Comparison of VEGF, VEGF-B, VEGF-C and Ang-1 mRNA ... - Nature

0 downloads 0 Views 601KB Size Report
VEGF-C, but not VEGF-B mRNA expression. Interest- ingly, these growth factors and hypoxia simultaneously downregulated the mRNA of another endothelial ...
Oncogene (1997) 14, 2475 ± 2483  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

Comparison of VEGF, VEGF-B, VEGF-C and Ang-1 mRNA regulation by serum, growth factors, oncoproteins and hypoxia Berndt Enholm1,*, Karri Paavonen1,*, Ari RistimaÈki2, Vijay Kumar1, Yuji Gunji1, Juha Klefstrom1, Laura Kivinen3, Marikki Laiho3, Birgitta Olofsson4, Vladimir Joukov1, Ulf Eriksson4 and Kari Alitalo1 1

Molecular/Cancer Biology Laboratory and 3Department of Virology, Haartman Institute, PL21 (Haartmaninkatu 3), 00014 University of Helsinki, Finland; 2Departments of Clinical Chemistry and Obstetrics and Gynecology, University of Helsinki, PL22 (Haartmaninkatu 2), 00290 Helsinki, Finland; 4Ludwig Institute for Cancer Research, Stockholm Branch, Box 240, S-171 77 Stockholm, Sweden

The vascular endothelial growth factor (VEGF) family has recently been expanded by the isolation of two additional growth factors, VEGF-B and VEGF-C. Here we compare the regulation of steady-state levels of VEGF, VEGF-B and VEGF-C mRNAs in cultured cells by a variety of stimuli implicated in angiogenesis and endothelial cell physiology. Hypoxia, Ras oncoprotein and mutant p53 tumor suppressor, which are potent inducers of VEGF mRNA did not increase VEGF-B or VEGF-C mRNA levels. Serum and its component growth factors, platelet-derived growth factor (PDGF) and epidermal growth factor (EGF) as well as transforming growth factor-b (TGF-b) and the tumor promoter phorbol myristate 12,13-acetate (PMA) stimulated VEGF-C, but not VEGF-B mRNA expression. Interestingly, these growth factors and hypoxia simultaneously downregulated the mRNA of another endothelial cell speci®c ligand, angiopoietin-1. Serum induction of VEGF-C mRNA occurred independently of protein synthesis; with an increase of the mRNA half-life from 3.5 h to 5.5 ± 6 h, whereas VEGF-B mRNA was very stable (T1/248 h). Our results reveal that the three VEGF genes are regulated in a strikingly di€erent manner, suggesting that they serve distinct, although perhaps overlapping functions in vivo. Keywords: VEGF; VEGF-B; VEGF-C; angiopoietin; angiogenesis

Introduction Angiogenesis, the sprouting of new blood vessels from pre-existing ones, is a complex process. During angiogenic sprouting, perivascular matrix proteins are degraded by proteases, endothelial cells migrate towards the angiogenic stimuli, proliferate and form the lumen of a new microvessel (Hanahan and Folkman, 1996). Angiogenesis is essential in normal physiological processes such as wound healing, tissue and organ regeneration and female reproductive

Correspondence: K Alitalo. e-mail: Kari.Alitalo@Helsinki.® *BE and KP should be considered equal ®rst authors and have been listed in alphabetical order. Received 4 December 1996; revised 10 February 1997; accepted 10 February 1997

functions. Angiogenesis is also involved in pathological processes such as the growth of tumors beyond the microscopic stage and subsequent metastasis, in rheumatoid arthritis, certain retinopathies and psoriasis (Folkman, 1995). Di€erent growth factors and cytokines may act alone or in synergy as stimuli for angiogenesis. Vascular endothelial growth factor (VEGF), ®broblast growth factors, hepatocyte growth factor, transforming growth factors (TGFs) and tumor necrosis factor-a (TNF-a) have been shown to induce angiogenesis in a variety of experimental models (Folkman, 1995; Dvorak et al., 1995). The newly formed vessels are apparently stabilized through the migration of pericytes, paracrine signalling and intercellular basement membrane matrix production. Some of these processes have been suggested to be mediated through angiopoietin-1 (Ang-1) via its endothelial cell speci®c Tie-2 receptor and plateletderived growth factor (PDGF)-BB via its receptor on pericytes (C. Betsholtz, personal communication; Davis et al., 1996; Dumont et al., 1992; Suri et al., 1996, Vikkula et al., 1996; Folkman and D'Amore, 1996). VEGF, also known as vascular permeability factor, is an important angiogenic agent (Thomas, 1996). Although encoded by a single gene, VEGF has several isoforms generated by alternative splicing. Of these, VEGF121 and VEGF165 are secreted soluble glycoproteins, whereas VEGF189 remains bound to heparan sulphate proteoglycans at the cell surface. VEGF regulates the normal development of the embryonic vasculature and apparently also the early differentiation of mesenchymal cells into hemangioblasts (Risau, 1995; Fong et al., 1995; Carmeliet et al., 1996). These precursor hemangioblasts are essential for embryonic vasculogenesis, especially in the formation of the dorsal aorta and blood-islands of the yolk sac (Breier et al., 1992). Transgenic mice with one disrupted VEGF allele have impaired blood vessel formation and die in utero, suggesting that VEGF concentrations may be critical in vasculogenesis and angiogenesis (Carmeliet et al., 1996; Ferrara et al., 1996). Tumors establish their microvasculature in part by secreting elevated amounts of VEGF. Increased levels of VEGF mRNA are found in hypoxic areas of many solid tumors (Plate et al., 1992; Shweiki et al., 1992). The hypoxia-induced rise in the steady-state levels of the VEGF mRNA occurs partly due to increased transcription, but mostly due to post-transcriptional mechanisms, including an increase in mRNA stability (Goldberg and Schneider, 1994; Stein et al., 1995).

