Introduction

Top Summary Introduction Procedures Results Discussion References

Neurons and pancreatic islet cells have much in common. Both types of cells are electrically excitable and both express a number of proteins found in the CNS. Recently, it has become apparent that machinery used by the CNS for rapid ionotropic synaptic transmission also is found in the endocrine pancreas. GABA, released from cells, has been postulated to activate GABA_A receptors on  cells and thereby attenuate glucagon secretion when glucose levels rise (Rorsman et al., 1989). Functional AMPA-type GluRs also are present in islets, but their role in islet physiology is not completely understood (Inagaki et al., 1995; Weaver et al., 1996).

The membrane potential of both neurons and islet cells is vital to their physiological function. Because both cell types express voltage-dependent Ca²⁺ channels, depolarizations of the membrane often lead to increases in [Ca²⁺]_i, which initiates hormone and neurotransmitter release and other long term biochemical alterations. Although excitatory (depolarizing) stimuli usually are associated with channels permeable to Na⁺ ions, anion channels also may depolarize cells under certain conditions. For example, activation of GABA_A receptors causes depolarization in neurons after excessive inhibitory synaptic inputs (Staley et al., 1995). GABA receptors activation in the gastropancreatic cell line, AR42J, also has been shown to increase [Ca²⁺]_i. The exact mechanism by which this occurs has not been completely elucidated. In one case, it was suggested that collapse of the normal Cl gradient coupled with outward fluxes of HCO₃, which also is permeable through GABA receptors, causes the depolarizing response (Staley et al., 1995).

Permanent lines derived from pancreatic islet cells often contain and secrete the appropriate hormone and maintain the electrical excitability of the parent cell types. In addition, some are known to express glutamate and GABA_A receptors (Rorsman et al., 1989; Von Blankenfeld et al., 1995). However, to our knowledge, naturally expressed functional glycine receptors have not been detected in any permanent cell line previous to this report.

We have characterized the expression patterns, functional properties, and structural features of amino acid receptors expressed in GK-P3 cells, which are derived from pancreatic  cells. These cells were found to express both strychnine-sensitive glycine receptors and AMPA-type GluRs. We used this cell line to examine how both glycine and AMPA receptors regulate [Ca²⁺]_i using fluorescent imaging with Fura-2. Our studies revealed that the kinetic and ionic permeability properties of glycine and AMPA receptors in GK-P3 cells were similar to those receptors found in neurons. However, we observed that the activation of either strychnine-sensitive glycine receptors, traditionally regarded as inhibitory, or AMPA receptors increased [Ca²⁺]_i. The Ca²⁺ seemed to enter the cells largely through voltage-gated Ca²⁺ channels. Thus, because of their ability to depolarize GK-P3 cells, glycine receptors and GluRs may be capable of stimulating or potentiating peptide hormone secretion from islets. In addition, our results suggest that because neither islet glycine receptors nor GluRs are Ca²⁺ permeable, their ability to increase intracellular Ca²⁺ depends on their coexpression with voltage-gated Ca²⁺ channels and on the ionic gradients present in the cells in which they are expressed.

	Experimental Procedures

Top Summary Introduction Procedures Results Discussion References

Materials. AMPA was obtained from Research Biochemicals (Natick, MA). Kainic acid, NMDA, leupeptin, pepstatin A, and amphotericin B were obtained from Sigma Chemical (St. Louis, MO). CNQX and dizocilipine maleate (MK-801) were obtained from Research Biochemicals. Cyclothiazide was obtained from Eli Lilly (Indianapolis, IN). The anti-GluRB/C antibody was a gift from Dr. Robert Wenthold. Anti-GluRA and anti-GluRD antibodies were purchased from Chemicon International (Temecula, CA). The anti-glycine receptor antibody was a gift from Dr. Heinrich Betz. Alkaline phosphatase-labeled goat anti-rabbit and goat anti-mouse antibodies were purchased from DAKO (Carpinteria, CA). Bicinchoninic acid protein assay reagents were purchased from Pierce Chemical (Rockford, IL). All other chemicals were of reagent grade or higher.

