Investigation of Conjunctival Goblet Cell and Tear MUC5AC Protein in Patients With Graves’ Ophthalmopathy

Abstract

Purpose: The aim of this study was to investigate conjunctival goblet cell density (GCD) and tear mucin-5AC (MUC5AC) protein levels in patients with Graves’ ophthalmopathy (GO) and their association with dry eye indicators.

Methods: A total of 99 patients with GO (54 active, 45 inactive) and 40 healthy controls were recruited. Comprehensive ophthalmic examinations, including the external eye, ocular surface, GCD, and tear MUC5AC ELISA, were performed. The GCD examination was performed in temporal bulbar conjunctiva, including IVCM GCD by in vivo confocal microscopy (IVCM) and filled GCD of cytokeratin-7 and MUC5AC-positive co-immunomarkers by impression cytology. Tear MUC5AC protein was detected using samples extracted from Schirmer strips.

Results: The GO group showed a significant decrease in IVCM GCD, filled GCD, and normalized tear MUC5AC protein compared to controls, with the active GO group showing the greatest decrease (all P < 0.05). Tear MUC5AC protein levels in GO correlated with those of IVCM GCD (r = 0.40, P < 0.001) and filled GCD (r = 0.54, P < 0.001, respectively). Higher ocular surface disease index (r = −0.22, P < 0.05; r = −0.20, P < 0.05; r = −0.21, P < 0.05) and lisamine green staining (r = −0.23, P < 0.05; r = −0.38, P < 0.001; r = −0.42, P < 0.001) were associated with lower tear MUC5AC protein levels, IVCM GCD, and filled GCD, respectively, which decreased with increasing clinical activity score (r = −0.24, P < 0.05; r = −0.28, P < 0.01; r = −0.27, P < 0.01) and conjunctival congestion score (r = −0.27, P < 0.01; r = −0.33, P < 0.001; r = −0.42, P < 0.001).

Conclusions: The goblet cell count and tear MUC5AC protein in GO eyes were decreased, possibly due to ocular surface inflammation.

Translational Relevance: This study observed the change of tear film mucin in GO patients.

Introduction

Graves’ ophthalmopathy (GO), also known as thyroid eye disease or thyroid-associated orbitopathy, is an autoimmune inflammatory disease that accounts for most orbital diseases.^1–3 Most patients with GO suffer from Graves’ disease with hyperthyroidism, but GO can also exist in patients with euthyroidism or hypothyroidism caused by Hashimoto thyroiditis.^4–6 Although its pathogenesis remains unknown, current evidence suggests that the orbital tissue and thyroid have cross-reactivity antigens.^7,8 Expression of the thyroid-stimulating hormone receptor and insulin-like growth factor 1 receptor in orbital tissues can be recognized as autoantigens,^9,10 resulting in extraocular muscle hypertrophy, orbital fat, and connective tissue (including lacrimal glands) hyperplasia.^11,12 Clinically, GO mainly presents as exophthalmos, eyelid retraction, eyelid and conjunctival edema, extraocular muscle dysfunction, diplopia, and strabismus.^13,14 The clinical activity score (CAS) is used to evaluate the activity of GO, with CAS ≥ 3/7 for active GO and CAS < 3/7 for inactive GO.¹⁵ 

The TFOS DEWS II Report defines dry eye disease (DED) as a multifactorial ocular surface disease characterized by loss of homeostasis of the tear film accompanied by ocular symptoms, the causes of which include tear film instability and hyperosmolarity, ocular surface inflammation and damage, and neurosensory abnormalities.¹⁶ DED is subdivided into aqueous-deficient dry eye and evaporative dry eye.¹⁶ Ocular surface impairment in patients with GO includes conjunctival hyperemia and chemosis, DED, and exposure keratopathy,^17,18 with DED being the most common disorder, causing eye discomfort in up to 65% to 85% of GO patients.^19,20 

