Nonanticoagulant heparin reduces myocyte Na⁺ and Ca²⁺ loading during simulated ischemia and decreases reperfusion injury

Abstract

Heparin desulfated at the 2-O and 3-O positions (ODSH) decreases canine myocardial reperfusion injury. We hypothesized that this occurs from effects on ion channels rather than solely from anti-inflammatory activities, as previously proposed. We studied closed-chest pigs with balloon left anterior descending coronary artery occlusion (75-min) and reperfusion (3-h). ODSH effects on [Na⁺]_i (Na Green) and [Ca²⁺]_i (Fluo-3) were measured by flow cytometry in rabbit ventricular myocytes after 45-min of simulated ischemia [metabolic inhibition with 2 mM cyanide, 0 glucose, 37°C, pacing at 0.5 Hz; i.e., pacing-metabolic inhibition (PMI)]. Na⁺/Ca²⁺ exchange (NCX) activity and Na⁺ channel function were assessed by voltage clamping. ODSH (15 mg/kg) 5 min before reperfusion significantly decreased myocardial necrosis, but neutrophil influx into reperfused myocardium was not consistently reduced. ODSH (100 μg/ml) reduced [Na⁺]_i and [Ca²⁺]_i during PMI. The NCX inhibitor KB-R7943 (10 μM) or the late Na⁺ current (I_Na-L) inhibitor ranolazine (10 μM) reduced [Ca²⁺]_i during PMI and prevented effects of ODSH on Ca²⁺ loading. ODSH also reduced the increase in Na⁺ loading in paced myocytes caused by 10 nM sea anemone toxin II, a selective activator of I_Na-L. ODSH directly stimulated NCX and reduced I_Na-L. These results suggest that in the intact heart ODSH reduces Na⁺ influx during early reperfusion, when I_Na-L is activated by a burst of reactive oxygen production. This reduces Na⁺ overload and thus Ca²⁺ influx via NCX. Stimulation of Ca²⁺ extrusion via NCX later after reperfusion may also reduce myocyte Ca²⁺ loading and decrease infarct size.

early reperfusion of ischemic myocardium by means of thrombolysis or primary angioplasty can reduce the size of the resulting myocardial infarction and thus reduce mortality and morbidity. However, the reperfusion process itself can cause myocardial injury, which can account for up to 50% of the ultimate infarct size (44). Experimental evidence from many sources has suggested that myocyte Ca²⁺ overload induced by reverse Na⁺/Ca²⁺ exchange (NCX) is an important contributor to reperfusion injury. Normally the sarcolemmal NCX functions primarily in a forward mode to extrude Ca²⁺ from the myocyte (3 Na⁺ in, 1 Ca²⁺ out). Ischemic myocytes are depolarized and have an elevated sodium concentration ([Na⁺]_i) due in part to impaired function of the Na⁺-K⁺-ATPase pump and activation of Na⁺/H⁺ exchange by intracellular acidosis. Depolarization and an increase in [Na⁺]_i promote reverse NCX (1 Ca²⁺ in, 3 Na⁺ out), but NCX is inhibited by acid pH and severe ATP depletion. Reverse NCX is activated during reperfusion by abrupt washout of extracellular H⁺, which restores extracellular pH to normal, and by reintroduction of oxygen, which allows resynthesis of ATP. The resulting Ca²⁺ overload contributes to hypercontracture with sarcolemmal rupture as ATP resynthesis occurs (contraction band necrosis). In combination with the production of reactive oxygen species (ROS) induced by resupply of oxygen into ischemic myocardium, Ca²⁺ overload can also induce the mitochondrial permeability transition (MPT) (8), which is a cause of irreversible myocyte injury (42). Furthermore, ROS may alter Na⁺ channel function, increasing Na⁺ influx, and thus contribute to increased [Na⁺]_i (46). Hours after reperfusion, neutrophils can accumulate in reperfused myocardium and cause injury via ROS generation, microvascular plugging, and release of proteolytic enzymes (44). As recently discussed by Hausenloy and Yellon (12), there are a number of promising approaches to prevent this cascade of events, including inhibition of the MPT by administration of cyclosporine immediately prior to reperfusion. However, it is apparent that multiple pathways are involved in mediating reperfusion injury. Therefore, maximally effective preventive therapy will have to be directed at many and not just one of these mechanisms.

The glycosaminoglycan heparin, when employed in higher-than-anticoagulant plasma concentrations, has been consistently demonstrated to reduce myocardial reperfusion injury (7, 10, 39). As an example, when given 5 min prior to reperfusion, heparin desulfated at the 2-O and 3-O positions (ODSH) decreases myocardial infarct size by 38% in open-chest dogs subjected to 90 min left anterior descending coronary artery (LAD) occlusion and 4 h of reperfusion (39). In dogs, this protective was attributed to the anti-inflammatory activity of heparins, since ODSH impairs neutrophil rolling through inhibition of P- and l-selectins (40) and also significantly reduces neutrophil influx into ischemic-reperfused myocardium (39). To determine whether these findings were reproducible in an animal model of ischemia-reperfusion injury more relevant to humans, we studied the effects of ODSH in closed-chest pigs. In addition, to examine other possible protective mechanisms, we studied effects of ODSH on [Na⁺]_i and [Ca²⁺]_i in suspensions of isolated adult rabbit ventricular myocytes during simulated ischemia and investigated the direct effects of ODSH on Na⁺ channel function and on NCX. To our surprise, we found that ODSH has a more proximal protective effect against reperfusion injury by preventing reverse-mode operation of NCX and Ca²⁺ accumulation during reperfusion through apparent inhibition of Na⁺ loading in ischemic myocardium.

METHODS

Closed-Chest Pig Ischemia-Reperfusion

Care of animals conformed with the Guide for the Care and Use of Laboratory Animals published by the U.S. National Institutes of Health (NIH Publication No. 85-23, revised 1996), and all relevant protocols were approved by the Institutional Animal Care and Use Committees of Emory University and the University of Utah.

Surgical preparation.

Yorkshire-cross pigs (Palmetto Research Swine, Reevesville, SC) of either sex, weighing 25–35 kg, were used for the experiment. Animals were premedicated with an intramuscular injection of ketamine (30 mg/kg), acepromazine (1.1 mg/kg), and atropine (0.05 mg/kg). Pigs were induced with an intravenous injection of thiopental (10 mg/kg) and maintained with continuous inhalation of isoflurane (1–1.5%). Aspirin (81 mg) was administered intravenously prior to the experiment.

