Hormonal Regulation of Cystathionine β-Synthase Expression in Liver

Homocysteine metabolism is altered in diabetic patients. Cystathionine β-synthase (CBS), a key enzyme involved in the transsulfuration pathway, which irreversibly converts homocysteine to cysteine, catalyzes the condensation of serine and homocysteine to cystathionine. Studies in streptozotocin-induced diabetic rats have shown that CBS enzyme activity is elevated in the liver but not in the kidney, and this effect is reversed by insulin treatment. To determine whether these effects resulted from alterations at the level of gene transcription, CBS mRNA was measured in diabetic and insulin-treated diabetic rats. CBS mRNA levels were found to be markedly higher in streptozotocin-induced diabetic rat livers; these were reduced by insulin administration. In H4IIE cells, a rat hepatoma cell culture model, glucocorticoids increased the cellular levels of CBS enzyme protein and CBS mRNA; insulin inhibited this stimulatory effect. Treatment with insulin also decreased CBS levels in HepG2 cells, a human hepatoma cell line. Nuclear run-on experiments in the rat cells confirmed that stimulation of CBS gene expression by glucocorticoids and the inhibition by insulin occurred at the transcriptional level. Transient transfections of HepG2 cells with a CBS-1b promoter luciferase reporter construct showed that the promoter activity was decreased by 70% after insulin treatment. These results show that insulin has a direct role in regulating homocysteine metabolism. Altered insulin levels in diseases such as diabetes may influence homocysteine metabolism by regulating the hepatic transsulfuration pathway.

(1). Homocysteine, a sulfur-containing nonprotein amino acid, is an intermediate in the metabolism of methionine. It is at a metabolic crossroads between its synthesis from methionine and its removal through the transsulfuration or remethylation pathways (2). Two pyridoxal 5Ј-phosphate-dependent enzymes comprise the transsulfuration pathway: CBS, which catalyzes the condensation of serine and homocysteine to cystathionine, and cystathionine ␥-lyase, which catalyzes the formation of cysteine, ␣-ketobutyrate, and ammonia (3,4). The regulation of CBS gene expression is important in a number of physiological situations. Feeding rats a high protein diet or a high methionine diet increases CBS activity (5). It is also known that flux through the transsulfuration pathway provides cysteine for glutathione synthesis, so that altered CBS levels may be of importance in oxidative stress (6). The well known sparing effect of cysteine on methionine requirements is mediated, at least in part, via alterations in CBS activity. Our own recent work shows that glucagon administration to rats increases hepatic CBS enzyme activity and mRNA levels (7). CBS enzyme deficiency, an autosomal recessively inherited disorder, is the leading cause of homocystinuria. Partial deficiency may lead to hyperhomocysteinemia, causing premature peripheral and cerebral occlusive arterial disease (8,9).
Elevation of plasma homocysteine levels is recognized as an independent risk factor for the development of cardiovascular disease. The plasma concentration of homocysteine is known to be perturbed in diabetes, being increased when renal insufficiency is evident (10 -13). However, patients with insulindependent diabetes mellitus with no clinical signs of renal insufficiency, have lower than normal levels of homocysteine (14). Our previous work has shown that the plasma homocysteine level was also decreased in the streptozotocin-induced rat diabetic model with a concomitant increase in hepatic CBS enzyme activity; insulin administration reversed these effects (15). Insulin-dependent diabetes mellitus is characterized not only by insufficient circulating levels of insulin but also by elevated levels of counterregulatory hormones such as glucagon and glucocorticoids (15)(16)(17). Moreover, cyclic AMPelevating agents and glucocorticoids have been shown to increase the level of CBS enzyme activity in rat hepatoma cells (18). It is therefore possible that the increase in the hepatic CBS activity in diabetic rat liver may be brought about by decreased insulin and/or by increased levels of counterregulatory hormones. The aim of this study was to examine the role of insulin and counterregulatory hormones on the expression of the CBS gene, using a diabetic rat liver model as well as rat and human hepatoma cell lines.
