Targeted spatial profiling identifies a CD8A − LARS axis associated with neoadjuvant chemotherapy resistance in triple-negative breast cancer

Targeted spatial profiling identifies a CD8A − LARS axis associated with neoadjuvant chemotherapy resistance in triple-negative breast cancer | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Targeted spatial profiling identifies a CD8A − LARS axis associated with neoadjuvant chemotherapy resistance in triple-negative breast cancer Misato Yamamoto, Ravi Velaga, Yuko Takano, Satoko Shimada, Madoka Iwase, and 9 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9572543/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 9 You are reading this latest preprint version Abstract Background Triple-negative breast cancer (TNBC) accounts for approximately 10–20% of all breast cancers and is characterized by aggressive clinical behavior and poor prognosis. Pathological complete response (pCR) after neoadjuvant chemotherapy (NAC) is associated with a good prognosis. Recently, the addition of pembrolizumab to NAC has improved pCR and overall survival. However, TNBC is biologically heterogeneous, and existing biomarkers, including tumor-infiltrating lymphocytes and PD-L1 provide limited prediction of chemoresistance, leaving an unmet need for markers that could guide integration of immunotherapy in resistant disease. In this study, we aimed to identify biomarkers of chemoresistance in TNBC and characterize their spatial features by integrating transcriptomic and spatial analyses. Methods We retrospectively analyzed 49 TNBC patients treated with anthracycline- and taxane-based NAC at Nagoya University Hospital between 2017 and 2023. Pre-treatment FFPE biopsies from eight patients (pCR n = 4, non-pCR n = 4) were profiled by nCounter gene expression analysis and Xenium spatial transcriptomics. The results were further validated using publicly available datasets, including TCGA. Results No clinicopathological factor was significantly associated with treatment response. nCounter analysis identified reduced expression of immune-related genes and increased expression of proliferation and cell cycle-related genes in non-responders. Spatial transcriptomic profiling revealed greater immune-cell abundance and diversity in the pCR group, with stronger immune-cell and immune-tumor cell interactions. In contrast, the non-pCR group showed enhanced stromal−epithelial interactions and reduced spatial proximity between immune and tumor cells. Cluster-and cell-type resolved analysis identified reduced expression of CD8A and other T-cell associated genes ( CD3E , GZMA ) and cell-cycle genes ( CCNA1 , CCNB1 , E2F3 ) in non-pCR tumors. The inverse spatial relationship between CD8A and LARS suggests as a candidate CD8A−LARS axis linking reduced cytotoxic T-cell presence with altered amino acid metabolism. Conclusions Reduced CD8A and elevated LARS expression in pretreatment TNBC tumors may contribute to chemoresistance through coordinated metabolic reprogramming, immune-cell exclusion, and tumor-cell proliferation. We propose the CD8A - LARS axis as a potential resistance axis warranting functional and prospective validation in independent TNBC cohorts. Triple negative breast cancer Neoadjuvant chemotherapy Pathological complete response Resistance marker Pretreatment specimen Gene expression profiling Spatial transcriptomics LARS Cell cycle Metabolic reprogramming Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Background Breast cancer is the most frequently diagnosed cancer among women worldwide. Triple-negative breast cancer (TNBC) accounts for 10–20% of breast cancer cases and is associated with a particularly poor prognosis owing to its aggressive clinical behaviour and lack of targeted therapies 1 , 2 . For early-stage TNBC, anthracycline- and taxane-based chemotherapy has long been the standard of care. 3 Neoadjuvant chemotherapy (NAC) has become standard in the management of stage II/III TNBC, supported by the finding that a pathological complete response (pCR) after NAC is associated with improved long-term outcomes 3 . NAC regimens including the addition of carboplatin have been devised to improve pCR rates 4 , 5 . In patients who fail to achieve a pCR, adjuvant capecitabine is a widely accepted additional treatment, as demonstrated in the CREATE-X trial 6 . Most recently, the KEYNOTE-522 (KN-522) trial showed that the addition of pembrolizumab to NAC significantly increased the pCR rate compared to chemotherapy alone (65% vs. 51%), with a 5-year overall survival rate of 86.6% in the pembrolizumab arm versus 81.7% in the placebo arm, leading to its adoption as a new standard of care for stage II/III TNBC 7 – 9 . Although the KN-522 trial demonstrated improved prognosis, long-term adverse events, especially immune-related adverse events, remain among approximately one-quarter of patients. Hypothyroidism and adrenal insufficiency occurred in approximately 13.7% and 2.3% of the patients, respectively, and some required long-term hormone replacement therapy 7 . Due to tumor heterogeneity and varying disease biology, it is difficult to confirm their therapeutic potential and use in a clinical setting. TNBC exhibits chemoresistance driven by the convergence of multiple signaling pathways (PI3K/AKT/mTOR, JAK/STAT, NF-κB), developmental programs (Notch, Wnt/β-catenin, Hedgehog, TGF-β), hypoxia-induced responses, and stemness-associated mechanisms, along with pronounced epigenetic and intratumoral heterogeneity 10 . During NAC, some dominant tumor clones are eliminated in certain patients but persist in others, contributing to treatment resistance 11 . Spatial profiling studies have revealed distinct immune “hot” and “cold” regions within TNBC tumors. Although TNBC exhibits the highest immune activation among breast cancer subtypes, only a subset of patients derive clinical benefit from immunotherapy. 12 Consequently, accurately characterizing tumor immune activation and evasion remains a major challenge to predicting response to NAC. Bulk transcriptomic profiling using RNA sequencing or multiplexed technologies such as the nCounter platform has provided insights into TNBC subtypes and potential mechanisms of chemoresistance 13 . However, these approaches lack spatial resolution and fail to capture the gene expression context in the tumor microenvironment (TME). Spatial transcriptomic technologies such as Xenium have enabled high-resolution mapping of gene expression within intact tissue sections, allowing for deeper interrogation of cell–cell interactions and spatially regulated resistance pathways 13 , 14 and elucidation of the TME. In this study, we performed transcriptomic analysis by combining gene expression profiling using nCounter and spatial transcriptomics using Xenium. We aimed to identify biomarkers and/or underlying mechanisms associated with chemoresistance in TNBC. Methods 1. Patient Selection and Data Inclusion This was a single-centre retrospective study of TNBC patients undergoing NAC and surgery at Nagoya University Hospital (January 2017 to December 2023). Inclusion required estrogen receptor (ER)-negative (< 1% of tumor cells), progesterone receptor (PgR)-negative (< 1%), and human epidermal growth factor receptor type 2 (HER2)-negative (IHC 0, 1+, or 2 + with negative in situ hybridization) status confirmed on pretreatment biopsy. Tumor-infiltrating lymphocytes (TILs) were assessed on H&E sections by two independent pathologists according to TILs Working Group (2014) guidelines. BRCA1/2 testing was performed using the BRCA analysis Test (Myriad Genetics) under the Japanese National Health Insurance system. NAC comprised epirubicin (90 mg/m²) and cyclophosphamide (600 mg/m²) every 21 or 14 days for four cycles, followed by docetaxel (75 mg/m², every 21 days for 4 cycles) or paclitaxel (80 mg/m² weekly for 12 cycles). pCR (ypT0/is ypN0) defined as responders; residual disease defined as non-responders. Univariate and multivariate analyses were conducted to explore the predictors of pCR among clinical factors. 2. nCounter Gene Expression Profiling Pretreatment formalin-fixed paraffin-embedded (FFPE) tumor sections from eight TNBC patients were analyzed, including four responders (achieving pCR after NAC; Y1–4) and four non-responders (with residual disease after NAC; Y5–8). Total RNA was extracted using the QIAGEN RNeasy FFPE Kit and assessed by spectrophotometry (260/280 and 260/230 ratios) and Agilent TapeStation. Gene expression profiling was performed using the NanoString nCounter Human Tumor Signalling 360 panel (760 genes, 13 housekeeping genes, and 6 positive control probes). Quality control (imaging quality, binding density, positive control linearity, housekeeping stability) and normalization (geometric mean of positive controls and stable housekeeping genes) were performed in ROSALIND® (OnRamp BioInformatics). Normalized counts were log₂-transformed. Differential expression analysis was performed using DESeq2-based methods within ROSALIND. Genes with P ≤ 0.05 and log₂ fold change ≥ 1.5 were considered significant and visualized using volcano plots and heatmaps. 3. Xenium Spatial Transcriptomics Targeted spatial transcriptomics was performed on matched FFPE sections (n = 8) using the 10x Genomics Xenium platform. A customized 330-gene Human Breast panel comprised the standard 280-gene panel supplemented with 50 differentially expressed genes identified from nCounter analysis. Gene annotations were based on Karaayvaz et al. (2018) 15 , Pal et al. (2021) 16 , and the Human Protein Atlas. Sections (5 µm) were processed according to the manufacturer’s protocol and imaged using Xenium Analyser Software v1.9.2.0. Data processing followed best practice guidelines 17 . Transcripts were excluded if they had (i) quality value < 20, (ii) duplicate coordinates, or (iii) anomalously high counts indicative of technical artifacts. Cell segmentation used Cellpose v2.0 18 for nuclear segmentation from DAPI images (fixed diameter 25 pixels), followed by Baysor 19 for transcript assignment using spatial coordinates and probabilistic modeling (scale 2.4 µm, SD 25%, minimum 30 molecules per cell, 4 clusters, prior segmentation confidence 0.8) (Fig. S1 A). Cells were retained if they met the following criteria: ≥ 12 transcripts, ≥ 10 unique genes, ≤ 5% negative control transcript rate, and nuclear area 6 − 80 µm², yielding 1,223,609 cells (responders 603,081; non-responders 620,528) (Fig. S1 B). Gene expression matrices were library-size normalized, and log₂-transformed. Clustering was performed using the Louvain algorithm in Seurat v5.2.1 20 . Cell type annotation integrated canonical marker expression, reference metadata (330_CellType.csv), and manual curation; 728 cells with missing spatial coordinates were excluded. Compositional differences in gene-of-interest (GOI)-expressing cells were assessed using Fisher’s exact test with Benjamini-Hochberg (BH) FDR correction (*FDR < 0.05, **< 0.01, ***< 0.001; † P < 0.05 nominal) for cellular heterogeneity and clustering patterns. Merged Seurat objects were analyzed using Principal Component Analysis (PCA) (50 PCs) and Uniform Manifold Approximation and Projection (UMAP) (dims 1 − 30). Spatial maps were generated using Baysor-derived coordinates to preserve tissue architecture. 4. Biomarker Selection and Spatial Differential Expression Resistance markers ( AURKA , VEGFA , KIT , WEE1 , LARS , WARS ) and cytotoxic immune markers ( CD3E , CD8A , CD247 , CD3G , GZMA ) were selected based on nCounter/Xenium differential expression using the following criteria: (i) P ≤ 0.05, (ii) expression in ≥ 2 clusters, (iii) presence in ≥ 2 tissues per group, and (iv) biological relevance. Pseudobulk analysis was performed using edgeR v3.40.2, with counts aggregated at the sample level and normalized using TMM. Genes with ≥ 10 counts in ≥ 2 samples were retained. Differential expression was assessed using quasi-likelihood F-tests with BH-FDR correction. Within-cluster comparisons used Welch’s t-tests (minimum 3 cells per group). Log₂ fold change (log₂FC) was calculated as log₂((mean_R + 0.01) / (mean_NR + 0.01)), where positive values indicate higher expression in responders. Results were visualized as signed -log₁₀( P ) heatmaps (* P < 0.05, **< 0.01, ***< 0.001; †FDR < 0.05, ††< 0.01, ††† 1.2 million cells. 5. Cell–Cell Communication and Spatial Neighborhood Analysis Intercellular communication was analyzed using CellChat v1.6.1 21 , CellChatDB.human and the 330-gene panel for broad interaction patterns, instead of any specific ligand−receptor (LR) pairings. Communication probabilities were computed using the triMean method under a law-of-mass-action model, with a minimum of 10 cells per cell type. Differential communication was visualized using circle and bubble plots (top 25 LR pairs). Spatially informed LR analysis was performed using Squidpy (sq.gr.ligrec; 100 permutations, BH-FDR; 154 testable pairs). Interactions with FDR < 0.05 were considered statistically significant. Pathway-level information flow was calculated as the sum of communication probabilities across all cell type pairs for each signaling pathway. To capture the physical proximity patterns between cell types in situ, spatial neighborhood graphs were constructed per sample using Delaunay triangulation (sq.gr.spatial_neighbors) 17 . Cell type co-localization was quantified using neighborhood enrichment (sq.gr.nhood_enrichment, 100 permutations). 6. Functional Module Analysis and Statistical Testing Cell type annotation was refined using Louvain clustering 22 at multiple resolutions (0.3, 0.5, 0.8) in Seurat v5.2.1 with SCTransform 23 normalization. Functional module scores were calculated using Seurat’s AddModuleScore() to capture immune ( CD8A ), metabolic ( LARS ), and cell cycle ( CCNA1 , CCNA2 , CCNB1 , CCND1 , CCNE1 , CDC14A , CDC20 , E2F3 ) axes. Per-cell module scores were analyzed spatially and by treatment response. Associations between clinical variables and pCR were assessed using bias-adjusted logistic regression (JMP18; P < 0.05 significant). Spatial comparisons used Wilcoxon rank-sum tests with BH-FDR correction. Correlations among module scores were evaluated using Spearman’s rank correlation. 7. Cross-Validation Using External Cohorts Pretreatment TNBC Visium data (Wu et al., 2021; GSE176078; samples CID4465 and CID44971) 24 were log-normalized in Seurat v4.3.0 (scale factor 10,000). Pearson correlations between Visium samples and Xenium samples (mean expression across 330 genes) were used to assign external samples to responder-like or non-responder-like profiles. LARS -correlated genes in CID44971 were identified using Spearman correlation (FDR < 0.05) and functionally annotated. PCA (centred and scaled; 68% confidence ellipses) was performed using the full 330-gene panel and nine GOI ( CD8A , LARS , CCNA1 , CCNA2 , CCNB1 , CCND1 , CDC14A , CDC20 , E2F3 ). LARS survival analysis was performed using the TCGA 2012 basal breast cancer cohort (n = 98) 25 in Xena 26 , with median-based stratification. Gene set enrichment analysis was conducted using blitzGSEA 27 , with significance defined as P ≤ 0.05 and FDR ≤ 0.05. Computational Tools Data processing was performed using Python (quality control and segmentation) and R (statistical analysis and visualization), including Cellpose v3.1.1, Baysor v0.6.2, Seurat v5.2.1, CellChat v1.6.1, Squidpy v1.2.0 and edgeR v3.40.2. Results Association of clinicopathological characteristics and treatment response: To determine the clinicopathological characteristics that serve as potential predictive markers, we evaluated 49 patients with stage I-III TNBC who underwent surgery after NAC. The median patient age was 50 years (range 26–73 years). Clinical stage II disease represented 55.1% of the cases and 34.7% had clinically node-positive disease. TILs were evaluated in 31 patients (Table. S1). The pCR rate in all patients was 34.7%. Univariate analysis revealed that age and BRCA status were significantly associated with pCR (age > 50; P = 0.0434, with BRCA1/2 pathogenic variants; P = 0.0267, Table. 1). Other clinicopathological characteristics, including TILs levels, were not significantly associated with pCR. Age was not a significant variable in the multivariable analysis. Although BRCA status was significantly associated with pCR in the univariate analysis, it was excluded from the multivariate model because of the small number of BRCA -pathogenic cases (n = 5), which could have resulted in unstable estimates or overfitting. Transcriptomic profiling using nCounter reveals immunosuppression and proliferative signatures associated with chemotherapy resistance in TNBC: To identify transcriptional changes associated with chemoresistance, we compared pretreatment gene expression between TNBC responders (n = 4) and non-responders (n = 4) by nCounter profiling (Table S2 ). Differential expression analysis was performed to identify 83 differentially expressed genes (DEGs) ( P ≤ 0.05, Fig. 1 , Table S3 ), comprising 52 downregulated and 31 upregulated genes in non-responders. Notably, the genes downregulated in non-responders included CD247, GZMA , CD27 , CXCR6 , CCR2 , and PRF1 markers associated with cytotoxic T-cell activation and immune signalling. Upregulated genes in non-responders included LOX , WEE1 , CDK6 , ARID1A , and RUNX1 , indicating cell cycle regulation and stromal remodelling in chemoresistance. Spatial transcriptomics highlights reduced immune composition and altered tumor architecture: Transcriptional enrichment and Cell type composition: To spatially contextualise the transcriptional differences observed in the nCounter analysis and for novel/additional findings, we profiled the same tumors using a 10x Genomics Xenium breast cancer targeted panel (330 genes consisting of 280 genes and 50 custom genes) (Table. S4). In a unified UMAP embedding of all Y1-Y8 samples. cells clustered primarily by cell type identity rather than treatment response group, indicating conserved transcriptional programs across responders and non-responders, preserving distinct immune, stromal, and epithelial compartments (Figs. 2 A-B). After detailed cell-type annotation, responders displayed a diverse mixture of immune subtypes, including T cells, NK cells, macrophages, dendritic cells, whereas non-responders showed reduced immune diversity and relatively homogeneous epithelial and stromal populations (Fig. 2 B-C), although statistically significant diversity was not observed. We acknowledge that the limited sample size (n = 4 per group) and the targeted gene panel constrain statistical power for detecting differences in less abundant populations. Marker gene analysis to differentiate responders and non-responders: Differential expression analysis identified CCR2 , a chemokine receptor critical for immune cell recruitment, as the most significantly downregulated gene in non-responders (log2FC = -1.79, P = 0.008; Fig. 2 D, Table. S5). This reduction was consistent across multiple cell populations, including T cells ( P = 0.007), breast cancer cells ( P = 0.004), and fibroblasts ( P = 0.014), suggesting a broad reduction in CCR2 -mediated signalling in the non-responding tumor microenvironment. Conversely, WEE1 , a cell cycle checkpoint kinase and emerging therapeutic target, was significantly upregulated in non-responders (log2FC = 1.15, P = 0.040), particularly in breast cancer cells ( P = 0.017) (Fig. 2 D, Table. S5). To resolve the cellular contribution of these and other candidate genes, we examined the spatial and cluster-level expression patterns of all panel genes across responders and non-responders. To validate genes identified through cluster-based discovery, we performed systematic differential expression analysis across clusters and cell types with multiple testing correction (Fig. 3 A-B). Marker genes were selected based on consistent differential expression across multiple clusters, within and between biological replicates (Fig. S2 C, Tables. S6-7). Cell type composition analysis revealed distinct cellular sources for immune, metabolic and cell cycle markers (Fig. S3 A, Table. S8). Immune-related genes ( CD3G , GZMA , CD3E , CD247 , and CD8A ) were predominantly expressed by T cells, whereas CCR2 showed a more distributed pattern in T cells, followed by B cells (notably in responders) and macrophages (Fig. S3 A). CXCL12 displayed the most heterogeneous profile, expressed largely by breast glandular cells, followed by fibroblasts, macrophages, and T cells. Resistance and metabolism associated markers WEE1 , VEGFA , KIT , AURKA , LARS and WARS were predominantly expressed by breast glandular cells (Fig. S3 A). Comparison of cell type composition between responders and non-responders revealed systematic differences in the cellular sources of gene expression (Fig. S3 B). T cells contributed a significantly higher proportion of expressing cells in responders across all 13 candidate genes (FDR < 0.001) (Fig. S3 B) whereas, breast glandular cells contributed significantly higher in non-responders for resistance markers including AURKA , LARS , VEGFA , and WEE1 (FDR < 0.001) (Fig. S3 B). Cluster- and cell-type resolved differential expression supported these patterns (Fig. 3 A, Table. S9-10). CD8A was consistently elevated in responders across multiple clusters (FDR < 0.001), supporting enhanced cytotoxic immune activity in treatment-sensitive tumors, and was higher in responders within breast cancer cells, breast myoepithelial cells, adipocytes, endothelial cells, and macrophages (5/14 cell types; FDR < 0.001). The single exception was within T cells themselves, where CD8A was higher in non-responders, likely reflecting differential T-cell functional states rather than