International Journal of Molecular Sciences

Article Comparative Analysis of Age-Related Changes in Lacrimal Glands and Meibomian Glands of a C57BL/6 Male Mouse Model

1,2,3 2 4, , 1,2,3, , Chang Ho Yoon , Jin Suk Ryu , Ho Sik Hwang * † and Mee Kum Kim * † 1 Department of Ophthalmology, Seoul National University College of Medicine, Seoul 03080, Korea; [email protected] 2 Laboratory of Ocular Regenerative Medicine and Immunology, Seoul Artificial Eye Center, Seoul National University Hospital Biomedical Research Institute, Seoul 03080, Korea; [email protected] 3 Department of Ophthalmology, Seoul National University Hospital, Seoul 03080, Korea 4 Department of Ophthalmology, College of Medicine, The Catholic University of Korea, Seoul 07345, Korea * Correspondence: [email protected] (H.S.H.); [email protected] (M.K.K.); Tel.: +82-2-3779-1025 (H.S.H.); +82-2-2072-2665 (M.K.K.); Fax: +82-2-761-6869 (H.S.H.); +82-2-741-3187 (M.K.K.) These authors contributed equally to this work. †


Received: 29 April 2020; Accepted: 2 June 2020; Published: 11 June 2020


Abstract: It is not known how biological changes in the lacrimal (LGs) and meibomian (MGs) glands contribute to dry eye disease (DED) in a time-dependent manner. In this study, we investigated time-sequenced changes in the inflammation, oxidative stress, and senescence of stem cells in both glands of an aging-related DED mouse model. Eight-week (8W)-, one-year (1Y)-, and two-year (2Y)-old C57BL/6 male mice were used. MG areas of the upper and lower eyelids were analyzed by transillumination meibography imaging. The number of CD45+, 8-OHdG+, Ki-67+, and BrdU+ cells was compared in both glands. Increased corneal staining and decreased tear secretion were observed in aged mice. The MG dropout area increased with aging, and the age-adjusted MG area in lower lids was negatively correlated with the National Eye Institute (NEI) score. Increased CD4+ interferon (IFN)-γ+ cells in LGs were found in both aged mice. An increase in 8-OHdG+ cells in both glands was evident in 2Y-old mice. Reduced Ki-67+ cells, but no change in CD45+ cells, was observed in the MGs of 1Y-old mice. Increased BrdU+ cells were observed in the LGs of aged mice. This suggests that age-dependent DED in C57BL/6 mice is related to inflammation of the LGs, the development of MG atrophy, and oxidative stress in both glands.

Keywords: dry eye; aging; lacrimal glands; meibomian glands; inflammation; oxidative stress; senescence

1. Introduction Aging is a biologically complex process associated with cellular senescence and dysregulation of the immune, endocrine, and metabolic systems, resulting in multiple organ pathologies [1,2]. Chronic oxidative stress affects the aging process by modifying interactions of these systems [2]. Given that dry eye prevalence increases with age [3], the aging process appears to significantly affect dry eye disease (DED). The aging-dependent changes in the ocular surface with DED are related to biological changes in the lacrimal glands (LGs) and meibomian glands (MGs). Although the pathogenesis of aging-associated DED has not yet been fully understood, recent studies have shown that autoimmune changes or oxidative stress in the LGs and acinar cell atrophies in the MGs occur with aging [4–8].

Int. J. Mol. Sci. 2020, 21, 4169; doi:10.3390/ijms21114169 www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2020, 21, 4169 2 of 19

However, it is not known how biological changes in both the LGs and MGs contribute to DED in Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 2 of 20 a time-dependent manner. With regard to the senescence of stem cells, oxidative stress, and gland inflammation,However, it has notit is not been known determined how biological which changes process in both contributes the LGs and to theMGs profound contribute ocularto DED surface changesin with a time aging-dependent when manner. the two With glands regard are to compared the senescence to each of stem other cells, over oxidative time. stress Therefore,, and gland we aimed to investigateinflammation the time-sequenced, it has not been determined changes which in the process inflammation, contributes to oxidative the profound stress, ocular and surface senescence of progenitorchanges cells with inaging both when the the LGs two and glands MGs, are compared as well asto each compare other over the time. changes Therefore, between we aimed them in an to investigate the time-sequenced changes in the inflammation, oxidative stress, and senescence of age-relatedprogenitor DED C57BLcells in /both6 mouse the LGs model. and MGs, as well as compare the changes between them in an age- related DED C57BL/6 mouse model. 2. Results 2. Results 2.1. DED Was Established in Aged C57BL/6 Male Mice 2.1. DED was Established in Aged C57BL/6 Male Mice Corneal punctate epithelial erosion was higher in the aged (1-year (1Y)-old and 2-year (2Y)-old) mice than inCorneal eight-week punctate (8W)-old epithelial mice. erosion The was mean higher ( SD) in the corneal aged (1 staining-year (1Y) scores-old and of 2 8W-,-year 1Y-,(2Y)- andold) 2Y-old mice than in eight-week (8W)-old mice. The mean± (± SD) corneal staining scores of 8W-, 1Y-, and 2Y- mice were 1.88 1.28, 6.25 2.01, and 5.95 2.85, respectively. The scores of the 1Y- and 2Y-old mice old mice± were 1.88 ± 1.28,± 6.25 ± 2.01, and± 5.95 ± 2.85, respectively. The scores of the 1Y- and 2Y-old groups weremice groups significantly were significantly higher than higher those than ofthose the of 8W-old the 8W-old mice mice (all (allp p< <0.0001; 0.0001; Kruskal Kruskal–Wallis–Wallis test followedtest by followed Dunn’s by post Dunn’s hoc post test); hoc however, test); however, no di ffnoerence difference was was observed observed between between 1Y 1Y-- and and 2Y 2Y-old-old mice (Figure1miceA,B). (Figure Tear secretion 1A,B). Tear was secretion not di ffwaserent not amongdifferent the among groups; the groups; however, however, when thewhen body the body weight was adjusted,weight the normalized was adjusted, value the normalized was significantly value was significantly lower in aged lower mice in aged than mice in than 8W-old in 8W mice-old mice (Figure 1C; all p < 0.0001;(Figure one-way 1C; all p < ANOVA 0.0001; one followed-way ANOVA by Tukey’s followed post by Tukey’s hoc test). post Periodic hoc test). acid–Schi Periodic acidff (PAS)–Schiff staining (PAS) staining of the conjunctiva showed that the number of goblet cells was significantly lower in of the conjunctiva showed that the number of goblet cells was significantly lower in 1Y-old mice than 1Y-old mice than in 8W-old mice (Figure 1D; p = 0.043; one-way ANOVA followed by Tukey’s post in 8W-oldhoc mice test). (FigureHowever,1D; thep number= 0.043; of one-waygoblet cells ANOVA in 2Y-old followed mice was bynot Tukey’sdifferent postfrom the hoc number test).However, in the number8W- or of 1Y goblet-old mice. cells Given in 2Y-old these miceresults, was DED not was di fullyfferent established from the at numberone year of in age 8W- in orC57BL/6 1Y-old mice. Given thesemale results,mice. DED was fully established at one year of age in C57BL/6 male mice.

Figure 1.Figure(A) Representative1. (A) Representative images images of of corneal corneal fluorescein fluorescein staining. staining. (B) (BThe) The National National Eye Institute Eye Institute (NEI) corneal(NEI) corneal staining staining score score was was significantly significantly higher higher in one one-year-year (1Y (1Y)-)- and and two two-year-year (2Y)-old (2Y)-old mice mice than in eight-week (8W)-old mice (all p < 0.0001; Kruskal–Wallis test followed by Dunn’s post hoc than in eight-week (8W)-old mice (all p < 0.0001; Kruskal–Wallis test followed by Dunn’s post hoc test; n = 26, 12, and 22 for 8W-, 1Y-, and 2Y-old mice, respectively). (C) Tear secretion did not change n C test; =over26, 12,time and (Kruskal 22 for–Wallis 8W-, test) 1Y-,. However, and 2Y-old when mice, the level respectively). was normalized ( )Tear by body secretion weight did(BW), not the change over timenormalized (Kruskal–Wallis value was test).lower in However, 1Y- and 2Y when-old mice the than level in was8W-old normalized mice (all p by< 0.0001; body o weightne-way (BW), the normalizedANOVA value followed was by lower Tukey’s in 1Y-post and hoc 2Y-old test; n = mice 36, 22, than and in 22 8W-old for 8W mice-, 1Y- (all, andp < 2Y0.0001;-old mice, one-way ANOVArespectively followed by). ( Tukey’sD) Representative post hoc photographs test; n = 36, of 22, periodic and 22 acid for– 8W-,Schiff 1Y-, (PAS) and staining 2Y-old of mice, the inferior respectively). (D) Representativeconjunctival photographsfornix (sagittal ofsection; periodic 200× acid–Schi). The numberff (PAS) of goblet staining cells was of the significantly inferior conjunctival lower in 1Y- fornix (sagittalold section; mice than 200 in ).8W The-old mice number (p = 0 of.043; goblet one-way cells ANOVA was significantlyfollowed by Tukey’s lower post in 1Y-oldhoc test; micen = 12, than in × 8W-old mice (p = 0.043; one-way ANOVA followed by Tukey’s post hoc test; n = 12, 6, and 3 for 8W-, 1Y-, and 2Y-old mice, respectively). Ns, not significant; * p < 0.05 and **** p < 0.0001. Data are presented as means standard error. ± Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 3 of 20

6, and 3 for 8W-, 1Y-, and 2Y-old mice, respectively). Ns, not significant; *p < 0.05 and ****p < 0.0001. Int. J. Mol.Data Sci. are2020 presented, 21, 4169 as means ± standard error. 3 of 19

2.2. MG Dropout Area Increased with Aging, and Age-Adjusted Lower MG Area was Negatively Correlated 2.2. MG Dropout Area Increased with Aging, and Age-Adjusted Lower MG Area Was Negatively Correlated withwith thethe SeveritySeverity ofof CornealCorneal EpithelialEpithelial ErosionErosion Transillumination meibography meibography showed showed less less dense dense MG MG in in aged aged mice mice (1Y- (1Y and- and 2Y-old 2Y-old mice) mice than) than in 8W-oldin 8W-old mice mice (Figure (Figure2A). 2A). The The MG areasMG areas in 2Y-old in 2Y mice-old weremice largerwere larger than those than inthose 8W-old in 8W mice-old ( pmice= 0.004; (p = Kruskal–Wallis0.004; Kruskal– testWallis followed test followed by Dunn’s by Dunn’s post hoc post test). hoc The test). MG The areas MG of areas the lower of the eyelid lower in eyelid 1Y-old in mice 1Y- wereold mice larger were than larger those than in 8W-old those micein 8W (p-old= 0.021; mice one-way (p = 0.021; ANOVA one-way followed ANOVA by Tukey’sfollowed post by hocTukey’s test). Thepost MGhoc dropouttest). The areas MG increaseddropout areas with increased aging. Upper with MGaging dropout. Upper areas MG dropout in 1Y- and areas 2Y-old in 1Y mice- and were 2Y- largerold mice than were those larger in 8W-old than those mice in (p 8W= 0.023-old mice and p(p= = 0.02,0.023 respectively; and p = 0.02,Kruskal–Wallis respectively; Kruska test followedl–Wallis bytest Dunn’s followed post byhoc Dunn’s test) post and lowerhoc test) MG and dropout lower MG areas dropout in 2Y-old areas mice in were2Y-old larger mice than were those larger in than 8W- andthose 1Y-old in 8W mice- and ( p1Y<-0.0001old mice and (p p< =0.00010.029, and respectively; p = 0.029, one-wayrespectively; ANOVA one-way followed ANOVA by Tukey’s followed post by hocTukey’s test). post Interestingly, hoc test). theInterestingly, lower MG the area lower was negativelyMG area was correlated negatively with correlated the corneal with staining the corneal score afterstaining age scoreadjustment after age (r = adjustment0.346; p = (0.034;r = −0 partial.346; p correlation= 0.034; partial analysis; correlation Figure2 analysis;C). However, Figure upper 2C). − MGHowever, areas andupper upper MG areas MG dropout and upper areas MG did dropout not show areas any did correlation not show any to the correlation corneal stainingto the corneal score afterstaining age sc adjustmentore after age in thisadjustment mouse model in this ( rmouse= 0.213, modelp = (0.200r = −0.213, and r p= = 00.195,.200 andp = r 0.240,= −0.195, respectively; p = 0.240, − − partialrespectively; correlation partial analysis). correlation analysis).

