TMPRSS11B promotes an acidified microenvironment and immune suppression in squamous lung cancer

Abstract

Lung cancer is the leading cause of cancer-related deaths worldwide. Existing therapeutic options have limited efficacy, particularly for lung squamous cell carcinoma (LUSC), underscoring the critical need for the identification of new therapeutic targets. We previously demonstrated that the Transmembrane Serine Protease TMPRSS11B promotes the transformation of human bronchial epithelial cells and enhances lactate export from LUSC cells. Here, we evaluate the impact of TMPRSS11B activity on the host immune system and the tumor microenvironment (TME). Tmprss11b depletion significantly reduces tumor burden in immunocompetent mice and triggers an infiltration of immune cells. RNA FISH analysis and spatial transcriptomics in the autochthonous Rosa26-Sox2-Ires-Gfp^LSL/LSL; Nkx2-1^fl/fl; Lkb1^fl/fl (SNL) model reveal an enrichment of Tmprss11b expression in LUSC tumors, specifically in Krt13⁺ hillock-like cells. Furthermore, utilizing ultra-pH-sensitive nanoparticle imaging and metabolite analysis, we identify regions of acidification, elevated lactate, and enrichment of immunosuppressive (M2-like) macrophages in LUSC tumors. These results demonstrate that TMPRSS11B promotes an acidified and immunosuppressive TME and nominate this enzyme as a therapeutic target in LUSC.

Keywords: Hillock Cells; Immune Suppression; Lactate-mediated TME Acidification; Squamous Cell Lung Cancer; Transmembrane Serine Protease TMPRSS11B.

Figure 1. Tmprss11b depletion reduces tumor growth and enhances immune cell infiltration.

(A) Schematic of the syngeneic experiments performed using KLN205 murine LUSC cells. Figure created in BioRender. (B) qRT-PCR analysis of Tmprss11b mRNA in KLN205 cells expressing control sgRNA or two independent Tmprss11b sgRNA. Ordinary one-way ANOVA with Dunnett’s multiple comparisons test was used for the analysis. Plot represents mean ± standard deviation (SD); n = 4 per group (technical replicates), ****P < 0.0001 for all comparisons. The experiment was repeated three times for confirmation (biological replicates). (C) Quantification of tumor volumes of KLN205 tumors expressing control or Tmprss11b sgRNA on day 45 (terminal measurement) post injection in syngeneic DBA/2 wild-type mice (n = 9 control sgRNA mice; n = 12 Tmprss11b sg1 mice; n = 12 Tmprss11b sg2 mice) (biological replicates), P = 0.0036 (T11b sgRNA1), P = 0.0022 (T11b sgRNA2). Brown–Forsythe and Welch ANOVA test with Dunnett’s T3 multiple comparisons test was used for the statistical analysis. The plot represents mean with 95% confidence interval (CI). (D) qRT-PCR analysis of Tmprss11b mRNA in KLN205 cells expressing doxycycline-inducible control shRNA or two independent shRNA sequences targeting Tmprss11b. RM one-way ANOVA with Dunnett’s multiple comparisons test was used for the analysis. Plot represents mean ± standard deviation (SD); n = 4 per group (technical replicates), P = 0.0012 (T11b shRNA1), P = 0.0013 (T11b shRNA2). The experiment was repeated three times for confirmation (biological replicates). (E) Quantification of tumor volumes of KLN205 cells expressing doxycycline-inducible control or Tmprss11b shRNA in syngeneic DBA/2 wild-type mice. (n = 9 control shRNA mice; n = 6 Tmprss11b sh1 mice; n = 8 Tmprss11b sh2 mice, biological replicates), P = 0.0013 (T11b shRNA1), P < 0.0001 (T11b shRNA2). Linear mixed-effects models were used to investigate the differences in tumor volume over time among the three groups. The plot represents the standard error of the mean (SEM). (F) Fluorescent immunohistochemistry (IHC-F) staining for CD4 + T cells in KLN205 tumor sections, with quantification. An unpaired t test with Welch’s correction was used to compare CD4 + T cells between the groups (n = 6 fields per tumor section, three tumors per group, biological replicates), P = 0.0042. Plot represents mean ± SD. Scale bar, 50 μm. (G) Fluorescent immunohistochemistry (IHC-F) staining for phosphorylated ERK (pERK) in KLN205 tumors, with quantification (right). An unpaired t test with Welch’s correction was used to compare phosphorylated ERK levels between the groups (n = 10–14 fields per tumor section, three tumors per group, biological replicates), P = 0.0008. Plot represents mean ± SD. Scale bar, 50 μm. Source data are available online for this figure.

