Tumor metastasis is the primary cause of mortality in patients with cancer. Several chemokines are identified as important mediators of tumor growth and/or metastasis. The level of CXCL13 has been reported to be elevated in serum or tumor tissues in patients, which mainly functions to attract B cells and follicular B helper T cells. However, the role of CXCL13 in cancer growth and metastasis is not fully explored. In the current study, we found that CXCL13 is not a strong mediator to directly promote tumor growth; however, the mice deficient in CXCL13 had far fewer pulmonary metastatic foci than did the wild-type mice in experimental pulmonary metastatic models. In addition, Cxcl13 -/- mice also had fewer IL-10-producing B cells (CD45+CD19+IL-10+) in the metastatic tumor immune microenvironment than those of wild-type C57BL/6 mice, resulting in an enhanced antitumor immunity. Notably, CXCL13 deficiency further improved the efficacy of a traditional chemotherapeutic drug (cyclophosphamide), as well as that of anti-programmed death receptor-1 immunotherapy. These results suggested that CXCL13 has an important role in regulating IL-10-producing B cells in tumor metastasis and might be a promising target for improving therapeutic efficiency and stimulating tumor immunity in future cancer therapy.

2 Materials and Methods

2.1 Animals

Cxcl13-/- Mice and Il10-/- mice were purchased from the Jackson Laboratory. 6-8 weeks old Wild Type C57BL/6 mice were purchased from Vital River (Beijing, China). All of these mice were bred in specified pathogen free conditions. Animal experiments were performed according to the guidelines and approved by the animal care and use committee of Sichuan University (Chengdu, China).

2.2 Cell Culture

The murine tumor cell lines B16-F10, LL/2, E.G7-OVA, ID8 were obtained from the American Typical Culture Collection (ATCC). B16-F10, LL/2, ID8 were cultured in DMEM medium (Gibco) with 10% fetal bovine serum (FBS) (Gibco) and 50μg/ml penicillin/streptomycin (Gibco). E.G7-OVA were maintained in RPMI 1640 medium (Gibco) suppled with 10% FBS, 50μg/ml penicillin/streptomycin and another 400 μg/ml G418. All of these cell lines were cultured in 37 oC with humidified 5% CO2.

2.3 Cell viability and Cell proliferation detection

The Cell Counting Kit-8 (MedChemExpress) was used to test the viabilities of cancer cells according to the manufacture protocol. Briefly, different numbers of cells were seeded in 12-well plates and cultured for 24 h before the treatment. Then the cells were incubated with 500 ng/ml of recombinant murine CXCL13 (rmCXCL13) (PeproTech) for indicated times followed by staining with 10% CCK-8 solution at 37℃ for another 2 h. Then the absorbance at 450 nm was measured by Microplate Reader.

For cell proliferation assay, tumor cells were seeded into 12-well plates (around 2.5 ×104 - 1×105 cells per well). After culturing overnight (except E.G7-OVA cells), cells were subsequently cultured in the presence of 500 ng/ml of rmCXCL13 and cell numbers were counted every day using a cell counter (Countstar).

2.4 Cell migration assay

B16-OVA and LL/2 cells (5 × 104 cells) were suspended in 200 μl serum-free medium on the upper chamber in the 24 well plates, while the lower chamber was filled with 600 μl of medium containing 10% FBS, antibiotics and 500 ng/ml of rmCXCL13. Then, the cells were incubated at 37 °C with 5% CO2 for 24 h. Cells were washed with PBS, then fixed with 2% paraformaldehyde for 10 min, stained with 0.5% crystal violet for 5 min at room temperature, and then washed with ddH2O twice. Finally, cells remained in the upper chamber were slightly removed using cotton swabs, and those that had migrated to the surface of the lower chamber were captured by a microscope (Olympus). The migrated cells in the pictures were counted by Image J software.

2.3 Flow cytometry (FCM) analysis

The mice under homeostatic conditions or in the experimental pulmonary metastasis models were sacrificed. The lungs containing tumor nodules or not and spleens of mice were collected.

