Peptidoglycan‐treated tumor antigen‐pulsed dendritic cells impart complete resistance against tumor rechallenge

Summary

Solid tumors elicit suppressive T cell responses which impair antigen‐presenting cell (APC) functions. Such immune suppression results in uncontrolled tumor growth and mortality. Addressing APC dysfunction, dendritic cell (DC)‐mediated anti‐tumor vaccination was extensively investigated in both mice and humans. These studies never achieved full resistance to tumor relapse. Herein, we describe a repetitive RM‐1 murine tumor rechallenge model for recurrence in humans. Using this newly developed model, we show that priming with tumor antigen‐pulsed, Toll‐like receptor (TLR)2 ligand‐activated DCs elicits a host‐protective anti‐tumor immune response in C57BL/6 mice. Upon stimulation with the TLR2 ligand peptidoglycan (PGN), the tumor antigen‐pulsed DCs induce complete resistance to repetitive tumor challenges. Intra‐tumoral injection of PGN reduces tumor growth. The tumor resistance is accompanied by increased expression of interleukin (IL)‐27, T‐box transcription factor TBX21 (T‐bet), IL‐12, tumor necrosis factor (TNF)‐α and interferon (IFN)‐γ, along with heightened cytotoxic T lymphocyte (CTL) functions. Mice primed four times with PGN‐stimulated tumor antigen‐pulsed DCs remain entirely resistant to repeat challenges with RM‐1 tumor cells, suggesting complete prevention of relapse and recurrence of tumor. Adoptive transfer of T cells from these mice, which were fully protected from RM‐1 rechallenge, confers anti‐tumor immunity to syngeneic naive recipient mice upon RM‐1 challenge. These observations indicate that PGN‐activated DCs induce robust host‐protective anti‐tumor T cells that completely resist tumor growth and recurrence.

To address APC dysfunction in various tumor malignancies and recurrence, we repetitively challenged mice with RM‐1 tumor cells after immunization with PGN‐activated tumor antigens‐pulsed DCs. The study reveals that twice or four times administration of such DC formulation enhance host‐survival by inducing host‐protective, adoptively transferable, robust anti‐tumor T cells. The mice also remain resistant to tumor rechallenge, instilling a protocol that removes a major bottleneck in anti‐tumor immunotherapy impeding the relapse and tumorigenesis.

Introduction

Cells with tumorigenic potential continuously emerge, but these cells are eliminated by the immune system. Two major cell types that mediate this anti‐tumor immune surveillance are dendritic cells (DCs) and T cells. The DCs acquire, process and present tumor antigens to tumor antigen‐specific T cells, which mediate anti‐tumor functions [1, 2]. However, generating DCs, which would elicit a robust T cell response such that the primed individual remains resistant to repeat tumor challenges, remains a bottleneck in DC‐mediated anti‐tumor therapy and prophylaxis. While identifying the DC phenotype that would elicit robust, long‐lasting, host‐protective anti‐tumor T cells failed, the emerging complexities of DC heterogeneity only exaggerated the problem. Therefore, we adopted an alternative approach, which was to activate the DCs – irrespective of their phenotypes – that efficiently elicit robust tumor‐eliminating T cells that remained host‐protective even in the face of repeated tumor challenges. The repeated tumor challenge protocols were developed to mimic the phenomena of tumor relapse and recurrence.

We treated the tumor antigen‐pulsed DCs with different Toll‐like receptor (TLR) ligands for priming the C57BL/6 mice, followed by challenge with prostate tumor cells RM‐1, a C57BL/6‐derived tumor [3]. TLR activation enhances the co‐stimulatory and antigen‐presenting capabilities of antigen‐presenting cells (APCs) bridging the innate and adaptive arms of the immune system [4]. Following the cue, poly I:C (TLR3‐ligand), imiquimod (TLR7/8‐ligand) and cytosine–phosphate–guanine (CpG) (TLR9‐ligand) were used to treat tumor antigen‐pulsed DCs, which were subsequently used for priming mice in melanoma models [5, 6, 7, 8]. Apart from these interventions, several studies have shown beneficial effects of synthetic TLR2 ligands in association with novel tumor antigens, radiotherapy or chemotherapy [9, 10, 11]. Some of them have shown adverse effects when injected systemically [12, 13]. By contrast, we showed that the TLR1/2 ligand {N‐palmitoyl‐S‐[2,3‐bis(palmitoloxy)‐(2RS)‐propyl]‐Cys‐Ser‐Lys4} (Pam₃CSK₄) and the TLR2/6 agonist S‐[2,3‐bispalmitoyiloxy‐(2R)‐propyl]‐R‐cysteinyl‐amido‐monomethoxyl polyethylene glycol (BPP) elicited counteractive immune responses in an infection model [14]. We observed that Pam₃CSK₄ elicited a strong T helper type 2 (Th2) response, whereas BPP elicited a strong Th1 response in a mouse Leishmania major infection model [14]. Tumor regression requires a strong Th1 response, whereas tumor growth is promoted by a Th2 response [15]. We examined whether or not these ligands play an anti‐tumor immunomodulatory role in a prostate tumor model [3, 16, 17]. In this model, bone marrow‐derived DCs from IL10^−/− mice are pulsed with irradiated RM‐1 cells and subcutaneously (s.c.) transferred into C57BL/6 mice, as RM‐1 cells were derived from a tumor in C57BL/6 mice. It was observed that such DC‐mediated antigenic priming reduced tumor growth upon challenge [16]. Herein, we tested whether treatment of the RM‐1 tumor antigen‐pulsed DCs with TLR2 ligands – Pam₃CSK₄, peptidoglycan (PGN) or BPP – would enhance the host‐protective priming efficiency of the DCs. We observed that PGN‐activated DCs elicited T cells in primed C57BL/6 mice, rendering them resistant to repeat challenges with RM‐1 prostate tumor cells. Transfer of these T cells to naive C57BL/6 mice converted the recipients entirely resistant to tumor challenge. We report a novel finding that priming with the PGN‐activated DCs, irrespective of their phenotype, imparts complete resistance to tumor rechallenge.

