This study was designed to test the hypothesis that long-term hypoxia (LTH) increases fetal plasma leptin and fetal adipose or placental leptin expression and alters hypothalamic and adrenocortical leptin receptor (OB-R) expression. Pregnant ewes were maintained at high altitude (3,820 m) from day 30 to ∼130 days of gestation. Reduced Po2 was maintained in the laboratory by nitrogen infusion through a maternal tracheal catheter. On day 132, normoxic control and LTH fetuses underwent surgical implantation of vascular catheters (n = 6 for each group). Five days after surgery, maternal and fetal arterial blood samples were collected for leptin, insulin, and glucose analysis. Placental tissue, periadrenal fat, and fetal hypothalami and adrenal glands were collected from additional control (n = 7) and LTH (n = 8) fetuses for analysis of leptin mRNA by quantitative, real-time, RT-PCR (qRT-PCR). There was a significant (P < 0.03) elevation in fetal plasma leptin in the LTH fetuses (3.5 ± 0.7 ng/ml) vs. control (1.1 ± 0.1 ng/ml). There were no differences in either glucose or insulin concentrations between the two groups. Periadrenal adipose leptin mRNA was significantly higher in the LTH group compared with control, as was placental leptin expression. The levels of leptin mRNA in adipose were ∼70 times higher vs. placenta. LTH significantly reduced expression of OB-Ra (short-isoform) in the hypothalamus (P = 0.0156), while resulting in a significant increase in adrenal OB-Rb (long-form) expression (P < 0.03). Our data suggest that leptin is a hypoxia-inducible gene in the ovine fetus and OB-R expression is altered by LTH. These changes may be responsible in part, for our previously observed alterations in fetal hypothalamic-pituitary-adrenal function following LTH.
hypoxia represents a potential threat to the developing fetus and is a potent stimulator of the fetal stress response, increasing both fetal plasma ACTH and cortisol (5, 8). This activation of the hypothalamic-pituitary-adrenal (HPA) axis helps the fetus survive the acute hypoxic insult. Both the duration and magnitude of the stress response are proportional to the degree and duration of the hypoxic insult. Different experimental methods have been used to induce hypoxia in the sheep fetus with the duration of hypoxia ranging from acute [minutes to hours (5)] to prolonged [several hours to several days (16)] to chronic [a few weeks (35)].
Our laboratory has focused on the effects of long-term hypoxia (LTH) using a model in which the ewe is maintained at an altitude of 3,820 m from day 30 of gestation to near term (1, 22). During this time, the maternal arterial Po2 is ∼60 mmHg and fetal arterial Po2 is ∼17–19 mmHg. We have shown that the HPA axis of fetal sheep undergoes significant adaptation following LTH. Under conditions of LTH, anterior pituitary processing of proopiomelanocortin (POMC) to ACTH1–39 and basal plasma ACTH1–39 concentrations are elevated (36) and there is enhanced cortisol secretion in response to a secondary stressor (1, 22). We also observed that the adrenal cortex of LTH fetuses exhibits reduced expression of the ACTH receptor coupled with decreased expression in key steroidogenic enzymes cytochrome P-450 17α-hydroxylase (CYP17) and cytochrome P-450 side-chain cleavage (CYP11A1) (37). As a result of these adaptations, the LTH fetus maintains normal basal plasma cortisol, whereas plasma cortisol concentrations achieved in response to a secondary stressor are greater. Cumulatively, this HPA axis adaptation likely aids in the survival of the LTH-compromised fetus confronted with a potentially threatening secondary stress, while preventing preterm delivery that would occur following a premature rise in basal fetal plasma cortisol concentrations. However, the mechanisms mediating this striking adaptation of the HPA axis remain to be elucidated.
Leptin is a 16-kDa polypeptide that is expressed primarily in adipose tissue. The product of the obese (ob) gene, this hormone has a wide range of physiological functions, including energy homeostasis, bone formation, reproduction, and effects on the cardiovascular system (2). Leptin also exerts effects on the HPA axis. For instance, leptin inhibits restraint stress-induced activation of the HPA in the adult rat (18). In vitro studies using cultured bovine adrenal cortical cells demonstrated that leptin suppressed cortisol output in response to ACTH stimulation and that this effect was through a reduction in CYP17 expression (9). Additional studies confirmed this observation in bovine adrenocortical cells and extended the inhibitory effect of leptin to include CYP11A1 (25). Leptin has also been shown to inhibit ACTH-stimulated cortisol production in isolated human adrenal cortical tissue (41).