Mouse lung tissues were scissored into small pieces and then digested into single cell suspension by 1 mg/mL collagenase Type I, 0.5 mg/mL collagenase Type IV (Gibco) and 0.05 mg/mL DNase I (Sigma) in RPMI 1640 basic medium for 1 h in 37oC. Resulting cell suspensions were passed through 70-μm MACS® SmartStrainers (Miltenyi Biotec) to remove clumps of cells and debris. The spleens were directly ground on the BD Falcon™ 70-μm nylon cell strainers with a syringe plunger to obtain single cell suspensions. The red blood cells (RBCs) were then lysed by RBCs lysis buffer (154 mM NH4Cl, 10 mM KHCO3, 0.1 mM EDTA·2Na, pH 7.4) for 5 min at room temperature. Then single cells without RBCs were washed with phosphate-buffered solution (PBS) for two times and resuspended in PBS.

For cell surface staining, cells (1 × 106 cells/100 ml/tube) were stained with 1 ml antibodies at 4 oC for 30 min in the dark; for intracellular or transcription factor cytokine staining, following the surface staining (2 × 106 cells/100 ml/tube), cells without extra stimulation were performed with BD Cytofix/Cytoperm™ Fixation/Permeabilization Kit (BD Bioscience) and Transcription Factor staining buffer Set (eBioscience), respectively. Then the fixed and permeabilized cells were stained with 5-10 ml intracellular antibodies at room temperature for 2 h. Finally, cells were washed and resuspended in PBS (500 ml/tube), then processed using FACS Calibur Flow Cytometer (BD Biosciences) or NovoCyte Flow Cytometer (ACEA Biosciences), and the results were analysed using FlowJo version 10 software or NovoExpress software respectively. Dead cells were excluded using Live/Dead® Fixable Dead Cell Stain Kits (Near-IR fluorescent reactive dye, Life Technologies).

The following anti-mouse antibodies purchased from BD Biosciences or BioLegend were used: PerCP/Cyanine5.5 or APC-conjugated anti-CD45, PerCP/Cyanine5.5-conjugated anti-CD3ε, APC or PE-conjugated anti-CD4, FITC or APC-conjugated anti-CD8α, PerCP Cyanine5.5-conjugated anti-CD69, PE-conjugated anti-IFN-γ, APC or PE-conjugated anti-CD19, FITC or PE-conjugated anti-IL-10, PE-conjugated anti-CD25, Alexa Fluor 488-conjugated anti-FoxP3, PE-conjugated anti-F4/80, FITC-conjugated anti-CD11b, APC-conjugated anti-CD11c, PE-conjugated anti-Ly-6G, APC-conjugated anti-Ly-6C, Brilliant Violet 421-conjugated anti-CD206, FITC-conjugated anti-CD185(CXCR5), PE.Cy7-conjugated anti-CD45R/B220. FITC-conjugated anti-CD40, PE.Cy7-conjugated anti-MHC II, PE-conjugated anti-CD24, Brilliant Violet 421-conjugated anti-CD64, PerCP/Cyanine5.5-conjugated anti-CD21, Alexa Fluor 647-conjugated anti-CD23, FITC-conjugated anti-CD80, FITC-conjugated anti-CD86, FITC-conjugated anti-CD1d, APC-conjugated anti-CD5, FITC-conjugated anti-IgM, PE-conjugated anti-IgD, Brilliant Violet 421-conjugated anti-TGF-b. APC-conjugated anti-Arginase-1 was purchased from Invitrogen. PE-conjugated anti-IL-35/EBI3 Subunit and Aexa Fluor 488-conjugated anti-IL-35/p35 Subunit were purchased from R&D Systems.

2.4 Animal experiments

To establish experimental pulmonary metastatic model, B16-OVA cells (2×105) were intravenously (i.v) injected into C57BL/6 mice. Mice were sacrificed on Day 14 or 21 for measuring lung metastatic foci or lung weight. Lung metastatic foci were calculated under the tissue microscope (Leica MZFL III). For the subcutaneous tumor model, B16-OVA cells (1×106) were subcutaneously (s.c) transplanted into the right back flank of C57BL/6 mice, tumor size were measured every two days and tumor volumes were calculated by the following formula: tumor volume = length × weigh2 × 0.52.

