Insulin requires A₁ adenosine receptors expression to reverse gestational diabetes-increased L-arginine transport in human umbilical vein endothelium

Abstract

Gestational diabetes mellitus (GDM) associates with increased L-arginine transport and extracellular concentration of adenosine in human umbilical vein endothelial cells (HUVECs). In this study we aim to determine whether insulin reverses GDM-increased L-arginine transport requiring adenosine receptors expression in HUVECs. Primary cultured HUVECs from full-term normal (n = 38) and diet-treated GDM (n = 38) pregnancies were used. Insulin effect was assayed on human cationic amino acid transporter 1 (hCAT1) expression (protein, mRNA, SLC7A1 promoter activity) and activity (initial rates of L-arginine transport) in the absence or presence of adenosine receptors agonists or antagonists. A₁ adenosine receptors (A₁AR) and A_2AAR expression (Western blot, quantitative PCR) was determined. Experiments were done in cells expressing or siRNA-suppressed expression of A₁AR or A_2AAR. HUVECs from GDM exhibit higher maximal transport capacity (maximal velocity (V_max)/apparent Michaelis Menten constant (K_m), V_max/K_m), which is blocked by insulin by reducing the V_max to values in cells from normal pregnancies. Insulin also reversed the GDM-associated increase in hCAT-1 protein abundance and mRNA expression, and SLC7A1 promoter activity for the fragment −606 bp from the transcription start point. Insulin effects required A₁AR, but not A_2AAR expression and activity in this cell type. In the absence of insulin, GDM-increased hCAT-1 expression and activity required A_2AAR expression and activity. HUVECs from GDM pregnancies exhibit a differential requirement of A₁AR or A_2AAR depending on the level of insulin, a phenomenon that represent a condition where adenosine or analogues of this nucleoside could be acting as helpers of insulin biological effects in GDM.

Introduction

Gestational diabetes mellitus (GDM) characterizes by abnormal maternal D-glucose metabolism and altered insulin signaling in the fetoplacental circulation [1–3]. GDM also associates with increased uptake of the cationic amino acid L-arginine [4], the substrate for nitric oxide (NO) synthesis by the endothelial NO synthase (eNOS) [5], in human umbilical vein endothelial cells (HUVECs), changes referred as GDM-associated fetoplacental endothelial dysfunction [3, 6]. The latter is reinforced with findings showing that uptake of the endogenous nucleoside adenosine, a potent vasodilator in the human fetoplacental vasculature [7, 8] and other vascular beds [9–11], is also reduced in HUVECs from GDM pregnancies [8, 12].

Reduced adenosine transport leads to increased extracellular levels of adenosine in HUVECs primary cultures from GDM pregnancies, a finding that agrees with the elevated adenosine plasma level detected in human umbilical vein blood from GDM pregnancies [8, 13]. We reported that activation of A_2A adenosine receptors (A_2AAR) is required for the increase caused by insulin on L-arginine transport via the human cationic amino acid transporter 1 (hCAT-1), a Na⁺- and pH-independent membrane transporter (apparent K_m ~100 μM) [14, 15], and eNOS activity in HUVECs from normal pregnancies [6, 12]. Additionally, insulin restores GDM-reduced adenosine transport [15] in this cell type. However, whether insulin effect requires activation of adenosine receptors (ARs) in the fetoplacental endothelium from GDM pregnancies is unknown [3, 11]. We hypothesize that insulin will reverse the GDM-associated increase in hCAT-1 expression and activity requiring expression and activation of ARs in HUVECs. Our results show that increased L-arginine transport in HUVECs from GDM pregnancies is reversed by insulin to values in cells from normal pregnancies via a mechanism where A₁AR play a role.

Materials and methods

Human umbilical cords and study groups

Umbilical cords were collected after delivery from 38 full-term normal or 38 full-term GDM pregnancies. Ethnicity of patients included in this study was Hispanic. The investigation conforms to the principles outlined in the Declaration of Helsinki. Ethics Committee approval from the Faculty of Medicine of the Pontificia Universidad Católica de Chile and informed consent of patients (all of them from the Hospital Clínico UC in Santiago de Chile) were obtained. Patients between the 24–28 weeks of gestation with basal glycemia >5 mM (>90 mg/dL, i.e., overnight starvation) and with >7.9 mM (>140 mg/dL) at 2 h after an oral glucose load (75 g) were diagnosed as gestational diabetes (according with the Perinatal Guide 2014 report from the Health Ministry of Chile) and subjected to dietary treatment with 1500 kcal/day and a maximum of 200 g per day carbohydrates (Table 1). The homeostasis model assessment for insulin resistance (HOMA-IR) was calculated from:

$$\mathrm{I}\mathrm{R}=\frac{\mathrm{Insulin}}{22.5\cdot e^{-ln\left(\mathrm{Glucose}\right)}}$$

where Insulin is in micro unit per millilitre and Glucose is basal glycemia in millimole per litre [16, 17]. Insulin sensitivity (IS) was derived from these values by $\mathrm{IS}=\frac{1}{\mathrm{I}\mathrm{R}}\cdot 100$ (expressed in %). Additionally, β cell function (β, expressed in %) was estimated from:

$$\beta =20\cdot \frac{\mathrm{Insulin}}{\mathrm{Glucose}-3.5}$$

Table 1.

Clinical characteristics of women with normal or GDM pregnancies and newborns

Variables	Normal pregnancies (n = 38)	GDM pregnancies (n = 38)
Maternal variables
Age (years)	28 ± 6 (18–38)	31 ± 6 (20–38)
Height (cm)	159 ± 7 (157–162)	161 ± 9 (157–163)
Weight (kg)
24–28 wg	52 ± 1.3 (50–55)	53 ± 1.2 (50–55)
38–40 wg	61 ± 1.1† (55–65)	63 ± 1.1† (59–66)
BMI (kg/m2)
24–28 wg	20.6 ± 0.2 (20.3–21.0)	20.5 ± 0.47 (20.3–20.7)
38–40 wg	24.1 ± 0.1 (22.3–24.8)	24.3 ± 0.44 (23.9–24.8)
Systolic blood pressure (mmHg)
24–28 wg	101 ± 4 (98–107)	103 ± 5 (97–106)
38–40 wg	106 ± 6 (95–112)	110 ± 3 (101–112)
Glycosylated hemoglobin A1c
24–28 wg (% of total) [mmol/mol]	4.1 ± 0.3 (3.6–4.7) [21.3 ± 1.6 (18.7–24.4)]	4.1 ± 0.2 (3.8–4.6) [21.3 ± 1.0 (19.7–23.9)]
38–40 wg (% of total) [mmol/mol]	4.0 ± 0.3 (3.3–5.0) [20.2 ± 1.5 (16.7–25.3)]	6.1 ± 0.1*† (5.6–6.9) [41.0 ± 0.7 (37.7–46.4)]
Glycemia basal at delivery (mmol/L)	4.4 ± 0.4 (4.1–5.1)	4.6 ± 0.5 (3.9–5.6)
OGTT (mmol/L)
Glycemia basal	4.5 ± 0.4 (3.9–5.2)	4.7 ± 0.4 (4.1–5.2)
Glycemia 2 h after glucose	5.1 ± 0.5 (3.7–6.1)	9.8 ± 1.2* (8.1–11.9)
Plasma insulin (μU/mL)	5.1 ± 0.1 (4.2–5.4)	7.9 ± 1.5* (6.2–12.1)
HOMA-IR	1.00 ± 0.05 (0.77–1.22)	1.62 ± 0.16* (1.37–1.97)
HOMA-IS (%)	101 ± 5 (82–130)	62 ± 6* (51–73)
HOMA-β (%)	114 ± 7 (64–140)	143 ± 14* (115–310)
Newborn variables
Sex (female/male)	12/16	13/15
Gestational age (weeks)	38.1 ± 0.3 (38–40)	38.2 ± 0.2 (38–39)
Birth weight (g)	3087 ± 67 (2417–3451)	4299 ± 63* (3571–4731)
Height (cm)	50 ± 2.2 (47–54)	51 ± 3.3 (48–54)
Ponderal index (g/cm3 × 100)	2.46 ± 0.02 (1.7–7.4)	3.44 ± 0.02* (2.9–6.0)
Umbilical vein D-glucose (mmol/L)	3.8 ± 0.5 (3.1–4.4)	4.4 ± 0.5 (4.2–4.8)
Umbilical vein insulin (μU/mL)	6.1 ± 0.7 (5.5–7.0)	11.6 ± 0.6* (8.9–13.1)
HOMA-IR	1.03 ± 0.09 (0.76–1.36)	2.27 ± 0.12* (1.66–2.80)
HOMA-IS (%)	97 ± 12 (74–132)	44 ± 4* (34–60)