ance with the guidelines of the Canadian Council on Animal Care. Male Sprague-Dawley rats weighing 250 -350 g were fed laboratory chow ad libitum and had free access to water. Diabetes was induced by a single injection of streptozotocin (100 mg/kg body weight; freshly dissolved in 10 mM citrate buffer (pH 4.5)) through the tail vein, under light ether anesthesia; an equal volume of citrate buffer was administered to control rats. The streptozotocin-diabetic rats were treated subcutaneously with insulin (Novolin Ultralente human insulin; Lilly) for 5 days to allow the animals to recover from any nonspecific effects of streptozotocin. Thereafter, insulin was withdrawn, and the rats received saline injections for up to 5 days. At this point, insulin was readministered to untreated-diabetic rats for up to 5 days. Control rats received saline throughout the experiment. The insulin dose (ϳ40 units/day) was adjusted to maintain blood glucose close to normal values as measured with an Ames Glucometer II, using a drop of blood obtained by tail prick. On the day of the study, animals were anesthetized with an intraperitoneal injection of sodium pentobarbital (65 mg/kg). Blood samples were taken from the abdominal aorta into heparinized tubes and placed on ice for a few minutes until they were centrifuged at 2700 ϫ g for 15 min for plasma separation. The liver was rapidly removed, freeze-clamped in liquid nitrogen, and stored at Ϫ70°C until used.
Cell Culture-Rat hepatoma H4IIE cells and human hepatoma HepG2 cells were obtained from the American Tissue Culture Collection (Manassas, VA). The CBS-negative a23 Chinese hamster Don fibroblast cell line has been described previously (19). H4IIE cells were grown in 75-cm 2 culture flasks in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, penicillin (100 units/ml), and streptomycin (100 g/ml) (Canadian Life Technologies, Burlington, Canada). HepG2 cells and a23 Chinese hamster Don fibroblast cells were cultured in Iscove's modified Dulbecco's medium supplemented with fetal bovine serum (10%), penicillin (100 units/ml), and streptomycin (100 g/ml). Cells were grown to 70 -75% confluence under 5% CO 2 in a humidified incubator at 37°C. H4IIE cells were made quiescent by overnight culture in medium containing 0.5% fetal bovine serum. Before the addition of hormones, cells were provided with fresh medium containing 0.5% fetal bovine serum for 1 h. The medium was changed once more before stimulating with triamcinolone and/or insulin (Sigma). At the end of the incubation, media were aspirated, and cells were washed once with phosphate-buffered saline. Cell viability was monitored by trypan blue exclusion (20).
Analytical Procedures-CBS enzyme activity in rat liver and in H4IIE cells was assayed using the method described by Miller et al. (21) in the presence of 0.95 mM S-adenosylmethionine, a positive allosteric regulator of the enzyme. The production of cystathionine was measured by HPLC (22). Homocysteine concentrations were determined using HPLC by the method of Vester and Rassmussen (23). This method measures total plasma homocysteine (both free and protein-bound forms). Plasma glucose was determined enzymatically (24). Protein concentration in the liver samples was determined by the Biuret method and by the method of Lowry (25) in cultured cells, using bovine serum albumin as standard. CBS enzyme activity in HepG2 cells was measured according to the method described by Kraus (26).
Western Blots-H4IIE cells were lysed in SDS sample buffer (62.5 mM Tris-HCl, 2% SDS, 10% glycerol, 50 mM dithiothreitol, and 0.1% bromphenol blue), scraped off the plate, and sonicated for 10 -15 s to shear DNA and reduce viscosity. Protein concentration of the cell lysates was determined by the bicinchoninic acid (27) method (Pierce) using bovine serum albumin as a standard. 15-20 g of total protein was separated by SDS-PAGE (10% polyacrylamide) and transferred to 0.2-m pore size nitrocellulose membranes (Bio-Rad). The membranes were blocked with 5% nonfat dry milk in Tris-buffered saline, pH 7.6, containing 0.1% Tween 20 (blocking buffer), incubated with anti-CBS (rat) polyclonal antibody (1:5000 dilution in blocking buffer), washed, and incubated with anti-rabbit IgG conjugated to horseradish peroxidase (1:2000 dilution in blocking buffer). The membranes were then washed and detected by exposing the membranes to Kodak Biomax ML film, employing a chemiluminescent detection system (New England Biolabs, Beverly, MA). Protein bands were quantitated by using Chemi-Imager™ 4000 (Alpha Innotech Corp., San Leandro, CA).
