Received: 13 April 2021 Revised: 17 June 2021 Accepted: 21 June 2021 Published online: 14 July 2021 DOI: 10.1002/ctm2.491 RESEARCH ARTICLE Blood biomarkers associated to complete pathological response on NSCLC patients treated with neoadjuvant chemoimmunotherapy included in NADIM clinical trial Raquel Laza-Briviesca1,2 Alberto Cruz-Bermúdez1 Ernest Nadal3 Amelia Insa4 María del Rosario García-Campelo5 Gerardo Huidobro6 Manuel Dómine7 Margarita Majem8 Delvys Rodríguez-Abreu9 Alex Martínez-Martí10 Javier De Castro Carpeño11 Manuel Cobo12 Guillermo López Vivanco13 Edel Del Barco14 Reyes Bernabé Caro15 Nuria Viñolas16 Isidoro Barneto Aranda17 Santiago Viteri18 Bartomeu Massuti19 Marta Casarrubios1,2 Belén Sierra-Rodero1,2 Carlos Tarín20 Aránzazu García-Grande21 Cara Haymaker22 Ignacio I. Wistuba22 Atocha Romero1 Fernando Franco1 Mariano Provencio1 1 Servicio de Oncología Médica, Instituto de Investigación Sanitaria Puerta de Hierro-Segovia de Arana (IDIPHISA), Hospital Universitario Puerta de Hierro-Majadahonda, Madrid, Spain 2 Doctoral School, Universidad Autónoma de Madrid, Madrid, Spain 3 Institut Català d’Oncologia, L’Hospitalet De Llobregat, Barcelona, Spain 4 Fundación INCLIVA, Hospital Clínico Universitario de Valencia, Valencia, Spain 5 Hospital Universitario A Coruña, A Coruña, Spain 6 Hospital Universitario de Vigo, Pontevedra, Spain 7 Hospital Universitario Fundación Jiménez Díaz, Madrid, Spain 8 Hospital de la Santa Creu i Sant Pau, Barcelona, Spain 9 Hospital Insular de Gran Canaria, Las Palmas, Spain 10 Hospital Universitario e Instituto de Oncología Vall dťHebron (VHIO), Barcelona, Spain 11 Hospital Universitario La Paz, Madrid, Spain 12 Hospital Universitario Regional de Málaga, Málaga, Spain 13 Hospital Universitario Cruces, Barakaldo, Spain 14 Hospital Universitario de Salamanca, Salamanca, Spain 15 Hospital Universitario Virgen del Rocio, Seville, Spain 16 Hospital Clínic, Barcelona, Spain 17 Hospital Universitario Reina Sofia, Córdoba, Spain 18 Instituto Oncológico Dr. Rosell, Hospital Universitario Quiron Dexeus, Grupo QuironSalud, Barcelona, Spain 19 Hospital General de Alicante, Alicante, Spain 20 Bioinformatics Unit, Instituto de Investigación Sanitaria Puerta de Hierro-Segovia de Arana, Madrid, Spain 21 Flow Cytometry Core Facility, Instituto de Investigación Sanitaria Puerta de Hierro-Segovia de Arana (IDIPHISA), Madrid, Spain 22 Departments of Translational Molecular Pathology, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA Clin. Transl. Med. 2021;11:e491. wileyonlinelibrary.com/journal/ctm2 https://doi.org/10.1002/ctm2.491 LAZA-BRIVIESCA et al. Correspondence Mariano Provencio and Alberto Cruz Graphical Abstract Bermudez, Medical Oncology Depart- ment, Puerta de Hierro Hospital, Manuel de Falla Street #1, 28222 Majadahonda, Madrid, Spain. Email: mprovenciop@gmail.com; alberto.cruz.bermudez@gmail.com 1. Non-small cell lung cancer (NSCLC) patients achieving complete pathologic response (CPR) after neoadjuvant chemoimmunotherapy seem to have a distinc- tive peripheral blood immune status at diagnosis and surgery. 2. At diagnosis, CPR patients are characterized by a stronger previously induced immune response with a higher cytotoxic profile and lower levels of inhibitory cytokines and cells. 3. This exploratory analysis of the NADIM study supports the use of blood as a valid source for response biomarkers and may serve as a first step to elucidate response mechanisms to chemoimmunotherapy in NSCLC. Received: 13 April 2021 Revised: 17 June 2021 Accepted: 21 June 2021 Published online: 14 July 2021 DOI: 10.1002/ctm2.491 RESEARCH ARTICLE Blood biomarkers associated to complete pathological response on NSCLC patients treated with neoadjuvant chemoimmunotherapy included in NADIM clinical trial Raquel Laza-Briviesca1,2 Alberto Cruz-Bermúdez1 Ernest Nadal3 Amelia Insa4 María del Rosario García-Campelo5 Gerardo Huidobro6 Manuel Dómine7 Margarita Majem8 Delvys Rodríguez-Abreu9 Alex Martínez-Martí10 Javier De Castro Carpeño11 Manuel Cobo12 Guillermo López Vivanco13 Edel Del Barco14 Reyes Bernabé Caro15 Nuria Viñolas16 Isidoro Barneto Aranda17 Santiago Viteri18 Bartomeu Massuti19 Marta Casarrubios1,2 Belén Sierra-Rodero1,2 Carlos Tarín20 Aránzazu García-Grande21 Cara Haymaker22 Ignacio I. Wistuba22 Atocha Romero1 Fernando Franco1 Mariano Provencio1 1Servicio de Oncología Médica, Instituto de Investigación Sanitaria Puerta de Hierro-Segovia de Arana (IDIPHISA), Hospital Universitario Puerta de Hierro-Majadahonda, Madrid, Spain 2 Doctoral School, Universidad Autónoma de Madrid, Madrid, Spain 3 Institut Català d’Oncologia, L’Hospitalet De Llobregat, Barcelona, Spain 4 Fundación INCLIVA, Hospital Clínico Universitario de Valencia, Valencia, Spain 5 Hospital Universitario A Coruña, A Coruña, Spain 6 Hospital Universitario de Vigo, Pontevedra, Spain 7 Hospital Universitario Fundación Jiménez Díaz, Madrid, Spain 8 Hospital de la Santa Creu i Sant Pau, Barcelona, Spain 9 Hospital Insular de Gran Canaria, Las Palmas, Spain 10 Hospital Universitario e Instituto de Oncología Vall dťHebron (VHIO), Barcelona, Spain 11 Hospital Universitario La Paz, Madrid, Spain 12 Hospital Universitario Regional de Málaga, Málaga, Spain 13 Hospital Universitario Cruces, Barakaldo, Spain 14 Hospital Universitario de Salamanca, Salamanca, Spain 15 Hospital Universitario Virgen del Rocio, Seville, Spain 16 Hospital Clínic, Barcelona, Spain 17 Hospital Universitario Reina Sofia, Córdoba, Spain 18 Instituto Oncológico Dr. Rosell, Hospital Universitario Quiron Dexeus, Grupo QuironSalud, Barcelona, Spain 19 Hospital General de Alicante, Alicante, Spain 20 Bioinformatics Unit, Instituto de Investigación Sanitaria Puerta de Hierro-Segovia de Arana, Madrid, Spain This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2021 The Authors. Clinical and Translational Medicine published by John Wiley & Sons Australia, Ltd on behalf of Shanghai Institute of Clinical Bioinformatics Clin. Transl. Med. 2021;11:e491. wileyonlinelibrary.com/journal/ctm2 1 of 16 https://doi.org/10.1002/ctm2.491 2 of 16 LAZA-BRIVIESCA et al. 21 Flow Cytometry Core Facility, Instituto de Investigación Sanitaria Puerta de Hierro-Segovia de Arana (IDIPHISA), Madrid, Spain 22 Departments of Translational Molecular Pathology, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA Correspondence Mariano Provencio and Alberto Cruz Abstract Bermudez, Medical Oncology Depart- Background: Immunotherapy is being tested in early-stage non-small cell lung ment, Puerta de Hierro Hospital, Manuel de Falla Street #1, 28222 Majadahonda, cancer (NSCLC), and achieving higher rates of complete pathological responses Madrid, Spain. (CPR) as compared to standard of care. Early identification of CPR patients has Email: mprovenciop@gmail.com; vital clinical implications. In this study, we focused on basal peripheral immune alberto.cruz.bermudez@gmail.com cells and their treatment-related changes to find biomarkers associated to Raquel Laza-Briviesca and Alberto Cruz- CPR. Bermúdez contributed equally. Methods: Blood from 29 stage IIIA NSCLC patients participating in the NADIM Present address Carlos Tarín, Basic Medical Sciences, trial (NCT03081689) was collected at diagnosis and post neoadjuvant treatment. Faculty of Medicine, Universidad CEU San More than 400 parameters of peripheral blood mononuclear cells (PBMCs) phe- Pablo, Madrid, Spain. notype and plasma soluble factors were analyzed. Funding information Results: Neoadjuvant chemoimmunotherapy altered more than 150 immune European Social Fund, Grant/Award parameters. At diagnosis, 11 biomarkers associated to CPR were described, with Numbers: PEJ16/MED/AI-1972, PEJD- an area under the ROC curve >0.70 and p-value <.05. CPR patients had sig- 2018-PRE/SAL-8641, PEJD-2019- PRE/BMD-17006; Spanish Lung Can- nificantly higher levels of CD4+ PD-1+ cells, NKG2D, and CD56 expression on cer Group; Bristol-Myers Squibb; Hori- T CD56 cells, intensity of CD25 expression on CD4+ CD25hi+ cells and CD69 zon 2020 Framework Programme, expression on intermediate monocytes; but lower levels of CD3+ CD56– CTLA- Grant/Award Number: CLARIFY 875160; Ministerio de Ciencia e Innovación, 4+ cells, CD14++ CD16+ CTLA-4+ cells, CTLA-4 expression on T CD56 cells and Grant/Award Numbers: RTC2017- lower levels of b-NGF, NT-3, and VEGF-D in plasma compared to non-CPR. Post 6502-1, RTC2019-007359-1; Instituto de Salud Carlos III, Grant/Award Num- treatment, CPR patients had significantly higher levels of CD19 expression on B bers: CD19/00170, PI19/01652; European cells, BCMA, 4-1BB, MCSF, and PARC and lower levels of MPIF-1 and Flt-3L in Regional Development Fund (ERDF), plasma compared