If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Britton Chance Center for Biomedical Photonics, Wuhan National Laboratory for Optoelectronics-Huazhong University of Science and Technology, Wuhan, Hubei, ChinaKey Laboratory of Biomedical Photonics of Ministry of Education, Collaborative Innovation Center for Biomedical Engineering, School of Engineering Sciences, Huazhong University of Science and Technology, Wuhan, Hubei, China
Britton Chance Center for Biomedical Photonics, Wuhan National Laboratory for Optoelectronics-Huazhong University of Science and Technology, Wuhan, Hubei, ChinaKey Laboratory of Biomedical Photonics of Ministry of Education, Collaborative Innovation Center for Biomedical Engineering, School of Engineering Sciences, Huazhong University of Science and Technology, Wuhan, Hubei, China
Britton Chance Center for Biomedical Photonics, Wuhan National Laboratory for Optoelectronics-Huazhong University of Science and Technology, Wuhan, Hubei, ChinaKey Laboratory of Biomedical Photonics of Ministry of Education, Collaborative Innovation Center for Biomedical Engineering, School of Engineering Sciences, Huazhong University of Science and Technology, Wuhan, Hubei, China
Britton Chance Center for Biomedical Photonics, Wuhan National Laboratory for Optoelectronics-Huazhong University of Science and Technology, Wuhan, Hubei, ChinaKey Laboratory of Biomedical Photonics of Ministry of Education, Collaborative Innovation Center for Biomedical Engineering, School of Engineering Sciences, Huazhong University of Science and Technology, Wuhan, Hubei, China
Britton Chance Center for Biomedical Photonics, Wuhan National Laboratory for Optoelectronics-Huazhong University of Science and Technology, Wuhan, Hubei, ChinaKey Laboratory of Biomedical Photonics of Ministry of Education, Collaborative Innovation Center for Biomedical Engineering, School of Engineering Sciences, Huazhong University of Science and Technology, Wuhan, Hubei, China
Britton Chance Center for Biomedical Photonics, Wuhan National Laboratory for Optoelectronics-Huazhong University of Science and Technology, Wuhan, Hubei, ChinaKey Laboratory of Biomedical Photonics of Ministry of Education, Collaborative Innovation Center for Biomedical Engineering, School of Engineering Sciences, Huazhong University of Science and Technology, Wuhan, Hubei, China
Qingming Luo, Wuhan National Laboratory for Optoelectronics-Huazhong University of Science and Technology, Wuhan, Hubei 430074, People’s Republic of China.
Britton Chance Center for Biomedical Photonics, Wuhan National Laboratory for Optoelectronics-Huazhong University of Science and Technology, Wuhan, Hubei, ChinaKey Laboratory of Biomedical Photonics of Ministry of Education, Collaborative Innovation Center for Biomedical Engineering, School of Engineering Sciences, Huazhong University of Science and Technology, Wuhan, Hubei, China
Correspondence: Zhihong Zhang, Wuhan National Laboratory for Optoelectronics-Huazhong University of Science and Technology, Wuhan, Hubei 430074, People’s Republic of China.
Britton Chance Center for Biomedical Photonics, Wuhan National Laboratory for Optoelectronics-Huazhong University of Science and Technology, Wuhan, Hubei, ChinaKey Laboratory of Biomedical Photonics of Ministry of Education, Collaborative Innovation Center for Biomedical Engineering, School of Engineering Sciences, Huazhong University of Science and Technology, Wuhan, Hubei, China
It remains unclear how monocytes are mobilized to amplify inflammatory reactions in T cell-mediated adaptive immunity. Here, we investigate dynamic cellular events in the cascade of inflammatory responses through intravital imaging of a multicolor-labeled murine contact hypersensitivity model. We found that monocytes formed clusters around hair follicles in the contact hypersensitivity model. In this process, effector T cells encountered dendritic cells under regions of monocyte clusters and secreted IFN-γ, which mobilizes CCR2-dependent monocyte interstitial migration and CXCR2-dependent monocyte cluster formation. We showed that hair follicles shaped the inflammatory microenvironment for communication among the monocytes, keratinocytes, and effector T cells. After disrupting the T cell-mobilized monocyte clusters through CXCR2 antagonization, monocyte activation and keratinocyte apoptosis were significantly inhibited. Our study provides a new perspective on effector T cell-regulated monocyte behavior, which amplifies the inflammatory reaction in acquired cutaneous immunity.