Materials and methods

Tumor model and TLR‐ligand treatment

RM‐1 prostate tumor cells, maintained in Ham’s F12K complete medium (gibco BRL, Grand Island, NY, USA) were injected into C57BL/6 mice [The Jackson Laboratory; Bar Harbor, ME, USA; s.c., 2×10⁵ RM‐1 cells/50 μl phosphate‐buffered saline (PBS)]. On days 5, 7 and 9 after RM‐1 cell injection, endotoxin‐free Pam₃CSK₄ (1 μg; InvivoGen, San Diego, CA, USA), BPP (1 μg; HZI, Braunschweig, Germany) or PGN (5 μg, InvivoGen) was injected intratumorally. Tumor volumes (mm³) on the day of injection were 12·37 ± 3·300 (day 5), 75·21 ± 16·554 (day 7) and 198·41 ± 46·183 (day 9). One group of mice received PBS as a control. The tumor size was measured using Vernier calipers on alternate days; tumor volume was calculated as (length × width × width)/2 [18]. Twenty‐one days later, tumors were weighed and photographed under a stereomicroscope (12×, Stemi DV4; Carl Zeiss, Hamburg, Germany). All experiments were performed according to the animal use protocols approved by the Institutional Animal Care and Use Committee.

DC generation

DCs were derived from bone marrow (BM) progenitor cells [16]. In brief, femur cells were harvested, fractionated on Histopaque^®‐1077 (Sigma, St Louis, MO, USA) and cells from the interface were washed twice with sterile PBS. The cells were seeded at a density of 1 × 10⁶ cells/ml/well in a 24‐well plate and adhered cells were maintained in DC culture medium [RPMI‐1640, 10% fetal calf serum (FCS), 100 U/ml penicillin, 100 µg/ml streptomycin, 50 µM β‐ mercaptoethanol (ME), 20 ng/ml granulocyte–macrophage colony‐stimulating factor (GM‐CSF) and 10 ng/ml interleukin (IL)‐4 (BD PharMingen, San Diego, CA, USA)]. Culture medium was replaced every 3 days. On day 6, dislodged cells were used as BM‐derived DC (BMDC).

Tumor antigen pulsing of DCs and immunization

BMDCs were pulsed with irradiated tumor cells (10 000 rad) at a ratio of 3 : 1 for 16 h together with different TLR2 ligands: Pam₃CSK₄ (1 µg/ml), PGN (5 µg/ml) or BPP (1 µg/ml). For immunization, pulsed DCs were collected, extensively washed and 1 × 10⁶ cells were injected s.c. at 7‐day intervals. As controls, pulsed DCs incubated without a TLR ligand were injected into the medium DC group (Med‐DC). One week after the last immunization, mice were challenged with 2 × 10⁵ RM‐1 tumor cells. Unprimed mice challenged with RM‐1 cells were considered as the control group.

Quantitative‐polymerase chain reaction (q‐PCR) assay

Total RNA isolated using TRI reagent (Sigma‐Aldrich, St Louis, MO, USA) was used for first‐strand cDNA synthesis using the Thermoscript reverse transcription–polymerase chain reaction (RT–PCR) system (Invitrogen Life Technologies, Carlsbad, CA, USA) [14], as per the manufacturer’s instructions. Quantitative PCR (q‐PCR) was performed in duplicate in 0·1 ml MicroAmp fast reaction strip tubes (Applied Biosystems, Foster City, CA, USA) in a 10µl reaction mixture containing 10 ng cDNA, 2 ng forward primer, 2 ng reverse primer and 2× SYBR Premix Ex Taq II (5 µl; Takara, Kusatsu, Shiga, Japan). The sequences of the forward and reverse primers are given in Table 1. qPCR was performed on StepOnePlus (Applied Biosystems): 95°C for 30 s, 40 cycles of 95°C for 5 s and 60°C for 35 s. Relative quantitation was computed using the comparative threshold (ΔΔC_t) method. mRNA expression levels of the target genes were normalized against glyceraldehyde 3‐phosphate dehydrogenase (GADPH) levels and expressed as relative fold change compared with untreated controls.

