Effect of Anoxia on Cyclic Nucleotides and Inositol Phosphate Turnover in Cardiac Myocytes

J Mol Cell Cardiol 28, 1857–1866

Effect of Anoxia on Cyclic Nucleotides and Inositol Phosphate Turnover in Cardiac Myocytes Timothy P. Geisbuhler and Tammie L. Schwager Department of Physiology, Kirksville College of Osteopathic Medicine, Kirksville, MO 63501, USA (Received 13 July 1995, accepted in revised form 25 March 1996) T. P. G  T. L. S. Effect of Anoxia on Cyclic Nucleotides and Inositol Phosphate Turnover in Cardiac Myocytes. Journal of Molecular and Cellular Cardiology (1996) 28, 1857–1866. The loss of 5′-nucleotides (especially ATP and GTP) from cardiac muscle cells is a distinguishing feature of myocardial ischemia. Isolated adult rat cardiac myocytes were used as a model system to determine whether GTP depletion could affect (1) the ability of the myocytes to synthesize cyclic GMP (cGMP), or (2) the ability of the myocytes to respond to aadrenergic challenge. Myocytes were made anoxic for 30- or 60-min periods, then challenged with either 1 m sodium nitroprusside (NaNP) for 1 min or 40 l norepinephrine (NE) for 20 min. The cells were extracted and the extracts assayed for cyclic GMP (NaNP challenge) or phosphoinositides (NE challenge). When challenged with NaNP, anoxic myocytes made up to five-fold more cGMP than aerobic controls (1401±353 fmol cGMP/ mg cell protein in anoxic cells v 121±23 fmol/mg in aerobic controls). Phosphoinositide turnover was reduced in anoxic cells v aerobic controls. Stimulation of this pathway by NE was reduced two-fold after 30 min of anoxia, and abolished after 60 min of anoxia. Similar results were obtained with 30 l and 60 l phenylephrine. The authors concluded that nucleotide depletion under anoxic conditions has no effect on the production of cyclic GMP, but may interfere with the linkage of a-adrenergic receptors to phosphatidylinositol breakdown.  1996 Academic Press Limited

K W: Heart; Rat; Cardiac myocytes; Cyclic GMP; Inositol phosphates; Anoxia; a-adrenergic stimulation; Norepinephrine; Phenylephrine.

Introduction

cyclic GMP (cGMP); depletion of GTP could leave the cell unable to synthesize this important second messenger. Both a- and b-adrenergic receptors transmit their signals using G-proteins. Anoxic GTP depletion could impair the cells’ ability to respond to an incoming adrenergic signal. The subject of myocardial a-adrenoceptors has been recently reviewed (Endoh, 1991; Lamers et al., 1992; Fedida et al., 1993; Terzic et al., 1993). Heart a-adrenergic receptors are of the a1 type, which when stimulated by phenylephrine, norepinephrine, or methoxamine produce inositol 1, 4,5-trisphosphate (InsP3) and diacylglycerol from phosphatidylinositol bisphosphate (PIP2; Endoh et al., 1991; Mouton et al., 1991a). Whether the production of InsP3 results in the release of calcium

The loss of 5′-nucleotides (especially ATP and GTP) from cardiac muscle cells is a distinguishing feature of myocardial ischemia. The depletion of adenine nucleotides is associated with reduced mechanical function in post-ischemic heart, although the specific factors responsible for this dysfunction are not clearly understood. Guanine nucleotides are also depleted from ischemic tissue along with adenine nucleotides, but the contribution of this phenomenon to the loss of post-ischemic function in heart has not been studied. GTP depletion in myocytes could be responsible for a number of cellular metabolic dislocations. Guanylate cyclase uses GTP as a substrate to form

Please address all correspondence to: Timothy P. Geisbuhler, Department of Physiology, Kirksville College of Osteopathic Medicine, 800 West Jefferson St., Kirksville, MO 63501, USA