HepG2 cell lysate samples were electrophoresed on a 9% SDS-polyacrylamide gel and electroblotted onto polyvinylidene difluoride membrane (Immobilon-P; Millipore Corp., Bedford, MA) using a semidry blotting technique. CBS-negative a23 Chinese Hamster Don fibroblast cell lysate was used as a negative control. The membranes were blocked with 5% nonfat dry milk in phosphate-buffered saline containing 0.2% Tween 20. CBS was detected using specific anti-CBS antibody (1:2000 dilution in 3% bovine serum albumin, 0.2% Tween 20 in phosphatebuffered saline, pH 7.0), washed, and incubated with secondary antirabbit IgG, conjugated to horseradish peroxidase (Jackson Immunoresearch Laboratories, West Grove, PA). The second antibody was diluted to 1:30,000 in 5% nonfat dry milk in phosphate-buffered saline, pH 7.0, and 0.2% Tween 20. The signal was visualized using the ECL Western blot system (Amersham Biosciences) and fluorography with X-Omat film.
Northern Blots-Total RNA was isolated by a rapid guanidinium isothiocyanate method as previously described (28). RNA was denatured in formamide (50%), formaldehyde (6.5%) sample buffer at 65°C for 15 min; separated on formaldehyde (1.8%), agarose (0.8%); and transferred to nylon (S & S Nytran) membranes (Schleicher & Schuell). RNA was fixed by UV cross-linking. The membranes were prehybridized and then hybridized with a cDNA probe specific for rat CBS (2.3-kb EcoRI fragment of rat CBS cDNA (29), which was random primerlabeled with [␣-32 P]dCTP (3000 Ci/mmol; PerkinElmer Life Sciences) using the MegaPrime labeling kit (Amersham Biosciences), according to the manufacturer's instructions. The membranes were stripped and hybridized with a random primer radiolabeled ␤-actin cDNA probe (CLONTECH Laboratories, Palo Alto, CA) to control for equal loading of RNA and washed at high stringency (final wash, 0.1% SSC, 0.1% SDS at 53°C). Autoradiography was carried out by exposure of the blot to Kodak XAR or Biomax film. Autoradiographs were scanned on Chemi-Imager™ 4000, and RNA levels were quantitated with Alpha Ease software (Alpha Innotech Corporation, San Leandro, CA).
Reverse Transcriptase-PCR-Total RNA was isolated from H4IIE cells after treatment with hormones as described. 2 g of RNA was reverse transcribed using a one-step reverse transcription kit (Qiagen, Mississauga, Canada) and amplified by 30 cycles. An upstream primer (5Ј-GCCTTCAGGACATCCAGTGT-3Ј) and a downstream primer (5Ј-TCTTTCCGGGTCTGCTCACGGGC-3Ј) corresponding to the coding sequence of type 3 rat CBS cDNA (28) were used to amplify a 1644-bp PCR fragment. A 768-bp fragment of the rat ␤-actin was co-amplified using amplimer set primers (CLONTECH Laboratories, Palo Alto, CA). PCR products were separated on 0.8% agarose. Ethidium bromidestained bands were visualized by UV illumination. The gel was immersed in 0.25 M HCl for 30 min, denatured in denaturing buffer (1.5 M NaCl, 0.5 M NaOH) for 30 min, rinsed with distilled water, soaked in neutralization buffer (0.5 M Tris-HCl, pH 7.0, 1.5 M NaCl) with gentle agitation, incubated in 10ϫ SSC, and transferred to S and S Nytran SuPer charge nylon membrane. After transfer, the DNA was immobilized by UV cross-linking in a Hoefer UVC 500 UV cross-linker (Amersham Biosciences). The blots were simultaneously probed with 32 Plabeled CBS cDNA and radiolabeled ␤-actin cDNA probe (CLONTECH Laboratories, Palo Alto, CA). Autoradiography was carried out using Kodak XAR or Biomax film.
