Involvement of cytotoxic immune cells has been studied in a variety of model systems with different radiation qualities. The most notable modifications of lymphocyte actions are summarized in Table 1. Activation of CTL (shown in Figure 2) is mainly triggered via the T-cell receptor. In 67NR and A20 tumor-bearing mice irradiated with γ-rays (2–6 Gy), increased CTL cytotoxicity was reported (54). In an in vitro study by Garnett et al. (55) using several carcinoma cell lines, it was shown that after irradiation with γ-rays (10–20 Gy), WiDr, Caco-2, SW620, SW1463, and HCT116 cells were more sensitive to CTL-mediated lysis primed against carcinoembryonic antigen (CEA), while A549 cells responded to Fas-mediated cell lysis. Increased expression of Fas (CD95) was also observed on tumor cells in a MC38 mouse adenocarcinoma cell model after γ-irradiation (20 Gy), which enhanced the lytic activity of CTL (56). Expression of the surface proteins Fas, CEA, intercellular adhesion molecule 1 (ICAM-1), mucin-1 (MUC-1), and MHC-1 was increased in those cell lines as well, enhancing their susceptibility to immune mediated lysis (55). ICAM-1 can engage in receptor-ligand binding between a T-cell and an antigen-presenting DC and thereby contribute to T-cell activation (21) as well as recruitment of immune cells from the blood stream to endothelial cells before extravasation to the tumor (57, 58).

Similar results were obtained using 200 MeV protons (produced using a passive scattering proton beam). In in vitro tumor cell models (human prostate (LNCaP), breast (MDA-MB-231), lung (H1703) carcinoma, and chordoma (JHC7) cells), expression of HLA-ABC, CEA, MUC-1, and ICAM-1 was increased after proton (8 Gy) and γ-irradiation (8 Gy), as well as sensitivity of the tumor cells to CEA-specific CTL-mediated lysis increased (66). Increased CTL activity has been partially allotted to the production of unique MHC-I antigenic peptides after γ-irradiation (1–25 Gy) leading to increased tumor recognition by T-cells (59).

In vivo studies with carbon ion irradiation (290 MeV/n, LET 77 keV/µm) of tumor-bearing mice revealed an increased CTL-associated lysis of isolated tumor splenocytes after carbon ion irradiation treatment with supplementary intratumoral DC injection (67).

The Natural Killer Group 2D [NKG2D, reviewed by Spear et al. (68)] receptor promotes amongst others the activation of NK cells. The human NKG2D receptor recognizes the ligands MHC class I chain-related protein A (MIC-A) and B (MIC-B) and HCMV UL16-binding proteins [ULBP1-6 (68)]. Expression of NKG2D ligands has been found to be increased in irradiated tumor cell lines [NCI-H23, A549 (61, 69)] resulting in enhanced activity of NK cells (summarized in Figure 3) toward tumor cells after X-irradiation (8 Gy). The response was presumed to be triggered by an upregulation of the NKG2D ligands MIC-A/B and ULBP1-3 and could be further increased by inhibition of histone deacetylase (61).

Upregulation of MHC-I molecules and heat-shock proteins may abolish this activation (Figure 3) by induction of an increased expression of inhibitory NK cell surface receptors (61). NK cell activity has also been found to be diminished after cleavage of NKG2D ligands via matrix metalloproteases (69, 70).

An enhanced radiotherapy effect mediated by NK cells has been reported after electron irradiation (8 Gy). The cytotoxic effect of NK cells was tested on various cancer stem cell lines (MDA-MB-231, U87MG, A673, and PANC-1) in vivo, where mice were inoculated with cultured tumor cells and locally irradiated, then injected with NK cells, and in vitro, assessing the NK cytotoxicity directly on irradiated tumor cell (62).

As mentioned above, in vivo studies can imply direct irradiation effects on immune cells. A way around this is to inject non-irradiated lymphocytes into the irradiated tumor-bearing host and analyze the effects.