For the adoptive transfer B cells model, the lymphocytes were isolated from the spleen of wild type (WT) or Il-10-/- C57BL/6 mice using the mouse lymphocyte separation medium (DAKEWE), the regulatory B cells were stimulated or unstimulated and then separated by the regulatory B cell isolation kit (Miltenyi Biotec) according to the manufacturer’s instruction. Briefly, the pre-enrichment of B cells from lymphocytes were stimulated in vitro with 10 mg/mL LPS (Sigma) for 24 h at 37 °C, 5% CO2 and added 50 ng/mL PMA and 500 ng/mL ionomycin (Sigma) for the last 5 h of stimulation. Then IL-10-producing regulatory B cells are specifically isolated. B cells from Il-10-/- mice were used as a negative control, which were isolated and stimulated exactly the same way as the regulatory B cells from WT mice. Subsequently, Il-10-/- B cells or regulatory B cells were i.v adoptively transferred (2 × 106 cells/200 mL/mouse) into tumor-bearing WT recipient mice on day 1, 3 and 6 after the i.v injection of B16-F10 tumor cells (2×105). Mice were sacrificed on day 14 for lung metastases measure, morphological analysis and FCM analysis.

For chemotherapeutics, WT or Il-10-/- C57BL/6 mice were intraperitoneal (i.p) injected with 100 mg/kg Cyclophosphamide (Melonepharma/Meilun biotech) or solvent on day 0, 3, 5 after the i.v injection of B16-F10 tumor cells. For anti-PD-1 (BioXCells) therapy, WT or Il-10-/- C57BL/6 mice were i.p injected with control IgG or anti-PD-1 monoclonal antibodies (250 mg/mouse) at day 0, 3 after the i.v injection of B16-F10 tumor cells.

2.5 Western blot analysis

Immunoblotting was performed using a standard protocol. After aspirating the supernatant, B16-F10 and LL/2 tumor cells in the plates were lysed and denatured by the addition of sample buffer and boiling for 10 min. Proteins were loaded (10 ul/lane) and resolved by 8% sodium dodecyl sulphate (SDS) - polyacrylamide gel electrophoresis (PAGE) and transferred onto PVDF membranes (Millipore). After blocking with 5% non-fat milk in tris-buffered saline containing 0.05 % Tween 20 (TBST) for 1 h at room temperature, the membranes were probed overnight at 4°C with primary antibodies used as follows: rabbit polyclonal anti-E-cadherin (Proteintech, 1:5,000), rabbit polyclonal anti-N-cadherin (Proteintech, 1:2,000), rabbit polyclonal anti-MMP-9 (Proteintech, 1:1,000) and rabbit polyclonal anti-GAPDH (Santa Cruz Biotechnology, 1:4,000). After washing with TBST three times, the membranes were incubated with horseradish peroxidase (HRP)-conjugated goat anti-rabbit secondary antibodies for 1 h at room temperature (CST, 1:5,000). Membranes were then developed using an enhanced chemiluminescence detection kit (Millipore) and exposed to autoradiography film with a film developer in a dark room. GAPDH was used as an internal control.

2.5 Quantitative Real-Time PCR

Total RNA was extracted by the RNAsimple Total RNA Kit (TIANGEN). The reverse transcription was achieved by the iScript reverse transcriptase (Bio-Rad). The real-time PCR was performed with the CFX Connect Real-Time PCR Detection System (Bio-Rad). KAPA SYBR FAST qPCR Master Mix (Kapa Biosystems) was used, and the expression of the above genes was normalized by 18S ribosomal RNA (rRNA). The mouse primers used were as follow:

Ifng, 5’-TGAACGCTACACACTGCATCT-3’ (forward) and 5’-GACTCCTTTTCCGCTTCCTGA-3’ (reverse);

Prfn, 5’-TGGAGGTTTTTGTACCAGGC-3’ (forward) and 5’-TAGCCAATTTTGCAGCTGAG-3’ (reverse);

Tgfb1, 5’-AAGTTGGCATGGTAGCCCTT-3’ (forward) and 5’-GGAGAGCCCTGGATACCAAC-3’ (reverse);

Grzb, 5’-CTCTCGAATAAGGAAGCCCC-3’ (forward) and 5’-CTGACCTTGTCTCTGGCCTC-3’ (reverse)；

Cxcr5, 5′-ATGAACTACCCACTAACCCTGG-3′ (forward) and 5′-TGTAGGGGAATCTCCGTGCT-3′(reverse);

18S rRNA, 5′-CGCCGCTAGAGGTGAAATTCT-3′ (forward) and 5′-CGAACCTCCGACTTTCGTTCT-3′(reverse).

