In vitro, M1 Raw264.7 macrophages were reprogrammed toward M2-like macrophages after HDI (20 Gy). Sustained M2 phenotype after irradiation resulted from NFκB p50 activation, leading to high IL-10 production and reduced TNFα secretion. These results were confirmed in vivo since HDI (3 Gy × 20 Gy over 3 days) promoted M2-like macrophages at the tumor site, in mice with Panc02 cell xenografts (101). In another study, high doses of X-rays (25 Gy in one shot or 60 Gy fractioned over 3 weeks) increased the number of M2-like TAMs in a murine model of prostate cancer (TRAMP-C1 cell line). This was evidenced by a low inducible nitric oxide synthase (iNOS) level in macrophages and by an increase in Arg1 and COX-2 mRNA expression. Such high doses induced angiogenesis and tumor growth in this cancer model (7). An increased number of M2-like macrophages after high doses of radiation (12 Gy) was also observed in an oral cancer model. The recolonization of tumor by M2-like macrophages elicited the secretion of pro-angiogenic factors that contribute to neoangiogenesis, favoring tumor growth (102). In addition, Seifert et al. similarly described accelerated tumor growth after irradiation (12 Gy) of murine pancreatic tumor models. This was driven by an early infiltration of M2-polarized TAMs into the tumor after radiotherapy and resulted in a subsequent T-cell suppressive response. This T cell response was linked to the upregulation of PD-L1 expression in tumor. The authors showed that RT accelerated the progression of pancreatic dysplasia to invasive carcinoma but also promoted tumor growth in invasive pancreatic ductal adenocarcinoma (PDA). Indeed, a decrease in iNOS, IRF5 and H2eb1 mRNA expression and a higher expression of Arg1, CD206, and PD-L1 were observed in TAMs from pancreatic dysplasia and PDA irradiated with hypofractionated (12 Gy) or fractionated (3 Gy × 6 Gy) doses (103). When TAMs were collected from tumor-bearing mice after RT and transferred to other tumor-bearing mice, tumor growth was accelerated to a higher extent than with unirradiated-macrophage transfer (7). Altogether, these results showed that HDI skew TAMs in a M2-like phenotype (Table 1). This is why the blockade of M-CSF, in order to prevent macrophage recruitment, in combination with HDI is attractive. Indeed, it revealed interesting results: M-CSF inhibition combined to RT promotes tumor regression. This reinforces the idea that HDI-irradiated TAMs take part to the tumor growth and tumor radioresistance. Furthermore, high levels of TAM-derived IL-10 elicited T cell anergy in irradiated tumor (103). This is why an adoptive CD4⁺ T cell transfer has not met any success, except with the use of anti-M-CSF therapy combined to RT. This observation highlights the necessity to reprogram macrophages toward a M1-like phenotype to overcome the immunosuppression and the tumor regrowth after RT.

