Mechanism of action of retinoids

Retinoic acids in their all-trans or 9-cis configuration are highly-potent activators of the retinoic acid receptors (RAR) and the retinoid-X receptors (RXR). By activation of these nuclear receptors retinoic acids can influence the transcription of various retinoid-response genes (De Luca, Reference De Luca1991). In addition to the retinoic acids several other retinoids, such as 13,14-dihydroretinoic acid (Moise et al. Reference Moise, Kuksa, Imanishi and Palczewski2004, Reference Moise, Kuksa, Blaner, Baehr and Palczewski2005), 3,4-didehydroretinoic acids (Allenby et al. Reference Allenby, Bocquel, Saunders, Kazmer, Speck and Rosenberger1993), 4-oxo-retinol (Achkar et al. Reference Achkar, Derguini, Blumberg, Langston, Levin and Speck1996) and also 4-oxo-retinoic acids (Baron et al. Reference Baron, Heise, Blaner, Neis, Joussen, Dreuw, Marquardt, Saurat, Merk, Bickers and Jugert2005), have been found to be potent activators of RAR (Fig. 2).

Fig. 2. Structural formulas and interrelationships between the various retinoids.

Another important pathway relevant to the immune system involves retinoids with a retro-structure such as anhydroretinol (4,5-didehydro-15,5-retro-deoxyretinol, AR) and 14-hydroxy-retro-retinol (14-HRR) (Fig. 2). 14-HRR has been shown to be a crucial factor for lymphocyte proliferation and AR is a factor responsible for the induction of apoptotic effects (Buck et al. Reference Buck, Derguini, Levi, Nakanishi and Hämmerling1991, Reference Buck, Grun, Derguini, Chen, Kimura, Noy and Hammerling1993; Derguini et al. Reference Derguini, Nakanishi, Hämmerling and Buck1994; O'Connell et al. Reference O'Connell, Chua, Hoyos, Buck, Chen, Derguini and Hämmerling1996). How these effects are mediated is still not clear, but one mechanism for 14-HRR is that it activates protein kinase Cα (Imam et al. Reference Imam, Hoyos, Swenson, Levi, Chua, Viriya and Hämmerling2001).

Mechanism of action of carotenoids

The exact mechanism of action of carotenoids has been only partially elucidated, with the focus mainly on β-carotene in the majority of investigations. Two mammalian enzymes have been identified so far, the cyclic cleavage enzyme 15,15′-β-carotene oxygenase 1 (BCO1) (von Lintig & Vogt, Reference von Lintig and Vogt2000; Redmond et al. Reference Redmond, Gentleman, Duncan, Yu, Wiggert, Gantt and Cunningham2001; von Lintig & Wyss, Reference von Lintig and Wyss2001) and the acyclic cleavage enzyme 2 (BCO2) (Kiefer et al. Reference Kiefer, Hessel, Lampert, Vogt, Lederer, Breithaupt and von Lintig2001; Fig. 1(b)). BCO1 divides β-carotene into two units of retinal, which can be either oxidised to retinoic acid or reduced to retinol (Redmond et al. Reference Redmond, Gentleman, Duncan, Yu, Wiggert, Gantt and Cunningham2001; von Lintig & Wyss, Reference von Lintig and Wyss2001), while BCO2 has been shown to transform β-carotene into apo-8-carotenal (Kiefer et al. Reference Kiefer, Hessel, Lampert, Vogt, Lederer, Breithaupt and von Lintig2001; Fig. 1(b)). It is not known whether apo-carotenals occur endogenously in mammals, but it has been shown that apo-8-carotenal can be oxidised to apo-8-carotenoic acid or degraded to other short-chain apo-carotenals via the β-oxidation pathway (Wang et al. Reference Wang, Russell, Liu, Stickel, Smith and Krinsky1996; Barua & Olson, Reference Barua and Olson2000; for review, see Wang, Reference Wang1994). Few investigations of the biological activity of apo-carotenals and apo-carotenoic acids have been undertaken, but it is known that apo-carotenals and apo-carotenoic acids are only weak activators of the retinoid receptors RAR and RXR (Tibaduiza et al. Reference Tibaduiza, Fleet, Russell and Krinsky2002).

