Kinetics of the proportion of CD3−/CD19+ B cells before, during RCT and in the follow-up period
As summarized in Fig. 1A and supplementary Table 2, the percentage of CD3-/CD19+ B cells in patients with non-recurring SCCHN (n = 23) was significantly lower compared to that of healthy volunteers (Ctrl) already before start of RCT at t0 (t0 vs. Ctrl: 8.88% vs. 10.83%, p ≤ 0.05), and remained low until 3 months after RCT at t2. The most striking drop in B cells was observed directly after the application of 20 to 30 Gy at t1 (t0 vs. t1: 8.88% vs. 4.66%, p ≤ 0.001; supplementary Table 3). A recovery to initial B cell levels was detected 6 months after RCT at t3 (Fig. 1A, supplementary Table 2). Patients with locoregional recurrence did not reach initial levels up to t5 (3–15 months) in case of tumor recurrence (Fig. 1A, supplementary Table 5).
Kinetics of the proportions of CD3 +/CD4 + T helper, CD3+/CD8+ cytotoxic and CD4+/CD25+/FoxP3+ regulatory T cell subsets before, during RCT and in the follow-up period
The percentage of CD3+/CD56− T cells in SCCHN patients without locoregional recurrence dropped significantly 3 (t2) and 6 (t3) months after RCT compared to initial levels (t0 vs. t2: 70.37% vs. 60.88%, p ≤ 0.01; t0 vs. t3: 70.37% vs. 56.05%, p ≤ 0.001; Fig. 1B). At t2 and t3, the values were also significantly lower than those of controls (Ctrl vs. t2: 67.29% vs. 60.88%, p ≤ 0.05, Ctrl vs. t3: 67.29% vs. 56.05%, p ≤ 0.001; Fig. 1B, supplementary Tables 2) and to initial levels (supplementary Table 3). In patients with locoregional recurrence T cell ratios showed a slight increase at t1 that dropped below initial levels until t5 (Fig. 1B, supplementary Table 4).
As CD3+/CD4+ T helper cells make up more than 2/3 of the total CD3+ T cell counts, the kinetics of these cells followed that of the T cells in the course of therapy and in the follow-up period of SCCHN patients without locoregional recurrence. T helper cells dropped significantly at t2 and t3 (t0 vs. t2: 46.7% vs. 25.2%, t0 vs. t3: 46.7% vs. 26.27%, p ≤ 0.001; Fig. 1C, supplementary Tables 3), and patients without locoregional recurrence had significantly lower percentages of T helper cells compared to healthy controls between t1 and t3 (Ctrl vs. t1: 48.82% vs. 46.7%, p ≤ 0.05; Ctrl vs. t2: 48.82% vs. 25.2%, Ctrl vs. t3: 48.82% vs. 26.27%, p ≤ 0.001; Fig. 1C, supplementary Tables 2). Interestingly, the proportion of T helper cells in patients with locoregional recurrence was below that of controls and non-recurrent patients already before start of RCT and remained unaltered low throughout the whole course of therapy (Fig. 1C, supplementary Table 4, 5).
In contrast to the T helper cells, non-recurrent SCCHN patients had significantly higher proportions of cytotoxic CD8+ T cells at all timepoints (t0, t1, t2, t3) compared to controls (Ctrl vs. t0: 10.83% vs. 17.12%, Ctrl vs. t1: 10.83% vs. 19.03%, Ctrl vs. t2; 10.83% vs. 25.44%, Ctrl vs. t3: 10.83% vs. 10.54%, p ≤ 0.001; Fig. 1D, supplementary Table 2). After RCT, the proportion of CD8+ T cells in patients without locoregional recurrence increased further (t0 vs. t2: 17.12% vs. 25.44%, p ≤ 0.001; t0 vs. t3: 17.12% vs. 10.54%, p ≤ 0.05; Fig. 1D, supplementary Tables 3). Notably, recurrent patients showed a trend towards higher percentages of cytotoxic T cells at all timepoints (Fig. 1D, supplementary Tables 4 and 5), but no significant increase after RCT.