Skin is the primary interface that protects the human body from foreign antigen challenges. Monocytes and macrophages (Mons/Mφs) play an important role in mediating efficient acquired cutaneous immunity, especially in defense against pathogen invasion (
). Once monocytes have infiltrated the extravascular dermis, they sense and identify danger signals, which guide their navigation toward inflammatory foci to mediate effector functions (
). CCR8 has been reported to mediate inflammatory monocyte migration into lymph nodes, and CCR6 has been reported to contribute to monocyte-derived dendritic cell (DC) migration toward the epithelium (
). Although cytokines can directly stimulate monocyte activation in vitro, in vivo activation of Mons/Mφs is much more complex because the effects of cytokines are exerted locally at the site of their production in an autocrine or paracrine manner (
). It is essential to dynamically interpret how monocytes are activated in situ. Moreover, keratinocyte apoptosis is integral to the pathogenesis of skin disorders (
). Investigating the relationship between monocyte activation and keratinocyte death will help explain the dynamic function of monocytes in acquired cutaneous immunity.
Recently, intravital imaging of immune cell motility and dynamic response have provided excellent insights into the mechanisms of cell behavior and function in innate and adaptive immune responses (
). Therefore, investigating the highly orchestrated behavior of monocytes will provide an avenue to uncovering how monocytes efficiently amplify the adaptive immune response.
Here, we intravitally imaged the murine contact hypersensitivity (CHS) model to dynamically investigate the spatiotemporal relationship between monocytes, effector T cells, and keratinocytes in vivo. We found that effector T cells mobilize monocytes into forming clusters around hair follicles and showed that T cell-regulated cluster formation of monocytes was an essential step for eliciting an efficient immune response in CHS.
Results
CX3CR1int monocytes accumulate and form clusters in the CHS model
To investigate the dynamic behavior of inflammatory monocytes in CHS, we used oxazolone (OXA) to induce a CHS response in CX3CR1-GFP transgenic mice (
). After challenging the immunized mice with 1% OXA (OXA challenge), a large number of GFP-positive cells were recruited into the dermis during 12–24 hours in CHS, mainly CX3CR1intCD11b+ cells (Figure 1a). In addition, the data in Supplementary Figure S1a online show that these infiltrated GFP cells were Ly6ChighCD115+F4/80+CCR2+, namely, inflammatory monocytes. The percentage of Ly6ChighCX3CR1int cells among all CD45+ cells in the dermis increased from 1.2% before OXA challenge to 20.3% and 23.3% at 12 and 24 hours, respectively, after OXA challenge (Figure 1c, red box). Conversely, the percentage of CX3CR1high cells among all CD45+ cells in the dermis decreased to 1.9% at 12 hours and 0.7% at 24 hours after OXA challenge (Figure 1c). Therefore, GFP-positive cells in the dermis mainly represent inflammatory monocytes detected at 12–24 hours after hapten challenge in CX3CR1-GFP mice.
Figure 1CX3CR1int monocyte accumulation and cluster formation around hair follicles in CHS mice. (a) Flow cytometric analysis of dermal cells from CX3CR1gfp/+ mice at steady state and 12 and 24 hours after OXA challenge. Representative flow cytometry plots are gated for live dermal CD45+ cells. (b) Flow cytometric analysis of CX3CR1+CD11b+ monocytes in blood. (c) Quantification of CX3CR1+ cells accumulating over time in CHS mice (n = 4 mice for each point). (d) Time-lapse imaging of interstitial migration and aggregation of CX3CR1int monocytes (green) in the dermis from 14–22 hours after OXA challenge. Scale bar = 100 μm. (e) Mean fluorescence intensity of the monocytes in the cluster. (f) Confocal imaging of ear sections at 0, 24, 48, and 72 hours after OXA challenge. Scale bars = 100 μm. (g) Three-dimensional constructs of CX3CR1int monocyte cluster around HFs 24 hours after OXA challenge. GFP (green/monocytes), tdTomato (red/keratinocyte), SHG (blue/collagen). Scale bar = 70 μm. (h) The schematic diagram of the DEJ in skin. The data are representative of three independent experiments (a–c, mean ± standard error of the mean in c). CHS, contact hypersensitivity; DEJ, dermal-epidermal junction; h, hour; MFI, mean fluorescence intensity; OXA, oxazolone; SHG, second harmonic generation.
We then explored the distribution of CX3CR1int-monocytes based on the GFP signal in the dermis. We chose the central site of a mouse dorsal ear as the imaging site and monitored the dynamic process of monocyte cluster formation. Time-lapse confocal imaging data showed that CX3CR1int monocytes rapidly migrated toward hair follicles (HFs), formed a cluster beginning at 15 hours after the OXA re-challenge, and reached a maximum at 21 hours (Figure 1d and e, and see Supplementary Movie S1 online). In addition, chimera mice without OXA challenge served as a control, for which monocyte accumulation did not occur around hair follicles even after 120 minutes of time-lapse imaging 1 minute apart. This result further excluded the influence of hair shafts on monocyte accumulation (see Supplementary Movie S2 online). Because Langerhans cells and dendritic epidermal T cells anchored in the epidermis also express high levels of GFP in CX3CR1gfp/+ mice, we then reconstructed the bone marrow in x-ray–irradiated (9 Gy) albino-C57 mice through adoptive transfer of CX3CR1gfp/+ bone marrow and confirmed that these aggregated GFP+ cells were derived from CX3CR1gfp/+ bone marrow but not from GFP+ cells in the epidermis (see Supplementary Figure S2 online). These results showed that bone marrow-derived CX3CR1int monocytes formed a cluster in CHS. The results from ear sections at 0 and 24 hours after OXA- challenge also showed large numbers of CX3CR1int monocytes accumulating in the dermis and forming clusters under the epidermis 24 hours after OXA challenge (Figure 1f).
