PARDOE, Ian, Stuart (Metropolitan House, 2 Salisbury RoadMoseley, Birmingham B13 8JS, GB)
1. Use of a cardiac glycoside and/or a diuretic in a solvent or carrier in the formulation of a pharmaceutical composition for the therapeutic treatment of neuropathic pain.
2. Use as claimed in Claim 1 , wherein the composition is for treatment of neuropathic pain associated with neuronal hyperexcitability.
3. Use as claimed in Claim 1 or Claim 2, wherein the diuretic is a loop diuretic, thiazide diuretic or sulphonyl urea.
4. Use as claimed in any preceding Claim, wherein the diuretic is furosemide.
5. Use as claimed in any preceding Claim, wherein the cardiac glycoside is selected from digoxin, digitoxin, ouabain, strophanthia or other known cardiac glycoside.
6. Use as claimed in any preceding Claim, wherein the cardiac glycoside is digoxin.
7. Use as claimed in Claim 4, wherein the concentration of furosemide is 0.5 mg/ml.
8. Use as claimed in Claim 6, wherein the concentration of digoxin is 0.05 μg/ml.
9. Use as claimed in any preceding Claim, wherein the carrier is a topical medicament which can pass through skin with lesions or skin absent such lesions.
10. Use as claimed in Claim 9, wherein the carrier is in a single phase formulation.
11. Use as claimed in Claim 10, wherein the carrier is in the form of a topical gel formulation.
12. Use as claimed in Claim 11 , wherein water is incorporated into the gel as a hydrogel.
13. Use as claimed in Claim 11 or Claim 12, wherein the gel comprises 40% water, 40% propylene glycol and 20% urea w/w.
14. Use as claimed in Claim 11 or Claim 12, wherein the gel comprises 50% propylene glycol, 40% ethanol and 10% water w/w.
15. Use as claimed in any preceding Claim, comprising 1 :14 digoxin:furosemide.
16. Use as claimed in any of Claims 1 to 10 wherein the solvent or carrier is at pH5..
17. Use as claimed in any of Claims 1 to 10, wherein the carrier is a pressure- sensitive adhesive.
18. Use as claimed in Claim 17, wherein the pressure-sensitive adhesive comprises part of a patch.
19. Use as claimed in any one of Claims 1 to 10, wherein the solvent comprises part of a lacquer or paint.
20. A method of making a gel-based pharmaceutical composition for the therapeutic treatment of neuropathic pain, comprising mixing 50% propylene glycol with 40%
ethanol and 10% water by weight, adding cardiac glycoside and/or diuretic in excess, forming a saturated solution and incorporating a gelling agent thereto.
21. A method as claimed in Claim 20, wherein the gelling agent is hydroxypropycellulose.
22. A method as claimed in Claim 20 or Claim 21 , including formulating the gel at pH5
23. Use of digoxin and furosemide in preparation of a composition for therapeutic treatment of neuropathic pain.
SE OF A CARDIAC GLYCOSIDE AND/OR A DIURETIC FOR THE TREATMENT OF
This invention relates to the treatment of pain, for example neuropathic pain. In particular, but not exclusively, it relates to the therapeutic treatment of neuropathic pain by the topical application of drugs.
Neuropathic pain is the result of an injury or malfunction in the peripheral or central nervous system. The pain is often triggered by an injury, but this injury may or may not involve actual damage to the nervous system. Nerves can be infiltrated or compressed by tumors, strangulated by scar tissue, or inflamed by infection. The pain frequently has burning, lancinating, or electric shock qualities. Persistent allodynia, pain resulting from a nonpainful stimulus such as a light touch, is also a common characteristic of neuropathic pain. The pain may persist for months or years beyond the apparent healing of any damaged tissues. In this setting, pain signals no longer represent an alarm about ongoing or impending injury, instead the alarm system itself is malfunctioning.
Examples of neuropathic pain include post herpetic (or post-shingles) neuralgia, reflex sympathetic dystrophy / causalgia (nerve trauma), components of cancer pain, phantom limb pain, entrapment neuropathy (e.g., carpal tunnel syndrome), and peripheral neuropathy (widespread nerve damage, most commonly caused by diabetes or chronic alcohol use).
The fundamental characteristic of neuropathic pain is the hypersensitivity of the pain detection mechanism. Actions, such as gentle touching or stroking of affected parts of the body, will result in a massively disproportionate level of pain being felt by the
patient. The exact mechanism by which this hypersensitivity develops is unknown, however the fundamental electrical mechanism by which pain is transmitted is the same in all cases- the sensitised nerve more readily fires its action potential than a normal axon, so consequently more pain messages will be received by the central nervous system. In some cases these damaged or sensitised neurons will fire spontaneously, sending painful messages to the brain in the absence of any direct stimulus.
The more detailed patho-physiology of neuropathic pain is not well understood. Neurophysiologic and neuroanatomic changes may occur in some types of neuropathic pain following injury to neural tissue. Injury to peripheral neural axons can result in abnormal nerve regeneration in the weeks to months following injury. The damaged axon may grow multiple nerve sprouts, some of which form neuromas. These nerve sprouts, including those forming neuromas, can generate spontaneous activity, which peaks in intensity several weeks after injury. Unlike normal axons, these structures are more sensitive to physical distention, which is clinically associated with tenderness and the appearance of Tinel's sign (i.e., sensation of tingling or "pins and needles"). After a period of time, atypical connections may develop between nerve sprouts or demyelinated axons in the region of the nerve damage, permitting "cross-talk" between somatic or sympathetic efferent nerves and nociceptors. This has been hypothesized as another mechanism sustaining a peripheral generator in some types of neuropathic pain. Dorsal root fibers may also sprout following injury to peripheral nerves.
With respect to pharmacological treatment of neuropathic pain, treatment can be difficult since most of the interventions used have side effects and may not work (for example, only two thirds of patients in controlled clinical trials typically respond, and those patients may only respond by a third). There are major concerns in the treatment
of patients with neuropathic pain because good treatment algorithms for treating a broad spectrum of neuropathic pains do not exist. Furthermore, current treatment guidelines are not written in terms of underlying patho-physiology, but rather by disease history (diabetes, herpes). It is difficult to predict whether a patient will respond, and consequently, many agents need to be tried before a patient can be managed appropriately.
Acute, inflammatory and neuropathic pain can be attenuated or abolished by local treatment with the sodium channel blocker lidocaine (or lignocaine). At present this is the only licensed topical treatment for neuropathic pain. It is thought to exert its effect by delivering amounts of lidocaine sufficient to block sodium channels on small damaged pain fibres but insufficient to interfere with normal conduction of impulses in larger sensory fibers. When applied to painful areas it is able to provide local analgesia. This may lead to the systemic absorption of drug, so some patients, who are taking other forms of treatment such as oral class 1 anti-arrhythmic drugs such as mexiletene may be unable to receive this form of treatment. Also adverse effects and allergic responses may occur, leading to rashes, redness and other adverse effects. Systemic treatments of neuropathic pain by drugs such as gabapentin or morphine often cause side effects such as sleepiness, nausea and more serious respiratory depression which could lead to death.
There is therefore a clinical need for other effective topical treatments of neuropathic pain.
Medications used to treat neuropathic pain have traditionally been categorized by their drug class (antidepressant, anticonvulsant, anti-arrhythmic, analgesic [opioid or nonsteroidal anti-inflammatory], topicals). Thus far, medication trials have not been
able to find different responses of symptoms to different drugs, except in postherpetic neuralgia (PHN) where patients with allodynia and no sensory loss responded to topical anesthetics, whereas patients with major sensory loss did not (Fields, Rowbotham, & Baron, 1998). A recent pilot study of a topical anesthetic in painful diabetic neuropathy (PDN), however, found comparable improvement with and without allodynia, so the literature is clearly still evolving (Barbano, et al., 2004). Although mechanistic stratification of neuromodulator medications is in its infancy, four categories have been suggested (Table 2), including voltage gated sodium channel modulators thought to be involved in peripheral sensitization, calcium flux modulators that operate at the level of the dorsal horn, drugs that enhance descending inhibition through actions on serotonin, norepinephrine, and opioid transmission (Beydoun & Backonja, 2003), and drugs that modulate central sensitization by their effects on NMDA receptors. No drugs that affect resting potential of nerves have been used as potential treatment options.
Mechanistic Stratification of Drugs Used to Treat Neuropathic Pain
Neuropathic pain is frequently chronic, and tends to have a less robust response to treatment with morphine or other systemic opiod drugs. Usually, neuropathic problems are not fully reversible so treatment is aimed at control of pain rather than a cure of the underlying condition.
All living cells, including cells of the human body, have an uneven distribution of ions between the inside and outside of the cell. The dominant ions outside cells (in an organ system) are positively charge sodium ions (Na+) and negatively charged chloride ions (Cl-).
The dominant ions inside a cell are the positively charged potassium ions (K+) and negatively charged proteins (Pr-). The concentration of potassium ions is as much as
20 times greater than the extracellular concentrations. Conversely, the extracellular fluid contains a concentration of sodium ions (Na + ) as much as 10 times greater than that within the cell.
The electrical gradient that exists between the inside and outside of a cell is called the resting membrane potential, and in human cells, the inside of the cell is negatively charged relative to the outside of the cell. The resting membrane potential in man is in the order of -7OmVoItS.
This imbalance is maintained by the active transport of ions across the membrane known as the sodium potassium pump. The sodium-potassium pump maintains this unequal concentration by actively transporting ions against their concentration gradients.
These concentration gradients are established by the active transport of both ions, and the same transporter, called the Na7K + ATPase (also known as the sodium pump), does both jobs. It uses the energy from the hydrolysis of ATP to actively transport 3 Na + ions out of the cell for each 2 K + ions pumped into the cell.
This accomplishes several vital functions:
• It helps establish the net negative charge across the plasma membrane with the interior of the cell being negatively charged with respect to the exterior.
• The accumulation of sodium ions outside of the cell draws water out of the cell and thus enables it to maintain osmotic balance (otherwise it would swell and burst from the inward diffusion of water).
• The gradient of sodium ions is harnessed to provide the energy to run several types of indirect pump.
The crucial roles of the Na + /K + ATPase are reflected in the fact that almost one-third of all the energy generated by the mitochondria in animal cells is used just to run this pump.
Nerve cells or axons are cells with a specific function of propagation of electrical messages. They have several important features which are responsible for their function;
When a sensory stimulus is received by a nerve cell there is a change in its resting potential- it converts the sensory input into an electrical signal, and in the case of a pain nerve, this will be a nociceptive stimulus.
Time , 'hi s
Changed polarity of the membrane, the action potential, results in propagation of the nerve impulse along the membrane. An action potential is a temporary reversal of the electrical potential along the membrane for a few milliseconds.
Sodium gates and potassium gates open in the membrane to allow their respective ions to cross. Sodium and potassium ions reverse positions by passing through membrane protein channel gates that can be opened or closed to control ion passage.
Sodium crosses first. At the height of the membrane potential reversal, potassium channels open to allow potassium ions to pass to the outside of the membrane. Potassium crosses second, resulting in changed ionic distributions, which must be reset by the continuously running sodium-potassium pump.
Eventually enough potassium ions pass to the outside to restore the membrane charges to those of the original resting potential. The cell begins then to pump the ions back to their original sides of the membrane.
The action potential begins at one spot on the membrane, but spreads to adjacent areas of the membrane, propagating the message along the length of the cell membrane. After passage of the action potential, there is a brief period, the refractory period, during which the membrane cannot be stimulated. This prevents the message from being transmitted backward along the membrane.
The steps in an Action Potential are:
1. At rest the outside of the membrane is more positive than the inside. 2. Sodium moves inside the cell causing an action potential, the influx of positive sodium ions makes the inside of the membrane more positive than the outside. This occurs at the voltage gated sodium channel.
3. Potassium ions flow out of the cell, restoring the resting potential net charges.
4. Sodium ions are pumped out of the cell and potassium ions are pumped into the cell, restoring the original distribution of ions.
The current treatment of neuropathic pain, which targets the peripheral sensitization of neurons, invariably targets the propagation of the action potential and more specifically the drugs are designed to aim at disruption of the voltage gated sodium channel.
It is an object of this invention to provide a composition for the treatment of neuropathic pain.
It is a further object of this invention to provide a composition which can affect the resting potential of a nerve cell.
Yet further objects of the invention relate to methods of treatment of neuropathic pain.
Accordingly, in a first aspect the invention provides the use of a cardiac glycoside and/or a diuretic in the manufacture of a composition for the treatment of pain, e.g. neuropathic pain.
A second aspect of the invention comprises a composition to treat pain, e.g. neuropathic pain, the composition comprising a cardiac glycoside and/or a loop diuretic and a carrier.
A further and/or more specific aspect of the invention provides a topical gel formulation for the treatment of pain, e.g. neuropathic pain, comprising at least one cardiac glycoside and/or diuretic in a gel carrier medium, said formulation being capable of transdermal delivery of the said diuretic and/or glycoside.
A yet further aspect of the invention provides a transdermal active principle delivery means comprising a skin adherent or skin-tolerant substrate applicable to a skin area, which substrate includes a composition for treating pain, e.g. neuropathic pain, comprising a transdermal^ effective carrier medium including at least one active principle selected from the group consisting of diuretics and/or cardiac glycosides.
A fifth aspect of the invention provides a method of treating pain, for example neuropathic pain, the method comprising applying a composition comprising one or both of at least one cardiac glycoside and/or at least one diuretic to a site on a patient.
A sixth aspect of the invention provides a method of affecting the resting potential of a nerve cell, the method comprising applying a composition comprising one or both of at least one cardiac glycoside and/or at least one diuretic to the nerve cell.
The neuropathic pain may be selected from post herpetic neuralgia, diabetic neurophathy, allodynia, phantom limb syndrome.
The diuretic may be selected from loop diuretics, thiazide diuretics or sulphonylureas.
Loop diuretics are substances which act on the ascending loop of Henle in the kidney. They are sulphonamides but may be other substances too. Typical examples include: acetazolamide mefruside ambuside methazolamide azosemide piretanide bumetanide torsemide butazolamide tripamide chloraminophenamide xipamide clofenamide clopamide ethacrynic acid clorexolone etozolin disulfamide ticrynafen ethoxzolamide furosemide
Preferably the loop diuretic is one or more of furosemide, bumetamide, ethacyrnic acid or torasemide.
Preferred is furosemide which is an anthrilic acid derivative, chemically 4-chloro-N- furfuryl-5-sulfamoylanthranilic acid. It is practically insoluble in water at neutral pH,
however is freely soluble in alkali. Furosemide exerts its physiological effect by inhibition of the transport of chloride ions across cell members. Furosemide is a loop diuretic with a short duration of action. It is used for treating oedema due to hepatic, renal, or cardiac failure and treating hypertension. The bioavailability of furosemide is between 60% to 70% and it is primarily excreted by filtration and secretion as unchanged drug. Furosemide acts on the Na+/K+/2CI- cotransformer. For its diuretic effect, its predominant action is in the ascending limb of the loop of Henle in the kidney. Loop diuretics markedly promote K + excretion, leaving cells depleted in intracellular potassium. This may lead to the most significant complication of long term systemic furosemide usage namely a lowered serum potassium. We postulate that it is this action however which makes furosemide a candidate for use as an agent against DNA viral infections.