Table 1.

List of reverse and forward primers used for qPCR analysis

S. no.	Gene	Forward primer	Reverse primer
1.	GAPDH	5′‐ATTGTCAGCAATGCATCCTG‐3′	5′‐ATGGACTGTGGTCATGAGCC‐3′
2.	IL‐10	5′‐AACATACTGCTAACCGACTCC‐3′	5′‐TCCTTGATTTCTGGGCCATG‐3′
3.	IL‐12p40	5′‐ CCTGAAGTGTGAAGCACCAA‐3′	5′‐ AGACAGAGACGCCATTCCA ‐3′
4.	TLR1	5′‐CAATGTGGAAACAACGTGGA‐3′	5′‐TGTAACTTTGGGGGAAGCTG‐3′
5.	TLR2	5′‐AAGAGGAAGCCCAAGAAAGC‐3′	5′‐CGATGGAATCGATGATGTTG‐3′
6.	TLR6	5′‐TTGTTTGCGCCCTGGCCTTAAT‐3′	5′‐TGTTCTTGGTGGCAGGTCTTTG‐3′
7.	IL‐4	5′‐GGTGTTCTTCGTTGCTGTGA‐3′	5′‐TCTCGAATGTACCAGGAGCC‐3′
8.	IL‐27p28	5′‐CAAAGGAGGAGGAGGACAA‐3′	5′‐CACTTGGGATGACACCTGA‐3′
9.	EBI3	5′‐TGCCATGCTTCTCGGTAT‐3′	5′‐TAAGTGGCAATGAAGGACG‐3′
10.	IL‐6	5′‐CTTGGGACTGATGCTGGTGA‐3′	5′‐TCACCAGCATCAGTCCCAAG‐3′
11.	IL‐1β	5′‐CAGGCAGGCAGTATCACTCA‐3′	5′‐TGGGAACGTCACACACCAG‐3′
12.	T‐bet	5′‐GTTTCTACCCCGACCTTCCA‐3′	5′‐TGGAAGGTCGGGGTAGAAAC‐3′
13.	TNF‐α	5′‐ATGAGCACAGAAAGCATGA‐3′	5′‐ATCATGCTTTCTGTGCTCAT‐3′
14.	IFN‐γ	5′‐GGCTGTTTCTGGCTGTTACTG‐3′	5′‐GTTGCTGATGGCCTGATTGT‐3′
15.	TLR‐9	5′‐ACTGAGCACCCCTGCTTCTA‐3′	5′‐AGATTAGTCAGCGGCAGGAA‐3′

qPCR =quantitative polymerase chain reaction; GAPDH = glyceraldehyde 3‐phosphate dehydrogenase; IL = interleukin; TLR = Toll‐like receptor; EBI3 = Epstein–Barr virus gene 3; T‐bet = T‐box transcription factor TBX21; TNF = tumor necrosis factor; IFN = interferon.

Flow cytometric analysis

Anti‐CD3‐phycoerythrin‐cyanin 7 (PE‐Cy7) (552774), anti‐CD8‐fluorescein isothiocyanate (FITC) (553031), anti‐tumor necrosis factor (TNF)‐α biotinylated (554415), allophycocyanin (APC) streptavidin (554067), anti‐CD11b‐PE‐Cy7 (552850), anti‐major histocompatibility complex (MHC)‐II‐PE (553552), anti‐CD80‐FITC (553768), anti‐CD86‐FITC (561962) and anti‐IFN‐γ‐PE (554412) antibodies were purchased from BD Pharmingen. Anti‐TLR1‐eF660 (50‐9011‐82) antibody was purchased from eBioscience (San Diego, CA, USA). Anti‐TLR2‐FITC (309707), anti‐CD11c‐PerCP‐Cy5.5 (117328) and anti‐F4/80‐APC‐Cy7 (123118) antibodies were purchased from BioLegend (San Diego, CA, USA). Anti‐TLR6 (SC‐30001) and anti‐rabbit immunoglobulin (Ig)G‐FITC (SC‐2359) antibodies were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). For multi‐color fluorescence activated cell sorter (FACS) analyses, cells from the lymph node and tumor were stained (after blocking with 10% FCS) with fluorescently labeled anti‐CD3, anti‐CD8, anti‐CD11b, anti‐CD11c, anti‐F4/80, anti‐MHC‐II, anti‐CD80, anti‐CD86, anti‐TLR1, anti‐TLR2 and anti‐TLR6 antibodies for 45 min at 4˚C in the dark and washed twice with FACS buffer (1× PBS, 10 mM HEPES buffer and 3% FCS). Intracellular cytokine staining for TNF‐α and IFN‐γ was performed using a Cytofix/Cytoperm‐Plus Kit with GolgiPlug (BD Pharmingen). Cells were acquired on a FACS Canto‐II flow cytometer (BD Biosciences) and analyzed in the CD3⁺CD8⁺ gate for CD8 T cells, CD11b⁺F4/80⁺ gate for macrophages and CD11c⁺MHC‐II⁺ gate for DCs using BD FACSDiva software (version 5.2; BD Biosciences).

