Epidermal Viral Immunity Induced by CD8α+ Dendritic Cells But Not by Langerhans Cells

See allHide authors and affiliations

Science  26 Sep 2003:
Vol. 301, Issue 5641, pp. 1925-1928
DOI: 10.1126/science.1087576


The classical paradigm for dendritic cell function derives from the study of Langerhans cells, which predominate within skin epidermis. After an encounter with foreign agents, Langerhans cells are thought to migrate to draining lymph nodes, where they initiate T cell priming. Contrary to this, we show here that infection of murine epidermis by herpes simplex virus did not result in the priming of virus-specific cytotoxic T lymphocytes by Langerhans cells. Rather, the priming response required a distinct CD8α+ dendritic cell subset. Thus, the traditional view of Langerhans cells in epidermal immunity needs to be revisited to accommodate a requirement for other dendritic cells in this response.

Langerhans cells (LCs) represent the dominant dendritic cell (DC) subset within the skin epidermis (1). They are a mobile cell population capable of efficient antigen capture (2) and migration out of the skin on activation by a variety of stimuli (35). At the same time, surface major histocompatibility complex (MHC) class II proteins and co-stimulatory molecules are up-regulated as a requirement for T cell priming (6, 7). These mature progeny of LCs contribute to DCs found within the T cell zones of skin-draining lymph nodes (LNs) (8, 9). This has led to the accepted view that LCs are the prototypic tissue-resident DCs, capable of acquiring antigen as a consequence of infection within peripheral compartments and migrating to draining LNs to directly prime T cells.

Consistent with this, early experiments using contact-sensitizing agents showed that antigen-bearing DCs thought to be of LC origin appeared within draining LNs within hours of exposure (5, 10). However, there have been few reports that show LC presentation after true infections of the skin. Further complicating this is the emerging realization that LCs are one of a number of distinct DC subsets that exist within the secondary lymphoid tissues (11). Thus, it becomes difficult to assume that DC presentation in skin-draining LNs necessarily depends on LC involvement. Given this, we examined the DC subsets involved in priming cytotoxic T lymphocytes (CTLs) after cutaneous herpes simplex virus 1 (HSV) infection. This route of inoculation results in virus replication within the epidermis, which should call resident LCs into play if this population is indeed involved in CTL priming.

At this time, at least six distinct LN-derived DC populations can be distinguished by differential expression of markers, including CD8α, CD4, CD205 (previously known as DEC205), and CD11b. Using a combination of CD205, CD11c, and CD8α, we compared the relative expression of these markers in cells that migrate out of ex vivo cultured epidermal sheets from flank skin and in DCs found in skin-draining LNs (brachial and inguinal LNs) or other tissues (mesenteric LNs) (Fig. 1A). DCs that emigrated from the skin epidermis consisted exclusively of a single CD205high population with little CD8α expression (Fig. 1A, left). A similar population was present in the skin-draining LNs, but not in the mesenteric LNs (Fig. 1A, middle and right). Some low-level CD8α expression was seen within this population, as reported previously for LCs once they reach the LNs (12). Combined, these data show that LCs are MHC class II+ (13) and that they express high levels of CD205 and either no or little CD8α.

Fig. 1.

(A) LCs are epidermal CD8αlow/–CD205high DCs found in the skin-draining (brachial and inguinal) LNs but not in the mesenteric LNs. Epidermal sheets from C57BL/6 mice were incubated for 2 days in medium that contained the chemokine 6Ckine (12). Migrating cells (left) were analyzed for expression of CD11c, CD205, and CD8α antigens by flow cytometry. DC subsets in the skin-draining LNs (middle) and the mesenteric LNs (right) were analyzed by flow cytometry. Single-cell suspensions were prepared and depleted of non-DCs and stained for CD11c, CD8α, and CD205. Cells were analyzed by flow cytometry gating on CD11c+ cells. The CD8αlow/–CD205high LCs are within the solid rectangles and their absence in the mesenteric LNs is shown by a dashed rectangle. (B) HSV infection is confined to the epidermal layer of skin. C57BL/6 mice were flank-inoculated withHSV. At various times after inoculation, skin from the infected site was removed and examined for the presence of HSV antigens (brown staining) by immunohistochemistry. Staining was performed on paraffin-embedded sections with rabbit antiserum to HSV and counterstained with hematoxylin. Magnification, 100×. Arrows denote the epidermal layer; arrowheads indicate the site of scarification. By 72 hours after infection, the epidermis had sloughed away from the underlying dermis because of infection.

