Blocks Differentiation of Transformable 3T3 LI and C3HIOT#{189}-derived Preadipocytes in a Dose- and Time-dependent Manner1

To investigate the functional relationship between the transforming ability of Ras and its role as an integral component of the differentiation-promoting insulin signaling pathway, we introduced a Ieu6l-activated ras gene into the Ras-transformable 3T3 LI (ATCC CCL92.1) and a number of C3HIOT1/ -derived preadipocytic cell lines. The results demonstrate a quantitative reciprocal regulation of differentiation and several transformation-associated properties in response to graded levels of ras gene expression, with the loss of differentiative capacity, morphological transformation, stimulation of proliferation, and anchorage-independent growth requiring increasing levels of Ras’#{176}” 1 protein. Furthermore, using novel, tightly regulatable 3T3 LI transfectants, we demonstrated that Ras ’#{176}1effectiveness in blocking adipocytic differentiation is strictly dependent on the timing of its expression relative to cell growth arrest, with ms’#{176} ’expression being ineffective at inhibiting differentiation or inducing morphological transformation once the differentiative process has commenced. Moreover, ras’#{176} 1induction failed to substitute for or enhance the c-Ras-dependent differentiative insulin signal, even under conditions in which it did not induce transformation. Therefore, although necessary for insulin signal transduction, the Ras signal alone is not sufficient to induce adipocytic differentiation in this system. Consistent with its established role as a downstream effector of Ras, v-Raf expression mirrored the Ra&#{176}’ 1 effects on adipocytic differentiation and transformation. Received 6/3/96; revised 8/29/96; accepted 11/6/96. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to mdicate this fact. 1 The financial assistance of the Cancer Research Society Inc., and the Natural Sciences and Engineering Research Council of Canada through grants to LA. and T.H. is gratefully acknowledged. 2 To whom requests for reprints should be addressed, at Department of Pathology, Beth Israel Hospital, Harvard Medical School, Boston, MA 02215. Phone: (617) 667-4344; Fax: (617) 667-2943; E-mail: thalioti@ warren.med.harvard.edu. Introduction An accruing body of evidence suggests that many types of cells regulate their growth rate in vivo by regulating their state of differentiation (1-3). Consequently, investigation of the functional relationship between cellular transformation and differentiation in a defined in vitro system may provide more physiologically relevant data than the study of each one in isolation. The ras genes are members of a ubiquitous eukaryotic multigene family encoding small guanine nucleotide (GTP, GDP)-binding proteins that are involved in cellular signaling (4). The observation that serum growth factor receptors, as well as receptorand non-receptor-type oncogene products with tyrosine kinase activity require Ras function to elicit mitogenic and oncogenic responses (5-8) suggests that Ras acts downstream of tyrosine kinases in proliferative pathways. Aside from its role in mitogenesis, Ras has also been shown to play a key role in a number of differentiation systems, including insulin-induced adipocytic differentiation (9-i 1). The peptide hormone insulin exerts a mitogenic effect on fibroblasts and cycling preadipocytes, whereas treatment of quiescent preadipocytes with insulin leads to a potent acceleration of adipocytic differentiation (i 2). A compelling body of evidence suggests that c-Ras is an integral and essential component of both proliferative and differentiationpromoting insulin signaling pathways. In fibroblasts, insulin treatment activates c-Ras and increases its rate of GTP/GDP exchange (5, 13, 14), whereas introduction of the ,. 5asn1 7 or rasIYsl2.serl8o dominant-negative mutants abrogates the insulin mitogenic signal (15, 16). In addition, c-Ras downregulation, through expression of its glyi2,asni 7 or lysi2,seri86 dominant-negative mutants in a non-ras-transformable 3T3 Li subline (1 5), blocks insulin-induced adipocytic differentiation. A system in which the insulin signal can be variably interpreted as differentiative or proliferative provides an opportunity to study the interrelatedness of these two processes. Because Ras participates in both types of responses (9, 15, 16), it may provide a molecular handle for the specificity of that signal. Previous reports investigated the effect of v-ras expression in adipogenesis (9, 17). In those studies, v-tas expression led to the induction of adipocytic differentiation, obviating the insulin requirement. Those experiments were carried out using lines (REF52 fibroblasts or a subclone of 3T3 Li), which are notable for the fact that they cannot be transformed by v-Ras. Consequently, although invaluable for the study of the differentiative function of v-Ras in isolation, such systems are inherently limited in their capacity to establish the relationship between the proliferative/transforming function of Ras and its role as an integral component of 12 Ras, Raf, Insulin, and Adipocytic Differentiation the differentiative insulin signal pathway. To address this question, we investigated the effect of activated rasIeUJ6l expression in the prototype 3T3 Li line, which is transformable by v-Ras. Our results demonstrate that eu6l expression in this line or in a series of C3H1OT#{189}(iOT#{189})-derived preadipocytes causes a quantitative reciprocal regulation of differentiation versus proliferation, with loss of insulin-dependent differentiative capacity and acquisition of several transformation-associated properties, in response to graded levels of Ras 61 . Anchorage-independent growth and stimulation of proliferation required the highest levels of Ras ’61, whereas morphological transformation and loss