Figure 2. Tmprss11b expression is enriched in mouse lung squamous tumors.

(A) RNA sequencing analysis of tumors from various mouse models of lung cancer. The y-axis represents the normalized counts for Tmprss11b. One-way ANOVA with Dunnett’s multiple comparisons test was used for the statistical analysis (RPR2 mice n = 5; RPM mice n = 15; SNL mice n = 4; LP mice n = 6; SL mice n = 9; KP mice n = 8, biological replicates), ****P < 0.0001 (RPR2, RPM, KP), *P = 0.0133 (LP). Plot represents mean ± SD. (B) Schematic representation of Ad-Cre mediated tumor induction in SNL mice. Figure created in BioRender. (C) Representative MRI images of the mice in (B), 3- & 4-months post infection. Red outlines denote tumors (Biological replicates n > 3). (D) Representative H&E images of SNL mouse lung, 7 months post infection with Ad-Cre showing distinct regions of LUSC and mucinous LUAD (Biological replicates n > 3). Scale bar, 100 μm. (E) H&E image of SNL mouse lung (11 months post infection with Ad-Cre) and RNAscope of Tmprss11b on a serial section. Left, red outline denotes squamous tumors based on H&E staining. Right, yellow outline denotes regions with Tmprss11b expression (red) corresponding to the regions of squamous tumors. The staining was repeated three times with serial sections (technical replicates) and with lung sections from different mice (n = 4, biological replicates). Scale bar, 2 mm. (F) Zoom-in of (E) showing Tmprss11b expression by RNAScope in squamous tumors (top panel) and normal lung (bottom panel). Scale bar, 200 μm. (G) Representative H&E image with annotations and RNAscope analysis of Tmprss11b (red) and Sox2 (green) in SNL lung sections. Scale bar, 400 μm. Source data are available online for this figure.

Figure 3. Tmprss11b expression correlates with an increase in squamous markers and oncogenic signaling pathways.

(A) Image depicting the different clusters identified in a SNL lung section based on spatial transcriptomic analysis. Scale bar, 2 mm. (B) H&E image of the lung section from A) (left) and manual annotation of different regions (right); LUSC (S1, S2, S3) and LUAD (A1 and A2). Scale bar, 2 mm. (C) H&E image of the lung section from A), annotated with regions of LUSC and LUAD, and corresponding spatial plots depicting the distribution of the indicated mRNAs. Scale bar, 1 mm. (D) Gene set enrichment analysis (GSEA) from T11b-high versus low squamous tumors (S1&S2 versus S3) with normalized enrichment scores (NES), false discovery rate (FDR) and P values for the indicated gene signatures. The nominal P and FDR values were obtained from the “GSEA Preranked” tool (from Broad Institute) using a weighted scoring scheme. Gene sets were evaluated based on the default normalized enrichment score method, and statistical significance was determined by bootstrapping with 1000 permutations. (E) Bar graph representing top squamous genes/oncogenes from differential gene expression (DEG) analysis of Tmprss11b-high versus low LUSC spatial transcriptomics data. Differential gene expression was calculated using Seurat’s FindAllMarkers function, and direct comparisons between two classes and corresponding P values were obtained using Seurat’s FindMarkers function with the Wilcoxon rank-sum test with Bonferroni P value correction. Source data are available online for this figure.

