CCR7-independent transport of skin antigens occurs in the dermis

Abstract

Under homeostatic conditions, skin DCs migrate to regional LNs transporting self-antigens (self-Ags). The transport of self-Ags is considered to be critical for maintaining peripheral tolerance. Although the chemokine receptor CCR7 potently induces the migration of skin DCs to regional LNs, Ccr7^−/− (Ccr7-KO) mice do not show skin auto-immune diseases. To resolve this inconsistency, we examined Ccr7-KO epidermis- or dermis-hyperpigmented transgenic (Tg) mice, in which the transport of skin self-Ags is traceable by melanin granules (MGs). Under CCR7-deficient conditions, the transport of epidermal MGs to regional LNs was impaired at 7 weeks of age. However, epidermal MGs could be transported when they had accumulated in the dermis. Ccr7-KO-dermis-pigmented Tg mice confirmed the presence of CCR7-independent transport from the dermis. Compared with WT-dermis-pigmented Tg mice, the amount of transported melanin and number of MG-laden CD11c⁺ cells were both approximately 40% of the WT levels, while the number of MG-laden CD205⁺ or CD207⁺ cells decreased to about 10% in skin regional LNs of Ccr7-KO-dermis-pigmented Tg mice. Cell sorting highlighted the involvement of CD11c⁺ cells in the CCR7-independent transport. Here, we show that CCR7-independent transport of skin self-Ags occurs in the dermis. This system might contribute to the continuous transport of self-Ags, and maintain peripheral tolerance.

Introduction

Under homeostatic conditions, tissue-specific self-antigens (self-Ags) traffic from the periphery to regional LNs [[1-4]]. DCs take part in the transport of these self-Ags by capturing them in the periphery, as is the case for foreign Ags [[5, 6]]. This transport is considered to play a critical role in the induction and maintenance of peripheral immune tolerance [[7, 8]]. DCs expressing the chemokine receptor CCR7 migrate to peripheral LNs through attraction by its ligand, CCL21, expressed in afferent lymphatics and the T-cell zone of peripheral LNs [[9-11]]. Both Ccr7^−/− (Ccr7-KO) mice and CCL21-deficient paucity of lymph node T-cell (plt/plt) mice display a 40% decrease of CD11c⁺ DCs in peripheral LNs compared with normal control mice [[9, 10]]. Furthermore, both mice are prone to develop autoimmune disorders in the gastrointestinal tract, exocrine glands, pancreas, or kidney [[12, 13]]. Therefore, CCR7 is a potent regulator of the steady-state migration of DCs.

Epidermal Langerhans cells (LCs) and dermal DCs are representative skin DCs, and they participate in the transport of skin self-Ags [[2, 4]]. The migration of these skin DCs is also strongly regulated by CCR7, but other potent molecules regulating the migration of DCs at steady state in vivo have not been reported [[11, 14, 15]]. It is likely that the abrogation of CCR7–CCL21 interaction would impair the transport of skin self-Ags, and result in the disruption of peripheral tolerance in the skin. Nevertheless, neither Ccr7-KO mice nor plt/plt mice show severe autoimmune disorders in the skin [[9, 10]]. These observations raise the possibility that the transport of skin self-Ags is not indispensable for the induction and maintenance of peripheral tolerance in the skin.

We have investigated the transport of skin self-Ags to regional LNs using two lines of skin-hyperpigmented transgenic (Tg) mice. KRT14-Kitl-Tg (Kitl-Tg) mice and KRT14-HGF-Tg (HGF-Tg) mice show the hyperaccumulation of melanin granules (MGs) in the epidermis and the dermis, respectively [[16, 17]]. Since MGs are not dissolved and remain in the LNs, the Tg mouse system enables us to trace the transport of skin self-Ags to regional LNs by employing MGs as a tracer. Using the Tg mice, we found that skin MGs are continuously transported to regional LNs under homeostatic conditions [[2, 18-20]]. The Tg mouse system also revealed that Tgfb1^−/− mice, which lack LCs [[21]] cannot transport skin self-Ags from both the epidermis and the dermis, and that transport of epidermal self-Ags to regional LNs in plt/plt mice, lacking a subset of CCR7 ligands, does occur [[18]]. It cannot be ruled out, however, that the transport results from a subset of CCL21 expressed in afferent lymphatics of plt/plt mice [[14, 22]]. Therefore, it is necessary to clarify whether CCR7-independent transport of skin self-Ags occurs.

In this study, we demonstrate that CCR7-independent transport of skin self-Ags in a steady state does occur, by using Ccr7-KO mice. The dermis supports the transport of not only dermal but also epidermal self-Ags to regional LNs in the absence of CCR7.