2.6 Gelatin zymography analysis for MMP-9 activity

The conditioned media obtained from B16-F10 and LL/2 cells stimulated with or without rmCXCL13 were collected for MMP-9 enzyme activity detection by Gelatin Zymography Analysis Kit according to the manufacturer’s protocol (Real-Times (Beijing) Biotechnology Co.,Ltd). Briefly, serum-free conditioned media samples were subjected to 10 % SDS-PAGE, containing 0.1 % gelatin under non-reducing conditions. The gels were carefully removed after electrophoresis and washed in renaturing buffer for 30 min at room temperature, and then incubated overnight at 37 oC in developing buffer with gentle agitation. After rinsing three times with deionized water, the gels were stained with FastBlue protein staining solution at room temperature. Next, the gels were carefully removed from the water and placed in a plastic sheet protector. A high resolution scanister was used for scanning the gels. The formation of clear bands around 100 kDa on the gelatin gels against the blue background indicated the gelatinolytic activity of MMP-9. Collagenase IV lane was added and used as a positive control.

2.7 Immunofluorescence staining

Lung specimens were fixed with 4% formaldehyde and embedded in paraffin. Sections (3 µm) were deparaffinized and then subjected to an EDTA buffer for antigen retrieval. The sections were incubated with blocking buffer (5% goat serum) for 30 min at room temperature, and subsequently incubated with secondary antibody (Life, Alexa Fluor 488 goat anti-mouse IgG, 1:800) for 1 h at room temperature, which was prepared in 1% BSA in PBS. Finally, the cell nucleus was stained with 1 ug/mL DAPI for 5 min at room temperature. Sections were mounted with fluorescent mounting medium (Sigma, F6182) and observed with Vectra 3 Automated Quantitative Pathology Imaging System (Akoya Biosciences).

2.8 Statistical Analysis

The statistical analysis was performed by the GraphPad Prism 7 software. The data are presented as means ± s.d. or means ± s.e.m. The P values was evaluated by a two-sided unpaired t-test or one- or two-way analysis of variance (ANOVA), in which P < 0.05 were considered as statistical significance among each group.

3.1 CXCL13 deficiency inhibits murine experimental pulmonary metastasis

Some types of tumor cells express chemokine receptors which subsequently affect the tumor growth and metastatic behavior. The cognate receptor of CXCL13, namely, CXCR5 has been reported to be expressed on CT26 cells in vivo and several pancreatic carcinoma cell lines (A818-4, AsPC1, BxPc3, Colo357, HPAF2, MiaPaCa2, Panc1, Panc89,PancTu-I, and PT45P1) (12). To study how CXCL13-CXCR5 signaling affects tumor metastasis, we first detected the CXCR5 expression in several cell lines including murine melanoma cell line (B16-F10), murine Lewis lung carcinoma cell line (LL/2), murine lymphoma cell line (E.G7-OVA), murine ovarian surface epithelial cell line (ID8). The flow cytometry (FCM) results showed that no CXCR5 was expressed on the cell lines mentioned above or B16-F10 cells derived from tumor tissues of murine pulmonary metastatic models (In vivo B16-F10) (Fig. 1A). In contrast, as the positive control, mouse lymphocytes highly express CXCR5 (Fig. 1A). Similar results were observed in these cells further detected by RT-PCR (Supplemental Fig. 1A).

Next, we managed to study whether CXCL13 could directly influence the tumor growth. Results indicated that recombinant murine CXCL13 (rmCXCL13) -treated tumor cells showed similar viability compared with PBS-treated tumor cells (Fig. 1B). Moreover, our data also showed that CXCL13 does not influence the proliferation of tumor cells (Supplemental Fig. 1B). Collectively, our results demonstrate that CXCL13 does not affect tumor cell growth in vitro, including cell viability and proliferation. To further study whether CXCL13 is associated with tumor growth in vivo, tumor cells including B16-F10 cells, LL/2 cells, or E.G7-OVA cells were injected subcutaneously (s.c.) into wild type (WT) or Cxcl13-/- mice. However, no significant differences in the tumor growth rates of WT and Cxcl13-/- mice were observed (Fig. 1C). These results suggested that CXCL13 may not be able to stimulate the growth of tumor cells by CXCL13-CXCR5 signaling in vitro and in vivo. Furthermore, we examined the role of CXCL13 on tumor migration or expression of ECM remodeling proteins. It was found that exogenous CXCL13 stimulation in vitro does not affect the migration and EMT proteins expression of B16-F10 and LL/2 (Supplemental Fig. 2A& B). In addition, not only the protein level of MMP-9 (one of ECM remodeling proteins) was not affected by CXCL13 stimulation, but also the enzyme activity of MMP-9 did not change (Supplemental Fig. 2C)

