1
Ingestion of bovine lactoferrin induces expression of interferon alpha in the human colon
Supplementary Text S2:
1. Neither bLF nor bLF-derived peptides directly exerted toxic effects against colorectal polyps
In all of our work with bLF, bLF itself (up to 1000 µg/ml) has never exhibited any direct toxic effect toward normal or transformed cells in culture (data not shown), and in all of the animal studies performed to date, ingestion of bLF had no toxic effects on the experimental animals. In addition, after ingestion of lactoferrin, neither lactoferrin nor lactoferrin fragments can be detected in the cecum or colon of rats (Kuwata et al., 2001) or the colon of humans (Troost et al., 2002): It is important to note that in the report by Kuwata et al., 2001, while significant levels of lactoferrin fragments were detected in the lower small intestine of rats orally administered lactoferrin, no lactoferrin fragments were detected in the cecum or colon of these animals. Thus, available evidence indicates that neither bLF itself nor bLF-derived peptide fragments had a direct toxic effect on colorectal polyp cells in the Tokyo-trial.
2. Neither bLF nor bLF-derived peptides directly induced expression of IFNA in the colon of the Tokyo-trial participants
To date, only two receptors, Toll-like receptor 4 (TLR4) and low-density lipoprotein receptor-related protein 1 (LRP1; also known as CD91), are known to bind lactoferrin and activate intracellular signaling pathways (Alexander et al., 2012). TLRs play a critical role in pathogen recognition by the immune system (Kumar et al., 2009; Moresco et al., 2011). TLRs are expressed by immune and epithelial cells throughout the small intestine and colon (Abreu et al., 2005; Artis, 2008; O'Hara & Shanahan, 2006). However, TLR4 expression and response to agonist is low in colon epithelial and dendritic cells (Abreu et al., 2005; Hart et al., 2005; Takenaka et al., 2007), with LPS tolerance being established shortly after birth (Lotz et al., 2006). Moreover, lactoferrin binds TLR4 via its carbohydrate moieties (Ando et al., 2010), and bovine lactoferrin does not bind human TLR4 (Ando et al., 2010). LRP1, the other known signaling receptor for lactoferrin (Grey et al., 2004; Naot et al., 2005; Takayama et al., 2003), is a widely expressed, multifunctional receptor having diverse biological roles (Boucher & Herz, 2011; Lillis et al., 2008). Unlike TLR4, bovine lactoferrin binds to and stimulates human LRP1 signal transduction (Takayama et al., 2003). However, there are no reports of lactoferrin stimulating interferon expression through LRP1 signaling, including a study using a low-density array to evaluate gene expression in osteoblastic cells treated with lactoferrin (Naot et al., 2011). In addition, lactoferrin is present at moderate to high levels in human milk, tears, upper respiratory tract mucosal fluid, and seminal plasma (Alexander et al., 2012), and there are no reports of specifically increased cytokine (including interferon) expression in the eye, upper respiratory tract, seminal vesicles, or infant intestine. Also, there are no reports of lactoferrin knockout mice (Ward et al., 2003; Ward et al., 2008) being deficient in cytokine (including interferon) production: It is important to note that in the report by Ward et al., 2008, while neutrophils from lactoferrin knockout mice have a defective oxidative burst when tested in vitro, all neutrophil responses are completely normal when tested in vivo (also see Alexander et al., 2012). Finally, as noted above, after ingestion of bLF, neither bLF nor bLF-derived peptide fragments reach the colon (Kuwata et al., 2001; Troost et al., 2002). Taken together, these reports indicate that neither intact nor peptide fragments of bLF are present in the colons of humans ingesting bLF, and that bLF and its peptide fragments do not bind to receptors in the human colon that mediate expression of type I interferons.