0022–2828/96/091857+10 $18.00/0

1857

 1996 Academic Press Limited

1858

T. P. Geisbuhler and T. L. Schwager

to the cytosol in heart is a matter of some debate (Endoh, 1991; Movsesian et al., 1985; Nosek et al., 1986). Recent studies with cardiac myocytes showed elevated cytosolic calcium after a-adrenergic stimulation (Eckel et al., 1991). Less dramatic results were obtained using skinned cardiac muscle fibers and caged InsP3 (Kentish et al., 1990), indicating a regulatory role for InsP3 rather than a direct stimulation of calcium release. InsP3 directly stimulates calcium release from intracellular stores in other tissues, particularly in smooth muscle. Cell injury in cardiac myocytes is generally divided into three stages: (1) the early phase of anoxia, lasting until the completion of rigor contracture; (2) a middle phase, during which there is severe energy depletion in the cells; and (3) the late phase of cell death and necrosis (Rim et al., 1990). Development of rigor contracture (stage 1) is the result of a catastrophic loss of cytosolic ATP (Bowers et al., 1992) and appears to be independent of extracellular calcium concentrations (Rim et al., 1990). The exact role of cytosolic calcium in anoxic or ischemic injury is still unclear. Rim et al. (1990) feel that cytosolic calcium may not be related to the development of irreversible injury (stages 2 and 3) in heart, while the data of Allshire et al. (1987) suggest that loss of calcium homeostasis is involved in this progression toward cell death. The authors developed a model to explain injury to anoxic cardiac myocytes based on the depletion of GTP in the cells. This model centered on three points: (1) depletion of GTP in anoxic myocytes degrades the cells’ ability to produce cyclic GMP (cGMP); (2) depressed levels of cGMP allow phospholipase C-dependent degradation of phosphatidylinositol bisphosphate (PIP2) to diacylglycerol (DG) and inositol 1,4,5-trisphosphate (InsP3), thereby allowing the release of calcium to the cytosol [cGMP has been proposed as an inhibitor of phospholipase C (Rapoport, 1986; Hirata et al., 1990)]; and (3) depletion of GTP interferes with the ability of the myocytes to accept incoming adrenergic signals by decreasing activity of the Gproteins (see Fig. 1). This model was tested using isolated adult rat cardiac myocytes in a previously developed anoxia protocol. The authors investigated whether anoxia-induced GTP depletion could affect (1) the ability of the myocytes to synthesize cGMP; or (2) the ability of the myocytes to respond to aadrenergic challenge with norepinephrine or phenylephrine. The data suggest that anoxia does not affect basal cGMP levels in cardiac myocytes; further, anoxic cells produce more cGMP than aerobic controls if stimulated with sodium nitroprusside. Finally, the data show that cardiac

myocytes lose the ability to respond to the a1 agonists norepinephrine and phenylephrine under anoxic conditions.

Materials and Methods Myocyte preparation Cardiac myocytes were isolated from adult rat hearts by collagenase digestion as previously described (Geisbuhler et al., 1992). The resulting cell population was tolerant to 1 m CaCl2. The cells were suspended in glucose-free Joklik-modified Minimal Essential Medium (MEM/Joklik) supplemented with 1.5% bovine serum albumin (BSA), 20 m creatine, 60 m taurine, and 1 m CaCl2 in experiments measuring cyclic nucleotides. The medium was changed to an identical medium lacking inositol when inositol phosphates were measured.

Anoxia protocol All incubations were conducted at 37°C in siliconized glass flasks, stoppered and gassed continuously with O2 or N2 as appropriate (Geisbuhler and Rovetto, 1990). Suspending buffer was equilibrated at least 1 h with 100% O2 or N2; the experiment was initiated by dividing the preparation into two parts, settling the cells, removing the supernatant, and replacing it with the appropriately gassed buffer. The cells were placed into siliconized glass flasks, stoppered securely, and incubated under a stream of either O2 (aerobic controls) or N2 (anoxia). After incubating for either 30 or 60 min, aliquots were removed from both aerobic and anoxic flasks for nucleotide or phosphoinositide (InsPx) analysis.

Production of cyclic nucleotides In experiments concerning cyclic GMP (cGMP) production, myocytes were preincubated with 150 l Zaprinast (M&B 22948, a cGMP-specific phosphodiesterase inhibitor), then exposed to anoxia as described above. After 0, 30 or 60 min of anoxia, aliquots were removed for 5′-nucleotide and prechallenge cGMP analysis; the remainder of the cells were challenged for 1 min with 1 m sodium nitroprusside (NaNP). Aliquots for post-challenge cGMP analysis were then removed. All aliquots of

1859

Second Messengers in Anoxic Cardiac Myocytes Extracellular Signal RECEPTOR ENZYME block

PI

R2

R1

ATP

R1

R2

P I

P I

GTP Depletion

PIP2

PIP ATP

R2

R1 P I

P

P P cGMP

CDP-DG

R2

R1

CTP

P

R2

–

DG

PA R1

ATP

P

R2

R1 IP3

OH

IP2

IP

Ca2+ increase

PKC

RESPONSE

Figure 1 Model for the turnover of inositol phospholipids in intracellular signaling. Active molecular species are diacylglycerol (DG) and inositol 1,4,5-trisphosphate (IP3); IP2=inositol diphosphate, IP=inositol monophosphate, PA= phosphatidic acid, CDP-DG=cytidine diphosphate diglyceride, PI=phosphatidylinositol, PIP=PI monophosphate, PIP2= PI diphosphate, PKC=protein kinase C.

cell material were extracted and assayed for cGMP as described below.

either for 20 min with 40 l norepinephrine (NE), 10 min with 30 l phenylephrine (PE), or 20 min with 60 l PE. Aliquots were removed for analysis of post-challenge InsPx and extracted as described below.