Figure 4. TMPRSS11B is enriched in the KRT13⁺ hillock-like cells and induced by KLF4.

(A) H&E image of the lung section from 3 A), annotated with regions of LUSC and LUAD, (left) and corresponding spatial plots depicting the distribution of the indicated mRNAs. Scale bar, 1 mm. (B) Representative H&E image and RNAscope analysis of Tmprss11b, Trp63, and Krt13 in SNL lung sections. The staining was repeated with lung sections from different mice (n = 2–3, biological replicates). Scale bar, 50 μm. (C) Tmprss11b transcript abundance in LUSC patient tumors (n = 500) relative to LUAD tumors (n = 512) (data obtained from TCGA-GDC (Cancer Genome Atlas Research et al, ; Heath et al, 2021)). Welch’s t test was used for the analysis, ****P < 0.0001. Plot represents median with 95% CI. (D) qRT-PCR analysis of Tmprss11b mRNA in BEAS-2B cells expressing doxycycline-inducible GFP or KLF4. Ordinary one-way ANOVA with Dunnett’s multiple comparisons test was used for the analysis. Plot represents mean ± SD; n = 3 (technical replicates) per group, ****P < 0.0001. The experiment was repeated two times to confirm the observations. Source data are available online for this figure.

Figure 5. Infiltration of M2-like macrophages in Tmprss11b-high squamous tumors.

(A) Spatial plots from cell deconvolution analysis showing the distribution of various immune cell populations in LUSCs (top) and mucinous LUADs (bottom). (B) Quantification of immune cell populations in (A). A two-sided Mann–Whitney–Wilcoxon test was used for the statistical analysis (T11b-high LUSC n = 344, LUAD n = 1148, biological replicates), ****P = 7.96E-78 (Classical monocyte), ****P = 4.74E-17 (Alveolar macrophage), ****P = 6.04E-133 (Leukocyte), ***P = 2.31E-05 (Non-classical monocyte). The box plots represent the distribution of values for each group, extending from the 25th percentile to the 75th percentile; T11b-high LUSC: Classical monocyte (minima=0.00035, median=0.32265, maxima=1, Q1 = 0.16193, Q3 = 0.52806), Alveolar macrophages (minima=0.01914, median=0.44117, maxima=0.91769, Q1 = 0.32846, Q3 = 0.56415), Leukocyte (minima=0.9.87E-05, median=0.41385, maxima=1, Q1 = 0.27138, Q3 = 0.61336), Non-classical monocyte (minima=0.00036, median=0.25014, maxima=0.92007, Q1 = 0.13621, Q3 = 0.44976), LUAD: Classical monocyte (minima=0, median=0.00788, maxima=0.47288, Q1 = 0.00108, Q3 = 0.18980), Alveolar macrophages (minima=0.04437, median=0.36127, maxima=0.67967, Q1 = 0.28261, Q3 = 0.44143), Leukocyte (minima=0, median=0.00264, maxima=0.33064, Q1 = 0.00058, Q3 = 0.13260), Non-classical monocyte (minima=0, median=0.22165, maxima=0.68314, Q1 = 0.10763, Q3 = 0.33783). (C) Gene set enrichment analysis (GSEA) of Tmprss11b-high LUSC versus LUAD spatial transcriptomics data with normalized enrichment scores (NES), false discovery rate (FDR) and P values for the indicated gene signatures. The nominal P and FDR values were obtained from the “GSEA Preranked” tool (from Broad Institute) using a weighted scoring scheme. Gene sets were evaluated based on the default normalized enrichment score method, and statistical significance was determined by bootstrapping with 1000 permutations. (D) Bar graph representing the top M2-like/TAM genes from the differential gene expression (DEG) analysis of Tmprss11b-high LUSC versus LUAD spatial transcriptomics data. (E) Representative H&E images and immunohistochemistry (IHC) for SOX2, MSR1 (CD204), HMOX1, and ARG1 in SNL LUSC (top) and mucinous LUAD (bottom). The staining was repeated with lung sections from different mice (n = 3–4, biological replicates). Scale bar, 100 μm. Source data are available online for this figure.