Results

Transport of epidermal MGs to regional LNs is not completely disrupted in KO-Kitl-Tg mice

Kitl-Tg mice exhibit hypermelanocytosis in the epidermis [[16]]. Since epidermal MGs were transported to regional LNs by LCs in these mice, we have employed them as a system for tracing the steady-state trafficking of epidermal self-Ags by defining the MG-laden cells as cells captured and/or transported with skin self-Ags [[2, 18-20]]. To examine whether the transport of epidermal self-Ags was CCR7-dependent, we investigated the skin regional LNs of Ccr7-KO mice carrying Kitl-Tg. In 7-week-old Ccr7^+/−-Kitl-Tg (hereafter referred to as Kitl-Tg control) mice, skin regional LNs were pigmented by accumulated MGs (Fig. 1A, upper middle panel). The pigmentation was not observed in nonskin regional LNs but restricted to skin regional LNs as we reported previously [[2, 18-20]] (Fig. 1A, lower panels). On the other hand, the regional LNs of Ccr7-KO-Kitl-Tg double-mutant (KO-Kitl-Tg) mice were not pigmented (Fig. 1A, upper right panel). In the regional LNs of KO-Kitl-Tg mice, MG-laden cells were hardly detected (Supporting Information Fig. 1A). Cells positive for CD205 (a marker of LC descendants and part of dermal DCs) [[23]] or CD207 (langerin, a marker of LCs and part of dermal DCs) [[15, 24-27]] were decreased (Supporting Information Fig. 1A). Cells positive for CD68 (macrosialin, a marker of tissue macrophages [[28]]) were observed in the LNs of KO-Kitl-Tg mice, the same as in the LNs of Kitl-Tg control mice, and the expressions of MHC class II (professional APCs) and CD90.2 (T cells) in the T-cell area were diminished (Supporting Information Fig. 1A). This phenotype was consistent with that of original Ccr7-KO mice [[9]].

Next, we investigated the regional LNs in aged KO-Kitl-Tg mice as well as young ones. Unexpectedly, the regional LNs of 30-week-old KO-Kitl-Tg mice were pigmented (Fig. 1B, right panel). The accumulation of MGs was evident in the LNs, although the distribution of CD205⁺ and CD207⁺ DCs was not evident like in 7-week-old KO-Kitl-Tg mice (Supporting Information Fig. 1B). Measurement of the amount of transported melanin in MGs accumulated in the regional LNs by alkaline tissue dissolution methods [[19]] showed that the amount in the regional LNs of 30-week-old KO-Kitl-Tg mice (22.9 ± 4.5 μg) was significantly greater than that in non-Tg control (9.4 ± 2.2 μg), and was 11.8% of that of 30-week-old Kitl-Tg control mice (193.5 ± 24.6 μg) (Fig. 1C). The amount of melanin from accumulated MGs increased linearly with age in both mice (the constant rates of increase were 6.8 μg per week in Kitl-Tg control mice and 0.7 μg per week in KO-Kitl-Tg mice, Fig. 1D). These results indicated that a small amount of melanin in epidermal MGs was continuously transported to regional LNs even under CCR7-deficient conditions, and regional LNs would become pigmented in aged KO-Kitl-Tg mice.

Figure 1. The transport of epidermal melanin granules (MGs) to regional LNs is severely impaired in young KO-Kitl-Tg mice, but not in aged KO-Kitl-Tg mice. (A) Inguinal LNs of 7-week-old Ccr7^+/−-Tg⁻ mice (Tg⁻), Kitl-Tg control mice (Kitl-Tg cont.), and KO-Kitl-Tg mice are shown (top). Nonskin regional mesenteric LNs of each mouse are also shown for comparison (bottom). Bar = 1 mm. Data are representative of three independent experiments using one to two mice of both genotypes per group. (B) Inguinal LNs of 30-week-old Tg⁻ mice, Kitl-Tg control mice, and KO-Kitl-Tg mice are shown. Bar = 1 mm. (C) The amount of melanin accumulated in the regional LNs of 30-week-old Tg⁻ mice, Kitl-Tg control mice, and KO-Kitl-Tg mice was determined by an alkaline solubilization of MGs and spectrophotometry. Skin regional LNs from half of the body (see Materials and Methods) were taken from one to three mice per genotype and analyzed. Data are shown as mean ± SD of three to nine mice pooled from three independent experiments. *p < 0.05, two-sided Student's t-test. (D) The increase in the amount of accumulated melanin in regional LNs is plotted as the amount of melanin in skin regional LNs over time. The numbers of Kitl-Tg control mice and KO-Kitl-Tg mice used are as follows: n = 2 and 1, 10-week-old mice; n = 1 and 2, 11-week-old mice; n = 4 and 2, 20-week-old mice; and n = 7 and 7, 30-week-old mice. Data are shown as mean ± SD of measurements pooled from one (10- and 11-week-old) or two (20- and 30-week-old) independent experiments per week. y in the graph: the functions of regression curve. R²: R squared.

