Species Variation in Small Molecule Components of Animal Vitreous

Abstract

Purpose.: We characterized differences in biochemical composition of the vitreous of different animal species with respect to small molecule constituents.

Methods.: Vitreous samples were extracted from sheep, pig, Dutch Belted rabbits, and New Zealand white rabbits. The vitreous samples were investigated for acetylcholinesterase (AChE) activity and, in addition, were subjected to metabolomics determination using mass spectrometry.

Results.: AChE activity varied across the species investigated with greater activity noted in larger animals. Principal component analysis demonstrated species differentiation in relation to metabolomic profile. Key peaks identified the importance of animal diet on small molecule composition of the vitreous.

Conclusions.: Our results highlighted principal and consistent differences in small molecule composition and enzymatic activity of the vitreous depending on species. Interesting differences were demonstrated, showing that diet potentially can impact on components of and metabolites contained within the vitreous. Material will be exchanged between vascular and retinal tissue with the vitreous compartment and as a nonvascular, slowly equilibrating “sink” might reflect changes in transporter activity. As a first step, understanding the differences in the metabolic profile of vitreous from different species may impact interpretation of such activity across different species.

Introduction

The vitreous is the largest structure of the eye, separating the lens from the retina. The vitreous is composed of approximately 99% water; however, unlike water it is viscoelastic in nature and has a gel-like structure, due to the presence of collagen, hyaluronan, and proteoglycans. ¹ The vitreous is responsible for the support and protection of the retina and lens. It also acts as a diffusional barrier to the free movement of molecules between the anterior and posterior eye, and ensures that light can pass through the posterior eye unrestricted. In addition, it is thought to act as a reservoir for compounds that are unable to cross over the blood-retinal barrier, such as proteins and other solutes, ² and it often is used as a drug reservoir in the treatment of posterior eye disease. Water molecules are associated with hyaluronan, which is dispersed between linear strands of collagen fibers, resulting in a cross-linked polymer network. ³ The distribution of hyaluronans varies depending on position within the vitreous, with higher concentrations of hyaluronans found just behind the lens, and this results in an increased viscosity compared to the vitreous closest to the retina. ⁴ Alongside viscosity, the rigidity of the vitreous also varies depending on position, with an increase in rigidity found around the perimeter, due to the predominance of collagen. ⁵ Although the structure of the vitreous permits small molecules to diffuse freely through, as a result of the vitreous structure created by collagen and hyaluronans, a diffusional barrier to the movement macromolecules exits. 

The biochemistry and physiology of the vitreous have continued to be of interest, not only focusing on collagen and hyaluronan, but also investigating the amino acid and protein composition of the vitreous. ^6–8 Various techniques have been used in an attempt to generate a metabolomic profile of the vitreous, including nuclear magnetic resonance (NMR) ⁹ and mass spectrometry (MS). ¹⁰ Using NMR and multivariate analysis to generate a metabolic profile of the human vitreous, differences in biochemical nature of the vitreous in two different forms of uveitis, lens-induced uveitis and chronic uveitis, were identified, enabling segregation of different forms of uvetitis. ⁹ MS also has been used to identify changes in the biochemistry of the vitreous in disease, with increased protein expression demonstrated in the vitreous of eyes affected by diabetic retinopathy. Alongside this work, specific enzymatic activity of the vitreous also has been considered and has identified the presence of enzymes, such as acetylcholinesterase (AChE) ¹¹ and degradative proteinases in the human vitreous. ¹² Although our understanding of the biochemical nature of the vitreous has improved, studies have focused mainly on differences occurring in the presence of disease. 

