Liver expression of epidermal growth factor RNA. Rapid increases in

THEJOI’RNAL OF BlULoGlCAL CHEMISTRY

Communication

Vol. 269, No. 31, Issue of August 5, pp. 19667-19670, 1994 Printed in U.S.A.

Liver Expression of Epidermal Growth Factor RNA

number of cellular substrates, some of which appear to change the activity of transcriptional regulators such as Fos, Jun, and Myc (3, 6). Recently, it has been shown that following EGF injection, three mouse liver cytoplasmic proteins, which become W I D INCREASES IN IMMEDIATE-EARLY PHASE OF rapidly phosphorylated, are translocatedto the nucleus; two of these proteins appearto be identical to a y-interferon-inducible L m R REGENERATION* complex (GAF), a DNA-binding factor that transactivates in(Received for publication, May 27, 1994) terferon-responsive genes (9). Within minutes following the Beat MullhauptS, Anna Feren, Eric Fodorl, and binding of EGF, the ligand-receptor complex becomes internalAlbert Jones ized and is largely degraded (10). However, a fraction of the From the Veterans Administration Medical Center and EGF-bound receptor has been localized to the cell nucleus, but the Department of Medicine, Anatomy and LiverCenter, its significance remains unknown (11,12). University of California, S u n In normalliver, hepatic cells are quiescent, yet within hours Francisco, California 94143 after liver injury or following a partial removal of liver tissue, The liver exhibits a remarkable capacity to regenerate they rapidly progress into mitosis (13, 14). A number of mitoits mass following partial removal or after injury. Transgens have been identified that will activate hepatic cellular membrane receptors for epidermal growth factor (EGF)division, but the underlying mechanism initiating the liver are highly expressed in liver cells, which quickly re- growth program remains unknown and likely occurs by more spond to this polypeptide mitogen by activating an in- than one pathway (13, 14). Gene expression profiles in immetrinsically low rate of cell division. Although EGF ap- diate-early times preceding the onset of liver cell replication pears to regulate liver growth, its significance has have been closely examined to provide clues to the identity of remained unclear, and only a small change in serum lev- initiatingstimuli (15, 16). Many changes seenwithinthe els can bedetected during hepatocellular proliferation. course of liver regrowth canbe reproduced in cell culture by the Using a reverse transcription-polymerasechain reaction addition of transforming growth factor a (TGF-a) and hepatoassay, we report here the novel finding that EGFRNA cyte growth factor (HGF), which are synthesized by liver cells, transcripts are synthesized in a hepatic cell-specificpattern, appearing in hepatocyte and lipocyte cell types. and by the addition of EGF (17, 18). These experiments and Our data revealthat within 15 min following 70% a liver others have led ato hypothesis that polypeptide growth factors removal, EGF RNA levels increase >lO-fold andthen di- function within the initiation phase of hepatic cellular prolifatemporal coincidence of minish below basal levels prior to the first wave of re- eration. Attempts to demonstrate generative cell division. Immunoanalysis of metaboli- increases in TGF-a or HGF mRNA levels or in their protein cally labeled hepatocytesshows that EGF accumulatesas levels with the immediate-early timeperiod of hepatic prolifa large 60-kDa peptide. Theseresults demonstrate that eration have been negative, although HGF and TGF-a levels EGF