Regulation of VEGF-B and VEGF-C B Enholm et al

In order to analyse growth factor regulation of the VEGF-B and VEFG-C genes, serum starved human IMR-90 ®broblasts were exposed to 5% fetal calf serum (FCS) for 6 h, after which the cells were lysed and total RNA was isolated, electrophoresed, blotted and hybridized with VEGF-B and VEGF-C probes and, for comparison, with a VEGF probe. The results of this analysis (Figure 1a), indicate that VEGF-C mRNA is increased, but VEGF-B mRNA is not responsive to serum stimulation. Of interest is also the broader VEGFC mRNA band in serum stimulated cells than in unstimulated cells, mainly due to the appearance of faster migrating RNA species. In contrast, the steadystate mRNA levels for Ang-1 (Davis et al., 1996), were relatively high under basal conditions but downregulated by serum treatment. Figure 1b shows a densitometric evaluation of the serum-induced VEGF-C mRNA steady-state levels (4.8+1.0-fold increase from ®ve independent experiments). This analysis also indicates that the unstimulated steady-state levels of VEGF-B mRNA are considerably higher than those for VEGF

0.5 5

0.5 5

0.5 5

28S —

18S — ➝

Regulation of mRNAs of the VEGF family by serum and growth factors in human ®broblasts

FCS (%) 0.5 5

➝ ➝

Results

a



Growth factors and certain cytokines also increase VEGF mRNA steady-state levels in e.g. keratinocytes (Frank et al., 1995). This may be relevant in processes, such as wound healing and in¯ammation, where angiogenesis and vascular permeability changes play major roles. In addition, induction of VEGF mRNA and protein is obtained, when cells are treated with TGF-b (Pertovaara et al., 1994), epidermal growth factor (EGF) (Frank et al., 1995), PDGF (Finkenzeller et al., 1992), interleukin (IL)-6 (Cohen et al., 1996) or a tumor promoting phorbol ester (Garrido et al., 1993). VEGF-B and VEGF-C are two recently discovered members of the VEGF family (Olofsson et al., 1996; Joukov et al., 1996). This family of growth factors thus far comprises four members: VEGF, VEGF-B, VEGFC and the placenta growth factor (PlGF) (Maglione et al., 1991). The members of the VEGF gene family are all located in di€erent human chromosomes (Vincenti et al., 1996; Paavonen et al., 1996; Maglione et al., 1993). VEGF-C was discovered as the ligand for the third member of the VEGF receptor family, which is expressed mainly on venous endothelium of early embryos and lymphatic endothelium of adult tissues (Kukk et al., 1996; Kaipainen et al., 1995; Joukov et al., 1996). VEGF-C is expressed at low levels in many tissues, most prominently in the heart, placenta, skeletal muscle, ovary and the small intestine. VEGFC mRNA also occurs in certain tumor cell lines. VEGF-B is also expressed in most tissues, but especially highly in the heart and skeletal muscle (Olofsson et al., 1996). The existence of a gene family consisting of several related growth factors suggests that these factors have overlapping but distinct functions and that they may be di€erentially regulated. We have therefore studied the regulation of the mRNAs for the VEGF-B and VEGF-C genes.



2476

VEGF

VEGF B

VEGF C

Ang-1

b

Figure 1 E€ect of FCS on VEGF, VEGF-B, VEGF-C and Ang-1 mRNA expression. (a) IMR-90 cells were incubated with 0.5% or 5% FCS for 6 h. Total RNA was isolated and analysed by Northern blot hybridization. Arrows show the gene-speci®c hybridization signals. The mobilities of the 28S and 18S ribosomal RNA bands are indicated (b) Quantitation of the autoradiographic signals from ®ve di€erent experiments. The values represent arbitrary densitometric units (means+s.e.m.). The asterisk indicates a signi®cant (P50.05) di€erence. The values have been normalized for the GAPDH signals

and VEGF-C in unstimulated conditions. This is consistent with the fact that VEGF-B mRNA was found to have the longest half-life of the three mRNAs. Kinetics and growth factor dose-dependence of VEGF mRNA regulation In order to understand the kinetics and dose-dependence of mRNA regulation by growth factors, IMR-90 cells were stimulated for various periods of time using FCS or the growth factors PDGF, EGF or TGF-b, which have been shown to increase VEGF mRNA expression in various cells. As shown in Figure 2, FCS induced the greatest increase of VEGF-C transcript levels of the three VEGF genes, starting at 4 h of stimulation and continuing throughout the experiment (24 h). In contrast, while VEGF mRNA was also increased, transcript levels returned to nearly unstimulated values after 8 h. PDGF stimulated VEGF-C and VEGF expression with a similar kinetics compared to FCS. Strikingly, EGF, which has relatively few receptors in ®broblastic cells also stimulated VEGF-C mRNA expression, but had only a modest e€ect on VEGF. Further speci®city is shown by the fact that TGF-b signi®cantly stimulated only VEGF mRNA levels in