Figure 2. 2. (A(A)) Representative Representative transillumination transillumination meibography meibography images. images. (B) Meibomian (B) Meibomian gland gland (MG) (areasMG) ofareas the of upper the upper eyelid eyelid in 2Y-old in 2Y mice-old (micep = 0.004; (p = 0 Kruskal–Wallis.004; Kruskal–Wallis test followed test followed by Dunn’s by Dunn’s post hocpost test)hoc andtest)lower and lower eyelid eyelid in 1Y-old in 1Y mice-old (micep = 0.021; (p = 0.021; one-way one- ANOVAway ANOVA followed followed by Tukey’s by Tukey’s post hoc post test) hoc were test) largerwere larger than thosethan those in 8W-old in 8W mice.-old mice. MG dropoutMG dropout areas areas of upper of upper eyelids eyelids were were larger larger in 1Y- in 1Y and- and 2Y-old 2Y- miceold mice than than 8W-old 8W mice-old (micep = 0.023 (p = and0.023p = and0.02, p = respectively; 0.02, respectively; Kruskal–Wallis Kruskal– testWallis followed test followed by Dunn’s by postDunn’s hoc post test). hoc MG test). dropout MG areasdropout of lower areas eyelidsof lower were eyelids larger were in 2Y-old larger mice in 2Y than-old 8W- mice and than 1Y-old 8W- mice and (1Yp <-old0.0001 mice and (p

2.3. Aging-Dependent Inflammation was not Evident, but Decreased Potential of Differentiation was Int. J. Mol. Sci.Suspected2020, 21 ,in 4169 the MGs 4 of 19 MG dropout tended to increase with age (Figure 3A). The distribution of neutral lipids in the MGs did not seem to differ significantly with age (Figures 3B and 4). It was found that PPARγ, which known to regulateis known meibocyte to regulate meibocytedifferentiation differentiation and lipid and synthesis, lipid synthesis, was distributedwas distributed in in both both the the cytoplasm and nucleuscytoplasm in 8W-old and nucleus mice, butin 8W confined-old mice, tobut theconfined nucleus to the in nucleus 2Y-old in mice2Y-old (Figure mice (Figure4). The 4). The number of Ki-67+ cellsnumber in the of MGs Ki-67 as+ cells a marker in the MGs for as proliferating a marker for proliferating cells was significantlycells was significantly decreased decreased in 1Y-old in mice compared1Y to- 8W-oldold mice micecompared (Figure to 8W3C;-old p mice= 0.036 (Figure for 3C; the p = upper 0.036 for eyelids the upper by the eyelids Mann–Whitney by the Mann– test and + Whitney test and 0.035 for the lower eyelids by the independent t-test). The number+ of CD45 cells in 0.035 for thethe lowerMGs as eyelids a marker by for the immune independent cells was nott-test). increased The in number1Y-old mice of relative CD45 tocells 8W-old in themice MGs as a marker for(Figure immune 3D). cells was not increased in 1Y-old mice relative to 8W-old mice (Figure3D).

Figure 3. (AFigure) Representative 3. (A) Representative images images ofcoronal of coronal sections sections of of the the MGs MGs stained stained with hematoxylin with hematoxylin-eosin-eosin (H&E). The(H&E). area betweenThe area between the dotted the dotted lines lines indicates indicates suspicious suspicious MG MG dropout dropout (200×). (200 (B) Representative). (B) Representative × images of coronalimages of sections coronal ofsections the MGs of the stained MGs stained with withLipidTOX LipidTOX and and PPAR PPARγγimmunofluorescence immunofluorescence (100 ). (100×). (C) The expression of Ki67+ cells in both MGs of the upper and lower eyelids was significantly + × (C) The expressionreduced in of 1Y Ki67-old micecells compared in both to MGs 8W-old of mice the upper(p = 0.036 and by the lower Mann eyelids–Whitney was test significantly and 0.035 by reduced in 1Y-old micethe independent compared t-test, to respectively 8W-old mice; 200×) (.p (D=) The0.036 expression by the ofMann–Whitney CD45+ cells in the MGs test of andthe lower 0.035 by the independenteyelidt-test, was respectively; not different between 200 ). 8W (-D and) The 1Y-old expression mice (p = of0.571; CD45 Mann+ –cellsWhitney in the test; MGs 200×). of All the lower × eyelid wasimages not di ffwereerent obtained between from 8W- the coronal and 1Y-old section. mice n = 5 ( pand= 0.571;3 for 8W Mann–Whitney- and 1Y-old mice, test;respectivel 200 y.). * Allp images < 0.05. × were obtainedInt. J. Mol. Sci. from 2020, the21, x coronalFOR PEER section.REVIEW n = 5 and 3 for 8W- and 1Y-old mice, respectively. *5 pof <20 0.05.

Figure 4. RepresentativeFigure 4. Representative images images of coronal of coronal sections sections of of the the MGs MGs stained stained with LipidTOX with LipidTOX and PPAR andγ PPARγ immunofluorescenceimmunofluorescence (coronal (coronal section; section; 400 400×).). Lipid Lipid droplets showed showed a similar a similar distribution distribution for 8W- and for 8W- and 2Y-old mice. PPARγ staining shows cytoplasmic× and nuclear localization in 8W-old mice. In contrast, γ 2Y-old mice.PPARγ PPAR stainingstaining shows predominant shows cytoplasmic nuclear localization and nuclear in 2Y-old localization mice. in 8W-old mice. In contrast, PPARγ staining shows predominant nuclear localization in 2Y-old mice. 2.4. Aging-Dependent Inflammation was Evident in the LGs Cross-sections of the LGs showed ductal dilatation and periductal infiltration of inflammatory cells in 1Y- and 2Y-old mice (Figure 5A,B). The focus score of the LGs sequentially increased with aging and was significantly higher in 1Y- and 2Y-old mice than in 8W-old mice (Figure 5A–C; p = 0.010 and < 0.0001, respectively; Kruskal–Wallis test followed by Dunn’s post hoc test). The weight of the LGs in 2Y-old mice was significantly higher than that of 8W-old mice (p = 0.003; one-way ANOVA followed by Tukey’s post hoc test). However, when the weight of the LGs was normalized by body weight, the LG density in aged mice (1Y- and 2Y-old) was significantly lower than that of 8W-old mice (Figure 5D; p = 0.002 and 0.001, respectively; one-way ANOVA followed by Tukey’s post hoc test). Flow cytometry analysis of the LGs revealed that the percentage of CD4+ interferon (IFN)-γ+ cells and CD4+ IL-17A+ cells increased in 1Y-old mice compared to 8W-mice (all p < 0.0001; one-way ANOVA followed by Tukey’s post hoc test), but decreased in 2Y-old mice (p = 0.0015 and 0.0004, respectively; one-way ANOVA followed by Tukey’s post hoc test) (Figure 5E). When an independent t-test was performed on only 8W- and 2Y-old mice, the percentages of CD4+ IFNγ+ cells and CD4+ IL-17A+ cells for 8W- and 2Y-old mice were significantly different (p = 0.010 and p < 0.001, respectively), although they were not significantly different in a post hoc test (p = 0.785 and 0.975, respectively).

Int. J. Mol. Sci. 2020, 21, 4169 5 of 19

2.4. Aging-Dependent Inflammation Was Evident in the LGs Cross-sections of the LGs showed ductal dilatation and periductal infiltration of inflammatory cells in 1Y- and 2Y-old mice (Figure5A,B). The focus score of the LGs sequentially increased with aging and was significantly higher in 1Y- and 2Y-old mice than in 8W-old mice (Figure5A–C; p = 0.010 and <0.0001, respectively; Kruskal–Wallis test followed by Dunn’s post hoc test). The weight of the LGs in 2Y-old mice was significantly higher than that of 8W-old mice (p = 0.003; one-way ANOVA followed by Tukey’s post hoc test). However, when the weight of the LGs was normalized by body weight, the LG density in aged mice (1Y- and 2Y-old) was significantly lower than that of 8W-old mice (Figure5D; p = 0.002 and 0.001, respectively; one-way ANOVA followed by Tukey’s post hoc test). Flow cytometry analysis of the LGs revealed that the percentage of CD4+ interferon (IFN)-γ+ cells and CD4+ IL-17A+ cells increased in 1Y-old mice compared to 8W-mice (all p < 0.0001; one-way ANOVA followed by Tukey’s post hoc test), but decreased in 2Y-old mice (p = 0.0015 and 0.0004, respectively; one-way ANOVA followed by Tukey’s post hoc test) (Figure5E). When an independent t-test was performed on only 8W- and 2Y-old mice, the percentages of CD4+ IFNγ+ cells and CD4+ IL-17A+ cells for 8W- and 2Y-old mice were significantly different (p = 0.010 and p < 0.001, respectively), although they were not significantly different in a post hoc test (p = 0.785 and 0.975, respectively). Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 6 of 20

