Kit is essential for PMA-inflammation–induced mast-cell accumulation in the skin

Connective tissue mast cells reside in a variety of tissues, notably skin and peritoneal cavity (reviewed in Galli et al¹ ). The exact hematopoietic pathways that fill and maintain mast cell compartments are still poorly understood. Mast cell–committed progenitors have been identified in fetal blood,²  adult bone marrow (BM),^3,4  and spleen,⁵  and their existence suggests that mast cell commitment can precede tissue immigration. A crucial growth factor receptor that drives mast cell development in vivo is the receptor tyrosine kinase Kit. The importance of Kit is documented by the drastic reduction of mast cell numbers in mice with inactivating mutations in Kit or Kit ligand (stem cell factor [Scf]) genes⁶  (reviewed in Besmer,⁷  Galli et al,⁸  and Broudy⁹ ). In the skin, Kit signals can regulate mast cell numbers as shown by local mast cell proliferation and activation after subcutaneous administration of exogenous Scf.^10,11  In contrast to steady state mast cell development, less information exists on the requirement for Kit signaling in mast cells under inflammatory conditions.

Kit^W/Wv mice have been widely used as an in vivo model for mast cell deficiency. These mice have only approximately 1% of normal mast cell numbers in their skin, and they lack peritoneal mast cells.⁶ Kit^W/Wv mice express the Kit^W protein together with the Kit^Wv protein. The Kit^W protein lacks the transmembrane domain and thus cannot be expressed on the cell surface. In Kit^W/Wv mice, Kit expression is reduced to about 5% of wild-type levels.¹²  In addition to this low cell-surface expression, the function of the Kit^Wv receptor is impaired because of a point mutation that attenuates its kinase activity.¹³ 

Skin mast cells are present constitutively (ie, in the absence of an immune response), but their numbers markedly increase under local inflammatory conditions. Despite impaired Kit expression and function, mast cell numbers can also increase strongly (10- to 100-fold) in the skin of Kit^W/Wv mice during the natural history of an idiopathic dermatitis¹⁴  or during phorbol-12-myristate-13-acetate (PMA)–induced chronic dermatitis.¹⁵ 

Mice homozygous for Kit^W alleles die within a few days after birth,¹⁶  a fact that has precluded analyses of mast cells in adult Kit^W/W mice. Kit^W/W mice suffer from severe anemia¹⁶  that is the cause of their lethality.¹⁷  Stimulation of erythropoiesis by transgenic overexpression of a human erythropoietin (EPO) gene is sufficient to generate viable Kit^W/W mice.¹⁷  In these adult Kit^W/WTg(EPO) mice (referred to as Kit^W/W) the transgene has no salvage effects on hematopoietic cells outside the erythrocyte lineage, including stem/progenitor cells^12,17  and mast cells (see Figure 2).

The fact that chronic inflammation induces mast cell accumulation even in Kit^W/Wv mice raises the possibility that Kit signaling is not essential under inflammatory conditions. Alternatively, the residual Kit expression/signaling in Kit^W/Wv mast cells is sufficient to increase mast cell numbers. To solve this question, we have challenged the skin of adult Kit^W/W mice by PMA to induce a chronic irritative dermatitis. In contrast to Kit^W/Wv mice that showed a massive mast cell accumulation, the skin of Kit^W/W mice remained free of mast cells. To address the role of lymphocytes during this inflammation-induced mast cell accumulation, we have also generated and analyzed new compound mutants. In these triple mutant mice, termed R⁻γ⁻Kit^W/Wv, the mast cell deficiency of Kit^W/Wv mice was combined with lack of all lymphocytes owing to mutations in the recombination activating gene (Rag-2) and the common cytokine receptor gamma chain (γ_c) gene. Of note, inflammation-driven mast cell expansion was severely impaired in these compound R⁻γ⁻Kit^W/Wv mutants. Adoptive transfer experiments suggest that this defect in mast cell reconstitution was caused by the absence of lymphocytes rather than lack of γ_c, a receptor subunit invoked in mast cell growth. Hence, in this report, we provide new insights into the cellular and molecular regulation of mast cell numbers in the skin.

