We investigated the effects of various culture conditions on the growth of normal human corneal endothelial cells in culture. Falcon Primaria tissue culture plastic ...
Investigative Ophthalmology & Visual Science, Vol. 30, No. 2, February 1989 Copyright © Association for Research in Vision and Ophthalmology
Growth of Human Corneal Endothelial Cells in Culture Beatrice Y. J. T. Yue, Joel Sugar, Jean E. Gilboy, and Judith L. Elvart
We investigated the effects of various culture conditions on the growth of normal human corneal endothelial cells in culture. Falcon Primaria tissue culture plastic was found to provide a more suitable surface for endothelial cell growth than the conventional Corning tissue culture plastic. Also, media containing 10% fetal bovine serum and 5% calf serum (complete media) facilitated the growth of human cells better than those containing Nu-serutn. Supplementation with epidermal or fibroblast growth factor (10 and 100 ng/ml) to the complete media had no effect on human endothelial cell growth. Chondroitin sulfate at low concentrations (100 jtg/ml to 1 mg/ml) also showed little effect. At high concentrations (13.5 and 25 mg/ml), however, chondroitin sulfate significantly promoted human corneal endothelial cell growth during a 1- to 2-week incubation period. From the 37 cultures initiated, outgrowth from explants appeared within 3 to 7 days. Cells were polygonal in shape and, at confluency, formed a continuous monolayer. We attained a success rate of 87% (7/8) growing primary cultures from donors under 20 years of age and a 59% (17/29) success rate from older donors. Invest Ophthalmol Vis Sci 30:248-253,1989 complete media3 with EGF, fibroblast growth factor (FGF), or endothelial cell growth factor (ECGF). These growth factors are mitogenic toward either corneal6"8 or vascular endothelial9 cells. In addition, the effects of exogenous chondroitin sulfate or hyaluronic acid were assessed.
Growth of pure human endothelial cells in tissue culture has been achieved in several laboratories1"5 from the endothelium of either the whole cornea3 or corneoscleral rims.5 Long-term serial cultivation, however, has generally been limited only to endothelia derived from donors 20 years and younger.3-5 Tissue culturing of cells from older human donors still remains difficult. In an attempt to improve human corneal endothelial cell growth, especially from older donors, we have tested a number of tissue culture conditions for their mitogenic effects. Experiments included culturing cells on the recently developed Falcon Primaria flasks. The more positively charged surface of these flasks simulates that of basement membranes. In another set of experiments, we replaced serum with commercially prepared Nu-serum, which contains epidermal growth factor (EGF), endothelial cell growth supplement, insulin, transferrin, triiodothyronine, progesterone, estradiol, testosterone, hydrocortisone, selenium, phosphoethanolamine, glucose, amino acids, vitamins as well as 25% vol/vol fetal bovine serum. A third set of experiments involved supplementing
Materials and Methods Normal human eyes were obtained from the Illinois Eye Bank. Corneoscleral rims from human donors were received following use of the central cornea for corneal transplantation. A total of 19 human eyes and 18 corneoscleral rims were obtained for the study. Donor age ranged from 2 months to 75 years. For the first group of tissues, the interval between death and enucleation varied from 1 to 9 hr (3.4 ± 2.3 hr), and the interval between enucleation and culture ing varied from 2 to 69 hr (18.2 ± 14.3 hr). For the second group, the corneoscleral rims, the interval between death and enucleation varied from 1 to 15 hr (4.4 ± 3.8 hr). Corneas were stored in K-Sol®10 (CILCO, Cooper Vision, Bellevue, WA) prior to keratoplasty for a period varying from 4 hr to 7 days (89.1 ± 48.0 hr). Corneal endothelial cultures were initiated 3 to 26 hr (6.2 ± 4.8 hr) from corneoscleral rims following excision of the central corneas.
From the Department of Ophthalmology, Lions of Illinois Eye Research Institute, University of Illinois at Chicago, Chicago, Illinois. Supported by research grants EY-O389O, EY-05628, and core grant EY-01792 from National Eye Institute, Bethesda, Maryland, and an unrestricted grant from Research to Prevent Blindness, Inc., New York, New York. Submitted for publication: December 7, 1987; accepted August 8, 1988. Reprint requests: Beatrice Y. J. T. Yue, Department of Ophthalmology, Lions of Illinois Eye Research Institute, University of Illinois at Chicago, 1905 West Taylor Street, Chicago, IL 60612.