abundance. At cluster level, LARS showed an inverse pattern, elevated in non-responders in half of the clusters (6/12 clusters FDR < 0.001), versus 4/12 clusters enriched in responders (Fig. 3 A). Cell-type analysis confirmed elevated LARS in non-responders within breast glandular cells, fibroblasts, and T cells (Fig. 3 B, Table. S10). Cell cycle genes ( CCNA1 , CCNA2 , CCNB1 , CCND1 , CDC14A , CDC20 , E2F3 ) were similarly elevated in non-responders at both cluster and cell-type levels (Fig. 3 B, Fig S3 ). Together, the cluster and cell-type gene expression patterns suggest a composite axis integrating immune activity, translational capacity, and proliferative stress that distinguishes responders from non-responders. Cell Communication and Spatial Rewiring distinguish chemotherapy response patterns: Having characterized transcriptional differences at cluster and cell-type levels, we next examined how these were reflected in tissue organisation and intercellular signalling. Spatial neighbourhood analysis (Fig. 3 C) and CellChat (Fig. 4 A) based ligand-receptor inference identified distinct microenvironmental compositions and intercellular communication architecture between groups. Both approaches identified NK cells and dendritic cells exclusively in responders, consistent with intact innate immune surveillance. In non-responders, stromal interactions dominated both spatial proximity and cell-cell interaction networks, with fibroblasts, macrophages, and mast cells forming the principal communication hubs. CellChat analysis revealed distinct communication architectures between groups (Fig. 4 B). In responders, endothelial cells and macrophages served as central communication hubs with active immune cell participation, including T cell–adipocyte crosstalk. In non-responders, adipocyte autocrine signalling and a breast cancer-B cell communication axis were prominent, while macrophage network engagement was diminished. Twenty significant ligand-receptor (L-R) pairs were idendified in responders (Table. S11) across six signalling pathways ( PECAM1 , CD45 , CXCL , CDH , VEGF ), compared with sixteen pairs across five pathways in non-responders ( PECAM1 , VEGF , CXCL , CDH , NCAM ). All detected interactions achieved statistical significance ( P < 0.001, FDR < 0.001). The strongest communication probability was observed for homophilic PECAM1 - PECAM1 interactions, with endothelial cell autocrine signalling dominant in both groups (responders: prob = 0.51; non-responders: prob = 0.523; Fig. 4 C, Table. S11). However, PECAM1 pathway information flow was substantially higher in responders (3.56) compared to non-responders (2.48). Notably, the CD45 pathway ( PTPRC - MRC1 ) (Fig. 4 C) was detected exclusively in the responders (information flow = 0.20), with NK cells, dendritic cells, T cells and macrophages signalling to macrophages (prob = 0.044–0.054; Fig. 4 C). Breast cancer cells in responders also exhibited CXCL12 - CXCR4 signalling to T cells (prob = 0.035, P < 0.001; Fig. 4 C, Table. S11), consistent with active chemokine-mediated immune recruitment. Intra- and inter-patient spatial Heterogeneity: Integration of spatial cell-type distributions (Fig. 5 A) with CellChat-derived communication networks (Fig. 5 B) revealed distinct patterns of inter-patient heterogeneity. Among responders, Y3 displayed a highly compartmentalised spatial architecture coupled with a dense, multi-cellular communication network in which T cells served as a central signalling hub (Fig. 5 A-B). In contrast, Y1 showed abundant T- and B-cell infiltration but limited communication activity, indicating a discordance between immune presence and signalling engagement. Y2 and Y4 exhibited intermediate patterns, with structured spatial organisation and less extensive communication networks (Fig. 5 B). Non-responders, by comparison, showed more uniform spatial and signalling profiles. Adipocyte autocrine signalling was observed across all four non-responders and was accompanied by statistically significant spatial clustering of adipocytes, particularly in Y7 and Y8 (Fig. 5 A-B). Fibroblast-mediated communication was likewise consistent across non-responders and corresponded to expanded fibroblast spatial domains in Y6 − Y8 (Fig. 5 A-B). Together, these patterns suggest that non-response is associated with a reproducible stromal-communication network dominated by adipocyte and fibroblast signalling, whereas responder tumors encompass a wider range of immune-stromal architectures, from limited communication despite immune infiltration (Y1) to highly active immune-stromal crosstalk (Y3), indicating that treatment sensitivity in TNBC may arise through diverse microenvironmental configurations. Sample-level transcriptional analysis revealed crucial insights into intra-tumor heterogeneity (Fig. 6 A, Fig. S4 ). Responder tumors contained distinct transcriptional domains with high cytotoxic T cell marker expression and coordinated immune-cell clustering were, whereas non-responder tumors were dominated by resistance marker-rich regions with minimal immune infiltration and enhanced stromal compartmentalisation (Fig. S4 ). This spatial segregation reflects the limitations of targeted expression analysis in underestimating the true nature of potential biomarkers in detecting and understanding intra-tumor and spatial heterogeneity. Despite this heterogeneity, group-level marker patterns were preserved. UMAP visualization showed robust expression of T-cell markers T cell markers ( CD3E , CD8A , CD247 , CD3G ) and the cytotoxic effector molecule GZMA in responders, with reduced expression in non-responders. Conversely, the cell-kinase WEE1 and the metabolic enzyme LARS were elevated in non-responders (Fig. 6 A). Spatial mapping of gene expression confirmed these patterns at the tissue level (Fig. 6 B). Responders harboured significantly more CD8A + (56,981 vs 33,049), CD3E + (67,696 vs 35,551), and GZMA + cells (45,777 vs 19,607) (Table. S12), with immune cells distributed throughout the tumor. Notably, CCR2 + cells, were 3.9-fold more abundant in responders (32,949 vs 8,403) (Table. S12). In contrast, non-responders showed expanded populations of WEE1 + cells (156,559 vs 78,759) and LARS + cells (120,720 vs 97,518) (Table. S12), suggesting enhanced cell cycle checkpoint activity and altered metabolic programming. Per-cell expression of T-cell markers was comparable between groups (Fig. 6 C), indicating that the differences reflected changes in cell abundance rather than per-cell transcriptional outputs (Table. S13). Cross-Cohort support of CD8A and LARS axis using Spatial and TCGA TNBC cohorts : To seek independent support for the CD8A and LARS patterns observed in our Xenium TNBC cohort, we examined publicly available TNBC spatial transcriptomics data from Wu et al 24 . Of ten TNBC pretreatment samples profiled by using 10x Visium, two samples (CID4465 and CID44971) contained spatial gene expression data overlapping our Xenium panel. Pearson correlation across all 330 panel genes (Fig. 7 A) showed that CID4465 most closely resembled responders in our cohort, while CID44971 displayed a mixed pattern with moderate correlations to both groups. Direct comparison of LARS and CD8A expression further showed that CID44971 (Fig. 7 B) most closely resembled Y7, a non-responder characterised by elevated LARS and reduced CD8A . To examine the transcriptional context of LARS expression in CID44971, we constructed a LARS -correlation network and conducted a pathway enrichment analysis. Significantly enriched (FDR < 0.05) included ECM remodelling, cell cycle, mTOR signalling, amino acid metabolism, and immune response (Fig. 7 C). We next asked whether a focused 9-gene set ( CD8A , LARS , CCNA1 , CCNA2 , CCNB1 , CCND1 , CDC14A , CDC20 and E2F3 ) could capture the responder- and non-responder distinction captured by the full panel (330 genes). PCA on the full 330 gene panel separated responders and non-responders along PC1 (95.5% variance; Fig. 7 D, left). Strikingly, the 9-gene set preserved this separation (PC1: 99.9% variance; Fig. 7 D, right). While these findings support further evaluation of the 9-gene set as a candidate stratifier of pretreatment TNBC tumors, validation in larger, independent cohorts will be required. Survival predictability of LARS expression with a median of 11.4 in TCGA basal subtype cohort showed a trending worse prognosis for disease free interval in LARS high tumors ( P = 0.0958, and log rank test statistic = 2.773; Fig. 7 E, left) also serves as supporting evidence. Blitz gene set enrichment analysis (blitzGSEA) found significant positive enrichment of HALLMARK_E2F_TARGETS, HALLMARK_G2M_CHECKPOINT, and HALLMARK_MTORC1_SIGNALING (NES = 8.123, 7.469 and 5.501; P values = 4.412e-16, 8.059e-14 and 3.766e-08; FDR = 2.206e-14, 1.007e-12 and 2.337e-07) (Fig. 7 E, middle and right, Table. S14) in LARS high tumors, recapitulating the cell-cycle and metabolic features observed in non-responders by Xenium. ImmuneSigDB analysis in the same cohort resulted in positive enrichment of two CD8 T-cell gene sets including, GSE10239_MEMORY_VS_DAY4.5_EFF_CD8_TCELL_DN (NES = 7.224; P = 5.018e-13; FDR = 4.075e-10) and GSE15750_DAY6_VS_DAY10_EFF_CD8_TCELL_UP (NES = 7.082, P = 1.411e-12; FDR = 6.646e-10), suggesting altered effector/memory CD8 T-cell dynamics in LARS -high tumors. Canonical pathway enrichments included, positively enriched REACTOME_SELENOAMINO_ACID_METABOLISM (NES = 4.389; P = 1.136e-05; and FDR = 0.003), REACTOME_GLUCOSE_METABOLISM, REACTOME_MITOCHONDRIAL_BIOGENESIS,REACTOME_MITOCHONDRIA_PROTEIN_IMPORT and REACTOME_CHEMOKINE_RECEPTORS_BIND_CHEMOKINES (NES = 3.291, 3.257, 3.201, and 2.886; P = 9.9e-4, 0.001, 0.001 and 0.003) in the LARS high tumors suggesting heterogeneous cell cycle, immune and metabolic reprogramming molecules in LARS differentially expressed tumors, which supports our Xenium data findings. Together, these cross-cohort observations support the CD8A − LARS axis as a candidate framework linking immune evasion as a predictor of treatment response in this small cohort. Immune evasion via low CD8A infiltration and high LARS activity revealed a potential alternative chemoresistance axis in TNBC. Validation in larger, prospectively collected cohorts will be required to assess its predictive utility. Discussion Clinicopathological analysis revealed that TILs, a widely reported prognostic feature of TNBC 28 , 29 , were not significantly associated with pCR in our cohort, suggesting that TIL density alone may be insufficient to predict chemotherapy response. This observation prompted us to interrogate the molecular and spatial features of pretreatment tumors to better identify potential predictive biomarkers and mechanisms underlying NAC resistance in TNBC. Our targeted transcriptomic and spatial analyses identified a previously unrecognised association between immune surveillance and metabolic reprogramming in chemoresistant TNBC. The inverse relationship between CD8A – LARS expression, together with cell-cycle dysregulation and the LARS -functional enriched pathways, provides hypothesis generating insights into the molecular and spatial determinants of differential treatment response. These findings present an apparent contextual contrast with the established role of leucyl-tRNA synthetase ( LARS ) in TNBC biology. Passarelli et al. recently demonstrated that LARS functions as a tumor suppressor in breast cancer, with genetic deletion enhanced tumor formation and LARS repression reduced translation of growth-suppressive genes, including EMP3 and GGT5 during mammary epithelial to mesenchymal cell transformation and in human and mouse breast cancers 30 . LARS is known to preferentially regulate RagD , but not RagC , thereby modulating mTORC1 recruitment and TFEB/TFE3 phosphorylation in lysosomes 31 . In contrast, our spatial data indicate that LARS is elevated in chemoresistant TNBC tumors, where significantly elevated LARS expression accompanied by immune exclusion ( CD8A low expression), and elevated proliferative markers. Consistent with this, high LARS expression tumors in the TCGA basal breast cancer cohort also demonstrated enriched activity of mTORC1, E2F and G2M checkpoint gene sets. These observations may reflect context dependent roles of LARS . Differences in tissue context, spatial cellular composition and the model organism in which LARS is regulated, likely contribute to this divergence. In addition, Passarelli et al. characterised the monoallelic deletion of LARS and the functional consequences on downstream t-RNAs which were beyond the scope of our study but remain important for mechanistic follow-up. This context-dependent gene expression pattern may reflect the broader functional plasticity of aminoacyl-tRNA synthetases in TNBC biology. While LARS canonically supports protein synthesis and has been shown to act as a tumor suppressor by prompting growth-inhibitory gene translation, metabolic stress conditions such as glucose starvation can trigger post-translational modifications like LARS phosphorylation and functional reprogramming 32 via mTORC1 activation 33 . Our data suggest that LARS may switch from a tumor-suppressive role during initiation toward a resistance-promoting factor, where enhanced leucine sensing and mTORC1 activation may support survival and proliferation under therapeutic stress. Our observations are consistent with previous findings that activated CD8 + T cells rely heavily on leucine uptake, which fuels mTORC1 activation and c-Myc expression to support clonal expansion and effector differentiation 33 . Elevated LARS activity in tumor cells may therefore create a competitive metabolic environment that constrains CD8⁺ T-cell expansion locally, contributing to the immune exclusion observed in our LARS-high, CD8A-low resistant tumors. The clinical efficacy of immune checkpoint inhibitors in TNBC, demonstrated by the KEYNOTE-522 trial, underscores the importance of the immune microenvironment in determining treatment response and provides a clinical rationale for resolving the spatial and cellular features that distinguish immunologically engaged from immunologically inactive tumors 7 , 8 and reported its critical use in estimating immune infiltration for treatment efficacy. Our spatial analysis indicates that chemotherapy resistance in TNBC is characterised by altered spatial architecture, with resistant tumors showing reduced T cell-cancer cell proximity and enhanced fibroblast-mediated stromal compartments. The exclusive presence of CD45 pathway signalling ( PTPRC - MRC1 , information flow = 0.20) in chemotherapy responders provides mechanistic insight into differential treatment outcomes. The potential PTPRC - MRC1 interactions from NK cells, dendritic cells, T cells, and macrophages with macrophages points to a distinct macrophage-engaged communication architecture in treatment-sensitive tumors. The absence of this signalling axis in non-responders may reflect an immunologically "cold" tumor microenvironment characterised by limited immune cell infiltration, consistent with reduced chemotherapy efficacy. The cell-type annotation distinction between "breast glandular cells" and “breast cancer cells” in the Xenium breast cancer panel warrants methodological consideration. The more extensive communication observed in the glandular-annotated population may therefore reflect heterogeneity within the malignant compartment rather than non-malignant epithelial activity. Both cell populations engaged in biologically meaningful signalling. Breast cancer cells engaged through CXCL12 - CXCR4 chemokine signalling 34 ,and breast glandular cells through VEGFA - VEGFR2 (KDR) angiogenic signalling 35 , consistent with prior spatial transcriptomic analyses of breast cancer microenvironments describing heterogeneous tumor cell states and context-dependent stromal interactions 24 , 36 . CCR2 and CXCL12 , identified through differential expression and ligand-receptor analyses respectively, showed concordant responder-enriched expression across clusters and cell types. The elevated CXCL12 in responders is consistent with the CXCL12-CXCR4 signalling identified by CellChat and supports a model of active chemokine-mediated T cell recruitment in responders. Future studies using expanded panels or whole-transcriptome spatial platforms will be required to resolve the cell-state distinctions, which are limited in a targeted panel study. Together, these findings are consistent with emerging evidence that TME architecture influences response to NAC in TNBC 37 . The inverse relationship between CD8A and LARS expression observed in our cohort suggests an exploratory metabolic-immune resistance axis candidate in TNBC. Recent studies have shown that aminoacyl-tRNA synthetases, particularly VARS, promote therapeutic resistance in melanoma through codon-biased translational reprogramming and fatty acid oxidation 38 . Whether LARS plays a comparable role in TNBC, remains to be established. Our findings present LARS-associated metabolic stress, as a potential candidate in TNBC resistance, suggesting that LARS -mediated metabolic reprogramming could antagonize cytotoxic T cell function, subject to further validation of T cell and metabolic cells states in a larger cohort. Amino acid metabolism, enriched in the LARS -associated network of CID44791 sample examined here, emerged as a critical determinant of tumor immunity, with amino acid sensors including mTORC1 integrating the metabolic status with immune cell differentiation and disease function 24 , 39 , 40 . The spatial segregation of LARS -high and CD8A -low regions in resistant tumors could result in metabolic reprogramming, creating immunosuppressive microenvironments, although the directionality of whether metabolic state shapes immune exclusion or vice-versa needs further examination. The coordinated upregulation of cell cycle genes ( CCNA1 , CCNB1 , E2F3 and others) in non-responders, together with HALLMARK_E2F_TARGETS enrichment in LARS -high TCGA basal tumors, indicates a proliferative axis associated with chemoresistance in our cohort. This pattern is notable because highly proliferative tumors have conventionally been associated with greater chemotherapy sensitivity; with elevated proliferation, metabolic ( LARS -high) and immune-excluded ( CD8A -low) features in non-responders. This suggests that proliferation alone is an insufficient predictor of response, and that the broader cellular context determines treatment outcome. The integration of metabolic reprogramming ( LARS ), immune dysfunction ( CD8A loss), and elevated proliferation in non-responders suggests a multi-layered resistance network that may benefit from combination therapeutic approaches. Our study revealed that these mechanisms are not uniformly distributed, demonstrating intra-tumor and spatial heterogeneity, and are organised into distinct resistance niches that could be targeted with spatially directed therapies. Building on these observations, we propose a candidate LARS -RagD- mTORC1 - CD8 T -cell axis as a working framework that could be distinct from the classical PD-1/PD-L1 directed therapeutic approaches. Limitations and future directions: Our study comprised relatively small cohort size (eight samples), different targeted gene candidates with some overlap between the nCounter and Xenium panels, and a focus on pretreatment samples alone. Cross-cohort validation against the Wu et al. spatial atlas was based on the two TNBC samples with available spatial data, with one sample (CID44971) contributing the principal cross-cohort similarity to non-responders; broader validation in independent spatial cohorts will be required. Future studies should examine the spatial evolution during treatment and validate the findings in larger cohorts and whole transcriptomes. Functional validation of the CD8A – LARS resistance axis and its therapeutic targeting need to be conducted in future studies. Realising any of the LARS-associated metabolic adaptation, modulating fibroblast-rich stromal compartments that support immune evasion and developing biomarker guided combinational immunotherapy directions will require functional validation in TNBC models and prospective evaluation in larger, well-characterised cohorts. Although the present cohort received chemotherapy without immune checkpoint inhibitors (ICIs), evaluating CD8A and LARS expression patterns in patients receiving chemoimmunotherapy regimens, represents one of the important next steps, particularly to determine whether these features inform response to ICI combination therapies. Conclusions Our spatial transcriptomics analysis of pretreatment TNBC tumors identifies a candidate metabolic-immune resistance axis defined by inverse spatial relationship between LARS and CD8A accompanied by elevated proliferative gene expression in non-responders. Although these findings are exploratory and require validation in larger, prospectively collected cohorts, they support the value of spatial transcriptomics in resolving the cellular and metabolic features of chemoresistance and provide a hypothesis-generating framework, the CD8A-LARS axis, for future mechanistic and biomarker focused investigation in TNBC. Abbreviations BH Benjamini-Hochberg DEGs Differentially expressed genes ER Estrogen receptor FFPE Formalin-fixed paraffin-embedded GOI Gene-of-interest HER2 Human epidermal growth factor receptor 2 ICI Immune checkpoint inhibitor KN-522 KEYNOTE-522 trial LARS Leucyl-tRNA synthetase Log 2 FC Log 2 fold change LR Ligand-receptor NAC Neoadjuvant chemotherapy PCA Principal component analysis pCR Pathological complete response PgR Progesterone receptor TILs Tumor-infiltrating lymphocytes TME Tumor microenvironment