Correspondence: Stefan Kippenberger, Department of Dermatology, Venereology and Allergy, Johann Wolfgang Goethe University, Theodor-Stern-Kai 7, D-60590 Frankfurt/Main, Germany, Tel.: +49 69 6301 7734, Fax: +49 69 6301 6466, e-mail: email@example.com
In today′s world, natural body odour is mostly considered as being unpleasant and combated by intensive cleansing, deodorants and perfumes. However, there is evidence that volatile body compounds provide the recipient with important information. Here, we present the chemical identity of odorous compounds derived from odourless precursors within sweat and sebum. Moreover, distinct volatile markers may be relevant for the clinical diagnosis of disease. Interestingly, ageing seems to correlate with the appearance of specific compounds that convey the so-called old man smell. Finally, it is discussed if human skin odour has the quality to act as pheromone transmitting information between individuals in terms of major histocompatibility complex type or reproductive status.
Skin Odour: nothing is more remarkable than a smell!
The sources of skin odour contributing to an individual ‘odour signature’ are diverse. A chief producer is the apocrine gland, located in the axillary, groin and anogenital regions, as well as in the umbilicus, eyelid (Moll's glands), areola and the external auditory meatus. They are fully developed up to reproductive maturity. The characteristic odour of anatomic sites rich in apocrine glands is formed from the interaction between odourless (water soluble) precursor molecules found in the glands secretion and the cutaneous microflora [1, 2]. In particular, aerobic Corynebacteria metabolize odourless steroids producing 16-androstenes (5α-androstenol, 5α-androstenone) with a pungent musk- and urine-like odour . It is, therefore, not surprising that men have been reported to have more numerous and larger apocrine glands than women . Correspondingly, axillary androstenone levels are much higher in men than in women . Moreover, volatile C6-C11 acids, the most prominent being 3-methyl-2-hexenoic acid (3M2H), are reported to contribute to axillary malodour . Interestingly, 3M2H is bound to carrier proteins (ASOP1, ASOP2) after secretion into the apocrine glandular lumen preventing the formation of smell. The characteristic goat-like axillary odour only appears after liberalization by bacteriolysis [7, 8]. There are also indications for a gender-specific odour signature: a cheesy, rancid odour derived from the metabolism of glutamine conjugates seems typical for men, whereas an onion-like smell from that of sulphur-rich conjugates seems typical for women . Very recently, it was found that a single nucleotide polymorphism (538G->A) in the gene ABCC11 is responsible for the transition from a strong axillary odour, as commonly present in Caucasians and Africans, to a faint acidic scent typical for Asians . Interestingly, this nucleotide substitution correlates with a white and dry earwax phenotype frequently (80–95%) seen among East Asians and being rare (0–3%) in populations from European and African origin .
In addition to apocrine glands, the eccrine sweat glands also contribute to skin odour. They are distributed all over the skin but particularly concentrated in the soles of the feet, the palms of the hands and the forehead. Eccrine sweat is mostly water, but also contains glycoproteins, lactic acid, sugars, amino acids and electrolytes  providing an ideal substrate for the growth of microorganisms. The sebaceous gland is also considered to be a player in skin odour formation. It was found that yeasts of the genus Pityrosporum when grown on a lipid substrate similar to human sebum produce γ-lactones recognized as having a fruity canned-peach odour . In particular, Pityrosporum ovale, the most frequent microorganism of the capillitium , may contribute to the typical human scalp odour, which is significant in smell recognition of newborns. Furthermore, it is speculated that human sebum acts as a concrete, a carrier compound that retards the liberation of odorous molecules .
Because of the diverse composition of odour producers within human skin, different body areas such as scalp, axillae or feet emit a specific smell which in toto generates a complex mixture of odorants. Tracking dogs, for example, trained to recognize a person′s scent on a garment worn on a particular body part are not reliably able to relate the person′s odour to other body parts .