“Secondary” distribution of MGs was observed in the dermis of aged KO-Kitl-Tg mice

The accumulation of MGs in the LNs of aged KO-Kitl-Tg mice implied the CCR7-independent migration of epidermal Ag-transporting cells (i.e. LCs). To investigate which cells contained MGs in the skin, the skin of aged KO-Kitl-Tg mice was examined. The distribution of MGs was restricted to the epidermis in 7-week-old KO-Kitl-Tg mice (Fig. 2A). Interestingly, MGs were distributed in both the epidermis and the dermis of 30-week-old KO-Kitl-Tg mice (Fig. 2B). The epidermis–dermis border of the aged mice was clear and not disrupted. The distribution of MGs in the dermis was also observed in aged Kitl-Tg control mice (see Discussion). It should be noted that the skin of KO-Kitl-Tg mice consistently contained a larger amount of melanin than that in Kitl-Tg control mice (Fig. 2B). We speculated that the distribution of MGs in the dermis was correlated with the CCR7-independent transport of MGs to regional LNs. The presence of unknown transport of skin self-Ags could be examined by investigation of mice exhibiting hyperpigmentation only in the dermis.

Figure 2. Melanin granules (MGs) are distributed in the dermis of aged KO-Kitl-Tg mice. (A) Skin from the torso of 7-week-old Kitl-Tg mice. Skin from the torsos of Ccr7^+/−-Tg⁻ mice (Tg⁻) either stained with H&E (top left) or unstained (top right, dashed line is the border between the epidermis and the dermis) are shown as controls. The restricted distribution of MGs in the epidermis in the skin of both Kitl-Tg control mice and KO-Kitl-Tg mice are shown unstained (bottom). Black arrows indicate pigmented epidermis. E, epidermis; D, dermis. Bars = 100 μm. (B) Skin from the torso of 30-week-old Kitl-Tg mice. The distribution of MGs is observed in the dermis of both Kitl-Tg control mice (left) and KO-Kitl-Tg mice (right). The table shows the amount of melanin found in a 15.7 mm² area of ear skin. The numbers in parentheses indicate the numbers of mice examined. Data are shown as mean ± SD of three to six mice pooled from three independent experiments. Dashed lines indicate the border between the dermis and subcutaneous tissue. E, epidermis; D, dermis; S, subcutaneous tissue. Bars = 100 μm.

Dermal MGs were transported to skin regional LNs in mice lacking CCR7

HGF-Tg mice exhibit hypermelanocytosis in the dermis and exhibit pigmented skin regional LNs by transported dermal MGs [[2, 17]]. In both Ccr7^+/−-HGF-Tg (HGF-Tg control) mice and Ccr7-KO-HGF-Tg (KO-HGF-Tg) mice, MGs were distributed in the dermis but not the epidermis (Fig. 3A). In contrast to the LNs of 7-week-old KO-Kitl-Tg mice (Fig. 1A), skin regional LNs of 7-week-old KO-HGF-Tg mice were pigmented (Fig. 3B, right panel). Compared with that in HGF-Tg control mice, sparse distribution of MG-laden cells was observed in the LNs of KO-HGF-Tg mice, and CD11c (a pan-DC marker)⁺ cells, CD205⁺ cells, and CD207⁺ cells were decreased (Supporting Information Fig. 2). The decreases in the numbers of these marker-positive cells in the LNs of KO-HGF-Tg mice were 64.8% in CD11c⁺ cells, 23.8% in CD205⁺ cells, and 17.6% in CD207⁺ cells compared with those of HGF-Tg control mice (Table 1). The amount of melanin in the regional LNs of 7-week-old KO-HGF-Tg mice (17.8 ± 1.6 μg) was significantly larger than that in the LNs of non-Tg mice (6.3 ± 0.1 μg), and accounted for 42.5% of that of age-matched HGF-Tg control mice (41.9 ± 14.3 μg, Fig. 3C). The amount of melanin in both LNs increased constantly (the constant rates of increase were 8.5 μg per week in HGF-Tg control mice and 3.7 μg per week in KO-HGF-Tg mice, Fig. 3D). In addition, the amount of melanin in the LNs of 7-week-old KO-HGF-Tg mice reached 77.7% of that in the LNs of 30-week-old KO-Kitl-Tg mice (Fig. 1C and 3C). These results indicated that approximately 40% of dermal MGs were transported to regional LNs, and the transport was constant and more efficient than the transport of epidermal MGs under CCR7-deficient conditions.