TNBC Triple-negative breast cancer UMAP Uniform manifold approximation and projection. Declarations Ethics approval and consent to participate The study was approved by the Nagoya University Hospital IRB (2022 − 0244, 2023-0066), and all patients provided informed consent. Consent for publication Not applicable Competing Interests Norikazu Masuda receives grants from Chugai Pharmaceutical Co.,Ltd., Eli Lilly Japan K.K. , AstraZeneca K.K., Pfizer Inc., Daiichi Sankyo Company, Ltd., MSD K.K., Eisai Co., Ltd., Gilead Sciences, Inc. and Ono Pharmaceutical Co., Ltd.; and honoraria from Chugai Pharmaceutical Co.,Ltd., Pfizer Inc., AstraZeneca K.K., Eli Lilly Japan K.K. , Daiichi Sankyo Company, Ltd., Eisai Co., Ltd., Gilead Sciences, Inc. and MSD K.K.; and serves as a representative of the Board of Directors of the Japan Breast Cancer Research Group (JBCRG) (unpaid, 2021–2025), has served as a member of the JBCRG Board of Directors (unpaid) since 2007, and is a member of the Board of Directors of the Japanese Breast Cancer Society (JBCS) (unpaid, 2021–2024), Japan Society of Clinical Oncology (JSCO) (unpaid, since 2023), Japan Association of Breast Cancer Screening (JABCS) (unpaid, since 2024), and Kyoto Breast Cancer Research Network (KBCRN) (unpaid, since 2025). Tomoki Ebata receives honoraria from AstraZeneca K.K. and MSD K.K.Yuko Takano receives grants from Eli Lilly Japan K.K. and honoraria from Chugai Pharmaceutical Co., Ltd., Daiichi Sankyo Company, Ltd., MSD K.K., Eli Lilly Japan K.K., and AstraZeneca K.K.Madoka Iwase receives grants from AstraZeneca K.K., Daiichi Sankyo Company, Ltd., Eli Lilly Japan K.K., MSD K.K., Ono Pharmaceutical Co., Ltd., Chugai Pharmaceutical Co., Ltd., Eisai Co., Ltd., Gilead Sciences, Inc., Novartis AG, and Pfizer Inc.; and honoraria from Chugai Pharmaceutical Co., Ltd., Eli Lilly Japan K.K., Pfizer Inc., Taiho pharmaceutical Co.Ltd., Nipro Corporation, Daiichi Sankyo, Co., Ltd., MSD K.K., Kyowa Kirin Co., Ltd., and Exact Sciences Co.The remaining authors declare no competing interests. Funding The Moonshot Research and Development Program(grant no. JP22zf0127009)from the Japan Agency for Medical Research and Development (AMED). Author Contribution Conceptualization, M. Y., Y. T., R. V., and N. M.; Methodology, M. Y., R. V., Y. T., and N. M.; Investigation, M. Y., R. V., S. S., and N. M.; Resources, K. F., S. M., K. I., A. E., and N. M.; Formal analysis, M. Y., R. V., and N. M.; Writing – Original Draft, M. Y., R. V., and N. M.; Writing – Review & Editing, M. Y., R. V., Y. T., and N. M.; Final manuscript approval, all authors; Supervision, N. M. Acknowledgement This study was supported by The Moonshot Research and Development Program (Grant No. JP22zf0127009) of the Japan Agency for Medical Research and Development (AMED). The authors wish to acknowledge the Division for Medical Research Engineering, Nagoya University Graduate School of Medicine, for the use of Xenium. We thank Mr. Yamaguchi (Technical Center, Nagoya University) for technical support (Xenium sample processing) and data acquisition. Data Availability This study did not report the original code. We adapted the segmentation and DAPI channel extraction steps from Cellpose's documentation and workflows18 and the Baysor GitHub repository, modified for Julia CLI integration with Xenium outputs, and incorporated STRtree-based transcript-to-cell assignment19. For UMAP embedding, Louvain clustering, cell type annotation, and marker gene identification, we adapted the code from Satija R et al., 201520. The CellChat scripts were adapted from the CellChat R package21. Our spatial neighborhood scripts (cell-cell distance, interaction matrices, and niche-DE style analysis) were adapted from Salas M et al., 202517. The unprocessed data used in this study are available upon reasonable request from the lead authors. The validation dataset is available from Wu SZet al24. (https://doi.org/10.5281/zenodo.4739739). References Huang M, Haiderali A, Fox GE, Frederickson A, Cortes J, Fasching PA, O'Shaughnessy J. (2022). Economic and Humanistic Burden of Triple-Negative Breast Cancer: A Systematic Literature Review. Pharmacoeconomics 40 , 519–558. 10.1007/s40273-021-01121-7 Won KA, Spruck C. Triple–negative breast cancer therapy: Current and future perspectives (Review). Int J Oncol. 2020;57:1245–61. 10.3892/ijo.2020.5135 . Jie H, Ma W, Huang C. Diagnosis, Prognosis, and Treatment of Triple-Negative Breast Cancer: A Review. Breast Cancer (Dove Med Press. 2025;17:265–74. 10.2147/bctt.s516542 . Poggio F, Bruzzone M, Ceppi M, Pondé NF, La Valle G, Del Mastro L, de Azambuja E, Lambertini M. Platinum-based neoadjuvant chemotherapy in triple-negative breast cancer: a systematic review and meta-analysis. Ann Oncol. 2018;29:1497–508. 10.1093/annonc/mdy127 . von Minckwitz G, Schneeweiss A, Loibl S, Salat C, Denkert C, Rezai M, Blohmer JU, Jackisch C, Paepke S, Gerber B, et al. Neoadjuvant carboplatin in patients with triple-negative and HER2-positive early breast cancer (GeparSixto; GBG 66): a randomised phase 2 trial. Lancet Oncol. 2014;15:747–56. 10.1016/s1470-2045(14)70160-3 . Masuda N, Lee SJ, Ohtani S, Im YH, Lee ES, Yokota I, Kuroi K, Im SA, Park BW, Kim SB, et al. Adjuvant Capecitabine for Breast Cancer after Preoperative Chemotherapy. N Engl J Med. 2017;376:2147–59. 10.1056/NEJMoa1612645 . Schmid P, Cortes J, Pusztai L, McArthur H, Kümmel S, Bergh J, Denkert C, Park YH, Hui R, Harbeck N, et al. Pembrolizumab for Early Triple-Negative Breast Cancer. N Engl J Med. 2020;382:810–21. 10.1056/NEJMoa1910549 . Schmid P, Cortes J, Dent R, Pusztai L, McArthur H, Kümmel S, Bergh J, Denkert C, Park YH, Hui R, et al. Event-free Survival with Pembrolizumab in Early Triple-Negative Breast Cancer. N Engl J Med. 2022;386:556–67. 10.1056/NEJMoa2112651 . Schmid P, Cortes J, Dent R, McArthur H, Pusztai L, Kümmel S, Denkert C, Park YH, Hui R, Harbeck N, et al. Overall Survival with Pembrolizumab in Early-Stage Triple-Negative Breast Cancer. N Engl J Med. 2024;391:1981–91. 10.1056/NEJMoa2409932 . Panda VK, Mishra B, Mahapatra S, Swain B, Malhotra D, Saha S, Khanra S, Mishra P, Majhi S, Kumari K et al. (2025). Molecular Insights on Signaling Cascades in Breast Cancer: A Comprehensive Review. Cancers (Basel) 17 . 10.3390/cancers17020234 Ferrari P, Scatena C, Ghilli M, Bargagna I, Lorenzini G, Nicolini A. (2022). Molecular Mechanisms, Biomarkers and Emerging Therapies for Chemotherapy Resistant TNBC. Int J Mol Sci 23 . 10.3390/ijms23031665 Imani S, Farghadani R, Roozitalab G, Maghsoudloo M, Emadi M, Moradi A, Abedi B, Kaboli J, P. Reprogramming the breast tumor immune microenvironment: cold-to-hot transition for enhanced immunotherapy. J Exp Clin Cancer Res. 2025;44:131. 10.1186/s13046-025-03394-8 . Jing SY, Wang HQ, Lin P, Yuan J, Tang ZX, Li H. Quantifying and interpreting biologically meaningful spatial signatures within tumor microenvironments. NPJ Precis Oncol. 2025;9:68. 10.1038/s41698-025-00857-1 . Mo CK, Liu J, Chen S, Storrs E, Targino da Costa ALN, Houston A, Wendl MC, Jayasinghe RG, Iglesia MD, Ma C, et al. Tumour evolution and microenvironment interactions in 2D and 3D space. Nature. 2024;634:1178–86. 10.1038/s41586-024-08087-4 . Karaayvaz M, Cristea S, Gillespie SM, Patel AP, Mylvaganam R, Luo CC, Specht MC, Bernstein BE, Michor F, Ellisen LW. Unravelling subclonal heterogeneity and aggressive disease states in triple-negative breast cancer through single-cell RNA-seq. Nat Commun. 2018;9:3588. Pal B, Chen Y, Vaillant F, Capaldo BD, Joyce R, Song X, Bryant VL, Penington JS, Di Stefano L, Tubau Ribera N et al. A single-cell RNA expression atlas of normal, preneoplastic and tumorigenic states in the human breast. (2021). EMBO J. 40. Marco Salas S, Kuemmerle LB, Mattsson-Langseth C, Tismeyer S, Avenel C, Hu T, Rehman H, Grillo M, Czarnewski P, Helgadottir S, et al. Optimizing Xenium In Situ data utility by quality assessment and best-practice analysis workflows. Nat Methods. 2025;22:813–23. 10.1038/s41592-025-02617-2 . Stringer C, Wang T, Michaelos M, Pachitariu M. Cellpose: a generalist algorithm for cellular segmentation. Nat Methods. 2021;18:100–6. 10.1038/s41592-020-01018-x . Petukhov V, Xu RJ, Soldatov RA, Cadinu P, Khodosevich K, Moffitt JR, Kharchenko PV. Cell segmentation in imaging-based spatial transcriptomics. Nat Biotechnol. 2022;40:345–54. 10.1038/s41587-021-01044-w . Satija R, Farrell JA, Gennert D, Schier AF, Regev A. Spatial reconstruction of single-cell gene expression data. Nat Biotechnol. 2015;33:495–502. 10.1038/nbt.3192 . Jin S, Guerrero-Juarez CF, Zhang L, Chang I, Ramos R, Kuan CH, Myung P, Plikus MV, Nie Q. Inference and analysis of cell-cell communication using CellChat. Nat Commun. 2021;12:1088. 10.1038/s41467-021-21246-9 . Blondel VD, Guillaume J-L, Lambiotte R, Lefebvre E. (2008). Fast unfolding of communities in large networks. Journal of statistical mechanics: theory and experiment 2008 , P10008. Hafemeister C, Satija R. Normalization and variance stabilization of single-cell RNA-seq data using regularized negative binomial regression. Genome Biol. 2019;20:296. 10.1186/s13059-019-1874-1 . Wu SZ, Al-Eryani G, Roden DL, Junankar S, Harvey K, Andersson A, Thennavan A, Wang C, Torpy JR, Bartonicek N, et al. A single-cell and spatially resolved atlas of human breast cancers. Nat Genet. 2021;53:1334–47. 10.1038/s41588-021-00911-1 . Comprehensive molecular portraits of human breast tumours. Nature. 2012;490:61–70. 10.1038/nature11412 . Goldman MJ, Craft B, Hastie M, Repečka K, McDade F, Kamath A, Banerjee A, Luo Y, Rogers D, Brooks AN et al. (2020). Visualizing and interpreting cancer genomics data via the Xena platform. In Nat Biotechnol, pp. 675–678. 10.1038/s41587-020-0546-8 Lachmann A, Xie Z, Ma'ayan A. blitzGSEA: efficient computation of gene set enrichment analysis through gamma distribution approximation. Bioinformatics. 2022;38:2356–7. 10.1093/bioinformatics/btac076 . Zhang L, Wang XI, Zhang S. Tumor-infiltrating lymphocyte volume is a better predictor of neoadjuvant therapy response and overall survival in triple-negative invasive breast cancer. Hum Pathol. 2018;80:47–54. 10.1016/j.humpath.2018.05.024 . van den Ende NS, Nguyen AH, Jager A, Kok M, Debets R, van Deurzen CHM. (2023). Triple-Negative Breast Cancer and Predictive Markers of Response to Neoadjuvant Chemotherapy: A Systematic Review. Int J Mol Sci 24 . 10.3390/ijms24032969 Passarelli MC, Pinzaru AM, Asgharian H, Liberti MV, Heissel S, Molina H, Goodarzi H, Tavazoie SF. Leucyl-tRNA synthetase is a tumour suppressor in breast cancer and regulates codon-dependent translation dynamics. Nat Cell Biol. 2022;24:307–15. 10.1038/s41556-022-00856-5 . Gollwitzer P, Grützmacher N, Wilhelm S, Kümmel D, Demetriades C. A Rag GTPase dimer code defines the regulation of mTORC1 by amino acids. Nat Cell Biol. 2022;24:1394–406. 10.1038/s41556-022-00976-y . Sung Y, Yoon I, Han JM, Kim S. Functional and pathologic association of aminoacyl-tRNA synthetases with cancer. Exp Mol Med. 2022;54:553–66. 10.1038/s12276-022-00765-5 . Sinclair LV, Rolf J, Emslie E, Shi YB, Taylor PM, Cantrell DA. Control of amino-acid transport by antigen receptors coordinates the metabolic reprogramming essential for T cell differentiation. Nat Immunol. 2013;14:500–8. 10.1038/ni.2556 . Feig C, Jones JO, Kraman M, Wells RJ, Deonarine A, Chan DS, Connell CM, Roberts EW, Zhao Q, Caballero OL, et al. Targeting CXCL12 from FAP-expressing carcinoma-associated fibroblasts synergizes with anti–PD-L1 immunotherapy in pancreatic cancer. Proc Natl Acad Sci U S A. 2023;110:20212–7. Griffiths JI, Cosgrove PA, Medina EF, Nath A, Chen J, Adler FR, Chang JT, Khan QJ, Bild AH. Cellular interactions within the immune microenvironment underpin resistance to cell cycle inhibition in breast cancers. Nat Commun. 2025;16:2132. Bassez A, Vos H, Van Dyck L, Floris G, Arijs I, Desmedt C, Boeckx B, Vanden Bempt M, Nevelsteen I, Lambein K, et al. A single-cell map of intratumoral changes during anti-PD-1 treatment of patients with breast cancer. ) Nat Med. 2021;27:820–32. Mi H, Varadhan R, Cimino-Mathews AM, Emens LA, Santa-Maria CA, Popel AS. (2023). Spatial and Compositional Biomarkers in Tumor Microenvironment Predicts Clinical Outcomes in Triple-Negative Breast Cancer. bioRxiv. 10.1101/2023.12.18.572234 El-Hachem N, Leclercq M, Susaeta Ruiz M, Vanleyssem R, Shostak K, Körner PR, Capron C, Martin-Morales L, Roncarati P, Lavergne A, et al. Valine aminoacyl-tRNA synthetase promotes therapy resistance in melanoma. Nat Cell Biol. 2024;26:1154–64. 10.1038/s41556-024-01439-2 . Wei Z, Liu X, Cheng C, Yu W, Yi P. Metabolism of Amino Acids in Cancer. Front Cell Dev Biol. 2020;8:603837. 10.3389/fcell.2020.603837 . Fernandes SA, Angelidaki DD, Nüchel J, Pan J, Gollwitzer P, Elkis Y, Artoni F, Wilhelm S, Kovacevic-Sarmiento M, Demetriades C. Spatial and functional separation of mTORC1 signalling in response to different amino acid sources. Nat Cell Biol. 2024;26:1918–33. 10.1038/s41556-024-01523-7 . Tables Table 1 Patient characteristics and clinical response n (%) pCR n pCR rate % Univariate P value* Age ≤50 25 (51.0) 12 48.0 >50 24 (49.0) 5 20.8 0.0434* BRCA1/2 status wild 16 (32.7) 4 25.0 pathogenic 5 (10.1) 4 80.0 0.0267* VUS 2 ( 4.1) 1 50.0 unknown a 26 (53.1) 8 30.8 Tumor size ≤2cm 15 (30.6) 8 53.3 >2cm 34 (69.4) 9 26.5 0.0721 Nodal status negative 32 (65.3) 14 43.8 positive 17 (34.7) 3 17.7 0.0593 TILs ≤20% 15 (30.6) 4 13.3 >20% 16 (32.7) 6 37.5 0.5179 unknown b 18 (36.7) 7 38.9 Ki-67 ≤20% 3 ( 6.1) 1 33.3 >20% 30 (61.2) 9 30.0 0.9054 unknown b 16 (32.7) 7 43.8 pCR pathological complete response, BRCA Breast Cancer Susceptibility Gene, VUS variant of unknown significance, TILs tumor infiltrating lymphocytes, * P ≤ 0.05 a BRCA1/2 status unknown includes those not covered by insurance and those for which patient consent was not obtained. b TILs and ki-67 unknown includes cases in which tissue was not available because the diagnosis was made by biopsy at another hospital. Additional Declarations Competing interest reported. Norikazu Masuda receives grants from Chugai Pharmaceutical Co.,Ltd., Eli Lilly Japan K.K. , AstraZeneca K.K., Pfizer Inc., Daiichi Sankyo Company, Ltd., MSD K.K., Eisai Co., Ltd., Gilead Sciences, Inc. and Ono Pharmaceutical Co., Ltd.; and honoraria from Chugai Pharmaceutical Co.,Ltd., Pfizer Inc., AstraZeneca K.K., Eli Lilly Japan K.K. , Daiichi Sankyo Company, Ltd., Eisai Co., Ltd., Gilead Sciences, Inc. and MSD K.K.; and serves as a representative of the Board of Directors of the Japan Breast Cancer Research Group (JBCRG) (unpaid, 2021–2025), has served as a member of the JBCRG Board of Directors (unpaid) since 2007, and is a member of the Board of Directors of the Japanese Breast Cancer Society (JBCS) (unpaid, 2021–2024), Japan Society of Clinical Oncology (JSCO) (unpaid, since 2023), Japan Association of Breast Cancer Screening (JABCS) (unpaid, since 2024), and Kyoto Breast Cancer Research Network (KBCRN) (unpaid, since 2025). Tomoki Ebata receives honoraria from AstraZeneca K.K. and MSD K.K. Yuko Takano receives grants from Eli Lilly Japan K.K. and honoraria from Chugai Pharmaceutical Co., Ltd., Daiichi Sankyo Company, Ltd., MSD K.K., Eli Lilly Japan K.K., and AstraZeneca K.K. Madoka Iwase receives grants from AstraZeneca K.K., Daiichi Sankyo Company, Ltd., Eli Lilly Japan K.K., MSD K.K., Ono Pharmaceutical Co., Ltd., Chugai Pharmaceutical Co., Ltd., Eisai Co., Ltd., Gilead Sciences, Inc., Novartis AG, and Pfizer Inc.; and honoraria from Chugai Pharmaceutical Co., Ltd., Eli Lilly Japan K.K., Pfizer Inc., Taiho pharmaceutical Co.Ltd., Nipro Corporation, Daiichi Sankyo, Co., Ltd., MSD K.K., Kyowa Kirin Co., Ltd., and Exact Sciences Co. The remaining authors declare no competing interests. Supplementary Files DocumentS1.docx FigureS1.jpg FigureS4.jpg FigureS3.jpg FigureS3.jpg FigureS2.jpg TableS4.xlsx TableS1.docx TableS5.xlsx TableS6.xlsx TableS7.xlsx TableS3.xlsx TableS8.xlsx TableS11.xlsx TableS12.xlsx TableS14.xlsx TableS10.xlsx TableS9.xlsx TableS13.xlsx TableS2.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 12 May, 2026 Reviewers agreed at journal 07 May, 2026 Reviewers agreed at journal 06 May, 2026 Reviewers agreed at journal 06 May, 2026 Reviewers agreed at journal 05 May, 2026 Reviewers invited by journal 04 May, 2026 Editor assigned by journal 01 May, 2026 Submission checks completed at journal 01 May, 2026 First submitted to journal 30 Apr, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9572543","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":636501904,"identity":"690783ce-5145-4a93-9022-83411e45b638","order_by":0,"name":"Misato Yamamoto","email":"","orcid":"","institution":"Nagoya University Graduate School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Misato","middleName":"","lastName":"Yamamoto","suffix":""},{"id":636501905,"identity":"e479d09a-66b4-4edb-b37b-e75a5b4f3069","order_by":1,"name":"Ravi Velaga","email":"","orcid":"","institution":"Shinshu University","correspondingAuthor":false,"prefix":"","firstName":"Ravi","middleName":"","lastName":"Velaga","suffix":""},{"id":636501906,"identity":"868b08bf-f05d-45c4-b206-a2785ce8e762","order_by":2,"name":"Yuko Takano","email":"","orcid":"","institution":"Nagoya University Graduate School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Yuko","middleName":"","lastName":"Takano","suffix":""},{"id":636501907,"identity":"6d6851a7-266f-410e-a708-229759ceeec5","order_by":3,"name":"Satoko Shimada","email":"","orcid":"","institution":"Nagoya University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Satoko","middleName":"","lastName":"Shimada","suffix":""},{"id":636501912,"identity":"3e89834c-5cbf-46a6-8789-b98f1e711418","order_by":4,"name":"Madoka Iwase","email":"","orcid":"","institution":"Nagoya University Graduate School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Madoka","middleName":"","lastName":"Iwase","suffix":""},{"id":636501913,"identity":"44469662-ec66-4071-a72d-06df4d09bc27","order_by":5,"name":"Takahiro Ichikawa","email":"","orcid":"","institution":"Nagoya University Graduate School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Takahiro","middleName":"","lastName":"Ichikawa","suffix":""},{"id":636501915,"identity":"e6f470d0-2a53-43d8-bbea-2d3968fa58ba","order_by":6,"name":"Dai Takeuchi","email":"","orcid":"","institution":"Nagoya University Graduate School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Dai","middleName":"","lastName":"Takeuchi","suffix":""},{"id":636501916,"identity":"b5ae6d01-ff25-475a-bb7d-f053ab4dabb7","order_by":7,"name":"Toyone Kikumori","email":"","orcid":"","institution":"Nagoya University Graduate School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Toyone","middleName":"","lastName":"Kikumori","suffix":""},{"id":636501917,"identity":"0d11f483-4a00-403f-91ac-3b6466e4a47d","order_by":8,"name":"Kazuhiro Furuhashi","email":"","orcid":"","institution":"Nagoya University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Kazuhiro","middleName":"","lastName":"Furuhashi","suffix":""},{"id":636501918,"identity":"000642bb-e626-40c7-9e0d-3aba6939aa2e","order_by":9,"name":"Shoichi Maruyama","email":"","orcid":"","institution":"Nagoya University Graduate School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Shoichi","middleName":"","lastName":"Maruyama","suffix":""},{"id":636501919,"identity":"ee1fc32e-a23a-4469-ac69-22070c179412","order_by":10,"name":"Atsushi Enomoto","email":"","orcid":"","institution":"Nagoya University Graduate School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Atsushi","middleName":"","lastName":"Enomoto","suffix":""},{"id":636501920,"identity":"68746c70-16f4-493e-b6f4-9a0322e7f939","order_by":11,"name":"Ken-ichi Ito","email":"","orcid":"","institution":"Shinshu University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Ken-ichi","middleName":"","lastName":"Ito","suffix":""},{"id":636501921,"identity":"5563bd9e-fb3a-4528-90c3-6cbb7590b9d3","order_by":12,"name":"Tomoki Ebata","email":"","orcid":"","institution":"Nagoya University Graduate School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Tomoki","middleName":"","lastName":"Ebata","suffix":""},{"id":636501922,"identity":"11e6bd6e-06ba-41ea-8e80-2eef08428a45","order_by":13,"name":"Norikazu Masuda","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7UlEQVRIiWNgGAWjYNACNgY5BnYgiQA8hLUYMzCTqiWxAVULHmAudvjxZ54yu/T+Zt5jD34w2MibM/AefMAgcwenFsvZaWbSPOeSc2cc5ks37GFIM9zZwJdswMDzDKcWg9sJZsy8bcy5DYd5zCR4GA4nGBwAMhh4DuPRkv75M29bfbo8UIvkH4gW8x/4teQYSPO2AVUCtUjDbGEgoKVMcs6544YbD/OlScsYpBluOMxjLJGA1y/pmz+8KauWlzvee0zyTYWNvMHxHsMPH3twhxgSAMWeARAzA3FizwFitcDBD6K0jIJRMApGwcgAAP8pTIZkR5iLAAAAAElFTkSuQmCC","orcid":"","institution":"Nagoya University Graduate School of Medicine","correspondingAuthor":true,"prefix":"","firstName":"Norikazu","middleName":"","lastName":"Masuda","suffix":""}],"badges":[],"createdAt":"2026-04-30 05:55:00","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9572543/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9572543/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":109250085,"identity":"61d4ade7-eb4e-4124-be13-5f6b214b864d","added_by":"auto","created_at":"2026-05-14 09:06:28","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":45755,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003enCounter transcriptomic profiling of differentially expressed genes with chemotherapy resistance in TNBC\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn the nCounter analysis, 83 differentially expressed genes were identified at a significance threshold of \u003cem\u003eP\u003c/em\u003e ≤ 0.05, with 52 downregulated and 31 upregulated in non-responders.