In vitro, human unpolarized monocyte-derived macrophages shifted toward M1-like macrophages after MDI (2 Gy × 5). This is highlighted by the upregulation of pro-inflammatory markers (M1 phenotype) such as human leukocyte antigen-cell surface receptor (HLA-DR) and CD86, but also by the downregulation of anti-inflammatory markers (M2 phenotype) such as decreased mRNA expression of CD163, MRC1 (C-type mannose receptor 1, CD206) and versican (ECM proteoglycan) and reduced secretion of IL-10. In these macrophages, phagocytosis, associated with the M1-like phenotype, was increased after MDI while irradiation did not influence the ability of cocultured macrophages to promote the invasion of cancer cells and angiogenesis, features of M2-like macrophages (104). In the same line, unpolarized Raw 264.7 macrophages exhibited a higher expression of M1 markers (iNOS, TNFα, IL-12p70, IFNγ RANTES, and MCP-1) and a higher abundance of the phosphorylated p65 subunit from NFκB after γ-irradiation (2 Gy) (105). The γ-irradiation (2 Gy) of a human macrophage cell line (U937) also provoked a higher expression of TNFα and IFNγ compared to unirradiated macrophages (106). In another study, CD11b⁺/Gr-1 peritoneal macrophages from RT5 mice were γ-irradiated (2 Gy) ex vivo and showed an increased iNOS expression, related to M1 phenotype (33). Finally, PMA-differentiated macrophages (THP-1), murine macrophages (Raw 264.7) and human monocyte-derived macrophages (hMDM) revealed a pro-inflammatory profile after moderate doses of γ-rays (2 and 4 Gy). This was evidenced by increased mRNA levels for TNFα, IFNγ, IL-6, and IL-23 and higher protein levels for IL-1β and IL-8. While most of the previous studies showed the activation of NFκB p65 for macrophage reprogramming, this last study revealed that IR-induced M1-like phenotype was promoted by the transcriptional expression of IRF5 and was ataxia telangiectasia mutated (ATM)-dependent (107). Altogether, the results from in vitro experiment showed that unpolarized macrophages tend to acquire a M1 phenotype after MDI.

In addition to program unpolarized macrophages to M1 phenotype, MDI also potentiated the already acquired M1 phenotype. Indeed, MDI (2 Gy × 5) promoted a pro-inflammatory profile in M1 macrophages stimulated with LPS/IFNγ, as indicated by an increased expression of HLA-DR (104). Another in vitro experiment revealed that doses under 1 Gy prevented the polarization of macrophages toward M1 phenotype whereas doses of 5 and 10 Gy promoted the M1 phenotype in murine Raw 264.7 macrophages when stimulated with LPS/IFNγ. This is emphasized by an increased expression of iNOS and a higher NO production in irradiated macrophages (105). In contrast, in hMDMs stimulated with M-CSF and IL-10 (M2 macrophages), MDI did not influence the expression of pro- and anti-inflammatory markers (104). In other words, the effect of MDI on polarized macrophages showed an enhanced M1 phenotype in M1 macrophages but no change in M2 macrophages, meaning that MDI could not reprogram TAMs in vitro.

Cocultured experiments were performed between radiosensitive (RKO cells) or radioresistant (SW1463 cells) colorectal cancer cell lines and human unpolarized monocyte-derived macrophages. After irradiation, the mRNA expression of some pro-inflammatory (CCR7, IL1β, CXCL8) markers was decreased whereas the mRNA expression of anti-inflammatory markers was unchanged when macrophages were cocultured with RKO cells. MDI also promoted cancer cell invasion and migration of cocultured RKO cells. However, when macrophages were cocultured with SW1463 cells, there was an upregulation of the expression of pro-inflammatory (CCR7, CD80) and anti-inflammatory markers (IL-10 and CCL18) but no change of cancer cell migration and invasion. It means that following MDI, unpolarized macrophages adopt a different phenotype according to the type of cancer cells with which they interact (108).

