Cytokine and chemokine levels in tears from healthy subjects

Authors


Amalia Enríquez-de-Salamanca
IOBA (Institute of Applied Ophthalmobiology)
University of Valladolid, Campus Miguel Delibes
Valladolid E-47011
Spain
Tel: + 34 983 186369
Fax: + 34 983 184762
Email: amalia@ioba.med.uva.es

Abstract

Acta Ophthalmol. 2010: 88: e250–e258

Abstract.

Purpose:  There is growing evidence for the existence of an ‘immune tone’ in normal tears. The aim of this study was to determine the levels of a large panel of cytokines and chemokines in tears obtained from healthy subjects. These levels can then serve as baseline values for comparison with patients suffering from ocular surface diseases.

Subjects and Methods:  Nine healthy subjects participated in this study, and normal ocular surface health was documented by the results of a dry eye questionnaire, Schirmer strip wetting, and vital staining of the cornea. Four microliters of tears were collected from each eye and analysed separately with multiplex bead-based assays for the concentration of 30 cytokines and chemokines.

Results:  Twenty-five cytokines/chemokines were detected. CCL11/Eotaxin1, GM-CSF, G-CSF, IFN-γ, IL-2, IL-3, IL-4, IL-5, IL-10, IL-13, IL-12p70, IL-15, CX3CL1/Fractalkine, TNF-α, epidermal growth factor, and CCL4/MIP-1β were present at 5–100 pg/ml. IL-1β, IL-6, IL-7A, CXCL8/IL-8, and CCL2/MCP-1 were present at 100–400 pg/ml. IL-1Ra, CXCL10/IP-10 and vascular endothelial growth factor were present at more than 1000 pg/ml.

Conclusion:  Multiplex bead-based assays are convenient for cytokine/chemokine detection in tears. Fracktalkine has been detected in human healthy tears for the first time. The knowledge of cytokine/chemokine concentrations in tears from normal subjects is an important reference for further comparison with patients suffering from ocular surface diseases. Variability in their levels can reflect a phenomenon of potential importance for the understanding of the ocular surface cytokine pattern.

Introduction

The tear film, as secreted by the lachrymal functional unit, plays an important role in maintaining a normal ‘homeostatic’ environment for the epithelium of the ocular surface (Stern et al. 1998). The tears provide a tightly regulated, optimal extracellular environment that is critical to numerous functions such as antimicrobial defence, wound healing and inflammatory responses. There is growing evidence that the normal tear film contains several pro- and anti-inflammatory cytokines. Cytokines and chemokines play an integral role in the coordination and maintenance of inflammatory processes, and multiple studies have shown the presence of several pro-inflammatory agents in tears of patients with ocular surface diseases (Fujishima et al. 1995; Barton et al. 1997, 1998; Leonardi et al. 1998, 2003, 2006; Thakur & Willcox 1998; Tishler et al. 1998; Uchio et al. 2000; Cook et al. 2001; Solomon et al. 2001; Sarac et al. 2003; Eperon et al. 2004; Lema & Duran 2005; Fodor et al. 2006; Narayanan et al. 2006; Sack et al. 2007; Yoon et al. 2007). These molecules are secreted not only by cells of the immune system but also by epithelial cells in many mucosal tissues, including the ocular surface (Paolieri et al. 1997; Hingorani et al. 1998; Hershberg & Mayer 2000; Davies & Holgate 2002; Irkec & Bozkurt 2003; Meyer-Hoffert et al. 2003; Heiman et al. 2005; Rimoldi et al. 2005; Mitsias et al. 2006). Although secretion of cytokines and chemokines by epithelial cells is normally increased upon cell stimulation, basal unstimulated secretion also occurs (Gamache et al. 1997; Smit et al. 2003; Stahl et al. 2003; Siemasko et al. 2005; Enriquez-de-Salamanca et al. 2008).