\u003c/p\u003e","description":"","filename":"Figure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/f4bec4d3780d3d13d58fe203.jpg"},{"id":109249808,"identity":"a67e3f83-c00b-4f38-93ca-cc27d6547f71","added_by":"auto","created_at":"2026-05-14 09:04:16","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":95318,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSpatial transcriptomics using 10x Genomics Xenium highlights reduced immune composition and altered tumor architecture\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) UMAP embedding of all Y1−Y8 samples demonstrates that cells cluster primarily by cell type identity, indicating conserved transcriptional programs while preserving distinct immune, stromal, and epithelial compartments.\u003c/p\u003e\n\u003cp\u003e(B, C) Individual group level UMAPs with detailed cell types are shown. Responders exhibited a diverse mixture of immune cell subtypes including T cells, NK cells, macrophages, and dendritic cells. In contrast, non-responders showed reduced immune diversity and relatively uniform epithelial and stromal populations.\u003c/p\u003e\n\u003cp\u003e(D) Volcano plot displaying differential gene expression based on the Xenium 330 gene panel spatial transcriptomics. \u003cem\u003eCCR2\u003c/em\u003e and \u003cem\u003eWEE1\u003c/em\u003e showed significant differential expression in non-responders.\u003c/p\u003e","description":"","filename":"Figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/ffafb946df4e4389f6379b2e.jpg"},{"id":109249777,"identity":"4de3a978-b15b-42f6-8b28-3acdf0e6c67b","added_by":"auto","created_at":"2026-05-14 09:04:00","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":151195,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe cluster and cell-type gene expression patterns suggested a composite axis integrating immune activity, translational capacity, and proliferative stress between responders and non-responders.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A, B) Genes of interest with significant differential expression across clusters and cell types respectively with multiple testing correction. Immune-related genes were predominantly expressed by T cells in responders and the resistance and metabolic markers were predominantly expressed by breast glandular cells in non-responders.\u003c/p\u003e\n\u003cp\u003e(C) Spatial neighborhood analysis showing differential cell−cell proximity patterns between treatment groups. Responders exhibited increased spatial proximity between immune cells and tumor-associated populations, whereas non-responders displayed enhanced stromal-dominated neighborhoods.\u003c/p\u003e","description":"","filename":"Figure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/a90a5a7556bdd8042769e3b8.jpg"},{"id":109250100,"identity":"cadf7dc7-afc7-403a-be68-570804f00825","added_by":"auto","created_at":"2026-05-14 09:06:38","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":92951,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCell communication and Spatial Rewiring distinguish chemotherapy response patterns\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Circle plots following CellChat analysis in responders and non-responders reveal stromal-dominated interaction networks centered on fibroblasts, macrophages, and mast cells in non-responders.\u003c/p\u003e\n\u003cp\u003e(B) Circle plots demonstrating cell signalling interactions in responders show that \u0026nbsp;endothelial cells and macrophages serve as central communication hubs with active immune cell participation, including T cell−adipocyte crosstalk. In contrast, non-responders exhibited prominent adipocyte autocrine signalling and a breast cancer−B cell communication axis.\u003c/p\u003e\n\u003cp\u003e(C) Bubble plot showing ligand−receptors pairs in both responders and non-responders. Notably, CD45 (PTPRC−MRC1) signaling was detected exclusively in responders\u003cstrong\u003e.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Figure4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/f724fc5264d9e1fd9df7e356.jpg"},{"id":109252490,"identity":"b0b2fa05-5b22-49fc-b0f3-c1f083ef022e","added_by":"auto","created_at":"2026-05-14 09:27:13","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":83810,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eInter-patient heterogeneity in spatial architecture and cell–cell communication networks\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Spatial plots demonstrating distribution of major cell types across individual TNBC samples (Y1−Y8). Individual spatial plots illustrate inter-patient heterogeneity in tissue organization. Responders exhibit heterogeneous spatial architectures with immune-enriched regions, whereas non-responders display more uniform tumor- and stroma-dominated structures, including adipocyte clustering in selected cases.\u003c/p\u003e\n\u003cp\u003e(B) CellChat-derived circle plots display intercellular communication networks for individual (Y1−Y8) samples. Individual responder tumors show variable but generally denser immune-centred communication patterns, with T cells acting as major signalling hubs in some cases (e.g., Y3). In contrast, non-responders demonstrate more uniform, stromal-dominated communication networks, including consistent adipocyte autocrine signalling, particularly prominent in Y7 and Y8.\u003c/p\u003e","description":"","filename":"Figure5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/1bb76cb585f5dc06a0e54bf0.jpg"},{"id":109250341,"identity":"3e87a9a3-dec4-4ea3-ab61-f6c4d1660bf9","added_by":"auto","created_at":"2026-05-14 09:07:58","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":124664,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSpatial transcriptomic profiling reveals immune-active and resistance-associated transcriptional programs\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Integrated UMAP embedding of Xenium spatial transcriptomics data showing differential expression of immune activation and resistance-associated markers between responders and non-responders. Responders exhibited high expression of T cell markers (\u003cem\u003eCD3E\u003c/em\u003e, \u003cem\u003eCD8A\u003c/em\u003e, \u003cem\u003eCD247\u003c/em\u003e, \u003cem\u003eCD3G\u003c/em\u003e) and the cytotoxic effector gene \u003cem\u003eGZMA\u003c/em\u003e, whereas non-responders showed elevated expression of the cell cycle checkpoint kinase \u003cem\u003eWEE1\u003c/em\u003e and the metabolic enzyme \u003cem\u003eLARS\u003c/em\u003e.\u003c/p\u003e\n\u003cp\u003e(B) Representative spatial maps of marker gene expression across tissue sections, confirming immune-enriched tumor microenvironments in responders and resistance marker-dominant regions with limited immune infiltration in non-responders.\u003c/p\u003e\n\u003cp\u003e(C) Violin plots showing expression distributions of key immune and resistance marker genes of interest between the responder and non-responder groups. Responders displayed significantly higher numbers of cells expressing \u003cem\u003eCD8A\u003c/em\u003e+, \u003cem\u003eCD3E\u003c/em\u003e+, \u003cem\u003eGZMA\u003c/em\u003e+, and \u003cem\u003eCCR2\u003c/em\u003e+, whereas non-responders showed expanded populations of \u003cem\u003eWEE1\u003c/em\u003e+ and \u003cem\u003eLARS\u003c/em\u003e+ cells (Tables S12−S13).\u003c/p\u003e","description":"","filename":"Figure6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/4bacb31fdf737f88ac19e793.jpg"},{"id":109249807,"identity":"0145b9ee-514f-49c7-b0e1-94e4f32be69f","added_by":"auto","created_at":"2026-05-14 09:04:15","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":76492,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCross-cohort analysis of \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eCD8A\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e and \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eLARS\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003eexpression using spatial TNBC atlas and TCGA basal breast cancer cohort as supporting findings\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Pearson correlations between CID4465 and CID44971 samples from Wu et al., 2021 (Nature Genetics) and the eight Xenium samples (Y1−Y8) using all 330 genes in the panel. While CID4465 showed strong similarity to the chemotherapy responder group, CID44971 exhibited a mixed characteristics pattern with moderate correlations across both responders and non-responders.\u003c/p\u003e\n\u003cp\u003e(B) Direct comparison of \u003cem\u003eLARS\u003c/em\u003e and \u003cem\u003eCD8A\u003c/em\u003eexpression showed that CID44971 aligned more closely with Y7, a non-responder tumor sample in our study.\u003c/p\u003e\n\u003cp\u003e(C) Pathway enrichment analysis showed significant enrichment in \u003cem\u003eLARS\u003c/em\u003e-correlated ECM remodelling, cell cycle, mTOR signalling, amino acid metabolism, and immune response, with statistical significance (FDR \u0026lt; 0.05).\u003c/p\u003e\n\u003cp\u003e(D) Principal component analysis based on the 9 genes of interest (\u003cem\u003eCD8A\u003c/em\u003e, \u003cem\u003eLARS\u003c/em\u003e, \u003cem\u003eCCNA1\u003c/em\u003e, \u003cem\u003eCCNA2\u003c/em\u003e, \u003cem\u003eCCNB1\u003c/em\u003e, \u003cem\u003eCCND1\u003c/em\u003e, \u003cem\u003eCDC14A\u003c/em\u003e, \u003cem\u003eCDC20\u003c/em\u003e, and \u003cem\u003eE2F3\u003c/em\u003e) alone captured 99.9% of the total variance, effectively reflecting responder−non-responder separation better than the 95% observed in the 330-gene panel.\u003c/p\u003e\n\u003cp\u003e(E) Survival analysis in the TCGA basal subtype showed that high \u003cem\u003eLARS\u003c/em\u003e expression (median = 11.4) tended to predict shorter disease-free interval (\u003cem\u003eP\u003c/em\u003e = 0.0958) (left). BlitzGSEA revealed significant enrichment of G2M checkpoint and mTORC1 signaling in \u003cem\u003eLARS\u003c/em\u003e-high tumors (middle and right).\u003c/p\u003e","description":"","filename":"Figure7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/565ff14d8af8a3f990dafc25.jpg"},{"id":109297289,"identity":"1d6092ec-f5c8-4c84-a0d4-ff0fd709a678","added_by":"auto","created_at":"2026-05-15 08:55:28","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1120269,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/6fa752d9-e781-4245-9f8b-4d993336663f.pdf"},{"id":109249815,"identity":"274d8c8e-9a5a-4a69-baf6-8ad9ab3f9ec3","added_by":"auto","created_at":"2026-05-14 09:04:21","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":20745,"visible":true,"origin":"","legend":"","description":"","filename":"DocumentS1.docx","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/b573ec80bbd9c72cce9d34ec.docx"},{"id":109249824,"identity":"170ad675-5c2b-4cfd-bda0-d8ad67bfa8d4","added_by":"auto","created_at":"2026-05-14 09:04:29","extension":"jpg","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":66347,"visible":true,"origin":"","legend":"","description":"","filename":"FigureS1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/4eab34f47d42205f5b8cf30f.jpg"},{"id":109250097,"identity":"63ddb43d-e787-4a09-a177-5475fa518004","added_by":"auto","created_at":"2026-05-14 09:06:37","extension":"jpg","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":130312,"visible":true,"origin":"","legend":"","description":"","filename":"FigureS4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/7edd38317f798ede4a599695.jpg"},{"id":109250072,"identity":"e47a6c21-a4ad-4772-bb22-ef8d8a4eb0f2","added_by":"auto","created_at":"2026-05-14 09:06:17","extension":"jpg","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":86546,"visible":true,"origin":"","legend":"","description":"","filename":"FigureS3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/b1fb2a8692acb43899b313a0.jpg"},{"id":109249819,"identity":"fa72a8de-03e3-4777-88f8-b6f2cc67b435","added_by":"auto","created_at":"2026-05-14 09:04:23","extension":"jpg","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":86546,"visible":true,"origin":"","legend":"","description":"","filename":"FigureS3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/0b104966a7d8161c3e63cdc7.jpg"},{"id":109250342,"identity":"3c88e6ac-838e-4298-ab89-978d69800f73","added_by":"auto","created_at":"2026-05-14 09:07:58","extension":"jpg","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":101021,"visible":true,"origin":"","legend":"","description":"","filename":"FigureS2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/a0289b1979c584a20ea439e7.jpg"},{"id":109249782,"identity":"2c24ab1f-facc-40ee-97b6-3c89eaa204fc","added_by":"auto","created_at":"2026-05-14 09:04:02","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":16405,"visible":true,"origin":"","legend":"","description":"","filename":"TableS4.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/0615ed9d2ed94c98117f94c7.xlsx"},{"id":109249820,"identity":"56ce3ea9-a78c-48b9-84ad-8f11effd43c0","added_by":"auto","created_at":"2026-05-14 09:04:23","extension":"docx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":36211,"visible":true,"origin":"","legend":"","description":"","filename":"TableS1.docx","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/9df36a00adecd5b049b87614.docx"},{"id":109250356,"identity":"c6cb67e9-79b4-4993-88d9-2b0b7378d6b5","added_by":"auto","created_at":"2026-05-14 09:08:10","extension":"xlsx","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":40866,"visible":true,"origin":"","legend":"","description":"","filename":"TableS5.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/31d7fa097a8ff44daef362e0.xlsx"},{"id":109249821,"identity":"30350278-9b16-4a8f-abb1-b47e6ec9c5b2","added_by":"auto","created_at":"2026-05-14 09:04:24","extension":"xlsx","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":103131,"visible":true,"origin":"","legend":"","description":"","filename":"TableS6.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/1a6a2d59ef4695cf5e407eb8.xlsx"},{"id":109249817,"identity":"f5b92c52-e255-46af-98a2-3dd54e59c7c2","added_by":"auto","created_at":"2026-05-14 09:04:23","extension":"xlsx","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":111814,"visible":true,"origin":"","legend":"","description":"","filename":"TableS7.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/c1f3e32cb83df319595f6041.xlsx"},{"id":109250351,"identity":"b93f26d3-4e33-4114-954c-52e0bdb0141c","added_by":"auto","created_at":"2026-05-14 09:08:09","extension":"xlsx","order_by":10,"title":"","display":"","copyAsset":false,"role":"supplement","size":146535,"visible":true,"origin":"","legend":"","description":"","filename":"TableS3.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/77c6c4b23d37be5bb469415b.xlsx"},{"id":109249818,"identity":"2cf611c5-9ec6-4d86-a23c-5b3f621e899f","added_by":"auto","created_at":"2026-05-14 09:04:23","extension":"xlsx","order_by":11,"title":"","display":"","copyAsset":false,"role":"supplement","size":31572,"visible":true,"origin":"","legend":"","description":"","filename":"TableS8.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/51f1c3425f62aff3986688eb.xlsx"},{"id":109250366,"identity":"1f298beb-ba7d-465c-b241-56c4156db853","added_by":"auto","created_at":"2026-05-14 09:08:16","extension":"xlsx","order_by":12,"title":"","display":"","copyAsset":false,"role":"supplement","size":13122,"visible":true,"origin":"","legend":"","description":"","filename":"TableS11.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/4810eb91549c6dd91dad1384.xlsx"},{"id":109296004,"identity":"f5aca1c4-d7a6-4914-b144-110f6f5db793","added_by":"auto","created_at":"2026-05-15 08:43:46","extension":"xlsx","order_by":13,"title":"","display":"","copyAsset":false,"role":"supplement","size":23307,"visible":true,"origin":"","legend":"","description":"","filename":"TableS12.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/8217738a6742c785d6ec77d1.xlsx"},{"id":109250431,"identity":"2590e1e5-7654-4961-8f0b-937cb9d1cf96","added_by":"auto","created_at":"2026-05-14 09:08:30","extension":"xlsx","order_by":14,"title":"","display":"","copyAsset":false,"role":"supplement","size":22694,"visible":true,"origin":"","legend":"","description":"","filename":"TableS14.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/162fc71dcecf453508683272.xlsx"},{"id":109249809,"identity":"2227b650-1628-4fcf-9ce3-da97a8d05245","added_by":"auto","created_at":"2026-05-14 09:04:16","extension":"xlsx","order_by":15,"title":"","display":"","copyAsset":false,"role":"supplement","size":34401,"visible":true,"origin":"","legend":"","description":"","filename":"TableS10.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/21d56d6e0c7db2ff8f2aa276.xlsx"},{"id":109249825,"identity":"dcb74d34-ead0-415f-bc10-a10d4a8736f5","added_by":"auto","created_at":"2026-05-14 09:04:29","extension":"xlsx","order_by":16,"title":"","display":"","copyAsset":false,"role":"supplement","size":30327,"visible":true,"origin":"","legend":"","description":"","filename":"TableS9.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/d0d7406b58815274aa34de60.xlsx"},{"id":109249823,"identity":"f5ba917c-1847-4a13-9f34-f80c91b7a671","added_by":"auto","created_at":"2026-05-14 09:04:29","extension":"xlsx","order_by":17,"title":"","display":"","copyAsset":false,"role":"supplement","size":12321,"visible":true,"origin":"","legend":"","description":"","filename":"TableS13.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/1d43dbbd640f659818ef4b11.xlsx"},{"id":109249781,"identity":"fffb8c7f-897b-4c76-941a-a08ae85c22e9","added_by":"auto","created_at":"2026-05-14 09:04:02","extension":"docx","order_by":18,"title":"","display":"","copyAsset":false,"role":"supplement","size":20240,"visible":true,"origin":"","legend":"","description":"","filename":"TableS2.docx","url":"https://assets-eu.researchsquare.com/files/rs-9572543/v1/70a959e1ed440ea94f08ab2b.docx"}],"financialInterests":"Competing interest reported. Norikazu Masuda receives grants from Chugai Pharmaceutical Co.,Ltd., Eli Lilly Japan K.K. , AstraZeneca K.K., Pfizer Inc., Daiichi Sankyo Company, Ltd., MSD K.K., Eisai Co., Ltd., Gilead Sciences, Inc. and Ono Pharmaceutical Co., Ltd.; and honoraria from Chugai Pharmaceutical Co.,Ltd., Pfizer Inc., AstraZeneca K.K., Eli Lilly Japan K.K. , Daiichi Sankyo Company, Ltd., Eisai Co., Ltd., Gilead Sciences, Inc. and MSD K.K.; and serves as a representative of the Board of Directors of the Japan Breast Cancer Research Group (JBCRG) (unpaid, 2021–2025), has served as a member of the JBCRG Board of Directors (unpaid) since 2007, and is a member of the Board of Directors of the Japanese Breast Cancer Society (JBCS) (unpaid, 2021–2024), Japan Society of Clinical Oncology (JSCO) (unpaid, since 2023), Japan Association of Breast Cancer Screening (JABCS) (unpaid, since 2024), and Kyoto Breast Cancer Research Network (KBCRN) (unpaid, since 2025). \nTomoki Ebata receives honoraria from AstraZeneca K.K. and MSD K.K.\nYuko Takano receives grants from Eli Lilly Japan K.K. and honoraria from Chugai Pharmaceutical Co., Ltd., Daiichi Sankyo Company, Ltd., MSD K.K., Eli Lilly Japan K.K., and AstraZeneca K.K.\nMadoka Iwase receives grants from AstraZeneca K.K., Daiichi Sankyo Company, Ltd., Eli Lilly Japan K.K., MSD K.K., Ono Pharmaceutical Co., Ltd., Chugai Pharmaceutical Co., Ltd., Eisai Co., Ltd., Gilead Sciences, Inc., Novartis AG, and Pfizer Inc.; and honoraria from Chugai Pharmaceutical Co., Ltd., Eli Lilly Japan K.K., Pfizer Inc., Taiho pharmaceutical Co.Ltd., Nipro Corporation, Daiichi Sankyo, Co., Ltd., MSD K.K., Kyowa Kirin Co., Ltd., and Exact Sciences Co.\nThe remaining authors declare no competing interests.","formattedTitle":"Targeted spatial profiling identifies a CD8A − LARS axis associated with neoadjuvant chemotherapy resistance in triple-negative breast cancer","fulltext":[{"header":"Background","content":"\u003cp\u003eBreast cancer is the most frequently diagnosed cancer among women worldwide. Triple-negative breast cancer (TNBC) accounts for 10\u0026ndash;20% of breast cancer cases and is associated with a particularly poor prognosis owing to its aggressive clinical behaviour and lack of targeted therapies\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. For early-stage TNBC, anthracycline- and taxane-based chemotherapy has long been the standard of care. \u003csup\u003e3\u003c/sup\u003e Neoadjuvant chemotherapy (NAC) has become standard in the management of stage II/III TNBC, supported by the finding that a pathological complete response (pCR) after NAC is associated with improved long-term outcomes\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. NAC regimens including the addition of carboplatin have been devised to improve pCR rates\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e,\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. In patients who fail to achieve a pCR, adjuvant capecitabine is a widely accepted additional treatment, as demonstrated in the CREATE-X trial\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eMost recently, the KEYNOTE-522 (KN-522) trial showed that the addition of pembrolizumab to NAC significantly increased the pCR rate compared to chemotherapy alone (65% vs. 51%), with a 5-year overall survival rate of 86.6% in the pembrolizumab arm versus 81.7% in the placebo arm, leading to its adoption as a new standard of care for stage II/III TNBC\u003csup\u003e\u003cspan additionalcitationids=\"CR8\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. Although the KN-522 trial demonstrated improved prognosis, long-term adverse events, especially immune-related adverse events, remain among approximately one-quarter of patients. Hypothyroidism and adrenal insufficiency occurred in approximately 13.7% and 2.3% of the patients, respectively, and some required long-term hormone replacement therapy\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eDue to tumor heterogeneity and varying disease biology, it is difficult to confirm their therapeutic potential and use in a clinical setting. TNBC exhibits chemoresistance driven by the convergence of multiple signaling pathways (PI3K/AKT/mTOR, JAK/STAT, NF-κB), developmental programs (Notch, Wnt/β-catenin, Hedgehog, TGF-β), hypoxia-induced responses, and stemness-associated mechanisms, along with pronounced epigenetic and intratumoral heterogeneity\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. During NAC, some dominant tumor clones are eliminated in certain patients but persist in others, contributing to treatment resistance\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. Spatial profiling studies have revealed distinct immune \u0026ldquo;hot\u0026rdquo; and \u0026ldquo;cold\u0026rdquo; regions within TNBC tumors. Although TNBC exhibits the highest immune activation among breast cancer subtypes, only a subset of patients derive clinical benefit from immunotherapy.