Regulation of β-Galactoside α 2,6-Sialyltransferase Gene Expression by Dexamethasone

The hepatic acute phase response is accompanied by increased levels of Gal@l-4GlcNAc cr2,6-sialyltransferase activity in liver and in circulation. Previous studies suggested that cytokines and glucocorticoids mediate the induction of this sialyltransferase activity. In this study the regulation of sialyltransferase expression by dexamethasone in H35 rat hepatoma cells is assessed by Northern hybridization and enzyme activity assays. Exposure of H35 cells to 1 PM dexamethasone for 24 h causes a 3-4-fold enrichment of sialyltransferase mRNA and a corresponding increase in enzymatic activity. The induction of sialyltransferase mRNA begins within 3 h of dexamethasone treatment and reaches a plateau within 24 h. Sialyltransferase mRNA induction is dose dependent; the minimum concentration of dexamethasone necessary for induction is lo-’ M, and induction was maximal t lo-‘ M. Induction is sensitive to actinomycin D, suggesting that regulation may be exerted by altering the rate of mRNA synthesis. Puromycin and cycloheximide are ineffective in blocking induction, suggesting that de novo protein synthesis is not required for induction. Finally, dexamethasone alone is sufficient for maximum induction of sialyltransferase mRNA. In contrast, maximal induction of al-acid glycoprotein, a well studied hepatic acute phase reactant, requires both dexamethasone and cytokines, implying that different pathways exist for the induction of participants in the acute phase response.

Regulation of ,&Galactoside ac2,6-Sialyltransferase Gene Expression by Dexamethasone* (Received for publication, September 29, 1988) XueCheng Wang, Terrance P. O The hepatic acute phase response is accompanied by increased levels of Gal@l-4GlcNAc cr2,6-sialyltransferase activity in liver and in circulation. Previous studies suggested that cytokines and glucocorticoids mediate the induction of this sialyltransferase activity. In this study the regulation of sialyltransferase expression by dexamethasone in H35 rat hepatoma cells is assessed by Northern hybridization and enzyme activity assays. Exposure of H35 cells to 1 PM dexamethasone for 24 h causes a 3-4-fold enrichment of sialyltransferase mRNA and a corresponding increase in enzymatic activity. The induction of sialyltransferase mRNA begins within 3 h of dexamethasone treatment and reaches a plateau within 24 h. Sialyltransferase mRNA induction is dose dependent; the minimum concentration of dexamethasone necessary for induction is lo-' M, and induction was maximal at lo-' M. Induction is sensitive to actinomycin D, suggesting that regulation may be exerted by altering the rate of mRNA synthesis. Puromycin and cycloheximide are ineffective in blocking induction, suggesting that de novo protein synthesis is not required for induction. Finally, dexamethasone alone is sufficient for maximum induction of sialyltransferase mRNA. In contrast, maximal induction of al-acid glycoprotein, a well studied hepatic acute phase reactant, requires both dexamethasone and cytokines, implying that different pathways exist for the induction of participants in the acute phase response.
The P-galactoside a2,6-sialyltransferase catalyzes the transfer of sialic acid onto exposed GalPl-4GlcNAc termini of N-linked oligosaccharides common to serum and cell surface glycoproteins (1,2). Although this sialyltransferase has widespread tissue distribution, it is particularly abundant in liver, the major site of serum glycoprotein synthesis (3). However, very little is known about the regulatory mechanisms that dictate the expression of this enzyme, especially within the context of its role in appropriate glycosylation of hepatic and serum glycoproteins. The sialyltransferase exists predominantly in a membrane-bound form within the Golgi and trans-Golgi network (4) where it participates in the posttranslational modification of newly synthesized secretory or cell surface glycoproteins. A soluble form of the sialyltransferase exists in the serum (5) and is thought to be derived from the liver (6, 7) by a proteolytic event that liberates the catalytic domain from its membrane anchor (6,s).