The mechanism underlying DED in patients with GO has not been fully elucidated, but multiple factors have been confirmed to be involved. Exophthalmos, eyelid retraction, and eyelid fissure height increases can lead to accelerated tear evaporation and increased tear osmotic pressure.^11,21 Aqueous tear production may decrease due to continuous inflammatory stimulation²² and lacrimal gland damage of GO.^23,24 Increased incomplete blinking may cause obstructive meibomian gland dysfunction, resulting in elevated evaporation of the tear film.²⁵ Increased corneal keratocyte density and decreased corneal nerve fiber density also worsen the ocular surface of active GO.¹⁷ In addition, a reduction in conjunctival goblet cell density (GCD) has been detected in patients with GO using in vivo confocal microscopy (IVCM) and conjunctival impression cytology.^26,27 

Goblet cells are specialized cells that produce the gel-forming mucin genes MUC5AC, MUC5B, and MUC2.²⁸ Among these, mucin-5AC (MUC5AC) is the most abundant mucin secreted by goblet cells; it can lubricate and protect the ocular surface as a component of the tear film.^29,30 However, although a reduction in GCD was found in patients with GO, it has not been confirmed whether the goblet cell–specific mucin MUC5AC is altered. Thus, in this study, we used IVCM and conjunctival impression cytology to determine whether patients with GO have altered conjunctival GCD and MUC5AC secretion and to explore the factors affecting the changes in GCD and mucin secretion. 

Methods

Participants

This prospective cross-sectional study was conducted in accordance with the tenets of the Declaration of Helsinki and was approved by the ethics committee of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology (2022-0293). Written informed consent was obtained from all participants. A total of 99 outpatients with GO (99 eyes) were included in this study. The control group consisted of 40 healthy volunteers (40 eyes) matched for sex and age. GO was diagnosed by Bartley criteria³¹ and divided into active and groups according to the European Group on Graves’ Orbitopathy (EUGOGO).¹⁵ The exclusion criteria for subjects were as follows: age, <18 years or >70 years; any uncontrolled systemic disease; presence of rheumatic diseases or drug use affecting the ocular surface; history of ocular trauma, surgery, or radiation; contact lens wearing; severe ocular surface disease, lacrimal passage, or eyelid disease; or the use of eye drops. 

Clinical Evaluation

Participants’ sex, age, and the type, duration, and treatment of thyroid disease were recorded. Patients with GO also provided the duration and CAS. Clinical measurements included the following: (1) proptosis and palpebral fissure height (PFH); (2) Ocular Surface Disease Index (OSDI; score 0–100), an effective scale for evaluating the severity of dry eye that assigns higher scores to more severe symptoms; (3) lipid layer thickness (LLT) and partial blinking rate (PBR) measured by a LipiView Ocular Surface Interferometer (Johnson & Johnson Vision Care, Jacksonville, FL; LLT was artificially set to 100 nm, as the instrument cannot display accurate LLT value if they exceed 100 nm); (4) tear meniscus height (TMH), noninvasive breakup time (NIBUT), conjunctival congestion score (CCS), and meibography scores for upper and lower eyelids, which were evaluated with the OCULUS Keratograph 5M (OCULUS, Wetzlar, Germany), with meibography scores calculated according to the meibomian gland dropout area (0, no dropout; 1, dropout area < 1/3; 2, dropout area = 1/3 to 2/3; 3, dropout area > 2/3); (5) Schirmer I test (SIT), measured for 5 minutes without anesthesia to assess tear secretion in each eye; (6) corneal fluorescein staining (CFS), where the cornea was divided into four quadrants and each quadrant was scored on a scale of 0 to 3, with a maximum score of 12; and (7) conjunctival lisamine green staining (LGS), which was scored from 0 to 3 in the exposed nasal and temporal bulbar conjunctiva, with a maximum total score of 6. 