Turnover of cellular phosphoinositides In experiments concerning InsPx turnover, myocytes representing 200–300 mg of cell protein were suspended in 12 ml MEM/Joklik medium containing 50 l CaCl2, 1.5% BSA, and 500 lCi 2-[3H]-myoinositol, and incubated 1 h at 37°C. (Because normal MEM/Joklik medium contains inositol as an additive, inositol-free medium was prepared for this incubation.) Calcium was then added to 1 m over 30 min. The cells were washed two to three times with medium containing 1 m CaCl2, no radiolabel, and 10 m LiCl (a general phosphatase inhibitor which prevents degradation of the phosphoinositides). The cells were divided and subjected to the anoxia protocol as described above. At 0, 30 and 60 min, aliquots were removed for pre-challenge InsPx analysis and cell counting. The remainder of the suspended cells were challenged

Extraction Aliquots were quickly transferred to cold 1.5-ml microcentrifuge tubes containing (from the bottom) 0.1 ml 2N perchloric acid, 0.4 ml 1-bromododecane, and 0.4 ml of buffer (this top layer acts as a “pad” which prevents disturbance of the lower layers as cells are injected). Cell suspensions were injected into the solution in the top layer and immediately centrifuged at 13 500 ×g. The aqueous layer above the oil was removed, the tube above the oil was washed, and the extract was removed and neutralized as previously described (Geisbuhler et al., 1987). The neutralized extract was then analyzed for 5′-nucleotides, cyclic nucleotides, or phosphoinositides.

1860

T. P. Geisbuhler and T. L. Schwager

Cell counting

Cell protein

Myocytes were stained and counted using a variation on the procedure described by Altschuld et al. (1985). “Trypan Blue” fixative was prepared by dissolving 30 mg Direct Blue 15 [an analog of Trypan Blue (Direct Blue 14)] in 10 ml of phosphatebuffered saline containing 8% glutaraldehyde. [This latter solution is marketed by Polysciences (Warrington, PA, USA) as electron microscopic-grade glutaraldehyde, sealed in 10-ml vials under argon.] To fix and stain the myocytes, 200 ll of cell suspension was rapidly injected into 100 ll of dye/ fixative solution, using enlarged-orifice pipette tips to avoid mechanical damage to the cells. An aliquot of the stained cells was placed on a hemacytometer and observed by light microscopy. The cells were photographed, and the prints used for counting stained cells (dead) v unstained (live) cells, as well as various morphological forms [described in Brierley et al. (1986) and Geisbuhler and Rovetto (1990)].

Cell pellets were dissolved in NaOH and assayed for protein essentially as described by Lowry et al. (1951).

Cyclic nucleotides Cyclic GMP was determined by radioimmunoassay, using kits designed for this purpose from New England Nuclear (Boston, MA, USA).

5′-nucleotides Nucleotides were measured using anion-exchange HPLC as described by Geisbuhler et al. (1984).

Phosphoinositides InsPx fractions were separated by anion exchange chromatography as described by Berridge et al. (1983). Neutralized extracts were applied to 0.2 ml columns of Dowex-1 × 8 formate. The columns were washed with, in sequence: 5 ml deionized water (elutes free inositol), 5 ml 5 m Na2B4O7/ 60 m ammonium formate (glycerophosphoinositol), 10 ml 0.1 formic acid/0.2  ammonium formate (InsP1), 10 ml 0.1  formic acid/0.4  ammonium formate (InsP2), 5 ml 0.1  formic acid/ 1.0  ammonium formate (InsP3), and 5 ml 0.1  formic acid/2.0  ammonium formate (InsP4–6). Tritium in each fraction was assayed by liquid scintillation spectrometry.

Statistics Mean values were compared for differences using Student’s t-test. All comparisons were made withintime (for example, 30-min anoxic values were compared with 30-min aerobic values) and withincondition (for example 30-min aerobic unstimulated values were compared with 30-min aerobic stimulated values). Comparison of values using one-way analysis of variance followed by a Newman–Keuls post-hoc analysis did not alter the results.

Results The effect of our anoxia protocol on cell viability and morphology has been reported previously (Geisbuhler and Rovetto, 1990). The present study showed similar results (Fig. 2). Thirty min of anoxia was sufficient to convert many rod-shaped cells to “contracted” (box-like) forms. Overall viability by trypan blue exclusion was slightly depressed by anoxia. Because our anoxic cell injury model depended on the depletion of guanine nucleotides, total guanine nucleotides were routinely assayed in all cell preparations. Anoxia depleted total guanine nucleotides at least 50% in cardiac myocytes (Fig. 3). GTP was especially degraded; less than 10% of the GTP pool remained after 60 min of anoxia. Native levels of cyclic GMP (cGMP) were low in our myocyte preparation (21.8±2.3 fmol/mg cell protein). Anoxia did not significantly alter the resting cGMP level (16.8±6.9 fmol/mg protein and 18.9±4.2 fmol/mg protein after 30 min and 60 min of anoxia, respectively). Challenge of freshly isolated myocytes with 1 m sodium nitroprusside (NaNP) for 1 min elevated cGMP 15-fold (Fig. 4). After 30 min of anoxia, the same pharmacologic challenge to the myocytes gave an 83-fold increase in cGMP; this level was 11 times the post-challenge cGMP level in the paired aerobic controls. Myocytes which were anoxic for 60 min showed the same pattern of guanylate cyclase stimulation, although the results were quantitatively less impressive than those at 30 min. There was no clear correlation of cGMP produced with the amount of GTP in the