After identifying CX3CR1int monocyte accumulation and cluster formation, we then combined the imaging of monocytes with skin structure (keratinocytes, dermal vasculature, and collagen) in vivo by constructing CX3CR1-mTmG chimera mice, the skin keratinocytes of which express RFP and the monocytes of which express GFP (Figure 1g and h, and see Supplementary Figure S3 online). Collagen was imaged based on the second harmonic generation signal, which distinguished the epidermis from the dermis (exhibiting the second harmonic generation signal). The results confirmed that the CX3CR1+ monocyte cluster presented a sandwich structure. In these clusters, some CX3CR1-GFP+ cells had migrated across the basal keratinocytes, and some CX3CR1-GFP+ cells were distributed under the epidermis and aggregated around HF keratinocytes at the dermal-epidermal junction. Thus, the imaging shows the amount of CX3CR1 monocytes tightly connected with the basal keratinocytes around the HFs 24 hours after the OXA challenge (see Supplementary Figure S3 and Supplementary Movie S3 online). The results also confirmed that the CD68+ cells formed clusters around the HF region (see Supplementary Figure S4a online). FACS data confirmed that the CD11b+CX3CR1-GFPint monocytes were CD68 positive in the dermis (see Supplementary Figure S4b).
CX3CR1int monocytes mediate keratinocyte death
In addition, through labeling basal keratinocytes (isolectin B4) and dead cells (propidium iodide) in vivo, we found that keratinocytes died around monocyte clusters. To investigate whether these CX3CR1int monocytes mediated keratinocyte death, clodronate liposomes (Clo-liposomes) were injected intravenously to deplete circulating monocytes. Flow cytometry confirmed the elimination of circulating CX3CR1+ monocytes in blood (see Supplementary Figure S5a online) and interstitial CX3CR1int monocytes in dermis (see Supplementary Figure S5b). Ear swelling decreased from 220 μm to 68 μm after monocyte depletion (Figure 2b). Through intravital imaging of propidium iodide-positive cells in the epidermis, we observed that mice that received the Clo-liposome injection showed less cell death (174 ± 29 cells/mm2) than CHS mice that received the phosphate buffered saline liposome injection (2,260 ± 436 cells/mm2) (Figure 2c). TUNEL staining of ear sections confirmed that the manner of death of keratinocytes in CHS was apoptosis (Figure 2e). Keratinocyte apoptosis was inhibited when CX3CR1int monocytes were eliminated in the Clo-liposome–treated group (Figure 2f). Thus, our results provide details about the spatial structure of monocyte-keratinocyte clusters and show that CX3CR1int monocytes mediated keratinocyte apoptosis in the CHS model.
Figure 2The accumulation of CX3CR1int monocytes mediating keratinocyte death. (a) Representative three-dimensional images of CX3CR1int monocyte clusters and keratinocytes 24 hours after OXA challenge. GFP (green/monocytes), isolectin (blue/keratinocyte), dead cells (red/PI), SHG (yellow/collagen). Scale bar = 50 μm. (b) Intravenous injection of Clo-liposome significantly inhibits ear swelling in the CHS model. (c) Quantification of dead cells (n = 4 mice per group). ∗∗P < 0.01, unpaired Student t test. (d–f) TUNEL staining of mice ear sections in different groups. (d) Untreated group, (e) PBS liposome-treated group, and (f) Clo-liposome treated group. The data are pooled from three experiments (d, mean ± standard error of the mean). Scale bars = 100 μm. CHS, contact hypersensitivity; Clo-liposome, clodronate liposome; h, hour; OXA, oxazolone; PBS, phosphate buffered saline; PI, propidium iodide; SHG, second harmonic generation.
CCR2 and CXCR2 regulate cluster formation in CX3CR1int monocytes
Having shown that CX3CR1int monocytes mediate keratinocyte apoptosis in CHS, we then investigated how monocytes migrate, arrest, and form clusters around HFs. The spatial distance between the epidermis and the dermal vasculature suggests that monocytes require at least two steps to form a stable cluster around HFs: first, they migrate toward the epidermis; then, they are arrested around the region.