Table 1. The number and ratio of total marker-positive cells and MG-laden marker-positive cells in the LNs of HGF-Tg control mice and KO-HGF-Tg micea

Marker	Number of total cells			Number of MG-laden cells
	HGF-Tg control	KO-HGF-Tg	Ratio: KO versus	HGF-Tg control	KO-HGF-Tg	Ratio: KO versus
			control			control
	(Number of cells per 0.106 mm^2)		(%)	(Number of cells per 0.106 mm^2)		(%)
^a The number of cells shown in graphs of Fig. 5 and the ratio of the number of cells in regional LNs of Ccr7-KO-HGF-Tg (KO-HGF-Tg) mice compared with that in regional LNs of Ccr7^+/–-HGF-Tg (HGF-Tg control) mice are shown.
CD11c	19.3 ± 4.9	12.5 ± 4.7	64.8 ± 24.5	6.8 ± 2.8	3.0 ± 1.6	44.1 ± 24.0
CD68	33.6 ± 6.4	22.7 ± 6.4	67.6 ± 19.0	23.0 ± 6.0	13.2 ± 4.6	57.4 ± 20.1
CD205	28.6 ± 5.5	6.8 ± 2.4	23.8 ± 8.4	15.7 ± 5.3	1.6 ± 0.8	10.2 ± 5.4
CD207	31.9 ± 6.0	5.6 ± 3.1	17.6 ± 9.7	19.5 ± 4.8	1.4 ± 1.2	7.2 ± 6.0
CD103	17.6 ± 2.9	8.5 ± 4.0	48.3 ± 22.5	7.0 ± 2.2	1.7 ± 1.3	24.3 ± 19.1

Figure 3. Transport of dermal melanin granules (MGs) to regional LNs is observed in 7-week-old dermis-pigmented KO-HGF-Tg mice. (A) Skin from the torsos of 7-week-old HGF-Tg control mice and KO-HGF-Tg mice. Skin from the torso of a Ccr7^+/−-Tg⁻ mouse (Tg⁻) was stained with H&E (left) or left unstained (right) and shown as controls (top). The restricted distribution of MGs in the dermis in both HGF-Tg control mice (left) and KO-HGF-Tg mice (right) are indicated by the arrows (bottom). E, epidermis; D, dermis; S, subcutaneous tissue. Bars = 100 μm. Data are representative of at least two independent experiments using one to two mice of both genotypes per group. (B) Inguinal LNs of 7-week-old Tg⁻ mice, HGF-Tg control mice, and KO-HGF-Tg mice are shown. Bar = 1 mm. The representative data from at least four independent experiments using one to two mice of both genotypes per group are shown. (C) Melanin accumulates in the regional LNs of 7-week-old HGF-Tg mice. Data are shown as mean ± SD of three to nine mice pooled from three independent experiments. *p < 0.05, two-sided Student's t-test. (D) Melanin accumulates in the regional LNs over time. The amount of melanin in skin regional LNs is shown. The numbers of mice used were as follows: n = 2 and 2 in 3-week-old HGF-Tg control mice and KO-HGF-Tg mice, n = 5 and 8 in 7-week-old HGF-Tg control mice and KO-HGF-Tg mice, n = 4 and 3 in 30-week-old HGF-Tg control mice and KO-HGF-Tg mice, respectively. Data were pooled from one (3-week-old), two (30-week-old), or three (7-week-old) independent experiments per week, and shown as mean ± SD. y in the graph: the functions of regression curve.

CD11c⁺ cells were involved in the CCR7-independent transport of dermal self-Ags

It is reported that DCs decreased in skin regional LNs of Ccr7-KO mice [[9, 14]]. Therefore, whether DCs take part in the CCR7-independent transport of skin MGs under steady-state conditions or not should be important. To investigate which cells were involved in the transport, we examined MG-laden cells in regional LNs by immunohistochemistry and cell sorting.

In Kitl-Tg control mice of both 7 weeks and 30 weeks of age, CD205⁺, CD207⁺, or CD68⁺ MG-laden cells were present in the regional LNs (Fig. 4B and C, and Supporting Information Figs. 3 and 4). These marker-positive MG-laden cells were hardly detected in the LNs of 7-week-old KO-Kitl-Tg mice because of the marked decrease of MG-laden cells themselves (Fig. 4B and Supporting Information Fig. 3, “KO-Kitl-Tg (7 wk)”). Although MG-laden cells were observed in the regional LNs of 30-week-old KO-Kitl-Tg mice, very few CD205⁺ MG-laden cells were detected, and CD207⁺ MG-laden cells were rarely found. CD68⁺ MG-laden cells were detected in LNs of KO-Kitl-Tg mice (Fig. 4C and Supporting Information Fig. 4, “KO-Kitl-Tg (30 weeks)”). These CD68⁺ MG-laden cells seemed to contain many MGs compared with either CD205⁺ or CD207⁺ MG-laden cells in KO-Kitl-Tg LNs (Fig. 4C).

