Pharmacological Inhibition of Protein Kinases in Intact Cells: Antagonism of Beta Adrenergic Receptor Ligand Binding by H-89 Reveals Limitations of Usefulness1

Experimental Procedures

Materials.

H-89 (N-[2-(p-bromocinnamylamino)ethyl]-5-isoquinolinesulfonamide·2HCl) was obtained from Alexis Corp. (San Diego, CA), BIOMOL (Plymouth Meeting, PA), and Calbiochem (La Jolla, CA). H-85 (N-[2-(N-formyl-p-chlorocinnamylamino)ethyl]-5-isoquinolinesulfonamide) was obtained from Seikagaka Corp. (Rockville, MD). H-7 [(±)-1-(5-isoquinolinesulfonyl)-2-methylpiperazine·2HCl], staurosporine (ST), bisindolylmaleimide IX [Ro 31-8220; 2-[1-[3-(amidinothio)propyl]-1H-indol-3-yl]-3-(1-methylindol-3-yl)-maleimide·CH₄O₃S] (Bis IX), and KT5720 were obtained from Alexis Corp. [¹²⁵I]Iodopindolol (IPIN) (2200 Ci/mmol), [¹²⁵I]adenosine-3′,5′-cyclic phosphoric acid (2200 Ci/mmol), [³H]dihydroalprenolol hydrochloride (82 Ci/mmol), [³²P]orthophosphoric acid (10,000 Ci/mmol), [γ-³²P]ATP, tetra(triethylammonium) salt (6000 Ci/mmol), [³H]scopolamine methylchloride (NMS) (84 Ci/mmol), [5,6-³H]SQ 29548 (51 Ci/mmol), andmyo-[1,2- ³H(N)]inositol (47 Ci/mmol) were obtained from DuPont-NEN (Boston, MA). [0-methyl-³H]Yohimbine (95 Ci/mmol) was obtained from Amersham (Arlington Heights, IL). FuGENE was obtained from Boehringer-Mannheim (Indianapolis, IN). The rat mast cell line RBL-2H3 stably expressing the human m1 muscarinic acetylcholine receptor (m1 mAchR; 40 fmol/mg whole cell protein) was provided by F. Santini. BEAS-2B human airway epithelial cells were provided by C. Harris. Phosphospecific antibody against p42/p44 MAP kinase was obtained from New England Biolabs (Beverly, MA). U46619 was obtained from Cayman Chemical Co. (Ann Arbor, MI). All other chemical reagents were obtained from Sigma Chemical (St. Louis, MO).

Cell Culture.

HASM cultures were established as described previously (Panettieri et al., 1989) from human tracheae obtained from lung transplant donors, in accordance with procedures approved by the University of Pennsylvania Committee on Studies Involving Human Beings. BEAS-2B (Penn et al., 1994), COS-1 (Goodman et al., 1996), and human embryonic kidney (HEK ) 293 (Krupnick et al., 1997) cultures were maintained as previously described. RBL-2H3 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 14% fetal bovine serum. Serum starvation of HASM cells consisted of 24-h incubation of cultures in DMEM supplemented with 5 μg/ml transferrin and 5 μg/ml insulin.

Spodoptera frugiperda (Sf9) membranes expressing either the human beta-2 AR or beta-1 AR were prepared as described previously (Penn et al., 1996).

Transfection Procedures.

pcDNA_3α2ARC10, encoding the human alpha-2C10 AR, was provided by A. Gagnon. pcDNA₃ TXA_2α receptor (TXA_2αR) was generated by reverse transcription-polymerase chain reaction (RT-PCR) of U937 (a human monocyte line) total mRNA using primers corresponding to the 5′ and 3′ flanking sequences of the published human TXA_2αR (Raychowdhury et al., 1994) and cloning of the ∼1300-bp product into the EcoRI/XhoI sites in pcDNA₃. COS-1 and HEK 293 cells were transfected using FuGENE and 0.1 to 0.5 μg/ml pcDNA_3α2ARC10 (COS-1) or pcDNA₃TXA_2αR (HEK 293) as per the manufacturer’s instructions. In studies designated for phosphoinositide analysis, HEK 293 cells were transfected in 6-well plates with FuGENE and 0.5 μg/ml pcDNA₃TXA_2αR. For receptor binding assays, cells were harvested 2 days after transfection, and membranes were prepared as described previously (Penn et al., 1994).