to non-CPR. Grant/Award Numbers: CD19/00170, PI19/01652 Conclusions: Patients achieving CPR seem to have a distinctive peripheral blood immune status at diagnosis, even showing different immune response to treatment. These results reinforce the different biology behind CPR and non-CPR responses. KEYWORDS biomarkers, chemoimmunotherapy, immune cells, neoadjuvant, non-small cell lung cancer 1 INTRODUCTION logical responses (CPR) near 4%,4 and 2-year overall sur- vival (OS) rates of 60%.5 Non-small cell lung cancer (NSCLC) accounts for 80%–85% Cancer immunology knowledge has remarkably of all lung cancer cases, being the primary cause of cancer- advanced with the discovery of immune checkpoint related death worldwide and the second by incidence.1 proteins. The immune checkpoint PD-1 in lymphocytes6 Over 30% of NSCLC patients are diagnosed at locally and myeloid cells7 binds to PD-L1 in the tumor cells, advanced stage known as stage III.2 Specifically, patients leading to immune cell inactivation and tumor escape.8 with resectable stage III disease, which are potentially This has led to anti-PD-1 or anti-PD-L1 antibodies use curable, are usually treated with multimodal neoadjuvant against cancer, bringing a new era of treatments known as treatment, based on chemotherapy and surgery.3 How- immunotherapies.9 The adaptive immune response lead ever, clinical outcomes of current treatments based on by T cells plays a major role on antitumor responses10 ; chemotherapy are still discouraging, with complete patho- however, macrophages,11 natural killer (NK) cells,12 B LAZA-BRIVIESCA et al. 3 of 16 cells,13,14 or NK-T cells15 are also involved in antitumor samples from 30 patients at baseline (pretreatment) and responses. 34 patients after three cycles of neoadjuvant treatment Immunotherapy treatment has become the standard of with nivolumab plus carboplatin (posttreatment) were treatment for advanced stages and is now being studied obtained. Blood samples at both time points were avail- for early and locally advanced stages.16 Recently, the com- able for 29 patients. Informed consent for the collection of bination of nivolumab (anti-PD-1) and platinum-based research samples and study protocol were approved by the chemotherapy as neoadjuvant treatment for resectable clinical research ethics committee of Hospital Puerta de stage III (NADIM clinical trial, NCT03081689) has shown Hierro and the Spanish Lung Cancer Group (SLCG) Board 63% of CPR for resected patients, and 2-year OS of 90%.17 in accordance with the International Conference on Har- These results open the possibility of transforming locally monization Guidelines on Good Clinical Practice and the advanced NSCLC to a curable disease in a substantial per- Declaration of Helsinki. Clinical characteristics of patients centage of patients, as CPR may be predictive of long-term with blood sample available for this study are described in survival following neoadjuvant therapy.18,19 Table S1. Hospitals and patients recruited are listed in Table Consequently, there is a clinical need for predictive S2. biomarkers to identify such patients achieving CPR, which Pathologic response evaluation was carried out as could potentially be free of disease after neoadjuvant treat- previously described following international guidelines ment. Current biomarkers such as PD-L1 tumor propor- recommendations.17,27 We classified patients in complete tion score (TPS) and Tumor mutational burden (TMB) pathological response (CPR, 0% of viable tumor cells remain controversial in clinical practice for anti-PD-1 in tumor bed or any lymph node tested) and non-CPR monotherapies.20,21 Moreover, their value for combined (patients with any percentage of viable tumor cells in resec- chemoimmunotherapy is more limited,22 and its power to tion specimens). Patients that did not undergo surgery identify patients achieving CPR insufficient.17,23 were excluded from the pathological response analysis; Thus, additional predictive biomarkers beyond PD-L1 however, they were included in comparisons of treatment and TMB are required. In this way, an increasing num- effect and correlation analysis. Additionally, PD-L1 TPS ber of investigations are focusing on the host antitumor and TMB were retrieved from previous report17 and cor- immune response. Accordingly, patients achieving CPR related with blood parameters analyzed at diagnosis from showed a distinct tumor infiltrating lymphocytes (TILs) this study. component at surgery evaluation.17 Additionally, an asso- ciation between TILs24 or peripheral blood mononuclear cells (PBMCs) characteristics,25,26 with clinical benefit of 2.2 Cell and plasma isolation lung cancer patients under immunotherapy, has been described. These differences seem to indicate that CPR Peripheral blood was diluted 1:1 in 1640 RPMI (Corn- may also be reflected in patient’s immune system at the ing, NY, USA), and mononuclear cells (PBMCs) were iso- peripheral level. This would allow a noninvasive alterna- lated by Lymphoprep (Stemcell, Vancouver, Canada) den- tive to tissue biopsy in predicting and monitoring treat- sity gradient centrifugation and cryopreserved using freez- ment outcomes. ing medium (1:1 RPMI/FBS, Linus, Spain) containing 10% Here, we present the first research study focused on DMSO (Carl Roth, Germany) until use. Plasma fraction the composition and phenotypic characteristics of differ- was collected after density gradient centrifugation was per- ent peripheral immune cell subpopulations and plasma formed and stored at −80◦ C until use. factors in NSCLC patients receiving chemoimmune neoad- juvant therapy, in order to characterize treatment effective- ness and identify biomarkers of CPR. 2.3 Immunophenotyping of PBMCs Cryopreserved pretreatment and posttreatment PBMCs 2 MATERIALS AND METHODS were thawed in parallel, washed with 5% FBS in 1× PBS, and surface stained with CD3-PerCP (clone BW264/56), 2.1 Patients and samples CD4-Viogreen (clone REA623), CD8-APCVio770 (clone BW135/80), CD14-APC (clone Tük4), CD16-Viogreen Patients were selected from NADIM clinical trial: neoad- (clone REA423), CD19-FITC (clone LT19), CD56- juvant chemotherapy plus nivolumab in resectable stage APCVio770 (clone REA196), CD69-FITC (clone REA824), IIIA non-small-cell lung cancer, an open-label, multicen- CD107a-FITC (clone H4A3), CTLA4-PE (clone BNI3), ter, single-arm, phase 2 trial.17 All patients with available PD1-PEVio770 (clone PD1.3.1.3), NKG2D-PE (clone peripheral blood sample were included. Peripheral blood REA797), NKp44-PEVio770 (clone 2.29) from Miltenyi 4 of 16 LAZA-BRIVIESCA et al. Biotec (Germany) and CD25-PeCy7 (clone M-A251) from curve analysis. Relationship between variables was done BD Bioscience (NJ, USA). Detection of PD-1 using anti- by Spearmanťs correlation considering R > ±0.80 as strong PD-1 clone PD1.3.1.3 was almost technically abrogated in linear correlation with significant p-value. For all statisti- posttreatment samples due to PD-1 binding of nivolumab cal tests, p-value <.05 was considered statistically signifi- as previously described.28 However, percentage of PD-1+ cant; *, **, and *** indicated p-values <.05, <.01, and <.001, cells and PD-1 MFI in PD-1+ cells was still analyzed in respectively. For p-values between .05 and .1, the numeri- posttreatment samples to discard possible differences cal value is included. Because the research was designed between groups. as a discovery study, p-values were not adjusted in order to Cells were acquired on a MACS Quant 10 cytometer and maximize the finding of new biomarkers and the genera- DAPI (ThermoFisher, CA, USA) staining prior acquisition tion of new hypothesis. IBM SPSS Statistics 25 was used for was used to exclude dead cells during analysis on FlowJo all statistical analyses. All authors had access to the data V10 software. The median number of viable cells after and certify the accuracy of the results presented. freezing, thawing, and staining was 93.6%, IQR 91.2–95.2. A minimum of 100,000 lymphocyte events were recorded. Expression of surface markers was evaluated by percent- 3 RESULTS age of positive cells and median fluorescence intensity (MFI) using fluorescence minus one (FMO) controls. Gat- 3.1 Patient characteristics ing strategy for all panels is shown in Figure S1A–F. Flow cytometry analysis was performed only in paired samples No statistical differences on age, sex, smoking status, (29 patients). histology, and lymph node involvement were observed Immunophenotyping of circulating NK cells according to the degree of pathological response (Table 1). (CD3– CD56+ ), T cells (CD3+ CD56– ), cytotoxic T cells Patients included in