A method for identifying and classifying tertiary lymphatic structures and its application in cancer treatment

Abstract

The invention belongs to the field of biomedical technology, and specifically discloses a method for identifying and classifying a tertiary lymphatic structure and an application thereof in cancer treatment, a tertiary lymphatic structure (TLS) identification method, comprising sample collection, sample spatial transcriptome sequencing, identifying TLS based on spatial transcriptome sequencing data, and identifying TLS based on spatial transcriptome sequencing data, comprising S1: obtaining an expression spectrum matrix retaining spatial position information of each sample after sequencing analysis, and analyzing with a resolution of Bin50; S2: marking a TLS region to be identified on the sample; S3: extracting a characteristic gene set related to TLS: using the AddModuleScore() function of the Seurat toolkit to calculate two different TLS characteristic gene sets, the characteristic gene sets being TLS-50 and 12-chemokine characteristic gene sets; S4: determining a spatial co-expression module related to TLS: using the Hotspot toolkit to identify a region with a TLS characteristic spatial co-expression pattern, and the like. The identification method of the invention can fully display the spatial characteristics of TLS in hepatocellular carcinoma.

Description

Technical Field

[0001] The invention belongs to the field of biomedical technology, and in particular relates to a method for identifying and classifying tertiary lymphatic structures and its application in cancer treatment. Background Art

[0002] Hepatocellular carcinoma (HCC) is the third leading cause of cancer-related death worldwide. According to the 2020 global cancer data from GLOBOCAN, the number of new HCC cases and deaths in China accounted for 45.3% and 47.1% of the world's total, respectively. In recent years, cancer immunotherapy, represented by immune checkpoint inhibitors, has achieved revolutionary breakthroughs in anti-tumor therapy, suggesting that interventions based on the tumor immune microenvironment have broad application prospects. However, the response rate of immune checkpoint inhibitors in HCC is only 20%, and most HCC patients still fail to benefit from this therapy. Therefore, it is crucial to predict the efficacy of immunotherapy in order to accurately screen the population that will benefit. Currently, the commonly used molecular markers for predicting the efficacy of immunotherapy include the expression level of tumor cell programmed death factor ligand-1 (PD-L1), tumor mutation load, and microsatellite instability, which have been widely used as molecular markers for screening the population that will benefit in tumors such as lung cancer and melanoma. However, in the relevant research results on liver cancer, there is no significant correlation between the expression level of PD-L1 and the efficacy of anti-PD-1 antibody treatment. The incidence of high tumor mutation load in liver cancer is less than 1%, and the only available data suggest that high tumor mutation load cannot predict the benefit of immunotherapy in liver cancer. Similarly, microsatellite instability is also extremely rare in liver cancer and is not valuable as a molecular marker for predicting efficacy. Currently, there is still a lack of molecular markers that can effectively predict immunotherapy for liver cancer.

[0003] Tertiary lymphatic structure (TLS) can be divided into two developmental stages: mature and immature, according to its pathological tissue structure. Compared with immature TLS with only sparse T and B lymphocyte aggregation, mature TLS marked by the appearance of follicular dendritic cells and germinal centers have been widely reported to be associated with better prognosis and immunotherapy efficacy. On the contrary, immature TLS only provides weaker protection and may even be associated with tumor progression, which may be attributed to the lack of germinal centers, which cannot support the formation of complete B cell function, or the expression of immunosuppressive molecules in immature TLS, and the lower activity of T cells in it, which ultimately leads to weaker or even tumor-promoting immune responses. These data all indicate that mature TLS has great potential as a marker for predicting the efficacy of immunotherapy for liver cancer. However, under the current commonly used pathological staining judgment, mature TLS accounts for only 9%, and the remaining TLS are classified as immature TLS, which limits the application of the former as a molecular marker for predicting the efficacy of immunotherapy for liver cancer. In addition, since TLS in tumors are short-lived, dynamically maturing and evolving immune structures, it is not reasonable to lump all immature TLS together, and it is difficult to distinguish which immature TLS are pre-mature TLS that will eventually enter the mature stage using only the current staining markers, and which immature TLS remain in the immature stage due to the lack of certain conditions or restrictions by certain factors. Therefore, further distinguishing the subtypes of immature TLS and finding key markers are of great significance for predicting the potential benefit population of immunotherapy and expanding the proportion of people who benefit from immunotherapy.

[0004] To gain a deeper understanding of the role of TLS, several spatial analysis methods have recently been used to characterize TLS in situ, such as CODEX, 10X Visium, and NanoString. However, these existing technologies have some limitations, such as limited analytical breadth, poor spatial resolution, or insufficient field of view, which restricts the comprehensive characterization of TLS and its microenvironment from multiple perspectives. Summary of the invention

[0005] An object of the present invention is to provide a tertiary lymphatic structure (TLS) identification method to solve the problems of limited analysis breadth, poor spatial resolution, or insufficient field of view under existing identification methods.

[0006] Specifically, the present invention provides the following solutions:

[0007] A method for identifying a tertiary lymphatic structure (TLS), comprising sample collection, sample spatial transcriptome sequencing, identifying TLS based on spatial transcriptome sequencing data, and identifying TLS based on spatial transcriptome sequencing data, comprising the following steps:

[0008] S1: After sequencing analysis, the expression matrix retaining the spatial position information of each sample is obtained, and the resolution of Bin50 is used.

[0009] Analyze the rate;

[0010] S2: Mark the TLS area to be identified on the sample;

[0011] S3: Extracting characteristic gene sets related to TLS: using the AddModuleScore() function of the Seurat toolkit to calculate two different TLS characteristic gene sets, the characteristic gene sets being TLS-50 and 12-chemokine characteristic gene sets;

[0012] S4: Identify spatial co-expression modules associated with TLS: Use the Hotspot toolkit to identify spatial co-expression modules with TLS characteristics.

[0013] regions of expression patterns;

[0014] S5: Identification of immune cell clusters: NormalizeData(), ScaleData(), RunPCA(), FindNeighbors(), FindClusters(), and RunUMAP() functions of the Seurat toolkit and SCTransform() function of the sctransform toolkit were used to reduce the dimension and cluster the spatial transcriptome data to obtain cluster subgroups. The FindAllMarkers() function was then used to calculate the characteristic genes of each cluster subgroup. Based on the characteristic genes, the immune cell clusters were identified.

[0015] immune cell clusters with immune cell characteristics;

[0016] 8: TLS region determination: using the OPTICS algorithm, the immune cell cluster region with high immune cell density is identified in the previously found immune cell clusters, which is the TLS region;

[0017] Under the above recognition method, preferably, the TLS area in step 8 can be further optimized by fine-tuning the OPTICS parameters to maximize TLS recognition, and using the closing operation in OpenCV to fill the small holes in the TLS area, and based on the cell density differences observed in the staining image, applying morphological operations to adjust the TLS boundaries, and further optimizing the TLS area dominated by non-immune cells.

[0018] Another object of the present invention is to provide a classification method for tertiary lymphatic structures for further subclassification of TLS and to deepen the understanding of TLS.

[0019] The present invention provides a method for classifying tertiary lymphatic structures, comprising the following steps:

[0020] (1) Differentiate between mature and immature TLS;

[0021] (2) transcriptome data of all naive B and germinal center B cells in the identified mature TLS;

[0022] (3) construct a TLS maturation trajectory with naive B cells in mature TLS as the starting point and germinal center B cells as the end point;

[0023] (4) All naive B cells and germinal center B cells in the immature TLS to be identified are included and mapped to the TLS maturation trajectory. A regression equation is constructed to predict the development trend of TLS based on the slope of the linear regression. The slope fits the transcriptional similarity of naive B cells and germinal center B cells along the maturation trajectory in each immature TLS. According to the trend of the slope of the linear fitting equation of each immature TLS to be identified, the TLS in the positive slope group is classified as TLS tending towards maturity, and the TLS in the negative slope group is classified as TLS deviating from maturity.