Tissue preparation and culture. GK-P3 cells were derived from an insulinoma removed from a transgenic mouse expressing a fusion gene containing the rat glucokinase upstream promoter (1000 to +14) linked to the SV40 large T antigen gene (Jetton TL, et al., 1998). Experiments were performed on GK-P3 cells grown for 6-20 passages in Dulbecco's modified Eagle's medium with 10% (v/v) fetal bovine serum.

Electrophysiological methods. Ionic currents were measured by whole-cell patch-clamp electrophysiological methods using either an Axopatch 200 (Axon Instruments, Foster City, CA) or a Dagan PC-1 Amplifier (Dagan Corporation, Minneapolis, MN). Signals were filtered at 2000 Hz (3 dB, eight-pole Bessel filter; Frequency Devices, Haverhill, MA), digitized at 5000 Hz using an ITC-16 interface (Instrutech, Great Neck, NY), and recorded on a Macintosh Quadra 800 or IIfx computer. The cells were perfused continuously with an extracellular solution containing 140 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl₂, 1.0 mM MgCl₂, and 5 mM HEPES, pH 7.2. In some experiments, an islet external solution was used consisting of 145 mM NaCl, 2.5 mM KCl, 2.5 mM CaCl₂, 1.0 mM MgCl₂, 2 mM glucose, and 10 mM HEPES, pH 7.3. Most currents were measured in conventional whole-cell mode using patch pipettes filled with a solution that consisted of 135 mM CsCl, 1 mM MgCl₂, 11 mM EGTA, and 10 mM HEPES, pH 7.3. In some cases, electrical access to the cell was obtained using the perforated patch method. In these experiments, patch pipettes were filled with a solution containing 10 mM KCl, 10 mM NaCl, 70 mM K₂SO₄, 2 mM MgCl₂, and 10 mM HEPES, pH 7.3, and 240 µg/ml amphotericin B. Rapid agonist application experiments were performed by lifting a single cell from the bottom of the culture dish after whole-cell recording was established. The cell was perfused continuously by control external solution through one side of a glass  tubing pipette. Agonist-containing solution flowed in the other side of the pipette. The solution bathing the cell was exchanged by stepping the application pipette with a piezoelectric manipulator so that the side of the  pipette containing agonist solution faced the cell. The time required for complete agonist exchange was 10 msec. In situations that did not require rapid application, cells remained attached to the bottom of the culture dish for the duration of the recording. In these cases, the cell was continuously bathed in control solution from a flow pipe connected to eight reservoirs. Solution exchange was accomplished by opening a single valve at a time.

Ratiometric imaging. The experimental setup for ratio imaging studies consisted of a Nikon Diaphot inverted microscope with a 40× oil immersion Fluor objective. Epifluorescence illumination was provided by a 100-W mercury burner. The excitation wavelength was set by bandpass filters of 340 or 380 nM and changed by a Ludl high-speed filter wheel. The intensity of illumination was set by neutral density filters, also controlled by the Ludl filter wheel. Images were collected through a Dage CCD72 video camera and a DAGE GenIISys Image Intensifier. Computer acquisition of the images and control of the filter wheel were provided by the program Simca (Compix), which also was used for analyzing the images. With this system, ratio images were recorded at a speed of ~1 frame/sec.

Immunological methods. GK-P3 cells were harvested from tissue culture flasks with a cell scraper and pelleted by centrifugation at 2000 rpm for 5 min at 4° in a Beckman TJ-6 centrifuge. The supernatant solutions were aspirated, and the cell pellets were rinsed twice with extraction buffer that consisted of 20 mM 3-(N-morpholino)propanesulfonic acid, pH 7.5, 500 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 0.002 mM leupeptin, and 0.002 mM pepstatin A. The pellets then were resuspended in 3 ml of extraction buffer. The resuspended samples were sonicated twice for 15 sec at 30-sec intervals and then centrifuged at 100,000 × g for 1 hr at 4°. The supernatant fractions were aspirated away, and the crude membrane pellets were rinsed with extraction buffer and resuspended in 200 µl of a buffer consisting of 20 mM 3-(N-morpholino)propanesulfonic acid, pH 7.5, 1 mM phenylmethylsulfonyl fluoride, and 0.1% (w/v) SDS. Islet and brain membranes were prepared as described previously (Weaver et al., 1996).

