The correlation of messenger RNA (mRNA) expression between chemokines and DC markers was analyzed using the cBioPortal (http://www.cbioportal.org/). Colorectal adenocarcinoma dataset of 592 samples with mRNA data (RNA Seq V2) from The Cancer Genome Atlas (TCGA), Pan-Cancer Atlas, was used for the analysis. Expression Z-scores of tumor samples compared to the expression distribution of all log-transformed mRNA expression of adjacent normal samples in the cohort (log RNA Seq V2 RSEM). The Z-score threshold was set at ± 2.0. Analytical methods included both Spearman’s rank correlation and Pearson correlation were calculated, and their correlation coefficients and respective P values were shown in the Results.

CT26 and MC38 cell lines were cultured in Roswell Park Memorial Institute Medium (RPMI 1640) and Dulbecco’s Modified Eagle Medium (DMEM), respectively, supplemented with 10% fetal bovine serum (FBS, Sigma-Aldrich) and 100 U/mL penicillin/streptomycin (Beyotime). Cell lines were authenticated by Shanghai Biowing Applied Biotechnology, and routinely tested for mycoplasma contamination. All cells were maintained in a 5% CO₂, humidified incubator at 37°C. CT26 and MC38 were infected with CCL3-, CCL19-, CCL21-, and XCL1-overexpressing lentiviruses co-expressing ZsGreen, in the presence of 6 μg/mL polybrene to generate cell lines stably expressing the chemokines or empty vector.

Total RNA was isolated by the RNA extraction kit (TIANGEN Biotech), and complementary DNA (cDNA) was synthesized by reverse transcription kit (Vazyme). Real-time polymerase chain reaction (PCR) was carried out using TB Green^® Premix Ex Taq™ (Takara^®) in the ABI 7900HT fast real-time PCR system (Applied Biosystems). mRNA expression levels of target genes were normalized to the glyceraldehyde 3-phosphate dehydrogenase (Gapdh) using the 2^−ΔΔCt method.

Proliferation/survival of CT26 and MC38 cells overexpressing chemokines or vector was determined by cell counting kit-8 (CCK-8, TargetMol^®) and measured by a microplate reader (Tecan Infinite^® M Plex) every 24 h according to the protocols from the manufacturer.

Cells were plated at densities of 500, 1,000, and 2,000 cells/well in 12-well plates using a two-layer soft agar setting. The bottom of each well was covered by an agar layer, consisting of 1× DMEM complete medium with 0.6% (w/v) agarose. Subsequently, different concentrations of cells suspended in 1× DMEM complete medium with 0.35% (w/v) agarose were added on top of the bottom agar layer. After 10 days, colonies were stained with 0.005% crystal violet solution and counted using ImageJ software.

Six- to eight-week-old male C57BL/6 mice and male BALB/c mice were purchased from Shanghai Lingchang BioTech Co., Ltd. The 0.5 × 10⁶ CT26 or 1 × 10⁶ MC38 cells overexpressing chemokines or vector were injected subcutaneously at the lower flank of mice in 100 μL phosphate-buffered saline (PBS) respectively. Tumor size was measured by a vernier caliper every three days and calculated as Tumor Volume (mm³) = Length (mm) × Width (mm) × Width (mm)/2. Mice were euthanized when tumor size reached 2,000 mm³ in volume or the tumor became ulcerated.

Whole tumors were fixed for 24 h in 4% formalin and subsequently embedded into paraffin, sectioned, and then mounted onto slides. After antigen retrieval, permeabilization, and blocking, anti-Ki-67 antibody (Abcam, ab16667) was incubated with the epitope at 4°C overnight, followed by the addition of horseradish peroxidase (HRP)-conjugated goat anti-rabbit immunoglobulin G (IgG; Abcam, ab6721, at 1:200 dilution) for another 30 min in dark at room temperature. Next, sections were treated with 3,3’ diaminobenzidine (DAB) substrate (Abcam, ab64238) and imaged by microscope (Leica DM6 B). The ratio of Ki-67⁺ cells was analyzed using ImageJ software. Four different fields were counted to calculate the average ratio for each tumor.

