Ocular Mucosal CD11b+ and CD103+ Mouse Dendritic Cells under Normal Conditions and in Allergic Immune Responses

Steady state dendritic cells (DC) found in non-lymphoid tissue sites under normal physiologic conditions play a pivotal role in triggering T cell responses upon immune provocation. CD11b+ and CD103+ DC have received considerable attention in this regard. However, still unknown is whether such CD11b+ and CD103+ DC even exist in the ocular mucosa, and if so, what functions they have in shaping immune responses. We herein identified in the ocular mucosa of normal wild-type (WT) and Flt3-/- mice the presence of a CD11b+ DC (i.e., CD11c+ MHCII+ CD11b+ CD103- F4/80+ Sirp-a+). CD103+ DC (i.e. CD11c+ MHCII+ CD11b low CD103+ CD8a+ DEC205+ Langerin+) were also present in WT, but not in Flt3-/- mice. These CD103+ DC expressed high levels of Id2 and Flt3 mRNA; whereas CD11b+ DC expressed high Irf4, Csfr, and Cx3cr1 mRNA. Additionally, the functions of these DC differed in response to allergic immune provocation. This was assessed utilizing a previously validated model, which includes transferring specific populations of exogenous DC into the ocular mucosa of ovalbumin (OVA)/alum-primed mice. Interestingly, in such mice, topical OVA instillation following engraftment of exogenous CD11b+ DC led to dominant allergic T cell responses and clinical signs of ocular allergy relative to those engrafted with CD103+ DC. Thus, although CD11b+ and CD103+ DC are both present in the normal ocular mucosa, the CD11b+ DC subset plays a dominant role in a mouse model of ocular allergy.


Introduction
As the most potent stimulators of T lymphocytes, dendritic cells (DC) are widely appreciated as a very unique subpopulation of antigen presenting cells and key in the generation, as well as the regulation, of adaptive immune responses. Much of their proficiency in T cell priming is attributed to high-level expression of MHC class II and costimulatory molecules such as CD80 and CD86. In addition, steady state DC, which take up residence in uninflamed non-lymphoid tissues such as the mucosa, are also highly efficient in mobilizing to lymphoid organs for Ag presentation to T cells. Now fully appreciated is the understanding that DC comprise multiple populations with differing ontogenies [1]. Two recently described subsets receiving considerable attention include, CD11b+ DC (i.e., CD11c+ MHCII+ CD11b+ CD103 low) and CD103+ DC (i.e., CD11c+ MHCII+ CD11b low CD103+) [1][2][3][4][5]. These cells have been described in mucosal tissues lining the lung and gastrointestinal tracts, in addition to tissues of other organs, e.g. liver and pancreas [1][2][3][4][5]. CD11b+ DC can be identified via their F4/80+ Sirp-a+, while CD103+ DC are CD8a+ DEC205+ langerin+ [2][3][4]. Transcriptional programs are also disparate, as CD11b+ DC express Irf4, Csfr, and Cx3cr1 [2][3][4], whereas CD103+ DC express Irf8, Id2, and Flt3 [2][3][4]. It is widely agreed that CD103+ DC arise from pre DC precursors in an Flt-3 dependent manner, whereas CD11b+ DC may be a heterogeneous population derived from multiple precursors, including pre DC and monocytes [2,3]. In addition, a third population has also been described in the gut lamina propria, which is CD103+ CD11b+ double positive [1].
Considerable attention has also been focused on attempting to elucidate the precise functional roles of CD103+ and CD11b+ DC; however, many questions still remain. Indeed, numerous reports have converged on the importance of viral Ag cross presentation by CD103+ DC [5][6][7][8]. In addition, their robust IL-12 production [9] has led to the tenet that CD103+ DC are preferential for the generation of Th1 responses. However, this has been challenged by reports demonstrating an importance for CD103+ DC in tolerance [10] and homeostasis [11]. Likewise, a recent report by Nakano et al demonstrated that pulmonary CD103+ DC are necessary to prime Th2 responses to inhaled allergens [12]-a finding which also challenges the notion that CD11b+ DC are preferential to stimulation of Th2 responses [13][14][15]. Thus, future work is required in order to further elucidate the nature and function of steady state CD103+ versus CD11b+ DC in adaptive immunity.
Many of the mucosal tissue sites (e.g., airway and gastrointestinal tracts) have already been probed for identification and function of CD11b+ and CD103+ DC, with an exception for the ocular mucosa (i.e., bulbar, fornix and palpebral conjunctiva) [1][2][3][4][5]. Investigating this would also be important in the ocular mucosa, which can indeed suffer from immunologic pathologies including those of autoimmune (e.g., conjunctival cicatrization), infectious (e.g., herpetic conjunctivitis) and/or allergic (e.g., atopic keratoconjunctivitis) origins. Steady state Langerhans cells and langerin+ DC have been identified in the normal mouse cornea [16], but have not been reported in the conjunctiva (conj). However, Ohbayashi et al have identified CD11c+ CD11b+ cells in the mouse ocular mucosa via immunohistochemistry [17,18]. Whether the lineage and phenotype of such CD11b+ DC are akin to steady state CD11b+ DC previously described [2][3][4][5], and if CD103+ DC also exist there, is completely unknown.
We therefore investigated this in the current study, and were able to determine that CD11b+ and CD103+ DC exist in the ocular mucosa of mice. Interestingly, use of a previously described model of ocular allergy, whereby exogenous DC (e.g., CD103+ versus CD11b+ DC) are engrafted into the ocular mucosa in topically challenged mice [19], revealed that CD11b+ DC were dominant in triggering allergic T cell responses and clinical signs of ocular allergy. Thus, we demonstrate herein that CD103+ and CD11b+ DC reside in the ocular mucosa; however, they appear to play different roles in mediating adaptive immune responses relevant in allergy.