Importantly, leptin also appears to play a potential role in the regulation of HPA function in the fetus. Intracerebral infusion of leptin blunted the increase in amplitude of plasma ACTH and cortisol pulses, as well as mean plasma ACTH and cortisol concentrations in near-term fetal sheep (21). Recent studies by McMillen et al. (32), showing that intravenous leptin infusions suppressed the prepartum rise in both ACTH and cortisol, also support an inhibitory role for leptin on the fetal ovine HPA axis during late gestation.
Leptin mediates its actions via the leptin receptor (OB-R). Two major splice variants of the leptin receptor (OB-R) exist, encoded by a single gene. The OB-Rb, or “long form”, of the receptor is considered the most biologically active form, whereas the short isoform (OB-Ra), lacking most of the intracellular domain involved in mediating signal transduction, represents the other major isoform and is generated by alternative exon use. Both OB-Rb and OB-Ra have identical extracellular ligand-binding domains and have been shown to exhibit similar affinity for leptin (7). The OB-Rb isoform is highly expressed in the hypothalamus, whereas the OB-Ra is the predominant form expressed in peripheral tissues in adults. (34, 43). Within the hypothalamus, the OB-R has been localized in the paraventricular nucleus, indicating that leptin may have direct effects on corticotropin releasing hormone/arginine vasopressin (AVP) neurons (4). Interestingly, both OB-Rb and OB-Ra are expressed in the adrenal cortex in rodents (30), humans (41), and ruminants (10). The presence of the OB-Rb in the adrenal gland is consistent with the direct effects observed for leptin on adrenocortical function.
Although the function(s) of leptin in the adult has been widely studied, the role of leptin in the fetus is not as well defined. Leptin is found in the fetal circulation and is expressed in fetal adipocytes (11, 32, 50). Leptin has also been found to be expressed in the human and ovine placental trophoblast (20), although placental expression of leptin in the ovine is considerably less than observed in humans. In the adult, leptin appears to be a hypoxia-inducible gene (23, 40, 47, 48). However, other than a correlation between fetal Po2 and plasma leptin levels in the sheep fetus (14), no data are currently available on the effects of hypoxia on fetal leptin. The present study was designed to test the hypothesis that expression and plasma levels of leptin are elevated in the LTH ovine fetus. We compared concentrations of plasma leptin and the expression of leptin in periadrenal white adipose tissue, as well as the placenta in near-term normoxic control and LTH fetal sheep. Additionally, we studied the effects of LTH on hypothalamic and adrenal leptin receptor mRNA.
MATERIALS AND METHODS
The animal protocols for this study were submitted to, reviewed, and aprroved by the Institutional Animal Care and Use Committee at Loma Linda Medical School prior to any of the studies being conducted. Time-dated pregnant ewes were maintained at the Barcroft Laboratory White Mountain Research Station (elevation 3,820 m) for ∼110 days beginning at day 30 of gestation (term = 146 days). The animals were then transported to Loma Linda University Medical Center Animal Research Facility (elevation: 346 m) where they were implanted with a nonocclusive tracheal catheter (4.0 mm OD) and an arterial catheter. Maternal Po2 for the LTH group was maintained at ∼60 mmHg (mean Po2 measured in animals at altitude) by adjusting humidified nitrogen (N2) gas flow through the tracheal catheter as previously described (1, 17, 22). Normoxic, age-matched pregnant ewes were used as controls. Between days 137 and 141 of gestation, ewes in each group were sedated with pentobarbital sodium, intubated, and maintained under general anesthesia with 1.5–2% halothane in oxygen while the fetuses were delivered through a midline laparoptomy. Fetal adrenal glands were collected, and periadrenal adipose tissue was obtained and snap-frozen in liquid nitrogen. Hypothalami were collected by using the anterior aspect of the optic chiasm as the forward boundary, posterior ∼1 cm to the mamillary body, vertically ∼1 cm, and laterally ∼0.5 cm from midline. Placental tissue was also collected and snap frozen. All tissues were stored at −80°C until analyzed. Basal maternal and fetal arterial blood samples were collected for plasma leptin determinations from control and LTH chronically catheterized animals from a previously described study (22). Fetal plasma was also analyzed for glucose and insulin concentrations.
