Primary congenital glaucoma (PCG) refers to thedisease condition that results from the elevation of the intraocular pressuredue to obstruction of aqueous outflow by an isolated maldevelopment of theaqueous outflow pathways (Stamper RL, Lieberman MF, Drake MV. Developmental and childhood glaucoma. In: Stamper RL, Lieberman MF, Drake MV, eds. Becker-Shaffer’sDiagnosis and Therapy of the Glaucomas, 8th ed. Amsterdam: Mosby, Elsevier;2009:294-329.).
The clinical features of the disease encompass thecausative pathology of goniodysgenesis –detected by gonioscopy- as well as thesecondary effects in the eye, including elevated intraocular pressure (IOP),enlarged –possibly cloudy- cornea with a broad limbus, a deep anterior chamber (AC)–readily accessible for aqueous humour (AH) sampling- and an enlarged axiallength, among others (Fernandez LM, Martinez-de-la-Casa JM, Garcia-Bella J,Mendez C, Saenz-Frances F, Garcia M, Escribano J, Garcia-Feijoo J. ClinicalVariability of Primary Congenital Glaucoma in a Spanish Family With Cyp1b1 GeneMutations. (J Glaucoma 2015;24:630–634).The eye develops in utero as an outpouching of theforebrain that forms the primitive optic vesicle. The anterior segment of theeye is a descriptive term for the part of the eye that includes and liesanterior to the pars plicata of the ciliary body. This anterior segment of theeye develops from all 3 germ layers of the embryo, namely surface ectoderm,mesoderm, endoderm, as well as neural crest cells (Abdo,M.
, Haddad, S. and Emam, M. (2017), Development of the New Zealand White RabbitEye: I. Pre- and Postnatal Development of Eye Tunics. Anat.
Histol. Embryol.,46: 423–430. doi:10.1111/ahe.12284423–430.
doi:10.1111/ahe.12284). Theangle of the AC and specifically the trabecular meshwork (TM) develops fromneural crest cells that invade the area of the angle of the AC around thesecond gestational month (Seefelder R, Wolfrum (1906) Zur Entwicklung derVorderen Kammer und des Kammerwinkels beim Menschen nebts Bemerkungen uber ihreEntstehung bei Tieren. Grafes Arch Ophthl 63:430–451). Development of theanterior segment of the eye is under the influence of a number of genes throughtheir secretory proteins, including PAX6, PAX2, BMP4 and the sonic hedgehogprotein, just to name some (Lupo G, Andreazzoli M, Gestri G, Liu Y, He RQ, Barsacchi G.
Homeobox genes in the genetic control of eyedevelopment. Int J Dev Biol. 2000;44(6):627-36.).The composition of the aqueous humour (AH) includeselectrolytes, organic solutes, growth factors, cytokines, and proteins (To CH, Kong CW, Chan CY, ShahidullahM, Do CW. The mechanism of aqueous humour formation.
Clin Exp Optom.2002;85:335–349, Freddo TF. The Glenn A. Fry Award Lecture 1992: aqueous humorproteins: a key for unlocking glaucoma? Optom Vis Sci. 1993;70:263–270, McLaren JW, Ziai N,Brubaker RF.
A simple three-compartment model of anterior segment kinetics. ExpEye Res. 1993;56:355–366), the concentration of the later being quite small, in the range of 120 to500 ng/?L and although derived from plasma proteins, are structurally andfunctionally different. Little is known about the protein composition of the AH,with less than 150 proteins identified so far by different techniques includingdifferential protein expression and Multidimensional Protein IdentificationTechnology (MudPIT) (Richardson MR, Price MO, Price FW, et al. Proteomicanalysis of human aqueous humor using multidimensional protein identificationtechnology. Mol Vis. 2009;15:2740–2750).
