nk cell inﬁltrates and hla class i expression in primary ... · ment in breast cancer....

Precision Medicine and Imaging

NK Cell Infiltrates and HLA Class I Expressionin Primary HER2þ Breast Cancer Predict andUncouplePathologicalResponseandDisease-freeSurvivalAura Muntasell1, Federico Rojo2,3, Sonia Servitja2,4, Carlota Rubio-Perez5,Mariona Cabo1, David Tamborero5,6, Marcel Costa-García7, María Martínez-Garcia2,4,Sílvia Men�endez2, Ivonne Vazquez8, Ana Lluch9,10, Abel Gonzalez-Perez5, Ana Rovira2,4,Miguel L�opez-Botet1,7, and Joan Albanell2,4,7

Abstract

Purpose:We investigated the value of tumor-infiltratingNK(TI-NK) cells and HLA class I tumor expression as biomarkersof response to neoadjuvant anti-HER2 antibody–based treat-ment in breast cancer.

Experimental Design: TI-NK cells and HLA-I were deter-mined by IHC in pretreatment tumor biopsies from twocohorts of patients with HER2-positive breast cancer[discovery cohort (n ¼ 42) and validation cohort (n ¼ 71)].Tumor-infiltrating lymphocytes (TIL)were scored according tointernational guidelines. Biomarker association with patho-logic complete response (pCR) and disease-free survival (DFS)was adjusted for prognostic factors. Gene set variation analysiswas used for determining immune cell populations concom-itant to NK-cell enrichment in HER2-positive tumors from theCancer Genome Atlas (n ¼ 190).

Results: TI-NK cells were significantly associated withpCR in the discovery cohort as well as in the validation

cohort (P < 0.0001), independently of clinicopathologicfactors. A �3 TI-NK cells/50x high-power field (HPF) cutoffpredicted pCR in the discovery and validation cohort [OR,188 (11–3154); OR, 19.5 (5.3–71.8)]. Presence of TI-NKcells associated with prolonged DFS in both patient cohorts[HR, 0.07 (0.01–0.6); P ¼ 0.01; HR, 0.3 (0.08–1.3); P ¼0.1]. NK-, activated dendritic- and CD8 T-cell gene expres-sion signatures positively correlated in HER2-positivetumors, supporting the value of NK cells as surrogates ofeffective antitumor immunity. Stratification of patients bytumor HLA-I expression identified patients with low andhigh relapse risk independently of pCR.

Conclusions: This study identifies baseline TI-NK cells as anindependent biomarker with great predictive value for pCR toanti-HER2 antibody–based treatment and points to the com-plementary value of tumor HLA-I status for defining patientprognosis independently of pCR.

IntroductionHER2 overexpression and/orHER2 gene amplification occur in

approximately 15% to 20% of breast tumors and are associatedwith aggressive disease (1). Combination of chemotherapy andanti-HER2mAbs, trastuzumab as standalone or trastuzumab andpertuzumab, is the prevailing neoadjuvant approach for patientswith primary HER2-positive breast cancer (2, 3). Achievement ofpathologic complete response (pCR) to neoadjuvant treatmenthas been associated with improved disease-free and overall sur-vival (OS); nonetheless, 35% to 50% of patients do not achievepCR and/or eventually relapse. On the other hand, a subgroup ofpatients presents excellent clinical outcomes to anti-HER2 mAbsin the absence of concomitant chemotherapy (2, 3). This hetero-geneity in clinical efficacy highlights: (i) the importance of furtherunderstanding the mechanisms of action underlying anti-HER2mAb–based treatment efficacy in different patients, and (ii) theneed for biomarkers aiding in patient stratification for tailoringtreatment escalation or deescalation.

Approved anti-HER2 mAbs are immunoglobulins of the G1subclass (IgG1) that specifically block HER2-mediated onco-genic signaling and can trigger antitumor immunity by engag-ing Fcg receptors (FcgR) expressed by immune cells (4, 5). In

1Immunity and Infection Lab, IMIM (Institut Hospital del Mar d'InvestigacionsM�ediques), Barcelona, Spain. 2Cancer Research Program, IMIM (Institut Hospitaldel Mar d'Investigacions M�ediques), Barcelona, Spain. 3Department of Pathol-ogy, IIS 'Fundaci�on Jim�enez Diaz', Madrid, Spain. 4Department of MedicalOncology, Hospital del Mar-CIBERONC, Barcelona, Spain. 5Institute for Researchin Biomedicine, Barcelona Institute of Science and Technology, Barcelona,Spain. 6Department of Oncology Pathology, Karolinska Institutet, Stockholm,Sweden. 7Pompeu Fabra University, Barcelona, Spain. 8Department ofPathology, Hospital del Mar, Barcelona, Spain. 9Department of Oncology,Hospital Clinico de Valencia-CIBERONC, Valencia, Spain. 10Universitat deValencia, Valencia, Spain.

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

A. Muntasell and F. Rojo contributed equally to this article.

M. L�opez-Botet and J. Albanell are co-senior authors of this article.

Corresponding Authors: Aura Muntasell, IMIM (Hospital del Mar MedicalResearch Institute), Doctor Aiguader, 88, Barcelona 08003, Spain. Phone: 349-3316-0811; Fax: 349-3316-0901; E-mail: [email protected]; and Joan Albanell,Hospital del Mar, IMIM, Passeig Maritim 25-29, Barcelona 08003, Spain.Phone: 349-3248-3137; Fax: 349-3248-3366; E-mail:[email protected]

doi: 10.1158/1078-0432.CCR-18-2365

�2018 American Association for Cancer Research.

ClinicalCancerResearch

www.aacrjournals.org 1535

on June 11, 2020. © 2019 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst December 6, 2018; DOI: 10.1158/1078-0432.CCR-18-2365

http://crossmark.crossref.org/dialog/?doi=10.1158/1078-0432.CCR-18-2365&domain=pdf&date_stamp=2019-2-5

http://clincancerres.aacrjournals.org/

experimental models, tumor antigen–specific mAbs rapidly killtumor targets via FcgR-mediated cytotoxicity (ADCC; refs. 6–8).This short-term process also induces a vaccine effect, linkingADCC with the development of long-term antitumor adaptiveimmunity (9). Natural killer (NK) cells are innate lymphocytescapable of recognizing antibody-coated targets through theactivating Fcg receptor CD16A (FcgRIIIA). CD16-mediatedNK-cell activation triggers the release of cytotoxic mediators,proinflammatory cytokines, and immune cell–recruiting che-mokines (10–12). The contribution of NK cells, together withCD8 T cells, to anti-HER2 mAb activity (13) and their role ascommanders of antitumor adaptive responses has been recentlyestablished in in vivo models (14, 15). In patients with breastcancer, the presence of tumor-infiltrating lymphocytes (TIL) hasbeen associated with response to neoadjuvant treatmentsincluding anti-HER2 mAbs (16, 17) and increased tumorinfiltration by NK cells after trastuzumab–docetaxel and theantibody–drug conjugate T-DM1 therapies has been reported(18–20). Despite these evidences, clinically relevant predictivebiomarkers are lacking.

