**6. Animal models of allergy and preclinical studies with probiotics**

Even though animal models do not completely recapitulate all clinical, histological and immunological features of human AD, they can offer valuable tools (i) to evaluate candidate probiotic strains and their ability to prevent/alleviate AD and (ii) to elucidate their cellular and molecular mechanisms of action. Indeed, assessing the effect of probiotic candidate strains directly in human trials is expensive and time-consuming. Moreover, the number of candidate strains, the importance of their preparation/formulation, the dose, routes and possible administration regimens/schedules increase the number of parameters to be evaluated before to launch a clinical trial. Also, food and safety agencies recommend to better characterize the mechanism of action of potential probiotics. Since access to biological materials other than blood is limited and functional studies are difficult to perform in human trials, research on probiotics may benefit from preclinical animal models (Kalliomaki et al., 2010).

In the context of AD, several animal models are used nowadays (Jin et al., 2009) (table 1). The historical model of Nc/Nga mice that spontaneously develop AD-like features has been commonly used to evaluate the capacity of candidate strains to prevent/manage AD (Matsuda et al., 1997). In this model, Nc/Nga mice housed under specific pathogen free conditions are protected from AD. Once these mice are transferred to air-unregulated conventional environment, they exhibit AD-like lesions by 7-8 weeks. Feeding (heat-treated) *Lactobacillus rhamnosus* GG or live *Lactobacillus johnsonii* NCC533 around weaning period prevented or delayed the onset of AD (Sawada et al., 2007; Tanaka et al., 2008). Offsprings exhibited lower clinical scores with shorter scratching duration/frequency, a reduced total IgE serum titer and a decreased number of mast cells infiltrating skin lesions. One of the caveats of the model of Nc/Nga mice is that the allergen is unknown. To circumvent this


children are affected in developed countries (Asher et al., 2006; Williams et al., 1999). The treatment of AD is mainly symptomatic and includes emollients to moisturize the skin, and topical corticosteroids along with a switch to an elimination diet (BuBmann et al.,

AD has a complex etiology. Recent investigations into mechanisms that drive the inflammatory response in AD have highlighted the crucial role of genetic predisposition. Many mutations have been associated to the development of AD. These mutations are located in two subsets of genes, in structural protein genes involved in the epidermal-barrier function or in genes involved in IgE production (Bieber, 2008; Demehri et al., 2009; Hsu et al., 2008; Suzuki et al., 2011). Also, environmental factors can contribute to the development of AD. As

As mentioned above, it has been postulated ("hygiene hypothesis") that the pattern of bacterial colonization of the gut during the early months after birth can contribute to the development of AD (Vael and Desager, 2009). During the first year of life the newborn immune system is still under development while exposure to novel dietary foods is increasing, especially around the weaning period. These multifactorial events in concert contribute to the establishment of oral tolerance, *i.e.* prevention or development of atopic sensitization to common foods such as cow's milk, eggs, wheat and nuts that predispose to the development of AD (Garcia et al., 2007; Gonzalez, I et al., 1971; Han et al., 2004;

Even though animal models do not completely recapitulate all clinical, histological and immunological features of human AD, they can offer valuable tools (i) to evaluate candidate probiotic strains and their ability to prevent/alleviate AD and (ii) to elucidate their cellular and molecular mechanisms of action. Indeed, assessing the effect of probiotic candidate strains directly in human trials is expensive and time-consuming. Moreover, the number of candidate strains, the importance of their preparation/formulation, the dose, routes and possible administration regimens/schedules increase the number of parameters to be evaluated before to launch a clinical trial. Also, food and safety agencies recommend to better characterize the mechanism of action of potential probiotics. Since access to biological materials other than blood is limited and functional studies are difficult to perform in human trials, research on probiotics may benefit from preclinical

In the context of AD, several animal models are used nowadays (Jin et al., 2009) (table 1). The historical model of Nc/Nga mice that spontaneously develop AD-like features has been commonly used to evaluate the capacity of candidate strains to prevent/manage AD (Matsuda et al., 1997). In this model, Nc/Nga mice housed under specific pathogen free conditions are protected from AD. Once these mice are transferred to air-unregulated conventional environment, they exhibit AD-like lesions by 7-8 weeks. Feeding (heat-treated) *Lactobacillus rhamnosus* GG or live *Lactobacillus johnsonii* NCC533 around weaning period prevented or delayed the onset of AD (Sawada et al., 2007; Tanaka et al., 2008). Offsprings exhibited lower clinical scores with shorter scratching duration/frequency, a reduced total IgE serum titer and a decreased number of mast cells infiltrating skin lesions. One of the caveats of the model of Nc/Nga mice is that the allergen is unknown. To circumvent this

such, AD can be classified into IgE-mediated and non IgE-mediated subtypes (Fig. 1.).

**6. Animal models of allergy and preclinical studies with probiotics** 

2009; Peserico et al., 2008).

Heratizadeh et al., 2011).

animal models (Kalliomaki et al., 2010).