In Vivo Confocal Microscopy

GCD in the temporal bulbar conjunctiva of all participants was measured under IVCM (HRT III Corneal Rostock Module; Heidelberg Engineering, Heidelberg, Germany), as shown in Figure 1. After topical anesthesia (0.4% Benoxil), a sterile disposable plastic cap for the anterior lens of the microscope was applied to the surface of the temporal bulbar conjunctiva about 3 mm from the corneal limbus. The patient was instructed to stare in the opposite direction from the measured area. Carefully scanned high-quality images were obtained of the conjunctival layer where the goblet cells were located. The detection area was within a small range, and target maps of four different positions were obtained. Image J 1.8.0 (National Institutes of Health, Bethesda, MD) was used to calculate the average density of goblet cells based on four images from each eye (expressed as cells/mm²). 

Figure 1.

 

IVCM microscopic images of goblet cells of the temporal bulbar conjunctiva in the control group and GO group. Goblet cells are larger than the surrounding epithelial cells, are round or oval in shape, and have medium or low reflexes (arrows). The IVCM GCD in the GO eye (B, C) was significantly lower than that in the control eye (A), and the IVCM GCD in the active GO eye (C) was lower than that in the inactive GO eye (B). Scale bars: 50 µm.

Mucin 5AC and Cytokeratin 7 Immunofluorescence Staining of Impression Cytology Samples

After the IVCM examination, impression cytology samples were collected from the same temporal bulbar conjunctiva using Millicell Cell Culture Inserts (0.4 µm, PICM01250; MilliporeSigma, Burlington, MA) under topical anesthesia (0.4% Benoxil). The insert was gently pressed on the conjunctiva about 3 mm from the corneal limbus for about 5 seconds.³² The samples were immobilized with 95% ethanol immediately after removal and moistened with phosphate-buffered saline (PBS) at 4°C until immunolabeling. 

The cell membrane was ruptured with 0.3% Triton X-100 for 30 minutes and washed with PBS three times for 5 minutes each time. The cells were then blocked in 10% donkey serum for 1 hour, followed by incubation with mouse monoclonal antibodies against MUC5AC (1:250, ab3649; Abcam, Cambridge, UK) and rabbit monoclonal antibodies against cytokeratin 7 (CK7; 1:250, ab181598; Abcam) overnight at 4°C. The next day, after a PBS wash for 15 minutes, they were incubated with Goat Anti-Mouse IgG H&L (1:500, ab150116; Abcam) and Goat Anti-Rabbit IgG H&L (1:500, ab150077; Abcam) at 37°C for 1 hour. Following a PBS wash for 15 minutes, nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; 28718-90-3; Sigma-Aldrich, St. Louis, MO). The cells were rinsed in PBS and sealed with anti-fluorescence quenching tablets. Images were obtained using a laser scanning confocal microscope (A1 HD25/A1R HD25; Nikon, Tokyo, Japan), and four fields were taken from each stained specimen to calculate the mean value. CK7, MUC5AC, and DAPI are shown in green, red, and blue, respectively, in Figure 2. MUC5AC is a mucin released by goblet cells, and CK7 has been utilized in numerous studies to detect conjunctival goblet cells. Cells that are positive for both CK7 and MUC5AC immunolabeling can represent goblet cells with MUC5AC secreting function.³³ Therefore, the filled goblet cell was identified by simultaneous immunofluorescence of CK7 and MUC5AC in our study, and their densities were calculated by image analysis software (NIS-Elements Viewer 5.21; Nikon). 

Figure 2.

 

Representative images of temporal bulbar conjunctiva impression cytology samples from the control and GO groups. These images were obtained from the eyes of control, inactive GO, and active GO. CK7 antibody was used to identify goblet cells (green), and MUC5AC antibody was used to target the MUC5AC protein (red). DAPI was utilized to visualize cell nuclei. Filled goblet cells were defined as positive for both CK7 and MUC5AC immunolabeling shown in the merged images. Scale bars: 50 µm.

Tear MUC5AC ELISA

The wet portion of each strip after the SIT was stored independently in an Eppendorf tube and immediately frozen at −80°C until analysis. Defrosted Schirmer strips were cut into pieces and stirred with 60 µL PBS in a sterile 1.5-mL microcentrifuge tube at 4°C for 1 hour and then centrifuged at 4°C at 14,000 RPM for 10 minutes to obtain tears. The total protein concentration of each sample was measured using a BCA Protein Analysis Kit (23225; Thermo Fisher Scientific, Waltham, MA). The MUC5AC ELISA was performed according to the manufacturer's protocol (RX104831H; Quanzhou Ruixin Biotechnology Co., Guangzhou, China). The absorbance was measured at 450 nm with an enzyme labeling instrument. Finally, the ratio of MUC5AC protein to total protein was used to normalize the protein level (expressed as ng/mg). 

