Molecular Modeling of Skin care Products: Application for the Design of Pepidomimetics

May 1, 2008 | Contact Author | By: Jean-Francois Nicolay, PhD, Exsymol S.A.M. and Mindy Goldstein, PhD, Estée Lauder
Your message has been sent.
(click to close)
Contact the Author
This item has been saved to your library.
View My Library
(click to close)
Save to My Library
Title: Molecular Modeling of Skin care Products: Application for the Design of Pepidomimetics
  • Article
  • Media
  • Keywords/Abstract

Editor’s note: Our regular columnist Mindy Goldstein, PhD, welcomes the following “Tech Edge” contribution from colleague Jean-Francois Nicolay, PhD, from Exsymol S.A.M.

From the column editor: Molecular modeling has been used in the pharmaceutical industry for years to predict how a new molecule may work based on research conducted with molecules of a similar structure. This approach can help to reduce the costs of determining which new molecules to test and target for commercialization. The cosmetics industry has also begun to use this technology to predict the behavior of new molecules on skin and in skin care applications. This month, I welcome Jean-Francois Nicolay, PhD, from Exsymol S.A.M. to present an overview of molecular modeling and its application in skin care. His educational background focused both on biology and chemistry, culminating in a doctoral degree in organic chemistry. He then joined the pharmaceutical company Mayoly-Spindler as a medicinal chemist, where he worked on morphine derivatives, antifungal macrolides and therapeutic peptides. After a short passage in the diagnostic industry, he joined Exsymol S.A.M. in 1991 as the head of its chemistry department. In this role, Dr. Nicolay has applied the techniques of medicinal chemistry to the design of cosmetic active ingredients. He became Exsymol’s research manager in 1999.

There is increasing demand for comprehensive data about the destiny of a cosmetic product following its application to the skin. For instance, data on percutaneous absorption is needed for efficient risk assessment to determine systemic exposure. This data is also very useful for efficacy assessment of skin care products.

The bioavailability of an active ingredient—i.e., its ability to reach a biological target within the skin—must be compatible with the claimed activity. In fact, the epidermis forms an effective barrier, and penetration into the skin is difficult to achieve. A painful reminder of this is the fact that a needle is necessary for the subcutaneous injection of a vaccine or a drug. Formulations that are adapted with barrier-disrupting additives can help to enhance penetration but undesirable “guests,” such as the constituents of the preservative system, may also take advantage of this permeation.

An alternative strategy is tailoring the active ingredient to improve its bioavailability. Molecular modeling, structure-activity relationship studies (SAR) and metabolization estimations are new approaches in the cosmetic field that can bring important benefits. Application of these techniques to peptides is of particular interest since they are endowed with many valuable biological properties but unfortunately have low bioavailabilities when topically applied. 

Tools for Molecular Modeling

Molecular modeling software: A number of software programs for the prediction of physicochemical parameters on the basis of the molecular structure have been developed.1,2 These so-called expert systems display predictive endpoints useful for structure-permeability relationship studies, such as ionization constants, water solubility and the octanol/water partition coefficient. Log P, the logarithm of the partition coefficient, or Log D, the pH-dependent coefficient for ionizable compounds, are key parameters for percutaneous absorption estimation.3 Roughly, negative Log P/Log D are unfavorable to penetration, while neutral to positive values (0, +3) are more favorable. 

The aim of in silico chemistry is to provide a fast estimation of the behavior of candidate ingredients when applied to skin. In silico modeling has been carried out by pharmaceutical companies for years in order to limit late-stage failures during drug development. Regulatory agencies recommend the use of expert systems for early detection of potential toxicological problems,4 which is quite indicative of the quality of their prediction.

3-D in vitro models: Commercially available human reconstituted epidermis (HRE), obtained from normal keratinocytes, presents high histomorphological and biochemical homologies with the normal epidermis. HRE has a fully differentiated stratum corneum and a well-defined barrier function.5 This easy-to-use, standardized material has much less variation than skin explants and can be adapted on diffusion cells for penetration studies (see Figure 1).

The stratum corneum predominantly controls penetration, thus HRE enables a rapid preliminary indication about the bioavailability of a candidate cosmetic ingredient.

Reconstituted skin or full thickness models consisting of a stratified epidermis superimposed on a collagen lattice are dermal-equivalent and now commercially available. Such models make it possible to examine the permeation of candidate ingredients and the distribution of the product, since the epidermis can be separated for the dermal equivalent following the permeation test.

Metabolization estimation: Products applied at the surface of the skin can be metabolized by extra-cellular or intra-cellular enzymes upon penetration or permeation. This may result in a rapid loss of both the biological availability and activity. In order to obtain early information about the sensitivity of compounds to cutaneous enzymes, in vitro experiments can be conducted with multi-enzymatic mixtures. Such substrate competition tests (see Figure 2) are procedures adapted to peptide metabolization studies.

In such tests, the candidate peptide—for example, in Figure 2, a dipeptide derivative—and a selected chromogenic probe are simultaneously exposed to a multi-enzymatic mixture. The hydrolysis of the probe is monitored by spectrophotometry. A strong signal, indicative of an important hydrolysis of the chromogenic probe, shows that the candidate peptide did not interfere with the hydrolysis and is not sensitive to proteolytic enzymes. 

A weak signal indicates that the peptide was hydrolyzed. Kinetic parameters such as the Michaelis-Menten constant (KM) and the maximum reaction velocity (Vmax), useful for making comparisons, can be obtained from this experiment. These parameters are obtained from the Lineweaver-Burke equation. KM roughly is an inverse measure of the affinity of an enzyme for its substrate. 

