Hepatocyte growth factor (HGF), a heparin-binding polypeptide mitogen, stimulates DNA synthesis in adult rat and human hepatocytes and in several other cells of epithelial origin. Recently, it was determined that scatter factor (SF), a protein that has been shown to cause the dispersion and migration of epithelial cells in culture, is identical to HGF. Moreover, the receptor for HGF was identified as the product of the proto-onco-gene, c-MET, a tyrosine kinase-containing transmem-brane protein. c-MET expression has been reported in a variety of adult and embryonic mouse tissues. Simi-larly, we and others have demonstrated that HGF is expressed in various adult rat and human tissues. In the present study, the tissue distribution of HGF during rat development was determined by immunohistochemistry using an HGF-specific polyclonal antiserum. Between day 12 and day 19, immunoreactivity for HGF was present in various locations such as hematopoietic cells, somites, squamous epithelium of the esophagus and skin, periventricular germinal matrix of the brain, bronchial epithelium, renal collecting tubules and chon-drocytes. After day 19, HGF immunoreactivity was also present in the pancreas, submaxillary glands and neural tissues. In addition to immunolocalizing HGF in tissue sections, bioreactive and immunoreactive HGF was extracted and purified from rat fetuses. Other studies demonstrated the presence of HGF and c-MET mRNA in total fetal rat, and in fetal and neonatal rat liver. Addition of purified HGF to fetal and neonatal rat liver cultures enriched for hepatocytes stimulated DNA syn-thesis up to six-fold over controls. These findings strongly suggest a pivotal role for this potent regulator of growth and development.

Hepatocyte growth factor (HGF) or scatter factor (SF) has the ability to elicit a variety of responses in cultured cells, especially cells of epithelial origin. These responses include mitogenesis (Nakamura et al., 1987; Gohda et al., 1988; Zarnegar and Michalopoulos, 1989), motogenesis (Stoker et al., 1987) and morphogenesis (Stern et al., 1990; and Montesano et al., 1991). Hepatocyte growth factor was first purified to homogeneity in the mid-1980s from the plasma and serum of rats, rabbits and humans on the basis of stim-ulating DNA synthesis in rat and human hepatocytes (Naka-mura et al., 1987; Gohda et al., 1988; Zarnegar and Michalopoulos, 1989). Scatter factor, meanwhile, was inde-pendently characterized by Stoker et al. (1987) due to its ability to disperse contiguous sheets of epithelial cells in culture. Recently, sequencing of cDNA clones of the two proteins proved that these two factors are indeed identical and that they share the same receptor, c-MET (Weidner et al., 1991; and Naldini et al., 1991b). With heterodimeric subunits of 70,000 and 35,000 (Mr), HGF is the largest known growth factor presently described in the literature (Nakamura et al., 1987; Gohda et al., 1988; Zarnegar and Michalopoulos, 1989).

HGF’s role in vivo is still obscure at this point; recent evidence, however, suggests HGF may be involved in cel-lular functions such as wound repair, organ growth and regeneration as in the liver (Kinoshita et al., 1989; Okajima et al., 1990; Noji et al., 1990; Lindroos et al., 1991; Shir-macher et al., 1992) and kidney (Tashiro et al., 1990; Oka-jima et al., 1990; Noji et al., 1990; Wolf et al., 1991b; Nagaike et al., 1991; Ishibashi et al., 1992), and develop-ment (Selden et al., 1990 and Stern et al., 1990). Of spe-cial interest are studies conducted by Stern et al. (1990) who implanted inert beads carrying HGF into developing chick embryos and found gross anomalies in the architec-ture of the primitive streak near the site of application. Other investigators have also linked HGF to developmen-tal processes. Selden et al. (1990) demonstrated that HGF mRNA is present in fetal human liver, while Rosen et al. (1990b) have demonstrated the presence of HGF in human amniotic fluid. Furthermore, others have shown that HGF protein (Rosen et al., 1990b; Wolf et al., 1991a) and its mRNA (Miyazawa et al., 1989 and Zarnegar et al., 1991) are present in human placenta while the protein has been localized in rat placenta (Wolf et al., 1991b). In fact, a human placental cDNA library was initially used to isolate the human HGF cDNA (Miyazawa et al., 1989). These find-ings prompted us to examine HGF’s cellular distribution in rat embryos.

HGF extraction and purification from adult rat serum

Extraction and purification of HGF from adult rat serum (Pel-Freez Biological, Rogers, AR) was carried out according to Zarnegar and Michalopoulos, 1989.

Preparation of chicken anti-rabbit HGF polyclonal antibody

Polyclonal antibodies to rabbit HGF were prepared in a chicken as described by Zarnegar et al. (1990). These antibodies were used for immunohistochemical and ELISA studies.

Rat embryo preparation for immunohistochemistry

Several timed pregnant F-344 rats purchased from Frederick Cancer Research and Development Center (Frederick, MD) were killed at various time points throughout pregnancy (days 12 through delivery) under ether anesthesia, and their uteruses were removed. The maternal tissues and placenta were separated from each embryo prior to placing the entire embryo in a formalin solu-tion of 3.7% paraformaldehyde buffered with phosphate-buffered saline (PBS) for 24 hours. Using a 27-gauge syringe, formalin was injected into the gut and thoracic cavities of larger fetuses (day 18 through birth) to facilitate fixation. The fixed embryos were then cut and routinely processed through a series of graded ethanol and xylene and embedded in paraffin. Sections were cut at 4 μm and mounted on slides using Histostik (Accurate Chemical and Scientific Corporation, Westbury, NY). After air drying, slides were used for immunohistochemistry.

Immunohistochemistry

Immunohistochemistry was performed as reported by Wolf et al. (1991a, b) with only slight modifications. Previously, we have shown that anti-HGF antiserum incubated with purified rabbit HGF for one hour prior to being used as the primary antibody in immunohistochemical studies abrogates all tissue staining (Wolf et al., 1991b). Briefly, slides were deparaffinized with xylene and placed in ethanol (100%, 95%). Slides were then incubated in a 1.875% solution of H2O2 in methanol for 30 minutes at room tem-perature, followed by rehydration in a series of graded ethanol and deionized water. Incubation with a solution of 0.1% trypsin [w/v] (Sigma Chemical Company, St Louis, MO) and 0.1% CaCl2 [w/v] in 0.05 M Tris (pH 7.6) was carried out for 15 minutes at room temperature. Slides were blocked using a 10% solution of goat serum in PBS for 20 minutes at room temperature; the goat serum solution was then removed. Primary polyclonal antiserum against rabbit HGF made in a chicken was prepared at a dilution of 1:600 in 1% BSA (bovine serum albumin) in PBS, and approx-imately 250 μl was used per slide. Incubation at room tempera-ture took place for 90 minutes. Slides were then rinsed 3 × 5 min-utes with PBS. Secondary antiserum consisting of biotinylated goat anti-chicken antibodies (Vector Laboratories, Burlingame, CA) was prepared in 1% BSA in PBS at a dilution of 1:200 and applied to the slides for 30 minutes at room temperature. The slides were rinsed 3 × 5 minutes with PBS. A biotin-avidin ‘com-plex solution’ consisting of 0.625% solution of avidin DH and biotinylated horseradish peroxidase H (Vector Laboratories, Burlingame, CA) was then prepared and applied to the slides for 30 minutes at room temperature. Slides were rinsed with PBS as mentioned above. Next, the sections were incubated for 2 min-utes with a solution of DAB (diaminobenzadine -0.5 mg/ml) in 0.05 M Tris buffer (pH 7.6) with the addition of approximately 20 μl/ml of 30% H2O2 (0.6% H 2O2) for 5 minutes. The sections were rinsed in deionized water, washed with tap water for 10 min-utes, stained with hematoxylin, dehydrated in ethanol, and embed-ded in Permount (Fisher Scientific, Pittsburgh, PA) or S/P Accu•Mount 60 mounting medium (Baxter Healthcare Corporation, McGaw Park, IL). Staining intensity was graded on a scale of – = no staining to +++ = very intense staining.