Figure 6. Tmprss11b-high squamous tumors and acidified regions of the TME are enriched for immunosuppressive macrophages.

(A) Representative ex vivo images of lungs from SNL mice, control (no Cre) or infected with Ad-Cre (11 months post infection), 24 h post injection with PDBA ICG 5.3 ultra pH-sensitive (UPS) nanoparticle. Bright-field images (left) and the corresponding ICG fluorescence images (right) are shown. (B) Quantification of mean ICG fluorescence intensity (control n = 5 mice; Ad-Cre n = 3 mice, biological replicates). Unpaired t test with Welch’s correction was used for the analysis, P = 0.0281. Plot represents mean ± SD. (C) H&E image of the lung section from Fig. 3A), with ICG signal on the same section, and RNAscope of Tmprss11b (red) on a serial section. Also shown are spatial plots depicting enrichment of Cd68 (macrophage marker) and Spp1 (immunosuppressive macrophage marker) transcripts in these regions. Scale bar, 1 mm. (D) Gene set enrichment analysis (GSEA) of immune cell pathways from the low pH versus high pH spatial transcriptomics data with normalized enrichment scores (NES), false discovery rate (FDR) and p values for the indicated gene signatures. The nominal P and FDR values were obtained from the “GSEA Preranked” tool (from Broad Institute) using a weighted scoring scheme. Gene sets were evaluated based on the default normalized enrichment score method, and statistical significance was determined by bootstrapping with 1000 permutations. (E) Bar graph representing top M2-like/TAM genes from the differential gene expression (DEG) analysis of low pH versus high pH spatial transcriptomics data. Differential gene expression was calculated using Seurat’s FindAllMarkers function, and direct comparisons between two classes and corresponding P values were obtained using Seurat’s FindMarkers function with the Wilcoxon rank-sum test with Bonferroni P value correction. (F) Representative ICG fluorescence images (left) and the corresponding immunohistochemistry (IHC) validation of HMOX1 and MSR1 (right) in SNL lung sections. The staining was repeated with lung sections from different mice (n = 2, biological replicates). Scale bar, 500 μm. Source data are available online for this figure.

Figure 7. Tmprss11b-high squamous tumors and acidified regions have elevated levels of lactate.

(A) Schematic representation of the experimental pipeline to assess metabolites in SNL lung tumors through laser capture microdissection (LMD) followed by mass spectrometry. Figure created in BioRender. (B) Representative ICG fluorescence images (depicting nanoparticle accumulation) (green), H&E image (middle), and bright-field image of the serial section (right) used for LMD with the same region annotated in green. Scale bar, 200 μm. (C) Quantification of lactate concentration (pg/μm²) in SNL lung tissues; normal, LUADs and LUSCs (n = 7–11 regions per group from a total of 2 mice, biological replicates). Ordinary one-way ANOVA with Dunnett’s multiple comparisons test was used for the statistical analysis, P = 0.0057. The plot represents mean with 95% CI. (D) Quantification of lactate concentration (pg/μm²) in the regions of low acidity (high pH) and high acidity (low pH), in the SNL lung tissues (n = 12–20 regions per group from a total of two mice, biological replicates). Ordinary one-way ANOVA with Dunnett’s multiple comparisons test was used for the statistical analysis, P = 0.0013. The plot represents the mean with 95% CI. (E) Schematic representation of the immunosuppressive niche established in Tmprss11b-high lung squamous tumors and the surrounding acidified regions in the tumor microenvironment. Figure created in BioRender. Source data are available online for this figure.

Figure EV1. Tmprss11b depletion inhibits tumor burden in a syngeneic mouse model of LUSC.

