Systemic modulation of stress and immune parameters in patients treated for prostate adenocarcinoma by intensity-modulated radiation therapy or stereotactic ablative body radiotherapy

Approximately 50% of all cancer patients will receive radiotherapy (RT) during their course of illness. RT aims to kill tumor cells while sparing as much as possible of the surrounding healthy tissues [1]. Intensity-modulated radiation therapy (IMRT) and stereotactic ablative body radiotherapy (SABR) are used to perform a more precise irradiation of the tumor volume [2]. While radiation delivery during IMRT is conventionally fractionated (e.g., 2 Gy/fraction, 5 days/week), it is hypofractionated during SABR (e.g., 7.25 Gy/fraction, 5 fractions every second day). Thus, widely different RT protocols are nowadays used for the treatment of similar tumor entities. RT has long been regarded as local therapy, but this point of view changed when data on systemic effects of locally delivered radiation were collected [3]. Biomarkers of radiation exposure have been discovered [4] in the blood of patients after local [5, 6] or total body [7‐9] irradiation. These bioindicators include genes such as FDRX (coding for ferredoxin reductase, a mitochondrial protein involved in electron transport), DDB2 (coding for the damage-specific DNA binding protein 2, involved in DNA repair), MDM2 (coding for the mouse double minute 2 proto-oncogene, involved in the regulation of p53 degradation), and SESN1, GADD45, and CCNG1 (coding, respectively, for the sestrin 1 protein, the growth arrest and DNA damage-inducible protein 45, and the cyclin G1 protein, involved in the inhibition of cell cycle and growth arrest in stressful conditions). Expression of these genes is induced by p53 after activation of the radiation-induced DNA damage response checkpoints. Interestingly, the expression of some inflammatory genes such as ARG1 (coding for arginase 1, involved in amino acid metabolism and cell proliferation), BCL2L1 (coding for a protein of the BCL2 family, involved in the control of apoptosis), and MYC (coding for the MYC proto-oncogene, involved in cell cycle progression, apoptosis, and cellular transformation) is also found to be dysregulated [8]. While expression of ARG1 and BCL2L1 was upregulated, MYC expression was decreased in the blood of cancer patients during RT. These findings show that RT can result in long-term modification of gene expression and potential immunomodulatory effects [3]. In fact, tumor development is intimately linked to the negative modulation of immune cell functions [10]. Tumor cells can, for example, up-regulate their expression of immune checkpoint molecules such as programmed death ligand 1 (PD-L1), making them, e.g., less susceptible to killing by cytotoxic T lymphocytes. It has been proposed that the anti-tumor efficiency of RT can benefit from a synergy of concomitant re-activation of the immune system [11‐13].

Systemic effects of ionizing radiation could also be observed at the level of the proteome, including cytokines, and metabolome in the blood of cancer patients treated with RT [14]. Early changes in the plasma level of a collection of cytokines were, for example, associated with toxicity in patients treated for lung cancer, and this response was modified by the combination of chemotherapy with RT [15]. The intensity of radiation-induced toxicity, including the inflammatory response, is one key factor affecting RT-related changes in the blood proteome [16]. RT-related changes could be also detected in the serum lipidome and dynamic radiation-induced changes in the levels of phosphatidylcholines (PCs) and lysophosphatidylcholines (LPCs) were observed in patients treated for head and neck cancer [17]. LPCs are pro-inflammatory lipids involved in atherosclerosis [18] and an increased plasma LPC/PC ratio has been reported in certain pro-inflammatory conditions [19]. It was recently reported that the LPC/PC ratio increased in the serum of whole-body-irradiated mice [20]. Hence, this parameter represents a potential inflammation-related metabolomic marker of the response to radiation.

However, only few data exist on joint analyses of the manifold stress and immune changes that might occur following local treatment of solid tumors such as prostate adenocarcinoma. Such information could shed further light on how a distinct RT scheme impacts at the systemic level and might help to improve multimodal therapies in the future.

