Lubricity is a basic requirement between 2 interfaces in relative motion, allowing them to slide smoothly against one other. The coefficient of friction (CoF) is used to describe the lubricity of interfaces and is defined as the ratio between friction and normal force.1 A low CoF value therefore indicates high lubricity. Tribology is the science of surface interactions during relative motion. Recently, microtribometers have been used in several studies in determining the CoF of biological tissues or highly lubricious materials, such as contact lenses (CLs).2–5 Advantages of microtribometer use are the ability to measure frictional force resolutions, to analyze microscopic sized samples, and to apply biologically relevant normal force ranges.
Roba et al5 developed an optimized method for CL CoF testing by determining a combination of parameters including lubricant solution [tear-mimicking solution (TMS) containing lysozyme and serum], contact pressure range (2.7 and 19.7 kPa), and counter surface composition (hydrophobized mucin-coated glass) that are suggested to be biologically relevant to the ocular environment. More recently, the CL CoF and comfort have been correlated, indicating an inverse relationship between them.3,6,7 Hence, determining the CoF of human corneas would provide a suitable comparator to CL CoF values and may facilitate the development of more comfortable lenses.8 Intuitively, the lubricity of the corneal tissue is expected to be as high as other naturally lubricated systems, such as natural articular joints.2
The CoF of corneal tissue has been the focus of previous studies.9,10 Cobb et al9 and Dunn et al10 both measured CoF values of monolayers of SV-40 immortalized live human corneal epithelial cells (HCE-T), reporting a direct relationship between the CoF and tribo-stress induced cell damage with a normal force range of 0.5 mN to 6 mN and 500 μN to 2 mN, respectively. Meyer et al,11 who used umbilical vein graft tissue as a surrogate for corneal tissue in CoF testing, suggested that cell cultures are not sufficiently robust for most friction tests. Other studies have used whole preserved human corneas.8,12–14 However, the corneal preservation medium does not prevent epithelial degradation and leads to changes in lubricity and other tissue qualities.15
To derive CoF values that may be correlated to future comfort studies, the use of a tear film–mimicking solution is sine qua non. The tear film overlays the corneal epithelium, forming a protective and refractive surface, as well as providing lubrication by forming a mucin-rich surface brush.16–18 Polymer brushes, both natural and man-made, are high water-containing thin films that have been recognized to decrease the frictional forces between 2 surfaces.19,20 Meyer et al11 determined that the lubricant involved in the interaction between 2 interfaces can alter the measured CoF. Cobb et al9 and Dunn et al10 used HEPES-buffered DMEM/F12 medium as the lubricant in CoF testing. However, HEPES buffer solutions have been associated with adverse effects on the corneal tissue resulting in tissue swelling.21 Gorbet et al22 suggested that corneal cell viability is adversely affected when borate-based buffers are used. Additionally, several other investigations suggested that borate can react with components present in CL materials, such as polyvinyl alcohol.23–25 This potential reaction may influence the ability to compare the cornea and CL CoF values. Literature review indicates the lack of a standard solution for use in CoF measurement procedures.
Determining the CoF of fresh human corneal tissue would fill a gap in the current knowledge of ocular tribology and would test the applicability (and sensitivity) of the devised technique. Furthermore, knowing the relationship between the cornea and CL CoF could allow for the development of products (ie, CLs) that are better suited for use on the human eye. Thus, the purpose of this study, which is largely unprecedented, was to measure the CoF of fresh human corneas and investigate the effects of different solutions on the tribological tissue response.
MATERIALS AND METHODS
Selection and Handling of Tissue Samples
Fresh human whole globes were provided by Lions VisionGift (Portland, OR) in a moist container. All samples were accompanied by a Non-Transplant Donor Information Report, which included the time of death, time of ocular cooling, and time of global enucleation (performed by Lions VisionGift). A medical and social history interview was conducted for each donor. Tissue samples from donors with a history of any of the following were excluded from this investigation: Sjörgen syndrome, keratoconus, corneal dystrophies, previous corneal surgery, any ocular condition for which eye drops were used on a daily basis (eg, glaucoma, severe allergies), and smoking. Lions VisionGift is an eye bank accredited by the Eye Bank Association of America. Informed consent was obtained for all samples, and the study was conducted in compliance with the Good Tissue Practice Guidance document.
