In vitro skin permeation evaluation: the only realistic option

Increasing requirements for cruelty-free risk assessment in the cosmetic industry have led to the development of several alternative experimental evaluation strategies. Quantification of the potential dermal absorption of ingredients of cosmetic and other formulations by determination of human skin permeation rates in vitro is particularly relevant. Using modifications of standard in vitro protocols the human skin permeation rates of several cosmetic ingredients and potential contaminants have been determined under conditions designed to mimic consumer use. Skin penetration and permeation of octyl salicylate (a sunscreen), nonylphenol ethoxylates (surfactants) and three nitrosamines (potential contaminants) is discussed. The data demonstrate the usefulness of this technique as a tool in the overall risk assessment of cosmetic formulations.

Content variability of active drug substance in compounded oral 3,4-diaminopyridine products

WHAT IS KNOWN AND OBJECTIVE

3,4-diaminopyridine (3,4-DAP; amifampridine) is used for symptomatic treatment of Lambert-Eaton myasthenic syndrome. Until recently, it was only available as a compounded product, which raises safety concerns because of possible high variability in active drug substance content. The objective of this study was to evaluate the variability in dosage form weight, active content variability and impurity of compounded oral 3,4-DAP drug products.

METHODS

Ten samples each of 9 oral 3,4-DAP compounded products were weighed, extracted with water and the 3,4-DAP content determined by ultra high-performance liquid chromatography.

RESULTS AND DISCUSSION

Variability in dosage form weight ranged from 0·81% relative standard deviation (RSD) to 4·82% RSD. In the 90 samples tested, 3,4-DAP content ranged from 22·2% to 125·2% of declared label content. All 10 samples of one compounded product had active drug substance content well below the declared label content (35·0%, 51·7% RSD). No compounded product achieved the Good Manufacturing Practice (GMP) standard of 95-105% range limit of declared label content; one achieved 90-110%, and four others achieved 80-120% of declared content for all 10 samples. There was no evidence of a significant presence of degradation products or related substances in any compounded product.

WHAT IS NEW AND CONCLUSION

Compounded 3,4-DAP products are subject to considerable variability in active drug substance content. This variability seems to be principally because of heterogeneous formulated material rather than variation in dosage form weight.

In-vitro human skin penetration of the fragrance material geranyl nitrile

In-vitro human skin permeation and distribution of geranyl nitrile (GN) was determined using epidermal membranes following application (5 microl/cm(2)) in 70% ethanol, under non-occlusive conditions, at maximum in-use concentration (1%). Permeation was measured (12 time-points over 24 h) using 6% (w/v) Oleth-20 in pH 7.4 phosphate buffered saline as receptor. Permeation of reference benzoic acid was assessed using the same skin donors. Overall recovery of GN at 24 h was low (14.1+/-0.4%) due to evaporation. Evaporative loss of GN from polytetrafluoroethylene (PTFE) sheet, under the same conditions was rapid (93% over 24h) although this overestimated loss during permeation where evaporation competed with uptake. At 24 h, 1.89+/-0.15 microg/cm(2) GN, (3.74+/-0.30% of applied dose) (mean+/-standard error, SE, n=12), had permeated. Following rapid initial permeation, the absorption plateaued due to depletion. Levels of GN in the epidermis (plus any remaining stratum corneum after tape stripping), filter paper membrane support and receptor fluid were combined (as per SCCNFP guidelines) to produce a total absorbed dose value of 4.72+/-0.32%. Systemic exposure resulting from the use of GN as a fragrance ingredient, under unoccluded conditions, would be low based on the currently reported use levels.

