In Vitro Effects of Three Blood Derivatives on Human Corneal Epithelial Cells

Abstract

Purpose.: Wecompared the effects of three blood derivatives, autologous serum (AS), platelet-rich plasma (PRP), and serum derived from plasma rich in growth factors (PRGF), on a human corneal epithelial (HCE) cell line to evaluate their potential as an effective treatment for corneal epithelial disorders.

Methods.: The concentrations of epidermal growth factor (EGF), fibroblast growth factor (FGF), vascular endothelial growth factor (VEGF), hepatocyte growth factor (HGF), platelet-derived growth factor (PDGF), and fibronectin were quantified by ELISA. The proliferation and viability of HCE cells were measured by an 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) colorimetric assay. Cell morphology was assessed by phase-contrast microscopy. The patterns of expression of several connexin, involucrin, and integrin α6 genes were analyzed by real-time RT-PCR.

Results.: We found significantly higher levels of EGF in PRGF compared to AS and PRP. However, AS and PRGF induced robust proliferation of HCE cells. In addition, PRGF cultured cells grew as heterogeneous colonies, exhibiting differentiated and non-differentiated cell phenotypes, whereas AS- and PRP-treated cultures exhibited quite homogeneous colonies. Finally, PRGF upregulated the expression of several genes associated with communication and cell differentiation, in comparison to AS or PRP.

Conclusions.: PRGF promotes biological processes required for corneal epithelialization, such as proliferation and differentiation. Since PRGF effects are similar to those associated with routinely used blood derivatives, the present findings warrant further research on PRGF as a novel alternative treatment for ocular surface diseases.

Introduction

The cornea is considered to be the most sensitive part of the corporeal surface and its integrity is critical for normal visual function. Many corneal surface disorders involve tear film dysfunction, the consequences of which are mainly inflammatory, but they also may include epithelial metaplasia, glandular dysfunction, and neurotrophic damage. ^1,2  

Conventional treatments for these disorders include pharmaceutical tear film substitutes, therapeutic contact lenses, anti-inflammatory agents, and oral antibiotics. However, to date, an ideal substitute for natural tear film has not been developed, essentially because natural tear is a dynamic and complex solution with specific antimicrobial, mechanical, and optical properties. In addition, natural tear contains many proteins, vitamins, and growth factors, which promote proliferation, migration, and differentiation. ³ It should not be forgotten that most artificial tears contain preservatives, which potentially make them toxic for the ocular surface. ^4,5  

To obtain an alternative tear substitute, several trials have been performed using recombinant nerve growth factor (NGF), ^6,7 epidermal growth factor (EGF), ⁸ or fibronectin ⁹ as topical treatments. However, although some preparations have yielded successful results, the stability of these preparations was low. The possibility of stimulating cell proliferation by different factors present in plasma also has been investigated. ¹⁰ These factors, acting together, could in principle lead to a more integrated response of the affected tissue. In this sense, the introduction of blood-derived products in the field of ophthalmology has been a breakthrough. The description of autologous serum (AS) ¹¹ was the first revolution, because its pH, osmolality, and biochemical properties were similar to those of natural tear. In addition, it contains essential nutrients, such as growth factors, vitamins, and bacteriostatic components, such as IgG and lysozyme. Since then, AS has been used frequently for topical therapy in patients with ocular disorders, as it is associated with enhanced healing. ^12–17  

Subsequently, platelets that are involved in tissue repair processes were identified as a major source of the growth factors present in plasma. This fact has motivated growing interest in the development of platelet-rich plasma (PRP) preparations, with the intention of increasing the concentration of these mediators and potentiating healing processes. ¹⁸ However, there presently is no consensus about the most adequate method to obtain AS and other blood derivatives. ¹⁹ Thus, different preparations have been reported by different research groups. ^{13,17,20–24} Furthermore, thrombin is used routinely to stimulate the release of growth factor content from platelets. However, this practice could lead to important adverse events, such as immune reaction and the appearance of coagulopathies. ²⁵  

More recently, a novel blood derivative has been developed that is characterized by easy and fast processing, and the ability to stimulate the release of platelet content in the absence of thrombin and leucocytes. This plasma rich in growth factors (PRGF) was introduced in maxillofacial surgery and trauma with excellent results. ^26,27 Recently, our group successfully and, to our knowledge, for the first time used serum derived from PRGF for the treatment of persistent epithelial defects ²⁸ and dry eye syndrome. ²⁹  

In this study, we compared the in vitro effect on human corneal epithelial (HCE) cells of three blood derivatives (AS, PRP, and PRGF), which are used in the treatment of corneal disorders. To this end, we performed several assays to characterize the growth factor content of the three different preparations, and to determine their effects on the growth and differentiation of a well characterized HCE cell line. 