In addition to their nuclear receptor-mediated effects carotenoids also exhibit antioxidant activity by quenching radicals such as singlet oxygen (for review, see Cantrell & Truscott, Reference Cantrell, Truscott, Krinsky, Mayne and Sies2004), and via these antioxidant effects carotenoids may inhibit radical- or peroxide-mediated biological effects such as fatty acid metabolism by lipoxygenase-mediated pathways (Bar-Natan et al. Reference Bar-Natan, Lomnitski, Sofer, Segman, Neeman and Grossman1996). Thereby, carotenoids may mediate gene activation via metabolism of the PUFA linoleic acid and arachidonic acid to their hydroxy-metabolites 15-hydroxyeicosatetraenoic acid (15-HETE) and 13-hydroxyoctadecadienoic acid (13-HODE), which are highly potent activators of PPAR (Huang et al. Reference Huang, Welch, Ricote, Binder, Willson, Kelly, Witztum, Funk, Conrad and Glass1999; Fig. 1(c)).

Regulation of carotenoids and retinoid concentrations in man

Two major pathways have been described for the regulation of the concentration of the active retinoid and RAR activator all-trans-retinoic acid (Fig. 3):

1. (a) organ-specific targetted temporal and spatial synthesis by retinaldehyde dehydrogenase (RALDH) isoforms (for review, see Duester, Reference Duester2000)). In reproductive tissues such as embryo, uterus, ovaries and testes retinoic acid is synthesised via controlled spatial and temporal expression of RALDH isoforms. In the adult retinoic acid synthesis mainly occurs in continually-differentiating tissues such as skin, hair and the immune and intestinal systems (for review, see Napoli Reference Napoli1999);

2. (b) a second means of regulation is non-specific regulation via nutrient bioavailability. High intakes of vitamin A, mainly in the form of retinyl esters, lead to increased concentrations of retinyl esters, retinol and retinoic acids (Arnhold et al. Reference Arnhold, Tzimas, Wittfoht, Plonait and Nau1996; van Vliet et al. Reference van Vliet, Boelsma, de Vries and van den Berg2001). High vitamin A intakes also lead to long-term increases in retinoic acid concentrations (Siegel et al. Reference Siegel, Craft, Roe, Duarte-Franco, Villa, Franco and Giuliano2004). In addition, provitamin A carotenoids such as β-carotene, α-carotene and β-cryptoxanthin are mainly stored in organs such as the liver and adipose tissue, and their release into the serum seems to be non-homeostatically regulated. This finding implies that high consumption of provitamin A carotenoids potentially leads to high organ concentrations and high serum concentrations of these carotenoids and, further, higher serum concentrations of all-trans-retinoic acid. For example, human subjects supplemented with β-carotene have been shown to have increased concentrations of all-trans-retinoic acid (Thürmann et al. Reference Thürmann, Steffen, Zwernemann, Aebischer, Cohn, Wendt and Schalch2002).

Fig. 3. Biologically-relevant pathways originating from provitamin A carotenoids and vitamin A for the generation of all-trans-retinoic acid. BCO, β-carotene oxygenase; RDH, retinol dehydrogenase; RALDH, retinaldehyde dehydrogenase.