WBKit^W/+ (strain, WB), C57BL/6Kit^Wv/+ (strain, C57BL/6) mice (Japan-SLC, Shizuoka, Japan) and WBB6F1Kit^W/Wv (strain, F1 of WB × C57BL/6) mice were continuously bred in our animal facility. Parental mouse strains for Rag-2^−/−γ_c⁻Kit^W/Wv (termed R⁻γ⁻Kit^W/Wv) triple mutant mice were generated by crossing Rag-2^−/−18 γ_c⁻¹⁹ double-mutant mice (strain, mixed C57BL/6 and C57BL/10 background; Taconic, Germantown, NY) to WBKit^W/+ and C57BL/6Kit^Wv/+ mice. White-spotted (W or Wv) F1 mice were intercrossed to obtain Rag-2^−/−γ_c⁻Kit^W/+ or Rag-2^−/−γ_c⁻Kit^Wv/+ mice. Mice were typed for Rag-2 deficiency by lack of T and B cells in the peripheral blood by flow cytometry and for γ_c deficiency by polymerase chain reaction (PCR) on genomic DNA as described.¹⁹ R⁻γ⁻Kit^W/Wv mice were obtained by crossing Rag-2^−/−γ_c⁻Kit^W/+ to Rag-2^−/−γ_c⁻Kit^Wv/+ mice. Kit^W/W Tg(EPO) mice (strain, mixed WB and C57BL/6 background) were bred as described.¹⁷  The EPO transgene, its properties, and the promoter have been described previously.²⁰ 

Cells were stained, analyzed, or sorted using Calibur analysis or Aria cell sorter instruments (BD Biosciences [BD], Heidelberg, Germany) as described.^21,22  Briefly, cells were blocked with 500 μg/mL mouse IgG (Dianova, Hamburg, Germany), prior to specific antibody staining. Peritoneal exudate cells were harvested by peritoneal lavage with 10 mL warm (37°C) PBS/5% FCS and stained with FITC-labeled anti-T1 and PE-labeled anti-FcϵRI antibodies. Peripheral-blood leukocytes were enriched as described²³  and analyzed for the presence of lymphocytes using biotinylated anti-CD45, PE-Cy7–labeled anti-CD3 (all BD Pharmingen, San Diego, CA) antibodies. Second-step reagent was streptavidin-APC (Molecular Probes, Carlsbad, CA). Splenic mast cells were analyzed by staining of spleen cell suspensions with a FITC-labeled lineage (lin) cocktail (anti-CD4, anti-CD8, anti-CD19, anti-Ter119), APC-labeled anti-Kit (all BD Pharmingen), and PE-labeled anti-FcϵRI (eBioscience, San Diego, CA). Mast cells were identified by their lin⁻Kit⁺FcϵRI⁺ phenotype. Numbers of lymphocytes in the blood and mast cells in the spleen were calculated by the addition of a defined number of flow cytometry beads (Calibrite; BD) to the stained cells. Lymphocytes were purified negatively by staining of spleen cells with a FITC-labeled myeloid-erythroid lineage cocktail (anti–Mac-1, anti–Gr-1, and anti-Ter119; all BD Pharmingen) followed by magnetic bead (Dynal Biotech, Oslo, Norway) depletion. Subsequently, cells were stained with APC-labeled anti-Kit, and Mac-1⁻Gr-1⁻Ter119⁻Kit⁻cells (ie, the nonmyeloid, nonerythroid, non–mast cell fraction) were cell-sorter purified. This fraction was composed of 53% B cells, 41% T cells, and the remaining cells being non-T and non-B cells (presumably natural killer [NK] cells). Kit⁺ cells were excluded from the sort to prevent transfer of contaminating wild-type mast cells together with the lymphocytes.

Each mouse received 3 × 10⁷ splenic lymphocytes by intravenous injections. Beginning on day 5 after lymphocyte transfer, mice were treated with PMA or acetone for 6 weeks. Numbers of lymphocytes were measured in the blood 13 days after transfer (not shown) and at the end of the experiment (6 weeks after transfer) in blood (Figure 6), spleen, and lymph nodes (not shown). Bone marrow mast cell cultures (BMMCs) were established as described.²²  Because the relative contribution of local mast cell expansion versus increased colonization of mast cell progenitors from the circulation to the total mast cell increase in the inflamed skin is not fully understood, we choose intravenous cell transfer. Mice received a total of 3 intravenous BMMC injections (cell numbers were 1 × 10⁷ on day 0, 2 × 10⁶ on day 14, and 2 × 10⁶ on day 22), and mast cell numbers in the skin (not shown) and the spleen (Figure 6) were determined on day 30.

