Hepatocyte Growth Factor, Keratinocyte Growth Factor, and Other Growth Factor-Receptor Systems in the Lens

Home , Fibroblast growth factor, Hepatocyte growth factor

Jian Weng,* Qianwa Liang,~\ Rajiv R. Mohan,~f Qian Li* and Steven E. Wilsonf

Purpose. To examine the expression and function of hepatocyte growth factor (HGF), keratinocyte growth factor (KGF), epidermal growth factor (EGF) and other growth factor- cytokine-receptor systems in lens epithelial cells. Methods. Reverse transcription-polymerase chain reaction (RT-PCR) and Northern blot analysis were used to examine the expression of messenger RNAs in primary cultured rabbit and human lens cells and in ex vivo rabbit lens tissue. Protein expression and the effect of HGF and KGF on crystallin expression in lens epithelial cells were evaluated by immunoprecipitation and Western blot analysis. The effect of exogenous HGF, KGF, and EGF and of the coculture of lens epithelial cells with corneal endothelial cells on the proliferation of rabbit lens cells in a Transwell system was determined by cell counting. Results. Messenger RNAs and proteins of HGF and KGF were expressed in primary rabbit lens epithelial cells and in ex vivo rabbit lens epithelial tissue. Human lens cells also expressed the mRNAs. Other growth factors and receptor messenger RNAs were also expressed. Hepato- cyte and keratinocyte growth factors, and coculture with corneal endothelial cells stimulated proliferation of rabbit lens epithelial cells. In first-passage rabbit lens cells, HGF, KGF, and EGF increased the expression of alpha and beta crystallins. Conclusions. Hepatocyte and keratinocyte growth factor-receptor systems are expressed in lens cells. HGF and KGF are not expressed in epithelial cells in such tissues as skin, cornea, and lacrimal gland in which fibroblastic and epithelial cells interact in the formation of an organ. Expression of these growth factors in the lens may have evolved because the lens cells are relatively isolated within the anterior chamber of the eye. Our results suggest, however, that growth factors released by the corneal endothelium also could modulate lens functions (aquecrine interactions). Invest Ophthalmol Vis Sci. 1997;38:1543-1554.

VFrowth factors modulate such functions as prolifera- factor (PDGF), and insulin-like growth factor (IGF) tion, motility, differentiation, and apoptosis in lens have been implicated in the regulation of lens cell epithelial cells.1"3 Acidic fibroblast growth factor functions.1"3 (FGF-1), basic fibroblast growth factor (FGF-2), epi- The anatomic position of the lens within the ante- dermal growth factor (EGF), platelet-derived growth rior segment presents a regulatory milieu that is unique to the eye. Because the adult lens is avascular From the "Department of Ophthalmology, University of Texas Southwestern Medical and lacks innervation, all of the regulatory signals Center at Dallas, Texas, and the -fEye Institute and the Department of Cell Biology, must be derived from the lens cells themselves or from The Cleveland Clinic foundation, Ohio. Presented in pan at the annual meeting of the Association for Research in Vision signals delivered through the aqueous humor. All of and Ophthalmology, Fort Lauderdale, Florida, May, 1995. the nutritional requirements of the lens are provided Supported in part by U. S. Public Health Service grant EY10056from the National Eye Institute, National Institutes of Health, Bethesda, Maryland and The Cleveland by the aqueous humor, and it therefore seems plausi- Clinic Foundation, Cleveland, OH. ble that important regulatory signals could also be Submitted for publication September 23, 1996; revised February 20, 1997; accepted February 21, 1997. delivered through this medium. Proprietary interest category: p(SEW); N (JW, QL, RRM, QL) has financial interest in a patent related to applications of HGF and KGF to the eye. None of the other In results of recent studies on the expression and authors has any commercial or proprietary interest. function of corneal hepatocyte growth factor (HGF) Reprint requests: Steven. E. Wilson, Eye Institute and Department of Cell Biology/ A31, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH and of keratinocyte growth factor (KGF), we were sur- 44195. prised to discover that corneal endothelial cells pro-

Downloaded from iovs.arvojournals.org on 09/26/2021 1544 Investigative Ophthalmology & Visual Science, July 1997, Vol. 38, No. 8