Differentiating and proliferating effects of retinoids on immune-competent cells

Retinoids and T-helper cell 1 – T-helper cell 2 balance

Probably the most important aspect of the role of retinoic acid in relation to the immune system is its effect on the T-helper (Th) 1–Th2 balance (Cantorna et al. Reference Cantorna, Nashold and Hayes1994; Tokuyama et al. Reference Tokuyama, Tokuyama and Nakanishi1995; Tokuyama & Tokuyama, Reference Tokuyama and Tokuyama1996; Hoag et al. Reference Hoag, Nashold, Goverman and Hayes2002; Rühl et al. Reference Rühl, Garcia, Schweigert and Worm2004). Several research groups (Tokuyama et al. Reference Tokuyama, Tokuyama and Nakanishi1995; Worm et al. Reference Worm, Krah, Manz and Henz1998, Reference Worm, Herz, Krah, Renz and Henz2001; Hoag et al. Reference Hoag, Nashold, Goverman and Hayes2002; Stephensen et al. Reference Stephensen, Rasooly, Jiang, Ceddia, Weaver, Chandraratna and Bucy2002; Iwata et al. Reference Iwata, Eshima and Kagechika2003) have reported that in in vitro models as well as in in vivo models retinoic acid exerts a direct effect on T-cells, suppressing Th1 development and enhancing Th2 development via an RAR-mediated response. In vitamin A-deficient mice, in particular, there is definite evidence for a Th2 defect (Carman et al. Reference Carman, Smith and Hayes1989). There are few Th2 cells and the addition of retinyl acetate restores Th2 cell numbers. Thus, vitamin A-deficient mice have an insufficiency of Th2 cells to drive B-cell proliferation and differentiation. Excessive interferon-γ (IFN-γ) synthesis may partly account for this Th2 cell insufficiency, because IFN-γ inhibits Th2 cell development (Abbas et al. Reference Abbas, Murphy and Sher1996), as summarised in Fig. 4. Various aspects of vitamin A deficiency that relate to the immune response are summarised in a review by Hayes et al. (Reference Hayes, Nashold, Enrique Gomez, Hoag, Nau and Blaner1999).

Fig. 4. Retinoic acid (RA)-modified pathways for T-helper (Th) 1 and Th2 regulation. APC, antigen-presenting cell; IFN, interferon; EO, eosinophil; +, , promoted; –, ⊥, inhibited; , effects of cytokines.

How all these effects are mediated has not been established, but it has been shown that retinoic acid regulates the expression of IFN-γ, IL-2, IL-10 and IL-12 production (Cantorna et al. Reference Cantorna, Nashold and Hayes1994; Tokuyama & Tokuyama, Reference Tokuyama and Tokuyama1996; Stephensen et al. Reference Stephensen, Rasooly, Jiang, Ceddia, Weaver, Chandraratna and Bucy2002, Reference Stephensen, Jiang and Freytag2004; Iwata et al. Reference Iwata, Eshima and Kagechika2003; Rühl et al. Reference Rühl, Garcia, Schweigert and Worm2004).

Retinoids and apoptotic effects

An important function of endogenous retinoids is the induction and inhibition of apoptotic effects (for review, see Szondy et al. Reference Szondy, Reichert and Fesus1998). Retinoids induce apoptosis of immune-competent cells during back regulation of immune reactions (see p. 464) and during thymic selection (Foerster et al. Reference Foerster, Sass, Rühl and Nau1996; Yagi et al. Reference Yagi, Uchida, Kuroda and Uchiyama1997; Szondy et al. Reference Szondy, Reichert and Fesus1998). In various cell lines it has been shown that apoptosis is a major effect induced by retinoids (for review, see Altucci & Gronemeyer, Reference Altucci and Gronemeyer2001). The apoptotic effects of retinoids are mainly induced via RAR/RXR-mediated effects (Szondy et al. Reference Szondy, Reichert and Fesus1998), other nuclear receptors such as PPAR (Theocharis et al. Reference Theocharis, Margeli, Vielh and Kouraklis2004) and Nur77 (Szegezdi et al. Reference Szegezdi, Kiss, Simon, Blasko, Reichert, Michel, Sandor, Fesus and Szondy2003; Toth et al. Reference Toth, Ludanyi, Kiss, Reichert, Michel, Fesus and Szondy2004), and also via non-nuclear receptor-mediated effects (for review, see Lotan, Reference Lotan2003).

Carotenoids and immune responses

Several effects of carotenoids are thought to be mediated by their metabolism to vitamin A and subsequent mediation of RAR/RXR-response pathways. Surprisingly, even non-provitamin A carotenoids such as lutein, canthaxantin and lycopene exhibit marked effects on the immune system (for summary, see Chew & Park, Reference Chew and Park2004; Hughes, Reference Hughes, Krinsky, Mayne and Sies2004).