this study showed similar clinical (CD3+ CD8+ ), helper T cells (CD3+ CD4+ ), T CD56 cells characteristics to NADIM trial complete cohort (Table S1). (CD3+ CD56+ ), B cells (CD3– CD19+ ), and monocytes At baseline and prior to surgery, similar proportions of (CD14+ ) was determined. circulating NK cells (p-value .864), T cells (p-value .845), CD56+ T cells (p-value .306), B cells (p-value .262), and monocytes (p-value .124) were found between CPR and 2.4 Immunoassays for detecting non-CPR patients (Figure S2A). However, deepening into cytokines and soluble factors in plasma their phenotype and activation levels, we found significant differences described below. Sixteen key soluble factors from plasma were measured in duplicate using the human immuno-oncology check- point protein panel (Cat. #HCKPMAG-11K, Millipore, MA, 3.2 Association of immunocheckpoints USA) following manufacturer’s instructions. Plasma lev- expressed in PBMCs and pathological els of 200 cytokines were measured on the same sam- response ples using G-Series Human Cytokine Antibody Array 4000 (RayBiotech, GA, USA) following manufacturer’s instruc- 3.2.1 PD-1 expression on lymphocytes tions. Cytokine array and soluble factor analysis were per- formed on all samples. PD-1 expression and percentages of PD-1+ cells on differ- ent lymphocytes subsets, as T cells, CD4+ T cells, CD8+ T cells, CD56+ T cells, and NK cells were measured. At 2.5 Statistical analysis diagnosis, CPR patients showed a higher percentage of PD- 1+ cells compared to non-CPR patients, for all lymphocyte Raw data generated from flow cytometry and cytokine subsets, reaching to statistically significant differences for analysis is available in Table S3. Nonparametric Wilcoxon CD4+ PD-1+ cells (p-value .045). PD-1+ cells were almost signed rank test was used for analysis of treatment vari- technically undetectable after initiating the treatment with ation and Mann–Whitney U-test was performed to deter- chemotherapy plus nivolumab (Figure 1A). The percent- mine differences between CPR and non-CPR patients. age of CD3+ CD4+ PD-1+ cells at diagnosis had a predic- Pearson’s chi-square test was used for association between tive value to distinguish CPR and non-CPR patients, with categorical groups (for baseline characteristics with patho- an AUC of 0.728 (p-value .045) (Figure 1B). No differences logical response). We considered area under the curve between pathologic response groups were found on MFI (AUC) >0.70 with significant p-value as relevant predictive PD-1 expression neither at diagnosis nor after neoadjuvant biomarker using receiver operating characteristic (ROC) treatment (Figure S2B). LAZA-BRIVIESCA et al. 5 of 16 Differences between CPR and non-CPR at diagnosis and post neoadjuvant treatment Percentage of lymphocytes subpopulations PD-1+ A T cells CD4 T cells CD8 T cells T-NK like NK cells B CD3+CD56-PD1+ CD3+CD4+PD1+ CD3+CD8+PD1+ CD3+CD56+PD1+ CD3-CD56+PD1+ 35 ns ROC curve of CD3+CD4+PD-1+ 30 Percentage of PD-1+ cells 1.0 25 ns ns * ns ns ns 0.07 ns ns 0.8 20 Sensitivity 0.6 15 AUC ROC 0.728 0.4 p-value 0.045 10 Cut-off >5.31 0.2 Sensitivity 67% 5 Specificity 92% 0.0 0 0.0 0.2 0.4 0.6 0.8 1.0 PRE POST PRE POST PRE POST PRE POST PRE POST 1 - Specificity MFI NKG2D on T lymphocytes and NK cells C T cells CD3+CD56-NKG2D+ CD4 T cells CD3+CD4+NKG2D+ CD8 T cells CD3+CD8+NKG2D+ T-NK like CD3+CD56+NKG2D+ NK cells CD3-CD56+NKG2D+ D ROC curve of MFI NKG2D 1350 0.06 ns ns 0.05 ns ns ns ns on CD3+CD56+NKG2D+ 1200 * * 1.0 1050 MFI of NKG2D+ 900 0.8 750 Sensitivity 0.6 600 AUC ROC 0.786 450 0.4 p-value 0.012 300 Cut-off > 541 0.2 Sensitivity 60% 150 Specifity 92% 0.0 0 PRE POST POST PRE POST PRE POST PRE POST 0.0 0.2 0.4 0.6 0.8 1.0 PRE 1 - Specificity Neurotrophins and angiogenesis factors E b-NGF NT-3 VEGF-D F ROC curve of 11000 b-NGF, NT-3 and VEGF-D ns ns ns 10000 * * * Levels of soluble expression 1.0 9000 8000 0.8 7000 Sensitivity 6000 0.6 5000 NT-3 4000 0.4 b-NGF 3000 VEGF-D 2000 0.2 1000 0.0 0 PRE POST PRE POST PRE POST 0.0 0.2 0.4 0.6 0.8 1.0 1 - Specificity CPR Non-CPR Diagnosis Post-treatment F I G U R E 1 PD-1 and NKG2D immunocheckpoints and pathological response. Differences between CPR and non-CPR patients at diagnosis and post neoadjuvant treatment. (A) Percentage of PD-1+ cells on T cells (CD3+ CD56– PD-1+ ), CD4 T cells (CD3+ CD4+ PD-1+ ), CD8 T cells (CD3+ CD8+ PD-1+ ), T-NK like (CD3+ CD56+ PD-1+ ) cells, and NK cells (CD3– CD56+ PD-1+ ). (B) AUC ROC curve for CD3+ CD4+ PD-1+ . (C) MFI of NKG2D on T cells (CD3+ CD56– NKG2D+ ), CD4 T cells (CD3+ CD4+ NKG2D+ ), CD8 T cells (CD3+ CD8+ NKG2D+ ), T-NK like (CD3+ CD56+ NKG2D+ ) cells, and NK cells (CD3– CD56+ NKG2D+ ). (D) AUC ROC curve for CD3+ CD56+ NKG2D+ . (E) Levels of b-NGF, NT-3, and VEGF-D in plasma. AUC ROC curve for b-NGF, NT-3, VEGF-D. Legend: Diagnosis (PRE, grey dots), prior to surgery (POST, black dots), CPR patients (yellow bars), and non-CPR patients (purple bar). Nonparametric Mann–Whitney test was used for comparisons (ns, not significant; *p-value <.05; **p-value <.01; ***p-value <.001) and receiver operating characteristic (ROC) curve analysis showing AUC ROC curve, p-value, and optimum cut-off 6 of 16 LAZA-BRIVIESCA et al. TA B L E 1 Baseline characteristics of patients according to pathological response p-Value (CPR Total (N = 27) CPR (N = 15) Non-CPR (N = 12) vs. non-CPR) Age (years)a 67 (59–72) 69 (60–70) 64 (57.5–72.5) .695 b Sex .756 Male 21 (77.8) 12 (80) 9 (75) Female 6 (22.2) 3 (20) 3 (25) Cigarette-smoking historyb .795 Former smoker (≥1 year) 15 (55.6) 8 (53.3) 7 (58.3) Current smoker 12 (44.4) 7 (46.7) 5 (41.7) b Histology .115 Adenocarcinoma 15 (55.6) 9 (60) 6 (50) Squamous 9 (33.3) 3 (20) 6 (50) NOS/undifferenciated 3 (11.1) 3 (20) 0 b Nodes .984 N0 7 (25.9) 4 (26.7) 3 (25) N1 2 (7.4) 1 (6.7) 1 (8.3) N2 18 (66.7) 10 (66.7) 8 (66.7) Note: Age, sex, smoking status, histology, and affected lymph nodes comparison between pathological responses groups (CPR, complete pathological response) and non-CPR using Pearsonťs chi-square test. p-Value of chi-square or Kruskall–Wallis when adequate. a Median and interquartile range (IQR). b N (%). 3.2.2 NKG2D expression on T lymphocytes 0.728 (p-value .045) for VEGF-D (Figure 1F and Figure and NK cells S3F–H). When analyzing NKG2D expression, we found that at diag- nosis, NKG2D expression was higher in CD56+ T cells in 3.2.3 CTLA-4 expression on lymphocytes CPR patients compared to non-CPR patients (p-value .012) and there was a trend for CD8+ T cells (p-value .051) (Fig- We analyzed CTLA-4+ surface expression on T cells, ure 1C). Post neoadjuvant treatment, higher expression of CD56+ T cells, NK cells, B cells, and monocytes and, NKG2D on CD4+ T cells was found in CPR patients (p- despite of the scarcity of CTLA-4+ cells in peripheral value .017) (Figure 1C). NKG2D expression at diagnosis on blood, significant differences were found on NK cells and CD3+ CD56+ NKG2D+ cells showed an AUC of 0.786 (p- monocytes (Figure 2). We observed lower pretreatment value .012) (Figure 1D). percentages of CTLA-4+ NK cells (CD3– CD56+ CTLA- On the contrary, no significant differences at diagno- 4+ ) in patients with CPR (Figure 2A). Similarly, lower sis or post treatment in the percentage of NKG2D+ cells percentages of CTLA-4+ classical monocytes (p-value for T cells, CD56+ T cells, and NK cells between CPR and .026) and nonclassical monocytes (p-value .071) were non-CPR patients were obtained. However, a trend toward determined in CPR patients (Figure 2C). At diagnosis, higher percentage of CD8+ NKG2D+ cells was observed in higher expression of CTLA-4 was found on CD56+ CPR patients (p-value .07) (Figure S2C). T cells (CD3+ CD56+ CTLA-4+ ), but not on T cells Furthermore, we noticed that NKG2D expression (CD3+ CD56– CTLA-4+ ) in non-CPR patients (Figure 2E). inversely correlate with plasma levels of neurotrophin 3 These three subsets of cells showed an AUC ROC of 0.750 (NT-3) on CD8 T cells and T CD56 cells, as well as with (p-value .028), 0.753 (p-value .026), and 0.844 (p-value vascular endothelial growth factor D levels (VEGF-D) on .011), respectively (Figure 2B,D,F). T CD56 cells (Figure S3A–D). Lower levels of these factors Summarizing these initial results, our study and the nerve growth factor beta (b-NGF) were found describes the relevance of PD-1, CTLA-4, and NKG2D on plasma from CPR patients at diagnosis (Figure 1E). immunocheckpoints on immune cells subsets and, even Moreover, these soluble factors can be used as biomarkers greater, their value as biomarkers associated to complete associated to CPR at diagnosis, shown by an AUC of 0.750 pathological response to neoadjuvant