Mice were purchased from the Jackson Laboratory (Bar Harbor, ME, USA) or Taconic. We used 8- to 12-week-old male and female wild-type BALB/c, C57BL/6J, or OT-I TCR transgenic mice on the RAG-1 knockout background. All animals were housed and handled according to National Institutes of Health institutional guidelines under an approved protocol by Case Western Reserve University Institutional Animal Care and Use Committee (No. 2012-0126 and 2015-0118).

For tumor measurements, mice were injected s.c. with 1 × 10⁶ of each tumor construct alone or mixed at 1:1 ratio (5 × 10⁵ each) in the left flank, and tumors were measured twice weekly. Tumor growth was measured using electronic calipers. The formula, V = π × D × d2, was used to calculate tumor volumes (16), where “D” is the largest diameter and “d” is the smallest diameter. Mice were sacrificed between days 21 and 50, depending on the experimental setup.

For gross LN measurements, mice received s.c. footpad injections of tumor cells (1 × 10⁶), and the popliteal TDLN and non-draining lymph node (NDLN) were removed on day 5 or 7. High-resolution pictures were taken of TDLN and NDLN from each group (NI, WTTU, L3TU, WTTU + αAsialo-GM1, or L3TU + αAsialo-GM1) prior to FACS analysis. The large and small diameters of the LNs were measured in Adobe Illustrator software. The formula, V = π × D × d2, was used to calculate LN volumes (16). In order to normalize measurements from different photo sessions, each experiment was divided by the average of the WTTU group within the same experiment, and the relative fold-change was calculated.

Antibodies were purchased from eBioscience, BD Pharmingen, or BioLegend and are as follows: rat anti-mouse CD4 FITC and PE (GK1.5), APC (RM4-5); CD8a FITC, PE, and APC (53-6.7); CD19 FITC, PE, and APC (1D3); CD11c FITC, PE, and APC (N418); CD49b PE and APC (DX5); CD3 PE and APC (145-2C11); CD103 FITC (2E7); PD-L1 PE and APC (10F.9G2); CD69 PE and APC (H1.2F3); and IFNγ APC (XMG1.2). Mice received s.c. footpad injections of either tumor constructs and the popliteal TDLN and NDLNs were removed 1, 3, 5, 7, or 10 days later. LNs were made into single-cell suspensions with ice-cold FACS buffer (0.5% FBS and 0.5% EDTA in 1× sterile PBS). For surface staining, unlabeled rat anti-mouse blocking Fc antibody was applied for 30 min on ice followed by primary antibody staining for 30 min on ice and protected from light. Cell viability test was conducted using 7-AAD (Biolegend), which was added 20 min into the primary staining. The samples were then washed twice with ice-cold FACS buffer and analyzed on an Accuri C6 flow cytometer. For intracellular IFNγ staining, cells were plated with 1 μl/ml of GlogiStop (eBioscience) for 6-h on a 96 well plate pre-coated with unlabeled anti-CD3 at 1μg/ml for a total of 90 min at 37°C prior to membrane permeabilization with Cytofix/Cytoperm (BD Biosciences), followed by staining with anti-IFNγ antibody. Analysis was performed using Accuri C6 and FlowJo software.

For NK cell deletion, mice received intraperitoneal (I.P.) injections of 50 μg of αAsialo-GM1 (Poly21460) in 100 μl of 1× PBS. One day later, mice received s.c. footpad injections of 1 × 10⁶ tumor cells. 5 days later, the popliteal TDLN and NDLN were removed for FACS analysis. For CCL3 neutralization, mice received I.P. injections of anti-CCL3 antibodies (50 μg/mouse; R&D System) concurrently with 1 × 10⁶ WTTU or L3TU. Anti-CCL3 antibody injection was repeated again 48 h later. The popliteal TDLN and NDLN were removed 5 days later for FACS analysis.

WTTU or L3TU were incubated for 24 h at 37°C in 95/5% O₂/CO₂ in 1 ml of complete media (RPMI 1640 with 10% fetal bovine serum, 1% HEPES, 1% non-essential amino acids, and 1% penicillin/streptomycin). The spent media from in vitro cultures or serum samples obtained from tumor-bearing mice were quantified for CCL3 protein contents by ELISA in accordance to the manufacture’s protocol (R&D systems, MMA00).

CT26 tumor cells were stably transfected with a PCDNA3.1 plasmid vector that contains the mouse CCL3 cDNA and maintained under Hygromycin (150 μg/ml) selection.

