Background
Although pregnancy complicated by diabetes may result in a normal delivery with few adverse maternal effects or the child’s long-term health, obesity and poor glycemic control during pregnancy can have deleterious maternal and embryo-fetal effects as well as delivery complications [
1]. Lifestyle choices, in particular caloric intake, contribute significantly to obesity and increase expression of pro-inflammatory markers that precede the onset of insulin resistance and hyperglycemia associated with T2DM [
2]. This apparent immune dysregulation in T2DM suggests that it may be a relevant therapeutic target [
2], especially under physiological conditions such as gestation where adverse pregnancy outcomes (APOs) may occur in face of improper glycemic control.
Maintaining adequate glycemic control is the focus of current therapeutic interventions in pregnancies complicated by gestational diabetes mellitus (GDM). However, lowering the risk of APOs has always been a challenge with the use of current therapeutic modalities such as insulin and/or metformin [
3,
4]. Although insulin treatment of GDM reduces serious perinatal morbidity and partially improves the woman’s health-related quality of life [
5], nonetheless, offspring of obese and diabetic women with adequate symptomatic control of their hyperglycemia have a five to tenfold increased risk of congenital anomalies and a fivefold greater risk of perinatal mortality than non-obese women based upon the population studied [
3,
4,
6].
When challenged with a high-fat diet, the New Zealand Obese (NONcNZO10/LtJ) mice become hyperglycemic, hyperinsulinemic and insulin resistant, making them an important tool for investigating the links between metabolic dysregulation and reproductive and developmental defects in the obese and diabetic subjects [
7,
8] and reviewed in [
9].
While the use of immunosuppressants is primarily indicated to prevent allograft rejection, mounting evidence suggests they may be used safely during pregnancy after solid organ transplant [
10,
11]. A recent recommendation suggests that in utero exposure to the immunosuppressant tacrolimus does not increase the risk of major congenital malformations, although there was an increased risk of low birth weight and pre-term birth [
12]. Based upon this safety profile and the clinical importance of tacrolimus, it was decided to use this agent to test the hypothesis that immunosuppression can improve pregnancy outcomes in the chronically overfed obese and diabetic murine model of the human obesity-associated T2DM.
Discussion
The present study demonstrates that tacrolimus, an immunosuppressant currently used in organ transplant procedures, significantly improved glycemic control with concomitant improvement of reproductive outcomes, and increased postnatal survival without increasing risks of major congenital malformations in a murine model of human obesity-associated T2DM. In insulin resistant diabetic HFD-dNONcNZO mice, metformin normalized blood glucose levels to the same extent as tacrolimus but did not offer the same degree of protection against reproductive health and developmental adversities. This suggests that the immune system, independent of glycemic control, is a critical component of the management of adverse pregnancy outcomes associated with chronic maternal overnutrition.