Figure 4. Melanin granules (MGs) are transported in the regional LNs of aged KO-Kitl-Tg mice, while most of these cells are either CD205- or CD207-negative. (A) Representative marker-positive (Marker(+)) cells, MG-laden cells, and marker(+) MG-laden cells. Marker(+) MG-laden cells were detected by merging bright field images with immuno-stained images. Left and middle panels are enlarged views from a CD205-stained merged picture, and a right panel is enlarged view from a CD68-stained merged picture of 7-week-old Kitl-Tg control LNs (indicated in Fig. 4B). Note a few MGs in CD205⁺ cells (DCs), while cells containing many MGs are not CD205⁺ but CD68⁺ (macrophages). Bar = 25 μm. (B) Marker-positive MG-laden cells in the LNs of 7-week-old Kitl-Tg control mice and KO-Kitl-Tg mice. Bar = 50 μm. (C) Marker-positive MG-laden cells in the LNs of 30-week-old Kitl-Tg control mice and KO-Kitl-Tg mice. Bar = 50 μm. Data shown are representative of at least three independent experiments using one to two mice of both genotypes. White arrowheads: marker(+) cells, black arrowheads: MG-laden cells, yellow arrowheads: marker(+) MG-laden cells.

In the regional LNs in 7-week-old HGF-Tg control mice, many CD11c⁺, CD205⁺, or CD207⁺ MG-laden cells were observed, while these cells were decreased in the LNs of KO-HGF-Tg mice (Fig. 5 and Supporting Information Fig. 5). Compared with that in HGF-Tg control mice, the number of CD11c⁺ MG-laden cells was decreased to 44.1% (6.8 ± 2.8 cells to 3.0 ± 1.6 cells per 0.106 mm² of LN specimen), while the decreases in the numbers of CD205⁺ MG-laden cells and CD207⁺ MG-laden cells were remarkable: 89.8% (15.7 ± 5.3 cells to 1.6 ± 0.8 cells) and 92.8% (19.5 ± 4.8 cells to 1.4 ± 1.2 cells) in the LNs of KO-HGF-Tg mice, respectively (Fig. 5 and Table 1). The presence of CD103 (αE integrin)⁺ MG-laden cells was also examined because CD103 was recently reported to be a marker of a subset of dermal DCs defined as CD207⁺CD103⁺ dermal DCs, which can cross-present skin Ags [[29]]. CD103⁺ cells were distributed in the T-cell area of the LNs of HGF-Tg control mice. These cells were also observed in the LNs of KO-HGF-Tg mice (Supporting Information Fig. 2). CD103⁺ MG-laden cells were observed in both HGF-Tg control and KO-HGF-Tg LNs, and the decrease of CD103⁺ MG-laden cells was 75.7% in KO-HGF-Tg mice (7.0 ± 2.2 cells to 1.7 ± 1.3 cells, Fig. 5B, Supporting Information Fig. 5, and Table 1). CD68⁺ MG-laden cells were decreased by 42.6% in the LNs of KO-HGF-Tg mice compared with that of HGF-Tg control mice (23.0 ± 6.0 cells to 13.2 ± 4.6 cells, Fig. 5A and Table 1). The CD68⁺ MG-laden cells also seemed to contain many MGs compared with either CD205⁺ or CD207⁺ MG-laden cells, the same as in 30-week-old KO-Kitl-Tg mice.

Figure 5. Both CD205⁺ melanin granule (MG)-laden cells and CD207⁺MG-laden cells are dramatically decreased in the regional LNs of 7-week-old KO-HGF-Tg mice. Marker⁺MG-laden cells in the LNs of 7-week-old HGF-Tg mice were evaluated as in Fig. 4, and the numbers of marker⁺ cells either with or without MGs (total cells) and that of marker⁺MG-laden cells (MG-laden cells) were counted in the specimens. (A) CD11c⁺MG-laden cells and CD68⁺MG-laden cells in the regional LNs of HGF-Tg control mice and KO-HGF-Tg mice. White arrowheads: marker(+) cells, black arrowheads: MG-laden cells, yellow arrowheads: marker(+) MG-laden cells. Bar = 50 μm. (B) CD205⁺, CD207⁺, and CD103⁺MG-laden cells in the regional LNs of HGF-Tg mice and KO-HGF-Tg mice. (A and B) The number of each cell type in the LNs is indicated by graphs below pictures. The number of cells in 0.106 mm² is indicated as the mean + SD of 2-4 mice/genotype, pooled from two independent experiments. *p < 0.05, two-sided Student's t-test.

