C57BL/6 WT and CCR6^−/− mice on a C57BL/6 background (B6.CCR6^−/−) were sourced and generated as described (13). Animal experiments were approved by the University of Tasmania Animal Ethics Committee. Immunization was carried out by injecting 1.0 × 10⁸ sheep red blood cells (SRBC) suspended in sterile PBS intraperitoneally.

Spleens from sacrificed mice were removed and homogenized in liquid nitrogen. Spleens were then added to 1 mL of TriReagent (Sigma-Aldrich, NSW, Australia) and RNA extraction carried out as per manufacturer’s instructions. cDNA was synthesized from total RNA using Tetro cDNA synthesis kit (Bioline, Alexandria, NSW, Australia). Template cDNA was stored at −20°C until further use. Quantitative real-time PCR was performed on cDNA templates using the SensiFAST SYBR No-Rox kit (Bioline). Results were normalized to β-actin (sense 5′- AGA GGG AAA TCG TGC GTG AC -3′ and anti-sense 5′- CAA TAG TGA TGA CCT GGC CGT -3′). The following primers were used: Ccr6 (sense 5′- TGT CCT CAC CCT ACC GTT CTG -3′ and anti-sense 5′- TAC AGG CCA GGA GCA GCA T -3′), and Bcl6 (sense 5′- CTG CAG ATG GAG CAT -3′ and anti-sense 5′- CGG CTG TTC AGG AAC -3′).

The following rat anti-mouse antibodies and conjugations were obtained from BioLegend (Australian Biosearch, WA, Australia), BD Biosciences (Sydney, NSW, Australia), or eBioscience (Sydney, NSW, Australia) and used for flow cytometry: B220-Biotin (clone RA3-6B2), CD19-APC Fire 750 (6D5), CCR6-PE (29-2L17), CCR6-AF647 (140706), CD11b-PerCP-Cy5.5 (M1/70), CD11b-BV510 (M1/70), CD4-APC (RM4-5), CD4-PerCP Cy5.5 (RM4-5), CD8α-PB (53-6.7), CXCR5-Biotin (2G8), CXCR5-PerCP-Cy5.5 (2G8), PD-1-PE (J43), PD-1-PE-Cy7 (J43), TCR-αβ-PB (HM3628, Thermo Fisher Scientific Australia, Soresby, VIC, Australia), hamster IgG₁-λ isotype-PE (G235–2356), and rat IgG₁-κ isotype-FITC (eBRG1). Cy5-conjugated streptavidin (Jackson Immuno Research, Pennsylvania, PA, USA) was used as secondary reagent. Unlabeled CCL20 (114906) was obtained from R&D Systems (Sydney, NSW, Australia) and labeled with DyLight 488 Microscale Antibody Labeling Kit (Thermo Fisher Scientific, Australia) according to the manufacturer’s instructions.

Murine spleens were dissected and pushed through a 40 µm nylon cell strainer to obtain a single cell suspension. After washing, the cells were resuspended in 10 mL of red blood cell lysis buffer and left to incubate at room temperature for 10 min. For cells undergoing intracellular cytokine staining, 0.5 µL of 200 µg/mL PMA (Sigma-Aldrich) and 0.5 µL of 10 mM ionomycin (Thermo Fisher Scientific, Australia) were added to a 5 mL resuspension of the cells in RPMI medium (Thermo Fisher Scientific, Australia) and were incubated at 37°C for 1 h, 5% CO₂. Following this, 1 µL of Golgi stop (BD Biosciences) (equivalent to 3.75 mM monensin) was added and the suspension incubated for further 3 h at 37°C.

Multicolor flow cytometry was performed on splenocytes using CyAn ADP and Gallios flow cytometers (Beckman Coulter, Inc., NSW, Australia). Post-acquisition analysis was performed using FlowJo v7.6 software (TreeStar, Inc., Ashland, OR, USA) and Kaluza 1.5a software (Beckman Coulter, Inc.). Fluorescent-activated cell sorting (FACS) was performed using the MoFlo Astrios cell sorter (Beckman Coulter, Inc.) with a sorting accuracy >95%.