Effects of Kinase Inhibitors on Receptor-Mediated cAMP Accumulation.

Experiments examining the effects of various protein kinase inhibitors on beta-2 AR, prostaglandin E₂ (PGE₂) receptor, and adenylyl cyclase responsiveness were performed on HASM and BEAS-2B cells in a manner similar to that described previously (Penn et al., 1998). Briefly, serum-starved, confluent cells grown in 24-well dishes were pretreated with or without various concentrations of either vehicle [0.1% dimethyl sulfoxide (DMSO)], H-89, H-85, H-7, staurosporine, KT5720, or Bis IX for up to 1 h, followed by a 30-min pretreatment with or without 1 μM isoproterenol (ISO). Cells were then washed four times in cold phosphate-buffered saline (PBS), and individual wells were stimulated with 500 μl of PBS containing 300 μM ascorbic acid, 1 mM isobutylmethylxanthine (IBMX), and either vehicle (basal), (−)-ISO, PGE₂, or forskolin (FSK) for 10 min (except time course studies). cAMP was isolated and quantified by radioimmunoassay as described previously (Penn et al., 1998).

Assessment of In Vivo PKA-Mediated Phosphorylation.

HASM cells grown to confluence in 6-well plates were serum starved for 24 h and then loaded for 1 h with 0.3 mCi of [³²P]orthophosphate (in 1.2 ml of phosphate-free DMEM) per well. Cells were subsequently treated with various kinase inhibitors for 1 h and then stimulated with either vehicle (0.01% ethanol) or 10 μM FSK for 30 min. Cells were then washed three times with cold Tris-buffered saline and harvested in buffer containing 20 mM Tris·HCl, pH 8.0, 5 mM EDTA, 5 mM EGTA, 0.5 mM phenylmethylsulfonyl fluoride, 10 μg/ml pepstatin, 20 μg/ml leupeptin, 20 μg/ml benzamidine, 10 mM sodium pyrophosphate, 10 mM sodium fluoride, and 1% Triton X-100. Then, 20 μg of protein of the Triton-soluble extract was subjected to electrophoresis on a 10% SDS-polyacrylamide gel, and the gel was stained with Coomassie blue, dried, and visualized by exposure to Fuji RX film.

Analysis of MAPK Activity.

HASM cells were grown in 6-well plates to 80 to 90% confluence and then serum-starved for 48 h. Cells were pretreated with the various protein kinase inhibitors, followed by 30-min stimulation with either 10 ng/ml EGF or 100 nM PMA. Cells were then washed with cold PBS and lysed by the direct addition of 125 μl of SDS sample buffer. Harvested lysates were subjected to Western analysis using a rabbit polyclonal IgG antibody (NEB) that specifically recognizes the phosphorylated forms (Thr202/Thr204) of p42 and p44 MAP kinase, goat anti-rabbit horseradish peroxidase-conjugated secondary antibody, and visualization by enhanced chemiluminescence (ECL; Amersham). Equal loading of samples was assessed by staining of blot with Ponceau S or by subsequent probing of stripped blots with a polyclonal antibody (NEB) that recognizes total p42/p44 MAPK (phosphorylation-state independent).

Effects of Kinase Inhibitors on GRK Activity.

Bovine GRK2 and human GRK5 were overexpressed and purified from Sf9 cells (Kim et al., 1993; Kunapuli et al., 1994). GRK-mediated phosphorylation was assayed by incubating 0.8 pmol of GRK with urea-treated rod outer segments membranes (120 pmol of rhodopsin) in 20 μl of 60 mM Tris·HCl, pH 8.0, 4 mM MgCl₂, 0.5 mM EDTA, 0.1 mM [γ-³²P]ATP, and varying concentrations of protein kinase inhibitors for 8 min at 30°C in room light. Reactions were stopped with SDS buffer and electrophoresed on a 10% SDS-polyacrylamide gel. Gels were stained with Coomassie Blue, dried, and visualized. ³²P-labeled bands were excised and counted.