The Role of Menthol in Skin Penetration from Topical Formulations of Ibuprofen 5% in vivo

In vivo plasma profiles from formulations containing 5% ibuprofen were compared after a single topical application in a randomised, double-blind, cross-over trial. Ibuleve gel (Dermal Laboratories, UK) contained only ibuprofen whilst Deep Relief gel (Mentholatum, UK) also contained 3% menthol. In contrast to results obtained when these products were compared under in vitro conditions, there was no statistically significant difference in vivo between delivery of ibuprofen. Estimated relative bioavailability fraction (Deep Relief gel/Ibuleve gel) from log-transformed AUC((0-24h)) was 0.99 (95% CI: 0.94-1.04), estimated C(max )ratio was 0.96 (95% CI: 0.91-1.00) and estimated t(max) ratio was 1.01 (95% CI: 0.81-1.20). Menthol produces local vasodilation, which reduces skin barrier function, and these data demonstrate that it is inappropriate to extrapolate from in vitro data where formulation components produce biologically-mediated enhancement of permeation which cannot be modelled ex vivo. In clinical use, these products deliver comparable amounts of ibuprofen, but only Deep Reliefgel provides the secondary immediate benefit of the direct analgesic action of menthol.

Molecular imprinted polymer sensors: implications for therapeutics

A biosensor is a sensor that uses biological selectivity to limit its perception to particular key molecules and can be defined as an analytical device possessing a biological or biologically derived sensing element integrated with or associated closely with a physicochemical transducer. In the future it is likely that a number of key developments in therapeutic monitoring and intelligent drug delivery will rely on real-time feedback information in order to deliver an appropriate response. However due to issues of integration and the fragility and unreliability of the bio-molecule, biosensors are currently unable to fulfil this role. Molecular imprinted polymers are viable alternatives to both antibodies and enzymes and this review considers the current position of molecular imprinted polymer sensing.

Percutaneous penetration of diethanolamine through human skin in vitro: Application from cosmetic vehicles

Concern has been raised over the safety of diethanolamine (DEA) which may be present as a minor component of alkanolamide ingredients of cosmetic formulations. Skin penetration data were therefore generated for a range of typical formulations under in-use conditions. Seven rinse-off formulations (A-E, G and H), a leave-on emulsion (F), representing prototype cosmetic formulations and containing representative levels of DEA were prepared. Target levels of DEA were attained by inclusion of DEA as either (14)C-DEA or a combination of (14)C-DEA and unlabeled DEA. Skin permeation and distribution were evaluated using human skin in vitro, static diffusion cells and phosphate buffered saline (pH 7.4) as the receptor phase. At least 12 replicate epidermal membranes were prepared from a minimum of four donors for each test group. Receptor phase samples were taken at appropriate time intervals. At the end of the test period, radioactivity remaining on the skin surface and on the diffusion cell donor cap was determined before the skin samples were tape-stripped. The remaining tissue was solubilized and radioactivity determined. Permeation was very low from all vehicles applied under in-use conditions (range 1-48 ng/cm(2) over 24 h). Comparison was also made between permeation and distribution of DEA from an infinite dose of a simple aqueous solution and the leave-on formulation (F) through paired samples of fresh and frozen full thickness skin from the same donors. When applied as an infinite dose in aqueous solution DEA permeation at 24 h was greater through frozen than through fresh skin. From the leave-on formulation, permeation was similar and very low for both fresh and frozen skin. Recovery of DEA after application of the aqueous solution to fresh human skin and subsequent aqueous and organic extraction of the epidermal and dermal tissue indicated that the majority (>98%) of DEA was in the aqueous extract, suggesting that DEA was in the free state and not associated with the lipid fraction. These data provide a basis for the estimation of the potential systemic exposure and safety margins for DEA in representative cosmetic formulations.

Pharmaceutical applications for molecularly imprinted polymers

Molecular imprinting is a means of introducing sites of specific molecular arrangement into an otherwise uniform polymeric matrix. This is achieved by formation of a pre-polymerisation complex between complementary monomers and the template molecule. Subsequent polymerisation in the presence of a crosslinker, in a porogenic environment, results in the production of a macroporous polymer capable of specific molecular recognition. This paper considers potential roles for molecularly imprinted polymers within a pharmaceutical remit. Applications including controlled release, drug monitoring devices and biological receptor mimetics are discussed. Histamine and ephedrine molecularly imprinted polymers (MIPs) were studied as potential biological receptor mimics whilst a propranolol MIP was investigated for its use as a rate attenuating selective excipient in a transdermal controlled release device. Preliminary studies concerning the preparation of a theophylline selective transcutaneous monitoring device, using a theophylline MIP, are also described.