Results
Characterization of steady state DC subsets of the ocular mucosa in wild type and Flt-3 KO mice Steady state CD11b+ and CD103+ DC were previously identified in numerous tissues of the mouse, including the liver and lung [1][2][3][4][5]. CD103+ DC arise from pre DC progenitors in a Flt-3 dependent manner; whereas, CD11b+ DC may partially arise from this pathway, as well as from monocyte precursors in a Flt-3 independent manner [1][2][3][4][5]. In the current study, we used flow cytometry to examine the ocular mucosa, i.e. conj, for the presence of distinct CD103+ and CD11b+ DC subsets, and compared these with known CD103+ and CD11b+ DC previously described in the liver and lung [1][2][3][4][5]. We also queried Flt-3 KO mice to this end, which have been previously shown to be absent of CD103+ DC in numerous organs such as the liver and lung [1][2][3][4][5]. In our study, we identified, similar to the liver and lung, the presence of a distinct population of CD103+ DC (CD11c+ I-A/I-E+ autofluorescent-CD11b low/neg CD103+) and another population of CD11b+ DC (CD11c+ I-A/I-E+ autofluorescent-CD11b+ CD103 low/neg) in the conj (Fig. 1a, b). The frequency of CD11b+ DC was greater than that of CD103+ DC for all tissues evaluated (Fig. 1b). Furthermore, as seen in the liver and lung, conj of Flt-3 KO mice were absent of CD103+, but not CD11b+ DC (Fig. 1b).

