Akt/mTOR Role in Human Foetoplacental Vascular Insulin Resistance in Diseases of Pregnancy

1. Introduction

Insulin modulates D-glucose homeostasis, and a reduced response or a lack of response to this hormone (hereafter referred as “insulin resistance”) is characteristic in several pathologies, including diabetes mellitus and obesity [1, 2]. Insulin resistance tightly relates with abnormal responses of the vascular endothelium, that is, endothelial dysfunction, to vasoactive molecules including insulin and the endogenous nucleoside adenosine [3, 4]. Human pregnancy courses with physiological maternal and foetal insulin resistance as an adaptive response to the increasing nutrient requirement by the pregnant women and the growing foetuses [5].

Insulin signalling involves preferential activation of the protein kinase B (PKB)/Akt (Akt) and mitogen-activated protein kinase (MAPK) signalling pathways [4, 6]. Vascular actions of insulin in the human placenta and umbilical cord vessels (hereafter referred as “foetoplacental vasculature”) are of relevance since this vascular bed lacks innervation, and the control of the blood flux results from local release of vasoactive molecules [4, 7]. The mechanisms behind vascular insulin effects include the synthesis of nitric oxide (NO) by the endothelial NO synthase (eNOS) isoform, ATP release, and adenosine-mediated increase of L-arginine transport and NO synthesis [4, 8, 9]. Pathologies of pregnancy, such as preeclampsia (PE) [10] and gestational diabetes mellitus (GDM) [4, 11], and abnormal maternal conditions, such as pregestational maternal obesity (PGMO) and maternal obesity in pregnancy [12], show with reduced insulin signalling in the foetoplacental vasculature. In this review, we propose that common signalling mechanisms result in insulin resistance of the human foetoplacental vasculature in these diseases.

2. Insulin Signalling

Insulin activates the splice variants A (IR-A) and B (IR-B) of insulin receptors (IRs) in the human foetoplacental vasculature [13]. IR-A and IR-B are expressed in this vascular bed with IR-A showing higher affinity for insulin than that with IR-B [4, 13, 14]. IR activation by β-subunit autophosphorylation recruits and phosphorylates two protein families, that is, the insulin receptor substrates (IRSs) and the Src homology 2 domain-containing transforming protein 1 (SHc) [15] (Figure 1). IRSs have at least six members (IRS-1 to IRS-6), where IRS-1 and IRS-2 are the most characterized [15]. SHc corresponds to at least three different proteins (SHcA, SHcB, and SHcC), with SHcA being expressed in mammals as the alternative splicing isoforms SHcA 46, SHcA 52, and SHcA 66 [16]. IRS-1 and IRS-2 are major activators of Akt via phosphatidylinositol 3 kinase (PI3K) compared with a minor effect on 44 and 42 kDa mitogen-activated protein kinases (p44/42^mapk); instead, SHcA preferentially activates p44/42^mapk via the growth factor receptor-bound protein 2 (Grb2) [17]. However, whether stimulation of IR-A or IR-B results in differential SHc or IRS activation and signalling is unknown. The physiological response of most tissues in the human body, including the foetoplacental vasculature, is that activation of p42/44^mapk and Akt signalling pathways results in increased eNOS expression and activity leading to vasodilation [4, 18]. However, under pathological conditions, the equilibrium between signalling associated with IR-A and IR-B activation by insulin is lost and a preferential activation of p42/44^mapk or Akt is reported. Several studies describe a variety of cell signalling mechanisms potentially involved in these alterations of insulin response; however, upstream- and downstream-associated signalling pathways are not addressed.

3. Insulin Resistance

Insulin resistance is seen in subjects where the metabolic handling of D-glucose is deficient [2]. PE [19, 20], GDM [21, 22], and obesity in pregnancy [23] show with insulin resistance in the mother, foetus, and newborn. However, whether insulin resistance results from or is the cause of these pathological conditions is still unclear.

