Meibomian Gland Dysfunction Model in Hairless Mice Fed a Special Diet With Limited Lipid Content

Abstract

Purpose: A novel meibomian gland dysfunction (MGD) model was developed to facilitate understanding of the pathophysiology of MGD and to evaluate treatment with azithromycin ophthalmic solution (azithromycin). MGD was induced in HR-1 hairless mice by feeding them a special diet with limited lipid content (HR-AD).

Methods: Male HR-1 hairless mice were fed an HR-AD diet for 16 weeks. Development of MGD was assessed by histopathology at 4-week intervals. The lid margin was observed by slit-lamp examination. After cessation of the HR-AD diet, the mice were fed a normal diet to restore normal eye conditions. Expression of cytokeratin 6 was determined by immunostaining. We evaluated the effects of topically applied azithromycin on the plugged orifice in this model.

Results: After mice were fed the HR-AD diet, histopathology analysis showed hyperkeratinization of the ductal epithelium in the meibomian gland. Ductal hyperkeratinization resulted in the loss of acini, followed by atrophy of the gland. Slit-lamp examination revealed a markedly plugged orifice, telangiectasia, and a toothpaste-like meibum compared with that of a normal eyelid. Cessation of feeding with HR-AD ameliorated both the MGD signs and the expression of cytokeratin 6, restoring the tissue to a histologically normal state. Azithromycin treatment significantly decreased the number of plugged orifices and ameliorated atrophy, as revealed by histopathologic analysis.

Conclusions: We developed a novel model that mimics human MGD signs in HR-1 hairless mice fed an HR-AD diet. Azithromycin treatment led to therapeutic improvement in this model. This MGD model could be useful for the evaluation of drug candidates for MGD.

Meibomian glands, a type of sebaceous gland, are arranged vertically within the upper and lower tarsal plates.¹ Meibomian glands secrete lipids, which form a superficial oily layer on the tear film. Meibomian gland dysfunction (MGD) is defined as a chronic and diffuse abnormality of the meibomian glands, and it is commonly observed with terminal duct obstruction and/or qualitative or quantitative changes in glandular secretion. MGD results in alteration of the tear film, eye irritation, clinically apparent inflammation, and ocular surface disease.² Some ophthalmic examinations have been developed to diagnose MGD.³ The most conventional and significant means of diagnosing MGD is to use a slit lamp to thoroughly examine the lid margin, particularly that surrounding the orifice. In a patient with characteristic signs of obstructive MGD, slit-lamp microscopy reveals meibomian gland orifices that are closed with plugs consisting of thickened, opaque secretions containing keratinized material with telangiectasia around the orifice and lid margin rounding.^1,4 MGD impairs quality of life and quality of vision because the oily layer plays an important role in maintaining the stability of the tear film by preventing tears from evaporating, reducing the surface tension of tears and lessening the friction of blinking.⁴ The causes and progression of MGD should be further investigated to develop better pharmacologic treatments. Thus, appropriate animal models mimicking MGD, with a pathogenesis similar to that observed in humans, are strongly desirable to understand the pathophysiology of the disease and to develop potential pharmacologic interventions. There are several spontaneous, genetic animal models and chemically induced models of MGD, including the rhino mouse model,⁵ ACAT-1 null mice,⁶ an adrenaline-induced rabbit model,^7,8 a polychlorinated biphenyl (PCB)-induced monkey model,⁹ an isotretinoin-induced blepharitis model,¹⁰ and TRAF6-deficient mice.¹¹ Although these models provide insight into the potential causes and key molecular events of disease, the cause and pathophysiology of MGD remains unclear. Evaluating treatments for abnormal meibomian glands is difficult using these models because the clinical presentation is late-onset and/or irreversible. The lack of an appropriate model substantially contributes to the current lack of pharmacologic treatments for MGD. 

Azithromycin ophthalmic solution (AzaSite; Merck & Co., Inc., Whitehouse Station, NJ, USA) has been approved for the treatment of bacterial conjunctivitis. In addition to anti-bacterial effects, azithromycin suppresses the mRNA expression of inflammatory mediators toward normal levels in MGD patients.¹² Azithromycin can also act directly to promote lipid production in human meibomian gland epithelial cells.¹³ It has been reported that azithromycin significantly reduces meibomian gland plugging and eyelid redness and improves meibomian gland secretion and restores meibomian gland lipid secretion.^14,15 

Tobramycin/dexamethasone ophthalmic suspension is indicated for acute anterior blepharitis, although it is often used for posterior blepharitis.¹⁶ However, no reports have demonstrated that tobramycin/dexamethasone can reduce meibomian gland plugging. 

