In the current study, despite a severe dietary iron restriction, only a moderate degree of anemia (hemoglobin ~ 80 g/L) was elicited. Interestingly, even a moderate induction of iron deficiency appears to contribute to elevations in both steady-state levels of serum glucose and insulin regardless of basal diet formulation. Others have postulated that an increase in serum glucose may be due, at least in part, to an elevation of cortisol observed in severely anemic rats[21, 22]. Interestingly, the relative decrease in cortisol in the ID groups in the current study suggests that other mechanisms may be responsible for the presence of hyperglycemia. In fact, the results of this study are in agreement with those obtained in a human study wherein patients with severe iron deficiency exhibited reduced cortisol secretion. In addition to the relative hyperglycemia, a relative hyperinsulinemia was observed in the ID animals in both groups as well. These metabolic adaptations presumably occur as a compensatory means as ID animals preferentially utilize glucose, rather than fat, as a metabolic substrate for peripheral tissues as a result of decreased hemoglobin levels and subsequent decreased oxidative capacity[4, 9, 10]. Thus, blood glucose levels likely remain elevated to ensure that adequate fuel substrate is available for energy production, and insulin levels may remain elevated to facilitate the entry of glucose into insulin-dependent tissues.
In an effort then to further interrogate the underlying metabolic changes that occur with iron deficiency, we examined the hepatic expression of 84 genes involved in maintenance of glucose homeostasis. The level of significance (fold change ≥ 1.3), while small, was considered significant as previous studies have shown that dietary intervention(s) tend to elicit somewhat modest, yet biologically meaningful, transcriptional responses[24, 25]. The ID rats in each group exhibited modest, but significant alterations in the expression of genes representative of glucose metabolism. Notable changes in gene expression include those genes associated with metabolic pathways including both glycolysis and gluconeogenesis and are in agreement with the findings of others[2, 10, 26]. The significant increase in Gck expression is likely due to the relative increase in circulating insulin levels observed in the ID groups, as insulin is a known inducer of hepatic Gck mRNA expression[27, 28]. Increased expression of Gck could potentially be very important as ID animals have been shown to have an increased reliance on glucose as a metabolic substrate, and Gck is able to rapidly increase the rate of glucose phosphorylation in the liver in response to the elevations in blood glucose levels. Furthermore, as Gck catalyzes the first step in hepatic glucose utilization it can contribute multiple pathways including glycogen synthesis, glycolysis, and de novo lipogenesis which could explain the enhanced glucose utilization and hyperlipidemia reported in response to dietary ID[2, 10, 26, 29, 30].
The significant increase in Gck mRNA expression, presumably in response to the elevated insulin levels, suggests that insulin signaling in ID animals is at least partially intact. However, previous observations suggest that alterations in metabolic gene expression are indicative of an impaired hepatic insulin response wherein ID animals exhibited a form of mixed insulin resistance. Under normal conditions (e.g., iron sufficiency) when insulin levels are elevated in a fed state, insulin acts both to repress gluconeogenic gene expression and to simultaneously activate lipogenic gene expression[31, 32]. In contrast, chronic hyperinsulinemia contributes to a combination of hepatic insulin resistance in which the insulin-dependent activation of lipogenic gene expression remains intact, but gluconeogenic gene expression is inadequately repressed[5, 30, 33]. In this model of mixed insulin resistance, insulin acts through the mammalian target of rapamycin complex 1 to activate lipogenesis via a Srebp1c-dependent increase in lipogenic gene expression, whereas insulin-induced phosphorylation of the transcription factor forkhead box protein O1 is diminished such that gluconeogenic gene expression remains inappropriately active. Thus, mixed insulin resistance remains a candidate mechanism explaining the relative hyperglycemia and hyperlipidemia reported in ID animals. Despite changes in hepatic insulin signaling, peripheral tissue insulin sensitivity as assessed by glucose clearance appears to be enhanced with ID[2, 4]. The extent to which there are tissue-specific differences in insulin signaling in response to iron deficiency warrants additional investigation.