Cytotoxic T lymphocyte (CTL) cell assay

For CTL assay [18], splenocytes were plated at 1·5 × 10⁷ cells/well in a six‐well plate containing 1·5 × 10⁶ irradiated RM‐1 cells. Five days later, viable CD8⁺ T cells were isolated using enrichment cocktail and were plated in triplicate against [³H]‐TdR‐incorporated target RM‐1 cells at a ratio of 6 : 1. The cytolytic activity was tested by a standard 4‐h [³H]‐thymidine release assay (JAM test).

Adoptive T cell transfer

Using a FACS sorter (Aria III SORP; BD Biosciences), total CD4⁺ and CD8⁺ T cells were purified from the spleen of the four times DC‐immunized, RM‐1 tumor cell‐injected mice, washed with Hanks’s balanced salt solution (HBSS) and pulled at a ratio of 3 : 1. 2 × 10⁶ T cells were intravenously injected into each of the recipient mice. These mice were next injected with RM‐1 cells (2 × 10⁵ cells/mouse).

AngioQuant analysis

On day 21 after tumor cell inoculation, tumor‐bearing mice were euthanized, opened carefully and tumor images were captured using a stereomicroscope (Stemi DV4; Carl Ziess). The number of blood vessels was quantified using AngioQuant (AngioQuant Freeware; www.cs.tut.fi/sgn/csb/angioquant/) [18].

Statistical analyses

Statistical analyses were performed using the SigmaPlot version 10.0. Statistical differences in mean values between one or more groups were determined by Student’s t‐test and one‐way analysis of variance (anova) (the Holm–Sidak method). The error bars show mean ± standard error of the mean (s.e.m.) unless indicated otherwise. Values of *P < 0·05, **P < 0·01 and ***P < 0·001 were considered statistically significant.

Results

Intra‐tumoral injection of the TLR2 ligand reduces tumor growth

TLR activation has been shown to mount immune responses against tumor cells [5, 6, 8]. However, tumor growth may also affect the expression of TLRs on tumor‐infiltrating immune cells, and such altered TLR expression may promote a protumor T cell response facilitating tumor development. Therefore, it was necessary to examine the alteration in TLR expression profile in a growing tumor. Using our previous RM‐1 tumor model [16], we examined the kinetics of TLR1, TLR2 and TLR6 expression in tumor tissues on different days after tumor cell inoculation. Tumor volume and tumor weight steadily increased during the growth of the RM‐1 tumor in C57BL/6 mice (Fig. 1a), accompanied by significantly increased TLR1 but diminished TLR2 and TLR6 transcripts in tumors (Fig. 1b). TLRs are mainly expressed by the APCs such as macrophages and DCs. Therefore, a single‐cell suspension was prepared from the tumor and examined for the tumor‐associated macrophages and DCs by FACS analysis. Similar expression profiles were observed in tumor‐infiltrating CD11c⁺ DCs, where percentage and mean fluorescence intensities (MFI) of TLR1 increased, but the MFI for TLR2 and TLR6 decreased (Fig. 1c,d); CD11b⁺ tumor‐associated macrophages (TAMs) did not show similar alterations (Supporting information, Fig. S1a,b). These data suggest tumor‐induced selective alterations in TLR1, TLR2 and TLR6 expression in tumor‐infiltrated DCs and provide the scientific rationale for targeting TLR for activating DCs.

Fig. 1.

Intratumoral administration of Toll‐like receptor (TLR)2 ligands reduces tumor growth. (a) The size and volume of each tumor were measured (n = 5). After euthanizing the mice, tumors were excised (n = 5) and weighed. (b) Total RNA from tumor tissue was isolated and the relative expression of TLR1, TLR2 and TLR6 during the progression of tumor was assessed by quantitative–polymerase chain reaction (q‐PCR). (c) On days 7, 14 and 21, tumors were collected, crushed and stained for TLR1, TLR2 and TLR6 expression on CD11c⁺major histocompatibility complex (MHC)‐II⁺ dendritic cells (DCs). Fluorescence activated cell sorter (FACS) histograms represent the percentage and mean fluorescence intensity (MFI) of TLR1, TLR2 and TLR6 positive DC infiltrates. Gating strategy and appropriately matched isotype controls are shown at the top. (d) Bar graphs represent the mean ± standard deviation (s.d.) of percentage (left) and MFI (right) of TLR1, TLR2 and TLR6 positive DCs. (e) Intratumoral injection of {N‐palmitoyl‐S‐[2,3‐bis(palmitoloxy)‐(2RS)‐propyl]‐Cys‐Ser‐Lys4} (Pam₃CSK₄) (P3C), peptidoglycan (PGN) or S‐[2,3‐bispalmitoyiloxy‐(2R)‐propyl]‐R‐cysteinyl‐amido‐monomethoxyl polyethylene glycol (BPP), the respective ligands for the TLR1/TLR2, TLR2/TLR2 and TLR2/TLR6 dimers, reduces the tumor growth. The tumor volumes were measured and the collected tumors (n = 5) were weighed (inset). (f) The photographs were taken using a stereomicroscope at 12× magnification to visualize angiogenesis and tumor size. (g) Graphical representation of the number of tubule complexes quantified by AngioQuant. The experiments were repeated three times, and the data from one representative experiment are shown. The values shown are mean ± standard error of the mean (s.e.m.). *P < 0·05; **P < 0·01; ***P < 0·001.