duce HGF and KGF.4 These two growth factors have Board of the University of Texas Southwestern Medi- been identified in many tissues as paracrine modula- cal Center approved the use of human tissues in this tors of stromal-epithelial interactions secreted by fi- study. broblastic cells to regulate epithelial cell functions. Proliferation experiments were performed by plat- What functions could be served by endothelial-derived ing 2 X 104 first-passage rabbit cells per well in six- HGF and KGF? Is it possible that these growth factors well cluster plates (Falcon, Lincoln Park, NJ). Twenty- could be released from the endothelium to modulate four hours after seeding, cells were transferred into 2 the functions of epithelial-derived lens tissue that is ml of keratinocyte-defined medium (KDM: modified continuously exposed to the aqueous humor? We have MCDB 153, 5 //g/ml insulin, 0.5 /Ltg/ml hydrocorti- speculated previously that these types of growth fac- sone, and 0.15 mM Ca2+ without EGF; Clonetics, San tor- or cytokine-mediated communications might oc- Diego, CA) without growth factors, serum, or other cur between different cell types exposed to the aque- extracts. Growth factors to be tested were added to ous humor; and we have termed these hypothetical the medium. 4 regulatory mechanisms aquecrine interactions because At the time of addition, growth factors were di- this mode of communication is not defined ade- luted in sterile, phosphate-buffered saline (PBS) with quately by terms that are currently available to de- 0.2% gelatin (Sigma, St. Louis, MO) to provide 5 /il scribe growth factor-cytokine function (that is, auto- of additional volume to each well and to obtain the crine, paracrine, endocrine, and juxtacrine). This final concentration of growth factor. Each growth fac- study examined the expression of the HGF and KGF tor was tested in six wells in each experiment. Five growth factor-receptor systems in the lens and the microliters of PBS with 0.2% gelatin as a vehicle was effect of HGF, KGF, and cocultured corneal endothe- added to control flasks. Media were changed at 3-day lial cells on the proliferation of lens cells. The expres- intervals with the addition of fresh growth factors. sion of other growth factor-receptor system messen- Human recombinant double-chain HGF was ob- ger ribonucleic acids (mRNAs) in the human and rab- tained from Genentech (South San Francisco, CA), bit lens was also evaluated. human EGF from Collaborative Biomedical Products (Bedford, MA), and human KGF from Bachem (Phila- delphia, PA). Each of these human growth factors METHODS stimulates proliferation of rabbit corneal epithelial cells (Wilson SE, unpublished data, 1994), demonstra- Cell Culture, Ex Vivo Tissue, and Proliferation ting the cross-species efficacy of the growth factors. Assays Concentrations of HGF and KGF were at 5 ng/ml and Primary cultures of human5 and rabbit6 lens epithelial EGF was at 10 ng/ml, because these concentrations cells and of rabbit corneal endothelial7 cells were es- are optimal for a wide variety of cell types.8"15 Aliquots tablished by previously described methods from ex- of stock solutions of the growth factors were placed plants in Dulbecco's minimal essential medium in siliconized micro tubes and were stored at — 80°C to (DMEM; Gibco, Gaithersburg, MD) with 10% fetal minimize denaturation caused by repeated freezing bovine serum (Gibco). Human cells were derived and thawing. from fresh eyes of donors less than 2 years old in Coculture of first-passage rabbit lens epithelial compliance with the Helsinki Accord on Human cells with rabbit corneal endothelial cells was per- Rights. Rabbit cells were obtained from 7-week-old formed in six-well Transwell plates (Falcon, Lincoln New Zealand White rabbits. Rabbit lens tissue for re- Park, NJ) with a 0.45-/L/M pore size. Two X 104 first- verse transcription-polymerase chain reaction (RT- passage rabbit lens cells per well were plated in the PCR) was obtained by euthanizing New Zealand White bottom of the well with a monolayer of first-passage rabbits, immediately removing the corneoscleral rim rabbit corneal endothelial cells on the Transwell mem- with sharp scissors, expressing the intact lens, strip- brane insert. Two milliliters of KDM were added to ping the lens capsule with jeweler's forceps, transfer- the wells and changed at 3-day intervals. Lens epithe- ring the capsule with associated lens epithelial tissue lial cells alone with an empty Transwell insert or cor- into a 1.5-ml microfuge tube containing 0.75 ml of neal endothelial cells alone on the insert served as guanidinium thiocyanate solution (GTC)7 for RNA ex- control specimens. traction, and pulverizing with a Teflon disposable pes- After a 2-week exposure to growth factors or to tle (Fisher Scientific, Pittsburgh, PA). Extraction of cocultured endothelial cells, lens cells were photo- RNA from the GTC solution was achieved as previously graphed with an inverted phase-contrast microscope, described.7 These studies were conducted according washed once, and dissociated with 1 ml 0.25% trypsin to the ARVO Statement for the Use of Animals in (approximately 10 minutes). The total number of cells Ophthalmic and Vision Research and the tenets of the for each well was determined from the average of Declaration of Helsinki. The Investigational Review three counts made with a Coulter counter (Model Zf;

Downloaded from iovs.arvojournals.org on 09/26/2021 TABLE i. Primers and Expected Amplification Sizes n Size Modulator mRNA/Genomic Upstream Primer Downstream Primer Accession Numbers

H and R 0 actin 350/790 AGGCCAACCGCGAGAAGATGACC GAAGTCCAGGGCGACGTAGCAC X00351 HandRHGF (set 1) 177/>177 GAGTACTGTGCAATTAAAACATGC CATGTCATGCTTGTGAGGATA D90318 I H HGF (set 2) 522/>522 CATGGATCATACAGAATCAGG CCATTCCCACGATAACAATCT D90318 n H HGF receptor 336/432 TGGTCCTTTGGCGTCGTCCTC CTCATCATCAGCGTTATCTTC J02958, U11813 R HGF receptor 302/unk AAAGTTTACCACCAAGTCAGA CGTTCACATAAGTAGCGTTCA J02958 472/unk (Rex vivo)* TGATTTTGGTCTTGCCAGAGA CTCATCATCAGCGTTATCTTC J02958 H KGF 685/unk CACTTTATATGTTCACCAATG TGACTTATACTACCAGGAATG M60828 RKGF 322/unk GGGGATATAAGGGTGAGAAGA GATTTAAGGCAACGAACATTT Z22703, U58503 H and R KGF receptor 172/1200 GGATCAAGCACGTGGAAAAGA GCCCTATATAATTGGAGACCT M80638 104/104 (Rex vivo)* GCACTCGGGGATAAATAGTTC GCCCTATATAATTGGAGACCT M80638 H and R EGF 394/unk TATGTCTGCCGGTGCTCAGAA AGCGTGGCGCAGTTCCCACCA X04571 H and R EGF receptor 224/unk CACCTGCGTGAAGAAGTGTCC AGGTGCAGT1T1TGAAGTGTT X0O588 H basic FGF 404/>404 TTCAAGGACCCCAAGCGGCT GGCCATTAAAATCAGCTCTT M27968, X04433 H FGFR-1 800, 1000, 1100f/>1100 CGAGCTCACTGTGGAGTATCCATG GTTACCCGCCAAGCACGTATAC M60485 H FGFR-2 199/2400 GGATCAAGCACGTGGAAAAGA ACCATGCAGAGTGAAAGGATA M55614 H TGF-/? 1 244/>244 CCGCAAGGACCTCGGCTGGAA GATCATGTTGGACAGCTGCTC M38449, X05849, X05850 H TGF 0 receptor IJ 384/unk TATGAGCAGGGGAAGATGACG CTGTACCAGGAAAGGAAGACC L15436 H TGF 0 receptor II 466, 543§/unk CCACCGCACGTTCAGAAGTCG CTTGCCGGTTTCCCAGGTTGA M85079 HIL-1 a 711/unk GTTCCAGACATGTTTGAAGACCTG TGGATGGGCAACTGATGTGAAATA X02531 H IL-1 receptor, type I 379/unk TGACAAAATTGGCCAGAGAG TTGTGCTAAACCAGTTTTAA M27492 H PDGF a 675/>675 GCGATGAGGACCTTGGCTTGC TGCGGCTCATCCTCACCTCA X03795, M020488, M20494 H glucocorticoid receptor 626/>626 ATGAGACCAGATGTAAGCTC AATGCCATAAGAAACATCCA X03225, X03348

H and R = primers that will amplify human and rabbit mRNA, respectively; Size = expected size of the PCR amplification from mRNA and genomic DNA; Unk = sequences in which the genomic sequence is not known. > = amplifications in which the upstream and downstream primers are known to be in separate exons although the exact size of the genomic amplification is not known. Accession numbers refer to the Genbank accession numbers assigned to each sequence used is designing primers. * Additional primer pairs for KGF receptor and HGF receptor were designed for ex vivo rabbit lens samples. t Primers for FGF receptor-1 yield three alternative mRNAs yielding amplification products of 800 base pairs for the (3 amino-terminal motif, 1000 base pairs for the y amino-terminal motif, and 1100 base pairs for the a amino-terminal motif. X Primers for TGF 0 receptor I were designed from the mouse sequence. § Primers for TGF 0 receptor II identify two alternative mRNAs. The 543 base pair transcript corresponds to TGF 0 receptor II mRNA with a 25 amino acid insertion relative to the better characterized TGF 0 receptor II mRNA.20 HGF = hepatocyte growth factor; KGF = keratinocyte growth factor; EGF = epidermal growth factor; FGF = fibroblast growth factor; TGF = transforming growth factor; IL-1 = interleukin-1; PDGF = platelet-derived growth factor.