Tissue Culture Corneal tissues used for cultures were obtained at a distance 2 mm or greater from the scleral rims to ensure that the corneas were not contaminated by any neighboring tissues. Endothelium-Descemet's membrane explants were stripped from the corneas
HUMAN CORNEAL ENDOTHELIAL CELL CULTURE / Yue er ol
and cultured in Falcon Primaria T-25 flasks (Falcon Laboratories, Becton-Dickinson & Co., Oxnard, CA) as previously described1' at 37°C with 5% CO2-95% air in complete media. All tissue culture media were obtained from Hazleton Dutchland Inc. (Lenexa, KS). The complete media contained modified Eagle's minimum essential medium with Earle's salt (MEM, cat. No. 51-41277), supplemented with 10% fetal bovine serum (cat. No. 12-10378), 5% calf serum (cat. No. 12-13377), 2% 200 mM glutamine (cat. No. #59-20176), 2% essential (cat. No. 58-25277, 50X) and 1% nonessential (cat. No. 58-57277, 100X) amino acids. The antibiotics used were 10 Mg/ml of gentamicin (Elkins-Sinn, Inc., Cherry Hill, NJ) and 1.2 Mg/ml of amphotericin B (E.R. Squibb & Sons, Inc., Princeton, NJ). Generally 3-5 pieces (each approximately 3 X 1 mm in size) from a single Descemet's membrane-endothelium complex were included in one T-25 flask. When cells grew out from the explants and had covered 10-20% of the flask surface area, the cells were dispersed by trypsinization. The dispersed cells were allowed to resettle and spread in the same flask. For trypsinization, cells in flasks were washed 2 times with Versene (Flow Laboratories, McLean, VA, cat. No. 28-203-49, 0.02% ethylenediaminetetraacetate disodium salt), and incubated with 1.5 ml of 0.25% trypsin solution (Hazleton Dutchland Inc., cat. No. 59-22777, 2.5% trypsin in modified Hank's balanced salt solution, without calcium and magnesium) at room temperature. After 45 to 60 seconds, the solution was poured off, leaving a residual amount of trypsin in the flask. Cells were then monitored closely with a phase contrast microscope. When they became rounded and refractile (approximately 5-7 min), 2-3 ml of complete media were added to the flasks and the cells were gently dislodged using a Pasteur pipette. The cells resettled and grew to cover the entire bottom of the flask. Cells were then subcultured with a 1:2 splitting ratio using the trypsinization procedure described above except that the incubation time with trypsin solution was between 1 and 2 min. Cell Growth Studies Cell growth experiments were conducted with cells between the first and third passages as follows: when primary or subcultures grown on Primaria T-25 flasks had reached confluency, the cells were trypsinized and suspended in complete media. Equal aliquots of single cell suspensions were inoculated onto each well of either Corning 24-well plates (Corning Scientific Products, Corning, NY) or Falcon Primaria 24-well plates (Falcon Laboratories, Becton-Dickinson & Co., Oxnard, CA). Fifteen thousand cells were
20.000 S3 3
Days Fig. 1. Effects of Falcon Primaria and Corning plates on human endothelial cell growth. Second-passage human corneal endothelial cells from a 2-month-old donor were used for the experiment. Cells were plated on either Falcon Primaria (•) or Corning (O) 24-well plates at 10,000 cells per well. Cell growth was followed by resuspending cells from quadruplicate wells and counting the cell number at various intervals. Values of mean ± SD (n = 4) are presented at each time point. Another experiment was performed with cells derived from a 15-month-old donor and a similar pattern was obtained.
added to each well, unless otherwise specified. Twenty-four hours later, when the cells were settled, fresh complete media with and without the substance under evaluation were added. The effects of each substance were assessed by trypsinizing the cells from triplicate or quadruplicate wells at various time intervals and counting the cell number with a ZBI Coulter counter. Each substance was evaluated with cells obtained from two to three different donors. Nu-serum, EGF, FGF, and ECGF were obtained from Collaborative Research, Inc. (Waltham, MA). Chondroitin sulfate (Sigma Co., St. Louis, MO, cat. No. C-4134, type A, from whale cartilage), heparin (Sigma Co., cat. No. H-5640, from porcine intestinal mucosa), hyaluronic acid (ICN Biomedicals, Naperville, IL, cat. No. 100735, from human umbilical cord), and the growth factors were reconstituted with MEM and filtered through 0.22 ^m Millipore filters
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / February 1989
Table 1. Effects of growth factors on rabbit and human corneal endothelial cell growth Rabbit cultures Control
0.83 ± 0.08f
0.93 ± 0.03f
The mean value of cell numbers from the control wells was normalized to 1.00. Experimental data were then expressed as mean ± standard deviation (n = 4) relative to the control. The control value (cell number) for rabbit cultures was 82,571, and that for human cultures was 16,384. * P < 0.0001 compared to controls. f Not significant compared to controls. % P < 0.03 compared to controls.
cells were used. The rabbit cells were obtained as previously described.11J2 Results Effects of Various Factors on Endothelial Cell Growth
Days Fig. 2. Effects of complete media and media containing 15% Nu-serum on human endothelial cell growth. Second-passage human corneal endothelial cells from a 2-month-old donor were used for the experiment. Cells were plated on Falcon Primaria plates at 12,000 cells per well and fed with either complete media (•) or media containing 15% Nu-serum (O). Cell growth was followed by resuspending cells from quadruplicate wells and counting the cell number at various intervals. Values of mean ± SD (n = 4) are presented at each time point. Another experiment was performed with cells derived from a 15-month-old donor, and a similar pattern was obtained.