chemoimmunother- (p-value .028) for b-NGF, 0.778 (p-value .015) for NT-3, and apy in lung cancer patients. LAZA-BRIVIESCA et al. 7 of 16 Differences between CPR and non-CPR at diagnosis and post neoadjuvant treatment A Percentage of lymphocytes subpopulations CTLA-4+ B T cells CD4 T cells CD8 T cells T-NK like NK cells B cells CD3+CD56-CTLA-4+ CD3+CD4+CTLA-4+ CD3+CD8+CTLA-4+ CD3+CD56+CTLA-4+ CD3-CD56+CTLA-4+ CD3-CD19+CTLA-4+ 16 ROC curve of CD3-CD56+CTLA-4+ ns ns ns ns ns ns ns ns ns 13 * 1.0 Percentage of CTLA-4 + cells 10 ns ns 7 0.8 4 Sensitivity 0.6 1 1.0 0.4 AUC ROC 0.750 0.8 p-value 0.028 0.6 Cut-off < 0.195 0.2 Sensitivity 67% 0.4 Specifity 92% 0.2 0.0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 PRE POST PRE POST PRE POST PRE POST PRE POST PRE POST 1 - Specificity C Percentage of monocytes subpopulations CTLA-4+ D Classical Intermediate Non-classical ROC curve of CD14++CD16+CTLA4+ 100 ns ns 0.07 ns 70 1.0 ns * Percentage of CTLA-4+ cells 40 10 0.8 10 Sensitivity 7 0.6 4 AUC ROC 0.753 0.4 p-value 0.026 1 1.0 Cut-off < 0.115 0.2 Sensitivity 60% 0.5 Specificity 100% 0.0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 PRE POST PRE POST PRE POST 1 - Specificity MFI CTLA-4 on CD3+ cells E T-NK like NK cells F ROC curve of MFI CTLA4 CD3+CD56+CTLA-4+ CD3-CD56+CTLA-4+ 6500 5000 * ns ns ns on CD3+CD56+CTLA4+ 3500 1.0 2000 500 MFI of CTLA-4+ 500 0.8 Sensitivity 400 0.6 300 AUC ROC 0.844 0.4 200 p-value 0.011 Cut-off < 234 0.2 Sensitivity 70% 100 Specifity 100% 0 0.0 PRE POST PRE POST 0.0 0.2 0.4 0.6 0.8 1.0 1 - Specificity CPR Non-CPR Diagnosis Post-treatment F I G U R E 2 CTLA-4 immunocheckpoint and pathological response. Differences between CPR and non-CPR patients at diagnosis and post neoadjuvant treatment. (A) Percentage of CTLA-4+ cells on T cells (CD3+ CD56– CTLA-4+ ), CD4 T cells (CD3+ CD4+ CTLA-4+ ), CD8 T cells (CD3+ CD8+ CTLA-4+ ), T-NK like (CD3+ CD56+ CTLA-4+ ) cells, NK cells (CD3– CD56+ CTLA-4+ ), and B cells (CD3– CD19+ CTLA-4+ ). (B) AUC ROC curve for CD3– CD56+ CTLA-4+ . (C) Percentage of CTLA-4+ cells on classical monocytes (CD14++ CD16– CTLA-4+ ), intermediate monocytes (CD14++ CD16+ CTLA-4+ ), and nonclassical monocytes (CD14+ CD16+ CTLA-4+ ). (D) AUC ROC curve for CD14++ CD16– CTLA-4+ . (E) MFI of CTLA-4+ on T-NK like cells (CD3+ CD56+ CTLA-4+ ), and T cells (CD3+ CD56– CTLA-4+ ). (F) AUC ROC curve for CD3+ CD56+ CTLA-4+ . Legend: Diagnosis (PRE, grey dots), prior to surgery (POST, black dots), CPR patients (yellow bars), and non-CPR patients (purple bar). Nonparametric Mann–Whitney test was used for comparisons (ns, not significant; *p-value <.05; **p-value <.01; ***p-value <.001) and receiver operating characteristic (ROC) curve analysis showing AUC ROC curve, p-value, and optimum cut-off 8 of 16 LAZA-BRIVIESCA et al. A1 MFI CD56 on CD56+ cells B1 MFI CD25 on Treg C1 MFI CD19 on B cells T-NK like NK cells T regs CD3-CD19+ CD3-CD19+CD19hi CD3+CD56+ CD3-CD56+ CD3+CD4+CD25hi 8000 1100 1800 1000 ** 0.06 ns ns 7000 0.06 * ns * ns 900 * MFI of CD1 9+ on B cells 1600 6000 800 MFI of CD56+ MFI of CD25+ 700 5000 1400 600 4000 500 1200 3000 400 300 2000 200 1000 1000 100 0 800 0 PRE POST PRE POST PRE POST PRE POST PRE POST A2 ROC curve of MFI CD56 B2 ROC curve of MFI CD25 D1 140000 BCMA on CD3+CD56+ on CD4+CD25hi * Levels of soluble expression 1.0 1.0 120000 0.8 0.8 100000 Sensitivity Sensitivity 0.6 0.6 80000 AUC ROC 0.836 AUC ROC 0.758 0.4 0.4 60000 p-value 0.003 p-value 0.023 Cut-off > 460 Cut-off > 1232 0.2 0.2 Sensitivity 33% Sensitivity 60% 40000 Specifity 100% Specifity 100% 0.0 0.0 20000 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 1 - Specificity 1 - Specificity 0 CPR Non-CPR Diagnosis Post-treatment F I G U R E 3 Immunophenotyping associated to pathological response. (A1) MFI of CD56+ cells on T-NK like (CD3+ CD56+ ) cells and NK cells (CD3– CD56+ ). (A2) AUC ROC curve of CD56 MFI on CD3+ CD56+ . (B1) MFI of CD25 on CD3+ CD4+ CD25hi . (B2) AUC ROC curve of CD25 MFI on CD3+ CD4+ CD25hi . (C1) MFI of CD19 on CD3– CD19+ and CD3– CD19hi . (D1) BCMA levels of expression. Legend: Diagnosis (PRE, grey dots), prior to surgery (POST, black dots), CPR patients (yellow bars), and non-CPR patients (purple bar). Nonparametric Mann–Whitney test was used for comparisons (ns, not significant; *p-value <.05; **p-value <.01; ***p-value <.001) and receiver operating characteristic (ROC) curve analysis showing AUC ROC curve, p-value, and optimum cut-off 3.3 Association between immune Moreover, the AUC of CD56 expression at diagnosis on phenotype of PBMCS and pathological CD56+ T cells was 0.836 (p-value .003). Thus, MFI higher response than 460 relative units predicts, with a 60% sensitivity and 100% specificity, patients achieving CPR. This population 3.3.1 CD56 expression on NK and T cells of patients entails one-third of analyzed patients (nine out of 27) (Figure 3A2). CD56 expression was used as a cytotoxic activation marker We did not find any differences in CD56+ T cells and in NK and CD56+ T cells. CPR patients had higher lev- NK cells percentages, even though CPR patients showed els of CD56 expression on CD56+ T cells at diagnosis, (p- a trend toward higher CD56+ T cells at diagnosis (p-value value .003) and post treatment (p-value .060) (Figure 3A1). .306) and post treatment (p-value .092) (Figure S2D). In LAZA-BRIVIESCA et al. 9 of 16 addition, CD56 expression on CD56+ T cells was inversely 3.4 Correlations between immune correlated to b-NGF (p-value .031) shown in Figure S3E. phenotype of PBMCs and cytokines To unravel potential association between immune cell sub- 3.3.2 CD25 expression on CD4 T populations and cytokines, we assessed the correlations lymphocytes among the described biomarkers and pathologic response. At diagnosis, there is a positive correlation between Regulatory T cells (Tregs) were determined selecting CD19 expression on B cells with CD4+ PD-1+ cells (R 0.419; CD4+ CD25high cells, as described in methods (Figure S1B). p-value .024), CD25 expression on Tregs CD4+ CD25high (R At diagnosis, CPR patients had higher levels of CD25 0.570; p-value .001), and CD56 expression on CD56+ T cells expression on CD3+ CD4+ CD25high (Figure 3B1). The value (R 0.512; p-value .005) (Figure S4A–C). A positive correla- of pretreatment CD25 MFI on regulatory T cells as a tion between CD56 expression on CD56+ T cells and CD25 biomarker associated to CPR is shown by an ROC AUC expression on Tregs (R 0.555; p-value .002) was found (Fig- of 0.758 (p-value .023) (Figure 3B2). Meanwhile, no differ- ure S4D). Furthermore, NKG2D expression on T cells cor- ences were seen in percentage of CD3+ CD4+ CD25hi cells, relate positively with CD56 expression (R 0.465; p-value at diagnosis or post treatment (Figure S2E). .011) and NKG2D expression (R 0.652; p-value .0001) on CD56+ T cells (Figure S4E,F). Following neoadjuvant treatment, we observed a nega- 3.3.3 CD19 expression on B cells tive correlation between CD19 expression on B cells and NKG2D on T cells (R −0.488; p-value .007) and CD56+ T B cells were subdivided, finding a larger size population cells (R −0.565; p-value .001). A positive correlation was named as CD19high , as described in methods (Figure S1B). found between NKG2D expression on T cells and CD56+ Post neoadjuvant treatment, CPR patients have higher lev- T cells (R 0.783; p-value .000) (Figure S4G–I). els of CD19 MFI on B cells and on CD19high cells post treat- Nevertheless, focusing on cytokines, on one hand, NT- ment (Figure 3C1). Furthermore, higher posttreatment lev- 3 correlated positively with VEGF-D and Flt-3L, and neg- els of TNF receptor superfamily member 17 (TNFRSF17; atively with PARC and MPIF-1 (Figure S5A–D). Simi- BCMA) were found on CPR patients (Figure 3D1). On the larly occurred with VEGF-D, as it correlated positively contrary, no differences were seen in percentage of B cells with b-NGF and Flt-3L and negatively with PARC (Figure or CD19hi cells at diagnosis or following neoadjuvant treat- S5E–G). On the other hand, there was a positive correlation ment for CPR patients or non-CPR patients (Figure S2F). between PARC and Flt-3L and a negative correlation with MPIF-1 (Figure S5H,I). Lastly, BCMA positively correlated with 4-1BB (Figure S5J). 