Regulation of VEGF-B and VEGF-C B Enholm et al

2477

a

PDGF

FCS 0

2

4

8

24

2

4

8

EGF 24

2

4

8 24

TGF-β 2

4

8

24

h — 28S

VEGF — VEGF-C — VEGF-B —

— 18S

Ang-1 —

GAPDH —

b

Figure 2 E€ect of growth factor stimulation time on VEGF, VEGF-B, VEGF-C and Ang-1 mRNA levels. (a) The cells were stimulated with 5% FCS or with 100 ng/ml PDGF, 100 ng/ml EGF or 10 ng/ml TGF-b in 0.5% FCS for the indicated timeperiods. Total RNA was isolated and analysed by Northern blot hybridization with the indicated probes. (b) Quantitation of the VEGF and Ang-1 gene speci®c signals using arbitrary densitometric units. VEGF is represented by open squares, VEGF-C by open triangles, VEGF-B by closed squares and Ang-1 by closed triangles. The values have been normalized for the GAPDH signals

these ®broblastic cells. In experiments where the three growth factors were added in combination to the IMR90 cells, the resulting e€ect at 4 h was similar to that of FCS (data not shown). In contrast, VEGF-B mRNA levels were only minimally a€ected by FCS or the growth factors and Ang-1 mRNA was consistently downregulated by about 50 ± 70% after 8 h treatment with serum or the growth factors.

Studies of the growth factor concentration-dependence of mRNA regulation are shown in Figure 3. These results indicate that maximal levels of VEGF-C expression are obtained by relatively low concentrations of the three di€erent growth factors, although serum shows a dose response e€ect at all concentrations tested. Maximal PDGF-induced expression of VEGF-C mRNA was about half of the level obtained by FCS stimulation,

Regulation of VEGF-B and VEGF-C B Enholm et al

and then transferred to medium containing 5% FCS for 4 h. Thereafter, half of the plates were returned to medium containing 0.5% FCS and actinomycin D was added to all cultures to inhibit further RNA synthesis. Messenger RNA was isolated and analysed after various time periods. The half-life of VEGF mRNA was about 1 h in IMR-90 cells in serum deprived conditions and somewhat prolonged in the presence of serum. VEGF-C mRNA had a half-life of about 3.5 h in the starvation medium and 5.5 ± 6 h in the presence

1

0.1

100

EGF 10

1

0.1

0.01

1000

TGF-β 100

1

0.1

PDGF 15

5

(%)

1.5

FCS

10

a

1000

The half-lives of the di€erent mRNAs were analysed in IMR-90 cells. Cells were starved for 48 h in 0.5% FCS

100

Stability of VEGF family mRNAs and e€ect of protein synthesis inhibitors

10

while VEGF mRNA levels were upregulated equally by PDGF and FCS (see Figure 3b). The VEGF-B mRNA signal remained relatively constant with all concentrations of the growth factors studied.

0.5

2478

(ng/ml) — 28S

VEGF — VEGF-C — — 18S VEGF-B — GAPDH —

b

Figure 3 FCS and growth factor concentration dependence of VEGF, VEGF-B and VEGF-C mRNA regulation. (a) IMR-90 cells were incubated with FCS (0.5 ± 15%) or with PDGF (0.1 ± 1000 ng/ml), EGF (0.1 ± 1000 ng/ml) or TGF-b (0.01 ± 100 ng/ml) in 0.5% FCS for 6 h and analysed for the respective RNAs as in Figure 2a. (b) Quantitation of the VEGF gene speci®c signals as in Figure 2b

Regulation of VEGF-B and VEGF-C B Enholm et al

of serum (Figure 4a). VEGF-B mRNA did not decay signi®cantly during the analysed time periods. Cycloheximide (CHX), added to cultures of IMR-90 cells to inhibit ongoing protein synthesis during serum stimulation did not prevent VEGF-C mRNA induction (Figure 4b). However, an increase in the VEGF-C mRNA signal was seen in starved cells treated with CHX. CHX also increased VEGF mRNA levels in serum-starved cells although an additional e€ect was seen when it was added together with serum. Again, VEGF-B mRNA levels were unchanged during the experiment. Tumor promoter PMA and Ras oncoprotein as regulators of VEGF genes The VEGF promoter has been shown to contain AP-1 binding sites, and VEGF mRNA expression is known to be upregulated by PMA and the Ras oncoprotein (Grugel et al., 1995; Finkenzeller et al., 1995). As can be seen in Figure 5a, IMR-90 ®broblasts and HT-1080 ®brosarcoma cells exposed to 40 ng/ml PMA for 4 h or longer showed increased VEGF and VEGF-C mRNA

levels. Furthermore, the latter mRNA band showed a faster migrating species. In contrast, VEGF-B mRNA was unaltered. In the ®brosarcoma cells, the unstimulated levels of VEGF mRNA were higher than in the ®broblasts. Furthermore, while the PMA-induced increase in VEGF and VEGF-C mRNAs was downregulated at 24 h in ®broblasts, it persisted in the ®brosarcoma cells. Cells transfected with mutant Ras oncogenes have been shown to express elevated levels of VEGF (Mazure et al., 1996). We measured the levels of VEGF, VEGF-B and VEGF-C in transfected NIH3T3 cells expressing the inducible v-Ha-Ras plasmid construct detailed in the Materials and methods section. After 48 h of treatment with 10 mM isopropylthiogalactoside (IPTG) to activate transcription of the v-Ha-Ras construct, there was a strong induction of VEGF mRNA, as has been described previously (Mazure et al., 1996). However, the expression of VEGF-B and VEGF-C mRNAs was unaltered in these conditions (Figure 5b). Regulation of VEGF genes by hypoxia Decreased partial oxygen pressure in tissues has been shown to be an important regulator of VEGF in angiogenesis (Stein et al., 1995). We were therefore interested in examining the possibility that the novel