Fresh corneal tissue was defined as a whole-globe enucleation time of ≤7 hours from the time of donor death. The minimum acceptable tissue quality was defined as an intact epithelium, which was verified by slit-lamp examination. Within a mean of 6 hours, 15 minutes (range: 3 hours, 30 minutes to 8 hours, 48 minutes), the corneal tissue was excised based on standard operating procedures by an eye bank–qualified technician at Lions VisionGift immediately before mounting and CoF measurement, as described below. After the excision, the cornea was handled using sterile tweezers clamped at the outer border to prevent damage to the measurement area. Because only 1 sample could be measured at a time, the time intervals for left eye corneal tissue measurements were on average 1.5 hours longer than those for right eye tissue measurements.
Test Solution Preparation
For the evaluation of the CoF in this study, 4 different test solutions were used: TMS in borate-buffered saline (PS) (provided by Johnson & Johnson VisionCare, Inc, Jacksonville, FL; see Table, Supplemental Digital Content 1; http://links.lww.com/ICO/A300), TMS in phosphate-buffered saline (PBS) [Dulbecco phosphate-buffered saline (Sigma-Aldrich, St Louis, MO); see Table, Supplemental Digital Content 2, http://links.lww.com/ICO/A301], TMS in HEPES-buffered saline (TMS-HEPES) [0.01 M solution, diluted from a 1 M solution (Sigma-Aldrich); see Table, Supplemental Digital Content 3, http://links.lww.com/ICO/A302], and tear-like fluid in PBS (TLF-PBS) (see Table, Supplemental Digital Content 4, http://links.lww.com/ICO/A303). For the TMS, serum [PreciNorm U Plus (COBAS INTEGRA/Roche Diagnostics, Mannheim, Germany)] was diluted in the selected buffer along with lysozyme [muramidase from chicken protein (Sigma-Aldrich)]. The TLF-PBS, which contains a mix of proteins, mucin, and added lipids, was provided by Johnson & Johnson VisionCare, Inc. Testing solutions were prepared before study initiation and were used sequentially as presented in supplementary tables. The osmolarity range was 420 to 430 mOsm/kg for the TMS-PS solution and was 300 to 310 mOsm/kg for the TMS-PBS, TMS-HEPES, and TLF-PBS solutions, as measured using an osmometer.
Instrument and Setup
All CoF measurements were performed using the methodology described by Roba et al.5 A Basalt Must brand microtribometer (Tetra GmbH, Ilmenau, Germany) was used to perform the friction tests. The applied normal force (FN) range was 0.2 to 4.0 mN, with a linear stroke length of 1.0 mm and a sliding speed of 0.1 mm/s. The instrument, setup, and counter surface preparation protocol used for the friction tests in this study have been described in detail by Roba et al.5
Sample Handling and Mounting
For all CoF measurements, the corneal tissue was carefully aligned on a cast silicone-rubber [polyvinylsiloxane (PVS), Provil Novo, Germany] sample holder (see Figure, Supplemental Digital Content 1, http://links.lww.com/ICO/A304), ensuring that no air was trapped between the holder and cornea. The PVS holder was then placed on a polyacetal copolymer holder and fixed with a separate PVS-stabilizing ring, similar to the method used for CL measurements by Roba et al.5 The cavity in the center of the ring was filled with the test solution to prevent the cornea from dehydrating. The polyacetal copolymer holder was fixed to the microtribometer using 2 screws. The CoF was determined individually for each cornea using the testing equipment; the right cornea was tested first when 2 corneas were accepted from a single donor.
Normal Forces and Contact Pressure
As previously demonstrated by Roba et al,5 the use of different FN allows for the calculation of the CoF as the slope of the linear regression of the absolute values for the applied FN versus measured lateral force plots. Similar to the method used with CLs,5 the contact area and the contact pressure between the spherical corneal tissue and the mucin-coated hydrophobic flat glass surface were estimated using the Hertzian contact model as described by Chaudhri and Yoffe.26 The radius of the corneal tissue was assumed to be the same as that of the CL holder used to cast the tissue holder (7 mm). The Poisson ratio (ν) and E-modulus were taken from the data published by Knox Cartwright et al.27 The resulting calculated contact pressure ranged between 3.5 kPa (at 0.2 mN) and 8.8 kPa (at 4 mN). The values calculated were in good agreement with eyelid contact pressures reported previously.28
Measurement Procedure and Data Evaluation
For each of the 6 applied normal loads, 2 sliding movements were performed, while recording the resulting lateral forces. For the evaluation of the CoF, the average of 50 points in the constant region was taken for both the trace and retrace track of the second sliding movement. The averages of the data were then used to generate the tangential force (FT) values at given FN values that served as the basis (“CoF-ramp”) for the calculation of the linear regression, the slope of which represented the CoF.
