Experimental Eye Research 76 (2003) 417–420 www.elsevier.com/locate/yexer
Minocycline effect on meibomian gland lipids in meibomianitis patients W.E. Shinea,*, J.P. McCulleya, A.G. Pandyab a
Department of Ophthalmology, The University of Texas, Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd., Dallas, TX 75235-9057, USA b Department of Dermatology, The University of Texas, Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd., Dallas, TX 75235-9057, USA Received 9 September 2002; accepted in revised form 20 December 2002
Abstract The objective of this research was to determine the effect of oral minocycline on the meibomian gland nonpolar and free fatty acid lipids of chronic blepharitis patients. Patients – seborrheic blepharitis (SBBL), acne rosacea (AR) without ocular involvement, and acne rosacea with meibomianitis (AR-MKC). Minocycline treatment – 50 mg orally for 2 weeks followed by 100 mg to the end of 3 months; this was followed by 3 more months with no treatment. Meibomian gland secretions (meibum) were collected before treatment, at the end of the 3 months on treatment, and 3 months after stopping treatment. Lipids were separated and analyzed for wax and sterol esters, triglycerides, diglycerides, free cholesterol and free fatty acids. Data were analyzed statistically by ANOVA. Minocycline treatment resulted in decreased diglycerides and free fatty acids in the group AR-MKC, which continued into the second 3 months (off treatment) and was significant. Cholesterol decreased, but triglycerides initially decreased with treatment and then increased when treatment in the group was discontinued (second 3 months); these results, however, were not significant. Thus, minocycline has its greatest effect on lipid types, which result from degradation (lipase) reactions, suggesting a lipase inhibition effect and/or direct effect on ocular flora. This minocycline effect continues even after treatment is discontinued, suggesting a more lasting effect on ocular microflora. Minocycline may be most effective when the treatment period is longer than 3 months. These results give insight into disease mechanisms associated with chronic blepharitis. q 2003 Elsevier Science Ltd. All rights reserved. Keywords: blepharitis; meibomianitis; minocycline; lipids; bacteria
1. Introduction Chronic blepharitis is one of the most difficult ocular diseases to treat. We have determined that there are six chronic blepharitis groups; they are staphylococcal, seborrheic alone, seborrheic with an associated staphylococcal infection, seborrheic with meibomian seborrhea, seborrheic with spotty inflammation of the meibomian glands, and primary meibomianitis (severe inflammation of the meibomian glands) (McCulley et al., 1982). Two-thirds of the patients with primary meibomianitis have the associated skin condition acne rosacea. Also a type of chronic blepharitis designated meibomian gland dysfunction (MGD) has been associated with abnormalities of keratinization (Gutgesell et al., 1982); it is not clear however, what percentage of chronic blepharitis patients present with this condition. Tetracycline analogues have been shown to be effective in treating primary meibomianitis. It has been demonstrated * Corresponding author. Dr Ward E. Shine, Department of Ophthalmology, The University of Texas, Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd., Dallas, TX 75235-9057, USA. E-mail address: [email protected]
that tetracycline can inhibit lipase activity and therefore decrease the release of noxious free fatty acids (Gutgesell et al., 1982). Previously we have shown that tetracycline results in decreased bacteriological lipase activity in vitro (Dougherty et al., 1991), as have others (Gulbenkian et al., 1980). We have been investigating the associations between chronic blepharitis and eyelid meibomian gland lipids and microflora for a number of years and have discovered important relationships between these lipids and chronic blepharitis disease states. Here we extend these studies to determine what effect oral minocycline treatment has on meibomian gland lipids (meibum) of primary meibomianitis patients.