Examination of CD11b+ versus CD103+ DC in triggering allergic immune responses in a model of ocular allergy
With the conclusive identification of CD103+ and CD11b+ DC in normal conj, we next moved on to assess the role of these respective DC subsets in a model of ocular allergy. We utilized our previously described model to accomplish this [19], which involves adoptive transfer of T cells from OVA/alum-primed mice and topical OVA challenge in hosts which have received exogenously derived DC by their engraftment into the conj. Such site-specific engraftment of DC leads to a robust augmentation of allergic immune responses and provides direct functional information of engrafted DC [19].
In the current study we utilized bone marrow (BM) precursors to differentiate CD11b+ DC via standard GMCSF conditioning [20,21], whereas CD103+ DC were differentiated via Flt-3 conditioned media (Fig. 3a), as previously described by Sathe et al [22]. DC were also magnetically sorted. Naïve hosts were adoptively transferred with T cells from OVA/alum-primed mice and subsequently engrafted with exogenously BM-derived and purified CD11b+ versus CD103+ DC prior to topical OVA challenge. To assess the in vivo functional role in triggering allergic T cell responses by respective engrafted DC subsets, harvested host LN were OVA stimulated in vitro for intracellular flow cytometry analyses of CD4+ expression of IL-4, IL-5, IL-13, and IFN-g (Fig. 3b). Using this approach, we found that CD11b+ DC led to increased allergic T cell responses, as indicated by elevated CD4+ IL-4+, IL-5+, IL-13+, and IFN-g+ frequencies (Fig. 3b). ELISA assays of select cytokines from supernatants of parallel run cultures also confirmed this (Fig. 3c).
We also examined the role of CD11b+ vs. CD103+ DC in mediating clinical signs of ocular allergy, which include conjunctival chemosis, hyperemia, lid edema, and tearing/discharge [19]. To accomplish this, CD11b+ DC, CD103+ DC, or sham HBSS were engrafted into the conj of adoptively transferred hosts, as described above. Mice were challenged once daily for 10 d, and masked scoring was performed on each day at 20 min post challenge (i.e. consistent with the time frame for an immediate hypersensitivity response [19]), as well as 6 hr and 24 hr post challenge (i.e. consistent with the time frame for a late phase response [19]). Using this approach, we found increased clinical signs in response to engrafted CD11b+ DC, starting as early as challenge day 2 (Fig. 4). This was maintained through termination of the experiment on challenge day 7 (Fig. 4).
We next determined whether these results were relevant with responses generated by already differentiated CD11b+ and CD103+ nonlymphoid tissue DC. To accomplish this, we FACS sorted lung CD11b+ vs. CD103+ DC (Fig. 5a) and heterotopically engrafted these respective subsets into the conj of adoptively transferred hosts. Clinical scores in response to instillation of OVA challenges were then ascertained. Interestingly, we found increased clinical signs in mice engrafted with purified lung CD11b+ DC, relative to mice engrafted with purified lung CD103+ DC (Fig. 5b). This was observed as early as challenge day 2 (Fig. 5b) and maintained through termination of the experiment on challenge day 7 (Fig. 5b).

