: The gene gas-1 encodes a subunit of complex I of the mitochondrial electron transport chain in Caenorhabditis elegans. A mutation in gas-1 profoundly increases sensitivity of C. elegans to volatile anesthetics. It is unclear which aspects of mitochondrial function account for the hypersensitivity of the mutant.
: Oxidative phosphorylation was determined by measuring mitochondrial oxygen consumption using electron donors specific for either complex I or complex II. Adenosine triphosphate concentrations were determined by measuring luciferase activity. Oxidative damage to mitochondrial proteins was identified using specific antibodies.
: Halothane inhibited oxidative phosphorylation in isolated wild-type mitochondria within a concentration range that immobilizes intact worms. At equal halothane concentrations, complex I activity but not complex II activity was lower in mitochondria from mutant (gas-1) animals than from wild-type (N2) animals. The halothane concentrations needed to immobilize 50% of N2 or gas-1 animals, respectively, did not reduce oxidative phosphorylation to identical rates in the two strains. In air, adenosine triphosphate concentrations were similar for N2 and gas-1 but were decreased in the presence of halothane only in gas-1 animals. Oxygen tension changed the sensitivity of both strains to halothane. When nematodes were raised in room air, oxidative damage to mitochondrial proteins was increased in the mutant animal compared with the wild type.
: Rates of oxidative phosphorylation and changes in adenosine triphosphate concentrations by themselves do not control anesthetic-induced immobility of wild-type C. elegans. However, they may contribute to the increased sensitivity to volatile anesthetics of the gas-1 mutant. Oxidative damage to proteins may be an important contributor to sensitivity to volatile anesthetics in C. elegans.
THE mechanisms for the anesthetic effects of volatile anesthetics are incompletely understood.1,2Using immobility as an endpoint, we established that the interactions of multiple genes are crucial in controlling the sensitivity of Caenorhabditis elegans to halothane.3The mutation fc21 in one gene, gas-1 (g eneral a nesthetic s ensitive), causes hypersensitivity to all straight chain alcohols up to C12 and to volatile anesthetics.3,4Nematodes with the gas-1(fc21) mutation move normally in air, indicating a functional neuromuscular system. However, they have a reduced lifespan, slow growth rates, and an increased sensitivity to the deleterious effects of paraquat and hyperoxia.5
Gas-1 encodes a homolog of the bovine 49-kd subunit of nicotinamide adenine dinucleotide (NADH) ubiquinone oxidoreductase (complex I), the first complex of the mitochondrial electron transport chain.6GAS-1 protein expression is consistent with mitochondrial localization. In isolated mitochondria, complex I–dependent oxidative phosphorylation and electron transport are specifically depressed by the gas-1(fc21) mutation, confirming that gas-1 encodes a functional subunit of complex I in C. elegans .7
In all organisms studied, complex I catalyzes the first step of electron transport, the transfer of electrons from NADH to a quinone.8–10The energy released is harnessed to establish a proton gradient across the inner mitochondrial membrane. In mitochondria, the proton gradient ultimately drives the production of adenosine triphosphate (ATP). The exact composition of complex I varies widely across species, but all forms of complex I contain homologs of the 15 subunits that constitute the enzyme in Paracoccus . One of these core subunits is the 49-kd protein, whose function is incompletely defined. A knockout mutant of the “49-kd gene” in Neurospora completely lacked complex I activity because the matrix arm of the enzyme complex failed to assemble.11The matrix arm of complex I contains the binding site for NADH as well as all but one of the redox centers. The 49-kd subunit from Rhodobacter is thought to bind the head group of quinones such as the electron acceptor, ubiquinone,12and may contribute to the movement of ubiquinone from the membrane to the matrix arm of complex I.13Recent data indicate that this subunit interacts with another subunit of complex I to bind ubiquinone.14Lastly, our previous studies indicated that mutations in the 49-kd subunit of C. elegans specifically decreased complex I–dependent metabolism.7Therefore, 49-kd proteins seem to be essential for the structure and function of complex I.
Electron transport to oxygen (aerobic respiration) in intact mitochondria is ordinarily tightly coupled to ATP production (phosphorylation). Measurement of oxygen consumption and adenosine diphosphate (ADP) utilization by isolated mitochondria provide an assessment of the ability of mitochondria to generate high-energy phosphate bonds for cellular metabolism. Individual complexes of the electron transport chain can then be assayed for their effects on oxidative phosphorylation.