Materials and Methods

Preparation of Blood-Derived Products

Blood from 16 healthy volunteers was collected by venipuncture (age range 30–60 years) after patient consent had been signed, in accordance with the Declaration of Helsinki. All volunteers were healthy and not taking any medication. The blood sample from each volunteer was processed according to the following three methods to obtain the corresponding blood derivatives: 

1.  
   Autologous serum (AS): Spontaneous coagulation for 2 hours at room temperature followed by centrifugation at 1,000 g for 15 minutes. Collection of the complete supernatant fraction. ¹⁷
2.  
   Platelet-rich plasma (PRP): Centrifugation at 460 g for 8 minutes and collection of the complete supernatant fraction. ²⁴
3.  
   Serum derived from plasma rich in growth factors (PRGF): Centrifugation at 460 g for 8 minutes, followed by collection of the complete supernatant fraction and induction of clot formation by adding calcium chloride at a final concentration of 22.8 mM (BTI Biotechnology Institute, S.L., Miñano, Álava, Spain), in the absence of red and white blood cells. After 2 hours at 36°C, the clot was retracted and the supernatant was collected. ^28,29

For blood collection, we use tubes with sodium citrate as anticoagulant for PRP and PRGF processing or without anticoagulant for AS processing. The complement factors of all blood derivatives were inactivated at 56°C for 30 minutes. Afterwards, we pooled samples from the different volunteers to obtain representative blood preparations that provided reproducible results and minimized inter-individual variability. These pools were stored at −20°C until its use for the in vitro assays. 

Quantification of Growth Factor Concentrations

The concentrations of epidermal growth factor (EGF), fibroblast growth factor (FGF), vascular endothelial growth factor (VEGF), hepatocyte growth factor (HGF), platelet-derived growth factor (PDGF), and fibronectin were measured from undiluted preparations using commercially available Quantikine colorimetric sandwich ELISA kits purchased from R&D (Minneapolis, MN), except for human fibronectin, which was acquired from Chemicon International Inc. (Temecula, CA). Results were expressed as mean ± SD for each age group for the three different preparations. 

Cell Culture Model

SV-40 immortalized HCE cells were provided kindly by K. Araki-Sasaki. ³⁰ Cells were cultured at 37°C under a 5% CO₂ atmosphere in Dulbecco's modified Eagle's mediu (DMEM): Ham's F12 (1:1 mix; Lonza, Verviers, Belgium) with 2 mM L-glutamine (Lonza) and 1% penicillin-streptomycin (Lonza), together with 10% fetal bovine serum (FBS; Lonza). We also added to the medium 10 ng/mL EGF (Sigma, St. Louis, MO), 5 μg/mL insulin (Sigma), 0.1 μg/mL cholera toxin (Gentaur Molecular Products, Brussels, Belgium), and 0.5% dimethylsulfoxide (DMSO; Sigma). From here onwards, we refer to EGF, insulin, cholera toxin, and DMSO as “supplements.” Upon confluence, cells were detached with 0.5% trypsin-0.2% EDTA (Sigma) and subcultivated 1:4–1:5 twice a week. 

Cell Proliferation Assays

The effect of the three blood derivatives on HCE cell proliferation was tested at several dilutions (10, 20, and 50%) and times (24, 48, and 72 hours). Proliferation was expressed as proliferation rate ± SD of viable cells with respect to viable cells just before exposure to blood derivatives (t = 0 hours). 

Cell viability was assessed using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay (Sigma-Aldrich). MTT is reduced by mitochondrial and cytosolic dehydrogenases in living cells to a purple formazan dye, which is detectable spectrophotometrically and whose absorbance is known to be directly proportional to the number of viable cells. 

