Mass Spectrometric Analysis of Meibomian Gland Lipids

Abstract

The precorneal tear film keeps the eye surface moist and helps to maintain normal eye function. The outermost lipid layer of the tear film, which attenuates tear film evaporation, contains meibum secreted from the meibomian gland. Most meibum lipids are neutral, including wax esters (WEs), cholesteryl esters (CEs), and diesters (DEs), along with some polar lipids including free fatty acids (FFAs), O-acyl-ω-hydroxy fatty acids (OAHFAs), and trace phospholipids. Detection of neutral lipids by mass spectrometry (MS) is challenging due to interference from impurities, particularly when working with minute-volume meibum samples. Here, we describe procedures for sample preparation and MS analysis of these elusive meibum lipids that can be used to examine dry eye disease mechanisms. Because the method described here minimizes impurity peaks for lipids generally, neutral and otherwise, it may be applied to high-sensitivity analysis of other biological samples.

1. Introduction

The cornea is covered by a tear film that imparts several functions, including lubrication and antimicrobial protection of the corneal surface [1]. The tear film, which has a total thickness of 2~5.5 μm, is composed of inner, intermediate, and outer strata called the mucin, aqueous, and lipid layers, with the inner two layers together forming a mucoaqueous phase [2,3]. In the absence of the outer lipid layer, the water in the mucoaqueous phase would evaporate rapidly (10.9 ± 0.1 μm/min) and be completely evaporated within 12~30 s [4]. This evaporation could not be compensated for by lacrimal fluid secretion, which has a normal flow rate of only 1.0 ± 0.4 μL/min [5] or 0.08 ± 0.03 μm/min for an average exposed eye surface area (approximate 1.25 cm²) [6]. Instead, the outermost lipid layer plays an essential role in attenuating tear film evaporation [7]. The lipid layer components are supplied by the eye’s meibomian gland secretions, collectively called meibum [2]. An abnormality in meibum composition will lead to a defect in the lipid layer, such that it cannot impede mucoaqueous phase evaporation effectively, resulting in dry eye disease [8,9]. To understand the mechanisms of dry eye disease, it is therefore essential to characterize lipid layer composition abnormalities by way of comprehensive identification and quantification of differences in meibum lipids between healthy and dry eyes.

Meibum is a holocrine secretion released upon bursting of a fully differentiated meibomian gland cell, at which time the meibum-containing cell’s full contents are discharged [10]. Mass spectrometry (MS) analysis studies have demonstrated that meibum is composed predominantly of lipids, the majority of which are neutral, including cholesteryl esters (CEs), wax esters (WEs), and diesters (DEs), along with a small amount of polar lipids, including free fatty acids (FFAs) and O-acyl-omega hydroxy fatty acids (OAHFAs) [11-15]. Other cell components, including phospholipids constituting cell membranes, are often below the limit of detection [12]. Each of these neutral lipid molecules (of up to 79 carbon atoms) contains two to three long or ultra-long carbon chains, with one or two ester groups as the only chemical functional group [14]. As a result, these lipids are highly nonpolar, and thus their dissolution typically requires a solvent mixture with a highly nonpolar solvent, such as chloroform [14]. However, potent organic solvents such as chloroform can extract contaminants from laboratory ware, particularly, additives from plastics [16,17]. Compared to other molecules (e.g., proteins/peptides or phospholipids that contain acidic or basic functional groups), these nonpolar lipids are more difficult to ionize for MS detection [18]. Therefore, interference from contaminants is a major concern when analyzing neutral lipids. The MS analysis of meibum lipids is particularly challenging, due to the minute size of meibum lipid samples, which are typically in the microgram range when collected with glass microcapillary tubes [14]. Thus, the same absolute amount of plasticizers or other impurities that are always a concern in lipid analysis become more substantial relative to the volume of a meibum sample, making the signals from lipids of interest more sensitive to interference. As a result, minimizing interference by careful handling is very important for acquiring high-quality meibum lipid composition data, particularly in shotgun MS, wherein all peaks are detected simultaneously. Without careful cleaning, it is not uncommon to detect high-intensity contaminant peaks, including from plastic additives and carryover of calibration standards (see Fig. 4 of reference [19] for an example).