Olfactory perception: smells like a disease
In times where the diagnostic arsenal of physicians was rather limited, patients′ odour served to identify ailments. The diagnosis by smell goes back to the observations of Hippokrates (460 BC – 370 BC), the famous physician of Ancient Greece, and was carried forward by Galenus (ca. 120 – 200) and Avicenna (980–1037). By mistake, in the Middle Ages, the smell itself was considered to be the cause of the disease resulting in the misguided attempts to combat diseases such as plague and typhus by carrying scented pouches or torches . It is not within the scope of this review to give a complete list of diseases associated with odour changes; instead, we would like to focus on some selected examples (see Table 1). For more detailed information please refer to the recent review by Shirasu and Touhara . For the dermatologist, the odour emitted by chronic wounds such as ulcus cruris provides clues about microbial colonization and hence the necessity for an antiseptic treatment which is reported to have a deodorizing effect . Moreover, bullous congenital ichthyosiform erythroderma, caused by mutations in keratins 1 and/or 10, is associated with a foul skin odour ; however, this criterion plays a minor part in diagnosis. Interestingly, animals are also reported to be useful in making a diagnosis by olfactory perception. Inspired by anecdotal findings where dogs, whose olfactory sensitivity is much superior to humans, alerted their owners to skin lesions that later became diagnoses such as melanoma  and basal cell carcinoma , more controlled studies were initiated. Recently, it was reported that a trained Labrador retriever accurately identified cancer patients by sniffing exhaled breath and stool samples . Besides the use of dogs, also an electronic nose, originally designed to monitor the air quality, was experimentally used to discriminate cancer cells in vitro . Moreover, the electronic nose is particularly useful to detect odours caused by pathologic microorganisms such as Helicobacter pylori .
Table 1. Metabolic and infectious diseases with typical odour
It is a common experience that body odour seems to change not only from childhood to puberty but also from middle ages to older age. The so-called ‘old lady smell', ‘old man smell' or ‘old person smell' is an idiom often used to describe an odour that is characteristically associated with the elderly. Decreased androgen production at an older age may contribute to a change in skin odour as the metabolic activity of apocrine and sebaceous glands is under control of androgenic hormones . Age-related changes can be perceived by others; the odour of postclimacteric women was frequently mistaken for that of men. On the other hand, unfamiliar smellers after being exposed to the same odours reported feeling relaxed . In an analytical study to identify volatile organic compounds that correlate with age using gas chromatography, three compounds (dimethylsulphone, benzothiazole and nonanal) were found to correlate with age . Interestingly, this extensive study failed to detect the unsaturated C9 aldehyde, 2-nonenal, which was reported by a Japanese group to increase with age . Haze et al. found a rapid increase of 2-nonenal in individuals older than 39 years. The compound was perceived to smell of orris, fat and cucumber. As in the first study mainly Caucasians participated, it seems likely that the conflicting data may be diet-linked reflecting a cultural phenomenon. Previous studies showed that diet can affect the perception of human body odour . For instance, consumption of red meat decreased the pleasantness′ of axillary odour . Unfortunately, the Japanese study does not tell us anything about the eating habits of the individuals tested; however, in this study, they found the ω7 monounsaturated acid, palmitoleic acid, which is a common constituent of fish oil , to be the origin of 2-nonenal. This supports the assumption that a specific diet is the cause of the occurrence of culture-specific odour compounds – also in old age. Furthermore, there is evidence that the occurrence of age-specific compounds in elderly persons is caused by lipid peroxidation of higher molecular weight unsaturated acids, which seems to be upregulated in older individuals [47, 48]. However, changes in physical activity, frequency and intensity of personal hygiene measures and changes in grooming habits with age may also contribute to the occurrence of an age-related odour.
Communication via pheromones – also in humans?