An in vivo study explained a reduced tumor volume (RMA-S lymphoma/B16 melanoma) in mice by injected NK cells after 5 Gy total body irradiation. The effect was even more pronounced after prestimulating the NK cells with IL-12, -15, and -18, with highly increased expression of interferon γ (IFN-γ), granzyme B, and perforin. Those prestimulated NK cells were found to have rapidly proliferated in dependence of IL-2 production by CD4+ Th-cells (60).

Enhanced antitumor response after X-irradiation (PANTAK Therapax DXT 300 Model X-Ray Unit, 42.5 Gy) has been linked to DCs. Intratumoral injection of DC was performed in mice bearing irradiated D5 tumors, resulting in reduced tumor size and increased IFN-γ secretion (71). As shown by Scholch et al. (65), in the in vivo (PANC-02 cells in mouse model) antitumor response of immune cells after irradiation (5 Gy × 2 Gy), the CTL mediated response dominates over NK cells, and was shown to be abrogated by depletion of DC, indicating a necessity of DC mediated antigen presentation for the immune cell effectiveness against tumor tissue. Although very promising, the described effects do not take the radiation effect on immune cells into account, since no immune cells were injected after irradiation (65). After X-irradiation (5 Gy × 2 Gy, 3 Gy × 5 Gy, 15 Gy), DC show an increased expression of IL-2R (CD25), which can mediate an increased activation of CD4+ T-cells via presentation of the activating IL-2 to the T-cell [and potentially CTL and NK cells as well, although not tested in the study (72)].

On the other hand, immunosuppressive Tregs were found to be increased in irradiated tumors in an in vivo mouse study, bearing lung and colon tumors (63) as well as in tumors and tumor draining lymph nodes of mice injected with mouse melanoma and lymphoma cell lines (64). The increased presence of Tregs was associated with increased tumor growth and has been hypothesized to depend on Langerhans cells, the DCs in the epidermis (64). Systemic inhibition of Tregs using cycloheximide (CHX) and anti-CD25 antibodies proved to increase the number of CD8+ and CD4+ non-Tregs. Along with those results, CHX and anti-CD25-antibody treatment resulted in enhanced tumor regression, indicating a suppressive function of Tregs (63). In other in vivo studies, Tregs were suppressed by blockage of cytotoxic T-lymphocyte associated protein 4 (CTLA4) in mice injected with 4T1 mouse mammary carcinoma or CD-26 murine colon cancer cells. Subsequent radiation exposure with 10 and 12 Gy of γ-irradiation resulted in tumor reduction that was associated with CTL-mediated cytotoxicity (73, 74). In the study by Son et al., the irradiation treatment was augmented with immature DC (74), but due to different irradiation parameters as well as different tumor application of the two studies, the effectiveness of this augmentation cannot be assessed.

Monocytes and T-cells were shown in vitro (THP-1 and Jurkat cell lines, respectively) to have increased viability after incubating them with conditioned medium from carbon ion irradiated neuronal tumor cells (SH-SY5Y and U87; Carbon ions 165 MeV/n, LET 30 keV/µm, 1–5 Gy), as well as decreased migration of THP-1, hinting at more in-depth interactions of immune cells in response to radiation (75).

Radiation therapy with an electron beam (fractionated 8 Gy on three consecutive days; Varian Truebeam linear accelerator) has been shown to slow tumor growth of mice bearing 67NR tumors in vivo in an abscopal manner (76). In the same model, after enrichment of DC using DC growth factor Flt3-L (Fms-related tyrosine kinase-3 ligand), abscopal tumor size reduction was observed after low doses (2–6 Gy) of γ-irradiation (⁶⁰Co source). The effect was proven to be T-cell dependent, as abscopal tumor size was not influenced in T-cell deficient mice (54).