Expression of HLA class I molecules is an additional parameterdetermining the susceptibility of cancer cells toNK andCD8T-cellrecognition. Indeed, tumor–antigen (neoantigen) presentationby HLA class I is required for triggering tumor-specific CD8 T-cellcytotoxicity, a pathway commonly hijacked in different tumortypes, including breast cancer (21–23). HLA-I downregulationfacilitates NK-cell–mediated tumor cell recognition, whereas highHLA-I expression may repress NK-cell activation against tumorsthrough interaction with inhibitory receptors of the KIR, CD94/NKG2, and LILRB1 families (10–12). Previous studies havereported high HLA-I expression as well as total HLA class I lossas good prognosis indicators in breast carcinomas (23–25).

We hypothesized that tumor-infiltratingNK cells andHLA classI expression in breast carcinomas could influence the efficacy ofanti-HER2 mAb–based neoadjuvant treatment.

Materials and MethodsClinical samples

This study includes pretreatment biopsy samples from twocohorts of patients with primary HER2-positive breast cancer.The discovery cohort included patients prospectively recruitedalong 2014–2016 at Hospital del Mar (Barcelona, Spain; n¼ 42).A retrospective cohort including patients recruited along 2008–

2013 at Hospital del Mar (Barcelona, Spain) and HospitalFundaci�on Jim�enez Díaz (Madrid, Spain) was used as validationcohort (n¼ 71; cohort diagram in Supplementary Fig. S1). HER2-positive subtype classificationwas defined following 2013ASCO/CAP guidelines (26). All patients received a neoadjuvant combi-nation treatment of standard chemotherapy and anti-HER2mAbs. The primary efficacy endpoint was pCR defined asypT0ypN0 based on histopathologic analysis of the resectionspecimen (27); disease-free survival (DFS) as the time fromsurgery until any breast cancer relapse or death by any causeand distant metastasis–free survival (DMFS) as the time fromsurgery to distant metastasis occurrence were secondary efficacyendpoints.

The study was conducted following Declaration of Helsinkiguidelines. All patients gave written informed consent for theanalysis of tumor biopsies for biomarker assessment. This studywas approved by the Hospital del Mar Ethics Committee (2013/5307) and is reported according to the REMARK guidelines.

Stromal TILs and NK-cell quantificationStromal tumor–infiltrating lymphocytes were quantified on

hematoxylin and eosin sections of pretreatment tumor biopsiesfollowing the guidelines of the International TIL Working Group(17). NK cells were identified as CD56þCD3� cells in tumorstroma–immune infiltrates by double IHC staining. Briefly,heat antigen retrieval was done in pH 9 citrate–based bufferedsolution and endogenous peroxidase was quenched. Mousemonoclonal anti-CD56 (clone 123-3, Dako-Agilent) and rabbitmonoclonal anti-CD3 (2GV6, Ventana-Roche) antibodies wereused, followed by incubation with a polymer coupled withperoxidase (UltraView, Ventana-Roche). Sections were visualizedwith 3,30-diaminobenzidine and alkaline phosphatase, and coun-terstained with hematoxylin. All incubations were performed atUltra Platform (Ventana-Roche). CD56 and CD3 were evaluatedby a computerized measurement using a DM2000 Leica micro-scope equippedwith theNuance FXMultispectral Imaging System(PerkinElmer). A computer-aided analysis yielded quantitativedata of CD56 and CD3 on pseudofluorescence images. CD56-positive CD3-negative cells were counted in 50 adjacent �400microscopic fields considering tumor stroma and expressed inabsolute numbers. TILs and NK-cell scoring was performed by anexpert pathologist (F. Rojo, Institut Hospital del Mar d'Investiga-cions M�ediques, Barcelona, Spain.) blinded to clinical data.

HLA class I IHC and quantificationHLA-I expression was evaluated in baseline biopsy sections

using the mouse mAb HC10, kindly provided by Prof. Dr. HiddePloegh (Whitehead Institute for Biomedical Research, Cam-bridge, MA). HC10 mostly reacts with HLA-B and HLA-C heavychains and some HLA-A (HLA-A10, -A28, -A29, -A30, -A31, -A32,and –A33; ref. 28). Staining of stromal cells served as internalpositive control for HLA-I antibody reactivity. A semiquantitativehistoscore was calculated for HLA-I by estimation of the percent-age of cells positively stained with low, medium, or high stainingintensity. The final score was determined after applying a weight-ing factor to each estimate. The following formula was used:histoscore ¼ (low%)X1 þ (medium%)X2 þ (high%)X3; theresults ranged from 0 to 300. HLA-I staining H score was sepa-rately quantified in tumor and stromal cells as well as in normalepithelium, when available. Four tumor groups were defined onthe basis of HLA-I H score quartiles in transformed cells: (i) HLA-I

Translational Relevance

Predictive biomarkers are needed for personalized breastcancer treatment. This study identifies tumor-infiltrating NKcells and tumorHLAclass I as complementary biomarkerswithpredictive and prognostic value in patients with primaryHER2-positive breast cancer treated with neoadjuvant anti-HER2mAbs. Our observations provide clues for future patientstratification and treatment adaptation as well as for rationallydesigning NK- and T-cell–targeted immunotherapy in a per-sonalized manner. In a general context, our study reveals theimportance of defining TIL subpopulations while integratingHLA class I tumor expression for optimizing the developmentof strategies for cancer treatment.

Muntasell et al.

Clin Cancer Res; 25(5) March 1, 2019 Clinical Cancer Research1536




negative (HLA-Ineg) group: tumors with HLA-I H scores �3 (n¼28); (ii) HLA-I low (HLA-Ilow) group: HLA-I H scores >3 and� 60 (n ¼ 25); (iii) HLA-I normal (HLA-I norm) group: H scores>60 and�180 (n ¼ 32), and (iv) HLA-I high (HLA-Ihigh) group:H score >180 (n ¼ 20).