* This work was supported by Grant GM38193 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed. Elevation in liver and serum P-galactoside a2,6-sialyltransferase activity is one of the hepatic responses to acute systemic injury (6,9,10) that include increased serum protein-bound carbohydrate (11,12) and the induction of a subset of serum glycoproteins, the acute phase reactants (13)(14)(15). Because many of the acute phase reactants, most notably al-AGP,' fibrinogen, and haptoglobin, are sialylated glycoproteins (15), it is not surprising that enhanced sialyltransferase activity is part of the hepatic acute phase response. Nevertheless, the precise molecular mechanism(s) that coordinate the overall hepatic acute phase reaction remain far from clear.
Activated monocytes (16), tissue macrophages (17), and growing keratinocytes (18) secrete factors that elicit acute phase protein production in vitro as well as in uiuo. In some cases, the "hepatocyte stimulating factors" (HSF) appear to be the systemic lymphokines, interleukin-1 (19) and interleukin-6 (20). For the expression of some acute phase proteins, maximum induction by HSF requires the synergistic cooperation of glucocorticoids (21). It has been proposed that glucocorticoids are required to maintain cells in a state receptive to HSF stimulation (22). Consistent with this view is the observation that dexamethasone induction of al-AGP on the mRNA level requires continued protein synthesis (23). This may suggest the participation of yet unidentified, short-lived protein intermediates in the acute phase response. On the other hand, the existence of sequences residing 5' of the al-AGP structural gene required for the response to glucocorticoid indicates that glucocorticoid may regulate acute phase protein gene transcription directly. There are additional mechanisms by which glucocorticoids could influence hepatic gene expression post-transcriptionally. For example, it was recently reported that dexamethasone facilitates the intracellular transport of secretory glycoproteins in hepatocytes (24).
Induction of sialyltransferase activity by glucocorticoids (7) as well as by HSF (25) has been reported. Owing to the lack of nucleic acid probes, these investigators have relied solely on enzyme assays to measure sialyltransferase gene expression. Consequently, the molecular pathways that coordinate sialyltransferase expression with the induction of other acute phase proteins remain unclear. In this report, we utilize a probe complementary to the coding region of the liver Pgalactoside a2,6-sialyltransferase mRNA to examine the regulation of this sialyltransferase in hepatoma and primary hepatocyte cell cultures. We report that dexamethasone induces sialyltransferase activity by elevating the cellular level of sialyltransferase mRNA. Furthermore, induction is not dependent on continuing protein synthesis but does require ongoing transcription. Together the data support the model The abbreviations used are: al-AGP, al-acid glycoprotein; DMEM, Dulbecco's modified Eagle's medium; HSF, hepatocyte stimulating factor; PMA, phorbol 12-myristate 13-acetate. whereby glucocorticoids mediate the direct induction of sialyltransferase expression at the transcriptional level.
[cx-~'P]~ATP (3000 Ci/mmol) and CMP [14C]NeuAc (18 mCi/mmol) were purchased from Amersham Corp. and Du Pont-New England Nuclear, respectively. The Dulbecco's modified Eagle's medium (DMEM), fetal bovine serum, glutamine, pyruvate, penicillin, and streptomycin were purchased from GIBCO Laboratories. All other biochemicals were of the highest quality commercially available, and the chemicals were of reagent grade or higher.
Conditioned Media-Colo-16 conditioned medium is unfractionated medium from Colo-16 (human squamous carcinoma) cells, which constitutively produce hepatocyte-stimulating factors (26). This conditioned medium was generously provided by Dr. H. Baumann, Roswell Park Memorial Institute, Buffalo, NY.
Cell Cultures-A rat liver hepatoma cell line (H35) (27), a gift from Dr. H. Baumann, was grown in monolayer culture in DMEM supplemented with 10% fetal bovine serum, 2 mM glutamine, 1 mM pyruvate, 100 units/ml of penicillin, and 100 pg/ml of streptomycin, and incubated at 37 'C in 5% COZ. Cells were plated at a density of approximately 6 X lo' cells/cm'. Stock solutions of dexamethasone and retinoic acid were diluted in ethanol to a concentration of

M.