1861

Second Messengers in Anoxic Cardiac Myocytes Viability

% Rods

% Contracted

100 *

% of total cells

80

** ***

60 *

40 *

20

*** 0

60

30 Time (min)

0

0

30 Time (min)

60

0

30 Time (min)

60

Figure 2 Myocyte morphology and viability under anoxic conditions. Values are expressed as a percentage of total cells counted, and represent the mean ± ... of four preparations (each mean is the result of triplicate determinations). Χ, aerobic; Β, anoxic. Asterisks indicate a significant change from aerobic controls at the same time point [PΖ0.05 (∗), 0.01 (∗∗), or 0.001 (∗∗∗)]. 2000

1.0 0.8 0.6

*

* *

0.4 *

0.2 0

* * 0

30

60

30

*** 1500

1000

60

Time (min)

Figure 3 GTP depletion in cardiac myocytes under anoxic conditions. GTP=guanosine triphosphate (Φ), TGN=total guanine nucleotides (Ε). All numbers are expressed as nmol nucleotide/mg cell protein, and represent the mean ± ... of six preparations (each mean is the result of triplicate determinations). Asterisks indicate a significant change from aerobic controls at the same time point [PΖ0.05 (∗) or 0.01 (∗∗)].

cells (that is, less GTP in the cells did not necessarily mean less cGMP produced; data not shown). Myocyte PIP2 was labeled by incubating the cells for 1 h with 10 lCi/ml[3H]-myo-inositol. The labeling buffer was washed from the cells and replaced with glucose-free buffer containing 10 m LiCl. (The latter buffer was pre-equilibrated with either O2 or N2, and the experiment initiated as described in Materials and Methods.) At t=0, 30 and 60 min, aliquots were removed for cell counting and pre-challenge phosphoinositide (InsPx) analysis; agonist [norepinephrine (NE) or phenylephrine (PE)] was then added and the incubation

*

500

0

Anoxic

Aerobic

fmol cGMP/mg cell protein

nmol/mg cell protein

1.2

0

30 Time (min)

60

Figure 4 Anoxia does not impair cyclic GMP production in cardiac myocytes. All numbers are expressed as fmol nucleotide/mg cell protein and represent the mean ± ... of six preparations (each mean is the result of triplicate determinations). Asterisks indicate a significant change from aerobic controls at the same time point [PΖ0.05 (∗), 0.01 (∗∗), or 0.001 (∗∗∗)]. (Ε, aerobic; Φ, aerobic + NaNP; ∆, anoxic; ;, anoxic + NaNP).

continued for either 10 or 20 min. Cells were extracted, and the phosphoinositides separated on Dowex-1 formate columns and assayed for radioactivity. The data from these incubations are presented in Figures 5, 6 and 7. Figure 5 shows data from cells which were incubated with norepinephrine. In cells which were unstimulated with norepinephrine, 57–75% of the label was recovered as inositol. Glycerophosphoinositol (GPI) accounted for 5–7% of the label; this did not change through the various conditions. LiCl is an imperfect inhibitor of the InsP3 and InsP2 phosphatases; products of PIP2 hydrolysis accumulate as InsP1, whose phosphatase is more strongly inhibited by LiCl (Berridge et al., 1982). For this

T. P. Geisbuhler and T. L. Schwager * 30

* *

*

20

10

0

30

0

Total dpm/mg in InsPx (Thousands)

Total dpm/mg in InsPx (Thousands)

1862

60

10

5

0

0

30

60

60

* *

* *

* *

40 20 0

0

30 Time (min)

60

* * % dpm/mg in InsPx

% dpm/mg in InsPx

80

60

* *

* *

40

*

* * *

20

0

0

30 Time (min)

60

Figure 5 Anoxia impairs InsPx production during norepinephrine challenge in cardiac myocytes. NE=norepinephrine. Top panel: InsPx expressed as total dpm[3H]/ mg cell protein in combined nsPx fractions. Bottom panel: dpm 3H in InsPx expressed as % of total recovered 3H (all fractions). Values are the mean ± ... of four preparations (each mean is the result of triplicate determinations). Asterisks indicate a significant change from unstimulated controls at the same time point [PΖ0.05 (∗) or 0.01 (∗∗)]. Closed triangles indicate a significant change from aerobic controls at the same time point [PΖ0.05 (&) or 0.01 (&&)]. (Ε, aerobic; Φ, aerobic + NE; ∆, anoxic; ;, anoxic + NE).