To verify the involvement of CD11c⁺ cells in the CCR7-independent transport, sorting of CD11c⁺ cells from the LN cells of KO-HGF-Tg mice was performed. MG-laden DC-like-shaped cells were enriched in CD11c⁺ fraction, but these cells were rarely detected in CD11c⁻ fraction of LN cells of KO-HGF-Tg mice as well as HGF-Tg control mice (Supporting Information Fig. 6). Since the dermis contains γδ T cells, and their migration was reported in bovines [[30-32]], the contribution of γδ T cells to the CCR7-independent transport was also investigated. FACS analysis revealed that γδ TCR⁺ cells constituted 1.4 and 0.9% of presorted LN cells in HGF-Tg control mice and KO-HGF-Tg mice, respectively (Supporting Information Fig. 7B, “presort”). Among sorted γδ TCR⁺ cells, 0.34% of γδ TCR⁺ cells of KO-HGF-Tg carried MGs. This proportion was not so different from that in HGF-Tg control (0.24%, Supporting Information Fig. 7C).

Altogether, these results indicated that CD11c⁺ DC-lineage cells should be involved in the CCR7-independent transport of dermal self-Ags in a steady state, although CD205⁺ DCs, CD207⁺ DCs, or γδ T cells might not contribute significantly to the transport.

Discussion

In this study, we demonstrated the presence of CCR7-independent transport of skin self-Ags to regional LNs in a steady state. The transport functions in the dermis, and may guarantee the continuous transport of skin self-Ags from both the epidermis and the dermis. Impaired accumulation of MGs (Fig. 1) and clear decrease of both CD205⁺ cells and CD207⁺ cells in the LNs of 7-week-old KO-Kitl-Tg mice (Fig. 4 and Supporting Information Fig. 1 and 3) suggested that these cells (probably LCs) should be responsible for the transport of MGs from the epidermis. These results may emphasize the strict regulation of the steady-state migration of LCs by CCR7 as reported previously [[14]]. It is reported that CCR7 is required for the entry of LCs into afferent lymphatics in the dermis but not for the mobilization of LCs from the epidermis to the dermis [[14]]. Our result that the amount of melanin in the skin of 30-week-old KO-Kitl-Tg mice was greater than that in 30-week-old Kitl-Tg control mice suggests that MG-laden Ccr7-KO LCs could not enter afferent lymphatics and remained in the dermis. CCR7-independent dermal Ag-transporting cells might capture and transport the remaining LCs, which resulted in the accumulation of MGs in the LNs of aged KO-Kitl-Tg mice. In contrast, dermal self-Ags were transported more efficiently than epidermal ones under CCR7-independent conditions. It is possible that the free entrance of dermal MGs into afferent lymphatics like soluble compounds and proteins [[33]] resulted in the accumulation of MGs in the regional LNs of KO-HGF-Tg mice. However, because the transport of dermal MGs to regional LNs was completely abolished in “LC-lacking” Tgfb1^−/−-HGF-Tg mice [[2]], the CCR7-independent transport of dermal self-Ags should be a cell-mediated process, and the cells should develop in a manner dependent on TGF-β₁.

It is well known that dermis contains several immune-competent cells [[34]]. Our results showed that the regional LNs of 7-week-old KO-HGF-Tg mice contained approximately 40% of the amount of melanin in HGF-Tg control mice. An interesting consistency was observed in the decrease of the number of CD11c⁺ MG-laden cells (44.1 ± 24.0%). Enrichment of MG-laden cells in CD11c⁺ fraction by cell sorting further suggested the involvement of CD11c⁺ cells in the CCR7-independent transport of dermal self-Ags. However, it is doubtful that either CD205⁺ dermal DCs or CD207⁺ dermal DCs transported approximately 40% of dermal MGs to regional LNs because of the remarkable decrease of these cells in regional LNs of KO-HGF-Tg mice and the few MGs in these cells histologically. Recent reports that the migration of CD207⁺ dermal DCs to regional LNs is CCR7-dependent, as well as LCs [[15]], or CD207⁺ dermal DCs could develop and be maintained without TGF-β₁ [[35, 36]], may also support our speculation. Recently, CD207⁺CD103⁺ migratory dermal DCs were reported as cells that cross-present skin self-Ags in regional LNs [[29]]. The decrease of both total and MG-laden CD103⁺ cells was not parallel to the decrease of CD207⁺ cells in our KO-HGF-Tg mice, suggesting that the CD103⁺ cells might not correspond with CD207⁺CD103⁺ dermal DCs. As CD103 is also a marker of a subset of CD4⁺ T cells in LNs [[37]], a part of CD103⁺ cells that we observed might include T cells. In any case, since a very small number of CD103⁺ MG-laden cells are present, the details of the CD103⁺ cells, including the precise lineage and the contribution to the transport of skin Ags under CCR7-sufficient/deficient conditions, should be clarified in future work. Skin contains γδ T cells in both the epidermis and the dermis. Although epidermal γδ T cells are known to be immobile, the CCR7-independent migration of dermal γδ T cells in bovines or skin T cells in mice to regional LNs was reported ([32] and M. Tomura, unpublished observations). In contrast, a significant decrease of γδ T cells in skin regional LNs of Ccr7-KO mice was also reported [[31]]. At least, we think that γδ TCR⁺ cells in the LNs of KO-HGF-Tg mice are too few to support 40% of transport of dermal self-Ags (Supporting Information Fig. 7). This result might also underscore the immobility of epidermal γδ T cells. Do CD68⁺ macrophages contribute to the CCR7-independent transport of self-Ags? Large MG-rich cells were exclusively CD68⁺ but DC lineage marker-negative. Although the migration of macrophages from the dermis or lung to regional LNs was reported [[34, 38]], the capacity of macrophages to transport self-Ags is still unclear. In Tgfb1^−/− mice, F4/80⁺ or CD11b⁺ macrophages are present in peripheral LNs as well as normal mice [[21, 39]]. These macrophages would include CD68⁺ cells because CD68 is a broad marker for tissue macrophages [[40]]. CD68⁺ macrophages might contribute to capture MGs transported by numerous migratory cells in the LNs rather than the MGs transported from the dermis [[2]].