Figure 5. Representative images of the lacrimal gland (LG) sections acquired by (A) H&E staining Figure 5. Representative images(50×) and (B) of CD45 the immunofluorescence lacrimal staininggland (100× (LG) and 400 sections×). Arrows indicate acquired lymphocytic by (A) H&E staining (50 ) infiltration foci. Red boxes of subfigure A are enlarged in subfigure B. (C) The inflammation focus × and (B) CD45 immunofluorescencescore was higher in 1Y staining- and 2Y-old mice (100 than in 8Wand-old mice 400 (p = 0)..010 Arrowsand p < 0.0001, indicaterespectively; lymphocytic infiltration Kruskal–Wallis test followed by Dunn’s post ×hoc test; n = 22, 6, and× 11 for 8W-, 1Y-, and 2Y-old mice, foci. Red boxes of subfigurerespectively A are). (D enlarged) LG weight was inhigher subfigure in 2Y-old mice than B. in (C 8W)-old The mice inflammation (p = 0.003; one-way focus score was higher ANOVA followed by Tukey’s post hoc test); however, the body weight (BW)-adjusted LG weight was in 1Y- and 2Y-old mice thanlower inin 2Y- and 8W-old 1Y-old mice than mice in 8W- (oldp mice= (0.010p = 0.002 and and 0.001, respectively;p < 0.0001, one-way AN respectively;OVA Kruskal–Wallis followed by Tukey’s post hoc test; n = 8, 6, and 6 for 8W-, 1Y-, and 2Y-old mice, respectively). (E) After test followed by Dunn’sforward post-scatter hoc and test; side-scattern = gating22, in 6, theand LGs, CD4 11+ IFNγ+ for and8W-, CD4+ 1Y-,IL-17A+ andcells were 2Y-old mice, respectively). identified. The percentages of CD4+ IFNγ+ cells were higher in 1Y-old mice than in 8W- and 2Y-old (D) LG weight was highermice in (p2Y-old < 0.0001 and micep = 0.0015, than respectively in; 8W-oldone-way ANOVA mice followed (p by= Tukey’s0.003; post hoc one-way test). ANOVA followed by The percentages of CD4+ IL-17A+ cells were also higher in 1Y-old mice than in 8W- and 2Y-old mice Tukey’s post hoc test); however,(p < 0.0001 and the p = body 0.0004; one weight-way ANOVA (BW)-adjusted followed by Tukey’s post LG hoc test). weight Although was the lower in 2Y- and 1Y-old mice than in 8W-old micepercentage (p =s of0.002 CD4+ IFNγ and+ cells and 0.001, CD4+ IL-17A respectively;+ cells of 8W- and 2Y-old one-way mice were not significantly ANOVA followed by Tukey’s different in a post hoc test (p = 0.785 and 0.975, respectively), there was a significant difference (p = post hoc test; n = 8, 6, and0.010 6 and for p < 8W-,0.001, respectively) 1Y-, and when an 2Y-old independent mice, t-test was performed respectively). on only 8W- and (2YE-old) After forward-scatter and mice. (n = 13, 6, and 3 for 8W-, 1Y-, and 2Y-old mice, respectively.) *p < 0.05, **p < 0.01, and ****p < side-scatter gating in the0.0001 LGs, for CD4 the post+ hocIFN test, andγ+ †p and = 0.010 and CD4 ††p + < 0.001IL-17A for the independent+ cells t- weretest. Data identified. are The percentages presented as means ± standard error. of CD4+ IFNγ+ cells were higher in 1Y-old mice than in 8W- and 2Y-old mice (p < 0.0001 and p =

0.0015, respectively; one-way ANOVA followed by Tukey’s post hoc test). The percentages of CD4+

IL-17A+ cells were also higher in 1Y-old mice than in 8W- and 2Y-old mice (p < 0.0001 and p = 0.0004; one-way ANOVA followed by Tukey’s post hoc test). Although the percentages of CD4+ IFNγ+ cells and CD4+ IL-17A+ cells of 8W- and 2Y-old mice were not significantly different in a post hoc test (p = 0.785 and 0.975, respectively), there was a significant difference (p = 0.010 and p < 0.001, respectively) when an independent t-test was performed on only 8W- and 2Y-old mice. (n = 13, 6, and 3 for 8W-, 1Y-,

and 2Y-old mice, respectively.) * p < 0.05, ** p < 0.01, and **** p < 0.0001 for the post hoc test, and † p = 0.010 and p < 0.001 for the independent t-test. Data are presented as means standard error. †† ± Int. J. Mol. Sci. 2020, 21, 4169 6 of 19

Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 7 of 20 2.5. Effector T Cells and Activated Antigen-Presenting Cells Were Increased in the Drainage Lymph Nodes of Aged2.5. MiceEffector T Cells and Activated Antigen-Presenting Cells were Increased in the Drainage Lymph Nodes of Aged Mice Flow cytometry of drainage lymph nodes (DLNs) showed that the percentage of CD8+ IL-17A+, + + CD4+ IFN-Flow γcytometry+, and CD4 of drainage+ IL-17A lymph+ T cells nodes was (DLNs) higher showed in 2Y-old that the mice percentage than in of 8W-CD8 andIL-17 1Y-oldA , mice, CD4+ IFN-γ + , and CD4+ IL-17A+ T cells was higher in 2Y-old mice than in 8W- and 1Y-old mice, with with no significant differences between 8W- and 1Y-old mice (Figure6A; p < 0.05; one-way ANOVA). no significant differences between 8W- and 1Y-old mice (Figure 6A; p < 0.05; one-way ANOVA). The + + γ+ Thepercentage percentage of CD4 of CD4+ and andCD8 CD8+ IFNγ+IFN T cellsT was cells also was increased also increased in 1Y-old in 1Y-oldmice compared mice compared to 8W-old to 8W-old + + + micemice (Figure (Figure6 6A;A; pp << 0.05;0.05; one one-way-way ANOVA ANOVA).). The percentages The percentages of the CD11b of the+ CD86 CD11b+ and CD11bCD86 + majorand CD11b + + majorhistocompatibility histocompatibility complex complex (MHC)II (MHC)II+ Ly6C+ cellsLy6C werecells significantly were significantly increased increased in 2Y-old in mice2Y-old mice comparedcompared to to 8W- 8W- andand 1Y 1Y-old-old mice mice (p ( p< <0.05;0.05; one one-way-way ANOVA ANOVA),), but equal but equalfor 8W for- and 8W- 1Y and-old mice 1Y-old mice (Figure(Figure6B). 6B).

FigureFigure 6. 6. Flow cytometry cytometry for for immune immune cells cellsin the in drainage the drainage lymph lymphnodes (DLNs) nodes. ( (DLNs).A) In DLNs, (A the) In DLNs, thepercentage percentage of CD8 of+ CD8 IFNγ++ IFN T cellsγ+ wasT cells higher was in 1Y higher-old mice in 1Y-old than in mice8W-old than mice in (p 8W-old = 0.032; Kruskal mice (p– = 0.032; Kruskal–WallisWallis test followed test followed by Dunn’s by Dunn’spost hoc posttest) hoc. The test). percentages The percentages of CD8+ IL of-17 CD8A+, +CD4IL-17A+ IFN+-,γ CD4+, and+ IFN-γ+, + + andCD4 CD4IL+-17AIL-17A T cells+ T cellsin 2Y in-old 2Y-old mice micewere weresignificantly significantly higher higher than in than the other in the groups other groups(one-way (one-way + + + + + ANOVA).ANOVA) (.B (B)) The The percentagespercentages of CD11b CD11b +CD86CD86 and+ and CD11b CD11b MHCII+ MHCII Ly6C+ Ly6C cells were+ cells significantly were significantly higher in 2Y-old mice than in the other groups in the DLNs (one-way ANOVA). Red boxes indicate higher in 2Y-old mice than in the other groups in the DLNs (one-way ANOVA). Red boxes indicate CD11b+ MHCII+ Ly6C+ cells. n = 15, 8, and 3 for 8W-, 1Y-, and 2Y-old mice, respectively. *p < 0.05 and CD11b+ MHCII+ Ly6C+ cells. n = 15, 8, and 3 for 8W-, 1Y-, and 2Y-old mice, respectively. * p < 0.05 and ****p < 0.0001. Data are presented as means ± standard error. **** p < 0.0001. Data are presented as means standard error. ± 2.6. Oxidative Stress Increased in both Glands of Aged Mice 2.6. Oxidative Stress Increased in Both Glands of Aged Mice The number of 8-OHdG+ cells/section was higher in both glands of 1Y-old mice than in those of + 8WThe-old numbermice (Figure of 8-OHdG 7A; p = 0.018cells in/ thesection MGswas and higher0.015 in inthe both LGs; glands independent of 1Y-old t-test). mice In addition, than in those of 8W-oldthe percentage mice (Figure of 8-OHdG7A; p hi= propidium0.018 in the iodide MGs (PI) andlo cells 0.015 was in significantly the LGs; independent higher in botht-test). glands In of addition, hi lo the2Y percentage-old mice compared of 8-OHdG to thosepropidium of 8W- and iodide 1Y-old (PI) mice cells (Figure was 7B; significantly p = 0.009 and higher 0.007 in in MGs, both glands ofrespectively, 2Y-old mice and compared p = 0.008 a tond those0.033 in of LGs, 8W- respectively; and 1Y-old one mice-way (Figure ANOVA7B; followedp = 0.009 by Tukey’s and 0.007 post in MGs, respectively,hoc test). There and werep = no0.008 differences and 0.033 between in LGs, 8W- and respectively; 1Y-old mice one-way in either gland. ANOVA Surprisingly, followed when by Tukey’s post hoc test). There were no differences between 8W- and 1Y-old mice in either gland. Surprisingly, when the oxidative stress of glands was compared at a given time point, the LGs always showed Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 8 of 20 Int. J. Mol. Sci. 2020, 21, 4169 7 of 19 the oxidative stress of glands was compared at a given time point, the LGs always showed a higher apercentage higher percentage of 8-OHdG ofhi 8-OHdG PIlo cellshi thanPIlo thecells MGs than, regardless the MGs, of regardless age (Figure of age7B; p (Figure = 0.0247 B;in 8Wp =-old0.024 mice in 8W-oldand p = mice0.004 andin 1Yp -=old0.004 mice; in paired 1Y-old- mice;t test). paired- t test).

Figure 7. (A) Immunofluorescence staining of 8-OHdG in the MGs (sagittal section, upper panel) and Figure 7. (A) Immunofluorescence staining of 8-OHdG in the MGs (sagittal section, upper panel) and LGs (lower panel; 400 ). The number of 8-OHdG+ cells was significantly higher in 1Y-old mice (n = 5) LGs (lower panel; 400××). The number of 8-OHdG+ cells was significantly higher in 1Y-old mice (n = 5) than in 8W-old mice (n = 3) in both the MGs and LGs (p = 0.018 and 0.015, respectively; independent than in 8W-old mice (n = 3) in both the MGs and LGs (p = 0.018 and 0.015, respectively; independent t-test). (B) The percentages of 8-OHdGhi PIlo cells were significantly higher in the LGs than in the t-test). (B) The percentages of 8-OHdGhi PIlo cells were significantly higher in the LGs than in the MGs MGs in 8W- and 1Y-old mice (p = 0.024 and 0.004, respectively; independent t-test). In both the MGs in 8W- and 1Y-old mice (p = 0.024 and 0.004, respectively; independent t-test). In both the MGs and and LGs, 8-OHdGhi PIlo cells were significantly higher in 2Y-old mice than in 8W- and 1Y-old mice LGs, 8-OHdGhi PIlo cells were significantly higher in 2Y-old mice than in 8W- and 1Y-old mice (one- (one-way ANOVA; n = 9, 5, and 3 for 8W-, 1Y-, and 2Y-old mice, respectively). Red boxes indicate way ANOVA; n = 9, 5, and 3 for 8W-, 1Y-, and 2Y-old mice, respectively). Red boxes indicate 8-OHdGhi 8-OHdGhi PIlo cells. * p < 0.05 and ** p < 0.01. Data are presented as means standard error. PIlo cells. *p < 0.05 and **p < 0.01. Data are presented as means ± standard error.± 2.7. Senescence of Stem Cells Was Not Evident in Both Glands 2.7. Senescence of Stem Cells was not Evident in Both Glands As label-retaining progenitors or S-phase cells, the percentage of BrdU+ cells during the cell cycle was notAs dilabelfferent-retaining in the MGs,progenitors depending or S- onphase age. cells, Interestingly, the percentage the percentage of BrdU+ ofcells BrdU during+ cells the in cell the cycle LGs waswas significantlynot different higherin the inMGs aged, depending mice (1Y- on and age. 2Y-old Interestingly, mice) than the in 8W-oldpercentage mice of (Figure BrdU+8 ;cellsp = in0.014 the andLGs 0.017,was significantly respectively; higher Kruskal–Wallis in aged mice test (1Y followed- and 2Y by-old Dunn’s mice post) than hoc in test).8W-old When mice comparing (Figure 8; thep = glands,0.014 and the 0.017, LGs respectively tended to have; Kruskal a higher–Wallis percentage test followed of BrdU by+ Dunn’scells than post the hoc MGs, test). irrespective When comparing of age (thep = glands,0.004 in the 8W-old LGs mice;tended Wilcoxon to have a signed-rank higher percentage test; p < of0.001 BrdU in+1Y-old cells than mice; the paired MGs,t -test).irrespective of age (p = 0.004 in 8W-old mice; Wilcoxon signed-rank test; p < 0.001 in 1Y-old mice; paired t-test).

