Isolation and Culture of Human Meibomian Gland Ductal Cells

Abstract

Purpose: Hyperkeratinization of meibomian gland (MG) ducts is currently recognized as the primary pathologic mechanism of meibomian gland dysfunction (MGD). This research figured out a method to isolate the MG ducts and established a novel system to culture the human meibomian gland ductal cells (HMGDCs) for investigating the process of MGD.

Methods: The MG ducts were obtained from the eyelids of recently deceased donors and subjected to enzymatic digestion. The acini were then removed to isolate independent ducts. These MG ducts were subsequently cultivated on Matrigel-coated wells and covered with a glass plate to obtain HMGDCs. The HMGDCs were further cultivated until passage 2, and when they reached 60% confluence, they were treated with IL-1β and rosiglitazone for a duration of 48 hours. Immunofluorescence staining and Western blot techniques were employed to identify ductal cells and analyze the effects of IL-1β on HMGDCs in an in vitro setting.

Results: Ophthalmic micro-forceps and insulin needles can be employed for the purpose of isolating ducts. Within this particular culture system, the rapid expansion of HMGDCs occurred in close proximity to the duct tissue. MG ducts specifically expressed keratin 6 (Krt6) and hardly synthesized lipids. Furthermore, the expression of Krt6 was significantly higher (P < 0.0001) in HMGDCs compared to human meibomian gland cells. Upon treatment with IL-1β, HMGDCs exhibited an overexpression of keratin 1, which was effectively blocked by the administration of rosiglitazone.

Conclusions: The present study successfully isolated human MG ducts and cultured HMGDCs, providing a valuable in vitro model for investigating the mechanism of MGD. Additionally, the potential therapeutic efficacy of rosiglitazone in treating hyperkeratinization of ducts in patients with MGD was identified.

The meibomian gland (MG), a type of sebaceous and holocrine gland located in the superior and inferior tarsal plates, is essential for the maintenance of the tear film and the health of the ocular surface.¹ MG acini synthesize meibum, and the MG ducts deliver it to the ocular surface during blinking, maintaining the tear film stability and preventing microbiologic and organic contamination.² 

Once the normal function of the MGs is impaired, diminished tear film lipid content results in dry eye, which has been recognized as a growing public health concern causing ocular discomfort and impairing quality of life.³ Meibomian gland dysfunction (MGD) is commonly defined as a chronic, diffuse abnormality of the MGs, commonly characterized by terminal duct obstruction and/or qualitative/quantitative changes in glandular secretion.⁴ Even if both low- and high-delivery states of MGs can lead to MGD,^5,6 the most common mechanism for MGD is a low-delivery state characterized by MG duct obstruction. Massive shedding of hyperkeratotic ductal cells is identified as the leading cause of duct occlusion.⁶ 

Our previous research demonstrated that crude extracts of Staphylococcus aureus caused hyperkeratinization of ducts and increased expression of proinflammatory cytokines in MG explants.⁷ Additionally, it was confirmed that IL-1β, a proinflammatory cytokine, induced duct hyperkeratinization in rat MG tissue culture.⁸ However, these findings were limited to explant observations, and the precise underlying mechanisms remain unknown. To gain further insight into MG ducts, the initial step involves isolating and cultivating human meibomian gland ductal cells (HMGDCs). 

The primary objective of this study was to demonstrate the isolation and independent culture of human MG ducts. The detailed procedures for separating MG ducts and modifying the culture system were described. Furthermore, the study revealed that IL-1β enhanced the expression of keratin 1 (Krt1) in HMGDCs, while rosiglitazone inhibited this effect of IL-1β. 

Materials and Methods

Human Tissue

The donors were selected based on their age, with a maximum age limit of 70 years, and the time since their demise, which was less than 4 hours. Additionally, the donors were required to have no atrophy of the MGs or any other apparent ocular and systemic diseases. The human eyelids utilized in this study were obtained from a total of 10 donors, consisting of 5 women and 5 men, with ages ranging from 53 to 68 years. The acquired eyelids were uniformly trimmed, without differentiation between the upper and lower eyelids. The removal of tarsal plates followed the established protocol of the eye bank. The use of human tissue was approved by the Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (2022-S191), and adhered to the tenets of the Declaration of Helsinki. 

