Laser-Assisted Intradermal Delivery of Adjuvant-Free Vaccines Targeting XCR1⁺ Dendritic Cells Induces Potent Antitumoral Responses

This article is featured in In This Issue, p.5577

Dendritic cells (DCs) capture Ags in body barriers and migrate to lymph nodes (LNs), where they trigger the differentiation of Ag-specific, naive T cells into effector T cells. Recent studies, based on ontogeny and global gene-expression profiles, showed that a few DC subsets exist in the mouse. Functional specialization probably accounts for their existence, and they can be aligned with functionally equivalent human subsets (1–4). The XC-chemokine receptor 1 (XCR1) binds to a single ligand known as XCL1 and is expressed by a DC subset that was characterized previously by the expression of CD8α, CD207, or CD103 (5–7). Mouse XCR1⁺ DCs excel in cross-presentation of self-Ags (8, 9) and dead tumor cell–associated Ags, a feature that is probably owing to their expression of the C-type lectin CLEC9A (also known as DNGR1), a receptor for damaged and apoptotic cell materials (10, 11). Targeting Ags to DCs has been used successfully to generate strong immune responses and recently entered clinical trials (12). For instance, when delivered i.v. in the presence of adjuvant, Ags chemically or genetically conjugated to CLEC9A Ab or to XCL1 elicit potent cytotoxic CD8⁺ T cell responses capable of destroying tumors (13–17).

Owing to its high numbers of DCs, the skin is a particularly attractive site for vaccine administration. However, for ease of application, most of the skin vaccines that are used in mass-vaccination programs are not delivered into the epidermis or the dermis but into the hypodermis, which is a layer of fat and connective tissue just below the dermis with poorly characterized APCs. Such s.c. delivery bypasses the rich network of DCs present in the epidermis and dermis. Several delivery systems were developed recently to exploit the potential of dermal DCs. For example, vaccination with microneedles, made from a biocompatible polymer, has been used to introduce influenza virus vaccine (18) or live recombinant human adenovirus type 5 (19) into the dermis. These approaches generated robust humoral and cellular immune responses. A portable laser, the Precise Laser Epidermal System (P.L.E.A.S.E), also was used to create micropores in the stratum corneum, the superficial impermeable layer of the skin, and the epidermis, allowing topically applied Ags to diffuse into the dermis and to induce potent immune responses (20).

Vaccibodies are homodimeric chimeric proteins consisting of a targeting unit, a hinge domain, and an Ag moiety (21). XCL1-based vaccibodies were developed recently to target Ags to XCR1⁺ DCs (22). When used in a DNA format, flu hemagglutinin–containing vaccibodies improved CD8⁺ T cell responses and protected mice against a lethal challenge with influenza A virus (22). In the current study, we used the P.L.E.A.S.E. portable laser to deliver XCL1-based vaccibodies, in a protein format, within the dermis and demonstrated that they permit the harnessing of unique T cell immunostimulatory properties of XCR1⁺ dermal DCs. Using B16F10 (B16), a highly aggressive metastatic and poorly immunogenic melanoma and its OVA-expressing variant (B16-OVA), we demonstrated the in vivo antitumor efficacy of a single round of laser-assisted, dermal delivery of OVA-containing XCL1-based vaccibodies in both prophylactic and therapeutic settings and in the absence of intentionally added adjuvants.

Mice were used at 8–12 wk of age, housed under specific pathogen–free conditions, and handled in accordance with French and European directives. All mice were maintained on a B6 background. OT-I mice (23), OT-II mice (24), Ccr7^−/− mice (25), Myd88^−/−Trif^Lps2/Lps2 mice deficient in both MyD88 and Trif (Ticam1) (called Myd88^−/−Trif^−/− mice hereafter) (26), and Xcr1^tm1Dgen mice (Xcr1-βGal, called Xcr1^−/− mice hereafter) (6) were described previously. C57BL/6J (B6) mice were purchased from Janvier (Le Genest-Saint-Isle, France).