TCGA data analysis for immune populationsRNAseq and clinical data of the HER2-positive breast cancer

cohort (n ¼ 190) generated by the Cancer Genome Atlas (TCGA,http://cancergenome.nih.gov/abouttcga) were downloaded fromthe Firebrowse latest available release (2016/01; http://firebrowse.org). Relative infiltration levels of 16 immune popula-tions were computed by gene set variation analysis (GSVA)through the gsva R package using nonoverlapping gene setsspecifically overexpressed in each cell type as described previously(29). The GSVA produced normalized enrichment scores rangingfrom�1 to 1, which represent the abundance of the immune cellpopulation in the sample relative to the remaining analyzedcohort. Hierarchical clustering of resulting immune infiltrationpatterns was done using theWard linkagemethod (through ScipyPython library). Correlations between the relative abundance ofcell populations across tumors were computed through a linearregression, using the Scipy Python library.

Statistical analysisBivariate analyses by Mann–Whitney U test were used to

compare continuous variables in patients achieving pCR or not.pCR OR for TILs and tumor-infiltrating NK (TI-NK) cell numberscalculated by binary logistic regression are reported for 10%increments or individual units, respectively. Multivariate analysiswas conducted with binary logistic regression in which patientbaseline clinical and tumor characteristics: age (as a continuousvariable), tumor size (T1 and T2 vs. T3 and T4), lymph node stage(LNþ vs. LN ¼ 0), estrogen receptor status (ERþ vs. ER�), tumorgrading (G1 and G2 vs. G3), and Ki67 status (Ki67<20% vs. Ki67�20%) were added in the model separately. Thresholds forstromal TILs and TI-NK cells that best discriminated pCR weredetermined using receiver operating characteristic (ROC) curveanalysis in the discovery cohort. Optimal cut-off points for TILscore and TI-NK cell numbers were determined by seeking themaximumYouden index (J¼ sensitivityþspecificity-1). Predictiveeffects over pCR of TILs and TI-NK cells as categorical variableswere calculated by Fisher exact test. Cox proportional hazardsregression was used to estimate the HRs of TILs and TI-NKbiomarkers in the analysis of DFS. Kaplan–Meier curves for DFSand DMFS were used to compare time to event in patientscategorized according to pCR, TILs, or TI-NK cells analyzed bythe log-rank test. All P values were two sided; P values lower than0.05 were considered significant. Statistical analysis was per-formed using GraphPad Prism version 6 (GraphPad software)and STATA version 15 (STATA Corp.).

ResultsBaseline tumor-infiltrating NK-cell numbers predict pCR inpatients with HER2-positive breast cancer receiving anti-HER2mAb–based neoadjuvant therapy

We hypothesized that TI-NK cell numbers could serve aspredictive biomarkers of response to anti-HER2 mAb–basedtherapy. For that purpose, TILs and TI-NK cells were analyzed indiagnostic core biopsies from two patient cohorts with HER2-

positive breast cancer presented as discovery and validationdatasets (cohort characteristics in Table 1). NK cells, identifiedasCD56þCD3� lymphocytes by double IHC,weremainly locatedin tumor stroma areas immersed within T-cell (CD3þ) infiltrates(Fig. 1A). Eventually, intratumor NK cells in close proximity totransformed epithelia could also be detected, yet enumeration ofTI-NK cells was restricted to stroma-associated NK cells.

In the discovery cohort, TIL and TI-NK cells were detected in100% and 65% of core biopsies. Median TIL score and TI-NKcell numbers were 20% (IQR 10–40) and 1.5 cells/50x HPF (IQR0–15), respectively. In the validation cohort, TIL and TI-NK cellswere detected in 93% and 72%of core biopsies with amedian TILscore of 20% (IQR 10–30) and 2 TI-NK cells/50xHPF (IQR 0–12).Higher TI-NK cell densities were detected in ER-negative as com-pared with ER-positive tumors in the discovery dataset, whereasTIL scores were comparable. TI-NK cell numbers were not con-sistently related to any other classic pathologic factor (Supple-mentary Table SI).

pCR rate was 48% and 42% in the discovery and validationcohort, respectively. Global pCR rate upon combining bothcohorts was 43%. ER was the only traditional clinicopathologicfactor associated with pCR in both cohorts. Baseline TIL scorepositively associated with pCR in the discovery, in the valida-tion, and in the global analysis combining both patient cohorts(Fig. 1B). Baseline TI-NK cell numbers showed a remarkableassociation with pCR in all analyzed datasets (Fig. 1C). TILscore and TI-NK cell numbers remained significantly associatedwith pCR when adjusted by ER status in multivariate analysis[TILs discovery OR, 1.05 (1–1.09), P < 0.015; TILs validationOR, 1.04 (1.01–1.08), P ¼ 0.006; TILs total OR, 1.04 (1.0–1.07), P < 0.0001; TI-NK discovery OR, 6.9 (1–40), P¼ 0.03; TI-NK validation OR, 1.45 (1.2–1.7), P < 0.0001; TI-NK total OR,1.49 (1.2–1.79), P < 0.0001]. Remarkably, the associationbetween TIL score and pCR vanished upon adjusting by TI-NKcell numbers in multivariate logistic regression [Discovery: TILsOR, 0.9 (0.76–1.08), P ¼ 0.3; Validation: TILs OR, 1.0 (0.95–1.05), P ¼ 0.8; Total: TILs OR, 0.9 (0.93–1.03), P ¼ 0.4],indicating that pCR was selectively associated to NK-cell–enriched TILs.

ROC curves and cutoff values providing best discriminationbetween pCR and non-pCR were calculated for TI-NK cells andTILs in the discovery cohort (Supplementary Fig. S2) and used forpatient stratification in subsequent analysis (Fig. 1D). ORs forpCR were higher in the group of patients with �50% TILs andremarkably high in tumors with �3 TI-NK cells/50xHPF uponadjusting by ER status in both patient cohorts (Fig. 1D). Tumorswith�3NK cells/50xHPF showed pCR rates of 100% and 77% inthe discovery and validation cohorts, respectively. Remarkably,�3 TI-NK cells/50xHPF cutoff performed as a biomarker withgreat sensitivity, specificity, positive, and negative predictive val-ue, superior to TILs cutoff, in both patient cohorts (Supplemen-tary Fig. S2).

We also assessed the relationship between baseline TILs andTI-NK cells with DFS (Fig. 2). Median clinical follow-up was of31 and 49 months in the discovery and validation cohort,respectively (Table I). Despite the limited follow-up, TI-NKcell presence (�1/50xHPF) associated to improved DFS in thediscovery, in the validation, as well as in the combined analysisof both patient cohorts (Fig. 2C, F, and I). TI-NK cell associ-ation with DFS was comparable with that of pCR in bothpatient datasets (Fig. 2A, D, and G). On the contrary, TIL score

NK Cells and HLA Class I as Biomarkers in HER2 Breast Cancer

www.aacrjournals.org Clin Cancer Res; 25(5) March 1, 2019 1537



http://cancergenome.nih.gov/abouttcga

http://firebrowse.org

http://firebrowse.org


was not significantly associated to DFS in any of the datasets(Fig. 2B, E, and H). In an exploratory analysis of the associationbetween TI-NK, TILs, and pCR with DMFS (Supplementary Fig.S3), TI-NK maintained its statistical significance (P ¼ 0.03) butnot pCR (P ¼ 0.06), perhaps due to the reduced number ofevents in DMFS as compared with DFS. Overall, TI-NK cellspredicted pCR and DFS better than TILs in patients withprimary HER2-positive breast cancer treated with neoadjuvantanti-HER2 mAbs and chemotherapy.

