Rapid quantification of free cholesterol in tears using direct insertion/electron ionization-Mass spectrometry

Purpose. To establish a simple and rapid analytical method, based on direct insertion/electron ionization–mass spectrometry (DI/EI-MS), for measuring free cholesterol in tears from humans and rabbits. Methods. A stable-isotope dilution protocol employing DI/EI-MS in selected ion monitoring mode was developed and validated. It was used to quantify the free cholesterol content in human and rabbit tear extracts. Tears were collected from adult humans (n = 15) and rabbits (n = 10) and lipids extracted. Results. Screening, full-scan (m/z 40–600) DI/EI-MS analysis of crude tear extracts showed that diagnostic ions located in the mass range m/z 350 to 400 were those derived from free cholesterol, with no contribution from cholesterol esters. DI/EI-MS data acquired using selected ion monitoring (SIM) were analyzed for the abundance ratios of diagnostic ions with their stable isotope-labeled analogues arising from the D6-cholesterol internal standard. Standard curves of good linearity were produced and an onprobe limit of detection of 3 ng (at 3:1 signal to noise) and limit of quantification of 8 ng (at 10:1 signal to noise). The concentration of free cholesterol in human tears was 15 ± 6 μg/g, which was higher than in rabbit tears (10 ± 5 μg/g). Conclusions. A stable-isotope dilution DI/EI-SIM method for free cholesterol quantification without prior chromatographic separation was established. Using this method demonstrated that humans have higher free cholesterol levels in their tears than rabbits. This is in agreement with previous reports. This paper provides a rapid and reliable method to measure free cholesterol in small-volume clinical samples.


Introduction
The human tear film is composed of 3 layers -the inner mucus layer which is believed to make the epithelium more hydrophilic, 1 the middle aqueous layer which contains most of the electrolytes and proteins, 2 and the outer lipid layer whose major function is to prevent evaporation of the aqueous layer. 3 The lipid layer is of great importance in dry eye and many treatments are aimed at stabilising this layer. Rabbits have a tear film structure similar to humans, however their tear film is much more stable 4 and consequently they have a much longer inter-blink time than humans. [5][6][7] The lipid layer in both human and rabbit tears is composed of non-polar lipids, such as wax esters, cholesterol esters, triacylgycerols, and smaller amounts of polar lipids, such as phospholipids and (O-acyl)-ω-hydroxy fatty acids (OAHFA). [8][9][10] However the relative proportions of each type of lipid differ between humans and rabbits. 9 The major non-polar lipids in human tears are wax esters and cholesterol esters, with OAHFA the major polar type, whereas the major non-polar lipids in rabbit tears are 24,25-dihydro- 8 -lanosterol esters and diacylated diols, with OAHFA again the major polar lipid type. 9 Reports of free cholesterol and free fatty acids in tears suggest that these are minor components at ~1.5% and 0.5-2.1%, respectively. [10][11][12] Although free cholesterol is in relatively low abundance compared to cholesterol esters and wax esters in tears, it is nonetheless important. Cholesterol has previously been shown to form a condensed and compact monolayer at the air-water interface. It orients vertically at the interface with its hydrophilic hydroxyl (OH) group anchored in the aqueous phase. The rigidity of the cholesterol film plays an important role in interacting with molecules in the tear aqueous layer. 13 Abnormal proportions of free cholesterol and cholesterol esters in the composition of the tear film may lead to tear film instability. 14 Clinical studies have shown that cholesterol levels change in keratoconjunctivitis sicca 15 , acne rosacea related meibomian keratoconjunctivitis 16 , Sjögrens syndrome, and chronic blepharitis 15,17 , conditions that are also associated with reduced tear film stability. Recent studies have found that cholesterol is deposited onto worn contact lenses, with deposition quantities ranging from 2-37 µg per lens. 18,19 Furthermore, Saville et al. 20 were able to show differences in the level of cholesterol deposition between lens types. The deposition of lipid on contact lenses may contribute to the discontinuation of lens wear due to discomfort, or affect visual acuity during lens wear. 18,21 Reliable measurements of free cholesterol at low concentrations in the presence of comlex matrices of other more complex lipids presents unique analytical challenges. A series of enzymatic methods have been used for cholesterol quantification in body fluids such as plasma. 22 Other methods for determination of cholesterol such as gas-chromatography require derivatisation to make cholesterol sufficiently volatile for analysis. 23 These traditional methods are not suitable for tear samples due to the very small volumes of tears and low concentrations of the target analytes. A highperformance liquid chromatographic (HPLC) method using ultraviolet absorption detection has been developed for quantitation of cholesterol. 24,25 This method however, is impacted by relatively high background and inconsistency between studies. 18,26 Takatsu et al. 27 have reported a stable-isotope dilution mass spectrometry method to determine serum cholesterol, however their method involved purification of lipid extracts using HPLC which took around 25 minutes. Butovich also used HPLC-MS to detect cholesterol in tears. 9 Here we have developed a rapid and reliable method to quantify free cholesterol, without any derivatisation or chromatography steps, using direct insertion/electron ionization-mass spectrometry (DI/EI-MS). 20,28 The purpose of this paper is to detail the validation of this method and use this method to compare levels of free cholesterol in human and rabbit tears. . HPLC-grade chloroform and methanol and analytical grade ammonium acetate and butylated hydroxytoluene (BHT) used in the lipid extraction procedure were purchased from Crown Scientific (Sydney, NSW, Australia). All lipid solutions in the experimental work were prepared using HPLCgrade chloroform and stored in glass vials at −80 °C until use.