To identify the chemokine receptor that mediates the interstitial migration of CX3CR1int monocytes, antagonists RS504393, maraviroc, and SB225002 were intravenously injected into mice 12 hours after OXA challenge to block the receptors CCR2, CCR5, and CXCR2, respectively. Pertussis toxin, a commonly used reagent that blocks the receptor-dependent chemokine signaling pathway (
Specific uncoupling by islet-activating protein, pertussis toxin, of negative signal transduction via alpha-adrenergic, cholinergic, and opiate receptors in neuroblastoma x glioma hybrid cells.
), was used as a positive control. Compared with phosphate buffered saline-treated mice, the motility of the monocytes in pertussis toxin-treated mice was significantly reduced; the mean displacement declined from 8.8 μm to 2.9 μm at 12 hours after OXA challenge (Figure 3a and b), and the mean speed decreased from 3.24 μm/minute to 1.75 μm/minute (Figure 3c). Among all mice treated with chemokine receptor antagonists, only blockade of the chemokine receptor CCR2 (with the antagonist RS504393) resulted in motility of monocytes that was similar to that of the pertussis toxin-treated group (Figure 3b and c). Taken together, these results show that CCR2 regulated the interstitial migration of CX3CR1int monocytes toward the epidermis. In Figure 3a–c, the data also show that CCR5 partly modulated monocyte arrest in the interstitial space. In allergic contact dermatitis, RANTES was induced by TNF-α in the early stages after hapten challenge (
). Thus, RANTES released by inflamed keratinocytes might partly modulate monocyte arrest in the interstitial space.
Figure 3CCR2 regulates CX3CR1int monocyte interstitial migration, and CXCR2 regulates CX3CR1int monocyte cluster formation. (a–c) Motility parameters of CX3CR1int monocytes in the dermis 12 hours after OXA challenge. (a) MD calculated from tracks over 9 minutes. (b, c) MD and MS of CX3CR1int monocytes in the CHS model individually treated with pertussis toxin, a CCR2 antagonist, a CXCR2 antagonist, a CCR5 antagonist, or PBS. ∗∗∗∗P < 0.0001 (one-way analysis of variance with Dunnett test). (d) Confocal imaging of CX3CR1int monocytes in fixed inflammatory dermal sheets. GFP (green/monocytes), DAPI (blue). Scale bars = 100 μm. (e) Quantification of the number of clusters in fixed dermal sheets. Samples were analyzed in four mice per group, and four to five regions were captured for each sample (n > 16 regions for each group). (f) Quantification of GFP cell numbers in the interstitial space in fixed dermal sheets. The data are (d) representative of three independent experiments or (a–c, e) pooled from three experiments (mean ± standard error of the mean). CHS, contact hypersensitivity; MD, mean displacement; min, minute; MS, mean speed; ns, not significant; OXA, oxazolone; PBS, phosphate buffered saline; PTX, pertussis toxin.
We then investigated the cause of the monocyte arrest to form clusters around HFs. According to a previous report that CXCR2 plays an important role in monocyte arrest on the endothelium in atherosclerotic plaques (
), we speculated that CXCR2 might be the key molecule initiating monocyte cluster formation. The antagonist SB225002 was then intraperitoneally injected to block the function of CXCR2. Dermal sheets were harvested to observe the distribution of CX3CR1int monocytes in the dermis 24 hours after OXA challenge. The imaging data clearly showed that CXCR2 antagonization attenuated the number of CX3CR1int monocytes in the dermis (Figure 3d). By calculating the number of monocyte clusters per field of view, we showed that the CXCR2 antagonist blocked the aggregation of CX3CR1int monocytes in a dose-dependent manner and that 10 mg/kg SB225002 treatment dramatically decreased the number of monocyte clusters from 21/mm2 to 6/mm2 (Figure 3e). By calculating the number of monocytes per field of view (see Supplementary Figure S6 online), we found that SB225002 treatment did not significantly decrease the monocyte number in the interstitial space (Figure 3f). Although infiltrated neutrophils expressed high levels of CXCR2, we confirmed that monocyte cluster formation was neutrophil independent (see Supplementary Figure S7 online). Therefore, these data show that CXCR2 regulated CX3CR1int monocyte cluster formation around HF regions and that this regulation may directly act on monocytes.
Cluster formation is a key node event for triggering monocyte activation and keratinocyte apoptosis
To verify whether cluster formation was essential for monocyte activation, we used flow cytometry to analyze the TNF-α expression levels in CX3CR1int monocytes when clustering occurred or was inhibited by the CXCR2 antagonist. The CD11b+Ly6G–Ly6Chi cell subset was used to identify CX3CR1int monocytes in OXA-challenged ears (Figure 4a), and the percentage of TNF-α–positive cells among CX3CR1int monocytes decreased from 16.7% to 5.7% when the mice were pretreated with the CXCR2 antagonist (Figure 4b). These data showed that CXCR2-regulated cluster formation of CX3CR1int monocyte was essential for monocyte activation to release TNF-α.
