Regulation of Cutaneous Immunity In Vivo by Calcitonin Gene–Related Peptide Signaling through Endothelial Cells

Calcitonin gene–related peptide (CGRP) is a 37-aa neuropeptide produced by an alternative splicing of the calcitonin gene (1, 2) frequently coexpressed with substance P or somatostatin in sensory neurons (3). CGRP is a pleiotropic signaling molecule with many functional roles, including important immunoregulatory functions (4–6). CGRP‐containing nerves are distributed throughout various tissues and organs (7), and CGRP is expressed in both the central and peripheral nervous systems (7, 8). Many other cell types, including monocytes/macrophages (9), Langerhans cells (LCs; dendritic APCs that reside within the epidermis) (10), and keratinocytes (11), among others, can produce CGRP. There are two isoforms of CGRP, αCGRP and βCGRP, that differ by 1 aa in the rat and 3 aa in humans (2).

With regard to immune and inflammatory processes, CGRP has been shown to regulate APC function and inhibit the acquisition of immunity to Th1-dominant haptens while augmenting immunity to Th2-dominant haptens (12–14). Although CGRP has proinflammatory activities, many studies have demonstrated that administration of exogenous CGRP can inhibit the elicitation of inflammation by inflammatory stimuli in vivo (15–18), and perhaps related to this phenomenon, CGRP inhibits the release of certain chemokines by stimulated endothelial cells (ECs) (19). In addition, recent studies have implicated CGRP in the induction of psoriatic inflammation by the TLR7 agonist imiquimod in mice (20, 21). In this regard, innervation regulates the expression of certain inflammatory disorders in humans, most notably psoriasis (22, 23) which improves or resolves when occurring in an area of skin that becomes denervated. This suggests the possibility that the pathway described in this study may represent a mechanism by which the nervous system regulates inflammatory skin disorders. The pathophysiology of psoriatic arthritis also involves innervation (24, 25).

Recently, we reported that CGRP can bias the outcome of Ag presentation by epidermal LCs or blood-derived dendritic cells to responsive T cells in vitro away from Th1-type immunity and toward the Th2 and Th17 poles through actions on ECs acting as bystanders (4). In these experiments, the presence of CGRP-treated ECs led to stimulated responding CD4⁺ T cells exhibiting significantly increased expression of IL-17A accompanied by inhibition of IFN-γ, IL-4, and IL-22 expression at both the protein and the mRNA levels compared with Ag-presenting cultures containing ECs not exposed to CGRP (4). Accordingly, RNA levels of retinoic acid receptor-related orphan receptor γ t (RORγt) were increased in these T cells along with decreases in mRNA levels for T-box expressed in T cells (T-bet) and GATA binding protein 3 (GATA3) (4). We also found that CD4⁺ cells expressing cytoplasmic IL-17A were increased in this cell population, whereas cells expressing cytoplasmic IFN-γ or IL-4 were decreased (4). These results involved, at least in part, induction of IL-6 release by ECs induced by CGRP (4). Based on these in vitro observations, we have used our inducible, conditional EC-specific receptor activity modifying protein 1 (RAMP1) knockout (KO) mouse model to examine the in vivo relevance of these findings. Functional CGRP receptors result from the association of the calcitonin receptor-like receptor and RAMP1 (26). Thus, removing RAMP1 disables the CGRP receptor. These mice were generated by crossing mice with floxed RAMP1 with other mice carrying vascular endothelial (VE)-cadherin-Cre ER^T2. Administration of tamoxifen (Tx) early in life to these mice results in loss of RAMP1 from ECs (EC RAMP1 KO mice). In these experiments, Cre⁻, floxed RAMP1 mice (Cre⁻ mice) were used as negative control animals.