In the next set of experiments, we investigated the effects of CXCL13 on tumor metastatic behaviors. To establish the experimental metastasis models, 2 ´ 105 B16-F10 cells were intravenously injected into WT and Cxcl13-/- mice. Remarkably, 14 days after the inoculation of tumor cells, the level of pulmonary metastasis was distinctly lower in Cxcl13-/- mice comparing to that in WT mice (Fig. 1D). On day 21, the lung weights of Cxcl13-/- mice were significantly lower than that of WT mice (Fig. 1E). In addition to melanoma, we also established this kind of model of lung cancer cells-LL/2. After 21 days, mice were sacrificed and the lungs isolated from Cxcl13-/- mice also exhibited lower weights (Fig. 1F). These results showed that the pulmonary metastases were inhibited in Cxcl13-/- mice and implied that the CXCL13 signaling pathway plays an important role in the process of pulmonary metastasis.

3.2 Cxcl13 deficiency reduced the infiltration of regulatory B cells into tumor metastatic microenvironment

Tumor microenvironment contains a wide range of cell types which interact with the complex cytokine/chemokine network. A previous study have shown that the inhibition of chemokine ligand-receptor signaling can enhance the migration of T cells to tumor metastatic sites to augment antitumor immunity(13). Moreover, it has been reported that B cells have the capability to secrete immunosuppressive cytokine interleukin-10 (IL-10), which can repress the T cell mediated immune response by reducing the secretion of interferon-γ (IFN-γ) from CD8+ T cells(14).

To see whether CXCL13 could impact the immunity against metastatic cancer, we analyzed B cells and IL-10 level from melanoma pulmonary metastatic sites of WT and Cxcl13-/- mice by FCM. The loss of Cxcl13 gene led to the reduction of the percentages of IL-10-secreting B cells (CD45+CD19+IL-10+) in metastatic sites on Day 7, Day 14 and Day 21 after the tumor cell inoculation (Fig. 2A-F), which suggests that Cxcl13 deficiency is responsible for the inhibited infiltration of IL-10-secreting B cells at metastatic sites. Besides, the total CD19+ B cell populations on day 14 and day 21 slightly declined in Cxcl13-deficient mice, which has no statistical significance compared to WT mice (Fig. 2D-F). Furthermore, we investigated the B cell population of WT or Cxcl13-/- mice under homeostatic condition. Similar data were observed in the lung and no difference was found in the proportion of B cells in the spleen (Supplemental Fig. 3).

Generally, regulatory B cells are characterized as a subset of IL-10-secreting B cells. However, previous studies have shown that regulatory B cells also could produce IL-35 and TGF-β to inhibit immune responses (15, 16), in which IL­35 is a heterodimeric cytokine composed of p35 subunit of IL-12 and Epstein­Barr virus induced gene 3 (EBI3) subunit of IL-27. As expected, tumor B cells from Cxcl13-/- mice dramatically reduced the IL-35 and TGF-b production than that from WT mice (Fig. 2G&H). The above observations support our notion that Cxcl13 deficiency mainly reduced the infiltration of IL-10, IL-35 and TGF-b-secreting regulatory B cells into tumor metastatic microenvironment.

To uncover the phenotypes of B cells infiltrated into pulmonary metastatic foci, we gated CD45+CD19+IL-10+ regulatory B cells in metastatic foci by multi-color FCM to analyze the expression of common B cell markers. The data shows that this group of B cells are negative for CD23, and express low level of CD21, CD25, CD40, CD80 CD86 and IgM. Besides, they are positive for CD1d, CD5, and express high level of CD24, MHC-II and B220/CD45R (Supplemental Fig. 4A). As previous studies revealed that phenotype of regulatory B cell in autoimmune diseases or cancer are CD19+CD1dhighCD5+, CD21highCD23lowCD19+ or B220+CD25+(10, 17, 18). In our study, the regulatory B cells in pulmonary metastasis sites are CD45+CD19+IL-10+CD1d+CD5+CD24highMHC-IIhighB220high (Supplemental Fig. 4A), whereas in previous studies, the regulatory B cells identified from autoimmune diseases or cancer were CD19+CD1dhiCD5+, CD21hiCD23lowCD19+ or B220+CD25+(10, 17, 18). To further detect the phenotype of IL-10+ regulatory B cells in metastasis, we measured the CXCR5 expression and found that loss of Cxcl13 modestly suppressed the expression of CXCR5 on the regulatory B cells (Supplemental Fig. 4B). Therefore, the immunophenotypes of B cells which are responsible for the inhibited metastasis in Cxcl13-/- mice in the present study are not completely the same as what previous studies have shown, but they do have some common features.