Animal models are extremely important in ocular research and are used widely, involving studies to understand ocular tissue physiology, ¹³ model ocular disease, ¹⁴ and investigate ocular drug distribution. ¹⁵ A wide range of species have been used in ocular vitreous studies, including the frog vitreous, ¹⁶ bovine vitreous, ¹⁷ ovine vitreous, ¹⁸ and the most commonly used rabbit vitreous. ¹⁹ Often data from animal models are used interchangeably, especially in extrapolating data from pharmacokinetic studies from one species to another. Although a wide range of animal models currently are used, information comparing directly the biochemical nature of vitreous and identifying differences across species is limited. Differences in the biochemistry of the vitreous tissue potentially could impact on the movement and metabolism of molecules around the vitreous. Our original reasoning was that differences in enzymatic activity of AChE might be reflected in the small molecule composition of the vitreous. However, during our study an opportunity for a broader examination of the biochemistry of the vitreous fortuitously presented itself. In our study, we highlighted, for the first time to our knowledge, the principal and consistent differences in small molecule composition and enzymatic activity in the rabbit, pig, and sheep vitreous tissues, demonstrating the potential influence of diet on the metabolic profile of the vitreous. In addition, we also identified key biochemical differences in the composition of the vitreous in different rabbit subspecies. 

Methods

Vitreous Preparation

Sheep and pig eyes were obtained from a local abattoir (Carluke, Scotland). The sheep and pigs used in the study originated from the same herd, and were of approximately the same age, with the pigs and sheep both aged six months. New Zealand White (NZW) rabbit eyes (British) were obtained from the in house animal unit (Glasgow, Scotland). Dutch Belted (DB) rabbit vitreous and NZW rabbit (American) vitreous were obtained (pre-extracted and stored at −80°C within 2 hours following death) from Allergan (Irvine, CA). All rabbits used in our study were 6–12 months old. All animals were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Visual Research. For the intact sheep, pig, and NZW rabbit eyes (British), the vitreous tissues were removed and frozen within 2 hours following death. In the first instance, the anterior section was removed by creating a small incision into the sclera, close to the edge of the cornea. Using the incision created as a positional marker, the anterior eye then was cut away in a circular dissection, around the edge of the cornea. The vitreous then was lifted carefully away from the exposed posterior section of the eye, ensuring that cross contamination from other posterior ocular tissues did not occur. The vitreous samples then were snap frozen in liquid nitrogen before storage at −80°C. Before analysis, the vitreous samples were thawed at room temperature and then vortexed for 2 minutes to break down the gel phase. Following this, the samples were centrifuged at 13,200 revolutions per minute (rpm) for 10 minutes and the supernatant removed for analysis. 

Before liquid chromatography-high resolution mass spectrometry (LC-HRMS) analysis, a further sample preparation was performed. For each species 250 mg of vitreous sample was weighed and 0.75 mL of acetonitrile (Fisher Scientific, Loughborough, UK) added to enable protein precipitation. The sample then was vortexed for 1 minute and then centrifuged at 10,000 rpm for 10 minutes to pellet the precipitated proteins. As small molecule constituents of the vitreous were of interest in our study, the pellet was discarded, and the supernatant then was removed and transferred to a glass vial for injection into the LC-HRMS. Analysis was performed on samples from 10 sheep, 10 NZW (British) rabbits, 6 pigs, 4 DB rabbits, and 4 NZW (American) rabbits. 

AChE Assay

AChE assay determination was performed using an AChE assay kit method (Amplex Red Assay Kit, Molecular Probes; Life Technologies, Paisley, UK).The assay kit provides a method for detecting AChE activity through the determination of hydrogen peroxide produced through a series of chemical reactions. The reaction is initiated by the conversion of acetylcholine to choline by AChE, and leads to the generation of hydrogen peroxide, which in the presence of horseradish peroxidase reacts with Amplex Red to produce the fluorescent compound resourufin. ²⁰ A standard protein calibration curve was prepared using AChE from electric eel at concentrations of 160, 120, 80, 60, 40, 30, 20, and 10 mg/mL in 50 mM Tris-HCl buffer (pH 7.5; Sigma-Aldrich, Dorset, UK) in a total volume of 100 μL. Then, 10 μL of each calibration standard were added to a row in a microwell plate and, to another row, 10 μL of 30 μM tacrine (Sigma-Aldrich) prepared in Tris-HCl buffer were added to each well as AChE standard to monitor activity. Then, 10 μL samples of prepared animals vitreous were added to three wells for each animal vitreous sample. To each well, 20 μL of the detection system containing 10 μM acetylcholine chloride, 20 U/mL choline oxidase, 400 U/mL horseradish peroxidase, 400 U/mL Amplex Red (prepared in 50 mM Tris-HCl buffer, pH 7.5) were added and mixed thoroughly on a shaker plate for 5 minutes. Following incubation for 20 minutes, the fluorescence was determined in a fluorescence microplate reader (Victor2; PerkinElmer, Waltham, MT) using an excitation wavelength of 560 nm and emission wavelength of 590 nm. 