RNA PCR analysis. GK-P3 cells were cultured as described. RT-PCR was performed using a two-step RNA PCR kit (Perkin-Elmer-Cetus) according to the supplier's directions. RNA was isolated according to the AGCP method (Chomczynski and Sacchi, 1987) from rat brain, GK-P3 cells, and rat liver and then poly(A)⁺ selected using the PolyA Tract system (Promega, Madison, WI). Poly(A)⁺ RNA (100 ng) was used for RT-PCR amplification. Primer pairs were designed to correspond to different exons so that unspliced precursor mRNA, if present, would amplify as larger fragments because of intron or introns. Poly(A)⁺ RNA was converted to cDNA at 42° using primers specific to either GluRB (nt 3051-CCATTGTGTAAGGCACTCAGAAGGTTCC-nt 3090) or GluRC (nt 2933-TAGGCCCGGGCAAAGCAAAAAGATTTCAATG-nt 2954) (GluRB- and GluRC-specific primers were a gift from R. Emeson). Glutamate receptor sequences then were amplified using a primer to a conserved transmembrane region designed to amplify all of the AMPA receptor subtypes (TM1 corresponds to GluRB "flip" nt 861-TATGAAATCTGGATGTGCAT-nt 1681) and either GluRB or GluRC primers by melting at 95° for 1 min, annealing at 60° for 1 min, and extending at 72° for 2.5 min for 30 cycles. PCR products were separated on 1.2% (w/v) agarose gels, blotted onto Zeta Probe nylon membrane (BioRad) using 0.4 N NaOH, probed with oligomers (100 ng) corresponding to a conserved region in AMPA receptors (GluRB "flip" nt 2635-GCTTCCCGAGTCCTTGGGTCC-nt 2615), and detected with autoradiography.

	Results

Top Summary Introduction Procedures Results Discussion References

Amino acid receptors in pancreatic GK-P3 cells. We detected functional receptors for glycine, glutamate, and GABA in the pancreatic  cell line GK-P3. Glycine receptors appeared in 98% of the GK-P3 cells tested. However, we failed to observe glycine-elicited currents in the glutamate and GABA_A receptor-containing pancreatic  cell lines  TC-6 and  TC-9 (34 cells). Responses to glutamate were seen in 60% of GK-P3 cells, and GABA_A receptors were found in 23% of the cells tested. The average amplitude of currents evoked by various ligands in GK-P3 cells is shown in Table 1. Although cells containing glycine receptors often coexpressed GluRs, GABA_A receptors and GluRs appeared together in only 1 of the 30 GK-P3 cells tested. Because the properties of GABA receptors in pancreatic cell lines (Tyndale et al., 1994; Von Blankenfeld et al., 1995) and in pancreas (Rorsman et al., 1989) have been more examined thoroughly, we focused our attention on characterization of glycine receptors and GluRs in these cells.

Pancreatic receptors are closely related to neuronal receptors. Currents elicited by 100 µM glycine in GK-P3 cells seemed to arise from the glycine receptor associated with chloride channels. These currents were not due to activation of the glycine receptor associated with the NMDA/receptor complex. Glycine-evoked currents were blocked completely by 1 µM strychnine (Fig. 1A) and were insensitive to the NMDA receptor channel blocker MK-801 (Fig. 1B). Currents also were activated by two other agonists for the native glycine receptor: 300 µM taurine and 300 µM -alanine (Fig. 1C, Table 1). Zinc was an effective antagonist of these receptors (Fig. 1D) with 50 µM Zn²⁺ blocking an average of 88 ± 3% of the currents evoked by 100 µM glycine (10 cells). The concentration-response relation for glycine gave an EC₅₀ value of 90.4 (95% confidence interval, 66.1-123) µM (11 cells) and a Hill coefficient of 2.34 ± 0.28. Ionic substitution experiments were performed in which the extracellular chloride was replaced with HCO₃. The magnitude of the rightward shift in the I-V relation indicated that the Cl/HCO₃ permeability ratio was 5.6 ± 0.4 (four cells), which is identical to that found previously for neuronal glycine receptors (Bormann et al., 1987).