It has been proposed that MGD is significantly associated with higher blood levels of total cholesterol and increased blood levels of low-density lipoprotein than similarly aged controls without MGD.^17,18 Oral linoleic and gamma-linolenic acid supplementation have been shown to be beneficial in the treatment of MGD.¹⁹ These data suggest that lipid composition in the diet may be implicated in the development of MGD. 

Recent investigations have demonstrated that HR-1 hairless mice fed an HR-AD diet with limited lipid content, but not those fed a normal diet, developed atopic dermatitis-like symptoms, which are characterized by severely dry skin and inflammatory cellular infiltration into the skin.²⁰ Because abnormal sebaceous glands are one of the causes of severely dry skin, HR-1 hairless mice with severely dry skin are likely to show not only abnormal sebaceous glands but also meibomian gland dysfunction. However, macroscopic and histopathologic observations of meibomian glands in HR-1 hairless mice fed an HR-AD diet have not yet been reported. 

In this study, we investigated the pathology of MGD induced in HR-1 hairless mice fed HR-AD, including the occurrence of and temporal changes in morphology, to establish a novel MGD model to evaluate the treatment efficacy of eye drops. 

Materials and Methods

Animals

Four-week-old male HR-1 hairless mice were obtained from Hoshino Experimental Animal Center (Yashio, Saitama, Japan). The mice were housed under 12-hour light-dark cycles (lights on at 07:00 AM) at room temperature (23 ± 1°C) with humidity of 55 ± 10%. Food and water were provided ad libitum. Mice were fed a standard laboratory diet (F-2; Funabashi Farm, Chiba, Japan) or a special diet with limited lipid content (HR-AD; Nosan Corp., Yokohama, Japan), which was prepared for the HR-1 hairless mice, for up to 16 weeks.²⁰ After a set period of time, the HR-AD diet was changed to a normal diet to assess recovery of the plugged orifices. Experimental procedures were performed in accordance with Association for Research in Vision and Ophthalmology principles concerning the use of animals in ophthalmic and vision research. All experimental procedures were approved by the Committee on Animal Research at Santen Pharmaceutical Co., Ltd. (Osaka, Japan). 

Slit-Lamp Examination

Meibomian gland orifices of mice under isoflurane anesthesia were assessed using slit- lamp examination (SL-D7; Topcon, Tokyo, Japan) and were imaged using a digital camera (D100; Nikon, Tokyo, Japan). All images were obtained using the same camera with the same settings. The number of plugged orifices among the eight meibomian gland orifices in the center of the upper eyelid was quantified. In this study, plugged orifices were defined as opaque and swollen meibomian gland orifices. 

Histologic Analysis

Eyelid tissues, which included the meibomian gland orifices, were dissected at 4, 8, 12, and 16 weeks. Tissues were fixed with 10% neutral buffered formalin solution, embedded in paraffin, and vertically cut into 2-μm-thick sections from the temporal, central, and nasal portions of the eyelid. Next, the sections were stained with hematoxylin-eosin (HE) and rabbit polyclonal anti-cytokeratin 6 (CK 6; Abcam, Cambridge, UK) using the Simple Stain mouse MAX-PO kit (Nichirei Bioscience, Tokyo, Japan). The CK 6 signal was visualized with the chromogen diaminobenzidine tetrahydrochloride. Sections were counterstained with hematoxylin. 

MGD Treatment

Saline or AzaSite was instilled into the right eye for 4 weeks (2 μL/eye; twice per day on days 1 and 2 of instillation, and once per day on day 3 and thereafter) after the plugged orifice had developed as a result of feeding with HR-AD for 4 weeks. Saline or Tobradex ST ophthalmic suspension (2 μL/eye; 0.3% tobramycin and 0.05% dexamethasone; Alcon Laboratories, Inc., Fort Worth, TX, USA) was instilled four times daily into the right eye for 4 weeks after the plugged orifice had developed as a result of feeding with HR-AD for 2 weeks. Meibomian gland orifices were examined using a slit lamp at 2 and 4 weeks after the treatment. The number of plugged orifices was quantified in the same manner as described above. 

Statistical Analysis

Data are expressed as mean ± SE. The statistical significance of differences was assessed using the Aspin-Welch t-test or the Student's t-test. P values less than 0.05 were considered statistically significant. 