Figure 6 Anoxia impairs InsPx production during 30 l phenylephrine challenge in cardiac myocytes. PE= phenylephrine. Top panel: InsPx expressed as total dpm 3 H/mg cell protein in combined InsPx fractions. Bottom panel: DPM 3H in Inspx expressed as % of total recovered 3 H (all fractions). Values are the mean ± ... of five preparations (each mean is the result of triplicate determinations). Asterisks indicate a significant change from unstimulated controls at the same time point [PΖ0.05 (∗) or 0.01 (∗∗)]. Closed triangles indicate a significant change from aerobic controls at the same time point [PΖ0.05 (&) or 0.01 (&&)] (Ε, aerobic; Φ, aerobic + 30 lPE; ∆, anoxic; ;, anoxic + 30 lPE).

reason, almost all of the remaining label was associated with inositol 4-phosphate (InsP1). [For example, the labeling pattern of the aerobic controls at 0 min was (in dpm/mg cell protein): inositol = 16958±1595, GPI=1645±368, InsP1= 3535±291, InsP2=350±51, and InsP3= 233±45.] Therefore InsP3, InsP2, and InsP1 were added together, and the products designated InsPx. Data generated from analysis of dpm/mg cell protein are shown in the top panel of Figure 5. Standard error was significantly improved if the dpm/mg protein in each fraction was expressed as a percentage of the actual radioactivity recovered. The amount of radioactivity incorporated into the PIP2 pool varied with the cell preparation; the percentage of the total counts associated with each fraction was more stable. [Again using the aerobic control at 0 min as an example, the radiolabeling pattern was (in % of total recovered dpm/mg cell protein): inositol = 74.8±0.9, GPI=6.0±0.4, InsP1=16.7±1.0, InsP2=1.4±0.3, and InsP3=

0.9±0.06.] Data expressed in this way are shown in the bottom panel of Figure 5. InsPx accumulated slowly over 60 min of aerobic incubation (Fig. 5, Ε); we attributed this accumulation to a slow turnover of the phosphoinositol cycle in these cells. Thirty min of anoxia slowed InsPx turnover; after 60 min of anoxia, turnover through the pathway had ceased (Fig. 5, ∆). Stimulation of the myocytes with NE increased accumulation of InsPx 3.5-fold in aerobic cells (Fig. 5, Φ). Anoxia impaired this stimulation (Fig. 5, ;); NE stimulated InsPx production only slightly after 30 min of anoxia, and not at all after 60 min of anoxia. Similar results were obtained using phenylephrine (PE) as a stimulant. Two different phenylephrine regimens were used: 10 min at 30 l PE (Fig. 6) and 20 min at 60 l PE (Fig. 7). As in the previous experiments (NE stimulation), InsPx accumulated slowly in unstimulated cells (Figs 6

Total dpm/mg in InsPx (Thousands)

Second Messengers in Anoxic Cardiac Myocytes

* *

* *

* * *

5

0

0

30

60

% dpm/mg in InsPx

60 * * 40

* *

* *

* * *

20

0

0

30 Time (min)

60

Figure 7 Anoxia impairs Inspx production during 60 l phenylephrine challenge in cardiac myocytes. PE= phenylephrine. Top panel: InsPx expressed as total DPM 3 H/mg cell protein in combined InsPx fractions. Bottom panel: dpm 3H in InsPx expressed as % of total recovered 3 H (all fractions). Values are the mean ± ... of five preparations (each mean is the result of triplicate determinations). Asterisks indicate a significant change from unstimulated controls at the same time point [PΖ0.05 (∗) or 0.01 (∗∗)]. Closed triangles indicate a significant change from aerobic controls at the same time point [PΖ0.05 (&) or 0.01 (&&)]. [Ε (solid bars), aerobic; Φ (open bars), aerobic+60 l phenylephrine (PE); ∆ (hatched bars), anoxic; ; (shaded bars), anoxic+60 l PE].

and 7, Ε); anoxia slowed or stopped this accumulation (Figs. 6 and 7, ∆). PE stimulated InsPx production 1.8-fold (10 min at 30 l; Fig. 6, Φ) and 3.1-fold (20 min at 60 l; Fig. 7, Φ) in aerobic cells; this stimulation was lessened after 30 min of anoxia, and abolished after 60 min of anoxia (Figs 6 and 7, ;).

Discussion The original model (as it related to second messengers) centered on three hypotheses: (1) depletion of GTP in anoxic myocytes degrades the cell’s ability to produce cyclic GMP (cGMP); (2) depressed levels of cGMP [which has been postulated as an inhibitor