Downloaded from iovs.arvojournals.org on 09/26/2021 1546 Investigative Ophthalmology & Visual Science, July 1997, Vol. 38, No. 8

Hialeah, FL). Growth factors and the effect of endothelial cells were tested in three independent experi- 1 ments. recepto recepto Errors were expressed as the standard error of the o Actin HGF mean. Statistical comparisons were performed with UJ KGF the analysis of variance Dunn's test. P < 0.05 was con- BP 1 2 C 1 2 C 1 2 C 1 2 C BP ^^^^^^^^ sidered statistically significant. 600- 1 •• " -600 300 -• \• Reverse Transcription -Polymerase Chain 9 I I -300 Reaction 100-1 -100 Total cellular RNA was isolated from primary cultures FIGURE 2. Epidermal growth factor receptor, hepatocyte of rabbit and human lens epithelial cells and comple- growth factor receptor, and keratinocyte growth factor re- mentary DNA (cDNA), synthesized as previously de- 7 16 ceptor messenger RNA expression in human lens cells. Pri- scribed. ' The quality of cDNA synthesis was moni- mary cultures of human lens epithelial cells from young tored through the amplification of beta actin. Only donors were evaluated for the production of messenger cDNA yielding beta-actin amplifications of the ex- RNAs of these factors by reverse transcription-polymerase pected size for beta-actin mRNA, without contamina- chain reaction. Beta-actin expression was also evaluated as tion with a genomic beta-actin amplification product a positive control: 1 and 2 represent two different primary of a larger size, was used for experimental amplifica- lens epithelial cell cultures for each primer pair; C is a simul- tion. Polymerase chain reactions were performed with taneous control reaction without added complementary a temperature cycler (MJ Research, Watertown, MA) DNA target for each polymerase chain reaction primer pair. according to a previously described method.4 A modi- Amplification bands of the expected size (Table 1) were detected for each growth factor receptor and for beta actin fied hot-start method was used in which anti-Taq poly- (arroxos). Note that in the amplifications of hepatocyte merase antibody (Clontech, Palo Alto, CA) was added growth factor receptors an alternative band (approximately to the reactions according to the manufacturer's in- 430 base pairs) is present just above the expected band structions to prevent initiation of amplification until (arrow). This alternative band corresponds to an alternative hepatocyte growth factor receptor messenger RNA that codes for a truncated form of the receptor.12 The smaller EGF HGF-1 HGF-2 KGF amplification products in hepatocyte growth factor receptor BP 1 were artifacts in the polymerase chain reaction. A 100- -800 base-pair marker with sizes in base pairs (BP) is indicated. -500

200- -200 after the reactions were raised to the antibody-dena- turing temperature of 70°C. Amplification products FIGURE l. Epidermal growth factor, hepatocyte growth fac- of PCR were run on agarose gels, as previously de- tor, and keratinocyte growth factor messenger RNA expres- scribed.4 Primers for the human and rabbit growth sion in human iens cells. Primary cultures of human lens factors and receptors are provided in Table 1. Human epithelial cells from young donors were evaluated for the primers were designed from human Genbank se- production of messenger RNAs of these growth factors by quences, excepting primers that amplified human reverse transcription-polymerase chain reaction: 1 and 2 transforming growth factor beta (TGF-/3) receptor represent two different primary lens epithelial cell cultures from different donors for each primer pair; C is a simultane- type I mRNA, which were designed from the mouse ous control reaction without added complementary DNA sequence (Table 1). Human primers were tested with target for each polymerase chain reaction primer pair. Am- rabbit liver cDNA to determine whether they would plification bands of the expected size (Table 1) were de- amplify the rabbit sequence with an annealing temper- tected for each growth factor {arrows). Hepatocyte growth ature of 50°C. The original human KGF and HGF factor messenger RNA was detected with two different poly- receptor primers would not amplify rabbit cDNA. merase chain reaction primer pairs. There are numerous Therefore, alternative primers that amplified the rab- alternative polymerase chain reaction products with the epi- bit sequence at an annealing temperature of 50°C dermal growth factor primers that are attributable to the were designed from the mouse (KGF) and human large number of messenger RNAs having sequence similarity (HGF) receptor sequences. In addition, second sets of to epidermal growth factor. The expected product for epi- PCR primers for rabbit HGF and rabbit KGF receptors dermal growth factor, however, was detected and confirmed were designed (Table 1). Human and rabbit PCR by nucleic acid sequencing. Minor bands in hepatocyte products of the expected sizes were cut from agarose growth factor-2, lane 2, and keratinocyte growth factor, lane gels, cloned into the II Cloning Vector (Invitrogen, 2; were artifacts in the polymerase chain reaction by nucleic acid sequencing. A 100-base-pair marker with sizes in base San Diego, CA), and sequenced (Sequenase 2.0; pairs (BP) is indicated. United States Biochemical, Cleveland, OH) according

Downloaded from iovs.arvojournals.org on 09/26/2021 Growth Factor-Receptor Systems in the Lens 1547