Isolation and Culture of Human Meibomian Gland Ductal Tissues

After being washed with phosphate-buffered saline supplemented with 100 U/mL penicillin and 100 mg/mL streptomycin (Hyclone, Logan, UT, USA), the subcutaneous tissue, muscle, and conjunctiva were carefully removed. The tarsal plates were then placed in 2.5% Collagenase Ι (Sigma-Aldrich, St. Louis, MO, USA) and digested at 37°C for 4 hours. Following digestion, the connective tissue of the eyelid became soft, allowing for the separation of each gland under an anatomic microscope. Once the intact gland was obtained using microscissors, the ducts could be acquired by gently peeling the acinar with an insulin needle (Figs. 1A–C). Approximately 30 to 50 ducts were procured from the four eyelids of a solitary donor. Culture plates were precoated with a 5% Matrigel solution (356234; Corning, Corning, NY, USA) for a duration of 1 hour. Subsequently, four acupuncture needles with a diameter of 0.25 mm were strategically positioned around the periphery of the culture plates, while the isolated MG ducts were carefully positioned at the center. Considering the necessity for adequate space to facilitate duct tissue proliferation, approximately 15 ducts were allocated per culture well. Considering the necessity for adequate space to facilitate duct tissue proliferation, approximately 15 ducts were allocated per culture well. To prevent floating, a glass plate was employed to cover the MG ducts. The MG ducts were subsequently covered with a glass plate to prevent buoyancy. Subsequently, an adequate amount of defined keratinocyte serum-free medium (DKSFM) (Thermo Fisher Scientific, Waltham, MA, USA), supplemented with 10 ng/mL epidermal growth factor (Gibco, Grand Island, NY, USA), was added to fully submerge the ducts (Fig. 1D). The MG ducts were then incubated at 37°C in a 5% CO₂ environment, with the medium being replaced every 2 days. Upon reaching 90% confluence, the cells were enzymatically detached using 0.25% Trypsin-EDTA (Thermo Fisher Scientific) for subsequent subculturing. 

Figure 1.

 

Isolation and culture of human meibomian gland ducts. (A) After being enzymatically digested, the subcutaneous tissue between the glands becomes more transparent and the acini can be seen more clearly. (B) The acini were removed with ophthalmic micro-forceps to expose the ducts. (C) The isolated MG duct was transferred to the culture plate. (D) Schematic of the tissue culture model. The ducts were carefully transferred to a six-well culture plate coated with 5% Matrigel. Supported by four acupuncture needles, a glass plate covered those MG ducts to avoid floating. The ducts were incubated at 37°C with 5% CO₂.

Treatment of IL-1β and Rosiglitazone on HMGDCs

IL-1β (211-11B; PeproTech, Rocky Hill, NJ, USA) was dissolved in DKSFM to prepare stocks with a concentration of 10 mg/mL. Rosiglitazone (HY-17386; MedChemExpress, Monmouth, Junction, NJ, USA) was dissolved in DKSFM to prepare stocks with a concentration of 5 mM. HMGDCs were exposed to medium containing 50 ng/mL IL-1β, 50 ng/mL IL-1β + 50 µM rosiglitazone, or DKSFM alone for a duration of 48 hours, when the cells reached approximately 60% confluency. These concentrations of pharmacologic reagents were selected based on previous studies conducted on mouse MG explants. The earlier investigations revealed that the treatment of mice MGs with 50 ng/mL IL-1β did not yield any noteworthy impact on cell activity, while a noticeable keratinizing effect was observed.^8,9 