The presence of cytokines and chemokines in normal tears has been described in several studies (Fujishima et al. 1995; Gupta et al. 1996; Kokawa et al. 1996; Barton et al. 1997, 1998; Malecaze et al. 1997; Vesaluoma et al. 1997, 1999; Leonardi et al. 1998, 2003, 2006; Nakamura et al. 1998; Thakur & Willcox 1998; Thakur et al. 1998; Tishler et al. 1998; Vesaluoma & Tervo 1998; Schultz & Kunert 2000; Uchio et al. 2000; Cook et al. 2001; Solomon et al. 2001; Lee et al. 2002; Sarac et al. 2003; Eperon et al. 2004; Moschos et al. 2004; Lema & Duran 2005; Sack et al. 2005, 2007; Fodor et al. 2006; Kallinikos et al. 2006; Kitaichi et al. 2006; Long et al. 2006; Narayanan et al. 2006; Sonoda et al. 2006; Uchino et al. 2006a,b; Malvitte et al. 2007; Yoon et al. 2007; LaFrance et al. 2008) in which a variety of methods were utilized such as enzyme-linked immunosorbent assay (ELISA) (Fujishima et al. 1995; Gupta et al. 1996; Kokawa et al. 1996; Barton et al. 1997, 1998; Malecaze et al. 1997; Leonardi et al. 1998, 2003; Nakamura et al. 1998; Thakur & Willcox 1998; Thakur et al. 1998; Tishler et al. 1998; Uchio et al. 2000; Lee et al. 2002; Sarac et al. 2003; Eperon et al. 2004; Moschos et al. 2004; Lema & Duran 2005; Fodor et al. 2006; Kallinikos et al. 2006; Kitaichi et al. 2006; Long et al. 2006; Narayanan et al. 2006; Yoon et al. 2007), cytometric bead arrays (CBA) (Cook et al. 2001; Sonoda et al. 2006; Uchino et al. 2006a,b), membrane-bound antibody arrays (MA) (Sack et al. 2005, 2007), double-antibody radioimmunoassay (Vesaluoma et al. 1999), PAGE and immunoblot analysis (Schultz & Kunert 2000), and multiplex bead analysis (Leonardi et al. 2006; Malvitte et al. 2007; LaFrance et al. 2008; Lam et al. 2009). In many of these studies, tears from different subjects had to be either pooled or highly diluted because of tear volume limitations, and normally only one to three cytokines could be measured. The development of flow cytometric bead-based technology has enabled the simultaneous quantisation of multiple cytokines and chemokines in single tear samples as small as 1 μl. This avoids the necessity of pooling the samples from different eyes or from the same eye at different times as typically done in other techniques such as traditional ELISA.

The purpose of this study was to use multiplex bead-based assays to measure the levels of a large panel of cytokines and chemokines in nonpooled tear samples from right and left eyes of healthy human subjects in which ocular surface disease was ruled out.

Materials and Methods

Subjects

The study population consisted of six female and three male volunteers with a mean age of 33.1  (range 25–51). All subjects were systemically healthy, were not pregnant, were not under any medication, had no previous history of ophthalmic disease (such as allergic conjunctivitis, retinal disease, cataracts…), did not have any ocular symptoms and were not contact lens users. Written consent was obtained from all subjects after explanation of the nature and possible consequences of the study. The study was approved by the Institutional Review Board of the IOBA and followed the Tenets of the Declaration of Helsinki.

Clinical tests

To assure that subjects had no ocular surface disease, the following tests were performed sequentially, plus a biomicroscopic examination to rule out any asymptomatic conjunctivitis or blepharitis. The presence of atopy was excluded by MC.

Dry eye questionnaire

The ‘symptoms of discomfort questionnaire’ (SODQ) (Gonzalez-Garcia et al. 2007) was used to exclude symptomatic subjects. The questionnaire included eight questions, each scored from 0 to 4, for a total of 32 points. Individuals with scores >10 points were excluded.

Tear production

Tear production was measured by the Schirmer’s test without anaesthesia. One sterile strip (Schirmer Tear Test Strips, 5 × 35 mm; Alcon Laboratories, Inc., Fort Worth, TX, USA) was placed in the lateral canthus of the inferior lid margin of both eyes (Halberg & Berens 1961), and the subjects were asked to maintain their eyes closed during the test. After 5 min, the length of wetting was measured in millimetres. Individuals with scores <5 mm were excluded (Anonymous 2007b).

Corneal and conjunctival vital staining

Corneal and conjunctival integrity were evaluated with fluorescein and Rose Bengal staining, respectively. Fluorescein (Fluorets, Chauvin, Aubenas, France) and Rose Bengal (Akorn, Inc., Buffalo Grove, IL, USA) strips were used. Corneal and conjunctival staining was scored in both cases following the Oxford Scheme (score 0–5) (Bron et al. 2003). Individuals with scores >1 for either test were excluded.