Design of the Measurement Program and Effect of Aging on the Corneal Tissue
In a preliminary phase of the project, the influence of cycling and FN on measured CoF values was investigated. First, the effect of mechanical aging on CoF measurements was evaluated. Similar to the work by Roba et al,5 we aimed to determine whether changes in the lenses due to uptake from the simulated tear solutions occurred during measurement or whether damage to the tissue occurred during measurement. To evaluate this, 50 cycles at 2-mN force were applied after the initial CoF measurement, and then another CoF measurement was taken. In all cases (data on file), the CoF after 50 cycles at 2 mN drastically increased compared with the first CoF measurement. When the FT versus position plot for the corneal tissue was compared with that for a CL (in this case, Acuvue Oasys, Johnson & Johnson Vision Care, Inc; see Figure, Supplemental Digital Content 2, http://links.lww.com/ICO/A305), a clear change as a function of aging was observed for the corneal tissue. The lissamine green investigation (described below) for those samples showed major staining on the tissue (data on file). For those reasons, the standard procedure was adapted such that the measurement protocol used for corneas involved only a single ramp in which the CoF was measured using 6 FN (0.2, 0.5, 1, 2, 3, and 4 mN).
After CoF measurements, corneal staining was performed using a lissamine green ophthalmic strip (HUB Pharmaceuticals, LLC, Rancho Cucamonga, CA) followed by using a fluorescein sodium ophthalmic strip (Akron, Inc, Lake Forest, IL). Each dye was applied to the cornea using a drop of eye wash (Mayor Pharmaceuticals, Phoenix, AZ). Slit-lamp examination was conducted for the lissamine green dye using a Haag-Streit BX-900 (Haag-Streit, Koeniz, Switzerland) before the application and observation of the fluorescein dye with a Wratten filter (Kodak, Rochester, NY).
Data Evaluation and Statistical Methods
A CoF measurement was considered successful if all of the following conditions were met: (1) a region with constant FN values for all FT values was recorded; (2) no experimental abnormalities (eg, spike resulting from incorrectly mounted lens) were noted; (3) the shape of the FT curve was normal and no crossing of data lines was observed, the value of the CoF was equivalent to the slope of the plot of FT versus FN, and the CoF did not vary greatly with the selection of the data region. Human corneal CoF values were analyzed using a 1-way analysis of variance (ANOVA) model, including the buffer as a fixed factor, to test for differences between the buffer solutions. The Kenward and Roger method29 was used to calculate the denominator degrees of freedom for the analysis. Adjustment for multiplicity to control the type 1 error rate was performed using the Dunnett method.30 All statistical tests were 2-sided and were conducted at a 5% level of significance. All data summaries and statistical analyses were performed using SAS software Version 9.3 (SAS Institute, Cary, NC). Descriptive statistical analysis tests of R2 values were also conducted to compare CoF values with age of the donor and CoF values with time from death to CoF measurement.
The fresh human corneal tissue CoF measured in differing solutions resulted in a value range of 0.006 to 0.015.
Corneal Tissue Samples
The time from a tissue donor's death to preservation of the donated corneal tissue has been shown to affect the integrity of the epithelium. A review of multiple sources showed an ideal death to corneal tissue preservation time (DP) of 6 hours for maintaining the epithelium integrity during transplantation31,32; however, this 6-hour time point was often determined by calculating an average DP that included very low (2 hours) and very high (11 hours) values. In addition to the DP, other factors that may influence corneal tissue quality have been identified. The following 2 additional factors that affect corneal tissue quality were recorded for each cornea used in this study because the tissue was not preserved: the time from death to ocular cooling (cooling minimizes cell metabolism to preserve the original properties of corneal tissue) and the time from death to enucleation (globe preserved in a moist container for optimal epithelial care). The authors determined 7 hours to be the appropriate DP maximum duration for defining “fresh” human corneal tissue for use in measuring CoF values. With 7 hours as the upper limit, it was hypothesized that the mean DP would fall within the ideal transplantation DP in view of the projected sample size of the investigation. Of the 28 corneas used for the CoF measurements, 1 exceeded the 7-hour time limit by 29 minutes. This sample was still used for CoF measurements because the epithelium of this sample, along with the other 27 samples, was confirmed to be intact using slit-lamp examination. The mean times calculated for death to ocular cooling and death to enucleation were within the common preservation standard for transplantation [death to ocular cooling: mean of 3 hours, 16 minutes (range: 50 minutes to 5 hours, 53 minutes); death to enucleation: mean of 4 hours, 55 minutes (range: 2 hours, 10 minutes to 7 hours, 29 minutes)]. Additional details of the donors for these corneal tissue samples are provided in Table 1.