To extract skin myeloid cells, ears were split into dorsal and ventral parts and incubated with a solution of PBS containing 1 mg/ml Dispase (Roche) for 2 h at 37°C or overnight at 4°C. The dorsal and ventral parts were then cut into small pieces and incubated for 90 min at 37°C with RPMI 1640 culture medium containing 1 mg/ml DNase and 1 mg/ml Collagenase IV (Worthington Biochemical). The resulting single-cell suspension was subjected to centrifugation on a Percoll gradient (Amersham-Pharmacia).

Myeloid cells from the skin were characterized by flow cytometry, as previously described (27). Briefly, single-cell suspensions were prepared by enzymatic digestion and gentle dissociation of ear skin. After excluding dead cells (SYTOX⁺), T cells (CD3⁺), NK cells (CD161c⁺), B cells (CD19⁺), and neutrophils (Ly-6G⁺), the remaining CD45⁺MHCII⁺CD11c⁺ cells contain Langerhans cells (LCs; CD11b⁺CD24⁺), XCR1⁺ (CD11b⁻CD24⁺) dermal DCs, and CD11b⁺CD24^low dermal cells. Analysis of CD11b⁺CD24^low dermal cells for the expression of Ly-6C and CD64 permits the identification of CD11b⁺ DCs on the basis of their Ly-6C⁻CD64⁻ phenotype. The remaining CD11b⁺CD24^low dermal cells include monocytes, monocyte-derived DCs (moDCs), and macrophages.

Cells were stained and analyzed using a FACS LSR II system with DIVA software (BD Biosciences). Cell viability was evaluated using SYTOX (Invitrogen), according to the manufacturer’s protocol. Anti-NK1.1 (PK136), anti-CD3 (17A2), anti–Ly-6G (1A8), anti-CD19 (6D5), and anti-CD64 (X54-5/7.1) were from BioLegend. Anti-CD11c (N418), anti–MHC class II (I-A/I-E) (M5/114.15.2), anti-CD45.2 (104), anti-CD45.1 (A20), anti-CD24 (M1/69), and anti-CD5 (53-7.3) were from eBioscience. Anti–Ly-6C (AL21), anti-CD4 (RM4-5), and anti-CD8α (53-6.7) were from BD Pharmingen. Prior to the analysis of monocytes, macrophages, DCs, B cells, T cells, NK cells, and neutrophils were systematically gated out using a “dump-channel” corresponding to cells positive for B220, CD3, NK1.1, or Ly-6G. Analysis was performed using FlowJo software (TreeStar).

XCL1-OVA vaccibodies contain the XCL1 chemokine and the OVA Ag linked through a dimerization domain consisting of the hinge and CH3 domain of human IgG3. XCL1-OVA vaccibodies were produced in HEK293E cells and purified using an affinity column containing an Ab directed against the vaccibody dimerization domain (22). After elution, vaccibodies were dialyzed against PBS.

A P.L.E.A.S.E. portable laser developed by Pantec Biosolutions (28) was used with the following settings: fluence: 11.9 J/cm², pulse duration: 75 μs, RepRate: 200 Hz, pulses/pore: 2, pore array size: 14 mm², and pore density: 8%. P.L.E.A.S.E.-assisted skin microporation was performed on the ear of anesthetized mice, and 20 μl sterile PBS, sterile PBS containing 1.5 μg (low) or 3.1 μg (high) OVA, or sterile PBS containing 2.5 μg (low) or 5 μg (high) XCL1-OVA was applied evenly on the microporated ear surface. The PBS solution percolated inside the micropores in <10 min.

For histological analysis, ears were harvested after laser poration and embedded in paraffin. Sections (5 μm) were stained with H&E for microscopic examination.

OT-I and OT-II T cells were isolated from pooled LNs and spleen of OT-I or OT-II mice kept on a Rag-2^−/−-B6 (CD45.1) background using Dynal CD8⁺ and CD4⁺ T Cell Negative Isolation Kits (Invitrogen), respectively. Purity was determined by staining with CD4, CD8, CD5, and TCR Vα2. For labeling with CellTrace Violet (CTV; Molecular Probes), purified OT-I and OT-II T cells were resuspended in PBS containing 2.5 mM CTV for 20 min at 37°C. A total of 10⁶ CTV-labeled OT-I and OT-II T cells was adoptively transferred into the specified mice. At the indicated times, single-cell suspensions were prepared from the auricular LNs draining the immunized ears, and OT-I and OT-II T cells were analyzed by FACS.