transcription is a previously unrecognized compo- have been shown to increase at times slightly preceding the nent of hepatic gene expression, and rapidincreases in first wave of regenerative cell division (19-22). It has been EGFRNA levels in the immediate-early phase of liver proposed that TGF-a, which also binds to the EGF receptor, regeneration pointto EGF as an autocrine factorin the and HGF participate as signals activating a G, to S transition prereplicative hepatic growth program. of a previously primed hepatocyte (23). Here, we report data that show for the first time that RNA encoding the strong hepatic mitogen EGF is expressed by the liver. In contrast to Epidermal growth factor (EGF)l is a 140-kDa glycosylated HGF and TGF-a, our data show that EGF RNA rapidly accutransmembrane protein that is normally processed to a small mulates within theimmediate-early phase of liver growth. 6-kDa peptide and released from the cell surface (1, 2). EGF EXPERIMENTAL.PROCEDURES binds a receptor found in a variety of cells and, after binding, Isolation of RNA-Freshly removed liver or kidney was rinsed in transmits signals for multiple physiological responses, including the differentiation of certain epithelial cells, inhibition of phosphate-buffered salineand combined on ice withRNAzolTM(Tel-Test gastric acid secretion, and induction of mitosis (3, 4). Following Inc., Friendswood,TX)at a ratio of 1 g/20 ml of reagent (24). RNA was immediately solubilized in an ice bath with a Brinkmann large probe ligand binding, the cytoplasmic tyrosine kinase domain of the Polytron homogenizer using three 4-s bursts. RNA was isolated essenEGF receptor is autophosphorylated, presentinga binding site tially as recommended with two additional chloroform extractions, folfor SH2 domain-containing protein(s) that are intermediaries lowed by precipitation with 1.5 volumes of isopropyl alcohol. RNA pelml in a complex EGF signal transduction pathway(5-8). Occupa- lets were collectedby centrifugation, air-dried, and resuspended5 in tion of the EGF receptor also leads to phosphorylation of a of H,O/g of tissue. Sodium acetate was added to0.3 M , and the RNA was stored as a precipitate at -20 “C after the addition of 2 volumes of EtOH. Fractionated liver cells were isolated as described (29, 301, and * This work was supported in part by a merit review grant from the RNA was extracted as described above at a ratio of 1 ml of RNAzolTM/5 Veterans Administration (to A. J.).The costs of publication of this article were defrayedin part by the payment of page charges, This article x lo6 cells. OligonucleotidePrimers-Sequences of oligonucleotides used for RTmust therefore be hereby marked “advertisement”in accordance with18 PCR were obtained from published cDNA sequences and are listed in U.S.C. Section 1734 solely to indicate this fact. the 5’ to 3’ direction with coordinates as reported: for EGF, antisense j: Supported in part by the Swiss National Science Foundation. 8 To whom correspondence should be addressed. Tel.: 415-221-4810 positions 3464 to 3449 (GCAGCTTCCACCAACG),sensepositions 3033-3048 (GCTCAGACTGTCCTCC), sense positions3199-3214 (GG(ext. 3449); Fax: 415-750-6947. I The abbreviations used are: EGF, epidermal growth factor; TGF-a, 2864-2883 (GCTGAGATCAGAGGCTACAACTGC), and sense positions transforminggrowth factor a; HGF, hepatocyte growthfactor;RT-PCR, TGGTGTCAGG) (25); for @actin, antisensepositions 3774-3756 (GGTreverse transcription-polymerase chain reaction; bp, base pair. CTCACGTCAGTGTACAGG) and sense positions 3119-3136 (CCGCA-