VE-cadherin-CreER^T2 mice were the generous gift of Dr. S. Raffi (Weill Cornell Medicine). Sperm containing floxed RAMP1 [full KO-First mutant allele (Tm1a) on C57BL/6 background; B6NTac;B6N-Ramp1 < tm1a(EUCOMM)Wtsi > /H] was obtained from The Mary Lyon Centre, Harwell Science and Innovation Centre, Oxfordshire, U.K. In vitro fertilization of C57BL/6 mice yielded mice heterozygous for the RAMP1 mutant. These mice were crossed with mice expressing Flp to generate mice with the wild-type RAMP1 allele (Tm1c) with loxP flanking exon 2. These mice, bearing the Tm1c allele, were then crossed with VE-cadherin-CreER^T2 RAMP1^fl/fl mice to generate Tm1c/+:Cre/+:flp/+ mice (∼1/8 pups). Tm1c/+, Cre/+, Flp/+ mice were then crossed with Tm1c/Tm1c mice to generate the Tm1c/Tm1c, Cre/+ line. Treatment with Tx (see later) yields the EC RAMP1 KO mice. Mice carrying the floxed RAMP1, but not the (VE)-cadherin-CreER^T2 allele, were used as controls (Cre⁻). Mice with the wild-type RAMP1 were also bred with VE-cadherin-CreER^T2 mice to yield wild-type Cre⁺ wild-type RAMP1 mice as controls for some experiments. These mice are on a C57BL6/N background and were generated by the Mouse Genetic Core transgenic facility of Memorial Sloan-Kettering Cancer Center (MSKCC). Bred mice were genetically initially assessed by agarose gel–based PCR. Toe or tail specimens were obtained and placed in 75 µL of lysis buffer (25 mM sodium hydroxide with 0.2 mM EDTA) and incubated at 94°C for 45 min. Then, 75 µl neutralization buffer (40 mM Tris HCl) was added, mixed well, and spun down; genomic DNA was then harvested. A total of 0.5 µl unpurified DNA from each sample was used for PCR. The PCR program was as follows: initial denature for 3 min at 94°C, followed by 40 cycles of denaturation for 15 s at 94°C, annealing for 90 s at 68°C, and completion for 7 min at 72°C. Primers for target gene recombination were Tm1a/Tm1c primers: RAMP1, 5′-GCCTTTTCAAAGCACAGTGG-3′ forward, 5′-GTCACATGGCATCCA CAGAC-3′ reverse; Mut-R1, 5′-GAACTTCGGAATAGGAACTTCG-3′; Tm1d primers, 5′CAS-F1 AAGGCGCATAACGATAC CAC, 3′LOXP-R1 ACTGATGGCGAGCTCAGACC (using these two primers, Tm1d is 174 bp and Tm1c is 1000 bp). PCR products were resolved on a 2% agarose gel at 130 V for 45–50 min. In later experiments, quantitative RT-PCR was used for genetic analysis (Transnetyx).

The mice were bred by the Colony Management Group at MSKCC. Some experiments were performed with BALB/c mice obtained from The Jackson Lab. All mice for experiments were maintained in the Weill Cornell Medical College animal facility under specific pathogen-free conditions, food and water ad libitum, and under a standard 12-h photoperiod at a constant temperature of 21°C. All experiments were conducted in compliance with all relevant ethical regulations and were approved by the Institutional Animal Care and Use Committee of the Weill Cornell Medical College.

Tx free base, sunflower seed oil, CGRP receptor antagonist BIBN4096BS (BIBN), 1-fluoro-2,4-dinitrobenzene (DNFB), collagenase A, and DNase 1 were purchased from Millipore Sigma; anti-mouse CD3 mAb and anti-mouse CD28 mAb were from BD Biosciences.

Complete medium consisted of RPMI 1640 (Mediatech), 10% FBS (American Type Culture Collection), 100 U/ml penicillin, 100 μg/ml streptomycin, 0.1 mM nonessential amino acids, 0.1 mM essential amino acids, 2 mM l-glutamine, 1 mM sodium pyruvate, and 10 mM HEPES buffer (all from Mediatech).

Tx was prepared in a vehicle of sunflower oil at a concentration of 20 mg/ml by shaking at 37°C for 1–2 h until completely dissolved and then stored in a light‐blocked condition at 4°C for the duration of injections. Experimental mice were transferred from MSKCC to the mouse hazard room at Weill Cornell Medicine at 10–12 d old for use. VE-cadherin-CreER^T2 RAMP1^fl/fl, Cre⁻ mice and Cre⁺ wild-type RAMP1 mice were injected i.p. with 50 or 100 µL Tx stock solution or with vehicle once every 24 h for a total of 3 or 4 consecutive days at age 2 wk and again at age 3 wk. Treatment of VE-cadherin-CreER^T2 RAMP1^fl/fl mice with Tx results in loss of RAMP1 from ECs (EC RAMP1 KO mice). Recombination was monitored either by the agarose gel method or quantitative RT-PCR, as described earlier.

