An aging-susceptible circadian rhythm controls cutaneous antiviral immunity

Abstract

Aged skin is prone to viral infections, but the mechanisms responsible for this immunosenescent immune risk are unclear. We observed that aged murine and human skin expressed reduced levels of antiviral proteins (AVPs) and circadian regulators, including Bmal1 and Clock. Bmal1 and Clock were found to control rhythmic AVP expression in skin, and such circadian control of AVPs was diminished by disruption of immune cell IL-27 signaling and deletion of Bmal1/Clock genes in mouse skin, as well as siRNA-mediated knockdown of CLOCK in human primary keratinocytes. We found that treatment with the circadian-enhancing agents nobiletin and SR8278 reduced infection of herpes simplex virus 1 in epidermal explants and human keratinocytes in a BMAL1/CLOCK-dependent manner. Circadian-enhancing treatment also reversed susceptibility of aging murine skin and human primary keratinocytes to viral infection. These findings reveal an evolutionarily conserved and age-sensitive circadian regulation of cutaneous antiviral immunity, underscoring circadian restoration as an antiviral strategy in aging populations.

Keywords: Aging; Cellular immune response; Dermatology; Innate immunity; Skin.

Figure 1. Aging skin exhibits diminished circadian, AVP, and IL-27 transcription.

(A) Quantitative PCR (qPCR) of Bmal1, Clock, and Per2 in aged (n = 14–16, >1 year old) and young (n = 32–34 1 month old) male murine skin. Graphs represent averages of relative mRNA ± SEM with GAPDH used for an internal control. P values were obtained via 2-tailed Student’s t test. (B) qPCR of BMAL1, PER2, and IFITM1 in human primary keratinocytes over serial passaging (n = 2–3 donors/passage). Graphs represent averages of relative mRNA ± SEM with GAPDH used for an internal control. (C) qRT-PCR of Oas1, Oas2, and Ifitm1 in aged and young murine back skins as described in A. P values were obtained via 2-tailed Student’s t test. (D) Immunostaining for OAS1a (orange) and nuclei (blue) in aged and young nonwounded skin. Scale bar: 25 μm. (E) qRT-PCR of AVP in young and old skin 24 hours after wounding. Graphs represent averages of relative mRNA ± SEM with GAPDH used for an internal control. P values were obtained via 2-way ANOVA with multiple comparison. NW, nonwounded skin; W, wounded skin. (F) Immunostaining for CD301b (green) and nuclei (blue) in aged and young unwounded skin. Scale bar: 25 μm. (G) Flow cytometry showing reduced numbers of CD301b⁺ cells in aged skin compared with young skin (n = 4 mice/group) as a percentage of total harvested live, single cells. P values were obtained via 2-way ANOVA with multiple comparisons. (H) Histogram displays IL-27 production from CD11b⁺ (yellow) and CD301b⁺ (blue) cells compared with CD45^– (red). The flow cytometry gating strategy is included in Supplemental Figure 1. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Figure 2. Circadian rhythm transcriptional networks include antiviral genes in mammalian skin.

(A and B) Line plots showing rhythmic expression of circadian factors and AVP genes in (A) murine skins (GSE38625) and (B) baboon skins (GSE98965). Heatmaps including additional genes can be found in Supplemental Figure 3. (C) qRT-PCR of Oas1a and Oas2 in intact skin of Bmal1^–/– mice and WT littermates. (n = 3 mice/group, with technical triplicates per mouse). Graphs represent averages of relative mRNA ± SEM with GAPDH used for internal control. (D) Immunostaining for OAS1a (purple) in WT and Bmal1^–/– skin. Nuclei are stained blue. Scale bar: 25 μm. (E) qRT-PCR of Oas1a and Ifitm1 in intact belly skin of ClockΔ19 mice and BALB/C WT littermates harvested at 5 am (AM wound) or 5 pm (PM wound). (F) qRT-PCR of Oas1 and Ifitm1 in skin wounds of C57BL/6 inflicted at times indicated and harvested 24 hours later (n = 4 mice/group). For Ifitm1, P = 0.0725. (G) qRT-PCR of Oas1a and Ifitm1 in skin wounds of ClockΔ19 mice and BALB/C WT littermates inflicted at the indicated time and harvested 24 hours later. (E and G) n = 3 mice/group with technical triplicates, except WT pm time group, which used 2 mice. P values reported in this figure were obtained via 2-tailed Student’s t test. *P ≤ 0.05, **P ≤ 0.01. FPKM, fragments per kilobase of transcript per million mapped reads.

Figure 3. Circadian rhythms of wound-induced AVPs require cytokine signaling.

(A) Heatmap of microarray expression of Bmal1^–/– compared with heterozygotes intact murine skin as measured at Zeitgeber time 22 (GSE38625). (B) Immunostaining for CD301b (green) and nuclei (blue) in WT and Bmal1^–/– intact skin. Scale bar: 25 μm. Skin samples are the same sections displayed in Figure 2C. (C) Flow cytometry of CD301b⁺ cells expressing IL-27 as a percentage of total harvested live single cells and MFI of Il27p28 in Bmal1^–/– skin and WT skin (n = 3–4 mice/group). (D) Histogram displays IL27p28 expression in Bmal1^–/– (red line) and WT (blue line) mouse skin. The gating strategy is shown in Supplemental Figure 4. (E) qRT-PCR of Ifitm1 in skin wounds of WT or LysM-Cre.IL27p28^fl/fl mice (n = 3 mice/group). (F) qRT-PCR of Ifitm1 in skin wounds of Ifnar1^–/– mice (n = 3–4 mice). (G) qRT-PCR of type I IFNs in WT C57BL6 mice wounded at 8 am or 8 pm (n = 4 mice/time point), as in (E) above. Graphs represent averages of relative mRNA ± SEM with GAPDH used for an internal control. P values reported in this figure were obtained via 2-tailed Student’s t test. *P ≤ 0.05. NW, nonwounded skin.