The volatile anesthetic halothane accumulates in mitochondria in C. elegans (Roderick Eckenhoff, M.D., Professor, Department of Anesthesiology, University of Pennsylvania, Philadelphia, Pennsylvania, written personal communication, February 2004) and in mammals.15Previous studies from other investigators indicated that complex I in mammalian mitochondria is the most sensitive complex of the electron transport chain to inhibition by volatile anesthetics.16,17In contrast, the remaining mitochondrial complexes (II–V) are less sensitive to these compounds. In addition, children with certain complex I deficiencies are very sensitive to induction with the volatile anesthetic sevoflurane.18However, it remains unclear whether volatile anesthetics actually cause anesthesia by inhibition of mitochondrial function. Because patients with mitochondrial disease may show altered sensitivities to volatile anesthetics, it is important to understand the mechanisms by which such changes may arise.
Gas-1 animals are hypersensitive to halothane, and their mitochondria have profoundly decreased rates of oxidative phosphorylation via complex I–specific substrates, suggesting that complex I–dependent oxidative phosphorylation may directly control the anesthetic response in worms. We hypothesize that the same minimum absolute value of complex I–dependent oxidative phosphorylation is necessary in both wild-type and gas-1 animals to sustain movement. This threshold value of complex I activity would occur at the EC50s for each strain, the concentrations of halothane that immobilize each strain of animals. We therefore compared the doses of halothane required to decrease oxidative phosphorylation in mitochondria to those required to immobilize both wild-type and gas-1 animals. We measured the effects of halothane on oxidative phosphorylation with electron donors specific for complexes I or II. We also compared the effects of halothane on state 3 respiration (high ADP availability and thus close to maximal respiration rates) and state 4 respiration (ADP-limiting conditions and thus respiration is constrained by the inability to efficiently transport electrons to generate ATP). By comparing the effects of halothane on state 3 and state 4, we determined whether decreases in respiration were the result of direct inhibition of the electron transport chain or inhibition of coupling of electron transport to the generation of ATP.
Other facets of mitochondrial function may contribute to the behavior of C. elegans in anesthetics. Energy stores may be a major determinant of mobility in the face of anesthetics, if anesthetics decrease rates of oxidative phosphorylation. Therefore, we measured the ratio of ATP to ADP and adenosine monophosphate in wild-type and gas-1 animals in air and in halothane. In addition, we have noted that gas-1 is very sensitive to increased oxygen tensions and paraquat.5Therefore, we examined whether oxidative damage correlated with changes in anesthetic sensitivity.
Materials and Methods
The conventions for C. elegans nomenclature are followed throughout.19The wild-type C. elegans , N2, was obtained from the Caenorhabditis Genetics Center (Minneapolis, MN). Gas-1(fc21) was isolated in a screen for immobile worms in 3.5% enflurane after mutagenesis of N2 with ethylmethanesulfonate (N2 is immobilized by 6.5% enflurane).20Standard techniques were used for growing and maintaining cultures of C. elegans .20,21
Preparation of Mitochondria
All preparations were performed on ice. Two to three grams of worms cleaned of Escherichia coli were suspended in MSM-E (220 mm mannitol, 70 mm sucrose, 5 mm MOPS, 2 mm EDTA, pH to 7.4 with KOH). A polytron (Brinkman Instruments, Westbury, NY) was used for initial rupture: 20 s at 14,000 rpm. To the homogenate, 5 mg/(g worms) proteinase type XXVII (Sigma, St. Louis, MO) was added, and the resulting mixture was stirred for 10 min. Immediately afterward, the slurry was homogenized in a glass Potter/Elvehjem tissue grinder with a Teflon pestle (Dupont, Wilmington, DE). After adding 1 volume MSM-E containing 0.4% bovine serum albumin, the homogenate was centrifuged (300g , 10 min, 4°C). The supernatant was filtered through three layers of gauze and recentrifuged (7,000g , 10 min, 4°C). The mitochondrial pellet was resuspended in MSM-E and washed twice by centrifugation (7,000g , 10 min, 4°C). The final mitochondria pellet was resuspended in 100 μl MSM-E. Total protein was determined by the Lowry assay with bovine serum albumin as a standard.22