Dermal microvascular ECs (DMECs) were prepared from skin obtained as described later (see preparation of LCs); skin samples were incubated 1 h in Ca²⁺/Mg²⁺-free PBS containing 1 U Dispase/ml. Epidermal sheets were mechanically removed; the remaining dermis was washed in PBS with Ca²⁺and Mg²⁺ five times and incubated for 2–3 h in a digestion buffer containing 2 mg/ml collagenase A, 100 µg/ml DNase 1, and 1% BSA in Ca²⁺/Mg²⁺-PBS. The resulting suspension was filtered through a 100-, 70-, and 40-μm nylon mesh sequentially (VWR) and washed three times with PBS with Ca²⁺and Mg²⁺ containing 1% BSA. Cells were incubated for 10 min with FcR Blocking Reagent (Cat no. 130-092-575; Miltenyi Biotec) to block Fc receptors. Then PE-conjugated anti-mouse CD31 (MEC 13.3; BD Biosciences) was added to a concentration of 5 µg/ml, and allophycocyanin-conjugated anti-mouse CD45 mAb (30-F11; BD Biosciences) was added to a concentration of 5 µg/ml followed by incubation for 30 min at 4°C. They were then washed twice and resuspended in 1× cold PBS with Ca²⁺ and Mg²⁺ containing 80 µg/ml DNase 1 and 1% BSA. DMECs were isolated by sorting for CD31⁺CD45⁻ dermal cells with the BD Influx cell sorter (BD Biosciences).

Epidermal cells were prepared using a modification of a standard protocol (12). Truncal skin of mice was shaved with electric clippers and chemically depilated. Subcutaneous fat and panniculus carnosus were removed by blunt dissection. Skin was floated dermis side down for 1 h in Ca²⁺/Mg²⁺-free PBS containing 0.5 U Dispase/ml (BD Biosciences) and 0.38% trypsin (Mediatech). Epidermal sheets were collected by gentle scraping, washed, and dissociated by repetitive pipetting in HBSS (Mediatech) supplemented with 2% FBS. Epidermal cells were filtered through a 40-μm cell strainer (BD Biosciences) to yield epidermal cells containing 2–3% LCs. Epidermal cells were incubated for 10 min with FcR Blocking Reagent (cat. no. 130-092-575; Miltenyi Biotec) to block Fc receptors. Then, PE-conjugated anti-mouse I-A^b mAb (clone AF6-120.1; BD Biosciences) was added followed by addition of Alexa Fluor 488 anti-mouse CD45 mAb (clone S18009D; BioLegend) and allophycocyanin anti-mouse CD31 mAb (clone MEC 13.3; BD Biosciences), each to 5 μg/ml, followed by incubation for 30 min at 4°C. They were then washed twice and resuspended in cold PBS containing 80 µg/ml DNase 1. LCs were isolated by sorting for epidermal cells expressing I-A^b and CD45, but not expressing CD31, and sorting with the BD Influx cell sorter (BD Sciences).

Dermis was obtained from skin as described earlier (see preparation of DMECs). Dermal fragments were washed in Ca²⁺/Mg²⁺-free PBS at least five times and incubated for 2–3 h in a digestion buffer containing 2 mg/ml collagenase A, 100 µg/ml DNase 1, and 1% BSA in Ca²⁺/Mg²⁺-free PBS. The resulting suspension was filtered through a 100-, 70-, and 40-μm nylon mesh subsequently and washed three times with Ca²⁺/Mg²⁺-free PBS containing 1% BSA. Dermal cells were incubated for 10 min with FcR Blocking Reagent (cat. no. 130-092-575; Miltenyi Biotec) to block Fc receptors. Then, PE-conjugated anti-mouse I-A^b (AF6-120.1; BD Biosciences) was added to 5 μg/ml, and allophycocyanin-conjugated anti-mouse CD11c mAb (117310; BD Biosciences) was added to 5 μg/ml; cells were then incubated for 30 min at 4°C, followed by washing twice and resuspension in cold PBS containing 80 µg/ml DNase 1 and 1% BSA. I-A^b+CD11c⁺ dermal dendritic cells were then isolated by FACS with the BD Influx cell sorter (BD Biosciences).