Splenocytes from B6 mice were pulsed with the SIINFEKL peptide or left untreated and labeled with a low (0.25 mM) or high (2.5 mM) dose of CTV. A total of 10⁷ splenocytes of each preparation was adoptively transferred into mice that had been immunized for 6 d. Thirty-six hours later, a single-cell suspension was prepared from spleen, and the ratio of CTV^high/CTV^low cells was determined by FACS.

Mice were injected s.c. into the flank with 10⁵ B16-OVA or B16 melanoma cells (29). Tumor size was assessed 14–16 d later using a caliper. Determining tumor volume after sacrificing the mice and careful dissection of the tumor mass constituted a more accurate and reproducible measure of tumor size compared with the measurement of s.c. tumor volume on live mice with a caliper. To adhere to institutional ethical guidelines, mice inoculated with B16 melanoma cells were sacrificed 14 d (therapeutic protocol) or 16 d (prophylactic protocol) after tumor inoculation. The presence of T cells infiltrating the tumor was assessed after enzymatic treatment of tumor mass with collagenase 2 (Worthington) and Percoll gradient (Amersham-Pharmacia).

T cells harvested from the tumor mass were incubated for 6 h at 37°C in the presence of PMA (5 ng/ml) and ionomycin (250 ng/ml). Monensin (GolgiStop; BD Pharmingen) was added to the suspension for the last 5 h. Cells were stained with anti-CD5, anti-CD4, anti-CD8, and H-2 K^b tetramers loaded with the SIINFEKL peptide (iTAg MHC tetramers; Beckman Coulter) and then permeabilized using the Cytofix/Cytoperm kit (BD Biosciences). Intracellular cytokines were detected by staining with anti–IFN-γ (XMG1.2; BD Pharmingen).

The magnitude of the experimental effect observed in preliminary experiments led to the estimation of the proper sample size in each experiment. Because some of the experimental groups did not pass the D’Agostino and Pearson omnibus normality test, a Mann–Whitney U test was used to assess the statistical significance within the different immunization settings. Exact probability values are shown in the figures. A p value > 0.05 was considered nonsignificant.

Prior to targeting XCR1⁺ dermal DCs in situ, we characterized the specificity of the XCL1-based vaccibodies on single-cell suspensions prepared by enzymatic digestion and gentle dissociation of ear skin. For that purpose, the antigenic moiety of XCL1-based vaccibodies was replaced by mCherry, a red monomeric fluorescent protein (Fig. 1A) that allows binding specificity to be assessed by flow cytometry. By combining CD24 and CD11b expression, XCR1⁺ dermal DCs (also known as CD103⁺ or CD8α⁺ DCs) and LCs can be readily identified among the CD45⁺MHCII⁺CD11c⁺ cells found in the skin of the ear (Fig. 1B) (6, 7, 27). Analysis of CD45⁺CD11b⁺CD24^low dermal cells for the expression of Ly-6C and CD64 further identified CD11b⁺ DCs on the basis of their Ly-6C⁻CD64⁻ phenotype (27). The remaining CD45⁺CD11b⁺CD24^low non-DCs included monocytes, moDCs, and macrophages (Fig. 1B). XCL1-mCherry vaccibodies specifically stained XCR1⁺ DCs, whereas no staining was observed on cells isolated from the ear of mice deficient in XCR1 (Fig. 1C).