HCE cells were seeded at 3000 cells/well in 96-well plates and left to attach to the plastic substrate. Subsequently, to synchronize cultures, we substituted the initial culture medium containing FBS and supplements for a medium with 1% bovine serum albumin (BSA), but without supplements, and incubation proceeded overnight. Then, at this starting time (t = 0 hours), cells were exposed to medium containing supplements, and FBS, AS, PRP, or PRGF. After 24, 48, and 72 hours, cells were washed and incubated with 0.5 mg/mL MTT for three hours. Then the MTT solution was removed and 100 μL/well DMSO was added. Optical densities were determined at 540 nm using a microplate reader (ELx800 Microplate Reader; BioTek Instruments, Winooski, VT). All experiments were performed in quadruplicate and repeated three times. 

Microscopy Assays

Phase-contrast microphotographs were taken to analyze culture morphology. Cells were incubated for 14 days with 10% FBS or with 20% of one of the three different blood preparations. Cultures were passed every three days. After washing with fresh culture medium, cell morphology was observed using a phase-contrast microscope (Nikon Eclipse TS 100; Nikon, Tokyo, Japan) and images were captured with ProgRes CapturePro 2.6 software (Jenoptik, Jena, Germany). 

Real-Time RT-PCR

To compare transcription levels of a panel of genes in the differently treated HCE cells, real-time RT-PCR was performed. Thus, HCE cells were grown for 14 days with 10% FBS or 20% AS, PRP, and PRGF, in the presence or absence of supplements. Total RNA was extracted according to the manufacturer's protocol (RNeasy minikit; Qiagen Inc., Valencia, CA), and treated with DNase for 30 minutes at 37°C and 10 minutes at 65°C (Promega, Madison, WI). Total RNA was quantified, tested on an agarose gel, and stored at −80°C until use. 

For each sample, cDNA was synthesized from 0.5 μg total RNA using the iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA). Real-time PCR assays were performed with an iCycler PCR platform (Bio-Rad). The reaction mixture contained 1 μL cDNA from the RT reaction, together with forward and reverse specific primers (250 nM each) and iQSYBR Green Supermix (Bio-Rad) in a final reaction volume of 20 μl. Thermal cycling conditions were as follows: An initial polymerase activating step at 95°C for 3 minutes, followed by 60 cycles for 20–30 seconds each at 95°C (denaturation step), 20–30 seconds at the corresponding annealing temperature for each gene (Table 1), and 20–30 seconds at 72°C (extension step), during which data were collected. Each assay included a negative control consisting of the absence of cDNA. Expression data were generated from 4 amplification reactions with samples and controls run in triplicate, and performed on different cDNA samples reverse transcribed from RNA prepared from independent culture assays. Optical data obtained by real-time PCR were analyzed using the MyiQ Single-Color Real-Time PCR Detection System Software, Version 1.0 (Bio-Rad). The dynamic range of detection for each gene was determined by preparing 5-fold serial dilutions of control HCE cells (undiluted, 1:5, 1:25, 1:125, and 1:625). The reliability of real-time PCR was defined by regression analysis of average Ct versus the log₁₀ of the target copy number. PCR efficiency was around 93% with all primer pairs. Melt Curve analysis of each PCR assay and 1.5% agarose gel electrophoresis analysis of randomly selected samples were performed to confirm the specificity of the amplification products. The expression of three different housekeeping genes (GAPDH, β-actin, and RIG/S15) also was analyzed. They were chosen after verifying their suitability by geNorm software (version 3.5). They were used to normalize expression data obtained from the studied genes, using the Bio-Rad Gene Expression Macro Software Version 1.1 derived from the algorithms outlined by Vandesompele et al. ³¹ All primers were synthesized commercially (Isogen Life Sciences, Barcelona, Spain) and a BLAST search was performed to verify their specificity for their target DNA sequences. The sequences of the primer pairs, as well as the size of the corresponding amplified fragments are detailed in Table 1. 

Table 1. 