2. Materials and methods 2.1. Selection of patients Patients were selected based on clinical appearance and categorized as having acne rosacea with or without
0014-4835/03/$ - see front matter q 2003 Elsevier Science Ltd. All rights reserved. DOI:10.1016/S0014-4835(03)00005-8
W.E. Shine et al. / Experimental Eye Research 76 (2003) 417–420
meibomianitis, or seborrhea blepharitis alone (McCulley et al., 1982). Informed consent for each subject was obtained using a consent form approved by the Institutional Review Board of the University of Texas Southwestern Medical Center at Dallas. All study methods conformed to the Declaration of Helsinki. Minocycline was administered orally, 50 mg daily for 2 weeks followed by 100 mg daily to the end of the three month treatment period; no other treatment was permitted during the six months of the study. The treatment groups were seborrheic blepharitis (SBBL, 2 patients), acne rosacea without ocular involvement (AR, 2 patients) and acne rosacea with meibomianitis (AR-MKC, 6 patients). 2.2. Meibum collection Meibum (meibomian gland secretion) was collected as previously reported (Osgood et al., 1989) before minocycline treatment was begun, at the end of the three month treatment period (i.e. while still taking minocycline), and three months after minocycline treatment was stopped. This collection method utilized a lid conformer and a cotton swab to gently press the lid; the excreted lipids were collected with a spatula. These lipids were then dissolved in a small volume of chloroform-metholol solvent. 2.3. Assays Meibum was separated by solid phase extraction on an aminopropyl column. Nonpolar lipids (wax and sterol esters, free cholesterol, triglycerides, diglycerides) and free fatty acids were collected and the nonpolar lipids were further separated by high-pressure liquid chromatography (HPLC) on a cyanopropyl column with UV detection. The free fatty acids were methylated and then analyzed by gas chromatography-mass spectroscopy (GC-MS).
2.4. Data analysis The data were analyzed statistically (ANOVA) and significant treatment differences ( p , 0·05) were determined by the Tukey honest significant difference (HSD) method. These and the nonparametric Kruskal – Wallis ANOVA were utilized to test for significance (Statistica, 1995).
3. Results Important differences in treatment effect on certain nonpolar lipids were observed. Differences in the effect of minocycline treatment on the relative amounts of wax and sterol esters in the three disease groups were noted (Table 1). The AR-MKC group’s esters were somewhat lower than in the other two groups before treatment was begun, and continued lower during the 6 months. The greatest treatment effects were observed in the diglycerides, free fatty acids and cholesterol (Table 1). Before treatment diglycerides were highest in the SBBL group. However, at the end of the treatment period (3 months) all three groups’ diglycerides were lower than at the beginning; 3 months after minocycline treatment was discontinued (6 months) diglycerides were even lower (45% of the initial value in the SBBL and AR-MKC groups). In contrast to the diglycerides, the free fatty acids initially were lowest in the SBBL group (and highest in the AR and AR-MKC groups). Three months after treatment was discontinued (6 months) free fatty acids were much lower (35%) than the initial value for the groups AR and AR-MKC; however in the SBBL group there was no decrease. A similar pattern was observed for free cholesterol – it increased in the SBBL group but decreased to 5% of the initial value in the groups AR and AR-MKC. On the other hand, although triglycerides increased somewhat in the SBBL group, in the AR-MKC they decreased only at the
Table 1 Effect of minocycline on meibum nonpolar lipids and degradation products Lipid
W/ch TG DG FFA CH
After 3 months
After 6 months
After 3 months
After 6 months
After 3 months
After 6 months
100 4·8 3·4 1·3 0·03
60 7·5 2·2 ND 1·9
67 7·1 1·4 1·5 0·68
100 5·9 1·8 2·9 0·66
51 4·9 1·6 ND 0·03
76 5·9 1·2 0·82 0·03
100 13 2·7 2·5 5·5
76 8·0 1·9a 1·2 2·9
55 11 1·2a 0·88a 0·20b
Disease conditions: SBBL, seborrheic blepharitis; AR, acne rosacea; AR-MKC, AR with meibomian keratoconjunctivitis. Treatment: Before, before beginning minocycline treatment; 3 months ¼ 3 months minocycline treatment; 6 months ¼ 3 months followed by 3 months without treatment. Lipids: W/ch, wax and cholesterol esters; TG, triglycerides; DG, diglycerides; FFA, free fatty acids; and CH, cholesterol. a Significantly different ( p , 0·05) from corresponding before minocycline (before) values. Unmarked values were not significant ( p . 0·05), except for the AR-MKC DG and FFA (before) values. b Significantly different ( p , 0·05) from the 3-month value as determined by the Kruskal–Wallis test.
W.E. Shine et al. / Experimental Eye Research 76 (2003) 417–420
end of treatment and then increased after treatment was discontinued; they remained unchanged in the AR group. However, only the AR-MKC diglyceride 3 months values and the 6 month values for diglycerides and free fatty acids were significantly different from the initial (before) values. Although AR-MKC cholesterol values were not significantly different, the nonparametric Kruskal – Wallis test showed a significant difference ( p , 0·05) between the 3-month and 6-month values (Table 1). Because of the small number of replicates in the SBBL and AR groups, statistical comparisons of lipids in these groups was not meaningful.
