Driven by economic demands and consumer preference for low fat and high protein sources, commercial broiler chickens have been intensively selected for high growth rate and high breast yield (Griffin and Goddard, 1994). Poultry meat and egg production has seen the largest increase during past decades and supports the livelihoods and food security of billions of people worldwide. However, this spectacular progress is also accompanied by several undesirable changes including hyperphagia (Denbow, 1994), metabolic disorders (Julian, 1998; Velleman, 2015), and hypersensitivity to high environmental temperature due to high metabolic activity and lack of sweat glands (Settar et al., 1999; Huang et al., 2015; Velleman, 2015).

Heat stress (HS) is detrimental to poultry production from its strong adverse effects on feed intake, growth, meat yield, welfare, and mortality (Dale and Fuller, 1980; Cahaner and Leenstra, 1992; Deeb and Cahaner, 2002; Deeb et al., 2002). HS resulted in an estimated total annual economic loss to the U.S. poultry industry of more than one-hundred million dollars and, therefore, represents a serious financial burden (St-Pierre et al., 2003). At the animal level, its adverse effects can range from discomfort (mild stress) to multiple organ damage and, under extreme conditions (sever stress), leads to mortality following spiraling hyperthermia. At the cellular level, HS induces many alterations such as, protein misfolding and aggregation (Ruan et al., 2017), cell cycle arrest (Trotter et al., 2001), redox state (Keyse and Emslie, 1992), and transcription modulation (Sanchez et al., 1994). Depending on the type, severity, and duration of the stress, cells can develop highly efficient stress response and protein quality control systems to ensure their survival or activate stress signaling cascades that proceed into cell-death pathways. These responses are controlled by a complex molecular network that is still not completely defined particularly in avian species. At the molecular level, a general rapid response to HS is increased synthesis of heat shock proteins (HSPs). Based on their monomeric molecular size, these ubiquitously expressed chaperones are classified into about six families, with HSP70 and HSP90 being the most extensively studied. Besides their classical roles as molecular chaperones and housekeepers (folding/unfolding, assembly/disassembly), HSPs are now understood to play a pivotal role in many cellular processes including transport and trafficking, protein degradation, and cell signaling (Drew et al., 2014). In turn, these cellular alterations induced by HS can lead to various neuroendocrine, physiological, and immunological adaptations including modulation of lipid and glucose metabolism (Sanz Fernandez et al., 2015; Victoria Sanz Fernandez et al., 2015).

The liver is a highly metabolic tissue that plays a vital role in digestion (bile production), metabolism, immunity, and storage of nutrients as well as detoxification. It has been shown that the liver is responsive and susceptible to HS (Lin et al., 2006; Tang et al., 2015) and that HS alters the hepatic lipid metabolism (Faylon et al., 2015; Dalvi et al., 2017; Jastrebski et al., 2017). Drinking Morinda citrifolia (Noni) juice has been shown to enhance hepatic antioxidant capacity, improve lipid homeostasis and protect the liver from environmental and chemical stressors (Wang et al., 2008a,b; Chang et al., 2013; Lin et al., 2017). The botanical name of this Rubiaceae plant is originally derived from the two Latin words “morus” ascribing to mulberry, and “indicus” imputing to Indian. It is known as Noni in Hawaii, mulberry in India, mengkudu in Malaysia, nhaut in Southeast Asia, painkiller bush or cheese fruit in Caribbean, Canary Wood in Australia, and many other names in different countries (for review see Potterat and Hamburger, 2007). This medicinal plant has been widely used in human nutrition and health (Issell et al., 2009). Due to the recent concerns in the sub-therapeutic use of antibiotics in food animal industries, Noni fruits and leaves have also gained considerable popularity in animal nutrition (Brooks et al., 2009; Sunder and Kundu, 2015; Sunder et al., 2015). However, the effect of Noni on chicken liver metabolism under control and HS conditions is not known.