(A) Agarose gel electrophoresis images of PCR amplified products from the Surveyor assay performed on the genomic DNA isolated from KLN205 cells expressing control or Tmprss11b sg1 or Tmprss11b sg2. The assay was repeated two times with different surveyor primers to confirm the observations (biological replicates). (B) Image showing the resected tumors at endpoint from the syngeneic experiment in Fig. 1C. (C) qRT-PCR analysis of Tmprss11b mRNA in KLN205 cells expressing doxycycline-inducible control shRNA or two independent shRNA sequences targeting Tmprss11b. Brown–Forsythe and Welch ANOVA test with Dunnett’s T3 multiple comparisons test was used for the statistical analysis, ****P < 0.0001. Plot represents mean ± SD; n = 4 per group (technical replicates). Experiment was repeated two times for confirmation (biological replicates). (D) Quantification of tumor volumes of KLN205 cells expressing doxycycline-inducible control or Tmprss11b shRNA on day 50 (terminal measurement) post injection in syngeneic DBA/2 wild-type mice (n = 10 control shRNA mice; n = 8 Tmprss11b sh1 mice; n = 10 Tmprss11b sh2 mice, biological replicates). Ordinary one-way ANOVA with Dunnett’s multiple comparisons test was used for the statistical analysis, P = 0.0235. Plot represents mean ± SD. (E) Image showing the resected tumors from the syngeneic experiment in Fig. 1D. (F) qRT-PCR analysis of Tmprss11b mRNA in the resected KLN205 tumors from D). Brown–Forsythe and Welch ANOVA test with Dunnett’s T3 multiple comparisons test was used for the statistical analysis, P = 0.0055 (T11b shRNA1), P = 0.0041(T11b shRNA2). Plot represents mean ± SD; n = 3 technical replicates, n = 3–6 tumors per group, biological replicates. (G) Quantification of CD8a staining from fluorescent immunohistochemistry (IHC-F) performed on KLN205 tumor sections. An unpaired t test with Welch’s correction was used for the statistical analysis (n = 10–14 fields per tumor section, 3 tumors per group, biological replicates). Plot represents mean ± SD. Source data are available online for this figure.

Figure EV2. Tmprss11b-high squamous tumors have increased expression of oncogenes and squamous markers.

(A) Spatial plots depicting the read counts before and after the quality control process. Scale bar, 1 mm. (B) Volcano plot showing the top differentially expressed genes in the Tmprss11b-high versus low LUSCs spatial data. A total of 4090 genes are plotted, of which 970 pass filter (colored). A two-tailed Wilcoxon rank-sum test with Bonferroni correction was used for the statistical analysis (n = 4280 Visium spots post-QC, biological replicates). (C) Violin plot depicting normalized counts for Tmprss11b transcript in the annotated regions (from Fig. 2B) (n = 1597, biological replicates). (D) Gene set enrichment analysis (GSEA) of the Tmprss11b-high LUSC versus LUAD spatial data with normalized enrichment scores (NES), false discovery rate (FDR) and P values for the indicated gene signatures. The nominal P and FDR values were obtained from the “GSEA Preranked” tool (from Broad Institute) using a weighted scoring scheme. Gene sets were evaluated based on the default normalized enrichment score method, and statistical significance was determined by bootstrapping with 1000 permutations. (E) Top squamous markers and known oncogenes from the differential gene expression (DEG) analysis of the Tmprss11b-high LUSC versus LUAD spatial data. Differential gene expression was calculated using Seurat’s FindAllMarkers function, and direct comparisons between two classes and corresponding P values were obtained using Seurat’s FindMarkers function with the Wilcoxon rank-sum test with Bonferroni P value correction. (F) Representative H&E and immunohistochemistry (IHC) validation for KRT16 and LYPD3 in LUSC (top) and mucinous LUAD (bottom). The staining was repeated with lung sections from different mice (n = 2–3, biological replicates). Scale bar, 100 μm. Source data are available online for this figure

Figure EV3. Modulation of TMPRSS11B expression is accompanied by coordinated changes in expression of Keratin genes.