PMA (0.33 pg/ml) and CAMP (0.1 M) were dissolved in dimethyl sulfoxide and deionized distilled water, respectively. Drugs were added to the culture 24 h later. Control cultures contained equivalent amounts of solvent.
Primary Culture of Hepatocytes-Primary hepatocytes were isolated by collagenase perfusion of male rat (Sprague-Dawley, 250-300 g weight) liver and suspended in DMEM (28); the cell viability was greater than 90% as determined by trypan blue exclusion (10% trypan blue in 1 X phosphate-buffered saline). Aliquots of the hepatocyte suspension were plated on collagen-coated dishes at a density of 1.2 X lo5 cells/cm'. Cells were cultured in DMEM plus 10% fetal bovine serum, 100 units/ml of penicillin and 100 pg/ml of streptomcyin for 24 h before use in experiments.
RNA Preparation and Analysis-Total RNA from H35 cells and primary hepatocytes was extracted by the guanidium isothiocynate method (29), separated on formaldehyde agarose gels (30), electrotransferred onto nylon filter (Zetabind, Cuno Inc., Meriden, CT), and hybridized with radiolabeled probe (8 X lo5 cpm/ml) (31). The blots were incubated overnight at 65 "C and washed at the same temperature. The blots were exposed to XAR-5 films (Eastman Kodak Co., Rochester, NY) for 1-2 days with an intensifying screen. The sialyltransferase mRNA was quantitated by scanning with a soft laser densitometer (LKB 222-010 UltroScan XL).
Probe Preparation-The probe for hybridization of sialyltransferase mRNA was a 780-base pair BstEII/BstEII fragment from a rat @galactoside a2,6-sialyltransferase cDNA sequence which represents the distal two-thirds of the coding region. The cDNA was isolated from a rat liver cDNA library (Clontech Laboratory, Palo Alto, CA) using synthetic oligonucleotides specified by partial peptide sequence analysis.' The sequence in the open reading frame of this cDNA is in complete agreement with data recently reported by Weinstein et al. (8). The a,-AGP probe was a PstIIPstI fragment (850 base pairs) of pILR-10, a,-AGP cDNA containing plasmid that was kindly provided by Dr. H. Baumann. The &actin probe was derived from p11/14, a chicken @-actin cDNA plasmid kindly provided by Dr. D. Cleveland (Johns Hopkins School of Medicine). The DNA fragments were purified by agarose gel electrophoresis and labeled with [a-3ZP]dATP by the random-primer method (32).
Sialyltransferase Enzyme Activity Assay-Sialyltransferase enzyme activities were assayed in cellular homogenates by a method modified from that described by Baxter  Assays were initiated by the addition of 10 p1 of cellular homogenate (approximately 100-200 pg of protein) and incubated at 37 'C for 0 or 4 min. Under these conditions, incorporation was linear with respect to incubation time and substrate concentration for at least 60 min. Assays were terminated by the addition of 1 ml of ice-cold 10% trichloroacetic acid. Reaction mixtures were applied to 2.4-cm Whatman filter discs (GF/B) and washed twice with ice-cold trichloroacetic acid. Filters were rinsed once with cold 95% ethanol and acetone.
Filters were dried, immersed in 10 ml of ACS I1 scintillant (Amersham Corp.), and processed in a liquid scintillation counter to quantitate [14C]NeuAc incorporation. Protein concentration in cellular homogenates was quantitated according to the method of Bradford using a commercially available protein assay reagent (Bio-Rad). Total counts incorporated were corrected for background and normalized to protein concentration in the assay. Data were expressed as activity relative to basal enzyme activity in control samples.