To target XCL1-based vaccibodies to XCR1⁺ dermal DCs in a needle-free manner, we used a P.L.E.A.S.E. portable laser. Application on the mouse ear skin using 75 μs pulse duration with two pulses/pore and an energy of 11.9 J/cm² resulted in the formation of an array of micropores via laser ablation (Fig. 2A). Histological analysis of ear sections indicated that the stratum corneum and the epidermis were removed, whereas the integrity of the dermis was preserved (Fig. 2B). Therefore, the regimen chosen created pores whose depth was compatible with topically applied XCL1-based vaccibodies reaching the dermis. Analysis of single-cell suspension prepared from ear skin 24 h after laser-assisted delivery of XCL1-mCherry vaccibodies showed that ∼15% of XCR1⁺ dermal DCs were stained; using Xcr1^−/− mice, this staining was shown to depend on XCR1 expression (Fig. 2C). Therefore, the combination of laser microporation and XCL1-based vaccibodies permits targeting of XCR1⁺ dermal DCs in a specific and needle-free manner.

Twenty-four hours after laser microporation, analysis of the ear skin showed a mild inflammation that occurred irrespective of the application of XCL1-OVA vaccibodies and involved infiltrates made of neutrophils and CD45⁺CD11b⁺CD24^low cells (Supplemental Fig. 1A, 1B). Resolution of the different subsets present among the infiltrating CD45⁺CD11b⁺CD24^low cells (6, 7, 27) showed that they were dominated by Ly-6C^high monocytes and moDCs. In contrast, 24 h after laser microporation the numbers of LCs, XCR1⁺ and CD11b⁺ dermal DCs, and macrophages remained unchanged (Supplemental Fig. 1C). Therefore, despite the rapid infiltration of laser microporated skin by neutrophils, monocytes, and moDCs, the topically applied XCL1-based vaccibodies still succeeded at targeting XCR1⁺ dermal DCs.

To test whether targeting XCR1⁺ dermal DCs with Ag-loaded XCL1-based vaccibodies induced the activation of Ag-specific T cells in vivo, the model Ag OVA was inserted into XCL1-based vaccibodies. Mice were adoptively transferred with CTV-labeled OT-I CD8⁺ T cells and OT-II CD4⁺ T cells; they express a TCR specific for SIINFEKL, an OVA-derived peptide presented by H-2K^b, and for an OVA-derived peptide presented by H2-A^b, respectively. One day later, the ears of the mice were subjected to laser-assisted microporation and topical application of XCL1-OVA vaccibodies. To compare the potency of the XCL1-OVA formulation with that of free OVA and, thus, determine the gain resulting from targeting OVA to XCR1⁺ DCs through XCL1-OVA vaccibodies, ears also were subjected to laser-assisted microporation and topical application of OVA in amounts equimolar to that present in XCL1-OVA vaccibodies. Three days after immunization, single-cell suspensions were prepared from ear-draining auricular LNs, and the extent of OT-I and OT-II cell proliferation was determined by CTV dilution (Fig. 3A). XCL1-OVA vaccibodies triggered a higher proliferation of OT-I and OT-II cells than did free OVA (Fig. 3B). Quantification of the data confirmed that XCL1-OVA vaccibodies were, on a per-molecule basis, 15 times more effective at eliciting CD8⁺ T cell proliferation and 3 times more effective at eliciting CD4⁺ T cell proliferation than was free OVA (Fig. 3C). When the same experiments were repeated with Xcr1^−/− mice, the beneficial effect observed following XCL1-OVA treatment was lost: the magnitude of proliferation induced with XCL1-OVA was comparable to that of OVA (Fig. 3C).

The XCL1 chemokine present in XCL1-OVA vaccibodies was shown to retain its chemotactic function (22). Therefore, the higher potency noted for XCL1-OVA Abs compared with OVA may not result from the targeting of OVA to XCR1⁺ DCs but from the ability of XCL1 to promote encounters between XCR1⁺ DCs and T cells (5, 30). However, coadministration of OVA and XCL1 in free forms and in amounts similar to those used in XCL1-OVA treatment resulted in levels of T cell proliferation similar to those elicited by OVA alone (Fig. 3C). Therefore, the physical linkage between OVA and XCL1 provided by the vaccibody format was essential to maximize the immunogenic potential of XCR1⁺ dermal DCs, and it is likely that XCR1 merely functioned as a target for delivery of Ag.