(A) Top downregulated Keratin genes from differential gene expression analysis of control shRNA versus Tmprss11b shRNA bulk RNA sequencing from the KLN205 syngeneic experiment in Fig. 1D. The log2FC change depicts the reduction in expression of the indicated genes in the Tmprss11b knockdown tumors compared to the control. (B) Top Keratin genes from the differential gene expression (DEG) analysis of the Tmprss11b-high versus low in LUSC spatial data from SNL lung tumors. (C) Top keratin genes from the differential gene expression (DEG) analysis of the Tmprss11b-high LUSC versus LUAD spatial data from SNL lung tumors. (D) Top Keratin genes from the differential gene expression (DEG) analysis of TMPRSS11B-high versus low LUSC human tumors from TCGA. (E) Venn diagram depicting overlapping Keratin genes from the gene lists in (A–D). Source data are available online for this figure

Figure EV4. Tmprss11b-high squamous tumors show enrichment for hillock and basal gene signatures.

(A) H&E image of the lung section from 3 A), annotated with regions of LUSC and LUAD, (left) and spatial plots from the transcriptomic data depicting the distribution of the indicated gene signatures (right). Scale bar, 1 mm. (B) Violin plots representing the enrichment for the indicated gene signatures in Tmprss11b-high LUSC and LUAD. A two-tailed Wilcoxon rank-sum test was used for the statistical analysis (n = 1492, biological replicates). P < 2.22e-16 for all comparisons. Source data are available online for this figure

Figure EV5. Tmprss11b-high squamous tumors have higher infiltration of neutrophils.

(A) Gene set enrichment analysis (GSEA) of Tmprss11b-high LUSC versus LUAD spatial transcriptomics data with normalized enrichment scores (NES) and false discovery rate (FDR) for the indicated gene signatures; analysis performed using WEB-based Gene Set Analysis Toolkit. (B) H&E image of the lung section from Fig. 3A), annotated with regions of LUSC and LUAD, and corresponding spatial plots depicting the distribution of the indicated mRNAs (neutrophil markers). (C) Representative H&E and immunohistochemistry (IHC) validation for MPO, CD11b (ITGAM) and LY6G in LUSC (top) and mucinous LUAD (bottom). The staining was repeated with lung sections from different mice (n = 2, biological replicates). Scale bar, 100 μm. Source data are available online for this figure

Figure EV6. Tmprss11b-high squamous tumors show enrichment for M2-like macrophage markers.