Sialyltransferase Expression in H35 Cells Is Induced by
Dexamethasone-Previous reports have demonstrated that pgalactoside ~u2,6-sialyltransferase activity is elevated in the liver and in primary hepatocyte culture upon stimulation by dexamethasone and specific cytokines (25). To examine the molecular basis of this induction, we used a probe complimentary to the distal two-thirds of the coding region of the rat liver a2,6-sialyltransferase mRNA to assess expression on the RNA level. To determine if induced sialyltransferase activity is accompanied by increased steady-state mRNA, total cellular RNA was isolated from H35 cells that express nearly all major acute phase reactants in response to HSF stimulation (26). Northern blot analysis revealed that the H35-derived sialyltransferase mRNA migrated as a single band consistent with the 4.7-kilobase mRNA observed in rat liver (8) (Fig. 1,  lune 1 ). Exposure of H35 cells to 1 PM dexamethasone for 24 h resulted in a 4-fold increase in the steady-state level of the sialyltransferase mRNA (Fig. 1, lane 2).
To assess the dosage dependent stimulation of sialyltransferase in H35 cells, total RNA isolated from cells that were exposed to increasing concentrations of dexamethasone was analyzed on Northern blots. The sialyltransferase mRNA signal was quantitated by densitometric scanning and displayed in Fig. 2. The minimum concentration of dexamethasone required for hormonal induction of sialyltransferase is lo-@ M, and optimal stimulation occurs at M. The induction of sialyltransferase mRNA appears to be rapid, with halfmaximal induction being achieved approximately 6 h after the addition of dexamethasone (Fig. 3A). By 24 h, the mRNA level was close to maximal. The time course of induction was also assessed by determination of sialyltransferase enzymatic activity (Fig. 3B). As expected, the increase in enzymatic activity lagged behind the rise in the mRNA level, although close to maximum sialyltransferase activity was achieved within 24 h. Overall, exposure of H35 cells to dexamethasone resulted in a reproducible 3-4-fold induction of sialyltransferase gene expression on both the mRNA and enzymatic levels.
The ability of retinoic acid, CAMP, and phorbol ester to stimulate sialyltransferase expression was also tested. These compounds failed to significantly alter the level of sialyltransferase mRNA in H35 cells (Fig. 4, lunes 3, 5, and 7, respectively). Furthermore, in conjunction with dexamethasone, retinoic acid, CAMP, and PMA were unable to promote sialyltransferase expression higher than that achieved by dexamethasone alone (Fig. 4, lanes 4, 6, and 8).
Effects of Inhibitors of Protein and RNA Synthesis on Sialyltransferase Induction-Heightened sialyltransferase mRNA levels in dexamethasone-stimulated H35 cells may result from increased mRNA synthesis or stabilization of existing message. We tested the ability of H35 cells to maintain the appropriate level of sialyltransferase mRNA in the  dexamethasone (lane 2 ) . RNA (5 pg) was electrophoresed in a 0.8% agarose gel in the presence of formaldehyde, electrotransferred to Zetabind filter, and hybridized to '*P-labeled 780-base pair sialyltransferase cDNA fragment (see "Experimental Procedures"). The blot was exposed to x-ray film with an intensifying screen for 24 h. absence of transcription. H35 cells were stimulated with dexamethasone in the presence or absence of 10 pg/ml actinomycin D for 12 h. While stimulation with dexamethasone for 12 h normally caused a 3-4-fold increase in sialyltransferase mRNA (Fig. 5, compare lanes 1 and 2), actinomycin D prevented dexamethasone-dependent induction of sialyltransferase activity (compare lanes 5 and 6 ) .
To determine if de rwuo protein synthesis is required for sialyltransferase induction, the effects of puromycin and cycloheximide inhibition were tested. H35 cells were cultured in the presence of either 60 pg/ml puromycin or 10 pg/ml cycloheximide; concentrations previously demonstrated as sufficient to block the dexamethasone-mediated induction of al-AGP (23). Inhibition of protein synthesis by puromycin did Total RNA was isolated and equal amounts were loaded onto gels. The blot was probed for sialyltransferase mRNA signal and quantitated as described under "Experimental Procedures." The blot was subsequently probed for 8-actin and likewise quantitated. The sialyltransferase signal was normalized against the @actin signal. The data represents the mean of two experiments. not effect the steady-state level of sialyltransferase mRNA (Fig. 5, lane 3 ) . More importantly, dexamethasone-mediated induction at the mRNA level was not blocked by puromycin (Fig. 5, lane 4 ) . Identical results were obtained when 10 pg/ ml cycloheximide was used in place of puromycin (data not shown).