Spleens of mice were mechanically disrupted and passed through a 70-μm nylon mesh to yield a single-cell suspension. CD4⁺ cells were isolated by depletion of nontarget cells. The nontarget cells were indirectly magnetically labeled with a mixture of biotin-conjugated mAbs (CD8a, Cd11b, CD11c, CD19, CD45R [B220], CD49b [DX5], CD105, anti-MHC class II, Ter-119, and TCR‐γ/δ) as primary labeling reagent and anti-biotin mAb conjugated to microbeads as secondary labeling reagent (CD4 T Cell Isolation Kit, Mouse, no. 130-104-454; Miltenyi Biotec). The magnetically labeled nontarget cells were held on a MACS column in the magnetic field of a MACS separator, whereas the unlabeled CD4⁺ T cells passed through the column (Miltenyi Biotec).

Ten EC RAMP1 KO mice and 10 Cre⁻ mice were shaved on the dorsum with electric clippers; half of both sets of mice were immunized by application of 25 μl of 0.5% DNFB in acetone:olive oil (4:1) to the shaved dorsum, and the other half painted with acetone:olive oil (4:1) alone. On day 7, 5 μl of 0.2% DNFB 0.2% in acetone:olive oil (4:1) was applied to each side of each ear of all mice. Ear thickness was measured before ear painting and 24 and 48 h later using an engineer’s spring-loaded micrometer (Mitutoyo). Mice were euthanized after the 48-h ear swelling assessment, and ears were removed and placed in 10% formalin for histological analysis.

Additional experiments were performed examining the ability of administration of the CGRP inhibitor BIBN to wild-type mice to influence the expression of contact hypersensitivity (CHS). Wild-type mice (BALB/c) were injected i.p. with 100 µl PBS containing 15 µg BIBN or PBS alone 1 h before and 1 h after immunization with DNFB as described earlier or were mock immunized with diluent alone. One week later, 15 µg BIBN in 20 µl of PBS was injected into the base of each ear followed 30 min later by challenge of each ear with application of DNFB and assessment of 24- and 48-h ear swelling as described earlier. Control mice were similarly treated but injected with diluent alone instead of BIBN.

The inflammatory infiltrate in ears of EC RAMP1 KO mice and Cre⁻ mice immunized and challenged as described earlier, harvested after the 48-h ear swelling measurement, was examined histologically in a blinded fashion with slides stained with H&E (slides were prepared at the Laboratory of Comparative Pathology of the Center for Comparative Medicine and Pathology at Weill Cornell Medicine). The total inflammatory infiltrate, the neutrophilic infiltrate, and the lymphocytic infiltrates were scored in a blinded fashion on a semiquantitative fashion as follows: for overall infiltrate, 1+ = baseline “normal”-appearing section, 2+ = single focus of mildly dense inflammation, 3+ = two or more discontinuous/separate foci of mildly dense inflammation, 4+ = two or more discontinuous/separate foci of moderately dense inflammation, 5+ = diffuse inflammation with up to moderate density, and 6+ = diffuse inflammation with high density; for individual neutrophil or lymphocyte scoring criteria, not scored = cells not identified or within normal limits, 1+ = present but rare cells, 2+ = mild density of cells, 3+ = moderate density of cells, and 4+ = high density of cells. Images for publication were prepared by scanning slides with the Aperio GT450 Automated High Capacity Digital Pathology Slide Scanner (Leica Biosystems).

Ten EC RAMP1 KO mice and 10 floxed Cre⁻, RAMP1 mice (Cre⁻ mice) were shaved on the dorsum with electric clippers; half of both sets of mice were immunized by application of 10 µl of 1% DNFB in acetone:olive oil (4:1) to each side of the shaved superior dorsum, and the other half were painted with vehicle alone. Three days after DNFB immunization, mice were sacrificed, and draining lymph nodes (cervical, brachial, and axillary) were removed. Lymph nodes were mechanically disrupted and passed through a 70-μm nylon mesh to yield a single-cell suspension. CD4⁺ T cells were isolated as described earlier. Ninety-six-well flat-bottom plates were treated with 10 μg/ml anti-mouse CD3ε mAb (145-2C11; BD Biosciences) in PBS overnight and washed. T cells were cultured (3 × 10⁵ cells/well) in 250 μl complete medium containing 2 μg/ml anti-mouse CD28 mAb (553294; BD Biosciences). Supernatants were collected 72 h after stimulation, and cytokine contents were determined.