Plasma leptin assay.
Plasma leptin concentrations were determined using the multispecies leptin radioimmunoassay (Linco Research, St. Charles, MO) using recombinant ovine leptin as standard (recombinant ovine leptin was generously provided by Dr. Duane Keisler, University of Missouri, Columbia, MO). The ovine leptin standard diluted in parallel with the human leptin standard with >90% recovery.
Plasma insulin and glucose.
Plasma insulin levels were assessed by a specific ELISA (Linco Research), while plasma glucose concentrations were measured with a YSI 2700 glucose analyzer (Yellow Springs International, Yellow Springs, OH).
Quantitative real-time PCR.
Methodology was performed using methods which we have previously described and validated (36, 37). Total RNA was prepared from fetal hypothalami, adrenal glands, periadrenal adipose tissue, and placenta (n = 7 or 8 for control and LTH groups) using an RNA preparation kit as per manufacturer's instructions (Qiagen). Subsequently, total RNA was DNase I (1 unit; Ambion) treated (60 min at 37°C), and the DNase I was removed via PCR clean-up columns (Qiagen). RT was performed using 1 μg total RNA, oligo(dT) primer, and Superscript II as RT. An initial denaturation step was performed for 5 min at 95°C before first-strand synthesis at 42°C for 50 min; the reaction was terminated by heating to 70°C for 15 min.
Real-time PCR was performed using 25–100 ng of cDNA (equal to input RNA) per PCR. All PCRs were performed in triplicate. Initial qRT-PCRs were performed using serial dilutions of cDNA ranging from 250 to 15.625 ng (250, 125, 62.5, 31.25, 15.625 ng) to determine that the quantity of cDNA used for analysis of each specific mRNA was within the linear range of amplification for each primer. For each mRNA, a starting amount of the cDNA reaction was thus chosen within the linear amplification range. For each primer set, the amplicon was subcloned into the TA cloning vector (Invitrogen) and subjected to Sanger dideoxysequencing (Oklahoma Medical Research Foundation Sequencing Core, Oklahoma City, OK) to confirm amplicon identity. SYBR Green (1× SYBR Green master mix; Bio-Rad, Hercules, CA) was used as the fluorophore, and PCR was performed using a Bio-Rad iCycler equipped with the real-time optical fluorescent detection system. The primer sequences, derived from the National Center for Biotechnology Information (NCBI) and NCBI accession numbers used are listed in Table 1, and a schematic representation of OB-Ra and OB-Rb isoforms and primer pair location are illustrated in Fig. 1. A three-step PCR was used: 95°C for 45 s, annealing (primer specific but typically 55–60°C) for 30 s, and 72°C extension for 30 s. A total of 35 cycles were performed. A melt-curve analysis was conducted on each sample after the final cycle to ensure that a single product was obtained, and agarose gel electophoresis confirmed that the single PCR product was of the expected size. We used cyclophilin as a “housekeeping” mRNA, using the identical first-strand cDNA used for quantification of specific mRNAs of interest and in the same PCR run as for the gene of interest to circumvent any between-run variation. After analysis of several candidate mRNAs, cyclophilin demonstrated no gestational age or LTH modulation of expression (28, 29). We have also found that cyclophilin and GAPDH were equally efficacious when used as internal housekeeping mRNAs in our real-time PCR applications (29). Control PCR for each primer pair and RNA source included 1) elimination of RT during first-strand cDNA synthesis (assures that PCR product depends upon RNA) and 2) no RNA/cDNA in RT reaction (assures that no amplicon contamination has occurred). Primers were used that provided 1) a single PCR product (sequenced positive), 2) dilution curve of cDNA exhibited a slope of 100 ± 10% “efficiency” where 100% = Δ3 Ct/log cDNA input (Ct is the threshold PCR cycle at which fluorescence is detected above baseline), and 3) the melt-curve analysis post-PCR must demonstrate one product. With respect to exon positions, both forward and reverse leptin primers were within exon 3; OB-Ra primers corresponded to murine exon 16 (forward primer) and murine exon 18a (reverse primer); OB-Rb primers corresponded to murine exon 18. For quantification purposes, a synthetic single-stranded DNA standard derived from the mouse Myb-1 cDNA (NM_010848) was used to generate a standard curve (100, 10, 1, 0.1, 0.01, and 0.001 pg of standard DNA) for extrapolation of starting cDNA concentrations per reaction. Each standard point was run in duplicate and in the same PCR block as the unknowns. Linear regression was used to quantify starting RNA (cDNA) based on Ct values as extrapolated from the standard curve. The efficiency of the standard and primers was 100% based on the above criteria. Ct values for leptin and OB-R isoforms were corrected for the Ct of cyclophilin in each sample based on the within-sample deviation for cyclophilin (ΔCt) from the average for the housekeeping gene before quantification of mRNA for these genes of interest. For cyclophilin, the mRNA values were determined using the actual Ct value for that gene.