Discussion
Little is known about steady state DC subsets in the ocular mucosa, particularly regarding CD103+ and CD11b+ DC. One report by Ohbayashi et al identified via immunohistochemistry the presence of CD11b expressing DC in the normal mouse conjunctiva [17,18]. Consistent with this, we found a population of CD11b+ DC, but we also identified another population, that are CD103+ DC. Interestingly, however, we found that CD103+ and CD11b+ DC play different roles in mediating adaptive immune responses. By using a previously described model of ocular allergy in which exogenous DC are engrafted into the ocular mucosa in topically challenged mice [19], we found that the CD11b+ DC subset was dominant in triggering allergic immune responses.
We were able to identify the presence of these DC in the ocular mucosa via direct comparison with CD103+ and CD11b+ DC previously described in the liver and lung [2][3]. Furthermore, in line with previous reports, we found that steady state CD11b+ DC in the ocular mucosa are F4/80+ and Sirp-a+, but negative to low for CD8a, DEC205, and langerin [2][3][4]. Conversely, we demonstrated that steady state CD103+ DC in the ocular mucosa are negative to low for F4/80 and Sirp-a, but CD8a+, DEC205+, and langerin+. In line with Edelson et al [4], we demonstrated that steady state CD11b+ DC in the ocular mucosa express significantly higher levels of Irf4, Csfr, and Cx3cr1. In contrast, steady state CD103+ DC in the ocular mucosa express significantly higher levels of Id2 and Flt3, as well as a marginal increase in Irf8. Finally, CD11b+ DC, but not CD103+, were readily detectible in the ocular mucosa of Flt3 KO mice.
To assess the functional role of these DC subsets in ocular allergy, we utilized our model previously described by Schlereth et al [19]. It involves adoptive transfer of T cells and topical OVA challenge in hosts that have been engrafted into the ocular mucosa with exogenously derived DC. Such site-specific engrafted DC were shown to capture topically instilled allergen and mobilize to eye-draining LN for Th cell stimulation [19]. This in turn leads to a robust augmentation of secondary allergic immune responses relevant in ocular allergy, which is absent when CCR7-/-DC are engrafted instead [19]. Thus, such an approach provides direct functional information of engrafted DC [19] and to this end was appropriate to query the role of CD11b+ vs. CD103+ DC in ocular allergy. Our approach revealed a dominant role for CD11b+ DC in triggering allergic immune responses in ocular allergy. This is supported by our identification in the eye-draining LN of increased Th cell responses, as well as increased clinical signs of ocular allergy (i.e. conjunctival hyperemia and chemosis, lid edema, and tearing/discharge [19]).
It is important to highlight, however, that our identification of a steady state CD11b+ DC dominance in the elicitation of allergic immune responses was only demonstrated here in a model of secondary immune responses. This is because CD11b+ or CD103+ DC were engrafted into mice adoptively transferred with OVA sensitized T cells, and consequent allergic responses to topical OVA challenges is predominantly due to reactivation of sensitized T cells [19]. Others have examined steady state CD103+ and/or CD11b+ DC in secondary allergic immune responses and corroborate our findings [13,14]. Medoff et al adoptively transferred sensitized T cells into CD11b+ DC depleted mice (via use of CD11b-DTR mice), and found impaired allergic airway reactivity [13]. Likewise, Raymond et al reported a crucial role for CD11b+ DC in secondary allergic responses by testing allergic airway inflammation in sensitized mice [14]. In contrast, Nakano et al demonstrated a dominant role for pulmonary CD103+ DC in triggering Th2 responses to inhaled antigens [12]. However, their work was focused on primary immune responses. Thus, it may be surmised that CD103+ DC play a role in in allergic sensitization, whereas CD11b+ DC are key in secondary allergic reactions. Further work is needed to examine this.
Also requiring further investigation is the exact reasons by which CD11b+ DC are dominant in secondary allergic immune responses, as seen here and by others [13][14]. This is somewhat perplexing given the numerous reports converging on the inferiority of CD11b+ DC in mobilizing to draining LNs relative to their CD103+ counterparts [5,6,24]. In addition, endogenous CD11b+ DC normally reside deeper (i.e. in the stroma), and (as compared to intraepithelial CD103+ DC) would thus seem less advantageous with respect to allergen capture [23,24]. On the other hand, as shown here and by others, endogenous CD11b+ DC are greater in absolute number than CD103+ DC [2][3][4][5], which could perhaps compensate for these deficits. Such differences in frequency, however, does not seem to play a particularly important role given that in the current study an equal number of CD103+ and CD11b+ DC were engrafted. Data from our study likewise put into question the importance of anatomic location given that both DC subsets were engrafted into the same space.
Also not offering a clear explanation as to the reasons for the dominant role for CD11b+ DC seen here, are other superior attributes previously described. CD11b+ DC are known to be proficient in phagocytosis, which could be advantageous with respect to allergen uptake [15]. CD11b+ DC are also poor at production of the Th1 polarizing cytokine, IL-12 [9]. However, such collective properties do not appear to be absolute as Nakano et al demonstrated that CD11b+ DC are inferior to their CD103+ counterparts in Th2 priming [12]. Thus future work is required to understand why CD11b+ DC are dominant in triggering secondary allergic immune responses in our model and others [13,14].
In conclusion, we show here that steady state CD11b+ and CD103+ DC are present in the ocular mucosa; however, these DC have disparate roles in mediating adaptive immune responses. In a model of ocular allergy, wherein CD11b+ vs. CD103+ DC were engrafted into the ocular mucosa of mice adoptively transferred with sensitized T cells and challenged topically with allergen, revealed that CD11b+ DC are dominant in triggering secondary allergic immune responses relevant in ocular allergy. Future work is required to understand why CD11b+ DC are dominant in secondary allergic immune responses and whether such cells can be targeted therapeutically in treatment of atopic conditions such as allergic eye disease.

Mice and Anesthesia
Female C57BL/6 and Flt-3 knockout mice 8-12 wk old were purchased from Charles River Laboratories (Wilmington, MA) and Jackson laboratories (Bar harbor, ME). Mice were housed in a specific pathogen-free environment at the Schepens Eye Research Institute animal facility. The Institutional Animal Care and Use Committee approved all procedures. All animals were treated according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Anesthesia was used for all surgical procedures with intraperitoneal administered ketamine/ xylazine suspensions (120 and 20 mg/kg, respectively).

Enzymatic Digestion of Conjunctiva, Liver and Lung
This procedure has been previously described [16,19]. Briefly, euthanized mice underwent transcardial perfusion with cold HBSS prior to collecting the organs assessed. For conjunctiva, all sites including bulbar through palpebral regions from both the superior and inferior areas were collected. Tissues were minced and digested in 2 mg/mL collagenase D (Roche, Indianapolis, IN) and 0.05 mg/mL DNase (Roche) for 2 to 3 hours at 37uC. Cell suspensions were then triturated in 20 mM EDTA and then passed through a 70-um filter (BD Falcon; Becton-Dickinson, Franklin Lakes, NJ). Cells were then thoroughly washed.