Modeling of Peptidomimetics

Molecular modeling of peptides is an attractive topic because these biomolecules generally have a low bioavailability. There are a number of reasons for this, including:

Very negative Log Ds—many bioactive peptides are hydrophilic;

High molecular weight: A tetrapeptide easily reaches the limit of 500 d; and

Short biological half-life: Peptides are readily hydrolyzed by cutaneous hydrolytic enzymes.

Thus, structural modifications or the synthesis of peptidomimetics that are related structurally to the parent peptides, can greatly improve the efficacy of topically applied peptides.

A representative example is the molecular modeling of the natural dipeptide L-carnosine. This very small peptide (see Figure 3) is a water-soluble, hydrophilic antioxidant, capable of preventing lipid peroxydation in vitro.6

Its bioavailability is very low due to its sensitivity to a specific dipeptidase, carnosinase, and nonspecific cutaneous dipeptidases—arylamidases.

Removal of the carboxylic acid moiety of L-carnosine (see Figure 3) yields a low molecular weight peptidomimetic that is more hydrophobic than its parent compound. Predictive Log Ds at physiological pH obtained from an expert systema accounted for a 100-fold hydrophobicity improvement: Log D (carnosine) = -5.5; Log D (decarboxycarnosine) = -3.3.

Nevertheless, this data underlines that the peptidomimetic remains strongly hydrophilic, with a negative Log D.

A preliminary penetration study was then conducted on HRE, showing effective access to the epidermal compartment but no transepidermal diffusion. This was confirmed by immunohistolocalization (see Figure 4), with a staining of the full thickness of the HRE.

The substrate competition test conducted on L-carnosine, decarboxycarnosine and a reference dipeptide, L-alanyl-L-histidine, gave the most striking results. The metabolization study showed that L-carnosine and related natural dipeptides are readily hydrolyzed by a pool of enzymes obtained from keratinocytes (see Figure 5). 

At the opposite, the decarboxylated peptidomimetic proved to be highly resistant to enzymatic hydrolysis—very limited interference was detected with the hydrolysis of the probe.

The data obtained from the modeling tools was thus concluded to provide a better understanding in regard to the impact of a limited structural modification or decarboxylation on the bioavailability of a bioactive dipeptide. In this example, the modification provided a major improvement against in situ metabolization responsible for a loss of activity. 

This approach of combining modeling tools highlights that several parameters must be taken into account when considering the bioavailability of a candidate compound; not only lipophilicity, but also the molecular weight and metabolization.

Further experiments have confirmed this preliminary data. A penetration study on skin explants has shown that formulated decarboxycarnosine has a good bioavailability for the epidermal compartment. The use of penetration enhancers may enable access to the dermis.

General Conclusions

Data on the percutaneous absorption of skin care products is important for a better understanding of their in vivo efficacy. Such data may single out formulation constraints if a deep layer of the skin is targeted; such as the need for added penetration enhancers. A new approach for the design of innovative skin care products includes modeling their bioavailability for optimum in vivo efficacy. 

Peptides are trendy active ingredients that have low bioavailability, e.g. low in vivo efficacy, when topically applied. Molecular modeling of peptidomimetics is an attractive issue for this class of compound. 

Among the tools available for molecular modeling, predictive software or expert systems certainly are the most powerful. Interestingly, expert systems also predict toxicity such as acute toxicity and genotoxicity. Early detection of toxicological problems also helps in the development of new products. 

Expert systems will likely advance tremendously in upcoming years. Predictive tools of the future will become skin-specific, focusing on different routes of penetration, as well as display additional parameters.8 Advanced tools and molecular modeling in general enable formulators to design safer and more reliable cosmetic ingredients for personal care. 


1. GP Moss, JC Dearden, H Patel and MTD Cronin, Quantitative structure permeability relationships (QSPRs) for percutaneous absorption, Toxicol In Vitro 16 299–317 (2002)

2. S Geinoz, RH Guy, PA Carrupt and B Testa, Quantitative structure permeation relationships (QSPeRs) to predict skin permeation. A critical review, Pharm Res 21 83–92 (2004)

3. J Arct, Lipophilicity and the percutaneous absorption in cosmetics SOFW J 129 2–9 (2003)

4. MTD Cronin, JS Jaworska, JD Walker, MHI Comber, CD Watts and AP Worth, Use of QSARs in international decision-making frameworks to predict health effects of chemical substances, Environ Health Perspectives 111 1391–1401 (2003)

5. N Garcia, O Doucet, M Bayer, D Fouchard, L Zastrow and J-P Marty, Characterization of the barrier function in a reconstituted human epidermis cultivated in a chemically defined medium, Int J Cosmetic Sci 24 25–34 (2002)

6. EA Decker and H Faraji, Inhibition of lipid oxidation by carnosine, JAOCS 67 650–652 (1990)

7. A Boldyrev, Hydrolysis of carnosine and related compounds by mammalian carnosinases, Comp Biochem Physiol B Biochem Mol Biol 127 443–446 (2000)

8. IT Degim, Understanding skin penetration: Computer aided modeling and data interpretation, Current Computer-aided Drug Design 1 11–19 (2005)                 



Figure 1. Schematic representation

Figure 1. Schematic representation  

Figure 2. Substrate competition test peptide sensitivity

Figure 2. Substrate competition test peptide sensitivity 

Figure 3. Chemical structure

Figure 3. Chemical structure 

Figure 4. Localization of decarboxy-carnosine

Figure 4. Localization of decarboxy-carnosine 

Figure 5. Subtrate competition test with L-carnosine

Figure 5. Subtrate competition test with L-carnosine 

Goldstein: Molecular Footnotes

a ADME is a product of Pharma Algorithms. 

Next image >