We have developed a method for shotgun lipidomics of meibum and have obtained high-quality MS and MS/MS spectra of meibum lipids with our method in both positive and negative ion modes. This method has also been applied to meibum collected from tree shrews [20]. This approach can be used to compare lipid compositions across meibum samples, between healthy and dry eye subjects [8,21,22] or between humans and other animals (e.g., tree shrews) with a different evaporation retardation capability [20]. The results of such comparisons can help us to understand the mechanism of dry eye disease. The method described here may be applied to high sensitivity analyses of various lipid-containing samples, including neutral-lipid-rich biological samples such as tears [23,24], sebaceous gland secretions (sebum) [25], vernix caseosa [26], cerumen [27], preen gland secretions [28], and seed oil [29].

2. Materials

Prepare all solutions using HPLC or LC/MS grade water and solvents when available, including HPLC-grade chloroform (>99.9%, with amylene stabilizer), LC/MS-grade methanol (>99.9%), ammonium hydroxide (NH₄OH) solution (25%), ammonium acetate (NH₄Ac), and HPLC-grade water. Wear gloves when performing the experiments. Use glass syringes to transfer organic solvents, including chloroform and methanol, or solutions containing these organic solvents. Standard pipettes are used to transfer water or aqueous solution.

1. Gloves: Kimberly-Clark Professional™ Purple Nitrile™ Exam Gloves.

2. Tissues: Kimberly-Clark Professional™ Kimtech Science™ Kimwipes™ Delicate Task Wipers, 1-Ply.

3. Airtight glass syringes

   1. Two 1 mL syringes with a removable point style 2 needle, for transferring solvents between storage glass vials or from storage glass vials to sample vials.
   2. One 100 μL syringe with removable point style 3 needle, for transferring solvent from storage glass vials to sample vials or between sample vials.
   3. One 25 μL syringe with removable point style 3 needle, for transferring solution from storage glass vials to sample vials or between sample vials.

4. 32-mm length, 0.5-μl capacity, 0.022 inch O.D., 0.0056 inch I.D. glass microcapillary tubes (Drummond; Broomall, PA).

5. Amber screw thread sample vials with PTFE-faced/white rubber lined caps (35 mm height) for microcapillary glass tubes.

6. 20 mL clear glass vials with PTFE-lined caps for solvent and solvent mixture storage.

7. Chloroform

   Pour ~18 mL HPLC grade chloroform into a 20 mL glass storage vial. Wipe the mouth of the original chloroform bottle and the glass vial with Kimwipes. Keep chloroform from contacting gloves or the plastic caps through wetted Kimwipes, and cap the vial and wrap with aluminum foil to avoid exposure to air and light to prevent chloroform from degradation during storage (see Note 1).

8. Methanol

   Pour ~ 18 mL HPLC grade methanol into a 20 mL glass vial. Wipe the edge of the original methanol bottle and the storage glass vial with one or more stacked pieces of folded Kimwipes. Take care not to let methanol permeated through the Kimwipes in contact with gloves or plastic caps.

9. Chloroform-methanol solvent mixture (2:1, V/V)

   Use a 1 mL glass syringe to transfer 8 mL chloroform (item 7), 4 mL methanol (item 8) to a 20 mL glass vial. Cap the vial and mix the solvent well, but not to contact the cap. Seal and wrap the vial with aluminum foil to keep it from exposure to air and light to prevent chloroform degradation (See Note 1).

10. 2 mL glass vials for sample solutions

11. 2mL storage glass vials with PTFE-faced/clear silicone-lined caps for long-term storage of solutions

12. Cleaning wires (originally for cleaning 22s, 25s, and 28-30 gauge needles) for meibum sample preparation.

13. 100 mM NH₄Ac solution

    Make a 1 M NH₄Ac solution by dissolving 385 mg NH₄Ac in 5 mL HPLC grade water in a 20 mL glass vial, and make a 10 times dilution to 100 mM by adding 900 μL methanol to 100 μL of the 1 M NH₄Ac solution in a 2 mL storage vial.