of differentiative capacity were apparent even at the lowest Ras levels. Moreover, ras expression in tightly regulatable raS u6l transfectants following the decision-making step of adipocytic differentiation (growth arrest) failed to induce differentiation in the absence of insulin, in the face of a complete refractoriness to transformation, demonstrating that the inability of Ras ’61 to substitute for the insulin signal in this system can clearly be uncoupled from its transforming effect. These results demonstrate that the timing of ra&#{176}’ 61 expression relative to the decision-making step of adipocytic differentiation is critical in determining the dominance hierarchy between transformation and differentiation. Furthermore, although c-Ras activation is a necessary requirement for insulin signal transduction, ras ’61 expression alone does not appear to be sufficient for induction of adipocytic differentiation in this system. Results Production of Preadipocytic Lines Expressing Varying Ras Levels. To examine the relationship between transformation and adipocytic differentiation, we expressed ras 61 in the 3T3 Li preadipocytic cell line (ATCC CCL92.i), which is transformable by activated Ras. To obtain a fine-tuned control of ras gene transcriptional regulation, we developed an inducible system of expression that relies on a combination of relief by IPTG3 of negative control by the iac operator-repressor system, with positive induction by heavy metal inducers of the hMT-llA promoter (1 8). The raS 61 gene was cloned downstream from the hMT-IIA lacO promoter in the hMT-IlAlacO(-55)CAT plasmid (1 8). This plasmid was named hMT(IacO)ras and was transfected into 3T3 Li cells, using the calcium phosphate procedure (19). The pY3 plasmid, expressing the gene conferring resistance to hygromycin B (20), was cotransfected and offered a selectable marker (see “Materials and Methods”). After selection, resistant colonies were picked, expanded into clones, and screened by Northern blotting for ras gene expression before or after promoter induction. To achieve better induction ratios, three inducible clones were supertransfected with the iaci gene, coding for the Lac repressor, with G4i 8-resistance coselection, using the pSV2-neo plasmid (21). A number of 3 The abbreviations used are: IPTG, isopropyl-/3-D-thiogalactoside; Dex, dexamethasone; hMT-llA, human metallothionein IIA; lacO, lac operator; GPD, glycerophosphate dehydrogenase; TGF-j3, transforming growth factor j3. G4i 8-resistant colonies were picked and screened for gene expression by Northern blotting, before or after promoter induction with 1 j.u i Dex, 50 M ZnSO4 and 20 m i IPTG. Three clones, with induced Ras levels similar to the 2Hi, ras-transformed line (21), and approximately 50 times lower background levels, were designated as LraslacOllIC4.a, LraslacOllA6-27, and LraslacOIIA6.i , respectively, and chosen for additional experiments. To follow the dose response of induction, ras-specific RNA levels were measured by Northern blotting in the three highly inducible clones by adding increasing amounts of Dex to the culture medium, in conjunction with 50 M ZnSO4 and 20 mr,i IPTG. A careful quantitation indicated that 1 M Dex can fully induce the promoter under these conditions, whereas the background is approximately 2% of the fully induced levels (e.g., clone LrasIacOlIA6-27; Fig. 1A, Fig. 2, and Table 1). To confirm that the phenotypic effect is indeed a consequence of increased ,. 5Ieu61 expression, rather than a nonspecific effect of the inducers, the leucine-6i -activated ras gene was introduced into 3T3 Li cells under control of the strong constitutive Moloney murine leukemia virus promoter, through infection with the culture supernatant from the NiJ6 packaging line, which secretes a virus carrying the ras 61 gene (see “Materials and Methods”). A number of resistant colonies were isolated, expanded into clones, and tested for ras gene expression levels by Northern blotting as above. Twelve colonies with varying Ras 61 levels were named LV1 to LV1 2 and chosen for additional experiments (Table 1). Reciprocal Regulation of Transformation and Differentiation by Ras. Previous results demonstrated a quantitative regulation of the manifestation of transformation-related properties in response to graded levels of expression of the middle tumor antigen of polyoma virus or v-src in murine fibroblasts (19, 22). Therefore, the highly inducible ras ’ 61expressing clones described above were used to examine the dependence of various parameters of transformation on the levels of ras ’61 gene expression, as follows. Anchorage-independent Growth and Growth Rate Acceleration Require Maximal Levels of ras Gene Expression. ras 61 regulatable LrasIacOlIA6-27 cells were plated in soft agar containing varying amounts of promoter inducers (see “Materials and Methods”). As shown in Fig. 2, upon maximal ras induction, resulting in Ras levels comparable to those present in the 2H1 , ms-transformed line (21), cells grew as efficiently in semisolid nutrient agar as did the 2Hi cells themselves, whereas no growth was observed in the absence of the inducers. Intermediate induction levels, obtamed through the addition of lower amounts of Dex as described above, were used to examine this dependence further. As shown in Fig. 2, some agar growth was observed when Ras levels were approximately 40% of the levels in the 2Hi line, and the growth response increased roughly proportionally to the amount of Ras present in the cell, up to 100%. Similarly, their growth rate on plastic increased with increasing Ras 61 levels with doubling times from 26.4 to 19 h, in a manner that paralleled their acquisition of anchorage-independent growth (Fig. 2). Similar results were obtamed with the other two inducible ras ’61-expressing lines (Table 1) and a mixed culture of hMT(lacO)ras-transfected