3.3.4 Monocytes and macrophages Remarkably, none of the blood parameters analyzed at activation diagnosis in this study were associated to basal PD-L1 TPS or TMB (data not shown). To measure activation of monocytes and macrophages, we used CD69 expression and levels of different soluble fac- tors. At diagnosis, only intermediate monocytes were early 3.5 Neoadjuvant treatment influence activated in CPR patients, showing higher CD69 expres- on immune cells and plasma factors sion (p-value .017) (Figure 4A). The use of these activated intermediate monocytes as CPR biomarker is shown by an We also describe the significant effect of chemoim- AUC ROC of 0.772 (p-value .017) (Figure 4B). munotherapy on immune cells (Table S4A,B) and plasma Moreover, posttreatment cytokine levels related to factors (Table S5A–C) in all patients and stratifying by immune system activation, TNF receptor superfamily pathologic responses. As previously shown, percentage of member 9 (TNFRSF9; 4-1BB) and macrophages function general population T cells, CD4+ T cells, CD8+ T cells, such as, macrophage colony-stimulating factor 1 (CSF-1; monocytes, B cells, NK cells, and CD56+ T cells did not MCSF), C–C motif chemokine ligand 18 (CCL18; PARC), vary with treatment (Figure S2A). Fms-related receptor tyrosine kinase 3 ligand (Flt-3L) and Independently of pathologic response and follow- C–C motif chemokine ligand 23 (CCL23; MPIF-1), seem to ing treatment initiation, a decrease of cells’ percent- have an association with pathological response. We found ages was observed for most of the subgroups except higher levels of 4-1BB, MCSF, and PARC but lower levels of for CD8+ PD-1– , CD8+ NKG2D– , CD3+ CD56+ CD16– , and Flt-3L and MPIF-1 in CPR patients after neoadjuvant treat- CD3+ CD56+ NKG2D– . Moreover, these cells are increased ment (Figure 4C). after treatment only in CPR patients. (Figure 5A). 10 of 16 LAZA-BRIVIESCA et al. A MFI of CD69 on monocytes B Classical Intermediate Non-classical ROC curve of MFI CD69 14000 on CD14++CD16+CD69+ ns 12000 1.0 MFI of CD69+ 10000 0.8 ns ns ns 0.054 Sensitivity 8000 * 0.6 6000 AUC ROC 0.772 0.4 p-value 0.017 4000 Cut-off >1454 0.2 Sensitivity 40% 2000 Specificity 100% 0.0 0.0 0.2 0.4 0.6 0.8 1.0 0 PRE POST PRE POST PRE POST 1 - Specificity C 4-1BB MCSF PARC Flt-3L MPIF 156000 8000 50000 12000 40000 106000 * * 11000 * * * 45000 Levels of soluble expression Levels of soluble expression Levels of soluble expression Levels of soluble expression Levels of soluble expression 7000 35000 10000 56000 40000 6000 9000 30000 6000 35000 6000 8000 5000 25000 30000 7000 5000 4000 25000 6000 20000 4000 5000 3000 20000 15000 3000 4000 15000 2000 3000 10000 2000 10000 2000 1000 1000 5000 5000 1000 0 0 0 0 0 CPR NO CPR Diagnosis Post-treatment F I G U R E 4 Monocytes and macrophages activation. (A1) MFI of CD69 on classical monocytes (CD14++ CD16– CD69+ ), intermediate monocytes (CD14++ CD16+ CD69+ ), and nonclassical monocytes (CD14+ CD16+ CD69+ ). (A2) AUC ROC curve of CD69 MFI on CD14++ CD16+ CD69+ . (B) 4-1BB levels of expression. (C) MCSF levels of expression. (D) PARC levels of expression. (E) Flt-3L levels of expression. (F) MPIF levels of expression. Legend: Diagnosis (PRE, grey dots), prior to surgery (POST, black dots), CPR patients (yellow bars), and non-CPR patients (purple bar). Nonparametric Mann–Whitney test was used for comparisons (ns, not significant; *p-value <.05; **p-value <.01; ***p-value <.001) and receiver operating characteristic (ROC) curve analysis showing AUC ROC curve, p-value, and optimum cut-off However, when looking at the expression of analyzed no differences found on soluble factors as a predictive markers by MFI, all patients have a slightly decrease biomarker or associated to pathological response. How- on PD-1 detection on monocytes, T cells, CD4+ T cells, ever, when comparing samples at diagnosis versus post- CD8+ T cells, and NK cells (Figure 5A). Regarding CD8 neoadjuvant treatment, major differences were found in expression, measured as MFI, there was an increase on non-CPR patients with a decrease, post treatment, of all CD8 expression in all patients (p-value .0004). Concerning soluble factors analyzed except for HVEM (p-value .049) CD56, on one hand, its expression on NK cells increased and TIM-3 (p-value .002) (Figure 5B). after treatment in all patients (p-value .002) and this Lastly, levels of many cytokines changed after treatment. significance was only maintained in CPR patients (p-value Considering the focus of this study and space limitation, .031). On the other hand, CD56 expression on CD56+ we describe here only the changes that are exclusive of T cells was increased only in non-CPR patients (p-value patients achieving CPR. These patients showed an increase .019) (Figure 5A). of ANG (p-value .027), angiostatin (p-value .041), CTACK Focusing on the effect of neoadjuvant chemoim- (p-value .031), HCC-1 (p-value .001), MICA (p-value munotherapy on cytokines and soluble factors, there were .041), RAGE (p-value .041), TARC (p-value .020), TIM-1 LAZA-BRIVIESCA et al. 11 of 16 Median Median A C ONLY CPR CPR Non-CPR ONLY CPR CPR N on-CPR NON-CPR NON-CPR TOTAL TOTAL MPR MPR MPR MPR MPR MPR MPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR MPR MPR MPR MPR MPR MPR MPR IPR IPR IPR IPR IPR NA NA CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR IPR IPR IPR IPR IPR NA NA NA ** * Clásicos * * CD14++CD16-CTLA4+ * ANG ** ** CD14++CD16+CD107a+ *** ** ANG-1 * CD14++CD16+CD69+ Angiostatin *** * * No clásicos % Monocytes * * ** * CD14+CD16+CD69+ *** * ** BDNF ** * CD14+CD16+NKp44+ * BMP-4 ** * CD14+CD16+PD1+ * BMP-5 * * CD3+CD56-CD16+ * b-NGF ** * CD3+CD56-CD107a+ ** * BTC * * CD3+CD56-CTLA4+ * * CD40L * * CD3+CD56-NKp44+ % T cells * * CTACK *** ** ** CD3+CD56-PD1+ *** * * DR6 *** ** ** CD3+NKG2D+PD1+ *** ** EGF *** *** ** CD3+NKG2D-PD1+ * * EGF R ** * CD3+CD4+CD69+ * ENA-78 *** ** * CD3+CD4+NKG2D+PD1+ % CD4 T cells * Eotaxin *** *** ** CD3+CD4+PD1+ ** Fas *** *** ** CD3+CD4+PD1+NKG2D- * Fcr RIIB/C * CD3+CD4+CD8d *** ** * Flt-3L * * CD3+CD8+NKG2D- * Follistatin * * CD3+CD8+NKG2D+ *** ** ** GDF-15 * * CD3+CD8+PD1- % CD8 T cells *** * ** GM-CSF *** ** ** CD3+CD8+PD1+ * * GRO *** ** ** CD3+CD8+NKG2D+PD1+ *** ** ** CD3+CD8+PD1+NKG2D- ** ** HCC-1 * * CD3+CD56+CD16- * HGF * * CD3+CD56+NKG2D- % T-NK like cells *** ** IFNg *** ** * CD3+CD56+PD1+ * IGFBP-4 *** * * CD3-CD56+CD107a+ *** * ** IL-10 * CD3-CD56+CTLA4+ *** * ** IL-11 * * CD3-CD56+NKp44+ * IL-12p40 * ** CD3-CD56+PD1+ % NK cells *** * ** IL-12p70 * CD3-CD56+NKG2D+PD1+ *** ** IL-13 * * CD3-CD56+NKG2D-PD1+ * IL-15 * * CD14++CD16-CTLA4+_MFICTLA4+ ** IL-16 * CD14++CD16-NKp44+_MFINKp44+ *** ** IL-17 * CD14++CD16+CD69+_MFICD69+ MFI on monocytes ** * IL-17B ** * CD14++CD16+NKp44+_MFINKp44+ * IL-17R ** * CD14++CD16+PD1+_MFIPD1+ ** * IL-1a * CD3+CD56-CD16+_MFICD16+ ** ** IL-1ra * CD3+CD56-CD69+_MFICD69+ *** ** IL-2 *** ** CD3+CD56-PD1+_MFIPD1+ MFI on T cells * IL-29 * * CD3+NKG2D+PD1+_MFIPD1+ *** ** IL-5 ** * CD3+NKG2D-PD1+_MFIPD1+ ** * CD3+CD4+PD1+_MFIPD1+ *** ** IL-6 * * CD3+CD4+PD1+NKG2D-_MFIPD1+ MFI on CD4 T cells * IL-6sR * * CD3+CD4+Treg_MFICD25+ *** ** IL-7 *** * ** CD3+CD8+_MFICD8+ ** * IL-8 ** * CD3+CD8+CD69+_MFICD69+ *** * ** IL-9 *** ** CD3+CD8+CD69-_MFICD8+ ** * * Insulin ** CD3+CD8+NKG2D-_MFICD8+ * IP-10 *** * ** CD3+CD8+NKG2D+_MFICD8+ MFI on CD8 T cells ** * LAP *** * ** CD3+CD8+PD1-_MFICD8+ * L-Selectin *** *** ** CD3+CD8+PD1+_MFICD8+ * ** MCP-1 * CD3+CD8+PD1+_MFIPD1+ * * MCP-3 * CD3+CD8+NKG2D+PD1+_MFIPD1+ ** * MCSF R * CD3+CD56+_MFICD56 * MICA * CD3+CD56+CD16+_MFICD16+ *** * ** MIF * CD3+CD56+CD16+_MFICD56+ *** ** MIG * * CD3+CD56+CD69-_MFICD56+ ** ** MIP-1d * CD3+CD56+CD69+_MFICD69+ MFI on T-NK like cells * MIP-3a * ** CD3+CD56+NKG2D+_MFICD56+ * MIP-3b * CD3+CD56+NKp44+_MFICD56+ ** * MSPa * * CD3+CD56+PD1-_MFICD56+ * CD3+CD56+PD1+_MFICD56+ * * NGF R ** * CD3-CD56+_MFICD56+ * * NT-3 ** * CD3-CD56+PD1+_MFIPD1+ ** ** PAI-I * * CD3-CD56+NKG2D+PD1+_MFIPD1+ MFI on NK cells *** * ** PDGF-AA * CD3-CD56+NKG2D-PD1+_MFIPD1+ *** ** ** PDGF-AB *** * ** PDGF-BB * PECAM-1 Median *** ** PF4 * PIGF CPR N on-CPR ** * RAGE ONLY CPR NON-CPR B *** *** ** SCF R TARC TOTAL * MPR MPR MPR MPR MPR MPR MPR * TGFb CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR CPR IPR IPR IPR IPR IPR NA NA NA * * TIM-1 ** * TIMP-1 ** ** BTLA *** ** TIMP-2 ** ** CD28 * ** CD80/B7-1 *** ** TNF RI * * CD86/B7-2 ** ** TNF RII ** ** CTLA-4 *** ** TNFa ** ** GITR *** ** TNFb * * GITRL ** * TRAIL R3 * HVEM * * ICOS * TRAIL-R4 ** ** LAG-3 * VCAM-1 ** * PD-L1 ** VEGF ** ** TIM-3 * VEGF R3 ** ** TLR-2 0 0.5 1.0 1.5 2.0 >2 No data F I G U R E 5 Neoadjuvant treatment influence on immune cells and plasma factors. (A) Percentage of immune cells from peripheral blood and MFI on immune cells from peripheral blood. (B) Soluble factors secretion in plasma. (C) Cytokine secretion on plasma. Legend: Fold-change (POST/PRE) from 0 to 2, 0 indicating (blue) a decrease after treatment, 1 (pink) indicates no changes, 2 (red) indicates increase, and >2 (dark red) indicates more than double of increase, and black squares (no data). Nonparametric Wilcoxon