The systemic inhibition of Tregs using CHX and anti-CD25-antibodies in an in vivo tumor-bearing mouse model (lung and colon carcinoma) or via CTLA4 blockage in an in vivo tumor-bearing mouse model (colon carcinoma) after irradiation of the tumor resulted in reduced growth of distant non-irradiated tumor cells (63, 74). The indicated suppressive action of Tregs on antitumor responses can thereby also be expected non-irradiated tumors.

These studies show that irradiation of tumor cells or tissue has long-ranging effects on different immune cell subpopulations. This results in activation of CTL and NK cells, supported by increased activity of DCs, which meets an orchestrated immune suppressive response initiated by Tregs. Activation of CTL and NK cells was shown in in vitro and in vivo studies, Treg activation only in vivo. As a broad variety of neoplastic cell lines activated these immune cell populations, the tumor cell type seems to have no apparent influence on immune cell activation.

The presence of immunosuppressing cyto- and chemokines is vital to the development and progression of tumor cells. The tumor cells themselves can secrete factors that protect them from lysis via CTLs or NK cells or elicit cytokine expression in other cells that enable tumor survival, most notably are TGF-β and IL-10.

Transforming growth factor β has been shown to reduce a wide variety of antitumor immune functions. It inhibits growth of immune cells and reduces IL-2, IL-2R, IFN-γ, and NKG2D expression resulting in impairment of their activity. Furthermore, downregulation of MHC-I molecules on the tumor cell surface reduces their susceptibility to CTL-mediated tumor cell lysis. Expression of TGF-β by several tumor types has been reported (77, 78).

Interleukin 10 is one of the immune system’s “Off-Switches,” known for its regulatory characteristics in suppressing inflammatory responses (79). It effectively reduces antigen-presentation, Th1 responses, NK cell cytokine expression, and functions of monocytes and macrophages. An important way of inactivating the inflammatory immune response is by reducing the abilities of DCs to present antigens and to produce proinflammatory cytokines such as IL-12 (80). IL-12 can promote NK-mediated actions against tumor tissue. Among other factors associated with tumor growth are TNF, IL-1, IL-6, IL-8 [C-X-C motif chemokine ligand 8 (CXCL8)], IL-11, IL-17a, IL-22, acute phase proteins, CCL20, PGE2, colony stimulating factor-1 (CSF-1)/macrophage CSF (M-CSF), vascular endothelial growth factor (VEGF), and granulocyte-macrophage CSF (GM-CSF) (81–86).

The field of cytokines promoting tumor development and progression is vast and has been reviewed elsewhere (81–85). The communicative relationship of cytokines in radiation biology, as well as general notions on their functions, has been extensively reviewed by Schaue et al. (87). One can suspect that modulation of cytokine expression may be able to accomplish long ranging effects in terms of non-targeted responses after irradiation exposure. In this review, the focus lies on the modulation of cytokine expression after exposure to ionizing radiation from differing radiation qualities and irradiation schemes.

Since the tumor environment is of immunosuppressive nature, the question arises, how irradiation of tumor cells modulates the cytokine responses that induce or further suppress the immune response. The cytokine and chemokine response of diverse tumor cell lines is shown in Table 2 and Figure 4.

Fractionated irradiation (5 Gy × 2 Gy, 3 Gy × 5 Gy, 15 Gy) of human colorectal adenocarcinoma cells (SW480 cell line) with X-rays has been reported to increase expression and secretion of proinflammatory cytokines IL-6, IL-8, IL-12p70, and TNF-α by DCs. The immunosuppressive cytokine IL-10 and the proinflammatory cytokine IL-1β were insignificantly increased without impeding antitumor response of Th1-cells (72).

The glioblastoma cell line T98G expressed and secreted an elevated level of IL-6 and IL-8 after γ-irradiation (1 Gy) (89). In addition to IL-6 and IL-8, CXCL1 expression was induced by high γ-radiation doses (10–30 Gy) in LN-229 glioma cells, this was observed for several days after irradiation (98).