Impact of Buffer Solution
CoF measurements were performed on intact corneas using a microtribometer according to the methodology described by Roba et al.5 Previous studies have shown that the test solution used, particularly the protein content of that solution, can have a substantial impact on measurements of friction, including CoF values, of the cornea.12,33 Therefore, CoF values were measured in 4 different solutions: TMS-PS, TMS-PBS, TMS-HEPES, and TLF-PBS. Each solution contains a mixture of proteins and mucin; however, lipid components are found only in TLF. Therefore, TLF is currently the most representative artificial solution that mimics natural tears. Seven successful corneal CoF measurements were obtained in TMS-PS solution, 6 in TMS-PBS solution, 5 in TMS-HEPES solution, and 10 in TLF-PBS solution. CoF values by buffer solution are summarized in Figure 1. ANOVA showed overall significant differences in human corneal CoF measurements between the 4 solutions. Least-squares mean differences between the solutions (TLF-PBS, TMS-HEPES, and TMS-PS vs. TMS-PBS) are summarized in Table 2. These results showed that there were statistically significant differences in the human corneal CoF measurements in TMS-PBS compared with those in TLF (P = 0.0424) and in CoF measurements in TMS-PBS compared with those in TMS-HEPES (P = 0.0179). For human corneal CoF measurements in TMS-PBS and TMS-PS, no statistically significant difference was determined (P = 0.2389).
Impact of Age of the Donor and Time From Death to CoF Measurement
Because this was a novel pilot investigation, the influences of 2 additional factors, age of the donor and time from death to CoF measurement, on corneal CoF measurements were also evaluated. The CoF versus the age of the donor and the CoF versus time from death to CoF measurement are presented in Figures 2, 3, respectively. No correlations were observed between the corneal CoF and age of the donor or between corneal CoF and time from death to CoF measurement. The time disparity between the right and left corneal CoF measurements, with a mean CoF of 0.015 (n = 16) for the right eye and 0.013 (n = 12) for the left, was not statistically different (P = 0.93).
To determine whether the tribological methodology caused tissue damage, 5 corneas were evaluated for corneal staining after CoF measurements. The tissue measured in the solution considered most similar to natural tears (TLF-PBS) was evaluated. Figure 4 shows 1 corneal tissue sample evaluated for staining. This image was captured approximately 10 hours after death and is a typical representation of the 5 corneas that underwent an excision from the whole globe, CoF measurement in TLF-PS solution by the microtribometer, and evaluation for corneal staining. No lissamine green or fluorescein staining was observed in the CoF test region. After the corneal excision, the tissue was kept moist by the testing solution (TLF-PS); however, no refrigeration or preservation medium was used, most likely causing edema-driven deep tissue folds.
To our knowledge, this was the first time that the CoF of fresh human corneal tissue was successfully measured. Despite the availability of preserved human corneal tissue, the authors decided against its use on the premise it may possibly influence tissue integrity and accuracy of CoF values obtained. Challenges faced with the use of fresh corneal samples included availability and timely acquisition to maintain ex vivo corneal tissue integrity. This resulted in the use of both right and left corneas from each donor, as well as establishing a definition of “fresh” human corneal tissue. The use of bilateral samples from each donor is a limiting factor; however, because of logistical constraints as mentioned above, each corneal tissue was considered as an independent sample regardless of its origin. An additional challenge was conducting CoF measurements using a solution that adequately mimics the natural tear film.18,34 To address this, 4 different test solutions were used.
To conduct this novel investigation, the authors used a microtribometer and an adaptation of the methodology described by Roba et al.5 The mean CoF values ranged from 0.006 to 0.015 (depending on the test solution). No clear trend was observed when the CoF was plotted as a function of donor age or time from death to CoF measurement, which suggests that CoF measurements in this study were not significantly affected by either of these factors. Review of the literature shows that there is currently no comparator and the CoF values presented here stand alone in the tribological field.