Figure 4. Keratinocyte autonomous circadian rhythm regulates antiviral activity.

(A) qRT-PCR of BMAL1, OAS1, OAS2, and MX1 in human primary keratinocytes synchronized via growth factor starvation and harvested every 6 hours (representative of 3 independent experiments). Graphs represent relative mRNA ± SEM with GAPDH used for an internal control and relative to that of 0-hour time point. (B) qRT-PCR of BMAL1, CLOCK, OAS1, and IFITM1 in NTERT keratinocytes transfected with siRNA: siCtrl, siCLOCK, or siBMAL1 (n = 3 biological replicates). Graphs represent averages of relative mRNA ± SEM with GAPDH used for an internal control. (C) qPCR of HSV-1 gene UL29 in cell culture–conditioned medium of primary keratinocytes transfected with siCtrl, siBMAL1, or siCLOCK 24 hours prior to infection with HSV-1 (MOI, 0.01). Knockdown efficacy is shown by qRT-PCR. Graphs represent averages of either relative DNA or mRNA ± SEM with GAPDH used for an internal control. Primary keratinocytes from donors were pooled (n = 3). (D) Immunofluorescence of HSV-1 (MOI, 0.01) in human NTERT keratinocytes transfected with siCtrl or siBMAL1. Scale bar: 160 μm. (E and F) Knockdown efficacy of BMAL1 in human NTERT cells is shown by qRT-PCR. ImageJ (Fiji) quantification of relative viral immunofluorescence normalized to nuclear staining. Box-and-whisker plots represent averages of relative immunofluorescence ± SEM (n = 3 samples/group in E; n > 400 cells/condition in F). P values reported in this figure were obtained via 2-tailed Student’s t test. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Figure 5. Pharmacological augmentation of circadian rhythm reduces HSV-1 infection in a BMAL1/CLOCK-dependent manner.

(A and B) Bmal1–luciferase reporter assay. Human NTERT keratinocytes were transduced with Bmal1–luciferase reporter construct and starved of growth supplement overnight before treatment with vehicle or 10 μM SR8278. Cells were harvested (A) every 6 hours and (B) at 24 hours for RLU measurements (n = 4 samples). (C) Immunostaining for HSV-1 antigen in human epidermal explants infected with HSV-1 and treated with vehicle or 10 μM SR8278. Scale bar: 500 μm. (D) ImageJ (Fiji) quantification of viral immunofluorescence normalized to nuclear staining (n > 5,000 cells quantified per condition). (E) qPCR of HSV-1 viral gene UL29 relative to human KRT14 in epidermal skin infection (n = 4 skin explants). (F) Immunostaining for HSV-1 antigen in human NTERT keratinocytes (MOI, 0.01) transfected with siCtrl or siBMAL1 and supplemented with vehicle or 10 μM SR8278. Scale bar: 160 μm. (A–F) P values were obtained via 2-tailed Student’s t test. (G) ImageJ (Fiji) quantification of relative viral immunofluorescence normalized to nuclear staining. Data for quantification of siCtrl (vehicle) are the same as used in Figure 4E (n > 400 cells/condition quantified). (G–H) P values were obtained using 1-way ANOVA with multiple comparisons. (H) qPCR of HSV-1 viral gene UL29 in cell culture–conditioned medium of primary keratinocytes (MOI, 0.01) transfected with siBMAL1 or siCLOCK and infected with HSV-1 (n = 3). Graphs represent averages of relative viral DNA or viral protein immunofluorescence normalized to human KRT14 or nuclear staining ± SEM. P values were obtained using 1-way ANOVA with multiple comparisons. *P ≤ 0.05, **P ≤ 0.01, ****P ≤ 0.0001. ZT, Zeitgeber time.

Figure 6. Antiviral immune decline of aging skin can be rescued by a circadian enhancer treatment.

(A) qRT-PCR of Oas1a, Oas2, and Ifitm1 in Bmal1^+/– and WT skin that was wounded and collected 24 hours later (n = 3–6 mice/genotype). P value was obtained with 2-way ANOVA with multiple comparisons. (B) qPCR of HSV-1 UL29 gene relative to murine Krt14 in HSV-1–infected aging (>365 days) and young (2–6 months) mouse epidermis (n = 3 skin explants per group). (C) qPCR of viral UL29 relative to murine Krt14 in HSV-1 infection of aging mouse (>365 days) epidermis treated with vehicle or 10 μM SR8278 (n = 3 skin explants per condition). (D and E) qPCR of HSV-1 UL29 in cell culture–conditioned medium of infected P2 and P8 keratinocytes treated with vehicle or 10 μM SR8278 (MOI, 0.01). Graphs represent averages of relative HSV-1 UL29 DNA normalized to Krt14 ± SEM. n = 3/group. (F) Working model of aging-associated decline of cutaneous innate antiviral immunity. Circadian rhythm factors BMAL1 and CLOCK regulate expression of cutaneous AVPs through both keratinocyte-autonomous and leukocyte-mediated processes. IL-27 conveys a time-of-day response and type I IFN signaling ensures a robust antiviral immunity. Aging-associated circadian decline decreases cutaneous antiviral immunity, whereas pharmacological means of circadian enhancement increases it. Except for A, P values reported in this figure were obtained via 2-tailed Student’s t test. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. NW, nonwounded skin; W, wounded skin.