NK-cell–related gene set in breast carcinomas overexpressingHER2 correlatewith activateddendritic cell andCD8T-cell genesignatures and are associated with increased OS

To get an insight into the immune context associated with TI-NK cells, we estimated by GSVA the relative abundance of 16immune cell populations [cytotoxic NK cells (NKdim), B cells,eosinophils, macrophages, mast cells, CD56bright NK cells

(NKbright), neutrophils, T-helper cells (Th), central memoryT cells (Tcm), effector memory T cells (Tem), follicular helperT cells (Tfh), activated dendritic cells (aDC), immature dendriticcells (iDC), activated CD8 T cell (CD8), gamma delta T cells(Tgd), and regulatory T cells (Treg)] across HER2-positive breasttumors in the public TCGA dataset (n ¼ 190). Figure 3A showsthe distribution of the relative enrichment of these cell popula-tions across 190 tumors. In agreement with IHC data, therelative enrichment in NKdim gene signature varied in distincttumors and tended to be higher in ER-negative as comparedwith ER-positive tumors (Mann–Whitney U test, P ¼ 0.052;Fig. 3B), yet no significant association was found with otherclinical parameters such as tumor size or lymph node status(data not shown). NKdim gene set enrichment showed a pos-itive correlation with gene signatures identifying activateddendritic and CD8 T cells, hallmarks of productive antitumorimmunity, as well as B and regulatory T lymphocytes (Fig. 3A).

Table 1. Cohort description

Discovery Validation P Overall

Type of cohort Prospective Retrospective Prospective þ retrospectiveClinical setting Neoadjuvant Neoadjuvant NeoadjuvantN of patients 42 71 113Median age 64 (36–88) 54 (36–87) 0.01a 57 (36–88)Tumor size, n (%)T1–T2 33(79%) 37 (52%) 0.01b 70 (62%)T3–T4 9 (21%) 31 (43%) 40 (35%)NA – 3 (6%) 3 (3%)

Lymph node status, n (%)N0 17 (40%) 26 (37%) 1b 43 (38%)Nþ 25 (59%) 39 (55%) 64 (57%)NA – 6 (8%) 6(5%)

Histologic type, n (%)DIC 38 (90%) 70 (99%) 0.06b 108 (96%)Others 4 (10%) 1 (1%) 5 (4%)

Histologic grading, n (%)G1–G2 17 (40%) 37 (52%) 0.3b 54 (48%)G3 23 (55%) 32 (45%) 55 (49%)NA 2 (5%) 2 (3%) 4 (3%)

Hormonal status, n (%)ERþ 31 (73%) 45 (63%) 0.3b 76 (67%)ER� 11(27%) 26 (37%) 37 (33%)PRþ 21 (50%) 31 (44%) 0.5b 54 (46%)PR� 21 (50%) 40 (56%) 62 (54%)

Ki67 Index, n (%)<20% 7 (17%) 7 (10%) 0.38b 14 (12%)�20% 35 (83%) 61 (86%) 99 (85%)NA – 3 (4%) 3 (3%)

Chemotherapy and hormonal therapyAnthracyclines þ taxanes 30 (71%) 30 (42%) 0.001c 60 (53%)Taxanes 9 (21%) 38 (53%) 47 (41%)Hormonal 3 (7%) – 3 (3%)Others – 3 (4%) 3 (3%)

Pathologic clinical responsepCR 20 (48%) 30 (42%) 0.5b 50 (43%)No pCR 21 (50%) 40 (56%) 61 (54%)NA 1 (2%) 1 (1%) 2 (2%)

Anti-HER2Trastuzumab 27 (64%) 71 (100%) <0.0001b 98 (87%)Trastuzumab and pertuzumab 15 (36%) – 15 (13%)

Follow-up (median, IQR; months)Clinical follow-up 31 (22–38) 49 (24–71) 0.006a 34 (24–55)DFS at 3 years (value, 95% CI) 0.90 (0.73–0.96) 0.91 (0.8–0.96) 0.583 0.91 (0.82–0.95)

aMann–Whitney discovery versus validation cohort.bFisher test discovery versus validation cohort.cAcþT versus T in discovery and validation cohort.

Muntasell et al.





Of note, tumor stratification according to the top quartileNKdim gene set enrichment identified patients showing goodprognosis (P ¼ 0.008; Fig. 3C), whereas regulatory T cell (P ¼0.07) or activated CD8 T-cell (P ¼ 0.34) gene expression

signatures showed no significant correlation with OS in thesame dataset. Hence, NKdim cell–associated gene expressionsignature may also perform as surrogate biomarker of effectiveantitumor immunity in primary HER2 breast carcinomas.

Figure 1.

Baseline tumor-infiltrating NK-cell numbers predict completeresponse in patients with HER2-positive breast cancer receivinganti-HER2 mAb–basedneoadjuvant therapy. Stainingof pretreatment tumor biopsyparaffin-embedded sectionsusing anti-CD3 (green), anti-CD56 (red), and nucleicounterstaining (blue). A,Merged image and compositionindicating tumor and stromalareas are shown. NK cells wereidentified as CD56þCD3� cells.Each image row corresponds toan independent biopsyrepresentative of TIL areaseither lacking or presenting lowor high NK-cell numbers. B andC, TIL score and tumor-infiltrating NK-cell numbers inbaseline biopsies of patientsstratified by their clinicalresponse to neoadjuvanttreatment in the discovery,validation, and total cohort.Mann–Whitney U test was usedto compare medians betweengroups. D,ORs and 95%confidence intervals for pCRprediction of clinicopathologicfactors, TIL score, and TI-NKcells as categorical variables inthe discovery, validation, andtotal cohort. Values adjusted byER status.