Receptor Binding Studies.

Competition of radioligand binding by various protein kinase inhibitors was performed using membranes prepared from cells expressing either the human beta-2 AR (Sf9 and BEAS-2B), human beta-1 AR (Sf9), human m1 mAchR (RBL-2H3), human m2 mAchR (HASM), humanalpha2ARC10 (COS-1), or human TXA_2αR (HEK 293). Typically, membranes were incubated with a 1- to 3-fold concentration of the established or empirically determinedK_d for [¹²⁵I]IPIN (beta-1 AR and beta-2 AR), [³H]dihydroalprenolol (beta-1 AR andbeta-2 AR), [³H]NMS (m1 and m2 mAchR), [0-methyl-³H]yohimbine (alpha2ARC10), or [5,6-³H]SQ 29548 (TXA_2αR ), in the presence of various concentrations of protein kinase inhibitors. For saturation binding analysis, BEAS-2B cells were pretreated with or without 10 μM H-89 for 30 min at 22°C in PBS containing 5 mM EDTA, 1 mM ethylene glycol bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid, 0.5 mM phenylmethylsulfonyl fluoride, 10 μg/ml pepstatin, 20 μg/ml leupeptin, and 20 μg/ml benzamidine and then washed four times in cold PBS. Cells were resuspended and incubated in PBS containing ∼8 to 200 pM [¹²⁵I]IPIN for 1 h at 22°C. All competition binding reactions were performed in 25 mM Tris·HCl, pH 7.5, and 2 mM MgCl₂, except [5,6-³H]SQ 29548 binding, which was performed as described previously (Raychowdhury et al., 1994) using 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid-buffered DMEM. The addition of compounds did not alter buffer pH, except H-7 at 500 μM, which was adjusted to pH 7.5 with NaOH. Nonspecific binding was determined using 1 μM alprenolol (beta-1 AR andbeta-2 AR), 1 μM atropine (m1 and m2 mAchR), 10 μM phentolamine (alpha2ARC10), or 1 μM SQ 29548 (TXA_2αR ). All binding reactions (except [5,6-³H]SQ 29548) were terminated by the addition of 5× 4 ml of ice-cold 25 mM Tris·HCl, pH 7.5, and 2 mM MgCl₂and filtration through Whatman GF/C filters using a Brandel Cell Harvester. IC₅₀ values were estimated by interpolation, andK_i values were calculated using the method of Cheng and Prusoff (1973).

Accumulation of Total [³H]Inositol Phosphates.

[³H]Inositol phosphate formation was determined as reported previously with minor modifications (Widdop et al., 1993). Near-confluent cell monolayers in 12-well plates were incubated for 24 h at 37°C with 500 μl of inositol-free DMEM containingmyo-[³H]inositol (47 Ci/mmol) at a concentration of 4 μCi/ml. After loading, cells were washed once with PBS. Inositol-free DMEM containing 10 mM LiCl, with or without kinase inhibitors, was added to each well, and the cells were incubated for 10 min at 37°C. Cells were then stimulated for 10 min with 20 μM (final concentration) histamine (HASM and BEAS-2B cells), 50 μM carbachol (RBL-2H3), or 20 nM U46619 (HEK 293). Reactions were stopped by aspiration of medium and the addition of 0.8 ml of ice-cold 0.4 M perchloric acid. One-half volume of 0.72 N KOH/0.6 M KHCO₃was added, and the sample was centrifuged to settle the precipitate. The supernatant was applied to 1 ml AG1-X8 (BioRad) columns (100–200 mesh, formate form), columns were washed with 10 ml of 0.1 N formic acid, and total inositol phosphates were eluted with 1.5 M ammonium formate/0.1 N formic acid and counted.

Data Analysis.

Except where noted, values are reported as mean ± S.D. cAMP accumulation was calculated by subtracting the matched basal value from that determined with agent stimulation, as described previously (Penn et al., 1994).