One ear of each animal was treated 3 times per week on Monday, Wednesday, and Friday over a period of 6 weeks with 10 μg PMA (Sigma, Taufkirchen, Germany) dissolved in acetone (solvent). For controls, the contralateral ears were treated with solvent alone. Before each PMA treatment, ear thickness was determined using a caliper (Käfer Messuhrenfabrik, Villingen Schwenningen, Germany). After 6 weeks (Figures 4–5), mice were killed, and ear skin was processed for histology. The study protocol was approved by the Regierungsprasidium Tubingen (permit no. 761).

Histochemical analysis of tissue mast cells was done as described.²²  Mast cells were also identified on acetone-fixed cryosections after blocking with 0.1 mg/mL mouse IgG in PBS/1% BSA using APC-labeled anti-Kit. The presence of granulocytes was determined by staining with FITC-labeled anti–Gr-1 (not shown). Tissue sections shown in Figures 2 and 4 were inspected with an Axioskop microscope using objectives with 10 × (0.30 aperture) and 20 × (0.50 aperture) magnification, respectively, at room temperature in air. Microscope and objectives were from Zeiss (Oberkochen, Germany). Fluorescence photomicrographs of berberine sulfate–stained cytospin preparations (Figure 2) were taken using the ORCA ER camera (Hamamatsu Photonics, Herrsching, Germany) and processed using Openlab software (Improvision, Coventry, England). Light microscopy photomicrographs of toluidin staining (Figure 4) were taken using the OM-11 color camera (Olympus, Hamburg, Germany).

Ears and peritoneal exudate cells were homogenized in RNAzol (WAK-Chemie, Steinbach, Germany), and cDNA was prepared as described.¹²  Primers for reverse transcription (RT)–PCR were used as follows: Hprt sense, 5′-GCTGGTGAAAAGGACCTCT-3′; Hprt antisense, 5′-CACAGGACTAGAACACCTGC-3′; Mc-cpa sense, 5′-ACACAGGATCGAATGTGGAG-3′; Mc-cpa antisense, 5′-TAATGCAGGACTTCATGAGC-3′. Southern blot probe for the Mc-cpa PCR product was 5′-CCTCCTAACCACCAGGACCT-3′.

To investigate the role of Kit signaling strength in mast cells and the role of lymphocytes in the regulation of mast cell numbers in chronically inflamed skin, we generated a series of informative single and compound mutants (Figure 1A). Skin inflammation was induced in the following mouse strains: (1) mice wild type for Kit, Rag-2, and γ_c in which normal Kit signaling was permissive, and all lymphocytes subsets were present; (2) Kit^W/Wv mice in which Kit expression and signaling were impaired but not completely abrogated, and all lymphocytes subsets were present; (3) Kit wild-type mice lacking T, B, and NK cells (Rag-2^−/−γ_c⁻ [termed R⁻γ⁻]); (4) Kit^W/Wv mice lacking T, B, and NK cells (Rag-2^−/−γ_c⁻Kit^W/Wv [termed R⁻γ⁻Kit^W/Wv]); and (5) Kit^W/W (ie, Kit cell surface null having peripheral T, B, and NK cells).

Mice of all genotypes where treated with PMA 3 times per week over a period of 6 weeks (Figure 1B). As a control, the contralateral ear of each mouse was treated with solvent alone. During the course of the treatment, ear thickness was determined as a measure of inflammation, and, after 4 to 6 weeks of treatment, mast cell numbers were counted by toluidine blue staining on skin sections of both ears.

Mice bearing the viable Kit^W/Wv genotype have strongly reduced numbers (approximately 1% of wild-type mice) of mature dermal mast cells and lack peritoneal mast cells^6,15  (Figure 2). In contrast to the widely used Kit^W/Wv mouse, up to now mast cells have not been analyzed in adult Kit^W/W mice. Cytospins of peritoneal exudate cells (PECs) were analyzed by toluidine blue reactivity which results in the mast cell–specific, metachromatic staining (Figure 2A-D) and by reactivity with the heparin-binding dye berberine sulfate²⁴  (Figure 2E-H). To include cell numbers larger than those accessible on cytospins, PECs were also analyzed by flow cytometry using the mast cell marker T1²⁵  and the high-affinity IgE receptor (FcϵRI) (Figure 2I-L). Kit was excluded as a mast cell marker because it is only weakly or not at all expressed on hematopoietic cells in Kit^W/Wv and Kit^W/W mice, respectively.²¹ 