Total protein determination also was done using the Bradford protein assay method described previously. ²¹ A standard protein calibration curve was prepared using BSA (Sigma-Aldrich) at concentrations of 1.4, 1.2, 1, 0.8, 0.6, 0.4, and 0.2 mg/mL in PBS (Sigma-Aldrich) in a total volume of 50 μL. Then, 10 μL of each calibration standard were added to a row in a microwell plate. In addition to the calibration standards, 10 μL samples of prepared animals vitreous were added to three wells for each animal vitreous sample. To each well, 200 μL of Bradford reagent (Sigma-Aldrich) were added and mixed thoroughly on a shaker plate. The absorbance of the prepared calibration samples and vitreous samples finally were measured at 595 nm using a UV/Vis microplate reader (Spectromax 190; Molecular Devices, Sunnyvale, CA). 

To test for statistically relevant changes in vitreous tissue across animal groups, one-way ANOVA followed by Tukey's post-hoc analysis was performed on the data using Minitab Statistical Software version 16. Differences in tissue concentrations were considered significant when the P value obtained was less than 0.05. 

LC-HRMS Analysis

Measurements were performed on a Dionex Ultimate 3000 high performance liquid chromatography (HPLC) system comprising a degasser, binary pump, and temperature-controlled autosampler (Dionex, Camberley, UK) combined with an Exactive (Orbitrap) mass spectrometer (Thermo Fisher Scientific, Bremen, Germany). Nitrogen, was used as the sheath and auxiliary gas, and was produced by a NM32LA nitrogen generator (Peak Scientific, Scotland, UK). Separation was achieved on a ZIC-HILIC column (150 × 4.6 mm, 5 μm; Merck, Hichrom, UK) coupled with a guard ZIC-HILIC column at a flow rate of 300 μL/min using an elution gradient of 0.1% (vol/vol) formic acid (Sigma-Aldrich) in H₂O (reagent A) and 0.1% (vol/vol) formic acid in acetonitrile (ACN; reagent B; Fisher Scientific). The gradient program started with the percentage of reagent B decreasing from 80–50% over the first 12 minutes, followed by an isocratic condition held for 14 minutes, before dropping down to 20% reagent B for 10 minutes to ensure column elution. Finally, the ratio of reagent B was increased back to 80% for column re-equilibration in the final 10 minutes. The HPLC stream flowed into an ESI interface operating in a positive/negative polarity switching mode. The spray voltage was 4.5 kv for positive mode and 4.0 kv for negative mode. The temperature of the ion transfer capillary was 275°C, and sheath and auxiliary gas flow rates were 50 and 17 arbitrary units, respectively. The full scan range was performed from 75–200 key mass peaks (m/z) for positive and negative modes, with settings of AGC target and resolution set to balanced and high, respectively. The data were recorded using Xcalibur 2.1.0 software package (Thermo Scientific, Northumberland, UK). Mixtures of standard compounds were measured simultaneously alongside real samples. 

Data Processing

Raw data were sliced to positive and negative data sets using the RecalOffline tool in the Xcalibur software package, before importing them into MZMine 2.2 software (VTT, Laskut, Finland). After the process of chromatogram building, chromatogram deconvolution, de-isotope, alignment, gap filling, and identification (adduct and complex search) were performed to convert each data set into hundreds of extracted unique chromatographic features (peaks), based on the combination of mass signal and retention time. Some important settings for these processes were noise level 10,000, minimum peak height 50,000, m/z tolerance 0.001, minimum time span 20 seconds, signal-to-noise level 3, and retention time tolerance 30 seconds. After removing a few non–peak-shape features by visualization, both data sets were exported from MZMine 2.2 to an Excel workbook (Microsoft, Berkshire, UK) for further statistical analysis. 