for sterilization. Aliquots of the solutions were added to the complete media to attain the desired final concentrations. The pH of the medium remained unchanged by addition of any of the test substances. The osmolarity for the complete media, complete media plus 13.5 mg/ml chondroitin sulfate, and complete media plus 25 mg/ml chondroitin sulfate was 302, 338 and 371 mOsm, respectively. In some experiments, rabbit corneal endothelial
Figure 1 shows the growth curve of normal human corneal endothelial cells on Corning or Falcon Primaria plates. An equal number of endothelial cells was inoculated on day 0. The next day approximately 80-90% of cells were settled on both surfaces. The Falcon Primaria plates ultimately facilitated endothelial cell growth better than did the Corning plates. When 15% Nu-serum instead of 10% fetal bovine serum and 5% calf serum was used in the complete media, corneal endothelial cell growth was reduced
Table 2. Effects of chondroitin sulfate on human corneal endothelial cell growth
Control Chondroitin sulfate (13.5 mg/ml) Chondroitin sulfate (25 mg/ml) Chondroitin sulfate (1 mg/ml) Hyaluronic acid (1 mg/ml)
1.00 + 0.12
1.54 + 0.15t
1.84 ± 0 . 1 2 *
0.97 ± 0.07* 1.14 + 0.15*
The mean value of cell numbers from the control wells was normalized to 1.00. Experimental data were then expressed as mean ± standard deviation (n = 4) relative to the control. The control value (cell number) for 1-week experiments was 36,770, and that for 2-week experiments was 41,661. * Not significant compared to controls. t P < 0.003 compared to controls. %P< 0.0002 compared to controls.
HUMAN CORNEAL ENDOTHEUAL CELL CULTURE / Yue er ol
(Fig. 2). EGF and FGF, which stimulated third-passage rabbit corneal endothelial cell growth at a high concentration (100 ng/ml), did not have a significant growth-promoting effect on first-passage human corneal endothelial cells derived from a 4-month-old donor (Table 1). Experiments were repeated with human cells derived from a 15-month-old and a 17year-old donor, yielding similar results. ECGF (100 ng/ml), with or without heparin (90 Mg/ml), inhibited cell growth (data not shown). Supplemental chondroitin sulfate and hyaluronic acid at 1 mg/ml exhibited little effect (Table 2) on human corneal cell growth (first-passage human cells derived from a 15-month-old donor). However, chondroitin sulfate at higher concentrations (13.5 and 25 mg/ml) enhanced endothelial cell growth (Table 2) during 1-week and 2-week incubations periods. These results were confirmed by two other experiments using cells derived from a 4-month-old and a 17-year-old donor. Morphology and Growth of Human Corneal Endothelial Cells Outgrowths from explants on Primaria plastic were seen between 3 and 10 days. The cells displayed a polygonal endothelial morphology under a phasecontrast microscope (Fig. 3A). Generally, cells immediately adjacent to the explants were compact and small in size. As they proliferated and migrated away from the explants, the cells retained the same polygonal appearance but became larger in size (Fig. 3A, arrow). Cell size remained essentially unchanged from these larger cells throughout subsequent culturing periods. Cultures that did not grow well, however, contained further enlarged and multinucleated cells that had many vacuoles. At confluency, cells formed a continuous monolayer (Fig. 3B). None of the growth factor and chondroitin sulfate supplementations changed the morphology of human endothelial cells. We used Falcon Primaria flasks and complete media for establishment of human corneal endothelial cell cultures. Among the 19 primary cultures that were initiated from whole-eye donors, 16 were successful with cells grown to 50% or more confluency (Table 3). The remaining 3 cultures were judged to be in the 10-49% confluency category (Table 3), having potential for further growth. Among the 18 cultures that were derived from corneoscleral rims, eight were successful. Nine cultures had outgrowth initially (less than 10% confluency) but failed to progress, and the cells eventually died (Table 4). The overall success rate for primary cultures therefore was 87.5% (7/8)
Fig. 3. Phase contrast micrographs of human corneal endothelial cells (XI35). (A) Outgrowth from an explant (E) after 7 days in culture. Cells were derived from a 15-year-old donor. Arrowheads indicate cells at the lead edge that appear larger than the cells immediately adjacent to the explant. (B) Confluent culture of firstpassage cells derived from a 15-year-old donor.
for donors under 20 years of age and 59% (17/29) for donors between 21 and 75 years of age. One of the seven cultures from the former group was maintained up to the seventh passage. Another one was up to the fifth passage, and two others were up to the third passage. Two cultures from the older donor group were carried through the first passage with viable and healthy cells.
INVESTIGATIVE OPHTHALMOLOGY & VI5UAL SCIENCE / February 1989
Table 3. Growth of normal human endothelial cells from whole corneas Donor age (years)
Total cultures 4 1 2 3
0 0 0 0 0 0 0