Differential permeation of propranolol enantiomers across human skin in vitro from formulations containing an enantioselective excipient

This work tested the hypothesis that a stereospecific topical formulation could be used to engineer differential permeation rates for each enantiomer of an applied racemate across human skin in vitro. Racemic and enantiomerically pure R or S propranolol HCI were formulated with cellulose tris(3,5-dimethyl phenyl carbamate) (CDMPC) and applied to excised human skin using side-by-side Franz-type diffusion cells. When the pure enantiomers were used, there was a marked difference between the penetration rates of R and S propranolol (flux ratio: 2.06; P = 0.04). When racemic propranolol was used, the difference was reduced, although still statistically significant (flux ratio: 1.2; P = 0.08), particularly in view of the differential activities of the two enantiomers. Control experiments, in which no CDMPC was present, produced equal permeation rates. The results can be rationalised in terms of differential adsorption onto CDMPC within the vehicle, whereby S-propranolol is preferentially bound relative to R-propranolol. This causes an imbalance in the apparent donor phase concentrations that (in accordance with Fickian diffusion laws and thermodynamic activity) gives rise to differences in permeation rates. The diminished differential observed when the racemate was used, rather than individual enantiomers, is less easily rationalised. In this work, it was the permeation of the eutomer (S-propranolol) that was retarded, although the general principle of stereoselectively retarded skin permeation has been established.

Comparison of the transdermal delivery of estradiol from two gel formulations

OBJECTIVE

Conventional oral oestrogen replacement therapy can relieve postmenopausal symptoms but is associated with undesirable side-effects which can be minimised by avoiding the fluctuating hormonal blood levels resulting from oral therapy and eliminating hepatic first-pass metabolism by the use of the transdermal route. The two commercially available transdermal gel formulations differ in composition and application recommendations. Sandrena Gel contains 0.1% (w/w) and Oestrogel 0.06% (w/w) estradiol and recommended dosages are 0.5-1.5 g over 200-400 cm2 (Sandrena Gel) and 2.5 g gel over 720 cm2 (Oestrogel). In transdermal therapy the formulation composition may have a significant effect on drug delivery and we have therefore compared the permeation of estradiol from these formulations across human skin in vitro.

METHODS

The in vitro percutaneous penetration of estradiol from the formulations through epidermal membranes prepared from excised female human thing skin was assessed over a 24 h period using static type Franz diffusion cells.

RESULTS

Permeation of the active was similar from each formulation representing (at 24 h) 18.2 +/- 3.5% of the applied dose from Sandrena Gel and 17.4 +/- 4.8% of the applied dose from Oestrogel. These percentages equate to cumulative skin permeations of 0.65 +/- 0.15 microgram/cm2 and 0.45 +/- 0.15 microgram/cm2 respectively.

CONCLUSION

The results suggest that the two formulations are bioequivalent at the recommended dose levels.

Percutaneous penetration of octyl salicylate from representative sunscreen formulations through human skin in vitro