As chicken liver is the main site for de novo fatty acid synthesis (lipogenesis) and also a site for fat storage, and as there is a subtle balance between hepatic lipogenesis and lipolysis, we aimed to determine in the present study the effects of Noni-enriched diet on growth performance, circulating metabolite and hormone levels as well as on the expression of hepatic lipogenesis- and lipolysis-related genes in broiler chickens exposed to acute and chronic HS.

HS adversely affects poultry production through depressed feed intake and growth as shown in many previous reports (Dale and Fuller, 1980; Cahaner and Leenstra, 1992; Leenstra and Cahaner, 1992; Deeb et al., 2002) and confirmed by the present study. These effects are mediated by numerous molecular, neuroendocrine, and physiological processes including nutrient portioning and lipid metabolism that are not completely defined, at least, in avian species.

As commercial broilers are fed lipid-poor diets (< 10%), the majority of the accumulated fat is derived from the liver which is the main site for lipogenesis (Goodridge and Ball, 1967; Leveille et al., 1968; Yeh and Leveille, 1971). Indeed, the avian liver is responsible for more than 90% of de novo fatty acid synthesis and is also a site of lipid storage in birds (Goodridge and Ball, 1967; Leveille et al., 1968; Yeh and Leveille, 1971). The increased levels of hepatic FASN and ACLY proteins, rate limiting enzymes in de novo fatty acid synthesis, indicates that HS induce hepatic lipogenesis in chickens which corroborates previous study reported by Geraert et al. (1996) and Jastrebski et al. (2017). However, the upregulation of p-ACCα^Ser79 by HS, which indicates that the enzyme is less active, is intriguing. Although it is not clear at this time-point why ACCα is down-regulated, it is possible that a compensatory increase in another ACC isoform such as, ACCβ (ACC2) is involved. Furthermore, as its regulation is complex, it is possible that the down regulation of ACCα might be due to a negative feedback loop. For instance, the ACCα decrease could result from an increase in glucagon and/or catecholamine levels which has been reported to be induced by stressful stimuli including HS and feed deprivation (Bloom et al., 1973; Iguchi et al., 2012). Although we didn't determine the plasma glucagon levels in the present study, the significant increase in circulating glucose concentration (23%) by HS, despite the significant reduction in feed intake, supports the above mentioned hypothesis. In fact, during HS along with reduced energy intake, glucagon plays a key role in maintaining glucose homeostasis via stimulating hepatic glycogenolysis and gluconeogenesis (Exton and Park, 1968; Garrison and Haynes, 1973; Watford, 1985) which is evidenced in our study by the upregulation of hepatic phosphoenolpyruvate carboxykinase (PEPCK), glycogene phosphorylase (GLGP), and fructose 2,6-biphosphatase 4 (PFKFB4) (data not shown). An additional possibility is that HS activates the energy sensor AMPKα1/2 (data not shown) and AMPK has been shown to phosphorylate ACCα^Ser79 (Park et al., 2002). We should note that we measured here only one ACC phosphorylation site (Ser79), however there are several other sites including Ser1200 and Ser1215 (Brownsey et al., 2006) that might be differently regulated and phosphorylated by HS.

HS down regulated the hepatic expression of all tested lipogenic genes as well as their related transcription factors which is different from previous integrated transcriptomic studies reported in rodents and chickens (Bhusari et al., 2007; Jastrebski et al., 2017). This discrepancy might due to the technique used and the choice of the normalization (housekeeping) control (real-time qPCR in our study vs. RNAseq or microarray in previous studies) and/or experimental conditions (species, age, duration, intensity, and/or severity of HS, diet composition, etc.). The absence of correlation between gene and protein expression in our study indicates that the transcription and translation of these target markers are differentially regulated by HS which is not surprising. It is plausible that under our experimental stress conditions, the pool of already synthesized lipogenic mRNA might be efficiently translated with a quick turnover or degradation, however its protein product accumulation increases and remain in the cellular pool due to its high half-life (Schwanhausser et al., 2011). The protein stability may also increase by post-translational modification like acetylation or glycosylation. An additional possibility is that mRNA translation can be upregulated by RNA binding proteins and/or down-regulation of micro (mi) RNA targeting that mRNA.