(A) Quantification of immune cell populations in Tmprss11b-high LUSC vs. LUAD using cell deconvolution analysis of the spatial data. A two-sided Mann–Whitney–Wilcoxon test was used for the statistical analysis (T11b-high LUSC n = 344, LUAD n = 1148, biological replicates), ****P = 7.96E-78 (Classical monocyte), ****P = 7.63E-16 (T cell), ****P = 4.74E-17 (Alveolar macrophage), ****P = 6.04E-133 (Leukocyte), ***P = 2.31E-05 (Non-classical monocyte), ****P = 2.84E-10 (Lung endothelial cell), ****P = 1.28E-111 (Stromal cell), ****P = 3.92E-09 (B cell), ****P = 4.41E-44 (Ciliated columnar cell of tracheobronchial tree). The box plots represent the distribution of values for each group, extending from the 25th percentile to the 75th percentile; T11b-high LUSC: Classical monocyte (minima=0.00035, median=0.32265, maxima=1, Q1 = 0.16193, Q3 = 0.52806), Natural killer cell (minima=0.04763, median=0.37114, maxima=0.82722, Q1 = 0.28091, Q3 = 0.49943), T cell (minima=0, median=0.18974, maxima=0.72977, Q1 = 0.06047, Q3 = 0.32819), Alveolar macrophages (minima=0.01914, median=0.44117, maxima=0.91769, Q1 = 0.32846, Q3 = 0.56415), Leukocyte (minima=0.9.87E-05, median=0.41385, maxima=1, Q1 = 0.27138, Q3 = 0.61336), Myeloid cell (minima=0.00012, median=0.04951, maxima=0.61895, Q1 = 0.00328, Q3 = 0.24955), Non-classical monocyte (minima=0.00036, median=0.25014, maxima=0.92007, Q1 = 0.13621, Q3 = 0.44976), Lung endothelial cell (minima=0, median=0.18464, maxima=0.52498, Q1 = 0.10807, Q3 = 0.27484), Stromal cell (minima=0, median=0.44905, maxima=0.94097, Q1 = 0.31863, Q3 = 0.56757), B cell (minima=0, median=0.24525, maxima=0.69056, Q1 = 0.13767, Q3 = 0.35883), Mast cell (minima=0.00046, median=0.09658, maxima=0.30038, Q1 = 0.06603, Q3 = 0.15977), Ciliated columnar cell of tracheobronchial tree (minima=0.05124, median=0.18792, maxima=0.34003, Q1 = 0.15953, Q3 = 0.23173), Type II pneumocyte (minima=0, median=0.28766, maxima=0.71305, Q1 = 0.19541, Q3 = 0.40246); LUAD: Classical monocyte (minima=0, median=0.00788, maxima=0.47288, Q1 = 0.00108, Q3 = 0.18980), Natural killer cell (minima=0.04933, median=0.39471, maxima=0.76680, Q1 = 0.31838, Q3 = 0.49774), T cell (minima=0.00029, median=0.28494, maxima=0.70883, Q1 = 0.19111, Q3 = 0.39820), Alveolar macrophages (minima=0.04437, median=0.36127, maxima=0.67967, Q1 = 0.28261, Q3 = 0.44143), Leukocyte (minima=0, median=0.00264, maxima=0.33064, Q1 = 0.00058, Q3 = 0.13260), Myeloid cell (minima=0, median=0.05014, maxima=0.64748, Q1 = 0.00252, Q3 = 0.26050), Non-classical monocyte (minima=0, median=0.22165, maxima=0.68314, Q1 = 0.10763, Q3 = 0.33783), Lung endothelial cell (minima=0, median=0.23560, maxima=0.51640, Q1 = 0.16346, Q3 = 0.30464), Stromal cell (minima=0.39036, median=0.75595, maxima=1, Q1 = 0.66414, Q3 = 0.84667), B cell (minima=0, median=0.30153, maxima=0.62486, Q1 = 0.22758, Q3 = 0.38650), Mast cell (minima=0, median=0.08179, maxima=0.20366, Q1 = 0.06479, Q3 = 0.12034), Ciliated columnar cell of tracheobronchial tree (minima=0.08851, median=0.25076, maxima=0.41768, Q1 = 0.21195, Q3 = 0.29424), Type II pneumocyte (minima=0, median=0.29401, maxima=0.63087, Q1 = 0.21685, Q3 = 0.38246). (B) Quantification of immune cell populations in Tmprss11b-high vs. Tmprss11b-low LUSC using cell deconvolution analysis of the spatial data. A two-sided Mann–Whitney-Wilcoxon test was used for the statistical analysis (T11b-high LUSC n = 344, Tmprss11b-low LUSC n = 104, biological replicates), ****P = 8.89E-08 (Classical monocyte), ****P = 3.22E-15 (Alveolar macrophage), ****P = 1.25E-32 (Leukocyte), ***P = 6.54E-05 (Non-classical monocyte), ****P = 