 

Primers and Conditions Used for Real Time RT-PCR

Gene	Forward Primer (5′-3′)	Reverse Primer (5′-3′)	Amplicon Size (bp)	Annealing Temperature °C
CX26	GCTGCAAGAACGTGTGCTAC	TGGGTTTTGATCTCCTCGAT	196 bp	65°C
CX30	TGCTTAACGTGGCAGAGTTG	GGTTGGTATTGCCTTCTGGA	244 bp	60°C
CX31.1	CCTGAGTGGGGTCAACAAGT	GGGACACAGGGAAGAACTCA	191 bp	65°C
CX43	CCTTCTTGCTGATCCAGTGGTAC	ACCAAGGACACCACCAGCAT	154 bp	60°C
Integrin α6	ATGCACGCGGATCGAGTTT	TTCCTGCTTCGTATTAACATGCT	160 bp	60°C
Involucrin	TCCTCCAGTCAATACCCATCAG	GCAGTCATGTGCTTTTCCTCTTG	126 bp	60°C
GAPDH	CCTGTTCGACAGTCAGCCG	CGACCAAATCCGTTGACTCC	102 bp	56°C
β-actin	AGATGACCCAGATCATGTTTGAG	GTCACCGGAGTCCATCACG	119 bp	60°C
RIG/S15	TTCCGCAAGTTCACCTACC	CGGGCCGGCCATGCTTTACG	361 bp	60°C

Statistical Analysis

Means and SDs of all variables were calculated using Windows NT Excel software. To determinate the degree of statistical significance, we performed a paired 2-tailed Student's t-test. 

Results

Quantification of Growth Factors in the Blood-Preparations

Three different preparations (AS, PRP, and PRGF) were obtained from the blood of each volunteer (n = 16). Then, we measured by colorimetric assays the concentration of several growth factors in each blood preparation. We found that EGF levels were significantly higher in PRGF compared to the other two preparations (Table 2). Furthermore, in PRP, in which clotting had not occurred, the concentration of EGF was significantly lower than in the other two. Significant differences in FGF levels were not found among the different preparations, but the concentration of this factor was found to be age-dependent. Thus, the mean concentration of FGF in the younger volunteers (88.94 ± 9.04 pg/mL) was higher than that in the 40–49-year-old group (41.47 ± 5.89 pg/mL) and in the 50–59-year-old group (39.70 ± 3.97 pg/mL). Mean VEGF levels also were age-dependent, being highest in the group of oldest volunteers. In addition, VEGF levels were significantly lower in PRP preparation than in the other two preparations. We also observed higher levels of HGF in AS compared to PRP and PRGF, indicating that the method for obtaining AS increases the concentration of HGF. Finally, the concentrations of PDGF and fibronectin were not statistically different among the different preparations or with respect to age. 

Table 2. 

 

Concentration of Growth Factors in the Three Different Blood Derivatives

Blood Derivatives	Age (y)	EGF (pg/mL)	FGF (pg/mL)	VEGF (pg/mL)	HGF (pg/mL)	PDGF (ng/mL)	Fibronectin (μg/mL)
AS	30–39	409.52 ± 69.04	102.50 ± 69.17	127.69 ± 100.84	282.47 ± 136.57	17.06 ± 4.24	30.83 ± 6.86
	40–49	460.39 ± 114.83	36.35 ± 20.42	143.68 ± 36.92	333.13 ± 62.59	16.71 ± 2.99	30.18 ± 9.86
	50–59	398.07 ± 76.16	36.73 ± 37.21	203.35 ± 97.65	222.90 ± 22.94	18.11 ± 4.15	32.34 ± 4.45
	Mean ± SD	417.94 ± 82.62	58.55 ± 54.64	160.06 ± 89.90	275.83 ± 101.65‡	17.36 ± 3.73	31.11 ± 6.76
PRP	30–39	287.79 ± 70.88	81.87 ± 52.93	57.97 ± 43.59	93.13 ± 92.45	17.48 ± 5.44	28.65 ± 8.37
	40–49	277.79 ± 85.61	50.30 ± 35.86	47.59 ± 39.76	48.47 ± 17.71	15.80 ± 4.52	31.36 ± 2.56
	50–59	273.49 ± 93.73	36.71 ± 32.83	70.75 ± 48.20	98.90 ± 11.49	20.68 ± 8.36	31.99 ± 1.90
	Mean ± SD	279.83 ± 78.18	57.04 ± 44.44	60.31 ± 42.50†	84.60 ± 63.87	18.26 ± 6.44	30.55 ± 5.10
PRGF	30–39	480.80 ± 94.88	82.45 ± 51.61	105.06 ± 80.33	117.32 ± 118.64	16.07 ± 5.01	32.05 ± 4.60
	40–49	522.64 ± 93.75	37.76 ± 19.33	81.49 ± 42.68	37.13 ± 42.88	14.49 ± 5.00	31.47 ± 3.02
	50–59	475.09 ± 109.06	45.66 ± 29.16	169.62 ± 84.88	79.07 ± 44.11	15.80 ± 4.83	31.98 ± 3.99
	Mean ± SD	489.12 ± 95.48*	58.27 ± 41.25	124.60 ± 79.64	83.74 ± 79.42	15.58 ± 4.65	31.83 ± 3.56