[0024] Furthermore, in step (1), TLS in which the proportion of CR2-positive B cells is not less than 5% of all cells in the TLS are defined as mature TLS, and for the remaining TLS, TLS with less than 5 naive B cells and germinal center B cells are filtered out, and those with more than 5 naive B cells and GC-B cells are classified as immature TLS.

[0025] Furthermore, the method for constructing the TLS maturity trajectory in step (3) is:

[0026] S1: After the transcriptome data were standardized and normalized, the genes related to the mature TLS co-expression module were used to perform PCA analysis on TLS by the Hotspot toolkit. The first 50 principal components were obtained from the PCA analysis for subsequent batch correction by the Harmony toolkit.

[0027] S2: Harmony's coordinates were then input into the SPRING toolkit to construct a k-nearest neighbor (kNN) network, and the Monocle3 toolkit was subsequently used to perform a two-dimensional visualization of the trajectory of naive B to germinal center B cells. Based on the coordinates of the kNN network, the learn_graph() function in the Monocle3 package was used to represent the cells on a pseudo-time trajectory;

[0028] S3: The order_cells() function in the Monocle3 package was used to assign a pseudo-time value to each cell along the trajectory from naive B to germinal center B cells. Finally, the slingshot() function in the Slingshot toolkit was applied to fit a smooth trajectory, which is the TLS maturation trajectory.

[0029] Furthermore, in step (4), the regression equation is constructed as follows:

[0030] S1: The TLS maturation trajectory is evenly divided into several bins. In each bin, the average expression of all cells is taken to calculate the gene expression profile and create a pseudo batch of cells;

[0031] S2: Then, the gene expression profiles associated with mature TLS were extracted to represent the reference cell state at each pseudo-time point in the pseudo-batch data and referred to as “reference cells”;

[0032] S3: Calculate the Spearman rank correlation between each B cell to be identified from the immature TLS and each reference cell, and align them to the most appropriate reference cell based on the maximum value of the correlation between the B cells to be identified and the reference cells;

[0033] S4: After comparing all the B cells to be identified with the reference cells, the least squares method was applied to fit the regression line based on the pseudo time points and correlation scores: the slope of the regression line represents the overall trend along the TLS maturation trajectory. Based on the linear regression results, the Mfuzz toolkit was used to divide immature TLS into two different groups. According to the positive or negative slope of the linear regression results, TLS with a positive slope were classified as TLS tending towards maturity, while those with a negative slope were classified as TLS deviating from maturity. The B cells to be identified were naive B cells or germinal center B cells.

[0034] As an embodiment, the present invention also provides a computer program, including a computer program / instruction, which implements any of the TLS identification or classification methods when executed by a processor.

[0035] As another embodiment, the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium is used to store the computer program.

[0036] As other broader embodiments, the present invention provides a computer system / device / apparatus, comprising a memory, a processor, and a computer program stored in the memory, wherein the computer program is the above-mentioned computer program.

[0037] Based on the sub-classified TLS obtained under the TLS classification method of the present invention, the present invention proposes some implementation schemes. Specifically, the present invention provides any of the following applications of the TLS tending to mature:

[0038] a1) Prepare products for predicting the risk assessment of cancer patients after ICB treatment;

[0039] a2) preparing products for stratification of tumor patients after ICB treatment;

[0040] a3) preparing products for enrichment of tumor patients after ICB treatment;

[0041] a4) preparing a product for predicting the survival rate of tumor patients after ICB treatment;

[0042] a5) Construct a model / system / device / computer program product for risk assessment after ICB treatment of tumors;

[0043] a6) Construct a model / system / device / computer program product for predicting the survival rate of cancer patients after ICB treatment

[0044] Preferably, the tumor is liver cancer; preferably, the ICB treatment is anti-PD-1 treatment.

[0045] Furthermore, the maturation-oriented TLS is an immature TLS that diffusely expresses CXCL9.

[0046] The technical effects achieved by the present invention are:

[0047] In this study, we developed a new strategy to comprehensively map TLS in situ and analyze its microenvironment through single-cell resolution large-field spatial transcriptomics Stereo-seq, which comprehensively demonstrated the spatial characteristics of TLS in hepatocellular carcinoma (HCCs). We proposed a new classification strategy to divide typical immature TLS into two groups, namely, those tending to mature and those deviating from mature, according to their developmental trajectory relative to the TLS maturation process, revealing the unique characteristics of immature TLS in anti-tumor immunity. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 In situ identification and transcriptomic characterization of tertiary lymph node structures (TLSs) in patients with hepatocellular carcinoma (HCC)

[0049] (A) Schematic overview of the experimental design and data acquisition process for TLS and their tissue environment. The study included 20 samples from 10 patients as a treatment-naive cohort. In addition, 5 samples from 4 patients who received anti-PD-1 treatment were included. For each sample, fresh frozen tissue blocks were serially sectioned and stained with hematoxylin and eosin (H&E). Serial sectioning was performed until a tissue section containing the maximum number of TLS determined by a pathologist was selected. Subsequent tissue sections were used for spatial transcriptomics (ST) analysis using Stereo-seq. Adjacent tissue blocks were used for single-cell RNA sequencing (scRNA-seq) and single-cell ATAC sequencing (scATAC-seq). The integrated data from adjacent H&E staining images, ST, scRNA-seq, and scATAC-seq help to identify and characterize TLS and their tissue environment in situ.

[0050] (B) Representative example illustrating the main steps of in situ TLS identification and cell type annotation. First, pathologists annotated TLS in H&E-stained images. Second, ST data obtained from whole tissue sections were subjected to unsupervised clustering to segment tissue regions. Third, TLS gene set scores were calculated for whole tissue sections using ST data. In situ identification of TLS was achieved by combining the segmented tissue regions with the TLS gene set scores. The identified TLS were validated by comparing them with the annotated TLS based on H&E-stained images. Finally, spatial cell type annotation of whole tissue sections was performed using scRNA-seq data. For better visualization, only annotated cells present in TLS are shown. Dark lines on the H&E-stained images indicate TLS annotated by pathologists. The legend at the bottom is labeled with a color code indicating the cell type. Scale bar, 1 mm.

[0051] Figure 2 Overview of spatial transcriptome data of representative tissue sections (A) Each H&E-stained image corresponds to the spatial transcriptome data of its adjacent sections. The sample name is marked at the top of each image. In the sample name, "T" represents tumor tissue, "B" represents marginal (border) tissue, and "P" represents non-tumor liver (para-cancer) tissue. Scale bar, 1 mm.

[0052] (BD) The spatial distribution of total unique molecular identifier (UMI) counts (B), the number of captured genes (C), and the unsupervised clustering results (D) were visualized in situ in representative tissue sections. The spatial resolution of each virtual detection point was set to 50 grids, representing 50x 50 points aggregated together, equivalent to a square size of 25 μm × 25 μm.

[0053] (E) Visualization of H&E staining images (left), spatial clustering results (middle), and regional annotations (right) of a representative sample HCC02-B at the tumor margin. “NM” indicates negative margin, “TC” indicates tumor capsule, and “PM” indicates positive margin.

[0054] (F) Visualization of spatial regions of a representative sample HCC02-B tumor margin using unsupervised cluster analysis of ST data (upper row) and marker genes representing different spatial regions (lower row). Different spatial regions were defined based on the expression of normal liver tissue marker CYP2E1, hepatocellular carcinoma tumor cell marker GPC3, and tumor-associated capsule and stromal marker COL1A1. The spatial expression pattern of MS4A1 highlights the lymphoid aggregation center, while CD3E is expressed in the peripheral area.