Myocardial infarction.

Arterial access was achieved via bilateral femoral artery cut downs for the insertion of 8-Fr sheaths. Central venous and carotid access was achieved via a neck incision to expose the external jugular vein and common carotid artery. Animals were then anticoagulated with 50 U/kg of heparin to maintain an activated clotting time between 250–350 s prior to ischemia. A 7- to 8-French pigtail catheter was placed in the left ventricular cavity under fluoroscopic guidance to measure pressure and for injection of microspheres. Angiography was performed to define coronary anatomy and measure the diameter of the LAD at the point of intended balloon occlusion. A coronary sinus catheter was placed via the external jugular vein under fluoroscopic guidance for coronary venous sampling. Baseline cardiodynamic and hemodynamic data were measured using a solid-state transducer-tipped catheter in the left ventricle (Millar Instruments, Houston, TX) to measure left ventricular pressure, and a fluid-filled transducer was connected to the side port of the femoral artery sheath to measure peripheral arterial pressure. Approximately 3–4 million 15-μm neutron-activated microspheres (BioPhysics Assay Laboratory, Worcester, MA) (15 μm) were delivered through a pigtail catheter in the left ventricle over a 30-s period to quantify regional myocardial blood flow. Simultaneous with injection, a reference blood sample was withdrawn at a rate of 7 ml/min from the femoral artery sheath for 90 s during and after injection of microspheres. A contrast ventriculogram (60° right anterior oblique) was obtained to assess global and regional myocardial function at baseline. An angioplasty balloon catheter sized to exceed ambient diameter by 1 mm (range chosen, 3.0–4.0 mm) was then guided into the proximal LAD after the first diagonal branch. Prior to ischemia, amiodarone (8 mg/kg) was administered to reduce the incidence of ventricular arrhythmias. In addition, 2% lidocaine (4–8 ml total iv) was administered during ischemia as needed to attenuate ventricular arrhythmias. Ischemic balloon occlusion time was 75 min with the site of coronary occlusion confirmed by contrast angiography, and ischemia confirmed by ST segment changes on electrocardiography and injection of microspheres delivered at the end of ischemia as described above. Episodes of ventricular fibrillation were immediately treated with electrical cardioversion delivered at 200 Joules. During the ischemic period, animals were randomly assigned to receive either saline vehicle or ODSH at a dose of 5 mg/kg, 15 mg/kg, or 45 mg/kg as an intravenous bolus at 2 min prior to deflation of the balloon (pharmacological postconditioning) and repeated at 90 min of reperfusion. Following deflation of the angioplasty balloon, animals underwent 3 h of reperfusion. Microspheres were again injected to measure myocardial blood flow at 15 min of reperfusion and at 180 min of reperfusion. At the end of 180-min reperfusion, animals were euthanized with an intravenous injection of pentobarbital sodium (100 mg/kg), and the heart was excised to quantify the area at risk (AAR), infarct size, regional myocardial blood flow, and myeloperoxidase (MPO) activity. Hemodynamic data (left ventricular and arterial blood pressure) and derived variables were recorded continuously using IOX and Datanalyst software (EMKA Technologies, Falls Church, VA).

Determination of AAR and infarct size.

After harvesting the heart, the LAD was ligated with a 2-0 silk suture placed at the site of balloon inflation, and diluted Unisperse blue dye was injected into the aortic root to stain the nonischemic region blue and thereby outline the AAR. The left ventricle was then cut into 5–6 transverse slices and the AAR was separated from the nonischemic zone and incubated in a 1% buffered solution of triphenyltetrazolium chloride at 37°C to differentiate the area of necrosis from the nonnecrotic AAR. The AAR, as a percent of the left ventricular mass (AAR/LV) and the area of necrosis (NEC) as a percent of the AAR (NEC/AAR) were calculated by tissue weight as reported previously (39).

Tissue MPO activity.

After determining infarct size, tissue samples from the nonischemic and AAR zones were saved for analysis of tissue MPO activity, an enzyme used as a marker of neutrophil accumulation. The samples were frozen and stored at −70°C until assayed. The samples were homogenized in hexadecyltrimethyl ammonium bromide and dissolved in potassium phosphate. After centrifugation, supernatants were collected and mixed with O-dianisidine dihydrochloride and hydrogen peroxide in phosphate buffer. The activity of MPO was measured spectrophotometrically at 460 nm absorbance (SPECTRAmax; Molecular Devices, Sunnyvale, CA) and expressed as Δabs·min⁻¹·g⁻¹ tissue (39).

Regional myocardial blood flow.

Regional myocardial blood flow in the subepicardial and subendocardial regions of the AAR and nonischemic left ventricular free wall was determined by neutron-activated microspheres at baseline and ischemia and at 15 min and 3 h of reperfusion using the reference sampling method as previously described (45). Samples were dessicated according to instructions from BioPal Laboratories and sent for activation and analysis.

Results are expressed as milliliters per minute per gram tissue determined from the equation: Flow_T = [(R_T × Flow_Ref)/R_Ref]/Weight_T, where T is tissue and R is radioactivity, which equals the number of microspheres, and Ref is the reference sample.

Experiments with Isolated, Paced Ischemic Cardiac Myocytes

Myocyte isolation.

Adult rabbit ventricular myocyte isolation was performed as previously described (46). Briefly, isolated hearts were perfused by an aortic cannula with nominally Ca²⁺-free bicarbonate-buffered balanced salt solution for 10 min, followed by 15- to 25-min perfusion with 0.3 mg/ml collagenase (Type 2; Worthington Biochemical, Freehold, NJ), 0.4 mg/ml hyaluronidase (Type-S; Sigma, St. Louis, MO), and 50 μM CaCl₂. The left ventricle was then minced in 0.24 mg/ml collagenase, 0.02 mg/ml trypsin, and 50 μM Ca²⁺ with the solution agitated by bubbling with 5% CO₂-95% O₂. The cell suspension was then centrifuged in 0.03 mg/ml of trypsin inhibitor (Sigma), and the myocytes resuspended in 200, then 500, then 1,000 μM Ca²⁺ solutions. Myocytes were filtered through a 300-μm filter to remove cell clumps. The yield of rod-shaped viable myocytes averaged 50%.

Measurement of [Ca^2±]_i and [Na^±]_i.