In this exploratory study, we therefore complementarily analyzed the expression of stress-response and inflammatory cytokine genes, the modulation of immune cell populations, changes in metabolites, and changes in cytokines levels in the blood of a group of 8 patients undergoing RT for prostate adenocarcinoma. As a first attempt to find out whether modulation of these parameters at the systemic level depends on factors such as the irradiated volume, the dose delivered per fraction, the dose rate, and the total dose received, biomaterial of patients who were treated with two markedly different RT modalities, namely IMRT and SABR, was used for the analyses.

This exploratory study presents for the first time a detailed analysis of several stress and immune parameters at a systemic level that might be modulated in vivo following local exposure to ionizing radiation in a small group of patients treated for prostate cancer by RT using IMRT and SABR. Both RT modalities have similar anti-tumor efficiency [30] and a low level of acute toxicity was observed in our patients: only one of the IMRT patients exhibited grade 2 gastrointestinal toxicity. Therefore, the observed modulations of immune or immune-related parameters reported in this study can be considered to reflect the effects of locally applied ionizing radiation on systemic immune alterations in patients with prostate adenocarcinoma rather than differences in radiation-induced toxicity.

Overall, the exposure of IMRT and SABR patients differed in terms of the total dose delivered over the RT course, the dose per fraction, the number of fractions, the dose rate, and the irradiated CTV. Patients treated with IMRT and one patient treated with SABR received androgen-deprivation therapy. ADT has been shown to induce a rapid recruitment of T lymphocytes and antigen-presenting cells to the tumor bed [31]. ADT can also induce a rebound in the production of naïve peripheral T lymphocytes in prostate cancer patients [32]. However, no effects of ADT on the homeostasis of B, T, and NK cell subsets in the peripheral blood of treated patients were observed in a comparison of healthy donors and treated patients [33]. In this study, immunophenotyping did not reveal any specific modulation of lymphocyte populations in ADT-treated patients before or during RT. One can therefore conclude that ADT did not interfere with the analyses of systemic modulations during IMRT and SABR. The observed differences should rather result from differences in radiation volume and/or dose and/or treatment technique, including the number of fractions delivered. Because of the small size of 4 patients per treatment group, putative differences between RT modalities might be masked by inter-individual differences between patients. To circumvent this potential problem, we compared the evolution of the modulation of the different endpoints during IMRT and SABR, and not their absolute level. Furthermore, effect size values were calculated. In contrast to statistical testing, effect size does not depend on sample size and thus can provide more reliable results. The use of a restrictive threshold (at least a large effect size) allowed identification of the strongest differences and variations between samples with a high level of confidence.

In this study, a large array of stress and immune parameters at the systemic level in the blood of IMRT- and SABR-treated prostate cancer patients was analyzed. The average total dose delivered to the patients’ blood during RT in the treating center (with an estimation based on the individual treatment plans, assuming that the blood volume corresponds to 8% of the irradiated body mass and that there is a homogenous vascularization of irradiated tissues) was 0.035 Gy per 100 mL and 0.009 Gy per 100 mL in the case of IMRT and SABR patients, respectively. Because of the specificities of IMRT and SABR, samples B and C correspond to different treatment doses and durations. In an attempt to directly compare the effects of these two treatment modalities for the same dose, we interpolated values obtained in IMRT patients in samples B and C to equivalent radiation doses (i.e., 7.25 and 36.25 Gy).

A large body of work has been devoted to the identification of biomarkers of radiation exposure. The level of transcription of selected p53-dependent genes was shown to be a good candidate for both total and partial body exposure over a wide range of ionizing radiation doses [34]. Data on the systemic effects of IMRT and SABR are, however, still scarce. Our analyses revealed a transient increase in the expression of FDXR and DDB2 in both IMRT and SABR patients during the course of RT, with, however, different kinetics. These different kinetics might be related to the total dose delivered at each specific blood collection timepoint and/or to the exposed volume. Interpolating the putative expression levels obtained in IMRT patients to similar doses delivered in SABR suggests that DDB2 transcription is induced in both treatments already after 7.25 Gy. It further increases until a cumulative dose of 36.25 Gy, suggesting similar induction mechanisms during the course of SABR and IMRT treatments.