Supernatant IL‐17A and IL‐4 levels were determined by sandwich ELISA following the manufacturer’s instructions (IL-17A, DY421 [R&D Systems]; IL-4, DY404 [R&D Systems]). IL‐22 ELISA kits were purchased from Antigenix America (cat. no. RMF222CK). IFN-γ production in cultured medium was measured by a sandwich ELISA using purified rat anti‐mouse IFN-γ capture mAb (XMG1.2; BD Biosciences), biotinylated rat anti‐mouse IFN-γ detection mAb (XMG1.2; BD Biosciences) avidin-HRP (no. 554066; BD Biosciences), and 3,3′,5,5′-tetramethylbenzidine substrate (BD OptEIA Reagent Set, no. 555214; BD Biosciences), read at 405 nm.

Total mRNA was extracted from ECs, dermal dendritic cells, LCs, and CD4⁺ T cells using the RNeasy Plus Mini Kit (Qiagen); DNA eliminator columns were used to eliminate any contamination with genomic DNA. cDNA was synthesized using a high-capacity RNA-to-cDNA kit according to the manufacturer’s instructions (SuperScript VILO cDNA Synthesis Kit; Invitrogen). The primers for murine RAMP1 were designed to span the junction between exon 1 and exon 2 (reference sequence number: NM_001168392): 5′-GGATGAGAGTCCCATAGTCAGG-3′ forward and 5′-GGGGCTCTGCTTGCCAT-3′ reverse. Expression was performed using PrimeTime primers (Integrated DNA Technologies) and was normalized to GAPDH (5′‐GTGGAGTCATACTGGAACATGTAG‐3′ forward; 5′‐AATGGTGAAGGTCGGTGTG‐3′ reverse). PCR was performed with Power SYBR Green PCR Master Mix (Applied Biosystems) and an ABI 7900HT instrument (Applied Biosystems).

EC RAMP1 KO and Cre⁻ control mice were sensitized with DNFB. Three days later, CD4⁺ T cells were isolated from draining lymph nodes and were stimulated with anti-CD3 mAb and anti-CD28 mAb. Twelve hours later, total RNA was isolated, and RT-PCR was performed using methods as described earlier. All of the PrimeTime primers were purchased from Integrated DNA Technologies: murine IL-17A (5′-GAGCTTCCCAGATCACAGAG-3′ forward; 5′-AGACTACCTCAACCGTTCCA-3′ reverse), IFN-γ (5′-GAGCTCATTGAATGCTTGGC-3′ forward; 5′-CAGCAACAACATAAGCGTCAT-3′ reverse), RORγt (5′-TCCCACATCTCCCACATTG-3′ forward; 5′- AATGTCTGCAAGTCCTTCCG -3′ reverse), GATA3 (5′-GTCCCCATTAGCGTTCCTC-3′ forward; 5′-CCTTATCAAGCCCAAGCGAAAATGTCTGCAAGTCCTTCCG-3′ reverse), T-bet (5′-CAAGACCACATCCACAAACATC-3′ forward; 5′-TTCAACCAGCACCAGACAG-3′ reverse), IL-4 (5′-TCTTTAGGCTTTCCAGGAAGTC-3′ forward; 5′-GAGCTGCAGAGACTCTTTCG-3′ reverse), and IL-22 (5′-AATCGCCTTGATCTCTCCAC-3′ forward; 5′-GCTCAGCTCCTGTCACATC-3′ reverse). Expression of each cytokine and transcription factor was normalized to GAPDH (earlier primer sequence).