To examine LC involvement in infection of the epidermis, where they represent the dominant DC subset, we employed a flank abrasion method of inoculation with HSV that results in an infection confined to the epidermal layer of the skin. Infection in this setting started at the point of scarification, with staining for virus antigen evident by 24 hours in the epithelial cells from the epidermis and hair follicles, peaking at around 2 days post-inoculation (Fig. 1B). Virus antigen expression never appeared in the dermis, remaining confined to the skin epidermis (Fig. 1B).

To determine which DC subset presented class I–restricted antigen after HSV skin infection, we used T cells derived from the gBT-I transgenic line, which express a T cell receptor specific for the immunodominant HSV glyco-protein B (gB) determinant gB498–505 (14). We have previously shown that only the conventional CD8αhigh DCs presented HSV-derived antigen to CTLs after subcutaneous footpad infection with this virus (15). To determine whether this population was also involved in antigen presentation after epidermal infection, CD11c+ DCs were isolated from the brachial LNs 2 days after flank inoculation with HSV, separated into three subsets on the basis of differential expression of CD45RA and CD8α, and used to stimulate gBT-I cells. Presentation was confined to a single CD8αhighCD45RA DC population (Fig. 2A). These cells represent the conventional CD8α+ DC subset (16) and are distinct from the CD45RA+ plasmacytoid DCs, some of which also express CD8α (17, 18). The lack of presentation by any other CD11c+ DC subset (Fig. 2A) suggested that LCs (found in the CD8αCD45RA population) were not responsible for priming to HSV. LC involvement was formally excluded by separation of CD11c+ DCs on the basis of CD205 and CD8α+ expression. In this experiment, CD8αlow/–CD205high LCs failed to stimulate T cells, whereas the CD8αhighCD205+ DCs provided good stimulation (Fig. 2B). Thus, all class I–restricted presentation of HSV antigen involved the conventional CD8αhigh DC subset with little or no LC involvement.

Fig. 2.

CD8αhigh DCs but not LCs or plasmacytoid DCs are capable of priming naïve T cells after HSV infection. DCs were isolated from the brachial LNs of C57BL/6 mice 2 days after flank inoculation with HSV. (A) The CD11c+ DCs were sorted into CD8 α+/– CD45RA+ plasmacytoid DCs, CD8αhighCD45RA DCs, and CD8α CD45RA DCs. (B) CD11c+ DCs were sorted into CD8αlow/– CD205high LC and CD8αhighCD205+ DC populations. All sorted DC subsets were then cocultured in vitro with gBT-I cells labeled with carboxyfluorescein diacetate succinimidyl ester (CFSE). After 60 hours of culture, gBT-I T cells were analyzed for proliferation by flow cytometry. When DC subsets as in (B) were tested for antigen presentation to a gB498–505-specific T cell hybridoma, no presentation by LCs was observed (fig. S1).

The recently described insensitivity of LCs to gamma irradiation (19) allowed us to next design an experiment that extended these ex vivo studies on isolated DC subsets to an in vivo approach. Combinations of the bm1 (Ly5.2+) and B6.Ly5.1 mouse strains, which differ in their Ly5 proteins, were used to generate bone marrow–chimeric mice to confirm the resistance of LCs to gamma irradiation. In this case, expression of Ly5 alleles was used to discriminate donor bone marrow–derived cells from radioresistant recipient cells. When B6.Ly5.1 mice were lethally irradiated and reconstituted with bm1 bone marrow (bm1 → B6.Ly5.1 chimeras), recipient B6.Ly5.1 DCs (Ly5.1+ cells) dominated the CD11c+ I-Ab+ LC emigrants from epidermal sheets (Fig. 3A). Recipient cells also represented the bulk of the CD8αlow/– CD205high LC subset in skin-draining LNs in these same mice, whereas the remaining LN CD11c+ DCs, including the CD8αhigh DCs, were predominantly of donor (bm1) origin and thus expressed the Ly5.2 marker (Fig. 3B). Given the radioresistance of LCs, the paucity of recipient DCs that expressed high levels of CD8α reinforces that this population is not derived from LC precursors.

Fig. 3.

