386 JOURNAL OF COSMETIC SCIENCE
contrast, with nonionic and zwitterionic surfactants which are milder towards proteins,
lipid interactions may play a dominant role.
Recently, in vitro systems based on reconstructed skin equivalent models have also become
popular for assessing the skin irritation potential of surfactants and other actives.43–45
Walters et al. used the 3-D EpiDerm™ model system to evaluate tissue viability and
primary cytokine interleukin-1α release to evaluate the potential dermal irritation of 224
nonionic, amphoteric and/or anionic surfactant-containing formulations, or individual
raw materials (see Figure 5).45 The authors showed a correlation between in vivo TEWL
measurements in a patch test and the IL1 alpha release in in-vitro studies (see Figure 6).
The results presented are consistent with the prevailing understanding that the order of
irritation potential follows: anionic amphoteric nonionic. The reasons for differences
within each category were not discussed in this paper. However, the data from such a large
list of surfactants is certainly worth exploring further. Since the quality of the SC barrier in
the EpiDerm model system is relatively weak compared to the real human SC, the results
by such tests systems will be indicative of the inherent irritation potential of an ingredient
rather than in a real-life situation in subjects with healthy SC under normal use conditions.
Since patch tests are rather exaggerated and enhance penetration under an occlusive patch,
the results from such reconstructed models may be reflective of the situation in subjects
with a compromised skin barrier.
STRUCTURE-FUNCTION RELATIONSHIPS GOVERNING SURFACTANT-
INDUCED SKIN IRRITATION
The various test methodologies described above are useful in predicting the irritation
potential of surfactants. It is important at this stage to go beyond just predicting irritation
potential to understanding structure-function relationships governing irritation potential
of surfactants. A simplistic analysis by looking at the trends in Figures 3–6 suggests that
the charge and the size of the surfactant head group plays a role in the irritation potential
Figure 5. In vitro assessment of skin irritation potential of surfactant-based formulations by using a 3D skin
reconstructed tissue model and cytokine response. Figure reproduced from Walters et. al.45

387 The Human Stratum Corneum
of surfactants, with anionic surfactants showing higher irritation tendency compared to
amphoteric and nonionic surfactants. This is consistent with the pragmatic approach
that formulators have been using over the years to increase mildness towards skin by
increasing the size of the polar head group. For example, it is well known that harshness
of ethoxylated sulfates follow the order: SLS SLES 1EO SLES 2 EO SLES 3EO.
This can be generalized further with more quantitative relationships, relating headgroup
structural parameters to protein denaturation tendency.
The molecular mechanism involved in denaturation of proteins by surfactants in aqueous
solutions is thought to be due to the binding of surfactants to the proteins resulting in
the formation of micelle-like structures on the protein backbone and causing significant
electrostatic repulsion within the network.34 This in turn unravels the tertiary structure
of proteins. This charging of the structure in the case of insoluble proteins such as zein
results in dissolution of the protein. In contrast, in the case of crosslinked proteins such
as keratin, this results in osmotic driven swelling of proteins. Based on this hypothesis,
Lips et al. examined a correlation between micelle charge density and the dissolution of
zein using a variety of surfactants and found a linear correlation between micelle charge
density and the protein dissolution.34 Micelle zeta potential also correlated linearly with
the protein dissolution. Since the zeta potential of micelles is related to its charge density,
both the parameters correlating with the dissolution are not surprising, and overall, these
correlations support the hypothesis that micelle charge density is a predictive measure of
the denaturing potential, and hence the irritation tendency of the surfactants.
The role of micelle charge density as a key predictive parameter is further evident from the
work of Morris et al., which showed that the penetration of anionic surfactants into human
skin as measured from Franz diffusion experiments correlated linearly with the micelle
Figure 6. The correlation between cytokine release in-vitro and clinical TEWL measurement in a patch test.
Data from Walters et al. TEWL measurements are from in-vivo patch testing.45