\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e Consequently, accurately characterizing tumor immune activation and evasion remains a major challenge to predicting response to NAC.\u003c/p\u003e \u003cp\u003eBulk transcriptomic profiling using RNA sequencing or multiplexed technologies such as the nCounter platform has provided insights into TNBC subtypes and potential mechanisms of chemoresistance\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. However, these approaches lack spatial resolution and fail to capture the gene expression context in the tumor microenvironment (TME). Spatial transcriptomic technologies such as Xenium have enabled high-resolution mapping of gene expression within intact tissue sections, allowing for deeper interrogation of cell\u0026ndash;cell interactions and spatially regulated resistance pathways\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e and elucidation of the TME.\u003c/p\u003e \u003cp\u003eIn this study, we performed transcriptomic analysis by combining gene expression profiling using nCounter and spatial transcriptomics using Xenium. We aimed to identify biomarkers and/or underlying mechanisms associated with chemoresistance in TNBC.\u003c/p\u003e "},{"header":"Methods","content":" \u003cp\u003e \u003cb\u003e1. Patient Selection and Data Inclusion\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThis was a single-centre retrospective study of TNBC patients undergoing NAC and surgery at Nagoya University Hospital (January 2017 to December 2023). Inclusion required estrogen receptor (ER)-negative (\u0026lt;\u0026thinsp;1% of tumor cells), progesterone receptor (PgR)-negative (\u0026lt;\u0026thinsp;1%), and human epidermal growth factor receptor type 2 (HER2)-negative (IHC 0, 1+, or 2\u0026thinsp;+\u0026thinsp;with negative in situ hybridization) status confirmed on pretreatment biopsy. Tumor-infiltrating lymphocytes (TILs) were assessed on H\u0026amp;E sections by two independent pathologists according to TILs Working Group (2014) guidelines. \u003cem\u003eBRCA1/2\u003c/em\u003e testing was performed using the BRCA analysis Test (Myriad Genetics) under the Japanese National Health Insurance system.\u003c/p\u003e \u003cp\u003eNAC comprised epirubicin (90 mg/m\u0026sup2;) and cyclophosphamide (600 mg/m\u0026sup2;) every 21 or 14 days for four cycles, followed by docetaxel (75 mg/m\u0026sup2;, every 21 days for 4 cycles) or paclitaxel (80 mg/m\u0026sup2; weekly for 12 cycles). pCR (ypT0/is ypN0) defined as responders; residual disease defined as non-responders.\u003c/p\u003e \u003cp\u003eUnivariate and multivariate analyses were conducted to explore the predictors of pCR among clinical factors.\u003c/p\u003e \u003cp\u003e \u003cb\u003e2. nCounter Gene Expression Profiling\u003c/b\u003e \u003c/p\u003e \u003cp\u003ePretreatment formalin-fixed paraffin-embedded (FFPE) tumor sections from eight TNBC patients were analyzed, including four responders (achieving pCR after NAC; Y1\u0026ndash;4) and four non-responders (with residual disease after NAC; Y5\u0026ndash;8). Total RNA was extracted using the QIAGEN RNeasy FFPE Kit and assessed by spectrophotometry (260/280 and 260/230 ratios) and Agilent TapeStation. Gene expression profiling was performed using the NanoString nCounter Human Tumor Signalling 360 panel (760 genes, 13 housekeeping genes, and 6 positive control probes).\u003c/p\u003e \u003cp\u003eQuality control (imaging quality, binding density, positive control linearity, housekeeping stability) and normalization (geometric mean of positive controls and stable housekeeping genes) were performed in ROSALIND\u0026reg; (OnRamp BioInformatics). Normalized counts were log₂-transformed. Differential expression analysis was performed using DESeq2-based methods within ROSALIND. Genes with \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026le;\u0026thinsp;0.05 and log₂ fold change\u0026thinsp;\u0026ge;\u0026thinsp;1.5 were considered significant and visualized using volcano plots and heatmaps.\u003c/p\u003e \u003cp\u003e \u003cb\u003e3. Xenium Spatial Transcriptomics\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTargeted spatial transcriptomics was performed on matched FFPE sections (n\u0026thinsp;=\u0026thinsp;8) using the 10x Genomics Xenium platform. A customized 330-gene Human Breast panel comprised the standard 280-gene panel supplemented with 50 differentially expressed genes identified from nCounter analysis. Gene annotations were based on Karaayvaz et al. (2018)\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e, Pal et al. (2021)\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e, and the Human Protein Atlas. Sections (5 \u0026micro;m) were processed according to the manufacturer\u0026rsquo;s protocol and imaged using Xenium Analyser Software v1.9.2.0.\u003c/p\u003e \u003cp\u003eData processing followed best practice guidelines\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. Transcripts were excluded if they had (i) quality value\u0026thinsp;\u0026lt;\u0026thinsp;20, (ii) duplicate coordinates, or (iii) anomalously high counts indicative of technical artifacts. Cell segmentation used Cellpose v2.0\u003csup\u003e18\u003c/sup\u003e for nuclear segmentation from DAPI images (fixed diameter 25 pixels), followed by Baysor\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e for transcript assignment using spatial coordinates and probabilistic modeling (scale 2.4 \u0026micro;m, SD 25%, minimum 30 molecules per cell, 4 clusters, prior segmentation confidence 0.8) (Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003eA).\u003c/p\u003e \u003cp\u003eCells were retained if they met the following criteria: \u0026ge; 12 transcripts, \u0026ge; 10 unique genes, \u0026le; 5% negative control transcript rate, and nuclear area 6\u0026thinsp;\u0026minus;\u0026thinsp;80 \u0026micro;m\u0026sup2;, yielding 1,223,609 cells (responders 603,081; non-responders 620,528) (Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003eB). Gene expression matrices were library-size normalized, and log₂-transformed. Clustering was performed using the Louvain algorithm in Seurat v5.2.1\u003csup\u003e20\u003c/sup\u003e. Cell type annotation integrated canonical marker expression, reference metadata (330_CellType.csv), and manual curation; 728 cells with missing spatial coordinates were excluded.\u003c/p\u003e \u003cp\u003eCompositional differences in gene-of-interest (GOI)-expressing cells were assessed using Fisher\u0026rsquo;s exact test with Benjamini-Hochberg (BH) FDR correction (*FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.05, **\u0026lt; 0.01, ***\u0026lt; 0.001; \u0026dagger;\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05 nominal) for cellular heterogeneity and clustering patterns. Merged Seurat objects were analyzed using Principal Component Analysis (PCA) (50 PCs) and Uniform Manifold Approximation and Projection (UMAP) (dims 1\u0026thinsp;\u0026minus;\u0026thinsp;30). Spatial maps were generated using Baysor-derived coordinates to preserve tissue architecture.\u003c/p\u003e \u003cp\u003e \u003cb\u003e4. Biomarker Selection and Spatial Differential Expression\u003c/b\u003e \u003c/p\u003e \u003cp\u003eResistance markers (\u003cem\u003eAURKA\u003c/em\u003e, \u003cem\u003eVEGFA\u003c/em\u003e, \u003cem\u003eKIT\u003c/em\u003e, \u003cem\u003eWEE1\u003c/em\u003e, \u003cem\u003eLARS\u003c/em\u003e, \u003cem\u003eWARS\u003c/em\u003e) and cytotoxic immune markers (\u003cem\u003eCD3E\u003c/em\u003e, \u003cem\u003eCD8A\u003c/em\u003e, \u003cem\u003eCD247\u003c/em\u003e, \u003cem\u003eCD3G\u003c/em\u003e, \u003cem\u003eGZMA\u003c/em\u003e) were selected based on nCounter/Xenium differential expression using the following criteria: (i) \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026le;\u0026thinsp;0.05, (ii) expression in \u0026ge;\u0026thinsp;2 clusters, (iii) presence in \u0026ge;\u0026thinsp;2 tissues per group, and (iv) biological relevance.\u003c/p\u003e \u003cp\u003ePseudobulk analysis was performed using edgeR v3.40.2, with counts aggregated at the sample level and normalized using TMM. Genes with \u0026ge;\u0026thinsp;10 counts in \u0026ge;\u0026thinsp;2 samples were retained. Differential expression was assessed using quasi-likelihood F-tests with BH-FDR correction.\u003c/p\u003e \u003cp\u003eWithin-cluster comparisons used Welch\u0026rsquo;s t-tests (minimum 3 cells per group). Log₂ fold change (log₂FC) was calculated as log₂((mean_R\u0026thinsp;+\u0026thinsp;0.01) / (mean_NR\u0026thinsp;+\u0026thinsp;0.01)), where positive values indicate higher expression in responders. Results were visualized as signed -log₁₀(\u003cem\u003eP\u003c/em\u003e) heatmaps (*\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05, **\u0026lt; 0.01, ***\u0026lt; 0.001; \u0026dagger;FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.05, \u0026dagger;\u0026dagger;\u0026lt; 0.01, \u0026dagger;\u0026dagger;\u0026dagger;\u0026lt; 0.001). Fifteen major cell populations (\u0026ge;\u0026thinsp;200 cells) were included for cell-type-level analysis. Distribution across cells demonstrate the full distribution of gene expression across \u0026gt;\u0026thinsp;1.2\u0026nbsp;million cells.\u003c/p\u003e \u003cp\u003e \u003cb\u003e5. Cell\u0026ndash;Cell Communication and Spatial Neighborhood Analysis\u003c/b\u003e \u003c/p\u003e \u003cp\u003eIntercellular communication was analyzed using CellChat v1.6.1\u003csup\u003e21\u003c/sup\u003e, CellChatDB.human and the 330-gene panel for broad interaction patterns, instead of any specific ligand\u0026minus;receptor (LR) pairings. Communication probabilities were computed using the triMean method under a law-of-mass-action model, with a minimum of 10 cells per cell type. Differential communication was visualized using circle and bubble plots (top 25 LR pairs). Spatially informed LR analysis was performed using Squidpy (sq.gr.ligrec; 100 permutations, BH-FDR; 154 testable pairs). Interactions with FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.05 were considered statistically significant. Pathway-level information flow was calculated as the sum of communication probabilities across all cell type pairs for each signaling pathway.\u003c/p\u003e \u003cp\u003eTo capture the physical proximity patterns between cell types in situ, spatial neighborhood graphs were constructed per sample using Delaunay triangulation (sq.gr.spatial_neighbors)\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. Cell type co-localization was quantified using neighborhood enrichment (sq.gr.nhood_enrichment, 100 permutations).\u003c/p\u003e \u003cp\u003e \u003cb\u003e6. Functional Module Analysis and Statistical Testing\u003c/b\u003e \u003c/p\u003e \u003cp\u003eCell type annotation was refined using Louvain clustering\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e at multiple resolutions (0.3, 0.5, 0.8) in Seurat v5.2.1 with SCTransform\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e normalization. Functional module scores were calculated using Seurat\u0026rsquo;s AddModuleScore() to capture immune (\u003cem\u003eCD8A\u003c/em\u003e), metabolic (\u003cem\u003eLARS\u003c/em\u003e), and cell cycle (\u003cem\u003eCCNA1\u003c/em\u003e, \u003cem\u003eCCNA2\u003c/em\u003e, \u003cem\u003eCCNB1\u003c/em\u003e, \u003cem\u003eCCND1\u003c/em\u003e, \u003cem\u003eCCNE1\u003c/em\u003e, \u003cem\u003eCDC14A\u003c/em\u003e, \u003cem\u003eCDC20\u003c/em\u003e, \u003cem\u003eE2F3\u003c/em\u003e) axes. Per-cell module scores were analyzed spatially and by treatment response.\u003c/p\u003e \u003cp\u003eAssociations between clinical variables and pCR were assessed using bias-adjusted logistic regression (JMP18; \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05 significant). Spatial comparisons used Wilcoxon rank-sum tests with BH-FDR correction. Correlations among module scores were evaluated using Spearman\u0026rsquo;s rank correlation.\u003c/p\u003e \u003cp\u003e \u003cb\u003e7. Cross-Validation Using External Cohorts\u003c/b\u003e \u003c/p\u003e \u003cp\u003ePretreatment TNBC Visium data (Wu et al., 2021; GSE176078; samples CID4465 and CID44971)\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e were log-normalized in Seurat v4.3.0 (scale factor 10,000). Pearson correlations between Visium samples and Xenium samples (mean expression across 330 genes) were used to assign external samples to responder-like or non-responder-like profiles.\u003c/p\u003e \u003cp\u003e \u003cem\u003eLARS\u003c/em\u003e-correlated genes in CID44971 were identified using Spearman correlation (FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.05) and functionally annotated. PCA (centred and scaled; 68% confidence ellipses) was performed using the full 330-gene panel and nine GOI (\u003cem\u003eCD8A\u003c/em\u003e, \u003cem\u003eLARS\u003c/em\u003e, \u003cem\u003eCCNA1\u003c/em\u003e, \u003cem\u003eCCNA2\u003c/em\u003e, \u003cem\u003eCCNB1\u003c/em\u003e, \u003cem\u003eCCND1\u003c/em\u003e, \u003cem\u003eCDC14A\u003c/em\u003e, \u003cem\u003eCDC20\u003c/em\u003e, \u003cem\u003eE2F3\u003c/em\u003e).\u003c/p\u003e \u003cp\u003e \u003cem\u003eLARS\u003c/em\u003e survival analysis was performed using the TCGA 2012 basal breast cancer cohort (n\u0026thinsp;=\u0026thinsp;98)\u003csup\u003e25\u003c/sup\u003e in Xena\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e, with median-based stratification. Gene set enrichment analysis was conducted using blitzGSEA\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e, with significance defined as \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026le;\u0026thinsp;0.05 and FDR\u0026thinsp;\u0026le;\u0026thinsp;0.05.\u003c/p\u003e \u003cp\u003eComputational Tools\u003c/p\u003e \u003cp\u003eData processing was performed using Python (quality control and segmentation) and R (statistical analysis and visualization), including Cellpose v3.1.1, Baysor v0.6.2, Seurat v5.2.1, CellChat v1.6.1, Squidpy v1.2.0 and edgeR v3.40.2.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eAssociation of clinicopathological characteristics and treatment response:\u003c/h2\u003e \u003cp\u003eTo determine the clinicopathological characteristics that serve as potential predictive markers, we evaluated 49 patients with stage I-III TNBC who underwent surgery after NAC. The median patient age was 50 years (range 26\u0026ndash;73 years). Clinical stage II disease represented 55.1% of the cases and 34.7% had clinically node-positive disease. TILs were evaluated in 31 patients (Table. S1). The pCR rate in all patients was 34.7%. Univariate analysis revealed that age and \u003cem\u003eBRCA\u003c/em\u003e status were significantly associated with pCR (age\u0026thinsp;\u0026gt;\u0026thinsp;50; \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.0434, with \u003cem\u003eBRCA1/2\u003c/em\u003e pathogenic variants; \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.0267, Table. 1). Other clinicopathological characteristics, including TILs levels, were not significantly associated with pCR. Age was not a significant variable in the multivariable analysis. Although \u003cem\u003eBRCA\u003c/em\u003e status was significantly associated with pCR in the univariate analysis, it was excluded from the multivariate model because of the small number of \u003cem\u003eBRCA\u003c/em\u003e-pathogenic cases (n\u0026thinsp;=\u0026thinsp;5), which could have resulted in unstable estimates or overfitting.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eTranscriptomic profiling using nCounter reveals immunosuppression and proliferative signatures associated with chemotherapy resistance in TNBC:\u003c/h3\u003e\n\u003cp\u003eTo identify transcriptional changes associated with chemoresistance, we compared pretreatment gene expression between TNBC responders (n\u0026thinsp;=\u0026thinsp;4) and non-responders (n\u0026thinsp;=\u0026thinsp;4) by nCounter profiling (Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). Differential expression analysis was performed to identify 83 differentially expressed genes (DEGs) (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026le;\u0026thinsp;0.05, Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, Table \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e), comprising 52 downregulated and 31 upregulated genes in non-responders. Notably, the genes downregulated in non-responders included \u003cem\u003eCD247, GZMA\u003c/em\u003e, \u003cem\u003eCD27\u003c/em\u003e, \u003cem\u003eCXCR6\u003c/em\u003e, \u003cem\u003eCCR2\u003c/em\u003e, and \u003cem\u003ePRF1\u003c/em\u003e markers associated with cytotoxic T-cell activation and immune signalling. Upregulated genes in non-responders included \u003cem\u003eLOX\u003c/em\u003e, \u003cem\u003eWEE1\u003c/em\u003e, \u003cem\u003eCDK6\u003c/em\u003e, \u003cem\u003eARID1A\u003c/em\u003e, and \u003cem\u003eRUNX1\u003c/em\u003e, indicating cell cycle regulation and stromal remodelling in chemoresistance.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eSpatial transcriptomics highlights reduced immune composition and altered tumor architecture:\u003c/h3\u003e\n\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eTranscriptional enrichment and Cell type composition:\u003c/h2\u003e \u003cp\u003eTo spatially contextualise the transcriptional differences observed in the nCounter analysis and for novel/additional findings, we profiled the same tumors using a 10x Genomics Xenium breast cancer targeted panel (330 genes consisting of 280 genes and 50 custom genes) (Table. S4). In a unified UMAP embedding of all Y1-Y8 samples. cells clustered primarily by cell type identity rather than treatment response group, indicating conserved transcriptional programs across responders and non-responders, preserving distinct immune, stromal, and epithelial compartments (Figs.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA-B). After detailed cell-type annotation, responders displayed a diverse mixture of immune subtypes, including T cells, NK cells, macrophages, dendritic cells, whereas non-responders showed reduced immune diversity and relatively homogeneous epithelial and stromal populations (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB-C), although statistically significant diversity was not observed. We acknowledge that the limited sample size (n\u0026thinsp;=\u0026thinsp;4 per group) and the targeted gene panel constrain statistical power for detecting differences in less abundant populations.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eMarker gene analysis to differentiate responders and non-responders:\u003c/h3\u003e\n\u003cp\u003eDifferential expression analysis identified \u003cem\u003eCCR2\u003c/em\u003e, a chemokine receptor critical for immune cell recruitment, as the most significantly downregulated gene in non-responders (log2FC = -1.79, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.008; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD, Table. S5). This reduction was consistent across multiple cell populations, including T cells (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.007), breast cancer cells (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.004), and fibroblasts (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.014), suggesting a broad reduction in \u003cem\u003eCCR2\u003c/em\u003e-mediated signalling in the non-responding tumor microenvironment. Conversely, \u003cem\u003eWEE1\u003c/em\u003e, a cell cycle checkpoint kinase and emerging therapeutic target, was significantly upregulated in non-responders (log2FC\u0026thinsp;=\u0026thinsp;1.15, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.040), particularly in breast cancer cells (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.017) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD, Table. S5). To resolve the cellular contribution of these and other candidate genes, we examined the spatial and cluster-level expression patterns of all panel genes across responders and non-responders. To validate genes identified through cluster-based discovery, we performed systematic differential expression analysis across clusters and cell types with multiple testing correction (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA-B). Marker genes were selected based on consistent differential expression across multiple clusters, within and between biological replicates (Fig. \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003eC, Tables. S6-7). Cell type composition analysis revealed distinct cellular sources for immune, metabolic and cell cycle markers (Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003eA, Table. S8). Immune-related genes (\u003cem\u003eCD3G\u003c/em\u003e, \u003cem\u003eGZMA\u003c/em\u003e, \u003cem\u003eCD3E\u003c/em\u003e, \u003cem\u003eCD247\u003c/em\u003e, and \u003cem\u003eCD8A\u003c/em\u003e) were predominantly expressed by T cells, whereas \u003cem\u003eCCR2\u003c/em\u003e showed a more distributed pattern in T cells, followed by B cells (notably in responders) and macrophages (Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003eA). \u003cem\u003eCXCL12\u003c/em\u003e displayed the most heterogeneous profile, expressed largely by breast glandular cells, followed by fibroblasts, macrophages, and T cells. Resistance and metabolism associated markers \u003cem\u003eWEE1\u003c/em\u003e, \u003cem\u003eVEGFA\u003c/em\u003e, \u003cem\u003eKIT\u003c/em\u003e, \u003cem\u003eAURKA\u003c/em\u003e, \u003cem\u003eLARS\u003c/em\u003e and \u003cem\u003eWARS\u003c/em\u003e were predominantly expressed by breast glandular cells (Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003eA). Comparison of cell type composition between responders and non-responders revealed systematic differences in the cellular sources of gene expression (Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003eB). T cells contributed a significantly higher proportion of expressing cells in responders across all 13 candidate genes (FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003eB) whereas, breast glandular cells contributed significantly higher in non-responders for resistance markers including \u003cem\u003eAURKA\u003c/em\u003e, \u003cem\u003eLARS\u003c/em\u003e, \u003cem\u003eVEGFA\u003c/em\u003e, and \u003cem\u003eWEE1\u003c/em\u003e (FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003eB).