Proteins and microRNAs are differentially expressed in tear fluid from patients with Alzheimer’s disease

Article metrics

Abstract

Alzheimer’s disease (AD) is characterized by a progressive loss of neurons and cognitive functions. Therefore, early diagnosis of AD is critical. The development of practical and non-invasive diagnostic tests for AD remains, however, an unmet need. In the present proof-of-concept study we investigated tear fluid as a novel source of disease-specific protein and microRNA-based biomarkers for AD development using samples from patients with mild cognitive impairment (MCI) and AD. Tear protein content was evaluated via liquid chromatography-mass spectrometry and microRNA content was profiled using a genome-wide high-throughput PCR-based platform. These complementary approaches identified enrichment of specific proteins and microRNAs in tear fluid of AD patients. In particular, we identified elongation initiation factor 4E (eIF4E) as a unique protein present only in AD samples. Total microRNA abundance was found to be higher in tears from AD patients. Among individual microRNAs, microRNA-200b-5p was identified as a potential biomarker for AD with elevated levels present in AD tear fluid samples compared to controls. Our study suggests that tears may be a useful novel source of biomarkers for AD and that the identification and verification of biomarkers within tears may allow for the development of a non-invasive and cost-effective diagnostic test for AD.

Introduction

Alzheimer’s disease (AD) is an age-related neurodegenerative disorder characterised by wide-spread neuron and synapse loss resulting in phenotypic cognitive deficits and eventual onset of dementia. The histopathologic hallmarks of AD include extracellular plaques containing the amyloid-β (Aβ) peptide and neurofibrillary tangles containing the hyperphosphorylated protein tau1. The early pathogenic process of AD is clinically silent, developing over many years before a clinical diagnosis is possible2. During the early pathogenesis of AD, one of the few detectable changes is a transition from cognitively normal to a mild cognitive impairment (MCI). MCI is a high risk state for the development of AD but is a heterogeneous condition common among many neurodegenerative diseases as well as occurring in unaffected individuals with a large percentage of cases never converting to dementia making this state very susceptible to misdiagnosis3.

The uncertainty in identifying the cognitively unimpaired members of a research cohort independent from biomarker profiles in the earliest stages of AD and reliance on the onset of cognitive loss for diagnosis where molecular pathology of AD has already reached a critical stage is a major limitation in clinical trials where earlier interventional therapies for AD show more promise4. Detection of these earlier stages of AD is reliant on neuroimaging biomarkers, based on magnetic resonance imaging (MRI), functional magnetic resonance imaging (fMRI) and computed tomography (CT) scanning techniques5,6,7,8 as well as protein biomarkers from cerebrospinal fluid (CSF)9,10. Neuroimaging and CSF biomarkers both have significant challenges in their practicality, with neuroimaging being severely limited by cost and availability while obtaining CSF biomarkers requires an invasive sampling procedure. This has driven investigation into possible biomarkers present in less invasive body fluids such as blood11,12,13 or urine14.

Tears represent a non-invasive biofluid but are a largely unexplored biomarker source. There have been very limited studies on tears as a source of biomarkers for neurological disorders although they are known to contain proteins, nucleotides and other molecules15,16,17. Currently only one study has investigated tears as a source of molecular biomarkers in AD18. This lack of investigation into tear fluid is striking due to its close relationship to the eye which has been shown to have strong connections to AD pathology with numerous putative markers for AD based on the effect of AD upon the eye19. These include distinct pupillary size changes and changes in the responses to light20 as well as plaque formation in retina and lenses of AD patients which illustrates the powerful effect AD pathology has on the various components of the eye21,22. Tears have also been identified as a source of biomarkers in other neurological disorder including multiple sclerosis in which tear-based proteins were suggested as a screening test before progressing to more invasive and definitive diagnostic techniques such as CSF-based ELISA23.

In the present work we investigated possible differences in tear composition of healthy aged controls and cognitively impaired patients. This was performed by high throughput proteomic and PCR-based platforms to assess protein content and relative expression of microRNAs. MicroRNAs are a class of short, non-coding molecules which function in the post-transcriptional regulation of mRNA and protein levels in the cell24. MicroRNAs have become an area of particular interest in AD research, with critical roles in the pathogenesis of the disease reported25,26. Due to their biochemical characteristics, including relative stability in extracellular environments and their presence in extracellular particles such as exosomes27, microRNAs have attracted much attention as possible biomarkers for brain diseases including AD28. Increasing evidence suggest AD-specific concentration changes of microRNAs in various biofluids including plasma and CSF29,30. The presence of microRNAs within tear fluid has been reported in healthy humans31, however, to date has not been investigated in AD or any other brain disease.

Here, by analysing samples from patients with MCI and AD, we report disease-specific differences in tear content at the protein level as well as changes in microRNA concentrations, thereby providing the proof-of-concept that tear fluid can potentially be used as novel biofluid for the diagnosis of AD.

Results

Protein composition changes in tear fluid of AD patients

The proteomic content of tear fluid has been previously reported in healthy volunteers as well within other neurological disorders such as multiple sclerosis23. Here we explored whether AD pathology led to a distinct proteomic fingerprint in tear fluid which could be used as potential biomarker for AD diagnosis. First, we analysed if secretion of tears was altered. Tear production (flow rate) was 12 ± 9 mm/5 min in controls which was not significantly different in patients with either MCI or AD (9.25 ± 6.3 and 8,5 ± 2,9/5 min respectively, Fig. 1A). Total protein content in tears was also similar between controls and the two patient groups (Fig. 1B).

Figure 1
figure1

Protein concentration and volume of collected tear fluids. (A) Tear secretion volume measured as the length of the wet strip (in mm) (ANOVA, p = 0.531). (B) Total protein concentration in 500 μl of PBS with a cocktail of peptidase inhibitors where strips were submerged (p = 0.622). (C) Number of different proteins present per sample in tears analysed in this study after analysis by LC-MS/MS (p = 0.067). (D) Number of different proteins present in each cohort. CON, control; MCI, Mild cognitive impairment; AD, Alzheimer’s disease. The bars represent mean values with the standard error.