Results

Our initial inquiry into kinase inhibition in intact cells arose from studies examining the role of PKA in the desensitization of thebeta-2 AR in HASM. Several different kinase inhibitors were tested for their ability to attenuate the agonist-specific desensitization of the beta-2 AR invoked by pretreating cells for 30 min with 1 μM ISO. Typically, the loss ofbeta-2 AR responsiveness after beta-agonist exposure, characterized by a loss of ISO-stimulated cAMP with subsequent rechallenge, is mediated by beta-2 AR phosphorylation by GRKs and, in some cell types, by PKA as well (Penn and Benovic, 1998). Figure 1 depicts the results of experiments in which cells were pretreated for 1 h with 10 μM H-89, 500 μM H-7, 10 μM KT5720, or 1 μM staurosporine before 30-min pretreatment with 1 μM ISO. H-85 (10 μM), an analog of H-89 that does not inhibit PKA, and Bis IX (1 μM), a PKC-specific inhibitor, were used as negative controls and to assess possible contributions of PKC. Among the suspected PKA inhibitors, concentrations used were ∼200 to 500 times theK_i values reported to inhibit PKA activity in in vitro assays. Interestingly, H-89 and, to a lesser extent, staurosporine were able to attenuate ISO-inducedbeta-2 AR desensitization in HASM. H-85 also had a small, significant effect, whereas KT5720 and H-7 exhibited no effect. Bis IX also had no effect, suggesting that PKC does not contribute to agonist-specific beta-2 AR desensitization in HASM and that the effect of staurosporine is likely mediated via inhibition of PKA, which we have reported previously (Penn et al., 1998).

The disparate effects of the various kinase inhibitors prompted us to more directly examine the effectiveness and specificity of these agents. The effects of inhibitors on PKA-mediated whole-cell phosphorylation were examined in HASM cells loaded with [³²P]orthophosphate and then challenged with 10 μM FSK for 30 min. Analysis of cell extracts by SDS-polyacrylamide gel electrophoresis (PAGE) (Fig. 2A) demonstrates that stimulation with FSK induces the appearance of bands at ∼40 and 20 kDa, whereas pretreatment with H-89 or staurosporine, but not KT5720, H-7, Bis IX, or H-85 (data not shown), eliminates this induction. Thus, only H-89 and staurosporine appear to be effective PKA inhibitors in HASM cells, a result consistent with the results of Fig.1 and our observation that H-89 and staurosporine are able to attenuate the cAMP-mediated desensitization of the beta-2 AR in HASM elicited by pretreatment with either FSK or PGE₂(a ∼25% loss of beta-2 AR responsiveness) (Penn et al., 1998, and data not shown). To examine compound selectivity, the capacity of these agents to inhibit p44/p42 MAPK phosphorylation by EGF (through tyrosine kinase activation) or the phorbol ester PMA (through PKC) was also tested (Fig. 2B). None of the agents inhibited EGF-stimulated p44/p42 phosphorylation. Staurosporine, H-7, and Bis IX effectively inhibited PMA-stimulated MAPK activation, whereas H-89 and KT5720 had no effect. Interestingly, despite a similar potency for inhibiting PKA and PKC in vitro (3–6 μM), H-7 appears to preferentially inhibit PKC (Fig. 2B) and not PKA (Figs. 1 and 2A) in HASM cells.

None of the agents demonstrated significant potency (i.e., were effective at concentrations that might be used to inhibit kinase activity in intact cells) in inhibiting either GRK2 or GRK5 when GRK activity was assessed by phosphorylation of rhodopsin in vitro (Table1). A slightly greater tendency of compounds to inhibit GRK5- versus GRK2-mediated phosphorylation was observed, with staurosporine being the most potent. Although methodology for direct assessment of GRK activity in intact cells has yet to be developed, the low potencies demonstrated in an in vitro assay conducted using a low concentration of ATP (0.1 mM) suggest that the examined compounds are unlikely inhibitors of GRKs in intact cells. Collectively, these studies suggest that of the compounds examined, only H-89 and staurosporine are effective inhibitors of PKA in HASM cells, that H-89 appears to be specific for PKA inhibition, and that the effects of H-89 and staurosporine on beta-2 AR responsiveness in HASM are likely via inhibition of PKA.