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eCluster- and cell-type resolved differential expression supported these patterns (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA, Table. S9-10). \u003cem\u003eCD8A\u003c/em\u003e was consistently elevated in responders across multiple clusters (FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.001), supporting enhanced cytotoxic immune activity in treatment-sensitive tumors, and was higher in responders within breast cancer cells, breast myoepithelial cells, adipocytes, endothelial cells, and macrophages (5/14 cell types; FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.001). The single exception was within T cells themselves, where \u003cem\u003eCD8A\u003c/em\u003e was higher in non-responders, likely reflecting differential T-cell functional states rather than abundance. At cluster level, \u003cem\u003eLARS\u003c/em\u003e showed an inverse pattern, elevated in non-responders in half of the clusters (6/12 clusters FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.001), versus 4/12 clusters enriched in responders (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). Cell-type analysis confirmed elevated \u003cem\u003eLARS\u003c/em\u003e in non-responders within breast glandular cells, fibroblasts, and T cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB, Table. S10). Cell cycle genes (\u003cem\u003eCCNA1\u003c/em\u003e, \u003cem\u003eCCNA2\u003c/em\u003e, \u003cem\u003eCCNB1\u003c/em\u003e, \u003cem\u003eCCND1\u003c/em\u003e, \u003cem\u003eCDC14A\u003c/em\u003e, \u003cem\u003eCDC20\u003c/em\u003e, \u003cem\u003eE2F3\u003c/em\u003e) were similarly elevated in non-responders at both cluster and cell-type levels (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB, Fig \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e). Together, the cluster and cell-type gene expression patterns suggest a composite axis integrating immune activity, translational capacity, and proliferative stress that distinguishes responders from non-responders.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eCell Communication and Spatial Rewiring distinguish chemotherapy response patterns:\u003c/h2\u003e \u003cp\u003eHaving characterized transcriptional differences at cluster and cell-type levels, we next examined how these were reflected in tissue organisation and intercellular signalling. Spatial neighbourhood analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC) and CellChat (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA) based ligand-receptor inference identified distinct microenvironmental compositions and intercellular communication architecture between groups. Both approaches identified NK cells and dendritic cells exclusively in responders, consistent with intact innate immune surveillance. In non-responders, stromal interactions dominated both spatial proximity and cell-cell interaction networks, with fibroblasts, macrophages, and mast cells forming the principal communication hubs.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eCellChat analysis revealed distinct communication architectures between groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). In responders, endothelial cells and macrophages served as central communication hubs with active immune cell participation, including T cell\u0026ndash;adipocyte crosstalk. In non-responders, adipocyte autocrine signalling and a breast cancer-B cell communication axis were prominent, while macrophage network engagement was diminished. Twenty significant ligand-receptor (L-R) pairs were idendified in responders (Table. S11) across six signalling pathways (\u003cem\u003ePECAM1\u003c/em\u003e, \u003cem\u003eCD45\u003c/em\u003e, \u003cem\u003eCXCL\u003c/em\u003e, \u003cem\u003eCDH\u003c/em\u003e, \u003cem\u003eVEGF\u003c/em\u003e), compared with sixteen pairs across five pathways in non-responders (\u003cem\u003ePECAM1\u003c/em\u003e, \u003cem\u003eVEGF\u003c/em\u003e, \u003cem\u003eCXCL\u003c/em\u003e, \u003cem\u003eCDH\u003c/em\u003e, \u003cem\u003eNCAM\u003c/em\u003e). All detected interactions achieved statistical significance (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001, FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e \u003cp\u003eThe strongest communication probability was observed for homophilic \u003cem\u003ePECAM1\u003c/em\u003e-\u003cem\u003ePECAM1\u003c/em\u003e interactions, with endothelial cell autocrine signalling dominant in both groups (responders: prob\u0026thinsp;=\u0026thinsp;0.51; non-responders: prob\u0026thinsp;=\u0026thinsp;0.523; Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC, Table. S11). However, \u003cem\u003ePECAM1\u003c/em\u003e pathway information flow was substantially higher in responders (3.56) compared to non-responders (2.48). Notably, the \u003cem\u003eCD45\u003c/em\u003e pathway (\u003cem\u003ePTPRC\u003c/em\u003e-\u003cem\u003eMRC1\u003c/em\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC) was detected exclusively in the responders (information flow\u0026thinsp;=\u0026thinsp;0.20), with NK cells, dendritic cells, T cells and macrophages signalling to macrophages (prob\u0026thinsp;=\u0026thinsp;0.044\u0026ndash;0.054; Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). Breast cancer cells in responders also exhibited \u003cem\u003eCXCL12\u003c/em\u003e-\u003cem\u003eCXCR4\u003c/em\u003e signalling to T cells (prob\u0026thinsp;=\u0026thinsp;0.035, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001; Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC, Table. S11), consistent with active chemokine-mediated immune recruitment.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eIntra- and inter-patient spatial Heterogeneity:\u003c/h3\u003e\n\u003cp\u003eIntegration of spatial cell-type distributions (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA) with CellChat-derived communication networks (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB) revealed distinct patterns of inter-patient heterogeneity. Among responders, Y3 displayed a highly compartmentalised spatial architecture coupled with a dense, multi-cellular communication network in which T cells served as a central signalling hub (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA-B). In contrast, Y1 showed abundant T- and B-cell infiltration but limited communication activity, indicating a discordance between immune presence and signalling engagement. Y2 and Y4 exhibited intermediate patterns, with structured spatial organisation and less extensive communication networks (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). Non-responders, by comparison, showed more uniform spatial and signalling profiles. Adipocyte autocrine signalling was observed across all four non-responders and was accompanied by statistically significant spatial clustering of adipocytes, particularly in Y7 and Y8 (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA-B). Fibroblast-mediated communication was likewise consistent across non-responders and corresponded to expanded fibroblast spatial domains in Y6\u0026thinsp;\u0026minus;\u0026thinsp;Y8 (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA-B). Together, these patterns suggest that non-response is associated with a reproducible stromal-communication network dominated by adipocyte and fibroblast signalling, whereas responder tumors encompass a wider range of immune-stromal architectures, from limited communication despite immune infiltration (Y1) to highly active immune-stromal crosstalk (Y3), indicating that treatment sensitivity in TNBC may arise through diverse microenvironmental configurations.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eSample-level transcriptional analysis revealed crucial insights into intra-tumor heterogeneity (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA, Fig. \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e). Responder tumors contained distinct transcriptional domains with high cytotoxic T cell marker expression and coordinated immune-cell clustering were, whereas non-responder tumors were dominated by resistance marker-rich regions with minimal immune infiltration and enhanced stromal compartmentalisation (Fig. \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e). This spatial segregation reflects the limitations of targeted expression analysis in underestimating the true nature of potential biomarkers in detecting and understanding intra-tumor and spatial heterogeneity. Despite this heterogeneity, group-level marker patterns were preserved.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eUMAP visualization showed robust expression of T-cell markers T cell markers (\u003cem\u003eCD3E\u003c/em\u003e, \u003cem\u003eCD8A\u003c/em\u003e, \u003cem\u003eCD247\u003c/em\u003e, \u003cem\u003eCD3G\u003c/em\u003e) and the cytotoxic effector molecule \u003cem\u003eGZMA\u003c/em\u003e in responders, with reduced expression in non-responders. Conversely, the cell-kinase \u003cem\u003eWEE1\u003c/em\u003e and the metabolic enzyme \u003cem\u003eLARS\u003c/em\u003e were elevated in non-responders (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA). Spatial mapping of gene expression confirmed these patterns at the tissue level (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB). Responders harboured significantly more \u003cem\u003eCD8A\u003c/em\u003e\u003csup\u003e+\u003c/sup\u003e (56,981 vs 33,049), \u003cem\u003eCD3E\u003c/em\u003e\u003csup\u003e+\u003c/sup\u003e (67,696 vs 35,551), and \u003cem\u003eGZMA\u003c/em\u003e\u003csup\u003e+\u003c/sup\u003e cells (45,777 vs 19,607) (Table. S12), with immune cells distributed throughout the tumor. Notably, \u003cem\u003eCCR2\u003c/em\u003e\u003csup\u003e+\u003c/sup\u003e cells, were 3.9-fold more abundant in responders (32,949 vs 8,403) (Table. S12). In contrast, non-responders showed expanded populations of \u003cem\u003eWEE1\u003c/em\u003e\u003csup\u003e+\u003c/sup\u003e cells (156,559 vs 78,759) and \u003cem\u003eLARS\u003c/em\u003e\u003csup\u003e+\u003c/sup\u003e cells (120,720 vs 97,518) (Table. S12), suggesting enhanced cell cycle checkpoint activity and altered metabolic programming. Per-cell expression of T-cell markers was comparable between groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC), indicating that the differences reflected changes in cell abundance rather than per-cell transcriptional outputs (Table. S13).\u003c/p\u003e \u003cp\u003e \u003cb\u003eCross-Cohort support of\u003c/b\u003e \u003cb\u003eCD8A\u003c/b\u003e \u003cb\u003eand\u003c/b\u003e \u003cb\u003eLARS\u003c/b\u003e \u003cb\u003eaxis using Spatial and TCGA TNBC cohorts\u003c/b\u003e:\u003c/p\u003e \u003cp\u003eTo seek independent support for the \u003cem\u003eCD8A\u003c/em\u003e and \u003cem\u003eLARS\u003c/em\u003e patterns observed in our Xenium TNBC cohort, we examined publicly available TNBC spatial transcriptomics data from Wu et al\u003csup\u003e24\u003c/sup\u003e. Of ten TNBC pretreatment samples profiled by using 10x Visium, two samples (CID4465 and CID44971) contained spatial gene expression data overlapping our Xenium panel.\u003c/p\u003e \u003cp\u003ePearson correlation across all 330 panel genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA) showed that CID4465 most closely resembled responders in our cohort, while CID44971 displayed a mixed pattern with moderate correlations to both groups. Direct comparison of \u003cem\u003eLARS\u003c/em\u003e and \u003cem\u003eCD8A\u003c/em\u003e expression further showed that CID44971 (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB) most closely resembled Y7, a non-responder characterised by elevated \u003cem\u003eLARS\u003c/em\u003e and reduced \u003cem\u003eCD8A\u003c/em\u003e. To examine the transcriptional context of \u003cem\u003eLARS\u003c/em\u003e expression in CID44971, we constructed a \u003cem\u003eLARS\u003c/em\u003e-correlation network and conducted a pathway enrichment analysis. Significantly enriched (FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.05) included ECM remodelling, cell cycle, mTOR signalling, amino acid metabolism, and immune response (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eC).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWe next asked whether a focused 9-gene set (\u003cem\u003eCD8A\u003c/em\u003e, \u003cem\u003eLARS\u003c/em\u003e, \u003cem\u003eCCNA1\u003c/em\u003e, \u003cem\u003eCCNA2\u003c/em\u003e, \u003cem\u003eCCNB1\u003c/em\u003e, \u003cem\u003eCCND1\u003c/em\u003e, \u003cem\u003eCDC14A\u003c/em\u003e, \u003cem\u003eCDC20\u003c/em\u003e and \u003cem\u003eE2F3\u003c/em\u003e) could capture the responder- and non-responder distinction captured by the full panel (330 genes). PCA on the full 330 gene panel separated responders and non-responders along PC1 (95.5% variance; Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eD, left). Strikingly, the 9-gene set preserved this separation (PC1: 99.9% variance; Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eD, right). While these findings support further evaluation of the 9-gene set as a candidate stratifier of pretreatment TNBC tumors, validation in larger, independent cohorts will be required.\u003c/p\u003e \u003cp\u003eSurvival predictability of \u003cem\u003eLARS\u003c/em\u003e expression with a median of 11.4 in TCGA basal subtype cohort showed a trending worse prognosis for disease free interval in \u003cem\u003eLARS\u003c/em\u003e high tumors (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.0958, and log rank test statistic\u0026thinsp;=\u0026thinsp;2.773; Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eE, left) also serves as supporting evidence.\u003c/p\u003e \u003cp\u003eBlitz gene set enrichment analysis (blitzGSEA) found significant positive enrichment of HALLMARK_E2F_TARGETS, HALLMARK_G2M_CHECKPOINT, and HALLMARK_MTORC1_SIGNALING (NES\u0026thinsp;=\u0026thinsp;8.123, 7.469 and 5.501; \u003cem\u003eP\u003c/em\u003e values\u0026thinsp;=\u0026thinsp;4.412e-16, 8.059e-14 and 3.766e-08; FDR\u0026thinsp;=\u0026thinsp;2.206e-14, 1.007e-12 and 2.337e-07) (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eE, middle and right, Table. S14) in LARS high tumors, recapitulating the cell-cycle and metabolic features observed in non-responders by Xenium. ImmuneSigDB analysis in the same cohort resulted in positive enrichment of two CD8 T-cell gene sets including, GSE10239_MEMORY_VS_DAY4.5_EFF_CD8_TCELL_DN (NES\u0026thinsp;=\u0026thinsp;7.224; \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;5.018e-13; FDR\u0026thinsp;=\u0026thinsp;4.075e-10) and GSE15750_DAY6_VS_DAY10_EFF_CD8_TCELL_UP (NES\u0026thinsp;=\u0026thinsp;7.082, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.411e-12; FDR\u0026thinsp;=\u0026thinsp;6.646e-10), suggesting altered effector/memory CD8 T-cell dynamics in \u003cem\u003eLARS\u003c/em\u003e-high tumors. Canonical pathway enrichments included, positively enriched REACTOME_SELENOAMINO_ACID_METABOLISM (NES\u0026thinsp;=\u0026thinsp;4.389; \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.136e-05; and FDR\u0026thinsp;=\u0026thinsp;0.003), REACTOME_GLUCOSE_METABOLISM, REACTOME_MITOCHONDRIAL_BIOGENESIS,REACTOME_MITOCHONDRIA_PROTEIN_IMPORT and REACTOME_CHEMOKINE_RECEPTORS_BIND_CHEMOKINES (NES\u0026thinsp;=\u0026thinsp;3.291, 3.257, 3.201, and 2.886; \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;9.9e-4, 0.001, 0.001 and 0.003) in the \u003cem\u003eLARS\u003c/em\u003e high tumors suggesting heterogeneous cell cycle, immune and metabolic reprogramming molecules in \u003cem\u003eLARS\u003c/em\u003e differentially expressed tumors, which supports our Xenium data findings.\u003c/p\u003e \u003cp\u003eTogether, these cross-cohort observations support the \u003cem\u003eCD8A\u003c/em\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026minus;\u003c/span\u003e\u003cem\u003eLARS\u003c/em\u003e axis as a candidate framework linking immune evasion as a predictor of treatment response in this small cohort. Immune evasion via low \u003cem\u003eCD8A\u003c/em\u003e infiltration and high \u003cem\u003eLARS\u003c/em\u003e activity revealed a potential alternative chemoresistance axis in TNBC. Validation in larger, prospectively collected cohorts will be required to assess its predictive utility.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eClinicopathological analysis revealed that TILs, a widely reported prognostic feature of TNBC\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e,\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e, were not significantly associated with pCR in our cohort, suggesting that TIL density alone may be insufficient to predict chemotherapy response. This observation prompted us to interrogate the molecular and spatial features of pretreatment tumors to better identify potential predictive biomarkers and mechanisms underlying NAC resistance in TNBC.\u003c/p\u003e \u003cp\u003eOur targeted transcriptomic and spatial analyses identified a previously unrecognised association between immune surveillance and metabolic reprogramming in chemoresistant TNBC. The inverse relationship between \u003cem\u003eCD8A\u003c/em\u003e\u0026ndash;\u003cem\u003eLARS\u003c/em\u003e expression, together with cell-cycle dysregulation and the \u003cem\u003eLARS\u003c/em\u003e-functional enriched pathways, provides hypothesis generating insights into the molecular and spatial determinants of differential treatment response.\u003c/p\u003e \u003cp\u003eThese findings present an apparent contextual contrast with the established role of leucyl-tRNA synthetase (\u003cem\u003eLARS\u003c/em\u003e) in TNBC biology. Passarelli et al. recently demonstrated that \u003cem\u003eLARS\u003c/em\u003e functions as a tumor suppressor in breast cancer, with genetic deletion enhanced tumor formation and \u003cem\u003eLARS\u003c/em\u003e repression reduced translation of growth-suppressive genes, including \u003cem\u003eEMP3\u003c/em\u003e and \u003cem\u003eGGT5\u003c/em\u003e during mammary epithelial to mesenchymal cell transformation and in human and mouse breast cancers\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e. \u003cem\u003eLARS\u003c/em\u003e is known to preferentially regulate \u003cem\u003eRagD\u003c/em\u003e, but not \u003cem\u003eRagC\u003c/em\u003e, thereby modulating mTORC1 recruitment and TFEB/TFE3 phosphorylation in lysosomes\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. In contrast, our spatial data indicate that \u003cem\u003eLARS\u003c/em\u003e is elevated in chemoresistant TNBC tumors, where significantly elevated \u003cem\u003eLARS\u003c/em\u003e expression accompanied by immune exclusion (\u003cem\u003eCD8A\u003c/em\u003e low expression), and elevated proliferative markers. Consistent with this, high \u003cem\u003eLARS\u003c/em\u003e expression tumors in the TCGA basal breast cancer cohort also demonstrated enriched activity of mTORC1, E2F and G2M checkpoint gene sets. These observations may reflect context dependent roles of \u003cem\u003eLARS\u003c/em\u003e. Differences in tissue context, spatial cellular composition and the model organism in which \u003cem\u003eLARS\u003c/em\u003e is regulated, likely contribute to this divergence. In addition, Passarelli et al. characterised the monoallelic deletion of \u003cem\u003eLARS\u003c/em\u003e and the functional consequences on downstream t-RNAs which were beyond the scope of our study but remain important for mechanistic follow-up.\u003c/p\u003e \u003cp\u003eThis context-dependent gene expression pattern may reflect the broader functional plasticity of aminoacyl-tRNA synthetases in TNBC biology. While \u003cem\u003eLARS\u003c/em\u003e canonically supports protein synthesis and has been shown to act as a tumor suppressor by prompting growth-inhibitory gene translation, metabolic stress conditions such as glucose starvation can trigger post-translational modifications like \u003cem\u003eLARS\u003c/em\u003e phosphorylation and functional reprogramming\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e via mTORC1 activation\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. Our data suggest that \u003cem\u003eLARS\u003c/em\u003e may switch from a tumor-suppressive role during initiation toward a resistance-promoting factor, where enhanced leucine sensing and mTORC1 activation may support survival and proliferation under therapeutic stress. Our observations are consistent with previous findings that activated CD8\u0026thinsp;+\u0026thinsp;T cells rely heavily on leucine uptake, which fuels mTORC1 activation and c-Myc expression to support clonal expansion and effector differentiation\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. Elevated \u003cem\u003eLARS\u003c/em\u003e activity in tumor cells may therefore create a competitive metabolic environment that constrains CD8⁺ T-cell expansion locally, contributing to the immune exclusion observed in our LARS-high, CD8A-low resistant tumors.