Spatial maps of hepatocellular carcinoma transcriptomes reveal spatial expression patterns in tumor immune microenvironment

Rationale: Hepatocellular carcinoma (HCC) is a highly heterogeneous and malignant disease with the complex immune microenvironment, which ultimately influence clinic outcomes of patients. However, the spatial expression patterns of diverse immune cells among tumor microenvironment remain to be further deciphered. Methods: Spatial transcriptomics sequencing (ST) was implemented on two portions of HCC specimens. Differentially expressed genes, cell cycle phases, epithelial-mesenchymal features, pseudo-time and immune infiltration analysis were applied to demonstrate the intratumor heterogeneity and define the specific immune-related regions, and the results were further validated by a second analysis on another ST study. In vitro and in vivo experiments were conducted to confirm the functional mechanisms of key molecules such as CCL15, CCL19 and CCL21. Clinical tissue samples were used to assess their potential prognostic and therapeutic values. Results: Totally, 7553 spots were categorized into 15 subsets by hierarchical clustering, and malignant subsets with intratumor heterogeneity phenotypes were identified. Spatial heterogeneity from distinct sectors highlights specific chemokines: CCL15 is remarkable in the core region of the carcinoma sector and facilitates the immunosuppressive microenvironment by recruiting and polarizing M2-like macrophages in vitro and in vivo; High expression of CCL15 and CD163 respectively predicts poor prognosis of HCC patients, and the combined application of them has better predictive value. CCL19 and CCL21, sharing similar spatial expression patterns, are highly-correlated and prominent in the immune infiltration enrichment and recruit CD3+ T cells and CD20+ B cells to inhibit the growth of HCC, indicating a good prognosis of HCC patients. Conclusions: Taken together, our studies preliminarily reveal intratumor heterogeneity of HCC based on ST techniques and unravel the previously unexplored spatial expression patterns in the immune microenvironment. We also highlight the clinical significance and spatial discrepancy of key molecules, providing novel insight for further developing therapeutic strategies in HCC.

antigen retrieval, blocked with 10% goat serum and then incubated with primary antibodies overnight at 4 ℃ as well as secondary antibodies at 37 ℃ for 45 mins, the details of antibodies were listed at supplementary Table S7. DAB treatment, hematoxylin staining and hydrochloric acid alcohol differentiation was used to dye nuclei. Finally, the sections were scanned by white light scanner or observed under the microscope. The relative quantification of targets in IHC depended on H-score, which integrated the percentage of positive cells and staining strength of targeted proteins.
As for immunofluorescence (IF) assays with two targets, the procedures before incubating with secondary antibodies were similar with IHC. Briefly speaking, tissue sections were incubated with CD3 and Goat Anti-Rabbit IgG (HRP), respectively. After CY3 staining, antigen retrieval and blocking, the first primary antibody was finished. Then tissue sections were incubated with CD20 and Alexa Fluor®647 goat anti-rabbit lgG (H+L), respectively. DAPI and antifluorescence quenching agents was used to dye nuclei and for storage. Tissue sections were scanned by 3DHistech (Pannoramic MIDI, Hungary) or observed with fluorescence microscope (Olympus#BX53, Japan).

Lentiviral infection
We constructed overexpressed CCL15 lentivirus named PGMLV-CMV-H_CCL15-3×Flag-EF1-mCherry-T2A-Puro. For lentiviral infection, lentivirus vectors were directly added to Huh7 cells inoculated in 6-well plates. After cells were cultured for 48 hours, the medium was replaced and the transfection efficiency was evaluated. Cells with lentivirus required 2 ug/ml puromycin treatments for 2 weeks, and then were used for further analyses.
Transwell migration assay 5 × 10 4 Cells were seeded in the upper chamber of transwell inserts (Corning, USA), which was lay upon the 24-well plates. PMA simultaneously was added into the upper chamber and an appropriate concentration of CCL15 was added into the lower chamber. Cells were cultured for 48 hours and then evaluated the migration proportions using the 20× microscope after cells were fixed by paraformaldehyde and stained by crystal violet (Beyotime, China).

Single-cell suspension preparation and flow cytometric analysis
Adherent macrophages models were digested into cell suspensions by trypsin and subcutaneous tumor tissue were treated with enzyme-assisted microdissection method by collagenase IV (Worthington, USA). Cultured cells or single cells from isolated tumor tissues were suspended in the dilution buffer.
The solutions were incubated with primary antibodies at 4℃ and secondary antibodies at room temperature or direct labeled antibodies at 4 ℃ for 30 minutes, respectively. The details and usage of antibodies were shown on Table   S7. Cells finally were detected by Flow Cytometer (CyAn ADP Analyzer, Beckman Coulter).

Construction of macrophage model in vitro
THP-1 and U937 were most commonly used cell lines to construct the macrophage model in vitro [1]. Generally, cells were inoculated in 6-well plates at suitable density. Phorbol 12-myristate 13-acetate (PMA) was added to the medium at 100 ng/μl and cultured for 48 hours. Suspension cell adherence and morphological changes can be used as signs of successful construction of macrophage model. The optimum concentration of CCL15 was determined by pre-experiment. THP-1 and U937 cells stimulated by CCL15 at appropriate concentration were cultured for a week and then collected for further studies.