Table 1. Common preclinical models of AD (adapted from (Jin et al., 2009))

Probiotics and Atopic Dermatitis 333

drawback, modified models have been set up. Repeated application of house dust mite *Dermatophagoides farina* extract induces human-like atopic skin lesions (Huang et al., 2003a). In this model, oral administration of live *Lactobacillus plantarum* from Kimchi, a traditional Korean fermented food (Won et al., 2011), or of live *L. johnsonii* NCC533 alleviated AD symptoms (Inoue et al., 2007). In another Nc/Nga-derived model, mice were sensitized by epicutaneous application of the hapten picryl chloride, and oral administration of heattreated *Lactobacillus breve* SBC8803 delayed the development of AD (Segawa et al., 2008). It is noteworthy that most of these experiments have been performed in young mice with administration of probiotics around weaning. However recent papers started to investigate the impact of different intervention periods. In this respect, a protective effect of live *L. rhamnosus* LPR was observed when the strain was fed to pregnant dams and their pups for 12 weeks. Protection was equally seen when probiotic treatment started at weaning but not

when it was initiated one week after the onset of the disease (Tanaka et al., 2009).

2005) and to induce regulatory T cells (Hacini-Rachinel et al., 2009).

subjects diagnosed with AD to better manage their symptoms.

symptoms was observed (Marsella, 2009).

**7.1 Prevention of AD** 

**7. Probiotics and allergy: Clinical studies** 

The general hypothesis evoked to explain the beneficial effect of specific probiotic strains in allergy mouse models is their capacity to establish or restore the Th1/Th2 balance. Indeed, allergic disorders such as AD are characterized by an immune response that is Th2-biased with increased production of cytokines IL-4, IL-5 and IL-13. Investigators working in the probiotic field thus started to screen strains according to their ability to polarize T cells responses into Th1 responses (with IL-12 and INFγ production) (Mohamadzadeh et al.,

Alternative models such as the canine model of AD might be of interest as the latter, for ex., displays features similar to human AD. Of note, Marsella et al. evaluated the efficacy of *L. rhamnosus* GG in the prevention of canine AD but no significant decrease in clinical

Studies have been conducted comparing the gut microbiota of infants with AD or food allergy to non-allergic infants and in these studies it was observed that allergic infants had reduced numbers of lactobacilli and bifidobacteria species in their gut (Adlerberth et al., 2007; Kirjavainen et al., 2002; Penders et al., 2007). A little more than a decade ago the first studies were published testing the hypothesis that probiotic intervention either in the pre- or post- natal period (pregnant women, offsprings or both) could influence the incidence of AD in the early years of life. Over 25 published studies have investigated similar hypotheses since then. However, these studies have largely differed in the choice of probiotic strain investigated, duration of administration of the strain, the population treated (mothers vs. newborns vs. both mothers and newborns) and in selecting the primary outcome addressing the efficacy of the trial. We have grouped the studies into 2 types- those in which the probiotic is given as a prevention strategy *i.e.* in at risk population (history of atopy in the family) and the others in which probiotics are administered as a therapeutic entity in

Around 15 clinical trials have investigated the efficacy of probiotic in prevention of AD (summarized in Table 2). We highlight some of the more relevant well designed trials that raised key questions in the field of probiotics. Kalliomaki *et al.* conducted the first long-term preventive study on probiotics in AD and found that supplementation with the *L. rhamnosus* 

Table 1. Common preclinical models of AD (adapted from (Jin et al., 2009))

drawback, modified models have been set up. Repeated application of house dust mite *Dermatophagoides farina* extract induces human-like atopic skin lesions (Huang et al., 2003a). In this model, oral administration of live *Lactobacillus plantarum* from Kimchi, a traditional Korean fermented food (Won et al., 2011), or of live *L. johnsonii* NCC533 alleviated AD symptoms (Inoue et al., 2007). In another Nc/Nga-derived model, mice were sensitized by epicutaneous application of the hapten picryl chloride, and oral administration of heattreated *Lactobacillus breve* SBC8803 delayed the development of AD (Segawa et al., 2008). It is noteworthy that most of these experiments have been performed in young mice with administration of probiotics around weaning. However recent papers started to investigate the impact of different intervention periods. In this respect, a protective effect of live *L. rhamnosus* LPR was observed when the strain was fed to pregnant dams and their pups for 12 weeks. Protection was equally seen when probiotic treatment started at weaning but not when it was initiated one week after the onset of the disease (Tanaka et al., 2009).

The general hypothesis evoked to explain the beneficial effect of specific probiotic strains in allergy mouse models is their capacity to establish or restore the Th1/Th2 balance. Indeed, allergic disorders such as AD are characterized by an immune response that is Th2-biased with increased production of cytokines IL-4, IL-5 and IL-13. Investigators working in the probiotic field thus started to screen strains according to their ability to polarize T cells responses into Th1 responses (with IL-12 and INFγ production) (Mohamadzadeh et al., 2005) and to induce regulatory T cells (Hacini-Rachinel et al., 2009).

Alternative models such as the canine model of AD might be of interest as the latter, for ex., displays features similar to human AD. Of note, Marsella et al. evaluated the efficacy of *L. rhamnosus* GG in the prevention of canine AD but no significant decrease in clinical symptoms was observed (Marsella, 2009).