\u003c/p\u003e \u003cp\u003eThe clinical efficacy of immune checkpoint inhibitors in TNBC, demonstrated by the KEYNOTE-522 trial, underscores the importance of the immune microenvironment in determining treatment response and provides a clinical rationale for resolving the spatial and cellular features that distinguish immunologically engaged from immunologically inactive tumors \u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e and reported its critical use in estimating immune infiltration for treatment efficacy. Our spatial analysis indicates that chemotherapy resistance in TNBC is characterised by altered spatial architecture, with resistant tumors showing reduced T cell-cancer cell proximity and enhanced fibroblast-mediated stromal compartments. The exclusive presence of \u003cem\u003eCD45\u003c/em\u003e pathway signalling (\u003cem\u003ePTPRC\u003c/em\u003e-\u003cem\u003eMRC1\u003c/em\u003e, information flow\u0026thinsp;=\u0026thinsp;0.20) in chemotherapy responders provides mechanistic insight into differential treatment outcomes. The potential \u003cem\u003ePTPRC\u003c/em\u003e-\u003cem\u003eMRC1\u003c/em\u003e interactions from NK cells, dendritic cells, T cells, and macrophages with macrophages points to a distinct macrophage-engaged communication architecture in treatment-sensitive tumors. The absence of this signalling axis in non-responders may reflect an immunologically \"cold\" tumor microenvironment characterised by limited immune cell infiltration, consistent with reduced chemotherapy efficacy.\u003c/p\u003e \u003cp\u003eThe cell-type annotation distinction between \"breast glandular cells\" and \u0026ldquo;breast cancer cells\u0026rdquo; in the Xenium breast cancer panel warrants methodological consideration. The more extensive communication observed in the glandular-annotated population may therefore reflect heterogeneity within the malignant compartment rather than non-malignant epithelial activity. Both cell populations engaged in biologically meaningful signalling. Breast cancer cells engaged through \u003cem\u003eCXCL12\u003c/em\u003e-\u003cem\u003eCXCR4\u003c/em\u003e chemokine signalling\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e ,and breast glandular cells through \u003cem\u003eVEGFA\u003c/em\u003e-\u003cem\u003eVEGFR2 (KDR)\u003c/em\u003e angiogenic signalling\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e, consistent with prior spatial transcriptomic analyses of breast cancer microenvironments describing heterogeneous tumor cell states and context-dependent stromal interactions\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. \u003cem\u003eCCR2\u003c/em\u003e and \u003cem\u003eCXCL12\u003c/em\u003e, identified through differential expression and ligand-receptor analyses respectively, showed concordant responder-enriched expression across clusters and cell types. The elevated \u003cem\u003eCXCL12\u003c/em\u003e in responders is consistent with the \u003cem\u003eCXCL12-CXCR4\u003c/em\u003e signalling identified by CellChat and supports a model of active chemokine-mediated T cell recruitment in responders. Future studies using expanded panels or whole-transcriptome spatial platforms will be required to resolve the cell-state distinctions, which are limited in a targeted panel study.\u003c/p\u003e \u003cp\u003eTogether, these findings are consistent with emerging evidence that TME architecture influences response to NAC in TNBC\u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThe inverse relationship between \u003cem\u003eCD8A\u003c/em\u003e and \u003cem\u003eLARS\u003c/em\u003e expression observed in our cohort suggests an exploratory metabolic-immune resistance axis candidate in TNBC. Recent studies have shown that aminoacyl-tRNA synthetases, particularly VARS, promote therapeutic resistance in melanoma through codon-biased translational reprogramming and fatty acid oxidation\u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e. Whether LARS plays a comparable role in TNBC, remains to be established. Our findings present LARS-associated metabolic stress, as a potential candidate in TNBC resistance, suggesting that \u003cem\u003eLARS\u003c/em\u003e-mediated metabolic reprogramming could antagonize cytotoxic T cell function, subject to further validation of T cell and metabolic cells states in a larger cohort. Amino acid metabolism, enriched in the \u003cem\u003eLARS\u003c/em\u003e-associated network of CID44791 sample examined here, emerged as a critical determinant of tumor immunity, with amino acid sensors including mTORC1 integrating the metabolic status with immune cell differentiation and disease function\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e,\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. The spatial segregation of \u003cem\u003eLARS\u003c/em\u003e-high and \u003cem\u003eCD8A\u003c/em\u003e-low regions in resistant tumors could result in metabolic reprogramming, creating immunosuppressive microenvironments, although the directionality of whether metabolic state shapes immune exclusion or vice-versa needs further examination.\u003c/p\u003e \u003cp\u003eThe coordinated upregulation of cell cycle genes (\u003cem\u003eCCNA1\u003c/em\u003e, \u003cem\u003eCCNB1\u003c/em\u003e, \u003cem\u003eE2F3\u003c/em\u003e and others) in non-responders, together with HALLMARK_E2F_TARGETS enrichment in \u003cem\u003eLARS\u003c/em\u003e-high TCGA basal tumors, indicates a proliferative axis associated with chemoresistance in our cohort. This pattern is notable because highly proliferative tumors have conventionally been associated with greater chemotherapy sensitivity; with elevated proliferation, metabolic (\u003cem\u003eLARS\u003c/em\u003e-high) and immune-excluded (\u003cem\u003eCD8A\u003c/em\u003e-low) features in non-responders. This suggests that proliferation alone is an insufficient predictor of response, and that the broader cellular context determines treatment outcome.\u003c/p\u003e \u003cp\u003eThe integration of metabolic reprogramming (\u003cem\u003eLARS\u003c/em\u003e), immune dysfunction (\u003cem\u003eCD8A\u003c/em\u003e loss), and elevated proliferation in non-responders suggests a multi-layered resistance network that may benefit from combination therapeutic approaches. Our study revealed that these mechanisms are not uniformly distributed, demonstrating intra-tumor and spatial heterogeneity, and are organised into distinct resistance niches that could be targeted with spatially directed therapies. Building on these observations, we propose a candidate \u003cem\u003eLARS\u003c/em\u003e-RagD-\u003cem\u003emTORC1\u003c/em\u003e-\u003cem\u003eCD8 T\u003c/em\u003e-cell axis as a working framework that could be distinct from the classical \u003cem\u003ePD-1/PD-L1\u003c/em\u003e directed therapeutic approaches.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eLimitations and future directions:\u003c/h2\u003e \u003cp\u003eOur study comprised relatively small cohort size (eight samples), different targeted gene candidates with some overlap between the nCounter and Xenium panels, and a focus on pretreatment samples alone. Cross-cohort validation against the Wu et al. spatial atlas was based on the two TNBC samples with available spatial data, with one sample (CID44971) contributing the principal cross-cohort similarity to non-responders; broader validation in independent spatial cohorts will be required. Future studies should examine the spatial evolution during treatment and validate the findings in larger cohorts and whole transcriptomes. Functional validation of the \u003cem\u003eCD8A\u003c/em\u003e\u0026ndash;\u003cem\u003eLARS\u003c/em\u003e resistance axis and its therapeutic targeting need to be conducted in future studies. Realising any of the LARS-associated metabolic adaptation, modulating fibroblast-rich stromal compartments that support immune evasion and developing biomarker guided combinational immunotherapy directions will require functional validation in TNBC models and prospective evaluation in larger, well-characterised cohorts. Although the present cohort received chemotherapy without immune checkpoint inhibitors (ICIs), evaluating CD8A and LARS expression patterns in patients receiving chemoimmunotherapy regimens, represents one of the important next steps, particularly to determine whether these features inform response to ICI combination therapies.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusions","content":"\u003cp\u003eOur spatial transcriptomics analysis of pretreatment TNBC tumors identifies a candidate metabolic-immune resistance axis defined by inverse spatial relationship between \u003cem\u003eLARS\u003c/em\u003e and \u003cem\u003eCD8A\u003c/em\u003e accompanied by elevated proliferative gene expression in non-responders. Although these findings are exploratory and require validation in larger, prospectively collected cohorts, they support the value of spatial transcriptomics in resolving the cellular and metabolic features of chemoresistance and provide a hypothesis-generating framework, the \u003cem\u003eCD8A-LARS\u003c/em\u003e axis, for future mechanistic and biomarker focused investigation in TNBC.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eBH\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eBenjamini-Hochberg\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDEGs\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eDifferentially expressed genes\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eER\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eEstrogen receptor\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eFFPE\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eFormalin-fixed paraffin-embedded\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGOI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eGene-of-interest\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eHER2\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eHuman epidermal growth factor receptor 2\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eICI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eImmune checkpoint inhibitor\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eKN-522\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eKEYNOTE-522 trial\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eLARS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eLeucyl-tRNA synthetase\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eLog\u003csub\u003e2\u003c/sub\u003eFC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eLog\u003csub\u003e2\u003c/sub\u003e fold change\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eLR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eLigand-receptor\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eNAC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eNeoadjuvant chemotherapy\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePCA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ePrincipal component analysis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003epCR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ePathological complete response\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePgR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eProgesterone receptor\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTILs\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eTumor-infiltrating lymphocytes\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTME\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eTumor microenvironment\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTNBC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eTriple-negative breast cancer\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eUMAP\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eUniform manifold approximation and projection.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e \u003cp\u003eThe study was approved by the Nagoya University Hospital IRB (2022\u0026thinsp;\u0026minus;\u0026thinsp;0244, 2023-0066), and all patients provided informed consent.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eConsent for publication\u003c/strong\u003e \u003cp\u003eNot applicable\u003c/p\u003e \u003c/p\u003e\u003cp\u003e\u003ch2\u003eCompeting Interests\u003c/h2\u003e\u003cp\u003eNorikazu Masuda receives grants from Chugai Pharmaceutical Co.,Ltd., Eli Lilly Japan K.K. , AstraZeneca K.K., Pfizer Inc., Daiichi Sankyo Company, Ltd., MSD K.K., Eisai Co., Ltd., Gilead Sciences, Inc. and Ono Pharmaceutical Co., Ltd.; and honoraria from Chugai Pharmaceutical Co.,Ltd., Pfizer Inc., AstraZeneca K.K., Eli Lilly Japan K.K. , Daiichi Sankyo Company, Ltd., Eisai Co., Ltd., Gilead Sciences, Inc. and MSD K.K.; and serves as a representative of the Board of Directors of the Japan Breast Cancer Research Group (JBCRG) (unpaid, 2021\u0026ndash;2025), has served as a member of the JBCRG Board of Directors (unpaid) since 2007, and is a member of the Board of Directors of the Japanese Breast Cancer Society (JBCS) (unpaid, 2021\u0026ndash;2024), Japan Society of Clinical Oncology (JSCO) (unpaid, since 2023), Japan Association of Breast Cancer Screening (JABCS) (unpaid, since 2024), and Kyoto Breast Cancer Research Network (KBCRN) (unpaid, since 2025). Tomoki Ebata receives honoraria from AstraZeneca K.K. and MSD K.K.Yuko Takano receives grants from Eli Lilly Japan K.K. and honoraria from Chugai Pharmaceutical Co., Ltd., Daiichi Sankyo Company, Ltd., MSD K.K., Eli Lilly Japan K.K., and AstraZeneca K.K.Madoka Iwase receives grants from AstraZeneca K.K., Daiichi Sankyo Company, Ltd., Eli Lilly Japan K.K., MSD K.K., Ono Pharmaceutical Co., Ltd., Chugai Pharmaceutical Co., Ltd., Eisai Co., Ltd., Gilead Sciences, Inc., Novartis AG, and Pfizer Inc.; and honoraria from Chugai Pharmaceutical Co., Ltd., Eli Lilly Japan K.K., Pfizer Inc., Taiho pharmaceutical Co.Ltd., Nipro Corporation, Daiichi Sankyo, Co., Ltd., MSD K.K., Kyowa Kirin Co., Ltd., and Exact Sciences Co.The remaining authors declare no competing interests.\u003c/p\u003e\u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eThe Moonshot Research and Development Program(grant no. JP22zf0127009)from the Japan Agency for Medical Research and Development (AMED).\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eConceptualization, M. Y., Y. T., R. V., and N. M.; Methodology, M. Y., R. V., Y. T., and N. M.; Investigation, M. Y., R. V., S. S., and N. M.; Resources, K. F., S. M., K. I., A. E., and N. M.; Formal analysis, M. Y., R. V., and N. M.; Writing \u0026ndash; Original Draft, M. Y., R. V., and N. M.; Writing \u0026ndash; Review \u0026amp; Editing, M. Y., R. V., Y. T., and N. M.; Final manuscript approval, all authors; Supervision, N. M.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eThis study was supported by The Moonshot Research and Development Program (Grant No. JP22zf0127009) of the Japan Agency for Medical Research and Development (AMED). The authors wish to acknowledge the Division for Medical Research Engineering, Nagoya University Graduate School of Medicine, for the use of Xenium. We thank Mr. Yamaguchi (Technical Center, Nagoya University) for technical support (Xenium sample processing) and data acquisition.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThis study did not report the original code. We adapted the segmentation and DAPI channel extraction steps from Cellpose's documentation and workflows18 and the Baysor GitHub repository, modified for Julia CLI integration with Xenium outputs, and incorporated STRtree-based transcript-to-cell assignment19. For UMAP embedding, Louvain clustering, cell type annotation, and marker gene identification, we adapted the code from Satija R et al., 201520. The CellChat scripts were adapted from the CellChat R package21. Our spatial neighborhood scripts (cell-cell distance, interaction matrices, and niche-DE style analysis) were adapted from Salas M et al., 202517. The unprocessed data used in this study are available upon reasonable request from the lead authors. The validation dataset is available from Wu SZet al24. (https://doi.org/10.5281/zenodo.4739739).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eHuang M, Haiderali A, Fox GE, Frederickson A, Cortes J, Fasching PA, O'Shaughnessy J. (2022). Economic and Humanistic Burden of Triple-Negative Breast Cancer: A Systematic Literature Review. Pharmacoeconomics \u003cem\u003e40\u003c/em\u003e, 519\u0026ndash;558. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s40273-021-01121-7\u003c/span\u003e\u003cspan address=\"10.1007/s40273-021-01121-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWon KA, Spruck C. Triple\u0026ndash;negative breast cancer therapy: Current and future perspectives (Review). Int J Oncol. 2020;57:1245\u0026ndash;61. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3892/ijo.2020.5135\u003c/span\u003e\u003cspan address=\"10.3892/ijo.2020.5135\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJie H, Ma W, Huang C. Diagnosis, Prognosis, and Treatment of Triple-Negative Breast Cancer: A Review. Breast Cancer (Dove Med Press. 2025;17:265\u0026ndash;74. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2147/bctt.s516542\u003c/span\u003e\u003cspan address=\"10.2147/bctt.s516542\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePoggio F, Bruzzone M, Ceppi M, Pond\u0026eacute; NF, La Valle G, Del Mastro L, de Azambuja E, Lambertini M. Platinum-based neoadjuvant chemotherapy in triple-negative breast cancer: a systematic review and meta-analysis. Ann Oncol. 2018;29:1497\u0026ndash;508. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/annonc/mdy127\u003c/span\u003e\u003cspan address=\"10.1093/annonc/mdy127\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evon Minckwitz G, Schneeweiss A, Loibl S, Salat C, Denkert C, Rezai M, Blohmer JU, Jackisch C, Paepke S, Gerber B, et al. Neoadjuvant carboplatin in patients with triple-negative and HER2-positive early breast cancer (GeparSixto; GBG 66): a randomised phase 2 trial. Lancet Oncol. 2014;15:747\u0026ndash;56. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/s1470-2045(14)70160-3\u003c/span\u003e\u003cspan address=\"10.1016/s1470-2045(14)70160-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMasuda N, Lee SJ, Ohtani S, Im YH, Lee ES, Yokota I, Kuroi K, Im SA, Park BW, Kim SB, et al. Adjuvant Capecitabine for Breast Cancer after Preoperative Chemotherapy. N Engl J Med. 2017;376:2147\u0026ndash;59. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1056/NEJMoa1612645\u003c/span\u003e\u003cspan address=\"10.1056/NEJMoa1612645\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchmid P, Cortes J, Pusztai L, McArthur H, K\u0026uuml;mmel S, Bergh J, Denkert C, Park YH, Hui R, Harbeck N, et al. Pembrolizumab for Early Triple-Negative Breast Cancer. N Engl J Med. 2020;382:810\u0026ndash;21. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1056/NEJMoa1910549\u003c/span\u003e\u003cspan address=\"10.1056/NEJMoa1910549\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchmid P, Cortes J, Dent R, Pusztai L, McArthur H, K\u0026uuml;mmel S, Bergh J, Denkert C, Park YH, Hui R, et al. Event-free Survival with Pembrolizumab in Early Triple-Negative Breast Cancer. N Engl J Med. 2022;386:556\u0026ndash;67. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1056/NEJMoa2112651\u003c/span\u003e\u003cspan address=\"10.1056/NEJMoa2112651\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchmid P, Cortes J, Dent R, McArthur H, Pusztai L, K\u0026uuml;mmel S, Denkert C, Park YH, Hui R, Harbeck N, et al. Overall Survival with Pembrolizumab in Early-Stage Triple-Negative Breast Cancer. N Engl J Med. 2024;391:1981\u0026ndash;91. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1056/NEJMoa2409932\u003c/span\u003e\u003cspan address=\"10.1056/NEJMoa2409932\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePanda VK, Mishra B, Mahapatra S, Swain B, Malhotra D, Saha S, Khanra S, Mishra P, Majhi S, Kumari K et al. (2025). Molecular Insights on Signaling Cascades in Breast Cancer: A Comprehensive Review. Cancers (Basel) \u003cem\u003e17\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/cancers17020234\u003c/span\u003e\u003cspan address=\"10.3390/cancers17020234\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFerrari P, Scatena C, Ghilli M, Bargagna I, Lorenzini G, Nicolini A. (2022). Molecular Mechanisms, Biomarkers and Emerging Therapies for Chemotherapy Resistant TNBC. Int J Mol Sci \u003cem\u003e23\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/ijms23031665\u003c/span\u003e\u003cspan address=\"10.3390/ijms23031665\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eImani S, Farghadani R, Roozitalab G, Maghsoudloo M, Emadi M, Moradi A, Abedi B, Kaboli J, P. Reprogramming the breast tumor immune microenvironment: cold-to-hot transition for enhanced immunotherapy. J Exp Clin Cancer Res. 2025;44:131. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s13046-025-03394-8\u003c/span\u003e\u003cspan address=\"10.1186/s13046-025-03394-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJing SY, Wang HQ, Lin P, Yuan J, Tang ZX, Li H. Quantifying and interpreting biologically meaningful spatial signatures within tumor microenvironments. NPJ Precis Oncol. 