Interestingly, in vivo experiments revealed promising results. A group of researchers suggested that moderated single dose (2 Gy) of γ-ray was able to completely reprogram TAMs in tumor. Klug and Prakash analyzed M1 (iNOS protein level, NO production) and M2 (Ym-1, Fizz-1 and Arg 1 protein level) markers in CD11b⁺ peritoneal macrophages from RT5 mice after whole-body irradiation (WBI). While untreated mice exhibited elevated expression of M2 markers in peritoneal macrophages, whole-body irradiated mice revealed increased M1 marker expression and decreased M2 marker expression. However, local irradiation was not able to modify the expression of M1 and M2 markers when measured in tumor tissue lysate. As local MDI (1, 2, and 6 Gy) induced a decline of T cell infiltration, the authors used CD8⁺ T cell transfer in addition to local irradiation (2 Gy) to induce a shift in the polarization of TAMs toward a M1 phenotype and to promote T cell infiltration in tumor. Local irradiation in combination with CD8⁺ T cell transfer induced an enhanced expression of IL-12p40 and IFNγ (M1 markers) and a reduced expression of IL-10 (M2 marker) in tumor tissue lysate. Indeed, the infiltration of tumor by T cells was exclusively dependent on the reprogrammed TAMs. When TAMs was depleted with CL treatment before local irradiation and CD8⁺ T cell transfer, there was an inhibition of T cell recruitment into irradiated tumors. The infiltration of CD8⁺ T cell into tumor is made feasible with the normalization of the vasculature. Indeed, pro-inflammatory macrophages elicited the normalization of the vasculature and allowed an efficient T cell transfer, leading to the eradication of the tumor and an increased survival in irradiated RT5 mice. However, it remains elusive how CD8⁺ T cell transfer impacts TAM reprogramming. Interestingly, local irradiation alone is not able to reprogram TAMs toward a M1-like phenotype. Then, it seems clear that CD8⁺ T cell transfer plays a role in macrophage reprogramming. The underlying mechanism needs to be further investigated (96) Table 1 summarized the effects of MDI on macrophage polarization described here above.

Low doses of irradiation represents a good alternative in order to bypass the toxicities observed during intensive radiotherapy (HDI) (Table 1). However, LDI rather favor the M2 phenotype of TAMs. Indeed, LDI decreased the iNOS level and the NO production in Raw 264.7 macrophages polarized in M1 phenotype, resulting in repolarization of M1 macrophages toward M2 phenotype (105). Other results in favor of a M2-like phenotype were described after low doses of γ-radiation in murine macrophages. Indeed, irradiation of murine macrophages (RAW 264.7) with a dose of 0.5 Gy inhibited MPK1 [mitogen-activated protein kinase (MAPK) phosphatase]. MPK1 dephosphorylates the MAPKs p38, c-Jun and ERK1/2 and the dephosphorylation of p38 was associated with a decrease in TNFα and IL1β levels. Irradiation (0.5 Gy) of RAW264.7 macrophages stimulated with LPS also led to a decrease in p38 phosphorylation and TNFα production (109). In addition, Frey and Lodermann showed that low doses of X-rays (0.5–0.7 Gy) induced an anti-inflammatory phenotype in LPS-activated human THP-1 cells as evidenced by reduced amounts of secreted IL-1β in the medium. This was linked to a lower p65 NFκB nuclear translocation and a weak p38 phosphorylation (110, 111). Closely related to these results, it was also demonstrated that doses under 2 Gy favored the repolarization of M1 macrophages into M2-like macrophages, as evidenced by an increased level of TGFβ in the culture medium. This was associated to a reduced nuclear translocation of p65 NFκB (112).

In conclusion, HDI and LDI show no effect on M2 to M1 macrophage polarization while MDI clearly evoke TAM reprogramming. However, most of studies performed in this field are in vitro experiments, i.e., macrophages alone with no contact with cancer cells. Further in vivo experiments should reinforce the results obtained by Klug et al. with MDI. Furthermore, to our knowledge, no clinical study has been designed to analyze the effect of MDI on TAM reprogramming. It is thus difficult to transpose these data to the human tumors.