[0055] Figure 3 Diagram showing the process of identifying tertiary lymphatic structures (TLSs)

[0056] (A) Schematic diagram of the TLS identification process: using H&E staining images as prior knowledge, the parameter settings in the unsupervised clustering analysis are optimized until an immune cell population characterized by the immune marker gene PTPRC is identified.

[0057] Then, potential immune aggregates were co-validated by a predetermined TLS-related gene set and a TLS-related spatial co-expression module. Immune aggregates were further confirmed by Optics3, using a minimum area threshold (default 20 grids) to filter sparse immune signals. Finally, TLS regions were refined by morphological operations. Scale bar, 1 mm.

[0058] (BE) In situ visualization of the spatial patterns of TLS-associated 50 gene sets (B), 12 chemokine gene sets (C), spatial co-expression modules (D), and identified TLSs (E) in five representative tissue sections.

[0059] Figure 4 Comparison of the spatial expression of TLS characteristic genes and the spatial distribution patterns of cells in Visium and Stereo-seq data.

[0060] Figure 5 (A) Annotation diagram of cell subtypes in TLS; (BF) Characteristic gene diagram of each cell subtype.

[0061] Figure 6 Cell annotations and corresponding HE staining images inside and around TLS.

[0062] Figure 7 Classification and characterization of developmental states based on TLS spatial transcriptome

[0063] (A) Schematic illustration of the classic TLS maturity classification strategy (left) and the new approach developed in this study to characterize the developmental heterogeneity of TLS using spatial transcriptome repertoires with cellular resolution (right). High endothelial venules (HEVs), separated B cell and T cell zones, and high expression levels of CD21 are the main features that distinguish mature TLS from other lymphoid aggregates. However, TLS have not yet formed fully functional structures, lack key features of mature TLS, and may have different developmental states. Some of these TLS may have the potential to mature and develop along a maturational trajectory toward mature TLS, while others may not, and their trajectories deviate from the maturation process.

[0064] (B) Construction from The differentiation trajectory of B cells to 41 germinal center (GC)-B cells defined by high transcription levels of CR2 (the gene encoding CD21). Upper left: The smoothed curve plot shows the differentiation trajectory along the line from Pseudo-temporal speed of B to GC-B cell differentiation. Grey shading indicates 95% confidence intervals for the fitted model. Bottom left: Heat map showing the time course of B to GC-B cell differentiation along Expression profiles of 1,314 differentially expressed genes (DEGs) along the differentiation pseudotime of B to GC-B cells. Right: Smoothed curves and 95% confidence intervals show the changes in representative gene expression profiles along the differentiation pseudotime. Vertical bars with light red indicate transition stages identified by pseudotime speed (upper left), in which cells rapidly switch to the functional state of GC-B cells.

[0065] (C) Schematic diagram illustrating the use of ST data (e.g. Figure 1 B) Classification of the developmental status of the identified TLS. and GC-B cells (left) were aligned to the differentiation trajectory of mature TLS as described in (B) to determine their transcriptional similarity with mature TLS in pseudotime (middle). Then, a linear regression model was used to fit the transcriptional similarity in pseudotime. Based on the slope of the regression line, TLS were divided into three categories (right): 1) mature TLS, whose CR2 transcription level was high and whose transcriptional similarity in differentiation pseudotime continued to increase; 2) maturation-oriented TLS, whose CR2 transcription level was low but whose transcriptional similarity in differentiation pseudotime steadily increased, consistent with the differentiation process; and 3) deviated from maturation-oriented TLS, whose transcriptional similarity in differentiation pseudotime showed a regular decrease, deviating from the differentiation process. In the middle figure, the gray line connecting the query cell and the reference cell (Ref) along the differentiation trajectory represents the transcriptional similarity, and the longer the line, the lower the transcriptional similarity.

[0066] (D) Heatmaps showing global alignments to quantify the similarity of global expression dynamics along two developmental trajectories for mature and maturation-oriented TLS (left), or for mature and deviated TLS (right). The solid purple line in each heatmap represents the best alignment between matching cells in the two trajectories while preserving pseudo-temporal ordering. Comparison of developmental trajectories derived from mature and maturation-oriented TLS yields an expected approximately diagonal alignment corresponding to the transcriptional identity of maturation-oriented TLS conforming to the differentiation trajectory of mature TLS. In contrast, the developmental trajectory of deviated TLS shows a diagonal alignment that deviates from the consistency of mature TLS, confirming that deviated TLS have an aberrant regulatory program during GC-B cell development. The color bar represents the scaled similarity of pairwise comparisons between cells in the two trajectories.

[0067] (E) Representative examples showing mature TLS (top), TLS approaching maturity (middle), and TLS deviating from maturity (bottom), including their H&E staining, spatial expression patterns of marker genes, and spatial cellular organization with cell subtype annotations. Dark lines on the H&E staining images indicate TLS identified by ST data analysis. Scale bar is 1 mm.

[0068] (F) Bubble plots showing the average transcript levels of marker genes in each mature, maturation-approaching, and deviated TLS (left), the enrichment scores of representative pathways (middle), and the spatial interaction strengths of representative ligand-receptor pairs (right). The size of the bubble represents the number of cells in the TLS, indicating the relative size of the TLS in the tissue section.

[0069] (G) Tumor region of patient HCC02 showing H&E staining (left), different TLS types superimposed with H&E staining (middle), and the spatial expression pattern of CR2 in TLS (right). The dotted lines divide the tumor region into three parts according to the composition of different types of TLS, thus highlighting the diversity of TLS developmental states within the tumor.

[0070] Figure 8 Characteristic profile of active adaptive immune responses in mature TLSs

[0071] (A) Spatial expression pattern of CR2 in 41 mature TLSs.

[0072] (B) Heat map showing the enrichment of mature TLSs, immature TLSs, and cell types outside of TLSs using ratio analysis (OR). Immature TLSs refer to all TLSs except mature TLSs. OR>1.5 indicates enrichment of the cell type in the corresponding spatial region. *OR>1.5, OR>2, ***OR>3.

[0073] (C) Box plots showing the differences in the proportions of GC-B (left) and naive B cells (right) between mature and immature TLSs. ****p<0.0001, using Wilcoxon rank sum test.

[0074] (D) Box plots showing the significant differential expression levels of representative marker genes between mature and immature TLSs. ****p<0.0001, **p<0.01, using Wilcoxon rank sum test.

[0075] (E) Immune-related GO pathway enrichment ranking of T cell upregulated genes in mature TLSs.

[0076] (F) Differential expression levels of selected genes involved in T cell cytotoxicity and memory function in mature TLSs and the outer regions of TLSs. ****p<0.0001, using Wilcoxon rank sum test.

[0077] (G) Line graphs depicting the average score of effector / memory and co-stimulatory gene sets for T cells as a function of spatial distance from the outside of TLSs to the border of mature TLSs (left). Representative examples show the score for the area surrounding mature TLSs (right). Scores were calculated within 20 grids (~450 μm) from the border of mature TLS. The dark lines indicate the borders of TLSs.

[0078] (H) Box plots show that the ratio of B cell MKI67-positive virtual checkpoints in mature TLSs was significantly higher than that in the area outside TLSs. *p < 0.05, using Wilcoxon rank sum test.

[0079] (I) Volcano plot showing differentially expressed genes (DEGs) in mature TLSs and B cells in the external regions of corresponding TLSs. Representative genes are highlighted in the figure.