Recent evidence suggests that the major biotransformation product of furosemide is a glucuronide. Furosemide is extensively bound to plasma proteins, mainly albumin. Plasma concentrations ranging from 1 to 400 mcg/ml are 91-99% bound in healthy individuals. The unbound fraction ranges between 2.3-4.1% at therapeutic concentrations. The terminal half life of furosemide is approximately 2 hours, and it is predominantly excreted in the urine.
Thiazide diuretics include the benzothiadriazines derivatives, also known as thiazides. Typical examples are: althiazide hydrobenzthiazide bemetizide hydrochlorothiazide bendroflumethiazide hydrofluoromethiazide benzthiazide indapamide benzylhydrochlorothiazide mebutizide buthiazide methylcyclothiazide chlorothiazide meticane chlorothalidone metalazone cyclopenthiazide paraflutizide
cyclothiazide polythiazide epithiazide quinethazone ethiazide teclothiazide fenquizone trichlormethiazide
Preferably the thiazide diuretic is one or more of chlorothiazide, hydrochlorothiazide, hydroflumethiazide, methyclothiazide, trichlormethazide, benzthiazide, bendroflumethazide, bendrofluazide, polythiazide or cyclothiazide.
Sulphonylureas are anti-diabetic drugs which influence ion transport across cell membranes. They are instanced by: acetohexamide glyburide
1 -butyl-3-metanilylurea glybuthiazole carbutamide glybuzole chlorpropamide glycycloamide glibenclamide glyclopyramide glibornuride glyhexamide gliclazide glymidine glimepiride glypinamide glipizide phenbutamide gliquidone tolazamide glisentide tolbutamide glisolamide tolcylamide glisoxepid
Preferably the sulphonylurea is one or more of tolbutamide, tolazamide, tolcyclamide, glibomuridum, acetohexamide, chlorpropamide, carbutamide, glyburide or glipizide .
By altering the cellular concentrations of ions, cellular ionic balances, cellular ionic milieu and cellular electrical potentials by the application of a cardiac glycoside and/or a diuretic or a sulphonylurea it is possible to change the metabolism of the cell without detriment to the cell. It has been shown that virus replication within such cells may be inhibited (see, for example, WO 01/49242). Anti-viral efficacy has been demonstrated
against the DNA viruses Herpes simplex virus type 1 and type 2, Feline Herpes virus, Cyclomegalovirus, Varicella zoster virus and Pseudorabies and Adenoviruses. The use of these species is equally of value in any other intracellular infection such as a bacterial infection as in Chlamydia. As stated above, we now propose that the same species may be used to treat pain, e.g. neuropathic pain, for example, by the alteration of the resting potential of a nerve cell.
The cardiac glycoside may be selected from digoxin, digitoxin, ovabain, strophanthia and other known cardiac glycosides. Digoxin is preferred
Digoxin is a cardiac glycoside obtained from the leaves of Digitalis lanata (Lond, 1994). It contains a steroid nucleus, with an unsaturated lactone essential for activity at the C17 position, and one or more glycoside residues at C3. The pharmacological activity of Digoxin is related to its ability to specifically bind to a site on the extracytoplasmic face of the α subunit of the enzyme Na+, K+ ATPase (regulates the concentration of sodium and potassium inside cells). This selectively inhibits the cellular active transport Na+ and K+ pump, which leads to an increase in the intracellular concentrations of sodium and calcium, and a decrease in the concentration of intracellular potassium (Henderson Morley, 2003). It is currently indicated in heart failure and supraventricular arrthymias. One main limitation to its clinical use is the risk of side effects such as nausea and vomiting associated with toxicity as it possesses a narrow therapeutic window.
One preferred application of this invention will be for the topical treatment of pain associated with post herpetic neuralgia, and acute varicella zoster infection.
Varicella zoster virus (VZV) is an alphaherpesviruses that is characterised by short growth cycle and rapid cell-to-cell spread resulting in cytocidal infection in a wide variety of cells/tissues. Following primary infection, VZV infection establishes a latent infection in sensory ganglia that persists for the lifetime of the host. A variety of stimuli can induce this latent virus to reactivate, travel back down the axons and produce a new round of productive infection at the site of initial infection. VZV is the causative agent of two human diseases: chicken pox following primary infection and herpes zoster or shingles following reactivation from a latent infection in sensory ganglia of affected dermatomes.
Primary and recurrent infection with VZV is frequently associated with pain and other sensory alterations. Herpes zoster is caused by VZV reactivation whereby the virus reactivates in the dorsal root ganglion and moves along the sensory nerve to the periphery. This results in a localised, painful, vesicular rash that can involve adjacent dermatomes. It is associated with more protracted and severe pain than other herpes virus infections during both the prodrome and clinical phase that may last as long as 3 weeks.
Postherpetic neuralgia (PHN) is the most common complication of herpes zoster. This has been defined as severe pain occurring one month after the onset of the typical blister like rash or that persists for greater than three months after the initial disease.
PHN is classed as a neuropathic pain that is associated with mechanical allodynia, where normally non-painful touch sensations are perceived as painful. In addition warm and cold allodynia and spontaneous pain have also been reported. PHN can result in impaired physical activity, disturbed sleep, social withdrawal and depression which many patients suffer for years. This pain state may be so severe that it has a major impact on quality of life and has been associated with suicide. A uniformly effective
treatment for PHN is not yet available. Current antiviral agents such as valaciclovir or famciclovir have been shown to reduce pain resulting from VZV providing treatment starts early after onset of the rash. Compounds such as anticonvulsants, tricyclic antidepressants, opioids and topical NSAIDs have had some efficacy in the treatment of PHN; however, there are a significant number of cases which remain intractable to current therapies. Therefore, there is a need for improved therapies that prevent or at least reduce, preferably substantially reduce PHN.
The main risk factor for the development of PHN is age, e.g. it has been reported that the incidence of PHN was 3-4% in the 30-49 age group but rose to 29% in those 70-79 and 34% in the over 80-year-old age group. The overall incidence of PHN in the population is difficult to determine; it has also been suggested that there are approximately 500,000 cases of PHN in the USA at any one time and it has further been suggested that a figure of 200,000 cases occurring at any one time in the UK was a conservative estimate.
The pathology of clinical zoster (shingles) is characterised by inflammation and severe damage to the nerves (haemorrhagic necrosis) of the ganglia with accompanying degeneration of motor and sensory roots. This suggests that some immune responses play a part in the development of zoster pain. The pathology of PHN is complex and not well understood. The involvement of the immune response in PHN has also been hypothesised, ongoing inflammation has been observed in post-mortem studies of some PHN patients and Gilden has been proposed that PHN may involve the persistence of VZV at levels above those normally seen during latency accompanied by continued inflammation. This point is important with regard to this invention, as antiviral effects of digoxin and furosemide have been demonstrated against VZV.
PHN is somewhat resistant to opioid and NSAIDs, which are the most commonly used treatments for acute and inflammatory pain states, suggesting that inflammatory responses per se are not sufficient to induce PHN.
Neuropathic pain (such as PHN) is characterised by neuronal hyperexcitability in damaged areas of the nervous system. This hyperexcitability is due to molecular changes (e.g. abnormal expression of sodium channels, changes in gaminobutyric acid (GABA) inhibition) at the level of the peripheral nociceptor, in the dorsal root ganglia (DRG), dorsal horn and brain. How latent VZV infection interacts with the neurone to induce such molecular changes is unclear.
Whilst we do not wish to be bound by any theory, we believe that the application of a cardiac glycoside, a diuretic or one or more drugs from each group in combination is an effective treatment for PHN for at least 2 reasons;
1. Replicating VZV has been implicated in the establishment and propagation of PHN. Topically applied cardiac glycosides, loop diuretics or a combination of both, have been demonstrated to inhibit VZV infection. Therefore topically applied drug will prevent the development of chronic VZV infection. 2. By altering the resting potential generated within a peripheral nerve, the perpetuation and propagation of hyperexcitability associated with VZV infection will be prevented and/or treated.
In order to demonstrate the efficacy of compositions of the invention, reference is made to the following Examples:
Example 1 :
Furosemide at a concentration of 1.0 mg/ml was very well tolerated by MRC5 cells in vitro; there was no adverse effect on cell morphology and cells replicated. Furosemide inhibited VZV plaque formation by 50% at this concentration.
Furosemide ID 50; 1.0 mg/ml. [see Table 1]
VZV replication was completely inhibited by Furosemide at a concentration of 2.0 mg/ml.
Digoxin at a concentration of 0.05 μg/ml was very well tolerated by MRC5 cells; there was no adverse effect on cell morphology and cells replicated. Digoxin inhibited VZV plaque formation by 50% at this concentration.
Digoxin ID 50; 0.05 μg /ml. [see Table 1]
VZV replication was completely inhibited by Digoxin at a concentration of 0.1 μg/ml.
Example 3: VZV replication was completely inhibited by Furosemide and Digoxin in combination at their individual ID 50 concentrations [see Table 1]. The combined dosage was equally well tolerated by MRC5 cells; there was no adverse effect on cell morphology and cells replicated.
The effects of Furosemide and Digoxin, individually and in combination, on Varicella Zoster virus replication in vitro NB. There was a ten-fold difference between adjacent multiplicities if infection (MOI)
* TNTC too numerous to count.
1 Furosemide 50% Plaque Inhibitory Dose [ID 50] 0.5 mg/ml.
2 Furosemide completely inhibited VZV at a concentration of 2.0 mg/ml. 3Digoxin 50% Plaque Inhibitory Dose ID 50; 0.05 μg /ml.
4 Digoxin completely inhibited VZV replication at a concentration of 0.1 μg/ml.
5 VZV replication was completely inhibited by Furosemide and Digoxin in combination at their individual ID 50 concentrations .
Comparative Example 1
Uninfected MRC5 cells replicated to normal yields in the presence of Furosemide at a concentration of 1.0 mg/ml, the same concentration as the VZV ID50.
Comparative Example 2
Uninfected MRC5 cells replicated to normal yields in the presence of Digoxin at a concentration of 0.05 μg/ml, the same concentration as the VZV ID50.
Comparative Example 3 Uninfected MRC5 cells replicated, though not to normal yields, in the presence of both Furosemide and Digoxin at their VZV ID50 concentrations. At these concentrations, VZV replication was completely inhibited.
Comparative Example 4
The effects of Furosemide and Digoxin on MRC5 cell metabolism were measured using the MTT assay. There were normal levels of metabolism in uninfected cells incubated with either Furosemide or Digoxin at their VZV ID50 concentrations. There was normal metabolism in uninfected cells incubated with both Furosemide and Digoxin at their VZV ID50 concentrations.
Viruses are intracellular parasites wholly dependent upon the infected host cell for survival. The earliest known anti-viral drugs were cytotoxic drugs, such as those used in cancer chemotherapy. It was postulated that these inhibitors of host cell metabolism would have an adverse effect upon the lifecycle of the virus. In the non-cancer patient cytoxic drugs were too toxic to the host to be of benefit, and as a consequence their use as anti-viral treatments was severely restricted.
The above results show efficacy against VZV.
In order to treat pain, for example neuropathic pain and, in particular, post herpetic neuralgia for example, a topical gel formulation may be applied to the body.
It is widely recognised that the stratum corneum of human skin is the rate limiting step in terms percutaneous absorption from topical applied agents. The stratum corneum in skin areas populated with viral warts is significantly different from that of 'regular' skin having as it does densely packed corneocytes with limited lipid intracellular pathways. This results in a stratum corneum with potentially even greater drug absorption rate limiting capabilities.
Whilst lesions are substantially absent in PHN we have sought to provide a topical medicament which can pass through skin with such lesions as well as skin absent such lesions. The rationale being that the passage of a medicament through lesioned skin is inhibited and will represent an arduous route for a medicament.
When formulating topical formulations, one approach is to chemically modify the barrier properties (making the barrier less efficient) by the incorporation of non- pharmacologically active (and 'skin friendly') excipients.
As was the case with each different type of anti-HPV formulations, the primary aim was to formulate topical formulations that can deliver the or ideally both of a cardiac glycoside (e.g. Digoxin) and a diuretic (e.g. Furosemide) (that, when applied together, act synergistically) to the site of action at efficacious amounts with minimal toxicity.
In combination, these two preferred drugs act synergistically and a primary aim is to provide topical gel formulations which is able to deliver them together in beneficial proportions. An effective way of achieving this is to employ a 'homogenous system' i.e. a single phase formulation. Furthermore, structured matrices (e.g. gels) provide better active release and minimal excipient intereaction. The occlusive nature of a gel also assists in coating the treatment site more effectively.
There are a wide range of topical formulation excipients available however the preferred ones selected took account of the following factors:
1 Degree of keratolytic activity of regular excipients. 2 Effects on 'Swelling' of keratin.
3 Relative Solubilites of the actives in the solvents bearing in mind both actives are relatively insoluble in H 2 O (to keep thermodynamic activity
high at relatively low concentrations thus maintaining efficacy and maintaining minimal toxicity).
UREA (BP, EP USP) Urea is commonly used in emollient formulations and has been shown to increase hydration of the stratum corneum at concentrations of 2 - 20% becoming keratolvtic at 20%.
In some preferred embodiments Urea is incorporated into the gel at 20% therefore providing hydration of the stratum corneum (additional effects on swelling) whilst simultaneously exerting mild keratolytic effects.
By hydrating the pores this may also facilitate movement of the active through openings in the SC. Another advantage of incorporating this excepient is that it is a natural product of metabolism and is excreted in the urine without systemic toxicity.
PROPYLENE GLYCOL (PG) (BP), EP USA)
PG is one of the most commonly used formulation excipients available and is frequently used at relatively low to medium concentration. The main reason for its use in topical formulations is excellent properties as a solvation agent. PG also increases water content in the stratum corneum encouraging an osmotic gradient through the stratum corneum.
If PG is used at concentrations of 40-70% it is acts as a keratolytic agent. PG is minimally absorbed and any systemic absorbance is oxidised in the liver to lactic acid pyruvic acid.
STERILE H 2 O (BP, EP, USP)
Swelling (and associated hydration of the tissue) is deemed highly preferred to permit optimal delivery of the two actives through the otherwise densely packed corneocytes/keratin to target zone the basal layer.
Ethanol (ETOH) (BP, EP, USP)
Ethanol is frequently used in topical formulation predominantly to improve drug solubility within the formulation. When exposed to the skin it causes dehydration, lipid mobilisation and potentially 'skin cracking' at relatively high concentrations.
PEG Average Molecular Weight 400 (BP, EP, USP)
Polyethylene glycol 400, also known as liquid macrogol, is a water miscible vehicle, co- solvent and a humectant and may withdraw moisture from the skin. It is a clear liquid with a slight alcoholic odour having a density of 1.13 g cm "3 at 20 0 C (water =1 ), and a viscosity of 43 Cs at 40 0 C.
PEG 400 is also known in topical formulations to improve drug solubility within the formulation. There is no scientific evidence that suggests PEG400 exhibits keratolytic activity, but there is some evidence to suggest it aids hydration of the skin.