Skin DCs capture incoming Ags and after 2–5 d, depending on the DC subset, migrate to skin-draining LNs to elicit T cell responses (31–33). However, Ags that enter the dermis and are small enough to penetrate the lymphatic vessels can reach the LN subcapsular sinus in a free form (34). Such lymph-borne Ags can be captured by DCs that line the subcapsular sinus (35). This led to a more rapid generation of effector T cell responses, independent of skin-derived migratory DCs. DC-free drainage of skin-delivered Ags is amplified during needle-based, intradermal, and s.c. immunization as a result of the excessive interstitial hydrostatic pressure created by fluid injections (34). To evaluate whether T cell responses elicited by laser-assisted, intradermal delivery of XCL1-OVA vaccibodies required the migration of skin-derived XCR1⁺ DCs, rather than the capture of free XCL1-OVA vaccibodies by the XCR1⁺ DCs that permanently reside in skin-draining LNs (36), we used Ccr7^−/− mice, in which the CCR7-dependent migration of DCs from the skin to the draining LNs is impaired (25). Analysis of T cell responses of wild-type and Ccr7^−/− mice 3 d after immunization showed that OT-I and OT-II responses were reduced 10–20-fold in the absence of CCR7 (Supplemental Fig. 2A, 2B). Therefore, CD4⁺ and CD8⁺ T cell activation induced by laser-assisted intradermal delivery of XCL1-OVA vaccibodies is primarily due to migration of OVA-presenting XCR1⁺ dermal DCs to draining LNs. Consistent with the view that following laser-assisted intradermal delivery free XCL1-OVA vaccibodies do not diffuse away from the ear immunization site via the blood, adoptively transferred OT-I and OT-II T cells present in the spleen and in LNs that drain territories distinct from the ear showed no sign of proliferation (Supplemental Fig. 2C, 2D).

Although XCL1-OVA vaccibodies were affinity purified and reconstituted in endotoxin-free PBS prior to delivery, we determined whether the extensive proliferation of Ag-specific T cells observed after targeting XCR1⁺ dermal DCs with XCL1-OVA vaccibodies persisted in Myd88^−/−Trif^−/− double-deficient mice that are deprived of two adaptors used in all of the TLR signal-transduction networks. Laser-assisted, intradermal delivery of XCL1-OVA vaccibodies in Myd88^−/−Trif^−/− double-deficient mice resulted in levels of OT-I and OT-II T cell proliferation similar to those elicited in wild-type mice (Supplemental Fig. 3). Therefore, T cell responses triggered by laser-assisted intradermal delivery of XCL1-OVA vaccibodies require the migration of dermal XCR1⁺ DCs and occur irrespective of TLR signals.

The B16-OVA melanoma is not rejected by immune-competent syngeneic B6 mice unless they have been subjected to prophylactic or therapeutic immunization. To evaluate the capacity of laser-assisted, intradermal delivery of XCL1-OVA vaccibodies to inhibit the growth of B16-OVA tumors, B6 mice were inoculated s.c. in the flank with B16-OVA cells (Fig. 4A). Three days later, the ears of the mice were subjected to laser-assisted, dermal delivery of XCL1-OVA vaccibodies, and tumor growth was monitored 14 d after tumor inoculation. For comparison, control mice received PBS or equimolar amounts of free OVA. A single intradermal immunization with a low or a high dose of XCL1-OVA vaccibodies significantly suppressed tumor growth compared with the OVA and PBS groups, and tumor growth was further reduced with the high dose of XCL1-OVA vaccibodies (Fig. 4B). To investigate the prophylactic effects on tumor growth induced by laser-assisted intradermal XCL1-OVA immunization, B6 mice were immunized with XCL1-OVA vaccibodies and inoculated s.c. with B16-OVA tumor cells 30 d later (Fig. 4C). Monitoring tumor growth 16 d after tumor inoculation showed that XCL1-OVA vaccibodies significantly slowed tumor growth compared with OVA and PBS (Fig. 4D). As expected, XCL1-OVA vaccibodies induced an OVA-specific immunity, and mice immunized with XCL1-OVA vaccibodies were not able to control the growth of B16 melanoma that did not express OVA (Fig. 5A). Moreover, the capacity of XCL1-OVA vaccibodies to inhibit B16-OVA tumor growth was dependent on the expression of XCR1 (Fig. 5B).

