3.1 Sleep deprivation exacerbates circadian and behavioral disturbances in SiNPs-treated mice
After two weeks of treatment, pallet and water intake, body weight, core body temperature, and locomotor activity were assessed in the NC, SiNPs-treated, and SD + SiNPs-treated groups. Locomotor activity, monitored via telemetry, revealed notable alterations in the SiNPs-treated and SD + SiNPs-treated mice compared to controls (Figs. 2A-B). During the daytime (ZT0-ZT12), the SD + SiNPs-treated group exhibited higher levels of activity than the NC and SiNPs-treated groups. Conversely, during the nighttime (ZT12-ZT24), the SD + SiNPs-treated group’s locomotor activity decreased to lower levels, while the NC and SiNPs-treated groups showed increased activity levels that were higher than those of the SD + SiNPs group. Although the SiNPs-treated group followed a similar 24-hour trend to the NC group, its average activity level over the day was lower than that of the NC group. The daily rhythms of the SiNPs-treated and SD + SiNPs-treated groups also differed significantly from the NC group (Figs. 2A-B). The core body temperature of the NC group followed a normal circadian pattern, increasing with locomotor activity during the dark cycle (Fig. 2C). In contrast, the SiNPs-treated and SD + SiNPs-treated group showed an abnormal pattern, their core body temperature was lower than the NC group in both the light and dark cycles (Figs. 2C-D).
Behavioral and physiological alterations in mice following SiNPs and SD + SiNPs treatments. (A-B) Locomotor activity patterns over a 24-hour period in the NC, SiNPs-treated, and SD + SiNPs-treated groups. Data were collected every 5 min. The gray shading indicates the dark phase. *P < 0.05, ***P < 0.001. (C-D) Core body temperature fluctuations over a 24-hour period in the NC, SiNPs-treated, and SD + SiNPs-treated groups. Data were recorded every 20 min. The gray shading indicates the dark phase. *P < 0.05, ***P < 0.001. (E) Pellet consumption in the NC, SiNPs-treated, and SD + SiNPs-treated groups. NS: not significant. (F) Water intake in the NC, SiNPs-treated, and SD + SiNPs-treated groups. NS: not significant. (G) Body weight changes in the NC, SiNPs-treated, and SD + SiNPs-treated groups. *P < 0.05, ***P < 0.001
To assess the effect of SD on the general behaviors of the mice, we collected data on pellet intake, water intake, and body weight over a two-week period. We observed no significant statistical differences in water intake and pellet intake among the three groups (Fig. 2E-F). Pellet and water intake were slightly lower in the SD + SiNPs-treated group compared to the NC group, but the difference was not statistically significant. The SiNPs-treated group exhibited the lowest intake among the three groups, yet these reductions were also not statistically significant (Fig. 2E-F). Despite the lack of significant differences in water intake, there was a clear trend in both pellet and water consumption, as well as a significant difference in body weight, with the SD + SiNPs-treated group weighing less than the other two groups (Fig. 2G).
Collectively, these results demonstrate that SD significantly exacerbates SiNPs-induced disturbances in circadian behavior and physiology, as evidenced by alterations in locomotor activity, core temperature regulation, and body weight.
3.2 Structural alterations in the ELGs triggered by sleep deprivation in SiNPs-treated mice
Our previous studies have shown that the size and weight of ELGs exhibit circadian oscillations, which can be modulated by both intrinsic (e.g., circadian disruption) [33] and extrinsic factors (e.g., type 1 diabetes) [48]. Building upon these findings, we investigated whether exposure to SiNPs and SD, individually or in combination, could alter the physiological characteristics of ELGs, particularly their structure and function.
ELG weights were measured at four ZT points (ZT0, ZT6, ZT12, and ZT18) across all experimental groups. In the NC group, ELG weight displayed robust diurnal variation, peaking at ZT18. This rhythmicity was abolished in both the SiNPs-treated and SD + SiNPs-treated groups. Notably, SD further exacerbated the disruption caused by SiNPs (Fig. 3A). Overall, ELG weights were significantly lower in both treatment groups compared to NC, indicating compromised lacrimal gland homeostasis (Fig. 3B).
Impact of SiNPs and SD + SiNPs treatment on ELG weight, tear secretion, and structural integrity. (A) Diurnal changes of ELG weight in the NC, SiNPs-treated, and SD + SiNPs-treated groups. *P < 0.05, ***P < 0.001. NS: not significant. (B) ELG weight measurements in the NC, SiNPs-treated, and SD + SiNPs-treated groups. **P < 0.01, ***P < 0.001. (C) Tear secretion was assessed using the phenol thread test at ZT 0, 6, 12, and 18 over a 24-hour cycle. Statistical significance is denoted as *P < 0.05 for comparisons between SiNPs-treated and SD + SiNPs-treated groups and ^P < 0.05 for comparisons between the NC and SD + SiNPs-treated groups. (D-F) Representative gross sections of ELGs from the NC (D), SiNPs-treated (E), and SD + SiNPs-treated (F) groups (scale bar: 500 μm, upper panels). Magnified views of the boxed regions show the acinar cell morphology (scale bar: 20 μm, lower panels). (G) Diurnal changes of ELG cell counts in the NC, SiNPs-treated, and SD + SiNPs-treated groups. *P < 0.05, ***P < 0.001. (H) Quantitative analysis of acinar cell counts in the NC, SiNPs-treated, and SD + SiNPs-treated groups. Statistical analysis was performed using Brown-Forsythe ANOVA. Multiple comparisons were conducted using the Games-Howell post hoc test. ***P < 0.001
Tear secretion, was assessed using phenol red thread tests, also followed a circadian rhythm in the NC group, with a peak at ZT18. This rhythmic pattern was absent in the SiNPs-treated and SD + SiNPs-treated groups. Importantly, no statistically significant differences in tear volume were observed between the two treatment groups, suggesting that SiNPs-induced dysfunction had already reached a plateau, with SD (Fig. 3C).