Bone marrows from C57BL6 mice were isolated and bone marrow-derived DCs (BMDCs) were generated in complete media at 3 × 10⁶ cells/3 ml/well in 6-well tissue culture plates supplemented with 15 ng/ml of granulocyte-macrophage colony-stimulating factor (GM-CSF) and 10 ng/ml interleukin-4 (IL-4) on days 0, 3, and 5. On day 3, the medium was removed and fresh media plus cytokines were added at 3 ml/well. On day 5, the cultures were replaced with fresh media plus cytokines. On day 7, non-adherent immature BMDCs were collected, washed with complete media and plated in a 6-well tissue culture plate with or without 100 ng/ml of CCL3 for 24 h. BMDCs were then collected and pulsed at 37°C in 95/5% O₂/CO₂ with various doses of SIINFEKL-peptide for 1 h, washed and cultured in triplicates with 100,000 CFSE-labeled (1 μM) naïve OT-I cells (at DC-to-OT-I ratio of 1:5) at 37°C in 95/5% O₂/CO₂ for 72 h. The percent of CFSE-dilution peaks, relative to non-pulsed and cultured BMDCs and OT-I cells, was calculated using FACS.

Total LN mRNA was isolated using TRIzol reagent in accordance with the manufacturer’s protocol (Gibco BRL, Carlsbad, CA, USA) and purified using an Illustra™ RNAspin Mini Kit (GE Healthcare Life Sciences). RNA quality was assessed by spectrophotometer absorption at 260/280 nm using the NanoDrop2000 spectrophotometer. RNA was converted to cDNA using EasyScript™ Reverse Transcriptase protocol consisting of 200 U/μl Moloney murine leukemia virus reverse transcriptase incubated for 60-min at 42°C in the presence of 50 mM Tris-HCl (pH 8.3), 100 mM NaCl, 0.1 mM EDTA, 5 mM DTT, 0.1% Triton X-100, 50% (v/v) glycerol, 10 μM of oligo (dT), 10 mM 29-deoxynucleoside 59-triphosphate, and 40 U/μl recombinant RNase inhibitor (Lamda BIOTECH, St. Louis, MO, USA). cDNA was amplified in the presence of FAM-labeled gene-specific primers and Bullseye EvaGreen qPCR Mastermix (MIDSCI™; Saint Louis, MO, USA) in a 96 well microtiter plate using the ABI Prism 7300 sequence detection system (Applied Biosystems). Each PCR was performed in triplicate and compared to WTTU. Relative levels of mRNA were determined using the cycle threshold (C_t). The gene expression was standardized according to cytochrome-c (CyC) expression within the TDLN. In order to compare the C_t values between target genes, we normalized each C_t to the average of the WTTU C_t using the following equation: 2^^{− (Target gene − CyC − target normalizer)}.

Significance analyses were performed using the standard t-test. Figures 1 and 2 were displayed as SEM for ease of visualization. SDs were shown for other figures.

We first constructed L3TU by stably transfecting WTTU CT26 with plasmids containing murine CCL3. No significant differences in the growth kinetics between WTTU and L3TU were observed in vitro (Figure S1A in Supplementary Material). ELISA analysis revealed that L3TU produced ~350 pg/ml of CCL3 per 1 × 10⁶ tumor cells in vitro over 24 h, whereas WTTU failed to secrete any detectable CCL3 (Figure S1B in Supplementary Material). Serum obtained from mice 7 days after 1 × 10⁶ L3TU inoculation contained ~150 pg/ml CCL3, while serum from non-injected (NI) and WTTU-injected mice had negligible (< 8 pg/ml) CCL3 (Figure S1C in Supplementary Material). Both L3TU and WTTU expressed similar surface programmed death-ligand 1 (PD-L1) at baseline and following IFNγ stimulation in vitro (Figures S1D,E in Supplementary Material). Next, we measured the tumor growth behavior in naïve BALB/c mice inoculated subcutaneously (s.c.) with 1 × 10⁶ WTTU, L3TU, or WTTU + L3TU mixture at 1:1 ratio (5 × 10⁵ each) in the flank. While WTTU grew aggressively in 100% of the recipient mice and became measureable by day 7 with an average tumor volume of ~12,000 mm³ after 3 weeks, 30% of mice in L3TU group and 20% of mice in the WTTU + L3TU group completely rejected the tumors with the remaining clinically evident tumor sizes being 25- and 6-fold smaller than WTTU, respectively (Figures 1A,B). A 20-fold reduction in the L3TU inoculum (5 × 10⁵ cells) resulted in a 40% tumor-free incidence (Figures 1C,D). As expected, mice that rejected primary L3TU tumors were capable of rejecting subsequent challenge with a lethal dose of WTTU, demonstrating the successful generation of antitumor immune memory (data not shown).