High tumor HLA class I expression identifies patients withprolonged DFS, independently of pCR

Because expression of HLA-I molecules regulates the suscepti-bility of cancer cells to NK and CD8 T-cell recognition, we nextevaluated whether HLA-I tumor expression could further modu-late the efficacy of anti-HER2 mAbs. A total of 107 biopsies fromboth patient cohorts were evaluated for HLA-I expression by IHCusing the HC10 antibody and combined in the analysis of HLA-Iassociation with clinical outcomes. The H score corresponding totumor, stroma cells, and normal epithelia was separately anno-tated. Themajority of transformed cells within a biopsy presenteda homogeneousHLA-I labeling, though occasionally, tumor areaswith distinct intensities were observed, likely reflecting tumorclone variants forHLA-I expression.HLA-I expressionwas variablein different tumors, ranging from complete absence to highintensity; independently of HLA-I levels in stroma or normalepithelial cells (Fig. 4A and B). As continuous variable, tumorHLA-I H score did not associate with pCR or with any tested

clinicopathologic factor (Fig. 4C; Supplementary Table SII). Ofnote, HLA-I expression was higher in stroma from tumors achiev-ing pCR in comparison with those not achieving it (Fig. 4C),correlating with higher TIL content (r ¼ 0.27; P ¼ 0.004). For theanalysis of extreme HLA-I tumor phenotypes, patients were strat-ified in four quartiles as defined in methods. pCR rate in HLA-Ineg

tumors was 57% and progressively decreased with increasingHLA-I tumor expression, being 40% in HLA-Ilow and 26% inHLA-Inorm tumors, suggesting a relationship between pCR andsusceptibility to NK-cell recognition. However, tumors in theHLA-Ihigh group showed the highest pCR rate (67%), perhapsreflecting tumor contexts in which tumor-specific CD8 T cellscontributed to the efficacy of anti-HER2 mAb–based treatment(Fig. 4D). Interestingly, the groupofpatientswithHLA-Ihigh tumorsremained disease-free along the follow-up, regardless of prior pCRachievement, whereas patients with HLA-Ineg tumors displayed atrend for delayed relapses as compared with patients with HLA-Ilow

and HLA-Inormal tumors, despite showing high-pCR rates (Fig. 4E).

Figure 2.

Baseline TI-NK cell number association with DFS in patients with HER2-positive breast cancer receiving anti-HER2 mAb–based neoadjuvant therapy. Kaplan–Meier analysis for DFS of patients in the discovery cohort stratified by pCR (A), baseline TILs (�25% cut-off; B), and baseline TI-NK cells (� 1 NK cells/50 HPFcutoff; C); validation cohort stratified by pCR (D), baseline TILs (E), or baseline TI-NK cells (F); total cohort stratified by pCR (G), baseline TILs (H), or baseline TI-NK cells (I).

Muntasell et al.





Patient stratification by baseline HLA-I tumor expressionenhances the predictive performance of TI-NK cells on pCR andDFS

We next analyzed whether TILs and TI-NK cell numbers per-formed equally well as pCR predictors in tumors expressingdifferent HLA-I levels. Average TIL scores and TI-NK cell numberswere comparable among tumors of all HLA-I expression groupsand TI-NK cell numbers were significantly associated with pCR inall HLA-I subgroups (Supplementary Fig. S4). For subsequentanalysis, the �3 Ti-NK cell/50xHPF cutoff was applied to theoverall patient population.

Stratification of patients by tumor HLA-I quartiles generatedthree contexts with distinct clinical behavior (Fig. 5): (i) patientswith HLA-Ihigh tumors: association between TI-NK cells and pCRwas complete in this group of tumors with all 13 cases withbaseline �3 TI-NK cells developing pCR posttreatment (Fig. 5A).Of note, all 20 patients remained relapse-free along the study,regardless of baseline TI-NK cell density and of pCR achievement(Fig. 5B); (ii) patients with HLA-Iinterm (lowþnorm) tumors: TI-NKcells predicted pCR in 83% (15/18) of cases (Fig. 5C) andperformed better than pCR in the prediction of DFS (Fig. 5D andE); tumors withHLA-I low and normal expression showed similar

clinical behavior (Supplementary Fig. S6); and (iii) patients withHLA-Ineg tumors: TI-NK cells predictive performance on pCR andDFS was lower in this group of patients as compared with otherHLA-I groups (Fig. 5F–H).

DiscussionDespite many efforts, clinically useful predictive biomarkers

of response to anti-HER2 mAb–based treatments in breastcancer remain elusive. Data of the herein reported study revealeda highly significant and independent association between tumor-infiltrating NK-cell numbers in the diagnostic core biopsy and theprobability of achieving complete pathologic response to anti-HER2 mAb–based neoadjuvant treatment. Moreover, stratifica-tion of patients according to HLA-I tumor expression levelsenhanced the prediction of long-term clinical outcomes, identi-fying patients with either decreased or increased relapse riskindependently of pCR. From the clinical standpoint, our studyidentifies two independent and complementary biomarkers,which might support clinical decision making. Mechanistically,our data support the combined action of NK and CD8 T cells asmajor contributors to anti-HER2mAb–based treatment efficacy in

Figure 3.

NK-cell–rich infiltrates in breast carcinomas overexpressing HER2 correlate with activated dendritic cell and CD8 T-cell densities and associated with longer OS.A, Hierarchically clustered heatmap showing the relative abundance of 16 immune populations across 190 HER2-positive breast carcinomas computed usingGSVA. Each column is a tumor sample and each row illustrates the relative enrichment for specific immune cell populations. GSVA scores range from�1 (blue) toþ1 (red). Coefficient of the correlation between NKdim GSVA scores and those corresponding to other immune subsets are indicated in the right column andcolored fromwhite (R2¼ 0) to dark magenta (R2¼ 1). B,Distribution of NKdim GSVA scores in HER2þ/ERþ (n¼ 147) and HER2þ/ER� (n¼ 43) breast tumors.Statistical significance assessed by Mann–Whitney U test. C, Kaplan–Meier survival curves of patients on the highest 25% percentile for NK-cell gene setenrichment versus remaining patients.






primary breast cancer and points to HLA-I expression as a tumor-intrinsic factor further influencing response to treatment.