19667

19668

Regulation of Liver EGF RNA during

AATGCTTCTAGGC) (26); andfor the EGFreceptor, antisense positions 2204 to2188 (CGTAGTGTACGCTTTCG) and sense positions 19821997 (GGAAGTATGCAGATGC)(27, 28). Reverse Dunscription-Immediately prior to use, RNA was precipitated from a n ethanolsuspension by centrifugationanddriedin vacuum. 5 pmol of the antisense oligonucleotide was added to up to 10 pg of RNAin a finalvolume of 20 p1 containing 50 mM Tris, pH 8.5,8 mM MgCl,, 30 mM KCI, 1 mM dithiothreitol, and 0.5 mM each dNTP and reverse-transcribed for 1 h a t 42 "C with 24 units of avian myeloblastosis virus reverse transcriptase (Boehringer Mannheim), followed by heat inactivation for 5 min at 95"C. Polymerase Chain Reaction-Up to 20% of the reverse transcript cDNA was added to a reaction mixture containing10 mM Tris, pH 8.0, 50 mM KCl, 2.5 mM MgCI,, 200 nM each dNTP, 100 pmol of the sense and antisense primers, and 1.75 units of Taq polymerase (Perkin-Elmer) and adjusted with distilled H,O to a final volume of 50 p1. 30 cycles of PCR was initiated after overlaying with mineral oil using a PerkinElmer 4800 thermal cycler. Single cycles consisted of 1 min at 95 "C, 1 min a t 65 "C, and 1 min a t 72 "C for EGF primers;1min at 95 "C, 1 min at 60 "C, and 1min a t 72 "C for actin primers; and1min at 95 "C, 1 min at 50 "C, and 1min at 72 "C for EGF receptor primer pairs. PCR bands were identified by size after electrophoresison a 2% agarose gel in 0.04 M Tris acetate, 0.001 M EDTA, run at 10 V/cm for 1 h; stained with ethidium bromide (0.1 pg/ml); and viewed under 300 nm transillumination. For quantification, a video image of the ethidium bromidestained gel was obtained and stored by computer using a U V P Model 5000 gel documentation system. The bands of interest were scanned, and peak areas were determined using Image 1.47 matching software (National Institutes of Health). Cell Culture,Immunoprecipitation, a n d Western Blot AnalysisHepatocytes were isolatedby collagenase perfusion (Liver Center Core Facility, San Francisco General Hospital, University of California, San Francisco, CA) and plated a t 4 x loficells/60-mm collagen-coated plate. Incubation and media compositions were as described (29). 48 h after plating, the medium was replaced with cysteine-free Dulbecco's modified Eagle's medium containing5% dialyzed fetal calf serum. After3 h, fresh medium was added containing 1 mCi of [3'S1cysteine. After 16 h, the medium was removed, and cells were washed with ice-cold phosphate-buffered saline, removed by scraping, and collected by centrifugation. Soluble protein was obtained after lysis in radioimmune prea type B Dounce homogenizer, followed by cipitationbufferusing centrifugation at 10,000 x g for 15 min. For immunoprecipitation, protein was precleared with Pansorbin (Calbiochem) for 2 at h room temperature, and 25 pl of serum was added per 200 p1 of cell extract, followed by incubation a t 4 "C overnight. Immunocomplexes were then precipitated with Pansorbin, washed, and dissociated in SDS sample buffer. For Western blot analysis, hepatocyte immunoprecipitates were prepared a s described above without ["Slcysteine and transferred to nitrocellulose after electrophoresis. EGF was detectedby ECL (Amersham Corp.) according to the instructions of the manufacturer using a 1:10,000 dilution of rat anti-EGF anda 1:50,000 dilution of horseradish peroxidase-linked secondary antibody.

Regenerative Liver Growth

I b.

M

Liver

640

W Kidnev

640

FIG.1.Analysis of EGF t r a n s c r i p t expression in liver and kidney by RT-PCR. a , schematic representationof the 4801-base pair rat EGF cDNA. The black box indicates the coding region for the mature EGF peptide, and shaded boxes represent EGF repeats6-8. Black bars depict primers used for analysis. PCR products of 266, 432, and 640 base pairs are indicated.b, RT-PCR analysis of liver and kidneyRNAs using the EGF coding region oligonucleotides. 40% of each reaction was electrophoresed on a 2% agarose gel. The sizesof the amplified products and molecular standards ( M ) are in base pairs.