1863

of phospholipase C (Rapoport, 1986; Hirata et al., 1990)] allow the degradation of phosphatidylinositol bisphosphate (PIP2) to diacylclycerol (DG) and inositol 1,4,5-trisphosphate (InsP3), thereby allowing the release of calcium to the cytosol; and (3) depletion of GTP interferes with the ability of the myocytes to accept incoming adrenergic signals by decreasing activity of the Gproteins (see Fig. 1). The data do not support hypotheses 1 and 2 of this model. Anoxia did not affect basal cGMP levels, and did not impair the ability of the myocytes to generate cGMP if challenged with NaNO (Fig. 4). Sixty min of anoxia decreased basal levels of cGMP from 21.82 fmol/mg (not significant, matched by aerobic controls). When stimulated with 1 m NaNP for 1 min, anoxic cells produced up to tenfold more cGMP than aerobic controls. Cytosolic guanylate cyclase is stimulated by Ca2+; the tendency of anoxic myocytes to accumulate cytosolic Ca2+ under anoxic conditions is well-documented. Therefore, elevation of cytosolic Ca2+ is probably responsible for the observed increase in cGMP production during anoxia. Why this stimulatory effect declines at 60 min is less clear. If large numbers of myocytes had died from 30 min to 60 min of anoxia, they would have contributed to protein without contributing to cyclic nucleotide levels; this would have artificially lowered the 60 min cGMP value. However, there was no such concurrent increase in non-viable myocytes, so that this artifact can be ruled out. Substrate depletion, intracellular acidosis, resequestration or leakage of cytosolic calcium, or buildup of an unidentified competitive or allosteric inhibitor (perhaps GMP) are all possible explanations. Such explanations cannot cause us to accept the hypothesis that GTP depletion interferes with cyclic GMP production, however, for there is clearly no impairment in the ability of the myocytes to make cGMP under anoxic conditions. It is interesting to contrast the present study using isolated myocytes with the whole-heart study done by Depre´ and Hue (1994). These investigators reported basal cGMP levels about ten-fold higher than those reported in this study (158 pmol/mg protein v 16–21 pmol/mg cell protein in the myocytes). They also reported an increase in cGMP to 267 pmol/mg protein within 10 min of the onset of anoxia (myocyte cGMP was unchanged through 60 min of anoxia). These differences are probably due to paracrine interaction of myocytes with atrial and endothelial cells in the intact heart model. Endothelial cells are rich in the constitutive form of nitric oxide synthetase (cNOS); these cells also can induce a second form of the enzyme (inducible

1864

T. P. Geisbuhler and T. L. Schwager

nitric oxide synthetase, iNOS) when stimulated by cytokines (Nathan and Xie, 1994). Nitric oxide generated by these enzymes stimulates guanylate cyclase activity. Atrial cells produce atrial natriuretic factor (ANF), which can elevate cGMP in myocytes through an R1 receptor (Anand-Srivastava and Trachte, 1993). Either endothelial cellgenerated nitric oxide or ANF from atrial cells could act on myocytes to generate cGMP in the intact heart model; no such opportunity exists in the isolated myocytes. Even though myocytes express both cNOS and iNOS (Schulz and Triggle, 1994), no stimulus is provided for these enzymes in our model. Anoxia in the absence of another cell type probably does not generate such a stimulus in the isolated myocytes. Comparison of studies using intact heart with similar studies using isolated cell models gives an excellent perspective on the types of data that each model provides. In this case, the present study with isolated myocytes gives unequivocal data on how the muscle cells of the myocardium respond to anoxia. Comparison of our results with a similar study using intact heart revealed a possible paracrine interaction of the myocytes with surrounding cells. Both types of information are important in determining the cellular pathology of ischemic heart disease. Such comparisons also highlight the caution that must be used in extrapolating from one model to the other. The presence of several isoforms of cyclic nucleotide phosphodiesterase in heart ventricle is well documented (Merkel, 1993; Endoh, 1995). Preliminary results (not shown) indicated that phosphodiesterase activity was degrading a significant portion of our cyclic nucleotides before they could be analysed, even in NaNP-stimulated cells. Inclusion of phosphodiesterase inhibitors in the suspending medium was necessary to ensure that this did not occur. Zaprinast (May and Baker 22948), a specific inhibitor of the cGMP-specific phosphodiesterase, did not affect cell viability or morphology (unlike some alternative compounds). Zaprinast at 150 l was sufficient to inhibit apparent phosphodiesterase activity; more inhibitor did not increase measured levels of cyclic nucleotide. Anoxia decreased turnover of phosphoinositide cycle components, as quantified by a buildup of the products InsP1–3. When the myocytes were stimulated with the a-adrenergic agonist norepinephrine, aerobic myocytes showed a two-fold increase in phosphoinositide turnover; this activity was abolished under anoxic conditions. The result is an agreement with results of similar studies in intact heart (Mouton et al., 1991a, b), and argues

in favor of the third hypothesis of the model (GTP depletion blocks signaling through receptors which are coupled to G-proteins). Impairment of the myocyte’s ability to accept incoming a-adrenergic signals could be caused by several deficiencies: (1) decrease in the expression of a-receptors; (2) decreased expression or activity of G-proteins independent of GTP depletion; or (3) decreased activity of G-proteins because of cytosolic GTP depletion. Allely et al. (1993) reported a twofold increase in the density of a1-receptors in ischemic rat ventricle over non-ischemic controls. This result agrees both with earlier studies (Corr et al., 1981) and with the observation that b- and receptors increase in number in ischemic heart (Strasser and Marquetant, 1990). Therefore, decreased expression of a-receptors is probably not an issue in this model. Expression and activity of G-proteins is more debatable. Early studies showed that Gsa decreased to about 30% of control after 15 min of ischemia in rat heart (Maisel et al., 1990). More recent work showed that expression of the G-proteins Gsa and Gia were unchanged in ischemic ventricles as compared to aerobic controls (Van den Ende et al., 1994). One might infer the presence of active Gproteins from the nearly universal observation that b-adrenergic activity is elevated under ischemic conditions. However, this increase in activity is due to increased numbers of b-receptors and increased activity of adenylate cyclase, and limited to the first 20 min of ischemia (Scho¨mig and Richardt, 1990). After that time, b-adrenergic activity decreases, presumably due to the lack of viable Gsa (Susanni et al., 1989; Opie, 1991). Thus, it is possible that the depression in a-adrenergic activity we observed under anoxic conditions is due to underexpression of Gqa or G11a (the G-proteins which couple the a1receptor to phospholipase C). No studies to date have examined expression of these G-proteins in heart during anoxia or ischemia. It is equally possible that substrate depletion (that is, depletion of GTP) interferes with G-protein activity, thus decreasing observed a-adrenergic activity. In the present study, GTP was degraded to 10% of control values after 60 min of anoxia (Fig. 3). Forty per cent of GTP is cytosolic in cardiac myocytes (0.2 nmol/mg protein); further, ATP is lost first from the cytosol after the onset of anoxia (Geisbuhler et al., 1984). If guanine nucleotides follow a similar pattern of post-anoxic degradation, cytosolic GTP would decline markedly after a relatively brief period of anoxia. Our estimate of 52 pmol GTP/mg protein after 60 min of anoxia represents 16–26 l GTP cell-wide [estimates for