In general, carotenoids modulate T-cell proliferation, e.g. β-carotene potentiates the increase in CD4+ cells and is suggested to be an immuno-enhancing agent in the management of HIV infections (Fryburg et al. Reference Fryburg, Mark, Griffith, Askenase and Patterson1995). Various supplementation studies with carotenoids in man (Watzl et al. Reference Watzl, Bub, Brandstetter and Rechkemmer1999) have found that enriching the diet with β-carotene (by carrot juice), lycopene (by tomato juice) or lutein (by spinach powder) to some extent mediates T-cell proliferation. In cell-culture experiments (Prabhala et al. Reference Prabhala, Maxey, Hicks and Watson1989; Jyonouchi et al. Reference Jyonouchi, Zhang, Gross and Tomita1994) the non-provitamin A carotenoids canthaxantin, astaxanthin and lutein have been shown to enhance T-cell proliferation.

Natural killer cell activity seems to be another important target of carotenoid action; supplementation of subjects with β-carotene enhances their natural killer cell activity as compared with subjects of a similar age given placebo treatment (Santos et al. Reference Santos, Meydani, Leka, Wu, Fotouhi, Meydani, Hennekens and Gaziano1996).

A proposed mechanism for carotene-mediated immunostimulation is related to its ability to suppress the generation of arachidonic acid cascade products in vitro (Halevy & Sklan, Reference Halevy and Sklan1987). It is suggested that the production of prostaglandin E₂, an immunosuppressive mediator, is down regulated (Halevy & Sklan, Reference Halevy and Sklan1987). This effect is possibly mediated via cyclooxygenase inhibition comparable with the lipoxygenase inhibition mechanism (Bar-Natan et al. Reference Bar-Natan, Lomnitski, Sofer, Segman, Neeman and Grossman1996; Fig. 1(c)).

Vitamin A and carotenoid transfer to milk and subsequently to the child

Retinol is the predominant retinoid in human serum and is mainly transported by the retinol-binding protein (RBP), although after high-vitamin A supplementation retinyl esters are incorporated and transported in lipoproteins (Mallia et al. Reference Mallia, Smith and Goodman1975). The transport of vitamin A from the maternal serum to the milk may be mediated via a number of possible transfer mechanisms. Vahlquist & Nilsson (Reference Vahlquist and Nilsson1979) have shown that RBP-mediated transfer is the most important source of milk vitamin A, providing a constant supply of the vitamin, while plasma lipoproteins become important in transport during increased intakes of vitamin A. The relative inefficiency of lipoprotein-mediated transfer may help to protect the offspring from ingestion of toxic levels of milk vitamin A in the case of maternal hypervitaminosis A (Vahlquist & Nilsson, Reference Vahlquist and Nilsson1979). Vitamin A appears in milk mainly as retinyl esters (Vahlquist & Nilsson, Reference Vahlquist and Nilsson1979). Later studies (Davila et al. Reference Davila, Norris, Cleary and Ross1985) in rats have shown that increased ingestion of vitamin A is not associated with increased maternal serum vitamin A concentrations. However, the liver vitamin A concentrations of the dams, their milk vitamin A concentrations and the liver vitamin A concentrations of their 14-d-old pups are higher when dams are fed higher-vitamin A (30 retinol equivalents/kg) diets during lactation. These findings indicate that the transfer of vitamin A in milk from mother to offspring and the vitamin A status of the dams and their suckling pups are influenced by maternal vitamin A intake during lactation (Fig. 5). More recent studies (Green et al. Reference Green, Green, Akohoue and Kelley2001a) have established that chylomicrons contribute at least one-third of the vitamin A in milk in rats fed at the higher level of vitamin A, while chylomicrons from rats fed at the lower level of vitamin A contain negligible amounts of vitamin A. In the animals fed the lower level of vitamin A holoRBP is able to deliver vitamin A to the lactating mammary tissue, since vitamin A is present even if rats are fed a vitamin A-free diet (Green et al. Reference Green, Snyder, Akohoue and Green2001b). Experiments with RBP-/- mice (Vogel et al. Reference Vogel, Piantedosi, O'Byrne, Kako, Quadro, Gottesman, Goldberg and Blaner2002) have shown that there is some variability in milk vitamin A levels with time, although there are no overall differences through the weaning period. The importance of postprandial vitamin A for retinyl ester incorporation into the mammary tissue and subsequently into the milk also involves the cellular RBP (CRBP) 3. In CRBP3-/- mice less vitamin A, particularly in the form of retinyl esters, is incorporated into the milk (Piantedosi et al. Reference Piantedosi, Ghyselinck, Blaner and Vogel2005).