However, several observations caused us to question the relatively large effect of H-89 in reversing ISO-mediated beta-2 AR desensitization in HASM. One concern was that although 10 μM H-89 appeared equally effective as 1 μM staurosporine in inhibiting PKA, H-89 was much more effective than staurosporine in reversing agonist-specific desensitization (Fig. 1). Second, this reversal (from ∼40 to 80% of control values) far exceeded the magnitude of cAMP-mediated desensitization invoked by FSK or PGE₂ treatment [a ∼25% loss (Penn et al., 1998)], even though FSK/PGE₂ stimulate higher levels of cAMP accumulation than does ISO. Third, the small effect of H-85 in attenuating beta-2 AR desensitization suggests the contribution of a nonspecific component mediating the effect of H-89. Last, H-89 had a similar effect in reversing agonist-specificbeta-2 AR desensitization in BEAS-2B human airway epithelial cells (see below), a cell line that exhibits little, if any, cAMP-mediated desensitization (Penn et al., 1994).

A more thorough analysis of the effects of H-89 on beta-2 AR responsiveness was therefore undertaken in both HASM and BEAS-2B cells. First, we examined the profile of cAMP production in HASM that occurs during the pretreatment phase with 1 μM ISO in the continued presence of H-89 or staurosporine after incubation with these agents for 1 h. Figure 3 reveals that although 1 μM staurosporine enhances 1 μM ISO-stimulated cAMP accumulation, 10 μM H-89 causes a slight inhibition, a result seemingly paradoxical with data from Fig. 1. This effect could not be attributed to alterations in phosphodiesterase activity because a similar inhibition was observed in the presence of 1 mM IBMX (data not shown). Next, we examined the dose-dependent response to ISO after H-89 and ISO pretreatment and extensive washing of cells. Oddly, H-89 pretreatment resulted in a reduction in cAMP accumulation at low ISO concentrations, whether cells were pretreated with or without ISO (Fig.4A), and significantly increased the EC₅₀ for ISO. Similar results were obtained using BEAS-2B cells (Fig. 4B).

These results were perplexing in light of early studies by Clark and coworkers (1987, 1988), demonstrating that PKA preferentially inhibitsbeta-2 AR responsiveness at low ISO concentrations; thus inhibition of PKA should enhance cAMP accumulation at low ISO concentrations and effectively decrease the EC₅₀for ISO. Indeed, this is the case in HASM cells treated with staurosporine (Penn et al., 1998). The observed effect of H-89 in HASM and BEAS-2B cells (Fig. 4, A and B) was more consistent with classic pharmacological receptor antagonism against a backdrop of intracellular PKA inhibition (the latter suggested by the enhancement of maximal response in ISO-pretreated cells, consistent with Fig. 1 results). When HASM cells were treated with H-89 for 1 h and then directly stimulated with ISO (no washing before ISO addition), a dose-dependent effect of H-89 antagonism of beta-2 AR activation is observed that appears competitive with ISO (Fig. 4C). Moreover, antagonism of ISO-stimulated cAMP accumulation is observed, albeit to a lesser extent, when H-89 is added to culture media immediately before stimulation with ISO (Fig. 4D). Basal levels of cAMP were not significantly altered by the acute addition of H-89 in either HASM or BEAS-2B cells (data not shown). These results suggest that H-89 antagonizes ISO binding to beta-2 AR, and the effects of H-89 in attenuating beta-2 AR desensitization can be explained in part as a result of the compound being retained despite extensive washing.

Accordingly, receptor binding analyses demonstrate that H-89 acts as abeta-2 AR antagonist of surprisingly high affinity (180 nM) (Fig. 5A and Table2). Sf9 membranes overexpressing the human beta-2 AR (∼20 pmol/mg protein) were incubated with 40 pM of [¹²⁵I]IPIN (∼2K_d) (Penn et al., 1996) and increasing concentrations of various kinase inhibitors. Competition binding using membranes prepared from BEAS-2B cells (∼100 fmolbeta AR/mg protein, >90% beta-2 AR) yielded nearly identical results (Table 2). Similar results were obtained in binding studies using intact BEAS-2B cells (data not shown). The addition of H-89 did not alter buffer pH, and results were similar whether binding was performed using 25 mM Tris, pH 7.5, 2 mM MgCl₂, PBS, pH 7.5, or 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid-buffered DMEM, pH 7.4 (data not shown). Interestingly, H-85, which is identical in structure to H-89 except for an N-formylamine substitution and a parachlorocinnamyl (replacing the parabromocinnamyl) group, inhibited beta-2 AR binding with approximately one-10th the potency (K_i = 1.5 μM). Much higher concentrations of H-7 (K_i > 500 μM) were required to significantly inhibit [¹²⁵I]IPIN binding to beta-2 AR, suggesting that the hydrophobic parahalocinnamyl moiety present in H-89 and H-85 is a key determinant of affinity.