Second Messengers in Anoxic Cardiac Myocytes

cell water range from 1.9 ll/mg cell protein (Geisbuhler et al., 1984) to 5.1 ll/mg cell protein (Kao et al., 1980)]. If cytosolic GTP is preferentially degraded under anoxic conditions (as the pattern for ATP depletion suggests), GTP levels at the cell membrane could fall to submicromolar levels during the course of a study such as this one. Lack of GTP at the cell membrane could then interfere with Gprotein function. The quantity of GTP necessary to drive heterotrimeric G-protein-mediated responses remains problematic. In vitro assays for G-proteins use a range of GTP concentrations from 10 l to 500 l in the assay system. In˜iguez-Lluhi et al. (1993) have modeled an apparent Km of 100 n GTP for Gsa, but suggest that factors both intrinsic and external to the G-protein may alter this value substantially upward. These factors might include the type of a-subunit being examined, its state of binding to other proteins, and the presence or absence of allosterically active molecules. Further work will be necessary to determine the anoxic depletion pattern of the guanine nucleotides in cardiac cells, and the influence of GTP depletion on G-protein function. In summary, this study showed that GTP and total guanine nucleotides are depleted in anoxic cardiac myocytes. This depletion has no effect on the ability of the myocytes to generate cGMP. Further, no correlation exists between cGMP levels and native levels of GTP, suggesting that substrate availability is not a major regulatory consideration for guanylate cyclase. Finally, stimulation of InsP1–3 production by norepinephrine and phenylephrine was inhibited by anoxia; GTP depletion may play a role in this latter phenomenon by depriving Gproteins of substrate, thus interfering with signal transduction.

Acknowledgements This work was supported by grants from the National Institutes of Health/Heart, Lung, and Blood Institute (HL-39025), the American Heart Association-Missouri Affiliate, and the KCOM WarnerFermaturo Research Fund.

References A MC, B CM, K BA, K AT, M A, S M, 1993. Modulation of a1adrenoceptors in rat left ventricle by ischemia and acyl carnitines: protection by ranolazine. J Cardiovasc Pharmacol 21: 869–873. A A, P HM, C KSR, C PH,

1865

1987. Cytosolic free Ca2+ in single rat heart cells during anoxia and reoxygenation. Biochem J 244: 381–385. A RA, W WC, L KG, K OR, C CC, M V, VH RS, B GP, 1985. Structural and functional properties of adult rat heart myocytes lysed with digitonin. J Biol Chem 260: 14325–14334. A-S MB, T GJ, 1993. Atrial natriuretic factor receptors and signal transduction mechanisms. Pharmacol Rev 45: 455–497. B MJ, D CP, H MR, 1982. Lithium amplifies agonist-dependent phosphatidylinositol responses in brain and salivary glands. Biochem J 206: 587–595. B MJ, D RMC, D CP, H JP, I RF, 1983. Changes in the levels of inositol phosphates after agonist-dependent hydrolysis of membrane phosphoinositides. Biochem J 212: 473–482. B KC, A AP, C PH, 1992. Bioluminescent measurement in single cardiomyocytes of sudden cytosolic ATP depletion coincident with rigor. J Mol Cell Cardiol 24: 213–218. B GP, W WC, A RA, 1986. Heart myocytes as models of the cellular response to ischemia. Adv Exp Med Biol 194: 303–314. C PB, S JA, K JB, K RJ, 1981. Increased a-adrenergic receptors in ischemic cat myocardium. J Clin Invest 67: 1232–1236. D´ C, H L, 1994. Cyclic GMP in the perfused rat heart. Effect of ischemia, anoxia and nitric oxide synthase inhibitor. FEBS Lett 345: 241–245. E J, G-E E, R H, 1991. Alphaadrenoceptor-mediated increase in cytosolic free calcium in isolated cardiac myocytes. J Mol Cell Cardiol 23: 617–625. E M, 1991. Myocardial a-adrenoceptors: multiplicity of subcellular coupling processes. Asia Pacific J Pharmacol 6: 171–186. E M, 1995. The effects of various drugs on the myocardial inotropic response. Gen Pharmacol 26: 1– 31. F D, B AP, G WR, 1993. a1-Adrenoceptors in myocardium: functional aspects and transmembrane signaling mechanisms. Physiol Rev 73: 469–487. G TP, R MJ, 1990. Lactate does not enhance anoxia/reoxygenation damage in adult rat cardiac myocytes. J Mol Cell Cardiol 22: 1325–1335. G T, A RA, T RW, A AZ, L K, B GP, 1984. Adenine nucleotide metabolism and compartmentalization in isolated adult rat heart cells. Circ Res 54: 536–546. G TP, J DA, R MJ, 1987. Cardiac myocyte guanosine transport and metabolism. Am J Physiol 253: C645–C651. G TP, S TL, R MJ, 1992. Guanosine metabolism in adult rat cardiac myocytes: inhibition by acyclovir and analysis of a metabolic pathway. J Mol Cell Cardiol 24: 683–690. H M, K KP, C C-H, I T, M F, 1990. Mechanism of cyclic GMP inhibition of inositol phosphate formation in rat aorta segments and cultured bovine aortic smooth muscle cells. J Biol Chem 265: 1268–1273.