Studies by Gretz and colleagues showed that rCCL3 administered s.c. in the footpad could readily drain though the afferent lymphatic ducts to associate within the high endothelial venules (HEVs) (17). Therefore, we examine how CCL3 produced by L3TU influence cellular traffic to the TDLN during the early phase (≤10 days) of immune priming following tumor inoculation. We examined the TDLN cellularity of CD4⁺ T cell, CD8⁺ T cells, CD19⁺ B cells, CD11c⁺ pAPCs, and CD49b⁺ cells. Both WTTU and L3TU inoculation led to significant increases in CD4⁺, CD8⁺, CD19⁺, CD11c⁺, and CD49b⁺ leukocytes in the TDLN compared to NI group (Figure 2). However, TDLNs draining L3TU showed 2- to 6-fold increases in total leukocyte sub-populations compared to WTTU TDLNs, with the enhancement of leukocyte accumulation reaching statistical significance after 3 days and peaking on days 5 and 7 (Figures 2A–E). These changes were accompanied by gross anatomical differences in LN sizes [Figure S2 in Supplementary Material subcutaneously (s.c.)]. These striking changes were not caused solely by CCL3 alone, as mice receiving direct s.c. injections of rCCL3 showed only transient (<2 days) changes in leukocyte accumulation in TDLN that were similar to PBS controls (Figure 2H). With the exception of CD4⁺ T cells, leukocyte numbers in both WTTU and L3TU TDLN returned to NI levels by day 10. Next, we interrogated whether the significant increases in T cell numbers correlated with changes in the T cell activation status. Mice were injected similarly as above, and both TDLN and contralateral non-draining lymph nodes (NDLNs) were removed on days 1, 3, 5, and 7 for CD69 expression assessment by flow cytometry (Figures 2F,G). In both CD4⁺ and CD8⁺ T cell subsets, we detected greater numbers of CD69⁺ T cells in L3TU TDLN. The number of CD69⁺ T cells began to dissipate after day 5. Interestingly, we also observed a significant and reproducible increase in leukocyte accumulation in the NDLNs on day 5, suggesting a systemic effect of CCL3 on immune cell mobilization and trafficking (Figure S3 in Supplementary Material). However, this effect was transient and quickly dissipated after day 5, and the transient cellularity increase was not accompanied by CD69⁺ T cell activation (Figures S3F,G in Supplementary Material), suggesting a requirement for the presence of tumor cells to sustain the accumulation of CD69⁺ T cell subset in TDLN. Despite differences in absolute cellularity of TDLN and NDLNs between WTTU and L3TU, the overall cellular compositions were similar between the two tumors. Normally, CD4⁺ T cells predominates among LN leukocytes in the naive mouse; however, B cells became the most prominent population proportionally within TDLN 3 days after tumor injections, and the trend continued until after day 7 when the relative composition returned to baseline (Figure S4 in Supplementary Material). The rate of composition reversal by day 10 in TDLN and NDLN between CD4+ T cells and B cells appeared to be more dramatic in L3TU compared to WTTU.

Next, we asked whether the enhanced number of activated CD69+ T cells would correlate with an anti-tumorigenic milieu in TDLN. To address this, following tumor inoculation we measured global cytokine mRNA differences, including TGFβ, TNFα, IL-10 and IFNγ, between WTTU and L3TU groups on day 5, the time point where we observed the greatest differences in cellular accumulation and CD69 positivity. While the average transcript levels of TGFβ, TNFα, and IL-10 in TDLN were similar between L3TU and WTTU, TDLN in L3TU contained a ~2.5-fold greater IFNγ mRNA transcript level as compared to that in WTTU (Figure S5 in Supplementary Material). We then examined intracellular IFNγ content among the cells found within TDLN on day 5. With the exception of CD8+ T cells, all other cell-types analyzed contained significantly elevated numbers of cells expressing intracellular IFNγ in L3TU TDLN (Figure 3). This finding was further confirmed by ELISPOT analysis (data not shown).