To assess structural changes in the lacrimal gland, HE staining was performed for histopathological examination (Figs. 3D-F). Figure 3G shows daily oscillations in the NC group, which is disrupted in both the SiNPs-treated and SD + SiNPs-treated groups. Quantitative analysis in Fig. 3H revealed significant differences in the number of cell nuclei among the three groups. The number of nuclei was significantly reduced in the SiNPs-treated group compared to the NC group, and further decreased in the SD + SiNPs-treated group relative to both the NC group and the SiNPs-treated group. These findings suggest that both SiNPs exposure and the combined intervention with SD induce structural abnormalities and functional impairment in ELGs, with more severe effects observed in the SD + SiNPs-treated group.
Collectively, these findings demonstrate that SD intensifies SiNPs-evoked disruptions in locomotor activity, core temperature regulation, and body weight.
3.3 Alterations in global gene expression of ELGs promoted by sleep deprivation in SiNPs-treated mice
To investigate the oscillations in the transcriptome of murine ELGs, we collected ELGs at three-hour intervals throughout a 24-hour period from the NC, SiNPs-treated, and SD + SiNPs-treated groups of C57BL/6J mice. RNA-seq was performed on the BGISEQ-500 platform, and transcriptome analysis was conducted using the JTK_CYCLE algorithm with a 24-hour oscillation period and a significance threshold of P < 0.05.
A total of 20,242 genes were identified through RNA-seq analysis. Genes were categorized into rhythmic (FPKM ≥ 0.1, adjusted P < 0.05), non-rhythmic (FPKM ≥ 0.1, adjusted P ≥ 0.05), and low-expressed (FPKM < 0.1) based on the obtained data. Among the 20,242 transcripts, rhythmic genes constituted 15.10%, 13.50%, and 16.26% of the NC, SiNPs-treated, and SD + SiNPs-treated groups, respectively. Low-expressed genes accounted for 33.30%, 33.02%, and 34.22% in the respective groups. Non-rhythmic genes were present in 51.60%, 53.48%, and 49.52% of the NC, SiNPs-treated, and SD + SiNPs-treated groups (Figs. 4A-C).
Comparative transcriptomic analysis of mouse ELGs in response to SiNPs and SD + SiNPs treatments. (A–C) The pie charts illustrate the transcriptomic composition of ELGs in the NC, SiNPs-treated, and SD + SiNPs-treated groups, categorized into low-expressed, rhythmic, and non-rhythmic genes. (D) The PCA scatterplot depicts the overall gene expression profiles of ELGs in the NC (blue), SiNPs-treated (red), and SD + SiNPs-treated (yellow) groups. Each dot represents an individual animal sampled at three-hour intervals over a 24-hour cycle (N = 3 per time point). The shaded regions indicate the distribution of each group. (E) The line graph illustrates the temporal distribution of peak gene expression across different ZT points in the NC, SiNPs-treated, and SD + SiNPs-treated groups. Differences in peak expression times highlight the impact of treatments on rhythmic gene regulation. (F) The volcano plot visualizes DEGs in ELGs between NC and SiNPs-treated mice. The x-axis represents ZT points, while the y-axis indicates fold change (FC). Red and gray dots denote genes with adjusted P-values < 0.01 and ≥ 0.01, respectively. Each group consisted of 24 mice
Compared to the NC group, the SiNPs-treated group experienced a decrease of 0.27% in low-expressed genes and 1.61% in rhythmic genes, whereas the SD + SiNPs-treated group showed an increase of 0.92% in low-expressed genes and 1.16% in rhythmic genes. The proportion of non-rhythmic genes increased by 1.88% in the SiNPs-treated group and decreased by 2.48% in the SD + SiNPs-treated group relative to NC (Figs. 4A-C).
Principal component analysis (PCA) was employed to discern distinct groupings among the treatments. The first three principal components accounted for 41.2%, 30.9%, and 16.4% of the variance, respectively (Fig. 4D). The PCA biplots demonstrated that the NC, SiNPs-treated, and SD + SiNPs-treated groups formed three distinct clusters, indicating that SiNPs and SD + SiNPs treatments have a significant impact on differential gene expression.
The distribution of gene expression peaks over a 24-hour period revealed the effects of SiNPs and SD + SiNPs treatments (Fig. 4E). The NC group exhibited peak gene expression at ZT12 and ZT15. In contrast, the SiNPs-treated group displayed peaks at ZT0, ZT3, ZT6, and ZT21. The SD + SiNPs-treated group showed peak expression at ZT9 and ZT18.
Volcano plot effectively highlights the top10 significantly differentially expressed genes (DEGs) between the NC group and the SiNPs-treated group (Fig. 4F and Table S1), providing insights into the effect of SiNPs treatment on ELGs. Volcano plots of significant DEGs between the NC group and the SD + SiNPs-treated, and between the SiNPs-treated and the SD + SiNPs-treated, are presented in Figure S2 and Tables S2-3.
Overall, these findings demonstrate SiNPs exposure—accentuated by SD—substantially alters the ELG transcriptome, differentially affecting rhythmic and non-rhythmic genes and highlighting the circadian sensitivity of the lacrimal gland to environmental and behavioral stressors.