The term pheromone is a neologism composed of the Greek words pherein (transfer) and hormone (excite) . Karlson and Lüscher define pheromones ‘as substances which are secreted to the outside by an individual and received by second individual of the same species, in which they released a specific reaction, for example, a definite behaviour or a developmental process’. A third class, so-called information pheromones, has since been defined to describe substances that indicate the identity or territory of an animal . First described in insects, pheromones were also reported in mammals. In mice, for example, a complex mixture of volatile compounds within urine contributes to pheromonal communication affecting reproduction. Particularly, the major urinary proteins (MUPs), components of adult male urine, bind to volatile pheromones and regulate their release into the air from urine marks. MUPs are involved in a plethora of reactions including synchronization of ovulation in a group of anoestrous female animals (‘Whitten effect’) and puberty acceleration in juvenile females . Of note, humans seem to be the only mammals with no active MUP genes yet analysed .
Now, we will focus on the evidence for human chemical communication in various social domains. The most often referred example of human pheromones is menstrual synchrony. It was initially found in the early 1970s that female students sharing apartments tend to synchronize the onset of their menstrual cycle . Subsequently, this phenomenon was tested in several dozen highly variable samples ranging from women sharing offices , basketball players , lesbian couples  or Bedouin mother–daughter dyads  showing a mixed pattern of results. It was suggested that time spent together, emotional closeness and cycle regularity can modulate cycle synchrony . Martha McClintock  in her pioneering study speculated that menstrual synchrony could be driven by chemicals produced in female armpits. This was originally tested , but the study was methodologically flawed . More recently, Stern and McClintock  collected odourless fluids derived from the axillae of women in the late follicular phase of their menstrual cycle. It was found that exposure to recipient women shortened their menstrual cycle and accelerated the preovulatory surge of luteinizing hormone (LH). Leaving aside that the chemicals responsible for the effect are not known, it should be pointed out that a majority of the studies on synchronization was criticized for methodological flaws especially concerning how synchrony was computed [60, 62].
Moreover, it is not clear what would be the functional significance of such a phenomenon, although it could have evolved in a context of different mating systems (e.g. polygyny) in our ancestors and was simply not selected out in recent human populations. It is thought that in some polygynous mating systems, male mating capacities are restricted. Thus, women might compete for access to men which in turn can result in oestrous synchrony. However, evidence for menstrual cycle synchrony in polygynous primates is equivocal .
There are also clues for olfactory communication between both sexes. In an experiment similar to the above, it was found that underarm secretions from men have the ability to stimulate the onset of the next peak of LH . Furthermore, such stimulation has an effect on the emotional state by reducing tension and increasing relaxation. Another approach was taken by scholars inspired by analytical studies. As mentioned previously, human axillary odour is partly due to the presence of some 16-androstenes (primarily 5α-androstenol, 5α-androstenone and androstadienone) that were originally found to affect mating behaviour in pigs . Some authors found a more positive perception of photographs impregnated with the androstenol  or a higher number of social interactions . On the other hand, after application of androstenol and androstenone, male participants found themselves less attractive  and other studies reported no effect . More consistent results come from studies employing androstadienone as a stimulus. This steroid was reported to affect mood, psychophysiology (e.g. skin conductance), cortisol levels and social perception especially in women (for review see ). However, it seems that in humans, the reaction to an olfactory stimulus is rather dependent on the context and so goes far beyond the original pheromone definition (i.e. pheromonal stimulus is followed by a specific behavioural reaction). Interestingly, a context which would most closely resemble pheromonal effects according to the original definition is the interaction between a mother and her newborn baby. Newborn infants when placed on mother's stomach spontaneously crawl to the breasts guided by the sense of smell (they do not show such behaviour when the areolas are washed) . Preference for odour of lactating breast was observed even in babies fed on formula, suggesting that the preference is not dependent on previous reward reinforcement . Furthermore, a recent detailed examination of areola morphology shows that a higher quantity of areolar skin glands is associated with faster onset of breastfeeding and sucking activity .