Next we performed liquid chromatography-mass spectrometry (LC-MS/MS) analysis to establish the possible disease-specific tear protein profile. To ensure quality of our samples, we examined the presence of well-established tear markers such as lysozyme, lipocalin 1, serotransferrin and retinal dehydrogenase 1 (ALDH1A1).

The number of proteins present in each tear sample was then analysed. No significant difference was observed between the average number of proteins present in each condition, although there was a trend toward higher numbers of unique proteins in tears from AD patients (Fig. 1C). To investigate what processes might be over-represented, we conducted a gene ontology analysis using the Database for Annotation, Visualization and Integrated Discovery (DAVID)32,33. However, comparing the three groups (controls, MCI and AD) no differences were found.

A more detailed analysis of proteins detected in all samples showed that around 40% of identified proteins were present only in one or two samples of each cohort (40% in control, 38% in MCI and 40% in AD). These proteins were eliminated from further analysis and proteins that appeared in at least 50% of samples of each cohort were retained. After filtering, the number of proteins that appeared in at least 50% of the samples of each cohort was 358 in control samples, 333 in MCI samples and 442 in AD samples (Fig. 1D). Surprisingly, peptides from tau protein or amyloid precursor protein (APP) were not detected in any of the samples. Proteins present in control, MCI and AD tears showed a similar proportion of protein distribution among different cellular components (Fig. 2A). In addition, no differences were found in levels of tear-specific enzymes between groups (Fig. 2B).

Figure 2
figure2

Functional genes and gene groups. (A) Gene Ontology (GO) analysis for categories ‘cellular component’ (http://www.uniprot.org/) for proteins present in tear samples. (B) Enzymes present in tear samples according to the six main divisions (classes). (C) Venn diagram showing which proteins are shared by each cohort. The overlapping region demonstrate that 298 proteins are present in the three tear categories. CON, control; MCI, Mild cognitive impairment; AD, Alzheimer’s disease.

We then analysed how these proteins (present in at least 50% of all samples) were distributed among the three groups. First, we observed that 298 proteins were present in control, MCI or AD samples, while 81 proteins were only found in samples from AD patients and 7 were unique to MCI (Fig. 2C). As before, to analyze what cellular processes might be overrepresented in these proteins, gene ontology analysis was conducted using DAVID. Figure 3A shows that GO term analysis by Biological process of those 298 common proteins were mainly related with retina homeostasis (Fig. 3A, p < 1 × 10−23) confirming the majority of proteins in tears are housekeeping proteins. Figure 3B shows GO term analysis by Cellular Component confirming that proteins present in tears fall mainly in the category of “Extracellular Exome”. GO analysis of the 81 proteins present in at least 50% of AD-samples did not identify any enrichment for a specific GO term with no pathways showing p values higher than 1 × 10−4 (Fig. 4). The most enriched categories identified contribute to the regulation of endopeptidase activity, protein folding, the regulation of cellular amino-acid metabolic processes and the regulation of mRNA stability (Fig. 4).

Figure 3
figure3

GO term analysis of common proteins present in tear samples. GO term analysis by Biological process (A) and by Cellular component (B) of those 298 common proteins present in the three tear groups. These proteins were mainly related with retina homeostasis and extracellular exome.

Figure 4
figure4

GO term analysis of proteins present in tear samples of AD patients. Proteins present in AD samples did not show a clear GO term pattern (Biological process), being the p values higher than 1 × 10−4. The most enriched categories contribute to regulation of endopeptidase activity, protein folding, regulation of cellular amino acid metabolic process and regulation of mRNA stability.

Identification of AD-specific proteins in tear fluid

To identify a potential tear diagnostic protein biomarker of AD, we analysed those proteins present in at least 50% of samples from each group. We identified 12 proteins present in at least five AD samples (55%) and only two control samples (13%) or less (Table 1). Interestingly, most of these proteins show an increase in MCI compared with control samples, suggesting an incremental change in disease progression. Interestingly, a GO term analysis of these 12 proteins revealed that 11 fall into the category of “metabolic processes”. However, how exactly these 12 proteins are mechanistically involved in AD remains unclear.

Table 1 Potential AD-specific tear biomarkers proteins.

The most relevant protein is eukaryotic translation initiation factor 4E (eIF4E) which appears in five AD samples (55%) but was absent in all control tear samples. Moreover, Tropomyosin alpha-4 chain and Triokinase/FMN cyclase are present in eight samples out of nine, although they are also present in some control and MCI samples (Table 1). eIF4E was selected for verification as potential biomarker for AD. Thus, we carried out Western blot with anti-eIF4E, however no signal was obtained with any of the tear samples analyzed (Supplementary Fig. 1).

Increased microRNA concentration in tears from AD patients

Analysis of microRNA content of tears was performed from the same group of patients with tear-fluid samples collected from the other eye. Quantitative analysis of RNA between 10–40 nucleotides (nt) in size revealed that tear fluid from AD patients contained significantly higher concentration of microRNA-sized molecules when compared to MCI and control (Fig. 5a,b). The extraction of microRNAs from 200 µl of diluted tear fluid produced on average 7.87 ng of RNA from AD patients, 4.78 ng in samples from patients with MCI and 3.79 ng in control (Fig. 5c). Varying volumes of tear fluid were collected from each of the volunteers ranging from 1–35 mm. The quantity of tear fluid sample showed no significant difference between conditions (p = 0.734) and no correlation with concentrations obtained after RNA extraction to mm of tear fluid obtained by Schirmer strip was observed (r2 = 0.010).

Figure 5
figure5

Tear RNA quality analysis and quantification. (a,b) RNA verification and quantification by AATI fragment analyser. RNA is separated out based on size (nucleotides (nt)) and quantified by signal intensity. (b) Quantification of the percentage of small-sized RNA extracted within the microRNA size range (10–40 nt) and the total quantity of microRNA-sized RNA molecules in each sample. (c) Comparison of microRNA concentration between conditions, control (n = 22), MCI (n = 16) and AD (n = 12) (p = 0.0189) ANOVA.

These results suggest quantitative alterations of small RNA levels occur within tear fluid of AD patients.