Mice were euthanized via CO₂ asphyxiation 15 days after the establishment of MC38 tumors. Tumor tissues were excised and cut into 1–2 mm³ pieces, and then treated with 1 mg/mL Liberase™ TL (Roche) and 100 mg/mL DNase I (Sigma-Aldrich) in gentleMACS™ Octo Dissociator (Miltenyi Biotec). After dissociation of tumor tissues, cell suspensions were passed through 70-μm cell strainers (BioFil^®) and washed twice with PBS containing 2% FBS. Inguinal lymph nodes were mechanically dissociated and passed through 70-μm cell strainers (BioFil^®) to obtain single-cell suspensions.

Before antibody staining, single-cell suspensions were stained with Zombie dyes (BioLegend^®) at a dilution of 1:200 for 10 min at room temperature to exclude dead cells. Markers on the cell surface were detected by staining cells with fluorochrome-labeled antibodies in PBS containing 2% FBS and 1 mmol/L ethylenediaminetetraacetic acid (EDTA) for 30 min on ice. For subsequent intracellular cytokine analysis, cells were washed once with PBS and treated with the True-Nuclear™ Transcription Factor Buffer Set (BioLegend^®). The following antibodies were used: CD45 (30-F11, BD Biosciences, dilution 1:200); CD11c (N418, BioLegend^®, dilution 1:100); CD11b (M1/70, BioLegend^®, dilution 1:100); CD103 (2E7, BioLegend^®, dilution 1:100); Ly6C (HK1.4, BioLegend^®, dilution 1:100); CD64 (X54-5/7.1, BioLegend^®, dilution 1:100); MHC-II (I-A/I-E, M5/114.15.2, BioLegend^®, dilution 1:100); CD3e (145-2C11, BioLegend^®, dilution 1:100); CD4 (GK1.5, BioLegend^®, dilution 1:100); NK1.1 (PK136, BioLegend^®, dilution 1:100); CD8a (53-6.7, BioLegend^®, dilution 1:100); Foxp3 (MF-14, eBioscience, dilution 1:100). The cDCs were defined as: CD45⁺CD11c⁺MHC-II⁺CD64^–Ly6C^–; the cDC1: CD45⁺CD11c⁺MHC-II⁺CD64^–Ly6C^–CD103⁺; the cDC2: CD45⁺CD11c⁺MHC-II⁺CD64^–Ly6C^–CD11b⁺; the macrophages: CD11b⁺CD64⁺Ly6C^–; the conventional CD4⁺ T cells: CD45⁺CD3e⁺CD4⁺Foxp3^–; the Treg cells: CD45⁺CD3e⁺CD4⁺Foxp3⁺; the CD8⁺ T cells: CD45⁺CD3e⁺CD8⁺; the NK cells: CD45⁺CD3e^–NK1.1⁺; the NKT cells: CD45⁺CD3e⁺NK1.1⁺. Fluorescence data were acquired on a BD LSRFortessa™ cell analyzer (BD Biosciences). Data were analyzed on FlowJo™ software.

Statistical tests were performed in GraphPad Prism 8. CCK-8 assay, colony formation assay, immunohistochemistry, and tumor growth in vivo were analyzed via two-way analysis of variance (ANOVA) followed by multiple comparisons; data from flow cytometry were analyzed via one-way ANOVA followed by multiple comparisons. Results were presented as mean values ± standard error of the measurement (SEM), and P < 0.05 was considered statistically significant.

To explore the potential in vivo role of the selected chemokines in regulating DCs in human CRC patients, we first analyzed the correlation of their mRNA expression with DC markers. Expression of all the four chemokine genes, including CCL3, CCL19, CCL21, and XCL1, was positively correlated with that of both ITGAX encoding the general DC marker CD11c and CLEC9A, which is a cDC1 marker, in human CRC samples (Figure 1) [34]. These findings implied that all the four chemokines played a positive role in DC infiltration in human colorectal tumors, although they might have discrepant activities in regulating different DC subsets and other immune cells.