Fig. 5. Simplified and schematic effects of human relevant mechanisms in the transfer of β-carotene and/or vitamin A to breast milk and subsequently to the child.

β-Carotene transport has not been investigated as extensively as the vitamin A transport mechanisms. A study in healthy lactating women (Schweigert et al. Reference Schweigert, Bathe, Chen, Buscher and Dudenhausen2004) has suggested that lipoprotein transport of carotenoids from serum to the milk is responsible for carotenoid transfer. β-Carotene levels in the maternal plasma and breast milk are increased after single-dose β-carotene supplementation, but retinol concentrations are not affected (Canfield et al. Reference Canfield, Giuliano, Neilson, Yap, Graver, Cui and Blashill1997). Long-term supplementation with 30 mg β-carotene/d is associated with a small but non-significant increase in human breast milk levels (Gossage et al. Reference Gossage, Deyhim, Yamini, Douglass and Moser-Veillon2002), while a study (Canfield et al. Reference Canfield, Giuliano, Neilson, Blashil, Graver and Yap1998) that provided 60 and 210 mg β-carotene/d has reported increases in both serum and milk retinol and β-carotene. Maternal β-carotene supplementation seems to be an important factor in the supply of vitamin A to the infant; in addition to increasing the levels of vitamin A in breast milk, maternal β-carotene supplementation also increases β-carotene in breast milk and can thereby supply retinol for the nursing infant (Canfield et al. Reference Canfield, Taren, Kaminsky and Mahal1999). However, it has not been established where the bioconversion of the milk-derived β-carotene to retinol takes place (for summary, see Fig. 5).

Whether this β-carotene or vitamin A can be converted to bioactive all-trans-retinoic acid in young mammals is not known, although a recent study (Rühl et al. Reference Rühl, Hamscher, Garcia, Nau and Schweigert2005) has shown that all-trans-retinoic acid is not present in the serum of rat pups aged 3 and 11 d, while other bioactive vitamin A metabolites are present at high concentrations.

Influence of Western nutrition on carotenoids and retinoid transfer in human subjects

Fig. 6 summarises how ‘Western’ nutrition, which is high in dietary fat, vitamin A and β-carotene, mediates metabolism to the active vitamin A metabolite all-trans-retinoic acid. Not only is the dietary intake of vitamin A (Olafsdottir et al. Reference Olafsdottir, Wagner, Thorsdottir and Elmadfa2001; Allen & Haskell, Reference Allen and Haskell2002; Mensink, Reference Mensink2002) and β-carotene (Hellenbrand et al. Reference Hellenbrand, Bauer, Boeing, Seidler and Robra2000) high, but also the high levels of dietary fat may increase various factors responsible for improved absorption and subsequent bioactivation of β-carotene and vitamin A to all-trans-retinoic acid.

Fig. 6. Influence of Western nutrition on some of the factors involved in carotenoid and retinoid metabolism to all-trans-retinoic acid. RAR, retinoic acid receptor; BCO, β-carotene oxygenase; RALDH, retinaldehyde dehydrogenase; CRBP, cellular retinol-binding protein; RDH, retinol dehydrogenase.

A high-fat diet up regulates the expression of RAR (Bonilla et al. Reference Bonilla, Redonnet, Noel-Suberville, Pallet, Garcin and Higueret2000), CRPB 1 and 2, which are responsible for retinol uptake (During et al. Reference During, Nagao and Terao1998; Takase et al. Reference Takase, Tanaka, Suruga, Kitagawa, Igarashi and Goda1998, Reference Takase, Suruga and Goda2000; Hellemans et al. Reference Hellemans, Rombouts, Quartier, Dittie, Knorr, Michalik, Rogiers, Schuit, Wahli and Geerts2003), and BCO1 (During et al. Reference During, Nagao and Terao1998; Boulanger et al. Reference Boulanger, McLemore, Copeland, Gilbert, Jenkins, Yu, Gentleman and Redmond2003). In addition, high intakes of fat and β-carotene result in increased β-carotene levels in several organs and increased levels of vitamin A in the serum and various organs (Schweigert et al. Reference Schweigert, Baumane, Buchholz and Schoon2000; van het Hof et al. Reference van het Hof, West, Weststrate and Hautvast2000; Ribaya-Mercado, Reference Ribaya-Mercado2002). In contrast, a diet low in energy reduces serum concentrations of retinoic acid and retinol (Berggren Soderlund et al. Reference Berggren Soderlund, Fex and Nilsson-Ehle2003).