Recombinant Plasmids, Transfection, and Luciferase Assays-The sequence and construction of the CBS-1b promoter luciferase reporter construct pCBS47 have been described previously (30). The control luciferase reporter constructs pGL-Control (positive control), pGL3basic (negative control containing no promoter), and pRL-CMV (internal transfection control) were all obtained from Promega. Cellular transfection was performed using Lipofectin (Invitrogen) according to the manufacturer's standard protocol for attached cells and typically used 0.5 g of the test construct and 50 ng of the internal pRL-CMV per well in a 24-well tissue culture plate. After transfection, cells were harvested in situ using direct cell lysis buffer (Promega) and were subsequently assayed for both Renilla and firefly luciferase activities using a dual luciferase system (Promega) and a Moonlight 2010 luminometer (Analytical Luminescence) according to the manufacturer's instructions. All promoter assay values were corrected for transfection efficiency by normalizing the firefly luciferase signal of the CBS-1b promoter construct to the Renilla luciferase values derived from the co-transfected, control vector pRL-CMV. All values are expressed as ratios of the test to the internal control reporter expression. Unless otherwise stated, the results shown are the average of at least three independent experiments performed in triplicate, with the error bars representing S.E.
Nuclear Run-on Analysis-Nuclear run-on assays were performed according to a previously described method (31). H4IIE cells were grown in 150-cm 2 flasks, and, for each assay, nuclei were isolated from 5 ϫ 10 7 cells. Nuclear suspensions (200 l) were incubated with 2.5 mM each of CTP, ATP, and GTP and with 200 Ci of [␣-32 P]UTP (3000 Ci/mmol; PerkinElmer Life Sciences). The 32 P-labeled RNA was isolated, after deoxyribonuclease I and proteinase K treatments, by phenol-chloroform extraction and ethanol precipitation; it was further purified by Micro Biospin 30 (Bio-Rad) chromatography. Equal amounts (10 7 cpm/ml) of labeled nuclear RNA were hybridized at 42°C for 72 h to Nytran membranes previously bound with 5 g of linearized plasmid DNAs. The immobilized plasmids used were CBS-pBluescript containing a 2.4-kb cDNA for rat liver CBS, ␤-actin-pBluescript (American Tissue Culture Collection), and pBluescript vector (for background control). After washing, the filters were subjected to autoradiography. Radioactive transcripts were quantified by densitometry. Data were normalized to transcription of the ␤-actin gene. Table I gives data on plasma homocysteine and glucose levels in diabetic rats. Plasma glucose levels were markedly higher in streptozotocin-injected rats, indicating profound diabetes. This returned to normal levels on insulin treatment. The animals were also ketotic. Mean blood pH in diabetic rats was 7.18 Ϯ 0.12, and the sum of ␤-hydroxybutyrate and acetoacetate was 5.5 Ϯ 0.8 mM. This ketosis was corrected by insulin treatment. Table I shows that plasma homocysteine level in the diabetic rats was only about 40% of that found in control rats. Treatment of diabetic rats with insulin increased plasma homocysteine levels within 1 day of treatment but did not return them to the control level even after 5 days of treatment. To examine whether the lowering of homocysteine levels in the untreated diabetic rat could be attributed to an increase in CBS enzyme activity and, by extension, an increase in homocysteine catabolism through the transsulfuration pathway, we measured the activity of this enzyme in livers from control, diabetic, and insulin-treated diabetic rats. CBS activity was elevated in the diabetic rats; this was returned to control values by insulin treatment (Fig. 1A). To determine whether this observed decrease in CBS enzyme activity is regulated at the level of transcription, we measured CBS mRNA in treated and untreated diabetic rat livers. CBS mRNA levels were higher in diabetic rat livers compared with control rat livers, and this was reversed by insulin treatment (Fig. 1B). When normalized for ␤-actin mRNA, the level of CBS mRNA was doubled in diabetic rat liver. Insulin treatment decreased the level of CBS mRNA to control values within a day (Fig. 1C).