Tear collection

Single tear samples were obtained from each eye of the nine subjects and were maintained separately, without pooling. All tear samples were obtained approximately at the same time of the day (16:00–19:00 hr) and from right eye first and then from left eye, in every subject. The samples were obtained before clinical tests to avoid any interference by the instilled vital dyes. The tears were collected nontraumatically from the lateral canthus so as to avoid the tear reflex as much as possible. They were drawn into 4 μl capillary tubes (Drummond, Broomall, PA, USA), and then placed separately into sterile collection tubes containing 36 μl of ice-cold Beadlyte® Cytokine Assay Buffer (Upstate-Millipore, Watford, UK) (final volume 40 μl; tear sample dilution = 1/10). The tubes were kept cold during collection and stored at −80°C until assayed.

Analysis of cytokine/chemokine concentrations

Cytokine/chemokine levels in the tear samples were determined by multiplex bead analysis in a Luminex IS-100 instrument (Luminex Corporation, Austin, TX, USA). The concentration of 30 cytokines and chemokines was measured in three separate assays: CCL11/Eotaxin-1, GM-CSF, IFN-γ, IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, CXCL8/IL-8, IL-10, IL-12(p40), IL-12(p70), IL-13, IL-15, CXCL10/IP-10, CCL2/MCP-1, CCL3/MIP-1α, CCL5/RANTES and TNF-α were measured simultaneously in a 22-plex assay (Upstate-Millipore). IL-17, IL-1Ra, TGF-α, CX3CL1/Fractalkine, G-CSF, CCL4/MIP-1β and epidermal growth factor (EGF) were measured in a 7-plex assay (Linco-Millipore, Watford, UK), and vascular endothelial growth factor (VEGF) was measured singly (Upstate-Millipore).

The samples were assayed following the manufacturer’s protocols with some modifications for low volume assays. Briefly, ten microliters of 1/10 diluted tear sample were incubated with antibody-coated capture beads for 2 hr at 20°C. Washed beads were further incubated with biotin-labelled anti-human cytokine antibodies for 1 hr, followed by streptavidin–phycoerythrin incubation for 30 min. Standard curves of known concentrations of recombinant human cytokines were used to convert fluorescence units to cytokine concentration (pg/ml). The minimum detectable concentrations in pg/ml were IL-15 = 0.1; IL-5 = 0.2; IL-17 = 0.25; IL-2 = 0.3; IL-1β, IL-3 and CXCL8/IL-8 = 0.4; TGF-α 0.69; IL-10 = 0.7; IL-12p70 = 0.8; IL-13 = 0.9; TNF-α, INF-γ, and IL-6 = 1.1; CXCL10/IP-10 = 1.14; GM-CSF and IL-7 = 1.2; IL-1α and IL-4 = 1.4; CCL5/RANTES = 1.7; VEGF = 2.6; CCL2/MCP-1 = 3.0; IL-12p40 = 3.8; CX3CL1/Fractalkine = 4.27; G-CSF = 4.91; CCL11/Eotaxin = 5.1; CCL3/MIP1-α = 7.4; IL-1Ra = 10.97; CCL4/MIP-1β = 27.67. Data were stored and analysed using the Bead View Software (Upstate).

Statistics

Data were expressed as means ± standard error of the means (SEM). Statistical significance for intergroup differences was assessed by nonparametric Wilcoxon signed-rank test and the Mann–Whitney U test. The level of statistical significance was established at p < 0.05.

Results

Clinical study

Based on the clinical evaluations, including four tests plus a biomicroscopic examination, all of the subjects had values for both eyes within the accepted normal range for the dry eye questionnaire, Schirmer’s test, corneal fluorescein staining and Rose Bengal staining (Table 1). In the absence of any other test altered, a+1 corneal or conjunctival staining score, or a 5- mm score in Schirmer test values were considered compatible with normality. Thus, patients no 4 and no 7, as well as no 3, no 5 and no 8 were included in the study.

Table 1.   Clinical characteristics of the study population.
Subject noGenderAgeDry eye questionnaire scoreEyeSchirmer’s test (mm)Corneal fluorescein stainingConjunctival Rose Bengal staining
  1. M, male; F, female; RE, right eye; LE, left eye.