HGF extraction and purification from day 19 rat embryos

Nine F-344 timed pregnant (day 15) rats purchased from Frederick Cancer Research and Development Center (Frederick, MD) were killed at day 19 of gestation under ether anesthesia, and their uteruses were removed. The embryos were carefully separated from maternal tissues and placenta, rinsed with PBS (phosphate-buffered saline) to remove any maternal blood, homogenized in two volumes of ice-cold 1 M NaCl for ninety seconds using a Polytron tissue homogenizer (Brinkman Instruments, Des Plaines, IL) in a total volume of 400 ml, and cleared by centrifugation at 15,000 g for 30 minutes. The amount of tissue used was approx-imately 90 grams (1 embryo/gram) wet weight. The NaCl con-centration of the extract was adjusted to 0.4 M, loaded onto a heparin-agarose column (80 ml total bed volume) that had been equilibrated with 0.4 M NaCl. The column was washed with approximately 4 l of 0.4 M NaCl and eluted with 1 l of 1.4 M NaCl. The volume of the eluate was then reduced to 50 ml by ultrafiltration (YM-10 filter, Amicon, Danvers, MA). The NaCl concentration was adjusted to 0.6 M NaCl and loaded onto a TSK heparin-agarose column (7.5 mm × 7.5 mm, Supelco, Bellefonte, PA) which had been equilibrated with 0.6 M NaCl. The column had previously been attached to a high-pressure liquid chro-matography pump (8800 series, Du Pont, Wilmington, DE). The column was then washed with 200 ml of 0.6 M NaCl. The sample was then eluted using a gradient of 0.6 M NaCl to 2.0 M NaCl in 110 minutes at a flow rate of 1 ml/minute. 20 μl of each frac-tion (2 ml/fraction) was checked for bioactivity on primary cul-tures of adult male rat hepatocytes based on [3H]thymidine incor-poration as described previously (Zarnegar and Michalopoulos, 1989). Bioactive fractions were pooled, and purity of bioactive and immunoreactive HGF was determined by SDS-PAGE. ELISA reactivity was then determined using the polyclonal chicken anti-HGF serum, as described previously (Zarnegar et al., 1990).

RNA extraction from rat fetuses and neonates

Day 19 rat fetuses were removed as described above. From other embryos (days 14, 15, 18 and 20) and neonates (day 7 after birth), livers were carefully excised using forceps and microscissors. Older fetuses (day 20) and neonates were euthanized by lethal injection of Nembutal sodium solution (Abbott Laboratories, North Chicago, IL) prior to extraction of their livers. Total RNA was then extracted using RNAzol B solution according to the man-ufacturer’s recommendations (Biotecx Laboratories, Inc., Hous-ton, TX). Each sample’s nucleotide:protein ratio was determined at absorbances 260 nm and 280 nm. All samples had at least a ratio of 1.8, and integrity was ascertained by 1% agarose/ethid-ium bromide gel.

Reverse transcriptase-polymerase chain reaction (RT-PCR)

For each reaction, 1 μg of total RNA from all samples (total embryo RNA, and embryonic and neonatal liver RNA) was reverse transcribed to cDNA using the GeneAmp preamplification system (Perkin Elmer Cetus, Norwalk, CT). The resulting cDNA

was then subjected to 40 cycles of polymerase chain reaction (PCR) using a DNA Thermal Cycler 480 (Perkin Elmer Cetus, Norwalk, CT), ‘AmpliTaq’ DNA polymerase (Perkin Elmer Cetus, Norwalk, CT), and primers specific for rat HGF, c-MET or actin. The rat-specific set of primers for HGF, which give an amplified fragment of 691 base pairs, consisted of a forward primer begin-ning at base pair 710 with nucleotide sequence 5′ATCAGA-CACCACACCGGCACAAAT3′ and a reverse primer beginning at base pair 1399 with nucleotide sequence 5′GAAATAGGGCAATAATCCCAAGGAA3′. The c-MET specific primers consisted of a forward primer of 5′TGACGCA-AGACTACACACTC3′ and a reverse primer of 5′TCAGTCA-GAAACTGGGAGACCTCT3′ giving a fragment length of 467 base pairs. Human ′-actin primers purchased from Clontech Lab-oratories, Inc. (Palo Alto, CA) were expected to give a fragment length of approximately 1000 base pairs. These primers were included to ensure that RNA from each sample was intact. After amplification and extraction using a one-step chloroform proce-dure, each sample was then applied to a 1% agarose/ethidium bro-mide gel at 100 V. The resulting gel was photographed with UV illumination. No bands were present when reverse transcriptase was omitted from the reaction indicating that amplified bands were not due to contamination of RNA samples with genomic DNA. Moreover, when genomic placental DNA was used as the tem-plate with the primers described above in the PCR reaction, no bands were observed (due to the presence of introns) again con-firming that amplified bands are due to the HGF cDNA.

Southern blot hybridization and restriction enzyme analysis of PCR-generated fragments

Southern hybridization was carried out on agarose gels contain-ing PCR-generated fragments to confirm the presence of HGF-specific fragments. Briefly, gels were denatured in 1.0 M NaCl and 0.5 M NaOH (pH 8.0) for 30 minutes, neutralized twice for 15 minutes each in 1.0 M NaCl and 0.5 M Tris (pH 7.2), and transferred overnight to nitrocellulose (Schleicher & Schuell, Inc., Keene, NH) using 10 × SSC (20 × SSC: 3.0 M NaCl, 0.3 M sodium citrate (pH 7.0)). DNA was then UV cross-linked to the nitrocellulose. Blots were prehybridized for 4 hours at 65°C in 6 × SSC, 5 × Denhardt’s solution, 100 μg/ml denatured salmon sperm DNA, and 0.1% SDS. Meanwhile, a full-length cDNA probe for human HGF was multiprime labelled (Amersham Corporation, Arlington Heights, IL) using α[32P]dCTP. Fresh hybridization solution was used when the labelled probe was added. Hybridization was carried out overnight at 65°C. On the following day, blots were washed once in 2 × SSC and 0.5% SDS for 15 minutes at 65°C, once in 1 × SSC and 1.0% SDS for 15 minutes at 65°C and twice in 0.1 × SSC and 1.0% SDS for 15 minutes each at 65°C. The blots were exposed to XAR [X-OMAT] film by Eastman Kodak Company (Rochester, NY) overnight. In addition to Southern blot hybridization, HGF-specific PCR-gen-erated fragments were subjected to restriction enzyme analysis using EcoRI restriction endonuclease (Gibco/BRL, Bethesda, MD). Approximately 10 μl of DNA from the RT-PCR reaction was incubated with 10 units of EcoRI and appropriate buffer for one hour at 37°C after which the DNA was run on a 1% agarose/ethidium bromide gel. EcoRI was expected to digest the PCR-generated fragment once resulting in two smaller fragments.

Preparation of fetal and neonatal rat liver cultures enriched for hepatocytes

Neonatal rat hepatocytes (day 14) were prepared with only a slight modification of the procedure described for adult rat hepatocytes (Zarnegar and Michalopoulos, 1989). However, preparation of fetal rat hepatocytes was performed as follows. Rat fetuses of var-ious gestational age (15, 19, and 20 days) were separated from maternal tissues as mentioned above. Livers were carefully excised and immediately rinsed several times in ice-cold sterile Minimum Essential Media (MEM) containing non-essential amino acids (Gibco/BRL, Bethesda, MD). After the final wash, livers were placed in approximately 10 ml of ice-cold MEM and finely minced with a scalpel. Meanwhile, 3 mg/ml collagenase H (Boehringer-Mannheim, Indianapolis, IN) and sterile CaCl2 solu-tion (at a final concentration of 4.3 mM) were added to 30 ml of 37°C MEM. Liver pieces were added to 15 ml of the 37°C col-lagenase solution in a sterile 100 ml beaker; the beaker contain-ing the mixture was then placed on a stir plate with a sterilized magnetic stir bar for 10 minutes at 37°C. After stirring, pieces were pipeted gently to facilitate dispersion of cells, and the super-natant containing dispersed cells was then added to an equal volume of ice-cold MEM. To the remaining pieces, another 15 ml of 37°C collagenase solution was added, and stirring continued for an additional 10 minutes. Once again, the pieces were dis-persed through pipeting, and the supernatant was added to an equal volume of ice-cold MEM. Cells were then passed through a mesh screen with pore diameter of 100 μm, added to a fresh 50 ml ster-ile conical tube and pelleted through centrifugation at 90 g for 5 minutes. After decanting the supernatant, remaining cells were resuspended in 5 ml of ice-cold MEM and allowed to sediment on ice for 20 minutes; this was repeated once more. Both fetal and neonatal cells were plated at approximately 200,000 cells per well on collagen-coated Falcon 6-well, flat-bottom plates (Becton-Dickinson, Lincoln Park, NJ) as described for adult rat hepato-cytes (Zarnegar and Michalopoulos, 1989). Approximately 85-90% of the fetal cells and greater than 95% of the neonatal cells were hepatocytes when morphology was examined. Typical mor-phology included characteristically clustered polygonal cells with pronounced, rounded nuclei and grainy cytoplasm. After a 48 hour incubation at 37°C, cells were precipitated with 5% trichloroacetic acid, and precipitated materials were counted in a ′-scintillation counter.