Histology

The tarsal plates were initially fixed in a 4% paraformaldehyde solution for a duration of 1 hour at room temperature. Subsequently, they underwent dehydration using a series of alcohol solutions before being embedded in paraffin for tissue sectioning at a thickness of 8 µm. The resulting sections were then subjected to staining with hematoxylin (Beijing Solarbio Biotech Co., Ltd., Beijing, China) for a period of 10 minutes, followed by rinsing with water. Subsequently, the sections were stained with a 1% hydrochloric acid ethanol solution (Zhongshan Jinqiao Biotechnology Co., Ltd., Beijing, China) and rinsed with water. Finally, eosin staining (Beijing Solarbio Biotech Co., Ltd.) was performed, followed by dehydration using a gradient ethanol solution (Kermel Biotech Co., Ltd., Tianjin, China). The sections were then rendered transparent using xylene (Kermel Biotech Co., Ltd.) and sealed with neutral gum (Zhongshan Jinqiao Biotechnology Co., Ltd.). 

Immunofluorescence Staining

The isolated MG tissues were snap-frozen in Tissue-Tek optimum cutting temperature compound (Sakura Finetek, Tokyo, Japan), cut into 8-µm-thick sections, and mounted on poly-L-lysine–coated glass slides. The MGs sections or cultured cells were fixed with 4% paraformaldehyde for 15 minutes, permeabilized with 0.5% Triton X-100 for 30 minutes, and blocked with 5% BSA for at least 1 hour at room temperature and then incubated with rabbit monoclonal antibody to keratin 6 (Krt6) (1:200, ab93279; Abcam, Cambridge, UK), Krt1 (1:200, ab185628; Abcam, Cambridge, UK) and mouse monoclonal antibody to Krt6 (1:200, ab218438; Abcam, Cambridge, UK), PPARγ (1:200, sc-7273, Arigo, Hamburg, Germany), and keratin 14 (Krt14) (1:200, ab7800; Abcam, Cambridge, UK) overnight at 4°C. After incubation, the samples were stained Cy3-conjugated Donkey Anti-Rabbit IgG (1:500, GB21403; Servicebio, Wuhan, China) or Alexa Fluor 488, Goat Anti-Mouse IgG (1:500, GB25301; Servicebio, Wuhan, China) at 37°C for 1 hour. DAPI (GDP1024; Servicebio, Wuhan, China), Nile Red (72485; Sigma-Aldrich, St. Louis, MO, USA), and LipdTOX (H34477; Thermo Fisher, Waltham, MA, USA) were used for nucleus and lipid staining, respectively. Images were obtained with a laser-scanning confocal microscope (Nikon, Tokyo, Japan). 

Western Blot

Cultured cells were lysed in RIPA Lysis and Extraction Buffer (78440; Thermo Fisher Scientific) with 1% PMSF Protease Inhibitor (36978; Thermo Fisher Scientific). Protein concentration was measured using a BCA assay kit (23225; Thermo Fisher Scientific). Electrophoresis separated proteins on 10% SDS-PAGE gels and transferred them to polyvinylidene difluoride membranes. The samples were blocked in 5% milk for 1 hour at room temperature and then incubated with rabbit monoclonal antibody against Krt6 (1:1000), rabbit polyclonal antibody to GAPDH (1:2000), mouse monoclonal antibody PPARγ (1:1000), and Krt1 (1:1000) at 4°C overnight. Immunodetection was performed with horseradish peroxidase–conjugated goat anti-rabbit IgG antibody (1:5000, ANT020; AntGene, Wuhan, China) and visualized with an ECL reaction kit (P0018, Beyotime Biotechnology, Shanghai, China). Bands were analyzed by normalizing to the corresponding GAPDH bands using ImageJ 1.52a software (National Institutes of Health, Bethesda, MD, USA). 

Statistical Analysis

All data were expressed as means ± standard deviations and analyzed using GraphPad Prism 8 software (GraphPad Software, La Jolla, CA, USA). One-way ANOVA test was used for multiple groups, and t-test was used for two groups. A value of P < 0.05 was considered statistically significant. 