Statistical Analysis

Statistical analysis was performed using SPSS Statistics 26.0 (IBM, Chicago, IL). Categorical variables were described by frequency and percentage, and comparisons between groups were made using continuity correction, χ² test, or Fisher's exact test. The Shapiro–Wilk test was used to test the normality of the continuous variables, and the Mann–Whitney U test was used to compare the continuous data with non-normal distribution between groups and was represented by the median (first, third quartiles). P < 0.05 was considered to be statistically significant. 

Results

Clinical Data

The demographic and clinical information for all participants is shown in Tables 1 and 2. We enrolled 99 patients with GO (99 eyes; 47 females and 52 males; median age, 53.00 years; range, 30.00–66.00 years), of which 45 eyes exhibited inactive GO and 54 eyes exhibited active GO. Of the patients with GO, 84.85% (84/99) had thyroid dysfunction, of whom 15 were treated with levothyroxine and 63 with antithyroid drugs (methimazole, 52; propylthiouracil, 11); six did not receive medication. We also recruited 40 healthy controls (40 eyes; 24 females and 16 males; median age, 53.50 years; range, 39.00–66.00 years). There were no significant differences in age or sex between the control and GO groups and no difference in disease duration between the inactive and active GO groups (all P > 0.05). 

Table 1.

 

Clinical Data for the GO and Control Groups

 

Table 2.

 

Comparison of Clinical Data Between Active and Inactive GO

 

Proptosis, PFH, and OSDI were more severe in patients with GO than in controls (all P < 0.001), with active GO being more severe than inactive GO (all P < 0.05). Patients with GO had significantly lower LLT, TMH, NIBUT, and SIT parameters than the controls, whereas PBR, meibography scores, CFS, LGS, and CCS were significantly higher (all P < 0.05). Except for PBR, TMH, NIBUT, and SIT (all P > 0.05), other ocular surface parameters were significantly different between active and inactive GO, with active GO having higher meibography scores; higher CFS, LGS, and CCS (all P < 0.05); and lower LLT (P < 0.05). 

Measurement of Goblet Cell Density

GCD measured using IVCM was referred to as the IVCM GCD. The IVCM GCD of the temporal bulbar conjunctiva in the GO group was 162.50 (125.00, 209.38) cells/mm², significantly lower than that in the control group, with 289.07 (260.16, 326.56) cells/mm² (P < 0.001) (Table 1, Fig. 3A). In addition, the active GO group had a greater loss of goblet cells than the inactive GO group, as shown by the lower IVCM GCD (active GO, 153.91 [105.86, 196.10] cells/mm²; inactive GO, 173.44 [142.19, 235.94] cells/mm²; P = 0.008) (Table 2, Fig. 3D). 

Figure 3.

 

Comparison of goblet cell density and tear MUC5AC protein between the control and GO groups and between the GO groups. IVCM GCD (A, D) was examined by confocal microscopy in vivo, and filled GCD (B, E) was measured by immunofluorescence using impression cytology samples. The normalized tear MUC5AC protein (C, F) was determined by ELISA (expressed in ng/mg). *P < 0.05, **P < 0.01, ***P < 0.001.

Similar results were observed in filled goblet cells co-immunomarkers CK7 and MUC5AC. The filled GCD of participants in the control group (271.00 [252.07, 342.41] cells/mm²) was significantly higher than in the GO group (141.60 [109.86, 195.31] cells/mm²; P < 0.001) (Table 1, Fig. 3B). The differences among the patient groups were also significant (active GO, 131.84 [99.49, 184.94] cells/mm²; inactive GO, 158.69 [124.52, 220.95] cells/mm²; P = 0.004) (Table 2, Fig. 3E). 

