Early Development, Identification of Mode of Action, and Use of Dantrolene Sodium: The Role of Keith Ellis, Ph.D.

DANTROLENE was first synthesized by Snyder et al. in 1967.¹  Snyder worked in Norwich Eaton Laboratories—a small pharmaceutical company based in New York State with an interest in urology antibiotics. Dantrolene resulted from the simple insertion of a phenyl ring between the nitro and furan (a five-membered aromatic ring with four carbons and one oxygen molecule) components of nitrofurantoin (a hydantoin derivative, a five-membered heterocyclic structure with carbonyl groups at positions 2 and 4 and –NH groups at positions 3 and 5; fig. 1). The drug was fed to rats and mice that were noticed to be paralyzed the following morning, although breathing, pupillary reflexes, and blood pressure were normal. When left for several days in this state, their abdomens developed a waffling effect caused by the grill of the cage. It was realized a new type of muscle relaxant had been found. Analogs were prepared but were found to be no more or less effective than the original drug. Addition of a sodium ion (to make a salt) did not diminish activity. The mode of action of dantrolene was unknown, and Keith Ellis—a research scientist with a strong background in skeletal muscle physiology and pharmacology and expertise in single skeletal muscle cell recordings—was employed by Norwich Eaton in 1967 to elucidate this.

The drug was initially earmarked for use in spastic disorders. Spasticity results from an imbalance between inhibitory and excitatory input to motor nerves at the spinal level, and a number of drugs had been used in its treatment including mephensin,²  baclofen, benzodiazepines, quinolone, and carbonic esters. Oral or intraperitoneal administration of this new relaxant to mice, rats, cats, and dogs resulted in dose-dependent muscle relaxation that was long lasting, relatively nontoxic, and free of the adverse effects that plagued other agents used in the treatment of spasticity. The median lethal dose was high, and although large doses produced in-coordination, lower doses preserved coordinated muscle activity while still providing reductions in muscle tone.^3,4  The agent had no apparent anticonvulsant or local anesthetic activity and proved an ineffective analgesic in rats. Cardiopulmonary depression was minimal. A reduction in muscle tone remained “the most pronounced and consistent feature” of this new agent.³  Over a 3-yr period, Ellis et al. published a number of papers investigating the means by which dantrolene sodium produced this effect.^3–9  Centrally acting muscle relaxants work at the level of the cortex, brainstem, and/or spinal cord, while peripheral agents exert their effects at various sites on the motor nerve, neuromuscular junction (NMJ), muscle cell membrane (i.e., sarcolemma), or the contractile apparatus itself.¹⁰  Using a series of classic physiologic models, Ellis employed a methodical stepwise approach to isolate and examine each of these potential sites of action along the neuromuscular pathway from the brain to the myocyte. The models included Straub tail mice, and decerebrate rigidity, flexor-reflex, cross circulation, isolated motor nerve, NMJ, and denervated muscle preparations.^5–7 

Initial tests were employed to determine if dantrolene acted centrally or peripherally. The “Straub tail response” of mice is a characteristic rigid elevation of the tail after subcutaneous morphine administration and is dependent on the presence of an intact lumbosacral spinal cord and its associated motor outflow.^11,12  This response has been used for initial assessment of agents with muscle relaxant potential and as a spastic muscle model for evaluating skeletal muscle relaxants.¹³  The model is particularly useful because it allows assessment of centrally and peripherally mediated muscle relaxation in addition to sedation and motor coordination (via a rotarod test) in the one species.¹⁴  The rotarod test is a sensitive measure of motor coordination. Trained mice balance “lumbar-jack” style on a horizontally mounted rod rotating at various speeds. Muscle in-coordination was defined as an inability to stay on the rod for more than 30 s. Ellis used measures of these assessments to create a “muscle relaxant efficacy” (MRE) index, allowing comparison of dantrolene against known relaxants. According to Ellis and Carpenter,⁴  the more peripherally acting an agent, the higher its MRE index. The MRE values for various muscle relaxants including pentobarbitone, diazepam, mephensin, and tubocurarine were less than 1: dantrolene had a value of 1.7, indicating its site of action was distal to the other agents tested.