Comparison of Induction of al-AGP and Sialyltransferase-While previously published data indicated that de rwuo protein synthesis is required for al-AGP induction (23), our data indicate that sialyltransferase induction by dexamethasone does not require ongoing protein synthesis (see above). To assess the differences that exist between the regulation of sialyltransferase and al-AGP, conditions known to stimulate al-AGP expression were examined for their stimulatory effects on sialyltransferase expression. As mentioned in the Introduction, maximal al-AGP induction requires glucocorticoid as well as hepatocyte stimulating factor (21). Consistent with these previous observations, a low level of ax-AGP expression was achieved when H35 cells were exposed to dexamethasone alone (Fig. 6, lane 2) or to a conditioned medium known to contain HSFs (26) (Fig. 6, lane 3 ) . Combined stimulation by both glucocorticoid and conditioned medium elicited maximal al-AGP induction (Fig. 6, lane 4 ) . In contrast, sialyltransferase expression is induced by dexamethasone (Fig. 6, lane 3) but remain unaffected by conditioned medium (Fig. 6, lane 4 ) . Furthermore, the combined exposure of H35 cells to dexamethasone and conditioned Kinetics of dexamethasone-induced sialyltransferase mRNA and enzyme activities in H35 cells. A, time course of dexamethasone induction of sialyltransferase mRNA in H35 cells. Cells were exposed to 1 p~ dexamethasone for the indicated times. Total cellular RNA was isolated and aliquots were analyzed as described in the legend to Fig. 2. B, time course of dexamethasone induction of sialyltransferase activity in H35 cells. Parallel dishes of cells were washed twice with ice-cold phosphate-buffered saline and the cells scraped from them in 1 ml of cold phosphate-buffered saline. The cells were resuspended in 100 pl of homogenization buffer (50 mM imidazole-HCI, pH 7.0, 1% Triton X-100 detergent, 2.5 mM MnC12) and sonicated (W-225; Ultrasonics, Inc.). 10 pl of cellular homogenate was assayed for sialyltransferase activity and normalized using variations in total cellular protein concentrations as described under "Experimental Procedures." Each point represents the mean of three experiments. medium resulted in a level of sialyltransferase expression that was not substantially different from treatment with dexamethasone alone (Fig. 6, lane 4 ) .
Since H35 cells appropriately express major acute phase proteins in response to glucocorticoid and HSF stimulation (26), we sought to demonstrate that primary hepatocytes exhibit the same pattern of regulated sialyltransferase expression observed in H35 cells. As shown in Fig. 7, the presence of 1 p~ dexamethasone caused identical elevations of sialyl-  6). After 12 h incubation, total cellular RNA was isolated, electrophoresed, blotted, and hybridized to the sialyltransferase cDNA sequence as already described.  lanes 1 and 2 ) or presence of one-third diluted colo-16 conditioned medium (lanes 3 and 4 ) . In addition, 1 p~ dexamethasone was included in cells represented by lanes 2 and 4. After 24 h, total cellular RNA was isolated and analyzed by Northern blot hybridization and probed with sialyltransferase and al-AGP cDNA sequence.
transferase mRNA levels in H35 cells (lanes 1 and 2) and in primary cultures of hepatocytes (lanes 3 and 4 ) . In addition, conditioned medium, whether alone or in conjunction with dexamethasone, has no further effect on the level of sialyltransferase mRNA in primary hepatocyte cultures (data not shown).