[0080] (J) Immune-related GO pathway enrichment ranking of B cell upregulated genes in mature TLSs.

[0081] Fig. 9 Characteristic diagrams of three different developmental states of TLSs

[0082] (A) UMAP visualization showing the distribution of GC-B and naive B cells (left) and the pseudo-temporal trajectory from naive B cells to GC-B cells (right).

[0083] (B) Each TLS was assigned to a different type by cluster analysis, and dynamic time warping was used to determine their transcriptional similarity with mature TLS along pseudotime based on the alignment of the differentiation trajectory of naive B and GC-B cells in TLS with that of mature TLS. A linear regression model was then used to fit the transcriptional similarity on pseudotime. Based on the slope of the regression line, all TLS were classified into three categories (mature TLS, TLS tending to mature, and TLS deviating from mature). Each gray line regressed the transcriptional similarity of naive B and GC-B cells in individual TLS. The x-axis represents the pseudotime of the developmental trajectory of GC-B cells from naive B cells, while the y-axis represents the transcriptional similarity between naive B and GC-B cells and CR2-positive TLS (representing mature TLS) in the target TLS. The colored lines in each figure represent the model-fitted transcriptional similarity trajectory based on the average similarity value on pseudotime.

[0084] (C) The proportions of mature, maturation-approaching, and deviating from maturity TLS in all samples.

[0085] (D) The proportion of mature, maturation-approaching, and deviating TLS in each sample. The number of TLS in the corresponding sample is shown on the top of the bar graph.

[0086] (E) Average expression levels of chemokines in the three TLS types.

[0087] (F) OR analysis showing the enrichment of cell types in the three TLS types. OR>1.5 indicates enrichment of the cell type. *OR>1.5 and **OR>2.

[0088] (G) Comparison of the cell proportions of B cells (upper) and T cells (lower) in the three TLS types. ****p<0.0001, **p<0.01, ns indicates non-significant results of the Wilcoxon rank sum test.

[0089] (H) Comparison of the spatial dependence of naive B and GC-B cells in the three TLS types using Moran's I as an indicator. ****p<0.0001, ns indicates non-significant results of the Wilcoxon rank sum test.

[0090] (I) Representative examples showing the spatial organization of B cells and T cells in three TLS types.

[0091] (J) Heatmap showing neighborhood enrichment scores for mature, maturation-oriented, and deviated TLS. The score represents the average number of target cells captured within a two-cell-width region around the central cell. For each cell type pair listed at the bottom of the heatmap, the cell type before the line represents the central cell, and the cell type after the line represents the target cell. The score is the standardized score in each column.

[0092] (K) Comparison of neighborhood enrichment scores for B and T cell pairs (left) and T and B cell pairs (right) in mature, maturation-prone, and maturation-departing TLS. Scores were calculated as described in (J). p < 0.0001, using the Wilcoxon rank sum test.

[0093] (L) Heatmap showing the enrichment of cell subtypes in the three TLS types. The color bar at the top indicates the TLS type as in c. OR>1.5, OR>2, and OR>3.

[0094] (M) Heatmap showing clustering of four feature types in mature TLS (top row), TLS approaching maturity (middle row), and TLS deviating from maturity (bottom row). Gene expression, i.e., average expression of individual genes; ligand-receptor, i.e., strength of ligand-receptor pairs; co-occurrence of cell subtypes, i.e., probability of each cell type surrounding the target cell type; functional status, gene set scores for representative pathways. The color in each image represents the average normalized signal in the TLS category. Key features associated with TLS maturation are highlighted. The color bar on the left represents the TLS type, as shown in (C).

[0095] Fig.10 Mature TLS responds to immune checkpoint therapy

[0096] (A) Circular Sankey diagram showing the different proportions of mature, maturation-oriented, and deviating from maturation TLS between the pre-treatment cohort (Pre, inner circle) and anti-PD-1 treatment cohort (Post, outer circle) in HCC patients.

[0097] (B) Box plots comparing the proportions of B cells (left) and T cells (right) in different TLS types between the anti-PD-1 treatment cohort and the pre-treatment cohort. **p<0.01, ns indicates not significant in the Wilcoxon rank sum test.

[0098] (C) Functional pathways significantly enriched in T cell genes upregulated in maturing TLS in the anti-PD-1 treated cohort compared with the pre-treatment cohort.

[0099] (D) Violin plots show the signature scores of specific gene sets for various T cell subtypes in different TLS types in the pre-treatment cohort (Pre) and anti-PD-1 treatment cohort (Post). ***p<0.001, *p<0.05 by Wilcoxon rank sum test.

[0100] (E) Heat map showing the average transcript levels of genes associated with T cell function in different TLS types in the pre-treatment cohort (Pre) and anti-PD-1 treatment cohort (Post).

[0101] (F) Circular plots showing the relationship between various cell subtypes and different TLS types in the pre-treatment cohort (Pre) and anti-PD-1 treatment cohort (Post). Strength of spatial cell interactions between B cells. The width and transparency of the lines, and the size of the nodes indicate the strength of the interaction between two cell populations.

[0102] (G) Heatmap showing gene expression of representative genes in Tfh cells of different TLS types in the pre-treatment cohort (Pre) and anti-PD-1 treatment cohort (Post) (left), spatial interaction strength of ligand-receptor pairs (middle), and Gene expression of representative genes in B cells (right). The color bar below each image represents the mean normalized signal.

[0103] (H) Spatial visualization of signals related to GC-B cell development in different TLS types in the pre-treatment cohort (Pre) and anti-PD-1 treatment cohort (Post). The top two rows show the transcription levels of CD40 and CD40LG in representative TLS, respectively. The third row shows the spatial interaction strength of the ligand-receptor pair CD40LG-CD40 in the corresponding TLS. The bottom row shows Transcript levels of BCL6 in B cells. DETAILED DESCRIPTION

[0104] The following will clearly and completely describe the concept of the present application and the technical effects produced in combination with the embodiments, so as to fully understand the purpose, features and effects of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of them. Based on the embodiments of the present application, other embodiments obtained by those skilled in the art without creative work are all within the scope of protection of the present application.

[0105] For commodities purchased in the test method, if the specific conditions are not specified, the test shall be carried out under conventional conditions or the conditions recommended by the manufacturer. If the manufacturer of the reagents or instruments used is not specified, conventional products can be purchased from the market.

[0106] Unless otherwise defined herein, scientific and technical terms used in conjunction with the present disclosure shall have the meanings commonly understood by those of ordinary skill in the art. Exemplary methods and materials are described below, but methods and materials similar or equivalent to those described herein can also be used in the practice and testing of the present disclosure.

[0107] As used herein, the following words and terms shall have the meanings indicated:

[0108] TLS-50 gene set: English name: TLS-50signature. Reference: Rui W, Wenbo G, XinyaoQ, et al. Comprehensive analysis of spatial architecture in primary liver cancer. [J]. Science advances, 2021, 7(51): eabg3750-eabg3750. This study identified the TLS region in the spatial transcriptome data by integrating 10XVisium spatial transcriptome data with H&E staining, and used the top 50 genes (TLS-50) with high expression in this region as the characteristic genes of TLS.

[0109] 12-chemokine gene subset: English name: 12chemokine gene subset. Reference: Coppola D, Nebozhyn M, Khalil F, et al. Unique Ectopic Lymph Node-Like Structures Present in Human Primary Colorectal Carcinoma Are Identified by Immune Gene Array Profiling [J]. The American Journal of Pathology, 2011, 179 (1): 37-45. This study used the gene expression data of 326 colorectal cancer patients to screen out a gene set containing 12 chemokine genes, and found that TLS was significantly correlated with the 12-chemokine gene expression characteristics, suggesting that if the patient's tumor tissue highly expresses these 12 chemokines, TLS is more likely to exist in the tumor tissue.