3.4 Disruption of circadian rhythmicity of ELGs promoted by sleep deprivation in SiNPs-treated mice
To explore the impact of SiNPs treatment and SD + SiNPs treatment on the circadian rhythmicity of ELGs, we employed Venn diagrams, heatmaps, pie charts, and rose diagrams to analyze the rhythmic genes among the three groups. Venn diagrams were utilized to identify unique and shared rhythmic genes across the NC, SiNPs-treated, and SD + SiNPs-treated groups. Among the 5,901 non-redundant gene transcripts, the proportion of unique rhythmic genes in each group was 18.53% for the NC, 17.32% for the SiNPs-treated, and 23.88% for the SD + SiNPs-treated groups (Fig. 5A and Table S4).
Effects of SiNPs and SD + SiNPs treatment on the rhythmic transcriptome of murine ELGs. (A) The venn diagram illustrates the overlap and divergence of rhythmic transcripts among the NC, SiNPs-treated, and SD + SiNPs-treated groups. (B) The heatmaps display the expression profiles of 1,094 rhythmic transcripts unique to the NC group at different ZT points. The left panel represents the NC group, while the middle and right panels correspond to the SiNPs-treated and SD + SiNPs-treated groups, respectively. Gene expression levels are normalized within a ± 2 range, as indicated by the color scale. (C) The heatmaps show the expression levels of 1,022 rhythmic transcripts exclusive to the SiNPs-treated group across various ZT points. The panel arrangement is consistent with (B), with the NC group on the left, the SiNPs-treated group in the middle, and the SD + SiNPs-treated group on the right. (D) The heatmaps illustrates the expression patterns of 1,409 rhythmic transcripts unique to the SD + SiNPs-treated group over multiple ZT points, following the same panel arrangement as in (B) and (C). (E) The venn diagram depicts the transition of rhythmic genes in the NC group to either non-rhythmic (blue) or low-expressed (red) genes in the SiNPs-treated group. (F) The venn diagram illustrates the conversion of rhythmic genes in the SiNPs-treated group into non-rhythmic (red) or low-expressed (yellow) genes in the SD + SiNPs-treated group. (G-H) The wind rose diagrams represent the mean vector and length of rhythmic genes unique to each group (G) and those shared across all three groups (H). The NC, SiNPs-treated, and SD + SiNPs-treated groups are positioned on the right, middle, and left, respectively
Heatmaps were constructed to illustrate the expression levels of 1,096 unique genes in the NC group (Figs. 5B), 1,022 in the SiNPs-treated group (Figs. 5C), and 1,409 in the SD + SiNPs-treated group (Figs. 5D) across eight ZT points. The heatmaps revealed significant alterations in gene expression patterns between the SiNPs-treated group and the NC group, with further pronounced changes observed in the SD + SiNPs-treated group.
Pie charts were used to depict the transformation of rhythmic genes post-treatment (Figs. 5E-F). Of the rhythmic genes unique to NC group, 1,470 maintained their rhythmic expression after SiNPs treatment, while 1,568 transitioned to non-rhythmic status (Fig. 5E). In the SiNPs-treated group, none of the unique rhythmic genes remained unchanged after SD + SiNPs treatment; 1,005 became non-rhythmic, and 17 became low-expressed (Fig. 5F).
The rose diagrams, generated using Oriana software, visually demonstrated the differences in the periodicity and phase distribution of unique and shared rhythmic genes across groups. The NC group exhibited a mean vector (µ) of 9:30 and a mean vector length (r) of 0.635, the SiNPs-treated group showed µ = 6:40 and r = 0.163, and the SD + SiNPs-treated group presented µ = 9:26 and r = 0.596 (Fig. 5G). Figure 5H illustrates the changes in the average vector and its length for shared rhythmic genes under different treatments: NC group (µ = 9:37, r = 0.55), SiNPs-treated group (µ = 8:29, r = 0.463), and SD + SiNPs-treated group (µ = 8:34, r = 0.532).
The composite pie charts revealed phase shifts among shared genes across groups. After SiNPs treatment, 33.83% of shared genes in the NC group maintained their phase, while 66.17% experienced phase shifts, with 68.98% advancing and 31.02% delaying. Following SD + SiNPs treatment, 34.45% of shared genes remained phase-stable, 65.55% underwent phase changes, with 43.83% delaying and 56.17% advancing. In the SD + SiNPs treatment group, 82.71% of shared genes exhibited phase alterations, 66.32% advanced, and 33.68% delayed, with only 17.29% remaining unchanged (Fig. S1).
The findings indicate that both SiNPs treatment and the combination of SD + SiNPs treatment significantly disrupt the circadian rhythmicity of ELGs, leading to substantial alterations in the expression and phase distribution of rhythmic genes, which underscores the profound impact of these treatments on the molecular clock of lacrimal gland function.
3.5 Reshaped KEGG and phase-set enriched pathways triggered by sleep deprivation in SiNPs-treated mice
To evaluate the functional implications of circadian gene expression changes, KEGG pathway enrichment analysis was performed on rhythmic genes unique to the NC, SiNPs-treated, and SD + SiNPs-treated groups. The genes exclusive to the NC group were significantly enriched in nine pathways (Fig. 6A and Table S5, P < 0.05), with a focus on cellular processes and metabolic pathways. However, the rhythmic genes of SiNPs-treated group enriched in different KEGG pathways from the NC group (Fig. 6B and Table S6, P < 0.05), so did the rhythmic genes of SD + SiNPs-treated group (Fig. 6C and Table S7, Q < 0.05). Notably, the cell cycle pathway was the only pathway shared between the SiNPs-treated and SD + SiNPs-treated groups.