Differences between control normoxic and LTH fetuses were compared using Student's t-test with significance chosen at P = 0.05.
Figure 2 illustrates the effect of LTH on maternal and fetal plasma leptin concentrations. Although there was a trend toward increased maternal plasma leptin in LTH ewes, the value compared with normoxic controls was not statistically significant. In contrast, LTH had a profound effect on fetal plasma leptin concentrations. Leptin levels were threefold higher in the LTH group (P < 0.03, compared with control). Since plasma insulin may regulate leptin (38, 39), we also measured fetal plasma glucose and insulin levels in control and LTH fetuses. There were no differences in either glucose or insulin concentrations between the two groups (Fig. 3).
Leptin mRNA in both periadrenal adipose (P = 0.025) and placenta (P = 0.022) was significantly elevated in the LTH animals (Fig. 4). It is important to note that the levels of expression in adipose tissue are ∼70 times higher than in placenta.
Similar to what has been reported in the adult, the predominant form of the leptin receptor in the fetal hypothalamus was the OB-Rb (Fig. 5). In the hypothalamus, the OB-Rb mRNA was ∼30- to 40-fold more abundant compared with OB-Ra mRNA (Fig. 5). LTH did not alter mRNA levels for the OB-Rb in the hypothalamus. However, hypothalamic OB-Ra expression was significantly lower in the LTH group compared with control (fg mRNA/100 ng cDNA, P = 0156, Fig. 5). In the adrenal gland, the OB-Ra splice variant was ∼10-fold more abundant (Fig. 6) than the OB-Rb splice variant in control fetuses (Fig. 6). OB-Rb mRNA was significantly elevated in the LTH fetal adrenals (fg mRNA/100 ng cDNA, P < 0.03, compared with control) with overall abundance being nearly equally compared with the OB-Ra isoform.
In the present study, we demonstrated that LTH exerts profound effects on circulating leptin concentrations, leptin expression in adipocytes, and leptin receptor expression in the near-term ovine fetus. We designed this study to identify a potential endocrine mechanism(s) that may mediate or contribute to the changes in function of the fetal HPA axis we previously observed in the LTH fetus (36, 37). Indeed, on the basis of reports describing the capacity of leptin to inhibit ACTH-induced adrenocortical function (9, 25, 41) and to modify function of the hypothalamic-pituitary arm of the HPA axis (4, 18), this study specifically addressed the question: does LTH alter expression of leptin in sites of its production, as well as endocrine leptin receptor expression?
In the present study, we observed a threefold increase in fetal plasma leptin concentrations in response to LTH coupled with a significant increase in white adipose (periadrenal) tissue leptin mRNA. Similarly, leptin mRNA in the placental levels was also elevated in the LTH animals. Masuzaki et al. (31) first demonstrated that the human placenta is a significant, extra-adipose source of leptin. However, species differences exist in the relative contribution of placental leptin to the fetal pool. In vitro studies with perfused human placenta demonstrated that over 95% of placental leptin is delivered to the maternal circulation (26). Similar to other reports of either no placental expression of leptin (12) in sheep or minimal placental leptin expression (present study), the level of placental leptin expression we observed was >50-fold lower compared with fetal adipose. Thus it is highly likely that the increased expression observed in fetal adipose tissue represents the source of the noted increase in circulating fetal leptin, even considering the larger mass of placental tissue compared with fetal adipose.