RAMP1 protein expression was detected by Western blotting with an mAb specific to RAMP1. DMECs were isolated from both Cre⁻ and RAMP1 KO mice by FACS, as described earlier. After washing two times with cold PBS, ECs were lysed by adding 50 ml of cell lysis buffer containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM Na₂ EDTA, 1 mM EGTA, 1% Triton, 2.5 mM sodium pyrophosphate, 1 mM glycerophosphate, 1 mM Na₃VO₄, 1 mg/ml leupeptin, and 1 mM phenylmethylsulfonylfluoride (Cell Signaling). Cells were then sonicated for 5 s for five times with cooling on ice between sonication. Cell lysates were then centrifuged for 5 min. The protein concentration in each sample was determined by a NanoDrop Spectrophotometer (ND-1000; NanoDrop Technologies). Equal amounts of protein (20 µg) of cell lysate from each sample were mixed with 3× Blue Loading Buffer (Cell Signaling) with 40 mM DTT and were separated on a Mini-PROTEAN Precast SDS-PAGE gel (Bio-Rad), followed by electrotransfer to a nitrocellulose membrane (Bio-Rad). The membrane was blocked with EveryBlot Blocking Buffer (Bio-Rad) for 15 min and then washed for 5 min three times with 1× TBST buffer (TBS buffer, 0.1% Tween 20; Bio-Rad), followed by probing with 1:1000 dilution of rabbit anti-mouse recombinant anti-RAMP1 Ab (clone EPR10867; Abcam) or with 1:10,000 dilution of rabbit anti-mouse recombinant anti-GAPDH (clone EPR16891; Abcam) overnight at 4°C. After washing three times with 1× TBST, the membrane was probed with 1:10,000 diluted HRP-conjugated goat anti-rabbit IgG (catalog number STAR208P; Bio-Rad) at room temperature for 1 h. GAPDH was used as the loading control. The protein signals were detected using an ECL mixture (Clarity Western ECL Substrate; Bio-Rad) and recorded by a Bright CL 1000 (Invitrogen) imaging system.

For analysis of the study endpoints, including supernatant cytokine contents, CHS (ear swelling) responses, cellular infiltrates observed on histologic sections, and expression of specific mRNAs, linear mixed-effects regression was used to estimate levels of an end point in each treatment group while accounting for potential within-experiment and within-mouse correlations depending on the design of the experiments. Simultaneous testing of general linear hypotheses testing was used to evaluate the contrasts of interest. The p values were adjusted for multiple comparisons by controlling the false discovery rate. Cytokine and mRNA expression data were log transformed before the analysis to ensure the underlying model assumptions were satisfied. All statistical analyses were conducted using R (https://www.r-project.org/). All p values were two sided.

Initial experiments examined the expression of RAMP1 mRNA in ECs derived from EC RAMP1 KO mice and ECs derived from identically treated RAMP1 floxed but Cre⁻ mice. ECs from EC RAMP1 KO mice had substantially reduced expression of RAMP1 compared with ECs from identically treated Cre⁻ mice (Supplemental Fig. 1A). Epidermal LCs and dermal dendritic cells isolated from these two categories of mice showed no difference in expression of RAMP1 (Supplemental Fig. 1B, 1C). Surprisingly, RAMP1 expression in CD4⁺ T cells isolated from the spleens of EC KO mice showed a small but significant increase in RAMP1 expression compared with CD4⁺ T cells from Cre⁻ mice (Supplemental Fig. 1D). By Western blot, ECs from EC RAMP1 KO mice also expressed greatly reduced expression of RAMP1 compared with ECs from identically treated Cre⁻ mice (Supplemental Fig. 1E).

Histologic sections of EC RAMP1 KO mouse skin and Cre⁻ mouse skin were stained with periodic acid–Schiff plus diastase to look for differences in vascular morphology by light microscopy. Examination of these sections by an experienced dermatopathologist was performed, but no differences in morphology were noted by this technique.

The next set of experiments examined cytokine expression in CD4⁺ cells isolated from draining lymph nodes of EC RAMP1 KO mice and Cre⁻ control animals immunized to DNFB. DNFB or diluent alone was applied to the shaved dorsa of groups of each type of mouse. Three days later, mice were euthanized and draining lymph nodes harvested. A single-cell suspension of CD4⁺ T cells from the draining lymph nodes was obtained by magnetic Ab techniques. Cells were placed in culture and stimulated with anti-CD3 and anti-CD28 mAbs. Supernatants were harvested after 72 h and assayed by ELISA for cytokine contents. Cells from immunized EC RAMP1 KO mice produced significantly less IL-17A compared with immunized Cre⁻ control mice, while production of IFN-γ, IL-4, and IL-22 was significantly increased (Fig. 1). A small but significant reduction in IL-17A production was also seen in nonimmunized mice, as was a significant increase in IL-22 production. Control experiments were performed in a similar manner after administering Tx to mice expressing the wild-type RAMP1 alleles (no floxing) with and without VE-cadherin-CreER^T2. No differences in production of IL-17A, IFN-γ, IL-4, or IL-22 were observed between mice expressing wild-type RAMP1 with or without the presence of VE-cadherin-CreER^T2 (Supplemental Fig. 2).