Profiling of microRNAs in tear fluid of AD patients

To identify distinct microRNAs present in AD, MCI and control patient tear fluid, we used a genome-wide high-throughput qPCR-based microRNA platform (OpenArray) which allows us to examine the expression of 750 microRNAs in humans34. This analysis was performed on each condition using a pool of three samples. To avoid a possible bias in our analysis, we chose samples with a consistently high concentration of small RNAs (10–40 nt) (Fig. 6a). Demonstrating the correct functioning of the OpenArray, amplification was observed in all internal positive controls (RNU48, RNU44 and U6) while the negative control ath-miR159a was undetectable.

Figure 6
figure6

OpenArray-based quantification of microRNA profile in tears of patients with MCI and AD. (a) Heatmap of microRNA profile of tear samples from control (Con), and patients with mild cognitive impairment (MCI) and Alzheimer’s disease (AD). (b) Total number of microRNAs detected by the OpenArray within each condition. (c,d) Altered levels of microRNAs in patients suffering from AD (c) and MCI (d) relative to control. (d) Venn Diagram of the total number of altered microRNAs within MCI and AD groups and the overlap of altered microRNAs between the 2 conditions.

Using a cut-off of 35 cycle threshold (CT), analysis of the OpenArray detected 282 microRNAs in control, 200 microRNAs in MCI and 266 microRNAs in AD samples (Fig. 6b). Of these, 162 microRNAs were common to all three conditions (Supplementary File 1). Among microRNAs detectable across the 3 conditions, the 5 most abundant microRNAs were microRNA-338-5p, microRNA-378, microRNA-657, microRNA-1243 and microRNA-1247b (Supplementary File 1). In total 194 individual microRNAs were identified with a difference of >1.5 or <0.6 relative to control levels between MCI (156 microRNAs) and AD (102 microRNAs). Interestingly, down-regulation of microRNAs was the main response in both MCI (105 down-regulated microRNAs vs. 6 upregulated microRNAs) and AD (58 down-regulated microRNAs vs. 32 upregulated microRNAs) compared to control (Fig. 6d). Numerous other microRNAs were exclusively detected in AD (38 microRNAs) and/or MCI (14 microRNAs) which included miR-200b-5p which was only detected in AD samples (Supplementary File 1).

MicroRNA-200b-5p in tear fluid as a potential biomarker of AD

MicroRNA-200b-5p was selected from the OpenArray profile for verification as potential biomarker for AD using individual RT-qPCRs. MicroRNA-200b-5p was chosen because it was exclusively detected in AD samples by the OpenArray (Fig. 7a). Notably microRNA-200b-5p was not detected in the only previous microRNA profile of tears which analysed young healthy control subjects, further suggesting that miR-200b-5p expression is distinct in tear fluid of AD patients31. To date, microRNA-200b-5p, however, has not been analysed as potential biomarker for AD. Due to the lack of an established control for normalization of microRNAs in tear fluids, we chose microRNA-106a as endogenous control microRNA showing a high expression as detected by the OpenArray with minimal differences between groups (Fig. 7a). Quantification of microRNA-200b-5p by individual RT-qPCR confirmed our OpenArray results with higher levels present in tear fluid of AD patients (Fig. 7b). Further, as expected, microRNA-106a showed similar levels between samples from control, MCI and AD patients demonstrating its suitability as control for normalization (Fig. 7b). Finally, when normalizing microRNA-200b-5p expression to microRNA-106, patients with AD showed a significant increase when compared to control (Fig. 7c).

Figure 7
figure7

Individual validation of disease-specific microRNA concentration changes within tear fluid. (a) Inverted Ct values (iCT = 40 − Ct) of microRNA-200b-5p and reference gene microRNA-106a detected by OpenArray. MicroRNA-200b was undetectable within control samples by the OpenArray. (b) Raw iCt values of microRNA-200b-5p and microRNA-106a by individual Taqman RT-qPCR between control (n = 8), MCI (n = 6) and AD (n = 6). (c) Relative expression of microRNA-200b-5p normalised to reference gene microRNA-106a (Mann-Whitney; *p < 0.05, **p < 0.01).

Taken together, these findings suggest the existence of distinct microRNA changes in tear fluid occurring during AD. This study establishes a strong precedent for further study of microRNAs in tear fluid within AD and other neurological diseases with increased cohorts of patients as potential source of minimally invasive biomarkers.

Discussion

The present study suggests that a unique proteomic and microRNA composition may be present in tear fluid of AD patients. Using proteomic analysis, we identified 12 proteins contained within the majority of AD samples and only present in a small minority of control samples (Table 1). Analysis of the microRNA composition in our AD group also revealed increases in total concentration of microRNAs within tear fluid as well as distinct changes in the relative expression of common microRNAs including microRNA-200b-5p. The current study establishes, therefore, a strong rationale for further investigation into biomarkers within tear fluid, in particular with the novel finding of altered microRNA content in tear fluid of AD patients.

Previous studies have shown extensive ocular alterations related to AD35,36 and that changes in cellular components present in tear-fluid may be used to diagnose AD18. Human tear proteins have been separated and identified in the past by using several analytical approaches such as SDS-PAGE and Western blot as well as mass-spectrometry techniques18. Neither tau nor β-amyloid peptides have been identified in our study which is notable for β-amyloid due to its invasive nature and widespread presence in various non-brain tissues37. Full-length β-APP has, however, been recognized by Western blot using different β-amyloid-specific antibodies in human tear fluid samples previously38. MS has been employed previously to study altered protein expression in AD using cortical brain tissue, CSF and even tear fluid18,39,40. Notably this previous investigation of tear fluid by MS was performed on excised bands that were separated out by SDS-PAGE and showed significant intensity difference between AD, thus giving a focused MS analysis on a subset of proteins in tears. This strategy is distinct from our study in which MS was performed on the total protein samples of tear fluid, allowing analysis of the total proteomic content of tear fluid in AD.