 * Statistically significant differences between PRGF and PRP, and/or AS (P ≤ 0.05).

†  Statistically significant differences between PRP and PRGF, and/or AS (P ≤ 0.05).

‡  Statistically significant differences between AS and PRP, and/or PRGF (P ≤ 0.05).

Quantification data showed important differences in the concentration of growth factors among volunteers. To avoid this variability in in vitro experiments, we used pools of the preparations for subsequent culture assays. 

Cell Proliferation

We measured the effect on HCE cell proliferation of exposure for 24, 48, and 72 hours to the different blood derivatives. Two of the three preparations (AS and PRGF) produced a clear dose-response growth pattern in HCE cells (Figs. 1A, 1C). However, PRP induced a weaker response, which did not appear to be dose-dependent (Fig. 1B). We observed a tendency for cells incubated with AS and PRP to exhibit reduced viability following 24 hours of incubation with different concentrations of the three preparations (Figs. 1A, 1B). 

Figure 1. 

 

Patterns of proliferation (measured by the MTT assay) of HCE cells exposed for 24, 48, and 72 hours to increasing concentrations of the three blood derivatives. (A) AS. (B) PRP. (C) PRGF. Results are expressed as proliferation rate ± SD of viable cells with respect to viable cells at t = 0 hours. In each case, n = 3.

When comparing cell viability in response to the three different preparations with that in response to FBS (Fig. 2), we found that PRGF exhibited the highest cell viability at 24 hours (Fig. 2A). However, the three blood derivatives showed similar cell viability at doses of 10 and 20% at 48 and 72 hours of treatment (Figs. 2B, 2C). In contrast, doses of 50% AS and PRGF after 48 and 72 hours showed similar cell viability, which was higher than that of PRP. In cultures with 50% PRGF, cell viability was not significantly different from that observed when cells were cultured with FBS (Figs. 2A– 2C). 

Figure 2. 

 

Viability (MTT assay) of HCE cells exposed to different concentrations of blood derivatives. Viability was measured 24, 48, and 72 hours after treatment. Results are expressed as percentage mean ± SD with respect to viability in FBS-treated cultures (n = 3 for each case). Statistically significant differences with respect to the FBS cultures *P ≤ 0.01. †P ≤ 0.05.

Cell Morphology

Phase-contrast micrographs of HCE cells cultured with the three blood derivatives revealed substantial differences associated with the different treatments (Fig. 3). Thus, cells cultured with AS or PRP grew as quite homogeneous colonies of medium- and large-sized cells. In addition, PRP cultures showed a lower cell density, probably due to the weaker growth pattern of cells under this condition. In contrast, PRGF and FBS cultures grew in a similar manner, and both exhibited cell heterogeneity. Thus, two clearly distinguishable types of cell populations were observed, those with small cells that were grouped together, and those with larger and more flattened cells. 

Figure 3. 