On the basis of toluidine blue and berberine staining, and on flow cytometry, peritoneal mast cells were detectable in Kit^+/+ (Figure 2A,E,I) and in R⁻γ⁻ mice (not shown). In contrast, peritoneal mast cells were undetectable in all Kit mutants, that is, Kit^W/Wv (Figure 2B,F,J), R⁻γ⁻Kit^W/Wv (Figure 2C,G,K), and Kit^W/W (Figure 2D,H,L). Hence, the absence of lymphocytes did not affect the peritoneal mast cell compartment. In addition, the data confirmed the massive reduction of mast cells in Kit^W/Wv mice. At this level of analysis, there was no difference comparing hypomorphic Kit^W/Wv and null Kit^W/W mice with regard to mast cell development.

Gene expression analysis for mast cell–specific genes is a sensitive way to detect mast cells.²⁶  The mast cell carboxypeptidase A (Mc-cpa) is strongly and specifically expressed in the mast cell lineage²⁷  (references in Feyerabend et al²² ). Therefore, Mc-cpa expression was analyzed by RT-PCR in PECs (Figure 2M-P) or ears (Figure 2Q-T). In wild-type mice (Figure 2M,Q), Mc-cpa mRNA was strongly expressed in PECs and in skin. Consistent with the absence of mast cells in the peritoneal cavity of Kit^W/Wv mice,⁶ Mc-cpa mRNA was undetectable in Kit^W/Wv (Figure 2N) and R⁻γ⁻Kit^W/Wv (Figure 2O) mice. Not surprisingly, Kit^W/W PECs also lacked Mc-cpa mRNA (Figure 2P), indicating a complete absence of peritoneal mast cells in Kit^W/W mice.

In the ears, Mc-cpa was again strongly expressed in wild-type mice (Figure 2Q). Confirming previous results,²⁶  we detected Mc-cpa mRNA in the skin of both Kit^W/Wv mutants (Figure 2R,S). Titration of cDNA templates suggested that Mc-cpa expression was 10- to 100-fold lower in the ears of Kit^W/Wv mutants compared with wild-type mice (Figure 2R,S). Interestingly, in the ears of Kit^W/W mice Mc-cpa expression was undetectable by RT-PCR (Figure 2T). These data suggest that the residual Kit function in Kit^W/Wv mice is required for the low but significant mast cell production under steady state conditions in the Kit^W/Wv mutant.

PMA treatment elicits acute and chronic inflammatory responses, akin to irritative dermatitis, by nonspecific activation of a variety of PMA receptors (reviewed in Boutwell²⁸  and Brose and Rosenmund²⁹ ). The resulting immunologically nonspecific inflammation, measured by immediate (< 36 hours) tissue swelling, is amplified by mast cells after PMA treatment.³⁰  To measure the impact of other immune cells on this inflammation process, we determined time course and extent of ear thickness in all mutants after PMA treatment (Figure 3).

Ear swelling of fully immune-competent mice peaked at day 10 and declined thereafter to approximately double the normal thickness compared with nontreated contralateral ears (Figure 3A). Such strong initial reactions were not observed in any of the immune-compromised mice (Figure 3B-E), suggesting that mast cells and lymphocytes can both contribute to the initial swelling in the first 10 days. Beginning on day 12, and for the remainder of the observation time, ear thickness of wild-type (Figure 3A), Kit^W/Wv (Figure 3B), R⁻γ⁻ (Figure 3C), and Kit^W/W (Figure 3E) mice had at least doubled compared with the starting point (swelling ≥ 0.5 mm). In contrast, PMA-treated ears of R⁻γ⁻Kit^W/Wv mice showed reduced swelling (∼ 0.35 mm) in the long term (Figure 3D). Collectively, initial ear swelling was augmented by the presence of immune cells (either T, B, NK, and/or mast cells) (compare Figure 3A with B-E). Long-term ear swelling occurred in all mutants, but in the absence of all adaptive (T and B cells) and parts of innate (NK and mast cells) immune cells long-term ear swelling was reduced.

To obtain information about other cell types that are recruited during this inflammatory response, we examined infiltration of Gr-1⁺ leukocytes to the sites of chronic dermatitis. Immunofluorescence analysis of skin sections revealed an accumulation of polymorphonuclear Gr-1⁺ granulocytes in all treated ears, indicating that granulocyte recruitment was mast cell and lymphocyte independent (not shown).