To clarify the anti- or pro-tumor effects of the DC-targeting chemokines, we generated murine CRC cell lines that stably expressed the four chemokines. Successful overexpression of mouse CCL3, CCL19, CCL21, and XCL1 (encoded by Ccl3/Ccl19/Ccl21/Xcl1) in CT26 and MC38 cell lines was confirmed by quantitative reverse transcription-PCR (Figure 2A and 2B). CCK-8 assay demonstrated that overexpression of these genes did not significantly change the total numbers of cells during the 72-h culture period, compared to empty vector control, except that CCL3 and CCL21 reduced the growth rate of CT26 cells but to a limited extent (Figure 2C and 2D). Moreover, the colony-forming ability of MC38 cells was not affected by these chemokines (Figure 2E). It suggested that the chemokines had trivial or no direct effect on the proliferation or survival of tumor cells.

We then established subcutaneous tumor models using CT26 and MC38 overexpressing the four chemokines. In both models, tumoral overexpression of CCL3, CCL19, CCL21, and XCL1 significantly repressed tumor growth (Figure 3A–C), despite the differential overexpression levels in these cell lines (Figure 2A and 2B). Consistently, these chemokines also decreased the ratio of Ki-67⁺ cells in MC38 tumors indicating a negative effect on tumor cell growth in vivo (Figure 4A and 4B). Taking the results from CT26 and MC38 models together, CCL19 seemed to be the most potent anti-tumor chemokine among the four. Since these chemokines had no or little effects on the proliferation or survival of tumor cells (Figure 2C and 2D), their anti-tumor activities in vivo were very likely to depend on their immunomodulatory effects in the TME.

Next, we sought to investigate the effects of these chemokines on DC migration in MC38 tumor models by fluorescence-activated cell sorting (FACS). MC38 cells expressed ZsGreen simultaneously as an indicator of tumor-specific antigen. Tumoral overexpression of CCL19, CCL21, and XCL1 significantly increased ratios of total cDCs, cDC1, and cDC2 subtypes among CD45⁺ leukocytes in the TME, as compared to the empty vector control tumors (Figure 5A and 5B). However, CCL3 did not change the ratios of total cDCs or the two subtypes. In contrast, the percentages of macrophages were not affected by either chemokine, suggesting these chemokines were DC-specific. In particular, the ratio of ZsGreen⁺ cDCs was upregulated in XCL1-overexpressing tumors, suggesting an enhanced uptake of tumor antigens (Figure 5C).

Interestingly, cDCs, including cDC1 and cDC2 in the tumor-draining lymph nodes, were significantly decreased by tumoral expression of these chemokines, implying an efflux of cDCs from the lymph nodes or a biased upregulation of other immune cells such as T cells upon antigen presentation and stimulation (Figure 6A and 6B). Higher percentages of ZsGreen⁺ cDCs were found in the lymph nodes of CCL3- and XCL1-overexpressing tumor models (Figure 6C). This finding suggested that more tumor antigens were delivered by cDCs from tumors overexpressing these two chemokines, and possibly more tumor-specific T cells would be generated. The increased tumor antigen ZsGreen in the lymph nodes of XCL1-overexpressing tumor models may be partially attributed to the increased antigen uptake in the tumor (Figure 5C). These data demonstrated that tumoral expression of these chemokines did not impair or even promote the antigen transfer from the tumor site to tumor-draining lymph nodes.

Confirming a positive effect on cDCs in TME and lymph nodes, we further analyzed the tumor-infiltrating effector lymphocytes by FACS (Figure 7A). CD45⁺ leukocytes in total live cells were upregulated only in CCL3-overexpressing tumors, although the ratios of lymphocytes in CD45⁺ cells were not changed by CCL3 (Figure 7B). Consistent with their effects on cDCs, CCL19, CCL21, and XCL1 significantly increased the ratios of conventional Foxp3^–CD4⁺ T and CD8⁺ T cells. These three chemokines also elevated ratios of Treg cells, although the effect of CCL19 was not statistically significant. NK cells were upregulated by CCL19 and XCL1, while CCL19, CCL21, and XCL1 upregulated NKT. Together, these results revealed that these chemokines boosted anti-tumor adaptive immunity and enrichment of other effector cells with anti-tumor activities such as NK and NKT cells.