Carrying a pregnancy after solid organ transplantation is considered a high risk for maternal, fetal and neonatal complications but has been successful. Multiple studies have demonstrated that clinically tacrolimus poses no increased risk of congenital malformations but may increase the incidence of low birth weight and pre-term delivery [
10,
11,
28,
29]. These clinical experiences were in conflict, to some extent, with the preclinical safety studies conducted in female rats dosed prior to mating or during organogenesis [
30,
31]. At a dose one-third of the maternally toxic dose based upon exposure, tacrolimus had adverse effects on female reproductive parameters causing embryolethality and significant reductions of pup weights [
30]. These observations resulted in a US FDA pregnancy class C which states that “
Animal reproduction studies have shown an adverse effect on the fetus and there are no adequate and well-
controlled studies in humans, but potential benefits may warrant use of the drug in pregnant women despite potential risks”. However, in over 200 exposed pregnancies and 236 babies born to organ transplant recipients where tacrolimus was administered as a multi-drug based therapy, no significant teratogenicity could be concluded from multiple clinical registries [
32‐
35]. Nonetheless, unlike the present study, preclinical safety reports in rats were in contradiction to each other. Saegusa and associates [
30] reported on reduced reproductive performance, increased embryolethality and decreased pup weights among pregnant rats orally receiving 1.0 and 3.2 mg/kg FK506 (tacrolimus) after organogenesis and during lactation [
30]. On the other hand, Farley and co-workers [
31] recounted normal maternal health but dose-dependant elevation in the percentages of post-implantation resorption that led to the conclusion of a relative safety to the use of tacrolimus in pregnancy [
31]. Similarly, Ramos et al. [
36] reported on normal maternal and fetal health in otherwise normoglycemic female Wistar rats orally receiving tacrolimus at 10–40 times the currently described dosage. Notwithstanding species and route of administration related-differences in interpreting tacrolimus pharmacokinetics among mice, rats and women, our data are generally supportive of previous reports from clinical registry leveraging records from 20 to 200 women receiving 0.1–0.4 mg/kg/day tacrolimus mono-therapy or in a combination of immunosuppression whereby higher fetal exposure ratios were found to be exponentially related to higher maternal trough blood concentrations of the compound but reported on normal birth weights and post-natal development in these pregnancies [
28,
35,
37‐
39]. This is an important consideration because tacrolimus, like most drugs, crosses the placenta with concentrations found in the cord blood being ~71, 23 and 19% of maternal concentrations for whole blood, plasma and unbound form, respectively [
28,
39]. Even with transfer across the placenta, tacrolimus neither causes gross congenital malformations in humans [
10,
32,
40,
41] nor does it in normoglycemic Wistar rats exposed for 15 days to about 10–40× the current dose
per oral during peri-implantation phase [
36]. Nevertheless, tacrolimus has been reported to cause transient and reversible materno-fetal comorbidities such as hyperkalemia and in some cases reduced renal mass (oligonephronia) thereby increasing the risk for further development of renal failure and hypertension in adult life [
40,
42]. Additionally, it should be pointed out that although the present study used tacrolimus to normalize glycemic control in the obese and T2DM mice, a paradoxical new-onset diabetes after transplantation (NODAT), a severe complication following organ transplantation, has been reported to occur in between 2 and 53% of transplanted patients receiving higher dosing of tacrolimus and/or having pretransplantation hyperglycemia (reviewed in [
43‐
45]. Nonetheless, NODAT has not been reported in recent studies on allogeneic uterine transplantation in rats receiving low dosing of tacrolimus [
46], neither was it associated with worsening of clinical outcomes during a mean follow-up of 3 years in kidney transplant recipients [
47]; nevertheless it is warranted that further pre-clinical safety studies on the potential use of tacrolimus as anti-diabetic agent should be conducted.
Among endometrial changes and mechanisms linked to the adverse pregnancy outcomes in diabetic mice and women are impaired spiral artery remodeling and trophoblast invasion defects during gestation [
48,
49]. In an attempt to elucidate how tacrolimus restored vascular adaptation to pregnancy in the obese and diabetic mice, we examined uterine arterial physiology, spiral artery remodeling and late gestational (GD 18.5) placental cytokines during pregnancy in the HFD-dNONcNZO dams. Incomplete spiral artery remodeling with lumen stenosis has been recognized among ominous pathophysiological signs of poor placentation in GDM [
50,
51], This heralding histopathological sign was evident in the placentas of HFD-dNONcNZO dams expressing higher levels of pro-inflammatory cytokines that include TNFα, IL16 and IL23 (Additional file