For surface-bound chemokine studies, cells were blocked with Fc-receptor block (BD Biosciences) for 30 min on ice and antibody staining performed under sodium azide-free conditions. Fluorescent-minus-one and isotype controls were used to minimize non-specific binding and autofluorescence. To dissociate CCL20 from the surface cells were washed with acid as described (14). Briefly, cells were pelleted and resuspended in 200 µL ice-cold acid buffer for 2 min. Samples were neutralized with 5× volumes of R10 medium and cells were washed twice with FACS buffer (2% FCS, 0,01% NaN₃ in PBS). In the following experiments, Fc-block and surface staining were performed as indicated.

Statistical analyses using the non-parametric two-tailed Mann–Whitney U test were performed and graphs constructed using Prism software v5 (GraphPad Software). Graph columns represent median with error bars representing range. The alpha value was set at 0.05. In our figures these have been summarized as: *p < 0.05, **p < 0.01, ***p < 0.001.

CC chemokine ligand 20 is an inflammatory chemokine that is upregulated by the pro-inflammatory cytokine TNF (3) or induced by TGF-β and IL-6, or IL-21 during T_H17 cell differentiation (5). To examine whether CCL20 is involved in humoral immune responses, we immunized mice with SRBC, known to trigger strong GC responses. Splenocytes from unchallenged mice expressed virtually no CCL20 (Figure S1A in Supplementary Material). After immunization with SRBC splenocytes showed an upregulation of CCL20 mRNA in a time-dependent manner reaching approximately ninefold expression by day 7 (Figure S1A in Supplementary Material). Interestingly, Ccl20 has a very low-constitutive expression in the spleen, relative to other organs such as Peyer’s patches (15) (Adrian Y. S. Lee and Heinrich Körner, unpublished data) suggesting a subordinate role in splenic immune homeostasis. In fact, the rapid and marked elevation of Ccl20 post-immunization suggests it has a very specific role in the inflammatory response. Therefore, we examined the expression of CCL20 in this organ for subsequent experiments.

To identify possible sources for splenic CCL20, we sorted day 5 SRBC-immunized splenocytes and performed quantitative real-time PCR on major cellular subpopulations found in the spleen: B cells (B220⁺TCR-β¯), CD4⁺ T cells (B220¯TCR-β⁺CD4⁺), CD8⁺ T cells (B220¯TCR-β⁺CD8⁺), and miscellaneous CD11b⁺ cell populations (B220¯TCR-β¯CD11b⁺) which predominantly consists of NK cells, splenic macrophages, and granulocytes (4, 16). An earlier study (10) demonstrated very low-constitutive Ccl20 expression in the CD4⁺ T cell compartment of the spleen and no expression in the CD8⁺ T cell or monocytic compartments. Our study, however, demonstrated Ccl20 mRNA highly expressed in CD11b⁺ cells after activation (Figure S1B in Supplementary Material). In comparison, total splenic Ccl20 expression was low suggesting a marked dilution by non-CCL20-producing cells, most likely the B cells. This is in line with early expression studies that found overall very low amounts of constitutive splenic Ccl20 compared with other organs (15). However, we show a substantially upregulated expression induced upon immunization (Figure S1A in Supplementary Material) or challenge using bacterial lipopolysaccaride (data not shown).

CC chemokine receptor 6, the CCL20 receptor, modulates GC reactions and antibody affinity maturation, which appears to be regulated by follicular T helper (Tfh) cells (17). Thus, the CCR6/CCL20 axis might be involved in the GC response via Tfh cells (12) and we decided to focus our attention on CD4⁺ T cells, and specifically, Tfh cells. To initiate a strong humoral immune response, we used SRBC as a T cell-dependent antigen. We identified the Tfh population in spleen (Figure 1A) and sorted Tfh cells (defined as CD4⁺CXCR5^hiPD-1^hi) from SRBC-immunized mice (day 5 p.i.), an intermediate population Tfh^int (CD4⁺CXCR5^loPD-1^lo) and the remaining CD4⁺ T cells (non-Tfh cells; CD4⁺CXCR5⁻PD-1⁻) (Figure 1A). These non-Tfh cells comprise other T helper subsets (e.g., Th1), Tregs, and memory T cells including memory Tfh cells (18). The classic effector Tfh cells were significantly increased following immunization, confirming their activation (Figure S2 in Supplementary Material). To validate the quality of the isolated cell populations, we subjected sorted Tfh, Tfh^int, and non-Tfh cells to qRT-PCR for Bcl6, demonstrating a non-significant trend toward highest mRNA expression in the Tfh cells, moderate expression in the intermediary population, and lowest expression in the non-Tfh cells (Figure 1B). The fact that we detected more Bcl6 in sorted CD4⁺TCRb⁺CXCR5⁺PD-1⁺ cells corroborates their identity as Tfh as shown previously (19). Similarly, we also show that mRNA expression of Ccr6 tends to be highest in Tfh cells (Figure 1C), which corresponds to protein expression as determined by flow cytometry (Figure 1D).