Oxidative Phosphorylation Assay
Polarographic measurement of oxidative phosphorylation was performed as previously described.23Briefly, mitochondrial oxygen uptake was followed with a Clark type electrode (Yellow Springs Instruments, Yellow Springs, OH) connected to a chart recorder via a YSI Oxygen monitor. The electrode chamber was loaded with 500 μl incubation medium (100 mm KCl, 50 mm MOPS, 1 mm EGTA, 5 mm potassium phosphate, 1 mg/ml defatted bovine serum albumin) and equilibrated at 30°C before the oxygen monitor was set to 100% oxygen saturation (222.5 natom O) (fig. 1). The following components were then added sequentially: mitochondria (maximum 500 μg); 25–50 nmol ADP, allowing the mitochondria to consume endogenous substrates; either 10 mm l-malate or 20 mm succinate, providing a defined electron donor specific for complex I or II, respectively; 25–100 nmol ADP to determine state 3 (phosphorylating conditions) and state 4 (nonphosphorylating conditions) respiration rates; 1,000 nmol ADP to determine “high ADP” rate (maximally phosphorylating); and 100 nmol 2,4-dinitrophenol (2,4 dinitrophenylhydrazine) to uncouple the mitochondria. Finally, N,N,N′,N′-tetramethyl-p-phenylenediamine (TMPD)–ascorbate (a complex intravenous substrate) is added as an internal control for mitochondrial quality.
For measurements in the presence of halothane, halothane-saturated incubation medium was created as follows. A three-phase system of liquid halothane, incubation medium, and air in a closed test tube was shaken and incubated at 30°C overnight, allowing the phases to separate and the concentrations to come to an equilibrium in all phases. Halothane concentration of the air over the saturated solution was determined to be 34% by gas chromatography. Dilutions of the saturated solution were made in a sealed chamber before mitochondria were added. The final concentration of halothane was calculated from the dilution. If the mitochondria acted as pure lipid, we estimated that this would decrease the concentration of halothane in solution by approximately 1% only (i.e. , from 34% to 33%). Therefore, further corrective calculations were unnecessary. Experimental data reported are the results from three independent experiments for N2 and four separate experiments for gas-1(fc21) .
Respiration (oxygen uptake) rates and ADP/O were calculated as previously described.7,24,25The rate of phosphorylation was calculated by multiplying the rate of state 3 oxygen consumption (natom O · min−1· mg protein−1) by ADP/O to obtain nmol ADP phosphorylated · min−1· mg protein−1. Values for calculating the mean and SD were the results from four separate preparations from each strain.
Worms were synchronized by allowing adult worms to lay eggs for 2 h on agar plates with bacterial lawns. The adults were then removed, and the resulting eggs were grown to young adults. The resulting adults were cleaned of E. coli by moving them to an agar plate devoid of bacteria. After 10 min, the animals were moved to a second agar plate without bacteria for 90 min and then placed into 20 μl S-basal solution, frozen in liquid nitrogen overnight. The frozen animals were lyophilized and stored at −80°C until use. Animals exposed to halothane were also synchronized young adults cleaned of E. coli and then placed on an agar plate for exposure to halothane for 2 h. These worms were then immediately placed in S-basal, transferred to a dry ice–acetone bath, and then transferred into liquid nitrogen and lyophilized. Gas-1 worms, which lag in development compared with the wild type, were synchronized as for the wild type but required an additional 24 h to reach young adulthood. Lyophilized animals were sonicated in 100 μl trichloroacetic acid, 2.5%, on ice and extracted for 10 min in room air. Five microliters of extract was added to 95 μl of a 20-mm HEPES buffer to measure ATP.26ATP was measured by light emission in a standard luciferin–luciferase assay. ADP was measured after adding phosphenol pyruvate and phosphenol pyruvate kinase to the same assay. Adenosine monophosphate was measured by further adding myokinase (Sigma) to the mixture with cytosine triphosphate to drive the reaction. In each case, the enzymes were allowed to convert their substrates to ATP for 30 min at room temperature. ATP was expressed as a ratio of ATP to the total adenosine species. All values reported are the results of 11 separate measurements.
