Langerhans cells transfer targeted antigen to dermal DC and acquire MHC-II in vivo

To the Editor

Dendritic cells (DC) in the skin are critical for the development of adaptive immune responses to pathogens and for the maintenance of peripheral tolerance (Kashem et al. 2017). Under steady-state conditions, at least four subsets of cutaneous DC have been characterized: epidermal Langerhans cells (LC), dermal classical DC1 (cDC1, also known as CD103+ dDC), dermal cDC2 (also known as CD11b+ dDC) and CD103− CD11b− double negative (DN) DC. Though functional specializations of each DC subset have been well studied, the functional interaction between these DC subsets under steady-state conditions is poorly understood.

Targeting antigen to DC using monoclonal antibodies specific for C-type lectin receptors (CLR) offers an efficient method to study adaptive responses to antigen presentation by targeted populations of DC in vivo and is being developed for therapeutic vaccinations (Steinman and Banchereau 2007). We have previously developed BAC transgenic mice (huLang) with selective expression of human Langerin, a CLR, on epidermal LC (Bobr et al. 2010). Immunization of these mice with as little as 0.05ug of α-huLangerin mAb conjugated to the 2W1S model antigen (α-huLang-2W1S) efficiently and selectively targeted LC resulting in expansion endogenous 2W1S-specific CD4+ T cells which could be detected using 2W1S:I-A^b tetramer (Yao et al. 2015).

To examine whether LC directly induce T cell responses after antigen targeting, we crossbred huLang mice with mice that were constitutively ablated of MHC-II selectively in LC (huLang LC^ΔMHC-II mice) (Igyarto et al. 2009). LC in these mice express human Langerin allowing for targeting with α-huLang-2W1S and should also lack expression of MHC-II. To confirm the loss of MHC-II, single cell epidermal suspensions were isolated from flank skin and examined by flow cytometry. As expected, LC (gated as MHC-II+, Langerin+, CD45+) from huLang LC^ΔMHC-II mice expressed levels of MHC-II equivalent to MHC-II^−/− mice (Figure 1A, B).

Figure 1. MHC-II depletion on huLang LC^ΔMHC-II mice is complete.

(A) MHC-II MFI and (B) percentage of MHC-II gated on epidermal Langerin+, CD11b+ LC from WT, MHC-II^−/−, and huLang LC^ΔMHC-II mice. (C) The number of 2W1S:I-A^b-specific T cells in huLang, huLang LC^ΔMHC-II and littermate control mice at day 7 after i.p. immunization with 1ug α-huLang-2W1S. (D) CFSE dilution of OT-II T cells co-cultured with FACS purified epidermal LC from WT, MHC-II^−/−, and huLang LC^ΔMHC-II mice for 4 days. LCs were pre-incubated with 10 ug/ml OVA_323–339 peptide for one hour before co-culture. (E) CFSE dilution and (F) percentage of CFSE diluted OT-II T cells 4 days after co-culture with epidermal LC from huLangerin and huLang LC^ΔMHC-II mice that have been immunized in vivo with 10 ug i.p. α-huLang-OVA_323–339. Each symbol represents data from an individual mouse. **** p<0.0001, *** p<0.001, n.s. not significant.

Next, huLang LC^ΔMHC-II and huLang mice were injected i.p. with 1 μg α-huLang-2W1S. As expected, endogenous 2W1S:I-A^b tetramer binding CD4+ T cells isolated from lymph nodes and spleen 7 days later had expanded ~10-fold in huLang mice compared to WT lacking huLang expression (Figure 1C). Surprisingly, 2W1S:I-A^b-specific T cells expanded to a similar extent in huLang LC^ΔMHC-II mice. To ensure that LC in huLang LC^ΔMHC-II mice did not retain levels of MHC-II expression below the sensitivity of flow cytometry, we next FACS sorted LC from the epidermis and cultured them in vitro with OVA_323–339 peptide and naïve CFSE labeled OT-II T cells. Sorting purity for this and all subsequent experiments was >95% (data not shown). After 4 days, robust dilution of CFSE was observed in OT-II cells incubated with LC isolated from WT mice but not in cells incubated with LC isolated from MHC-II^−/− or huLang LC^ΔMHC-II mice (Figure 1D). This was further confirmed by targeting LC in vivo with 10 ug i.p α-huLang-OVA_323–339. Three days later epidermal LC were FACS sorted and co-cultured with CFSE labeled OT-II cells. Four days later after co-culture, LC from huLang mice induced OT-II proliferation, but LC from huLang LC^ΔMHC-II mice did not (Figure 1E,F). Targeted LC from huLang LC^ΔMHC-II mice incubated with OT-II and unfractionated lymph node (LN) cells from WT mice as a source of MHC-II+ cells were also unable to present antigen. From these experiments, we conclude that epidermal LC in huLang LC^ΔMHC-II mice fully lack functional MHC-II, yet targeting these cells using α-huLang-2W1S results in expansion of antigen-specific T cell responses in vivo.