RESULTS
Insulin-dependent diabetes mellitus is marked not only by a decrease in insulin levels but also by an increase in the circulating levels of the so-called counterregulatory hormones, glucocorticoids and glucagon (16,17). To evaluate whether insulin and its counter-regulatory hormones exert a direct and specific effect on liver cells, we used hepatoma cell lines. The basal CBS enzyme activity is low in H4IIE cells (rat hepatoma). However, CBS activity was elevated in these cells when stimulated with glucocorticoids. Specifically, the CBS activity in H4IIE cells, stimulated with the glucocorticoid analogue triamcinolone, exhibited an ϳ1.5-fold increase in CBS enzyme activity, which fell to normal levels when subjected to insulin treatment ( Fig.  2A). To examine whether this increase in enzyme activity in response to triamcinolone was due to an increased quantity of CBS protein, total cell lysate proteins were analyzed by Western blot. A significant increase in the level of CBS protein was evident in triamcinolone-stimulated H4IIE cells, and this was dependent on the concentration of triamcinolone. Fig. 2B shows the increase in CBS enzyme protein expression in response to increasing concentrations of triamcinolone. The 63-kDa CBS protein was evident in cells stimulated with 50 nM of triamcinolone, increased in a dose-dependent manner, and reached a maximum at a triamcinolone concentration of 1 M. H4IIE cells were therefore stimulated by a fixed concentration of triamcinolone (100 nM) in the presence of various doses of insulin (10 nM to 1 M). Fig. 2C shows that this glucocorticoid-mediated CBS expression was markedly inhibited by insulin at a concentration of 10 nM, and complete inhibition of CBS expression was seen at a concentration of 1 M.
To examine the effects of triamcinolone and insulin on CBS gene expression, total RNA was isolated from hormone-treated and control H4IIE cells. Total RNA (2 g) was reverse transcribed and amplified as described under "Experimental Procedures." Amplified products were analyzed by agarose gel electrophoresis. Ethidium bromide staining showed the expected 1642-bp CBS and 764-bp ␤-actin. Control cells expressed only low levels of CBS mRNA. To confirm that the products seen on ethidium bromide-stained gels were indeed CBS and ␤-actin, agarose gels were Southern blotted and  1. A, CBS enzyme activity in control, untreated diabetic, and insulin-treated diabetic rats. Data are expressed as means Ϯ S.D. Differences in letters between columns signify significant differences (p Ͻ 0.05). B, Northern blot analysis of total liver RNA in control, diabetic, and insulin-treated diabetic rats. Total RNA was isolated and separated on 0.8% agarose gels as described under "Experimental Procedures." Membranes were hybridized to 32 P-labeled CBS cDNA and ␤-actin cDNA. The blot is typical of three independent experiments. C, graphic representation of relative amounts of CBS mRNA measured by spot densitometry. CBS mRNA (2.4 kb) level was normalized to ␤-actin mRNA (2 kb) control. probed with 32 P random primer labeled CBS cDNA and ␤-actin cDNA (Fig. 3A). Southern blotting and densitometric analysis clearly showed that there was a 3-4-fold increase in CBS mRNA levels when H4IIE cells were stimulated with triamcinolone (1 M) for 18 h. This stimulation of mRNA expression by glucocorticoids was completely inhibited by insulin (1 M). Stimulation of these cells by CPT cAMP (200 M) resulted in a modest increase in CBS mRNA (by about 75%). However, no additive effect was seen when cells were stimulated by a combination of triamcinolone and cAMP (Fig. 3B). Stimulation by glucagon showed no effect on CBS mRNA expression. Our results show that triamcinolone and insulin exert opposing effects on CBS enzyme activity, protein concentration and mRNA levels in H4IIE cells. Glucocorticoids are known to function primarily by changing the rate of transcription of target genes (32). To determine whether glucocorticoids and insulin exerted direct control over the transcription of the CBS gene, nuclear run-on assays were performed on isolated nuclei from H4IIE cells that had been incubated with 1 M triamcinolone in the presence and absence of insulin for 6 h. Labeled transcripts show that the relative transcription rates were increased in cells that had been stimulated with triamcinolone. When normalized to the ␤-actin transcription rate, there was an approximately 2.5-fold increase in the CBS transcription rate after triamcinolone stimulation. Insulin treatment reversed this increase, suggesting that the effect of both hormones is mainly exerted at the level of transcription (Fig. 4, A  and B).