The canonical CD4⁺CCR6⁺ T cell population has been attributed to Th17 cells (20, 21); however, we show approximately 90% of Tfh cells are positive for CCR6 (Figure 1E), consistent with our mRNA data for strong Ccr6 expression in this population (Figure 1C). We performed intracellular cytokine staining and show that of the Tfh^int population, a marked proportion of CCR6⁻ cells are CCL20⁺, whereas very little CCR6⁺Tfh^int cells are positive (Figure 1E). In contrast, the CCL20 expression is less marked in Tfh cells, being present in approximately 30% of cells (Figure 1E). The dominant intracellular CCL20 expression in the intermediary population (Figure 1F) is congruent to microarray data that demonstrated that Ccl20 is upregulated in intermediate population and down-regulated in Tfh cells (19). Interestingly, in a second study, CCL20 expression has also been described to be higher in Tfr cells than in regular Tfh cells (22).

Since Tfh cells are known to form conjugates with CCR6⁺ B cells, we examined surface expression of CCL20 on these cells. After eliminating autofluorescence (Figure S3 in Supplementary Material) and non-specific binding through isotype and fluorescence-minus-one controls, we show that compared with non-Tfh cells, approximately 4% of Tfh^int and Tfh cells show CCL20 on their surface (Figures 2A,B). To eliminate the possibility of the flow cytometric assay detecting bound CCL20 to CCR6, we used CCR6^−/− mice and repeated the experiments. CCR6^−/− mice similarly demonstrated surface CCL20 expression indicating that CCL20 is bound independent of CCR6 (Figure 2B). The percentage of T cells with surface-bound CCL20 did not differ significantly between parent populations of WT and CCR6^−/− mice (Figures 2B,C).

To validate our differential flow cytometric detection of intracellular versus surface CCL20, we analyzed CCL20 expression on in vitro-generated Th17 cells (23, 24). As expected, after activation, a major percentage of Th17 cells (circa 90% of Th17 cells) display a strong expression of intracellular CCL20 as published previously (5). However, additionally we were able to detect surface CCL20 on about 25% of Th17 cells (Figure S4 in Supplementary Material).

Chemokines are known to bind sulfated polysaccharides [glycosaminoglycans (GAGs)] found on cell surfaces (25). The presence of GAGs cause polymerization of chemokines, thereby increasing local concentration and facilitating chemokine-receptor binding (26). In this case, the surface-bound CCL20 likely functions to increase local concentration of the chemokine in order to create a concentration gradient for CCR6⁺ cells [haptotaxis (27)]—a likely candidate being partnering CCR6⁺ B cells to T cells as T–B cell conjugates.

The likely source of surface-bound CCL20 appears to be Tfh^int and Tfh cells as indicated by protein (Figures 1E,F) and mRNA expression studies (19, 22). However, it is equally possible that CCL20 originates from other sources and becomes tethered on their surfaces. The closely related CCL19 and CCL21, for example, could be detected on the luminal side of high-endothelial venules despite the fact that these cells do not seem to express mRNA for these chemokines. Instead, they rely on transcytosis of the ligands from the surrounding T cell zone, presenting them on their surface to attract CCR7⁺ T cells to secondary lymphoid organs (28, 29). Therefore, it is possible that the source of CCL20 is the GC as both CCR6 and CCL20 are found in GC in lymphoid tissue (30) (Adrian Y. S. Lee and Heinrich Körner, unpublished data). To test whether CCL20 on Tfh^int and Tfh cells is covalently expressed on the cell surface or bound via GAGs or others, we treated cells with an acid wash before staining. This experiment confirmed the increase in CCL20 expression proportional to Tfh differentiation, with acid wash leaving the frequency of CCL20 expressing cells unchanged (Figure 2C). Unexpectedly, acid wash treatment enhanced the intensity of CCL20 staining (Figure 2D). Careful examination of the histograms showed that acid treatment removed the low-/intermediate surface CCL20 surface expression but not the high-CCL20 expression in the CCL20 positive cell fraction (Figure 2E). Hence, high-CCL20 surface expression represents covalently bound CCL20 and intermediate/low-expression represents CCL20 bound to GAG or others, but likely not in an autocrine or paracrine fashion to CCR6 (see Figure 2B). These data infer that a cell population with a low frequency of CCL20⁺ cells, for instance at a certain differentiation stage such as non-Tfh cells, can eventually provide as many CCL20 ligand binding sites as a population with a higher frequency of CCL20 positive cells. Cells displaying acid-sensitive CCL20 might serve as a mobile source of CCL20, thereby, providing spatio-temporal information during immune responses. Hence, even non-Tfh cells could have the capability of interacting with CCR6-expressing B cells; for instance at the entry point of the GC (12).