Int. J. Mol. Sci. 2020, 21, 4169 8 of 19 Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 9 of 20

Figure 8.8. BrdUBrdU/7/7-AAD-AAD cell cycle analysis by flowflow cytometry. We excluded the apoptotic cells andand + + gated 2n to 4n 4n BrdU BrdU+ cellscells by by 7- 7-AADAAD gating. gating. The The percentage percentage of BrdU of BrdU+ (2n (2nto 4n) to 4n)cells cells in the in MGs the MGs was wasnot different not different among among the thegroups groups (one (one-way-way ANOVA; ANOVA; n =n 9=, 9,5, 5,and and 3 3for for 8W 8W-,-, 1Y 1Y-,-, and and 2Y 2Y-old-old mice, + respectively);respectively); however, however, the the percentage percentage of of BrdU BrdU+ cellscells in the in theLGs LGs was was significantly significantly higher higher in aged in agedmice mice(1Y- and (1Y- and2Y-old 2Y-old mice mice)) than than in 8W in 8W-old-old mice mice (p (=p =0.0140.014 and and 0.017, 0.017, respectively; respectively; Kruskal Kruskal–Wallis–Wallis test followed by Dunn’s post hoc test; n = 9, 5, and 3 for 8W-, 1Y-, and 2Y-old mice, respectively). Red boxes followed by Dunn’s post hoc test; n = 9, 5, and 3 for 8W-, 1Y-, and 2Y-old mice, respectively). Red indicate 2n to 4n BrdU+ cells. * p < 0.05. Data are presented as means standard error. boxes indicate 2n to 4n BrdU+ cells. *p < 0.05. Data are presented as means± ± standard error. 3. Discussion 3. Discussion Our study showed that (1) dry eye and related ocular surface damage developed after one year of agingOur instudy a C57BL showed/6 male that mouse (1) dry model, eye and (2) related inflammation ocular surface of the LGsdamage was developed evident after after one one year year of agingof aging with in aa consistentC57BL/6 male increase mouse of themodel, focus (2) score inflammation and CD4 + ofIFN theγ +LGscells, was (3) evident atrophy after of the one MGs year was of + + evidentaging with after a consistent two years increase of aging, of (4) the oxidative focus score stress and was CD significantly4 IFNγ cells increased, (3) atrophy in both of the glands MGs after was twoevident years after of aging,two years and (5)of aging, senescence (4) oxidative of the stem stress cells was was significantly not evident increased in either gland.in both This glands suggests after thattwo aging-dependentyears of aging, and dry (5) eye senescence may be related of the tostem the cells inflammation was not evident of LG, the in either development gland. This of atrophy suggests of MG,that aging and oxidative-dependent stress dry in eye both may glands be related in this to mouse the inflammation model. The pathogenesisof LG, the development of age-related of atrophy dry eye isof aMG complex, and oxidative process. Itstress is well-known in both glands that in inflammation, this mouse model. senescence The pathogenesis of the stem cells, of age atrophy-related of dry the functionaleye is a complex cells, andprocess. oxidative It is well stress-known in the that LGs inflammation, or MGs contribute senescence to the pathogenesisof the stem cells, of age-related atrophy of DEDthe functional [6]. However, cells, it and is not oxidative known howstress these in the factors LGs aorff ectMGs age-related contribute DED to the sequentially pathogenesis in the of aging age- process.related DED Therefore, [6]. However, our study it isis worthy not known of notice how and these may factors shed lightaffect on age the-related pathogenesis DED sequentially of age-related in drythe aging eye by process. showing Therefore, how the LGsour andstudy MGs is worthy deteriorate of notice over and time may with shed regard light to on inflammation, the pathogenesis stem cellof age properties,-related anddry oxidativeeye by showing stress. how the LGs and MGs deteriorate over time with regard to inflammation,First, we evaluatedstem cell properties, age-related and changes oxidative in the stress. LGs. Given that tear and protein secretion are energy-requiringFirst, we evaluated active procedures,age-related thechanges LGs appearin the LGs. to be Given highly that susceptible tear and toprotein pathological secretion aging. are Emergingenergy-requiring evidence active suggests procedures, that the the LGs LGs undergo appear inflammation,to be highly susceptible atrophy, fibrosis,to pathological or ductal aging. cell proliferationEmerging evidence with aging suggests in humans, that the as LGs well undergo as rodents inflammation, [6,9–11]. Aging atrophy, impairs fibrosis the LGs, or by ductal altering cell multipleproliferation mechanisms, with aging such in humans as neuroendocrine,, as well as rodents immunological, [6,9–11]. Aging apoptotic, impairs protein the LGs secretion, by altering and oxidativemultiple mechanisms, stress pathways such [11 as]. Chronological neuroendocrine, changes immunological, in the LGs apoptotic, were reported protein in mice secretio withn, mast and oxidative stress pathways [11]. Chronological changes in the LGs were reported in mice with mast cell infiltration, decreased peroxidase secretion, and lipofuscin accumulation at 8 to 12 months of age,cell infiltration, followed by decreased T and B peroxidase lymphocytic secretion infiltration,, and decreasedlipofuscin accumulation innervation,decreased at 8 to 12 months acetylcholine of age, followed by T and B lymphocytic infiltration, decreased innervation, decreased acetylcholine release, release, and structural changes up to 24 months of age [5,11,12]. Given the post-mortem changes in the LGsand structural in humans, changes C57BL /up6 mice to 24 can months be used of age as a[5,11,12 model]. to Given evaluate the post age-related-mortem DED, changes since in C57BL the LGs/6 in humans, C57BL/6 mice can be used as a model to evaluate age-related DED, since C57BL/6 mice mice develop age-related dacryoadenitis, MG dysfunction, and corneal staining [5]. Interestingly, develop age-related dacryoadenitis, MG dysfunction, and corneal staining [5]. Interestingly, an an aging-related predominance of IFN-γ-secreting CD4+IFNγ+ or CD4+IL17+ cells in the LGs may aging-related predominance of IFN-γ-secreting CD4+IFNγ+ or CD4+IL17+ cells in the LGs may depend depend on the sex of the rodent [5,12]. This study corroborated the previous findings on age-related on the sex of the rodent [5,12]. This study corroborated the previous findings on age-related inflammatory changes in the LGs with a predominance of CD4+IFNγ+ cells in male mice. The focus inflammatory changes in the LGs with a predominance of CD4+IFNγ+ cells in male mice. The focus score in the LGs continuously increased with time, and the level of CD4+IFNγ+ cells or CD4+IL17+ was score in the LGs continuously increased with time, and the level of CD4+IFNγ+ cells or CD4+IL17+ was higher in the LGs for up to 2 years compared to 8W-old mice, although the level of both cells decreased higher in the LGs for up to 2 years compared to 8W-old mice, although the level of both cells in 2Y-old mice when compared with 1Y-old mice. Meanwhile, the DLNs of 2Y-old mice showed decreased in 2Y-old mice when compared with 1Y-old mice. Meanwhile, the DLNs of 2Y-old mice an increase in CD4+IFNγ+ and CD4+IL17+ cell populations, along with an increased population of showed an increase in CD4+IFNγ+ and CD4+IL17+ cell populations, along with an increased population of CD8+IL17+ cells, suggesting that both IFN-γ- and IL17-secreting cells are involved in

Int. J. Mol. Sci. 2020, 21, 4169 9 of 19

CD8+IL17+ cells, suggesting that both IFN-γ- and IL17-secreting cells are involved in age-related systemic autoimmune features, unlike the LGs. This finding may present different age-related changes of IL-17+ or IFNγ+ cells, depending on the location. A previous report also showed that the percentage of IFNγ+ cells increased across all tissues (conjunctiva, LG, and lymph nodes) at 24 months, while, on the contrary, the percentage of IL-17+ cells decreased in the LG and conjunctiva, unlike in the lymph nodes, at 24 months [4]. A healthy human study has demonstrated that age deeply impacts the contraction of CD4+IL-17+ cell populations in peripheral blood, while the population of CD4+IFNγ+ cells are relatively maintained, regardless of age [13]. In addition, the contraction rate differs depending on age, exhibiting a temporary increase of the percentage of CD4+IL-17+ cells in blood from those aged 35 to 49-years old, a decrease of the percentage of CD4+IL-17+ cells in the aged group (50 to 65-years old), and an increase of the percentage of CD4+IL-17+ cells again in the further-aged group (>65-years old) [14]. However, the secreting ability of IL-17 after in vitro stimulation with phytohemagglutinin continuously decreased, depending on age, owing to cellular senescence [14]. Given that the survival of immune cells may be affected differentially by the surrounding environment where the resident tissue cells provide survival factors, the shifting of IL-17+ cell or IFN γ+ cell changes from 1 year of age to 2 years of age is presumably affected by the (1) pathogenic increase of autoimmunity, (2) age-dependent cellular senescence, or (3) different survival rates based on where they reside. The changes in effector T cells were accompanied by increased populations of activated antigen-presenting cells in the DLNs, suggesting that an increased presence of the antigen is crucial to age-related dry eye, as previously reported [4]. High levels of IFN-γ in the conjunctiva are known to impact the loss of goblet cells [5,12]. Our study showed a decreased goblet cell density in 1Y-old mice, along with an increase in CD4+IFNγ+ cells in LG. However, there was no significant difference in the goblet cell density in 2Y-old mice compared to 8W-old mice, although the level of CD4+IFNγ+ cells in LG was still higher than that in young mice. We did not evaluate conjunctival Th1 cells directly, and this may have caused the difference between our results and those presented in a previous report [12]. Given that the relationship between goblet cell loss and aging remains controversial, depending on the detection method employed for goblet cells in humans [15,16], aging-related goblet cell changes should be further investigated. On the other hand, oxidative damage may be a key player in the pathogenesis of age-related and non-age-related dry eye [8,17,18], and our flow cytometric data were consistent with previous studies showing that 8-OHdG+ cells significantly increased in both the LGs and MGs of 2Y-old mice. This may support the study suggesting that calorie restriction can be a treatment option for age-related dry eye by reducing oxidative stress [19]. In 1Y-old mice, an inconsistency was observed between the histological findings and flow cytometric data. This occurrence might have been caused by an equivocal increase in 8-OHdG+ cells or by the semi-quantitative counting of histological sections with limited areas. Taken together, significant oxidative stress was evident in both the LGs and MGs after two years of age. Interestingly, the population of BrdU+ cells (label-retaining progenitor or S-phase cells) tended to increase over time in the LGs, while no significant changes in BrdU+ cells were observed over time in the MGs. When we gated out the CD45+ BrdU+ cells to remove a portion + of the immune cells (AppendixA Figure A1A), the population of CD45 − BrdU cells still tended to increase. Given that BrdU+ label-retaining or S-phase cells can be observed in acini, duct, and myoepithelial cells, this change might have been caused by ductal or myoepithelial cell proliferation that occurs in aged mice or compensatory proliferative responses against inflammatory damaging signals in the LGs (AppendixA Figure A1B) [ 6,20]. One study reported that the BrdU+ label-retaining acinar cell population was 11.98 cells/mm2 in the LGs following a 2-week chase period after 7 days of consecutive BrdU injections [20]. We observed flow cytometric data. Therefore, a direct comparison with the histological data presented in the report mentioned above was not possible. The interpretation and data shape of the BrdU+ cells may differ, depending on the BrdU injection time. The longer the chase period after BrdU injection, the larger the population of slow-cycling progenitor cells that can be observed relative to the population of transient amplifying cells. In the <24 h interval between the observation and BrdU injection, the majority of BrdU+ cells indicated S-phase cells when gated together Int. J. Mol. Sci. 2020, 21, 4169 10 of 19