Data are presented as mean ± SD (range). All women included in this study were Hispanic. GDM patients were treated with diet. OGTT was measured between 24 and 28 wg. HOMA-IR, HOMA-IS, and HOMA-β (β cell function) were estimated as described in “Materials and methods”

OGTT oral glucose tolerance test, BMI body mass index, HOMA-IR homeostasis model assessment for insulin resistance, HOMA-IS homeostasis model assessment for insulin sensitivity, wg weeks of gestation

*P < 0.05 versus values in Normal pregnancies; †P < 0.05 versus values at 24–28 weeks of gestation in Normal or GDM pregnancies

Cell culture

Confluent HUVECs primary cultures (37 °C, 5 % CO₂) were cultured in medium 199 (M199; Gibco Life Technologies, Carlsbad, CA, USA) up to passage 3 as described [6]. The culture medium was supplemented with 0.04 nM (~5.1 μU/mL) or 0.07 nM (~7.9 μU/mL) insulin for normal or GDM pregnancies, respectively [8]. Experiments were performed in the absence (referred as ‘without insulin’ or ‘in the absence of insulin’) or presence (referred as ‘with insulin’ or ‘in the presence of insulin’) of exogenous 1 nM insulin for 8 h. Cells were also exposed to 30 nM (2R,3R,4S,5R)-2-[6-(cyclopentylamino)purin-9-yl]-5-(hydroxymethyl)oxolane-3,4-diol (CPA, A₁AR agonist) (Sigma, Atlanta, GA, USA), 30 nM 8-cyclopentyl-1,3-dipropylxanthine (DPCPX, A₁AR antagonist) (Sigma), 30 nM 2-[p-(2-carbonyl-ethyl)-phenylethylamino]-5′-N-ethylcarboxamidoadenosine (CGS-21680, A_2AAR agonist) (Sigma), 10 nM 4-(2-[7-amino-2-[2-furyl]-[1, 2, 4]triazolo[2,3-a]{1,3,5}triazin-5-yl-amino]ethyl)phenol (ZM-241385, A_2AAR antagonist) (Sigma), 100 nM 2-[[6-Amino-3,5-dicyano-4-[4-(cyclopropylmethoxy)phenyl]-2-pyridinyl]thio]-acetamide (BAY60-6583, A_2BAR agonist) (Tocris Bioscience, Bristol, UK), 30 nM N-(4-cyanophenyl)-2-[4-(2,3,6,7-tetrahydro-2,6-dioxo-1,3-dipropyl-1H-purin-8-yl)phenoxy]-acetamide (MRS-1754, A_2BAR antagonist) (Tocris Bioscience), 30 nM 1-[2-chloro-6-[[(3-iodophenyl)methyl]amino]-9H-purin-9-yl]-1-deoxy-N-methyl-ß-D-ribofuranuronamide (2-Cl-IB-MECA, A₃AR agonist) (Tocris Bioscience), and 30 nM 3-propyl-6-ethyl-5[(ethylthio)carbonyl]-2-phenyl-4-propyl-3-pyridinecarboxylate (MRS-1523, A₃AR antagonist) (Tocris Bioscience) [6].

L-Arginine transport

Overall 0–1000 μM L-arginine transport (3 μCi/mL L-[³H]arginine (NEN, Dreieich, FRG), 1 min incubation, 37 °C) was measured as described [6]. Briefly, transport assays were performed in Krebs ((mM): NaCl 131, KCl 5.6, NaHCO₃ 25, NaH₂PO₄ 1, Hepes 20, CaCl₂ 2.5, MgCl₂ 1 (pH 7.4, 37 °C)) in cells preincubated (12–14 h) in phosphate-buffered saline (PBS) solution ((mM): 130 NaCl, 2.7 KCl, 0.8 Na₂HPO₄, 1.4 KH₂PO₄ (pH 7.4, 4 °C)) containing 0.25 % newborn and 0.25 % fetal calf sera. PBS and Krebs solutions were also supplemented with 2 U/mL adenosine deaminase 1. ATP, ADP, AMP, or adenosine was not detected in PBS or Krebs solutions as assayed by high-performance light chromatography (not shown) as described [12]. Cell monolayers were rinsed with ice-cold Krebs to terminate tracer uptake. Radioactivity in formic acid cell digests was determined by liquid scintillation counting, and uptake was corrected for D-[³H]mannitol (NEN) disintegrations per minute (d.p.m.) in the extracellular space [6].