A relationship between baseline TILs and anti-HER2 mAbefficacy has been described in primary HER2-positive breastcancer. However, the magnitude of the association and itsclinical usefulness remains controversial. Some studies, includ-ing a recent analysis of 6 patient cohorts from Gepar trials (16),have reported a positive association between TILs as continu-ous variable and likelihood of pCR achievement (30, 31) orDFS (17, 32) upon anti-HER2 mAb–based treatment, whereasin others, the association with pCR was more evident inlymphocyte predominant breast tumors, defined by variableTIL cutoffs ranging from 30% to 60% in different studies(33–35). Hence, a consensus TILs cutoff associated to pCR orDFS prediction across different patient cohorts with HER2-positive breast cancer is still lacking. In our study, baseline TILscore was associated to higher pCR rates as continuous variable.Nonetheless, the optimal cutoff value determined in the dis-covery cohort performed with low sensitivity in the predictionof pCR and was not associated to DFS in any studied cohorts. Itis recognized that TILs comprise diverse immune cell popula-tions that might differentially influence tumor developmentand treatment outcome. Our study discloses TI-NK cells as an

excellent biomarker labeling TIL contexts with antitumorpotential in response to anti-HER2 mAb–based treatment. Infact, TI-NK cells not only associated to pCR but also underliethe association between TILs and pCR in both analyzed patientcohorts, as indicated by multivariate analysis. In addition, thequantitative approach for TI-NK cell determination allowed theidentification of a precise cut-off value in the discovery cohortwith a very significant predictive potential for pCR in thevalidation cohort, as well as in the global analysis includingboth patient cohorts. TI-NK cell numbers also associated withDFS in both patient cohorts. However, the follow-up of thepatients was limited, precluding definite conclusions regardingthe durable clinical outcome, particularly in the ER-positivesubgroup. Nonetheless, the association between OS andNK-cell–rich tumors could also be detected by the analysis ofNK-cell gene expression signature in HER2 tumors from theTCGA.

Beyond the biological difference between TILs andNK cells thatcan justify their different prognostic and predictive value, the factthat prognostic and predictive differences between TILs and NKcells could be also due to the precision on their semiquantitativeversus quantitative enumeration cannot be ruled out. However,TI-NK cell numbers ranged from 0 to 30 cells in 50 high-power

Figure 4.

High HLA class I tumor expression in diagnostic biopsies identifies patients with good prognosis, independently of prior pCR. HLA class I expressionwas analyzed by IHC in baseline tumor biopsies using the HC10 mAb. HLA-I H score was separately analyzed in tumor, stroma cells, and normalepithelia (n ¼ 107). A, Representative examples of tumors with high (case A), intermediate (case B), and negative (case C) HLA-I staining.B, Distribution of HLA-I H scores in tumor, stroma, and normal epithelia. C, Tumor and stroma HLA-I H scores in patients achieving or not pCR toneoadjuvant treatment. D, pCR rate in tumors stratified by HLA-I H score quartiles. Dashed line indicates 50% pCR rate. E, Kaplan–Meier curve forDFS in patients stratified according to tumor HLA-I H score quartiles.

Muntasell et al.





fields (which include roughly a total of 3 � 103 cells) precludingtheir semiquantitative assessment.

In a complementary perspective, analysis of baselineHLA class I expression on tumor cells identified patients withdistinct long-term outcomes, in some cases discordant forpCR (patients in tumor HLA-I high and negative quartiles).Stratification of intermediate HLA-I–expressing tumors enhancedTI-NK cell performance for pCR and DFS prediction. However,due to the very small numbers of patients in each subgroup, resultsshould be considered hypotheses-generating and no clinical con-clusions should be drawn from that.

From a mechanistic perspective, our data support the com-plementary action of NK and CD8 T-cell responses along tumorcontrol by anti-HER2 mAb–based treatment. Baseline TI-NK cellpresence may indicate tumor permissiveness to NK-cell infil-tration/proliferation/survival. Indeed, an increase in NK-cellinfiltration in tumors treated with anti-HER2 mAbs has beenreported previously (18, 20, 36). Enhanced NK-cell infiltrationcould directly contribute to anti-HER2 mAb–triggered ADCC as

well as facilitate the development of tumor-specific T-cellimmunity, as recently described in preclinical models (9, 14,15). In this regard, the positive correlation between geneexpression sets identifying NK cells, activated dendritic cells,and CD8 T cells would support a role of TI-NK cells as orches-trators of effective antitumor immunity in HER2þ breasttumors, as recently proposed in patients with melanoma(14). In our study, the association between pCR and TI-NKcells in HLA-Ineg tumors indicates the dispensable action of CD8T cells along tumor control in some neoadjuvant contexts. Incontrast, the elevated pCR rates together with the absence ofrelapses in patients with high HLA-I–expressing tumors sup-ported the decisive contribution of CD8 T-cell memory both fortreatment-dependent tumor control as well as for metastasissurveillance in HER2þ breast cancer along anti-HER2 mAbtherapies. Hence, patients with HLA-I negative or low tumorsat diagnosis may benefit from treatments enhancing NK cellrather than T-cell antitumor immunity, whereas in patients withhigh HLA-I–expressing tumors substitution of chemotherapy

Figure 5.

Patient stratification by baseline HLA-I tumor expression enhances the predictive performance of Ti-NK cells on pCR. The predictive and prognostic valueof TI-NK cell cutoffs were analyzed in tumors stratified according to HLA-Ihigh, HLA-I interm (low-normal) as superindex, and HLA-Ineg expression.A, C, andF, pCR rates of tumors with TI-NK�3 versus those with TI-NK<3 in the three HLA-I groups. Insets indicate the number of patients in each group. B, D, and G,Kaplan–Meier curve for DFS in patients stratified by TI-NK cell�1 cutoff in the three HLA-I groups. E and H, Kaplan–Meier curve for DFS in patients stratified bypCR in HLA-I intermediate and negative groups.






treatment by PD1/PD-L1 checkpoint blockade, enhancingtumor T-cell immunity, could be a reasonable option.

Limitations of this study include a modest cohort size, theheterogeneity in the anti-HER2 mAb regimen as well as the lackof standardized methodology for TI-NK and HLA-I assessment.In contrast, its hypothesis-driven prospective–retrospectivedesign and the strength of the associations reported enhanceits intrinsic value. For HLA-I staining by IHC, the use of HC10mAb, provided a general yet partial view biased toward HLA-Band HLA-C expression assessment, including b2 microglobu-lin–free HLA-I heavy chains (37). Despite the fact that furtherstudies addressing the molecular mechanisms underlying thespecific HLA class I alterations in distinct tumors (21, 22, 38)would be of interest, the analysis implemented here mightfacilitate a meaningful stratification of patients with distinctclinical behaviors for treatment adaptation. The predictivevalue of both biomarkers deserve validation for their potentialclinical utility in HER2-positive breast cancer, likely extendingto general biomarkers for patients' stratification and selectionfor immunotherapy.

Disclosure of Potential Conflicts of InterestA. Muntasell reports receiving speakers bureau honoraria from Roche. S.

Servitja reports receiving speakers bureau honoraria from and is a consultant/advisory board member for Roche. M. Martínez-Garcia is a consultant/advisoryboard member for Roche. A. Lluch is a consultant/advisory board member forNovartis, Pfizer, Roche, Eisai, and Celgene. J. Albanell reports receiving com-mercial research grants and speakers bureau honoraria from and is a consultant/advisory board member for Roche. No potential conflicts of interest weredisclosed by the other authors.