Impact of SiNPs and SD + SiNPs treatment on KEGG and phase-clustered pathways in mouse ELGs. (A–D) Gene annotation of KEGG pathways significantly enriched in rhythmic genes unique to the NC group (A), SiNPs-treated group (B), SD + SiNPs-treated group (C), and those shared among all three groups (D), with P < 0.01. The top 10 enriched pathways are presented. The upper horizontal axis, aligned with the orange line graph, represents the number of term candidate genes, while the lower horizontal axis, aligned with the blue histogram, represents the -log10 (P-value), indicating the statistical significance of enrichment. (E–G) Summary of significantly phase-clustered pathways (P < 0.05) unique to the NC group (E), SiNPs-treated group (F), and SD + SiNPs-treated group (G). The inner circle and column length represent the phase distribution of rhythmic genes specific to each group. The outer red line marks KEGG pathways (P < 0.05) associated with rhythmic genes unique to each group, indicating enrichment at distinct ZT points, as determined by the phase distribution in the inner circle. Gray shading indicates dark cycles
A total of 18 pathways were enriched among genes common to all groups (Fig. 6D and Table S8, Q < 0.05). The top 10 pathways identified for each group were categorized into five groups: Cellular Processes, Genetic Information Processing, Metabolism, Organismal Systems, and Environmental Information Processing. The SiNPs-treated group’s pathways fell within Cellular Processes, Genetic Information Processing, and Metabolism (Table S5). After the SD + SiNPs treatment, the enriched pathways expanded to include all five categories (Table S6). Although the SD + SiNPs-treated group’s unique pathways matched the control group’s categories, the SD + SiNPs treatment did not fully counteract the effects of SiNPs treatment (Table S7). According to KEGG pathway level 2, the biological process categories for the SD + SiNPs-treated group diverged from those of the NC and SiNPs-treated groups, suggesting that SD + SiNPs treatment may amplify the effects of SiNPs treatment.
PSEA was employed to investigate the temporal expression patterns of these pathways. The NC group’s significantly enriched pathways were predominantly expressed between ZT9 and ZT12 (Fig. 6E and Table S9, Kuiper Q < 0.05). In contrast, the SiNPs-treated group had three pathways peaking near ZT4, ZT5, and ZT9 (Fig. 6F and Table S10, Kuiper Q < 0.05). The SD + SiNPs-treated group displayed enriched pathway expression primarily between ZT6 and ZT12, with two pathways, B cell receptor signaling and cytokine receptor interaction, exhibiting expression during the dark phase (Fig. 6G and Table S11, Kuiper Q < 0.05).
Overall, these findings demonstrate that SiNPs disturb the phase distribution of multiple ELG pathways, and SD broadens and deepens these shifts, showing that combined environmental and behavioral stressors can seriously compromise circadian and functional homeostasis in the lacrimal gland.
3.6 Alterations in cluster-dependent transcriptomic map and KEGG pathways triggered by sleep deprivation in SiNPs-treated mice
To investigate the temporal trends in rhythmic gene expression, we applied the Mfuzz soft clustering package for analysis. Our findings revealed that genes from the NC group (Figs. 7A-D), the SiNPs-treated group (Figs. 7E-H), the SD + SiNPs-treated group (Figs. 7I-L) were each categorized into four distinct clusters, each exhibiting unique expression profiles and enriched pathways.
Impact of SiNPs and SD + SiNPs treatment on the circadian gene clustering profile and KEGG pathways in murine ELGs. (A–L) Temporal gene expression Z-scores for four distinct enriched clusters unique to the NC group (A–D), SiNPs-treated group (E–H), and SD + SiNPs-treated group (I–L). Blue, orange, and yellow lines represent genes with low membership values, while pink, purple, and green lines indicate genes with high membership values. Gray shading denotes dark cycles. The corresponding right-side panels display the top 10 KEGG pathways (P < 0.05) enriched for circadian genes unique to each group and cluster
In Cluster 1, significant disparities were observed among the NC, SiNPs-treated, and SD + SiNPs-treated groups, with gene counts of 439, 242, and 289, respectively. The NC group exhibited a notably higher number of clustered genes. The genes in the NC group peaked at ZT9 and troughed at ZT3 and ZT18 (Fig. 7A, left). In contrast, the SiNPs-treated group peaked at ZT3 and troughed at ZT12 (Fig. 7E, left), whereas the SD + SiNPs-treated group peaked at ZT12 (Fig. 7I, left). These distinct global patterns underscore the divergent regulatory mechanisms at play.
Cluster 2 showed 367, 157, and 503 genes in the NC, SiNPs-treated, and SD + SiNPs-treated groups, respectively. The NC and SD + SiNPs-treated groups shared similar expression patterns, peaking at ZT9 and troughing at ZT18 (Figs. 7B and F, left). However, the SiNPs-treated group deviated from this trend, peaking at ZT12 and troughing at ZT3 (Fig. 7J, left).
The gene expression patterns in Cluster 3 were largely concordant across the groups, with gene counts of 165, 269, and 231, respectively, albeit with variations in the timing of peaks and troughs (Figs. 7C, G and K, left). In Cluster 4, the SiNPs-treated and SD + SiNPs-treated groups exhibited similar gene counts (358 and 392, respectively) (Figs. 7H and L, left), in contrast to the NC group, which had a significantly lower count of 127 (Fig. 7D, left). The expression patterns of enriched genes in the SiNPs-treated and SD + SiNPs-treated groups were broadly similar but diverged from those of the NC group.
Furthermore, the KEGG pathways enriched in genes belonging to each cluster varied across groups, indicating a reprogramming of the transcriptomic landscape in response to SiNPs treatment and SD (Figs. 7A-L, right). The KEGG pathways associated with the NC, SiNPs-treated, and SD + SiNPs-treated groups, along with their respective clusters, are shown in Tables S12-14, respectively. This finding suggests a complex interplay between these factors in modulating the biological pathways of ELGs.