Initial rate for transport (i.e., linear uptake up to 1 min) was derived from slope of linear phase of 100 μM L-arginine transport. Values for L-arginine transport were adjusted to the one phase exponential association equation considering the least squares fit:

$$v_{\mathrm{i}}=V_{\mathrm{m}}\cdot \left(1-e^{-\left(k\cdot t\right)}\right)$$

where v_i is initial velocity, V_m is mayor velocity at a given time (t) and L-arginine concentration, and e and k are constants. Overall L-arginine transport at initial rates was adjusted to the Michaelis-Menten hyperbola plus a nonsaturable, linear component as described [6]. Saturable L-arginine transport kinetic parameters maximal velocity (V_max) and apparent Michaelis-Menten constant (K_m) of transport were calculated as described [6]. The relative contribution of GDM, insulin, or ARs to the saturable L-arginine transport kinetic parameters was estimated from the maximal transport capacity (V_max/K_m) values for L-arginine transport by:

$$\frac{1}{{}^{C/X}F}=\frac{{}^{C}K_m\cdot {}^{X}V_{max}}{{}^{C}V_{max}\cdot {}^{X}K_m}$$

where ^CV_max and ^CK_m are the kinetics parameters for L-arginine transport in control conditions in cells from normal or GDM pregnancies, and ^XV_max and ^XK_m are kinetics parameters of L-arginine transport in HUVECs from normal or GDM pregnancies exposed to different experimental conditions [6, 8].

Trans-stimulation experiments for saturable L-arginine transport (100, 250, 500 μM) were performed in cells incubated (2 h) with Krebs solution containing 10 mM L-lysine. Incubation medium was removed and replaced by L-lysine-free Krebs and transport was determined as above [18].

Reverse transcription and quantitative RT-PCR

Experiments were performed using a Step One real time PCR system (Applied Biosystem, CA, USA) in a reaction mix containing 0.2 μM primers and master mix provided in the brilliant SYBR green qPCR Master Mix (Applied Biosystem, CA, USA) as described [6]. Hot Start Taq DNA polymerase was activated (15 min, 95 °C), and assays included a 95 °C denaturation (15 s), annealing (20 s) at 54 °C (hCAT-1 and 28S), and extension at 72 °C (hCAT-1 and 28S, 10 s). Fluorescent product was detected after 3-s step to 5 °C below the product melting temperature (T_m). Product specificity was confirmed by agarose gel electrophoresis (2 % w/v) and melting curve analysis. The product T_m values were 79.1 °C for hCAT-1 and 86.7 °C for 28S. hCAT-1 and 28S standards were prepared as described [6]. Oligonucleotide primers are as follows: hCAT-1 (sense) 5′-GAGTTAGATCCAGCAGACCA-3′, hCAT-1 (anti-sense) 5′-TGTTCACAATTAGCCCAGAG-3′, 28S (sense) 5′-TTGAAAATCCGGGGGAGAG-3′, 28S (anti-sense) 5′-ACATTGTTCCAACATGCCAG-3′. Expected size products for hCAT-1 (151 bp) and 28S (105 bp) were confirmed in PCR experiments. The 28S rRNA number of copies was unaltered (P > 0.05, n = 16) in all experimental conditions (not shown).

Western blotting

Proteins (70 μg) separated by polyacrylamide gel (10 %) electrophoresis were probed with a primary polyclonal goat anti-hCAT-1 (1:500), goat anti-A₁AR (1:1000), mouse anti-A_2AAR (1:1000) (Santa Cruz Biotechnology, Santa Cruz, CA, USA), or rabbit anti-A_2BAR (1:1000) or anti-A₃AR (1:1000) (Abcam, Cambridge, UK), or monoclonal mouse anti-β-actin (1:3000) (Sigma Aldrich, St Louis, MO, USA) antibodies [6]. Proteins were detected by enhanced chemiluminescence in a ChemiDoc-It® 510 Imagen System (UVP, LCC Upland, CA, USA) and quantified by densitometry.