In vivo studies showed that exposure to low doses of WBI in healthy mice activated the innate immune cells including macrophages (113–115) (Table 2). In healthy mice, a single dose of WBI resulted in a T_H1 cytokine expression profile whereas fractionated dose irradiation drove a T_H2 shift in spleen and blood of mice (116). In the same line, single low doses (0.075 and 2 Gy) of total body radiation in healthy mice polarized peritoneal macrophages toward M1 phenotype as indicated by increased IL-12 and IL-18 secretion. This polarization is strongly linked to the activation of p65 NFκB and MyD88 a few hours after WBI (117). Only a few studies were published on WBI of tumor-bearing mice, as this treatment is irrelevant for human. However, as WBI would reprogram TAMs, studying the underlying mechanisms could be very helpful. A first study indicated that low doses (0.01–0.1 Gy) of WBI induced opposite effects on macrophages and NK cells in BALB/c radiosensitive mice and C57BL/6 radioresistant mice. Peritoneal macrophages exhibited a M1 phenotype in BALB/c mice, evidenced by an increased production of IL1β, IL-12 and TNFα. However, M2 macrophages were observed in tumor from C57BL/6 mice after WBI, partly elucidating the radioresistance of this mouse strain (118). Klug et al. also showed that peritoneal macrophages from RT5 mice, irradiated with a systemic dose of 2 Gy, revealed a higher iNOS level and NO production whereas the expression of M2 markers was decreased in these macrophages, indicating macrophage reprogramming. Furthermore, macrophages from total body irradiated mice, when transferred to tumor-bearing mice in parallel with CD8⁺ T cell adoptive transfer, induced tumor regression, similar to local irradiation in combination with CD8⁺ T cell adoptive transfer (96). In parallel, Prakash et al. described a switch of TAM polarization toward M1 phenotype when RT5 insulinoma bearing mice were treated by systemic irradiation (2 Gy/week on 2 weeks). Indeed, tumor lysates presented a higher expression of M1 markers [iNOS, pSTAT3, TNFα, IL-12 (p70)] and a lower expression of M2 markers (CD206, Fizz-1, Ym-1 and Arg1). Peritumoral macrophages showed an elevated iNOS expression and NO production after WBI, modifying the tumor microenvironment with the normalization of the tumor vasculature.

To explain the discordance between local irradiation and WBI on the reprogramming of TAMs, the authors suggested that WBI allowed the mobilization of fresh reprogrammed macrophages from various lymphoid organs to infiltrate the tumor site and the surrounding microenvironment. Infiltration of tumor by reprogrammed macrophages can result in the activation of antitumor T-cell responses and can thereby act as an “endogenous vaccine” (33) On the other hand, local irradiation is not able to invite macrophages from lymphoid organs to the tumor site. It could be hypothesized that the transfer of CD8⁺ T cells would induce macrophage reprogramming in lymphoid organs. The combination of T cell transfer to irradiation would promote the normalization of the tumor vasculature, allowing tumor perfusion by CD8⁺ T cells and reprogrammed macrophages.

Nuclear factor kappa B plays an important role in the polarization of macrophages. It was shown that M2 macrophages acquired their phenotype under the effect of p50–p50 NFκB homodimer while M1 macrophages are answerable to p50–p65 NFκB heterodimer (24). Once p65–p50 NFκB is translocated into the nucleus, it allows the expression of pro-inflammatory mediators [TNFα, IL-1β, IL-6, IL-12(p40), IFNγ, CXCL10 and NOS2 (14)] whereas p50–p50 NFκB inhibits the expression of these genes. Indeed, several experiments revealed the need for p65 translocation to activate M1 polarization (24, 119). Crittenden and his colleagues evidenced that HDI (60 Gy) produced M2 phenotype through p50–p50 dimer activation, leading to a subsequent IL-10 production (101). In the same line, low X-ray radiation dose (<1 Gy) reduced the translocation of p50–p65 NFκB in M1 activated macrophages (112). LDI on human macrophages and LPS-activated THP-1 macrophages reduced the nuclear amount of p65 NFκB, a phenomenon also correlated to a decreased secretion of IL-1β (M1 cytokine) (110). Interestingly, Teresa Pinto et al. revealed an enhanced phosphorylation of RelA (p65) in macrophages following MDI (2, 6, and 10 Gy), correlated to M1 repolarization (104). Hildebrandt et al. also evidenced the role of NFκB in TAM reprogramming in Raw264.7 macrophages after 2 Gy irradiation (105). In an in vivo study, whole-body γ-irradiation (2 Gy)-induced the phosphorylation of NFκB p65 in tumor lysate. The activation of NFκB p65 is closely related to a reprogramming of TAMs toward M1-like macrophages (33) (Figure 3).