Effects of carotenoids and retinoids on immune development

Immune development after birth in man involves three major processes in which retinoids are involved: the differentiation of immune-competent cells; thymic selection; the proliferation and expansion of lymphocytes (summarised in Fig. 7).

Fig. 7. Influence of the human diet, via reduction in, and induction of, cytokine release, on some of the factors involved in postnatal immune development. IFN, interferon; Th, T-helper.

During thymic selection T-cells develop in the thymus through a series of stages defined by the expression of the cell-surface markers CD4 and CD8. The development of the human thymus starts before birth and ceases during puberty with involution of the thymus (Sen, Reference Sen2001). Vitamin A deficiency is known to be accompanied by immune deficiency and a susceptibility to a wide range of infectious diseases (for review, see Reifen, Reference Reifen2002; Semba, Reference Semba1994)). In vitamin A-deficient animals a marked atrophy of the thymus and spleen has been observed (West et al. Reference West, Howard and Sommer1989); on the other hand, retinoids at higher concentrations are toxic and cause involution of lymphoid organs, in particular the thymus (Makori et al. Reference Makori, Peterson, Lantz and Hendrickx2002). A recent study (I Kiss, R Rühl, B Fritzsche, T Nemeth, E Szegezdi, T Perlmann and Z Szondy, under review) has shown that retinoic acid-synthesising enzymes peak at the same time as RAR response, when thymic cellularity is highest and the T-cell selection process, as indicated by a high rate of apoptosis, is most effective. It can be concluded that thymic selection is a direct target of retinoic acid during thymocyte development, and it can be postulated that high maternal dietary intake of vitamin A or provitamin A carotenoids may modify thymic selection processes. Until now there has been no research on the effects of carotenoids on thymic selection processes.

The second important process in which retinoids are involved is the proliferation of lymphocyte populations; the lymphocyte response to mitogens, in particular, is retinoid dependent (Wang & Ballow, Reference Wang and Ballow1993; Ballow et al. Reference Ballow, Wang and Xiang1996a,b). In a recent study (Garcia et al. Reference Garcia, Rühl, Herz, Koebnick, Schweigert and Worm2003) of pregnant mice fed a basal (control) diet or different retinoid- and carotenoid-enriched (4500 retinol equivalents/kg) diets from day 1 of conception the percentage and total numbers of splenic mononuclear cells were determined on days 1, 3, 5, 7, 14, 21 and 65 of pregnancy. Increases were observed in the early days of pregnancy (3 and 5) with vitamin A (retinyl palmitate) supplementation, while β-carotene supplementation was found to mainly increase CD3+ cell numbers from day 5 to day 14. At day 7 increases were found in CD4:CD8 after vitamin A supplementation and in T-cell:B-cell after vitamin A and β-carotene supplementation. In general, IgG levels were not found to be altered by the different diets. These results confirm that supplementation with vitamin A and β-carotene affects immune cell functions during ontogenesis. However, maternal vitamin A supplementation via intraperitoneal injections has been shown to increase serum IgM and Th2-specific IgG1 levels in the progeny (Guzman & Caren, Reference Guzman and Caren1991). Furthermore, in a human supplementation study (Gossage et al. Reference Gossage, Deyhim, Moser-Veillon, Douglas and Kramer2000), which investigated the effects of β-carotene supplementation during early lactation (days 4–32 post partum) on circulating carotenoids and the T-cell proliferative response to phytohaemagglutinin, it has been found that neither lactation nor β-carotene supplementation affects T-cell proliferation.