Skin odour as a mating marker
Sexual selection theories commonly propose mate preferences based on two different systems: (i) gene products reflecting quality of the organism; and (ii) complementary genes. In the first case, frequently referred to as honest signals, it is expected that a majority of the population would follow such a preference . An example from human odour studies is a preference for odour of individuals with low fluctuating asymmetry (FA) . The level of FA reflects randomly distributed deviations from perfect bilateral symmetry. As traits on both sides of the body are based on the same genetic make-up, it is thought that individuals with low-level FA have more finely tuned coordination of the developmental machinery, and in consequence, they are also able to cope with environmental stress such as pathogens or toxins . In contrast, in the case of preferences for complementary genes, individuals are expected to vary in their preferences based on their own genetic make-up. One of the most widely studied models of the preference for complementarity are genes of the major histocompatibility complex (MHC) . Genes of this complex show high diversity, reaching over 200 different alleles in some loci in humans. Products of the MHC genes play a central role in immune system functioning as they specifically bind peptides of foreign origin and present them on the cell surface to other elements of the immune system . High diversity of the MHC genes is usually explained as a result of balancing selection and sexual selection . As inheritance of the MHC alleles is co-dominant, offspring of parents with different alleles show immune responses to wider spectrum of infections. One can, therefore, expect preferences for individuals of different MHC profile than ones own. Following pioneering work by Yamazaki et al.  who found that mice prefer odour of conspecifics different in the MHC genes, this effect was replicated in numerous vertebrate taxa . Results of several studies indicate that in general humans too prefer the odour of MHC different individuals [82, 83]. This pattern is, however, reversed in women using hormonal contraception [84, 85], which might have significant consequences on relationship quality and frequency of break-up . It was recently found that couples who met when the female partner was using hormonal contraception show lower sexual satisfaction; on the other hand, they report higher level of overall relationship satisfaction . Furthermore, there is ongoing debate on what level of dissimilarity we should expect as mating with highly dissimilar individuals may suffer from outbreeding effect, that is, disruption of locally selected gene complexes .
Perspectives: ‘follow your nose!’
We have outlined that human skin is an emitter of olfactory compounds that have the capability to direct our behaviour, consciously or unconsciously. In addition, physicians formerly used olfactory information for diagnosis. Unfortunately, in today′s high-tech medicine, this knowledge seems to have vanished. Among physicians of different medical specialties, the dermatologist is the one who approaches the patients closest. In this context, the nose can still provide valuable information in respect to disease, emotion, age and diet. Furthermore, there is accumulating evidence that odours from skin glands exert physiological and behavioural effects. The original definition of a pheromone implies a stereotypical stimulus-response reaction, as known from many insects. In contrast, the effects in humans seem to be more complex. In the majority of cases, the reaction towards an odour is dependent on individual history and actual motivation and, therefore, hardly ever stereotypical. Furthermore, it is an open question what social and biological impact the increasing personal hygiene measures typical for modern societies will have on human life. Could it be that olfaction in humans is just an aesthetic sense? In most mammals, odours are perceived by two distinct organs within the nasal cavity, namely the main olfactory epithelium (MOE), primary responsible for the perception of odours, and the vomeronasal organ (VNO), mainly specialized on pheromone-driven effects . The VNO binds volatile and non-volatile chemicals and transmits their signal to areas of the limbic system . Interestingly, recent findings describe exceptions from the functional separation between VNO and MOE by detecting pheromone receptors also in the MOE and vice versa vomeronasal responses to non-pheromonal stimuli [90, 91]. In this light, the probable absence of a functional VNO in adult humans  does not necessarily exclude communication via pheromones. However, it seems that olfactory communication in humans plays not that essential role as in most other mammals (see Fig. 1). Of note, recent studies add a new aspect on olfaction by showing the extra-nasal expression of olfactory receptors. In particular, functional olfactory receptors are expressed in human gastrointestinal cells , sperm  and prostate cells . These amazing findings amend the concept of olfaction. In this context, it would be intriguing to investigate whether those receptors are also functionally expressed in human skin.
We are grateful to Dr. Adrian Sewell and Jindra Havlíčková for critically reading the manuscript. JH is supported by the Grant Agency of Czech Republic (GACR P407/10/1303) and Charles University Research Centre (UNCE 204004). All of the authors took part in writing the manuscript and revising it for final publication.
Conflict of Interest
The authors have declared no conflicting interests.