Figure 4Monocyte cluster formation is necessary for TNF-α expression of monocytes and keratinocyte apoptosis. (a) Flow cytometric analysis of CX3CR1int monocytes expressing TNF-α 24 hours after OXA challenge. (b) The percentage of TNF-α–positive cells decreased in SB225002-pretreated CHS mice (n = 6 mice per group). (c) Detection of cell apoptosis in the epidermis by the TUNEL assay. The sensitized mice were OXA challenged or pretreated with Clo-liposomes, pretreated with antagonist of CXCR2, or untreated. TUNEL (red) and DAPI (blue) staining. Scale bars = 100 μm. (d) Quantification of apoptotic cells in c (n = 4 mice per group). ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.01. NS indicates not significant (P > 0.05), ND indicates not detected; unpaired Student t test. The data are representative of three independent experiments (a–c, mean ± standard error of the mean in b) or pooled from three experiments (d, mean ± standard error of the mean). CHS, contact hypersensitivity; Clo-liposome, clodronate liposome; ND, not detected; NS, not significant; OXA, oxazolone; Wt, wild type.
To verify whether clustering was essential for mediating keratinocyte death, the ear epidermis, which is mainly composed of keratinocytes, was separated and stained with a TUNEL kit to detect apoptosis (Figure 4c). As a control, the mice were injected intravenously with Clo-liposomes to deplete circulating monocytes 24 hours before OXA- challenge or pretreated with an antagonist of CXCR2 to inhibit the formation of CX3CR1int monocyte clusters. Confocal imaging of TUNEL-stained epidermal sheets showed numerous apoptotic cells (729 ± 162/mm2) in CHS mice 24 hours after OXA challenge (Figure 4d). Conversely, few apoptotic cells were observed in Clo-liposome–pretreated mice (41 ± 29/mm2) or in mice treated with an antagonist of CXCR2 (99 ± 36/mm2); the number of apoptotic cells was not a significant difference in these two groups (Figure 4d and e). These results showed CX3CR1int monocyte clustering as the critical factor mediating monocyte activation and keratinocyte death.
Antigen recognition induces CX3CR1int monocyte arrest and cluster formation
Having monitored the cluster formation of CX3CR1int monocytes in the dermis, we then wondered what triggers this migration of monocytes in CHS. Here, two groups of mice were sensitized with OXA (sensitized group) or a vehicle solution (vehicle control group), and the CX3CR1int monocyte behavior was investigated (see Supplementary Movie S4 online). Imaging data confirmed that 12 hours after the OXA challenge, CX3CR1int monocytes were evenly distributed throughout the dermis of the OXA-sensitized or vehicle control mice (Figure 5a). However, CX3CR1int monocytes formed clusters only 24 hours after OXA challenge in OXA-sensitized mice (Figure 5a). In addition, monocytes in the sensitized group displayed a confined migration with a higher arrest coefficient and lower speed compared with the vehicle control group (Figure 5b and c), which indicated that some mediators were released after antigen recognition-modulated interstitial monocyte arrest around HFs.
Figure 5Antigen recognition contributes to CX3CR1int monocyte arrest and cluster formation in the dermis. (a) Representative confocal images of CX3CR1int monocytes in the dermis of mice sensitized with (left) OXA or (right) vehicle control. n = 3 mice per group. Scale bars = 100 μm. (b–c) Motility parameters of CX3CR1int monocytes 12 hours or 24 hours after OXA challenge. (b) MS (μm/minute), (c) arrest coefficient. (d) Representative confocal images of monocytes, DCs, and OXA-sensitized T cells at different depths between 40 μm and 80 μm 12 and 24 hours after OXA challenge in CHS. The white arrows represent the DC-T clusters around the HFs. Scale bars = 100 μm. (e) Representative confocal images of monocytes, DCs, and DNFB-sensitized T cells at different depths between 40 μm and 80 μm 12 and 24 hours after OXA challenge in CHS. The white arrows represent the DNFB-sensitized T cells; GFP (green/monocytes), YFP (yellow/DCs), CMTMR (red/T cells). n = 4 mice. Scale bars = 100 μm. (f) Quantification of the recruitment of DCs and effector T cells under monocyte clusters in CHS mice. (g–i) Motility parameters of T cells in the dermis 12 and 24 hours after OXA challenge: (g) MS (μm/minute), (h) arrest coefficient, and (i) confinement ratio. ∗∗∗∗P < 0.0001; ns indicates not significant (P > 0.05). One-way analysis of variance with Dunnett test in b and c, unpaired Student t test in g–i; each dot represents one cell at the given time point. The data in b and c and g–i are pooled from three experiments (mean ± standard error of the mean). CHS, contact hypersensitivity; d, day; DC, dendritic cell; h hour; HF, hair follicle; MS, mean speed; ns, not significant; OXA, oxazolone.