14. Methanol with 2.5% NH₄OH solution: mix 100 μL 25% NH₄OH and 900 μL methanol in a 2 mL glass vial

3. Methods

3.1. Meibum sample collection

1. Roll the lower eyelid (see Note 2) away from the eyeball and apply gentle pressure to the eyelid [30,31]. Oily secretions will exit the meibomian gland orifice at the lid margin [32].

2. With the eyelid held in a rolled-away position relative to the eyeball with one hand, use the thumb and index finger of the other hand to hold a 32-mm, 0.5-μl glass microcapillary tube (see Notes 3), using the tube to tap secreted meibum droplets gently and repeatedly to fill the tube. The typical fill length of meibum samples in the tube is 0.5–1 mm, corresponding to approximately 8–16 nL or 7–13 μg (see Note 4).

3. Store meibum sample-filled microcapillary tubes individually in glass vials at −80 °C or −20 °C until the time of analysis (see Notes 5 and 6).

3.2. Sample solution preparation

1. Push meibum from the microcapillary tube into the sample vial using the cleaning wire.

2. Prepare sample stock solutions on the same day that MS analysis is to be performed (see Note 6). Dissolve single human meibum samples directly in a chloroform-methanol solvent mixture (2:1, V/V) within a 2-mL glass sample vial (see Notes 7 and 8). Each millimeter of meibum within the 32-mm capillary is dissolved directly in 100 μL of the chloroform-methanol mixture.

3. Dilute the above stock solution 5-fold or 10-fold with four or nine volumes of methanol, then add 1% (V/V) 100 mM NH₄Ac or 2.5% NH₄OH additive to the diluted solution. The final concentration of the additive in the working solution is 1 mM NH₄Ac or 0.025% NH₄OH (see Notes 9 and 10).

3.3. MS and MS/MS data acquisition

1. Thoroughly clean the setup, including syringes, infusion lines, sample preparation glass vials, and electrospray ionization probes, by alternately rinsing/infusing with chloroform-methanol 2:1 mixture and methanol, and occasionally with straight chloroform several times to ensure no significant background peaks from impurities in the system or carryover from previous experiments. (see Note 8).

2. Infuse the working solution at 7 μL/min into an electrospray-ionization (ESI) quadrupole time-of-flight (qTOF) ion source in positive ion mode until the total ion signal is stable (<1% variation) (see Note 11).

3. Acquire the MS spectra at 7 μL/min with a typical acquisition time ~3 min per MS run, which strikes a good balance between maximizing the signal-to-noise ratio and acquisition time. Sum all MS spectra to an averaged spectrum (Fig. 1).

4. Calibrate the TOF mass analyzer by using the known endogenous meibum lipid peaks as internal calibrants to correct for mass accuracy drift due to any temperature fluctuation (see Notes 12 and 13). Calibration peaks may include the ammonium adducts of the molecular species of two WEs (40:1 and 44:2), one CE (24:1), one triacylglycerol (TG)(54:3), one α, ω type-II DE (DE-II)(68:3), one ω type I-St DE (DE-I)(50:2), and the [M + H – H₂O]⁺ ion of cholesterol (see Note 14).

5. Select ions of interest for collision-induced dissociation MS/MS analysis (Fig. 2 a-g). Set up acquisition times from <1 s to up to 30 min, depending on the intensity of the precursor ion and the quality of the fragment ions, with which the averaged sum MS/MS spectra can meet the requirement, or the particular experimental design (e.g., MS/MS^all, see Note 15).

6. Calibrate the TOF mass analyzer for MS/MS acquisition with peaks from ammoniated DE-I 50:2 and its fragments (see Note 16).