4. Discussion The results suggest that one effect of minocycline treatment is to decrease the meibomian gland lipid degradation products free fatty acids, diglycerides and perhaps free cholesterol. The formation of these products is consistent with the enzymatic action of lipases and esterases on triglycerides, cholesterol esters, and fatty wax esters, which are major constituents of meibomian secretions. The effects of these free fatty acids, especially the monoand polyunsaturated ones derived from triglycerides (Shine and McCulley, 1996), are quite important. 4-hydroxynonenal, formed from linoleic acid, is a chemoattractant for neutrophils, thus promoting inflammation (Schaur et al., 1994). Furthermore, not only polyunsaturated fatty acids but also monounsaturated fatty acids such as oleic acid stimulate neutrophils to produce inflammatory mediators collectively called reactive oxygen species (ROS) (Li et al., 1996). Importantly, minocycline treatment not only decreases free fatty acid production, as in the present study (Table 1), but also significantly decreases ROS produced by neutrophils (Akamatsu et al., 1991). Thus minocycline affects inflammation in two ways, first by decreasing neutrophil recruitment and secondly by decreasing formation of ROS by neutrophils. Theoretically, free fatty acids can also destabilize the preocular tear film. The second degradation product, diglycerides, can also have an effect, although indirectly, on the inflammatory process. As an activator of protein kinase C, which in turn activates phospholipase A2 (PLA2), diglycerides can cause the release of the proinflammatory polyunsaturated fatty acid, arachidonic acid, from neutrophil phospholipids (Seeds et al., 1998). Again, minocycline has been reported to inhibit PLA2 activity (Pruzanski et al., 1992). Another breakdown product is free cholesterol that can be formed from cholesterol esters (present in meibum) by esterases. Also, we have previously shown that free cholesterol, stimulates bacterial growth in vitro (Shine et al., 1993). Finally, abnormal keratinization of the meibomian gland is observed with some chronic blepharitis patients (MGD). It should be noted however that keratin proteins have been identified in both normal and patient meibum (Ong et al., 1991). The causes of hyperkeratinization are important
since one result appears to be the plugging of the meibomian gland orifice. A detailed study has related epinephrine treatment to hyperkeratinization (Jester et al., 1989). More recent investigations in other tissues have related the epinephrine effect to an increase in cellular calcium resulting in keratinocyte differentiation (Schallreuter et al., 1995). Another more related study has demonstrated that in human epidermal keratinocytes, even at low calcium levels, oleic acid is an effective activator of keratinocyte differentiation (Hanley et al., 1998). We and others have reported high oleic acid levels in both meibum triglycerides (Shine and McCulley, 1996) and free fatty acids (Dougherty and McCulley, 1986); as we have reported here (Table 1) both triglycerides and free fatty acids in the AR-MKC group were initially much higher than in the AR and SBBL groups. With minocycline treatment free fatty acids were much lower in the AR-MKC group. We suggest that there are numerous causes of hyperkeratinization of the meibomian gland duct but one possible cause, high free oleic acid, can be controlled with minocycline treatment. Since the ultimate destination of meibum is the lipid layer of the tear film, the effect of large changes in nonpolar lipids and free fatty acids should be considered. We would expect the greatest effect from the free fatty acids and free cholesterol. For example, free cholesterol can stimulate staphylococcus growth (Shine et al., 1993). High cholesterol could also result in a lipid layer with a much more rigid polar lipid phase than normal. On the other hand a large proportion of the shorter chain free fatty acids could leave the lipid phase and enter the aqueous phase, thus changing its polarity. As we have reported (Table 1) these detrimental effects can be counteracted with minocycline treatment. The results of this study suggest how tetracycline type drugs may be effective in treating some types of chronic blepharitis. Although adverse events do occur with minocycline, we observed few. In fact, no adverse events were encountered in this (AR-MKC) treatment group, such as nausea, diarrhea, problematic photosensitivity, etc. However, the possibility (even at the 100 mg level) should be considered (Physicians’ Desk Reference, 2002). The results also suggest other treatments such as topical flavonoids with free-radical scavenging ability, e.g. silymarin (De La Puerta et al., 1996).
Acknowledgements Supported in part by an unrestricted research grant from Research to Prevent Blindness, Inc., New York, NY, USA, and NIH EY12430.