a

PMA

IMR-90

HT-1080

b

0

2

4

8

24

h

4.5 — 3.7 —

— VEGF —

2.4 —

— VEGF-C —

1.4 —

— VEGF-B

4.5 — 3.7 —

— VEGF —

2.4 —

— — VEGF-C

1.4 —

— VEGF-B

kb

— β-actin

c-Ha-Ras



+

28S —

18S —

— VEGF — VEGF-C — — VEGF-B

CHX FCS

– –

– +

+ –

+ +

Figure 4 (a) Decay of VEGF-C mRNA in the presence of actinomycin D in IMR-90 cells. The graph shows the decay of VEGF-C mRNA in the presence of 0.5% or 5% FCS supplemented with 10 mg/ml of actinomycin D. Total RNA was analysed by Northern blotting and the mRNA hybridization signals were quanti®ed using densitometric scanning. The values in the graph indicate the percentage of initial VEGF-C mRNA signal remaining in the speci®ed conditions. (b) E€ect of CHX on serum stimulation of VEGF family mRNAs. Starved cells were incubated in the indicated conditions for 6 h, after which the mRNAs were analysed

— β-actin Figure 5 E€ect of PMA and Ras oncoprotein on the expression of VEGF family mRNAs. (a) IMR-90 and HT-1080 cells were incubated overnight in the absence of serum and then treated with 40 ng/ml PMA for various periods of time. Five mg of isolated cellular poly(A)+ mRNA was analysed by Northern blot hybridisation. (b) Cells containing an inducible Ras construct were treated with IPTG for 48 h, and total cellular RNA was analysed. Control experiments indicated that IPTG induction did not increase VEGF mRNA levels in the neo-expressing control cells (data not shown)

2479

Regulation of VEGF-B and VEGF-C B Enholm et al

endothelial growth factor genes VEGF-B and VEGF-C might also be regulated by hypoxia. Rat C6 glioblastoma cells, which have previously been shown to upregulate expression of VEGF mRNA in low oxygen conditions were used for the experiment (Stein et al., 1995; Shweiki et al., 1992). The results, shown in Figure 6a indicate that hypoxia strongly induces VEGF mRNA expression, but has no signi®cant e€ect on VEGF-B mRNA levels. Although the major VEGF-C mRNA band of 2.4 kb was not markedly induced, an extra band of faster mobility was seen in the analysis of the hypoxia sample. Similar results were obtained in experiments with the human HT-1080 ®brosarcoma cell line, although basal expression of VEGF mRNA was high and induction was somewhat less pronounced (data not shown). Interestingly, Ang-1 mRNA was decreased under the hypoxic conditions in C6 cells (Figure 6a and data not shown).

a

Hypoxia



+

4.5 — 3.7 —

— VEGF —

2.4 —

— VEGF-C —

1.4 —

— VEGF-B

1.5 —

— Ang-1

kb

— WAF-1

b Hypoxia

0

2

4

8

24

h

— VEGF — — — VEGF-C 18S — — VEGF-B — β-actin

Hypoxia 4.5 — 3.7 — 2.4 — 1.4 — kb

p53 –/–

DN p53





+

E€ect of p53 on the hypoxic regulation of VEGF family mRNAs A temperature sensitive mutant of the p53 tumor suppressor protein upregulates VEGF gene expression in the mutant conformation, which blocks the transcriptional activation function of p53 (Kieser et al., 1994). We wanted to know whether the presence of the p53 tumor suppressor protein a€ects the hypoxic regulation of the VEGF, VEGF-B and VEGF-C genes. Embryonic ®broblasts from p537/7 mice (Donehower et al., 1992) were subjected to hypoxia for 16 h and analysed. In addition, Rat-1 ®broblasts expressing a dominant-negative form of p53 (Shaulian et al., 1992; Klefstrom et al., manuscript in preparation) and parental Rat-1 ®broblasts were treated in the same manner. As can be seen from Figure 6c, upon exposure to hypoxia all these cells showed increased VEGF mRNA expression but no change in VEGF-B or VEGF-C mRNA levels. Expression of VEGF family mRNAs in human tumor cell lines

28S —

c

In order to test whether normal cells also show a similar regulation of VEGF mRNAs, human IMR-90 ®broblasts were subjected to hypoxia treatment. As can be seen from Figure 6b, the hypoxic indution of VEGF mRNA occurs already at 8 h of treatment and is further increased at 24 h. In contrast, the VEGF-B and VEGF-C mRNA levels were not substantially altered.

We also analysed a group of human tumor cell lines cultured in the presence of 10% FCS for the expression of VEGF, VEGF-B and VEGF-C mRNAs. Among 12 tumor cell lines, including several leukemia cell lines, VEGF-B mRNA was detected in all and VEGF mRNA in most samples (Figure 7). In contrast, VEGF-C mRNA levels were low or undetectable in most of the cell lines, being abundant only in the PC-3, WI26, A549 and A498 cell lines and in the normal lung ®broblasts (WI38).