  2. Corneal fluorescein and conjunctival rose Bengal staining scored based upon Oxford Scale (Bron et al. 2003).

1M270RE3000
LE3000
2F519RE1900
LE1000
3F354RE1000
LE500
4M376RE3000
LE2701(3 dots)
5M274RE600
LE1500
6F292RE3000
LE3000
7F362RE301(2 dots)0
LE3000
8F252RE700
LE700
9F261RE3000
LE3000

Cytokine/chemokine concentrations in tears

Twenty-five of the 30 cytokines and chemokines analysed were detected in normal tears (Fig. 1). IL-1α, IL-12p40, IL-17, CCL3/MIP-1α and TGF-α were not detected in any sample. Of the 25 detected, 13 were present in 100% of the samples, and the remaining 12 were detected in 6–94% of the subjects (Table 2). The percentage of left and right eyes showing the presence of each cytokine and chemokine was very similar, although some discrepancies were found. For instance, CCL11/Eotaxin-1 was detected in 78% of the right eyes but in only 11% of the left eyes (Table 2). Similarly, IL-4 was detected in only 11% of the right eyes and 33% of the left eyes. For neither case was the difference significant. Table 2 also shows the percentage of detection reported by others (LaFrance et al. 2008).

Figure 1.

 Cytokine and chemokine concentrations in tears from normal healthy subjects as determined by multiplex immunobead-based assays. (A) Molecules detected in a range between 5 and 100 pg/ml; (B) molecules detected in a range between 100 and 400 pg/ml; (C) molecules with detected values higher than 2000 pg/ml. Values are mean ± SEM of both RE and LE eye samples.

Table 2.   Number of eyes with detected cytokines and chemokines.
Cytokine/ChemokineEyen/18% Previously reported [LaFrance et al. 2008]Cytokine/chemokineEyen/18% Previously reported [LaFrance et al. 2008]
  1. ND, not detected; –, not determined; RE, right eye; LE, left eye.

IL-1βRE9/961CCL2/MCP-1RE8/967
LE9/9LE8/9
Mean18/18Mean16/18
IL-2RE9/988 IFN-γRE7/991
LE9/9LE8/9
Mean18/18Mean15/18
IL-5RE9/994CCL4/MIP1βRE8/982
LE9/9LE7/9
Mean18/18Mean15/18
IL-6RE9/994IL-15RE6/9 6
LE9/9LE5/9
Mean18/18Mean11/18
IL-7RE9/994TNF-αRE4/912
LE9/9LE5/9
Mean18/18Mean9/18
CXCL8/IL-8 RE9/991 CCL11/EotaxinRE7/985
LE9/9LE1/9
Mean18/18Mean8/18
IL-10RE9/979IL-3RE2/9
LE9/9LE2/9
Mean18/18Mean4/18
IL-13RE9/982IL-4RE1/997
LE9/9LE3/9
Mean18/18Mean4/18
CXCL10/IP-10RE9/991IL-12 (p-70)RE1/973
LE9/9LE0/9
Mean18/18Mean1/18
VEGFRE9/997CCL5/RANTESRE0/985
LE9/9LE1/9
Mean18/18Mean1/18
IL-1RaRE9/994IL-1αRE0/9
LE9/9LE0/9
Mean18/18Mean0/18
CX3CL1/FractalkineRE9/9IL-12 (p-40)RE0/9
LE9/9LE0/9
Mean18/18Mean0/18
EGFRE9/9IL-17RE0/961
LE9/9LE0/9
Mean18/18Mean0/18
G-CSF RE9/988CCL3/MIP1-αRE0/988
LE8/9LE0/9
Mean17/18Mean0/18
GM-CSFRE9/930TGF-αRE0/9
LE7/9LE0/9
Mean16/18Mean0/18

The concentrations varied considerably among the cytokines and chemokines (Fig. 1, Table 3). For CCL11/Eotaxin1, GM-CSF, C-GSF, IFN-γ, IL-2, IL-3, IL-4, IL-5, IL-10, IL-13, IL-12p70, IL-15, CX3CL1/Fractalkine, TNF-α, EGF, and CCL4/MIP-1β, the levels ranged between 5 and 100 pg/ml (Fig. 1A). For IL-1β, IL-6, IL-7, CXCL8/IL-8 and CCL2/MCP-1, the levels were between 100 and 400 pg/ml (Fig. 1B). For IL-1Ra, CXCL10/IP-10 and VEGF, more than 2000 pg/ml were measured (Fig 1C). No significant differences between right and left eyes, or between males and females, were observed for any of the 25 cytokines and chemokines detected. No significant differences in cytokine/chemokine levels were observed if patients with +1 corneal or conjunctival staining score were eliminated from the analysis.