[0110] Immature B cells ( B Cell): A type of B lymphocyte that has not been immune stimulated and has not yet undergone antigen activation. They are usually found in lymphoid organs (such as lymph nodes, spleen) and peripheral blood. Naive B cells express B cell receptors (BCR), but these receptors have not yet been exposed to the corresponding antigens. When naive B cells bind to their specific antigens, they will be activated and begin to differentiate, transforming into germinal center B cells or memory B cells, thereby participating in the humoral immune response. Typically, naive B cells have lower expression levels because they have not yet experienced antigen selection and stimulation, and therefore lack memory for specific antigens. Naive B cells usually highly express genes such as IGHD, TCL1A, CD22, and CD72.

[0111] Germinal center B cells: A specific type of B lymphocyte in lymph nodes, spleen, and other lymphoid organs. They usually appear in the germinal center of the lymph node. The germinal center is a specific area in the lymph node reaction that promotes the proliferation, differentiation, and affinity maturation of B cells in immune responses. Germinal center B cells have highly active metabolism, DNA recombination, and highly expressed surface molecules, such as highly expressed AICDA, MKI67, BCL6, RGS13 and other genes. They play a key role in antibody affinity maturation and the generation of memory B cells.

[0112] Example 1 Spatial transcriptome sequencing and TLS identification based on liver cancer specimens

[0113] 1.1. Collection of liver cancer pathological specimens

[0114] Inclusion criteria: Meet the diagnostic criteria for hepatocellular carcinoma in the "Guidelines for the Diagnosis and Treatment of Primary Liver Cancer (2019 Edition)", and exclusion criteria: Combined with cholangiocarcinoma or other malignant tumors. Surgical resection specimens of the above patients were collected, including tumor tissue, junctional tissue, and paracancerous liver tissue. The junctional tissue is the tissue with a width of 0.5 cm from the tumor invasion edge to the inside and outside. The paracancerous liver tissue is the liver tissue at least 2 cm away from the tumor invasion edge. When taking the sample, avoid taking ischemic necrotic tissue as much as possible. The fresh tissue samples were placed on ice within 30 minutes of ex vivo, trimmed into tissue blocks with a length × width of <10 mm × 10 mm, 3 mm ≤ height ≤ 8 mm, and then placed horizontally in a tissue embedding box filled with OCT embedding agent, and then the latter was placed on the surface of liquid nitrogen for quick freezing until the embedding agent hardened and turned white, and finally stored in a -80 ° C refrigerator. In addition, the target samples need to be sliced ​​and RNA quality assessed, requiring the tissue to be intact, the embedding direction to be clear, and the RNA integrity RIN ≥ 7. The sample collection process was reviewed by the hospital ethics committee and complied with international bioethical standards.

[0115] 1.2 Spatial transcriptome sequencing

[0116] Each tissue sample is sequenced using BGI Stereo-seq spatial transcriptome technology (each detection point has a diameter of 220nm, and the center distance between two detection points is 500nm). The specific process is as follows: slice the OCT-embedded sample, and use fluorescence imaging technology to microscopically observe the target tissue slices. The fluorescence imaging results can confirm the basic information such as the shape contour, tissue structure, and completeness of the target sample tissue. The staining effect requires that the cell resolution rate is ≥90%. The target slice is then attached to the surface of the sequencing chip, and the tissue attached to the chip is sliced, fixed, and infiltrated. The chip will capture the mRNA molecules of the target sample tissue cells and synthesize cDNA to construct a sequencing library, and finally the sequencing library is sequenced.

[0117] 1.3 TLS identification based on spatial transcriptome data

[0118] After quality control, alignment, and quantification of the original FASTQ sequence (for the specific implementation of this process, refer to https: / / github.com / BGIResearch / SAW), the expression spectrum matrix that retains the spatial position information of each sample is obtained, and downstream analysis is performed using a resolution of Bin50 (50×50 DNA nanospheres / Bin, 25 μm in length).

[0119] Then, TLS was identified in situ on the spatial transcriptome data. The specific process is as follows:

[0120] 1) Pathological identification: Pathologists marked TLS candidate regions on H&E stained sections. This manual annotation provided additional evidence for the subsequent confirmation of the presence of TLS in spatial transcriptome data;

[0121] 2) TLS-related signature gene sets: Two different TLS signature gene sets, namely TLS-50 and 12-chemokine signature gene sets, were calculated using the AddModuleScore() function of the Seurat toolkit (v4.0.5), which showed spatial expression patterns associated with TLS.

[0122] 3) Spatial co-expression modules associated with TLS: The Hotspot toolkit (v1.1) was used to identify regions with characteristic spatial co-expression patterns of TLS.

[0123] 4) Identification of immune cell clusters: Seurat toolkit (v4.0.5) NormalizeData(), ScaleData(), RunPCA(), FindNeighbors(), FindClusters() and RunUMAP() functions and SCTransform() function in sctransform toolkit (v0.3.2) were used to reduce the dimension and cluster the spatial transcriptome data to obtain cluster subgroups (cluster). Then, FindAllMarkers() function was used to calculate the characteristic genes of each cluster. Based on the characteristic genes, clusters with immune cell characteristics were identified and defined as "immune cell clusters". The OPTICS algorithm was used to identify immune cell clusters with high immune cell density in the previously found immune cell clusters.

[0124] 5) TLS region determination: The presence of immune cell clusters with high cell density is the TLS region, that is, most of the characteristic markers including gene signatures, co-expression modules and H&E staining support are present in these regions. The OPTICS parameters were fine-tuned to maximize TLS recognition while considering these criteria, and the closing operation in OpenCV (v4.6.0.66) was used to fill the small holes in the TLS region.

[0125] 6) Optimize the TLS area: Based on the differences in cell density observed in the staining images, morphological operations were applied to adjust the TLS boundaries and further optimize the TLS area dominated by non-immune cells.

[0126] 7) Filtering and analysis: TLS with an area less than 20 Bin50 are filtered out, and the remaining TLS are used for downstream TLS maturity analysis.

[0127] 1.4TLS recognition effect

[0128] We developed a systematic identification process to capture spatial transcriptomic data from tissue sections (1 cm × 1 cm area) containing multiple TLS and their corresponding TME using Stereo-seq ( Figure 1 A); To identify tissue sections containing the most TLS for subsequent spatial transcriptomic analysis, we performed serial sectioning and combined H&E staining ( Figure 2 A); To resolve tissue structure at cellular resolution and perform spatial cell type annotation, we converted the original spatial expression matrix of Stereo-seq into virtual checkpoints (pseudospots) with a size of 25μm×25μm squares (each virtual checkpoint contains 50×50 DNA nanoballs, which is approximately the size of 1 cell). Therefore, this conversion allows us to assign transcripts to each defined cell area, achieving a cellular resolution of 1,006 genes per cell area on average ( Figure 2 B and 2C).