Interestingly, our study showed that the lipolytic pathway is also expressed in the chicken liver which is a considerable site for lipid storage and secretion (Hermier, 1997). HS seems to regulate the expression of hepatic lipolysis-related genes in a time-specific manner. For instance, LPL mRNA abundance was not affected by acute HS, but it was decreased by chronic HS. Both acute and chronic HS up regulate hepatic ATGL, but decreased LIPC expression which may explain the unchanged levels of plasma triglycerides. As LPL has other functions beside its hydrolytic activity, it is conceivable that HS might affect the maturation of plasma HDL (Strauss et al., 2001), uptake of lipoprotein-associated lipids and vitamins (Rinninger et al., 1998, 2001; Goti et al., 2002), and/or degradation of lipoprotein (Heeren et al., 2002) which warrant further investigations. Of particular interest, a novel function of LPL in the regulation of appetite and energy balance regulation has recently been postulated (Wang et al., 2011). Thus, it is possible that the reduced feed intake and body weight induced by HS in this study might be mediated via down regulation of LPL expression and this merit further in depth studies.

As expected and in agreement with previous studies (Xie et al., 2014; Wang et al., 2015), HSPs (HSP60, HSP70, and HSP90) were all upregulated by acute HS, but remained unchanged after longer HS exposure compared to TN conditions. This later observation is not surprising and indicates that the birds become acclimated to chronic HS which in turn results in no further increase in cellular HSPs (Givisiez et al., 1999). Noni supplementation reduced the expression of HSP90, but not that of HSP60 and HSP70-induced by acute HS suggesting that the regulation of HSP by Noni is family (molecular weight)-dependent. However, contrary to our expectation, Noni-enriched diet upregulated the hepatic expression of HSP60 and HSP90 in broilers maintained under TN conditions, as well as that of HSP60 in chronic heat-stressed broilers without affecting the expression of HSP70. The down regulation of hepatic HSP90 during acute HS supports the notion that Noni might alleviate stress induced by heat load and this has been substantiated by the reduction of core body temperature-induced by HS (data not shown). However, the upregulation of both HSP60 and HSP90 during the chronic but not acute TN conditions is intriguing. It has been reported that Noni juice induced stress and reduced body weight in mice (Muto et al., 2010). This is an unlikely scenario in our experimental conditions because birds fed control or Noni diet showed similar feed intake, body weight, and body weight gain without any sign of discomfort or stress (physical activity and behavior). In addition, necropsy analysis of all major organs in both groups revealed no pathology. Although the physiological meaning and biological determinants are not known at this time point, it is foreseeable that age (3 vs. 6 weeks in acute vs. chronic conditions, respectively) might impact the effect of Noni on the hepatic HSP60 expression. Age-dependent differences in liver gene expression has been recently reported (Uno et al., 2017) and HSP60 levels have been found to progressively decline with increasing age in humans (Rea et al., 2001). Besides their role in stress response, HSPs act also as molecular chaperones under normal physiological conditions where they carry out such essential functions as protein translocation, folding, and assembly (Hendrick and Hartl, 1993). It is, therefore, possible that HSP might mediate the effect of Noni in regulating these cellular processes, however owing to the gap in avian HSP biology, experimental evidence using mechanistic and functional studies supporting these possibilities needs to be provided.