One robust method currently availableto analyse peptides, proteins and most other biomolecules –up to the femtomolelevel– is the Matrix-assisted laser desorption/ionization time-of-flight massspectrometry (MALDI-TOF-MS) (Matrix-assisted Laser Desorption/Ionization MassSpectrometry in Peptide and Protein Analysis J. Kathleen Lewis, Jing Wei,and Gary Siuzdak in Encyclopedia of Analytical Chemistry R.A. Meyers (Ed.
) pp.5880-5894 . John Wiley & Sons Ltd, Chichester, 2000, Int J Med Sci. 2011; 8(1): 39–47. MALDI-TOF MSCombined With Magnetic Beads for Detecting Serum Protein Biomarkers andEstablishment of Boosting Decision Tree Model for Diagnosis of ColorectalCancer, Chibo Liu, Chunqin Pan, Jianmin Shen, Haibao Wang, and Liang Yong, Lin Wang, ChuanhaoTang, Bin Xu, Lin Yang, Lili Qu, LiangliangLi, Xiaoyan Li, Weixia Wang, Haifeng Qin, Hongjun Gao. Massspectrometry-based serum peptidome profiling accurately and reliably predictsoutcomes of pemetrexed plus platinum chemotherapy in patients with advancedlung adenocarcinomaPLOS, Published: June 8, 2017). MALDI-TOF-MSutilizes matrix-assisted laser desorption ionization of functionalized magneticbeads to provide high sensitivity mass analyses to separate the differentprotein components of a study sample (F.
Magni, Y. E. M. van der Burgt,C.
Chinello et al. Biomarkers discovery by peptide and protein profiling inbiological fluids based on functionalized magnetic beads purification and massspectrometry,” Blood Transfusion, vol. 8, no.
3, pp. s92–s97, 2010. J. F. Peterand A.
M. Otto, “Magnetic particles as powerful purification tool for highsensitive mass spectrometric screening procedures,” Proteomics, vol. 10, no. 4,pp. 628–633, 2010.
). Highlyconsistent high-throughput is assured by automated use of these magnetic beads.The AH is the secretory product of the non-pigmentedciliary epithelium of the ciliary body (Civan MM, Macknight AD. The ins and outs of aqueous humour secretion. ExpEye Res.
2004 Mar;78(3):625-31.). Its composition is a function ofthe ciliary body as well as of the other structures of the anterior segment ofthe eye consuming and producing different compounds according to theirfunctions and normal metabolism (Goel M, Picciani RG, Lee RK, Bhattacharya SK.Aqueous Humor Dynamics: A Review. The Open Ophthalmology Journal.2010;4:52-59. doi:10.2174/1874364101004010052.
). The AH as such could be viewed asthe tissue fluid of the anterior segment of the eye. Having said so, it wouldbe reasonable to hypothesise that the AH composition could be an indicator ofthe anterior segment development and functioning, and studying the compositionof this majestic fluid is likely to yield valuable information about theanterior segment of the eye, and proteins are but one component of the AH thatcan be studied for this aim. Hence, the aim of this study was to study theproteomic profile of AH in a number of PCG children using MALDI-TOF MS and SPE and compare this to a control group of asnearest as possible –at least hypothetically- to the normal AH composition. Wepresent here a preliminary report of the study findings.Materials and MethodsPatients and samplesThe study was conducted on 37 eyes of 37 childrenwith the diagnosis of PCG that presented to and were operated upon in AlexandriaMain University Hospital, Alexandria, Egypt. These PCG children were designatedas ‘cases’.
Additionally, 22 eyes of 22 patients undergoing surgery for senilecataract were included in the study and were designatedas ‘controls’. These control patients were otherwise generally normal. For allstudy patients, the sampling technique was the same. After the administrationof general anaesthesia, prepping of the skin with 10 % povidone iodine andadequate surgical draping, an initial paracentesis is made into the AC using asupersharp blade. Then, a tuberculin syringe fitted onto a 27 gauge needle is usedto puncture the cornea, tangential to the limbus, to enter the AC and withdraw0.1 ml of AH. The AC is then inflated with balanced salt solution (BSS) throughthe initial paracentesis and the planned surgical procedure continued. Thewithdrawn sample of AH was immediately emptied into an Eppendorf tube which wasclosed and sealed firmly and transferred in ice immediately to the laboratory, whereit was aliquoted into 20 uL samples and stored at -80 °C.