Further work supported this finding. IV dantrolene produced relaxation in decerebrate cats³  with rigid leg extension and catatonic contraction of the trunk: conclusive evidence the cortex was not the drug’s site of action. Dorsal root–ventral root preparations were also used, and all experienced a reduction in rigidity, which was abolished with high doses. Both of these experiments indicated the dantrolene effect was at a level lower than the cerebrum such as the spinal cord or even more distal.

Techniques for testing spinal cord monosynaptic and polysynaptic reflex responses had been well established.¹⁵  Monosynaptic reflexes were principally studied and assessed with variations in temperature. The anesthetized cat “flexor-reflex” model allows simultaneous assessment of drug effects on the spinal-based monosynaptic and polysynaptic reflexes that control muscle tone. Centrally acting agents such as mephensin (a membrane-stabilizing agent) and diazepam and baclofen (both γ-aminobutyric acid agonists) produce their effects by modulating (i.e., depressing excitatory or enhancing inhibitory) spinal neurons. Muscle relaxant effects are produced by inhibition of polysynaptic pathways in the spinal cord, but these effects did not occur with dantrolene, implying the drug had a peripheral rather than a central mode of action.^3,4  Cross-circulation studies in anesthetized dogs confirmed this view. In this model, the hind limb of the “recipient” dog was isolated neurally and vascularly with “trunk to limb” perfusion of the limb provided by a second “donor” dog. Contraction of the cranial tibial muscle of the isolated limb was produced via electrical stimulation of the cut motor nerves and the contractile response measured. This model separates the central and peripheral effects of muscle relaxants: agents administered to the recipient dog can reach the brain and spinal cord but not the isolated limb, while those administered to the donor are restricted to the isolated hind leg of the recipient. Dantrolene reduced the contractile response by 84% when administered to the donor dog but had no effect when administered to the recipient, providing conclusive evidence of a peripheral site of action.^3,4 

Ellis then turned his attention to the motor nerve. Conduction velocity—a sensitive measure of nerve function⁴ —did not change when isolated rat sciatic nerves were bathed in dantrolene-saturated Krebs solution, indicating a lack of any effect on neuronal conduction and narrowing dantrolene’s possible site(s) of action to the NMJ or the muscle cell itself. The NMJ and postjunctional muscle membrane were then examined. Dantrolene attenuated muscle contraction in response to sciatic nerve stimulation. When an equivalent dose of dantrolene was administered to both noncurarized and curarized muscle preparations, direct stimulation failed to show any difference. Dantrolene was also effective in denervated muscle. This eliminated the NMJ as the site of action of the drug. The effect of dantrolene on the postjunctional membrane was tested in a frog sartorial muscle preparation stimulated with “high potassium” Ringer’s solution, prepared by mole-for-mole replacement of the NaCl in Ringer’s solution (i.e., 6.5 g/l or 111 mM/l) with KCl. Increasing the potassium concentration of the extracellular fluid hypopolarizes the muscle cell and thus increasing excitability results. Dantrolene had no effect, indicating no change in the postjunctional membrane.⁴ 

Although preliminary studies had noted action potentials in muscle cells with dantrolene-induced reductions in twitch response, the nature of these action potentials had not been examined. Action potentials in dantrolene-treated muscle appeared normal, were propagated over the cell surface, and were of sufficient magnitude to elicit a contraction.⁴  Dantrolene had no effect on the rate of rise of the action potential—a sensitive measure of local anesthetic-induced membrane stabilization, while the failure of dantrolene to alter either membrane resistance or capacitance also suggested a normal flow of current across the cell membrane and into the T-tubules. Cells treated with dantrolene were electrically excitable but still exhibited significant reductions in contractility: Ellis deduced dantrolene’s site of action must lie a step beyond excitation of the cell membrane and the spread of this current into the T-tubules viz. at the level of the intracellular calcium-release mechanism.