In our study, in line with previous reports18, we have not observed a significant increase in tear flow rate and tear protein concentration, although an increase in the total number of distinct proteins present in tear samples from AD patients suggest that some differences in tear fluid protein composition exist. Notably, a large number of proteins were present only in few samples and, although the source of these sporadically present proteins is unknown, we speculate that slight epithelial damage caused by filter paper-based sampling might have contributed to the contamination of samples along with disintegration of cellular contamination during the freeze-thaw cycle that takes place during transport of the samples as has been previously proposed41. The intracellular proteins detected as differentially expressed might be due to the Schirmer paper method itself because this method contains tear fluid and possibly the cells as well. Thus, in addition to the individual variability of tear proteins, this may contribute to the inconsistent protein detection in the samples.

We identified a panel of 12 proteins differentially expressed in tear fluid of AD patients when compared to controls. Among these proteins is eIF4E, which recognizes and binds the 7-methylguanosine-containing mRNA cap during the initiation of protein synthesis facilitating ribosome binding. Previous data has shown an increase of phosphorylated eIF4E in brain tissue of AD patients42. Moreover, a reduction of CYFIP1 levels, a protein that represses cap-dependent translation of mRNA by interacting with the initiation factor eIF4E, triggers a cascade of changes towards AD43. Interestingly, protein levels of translation initiation factors including eIF2a, eIF3η, eIF5, and elongation factor eIF2, are altered in the hippocampus of AD patients44. Complement activation occurs in the brain of patients with AD and contributes to the development of an inflammatory state45. In our study we have detected two related proteins; complement factor H-related protein 146 and complement component C8 alpha chain. SERPINB3 seems to be enriched in AD tear fluid samples and this protein and other serpins are found within the fibrillary amyloid plaques of brains from patients with AD47. However, exactly how these 12 proteins are mechanistically involved in AD remains unclear. Notably 4 of the 12 proteins are among the 1651 proteins previously differentially expressed in various regions of the brain by MS; Serine-threonine kinase receptor-associated protein, Complement factor H-related protein 1, Tropomyosin alpha-4 chain and Complement component C8 alpha chain48. Asparagine-tRNA ligase, cytoplasmic and Corticosteroid-binding globulin were not detected in AD and the remaining 6 were unchanged from controls48.

Although this is the first investigation of microRNAs in tear fluid of subjects with a neurological condition, microRNAs as biomarkers have been extensively investigated in blood and CSF. CSF is the optimal biofluid for the study of neurological diseases with it directly interacting with neuronal tissues. Several putative microRNA biomarkers have been identified in CSF but its utility is ultimately limited by the practicality of its invasive method of collection49. Blood, on the other hand, only requires a minimally invasive collection which has led to it being the primary focus in investigating microRNA-based biomarkers for AD with some success50,51,52. One major complication to the use of blood-based microRNAs for neurological conditions is its circulating nature with blood interacting with almost every tissue in the body and thus potentially absorbing microRNAs released from these tissues. This is an issue particularly in age-related diseases such as AD where co-morbidities are common further complicating the identification of microRNA changes unique to a neurological disorder52. Tears may present a biofluid which avoids both of these complications with non-invasive collection and relatively isolated interactions. Quantitative analysis of microRNAs in tear samples revealed significantly higher concentrations of microRNA-sized molecules present within AD patients compared to control. The reason for the increased presence of microRNAs in AD is unclear but microRNA expression is known to change rapidly in response to cellular disruptions and is likely to be affected by the numerous degenerative processes associated with AD53,54,55. Increased degeneration of the eye through choroidal thinning and retinal degeneration all occur at increased rates within AD with putative links to AD pathology56,57. The destructive processes involved in addition to increased neuro-inflammatory response in AD could potentially drive the cellular responses which increase the expression of microRNA found in AD tears56,57.

Tear fluid is a seldom investigated biofluid and currently only one previous study has analysed microRNA content in tears31. In this study, microRNAs from tear fluid samples of 5 healthy donors were pooled and analysed by miScript microRNA QC PCR Array (Qiagen). A total of 320 microRNAs were detected, compared to the 284 microRNAs within our OpenArray of control samples with 133 microRNAs common between both profiles. The large percentage of miRNAs present in both profiles may indicate a more representative population of microRNAs consistently present in tear fluid.

MicroR200b-5p was verified as a potential novel biomarker for AD found to be elevated in tear fluid of AD patients. Importantly, when quantified by individual RT-qPCR, results mirrored observed changes within the OpenArray. MicroRNA-200b-5p’s function or relationship with AD or any process related to the eye or CNS is relatively unknown with it only having been studied functionally in oncogenic tissue with putative targeting of several elements of Ras homolog gene family pathways. It was also identified as an enriched microRNA within exosomes of aged mice which could contribute to microRNA-200b-5p’s increased presence in tear fluid with the ease at which exosomes pass membrane barriers facilitating its entry into tear fluid49. Notably microRNA-200b’s more commonly expressed isomiR miR-200b-3p58, was present at high concentrations in both conditions. MiR-200b-3p has putative roles in reducing beta-amyloid toxicity59 as well as potential biomarker in serum for AD52 but showed no significant difference between conditions in the OpenArray.

Results in total microRNA concentration and microRNA-200b-5p levels showed, however, no significant differences between MCI compared to control and with only limited differences identified in their proteomic composition. This was not entirely surprising with MCI being a heterogeneous condition both occurring as a result of neurodegeneration during the prodromal stage of most neurodegenerative diseases as well as cognitive impairments due to natural aging3. Future studies should be undertaken with larger MCI patient cohorts, preferably from longitudinal studies with a confirmed conversion from MCI to AD. Nevertheless, despite the limited sample size of our profiling study, our findings have demonstrated tear fluids as a source of highly concentrated microRNA and identified that both total concentration of microRNA and individual microRNAs within tear fluid become altered in AD. Thus, the development of novel qPCR-independent microRNA-measuring devices, already in development60 may be a plausible new approach for a quick and non-invasive detection of AD development.

In conclusion, we have demonstrated that tears, collected with a minimal discomfort for the patient, are an appropriate source of microRNA and may be used along with tear proteins as biomarkers for AD. The identification of these biomarkers in tear samples collected with Schirmer strips may allow for the development of a simple, non-invasive and cost-effective test for AD. Future studies with a greater sample size should be completed to replicate and extend these findings.