(A) Recipient epidermal LCs are not replaced after lethal irradiation and bone-marrow reconstitution. The Ly5 phenotype was analyzed on DCs that migrated from epidermal sheets from bone marrow–chimeric mice. Epidermal sheets obtained from the flank skin of bm1 (Ly5.2) → B6.Ly5.1 or B6.Ly5.1 → B6 chimeric mice (not shown) were incubated for 2 days in medium that contained 6Ckine. Cells migrating into the medium were collected in separate 24-hour pools and combined for analysis. Gates were generated on the CD11c+I-Ab+ LCs, and these cells were analyzed for expression of Ly5.1 and Ly5.2. (B) Recipient DCs dominate the LC population in the skin-draining LNs of bone marrow–chimeric mice. DCs were isolated from the skin-draining LNs of naïve bm1 → B6.Ly5.1 and B6.Ly5.1 → B6 (not shown) chimeric mice. Single-cell suspensions were prepared and depleted of non-DCs, then stained with CD11c, CD8α, CD205, and Ly5.2 and analyzed by flow cytometry to determine the contribution of recipient and donor cells to the LN DC populations. (C) bm1 → B6.Ly5.1 mice, which possess epidermal H-2Kb+ LCs, were unable to prime a CTL response after HSV epithelial infection. Lethally irradiated recipients engrafted with either Kb+ B6.Ly5.1 bone marrow or Kbm1+ bone marrow from bm1 mice were flank-inoculated with HSV. Seven days later, gB-specific cytotoxicity was analyzed withan in vivo CTL assay. gB498–505 peptide–pulsed splenocytes were labeled with a high concentration of CFSE (CFSEhigh), and unpulsed control targets were labeled witha low concentration of CFSE (CFSElow). Four hours after transfer, spleen cells were analyzed for removal of the CFSEhigh population. The percent specific lysis was calculated as described in (20). Error bars represent the mean and SD of 3 to 5 mice per group. (D) bm1 → B6 and B6 → B6 mice were injected subcutaneously in the footpad with 106 in vitro–derived, B6, splenic, cultured DCs coated withgB498–505. Mice were assayed 5 days later for in vivo gB-specific CTL killing as in (C). In this case, however, killing was examined in the popliteal LN rather than in the spleen.

The bm1 mice were specifically chosen because they have mutations in the Kb MHC class I molecule that eliminates the CTL response to the gB498–505 determinant by altering presentation of this immunodominant peptide (13). Thus, if LCs are involved in antigen presentation, then gB-specific CTL priming should occur in the bm1 → B6.Ly5.1 chimeras, where the majority of LCs are of recipient (B6.Ly5.1) origin and therefore bear the responder Kb MHC class I element. Importantly, the other DC subsets, which are all radiosensitive, will be donor (bm1)–derived in these chimeras and not able to contribute to CTL priming because they express Kbm1 rather than Kb. In these animals, CTL priming (20) did not take place (Fig. 3C), despite the presence of Kb+ LCs in the skin-draining LNs (Fig. 3B). In marked contrast, CTL activity was found in control B6.Ly5.1 → B6 chimeras, where all DCs expressed Kb (Fig. 3C). The lack of response in the bm1 → B6.Ly5.1 chimeras was not due to the deletion of gB-specific CTLs by bm1-derived bone marrow cells, because bm1 → B6 chimeric mice could generate gB-specific CTLs when primed subcutaneously with B6 DCs that were coated with the gB498–505 determinant (Fig. 3D). In addition to a lack of virus-specific cytotoxicity, preliminary results have shown that these same chimeras failed to show detectable expansion of virus-specific CD8+ T cells, as measured by Kb-gB498–505 tetramer binding or intracellular interferon-γ staining (fig. S2).

These experiments provide in vivo evidence that a mechanism exists for priming HSV-specific CTLs after skin infection that does not require antigen presentation by LCs. This builds on work by Zhao et al. (21) that used mucosal infection with HSV and excluded LC presentation to virus-specific helper T cells. Combined, these studies suggest that LCs may not directly present class I– or class II–restricted antigens in this type of skin infection. Although some dissenting evidence exists (22), an absence of LC-mediated presentation would be a departure from the usual assumptions made about LC function, which are based upon a role in direct T cell stimulation (1). However, this long-standing notion has largely come from indirect evidence, with many studies that focus on individual aspects of LC acquisition, migration, or presentation (2, 23, 24) without formally proving that LCs actually present skin-acquired antigen for in vivo T cell priming.

We stress that we do not exclude LC migration from the skin, nor antigen uptake by these cells, only the idea that these processes result in direct class I–restricted antigen presentation and CTL priming to HSV. However, given their prominence in skin epidermis and their antigen uptake and migratory capacity, it is hard to imagine that LCs play no role in response to cutaneous infection with HSV. LCs may have an indirect role in CTL priming, possibly as specialized antigen acquisition elements that transfer their cargo to the priming CD8α+ DC population. Indeed, the bone-marrow chimeras show that the overwhelming bulk of skin DCs are of recipient origin, whereas presentation appears to be confined to the donor subset, raising the possibility that the presenting DCs do not directly acquire their antigen within the site of infection. The details of any such antigen transfer and the importance of LCs in this or other immune events will require a better understanding of the role of the individual DC subsets.

Supporting Online Material

Materials and Methods

Figs. S1 and S2

References and Notes

References and Notes

Stay Connected to Science

Navigate This Article