In summary, soft clustering and KEGG pathway analysis revealed that SD, particularly when combined with SiNPs exposure, reprograms the temporal architecture of the ELG transcriptome. These alterations reflect a complex interaction between environmental and behavioral stressors and have profound implications for the regulation of lacrimal gland function and circadian homeostasis.
3.7 Alterations in core clock genes of ELGs promoted by sleep deprivation in SiNPs-treated mice
To investigate the effects of SiNPs and SD on the molecular circadian machinery of the lacrimal glands, we analyzed the temporal expression profiles of 12 core clock genes (Nr1d1, Nr1d2, Clock, Per1, Per2, Per3, Arntl, Cry1, Cry2, Npas2, Rora, and Rorc) across the NC, SiNPs-treated, and SD + SiNPs-treated groups(Fig. 8A). A two-way repeated measures ANOVA was performed to evaluate the effects of group, time point, and their interaction on gene expression levels. The results are summarized in Table S15. A significant main effect of time was observed for Per3, Cry2, and Npas2, while no significant effect of group or group × time interaction was found. These results indicate that gene expression varied over time, consistent with circadian regulation, but was not significantly affected by SiNPs or SD treatments. For Nr1d1, Arntl, Per1, Cry1, and Rora, significant main effects of time and group × time interaction were detected, but not for group alone, indicating that temporal expression dynamics differed between treatments despite similar overall expression levels. Although significant main effects of group, time, and group × time interaction were observed for Nr1d2, Clock, Per2, and Rorc, post hoc comparisons between groups did not reveal statistically significant differences. These findings suggest that treatment influenced the temporal expression dynamics of these genes, but differences between groups were relatively modest and did not reflect persistent changes in average expression levels. A separate two-way ANOVA confirmed that treatment, time point, and their interaction significantly influenced gene expression levels. Independent-sample t-tests at individual time points showed that, although not all comparisons reached significance, both SiNPs and SD + SiNPs treatments independently or synergistically affected gene expression at several time points. Furthermore, the SD + SiNPs-treated group exhibited distinct expression patterns compared to the SiNPs-treated group at several time points.
Impact of SiNPs and SD + SiNPs treatment on circadian transcription in murine ELGs. (A) The 24-hour expression profiles of 12 core clock genes, including Nr1d1 (REV-ERBα), Nr1d2 (REV-ERBβ), Clock, Per1, Per2, Per3, Arntl (Bmal1), Cry1, Cry2, Npas2, Rora, and Rorc, are presented. The x-axis represents the sampling time points, while the y-axis indicates gene expression levels at specific ZT points. The green, orange, and yellow lines correspond to the NC, SiNPs-treated, and SD + SiNPs-treated groups, respectively. Gray shading denotes the dark phase of the LD cycle. Three animals per group were sampled every three hours. At each time point, independent samples t-tests were applied to assess differences between the NC and SiNPs-treated groups, the NC and SD + SiNPs-treated groups, and the SiNPs-treated and SD + SiNPs-treated groups. Statistical significance is indicated as follows: *P < 0.05, **P < 0.01, ***P < 0.001 for comparisons between the NC and SiNPs-treated groups; ^P < 0.05, ^^P < 0.01, ^^^P < 0.001 for comparisons between the NC and SD + SiNPs-treated groups; and sP < 0.05, ssP < 0.01, sssP < 0.001 for comparisons between the SiNPs-treated and SD + SiNPs-treated groups. (B) The distribution of peak phases for core clock genes is shown for the NC, SiNPs-treated, and SD + SiNPs-treated groups. Gray shading represents the dark phase of the circadian cycle
Figure 8B illustrates the expression phases of the core clock genes across the three groups. Following SiNPs treatment and SD + SiNPs treatment, the expression phases of 11 out of 12 core clock genes, excluding Per2, were altered relative to the NC group. Specifically, the phases of Nr1d1, Nr1d2, Clock, Per1, Per3, and Cry2 were affected by SiNPs treatment and remained consistent with those of the SiNPs-treated group following SD + SiNPs treatment. The phases of Rorc, Arntl, Cry1, and Npas2 remained unchanged after SiNPs treatment but shifted following SD + SiNPs treatment. The phase of Rora was influenced by both treatments. Specifically, in the NC group, four clock components are present in the light cycle, exhibiting four peak phases: 01:30 (Npas2), 04:30 (Artnl), 06:00 (Clock), and 10:30 (Nr1d1). Six clock components are found in the dark cycle, with peak phases at 13:30 (Cry2 and Per3), 15:00 (Per1 and Per2), 19:30 (Cry1), and 22:30 (Rorc). Additionally, two peak phases at 12:00 (Rora and Nr1d2). Furthermore, in the SiNPs-treated group, five clock components are distributed in the light cycle, with four peak phases: 01:30 (Npsa2), 03:00 (Clock and Artnl), 04:30 (Rora), and 09:00 (Nr1d1). In the dark cycle, five clock components are present, showing four peak phases at 13:30 (Per1 and Nr1d2), 15:00 (Per2), 19:30 (Cry1), and 22:30 (Rorc). At the transition between the light and dark cycles, two clock components are observed, peaking at 12:00 (Per3 and Cry2). Moreover, in the SD + SiNPs-treated group, four clock components are found in the light cycle, with three peak phases at 03:00 (Clock and Artnl), 09:00 (Nr1d1), and 10:30 (Nr1d2). In the dark cycle, four clock components show peak phases at 13:30 (Per1), 15:00 (Per2), 18:00 (Cry1), and 22:30 (Rorc). At the junction of the light and dark cycles, four clock components peak at 00:00 (Npas2 and Rora) and 12:00 (Cry2 and Per3).