Interestingly, the distribution of MGs in the dermis was also observed in CCR7-sufficient aged Kitl-Tg control mice, implying that not all LCs enter afferent lymphatics even under CCR7-sufficient conditions in a steady state. Our Tg system using insoluble MGs has an advantage for tracing the trafficking and retention of skin self-Ags in vivo. Most of the remaining LCs might be captured by dermal resident cells such as macrophages, while a part of the remaining LCs might be captured by dermal Ag-transporting cells as occurs in CCR7-deficient mice. Under CCR7-sufficient conditions, epidermal self-Ags might be transported by both LCs and dermal DCs, which may contain CCR7-dependent or -independent ones.

We previously reported the transport of epidermal MGs to regional LNs of 7-week-old plt/plt-Kitl-Tg mice as well as control +/plt-Kitl-Tg mice and proposed the CCR7-independent trafficking of skin self-Ags [[18]]. This transport should be caused by the CCR7-dependent migration of LCs by another subset of CCL21 (CCL21-leucine) expressed in afferent lymphatics of plt/plt mice [[14, 22]]. However, epidermal self-Ags have actually been transported by CCR7-independent dermal system. Whereas we do not have any information about other organs, the skin might have an advantage in the transport of self-Ags by exhibiting distinct mechanisms in the epidermis and the dermis. The cause of auto-immune disorders observed in Ccr7-KO mice might be a lack of the CCR7-independent transport of self-Ags in affected organs, although the cause is also explained as the disruption of central tolerance [[12, 13]]. Meanwhile, our results showed that the numbers of both MG-laden cells and transported self-Ags in regional LNs of Ccr7-KO-Tg mice were smaller than those of Ccr7-sufficient Tg mice. This raises the question of whether the small number of self-Ags is actually sufficient to maintain peripheral tolerance. The level of self-Ags presented in the LNs influenced the induction of peripheral tolerance [[41, 42]]; however, the constant transport of skin self-Ags might play a critical role in the tolerance even if they are few in number. Subset(s) of Ag-transporting cells might also influence the induction of peripheral tolerance. Further studies should be undertaken to clarify this issue.

In conclusion, although molecule(s) regulating the CCR7-independent transport and the transporting cells should be examined more precisely, our results revealed a novel mechanism for the transport of skin self-Ags. This mechanism should provide a clue to understanding the relationship between peripheral self-Ags and immune regulation.

Materials and methods

Mice

C57BL/6JJcl (B6Jcl) was purchased from Clea Japan, Inc. Tg(KRT14-Kitl)1Takk and Tg(KRT14-HGF) are Tgs expressing mouse Kitl and human HGF driven by KRT14, respectively (herein called Kitl-Tg and HGF-Tg) [[16, 17]]. B6-Ccr7-KO mice (kindly provided by Dr. M. Lipp, MDC, Berlin) were twice crossed with Kitl-Tg mice or HGF-Tg mice to obtain F2 progeny. Genotyping was carried out by PCR of genomic DNA. All animals were maintained under SPF conditions at The Division of Laboratory Animal Science, Research Center for Bioscience and Technology, Tottori University. Experiments were approved and performed in accordance with the guidelines of the Animal Care and Use Committee of Tottori University.