To assess whether the reduction in melanoma tumor growth observed upon treatment with XCL1-OVA vaccibodies was associated with the induction of endogenous, Ag-specific CD8⁺ T cells, wild-type and Xcr1^−/− mice were immunized via laser-assisted intradermal delivery of XCL1-OVA vaccibodies. Six days after immunization, the magnitude of OVA-specific T cell cytotoxicity was assessed using an in vivo cytotoxic assay (Fig. 5C). XCL1-OVA vaccibodies induced a significant T cell cytotoxicity compared with PBS or OVA treatment. The absence of detectable lysis in Xcr1^−/− mice indicates that such T cell cytotoxicity was dependent on the expression of XCR1. Moreover, analysis of the CD8⁺ and CD4⁺ T cells that infiltrated the B16-OVA tumor mass of mice treated with XCL1-OVA vaccibodies showed that they were capable of producing IFN-γ (Fig. 5D). In contrast, OVA treatment did not induce IFN-γ⁺ T cells over PBS control, a finding consistent with the observation that OVA treatment was unable to slow tumor growth (Fig. 4). In addition, staining of CD8⁺ T cells that infiltrated the regressing B16-OVA tumor mass found in XCL1-OVA–treated mice with H-2 K^b tetramers loaded with the OVA-derived, SIINFEKL peptide showed an ∼100-fold increase in the numbers of OVA-specific, IFN-γ−producing CD8⁺ T cells compared with OVA-treated mice (Supplemental Fig. 4). Therefore, laser-assisted, intradermal delivery of XCL1-OVA vaccibodies induced antitumoral responses that can be correlated with the presence of OVA-specific IFNγ⁺ CD8⁺ T cells in the regressing tumor mass.

In the current study, we show that, by creating laser-generated micropores in the stratum corneum and epidermis of the skin, it is possible to target XCL1-OVA vaccibodies to XCR1⁺ dermal DCs and to induce Ag-specific CD8⁺ and CD4⁺ effector T cells. This process required migration of XCR1⁺ dermal DCs to draining LNs and occurred irrespective of TLR signals. Moreover, a single intradermal immunization with XCL1-OVA vaccibodies protected mice against melanoma tumor growth in both prophylactic and therapeutic settings and in the absence of intentionally added adjuvant. Whether Ag targeting to DC results in tolerance or immunity depends on parameters such as the immunogenicity of the targeting Ab (16) and the coadministration of adjuvants (37, 38). Adjuvants are intended to trigger the pattern-recognition receptors that are expressed by the targeted DCs and that are normally used to detect invading microorganisms or endogenous “danger” signals. In contrast to other studies that targeted XCR1⁺ DCs via needle-based, i.v., or cutaneous injection (14, 15, 39), we achieved Ag-specific protection against the B16-OVA melanoma in the absence of intentionally added adjuvant. Moreover, TLR signals were dispensable for the Ag-specific T cell responses resulting from laser-assisted intradermal delivery of XCL1-OVA vaccibodies. The mode of Ag delivery itself by skin laser microporation most likely explains such a marked difference in adjuvant requirement. The fractional Er:YAG laser operating in the P.L.E.A.S.E. portable laser creates microcoagulated areas in the skin that include dying cells (40). STING (stimulator of IFN genes) is a protein that resides in the endoplasmic reticulum of many cells, including DCs. It cooperates with the nucleotidyltransferase cGAS to trigger the production of type I IFNs in response to the presence of pathogen- or self-derived DNA in the cytosol. It was shown recently that XCR1⁺ DCs contribute to trigger T cell responses against tumors in a STING-dependent manner (41–43). XCR1⁺ DCs use STING to sense the self-DNA released by dying tumor cells (44) and, as a result, produce type I IFN that contributes to boost their Ag-presenting function and T cell costimulatory properties. Therefore, it is likely that, in our model, the death of keratinocytes induced by laser microporation acts as a STING-dependent adjuvant. We also noted that laser microporation triggers a rapid infiltration of the treated skin area with granulocytes, monocytes, and moDCs. Therefore, altogether, the adventitious phenomena resulting from skin laser microporation itself create an inflammatory milieu that most likely favors the development of potent T cell immune responses in the absence of exogenous adjuvants. In a mouse model, a rat anti-CLEC9A Ab used to deliver OVA to XCR1⁺ DCs induced CD4⁺ T cell and humoral responses against OVA in the absence of adjuvant (16), a property resulting from the presence of helper epitopes on the rat Ab that were recognized as foreign by the mouse immune system. Along the same line, the dimerization unit derived from human IgG3 domains and present in vaccibodies likely contributes to enhance mouse immune responses against the antigenic cargo (21). Finally, considering that the use of adjuvants in vaccines is often associated with safety issues, the ability to induce protective responses against melanoma tumor growth independently of the administration of exogenous adjuvants should facilitate the development of safer vaccines.