Data Analysis

Positive and negative data sets were merged, and the features were assigned as metabolites based on the accurate masses with a threshold of ±3 parts per million (ppm) from an in-house database, which included The Human Metabolome Database, KEGG Database, Metlin Database, and MassBank Database. Some assignments were identified by comparison of their retention times with the standards. Species were compared for those assigned features using ratios of averages using the peak areas. Relative standard deviations (RSDs) of peak areas within one species also were calculated. Finally, the features were signed by detectable polarity (P = positive only, N = negative only and P/N = both). To get a perceivable understanding of the results, principle component analysis also was done using SIMCA-P+ 12.0 (Umetrics AB, San Jose, CA). The obtained score plot and loading plot were analyzed. Interesting peaks identified from the loadings plot then were subject to one-way ANOVA followed by Tukey's post-hoc analysis. Differences in tissue concentrations were considered significant when the P value obtained was less than 0.05. 

Results

To generate a biochemical profile of the vitreous for each animal type, investigations using various techniques were done on each tissue. In the first instance the AChE activity of the vitreous was determined in relation to the total protein activity of the tissue (Fig. 1). AChE determination was performed on vitreous samples from NZW British rabbits (n = 10), DB rabbits (n = 4), NZW American rabbits (n = 4), sheep (n = 10), and pigs (n = 5), in triplicate. AChE activity varied across the range of animal vitreous investigated, with the vitreous tissue of the pig shown to have the greatest AChE activity, followed by the sheep vitreous. Differences were noted across animal subtype, with DB rabbits shown to have greater activity compared to the NZW rabbits. Similar levels of AChE activity were noted in both subtypes of NZW rabbits with no significant difference noted between the American and British NZW rabbits. 

Figure 1. 

 

Mean AChE activity (mU) of 10 μL of vitreous tissue standardized by total protein activity (mg) of 10 μL of vitreous tissue, for each animal subtype. NZW-B, NZW British-sourced rabbits (n = 10); DB, DB rabbits (n = 4); NZW-A, NZW American-sourced rabbits (n = 4). Sheep vitreous (n = 10) and pig vitreous (n = 5) results also are shown. Error bars show the calculated SE from the mean. *Significant difference in mean result compared to all other groups is denoted.

In addition to AChE determination and to generate a metabolic profile of each of the tissue samples, LC-HRMS analysis was performed on each of the animal vitreous tissues for NZW British rabbits (n = 10), DB rabbits (n = 4), NZW American rabbits (n = 4), sheep (n = 10), and pigs (n = 5). Following data acquisition and initial processing to identify and align each peak, data analysis was performed using the in-house metabolomics database. Chromatographic peaks were assigned as metabolites based on the accurate masses. For each assigned peak for all of the vitreous samples, the ion intensity obtained was derived and tabulated. Due to the volume of data obtained, principal component analysis (PCA) was used to determine similarities across vitreous samples by group clustering, and identify the metabolites responsible for group clustering. PCA is a useful method of establishing variance within a data set, shown by the scores plot, and can determine the origin of the variance, through the loadings plot. Principal components (PC) 1 and 2 demonstrated the best discrimination across vitreous animal species while covering 65% of the cumulative variance. The scores plot and loadings plot obtained from performing PCA on the complete data set are shown in Figures 2 and 3, respectively. From the scores plot it was clear that the vitreous samples cluster together depending on species, with the sheep samples grouping together, the pig samples grouping together, and the rabbit vitreous samples grouping together, demonstrating that clear differences in the metabolomic profile existed between vitreous species. 

Figure 2. 