viously described method.17 RNA size markers and 9 jig poly(A)+ RNA per lane were resolved on 1.2% E agarose formamide gels. The RNA was transferred to o Duralon-UV (Stratagene, San Diego, CA) using a vac- UL BP 12C12C12C12C12C12C BP uum blotter (Pharmacia, Piscataway, NJ) and cross- 900- -900 linked to the Duralon-UV with the UV-Stratalinker 600- -600 (Stratagene). 300- -300 Probes for HGF and KGF were generated by PCR from rabbit corneal epithelial cells or from the total FIGURE 3. Basic fibroblast growth factor, fibroblast growth cellular RNA of stromal fibroblast cells, using the spe- factor receptor-1, fibroblast growth factor receptor-2, trans- cific primers (Table 1). Amplification products of the forming growth factor beta-1, transforming growth factor beta receptor type I, and transforming growth factor beta expected size were cut from agarose gels, cloned into receptor type II messenger RNA expression in human lens the PCR II Cloning Vector, and sequenced according cells. Primary cultures of human lens epithelial cells from to the manufacturer's protocols. Sequences were con- young donors were evaluated for the production of messen- firmed to be specific for the mRNA for HGF and KGF. ger RNAs of these factors by reverse transcription-polymer- Each sequence was compared with all known se- ase chain reaction: 1 and 2 represent two different primary quences in the Genbank and EMBL databases using lens epithelial cell cultures from different donors for each die Mac Vector 5.0 System (IBI, New Haven, CT). No primer pair; C is a simultaneous control reaction without other nucleic acid sequences have been reported that added complementary DNA target for each polymerase would be expected to hybridize to die HGF or KGF chain reaction primer pair. Amplification bands of the ex- probes under the conditions used for Northern blot pected size (Table 1) were detected for each growth factor and receptor (arrows). The fibroblast growth factor receptor- analysis. 1 (fig) primers yielded three alternative polymerase chain Purified radiolabeled DNA probes with a specific reaction products that correspond to alternative fibroblast activity of approximately 2X10° cpm///g were pre- growth factor receptor-1 transcripts.19 The 1100-base-pair pared with the cloned HGF and KGF sequences, using polymerase chain reaction amplification product corre- the random primer-labeling kit Rediprime and 32P- sponds to the alpha amino-terminal motif of fibroblast dCTP (Amersham Life Sciences, Arlington Heights, growth factor receptor-1 containing three extracellular im- IL)) according to die manufacturer's protocol. Probe munoglobulin G-like disulfide loops. The 800-base-pair for the RNA size markers was prepared by 32P-labeling product corresponds to the beta amino-terminal motif of lambda genomic DNA. A total of 5 X 107 cpm of probe fibroblast growth factor receptor-1 containing two extracellular immunoglobulin G-Hke disulfide loops. Finally, the was used in each hybridization. Hybridization was per- 1000-base-pair amplification product corresponds to the gamma amino-terminal motif of fibroblast growth factor receptor-1. This motif does not contain a known signal sequence for membrane translocation and is thought to represent an intracellular form of fibroblast growth factor receptor-1. Similarly, two alternative transcripts for transforming growth factor beta receptor type II were detected in human lens cells.20 Within a lane the ratio of the alternative amplifi- BP cation products should be similar to the ratios of the target messenger RNAs in the complementary DNA. Therefore, the alternative transforming growth factor beta receptor type II messenger RNAs appear to be expressed at approxi- 300 -300 mately equal levels in cultured human lens cells. A 100-base pair marker with sizes in base pairs (BP) is indicated. FIGURE 4. Platelet-derived growth factor alpha, glucocorticoid receptor, interleukin-1 alpha, and interleukin-1 receptor type 1 messenger RNA expression in human lens cells. to the manufacturer's protocols, to verify that they Primary cultures of human lens epithelial cells from young were specific. donors were evaluated for the production of the messenger RNAs of these cell factors by reverse transcription-polymer- Northern Hybridization to Detect Hepatocyte ase chain reaction: 1 and 2 represent two different primary Growth Factor and Keratinocyte Growth Factor lens epithelial cell cultures from different donors for each mRNA in Rabbit Lens Cells primer pair; C is a simultaneous control reaction without added complementary DNA target for each polymerase Total cellular RNA was isolated using TRIzol (BRL, chain reaction primer pair. Amplification bands of the ex- + Gaithersburg, MD). Poly(adenylic acid; A) RNA was pected size (Table 1) were detected for each growth factor isolated using an mRNA Isolation Kit (Boehringer and receptor (airoius). A 100 base pair marker with sizes in Mannheim, Indianapolis, IN), according to a pre- base pairs (BP) is indicated.

Downloaded from iovs.arvojournals.org on 09/26/2021 1548 Investigative Ophthalmology & Visual Science, July 1997, Vol. 38, No. 8

blot analysis. Protein molecular weight standards (Am- ersham Life Sciences) were run simultaneously on the Western blot analysis gels (4% to 20% gradient poly- acrylamide; Novel Experimental Technology, San BP 2 C Diego, CA). 600 400 Western Blot Analysis to Evaluate the Effect of 200-1 Growth Factors on Crystallin Expression Fourth-passage rabbit lens cells from a single batch of FIGURE 5. Epidermal growth factor, epidermal growth factor rabbit eyes were subcultured into KDM without growth receptor, hepatocyte growth factor, hepatocyte growth fac- factors, serum, or other extracts and exposed to vehi- tor receptor, keratinocyte growth factor, and keratinocyte cle solution only and to 20 ng/ml EGF, 10 ng/ml growth factor receptor messenger RNA expression in rabbit KGF, or 10 ng/ml double-chain HGF. The cytosolic lens cells. Primary cultures of rabbit lens epithelial cells were evaluated for the production of messenger RNAs of these protein fractions were isolated after 1, 3, or 7 days of exposure to vehicle solution or to growth factors, as factors by reverse transcription-polymerase chain reaction: 1 1 and 2 represent two different primary lens epithelial cell previously described. ' Gels were run under reducing cultures for each primer pair; C is a simultaneous control conditions with SDS." One hundred micrograms of reaction without added complementary DNA target for each polymerase chain reaction primer pair. Amplification bands of the expected size (Table 1) were detected for each growth HGF KGF EGF factor and receptor (arrows). There were alternative poly- M L1 L2 Li merase chain reaction products with the epidermal growth 500-- factor primers that were attributable to the large number of messenger RNAs that have sequence similarity to epidermal growth factor. The expected product for epidermal growth factor, however, was detected and confirmed by nucleic acid 100- sequencing. A 100 base pair marker with sizes in base pairs (BP) is indicated.

