Patients diagnosed with locally advanced NSCLC receiving standard-of-care radiation therapy with concurrent chemotherapy (ccRTCT) were eligible for enrollment in the study. Patients were excluded if they had any of the following contraindications: recent (<6 months) stroke or myocardial infarction, severe heart failure (class III or IV), patients with an active implanted medical device (pacemaker, defibrillator or hearing aid implant), recurrent vasovagal syncope, unilateral or bilateral vagotomy, sick sinus syndrome (without a pacemaker), 2nd or 3rd-degree atrioventricular block and pregnancy or nursing. There were no differences between the two treatment groups in regards to age, tumor size, BMI and smoking history. The clinical study was approved by the local medical ethical committee (2018/016) and registered on ClinicalTrials.gov (identifier: NCT03553485). All patients signed informed consent before inclusion in the study.

All mice underwent submandibular blood sampling prior to (D₇), during (D₁₃) and after completion (D₂₀) of their treatments. Blood (200μl) was collected in heparin-coated tubes (Sarstedt, Germany), centrifuged for 10 minutes at 2000g to separate blood cell pellets from plasma. While cell pellets were immediately analyzed, plasma samples were stored at -20°C for further analysis. At day 20 after tumor injection, mice were sacrificed by cervical dislocation and single cell suspensions from lung and spleen were prepared. Lungs were first perfused with 5ml PBS and transferred to 1ml Roswell Park Memorial Institute-1640 medium (RPMI-1640, Sigma-Aldrich) containing 300U/ml collagenase-I (Sigma-Aldrich). Tissues were cut to small pieces using scissors, incubated at 37°C for 45 minutes, and finally mechanically reduced using an 18G syringe until single cell suspensions could be passed through a 40μm strainer. Spleens were transferred to 1ml PBS, stamped with the plunger of a 3cc syringe and passed through a 40μm strainer. Cell pellets from blood, lung and spleen, were resuspended in 1ml red blood cell lysis buffer, incubated for 5 minutes, followed by a centrifugation and wash step with PBS before further analysis.

Cell suspensions from murine lung tissue were cultured at 3 x 10⁵ cells/100µl complete RPMI medium with 30ng/ml recombinant IL-2 (Peprotech) in the presence of 3 x 10⁴ LLC cells. After 24 hours, supernatants were collected for evaluation of IFN-γ secretion via a mouse IFN-γ ELISA kit (Invitrogen, in accordance with the manufacturer’s guidelines). Cells were treated for an additional 4 hours with the protein transport inhibitor Golgi-stop (Monensin, BD Biosciences) prior to surface and intracellular staining for CD137, IL-2 and IFN-γ within the CD8⁺ T lymphocytes.

Peripheral blood (8ml) was collected in EDTA-coated tubes before (D₀), during (D₂₁) and at the end (D₄₂) of the RT regimen. To isolate the peripheral blood mononuclear cells (PBMCs), samples were centrifuged at 2200 rpm for 10 minutes. While plasma was immediately stored at -80°C for further analysis, PBMCs were purified using LeucoSep tubes according to the manufacturers’ instructions (Greiner Bio-One). In brief, cell pellets were diluted with equal volumes of PBS before transfer to LeucoSep tubes and centrifugation at room temperature for 15 min at 800 rcf without brake. Next, the PBMCs were collected, washed twice in PBS and aliquoted before cryopreservation in liquid nitrogen.

Concentrations of cytokines, chemokines and growth factors were evaluated on 12,5μl of murine plasma sample via cytokine bead array technology (BIO-RAD, Bio-Plex 200 System). More specifically the Bio-Plex Pro Mouse Chemokine Panel, 31-plex (BIO-RAD) and a Mouse TGF-β1 ELISA kit (Thermo Fisher Scientific) were used.

Concentrations of human TGF-β1, I-309/CCL1, ENA-78/CXCL5 and MDC/ADAM11 were measured using the respective human ELISA kits (Thermo Fisher Scientific) in accordance with the manufacturer’s instructions.

The large left lobe of each murine lung was fixed in buffered 4% formaldehyde and paraffin embedded. Tissue sections of 4μm were deparaffinized, hydrated and stained with haematoxylin, eosin and saffron (HES). The tissue sections were then dehydrated and mounted to allow histological evaluation. Immunohistochemistry images were acquired using a Leica DM 4000 microscope at 20x magnification.

