An ABCC11 Variant Reduces Underarm Odor, but Does Not Eliminate It

ABCC11 is known for its role in underarm odor. One variant is associated with lower amounts of substances that become odor compounds. But does this mean people can barely smell the resulting underarm odor?

A 2014 study examined genotype alongside trained assessors’ odor ratings, skin compounds, bacteria, and product-use habits. It found that underarm odor in the A/A group was weaker than in the other groups, but remained perceptible, and short-chain fatty acids were also detected.

Connecting genes, compounds, and actual odor

ABCC11 encodes a protein involved in transporting substances out of cells. The paper examined whether one position in DNA contained G or A. The combinations inherited from the two parents were classified into three genotypes: A/A, G/A, and G/G.

This variation is also associated with earwax type. Earlier studies had found that people with the A/A genotype had fewer odor precursors secreted by the underarm apocrine glands. Precursors are materials that do not themselves have a strong odor but can be converted into volatile odor compounds through the action of skin bacteria and other processes.

The study asked how closely this smaller supply of starting material corresponded to weaker odor perceived by people. It also examined odors arising from other materials, bacterial composition, and relationships with everyday grooming.

Selecting 25 people from each of three genotypes

The researchers initially screened 200 women aged 18–40 in Manila, the Philippines, and analyzed genotypes and questionnaires from 164. These comprised 46 A/A, 87 G/A, and 31 G/G participants. They randomly selected 25 from each genotype, 75 in total, for detailed measurements.

Earwax type was checked during recruitment. Once 60 people with dry earwax had been recruited, no more people with that type were added. Consequently, the proportions of the three genotypes in this sample cannot be read as genotype frequencies for the wider region.

For odor assessment, the underarms were washed in a standardized way. Six trained assessors rated both underarms on a 0–5 scale at 5 and 24 hours after washing. Assessors were not told participants’ genotypes, and measurements were made over four consecutive days.

A/A odor was weaker, but still perceptible

Mean ratings 5 hours after washing were 2.59 for A/A, 3.26 for G/A, and 3.21 for G/G. At 24 hours, the corresponding values were 2.60, 3.40, and 3.50. At both times, A/A ratings were significantly lower than those for the other two genotypes, while the G/A–G/G difference was not significant.

The authors emphasize that the A/A ratings still represented odor perceptible to other people. Having fewer particular precursors was not the same as eliminating underarm odor as a whole.

Fewer precursors, but short-chain fatty acids in all three groups

Liquid chromatography and mass spectrometry of samples collected by washing the skin with water showed lower levels of two precursors associated with the odor compounds HMHA and 3M2H in the A/A group. They had not disappeared completely: small amounts remained detectable. A precursor of the sulfur-containing compound 3M3SH was not detected in A/A participants.

An analysis comparing broader patterns of skin metabolites also tended to separate A/A participants from the two G-containing groups, with those two precursors contributing substantially to the difference.

Next, patches were used to collect volatile compounds from the underarm skin by adsorption, followed by gas chromatography and other analyses. The targeted sulfur-containing compounds and 3M2H were not detected in the A/A group but were detected in many G/A and G/G participants.

In contrast, short-chain fatty acids such as butyric, isovaleric, valeric, and caproic acids were detected in all three groups, with no difference in detection frequency. This does not mean their quantities were identical, but it supports the existence of odor-producing pathways beyond the particular precursors associated with ABCC11.

The authors suggest that fatty acids produced when bacteria break down amino acids from sweat or skin were leading candidates for the odor perceived in A/A participants. Because they did not test odor after removing individual compounds, the relative contributions of specific compounds remain unconfirmed.

Bacterial composition and product-use habits were also associated

Staphylococcus, Corynebacterium, and Anaerococcus accounted for large proportions of the skin bacterial community, with many bacterial groups shared across participants. Within that pattern, the A/A group had a lower proportion of Staphylococcus and a higher proportion of Corynebacterium than the G-containing groups. The difference in Anaerococcus was not significant.

A higher proportion of odor-related bacteria did not simply translate into stronger overall odor. The authors consider the amount of precursor material available to bacteria important alongside differences in bacterial composition. The study measured associations between groups, rather than replacing bacteria experimentally to see how odor changed.

Among the 164 questionnaire respondents, regular antiperspirant or deodorant use was reported by 50% of the A/A group, 86% of the G/A group, and 97% of the G/G group. The authors suggest that awareness of one’s own odor may influence product-use habits.

However, reasons for use can also include managing sweat, personal preferences, and social habits. Because the study did not directly measure how accurately people perceived their own odor, it cannot establish that genotype determines product-use behavior.

One genotype does not determine the whole odor picture

This was a comparison of women from one region. It did not manipulate genes experimentally, and the numerical results cannot simply be generalized to populations differing in age, living conditions, or grooming practices.

The paper linked ABCC11 variation with particular precursors, bacterial community composition, and product use. At the same time, it demonstrated that underarm odor arises through multiple materials and pathways, so a single genetic characteristic cannot divide people into odorless and odorous categories.

Summary

The A/A group had lower mean underarm-odor ratings than the G/A and G/G groups, along with lower levels of particular precursors and volatile compounds. Nevertheless, their odor remained perceptible, and short-chain fatty acids were detected in all three groups.

Associations were also found with skin bacterial composition and the proportion using products. Understanding body odor requires looking beyond genes to the substances secreted and the activity of bacteria. Connecting chemical analysis with actual smelling was an important contribution of this study.

Most authors were affiliated with Unilever at the time, with additional authors affiliated with Waters and a clinical-trials company. The original paper’s funding section disclosed these employment relationships and stated that the funders were not directly involved in data collection or manuscript preparation. A separate conflict-of-interest section could not be identified.

References

Harker M, et al. Functional characterisation of a SNP in the ABCC11 allele—Effects on axillary skin metabolism, odour generation and associated behaviours. Journal of Dermatological Science. 2014;73(1):23–30. Original article.