBRT (20 Gy per fx) administered before IL-2 resulted in ORRs of 71% and 60%, respectively. The responders demonstrated higher peripheral blood CD4⁺ effector memory T cells than nonresponders (6). In a randomized phase II trial of SBRT (1 or 2 doses, 20 Gy per fx) with IL-2 in 44 patients with advanced melanoma, SBRT with IL-2 resulted in 54% ORR in contrast to 35% in IL-2 alone cohort. However, enhanced ORR did not translate to improvement in PFS or OS. The study showed no difference in ORR in radiating 1 vs more than 1 tumor sites (7).

Golden et al. (8) combined GM-CSF with RT (35 Gy in 10 fx) targeting rate of abscopal responses as a primary objective of their study. The combination resulted in nonirradiated tumor response in 27.6% of patients with improved OS among abscopal responders. Patients with progressive disease had higher neutrophil to lymphocyte ratio at baseline in comparison with those in responders. The abscopal response was also described in a patient with metastatic pancreatic cancer patients when combining GM-CSF with local RT (45 Gy in 15fx) (54).

In the initial phase I study, intratumoral injection of TNFerade (adenovector containing TNF-α with a RT-inducible promoter) was combined with RT (20-70 Gy) and demonstrated safety of this approach along with ORR of 70% in patients with advanced solid tumors (149,150). However, follow-up study did not demonstrate with recurrent head and neck cancer and pancreatic cancer did not demonstrate superiority in OS (148,151).

In contrast to the above-mentioned cytokines, which are not widely used in clinical practice, targeting IL-6 and IL-8 are of particular interest for studying in investigational settings. IL-6 and IL-8 promote recruitment of immunosuppressive MDSCs, and overcoming RT-induced expansion of MDSCs and M2 macrophage is an important area of investigation to expand the clinical benefit of RT.

Overcoming immunosuppressive effects of RT

MDSCs induced immunosuppression and M2 macrophage polarization

Various preclinical studies proposed targets such as colony-stimulating factor 1 receptor (CSFR1), IL-6, and IL-8 to be combined with RT to address this immunosuppressive effect (62,141,152-154). RT resulted in an increase in CSF1 gene expression in irradiated tumor and serum CSF1 cytokine levels in patients with prostate cancer. The addition of CSF1R kinase inhibitor PLX3397 to RT resulted in a decrease in intratumoral and splenic MDSCs in RM-1 prostate cancer mouse model (62). In GL261 glioma cells, this combination also resulted in an increase in number of tumor-infiltrating lymphocytes, CD8+ and CD4+ T cells, and tumor-specific IFN-γ production (153). In vitro studies of RT-resistant esophageal squamous cell carcinoma cells demonstrate increased expression of IL-8 levels, suggesting that elevated IL-8 is one of the mechanisms of RT resistance (155). IL-6 enhanced recruitment of MDSCs in irradiated prostate tumor microenvironment and inhibition of IL-6 resulted in antitumor effect (156). The elevation of IL-6 after irradiation has been reported in C6 glioma cells resulting in radioresistance (157).

In the clinical setting, the safety and efficacy of combination of cabiralizumab (anti-CSF1R mAb) with nivolumab and hypofractionated RT was evaluated in patients with advanced solid tumors. The therapy was safe, resulting in a DLT rate of 12% and requiring no dose reduction of hypofractionated RT. The ORR was 22%, and responses were driven mainly by a decrease in size of irradiated tumor (158,159). IL-8 levels prior to irradiation were inversely associated with response to IO-RT in this study. The phase I clinical trial of anti–IL-8 therapy with RT (30-50 Gy in 3-5 fx) and anti-PD1 therapy (NCT04572451) is ongoing in patients with advanced solid tumors and melanoma. The trial of anti–IL-6 tocilizumab with ipilimumab and nivolumab and RT is ongoing in patients with advanced pancreatic cancer (NCT04258150).