Methods

Sample collection

The present study consisted of 32 donors (11 men [36.4%] and 21 women) including 9 patients with AD, 8 with MCI and 15 age-matched controls. Table 2 describes the demographics of participants. All age-matched controls (relatives of the patients) were seen by a neurologist, who did an interview and a complete physical examination ruling out any clinical neurological disease. Sample collection complied with the guidelines of the Helsinki Declaration and ethical approval was obtained from the Carlos III Institute of Health Ethics Committee (No. CEI PI 20_2014), while subjects gave informed written consent. In the case of AD patients, the consent was acquired in the presence of a caregiver, although none of them were placed under guardianship. All donors were from Soria (Spain) and the assessment of AD was carried out by a neurologist. MCI was based on Petersen criteria61 and AD was based on the NIA-AA criteria62. The diagnosis was AD with no biomarkers (imaging or biofluids) available. Besides neurological, neuropsychological and functional assessment, patients underwent an ophthalmic revision.

Table 2 Human samples used in the study.

Tear secretion was measured by using the Schirmer’s Test (Schirmer strips; Whatman, Maidstone, UK). Schirmer’s paper was kept in the lower lid margin for 5 min. Tear secretion was measured as the length of the wet strip (in mm). The right eye was used to analyse protein content while the left eye was used to analyse microRNA content. Schirmer’s tear fluid collection was carried out simultaneously in both eyes. Paper strips from the right eyes were then placed in Eppendorf tubes containing 500 μl of PBS with a cocktail of peptidase inhibitors (Roche, Basel Switzerland), strongly vortexed for 5 min and frozen in dry ice. Protein content was determined by the Bradford method (Bio-Rad, Hercules, CA, USA). The paper strips from left eyes were also placed in Eppendorf tubes containing 500 μl of water with RNase inhibitors (Diethyl pyrocarbonate (DEPC)-treated water, Ambion, CA, United States) and strongly vortexed for 5 min and frozen in dry ice. Upon arrival to the laboratory both samples were stored at −80 °C.

Protein analysis by RP-LC-MS/MS

In-Gel Digestion (Stacking gel): 100 µl of tear samples were mixed with sample buffer (4x) and boiled for 2 min. 50 µl of samples were then applied onto 1.2-cm wide wells of a conventional SDS-PAGE gel (0.75 mm-thick, 4% stacking, and 10% resolving). The run was stopped as soon as the front entered 3 mm into the resolving gel, so that the whole proteome became concentrated in the stacking/resolving gel interface. The unseparated protein bands were visualized by Coomassie staining, excised, cut into cubes (2 × 2 mm), and placed in 0.5 ml microcentrifuge tubes63. The gel pieces were destained in acetonitrile:water (ACN:H2O, 1:1), reduced and alkylated (disulfide bonds from cysteinyl residues were reduced with 10 mM DTT for 1 h at 56 °C, and thiol groups were alkylated with 50 mM iodoacetamide for 1 h at room temperature in darkness) and digested in situ with sequencing grade trypsin (Promega, Madison, WI) as described by Shevchenko et al.64 with minor modifications. The gel pieces were shrunk by removing all liquid using sufficient ACN. Acetonitrile was pipetted out and gel pieces were dried in a speedvac. The dried gel pieces were re-swollen in 50 mM ammonium bicarbonate pH 8.8 with 60 ng/µl trypsin at 5:1 protein:trypsin (w/w) ratio. The tubes were kept in ice for 2 h and incubated at 37 °C for 12 h. Digestion was stopped by the addition of 1% TFA. Whole supernatants were dried down and then desalted onto OMIX Pipette tips C18 (Agilent Technologies) before the mass spectrometric analysis was carried out.

Reverse phase-liquid chromatography RP-LC-MS/MS analysis (Dynamic Exclusion Mode): The desalted protein digest was dried, suspended in 10 µl of 0.1% formic acid and analyzed by RP-LC-MS/MS in an Easy-nLC II system coupled to an ion trap LTQ-Orbitrap-Velos-Pro hybrid mass spectrometer (Thermo Scientific). The peptides were concentrated (on-line) by reverse phase chromatography using a 0.1 mm × 20 mm C18 RP precolumn (Thermo Scientific), and then separated using a 0.075 mm × 250 mm C18 RP column (Thermo Scientific) operating at 0.3 μl/min. Peptides were eluted using a 180-min dual gradient from 5 to 25% solvent B in 135 min followed by gradient from 25 to 40% solvent B over 180 min (Solvent A: 0,1% formic acid in water, solvent B: 0,1% formic acid, 80% acetonitrile in water). ESI ionization was carried out using a Nano-bore emitters Stainless Steel ID 30 μm (Proxeon) interface. The Orbitrap resolution was set at 30.000. Peptides were detected in survey scans from 400 to 1600 amu (1 μscan), followed by twenty data dependent MS/MS scans (Top 20), using an isolation width of 2 u (in mass-to-charge ratio units), normalized collision energy of 35%, and dynamic exclusion applied during 30 seconds periods. Peptide identification from raw data was carried out using the SEQUEST algorithm (Proteome Discoverer 1.4, Thermo Scientific) and PEAKs 8.5 software. Database search was performed against Uniprot_Homo_sapiens_SwissProt. The following constraints were used for the searches: tryptic cleavage after Arg and Lys, up to two missed cleavage sites, and tolerances of 20 ppm for precursor ions and 0.8 Da for MS/MS fragment ions and the searches were performed allowing optional Met oxidation, Cys carbamidomethylation and Ser-Thr-Tyr Phosphorylation. Search against decoy database (integrated decoy approach) was carried out using false discovery rate (FDR) < 0.01. The list of identified proteins with at least two identified peptides are provided as Supplementary Data (Supplementary File 2). The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD01447565.