Next, we treated tumors with the ligands for TLR1, TLR2 and TLR6 to check their effects on tumor pathogenesis. Intratumoral injection of Pam₃CSK₄, PGN or BPP – the ligands for the TLR1, TLR2 and TLR6, respectively – revealed that PGN was the most effective anti‐tumor TLR2 ligand, as PGN treatment reduced the tumor size by three‐fold in comparison to the untreated group (Fig. 1e). Tumor images (Fig. 1f) were examined for tumor size and angiogenesis. We also quantified the number of blood vessels using AngioQuant and observed fewer tubules in the TLR‐treated groups compared to the control group (Fig. 1g). These findings suggest that PGN treatment was the most efficacious, as it significantly reduced tumor size, tumor volume and angiogenesis.

Twice‐immunization with PGN‐treated tumor antigen‐pulsed DCs prevents tumor growth

BMDCs were pulsed with RM‐1‐derived tumor antigens and treated with Pam₃CSK₄, PGN or BPP, followed by s.c. transfer to C57BL/6 mice on days 1 and 7. On day 14, these mice received 2 × 10⁵ RM‐1 tumor cells. PGN‐treated DCs significantly prevented tumor growth, as assessed by reduced tumor volume (Fig. 2a) and tumor weight (Fig. 2b) compared to the Pam₃CSK₄‐ or BPP‐treated DC recipients. Reduced tumor volume was accompanied by reduced angiogenesis (Fig. 2c,d). Reduced numbers of blood vessels limit the nutrients and oxygen supply to rapidly growing tumors, thereby shrinking the tumor size. IL‐1β, IL‐27, IL‐12 and IFN‐γ have been shown to possess anti‐tumor and anti‐angiogenic activities [19, 20, 21]. Therefore, we examined their expression in tumor tissue of the immunized mice. We observed increased expression of proinflammatory cytokines IL‐27, IL‐1β, IL‐6, IL‐12, TNF‐α and IFN‐γ in antigen‐pulsed and PGN‐treated, but not in antigen‐pulsed DC‐only injected mice (Med‐DC) (Fig. 2e). DCs express TLRs that recognize the pathogens, allow the maturation of DCs and up‐regulate the co‐stimulatory molecules required for T cell activation. However, the tumor microenvironment alters TLR expression on DCs and impairs DC‐activated anti‐tumor T cell functions [22, 23]. Therefore, we analyzed in vitro generated and the indicated TLR ligand‐treated DCs for the expression of co‐stimulatory molecules. We observed high expression of CD80, CD86 and MHC‐II in PGN‐treated DCs when compared with P3C, BPP and untreated DCs (Supporting information, Fig. 2a,b), suggesting efficient maturation of DCs by PGN. Once matured, tumor antigen‐loaded DCs present tumor antigens to T cells and allows their expansion to prevent tumor growth [24, 25, 26]. CD8 T cells play a crucial role in mounting sustainable anti‐tumor immunity [27, 28]. Draining lymph nodes of twice‐immunized tumor‐injected mice were examined for TNF‐α and IFN‐γ secreting CD8 T cells. Upon flow cytometric analyses of CD3⁺CD8⁺ T cells in draining lymph nodes of the tumor‐bearing mice, we observed increased numbers of IFN‐γ and TNF‐αsecreting CD8 T cells in the PGN‐DC injected group (Fig. 2f,g). These observations indicated that PGN treatment of the tumor antigen‐pulsed DCs conferred the ability to induce strong anti‐tumor immunity and tumor regression on these cells.

Fig. 2.