Table 3.   Concentrations of cytokines and chemokines.
Cytokine/chemokineConcentration (pg/ml)Reported concentrations* (pg/ml)Cytokine/chemokineConcentration (pg/ml)Reported concentrations* (pg/ml) 
RELEMeanRELEMean
  1. Values (mean ± SEM) correspond to the concentrations for corresponding % of detection detailed in Table 2.

  2. RE, right eye; LE, left eye; ND, not detected.

  3. * References: Thakur & Willcox (1998), Fujishima et al. (1995), Barton et al. (1997), Tishler et al. (1998), Uchio et al. (2000), Cook et al. (2001), Leonardi et al. (2006), Uchino et al. (2006 a,b), Malvitte et al. (2007), Nakamura et al. (1998), Lam et al. (2009), Sonoda et al. (2006), LaFrance et al. (2008).

CCL4/MIP1β4.8 ± 0.75.2 ± 0.55.0 ± 0.448–2 500IL-355.7 ± 4.071 ± 3.263.3 ± 3.0None
G-CSF8.4 ± 1.812.9 ± 5.010.5 ± 2.543–300CCL5/RANTESND67.5 (n = 1)67.510–580
IL-426.0 (n = 1)19.7 ± 4.321.3 ± 1.60–60IL-580.2 ± 3.379.9 ± 3.180.0 ± 2.22–45
GM-CSF33.9 ± 5.121.3 ± 5.928.4 ± 4.015–96EGF91.4 ± 13.274.3 ± 16.282.8 ± 10.4580–2 460
IL-1035.8 ± 0.938.9 ± 1.437.4 ± 0.91–65IL-1β103.3 ± 4.499.6 ± 3.6101.4 ± 2.80–209
IFN-γ48.8 ± 5.435.9 ± 4.141.9 ± 3.65–550IL-6130.3 ± 17.3130.6 ± 18.5130.4 ± 12.38–226
CX3CL1/Fractalkine44.7 ± 4.143.5 ± 3.244.1 ± 2.5NoneIL-7160.3 ± 9.8152.5 ± 8.6156.4 ± 6.4300–382
IL-1548.0 ± 3.742.5 ± 1.445.5 ± 2.114CCL2/MCP-1158.6 ± 59.5162.2 ± 42.8160.4 ± 35.410–3 500
IL-1345.9 ± 5.148.4 ± 12.647.1 ± 3.22–300CXCL8/IL-8325.0 ± 58.4320.4 ± 36.8322.7 ± 33.5148–414
IL-12 p7047.3 (n = 1)ND47.310–236VEGF3095.6 ± 473.22472.9 ± 458.42784.2 ± 328.42 608
TNF-α45.1 ± 1.249.3 ± 6.447.5 ± 3.310–525CXCL10/IP-103243.3 ± 564.62842.3 ± 606.93042.8 ± 405.023 622
IL-249.4 ± 3.652.8 ± 3.751.1 ± 2.520–1500IL-1Ra3933.8 ± 1392.13832.1 ± 1161.63882.9 ± 879.99589–29500
CCL11/Eotaxin63.9 ± 6.446.8 (n = 1)61.8 ± 4.42–294     

Discussion

One of the most important functions of the tear film is to protect the ocular surface. Thus, a healthy tear film is loaded with an ample variation of anti-infectious molecules such as IFN-γ, lactoferrin and lysozyme among others (Chandler & Gillette 1983). In the last 20 years, the presence of cytokines and chemokines in healthy tears has also been demonstrated (Nakamura et al. 1998; Thakur et al. 1998; Sack et al. 2005, 2007; Sonoda et al. 2006; Uchino et al. 2006a,b). It is well established that cytokines and chemokines are increased in ocular surface chronic inflammatory diseases such as dry eye disease, ocular allergy and others. These molecules play an integral role in the coordination and persistence of the inflammatory process. The characterization of cytokine and chemokine concentrations in normal tears is fundamental for the further comparison with tears from patients with ocular inflammatory diseases. However, because of volume limitations in tear samples in many previous studies, samples were pooled from different individuals or different eyes, and in many cases only a few cytokines could be determined.