+ withInt. J. 7-aminoactinomycinMol. Sci. 2020, 21, x FOR PEER D (7-AAD). REVIEW After two weeks of the chase period in our study, the BrdU11 of 20 cells included both progenitors and proliferating S-phase cells. Thus far, senescence of the progenitor cellswas wasnot evident not evident in the in LGs. the LGs.However, However, the cells the could cells couldnot be notfurther be further discriminated discriminated into acinar, into acinar,ductal, ductal,or myoepithelial or myoepithelial cells. cells.A further A further investigation investigation using using acinar acinar-specific-specific or oracinar acinar-excluding-excluding markers markers is ispending pending to to confirm confirm whether whether senescence senescence of of the the acinar acinar progenitor cells isis significantlysignificantly increasedincreased inin agedaged LGs.LGs. Second,Second, we we assessedassessed age-relatedage-related morphological morphological changes changes in in the the MGs. MGs. ForFor highlyhighly qualifiedqualified imagesimages ofof dropout,dropout, infraredinfrared transilluminated images images of of the the MGs MGs were were taken taken after after each each eyelid eyelid was was excised excised en enbloc. bloc. A wide A wide spectrum spectrum of white of white light light was was used used as the as light the light source, source, and andan infrared an infrared transmitting transmitting filter filterwas attached was attached in front in of front the ofobjective the objective lens to lensobtain to infrared obtain infrared images. images.As a result, As compared a result, compared to 8W-old tomice, 8W-old 1Y- and mice, 2Y 1Y--old and mice 2Y-old showed mice definite showed MG definite dropout MG as dropoutatrophy asin the atrophy lower in eyelid the lower (Figure eyelid 2A). (FigureTo quantify2A). To the quantify degree of the MG degree atrophy, of MG the atrophy,area of the the MGs area in of the the central MGs in 3 mm the centraleyelid was 3 mm measured eyelid was(Figure measured 9). The (Figure increasing9). The tendency increasing of MG tendency areas ofmay MG be areas related may to be growth related according to growth to according the aging toprocess. the aging Meanwhile, process. Meanwhile,lower MG dropout lower MG areas dropout increased areas sharply increased between sharply 1 and between 2 years 1 and(Figure 2 years 2B). (FigureThis indicates2B). This that indicates the lower that MGs the lowerseem MGsto be more seem toaffected be more by atheffected aging by process the aging than process the upper than MGs. the upperThe result MGs. corroborates The result corroborates previous human previous studies human suggest studiesing increased suggesting dropout increased of dropout the lower of MGs the lowerrelative MGs to the relative upper to MGs the upper [21–24 MGs]. In [ 21addition,–24]. In we addition, found wea negative found a correlation negative correlation between the between lower theMG lower area MGand areaocular and surface ocular staining, surface staining,which is whichalso a novel is also finding a novel and finding supports and supports clinically clinically relevant relevantdry eye drychanges eye changeswith age. with Further age. Furtherstudies to studies confirm to confirmthis outcome this outcome are in development. are in development.

Figure 9. Ex vivo transillumination meibography images of the central 3-mm area of interest were Figure 9. Ex vivo transillumination meibography images of the central 3-mm area of interest were analyzed. “The standardized MG outline” was defined with an imaginary line starting at the furthest analyzed. “The standardized MG outline” was defined with an imaginary line starting at the furthest endpoint of the MG and parallel to the eyelid margin. The area of the MG, including the eyelid margin, endpoint of the MG and parallel to the eyelid margin. The area of the MG, including the eyelid margin, was defined as “the MG area”, and the area where the MG does not exist inside the standardized MG was defined as “the MG area”, and the area where the MG does not exist inside the standardized MG outline was defined as “the dropout area”. outline was defined as “the dropout area”. To the best of our knowledge, this is the first study to quantify MG atrophy using mouse To the best of our knowledge, this is the first study to quantify MG atrophy using mouse meibography. If we measure the MG area in a histological section of the eyelid, the MG area will meibography. If we measure the MG area in a histological section of the eyelid, the MG area will vary, depending on the position of the section. Therefore, it seems best to measure the MG area with vary, depending on the position of the section. Therefore, it seems best to measure the MG area with meibography, just like in humans. In human studies, the dropout area can be directly measured by meibography, just like in humans. In human studies, the dropout area can be directly measured by meibography after eversion of the eyelid [25,26]. However, in this study, black pigmentation of the meibography after eversion of the eyelid [25,26]. However, in this study, black pigmentation of the C57BL/6 mice still blocked the distal MG, even in the infrared image. Therefore, this ex vivo study C57BL/6 mice still blocked the distal MG, even in the infrared image. Therefore, this ex vivo study requires an arbitrary reference line, such as the “standardized MG outline”. Although the quantification requires an arbitrary reference line, such as the “standardized MG outline”. Although the quantification of dropout is quite a difficult process to achieve using mouse meibography, the dropout area could be measured inside the standardized MG outline. This study corroborated a human study by showing that atrophy increased with age by non-contact meibography [22,27]. It

Int. J. Mol. Sci. 2020, 21, 4169 11 of 19 of dropout is quite a difficult process to achieve using mouse meibography, the dropout area could be measured inside the standardized MG outline. This study corroborated a human study by showing that atrophy increased with age by non-contact meibography [22,27]. It appears that atrophic changes affect the MGs slowly over time, and inflammatory changes affect the LGs at around one year. Ki-67 is a nuclear antigen that is expressed during the cell cycle, present at stage G1, and used as a proliferation marker [28]. In this study, Ki-67 expression was lower in the acini of the MGs of 1Y-old mice compared to 8W-old mice (Figure3C), which was consistent with a previous study showing that many Ki-67+ basal nuclei were present in the acini of MGs from young mice, but reduced in 1Y- to 2Y-old mice [28]. PPARγ is a member of the closely homologous genes ubiquitously expressed in various tissues and the major subtype expressed in sebocytes, where it regulates the expression of genes involved in lipogenesis [29,30]. In this study, the nuclear staining of PPARγ was similar in meibocytes between young and old mice, but the distribution of PPARγ staining in the cytoplasm was only noted in young mice (Figure4) and correlated well with previous studies [ 28,31]. In summary, morphological atrophy increased over time with a decrease in the proliferation and differentiation of meibocytes. These changes were very similar in humans [7,31,32], which is a finding supported by this study. In addition, we compared time-dependent inflammatory, oxidative, and proliferative capacity changes between the LGs and MGs. Regarding inflammation, the LGs presented more inflammation over time than the MGs based on the histological findings with CD45 staining (Figures3 and5) or by comparative gating with CD45+ cells in flow cytometry (AppendixA Figure A1). Notably, the 8-OHdGhi PIlo cells in the LGs were significantly higher than in the MGs of both young mice and 1Y-old mice, suggesting that the LGs are susceptible to oxidative damage. This statistical difference between both glands disappeared after two years of age, at which point the population of 8-OHdG+ cells was equally high in both glands. The BrdU+ cell assay was used to detect label-retaining progenitors or S-phase cells. The senescence of these cells was not evident in both the LGs and MGs up to two years. There is an inconsistency between the histological counting with KI-67 staining and flow cytometric data with BrdU cells employed to evaluate the proliferative capacity in MGs. Although KI-67+ expression was decreased in histological sections at one year, the population of BrdU+ cells in flow cytometry was not different over the first two years. Different detection methods and the usage of different biomarkers may have caused the inconsistency. KI-67+ cells indicate proliferating cells, such as transient amplifying cells, while BrdU+ cells after the chase period indicate progenitors, as well as proliferating cells. Therefore, the proliferative capacity may be reduced, but the population of progenitors appears to be equal in the MGs. On the contrary, the LGs did not show any reduction in proliferative capacity, even with an exclusion of CD45+ BrdU+ cells (Figure8 and AppendixA Figure A1). In summary, aging-related inflammation a ffects the LGs more than the MGs, and aging-related oxidative stress affects the LGs more than the MGs in the early period, but equally in later periods. Aging-related stem cell senescence was not evident in both the LGs and MGs. Our study had several limitations. As small numbers of animals were used owing to the limited budget, certain analyses could not include all 8W-, 1Y-, and 2Y-old mice (AppendixA Tables A1 and A2) and the analyses with obvious differences were not statistically significant (e.g., the percentage of BrdU+ cells in LG between 1-year-old and 2-year-old mice). In addition, it might be better to include data from 6-month-old mice. Given that structural changes or innervation changes of LG were not evident until 8 to 12-months old [11] and that the expression of either PPAR-γ or Ki67 of MG in 2-month-old mice was similar to that in 6-month-old mice [28], we only analyzed 8-week-old mice as representatives of young mice, and 1-year-old and 2-year-old mice. Second, although C57BL/6 mice are a valuable tool for assessing age-related dry eye, meibographic illustration in these mice is challenging due to dense pigmentation of the lid. Third, in our study, only male mice were used. Given that the female sex contributes differently to the phenotypes of age-related dry eye due to hormonal influences, our data on the MGs and LGs are different from those that would be obtained from female mice. It is known that goblet cell loss and corneal barrier disruption are less common in males than Int. J. Mol. Sci. 2020, 21, 4169 12 of 19 in females, and T cell infiltration in the LGs is greater in males than in females [12]. Inconsistencies in goblet cell data compared with previous studies may also be influenced by sex. Likewise, later atrophic changes in the MGs may be influenced by the male sex and hormone levels. Therefore, our data should be interpreted with a consideration of sex. In addition to CD4+ cells, CD20+ cells (B cells) play an important role in the pathogenesis of Sjögren’s syndrome [33]. In this study, we did not evaluate CD20+ cells. Further evaluation will be needed in future studies. Although PAS was only applied for goblet cell identification, based on the previous study’s protocol [12,34], additional nuclear counterstaining may be a better option for goblet cell identification. Finally, we used whole eyelid tissue for meibomian gland analysis with flow cytometry. We did our best to remove the skin and eyebrows, but the tissue may have still included skin epithelial cells or stromal fibroblasts because the specific markers for meibocytes are not yet known. Nevertheless, to the best of our knowledge, this is the first report to (1) present comparative data on age-related changes in the LGs and MGs regarding dry eye and (2) quantify MG atrophy using mouse meibography. These findings may be helpful in understanding the pathophysiology of age-related dry eye.