A pilot study on 10patient and 10 control samples was first conducted to choose betweenhydrophobic interaction and weak cation exchange chromatography magnetic beads.Hydrophobic interaction chromatography C8 beads produced more peaks andachieved better cross validation and recognition capability. Patients andcontrols were then stratified into the training and validation sets. Thetraining group was used to develop peptide models that could discriminatepatients from controls. The validation group was then used to test thepredictive power of the model derived from the training set.
Samplepreparation and mass analysis (peptide profiling)Using magnetic beads based on HydrophobicInteraction Chromatography (MB-HIC8, Bruker Daltonics Inc., Bremen, Germany) AHsamples were thawed on ice and fractionated before mass spectrometry (MS)analysis, which was conducted in the period from January to September 2017.These magnetic beads have good peptide-capturing performance and were used forprotein and peptide enrichment of AH samples according to the manufacturer’sinstructions. We added 20 ?L of the AH sample to 40 ?L of binding buffer then 5?L of MB-WCX beads in a polymerase chain reaction tube which was placed on a magnetic separator to isolate the unbound solution after careful mixing and incubation for 1 minute. The bound peptides were eluted fromthe magnetic beads after three rounds of bead separation and washing. Finally,1 ?L of the peptide eluate was mixed with 1 ?L of HCCA matrix (1.2mg of ?-Cyano-4-hydroxycinnamicacid in 50% acetonitrile with 1% trifluoroacetic acid), which was then spottedonto the sample anchor spots of a polished steel target plate (Bruker DaltonicsInc., Bremen, Germany).
The following settings were used for the MALDI-TOF MSanalyses on an Ultraflex III MALDI-TOF MS device (Bruker Daltonics Inc.): linearpositive ion mode, repetition rate of 200 Hz, ion source voltages of 25 kV and23.65 kV, lens voltage of 6.8 kV, pulsed ion extraction time of 300 ns.
Allsignals with a signal-to-noise ratio of 3 in a mass range of 1,000–20,000 Dawere recorded using FlexAnalysis software (version 3.4; Bruker Daltonics Inc.).The peptidomic patterns and models were processed using ClinPro Toolsbioinformatics software (version 3.0; Bruker Daltonics Inc.).
TheAnderson-Darling test was used for testing for specific distributions in thetraining group. Subsequently, peptide peaks, expressing the same mass-to-chargeratios (m/z) were compared between the patients and controls using Wilcoxontest for nonparametric data and t-test for parametric data. The ClinPro Tools3.0 software package (Bruker Daltonics) was used to analyze all serum sampledata derived from the training set. The peptidome MS data from the patients andcontrols in the training set were compared. Using ClinPro and to establish theprediction models, tools data from the training set were subjected to threedifferent mathematical model algorithms: the Genetic Algorithm (GA), SupervisedNeural Network (SNN), and Quick Classifier (QC).
Subsequently, each model was appliedto the validation set to test its ability to identify patients and controls. Dataprocessing and statistical analysisTheClinPro Tools software was used for the data processing and analysis accordingto the following workflow: spectra were normalized, baseline subtracted, peakssmoothed and peak areas were calculated for each spectrum. All peak signalswere processed for noise reduction using a top-hat baseline in the 800–20,000Da range. Theexpressions of the same mass-to-charge ratios (m/z) for the peptide peaks werecompared between the patients and controls. Spectra of the training set wereused for model construction. Three algorithms (genetic algorithm GA,supervised neural networks SNN, and quick classifier QC) were used toestablish the prediction models.