DISCUSSION
In rat hepatoma cells (H35) that express most major acute phase proteins in response to cytokine-mediated stimulation (26), @-galactoside a2,6-sialyltransferase activity is stimulated 3-4-fold in response to dexamethasone. This increase in enzymatic activity is preceded by a 3-4-fold elevation in the level of sialyltransferase mRNA. The dosage for minimal and maximal induction are consistent with that required for induction of other glucocorticoid responsive systems, such as murine mammary tumor virus (23) and tyrosine aminotransferase (34). Actinomycin D, an inhibitor of transcription, abolishes the induction of sialyltransferase mRNA by dexamethasone. Although steady-state mRNA levels are influenced by both the rate of message degradation as well as the rate of synthesis, our data suggest that transcriptional enhancement is the principal mechanism for dexamethasonedependent induction of sialyltransferase expression.
Gene products within the "glucocorticoid domain" in hepatoma and primary hepatocyte cultures include tyrosine aminotransferase (34), tryptophan oxygenase (35), el-AGP, and al,-globulin (36). Although in each of these cases glucocorticoids modulate mRNA levels, the mechanism for hormone responsiveness may be substantially different. For example, induction of al-AGP and al,-globulin requires ongoingprotein synthesis, suggesting the participation of intermediate regulatory factors (23). In contrast, the glucocorticoid-dependent stimulation of tyrosine aminotransferase is not affected by the presence of protein synthesis inhibitors such as cycloheximide and emitine (34). Induction of a2,6-sialyltransferase by dexamethasone is not inhibited by puromycin or cycloheximide. In addition, there is no appreciable lag time (less than 3 h) between introduction of the hormone and the first observable increase in steady-state mRNA level. Together these observations strongly suggest that dexamethasone directly participates in enhancing sialyltransferase expression by increasing mRNA synthesis.
One of the hepatic responses to acute systemic injury is an elevation of a2,6-sialyltransferase activity in the liver and circulation (6). Little is known about the signal(s) mediating this hepatic response and the mechanisms that exist for the coordinated modulation of this sialyltransferase with other acute phase gene products. For most major acute phase reactants (i.e. a1-AGP (23), az-macroglobulin (22), and alantichymotrypsin (37)), maximum induction requires the synergistic action of glucocorticoid and HSF (22). In contrast, dexamethasone alone is sufficient to achieve maximum 3-4fold elevation in the sialyltransferase mRNA level; the presence of conditioned medium in addition to dexamethasone results in further induction of al-AGP but not sialyltransferase. In addition, retinoic acid, CAMP, and phorbol ester, either alone or in concert with dexamethasone were ineffective in further stimulation of sialyltransferase gene expression.
Our observations contrast with those of Woloski et al. (25), who reported elevated sialyltransferase activity in response to a hepatocyte stimulating factor produced by human peripheral blood monocytes (25). In our hands, exposure of cells to conditioned medium derived from human squamous carcinoma cells (colo-16) was ineffective in stimulating sialyltransferase mRNA levels, yet colo-16 medium did promote the expression of al-AGP, an acute phase protein. A possible explanation for the discrepancy is that an HSF that stimulates 02,6-~ialyltransferase expression is produced by peripheral blood monocytes but not by squamous carcinoma cells. This is a plausible possibility, since the hepatic acute phase response is apparently mediated by multiple HSFs, each with a slightly different range of target specificities (22). The regulation of sialyltransferase gene expression by different factors has been postulated by Jamieson et al. (38) based on the response kinetics of sialyltransferase, aI-AGP, and albumin when animals were challenged by a monokine derived from peripheral monocytes. However, the 3-4-fold elevation in sialyltransferase mRNA achieved by dexamethasone alone is consistent with reported increases in sialyltransferase activity using both in vivo and cell culture models that mimic the acute phase response (6,7). Thus our data support a model whereby the increased sialyltransferase activity present in inflamed liver is the result of mRNA synthesis, and that this induction is dictated by the concentration of circulating glucocorticoids.