Analysis of the effect of 10 μM H-89 on ISO-stimulated cAMP generation from Fig. 4C using the Schild equation predicts a significantly higher (∼7 fold) K_ivalue than that determined by radioligand binding. However, the value of the slope of the Schild regression (0.57, r = 0.99) suggests that the effect of H-89 on the beta-2 AR-adenylyl cyclase signaling cascade is not via strict competitive antagonism of the beta-2 AR and likely involves an effect occurring distinct from the beta-2 AR binding site. To investigate the extent to which H-89 binding to beta-2 AR is reversible and competitive, saturation binding isotherms were performed using BEAS-2B cells (Fig. 5B). These experiments demonstrate that H-89 binding is largely surmountable, although binding at submaximal concentrations of [¹²⁵I]IPIN is still significantly inhibited in the pretreated and washed cells (a finding consistent with data in Fig. 4, A and B), suggesting retention of H-89 despite extensive washing. The collective findings from Figs. 2 through 5 suggest a multifactorial effect of H-89 on beta-2 AR signaling that involves competitive antagonism of agonist binding (inhibitingbeta-2 AR activation) combined with inhibition of intracellular PKA (serving to enhance beta-2 AR responsiveness). Additional mechanisms underlying H-89 effects might also include PKA-independent alterations on the receptor environment, perhaps influenced by the lipophilic nature of the compound.

Additional studies examining the selectivity of kinase inhibitors were performed. H-89 (K_i = 350 nM) and H-85 (K_i = 1.7 μM) also exhibited the ability to antagonize beta-1 AR binding (Fig. 5C), albeit with a slightly lower affinity than that observed for thebeta-2 AR. Similar inhibition of [³H]dihydroalprenolol binding by H-89 to bothbeta-2 AR and beta-1 AR was observed (K_i values of ∼240 nM and 500 nM, respectively). H-89 and H-85 appeared to be fairly selective forbeta ARs, exhibiting inhibition of other receptors at much higher concentrations (Table 2). None of the other kinase inhibitors tested significantly inhibited beta AR binding at concentrations that might be reasonably used to inhibit kinase activity in intact cells (i.e., up to 1000-fold theK_i established in in vitro assays), although it should be noted that some inhibition by H-7 and Bis IX ofbeta ARs and other receptors occurred at high concentrations.

To further investigate the specificity of H-89 for antagonizingbeta AR, we examined the effects of H-89 on the functional responses of different GPRs linked to either G_sand adenylyl cyclase (beta-2 AR and PGE₂ receptor) or G_q and phospholipase C (m1 mAchR, H1 histamine receptor, and TXA_2αR). Various cells were treated with 10 μM H-89 before stimulation with submaximal concentrations of agonists approximating 1 to 3 times the EC₅₀ of activation. Figure 6A demonstrates the lack of effect of H-89 on PGE₂-stimulated adenylyl cyclase in BEAS-2B cells. In a similar manner, total inositol phosphate generation by either H1 histamine receptor activation in HASM and BEAS-2B cells, m1 mAchR activation in RBL-2H3 cells, or TXA_2α activation in HEK 293 cells was not significantly influenced by H-89. Thus, analyses of various G_s- and G_q-coupled receptor function suggest that receptor antagonism by H-89 appears selective forbeta ARs.