2025;9:68. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41698-025-00857-1\u003c/span\u003e\u003cspan address=\"10.1038/s41698-025-00857-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMo CK, Liu J, Chen S, Storrs E, Targino da Costa ALN, Houston A, Wendl MC, Jayasinghe RG, Iglesia MD, Ma C, et al. Tumour evolution and microenvironment interactions in 2D and 3D space. Nature. 2024;634:1178\u0026ndash;86. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41586-024-08087-4\u003c/span\u003e\u003cspan address=\"10.1038/s41586-024-08087-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKaraayvaz M, Cristea S, Gillespie SM, Patel AP, Mylvaganam R, Luo CC, Specht MC, Bernstein BE, Michor F, Ellisen LW. Unravelling subclonal heterogeneity and aggressive disease states in triple-negative breast cancer through single-cell RNA-seq. Nat Commun. 2018;9:3588.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePal B, Chen Y, Vaillant F, Capaldo BD, Joyce R, Song X, Bryant VL, Penington JS, Di Stefano L, Tubau Ribera N et al. A single-cell RNA expression atlas of normal, preneoplastic and tumorigenic states in the human breast. (2021). EMBO J. 40.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarco Salas S, Kuemmerle LB, Mattsson-Langseth C, Tismeyer S, Avenel C, Hu T, Rehman H, Grillo M, Czarnewski P, Helgadottir S, et al. Optimizing Xenium In Situ data utility by quality assessment and best-practice analysis workflows. Nat Methods. 2025;22:813\u0026ndash;23. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41592-025-02617-2\u003c/span\u003e\u003cspan address=\"10.1038/s41592-025-02617-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStringer C, Wang T, Michaelos M, Pachitariu M. Cellpose: a generalist algorithm for cellular segmentation. Nat Methods. 2021;18:100\u0026ndash;6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41592-020-01018-x\u003c/span\u003e\u003cspan address=\"10.1038/s41592-020-01018-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePetukhov V, Xu RJ, Soldatov RA, Cadinu P, Khodosevich K, Moffitt JR, Kharchenko PV. Cell segmentation in imaging-based spatial transcriptomics. Nat Biotechnol. 2022;40:345\u0026ndash;54. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41587-021-01044-w\u003c/span\u003e\u003cspan address=\"10.1038/s41587-021-01044-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSatija R, Farrell JA, Gennert D, Schier AF, Regev A. Spatial reconstruction of single-cell gene expression data. Nat Biotechnol. 2015;33:495\u0026ndash;502. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/nbt.3192\u003c/span\u003e\u003cspan address=\"10.1038/nbt.3192\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJin S, Guerrero-Juarez CF, Zhang L, Chang I, Ramos R, Kuan CH, Myung P, Plikus MV, Nie Q. Inference and analysis of cell-cell communication using CellChat. Nat Commun. 2021;12:1088. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41467-021-21246-9\u003c/span\u003e\u003cspan address=\"10.1038/s41467-021-21246-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBlondel VD, Guillaume J-L, Lambiotte R, Lefebvre E. (2008). Fast unfolding of communities in large networks. Journal of statistical mechanics: theory and experiment \u003cem\u003e2008\u003c/em\u003e, P10008.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHafemeister C, Satija R. Normalization and variance stabilization of single-cell RNA-seq data using regularized negative binomial regression. Genome Biol. 2019;20:296. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s13059-019-1874-1\u003c/span\u003e\u003cspan address=\"10.1186/s13059-019-1874-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu SZ, Al-Eryani G, Roden DL, Junankar S, Harvey K, Andersson A, Thennavan A, Wang C, Torpy JR, Bartonicek N, et al. A single-cell and spatially resolved atlas of human breast cancers. Nat Genet. 2021;53:1334\u0026ndash;47. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41588-021-00911-1\u003c/span\u003e\u003cspan address=\"10.1038/s41588-021-00911-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eComprehensive molecular portraits of human breast tumours. Nature. 2012;490:61\u0026ndash;70. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/nature11412\u003c/span\u003e\u003cspan address=\"10.1038/nature11412\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoldman MJ, Craft B, Hastie M, Repečka K, McDade F, Kamath A, Banerjee A, Luo Y, Rogers D, Brooks AN et al. (2020). Visualizing and interpreting cancer genomics data via the Xena platform. In Nat Biotechnol, pp. 675\u0026ndash;678. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41587-020-0546-8\u003c/span\u003e\u003cspan address=\"10.1038/s41587-020-0546-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLachmann A, Xie Z, Ma'ayan A. blitzGSEA: efficient computation of gene set enrichment analysis through gamma distribution approximation. Bioinformatics. 2022;38:2356\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/bioinformatics/btac076\u003c/span\u003e\u003cspan address=\"10.1093/bioinformatics/btac076\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang L, Wang XI, Zhang S. Tumor-infiltrating lymphocyte volume is a better predictor of neoadjuvant therapy response and overall survival in triple-negative invasive breast cancer. Hum Pathol. 2018;80:47\u0026ndash;54. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.humpath.2018.05.024\u003c/span\u003e\u003cspan address=\"10.1016/j.humpath.2018.05.024\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan den Ende NS, Nguyen AH, Jager A, Kok M, Debets R, van Deurzen CHM. (2023). Triple-Negative Breast Cancer and Predictive Markers of Response to Neoadjuvant Chemotherapy: A Systematic Review. Int J Mol Sci \u003cem\u003e24\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/ijms24032969\u003c/span\u003e\u003cspan address=\"10.3390/ijms24032969\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePassarelli MC, Pinzaru AM, Asgharian H, Liberti MV, Heissel S, Molina H, Goodarzi H, Tavazoie SF. Leucyl-tRNA synthetase is a tumour suppressor in breast cancer and regulates codon-dependent translation dynamics. Nat Cell Biol. 2022;24:307\u0026ndash;15. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41556-022-00856-5\u003c/span\u003e\u003cspan address=\"10.1038/s41556-022-00856-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGollwitzer P, Gr\u0026uuml;tzmacher N, Wilhelm S, K\u0026uuml;mmel D, Demetriades C. A Rag GTPase dimer code defines the regulation of mTORC1 by amino acids. Nat Cell Biol. 2022;24:1394\u0026ndash;406. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41556-022-00976-y\u003c/span\u003e\u003cspan address=\"10.1038/s41556-022-00976-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSung Y, Yoon I, Han JM, Kim S. Functional and pathologic association of aminoacyl-tRNA synthetases with cancer. Exp Mol Med. 2022;54:553\u0026ndash;66. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s12276-022-00765-5\u003c/span\u003e\u003cspan address=\"10.1038/s12276-022-00765-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSinclair LV, Rolf J, Emslie E, Shi YB, Taylor PM, Cantrell DA. Control of amino-acid transport by antigen receptors coordinates the metabolic reprogramming essential for T cell differentiation. Nat Immunol. 2013;14:500\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/ni.2556\u003c/span\u003e\u003cspan address=\"10.1038/ni.2556\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFeig C, Jones JO, Kraman M, Wells RJ, Deonarine A, Chan DS, Connell CM, Roberts EW, Zhao Q, Caballero OL, et al. Targeting CXCL12 from FAP-expressing carcinoma-associated fibroblasts synergizes with anti\u0026ndash;PD-L1 immunotherapy in pancreatic cancer. Proc Natl Acad Sci U S A. 2023;110:20212\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGriffiths JI, Cosgrove PA, Medina EF, Nath A, Chen J, Adler FR, Chang JT, Khan QJ, Bild AH. Cellular interactions within the immune microenvironment underpin resistance to cell cycle inhibition in breast cancers. Nat Commun. 2025;16:2132.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBassez A, Vos H, Van Dyck L, Floris G, Arijs I, Desmedt C, Boeckx B, Vanden Bempt M, Nevelsteen I, Lambein K, et al. A single-cell map of intratumoral changes during anti-PD-1 treatment of patients with breast cancer. ) Nat Med. 2021;27:820\u0026ndash;32.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMi H, Varadhan R, Cimino-Mathews AM, Emens LA, Santa-Maria CA, Popel AS. (2023). Spatial and Compositional Biomarkers in Tumor Microenvironment Predicts Clinical Outcomes in Triple-Negative Breast Cancer. bioRxiv. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1101/2023.12.18.572234\u003c/span\u003e\u003cspan address=\"10.1101/2023.12.18.572234\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEl-Hachem N, Leclercq M, Susaeta Ruiz M, Vanleyssem R, Shostak K, K\u0026ouml;rner PR, Capron C, Martin-Morales L, Roncarati P, Lavergne A, et al. Valine aminoacyl-tRNA synthetase promotes therapy resistance in melanoma. Nat Cell Biol. 2024;26:1154\u0026ndash;64. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41556-024-01439-2\u003c/span\u003e\u003cspan address=\"10.1038/s41556-024-01439-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWei Z, Liu X, Cheng C, Yu W, Yi P. Metabolism of Amino Acids in Cancer. Front Cell Dev Biol. 2020;8:603837. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fcell.2020.603837\u003c/span\u003e\u003cspan address=\"10.3389/fcell.2020.603837\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFernandes SA, Angelidaki DD, N\u0026uuml;chel J, Pan J, Gollwitzer P, Elkis Y, Artoni F, Wilhelm S, Kovacevic-Sarmiento M, Demetriades C. Spatial and functional separation of mTORC1 signalling in response to different amino acid sources. Nat Cell Biol. 2024;26:1918\u0026ndash;33. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41556-024-01523-7\u003c/span\u003e\u003cspan address=\"10.1038/s41556-024-01523-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1\u003c/strong\u003e Patient characteristics and clinical response\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"533\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003en\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003epCR\u003c/p\u003e\n \u003cp\u003en\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003epCR rate\u003c/p\u003e\n \u003cp\u003e%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eUnivariate\u003c/p\u003e\n \u003cp\u003e\u003cem\u003eP\u003c/em\u003e value*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e≤50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e(51.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e48.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026gt;50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e(49.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e20.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.0434*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eBRCA1/2\u0026nbsp;\u003c/em\u003estatus\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ewild\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e(32.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e25.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003epathogenic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e(10.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e80.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.0267*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eVUS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e( 4.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e50.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eunknown\u0026nbsp;\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e(53.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e30.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eTumor size\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e≤2cm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e(30.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e53.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026gt;2cm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e(69.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e26.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.0721\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eNodal status\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003enegative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e(65.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e43.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003epositive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e(34.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e17.7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.0593\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eTILs\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e≤20%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e(30.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e13.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026gt;20%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e(32.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e37.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.5179\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eunknown\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e(36.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e38.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eKi-67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e≤20%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e( 6.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e33.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026gt;20%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e(61.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e30.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.9054\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eunknown\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e(32.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e43.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003epCR\u0026nbsp;\u003c/em\u003epathological complete response, \u003cem\u003eBRCA\u0026nbsp;\u003c/em\u003eBreast Cancer Susceptibility Gene,\u003cem\u003e\u0026nbsp;VUS\u003c/em\u003e variant of unknown significance, \u003cem\u003eTILs\u0026nbsp;\u003c/em\u003etumor infiltrating lymphocytes,\u0026nbsp;\u003csup\u003e*\u003c/sup\u003e\u003cem\u003eP\u003c/em\u003e ≤ 0.05\u003c/p\u003e\n\u003cp\u003e\u003csup\u003ea\u003c/sup\u003eBRCA1/2 status unknown includes those not covered by insurance and those for which patient consent was not obtained.\u003c/p\u003e\n\u003cp\u003e\u003csup\u003eb\u003c/sup\u003eTILs and ki-67 unknown includes cases in which tissue was not available because the diagnosis was made by biopsy at another hospital.\u0026nbsp;\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"breast-cancer-research","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"brcr","sideBox":"Learn more about [Breast Cancer Research](http://breast-cancer-research.biomedcentral.com)","snPcode":"13058","submissionUrl":"https://submission.nature.com/new-submission/13058/3","title":"Breast Cancer Research","twitterHandle":"@BCRJournal","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Triple negative breast cancer, Neoadjuvant chemotherapy, Pathological complete response, Resistance marker, Pretreatment specimen, Gene expression profiling, Spatial transcriptomics, LARS, Cell cycle, Metabolic reprogramming","lastPublishedDoi":"10.21203/rs.3.rs-9572543/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9572543/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eTriple-negative breast cancer (TNBC) accounts for approximately 10\u0026ndash;20% of all breast cancers and is characterized by aggressive clinical behavior and poor prognosis. Pathological complete response (pCR) after neoadjuvant chemotherapy (NAC) is associated with a good prognosis. Recently, the addition of pembrolizumab to NAC has improved pCR and overall survival. However, TNBC is biologically heterogeneous, and existing biomarkers, including tumor-infiltrating lymphocytes and PD-L1 provide limited prediction of chemoresistance, leaving an unmet need for markers that could guide integration of immunotherapy in resistant disease. In this study, we aimed to identify biomarkers of chemoresistance in TNBC and characterize their spatial features by integrating transcriptomic and spatial analyses.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eWe retrospectively analyzed 49 TNBC patients treated with anthracycline- and taxane-based NAC at Nagoya University Hospital between 2017 and 2023. Pre-treatment FFPE biopsies from eight patients (pCR n\u0026thinsp;=\u0026thinsp;4, non-pCR n\u0026thinsp;=\u0026thinsp;4) were profiled by nCounter gene expression analysis and Xenium spatial transcriptomics. The results were further validated using publicly available datasets, including TCGA.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eNo clinicopathological factor was significantly associated with treatment response. nCounter analysis identified reduced expression of immune-related genes and increased expression of proliferation and cell cycle-related genes in non-responders. Spatial transcriptomic profiling revealed greater immune-cell abundance and diversity in the pCR group, with stronger immune-cell and immune-tumor cell interactions. In contrast, the non-pCR group showed enhanced stromal\u0026minus;epithelial interactions and reduced spatial proximity between immune and tumor cells. Cluster-and cell-type resolved analysis identified reduced expression of \u003cem\u003eCD8A\u003c/em\u003e and other T-cell associated genes (\u003cem\u003eCD3E\u003c/em\u003e, \u003cem\u003eGZMA\u003c/em\u003e) and cell-cycle genes (\u003cem\u003eCCNA1\u003c/em\u003e, \u003cem\u003eCCNB1\u003c/em\u003e, \u003cem\u003eE2F3\u003c/em\u003e) in non-pCR tumors. The inverse spatial relationship between \u003cem\u003eCD8A\u003c/em\u003e and \u003cem\u003eLARS\u003c/em\u003e suggests as a candidate CD8A\u0026minus;LARS axis linking reduced cytotoxic T-cell presence with altered amino acid metabolism.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eReduced \u003cem\u003eCD8A\u003c/em\u003e and elevated \u003cem\u003eLARS\u003c/em\u003e expression in pretreatment TNBC tumors may contribute to chemoresistance through coordinated metabolic reprogramming, immune-cell exclusion, and tumor-cell proliferation. We propose the \u003cem\u003eCD8A\u003c/em\u003e-\u003cem\u003eLARS\u003c/em\u003e axis as a potential resistance axis warranting functional and prospective validation in independent TNBC cohorts.\u003c/p\u003e","manuscriptTitle":"Targeted spatial profiling identifies a CD8A − LARS axis associated with neoadjuvant chemotherapy resistance in triple-negative breast cancer","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-13 17:56:14","doi":"10.21203/rs.3.rs-9572543/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-05-13T02:07:07+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"261724904306495759966060263921504343368","date":"2026-05-07T11:20:56+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"292942193483323882195409835708010569367","date":"2026-05-07T03:26:51+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"264310533626455455429508175123811708165","date":"2026-05-06T08:04:24+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"12810686272051371034696230433528306727","date":"2026-05-05T12:06:10+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-05-05T03:00:23+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-05-01T13:12:24+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-05-01T08:54:49+00:00","index":"","fulltext":""},{"type":"submitted","content":"Breast Cancer Research","date":"2026-04-30T05:40:20+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"breast-cancer-research","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"brcr","sideBox":"Learn more about [Breast Cancer Research](http://breast-cancer-research.biomedcentral.com)","snPcode":"13058","submissionUrl":"https://submission.nature.com/new-submission/13058/3","title":"Breast Cancer Research","twitterHandle":"@BCRJournal","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4e3515c4-1b4e-4d55-9fc1-bfe8de5e6f86","owner":[],"postedDate":"May 13th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-05-13T02:07:07+00:00","index":31,"fulltext":""},{"type":"reviewerAgreed","content":"261724904306495759966060263921504343368","date":"2026-05-07T11:20:56+00:00","index":30,"fulltext":""},{"type":"reviewerAgreed","content":"292942193483323882195409835708010569367","date":"2026-05-07T03:26:51+00:00","index":28,"fulltext":""},{"type":"reviewerAgreed","content":"264310533626455455429508175123811708165","date":"2026-05-06T08:04:24+00:00","index":26,"fulltext":""},{"type":"reviewerAgreed","content":"12810686272051371034696230433528306727","date":"2026-05-05T12:06:10+00:00","index":23,"fulltext":""},{"type":"reviewersInvited","content":"16","date":"2026-05-05T03:00:23+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-05-01T13:12:24+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-05-01T08:54:49+00:00","index":"","fulltext":""},{"type":"submitted","content":"Breast Cancer Research","date":"2026-04-30T05:40:20+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-13T17:56:14+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-13 17:56:14","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9572543","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9572543","identity":"rs-9572543","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}