Each model was then applied to the validationset to test its ability to identify patients and controls.PatientcharacteristicsThestudy included 56 subjects which were equally divided (28 subjects each) intotraining and validation groups, stratified as 17 patients and 11 controls ineach group. The demographic and clinical characteristics of the study subjectsare presented in Table 1.Comparisonof peptidomic data TheClinPro Tools 3.
0 software package identified 109 peptide peaks of which 69 (63%)had significantly different intensities between the 2 groups as shown in figure1. The pseudogel view of the results is presented in figure 2.Constructionof predictive peptide models in the training setTheperformances of the three models, which were assessed by considering thecross-validation and recognition capability, are presented in table 2. The QCalgorithm yielded the best results and was used to set a prediction model, anda peptidome pattern classification was constructed. This model generated eightsignificantly different peaks at m/z 4257.
84, 4307.89, 4896.5, 4938.35, 5569.3, 7345.
89, and 9493.65Da, which provided a recognition capability of 82.1% and cross-validation value of 81.1% as presented in table 3. Seven peptides were over-expressedin PCG patients; m/z 4257.
84, 4307.89, 4896.5, 4938.35, 5569.3, 5583.
12 and 9493.65 Da while only one peak wasunder-expressed in patients; m/z 7345.89Da. This classification model correctly identified 71.
3% of PCG patients and 90.9%of controls in the training set (Table 3 and Figure 3). The peaksachieved areas under the curve of 0.824641, 0.
804611, 0.860922, 0.818216,0.
830688, 0.866213, 0.804233 and 0.812925 respectively. (Figure 4)Testingof the predictive peptide model in the validation setThepeptide prediction model thus established was tested in the validation set of28 samples in a blinded manner. The model successfully classified 80.
3 % as thepatients and 82.9% as the controls in the validation set (Table 4).Diagnosis of PCG is largelyclinical, and the molecular mechanisms underlying the responsible geneticdefects are not clearly understood so far. The concept of biological markers,many of which are proteins, has facilitated and improved the accuracy of diseasediagnosis in recent years as well as directing treatment regimens in some cases(PandeyA, Mann M. Proteomics to study genes and genomes. Nature.
2000 Jun15;405(6788):837-46.). Besides, studying the composition of AH might shed light on theexpression of genes involved in the development of the anterior segment of theeye. Indeed, relative quantification of protein expression has been made possiblethrough gel-based and gel-free techniques of protein analysis using massspectrometry and isotope labeling-based or label-free techniques. The aim ofthe current study was to characterise the protein content of AH in eyes withPCG. A preliminary report is presented herein.
The study identified 109 proteinpeaks of which 63 % were different between PCG patients and controls. Therehave been published reports about the protein content of AH in primary openangle glaucoma (Grus, F.H.
, Joachim, S.C., Sandmann, S.
, Thiel, U., Bruns, K., Lackner,K.J., Pfeiffer, N.,2008.Transthyretin and complex protein pattern in aqueoushumor of patients with primary open-angle glaucoma. Mol.
Vis. 14, 1437e1445.), uveitic glaucoma(Ladas,J.
G., Yu, F., Loo, R., Davis, J.
L., Coleman, A.L.
, Levinson, R.D., Holland,G.N., 2001.
Relationship between aqueous humor protein level and outflowfacility in patients with uveitis. Invest. Ophthalmol. Vis. Sci.
42, 2584e2588.) and myopia (Duan, X., Lu, Q.,Xue, P., Zhang, H., Dong, Z.
, Yang, F., Wang, N., 2008. Proteomic analysis ofaqueous humor from patients with myopia. Mol.
Vis. 14, 370e377.) (Kliuchnikova AA,Samokhina NI, Ilina IY, Karpov DS, Pyatnitskiy MA, Kuznetsova KG, Toropygin IY,Kochergin SA, Alekseev IB, Zgoda VG, Archakov AI, Moshkovskii SA. Human aqueoushumor proteome in cataract, glaucoma, and pseudoexfoliation syndrome.