Tear collection
Both human and rabbit tears were collected and analysed. Human and animal ethics were approved by the Ethics Review Panel of the University of New South Wales and Brien Holden Vision Institute/Vision CRC, respectively. The Tenets of the Declaration of Helsinki were adhered to for human studies, and informed consent was obtained prior to enrolling human subjects into the study. Basal tear samples (5-10 µL) from humans (n = 15) and rabbits (n = 16) were collected using calibrated and fire polished disposable glass micro-capillary tubes (Brand, Wertheim, Germany). The glass micro-capillary tube was gently placed just above the lower tear meniscus, minimizing contact of the tip with eye surface to avoid reflex tears. No sedation or anesthesia was required. Samples from either species were centrifuged for 5 mins at 5,000 g after collection to remove cells. The tear samples were stored at −80 °C until further use. Fresh tear samples collected from humans and rabbits were added directly to a pre-tared micro-extraction vial (average of 2.6 mg) and spiked with 5 µL (98 ng) of D 6 -cholesterol internal standard solution (0.05 mM in chloroform). Chloroform:methanol (200 µL; 2:1 vol/vol) containing 0.01% BHT was added to the tear sample and the mixture vortexed. Aqueous ammonium acetate (0.15 M; 25 µL) was added and the mixture was then centrifuged (800 g, 5 mins). The aqueous phase was removed and the organic phase dried under nitrogen. The sample was reconstituted in 50 µL chloroform:methanol (1:2 vol/vol; 0.01% BHT). 28 All samples were stored at -80 °C until they were analysed.

Sample preparation and lipid extraction
To test for matrix effects 3 tear samples (12.1 mg) were pooled in 1 ml of chloroform:methanol (200 µL; 2:1 vol/vol) containing 0.01% BHT. A stock solution of chloroform (250 µL) containing D 0 -cholesterol (931 ng) and D 6 -cholesterol (952 ng) was added and the mixture was vortexed, divided into 5 aliquots (250 µL) and extracted as described above. The extract results were compared with each other and with those obtained for the stock solution used to spike the pooled sample to evaluate the reproducibility of the assay and to identify potential matrix effects.
Absolute sample recovery was established using a 15 µL aliquot of standard solution of cholesterol (0.05 mM).

Mass spectrometry
Electron ionization mass spectrometry was conducted using a single quadrupole GC-MS system (Shimadzu QP5050, Kyoto Japan) fitted with a heated direct-insertion sample probe (DI-50).
Pasteur pipettes were used to load a drop (ca. 8 µL) of each tear extract or standard sample to one end of separate and disposable heat-sealed melting-point tubes. The hanging drop was allowed to dry to a film at the tip of the probe prior to introduction into the ion source via a vacuum lock where any residual solvent is removed. Programmed heating (40-250 °C at 80 °C/min) of the probe in the vacuum environment of the ion source resulted in the sublimation of free cholesterol. The maximum release occurred at ca. 0.9 ± 0.2 min (112-128 °C) with small inter-sample variations in the thermaldesorption profiles arising from differences in lipid coating thickness and film distribution after drying. Electron-ionisation mass spectrometry was undertaken during the temperature-programmed desorption using either (i) a full scan (m/z 40-600, DI/EI-MS) for qualitative comparison of the analytes present in the sample that are thermally stable and sufficiently volatile to be vaporized into 8 the ion source 20, 28 or (ii) selected ion monitoring (DI/EI-SIM) of specific ions diagnostic for cholesterol and its D 6 -isotopologue for the purposes of quantification.
For comparison to electrospray ionization-mass spectrometry (ESI-MS), 3 samples of meibum (collected as previously described 31 ) were dissolved in chloroform (1000 µL) containing 0.01% BHT were prepared in glass vials. 100 µL of sample was added to 10 µL   Figure 1c). An equivalent desorption profile obtained from an extract of a pooled tear sample is shown in Figure 2(d). These data clearly indicate that even in this complex matrix the early desorption feature in the m/z 368 and 386 channels arises exclusively from cholesterol with no contribution from other compounds, including cholesterol esters that are clearly resolved at longer analysis times (i.e., higher desorption temperatures). Furthermore, the results provided as supporting information ( Figure S2) show that for the analysis of free cholesterol each DI/EI-SIM analysis can be aborted after ca. 1 min (corresponding to ca. 120 °C) immediately following the maximum current for the diagnostic ions. This significantly reduced analysis time allowing the DI-probe to be withdrawn, the probe tip discarded and the probe quenched in cold ethanol ready for the next analysis. The procedure thus designed, was found to reduce analysis time, minimise sample carry-over.