In sepsis, a sustained activation of macrophages by LPS leads to a tolerance and a reprogramming of macrophages toward M2 phenotype. This reprogramming is driven by p21, involved in the shifting of the balance between p65–p50 and p50–p50 NFκB dimers. Indeed, p21 deficiency is linked to a reduced DNA-binding affinity of the p50–p50 homodimer and a prevalence for the p65–p50 heterodimer in macrophages stimulated with LPS (120, 121). In some cases, ubiquitination is controlled by the phosphorylation of a protein substrate (122). Indeed, some effectors of the DNA repair machinery are described to be involved in the ubiquitin proteasome pathway and regulate p21 protein level in ML-1 cells (myelocytic leukemia cell line) (123). Therefore, a high ubiquitination of p21 would lead to a lower level of p21 protein. This could be correlated with an increased p65–p50 DNA-binding affinity and could potentially favor a M1 phenotype after IR. Further investigations are, however, required to understand the possible role of p21 in regulating the NFκB balance after IR.

Radiotherapy induces major effects on cells by creating DNA damage. DNA damage and ROS produced by γ-radiation (2–20 Gy) induced the phosphorylation of ATM in U937 human macrophage cell line (106). When DNA damage appeared, ATM plays a central role in the detection and the activation of the DNA repair machinery [for review see Ref. (124)]. The activation of ATM notably leads to the ubiquitination of NEMO (NFκB Essential Modulator), a subunit of the IKK complex, a few hours after macrophage irradiation. The ubiquitination of NEMO is ATM-dependent as indicated by the reduced ubiquitination of NEMO when an ATM inhibitor (Ku5593) was used in combination with γ-rays (125). Ubiquitinylated NEMO then drives the activation of IKK complex in the cytoplasm. This activation goes through ubiquitination of IKβ and the subsequent degradation of this inhibitor by the proteasome. p65–p50 NFκB is, therefore, released in the cytoplasm and free to move into the nucleus where the heterodimer acts as a transcriptional factor [for review see Ref. (126)].

After γ-radiation (2 Gy), the phosphorylation of NEMO was followed by the activation of NFκB (p65). A subsequent upregulation of M1 markers (CD86, CD40, HLA-DR, TNFα, IFNγ) and a reduced secretion of IL-4 (M2 marker) followed the activation of NFκB. These changes are partially generated by elevated ROS content in cells, as indicated by the increase or decrease in the phosphorylation of NEMO when U937 monocytic cells were irradiated and treated with l-buthionine sulphoximine or N-acetyl cysteine (NAC), respectively. Indeed, l-buthionine sulphoximine irreversibly inhibits γ-glutamylcysteine synthase, hence depleting GSH level and increasing ROS content. NAC is a potent antioxidant, known to decrease ROS level by increasing GSH levels (106). However, another team showed that γ-rays (0, 3, 6, and 10 Gy) induced NFκB activation in Raw 264.7 macrophages, as highlighted by the degradation of the IKβ protein. The activation of NFκB was related to a higher NO production in macrophages. This phenomenon is dependent on DNA damage more than ROS level, since NAC did not affect NO production and IKβ degradation in irradiated cells (127).

Moreover, modulators of DNA repair, such as Olaparib (PARP inhibitor), increased ATM activation and upregulated IRF5 transcription, resulting in macrophage activation toward a pro-inflammatory phenotype (107). In conclusion, IR is able to induce DNA damage by direct or ROS-dependent interactions with DNA. DNA damage leads to the recruitment of the ATM kinase and DNA repair machinery. Consequently, ATM contributes to the reprogramming of macrophages after irradiation.