Measurement of ventricular myocyte [Ca²⁺]_i and [Na⁺]_i by flow cytometry in our laboratory has been described (23, 46). Briefly, HEPES-buffered balanced salt solutions were used, and [Ca²⁺]_i was measured with Fluo-3 and [Na⁺]_i with Na Green (Molecular Probes, Eugene, OR). Fluorescence in myocytes was excited with an argon laser at wavelength 488 nm and detected at 530 nm using flow cytometry (FACScan; Becton-Dickinson, Boston, MA). Propidium iodide (excitation 488 nm, fluorescence 640 nm) was used to identify nonviable myocytes. Approximately 2 × 10³ myocytes were analyzed to calculate average emission fluorescence intensity in each sample. The sample [Ca²⁺]_i values were calculated by Mn²⁺ quenching (37, 43) and [Na⁺]_i values by comparison with a standard curve of varied [Na⁺]_i vs. Na Green fluorescence.

Simulation of demand ischemia in isolated ventricular myocytes.

The techniques used have been previously described in detail (46). Partial metabolic inhibition of ATP production was caused by exposure to 2 mM CN in 0 glucose solution. To increase demand for ATP, myocytes were paced at 0.5 Hz in a custom-made five-well chamber. Each well had a volume of 5 ml and was fitted with platinum sheet electrodes on two sides for field stimulation. Each well was water jacketed and maintained at 37°C, and each well was bubbled with air to keep myocytes in suspension. Pacing electrodes were connected in series to a constant-current pulse generator that delivered a two-ampere pulse for 10 ms. Equal volume aliquots of a myocyte suspension from a single heart dissociation were placed in each well. After an appropriate time, a sample from a well was taken for determination of [Ca²⁺]_i by flow cytometry as described above. The average value for each well determined by flow cytometry constituted an “n” of 1, and because in a given experiment all myocytes were from the same dissociation, paired comparisons were possible when assessing drug or experimental condition effects.

Measurement of Na⁺ Channel Ionic Currents

Fused tsA201 cells (SV40 transformed HEK293 cells) expressing the cDNA for the human heart voltage-gated Na⁺ channel, Na_v1.5 (hH1a) were trypsinized and studied electrophysiologically as described previously (31). For Na⁺ current (I_Na) measurements the extracellular solution was (in mM): 15 Na⁺, 185 tetramethylammonium (TMA⁺), 200 2-(N-morpholino)-ethanesulfonic acid, 10 HEPES, 3 CaOH₂, pH 7.2 with TMA⁺-hydroxide (TMA⁺-OH). The internal solution contained (in mM): 200 TMA⁺, 200 F⁻, 10 EGTA, and 10 HEPES (pH 7.2 by hydrofluoric acid). For saxitoxin (STX) (Calbiochem, San Diego, CA) subtraction experiments, 1 μM STX was added to the extracellular solution, and 1 mM Ca²⁺ was added to all external solutions. The hypertonicity compensated for the lower conductivity of TMA⁺ and 2-(N-morpholino)-ethanesulfonic acid solutions. The Na⁺-free heparinic acid of ODSH was generated by passage over an ion exchange column, followed by lyophilization, and a concentration of 1 mg/ml was added, after which the pH was adjusted with TMA⁺-OH.

I_Na recordings were made with a large-bore, double-barreled, glass suction pipette for both voltage clamp and internal perfusion as previously described (31). Currents were measured with a virtual ground amplifier (Burr-Brown OPA-101) by using a 2.5-MΩ feedback resistor, and voltage protocols were imposed from a 16-bit DA converter (National Instruments, Austin, TX) over a 30:1 voltage divider. Data were filtered by the inherent response of the voltage-clamp circuit (corner frequency near 125 kHz) and recorded with a 16-bit AD converter at 200 kHz. A fraction of the current was fed back to compensate for series resistance. Cells were studied at room temperature. Leak resistance was calculated as the reciprocal of the linear conductance between −180 mV and −110 mV, and cell capacitance was measured from the integral of the current responses to voltage steps between −150 mV and −190 mV. Peak I_Na was taken as the mean of four data samples clustered around the maximal value of data digitally filtered at 5 kHz and leak corrected by the amount of the calculated time-independent linear leak. Data were capacity corrected using 4 to 8 scaled current responses recorded from voltage steps typically between −150 mV and −190 mV.

For peak current-voltage (I-V) relationships, the holding membrane potential (V_hp) was either −150 or −110 mV, step depolarizations were for 50 ms, and the pulse frequency was 0.5 s. To account for any time-dependent shifts in I_Na kinetics from a V_hp of −150 mV, the control values represent the means of the peak I_Na before ODSH heparinic acid and wash. From a V_hp of −110 mV, to eliminate any leftward time-dependent shift in Na channel kinetics as a cause of a decrease in I_Na at −110 mV, peak I-V relationships were first recorded in ODSH heparinic acid before washing to control. Normalized peak I-V relationships were fit with a Boltzmann distribution: I_Na = (V_t − V_rev)G_max/[1 + exp (V_t − V_1/2/s)], where V_t is test voltage, V_rev is reversal potential, G_max is maximum conductance change, and V_1/2 is voltage at the half point of the relationship.

For steady-state voltage-dependent Na⁺ channel availability curves the V_hp was −150, the duration of the conditioning steps were 500 ms, with a test step to 0 mV for 25 ms using an interpulse interval of 2.5 s. To account for any leftward time-dependent shift in Na⁺ channel kinetics as a cause of a leftward shift in V_1/2, control values represent the means of peak I_Na before ODSH heparinic acid and afterwash. Normalized steady-state voltage-dependent Na⁺ channel availability curves were fit with a Boltzmann distribution: I_Na = (I_max − I_r)/[1 + exp (Vc − V_1/2/s)] + I_r, where I_r is residual current set to 0, Vc is command voltage, and s, is slope.

For STX subtraction experiments the holding membrane potential was −110 mV, the test step duration was 100 ms, and the pulse frequency was 1 Hz. The 3-ms interval (from 96 ms to 99 ms in the test step) was averaged from the raw current recordings at each test potential for cells exposed to control solutions and to ODSH heparinic acid. The leak for each cell was measured by adding 1 μM STX to both the control and ODSH heparinic solutions, repeating the same voltage-clamp protocol, and meaning the data from the 3-ms interval between 96 and 99 ms of the 100-ms step. These leak values in STX were subtracted from those in control solutions. To account for any time-dependent shift in kinetics, the late I_Na in control was taken as the mean of values measured before exposure to ODSH heparinic acid, and after its wash, except in one of four cells only the wash measurement was used to compare to that in heparinic acid.