HGFR KGFR EGFR formed, after 4 hours of prehybridization, at 42°C for 16 hours in a hybridization solution containing 5 X SSPE, 50% form amide, 5 X Denhardt's solution, 0.5% sodium dodecyl sulfate (SDS), and 500 /^g/ml sheared and denatured salmon sperm DNA. Washes were performed with 2 to 0.2 X SSPE with 0.1% SDS at 37°C to 56°C. Blots were exposed to X-OMAT 5 (Eastman Kodak, Rochester, NY) film and were developed after 14 hours to 10 days of exposure at —85°C. FIGURE 6. Hepatocyte growth factor, keratinocyte growth fac- Immunoprecipitation and Western Blot Analysis tor, epidermal growth factor, hepatocyte growth factor re- to Detect Growth Factor Proteins ceptor, keratinocyte growth factor receptor, and epidermal growth factor receptor messenger RNA expression in ex vivo The soluble protein fraction used for immunoprecipi- rabbit lens tissue. Growth factor and receptor results are 17 tation of HGF or KGF was isolated from rabbit pri- shown in (A) and (B), respectively. Ex vivo rabbit lens tissue mary lens cells, ex vivo rabbit lens tissue, and rabbit was evaluated for the production of the growth factor and corneal stromal fibroblast cells. Immunoprecipitation receptor messenger RNA by reverse transcription-polymer- of HGF from the cytosolic protein fraction and West- ase chain reaction: LI and L2 represent two different ex ern blot analysis under nonreducing conditions were vivo lens tissue specimens from different animals for each performed as previously reported17 using the 1A3.1.2 primer pair; Li is a positive control reaction performed with anti-HGF antibody that has been previously character- complementary DNA from rabbit liver; C is a simultaneous control reaction without added complementary DNA target ized.18 The 1A3.1.2 antibody was used at a final con- 4 for each polymerase chain reaction primer pair. Amplifica- centration of 5 X 10~ mg/ml in Western blot analysis. tion bands of the expected size (Table 1) were detected for Similarly, immunoprecipitation and Western blot each growth factor and receptor {arrojus). M indicates a 100 analysis for KGF were performed using a commercially base pair marker with sizes in base pairs indicated. Alterna- available anti-KGF antibody (AF251-NA; RD Systems, tive bands for epidermal growth factor and hepatocyte Minneapolis, MN). The anti-KGF antibody was used growth factor receptor were determined to be artifacts of at a final concentration of 1 X 10~3 mg/ml in Western polymerase chain reaction.

Downloaded from iovs.arvojournals.org on 09/26/2021 Growth Factor-Receptor Systems in the Lens 1549

HGF KGF Two alternative transcripts for TGF-/3 receptor type 20 L L II mRNA were expressed at approximately equal lev- M 1 2 els. Messenger RNAs for platelet-derived growth factor alpha, the glucocordcoid receptor, interleukin-1 • 4.4 4.4- alpha, and interleukin-1 receptor type I were also detected in primary cultured human lens cells (Fig. 4). Detected in primary cultured rabbit lens epithelial 2.4- 2.1 •1.8 cells (Fig. 5) and in fresh ex vivo rabbit lens tissue (Figs. 6A, 6B) were mRNAs for EGF, HGF, KGF, and 1.4- their receptors. Messenger RNA was detected by RT- B PCR for both HGF receptors and KGF receptors with FIGURE 7. Northern blot analysis to detect hepatocyte growth two different sets of PCR primers (Table 1). The PCR factor and keratinocyte growth factor messenger RNA in obtained with the second set of primers for each of rabbit lens epithelial cells. (A) Transcripts approximately 4.4 these receptors is shown in Figure 6B. kb and 1.8 kb in size were detected with a rabbit hepatocyte- All human and rabbit PCR products were con- growth factor-specific probe in first-passage rabbit lens (LI firmed to be specific by nucleic acid sequencing. and L2) and in positive control rabbit corneal fibroblast Northern blot analysis was performed to confirm (SF) cells. M indicates RNA markers with sizes indicated, in that HGF and KGF mRNAs are produced in rabbit kilobases, to the left. (B) Transcripts approximately 5.1 kb lens epithelial cells. Two HGF transcripts (approxi- and 2.1 kb in size were detected with a rabbit keratinocyte mately 4.4 kb and 1.8 kb in size) were detected in growth factor-specific probe in first-passage rabbit lens (LI and L2) and positive control rabbit corneal fibroblast (SF) rabbit lens epithelial cells and in positive-control rab- cells. M indicates RNA markers with sizes indicated, in kilo- bit corneal stromal fibroblast cells (Fig. 7A). Two KGF bases, to the left. transcripts (approximately 5.1 kb and 2.1 kb in size) were detected in rabbit lens cells and in positive-control rabbit corneal fibroblast cells. (Fig. 7B). total protein were added per lane. Western blot analysis was performed with rabbit polyclonal antibody to total human alpha-crystallin fraction and rabbit poly- KGF HGF clonal antibody to total human beta-crystallin fraction L SF L SF that were kindly provided by Larry J. Takemoto, PhD. 90 KDa The anti-alpha-crystallin and anti-beta-crystallin anti- 28 KDa—• __ jg body solutions provided were used in final dilutions i of 1:5000. Otherwise, Western blot analysis was performed as described earlier. A HGF HGFR KGF RESULTS 1 2 1 2 1 2 90 KDa —• *~ Messenger RNAs of EGF, HGF, and KGF were de- 180 KDa tected by PCR in primary cultured human lens epithe- 28 KDa lial cells (Fig. 1). Hepatocyte growth factor mRNA was detected using two different sets of PCR primers. B Receptor mRNAs for EGF, HGF, and KGF were also FIGURE 8. Tmmunoprecipitation and Western blot analysis. produced by primary cultured human lens epithelial (A) Keratinocyte growth factor and hepatocyte growth factor cells (Fig. 2). In addition to the well-characterized met proteins were detected in cultured rabbit lens epithelial 12 transcript, a previously reported alternative HGF re- cells. Proteins of the expected size for keratinocyte growth ceptor transcript corresponding to a variant receptor factor (28 kDa) and hepatocyte growth factor (90 kDa) were that is truncated in the intracytoplasmic domain was detected in primary cultured rabbit lens epithelial (L) and detected (Fig. 2). corneal stromal fibroblast (SF) cells. (B) hepatocyte growth Basic FGF and FGF receptor-1 mRNAs were de- factor, hepatocyte growth factor receptor, and keratinocyte tected in human lens epithelial cells (Fig. 3). Three growth factor proteins were detected in ex vivo rabbit lens previously reported alternative transcripts for FGF re- tissue: 1 and 2 represent two ex vivo lens epithelial speci- 19 mens. Proteins of the expected size for hepatocyte growth ceptor-1 that are resolved by these PCR primers were factor (90 kDa), hepatocyte growth factor receptor (ISO detected in human lens epithelial cells; FGF receptor- kDa), and keratinocyte growth factor (28 kDa) were de- 2 mRNA was not detected. Messenger RNA for TGF- tected. Alternative bands in the ex vivo hepatocyte growth (3\ and mRNA for TGF-/? receptor type I and TGF-/? factor and hepatocyte growth factor R blots were derived receptor type II were detected in human lens cells. from immunoglobulin.

Downloaded from iovs.arvojournals.org on 09/26/2021 1550 Investigative Ophthalmology & Visual Science, July 1997, Vol. 38, No. 8

FIGURE 9. Primary cultures of human (A) and rabbit (B) lens epithelial cells. Rabbit lens epithelial cells in a Transwell culture without rabbit corneal endothelial cells (C) or with rabbit corneal endothelial cells (D) after 2 weeks of culture. There was a much larger number of epithelial cells in the Transwell culture with cornea! endothelial cells (Fig. 11) after 2 weeks in coculture. The matted background in C and D is generated by photo- graphing through the Transwell membrane.

