Cutting Edge: Migration of Langerhans Dendritic Cells Is Impaired in Autoimmune Dermatitis

Mice

MRL-lpr-, MRL^+/+-, and MHC (H-2^k)-matched healthy strains, including B10.BR, C3H/HeOuJ (C3H) mice, C57BL/6 (B6), and B6-Fas^lpr/lpr (B6-lpr) mice were purchased (The Jackson Laboratory) or bred locally. Animals were maintained in specific pathogen-free conditions and under institutional guidelines.

Flow cytometry

Gated FSC^highSSC^high CLN cells were analyzed for LC that were defined as CD207⁺CD11c^int/high cells among CD8α^−/low cells (where “FSC” is forward scatter, “SSC” is side scatter, and “int” is intermediate) (supplemental figure 1).⁴

In situ studies

Ear skin explants were harvested and dorsal halves separated with the aid of forceps. The dorsal halves were floated in 4% NH₄SCN to recover epidermal and dermal sheets that were fixed in acetone, blocked, and stained with conjugated Abs.

Skin cultures

Ear skin explants were cultured in the presence of IL-1β, TNF-α, and CCL19 for 4 days. Epidermal and dermal sheets were separated and stained slides were analyzed using a confocal microscope (supplemental figure 2).

In vivo migration

The dorsal side of the ear was painted with 1% FITC in dibutylphtalate-acetone. At the indicated time-points, mice were euthanized and their CLN harvested. The CLN cells were stained with indicated Abs to identify FITC⁺ cells by flow cytometry.

Statistics

Statistical comparisons between groups were made using two-tailed Student’s t test and ANOVA for comparison of two groups, and one-way ANOVA with Tukey as post hoc test were used when more than two groups were compared.

Additional information is provided in the supplemental material.

LC are reduced in the CLN of mice prone to develop autoimmune dermatitis

MRL-lpr mice that lack TCR α-chain have delayed skin inflammation (10), suggesting a pathogenic role of αβ T cells in lupus dermatitis. Because LC can activate self-Ag-specific T cells and initiate an autoimmune response (7), we surmised that autoimmune dermatitis-prone mice will have increased numbers of LC in the CLN. It was surprising, however, to find that the proportions of LC were markedly lower in MRL-lpr mice than in all control strains at all ages tested, whereas the proportions of myeloid DC (mDC) were equivalent between different strains (Fig. 1⇓, A and B). Total numbers of LC in CLN were also reduced in MRL-lpr mice at 5 wk of age before the onset of dermatitis that develops at ∼4–6-mo of age (Fig. 1⇓C, left panel) but similar to those in control strains at 12–14 wk of age when mDC numbers were 5- to 10-fold higher in MRL-lpr mice (Fig. 1⇓C, right panel). Consistently, CD8α⁻CD11c⁺CD205⁺ CD40^highMHCII^high cells that comprise LC were fewer in MRL-lpr CLN than in normal CLN (supplemental figure 1B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Analysis of LC in the CLN of autoimmune dermatitis-prone MRL-lpr and MRL^+/+ mice. A–C, Freshly isolated CLN cells were analyzed for LC that were defined as CD11c⁺CD207⁺ cells among large CLN cells with CD8α^high cells excluded (supplemental figure 1). Numbers represent LC (blue gate, CD11c⁺CD207⁺) and mDC (red gate, CD11c^highCD207⁻) as a percentage of gated cells. Proportions (B) and total numbers (C) of LC and mDC, gated as in A, of 5-wk-old and 12- to 14-wk-old mice are shown (∗, p < 0.01 vs C3H and B10.BR mice; n = 3–7 per group). B10, B10.BR; M^+/+, MRL^+/+; M-lpr, MRL-lpr. D–F, CLN cells from B6-lpr and B6 mice were analyzed for LC as in A (n = 2 per group, 10-wk old). Error bars represent mean ± SD. Results are representative of three independent experiments.

The reduced proportion of LC in the CLN of MRL-lpr mice is not a dilution effect due to lymphoproliferation in these mice (9), because hypercellular CLN of B6-lpr mice had a marked increase in total numbers of both LC and mDC while their proportions remained comparable to those in B6 mice (Fig. 1⇑, D–F). Furthermore, the reduction of LC in CLN precedes the onset of dermatitis (Fig. 1⇑, A–C) and worsens with age and disease development (supplemental figure 3, A–C). In fact, LC are rarely detected in the CLN of MRL-lpr mice with severe dermatitis (supplemental figure 3, A–C).