To further investigate potential factors contributing to hypertriacylglycerolemia that has been reported in iron-deficient animals, changes in hepatic lipogenic gene expression were assessed in the PF and ID groups on both diets. In fact, it was the finding of increased lipogenic gene expression in iron-deficient animals consuming an AIN-76-based diet that stimulated the comparison of the AIN-76 and AIN-93 diets. As early as 1982 it was speculated that causes of metabolic complications observed in animals consuming an AIN-76 diet were related to sucrose, the primary carbohydrate source in the AIN-76 formulation. Not surprisingly, basal levels of the lipogenic genes Acly and Fas were higher in the 76-PF group compared to the 93-PF group, though the expression of the lipogenic master transcriptional regulator Srebp1c remained largely unchanged. Nonetheless, compared to the PF group, lipogenic gene expression was further significantly enhanced only in animals consuming an iron-deficient AIN-76 diet. Elevated levels of TAG in the 76-ID group suggest a functional consequence of enhanced lipogenic gene expression resulting is de novo fatty acid synthesis and packaging into TAG destined for secretion from the liver. Although the levels of liver TAG were not determined in the present study, taken together these data provide compelling evidence that consumption of an iron-deficient yet high-sucrose diet may result in more severe metabolic complications leading to hyperglycemia, hyperinsulinemia, and hyperlipidemia.
Despite the differences observed in lipogenic gene expression and serum TAG between diets and consistent with the findings of others that β-oxidation is decreased in response to dietary ID, the mRNA expression of Cpt1 was found to be significantly reduced in both the 76- and 93-ID groups. The expression of the iron-containing enzyme Scd1 was also decreased in the 76-ID group, with the same trend being observed in the 93-ID group. Importantly, expression of the mRNA encoding for the iron uptake protein Tfr1 was significantly increased while expression of the iron sensor Hamp1 was significantly decreased in both ID groups. Furthermore, the relative abundance of both Tfr1 and Hamp1 mRNA was not different between diets for either the PF or ID animals indicating a similar response to dietary ID, regardless of basal diet. No differences in serum iron or non-heme liver iron between the ID groups on either diet also show that a similar degree of iron deficiency was induced in both ID groups. The moderate level of iron deficiency achieved in this study may serve as an explanation for the differences noted in lipogenic gene expression and serum TAG. Previous work investigating the metabolic lipid response to varying degrees of iron deficiency on an AIN-93 based diet only observed a significant increase in serum TAG at hemoglobin ≤ 66 g/L, a level which is consistent with data suggesting that the severity of metabolic responses to iron deficiency is most highly associated with reductions in hemoglobin[2, 4, 8, 12, 35]. Interestingly, despite a more modest induction of iron deficiency for the current study (hemoglobin = 80 g/L), relative elevations in serum glucose and insulin levels were still observed in both the 76-ID and 93-ID groups.
Our findings support the work of others that have clearly demonstrated that even a moderate induction of iron deficiency is sufficient to disrupt normal glucose homeostasis in rodents[2, 3, 8–10, 12, 26]. To date, relatively little evidence is available for humans, but an association between ID and elevated hemoglobin A1C (HbA1c) levels has been observed. Indeed, as iron status is improved, levels of HbA1c return to more normal levels. Interestingly, despite what some attribute to so-called over-nutrition, ID is also commonly observed in overweight and obese individuals, a population in which metabolic homeostasis is often disrupted[37–39]. The potential for ID to contribute to or exacerbate conditions wherein normal glucose and fat metabolism are already impaired will be the focus of future investigations.
Currently, the specific mechanisms contributing to iron deficient-induced hyperglycemia remain elusive, but the findings presented herein support the hypothesis that a depletion of iron status, regardless of dietary carbohydrate source, is sufficient to disrupt systemic glucose homeostasis in a weanling rat model of iron deficiency. Intriguingly, enhanced expression of the lipogenic genes Srebp1c, Acly, and Fas was only observed in the ID-76 group. The extent to which these results are suggestive of an iron-level by carbohydrate (i.e., sucrose) interaction remains unknown and is the focus of future studies. To this end, it will also be important to closely examine the changes in nutrient sensing and insulin signaling that occur in response to dietary iron deficiency across different tissues such as the liver and skeletal muscle. The significant increase in Gck expression in both ID groups provides some of the first insight into the means by which glucose uptake and utilization is altered in response to iron deficiency[2, 9, 26]. Future work should focus on the biological significance of this and the other somewhat modest transcriptional changes that occurred with particular attention to the allocation of glucose to various pathways.