Absorption Experiment 1
The swelling effect (i.e. absorption) of excipients ethanol (EtOH), polyethylene glycol (PEG), propylene glycol (PG) and water (H 2 O) on human callous skin was determined by submersing pre-weighed skin samples in those solvents for 24 hours. The skin samples were then removed, surface solvent gently removed with tissue, and re- weighed. Results show that only water was absorbed into the skin to any great extent. For this reason water is preferably incorporated into the gel, in the form of a hydrogel.
It was also found that in polyethylene glycol/water solutions, an increase in PEG resulted in a decrease in the swelling of callous skin. Incorporation of significant PEG into a gel is therefore less preferred.
Absorption Experiment 2
Propylene glycol (PG)/water solutions were made and show that PG had no undesirable effects on the swelling effects of water. There was no significant difference in swelling between a 100% water solution and a 50:50 solution of PG and water. Incorporation of 40% PG into a gel could therefore provide keratolytic action without affecting skin swelling.
Absorption Experiment 3
Literature indicates that a 20% urea solution exhibits keratolytic properties, including urea into a gel formulation is preferable for some embodiments (urea also makes gels/creams feel less greasy). By maintaining propylene glycol at 40%, excipients mixes were prepared with varying amounts of urea (up to 20%) to determine any effects on swelling. Results showed that urea had no adverse effect on the swelling (amount of uptake of water) of callous skin to any significant extent. The incorporation of 20% urea into the gel is therefore preferred in some embodiments, keeping it at a known (yet minimal) keratolytic concentration.
These results show that propylene glycol had no effects on the swelling which appears to be solely due to water. Urea has also been shown to have no undesirable effect on uptake. These results suggest that gels containing in the region of 40% water, 40% propylene glycol and 20% urea represent some preferred embodiments with dual keratolytic propertes of propylene glycol and urea, together with optimum swelling properties.
When developing topical dosage forms and prior to conducting any in vitro release or permeation experiments, it is important to carry out pre-formulation studies to acquire information on the physical and chemical properties of the prospective drug and solvent candidates. The solubility of the permeant in both the vehicle and in the various phases present in the skin, among other factors, determines its relative rate of penetration.
Typically, when applied to the skin the active or permeant will be present either dissolved or dispersed in a solvent or vehicle. Although the concentration of permeant present within the skin generally controls the rate of transport, that particular concentration is dependent on the solubility (i.e. thermodynamic activity) of the permeant present in the vehicle present on the surface of the skin.
Example 4 Relative solubilities of Digoxin and Furosemide permeant drugs were evaluated in a series of regular solvents/excipients. The solubilities were investigated at, 32°C, the average temperature at the surface of the skin. The results were then used to determine preferred permeant solvent combination for in vitro release and permeation to arrive at preferred gel formulations delivering optimum amounts of the 2 actives.
Using a metal spatula, small amounts of each model permeant were separately placed into 1.5ml vials (approximately 50mg. occupying ~ % volume of vial). The vials were then individually transferred onto a previously calibrated electronic balance and subsequently tared. Precisely 1g (weight/volume w/v) of each solvent was then carefully added to the vial using a calibrated 1000μl Gilson pipette, as the mass became closer to required 1g required, the 1000μl Gilson pipette was exchanged for a
200μl and 20μl Gilson pipette to enable accurate displacement of smaller volumes of the solvents. This procedure was carefully conducted to account for the different densities exhibited by the various solvents being studied.
Given that the solubility example was conducted at, 32°C, before solvent was added to the vial, all solvents were temperature-equilibrated at their designated temperatures in a thermostatically controlled incubator with a digital temperature reading (this was validated by placing a thermometer inside the incubator alongside the solvents). Subsequent weighing of the solvent into the respective vials containing the permeants was also carried out on an electronic balance at the appropriate study temperatures.
Negative control vials were present, containing each of the solvents with no drug added. For each permeant/solvent combination, a total of four replicates where carried out.
Once all solvents had been accurately weighed, the vials were secured into a blood cell rotator. Again this procedure was carried out at the predetermined temperature. An incubator was used to maintain a constant temperature of 32°C. The thermometer readings were periodically noted to ensure a stable consistent temperature was being maintained.
On commencement of the saturated solubility examples, the vials were periodically examined to ensure there was excess drug present within the solvent, i.e. visible solid particles or a suspension rather than a clear solution. If any vial visually showed no particles present then additional amounts of permeant where added until complete saturation was achieved for that specific permeant solvent combination was achieved. The vials were allowed to rotate for a total period of 24 hours.
Following a final visual inspection to ascertain presence of excess solid and the assumption that the equilibrium had been attained, the vials were immediately transferred to a centrifuge and spun at 12,500rpm for 10 minutes. The centrifuge had been temperature pre-equilibrated at 32 0 C. By centrifuging the vials, separation of saturated permeant solvent from excess permeant was achieved, with the excess solid forming a pellet at the bottom of the vial.
750μl (approximately 75%) of the supernatant was then transferred into another vial (pre warmed), and spun for a second time in a centrifuge at 12,500 rpm for 10 minutes. This ensured that any excess particulate matter from the pellet that could have occurred when removing the supernatant using the Gilson pipette was not present. Great care and attention was taken to ensure all equipment used including pipettes and tips were equilibrated at the corresponding temperature prior to their use to avoid any changes in permeant solubility with potential small changes in temperature. The vials were then sampled using pre-warmed pipette tips and analysed immediately by HPLC.
Absorption Experiment 4
From the results gained in the initial solubility examples of each of the permeants in each individual solvent, an additional solubility example was conducted using the same procedure as described previously, except using the following co-solvent mixes. Prototype Co-solvent mix: 40:40:20 Propylene Glycol: Water: Urea
(Maximum Keratolytic Potential) Revised Co-solvent mix 1 : 50:40:10 PG: EtOH: Water
(Maximum Digoxin Potential) Revised Co-solvent mix 2: 50:20:20:10 PG: EtOH :PEG400: Water
(Maximum Fru /Optimum Digoxin)
Figures 1 to 5 show the results obtained from each of the solubility tests. Figures 6 and 7 highlight comparisons. The solubility test results shown in Figures 1 to 3 indicated that both actives Furosemide (F) and Digoxin (D) were relatively insoluble in water, but their solubility increased in the presence of propylene glycol (PG). The presence of urea reduced solubility of both actives F and D.
From the solubility studies of the permeants in each individual solvent Digoxin was shown to have the highest solubility in ethanol and propylene glycol with relatively low solubility in PEG400 and water. Furosemide was shown to have significantly highest solubility in PEG400, relatively good solubility in PG and ethanol and low solubility in water.
The highest solubility of Digoxin in the presence of Furosemide in the 40:40:20 (PG: Water: Urea) co-solvent mix (prototype) was 433 and 5979 μg ml "1 respectively.
It was hypothesised that although the 40:40:20 (PG: Water: Urea) co-solvent mix that had significantly greater 'keratolytic' potential than the 50:40:10 (PG: EtOH: Water) co- solvent mix and still appeared to enable delivery of the actives through the skin tissue, (supported by in-vitro permeation studies) the chemical potential within the formulation may have been insufficient to diffuse to the location of the virus in efficacious amounts. It was also hypothesised that this may have been exacerbated by water in the formulation and hydrated lesions which may present an additional barrier to the ingress of the lipophilic Digoxin and Furosemide.
From the solubility studies pertaining to the later revised gel co-solvent mixes (50:40:10 PG: EtOH: Water and 50:20:20:10 PG: EtOH: PEG: Water) it was apparent that the co- solvent mix specifically designed to maximise the concentration primarily of Digoxin
(50:40:10) also maximised the concentration of Furosemide (more than the gel formulation that contained PEG400). On this basis only the 'Maximum Digoxin' revised gel was subjected to in vitro release testing.
In the Maximum Digoxin revised co-solvent mix the solubility of Digoxin increased from 433 μg ml "1 to 6267μg ml "1 (~ 15 fold). The solubility of Furosemide also increased from 5979 μg ml "1 to 95001 μg ml "1 (~ 16 fold). This unexpected increase in penetrant solubility was attributed to several factors:
incorporation of 40% ethanol. removal of urea from the system, reduction in the H 2 O content to 8%,
Loss of urea from the formulation would be expected to reduce keratolysis within the revised gel formulation. However, to compensate for this factor, the PG content was increased to 50%.
One means of optimising the delivery of permeants into the skin from dermatological formulations is to improve the thermodynamic activity or 'chemical potential' of the permeant in the delivery system. Optimum release and subsequent delivery of the permeant into and across the skin can usually be obtained when the thermodynamic activity is at its highest achievable level, i.e. at saturation (=1 ). Being above (drug crystallisation effects) and below (lower chemical potential) the saturation level results in reduction in release from the formulation and subsequent delivery across the skin. Therefore, achieving thermodynamic activity level 1 in the topical formulation is desirable to optimising delivery of the permeants into the skin. The respective levels of Digoxin and Furosemide required to achieve saturation in the various co-solvent mixes was investigated in the solubility tests. Other formulation factors must also be
considered in addition to the thermodynamic for e.g. Furosemide is amphiphillic in chemical nature and so formulation pH effects could affect performance.
Example 5 To investigate phenomena such as this, and others, batches of several different types of gel were formulated to probe the effect these parameters may have on the in vitro release characteristics of Digoxin and Furosemide.
The effects of different types of gel thickeners for example carbomers such as Carbopol 981NF Carbopol and ULTREZ (presently used in commercially available dermatological preparation due to excellent gelling, stability, and low skin toxicity characteristics). The cellulose derivative, Hydroxypropylcellulose was also investigated as a suitable thickening agent.
In addition to the types of thickener (and the respective am required to gain a 'desirable' viscosity) the effects of modifying the pH of the gel to pH 5, 7 and 9 was investigated.
Finally, the effect of altering the molar ratio of the permeants to a 1 :1 and 1 :14 molar ratio of Digoxin: Furosemide were investigated. The rationale behind these choices is discussed herein;
Rationale 1: Gel formulation containing a saturated solution of Digoxin and Furosemide in the 40:40:20 co-solvents mix at a 1 :1 molar ratio
This was achieved by formulating saturated solutions of Digoxin and Furosemide separately in the gel co-solvent mix (40:40:20 Water: Propylene Gycol: Urea) and then
combining them at pre-determined volumes such that equimolar amounts of the two drugs were contained in the solution.
The saturation limit of Digoxin in the 40:40:20 (PG: Wata: Urea) co-solvents mix was 433 μg cm "3 (see saturated solubility results section). The number of moles in solution is 433/780.9 = 0.55 μmol cm "3 . The saturation limit of Furosemide in co-solvents = 5979 μg cm "3 . The number of moles in solution is moles 5979/330.7 = 18.1 μmol cm "3 . Therefore, for a 1 :1 molar ratio we require 18.1/0.55 = 32.9 cm "3 of a saturated Digoxin solution to 1cm "3 of the equivalent Furosemide solution. By combining saturated solutions of Digoxin and Furosemide at a volumetric ratio of 33:1 respectively, a solution containing both in a 1 :1 molar ratio should result. (This is only true however if no drugs drop from solution on mixing - experimentally this was found to be the case). Both drugs in this case are at saturation and thus highest thermodynamic activity.
Rationale 2: Gel formulation containing a combined saturated solution of Digoxin and Furosemide in the 40:40:20 co-solvents mix, Molar Ratio 1:14.
The basis for this rationale was to obtain maximal amounts of both drugs in the mix independently and with no control of stoichiometry. Essentially, the presence of excess of both drugs would establish equilibrium at saturation for each component in the presence of the other.
When excess amounts of each drug were added to the co-solvent mix subsequent HPLC analysis resulted in a respective 1 :14 Digoxin: Furosemide molar ratio. This molar ratio was achieved due to the relative chemical potentials of each permeant in each others presence within the co-solvent mix. Although the molar ratio is in some ways surprisingly different to rationale 1 , again each permeant will be at saturation and
thus its highest level of thermodynamic activity. Furthermore, if Digoxin is the more potent of the two actives, excess Furosemide may be beneficial at the site of action sodium/potassium pump). It was hypothesised that pH variation would have minimal effect on the release of Digoxin, as it is a neutral molecule.
The following table details the different types of gels manufactured that were subsequently subjected to in vitro release testing and the proportion of each ingredient used (w/w).
Empirical formulae of each of the gels.
Example 6 Gel Preparation
All calculations were based on the relative proportions of the different formulation co- solvent mixes. In each different gel formulation, the thickener was added last as a percentage of the total w/w formulation mix.
The following sequence of events describes the preparation of the 'revised maximum Digoxin' gel. All other gels were also manufactured using the same sequence of events except using the appropriate components as described in the previous table. All gel manufacture was carried out to GLP standards. pH adjustment was carried out following addition of the thickener by the addition of appropriate amounts of NaOH or acetic acid. For those gels that contained Urea, the predetermined amounts of Urea was added following the combination of each co-solvent and before the permeants were added.
Into a clean one litre glass beaker, the 'Max Digoxin' gel formulation was prepared by firstly combining 25Og of Propylene glycol with 20Og of Ethanol and 5Og of Water (to constitute the 50:40:10 ratio). The amounts were weighed accurately on a pre- calibrated electronic balance and continually mixed with the use of a magnetic stirrer. Excess amounts of Furosemide and Digoxin were added to the co-solvent mix, to ensure a saturated solution of both Digoxin and Furosemide was obtained. The beaker was immediately sealed using para-film (to stop any solvent evaporation) and left continually stirring at room temperature over night. The resulting suspension was then centrifuged at 25,000rpm for 20 minutes to separate excess Furosemide and Digoxin drug, from the resultant saturated co-solvent mix. The resultant saturated solution was transferred into another clean one litre beaker, which had been previously placed on an electronic balance, with care taken in ensuring no excess permeant was also transferred, and the total weight of the saturated solution recorded. Using an electronic balance, HPC (hydroxypropycellulose) was subsequently weighed that corresponded to 8.0% w/w of the total saturated solution (w/w).
Whilst the saturated co-solvent mix was agitating vigorously (with the use of an overhead stirrer), with the use of a spatula, HFC was slowly added over a period of five
minutes. Following visual inspection to confirm that the HFC had been fully dispersed, the container containing the gel formulation was placed onto a blood tube rotator and left overnight to allow the HPC gel network to form. A resultant homogenous batch gel formulation was obtained. This procedure was repeated, at a smaller scale, to obtain a gel formulation consisting of 50:20:20:10 Propylene Glycol: Ethanol: PEG 400: Water co-solvent mix.
To formulate a gel from the Digoxin and Furosemide ampules, equal amounts of the Digoxin ampule (i.e. digoxin for IV injection 62.5 mcg/mu) and the Furosemide ampule (i.e. Furosemide for IV injection 20 mg/mt) were accurately weighed onto an electronic balance. When combining the ampules there is a reduction in the total ethanol content (as only the Digoxin ampule contains ethanol). To compensate for this (to ensure Digoxin does not precipitate out when combined with Furosemide) for each 2g of Digoxin that was combined with 2g of Furosemide 200μl of ethanol was also added. Again 8% w/w of HPC was added and the solution was left to form a gel following the same procedures as described above.
In Vitro Release In vitro release testing is a commonly used technique to examine the performance of topical drug formulations. It is a basic requirement as dictated by regulatory bodies such as the FDA and SUPACS. We utilised a recently developed improved in vitro model that eliminated some pitfalls in previously used methods.
