The power of tears: how tear proteomics research could revolutionize the clinic

KEYWORDS:

• Tear proteomics
• tear fluid
• tears
• eye disease

A thin layer of tears covers the ocular surface. The tears are a complex fluid comprised of secretions from a number of sources including the lacrimal gland, goblet cells, cornea, and vascular sources. The tear film provides an optically smooth surface for focusing light onto the retina as well as lubrication, prevention of dehydration of the mucosal surface, protection against pathogens, and nutrition for the underlying corneal and conjunctiva epithelial cells [1]. The tears contain thousands of molecules including proteins/peptides, lipids, electrolytes, and small molecule metabolites secreted from the main and accessory lacrimal glands, meibomian glands, goblet cells, and ocular surface epithelial cells [1]. The updated three-layer model of tear film consists of an outer lipid layer (outer nonpolar lipid sub-layer and inner polar lipid sub-layer with intercalated proteins), a middle aqueous layer, and an inner mucin layer with molecular interactions with the membranes of the corneal and conjunctival epithelial cells. In healthy eyes, the thickness of the tear film is around 5 µm [2] and the volume is about 5–10 µl [1].

Despite the small volume for sampling, tear fluid offers several advantages for biochemical analysis. Tears are usually collected noninvasively using Schirmer’s strips, other absorbent materials, or fire-polished microcapillary tubes. However, care must be taken not to activate the corneal nerves and induce reflex tears as there is a dramatic difference in the tear protein profile between the two, with reflex tending to be more dilute than basal tears [1]. Tears can be easily obtained from healthy subjects, which is often required as a control group for ‘Omics’ studies, whereas other ocular fluids (i.e. aqueous humor and vitreous) are not feasible for collection from healthy people. The biggest advantage is that tears are close to the disease site (for instance, ocular surface disease, lacrimal gland disease, etc.) as compared to detecting cancer biomarkers in blood, where biomarker molecules are highly diluted. The total protein concentration of human tears ranges from 6 to 11 mg/ml. Lysozyme, the most abundant tear protein, is approximately 1–mg/ml. Unlike plasma, depletion of abundant tear proteins such as albumin is not necessary and good quality tear proteome coverage can be obtained. This simplifies sample preparation without worrying about the possible loss of tear proteins during the depletion process.

Over the past decade, advances in mass spectrometry (MS) and proteomics/metabolomics/lipidomics/glycomics technology have significantly improved our understanding of the chemical composition of the tears. Close to 2000 tear proteins have been reported in humans [3,4]. At least 90 small molecule metabolites have been seen in human tears, but that number should increase as studies are published [5]. A few hundred individual lipid species across seven lipid classes have been documented to date [6]. Characterization of glycans in human tears has been reported recently [7,8]. This comprehensive biomolecule repertoire in human tears serves as a good source for biomarker discovery for diseases.

The biochemical changes in tear fluids have been explored in many eye diseases using ‘Omics’ approaches over the last decade. Dry eye syndrome is a tear film disease and among the first eye disease studied using proteomics. Alterations in lacrimal gland secreted proteins, and inflammation- and stress-related proteins were found to be associated with dry eye [9]. A number of studies have verified these basic proteomic findings [10–12]. Furthermore, a similar change in the tear proteome was also observed in a rabbit model of Sjögren’s syndrome-associated dry eye [13]. Proteomic studies have consistently found that levels of pro-inflammatory proteins indicating ocular surface inflammation were marked by well-known inflammation-related proteins (S100 A8 and S100 A9) [14]. Meibomian gland dysfunction also contributes to dry eye and a lipidomics study demonstrated that the levels of some amphiphilic lipids increased after eyelid-warming treatment in meibomian gland disease (MGD) [15]. Those amphiphilic lipids are believed to have critical roles in maintaining tear film stability [6,15]. Dry eye symptoms are common after refractive surgery procedures, but most patients recover after a few months. The tear proteomic profile was used to assess two refractive surgery platforms and we found that recovery toward a normal tear protein profile was depended on which treatment was used [16]. However, when the same study cohort was investigated using clinical assessments, the clinical parameters were not sensitive enough to show the statistical differences between the two platforms [17] though they observed similar trends as with the proteomics results [16]. Thus, tear proteomics or lipidomics may have some advantages over conventional clinical assessments (i.e. Schirmer test, tear break-up time, corneal staining) for patient stratification diagnosis, prognosis, and monitoring the response to treatment. These methods can be used to provide sensitive and quantitative procedures for patient stratification.