4. Materials and Methods The experimental protocol was approved on 29 May 2017 by the Institutional Animal Care and Use Committee of the Seoul National University Biomedical Research Institute (IACUC no. 17-0093-C1A1). Animal experiments were performed in accordance with the ARVO Statement for Use of Animals in Ophthalmic Vision and Research.

4.1. Animals A total of 56 C57BL/6 male mice were used in this study. Twenty-nine 8W (young), sixteen 1Y (middle aged), and eleven 2Y (elderly) mice were included. The mice were bred in a specific pathogen-free facility at the Biomedical Research Institute of Seoul National University Hospital (Seoul, Korea), maintained at 22–24 C with 55 5% relative humidity, and given free access to food and water. ◦ ± The number of samples (eyes, LGs, and eyelids) used in each experiment and list of groups included in each experimental analysis are summarized in AppendixA Tables A1 and A2, respectively.

4.2. Clinical Evaluation The corneal staining score and tear secretion test were performed under anesthesia (using a mixture of zoletil and xylazine at a ratio of 1:3). The corneal staining scores were blindly assigned by a single experienced ophthalmologist (CH. Y.) according to the National Eye Institute (NEI) scoring scheme. After a drop of 0.25% fluorescein dye was applied to the conjunctival sac for 30 s, the ocular surface was gently washed with 1 mL of normal saline, and corneal staining was observed using a microscope (Olympus SZ61; Olympus Corporation, Tokyo, Japan) under cobalt blue illumination. For the tear secretion test, phenol red-impregnated cotton threads (FCI Ophthalmics, Pembroke, MA, USA) were inserted into the lateral canthus of mice for 60 s. The amount of secreted tears was determined by measuring the length of the wet thread in millimeters. The weights of LG were measured and divided by the body weight to adjust the aging-related confounding factor based on a previous study protocol [5,35–37], and tear secretion was also divided by body weight before comparison [5,38].

4.3. Histopathology The whole eyeball and LGs were excised and fixed in formalin. The samples were sliced into 4 µm sections and subjected to hematoxylin-eosin (H&E), PAS, and immunohistochemical staining for CD45, Ki-67, LipidTOX, PPAR-γ, and 8-OHdG. The number of goblet cells was calculated as the sum of PAS-stained cells in the upper and lower eyelids under a microscope using a 200 objective × (8W, n = 12; 1Y, n = 6; 2Y, n = 3). The inflammatory focus score in the LGs was measured by counting CD45+ inflammatory cells in a blinded manner (>50 inflammatory cells/focus = 1, 25–50 inflammatory Int. J. Mol. Sci. 2020, 21, 4169 13 of 19 cells/focus = 0.5) (8W, n = 22; 1Y, n = 6; 2Y, n = 11) [39]. The number of CD45+, Ki-67+, and 8-OHdG+ cells was counted in the medium-(200 ; CD45+ and Ki-67+) or high-power fields (400 ; 8-OHdG+) × × of two to three different sections from the same animal and averaged as cell counts/section. Each staining protocol was adapted from the method used in previous studies [32,40–42]. The following primary anti-mouse antibodies were used: CD45 (ab10558; Abcam, Cambridge, MA, USA), Ki-67 (Cat no. 14-5698-82; eBioscience, San Diego, CA, USA), LipidTOX (H34475; Invitrogen, Carlsbad, CA, USA), PPAR-γ (Ab59256; Abcam, Cambridge, MA, USA), and 8-OHdG (GTX41980; GeneTex, Irvine, TX, USA).

4.4. Transillumination Meibography of Mice After euthanasia, both upper and lower eyelids were gently excised (8W, n = 27; 1Y, n = 6; 2Y, n = 16). The skin and muscles were removed from the excised eyelids as much as possible. The eyelids were carefully placed between two standard microscopic slides, and the slide was then placed over a wide spectrum white light source (LED plate), which was modified from the study of Reyes et al. [43]. The light passing through the eyelids was captured using a microscope (Olympus SZ61; Olympus Corporation) equipped with an infrared filter (Hoya R72; TOKINA Co. Ltd., Tokyo, Japan) and a Charge-Coupled Device (CCD) camera (acA1600-20 um; Basler Inc., Ahrensburg, Germany). A region of interest was identified in the central 3 mm area and analyzed using ImageJ software (Version 1.52a; NIH Image, Bethesda, MD, USA). The MG area, including the eyelid margin, was semi-automatically selected using the Wand tool (tolerance range of 5 to 30) of ImageJ (NIH Image). After defining the farthest point from the eyelid margin within the MG area, an imaginary line parallel to the eyelid margin was drawn at that point. As this ex vivo measurement method in mice is different from the in vivo method in humans, “the standardized MG outline” was defined with an imaginary line starting at the furthest endpoint of the MG and parallel to the eyelid margin. The area of the MG, including the eyelid margin, was defined as “the MG area” and the area where the MG does not exist inside the standardized MG outline was defined as “the dropout area” (Figure9).

4.5. 5-Bromo-20-Deoxyuridine (BrdU) Injection and Cell Cycle Analysis For cell cycle analysis, we performed BrdU/7-AAD analysis with the FITC BrdU Flow Kit (Cat no. 559619; BD Pharmingen, San Diego, CA, USA). BrdU is a thymidine analogue, and 7-AAD binds to double stranded DNA by intercalating between base pairs in G-C-rich regions [44]. With this combination, flow cytometric analysis enables the enumeration and characterization of cells that are actively synthesizing DNA (BrdU incorporation) in terms of their cell cycle position (i.e., apoptotic, G0/1, S, or G2/M phase) [45,46]. To evaluate the proportion of progenitor cells in the MGs and LGs, mice (8W, n = 9; 1Y, n = 5; 2Y, n = 3) were intraperitoneally injected with 50 µg BrdU (51-2420KC; BD Pharmingen) per g body weight twice a day 14 and 13 days before sacrifice (four injections in total). After cell acquisition by forward and side scattering (FSC-A and SSC-A, respectively), we excluded the apoptotic cells by 7-AAD gating and gated progenitors (2n) or present proliferating cells synthesizing DNA in the S-phase of the cell cycle (2n to 4n BrdU+ cells) [47]. The cells were assayed using a FACSCanto flow cytometer (BD BioSciences, Mountain View, CA, USA). Data were analyzed using Flowjo software (version 10.6.1; Tree Star, Ashland, OR, USA).

4.6. Flow Cytometry for Immune Cells and Oxidatively Stressed Cells The LGs (number of samples: 8W, n = 18; 1Y, n = 11; 2Y, n = 3), and cervical lymph nodes as DLN (number of mice: 8W, n = 15; 1Y, n = 8; 2Y, n = 3), were extracted and collected. Bilateral cervical lymph nodes were pooled for analysis. To obtain the MGs, the skin was peeled off and the eyebrows were removed as much as possible after excision of the upper and lower eyelids. For MG analysis, upper and lower samples of each eye were pooled (8W, n = 9; 1Y, n = 5; 2Y, n = 3). The proportion of IFN-γ and IL-17A in CD4+ and CD8+ T cells was assessed as a subset of effector T cells in the LGs and DLNs [48]. Int. J. Mol. Sci. 2020, 21, 4169 14 of 19

CD11bhiCD86hi cells were gated as activated antigen-presenting cells, and MHCII hiCD11bhiLy6Chi cells were gated as activated antigen-presenting monocytic cells in DLNs [39]. 8-OHdGhi/PIlo cells were gated as oxidatively stressed cells in both MGs and LGs based on the methods described in previous studies [49]. To obtain cell suspensions, the collected LGs and DLNs and eyelids were minced between the frosted ends of two glass slides in media containing RPMI (WelGENE, Daegu, Korea), 10% fetal bovine serum, and 1% penicillin-streptomycin. For intracellular IFN-γ and IL-17A staining, the cell suspensions were stimulated for 5 h with 50 ng/mL phorbol myristate acetate and 1 µg/mL ionomycin in the presence of GolgiPlug (BD Pharmingen). The cells were stained with the following antibodies: CD4-PE-Cy7 (Cat no. 25-0041; eBioscience), CD8-PerCP-Cy5.5 (Cat no. 45-0081; eBioscience), IFN-γ-FITC (Cat no. 11-7311; eBioscience), and IL17A-APC (Cat no. 17-7177; eBioscience). For the antigen-presenting cells and oxidatively stressed cells, cell suspensions were collected and incubated for 30 min at 4 ◦C with fluorescein-conjugated anti-mouse antibodies as follows: CD11b-PE-Cy7 (Cat no. 25-0112; eBioscience), CD86-APC (Cat no.17-0862; eBioscience), MHC class II - FITC (Cat no. 11-5321; eBioscience), Ly6c–PerCp-Cy5.5 (Cat no. 45-5932, eBioscience), 8-OHdG (GTX41980; GeneTex), and PI (Cat no.51-66211E; BD Pharmingen). The cells were assayed using a FACSCanto flow cytometer (BD BioSciences). Data were analyzed using Flowjo software (version 10.6.1; Tree Star).

4.7. Statistical Analysis The Shapiro–Wilk test was used to test the normality of the data. Appropriate parametric (independent or paired t-test) or non-parametric (Mann–Whitney or Wilcoxon signed-rank test) statistical tests were used to compare the two groups. A one-way ANOVA followed by Tukey’s post hoc test or a Kruskal–Wallis test followed by Dunn’s post hoc test was used for a comparison of more than two groups. A partial correlation analysis was conducted for the MG parameters and corneal staining score after the removal of confounding age factors. Statistical analyses were performed using GraphPad Prism software (version 8.2.0; GraphPad Software, La Jolla, CA, USA). Differences were considered statistically significant at p < 0.05.

5. Conclusions This study compared aging-dependent changes between the MGs and LGs in male C57BL/6 mice and it suggests that the inflammation of LG, the development of atrophy of MG, and oxidative stress in both glands may be involved in aging-dependent dry eye in this mouse model.