Results

Morphology, Biomarkers Expression, and Lipid Distribution of Human MGs

In 1666, Heinrich Meibom provided the initial description of MGs, accompanied by an illustrative depiction highlighting their fundamental characteristics.¹ Subsequent to the removal of the donor's eyelids, the subcutaneous tissue and muscle were excised to enhance the clarity of MG morphology. The arrangement of MGs was observed to be uniform and linear along the eyelid, with the upper eyelid housing approximately 25 to 30 MGs. The superior MGs exhibited a symmetrical distribution in a semiorbicular pattern, measuring 0.73 ± 0.24 cm, while the inferior MGs were comparable in structure but shorter in length than their upper counterparts (0.42 ± 0.06 cm, Figs. 2A, 2B). Upon removal of the connective tissue, a fully intact gland containing acini and ducts was observed under the microscope. The duct was situated centrally, surrounded by acini (Fig. 2C). The morphologic characteristics and details were evident in frozen sections stained with hematoxylin and eosin (Fig. 2D). 

Figure 2.

 

Morphology of the human MGs. (A) The morphology of the human MGs by removing subcutaneous tissue, muscle, and palpebral conjunctiva. (B) The average length of MGs in the upper eyelid was 0.73 ± 0.24 cm, and the average length of MGs in the lower eyelid was 0.42 ± 0.06 cm. (C) The acini (Ac) were distributed around the central duct (Du) in a single MG after enzymatic isolation and by (D) hematoxylin and eosin staining.

Immunofluorescence staining revealed the presence of Krt14 in all MG cells, while intense expression of Krt6 was limited to ductal cells (Figs. 3A–C), consistent with findings from a previous study.¹⁰ Additionally, a high concentration of lipids was predominantly detected in acinar cells (Figs. 3D–F). 

Figure 3.

 

Biomarker expression and lipid distribution in the human MGs. (A) Krt6 (green) was only intensively expressed in the ducts. (B) Krt14 (red) was detected in all the MG cells. (C) Merge of fluorescence signals are shown. The expression of (D) LipidTOX (green) and (E) Nile Red (red) with nuclei labeled by DAPI (blue) was examined in cryosections. (F) Merge of fluorescence signals are shown. The lipids in MGs were mainly distributed in acini.

Culture of Human Meibomian Gland Ductal Tissues and HMGDCs

In this study, DKSFM was utilized as a cell proliferation medium.¹¹ The cells exhibited proliferation primarily in the vicinity of the duct. Within 24 hours of introducing the medium to the MG duct culture system, HMGDCs initiated migration and rapid proliferation. After 48 hours, a significant number of HMGDCs demonstrated a remarkable capacity for multiplication (Figs. 4A, 4B). However, after approximately 96 hours, no discernible cell proliferation was observed in the vicinity of the duct. Consequently, these cells were subjected to sequential digestion and subsequent culture. Upon subculturing, the HMGDCs exhibited substantial proliferation similar to the initial generation. As illustrated in Fig. 5, HMGDCs were cultured to the fifth passage, and the proliferation capacity of HMGDCs gradually declined with the progression of passage times. Consequently, the cells from the second passage were employed for follow-up experiments. 

Figure 4.

 

The culture of the human meibomian gland ductal cells. (A) After 24 hours, cells grew out from ductal tissue. (B) More duct cells showed great multiplication capacity after 72 hours. (C–E) HMGDCs were cultivated to passage 2 and observed at (C) 24 hours, (D) 48 hours, and (E) 72 hours.

Figure 5.

 

The passage of human Meibomian gland ductal cells. (A–D) After being subcultured for 48 hours, the HMGDCs were investigated using microscopy. (E) The proliferation capacity of cells gradually declined with the progression of passage times.

Identification of Human Meibomian Gland Ductal Cells

The identification of HMGDCs was designed to compare with intact human meibomian gland cells (HMGCs), which were not dissociated and therefore contained both ductal cells and acinar cells. In this study, Krt6 was utilized as a specific marker for HMGDCs.¹² As anticipated, immunofluorescence staining (Fig. 6C) revealed that over 98.1% of the cells exhibited positive expression of Krt6, and Western blot (Figs. 6G, 6H) demonstrated that the expression of Krt6 in the isolated duct cells was higher than that in HMGCs (P < 0.0001). Additionally, the Western blot results (Figs. 6I, 6J) indicated that the expression of PPARγ, a crucial regulator of lipid secretion, was lower in the isolated duct cells compared to HMGCs (P < 0.001). Overall, these findings provide confirmation of successful in vitro cultivation of HMGDCs. 