We used two antibody preparations for identifying oxidative damage to protein. The first is the OxyBlot protein oxidation detection kit (InterGen, Burlington, MA). As a consequence of protein oxidation, carbonyl groups are introduced into protein side chains.27The carbonyl groups are derivatized to 2,4-dinitrophenylydrazone by a reaction with 2,4 dinitrophenylhydrazine. The derivatized protein samples are then run on polyacrylamide gels and Western blotted. The blots are exposed to the primary antibody, specific to the 2,4 dinitrophenylhydrazine moiety of the proteins, and followed by a horseradish peroxidase antibody secondary antibody.
The second procedure does not require a derivatization step. Instead, mitochondrial proteins are transferred to a Western blot and incubated with an antibody specific for 4-hydroxy-2-nonenal (HNE) Michael adducts.28HNE is a common reactive intermediate released by reactive oxygen species and interacts with cysteine, lysine, or histidine. Probing with the antibody uses common immunohistochemistry techniques.
EC50s were calculated using the methods described by Waud.29Values for oxidative phosphorylation rates and for EC50s were compared using analysis of variance for multiple comparisons.30Mean values in tables 1 and 2were compared by a Student t test. Significance was defined as P < 0.05.
Mitochondria from wild-type nematodes can metabolize malate or succinate.7,31Malate is a complex I–dependent electron donor, whereas electrons enter complex II specifically via succinate. Complex I–dependent oxidative phosphorylation was decreased in mitochondria from gas-1 animals compared with those from N2. Complex II–dependent oxidative phosphorylation rates were increased in gas-1 animals compared with those from N2 (table 1). These results were previously published7and are presented here to serve as controls for the effects of halothane.
Halothane and Oxidative Phosphorylation
To address the question of whether halothane directly affects respiration, we used malate or succinate to specifically isolate the effects of halothane on complex I or II, respectively. Halothane inhibited complex I–dependent oxidative phosphorylation in mitochondria from both N2 and gas-1 animals (fig. 2A) with IC50s (IC50is that concentration causing 50% inhibition of activity) of 4.1% and 4.3% halothane, respectively (table 2). At all concentrations of halothane (excepting the highest one, 6.4% halothane) the rates of state 3 respiration were significantly lower in gas-1 than they were in N2. In contrast, rates of state 4 (ADP-limited) respiration were not strongly affected by halothane (fig. 2B). In uncoupled mitochondria, rates were again inhibited by halothane (IC50s of 4.7% for N2 and 4.8% for gas-1 ), indicating that phosphorylation of ADP to ATP was not limiting in the presence of halothane (fig. 2C).
Fifty percent of N2 animals are reversibly immobilized at 3.2% halothane.4At this concentration, mitochondria isolated from N2 consume 80 natom O · min−1· mg mitochondrial protein−1in malate-dependent state 3 respiration (fig. 2A). Gas-1(fc21) animals are hypersensitive to halothane; the EC50of halothane is 1.2% for this mutant. When exposed to 1.2% halothane, mitochondria from gas-1(fc21) animals respired at 24 natom O · min−1· mg protein−1in malate-dependent state 3 respiration (fig. 2A). At concentrations of halothane that immobilized N2 animals, complex I–dependent rates were higher than those seen in gas-1 mitochondria in air. Therefore, no minimum threshold value of oxygen consumption in mitochondria correlated with the concentrations of halothane that immobilized animals in the two strains.
Halothane also inhibited complex II–dependent oxidative phosphorylation in mitochondria from N2 and gas-1 animals (fig. 3). At all concentrations of halothane tested, the rates for state 3 respiration and 2,4 dinitrophenylhydrazine uncoupled respiration were higher in gas-1 than in N2 (figs. 2A and C). State 4 respiration was not significantly different between the two strains (fig. 3B). Notably, the rates in mitochondria from N2 in room air (when the animal moved well) were considerably lower than the rates in mitochondria from gas-1 animals at concentrations that anesthetize the mutant animal (1% halothane). At concentrations of halothane causing immobility in gas-1 , no significant change in complex II–dependent rates was observed. As reported earlier, when an uncoupler (2,4 dinitrophenylhydrazine) of respiration to ATP synthesis was added, there was no increase in respiration with any substrate compared with respiration without 2,4 dinitrophenylhydrazine.7The addition of 2,4 dinitrophenylhydrazine also did not increase respiration at any halothane concentration (data not shown). These results again demonstrate that phosphorylation is not limiting in the presence of halothane.