Collectively, these findings suggest that while SiNPs and SD + SiNPs treatments do not drastically alter the average expression levels of core circadian genes, they significantly affect their phase relationships and temporal dynamics. This reprogramming of clock gene oscillations highlights a nuanced yet critical mode of circadian disruption within the ELGs under environmental and behavioral stress.
3.8 Immune cell infiltration triggered by sleep deprivation in SiNPs-treated mice
Leukocyte migration into peripheral tissues is tightly regulated by circadian rhythms [49]. To examine the impact of SD on the immune microenvironment of the ELGs, the temporal dynamics of CD4⁺ and CD8⁺ T cell populations were analyzed over a 24-hour period. A significant increase in CD4⁺ T cell numbers was observed between the SD + SiNPs-treated group and the NC group, with SD altering the peak time of CD4⁺ T cell accumulation in the ELGs (Figs. 9A, C-D). Similar trends were observed for CD8⁺ T cells (Figs. 9B, E-F).
Impact of SiNPs and SD + SiNPs treatment on immune cells and genes in murine ELGs. (A) Representative immunohistochemical images of CD4+ T cells in murine ELGs at ZT0 and ZT12, from NC, SiNPs-treated and SiNPs + SD-treated groups. Scale bar: 20 μm. (B) Representative immunohistochemical images of CD8+ T cells in murine ELGs at ZT0 and ZT12, from the NC, SiNPs-treated, and SD + SiNPs-treated groups. Scale bar: 50 μm. (C) Quantitative analysis of CD4+ T cell in murine ELGs, comparing the diurnal variation of positive cell ratio among the NC group, the SiNPs-treated group and the SD + SiNPs-treated group. **P < 0.01. (D) Average abundance of CD4+ T cell in murine ELGs from the NC, the SiNPs-treated and the SD + SiNPs-treated groups. Statistical analysis was performed using the Kruskal–Wallis test (non-parametric), followed by Dunn’s post hoc test for multiple comparisons. *P < 0.05, ***P < 0.001. (E) Quantitative analysis of CD8+ T cell in murine ELGs, comparing the diurnal variation of positive cell ratio among NC group, SiNPs-treated group and SD + SiNPs-treated group. For NC group, P = 0.3391. For SiNPs-treated group, P = 0.4931. For SD + SiNPs-treated group, P = 0.0001. *P < 0.05, **P < 0.01, ***P < 0.001. NS: not significant. (F) Average abundance of CD8+ T cells in murine ELGs from NC, SiNPs-treated and SD + SiNPs-treated groups. **P < 0.01. (G) Heatmaps of diurnal expression for immune-related DEGs between the NC group and SiNPs-treated group in murine ELGs. The expression levels of immune-related genes were obtained from RNA-Seq and expression range of DEGs was normalized to ± 3. (H) The PPINs and functional clusters (cluster 1–3) with relevant KEGG pathways of immune-related genes between the SiNPs-treated group and SD + SiNPs-treated group. (I) The top 10 KEGG pathways enriched histogram of immune-related genes with P < 0.05 were displayed. (J) Immunoblotting of phosphorylation of STAT3, JAK2, phosphorylation of IκBα and p65, and IL17A in ELGs at ZT0 and ZT12, from NC, SiNPs-treated and SD + SiNPs-treated groups
Further analysis of immune-related gene expression revealed significant transcriptional alterations in the SiNPs-treated group compared to the NC group, with SD further exacerbating these changes, as illustrated by the heatmap (Fig. 9G). Compared to NC group, SiNPs-treated group exhibited 134 upregulated and 69 downregulated genes, while SD + SiNPs-treated group showed 81 upregulated and 122 downregulated genes (Table S16). Protein-protein interaction network (PPIN) analysis highlighted the central roles of multiple core genes within these pathways (Fig. 9H), suggesting their potential involvement as key regulatory factors in SD- and particulate matter (PM)-induced lacrimal gland damage. Additionally, KEGG pathway enrichment analysis identified key immune pathways, including the NF-κB signaling pathway, IL-17 signaling pathway, JAK-STAT signaling pathway, chemokine signaling pathway, and Th1/Th2 cell differentiation (Fig. 9I and Table S17). To further evaluate the activation status of these pathways, we measured specific molecular markers: phosphorylation of STAT3 and JAK2 for the JAK-STAT pathway, phosphorylation of IκBα and p65 for the NF-κB pathway, and IL-17A expression levels for the IL-17 signaling pathway. As shown in Fig. 9J and Figures S3-S4, SiNPs significantly increased the phosphorylation of STAT3, JAK2, IκBα, and p65, as well as the expression of IL-17A. Furthermore, SD augmented these SiNPs-induced increases (Fig. 9J and Figures S3–S4). The dysregulation of these key immune pathways likely contributes to chronic inflammation and disruption of immune homeostasis, thereby exacerbating lacrimal gland injury and dysfunction.
Collectively, these results demonstrate that SD exacerbates SiNPs-induced immune perturbations by enhancing T cell infiltration and amplifying pro-inflammatory signaling in the lacrimal gland. The combined activation of JAK-STAT, NF-κB, and IL-17A pathways suggests a mechanism for sustained immune activation and tissue injury, contributing to the progression of dry eye pathology.
3.9 Neuro-related transcriptome profile alterations in the ELGs triggered by sleep deprivation in SiNPs-treated mice
To evaluate the impact of SD and SiNPs exposure on neural functions in murine ELGs, nerve-related genes were identified and analyzed. A total of 98 upregulated and 74 downregulated DEGs were identified between the NC and SiNPs-treated groups, and 96 upregulated and 75 downregulated DEGs between the NC and SD + SiNPs-treated groups (Table S18). The heatmap illustrates distinct expression patterns of neuro-related genes among the NC, SiNPs-treated, and SD + SiNPs-treated groups within the LD cycle (Fig. 10A). KEGG enrichment analysis identified the top 10 significantly enriched pathways, primarily associated with signal transduction, circadian rhythm synchronization, and neural regulation (Fig. 10C and Table S19).