 

Phase-contrast microphotographs of HCE cell monolayers after 14 days of exposure to (A) AS, (B) PRP, (C) PRGF, (D) FBS, (E) PRGF without supplements (– Supp.), and (F) SBF – Supp. AS, PRP, and PRGF, each at a dilution of 20%, and FBS at 10%. Morphologic differences among cultures are apparent. Compacted colonies of small and roundish cells (arrows), and bigger and more flattened cells (arrowheads) are indicated. In the absence of supplements, lower density colonies were seen, and big and flattened cells were predominant. Fields illustrated in these images are representative of the whole culture. Scale bar: 200 μm.

The absence of supplements from cultures was associated with a more differentiated phenotype in all cases. However, this was more evident in PRGF and FBS cultures, in which colonies of small-sized cells were observed rarely. Instead, medium- and large-sized cells coexisted in expanded colonies (Fig. 3). 

Differentiation Analysis by Gene Expression Patterns

The passage of specific growth modulation signals through gap junctions may regulate the proliferation and differentiation of human epithelial cells. For this reason, we performed quantitative RT-PCR to study the expression level of several connexins, such as connexin 26 (CX26), connexin 30 (CX30), connexin 31.1 (CX31.1), and connexin 43 (CX43 ). These gap junction proteins are transmembrane proteins, which are differentially distributed along the corneal epithelium. In addition, we analyzed the expression of integrin α6, which is expressed mainly in the basal epithelial cell layer as a component of hemidesmosomes, and involucrin, which is expressed specifically in superficial epithelial cells in the human cornea. To this end, HCE cells were cultured for 14 days with 10% FBS, 20% AS, 20% PRP, or 20% PRGF, in the presence and absence of supplements. 

Analysis of quantitative RT-PCR results revealed very distinct gene expression patterns in HCE cells cultured with PRGF in comparison to cells cultured with AS or PRP. PRGF treatment was found to induce higher expression of CX26, CX43, and involucrin, with differences being statistically significant in the case of the latter two genes (Fig. 4). In contrast, expression of CX31.1 was found to be the most variable among preparations, being significantly down-regulated in PRGF-exposed cultures in comparison to PRP and AS cultures. In addition, integrin α6 was more highly expressed in PRGF-cultured cells with respect to those cultured with AS or PRP, this difference being statistically significant for the latter case. 

Figure 4. 

 

Real-time RT-PCR of integrin α6, several connexin and involucrin gene expression in HCE cells after 14 days of exposure to the three blood derivatives in media including supplements. Results are expressed as fold gene expression ± SD versus FBS cultures (n = 4). Statistically significant differences with respect to PRGF cultures. *P ≤ 0.01. †P ≤ 0.05.

Finally, we also examined the effect of the absence or presence of supplements on the expression of these genes in HCE cells by real-time RT-PCR, since it is known that these factors are necessary to maintain the stable morphology of cultures. We observed that in the absence of supplements, the expression of involucrin and particularly CX31.1 increased in all treatments (data not shown) in comparison to the corresponding treatment in the presence of supplements. 

Discussion

We investigated the effect of three blood derivatives on HCE cell line. Two of the preparations are being used currently as therapeutic agents for several ocular surface diseases, ^17,24 while the third preparation (PRGF) recently has been introduced into the field of ophthalmology. ^28,29  

It is known that platelets are a source of a variety of growth factors with important functions in wound healing. ^26,32–34 The PRP preparation that we studied is a platelet concentrate in which platelets are neither activated nor removed. However, platelet concentrates cannot be stored for more than a few days unless frozen. When PRP preparations are frozen, platelets break and their membranes produce a cloudy aspect in PRP after thawing. This debris is thought to induce apoptotic cell death. ³⁵ As an alternative, our group has been using another blood derivative (serum derived from PRGF) in which the growth factors are released from platelets into the supernatant by their activation with calcium chloride, which is preferred over thrombin, because it enables a more sustained and physiologic release of platelet constituents, without immunologic reactions and the appearance of coagulopathies. ²⁵ In addition, the release of growth factors occurs in the absence of leukocytes, thereby avoiding the pro-inflammatory effects of proteases and acid hydrolases contained in white blood cells. ^36,37 Finally, we also used another frequently used blood derivative, AS. This preparation is obtained using extraction tubes that facilitate the subsequent coagulation process. 