test was used for comparisons, (empty, not significant; *p-value <.05; **p-value <.01; ***p-value <.001) (p-value .020), and TIMP-2 (p-value .003), and a decrease of statistically significant only for CD4+ , but not for CD8+ Fcr RIIB/C (p-value .045), GRO (p-value .036), and TRAIL- T cells. R4 (p-value .041) (Figure 5C). To our knowledge, there are no studies focused on predictive peripheral biomarkers of complete patholog- ical responses in the neoadjuvant setting. Besides, as 4 DISCUSSION chemoimmunotherapy seems superior to monotherapy16 and chemotherapy modifies the antitumor immune Immunotherapy alone or in combination with chemother- response,29,30 the potential biomarkers described for apy is revolutionizing the therapeutic approach of early- anti-PD(L)-1 monotherapy may be useless for chemoim- stage lung cancer patients, achieving a remarkable rate munotherapy. In fact, it is likely that these biomarkers of CPR in potentially resectable patients. In lung can- may vary also depending on tumor stage, treatment timing cer, CPR has been proposed as a surrogate of OS ben- (neoadjuvant or adjuvant), and the pathologic response efit in the neoadjuvant setting. In our study, patients rate used as threshold (CPR vs. non-CPR analysis identi- achieving CPR showed a trend toward higher basal lev- fied different parameters compared to MPR vs. non-MPR els of PD-1+ cells for T, NK, and T subsets, which was in our cohort), making the contrast of our results an 12 of 16 LAZA-BRIVIESCA et al. arduous task. Additionally, PD-L1 TPS and TMB at diag- levels of NKG2D in CD56+ T cells and nearly significant nosis were not associated to any parameter analyzed in in CD8+ T cells, but not for NK cells. Furthermore, the this study, indicating an independent value of these blood levels of MICA and MICB ligands were similar between biomarkers. CPR and non-CPR patients, suggesting that this evasion Although PD-1 interaction with its ligands leads to T- mechanism is not relevant in our cohort of patients. On cell inhibition, PD-1 expression is induced after antigen the other hand, NKG2D levels in CD56 T cells and CD8+ recognition. Thus, PD-1 levels may reflect chronic activa- T cells were inversely correlated with NT-3 plasma levels, tion of antitumor-specific lymphocytes.31 Few studies have which could imply a mechanism of inhibition by this described the importance of early proliferation of PD-1+ neurotrophin on immune cells.47,48 Similarly, an inverse cells, mainly CD8+ , after anti-PD-1 treatment in NSCLC correlation between VEGF-D and NKG2D was observed, patients.32,33 although this was only statistically significant for CD56+ Another population of T cells that seems to recapitu- T cells. To our knowledge, there is no specific information late the basal immune status of patients achieving CPR is on the role of VEGF-D and NKG2D regulation, although Tregs (CD4+ CD25hi ). The levels of these cells are higher a protumoral role of VEGF-D in lung cancer has been in the blood of NSCLC patients compared to healthy described.49 individuals,34,35 they are implicated on immune self- Other receptor involved in tumor evasion mechanisms tolerance36 and have been associated with favorable clin- is CTLA-4, originally described on T cells. There are few ical outcomes to anti-PD-1 therapy.37 We have previously studies on other cell types, such as monocytes and NK described in NSCLC patients that tumor-infiltrating Tregs cells, pointing in the same direction of inhibiting immune after chemoimmunotherapy are associated with CPR.17 In responses.50,51 Supporting this, we have seen that CPR this study, we describe their predictive value for CPR when patients presented reduced basal levels of CTLA-4+ on measured in peripheral blood at diagnosis. The associa- NK and classic monocytes. Despite CTLA-4 and PD-1 are tion between CPRs and basal PD-1+ or Tregs cells may inhibitory receptors and markers of antigen experienced reflect the immunogenicity of the tumor and the tolerance cells, in our study these receptors behave in the oppo- to chronic immune responses. site way, which could be explained by their participation In addition to T cells, CD56+ T cells also appear to in different pathways in which CTLA-4 is not affected by be characteristic of the immune system of CPR patients. nivolumab treatment.52 Reduced blood levels of CD56+ T cells has been described In this way, at diagnosis, antitumor immune status on NSCLC38 and other solid tumors,39–41 compared to for CPR patients seems to be characterized by a previ- healthy controls. Although, in a similar way to other ously induced immune response (PD-1+ cells and Tregs), groups,42 the levels of CD56+ T cells were not related with with a cytotoxic profile (NKG2D and CD56 receptors), response, but we observed that CPR patients had higher and lower levels of inhibitory cytokines and cells (neu- expression of activating receptors, such as NKG2D and rotrophins, VEGFD, and CTLA-4+ cells). This intense CD56. Besides the lack of knowledge in CD56+ T cells, and immune response would be hampered by some therapy- in particular of these receptors in responses to anti-PD-1 sensitive mechanisms, such as PD-1+ and Tregs, but not therapy, CD56+ T cells have been described as modulators by CTLA-4+ or determined inhibitory cytokines, that of the immune system.43 after chemoimmunotherapy would lead to complete tumor Traditionally, CD56 receptor is used to classify between elimination. Importantly, we were able to characterize this two types of NK cells, CD56bright (cytokine-producing cells) phenotype through peripheral blood, and moreover, to and CD56dim (cytotoxic cells). However, its expression is identify some biomarkers that could be useful as baseline also widely accepted as activation marker, associated with biomarkers for CPR. potent effector functions, in NK, classical αβ CD8 T cells, Most of these differences for pretreatment biomarkers and CD56+ T cells.44 Supporting this, we found a correla- are lost posttreatment, probably due to treatment effect, tion between CD56 and NKG2D for CD56+ T cells. that is able to modulate the immune cell repertoire. The NKG2D is a cytotoxic-related activation receptor main cell-related biomarker associated with CPR after expressed in NK, CD8+ T cells, and T CD56 cells. NKG2D treatment is CD19 expression. CD19 is a classic marker ligands (MICA, MICB, and ULBP1-6) are expressed on for B-cell identification, but also plays a key role as a stressed or tumor cells, favoring their destruction. Inter- coactivator for B-cell function and proliferation.53 Elevated estingly, a tumor evasion mechanism involves releasing levels of this marker may be associated with increased these ligands from the tumor cell surface, inducing humoral response and better clinical outcomes, because receptor internalization on immune cells, and therefore tertiary lymphoid structures have been associated with reducing their activation capacity.45,46 In our study, favorable prognosis.13 In fact, we found a correlation patients with CPR showed statistically significant higher between CD19 expression and other immune cells, such as LAZA-BRIVIESCA et al. 13 of 16 CD4+ PD-1+ and Tregs and CD56+ T cells, similar to previ- García Campelo reports personal fees from BMS, MSD, ous results.14,54 Roche, Pfizer, and AstraZeneca, outside the submitted Although neoadjuvant chemotherapy for cancer has his- work. Manuel Dómine reports personal fees from Astra- torically been considered immunosuppressive, we see that Zeneca, BMS, Boehringer Ingelheim, MSD, Pfizer, and patients maintain general lymphocyte populations during Roche, outside the submitted work. Margarita Majem treatment, similar to other studies evaluating monother- reports grants and personal fees from BMS, personal apy with anti-PD-1.32 In our study, neoadjuvant chemoim- fees and nonfinancial support from MSD, BOEHRINGER munotherapy modified many subpopulations of lympho- INGELHEIM, ASTRA ZENENCA, ROCHE, and personal cytes and monocytes in peripheral blood, as well as more fees from KYOWA KYRIN and PIERRE FABRE, out- than 50 cytokines and soluble factors in plasma. Addition- side the submitted work. Delvys Rodríguez-Abreu reports ally, to general changes that occurred in all patients, our grants and personal fees from Bristol Myers Squibb, per- results show that some changes were exclusive of patients sonal fees from GENENTECH/ROCHE, MSD, ASTRA who achieved CPR or non-CPR. This once again reinforces ZENECA, BOEHRINGER INGELHEIM, Novartis, and that they are distinct biological entities and the poten- Lilly, outside the submitted work. Alex Martínez-Martí tial role of these changes as biomarkers