Measurement of NCX

Suction pipettes were constructed from borosilicate capillary glass (Corning 8250 glass) and had resistances of ∼1.5 MΩ when filled. The filling solution contained (in mM): 50.0 aspartic acid, 20.0 NaCl, 20.0 HEPES, 20.0 BAPTA, 9.0 CaCl₂ (calculated free Ca²⁺ concentrations = 184 nM), 3.0 MgCl₂, 5.0 MgATP, 120.0 CsOH, pH 7.1 (15). Voltage-clamping was achieved with an Axopatch 200B clamp system using a CV203BU headstage. Membrane potential and membrane current were filtered at 50 kHz and digitized at 10–20 kHz with a 16-bit A/D converter (Digidata 1322A) and analyzed using pCLAMP 8 software (Axon Instruments). Myocytes were initially bathed in a normal solution containing (in mM): 126.0 NaCl, 4.4 KCl, 1.1 CaCl₂, 1 MgCl₂, 11 glucose, 24.0 HEPES, 12.9 NaOH (pH 7.4). Following break-in, cells were first bathed in the control NCX solution containing (in mM): 126.0 NaCl, 1.1 CaCl₂, 1 MgCl₂, 11 glucose, 24.0 HEPES, 12.9 NaOH (pH 7.4). The solution also contained (in μM): 100 niflumic acid, 10 nifedipine, 10 cyclopiazonic acid, and 200 tetracaine to block, respectively, Ca²⁺ -activated Cl⁻ current, Ca²⁺ current, sarcoplasmic reticulum Ca²⁺ pump, and voltage-sensitive release of Ca²⁺ from sarcoplasmic reticulum.

I-V relationships for NCX were measured by voltage-clamping myocytes with a ramp protocol as previously described (15) with some modification. Ramp pulses of 1,770-ms duration were given with 10-s intervals. Cells were held initially at −80 mV and then hyperpolarized to −100 mV for 0.5 ms followed by a linear ramp to +50 mV over 1.5 s, and then hyperpolarized to −100 mV over 220 ms and depolarized back to holding potential. An ascending ramp was chosen since this has been chosen to cause less perturbation of subsarcolemmal [Ca²⁺]_i than a descending ramp (9). The ramp was applied first in the control NCX solution and then after 5 min of superfusion in solution with 100 μg/ml ODSH. The protocol was repeated in the presence of 5 mM Ni²⁺ to obtain the background current (15), and this was subtracted to obtain the current attributable to NCX.

Statistical Analysis

Results are expressed as means ± SE. In experiments with multiple groups or treatments, a one-way ANOVA followed by Student-Newman-Keuls post hoc test was used to analyze for group differences in single-point data, such as infarct size, regional blood flow, myocardial MPO or effects of ODSH concentration in isolated myocytes. For longitudinal data, such as myocardial blood flow, two-way ANOVA for repeated measures was used. In experiments with two groups, the paired Student's t-test was used to assess difference. Significance was assigned at P < 0.05.

RESULTS

Effect of ODSH on Reperfusion Injury in the Pig

Forty-three pigs were initially entered into the study. A priori exclusion criteria were established to exclude cases in which the AAR/LV was < 20% or > 50%. Based on these exclusion criteria, three animals were excluded for AAR/LV < 20% and one for AAR/LV > 50%. In addition, one animal was excluded because the distal microcirculation failed to reperfuse following balloon deflation (confirmed by angiography and microspheres), and three were excluded because of technical complications (perivascular hematoma, cardiac tamponade, and intractable reperfusion arrhythmias). Six animals died during ischemia from intractable ventricular fibrillation. Data from 29 pigs are included in the final analysis: eight vehicle (control), six ODSH 5 mg/kg (ODSH 5), eight ODSH 15 mg/kg (ODSH 15), and seven ODSH 45 mg/kg (ODSH 45).

Regional myocardial blood flow was equivalent in the nonischemic myocardium at baseline in all groups studied. Myocardial blood flow in the nonischemic left ventricular myocardium remained unchanged during ischemia and reperfusion. LAD occlusion reduced subendocardial blood flow in the AAR by > 99% for all groups with no group differences. There were no significant group differences in the AAR regional blood flow in either the subepicardial or subendocardial regions at baseline, end ischemia, or at 15, 60, or 180 min of reperfusion (Table 1). Ejection fraction, determined by contrast angiography, was similar at baseline among all groups (Table 2). Moreover, ejection fraction was comparably reduced by ∼50% at the end of ischemia for all groups compared with their respective baseline and at 3 h remained ∼30% lower than baseline. There were no significant group differences in left ventricular systolic or end diastolic pressure, heart rate, or mean arterial pressure at any of the time points (Table 3).

Table 1.

Regional myocardial blood flow data in subepicardial and subendocardial regions of the area at risk

	Baseline	Ischemia	R15	R60	R180
Area at Risk, Subepicardial Blood Flow
Control	1.05±0.13	0.06±0.04	1.10±0.25	2.07±0.64	1.20±0.18
ODSH 5	1.53±0.34	0.07±0.06	1.82±0.67	2.20±0.43	1.05±0.14
ODSH 15	0.72±0.06	0.02±0.01	2.01±0.44	1.64±0.20	0.82±0.17
ODSH 45	1.05±0.16	0.08±0.03	1.95±0.32	2.31±0.39	1.11±0.16
Area at Risk, Subendocardial Blood Flow
Control	1.08±0.11	0.01±0.002	0.76±0.26	1.54±0.53	1.10±0.20
ODSH 5	1.57±0.35	0.02±0.01	1.07±0.37	1.92±0.40	0.90±0.12
ODSH 15	0.81±0.09	0.01±0.003	0.94±0.23	1.42±0.22	0.96±0.19
ODSH 45	1.05±0.17	0.01±0.001	1.48±0.37	2.21±0.50	1.1±0.15
Nonischemic Blood Flow
Control	1.37±0.22	0.94±0.14	0.70±0.06	0.89±0.18	0.92±0.08
ODSH 5	1.48±0.25	1.11±0.27	1.16±0.17	1.11±0.16	0.94±0.14
ODSH 15	1.06±0.25	0.69±0.08	0.79±0.13	0.80±0.12	0.83±0.11
ODSH 45	1.11±0.22	0.90±0.11	0.98±0.15	1.19±0.18	0.94±0.11

Data are means ± SE (ml·min⁻¹·g tissue⁻¹). R15, 60, and 180 = 15, 60, and 180 min of reperfusion. ODSH 5, 15, and 45, 2-O, 3-O desulfated heparin at 5, 15, and 45 mg/kg. Nonischemic myocardium data is shown for transmural samples.