To assess the contribution of Kit-mediated signals to the mast cell accumulation during chronic dermatitis, ears of PMA-treated mice of all genotypes were analyzed histochemically for the presence of mast cells (Figures 4–5). Skin thickness of PMA-treated ears (Figure 4B,D,F,H) was clearly increased compared with solvent-treated control ears (Figure 4A,C,E,G) visualizing the findings obtained by micrometer measurements (Figure 3). Mast cells were identified by their morphology and by their metachromatic staining with toluidine blue. Mast cells in control ears were only present in wild-type (Figure 4A) and R⁻γ⁻ (not shown) mice. Following control solvent treatment, we could not detect skin mast cells in Kit^W/Wv (Figure 4C), R⁻γ⁻Kit^W/Wv (Figure 4E), or Kit^W/W (Figure 4G) mice, demonstrating the paucity or absence of mast cells in the healthy skin of these mutants. As expected, PMA treatment resulted in increased mast cell numbers in wild-type mice (Figure 4B). Mast cell numbers increased not only in wild-type mice but also in Kit^W/Wv (Figure 4D) and R⁻γ⁻Kit^W/Wv (Figure 4F) mice. Interestingly, mast cells were undetectable in skin sections of Kit^W/W mice (Figure 4H).

Next, we obtained mast cell counts that were normalized on cartilage length (Figure 5). In wild-type mice (Figure 5A) and in R⁻γ⁻ mice (Figure 5C), 6 of 6 mice analyzed for each genotype showed a 3- to 4-fold increase in mast cells after PMA treatment (Kit^+/+ mice control ears [mean ± SD, 540 ± 274 mast cells/cm; Kit^+/+ mice PMA-treated ears, 1733 ± 454 mast cells/cm; R⁻γ⁻ mice control ears, 372 ± 96 mast cells/cm; R⁻γ⁻ mice PMA-treated ears, 999 ± 373 mast cells/cm). Mast cells also accumulated in 6 of 6 PMA-treated ears from Kit^W/Wv mice (Figure 5B) (Kit^W/Wv mice control ears, 1.3 ± 0.8 mast cells/cm; Kit^W/Wv mice PMA-treated ears, 179 ± 136 mast cells/cm) as reported previously.¹⁵  Although absolute mast cell numbers remained lower in the skin of Kit^W/Wv mice compared with wild-type or R⁻γ⁻ mice, the relative increase was larger in Kit^W/Wv compared with wild-type or R⁻γ⁻ mice. Interestingly, PMA treatment did not lead to the appearance of mast cells in the skin of Kit^W/W mice (Figure 5E) (Kit^W/W mice control ears, 0 mast cell/cm; Kit^W/W mice PMA-treated ears, 0 mast cell/cm), demonstrating an essential role for Kit-mediated signals in mast cell development in healthy and inflamed skin.

Analysis of R⁻γ⁻ mice (Figure 5C) indicated that the absence of lymphocytes and lack of γ_c expression did not impair the number of constitutive skin mast cells or their increase following skin inflammation in mice expressing wild-type levels of Kit. Thus, in contrast to peritoneal mast cells,³¹  γ_c expression is not required for normal numbers of skin mast cells. Because Kit signaling can act in marked synergy with γ_c signaling in other cell types such as thymocytes,³²  we next examined the inflammation-driven mast cell increase in combined Rag-2 and γ_c mutants on a Kit^W/Wv background (R⁻γ⁻Kit^W/Wv mice). In these mutants, the overall increase in mast cell numbers was strongly blunted (Figure 5D) (control ears, 2 ± 2 mast cells/cm, PMA, 12 ± 19 mast cells/cm) compared with Kit^W/Wv mice (Figure 5B). Moreover, 2 of 6 R⁻γ⁻Kit^W/Wv mice did not respond at all. Thus, on a Kit^W/Wv background, the presence of lymphocytes or γ_c expression is important for the regulation of mast cell numbers in inflamed skin.