Discussion

For several reasons, the use of pharmacological inhibitors of PKA (and H-89 in particular) represented an attractive means by which to address an a priori hypothesis questioning the role of PKA in beta-2 AR desensitization in human airway smooth muscle. Most importantly, pharmacological agents could affect the entire population of cells, whereas the effectiveness of transfection-based strategies would be limited by the level of transfection efficiency (∼50% using a replication-deficient adenovirus) obtainable in HASM (Penn et al., 1998). This was an important consideration because the alterations in maximalbeta-2 AR responsiveness induced by PKA are frequently small (Clark, 1986) and the identification of an experimental effect caused by inhibition of PKA would likely require significant PKA inhibition in all cells. H-89 was among the most attractive compounds considered, having been characterized in vitro as a highly potent and selective inhibitor of PKA (Chijiwa et al., 1990). Moreover, H-89 has been used in numerous studies to date examining the role of PKA in the regulation of cellular signaling and physiological effects induced by activation of a wide range of GPRs.

As mentioned, the principal drawback in using most pharmacological kinase inhibitors in intact cells is the apparent requirement for 100- to 1000-fold higher levels than those determined effective in cell-free assays. Effective inhibition of a target kinase in a given cell type is dependent in part on 1) the permeability of the inhibitor; 2) the relative stoichiometry of inhibitor, the intended target kinase, and any competing target enzymes; 3) the relative potencies of the inhibitor for all potential enzyme targets; and 4) the competitive nature of the inhibitor with intracellular ATP (MacKintosh and MacKintosh, 1994). This last consideration alone dictates that among those kinase inhibitors used in this study, intracellular concentrations of inhibitor must be orders of magnitude higher than those determined for kinases in vitro, where reportedK_i values are determined under conditions of low micromolar levels of ATP. One of the first investigations into the effect of H-89 in intact cells demonstrates that despite an in vitro K_i of H-89 for PKA of 48 nM, inhibition of cAMP-induced neurite outgrowth in PC12D cells was only observable at micromolar concentrations of H-89, with aK_i of ∼10 μM (Chijiwa et al., 1990).

With such caveats in mind, we examined the effects of various kinase inhibitors in a frequently used paradigm for characterizing agonist-specific beta-2 AR desensitization. HASM cells were treated with a saturating concentration of beta agonist, washed, and then challenged again with a saturating concentration ofbeta agonist, with the diminution of cAMP accumulation that occurs due to beta agonist pretreatment serving as an index of beta-2 AR desensitization. Pretreatment with high levels of various kinase inhibitors before and throughout pretreatment with ISO might presumably provide insight into the role of PKA in agonist-specific desensitization. We hypothesized that effective inhibition of PKA would attenuate agonist-specific desensitization to a degree commensurate with the cAMP-mediated beta-2 AR desensitization we had observed in these cells (Penn et al., 1998). The surprisingly large effect of H-89, relative to that observed for staurosporine, raised questions regarding potential nonspecific effects of H-89, which were also suggested by the small effect of H-85 in attenuating beta-2 AR desensitization (Fig. 1).

The subsequent finding that pretreatment with H-89 caused a significant inhibition of cAMP accumulation in HASM stimulated with submaximal concentrations of ISO strongly implied that H-89 antagonizedbeta-2 AR responsiveness to agonist at some level upstream of adenylyl cyclase. Indeed, a very recent study by Clark and coworkers (January et al., 1997) noted that HA6 cells pretreated with 5 μM H-89 alone exhibited a 4- to 5-fold increase in the EC₅₀ for epinephrine activation of adenylyl cyclase, whereas the addition of 5 μM H-89 directly to assays of adenylyl cyclase caused a 20- to 50-fold increase in the EC₅₀ of epinephrine stimulation; these data led the authors to conclude that H-89 was unsuitable for use in desensitization experiments because of its uncoupling ofbeta-2 AR activation.

Our studies extend this observation to define H-89 as a potent and selective inhibitor of beta-2 AR and beta-1 AR ligand binding. A reproducible inhibition of both [¹²⁵I]IPIN and [³H]dihydroalprenolol binding tobeta ARs was observed, with similar results obtained in binding to both overexpressed recombinant beta-2 AR (in Sf9 membranes) and endogenously expressed beta-2 AR (in BEAS-2B cells). The inhibition of binding by H-89 was not caused by alterations in binding buffer pH (10 μM H-89 had no effect on pH) and was reproducible using different types of binding buffers. Results were independent of supplier or lot number of H-89.