7. Repeat the procedures 1-3, and 5 for each working solution.

8. Change the MS polarity to detect negative ions and repeat procedures 1–7 for working solutions that contain the NH₄OH additive (Fig. 3; Fig. 2h; see Notes 12, 14, 15, 17 -19). Calibration peaks for MS in negative ion mode may include four deprotonated FAs (14:0, 16:0, 18:1, 21:0, and 24:1), cholesteryl sulfate (C₂₇H₄₅O₄S⁻), and three deprotonated OAHFAs (42:2, 50:3, and 52:3); for MS/MS in negative ion mode may include the OAHFA 50:2 ion and four of its fragment ions.

9. Calibrate the spectra after the acquisition using the internal calibrants described in procedures 4, 6, and 8 (see Note 12).

Fig. 1.

Comparison of two additives for ESI-MS analysis of meibum lipids in positive ion mode. a) 0.025% NH₄OH; and b) 1 mM NH₄Ac. The 0.025% NH₄OH additive, ideal for negative ion mode detection, also works well in positive ion mode and can therefore be used as the additive for both negative and positive ion detection. The two numbers separated by a colon labeling each peak for wax ester (WE) and α, ω Type II diester (DE-II) species are the total number of carbon atoms and the number of double bonds, respectively. For cholesteryl esters (CEs) and ω Type I-St diesters (DE-Is), the total number of carbon atoms and the number of double bonds of the fatty acyl substituent are shown. For triacylglycerols (TGs), the total number of carbon atoms and the summed number of double bonds of the three fatty acyl chains are shown. This research was originally published in the Journal of Lipid Research (Reference 38). © the Authors.

Fig. 2.

Representative MS/MS spectra extracted from MS/MS^all acquisition. a) wax ester (WE) 44:1; b) WE 43:0; c) α, ω Type II diester (DE-II) 68:3; d) sphingomyelin (SM) d34:1; e) cholesteryl ester (CE) 24:0; f) phosphatidylcholine (PC) 34:1; g) ω Type I-St diesters (DE-I) 50:2; and h) (O-acyl)-ω-hydroxy fatty acid (OAHFA) 50:2. Spectra a–g were acquired in positive ion mode; spectrum h was acquired in negative ion mode. Characteristic product ions are labeled with asterisks (*). The contents of the composition of the major molecule(s) of the ions are shown in parentheses. Moiety combinations for the major molecular species, including isomers, are also shown. Ions corresponding to ammoniated, protonated, and deprotonated lipids are labeled as M+NH₄⁺, M+H⁺, and M-H⁺, respectively. The abbreviations for the product ions and moieties of the lipids are as follows: FA, fatty acid; FAl, fatty alcohol; HO-FA, hydroxyl fatty acid; Chol, cholesteryl. This research was originally published in the Journal of Lipid Research (Reference 38). © the Authors.

Fig. 3.

ESI-MS analysis of meibum lipid sample with 0.025% NH₄OH in the negative ion mode. The two numbers labeling each peak, separated by a colon, represent the total number of carbon atoms and the number of double bonds for free fatty acids (FFA), hydroxy fatty acids (HO-FAs), and (O-acyl)-ω-hydroxy fatty acids (OAHFAs). This research was originally published in the Journal of Lipid Research (Reference 38). © the Authors.

4. Notes

1. When exposed to air and light, chloroform degrades and introduces impurity peaks as well as side reactions that interfere with mass spectrometry analysis [33-35]. Typically, commercially available chloroform includes a stabilizer such as amylene (~100 ppm) or ethanol (~1%) to minimize the degradation. A minimum exposure to light and air further keeps chloroform from degradation.

2. The protocol presented here is for human-related research. Study protocols for experiments involving human meibum collection must be approved by the researcher’s Institutional Review Board and developed in accordance with the Declaration of Helsinki.

3. Microcapillary collection of meibum directly from orifices minimizes the inclusion of contaminants from surrounding tissues [36]. The application of only gentle digital pressure minimizes the potential expression of immature meibum and avoids the discomfort of the subject encountered by other collection methods [11,37], such as with cotton buds, with meibomian gland forceps, or with a device called meibomian gland evaluator. These other approaches also collect more polar lipids, likely from immature meibomian gland cells or dead cells from the cornea surface, because more force is applied to the eyelid and a wider eyelid area is sampled [37].