Authors' ContributionsConception and design: A. Muntasell, F. Rojo, S. Servitja, M. L�opez-Botet,J. AlbanellDevelopment of methodology: A. Muntasell, F. Rojo, S. Men�endez, J. Albanell

Acquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): F. Rojo, S. Servitja, M. Martínez-Garcia, S. Men�endez,I. Vazquez, A. Lluch, J. AlbanellAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): A. Muntasell, F. Rojo, S. Servitja, C. Rubio-Perez,M. Cabo, D. Tamborero, M. Costa-García, A. Gonzalez-Perez, M. L�opez-Botet,J. AlbanellWriting, review, and/or revision of the manuscript: A. Muntasell, F. Rojo,S. Servitja, M. Cabo, D. Tamborero, M. Martínez-Garcia, A. Lluch, A. Gonzalez-Perez, A. Rovira, M. L�opez-Botet, J. AlbanellAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): F. Rojo, S. Servitja, S. Men�endez, J. AlbanellStudy supervision: A. Muntasell, S. Servitja, A. Gonzalez-PerezOthers (Interpretation and discussion of results): A. Rovira

AcknowledgmentsThe authors thank the technical help of Andrea Vera, Gemma Heredia, Sara

Santana, and Laura Soria.We acknowledge Xavier Duran [Consulting Service onMethodology for Biomedical Research (AMIB) at IMIM] for assistance instatistical analysis. We also thank the patients for their generous participationin the study. The authors are supported by coordinated research projects fromAsociaci�on Espa~nola contra el C�ancer (GCB15152947MELE) and ProyectoIntegrado de Excelencia ISCIII/FEDER (PIE 2015/00008); M. L�opez-Botet andA. Muntasell are supported by Worldwide Cancer Research (15-1146) and byPlan Estatal IþD Retos (SAF2013-49063-C2-1-R; SAF2016-80363-C2-1-R),Spanish Ministry of Economy and Competitiveness (MINECO, FEDER) andby Generalitat de Catalunya (2017 SGR 888); J. Albanell, A. Rovira, and F. Rojoare supported by ISCiii/FEDER (PI15/00146 and CIBERONC), by Generalitatde Catalunya (2017 SGR 507) and by Fundaci�o Privada Cellex (Barcelona).A.Gonzalez-Perez is supported by aRam�on yCajal contract (RYC-2013-14554).The results shownhere are in part basedondata generatedby the TCGAResearchNetwork.

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received July 24, 2018; revised October 17, 2018; accepted November 30,2018; published first December 6, 2018.

References1. SlamonDJ, Clark GM,Wong SG, LevinWJ, Ullrich A,McGuireWL. Human

breast cancer: correlation of relapse and survival with amplification of theHER-2/neu oncogene. Science 1987;235:177–82.

2. Baselga J, Bradbury I, Eidtmann H, Di CS, de AE, Aura C, et al. Lapatinibwith trastuzumab for HER2-positive early breast cancer (NeoALTTO): arandomised, open-label, multicentre, phase 3 trial. Lancet 2012;379:633–40.

3. Gianni L, Pienkowski T, Im YH, Roman L, Tseng LM, Liu MC, et al. Efficacyand safety of neoadjuvant pertuzumab and trastuzumab in women withlocally advanced, inflammatory, or early HER2-positive breast cancer(NeoSphere): a randomised multicentre, open-label, phase 2 trial. LancetOncol 2012;13:25–32.

4. Bianchini G, Gianni L. The immune system and response toHER2-targetedtreatment in breast cancer. Lancet Oncol 2014;15:e58–e68.

5. Savas P, Salgado R, Denkert C, Sotiriou C, Darcy PK, Smyth MJ, et al.Clinical relevance of host immunity in breast cancer: fromTILs to the clinic.Nat Rev Clin Oncol 2016;13:228–41.

6. Clynes RA, Towers TL, Presta LG, Ravetch JV. Inhibitory Fc receptorsmodulate in vivo cytotoxicity against tumor targets. Nat Med 2000;6:443–6.

7. Jaime-Ramirez AC, Mundy-Bosse BL, Kondadasula S, Jones NB, Roda JM,Mani A, et al. IL-12 enhances the antitumor actions of trastuzumab via NKcell IFN-gamma production. J Immunol 2011;186:3401–9.

8. Spiridon CI, Guinn S, Vitetta ES. A comparison of the in vitro and in vivoactivities of IgG and F(ab')2 fragments of a mixture of three monoclonalanti-Her-2 antibodies. Clin Cancer Res 2004;10:3542–51.

9. DiLillo DJ, Ravetch JV. Differential Fc-receptor engagement drives an anti-tumor vaccinal effect. Cell 2015;161:1035–45.

10. Cerwenka A, Lanier LL. Natural killer cell memory in infection, inflam-mation and cancer. Nat Rev Immunol 2016;16:112–23.

11. Moretta L,Moretta A.Unravelling natural killer cell function: triggering andinhibitory human NK receptors. EMBO J 2004;23:255–9.

12. Muntasell A,OchoaMC,Cordeiro L, BerraondoP, Lopez-Diaz deCA,CaboM, et al. Targeting NK-cell checkpoints for cancer immunotherapy. CurrOpin Immunol 2017;45:73–81.

13. Stagg J, Loi S, Divisekera U, Ngiow SF, Duret H, Yagita H, et al. Anti-ErbB-2mAb therapy requires type I and II interferons and synergizes with anti-PD-1 or anti-CD137 mAb therapy. Proc Natl Acad Sci USA 2011;108:7142–7.

14. Barry KC, Hsu J, Broz ML, Cueto FJ, Binnewies M, Combes AJ, et al. Anatural killer-dendritic cell axis defines checkpoint therapy-responsivetumor microenvironments. Nat Med 2018;24:1178–91.

15. Bottcher JP, Bonavita E, Chakravarty P, Blees H, Cabeza-Cabrerizo M,Sammicheli S, et al. NK cells stimulate recruitment of cDC1 into the tumormicroenvironment promoting cancer immune control. Cell 2018;172:1022–37.

16. Denkert C, von MG, Darb-Esfahani S, Lederer B, Heppner BI, Weber KE,et al. Tumour-infiltrating lymphocytes and prognosis in different subtypesof breast cancer: a pooled analysis of 3771 patients treated with neoadju-vant therapy. Lancet Oncol 2018;19:40–50.

17. Salgado R, Denkert C, Campbell C, Savas P, Nuciforo P, Aura C, et al.Tumor-infiltrating lymphocytes and associations with pathological com-plete response and event-free survival in HER2-positive early-stage breast

Muntasell et al.





cancer treated with lapatinib and trastuzumab: a secondary analysis of theNeoALTTO trial. JAMA Oncol 2015;1:448–54.