HCAT-1 promoter cloning

Genomic DNA was isolated using the Wizard SV Genomic DNA Purification System (Promega, Madison, WI, USA). The sequences −1606 and −650 bp from the transcription start point of the SLC7A1 gene (GenBank: AL596114) were PCR-amplified using Elongase Enzyme System (Invitrogen) and cloned into pGL3-basic reporter system [6, 18]. The pGL3–hCAT-1 reporter constructs generated were pGL3–hCAT-1⁻¹⁶⁰⁶ and pGL3–hCAT-1⁻⁶⁵⁰.

Transient transfection

Sub-confluent (60–80 %) HUVECs primary cultures were resuspended in serum-free M199. Aliquots of cell suspension (0.5 mL, 3.2 × 10⁶ cells/mL) were mixed with 10 μg of pGL3–hCAT1⁻¹⁶⁰⁶ or pGL3–hCAT1⁻⁶⁵⁰ constructs, pGL3-Basic (empty pGL3 vector), pGL3-Control (Simian Virus 40 promoter (SV40) pGL3 vector), and the internal transfection control vector pRL-TK expressing Renilla luciferase (Promega) [6]. After electroporation (300 V, 700 μF, 5–10 ms) (Gene Pulser II System, BioRad, CA, USA), cells were cultured (48 h) in M199 containing 2 % FCS. Transfection efficiency was estimated by transfection of the pEGFP-N3 vector (Clontech, Mountain View, CA, USA) and fluorescent cells were counted under an inverted fluorescent microscope (Leica DMIL; Wetzlar, Germany).

Luciferase assay

Firefly and Renilla luciferase activity was measured using Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA) in a Sirius luminometer (Berthold Detection System; Oak Ridge, TN, USA) [6].

Adenosine receptors suppression

Since the pharmacological approach used in this study suggested that A_2BAR and A₃AR may not be involved in the effect of insulin, we focused in generating cells knockdown for A₁AR and A_2AAR. Suppression of A₁AR and A_2AAR expression was done using the commercially available short interference RNAs (siRNA) Adenosine A₁-R siRNA(h) and Adenosine A_2A-R siRNA(h) (Santa Cruz Biotechnology) following manufacturer’s instructions (http://datasheets.scbt.com/siRNA_protocol.pdf). Cells knockdown for A₁AR (^KDA₁AR) and A_2AAR (^KDA_2AAR) were generated.

Statistical analysis

Values are mean ± SD (range) or SEM, with n = 38 different cell cultures (3–4 replicates) from normal or GDM pregnancies. The normality of the data was determined with Kolmogorov-Smirnov test. Comparisons between two and more groups were performed by means of Student’s unpaired t test and analysis of variance (ANOVA), respectively. If the ANOVA demonstrated a significant interaction between variables, post hoc analyses were performed by the multiple-comparison Bonferroni correction test. P < 0.05 was considered statistically significant. The statistical software GraphPad Instat 3.0b and Graphpad Prism 7.0a.65 (GraphPad Software Inc., San Diego, CA, USA) were used for data analysis. P < 0.05 was considered statistically significant.

Results

Patients and newborns

Normal or GDM pregnancies were singleton and pregnant women were normotensive, nonsmoking, non-alcohol or drug consuming, and without intrauterine infection or any other medical or obstetrical complications (Table 1). Newborns to GDM pregnancies were heavier at birth, with higher umbilical vein insulin and HOMA-IR, but lower HOMA-IS compared with normal pregnancies.

L-Arginine transport

In the absence of insulin, overall transport of L-arginine was unaltered by ARs agonists or antagonists in cells from normal pregnancies (Fig. 1a). Insulin increased overall transport in this cell type, an effect blocked by an A_2AAR antagonist. In the absence of insulin, overall L-arginine transport was higher in cells from GDM compared with normal pregnancies. GDM effect was reversed by an A_2AAR antagonist to values in cells from normal pregnancies. Insulin reversed GDM-increase in transport to values in normal pregnancies, an effect blocked by an A₁AR antagonist.

Fig. 1.