Diffusion experiments were performed using all glass Franz-type cells (nominal receptor phase volume, 3ml). The membranes were soaked in the receptor medium (appropriate co-solvent mix) for 24 hours prior to commencement. The membranes
were then taken out of the receptor fluid, the surface dried and carefully placed onto the pre-greased flange of the receptor compartment. The donor chamber was then placed onto the corresponding receptor compartment and pinch clamped in position. To each receptor compartment a micro stirrer was added. The effectiveness of this technique for PBS and the more viscous PEG400 was previously validated by applying small aliquots of dye to receptor compartments filled with solution.
An infinite dose was added by adding the formulations into each of the appropriate donor caps until the amount of gel added reached the top of the donor chamber. Pre- greased glass cover slips were then placed onto each of the donor chambers to form an occlusive airtight seal. The receptor compartments were then filled with the appropriate degassed co-solvent mix (either 40:40:23 Water: Propylene Glycol: Urea, 50:40:10 PG: ETOH: Water or 10% EtOH/Buffered Solution - salt concentration identical to that present in the ampules) and the cells placed on a multiple stirrer plate in a thermostatically controlled water bath, where the temperature at the surface of the membrane was maintained at 32 0 C. 200μl samples were collected at 1 , 2, 4, 6, 12 and 24 hrs and replaced with the appropriate temperature-equilibrated co-solvent mix. A total of 6 replicates were carried out for each treatment.
A Dual HPLC-UV analytical method for separation and detection of the 2 activities was developed. HPLC analysis was performed using a Hewlett Packard 1100 HPLC automated system fitted with a Phenomenex Kingsorb 5μm C18 Column (250 x 4.6mm). The mobile phase consisted of 40:30:30 Water: MeOH: MECN. The UV detector was set to 200nm and a 20μl injection volume was used. The flow rate was 1ml min "1 , the run time was 10 minutes and the retention time of furosemide was
typically 3.2 minutes and Digoxin 5.4 minutes. Standard calibration curves were constructed from standard solutions (range 0.1 , 1 , 10, 20, 40, 80 and 100μg ml "1 ) that contained the relative same proportions of the solvents or co-solvent mixes. A Chromatograph to illustrate the dual HPLC assay for the separation and detection of F 5 (first peak observed at 2.6 minutes) and D (second peak observed at 5.2 minutes) was obtained. This Chromatograph was obtained following an injection of a combined standard solution that contained D and F (both at a concentration of 50 μg ml "1 ).
Results are indicated in Figures 8 to 14. 0
Summary of the in vitro Release data produced from each gel (taken from Figures 8 through 11).
Summary of in vitro Release data produced from each gel (taken from Figures 12 through 14).
Data Processing Q24: This value is the total concentration of active per unit area (μg cm "2 ) that has been released into the receptor phase in 24 hours.
Flux: This is concentration of active that has been release per hour per unit area (μg cm h-1 ). Calculated by dividing Q24 by 24.
Release Rate: This is the gradient of the line taken from the Square Root of Time (hours) plot (μg cm-2 hθ.5). The R2 value is also reported indicating line rarity.
% Applied Dose The amount of active present in the receptor phase calculated as a percentage of the dose applied. Assuming the donor phase contains fixed 2g of gel
It has become a matter of routine that in vitro studies that investigate percutaneous permeation of topical creams, gels, ointments through human skin typically uses a phosphate buffered saline (PBS) receptor phase. PBS receptor phases are designed
to provide similar physiological solute conditions to those within and underlying the dermis. Many other different types of receptor phases have also been employed during in vitro studies, with one individual study utilising four different receptor phases being HBS, HEPES buffer (1.5mM CaCI 2 ), HEPES buffer (10% Bovine Calf Serum) and trizma base 45mM (4OmM sodium cholate).
However, the main purpose of in vitro release studies is to gain quantitative data on the release of the active from the topical formulation.
Both gels exemplified used the same improved in vitro release system, in which the co- solvents contained within the topical formulation and the contents of the receptor phase were both the 40:40:20 Water: Propylene Glycol: Urea gel mix. By employing the 40:40:20 gel mix as the receptor phase instead of the traditional phosphate buffered saline (PBS) co-elution of excipients in the formulation are minimal allowing only net translocation of Furosemide and Digoxin from the donor. Also soaking the membranes in this mix prior to cell assembly ensured that the only net migration would be the actives.
The maximum concentrations of Furosemide and Digoxin observed in the receptor phase were 543.60μg ml "1 , 42.47μg ml "1 respectively significantly lower than both the saturated solubilities (determined from earlier solubility studies). Therefore sink conditions were maintained throughout the experiments..
When plotted as the square of root time (hours) the release profiles obtained for Furosemide and Digoxin show excellent linear ties. R 2 values calculated were; 0.9993,
0.9974, 0.9993 for Furosemide at pH 5, 7 and 9 from the 1 :14 gel, 0.9836, 09961 ,
0.9742 for Digoxin at pH 5, 7, and 9 from the 1 :14 gel, 0.9844 and 0.9981 for
Furosemide and Digoxin respectively from the 1:1 gel. This suggests that the nylon release membrane chosen exerted minimal/zero rate limiting effects for both formulation.
Differences in release Characteristics with changing pH
For both Furosemide and Digoxin, as pH of the gel increased, there was a rank order reduction in the release rates. Furosemide at pH 5 had a release rate of 124.51 μg crn ' V 5 , yet at pH 9 this had been reduced to 43.389 μg cm "2 h 05 almost a 3 fold reduction. The reduction in the rate of release was more substantial for Digoxin at pH 5 10.18 μg cm "2 h 05 was release but was reduced to 086 μg cm "2 h α5 at pH 9 12 fold reduction.
Given that Digoxin is a neutral permeant the observations of reduced release with increasing pH was attributed to physiochemical properties of the gel (given that thickening characteristics of the gel can be affected via changes in pH). Furosemide is chemically amphiphillic which may explain the less significant changes in the rates of release with changes in pH.
These results would support formulating the gel at pH 5 to obtain optimum simultaneous release characteristics for both permeants.
Differences in release characteristics between gels of different permeant stoichiometrv. The in vitro Release Tests were extended to also investigate the differences between two gels that differed by the relative stoichiometry of Furosemide to Digoxin. The first gel was formulated as described in 'Rationale V having a 1 :1 molar ratio of Digoxin: Furosemide. The second gel was formulated according to 'Rationale 2' result in a 1 :14
molar ratio of Digoxin: Furosemide. Both gels were formulated at a pH of 7. Significant differences in release of both actives from the gels was apparent.
The 1 :14 Gel resulted in release rates of 77.482 μg cm "2 h 0 5 for Furosemide and 430 μg cm "2 h 05 . However, release rates produced the 1:1 Gel were significantly lower for Furosemide at 7.18 μg cm "2 h α5 and also lower at 3.48 μg cm '2 h 05 for Digoxin. This is a respective 10.8 fold and 1.2 fold reduction. This observation was attributed to the applied doses of the two formulations.
At 24hrs the amount of Furosemide and Digoxin release from both revised gel formulations (ULTREZ and HPC) was significantly higher than the corresponding amounts released from the Ampule gel formulation. The revised formulations, at 24hrs showed similar amounts being released in terms of Digoxin, however the HPC thickened gel formulation released almost 1.5 times more Furosemide than the Ultrez formulation.
When observing the Digoxin release rates for the 3 gels, the Ultrez formulation showed the highest rate of release with 274 μg crτϊ 2 h α5 , followed by a similar rate of release from the HPC formulation with 245 μg cm "2 h α5 , however significantly lower rates of release were gained from the Ampule gel formulation with 8.1 μg cm "2 h 05 -
In terms of rates of release of Furosemide from the 3 gels, the highest rate of release was observed from the HPC formulation with 2673 μg cm "2 h α5 followed by a significantly lower rate of release produced from the ULTREZ formulation with 1527μg cm "2 h α5 , again significantly lower rates of release were gained from the Ampule formulation.
In terms of the overall release characteristics exhibited by each of the gels, given that the rates of release for Digoxin from the HPC and Ultrez thickened formulation were similar however Furosemide rates of release were significantly higher in for the HPC gel, overall the HPC thickened gel performed better than the Ultrez gel. The performance of the ampule gel formulation was significantly poor when compared to both revised gel formulations.
It can be concluded from these results that a most preferred formulation for the simultaneous co-release of Furosemide and Digoxin from these formulations would be the 1:14 Digoxin: Furosemide at a pH of 5 in the 50:40:10 co-solvent mix thickened with 8% HPC.
Due to the differences of the co-solvent mixes in terms of their relative keratolytic activity, it is important to conduct in vitro permeation experiments to determine which formulation delivers the most amounts of Digoxin and Furosemide through human skin.
In Vitro Skin Permeation Example 9
Obtaining in vitro permeation data using diffusion cells is a commonly used technique to examine the performance of topical drug formulations.
Permeation experiments were carried out using all glass Franz-type cells incorporating plantar skin between a donor and receptor phase. Skin cuttings of approximately 0.6cm diameter were required to fit the customised 'Franz' cells used. Consequently, an in vitro permeation study was conducted that used 5 cells per gel treatment, with no more than a total of 15 cells being run simultaneously. To each receptor compartment a micro stirrer was added. An infinite dose of gel (~1g) was added to each donor
compartment via syringe and PEG400, being used as a receptor phase in each, thereby enabling adequate 'sink' conditions, limiting hydrodynamic boundary effects and tissue swelling effects. The applied gel was occluded with glass cover slips and the cells stirred in a thermostatically controlled water bath, maintaining the skin temperature at 32 0 C. 200μl samples were collected at 1 , 3, 6, 12, 24, 36, 48, 72, 96, 120 and 144hrs (6 days total) and replaced with temperature-equilibrated receptor phase. A total of five replicates were carried out for each gel. All samples were immediately analysed by HPLC.
Initial in vitro permeation trials were conducted on the following prototype gel:
1) Prototype Gel
1 :14 (Digoxin: Furosemide) in 40:40:20 (PG: Water: Urea) pH 5 (1.5% Carbomer - 98 1 NF) Subsequent in vitro permeation trials were conducted on the following revised gels:
2) Maximum Digoxin Gel
1 :14 (Digoxin: Furosemide) in 50:40:10 (PG: EtOH: Water) pH 5 (8% HPC)
3) Maximum Furosemide (optimum Digoxin gel)
1 :14 (Digoxin: Fursoemide) in 50:20:20:10 (PG: EtOH: PEG400: Water) pH 5 (8% HPC)
4) Digoxin and Furosemide Ampule Gel
2g of solution taken from the Dioxin ampule was combined with 2g of solution taken from the Furosemide ampule and 200μl of ethanol was added. Again, 8% w/w of HPC was added to thicken the gel.
The steady state flux of both Digoxin and Furosemide for each gel was calculated from the gradient of the linear section of the cumulative permeation profiles. The breakthrough time was taken from the point at which permeant was first detected in the receptor phase (dependent on the MDL for each permeant). Lag times were calculated by extrapolating the linear section of the cumulative amount permeated (steady state flux) to the X-axis (time).
Results (Prototype Gel) Figures 16 to 20 illustrate the permeation profiles for Furosemide and Digoxin through callous plantar skin from an infinite dose of gel.
Table 5 Summary of the Data taken from the 40:40:20 Prototype gel.
Conclusions (Prototype Gel)
It was established in an earlier study that at a pH of 5.5 the formulation displays the greatest release rates for Furosemide and Digoxin. This study assessed the capacity of the formulation to deliver both drugs via permeation through skin.
The breakthrough time for Digoxin was approx 36 hrs and for Furosemide approx 6hrs. It is hypothesised that these differences reflect several things:
1) Relatively higher concentration (hence driving force) of Furosemide in the formulation
2) Lower limit of detection for Digoxin (88ng) relative to furosemide (3ng) due to the presence of a weaker chromophore in former
3) Significant Differences in permeant molecular weight (Furosemide 330.7 Digoxin 780.9) 4) Barrier properties of the skin
In terms of the physical properties of the skin it is apparent that this study utilised relatively thick human callous plantar skin with an average thickness of 0.92mm. As seen in the previous 'Swelling Study', tissue used in this study again was shown to increase in mass, probably due to hydration and resultant swelling of the keratin within the plantar skin. The average % increase in weight was shown to be 43%.
Overall, Furosemide exhibited a typical permeation profile through the plantar skin (Figure 16). First order kinetics were observed between 120 and 312 hours from which a steady state of flux of 13.83 ± 1.18 μg cm "2 hr "1 was calculated. For Digoxin a typical permeation profile was also obtained. Steady state was attained at 144 hours to which a flux of 1.14 ± 0.28 μg cm "2 hr '1 was calculated (Figure 17). The relative steady state flux for Furosemide and Digoxin may not be entirely due to the physical characteristics of the permeants but also the relative keratolytic effects of the formulation.
Figure 19 illustrates the molar permeation of both drugs. It can be seen that the ratio of Furosemide and Digoxin (approx. 10:1) does not differ to any great extent from the molar ratio of the drugs found in the formulation (14:1). The higher concentration of Furosemide in the gel is also mirrored in the relative percentages of drug permeated (Figure 18). The greater propensity for Furosemide to permeate is expected, given Furosemide has a significantly lower molecular weight relative to Digoxin. This is reflected in the calculated apparent steady state flux values (Table 5). However, this
effect may not be solely due to the physical properties of the permeants but may also relate to the relative (and perhaps temporary) pore sizes created by the keratolytic effects of the formulation. The specific diameter/3-dimensional shape of the created pores may allow easier permeation of Furosemide rather than Digoxin relative to the permeant molecular weight.
It can be concluded from these results that a gel formulation containing water, propylene glycol and urea (40:40:20) thickened with 1.5% Carbopol 981 NF at pH 5 will successfully deliver Furosemide and Digoxin through skin. These results suggest that topical application of such a gel would be beneficial in treating neuropathic pain in the skin or superficial soft tissue.
Results from this study provide compelling evidence that Furosemide and Digoxin contained within this specific gel formulation can be delivered through human plantar skin and useful for the treatment of skin conditions associated with neuropathic pain.
Results - Revised Gels
Figures 21 and 22 illustrate the cumulative permeation profiles for both Furosemide and Digoxin respectively, through callous plantar skin from about 1g of each of the three different gel formulations.
Figures 23 and 24 provide a direct comparison between the cumulative permeation of both Digoxin and Furosemide from the prototype gel and revised maximum Digoxin gel.
Table 6a Summary of the Permeation data for the 'Max' Digoxin Gel Formulation
Table 6b Summary of the Permeation data for the 'Max' Furosemide Gel Formulation
Table 6c Summary of the Permeation data for the Dig/Fur Ampule Gel Formulation
Conclusions (Revised Gels)
This study assessed the ability of three (revised) gel formulations to simultaneously deliver both Furosemide and Digoxin through human callous skin. To highlight potential differences in the ability of the various gels to deliver the actives through the skin, a comparison of the permeation data produced by the revised gel formulations, relative to the original prototype gel formulation was conducted.
From the three formulations studied, greatest delivery of both Furosemide and Digoxin through human skin was observed from the 'Max' Digoxin gel formulation, followed by the Max Furosemide gel, with lowest Furosemide and Digoxin delivery observed from the Digoxin/Furosemide Ampule gel formulation. See Figures 21 and 22.