18. Arnould L, Gelly M, Penault-Llorca F, Benoit L, Bonnetain F, Migeon C,et al. Trastuzumab-based treatment of HER2-positive breast cancer: anantibody-dependent cellular cytotoxicity mechanism? Br J Cancer2006;94:259–67.

19. Gennari R, Menard S, Fagnoni F, Ponchio L, Scelsi M, Tagliabue E, et al.Pilot study of the mechanism of action of preoperative trastuzumab inpatients with primary operable breast tumors overexpressing HER2. ClinCancer Res 2004;10:5650–5.

20. Muller P, Kreuzaler M, Khan T, Thommen DS, Martin K, Glatz K, et al.Trastuzumab emtansine (T-DM1) renders HER2þ breast cancer highlysusceptible to CTLA-4/PD-1 blockade. Sci Transl Med 2015;7:315ra188.

21. Aptsiauri N, Ruiz-Cabello F, Garrido F. The transition from HLA-I positivetoHLA-I negative primary tumors: the road to escape fromT-cell responses.Curr Opin Immunol 2018;51:123–32.

22. Cai L, Michelakos T, Yamada T, Fan S, Wang X, Schwab JH, et al. DefectiveHLA class I antigen processing machinery in cancer. Cancer ImmunolImmunother 2018;67:999–1009.

23. de Kruijf EM, van Nes JG, Sajet A, Tummers QR, Putter H, Osanto S, et al.The predictive value of HLA class I tumor cell expression and presence ofintratumoral Tregs for chemotherapy in patients with early breast cancer.Clin Cancer Res 2010;16:1272–80.

24. Kaneko K, Ishigami S, Kijima Y, Funasako Y, Hirata M, Okumura H, et al.Clinical implication ofHLA class I expression in breast cancer. BMCCancer2011;11:454.

25. Madjd Z, Spendlove I, Pinder SE, Ellis IO, Durrant LG. Total loss of MHCclass I is an independent indicator of good prognosis in breast cancer.Int J Cancer 2005;117:248–55.

26. Wolff AC,HammondME,HicksDG,DowsettM,McShane LM, Allison KH,et al. Recommendations for human epidermal growth factor receptor 2testing in breast cancer: American Society of Clinical Oncology/College ofAmerican Pathologists clinical practice guideline update. J Clin Oncol2013;31:3997–4013.

27. Ogston KN, Miller ID, Payne S, Hutcheon AW, Sarkar TK, Smith I, et al. Anew histological grading system to assess response of breast cancers toprimary chemotherapy: prognostic significance and survival. Breast2003;12:320–7.

28. Perosa F, Luccarelli G, Prete M, Favoino E, Ferrone S, Dammacco F. Beta 2-microglobulin-free HLA class I heavy chain epitope mimicry by monoclo-nal antibody HC-10-specific peptide. J Immunol 2003;171:1918–26.

29. Tamborero D, Rubio-Perez C, Muinos F, Sabarinathan R, Piulats JM,Muntasell A, et al. A pan-cancer landscape of interactions between solidtumors and infiltrating immune cell populations. Clin Cancer Res 2018;24:3717–28.

30. Ingold HB, Untch M, Denkert C, Pfitzner BM, Lederer B, Schmitt W, et al.Tumor-infiltrating lymphocytes: a predictive and prognostic biomarker inneoadjuvant-treated HER2-positive breast cancer. Clin Cancer Res 2016;22:5747–54.

31. Solinas C, CeppiM, Lambertini M, Scartozzi M, Buisseret L, Garaud S, et al.Tumor-infiltrating lymphocytes in patients with HER2-positive breastcancer treated with neoadjuvant chemotherapy plus trastuzumab, lapati-nib or their combination: a meta-analysis of randomized controlled trials.Cancer Treat Rev 2017;57:8–15.

32. Loi S, Michiels S, Salgado R, Sirtaine N, Jose V, Fumagalli D, et al. Tumorinfiltrating lymphocytes are prognostic in triple negative breast cancer andpredictive for trastuzumab benefit in early breast cancer: results from theFinHER trial. Ann Oncol 2014;25:1544–50.

33. Bianchini G, Pusztai L, Pienkowski T, Im YH, Bianchi GV, Tseng LM,et al. Immune modulation of pathologic complete response afterneoadjuvant HER2-directed therapies in the NeoSphere trial. AnnOncol 2015;26:2429–36.

34. Denkert C, vonMG, Brase JC, Sinn BV, Gade S, Kronenwett R, et al. Tumor-infiltrating lymphocytes and response to neoadjuvant chemotherapy withor without carboplatin in human epidermal growth factor receptor 2-positive and triple-negative primary breast cancers. J Clin Oncol2015;33:983–91.

35. Liu S, Duan X, Xu L, Xin L, Cheng Y, Liu Q, et al. Optimal thresholdfor stromal tumor-infiltrating lymphocytes: its predictive and prog-nostic value in HER2-positive breast cancer treated with trastuzumab-based neoadjuvant chemotherapy. Breast Cancer Res Treat 2015;154:239–49.

36. Musolino A, Naldi N, Bortesi B, Pezzuolo D, Capelletti M, Missale G, et al.Immunoglobulin G fragment C receptor polymorphisms and clinicalefficacy of trastuzumab-based therapy in patients withHER-2/neu-positivemetastatic breast cancer. J Clin Oncol 2008;26:1789–96.

37. Perosa F, Luccarelli G, Prete M, Favoino E, Ferrone S, Dammacco F. Beta2-microglobulin-free HLA class I heavy chain epitope mimicry bymonoclonal antibody HC-10-specific peptide. J Immunol 2003;171:1918–26.

38. Ribas A, Wolchok JD. Cancer immunotherapy using checkpoint blockade.Science 2018;359:1350–5.






2019;25:1535-1545. Published OnlineFirst December 6, 2018.Clin Cancer Res Aura Muntasell, Federico Rojo, Sonia Servitja, et al. Disease-free SurvivalBreast Cancer Predict and Uncouple Pathological Response and

+NK Cell Infiltrates and HLA Class I Expression in Primary HER2

Updated version

10.1158/1078-0432.CCR-18-2365doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2018/12/06/1078-0432.CCR-18-2365.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/25/5/1535.full#ref-list-1

This article cites 38 articles, 15 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/25/5/1535.full#related-urls

This article has been cited by 1 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/25/5/1535To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-18-2365

http://clincancerres.aacrjournals.org/content/suppl/2018/12/06/1078-0432.CCR-18-2365.DC1

http://clincancerres.aacrjournals.org/content/25/5/1535.full#ref-list-1

http://clincancerres.aacrjournals.org/content/25/5/1535.full#related-urls

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/25/5/1535


nk cell inﬁltrates and hla class i expression in primary ... · ment in breast cancer....

Documents