Currently, the regulation of plasma leptin concentrations is poorly defined in fetal sheep, although a number of factors may contribute to changes in circulating leptin, including fetal plasma insulin and glucose (11), cortisol, and Po2 (14). In the present study, it does not appear that changes in plasma glucose or insulin are responsible for elevated leptin concentrations observed in the LTH fetuses because no differences were observed in these parameters compared with normoxic controls (Fig. 3). Although expression of key steroidogenic genes is suppressed in the adrenal of the LTH fetus, compensatory increases in fetal plasma ACTH result in maintained fetal plasma cortisol concentrations (1, 17, 22). Thus it is doubtful that cortisol is involved in mediating the observed changes in leptin expression.
Interestingly, Forhead et al. (14) noted an inverse relationship between fetal Po2 and plasma leptin, indicating that adipocyte release of leptin may be acutely regulated by changes in Po2. In the adult, leptin is a hypoxia-inducible gene. Studies in mice exposed to intermittent hypoxia demonstrated an upregulation of leptin mRNA expression (40). Elevated leptin levels in individuals following hypobaric hypoxia (47, 48) or in patients with sleep apnea (23) were also observed. Hypoxia in adipocytes markedly enhances expression of leptin and also stimulates the hypoxia-inducible factor-1 pathway (28). Our data show that leptin is elevated in response LTH in the ovine fetus. Therefore, it appears that the major factor mediating the increased leptin expression observed in fetal adipose tissue is likely the exposure to LTH. As such, the present study represents the first report that leptin may be a hypoxia-inducible gene in the fetal sheep adipocyte.
At present, there is little information regarding the expression of leptin receptors and their isoforms in the developing fetus. A recent study (34) demonstrated that hypothalamic expression of the OB-Rb isoform was largely restricted to the ventromedial hypothalamus and arcuate nuclei in the late-gestation sheep fetus, consistent with its expression in other species. The authors also noted that OB-Rb mRNA increased in response to hyperglycemia. In the present study, we did not observe an effect of LTH on OB-Rb mRNA in the hypothalamus but did note a significant decrease in OB-Ra mRNA in the hypothalamus of LTH fetal sheep. The OB-Rb is considered the biologically relevant form of the receptor, activating the JAK2/STAT3 signaling pathway (6). While the OB-Ra lacks a large region of the cytoplasmic domain, including the second JAK site (Box-2 domain), recent findings indicate that this receptor can activate ERK-mediated signaling and thus, while being incapable of full leptin signaling, is, nonetheless, not an inactive OB receptor (49). However, considering the ∼50-fold lower expression of the OB-Ra compared with the OB-Rb in the hypothalamus of fetal sheep, the physiological significance of the decline in OB-Ra in the LTH hypothalamus, if any, remains to be assessed. If OB-Ra expression in the hypothalamus of fetal sheep is restricted to certain neuronal phenotypes, then this change in expression may play a role in the adaptation to LTH.
A soluble leptin receptor, circulating in a complex with bound leptin, has been described in both humans and rodents. The plasma-soluble receptor serves as a leptin-binding protein and, as such, plays an important role regulating the bioavailability of circulating leptin. In rodents the soluble leptin receptor arises from yet another alternatively spliced OB-R transcript (OB-Re), in which alternative splicing leads to a truncated receptor lacking the transmembrane domain (46). In humans, the soluble leptin receptor is generated by ectodomain shedding of plasma membrane-bound OB-R via limited protease cleavage rather than via alternative splicing (29). Rodents also appear capable of forming soluble leptin receptor by ectodomain shedding as well (15). However, to date, a soluble leptin receptor generated by ectodomain shedding or alternative splicing of the OB-R leading to the OB-Re has not been described in ruminants. Although we were unable to detect the OB-Re mRNA isoform in either hypothalamic or adrenal RNA using rodent-specific primers for the OB-Re splice variant (Myers DA, unpublished observations), further studies clearly need to be performed to assess soluble leptin receptors in sheep.