When comparing the amounts of Digoxin and Furosemide delivered from the 'Max' Digoxin gel versus the prototype gel, significant differences can be observed. For the 'M' Digoxin gel, at 144 hours 620μg cm "2 of Furosemide and 101μg cm '2 of Digoxin permeated. However, for the prototype gel, at I44 hours only 574μg cm "2 of Furosemide and 30μg cm "2 of Digoxin permeated a 1.1 and 3.36 fold reduction respectively.
Additionally, for the original gel prototype, the breakthrough time for Digoxin was approx 36 hours and Furosemide approximately six hours. However, the breakthrough times gained from the 'Max' Digoxin gel formulation were significantly quicker, at one hour for both Digoxin and Furosemide.
It is hypothesised that these differences reflect several things:
1) Relatively higher concentration (hence driving force), primarily of Digoxin and also Furosemide present in the formulation 2) Lower limit of detection for Digoxin (880ng) relative to Furosemide (3ng), due to the presence of a weaker chromophore in former
3) Significant Differences in permeant molecular weight (Furosemide 330.7 Digoxin 7809)
4) Baffler properties of the callous membrane
In terms of the physical properties of the skin, it is apparent that this study utilised relatively thick human callous plantar skin, indicating that skin subject to, or in the region of underlying tissue subject to, neuropathic pain would be readily permeated.
Furosemide and Digoxin exhibited 'typical' permeation profiles through the callous plantar skin (Figure 2). For the 'Max' Digoxin gel formulation, first order kinetics were observed for Furosemide between 72 and 144 hours, and for Digoxin between 48 and
120 hours from which a steady state of flux of 6.759 and 1.015 μg cm "2 hr "1 accordingly was calculated.
It can be concluded from these results that a gel formulation containing 50:40:10 Propylene Glycol: Ethanol: Water thickened with 8% HPC enables simultaneous delivery of Furosemide and Digoxin through human callous skin. The permeation data produced from the revised 'Max' Digoxin gel formulation highlight significant improvements in the relative permeation characteristics for both permeants, when compared to the original prototype formulation. These results suggest that topical application of such a gel would be beneficial in controlling neuropathic pain.
Results provide compelling evidence that Furosemide and Digoxin contained within this specific gel formulation can be delivered through human skin, even human callous plantar skin and be useful for the treatment of PHN and other neuropathic pains.
A further formulation route which was considered was the use of an active in, for example, a glue.
Examples 10 to 12 are included by way of illustration to show the effects including synergistic effects of compositions comprising Digoxin and Furosemide against cells infected with HSV virus. It should be emphasised here that such examples are not however demonstrating transdermal^ effective delivery means entirely within the scope of the invention, but are nonetheless useful indicators of efficacy.
Bioassays with herpes simplex virus in vitro were undertaken to follow the anti-viral activity of the simultaneous administration of furosemide (1mg/ml) and digoxin (30 mcg/ml). Culture and assay methods follow those described by Lennette and Schmidt (1979) for herpes simplex virus and Vero cells with minor modifications.
Herpes simplex strains used:
Type 1 herpes simplex strain HFEM is a derivative of the Rockerfeller strain HF (Wildy 1955), and Type 2 herpes simplex strain 3345, a penile isolate (Skinner et a! 1977) were used as prototype strains. These prototypes were stored at -80 0 C until needed.
African Green Monkey kidney cells (vero) were obtained from the National Institute of Biological Standards and Control UK and were used as the only cell line for all experiments in the examples.
Cells and viruses were maintained on Glasgows modified medium supplemented with 10% foetal bovine serum.
Inhibition of HSV1
Multiplicity of Effect of Effect of digoxin Effect of infection furosemide alone furosemide
(dose of alone and digoxin in virus) combination
Medium + + ++++
Low + ++ ++++
This example demonstrates that virus activity was almost eliminated by applying low concentrations of the stock furosemide and glycoside solution to Vero cells infected with HSVl At higher concentrations virus activity was completely prevented. The antiviral effects of this stock solution were far greater than the effects of Furosemide or Digoxin alone. There was no direct virucidal activity on extracellular virus.
These experiments were repeated using a hsv2 strain, and almost identical results were obtained.
The method of Example 10 was repeated using type 1 herpes virus strain kos. Similar results were obtained.
In vitro bioassays were undertaken to follow the anti-viral activity of furosemide and digoxin when applied both simultaneously and alone.
The compositions were applied to different types of Vero cells (African green monkey kidney cells and BHK1 cells) and infected with type 2 herpes simplex virus (strains 3345 and 180) at low, intermediate, and high multiplicities of infection (MOI). Inhibition of virus replication was scored on the scale:
no inhibition 20% inhibition +
40% inhibition ++
60% inhibition +++
80% inhibition ++++
100% inhibition +++++
T denotes drug toxicity.
The following results were obtained using African green monkey kidney cells and type 2 herpes simplex strain 3345:
Agents ref: P5159PCT
The greatest effect of Digoxin alone (+++) occurred on application of 30 mcg/ml Digoxin at low multiplicity of infection only.
The greatest effect of Furosemide alone(+++) occurred on application of 1 mg/ml Furosemide at low and intermediate multiplicities of infection.
When the loop diuretic and cardiac glycoside were simultaneously applied to the infected cells, the greatest effect (+++++) was achieved using Dioxin at 30 mcg/ml and Furosemide at 1 mg/ml. 100% inhibition of hsv2 replication was shown at low, intermediate and high multiplicities of infection.
Similar results were, obtained using other combinations of Vero cells and type 2 herpes simplex strains.
This example demonstrates that replication of HSV-2 is not maximally inhibited by applying Furosemide or Digoxin alone. However, in combination Furosemide and Digoxin completely inhibited HSV2 replication.
This example demonstrates the in vitro release and permeation of Digoxin and Furosemide from transdermal delivery devices. Delivery systems were evaluated as formulations for this application in the presence and absence of additional excipients to aid both release and penetration. Acrylic polymer-based glues are the most common in currently marketed products and appear the best formulation, thus three glues were utilised.
Digoxin and Furosemide were purchased from Sigma, UK. Durotak acrylic glues were sourced from National Starch and Chemical Company. Duro-tak 87-900A (Glue 1), Duro-tak 87-2052 (Glue 2) and 87-201 A (Glue 3) were used. All solvents and chemicals used for the release and permeability were purchased from Sigma. The silicone sheeting that was used as a synthetic skin barrier was purchased from Advanced Biotechnologies, USA.
Methods Formulation and in vitro evaluation of a transdermal patech for the delivery of Digoxin and Furosemide is outlined below.
Development of an HPLC Method for Digoxin and Furosemide
For effective therapy the drug(s) must initially be released from a formulation prior to penetration of the skin; in each case the amount of drug release or the rate of penetration will need to be quantified. HPLC offers a reliable means of quantifying the amount of drug that has been released. There are several published methods that detail HPLC analysis of both drugs. The HPLC used was Agilent Series 1100 with a Phenomenex C18 (150 x 4.60 mm 5μ micro) column. The mobile phase was water, methanol and acetonitrile (40:30:30) and flowed at 1 ml/min. 20 μl of sample was injected and detected at 220 nm with a variable wavelength detector (VWD).
Figure 25 shows a calibration curve of Digoxin concentration according to the HPLC method used.
The HPLC was not able to detect Digoxin released from Glue 3 indicating that the Digoxin is preferentially bound within this glue.
Glue 1 showed the most favourable release with both drugs releasing at a rapid rate. It was considered that the profile of release indicated that all drug was released over the three day period thus an increased loading of drug within this glue would lead to increased drug release.
Figure 26 shows a calibration curve of Furosemide concentration according to the HPLC method used.
Example 14 - Manufacture of the Delivery Device Acrylic based pressure sensitive adhesives were sourced from National Starch and Chemical Company with properties that would be appropriate for use with Digoxin and Furosemide. A study was performed that measure the solubility of the drugs in a range of solvents.
Solvent Solubility of Digoxin (mg/m£) Solubility of Furosemide (mg/mC) Ethanol 5O8 10.15
Methanol 8.2 15.3
Ethyl acetate 20.4 35.6
After mixing the dissolved drug in solvent with glue; a film of 400 μm thickness was cast onto the backing membrane (Scotchpak 1109). This was left uncovered (yet protected from light) for the solvent to evaporate at room temperature for a period of approximately 45 minutes. Once sufficiently dry (approximately 45 minutes) the exposed surface was covered with liner (Stotchpak 1020) to prevent further solvent loss. All materials were cut to a measured size and stored in an airtight container at room temperature. Each patch of known weight had a known drug content, in this case a high loading per surface area is required.
Solvents used in conjunction with drug included, ethylacetate, methanol, ethanol, propanol and combining the dry drug powder with the glue directly.
Example 15 - Measurement of drug release from formulated patches Drug release studies were performed as a screening exercise prior to penetration studies. A circular patch of 1 cm diameter of the formulation was taken and placed into a sealed container containing an excess of release medium (2mC). The vial was sealed and shaken at a controlled speed and temperature (37 0 C) for a period of 48 hours. At set time points; 1 , 2, 4, 6, 8, 12, 24 and 48 hours a sample (0.5ml) was removed for analysis. Each time a sample was removed it was. replaced with fresh release medium to maintain an overall volume of 2m£ HPLC analysis of each sample allowed drug release over time to be plotted. The formulations were compared to note those that demonstrate the best release. In the clinical setting the patch will be approximately 0.25cm 2 and the release required is 25μg per 24 hours thus the release rate must be greater than 100μg/cm 2 /24hours.
Figure 27 shows the release of both drugs from Glue 1 (87900A);
Figure 28 shows the release of both drugs from Glue 2 (872677);
Figure 29 shows the release of both drugs from Glue 3 (87201 A);
Figures 30 to 34 shows an HPLC trace of the drugs release from the film in the solvent described releasing into a buffer solution as described.
A comparison of the graphs (Figures 35 and 36) above show that the drugs are released better when they are formed using methanol to dissolve the drugs rather than propylene glycol.
Example 16 - Measurement of drug permeation from formulated patches
The pressure sensitive adhesive incorporating the drug that demonstrates the greatest release was selected and the penetration into skin was evaluated. Franz cell apparatus was used to measure the penetration of the drug from the adhesive formulation into the skin membrane.
In the Franz cell, the upper layer represents the transdermal formulation and the lower layer the skin. The vessel below the skin is filled with fluid (the same as used in the release study) and stirred at a constant rate. At designated time intervals a sample from the lower vessel is taken using the side port and analysed using HPLC for drug content. The permeation of drug across the membrane over time can thus be calculated.
The membrane used in this study was a synthetic silicone based skin membrane purchased from Advanced Biotechnologies, USA.
Data from the penetration example suggests that the drug does penetrate the synthetic membrane.
Example 17 - Digoxin and Furosemide composition
The drug powders were mixed at a 1 :1 ratio and 500mg of this mix was blended with 10m€ of Glue 1. This mixture was then cast onto 3M Scotchpak 1020 release liner over an area of 80 by 120 mm. The solvents were left to evaporate and the film was covered with 3M Scotchpak 1109 polyester film laminate backing.
The drug loading is there 2.6mg/cm 2 of both drugs within the formulation.
The surface area of the 1cm diameter patches is 0.785cm 2 .
Each small patch contains 1.02mg of Digoxin and 1.02mg of Furosemide.
Examples 18 et seq
The high desirability of >1 dosage form for Digoxin and Furosemide to address the widely varying anatomical locations of the HPV infection was investigated, proposed variances included:
Plantar warts: drug-in-glue plaster-type application
Hand/finger warts: lacquer/paint
The aim of these later examples is to show both the feasibility of drug-in-glue formulations based on transdermal adhesive and the feasibility of lacquer/paint formulations based upon flexible collodion BP.
Example 18 - Materials
Digoxin (D) batch number 181104 and Furosemide (F) batch number 114310 were obtained from BUFA Pharmaceutical Products bv (Vitgeest, Netherlands). Centrimide lot no. A012633401 was obtained from Acros Organics (New Jersey, USA). Duro-tak®
387-2287 adhesive was a gift from National Starch and Chemical (Zutphen,
Netherlands). Flexible Collodion BP was obtained from JM Loveridge pic
(Southampton, UK). HPLC grade acetonitrile, ethanol and methanol were obtained from Fisher Scientific (Loughborough, UK). Pig ears were obtained from a local abattoir, prior to steam cleaning. Water was drawn from an ELGA laboratory still.
Example 18 - Drug-in-adhesive formulations
The ratios of F: D selected mix were 1 :1 , 1:25 and 1 :100, thus providing a sizeable excess of D. This was based on anecdotal evidence which suggested that D has substantially greater virostatic power than F, indicating that a formulation that delivered an excess of D may be more effective in reducing viral load and thus be effective in treating PHN. The effect each ratio had on the release of D and F is illustrated and ratios investigated which may produce optimum release of each active.
A drug-in-adhesive formulation is a type of matrix system in which drug and excipients can be dissolved or dispersed depending on the amount of drug required for the desired delivery profile (Venkatramann and Gale, 1998). As the solvent in the adhesive evaporates to form a solid matrix product, the concept of thermodynamic activity does not apply. However, the solvent is an important component as it creates microchannels in the matrix upon drying, to form a 'pathway' for the drugs to the skin. Generally, the limiting factor in the amount of drug that can be incorporated is the point at which bioadhesive properties are lost.
Preliminary work was performed to refine the composition of the model patches and the method of preparation. A loading dose of 0.5g of drug mix to 5g of adhesive was found to be optimum because further addition of drug mix decreased the adhesive properties of the patches. The drug mix was directly added to the adhesive, although 2.5ml of methanol was added to the mixture in order to decrease viscosity and aid casting out of the patches.
From a practical standpoint the method used to case out the patches was based on trial and error. It was consequently determined that to achieve a constant patch thickness, it was preferable to pour the drug-adhesive mixture onto a polymer-lined
paper in a horizontal line and then hold the paper vertically allowing the mixture to flow down the paper. This method was found to be reproducible and the drug-in-adhesive covered a surface area of approximately 8cm 2 with a depth measured to be almost exactly 1mm.
Example 19 - Preparation of drug-in-adhesive patches
Patches were prepared by the direct addition of 0.5g of drug mix, to 5g of adhesive (wet weight). Three drug mixes were prepared containing different molar ratios of F: D, the compositions of the drug mixes are displayed in Table 2.1. The appropriate amounts of drug mix and adhesive wee accurately weighed directly into glass vials using an analytical balance and 2.5ml of methanol was added to the mixture. Each vial was vortex-mixed for three minutes and left to rotate on a blood serum rotator overnight, ensuring that the drug mixture was homogeneously dispersed. Control patches were also prepared by the same method, containing no drug mix. Each adhesive mixture was then cast out onto polymer-lined paper as described above. The patches were covered and left for 48 hours to allow the solvent to evaporate (Chedgzy et al 2001 ). Clear polyethylene film was then attached to the exposed side of the patch to act as patch backing. Individual spherical patches were excised using a cork borer with a diameter of 1 cm (approximately 0.785cm 2 ).