Furthermore, in the tumor microenvironment, antibody produced by B cells also influence tumor development. Thus, we examined whether antibody abundances are changed in Cxcl13-/- mice by immunofluorescence. As expected, Cxcl13 deficiency enhanced the antibody abundances in the lung metastatic tumor microenvironment (Supplemental Fig. 5). This data implied that the increase in antibody produced by B cells may also contribute to the antitumor immune response in Cxcl13-/- mice, together with the aforementioned reduction in IL-10, IL-35 and TGF-b-producing regulatory B cells.

3.3 Reduction of regulatory B cell leads to boosted anti-tumor immunity in metastatic microenvironment

In tumor microenvironment, there are some kinds of important immunosuppressive cells such as type II macrophage (M2 Mac), myeloid-derived pro-tumor neutrophils, regulatory T cells (Treg), as well as regulatory B cells(19). To understand the mechanism by which the CXCL13 influences the metastatic microenvironment, we measured the changes in immune cell percentages at pulmonary metastatic sites in WT and Cxcl13-/- mice. We firstly found that the CD45+CD11b+ myeloid cell population obviously increased in the Cxcl13-/- mice. To further characterized these increased myeloid cell population, we utilized antibodies repectively against Ly-6G and Ly-6C for FCM staining. Date showed that the anti-tumor monocytes (CD45+CD11b+Ly-6C+Ly-6G-) in tumor metastatic microenvironment was severely elevated, while the pro-tumor neutrophils (CD45+CD11b+Ly-6G+) had no visible difference in Cxcl13 deficient mice (Fig. 3A and B). The above data indicated that the increased CD45+CD11b+ myeloid cell population was predominantly anti-tumor monocytes rather than pro-tumor neutrophils. Besides, loss of Cxcl13 strongly reduced the infiltrating of tumor growth-promoting M2 Mac (CD45+CD11b+F4/80+Arginase-1+CD206+) to the tumor microenvironment (Fig. 3C and D). In addition, other types of anti-tumor immune cells including DCs (CD45+CD11b+CD11c+CD24+CD64-) (Fig. 3E&F), tumor infiltrating activated CD4+ T lymphocytes (CD3+CD4+CD69+) (Fig. 3G) and CD8+ T lymphocytes (CD3+CD8+CD69+) (Fig. 3H), Th17 cells (CD3+CD4+IL-17A+) (Fig. 3I) and cytotoxic T lymphocytes (CTLs) (CD3+CD8+IFN-γ+) (Fig. 3J) were found to be increased in metastatic sites. Moreover, in Cxcl13 knockout mice, the percentages of pro-tumor regulatory T cells (CD3+CD4+FoxP3+) decreased compared with that in WT mice (Fig. 3K). Taken together, the FCM data suggested that the reduction of CD19+IL-10+ regulatory B cells could lead to the reduction of immunosuppressive cells and thereby enhance the functioning of anti-tumor cells, which might contribute to the inhibitory effect of pulmonary metastasis observed in Cxcl13-/- mice.

Furthermore, we also analyzed the mRNA transcription level of transforming growth factor-β (Tgfb1), perforin (Prfn), interferon-γ (Ifng) and granzyme-b (Grzb) in pulmonary metastatic foci by quantitative RT-PCR. As an immunomodulating cytokine, the transcription level of Tgfb1 in Cxcl13-/- models is not as strong as that in WT models (Fig. 3L). Besides, the transcription levels of Prfn, Ifng and Grzb were increased in Cxcl13-/- models than those in WT models (Fig. 3L). Notably, all of these cytokines or functional proteins are associated with the cytotoxic T cell activity, which further implies that the loss of CXCL13 could enhance the antitumor immune response in the metastatic tumor.