Results

MGD Assessed by Slit-Lamp Examination

Figure 1 shows an eyelid after 11 weeks of feeding with an HR-AD diet. Marked differences were observed between the lid margins of mice fed the HR-AD diet and those fed the normal diet. Rounding of the posterior lid margin was observed in the mice fed the HR-AD diet (Fig. 1C), and many plugged orifices and instances of telangiectasia were observed around the orifices in the eyelid margin (Fig. 1D). These changes became more distinct when the HR-AD feeding period was extended. After 16 weeks, thickened secretions and a toothpaste-like meibum appeared (Fig. 1E). No changes in the eyelid or lid margin were observed in the mice fed a normal diet (Figs. 1A, 1B). 

Figure 1

 

The eyelids of mice fed a normal diet (A, B) and the HR-AD diet (C, D, E) for 11 weeks, from 5 to 16 weeks of age. Rounding (C), plugging, and telangiectasia (D, arrowheads), and toothpaste-like meibum (E, arrowheads) can be seen.

Histopathologic Test for MGD

We evaluated histologic changes in the eyelid after 4, 12, and 16 weeks of feeding with HR-AD (Fig. 2). In the mice fed the HR-AD diet, the meibomian glands exhibited dramatic changes at all time points. At 4 weeks, the mice fed the HR-AD diet exhibited a thickening and hyperkeratinization of ductal epithelium in the meibomian glands (Fig. 2C). Keratin accumulation was detected in the ducts, and the orifices were plugged with a keratinized substance. In addition, a marked loss or atrophy of the acinar component was also observed in the glands. The thickness of the ductal epithelium increased, and the losses in acinar volume were enhanced as the HR-AD feeding period continued (Figs. 2C–E). Furthermore, the thickening of the epidermis was followed by hyperkeratosis in the HR-AD-fed mice (Figs. 2D, 2E). These histological changes in the meibomian glands were observed in the whole eyelid. However, in control mice fed the normal diet, the meibomian glands did not show any changes over the experimental period (Figs. 2A, 2B). A few inflammatory cells infiltrated the tarsal plates in both the HR-AD-fed and the normal diet-fed mice. 

Figure 2

 

Histology of the meibomian gland in mice fed a normal diet for 4 (A) and 16 (B) weeks, an HR-AD diet for 4 (C), 12 (D), and 16 (E) weeks. HE staining. Scale Bars: 200 μm.

Amelioration of MGD Signs After Diet Alteration

Next, we examined whether a normal diet could reduce the MGD signs induced by the HR-AD diet in mice. Five-week-old HR-1 hairless mice were fed an HR-AD diet for 14 days to develop plugged orifices. Once the mice had developed MGD signs, they were randomly divided into two groups. Mice in one group continued to be fed the HR-AD diet (HR-AD group), and mice in the other group were fed the normal diet (ND group) for an additional 4 weeks. In the HR-AD group, the number of plugged orifices continued to increase for 6 weeks (Fig. 3). In contrast, mice in the ND group had a decreasing number of plugged orifices. Significant improvements in MGD were observed by reducing the number of plugged orifices 4 weeks after the diet conversion. Moreover, immunohistological testing showed that mice in the ND group exhibited a histologic reversal of cytokeratin 6 staining in the meibomian ductal epithelium compared with the mice in the HR-AD group (Figs. 4B, 4C). The acinar atrophy was also recovered by diet conversion (data not shown). 

Figure 3

 

Effect of diet conversion on the number of plugged orifices in the MGD model induced in HR-1 hairless mice fed the HR-AD diet. The number of plugged orifices was recovered by switching from the HR-AD diet to a normal diet at 2 weeks. Each bar represents the mean ± SE of 6 eyes. *P < 0.05 versus mice fed a normal diet (Student's t-test). ##P < 0.01 versus mice fed the HR-AD diet (Student's t-test).

Figure 4

 

Immunostaining of the meibomian gland with keratin 6. HR-1 hairless mice fed a normal diet ([A] normal); MGD model mice fed the HR-AD diet (B); and HR-1 hairless mice fed a normal diet for 4 weeks after HR-AD-induced MGD (C). Overexpressed cytokeratin 6 in the meibomian ductal epithelium was recovered to normal levels (compare [C] to [A]). Scale Bars: 200 μm.