The human skin penetration of [14C]octyl salicylate from two representative sunscreen vehicles was determined in vitro. 3H-sucrose was incorporated into all formulations and provided a marker for membrane integrity. When applied as a finite dose in an oil-in-water emulsion vehicle containing 5% (w/w) octyl salicylate, the average total absorption of 14C over 48 hr was 0.65+/-0.16% of the applied dose (representing a total amount permeated of 1.58+/-0.36 microg/cm2). When applied as an infinite dose in the oil-in-water emulsion vehicle the average total absorption of 14C over 48 hr was 0.47+/-0.22% of the applied dose (representing a total amount permeated of 27.54+/-13.91 microg/cm2). When applied as a finite dose in a representative hydroalcoholic formulation containing 5% (w/w) octyl salicylate, the average total absorption of 14C over 48 hr was 0.59+/-0.09% of the applied dose (representing a total amount permeated of 1.58+/-0.25 microg/cm2). When applied as an infinite dose in the hydroalcoholic formulation the average total absorption of 14C over 48 hr was 0.23+/-0.05% of the applied dose (representing a total amount permeated of 11.28+/-2.55 microg/cm2). The penetration of [14C]salicylic acid [applied at a concentration of 2.7% (w/w), in the oil-in-water emulsion] was also determined. When applied as a finite dose the average total absorption of 14C over 48 hr was 1.14+/-0.23% of the applied dose (representing a total amount permeated of 1.65+/-0.39 microg/cm2). These results suggest that the in vitro human skin permeation of octyl salicylate is relatively low. The amounts of octyl salicylate and salicylic acid permeated when applied in similar vehicles were remarkably similar over 48 hr (1.58 microg/cm2 and 1.65 microg/cm2, respectively). This suggests the possibility that the 14C label appearing in the receptor fluid may, in both cases, represent salicylic acid. If this is the case, then it is possible that the amount of octyl salicylate permeating through the skin is much less than that suggested by the data obtained here. This supposition is, however, entirely speculative and has yet to be confirmed experimentally.

Percutaneous penetration of N-nitroso-N-methyldodecylamine through human skin in vitro: application from cosmetic vehicles

The human skin penetration of N-nitroso-N-methyldodecylamine (NDOMA) from isopropyl myristate (IPM) and two vehicles representative of cosmetic/personal care formulations was determined in vitro. When applied as an infinite dose in IPM (1 microgram/microliter) the average total absorption over 48 hr was 0.10 +/- 0.01% of the applied dose (all data are expressed as means +/- SE). When applied as a finite dose in a representative oil-in-water emulsion formulation the average total absorption over 48 hr was 4.66 +/- 0.76% of the applied dose. When applied as a finite dose in a representative shampoo formulation for 10 min, followed by rinsing (to represent in-use exposure conditions), the average total absorption over 48 hr was 0.75 +/- 0.17% of the applied dose. Approximately 72% of the NDOMA in the applied shampoo formulation was removed by rinsing. The overall data indicated that NDOMA could penetrate the skin but that penetration was low. The rate and extent of absorption, however, could be affected by differences in the vehicle of application, time of exposure and whether the formulation is (and the conditions are designed to mimic) a rinse-off or leave-on product.

Percutaneous penetration of dimethylnitrosamine through human skin in vitro: application from cosmetic vehicles

Human skin penetration of N-dimethylnitrosamine (DMN) from three vehicles has been determined in vitro. When applied as an infinite dose in isopropyl myristate (IPM, 1 microgram/microliter) the average total absorption over 48 hr was 2.6 +/- 1.2% of the applied dose (all data presented are expressed as means +/- standard errors). When applied as a finite dose in a representative oil-in-water emulsion vehicle the average total absorption over 48 hr was 4.0 +/- 0.3% of the applied dose. When applied as a finite dose in a representative shampoo vehicle for 10 min followed by rinsing (i.e. to represent in-use exposure conditions) the average total absorption over 48 hr was 1.1 +/- 0.1% of the applied dose. Approximately 72% of the DMN in the applied shampoo vehicle was removed by rinsing. There was considerable evaporative loss of DMN from the IPM and oil-in-water emulsion vehicles, such that absorption was complete within 3 hr of application. The overall data indicate that DMN can penetrate the skin rapidly but that in practice the amount actually available for penetration is significantly reduced by high permeant volatility. In contrast, application of N-nitrosodiethanolamine (NDELA) at a concentration of 1 microgram/microliter as an infinite dose generated an average total absorption over 48 hr of 23.6 +/- 6.4%, representing a total flux of 103.9 +/- 28.4 micrograms/cm2. In the case of NDELA, no evaporative loss was evident.