Several studies show that IRS-1-mediated activation of PI3K leads to formation of phosphatidylinositol triphosphate, the substrate for the human 3-phosphoinositide-dependent protein kinase 1 (PDK1), which activates Akt [15] (Figure 2). However, in insulin resistance, IR-B preferential activation by insulin results in IRS1/2-mediated increase in the activity of the p85α regulatory subunit of PI3K (PI3K p85α), which inhibits Akt thus reducing NO generation. Other reports show that Akt activation mediates increased activity of the mammalian target of rapamycin (mTOR), a regulator of cell proliferation, adhesion, migration, invasion, metabolism, and survival [24]. Interestingly, mTOR signals through p70 S6 kinase 1 (S6K1) which reduces insulin signalling by inhibiting IRSs-activity-mediated activation of Akt [25, 26]. Thus, a modulatory loop to keep a physiological Akt activity and therefore insulin signalling to cause vasodilation involves mTOR activation/deactivation depending on the state of activation of Akt. When mTOR is upregulated, the physiological consequences are reduced Akt-mediated, NO-dependent vascular responses to insulin.

Other reports address that mTOR activity is inhibited by the adenosine monophosphate kinase (AMPK) [27], a molecule considered as general sensor of the cell energy state getting activated in response to a lower ATP/AMP ratio [28, 29]. AMPK activation results in increased eNOS activator phosphorylation at serine 1177 (Ser¹¹⁷⁷) and serine 615 (Ser⁶¹⁵) in the vasculature [30]. Interestingly, AMPK activation increased the activity of PI3K/Akt/eNOS signalling cascade leading to higher NO generation and prevented the high D-glucose-impaired response to insulin in human umbilical vein endothelial cells (HUVECs) [31]. Thus, it is suggested that AMPK will increase insulin signalling due to its capacity to inhibit mTOR in the human foetoplacental vasculature.

Activation of p44/42^mapk triggers c-Jun N-terminal kinase (JNK) signalling in HUVECs, resulting in IRS inhibition [32, 33] (Figure 2). Since S6K1 activation by mTOR results in p44/42^mapk- and Akt-reduced activity in HUVECs [34] and insulin-dependent activation of p44/42^mapk inhibits AMPK in the rat skeletal muscle cell line L6 [35], a functional dependency between p44/42^mapk, AMPK, and mTOR activity may also be a phenomenon involved in impaired insulin sensitivity in the foetoplacental vasculature.

It is well described that proinflammatory cytokine tumour necrosis factor α (TNFα) [36] and the adipocytokine adiponectin [37] and leptin [38] play crucial roles in insulin resistance. TNFα activates the JNK signalling pathway in HUVECs [39] resulting in inhibition of IRS-1 and reduced Akt-mediated insulin signalling [40–42] (Figure 2). Interestingly, higher plasma TNFα is found late in pregnancy (34–36 weeks of gestation) suggesting a likely reduced insulin biological action at this stage of pregnancy [43]. Adiponectin keeps insulin signalling (i.e., acts as insulin sensitizer) increasing the IRS-dependent signalling pathway by activating AMPK [37] and, subsequently, inhibiting mTOR [44]. Interestingly, a reduced plasma level of adiponectin is reported in pregnant women with diabetes mellitus [36]. Since the maternal plasma TNFα level is elevated in PE [45], GDM [46], or obese pregnant women [47], a potential TNFα-dependent inhibition of adiponectin release in insulin resistance in pregnant women, and perhaps the foetus, is likely. However, whether TNFα regulates adiponectin release in pregnancy is still unknown. Leptin is released in obesity in response to accumulating subcutaneous fat and increased fatty acid oxidation [38], a phenomenon regarded as a state of higher insulin resistance [38, 48]. Additionally, leptin activates JNK leading to inhibition of IRS1/2 and reduced insulin sensitivity [32, 33]. Since (i) leptin also increases the generation of reactive oxygen species (ROS) in HUVECs [49], (ii) superoxide anion (O₂⁻), the most reactive ROS, scavenges NO [30], and (iii) ROS activates JNK in this cell type [49], a leptin/ROS (probably O₂⁻)/JNK pathway is likely described as a mechanism leading to reduced insulin sensitivity in the human foetoplacental vasculature. Interestingly, increased leptin concentration in the maternal circulation is reported in GDM pregnancies [50, 51], a disease that also shows with increased ROS generation [9, 11]. Thus, this adipocytokine may also play a role in insulin resistance particularly in diseases of pregnancy where ROS generation is increased.