Desai et al. (97) analyzed cytokine expression in the tumor cell lines HT1080, U373MG, HT29, A549, and MCF-7, using a single dose (2 and 6 Gy) and fractionated doses (2 Gy × 3 Gy) of γ-rays. Amongst the cytokines tested were TNF-α, IL-1β, IL-6, TGF-β, monocyte chemotactic protein 1 (MCP-1/CCL2), IL-15, VEGF, G-CSF, GM-CSF, Flt3-L, and IFN-γ-induced protein 10 (IP-10). While some cytokines (TNF-α, IL-8, IL-15, GM-CSF, and TGF-β) were highly upregulated after 6 Gy single-dose γ-irradiation, the expression profile strongly depended on the dose (TGF-β was downregulated in HT1080 cells at 2 Gy, highly upregulated at 6 Gy, and moderately upregulated in the fractionated irradiation scheme), cell line (downregulation of IL-6 in every irradiation scheme of U373MG cells but upregulation in HT1080 and A549 cells) and fractionation (IL-1β was downregulated in HT29 cells at 2 and 6 Gy single-dose irradiation but upregulated in the fractionated irradiation scheme).

In a mouse tumor model (RipTag-5 transgenic mice), TNF-α, IL-12p70, and INF-γ expression was found to be elevated, while VEGF and TGF-β were decreased after irradiation with 2 Gy γ-rays (100).

Tumor necrosis factor α and IL-1α were also reported to be released in H446 lung cancer cells after irradiation with γ-rays (8 Gy), but only TNF-α after irradiation with accelerated carbon ions (290 MeV/n, LET 13 keV/µm, 2 Gy) (99). IL-1 can act as stromal growth factor in tumors (103).

In conclusion, exposure to X-rays or γ-rays in therapeutic dose ranges (in fractionated or single-dose regime) modulates the expression of cytokines in many different tumor cell lines and also spontaneous tumor models.

Monocytes (THP-1 cell line) expressed reduced levels of the activating factors IL-15, IL-17, macrophage inflammatory protein 1β (MIP-1β, also known as CCL3) and IL-2 as well as increased levels of Treg-attracting IP-10 [CXCL10 (104)], Rantes (CCL5) and immunosuppressive VEGF (105) 24 h after irradiation with 1.5 Gy α-particles (²⁴¹Am Source, LET 127 keV/µm) (106). Irradiation of THP-1 derived macrophages with 0.5–20 Gy carbon ions (18.3 MeV/n, LET 108 keV/µm) has been shown to result in decreased TNF-α and IL-6 expression. Only extremely high doses (50 Gy) of carbon ions resulted in this study in an increased IL-6 expression (107). Irradiation of monocytes and macrophages with α-particles or accelerated carbon ions in therapeutic dose ranges (fractionated scheme) may therefore negatively modulate the immune response against tumor cells.

Damage-associated molecular patterns are secreted or released biomolecules that can initiate inflammatory responses upon binding to recipient receptors. Among those biomolecules are DNA molecules that are recognized by PRR anywhere outside the cell nucleus, or damaged RNA, which may be released in response to damages induced by ionizing radiation. Certain signaling proteins may also be recognized by PRR and stimulate immune functions.

One of those proteins is the high mobility group box 1 (HMGB1), a protein that under normal conditions binds to chromosomal DNA and facilitates nucleosomal structure maintenance and regulates gene expression. Acting as a DAMP (Figure 4), HMGB1 can support recruitment of immune cells via the chemokine receptor CXCR4 (which is bound by CXCL12) and activate immune responses via toll-like receptor 4 (TLR4) or induce caspase-1-dependent apoptosis (93, 95, 108) as well as DC maturation, Th1 polarization (109), and IFN-γ release of NK cells (110). It has been shown to be released after 2–15 Gy, X- and 0.9–3.5 Gy carbon ion irradiation by normal human fibroblasts (GM0639) and human bronchial epithelial cells (16HBE), as well as by the tumor cell lines A549, TE2, KYSE70, NCI-H460, WiDr, and the mouse melanoma cell line B16-F10 (93–95). Similar findings were reported in tumor cell lines of various tissue origins (MCF-7, NCI-H1703, DU-145, PC-3, HCT-15, SW480, T98G, and U251MG cells) (91, 92). Upon TLR9 stimulation, HMGB1 induces expression of IL-12p70, IL-12p40, IFN-α, IFN-γ, and TRAIL in DCs (111).