RT with anti-TGF-β therapy

TGF-β signaling pathway activation is one of major immunosuppressive TGF-β 1 in irradiated lesions, which can create an immunosuppressive environment limiting activity of RT in addition to enhancing fibrosis and inducing chronic organ damage (63,160).TGF-β creates immunosuppressive environment via inhibition of dendritic cell activity, cytotoxic T-lymphocyte function, decreased production of IFN-γ, and activation of Treg cells, along with decreasing the effectiveness of RT via inducing DNA damage response (161-164).Targeting TGF-β pathway can be promising in not only increasing efficacy of RT but also preventing RT-induced pulmonary and rectal fibrosis (165,166).The inhibition of TGF-β with TGF-β kinase receptor 1 inhibitor increased radiosensitivity of glioblastoma cells, resulted in increase in DNA double-strand breaks, and reduced DNA damage repair. Inhibition of TGF-β pathway impacted RT-induced gene expression changes with the largest effect of combination therapy being on cellular growth and proliferation (167). In 4T1 mammary breast cancer mouse model, TGF-β inhibition along with RT resulted in increase in CD8+ T cells and CD4+ T cells, gene expression profile of IFN-γ network and immune effector functions, and regression of irradiated and nonirradiated pulmonary metastatic lesions. The combination therapy resulted in increased expression of PD-L1 and PD-L2 in some proportion of 4T1 cell lines, and further anti-PD1 blockade resulted in complete regression of tumor (168). In murine CT26 colorectal cancer (CRC) model, oral inhibitor of TGF-β SM16 was given prior to RT resulting in CD8+ T cells in the tumor and CD4+ T cells in TDLNs. Preclinical data provide evidence of combination of RT with TGF-β inhibitors, but safety and efficacy are yet to be confirmed in human clinical trials.

Targeting Indoleamine 2,3-dioxygenase (IDO) pathway

Several preclinical studies support the enhanced immune activation while combining RT with indoleamine 2,3-dioxygenase pathway targeting molecules. The main function of IDO1 is the regulation of acquired local and peripheral immune tolerance in normal and pathological scenarios (169). As such, IDO1 activity induced the Treg compartment of the immune system, with concomitant inhibition of the effector T-cell compartment (170,171). Combination of IDO inhibitor with RT resulted in upregulation of CD80, CD86, and the major histocompatibility complex II in spleen DCs and the concurrent downregulation of CD4, CD25, and FOXP3 Tregs in Lewis lung cancer model (172). IDO1 enzyme inhibitor combination with anti-PD1mAb (PD1- monoclonal antibody) and RT resulted in tumor regression in a preclinical glioblastoma model (173). In the CRC model, IDO1 inhibition sensitized CRC to RT-induced cell death, whereas the IDO1 metabolite kynurenine promoted radioprotection. IDO1 inhibition also potentiated Th1 cytokines and myeloid cell-modulating factors in the tumor microenvironment and promoted an abscopal effect on tumors outside the RT field. Conversely, IDO1 blockade protected the normal small intestinal epithelium from RT toxicity and accelerated recovery from RT-induced weight loss, indicating a role in limiting side effects. These data demonstrated that IDO1 inhibition potentiates RT effectiveness in CRC (174). Overall, there are no clinical data about combining RT with IDO inhibitors other than a study in children with diffuse intrinsic pontine glioma combination of IDO pathway inhibitor indoximod with RT (54 Gy in 30 fx), and chemotherapy was well tolerated and improved OS in comparison with historical control (175).

Future directions and summary

With the rapid development of novel cancer immunotherapies, a number of new immune mechanisms are coming into the clinical setting, which may have the potential to enhance and overcome resistance mechanisms of RT. In particular, the introduction of a new generation of ICB agents in clinical trials such as T-cell immunoglobulin and ITIM domain (TIGIT), lymphocyte activation gene 3 (LAG3), and T cell immunoglobulin and mucin domain-containing protein 3 (TIM3) brings a new potential for combining these agents with hypofractionated RT. In a CT26 colon cancer mouse model, the combination of anti-PD1 and anti-TIGIT with 8 Gy RT in 3 fractions resulted in complete regression of tumor in 90% of cases. The antitumor effect of anti-TIGIT was not seen while combining with 2 Gy RT in 18 fractions. The results highlight that different regimens of RT may be required while combining with various ICBs (10). In a murine hepatocellular cancer model, RT induced increased expression of TIM3 on CD4+ and CD8+ TILs. The combination of RT and anti-TIM3 resulted in improved responses mainly driven via CD8+ T-cell activation (176). In a murine glioma model, the combination of RT with anti-TIM3 and anti-PD1 resulted in increase in CD8+ to Treg ratio, IFN-γ–secreting CD8+ cells and improved survival in comparison with anti-TIM3, anti-PD1, and RT monotherapies (177). Another promising combination therapy includes RT with targeting myeloid checkpoint “don’t-eat-me” molecule CD47. In a mouse KP1 small-cell lung cancer model, anti-CD47 antibody enhanced antitumor activity of local RT as well as abscopal responses in a macrophage-mediated T-cell independent manner (178). In addition to novel IO-based therapies, an intriguing combination approach may be using radioconjugates, which deliver RT with tumor-targeting agents to facilitate new IO-RT combinations. In ovarian and breast preclinical models, ²²⁵Ac-anti-HER2 and anti-CD47 antibodies demonstrated improved efficacy over each single agent alone. The underlying mechanism of this enhanced response is potentially driven by upregulation of calreticulin and prophagocytic signaling activity (179). Another developing approach for broadening the systemic impact of RT is biology-guided RT (BgRT), which uses positron emission tomography as an imaging tool to characterize the biological characteristics of the tumor and to deliver RT. The fast-rotating linear accelerator obtains signals in a real-time fashion from different metastatic sites allowing treatment of multiple tumor locations during the same session with reduced RT dose. BgRT for lung SBRT has been demonstrated to reduce planned target volumes and organs at risk (180,181). Lastly, nanoparticle-based drug delivery is an evolving approach with the goal to deliver the drug to the tumor region with less systemic toxic effects. This has gained a particular importance in non–T-cell inflamed tumors and for drugs that demonstrated systemic toxicities otherwise. Zn-CDA is a tumor-targeted STING agonist that demonstrated improved efficacy in combination with RT and anti–PD-L1 over either strategy alone in T-cell noninflammed pancreatic cancer and glioma models. The treatment was associated with enhanced gene expression related to antigen presentation and M1 polarization (182).