When comparing test solutions (TMS-PS, TMS-PBS, and TMS-HEPES) having the same protein composition (TMS), differences were observed. The CoF measured in TMS-PBS was significantly lower than the CoF measured in TMS-HEPES and TMS-PS; however, the difference was statistically significant only for TMS-HEPES. The difference did not reach statistical significance for TMS-PS. We hypothesize that PBS may induce particular protein confirmations that favor low CoF values for the corneal surface. Indeed, different buffer solutions have been shown to affect the conformation of proteins.35 CoF values measured in TMS-PBS were also significantly lower than those measured in TLF-PBS. In this case, the differences are to be found in the composition of the test solutions (TMS-PBS and TLF-PBS). Specifically, we hypothesize a friction-increasing effect of the added lipid component present in the TLF formulation. The testing solution containing both proteins and lipids, most similar to human natural tears, is the TLF formulation. At this time, it is not clear whether the high degree of variability seen within the TMS-HEPES and TLF-PBS measurements was induced by the solution, corneal tissue, or both. Molecular-level interactions between the components of the buffer solutions and the corneal epithelial surface need to be further explored to assess the higher variability observed for certain buffer types. Randomization of the solution types may further minimize potential bias.
Corneal staining evaluations after CoF measurements showed no cellular damage in the CoF testing region. In contrast, other corneal tissue models have shown substantial cellular damage after friction testing.9,10,12,33 Our results indicate that the corneal samples used in this analysis were fresh and had not undergone significant degradation due to aging32 and that the method of measuring the corneal CoF did not inflict damage to the tissue.
The results presented by this investigation provide the first CoF values determined for fresh human corneal tissue in multiple test solutions (range: 0.006–0.015). Because the CoF of corneas may influence the frictional forces delivered to the eye by CLs, the values reported here may provide insight into CL comfort. Thus, the CoF of the human cornea may serve as a physiological point of reference for the development of medical devices that are more compatible with the ocular environment.
The authors gratefully acknowledge Lions VisionGift for their dedication to research and for being instrumental in the success of this investigation. Editorial support for the writing of this article was provided by Megan Knagge, PhD, of MedErgy, and was funded by Johnson & Johnson VisionCare, Inc. The authors would also like to thank Aisha Elfasi, BS, for her contributions to the preparation of the manuscript.
1. Subbaraman LN, Jones LW. Measuring friction and lubricity of soft contact lenses: a review. Contact Lens Spectr. 2013;28:28–33.
2. Dunn AC, Sawyer WG, Angelini TE. Gemini interfaces in aqueous lubrication with hydrogels. Tribol Lett. 2014;54:59–66.
3. Kern J, Rappon J, Bauman E, et al.. Relationship between contact lens coefficient of friction and subjective lens comfort. Cont Lens Anterior Eye. 2013;36:e26.
4. Rennie AC, Dickrell PL, Sawyer WG. Friction coefficient of soft contact lenses: measurements and modeling. Tribology Lett. 2005;18:499–504.
5. Roba M, Duncan EG, Hill GA, et al.. Friction measurements on contact lenses in their operating environment. Tribol Lett. 2011;44:387–397.
6. Brennan NA, Coles C. Supportive data linking coefficient of friction and comfort. Cont Lens Anterior Eye. 2013;36:e10.
7. Coles-Brennan C, Brennan N. Coefficient of friction and soft contact lens comfort. Optom Vis Sci. 2012;89; e-Abstract 125603.
8. Ross G, Nasso M, Franklin V, et al.. Silicone hydrogels: trends in products and properties. Poster presented at: BCLA 29th Clinical Conference & Exhibition; June 3-5, 2005; Brighton, UK.
9. Cobb JA, Dunn AC, Kwon J, et al.. A novel method for low load friction testing on living cells. Biotechnol Lett. 2008;30:801–806.
10. Dunn AC, Cobb JA, Kantzios AN, et al.. Friction coefficient measurement of hydrogel materials on living epithelial cells. Tribology Lett. 2008;30:13–19.
11. Meyer AE, Baier RE, Chen H, et al.. Differential tissue-on-tissue lubrication by ophthalmic formulations. J Biomed Mater Res B Appl Biomater. 2007;82:74–88.