The percentage of regulatory T cells in SCCHN patients without locoregional recurrence at t0 was lower than that in healthy controls and dropped further directly after application of 20 to 30 Gy (t1) (Ctrl vs. t0: 9.92% vs. 6.77%, p ≤ 0.01; Ctrl vs. t1: 9.92% vs. 5.98%, p ≤ 0.001; Fig. 1E, supplementary Table 2). However, in the follow-up period (t2, t3) the percentage of Tregs increased significantly above initial levels (t0 vs. t2: 6.77% vs. 10.34%, p ≤ 0.01; t0 vs. t3: 6.77% vs. 10.06%, p ≤ 0.05; Fig. 1E, supplementary Table 3). In patients with locoregional recurrence Tregs remained below that of controls throughout the course of therapy (Fig. 1E, supplementary Table 4).
Kinetics of the proportion of CD3 +/CD56 + NKT cells before, during RCT and in the follow-up period
Similar to CD8+ T cells, NK-like T (NKT) cells increased significantly during therapy in patients without locoregional recurrences (t0 vs. t1: 4.41% vs. 7.18%, p ≤ 0.001; t0 vs. t3: 4.41% vs. 7.07%, p ≤ 0.01; Fig. 1F, supplementary Table 3). Notably, before start of therapy at t0, the percentage of NKT cells was almost twice as high compared to healthy controls (Ctrl vs. t0: 2.46% vs. 4.41%, p ≤ 0.05, Ctrl vs. t1: 2.46% vs. 7.18%, p ≤ 0.001; Ctrl vs. t2: 2.46% vs. 6.69%, p ≤ 0.001; Ctrl vs. t3: 2.46% vs. 7.07%, p ≤ 0.01; Fig. 1F, supplementary Table 2). Patients with locoregional recurrences showed stably higher proportions of NKT cells than controls throughout the therapy (Fig. 1F, supplementary Table 4).
Kinetics of the proportion of NK cell subsets (CD56+/CD69+, CD3−/CD56+ , CD3−/CD94+, CD3−/NKG2D+, CD3−/NKp30+, CD3−/NKp46+) and CD56bright/dim/CD16 NK subset analysis before, during RCT and in the follow-up period
Already during RCT the proportion of NK cells continuously increased up to t3 in patients without, but not in patients with locoregional recurrence (Fig. 2B). Before RCT at t0 none of the tested NK cell subpopulations differed between healthy controls and patients without locoregional recurrence (Fig. 2, supplementary Table 6). However, the percentage of CD3−/CD56+ NK cells in patients without locoregional recurrence increased significantly after RCT (t0 vs. t2: 11.1% vs. 15.83%, t0 vs. t3: 11.1% vs. 15.79%, p ≤ 0.01; Fig. 2B, supplementary Table 7) In contrast, patients with recurrence had lower proportions of all NK cell subpopulations already before RCT than controls that remained at low levels at all other timepoints (Fig. 2B-F, supplementary Table 8, 9).
Overall, the percentages of all NK cell subsets increased significantly after RCT in patients without locoregional recurrences (Fig. 2A-2F, supplementary Table 7). At t3, values of all NK cell subsets including those bearing activating receptors such as NKG2D, NKp30 and NKp46 were significantly higher than initial levels and those of healthy controls (Fig. 2A-2F, supplementary Tables 6, 7). In contrast, all NK cell subgroups in recurrent patients remained below that of patients without locoregional recurrence (Fig. 2A-2F, supplementary Table 8). Furthermore, recurrent patients had lower NK cell counts at time of locoregional recurrence (t5) compared to patients without recurrence at t3 (Fig. 2A-2F, supplementary Table 8).
A subgroup analysis of CD56bright/CD16+ NK cells in SCCHN patients without locoregional recurrence revealed significantly higher proportions of CD56bright/CD16+ NK cells compared to healthy controls, which dropped after application of 20 to 30 Gy at t1, but increased until t3 (Table 1A). A similar course was observed with respect to the CD56bright/CD16- NK cell subset in patients without locoregional recurrence. In recurrent patients both CD56bright NK cell subsets remained below that of patients without recurrence throughout the whole course of therapy (t0–t5) (Table 1B). In contrast, values of the CD56dim NK cell subset appeared to be elevated in recurrent patients (Table 1B).