Calcium was first thought to be involved in muscle contraction in the 1950s,¹⁶  but this was not proven until 1963 when Weber et al.¹⁷  demonstrated that very low concentrations of calcium were sufficient to activate muscle contraction.

Caffeine was known to produce muscle contracture. It can produce contractures in depolarized muscle, in muscles with disrupted tubules, and in media that is calcium free. Weber and Herz¹⁸  also demonstrated the mechanism in 1968, showing caffeine caused contractures by first releasing calcium from stores in the sarcoplasmic reticulum (SR) via terminal cisternae and by reducing uptake into the SR. Ellis used this effect to demonstrate dantrolene had its effect at the cellular level in 1971.⁵  The effect of dantrolene on caffeine stimulation is shown in figure 2. Caffeine sensitizes the calcium channel resulting in prolonged channel openings, but the precise mechanism by which this occurs is still unknown.¹⁹ 

Malignant hyperthermia (MH) was first described by Denborough in 1960, and its frequent autosomal dominant inheritance pattern was identified 2 yr later. As late as 1968, the basic underlying mechanism of MH was still unknown.²⁰  It was thought to be related to calcium metabolism, but this had not been demonstrated. A landmark paper was produced by Kalow et al. in 1970,²¹  indicating that sensitivity of MH muscle to both caffeine and halothane is probably caused by an increased release of calcium from SR stores. Up until this time, calcium release was found to be partially blocked by the local anesthetic procaine, most likely by decreasing channel opening, but the therapeutic effect of this in the treatment of MH was variable.

First coined by Topel et al. in 1968,²²  the term porcine stress syndrome (PSS) refers to a sudden death disorder in pigs of marketable weight subjected to various stresses including fighting, high ambient temperatures, and most commonly, transportation and handling before slaughter.^23–25  Rare occurrences of a similar condition had previously been reported in the 1950s and early 1960s in Germany, Denmark, and the Netherlands.^22,26–28  However, the syndrome rose to prominence in the United States in the 1960s after the producer-led drive to select for meatier pigs with better feed conversion, greater muscling, and leaner carcass weights—traits found in muscular European breeds such as the Pietrain and Landrace.^22,25,29,30  Pigs experiencing PSS typically die within minutes of the onset of clinical signs. Death is preceded by anxiety, tail twitching, muscular rigidity, tachycardia, dyspnea, blotchiness and cyanosis of the skin, and a rapid and uncontrollable increase in body temperature. Rigor mortis occurs almost immediately.^22,24 

Meat from PSS pigs stressed at slaughter is often pale, mushy, and watery—a condition known as pale, soft, and exudative (PSE) pork.^{22,24,25,29,30}  PSE pork is of inferior quality, has poor processing characteristics, and is unappealing to consumers. The problem had long plagued the pork industry with reports from German sausage manufacturers dating back to 1914; however, the rise of PSS-susceptible pigs caused a marked escalation in its occurrence, with huge costs to the U.S. pork industry.^25,28  The changes responsible for PSE pork occur in the period immediately after death and result from an accelerated glycolytic rate in the face of high carcass temperatures.^29,30  This leads to abnormal muscle tissue acidification and subsequent protein denaturation. The link between PSS and PSE pork was first described by Ludvigsen in 1954, but PSE pork can also occur in normal pigs subjected to severe stress.^22,25,29 