The development of ‘multiplex analysis’ technologies has made possible the simultaneous measurement of many molecules in very small amounts of sample. With the use of this technology, sample pooling can be avoided and individual information from each eye of each patient can be obtained. Several studies have already used this technology for cytokine measurement in tear samples from patients suffering from different ocular surface diseases. From all these studies, information regarding cytokine concentration in tears from normal subjects, used as control groups, was obtained. However, in many of these studies, control subjects were only described as subjects without previous history of ocular surface disease. Here, we quantified the presence of 30 cytokines and chemokines in each eye of healthy subjects. Assessments based on the dry eye questionnaire, Schirmer’s test, and fluorescein and Rose Bengal staining of the ocular surface were used to assure that subjects were in fact normal controls, therefore ruling out any asymptomatic ocular surface pathology that could have influenced any conclusions based on the measured levels.

One potential drawback of this study was that population size was limited (n = 9). However, although it was not big, it was large enough to establish the level of cytokines/chemokines in healthy tears, and serve as the basis for further studies. It should be also taken in consideration that in this study tear samples from both eyes were analysed separately (number of samples = 18), in contrast to other studies in which samples where either pooled or only taken form one eye.

In this study, ocular surface healthiness from the subjects to be included was specially checked. This evaluation was performed as a whole, based in four different test results plus a biomicroscopic examination. Thus, although some of the subjects included had a + 1 in corneal or conjunctival staining score, or a 5- mm Schirmer value, in the absence of any other test altered these values were considered as perfectly valid for a healthy ocular surface. In fact, it has been described that small amounts of staining in cornea or conjunctiva can be found in normal subjects (Xu et al. 1995; Anonymous 2007a). It has also been reported that Schrimer value can have large variations in normal, asymptomatic eyes (Henderson & Prough 1950) that can even have a value of less than the more accepted cut-off value (≤5 mm/5 min) for Schirmer test (Cho & Yap 1993; Anonymous 2007b).

This study confirmed that normal tears from healthy subjects do contain many cytokines and chemokines in a range of concentrations with no differences between males and females. We neither found any significant differences if patients with corneal or conjunctival +1 staining score were eliminated from the analysis. Our data are consistent with molecules previously reported by others authors, determined by other techniques such as CBA (Cook et al. 2001; Sonoda et al. 2006; Uchino et al. 2006a,b; LaFrance et al. 2008) or ELISA (Fujishima et al. 1995; Barton et al. 1997; Nakamura et al. 1998; Thakur & Willcox 1998; Tishler et al. 1998; Uchio et al. 2000), or with cytokine multiplex kits from manufacturers (Leonardi et al. 2006; Malvitte et al. 2007; LaFrance et al. 2008; Lam et al. 2009) different from the ones that we used. Particularly, comparison of our values with those of previously reported studies, in which multiplex assays were performed, revealed that our levels are between those reported by Leonardi et al. and those reported by Malvitte et al., and quite similar to those reported by LaFrance et al. and Lam et al., for control population. However, there were some cytokines/chemokines, such as CCL4/MIP-1β, CCL2/MCP-1, CXCL10/IP-10 and IL-1Ra, whose levels showed a big variability (>1000 pg/ml of difference) among these studies. Particularly, the most different values from ours are those reported by Malvitte et al., with almost the highest levels for each cytokine/chemokine. In the case of CCL4/MIP-1β, Malvitte et al. reported a concentration around 2000 pg/ml, in contrast to 48.5 pg/ml reported by LaFrance et al. and 5.0 pg/ml in our study. For CCL2/MCP-1 Malvitte et al. reported a concentration around 3000 pg/ml, whereas Leonardi et al. reported around 10 pg/ml, LaFrance et al. 132.5 pg/ml and we found 160.4 pg/ml. For CXCL10/IP-10 the reported concentrations by LaFrance et al. were 23 622.3 pg/ml while we found 3042 pg/ml in this study, and in the case of IL-1Ra, LaFrance et al. reported a concentration of 9589.4 pg/ml in comparison with 3882.9 pg/ml of this study. Among other factors, differences in the tear cytokine levels in these studies, could be because of differences in the mean age of control groups: 31 ± 7 years in the study of Leonardi et al., a range of 19–59 years in the study by LaFrance et al., 45 ± 17.3 years in the study of Lam et al. and 69.4 ± 9.6 years in the study of Malvitte et al. In our study, the mean age of the subjects was 33.1 ± 8, close to that of the study by Leonardi et al. Moreover, differences in cytokine/chemokine levels among these studies could also be because of the specific cytokine/chemokine antibody clones included in the different kits used in each study. This variability observed in cytokine/chemokine tear levels could reflect as well a phenomenon of potential importance for the understanding of the ocular surface cytokine/chemokine pattern.