[0129] The heterogeneity of TLS morphology, cellular composition, and functional status makes their identification difficult using traditional methods based on uniform markers or anatomical locations. To overcome these challenges, we developed a novel computational strategy for unsupervised TLS identification based on their spatial transcriptional properties ( Figure 1 B and Figure 3 A). This approach allows for a comprehensive and unbiased characterization of TLS that encompasses their complexity and captures their diverse features. We first performed unsupervised clustering of the spatial transcriptomics (ST) data for each tissue section, segmenting the tissue section into distinct spatial regions ( Figure 2 D), segmented tissue regions resemble histologically recognizable structures ( Figure 2E). Notably, we found a set of compartmentalized tissue regions scattered throughout the tissue sections that highly expressed genes associated with TLS at the transcriptional level (e.g., MS4A1, CD3E, and PTPRC) ( Figure 2 F and Figure 3 A). To validate these target tissue regions with TLS transcriptional signatures, we examined transcript levels of 50 TLS-associated genes and a set of 12 chemokine genes in whole tissue sections. Target tissue regions identified by unsupervised clustering exhibited TLS-associated signatures and were classified as TLS ( Figure 3 B and Figure 3 C). In addition, we applied Hotspot to further verify whether these TLS have similar spatial co-expression modules, which contain most of the TLS-related gene signatures ( Figure 3 D). To accurately determine the true boundaries of candidate TLS, we applied the Optics algorithm to identify clustered TLS signals and distinguish them from scattered signals that may be caused by sparsely infiltrating immune cells near the TLS ( Figure 3 E).

[0130] Finally, according to the above strategy, we identified 461 TLS of different morphology and size in 20 tissue sections of the untreated cohort. Furthermore, pathologists (who were not aware of the TLS identified by the above strategy in advance) manually identified TLS in the HE images corresponding to the above 20 spatial transcriptome sections. The process is as follows: observe the HE staining results under an ordinary optical microscope and identify the lymphocyte cluster structure in the tumor area (defined as lymphocyte clusters with vague and slightly sparse immune cell aggregation or obvious lymphocyte clusters (TLS cells are dense and TLS shape is round) with or without light-stained germinal center formation). The results showed that 80% of the TLS identified by the new TLS identification strategy we developed were also identified by pathologists. The other 20% of TLS that were not identified by pathologists in the first round of identification were confirmed in the second round of review. This is because these TLS have atypical morphology on HE and were ignored by pathologists in the initial evaluation.

[0131] In summary, we preliminarily verified the accuracy of TLS identification by using TLS characteristic genes found by unsupervised clustering, TLS characteristic gene set scores, TLS-related spatial co-expression modules, and the Optics algorithm. At the same time, experienced pathologists further confirmed the TLS identified by the above strategies, laying the foundation for subsequent analysis.

[0132] Furthermore, we conducted follow-up analysis on the TLS identified by the above identification strategy and achieved the following results:

[0133] (1) For the first time, the cellular and molecular characteristics of TLS were deconstructed in situ at the single-cell level: The resolution of currently commonly used spatial transcriptome technologies, such as 10x Genomics' Visium, is insufficient. Each detection point is 55um in diameter, and this range may include several to dozens of cells. It is not single-cell resolution and is not suitable for densely packed cell aggregates such as TLS. However, our new computational strategy for unsupervised TLS identification at single-cell resolution, developed based on the nanometer-level precision Stereo-seq sequencing technology launched by BGI, is able to deconstruct TLS at the single-cell level, making it possible to resolve TLS-related key molecular targets at the single-cell level. Compared with Visium data, our TLS identification strategy more accurately restores the unique spatial distribution and expression patterns of cells and characteristic genes within TLS in Stereo-seq data ( Figure 4 ), which are harder to discern in Visium’s data.

[0134] (2) For the first time, TLS was deconstructed at the level of unbiased, whole-transcriptome data: Currently, technologies that can achieve single-cell resolution, such as CODEX, like Stereo-seq, can only detect about 50 proteins on a slice at the same time. The cell composition in TLS is complex, and detecting only 50 proteins cannot fully clarify the composition of cells in TLS and the functions of cells. However, by combining Stereo-seq technology with our new computational strategy for unsupervised TLS identification, we can accurately annotate spatial in situ cell types to the cell subclass level based on accurately identified TLS, and perform unbiased, whole-transcriptome functional analysis on each cell subclass in TLS ( Figure 5 ).

[0135] (3) For the first time, TLS and its surrounding microenvironment were analyzed in a large field of view: Nanostring's GeoMx DSP technology has a detection area of ​​only tens to hundreds of microns, while the diameter of TLS is at least 200um. Therefore, this type of technology cannot fully cover the information of TLS and its surrounding microenvironment. However, by combining Stereo-seq technology with our new computational strategy for unsupervised TLS identification, we can distinguish the microenvironment inside and around TLS through clear TLS boundaries, and retain the information of the microenvironment around TLS in a large field of view ( Figure 6 ).

[0136] Example 2 Maturity classification of TLS in liver cancer

[0137] 2.1. Distinguishing between mature and immature TLS in in situ spatial transcriptome data

[0138] TLS with a CR2-positive B cell ratio of no less than 5% of all cells in the TLS were defined as mature TLS. For the remaining TLS, naive B cells and TLS with less than 5 germinal center B cells were filtered out, and the remaining TLS after filtering were classified as immature TLS.

[0139] 2.2. Constructing a Pseudo-Timing Trajectory of Mature TLS

[0140] Taking naive B cells in mature TLS as the starting point and germinal center B cells as the end point, a TLS maturation trajectory can be constructed to simulate the process of TLS maturation at the transcriptome level. First, the transcriptome data of all naive B and germinal center B cells in mature TLS identified from 2.1 were included. After standardizing and normalizing the data, we performed principal component analysis (PCA) on 1,118 genes associated with TLS co-expression modules (identified by Hotspot toolkit (v1.1)). The first 50 principal components (PCs) obtained from the PCA analysis were used for subsequent batch correction by Harmony toolkit (v0.1.1).

[0141] The coordinates of Harmony were then input into the SPRING toolkit to construct a k-nearest neighbor (kNN) network, and the Monocle3 toolkit (v1.0.0) was subsequently used to visualize the trajectory of naive B to germinal center B cells in two dimensions. Based on the coordinates of the kNN network, we used the learn_graph() function in the Monocle3 package to represent the cells on a pseudo-time trajectory. We then used the order_cells() function in the Monocle3 package to assign a pseudo-time value to each cell along the trajectory from naive B to germinal center B cells. Finally, we applied the slingshot() function in the Slingshot toolkit (v2.6.0) to fit a smoothed trajectory, which is therefore called the reference trajectory during TLS maturation. The above analysis was completed under the R programming language (version 4.1.0).

[0142] 2.3. Further incorporate naive B and germinal center B cells in each TLS and map them to the TLS maturation trajectory. The next goal is to predict the development trend of TLS based on the slope of the linear regression, which simulates the transcriptional similarity of naive B and germinal center B cells along the maturation trajectory in each immature TLS. First, the reference TLS maturation trajectory is evenly divided into 100 segments. Within each segment, the average expression of all cells is taken to calculate the gene expression profile to create a pseudo-bulk of cells. Then, the 1,118 TLS-related gene expression profiles in 2.2 are extracted to represent the reference cell state at each pseudo-time point in the pseudo-bulk data and are referred to as "reference cells". Next, the Spearman rank correlation between each B cell to be identified from the TLS (naive B or germinal center B cell) and each reference cell is calculated. By aligning the B cell to be identified with the reference cell with the highest correlation, its most likely cell state is determined. This alignment process is used for all B cells to align them with the reference cells along the maturation trajectory. After all the B cells to be identified were aligned with the reference cells, the least squares method was applied to fit the regression line based on the pseudo time points and the correlation scores. The slope of the regression line represents the overall trend along the TLS maturation trajectory. Based on the linear fitting equation of each TLS (y = βx + α, each TLS has its own intercept α and slope β), we used the Mfuzz toolkit (v2.58.0) to cluster all immature TLS into two groups. According to the trend of the slope of the linear fitting equation of all TLS in each group, the TLS in the positive slope group was named as TLS tending to mature, and the TLS in the negative slope group was named as TLS deviating from mature. The above analysis was completed in the R programming language (version 4.1.0).