MicroRNA extraction

Small RNAs were isolated from patient tear fluid using a miRNeasy Serum/Plasma kit following manufacturer’s instruction. Briefly, RNA were isolated by Qiazol/chloroform phase separation, then small RNA were further isolated by serum/plasma small RNA kit (Qiagen, West Sussex, UK) and diluted in 142 μl of RNase free water. 2 µl of purified small RNA isolate was analysed on AATI Fragment analyser (AATI, Iowa USA) for quantification of small RNA concentration. Quantity of RNA was assessed on the AATI Fragment analyser considering the range of 10–40 nucleotide long molecules as miRNAs. For profiling on the OpenArray, the three samples with higher concentration of small RNA in each condition were selected and pooled by condition.

OpenArray microRNA profiling

MicroRNA profiling was performed using the OpenArray platform from Thermo Fisher, as described previously66. Briefly, the OpenArray reverse transcription reaction was performed according to the manufacturer’s protocol using 3 µl of total RNA pooled from 3 samples. Before loading onto the OpenArray, cDNA was pre-amplified following the manufacturer’s recommendation. Then, the pre-amplified product was diluted with 0.1X TE (1/40) and 22.5 µl of diluted pre-amplified product was added to the same volume of 2X Taqman OpenArray Real time PCR Master Mix (Cat No. 4462164, AB). Finally, the mix of Pre-Amp product and Master-Mix was loaded onto a 384-well OpenArray plate. OpenArray panels were automatically loaded by the OpenArray AccuFill System (Thermo Fisher Scientific, Waltham, Massachusetts) and run on a QuantStudio 12 K Flex Real-Time PCR system. Each panel enables the quantification of microRNA expression of three samples. 754 microRNAs were amplified from each sample together with 16 replicates of four internal controls (ath-miR159a (negative control), RNU48, RNU44 and U6 rRNA).

Analysis of microRNA profile

The resulting data from OpenArray profiling was processed by an initial cleaning step with data being filtered according to “AmpScore” or “CqConf” values provided by Expression Suite software (Thermo Fisher Scientific, Waltham, Massachusetts) to avoid false positives. Cut off thresholds of 35 Ct were considered to be absent. After the initial cleaning, the OpenArray qPCR profile were normalised based on Geometric Mean Normalization, one of the most reliable normalisation strategies available and was performed as previously described67 with the normalised Ct (ΔCt) calculated as ΔCt = CtmicroRNA − CtGlobal-Mean. The normalised MCI and AD patient samples were then quantified as relative expression to the mean of the normalised control. Verification of OpenArray was performed within a separate cohort of patients and quantified by Taqman small-scale microRNA assays34 on 1.69 µl of each sample. Each of the selected microRNAs were reverse transcribed by Taqman microRNA reverse transcription kit with Taqman assay primers specific and quantified with qPCR Taqman fast universal master mix qPCR in a Quant Studio 12 K Flex real-time qPCR system. QPCR was performed in triplicate with a negative control for each qPCR. This protocol was followed to quantify microRNA levels of microRNA-200b-5p (Thermofischer scientific, assay ID: 478753). All Ct values were normalised to miR-106a (Thermofischer scientific, assay ID: 478225) by ΔCt method. Relative expression (RE) was calculated as RE = 2−ΔΔCt, where ΔΔCt = ΔCttarget microRNA − ΔCtreference gene.

Statistical analysis

Comparison of groups was performed using ANOVA test and post unpaired t-test. Kurskal-wallis followed by Dunn’s post-hoc test was carried out to test for significance between groups for a specific microRNA. Significance was accepted at p < 0.05.

References

  1. 1.

    Masters, C. L. et al. Alzheimer’s disease. Nat Rev Dis Primers 1, 15056 (2015).

  2. 2.

    Price, J. L. & Morris, J. C. Tangles and plaques in nondemented aging and “preclinical” Alzheimer’s disease. Ann Neurol 45, 358–368 (1999).

  3. 3.

    Mitchell, A. J. & Shiri-Feshki, M. Rate of progression of mild cognitive impairment to dementia–meta-analysis of 41 robust inception cohort studies. Acta Psychiatr Scand 119, 252–265 (2009).

  4. 4.

    Jack, C. R. Jr. et al. NIA-AA Research Framework: Toward a biological definition of Alzheimer’s disease. Alzheimers Dement 14, 535–562 (2018).

  5. 5.

    Johnson, K. A., Fox, N. C., Sperling, R. A. & Klunk, W. E. Brain imaging in Alzheimer disease. Cold Spring Harb Perspect Med 2, a006213 (2012).

  6. 6.

    Liu, Y. et al. Regional homogeneity, functional connectivity and imaging markers of Alzheimer’s disease: a review of resting-state fMRI studies. Neuropsychologia 46, 1648–1656 (2008).

  7. 7.

    Rabinovici, G. D. et al. Amyloid vs FDG-PET in the differential diagnosis of AD and FTLD. Neurology 77, 2034–2042 (2011).

  8. 8.

    Klunk, W. E. et al. Imaging brain amyloid in Alzheimer’s disease with Pittsburgh Compound-B. Ann Neurol 55, 306–319 (2004).

  9. 9.

    Hansson, O. et al. Association between CSF biomarkers and incipient Alzheimer’s disease in patients with mild cognitive impairment: a follow-up study. Lancet Neurol 5, 228–234 (2006).

  10. 10.

    Rosen, C., Hansson, O., Blennow, K. & Zetterberg, H. Fluid biomarkers in Alzheimer’s disease - current concepts. Mol Neurodegener 8, 20 (2013).

  11. 11.

    Baird, A. L., Westwood, S. & Lovestone, S. Blood-Based Proteomic Biomarkers of Alzheimer’s Disease Pathology. Front Neurol 6, 236 (2015).

  12. 12.

    Koyama, A. et al. Plasma amyloid-beta as a predictor of dementia and cognitive decline: a systematic review and meta-analysis. Arch Neurol 69, 824–831 (2012).

  13. 13.

    O’Bryant, S. E. et al. Comparing biological markers of Alzheimer’s disease across blood fraction and platforms: Comparing apples to oranges. Alzheimers Dement (Amst) 3, 27–34 (2016).

  14. 14.

    Ghanbari, H. et al. Biochemical assay for AD7C-NTP in urine as an Alzheimer’s disease marker. J Clin Lab Anal 12, 285–288 (1998).