Halothane and ATP Concentrations
The percent of adenosine species present as ATP was measured in the two strains. In the absence of halothane, we found no significant difference between N2 animals and gas-1 animals (fig. 4). The percent adenine present as ATP was 68% for N2 and 66% for gas-1 , in both cases consistent with values found in other species.32N2 animals exposed to approximately 150% of their EC50(4.9% halothane) showed a slight decrease in their ratio of ATP to total adenine nucleotides (68% to 63%). In contrast, gas-1 animals showed a more pronounced decrease in their ATP levels (66% to 43%) when exposed to 150% of their EC50concentration of halothane (1.9% halothane).
To identify whether oxidative damage to anesthetic targets or effectors play a role in anesthetic sensitivity, we grew N2 and gas-1 animals in low and high concentrations of oxygen. Gas-1 animals grown in 3% oxygen had a significant decrease in their sensitivity to halothane (EC50of 1.6%) (fig. 5) compared with animals grown in room air (EC50of 0.9%). N2 animals grown in 60% oxygen had a significant increase in their sensitivity to halothane (EC50of 2.2%) compared with those grown in room air (EC50of 3.2%) (fig. 5). Gas-1 animals did not survive exposure to 60% oxygen for a sufficient time to test the anesthetic effect. N2 animals grown in 3% oxygen had no significant change in their halothane sensitivity (data not shown).
Exposure to oxygen is expected to increase oxidative damage to proteins. We examined whether a difference existed in oxidative damage to mitochondrial proteins from N2 and gas-1 when grown in room air. When an antibody that detects oxidative damage to proteins was used as a probe to mitochondrial proteins (OxyBlot), increased oxidation of specific proteins from gas-1 mitochondria was seen compared with mitochondria from N2 (fig. 6A). When an antibody specific for the free radical damage inflicted by 4-hydroxynonenol, a product of lipid peroxidation, was used, one protein of approximately 26 kd was noted to be more damaged in gas-1 than in N2 (fig. 6B). This is the same size as a major band seen on OxyBlot (fig. 6A). Therefore, the free radical damage resulting from the gas-1 mutation is increased from that in N2 and specific to certain proteins. In addition, we previously showed that a mutation in a complex II subunit, known as mev-1 ,33did not alter anesthetic sensitivity.5We found that increased oxidative damage was also seen in mitochondrial protein from mev-1 but that the distribution of damaged proteins was different than that seen in gas-1 (data not shown).
We have noted that changes in mitochondrial function may alter sensitivity to volatile anesthetics in nematodes and in man.6,18Alterations in mitochondrial function could affect an the phenotype of an animal in any of several ways. We studied three such possible mechanisms by which mitochondrial function could alter anesthetic sensitivity.
A straightforward mechanism may be that an acute decrease in metabolism would lead to inadequate energy supplies to the cell. In turn, this would lead to a decrease in neuromuscular function, causing the anesthetized state. If such a mechanism were solely responsible for immobility in the nematode, an identical rate of mitochondrial oxygen consumption should be reached at the concentration of halothane that produces immobility in each strain of worms. We found that the rate of oxidative phosphorylation was greatly decreased in gas-1 mitochondria at all halothane concentrations when compared with those from N2. However, a common threshold value for oxygen consumption at halothane concentrations causing immobility was not seen. Notably, the rate of oxygen consumption seen in N2 mitochondria at its EC50was greater than that seen in gas-1 mitochondria in room air. Because gas-1 animals move well in room air (and their respiration must be adequate for movement), the rate of complex I–dependent oxidative phosphorylation cannot completely account for halothane-induced immobility in C. elegans . This is in contrast to our reported results for sensitivity to ethanol in these strains.34In ethanol, both gas-1 and N2 mitochondria reached a similar threshold rate for respiration at the concentrations of drug causing immobility. We interpreted these data as consistent with the possibility that the decrease in respiration was directly causative of the ethanol induced immobility. Currently, we do not know whether other volatile anesthetics, such as isoflurane, will demonstrate patterns similar to halothane or to ethanol. Because gas-1 is hypersensitive to all volatile anesthetics and to ethanol, such data are of interest.
At all concentrations of halothane studied, the rates of complex II–dependent respiration were higher in gas-1 than in N2. This is a significant disparity compared with the hypersensitivity of gas-1(fc21) animals to halothane3,4; therefore, an absolute rate of succinate-dependent respiration could not be linked to halothane concentrations that cause immobility of whole animals.