PC-3 HeLa WI38 WI26 G-401 A549 CHRF CEM DAMI HEL K562 MOLT4 U937 A498

2480

+ — — VEGF — — VEGF-C — VEGF-B β-actin

Figure 6 Regulation of VEGF, VEGF-B, VEGF-C and Ang-1 mRNAs by hypoxia. (a) C6 rat glioma cells were exposed to normoxia or hypoxia for 18 h. Poly(A)+ mRNA was isolated and 10 mg was analysed by Northern blotting and hybridization with a mixture of probes for VEGF, VEGF-B and VEGF-C or with Ang-1. Hybridization with a probe for the rat WAF-1 (p21) cyclin kinase inhibitor (a kind gift from Dr Bert Vogelstein; el-Deiry et al., 1993) RNA was used to control for equal loading of mRNA. In addition the ®lter was later controlled for equal loading using a probe for b-actin. (b) IMR-90 human ®broblasts were treated with hypoxia for the indicated periods and analysed as in (a). (c) Mouse embryo ®broblasts and RAT-1 cells expressing a dominant negative p53 with a homozygous deletion of p53 and RAT-1 cells expressing a dominant negative p53 were incubated in hypoxic conditions for 16 h, 20 mg of total RNA was isolated and analysed by Northern blot hybridization with the probes indicated

kb 4.5 — 3.7 — 2.4 —

— VEGF —

1.4 —

— VEGF-B

— VEGF-C

— β-actin Figure 7 Expression of VEGF, VEGF-B and VEGF-C mRNAs in tumor cell lines. Steady-state levels of mRNAs were analysed in a panel of human tumor cell lines. The cell lines represented are: PC-3 prostatic adenocarcinoma, HeLa cervical epitheloid carcinoma, WI38 diploid lung ®broblasts, WI26 SV40 transformed epithelial-like lung cells, G-401 Wilms' tumor, A549 lung carcinoma, CHRF megakaryoblastic leukemia, CEM T-lymphoblastoid leukemia, DAMI megakaryoblastic leukemia, HEL erythroblastic leukemia, K562 chronic myelogenous leukemia, MOLT4 T-cell leukemia, U937 histiocytic lymphoma, A498 kidney carcinoma

Regulation of VEGF-B and VEGF-C B Enholm et al

Discussion We studied the regulation of the mRNAs of two novel members of the VEGF family in response to several cellular stimuli. Several growth factors, such as TGF-b, EGF, and PDGF-BB, have been found to induce VEGF mRNA expression and may thus contribute to angiogenic stimuli and changes in vascular permeability. Serum, which contains di€erent growth stimulatory agents, induced an approximately ®vefold increase of VEGF-C mRNA in starved human ®broblasts. Serum also upregulated VEGF mRNA, as has been shown previously (Frank et al., 1995). VEGFC mRNA levels were also increased in response to PDGF and, to a lesser extent by EGF and TGF-b. Our unpublished experiments indicate that the changes of mRNA regulation are also re¯ected at the level of biosynthesis and secretion of the VEGF-C protein. As PDGF is a major connective tissue mitogen in serum and is released from platelets upon tissue injury, our mRNA results suggest that VEGF-C could be induced in vivo in wounds and contribute to repair of tissue injury, similar to what has been described for VEGF induction by PDGF in ®broblasts (Finkenzeller et al., 1992) and by keratinocyte growth factor in epidermal keratinocytes (Frank et al., 1995). In contrast, Ang-1 mRNA was decreased upon serum stimulation. This is consistent with the possibility borne out from Tie-2 and Ang-1 gene knockout and mutant Tie-2 studies (Suri et al., 1996; Vikkula et al., 1996) that Ang-1 has no role in the immediate response to growth factors released during the wound repair. Instead, Ang-1 could have a role later on in neovascularization processes, perhaps during the stabilization of the newly formed blood vessels by pericyte migration and the formation of a basement membrane matrix (see Folkman and D'Amore, 1997). We also investigated the e€ect of Ras oncogene activation on the steady-state mRNA levels of VEGF, VEGF-B and VEGF-C. However, induction of Ras only elevated the levels of VEGF mRNA, as reported previously (Mazure et al., 1996). This is interesting, as many growth factor signals are relayed via the Grb2Shc-Sos-Ras pathway and constitutive activation of Ras is considered to partially mediate the mitogenic signals induced by serum, including the upregulation of many serum-responsive genes (Pawson, 1995). Furthermore, Ras-regulated elements include AP-1 transcription factor binding sites, which are found in the promoters of the VEGF (Tischer et al., 1991) and VEGF-C genes (unpublished data of Dimitri Chilov and the authors). Upregulation of VEGF and VEGFC mRNA and protein levels (our unpublished observations) upon stimulation with PMA are consistent with the possibility that the AP-1 elements are involved in regulating VEGF and VEGF-C in IMR-90 cells, while in HT-1080 cells, mutant N-ras oncogene (Brown et al., 1984) may account for some of the di€erences in mRNA regulation upon PMA treatment. However, more detailed analyses of the mechanism of serum-induced VEGF-C mRNA expression such as RNA stability measurements and analysis of the expression of VEGF-C promoter-reporter constructs would be needed to de®ne the relevant signalling pathways, the responsive transcription factors and the binding sites.