The OB-Rb is the predominant OB-R isoform found in hypothalamic neurons that express neuropeptide Y (NPY), agouti-related peptide, cocaine- and amphetamine-regulated transcript, and POMC (13, 33, 42). In adult sheep, intracerebroventricular infusion of leptin resulted in a decreased expression of mRNA for NPY in the hypothalamic arcuate nucleus (19). This is consistent with isolation of OB-Rb in 60% of NPY-containing cells in sheep hypothalamus (24). In the sheep hypothalamus, NPY can also regulate ACTH secretagogues corticotropin-releasing factor (CRF) and AVP (27). Furthermore, the absence of leptin is responsible for the obese phenotype of ob/ob mice and the HPA axis is activated in these animals as evidenced by elevated corticosterone levels (45). Chronic administration of leptin to ob/ob mice also decreases plasma corticosterone levels, suggesting that the adipose hormone is capable of regulating the HPA axis (3). In near-term fetal sheep, intracerebral infusion of leptin blunted the size of increase that occurred in amplitude and mean value of plasma ACTH and cortisol pulses near term (21). These data are consistent with studies on the adult mouse, in which leptin attenuates restraint-induced activation of the HPA (18). Although we found enhanced POMC processing/release, the CRF and POMC message were not enhanced under conditions of LTH (36). Overall, our data on HPA function cumulatively indicate that, in addition to ACTH1–39 and ACTH precursors, other factors govern cortisol production in LTH fetuses. Under basal conditions, despite increased hypothalamic drive, there is normal basal cortisol output in the LTH fetuses. However, in response to a secondary stressor, this biochemical/molecular “brake” on cortisol production can be overridden, allowing the LTH fetus to mount an enhanced cortisol response. We propose that leptin plays a key role in the adaptive changes that take place in the HPA axis. Admittedly, to date, our data are correlative. However, these key observations, coupled with data in the literature, lend strong support for our novel hypothesis. Although the effects of hypoxia-induced leptin on fetal growth/metabolism and placental development are also fascinating, we will maintain our focus on the effects on the HPA axis (36).
In addition to potential effects at the hypothalamic level, leptin may be involved in the changes in adrenal expression of key steroidogenic genes observed in the LTH fetal sheep (37). Studies from our group demonstrated that leptin is capable of exerting an inhibitory effect on ACTH-stimulated CYP17 expression (44) consistent with studies in cultured bovine adrenal cortical cells where leptin suppressed cortisol output in response to ACTH stimulation through a reduction in CYP17 and CYP11A1 expression (25). Leptin has also been shown to inhibit ACTH-stimulated cortisol production in isolated human adrenal cortical tissue (41). Recent studies by McMillen et al. (32) also demonstrated an inhibitory effect of leptin on the fetal ovine HPA, potentially at both the hypothalamic-pituitary level and at the adrenal gland. Intravenous leptin infusion suppressed the prepartum rise in both ACTH and cortisol. Infusion of leptin after day 144 did not affect ACTH, but there was still marked suppression of cortisol up to three days before birth of the fetus, indicative of direct effects of leptin on cortisol production in the ovine fetus. In the present study, mRNA for the OB-Rb in the adrenal was elevated in the LTH fetuses, whereas no differences were observed for the OB-Ra form in the LTH fetal adrenal. Because leptin has been shown to inhibit ACTH-induced steroidogenesis coupled with increased plasma leptin, an increase in OB-Rb would serve to limit the capacity of the adrenal cortex response to the elevated ACTH observed in the LTH fetus.
Data from the present study indicate that leptin may play a key role in the adaptive changes in the HPA following LTH. In the hypothalamus, levels of OB-Rb were similar between normoxic and LTH fetal sheep, whereas levels of OB-Ra declined in the hypothalami of the LTH fetus. This indicates a potential enhancement of LTH signaling at the hypothalamic level. Considering the inhibitory nature of leptin at the level of the adrenal gland, enhanced leptin signaling may provide a “brake” against overstimulation of the HPA axis by conditions of sustained hypoxia. Components of the leptin signaling pathway are expressed in the ovine fetal adrenal (Myers DA, unpublished observations) supporting the hypothesis that OB-Rb signaling could serve as an important regulator of cortisol biosynthesis in the ovine fetal adrenal.
This study was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants HD-33147 (to D. A. Myers) and HD-31226 (to C. A. Ducsay)
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society