Table 7 Composition of F and D in 0.5 g drug mix - used to prepare patches
Ratio of F : D Mass of F (g) Mass of D (g)
1 :1 0.14885 0.35115
1 :25 0.0084 0.4916
1 :100 0.0021 0.4979
Example 20 - Receptor phase
The function of a receptor phase is to provide an efficient sink for the released or permeated drug. A rule of thumb is that the amount of drug should not exceed 10% of its solubility in a given sink. Furthermore, the sink must not interfere with the release or permeation process (Heard et al, 2002). Two receptor phases wee considered in this work. These were aqueous cetrimide 30 mg ml, an ionic surfactant and EtOH/water 10:90 v/v, chosen as both drugs wee known to be freely soluble in each medium.
Stock solutions of each were prepared in a volumetric flask and degassed by drawing through a 0.45 membrane before use. However, it was subsequently found that cetrimide interfered significantly with the HPLC analysis and for the rest of this work EtOH/water 20:90 v/v was used as a receptor phase.
Diffusional release of D and F mix from Example 19 patches The aim of this example was to determine whether or not different molar ratios of the two drugs would affect the extent and rate of the release of each drug. The polymer- lined paper was prized from the patches to expose one side of the patch. Each patch was then individually immobilised to the bottom of a general 7ml glass screw cap vial with a small daub of Duro-tak® 387-2287 adhesive to the polymer film and allowed to dry for 30 minutes. The dissolution media used were cetrimide 30mg mt 1 or EtOH/water 10:90 v/v, 5 mt of each was added individually to each vial. The vials were then placed on a Stuart Scientific Gyro-Rocker (Fisher, UK) set at 70rpm to ensure adequate mixing of the dissolution medium and incubated at 32 0 C (the temperature of the skin) in a laboratory incubator (Genlab). At time points of 1 , 3, 6, 12 and 24hr, (expected period of application) O.δmt of dissolution medium was sampled and placed in HPLC auto sampler vials. After each sample was taken, the receptor phase was replenished with 0.5mf of stock dissolution medium also at 32 0 C. The samples were
refrigerated at 2 - 4 0 C until HPLC analysis 24 hrs later. A total of 3 replicates were performed for each treatment in each receptor phase. The formulation that demonstrated the optimum release was used during permeation examples.
Rationale for membrane selection
To investigate novel topical formulations for treating pain, for example neuropathic pain, the delivery of across human callous skin would be the most appropriate in vitro model (that is it is believed to provide the most substantial barrier to the passage of actives). However, such material was not available and so an appropriate model was required. The use of pig skin as a suitable substitute has been demonstrated in several works, with the ear being the part that provides the closest permeability characteristics to human skin (Dick and Scott, 1992; Simon and Maibach, 2000). Permeation experiments were used to study this dermatological drug delivery system, because permeation can predict localisation (percutaneous absorption in the basal layer) the greater the flux, the greater the permeation through the stratum corneum including keratinocytes, which are of greater number in warts than healthy skin. Wart lesions are relatively more keratinised compared to 'normal' skin. However, determination of permeation across normal skin could be predictive of permeation through warts, particularly in a screening mode. This is justified as there is some evidence that keratin in skin plays an important part in determining rates of skin permeation (Hashiguchi et al, 1998; Heard et al, 2003).
Freshly slaughtered pigs are routinely subjected to sterilisation by steam cleaning, which has the effect of removing the entire epidermis. The pig ears used in this work were obtained prior to steam cleaning, with epidermis and stratum corneum intact.
Example 21 - Preparation of pig ear skin
The ears were washed under running water and full-thickness dorsal skin was separated from the cartilage via blunt dissection using a scalpel, then hair was removed using an electric razor. The skin was cut into samples of approximately 2cm 2 and visually inspected to ensure that each piece was free from abrasions and blood vessels. Specimens were then stored in a crease free state on aluminium foil at -20 0 C until required.
Example 22 - Permeation of D and F mix across pig ear skin from patches The skin samples were removed from the freezer and left to fully defrost. The donor and receptor compartments of Franz-type diffusion cells (see Figure 13) were greased, to provide a tight seal and prevent any leakage from the receptor phase. The polymer- lined paper was removed from the patches to expose one side and firmly pressed centrally onto the surface of each piece of skin. After adhesion was established, the skin was mounted onto the flange of a receptor compartment (nominal volume 2.5ml) of the diffusion sell, ensuring that the patch was placed directly over the flange aperture. The donor compartment was then placed on top and clamped to the receptor compartment using a pinch clamp. EtOH/water 10:90 receptor phase (maintained at 37°C) was used to fill the receptor compartment carefully to ensure that no air bubbles were in contact with the underside of the skin and the receptor phase was in contact with the skin. A small magnetic stirrer was added to ensure homogeneous mixing of the receptor phase. The Franz cells were placed on a magnetic stirrer immersed in a water bath (containing vercon) and maintained at a constant temperature of 37 0 C (therefore the surface of the skin was approximately 32 0 C). The donor aperture was occluded to mimic the backing layer of a commercial patch protecting it from moisture and the sampling arms were occluded to prevent evaporation of the receptor phase. At time points of 3, 6, 12, 24, 48 hours, 0.2μ£ of receptor phase was sampled and
transferred into auto sampler vials which were refrigerated at 2 - 4°C until required for analysis. The receptor phase was then replenished. The total number of replicates for each treatment was five.
Selection of paint medium
Of the array of vehicles available for the topical administration of Digoxin and Furosemide, a paint-like or lacquer formulation was considered particularly attractive for the treatment of various anatomical sites. This is because such treatments are relatively simple and offer a degree of resistance to abrasion. Also, such products are currently commercially available, for example, Salicylic Acid Collodion BP.
Example 23 - Collodion formulation
Commercially prepared Collodion BP is a liquid, with a high solvent content (mainly diethyl ether). On application to the skin the volatile components of the collodion rapidly evaporate transforming the liquid solution into a dry, solid film which will adhere to the skin. As with drug-in-glue adhesives, the change in physical state of the vehicle means that the thermodynamic activity, of liquid/semi-solid dermatological systems, only applies to the initial liquid formulation and is irrelevant to the formulation in a solid state. Therefore, the solubility of the actives to a certain extent is arbitrary, as more drug mix can be added by increasing the proportion of solvent to the liquid formulation.
After evaporation of the solvents in the formulation on solidification, crystallisation of the compounds will occur however; they will be retained in the matrix of the formulation. This could increase rates of delivery, as direct contact between crystallisation and the skin often provides good delivery, although the precise mechanism of this is unknown. Also affect the ability of the collodion to maintain intimate contact with the skin at a microscopic level effecting drug delivery i.e. the limiting factor, would be adhesion to the skin.
Several preliminary experiments were conducted to determine the maximum loading of drug mix in collodion. Problems encountered included sedimentation of drug mix due to limited solubility in collodion. The drug mix did not easily re-suspend on shaking; meaning that only a small amount of drug mix would dissolve in the collodion. To overcome this problem, and increase the solubility of the drug mix in collodion, various amounts of ethanol were added to the formulations until a balance between drug dissolving/reduced rate of sedimentation (which would increase if viscosity decreased) and the rate of drying (solvent evaporating) was found. It was concluded that 0.01 g of drug mix in 5 ml of collodion and 5 ml of ethanol was a good compromise. This formulation also showed good adhesive properties.
Example 24 - Preparation of collodion formulations
Drug mix (for composition see Table 7) 0.02g (a stock was made) was weighted on an analytical balance (accurate to 5 decimal places) and added directly to 10m£ of collodion and 10mi of ethanol in a McCartney bottle. The molar ratios used were F: D; 1 :1 , 1:2.5 (2:5) and 1 :10 because a smaller amount of drug mix was used, compared to the drug-in-adhesive and this allowed measurable amounts of F to be used. Each of the McCartney bottles was vortexed for three minutes and left to rotate on a blood serum rotator overnight, to ensure that the mixture was homogeneous and that any air bubbles present had dispersed. Control collodions were also prepared by the same method, however, no drug mix was added.
Table 8: Composition of F and D in 0.01 g drug mix - used to prepare collodions. Ratio of F: D Mass of F (g) Mass of D (g)
1 :1 2.977 x10 "3 7.023 x10 -3
1:2.5 1.447 x10 "3 8.553 x10 "3
1 :10 4.058 x10 '4 9.594 xiO "3
Example 25 - Diffusional release of D and F from collodions
Different molar ratios of the two drugs were used to determine affect upon release rate and the extent of the release of each drug. The collodion, 200μ£, was dispensed to the bottom of general 7ml glass screw cap vials using a Gilson Pipette and left to dry for three hours. Then 2ml of dissolution medium, again de-gassed EtOH/water 10:90, was added to each vial. The amount of receptor phase sampled and replenished was 200μ£, with a total of five replicates performed for each treatment. The formulation that demonstrated optimum release was selected for skin permeation experiments.
Example 26 - Permeation of D and F across pig ear skin from collodion
The method was essentially the same as described in Example 25. Mounted skin membranes were does with 200μl of collodion and left for thirty minutes to dry before the receptor phase was added. A total number of 4 replicates were performed for each treatment.
High Pressure Liquid Chromatography (HPLC) analysis
HPLC analysis was performed using the same method as described previously i.e. an Agilent series 1100 automated system, fitted with a Phenomenex Kingsorb 5mm C18 Column 250 x 4.6mm (Phenomenex, Macclesfield, UK) and a Phenomenex Securiguard guard column. D and F were detected using an ultraviolet (UV) detector set at wavelength 220 nm. The mobile phase consisted of 40:30:30 Water: MeOH:MeCN, de-gassed by drawing through a 0.45 membrane and run isocratically for 10min at a flow rate of 1 ml min "1 . The injection volume of each sample was 20μl. The retention time of F and D was typically 2.6 minutes and 5.2 minutes respectively. Data were acquired using Agilent software. Standard calibration curves were determined using standard solutions of 5, 10, 20, 40, 80 and 100μg mf 1 in the receptor phase, to prevent solvatochronic effects. The limit of detection was 0.1 μg rnf 1 .
Chromatogram peaks were integrated manually, and the data corrected for dilution effects. Cumulative release was determined and plotted against the square route of time to determine release rates. Cumulative permeation data were determined and plotted against time to order to obtain flux. Excel was used for data processing and Minitab for statistical analysis.
Example 27 - Diffusional release of D from patches Cumulative mass of Digoxin (D) released Cumulative release (mass/area) profiles of D from adhesive containing molar ratios of F: D; 1:1, 1:25, 1:100 were determined over 24hr and are illustrated in Figure 14. D was released from all the patches. The trend in the greatest cumulative release after 24hr (table 3) was 1 :100>1 :1>1 :25. The patches containing ratios of 1:1 and 1 :100 had similar profiles, and up to 12hr the greatest release was observed form the patches containing a molar ratio of 1 :1. Error bars were small.
Percentage release of loading dose of D from model patches
The percentage release of the loading dose of D from adhesives containing molar of F: D; 1:1 , 1:25 and 1 :100 was determined over 24hr and are displayed in Figure 39. The percentage release mimics the trend observed in Figure 38. Maximum percentage release values of D after 24hr are illustrated in table 3. Error bars were small.
Table 9 Maximum release values of D from patches at 24hr
Ratio Q 24 release Mass/Area (μg/cm 2 ) Q 24 release %
1 :1 130.03 3.17
1 :25 25.25 3.49
1 :100 136.18 0.56
Example 28 - Main effects plot illustrating D release data from patches
The main effects plot illustrated in Figure 40 used to visually summarise the data from the diffusional release of D from model patches. It illustrates the trend in ratio of percentage release of the loading dose of D and how this increases over time.
Example 29 - Determination of rate of release of D from patches
Linearity denoted by the cumulative release (mass/area) profiles in Figure 38 indicated zero order release kinetics from all three molar ratios. Rate of release was determined from the gradient of a trend line for each profile. For ideal linearity R 2 =1. Release values are illustrated in table 6.
Table 10 Release rate of D from model patches and R 2 values for each molar ratio
Ratio Release rate (mcgcm '2 h "1 )
1 :1 4.8353 0.9858
1 :25 1.0844 0.9916
1:100 5.1899 0.9945
Example 30 - Diffusional release of Furosemide (F) from model patches Cumulative mass of F released
Cumulative release (mass/area) profiles of F from adhesive containing molar of F: D: 1:1 , 1:25, 1:100 were determined over 24hr and are illustrated in Figure 41. F is released from all the patches. The 1 :1 ratio demonstrates a typical release profile, whereas release from 1 :25 and 1 :100 is linear. The trend in greatest cumulative release after 24hrs was 1 :1>1 :25>1 :100. Error bars were small.
Example 31 - Percentage release of loading dose of F from model patches
The trend in percentage release of loading dose of F (Figure 42) mimics the trend observed above, for maximum percentage release after 24hr refer to table 11. Error bars were small.
Table 11 : Maximum release values of F from model patches at 24hr
Ratio Q 24 release Mass/Area (μg/cm 2 ) Q 24 release %
1 :1 432.02 22.82
1 :25 10.77 17.23
1 :100 2.85 3.85
Example 32 - Main effects plot to illustrate release data of F from patches The main effects plot illustrated in Figure 43 summaries the data from the diffusional release of F from model patches. It illustrates the trend in ratio of percentage release of loading dose of F and how percentage release of loading of F increased over time.
Example 33 - Permeation of D and F mix across pig ear skin from patches
Permeation of D across pig ear skin from patches
Permeation of D across pig skin is illustrated as both cumulative mass/area and percentage permeation of loading of D and is shown in Figures 44 and 45 respectively. The profiles are of a similar shape and are atypical permeation profiles. However, they do illustrate that D is permeated the pig skin. Error bars are larger than for release results. Apparent maximum flux (Table 12 along with maximum permeation values) was calculated from Figure 45 however lag time and Kp could not be calculated from these profiles-.
Example 34 - Permeation of F across pig ear skin from patches
Permeation of F across pig skin is illustrated as both cumulative release (mass/area) of loading and percentage permeation of loading of F and is shown in Figures 46 and 47 respectively. Both of the profiles are of a similar shape and are atypical permeation profiles. However, they do show that F permeated the pig skin. Error bars are larger than for release and permeation of D across pig skin. Apparent flux maximum (Table 8 and maximum permeation values) was calculated, however lag time and Kp could not be calculated from Figure 46.
Table 12 Maximum permeation values of Digoxin and Furosemide from patches across pig skin
Active Q 24 permeation Q 24 Apparent flux SEM
Mass/Area permeation % maximum μgcm '2 h '1 (μg/cm 2 )
F 101.92 6.07 0.158 0.072
D 5.81 0.12 3.499 0.372
Example 35 - Comparison between mass released from the patches containing a F: D in a 1:1 ratio and mass permeated through the skin
Comparison between the mass/area of Digoxin released from the patches and mass/area of Digoxin that permeated the skin
Figure 48 illustrates the mass/area of D released from the patches and also the mass/area of D that permeated the skin and allows a comparison to be made. A larger mass of D was released from the patches that permeated the skin.
Example 36 - Comparison between the mass/area of Furosemide released from the patches and mass/area of Furosemide that permeated the skin
Figure 49 illustrates the mass/area of F released from the patches and also the mass/area of F that permeated the skin and allows a comparison to be made. A larger mass of F was released from the patches that permeated the skin.