VEGF-B mRNA is expressed at low a low level in many tissues, being abundant only in the heart, skeletal muscle and brown fat, tissues with high metabolic demands (Lagercrantz et al., 1996; Olofsson et al., 1996). None of the growth factors nor serum induced changes in VEGF-B mRNA levels. This raises interesting questions about the high and stable constitutive expression of VEGF-B mRNA in cultured cells. The strong, invariant expression of the VEGF-B gene in cultured cells suggests that its mRNA might already be upregulated due to explantation of cells into tissue culture. VEGF mRNA has a half-life of less than 1 h (Shima et al., 1995, and the present results). In contrast, the half-life of VEGF-B mRNA was found to be greater than 8 h. This is consistent with the fact that VEGF-B mRNA did not respond to stimulation with growth factors or serum. Indeed, rapid fluctuations of steady-state mRNA levels require a short mRNA half-life that can be regulated via changes in transcription rate and RNases and their inhibitors, as in the case of VEGF. VEGF-C mRNA half-life was estimated to be about 3.5 h in serum-starved lung ®broblasts. The half-life was increased to 5.5 ± 6 h in the presence of serum, suggesting that serum may increase the transcription rate of the VEGF-C gene and maintain a cellular factor slightly increasing its mRNA stability. According to its induction pattern, requiring 2 ± 4 h of serum stimulation but not ongoing protein synthesis, the VEGF-C gene may then be classi®ed as an immediate early serum response gene (Nathans et al., 1988). Hypoxia is the most speci®c signal inducing transcription of the VEGF gene and its mRNA stabilization (Stein et al., 1995). In contrast, VEGF-B and VEGF-C mRNAs did not appear to be regulated by hypoxia, nor was PlGF mRNA hypoxia sensitive (Gleadle et al., 1995) and Ang-1 mRNA was even decreased in the present experiments. Hypoxia, as well as serum and particularly PMA induced a faster migrating RNA species, which hybridized with the VEGF-C probe. The identity of this band is currently unknown, although our unpublished studies and those of others (Lee et al., 1996) have revealed that alternatively spliced forms of VEGF-C mRNA lacking exons 2 ± 4 or only exon 4 may exist in small amounts. Alternatively, this band could represent a crosshybridizing transcript. Furthermore, the possibility exists that the translation or secretion of VEGF factors or the proteolytic activation of the VEGF-C (Joukov et al., 1996) are altered in hypoxic conditions. These possibilities will be further investigated. It has been proposed that tumor cells harbouring a mutant p53 tumor suppressor gene are more resistant to apoptosis induced by tissue hypoxia, thus implying that hypoxic stress can select for tumor cells that have lost their apoptotic program. Transfection of an expression vector for the wild-type p53 oncoprotein into tumor cells has been shown to downregulate the expression of VEGF (Mukhopadhyay et al., 1995). Since hypoxia increases the levels of the p53 protein (Graeber et al. 1994), the question arises whether cells de®cient of p53 could display an enhanced VEGFresponse upon hypoxia (Graeber et al., 1996). We therefore studied the hypoxia-induced expression of genes of the VEGF family in cells de®cient for p53.

2481

Regulation of VEGF-B and VEGF-C B Enholm et al

2482

Fibroblasts de®cient for p53 retained their VEGF mRNA inducibility in response to hypoxia, but VEGFB and VEGF-C mRNA levels remained unaltered. Taken together, these results reveal the complexity of regulation of the VEGF gene family. Instead of one VEGF we need to consider at least three distinct e€ectors of vascular proliferation, which have differences in their polypeptide structures, homo- and heterodimerization properties, proteolytic processing, interaction with cell surface heparan sulfate proteoglycans, secretion and solubility, as well as receptor binding properties. As shown here, there are significant di€erences in the regulation of the steady-state mRNAs for VEGF, VEGF-B and VEGF-C. While VEGF has been shown to be controlled mainly by stabilization of its mRNA and to a lesser degree by transcriptional induction by growth factors, oncogenes and hypoxia, the VEGF-C mRNA is also highly inducible by several growth factors, but not by oncogenes or hypoxia and VEGF-B mRNA is unusually stable in cultured cells in vitro and its steady-state levels were not regulated in our experiments. These results suggest that similarities and di€erences may occur also in the expression of the three genes in vivo. Such mechanisms would maintain an elaborate heterogeneity of the vascular endothelium and a complex network of speci®c, but partially redundant signalling to keep blood vessel growth and their di€erentiated functions in check. In tumors, on the other hand, several molecular forms of VEGFs may be activated, resulting in the induction of tumor angiogenesis. Materials and methods Cell culture and growth factor treatments Diploid human lung ®broblasts (IMR-90, American Type Tissue Collection ATCC CCL-186) were grown in Dulbecco's modi®ed Eagle's medium (DMEM) containing 10% FCS, glutamine and antibiotics. Prior to initiation of the experiments, the cells were maintained in 0.5% FCS for 48 h. Cells were then incubated with human recombinant PDGF-BB (0.1 ± 1000 ng/ml), EGF (0.1 ± 1000 ng/ml), or TGF-b1 (0.01 ± 100 ng/ml) in 0.5% FCS for the time periods indicated. All growth factors were obtained from R&D Systems. Human HT-1080 ®brosarcoma cells and rat C6 glioblastoma cells were cultured as recommended by the supplier (ATCC). Generation and use of conditionally v-Ha-Ras expressing ®broblasts Mouse NIH3T3 ®broblasts (ATCC CRL 1658) were transfected by lipofection (Lipofectamine, GIBCO-BRL) with plasmids pSVlac0ras and pHbINLSneo (kindly provided by N Denko, U Cincinnati Medical Center) or pHbINLSneo. The plasmid pSVlac0ras consists of the SV40 promoter and lac operator enhancer elements driving expression of the human c-Ha-ras (Val-12) oncogene and the plasmid pHINLSneo contains the human b-actin promotor driving the expression of lac repressor and neomycin resistance genes (Liu et al., 1992). Following transfection, NIH3T3 cells were incubated in 0.6 mg/ml G418 (GIBCO-BRL) for 14 days, after which clones were isolated by ring-cloning. Cells transfected with pHbINLSneo and pSVlac0ras were ®rst grown in the presence of G418 for 14 days, after which they were transferred to