Author Contributions: Conceptualization, M.K.K.; methodology, M.K.K., H.S.H. and C.H.Y.; software, C.H.Y. and H.S.H.; validation, M.K.K., H.S.H. and C.H.Y.; formal analysis, M.K.K., H.S.H., J.S.R. and C.H.Y.; investigation, M.K.K., H.S.H., J.S.R. and C.H.Y.; resources, M.K.K. and H.S.H.; data curation, J.S.R. and C.H.Y.; writing—original draft preparation, C.H.Y.; writing—review and editing, M.K.K. and H.S.H.; visualization, C.H.Y.; supervision, M.K.K. and H.S.H.; project administration, M.K.K.; funding acquisition, M.K.K. and H.S.H. All authors have read and agreed to the published version of the manuscript. Funding: This research was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (No. NRF-2017R1A2B2007209) and a grant from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI) funded by the Ministry of Health & Welfare, Republic of Korea (No. HI17C0659). Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. Int. J. Mol. Sci. 2020, 21, 4169 15 of 19

Abbreviations

1Y One-year-old 2Y Two-year-old 7-AAD 7-aminoactinomycin D 8W Eight-week-old BrdU Bromo-20-deoxyuridine BW Body weight CCD Charge-Coupled Device DED Dry eye disease

DLNInt. J. Mol. Sci. 20 Drainage20, 21, x FOR PEER lymph REVIEW node 16 of 20 H&E Hematoxylin-eosin IFNCCD InterferonCharge-Coupled Device DED Dry eye disease LGDLN LacrimalDrainage lymph gland node MHCH&E MajorHematoxylin histocompatibility-eosin complex MGIFN MeibomianInterferon gland LG Lacrimal gland NEIMHC NationalMajor histocompatibility Eye Institute complex PASMG PeriodicMeibomian acid-Schi gland ff PINEI PropidiumNational Eye Institute iodide PAS Periodic acid-Schiff PI Propidium iodide Appendix A Appendix

+ Figure A1. (A)Figure CD45 A1./BrdU (A) CD45 gating/BrdU gating by flow by flow cytometry. cytometry. The The percentage percentage of CD45+ ofcells CD45 in the LGscells was in the LGs was significantly highersignificantly in 1Y-oldhigher in mice1Y-old mice than than in in 8W-old 8W-old mice mice (p = (0p.044;= 0.044;one-way one-wayANOVA followed ANOVA by followed by Tukey’s post hoc test; n = 9, 5, and 3 for 8W-, 1Y-, and 2Y-old mice, respectively). The percentage of Tukey’s post hocCD45 test;+ cellsn in =the9, LGs 5, was and also 3 significantly for 8W-, higher 1Y-, andthan that 2Y-old in the MGs mice,, irrespective respectively). of age (p = 0.019 The percentage of CD45+ cells inin the 8W LGs-old mice was and also 0.0005 significantly in 1Y-old mice, respectively; higher than paired that t-test; in n the= 5 per MGs, group). irrespective The percentage of age (p = 0.019 - + in 8W-old miceof and CD45 0.0005 BrdU cells in in 1Y-old the MGs mice, was not respectively; different between paired 8W- and t1Y-test;-old micen = (p5 = 0 per.483); group). however, The percentage the percentage of CD45- BrdU+ cells in the LGs was significantly higher in 1Y-old mice than in 8W- + of CD45− BrdUold cellsmice (pin < 0.0001; the MGs independent was nott-test;di n =ff 5erent per group). between (B) Representative 8W- and images 1Y-old of the mice LG sections (p = 0.483); however, by Ki-67 immunofluorescence+ staining (400×). Arrows indicate inflammatory foci. The expression of the percentage of CD45− BrdU cells in the LGs was significantly higher in 1Y-old mice than in 8W-old mice (p < 0.0001; independent t-test; n = 5 per group). (B) Representative images of the LG sections by Ki-67 immunofluorescence staining (400 ). Arrows indicate inflammatory foci. The expression × of Ki67+ cells in both the acinar and inflammatory cell infiltrating areas tends to increase with aging. * p < 0.05, *** p < 0.001, and **** p < 0.0001. Data are presented as means standard error. ± Int. J. Mol. Sci. 2020, 21, 4169 16 of 19

Table A1. Number of samples in each experiment. The eyes, LGs, and eyelids for MGs that were used in each experiment were taken from 58 samples of twenty-nine 8-week-old mice, 32 samples from sixteen 1-year-old mice, and 22 samples from eleven 2-year-old mice.

Number of Experiment Samples 8W 1Y 2Y Corneal staining (eyes) 26 12 22 Dry eye Tear secretion (eyes) 36 22 22 assessment Conjunctival goblet cell count (eyes) 12 6 3 Gross MG area analyzed by meibography 27 6 16 examination LG weight 8 6 6 Histology H&E staining (coronal section) 8 6 0 H&E staining, and LipidTOX/PPARγ 10 0 8 IHC (coronal section) MG analysis Ki-67 and CD45 IHC (coronal section) 5 3 0 8-OHdG IHC (sagittal section) 5 3 0 Flow cytometry * 8-OHdGhiPIlo and BrdU+ 4 0 3 8-OHdGhiPIlo, BrdU+, and CD45+ 5 5 0 Histology H&E staining and CD45 IHC 9 0 3 H&E staining, CD45 IHC, and Ki-67 IHC 9 3 8 H&E staining, CD45 IHC, and 8-OHdG 4 3 0 IHC LG analysis 8-OHdG IHC 1 0 0 Flow cytometry CD4+ IFNγ+ and CD4+ IL-17A+ 9 6 0 8-OHdGhiPIlo, BrdU+, and CD45+ 4 5 0 CD4+ IFN-γ+, CD4+ IL-17A+, 5 0 3 8-OHdGhiPIlo, BrdU+, and CD45+ Secondary lymphoid Flow cytometry of DLN † (mice) 15 8 3 organs

* Upper and lower samples of each eye were pooled and analyzed. † Bilateral cervical lymph nodes were pooled for analysis. Abbreviations: W, week-old; Y, year-old; MG, meibomian gland; LG, lacrimal gland; H&E, hematoxylin-eosin; IHC, immunohistochemical staining; DLN, drainage lymph node.

Table A2. List of groups included in each experimental analysis.

Experiment Included Groups Corneal staining 8W, 1Y, and 2Y Dry eye assessment Tear secretion 8W, 1Y, and 2Y Conjunctival goblet cell count 8W, 1Y, and 2Y MG area analyzed by meibography 8W, 1Y, and 2Y Gross examination LG weight 8W, 1Y, and 2Y Histology H&E staining 8W, 1Y, and 2Y LipidTOX/PPARγ IHC (coronal section) * 8W and 2Y MG analysis Ki-67 and CD45 IHC * 8W and 1Y 8-OHdG IHC * 8W and 1Y Flow cytometry 8-OHdGhiPIlo 8W, 1Y, and 2Y BrdU+/7-AAD 8W, 1Y, and 2Y Int. J. Mol. Sci. 2020, 21, 4169 17 of 19

Table A2. Cont.

Experiment Included Groups Histology H&E staining and CD45 IHC 8W, 1Y, and 2Y Ki-67 IHC 8W, 1Y, and 2Y LG analysis 8-OHdG IHC * 8W and 1Y Flow cytometry CD4+ IFNγ+ and CD4+ IL-17A+ 8W, 1Y, and 2Y 8-OHdGhiPIlo 8W, 1Y, and 2Y BrdU+/7-AAD 8W, 1Y, and 2Y Secondary lymphoid organs Flow cytometry of DLN (mice) 8W, 1Y, and 2Y * Experiments performed in only two groups. Abbreviations: W, week-old; Y, year-old; MG, meibomian gland; LG, lacrimal gland; H&E, hematoxylin-eosin; IHC, immunohistochemical staining; DLN, drainage lymph node.

References

1. van Deursen, J.M. The role of senescent cells in ageing. Nature 2014, 509, 439–446. [CrossRef] 2. Muller, L.; Di Benedetto, S.; Pawelec, G. The Immune System and Its Dysregulation with Aging. Subcell. Biochem. 2019, 91, 21–43. [PubMed] 3. Schaumberg, D.A.; Sullivan, D.A.; Buring, J.E.; Dana, M.R. Prevalence of dry eye syndrome among US women. Am. J. Ophthalmol. 2003, 136, 318–326. [CrossRef] 4. Bian, F.; Xiao, Y.; Barbosa, F.L.; de Souza, R.G.; Hernandez, H.; Yu, Z.; Pflugfelder, S.C.; de Paiva, C.S. Age-associated antigen-presenting cell alterations promote dry-eye inducing Th1 cells. Mucosal Immunol. 2019, 12, 897–908. [CrossRef][PubMed] 5. de Souza, R.G.; de Paiva, C.S.; Alves, M.R. Age-related Autoimmune Changes in Lacrimal Glands. Immune Netw. 2019, 19, e3. [CrossRef][PubMed] 6. de Paiva, C.S. Effects of Aging in Dry Eye. Int. Ophthalmol. Clin. 2017, 57, 47–64. [CrossRef][PubMed] 7. Nien, C.J.; Massei, S.; Lin, G.; Nabavi, C.; Tao, J.; Brown, D.J.; Paugh, J.R.; Jester, J.V. Effects of age and dysfunction on human meibomian glands. Arch. Ophthalmol. 2011, 129, 462–469. [CrossRef] 8. Tsubota, K.; Kawashima, M.; Inaba, T.; Dogru, M.; Matsumoto, Y.; Ishida, R.; Kaido, M.; Kojima, T.; Uchino, M.; Uchino, Y.; et al. The antiaging approach for the treatment of dry eye. Cornea 2012, 31 (Suppl. 1), S3–S8. [CrossRef] 9. Nasu, M.; Matsubara, O.; Yamamoto, H. Post-mortem prevalence of lymphocytic infiltration of the lacrymal gland: A comparative study in autoimmune and non-autoimmune diseases. J. Pathol. 1984, 143, 11–15. [CrossRef] 10. Obata, H.; Yamamoto, S.; Horiuchi, H.; Machinami, R. Histopathologic study of human lacrimal gland. Statistical analysis with special reference to aging. Ophthalmology 1995, 102, 678–686. [CrossRef] 11. Rocha, E.M.; Alves, M.; Rios, J.D.; Dartt, D.A. The aging lacrimal gland: Changes in structure and function. Ocul. Surf. 2008, 6, 162–174. [CrossRef] 12. McClellan, A.J.; Volpe, E.A.; Zhang, X.; Darlington, G.J.; Li, D.Q.; Pflugfelder, S.C.; de Paiva, C.S. Ocular surface disease and dacryoadenitis in aging C57BL/6 mice. Am. J. Pathol. 2014, 184, 631–643. [CrossRef] [PubMed] 13. van der Geest, K.S.; Abdulahad, W.H.; Tete, S.M.; Lorencetti, P.G.; Horst, G.; Bos, N.A.; Kroesen, B.J.; Brouwer, E.; Boots, A.M. Aging disturbs the balance between effector and regulatory CD4+ T cells. Exp. Gerontol. 2014, 60, 190–196. [CrossRef][PubMed] 14. Schmitt, V.; Rink, L.; Uciechowski, P. The Th17/Treg balance is disturbed during aging. Exp. Gerontol. 2013, 48, 1379–1386. [CrossRef] 15. Paschides, C.A.; Petroutsos, G.; Psilas, K. Correlation of conjunctival impression cytology results with lacrimal function and age. Acta Ophthalmol. 1991, 69, 422–425. [CrossRef] 16. Wei, A.; Hong, J.; Sun, X.; Xu, J. Evaluation of age-related changes in human palpebral conjunctiva and meibomian glands by in vivo confocal microscopy. Cornea 2011, 30, 1007–1012. [CrossRef] Int. J. Mol. Sci. 2020, 21, 4169 18 of 19