Figure 6.

 

The expression of Krt6 and PPARγ in HMGDCs and HMGCs. (A–F) Immunofluorescence of Krt6 (green) and Krt14 (red), with nuclei labeled by DAPI (blue) in (A–C) HMGDCs and (D–F) HMGCs. (G, H) Representative Western blot result of Krt6 protein of HMGDCs and HMGCs in DKSFM. GAPDH was used as the control. HMGDCs showed more Krt6 (n = 3 in each group) (****P < 0.0001). (I, J) Representative Western blot result of PPARγ of HMGDCs and HMGCs in DKSFM. GAPDH was used as the control. HMGDCs showed less PPARγ (n = 3 in each group) (***P < 0.001).

Effects of IL-1β on Hyperkeratosis and Rosiglitazone on Antihyperkeratosis of HMGDCs

It has been reported that keratinized epithelium appears limited to ductal orifice in the normal eyelid.¹³ Hyperkeratinization of the ductal orifice is the main pathologic change for obstructive MGD.⁶ Krt1, a biomarker for epithelial keratinization,¹⁴ is increased in mouse MG explants after IL-1β treatment.⁸ This study aimed to examine the antikeratinization effect of rosiglitazone on HMGDCs stimulated by inflammation by treating them with 50 ng/mL IL-1β or 50 ng/mL IL-1β + 50 µM rosiglitazone for 48 hours. To ensure the reliability of the experimental results, we first demonstrated that there was no significant alteration in the expression of Krt14 subsequent to treatment with IL-1β and rosiglitazone based on immunofluorescence and Western blot (Figs. 7A–H). Immunofluorescence staining (Figs. 8A–I) demonstrated that IL-1β induced the expression of Krt1 in the HMGDCs. Additionally, Western blot analysis (Figs. 8J, 8K) confirmed the upregulation of Krt1 protein expression in IL-1β–treated HMGDCs. The potential anti-inflammatory effects of rosiglitazone on the ocular surface have been observed.¹⁵ When simultaneously treated with IL-1β and rosiglitazone, the impact of IL-1β on HMGDCs was mitigated. Additionally, rosiglitazone resulted in a downregulation of Krt1 protein expression in IL-1β–treated HMGDCs. 

Figure 7.

 

Impact of IL-1β and rosiglitazone on the expression of Krt14 in HMGDCs. (A–F) Immunofluorescence of Krt14 (red) with nuclei labeled by DAPI (blue) in ducal cells exposed to (A, B) vehicle, (C, D) IL-1β, or (E, F) rosiglitazone. (G, H) The relative expression of Krt14 protein demonstrated that there was no significant alteration in the expression of Krt14 subsequent to treatment with IL-1β and rosiglitazone.

Figure 8.

 

Impact of IL-1β and rosiglitazone on keratinization HMGDCs. (A–I) Immunofluorescence of Krt1 (green) and Krt6 (red) with nuclei labeled by DAPI (blue) in ducal cells exposed to (A–C) vehicle, (D, E) IL-1β, or (G–I) rosiglitazone. (J, K) The relative expression of Krt1 protein slightly increased and declined in 50 ng/mL IL-1β and only a little increased in 50 ng/mL IL-1β + 50 µM rosiglitazone. GAPDH was used as the control (n = 3 in each group) (**P < 0.01, ****P < 0.0001).

Discussion

The findings of this study indicate that it is possible to isolate and culture human MG ducts in vitro. The identification of HMGDCs was achieved through the detection of Krt6, a protein that is specifically expressed in MG ducts. Furthermore, this study provides evidence that the expression of Krt1 in HMGDCs can be induced by IL-1β in vitro. Additionally, the application of rosiglitazone was found to reduce the keratinization of HMGDCs. 