Recognition of rat HGF by anti-rabbit HGF antiserum

To investigate whether chicken anti-rabbit HGF antiserum cross-reacts with rat HGF, we purified HGF from adult rat serum and determined by ELISA that adult rat HGF is immunoreactive (Fig. 1). Therefore, chicken anti-rabbit HGF antiserum was used to immunolocalize HGF in rat embryos of various gestational age. Previously, we have shown that anti-HGF antiserum incubated with purified rabbit HGF for one hour prior to being used as the primary antibody in immunohistochemical studies abrogates all tissue staining (Wolf et al., 1991b).

Fig. 1.

Recognition of rat HGF by chicken anti-rabbit HGF polyclonal antibody. To investigate whether chicken anti-rabbit HGF antiserum cross-reacts with rat HGF, we purified HGF from adult rat serum and determined by ELISA that adult rat HGF is indeed immunoreactive as described in Materials and methods. Various concentrations of rat HGF (shown as solid bars) as compared to the same concentrations of rabbit HGF (shown as striped bars) used to prepare the antibody were tested for ELISA immunoreactivity at 492 nm. Fig. 1 shows that rat and rabbit HGF are comparably recognized by the chicken anti-rabbit HGF polyclonal antibody. Therefore, chicken anti-rabbit HGF antiserum was used to immunolocalize HGF in rat embryos of various gestation.

Fig. 1.

Recognition of rat HGF by chicken anti-rabbit HGF polyclonal antibody. To investigate whether chicken anti-rabbit HGF antiserum cross-reacts with rat HGF, we purified HGF from adult rat serum and determined by ELISA that adult rat HGF is indeed immunoreactive as described in Materials and methods. Various concentrations of rat HGF (shown as solid bars) as compared to the same concentrations of rabbit HGF (shown as striped bars) used to prepare the antibody were tested for ELISA immunoreactivity at 492 nm. Fig. 1 shows that rat and rabbit HGF are comparably recognized by the chicken anti-rabbit HGF polyclonal antibody. Therefore, chicken anti-rabbit HGF antiserum was used to immunolocalize HGF in rat embryos of various gestation.

Immunohistochemistry

Immunohistochemical studies using the polyclonal anti-serum demonstrated that HGF is present in all three germ layers in the developing rat (Fig. 2A—day 20 rat embryo stained with chicken anti-rabbit HGF antiserum; Fig. 2B—day 20 rat embryo stained with chicken non-immune serum). Whenever HGF immunoreactivity was present, it showed a diffuse cytoplasmic pattern. Strong staining for HGF at day 12 was evident in the somites (Fig. 3A). In addition, hematopoietic cells were positive at day 12 and remained positive throughout gestation (Fig. 3I, J—day 20 rat fetus stained with chicken anti-rabbit HGF antiserum or chicken non-immune serum, respectively). Positive staining was also visible in the neural ectoderm, aortic arches and the squamous epithelium of the esophagus. On day 14, the squamous epithelium of the esophagus and skin, and the columnar epithelium of the bronchi appeared to stain more strongly and remained intense throughout gestation. In the lung, macrophages appeared positive, also. The somites that stained strongly positive at day 12 were weakly stained by day 14. Strong staining of the periventricular germinal matrix of the brain (Fig. 3B) and the collecting tubules of the kidney (Fig. 3E, F) for immunoreactive HGF was evi-dent at day 17, while chondrocytes stained weakly from day 17 throughout gestation. The pancreas (Fig. 3H) and the submaxillary all stained strongly positive at day 19 and remained positive during gestation. Muscle, bone, fibrous tissue and adipose tissue were consistently negative for HGF. The immunohistochemical results are compiled in Table 1.

Table 1.

Immunohistolocalization of HGF in rat fetuses

Immunohistolocalization of HGF in rat fetuses
Immunohistolocalization of HGF in rat fetuses

HGF purification from day 19 rat embryos

An extract of 90 embryos that were at 19 days of gestation was fractionated by heparin-agarose affinity chromatogra-phy as described in Materials and methods. As shown in Fig. 4, two major peaks and one minor peak of growth-pro-moting activity (as determined by DNA incorporation of [3H]thymidine in adult rat hepatocytes in primary culture) were eluted at approximately 0.9 M, 1.3 M and 1.2 M NaCl, respectively. However, only the first peak at 0.9 M NaCl was immunoreactive with chicken anti-rabbit HGF anti-serum shown by dotted bars in Fig. 4. HGF extracted from adult rabbit and human placenta elutes from a heparin-agarose column between 0.9 M and 1 M NaCl (Zarnegar and Michalopoulos, 1989). Since the immunoreactive and bioactive fractions of the embryo extract elute from the heparin-agarose column at nearly the same NaCl concen-tration (0.9 M NaCl), it strongly suggests that this peak con-tained HGF. The identities of the growth-promoting activ-ities contained in the other peaks have not been investigated at this point. The purity of the embryonic HGF from the 0.9 M peak was analyzed by SDS-PAGE under non-reduc-ing conditions and stained with silver. Only one band with Mr of approximately 80,000 was evident (data not shown).

Fig. 2.

(A, B) Immunohistochemical analysis of a day 20 rat fetus using chicken anti-rabbit HGF antiserum. A shows a sagittal section of a day 20 rat fetus stained with chicken anti-rabbit HGF antiserum and B shows an identical section stained with non-immune serum.

Fig. 2.

(A, B) Immunohistochemical analysis of a day 20 rat fetus using chicken anti-rabbit HGF antiserum. A shows a sagittal section of a day 20 rat fetus stained with chicken anti-rabbit HGF antiserum and B shows an identical section stained with non-immune serum.

Fig. 3.

Immunohistochemical analysis of embryonic rat tissues of various gestational age using either a chicken anti-rabbit HGF antiserum or chicken non-immune serum. (A) The somites of a day 12 rat embryo stained with anti-rabbit HGF antiserum. Staining for HGF is seen throughout the somite. (B) The periventricular germinal matrix of the brain of a day 17 rat fetus stained with anti-rabbit HGF antiserum. Extremely intense staining is noted in this tissue. (C) The eye of a day 17 rat fetus stained with anti-rabbit HGF antiserum. (D) Columnar epithelium in the lung of a day 17 rat fetus stained with anti-rabbit HGF antiserum. Strong staining for HGF is indicated by an arrow. (E) Kidney of a day 17 rat fetus stained with anti-rabbit HGF antiserum. (F) Kidney of a day 17 rat fetus stained with non-immune serum. No appreciable staining is observed.===(G) Nose whisker follicle stained with anti-rabbit HGF antiserum. (H) Pancreas of a day 19 rat fetus stained with anti-rabbit HGF antiserum. Note intense staining in the acinar cells. (I) The liver of a day 20 rat embryo stained with anti-rabbit HGF antiserum. Arrows indicate the hematopoietic cells which stain strongly for the presence of HGF. (J) The liver of a day 20 rat fetus stained with chicken non-immune serum. No staining is observed in the hematopoietic cells indicated by the arrow. (K) Epidermis of a day 20 rat fetus stained with anti-rabbit HGF antiserum. Note the intense staining in the apical portion of the epidermis and the complete absence of staining in the hair follicle. (L) Epidermis of a day 20 rat fetus stained with chicken non-immune serum. Previously, we have shown that anti-HGF antiserum incubated with purified rabbit HGF for 1 hour prior to being used as the primary antibody in immunohistochemical studies abrogates all tissue staining (Wolf et al., 1991b)

Fig. 3.

Immunohistochemical analysis of embryonic rat tissues of various gestational age using either a chicken anti-rabbit HGF antiserum or chicken non-immune serum. (A) The somites of a day 12 rat embryo stained with anti-rabbit HGF antiserum. Staining for HGF is seen throughout the somite. (B) The periventricular germinal matrix of the brain of a day 17 rat fetus stained with anti-rabbit HGF antiserum. Extremely intense staining is noted in this tissue. (C) The eye of a day 17 rat fetus stained with anti-rabbit HGF antiserum. (D) Columnar epithelium in the lung of a day 17 rat fetus stained with anti-rabbit HGF antiserum. Strong staining for HGF is indicated by an arrow. (E) Kidney of a day 17 rat fetus stained with anti-rabbit HGF antiserum. (F) Kidney of a day 17 rat fetus stained with non-immune serum. No appreciable staining is observed.===(G) Nose whisker follicle stained with anti-rabbit HGF antiserum. (H) Pancreas of a day 19 rat fetus stained with anti-rabbit HGF antiserum. Note intense staining in the acinar cells. (I) The liver of a day 20 rat embryo stained with anti-rabbit HGF antiserum. Arrows indicate the hematopoietic cells which stain strongly for the presence of HGF. (J) The liver of a day 20 rat fetus stained with chicken non-immune serum. No staining is observed in the hematopoietic cells indicated by the arrow. (K) Epidermis of a day 20 rat fetus stained with anti-rabbit HGF antiserum. Note the intense staining in the apical portion of the epidermis and the complete absence of staining in the hair follicle. (L) Epidermis of a day 20 rat fetus stained with chicken non-immune serum. Previously, we have shown that anti-HGF antiserum incubated with purified rabbit HGF for 1 hour prior to being used as the primary antibody in immunohistochemical studies abrogates all tissue staining (Wolf et al., 1991b)

Fig. 4.