17. Dogru, M.; Kojima, T.; Simsek, C.; Tsubota, K.; Tsubota, K.; Kawashima, M.; Inaba, T.; Dogru, M.; Matsumoto, Y.; Ishida, R.; et al. Potential Role of Oxidative Stress in Ocular Surface Inflammation and Dry Eye Disease. The antiaging approach for the treatment of dry eye. Invest. Ophthalmol. Vis. Sci. 2018, 59, DES163–DES168. [CrossRef] 18. Kojima, T.; Wakamatsu, T.H.; Dogru, M.; Ogawa, Y.; Igarashi, A.; Ibrahim, O.M.; Inaba, T.; Shimizu, T.; Noda, S.; Obata, H.; et al. Age-related dysfunction of the lacrimal gland and oxidative stress: Evidence from the Cu,Zn-superoxide dismutase-1 (Sod1) knockout mice. Am. J. Pathol. 2012, 180, 1879–1896. [CrossRef] 19. Kawashima, M.; Tsubota, K. Effect of calorie restriction on change in lacrimal gland with age. Cornea 2011, 30 (Suppl. 1), S29–33. [CrossRef] 20. You, S.; Tariq, A.; Kublin, C.L.; Zoukhri, D. Detection of BrdU-label retaining cells in the lacrimal gland: Implications for tissue repair. Cell Tissue Res. 2011, 346, 317–326. [CrossRef] 21. Eom, Y.; Choi, K.E.; Kang, S.Y.; Lee, H.K.; Kim, H.M.; Song, J.S. Comparison of meibomian gland loss and expressed meibum grade between the upper and lower eyelids in patients with obstructive meibomian gland dysfunction. Cornea 2014, 33, 448–452. [CrossRef] 22. Pult, H. Relationships Between Meibomian Gland Loss and Age, Sex, and Dry Eye. Eye Contact Lens 2018, 44 (Suppl. 2), S318–S324. [CrossRef][PubMed] 23. Pult, H.; Riede-Pult, B.H.; Nichols, J.J. Relation between upper and lower lids’ meibomian gland morphology, tear film, and dry eye. Optom. Vis. Sci. 2012, 89, E310–315. [CrossRef][PubMed] 24. Srinivasan, S.; Menzies, K.; Sorbara, L.; Jones, L. Infrared imaging of meibomian gland structure using a novel keratograph. Optom. Vis. Sci. 2012, 89, 788–794. [CrossRef] 25. Arita, R.; Minoura, I.; Morishige, N.; Shirakawa, R.; Fukuoka, S.; Asai, K.; Goto, T.; Imanaka, T.; Nakamura, M. Development of Definitive and Reliable Grading Scales for Meibomian Gland Dysfunction. Am. J. Ophthalmol. 2016, 169, 125–137. [CrossRef][PubMed] 26. Yeh, T.N.; Lin, M.C. Risk factors for severe Meibomian gland atrophy in a young adult population: A cross-sectional study. PLoS ONE 2017, 12, e0185603. [CrossRef] 27. Ban, Y.; Shimazaki-Den, S.; Tsubota, K.; Shimazaki, J. Morphological evaluation of meibomian glands using noncontact infrared meibography. Ocul. Surf. 2013, 11, 47–53. [CrossRef] 28. Nien, C.J.; Paugh, J.R.; Massei, S.; Wahlert, A.J.; Kao, W.W.; Jester, J.V. Age-related changes in the meibomian gland. Exp. Eye Res. 2009, 89, 1021–1027. [CrossRef] 29. Rosen, E.D.; Spiegelman, B.M. PPARgamma: A nuclear regulator of metabolism, differentiation, and cell growth. J. Biol. Chem. 2001, 276, 37731–37734. [CrossRef] 30. Rosen, E.D.; Sarraf, P.; Troy, A.E.; Bradwin, G.; Moore, K.; Milstone, D.S.; Spiegelman, B.M.; Mortensen, R.M. PPAR gamma is required for the differentiation of adipose tissue in vivo and in vitro. Mol. Cell 1999, 4, 611–617. [CrossRef] 31. Hwang, H.S.; Parfitt, G.J.; Brown, D.J.; Jester, J.V. Meibocyte differentiation and renewal: Insights into novel mechanisms of meibomian gland dysfunction (MGD). Exp. Eye Res. 2017, 163, 37–45. [CrossRef][PubMed] 32. Jester, J.V.; Potma, E.; Brown, D.J. PPARgamma Regulates Mouse Meibocyte Differentiation and Lipid Synthesis. Ocul. Surf. 2016, 14, 484–494. [CrossRef][PubMed] 33. Nocturne, G.; Mariette, X. B cells in the pathogenesis of primary Sjogren syndrome. Nat. Rev. Rheumatol. 2018, 14, 133–145. [CrossRef][PubMed] 34. Lee, H.; Shim, W.; Kim, C.E.; Choi, S.Y.; Lee, H.; Yang, J. Therapeutic Efficacy of Nanocomplex of Poly(Ethylene Glycol) and Catechin for Dry Eye Disease in a Mouse Model. Invest. Ophthalmol. Vis. Sci. 2017, 58, 1682–1691. [CrossRef][PubMed] 35. Sullivan, D.A.; Block, L.; Pena, J.D. Influence of androgens and pituitary hormones on the structural profile and secretory activity of the lacrimal gland. Acta Ophthalmol. 1996, 74, 421–435. [CrossRef][PubMed] 36. Sato, E.H.; Sullivan, D.A. Comparative influence of steroid hormones and immunosuppressive agents on autoimmune expression in lacrimal glands of a female mouse model of Sjogren’s syndrome. Invest. Ophthalmol. Vis. Sci. 1994, 35, 2632–2642. [PubMed] 37. Modulo, C.M.; Machado Filho, E.B.; Malki, L.T.; Dias, A.C.; de Souza, J.C.; Oliveira, H.C.; Jorge, I.C.; Santos Gomes, I.B.; Meyrelles, S.S.; Rocha, E.M. The role of dyslipidemia on ocular surface, lacrimal and meibomian gland structure and function. Curr. Eye Res. 2012, 37, 300–308. [CrossRef][PubMed] Int. J. Mol. Sci. 2020, 21, 4169 19 of 19

38. Kojima, T.; Nagata, T.; Kudo, H.; Muller-Lierheim, W.G.K.; van Setten, G.B.; Dogru, M.; Tsubota, K. The Effects of High Molecular Weight Hyaluronic Acid Eye Drop Application in Environmental Dry Eye Stress Model Mice. Int. J. Mol. Sci. 2020, 21, 3516. [CrossRef] 39. Kim, J.; Choi, S.H.; Kim, Y.J.; Jeong, H.J.; Ryu, J.S.; Lee, H.J.; Kim, T.W.; Im, S.H.; Oh, J.Y.; Kim, M.K. Clinical Effect of IRT-5 Probiotics on Immune Modulation of Autoimmunity or Alloimmunity in the Eye. Nutrients 2017, 9, 1166. [CrossRef] 40. De Roo, A.K.; Wouters, J.; Govaere, O.; Foets, B.; van den Oord, J.J. Identification of Circulating Fibrocytes and Dendritic Derivatives in Corneal Endothelium of Patients With Fuchs’ Dystrophy. Invest. Ophthalmol. Vis. Sci. 2017, 58, 670–681. [CrossRef][PubMed] 41. Kim, Y.H.; Kim, D.H.; Shin, E.J.; Lee, H.J.; Wee, W.R.; Jeon, S.; Kim, M.K. Comparative Analysis of Substrate-Free Cultured Oral Mucosal Epithelial Cell Sheets from Cells of Subjects with and without Stevens-Johnson Syndrome for Use in Ocular Surface Reconstruction. PLoS ONE 2016, 11, e0147548. [CrossRef][PubMed] 42. Subastri, A.; Suyavaran, A.; Preedia Babu, E.; Nithyananthan, S.; Barathidasan, R.; Thirunavukkarasu, C. Troxerutin with copper generates oxidative stress in cancer cells: Its possible chemotherapeutic mechanism against hepatocellular carcinoma. J. Cell Physiol. 2018, 233, 1775–1790. [CrossRef] 43. Reyes, N.J.; Yu, C.; Mathew, R.; Kunnen, C.M.; Kalnitsky, J.; Redfern, R.L.; Leonardi, A.; Perez, V.L.; MacLeod, A.S.; Gupta, P.K.; et al. Neutrophils cause obstruction of eyelid sebaceous glands in inflammatory eye disease in mice. Sci. Transl. Med. 2018, 10, eaas9164. [CrossRef] 44. Biosciences, B. Cell Cycle Analysis Using the BD BrdU FITC Assay on the BD FACSVerse™ System. Growth 2011, 2, G1. 45. Lima, K.G.; Krause, G.C.; Schuster, A.D.; Catarina, A.V.; Basso, B.S.; De Mesquita, F.C.; Pedrazza, L.; Marczak, E.S.; Martha, B.A.; Nunes, F.B.; et al. Gallic acid reduces cell growth by induction of apoptosis and reduction of IL-8 in HepG2 cells. Biomed. Pharmacother. 2016, 84, 1282–1290. [CrossRef][PubMed] 46. Paral, P.; Bajecny, M.; Savvulidi, F.; Necas, E. Cell Cycle Analysis Using In Vivo Staining of DNA-Synthesizing Cells. Methods Mol. Biol. 2019.[CrossRef] 47. Braumuller, H.; Wieder, T.; Brenner, E.; Assmann, S.; Hahn, M.; Alkhaled, M.; Schilbach, K.; Essmann, F.; Kneilling, M.; Griessinger, C.; et al. T-helper-1-cell cytokines drive cancer into senescence. Nature 2013, 494, 361–365. [CrossRef][PubMed] 48. Lee, H.K.; Kim, K.W.; Ryu, J.S.; Jeong, H.J.; Lee, S.M.; Kim, M.K. Bilateral Effect of the Unilateral Corneal Nerve Cut on Both Ocular Surface and Lacrimal Gland. Invest. Ophthalmol. Vis. Sci. 2019, 60, 430–441. [CrossRef] 49. Cambi, M.; Tamburrino, L.; Marchiani, S.; Olivito, B.; Azzari, C.; Forti, G.; Baldi, E.; Muratori, M. Development of a specific method to evaluate 8-hydroxy, 2-deoxyguanosine in sperm nuclei: Relationship with semen quality in a cohort of 94 subjects. Reproduction 2013, 145, 227–235. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).