To distinguish between effects mediated by lack of Rag-2 (ie, loss of T and B cells) or γ_c expression (ie, loss of signaling via γ_c-dependent cytokines), we performed 2 sets of further experiments. In the first set, cell-sorter purified, wild-type (R⁺γ⁺Kit^+/+) lymphocytes (see “Materials and methods”) were injected intravenously into R⁻γ⁻Kit^W/Wv mice. This adoptive transfer lead to reconstitution of circulating T cells in R⁻γ⁻Kit^W/Wv mice reaching about 50% (2157 ± 819 T cells/μL blood) (Figure 6D) of the numbers found in Kit^+/+ (3960 ± 1017 T cells/μL blood) (Figure 6A) or Kit^W/Wv (4250 ± 780 T cells/μL blood) (Figure 6B) mice. No lymphocytes were detectable in R⁻γ⁻Kit^W/Wv mice (Figure 6C). Ears in these groups of mice were subjected to PMA treatment over a period of 6 weeks. In agreement with the experiments shown in Figure 5, Kit^+/+ showed a robust increase in skin mast cell numbers (control ears, 318 ± 64 mast cells/cm; PMA-treated ears, 1526 ± 139 mast cells/cm) (Figure 6E). Likewise, Kit^W/Wv demonstrated a marked response (control ears, 1.7 ± 2.8 mast cells/cm; PMA-treated ears, 216 ± 80 mast cells/cm) (Figure 6F), whereas R⁻γ⁻Kit^W/Wv mice showed only a very modest mast cell response (control ears, 1 ± 1.7 mast cells/cm; PMA-treated ears, 22 ± 6 mast cells/cm) (Figure 6G). Lymphocyte-reconstituted R⁻γ⁻Kit^W/Wv mice harbored 4 ± 4.5 mast cells/cm in control ears and 95 ± 71 mast cells/cm in PMA-treated ears (Figure 6H). Although the comparison of the PMA-induced cell numbers in non–lymphocyte-reconstituted R⁻γ⁻Kit^W/Wv (open symbols in Figure 6G) versus in lymphocyte-reconstituted R⁻γ⁻Kit^W/Wv mice (open symbols in Figure 6H) has a P value of only .2 (Mann Whitney rank test) and hence does not reach statistical significance, it should be noted that 2 of 3 mice showed mast cell numbers above 100 cells/cm (106 and 160). We did not observe similarly high mast cell numbers in any of the nonreconstituted, PMA-treated R⁻γ⁻Kit^W/Wv ears (Figures 5–6), suggesting that the addition of lymphocytes to R⁻γ⁻Kit^W/Wv mice might promote mast cell accumulation in the inflamed skin.

In the second set of experiments, we addressed the possibility that the poor mast cell increase in R⁻γ⁻Kit^W/Wv mice was due to lack of γ_c expression in mast cells³¹  in these mutants. To this end, we transplanted γ_c-competent (R⁺γ⁺Kit^+/+) BMMCs into R⁻γ⁻Kit^W/Wv and Kit^W/Wv hosts. Despite repeated intravenous injections of BMMCs (see “Material and methods” for reasoning for route of administration, cell numbers, and frequency of injections), we failed to detect a significant increase in mast cell numbers in the ears of both types of recipients (R⁻γ⁻Kit^W/Wv without BMMCs [n = 3], 1.3 ± 1.5 mast cells/cm; R⁻γ⁻Kit^W/Wv with BMMCs [n = 5], 3.2 ± 5.5 mast cells/cm; Kit^W/Wv without BMMCs [n = 3], 7.7 ± 6.5 mast cells/cm; Kit^W/Wv with BMMCs [n = 5], 5.4 ± 3.4 mast cells/cm). Moreover, R⁻γ⁻Kit^W/Wv mice transplanted with wild-type (R⁺γ⁺Kit^+/+) BMMCs and challenged with PMA (as described for Figure 6E-H) also failed to generated skin mast cells (not shown).

These findings are in agreement with previous experiments that revealed very inefficient mast cell reconstitution in the skin following intravenous transfer.³³  However, the spleen has been reported to be a more effective site of mast cell reconstitution following intravenous transfer of BMMCs.³³  We therefore compared by flow cytometry mast cell numbers in the spleens of the same R⁻γ⁻Kit^W/Wv and Kit^W/Wv recipient mice that had received γ_c-competent (R⁺γ⁺Kit^+/+) BMMCs. Mast cells were present in the spleens of both types of recipients, but, interestingly, numbers of mast cells were at least one order of magnitude higher in Kit^W/Wv (1648 ± 780 mast cells/cm) compared with R⁻γ⁻Kit^W/Wv (87 ± 112 mast cells/cm) recipients (Figure 6J,K) (P = .004). Because γ_c-competent BMMCs reconstituted Kit^W/Wv mice well, but R⁻γ⁻Kit^W/Wv only poorly, these experiments support the notion that lymphocytes contribute to the regulation of mast cell compartments upon reconstitution of mast cell–deficient mice. Under these conditions, γ_c-competent mast cells did not rescue the defective mast cell reconstitution in R⁻γ⁻Kit^W/Wv mice.