In contrast, the effects of local radiation exposure on immune parameters at the cellular and humoral levels were different in patients treated with IMRT and SABR. The induction of transcription of IL1B and TGFB1 genes took place only in IMRT patients. Generally, IMRT resulted in modulation of a much larger number of cytokines and proteins in the serum compared to SABR. Evolution of the LPC/PC ratio, a postulated metabolomic indicator of inflammation [19, 20], was strikingly different during the course of IMRT and SABR. Altogether, these observations suggest that inflammatory mechanisms are differentially regulated during the course of IMRT and SABR. In contrast, the effects of RT on the expression of certain activation markers in circulating immune cell subsets was very homogeneous. In all patients, an increase in the percentage of PD-1-expressing CD4⁺ T lymphocytes was for example observed at the end of the treatment.

The impact of even low doses of ionizing radiation on the expression of activation markers by immune cell subsets has already been observed [9]. Modulation of anti-tumor immune responses by interfering with PD-L1/PD‑1 interactions has systemic effects on the T cell compartment. Thus, an increase in circulating PD‑1⁺ CD4⁺ T cells (and in myeloid-derived suppressor cells) following RT completion was, for example, associated with reduced T cell activity in colorectal cancer patients [35]. However, on the other hand, RT can also result in an increased frequency of circulating tumor-specific T cells in the CD8⁺ population, as observed in colorectal and prostate cancer patients [36], and the frequency of circulating HLA-DR^hi monocytes is a strong predictor of progression-free and overall survival in response to therapy in stage IV melanoma patients [37]. It was already described in 1999 that even though RT reduces immune cell numbers in the peripheral blood, the remaining lymphocyte function was still within the normal range [38]. Our findings confirm such immune modulations in prostate adenocarcinoma patients. In addition, the results of this explorative study provide—for the first time—indications that different RT protocols and application methods such as IMRT and SABR modulate several components of the immune system differently, as well as their dynamics. Interestingly, in patients treated with IMRT, changes in the level of BAFF, a survival factor of B cells, correlated with absolute numbers of B cells. Thus, local radiation exposure partially exerts its systemic effects by altering immune cell subsets, their activation status, and also their microenvironment.

The immunological consequences of DNA damage are becoming more and more evident [39]. Interestingly, during IMRT, expression of the DDB2 gene appears to be already increased after 7.25 Gy, whereas induction of the expression of IL1B and TGFB1 genes in the same samples is only evident after 36.25 Gy. Similarly, the serum concentration of the inflammatory cytokines and proteins modulated during IMRT was mostly upregulated only between samples B and C, but not in sample B compared to the pre-RT sample A. This delay in the induction of stress-response genes and of inflammatory parameters in the peripheral blood cells suggests that these events are differentially regulated in blood cells with individual dynamics. It may be that inflammatory genes are not directly induced by radiation, but rather that their up-regulation is a consequence of the biological effects elicited by radiation exposure, namely tumor cell death and the death of at least a fraction of irradiated circulating blood cells [40]. Irrespective of their identity, these mechanisms and their outcome(s) are complex, as shown by the findings that both IL-1β, a prototypic pro-inflammatory cytokine, and TGFβ1, a prototypic anti-inflammatory cytokine, were coregulated during IMRT. Another example is the coregulation of the chemokine RANTES and of the exopeptidase DDPIV, as proteolytic cleavage of RANTES by DDPIV can modulate its activity. The simultaneous upregulation of effectors with opposing activities indicates that radiation-induced stress and immune effects are multifaceted. The balance of these activities is finely regulated during the responses elicited by IMRT and SABR. In the rare cases of a similar factor being modulated in patients treated with either IMRT or SABR, this modulation appears to be in opposite directions in the different RT modalities, as shown in our study of GDF-15 and Tim‑3. GDF-15 is a member of the TGFβ family with both pro- and anti-tumor activities [41]. Its expression is associated with cellular stress conditions like irradiation in human PBLs [42], primary fibroblasts exposed ex vivo [43, 44], and murine splenocytes following whole-body exposure [45]. Our results now show that it is also transiently induced in vivo during RT, with specific regulation in each RT modality. This difference may be linked to the different initial levels of GDF-15 in both groups of patients, or to differences in cell death/stress or tissue damage. We did indeed observe regulation of several factors denoting endothelial dysfunction (CD31, VCAM, thrombospondin) and tissue remodeling (MMP9, chitinase 3‑like 1) in IMRT, but not in SABR patients. The other differently regulated protein common to both groups of patients is Tim‑3, a negative regulator of Th lymphocytes and innate immune cell activation when expressed on the cell surface [46]. In vitro, Tim‑3 has been found to be shed by ADAM proteases from the surface of human TLR-activated CD14⁺ monocytes [47, 48]. Here, we observed a low but detectable persistent increase in the circulating level of Tim‑3 in the weeks following the end of the treatment in both IMRT and SABR patients. So far, no function has been attributed to this soluble Tim‑3 protein, but its increase may reflect innate immune cell activation during IMRT and SABR treatments.