1866

T. P. Geisbuhler and T. L. Schwager

I˜ -L J, K C, G AG, 1993. The importance of G-protein bc-subunits. Trends Cell Biol 3: 230– 236. K RL, C EW, L SL, K JM, T GFO, W EH, 1980. The effects of insulin and anoxia on the metabolism of isolated mature rat cardiac myocytes. Arch Biochem Biophys 203: 587–599. K JC, B RJ, L TJ, M IP, P JR, F MA, 1990. Calcium release from cardiac sarcoplasmic reticulum induced by photorelease of calcium or Ins(1,4,5)P3. Am J Physiol 258: H610– H615. L JMJ, D DHW, B K, M JTA, V H HAA, 1992. Occurrence and functions of the phosphatidylinositol cycle in the myocardium. Mol Cell Biochem 116: 59–67. L OH, R NJ, F AL, R RJ, 1951. Protein measurement with the Folin phenol reagent. J Biol Chem 193: 265–275. M AS, R¨  LA, I PA, 1990. b-Adrenergic receptors and the G8 protein in myocardial ischemia and injury. Basic Res Cardiol 85 [Suppl. 1]: 47–56. M L, 1993. Zaprinast: a cGMP-selective phosphodiesterase inhibitor. Cardiovasc Drug Rev 4: 501–515. M R, H B, L A, 1991a. Increased myocardial inositol trisphosphate levels during a1-adrenergic stimulation and reperfusion of ischaemic rat heart. J Mol Cell Cardiol 23: 841–850. M R, H B, L A, 1991b. The effect of ischaemia and reperfusion on sarcolemmal inositol phospholipid and cytosolic inositol phosphate metabolism in the isolated perfused rat heart. Mol Cell Biochem 105: 127–135. M MA, T AP, S M, W JR, 1985. Inositol trisphosphate does not release Ca2+ from permeabilized cardiac myocytes and sarcoplasmic reticulum. FEBS Lett 185: 328–332. N C, X Q-W, 1994. Regulation of biosynthesis of nitric oxide. J Biol Chem 269: 13725–13728.

N TM, W MF, Z ST, G RE, 1986. Inositol trisphosphate enhances calcium release in skinned cardiac and skeletal muscle. Am J Physiol 250: C807–C811. O L, 1991. The Heart: Physiology and Metabolism. Chapter 17, p. 435. Raven Press, New York. R RM, 1986. Cyclic guanosine monophosphate inhibition of contraction may be mediated through inhibition of phosphatidylinositol hydrolysis in rat aorta. Circ Res 58: 407–410. R DS, A RA, G CE, 1990. Effects of calcium depletion and loading on injury during metabolic inhibition of isolated adult rat myocytes. J Mol Cell Cardiol 22: 227–237. S¨  A, R G, 1990. Cardiac sympathetic activity in myocardial ischemia: release and effects of noradrenaline. Basic Res Cardiol 85 [Suppl. 1]: 9–30. S R, T CR, 1994. Role of NO in vascular smooth muscle and cardiac muscle function. Trends Pharmacol Sci 15: 255–259. S RH, M R, 1990. Supersensitivity of the adenylyl cyclase system in acute myocardial ischemia: evaluation of three independent mechanisms. Basic Res Cardiol 85 [Suppl. 1]: 67–78. S EE, M T, K DR, V DE, V SF, H CJ, 1989. One hour of myocardial ischemia decreases the activity of the stimulatory guaninenucleotide regulatory protein G8. Circ Res 65: 1145– 1150. T A, P´  M, V G, V SM, 1993. Cardiac a1-receptors: an overview. Pharmacol Rev 45: 147–175. V  E R, B HD, M MC, V Z PA, 1994. Influence of ischemia and reperfusion on cardiac signal transduction. G protein content, adenylyl cyclase activity, cyclic AMP content, and forskolin and dibutyryl cyclic AMP-induced inotropy in the rat Langendorff heart. Fundam Clin Pharmacol 8: 408–416.