Stereotactic body radiotherapy (SBRT) is a key treatment modality for early stage and oligo-metastatic non-small cell lung cancer (NSCLC)1, 2. SBRT induces DNA double strand breaks, leading to cell killing, and may also modulate systemic immunity. Understanding the immunomodulatory mechanisms of SBRT may have a significant impact on the strategies for combining immune checkpoint blockade (ICB) and SBRT.
CD8+ cytotoxic T lymphocytes (CTLs) can mount responses against many human cancer types, but are often insufficient to eradicate tumors, as they become exhausted3-5. ICBs may reduce CTL exhaustion, while SBRT may further promote systemic immune activation. But how to achieve activation of anti-tumor CTL responses using SBRT and what the optimal radiation dose (ablative versus non-ablative), fractionation schedule and treatment sequence in combination with ICB are, remain outstanding questions.
We evaluated the impact of ablative SBRT on systemic immunity in patients with early stage NSCLC in a prospective study. We used immune profiling of peripheral blood by longitudinal assessment at first SBRT fraction (baseline), during and at the end of SBRT as well as at first (FU1) and second (FU2) follow-up (six weeks and another 4 and a half months after the last SBRT fraction, respectively). The pre-specified primary endpoint was increase (yes/no) in circulating CD8+ CTL counts at FU1 compared to pre-treatment, and secondary endpoints included changes in other T-cell subsets at all time-points. Study accrued 56 early-stage NSCLC patients between 2016-2021, of whom 50 were evaluable (4 dropped out and 2 withdrew consent). Patients and treatment characteristics are shown in eTable 1.
The absolute counts of circulating CD8+ CTLs at FU1 compared to baseline increased only in 21% of the patients (not significant). Moreover, there was a significant decrease in the mean absolute counts of CD8+ CTLs and CD4+ T-cells at all time-points compared to pre-treatment values (Figure 1A, B). These data show that there is a significant lymphodepletion during and after SBRT, despite the smaller irradiated volumes and no nodal irradiation.
We then sought to examine the changes in Ki-67, a marker of cellular proliferation and T-cell reinvigoration, expressed by cycling or recently divided cells4, 6, 7. Interestingly, the proportion of proliferating CD4+ and CD8+ T-cells among peripheral blood lymphocytes (CD3+ T-cells) significantly increased immediately after SBRT (at the end of the treatment) (Figure 1C). These increases occurred in the proliferating CD4+ and CD8+ T-cell subsets expressing PD-1, containing tumor-specific T-cells8, 9, but also in the PD-1– subsets (Figure 1D, E). Moreover, median fluorescence intensity of PD-1 immunostaining was also higher at the end of treatment in the CD8+ and CD4+ T-cells, indicating an increased expression level for this activation marker (Figure 1F). Additionally, the proportion of T-cells expressing the activation markers IFN-γ and IL-17A was increased during and after SBRT (Figure 1G-I).
Overall, there was a significant decrease in the number of naïve and memory CD8+ and CD4+ T-cell subpopulations after SBRT compared to pre-treatment values (eFigure 1A-B). Nevertheless, the fractions of CD8+ and CD4+ T-cells expressing inducible costimulatory (ICOS) increased at FU1 (eFigure 1C). Regulatory T cells (Treg), which are considered more radioresistant than other lymphocyte subsets, showed a significant decrease at FU1 (eFigure 1D). Similarly, myeloid-derived suppressor cells (MDSC), which can modulate tumor progression10, decreased at post-treatment time-points (eFigure 1E). TIM3 and CTLA-4 expression was detected only on a small minority of circulating T-cells, indicating that most circulating T-cells were not terminally exhausted after SBRT (eFigure 1F-G). All results are summarized in eTable 2.
As high radiation doses per fraction seem to attenuate immunogenicity11, we stratifyed for dose per fraction using as cut-off 10Gy. Patients treated with 10Gy or less (n=25) showed significant increases in proportion of proliferating CD8+ and CD4+ T-cells compared to pre-treatment values, but we detected no changes in patients who received more than 10Gy per fraction (n=19) (Figure 2A). The same results were obtained for the proliferating PD1+ and PD1– CD8+ T-cell fractions (Figure 2B).
Clinically, with a median follow-up of 31 months, median overall survival (OS) was not reached. At 2 and 4 years, OS rate was 75% and 51%, respectively, and progression-free survival (PFS) was 56% and 25%, respectively, with a median PFS of 36 months (eFigure 2). Only one patient developed local progression with a regional and distant progression, four patients developed distant metastases (8%), 3 patients regional and distant metastases (6%) and nine patients developed a regional recurrence (8%). There was no correlation between the biological effective dose (BED) and OS (hazard ratio per Gy [HR]=0.99, 95%CI: 0.98-1.01, p=0.4) or PFS ([HR]=0.99, 95% CI: 0.98-1.01, p=0.2). Moreover, there was no difference in outcomes between patients treated with more than 3 fractions versus those who received three fractions (OS: [HR]=1.50, 95%CIs: 0.56-3.99, p=0.4; PFS: [HR]=1.60, 95%CIs: 0.71-3.38, p=0.3). Thus, we found no signal of superior efficacy based on BED or dose per fraction. However, in an exploratory analysis, we found that a longer PFS was associated with an absolute increase of the CD8+ CTLs at FU1 compared to pre-treatment values (p=0.043, log-rank test, eFigure 2C).
Taken together, these data show that SBRT can lead to lymphopenia in early-stage NSCLC, despite the smaller irradiated volumes. Of note, an absolute increase in circulating CD8+ CTLs at follow-up compared to pre-treatment values was associated with longer PFS. Interestingly, SBRT-induced lymphopenia was associated with increased T-cell proliferation, which included tumor-specific T-cells12. Use of ablative SBRT has been reported to decrease the inhibitory signals from the tumor, reduce T-cell exhaustion and promote T-cell activation13, 14. In patients with treated with ICBs clinical failure was due to an imbalance between T-cell reinvigoration and tumor burden 4. The magnitude of T-cells reinvigoration in relation to pre-treatment tumor burden correlated with clinical response4. Moreover, the Ki-67 response in the PD-1+CD8+ T-cell subset peaked at 3-4 weeks after initiating ICB treatment4, 7, while in our study the Ki-67 response in the PD-1+CD8+ subset peaked at the end of SBRT. Optimal integration of ICB with SBRT should take into consideration these T-cell responses.
The increased proliferation of circulating CD8+ and CD4+ T-cells in patients treated with more than 10 Gy per fraction could be due to immunogenic cancer cell death. In pre-clinical studies cytoplasmic leakage of DNA after 8-12Gy was detected by cGAS/STING and activated primordial viral response pathways leading to production of type I IFN activation, while at higher doses there was a reduction in the IFN response and T-cell priming leading in lack of synergy with ICB11. This concept was tested in a study, in which sub-ablative total doses of 3x8 Gy were used for the combination with ICBs showing a prolongation of survival but did not meet the pre-specified endpoints15. Our results show that both systemic immune modulation and reduction of tumor burden can be achieved when using ablative SBRT with less than 10Gy per fraction.
Our study has limitations. Due to the different duration of SBRT regimens, post-treatment evaluations were not time matched.
In conclusion, our study shows that SBRT alone can significantly increase the fraction of proliferating CD4+ and CD8+ T cells, most prominently at the end of treatment and only when using 10Gy or less per fraction. These data have direct implications for the optimal integration of ICBs with SBRT in NSCLC and potentially other malignancies.