Elution profile, stimulation of DNA synthesis in adult rat hepatocytes and ELISA immunoreactivity of embryonic rat HGF. Embryonic rat HGF was purified from approximately 90 day 19 rat embryos by two rounds of heparin-affinity chromatography as described in Materials and methods. The elution profile of HGF from a TSK heparin-agarose column (7.5 mm × 7.5 mm, Supelco, Bellefonte, PA) indicated by the solid line was measured at absorbance 280 nm. The sample was eluted using a gradient of 0.6 M NaCl to 2.0 M NaCl in 110 minutes at a flow rate of 1 ml/minute which is shown as the dashed line. 20 μl of each fraction (2 ml/fraction) was checked for bioactivity on primary cultures of adult male hepatocytes based on [3H]thymidine incorporation (disintegrations per minute [DPM] × 10−4/105 cells) as has been described previously (Zarnegar and Michalopoulos, 1989). The stimulation of DNA synthesis is represented by the dotted line. Two major peaks (at approximately fraction numbers 12 and 32, respectively) and one minor peak (at approximately fraction number 23) of growth promoting activity were detected by [3H]thymidine incorporation. Fractions corresponding to the peaks of DNA synthesis were pooled and subjected to ELISA using the polyclonal antibody described previously. ELISA showed that only one of the peaks was immunoreactive, namely the peak corresponding to fraction number 12. The dotted bars indicate relative ELISA immunoreactivity which was measured at an absorbance of 492 nm. Pooled fractions corresponding to the three peaks were also subjected to SDS-PAGE under non-reducing conditions, and the resulting gel was silver stained. SDS-PAGE revealed that only the pooled fractions corresponding to the peak at approximately fraction number 12 had a migration pattern characteristic of HGF: only a single band visible at an Mr of 80,000 (data not shown).

Fig. 4.

Elution profile, stimulation of DNA synthesis in adult rat hepatocytes and ELISA immunoreactivity of embryonic rat HGF. Embryonic rat HGF was purified from approximately 90 day 19 rat embryos by two rounds of heparin-affinity chromatography as described in Materials and methods. The elution profile of HGF from a TSK heparin-agarose column (7.5 mm × 7.5 mm, Supelco, Bellefonte, PA) indicated by the solid line was measured at absorbance 280 nm. The sample was eluted using a gradient of 0.6 M NaCl to 2.0 M NaCl in 110 minutes at a flow rate of 1 ml/minute which is shown as the dashed line. 20 μl of each fraction (2 ml/fraction) was checked for bioactivity on primary cultures of adult male hepatocytes based on [3H]thymidine incorporation (disintegrations per minute [DPM] × 10−4/105 cells) as has been described previously (Zarnegar and Michalopoulos, 1989). The stimulation of DNA synthesis is represented by the dotted line. Two major peaks (at approximately fraction numbers 12 and 32, respectively) and one minor peak (at approximately fraction number 23) of growth promoting activity were detected by [3H]thymidine incorporation. Fractions corresponding to the peaks of DNA synthesis were pooled and subjected to ELISA using the polyclonal antibody described previously. ELISA showed that only one of the peaks was immunoreactive, namely the peak corresponding to fraction number 12. The dotted bars indicate relative ELISA immunoreactivity which was measured at an absorbance of 492 nm. Pooled fractions corresponding to the three peaks were also subjected to SDS-PAGE under non-reducing conditions, and the resulting gel was silver stained. SDS-PAGE revealed that only the pooled fractions corresponding to the peak at approximately fraction number 12 had a migration pattern characteristic of HGF: only a single band visible at an Mr of 80,000 (data not shown).

HGF and cMET mRNA expression in fetal and neonatal rat RNA using the reverse transcriptase-polymerase chain reaction

We examined total embryonic RNA, RNA from livers of fetuses (days 14, 15, 18 and 20), and RNA from neonatal rat liver (day 7 after birth) by RT-PCR in an effort to demonstrate that the HGF and c-MET genes are both expressed during rat development. The expression of HGF mRNA indirectly supports that the HGF protein observed in immunohistological studies or purified from rat fetuses is the result of expression of the HGF gene in the embryo proper. Since HGF mRNA is suspected of existing in rel-atively low-abundance and since isolation of large quanti-ties of fetal tissue is often difficult, we took an alternate approach, namely the polymerase chain reaction, to detect the HGF transcript. Total RNA extraction from rat embryos and placenta, reverse transcription, and PCR were carried out as explained in Materials and methods. After PCR prod-ucts were fractionated on a 1% agarose/ethidium bromide gel, expected amplified fragments of 691 base pairs for HGF, 467 base pairs for c-MET, and approximately 1000 base pairs for ′-actin were observed. Southern blot analy-sis of the gel was carried out using a 32P multiprime-labelled human HGF cDNA as a probe as described in Materials and methods; it showed hybridization to the 691 base pair fragments of all samples in which HGF primers were used (Fig. 5A—lanes B, E, H, K, N and Q). How-ever, hybridization was unexpectedly seen with other frag-ments within the same lanes that were not visible on the 1% agarose/ethidium bromide gel by UV illumination (Fig. 5B—lanes B, E, H, K, N and Q). Characterization of the additional fragments is underway presently.

Fig. 5.

(A, B) Analysis of fetal rat RNA by Reverse-Transcriptase PCR and subsequent Southern blot analysis of HGF-specific DNA fragments generated by RT-PCR. (A) Lanes B, C and D show day 14 fetal rat liver mRNA subjected to RT-PCR using either HGF-specific primers, c-MET specific primers, or ′-actin specific primers, respectively, as described in Materials and methods. Lanes E, F, and G show day 15 fetal rat liver mRNA subjected to RT-PCR using either HGF-specific primers, c-MET specific primers, or ′-actin specific primers, respectively. Lanes H, I, and J show day 18 fetal rat liver mRNA subjected to RT-PCR using either HGF-specific primers, c-MET specific primers, or ′-actin specific primers, respectively. Lanes K, L, and M show day 20 fetal rat liver mRNA subjected to RT-PCR using either HGF-specific primers, c-MET specific primers, or ′-actin specific primers, respectively. Lanes N, O, and P show day 7 neonate rat liver mRNA subjected to RT-PCR using either HGF-specific primers, c-MET specific primers, or ′-actin specific primers, respectively. Lanes Q, R, and S show total day 19 fetal rat mRNA subjected to RT-PCR using either HGF-specific primers, c-MET specific primers, or ′-actin specific primers, respectively. Lanes A and T represent molecular size standards. The size of each marker is indicated at the far left. Expected fragment lengths are as follows: HGF = 691 bp; c-MET = 497 bp; and ′-actin = ~1000 bp. (B) Southern blot hybridization using a human HGF full-length cDNA as probe was performed on the 1% agarose/ethidium bromide gel in A. One intensely hybridizing band is noted at approximately 691 bp in lanes B, E, H, K, N, and Q. These lanes contain HGF-specific RT-PCR generated fragments. Other hybridizing bands are also noted in these lanes which may possibly represent alternatively spliced forms of HGF mRNA as described under Results. Note that no non-specific hybridization to fragments generated from RT-PCR using either c-MET specific or ′-actin primers or to molecular size standards is seen (lanes A, C, D, F, G, I, J, L, M, O, P, R, S, or T).

Fig. 5.

(A, B) Analysis of fetal rat RNA by Reverse-Transcriptase PCR and subsequent Southern blot analysis of HGF-specific DNA fragments generated by RT-PCR. (A) Lanes B, C and D show day 14 fetal rat liver mRNA subjected to RT-PCR using either HGF-specific primers, c-MET specific primers, or ′-actin specific primers, respectively, as described in Materials and methods. Lanes E, F, and G show day 15 fetal rat liver mRNA subjected to RT-PCR using either HGF-specific primers, c-MET specific primers, or ′-actin specific primers, respectively. Lanes H, I, and J show day 18 fetal rat liver mRNA subjected to RT-PCR using either HGF-specific primers, c-MET specific primers, or ′-actin specific primers, respectively. Lanes K, L, and M show day 20 fetal rat liver mRNA subjected to RT-PCR using either HGF-specific primers, c-MET specific primers, or ′-actin specific primers, respectively. Lanes N, O, and P show day 7 neonate rat liver mRNA subjected to RT-PCR using either HGF-specific primers, c-MET specific primers, or ′-actin specific primers, respectively. Lanes Q, R, and S show total day 19 fetal rat mRNA subjected to RT-PCR using either HGF-specific primers, c-MET specific primers, or ′-actin specific primers, respectively. Lanes A and T represent molecular size standards. The size of each marker is indicated at the far left. Expected fragment lengths are as follows: HGF = 691 bp; c-MET = 497 bp; and ′-actin = ~1000 bp. (B) Southern blot hybridization using a human HGF full-length cDNA as probe was performed on the 1% agarose/ethidium bromide gel in A. One intensely hybridizing band is noted at approximately 691 bp in lanes B, E, H, K, N, and Q. These lanes contain HGF-specific RT-PCR generated fragments. Other hybridizing bands are also noted in these lanes which may possibly represent alternatively spliced forms of HGF mRNA as described under Results. Note that no non-specific hybridization to fragments generated from RT-PCR using either c-MET specific or ′-actin primers or to molecular size standards is seen (lanes A, C, D, F, G, I, J, L, M, O, P, R, S, or T).