 

The scores plot for PC 1 versus PC 2 using positive ion spectra obtained for all vitreous tissue samples. Grouping NZW-B, NZW (British) rabbits; Grouping DB, DB rabbits; Grouping NZW-A, NZW (American) rabbits; Grouping S, sheep, and Grouping P, pig.

Figure 3. 

 

The loadings plot for PC 1 versus PC 2 demonstrating ion fragments controlling sample discrimination for positive and negative ion spectra.

The loadings plot (Fig. 3) contains the mass peak data used to generate the scores plot and identifies the mass peaks responsible for variation between sample clusters. The distance of the mass peak from the origin gives an indication of the mass peaks responsible for pulling the samples to their designated position on the scores plot. A range of metabolites were identified in the loadings plot and included a number of different types of analytes, including amino acids, lipids, and sugars. A number of key interesting mass peaks were identified from the loadings plot, and were selected to focus on the structural assignments of the mass peaks (see Table). 

Table. 

 

Key Mass Peaks (m/z) Responsible for Sample Separation on the Scores Plot, Identified by the Loadings Plot

ESI Mode	Key Mass Peaks (m/z)	Assignment	Role
Negative	111	Uracil	DNA base
Negative	131	6-Hydroxyhexanoic acid	Fatty acid
Positive	137	4-Hydroxybenzoate	Antioxidant
Positive	144	Stachydrine	Plant metabolite
Positive	152	Guanine	DNA base
Positive	157	Allantoin	Uric acid metabolite
Negative	215	2-C-methyl-D-erythritol-4-phosphate	Enzyme: steroid production

 Mass peak assignment is detailed and the role of the fragment in the tissue is listed.

In positive ion spectra m/z of 137, 144, 152, and 157 were identified, and in the negative spectra 111, 131, and 215 were selected. An m/z of 111 on the negative spectra and 152 on the positive spectra were confirmed to be DNA base pairs uracil and guanine, respectively, and were eluted at retention times of 8.25 and 12.5 minutes. Figure 4 shows the mean ion count obtained across the animal species. Although greater levels were seen in the pig and DB rabbits, this result was not statistically significant. Two important metabolites were identified on the positive spectra to be stachydrine (m/z 144, retention time 15.19 minutes), a metabolite originating from plant material, and allantoin (m/z 157, retention time 13.07 minutes), a metabolite of uric acid. Figure 5 shows the mean ion count obtained for these two metabolites across the species. Interesting, although not statistically significant, differences were shown for allantoin, with the pig showing a 3-fold increase in ion count compared to the sheep. The DB rabbits again showed greatest similarity in ion count to the pig and a 4-fold increase in ion count compared to both groups of NZW rabbits. On the other hand, stachydrine showed considerable differences across species, with low levels obtained in the pig and all three rabbit types, and statistically significant greater levels in the sheep vitreous. 

Figure 4. 

 

Mean uracil and guanine ion count for each animal subtype. Error bars show the calculated SE from the mean.

Figure 5. 

 

Mean allantoin and stachydrine ion count for each animal subtype. Error bars show the calculated SE from the mean. *Significant difference in mean result compared to all other groups is denoted.

A similar relationship was demonstrated for the antioxidant 4-hydroxybenzoate (m/z 137, retention time 5.9 minutes) with statistically significant higher levels shown in the sheep vitreous compared to the vitreous from the other animal species (Fig. 6). In the negative spectra, m/z 131 was eluted at a retention time of 6.23 minutes and determined to represent the fatty acid, 6-hydroxyhexanoic acid (Fig. 6). An interesting difference was noted for this compound with a statistically significantly greater ion count found in the samples from the American NZW rabbits. Across the other species, similar levels of 6-hydroxyhexanonic acid were seen, with the ion count in the American NZW rabbits 4-fold greater than the ion count obtained in the British NZW rabbits. The final ion identified was m/z 215 on the negative spectra, which eluted at a retention time of 14.52 minutes and was determined to be 2-C-methyl-D-erythritol 4-phosphate (Fig. 7). No statistically significant difference across mean ion count for the difference species types was shown, with similar mean values obtained for the sheep and British NZW rabbits, and similar means obtained for the pig, DB rabbits and American NZW rabbits. 