Involvement of adenosine receptors on GDM and insulin effect on L-arginine transport. a Overall (100 μM) L-arginine transport in HUVECs from normal (Normal) or gestational diabetes mellitus (GDM) pregnancies incubated without or with insulin (1 nM, 8 h) in the absence (−) or presence (+) of A₁AR, A_2AAR, A_2BAR, or A₃AR agonists (Ag) or antagonists (Antag). b Saturable L-arginine transport as in a in cells incubated without (Control) or with A₁AR or A_2AAR antagonists. c L-Arginine transport (1 min) in cells preloaded with a solution without L-lysine (Control) or with 1 mM L-lysine (L-Lysine) (see “Materials and methods”). In a, *P < 0.05 vs. corresponding values in Without insulin. †P < 0.05 vs. corresponding values in the absence of Ago and Antag. In c, *P < 0.05 vs. all other values. Values are mean ± SEM (n = 38)

Overall L-arginine transport kinetics was semisaturable (not shown), and non-saturable transport component (i.e., K_D) was similar in normal or GDM pregnancies (Table 2) and unaltered by insulin or agonists or antagonists for ARs. However, insulin increased the v_i of transport in normal pregnancies, an effect blocked by the A_2AAR antagonist. In the absence of insulin, GDM-increase in v_i was blocked by the A_2AAR antagonist. Insulin reversed GDM effect on v_i, an effect blocked by the A₁AR antagonist (Table 2).

Table 2.

Kinetic parameters for L-arginine transport in HUVECs from normal or GDM pregnancies