4. Because meibum is hydrophobic, it is not subject to the typical capillary action that would draw a hydrophilic aqueous solution into a glass (also hydrophilic) microcapillary tube automatically. Moreover, because meibum solidifies as it cools, it will congeal when it moves away from the warm human body during the collection procedure. These two issues reduce the amount of meibum lipids collected. Recently, we showed that meibum can be collected more efficiently in polytetrafluoroethylene (PTFE) microcapillary tubes owing to the hydrophobicity of PTFE [24]. Importantly, this collection method is compatible with lipid extraction by chloroform or a chloroform-containing solvent mixture for MS analysis [24].

5. Keep the sample vials at −80 °C or −20 °C until processing to minimize potential degradation. Keep meibum in microcapillary tubes in the capped glass vials also minimize exposure of meibum lipids to the oxygen in the atmosphere. These storage conditions are satisfactory because most of the lipid species in meibum are saturated or monounsaturated, and less prone to oxidation. Despite that there is no significant change in lipid profiles in positive ion mode mass spectra of meibum solutions stored in solvents at 4 °C for one month, the glass vials used for sample storage can be filled with inert gas (e.g. nitrogen) for long term storage.

6. Although meibum lipids are typically stable, OAHFA peak intensities appear to decrease when samples are stored as a solution in a chloroform-methanol mixture overnight, either at room temperature or at −20 °C, possibly due to that OAHFA sticks to the glass vial walls. To minimize loss, prepare meibum solution immediately before MS analysis.

7. Multiphase separation is not performed because human meibum is composed almost exclusively of neutral lipids, with negligible amounts of other species [14,38]. There is no significant difference in lipid profiles in samples with [23] and without [38] multiphase separation, thus, lipid extraction methods, such as Folch extraction [39], is optional and can be applied if desired.

8. It is important to clean the experimental setup before commencing with data acquisition. Plastics, with the exception of Teflon (e.g., in syringe plungers) which is known to be inert to organic solvent, can introduce plastic additives (i.e., plasticizers) that interfere with MS analysis of lipids [17]. Therefore, glass vials and syringes should be used throughout the workup. Plastic caps for glass vials need to have a PTFE (or aluminum) liner.

9. We found that samples with 1 mM NH₄Ac or 0.025% NH₄OH are most suitable for ESI-MS analysis. The concentration may vary depending on the experimental conditions, but a higher concentration may cause ion suppression due to surface charge competition [14,40].

10. Samples with 0.25 % NH₄OH is ideal for negative ion mode analysis, which shows a much higher sensitivity for detection of negative ions, such as meibum FFAs [12]. The solution also works well for meibum lipid detected as positive ions which are otherwise formed in the presence of acidic additive. Thus, the same sample solution with NH₄OH can be used for lipid detection as both positive- and negative-ions [38].

11. High resolution mass spectrometry data can be acquired by any high-resolution mass spectrometers, including the q-TOF instruments from Waters [8,12,14], Bruker [21,22], and Sciex [20,23,24,38]. The flow rate should be optimized by the experimental conditions, which can vary in the range of 3 to 40 μL/min.

12. This chapter focuses on MS and MS/MS data acquisition using a high resolution qTOF instrument. Because the calibration of a TOF instrument may drift due to for example, temperature fluctuation [41], calibration should be made daily or whenever re-calibration is necessary.

13. The theoretical values of these peaks and their corresponding elementary composition (in parenthesis) are: m/z 369.35158 (C₂₇H₄₅⁺), m/z 608.63401 (C₄₀H₇₈O₂NH₄⁺), m/z 662.68096 (C₄₄H₈₄O₂NH₄⁺), m/z 752.72791 (C₅₁H₉₀O₂NH₄⁺), m/z 902.81712 (C₅₇H₁₀₄O₆NH₄⁺), m/z 1027.01509 (C₆₈H₁₂₈O₄NH₄⁺), and m/z 1145.09334 (C₇₇H₁₃₈O₄NH₄⁺). Depending on experimental conditions and the specific adduction, the corresponding protonated, or sodiated lipid [12] peaks may be used for MS calibration in positive ion mode.