Histological analysis

Trunk skin was embedded in polyester wax (BDH) and cut at 7 μm thickness. The specimens were stained with H&E or unstained for the detection of MGs. Skin regional LNs (axillary, brachial, and inguinal LNs) were embedded in optimal cutting temperature compound (Sakura Finetec), and cryostat sections were cut at 7 μm thickness. The specimens were stained with antibodies (Abs): anti-CD205 (NLDC145) kindly provided by Dr. K. Inaba (Kyoto University, Kyoto, Japan) and anti-CD207 (eBioL31) kindly provided by Dr. K. Kabashima (Kyoto University); anti-CD68 (FA11) (Serotec), anti-MHC class II (M5/114.15.2), anti-CD207 (eBioRMUL.2), and anti-CD103 (2E7) (eBioscience); anti-CD11c (HL3) and anti-CD45R-FITC (B220, RA3-6B2) (BD); anti-CD90.2 (Thy1.2, 30-H12) and anti-CD117 (Kit, ACK2), prepared in our laboratory. Biotinylated rabbit anti-rat IgG (H+L) Ab (BA-4001, Vector) and goat anti-Armenian hamster IgG Ab (SC-2445, Santa Cruz, CA, USA) were used as secondary Abs. ExtrAvidin®-FITC or ExtrAvidin®-tetramethylrhodamine-5-(and 6-)isothiocyanate (TRITC, Sigma-Aldrich) was used for detection. The number of marker-positive MG-laden cells was counted per defined field area (0.106 mm²) under a microscope (BX60, Olympus), and a total of 10 fields were recorded. B-cell follicles and medullary sinuses were excluded from the count area. Marker-positive MG-laden cells were defined on immunostained views merged with bright field views. Images were taken using a microscope (BX60 and Leica DMRB, Leica Microsystems) and digital cameras (Polaroid PDMC II I, Nihon Polaroid, and Nikon DS-U2/ DS-Ri1, Nikon). All images were processed using Adobe® Photoshop® CS3 software (Adobe Systems Inc.).

Measurement of accumulated melanin in tissues

The amounts of melanin accumulated in skin regional LNs and ear skin were measured by an alkaline solubilization technique and spectrophotometry as described previously [[19]]. Skin regional LNs taken from the right side of the body from a mouse (submandibular, axillary, brachial, inguinal, and popliteal LNs) and 15.7 mm² of ear skin after the removal of coating were used for the assay.

Flow cytometric analysis and cell sorting

Skin regional LNs were homogenized with slide glasses. For sorting CD11c⁺ cells, cell suspension was labeled with an anti-CD11c-phycoerythrin (PE) Ab (GK1.5, BD), and PE-single-positive cells were sorted using a Mo-Flo XDP cell sorter (Beckman Coulter). For sorting γδ T cells, cell suspension was labeled with a cocktail of biotinylated anti-CD4 (GK1.5, prepared in our laboratory), CD8α (53-6.7, BD), CD11b (M1/70, Beckman Coulter), and CD11c (N418, AbD Serotec) Abs to deplete αβ T cells, macrophages, and DCs. Labeled cells were separated using MACS® system (Miltenyi Biotec). Cocktail-negative cells were then labeled with anti-γδ TCR-PE Ab (GL3, eBioscience) and anti-B220 FITC Ab (RA3-6B2, BD). PE-single-positive cells were then sorted. Cell analysis and sorting were carried out using an EPICS®-XL™ cell analyzer and a Mo-Flo XDP cell sorter. Data were analyzed using WinMDI version 2.8 software.

Statistical analysis

Statistical difference was assessed by two-sided Student's t-test using Microsoft® Excel® 2008 for Mac (Microsoft). Data are indicated as the mean ± SD.

Acknowledgments

We thank Dr. M. Lipp (Max-Delbrück Center for Molecular Medicine, Germany) and Dr. Y. Takahama (Tokushima University, Japan) for providing Ccr7-KO mice, Dr. T. Kunisada (Gifu University, Japan) for providing Kitl-Tg mice and HGF-Tg mice, Dr. K. Inaba and Dr. K. Kabashima (Kyoto University, Japan) for providing Abs, and Dr. H. Yamazaki and Dr. T. Yamane (Mie University, Japan) for warm encouragement and helpful discussions. We also thank Dr. Y. Nakayama and Ms. H. Miyauchi (Research Center for Bioscience and Technology, Tottori University) for technical assistance in the cell sorting, and Ms. T. Shinohara for technical assistance. This work was supported by the discretionary expenses of the president of Tottori University (M.Y.) and a Grant-in-Aid for Scientific Research (C) from the Ministry of Education, Culture, Sports, Science, and Technology (S.-I.H). A.M. and K.O. are Research Fellows of the Japan Society for the Promotion of Science.

Conflict of interest

The authors declare no financial or commercial conflict of interest.

Abbreviations

Kitl

kit ligand

LC

Langerhans cell

MG

melanin granule

TRITC

tetramethylrhodamine-5-(and 6-)isothiocyanate