In 1966, Hall et al.³¹  described a fatal hyperthermia accompanied by muscle rigidity in three pigs from one litter after the use of halothane/suxamethonium. They recognized the probable genetic basis of this phenomenon and proposed the reaction might be related to similar fatal reactions reported in humans. This was confirmed as being almost identical to the human form of MH in 1968.³²  In the first International Symposium on Malignant Hyperthermia in 1971, Nelson et al.³³  proposed the syndromes PSS, PSE pork, and MH appeared to be a susceptibility to a loss of control of skeletal muscle metabolism and that the similarities could be useful for research. They proposed the three conditions were manifestations of a single genetically based myopathic syndrome.³³ 

Lauren Christian was an animal scientist and geneticist with an interest in pig production. A faculty member of the Animal Sciences Department at Iowa State University of Science and Technology (Ames, Iowa), who went on to become an internationally recognized expert in swine genetics, Christian was a coauthor of the paper that first gave name to the disorder PSS.²²  He was the first to propose an autosomal recessive mechanism of inheritance for the trait and was instrumental in developing “the halothane field test” for identification of PSS pigs.^34,35  The test involved exposure of weanling pigs to halothane—6% for 1 min followed by 2% for 2 min—delivered in 650 ml/min oxygen via a tightly fitting facemask. The pigs were watched closely for signs of pelvic and/or thoracic limb rigidity and the facemask removed immediately if this occurred to prevent the onset of a fulminant MH episode. Christian contributed to more than 15 publications investigating PSS, its mode of inheritance, and its contribution to the phenomenon of PSE pork—the postmortem manifestation of PSS. He also acted as a consultant for private and corporate pig production firms.

In about 1973, Keith Ellis received a general scientific bulletin across his desk containing a short article written by Lauren Christian. Ellis describes this in the following way, “The publication was in full colour, was really an advertisement with copies sent to anybody and everybody. Not only was there biological data in the publication but also mechanical equipment, testing materials etc” (personal communication via email on October 9, 2000, and November 10, 2016). This communication—perhaps about a paragraph long—outlined a syndrome of muscle rigidity and sudden death in certain pigs, triggered by a variety of causes. The article noted the economic implications of the syndrome and indicated that abnormal calcium release was the likely cause. Unfortunately, the bulletin has not been able to be traced. Dantrolene at this time was “a drug looking for a disease,” although it had been used for spastic disorders as indicated previously and offered some promise for the treatment of tetanus (K. Ellis, personal communication). Ellis immediately realized dantrolene may be effective in the management of this disorder. He did a literature search and wrote to five or six researchers interested in the problem of MH but only one—Gaisford Harrison in South Africa—replied. Samples were sent to South Africa. MH was induced in eight Landrace pigs and seven survived after treatment with IV dantrolene. The one that did not probably received too low a dose. The finding was published in the British Journal of Anaesthesia, but the author included only one tangential reference to Ellis’s involvement.³⁶ 

The drug still had to be introduced into anesthetic practice. The work of Dr. Ralph White with solubility and Dr. Mary Kolb with early clinical usage were prominent factors in the introduction of dantrolene.³⁷  Dr. Kolb was the first to organize clinical trials in 1976, proposing a skew-type sequential design. One hundred anesthetic departments were canvassed (personal communication), 65 institutions took part, and eventually trials were undertaken on 11 patients with successful treatment in all. The drug was approved for use in the management of MH by the Federal Drug Administration in August 1979. Dantrolene’s chemical composition is 1-{[5-(p-nitrophenyl)furfurylidene]amino}hydantoin. The name was derived from dan (hydantoin derivative), tro (nitrofurantoin), and olene, a muscle relaxant effect designated by the United States Adopted Names Council.

What is now known about how dantrolene works? Dantrolene binding to a specific N-terminal region of type-1 ryanodine receptor (RyR1) involving residues 590 to 609³⁸  reduces RyR1-dependent calcium flux in vesicle studies³⁹  and dihydropyridine receptor and RyR1 coupling in muscle cells.⁴⁰  Recent evidence indicates that the inhibitory effect of dantrolene on RyR1 channel activity requires calmodulin binding to RyR1.⁴¹  Additional studies suggest that dantrolene reduces RyR1-dependent calcium influx via store-operated calcium entry,⁴²  although this may not be due to a direct inhibition of store-operated calcium entry channels.⁴³  It is possible that other cellular factors (e.g., proteins, oxidative state, posttranslational modification) may also play an important role in facilitating dantrolene inhibition, although relative importance of these different factors is as yet undefined.