Despite the important contributions of mast cells to the pathophysiology of skin inflammation, little is known about the cellular and molecular signals that regulate mast cell numbers in inflamed skin. The crucial role of Kit and its ligand, Scf, for normal mast cell development is evident in mice bearing hypomorphic mutations in the genes encoding Kit⁶  or Scf.³⁴  The mast cell paucity is, however, not absolute in the Kit^W/Wv mutant as shown by a significant enrichment of mast cells at sites of idiopathic¹⁴  or chemically¹⁵  induced chronic dermatitis. These findings raised the question whether residual Kit signaling, mediated by the Kit^W/Wv receptor, was facilitating this response or whether mechanisms unrelated to Kit signaling were able to drive mast cell accumulation under these conditions. Whatever the nature of these signals, they are insufficient to promote significant numbers of skin mast cells in Kit^W/Wv mice under nonpathologic conditions.

Here, we have taken advantage of the recently developed viable Kit^W/W mutants¹⁷  to address the role of Kit in the regulation of mast cell numbers during chronic dermatitis. Adult Kit^W/W mice have previously given insight into the crucial role of Kit in adult lymphopoiesis.^12,35  Although in vitro–derived mast cell lines were used to characterize the loss of Kit expression in viable Kit^W/W mutants,¹²  mast cell compartments have not been analyzed in these mice. Because the hypomorphic Kit^W/Wv mutant has been widely used as a “mast cell–deficiency model,” we have now compared mast cells in Kit^W/W and Kit^W/Wv mice. The results show that Kit^W/W mice differ in several aspects from Kit^W/Wv mice. On the basis of mast cell protease expression (Figure 2) and histologic analysis (Figure 4), we confirm previous reports (reviewed in Tsai et al²⁶ ) that Kit^W/Wv mice have very low but measurable numbers of skin but not peritoneal mast cells. In contrast, using the same criteria, we now find no evidence for skin mast cells in Kit^W/W mice. The absence of mast cells in the peritoneal cavity in both strains, but the presence of mast cells in the skin of Kit^W/Wv mice, suggests that skin mast cells are slightly less dependent on Kit signaling compared with peritoneal mast cells. However, the most dramatic difference between mast cells in Kit^W/Wv and Kit^W/W mice became only apparent following experimentally induced chronic dermatitis. Although mast cell numbers increased strongly in Kit^W/Wv mice, mast cells remained undetectable in Kit^W/W mice. These experiments demonstrate an active role of the Kit^Wv protein in the regulation of mast cell numbers in response to skin inflammation. Collectively, Kit-mediated signals are essential for skin mast cell formation under steady state conditions and under inflammatory conditions.

Short- and long-term reactions during PMA-induced chronic dermatitis include ear swelling. PMA nonspecifically activates a wide variety of cell types²⁸  by directly interacting with the signaling component protein kinase C and other PMA receptors.²⁹  To determine whether the residual Kit function in Kit^W/Wv mast cells was in itself sufficient for the mast cell response to inflammation, we eliminated T, B, and NK cells from Kit^W/Wv mice by generating R⁻γ⁻Kit^W/Wv mice. In short-term reactions (≤ 30 hours) mast cells play a role in ear swelling.³⁰  After 3 days of PMA treatment we observed a slower onset of ear swelling in all analyzed mutants. Thus, not only mast cells, but also T, B, or NK cells, are required for initial ear swelling. In contrast, after 15 days, only mice lacking T, B, and NK cells showed reduced swelling. Ear thickness was indistinguishable after 6 weeks of PMA treatment comparing Kit^+/+ and Kit^W/Wv mice¹⁵  (Figure 3). The fact that R⁻γ⁻ mice showed a similar swelling demonstrates that lymphocytes are not required for this skin reaction. Moreover, R⁻γ⁻Kit^W/Wv that had initially very few mast cells and constantly lacked lymphocytes showed the same degree of swelling.