3. A model by which ingestion of bLF is able to induce the expression of interferon alpha (IFNA)
Lactoferrin is an important component of the innate immune system involved in mucosal and neutrophil-mediated immunity (reviewed in Alexander et al., 2012). The primary function of endogenous human LF in mucosal fluids is to promote the non-lethal, non-inflammatory removal of microbial pathogens away from cells and tissues. Ingestion of bovine lactoferrin has effects distinct from those of endogenous human LF. In contrast to endogenous LF, which is microbiostatic, ingestion and subsequent gastric digestion of bLF results in the generation of peptides (Furlund et al., 2013; Kuwata et al., 1998a; Kuwata et al., 1998b; Kuwata et al., 2001) with microbicidal activity (Gifford et al., 2005; Haney et al., 2007; Liu et al., 2011; Mirza et al., 2011; van der Kraan et al., 2004; Yamauchi et al., 1993). Importantly, the cell killing activity of naturally occurring cationic anti-microbial peptides is targeted primarily toward microorganisms rather than the host cell (Fadnes et al., 2009; Matsuzaki, 1999; Seo et al., 2010; Yeaman & Yount, 2003). In the small intestine for example, paneth cells, which are located in the crypts of Lieberkühn, secrete large amounts of antimicrobial peptides and concentrations of cationic anti-microbial peptides can reach very high levels in the crypts of Lieberkühn (Ayabe et al., 2000; Bevins, 2005; Muller et al., 2005; Ouellette & Bevins, 2001; Ouellette, 2005, 2011) without any reported adverse effects to the host. bLFcin, like other naturally occurring cationic anti-microbial peptide, is relatively non-toxic toward mammalian cells (Fadnes et al., 2009; Fadnes et al., 2011). A hypothetical mechanism by which ingestion of bLF is able to promote expression of IFNA in the colon is the following: Microbicidal peptides generated as a consequence of digestion of bLF in the stomach kill a small fraction of intestinal microbes. This produces microbial antigens such as lipopolysaccharide, lipoproteins, flagellin, and unmethylated CpG-containing DNA. These antigens are able to interact with pattern recognition receptors (PRRs) (Kawai & Akira, 2010; Kumagai & Akira, 2010; Takeda et al., 2003; Takeda & Akira, 2005) expressed by intestinal epithelial and immune cells (Abreu et al., 2005; Artis, 2008; Hall et al., 2008; O'Hara & Shanahan, 2006; Yrlid et al., 2006) and induce expression of type I interferons (Yrlid et al., 2006), a target of PRR/TLR signaling pathways (Kawai et al., 2004; Kawai & Akira, 2010; Paun et al., 2008; Takeuchi & Akira, 2010). In addition, generation of microbial antigens in the small intestine could enhance the microbicidal activity of Paneth cells (Ayabe et al., 2000; Muller et al., 2005; Santaolalla et al., 2011; Santaolalla & Abreu, 2012), generating additional microbial antigens able to promote signaling in the colon. In support of this model, it is known that microbial antigens induce expression of IFNA in the developing intestine in mice (Mirpuri et al., 2010).
Incidentally, this model of the mechanism of action of ingested bLF predicts that persons exposed to higher levels of intestinal microbial antigens, for example through ingestion and gastric digestion of impure food and water, would have higher levels of cytokine expression in their intestinal mucosa compared to persons living in a more pristine environment. This leads to the possibility that for microbial antigen exposed persons, ingestion of bLF would have less (or no) effect on intestinal cytokine expression.
4. Our model can explain the results of bLF in animal studies
Studies using animal models have demonstrated that ingestion of bovine lactoferrin (bLF) is able to induce cytokine expression (including expression of type I interferons) in the intestine, enhance immune function, and inhibit carcinogenesis in the colon and other organs of experimental animals (Artym et al., 2003; Artym et al., 2004; Artym et al., 2005; Bhimani et al., 1999; Iigo et al., 2004; Iigo et al., 2009; Kuhara et al., 2000; Kuhara et al., 2006; Masuda et al., 2000; Sekine et al., 1997a; Sekine et al., 1997b; Shin et al., 2005; Takakura et al., 2006; Teraguchi et al., 2004; Tsuda et al., 2010; Ushida et al., 1999; Wakabayashi et al., 2004; Wakabayashi et al., 2006; Wang et al., 2000; Welsh et al., 2011). It is highly unlikely that ingested bLF directly inhibits carcinogenesis at remote sites in experimental animals: As discussed by Yao et al., 2013, transport of ingested bLF or bioactive bLF fragments from the intestine to a remote site of action is extremely unlikely. For example, ingested bLF can not be detected in the blood of adult rats (Teraguchi et al., 2004). In addition, in the trial reported by Kozu et al., 2009, bLF was undetectable in the serum of patients ingesting 3.0 g bLF daily for one year. Also, as discussed in section 1, neither ingested bLF nor its fragments reach the colon. Finally, as noted above, in all of our work with bLF, bLF itself (up to 1000 µg/ml) has never exhibited any direct toxic effect toward normal or transformed cells in culture. Therefore, available evidence indicates that levels of lactoferrin required to directly inhibit carcinogenesis in the colon or at remote sites are not attained through ingesting bLF. Rather, it is far more likely that ingested bLF inhibits carcinogenesis via modulation of immune activity.