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Glycine-evoked currents in GK-P3 cells are indicative of neuronal inhibitory glycine receptors. In all cases, currents were recorded under conventional whole-cell voltage-clamp (Vhold = 60). Solid bars, application of agonists. Stippled bars over the current traces, application of antagonists. A, Inward currents evoked by glycine were completely and reversibly blocked by 1 µM strychnine. B, Glycine-evoked currents were completely insensitive to block by the NMDA receptor antagonist, MK-801. C, Currents also were evoked by glycine receptor agonists taurine and -alanine. D, Zinc (50 µM) blocked currents mediated by GK-P3 glycine receptors.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   AMPA receptors in GK-P3 cells showed properties similar to those of neuronal AMPA receptors. Solid bars, application of agonist. Stippled bars over the current traces, antagonists or modulators. All currents were recorded in standard whole-cell mode with voltage clamped at 60 mV. A, Rapid application of 1 mM glutamate produced currents with rapid desensitization. B, Steady state currents evoked by 300 µM glutamate by the AMPA receptor were potentiated by cyclothiazide. C, Currents evoked by the AMPA receptor agonist kainate (300 µM) were blocked by 10 µM CNQX, a competitive antagonist of neuronal AMPA receptors. D, Application of 300 µM glutamate depolarized GK-P3 cells and elicited action potentials typical of pancreatic  cells. Voltage recordings were made using perforated patch mode at the resting membrane potential, which in this cell was 52 mV.

View larger version (35K): [in this window] [in a new window]   Fig. 3.   Western blots and RT-PCR detected glycine receptor (GlyR) and AMPA receptor subunits in GK-P3 cells. A, Shown are Western blots of membrane proteins isolated from brain (20 µg), GK-P3 cells (30 µg), and isolated rat pancreatic islets (25 µg) probed with an antibody that recognizes glycine receptor subunits, mAB4a. Bands of similar size (50 kD) were seen in all three preparations. Islet membranes also contain some bands of lower apparent molecular weight. B, Western blots of membrane proteins from brain (20 µg) and GK-P3 cells (30 µg) probed with the subunit-specific anti-GluR antibodies GluRA, GluRB/C, and anti-GluRD. C, Results of RT-PCR analysis of GluR expression in GK-P3 cells are illustrated. PCR amplicons were generated from poly(A)+ RNA from rat brain, GK-P3 cells, and rat liver. RNA was converted to cDNA using primers specific to GluRB or GluRC (3' end). GluR sequences were amplified using TM1 and GluRB or GluRC primers, separated on a 1.2% (w/v) agarose gel, transferred to nitrocellulose, then probed with GluR-specific oligonucleotides.

Receptor activation regulates [Ca²⁺]_i. Membrane depolarization is a critical step in the signal transduction mechanism that links serum glucose levels to insulin secretion from pancreatic  cells. Therefore, we examined the ability of glycine and glutamate to alter the membrane potential of GK-P3 cells through actions at their respective receptors. In cases where both whole-cell capacitance and current amplitudes were measured, GK-P3 cells exhibited an average steady state current density of 21 ± 2.7 pA/pF (63 cells) on application of 100 µM glycine, whereas 300 µM glutamate gave an average current density of 1.99 ± 0.32 pA/pF (11 cells). Whole-cell recordings of membrane potential using perforated patch revealed that the application of 300 µM glutamate depolarized these cells an average of 27 ± 6.8 mV (five cells) (Fig. 2D) from their normal resting membrane potential of 50.5 ± 3.6 mV (five cells). This depolarization led to the firing of repetitive action potentials similar to those observed in pancreatic  cells on exposure to high glucose (Rorsman and Trube, 1986). Because of uncertainty in the effects of whole-cell patch-clamp on the normal Cl gradients in GK-P3 cells, we did not conduct a similar experiment with glycine receptor agonists but instead turned to ratiometric imaging of [Ca²⁺]_i.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Glutamate and glycine receptor activation increased [Ca2+]i in GK-P3 cells. Shown are ratiometric measurements of [Ca2+]i in Fura-2-loaded GK-P3 cells. Solid bars, agonists were perfused on the cells during the time indicated by the. Stippled bars, perfusion of modulators or antagonists. Each trace, measurements from representative cells in the same field of view to illustrate the variety of response observed. Typically, 20-30 cells were measured at the same time.