We examined the in vitro effects of each preparation on the growth, morphology, and cell-to-cell communication features of immortalized HCE cells, using concentrations of blood derivatives that are used commonly in clinical practice. 

First, we determined the concentrations of several growth factors in these preparations to characterize their content. These factors can influence corneal healing and may explain the different effects of the blood derivatives. It has been demonstrated that EGF supports epithelial cell proliferation ^38,39 and, as in all tissue repair, cell proliferation is a necessary step for corneal healing. Accordingly, we observed that EGF is critical for HCE cell proliferation because removal of EGF from the medium leads to almost 50% reduction in proliferation (data not shown). We found the highest concentrations of EGF in PRGF. Thus, the presence of high levels of EGF is likely to be one of the reasons for the robust proliferative activity of HCE cells in PRGF cultures. Consistent with this idea is the finding that in PRP, in which platelet activation has not occurred, EGF levels are lower than in other preparations and there is a weaker effect on corneal epithelial cell growth. 

HGF has been shown to be implicated in the paracrine enhancement of angiogenesis by inducing VEGF expression. Thus, VEGF and HGF may be secreted or synthesized in the same local environmen. ⁴⁰ Accordingly, the significantly higher concentrations of HGF in AS could induce angiogenic effects. However, despite these levels of VEGF and HGF, we did not detect neovascularization in any of the patients treated with AS. ²⁸ This absence of neovascularization may be due to an inhibitory effect or to the existence of a balance between pro-angiogenic (VEGF, HGF, and so forth) and anti-angiogenic (thrombospodin-1, platelet factor-4, endostatin) factors in the preparations. 

The mitogenic activity of PDGF has been demonstrated recently. ^40,41 However, this activity is inhibited only modestly by an anti-PDGF antibody, ⁴⁰ suggesting that other proteins may contribute to the total cell proliferation. Consistently, our data showed similar levels of PDGF in all three blood derivatives, which nevertheless exhibited very different cell proliferation patterns. 

We detected high concentrations of fibronectin in the three preparations. During wound healing, fibronectin/fibrinogen receptors are upregulated on epithelial cells, which migrate over the bare wound. ⁴² Therefore, the high concentrations of fibronectin found in blood derivatives could favor cell migration. Furthermore, high levels of EGF receptor have been reported to be expressed in the cells migrating over the wound. ⁴³ Thus, the relatively high EGF levels in PRGF also may contribute to improved wound healing. 

AS and PRGF were found to induce robust proliferation in HCE cells. However, cultures incubated in the presence of PRGF did not exhibit the tendency to reduced viability exhibited by cells 24 hours after incubation with AS and PRP. These findings raise the possibility that PRGF may be tolerated better initially by the corneal epithelium than the other two treatments. 

MTT assays showed that the proliferative response of cells exposed to AS and PRGF (but not PRP) was dose-dependent. On the other hand, HCE cells cultured with similar preparations produced by others have been shown to exhibit substantially reduced viability after 24 hours. ²² However, we showed here that the viability of cultures exposed to these three preparations does not decrease with respect to FBS cultures, following 24–72 hours of incubation, supporting the potential safety of these derivatives in clinical applications. 

We also found that the three blood derivatives induced different morphologies in cultured cells. AS and PRP appeared to induce cell growth in quite homogeneous colonies of medium- or large-sized cells. In contrast, HCE cells cultured with PRGF exhibited a more heterogeneous morphology; they included a subpopulation of small cells organized in compact colonies. These may represent phenotypically less differentiated cells that may be a cellular source for the renewal of the epithelium. These cultures also exhibited another subpopulation of large and flattened cells, which likely represent a more differentiated phenotype. In agreement with these data, it has been reported that the differentiation of corneal cultured cells results in cell flattening and the formation of large epithelial sheets with increased intercellular communication. ⁴⁴  

Corneal epithelial cells communicate with each other through gap junctions. ^45–48 These gap junctions, which are made up of connexins, may contribute to the regulation of corneal epithelial cell functions, such as cell growth, differentiation, adhesion, and migration. ^47,49 A number of connexins are known to be expressed in HCE cells. Cx26 is expressed in the basolateral plasma membranes of basal cells. ⁴⁴ A suprabasal distribution has been found for Cx30. In contrast, Cx43 localizes to the apical surface in the basal layer and is found in all plasma membranes throughout the suprabasal layers. Finally, Cx31.1 has been localized to the apical cell surface of basal cells. ⁴⁴ However, neither CX26, 30, 31.1, nor 43 are expressed in the upper and most differentiated cells of epithelia; cells in the latter intensely express the late differentiation marker involucrin, ⁵⁰ which is a structural protein found in the cytosol. 