Immune alterations in murine ELGs following SiNPs and SD + SiNPs treatments. (A) Heatmaps of diurnal expression for nerve-related DEGs between the NC group and SiNPs-treated group in murine ELGs. The expression levels of immune-related genes were obtained from RNA-Seq and expression range of DEGs was normalized to ± 4. (B) The PPINs and functional clusters (cluster 1–3) with relevant KEGG pathways of nerve-related genes between the NC group and SiNPs-treated group. (C) The top 10 KEGG pathways enriched histogram of nerve-related genes with P < 0.05 were displayed. (D) Representative images of anti-β-III tubulin immunostaining in ELGs from the NC, SiNPs, and SD + SiNPs groups. Scale bar:20 μm. (E) Quantitative analysis of anti-β-III tubulin staining in ELGs from the NC, SiNPs, and SD + SiNPs groups. Statistical analysis was performed using Brown-Forsythe ANOVA. Multiple comparisons were conducted using the Games-Howell post hoc test. For NC vs. SiNPs, P = 0.0007. For NC vs. SD + SiNPs, P < 0.0001. For SiNPs vs. SD + SiNPs, P < 0.0001. ***P < 0.001
To further explore the interactions among these genes, PPINs were constructed, categorizing neuro-related genes into three clusters (Fig. 10B). According to KEGG functional annotation, SiNPs exposure may induce neurological dysfunction in ELGs, characterized by disruptions in signal transduction pathways, inhibition of neurogenesis, and neurotransmitter cycle disorders.
Immunofluorescence analysis of β-III tubulin, a neuronal marker, revealed decreased signal intensity in the ELGs of the SiNPs-treated and SD + SiNPs-treated mice, indicating reduced innervation or neurodegeneration (Fig. 10D). Quantification showed a significant reduction in the β-III tubulin-positive density in the SiNPs-treated group compared to the NC group, and was further diminished under SD (Fig. 10E). These findings suggest that SiNPs compromise the neural density of the ELGs in mice, potentially contributing to decreased tear secretion. SD intensifies SiNPs-induced neural damage, thereby promoting lacrimal dysfunction and dry eye progression.
3.10 ROS accumulation and DNA damage in the ELGs triggered by sleep deprivation in SiNPs-treated mice
To examine the impact of SD on oxidative stress in the ELGs, immunofluorescence staining for ROS was performed. Compared to the NC group, the SiNPs-treated group exhibited a significant increase in ROS accumulation (P < 0.01). This increase was further amplified by SD, as evidenced by enhanced fluorescence intensity (Figs. 11A-C). We also measured MDA levels, a marker of lipid peroxidation, in the ELGs. Consistent with the ROS findings, MDA levels were significantly elevated in the SiNPs group compared to the NC group (P < 0.001). SD further augmented this increase in MDA (Figs. 11D-E). Furthermore, both SiNPs exposure and the combined SiNPs + SD treatment significantly disrupted oscillatory rhythms compared with the NC group (Figs. 11C and E).
Oxidative stress and DNA damage in murine ELGs following SiNPs and SD + SiNPs treatments. (A) Representative images showing ROS levels at ZT0, ZT6, ZT12, and ZT18 in murine ELGs from the NC, SiNPs-treated, and SD + SiNPs-treated groups. Scale bar: 50 μm. (B) Quantitative analysis of ROS levels in murine ELGs from NC, SiNPs-treated, and SD + SiNPs-treated groups. Mean fluorescence intensity (MFI) values of ROS were log-transformed using the natural logarithm (ln) to improve normality before statistical analysis. Statistical analysis was performed using Brown-Forsythe ANOVA. Multiple comparisons were conducted using the Games-Howell post hoc test. For NC vs. SiNPs, P = 0.002. For NC vs. SD + SiNPs, P < 0.0001. For SiNPs vs. SD + SiNPs, P < 0.0001. **P < 0.01, ***P < 0.001. (C) Diurnal changes of ROS levels in murine ELGs from NC, SiNPs-treated, and SD + SiNPs-treated groups. Statistical analyses were performed on natural log-transformed data. For the NC group, F = 19.65, P < 0.0001. For the SiNPs-treated group, F = 3.794, P = 0.0265. For the SD + SiNPs-treated group, F = 1.957, P = 0.1499. **P < 0.01, ***P < 0.001. (D) Quantitative analysis of MDA levels in murine ELGs from the NC, SiNPs-treated, and SD + SiNPs-treated groups. Statistical analysis was performed using Brown-Forsythe ANOVA. Multiple comparisons were conducted using the Games-Howell post hoc test. For NC vs. SiNPs, P < 0.001. For NC vs. SD + SiNPs, P < 0.001. For SiNPs vs. SD + SiNPs, P < 0.001. ***P < 0.001. (E) Diurnal changes of MDA levels in murine ELGs from the NC, SiNPs, SD + SiNPs. Statistical analyses were performed on natural log-transformed data. For the NC group, F = 57.601, P < 0.001. For the SiNPs-treated group, F = 4.732, P = 0.012. For the SD + SiNPs-treated group, F = 2.559, P = 0.084. ***P < 0.001. (F) Representative images showing γ-H2AX levels at ZT0, ZT6, ZT12, and ZT18 in murine ELGs from the NC, SiNPs-treated, and SD + SiNPs-treated groups. Scale bar: 20 μm. (G) Quantitative analysis of γ-H2AX levels in murine ELGs from the NC, SiNPs-treated, and SD + SiNPs-treated groups. Statistical analysis was performed using the Kruskal–Wallis test (non-parametric), followed by Dunn’s post hoc test for multiple comparisons. For NC vs. SiNPs, P = 0.8374. For NC vs. SD + SiNPs, P < 0.0001. For SiNPs vs. SD + SiNPs, P = 0.0006. ***P < 0.001. NS: not significant. (H) Diurnal changes of γ-H2AX levels in murine ELGs from the NC, SiNPs-treated, and SD + SiNPs-treated groups. For the NC group, F = 32.35, P < 0.0001. For the SiNPs-treated group, F = 7.167, P = 0.0017. For the SD + SiNPs-treated group, F = 20.02, P < 0.0001. ***P < 0.001. NS, not significant
To further evaluate the impact of oxidative stress on DNA damage, immunohistochemical staining for γ-H2AX, a marker of DNA double-strand breaks, was conducted. The SD + SiNPs-treated group exhibited significantly higher γ-H2AX-positive staining compared to the NC group, indicating that ROS accumulation contributed to DNA damage in the ELGs (Figs. 11F-G). Moreover, the oscillatory rhythm of γ-H2AX levels was disrupted by SiNPs treatment and totally changed by SD + SiNPs treatment (Fig. 11H).