  15. 15.

    Dartt, D. A., Knox, I., Palau, A. & Botelho, S. Y. Proteins in fluids from individual orbital glands and in tears. Invest Ophthalmol Vis Sci 19, 1342–1347 (1980).

  16. 16.

    Ubels, J. L., Rismondo, V. & Osgood, T. B. The relationship between secretion of retinol and protein by the lacrimal gland. Invest Ophthalmol Vis Sci 30, 952–960 (1989).

  17. 17.

    Crooke, A., Guzman-Aranguez, A., Peral, A., Abdurrahman, M. K. & Pintor, J. Nucleotides in ocular secretions: their role in ocular physiology. Pharmacol Ther 119, 55–73 (2008).

  18. 18.

    Kallo, G. et al. Changes in the Chemical Barrier Composition of Tears in Alzheimer’s Disease Reveal Potential Tear Diagnostic Biomarkers. PLoS One 11, e0158000 (2016).

  19. 19.

    Lim, J. K. et al. The Eye As a Biomarker for Alzheimer’s Disease. Front Neurosci 10, 536 (2016).

  20. 20.

    Scinto, L. F. et al. A potential noninvasive neurobiological test for Alzheimer’s disease. Science 266, 1051–1054 (1994).

  21. 21.

    Kerbage, C. et al. Detection of Amyloid beta Signature in the Lens and Its Correlation in the Brain to Aid in the Diagnosis of Alzheimer’s Disease. Am J Alzheimers Dis Other Demen 30, 738–745 (2015).

  22. 22.

    Jiang, Q. et al. Long non-coding RNA-MIAT promotes neurovascular remodeling in the eye and brain. Oncotarget 7, 49688–49698 (2016).

  23. 23.

    Salvisberg, C. et al. Exploring the human tear fluid: discovery of new biomarkers in multiple sclerosis. Proteomics Clin Appl 8, 185–194 (2014).

  24. 24.

    Lewis, B. P., Burge, C. B. & Bartel, D. P. Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets. Cell 120, 15–20 (2005).

  25. 25.

    Hebert, S. S. et al. Genetic ablation of Dicer in adult forebrain neurons results in abnormal tau hyperphosphorylation and neurodegeneration. Hum Mol Genet 19, 3959–3969 (2010).

  26. 26.

    Bernstein, E. et al. Dicer is essential for mouse development. Nat Genet 35, 215–217 (2003).

  27. 27.

    Mitchell, P. S. et al. Circulating microRNAs as stable blood-based markers for cancer detection. Proc Natl Acad Sci USA 105, 10513–10518 (2008).

  28. 28.

    Quinlan, S., Kenny, A., Medina, M., Engel, T. & Jimenez-Mateos, E. M. MicroRNAs in Neurodegenerative Diseases. Int Rev Cell Mol Biol 334, 309–343 (2017).

  29. 29.

    Denk, J. et al. MicroRNA Profiling of CSF Reveals Potential Biomarkers to Detect Alzheimer’s Disease. PLoS One 10, e0126423 (2015).

  30. 30.

    Martinez, B. & Peplow, P. V. MicroRNAs as diagnostic and therapeutic tools for Alzheimer’s disease: advances and limitations. Neural Regen Res 14, 242–255 (2019).

  31. 31.

    Weber, J. A. et al. The microRNA spectrum in 12 body fluids. Clin Chem 56, 1733–1741 (2010).

  32. 32.

    Huang da, W., Sherman, B. T. & Lempicki, R. A. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4, 44–57 (2009).

  33. 33.

    Huang da, W., Sherman, B. T. & Lempicki, R. A. Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res 37, 1–13 (2009).

  34. 34.

    Kenny, A., Jimenez-Mateos, E. M., Calero, M., Medina, M. & Engel, T. Detecting Circulating MicroRNAs as Biomarkers in Alzheimer’s Disease. Methods Mol Biol 1779, 471–484 (2018).

  35. 35.

    Koronyo-Hamaoui, M. et al. Identification of amyloid plaques in retinas from Alzheimer’s patients and noninvasive in vivo optical imaging of retinal plaques in a mouse model. Neuroimage 54(Suppl 1), S204–217 (2011).

  36. 36.

    Goldstein, L. E. et al. Cytosolic beta-amyloid deposition and supranuclear cataracts in lenses from people with Alzheimer’s disease. Lancet 361, 1258–1265 (2003).

  37. 37.

    Joachim, C. L., Mori, H. & Selkoe, D. J. Amyloid beta-protein deposition in tissues other than brain in Alzheimer’s disease. Nature 341, 226–230 (1989).

  38. 38.

    Van Setten, G. B. et al. Beta-amyloid protein protein precursor expression in lacrimal glands and tear fluid. Invest Ophthalmol Vis Sci 37, 2585–2593 (1996).

  39. 39.

    Zhang J, et al. Quantitative proteomics of cerebrospinal fluid from patients with Alzheimer disease. J Alzheimers Dis 7, 125–133; discussion 173–180 (2005).

  40. 40.

    Chen, J., Wang, M. & Turko, I. V. Mass spectrometry quantification of clusterin in the human brain. Mol Neurodegener 7, 41 (2012).

  41. 41.

    van Haeringen, N. J. & Glasius, E. The origin of some enzymes in tear fluid, determined by comparative investigation with two collection methods. Exp Eye Res 22, 267–272 (1976).

  42. 42.

    Li, X. et al. Phosphorylated eukaryotic translation factor 4E is elevated in Alzheimer brain. Neuroreport 15, 2237–2240 (2004).

  43. 43.

    Tiwari, S. S. et al. Alzheimer-related decrease in CYFIP2 links amyloid production to tau hyperphosphorylation and memory loss. Brain 139, 2751–2765 (2016).

  44. 44.

    Hernandez-Ortega, K., Garcia-Esparcia, P., Gil, L., Lucas, J. J. & Ferrer, I. Altered Machinery of Protein Synthesis in Alzheimer’s: From the Nucleolus to the Ribosome. Brain Pathol 26, 593–605 (2016).

  45. 45.