of pathological reports personal fees and nonfinancial support from Bris- response. tol Myers Squibb, F. Hoffmann La Roche AG, Merck This study has several limitations such as the lim- Sharp & Dohme, Pfizer, Boehringer Ingelheim, MSD ited number of patients included in this analysis or the Oncology, and AstraZeneca, outside the submitted work. lack of control arm to discern between prognostic and Javier De Castro Carpeño reports personal fees from Astra predictive biomarkers. Additionally, the scarcity of dis- Zeneca, Boehringer Ingelheim, Merck Sharp and Dohme, ease progression events or deaths in these patients,17 pre- Hoffmann-la Roche, Bristol Myers Squibb, Takeda, Pfizer, cluded the association of these parameters to patient’s and Novartis, outside the submitted work. Edel Del Barco survival. reports nonfinancial support from ROCHE, BMS, PFIZER, To conclude, this is the first study focused on the compo- ASTRA-ZENECA, and MERCK, during the conduct of the sition and phenotypic characteristics of PBMCs in NSCLC study. Isidoro Barneto Aranda reports consulting or advi- patients treated with neoadjuvant chemoimmunotherapy. sory board for Bristol Myers, Takeda, Roche, Astra Zeneca, Patients achieving CPR seem to have a characteristic and Behringer Inngelheim. Santiago Viteri reports per- peripheral blood immune status at diagnosis that allows sonal fees and nonfinancial support from BMS, ROCHE, the identification of baseline blood-based biomarkers asso- personal fees from MSD and ABBVIE, and nonfinan- ciated to CPR. Additionally, CPR patients also showed a cial support from OSE IMMUNOTHERAPEUTICS and distinctive modification of their immune status as a con- MERCK SERONO, outside the submitted work. Bartomeu sequence of neoadjuvant treatment with a different pro- Massuti reports grants and personal fees from Roche, file prior surgery. Further studies using larger cohorts of and personal fees and other from BMS, Takeda, MSD, patients and including a control arm are warranted to and Boehringer, outside the submitted work. Cara Hay- validate these biomarkers in the neoadjuvant chemoim- maker participated in scientific advisory board from Bri- munotherapy setting. acell. Ignacio I. Wistuba reports grants and personal fees from Genentech/Roche, Bayer, Bristol Myers Squibb, AstraZeneca/Medimmune, Pfizer, HTG Molecular, Merck, AC K N OW L E D G M E N T S Guardant Health, and personal fees from GlaxoSmithK- We thank the patients, their families, all the participating line, MSD, and Oncocyte, and grants from Oncoplex, clinical teams, and all the Spanish Lung Cancer Group, DepArray, Adaptive, Adaptimmune, EMD Serono, Takeda, BMS. A special thanks to Hospital Puerta de Hierro staff Amgen, Johnson & Johnson, Karus, Iovance, 4D, Novar- and Maria del Rocio Moreno Villa for their technical assis- tis, and Akoya, outside the submitted work. Atocha tance. Romero reports personal fees from Boehringer Ingelheim, outside the submitted work. Mariano Provencio reports CONFLICT OF INTEREST grants, personal fees, and nonfinancial support from BMS, Ernest Nadal reports personal fees from Bristol Myers ROCHE, ASTRAZENECA, and personal fees from MSD Squibb, Merck Sharpe & Dohme, AstraZeneca, Lilly, and TAKEDA, outside the submitted work. Raquel Laza- Amgen, and Boehringer Ingelheim, and grants and per- Briviesca, Alberto Cruz-Bermúdez, Gerardo Huidobro, sonal fees from Roche and Pfizer, outside the submit- Manuel Cobo, Guillermo López Vivanco, Reyes Bernabé ted work. Amelia Insa reports personal fees from Bris- Caro, Nuria Viñolas, Marta Casarrubios, Belén Sierra- tol, BOERINGHER, MSD, PFIZER, Roche, and ASTRA Rodero, Carlos Tarín, Aránzazu García-Grande, and Fer- ZENECA, outside the submitted work. María del Rosario nando Franco declare no conflict of interest. 14 of 16 LAZA-BRIVIESCA et al. AU T H O R CO N T R I B U T I O N S Bartomeu Massuti https://orcid.org/0000-0002-6247- Alberto Cruz-Bermúdez and Mariano Provencio conceived 4493 and designed the study. Ernest Nadal, Amelia Insa, María Marta Casarrubios https://orcid.org/0000-0002-2872- del Rosario García Campelo, Gerardo Huidobro, Manuel 9552 Dómine, Margarita Majem, Delvys Rodríguez-Abreu, Alex Belén Sierra-Rodero https://orcid.org/0000-0003-0276- Martínez-Martí, Javier De Castro Carpeño, Manuel Cobo, 6088 Guillermo López Vivanco, Edel Del Barco, Reyes Bern- Carlos Tarín https://orcid.org/0000-0002-1056-4024 abé Caro, Nuria Viñolas, Isidoro Barneto Aranda, San- Aránzazu García-Grande https://orcid.org/0000-0001- tiago Viteri, Bartomeu Massuti, Fernando Franco, and 5737-8918 Mariano Provencio recruited and treated patients. Raquel Cara Haymaker https://orcid.org/0000-0002-1317-9287 Laza-Briviesca, Marta Casarrubios, Belén Sierra-Rodero, Ignacio I. Wistuba https://orcid.org/0000-0003-3365- Alberto Cruz-Bermúdez, Carlos Tarín, Atocha Romero and 6340 Aránzazu García-Grande carried out the experiments and Atocha Romero https://orcid.org/0000-0002-1634-7397 analyzed the data. Raquel Laza-Briviesca, Alberto Cruz- Fernando Franco https://orcid.org/0000-0001-8964- Bermúdez, Marta Casarrubios, Belén Sierra-Rodero, Cara 5136 Haymaker, Ignacio I. Wistuba, Atocha Romero, and Mar- Mariano Provencio https://orcid.org/0000-0001-6315- iano Provencio interpreted the data. All the authors have 7919 read and contributed to the final version of the manuscript and approved its submission for publication. REFERENCES 1. Siegel RL, Miller KD, Fuchs HE, Jemal A. Cancer statistics, 2021. D A T A AVA I L A B I L I T Y S T A T E M E N T CA Cancer J Clin. 2021;71:7-33. The data that support the findings of this study are avail- 2. Morgensztern D, Ng SH, Gao F, Govindan R. Trends in stage dis- able in the supplementary material of this article. tribution for patients with non-small cell lung cancer: a national cancer database survey. J Thorac Oncol. 2010;5:29-33. 3. Majem M, Hernández-Hernández J, Hernando-Trancho F, et al. ORCID Multidisciplinary consensus statement on the clinical manage- Raquel Laza-Briviesca https://orcid.org/0000-0003- ment of patients with stage III non-small cell lung cancer. Clin 2369-567X Transl Oncol. 2020;22:21-36. Alberto Cruz-Bermúdez https://orcid.org/0000-0001- 4. Crinò L, Weder W, van Meerbeeck J, Felip E. Early stage and 5136-7011 locally advanced (non-metastatic) non-small-cell lung cancer: Ernest Nadal https://orcid.org/0000-0002-9674-5554 ESMO clinical practice guidelines for diagnosis, treatment and Amelia Insa https://orcid.org/0000-0002-3438-6170 follow-up. Ann Oncol. 2010;21:103-115. 5. Roth JA, Fossella F, Komaki R, et al. A randomized trial compar- María del Rosario García-Campelo https://orcid.org/ ing perioperative chemotherapy and surgery with surgery alone 0000-0003-2113-1504 in resectable stage IIIA non-small-cell lung cancer. J Natl Cancer Manuel Dómine https://orcid.org/0000-0003-1634-9832 Inst. 1994;86:673-680. Margarita Majem https://orcid.org/0000-0002-9919- 6. Ishida Y, Agata Y, Shibahara K, Honjo T. Induced expression 7485 of PD-1, a novel member of the immunoglobulin gene super- Delvys Rodríguez-Abreu https://orcid.org/0000-0003- family, upon programmed cell death. EMBO J. 1992;11:3887- 0506-1366 3895. Alex Martínez-Martí https://orcid.org/0000-0002-6714- 7. Strauss L, Mahmoud MAA, Weaver JD, et al. Targeted dele- tion of PD-1 in myeloid cells induces antitumor immunity. Sci 9526 Immunol. 2020;5:1-15. Javier De Castro Carpeño https://orcid.org/0000-0002- 8. O’Donnell JS, Teng MWL, Smyth MJ. Cancer immunoediting 3622-6306 and resistance to T cell-based immunotherapy. Nat Rev Clin Manuel Cobo https://orcid.org/0000-0003-3402-1144 Oncol. 2019;16:151-167. Guillermo López Vivanco https://orcid.org/0000-0003- 9. Sharma P, Allison JP. Immune checkpoint targeting in cancer 3800-6224 therapy: toward combination strategies with curative potential. Edel Del Barco https://orcid.org/0000-0001-6443-0585 Cell. 2015;161:205-214. 10. Palucka AK, Coussens LM. The basis of oncoimmunology. Cell. Reyes Bernabé Caro https://orcid.org/0000-0003-0312- 2016;164:1233-1247. 9195 11. Lou Y, Diao L, Cuentas ERP, et al. Epithelial-mesenchymal Nuria Viñolas https://orcid.org/0000-0002-8287-2739 transition is associated with a distinct tumor microenviron- Isidoro Barneto Aranda https://orcid.org/0000-0001- ment including elevation of inflammatory signals and multiple 6302-1034 immune checkpoints in lung adenocarcinoma. Clin Cancer Res. Santiago Viteri https://orcid.org/0000-0002-7572-1123 2016;22:3630-3642. LAZA-BRIVIESCA et al. 15 of 16 12. Pesce S, Greppi M, Grossi F, et al. PD/1-PD-Ls checkpoint: 28. Zelba H, Bochem J, Pawelec G, et al. Accurate quantification of insight on the potential role of NK cells. Front Immunol. T-cells expressing PD-1 in patients on anti-PD-1 immunother- 2019;10:1-8. apy. Cancer Immunol Immunother. 