Staining of surface markers was performed for 30 minutes at 4°C in cold PBS containing 1% bovine serum albumin (BSA, Sigma-Aldrich) and 0.02% sodium azide (FACS buffer; Sigma-Aldrich). The respective fluorescently labeled antibodies are listed in Supplementary Tables 1 and 2 ( Supplementary Material ). To avoid non‐specific antibody binding, Fc receptors were first blocked by incubating all samples with CD16/32 antibody (BD Biosciences). For intracellular staining, cells were subsequently fixed and permeabilized using Cytofix/Cytoperm™ kit (BD Biosciences). Therefore, cells were washed using Perm/Wash buffer, resuspended in Fixation/Permeabilization solution for 20 minutes at 4°C. Next, samples were washed twice using the Perm/Wash buffer prior to staining with the anti-arginase-1-PE antibody (R&D systems, Abingdon, United Kingdom), diluted in Perm/Wash buffer for 30 minutes at 4°C. The stained cells were evaluated on an LSRFortessa flow cytometer (Beckton Dickinson) while analysis was performed with the FlowJo 10.5.3 software (Tree Star Inc., Ashland, OR, USA).

Two-tailed unpaired t-tests were used to determine the significance of differences. A P value ≤ 0.05 was used as cut-off value for significance. Aggregated data are presented in figures using mean values to represent the central tendency and standard error of the mean (SEM) to represent variability. All statistical analysis was computed using GraphPad Prism v 7.0.

The delicate balance between immune stimulatory effector cells and pro-tumoral suppressive immunocytes represents a critical determinant for lung cancer patient prognosis. Therefore, we first investigated the impact of VNS monotherapy on this balance within the lung TME and periphery (blood and spleen) of lung-tumor-bearing mice. More specifically, 11 days after C57BL/6 mice were challenged i.v. with LLC-Fluc cells, VNS (25Hz) or sham treatment was executed twice daily, for 4 consecutive days ( Figure 1A ). Blood samples were collected at D₇ (baseline), 3 days after the first (D₁₃) and 6 days after the last (D₂₀) treatment, prior to euthanization. Aside from blood at D₂₀, perfused lungs and spleens were also collected to evaluate the following immune fractions in flow cytometry: CD8⁺ CTLs, CD4⁺ T cells, CD19⁺ B cells, CD56⁺ NK cells, CD11c⁺ DCs, SiglecF⁺ alveolar macrophages (AMs), F4/80⁺ TAMs, CD25⁺/CD127^- Tregs, Ly6G⁺ neutrophils and Ly6C⁺ monocytes. Gating strategies of lymphoid and myeloid cell populations are provided as Supplementary Material ( Figure S1 ).

No marked lymphoid differences were observed for the fractions of CD8⁺ CTLs ( Figure 1B ), CD4⁺ helper T cells ( Figure 1C ) between sham and VNS treated animals. Interestingly, VNS did result in a non-significant trend towards more NK cells in the lung TME ( Figure 1D ) and a significantly reduced percentage of splenic CD4⁺/CD25⁺/CD127^- Tregs ( Figure 1E ). To investigate the functional impact of VNS monotherapy on the tumor infiltrated CD8⁺ lymphocytes, we co-cultured lung tumor derived cells from sham and VNS treated animals with LLC cells at ratio 10:1 in vitro. Twenty-four hours later, expression of the following functional markers on the CD8⁺ TILs was assessed via flow cytometry: CD137 (4-1BB), IL-2, IFN-γ and PD-1. We observed a significant increase of IFN-γ and CD137 in the CTLs derived from VNS treated mice ( Figure 1F ). Moreover, within the supernatants a confirmative -yet not significant- trend for enhanced IFN-γ secretion upon VNS was shown by ELISA ( Figure 1G ), suggesting that VNS treatment ameliorated the cytotoxic profile of lung tumor infiltrated CTLs.

When we assessed changes in the myeloid compartment of blood, spleen and lung, VNS treatment did not seem to result in any significant changes in the abundance of DCs, AMs, TAMs nor neutrophils (data not shown) except for a non-significant trend towards more monocytes in the lung TME ( Figure 1H ). To further decipher the systemic immunomodulatory effects of VNS, we analyzed plasma levels of 32 cytokines and chemokines ( Figures 1I, J ). As depicted in Figure 1J , VNS treatment resulted in a significant reduction of I-309/CCL1 and MDC/CCL22, which is in line with peripheral reduction of Tregs, since both have been described to regulate Treg recruitment in murine cancer models (51–55). Although we did not see any changes in the neutrophilic fraction, multiplex analysis did show a significant reduction in the plasma levels of the neutrophil-attracting chemokine ENA-78/CXCL5 ( Figure 1J ). In addition, there was a non-significant trend towards reduced plasma levels of TGF-β, renowned for its negative effect on anti-tumor immunity by suppressing immune effector cell functions of, amongst others, CTLs, CD4⁺ T cells and NK cells while promoting the generation and recruitment of Tregs (56–61).