), we then explored the spatiotemporal relationship between DC–T-cell interactions and CX3CR1int monocyte clustering in CHS. T cells from OXA-sensitized mice labeled with 5-(and-6)-(((4-Chloromethyl)Benzoyl)Amino) tetramethylrhodamine (CMTMR) were transferred into CD11cEYFP-CX3CR1GFP mice, in which DCs expressed EYFP and monocytes expressed GFP. As shown by intravital imaging (Figure 5d, and see Supplementary Movies S5 and S6 online), OXA-sensitized effector T cells encountered dermal DCs 12 hours after OXA challenge and formed DC–T-cell clusters (n = 5.9 ± 0.5 DCs per cluster, n = 4.9 ± 0.8 T cells per cluster) around HFs at a depth of 60–80 μm. At 24 hours after the OXA challenge, monocytes formed clusters (depth of 40–60 μm) on top of the DC–T-cell clusters (see Supplementary Figure S8 online). As a control, few DNFB-sensitized T cells infiltrated after the OXA challenge compared with OXA-sensitized T cells. In addition, fewer monocyte clusters were observed in the DNFB T-cell adoptive transfer group (Figure 5e). In addition, quantitative analysis of the cell motility indicated that T cells displayed a resting state to recognize the antigen 12 hours after OXA challenge and displayed an active state 24 hours after OXA challenge (Figure 5g–i, and see Supplementary Movie S7 online). These data show that CX3CR1int monocytes formed clusters after T-cell recognition of the antigen. Although the percentage of transferred cells was lower than that of CX3CR1 monocytes, these activated T cells in clusters have been reported to be critical for IFN-γ release by T cells (
). In addition, we found that 41 ± 14 endogenous T cells accumulated in one monocyte cluster (see Supplementary Figure S4c–e). Taken together, these data provide spatiotemporal evidence supporting the hypothesis that effector T cells are involved in the modulation of monocyte clusters around HFs in skin.
IFN-γ secreted by effector lymphocytes modulates the formation of monocyte clusters
Because dermal DC–T-cell clusters have been reported to activate T cells to express IFN-γ, our results suggest that the IFN-γ secreted by T cells after antigen recognition in the early stage (12 hours after the OXA challenge) might have induced monocyte cluster formation in the late stage (24 hours after the OXA challenge). Four experimental groups were then designed to explore the effect of lymphocyte-mediated secretion of IFN-γ on the CX3CR1int monocyte cluster: (i) a normal CHS model, (ii) an adoptive CHS model, (iii) an IFN-γ–deficient adoptive CHS model, and (iv) an irritation model (Figure 6a). Through calculating the number of monocytes in the normal CHS model, we observed that 130 ± 11 GFP+ monocytes formed one cluster. Through calculating the number of monocytes in the adoptive CHS model, we observed that the number of monocytes in each cluster was significantly reduced in the IFN-γ–deficient adoptive CHS model (38 ± 3 cells/one cluster) compared with the number in the adoptive CHS model (89 ± 5 cells/one cluster) (Figure 6b and c). In the irritational model, only 20 ± 2 monocytes per cluster were calculated. These results clearly indicated that the IFN-γ secreted by hapten-specific lymphocytes participated in modulating monocyte cluster formation.
Figure 6Evaluation of hapten-specific lymphocytes secrete IFN-γ to modulate monocyte cluster formation. (a) Schematic of four experimental groups. (b) Representative confocal images of monocyte clusters around HFs 24 hours after OXA challenge (n = 4 mice per group). Green represents dispersed CX3CR1-GFP monocytes, and yellow represents aggregated cells. Scale bars = 100 μm (c) Quantification of the number of cell clusters around HFs (n = 4 mice per group). The data in each group were derived from at least 12 imaged regions represented in b. ∗∗∗∗P < 0.0001 (unpaired Student t test). (d–j) IFN-γ up-regulates CXCL1, CXCL2, MIF, and CCL2 expression in inflammatory skin in the CHS mouse model. (d) Quantification of CXCL1, (e) CXCL2, (f) MIF, and (g) CCL2 expression in ear homogenates from four groups: normal ear (untreated), irritation model (1% OXA), wild-type CHS model, and IFN-γ–deficient CHS model. (h) Distinction of HF keratinocytes (HF, Sca-1–) from interfollicular epidermal keratinocytes (IE, Sca-1+) among tail skin keratinocytes using flow cytometry. (i) Sorting of HF cells and IE cells based on Sca-1 as a marker. (j) Quantification of CCL2 expression in keratinocytes before and 1 or 8 hours after IFN-γ stimulation (n = 3 measurements per group). ∗P < 0.05; ∗∗ P < 0.01; ∗∗∗P < 0.001; ns, not significant (P > 0.05) (unpaired Student t test). The data are representative of three independent experiments (b, d–j, mean ± standard error of the mean) or pooled from three experiments (c, mean ± standard error of the mean). CHS, contact hypersensitivity; DC, dendritic cell; h, hour; HF, hair follicle; IE, interfollicular epidermis; LN, lymph node; MIF, macrophage migration inhibitory factor; OXA, oxazolone; WT, wild type.