These additional fragments may represent alternatively spliced forms of HGF mRNA. For example, Seki et al. (1990) screened a human leukocyte cDNA library to iso-late clones for HGF and revealed a clone that contained a 15 base pair in-frame deletion. When this cDNA was tran-siently expressed in COS-1 cells and the resulting protein isolated, it was found that the protein containing the dele-tion was equally bioactive when compared to HGF. Seki et al. (1990) attribute this deletion to alternative splicing. Additional HGF mRNAs have been observed (Selden et al., 1990; Rubin et al., 1991; and Zarnegar et al., 1991), and HGF cDNAs have been isolated from human placental (Miyazawa et al., 1991) and fetal fibroblast cell line (M426) libraries (Chan et al., 1991). Three different sized cDNAs (6.0 kb, 3.0 kb and 1.3 kb) were observed (Chan et al., 1991). The two larger cDNAs encoded full-length HGF molecules; the smallest transcript, however, resulted from an alternative splicing event and encoded a truncated ver-sion of HGF with Mr of 28,000 which bound to the HGF receptor, c-MET, competed with native HGF for binding to c-MET, but did not appear to have stimulatory activity (Chan et al., 1991). Chan et al. (1991) demonstrated that the expression levels of each of the HGF transcripts varied depending upon from which tissue or cell line mRNA was extracted. The highest levels of 1.3 kb mRNA transcript were observed in human foreskin fibroblasts as compared to human placenta or other fibroblast types. Perhaps this variant of HGF acts as an antagonist of native HGF in normal cellular functions and modulates one or all of HGF’s potent growth-, motility-, morphogenic-potentiating activi-ties. The differential expression of the HGF variant during embryogenesis deserves further investigation.

HGF-stimulated [3H]thymidine incorporation in DNA from rat embryonic and neonatal liver cultures enriched for hepatocytes

HGF is a known mitogen for adult rat and human hepato-cytes in culture; thus, we wished to examine whether HGF stimulates DNA synthesis in fetal liver cultures enriched for hepatocytes. We found that fetal liver cultures enriched for hepatocytes (days 15, 19 and 20) and neonatal hepato-cytes (day 14 after birth) respond in a dose-dependent manner to HGF with increased DNA synthesis as deter-mined by [3H]thymidine incorporation (Fig. 6). Fold increase in [3H]thymidine incorporation in trichloroacetic acid precipitable materials was determined showing that HGF at 15 ng/ml stimulated anywhere from 2.3 to 6.3-fold increase in fetal and neonatal hepatocyte cultures (Fig. 6). Similar results were obtained when epidermal growth factor was added to cultures as a positive control (data not shown).

Fig. 6.

Stimulation of DNA synthesis in fetal liver cultures enriched for hepatocytes (days 15, 19, and 20) or neonatal rat hepatocytes (day 14) by adult rabbit HGF. Fetal liver cultures enriched for hepatocytes (days 15, 19, and 20) or neonatal rat hepatocytes (day 14) were prepared as described in Materials and methods. Adult rabbit HGF was added to duplicate cultures at various concentrations. DNA synthesis was determined by [3H]thymidine incorporation into trichloroacetic acid-precipitable materials and is represented in Fig. 6 as fold increase over control. (Control is no addition of growth factor.) A dose-dependent stimulation of DNA synthesis is noted in cultures from the various time points with a maximum stimulation in nearly all cultures at an HGF concentration of 15 ng. Fetal day 15 (F-DAY 15) is indicated by light dotted bars; fetal day 19 (F-DAY 19) is indicated by black bars; fetal day 20 (F-DAY 20) is indicated by diagonally striped bars; and neonatal day 14 (N-DAY 14) is represented by horizontally striped bars.

Fig. 6.

Stimulation of DNA synthesis in fetal liver cultures enriched for hepatocytes (days 15, 19, and 20) or neonatal rat hepatocytes (day 14) by adult rabbit HGF. Fetal liver cultures enriched for hepatocytes (days 15, 19, and 20) or neonatal rat hepatocytes (day 14) were prepared as described in Materials and methods. Adult rabbit HGF was added to duplicate cultures at various concentrations. DNA synthesis was determined by [3H]thymidine incorporation into trichloroacetic acid-precipitable materials and is represented in Fig. 6 as fold increase over control. (Control is no addition of growth factor.) A dose-dependent stimulation of DNA synthesis is noted in cultures from the various time points with a maximum stimulation in nearly all cultures at an HGF concentration of 15 ng. Fetal day 15 (F-DAY 15) is indicated by light dotted bars; fetal day 19 (F-DAY 19) is indicated by black bars; fetal day 20 (F-DAY 20) is indicated by diagonally striped bars; and neonatal day 14 (N-DAY 14) is represented by horizontally striped bars.

The tissue distribution of HGF during rat development was investigated showing that HGF is present in many embry-onic epithelial tissues. These findings are of interest since HGF was demonstrated by us and others to stimulate DNA synthesis and/or induce process formation and migration in a multitude of adult epithelial cells in culture such as rat kidney proximal tubule epithelial cells, rat non-parenchy-mal liver cells, human melanoma cells, mouse keratinocytes (Kan et al., 1991), melanocytes (Kan et al., 1991; Rubin et al., 1991; Matsumota et al., 1991), mammary epithelial cells (Rubin et al., 1991) and vascular endothelial cells (Rosen et al., 1990a). In addition, we have shown that HGF is present in most adult epithelial tissues (Zarnegar et al., 1990 and Wolf et al., 1991a, b) and is detected in unfertilized human oocytes (Wolf et al., 1991b). Several investigators have also demonstrated that HGF is produced by a variety of fetal fibroblasts (Stoker et al, 1987; Rosen et al., 1990b; and Rubin et al., 1991) and human placental fibroblasts (Rosen et al., 1990b).

Recently, the receptor for HGF was discovered as the product of the proto-oncogene, c-MET (Bottaro et al., 1991; Naldini et al., 1991a), which is a transmembrane protein with an Mr of 190,000 endowed with tyrosine kinase activity and consists of a large and small subunit with Mrs of 145,000 and 50,000. In addition to being present in sev-eral adult tissues including liver, c-MET is also expressed in some of the same embryonic tissues as HGF. Chan et al. (1988) isolated c-MET mRNA from mouse embryonic tissues of various gestational age such as brain, liver and kidney with expression of c-MET apparently rising towards the end of gestation.

Since HGF was first characterized by its ability to stim-ulate DNA synthesis in hepatocytes, studies were performed to determine HGF and c-MET mRNA expression in fetal rat liver at various gestational time points and to examine the capacity of fetal rat liver cultures to respond to HGF by increased DNA synthesis. Selden et al. (1990) have demonstrated that HGF mRNA is detectable in fetal human liver by northern blot analysis. We show by RT-PCR that both HGF and c-MET mRNA are present in liver as early as day 14 (earliest time point examined) of rat embryoge-nesis, while liver cultures from as early as day 15 (earliest time point examined), which contained approximately 85% hepatocytes, respond to adult rabbit HGF in a dose-depen-dent fashion. HGF immunoreactivity was also observed in hematopoietic cells of the liver. Thus, the presence of HGF in day 19 rat liver, the expression of HGF mRNA during human and rat fetal liver development (Selden et al., 1990), and the expression of HGF’s receptor, c-MET, during mouse (Chan et al., 1988) and rat embryogenesis all suggest that HGF helps regulate fetal liver development.