Using a similar experimental setup and lung tumor model as depicted in Figure 1A , we previously observed that fractionated RT results in significant reduction of lung tumor growth (22). However, we and others showed that fractionated RT results in a reduction of the tumor infiltrated CD8⁺ CTL fraction, while the abundance of immunosuppressive myeloid cells increased. To investigate the impact of VNS on these fluctuations, LLC-Fluc bearing mice were treated with RT in combination with sham (four consecutive daily fractions of 3.2 Gy) or in combination with VNS (25Hz) 2 times daily ( Figure 1A ). Flow cytometric analysis of blood, spleen and the lung TME showed that VNS had no impact on the RT-induced reduction of CD8⁺ T cells ( Figure 2A ) nor on the RT-mediated rise in tumor infiltrated CD4⁺ T cells ( Figure 2B ). In addition, no significant changes were found for the numbers of NK cells nor Tregs between mice treated with RT + sham and RT + VNS ( Figures 2C, D ). Notably, VNS did ameliorate the RT-induced increase of IFN-γ⁺ TILs and VNS of RT treated animals further resulted in a significantly reduced expression of PD-1 in the lung TME-residing CTLs ( Figure 2E ). While for the myeloid subsets evaluated in the periphery and lung TME, no significant changes were found, the levels of splenic CD11b⁺ Ly6G^- Ly6C^hi monocytes did reduce significantly upon RT + VNS compared to RT + sham ( Figure 2F ). Finally, no significant changes in any of the 32 different chemo- and cytokines were found in the plasma of mice treated with RT + sham and RT + VNS (data not shown). To evaluate the therapeutic impact of VNS alone or on RT treatment, orthotopic lung tumor growth was assessed via BLI ( Figures 3A, B ) and immunohistochemistry ( Figures 3C, D ). VNS monotherapy as well as in combination with RT did not result in tumor growth reduction compared to sham alone or RT + sham respectively ( Figures 3B, D ).

To gain more insights into the potential immunomodulatory effect of VNS on RT treated NSCLC patients, we performed a pilot clinical study. Here we analyzed the effect of VNS on the composition of PBMCs in patients with locally advanced NSCLC receiving ccRTCT. We collected blood samples from 7 patients (n=3 standard ccRTCT, n=4 standard ccRTCT + VNS) before (D₁), during (D₂₁) and at the end of their treatment (D₄₂) ( Figure 4A ) to analyze cellular and molecular changes. Patients’ and tumor characteristics are summarized in Table 1 .

To determine the numbers of circulating T lymphocytes, DCs, NK cells, monocytes (CD14⁺CD15⁻HLA-DR^lo/–) and neutrophils (CD11b⁺CD14⁻CD15⁺), we performed flow cytometry on freshly isolated blood samples. The gating strategies of the different lymphoid and myeloid cell populations are provided as Supplementary material ( Figure S2 ). In line with our preclinical murine findings, the proportions of CD8⁺ and CD4⁺ T cells did not differ between the standard ccRTCT group and the investigational ccRTCT + VNS group ( Figures 4B, C ). While we did find a reduction in splenic Tregs upon VNS treatment of tumor bearing animals, no changes in murine blood Tregs were found, again in line with the lack of changes in blood Tregs upon VNS treatment of NSCLC patients ( Figure 4D ). To test the functionality of CD8⁺ T cells, we determined the expression of CD107a, IFN-γ, Granzyme-B and PD-1, showing no substantial differences between the two treatment groups (data not shown). The number of anti-tumoral NK cells and DCs within the fraction of blood mononuclear cells remained stable during the 6 weeks of ccRTCT while these populations slightly increased in the ccRTCT + VNS group ( Figures 4E, F ). Whereas VNS had systemic immunomodulatory effects on the following 3 murine chemokines, I-309/CCL1, MDC/CCL22 and ENA-78/CXCL5, we found no indication for an effect on their human isoform (data not shown) nor any changes in the fraction of monocytes ( Figure 4G ). Comparing levels of neutrophils, we found a more pronounced increase of neutrophils in the standard treatment group compared to the investigational group ( Figure 4H ). In advanced NSCLC, peripheral blood biomarkers including the tumor marker carcinoembryonic antigen (CEA) and the neutrophil over lymphocyte ratio (NLR) have been proposed as prognostic biomarkers useful for treatment monitoring (62). In particular, higher CEA and NLR have shown a significant association with worse outcomes (15, 62, 63). When serum levels of CEA were extracted from the electronic patient record system, no differences in the frequency between NSCLC patients receiving the conventional treatment compared to the experimental combination treatment were found ( Figure 4I ). While ccRTCT treated patients showed a slight increase in NLR, this level decreased in the ccRTCT + VNS group, suggestive for a trend towards VNS-installed reduction of the NLR ( Figure 4J ).