Even though this explorative study was already quite comprehensive, future studies should be even more detailed and include more immune parameters such as danger signals that reflect radiation-induced immune damage response [49] or HSP70 abundance [50], which give indications about the tumor status.

In addition, future studies should also consider additional endpoints such as DNA damage, including complex and oxidative lesions, apoptosis, the use of sensitive methods such as enzyme-linked immunosorbent assay (ELISA) to measure the modulation of circulating inflammatory cytokines, and of course clinical outcome. The level of a collection of cytokines was found to be differentially regulated in lung cancer patients treated by RT only vs. patients treated with RT + chemotherapy, and the early modulation of some of these cytokines was correlated with lung toxicity [15]. It was also shown that DNA damage could be detected in non-irradiated, out-of-field tissues in a different cohort of patients treated by RT or RT + chemotherapy for lung cancer [51]. This abscopal effect is similarly observed in pre-clinical mouse models locally exposed to a single dose of high-dose-rate synchrotron radiation [52, 53]. In mice, the generation of genotoxic lesions at distant sites was found to depend on a functional immune system and was attenuated in the absence of the CCL2/MCP‑1 cytokine. It would be interesting to find out whether the different modulation of immune parameters that we observe in IMRT and SABR patients translates into a differential pattern of genotoxic stress at distant sites according to RT modality in patients. Pre-clinical mouse studies could be designed to investigate the eventual relations between dose, dose rate, repeated exposure, immune modulation, and genotoxic effects at distant sites in the context of local irradiation of normal tissue or tumors.

In conclusion, the data presented herein depict that localized irradiation results in systemic modulation of a large range of cellular and humoral immune parameters. The radiation-induced effects on inflammatory gene transcription, immune cell homeostasis, and serum concentration of lipids and cytokines appear to be different in patients treated by different RT modalities. These differences are qualitative (e.g., genes induced only in IMRT, more cytokines modulated in IMRT, myeloid cells modulated only in SABR) and quantitative (e.g., different evolution of LPC/PC ratio). This exploratory study suggests that RT modalities with similarly high therapeutic efficiency and low clinical toxicity can have different effects on immune system homeostasis and activation at the systemic cellular and molecular level for up to 5 weeks following RT completion. Late effects of RT such as a decrease in circulating CD3⁺ T lymphocytes have previously been observed in prostate cancer patients 12 months post treatment [54]. Thus, the findings presented here need to be confirmed in larger groups of patients in the future, also taking into account further analyses at even later timepoints after completion of RT.