Twice‐immunization with the tumor antigen‐pulsed Toll‐like receptor (TLR) ligand‐treated dendritic cells (DCs) results in a reduction of tumor growth. (a) The size and volume of each tumor were measured (n = 5). (b) After euthanizing the mice, tumors were excised (n = 3), weighed and presented as mean ± standard error of the mean (s.e.m.). (c) The photographs were taken using a stereomicroscope at 12× magnification to visualize angiogenesis and tumor size. (d) The AngioQuant free software was used and the quantified blood vessel complexes are represented graphically. (e) Relative fold changes in the expression of various proinflammatory molecules in the tumor tissue were assessed by quantitative–polymerase chain reaction (q‐PCR). (f) Single‐cell suspensions were made from lymph nodes and cells were stained for fluorescence activated cell sorter (FACS) analysis. The flow cytometric histograms represent the expression of tumor necrosis factor (TNF)‐α and interferon (IFN)‐γ in CD3⁺CD8⁺ T cells in the draining lymph node of twice‐immunized, tumor‐bearing mice (n = 3). Data shown are from one representative experiment that was repeated three times. The values shown are mean ± standard error of the mean (s.e.m.). *P < 0·05; **P < 0·01; ***P < 0·001.

Mice primed four times with tumor antigen‐pulsed, PGN‐treated DCs are resistant to tumor rechallenge

Addressing whether the priming regimen might result in complete prevention of tumor growth, we administered the TLR2 ligand‐treated tumor antigen‐pulsed DCs four times into the C57BL/6 mice, followed by challenge with RM‐1 tumor cells. We observed that four times‐priming rendered the mice completely resistant to RM‐1 cells, whereas all control mice receiving Med‐DC and untreated mice (n = 5) developed large tumors (Fig. 3a,b). The PGN‐DC and P3C‐DC recipients showed 100% survival and the protection from tumor rechallenge was accompanied by an enhanced cytotoxic T cell (CTL) activity and very low IL‐10, but high IL‐12 and IFN‐γ production (Fig. 3c–e). To test the durability of the anti‐tumor resistance, the mice resistant to the first tumor challenge were again rechallenged after 1 month with RM‐1 cells. The PGN‐DC and P3C‐DC treated mice continued to exhibit complete resistance to the second challenge, while the BPP‐DC treated mice significantly delayed the tumor growth (Fig. 3f). To assess the robustness of the anti‐tumor functions of these T cells, CD4⁺ and CD8⁺ T cells were isolated from the rechallenged mice, pooled at a ratio of three CD4⁺ T cells to one CD8⁺ T cell and adoptively transferred to naive syngeneic recipients, followed by RM‐1 challenge. Mice receiving PBS (saline), mice receiving T cells from naive mice and the untreated group (control) had grown large tumors, whereas the mice receiving T cells from the PGN‐DC primed rechallenged group showed complete protection against tumor (Fig. 3g–i), indicating elicitation and establishment of robust anti‐tumor memory T cells by appropriate priming with TLR2 ligand‐treated tumor antigen‐pulsed DCs. While pulsing with whole tumor antigens instead of a single antigen may induce broader immunity against larger tumor burden, the complete resistance to repeated tumor challenges establishes the protocol for testing tumor recurrence. In the process, we have also identified PGN as a potential adjuvant for anti‐tumor effect.

Fig. 3.

Four times dendritic cell (4× DC) immunization completely abolishes tumor growth in mice challenged with live RM‐1 tumor cells. Tumor antigen‐pulsed, Toll‐like receptor (TLR)2 ligand‐treated DCs were injected four times at an interval of 7 days into mice. After 7 days, these mice were challenged by tumor cells. (a) Mice were euthanized 21days after RM‐1 cells injection, tumors were excised (n = 4) and weighed. (b) The size and volume of each tumor were measured (n = 4). (c) Splenocytes were co‐cultured with irradiated RM‐1 cells. After 5 days, viable CD8⁺ T cells were purified and tested for their cytolytic activity against [³H]‐thymidine‐incorporated target cells (RM‐1) in a standard 4h Just another method (JAM) test; the effector to target ratio was 6 : 1 and each data set is the mean of triplicate samples. (d) The kinetics of percentage survival of RM‐1 challenged control and immunized groups (n = 5) is shown. (e) Relative fold changes in the expression of T cell‐specific anti‐inflammatory: interleukin (IL)‐4 (e,i) and IL‐10 (e,ii) and proinflammatory molecules IL‐12 (e,iii) and interferon (IFN)‐γ (e,iv), in the splenocytes were checked by quantitative–polymerase chain reaction (q‐PCR) after normalizing mRNA of the genes against the glyceraldehyde 3‐phosphate dehydrogenase (GAPDH). (f) Thirty days after the primary challenge, the mice showing resistance against the tumor were rechallenged with live RM‐1 tumor cells. Mice were observed until day 19 and euthanized; tumor weight (inset) and volume measured on alternate days is shown. (g) CD4 and CD8 T cells were isolated from the spleen of the rechallenged mice and transferred to the naive mice, followed by RM‐1 cells injection. Naive and control mice were injected with T cells from the naive and untreated tumor‐bearing mice, respectively, while mice in the saline group were injected with sterile phosphate‐buffered saline (PBS). Mice were euthanized on day 19 after tumor cell inoculation, tumors were excised and volume and weight (h) were measured. (i) The photographs were taken using stereomicroscope at 12× magnification to visualize angiogenesis and tumor size. Data shown are from one representative experiment that was repeated twice. The values shown are mean ± standard error of the mean (s.e.m.). *P < 0·05; **P < 0·01; ***P < 0·001.