Example 37 - Diffusional release of D from collodion
Cumulative mass/area of D released from collodions
Cumulative release profiles of D from collodions containing molar ratios of F: D, 1:1 , 1 :2.5 and 1 :10 were determined over 24hr and are illustrated in Figure 50 released from each of the collodions. The trend the in greatest cumulative release after 24hr (see table 7) was 1 :100.1 :2.5>1 :10. The shape of the three profiles were similar and error bars small.
Example 38 - Percentage release of loading dose of D from collodion The percentage release of the loading dose of D from collodions containing molar ratios of F: D; 1 :1, 1 :2.5 and 1 :10 was determined over 24hr and are displayed in Figure 51. The percentage release mimics the trend observed in Figure 50. Maximum percentage release values of D after 24hr are illustrated in table 13. Error bars were small.
Table 13: Maximum release values of Digoxin from collodions after 24hr
Ratio Q 24 release Mass/Area (μg/cm 2 ) Q 24 release %
1 :1 25.78 32.54
1 :2.5 29.32 25.89
1 :10 34.01 30.36
Example 39 - Determination of rate of release of loading of D from collodion
Figure 52 illustrates the cumulative release of Digoxin from the three different collodions plotted against the square root of time. Linearity of the plots indicates first order release kinetics, 1 :10 shows the greatest rate of release. R 2 and rate of rate of release are illustrated in table 11.
Table 14: Rate of release values of D from collodion
Ratio Release rate (mcgcm h )
1:1 4.5393 0.9859
1:25 4.8852 0.9816
1:100 6.5231 0.9709
Example 40 - Diffusional release of Furosemide from collodion Cumulative mass/area released of F from collodion
The cumulative release profiles of F from collodions containing molar ratios of F: D; 1 :1 , 1 :2.5 and 1 :10 were determined over 24hr and are shown in Figure 53. F is released from all the different collodions producing a typical release profile. The trend in greatest cumulative release after 24hr was 1 :1>1 :2.5>1 :10 (see Table 15 for maximum release values). The size of the error bars varied.
Example 41 - Percentage release of loading dose of F from collodions The trend in percentage release of loading dose of F (Figure 54) mimics that of cumulative release. For maximum percentage release after 24hr see table 12. Error bars were small.
Table 15 Maximum release values of F from collodion after 24hr
Ratio Q 24 release Mass/Area (μg/cm 2 ) Q 24 release %
1:1 6.02 18.33
1:2.5 3.27 9.95
1:10 0.77 3.33
Example 42 - Release rates of Furosemide from collodion Figure 55 depicts cumulative release of F from the collodions containing the three different molar ratios plotted against the square root of time. Linearity was reported from reported from 1 :1 indicating first order kinetics. For release values refer to Table 16.
Table 16: Rate of release data of F from collodion
Ratio Release rate (mcgcm h )
1:1 1.4811 0.9438
1:2.5 1.0043 0.8742
1:10 0.0575 0.1356
Example 43 - Permeation of Digoxin and Furosemide mix across pig ear skin from collodions
Permeation of D across pig ear skin from collodions
Permeation of D across pig skin is illustrated as both cumulative mass/area and cumulative percentage of loading of D and are illustrated in Figures 56 and 57 respectively. Both of the profiles are similar in shape and are atypical of permeation profiles. However they do illustrate that D from collodion is permeated the skin. Error
bars were larger than for collodion release results. For AFM and maximum permeation values refer to Table 17. Lag time and Kp could not be calculated from these profiles.
Example 44 - Permeation of Furosemide across pig ear skin from collodion Permeation of F across pig ear skin is illustrated as both cumulative mass/area and cumulative percentage and shown in Figure 58 and 59 respectively. The profiles are of a similar shape and are atypical permeation profiles. However, they do show that F permeated the pig skin. Error bars are large. AFM and maximum permeation values are displayed in Table 17. However, lag time and Kp could not be calculated from Figure 58.
Table 17 Maximum permeation values of D and F mix from collodion
Active Q 24 permeation Q 24 Apparent maximum SEM
Mass/Area permeation % flux μgcm "2 h "1 (μg/cm 2 )
F 39.45 79.64 4.3423 2.05
D 8.03 5.39 0.313 0.83
Example 45 - Comparison between mass released from the collodion containing F: D in a 1:1 molar ratio and mass permeated through the pig skin
Controls were used throughout this work. During the release studies, formulations containing no actives were used as controls. The corresponding chromatograms illustrated no peaks at the wavelength of detection.
During permeation studies formulations containing no actives and skin without a formulation applied to it were used as controls. The corresponding chromatograms illustrated no peaks at the wavelength of detection.
Diffusional release of Digoxin and Furosemide from patches
Dermatological formulations are required to release the active compound(s) at the surface of the skin. Generally, the rate-limiting step in skin permeation is transport across the stratum corneum, although in some cases the rate-limiting step can be release of the active compound(s) from the formulation. If this occurs the bioavailability of the compound(s) may be affected. This is less likely to happen during the permeation of D and F through callous material.
The release of Digoxin and Furosemide from the adhesive could potentially be limited by three parameters: molar ratio, drug loading and the interaction of the drugs with adhesive. The aim of this investigation was to establish which molar ratio would release the maximum mass of D and a sufficient mass F and could therefore be used in subsequent permeation studies. Overall the release of D would have a greater influence in the choice of ratio than F, refer to example 22.
Diffusional release of Digoxin from patches
These results showed that a proportion of the loading mass of D was released from all of the patches. The extent of release was observed in terms of cumulative release (mass/area), to establish the maximum mass/area of D released. From this the maximal dose that could potentially come in contact with the surface of the patients' skin could be estimated. This was found to be in the order of 136.18μgcm "2 .
An initial burst in the release of D was observed from all of the patches. This was most prominent from the patches containing 1 :1 and 1:100 molar ratio. This may be due to release of D molecules at or near the surface of the patch. The release from all three ratios was linear, displaying zero order release kinetics, which are desirable of a topical delivery device. The trend for greatest release (mass/area) was 1 :100>1:1>1:25. The 1:100 ratio gave the greatest mass/area released as expected because it contained the largest mass/area of D. The 1 :1 ratio gave similar results, which was not expected as it contained the smallest mass of D, suggesting that loading, was not the rate-limiting factor of release.
Percentage release of the loading dose was calculated to allow, for slight variation in patch preparation, and comparison between the formulations. Percentage release was expected to be small with a large amount of drug retained in the matrix.
The rate of release was examined, in order to distinguish between 1:1 and 1:100 in terms of which formulation would give the maximum delivery of D in the shortest time period. Although the rate of release from 1 :100 was the greatest at 5.19μg cm "2 h '1 it was surprisingly similar to that of 1 :1 at 4.84μg cm "2 h "1 .
Diffusional release of F from patches
A proportion of F was released from all the patches and this confirmed that both drugs were released simultaneously from the matrix and therefore could potentially simultaneously permeate the skin.
Again the extent of release was observed as cumulative release (mass/area) to establish the maximum mass released, and hence the maximal dose of F that could
potentially come into contact with a patients' skin. This was found to be in the order of 432.02μg cm "2 .
In summary, this data provided sufficient information to allow the rational selection of the most promising formulation for permeation studies. Thus patches containing D: F in a 1:1 molar ratio were selected. Percentage release of both D and F is greater than from the other ratios. The 1:1 ratio also released the greatest mass/area of both drugs.
The larger the concentration gradient, the higher the rate of permeation. This ratio also provided the greatest rate of release i.e. an optimal mass is released in the shortest time.
Permeation of D and F mix across pig skin from model patches containing 1:1 molar ratio Dermal absorption involves several processes. Firstly the actives are released from the formulation; they then encounter the surface of the skin and establish a SC reservoir. This leads to penetration of the barrier and finally diffusion into another compartment of the skin (Schaefer and Redelmeler, 1996).
Permeation profiles were presented as cumulative mass/area and cumulative percentage permeation of total loading. Cumulative permeation results illustrated that both D and F permeated the skin and therefore have potential as a future localised neuropathic pain treatment. Permeation through the skin can predict localisation and therefore it is possible that both D and F are coming in to contact with the basal layer of the epidermis.
Comparison between the mass of Digoxin and Furosemide released from model patches containing F: D 1:1 and mass permeated through the skin
Differences were observed in the mass/area of D and F released from the patches and the mass/area of D and F permeated across the skin, in that mass released was greater than that permeated. Assuming that the mass released of D and F from the patches into the dissolution medium is approximately the same as that released at the SC. This suggests that a quantity of the each of the actives could be retained in the skin. From visual inspection of Figures 60 and 61 it is possible to observe that a higher proportion of D than F is retained in the skin. This was a positive result as it is desirable to have an excess of D at the site of infection.
Diffusional release of Digoxin and Furosemide from collodion
As with the patches, the release of D and F from the collodion could be potentially limited by three parameters, molar ratio, drug loading and interaction between the drugs and the collodion matrix. The aim of this experiment was to establish which collodion contained the molar ratio of D: F that released the maximum amount of D and a sufficient amount of F. This would be used for further permeation studies. Overall the release of D would have a larger influence in choice of ratio over release of F (Example 23).
Diffusional release of Digoxin from collodion
The results illustrated that a proportion of the loading mass of D was released from all three of the collodions, and release increased over time. Cumulative release (mass/area) plots depicted extent of release and illustrated the maximum dose released after 24hr. The maximal dose of D released after 24hr was in the order of 34.01 μg cm "2 and is in theory the dose delivered to the surface of the patients' skins.
Cumulative release (mass/area) profiles for the three ratios, were typical of release, and began to plateaux after six hours. The trend for release was 1 :10> 1 :2.5 >1 :1 , and was expected demonstrating a proportional relationship between the initial mass of D in the collodion and the mass released from it. From these results it is possible that loading mass, molar ratio or interaction with the vehicle (collodion) could be the limiting factor in mass released.
Release profiles for percentage release of loading dose were also plotted, to allow for variation in volume of collodion pipette into each vial and to allow comparison between formulations. Percentage release ranged from 25.54 - 30.36%, which was relatively high compared to approximate 10%, expected and compared to the patches. This suggested that differences between the adhesive and collodion matrix could be responsible. A possible explanation could be the formation of larger micro channels in the matrix of the collodion as the solvent evaporates on drying, or a greater number may be formed than in the patches due to the higher solvent content of collodion.
Percentage release of loading dose did not follow the same trend as cumulative release mass/area, and instead was 1 :1>1 :10>1.2.5. This was no pattern was followed. However, this trend correlated with the trend in cumulative mass/area released of D from the patches. This suggested that the effect of the vehicle would only have an influence on the over all extent of release from all three of the collodions, and that the difference in molar ratios contribute towards the trend.
Statistical evaluation by a two-way ANOVA, illustrated that there was a significant difference between 1 :1 and the other ratios. Optimum percentage release was attained from 1 :1, however this did not give the largest mass/area released. A significant difference in percentage release at each time point was observed (as with D) which
increased over time, concluding that frequency of administration of the collodions for the delivery of D, like the patches would be at the most once every 24hr.
Error bars were small indicating good reproducibility between samples. In summary at this stage of the investigation, likewise with the patches the decision of which collodion will be used for permeation studies lay between 1 :1 and 1 :10 (i.e. the lowest and greatest excess moles of D).
Linear plots indicated first order release kinetics. In general the rates of release were similar, although 1 :10 gave the greatest rate of release whilst 1:1 gave the smallest, the optimum molar ratio could not be determined from this data.
Diffusional release of Furosemide from collodion
Furosemide was released form all the collodions, indicating that all the collodions could be potentially used in permeation studies, as they illustrated simultaneous release of Digoxin and Furosemide. Maximal dose released after 48hr was in the order of 6.02μg cm "2 .
Cumulative release (mass/area) of Furosemide from collodion was lower than that of Digoxin, unlike the patches, thereby potentially delivering more of Digoxin to the site of infection or pain, which was desirable. The profiles from all the molar ratios were typical of release, an initial burst was observed between 1-6hr, and plateau in the profile at 6hrs, which was comparable with the Digoxin release profiles. This was most likely to be due to depletion, because it was observed from both drugs and to a lesser extent in the patches (which contained a higher dose of D and F). The trend in cumulative release (mass/area) was 1:1>1:2.5>1:100 and was unexpected as 1:1 contained the lowest (mass/area) of F. This trend was also observed in the percentage
release data which indicates that D having an effect on the release of F as otherwise one would expect the percentage release of F to be the same for each ratio.
Statistical evaluation by a two-way ANOVA, indicated a significant difference between 1 :1 and the other ratios. Optimum percentage release was obtained from 1 :1 , which also released the greatest mass. A significant difference in percentage release at each time point was illustrated as with D, less of an increase as observed within time points after 6hr. This suggests that administration of collodion may be required more frequently for optimum delivery of F.
Error bars throughout this part of the investigation were small indicating good reproducibility between samples. In summary of this data, for delivery of F, the 1 :1 ratio appeared to be the strongest candidate.
Cumulative mass/area released of F against the square root of time, depicted linearity for 1 :1 ratio with R 2 value close to 1. This ratio also illustrated the highest rate of release. However R 2 values for the other ratios were not close to unity indicating poor correlation.
Comparison between Digoxin and Furosemide release data from collodion
In summary, a decision of which ratio would potentially provide optimum delivery of D and F was not as clear as for the patches, especially regarding the release of D.
This investigation provided enough information for a molar ratio to be chosen for permeation studies. Patches containing D: F in a 1 :1 molar ratio were used as, percentage release of both D and F was essentially greater than from the other ratios.
The 1 :1 ratio also released the greatest mass/area of F. Providing the greatest concentration gradient.
Permeation of D and F across pig skin from collodion containing D and F in a 1 :1 molar ratio
Permeation data was shown as cumulative mass/area and percentage permeation of total loading. The permeation data illustrated that both F and D simultaneously permeated the skin, and can be used as a prediction of localisation.
The permeation profiles for both D and F were atypical as were the permeation profiles for the patches. Therefore suggests this could be related to the nature of the actives individually or in combination. The profile for D is however different to that of F differing from a typical profile only during phase 1. The percentage release profile for D mimicked this shape. The profiles for F were a similar shape to that seen from the patches.
The SEM for the permeation profiles was larger in magnitude than those for the release profiles. This indicated less reproducibility in data compared to the release data. The major difference between the release experiments and the permeation was the introduction of the skin, therefore this may have had an impact on the results. The SEM was also of a larger magnitude for F compared to D. A reason for this could be the amount of solvent present in the liquid state of the collodion (all solvent had evaporated from the patches during preparation) could affect the integrity of the skin and reduce reproducibility between replicates.
The atypical nature of these profiles meant that SSF could not be accurately measured and AMF was measured instead. For D this was calculated between 12 - 24hr to be
0.313μg cm "2 h '1 and for F between 6 - 12hr to be 4.3423μg cm "2 h "1 . It was not possible to measure lag time and only an estimation of kp was calculated.
The mass/area of D that permeated the skin was 8.02μg cm "2 (1.03 x 10 '8 μg cm "2 ) compared to 28.49μg cm "2 (8.62 x 10 "8 μg cm "2 ) of F, suggesting that drug delivery to the basal layers is a reality. The observation that a greater mass/area of F permeated may be associated with the large SEM indicating that these results lacked reproducibility between samples. If integrity of the skin had decreased as F is smaller than D it is possible that it would penetrate the skin more effectively. It is also less lipophilic and therefore less likely to become trapped in a compartment of the skin. A larger percentage of loading of F permeated the skin than D, which was the same for the patches.