	Saturable transport			Overall transport
	V max (pmol/μg protein/ min)	K m (μM)	V max/K m (pmol/μg protein/min/(μM))	K D (pmol/μg protein/min/(μM))	v i (pmol/μg protein/ 0.5 s)
Normal pregnancies
Without insulin
Control	1.2 ± 0.2	89 ± 20	0.0135 ± 0.0027	0.0011 ± 0.0004	0.006 ± 0.0001
A1AR ag	1.2 ± 0.1	103 ± 21	0.0117 ± 0.0016	0.0012 ± 0.0005	0.006 ± 0.0001
A1AR antag	1.0 ± 0.2	77 ± 40	0.0130 ± 0.0047	0.0011 ± 0.0004	0.005 ± 0.0002
A1AR ag/antag	1.1 ± 0.3	79 ± 26	0.0139 ± 0.0042	0.0012 ± 0.0003	0.006 ± 0.0002
A2AAR ag	1.1 ± 0.2	108 ± 15	0.0102 ± 0.0016	0.0011 ± 0.0003	0.005 ± 0.0001
A2AAR antag	1.0 ± 0.1	99 ± 33	0.0101 ± 0.0022	0.0010 ± 0.0003	0.007 ± 0.0001
A2AAR ag/antag	1.3 ± 0.4	87 ± 24	0.0149 ± 0.0044	0.0013 ± 0.0007	0.008 ± 0.0003
A2BAR ag	1.5 ± 0.3	98 ± 21	0.0153 ± 0.0031	0.0012 ± 0.0004	0.008 ± 0.0002
A2BAR antag	1.2 ± 0.1	91 ± 16	0.0132 ± 0.0017	0.0011 ± 0.0004	0.006 ± 0.0001
A2BAR ag/antag	1.1 ± 0.3	87 ± 21	0.0126 ± 0.0032	0.0012 ± 0.0006	0.006 ± 0.0002
A3AR ag	0.9 ± 0.2	62 ± 23	0.0145 ± 0.0043	0.0012 ± 0.0005	0.006 ± 0.0002
A3AR antag	1.3 ± 0.1	97 ± 21	0.0134 ± 0.0020	0.0013 ± 0.0004	0.007 ± 0.0001
A3AR ag/antag	1.2 ± 0.3	96 ± 31	0.0125 ± 0.0036	0.0012 ± 0.0005	0.006 ± 0.0002
With insulin
Control	2.3 ± 0.1*	93 ± 22	0.0247 ± 0.0035*	0.0012 ± 0.0004	0.012 ± 0.0002*
A1AR ag	2.7 ± 0.1*	104 ± 21	0.0260 ± 0.0031*	0.0012 ± 0.0004	0.013 ± 0.0003*
A1AR antag	2.2 ± 0.1*	88 ± 20	0.0250 ± 0.0035*	0.0011 ± 0.0005	0.011 ± 0.0003*
A1AR ag/antag	2.2 ± 0.2*	81 ± 16	0.0272 ± 0.0039*	0.0012 ± 0.0006	0.012 ± 0.0003*
A2AAR ag	2.9 ± 0.2*	137 ± 20	0.0212 ± 0.0023*	0.0010 ± 0.0004	0.011 ± 0.0002*
A2AAR antag	0.9 ± 0.4†	95 ± 29	0.0095 ± 0.0036†	0.0011 ± 0.0004	0.004 ± 0.0002†
A2AAR ag/antag	1.1 ± 0.2†	82 ± 27	0.0134 ± 0.0034†	0.0013 ± 0.0005	0.007 ± 0.0002†
A2BAR ag	1.7 ± 0.1*	95 ± 21	0.0179 ± 0.0025*	0.0010 ± 0.0005	0.011 ± 0.0002*
A2BAR antag	2.3 ± 0.2*	110 ± 29	0.0209 ± 0.0031*	0.0013 ± 0.0003	0.012 ± 0.0002*
A2BAR ag/antag	1.9 ± 0.3*	89 ± 21	0.0213 ± 0.0043*	0.0013 ± 0.0005	0.011 ± 0.0003*
A3AR ag	2.5 ± 0.1*	84 ± 12	0.0298 ± 0.0027*	0.0012 ± 0.0005	0.014 ± 0.0003*
A3AR antag	2.2 ± 0.1*	88 ± 18	0.0250 ± 0.0033*	0.0011 ± 0.0003	0.011 ± 0.0002*
A3AR ag/antag	2.4 ± 0.3*	99 ± 18	0.0242 ± 0.0038*	0.0013 ± 0.0006	0.013 ± 0.0003*
GDM pregnancies
Without insulin
Control	4.3 ± 0.3	120 ± 31	0.0358 ± 0.0059	0.0013 ± 0.0003	0.021 ± 0.004
A1AR ag	4.3 ± 0.3	143 ± 33	0.0301 ± 0.0045	0.0014 ± 0.0004	0.021 ± 0.004
A1AR antag	4.2 ± 0.3	125 ± 25	0.0336 ± 0.0045	0.0012 ± 0.0002	0.019 ± 0.003
A1AR ag/antag	4.1 ± 0.5	119 ± 34	0.0345 ± 0.0071	0.0013 ± 0.0003	0.020 ± 0.004
A2AAR ag	4.0 ± 0.3	150 ± 35	0.0267 ± 0.0041	0.0012 ± 0.0001	0.016 ± 0.002
A2AAR antag	2.1 ± 0.5†	135 ± 42	0.0156 ± 0.0043†	0.0012 ± 0.0001	0.009 ± 0.002†
A2AAR ag/antag	2.2 ± 0.4†	127 ± 26	0.0173 ± 0.0034†	0.0011 ± 0.0003	0.010 ± 0.002†
A2BAR ag	4.3 ± 0.3	134 ± 33	0.0321 ± 0.0051	0.0012 ± 0.0001	0.018 ± 0.002
A2BAR antag	4.1 ± 0.3	120 ± 34	0.0342 ± 0.0060	0.0012 ± 0.0002	0.019 ± 0.003
A2BAR ag/antag	3.9 ± 0.6	129 ± 37	0.0302 ± 0.0067	0.0012 ± 0.0004	0.017 ± 0.004
A3AR ag	4.1 ± 0.3	142 ± 53	0.0289 ± 0.0064	0.0012 ± 0.0001	0.017 ± 0.003
A3AR antag	4.2 ± 0.3	135 ± 43	0.0311 ± 0.0061	0.0012 ± 0.0001	0.018 ± 0.003
A3AR ag/antag	4.0 ± 0.5	121 ± 31	0.0331 ± 0.0064	0.0013 ± 0.0003	0.020 ± 0.004
With insulin
Control	2.2 ± 0.5*	172 ± 40	0.0128 ± 0.0029*	0.0013 ± 0.0003	0.009 ± 0.002*
A1AR ag	2.2 ± 0.4*	142 ± 35	0.0155 ± 0.0033*	0.0014 ± 0.0040	0.011 ± 0.012*
A1AR antag	4.4 ± 0.3†	162 ± 46	0.0272 ± 0.0048†	0.0012 ± 0.0002	0.017 ± 0.003†
A1AR ag/antag	4.1 ± 0.4†	149 ± 35	0.0275 ± 0.0047†	0.0013 ± 0.0003	0.018 ± 0.003†
A2AAR ag	1.9 ± 0.3*	134 ± 31	0.0142 ± 0.0028*	0.0012 ± 0.0001	0.008 ± 0.001*
A2AAR antag	1.9 ± 0.4*	163 ± 95	0.0117 ± 0.0046*	0.0012 ± 0.0001	0.007 ± 0.002*
A2AAR ag/antag	2.2 ± 0.4*	135 ± 25	0.0163 ± 0.0030*	0.0013 ± 0.0003	0.010 ± 0.002*
A2BAR ag	2.2 ± 0.2*	142 ± 37	0.0155 ± 0.0027*	0.0012 ± 0.0001	0.009 ± 0.001*
A2BAR antag	2.0 ± 0.3*	173 ± 72	0.0116 ± 0.0033*	0.0012 ± 0.0001	0.007 ± 0.002*
A2BAR ag/antag	2.2 ± 0.3*	165 ± 25	0.0133 ± 0.0019*	0.0014 ± 0.0005	0.010 ± 0.002*
A3AR ag	2.3 ± 0.2*	162 ± 34	0.0142 ± 0.0021*	0.0012 ± 0.0001	0.009 ± 0.001*
A3AR antag	2.3 ± 0.3*	163 ± 53	0.0141 ± 0.0032*	0.0012 ± 0.0001	0.009 ± 0.002*
A3AR ag/antag	2.2 ± 0.2*	152 ± 25	0.0145 ± 0.0018*	0.0011 ± 0.0005	0.008 ± 0.002*

