90 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS is now being recognized. Elias et al. (1) have proposed that these lipids form multiple layers in a lameliar liquid crystalline arrangement. Friberg and Osborne (2) developed a model stratum corneum lipid system, based on Elias' model, in which neutral and polar lipids are incorporated into a lameliar liquid crystalline host of a fatty acid/soap mixture (pH = 4.5-6). The complete model is a mixture of liquid crystalline and solid phases the physical state of the lipids in this model system is correlated with the maintenance of proper barrier function (3). A liquid crystalline state permits liquid-like diffusion across the bilayer, while a solid state permits rapid water loss via cracks. It has been proposed (3) that the mixed solid and liquid crystalline states produce the optimal barrier to water loss. The occurrence of dry skin associated with cold, dry weather may stem from an exten- sive, elevated level of skin lipids in the solid state. Therefore, a material that maintains a higher proportion of lipid in the liquid crystalline state may be an effective moistur- izer. Recently Froebe et al. (4) demonstrated that glycerol interacts with the model iipids to maintain the liquid crystalline state even at low relative humidity (RH). Thus glycerol may condition the skin by this alternative mechanism at low RH, since it did not exhibit humectant activity under these conditions. In the above studies, the liquid crystalline state was observed by polarized light mi- croscopy, while humectancy was determined gravimetrically from water evaporation. To provide a quantitative way of monitoring the liquid crystalline state, we now add the technique of differential scanning calorimetry (DSC), along with optical microscopy, to examine the effects of putative moisturizers at maintaining the liquid crystalline state of model stratum corneum lipids after incubating at 6% and 92% relative humidities for varying periods of time. Previous studies using high-sensitivity DSC of human as well as porcine stratum cor- neum (5-7) have shown three broad transitions at 65, 75, and 105øC. These have been attributed to thermal transitions of intercellular lipids, lipid-protein complexes, and intracellular keratin, respectively. If the model lipid system as proposed by Friberg and Osborne (2) is valid, we should expect to see thermal transitions due to the lipids only. It is well known that lipid transitions are due to changes in hydrocarbon chain packing within the bilayer endothermic changes are due to changes in packing from an ordered to a more disordered state as more gauche rotamers are formed. These phase transitions are hydration-dependent. For example, at low hydration, where the hydrophilic head groups are not fully hydrated, the hydrocarbon chains are more closely packed and require more energy for melting. Hence the transition will occur at a higher temperature and have a greater enthalpy, as reported earlier for dimyristoylphosphatidylcholine (8). In the present work, the enthalpy of the phase transition of the model lipid system was quantitatively determined by DSC, at 6% and 92% RH as a function of time and water loss the phase behavior of the model lipid at these humidities was also tracked by polarizing microscopy, which qualitatively determined the extent of crystal formation in a liquid crystalline matrix. The modification of the phase behavior of the model lipid system by maleated soybean oil, identified here as Glyceridacid, at 5-15 wt% concen- tration, has also been investigated. Glyceridacid is a modified triglyceride with the chemical name 2-(alkoyloxy)1-[(alkoyloxy)methyl]-ethyl-7-(4-heptyl-5, 6)-dicarboxy-2- cyclohexene-l-yl) heptanoate. It has previously been shown to penetrate the stratum corneum (9), and the skin softening properties of this compound have been reported

STRATUM CORNEUM LIPIDS 91 (10). The epidermal lipids were the primary site of action for Glyceridacid as a skin softener (10). The effects of glycerol on the model lipid were also investigated as a quantitative check on our previous work (4), which showed via microscopy that glycerol is able to maintain the liquid crystalline state of the model lipid at low humidity. This was suggested as an alternative mechanism to humectancy at low humidity for glycerol, which is known to prevent/reverse dry skin in vivo (11-13). MATERIALS AND METHODS Maleated soybean oil, the fumaric acid adduct of soybean oil, is marketed by Van Dyk (Belleville, NJ) under the tradename Ceraphyl GA. © This substituted triglyceride, identified here as Glyceridacid, was originally marketed by Westvaco under the trade- name Glyceridacid 100. © Phosphatidylethanolamine was obtained from Avanti Polar Lipids (Birmingham, AL), while the other lipids shown in Table I and glycerol were obtained from Sigma Chemical Company (St. Louis, MO). All lipids were of the highest grade (98-99%) and used without further purification. The composition of the model lipid, based on the analysis of stratum corneum lipids by Elias et al. (1), is shown in Table I. The procedure for preparing host lipid, and subsequent incorporation of other lipids to form the model lipid at 32% dehydration, is described elsewhere (2,4). The model lipid thus prepared consisted of two phases--a liquid crystalline phase and an excess liquid phase, presumably due to 24% triolein in the model lipid. This two-phase system is not adequate for DSC studies, and we have modified the model lipid system by a reduction of the triolein to 12%, resulting in almost a single phase. The new composition of the model lipid system is shown in Table Table I Composition of Model Epidermal Lipid (1) and Modified Model Epidermal Lipid* Wt% in mixture Wt% in mixture Component (model lipid) (modified model lipid) Free fatty acids 19.0 22.2 Oleic acid 33.1 33.1 Linoleic acid 12.5 12.5 Palmitic acid 36.8 36.8 Palmitoleic acid 3.6 3.6 Stearic acid 9.9 9.9 Myristic acid 3.8 3.8 Phosphatidylethanolamine 5.0 5.8 Cholesterol 14.0 16.4 Cholesterol sulfate 2.0 2.4 Triolein 25.0 12.0 Oleic acid palmityl ester 6.0 7.2 Squalene 7.0 8.3 Pristane 4.0 4.7 Ceramides (type III) 18.0 10.5 Ceramides (type IV) -- 10.5 * Model lipid was prepared containing 32% water.