3.4 Adoptive transfer of regulatory B cells promotes tumor pulmonary metastasis in WT mice

To further confirm underlying mechanism and investigate the role of regulatory B cell, the adoptive transfer experiments were carried out. Followed by the establishment of B16-F10 pulmonary metastasis model, Il-10-/- B cells or regulatory B cells were intravenously injected into WT tumor-bearing recipient mice. The phenotype of these adoptively transferred B cells was analysed by FCM and data indicated that IL-10-secreting regulatory B cells were well enriched by the kit from 27% to 84% after in vitro stimulation (Supplemental Fig. 6). The adoptive transfer results showed that, the transfer of IL-10-producing B cells remarkably promoted pulmonary metastasis of B16-F10 cells, while the negative control of Il-10-/- B cells exhibited minor effects on tumor metastasis compared with the PBS group (Fig. 4A). Additionally, the H&E staining further confirms that adoptive transfer of WT regulatory B cells led to the increased numbers of metastatic foci in lungs (Fig. 4B). We further explored the role of untreated normal B cells in cancer metastasis. However, data indicated that adoptive transfer of unstimulated normal B cells has no effect on tumor pulmonary metastasis (Supplemental Fig. 7).

To further look into how the transferred regulatory B cells shaped the tumor microenvironment, the percentages of IL-10 producing B cells, regulatory T cells, activated T cells were detected by FCM. Despite the increased counts of CD45+CD19+IL-10+ B cells in pulmonary metastatic sites after adoptive transfer of regulatory B cells (Fig. 4C), the immune suppressive CD3+FoxP3+ regulatory T cells were also enriched in metastatic sites while compared with that of the control group (Fig. 4D). In contrast, the percentages of CD4+CD69+ T cells, CD8+CD69+ T cells, and CTLs were reduced after the transfer of regulatory B cell (Fig. 4E-G). Hence, these results demonstrate that regulatory B cells could promote pulmonary metastasis by the secretion of IL-10, which might be also responsible for the increased number of regulatory T cells, the inhibition of CD4+ and CD8+ T cell activation, and down regulation of the IFN-γ level, resulting in the favorable metastatic tumor microenvironment with suppressed antitumor immunity. 

3.5 Deficiency of CXCL13 enhanced the anti-tumor efficacy of chemotherapy

It has been revealed that some types of immune cells, including B cells, in tumor microenvironment contribute to the chemotherapy drug resistance(19). To investigate whether CXCL13 is associated with chemo-resistance, we treated tumor-bearing Cxcl13-/- mice with chemotherapy using cyclophosphamide (CTX). Interestingly, the data showed that both Cxcl13-/- mice without chemotherapy and WT mice treated with CTX did not show significant inhibition on tumor metastasis while compared with control group. However, the effects of chemotherapy in Cxcl13-/- mice exhibited a remarkably decreased number of metastatic foci while compared with other groups (Fig. 5A&B). The pathological examination of H&E staining also confirmed the results (Fig. 5C). Besides, the FCM analysis revealed the significantly elevated percentages of activated CD4+ T cells and CTLs in the CTX treated Cxcl13-/- mice compared to that of other groups (Fig. 5D&E). In conclusion, these results demonstrated that CXCL13 might be closely associated with chemo-resistance and is a potential target for enhancing the antitumor effects of chemotherapy with a combinational therapeutic strategy.

3.6 Deficiency of CXCL13 enhanced the efficacy of cancer immunotherapy targeting PD-1

So far, targeted therapy against immune checkpoint has been intensively studied, such as targeting programmed death receptor-1 (PD-1) and cytotoxic T lymphocytes associated antigen-4 (CTLA-4) in the treatment of melanoma and non-small cell lung carcinoma (NSCLC)(20). It has been proven that pembrolizumab (anti-PD-1 antibody) could prolong progression-free and overall survival of patients with melanoma.(21, 22) However,  immunotherapy targeting PD-1 is still sometimes limited by the tumor microenvironment. In the present study, we investigated whether CXCL13 and regulatory B cell affects the anti-tumor effects of immune checkpoint-targeted therapy. The monoclonal antibodies against PD-1 were injected into Cxcl13-/- mice and WT mice were inoculated with B16-F10 melanoma cells. The data showed that, although the monoclonal antibodies against PD-1 were effective in the treatment of melanoma pulmonary metastasis in WT mice, the Cxcl13-/- mice manifested better outcomes (Fig. 6A&B). However, the most favorable outcomes were observed in Cxcl13-/- mice treated with the monoclonal antibodies against PD-1 (Fig. 6A&B). Moreover, the results from H&E staining analysis also suggest that the monoclonal antibodies against PD-1 is more effective for the treatment of melanoma pulmonary metastasis in Cxcl13-/- mice than that in WT mice (Fig. 6C). The results showed that more activated CD4+ T cells and CTLs were observed in the Cxcl13-/- + Anti-PD-1 mAb group (Fig. 6D&E). These results indicate that the monoclonal antibodies against PD-1 is more effective for the augmentation of T cell responses in Cxcl13-/- mice than that in WT mice. In addition, comparing to control group, anti-PD-1 antibody has made no significant difference between the proportion of CD4+FoxP3+ regulatory T cells in the the metastatic sites of WT mice, but it led to the decreased regulatory T cells in the metastatic sites of Cxcl13-/- mice (Fig. 6F). In conclusion, our study shows that the monoclonal antibodies against PD-1 is more effective for the treatment of melanoma pulmonary metastasis in Cxcl13-/- mice than in WT mice, which implies that CXCL13 may be considered as an combinational therapeutic target for enhancing the efficacy of immune checkpoint inhibitors in the treatment of pulmonary metastatic tumors.