In other experiments, examination of mRNA levels for cytokines and relevant transcription factors in CD4⁺ T cells from draining lymph nodes was performed. Groups of EC RAMP1 KO mice and Cre⁻ control mice were similarly immunized. Three days later, draining lymph nodes were harvested. CD4⁺ T cells were isolated and stimulated for 12 h with anti-CD3 mAb and anti-CD28 mAb as in the Materials and Methods. Then, total mRNA was prepared, and quantitative RT-PCR was performed for mRNAs of interest. As shown by the data in (Fig. 2, draining lymph nodes from EC RAMP1 KO mice showed significantly less expression of mRNAs for IL-17A and the Th17-associated transcription factor RORγt compared with Cre⁻ control mice, while expression of mRNAs for IFN-γ and the Th1-associated transcription factor T-bet, mRNAs for IL-4 and the Th2-associated transcription factor GATA3, and IL-22 mRNA were significantly increased.

We also examined CHS responses in EC RAMP1 KO mice. Groups of EC RAMP1 KO mice and Cre⁻ control mice were each immunized by application of DNFB to the shaved dorsum or were mock immunized with diluent alone. Seven days later, mice were challenged by application of DNFB to the ears and the CHS response quantified by measurement of 24- and 48 h ear swelling. As shown by the data in (Fig. 3A, EC RAMP1 KO mice had a substantially and significantly lesser CHS response, measured by 24- and 48-h ear swelling, compared with Cre⁻ control mice. Histologic assessment of the ears showed a significant reduction in the total inflammatory cell response, a significant reduction in the neutrophil infiltrate, and a trend toward a lesser lymphocyte infiltrate in ear skin from immunized EC RAMP1 KO mice compared with Cre⁻ mice (Fig. 3B). In (Fig. 3C, photomicrographs of a representative section from the ear of a mouse from each group taken 48 h after challenge (immunized Cre⁻, immunized EC RAMP1 KO, nonimmunized Cre⁻, and nonimmunized EC RAMP1 KO) are shown. Note that the ear from the immunized Cre⁻ mouse is thicker than the ear from the immunized EC RAMP1 KO with more edema and a thicker dermis with greater inflammatory cell infiltrate. Also, note that we have chosen representative areas for each photomicrograph. There is variation in the density of the inflammatory infiltrate in each section. The dataset from the blinded evaluation of the infiltrate in sections from each ear taken from each mouse is shown in Supplemental Table I.

A control experiment was also performed in which VE-cadherin-CreER^T2 RAMP1^fl/fl mice and RAMP1^fl/fl without Cre mice (Cre⁻ mice) were injected with vehicle alone instead of Tx at 2 and 3 wk of age and then immunized and challenged as described earlier. As shown in Supplemental Fig. 3, no difference in ear swelling was seen in immunized VE-cadherin-CreER^T2 RAMP1^fl/fl mice compared with RAMP1^fl/fl no Cre mice treated with vehicle alone instead of Tx.

To further examine the role of CGRP signaling in CHS, we used the small molecule competitive inhibitor of CGRP BIBN (also known as olcegepant). Additional experiments were performed examining the ability of administration of this CGRP inhibitor to wild-type mice to influence the expression of CHS. Mice were injected i.p. with BIBN 1 h before and 1 h after immunization with DNFB by application to the shaved dorsum, and 1 week later BIBN was injected into the base of each ear followed 30 min later by challenge of each ear with application of DNFB. Control mice were similarly treated but injected with diluent alone instead of BIBN. Other groups of control mice were treated identically to the groups described earlier but were mock immunized with vehicle alone. As shown in (Fig. 4, mice injected with BIBN at immunization and elicitation of immunity had a significantly reduced CHS response compared with control mice injected with PBS alone instead of BIBN.