Cultured fetal hepatocytes have also been shown to respond to other growth factors. For instance, Hoshi et al. (1987) cultured human fetal hepatocytes and added a vari-ety of growth promoting substances (including epidermal growth factor [EGF], acidic fibroblast growth factor [aFGF], platelet-derived growth factor [PDGF] and trans-forming growth factor beta [TGF′]) to determine their growth factor requirements. It was found that epidermal growth factor, insulin, dexamethasone, bovine neural extract and conditioned medium from hepatocellular carci-noma cells added separately to cultures all maintained hepa-tocyte viability, but the addition of EGF in the presence of insulin, the addition of conditioned hepatocellular carci-noma media, or the addition of 10% dialyzed fetal bovine serum (dFBS) stimulated DNA synthesis. Hoshi et al. (1987) subsequently concluded that a factor(s) promoting embryonic hepatocyte growth may either be EGF or an EGF-like substance and/or other factors that are present in dFBS and conditioned hepatocellular carcinoma media. They also suggested that hepatocyte growth factor may indeed be one of the serum-borne factors. We and others have isolated HGF from rat, rabbit and human serum in earlier studies lending support to the finding of Hoshi et al. (1987).

Based on immunohistochemical results and on studies demonstrating that HGF is capable of stimulating DNA syn-thesis, cell migration and morphogenesis in a variety of epithelial cell types in culture, one may postulate that HGF helps to maintain mesenchymal-epithelial homeostasis. Mesenchymal-epithelial interactions encompass two types: ‘instruction,’ whereby one cell type via soluble or mem-brane-bound molecules induces a second cell type to dif-ferentiate, or ‘permission,’ in which one cell type produces substances that maintain a fully differentiated or already differentiating cell type on a predicted course. Once the initial signal is sent to the neighboring tissue, an interplay between the two tissues ensues regulating changes occur-ring in both. Interactions as such are especially evident during development, and the signals transmitted from one cell type to another often include growth factors (Sawyer and Fallon, 1983). Transforming growth factor beta, for example, is a known inhibitor of epithelial cell growth in culture (Sporn and Roberts, 1988). [In fact, it is also a potent inhibitor of the stimulatory activity of HGF (Zarne-gar and Michalopoulos, 1989).] Akhurst et al. (1990) have shown that during mouse development TGF′ mRNA expression is limited primarily to epithelial cell types. Yet, TGF′ protein is localized only in mesenchymal tissues (Heine et al., 1987). This has fueled speculation that TGF′ is produced and secreted by epithelial cells to modulate epithelial and mesenchymal growth. As it has been pointed out, HGF immunolocalization in the developing rat is detected primarily in epithelial cell types, while HGF mRNA has been detected by northern blot analysis and in situ hybridization in mesenchymal tissues of adult rats such as Kupffer cells and Ito cells of the liver, macrophages of the lungs (Noji et al., 1990) and mesangial cells of the kidney (Ishibashi et al., 1992). HGF protein, meanwhile, has been isolated from several fetal fibroblast cell lines such as MRC5 (Stoker et al., 1987; Rosen et al., 1990b; Rubin et al., 1991). This may suggest that HGF and TGF′ are secreted in a paracrine mechanism to elicit their effects in development.

Perhaps some of the most compelling evidence indicat-ing that HGF may have inductive effects during develop-ment comes from the studies of Stern et al. (1990), who grafted either pellets of cells producing HGF or inert car-riers loaded with HGF such as agar, AG1-X2 formate beads or Mono-S-beads into chick embryos. They found a statistically significant number of unusual axial structures result-ing in embryos with such implants. Typical anomalies included bent embryonic axes, secondary mesodermal or neural structures, disruption of somite structure reminiscent of ‘de-epithelialization’ and extrogastrulae. Embryos receiving implants containing cells which do not produce HGF or inert carriers loaded with protein isolated from the supernatant of non-producing cells had fewer anomalies. The means by which HGF induces the formation of such structures is unknown at present; however, as Stern et al. (1990) point out, HGF’s ability to disperse cells is not suf-ficient to explain the development of additional neural structures arguing for the possibility that HGF elicits ‘inductive’ responses in cells. HGF is also believed to induce morphogenesis in adult MDCK (Madin-Darby Canine Kidney epithelial cells) cells. Montesano et al. (1991) have shown that addition of purified HGF to MDCK cells plated in collagen gels stimulates formation of tubule-like structures. This morphology could not be induced by the addition of other growth factors such as EGF, PDGF or acidic FGF.

Investigations detailing the cell types in which HGF is expressed should prove quite insightful to discovering the role of HGF in development, especially of epithelial tissues. In light of this, however, one must address the influence other growth factors may have on epithelial growth and morphogenesis. As mentioned earlier, TGF′ is most likely involved in mesenchymal-epithelial homeostasis along with a plethora of other substances including epidermal growth factor/transforming growth factor α (EGF/TGFα)? ker-atinocyte growth factor (KGF), and insulin-like growth factor I (IGF-I). As an example, both EGF and TGFα have been demonstrated to stimulate DNA synthesis in adult, neonatal and fetal epithelial cells in vitro (For Review, See Carpenter and Wahl, 1991). In addition to stimulating DNA synthesis in epithelial cells, EGF and TGFα are also capable of promoting morphogenesis in a variety of fetal epithelial tissues such as chick lung (Goldin and Opper-man, 1980) and mouse mammary glands (Tonelli and Sorof, 1980; and Taketani and Oka, 1983) while preventing mouse tooth morphogenesis as determined by organ culture exper-iments (Partanen et al., 1985; for review, See Partanen, 1990). Studies to determine the presence of EGF and TGFα transcripts in the embryo showed that only TGFα? and not EGF, mRNA was detected in embryonic tissues (Matrisian et al., 1982, Proper et al., 1982; Twardzik et al., 1982, 1985; Han et al., 1987; and Wilcox and Derynck, 1988) whereas EGF mRNA can be found in neonates beginning at two weeks after birth (Popliker et al., 1987) and in several adult tissues with the highest expression occurring in the sub-maxillary gland and kidney suggesting that TGFα may be a fetal form of EGF. Very few adult tissues express TGFα; however, northern blot and in situ hybridization techniques carried out on human neonatal foreskin and on adult ker-atinocytes indicated that TGFα mRNA is detectable in ker-atinocytes but not in melanocytes (Coffey et al., 1987). Also discovered was a dose-dependent induction of TGFα mRNA expression in human foreskin keratinocyte cultures by EGF suggesting that, in vivo, an autocrine mechanism for TGFα production in these cells may exist (Coffey et al., 1987). This type of expression is in contrast to that of HGF which appears to be expressed during development and throughout adulthood.

KGF, a member of the fibroblast growth factor family, has been shown to be mitogenic for a mouse keratinocyte cell line (BALB/MK) and human neonatal foreskin ker-atinocytes in culture and is believed to stimulate ker-atinocyte proliferation in vivo. KGF was initially purified from supernatants of stromal cell cultures and characterized by its ability to stimulate BALB/MK cells in culture. Marchese et al. (1990) have shown that when KGF or EGF was added to human keratinocytes cultured in subphysio-logical concentrations of calcium, both EGF and KGF stim-ulated cell proliferation. However, at physiological con-centrations of calcium which induce terminal differentiation and cessation of cell growth in keratinocytes, no net increase in DNA synthesis was noted in the presence of KGF or EGF but only KGF permitted the expression of markers indicative of terminal differentiation such as K1 or filaggrin. Remarkably, Matsumoto et al. (1991) have demonstrated that HGF overcomes calcium-induced inhibi-tion of cell growth and actually stimulates DNA synthesis and motility of keratinocytes which have been induced to differentiate with physiological levels of calcium. HGF was also shown to suppress DNA synthesis but not motility in keratinocytes cultured at low calcium levels. Perhaps it is important to mention at this point that HGF immunoreac-tivity in the skin of rat fetuses was seen not in the basal stem cells but in the terminally differentiating keratinocytes. This may suggest that HGF is somehow involved in this process in vivo.

In summary, immunohistological studies on rat fetuses demonstrate that HGF protein is localized in most epithe-lial tissues including the squamous epithelium of the skin and upper digestive tract, the columnar epithelium of the lung, the exocrine portion of the pancreas, the periventric-ular germinal matrix of the brain, among many others. Non-epithelial cell types demonstrating immunoreactivity for HGF are the hematopoietic cells of the liver, the megakary-ocytes of the liver and lung, and chondrocytes. We also detected HGF and c-MET mRNA in fetal rat liver as early as day 14 of gestation, and demonstrated that fetal rat cul-tures respond to HGF via increased DNA synthesis. These data suggest that HGF accompanied by a myriad of other substances participates in embryogenesis via functions, yet undefined; and, aberrant expression either of HGF or its receptor, c-MET, may result in abnormal development of the fetus. Further investigations will help elucidate the role(s) of HGF in development.

This work was supported by NIH grants CA43632, CA35373 and CA30241 to G.K.M and ACS grant CN55 to R.Z.