Table 2.

Global ejection fraction determined by contrast ventriculography

	Baseline	Ischemia	R180
Control	52±2%	25±2%	34±3%
ODSH 5	61±2%	30±3%	42±5%
ODSH 15	57±4%	23±3%	42±4%
ODSH 45	60±2%	30±3%	38±6%

Values are means ± SE.

Table 3.

Cardio- and hemodynamic data in control and ODSH-treated pigs

	Baseline	Ischemia	R60	R120	R180
Left Ventricular Systolic Pressure, mmHg
Control	84±5	66±5	76±7	80±5	79±6
ODSH 5	74±6	59±6	72±5	72±10	70±9
ODSH 15	73±9	68±6	75±11	76±6	76±6
ODSH 45	79±13	72±7	73±9	78±7	77±5
Left Ventricular End Diastolic Pressure, mmHg
Control	8±4	16±6	18±5	13±5	11±7
ODSH 5	8±5	18±10	24±3	17±3	16±1
ODSH 15	7±5	17±6	14±7	10±7	10±6
ODSH 45	9±4	16±5	14±7	10±6	9±6
Heart Rate, beats/min
Control	113±16	88±24	97±21	109±18	111±22
ODSH 5	104±10	93±8	96±15	101±7	105±13
ODSH 15	123±28	96±18	99±12	106±15	109±16
ODSH 45	122±19	105±5	99±13	106±12	110±11
Mean Arterial Pressure, mmHg
Control	59±12	51±11	57±9	62±8	59±10
ODSH 5	30±19	42±9	50±11	61±5	61±5
ODSH 15	50±16	54±7	54±8	58±6	60±5
ODSH 45	55±11	55±9	51±9	57±6	57±6

Data expressed as means ± SE. R60, 120, and 180 = 60, 120, and 180 min of reperfusion, respectively.

The average data for AAR/LV was similar among all groups (Fig. 1A, left). No significant reduction in infarct size was observed in the 5 mg/kg ODSH group compared with control. However, there was a significant infarct size reduction (NEC/AAR) with both 15 mg/kg and 45 mg/kg ODSH relative to control, with no difference between groups receiving 15 or 45 mg/kg ODSH (Fig. 1A, right). Since collateral blood flow during ischemia was comparable in all groups, the significant reduction in infarct size in the two treatment groups was not due to greater values in collateral blood flow during coronary occlusion. There could be no effect of the ODSH heparin on collateral blood flow since the compound was not administered until 5 min before reperfusion. ODSH did not significantly increase ejection fraction compared with untreated controls at 180 min, although there was a trend toward improvement (Table 2).

Fig. 1.

Pharmacological postconditioning with 2-O, 3-O desulfated heparin (ODSH) reduces infarct size in ischemic-reperfused porcine myocardium. A: area at risk in the left ventricle (AAR/LV, left) and infarct size expressed as area of necrosis relative to the AAR (NEC/AAR). AAR was comparable in all groups. Infarct size was significantly reduced in the 15 mg/kg and 45 mg/kg ODSH groups. *P < 0.05 vs. control. B: myeloperoxidase activity (MPO) in ischemic-reperfused myocardium, expressed as Δabsorbance at 460 nm·min⁻¹·g tissue⁻¹ (A460·min⁻¹·g tissue⁻¹), was significantly reduced in the 45 mg/kg but not in 5 or 15 mg/kg ODSH groups. *P < 0.05 vs. other groups. ODSH 5, 5 mg/kg ODSH; ODSH 15, 15 mg/kg ODSH; ODSH 45, 45 mg/kg ODSH.

We have previously demonstrated that ODSH reduces canine myocardial reperfusion injury accompanied by a significant reduction in neutrophilic infiltration into ischemic-reperfused myocardium (39). We therefore expected to find similar results in ischemic-reperfused pigs. Surprisingly, MPO activity in ischemic-reperfused myocardium was significantly reduced compared with controls only in pigs treated with 45 mg/kg ODSH and not in pigs receiving 15 mg/kg drug (Fig. 1B). In nonischemic myocardium, MPO activity was comparable among all groups. These results confirm our previous findings that ODS reduced infarct size in open-chest dogs when given 5 min prior to reperfusion (39). However, reduction in neutrophil influx does not appear to be the only mechanism of protection, since significant infarct size reduction was observed with 15 mg/kg ODSH (Fig. 1A), despite no decrease in myocardial MPO (Fig. 1B).

Effect of ODSH on Ca²⁺ and Na⁺ Loading in Myocytes

There is considerable support for the idea that myocyte Ca²⁺ loading via reverse NCX, triggered by increased Na⁺ loading and myocyte depolarization, is an important cause of reperfusion injury. Isolated adult ventricular myocytes subjected to simulated ischemia (metabolic inhibition with CN and 0 glucose) provide a model in which we have shown Ca²⁺ influx via NCX contributes to Ca²⁺ loading and in which the degree of rise in [Ca²⁺]_i directly correlates with the degree of injury in the intact heart during ischemia-reperfusion (46). We therefore examined the effects of ODSH on [Ca²⁺]_i in this model. Figure 2 shows the effects of different concentrations of ODSH on myocyte [Ca²⁺]_i during 45 min of simulated ischemia [pacing-metabolic inhibition (PMI)]. ODSH induced a dose-dependent reduction of [Ca²⁺]_i, with a substantial effect at 100 μg/ml, a concentration similar to that present in the serum of humans when therapeutic anticoagulation is achieved during Phase I dose-escalation studies (data on file with FDA for IND No. 72,247, submitted by ParinGenix, Weston, FL).

Fig. 2.

Effects of ODSH on [Ca²⁺]_i during pacing-metabolic inhibition (PMI) in rabbit ventricular myocytes. PMI for 45 min caused a marked increase in [Ca²⁺]_i, and the magnitude of the increase was reduced by ODSH at concentrations of 10 (ODSH 10) and 100 μg/ml (ODSH 100). *P < 0.05; **P < 0.01; vs. PMI; n = 7. ODSH 1, 1 μg/ml ODSH.