Based on findings from the present study and those of others, the following characteristics can be ascribed to H-89: for PKA inhibition, an in vitro K_i of 48 nM and an in vivoK_i of ∼10 μM (see discussion above); K_i values of 180 and 350 nM for inhibition of beta-2 AR and beta-1 AR binding, respectively; and a variableK_i value for beta-2 AR antagonism predicted by Schild analysis of H-89-mediated inhibition of cAMP generation (Fig. 4C) that varies directly with H-89 concentration (e.g., a predicted K_i of ∼200 nM with 0.1 μM H-89 but a K_i of ∼1.3 μM with 10 μM H-89).

The receptor binding properties of H-89 could possibly be inferred from our understanding of the structurally related tetrahydroisoquinolines and their analogs, which have been shown to be effective agonists/antagonists for both beta-adrenergic and TXA₂ receptors (Christoff et al., 1997;Fraundorfer et al., 1994; Shams et al., 1997). Although the structural characteristics of H-89 that confer its affinity to beta ARs are not obvious, one might predict, based on the analysis of ligand/beta-2 AR interactions described by Jasper and Insel (1992), that the sulfonamide or amine group could interact with Asp¹¹³ of the beta-2 AR. It should also be noted that H-89 lacks the hydroxyl groups on the aromatic ring (isoquinoline group in H-89) that appear to be required forbeta-2 AR activation by ligand. Based on the relative affinities of H-7, H-85, and H-89 for beta ARs, it is also plausible to suggest that the large hydrophobic side chain possessed by H-89 and H-85 (Fig. 7) may also serve to anchor these molecules in the plasma membrane (or to a region of thebeta AR distinct from the ligand binding domain) and thus increase their effective affinities for the beta ARs, perhaps in a manner not unlike that demonstrated for thebeta agonist salmeterol (Clark et al., 1996; Green et al., 1996).

Although experiments examining the effects of H-89 on both receptor binding and receptor-mediated signaling suggest that the action of H-89 at the receptor level are selective for beta ARs, we did observe that binding to other receptors could be slightly inhibited by concentrations of H-89 that might be used to inhibit PKA in intact cells. Although this may not be problematic under many experimental conditions, it does suggest that H-89 as well as other inhibitors (e.g., Bis IX and H-7) used at high concentration in intact cell or organ system models may inhibit the function of other receptors, particularly those activated by low levels of agonist (e.g., by circulating hormones, or in a paracrine or autocrine manner). Thus, it would appear prudent in such circumstances not only to establish the selectivity of agents with respect to intracellular kinases but also to consider potential effects on receptors that are relative to the system.

An additional concern associated with the use of high levels of kinase inhibitors is cellular toxicity. Whether enhanced cell morbidity or mortality represents a nonspecific side effect or is necessarily a consequence of effective inhibition of target kinases is unclear. In both HASM and BEAS-2B cells, prolonged incubation (>2 h) with 10 μM H-89 leads to a reduction in FSK-stimulated cAMP accumulation (R. B. Penn and J. L. Benovic, unpublished observations). A 16-h incubation of both BEAS-2B and HASM cells (but not COS-1 cells) with 1 μM H-89 causes an ∼20% loss of cells in culture by detachment compared with DMSO-treated control, whereas 16-h treatment with 500 μM H-7 causes a small loss (∼20%) in COS-1 cells and a larger loss (∼30%) in HASM.

In summary, the present study demonstrates that the significant attenuation of agonist-specific desensitization of thebeta-2 AR by the PKA inhibitor H-89 in two different physiologically relevant cell types is primarily a result ofbeta-2 AR antagonism. Two other inhibitors, KT5720 and H-7, shown effective in inhibiting PKA-mediated functions in numerous other cell types, failed to influence beta-2 AR desensitization or inhibit PKA-induced protein phosphorylation in HASM, thus suggesting the cell-specific nature of inhibitors. Additional studies examining the pharmacological and functional properties of various kinase inhibitors demonstrate the need to think beyond the obvious requirements for kinase selectivity and potency and consider the numerous nonspecific effects that may occur with use of such agents in investigation using intact cells.