HMGB1 has also been indicated to induce NF-κB activity, as measured by p65 translocation as well as IκBα degradation, in presence of PRR CD14 and TLR4 (112). HMGB1 also induces increased TNF-α expression in human peripheral blood mononuclear cells, as well as TNF-α, IL-1α, IL-1β, IL-1RA, IL-6, IL-8, MIP-1α, and MIP-1β in human monocytes, but not IL-10 or IL-12 (113).

This indicates that inflammatory protein expression of immune cells may be in part due to stimulation via HMGB1 acting as DAMP after irradiation injury. As part of the bystander response of immune cells, this HMGB1 induced expression of cytokines may lead to prompting further immune action such as CD8+ T-cell or NK cell mediated killing of irradiated tumor tissue. There is very little data available regarding HMGB1 modulation by proton or carbon ion irradiation. This calls for more research in this upcoming and promising radiotherapy approach—especially in context of DAMP interaction with immune cells. While the primary signaling pathways for interactions of HMGB1 with any leukocyte population can be elucidated using X-irradiation, the effect of particle irradiation on this intercellular communication can only be assessed with specifically designed experiments for this question.

Besides the cyto- and chemokine expression of irradiated tumor or immune cells, bystander cells not directly hit by radiation might modify their gene expression profile. In coculture, U937 macrophages have been shown to secrete TNF-α and IL-1α (IL-1α not at high doses) after irradiation of NCI-H446 lung cancer cells with γ-rays (¹³⁷Cs Source, 8 Gy) but only TNF-α after irradiation with accelerated carbon ions (290 MeV/n, LET 13 keV/µm, 2 Gy) (99). Microbeam irradiation of 0.45% of a THP-1 derived macrophage population with 5 Gy carbon ions (18.3 MeV/n, LET 108 keV/µm) using a heavy ion microbeam resulted in significantly reduced expression of TNF-α and IL-6 (107).

The release of chemokines by irradiated cells and build-up of a chemokine gradient results in recruitment of selected immune cell populations to the irradiation site.

The chemokine CX3CL1 can recruit osteoclasts which are formed by fusion and differentiation of monocytes (114). This might have clinical relevance for osteolytic tumors.

Expression of CXCL16, the only known ligand for chemokine receptor CXCR6—expressed on NK cells, is increased in various breast cancer (MCF7, SKBR3, and MDA-MB231, 4T1, 67NR, HTB-20), melanoma (B16/F10), fibrosarcoma (MC57), and colon carcinoma (MCA38) cell lines after γ-irradiation (2–20 Gy) (88, 90, 101). This increased expression of CXCL16 facilitates an enhanced migration NK cells (Figure 4) toward the tested tumor cells (101).

Macrophage inflammatory protein 1α (CCL3) is a T-lymphocyte/monocyte derived chemokine recruiting CCR1/CCR5 expressing leukocytes (monocytes, DC, NK-, and T-cells). Administration of an active CCL3 agent (ECI301) resulted in reduction of tumors in vivo (Colon26 adenocarcinoma, MethA fibrosarcoma and Lewis lung carcinoma cells in mice) after electron irradiation (6 MeV electron beam, 6 Gy). Depletion of CD8+ T-cells reduced the antitumor effect of CCL3 administration indicating radiation induced recruitment of this cell population to the tumor site (115).

CCL3 administration also served to reduce tumor size of non-irradiated tumors in the in vivo model used by Shiraishi et al. (115). This effect was dependent CD4+ Th cells and NK cells, as depletion of those cell populations has shown. This indicates a CCL3 dependent recruitment of those populations to the non-irradiated tumor after irradiation (115).