To determine whether Ca²⁺ influx via reverse-mode NCX was involved in this effect of ODSH, we measured the reduction in [Ca²⁺]_i induced by ODSH in the presence of the NCX inhibitor, KB-R7943 (KB-R, 10 μM) (2, 41). KB-R has inhibitory effects on both forward and reverse modes of NCX, but under our experimental simulated ischemia conditions (myocytes depolarized, elevated [Na⁺]_i ) it is assumed that the predominant mode of function of NCX is reverse (Ca²⁺ influx). Therefore, any effect of KB-R is due primarily to an inhibition of reverse-mode NCX. The results are shown in Fig. 3. KB-R and ODSH caused a similar reduction in myocyte [Ca²⁺]_i during 45 min of simulated ischemia, and in the presence of KB-R, ODSH caused no significant further reduction in [Ca²⁺]_i. We have previously reported that 10 μM KB-R also decreases Na⁺ loading during PMI (46), an effect that could reflect the ability of KB-R to partially inhibit the Na⁺ current (41). Thus the ability of KB-R to decrease Ca²⁺ could be due to a combination of direct inhibition of reverse-mode NCX, and indirect inhibition of NCX due to a decrease in Na⁺ loading. However, as shown in Table 4, a more potent and selective inhibitor of NCX, SN-6, which has a less significant inhibitory effect in I_Na, (26), had an effect at 1 μM similar to that produced by 10 μM KB-R.

Fig. 3.

Influence of the Na⁺/Ca²⁺ exchange (NCX) inhibitor KB-R7943 (KBR) on [Ca²⁺]_i during PMI, and on the effects of ODSH. KBR (10 μM) and ODSH (100 μg/ml) caused a similar reduction in [Ca²⁺]_i, and in the presence of KB-R there was no further reduction in [Ca²⁺]_i induced by exposure to ODSH. *P < 0.05 vs. PMI; n = 6.

Table 4.

Effects of Na⁺/Ca²⁺ exchange inhibitor SN-6 and ODSH on [Ca⁺]_i during pacing-metabolic inhibition (PMI)

Groups	Means ± SE, n = 6
HEPES	198.3±24.6
PMI	589.6±71.3
PMI + SN-6, 1 μM	402.6±54.6*
PMI + ODSH, 100 mg/ml	396.9±64.0*
PMI + SN-6 (1 μM) + ODSH (100 mg/ml)	377.9±58.6*,†

^*P < 0.01 vs. PMI;

^†P (not significant) vs. PMI + SN-6.

These observations suggested that ODSH could be reducing Ca²⁺ loading either by directly inhibiting NCX, or by reducing Na⁺ loading and thereby indirectly inhibiting reverse-mode NCX. We therefore measured the direct effects of ODSH on NCX by means of voltage-clamp studies in intact isolated rabbit ventricular myocytes. The results are shown in Fig. 4. Rather than inhibiting NCX, ODSH caused a significant stimulation of exchange. This finding was initially surprising but is consistent with a previous report that heparin and heparan sulfate disaccharides can stimulate Ca²⁺ extrusion by NCX in smooth muscle cell lines (32). These results also indicated that ODSH was not reducing Ca²⁺ loading during PMI by direct inhibition of the exchanger. We therefore studied the effects of ODSH on Na⁺ loading, with results shown in Fig. 5. As previously demonstrated in this model, PMI caused a marked increase in Na⁺ loading, which was almost completely eliminated by 100 μg/ml ODSH.

Fig. 4.

Effects of ODSH on NCX current (I_NCX). A: ODSH (100 μg/ml) increased I_NCX over the voltage range of approximately −60 mV to +50 mV. B: Summary IV curve showing the stimulatory effect of ODSH on I_NCX. *P < 0.05; n = 5.

Fig. 5.

Effects of ODSH on [Na⁺]_i during PMI. PMI caused a significant rise in [Na⁺]_i, and this was reduced by exposure to 100 μg/ml ODSH. *P < 0.05 vs. HEPES control, **P < 0.05 vs. PMI; n = 6.

Recent work in our laboratory has indicated that a major component of Na⁺ loading that occurs in this model during PMI is inhibited by ranolazine (46) and thus appears to be mediated by Na⁺ influx via the late Na⁺ current (I_Na-L) (1). This increased Na⁺ loading causes increased Ca²⁺ loading via NCX. To determine whether ODSH could be altering Na⁺ loading via a similar inhibitory effect on I_Na-L, we examined the effects of ODSH on [Ca²⁺]_i during PMI in the presence of ranolazine. The results are shown in Fig. 6. ODSH and ranolazine reduced [Ca²⁺]_i to a similar degree, and in the presence of ranolazine, ODSH had no additional effect on Ca²⁺ loading. These observations provided strong indirect evidence that ODSH was decreasing Na⁺ (and Ca²⁺) loading via an inhibition of I_Na-L. To provide more support for this idea, we examined the effects of a selective activator of I_Na-L, sea anemone toxin II (30), on [Na⁺]_i in paced myocytes in the absence of metabolic inhibition. The results are shown in Fig. 7. Exposure to 10 nM anemone toxin II caused a substantial increase in [Na⁺]_i in paced myocytes that was almost completely inhibited by ODSH 100 μg/ml. ODSH also reduced [Na⁺]_i in myocytes in the absence of anemone toxin II but to a much smaller extent. ODSH 100 μg/ml had no effect of fluorescence intensity of fluoresceine-labeled microspheres, indicating there was no quenching of Na Green (or Fluo-3) fluorescence by ODSH.

Fig. 6.

Influence of ranolazine (Ran) on [Ca²⁺]_i during PMI, and on the effects of ODSH. Ran (10 μM) plus ODSH 100 (μg/ml) caused a similar reduction in [Ca²⁺]_i, and in the presence of Ran there was no further reduction in [Ca²⁺]_i induced by exposure to ODSH. *P < 0.05 vs. PMI; n = 7.

Fig. 7.

Effect of ODSH on the rise in [Na⁺]_i-induced by exposure to sea anemone toxin II (ATX). Compared with control conditions [HEPES + pacing (HP)] exposure to ATX 10 nM caused a highly significant increase in [Na⁺]_i, and this was reduced by exposure to 100 μg/ml ODSH. ODSH also caused a small but significant decrease in [Na⁺]_i during control conditions (no ATX, HP alone). *P < 0.05, **P < 0.01 vs. HP; ***P < 0.01 vs. HP + ATX; n = 6.