To further establish a link between LC reduction and dermatitis, we monitored MRL^+/+ mice that develop skin disease at ∼8–12 mo of age for dermatitis and analyzed LC at different ages. Results show that as compared with 5- or 13-wk-old MRL^+/+ mice (Fig. 1⇑, A–C), 8-mo-old disease-free MRL^+/+ mice have a 3- to 4-fold reduction and those with dermatitis have a 10- to 12-fold reduction of LC in CLN (supplemental figure 3, D–F). These data establish a link between the reduction in LC and the development and severity of dermatitis in MRL-lpr and MRL^+/+ mice. Thus, MRL-lpr mice that develop severe autoimmune dermatitis in early life have a more profound early defect in LC migration than MRL^+/+ mice that develop a delayed disease.

LC are abundant in the epidermis of MRL-lpr mice

The apparent lack of LC in CLN of MRL-lpr mice could be due to their reduced numbers in the skin. However, in situ staining of ear epidermis (Fig. 2⇓, A and B) as well as flow cytometry of epidermal cell suspensions (Fig. 2⇓C) showed no reduction in LC in the skin of MRL-lpr mice compared with control strains. In fact, LC numbers in the epidermis were significantly higher in MRL-lpr mice than in normal mice (Fig. 2⇓B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Analysis of LC in the epidermis. A, Fresh epidermal sheets recovered from dorsal ear halves of 12-wk-old mice were stained for MHCII and CD207 and analyzed using confocal microscopy. Shown are maximum projections of eight sections spanning ∼8–10 μm. Scale bars, 40 μm. All MHCII⁺ cells in the epidermis express CD207. Arrows indicate large activated LC. B, Epidermal LC numbers are shown as the mean ± SD of MHCII⁺CD207⁺ cells per mm² (∗, p < 0.05 vs C3H and B10.BR mice; n = 7–9 per strain). C, Epidermal cell suspensions were obtained by trypsin digestion of epidermal sheets harvested from 10- to 12-wk-old mice (n = 10–15 per strain). After an overnight culture, pooled cells were analyzed for MHCII and CD11c.

Most epidermal LC in 8- to 12-wk-old mice were evenly distributed and displayed an immature phenotype as shown by mostly intracellular MHCII and low expression of CCR7 and CD86 (Figs. 2⇑A and 3⇓A). Occasional large LC with MHCII redistributed to cell surface were seen in the epidermis of healthy strains (arrow in Fig. 2⇑A, B10.BR), likely representing matured LC before emigration. MRL-lpr epidermis, however, had large clusters of such activated LC (arrow in Fig. 2⇑A, MRL-lpr). Similar aggregates of large LC were also seen in the fresh epidermis of 8-mo-old MRL^+/+ mice several weeks before the onset of dermatitis (data not shown). We posit that such accumulation of activated LC reflects their inability to emigrate from the skin.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Ex vivo LC migration. Dorsal ear half explants from mice were cultured with cytokines for 4 days, after which epidermal (A–C) and dermal (D) sheets were removed, stained, and analyzed using confocal microscopy as described in supplemental Fig. 2. CCR7, MHCII, and CD86 staining in freshly isolated (A) and cultured (B) epidermal sheets are shown at maximum projections of eight sections spanning ∼8 μm. Scale bars, 40 μm. C, LC were counted in these epidermal sheets before and after culture. The emigration of LC from the epidermis was assessed by a reduction in epidermal LC counts after culture. Results are expressed as the percentage of emigrating cells that was calculated as (1 − [numbers of LC in cultured epidermal sheets/numbers of LC in fresh epidermal sheets]) × 100 (∗, p < 0.001, MRL-lpr vs all control strains; n = 4–6 per group, 9- to 12-wk old). Error bars represent mean ± SD. Results represent five independent experiments. D, Dermal sheets recovered from skin explants after culture were stained for MHCII and CD207. Confocal images show maximum projections of 8–10 sections spanning 15–20 μm. Original magnification was ×40. Note the cord-like formations in normal mice, but not in MRL-lpr mice. Data represent four separate experiments.