Figure 6. 

 

Mean 4-hydroxybenzoate and 6-hydroxyhexanoic acid ion count for each animal subtype. Error bars show the calculated SE from the mean. *Significant difference in mean result compared to all other groups is denoted.

Figure 7. 

 

Mean 2-C-methyl-D-erythritol-4-phosphate ion count for each animal subtype. Error bars show the calculated SE from the mean.

Discussion

AChE is an enzyme responsible for the degradation of the neurotransmitter acetylcholine into choline and acetic acid, but contributes to other nonspecific esterase activity. Acetylcholine is found in significant concentrations in the retina, as the transmitter is released during stimulation of nicotinic and muscarinic receptors. ^22,23 AChE has been shown to be present in the connective tissues of the eye, and is responsible for 70% of the total cholinesterase activity of the vitreous and 51% of the total cholinesterase activity of the aqueous humor. ¹¹ AChE found in the vitreous is likely to originate from transmitter termination activity in the vitreous. The species differences in AChE content is interesting, with larger animals having greater AChE activity, which may reflect the volumetric differences between species, since the enzyme will clear more slowly in a larger eye. It suggested that the vitreous might conserve evidence of retinal function, for example associated with transporter activity. Differences in the biochemical composition of the vitreous were detected across the samples. In part, this was expected as a result of the differences in AChE, but the analysis revealed unrelated components showing consistent differences that alerted us to the “finger-printing” of vitreous according to species. This was carried out by a “blind” screen. 

Before performing PCA on the chromatographic peak data set obtained, a data processing set first was carried out. The application of an appropriate data processing step can improve the quality of data obtained vastly by PCA, and selection of the most suitable method often is dependent on the individual data set. ²⁴ In this instance a mean centering approach to data processing was used, and enabled discrimination between the vitreous tissues of different species and the biochemical markers controlling this discrimination to be identified. Obvious candidates were nucleic acids, and the vitreous content of DNA and RNA base pairs guanine, a purine, and uracil, a pyrimidine, was found to vary between species. Previously, vitreous sample storage conditions following death have been shown to have an impact on uracil levels. In the chicken eye, uracil content has been shown to differ depending on the postmortem holding temperature, with uracil content shown increased 3-fold when the tissue storage temperature was increased from 6°C–30°C. ²⁵ In addition, uracil concentration in the vitreous has been demonstrated to differ with age in the chicken eye, with higher concentrations found in vitreous tissue extracted at birth compared to a 10-week-old chicken vitreous, when both samples were stored at 20°C; however, this difference was not statistically significant. ²⁶ Although it is apparent that changes in the components of the vitreous can occur post mortem, in both literature cases the vitreous tissues were exposed to extremes of temperature, which was not the case in our study. As quickly as possible following death, the vitreous tissues were extracted and frozen until immediately before use. The effect of chilling and freezing on the chemical constituents of the vitreous has been shown previously to have little effect on the concentrations of constituents of the vitreous, with the exception of some ions, namely potassium and magnesium. ²⁷ It is anticipated that, as the samples were immediately frozen within 2 hours following death, the effect of storage time would not be as significant as the effect in tissues stored at high temperatures. Although the stability of metabolites from biologic samples can vary during freezing between compounds, biologic variability experienced in metabolomic studies using mass spectrometry has been shown previously to be greater than analytical variability experienced during the freeze thaw process. ^28,29 However it would be beneficial in future studies to investigate the impact of the freeze-thaw process on the metabolomic profile of the vitreous of different animal species. 