The ratio of moles that permeated the skin was D: F 1 :8, supporting suggestions that F permeated the skin more easily.
Comparison between patches and collodion
It was not possible to statically compare the patch formulation to the collodion formulation, as although the rational behind the choice of ratio was the same, the actual ratios chosen for each formulation were slightly different. The discussion so far has compared the data obtained from the patches and collodion, this next part of the discussion compares qualitative difference between the formulations.
Vehicle differences A large amount of ethanol was present in the collodion on application to the skin, comparatively there was no ethanol present in the patches. The ethanol can cause pain and irritation to some skin types. There are possible formulation solutions to
overcome this, for example the inclusion of a local anaesthetic such as lignocaine to the formulation. However this would increase the number of actives in the formulation and could complicate the licensing of the product. The inclusion of ethanol might aid percutaneous absorption to the basal cells. Dehydration of the skin, for example, keratinised skin may cause it to crack and forming microscopic pathways to the site of action. Ethanol is also known to act as a permeation enhancer by solubilising the lipids in regular skin. The extent of this in skin is unknown, but perhaps will be reduced due to a lower proportion of lipids in this type of tissue.
Properties of the dosage form
The patch offers a thicker film than the collodion, meaning that a larger mass of one or both of the binary drug combination can be incorporated into the formulation, and perhaps offer a prolonged duration of treatment, increasing compliance. Thickness of film of collodion is approximately 5 - 20μcl limiting the amount of actives applied to the skin (Schaefer and Redelmirer, 1996) compared to approximately 1mm of the patches. This suggests that movement of molecules from the upper surface of the patch through the bulk matrix to a greater extent in the patches, reducing frequency of dosing and aiding compliance. Both dosage forms are flexible. The suitability of these patches in the treatment of neuropathic pain has been demonstrated and will be further established in forthcoming clinical trials. Overall the formulation determines the kinetics and extent of percutaneous absorption, which has an impact upon the onset of action, duration and extent of a biological response.
In addition to the above described examples, the following additional embodiments demonstrate the in vitro release and permeation of Digoxin and Furosemide from transdermal delivery devices. Several drug-in-glue formulations containing differing amounts of Digoxin and Furosemide were compared for their rates of drug release,
rates of drug permeation through porcine skin and the concentration of drug within the skin sample. The ratios of the active principles were varied to investigate optimum formulations for delivery of Furosemide and Digoxin to provide dermal saturation.
Digoxin and Furosemide were purchased from Sigma, UK. Glue 1 was sourced from National Starch and Chemical Company. All solvents and chemicals used for the release and permeability studies were purchased from Sigma. The porcine ear skin that was used as a skin barrier was purchased from a local abattoir.
A convenient drug loading is 25mg/m£ of both Digoxin and Furosemide within the acrylate glue at a 1:1 ratio. If the total concentration of drug is maintained at 50mg/mJ! then the following systems can be examined: 50mg/mJ! Digoxin
46.7mg/mf Digoxin and 3.3mg/mt Furosemide (14:1 ratio)
40mg/mJ! Digoxin and 10mg/ml! Furosemide (4:1 ratio)
30mg/m£ Digoxin and 20mg/m^ Furosemide (3:2 ratio)
25mg/mJ! Digoxin and 25mg/m£ Furosemide (1:1 ratio) 20mg/mJ! Digoxin and 30mg/m£ Furosemide (2:3 ratio)
10mg/m£ Digoxin and 40mg/mJ! Furosemide (1:5 ratio)
3.3mg/m(! Digoxin and 46.7mg/mJ! Furosemide (1:14 ratio)
Plus a control using the glue only
The above systems measure ratios in a mass by mass form. Molar ratios of drugs were also examined at a 1 :1 ratio of F:D, a 1 :25 and a 1:100 ratio, and results provided in Table 18. Table 18
Drug release studies
Drug release from the patches into a solution of mobile phase was measured for the nine mass-ratio formulations. This was done to compare how the drug loading affects drug release.
Drug Permeation Studies
Drug permeation through porcine ear skin was measured using Franz Diffusion cells where the amount of both drugs that permeated the tissue was measured over time and compared to the initial drug loading within the patch. The molar-ratio patches were used in this study. Pig's ear skin was used as a model membrane and the drug release through this tissue was measured using Franz cell apparatus. The skin was mounted above the receptor fluid that contains water:methanol:acetonitrile (40:30:30) as used for the mobile phase within the HPLC analysis.
The entire system was sealed to avoid moisture loss and samples were taken from the receptor fluid at intervals of 0, 4, 8, 12, 24, 48 and 72 hours. The receptor fluid was stirred continuously to ensure a homogenous receptor solution. The concentrations of both furosemide and digoxin within this fluid were measured via HPLC analysis. After 72 hours the skin was homogenised and the concentration of both drugs within this tissue was determined (via extraction) to note the "saturation" levels.
Skin Saturation studies
It has been well documented that skin has a capacity for the retention of drugs. It is generally thought that drugs with a higher logP value are retained to a greater extent within the skin. The amount of drug that was present in the skin sample at the end of the 72 hour period was measured via homogenisation of the skin onto which the patch had been administered and extraction of the drug. Each Franz cell was loaded with a patch of 2cm diameter that would Q contain.
The cumulative amount of drug that is released from the glue or has penetrated the skin, Q (μg/cm 2 ) was plotted against time in Figure 64. The linear portion of such a slope (at least 5 data points used) was taken as being the steady state flux, Jss. The permeability coefficient, Kp (units = cm per time), the constant for each drug that determines how fast it is able to diffuse either through the glue to allow release or through the skin was then calculated as:
Kp = Jss/Cv
Where Cv is the concentration of the penetrant in the donor compartment (concentration of digoxin or furosemide within the patch, units = μg/cm 3 )
Drug release studies:
Patches were made of the initial nine formulations and the drug release from these formulations into a solution of the mobile phase was measured.
Some example data is shown below, the mass of digoxin released from each formulation was plotted against time in Figure 64. A similar plot was constructed for furosemide.
The gradient of these results was calculated and is a measure of the steady state flux from the patches, Jss. Division of the steady state flux by the initial concentration gives the permeability coefficient, this value is a constant that determines the rate of drug release from the patch. The table below provides the data that measures both the amount of drug release from each patch at 4 days, the steady state flux and the permeation coefficient for each formulation.
The rates of both digoxin and furosemide release from the patches are listed in the table below.
Table 19 shows that at similar concentration values, furosemide is released to a greater extent than digoxin, e.g. compare formulations 1 and 9. The steady state flux for each drug increases as the initial loading of drug within the patch increases. This is as expected as the drug is released from the patch due to a concentration gradient that exists between the drug loading and release medium. The permeation coefficient is a measure of the rate of drug release in cm per second of each drug from the patch.
These values are relatively constant for all formulations which indicates that the two drugs do not interfere in the release of one another. The Kp values for each drug alone are similar to the values in patches that contain both drugs. Kp for furosemide is approximately four times greater than Kp for digoxin, this is likely to be due to the comparatively smaller size of furosemide.
Table 20 below shows the data for the drug released from the patches that has penetrated the skin.
Table 20 shows the penetration of the skin, both the flux values and permeation coefficient values are much lower than the release of the drug from the formulations listed in the table above. This is expected and reflects the barrier properties of the skin. Furosemide penetrates the skin to a greater extent than digoxin as demonstrated by the permeation coefficient which is nearly eight times higher than digoxin.
The drug that accumulated in the skin was also measured. The drug that was present in a 2 cm diameter cross section of skin was calculated for all four formulations.
The level of digoxin appeared to be independent of the loading formulation, indicating that the skin was saturated with digoxin at a concentration of 40 μg over 3.14 cm 2 or 12.73 μg/cm 2 . Furosemide did not accumulate within the skin and permeated directly through the skin. The concentration measured at 72 hours was a transient indication of furosemide within the skin that was dependant upon the loading concentration. Results are shown in Figure 65.
The rate of furosemide release from the patch, Kp for the patch was 6.53 x10 "10 cm per second, this was not greatly faster than the rate of furosemide penetrating porcine ear skin at 4.32 x10 "8 cm/second.
Digoxin was considerably slower both in terms of drug release and also in terms of skin penetration with permeation coefficients of 1.60 x 10 "7 cm/s and 5.52 x 10 '8 cm/s for the patch and skin respectively.
If the initial patch concentration for digoxin is plotted against the steady state flux rate through the skin, as shown in Figure 66, it can be seen that for the flux to be greater than 0 the initial concentration within the patch must be 804.5 μg/cm 3 .
25 000 μg/cm3 was the lowest concentration used within the skin study. The required flux for effective therapy was 25 μg per day, if this is assumed to come from a patch with a surface area of 1cm 2 then the loading dose should be:
Flux = 25 μg per day per cm 2 = 1.04 μg per cm 2 per hour, thus a loading dose of 6004.5 μg per cm 3 is required.
However, this study enhanced the overall penetration of digoxin through the skin as a very lipophilic substance was used in the donor phase to enhance the concentration gradient to maximise skin penetration of both digoxin and furosemide.
Whilst the above examples all relate to the use of Digoxin and Furosemide, it is believed that one or both of diuretics and cardiac glycosides can be used in the treatment of neuropathic pain.
It is believed that other actives similar to digoxin and furosemide are able to alter the resting potential of damaged or other nerve cells to treat pain, e.g. neuropathic pain and/or HZV or the symptoms thereof . The comparative abilities of these species to act as Ionic contra viral therapy (ICVT) agents is shown below.
The comparative solubilities and ICVT-potencies of Digoxin, Diqitoxin and Lanoxin (IV)
1 ) Comparative 'ICVT-ivities' (Ionic contra-viral therapy-activities) Solutions of Digoxin and Digitoxin were prepared from powder to a concentration of 250μg per ml in 70% ethanol and their ICVT-ivities compared with the 'standard' Digoxin preparation; i.e. IV Lanoxin, which is supplied at 250 μg per ml in 10% ethanol.
The ID 50 values of Digoxin prepared from powder and Lanoxin (circles) (Figure 67) were very similar, i.e. 60ng per ml. Digitoxin (squares) appeared to be marginally better with an ID50 of 30ng per ml.
2) Comparative solubilities
Saturated solutions of Digoxin and Digitoxin (were prepared in 90% ethanol and their 'ICVT-ivities' compared with the 'standard' Digoxin preparation; i.e. Lanoxin.
Digoxin solution prepared from powder was as effective as Lanoxin (circles) (Figure 68).
Digitoxin (squares) was again more effective than Digoxin.
Digitoxin is more soluble than Digoxin; preparation of a saturated solution (17.5mg per ml) in 90% ethanol will enable use at a maximum concentration of 486μg per ml in a 'safe-ocular- concentration (2.5%) of ethanol.
Digoxin was previously used at a concentration of 62.5μg per ml.
486μg per ml is approximately eight times more concentrated and if Digitoxin is indeed twice as potent then it might be possible to use what would effectively be 16 X the previous 'dose'. Toxicity at this higher concentration will, of course, need to be examined.
3) Comparative 'ICVT-ivities'
Fresh solutions of Digoxin and Digitoxin were prepared from powder to a concentration of 250μg per ml in 70% ethanol and their ICVT-ivities again compared with the 'standard' Digoxin preparation; i.e. IV Lanoxin in order to further examine their relative potencies. Results are depicted in Figure 69.
Examples of Thiazide (Hydrochlorothiazide and Metolazone), Sulphonylurea (Tolbutamide), Sulphonamide (Furosemide, Acetazolamide, Bumetanide, Torasemide and Ethacrynic acid) and K sparing diuretic (Amiloride) were tested for ICVT activity. The cardiac glycosides Digoxin, Digitoxin, Lanoxin and Strophanthin G were also tested.
Using Herpes simplex virus (HSV), 50% plaque inhibitory dose (ID50) were established using the standard plaque inhibition assay. Various solvents were required to facilitate testing and these were sometimes detrimental to tissue culture, depending upon their concentration. Certain compounds elicited potent ICVT activity (Furosemide, Digoxin, Lanoxin and Digitoxin) and these were active at high dilution; experimental conditions in which solvent toxicity was excluded.
Other compounds elicited only 'borderline' CVI activity. These compounds (Acetazolamide, Tolbutamide and Hydrochiorthiazide) were further tested using alternative solvents in the same test system (i.e. the plaque inhibition assay) and others (Bumetanide, Torasemide, Tolbutamide and Hydrochlothiazide) in a more sensitive test for ICVT activity in which the effects on virus yields were determined. The effects of cardiac glycosides Digoxin and Strophanthin on virus yields were also tested in this assay.
Hvdrochiorothiazide Solvent: Ethanol 10% 5 mg/ml HSV Plaque 1 D50 Negative @ 2.5 mg/ml
Solvent: NaOH 1 % aqueous 1 0 mg/ml
HSV Plaque 1 D50 400 μg/ml Borderline +/
HSV yield reduced to zero at 600 μg/ml +
Solvent: PEG 10 mg/ml Solvent: PG 0 mg/ml
Solvent: NaOH 1 % aqueous 10 mg/ml
HSV Plaque ID50 500 μg/ml Borderline +/
Solvent: PEG 10 mg/ml
HSV Plaque 1 D50 500 μg/ml Borderline +/ HSV yield reduced to zero 300 μg/ml +
Solvent: PG 10 mg/ml
HSV Plaque ID50 500 μg/ml Borderline +/-
HSV yield reduced to zero 300 μg/ml +
Solvent IPA 10 mg/ml
HSV Plaque 1D50 250 μg/ml Borderline +/
Sulphonamide Furosemide +
Solvent: aqueous (IV) 10 mg/ml HSV Plaque 1 D50 1 mg/ml
Solvent: PEG 40 mg/ml
HSV Plaque 1 D50 Negative @ 500 μg/ml
Solvent: PG 7mg.ml
HSV Plaque 1 D50 Negative @ 100 μg/ml
Bumetanide Solvent: (IV) Aqueous 500 μg/ml
HSV Plaque 1 D50 Negative @ 100 μg/ml -
HSV yield reduced Borderline +/-
Solvent: NaOH 1 % aqueous 5 mg/ml
HSV Plaque 1 D50 60 μg/ml Borderline +/
HSV yield unaffected at 90 μg/ml
Solvent; (IV) Aqueous 100 μg/ml HSV Plaque 1 D50 25 μg/ml Negative
K sparing diuretic Amiloride
Solvent: Aqueous 500 μg/ml
HSV Plaque ID5O 250 μg/ml +/-
Diqoxin (IV) 250 μg/ml
HSV Plaque 1 D50 60ng/ml +
HSV yield reduced +
HSV Plaque ID50 30 ng/ml +
HSV yield reduced +
Lanoxin (IV) 250 μg/ml
HSV Plaque ID5O 60ng/ml + HSV yield reduced +
Strophanthin G Solvent: Aqueous
HSV Plaque 1 D50 1 mg/ml Cytotoxic HSV yield reduced Borderline +/-
Other delivery methods may be used to treat neuropathic pain. Depot applications which release actives to alter the central nervous system, spinal cord or the principal or major nerve groups which extend from the spinal column.