Discussion

Surgical, radiological and chemotherapeutic interventions of solid tumors reduce the tumor burden beyond detection, but finally fail as the tumors relapse. The problem of tumor recurrence is often associated with enhanced survival and drug‐resistance of the tumor cells. Therefore, tumor recurrence is a major problem associated with the current cancer remedies and requires an alternative that bypasses the cancer cells conquest of the therapies. Herein, we show that the tumor antigen‐pulsed, peptidoglycan‐activated BMDCs induce robust host‐protective T cells that prevent tumor growth even after rechallenge. The major limitation of the rechallenge experimental mouse model is short‐duration (threee‐four month), where we need to inject live tumor cells for assessment of tumor relapse. Tumor cell heterogeneity and stemness are also important parameters for natural tumor relapse, which are difficult to consider while rechallenging mice with cultured tumor cells.

TLRs were first proposed to recognize pathogen‐associated molecular patterns (PAMPs) [20, 29]. TLR activation by their specific ligands trigger myeloid differentiation primary response 88 (MyD88)‐ and Toll/IL‐1 receptor domain‐containing adapter protein (TIRAP)‐dependent signaling cascades regulating the expression of TNF‐α, IL‐12, IL‐6 and type I IFNs. These cytokines primarily govern the inflammatory responses, which can be controlled by another TLR‐inducible protein called activating transcription factor 3 (ATF‐3) [30]. These proinflammatory responses are crucial for elimination of pathogens [31]. Besides pathogens, several reports have discussed anti‐tumor properties of TLR2, TLR3, TLR7, TLR8 and TLR9 [5, 6, 7, 8, 9, 10, 11], while priming with DCs treated with their ligands, including TLR7‐based ointment (Aldara^TM) approved for topical application [32], failed to induce complete immunity to tumor challenge. The reason Aldara fails is the metastasis of the melanoma cells to bone marrow and lungs. The use of TLR2 ligands with tumor antigens has also been widely tested, which provided protection to a certain level, but showed undesired harmful effects. Akazawa et al. prepared BMTCs (bacteria‐mimicking tumor cells) by conjugating modified TLR2 ligands to irradiated tumor cells, but their formulations failed as a vaccine candidate [33]. Another group demonstrated the anti‐tumor effects of Pam₃CSK₄‐long synthetic peptide conjugate in the B16F10 melanoma model. They observed a high efficiency of conjugates compared to the co‐injection of free forms of ligand and peptide [9]. Although TLR ligands were examined as adjuvants connecting APCs with T cells upon activation by synthetic or natural PAMPs [4], devising effective immunotherapy by establishing tumor‐eliminating T cells faces major challenges.

In our tumor model, we dissected tumor on different days and identified differential expression of TLR1, TLR2 and TLR6 in a phase‐specific manner. We report for the first time that tumor growth alters TLR expression on tumor‐infiltrated DCs. The uninterrupted interactions between the tumor cells and DCs regulate this process, which could be responsible for DC paralysis resulting in partial failure of recent DC vaccination approaches [22, 23]. A significant number of studies is required for understanding the tumor microenvironment‐induced inhibitory effects on the DCs. DC maturation is characterized by increased expression of MHC molecules, CD80 and CD86 [23]. We found the expression of these molecules in DCs was enhanced upon treatment with the TLR2 ligand PGN. Intratumoral injection of PGN reduces tumor growth, perhaps by stimulating the tumor‐infiltrating DCs, acting as altered APCs to induce efficient anti‐tumor T cell responses. To examine this hypothesis, we generated the DCs in vitro, pulsed these cells with gamma‐irradiated tumor cells in the presence of TLR2 ligands and immunized C57BL/6 mice with these cells. Besides reducing the tumor growth, the PGN‐treated antigen‐pulsed DCs also increased the number of IFN‐γ⁺TNF‐α⁺CD8 T cells, the professional killers [34]. High IL‐27, T‐bet and IFN‐γ expression in tumor tissue suggests tumor regression by the infiltrating IFN‐γ‐secreting Th1 cells.