We next asked whether the effects of hormones on CBS enzyme activity and expression seen in rodent cells could be demonstrated in a human cell model. There was a strong basal expression of CBS in nonconfluent HepG2 cells. When these cells were subjected to insulin treatment, CBS enzyme activity was decreased by about 70%. Treatment with the counterregulatory hormone, glucagon, had no effect on CBS enzyme activity in these cells (Fig. 5A). These results were very similar to those obtained from the rat model. Western blotting analysis showed that this decrease in CBS activity is accompanied by decreased levels of CBS protein in insulin-treated cells compared with control. There was no change in the level of CBS protein in glucagon-treated cells (Fig. 5B). To examine whether insulin exerted its effect by controlling the promoter elements that mediate CBS gene transcription, HepG2 cells were transfected with the CBS-1b promoter reporter construct pCBS47 fused to a luciferase reporter gene (30). Treatment of the transfected cells with 0.1 M insulin for 48 h resulted in a marked decrease in luciferase activity (Fig. 5C), showing that the decreased CBS activity brought about by insulin is mediated through insulin-sensitive sequence(s) on the CBS gene regulatory region. Stimulation by glucagon treatment had no effect on CBS promoter activity.

DISCUSSION
It is now clear that homocysteine metabolism is impaired in diabetic patients (15,33). Those with nephropathy have increased plasma homocysteine levels (10,12,13). This may be due to the fact that the kidney is an important organ in the removal of homocysteine (33,34). In contrast, diabetic patients with normal creatinine levels (indicator of kidney function) have lower plasma homocysteine levels (13). The present study shows that plasma homocysteine levels are decreased in the streptozotocin-induced diabetic rat model. Streptozotocintreated rats were frankly diabetic as indicated by elevated plasma glucose levels, which were well controlled in the insulintreated diabetic group. We have previously shown that plasma creatinine levels were not elevated in diabetic rats (15). Hence, the changes in plasma homocysteine levels were not due to changes in renal function. Our data show that in diabetic rats there is an increased CBS enzyme activity accompanied by a concomitant increase in CBS mRNA expression, which can be reversed by insulin treatment. Thus, the decreased homocysteine levels in the Type 1 diabetic model are clearly associated with an increase in CBS activity, a key enzyme involved in the catabolism of homocysteine. This, in turn, is regulated by an increase in CBS gene expression, suggesting that the hepatic transsulfuration pathway is regulated by insulin possibly at the level of transcription. Increased flux through the hepatic transsulfuration pathway is also suggested by earlier work from this laboratory, which demonstrated that the hepatic content of cysteine is increased 3-fold in diabetic rats, whereas the level of methionine is significantly lowered (35). This enzyme is also regulated by S-adenosylmethionine, which is a positive allosteric regulator of CBS (36). However, we have measured the levels of this metabolite in freeze-clamped livers and have found no significant change in streptozotocin-diabetic rats with or without insulin treatment (control, 79.1 Ϯ 7.5; diabetic, 75.3 Ϯ 1.5; diabetic ϩ insulin, 80.8 Ϯ 6.9 nmol/g liver).
CBS enzyme activity is known to play an important role in determining plasma homocysteine levels. The human CBS gene has been localized to chromosome 21 (21q22.3) and is overexpressed in the trisomy 21 of Down's patients (37,38). These patients have lower than normal plasma homocysteine (39). CBS knockout mice have also been produced (40). Heterozygous CBS ϩ/Ϫ mice have a 50% reduction in CBS expression and twice the normal homocysteine levels (39). They also exhibit marked impairment in endothelial function (41,42).  2,5,7,9,11, empty. C, the human CBS-1b promoter is down-regulated by insulin but not affected by glucagon in HepG2 cells. HepG2 cells were transfected with the CBS-1b promoter reporter construct pCBS47. After transfection, these cells were induced to proliferate by the addition of 10% fetal calf serum and were treated with either 0.1 M insulin or glucagon for 48 h. Transfection and subsequent Luciferase assays were performed as described under "Experimental Procedures." Values shown represent the mean and S.D. of three independent experiments performed in triplicate.