Unpublished correlation studies from our group have shown differences in cytokines/chemokines in tears with clinical parameters depending on which eye (right or left) was compared. In this study, tear samples were collected separately from right and left eyes, and they were not pooled so as to preserve any differences that might exist. We found no significant differences between concentration of cytokines and chemokines in tears from right and left eyes, although some of them had different percentages of detection, as in the case of CCL11/Eotaxin and IL-4. This could be because of the fact that in both cases the percentage of detection was low and the study population size was not large. Also, although tear samples were collected trying to avoid reflex tearing as much as possible, it is possible that some of it happened in left eyes, that were obtained always in second place; however, it seems this was not the case, as dilution effect was not observed for every molecule determined, and even in some cases, left eye concentration was higher than right eye. Further analyses regarding differences in cytokine/chemokine levels between both eyes are warranted, and although no significant differences were observed between eyes in this study, based on our other results, we think that sample pooling between both eyes should be always avoided when possible.

We also found that some of the cytokine and chemokines analysed were detected in <100% of the study population (Table 2) in agreement with previously reported values (LaFrance et al. 2008). In some cases, our values were different than those found by others. For instance, we found that all of the subjects had GM-CSF while La France et al. reported detection in only 30% of their subjects (LaFrance et al. 2008). Similarly, we found that some molecules, such as IL-4, IL-12 (p70) and CCL5/RANTES, were present in a low percentage of the population (<25%), and some others not detected at all, such CCL3/MIP-1α and IL-17, while La France et al. found them in the range of 61–97% of samples (LaFrance et al. 2008). These differences could be related to the minimum detectable concentrations for each particular molecule and/or with the personal characteristics of the subjects studied. Other influences that affect the percentage outcome could be the size of the study population and differences in the age or distribution between sexes of the population in the study. Further studies are warranted to settle the basis of variability in percentage of detection among healthy subjects, to determine which are the factors that could influence it. This, along with further studies of cytokine levels in a bigger population with a wider age range, will be crucial for an efficient comparison of healthy subjects and patients suffering from ocular surface disease.

To our knowledge, this is the first time that the presence of CX3CL1/Fractalkine has been described in human tears. CX3CL1/Fractalkine is an epithelial and endothelial cell-derived chemoattractant for intraepithelial lymphocytes in the mucosa of the small intestine (Muehlhoefer et al. 2000). It is the only known member of the CX3C chemokine family, and it differs greatly from other chemokines in that it can exit the cell either in a soluble form or in a membrane-bound form (Bjerkeli et al. 2007). It is expressed in various organs including the iris, ciliary body, and retina of rats and humans (Silverman et al. 2003). It interacts with the unique receptor CX3CR1 that is expressed on monocytes, natural killer cells and some T cells. Soluble CX3CL1/Fractalkine is a potent chemoattractant for CX3CR1+ leucocytes. The CX3CL1/Fractalkine-CX3-CR1-mediated mechanism may direct lymphocyte chemoattraction and adhesion within the healthy and diseased human small intestinal mucosa (Muehlhoefer et al. 2000). Numerous studies have demonstrated that CX3CL1/Fractalkine-CX3CR1 interaction contributes to the development of various inflammatory diseases such as rheumatoid arthritis, asthma, Wegener’s granulomatosis, Crohn’s disease, psoriasis, glomerulonephritis, experimental autoimmune anterior uveitis and even atherosclerosis (Matsunawa et al. 2006; Tremblay et al. 2006; Umehara et al. 2006; Bjerkeli et al. 2007; Ramos et al. 2007; Volger et al. 2007; Yoshimoto et al. 2007). Very recently, fractalkine expression in lachrymal glands of NFS7NZWF1 mice thymectomized 3 days after birth (in which Sjögren′s Syndrome spontaneously occurs) have been described (Tsubota et al. 2009); authors from this work conclude that fractalkine possibly contribute to Sjögren′s syndrome development. The role of CX3CL1/Fractalkine in normal tears is as yet unknown and deserves further analysis.