Since i.p administration of α-huLang-2W1S targets LC in both the epidermis and skin draining LNs (dLN) (Yao et al. 2015), we next checked whether LC from skin dLNs also lack functional MHC II. LC from skin dLNs were sorted from huLang and huLang LC^ΔMHC-II mice three days after targeting with 10 μg α-huLang-OVA_323–339. LC were then cultured in vitro with OT-II cells as above. To our surprise, LC from skin dLNs isolated from huLang LC^ΔMHC-II mice induced OT II T cell proliferation, though less efficiently than LC from WT mice (Figure 2A). Since LC cannot acquire MHC-II from MHC-II+ cells in culture (Figure 1E–F), LC in huLang LC^ΔMHC-II mice must have acquired functional MHC-II in vivo once they exited the epidermis. Studies have shown that APCs can acquire MHC-II through trogocytosis when in contact with other APCs (Herrera et al. 2004; Segura et al. 2007). To test if MHC-II-deficient LC can acquire MHC-II from other cells, we compared MHC-II expression of epidermal LC and LN LC isolated from huLangerin and huLang LC^ΔMHC-II mice. Indeed, MHC-II was evident on LN LC, but not epidermal LC isolated from huLang LC^ΔMHC-II mice (Figure 2B). Since ‘cross-dressing’ of MHC-II leads to characteristic clustered molecule distribution (Liu et al. 2016) (Nakayama 2015), we checked the expression pattern of MHC-II on LC from huLang and huLang LC^ΔMHC-II mice using ImageStream. As expected, in control huLang mice expression of MHC-II was higher and expression of both muLangerin and huLangerin was lower on dLN LC than epidermal LC (Figure 2C). In huLang LC^ΔMHC-II mice MHC-II was absent from epidermal LC but “patchy” expression was evident on the LN LC suggesting that their MHC-II was acquired from other cells.

Figure 2. LC acquire MHC-II in the LN and transfer antigen to dermal cDC2 and DN DC.

(A) Percentage of CFSE diluted OT-II T cells co-cultured with FACS sorted LC (gated as CD11c⁺, huLangerin⁺, muLangerin⁺, CD11b⁺, CD103⁻) from skin dLNs of huLangerin and huLang LC^ΔMHC-II mice that have been immunized in vivo with 10 ug i.p. α-huLang-OVA_323–339. (B) MHC-II MFI of epidermal LC and skin dLN LC from huLangerin and huLang LC^ΔMHC-II mice. (C) ImageStream analysis showing MHC-II, huLangerin and muLangerin distribution of epidermal LC and skin dLN LC from huLangerin and huLang LC^ΔMHC-II mice. (D) Percentage of proliferated OT-II T cells 4 days after co-cultured with FACS sorted APC subsets from skin dLNs of huLang mice and (E) huLang LC^ΔMHC-II mice that have been immunized in vivo with 10 ug i.p. α-huLang-OVA_323–339 for three days. Each symbol represents data from an individual mouse. *** p<0.001.

Since MHC-II transfer has been shown to be bidirectional in a DC adoptive transfer model (Herrera et al. 2004) and since LC can transfer antigen to other DC for cross presentation in vitro (van den Berg. 2015), we examined whether antigen targeted LC could also transfer antigen to other DCs in the LN. To identify the subset of APCs acquiring antigen from targeted LC, we FACS-sorted APCs subsets from LN of α-huLang-2W1S immunized huLangerin mice, and co-cultured them in vitro with CFSE labeled OT-II T cells. Only migratory CD11b+ cDC2 and DN dDC induced potent OT-II T cell proliferation indicating that they can acquire antigen from LC (Figure 2D). A similar, but less robust result was obtained from these DC subsets in huLang LC^ΔMHC-II mice (Figure 2E). These results demonstrate that CD11b+ dDC and DN dDC can acquire antigen from LC in vivo.

In summary, these data have demonstrated that in vivo LC can acquire MHC-II from other cells and that CD11b+ dDC and DN dDC can acquire antigen from LC in the steady state. This finding reveals, to our knowledge, a previously unreported mechanism of DC cooperation in vivo and provides important insights on the design of DC vaccines aiming to target specific DC subtypes. These mechanisms of antigen and MHC-II transfer may also occur in the contexts of skin infection, vaccination and/or autoimmune disease though this remains to be elucidated.