In addition, four experimental groups were designed to investigate how IFN-γ controls the secretion of CCR2 and CXCR2 ligands in CHS: (i) normal ear (untreated), (ii) irritation model (1% OXA), (iii) wild-type CHS model, and (iv) IFN-γ–deficient CHS model. As shown in Figure 6d–g, CXCL1, CXCL2, MIF, and CCL2 expression increased significantly after antigen challenge and recognition in CHS. In addition, the increase of these chemokines was significantly inhibited in IFN-γ–deficient CHS group. Because monocyte cluster formation was mainly located around HFs in the mouse CHS model, we then wondered whether the chemokine expression in keratinocytes differed in distinct regions after IFN-γ stimulation. We isolated mouse epidermal keratinocytes and sorted the HF keratinocytes according to their Sca-1 expression levels (Figure 6h–j) (
). Quantitative real-time reverse transcriptase–PCR indicated that the CCL2 expression levels in the HF keratinocytes reached 17-fold higher values relative to those in interfollicular epidermal keratinocytes after 8 hours of IFN-γ stimulation (Figure 6j). This result indicated that IFN-γ stimulated HF keratinocytes to release high CCL2 levels, attracting interstitial CX3CR1int monocyte migration toward the keratinocytes around HF regions in CHS. Taken together, IFN-γ release after antigen recognition triggers monocyte migration to HF regions and cluster formation through up-regulating CXCL1, CXCL2, MIF, and CCL2 expression.
Discussion
Here, we developed a multicolor-labeled murine CHS model to unravel the mechanism of the mobilization of monocytes for amplifying the adaptive immune response in the skin. We characterized effector T cell-regulated interstitial monocyte cluster formation around HFs after antigen recognition. Close contact between monocytes in the cluster provides an important microenvironment for accelerating monocyte activation in situ, which induces keratinocyte apoptosis in CHS.
Although effector T cells initiate the immune reaction at the periphery after antigen re-challenge in adaptive immunity, it remains unclear how monocytes dynamically amplify an efficient adaptive immune response. By combining spatiotemporal information of immune cell migration, we found that effector T cells encounter DCs under HF regions and release IFN-γ to mobilize monocyte cluster formation around HFs. This presents a change in the spatial distribution of monocytes in the interstitial space. In this process, CCR2 and CXCR2 regulate the formation of monocyte clusters (Figure 3a–e). Specifically, we found that diverse expression of chemokines in keratinocytes triggers cluster formation of interstitial monocytes. The CCL2 level in HF keratinocytes was 17-fold higher than that in interfollicular keratinocytes after IFN-γ challenge (Figure 6h–j). This result confirms that IFN-γ not only mediated the accumulation of monocytes from blood but also mobilized interstitial monocytes to form clusters at the dermal-epidermal junction in the HF region. Previously,
showed that DC–T-cell clusters form around blood capillaries. In our study, HF keratinocytes might secrete chemokines such as CCL20 to attract DC migration toward HFs after hapten challenge, causing the difference in DC–T-cell cluster locations in our model.
Mons/Mφs activation is very important for disease pathogenesis. However, their activation is currently controversial and confusing (
). Our results confirm that cluster formation is a key node event for TNF-α expression in monocytes (Figure 4b). We speculated that tight contact among monocytes in the clusters provided a microenvironment for accelerating monocyte activation in situ. Although previous studies have shown that IFN-γ modulates myeloid cell infiltration and stimulates their activation (
), our results showed that mobilizing the formation of monocyte clusters represents another pathway leading to the stimulation of monocyte activation in addition to direct stimulation by IFN-γ.
Keratinocyte apoptosis is integral to the pathogenesis of dermatitis, and targeting keratinocyte apoptosis has been an effective strategy for treating allergic contact dermatitis (
). Our results show that monocytes are the inflammatory cells that closely connect with keratinocytes in spatial structure of skin (see Supplementary Figure S3 and Supplementary Movie S3). We found that interstitial monocytes mediated keratinocyte death after forming clusters in the dermis in the mouse CHS model (Figure 4c and d). In previous reports, effector T cells have been regarded as the main effector cells that mediate keratinocyte apoptosis through Fas protein (
). Our results show that monocyte cluster formation around HFs was also a necessary step for mediating monocyte activation and keratinocyte apoptosis. Activated monocytes might express mediators like TNF-α, which cooperate with FasL or TWEAK to mediate keratinocyte apoptosis (
). Further work should focus on the molecular mechanism of monocytes that mediate keratinocyte death in CHS.