4. Insulin Resistance in Pregnancy Diseases

4.1. Preeclampsia

Preeclampsia (PE) is a heterogeneous pregnancy-specific multisystemic syndrome, defined by the occurrence of new onset hypertension (≥140/90 mmHg) and proteinuria (≥300 mg/24 hours) after 20 weeks of gestation [10, 52]. PE is of early onset (EOPE, <34 weeks of gestation) or late onset (LOPE, ≥34 weeks of gestation) [10, 53, 54]. EOPE and LOPE pregnancies associate with impaired insulin response of the maternal [55] and foetoplacental vasculature [20, 56]. However, not a clear mechanism explaining these alterations in EOPE and LOPE is yet available.

Preferential activation of p42/44^mapk and Akt is described in the foetoplacental vasculature from PE. Preterm PE (<37 wg) with HELLP (Hemolysis, Elevated Liver enzymes, and Low Platelet count) courses with increased phosphorylated p42/44^mapk activation in villous trophoblast [57]. In addition, the maternal plasma level from women with EOPE shows a higher level of endothelin-1 (ET-1) [58], but reduced Akt activity in the placenta [59] (Figure 3). Thus, an ET-1-dependent inhibition of Akt reducing insulin signalling is likely in this disease. Furthermore, since Akt activity positively correlates with NO generation in human foetal endothelial cells [60], EOPE-associated foetoplacental vascular dysfunction due to reduced NOS activity may involve p44/42^mapk/ET-1/Akt signalling. On the other hand, LOPE pregnancies show with unaltered p42/44^mapk [57] and unaltered [57] or decreased [61] Akt activity in the placenta. Intriguingly, eNOS protein abundance and activator phosphorylation (Ser¹¹⁷⁷) are higher in HUVECs from LOPE pregnancies [20], findings complemented by elevated nitrate/nitrite ratio in human umbilical vein serum [62, 63], but contrary to the reported lower nitrate/nitrite ratio [64] and NOS-generation of L-citrulline from L-arginine (index of NOS activity) [20] in this cell type. One plausible explanation for reduced NOS activity in HUVECs from LOPE pregnancies is a predominant functional effect of an increase of eNOS inhibitor (Thr⁴⁹⁵) compared with the effect of an activator (Ser¹¹⁷⁷) phosphorylation of this enzyme [20]. Earlier studies show increased IRS-1 (Ser³¹²) and IRS-2 (Ser⁷³¹) inhibitor phosphorylation in response to insulin in the placenta from LOPE pregnancies [65]. Since IRS1/2 are key activators of Akt, LOPE-reduced Akt and NOS activity could involve IRS1/2 inhibition. Thus, LOPE-associated impaired insulin response could result from reduced IRS1/2/Akt/eNOS signalling in the human foetoplacental vasculature. Since activation of mTOR results in reduced IRS1/2 activity, it is likely that this signalling molecule is involved in the effect of EOPE and LOPE on NOS activity. However, there is no information regarding the potential role of mTOR in the aetiology of EOPE or LOPE in this vascular bed.