0.34% soft agar containing 10% FCS, 10% tryptose phosphate broth, 0.6 mg/ml G418 and 20 mM IPTG (Promega) in 26DMEM and plated on top of 0.5% agar containing 10% FCS, 10% tryptose-phosphate broth, 26DMEM and 10 mM IPTG. After 2 weeks, colonies with a transformed morphology were picked and expanded in the absence of IPTG. Clones used in the experiments were NIH3T3 lac0ras8 and NIH lacIneo4. Hypoxia treatment Con¯uent cultures of tumor cells were grown on 10 cm diameter tissue culture plates containing 10 ml of DMEM and 5% fetal calf serum plus antibiotics. The cultures were exposed to normoxia or hypoxia for 16 h or 18 h. The hypoxic environment was achieved by incubating the cell cultures in an Anaerocult A anaerobic culture jar supplied by Merck. Isolation and analysis of RNA Total RNA was isolated using the Trizol Reagents (GIBCO-BRL). For Northern blots, 20 mg of total RNA was denatured in 1 M glyoxal, 50% dimethylsulfoxide, and 10 mM phosphate bu€er at 508C for 60 min and electrophoresed through 1.2% agarose gels. Alternatively, the RNA was isolated using the RNEasy total RNA isolation kit (Qiagen Gmbh), and electrophoresed through 1.6% formaldehyde-agarose gels. In some cases, the poly(A)+ mRNA fraction was isolated by binding to oligo(dT) cellulose. About 8 mg of the RNA was electrophoresed, the RNA gels were transferred to nylon membranes, which were u.v.-irradiated and then hybridized with a mixture of radiolabeled fragments of human VEGF cDNA (nucleotides nt 57 ± 638, Genbank Acc. No. 15997), VEGF-B167 cDNA (nt 1 ± 382, Genbank Acc. No. U48800) and VEGFC cDNA (nt 495 ± 1661, Genbank Acc. No X94216). The human Ang-1 cDNA probe was a 1.4 kb RT ± PCR fragment containing the open reading frame (Davis et al., 1996). DNAs were labelled using a standard technique involving enzymatic extension reactions of random primers. As controls for equal gel loading, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or b-actin was used for normalisation. The probed blots were exposed to Fuji Medical X-Ray ®lm after quantitation of the signals with Fuji®lm IP reader Bio-Imaging Analyser BAS1500 and the MacBas software supplied by the manufacturer. Analysis of mRNA half-life mRNA half-life experiments were carried out using human IMR-90 lung ®broblasts. The cells were starved in 0.5% FCS DMEM for 48 h and shifted to medium containing 5% FCS for 4 h prior to mRNA stability measurements. Transcription was inhibited by adding actinomycin D (Sigma) at a concentration of 10 mg/ml. For inhibition of protein synthesis, the cells were treated with 10 mg/ml cycloheximide for the indicated periods of time.

Acknowledgements We would like to thank Raili Taavela, Mari HelanteraÈ , Ritva Javanainen and Tapio Tainola for excellent technical assistance. This study was supported by the Finnish Cancer Organizations, the Finnish Academy, the Sigfrid Juselius Foundation, the University of Helsinki and the State Technology Development Centre. BE was supported in part by Finska LaÈkarsaÈllskapet and, KP in part by the Ida Montin Foundation.

Regulation of VEGF-B and VEGF-C B Enholm et al

2483

References Breier G, Albrecht U, Sterrer S and Risau W. (1992). Development., 114, 521 ± 532. Brown R, Marshall C, Pennie S and Hall A. (1984). EMBO J., 3, 1321 ± 1326. Carmeliet P, Ferreira V, Breier G, Pollefeyt S, Kieckens L, Gertsenstein M, Fahrig M, Vandenhoeck A, Harpal K, Ebenhardt C, Declercq C, Pawling J, Moons L, Collen D, Risau W and Nagy A. (1996). Nature, 380, 435 ± 439. Cohen T, Nahari D, Cerem LW, Neufeld G. and Levi BZ. (1996). J. Biol. Chem., 271, 736 ± 741. Davis S, Aldrich TH, Jones PF, Acheson A, Compton DL, Jain V, Ryan TE, Bruno J, Radziejewski C, Maisonpierre PC and Yancopoulos GD. (1996). Cell, 87, 1161 ± 1169. Donehower LA, Harvey M, Slagle BL, McArthur MJ, Montgomery CAJ, Butel JS and Bradley A. (1992). Nature, 356, 215 ± 221. Dumont DJ, Yamaguchi TP, Conlon RA, Rossant J and Breitman ML. (1992). Oncogene, 7, 1471 ± 1480. Dvorak HF, Brown LF, Detmar M and Dvorak AM. (1995). Am. J. Pathol., 146, 1029 ± 1039. el-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R, Trent JM, Lin D, Mercer WE, Kinzler KW and Vogelstein B. (1993). Cell, 75, 817 ± 825. Ferrara N, Carver-Moore K, Chen H, Dowd M, Lu L, O'Shea KS, Powell-Braxton L, Hillan KJ and Moore MW. (1996). Nature, 380, 439 ± 442. Finkenzeller G, Marme D, Weich HA, Hug H. (1992). Cancer Res., 52, 4821 ± 4823. Finkenzeller G, Technau A and Marme D. (1995). Biochem. Biophys. Res. Comm., 208, 432 ± 439. Folkman J. (1995). Nature Med., 1, 27 ± 31. Folkman J and D'Amore PA. (1996). Cell, 87, 1153 ± 1155. Fong GH, Rossant J, Gertsenstein M and Breitman ML. (1995). Nature, 376, 66 ± 70. Frank S, Hubner G, Breier G, Longaker MT, Greenhalgh DG and Werner S. (1995). J. Biol. Chem., 270, 12607 ± 12613. Garrido C, Saule S and Gospodarowicz D. (1993). Growth Factors, 8, 109 ± 117. Gleadle JM, Ebert BL, Firth JD and Rathcli€e PJ. (1995). Am. J. Physiol., 268, C1362 ± C1368. Goldberg MA and Schneider TJ. (1994). J. Biol. Chem., 269, 4355 ± 4359. Graeber TG, Osmanian C, Jacks T, Housman DE, Koch CJ, Lowe SW and Giaccia AJ. (1996). Nature, 379, 88 ± 91. Graeber TG, Peterson JF, Tsai M, Monica K, Fornace AJ and Giaccia AJ. (1994). Mol. Cell. Biol., 14, 6264-6277. Grugel S, Finkenzeller G, Weindel K, Barleon B and Marme D. (1995). J. Biol. Chem., 270, 25915 ± 25919. Hanahan D and Folkman J. (1996). Cell, 86, 353 ± 364. Joukov V, Pajusola K, Kaipainen A, Chilov D, Lahtinen I, Kukk E, Saksela O, Kalkkinen N and Alitalo K. (1996). EMBO J., 15, 290 ± 298. Kaipainen A, Korhonen J, Mustonen T, van Hinsbergh VWM, Fang G-H, Dumont D, Breitman M and Alitalo K. (1995). Proc. Natl. Acad. Sci. USA, 92, 3566 ± 3570. Kieser A, Weich HA, Brandner G, Marme D and Kolch W. (1994). Oncogene, 9, 963 ± 969.