In conclusion, we characterized a T cell-regulated monocyte cluster formation in CHS. We showed the molecular mechanism involved in monocyte movement and visualized the dynamic spatiotemporal regulation of cell migration, cluster formation, and activation. This study provides intuitive evidence for how effector T cells modulate inflammatory monocytes that amplify the adaptive immune response in a pathological environment. These findings also provide a clue for immunotherapeutic strategies in which the disruption of monocyte clusters could be a target for the treatment of T cell-mediated allergic contact dermatitis.
Materials and Methods
Mice
C57BL/6 female mice were obtained from the Hubei Research Center of Laboratory Animals (Hubei, China). CX3CR1GFP/GFP, mT/mG, albino-C57BL/6, IFN-γ–deficient, and CD11cEYFP transgenic mice were obtained from the Jackson Laboratory (Bar Harbor, ME). All mice were bred and maintained in a specific pathogen-free barrier facility and used at the age of 6–12 weeks. All animal studies were approved by the Animal Experimentation Ethics Committee of Huazhong University of Science and Technology (Hubei, China).
Induction and assessment of CHS responses
Mice were sensitized with 50 μl of 2% (weight/volume) oxazolone (Sigma-Aldrich, St. Louis, MO) diluted in vehicle (acetone:olive oil = 4:1) at the abdomen and then challenged with 10 μl of 1% (weight/volume) OXA at both the dorsal and ventral sides of the ears on day 5. For the methods of adoptive transfer model, whole-mount immunofluorescence, and TUNEL staining to assess apoptosis of the epidermal sheet, please see additional information in the Supplementary Materials online.
Intravital imaging of monocytes, T cells, and keratinocytes
Keratinocytes were visualized by intradermal injection of 1 μg of Alexa Fluor 647-conjugated isolectin B4 (Invitrogen, Waltham, MA). For the detection of dead cells in vivo, 100 μg propidium iodide (Sigma-Aldrich) in phosphate buffered saline was slowly injected intravenously into mice. To simultaneously capture the movement of the three types of cells (DCs, T cells, and monocytes), T cells were isolated from OXA-sensitized mice, labeled with cell tracker dye CMTMR (Invitrogen), and injected intravenously into CX3CR1-CD11c mice. Images were processed and analyzed using the Imaris 7.2.1 software (Bitplane, Belfast, UK). For more detailed parameters of intravital imaging, please see the additional information in the Supplementary Materials.
Statistics
Statistical comparisons between the two groups were evaluated using unpaired Student t tests or analysis of variance corrected for multiple comparisons.
The authors thank Guanxin Shen from Tongji Medical College, Li Zhang from Toronto General Research Institute, and Yanmei Han from Second Military Medical University for paper discussions. We thank the Optical Bioimaging Core Facility of Wuhan National Laboratory for Optoelectronics-Huazhong University of Science and Technology for support in data acquisition and the Analytical and Testing Center of HUST for spectral measurements. This work was supported by the Major Research plan of the National Natural Science Foundation of China (grant no. 91442201), the National Science Fund for Distinguished Young Scholars (81625012), the Science Fund for Creative Research Group of the National Natural Science Foundation of China (grant no. 61721092), the Fundamental Research Funds for the Central Universities (HUST: 2015ZDTD014), the Director Fund of Wuhan National Laboratory for Optoelectronics, and the 111 Project (no. B07038).
Author Contributions
ZL, ZHZ, and QML initiated and designed the project. ZL, FY, ZF, SQ, and LL performed the experiments. ZL and HZ assisted with data processing. JT assisted with paper discussion. ZL and ZHZ analyzed the data and wrote the paper.
Specific uncoupling by islet-activating protein, pertussis toxin, of negative signal transduction via alpha-adrenergic, cholinergic, and opiate receptors in neuroblastoma x glioma hybrid cells.
Hair follicles have recently emerged as immunologically active organs that orchestrate recruitment and trafficking of immune cells within skin. Liu et al. (2018) expand our knowledge in this growing area of research by characterizing the network of immune cell interactions during experimental contact hypersensitivity that, interestingly, is centered around hair follicles.
30038-1&id=gr1.jpg){kind=link}
30038-1&id=gr2.jpg){kind=link}
30038-1&id=gr3.jpg){kind=link}
30038-1&id=gr4.jpg){kind=link}
30038-1&id=gr5.jpg){kind=link}
30038-1&id=gr6.jpg){kind=link}