Akhurst
,
R.J.
,
Lehnert
,
S.A.
,
Gatherer
,
D.
and
Duffie
,
E.
(
1990
).
The role of TGF′ in mouse development
.
Annals New York Academy of Sciences
593
,
259
271
.
Bottaro
,
D.P.
,
Rubin
,
J.S.
,
Faletto
,
D.L.
,
Chan
,
A. M-L.
,
Kmiecik
,
T.E.
,
Vande Woude
,
G.F.
and
Aaronson
,
S.A.
(
1991
).
Identification of the hepatocyte growth factor receptor as the c-MET proto-oncogene product
.
Science
251
,
802
804
.
Carpenter
,
G.
and
Wahl
,
M.I.
(
1991
).
The epidermal growth factor family
.
In Peptide Growth Factors and Their Receptors
(ed.
M.B.
Sporn
and
A.B.
Roberts
)
New York
:
Springer-Verlag
.
Chan
,
A.M-L.
,
King
,
H.W.
,
Deakin
,
E.A.
,
Tempest
,
P.R.
,
Hilkens
,
J.
,
Kroezen
,
V.
,
Edwards
,
D.R.
,
Wills
,
A.J.
,
Brookes
,
P.
and
Cooper
,
C.S.
(
1988
).
Characterization of the mouse met proto-oncogene
.
Oncogene
2
,
593
599
.
Chan
,
A.M-L.
,
Rubin
,
J.S.
,
Bottaro
,
D.P.
,
Hirschfield
,
D.W.
,
Chedid
,
M.
and
Aaronson
,
S.A.
(
1991
).
Identification of a competitive HGF antagonist encoded by an alternative transcript
.
Science
254
,
1382
1384
.
Coffey
,
R.J.
, Jr
.,
Derynck
,
R.
,
Wilcox
,
J.N.
,
Bringman
,
T.S.
,
Goustin
,
A.S.
,
Moses
,
H.L.
and
Pittelkow
,
M.R.
(
1987
).
Production and auto-induction of transforming growth factor-α in human keratinocytes
.
Nature
328
,
817
820
.
Gohda
,
E.
,
Tsubouchi
,
H.
,
Nakayama
,
H.
,
Horono
,
S.
,
Sakiyama
,
O.
,
Takahashi
,
K.
,
Miyazaki
,
H.
,
Hashimoto
,
S.
and
Daikuhara
,
Y.
(
1988
).
Purification and partial characterization of hepatocyte growth factor from plasma of a patient with hepatic failure
.
J. Clin. Invest
.
81
,
414
419
.
Goldin
,
G.V.
and
Opperman
,
L.A.
(
1980
).
Induction of supernumerary tracheal buds and the stimulation of DNA synthesis in the embryonic chick lung and trachea by epidermal growth factor
.
J. Embryol. Exp. Morph
60
,
235
243
.
Han
,
V.K.
,
D’Ercole
,
A.J.
and
Lee
,
D.C.
(
1988
).
Expression of transforming growth factor alpha during development
.
Can. J. Physiol. Pharmacol
.
66
,
1113
1121
.
Heine
,
U.I.
,
Flanders
,
K.
,
Roberts
,
A.B.
,
Munoz
,
E.F.
and
Sporn
,
M.B.
(
1987
).
Role of transforming growth factor-beta in the development of the mouse embryo
.
J. Cell Biol
.
105
,
2861
2876
.
Hoshi
,
H.
,
Kan
,
M.
and
McKeehan
,
W.L.
(
1987
).
Direct analysis of growth factor requirements for isolated human fetal hepatocytes
.
In Vitro Cellular & Devel Biology
23
,
723
32
.
Ishibashi
,
K.
,
Sasaki
,
S.
,
Sakamoto
,
H.
,
Nakamura
,
Y.
,
Hata
,
T.
,
Nakamura
,
T.
and
Marumo
,
F.
(
1992
).
Hepatocyte growth factor is a paracrine factor for renal epithelial cells: stimulation of DNA synthesis and Na+/K+ ATPase activity
.
Biochem Biophys Res Commun
182
,
960
965
.
Kan
,
M.
,
Zhang
,
G.H.
,
Zarnegar
,
R.
,
Michalopoulos
,
G.K.
,
Myoken
,
Y.
,
McKeehan
,
W.
and
Stevens
,
J.I.
(
1991
).
Hepatocyte growth factor/hepatopoietin A stimulates the growth of rat kidney proximal tubule epithelial cells (RPTE), rat non-parenchymal liver cells, human melanoma cells, mouse keratinocytes and stimulates anchorage-independent growth of SV-40 transformed RPTE
.
Biochem. Biophys. Res. Commun
.
174
,
331
337
.
Kinoshita
,
T.
,
Tashiro
,
K.
and
Nakamura
,
T.
(
1989
).
Marked increase of HGF mRNA in non-parenchymal liver cells of rats treated with hepatotoxins
.
Biochem. Biophys. Res. Commun
.
165
,
1229
1234
.
Lindroos
,
P.M.
,
Zarnegar
,
R.
and
Michalopoulos
,
G.K.
(
1991
).
Hepatocyte growth factor (Hepatopoietin A) rapidly increases in plasma before DNA synthesis and liver regeneration stimulated by partial hepatectomy and carbon tetrachloride administration
.
Hepatology
13
,
743
750
.
Marchese
,
C.
,
Rubin
,
J.
,
Ron
,
D.
,
Faggioni
,
A.
,
Torrisi
,
M.R.
,
Messina
,
A.
,
Frati
,
L.
and
Aaronson
,
S.A.
(
1990
).
Human keratinocyte growth factor activity on proliferation and differentiation of human keratinocytes: differentiation response distinguishes KGF from EGF family
.
J. of Cell. Phys
.
144
,
326
332
.
Matrisian
,
L.M.
,
Pathak
,
M.
, and
Magun
,
B.E.
(
1982
).
Identification of an epidermal growth factor-related transforming growth factor from rat fetuses
.
Biochem. Biophys. Res.Commun
.
107
,
761
769
.
Matsumoto
,
K.
,
Hashimoto
,
K.
,
Yoshikawa
,
K.
and
Nakamura
,
T.
(
1991
).
Marked stimulation of growth and motility of human keratinocytes by hepatocyte growth factor
.
Exp.Cell Res
.
196
,
114
120
.
Miyazawa
,
K.
,
Tsubouchi
,
H.
,
Naka
,
D.
,
Takahashi
,
K.
,
Okigaki
,
M.
,
Arakaki
,
H.
,
Nakayama
,
H.
,
Hirono
,
S.
,
Sakiyama
,
O.
,
Gohda
,
E.
,
Daikuhara
,
Y.
and
Kitamura
,
N.
(
1989
).
Molecular cloning and sequence analysis of cDNA for human hepatocyte growth factor for hepatocytes
.
Bioch. Biophys. Res. Commun
.
163
,
967
973
.
Miyazawa
,
K.
,
Kitamura
,
A.
,
Naka
,
D.
and
Kitamura
,
N.
(
1991
).
An alternatively processed mRNA generated from human heptocyte growth factor gene
.
Eur. J. Biochem
.
197
,
15
22
.
Montesano
,
R.
,
Matsumoto
,
K.
,
Nakamura
,
T.
and
Orci
,
L.
(
1991
).
Identification of a fibroblast-derived epithelial morphogen as hepatocyte growth factor
.
Cell
67
,
901
908
.
Nagaike
,
M.
,
Hirao
,
S.
,
Tajima
,
H.
,
Noji
,
S.
,
Taniguchi
,
S.
,
Matsumoto
,
K.
and
Nakamura
,
T.
(
1991
).
Renotropic functions of hepatocyte growth factor in renal regeneration after unilateral nephrectomy
.
J. Biol. Chem
.
266
,
22781
22784
.
Nakamura
,
T.
,
Nawa
,
K.
,
Ichihara
,
A.
,
Kaise
,
A.
and
Nishino
,
T.
(
1987
).
Purification and subunit structure of hepatocyte growth factor from rat platelets
.
FEBS. Letters
224
,
311
318
.
Nakamura
,
T.
,
Nishizawa
,
T.
,
Hagiya
,
M.
,
Seki
,
T.
,
Shimonishi
,
M.
,
Sugimura
,
A.
,
Tashiro
,
K.
and
Shimizu
,
T.
(
1989
).
Molecular cloning and expression of human hepatocyte growth factor
.
Nature
342
,
440
443
.
Naldini
,
L.
,
Vigna
,
E.
,
Narsimhan
,
R..P
,
Gaudino
,
G.
,
Zarnegar
,
R.
,
Michalopoulos
,
G.K.
and
Comoglio
,
P.M.
(
1991a
).
Hepatocyte growth factor (HGF) stimulates the tyrosine kinase activity of the receptor encoded by the proto-oncogene c-MET
.
Oncogene
6
,
501
504
.
Naldini
,
L.
,
Weidner
,
K.M.
,
Vigna
,
E.
,
Gaudino
,
G.
,
Bardelli
,
A.
,
Ponzetto
,
C.
,
Narsimhan
,
R.P.
,
Hartmann