Among all of the molecules, IL-1Ra was present in the highest concentration. This is consistent with previous reports (Solomon et al. 2001) that found high levels in normal tear fluid. It is a naturally occurring cytokine receptor antagonist (Arend et al. 1998) that serves as a modulator of immune responses. IL-1Ra blocks cellular responses to IL-1 by binding with very high affinity to the IL-1 receptor without triggering signal transduction; thus, IL-1Ra regulates the agonist effects of IL-1 during chronic inflammatory and infectious diseases. Neutralizing anti-IL-1Ra antibodies inhibit endogenous IL-1Ra and demonstrate its importance as a natural anti-inflammatory protein in arthritis, colitis and granulomatous pulmonary disease (Arend et al. 1998). The relevance of IL-1 in corneal inflammation is further supported by observations that corneal cells themselves can produce this cytokine in response to various stimuli (Shams et al. 1989). In addition to its role as a modulator of corneal inflammation, IL-1Ra may also regulate corneal repair processes (Kennedy et al. 1995). Our hypothesis is that the high levels of IL-1Ra (as well as those of VEGF that was also detected in high concentration) may have a protective role for normal ocular surface.

We also detected IL-15 presence in tears from healthy subjects. This has been also reported in other studies, as those from Shoji et al. (2006) and LaFrance et al. (2008). The only of these studies that report IL-15 concentration levels is that of LaFrance (14 pg/ml), while in the other study by Shoji et al., IL-15 presence was detected by a membrane array, and only densitometric value was shown. Interestingly, in their study Shoji et al. point the fact that, among some other cytokines, IL-15 was strongly expressed in all healthy individuals; for that reason, the authors propose that these cytokines are inferred to play a role in maintaining homeostasis in normal ocular surface (Shoji et al. 2006). A protective role of IL-15 for intestinal epithelial cells has also been reported by Obermeier et al. (2006) where they describe that IL-15 has the potential to reduce mucosal damage by preventing intestinal epithelial cells apoptosis.

IL-1α, IL-12 p40, CCL3/MIP-1α, IL-17, and TGF-α were not detected in any sample. The absence of IL-17 in normal tears from healthy subjects is consistent with its role as a potent pro-inflammatory molecule that is present in several autoimmune diseases (Nashan & Schwarz 2003), although some other proinflammatory molecules, such as TNF-α or IL-6, were detected. Some other studies have described the presence of IL-1α, CCL3/MIP-1α, and IL-17 in normal tears (Nakamura et al. 1998; Malvitte et al. 2007; LaFrance et al. 2008). This could be because of the use of different detection antibodies, sampling procedures and/or differences in the populations studied.

Among others, the ocular surface cells are a source for these molecules. In fact, previous results from our group (Siemasko et al. 2005; Enriquez-de-Salamanca et al. 2008) and others (Gamache et al. 1997; Smit et al. 2003; Stahl et al. 2003) have shown that both corneal and conjunctival epithelial cells, in the absence of any stimulus, secrete many of the cytokines and chemokines determined here. Ocular surface epithelial cells may also contribute to the presence of specific patterns of cytokines and chemokines in inflammatory ocular surface diseases (Irkec & Bozkurt 2003; Lam et al. 2009). The presence of cytokines and chemokines in tears from healthy subjects without any ocular surface pathology further support the role of the conjunctival and corneal epithelia as immunomodulators of the ocular surface. Other sources of these molecules can be the main and accessory lachrymal glands, the corneal and conjunctival fibroblasts and the immunovigilant cells normally present in the ocular surface (Knop & Knop 2000).

In conclusion, this study reports concentrations of 30 cytokines (including one never described before, fractalkine) in tears from healthy subjects in which ocular surface disease was specifically ruled out. These reference values can be important for further comparison with patients suffering from ocular surface disease, surgery or are subjected to any other manipulation i.e. contact lens wear, chronic instillation of drugs, etc.

Acknowledgements

The authors thank I.Fernández (statistician) for statistical assistance. This article was presented in part as a poster at ‘World Immune Regulation Meeting –III’; special focus on ‘regulatory and effector mechanisms’ Annual Meeting, March 2009, Davos, Switzerland. The authors alone are responsible for the content and writing of the paper.

Ancillary