Effect of ODSH on Myocardial Na⁺ Current

The above results provide strong, but indirect, evidence that ODSH is decreasing Na⁺ influx via I_Na-L. To provide more direct evidence for this effect, we studied the influence of ODSH on Na⁺ channel I-V relationships. The results of those experiments are shown in Fig. 8. At a relatively high concentration of ODSH heparinic acid (1 mg/ml) we were able to detect a rightward shift in the I-V relationship, which results in a decrease in the inward Na⁺ current magnitude.

Fig. 8.

Effects of ODSH on Na⁺ channel ionic currents (I_NA). A: peak Na⁺ current-voltage (I-V) relationships from a holding membrane potential (V_hp) of −150 mV; values in the presence of 1 mg/ml ODSH heparinic acid (○). All Na⁺ currents were normalized to the maximal inward I_Na in control. Lines represent the fits to the Boltzmann equation for peak intravenous relationships (see methods). For n = 6 cells, there were no significant differences between control and ODSH heparinic acid. Control values were maximal conductance (G_max) = 1, voltage at the half point of the relationship (V_1/2) = −57 ± 3 mV, and slope (s) = −6.3 ± 0.6. In ODSH heparinic acid values were G_max = 1 ± 0.04, V_1/2 = −57 ± 3 mV, and slope (s) = −6.4 ± 0.6 . B: peak I-V relationships from a holding potential of −110 mV for control (•) and in 1 mg/ml ODSH heparinic acid (○). All I_Na were normalized to the maximal inward I_Na in control. Lines represent the fits to the Boltzmann equation for peak intravenous relationships (see methods). For n = 6 cells, G_max was significantly decreased from 1 to 0.87 ± 0.05 in ODSH heparinic acid, the slope was not significantly changed (−6.1 ± 0.7 in control vs. −6.2 ± 0.8 in ODSH heparinic acid), while there was a small leftward shift in the V_1/2 in control (−57 ± 3 mV) compared with ODSH heparinic acid (−56 ± 3 mV). C: steady-state voltage-dependent Na⁺ channel availability (SSI) curves in control (•) and in 1 mg/ml ODSH heparinic acid (○). All I_Na in each cell were normalized to its I_max from the fit of a Boltzmann relationship to SSI curve in control (see methods). Lines represent the fits to the Boltzmann equation. For n = 6 cells, V_1/2 was significantly shifted from −102 ± 5 mV in control to −104 ± 2 mV in ODSH heparinic acid. There were minimal but significant shifts in the slope from 7.4 ± 7 in control to 7.6 ± 6 in ODSH heparinic acid but not in I_max from 1 to 0.99 ± 0.01 in heparinic acid. D: late I_Na determined by saxitoxin subtraction of leak currents from a holding potential of −110 mV to step potentials from −100 to 20 mV (see methods) for 100 ms. •, means ± SE I_Na in control; ○ means ± SE in ODSH heparinic acid for 4 cells. Note the decrease in mean current in ODSH heparinic acid compared with control across the range of potentials. The values at −50, −40, and −30 mV were significantly different between control and ODSH heparinic acid.

DISCUSSION

The anticoagulant glycosaminoglycan heparin is well-known to reduce myocardial reperfusion injury through nonanticoagulant mechanisms. Previously, this effect was attributed to the ability of heparins to inhibit complement (7, 10) and to inhibit neutrophil migration into ischemic-reperfused myocardium (39) through blockade of selectin-mediated neutrophil rolling (40). In this study, we extend the understanding of heparin's protective pharmacology in reperfusion injury to include important effects on Na⁺ and Ca²⁺ loading during early reperfusion. Similar to results in dogs (39), we showed that ODSH significantly inhibits myocardial reperfusion injury in closed-chest pigs (Fig. 1A). However, the mechanism of protection in pigs cannot be attributed entirely to reduction in late neutrophil-mediated myocyte injury, since doses of 15 mg/kg and 45 mg/kg ODSH provided almost the same degree of reduction in infarct size (Fig. 1A), but, while the 45 mg/kg dose reduced MPO activity in necrotic infarcted myocardium, the 15 mg/kg dose had no effect (Fig. 1B). In searching for an alternative mechanism, we studied adult rabbit myocytes subjected to simulated ischemia. Rabbit ventricular myocytes have electrophysiologic and Ca²⁺ homeostasis characteristics that resemble those of larger mammals including humans (35, 36). Our recent study of ranolazine (46), showed that the IC₅₀ for an effect of ranolazine to reduce Ca²⁺ loading in rabbit myocytes during simulated ischemia (2.1 μM plasma concentration) was almost identical to the IC₅₀ for an effect of ranolazine to prolong treadmill exercise time in patients with coronary artery disease and angina pectoris (2.5 μM plasma concentration). This observation further supports the validity of our use of rabbit myocytes in these studies.

We noted that ODSH significantly reduced Ca²⁺ accumulation in paced rabbit myocytes subjected to simulated ischemia in a dose-dependent fashion (Fig. 2). The maximally observed reduction in [Ca²⁺]_i occurred at 100 μg/ml, an ODSH concentration that produces therapeutic anticoagulation in normal human volunteers. The decrease in [Ca²⁺]_i was produced by preventing reverse-mode operation of the NCX during ischemia with a similar degree of effectiveness as the direct NCX inhibitors KB-R (Fig. 3) or SN-6 (Table 4). On direct testing in voltage-clamped myocytes, ODSH stimulated rather than inhibited NCX activity (Fig. 4). However, ODSH reduced [Na⁺]_i in paced ischemic myocytes (Fig. 5) and decreased [Ca²⁺]_i during PMI as effectively as the I_Na-L inhibitor ranolazine (Fig. 6). ODSH also inhibited Na⁺ loading stimulated in nonischemic myocytes by the direct Na⁺ channel opener ATX (Fig. 7). These results suggest that at 100 μg/ml, ODSH causes a marked reduction in Na⁺ loading mediated by I_Na-L, which results in a marked decrease in Ca²⁺ loading via reverse NCX. In the intact ischemic heart, reperfusion results in a burst of ROS production (48), which would be expected to increase Na⁺ influx via I_Na-L (33) and thus increase Ca²⁺ loading, effects inhibited by ODSH. We also demonstrated direct inhibition by ODSH of the human heart Na⁺ channel expressed in HEK293 cells, but only at a 10-fold higher concentration of 1 mg/ml (Fig. 8). The higher concentration required in this more simple reconstituted system may suggest that in intact ventricular myocytes, heparin does not directly inhibit Na⁺ channels but influences membrane-associated Na⁺ channel regulatory relationships that are still poorly characterized.