Western blot analysis was performed to detect p\\ TGF-/? receptor type I; TGF-/3 receptor type II; HGF, HGF receptors, and KGF protein in primary platelet-derived growth factor alpha; interleukin-la; cultured rabbit lens epithelial cells and in ex vivo rab- interleukin-1 receptor type I; and the glucocorticoid bit lens tissue. Primary rabbit corneal fibroblasts were receptor were detected by RT-PCR in primary cul- used as positive controls, because they are known to tures of human lens epithelial cells from donors new- produce HGF and KGF.418 Both HGF and KGF pro- born to 2 years old. Messenger RNAs, coding for EGF, teins were detected in cultured rabbit lens epithelial HGF, KGF and their receptors, were also detected in cells (Fig. 8A) and in ex vivo rabbit lens tissue (Fig. rabbit primary lens epithelial cells. Importantly, we 8B). Hepatocyte growth factor receptor protein was confirmed that mRNAs, coding for HGF, KGF, EGF, detected in ex vivo rabbit lens tissue (Fig. 8B). and their receptors, were detectable in ex vivo rabbit First-passage rabbit lens epithelial cells (Fig. 9B) lens tissue, proving that expression of these mRNAs were stimulated (Fig. 10) to proliferate by EGF and is not an artifact of tissue culture. Other investigators HGF, but not by KGF. First-passage rabbit lens epithe- have demonstrated production of growth factors, cyto- lial cells were also stimulated to proliferate by a con- kines, and their receptors in lens epithelial cells, in- fluent monolayer of first-passage rabbit corneal endo- cluding TGF-/?/23 acidic FGF,24 basic FGF,25 insulin- thelial cells in a Transwell culture (Figs. 9C, 9D, 11). like growth factor-1,26 and EGF receptors.27 Thus, lens Increased expression of alpha- (Fig. 12A) and beta- epithelial cells transcribe mRNAs for a large number (Fig. 12B) crystallin proteins was stimulated by EGF, KGF, of growth factors and receptors that may modulate and HGF in cultured rabbit lens epithelial cells compared functions of lens cells. with that in control vehicle, detected by Western blot Both HGF and KGF have been best characterized analysis. The difference was especially prominent for as paracrine mediators expressed by fibroblastic cells alpha A crystallins 1 day after exposure but could also be to modulate such functions as proliferation, motility, noted at each time point, with each growth factor, for and differentiation in the epithelial cells of many tis- alpha A, alpha B, and beta crystallins. sues. Results of recent studies,4'18 showed that HGF and KGF were produced in corneal endothelial cells. DISCUSSION Although the HGF and the KGF produced by these Messenger RNAs (mRNA) coding for HGF, KGF, EGF, cells could have autocrine effects on endothelial cells and their receptors; basic FGF; FGF receptor-1; TGF- or paracrine effects on other cells of the cornea, we

Downloaded from iovs.arvojournals.org on 09/26/2021 Growth Factor-Receptor Systems in the Lens 1551

by RT-PCR were produced at physiologically significant levels and whether they were translated into proteins. We used rabbit lens epithelial cells for the Northern and Western blot analysis experiments because we did not have access to enough human lens tissue to generate the quantities of first-passagehuma n lens epithelial cells needed for these experiments.2 Messenger RNAs for HGF and KGF were detected in rabbit lens epithelial cells by Northern blot analysis. The sizes of the transcripts were identical to those detected in rabbit corneal stromal fibroblasts (Fig. 7). Two different sizes of mRNA for both HGF and KGF were detected. Alternative sized bands for HGF1112 and KGF15 that code for the corresponding proteins have been detected previously in other systems. The significance of the alternatively sized mRNAs is unknown, but in human cells both the larger and smaller HGF transcripts code for the full-length HGF precur- FIGURE 10. Effect of epidermal growth factor, hepatocyte sor protein.11'30 Proteins of HGF and KGF of the ex- growth factor, and keratinocyte growth factor on the prolif- pected sizes were also detected in rabbit lens epithelial eration of first-passage rabbit lens epithelial cells. The effect cells (Fig. 8A) and in ex vivo rabbit lens epithelial of each growth factor and the control was tested in 12 wells. Epidermal growth factor (10 ng/ml) and hepatocyte growth tissue (Fig. 8B) by immunoprecipitation and Western factor (5 ng/ml) increased the proliferation of rabbit lens epithelial cells. The increase was statistically significant (**) 40 -f compared with the increase in control cells (epidermal growth factor, P < 0.0001; hepatocyte growth factor, P = 0.002). The effect of epidermal growth factor was signifi- o cantly greater than that of hepatocyte growth factor (P < 0.0001). Keratinocyte growth factor at 10 ng/ml was not different from vehicle (con). There was a trend toward inhibition of proliferation of rabbit lens epithelial cells by keratinocyte growth factor, but the decrease was not statistically significant (P = 0.16). The results in three experiments were Q. similar for each growth factor. 0) O hypothesized that HGF and KGF produced by the corneal endothelium could regulate functions of the epithelial lens cells through the aqueous humor. We have 4 termed such communications aquecrine interactions RLE ONLY RLE/RCN and have hypothesized that these interactions may oc- Group cur not only between corneal endothelial cells and FIGURE 11. Effect of cocultured rabbit corneal endothelial lens cells, but also between endothelial cells, trabecu- cells (RCN) on the proliferation of first-passage rabbit lens lar meshwork cells, and other cells that are immersed 4 28 epithelial (RLE) cells. Keratinocyte-defined medium within the aqueous humor. ' Communication by aque- out growth factors was added to the wells and changed at crine mechanisms from endothelial cells to lens cells 3-day intervals. Rabbit lens epithelial cells alone and in co- would necessarily occur retrograde to the direction of culture with rabbit corneal endothelial cells were tested in flow of the aqueous humor. Fluorophotometry has six wells each. Lens epithelial cells with a confluent mono- demonstrated, however, that the diffusion of soluble layer of rabbit corneal endothelial cells on the Transwell factors from the cornea to the lens and even to the insert (RLE/RCN) had significantly (**) increased prolifer- vitreous humor can occur.29 ation compared with lens cells with an empty Transwell insert (P = 0.0001). No cells were detected in the bottom We were surprised to detect HGF and KGF mRNA of the Transwell when the confluent monolayer of rabbit by RT-PCR in primary cultured human and rabbit endothelial cells were cultured on the insert without lens lens epithelial cells and ex vivo rabbit lens tissue. Be- cells in the bottom of die six-well plate, indicating that endo- cause HGF and KGF are not normally produced by thelial cells could not pass through the pores in the 49 15 epithelial cells, " it was important to confirm Transwell membrane. Similar results were obtained in two whether the HGF and KGF mRNA transcripts detected experiments.