The lipogenic proteins followed a similar hepatic expression pattern as HSPs across our experimental conditions. For instance, HSPs (HSP70 and HSP90), FASN, and ACLY showed similar expression patterns in the liver of acute heat-stressed broilers. Similarly, the hepatic expression of HSP60 and FASN follow the same trend under chronic HS conditions. As HSPs are now understood to be involved in many cellular processes from transcription regulation to cell death signaling (Rutherford and Zuker, 1994), our data suggest that HS might modulate the expression of lipogenic protein through HSPs. Tang and co-workers (Tang et al., 2016) have shown that HSP60 silencing alter directly the expression of FASN and ACCα in glioblastoma cells. FASN has also been recently identified as a bona fide HSP90 client (Fierro-Monti et al., 2013). Zhang et al. (2012) reported that HSP90 binds to AMPK and regulates ACCα expression.

To gain a better understanding of HS responses, we next evaluated the effect of HS alone or in combination with QCT on the same parameters by using chicken hepatocyte cell culture. We chose QCT because it is a highly bioactive ingredient of Noni with an average concentration of 7.4 mg/g (CP Foods, personal communication, Kampkotter et al., 2007). Although the optimal temperature of cell culture (37°C) and chicken body temperature (~41°C) are quite different, HS induced the expression of HSP60 and HSP70 in sim-CEL cells comparably to chicken liver tissues. However, it down-regulated the expression of lipogenic (ACCα, FASN, and ACLY) proteins in vitro which is opposite to that in animals. This indicates, firstly, that HS may directly alter the expression of hepatic lipogenic genes independently of the reduction of feed intake. Secondly, the differential expression profile of hepatic lipogenic proteins between the in vivo and in vitro HS studies might be due to the involvement of other organs such as, the brain (central neural pathways for thermoregulation) (Morrison and Nakamura, 2011) and the hormonal milieu. In fact, we have previously shown that HS altered the hepatic expression of leptin (Dridi et al., 2008) which is a key regulator of lipogenic genes (Dridi et al., 2005). Additionally, HS has been shown to dysregulate several other appetite-related hormones including ghrelin, thyroid hormone, and adipocytokines (i.e., adiponectin) and these hormones orchestrate the regulation of hepatic lipogenic genes (Yin et al., 2005; Tao et al., 2006; Buyse et al., 2009; Morera et al., 2012). QCT treatment induces the expression of lipogenic protein in sim-CEL cells maintained under both control and short (4 h) HS exposure, indicating that a short-term control mechanism of lipogenesis by QCT is effective. Moreover, the prompt responsiveness of cultured hepatocytes to QCT suggest that the QCT effect may be direct rather than secondary. The disaccord of QCT effects between chicken and mammalian hepatocytes is striking. It has been reported that QCT administration inhibits FASN expression in human HepG2 (Li et al., 2013; Zhao et al., 2014), rat H4llE (Zhou et al., 2014), and rat BRL-3A cells (Wang et al., 2013). This discrepancy might be due to dose and/or species-specific regulation of lipogenic genes by QCT (Dridi et al., 2006). In fact, QCT has been shown to have a dose-dependent biphasic effect on mammalian cell proliferation (van der Woude et al., 2003) and on CYP1A2 expression (Kang et al., 2004).

In conclusion (Figure 8), HS induces the hepatic expression of HSPs and stimulates de novo fatty acid synthesis in broilers which may explain the increased fat deposition observed during hot season (Geraert et al., 1996). In this study, Noni supplementation did not ameliorate feed intake and growth depression under HS conditions, but it did modulate the hepatic expression of HSPs in a time-specific manner suggesting a potential role in stress response that merit further investigations. Possibly, the lack of effect of Noni in the present study could be attributed to clean litter and low relative humidity (20%) in the experimental environmental chambers. A less pristine environment such as, that found in commercial poultry houses may be required to demonstrate potential benefits of Noni.

JF, HR-S, and EG performed the experiments. LB and BH generate the immunofluorescence data. LE and TP measured the thyroid hormone levels. SD, AD, and WB purchased the reagents. SD designed the experiment and wrote the manuscript.