Natural Killer Group 2D Ligand Depletion Reconstitutes Natural Killer Cell Immunosurveillance of Head and Neck Squamous Cell Carcinoma

Head and neck squamous cell carcinoma (HNSCC) is a highly heterogeneous and aggressive tumor originating from the epithelial lining of the upper aero-digestive tract accounting for 300,000 annual deaths worldwide due to failure of current therapies. The natural killer group 2D (NKG2D) receptors on natural killer (NK) cells and several T cell subsets play an important role for immunosurveillance of HNSCC and are thus targeted by tumor immune evasion strategies in particular by shedding of various NKG2D ligands (NKG2DLs). Based on plasma and tumor samples of 44 HNSCC patients, we found that despite compositional heterogeneity the total plasma level of NKG2DLs correlates with NK cell inhibition and disease progression. Strikingly, based on tumor spheroids and primary tumors of HNSCC patients, we found that NK cells failed to infiltrate HNSCC tumors in the presence of high levels of NKG2DLs, demonstrating a novel mechanism of NKG2DL-dependent tumor immune escape. Therefore, the diagnostic acquisition of the plasma level of all NKG2DLs might be instrumental for prognosis and to decipher a patient cohort, which could benefit from restoration of NKG2D-dependent tumor immunosurveillance. Along these lines, we could show that removal of shed NKG2DLs (sNKG2DLs) from HNSCC patients’ plasma restored NK cell function in vitro and in individual patients following surgical removal of the primary tumor. In order to translate these findings into a therapeutic setting, we performed a proof-of-concept study to test the efficacy of adsorption apheresis of sNKG2DLs from plasma after infusion of human MICA in rhesus monkeys. Complete removal of MICA was achieved after three plasma volume exchanges. Therefore, we propose adsorption apheresis of sNKG2DLs as a future preconditioning strategy to improve the efficacy of autologous and adoptively transferred immune cells in cellular cancer immunotherapy.

Solid tumor spheroids were generated as described previously (1) by seeding 5x10 3 -1x10 4 cells/well in a volume of 150 µl/well of culture medium in 96-well plates coated with 1.5% agarose in basal medium. Solid tumor spheroids formed within approximately 24-48 h after initial seeding (d0) and were used in functional assays. For tumor spheroid growth curves, phase contrast pictures of independent solid spheroids were analyzed for each condition from six independent experiments by Fiji software (2).

Tumor Spheroid Infiltration of NK Cell Subpopulations
For infiltration studies with NK cell subpopulations, primary human NK cells on d7 of culture were stained with anti-human CD56 (clone B159; BD Bioscience, USA) and sorted on a FACS ARIAII instrument (BD Bioscience, USA) for CD56 bright and CD56 dim subpopulations. Solid FaDu tumor spheroids were co-cultured at an E:T of 1:1 with the individual NK cell subpopulations for 48 h. For histology, individual spheroids (n > 8) were collected and embedded into OCT compound (Sakura, Japan). Cryosections were stained with mouse anti-human CD45 (clone 2B11+PD7/26; DAKO, Germany) and analyzed by standard two-step polymer methods with Envision kits (DAKO, Germany). For counterstaining, Hematoxylin (DAKO, Germany) was used according to manufacturer's instructions.

Cytotoxicity Assays with NKG2D Blocking Antibodies
For blocking experiments, NK cells were pre-incubated with mouse anti-human NKG2D antibodies (MAB139; R&D Systems, USA) for 30 minutes prior to co-incubation with CFSE-labeled HNSCC (co-incubation for 16 h) and SiHa (co-incubation for 2 h) cell lines at appropriate effector-to-target (E:T) ratios in NK cell medium without IL-2. NK cells were labeled with mouse anti-human CD45-APC antibodies (clone 5B1; Miltenyi Biotec, Germany) and live/dead cell discrimination was achieved by staining with SytoxBlue (Life Technologies, USA) prior to the measurement on a BD FACSCanto II instrument equipped with a 96-well plate HTS Sampler. Cellular event counts were analyzed with FlowJo software. The percentage of cell lysis was calculated by evaluation of viable (CFSE + /SytoxBlue -) and dead (CFSE + /SytoxBlue + ) target cells. Data is represented as mean ± SEM of triplicates. Statistical significance was determined by one-way ANOVA.
Rhesus monkey lymphocytes were stained from whole blood samples by incubation with a cocktail of conjugated mouse anti-human CD3-APC.Cy7 (clone HIT3a), anti-human CD20-FITC (clone 2H7; both BioLegend, USA) and anti-human CD314-APC (clone BAT221; Miltenyi Biotec, Germany) following erythrocyte lysis using ACK lysis buffer (Thermo Fisher Scientific, USA) according to manufacturer's instructions. Cells were measured on a LSRII instrument (BD Bioscience, USA) and analyzed with FlowJo software to calculate the MFI over background.

Deglycosylation Assay
To check for glycosylation, sMICA*04 was treated with either peptide-N-glycosidase F (PNGaseF) to remove N-linked glycans, or protein deglycosylation mix (both New England Biolabs, USA) for removal of N-and O-linked oligosaccharides according to the manufacturer's protocol. Deglycosylation was analyzed by molecular mass reduction via SDS-PAGE and immunoblot with mouse anti-MICA antibody (MAB1300; R&D Systems, USA) and anti-mouse IgG HRP-conjugated secondary antibody (Sigma-Aldrich, USA). Recombinant MICA*04 produced from E.coli as previously described (5) served as nonglycosylated control.

Peptide Spot Arrays
Peptide spot arrays of human MICA (Q29983 aa23-308; 18 amino acid peptides; off-set by 4 amino acids) were synthesized by Fmoc chemistry at activated PEG spacers on cellulose membranes by automated parallel peptide synthesis on a MultiPep RS instrument (Intavis, Germany) and used for binding experiments as described previously (6, 7). After blocking with 5% milk/TBS-T, peptide membranes were incubated with anti-MICA antibody (clone AMO1). Peptide membranes were analyzed by chemiluminescence imaging after incubation with anti-mouse IgG HRP-conjugated antibody (Sigma-Aldrich, USA).

Rhesus Antibody Detection
For the analysis of anti-MICA monkey IgG/IgM antibodies the animals' plasma were analyzed three weeks after sMICA*04 injection. Therefore, ELISA plates were coated with sMICA*04, blocked with BSA blocking solution (Candor, Germany) and probed with rhesus macaque plasma in different dilutions and subsequent washing steps with 0.05% Tween20-PBS. Bound antibodies were detected with either anti-monkey-IgG (SAB3700766) or antimonkey-IgM (SAB3700780) HRP-conjugated antibodies (both Sigma-Aldrich, USA). Following visualization with TMB substrate (KPL, Germany), signals were measured in a microtiter plate reader (λ = 450 nm) and analyzed with Prism 6 software (GraphPad, USA). Data of plasma samples were calculated as mean ± SEM.          Supplementary Figure S7