Tear MUC5AC Protein

Compared with the control group, the normalized tear MUC5AC protein level in the GO group was significantly decreased (Table 1, Fig. 3C), and that in the active GO group was significantly lower than that in the inactive GO group (Table 2, Fig. 3F). In all patients with GO, the tear MUC5AC protein level was significantly correlated with IVCM GCD (r = 0.40, P < 0.001) and filled GCD (r = 0.54, P < 0.001). Additionally, there was a significant negative correlation between tear MUC5AC level and OSDI (r = –0.22, P < 0.05), CAS (r = –0.24, P < 0.05), LGS (r = –0.23, P < 0.05), and CCS (r = –0.27, P < 0.01). Moreover, the IVCM GCD and filled GCD were negatively correlated with CAS (r = –0.28, P < 0.01; r = –0.27, P < 0.01, respectively), OSDI (r = –0.20, P < 0.05; r = –0.21, P < 0.05, respectively), CCS (r = –0.33, P < 0.001; r = –0.42, P < 0.001, respectively), and LGS (r = –0.38, P < 0.001; r = –0.42, P < 0.001, respectively) (Fig. 4). 

Figure 4.

 

Spearman correlation matrix between clinical parameters and goblet cell density and tear MUC5AC in all patients with GO. *P < 0.05, **P < 0.01, ***P < 0.001.

Discussion

In this study, we observed a reduction in the density of conjunctival goblet cells in patients with GO, including the IVCM GCD, using in vivo confocal microscopy and filled GCD assessed by positive co-immunolabeling of CK7 and MUC5AC on conjunctival impression cytology samples. We also found that the levels of normalized tear MUC5AC protein in patients with GO were lower than those in the controls. The IVCM GCD, filled GCD, and tear MUC5AC levels were lower in the active GO group than in the inactive GO group. Additionally, tear MUC5AC protein levels in GO correlated with IVCM GCD and filled GCD. To our knowledge, our study is the first to assess the number of goblet cells with mucin-secreting function using co-immunolabeling in GO patients and to evaluate changes in tear MUC5AC protein levels in GO patients, as well as to analyze differences in the number and function of goblet cells in patients with active and inactive GO. 

DED occurs in 65% to 85% of patients with GO, but its exact pathogenesis is unclear. A variety of factors may change the composition of the tear film, resulting in a homeostasis imbalance in the ocular surface microenvironment of patients with GO and increasing the occurrence of dry eye. In patients with GO, decreased tear production and increased tear evaporation may also result in an impaired ocular surface. The tear function test and NIBUT were assessed in the GO and control groups. The findings showed that patients with GO had reduced tear secretion and tear film stability compared with controls, consistent with other investigations.^20,27 Previous research on patients with GO has revealed considerable PBRs and loss of meibomian glands.^34,35 The alterations of meibomian glands in patients with GO may be intimately related to PBR and CAS.²⁰ This notion is further supported by the results of our study, which showed that patients with GO had higher PBRs, greater meibomian glands dropout, and thinner lipid layers. Additionally, active GO showed a significantly higher loss of meibomian glands and lipid layer than inactive GO. Therefore, we suggest that the loss of meibomian glands and modifications to the lipid layer may also accelerate the evaporation of tears in patients with GO, in addition to the increased exophthalmos and palpebral fissure.^21,34 

The TOFS DEWS II Report states that evaporative dry eye may be related to a decline in mucin.¹⁶ Approximately 20 human mucins have been identified that are classified as secreted mucins and membrane-associated mucins according to their structure and function, with the former including gel-forming mucins and soluble mucins.³⁶ Human conjunctival goblet cells are distributed as single cells in the conjunctival epithelium and secrete gel-forming mucin MUC5AC.³⁷ Gel-forming mucins convert aqueous tears into mucoaqueous gels through their water-binding capacity, maintain moisture in the ocular surface, and provide lubrication.³⁸ A decrease in the number of goblet cells or mucins levels may exacerbate ocular surface damage. Therefore, studying goblet cells and mucins in patients with GO is crucial. 