Use of ATR-FTIR spectroscopy to study the diffusion of ethanol through glycerogelatin films

The use of ATR-FTIR spectroscopy to study the permeability of a glycerogelatin film is described. Measurement of the diffusion coefficient of ethyl alcohol-d in the film showed excellent reproducibility. Comparison of results from this technique with those previously obtained using an air-flow receptor phase diffusion cell show good agreement in terms of lag time assessed diffusion coefficients. ATR-FTIR spectroscopy revealed time-dependent changes in the composition of the glycerogelatin film during the diffusional process. It was also demonstrated that the concurrent assessment of both diffusant penetration and film composition is feasible.

Does solute stereochemistry influence percutaneous penetration?

The stratum corneum, the rate-limiting barrier to percutaneous penetration, is made up of several components, principally keratin and ceramides. These are potential sources of chiral discrimination that could result in differential diffusion rates, dependent upon the stereochemistry of the solute. Although binding to keratin can occur it is not a stereoselective process [percent binding to solubilised epidermal keratin: (R)-propranolol 7.9 +/- 1.7, (S)-propranolol 8.3 +/- 2.0]. On the other hand, studies with ceramide monolayers produced qualitative evidence of dose-dependent stereoselective interaction when the pure diastereomers of ephedrine were present in the aqueous subphase which suggested that differences in diffusion rates might occur in skin. However, the differences in permeation rates in vitro for these diastereomers through human skin were not statistically different [(+)-(1S,2R)-ephedrine 119.1 +/- 2.6 micrograms/cm2, (-)-(1R,2S)-ephedrine 107.0 +/- 3.9 micrograms/cm2, 12 h]. Time averaging, involving contributions from binding to all lipid headgroups present in the intercellular channels, may obscure specific differential interactions. Further, any stereospecific interaction may be subtle and readily overwhelmed if diffusant concentration is greater than the capacity of the skin to differentiate between stereoisomers. Evidence for intrinsic stereoselectivity in skin permeation has therefore yet to be obtained.

Steroidal Sapogenins from Tacca leontopetaloides

Diosgenin and its ring-F-hydroxylated derivatives isonarthogenin {spirost-5-ene-3,27-diol {3beta, 22 R, 25 S)} and isonuatigenin {spirost-5-ene-3,25-diol (3beta, 22 R, 25 S)}, together with the 22,25-epoxyfurost-5-ene isomer nuatigenin {furost-5-ene-3,26-diol-22,25-epoxy (3beta, 22 R, 25 S)}, were identified as the major steroidal sapogenins of the acid hydrolysate of an extract of leaves of TACCA LEONTOPETALOIDES by IR, (1)H-NMR, (13)C-NMR, and mass spectroscopy. A diosgenin ester was isolated which apparently arose by ring closure of a glycosylated furostene during hydrolysis.

The effect of gelatin grade and concentration on the migration of solutes into and through glycerogelatin gels

The diffusion of 4-hydroxybenzoic acid and phenobarbitone through glycerogelatin gels was found to be independent of the type of gelatin used. Three types of gelatin, two acid-processed and one alkali-processed were studied, and the bulk viscosities of gels prepared from them was seen to vary considerably. However, the microviscosities of the gels, as measured by ESR, showed no significant differences. Thus microviscosity was the factor governing diffusion. Gelatin concentration in aqueous solutions without glycerol influenced microviscosity and hence diffusion. This is believed to be caused by dissolution of water-soluble fractions of the gelatin. Interstices in the gelatin matrix, though reduced in size when gelatin concentration is raised, are still too large to act as physical barriers to diffusing molecules. It is suggested that hydrated gelatin forms the matrix of glycerogelatin mixtures and that the interstitial fluid, through which migration occurs, consists almost entirely of glycerol and water.