	Discussion

Top Summary Introduction Procedures Results Discussion References

We characterized the amino acid neurotransmitter receptors expressed in a cell line derived from the endocrine pancreas. These cells were used to examine the mechanisms of Ca²⁺ signaling mediated by these ligand-gated ion channels. The major findings of our work are (1) strychnine-sensitive glycine receptors and AMPA-type GluRs were expressed in the  cell-derived cell line, GK-P3; (2) the structural, pharmacological, and ionic permeability properties of GK-P3 glycine and AMPA receptors were very similar to those of neuronal receptors, (3) activation of glycine receptors, which are anion channels, depolarized GK-P3 cells and activated voltage-gated Ca²⁺ channels, causing increases in [Ca²⁺]_i, (4) AMPA receptor cation channels also were capable of increasing [Ca²⁺]_i through a similar mechanism.

In addition to neuronal cells derived from pluripotent precursors (Turetsky et al., 1993; Younkin et al., 1993), a number of immortalized lines of islet cells have been shown to express functional ligand-gated ion channels. Glutamate receptors have been reported in  cell-like MIN6 cells (Gonoi et al., 1994), and GABA receptors have been reported in native  and  cells based on immunological analysis and in  cells based on electrophysiological measurements (Rorsman et al., 1989). GABA receptors also have been detected in other gastropancreatic cell lines, such as AR42J and RIN (Tyndale et al., 1994; Von Blankenfeld et al., 1995). However, the GK-P3 cells represent the first transformed line reported to harbor functional, strychnine-sensitive glycine receptors.

The expression pattern of amino acid receptors in islet cell lines often approximates that of native islets. Previously, we have shown that AMPA receptors are distributed in all islet cells, except somatostatin-containing  cells, and that kainate receptors are found in TC cells as well as native  cells (Weaver et al., 1996). Others have found immunoreactivity for GABA receptor subunits in native  cells (Rorsman et al., 1991). Western blots of GK-P3 and islet cell membrane proteins using a glycine receptor antibody revealed proteins of an apparent molecular weight consistent with that of neuronal glycine receptors. These data suggest that glycine receptors expression in GK-P3 cells may be indicative of glycine receptor expression in native pancreatic islets. Previous reports have demonstrated that perfusion of islets with relatively high concentrations of glycine transforms glucose-induced fluctuations of [Ca²⁺]_i into sustained increases (Tengholm et al., 1992; Hellman et al., 1994). Our data suggest that these effects may be mediated by islet glycine receptors. Determining the strychnine sensitivity of the effects of glycine on insulin secretion clearly will be an area of interest for further research.

As mentioned, the mouse insulinoma cell line MIN6 has been shown to possess functional ionotropic GluRs (Gonoi et al., 1994). In contrast to the current results, MIN6 cells were reported to express active AMPA and NMDA receptor types when measured by patch-clamp. Analysis of mRNA also showed transcripts for a variety of GluR subunits, including NMDAR1, NMDAR2D, GluRB, GluRC, and KA2. We did not extensively analyze the expression of all these mRNA species and thus cannot rule out the possibility that GK-P3 cells harbor mRNAs encoding other GluR types. However, the evidence presented here clearly shows that the only functional GluRs detectable in GK-P3 cells belong to the AMPA receptor subtype. Our previous immunological studies on isolated rat islets did not reveal expression of NMDAR1 protein (Weaver et al., 1996), but islets did show GluRB/C and GluRB immunoreactivity very similar to that seen in GK-P3 cells. Thus, it seems that the constellation of GluRs expressed in GK-P3 cells is more similar to native islets than to those found in MIN6 cells and that AMPA receptors and not NMDA receptors are the main type of GluRs in pancreatic islets. Consistent with this proposal, perfusion experiments with whole pancreas have shown that activation of AMPA receptors, but not NMDA receptors, potentiates secretion of insulin (Bertrand et al., 1992) and glucagon (Bertrand et al., 1993).