RNA. Control amplifications in which the reverse transcription step was omitted complete or RT-PCR amplifications with yeast tRNA as template were negative. To confirm that hepatic RTPCR products were arising from the EGF coding region, the 432- and 266-base pair bands were isolated and digested with restriction endonucleases, and after electrophoresis, the products were compared to a map of the EGFcDNA sequence in the same region (25). Our resultsshow that the432- and 266-base pair PCR products from liver RNA were cleaved by RsaI and HinfI to fragments of a size expected from the EGF cDNA sequence (data not shown). To confirm further the identity of the RT-PCR signal, we subcloned the 432-base pair product and obtained its DNA sequence, which was identical to the published rat EGF sequence between the primer pairs (data not shown) (25). The linear response of the RT-PCR assay was determined from a log plot of the PCR product area versus the log template input asdescribed (31). As shown in Fig. 2, this analysisyields a linear correspondence for the abundant p-actin RNA up to a value of 40 ng. The relative level of EGF transcripts in the adult liver RNA was next estimatedby constructing a series of liver EGF cDNA template dilutions andcomparing PCR prodRESULTS ucts from each to a parallel series using adultkidney RNA, in Given the pronounced effect of EGF on hepatic growth, we which EGF mRNA is highly expressed (30). Fromthe analysis shown in Fig. 2, we can estimate that the dilution of kidney developed a quantitative assay using RT-PCR to determine whether the liver may constitute an uncharacterized site of RNA template that is necessary to achieve a PCR product EGF geneexpression and further sought to determine whether equivalent to thatof the liver is in the rangeof 15-fold. Thus, such expression is responsive to theactivation phase of hepatic while we cannot extract an absolute value from this comparison growth. Our strategy was identify to RNA molecules containing ( i e . nonequivalent ratios ofRNA to cell) following identical the mature EGF coding region using anoligonucleotide comple- conditions of hybridization, cDNA synthesis, andamplification, mentary to the C-terminal region. After hybridizing the 3'- liver EGF transcripts appearat a 15-fold lower level per mass antisense oligonucleotide to totalliver RNA, cDNAwas s p t h e - of RNA than isfound in equivalent adult kidney RNA samples. sized and used as a template for amplification by PCR using The liver is composed primarily of hepatocytes, but with a sense primers designed to hybridize a t sequential positions significantpopulation of specialized non-parenchymal cells, upstream covering the maturepeptide coding region shown in predominately of endothelial, Kupffer, and lipocyte cell types. Fig. la (25). As shown in Fig. lb, electrophoresis of PCR prod- To determine whether EGF transcriptsmay be expressed in a ucts from liver RNA yielded bands of a size predictedfrom the cell-specific pattern, purified liver cell fractions were isolated EGF cDNA sequence. As a control, parallel RT-PCRs were per- (29,321, and identical amounts of RNAfrom each were used for RT-PCR as described above using the432-base pair primer pair formed with kidney RNA, which has previously been characterized to contain high levels of the EGF transcript (30). As (Fig. 1).As shown in Fig. 3a, EGF transcripts were detected shown in Fig. lb, a comparison of liver and kidney RNAs shows only in hepatocyte and lipocyte cell types. Control RT-PCR asEGF PCR products of identical size, although kidney RNA pro- says with a rat p-actin primer pair produced an equivalent duced a considerably higher signal per mass of RNA than liver 656-base pair product in all cell types (Fig. 3b) (26).

Regulation of Liver EGF RNA during Regenerative Liver Growth

19669

EGF Hpx

A.

"2.55 J 0.05

Hrs o

.25 .5

1 4

M

8 24 72

I

0.55 2.55 2.05 1.55 1.05

log R N A ng

B.

1

2

3

C.

M

FIG.2. Quantitative analysis of p-actin and EGF RNA levels by RT-PCR Shown are the results from RT-PCR analysis with the 6-actin primer pair (0) using dilutions of the reverse transcript corresponding FIG.4. Time course of EGF RNA expression in regenerating to 40to 1.25 ngof template RNAand with the 434-base pair EGF primer liver following partial hepatectomy. Male adult Fisher 344 rats pair using a range of 800 to 25 ng of kidney ( A ) and liver (0) RNA underwent 70% hepatectomies ( H p x )(11);a t 0,0.25,0.5, 1,4,8,24, and templates. After electrophoresis, bands were photographed, and video a 72 h, livers were removed, and total RNA was isolated. a , RT-PCR image was obtained and storedby computer. A, the peak area of each analysis of 800 ng of liver RNA using the432-bp EGF primer pair (Fig. band was determined with matching software, and the log value (in 1);b, RT-PCR analysis of 20 ng of liver RNA using the 656-bp actin relative units) is plotteduersus the log of the corresponding RNA tem- primer pair (Fig. 3);c, RT-PCR analysis of 800 ng of RNA from sham plate asdescribed (31).B , shown are ethidium bromide-stained bands of controls, isolated at various time points as described above, amplified the RT-PCR products. with the 432-bp EGF primer pair. The sizes of molecular standards( M I are in base pairs.