By day 5, CD49b⁺ cells not only make up the cell-type with the most significant difference between WTTU and L3TU, but they also constituted the most statistically significant IFNγ+ population differences (Figures 2E and 3F). CD49b (VLA-2) is an integrin protein expressed on NK cells, and a subpopulation of NK T (NKT) cells, CD4+ and CD8+ cells (18, 19). NK, NKT, and CD8+ cells have been associated with tumor rejection and IFNγ upregulation after stimulation (20–22). Recently, CD49b+ CD4+ cells were defined as a subpopulation of regulatory T cells that produced IL-10 in response to stimulation rather than IFNγ (23). We hypothesized that the majority of the global IFNγ production within TDLN on day 5 is derived from activated CD49b+ CD3− NK cells, which express CCR5 and could potentially be recruited directly by CCL3 to the TDLN (22). Under homeostatic conditions, NK cells represent a small population (≤1%) in the LN but can accumulate and supply significant IFNγ to drive T cell activation and differentiation upon stimulation (Figure S6 in Supplementary Material) (24). We examined the dependence of cellular accumulation in the L3TU TDLN on CCL3 and NK cells by administering anti-CCL3-neutralizing monoclonal Abs (mAbs) or NK-depleting αAsialo-GM1 antibody prior to footpad inoculation with WTTU or L3TU (Figure S6 in Supplementary Material). We then examined the cellularity of TDLN and NDLN on day 5 as previously described (Figure 4). Blocking CCL3 resulted in diminished cellular accumulation in both TDLN and NDLN (Figure 4 and data not shown). While NK cell depletion over the course of 5 days was associated with a decrease in global IFNγ expression in TDLN (Figure S5 in Supplementary Material), however, counter to expectation, NK cell depletion also potentiated the enhanced cellular accumulation observed with aCCL3 (Figure 4).

Interferon-gamma expression at primary tumor sites can drive DC maturation and rapid migration to DLN for Ag-presentation (25). IFNγ has also been shown to induce the production of CXCL9 and CXCL10, which can aid in tumor rejection through the recruitment of activated CD8⁺ T cells and IFNγ-secreting Th1 cells to the primary tumor sites (26–28). Recently, studies have shown that CD103⁺ CD11c⁺ DCs, a subpopulation of dermal- and gut-resident pAPCs, respond to tumor-derived CCL4 and are the chief cells that produce CXCL9 and CXCL10 to recruit tumor-infiltrating T cells (29, 30). Indeed, we observed an accumulation of CD103⁺ CD11c⁺ DCs in L3TU TDLN, and such enhanced accumulation was dependent on CCL3, not NK cells or associated IFNγ (Figures 5A,B; Figure S5 in Supplementary Material). However, we show that TDLN NK cells are crucial for CXCL9 and CXCL10 production on day 5 (Figures 5C,D), as NK depletion resulted in a dramatic decrease in the mRNA transcripts of both chemokines.

Finally, we examined the immune-modulatory effects of CCL3 on pAPC function. Several reports have observed the direct modulatory effect of CCL3 on pAPC function. Watanabe and colleagues showed that macrophages pretreated with CCL3 exhibit strengthened adhesion to osteoblasts leading to the formation of pre-osteoclast cells in vitro, an important step in the process of bone reabsorption (31). Previously, we showed that blocking CCL3 (and its paralog CCL4) in vivo decreased CD8⁺ T cell and DC contacts in the vaccine-draining LN and diminishes the magnitude of the overall CD8⁺ T cell memory pool (5). In two separate studies, Park and colleagues showed that pretreatment of DCs with CCL3 in combination with CCL19 and LPS stimulation led to enhanced OVA-specific CD4⁺ (OT-II) T cell proliferation in vitro. To address a potential functional modulatory role of CCL3 on pAPC, we pretreated BMDCs from C57BL/6 mice with rCCL3 for 24 h, then co-cultured them with OT-I T cells for 3 days with and without Ags in vitro. CCL3-conditioned BMDC displayed a significantly enhanced capacity to drive OT-I proliferation when pulsed with SIINFEKL peptides at doses of 0.1–10 μg/ml (Figure 6A). Interestingly, CCL3-conditioned BMDC also exhibited enhanced cross-presentation capacity to drive OT-I proliferation when cultured with whole OVA-coupled beads, especially at low Ag doses of 0.1–1 ng/ml (Figure 6B).