References Akamatsu, H., Niwa, Y., Kurokawa, I., Masuda, R., Nishijima, S., Asada, Y., 1991. Effects of subminimal inhibitory concentrations of minocycline
W.E. Shine et al. / Experimental Eye Research 76 (2003) 417–420
on neutrophil chemotactic factor production in comedonal bacteria, neutrophil phagocytosis and oxygen metabolism. Arch. Dermatol. Res. 283, 524–528. De La Puerta, R., Martinez, E., Bravo, L., Ahumada, M.C., 1996. Effect of silymarin on different acute inflammation models and on leukocyte migration. J. Pharm. Pharmacol. 48, 968– 970. Dougherty, J.M., McCulley, J.P., 1986. Analysis of the free fatty acid component of meibomian secretions in chronic blepharitis. Invest. Ophthalmol. Vis. Sci. 27, 52–56. Dougherty, J.M., McCulley, J.P., Silvany, R.E., Meyer, D.R., 1991. The role of tetracycline in chronic blepharitis. Invest. Ophthalmol. Vis. Sci. 32, 2970–2975. Gulbenkian, A., Myers, J., Fries, D., 1980. Hamster flank organ hydrolase and lipase activity. J. Invest. Dermatol. 75, 289 –292. Gutgesell, V.J., Stern, G.A., Hood, C.I., 1982. Histopathology of meibomian gland dysfunction. Am. J. Ophthalmol. 94, 383–387. Hanley, K., Jiang, Y., He, S.S., Friedman, M., Elias, P.M., Bikle, D.D., Williams, M.L., Feingold, K.R., 1998. Keratinocyte differentiation is stimulated by activators of the nuclear hormone receptor PPARalpha. J. Invest. Dermatol 110, 368–375. Jester, J.V., Nicolaides, N., Kiss-Palvolgyi, I., Smith, R.E., 1989. Meibomian gland dysfunction II. The role of keratinization in a rabbit model of MGD. Invest. Ophthalmol. Vis. Sci. 30, 936–945. Li, Y., Ferrante, A., Poulos, A., Harvey, D.P., 1996. Neutrophil oxygen radical generation. Synergistic responses to tumor necrosis factor and mono/polyunsaturated fatty acids. J. Clin. Invest. 97, 1605–1609. McCulley, J.P., Dougherty, J.M., Deneau, D.G., 1982. Classification of chronic blepharitis. Ophthalmology. 89, 1173–1180.
Ong, B.-L., Hodson, S.A., Wigham, T., Miller, F., Larke, J.R., 1991. Evidence for keratin proteins in normal and abnormal human meibomian fluids. Curr. Eye. Res. 10, 1113–1119. Osgood, J.K., Dougherty, J.M., McCulley, J.P., 1989. The role of wax and sterol esters of meibomian secretions in chronic blepharitis. Invest. Ophthalmol. Vis. Sci. 30, 1958–1961. Physicians’ Desk Reference, 2002. Medical Economics Company, Montvale, NJ, p. 1863–1865. Pruzanski, W., Greenwald, R.A., Street, I.P., Laliberte, F., Stefanski, E., Vadas, P., 1992. Inhibition of enzymatic activity of phospholipases A2 by minocycline and doxycycline. Biochem. Pharmacol. 44, 1165– 1170. Schallreuter, K.U., Lemke, K.R., Pittelkow, M.R., Wood, J.M., Korner, C., Malik, R., 1995. Catecholamines in human keratinocyte differentiation. J. Invest. Dermatol. 104, 953 –957. Schaur, R.J., Dussing, G., Kink, E., Schauenstein, E., Posch, W., Kukovetz, E., Egger, G., 1994. The lipid peroxidation product 4-hydroxynonenal is formed by—and is able to attract—rat neutrophils in vivo. Free Rad. Res. 20, 365 –373. Seeds, M.C., Nixon, A.B., Wykle, R.L., Bass, D.A., 1998. Differential activation of human neutrophil cytosolic phospholipase A2 and secretory phospholipase A2 during priming by 1,2-diacyl- and 1O-alkyl-2-acylglycerols. Biochim. Biophys. Acta. 1394, 224–234. Shine, W.E., McCulley, J.P., 1996. Meibomian gland triglyceride fatty acid differences in chronic blepharitis patients. Cornea 15, 340 –346. Shine, W.E., Silvany, R., McCulley, J.P., 1993. Relation of cholesterolstimulated staphylococcus aureus growth to chronic blepharitis. Invest. Ophthalmol. Vis. Sci. 34, 2291–2296. Statistica, 1995. General Conventions and Statistics I. Statsoft, Tulsa, OK.