levels by 24 h, suggesting a down-regulation preceding DNA replication. In contrast to this transient expression, EGF transcripts remained constant throughout the same timeperiod in the sham-operatedliver RNA samples (Fig. 4c). To verify that this differential expressionof EGF did not resultfrom a general activation and repression of total hepatic RNA synthesis, we evaluated thelevels of p-actin RNA as a control throughout the regeneration timecourse using anRNA input corresponding to the mid-value of the linear p-actin PCR amplification range, FIG.3. Liver cell-type specificity of EGF, p-actin, and EGF re- shown in Fig. 2. As shown in Fig. 4b, liver p-actin transcript ceptor RNAs. 800 ng of RNA from each livercell fraction was analyzed by RT-PCR a s described for Fig. 1. a, RT-PCR with the 432-bp EGF levels remained relatively constant throughout the regrowth primer pair;b, RT-PCR of actin ( A C T )RNA, yielding a 656-bp RT-PCR time period. product; c, expression of EGF receptor (EGF,) mRNA using primer The presence of EGF transcripts in thenormal steady-state pairs, yieldinga 223-bp RT-PCR product. H p , hepatocytes; Lp, lipocytes; population of rodent liver RNA implies that this organ also En, endothelial cells; Kp, Kupffer cells; Lu, liver cells. The sizes of represents an uncharacterized site of EGF protein synthesis. molecular standards ( M ) are in base pairs. To identify de novo synthesis of hepatic EGF and to distinguish i t from that which may be endocytosed from serum, we used To determine whether the cell type-restricted patternof EGF cultured hepatocytes to label nascent protein synthesis. Here, expression could affect each liver cell type, we next designed hepatic cells were obtained from adult liver by collagenase RT-PCR assays to measure EGF receptor transcript levels in perfusion, and thehepatocyte fraction was isolatedby centrifuRNA from individual cell fractions usinga primer pair specific gation (29). Monolayer cultures were then incubatedwith for the receptor extracellular ligand-binding domain (27, 28). [3sS]cysteine-containingmedium, and at 16 protein h, wassoluAs shown in Fig. 3c, our assays demonstrate that while only bilized and immunoprecipitated with purifiedrat anti-EGF IgG hepatocyte and lipocyte cell populations express constitutive or mouse anti-EGF IgG, followed by SDS-polyacrylamide gel levels of the EGF RNA, each liver cell type maintainslevels of electrophoresis and autoradiography. As shown in Fig. 5A, an the receptor mRNA indicating each to be potentially responsive autoradiograph of the immunoprecipitates, using either rator to EGF signaling. mouse anti-EGF antibody, reveals that hepatocyte monolayer These data suggest that the liver constitutes a previously cultures accumulate EGF as a large 60-kDa peptide. To confirm unrecognized source of EGF, and expression of the EGFrecep- the identity of the hepatic EGFpeptide, Western blot analysis tor mRNA in each hepatic cell type supports the notion that was performed using rat anti-EGF antibody, As shown in Fig. EGF functions as an autocrine and paracrine regulator of liver 5B,this analysis reveals a similar 60-kDa EGF immunoreacgrowth. We next sought to determine whether the basal level of tive peptide. EGF ismodified within the timecourse of hepatic growth. For DISCUSSION this experiment, we evaluated EGFRNA levels a t various time intervals following a 70% hepatectomy, which is sufficient to This study shows that RNA transcripts that encode the reinduce massive and rapid cellularproliferation in the remain- ceptor-binding domain of EGF are a normal constituent of heing tissue mass.As shown in Fig. &, within 15 min following a patic gene expression and that rapid increases in these tran70% hepatectomy, EGF RNA rapidly accumulated in theliver scripts appear coincident with the immediate-early phase of remnant. By comparison to a dilution standard (see Fig. 21, liver growth. Our results strongly support the conclusion that within 15 min of a 70% hepatectomy, EGF RNAlevels increased the liver is a site of EGF synthesis and further raise the pos-10-fold (average value of three independent assays). Transibility that amplification of EGF represents a key element in script accumulation stopped after -1 h, falling below base-line the hepatocellularproliferationprogram.Previous studies 432

Regulation of Liver EGF RNA during Regenerative Liver Growth

19670 A.

ANTI PRE EGF IMM

“

M

2w

-

1 2 3

6.

ANTI PRE EGF IMM. _ _

1

2

After binding its receptor, EGF transmits mitogenic signals

- through secondary cytoplasmic messengers, some of which are

9;7 E24

, 69 2 , ”

4

directly linked to changes in gene expression. Indeed, recent data reveal that components of the cytokine and EGF pathways converge to control common sets of genes functioning within the prereplicative phase of cell growth (9). The initiation of liver regeneration proceeds through a sequential and ordered program of gene expression with clear parallels toprograms of mitogen stimulation seen in other cell types (16,23). The transient profile of EGF RNA accumulation reported here suggests that increases in EGF levels may well play a role within the Go-G, period and catalyze a cascade of events preceding the first wave of hepatic DNA replication. Acknowledgments-We gratefully acknowledge Dr. P. Schaudies for the generousgift of purified anti-rat EGF andDr.M. Bissell and associates (Liver Center, University of California, San Francisco, CA) for kindly supplying purified livercell fractions.