Allantoin is a metabolite formed during the breakdown of purines in animals. Guanine is converted to uric acid by xanthine oxidase, which then is broken down further into allantoin by uridase. In humans, this final step is not present and uric acid is excreted without further breakdown. Allantoin has been measured previously in human plasma; however, the origin is thought to be through the free radical activity of uric acid, in response to oxidative stress. ³⁰ The measurement of guanine and allantoin in the vitreous indicates purine metabolism in the vitreous or, alternatively, that the end products of metabolism in the neighboring retinal tissue are pumped from the retina into the vitreous. In sheep urine, allantoin exists from three possible sources, from fed purines from the diet, from purine bases generated from ruminal microorganisms, or purines originating from tissue turnover. ³¹ The differences found in allantoin content of vitreous across species could be related to diet and coincide with the differences noted in stachydrine levels measured. Stachydrine (also know as proline betaine) is an alkaloid constituent in plant material, with high content found in some legumes and citrus fruits. ³² Diet has been shown to have an effect the concentration of stachydrine measured in human biologic fluid, alongside glycine betaine. In a three-year follow up study, correlation in the concentrations of glycine betaine measured in plasma were demonstrated with an R ² value of 0.65 determined using Pearson's correlation coefficient. Conversely for stachydrine, no correlation was shown (R ² = 0.02) leading Lever et al. to suggest that stachydrine measured in biologic tissue was related to dietary factors and was not thought to have a role in metabolism or be subject to homeostatic control. ³³ It is likely that the differences seen across the species investigated are due to dietary factors. Sheep are, of course, vegetarian and in Scotland mainly live outside feeding on plants and grass. On the other hand, rabbits are fed controlled laboratory diets and this will vary between laboratories, whereas pigs are omnivorous and, as a result, eat a wide range of food groups. Typically, the pig diet would include grains and some form of protein. Another compound identified to have the same relationship as stachydrine in abundance across the vitreous species studied was 4-hydroxybenzoate; 4-hydroxybenzoate is formed by the oxidation of 4-hydroxybenzoic acid, which is a metabolite of a group of compounds called the parabens. The parabens are found commonly in plant material, mainly fruits, but also are used as an antimicrobial preservative in some foods. The parabens are absorbed readily from the gastrointestinal tract and are metabolized by esterases into 4-hydroxybenzoic acid. ³⁴ It is likely that this ion originates from a paraben ingested and then absorbed by the gastrointestinal tract; however, it would be difficult to determine exactly which paraben it has came from. Following absorption, the paraben then would be subject to enzymatic cleavage in the plasma to 4-hydroxybenzoic acid, before reaching the vitreous. ³⁵ As a similar relationship to stachydrine was demonstrated, it is likely that the difference in 4-hydroxybenzoate prevalence seen across the species investigated again is related to the diet. In addition, 2-C-methylerythritol-4-phosphate also was identified as an important mass peak in the PCA; 2-C-methylerythritol-4-phosphate is an intermediate compound involved in the nonmevalonate metabolic pathway, which occurs in plants and bacteria. The differences in levels of this metabolite across the species investigated could be associated with the presence of bacteria in the tissue; however, as the samples came from different sources and were frozen immediately after enucleation, the presence of this compound likely is to be again in relation to diet. 

The final metabolite influencing the PCA discussed is 6-hydroxyhexanoic acid. This metabolite is formed during the breakdown of the polymers, such as polycaprolactone. ³⁶ It is interesting that NZW (American) rabbits had significantly greater levels of 6-hydroxyhexanoic acid than all the other species investigated, especially the NZW (British rabbits), where for all the other metabolites identified, mean levels of NZW rabbits appeared similar. The origin of this material is unknown, but identification helped us to speculate that PCA might be a useful technique in understanding the remnant chemistry of polymeric implants. 

Our results highlighted, for the first time to our knowledge, principal and consistent differences in small molecule composition in the rabbit, pig, and sheep vitreous tissues. Striking differences were demonstrated with regard to the nature of soluble components contained within the vitreous, in part perhaps related to differences in enzymatic activity of the vitreous, but also to diet. Understanding these consistent differences in the small molecule profile of the vitreous composition might assist in the interpretation of data derived from different animal models. 

Footnotes

 Disclosure: J. Mains, None; L.E. Tan, None; T. Zhang, None; L. Young, None; R. Shi, None; C. Wilson, None