Downloaded from iovs.arvojournals.org on 09/26/2021 1552 Investigative Ophthalmology 8c Visual Science, July 1997, Vol. 38, No. 8 Days Days 1 3 7 Unk 1 3 7 Unk alpha B CON beta (28 KDa) CON alpha A (21 KDa)

Unk Unk alpha B beta (28 KDa) alpha A (21 KDa) EGF

Unk Unk alpha B beta (28 KDa) KGF alpha A (21 KDa) KGF

Unk Unk alpha B HGF alpha A (21 KDa) HGF beta (28 KDa) B FIGURE 12. Effect of epidermal growth factor, keratinocyte growth factor, and hepatocyte growth factor on alpha- and beta-crystallin expression in first-passage rabbit lens cells. (A) The expected approximately 21-kDa alpha A and 22-kDa alpha B crystallins were detected,21'22 and the expression of both was increased by epidermal growth factor, keratinocyte growth factor, or hepatocyte growth factor. The origin of a somewhat larger band (Unk) is unknown. (B) A 28-kDa beta-crystallin protein (horizontal arrow) was detected and its expression was also increased by epidermal growdi factor, keratinocyte growth factor, or hepatocyte growth factor, A slighdy larger band (diagonal arrow) of unknown origin is very faint.

blot analysis. The HGF receptor protein of the ex- diffusable factor released by the corneal endothelial pected size was also detected in ex vivo rabbit lens cells is HGF or KGF. tissue (Fig. 8B). These data demonstrate that rabbit What functions of lens epithelial cells are regu- lens cells produce HGF, HGF receptors, and KGF. lated by HGF and KGF? Double-chain, mature HGF One possibility is that lens epithelial cells developed stimulates the proliferation of first-passage rabbit lens the capacity to produce HGF and KGF as the lens epithelial cells. It is important that double-chain HGF evolved into a relatively isolated epithelial structure be used, in that single-chain precursor HGF has no suspended within the anterior segment of the eye. effect on cultured lens epithelial cells (Wilson SE, un- Hepatocyte and keratinocyte growth factors could published data, 1994). Whether lens-cell produced function within lens cells by autocrine or paracrine HGF is processed into the active growth factor could mechanisms. These data, however, do not eliminate not be determined in the current study. Human KGF the possibility that HGF and KGF secreted into the had no effect on the proliferation of rabbit lens cells, anterior chamber by the corneal endothelium also although there was a trend toward inhibition. Human modulate lens epithelial cell functions. In fact, the KGF stimulates the proliferation of rabbit corneal epi- results of the Transwell experiments performed in this thelial cells (Wilson SE, unpublished data, 1994), and study demonstrate that a diffusable factor released it is therefore unlikely that the species difference be- from corneal endothelial cells stimulates lens epithe- tween the growth factor and cultured cells accounts lial cell proliferation. Future studies using neutralizing for the lack of stimulation. The expression of alpha antibodies will be needed to determine whether the and beta crystallins in first-passage rabbit lens epithe-

Downloaded from iovs.arvojournals.org on 09/26/2021 Growth Factor-Receptor Systems in the Lens 1553

lial cells was was stimulated by HGF, KGF, and EGF. Institute, Bethesda, MD) for helpful discussion and sugges- The effect of each of the growth factors was most tions, and Larry J. Takemoto (Kansas State University, Man- prominent for alpha crystallin after 1 day of exposure. hatten, KS) for antibodies to lens crystallins. There was, however, increased expression of alpha References and beta crystallin in cultured rabbit lens cells with HGF, KGF, and EGF compared with that expressed 1. Lovicu FJ, McAvoy JW. Localization of acidic fibroblast growth factor, basic fibroblast growth factor, and hep- by the vehicle solution after 1, 3, or 7 days of exposure. arin sulphate proteoglycan in rat lens: Implications Epidermal growth factor has been shown previously for lens polarity and growth patterns. Invest Ophthalmol to increase production of gamma crystallin in human 2 VisSci. 1993; 34:3355-3365. lens epithelial cells. We were unable to obtain an 2. Ibaraki N, Lin LR, Reddy VN. Effects of growth factors antibody that reliably detected rabbit gamma crystallin on proliferation and differentiation of human lens to evaluate the effect of HGF or KGF on gamma-crys- epithelial cells in early subculture. Invest Ophthalmol tallin expression. VisSci. 1995;36:2304-2312. Alternative splicing of mRNA is a common mecha- 3. Robinson ML, MacMillan-Crow LA, Thompson JA, nism used by cells to generate growth factor and cyto- Overbeek PA. Expression of a truncated FGF receptor kine receptors with varying functions. Alternative re- results in defective lens development in transgenic mice. Development. 1995; 121:3959-3967. ceptors generated from a specific genomic sequence 4. Wilson SE, Walker JW, Chwang EL, He Y-G. Hepato- may have differences in ligand specificity, intracellular cyte growth factor (HGF), keratinocyte growth factor and extracellular localization, phosphorylation, tyro- (KGF), their receptors, FGF Receptor-2, and the cells sine kinase activity, and intracellular substrate interac- of the cornea. Invest Ophthalmol Vis Sci. 1993; 34:2544- tion that modulate the effects produced by the growth 2561. factor, cytokine, or hormone ligand.1219 In the current 5. AritaT, Lin L-R, Reddy VN. Differentiation of human study, alternative mRNAs coding for HGF receptor- lens epithelial cells in tissue culture. Exp Eye Res. 1, FGF receptor-1, and TGF-/? receptor type II were 1988;47:905-910. detected by RT-PCR. The alternative HGF receptor 6. Reddan JR, Chepelinsky AB, Dziedzic DC, Piatogorsky mRNA has been shown to code for a receptor having J, Goldenberg EM. Retention of lens specificity in long HGF ligand-binding specificity with a truncated intra- term cultures of diploid rabbit lens epithelial cells. cytoplasmic domain. Whether this alternative receptor Differentiation. 1986; 33:168-174. 7. Wilson SE, Lloyd SA. Epidermal growth factor and its functions as a dominant negative inhibitor of HGF receptor, basic fibroblast growth factor, transforming stimulation or activates alternative signal transduction 12 growth factor beta-1, and interleukin-1 alpha messen- pathways is under investigation. The alternative FGF ger RNA production in human corneal endothelial receptor-1 mRNAs code for receptor variants having cells. Invest Ophthalmol Vis Sci. 1991;32:2747-2756. differing extracellular domains (Fig. 3). The functions 8. Wilson SE, He Y-G, Weng J, Zeiske JD, JV Jester, of these alternative FGF receptors have not been well Schultz GS. Effect of epidermal growth factor, hepato- characterized. Although the TGF-/3 receptor type II cyte growth factor, and keratinocyte growth factor, on variant we detected in lens cells had been previously proliferation, motility, and differentiation of human reported,20 its function remains unknown. corneal epithelial cells. Exp Eye Res. 1994; 59:665-678. Thus, the HGF and KGF growth factor-receptor 9. Nakamura T, Nishizawa T, Hagiya M, et al. Molecular cloning and expression of hepatocyte growth factor. systems appear to function locally within the lens. Fu- Nature. 1989; 342:440-443. ture investigations into endogenous growth factor, cy- 10. Matsumoto K, Hashimoto K, Yoshikawa K, Nakamura tokine, and receptor expression in lens cells and the T. Marked stimulation of growdi and motility of hu- functions that they regulate should lead to a better man keratinocytes by hepatocyte growth factor. Exp understanding of the development, homeostasis, and Cell Res. 1991; 196:114-120. pathophysiology of the lens. In addition, investigations 11. Rubin JS, Chan AM-L, Bottaro DP, et al. A broad- into communication between cells of the anterior seg- spectrum human lung fibroblast-derived mitogen is a ment of the eye should provide new insights into the variant of hepatocyte growth factor. Proc Natl Acad Sci physiology and pathophysiology of the eye. USA. 1991;88:415-419. 12. Wilson SE, Weng J, Chwang E, et al. Hepatocyte Key Words growth factor (HGF), keratinocyte growth factor (KGF), and their receptors in human breast cells and cell-cell interactions, crystallins, differentiation, hepatocyte tissues: Alternative HGF and KGF receptors. Molecular growth factor, keratinocyte growth factor, lens, proliferation and Cellular Biology Research. 1994;40:337-350. 13. Matsumoto K, Hashimoto K, Yoshikawa K, Nakamura Acknowledgments T. Marked stimulation of growth and motility of hu- The authors thank Ralph Schwall of Genentech for his assis- man keratinocytes by hepatocyte growth factor. Exp tance with this study, David C. Beebe (Washington Univer- Cell Res. 1991; 196:114-120. sity, St. Louis, MO) and Joram Piatigorsky (National Eye 14. Marchese C, Rubin J, Ron D, et al. Human keratino-