It is possible that the combined rate of complex I and complex II oxidative phosphorylation controls the anesthetic response. Because complex II function actually increases in gas-1 mitochondria at the halothane concentration causing immobility in the corresponding animals, this is an interesting possibility. However, because we do not know the relative contributions of complexes I and II to the nematode, we are unable to rule this in or out. Our measurements indicate that ATP ratios do not decrease in wild-type animals on exposure to an anesthetizing concentration of halothane. Therefore, an acute lack of ATP is not necessary for immobilization of N2. However, a striking decrease is seen in gas-1 ATP concentrations when they are exposed to halothane. In the case of gas-1 animals, ATP production may be rate limiting in causing immobility. Such a decrease in ATP production may explain why some patients with mitochondrial disease have an increased sensitivity to sevoflurane.18We have not ruled out that both ATP production and utilization are both decreased by volatile anesthetics in a congruent manner, leaving ATP stores intact before and after exposure to anesthetics in the normal worm. Our results also do not allow us to measure absolute ATP concentrations. Therefore, if the total pool of adenylates were decreased in gas-1 compared with N2, the amount of ATP would also be decreased.
However, even if ATP production is rate limiting in the mutant, it does not seem to represent the only reason these animals are hypersensitive to halothane. We can make the mitochondrial mutant relatively resistant to halothane by growing it in a low concentration of oxygen; we can make wild-type animals hypersensitive to halothane by growing them in high concentrations of oxygen. In fact, the sensitivities of N2 and the mitochondrial mutant gas-1 were almost identical after growing N2 in high oxygen concentrations and gas-1 in low oxygen concentrations. However, we do not know whether these effects are reversible. Mitochondria are known to be a major source of free radical production with resulting oxidative damage to proteins.35It seems, then, that free radical damage may also mediate anesthetic sensitivity. By staining Western blots with antibodies specific for free radical damage, we found that, when grown in room air, oxidative damage was increased in gas-1 compared with N2. Surprisingly, this damage was fairly specific, with the primary effect being seen in a 26-kd mitochondrial protein. Because a second mitochondrial mutant with increased oxidative damage, mev-1 , had no changes in anesthetic sensitivity, nonspecific oxidative damage does not seem to alter anesthetic sensitivity. Instead, it may be that certain specific proteins are damaged in gas-1 mutants and that the accumulation of these damaged molecules is responsible for the increase in anesthetic sensitivity. This is consistent with our observation that gas-1 larvae are less sensitive to volatile anesthetics that adults. We are in the process of identifying the 26-kd protein as a first candidate for such a protein.
Oxidative damage resulting from mitochondrial function is currently thought to be one of the major causes of aging in all animals, including humans.35Because it is also known that anesthetic sensitivity increases with advancing age,36,37the current results may also indicate that hypersensitivity of the aged might be the result of deteriorating mitochondrial function. The results in children with mitochondrial myopathies indicate that mitochondrial changes can cause anesthetic hypersensitivity.18
Our results are most consistent with mitochondrial function in gas-1 affecting anesthetic sensitivity via at least two mechanisms. Acute changes in respiration contribute to immobility as indicated by decreased oxidative phosphorylation and changes in ATP concentrations in the mutant. Overlaid on the changes in respiration are long-term changes in oxidative damage by reactive oxygen species. Future studies will be directed at identification of the proteins damaged by oxidation and understanding whether free radical production is increased in gas-1 compared with wild-type animals. These results should further our understanding of how a mutation in a mitochondrial protein can profoundly change behavior in volatile anesthetics.
The authors are indebted to Charles L. Hoppel, M.D. (Professor, Departments of Pharmacology and Medicine, Department of Veterans Affairs Medical Center, Case Western Reserve University, Cleveland, Ohio), for his ongoing support and discussions. The authors thank Judy Preston (Research Assistant, Department of Anesthesiology, Case Western Reserve University, Cleveland, Ohio) for excellent technical assistance. They also thank Pam Szweda (Research Associate, Department of Physiology and Biophysics, Case Western Reserve University) and Luke Szweda (Associate Professor, Department of Physiology and Biophysics, Case Western Reserve University) for the generous sharing of their antibodies and continued advice and support in these studies.