These data demonstrate that CGRP signaling through ECs has a number of profound immunologic effects. With regard to CD4⁺ T cell responses, CGRP signaling through ECs appears to play an important role in the spectrum of Th cells that result from immunization. These in vivo data demonstrate that CGRP signaling through ECs favors the differentiation of IL-17A–expressing Th cells (Th17 cells) while reducing the generation of IL-4–expressing (Th2), IFN-γ–expressing (Th1), and IL-22–expressing (Th22) Th cells. These observations are quite similar to in vitro experiments in which ECs exposed to CGRP in vitro and then added to Ag-presenting cultures led to similar changes in the spectrum of Th cells generated in these cultures compared with the addition of ECs not pretreated with CGRP (4). Thus, we believe that the changes in the generation of subsets of Th cells in the regional lymph nodes reported in this article are the in vivo counterpart to these in vitro results. Because generation of Th subsets is likely occurring in the lymph nodes, the CGRP signaling effects we are observing in vivo are likely due to ECs in the lymph nodes. Of course, in our current in vivo experiments, this is uncertain. We plan additional experiments using tissue transplants between control and EC RAMP1 KO mice for the near future to attempt to determine whether this is the case. This report demonstrates the in vivo significance of the in vitro observations and suggests that CGRP may play a regulatory role in disorders characterized by inappropriate Th17 cell activity. In this regard, it has been reported that ECs in lesions of psoriasis, a disease involving Th17 cell activity (27), have CGRP on their surfaces (10) and that patients with psoriasis have elevated plasma levels of CGRP (28). The possibility that therapeutic manipulation of this pathway may benefit disorders characterized by inappropriate Th17/IL-17A activity should be examined.

The profound changes in CHS response in the EC-RAMP1 KO mice are also of considerable interest. Whether this observation relates to the changes in Th cell generation discussed earlier remains unknown. However, because CGRP is a potent vasodilator, it is also possible that CGRP signaling in the skin vasculature at the sites of immunization and/or elicitation of CHS results in differences in cell trafficking, vasodilatation, or other local changes altering the responses that are unrelated to changes in Th cells. The actual mechanisms responsible for the changes in CHS response in these mice are of considerable interest and should be investigated in the near future. A more complete understanding of the role of CGRP in CHS may have important therapeutic implications. Also, future experiments with reciprocal skin grafts between control and EC RAMP1 KO mice followed by hapten application to the grafts for immunization may allow for a determination of whether CGRP signaling in ECs in the skin itself or elsewhere in the animals (secondary lymphoid organs?) is responsible for the functional changes observed in vivo. The possibility that CGRP signals through circulating endothelial progenitor cells (EPCs) could also be considered. These cells have immunoregulatory properties, including downregulation of CD4⁺ T cell activation (29). The relevance of EPCs to CHS responses in vivo and the role of possible CGRP signaling through EPCs remains unknown but is an important area for future study.

Release or nonrelease of CGRP by peripheral nerves likely represents an important locus of modulation of immunity by the nervous system. In this regard, innervation is important for the expression of certain inflammatory skin disorders, most notably psoriasis, and this pathway may be a mechanism by which the nervous system regulates inflammatory skin disorders. Interruption of CGRP signaling at ECs may be a potential therapeutic target for the treatment of Th17-mediated skin diseases such as psoriasis, as well as CHS. It is possible that therapeutic manipulation of this pathway might also be useful for inflammatory or immune disorders in organs other than the skin.

These results convincingly demonstrate the role of CGRP regulation of immunity in vivo. However, they leave a number of questions unanswered. With regard to the experiments involving EC-RAMP1 KO mice, the locus of CGRP EC signaling within the mouse is unknown. It may reside in the skin, secondary lymph node organs, or perhaps elsewhere. This must be worked out. In the experiments utilizing the inhibitor of CGRP signaling, the locus of action also remains unknown. Further investigations of regulation of immunity by CGRP signaling will shed additional light on the role of this molecule in physiologic and pathophysiologic immune processes and may lead to novel therapeutic approaches for some inflammatory disorders.

We thank Drs. Willie Mark and Yun You for invaluable contributions in helping to engineer and breed the inducible, conditional RAMP1 KO mouse used in this project. We also thank Dr. Xuan Wang for reviewing the histologic sections of EC RAMP1 KO and Cre⁻ mouse skin.

R.D.G. has a research agreement with Biohaven Pharmaceuticals, Inc. for a clinical study involving the peptide being studied in this article. R.D.G. is on the scientific advisory board of Elysum Health, Inc. and holds equity and options; is on the scientific advisory board of Hoth Therapeutics, Inc. and receives fees for this service; has research agreements with Galderma, Inc., Leo Pharma, Inc., Pfizer, Inc., and Elysium Health, Inc.; and is also an advisor to Gore Range Capital. R.D.G. is also an informal advisor to BelleTorus Corporation but has no financial ties to BelleTorus at this time. The other authors have no financial conflicts of interest.