Migratory DCs originating from tissues such as the skin and intestine are thought to “instruct” Ag-specific naive T cells in a way that confers them with a propensity to home to the tissue from which the migratory DCs originated (45). Such tropism allows primed T cells to exert their effector functions in the tissue subjected to the Ag challenge. When injected i.v., Ag-conjugated anti-CLEC9A Abs target the XCR1⁺ DCs that permanently reside in the spleen and, thereby, initiate T cell responses in this organ (14). In contrast, following laser-assisted, intradermal delivery of XCL1-OVA vaccibodies, the onset of T cell responses depended on the migration of XCR1⁺ dermal DCs and remained limited to the skin-draining LNs. Whether skin-derived XCR1⁺ DCs confer to Ag-specific T cells a skin tropism superior to that elicited by spleen-resident XCR1⁺ DCs constitutes an important issue when treating conditions such as cutaneous melanoma. Although the T cell priming resulting from laser-assisted, intradermal delivery of XCL1-OVA vaccibodies is limited to the LN draining the treated skin area, a systemic T cell response ensued that was capable of protecting against cutaneous melanoma developing at a site distant from the one used for immunization. When translated to humans, this approach should limit the systemic side effects resulting from the administration of i.v. vaccines while achieving systemic protective immunity.

In conclusion, using laser-assisted intradermal delivery and a model Ag fused to the XCL1 chemokine, we showed that it is possible to target dermal XCR1⁺ DCs and harness their cross-presentation capacity. Although the efficiency of many vaccines relies on multiple rounds of administration in the presence of adjuvants, we showed that a single intradermal immunization with XCL1-based vaccibodies is sufficient to protect mice against melanoma tumor growth in the absence of exogenous adjuvants. Therefore, topic, laser-assisted intradermal delivery of Ags targeting XCR1⁺ DCs constitutes a promising route for the delivery of vaccines. In humans, XCR1 expression defines a DC subset that shows a similar anatomical distribution to mouse XCR1⁺ DCs and is endowed with cross-presentation capacity (30, 46–48), a feature that should facilitate the translation of the present mouse model to human settings.

We thank M. Dalod, A.-M. Schmitt-Verhulst, and C. Coscia for discussions, R.J. Driscoll for the Xcr1^tm1Dgen mice, L. Alexopoulou for the Myd88^−/−Trif^Lps2/Lps2 mice, M. Dalod for importing the Xcr1^tm1Dgen mice, and L. Chasson for assistance with histology.

R.B. holds shares in Pantec Biosolutions. B.B. is the inventor on United States patent 8,932,603 B2 on vaccibody filed by Oslo University Hospital, according to institutional rules. B.B. and E.F. are inventors on a patent application on XCL1 vaccibodies. B.B. is head of the scientific advisory board and hold shares in Vaccibody. The other authors have no financial conflicts of interest.