REFERENCES FIG.5. Epidermal growth factor protein synthesisin hepatocyte monolayer cultures. A, hepatocyte monolayer cultures were 1. Scott, J., Urdea, M., Quiroga, M., Sanchez-Pescador, R., Fong, N., Selby, M.. maintained on collagen-coated plates for 48 h prior to metabolically Rutter, W. J., and Bell G. I. (1983) Science 221,236-240 labeling with[35Slcysteine.After 16 h, soluble protein was prepared and 2. Gray, A,, Dull, T. J., and Ullrich, A. (1983)Nature 303,722-725 Carpenter, G., and Cohen, S. (1990) J. Biol. Chem. 266,7709-7712 3. immunoprecipitated with anti-EGF antibodies or nonimmune serum. Immunoprecipitates were solubilized and fractionated on a 7.5% SDS- 4. Marti, U., Burwen, S. J., and Jones, A. L. (1989)Hepatology 9, 126-138 5. Carpenter, G . (1987)Annu. Rev. Biochem. 66,881-914 ENHANCE (DuPontNEN), polyacrylamide gel, impregnatedwith 6. Carpenter, G. (1992) FASEB J. 6,32833269 dried, and exposed for 7 days. Lane 1, anti-rat EGF (38); lane 2, anti7. Ullrich, A., and Schlessinger, J. (1990) Cell 61,203-212 mouse EGF (Austral Biologicals); lune 3, normal serum (PREI”., 8. Yarden, Y., and Ullrich, A. (1988)Annu. Reu. Biochern. 67,443478 preimmune). B, immunoprecipitates from nonradiolabeled hepatocyte 9. RuffJaminson, S., Chen, K, and Cohen, S. (1993) Science 261, 1733-1736 monolayers were electrophoresed a s described for A, transferred to ni- 10. Carpenter, G., and Cohen, S. (1979)Annu. Rev. Biochem. 48, 193-216 trocellulose, and detected by chemiluminescence using the ECL detec- 11. RaDer. S. E.. Burwen. S. J., Barker, M. E., and Jones, A. L. (1987) Gastroenierology 92,1243-1250 tion kit (Amersham Corp.). Exposure was for 5 min at room tempera12. Marti, U., Burwen, S. J., Wells, A,, Barker, M. E., Huling, S., Feren, A. M., and ture. Molecular mass standards are in kilodaltons.

have indicated that EGF transcripts are absent or fall below detection limits of Northern analysis (30) inliver RNA. Since, in the presentwork, we have notmapped the entire EGF transcript unit,we cannot exclude the possibility that hepatic EGF transcripts arise from a gene related to EGF and are not detectable by Northern blot analysis or that alternate splicing pattern(s) of hepatic EGF RNA occur, as is described for a fraction of kidney EGF gene expression (25). Following processing and extracellular release, EGF is believed to function primarily as a n exocrine factor (10).However, other workers have identified incompletely processed forms of the EGF peptide that bind the EGF receptor and can support growth of receptor-dependent cell lines (33, 34). Thus, it has been suggested that molecular forms of EGF other than the 6-kDa peptide may also possess biological activity in vivo. It has been pointed out that EGF repeatmotifs found within the precursor coding region are present as conserved elements ina diverse groupof transmembrane ligand receptors, and by analogy, it has been proposed that largerforms of the EGFprecursor mayfunction as a receptor for a n as yet unidentified ligand (35).As we have shown here, hepatocytes accumulate EGF as a large, presumably incompletely processed, 60-kDa protein. In our experiments, we were unable to detect the6-kDa form of EGF unless the putative precursorfirst was extensively treated with trypsin (data notshown). However, it is possible that further processing of the EGFpeptide could occur by an activity extrinsic to thehepatocyte cultures employed here. Nevertheless, similar sized, partially processed EGF peptides havebeen identified as a major species in the kidney and as a predominant storageform in platelets, but as for the hepatocyte form, their functional significance remains unknown (36, 37).