14. This calibration strategy using internal lipid peaks can prevent the undesired long-lasting presence of external calibrants in the system, and mass accuracy can be significantly improved [12,14,38]. The selected peaks should be across the range of meibum lipid peaks with relatively high intensity and well resolved (minimal overlapping). The calculation of theoretical m/z values of the peaks should count the mass of an electron (loss or gain) [42].

15. MS/MS^all data (i.e. tandem MS spectra at 1-Da steps) can be acquired for all ions in the m/z range of interest with a Sciex Triple TOF 5600 instrument. A list of manufacturer’s predefined m/z values (including an appropriate mass defect) is employed for both positive and negative ion modes to define precursor ions for MS/MS^all acquisition for general application [20,38]. The same sample solution with 0.025% NH₄OH can be used for analyses in both ionization modes [20,38]. Typically, a 3-min MS scans are followed by MS/MS^all acquisition. The default MS/MS^all acquisition time is approximately 6 min [38] in each ionization mode, but can be changed as needed [20]. In the present study, MS/MS^all acquisition collects a total of 1,000 MS/MS spectra covering precursor ions in the m/z 200–1200 range at 1-Da steps in each ionization mode, and a total of 2,000 MS/MS spectra are collected [20,38]. Unit mass isolation window is applied to the precursor ions for the MS/MS spectra to match the total ion chromatograms of MS/MS^all in the corresponding MS spectra [38]. Pseudo precursor ion scanning and pseudo neutral loss scanning [20,38] spectra are extracted from MS/MS^all acquisition datasets with increased sensitivity for the detection of specific classes/subclasses of lipids, such as CEs and phospholipids [20,38].

16. The theoretical values of these peaks and their corresponding elementary composition (in parenthesis) are: m/z 1145.09334 (C₇₇H₁₃₈O₄NH₄⁺), m/z 369.35158 (C₂₇H₄₅⁺), m/z 459.45604 (C₃₂H₅₉O⁺), m/z 477.46661 (C₃₂H₆₁O₂⁺), m/z 741.71192 (C₅₀H₉₃O₃⁺), m/z 759.72249 (C₅₀H₉₅O₄⁺), and m/z 1128.06679 (C₇₇H₁₃₈O₄H⁺).

17. The theoretical values of meibum lipid peaks for MS calibration in negative ion mode and their corresponding elementary composition (in parenthesis) are: m/z 227.20165 (C₁₄H₂₇O₂⁻), m/z 255.23295 (C₁₆H₃₁O₂⁻), m/z 281.24860 (C₁₈H₃₃O₂⁻), m/z 325.31120 (C₂₁H₄₁O₂⁻), m/z 365.34250 (C₂₄H₄₅O₂⁻), 465.30441 (C₂₇H₄₅O₄S⁻), m/z 645.58273 (C₄₂H₇₇O₄⁻), m/z 755.69228 (C₅₀H₉₁O₄⁻), and m/z 783.72358 (C₅₀H₉₁O₄⁻).

18. Impurity peaks of linear alkylbenzene sulfonates [43] are often of high intensity and may be used as reference peaks for MS calibration. Because some impurity peaks may be close to a meibum lipid peak used for calibration [e.g. deprotonated linear alkylbenzene sulfonate with alkyl chain length C12 (C₁₈H₂₉O₃S⁻; m/z 325.18429) is close to deprotonated fatty acid 21:0 (m/z 325.31120)], care should be taken to avoid a mismatch.

19. The theoretical values for MS/MS calibration in negative ion mode are: m/z 757.70794 (C₅₀H₉₃O₄⁻), m/z 253.21731 (C₁₆H₂₉O₂⁻), m/z 281.24860 (C₁₈H₃₃O₂⁻), m/z 475.45206 (C₃₂H₅₉O₂⁻), and m/z 493.46262 (C₃₂H₆₁O₃⁻).