Taken together, these findings demonstrate that SiNPs exposure induces oxidative stress in the lacrimal glands, characterized by increased ROS accumulation, lipid peroxidation, and DNA damage. Importantly, SD significantly amplifies this oxidative injury and disrupts circadian redox homeostasis, likely contributing to the structural and functional decline of ELGs.
3.11 Activation of NLRP3 inflammasome triggered by sleep deprivation in SiNPs-treated mice
The NLRP3 inflammasome plays a critical role in inflammation and oxidative stress. To investigate the involvement of the ROS/NLRP3 pathway in SiNPs-induced dry eye exacerbated by SD, we evaluated the expression patterns of NLRP3 and its adaptor protein ASC in ELGs. Immunohistochemical staining of NLRP3+ cells was illustrated in Fig. 12A. Analysis of NLRP3 expression revealed that SiNPs treatment alone did not significantly affect NLRP3 production in ELGs, while a marked increase was observed when SiNPs were combined with SD (Fig. 12B), suggesting a possible synergistic effect between SiNPs treatment and SD treatment. Figure 12C showed the oscillation of NLRP3+ cells in the NC, the SiNPs-treated, and the SD + SiNPs-treated groups. SiNPs treatment altered the daily oscillation of NLRP3 compared to the NC group, with both groups peaking at ZT12, but exhibiting different trough values (ZT6 in the NC group vs. ZT18 in the SiNPs group). SD treatment further disrupted the daily rhythm of NLRP3, with the peak occurring at ZT6 and the trough at ZT18. Additionally, ASC levels were significantly elevated in the SiNPs-treated group compared to the NC group (Figs. 12E-I), with its oscillatory rhythm maintained (Fig. 12D). SD further increased ASC levels and disrupted its daily rhythmicity.
SD potentiates SiNPs-induced NLRP3 inflammasome activation in mouse ELGs. (A) Representative immunohistochemical images of NLRP3+ cells in mouse ELGs at ZT0 and ZT12 for the NC, SiNPs-treated, and SD + SiNPs-treated groups. Scale bar: 20 µm. (B) Average abundance of NLRP3+ cells of ELGs from NC, SiNPs-treated, and SD + SiNPs-treated groups. Statistical analysis was performed using the Kruskal–Wallis test (non-parametric), followed by Dunn’s post hoc test for multiple comparisons. Dunn’ s test showed that the SD + SiNPs-treated group differed significantly from the NC group (***P < 0.001) and the SiNPs-treated group (*P < 0.05), while the NC group, and the SiNPs-treated group were not significantly different (NS). (C) Diurnal variation analysis of NLRP3+ cell ratio in murine ELGs across the NC, SiNPs-treated group, and the SD + SiNPs-treated groups. Statistical analysis showed no significant diurnal variations in any group (ZT0: F = 1.419, P = 0.2665; ZT6: F = 19.05, P < 0.001; ZT12: F = 1.227, P = 0.3134; ZT18: F = 17.68, P < 0.001). ***P < 0.001. NS: not significant. (D) Diurnal variation analysis of ASC+ cells ratio in murine ELGs across NC group, SiNPs-treated group, and SD + SiNPs-treated group. For ZT0, F = 11.50, P = 0.0005; For ZT6, F = 10.98, P = 0.0008; For ZT12, F = 1.227, P = 0.005; For ZT18, F = 17.68, P < 0.0001. **P < 0.01, ***P < 0.001. (E) Representative immunofluorescence images of ASC+ cells in mouse ELGs at ZT0, ZT6, ZT12, and ZT18 time points for the NC group, the SiNPs-treated group, and the SD + SiNPs-treated group. Scale bar: 50 μm. (F-I) Quantitative analysis of average fluorescence signal intensity of ASC in mouse ELGs at ZT0 (G), ZT6 (H), ZT12 (I), and ZT18 (J) for the NC group, SiNPs-treated group, and SD + SiNPs-treated group. Statistical analysis revealed significant differences at all time points: ZT0 (F = 11.50, P = 0.0005), ZT6 (F = 10.98, P = 0.0008), ZT12 (F = 1.227, P = 0.005), and ZT18 (F = 17.68, P < 0.0001). **P < 0.01, ***P < 0.001
Taken together, these findings suggest that SiNPs trigger a mild activation of the NLRP3 inflammasome, which SD amplifies while deranging the rhythmic expression of NLRP3 and ASC, presumably via ROS build-up and immune imbalance.