1: Table S4). This is consistent with the reported human and non-human primate data that the placenta develops exaggerated pro-inflammatory response to obesity, which contributes to or results from placental vascular insufficiency [
52,
53]. Therefore, important implications came from findings of restricted uterine artery pulsatility and poor umbilical flow dynamics during gestation in the HFD-dNONcNZO dams. Vascular indices (RI and/or PI) are used to investigate impedance of the vascular bed distal to the vessel being examined, and a large quantity of continuous forward flow is generally observed throughout the diastole in low-resistance arterial waveforms [
54]. Contrary to this, higher vascular indices are characteristics of vascular flow supplying high-resistance and leaky vascular beds [
54]. Through mechanisms linked to chronic systemic hyperglycemia, local inflammation and the release of pro-inflammatory and pro-angiogenic molecules such as vascular endothelial growth factor (VEGF), advanced glycation end-products (AGE) and alterations to de novo synthesis of nitric oxide (NO), maternal hyperglycemia is known to induce vascular defects characterized by increased angiogenesis, increased vascular resistance, hyper-coagulability and preponderance of highly permeable vessels [
55‐
59]. On the other hand, it has been shown that umbilical artery flow velocity negatively correlates with umbilical artery resistance and is reflective of alterations to the uterine artery PI and placental vascular resistance [
60]. Therefore, the overall lower uterine artery PI throughout gestation and higher RI in the untreated HFD-dNONcNZO dams dictates their poor breeding performance and the higher rates of inflammation-mediated fetal demise observed in this mouse model. Also, considering that the placenta is an organ requiring constant perfusion and has low vascular indices with uterine artery PI values decreasing with gestational age [
61], the present findings that therapeutic interventions with either tacrolimus or metformin normalized uterine artery PI in the treated HFD-dNONcNZO dams have clinical inferences suggestive of a pivotal role for adequate maternal glycemic control and immuno-modulation relative to pregnancy outcomes in conditions of chronic maternal overnutrition.
Taken together, the present study support the observation that adverse fetal health among neonates born to diabetic mothers cannot be explained solely on the basis of maternal hyperglycemic control [
62]. Our data are also in parallel with the notions made by Leach et al. and Ericsson et al. that hyperglycemia should only be seen as a player within a spectrum of diabetic complications whereby maternal inflammation might have the upper hand [
62,
63]. Nonetheless, the present data are in conflict with the primate studies reported by Frias and associates [
52] where in the relatively wider placenta exchange area and high functional reserve are to be blamed for the lower rates of fetal growth restriction reported in the primate studies. However, the high rate of fetal demise at necropsy reported in the present study and by Frias and associates points to the uteroplacental flow restriction and inflammation to be the plausible culprits. Indeed, similar to the non-human primate model of obesity-induced placental growth restriction [
52], our findings suggest that inflammation in the pregnant adult HFD-dNONcNZO female mice is likely localized in tissues involving at least the placenta. This supports previous studies in that the end-organ response to HFD is local yet associated with dysregulation of metabolic, vascular and inflammatory pathways requiring an immunosuppression [
64].
A limitation of the present study is the chance of investigating renal functions among the tacrolimus-treated HFD-dNONcNZO dams. It has been reported in ordinary kibble-fed male Wistar rats that tacrolimus at a dose of 0.6 mg/kg/day for 9 consecutive weeks is nephrotoxic [
65]. Important to tacrolimus pharmacokinetics and mechanism of action is its bioavailability and the ratio of bound:free form in the blood. The latter two maternal variables are greatly influenced by a multitude of factors among which the renal functions have long been recognized [
66]. Therefore, future studies are warranted to establish whether or not our treated mice suffered tacrolimus-related nephrotoxicity. A possibility to overcome this limitation is the titration of dosage to maintain whole blood tacrolimus trough concentrations in the usual therapeutic range in plasma or urine of treated dams. This is particularly important since tacrolimus pharmacokinetic changes during pregnancy with maternal factors such as anemia and hypoalbuminemia significantly contribute to at risk pregnancy.
Authors’ contributions
AJHA, TRSO, MW and FWKK designed the research study, wrote, critically reviewed and approved the final version of the manuscript. AJHA performed all experiments. AJHA and TRSO collected and analyzed the data; SP and TRSO performed postnatal fetal analysis; MH analyzed ultrasound data. All authors read and approved the final manuscript.