2.42E-36 (Ciliated columnar cell of tracheobronchial tree). The box plots represent the distribution of values for each group, extending from the 25^th percentile to the 75^th percentile; T11b-high LUSC: Classical monocyte (minima=0.00029, median=0.30549, maxima=1, Q1 = 0.15744, Q3 = 0.49527), Natural killer cell (minima=0, median=0.41140, maxima=0.97370, Q1 = 0.26653, Q3 = 0.5490), T cell (minima=0, median=0.18983, maxima=0.81820, Q1 = 0.05084, Q3 = 0.35778), Alveolar macrophages (minima=0.00293, median=0.43913, maxima=0.91408, Q1 = 0.34461, Q3 = 0.57240), Leukocyte (minima=0.00068, median=0.40974, maxima=1, Q1 = 0.26157, Q3 = 0.62635), Myeloid cell (minima=4.55E-05, median=0.09014, maxima=0.65331, Q1 = 0.00217, Q3 = 0.26263), Non-classical monocyte (minima=1.67E-05, median=0.26527, maxima=0.85779, Q1 = 0.14324, Q3 = 0.42906), Lung endothelial cell (minima=0, median=0.17896, maxima=0.55746, Q1 = 0.09927, Q3 = 0.28255), Stromal cell (minima=0, median=0.44430, maxima=1, Q1 = 0.26373, Q3 = 0.62252), B cell (minima=0, median=0.24471, maxima=0.69114, Q1 = 0.14794, Q3 = 0.36522), Mast cell (minima=0, median=0.10069, maxima=0.30702, Q1 = 0.06993, Q3 = 0.16477), Ciliated columnar cell of tracheobronchial tree (minima=0.03638, median=0.19021, maxima=0.35185, Q1 = 0.15468, Q3 = 0.23355), Type II pneumocyte (minima=0, median=0.24117, maxima=0.67966, Q1 = 0.1531, Q3 = 0.36374); Tmprss11b-low LUSC: Classical monocyte (minima=0.00020, median=0.19138, maxima=0.60530, Q1 = 0.04595, Q3 = 0.26969), Natural killer cell (minima=0, median=0.37341, maxima=0.84796, Q1 = 0.29730, Q3 = 0.51756), T cell (minima=7.77E-05, median=0.24393, maxima=0.74889, Q1 = 0.10171, Q3 = 0.36058), Alveolar macrophages (minima=0.08051, median=0.30265, maxima=0.60637, Q1 = 0.23290, Q3 = 0.38229), Leukocyte (minima=0, median=0.10504, maxima=0.43837, Q1 = 0.03357, Q3 = 0.19549), Myeloid cell (minima=0.00048, median=0.20803, maxima=0.90453, Q1 = 0.00571, Q3 = 0.36524), Non-classical monocyte (minima=0.00013, median=0.19676, maxima=0.60907, Q1 = 0.08942, Q3 = 0.29728), Lung endothelial cell (minima=0, median=0.18352, maxima=0.39994, Q1 = 0.10900, Q3 = 0.23955), Stromal cell (minima=0.10516, median=0.53043, maxima=0.88675, Q1 = 0.32928, Q3 = 0.69341), B cell (minima=0.00077, median=0.28140, maxima=0.64062, Q1 = 0.17045, Q3 = 0.35852), Mast cell (minima=0, median=0.09511, maxima=0.21496, Q1 = 0.06911, Q3 = 0.12745), Ciliated columnar cell of tracheobronchial tree (minima=0.13684, median=0.37548, maxima=0.66299, Q1 = 0.30862, Q3 = 0.45037), Type II pneumocyte (minima=0.06470, median=0.26157, maxima=0.47461, Q1 = 0.21841, Q3 = 0.32089). (C) Gene set enrichment analysis (GSEA) of the Tmprss11b-high versus low LUSC spatial data with normalized enrichment scores (NES), false discovery rate (FDR) and p values for the indicated immune cell gene signatures. The nominal P and FDR values were obtained from the “GSEA Preranked” tool (from Broad Institute) using a weighted scoring scheme. Gene sets were evaluated based on the default normalized enrichment score method, and statistical significance was determined by bootstrapping with 1000 permutations. (D) Bar graph representing top M2-like/TAM genes from the differential gene expression (DEG) analysis of Tmprss11b-high versus low LUSC spatial transcriptomics data. Differential gene expression was calculated using Seurat’s FindAllMarkers function, and direct comparisons between two classes and corresponding P values were obtained using Seurat’s FindMarkers function with the Wilcoxon rank-sum test with Bonferroni P value correction. Source data are available online for this figure