In this study, we detected the expression of CX26, 31.1, and 43, but not CX30, in cultured HCE cells. An increase in connexins expression in corneal epithelial cells is indicative of cell differentiation, ^43,44,51 and the formation of a well-structured and organized epithelium. We found that the expression of CX26, CX43, and even involucrin are upregulated in PRGF-treated cultures, in comparison to those treated with AS and PRP. Consistently, at a morphologic level, we observed a subpopulation of flattened and differentiated cells in PRGF-treated cultures. 

In response to corneal epithelium debridement, a dynamic modulation of gap junction expression occurs, accompanied by epithelial proliferation and differentiation. ⁵² In addition, migrating epithelial cells show reduced expression of Cx43, together with a different cellular distribution of this protein. ^43,47,53,54 Here, we found that treatment of HCE cells with any of the three blood derivatives is associated with a lower expression of CX43 in comparison to that found in FBS-treated cultures. Down-regulated CX43 expression may reflect active migration processes under these culture conditions. 

On the other hand, integrin α6 is an adhesion molecule that attaches cells to extracellular matrix proteins. ⁵⁵ It is expressed mainly in less differentiated basal epithelial cells, as a component of hemidesmosomes. It also has been demonstrated that during regeneration of corneal epithelium, the hemidesmosomes relocate, allowing basal cells to migrate towards denuded areas, and it probably is a prerequisite for the formation of new functional gap junctions. ^43,56 In this context, we detected the highest levels of integrin α6 expression in PRGF-treated cultures, in which compact colonies of small and roundish cells could be observed. Cells with this more undifferentiated aspect were detected predominantly in PRGF-treated cultures. These cells are likely to be those that underlie the proliferation capacity of HCE cultures treated with PRGF, and may be the in vitro homologs that are responsible for the renewal of the epithelium. 

Enhanced expression of involucrin and CX31.1 was detected when supplements were removed from culture media. Increased CX31.1 expression usually is associated with epithelial cell differentiation. ^57,58 Consistently, we also observed a more differentiated phenotype of HCE cells and reduced proliferation under these culture conditions (not shown). In this sense, we observed that the proliferation of cultures incubated with FBS in the absence of supplements was reduced by 70%. However, this decrease was almost unappreciated in PRGF-treated cultures, possibly due to the presence of the aforementioned undifferentiated cells, which do not appear in cultures treated with other blood derivatives. 

In summary, AS and PRGF induce dose-dependent cellular proliferation to a similar degree. However, PRGF exhibits a higher concentration of EGF, and upregulates the expression of several genes involved in communication and cell differentiation, in comparison with the other blood derivatives studied in our report. In addition, PRGF-incubated HCE cells present a heterogeneous morphology, exhibiting differentiated and nondifferentiated cell phenotypes, whereas AS- and PRP-incubated cells are more homogeneous. The clinical relevance of the present results resides in their support of the use of a novel serum derived from PRGF, which is produced in the absence of leucocytes, as a promising alternative treatment for ocular surface disorders. Although a well-characterized HCE cell line represents a very useful model to understand better and simulate in vivo situations, in vivo experiments will be necessary to confirm that our observations in culture correlate with biological outcomes. 

Footnotes

 Supported by a fellowship from the Jesus de Gangoiti Barrera foundation and Torres-Quevedo Program (VF), by Grants UE07/10 and UE09/12 from the University of the Basque Country/Instituto Clínico-Quirúrgico de Oftalmología, Grant No. S-PE10UN60 from the Saiotek Program of the Basque Government, and Grant UFI11/44 from the University of the Basque Country.

Footnotes

 Disclosure: V. Freire, None; N. Andollo, None; J. Etxebarria, None; J.A. Durán, None; M.-C. Morales, None