Evidence for an alternative route from sterol to sapogenin in suspension cultures from Trigonella foenumgraecum

Incorporation of [4-(14)C]cholesterol, [26-(14)C]-cholesterol, [4-(14)C]sitosterol and [22,23-(3)H]-sitosterol into sapogenin in suspension cultures from Trigonella foenumgraecum differed when substrates were added at subculture and 10 d after subculture, suggesting alternative biosynthetic routes were in operation.

Time course of metabolism of aldrin and dieldrin by suspension cultures

Time course, up to 100d, of uptake and metabolism of aldrin and dieldrin added at subculture to suspension cultures from Phaseolus vulgaris (French bean) root and shoot, and Solanum tuberosum (potato) tuber comparable, with rapid dieldrin production and delayed appearance of other metabolites. When aldrin and dieldrin not added to Phaseolus cultures until 10 or 20d after subculture usual extent of conversion of aldrin to dieldrin, but reduced production of other metabolites, and growth inhibition negated. Increasing volumes of 2-methoxyethanol had detrimental effect on growth and uptake and metabolism. Dieldrin production maximal during rapid growth phase and probably independent of other conversions.

Consistency of metabolism of aldrin and dieldrin in suspension cultures from Phaseolus vulgaris root

Reproducibility of metabolism of aldrin and dieldrin in cell suspension cultures from Phaseolus vulgaris (French Bean) root, and effects of variations in hormonal supplementation and light, studied. Aldrin always converted to dieldrin, aldrin-transdihydrodiol, and other unidentified more polar metabolites, but not to dihydrochlordene-dicarboxylic acid. Dieldrin always converted to photodieldrin and more polar metabolites, but not to aldrin-transdihydrodiol. Metabolism reasonably consistent from one subculture to next, basically unaffected by alterations in medium tested, or by presence or absence of light. Variations observed quantitative rather than qualitative supporting validity of use of tissue culture in metabolic studies on xenobiotics. Photodieldrin shown to be a direct product of dieldrin metabolism and not the result of microbial or environmental interference.

Metabolism of [(14)C]Flamprop-isopropyl in suspension cultures from Medicago sativa and Phaseolus sp

Metabolism of the selective herbicide flamprop-isopropyl (isopropyl-N-benzoyl-N-(3-chloro-4-fluorophenyl)-2-aminopropionate) in suspension cultures of the dicotyledons Phaseolus vulgaris, P. multiflorus and Medicago sativa showed general similarities be, but particular variations from, that reported in intact plants of Hordeum vulgare. Hydrolysis of the ester and conjugation were major routes, but some hydroxylation and hydrolysis of the amide was also observed. The implications of these variations are discussed.

Causes and effects of sample loss using a dry capsule injection method for the automated gas chromatographic analysis of pesticide samples

The automatic capsule sampling system is said to have many advantages for accurate gas chromatographic analysis. One is that the solvent may be removed by evaporation before injection, so that no solvent peak is obtained to interfere with the separation. However, working with aldrin and dieldrin, we have shown that up to 50% of the material can be lost from the capsules by air-drying at normal ambient temperature before injection. These losses occurred only with aldrin and it is suggested that they are connected with its relatively high vapour pressure. Significant amounts of material enter the top third of the capsule which is removed during the sealing process. The losses appear to be due to a combination of adsorption to the capsules, creep-up the capsule walls, co-distillation with the solvent, and direct volatilisation. Silylation of the capsules was quite effective in reducing these losses.

Application of densitometry to the qualitative and quantitative evaluation of pharmaceutical colourants

Samples from all the British manufacturers of food grade Amaranth and Sunset Yellow FCF were examined by a densitometric thin-layer chromatographic procedure which provides a rapid and convenient method for producing a qualitative and quantitative statement on the impurity profiles of dye materials. Wide variations in the impurities were found which could account for variations in toxicological results. The method could be incorporated into official specifications to control more precisely the quality of food grade dyes.