Several reports support the involvement of circulating factors in the aetiology of PE including increased soluble Fms-like tyrosine kinase 1 (sFlt1), soluble endoglin (sEng), and reduced vascular endothelial growth factor (VEGF) plasma levels [66, 67]. The increased plasma levels of ET-1 and sEng result in a higher sFlt1 plasma level [68]. The latter reduces the availability of free VEGF-A to bind VEGF plasma membrane receptors and inhibition of PI3K/Akt signalling, including eNOS activity, in HUVECs [61, 69]. However, inhibition of the PI3K/Akt signalling does not alter sEng release from placenta explants or primary trophoblast in PE [59]; therefore, a differential response to PI3K/Akt-mediated insulin signalling in human foetoplacental endothelium versus trophoblast is likely. Interestingly, PI3K p85 phosphorylation at Tyr⁶⁸⁸ results in increased PI3K activity and Akt signalling in placental tissue from EOPE pregnancies [70]. The latter was proposed as a compensatory mechanism to the VEGF-reduced activation of PI3K/Akt signalling in this disease. However, PI3K p85 activator phosphorylation is unaltered in placentas from LOPE pregnancies [71], suggesting a different adaptive mechanism for insulin signalling in EOPE and LOPE pregnancies.

5. Concluding Comments

Insulin regulates canonical signal transduction pathways initiated by activation of IR-A/p44/42^mapk and IR-B/Akt in human foetoplacental vasculature in healthy pregnancies (Figure 3). IRS-1 and IRS-2 are upstream activators of the PI3K/Akt signalling pathway leading to activation of mTOR. SHcA 42 and SHcA 56 activate p44/42^mapk leading to increased release of vasoconstrictors, such as ET-1. Insulin resistance associated with PGMO, PE, and GDM results in foetoplacental vascular dysfunction and altered vascular reactivity to insulin. A likely potential common point in insulin resistance in these diseases is a reduced Akt signalling resulting in lower activation of mTOR and eNOS. A role for AMPK in this phenomenon is not clear, but the involvement of this molecule is likely since its activation positively correlates with mTOR activity. A role of NO in the response to insulin in the foetoplacental endothelium in diseases of pregnancy is well described [4, 10, 12]. Thus, modulation of NO generation could be a final target of an abnormal IR-A/SHcA/p44/42^mapk- and IR-B/IRSs/Akt-mediated signalling via Akt/mTOR in insulin resistance at the human foetoplacental vasculature. A therapy targeting these signalling molecules could be beneficial to improve insulin response in these diseases. PGMO is a risk factor for developing PE [106, 107] and GDM [107]. Thus, characterizing potential common signalling mechanisms for PGMO, PE, and GDM will facilitate the design of an approach to prevent insulin resistance in the co-occurrence of these or other disorders in pregnancy, thus reducing or abolishing their deleterious consequences for the mother, the foetus, and the newborn.

Conflicts of Interest

The authors confirm that there are no conflicts of interest.

Authors’ Contributions

Roberto Villalobos-Labra, Luis Silva, and Luis Sobrevia conceived and designed the study. Roberto Villalobos-Labra, Mario Subiabre, Luis Silva, Joaquín Araos, Tamara Sáez, Bárbara Fuenzalida, Marcelo González, Rocío Salsoso, and Andrea Leiva acquired the data/information. Roberto Villalobos-Labra, Mario Subiabre, Luis Silva, Fernando Toledo, Delia I. Chiarello, Joaquín Araos, Tamara Sáez, Bárbara Fuenzalida, Marcelo González, Fabían Pardo, Rocío Salsoso, Claudia Quezada, Andrea Leiva, and Luis Sobrevia analyzed the data/information. Roberto Villalobos-Labra, Mario Subiabre, Luis Silva, Rocío Salsoso, Joaquín Araos, Bárbara Fuenzalida, Fabían Pardo, Claudia Quezada, Andrea Leiva, and Luis Sobrevia interpreted the data/information. Roberto Villalobos-Labra, Mario Subiabre, Luis Silva, Rocío Salsoso, and Luis Sobrevia compiled the tables. Roberto Villalobos-Labra, Luis Silva, and Luis Sobrevia designed the figures. Roberto Villalobos-Labra, Luis Silva, and Luis Sobrevia wrote the manuscript.