An ocular surface issue receiving a considerable attention has been the clinically observed ocular surface inflammation seen in patients on topically applied glaucoma medications [18]. The concern is that a more inflammatory ocular surface may lead to a greater postoperative scarring response which could result in a higher risk of surgical failure [18]. Thus, monitoring and assessment of the ocular surface inflammation is crucial and may be useful for observing the outcomes of glaucoma surgery. A simple test, measuring S100 A8 or S100 A9 proteins in tears, was found to reflect the inflammatory status of the ocular surface and could be clinically useful. Clearly, these types of tests could be translated into office devices for greater access.

From the interaction of contact lens wear with ocular surface structures, it is not surprising that tear proteome is subject to contact lens wearers. However, the changes are found to be associated with different types of contact lenses and even contact lens care solutions [19]. Tear proteomics could be useful to understand these modifications to different types of contact lenses and to improve the contact lens tolerance and complications of wear.

The tear film proteomic profile has been found to provide basic biological information for many ocular diseases, such as keratoconus [20], thyroid eye disease [21], vernal keratoconjunctivitis [22], diabetic retinopathy [23], and primary open angle glaucoma [24]. One key question is whether those tear biomarkers are specific to certain diseases. By reviewing our own data on many different eye diseases and reports from other groups, we found there are significant differences in tear profiles in different diseases. Recently, our group conducted a mega-analysis of the tear proteome in four diseases and conditions including keratoconus (n = 57), thyroid eye disease (n = 23), systemic lupus erythematosus (n = 24), and post-laser-assisted in situ keratomileusis (LASIK) (n = 87). Distinct tear proteomic patterns were observed for each disease (unpublished data). Another example by Matheis et al. [21] also showed that a significantly different protein panel has been found in thyroid-associated orbitopathy as compared to dry eye and controls.

Based on the current knowledge, the tear proteome is clearly different from plasma proteome with about 500–600 plasma proteins being seen in tears [3]. This is striking as there are more than 8000 proteins in the plasma proteome. Due to the overlap between the tear proteome and plasma proteome, there may be opportunities to observe systemic responses in the tears. Several such examples thus far include breast cancer [25], type 2 diabetes [26], Alzheimer’s disease [27], and rheumatoid arthritis [28]. Some systemic diseases may affect the eye so that we can use ‘tears’ as a ‘window’ to assess systemic as well as ocular disease.

In summary, tear ‘omics’ research over the past decade has demonstrated the future applications of tear biomarkers for patient stratification or what is now often referred to as precision medicine. Tear collection is fast, safe, and noninvasive and offers a chance to determine the local pathology close to the disease site. The relatively simple chemical composition and sample preparation procedures make tear fluid an ideal source for diagnosis and prognosis. Proteomic studies can easily be translated into antibody-based assays for clinical use. There are already some tear-based diagnosis kits available in the market (Advanced Tear Diagnostics, Birmingham AL) for use with dry eye and ocular allergies. We believe that there will be more similar tear-based products available in the future. On the other hand, with the recent multiple reaction monitoring MS technology [4] and routine-use and miniaturized MS, targeted MS will be an alternative platform because of its advantages for multiplexing, higher specificity, and faster run times. Emerging ‘Omics’ technologies, such as data-independent acquisition (DIA or sequential window acquisition of all theoretical fragment ion spectra [SWATH]), are particularly useful in clinical proteomics because DIA or SWATH has higher reproducibility with much less missing data across many samples [29]. Ion-mobility MS [30] offers another dimension of separation based on ion mobility (related to the ‘shape’ of the molecule) in addition to m/z. This method is valuable for tear lipids because interfering isobaric species is the major challenge in lipidomics research. We look forward to a ‘tear test’ that will eventually become like a ‘blood test’ or ‘urine test’ used in eye clinics in the near future.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Additional information

Funding

The authors were supported by the National Medical Research Council (NMRC), Singapore (grant CG 2013) and the SingHealth Foundation which supported the core proteomics facility.