Recent work from Loscalzo's group (43,44) has provided strong evidence that the adverse effects of homocysteine are at least partly mediated by oxidative inactivation of nitric oxide (NO). They showed that the impaired endothelium-dependent vasodilation seen in CBS-deficient mice can be attenuated either by the overexpression of glutathione peroxidase or by increased cellular thiol pools (43,44). Therefore, CBS expression plays an important role in determining plasma homocysteine levels and vascular function. This is confirmed in our study, where diabetes increases CBS activity, insulin treatment decreases it, and both of these changes are reflected in plasma homocysteine levels.
Our data clearly indicate that the effects of glucocorticoids and insulin on CBS expression occur at the transcriptional level. First, changes in enzyme activity are paralleled by changes in the levels of CBS protein and mRNA abundance. Second, in H4IIE cells, nuclear run-on assays clearly showed that stimulation of CBS expression by glucocorticoids and inhibition by insulin occurred at the level of transcription. In addition, actinomycin D inhibited the glucocorticoid-mediated induction of CBS mRNA in H4IIE cells (results not shown). Finally, transfection of a CBS-1b promoter-luciferase construct into HepG2 cells clearly showed that insulin repressed the promoter activity. This reflected the changes in CBS enzyme activity and the relative abundance of CBS protein, indicating a direct regulation of the promoter activity by insulin.
The basal expression of CBS is fairly strong in rat liver cells (15), but in H4IIE cells it was quite weak. In these cells, CBS expression was markedly enhanced by glucocorticoid stimulation, whereas cAMP elicited a moderate stimulation. An earlier report has shown that the level of CBS enzyme activity in H15 and Ad-1 cells (rat hepatoma variants) was elevated when stimulated with a combination of dexamethasone and cyclic AMP (18). However, the molecular control of CBS gene expression was not investigated. Our results indicate that CBS enzyme protein and mRNA can be induced by triamcinolone alone in H4IIE cells. Stimulation of H4IIE and HepG2 cells by glucagon alone had no effect on CBS gene expression, although in HII4E cells there was a small but significant stimulation of CBS by a cAMP analogue. To our knowledge, this is the first report that shows that insulin exerts a direct effect on the CBS transcription rate.
Glucocorticoids play an important role in the control of hepatic genes encoding regulatory enzymes of intermediary metabolism, including those enzymes involved in amino acid metabolism. Well documented examples are phosphoenolpyruvate carboxykinase (PEPCK), tyrosine aminotransferase, and enzymes of the urea cycle (45,46). Transcription of these genes is stimulated by glucocorticoids and cAMP, and these actions are opposed by insulin (47)(48)(49). A comparison of PEPCK expression with that of CBS is of interest. In rats, hepatic PEPCK is induced in diabetes and repressed by insulin treatment. In rat liver cells, both PEPCK and CBS genes are enhanced by cAMP, although the effect on PEPCK is much stronger. The well characterized negative insulin response sequence (IRS), the so called PEPCK motif (T(G/A)TTT(T/G)(G/T)), present on the promoters of PEPCK, glucose-6-phosphatase, and tyrosine aminotransferase among others, is also present on the human CBS gene promoter at nucleotides Ϫ4430 to Ϫ4436 (TGTTTGG) with respect to the ATG start site of the CBS coding region (30,50). Similar sequences are also found on the 5Ј-untranslated region of rat CBS (51). However, this element is absent from pCBS47 whose expression is suppressed by insulin (Fig. 5). Maclean et al. have provided evidence that the regulation of CBS is sensitive to intracellular redox state, whereas recent studies by Ge et al.
have shown that the gene is regulated by the redox-sensitive transcription factors Sp1/Sp3 and NF-Y (52,53). Sp1 has been reported to play a dominant role in the regulation of CBS (54), and a number of indirect mechanisms have been described whereby insulin and glucocorticoids can indirectly modulate the expression of Sp1-regulated genes (55,56). Further work is required to investigate the exact mechanism by which insulin represses the expression of CBS.