Reduced emigration of MRL-lpr LC ex vivo

Abundance of LC in the epidermis, but their reduced numbers in the CLN, of MRL mice might reflect the reduced capacity of LC to emigrate from skin. To test this, we used an ex vivo LC migration assay in which explants of dorsal ear halves were cultured in medium supplemented with IL-1β, TNF-α, and the CCR7 ligand CCL19 for 4 days. LC in the cultured epidermis matured and expressed higher levels of MHCII and CD86 than LC in fresh epidermis (Fig. 3⇑, A and B). LC in normal mice, but not in MRL-lpr mice, could be seen migrating through the dermis and into culture chambers (supplemental figure 2), suggesting a reduced emigration of MRL-lpr LC. These data also suggest that the impaired LC mobilization in these mice is not due to a lack of CCL19, IL-1β, and TNF-α, as also seen in psoriatic skin (11). We quantified this defect by enumerating LC numbers in the epidermis before and after cultures (Fig. 3⇑C). Whereas >70% of LC emigrated out of floating ear skin explants in control animals, the reduction of LC in epidermis was <30% in MRL-lpr mice.

Because LC migration to CLN depends on the chemokine receptor CCR7 (12), we determined CCR7 expression. Intriguingly, after cytokine exposure almost all LC express CCR7 in MRL-lpr epidermis as compared with none in normal epidermis (Fig. 3⇑B). In normal mice, CCR7 expression is instead seen in MHCII⁺ cells in the dermis and in culture elutes (supplemental figure 2). Thus, in normal mice all CCR7⁺ LC leave the epidermis, whereas MRL-lpr LC do not emigrate despite enhancement of CCR7. Similar findings, i.e., reduced LC migration despite normal expression of CCR7, are seen in apolipoprotein E^−/− mice that have increased low density lipoprotein (13), a feature they share with MRL-lpr mice (14). It is thus enticing to speculate that altered lipid metabolism in MRL-lpr mice (14) might contribute to reduced LC migration in these mice.

Impaired migration of LC through dermis (“dermal cord” formation) in MRL-lpr mice

When LC migrate through dermis along lymphatics (12), they appear in formations called “dermal cords” (Fig. 3⇑D, upper two rows). Such cords were not seen in the dermis of MRL-lpr mice (Fig. 3⇑D, lower row). Instead, there were striking aggregates of DC in the dermis, indicative of dysregulated migration. Furthermore, few dermal cells expressed CD207 in MRL-lpr mice, suggesting that LC do not reach the dermis in these mice.

Reduced in vivo migration of LC in MRL-lpr mice

To ascertain whether LC do not migrate from skin to CLN in vivo, we used an in vivo migration assay. Upon application of the irritant (FITC in dibutylphtalate-acetone) on the ear, skin DC acquire the dye, mobilize, and migrate to CLN where they can be visualized as FITC⁺ DC (Fig. 4⇓A). Results show that the proportion and total numbers of FITC⁺MHCII^high cells were lower in MRL-lpr mice than in control strains at all time points after painting (Fig. 4⇓, A and B). MRL^+/+ mice that develop delayed dermatitis also have a defect, albeit less profound than that in MRL-lpr mice.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

In vivo migration of LC. CLN cells from MRL-lpr and control strains were analyzed for FITC and MHCII at the indicated time points after FITC painting. A, Frequency of FITC⁺MHCII^high cells is denoted as a percentage of gated cells on dot plots. All FITC⁺MHCII^high cells were CD11c⁺ (not shown). Dot plot from a non-painted C3H mouse is shown as a control. ND, not done. B, Total numbers of FITC⁺MHCII^high cells, gated as in A, per CLN at the indicated time points after painting (∗, p < 0.05; n = 4 mice per group, 10- to 13-wk-old). Total CLN cellularity increased with time, particularly at 96 h, after FITC painting in normal mice (not shown). Data represent two independent experiments.

In summary, autoimmune dermatitis-prone mice exhibit a reduced migration of LC in the steady state as well as upon activation, resulting in accumulation of LC in the skin and their deficiency in the CLN. Such disruption in the homeostasis of LC precedes the onset of dermatitis and correlates with the development and severity of skin inflammation. Because LC play a role in the maintenance of peripheral tolerance (4, 15), such lack of LC in the CLN may cause loss of self-tolerance, thus initiating skin autoimmunity. However, B6 mice that lack LC do not develop dermatitis (8, 16), suggesting that the loss of LC alone is not sufficient to cause dermatitis. We speculate that the immobilized but activated LC in the skin of MRL mice activate autoreactive T cells directly in the peripheral tissue, thus perpetuating chronic skin inflammation. In fact, our preliminary data show that splenic T cells infiltrate the skin explants of MRL-lpr mice more efficiently than those of normal mice. Additionally, LC isolated from the epidermis of MRL-lpr mice produce higher levels of proinflammatory cytokines than LC from the epidermis of normal mice (data not shown). Improved understanding of mechanisms and consequences of impaired LC migration may open new therapeutic avenues for inflammatory skin diseases.