The unexpected mast cell reconstitution defect in R⁻γ⁻Kit^W/Wv, when compared with Kit^W/Wv mice, might be due to lack of γ_c expression on mast cells or due to the absence of lymphocytes. Mice deficient for γ_c or a downstream signaling molecule Jak-3 were reported to have 2-fold reduced numbers of peritoneal mast cells.³¹  We did not notice a reduction in skin mast cells in R⁻γ⁻ mice, thus ruling out an important function for γ_c in this mast cell population. However, γ_c signaling in mast cells might be important if the Kit signal is limiting. Although we cannot definitively exclude a role for γ_c under such conditions, the relative inability of adoptively transferred γ_c⁺ BMMCs to reconstitute splenic mast cells in R⁻γ⁻Kit^W/Wv mice (Figure 6) argues against a major role of γ_c in mast cell reconstitution. The inability of robust reconstitution of skin mast cells following intravenous transfer of BMMCs prevented a more direct test whether the same rules apply to the skin as shown here for the spleen.

Lack of lymphocytes played no role for the accumulation of mast cells in Kit^+/+ mice but T, B, or NK cells were required on the Kit^W/Wv background. These data suggest that mast cells can expand in the skin by 2 mechanisms. The first possibility is a strong signal via Kit alone that can be transmitted by the wild-type (Kit⁺) but not by the kinase-weak (Kit^Wv) receptor. The second route would require 2 signals, a weak Kit signal that could be transmitted via Kit^Wv in Kit^W/Wv mice and a second signal that originates directly or indirectly from lymphocytes. It is conceivable that a weak Kit signal could also be generated in normal mice (Kit^+/+) under conditions of limiting availability of Kit ligand. The strongly blunted mast cell response in R⁻γ⁻Kit^W/Wv compared with Kit^W/Wv mice agrees with a supportive role of lymphocytes in mast cell recruitment, generation, or local proliferation. In addition to T and B cells, NK cells that are absent in R⁻γ⁻Kit^W/Wv mice could also play a role. The lymphocyte support for mast cell expansion could occur via direct cellular interactions or lymphocyte-derived factors may drive mast cell expansion. In vitro, IL-3 allows outgrowth of mast cells from bone marrow or peripheral blood of Kit^W/Wv36 and Kit^W/W12 mice. Furthermore, perfusion of Kit^W/Wv mice with IL-3 normalizes skin mast cell numbers,³⁷  and Kit ligand and IL-3 cooperate in the mast cell–driven worm expulsion.³⁸  In Kit^W/Wv mice, T-cell–derived IL-3 might, therefore, be a crucial factor for mast cell accumulation during the course of chronic dermatitis.

Presently, it is difficult to distinguish whether the defect in Kit^W/W mice is due to reduced availability of circulating mast cell progenitors or due to impaired skin colonization of mast cell progenitors or due to impaired survival or abrogated proliferation of mast cells within the skin. After 6 weeks of PMA treatment of normal mice, histologic analysis revealed that skin mast cells lacked expression of a proliferation marker, phospho-histone H3 (Ser10) (data not shown), indicating that, at least at this time point, mast cells did not proliferate locally. It is, however, possible that mast cells proliferated locally at earlier time points during the course of the inflammation. Progenitors with in vitro mast cell potential are present in peripheral blood of adult Kit^W/W mice,¹²  but these cells have not been accessible for direct analysis, partly due to lack of Kit expression that is an important marker of stem/progenitor cells. Hence, the precise stage at which mast cell development is absolutely blocked in adult Kit^W/W mice has to await the identification or development of specific mast cell progenitor markers. These could also be useful to study mast cell progenitor mobilization and their recruitment to healthy and inflamed tissues.

We thank Dr Michael Huber, Freiburg, for generous supply of BMMC, Thorsten Feyerabend for proficient cell transfers, Dr Hans Jörg Fehling for critical reading of the manuscript, and members of the animal facility for careful mouse keeping.

This work was supported by the Landesstiftung (P-LS-AL/11) (C.W. and H.-R.R.) and the Landesgraduierten-Förderung (S.M.S.) Baden-Württemberg and by the Deutsche Forschungsgemeinschaft (DFG-RO754/2-2) (H.-R.R.).

Contribution: C.W. and H.-R.R. designed research, analyzed data, and wrote the paper; C.W., S.B., S.M.S., and C.C. performed research.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Hans-Reimer Rodewald, Institute for Immunology, University of Ulm, D-89081 Ulm, Germany; e-mail: hans-reimer.rodewald@uni-ulm.de.