Downloaded from iovs.arvojournals.org on 09/26/2021 1554 Investigative Ophthalmology & Visual Science, July 1997, Vol. 38, No. 8

cyte growth factor activity on proliferation and differ- 23. Potts JD, Bassnett S, Beebe DC. Expression of trans- entiation of human keratinocytes: Differentiation re- forming growth factor beta in the embryonic avian sponse distinguishes KGF from EGF family. J CellPhys- lens coincides with the presence of mitochondria. Dev iol. 1990; 144:326-332. Dyn. 1995;203:3l7-322. 15. Finch PW, Rubin JS, Miki T, Ron D, Aaronson SA. 24. Renaud F, Oliver L, Desset S, et al. Up-regulation of Human KGF is FGF-related with properties of a para- aFGF expression in quiescent cells is related to cell crine effector or epithelial cell growth. Science. survival. / Cell Physiol. 1994; 158:435-443. 1989; 245:752-755. 25. Schweigerer L, Ferrara N, Haaparanta T, Neufeld G, 16. Wilson SE, Li Q, Weng J, et al. The Fas/Fas ligand Gospodarowicz D. Basic fibroblast growth factor: Ex- system and other modulators of apoptosis in the pression in cultured cells derived from corneal endo- cornea. Invest Ophthalmol Vis Sci. 1996; 37:1582- thelium and lens epithelium. ExpEyeRes. 1988; 46:71- 1592. 80. 17. SambrookJ, Fritsch EF, Maniatis T. Molecular Cloning. 26. Caldes T, Alemany J, Robcis HL, de Pablo F. Expres- A Laboratoiy Manual. 2nd ed. Cold Spring Harbor, sion of insulin-like growth factor I in developing lens NY: Cold Spring Harbor Laboratory Press; 1989:7.37- is compartmentalized. / Biol Chem. 1991;266:20786- 7.50, 8.60-8.74, 18.30-18.75. 20790. 18. Li Q, Weng J, Mohan R, et al. Hepatocyte growth 27. Hongo M, Itoi M, Yamamura Y, Yamaguchi N, Iman- factor (HGF) and HGF receptor protein in lacrimal ishi J. Distribution of epidermal growth factor recep- gland, tears, and cornea. Invest Ophthalmol Vis Sci. tors in rabbit lens epithelial cells. Invest Ophthalmol Vis 1996;37:727-739. Sci. 1993; 34:401-404. 19. Hou J, Kan M, McKeehan K, McBride G, Adams P, 28. Wordinger RJ, Clark AF, Wilson SE. Expression of McKeehan WL. Fibroblast growth factor receptors EGF, HGF, KGF, basic FGF, TGF, and their receptor from liver vary in three structural domains. Science. mRNAs in cultured human trabecular meshwork 1991;251:665-668. (HTM) cells. ARVO Abstracts. Invest Ophthalmol Vis 20. Nikawa J. A cDNA encoding the human transforming Sci. 1996; 37:4147. growth factor beta receptor suppresses the growth de- 29. Reidy JJ, Limberg M, Kaufman HE. Delivery of fluo- fect of a yeast mutant. Gene. 1994; 149:367-372. rescein to the anterior chamber using the corneal 21. Kantrorow M, Piatigorsky J. Alpha-crystallin/small collagen shield. Ophthalmology. 1990;97:1201-1203. heat shock protein has autokinase activity. Proc Nat 30. Chan A, Rubin J, Bottaro D, Hirschfield D, Chedid AcadSci USA. 1994;91:3112-3116. M, Aaronson SA. Isoforms of human HGF and their 22. BoursJ. Calf lens alpha-crystallin, a molecular chaper- biological activities. In: Hepatocyte Growth Factor-Scatter one, builds stable complexes with beta s- and gamma Factor (HGF-SFand C-met Receptor. Goldberg ID, Rosen crystallins. Ophthalmic Res. 1996;28(suppl):23-31. EM, eds. Basel: Birkhauser Verlag; 1993:67-79.

Downloaded from iovs.arvojournals.org on 09/26/2021