When microinjected into cells, heparin is well recognized to competitively inhibit inositol 1,4,5-triphosphate-activated Ca²⁺ release (11), and is also known to modulate the ryanondine receptor (6) and inhibit binding of ligands to voltage-dependent myocardial l-type Ca²⁺ channels (21). In nonexcitable cells, extracellular heparin has also been reported to suppress Ca²⁺ by multiple mechanisms, but only in milligram per milliliter concentrations (25). However, effects of extracellular heparin on Na⁺ channels are less well investigated. Recently, the epithelial Na⁺ channel (ENaC) has been shown to be constitutively phosphorylated at its COOH terminus by the ubiquitous membrane-associated protein kinase casein kinase-2 (CK2), an effect which maintains ENaC activity by preventing binding of the ubiquitin ligase Nedd4–2 to ENaC, with subsequent ubiquitination and degradation (4). Heparin is a potent CK2 inhibitor (27) that might be able to access membrane-associated CK2 in some cells. Extracellular heparin inhibits protein kinase C activity in smooth muscle cells (14) and NF-κB activation in perfused rat myocardium (39). Whether CK2 phosphorylation regulates stability of myocardial Na⁺ channels is presently unknown, but provides a future site for investigating the molecular mechanisms of our current observations.

As recently reviewed by Yellon and Hasenloy (44), there is extensive experimental evidence in animal models that significant myocardial injury can occur as a consequence of reperfusion. Although there has been skepticism as to whether or not this phenomenon occurs in humans, several recent studies have supported the presence of reperfusion injury in humans. Ischemic postconditioning, achieved by repetitive occlusion and reperfusion of the coronary artery in the early minutes after reperfusion of acute myocardial infarction (47), has been studied in 30 patients undergoing coronary angioplasty for acute myocardial infarction (34). At the beginning of reperfusion by direct stenting, postconditioning was performed within 1 min of reflow by four episodes of 1-min inflation and 1-min deflation of the angioplasty balloon to produce four brief periods of ischemia and reflow. Compared with control subjects, this simple procedure produced a 36% reduction in infarct size as measured by the magnitude of cardiac enzyme release. Ischemic postconditioning has been shown to decrease myocyte injury by reducing intracellular Ca²⁺ overload during the early minutes of flow restoration (38). The benefit of postconditioning is dependent upon activation of the cardioprotective enzyme protein kinase Cε (PKCε), demonstrated in animal models by the fact that the infarct-sparing effect of postconditioning is abolished by PKCε inhibition (45). Because PKC enzymes can enhance both forward and reverse modes of NCX operation (17), it is possible that ischemic postconditioning functions to decrease myocardium Ca²⁺ overload by stimulation of PKCε-mediated phosphorylation of the exchanger and thus enhance the forward NCX, thereby providing an overall decrease in myocyte Ca²⁺ concentration after reperfusion. More recently, Piot et al. (28) studied 58 patients with acute ST-elevation myocardial infarction who were randomly assigned to receive intravenous cyclosporine 2.5 mg/kg or placebo < 10 min prior reperfusion achieved by stenting of the occluded coronary artery. These investigators found that administration of cyclosporine, which inhibits the MPT, reduced creatine kinase release and reduced infarct size assessed by MRI by ∼25%. This important preliminary clinical study, as well as a number of studies in animal (3, 18) and isolated myocyte (29) models of ischemia-reperfusion injury, support the importance of induction of the MPT in the reperfusion injury cascade. These studies are relevant to our investigations of pharmacological postconditioning with ODSH.

The MPT is thought to contribute to irreversible injury caused by ischemia-reperfusion, and is proposed to result from opening of a pore in the inner mitochondrial membrane, consisting in part of the voltage-dependent anion channel, the adenine nucleotide translocator, and cyclophilin (5). Initiation of the MPT is thought to result from increased mitochondrial Ca²⁺ loading and increased ROS in mitochondria (8, 42), and results in release of mitochondrial Ca²⁺ and cytochrome c, as well as dissipation of the mitochondrial membrane potential, with increased consumption and decreased production of ATP. These events can cause necrotic and apoptotic myocyte death (20, 24). Our findings in isolated myocytes provide two mechanisms by which ODSH could indirectly inhibit the MPT in reperfused ischemic myocardium: 1) a reduction in Na⁺ loading via the Na⁺ channel during early reperfusion, which would indirectly decrease Ca²⁺ loading via reverse-mode NCX; and 2) direct stimulation of Ca²⁺ extrusion via forward-mode NCX later, after reperfusion. Both of these effects would decrease myocyte [Ca²⁺]_i after reperfusion, and thus decrease mitochondrial Ca²⁺ loading and reduce the induction of the MPT and irreversible injury.

Our results, suggesting that pharmacological postconditioning with ODSH primarily decreases reperfusion injury by reducing Na⁺ influx via the I_Na-L, are consistent with a recent reports that ranolazine (13) and other recently reported (22) I_Na-L inhibitors also reduce infarct size after ischemia-reperfusion. Based on the work of Inserte et al. (16), an agent such as ODSH, which decreases myocyte Ca²⁺ influx via reverse-mode NCX early during reperfusion and increases Ca²⁺ extrusion via forward NCX later after reperfusion, would be expected to provide substantial protection from reperfusion injury. Heparin is a well-known biologic product with a long history of safe and beneficial use as an anticoagulant during interventional cardiology. The surprising demonstration that it also significantly reduces reperfusion injury through ion channel effects early during reperfusion makes heparin an interesting molecular platform structure from which novel new analogs might be developed to provide both the anticoagulant activity and protection from reperfusion injury needed for rescue percutaneous coronary intervention.

GRANTS

Pig infarct studies were supported by a sponsored research agreement between Emory University and ParinGenix, the owner of 2-O, 3-O desulfated heparin. K. W. Spitzer was supported by National Heart, Lung, and Blood Institute Grant R37-HL-42873.

DISCLOSURES

Dr. Kennedy owns stock in ParinGenix, which is developing 2-O, 3-O desulfated heparin for use in the treatment of respiratory diseases. In vitro studies performed at the University of Utah were not supported by ParinGenix, but 2-O, 3-O desulfated heparin was a gift to W. H. Barry.