Tumor metastasis is the primary cause of mortality in patients with cancer. However, the currently available therapeutic approaches still have limited therapeutic effects. Thus, in the past few decades, considerable efforts have been made to discover prospective methods for the treatment of tumor metastasis. The relationship between chemokine network and tumor metastasis have been reported by numerous studies. Importantly, the CXCL12-CXCR4 axis has drawn substantial attention because CXCR4 is the most common chemokine receptor expressing on tumor cells.(2) Previous studies revealed that CXCL12-CXCR4 axis is responsible for lung, bone marrow and liver metastasis in a variety of types of cancers(23). Moreover, CCL21-CCR7 axis(2, 24) and CCL2-CCR2 have also been reported to act as contributors to metastasis(25). In the present study, it was shown that CXCL13 is able to recruit CXCR5+ B cells to metastatic sites, and B cells could secrete IL-10 to inhibit effective antitumor immune response, which facilitate pulmonary metastasis.

B cells are involved in multiple immunological processes such as the secretion of immunoglobulins, antigen presentation and immunosuppression. B cells that secrete IL-10 are generally considered as regulatory B cells. However, there are no generally accepted marker antibody panel for the characterization of regulatory B cells. Thus, more investigations are needed for the further understanding and characterization of regulatory B cells. In our study, the regulatory B cells in pulmonary metastasis sites are CD45+CD19+IL-10+CD1d+CD5+CD24hiMHC-IIhiB220hi, which have some common features as the reported ones. Moreover, our study has shown that B cell-derived IL-10 could inhibit effective antitumor responses by affecting immune cells in tumor microenvironment. The reduction of IL-10 producing B cells enhanced antitumor immunity, increased the amount of activated Th and Tc cells that infiltrated into metastatic sites, and led to the elevation of CD8+ cell-derived IFN-γ. Besides, we also found that the mRNA level of granzyme B, interferon-γ and perforin were increased while the mRNA level of TGF-β was reduced. To confirm that IL-10 secreted by regulatory B cells contributes to the development of pulmonary metastasis, we performed adoptive transfer of IL-10-producing B cells or IL-10-/- B cells to pulmonary metastasis modles established in WT mice. It was shown that the transfer of IL-10-producing B cells led to the inhibition of antitumor immunity and increased the proportion of pro-tumor CD4+FoxP3+ regulatory T cells in metastatic sites.

The relationship between CXCL13 and cancer advancing has been reported by previous studies. For instance, CXCL13 derived from myofibroblasts could promote the progression of malignant prostate cancer(26) and CXCL13 derived from human bone marrow endothelial cells could induce the invasion of prostate cancer cells to bone (27),(28) . It has also been observed that, in breast cancer and lung cancer patients, the concentration of CXCL13 in the serum and cancer specimens are increased (29),(30). However, few studies have focused on the chemo-resistance and immune-inhibitory properties concerned with CXCL13 expression in tumor treatment. In the present study, we evaluated the therapeutic effects of cyclophosphamide or antibody targeting PD-1 in Cxcl13-/- mice and WT mice. The therapeutic effects of chemotherapy and checkpoint-based immunotherapy were both significantly increased in pulmonary tumor metastasis model of Cxcl13-/- mice while compared with that of WT control. These results indicated that targeted therapies against CXCL13 or IL-10-producing B cells might be a promising therapeutic strategy to enhance the efficacies of traditional chemotherapy or immunotherapy against advanced pulmonary cancer metastasis.