Recently, a study (R Rühl, A Hänel, A Garcia, U Herz, FJ Schweigert and M Worm, under review) has been performed in mice in which vitamin A-supplemented (30 mg/d) or vitamin A-free diets were fed to the dams throughout lactation and directly to the pups after weaning with or without ovalbumin sensitisation. It was found that vitamin A supplementation decreases splenic T-cell and B-cell numbers and also enhances IL-4 production and specific IgE after sensitisation. By contrast, the mice fed the vitamin A-free diet were found to show no alteration in lymphocyte cell numbers, a slightly increased IL-4 production and no decrease in specific IgE levels. Together these findings show that the severity of allergic sensitisation depends on the vitamin A content of the maternal diet during lactation. In addition, vitamin A strongly enhances immune responses only after mitogen stimulation, while in the absence of immune stimulation vitamin A only has marginal effects.

The Th1→Th2 switch is another important process that is affected by vitamin A during postnatal development. Initial evidence is available from a study (Guzman & Caren, Reference Guzman and Caren1991) that describes increased Th2-specific IgG1 concentrations after vitamin A supplementation and from a recent study (R Rühl, A Hänel, A Garcia, U Herz, FJ Schweigert and M Worm, under review) showing enhanced IL-4 secretion and increased specific IgE levels after a vitamin A-supplemented diet and ovalbumin sensitisation. Whether these outcomes are mediated by lymphocyte-mediated effects, or via antigen-presenting cell-mediated effects, is as yet unknown. Previous studies have shown that, as antigen-presenting cells, dendritic cells are the early regulators of the Th1–Th2 response (Ridge et al. Reference Ridge, Fuchs and Matzinger1996).

The effects of carotenoids are quite difficult to investigate because of the pronounced differences in carotenoid absorption, kinetics and metabolism between man and rodent laboratory animals (for review, see Lee et al. Reference Lee, Boileau, Boileau, Williams, Swanson, Heintz and Erdman1999). A collaborative study (Rühl et al. Reference Rühl, Schweigert, Wahn and Grüber2006) has shown that carotenoids with provitamin A activity or non-provitamin A activity are present at different levels in children of different ethnicity and with a different risk of allergic sensitisation. Provitamin A carotenoids are low in children of Turkish origin who live in Germany and have low cultural adaptation to the German lifestyle, but they are higher in well-adapted Turkish children and are highest in German children. On the other hand, the levels of non-provitamin A carotenoids are high in Turkish children and low in German children. This study suggests that in man different levels of carotenoids may be associated with an increased prevalence of allergic diseases; whether these differences in carotenoid distribution between different groups is associated with the bioactive vitamin A metabolite all-trans-retinoic acid is under investigation.

Human breast milk v. milk formulas

In addition to a different pattern of proteins, milk formulas also have a different profile of carotenoids and retinoids. Breast milk mainly contains retinyl esters conjugated with different fatty acids such as palmitic, stearic, oleic and linoleic acids (Ross et al. Reference Ross, Davila and Cleary1985; Piantedosi et al. Reference Piantedosi, Ghyselinck, Blaner and Vogel2005), as well as retinol (Olafsdottir et al. Reference Olafsdottir, Wagner, Thorsdottir and Elmadfa2001), while milk formulas contain only retinyl palmitate or retinyl acetate as the vitamin A source (Landen et al. Reference Landen, Hines, Hamill, Martin, Young, Eitenmiller and Soliman1985). However, the plasma vitamin A concentrations for the formula-fed baby and the breast-fed baby are comparable (Ghebremeskel et al. Reference Ghebremeskel, Burns, Costeloe, Burden, Harbige, Thomas and Temple1999). Breast milk contains the whole spectrum of carotenoids present in the human diet and serum (Khachik et al. Reference Khachik, Spangler, Smith, Canfield, Steck and Pfander1997; Canfield et al. Reference Canfield, Clandinin, Davies, Fernandez, Jackson and Hawkes2003; Schweigert et al. Reference Schweigert, Bathe, Chen, Buscher and Dudenhausen2004), while milk formulas contain no carotenoids, very low concentrations of carotenoids or a limited variety of carotenoids (Sommerburg et al. Reference Sommerburg, Meissner, Nelle, Lenhartz and Leichsenring2000).

The importance of these different retinoid and carotenoid patterns in relation to the development of the immune system in Western countries is hard to predict, particularly as many aspects of the function of carotenoids remain elusive.