Radiation therapy promotes the formation of radicals, such as ROS or RNS in cells. ROS quantities depend on the dose applied to the cells. In macrophages, different ROS concentrations can differently influence the cellular responses and the polarization of macrophages. Indeed, the role of ROS in the polarization of macrophages is not so clear but the most accepted idea is that ROS play key roles in M1 responses (e.g., defense against invading microbes) (128) and regulates M2 polarization (129, 130).

Interferon-regulatory factor/signal transducer and activator of transcription (IRF/STAT) signaling is a central pathway in macrophage polarization. IRF can interact with two adaptors, MyD88 and TRIF, downstream of cytokine receptors or TLR. IRF5 and IRF4 are in competition to the binding of MyD88 and the subsequent activation of pro-inflammatory transcription factors, including NFκB. The activation of IRF5 through MyD88 then promotes M1 phenotype in macrophages. As a competitor for MyD88, IRF4 acts as an antagonist and can suppress M1 macrophage polarization. The balance between IFR4 and IRF5 in cells is critical for M2 and M1 polarization, respectively (24, 139).

Only one in vitro study established the role of IRF5 in macrophage reprogramming after irradiation. In this study, the authors showed that macrophages acquired a M1-like phenotype after exposition to γ-irradiation (2 Gy). This polarization is made feasible by the activation of NOX2 and ATM, hence driving the transcription of IRF5 (107). While these results are very interesting, the contribution of IRF/STAT pathway to macrophage polarization mediated by irradiation needs to be further investigated in vivo and in patients. Furthermore, IRF5 modulation is also known to drive M1 polarization through the activation of NFκB. The link between IRF and NFκB in macrophage reprogramming after radiotherapy should also be considered.

Nitric oxide mediator is a double-edge sword key regulator in tumor progression. On the one hand, NO has antitumoral effects as it promotes DNA damage, elicits cell death, and enhances anticancer therapeutic efficacy. Indeed, patients with lower iNOS expression levels, hence low NO production, showed recurrence of tumors and metastasis after RT. On the other hand, NO mediator is also known for its pro-tumoral effects since it induces antiapoptotic effects and promotes cell cycle, cancer progression, metastasis, angiogenesis, and chemoprotection. High level of iNOS is associated with poor prognosis and tumor-promoting effects (140). It thus seems that critical NO concentrations and iNOS levels finely modulate the fate of the tumor.

High doses of irradiation induced NO production and high iNOS expression level in macrophages. For instance, high doses of γ-irradiation (20 Gy) led to iNOS activation and NO production, favoring tumor growth and recurrence after RT (141). Another in vivo study using high doses (25 Gy or 3 Gy × 20 Gy) also generated high level of iNOS in murine macrophages after IR, contributing to the tumor growth (7). NO production contributes to DNA repair by inducing p53, poly(ADP ribose) polymerase and DNA-dependent protein kinase activation, leading to tissue regeneration and tumor regrowth. In addition, NO mediator is related to tumor angiogenesis and promotes reoxygenation. Therefore, targeting iNOS in murine macrophages enhanced postradiotherapeutic efficacy after HDI (141).

On the other hand, WBI (2 Gy × 2 Gy) on RT5 mice induced an increased NO production and iNOS level, enough to induce tumor regression. In addition, ex vivo radiation with similar doses on peritoneal macrophages from RT5 mice also stimulated an increased in iNOS level and NO production (33). Indeed, NO can be considered as a free radical and as a source of reactive oxygen and nitrogen species, inducing DNA damage (142). Another team hypothesized that iNOS and NO production induced by MDI (2 Gy) remodeled vasculature allowing cytotoxic lymphocyte recruitment, and subsequent tumor eradication (15). NO leads to higher tumor oxygenation by promoting vasodilatation and hence radiosensitization (142). It can diffuse toward bystander cancer cells leading to their radiosensitization (143).

High doses of irradiation would induce high level of NO in macrophages, leading to cancer progression. Contrarily, MDI drive a slight increase in NO production in macrophages, enough to participate to the reprogramming of TAMs and favoring tumor regression.