Jones, A. L. (1991) Hepatology 13, 15-20 13. Fausto, N., and Mead, J. E. (1989) Lab. Invest. 60,4-I3 14. Michalopoulos, G. K. (1990) FASEB J. 4, 176-187 15. Thompson, N. L., Mead, J. E., Braun, L., Goyette, M., Shank, P. R., and Fausto, N. (1986) Cancer Res. 46,3111-3117 16. Haber, B. A,, Mohn, K. L., Diamond, R. H., and Taub, R. (1993) J. Clin. Invest 91,1319-1326 17. Kruijer, W., Skelly, H., Botteri, F., van der Putten,H., Barber, J. R., Verma, I. M., and Leffert, H. L. (1986) J. Biol. Chem. 261,7929-7933 18. Koch, K S., Lu, X. P., Brenner, D. A., Fey, G. H., Martinez-Conde, A., and Leffert, H. L. (1990) In Vitro Cell. & Deu. Biol. 26, 1011-1023 19. Mead, J. E., and Fausto, N. (1989) Proc. Natl. Acad. Sci. U.S. A. 86, 15581562 20. Evarts, R. P., Nakataukasa, H., Marsden, E. R., Hu, Z., and Thorgeinson,S. S. (1992) Mol. Carcinog. 6, 25-31 21. Kinoshita, T., Hirao, S., Matsumoto, K, and Nakamura, T. (1991) Biochem. Biophys. Res. Commun. 177,330-335 22. Zarnegar, R., DeFrances, M. C., Kost, D. P., Lindroos, P., and Michalopoulos, G. K (1991) Biochem. Biophys. Res. Commun. 177,559-565 23. Fausto, N., and Webber, E.M. (1993) Crit. Rev. Eukaryotic Gene Expression3, 117-135 24. Chomczynski, P., and Sacchi, N. (1987) Anal. Biochem. 162,156-159 25. Saggi, S. J., Safirstein, R., and Price, P. M. (1992)DNA Cell Biol. 11,481487 26. Nudel, U., Zakut, R., Shani, M., Neuman, S., Levy, 2.. and Yaffe, D. (1983) Nucleic Acids Res. 11, 1759-1771 27. Avivi, A,, Lax, I., Ullrich, A., Schlessinger, J., Givol, D., and Morse, B. (1991) Oncogene 6,673-676 28. Petch, L. A., Harris, J., Raymond, V. W.,Blasband, A., Lee, D. C., and Earp,H. S. (1990)Mol. Cell. Biol. 10, 2973-2982 29. Bissell, D.M., Hammaker, L. E., and Meyer, U.A. (1973) J. Cell Biol. 59, 722-734 30 Rall, L. B., Scott, J., and Bell, G. I. (1985) Nature 313,228-231 31 Murphy, L. D., Herzog, C. E., Rudick, J. B., Fojo, A. T., and Bates, S. E. (1990) Biochemistry 29,10351-10356 32 Friedman. S. L..and Roll. F. J. (1987)Anal. Biochern. 161.207-218 Reich, M., Chen, K., Bell, G. I., and Cohen, S. (1989) Mol. 33. Mroczkowski, Cell. Biol. 9, 2771-2778 34. Breyer, J. A,, and Cohen, S. (1990) J. Biol. Chern. 266,16564-16570 35. Bell, G. I., Fong, N. M., Stempien, M. M., Wormsted, M. A,, Caput, D., Ku, L., Urdea, M. S., Rall, L. B., and Sanchez-Pescador, R. (1986)NucleicAcidsRes. 14,8427-8446 36. Lakshmanan, J., Salido, E. C., Lam, R., Barajas, L., and Fisher, D. A. (1990) Biochem. Biophys. Res. Comrnun. 173,902-911 37. Pesonen, K, Viinikka, L., Myllylli, G., Kiuru,J., and Perheentupa,J. (1989)J. Clin. Endocrinol. Metab. 68, 486491 38. Schaudies, R. P., and Johnson, J. P. (1993)Am. J. Physiol. 264, F523-F531 ~

~~~

B:,