Effects of Drug Treatment on Plugged Meibomian Gland Orifices

Azithromycin was therapeutically administered to HR-1 hairless mice with plugged orifices induced by the HR-AD diet. Azithromycin significantly reduced the number of plugged orifices during treatment for 4 weeks (Fig. 5). Azithromycin also reduced the thickened and hyperkeratinized meibomian ductal epithelium and acinar atrophy (Fig. 6). However, tobramycin/dexamethasone, which is a mixture of steroids and antibiotics, did not show any efficacy on the plugged orifices under the experimental conditions (Fig. 7). 

Figure 5

 

Effect of azithromycin on the number of plugged meibomian orifices in the HR-AD-induced MGD model mice in HR-1 hairless mice. Mice were treated with azithromycin for 4 weeks after development of plugged orifices after being fed the HR-AD diet for 4 weeks. Each bar represents mean ± SE of 6 eyes. ##P < 0.01 versus normal (Aspin-Welch t-test). **P < 0.01 versus normal (Student's t-test). $$P < 0.01 versus saline (Student's t-test).

Figure 6

 

Histology of the meibomian gland in the MGD model, induced by the HR-AD diet. The MGD model was maintained with HR-AD and treated with saline (A) or azithromycin (B) and was maintained with the normal diet (C) for 4 weeks. Azithromycin decreased the thickness and keratinization of the meibomian ductal epithelium. The acinar size also improved to a size similar to that observed in panel C (compare [B] to [C]). HE staining. Scale Bars: 200 μm.

Figure 7

 

Effect of tobramycin/dexamethasone on the number of plugged orifices in the MGD model mice, induced with the HR-AD diet. Mice were treated with tobramycin/dexamethasone for 4 weeks after development of plugged orifices by the HR-AD diet for 2 weeks. Each bar represents the mean ± SE of 6 eyes. **P < 0.01 versus normal (Student's t-test).

Discussion

HR-AD-fed HR-1 hairless mice showed plugged orifices, telangiectasia surrounding the orifices, posterior lid margin rounding, and toothpaste-like meibum, by slit-lamp examination. These pathophysiologic characteristics are similar to those observed in MGD patients.²¹ The mice fed on the HR-AD diet for 4 weeks exhibited acinar atrophy, which was diagnosed using a histologic test. The onset of these MGD signs was also observed earlier than in previous reports.^5,7,9 These data suggested that daily diet could be the cause of MGD, and our model should soon be available to evaluate drug candidates and to elucidate the pathophysiology of MGD. 

Because MGD is defined as a “diffuse abnormality” of the meibomian glands, a wide range of meibomian gland conditions could be evaluated using this animal model. To the best of our knowledge, a mouse has approximately 10 or more meibomian glands in the upper eyelid. In the present study, we evaluated plugged orifices and selected 8 meibomian gland orifices that were close to the center of the upper eyelid due to the difficulty in evaluating the orifice that was proximately close to the lid aperture of the slit lamp. Orifices in this animal model closely resembled those of MGD patients, using microscopy. Because microscopic observation with a slit lamp is important for diagnosing MGD in clinical practice, the animal model described in this study should be useful for studying the clinical state of the orifices of multiple meibomian glands by slit lamp. 

Thickening and hyperkeratinization of the ductal epithelium and atrophy of acinar cells in the meibomian gland were observed by histological analysis in this model. Jester et al.⁵ reported keratinized epithelium of the epidermis and markedly thickened and hyperkeratinized meibomian glands in rhino mice. Adrenaline treatment induced the accumulation of cell debris in the ducts of rabbits.⁸ In rhesus monkeys, PCB induced abnormal hyperkeratosis in the ductal epithelium, which was observed by histopathological analysis.⁹ Taken together, these previous observations in MGD animal models are consistent with the data from our model. 

Although the mechanisms underlying the induction of MGD signs by HR-AD remain unclear, these signs resemble the characteristic changes observed in aging humans.²² HR-1 hairless mice fed the HR-AD diet have also been reported to develop skin barrier dysfunction characterized by increased transepidermal water loss and thickening of the epidermis.²⁰ These observations suggest that breakdown of the meibomian ductal epithelium barrier may cause epithelial cell thickening and hyperkeratinization. Our data revealed that cytokeratin 6 was significantly overexpressed and that keratinized substances accumulated and plugged the meibomian orifices, resulting in an apparent extreme loss of mature acini followed by atrophy of the gland. Expression of the keratin 6 gene is also significantly increased in the meibomian glands of MGD patients.²³ 