    Shen, Y., Yang, L. & Li, R. What does complement do in Alzheimer’s disease? Old molecules with new insights. Transl Neurodegener 2, 21 (2013).

  46. 46.

    Hu, W. T. et al. CSF complement 3 and factor H are staging biomarkers in Alzheimer’s disease. Acta Neuropathol Commun 4, 14 (2016).

  47. 47.

    Nielsen, H. M. et al. Plasma and CSF serpins in Alzheimer disease and dementia with Lewy bodies. Neurology 69, 1569–1579 (2007).

  48. 48.

    McKetney, J. et al. Proteomic Atlas of the Human Brain in Alzheimer’s Disease. J Proteome Res 18, 1380–1391 (2019).

  49. 49.

    Grapp, M. et al. Choroid plexus transcytosis and exosome shuttling deliver folate into brain parenchyma. Nat Commun 4, 2123 (2013).

  50. 50.

    Geekiyanage, H., Jicha, G. A., Nelson, P. T. & Chan, C. Blood serum miRNA: non-invasive biomarkers for Alzheimer’s disease. Exp Neurol 235, 491–496 (2012).

  51. 51.

    Xie, B. et al. Increased Serum miR-206 Level Predicts Conversion from Amnestic Mild Cognitive Impairment to Alzheimer’s Disease: A 5-Year Follow-up Study. J Alzheimers Dis 55, 509–520 (2017).

  52. 52.

    Liu, C. G. et al. MicroRNA-135a and -200b, potential Biomarkers for Alzheimers disease, regulate beta secretase and amyloid precursor protein. Brain Res 1583, 55–64 (2014).

  53. 53.

    Moschos, S. A. et al. Expression profiling in vivo demonstrates rapid changes in lung microRNA levels following lipopolysaccharide-induced inflammation but not in the anti-inflammatory action of glucocorticoids. BMC Genomics 8, 240 (2007).

  54. 54.

    O’Connell, R. M., Taganov, K. D., Boldin, M. P., Cheng, G. & Baltimore, D. MicroRNA-155 is induced during the macrophage inflammatory response. Proc Natl Acad Sci USA 104, 1604–1609 (2007).

  55. 55.

    Nudelman, A. S. et al. Neuronal activity rapidly induces transcription of the CREB-regulated microRNA-132, in vivo. Hippocampus 20, 492–498 (2010).

  56. 56.

    Cunha, J. P. et al. OCT in Alzheimer’s disease: thinning of the RNFL and superior hemiretina. Graefes Arch Clin Exp Ophthalmol 255, 1827–1835 (2017).

  57. 57.

    Heneka, M. T. et al. Neuroinflammation in Alzheimer’s disease. Lancet Neurol 14, 388–405 (2015).

  58. 58.

    Rhodes, L. V. et al. Dual regulation by microRNA-200b-3p and microRNA-200b-5p in the inhibition of epithelial-to-mesenchymal transition in triple-negative breast cancer. Oncotarget 6, 16638–16652 (2015).

  59. 59.

    Higaki, S. et al. Defensive effect of microRNA-200b/c against amyloid-beta peptide-induced toxicity in Alzheimer’s disease models. PLoS One 13, e0196929 (2018).

  60. 60.

    McArdle, H. et al. “TORNADO” - Theranostic One-Step RNA Detector; microfluidic disc for the direct detection of microRNA-134 in plasma and cerebrospinal fluid. Sci Rep 7, 1750 (2017).

  61. 61.

    Petersen, R. C. Mild cognitive impairment as a diagnostic entity. J Intern Med 256, 183–194 (2004).

  62. 62.

    McKhann, G. M. et al. The diagnosis of dementia due to Alzheimer’s disease: recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 263–269 (2011).

  63. 63.

    Moreno, M. L. et al. Disulfide stress: a novel type of oxidative stress in acute pancreatitis. Free Radic Biol Med 70, 265–277 (2014).

  64. 64.

    Shevchenko, A., Wilm, M., Vorm, O. & Mann, M. Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal Chem 68, 850–858 (1996).

  65. 65.

    Perez-Riverol, Y. et al. The PRIDE database and related tools and resources in 2019: improving support for quantification data. Nucleic Acids Res 47, D442–D450 (2019).

  66. 66.

    Jimenez-Mateos, E. M. et al. microRNA targeting of the P2X7 purinoceptor opposes a contralateral epileptogenic focus in the hippocampus. Sci Rep 5, 17486 (2015).

  67. 67.

    Raoof, R. et al. Cerebrospinal fluid microRNAs are potential biomarkers of temporal lobe epilepsy and status epilepticus. Sci Rep 7, 3328 (2017).

Download references

Acknowledgements

This study was funded by grants from Spanish Ministry of Economy and Competitiveness (SAF 2006-02424, BFU-2008–03980, BFU2016-77885-P), Comunidad de Madrid (S2017/BMD-3700), Centro de Investigación Biomédica en Red sobre Enfermedades Neurodegenerativas (CIBERNED, ISCIII), institutional grant from the Fundación R. Areces and Science Foundation Ireland (CDA/4708; 12/COEN/18 and co-funded under the European Regional Development Fund and by FutureNeuro industry partners 16/RC/3948). Analysis using LC-MS/MS and database search was carried out in the ‘CBMSO Protein Chemistry Facility’, belonging to ProteoRed, PRB2-ISCIII and which was supported by grant PT13/0001. Authors would like to thank AFA León, AFA Soria and the people who have collaborated as subjects in the study.

Author information

A.K. carried out experiments with microRNAs, E.M.J.-M. run the OpenArray, M.A.-Z. and A.R. collected human tears, P. G.-M. carried out eye examination, J.A., T.E. and F.H. designed the experiment and J.H.M.P., D.C.H, J.A., T.E. and F.H. wrote the paper.

Correspondence to Tobias Engel or Félix Hernández.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Kenny, A., Jiménez-Mateos, E.M., Zea-Sevilla, M. et al. Proteins and microRNAs are differentially expressed in tear fluid from patients with Alzheimer’s disease. Sci Rep 9, 15437 (2019) doi:10.1038/s41598-019-51837-y

Download citation

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.