To assess whether this surface-bound CCL20–CCR6 interaction is essential for the formation of T–B cell conjugates, we compared the formation of T–B cell conjugates in SRBC-immunized spleens from WT and CCR6^−/− mice by flow cytometry. By gating on doublets and CD4⁺CD19⁺ cells, we determined a surrogate measure of T–B cell conjugates (Figure 3A)—a method performed previously (31). We found a significantly higher proportion of CD4⁺CD19⁺ conjugates in CCR6^−/− mice compared with WT mice, indicating that the CCR6–CCL20 chemokine axis is not essential for the formation of these T–B cell conjugates (Figure 3B). This would be in accordance with the observation that mice deficient in CCR6 appeared to form more GCs in an accelerated fashion in the spleen than their WT counterparts (11). To determine which subpopulation of CCL20 expressing T cells (see Figures 2C,D) contributes to T–B-conjugate formation, we analyzed Tfh and Tfh^int populations cells in CD11b-negative singlets or aggregates (Figure 3A, gating strategy). Whereas Tfh cells represent only 6–10% of CD4^hiTCR^hi cells within singlets or T cell duplets (Figures 3A,C), they are highly enriched over non-Tfh or Tfh^int cells in T–B-conjugates (Figures 3D,E). The stability of the conjugates was addressed in control experiments using acid washing and treatment with EDTA. The proportion of T and B cells in conjugates was not changed by these treatments (data not shown).

Thus, although the presence of CCR6 regulates T–B-conjugate formation negatively (Figure 3B), CCL20⁺ Tfh cells are most enriched in T–B-conjugates. Whether the B cells interacting with CCL20^hi Tfh here represent light zone B cells that express CCR6 (17) remains to be tested. In this case, the CCR6–CCL20 axis could be responsible for maintenance of the pairing between Tfh and cognate light zone B cell, thereby, contributing to the selection of high-affinity B cells since CCR6^−/− mice display poorer antibody quality production (11, 12).

On the other hand, our data show that Tfh^int cells have a relatively high frequency of low-/intermediate CCL20 expressing cells (Figure 2E). It is known that T cells near the T–B border already express Bcl6 and are PD-1^lo, preceding a Bcl6⁺PD-1^hi phenotype (32), and therefore may correspond to our Tfh^int. This Tfh^int population appears to precede cognate B cell interaction (33) and appears to be a necessary step in the selection of B cells and hence, formation of a GC (34). This notion together with our data implicates a fine-tuned CCR6–CCL20 axis in this critical conjugation, which might be modulated by the presentation of non-covalently CCL20 on the surface. Thus, the CCL20–CCR6 axis likely functions at several GC checkpoints governing the efficacy of the GC reaction via B cells. It is tempting to speculate that two different mechanisms may act at the GC entry and at the GC selection point, depending on quantity and quality of surface-presented CCL20.

To conclude, we have shown that CCL20 is a chemokine that is produced by monocytes and T cells in the spleen. As subsets of the latter, Tfh and Tfh^int cells are producers of CCL20 which is also tethered on their surfaces, an observation we have validated in vitro using Th17 cells. A likely function of this CCL20 is in the role of coordinating cognate CCR6⁺ B cells to receive T cell help for the formation of antibodies. As optimal humoral immune responses rely on this critical T–B cell interaction, further research into the role of the CCR6–CCL20 chemokine axis in this conjugation will be helpful in understanding suboptimal immune responses and the pathogenesis of autoimmune disease.