The mechanism by which ingested bLF is able to modulate immune activity is currently unknown. However, after decades of study, there is no direct evidence that ingested bLF binds to a lactoferrin receptor that is able to transduce a signal into the cell and stimulate immune cell activity or cytokine production. The only known lactoferrin binding receptor in the intestine capable of stimulating immune cell activity and cytokine production is TLR4 (Ando et al., 2010). However, two points argue against bLF-mediated TLR4 signaling being a primary effector of bLF activity. First, TLR4 expression and response to agonist is low in adult animals (Abreu et al., 2005; Lotz et al., 2006; Takenaka et al., 2007). Second, while the effect that digestion of bLF has on its binding to TLR4 is unknown, it is almost certainly detrimental; although, the binding affinities of bLF and its peptide fragments to mouse or rat TLR4 and bLF activation of TLR4-mediated signaling pathways have yet to be reported. In addition, as discussed above, while ingested bLF is able to promote expression of IFNA in the human colon, available evidence indicates that promotion of IFNA expression in the human colon is not mediated by bLF or bLF fragments binding to TLR4, suggesting the possibility that promotion of cytokine expression in the intestines of experimental animals may also not be mediated by binding of bLF or bLF fragments to TLR4.
In contrast to hypothetical signaling mediated by TLR4 or an as yet uncharacterized bLF receptor, each step in the model proposed in section 3 is well documented: Gastric digestion of ingested bLF generates peptides with microbicidal activity; killing of a small fraction of intestinal microflora by these peptides will generate microbial antigens; these antigens are able to activate pattern recognition receptor (PRR) signaling and induce expression of cytokines. This pathway can account for the induction of cytokine expression seen in the intestines of experimental animals fed bLF (Iigo et al., 2004; Iigo et al., 2009; Kuhara et al., 2000; Kuhara et al., 2006; Takakura et al., 2006; Teraguchi et al., 2004; Tsuda et al., 2010; Wakabayashi et al., 2006; Wang et al., 2000). Induction of cytokine expression, such as type I interferons, would explain how bLF in the diet results in priming/activation of immune cells in experimental animals and enhancement of immune function (Artym et al., 2003; Artym et al., 2004; Artym et al., 2005; Bhimani et al., 1999; Shin et al., 2005; Teraguchi et al., 2004; Wakabayashi et al., 2004; Wang et al., 2000; Welsh et al., 2011). Finally, the cytostatic activity of IFNA (Bekisz et al., 2010) coupled with cytokine induction and priming of immune effector cells has the potential to inhibit progression of precancerous intestinal polyps into cancers and primed immune cells are more likely to become activated by tumor antigens, explaining how bLF can inhibit carcinogenesis in the colon and other organs of experimental animals (Iigo et al., 2009; Kuhara et al., 2000; Masuda et al., 2000; Sekine et al., 1997a; Sekine et al., 1997b; Tsuda et al., 2010; Ushida et al., 1999).
As noted at the end of section 3, our model of the mechanism of action of ingested bLF suggests that for persons exposed to higher levels of intestinal microbial antigens, ingestion of bLF would have less (or no) effect on intestinal cytokine expression. Our model also predicts that experimental animals housed in less sanitary conditions may not respond as well to ingested bLF as animals housed in SPF facilities.
5. Induction of IFNA can explain the results of the Tokyo trial
IFNA proteins are type I IFNs (Pestka, 2007). All type I IFNs bind to the same receptor, which is composed of IFNR1 and IFNR2 (de Weerd et al., 2007). IFNA-induced signaling impinges on the expression of more than 100 genes (Der et al., 1998); overall, however, the actions of IFNA on target cells can be divided into two basic activities: (i) induction of a state in which the cell does not support the proliferation of infective microorganisms (Goodbourn et al., 2000; Stetson & Medzhitov, 2006) and (ii) activation of immune effector cell function (Hervas-Stubbs et al., 2011).
In the Tokyo trial reported by Kozu et al., 2009, ingestion of 3.0 g bLF daily for 1 year resulted in inhibition of polyp growth in trial participants 63 years old or younger. The first activity ascribed to interferons was inhibition of viral proliferation (Isaacs & Lindenmann, 1957; Isaacs et al., 1957; Nagano et al., 1954; Watanabe, 2004). IFNA mediates this effect by modulating numerous metabolic and apoptotic pathways, resulting in (i) direct activity by the cell against the virus (many of these activities will also inhibit proliferation of the infected cell); (ii) direct inhibition of cell growth and proliferation, which will also inhibit the proliferation of the virus; and (iii) up-regulation (priming) of pro-apoptotic pathways, which will also inhibit virus proliferation in IFNA primed cells when they become infected by the virus (Bekisz et al., 2010; Borden et al., 2007; de Veer et al., 2001; Goodbourn et al., 2000; Sadler & Williams, 2008; Trinchieri, 2010). In general, IFNA alone is not sufficient to induce apoptosis (Trinchieri, 2010); therefore, in the absence of additional pro-apoptotic stimuli (such as viral infection), IFNA signaling will be cytostatic. Induction of IFNA in the intestine by bLF and subsequent IFNA-mediated inhibition of cell growth and proliferation can account for the inhibition of polyp growth seen in the Tokyo trial. While not examined in the study, the cytostatic activity of IFNA2 is also expected to affect the proliferation of intestinal cells in general.