Mitogen-activated protein kinases are serine threonine kinases activated by diverse stimuli including cytokines and ROS. MAPKs regulate diverse pathways such as cell proliferation, apoptosis, and differentiation. MAPKs include extracellular signal-regulated kinase 1 and 2 (ERK 1/2), JNK and p38. They mediate immune cell functional responses to diverse signals and play a role in macrophage polarization. MAPK kinases (MAPKKs) activate MAPKs through dual phosphorylation on tyrosine and threonine residues (Thr–X–Tyr motif) located in the activation loop of the kinase domain (144). The phosphorylation of MAPK activates pro-inflammatory gene transcription and is correlated with M1-like macrophages. At the opposite, MAPKs are inactivated through dephosphorylation of threonine and/or tyrosine residues within the activation loop by MAPK phosphatase (MKP-1) (145). MKP-1-deficient macrophages exhibit skewed activation profiles: macrophages showed an enhanced pro-inflammatory phenotype in response to IFNγ and TNFα while this deficiency severely suppressed “M2-like” phenotype after IL-4 stimulation (146). Compared to wild-type mice, MKP-1-deficient mice produced greater amounts of TNFα, IL-1β, CCL2, granulocyte/macrophage colony-stimulating factor (GM-CSF), IL-6, IL-10, and IL-12p70 (147, 148).

Mitogen-activated protein kinases are also regulated by ROS and DNA damage (132). Within a few minutes, low doses (0.5 and 1 Gy) of γ-radiation induced a dephosphorylation of both ERK 1/2 and p38 MAPKs. This dephosphorylation depended on MKP-1 and induced the suppression of TNFα production in Raw 264.7 cells. These results correlated with the inability of LDI to reprogram M2 macrophages (109). Another study showed enhanced phosphorylation of ERK 1/2 and p38 following moderated doses (2 Gy) of γ-radiation in resident peritoneal exudate cells. This activation was related to a decreased dephosphorylation exerted by MKP-1 and was correlated with TNFα production. The explanation relies on the oxidation of the catalytic cysteine of MKP that inactivates the phosphatase, then triggering the activation of MAPK cascade including JNK and p38. These results correlated with the reprogramming of macrophages induced by MDI (149). Conflicting results revealed the inhibition of p38 phosphorylation in macrophages after WBI of tumor-bearing mice. The inhibition of p38 phosphorylation in macrophages allowed the reprogramming of TAMs toward M1 phenotype in tumor. This inhibition could be linked to the upregulation of MKP-1 (33).

The Nrf2 is involved in the management of intracellular oxidative stress. It is also known to repress inflammation by inhibiting pro-inflammatory cytokine expression. Indeed Nrf2 is able to bind the close proximity of IL-6, IL-1β, and IL-1α genes in LPS-stimulated macrophages. This binding hence inhibits the recruitment of RNA polymerase II to the transcription start site without affecting the p65 NFκB recruitment. The Nrf2 binding to these genes is independent on ROS level since the use of antioxidant (NAC) did not affect the inhibition of pro-inflammatory gene transcription mediated by Nrf2 (150). The transcription factor Nrf2 was translocated into the nucleus even after 0.1 Gy γ-radiation, leading to the upregulation of antioxidant proteins such as heme oxygenase-1 in Raw 264.7. Nrf2 translocation increased with the dose (0.1–2.5 Gy). The regulation of Nrf2 is mediated by ERK 1/2 pathway after the irradiation of macrophages (121). It is also known that in oxidative stress conditions, p21 interacts with Nrh2 domain of Nrf2, upregulating Nrf2-mediated antioxidant response (151). This could link the Nrf2 pathway to NFκB in the reprogramming of TAMs after IR: the absence of p21 prevents the anti-inflammatory response mediated by Nrf2 and favors the p50–p65 NFκB DNA-binding. The fate of irradiated macrophages would then depend on the connection between ROS level/DNA damage and the downstream-activated pathways.