[0143] 2.4 Verify the accuracy of the TLS maturity classification system based on the new TLS identification strategy

[0144] The remarkable heterogeneity observed in TLS highlights the dynamic and multi-layered nature of their development. To elucidate the developmental state of TLS, we used single-cell resolution spatial transcriptomics to assess the developmental potential of TLS and determine whether they are moving toward or away from mature TLS maturation ( Figure 7 A).

[0145] We first extracted naive B cells and GC-B cells from mature TLS to construct developmental trajectories, because the formation of germinal centers is a key marker of TLS maturation. We identified 41 mature TLS based on their high expression levels of CR2 (encoding CD21). Figure 8A), CR2 plays an important role in germinal center reaction and formation. In addition to CR2, these mature TLS are enriched in activated naive B cells and GC-B cells, and highly express marker genes associated with TLS maturation, including CXCL13, FCER2, and AICDA ( Figure 8 B-8J). Next, we collected 2,155 naive B cells and 446 GC-B cells in mature TLS and constructed the developmental trajectory from naive B cells to GC-B cells, which represents the process of TLS maturation ( Fig. 9 A). Collectively, these results elucidate the molecular features of TLS maturation and provide valuable clues for characterizing other TLS types that lack key features of maturation.

[0146] Next, we constructed the developmental trajectory from naive B cells to GC-B cells, which represents the process of TLS maturation ( Fig. 9 A, right). By examining the speed of transcriptome changes during differentiation from naive B cells to GC-B cells, we discovered a unique transition phase characterized by slow transcriptome changes during the early stages of naive B cell development, followed by dramatic external stimulation and conversion to GC-B cells ( Figure 7 B).

[0147] Next, we developed a method to quantify TLS developmental propensity by mapping the transcriptional profiles of naive B cells and GC-B cells from each TLS to the developmental trajectory established above. By examining the linear regression of the transcriptional profile mapped to the normal developmental trajectory, we assessed whether the developmental propensity was toward or away from the established TLS maturation trajectory ( Figure 7 C). To validate the accuracy of our method, we calculated the developmental trends of 41 mature TLS defined by CR2 expression levels. As expected, all of these TLS showed convergence in the pseudo-time of GC-B cell differentiation, confirming the accuracy of our method ( Fig. 9 B). Next, we calculated the developmental trends of the remaining TLS (258 typical immature TLS) and divided them into two categories according to their consistency with the differentiation trajectory (see 2.3 for specific methods): “TLS tending to mature” and “TLS deviating from mature” ( Fig. 9 B). Based on these analyses, we classified TLS in HCC into 41 mature TLS, 136 TLS tending to mature, and 122 TLS deviating from maturity ( Fig. 9 C and Fig. 9 D).

[0148] To verify the accuracy of TLS classification, we used cellAlign to compare the developmental trajectories of cells and compared the expression dynamics between the developmental trajectories of naive B cells to GC-B cells in mature, maturation-oriented, and deviated TLS. The analysis results showed that the maturation-oriented and mature types showed significant correlation in developmental pseudo-time, and the simulated developmental trajectories were very similar ( Figure 7 D), verifying that the maturation-prone type shares similar developmental processes with the mature type and has the potential to mature. In contrast, the deviated-maturity type shows a developmental trajectory that gradually deviates from maturity and does not undergo a transitional stage from a naive B cell to a GC-B cell ( Figure 7 D) This finding confirms that the deviant-maturity type deviates from the mature type during development, indicating that it ultimately fails to achieve maturity.

[0149] In general, we divided TLS into three groups according to their developmental potential to achieve maturity: mature type, maturation-approaching type, and maturity-departing type.

[0150] 2.5 Constructing a dynamic development path to further verify the accuracy of TLS classification

[0151] To verify the reliability of TLS classification results and explore the dynamic development paths of different TLS groups, we compared the pseudo-temporal expression dynamics of naive B cells to GC-B cells differentiation in mature TLS, conforming TLS, and deviating TLS groups. This analysis was performed using cellAlign (v0.1.0) software.

[0152] First, we estimated the pseudotime of naive B cells and GC-B cells in conforming / deviant TLS in the same pseudotime coordinates as mature TLS. The expression matrices of naive B cells and GC-B cells were imported into Seurat to obtain a two-dimensional latent embedding based on the structure of mature TLS. In brief, we applied the "FindTransferAnchors" function to find anchors between cells of each group and treated cells of mature TLS. This set of anchors was then applied in the "MapQuery" function to project the cells onto a reference two-dimensional kNN structure built using mature TLS. Subsequently, the projected structure was imported into Slingshot to construct the developmental trajectory of each TLS group and estimate pseudotime for each cell.

[0153] The calculated pseudo-time and expression matrices of the three TLS groups were imported into cellAlign for comparison. The "interWeights" function was first executed to evenly place 200 artificial points along the pseudo-time path of each TLS group. This interpolation step involves estimating gene expression at each time point using all cells, with each cell assigned a weight according to a Gaussian distribution centered at the interpolated point. The interpolated expression matrix was then scaled using the "scaleInterpolate" function. The distance of the matching of conforming TLS and deviating TLS to the mature TLS throughout the trajectory was measured using the "calcDistMat" function, where a distance matrix was calculated between each pair of points in gene expression space. To convert the distance measure (d) to a similarity measure (s), we first normalized d to the range [0 1] using d_norm = d / max(d). The similarity measure is then given by s = 1-d_norm.

[0154] To visually demonstrate the differences between the developmental trajectories of different TLS groups, we simulated the developmental trajectories of the conforming / deviant TLS groups and compared them with the mature TLS group. Specifically, we first binarized the distance matrix generated by cellAlign, marking the points with distances less than 25% of the total distance as 1 and the rest as 0. Using this virtual matrix, we applied OpenCV's morphological transformation to extract the contours. We then used the digitization function "st.dd.digitize" in Spateo (v1.0.2) to find the relevant curve along pseudotime, which corresponds to the central curve in the contour.

[0155] 2.6 Systematic characterization of different TLS subclasses

[0156] We characterized the classified TLS in terms of gene expression, cellular composition, and pathway enrichment. As shown in representative examples, mature TLS exhibited a distinct spatial pattern of gene expression, with AICDA expression concentrated in the center of the TLS, accompanied by high expression of CR2 ( Figure 7 E). The spatial co-localization of CR2 and AICDA expression is consistent with the function of CR2 in promoting the differentiation of germinal center B cells into antibody-producing plasma cells. Maturation-type TLS showed the characteristics of overall high expression of CXCL9, and CXCL9-positive macrophages were also significantly enriched in this type of TLS. In addition, judging from the spatial pattern of CXCL9 expression, its expression is not limited to macrophages, but diffusely expressed in TLS. Furthermore, maturation-type TLS also contains abundant CXCL9-CXCR3 and other interactions related to myeloid and T cell chemotaxis, suggesting stronger cell recruitment, which indicates that the chemokine encoded by CXCL9 plays a key role in mediating the maturation of TLS ( Figure 7 E and Fig. 9E). In addition to the representative examples, mature and maturation-prone TLS have higher enrichment of characteristic genes, pathways, and spatially interacting receptor-ligand pairs associated with promoting TLS maturation, while TLS deviating from maturation lack these features ( Figure 7 F and Fig. 9 F-9M), these results reveal that TLS at different developmental states have unique spatial properties.