Cholesterol quantification and establishing recovery and detection limits
As the full scan D 6 -cholesterol mass spectrum showed fragmentation behaviour analogous to To investigate the accuracy of the DI/EI-SIM approach to quantification a parallel analysis of free cholesterol was undertaken using electrospray ionization tandem mass spectrometry (ESI-MS/MS). Lipid extracts from human meibum were used in this comparative analysis as these represent an even more complex matrix that that found in tears. Comparison of the same three meibum extracts by DI/EI-SIM and ESI-MS/MS gave a regression coefficient (R 2 ) value of 0.987 and slope of 1.075 between the two assay techniques (see Supporting Information Figure S).

Cholesterol in human and rabbit tears
Free cholesterol in human and rabbit tears was quantified using the developed method. Tears (ca. 5 µL) were extracted using the protocol described above. The concentration of cholesterol in the human tears was found to be 15 ± 6 µg/g which was higher than that of the rabbit tears (10 ± 5 µg/g; p < 0.05; Figure 5).

Discussion
This study has established a method to rapidly quantify free cholesterol in lipid extracts from tears using DI/EI-SIM. Using this method, the lipids in tear samples were extracted directly into organic solvents and the crude extracts were analysed directly with no requirement for chromatographic separation. This represents a significant time-saving over traditional GC and HPLC based methods, i.e., less than 2 versus up to 25 mins. The simplicity of the purification steps minimise the potential loss of analyte during sample preparation with the absolute recovery found to be 72.1%. Since the sample is spiked with the stable isotope D 6 -cholesterol as internal standard prior to extraction any loss of analyte during extraction is accounted for in the quantification. The loss of any sample during extraction will not affect the analytical results, although there may be a slight reduction in sensitivity.
The LOQ for the analysis was found to be 8 ng (~20 pmol) of cholesterol, for which a minimum volume of 1 µL of tears was required. The elution profiles of the DI/EI-SIM data for each sample varied only slightly due to differences affecting the thermal desorption of the sample such as the distribution of the cholesterol film on the surface of the probe. The method provides a highly sensitive way to measure free cholesterol in small volumes of clinical samples and has the potential to provide new insight into the role of free cholesterol in tear film instability and dry eye syndrome.
If necessary, a concentration step (solvent evaporation) could be used to increase the detection limit for clinical samples (i.e., if dry-eye samples had much less cholesterol than normal samples).
Another advantage of this method is the ability to increase the sensitivity of analysis by adding another drop of sample from the same extract on top of a previously dried sample film. This means that the whole extract (ca. 50 µL) is potentially available for analysis. However, in each case high purity of the solvents is essential to avoid increasing background signals, particularly with multiple sample loadings. Nevertheless, as quantification was based on the ratio of ion peaks in the same sample, the sample loading volume will not affect the results, again improving assay accuracy.
One of the commonly used methods for measuring low concentration of cholesterol is GC-MS. Although GC-MS is sensitive, the analysis is time-consuming and not suitable for a large number of clinical samples. Additionally, since the process involves silylation and hydrolysis of cholesterol esters, 29 the measurements will be of free cholesterol and cholesterol generated from its esters. By comparison, the method outlined here provides a much more rapid means to analyse samples for evaluating only free cholesterol without derivatisation. Ruiz et al. 29 determined the cholesterol concentration by comparing the silylated cholesterol to 5-α-cholestane by peak-area ratio and found a mean cholesterol concentration of 146 ± 58 ppm in human tears and 59 ± 30 ppm in rabbit tears. Whilst the results of the current study agree with their findings that human tears contain a higher level of cholesterol content than rabbit tears, the actual concentrations do not agree. This may be because they measured free cholesterol and cholesterol hydrolysed from its esters during extraction and/or derivatization procedures. As has recently been shown, cholesterol esters are a major component of the tear and meibum lipidomes 30,31 and the concentration of cholesterol esters in tears is significantly higher than free cholesterol 11 . Given these complex matrix effects, understanding and rigorously excluding contributions from cholesterol esters to measurement of free cholesterol, as shown here, is essential for reliable quantification in tears.
Human and rabbit tear films differ markedly in their relative stability, 4 with an inter-blink time of several minutes for rabbit compared to about 10 seconds for human. [5][6][7] The factors which contribute to the stability of the rabbit tear film are not well understood and an understanding of this may lead to the development of novel therapeutic interventions for human diseases resulting from tear film instability. Although cholesterol is of relatively low abundance in meibomian gland secretions, cholesterol has been associated with the tear film instability in several eye diseases. [15][16][17] This study demonstrated that cholesterol levels were lower in the tear film of rabbits compared to humans. The implications of the difference need further study.   Figure 5. The concentration of free cholesterol in human (n = 15) and rabbit (n = 10) tears. The concentration of free cholesterol in human tears was significantly higher than that in rabbit tears (p < 0.05).