2018;67:1845-1851. 13. Helmink BA, Reddy SM, Gao J, et al. B cells and tertiary lym- 29. Parra ER, Villalobos P, Behrens C, et al. Effect of neoadjuvant phoid structures promote immunotherapy response. Nature. chemotherapy on the immune microenvironment in non-small 2020;577:549-555. cell lung carcinomas as determined by multiplex immunofluo- 14. Del Mar Valenzuela-Membrives M, Perea-García F, Sanchez- rescence and image analysis approaches. J Immunother Cancer. Palencia A, et al. Progressive changes in composition of lym- 2018;6:1-11. phocytes in lung tissues from patients with non-small-cell lung 30. Gaudreau PO, Negrao MV, Mitchell KG, et al. Neoadjuvant cancer. Oncotarget. 2016;7:71608-71619. chemotherapy increases cytotoxic T cell, tissue resident mem- 15. Hakanen H, Hernberg M, Mäkelä S, et al. Anti-PD1 ther- ory T cell, and B cell infiltration in resectable NSCLC. J Thorac apy increases peripheral blood NKT cells and chemokines in Oncol. 2020;16:127-139. metastatic melanoma patients. Ann Oncol. 2018;29:x3. 31. Topalian SL, Taube JM, Anders RA, Pardoll DM. Mechanism- 16. Calvo V, Sierra-Rodero B, Cruz-bermúdez A, Provencio M. Role driven biomarkers to guide immune checkpoint blockade in of immunotherapy in stage IIIA non-small cell lung cancer: a cancer therapy. Nat Rev Cancer. 2016;16:275-287. narrative review. Curr Challenges Thorac Surg. 2020:1-17. https: 32. Kamphorst AO, Pillai RN, Yang S, et al. Proliferation of PD-1+ //ccts.amegroups.com/article/view/42427 CD8 T cells in peripheral blood after PD-1-targeted therapy in 17. Provencio M, Nadal E, Insa A, et al. Neoadjuvant chemother- lung cancer patients. Proc Natl Acad Sci U S A. 2017;114:4993- apy and nivolumab in resectable non-small-cell lung cancer 4998. (NADIM): an open-label, multicentre, single-arm, phase 2 trial. 33. Kim KH, Cho J, Ku BM, et al. The first-week proliferative Lancet Oncol. 2020;21:1413-1422. response of peripheral blood PD-1þCD8þ T cells predicts the 18. Hellmann MD, Chaft JE, William WN, et al. Pathological response to Anti-PD-1 therapy in solid tumors. Clin Cancer Res. response after neoadjuvant chemotherapy in resectable non- 2019;25:2144-2154. small-cell lung cancers: proposal for the use of major pathologi- 34. Qiu J, Che G, Liu F, et al. The detection and clinical signifi- cal response as a surrogate endpoint. Lancet Oncol. 2014;15:e42- cance of peripheral regulatory CD4+CD25hiCD127low T cells e50. in patients with non-small cell lung cancer. Clin Transl Oncol. 19. Pataer A, Kalhor N, Correa AM, et al. Histopathologic response 2019;21:1343-1347. criteria predict survival of patients with resected lung cancer 35. Hu X, Gu Y, Zhao S, et al. Elevated circulating CD4+CD25- after neoadjuvant chemotherapy. J Thorac Oncol. 2012;7:825-832. Foxp3+ regulatory T cells in patients with nonsmall cell lung 20. Wojas-Krawczyk K, Kubiatowski T. Imperfect predictors for cancer. Cancer Biother Radiopharm. 2019;34:325-333. lung cancer immunotherapy—a field for further research. Front 36. Sakaguchi S, Ono M, Setoguchi R, et al. Foxp3+CD25+CD4+ Oncol. 2020;10:568174. natural regulatory T cells in dominant self-tolerance and 21. Kim KH, Kim CG, Shin EC. Peripheral blood immune cell- autoimmune disease. Immunol Rev. 2006;212:8-27. based biomarkers in anti-PD-1/PD-L1 therapy. Immune Netw. 37. Koh J, Hur JY, Lee KY, et al. Regulatory (FoxP3+) T cells 2020;20:1-15. and TGF-β predict the response to anti-PD-1 immunotherapy in 22. Garassino M, Rodriguez-Abreu D, Gadgeel S, et al. OA04.06 patients with non-small cell lung cancer. Sci Rep. 2020;10:1-10. evaluation of TMB in KEYNOTE-189: pembrolizumab plus 38. Dockry É, O’Leary S, Gleeson LE, et al. Epigenetic induction of chemotherapy vs placebo plus chemotherapy for nonsquamous CD1d expression primes lung cancer cells for killing by invariant NSCLC. J Thorac Oncol. 2019;14:S216-S217. natural killer T cells. Oncoimmunology. 2018;7:1-14. 23. Shu CA, Gainor JF, Awad MM, et al. Articles neoadjuvant ate- 39. Krijgsman D, de Vries NL, Skovbo A, et al. Characterization of zolizumab and chemotherapy in patients with resectable non- circulating T-, NK-, and NKT cell subsets in patients with col- small-cell lung cancer : an open-label, multicentre, single-arm, orectal cancer: the peripheral blood immune cell profile. Cancer phase 2 trial. Lancet Oncol. 2020;2045:1-10. Immunol Immunother. 2019;68:1011-1024. 24. Uryvaev A, Passhak M, Hershkovits D, et al. The role of tumor- 40. Molling JW, Langius JAE, Langendijk JA, et al. Low levels of cir- infiltrating lymphocytes (TILs) as a predictive biomarker of culating invariant natural killer T cells predict poor clinical out- response to anti-PD1 therapy in patients with metastatic non- come in patients with head and neck squamous cell carcinoma. small cell lung cancer or metastatic melanoma. Med Oncol. J Clin Oncol. 2007;25:862-868. 2018;35:25. 41. Singh AK, Shukla NK, Das SN. Altered invariant natural killer 25. Li S, Zhang C, Pang G, Wang P. Emerging blood-based biomark- T cell subsets and its functions in patients with oral squamous ers for predicting response to checkpoint immunotherapy in cell carcinoma. Scand J Immunol. 2013;78:468-477. non-small-cell lung cancer. Front Immunol. 2020;11:1-11. 42. Youn JI, Park SM, Park S, et al. Peripheral natural killer cells 26. Ottonello S, Genova C, Cossu I, et al. association between and myeloid-derived suppressor cells correlate with anti-PD-1 response to nivolumab treatment and peripheral blood lympho- responses in non-small cell lung cancer. Sci Rep. 2020;10:1-7. cyte subsets in patients with non-small cell lung cancer. Front 43. Krijgsman D, Hokland M, Kuppen PJK. The role of natural killer Immunol. 2020;11:125. T cells in cancer–a phenotypical and functional approach. Front 27. Travis WD, Dacic S, Wistuba I, et al. IASLC multidisciplinary Immunol. 2018;9:367. recommendations for pathologic assessment of lung cancer 44. Van Acker HH, Capsomidis A, Smits EL, Van Tendeloo VF. resection specimens following neoadjuvant therapy. J Thorac CD56 in the immune system: more than a marker for cytotox- Oncol. 2020;15:709-740. icity? Front Immunol. 2017;8:892. 16 of 16 LAZA-BRIVIESCA et al. 45. Skrzypski M, Marczyk M, Szymanowska A, et al. Abstract 3700: 52. Wei SC, Levine JH, Cogdill AP, et al. Distinct cellular mech- prognostic value of the expression of NKG2D and CD96 in early anisms underlie anti-CTLA-4 and anti-PD-1 checkpoint block- stage non-small cell lung cancer (NSCLC). Cancer Res. 2017;77. ade. Cell. 2017;170:1120-1133.e17. https://doi.org/10.1158/1538-7445.AM2017-3700 53. Wang K, Wei G, Liu D. CD19: a biomarker for B cell develop- 46. Cascone R, Carlucci A, Pierdiluca M, et al. Prognostic value ment, lymphoma diagnosis and therapy. Exp Hematol Oncol. of soluble major histocompatibility complex class I polypep- 2012;1:36. tiderelated sequence A in non-small-cell lung cancer - signifi- 54. Patel AJ, Richter A, Drayson MT, Middleton GW. The role of B cance and development. Lung Cancer Targets Ther. 2017;8:161- lymphocytes in the immuno-biology of non-small-cell lung can- 167. cer. Cancer Immunol Immunother. 2020;69:325-342. 47. Prakash Y, Thompson MA, Meuchel L, et al. Neurotrophins in lung health and disease. Expert Rev Respir Med. 2010;4:395- 411. S U P P O RT I N G I N F O R M AT I O N 48. Ricci A, Greco S, Mariotta S, et al. Neurotrophins and neu- Additional supporting information may be found online rotrophin receptors in human lung cancer materials and meth- ods subjects and tissue preparation. Am J Respir Cell Mol Biol. in the Supporting Information section at the end of the 2001;25:439-446. article. 49. Feng Y, Wang W, Hu J, et al. Expression of VEGF-C and VEGF- D as significant markers for assessment of lymphangiogenesis and lymph node metastasis in non-small cell lung cancer. Anat Rec (Hoboken). 2010;293:802-812. How to cite this article: Laza-Briviesca R, 50. Laurent S, Carrega P, Saverino D, et al. CTLA-4 is expressed Cruz-Bermúdez A, Nadal E, et al. Blood biomarkers by human monocyte-derived dendritic cells and regulates their associated to complete pathological response on functions. Hum Immunol. 2010;71:934-941. NSCLC patients treated with neoadjuvant 51. Davis-Marcisak EF, Fitzgerald AA, Kessler MD, et al. A novel chemoimmunotherapy included in NADIM clinical mechanism of natural killer cell response to anti-CTLA-4 ther- trial. Clin Transl Med. 2021;11:e491. apy identified by integrative analysis of mouse and human https://doi.org/10.1002/ctm2.491 tumors. 2020. https://doi.org/10.1101/2020.05.31.125625