The effectiveness of the DC vaccination depends upon the type of DC and antigens such as the proteins/peptides used for DC loading and TLR adjuvants. Doses of TLR adjuvants also play a key role in DC maturation, as different TLRs have different potentials for the activation of effector functions or immune responses. Moreover, different cell types show different TLR expression. Thus, all TLRs do not necessarily require the same amount of ligand for their effectiveness. Considering these propositions, we used the previously standardized optimal doses of Pam₃CSK₄, PGN and BPP. The degree of maturation is a key feature of DC vaccination, as immunization of patients with immature DCs may result in tolerance. Immature DCs are also incompetent to migrate to T cell areas in the lymphoid organs. Plasmacytoid DCs are also called type 2 DC because of their ability to induce Th2 differentiation; in contrast, myeloid DCs favor Th1 differentiation [35, 36]. Moreover, due to the following reasons, plasmacytoid DCs are not as effective in processing and presenting antigens as myeloid DCs: (1) low expression of lysosomal proteases cathepsin S and cathepsin D involved in antigen processing and (2) low expression of MHC‐II and co‐stimulatory molecules [36]. Dose, frequency and root of DC administration also affect the nature of T cell priming [35]. The observations reported here clearly demonstrate that these plausible problems are overcome by our current DC vaccination protocol. We injected TLR ligand‐treated tumor antigen‐pulsed DCs s.c. at an interval of 7 days. The mature DCs show up‐regulated expression of C‐C chemokine receptor type 7 (CCR7) and migrate to draining lymph nodes to interact with naive T cells for establishing anti‐tumor T cell responses [4, 23]. In our case, PGN‐treated tumor antigen‐loaded DCs reduced the numbers of immunosuppressive regulatory T cells (T_regs) and Th2 cells (data not shown) in draining lymph nodes of RM‐1 tumor cell‐challenged mice. The reduced numbers of T_reg and Th2 cells may lead to withdrawal of their inhibitory effects on host‐protective anti‐tumor CD8 T cells; as a result, IFN‐γ and TNF‐α secreted by CD8 T cells exert a strong anti‐tumor effect, causing the observed tumor regression.

The fourtimes DC priming also reinstated a strong Th1 response with concurrent very low IL‐10 production, implying a reduction of protumor T cells, such as T_reg cells. Elevated expression of IL‐12, TNF‐α and IFN‐γ maintained high cytotoxic T cell activation, which was strong enough to resist and eliminate RM‐1 tumor cells during the rechallenge, resulting in enhanced survival of mice. The inflammatory cytokine induced by vaccination acts on T cells to modulate their expansion and attain the memory characteristics. The difference in resistance to tumor by the two times‐primed and four times‐primed mice could arise from differences in memory T cells, the persistence of the tumor antigens relating to the threshold for memory T cell activation and sustenance or repeated activation of the antigen‐specific T cells and more robust T cell differentiation, such that they do not revert to tumor‐promoting T cells. However, deciphering all these molecular and cellular mechanisms call for independent studies.

To the best of our knowledge, this is the first study showing rechallenge experiments, where we translate the strength of anti‐tumor immune response against recurrence. These types of immune responses are crucial to completely prevent periodic tumor relapse following chemo‐ or radio‐therapeutic interventions. Adoptive transfer of T cells from the four times PGN‐DC‐immunized mice to the naive recipients also exerted host‐protective anti‐tumor effects, as tumor size and weight were radically reduced in these recipient mice. Thus, this study shows that PGN‐activated tumor antigen‐pulsed DCs enhance host survival by establishing host‐protective, adoptively transferable, robust anti‐tumor T cells, as the mice remain resistant to tumor rechallenge, establishing a protocol that removes a major bottleneck in anti‐tumor immunotherapy impeded by the relapse and tumorigenesis.

Disclosures

C. A. G. and T. E. are named as inventors in patents covering the use of BPPcysMPEG as adjuvant (PCT/DE03/03497 and PCT/EP2006011182). All other authors declare no conflicts of interest.

Author contributions

A. P. performed all in vitro and animal studies, analyzed data and performed statistical analyses. S. S. and P.C. contributed to animal studies and manuscript writing. C. A. G. and T. E. synthesized BPP and edited the manuscript. A. S. and D. C, helped in drafting the manuscript. B. S. designed the study, analyzed data and prepared the manuscript.

Supporting information

Fig. S1. After tumor cell injection, mice (n = 3 mice/group) were sacrificed on day‐7, ‐14 and ‐21, tumors were collected, crushed and stained for TLR1, TLR2 and TLR6 expression on CD11b⁺F4/80⁺ macrophages. (a) Representative FACS histograms showing the percentage and MFI of TLR1, TLR2 and TLR6 positive macrophages associated with tumors on day‐7, ‐14 and ‐21 after tumor cell injection along with isotype controls. (b) Bar graphs represent percentage and MFI of the TLR1, 2 and 6 expressing macrophages in the tumor tissue on different days. The experiment was repeated thrice and data from one experiment are shown. The data shown are mean ± SD. *P < 0·05, **P < 0·01, ***P < 0·001 and nonsignificant as ns.

Fig. S2. TLR2 ligands enhance the costimulatory activities of the DCs. Bone marrow derived DCs were generated and treated by Pam₃CSK₄ (1 μg/ml), PGN (5 μg/ml) and BPP (1μg/ml) for 16 hours followed by FACS staining. (a) The MFI and percentage of CD80, CD86 and MHC‐II positive DCs are shown in the histograms along with appropriately matched isotype controls. The right upper box represents gating strategy used for FACS analyses. The experiments were repeated thrice and the data from one representative experiment are shown. (b) Bar graphs represent percentage and MFI of DCs expressing CD80, CD86 and MHC‐II. The data shown are mean ± SD. *P < 0·05, **P < 0·01, *** P < 0·001.