Meibomian gland obstruction is reportedly influenced by endogenous factors, such as age, sex, and hormonal disturbances, as well as by exogenous factors, such as topical medications.¹ However, the relevance of inflammation in the pathophysiology of MGD remains controversial.¹⁶ In previous histopathologic studies of humans, the infiltration of inflammatory cells was not observed in specimens of either cystic dilation or acinar atrophy of the meibomian gland.^22,24 Few inflammatory cells were detected in the meibum of elderly patients, whereas keratinized material was observed in nearly all cases (Obata H, et al. IOVS 2002;43:ARVO E-Abstract 60). However, inflammatory cytokines were recently detected in tear samples obtained from MGD patients.²⁵ The proinflammatory state might have pathophysiologically occurred in these cases of MGD. Although minimal infiltration of inflammatory cells into the meibomian gland was observed under the experimental conditions in our model, the instillation of steroid eye drops did not improve the obstructed meibomian gland orifices. The infiltration of inflammatory cells did not appear to be essential in the development of plugging. Taking these results into consideration, our model should pathophysiologically mimic human MGD. 

More importantly, in this study, we found that a normal diet ameliorated the plugged orifices and acinar atrophy induced by special diet feeding. This is the first study to demonstrate improvements in the obstructed orifice and acinar atrophy using an animal model. Although the HR-AD diet is known to be deficient in lipids and some types of minerals, few detailed analyses of the components of this diet have been reported. According to an investigation by Fujii et al.²⁶ deficiency of n-6 polyunsaturated fatty acids, particularly linoleic acid, is mainly responsible for abnormal skin conditions because supplemental linoleic acid in HR-AD improved the dry skin symptoms in HR-1 hairless mice. Furthermore, the authors could not improve the skin condition by supplementation of minerals, including magnesium.²⁰ Meibum consists of substances such as wax esters, cholesteryl esters, triacylglycerols, and free fatty acids.²⁷ Arita et al.²⁸ reported that the free fatty acid composition of human meibum correlated with meibum color, as determined using slit-lamp microscopy. Oral fatty acids, particularly linoleic acid and gamma-linolenic acid, significantly reduced secretion of turbid substances and meibomian gland obstruction in MGD patients.¹⁹ In this context, free fatty acids, particularly linoleic acid, should play an important role in the development of MGD signs in this model, and the mechanism should be similar to that of MGD in humans. 

Azithromycin ophthalmic solution has been proposed as a novel treatment for posterior blepharitis, including MGD.²⁹ Foulks et al.¹⁵ reported that topical therapy with azithromycin relieved the signs and symptoms of MGD and restored the lipid properties of meibomian gland secretion to a normal state. In our model, azithromycin improved plugged meibomian orifices. This is the first report to show an improvement in obstructed meibomian orifices using therapeutic treatment with any drugs in animal models. This finding indicates that our model should be useful for evaluating drug candidates for MGD. Azithromycin exerts not only an antibacterial action but also anti-inflammatory effects.³⁰ In addition, azithromycin treatment has been shown to result in lipid accumulation in immortalized human meibomian gland epithelial cells.¹³ These properties contribute to the efficacy of azithromycin in this model. However, further study is required to reveal the pharmacologic mechanisms of the effect of azithromycin on plugged meibomian gland orifices and acinar atrophy. 

In this study, although MGD is considered a leading cause of evaporative-type dry eye in humans, corneal fluorescein staining did not change with the HR-AD diet. It is assumed that mice already showed variable staining before the HR-AD feeding period. Abnormal changes were thought to result from alterations to the ocular surface caused by small substances, such as dusts and chips because these mice lacked eyelashes. Therefore, the effects of plugging on tear stability in this model remain unclear. Although our model showed toothpaste-like meibum as a characteristic feature of MGD, it appears difficult to elucidate the components of the meibum due to the limited meibum volume in mice. 

Here we report the successful development of a novel MGD model induced by HR-AD in HR-1 hairless mice, demonstrating early onset of the characteristic clinical signs and atrophy of acinar cells in meibomian glands. A normal diet ameliorated the plugged orifices and restored the tissues to normal. Topically applied azithromycin was also shown to have therapeutic effects on the number of plugged meibomian gland orifices and atrophy of acinar cells. Taken together, the results indicate that this model could be useful not only for the elucidation of the pathogenesis of MGD but also for the evaluation of the efficacy of drug candidates for treating MGD. 

Acknowledgments

The authors thank Toru Shibata, Miwa Yoshimi, and Yumi Kuriki-Yamamoto for performing histologic tests. 

Disclosure: H. Miyake, Santen (E), P; T. Oda, Santen (E), P; O. Katsuta, Santen (E); M. Seno, None; M. Nakamura, Santen (E)