In summary, current data support the safety of combination therapy with RT and immunotherapy along with biological correlates showing enhanced immune response to combinatorial therapy. Furthermore, preclinical data support further efficacy investigation of novel ICB agents with RT along with targeting TGF-β, anti–IL-8, IDO, TLRs, and STING. Apart from incorporation of novel IO treatments, approaches such as radioconjugates, BgRT, and nanoparticles are advanced strategies to use for future combinatorial therapies. These as well as other RT combinatorial partners represent a bright future for clinical investigation of novel RT-IO clinical trials.

Data availability

No new data were generated or analyzed in support of this research.

Author contributions

Lilit Karapetyan, MD, MS, FACP (Conceptualization; Writing—original draft; Writing—review & editing), Uzoma K. Iheagwara, MD, PhD (Writing—original draft; Writing—review & editing), Adam C. Olson, MD, MS (Writing—original draft; Writing—review & editing), Steven J. Chmura, MD, PhD (Writing—review & editing), Heath K. Skinner, MD, PhD (Writing—original draft; Writing—review & editing), and Jason J. Luke, MD, FACP (Conceptualization; Writing—original draft; Writing—review & editing)

Funding

This work was supported by the National Institutes of Health (UM1CA186690-06, P50CA254865-01A1, P30CA047904-32, R01DE031729-01A1 to JJL).

Conflicts of interest

Jason J. Luke reports: Outside of submitted work: Data and Safety Monitoring Board: Abbvie, Immutep, Evaxion; Scientific Advisory Board: (no stock) 7 Hills, Affivant, Bright Peak, Exo, Fstar, Inzen, RefleXion, Xilio (stock) Actym, Alphamab Oncology, Arch Oncology, Duke Street Bio, Kanaph, Mavu, NeoTx, Onc.AI, OncoNano, physIQ, Pyxis, Saros, STipe, Tempest; Consultancy with compensation: Abbvie, Agenus, Alnylam, Atomwise, Bayer, Bristol-Myers Squibb, Castle, Checkmate, Codiak, Crown, Cugene, Curadev, Day One, Eisai, EMD Serono, Endeavor, Flame, G1 Therapeutics, Genentech, Gilead, Glenmark, HotSpot, Kadmon, KSQ, Janssen, Ikena, Inzen, Immatics, Immunocore, Incyte, Instil, IO Biotech, Macrogenics, Merck, Mersana, Nektar, Novartis, Partner, Pfizer, Pioneering Medicines, PsiOxus, Regeneron, Replimmune, Ribon, Roivant, Servier, STINGthera, Synlogic, Synthekine; Research Support: (all to institution for clinical trials unless noted) AbbVie, Astellas, Astrazeneca, Bristol-Myers Squibb, Corvus, Day One, EMD Serono, Fstar, Genmab, Ikena, Immatics, Incyte, Kadmon, KAHR, Macrogenics, Merck, Moderna, Nektar, Next Cure, Numab, Palleon, Pfizer, Replimmune, Rubius, Servier, Scholar Rock, Synlogic, Takeda, Trishula, Tizona, Xencor; Patents: (both provisional) Serial #15/612,657 (Cancer Immunotherapy), PCT/US18/36052 (Microbiome Biomarkers for Anti-PD-1/PD-L1 Responsiveness: Diagnostic, Prognostic and Therapeutic Uses Thereof). Steven J Chmura: Outside of submitted work: Royalties from Wolters Kluwer (UpToDate). Adam C. Olson reports: Outside of submitted work: research funding from Varian Medical Systems, Reflexion Medical and serves as a consultant for RenovoRx.

The other authors (LK, UKI, HKS) declare no disclosures or conflicts of interest.