,
G.
,
Zarnegar
,
R.
,
Michalopoulos
,
G.
,
Birchmeier
,
W.
and
Comoglio
,
P.M.
(
1991b
).
Scatter factor and hepatocyte growth factor are indistinguishable ligands for the MET receptor
.
EMBO J
.
10
,
2867
2878
.
Noji
,
S.
,
Tashiro
,
K.
,
Koyama
,
E.
,
Nohno
,
T.
,
Ohyama
,
K.
,
Taniguchi
,
S.
and
Nakamura
,
T.
(
1990
).
Expression of hepatocyte growth factor gene in endothelial and Kupffer cells of damaged rat livers, as revealed by in situ hybridization
.
Biochem. Biophys. Res. Commun
.
173
,
42
47
.
Okajima
,
A.
,
Miyazawa
,
K.
and
Kitamura
,
N.
(
1990
).
Primary structure of rat hepatocyte growth factor and induction of its mRNA during liver regeneration following hepatic injury
.
Eur. J. Biochem
.
193
,
375
381
.
Partanen
,
A-M.
,
Ekblom
,
P.
and
Thesleff
,
I.
(
1985
).
Epidermal growth factor inhibits morphogenesis and cell differentiation in cultured mouse embryonic teeth
.
Dev. Biol
.
111
,
84
94
.
Partanen
,
A-M.
(
1990
).
Epidermal growth factor and transforming growth factor α in the development of epithelial-mesenchymal organs of the mouse
.
In Current Topics in Development Biology
24
,
31
55
.
Popliker
,
M.
,
Shatz
,
A.
,
Avivi
,
A.
,
Ullrich
,
A.
,
Schlessinger
,
J.
and
Webb
,
C.G.
(
1987
).
Onset of endogenous synthesis of epidermal growth factor in neonatal mice
.
Dev. Bio
.
119
,
38
44
.
Proper
,
J.A.
,
Bjornson
,
C.L.
and
Moses
,
H.L.
(
1982
).
Mouse embryos contain polypeptide growth factor(s) capable of inducing a reversible neoplastic phenotype in nontransformed cells in culture
.
J. Cell. Phys
.
110
,
169
174
.
Rosen
,
E.M.
,
Mermosky
,
L.
,
Setter
,
E.
,
Vinter
,
D.W.
and
Goldberg
,
I.D.
(
1990a
).
Purified scatter factor stimulates epithelial and vascular endothelial cell migration
.
Proc. Soc. Exp. Biol. Med
.
195
,
34
43
.
Rosen
,
E.M.
,
Meromsky
,
L.
,
Romero
,
R.
,
Setter
,
E.
and
Goldberg
,
I.D.
(
1990b
).
Human placenta contains an epithelial scatter protein
.
Biochem. Biophys. Res. Commun
.
168
,
1082
1088
.
Rubin
,
J.S.
,
Chan
,
A.M-L.
,
Bottaro
,
D.
,
Burgess
,
W.
,
Taylor
,
W.J.
,
Cech
,
A.C.
,
Hirschfield
,
D.W.
,
Wong
,
J.
,
Miki
,
T.
,
Finch
,
P.
and
Aaronson
,
S.A.
(
1991
).
A broad-spectrum human lung fibroblast-derived mitogen is a variant of Hepatocyte Growth Factor
Proc. Natl. Acad. Sci., USA
88
,
415
419
.
Sawyer
,
R.H.
and
Fallon
,
J.F.
(
1983
).
Epithelial-Mesenchymal Interactions in Development
New York
:
Praeger
.
Schirmacher
,
P.
,
Geerts
,
A.
,
Pietrangelo
,
A.
,
Dienes
,
H.P.
and
Rogler
,
C.E.
(
1992
).
Hepatocyte growth factor/hepatopoietin A is expressed in fat-storing cells from rat liver but not myofibroblast-like cells derived from fat-storing cells
.
Hepatology
15
,
5
11
.
Seki
,
T.
,
Ihara
,
I.
,
Sugimura
,
A.
,
Shimonishi
,
M.
,
Nishizawa
,
T.
,
Asami
,
O.
,
Hagiya
,
M.
,
Nakamura
,
T.
and
Shimizu
,
S.
(
1990
).
Isolation and expression of cDNA for different forms of hepatocyte growth factor from human leukocyte
.
Biochem. Biophys. Res. Commun
.
172
,
321
327
.
Selden
,
C.
,
Jones
,
M.
,
Wade
,
D.
and
Hodgson
,
M.
(
1990
).
Hepatotropin mRNA expression in human foetal liver development and in liver regeneration
.
FEBS Letters
270
,
81
84
.
Sporn
,
M.B.
and
Roberts
,
A.B.
(
1988
).
Transforming growth factor-beta: new chemical forms and new biological roles
.
Biofactors
1
,
89
93
.
Stern
,
C.D.
,
Ireland
,
G.W.
,
Herrick
,
S.E.
,
Gherardi
,
E.
,
Gray
,
J.
,
Perryman
,
M.
and
Stoker
,
M.
(
1990
).
Epithelial scatter factor and development of the chick embryonic axis
.
Development
110
,
1271
1284
.
Stoker
,
M.
,
Gherardi
,
E.
,
Perryman
,
M.
and
Gray
,
J.
(
1987
).
Scatter factor is a fibroblast-derived modulator of epithelial cell mobility
.
Nature
327
,
239
242
.
Taketani
,
Y.
and
Oka
,
T
, (
1983
).
Possible physiological role of epidermal growth factor in the development of the mouse mammary gland during pregnancy
.
FEBS Letters
152
,
256
260
.
Tashiro
,
K.
,
Hagiya
,
M.
,
Hishizawa
,
T.
,
Seki
,
T.
,
Shimonishi
,
M.
,
Shimizu
,
S.
and
Nakamura
,
T.
(
1990
).
Deduced primary structure of rat hepatocyte growth factor and expression of the mRNA in rat tissues
.
Proc. Nat. Acad. Sci., USA
87
,
3200
3204
.
Tonelli
,
Q.J.
and
Sorof
,
S.
(
1980
).
Epidermal growth factor requirement for development of cultured mammary gland
.
Nature
285
,
250
252
.
Twardzik
,
D.R.
,
Ranchalis
,
J.E.
and
Todaro
,
G.J.
(
1982
).
Mouse embryonic transforming growth factors related to those isolated from tumor cells
.
Cancer Res
.
42
,
590
593
.
Twardzik
,
D.R.
(
1985
).
Differential expression of transforming growth factor-alpha during prenatal development of the mouse
.
Cancer Res
.
45
,
5413
6
.
Weidner
,
K.M.
,
Arakaki
,
N.
,
Hartmann
,
G.
,
Vandekerckhove
,
J.
,
Weingart
,
S.
,
Rieder
,
H.
,
Fanatsch
,
C.
,
Tsubouchi
,
H.
,
Hishida
,
T.
,
Daikuhara
,
Y.
and
Birchmeier
,
W.
(
1991
).
Evidence for the identity of human scatter factor and human hepatocyte growth factor
.
Proc. Nat. Acad. Sci., USA
88
,
7001
7005
.
Wilcox
,
J.N.
and
Derynck
,
R.
(
1988
).
Developmental expression of transforming growth factors alpha and beta in mouse fetus
.
Mol. and Cell. Bio
.
8
,
3415
3422
.
Wolf
,
H.K.
,
Zarnegar
,
R.
,
Oliver
.,
L.
and
Michalopoulos
,
G.K.
(
1991a
).
Hepatocyte growth factor in human placenta and trophoblastic disease
.
Am. J. Path
.
138
,
1305
1043
.
Wolf
,
H.K.
,
Zarnegar
,
R.
and
Michalopoulos
,
G.K.
(
1991b
).
Localization of hepatocyte growth factor in human and rat tissues: an immunohistochemical study
.
Hepatology
14
,
488
494
.
Zarnegar
,
R.
and
Michalopoulos
,
G.
(
1989
).
Purification and biological characterization of human hepatopoietin A: a polypeptide growth factor for hepatocytes
.
Cancer Res
.
49
,
3314
3320
.
Zarnegar
,
R.
,
Muga
,
S.
,
Rahija
,
R.
and
Michalopoulos
,
G.
(
1990
).
Tissue distribution of HPTA: a heparin-binding polypeptide growth factor for hepatocytes
.
Proc. Natl. Acad. Sci., USA
87
,
1252
1256
.
Zarnegar
,
R.
,
DeFrances
,
M.C.
,
Kost
,
D.
,
Lindroos
,
P.
and
Michalopoulos
,
G.K.
(
1991
).
Expression of hepatocyte growth factor (HGF) in regenerating rat liver after partial hepatectomy
.
Biochem. Biophys. Res. Commun
.
177
,
559
565
.