12. Morrison S, Sullivan DA, Sullivan BD, et al.. Dose-dependent and syndergistic effects of proteoglycan 4 on boundary lubrication at a human cornea-polydimethylsiloxane biointerface. Eye Contact Lens. 2012;38:27–35.
13. Schmidt T, Masala N, Morrison S, et al.. Characterization of recombinant human PRG4 as an ocular surface boundary lubricant. Invest Ophthalmol Vis Sci. 2012;53; e-Abstract 552.
14. Samsom M, Chan A, Jones L, et al.. PRG4 as a natural boundary lubricant for commercial silicone hydrogel contact lenses. Invest Ophthalmol Vis Sci. 2013;54; e-Abstract 5468.
15. Greenbaum A, Hasany SM, Rootman D. Optisol vs Dexsol as storage media for preservation of human corneal epithelium. Eye (Lond). 2004;18:519–524.
16. Nichols BA, Chiappino ML, Dawson CR. Demonstration of the mucous layer of the tear film by electron microscopy. Invest Ophthalmol Vis Sci. 1985;26:464–473.
17. Sharma A, Khanna R, Reiter G. A thin film analog of the corneal mucus layer of the tear film: an enigmatic long range non-classical DLVO interaction in the breakup of thin polymer films. Elsevier 1999;14:223–235.
18. Albarran C, Pons AM, Lorente A, et al.. Influence of the tear film on optical quality of the eye. Cont Lens Anterior Eye. 1997;20:129–135.
19. Lee S, Spencer ND. Materials science. Sweet, hairy, soft, and slippery. Science. 2008;319:575–576.
20. Giasson S, Spencer ND. Aqueous lubrication with polymer brushes. In: Spencer ND, ed. Aqueous Lubrication. New Jersey, NJ: World Scientific; 2014.
21. Hull DS, Green K, Thomas L. Effect of HEPES buffer on corneal storage in MK medium. Acta Ophthalmol (Copenh). 1984;62:900–910.
22. Gorbet MB, Tanti NC, Jones L, et al.. Corneal epithelial cell biocompatibility to silicone hydrogel and conventional hydrogel contact lens packaging solutions. Mol Vis. 2010;16:272–282.
23. Schultz RK, Meyers RR. The chemorheology of poly (vinyl alcohol)-borate gels. Macromolecules. 1969;2:281–285.
24. Manna U, Patil S. Borax mediated layer-by-layer self-assembly of neutral poly(vinyl alcohol) and chitosan. J Phys Chem B. 2009;113:9137–9142.
25. Angelova LV, Terech P, Natali I, et al.. Cosolvent gel-like materials from partially hydrolyzed poly(vinyl acetate)s and borax. Langmuir. 2011;27:11671–11682.
26. Chaudhri MM, Yoffe EH. The area of contact between a small sphere and a flat surface. Philos Mag A. 1981;44:667–675.
27. Knox Cartwright NE, Tyrer JR, Marshall J. Age-related differences in the elasticity of the human cornea. Invest Ophthalmol Vis Sci. 2011;52:4324–4329.
28. Shaw AJ, Collins MJ, Davis BA, et al.. Eyelid pressure and contact with the ocular surface. Invest Ophthalmol Vis Sci. 2010;51:1911–1917.
29. Kenward MG, Roger JH. Small sample inference for fixed effects from restricted maximum likelihood. Biometrics. 1997;53:983–997.
30. Dunnett CW. New tables for multiple comparisons with a control. Biometrics. 1964;20:482–491.
31. Kryczka T, Szaflik JP, Szaflik J, et al.. Influence of donor age, post-mortem time and cold storage on metabolic profile of human cornea. Acta Ophthalmol. 2013;91:83–87.
32. Van Meter WS, Katz DG, White H, et al.. Effect of death-to-preservation time on donor corneal epithelium. Trans Am Opthalmol Soc. 2005;103:209–224.
33. Schmidt TA, Sullivan DA, Knop E, et al.. Transcription, translation, and function of lubricin, a boundary lubricant, at the ocular surface. JAMA Opthalmol. 2013;131:766–776.
34. DelMonte DW, Kim T. Anatomy and physiology of the cornea. J Cataract Refract Surg. 2011;37:588–598.
35. Ugwu SO, Apte SP. The effect of buffers on protein conformational stability. Pharm Technology. 2004;28:86–109.