The current preparation of dantrolene (Dantrium sodium, Par Pharmaceutical Companies, Inc., USA.) contains 20 mg dantrolene as a lyophilized powder, 3 g mannitol, and sufficient sodium hydroxide to produce a pH of 9.5 in a 65-ml glass vial.⁴  Solubility may be prolonged, although a new North American preparation, Revonto⁴⁴  (U.S. World Meds, USA), can result in more rapid solubility, using tert-butyl alcohol (U.S. World Meds, personal communication). Both preparations however have cumbersome reconstitution and could potentially result in an excess fluid load.

Ryanodex⁴⁵  (Eagle Pharmaceuticals, USA) is a new, more rapidly prepared dantrolene formulation. Each Ryanodex vial contains 250 mg dantrolene sodium in lyophilized powder form that can be rapidly reconstituted as a uniform nanoparticle suspension (less than 1 min) using only 5 ml sterile water (without a bacteriostatic agent). It contains 125 mg mannitol and has a pH of 10.3. Expense however may limit its use.

All IV preparations of dantrolene may cause skeletal muscle weakness, gastrointestinal symptoms, thrombophlebitis, and if extravasated, tissue necrosis. The shelf life of Ryanodex is 2 yr and that of Dantrium and Revonto 3 yr.

Keith Osborne Ellis, Ph.D., was born in New York State. After earning a bachelor degree in biology from Heidelberg College (Tiffin, Ohio) and a doctorate in pharmacology from the University of Cincinnati (Cincinnati, Ohio), he went on to work in various postdoctorate positions before undertaking his dantrolene investigations at Norwich Eaton Laboratories. He was a colleague of Dr. Mary Kolb and played a minor role in the gaining of Federal Drug Administration approval of dantrolene, including determination of the effective dose. After his identification of dantrolene as a potentially effective treatment for MH, Ellis continued an association with other researchers in the field including Drs. Beverly Britt and Thomas Nelson. He continued with pharmacology-based research projects at Norwich Eaton Laboratories for some time, served on various university faculties in New York State, and later pursued a business career in pharmaceutical technology licensing and acquisitions. He was a member of a number of professional bodies including the American Society of Pharmacology and Experimental Therapeutics (Bethesda, Maryland), the Society for Experimental Biology (London, United Kingdom), and the Licensing Executives Society (Reston, Virginia).

Ellis’s vision has saved many lives over the past 37 yr^46–48 ; however, very little credit has been given to him over the decades and it is unlikely his name is known by many anesthetists. This article is an attempt to recognize the important role played by Keith Ellis in the discovery of this mainstay treatment for MH. Dr. Ellis is now retired; however, a video interview discussing his role in the Origins of Dantrium can be found at https://www.youtube.com/watch?v=h4CDcbKctfY to https://www.youtube.com/watch?v=4rFyEbIpkmc.

The authors gratefully acknowledge the advice of Robert Dirksen, Ph.D. (University of Rochester Medical Center, Rochester, New York); Dianne Daugherty, Malignant Hyperthermia Association of the United States (MHAUS; Sherburne, New York), for administrative assistance; Rose Hansen, librarian and administrative assistant, Midcentral Health, Palmerston North, New Zealand; MHAUS for access to a malignant hyperthermia video; and Keith Ellis, Ph.D. (Chestertown, New York), for providing personal information.

No funding received from National Institutes of Health, Wellcome Trust, Howard Hughes Medical Institute, or any departmental or institutional sources.

The authors declare no competing interests.