MACS Sorting of BM-Derived DC Subsets used for Engraftment
For BM-derived DC, cells were harvested on day 7 of cultures and prepared for magnetic sorting according to manufacturer's instructions (Miltenyi Biotec, Auburn, USA). For CD11b+ DC, GM-CSF cultivated cells were MACS sorted using anti-CD11c beads alone, as all CD11c+ cells co-express CD11b+. For CD103+ DC, Flt-3 cultivated cells were sorted using anti-CD103 beads, as all CD103+ cells co-express CD11c+; ,90% purity was achieved for both subsets. Purified cells were washed thoroughly and prepared for subconjunctival injection.

FACS Sorting of Lung DC Subsets used for Engraftment
Following enzymatic digestion, DC were first purified magnetically via anti-CD11c beads, according to manufacture's instructions (Miltenyi Biotec). Cells were then thoroughly washed and enumerated. Cells were then stained for FACS sorting as detailed above.

T Cell Adoptive Transfer
This method in allergic eye disease has been previously described [19]. Briefly, T cells were obtained from donor wildtype (WT) C57BL/6 mice that were immunized once intraperitoneal with a 100 ul suspension containing 1 mg aluminum hydroxide (Sigma Aldrich, St. Louis, USA) diluted in HBSS, 300 ng pertussis toxin (Sigma Aldrich), and 100 ug ovalbumin (Sigma Aldrich). Donor mice 2 wk post immunization were euthanized and spleens were collected. Donor spleens were prepared into single-cell suspensions by tissue press using a sterile 70 um sieve, and cells were then treated with red blood cell lysis buffer according to manufacturer's instructions (Sigma Aldrich) and washed thoroughly. Donor T cells were enriched via MACS sorting using anti-CD90.2 Ab according to manufacturer's instructions (Miltenyi Biotec, Auburn, USA). The sorted donor population was then enumerated via trypan blue exclusion assay, and donor T cells were set at a concentration 16107/ml of sterile HBSS. Recipient mice were adoptively transferred IV with 46106 donor T cells.

Conjunctival Engraftment and Ocular Allergy
This model was previously described [19]. Briefly OVA primed T cells were prepared as described above, and 46106 T cells were adoptively transferred into naïve hosts. Host mice were then anesthetized 16 hr later for unilateral injection of cells into the ocular mucosa by way of subconjunctival injection. Injection volume was 10 ul of sterile HBSS and contained 16105 purified BM derived CD11b+ or CD103+ DC. In another experiment, mice were similarly injected with 66104 purified CD11b+ or CD103+ DC from enzymatically digested lungs from donor mice. Challenge via topical OVA instillation (250 ug/5 ul eye drop) was administered immediately following subconjunctival (SCJ) injection, and then challenged additionally (to account for significant tearing post SCJ injection) twice more in 20 minute intervals. Challenges were subsequently administered once daily, for 7 days.

Clinical Scoring
This procedure has been described previously [19] and performed here in a masked fashion by two independent observers. Briefly, scoring was performed 20 min post challenge and done once daily from day 1 (i.e. 24 hr following SCJ injection) to day 7. Mice where examined biomicroscopically based on four independent parameters, which include: 1) lid swelling; 2) tearing; 3) chemosis; and 4) conjunctival vasodilation (redness). Each parameter was ascribed 0 (i.e. absent) to 3+ points (i.e. maximal) and were summed to yield a maximum score of 12+.

T cell Cytokine Analyses
Regional LN (cervical and submandibular) were collected and pooled from freshly euthanized mice. Single-cell suspensions were prepared and enumerated via trypan blue exclusion assay. Cells were plated in round-bottom 96-wells at a concentration of 2.06106/ml in triplicate wells of RPMI (10% FBS) with OVA (1 mg/ml) for up to 72 hr and restimulated with PMA/Ionomycin with Golgiplug (BD Pharmingen) for up to 4 hours. Cells were harvested and stained for extracellular CD4, and intracellular IL-4, IL-5, IL-13, and IFN-g. Parallel cultures were established for ELISA analyses of culture supernatant, and thus restimulated with PMA/Ionomycin only. ELISAs (eBioscience) for IFN-g and IL-13 were analyzed.

Statistical analysis
Statistical analyses included 1-way ANOVA and Bonferroni's Multiple Comparison Test, in addition to two-tailed student's ttest. Standard error and standard deviation of the mean were calculated. A p-value ,0.05 was considered statistically significant.