Glycine and GABA receptors in the CNS generally are considered to be inhibitory because of their selectivity for anions. However, a number of recent observations suggest that in some situations GABA receptor activation also can be excitatory (Staley et al., 1995; Von Blankenfeld et al., 1995). The exact reason for this remains debatable, but the two most attractive hypotheses suggest that depolarizing responses to GABA are caused by variations in the concentration gradient for the permeant anions Cl (Misgeld et al., 1986) or HCO₃ (Staley et al., 1995). The permeability properties of GK-P3 glycine receptors are identical to those of glycine or GABA receptors expressed in spinal neurons (Bormann et al., 1987) and thus represent a convenient system in which this issue can be examined. In many GK-P3 cells, glycine receptor activation depolarized the cells sufficiently to activate voltage-dependent Ca²⁺ channels, leading to increases in [Ca²⁺]_i. La³⁺, which blocks Ca²⁺ channels, but not glycine receptor currents, completely inhibited the increase in [Ca²⁺]_i normally seen with glycine. By contrast, the extracellular concentration of HCO₃ had little influence on the Ca²⁺ signals observed in these studies. It seems likely that normal intracellular concentrations of Cl and HCO₃, which were undisturbed during our Ca²⁺ imaging experiments, are sufficient to generate a net efflux of anions on receptor activation. This depolarizes the cell and opens voltage-dependent Ca²⁺ channels, allowing influx of Ca²⁺. GABA receptor activation in AR42J cells (Von Blankenfeld et al., 1995) likewise has been shown to increase [Ca²⁺]_i. Further investigations into the properties of GK-P3 and AR42J cells may help to reveal the mechanisms by which glycine and GABA can become excitatory neurotransmitters. In the future, it will be important to investigate the expression of strychnine-sensitive glycine receptors in native islets and to attempt to determine whether the anion gradients in native islet cells would cause glycine receptors to function as observed in GK-P3 cells.

Glutamate also was able to depolarize GK-P3 cells enough to increase [Ca²⁺]_i. Because La³⁺ blocked the responses to glutamate, the majority of the signals generated by AMPA receptor activation were probably due to flux of Ca²⁺ through voltage-gated Ca²⁺ channels. In other experiments, extracellular Na⁺ was replaced with the impermeant cation N-methylglucamine, and this also prevented AMPA receptor activation from increasing [Ca²⁺]_i (Partridge JG and Verdoorn TA, unpublished observations). This result is not particularly surprising based on the low average Ca²⁺ permeability exhibited by the AMPA receptors in GK-P3 cells. However, human embryonic kidney 293 cells expressing recombinant AMPA receptors that contain an edited GluRB subunit and exhibit low Ca²⁺ permeability mediate large increases in [Ca²⁺]_i as measured by the same assay (Utz and Verdoorn, 1997). Thus, factors other than Ca²⁺ permeability may dominate the ability of AMPA receptors to alter [Ca²⁺]_i. These factors may include receptor density, the other ion channels that are coexpressed with AMPA receptors in different cell types, and the differential ability of cell types to resist changes in measurable [Ca²⁺]_i through Ca²⁺-buffering proteins or by the ability of the cell to extrude Ca²⁺ from intracellular solutions.

The exact role of amino acid neurotransmitter systems in pancreatic islets remains uncertain. The postulated role of GABA in communicating between  and  cells (Rorsman et al., 1989) remains an important model for exploration of this issue. In this vein, glutamate and glycine could mediate signaling between different cells within an islet in addition to subserving communication between islets and the central nervous system. Recently, we detected high affinity, Na⁺-dependent glutamate uptake in isolated pancreatic islets, which has the capability of clearing glutamate from extracellular spaces and preventing chronic depolarization of the receptors there (Weaver et al., 1998). Similar processes may facilitate GABA and glycine signaling in islets.

It seems likely that fine tuning of islet hormone secretion under normal physiological conditions may be due in part to the presence of amino acid neurotransmitter receptors. A number of modulatory processes alter the properties of amino acid neurotransmitter receptors, and these may contribute to subtle regulation of islet hormone secretion. For example, zinc, which blocks glycine receptors and some types of GABA_A receptors, is found at high concentrations in the insulin-containing secretory vesicles of  cells where it is complexed with insulin. Changes in extracellular zinc concentration on insulin secretion could modulate the activity of glycine receptors and thereby modulate the electrical activity of islets.

Finally, the presence of amino acid neurotransmitter receptors in islets offers potential pharmacological targets for modifying islet function in pathological situations. Indeed, block of GluR desensitization by compounds such as diazoxide, which also is an insulin secretagogue acting at ATP-sensitive K⁺ channels, may underlie some of the therapeutic benefits of this class of drugs.