Kukk E, Lymboussaki A, Taira S, Kaipainen A, Jeltsch M, Joukov V and Alitalo K. (1996). Development, 122, 3829 ± 3837. Lagercrantz J, Larsson C, Grimmond S, Fredriksson M, Weber G and Piehl F. (1996). Biochem. Biophy. Res. Commun., 220, 147 ± 152. Lee J, Gray A, Yuan J, Luoh S-M, Avraham H and Wood WI. (1996). Proc. Natl. Acad. Sci. USA, 93, 1988 ± 1992. Liu HS, Scrable H, Villaret DB, Lieberman MA and Stambrook PJ. (1992). Cancer Res., 52, 983 ± 989. Maglione D, Guerriero V, Viglietto G, Delli-Bovi P and Persico MG. (1991). Proc. Natl. Acad. Sci. USA, 88, 9267 ± 9271. Maglione D, Guerriero V, Viglietto G, Ferraro MG, Aprelikova O, Alitalo K, Del Vecchio S, Lei K-J, Chou JY and Persico MG. (1993). Oncogene, 8, 925 ± 931. Mazure NM, Chen EY, Yeh P, Laderoute KR and Giaccia AJ. (1996). Cancer Res., 56, 3436 ± 3440. Mukhopadhyay D, Tsiokas L and Sukhatme VP. (1995). Cancer Res., 55, 6161 ± 6165. Nathans D, Lau LF, Christy B, Hartzell S, Nakabeppu, Y and Ryder K. (1988). Cold Spring Harbor Symp. Quant. Biol., 53, 893 ± 900. Olofsson B, Pajusola K, Kaipainen A, von Euler G, Joukov V, Saksela O, Orpana A, Pettersson RF, Alitalo K and Eriksson U. (1996). Proc. Natl. Acad. Sci., 93, 2576 ± 2581. Paavonen K, Horelli-Kuitunen N, Chilov D, Kukk E, Pennanen S, Kallioniemi O-P, Pajusola K, Olofsson B, Eriksson U, Joukov V, Palotie A and Alitalo K. (1996). Circulation, 93, 1079 ± 1082. Pawson T. (1995). Nature, 373, 573 ± 580. Pertovaara L, Kaipainen A, Mustonen T, Orpana A, Ferrara N, Saksela O and Alitalo K. (1994). J. Biol. Chem., 269, 6271 ± 6274. Plate KH, Breier G, Weich HA and Risau W. (1992). Nature,, 359, 845. Risau W. (1995). FASEB J., 9, 926 ± 933. Shaulian E, Zauberman A, Ginsberg D and Oren M. (1992). Mol. Cell. Biol., 12, 5581 ± 5592. Shima DT, Deutsch U and D'Amore PA. (1995). FEBS Lett., 370, 203-208. Shweiki D, Itin A, So€er D and Keshet E. (1992). Nature (Lond.), 359, 843 ± 848. Stein I, Neeman M, Shweiki D, Itin A and Keshet E. (1995). Mol. Cell. Biol., 15, 5363 ± 5368. Suri C, Jones PF, Patan S, Bartunkova S, Maisonpierre PC, Davis S, Sato TN and Yancopoulos GD. (1996). Cell, 87, 1171 ± 1180. Thomas K. (1996). J. Biol. Chem., 271, 603 ± 606. Tischer E, Mitchell R, Hartman T, Silva M, Gospodarowicz D, Fiddes JC and Abraham JA. (1991). J. Biol. Chem., 266, 11947 ± 11954. Vikkula M, Boon LM, Carraway IKL, Calvert JT, Dlamonti AJ, Goumnerow B, Pasyk KA, Marchuk DA, Warman ML, Cantley LC, Mulliken JB and Olsen RR. (1996). Cell, 87, 1181 ± 1190. Vincenti V, Cassano C, Rocchi M and Persico MG. (1996). Circulation, 93, 1493 ± 1495.