As in previous studies, we used IVCM to examine GCD alterations in patients with GO. IVCM has been an effective tool for evaluating the cornea,³⁹ conjunctiva,^27,40 and meibomian glands^20,41 of patients with GO. Through confocal microscopy, Wei et al.²⁷ demonstrated a decrease in GCD in the GO group, which was consistent with the findings of this investigation. Our research revealed that the IVCM GCD in the GO group was much lower than that in the control group, with a greater loss of goblet cells in the active GO. To verify the changes in the secretory function of goblet cells in patients with GO, impression cytology samples were used in this study. Goblet cells can be specifically stained by CK7 antibody.⁴² The identification of goblet cell–secreted products with periodic acid–Schiff (PAS), helix pomatia agglutinin (HPA), or MUC5AC antibody has been used to evaluate ocular surface health.^33,43,44 Using the co-immunomarkers of CK7 and MUC5AC antibodies, we discovered that patients with GO showed a decrease in filled GCD, which was accompanied by a decrease in the level of tear MUC5AC protein. Active GO has lower filled GCD and tear MUC5AC protein levels than inactive GO. Furthermore, greater dry eye severity and conjunctival staining score were found to be associated with lower IVCM GCD, filled GCD, and tear MUC5AC levels, which decreased with increased CAS or CCS. 

Inflammation has been recommended as a stable indicator of DED severity, as it promotes the pathophysiological mechanism of DED.⁴⁵ Studies have shown that the number of conjunctival goblet cells is reduced in aqueous-deficient dry eye⁴⁶ and some ocular surface inflammatory diseases (such as Sjögren syndrome, Stevens–Johnson syndrome, and graft-vs.-host disease).^47,48 Additionally, it has been observed that inflammatory cytokines are associated with the loss of goblet cells in the conjunctiva. The proinflammatory cytokine interferon (IFN)-γ secreted by Th1 cells can cause squamous metaplasia of the ocular surface epithelial cells and a decrease in GCD.⁴⁹ GCD was reduced as a result of the increased IFN-γ/interleukin (IL)-13 ratio.⁵⁰ Goblet cell apoptosis has also been found to be induced by tumor necrosis factor (TNF)-α and IFN-γ in animal experiments.⁵¹ Multiple cytokines, including IL-1β, IL-6, IL-7, IL-8, IL-10, IL-13, IL-17A, and TNF-α, are known to be highly elevated in GO tears, some of which are positively correlated with CAS.⁵² CAS and CCS, which can macroscopically represent the severity of ocular surface inflammation in patients with GO, were strongly correlated with GCD and tear MUC5AC protein in our study. Therefore, we hypothesized that ocular surface inflammation may be significantly related to conjunctival goblet cell apoptosis and decreased mucin secretion, which may be attributed to the changes in cytokines levels in patients with GO. 

This study had some limitations. First, the goblet cells were examined only in the temporal bulbar conjunctiva to minimize patient discomfort. Second, due to the small sample size of tears obtained, we did not detect markers of oxidative stress and inflammation on the ocular surface and could not directly evaluate the correlation among cytokines, goblet cells, and mucin secretion. We aimed to explore the possible mechanism of goblet cell apoptosis and reduced mucin secretion in patients with GO by studying quantitative markers. Experimental model studies were used to elucidate pathophysiological mechanisms and specific pathways. 

Conclusions

We found a decrease in GCD in GO eyes both in vivo and in vitro, especially in active GO. The GCD evaluated by IVCM and co-immunomarkers in patients with GO was significantly correlated with the tear MUC5AC protein levels, making it a reliable method to study goblet cell count and function. Decreased GCD and tear MUC5AC protein in patients with GO resulted in greater dry eye symptoms and conjunctival staining scores and were strongly associated with increased levels of inflammation (CAS and CCS). 

Acknowledgments

Supported by a grant from the National Natural Science Foundation of China (81900912). 

Disclosure: Y. You, None; J. Chen, None; H. Chen, None; J. Wang, None; H. Xie, None; X. Pi, None; X. Wang, None; F. Jiang, None