[0157] Example 3 Response of mature TLS to PD1

[0158] The present invention explores how ICB (immune checkpoint blockade) treatment affects the characteristics of TLS with different developmental states. First, our TLS classification was applied to classify 63 TLS in four patients receiving ICB treatment. Among these TLS, we identified 17 mature TLS, 27 TLS tending to be mature, and 19 TLS deviating from maturity. Interestingly, we observed a significant increase in the proportion of mature TLS in the anti-PD-1 treatment cohort compared with the pre-treatment cohort (17 / 63, 27.0% vs. 41 / 299, 13.7%; p = 0.015) ( Fig.10 A). In addition, the proportion of T cells in mature TLS increased significantly after ICB treatment ( Fig.10 B). Further analysis showed that compared with the pre-treatment cohort, T cells in the maturing TLS in the anti-PD-1 treatment cohort were enriched in pathways related to T cell activation and immune response activation ( Fig.10 C). In addition, specific T cell subtypes in maturing TLS in the anti-PD-1 treated cohort showed associations with T cell activation (for The scores of significantly increased gene sets related to T cells), effector memory function (precursor exhausted T cells, Tpex), and Tfh function (follicular helper T cells, Tfh) were Fig.10 D and 10E). These findings suggest that T cell activity in TLS that are moving toward maturity is significantly enhanced after ICB treatment. In contrast, TLS that are deviating from maturity have limited response to ICB treatment ( Fig.10 D and 10E).

[0159] To investigate how maturing TLS respond to anti-PD-1 therapy, we performed a comparative analysis of cellular interactions based on their spatial locations in TLS before and after ICB treatment. The spatial interactions between B cells were significantly enhanced ( Fig.10 F) It is worth noting that in the mature TLS, Tfh and Spatial ligand-receptor interactions of CD40L-CD40 between B cells were significantly increased in response to ICB treatment, consistent with upregulation of BCL6, a regulator of TLS maturation. Marker genes for B cell differentiation into GC-B cells ( Fig.10 G and 10H).

[0160] Our results suggest that maturational TLS show a robust response to ICB treatment.

Claims

1. A method for identifying tertiary lymphatic structures, i.e., TLS, comprising sample collection, sample spatial transcriptome sequencing, and identifying TLS based on spatial transcriptome sequencing data, characterized in that: Identification of TLS based on spatial transcriptome sequencing data includes the following steps: S1: After sequencing analysis, the expression matrix retaining the spatial position information of each sample is obtained and analyzed with a resolution of Bin50; S2: Mark the TLS area to be identified on the sample; S3: Extraction of characteristic gene sets related to TLS: Use the AddModuleScore() function of the Seurat toolkit to calculate two different TLS characteristic gene sets, which are TLS-50 and 12-chemokine characteristic gene sets; S4: Identify spatial co-expression modules associated with TLS: Use the Hotspot toolkit to identify regions with characteristic spatial co-expression patterns of TLS; S5: Identification of immune cell clusters: Use the NormalizeData(), ScaleData(), RunPCA(), FindNeighbors(), FindClusters(), and RunUMAP() functions of the Seurat toolkit and the SCTransform() function in the sctransform toolkit to perform dimensionality reduction and clustering on the spatial transcriptome data to obtain cluster subgroups. Then use the FindAllMarkers() function to calculate the characteristic genes of each cluster subgroup, and identify immune cell clusters with immune cell characteristics based on the characteristic genes. S6: TLS region determination: using the OPTICS algorithm, the immune cell cluster region with high immune cell density is identified in the previously found immune cell clusters, which is the TLS region.

2. A classification method for the three-level lymphatic structure according to the identification method of claim 1, characterized in that: The following steps are involved: (1) Distinguish between mature and immature TLS; (2) transcriptome data of all naive B cells and germinal center B cells in the identified mature TLS; (3) construct a TLS maturation trajectory with naive B cells in mature TLS as the starting point and germinal center B cells as the end point; (4) All naive B cells and germinal center B cells in the immature TLS to be identified are included and mapped to the TLS maturation trajectory. A regression equation is constructed to predict the development trend of TLS based on the slope of the linear regression. The slope fits the transcriptional similarity of naive B cells and germinal center B cells along the maturation trajectory in each immature TLS. According to the trend of the slope of the linear fitting equation of each immature TLS to be identified, the TLS in the positive slope group are classified as TLS tending towards maturity, and the TLS in the negative slope group are classified as TLS deviating from maturity.

3. The classification method according to claim 2, characterized in that: In step (1), TLS with a CR2-positive B cell ratio of not less than 5% of all cells in the TLS are defined as mature TLS. For the remaining TLS, TLS with less than 5 naive B cells and germinal center B cells are filtered out, and those with more than 5 naive B cells and germinal center B cells are classified as immature TLS.

4. The classification method according to claim 2, characterized in that: The construction method of TLS maturity trajectory in step (3) is: S1: After the transcriptome data were standardized and normalized, the genes related to the mature TLS co-expression module were used to perform PCA analysis on TLS by the Hotspot toolkit. The first 50 principal components were obtained from the PCA analysis for subsequent batch correction by the Harmony toolkit. S2: Harmony's coordinates were then input into the SPRING toolkit to construct a kNN network, and the Monocle3 toolkit was subsequently used to perform a two-dimensional visualization of the trajectory of naive B cells to germinal center B cells. Based on the coordinates of the kNN network, the learn_graph() function in the Monocle3 package was used to represent the cells on a pseudo-time trajectory; S3: The order_cells() function in the Monocle3 package was used to assign a pseudo-time value to each cell along the trajectory from naive B cells to germinal center B cells. Finally, the slingshot() function in the Slingshot toolkit was applied to fit a smooth trajectory, which is the TLS maturation trajectory.

5. The classification method according to claim 2, characterized in that: In step (4), the regression equation is constructed as follows: S1: The TLS maturation trajectory is evenly divided into several bins. In each bin, the average expression of all cells is taken to calculate the gene expression profile and create a pseudo batch of cells; S2: Then, the gene expression profiles associated with mature TLS were extracted to represent the reference cell state at each pseudo time point in the pseudo-batch data and referred to as reference cells; S3: Calculate the Spearman rank correlation between each B cell to be identified from the immature TLS and each reference cell, and align them to the most appropriate reference cell based on the maximum value of the correlation between the B cell to be identified and the reference cell; S4: After comparing all the B cells to be identified with the reference cells, the least squares method was applied to fit the regression line based on the pseudo time points and correlation scores. The slope of the regression line represented the overall trend along the TLS maturation trajectory. Based on the linear regression results, the Mfuzz toolkit was used to divide immature TLS into two different groups. According to the positive or negative slope of the linear regression results, TLS with a positive slope were classified as TLS tending towards maturity, while those with a negative slope were classified as TLS deviating from maturity. The B cells to be identified were naive B cells or germinal center B cells.

6. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store the computer program product of claim 6.

8. A computer system comprising a memory, a processor and a computer program stored in the memory, characterized in that: The computer program is the computer program product according to claim 6.

9. The application of the TLS tending to mature identified by the classification method described in any one of claims 2 to 5, characterized in that: The applications include: a1) Preparation of products for predicting the risk of cancer patients after ICB treatment; or a2) preparing products for stratification of cancer patients after ICB treatment; or a3) Preparation of products for use in enriched populations of cancer patients after ICB treatment; or a4) preparing products for predicting survival of cancer patients after ICB treatment; or a5) constructing a computer program product for risk assessment after ICB treatment of tumors; or a6) Construct a computer program product to predict the survival rate of cancer patients after ICB treatment.

10. The use according to claim 9, characterized in that: The tumor is liver cancer.

11. The use according to claim 9, characterized in that: The ICB treatment is anti-PD-1 treatment.

12. The use according to any one of claims 9 to 11, characterized in that: The maturation-oriented TLS is an immature TLS that diffusely expresses CXCL9.

Images