Proteomics. 2016 Jul;16(13):1938-46. doi: 10.1002/pmic.201500423.). Bouhenni et al (Bouhenni RA, Al Shahwan S, Morales J, Wakim BT, Chomyk AM, Alkuraya FS, Edward DP.
Identification ofdifferentially expressed proteins in the aqueous humor of primary congenitalglaucoma. Exp Eye Res. 2011Jan;92(1):67-75.
doi: 10.1016/j.exer.2010.11.004. Epub 2010 Nov 13.) attempted to identifyproteins that may be altered in PCG.
The report by Bouhenni et al included 7PCG cases and 4 controls. To the best of the authors’ knowledge, the number ofcases and controls reported in the current study is the biggest so farpublished. The identification of 109 different peaks in AH samples highlightsthe diversity of the protein content of AH; and confirms the notion thatalthough the protein content of AH is quite minute under normal conditions, itis quite diverse in its constituents.
The fact that 63 % of protein peaks weredifferent between cases and controls points out how different AH protein compositioncan be in PCG cases from older controls. This difference could be attributed toeither a difference in age between cases and controls or to the presence of thepathology of PCG in the cases. In this context it is important to emphasize a verycrucial limitation in the current study and similar studies involving the samepediatric age group; the practical impossibility to sample true controls foraccurate comparison in the study. The authors draw the attention that truecontrols are children of the same age group and who are ophthalmologicallytotally free.
Obviously, for ethical reasons, such sampling is practicallyimpossible. Hence, the authors of this study (and similar studies involvingpediatric patients) are obliged to include as controls adults with senile cataract,assuming the AH composition to be the closest to the normal, though the age rangedifference remains an issue.In the currentstudy, the QC algorithm provided the best result as a prediction model. The efficacyof the QC algorithm as a prediction model is in accordance with the studies by Liet al (Li, B.
, Li, B., Guo, T., Sun, Z., Li, X.,Li, X., … Mao, Y. (2017).
Application Value of Mass Spectrometry in theDifferentiation of Benign and Malignant Liver Tumors. MedicalScience Monitor?: International Medical Journal of Experimental and ClinicalResearch, 23, 1636–1644. http://doi.org/10.12659/MSM.
901064.) and Fan et al (Fan, N.-J., Gao, C.
-F., Wang,X.-L., Zhao, G., Liu, Q.
-Y., Zhang, Y.-Y., & Cheng, B.-G. (2012). SerumPeptidome Patterns of Colorectal Cancer Based on Magnetic Bead Separation andMALDI-TOF Mass Spectrometry Analysis. Journal of Biomedicine and Biotechnology, 2012, 985020.
http://doi.org/10.1155/2012/985020.). Studying the protein peaks in the cases revealed 7 out of 8 peptides thatare over-expressed in PCG eyes. It remains to be proven whether these peptidesrepresent over-expression of genes that regulate development of the anteriorsegment of the eye or the angle of the anterior chamber and that should havebeen suppressed or down-regulated during normal development, or whether theyrepresent metabolic bi-products of structures of the anterior segment of theeye under the stressful situation of elevated IOP.
After all, PCG is theclinical result of goniodysgenesis, and exhibits an elevated IOP as an importantpathology. The significance of the under-expressed peak remains to beelucidated as well. It is hoped that identification of these peptide peaks andcharacterisation of these proteins would help in a better understanding of thedysgenetic process responsible for PCG. This knowledge, substantiated withknowledge of the genetic backgrounds of gene mutations in PCG, and perhaps witha study of the transcriptomics in PCG, might present hope for proper geneticcounselling and possible gene therapy in the near future. The limitation ofthis preliminary report is already highlighted which is the nature of thecontrols being in a different age group and demonstrating an obvious agerelated disease in the eye. The excuse –as explained- is the ethical issue ofsampling a normal eye of a child.In conclusion,AH of children with PCG is significantly different from the AH of adultpatients, with the majority of proteins over-expressed in PCG eyes.