Over the past few decades, significant efforts have been made to elucidate the involvement of gland ducts in secretory processes in pancreatic and salivary glands.^16,17 Researches have also indicated that lacrimal gland duct cells play a significant role in the production of lacrimal gland fluids, similar to previous findings.¹⁸ Additionally, it has been suggested that the effectiveness and extent of glandular restoration in MGD are closely linked to the integrity of the MG ductal system.¹⁹ The surviving MG ductal network seems to influence the ability of acinar regeneration following induced MG atrophy.¹⁹ Consequently, despite the lower quantity of ductal cells compared to acinar cells, ducts are crucial for the normal physiologic functioning of MGs. 

According to the research, the gland orifice, central ducts, and meibum themselves exhibit abnormal keratinization in patients with MGD.²⁰ The hyperkeratinization of MG ducts can lead to obstruction in various ways, such as duct outlet stenosis. Maskin and Alluri²¹ achieved immediate and significant relief of symptoms in over 95% of cases by using fine nonsharp probes to restore a patent orifice and central duct, thereby relieving the obstruction. Further understanding of the pathologic changes associated with duct hyperkeratinization and the alleviation of ducts stenosis would be beneficial for the treatment of MGD. 

Inflammation plays a significant role in the pathogenesis of MGD, leading to ductal hypercornification and disruption of normal duct function.⁶ Posterior blepharitis is commonly linked to MGD, with inflammation from anterior blepharitis potentially extending to the posterior lid margin and causing secondary MGD.²² These observations highlight the strong association between inflammation and MGD. To simulate the disease state in vivo, we exposed HMGDCs to proinflammatory factors, aiming to look for potential therapeutic interventions for MGD. 

Previous studies have successfully cultured MG cells derived from rat and mouse.^7,23 However, when digesting the entire MGs, ductal cells became mixed with acinar cells and could not be isolated. Similarly, the immortalized human meibomian gland epithelial cell line, which is currently the only available immortalized cell line for in vitro studies on the physiology of meibomian gland epithelial cells, cannot be clearly classified as either ductal or acinar cells due to exhibiting characteristics of both.²⁴ Consequently, this poses difficulties in investigating the independent role of ductal cells in the pathologic mechanism of MGD. 

The initial and pivotal stage involves the isolation of MG ducts and the subsequent proliferation of HMGDCs in vitro. It has been observed that the activity of cells diminishes as the duration of treatment with digestive enzymes increases. This study made several attempts to determine the optimal digestion time, followed by the removal of acini to expose the ducts using a stereomicroscope. Considering the tendency of ducts to adhere to pipette tips during pipetting, the tissue culture method was preferred over the single-cell culture. Cultivating floating MG ducts, which are abundant in meibum, was difficult in the past. This study introduces a novel method that significantly contributes to the advancement of floating tissue culture. Additionally, the separation process yielded relatively pure acinar cells. In future investigations, we aim to further explore the distinct characteristics of the duct and acini. 

This study presents novel findings of a solitary MG, marking the first instance of such observation. The enzyme concentration and digestion duration were meticulously confirmed to facilitate future experiments involving the separation of ducts. Additionally, we have introduced a highly effective approach for cultivating duct cells, which can be readily adapted for the cultivation of other types of suspended tissues. 

Acknowledgments

Supported by the National Natural Science Foundation of China (82171025, 82070934), the Fundamental Research Funds for the Central Universities (HUST: 2019kfyXMBZ065), the Key Research and Development Program of Hubei Province (2021BCA146), the Clinical Research Foundation of Wuhan Union Hospital (2021xhlcyj03), and the Hubei Province Health and Family Planning Scientific Research Project (WJ2023M035). 

Disclosure: X. Peng, None; Y.-L. Du, None; S.-T. Liu, None; H. Chen, None; J.-S. Wang, None; C. Wang, None; H.-T. Xie, None; M.-C. Zhang, None