Data are presented as mean ± SEM (n = 38). L-Arginine transport (0–1000 μM, 1 min, 37 °C) was measured in HUVECs from normal (Normal) or gestational diabetes mellitus (GDM) pregnancies exposed (8 h) to culture medium without (Without insulin) or with (With insulin) insulin (1 nM). Transport assays were done in the absence (Control) or presence of the A₁AR agonist CPA (30 nM, A₁AR ag), antagonist DPCPX (30 nM, A₁AR antag), or both (A₁AR ag/antag), the A_2AAR agonist CGS-21680 (30 nM, A_2AAR ag), antagonist ZM-241385 (10 nM, A_2AAR antag), or both (A_2AAR ag/antag), the A_2BAR agonist BAY60-6583 (100 nM, A_2BAR agonist), antagonist MRS-1754 (30 nM, A_2BAR antag), or both (A_2BAR ag/antag), the A₃AR agonist 2-Cl-IB-MECA (30 nM, A₃AR ag), antagonist MRS-1523 (30 nM, A₃AR antag) or both (A₃AR ag/antag) (see “Materials and methods”). Maximal velocity (V _max) and apparent Michaelis-Menten constant (K _m) of saturable transport were calculated assuming a single Michaelis-Menten hyperbola. V _max/K _m represents maximal L-arginine transport capacity. The lineal phase of overall transport of L-arginine (K _D) was obtained from transport data fitted to a Michaelis-Menten equation increased in a lineal component. Initial velocity (v _i) was calculated for 0.5 s with 100 μM L-arginine transport

*P < 0.05 versus corresponding values in Without insulin; †P < 0.05 versus corresponding Control values

After subtracting the K_D from overall transport, the resulting saturable transport of L-arginine (Fig. 1b) coursed with higher V_max and V_max/K_m values, without variations in the apparent K_m in cells from normal pregnancies exposed to insulin (Table 2). Similar results were found in cells from GDM pregnancies in the absence of insulin. However, insulin reversed GDM-increase in V_max and V_max/K_m to values in normal pregnancies. The A_2AAR antagonist blocked insulin effect in normal pregnancies. In addition, this antagonist blocked GDM effect on saturable transport in the absence of insulin. However, the A₁AR antagonist blocked insulin effect on saturable transport in GDM pregnancies. Parallel experiments show that saturable transport of 100, but not 250 or 500 μM L-arginine, was trans-stimulated by L-lysine reaching similar values in both conditions (Fig. 1c).

HCAT-1 expression

Insulin increased hCAT-1 protein abundance in HUVECs from normal pregnancies to values in GDM (Fig. 2a), but reversed GDM-increase in hCAT-1 to values in normal pregnancies in absence of this hormone. Insulin effect in normal pregnancies was unaltered by the A₁AR antagonist, but this antagonist blocked this hormone’s effect in GDM. However, the A₁AR antagonist in the absence of insulin did not alter the GDM-increased hCAT-1 protein abundance. Insulin effect in cells from GDM was unaltered by an A_2AAR antagonist; however, this antagonist alone blocked the GDM-increase in hCAT-1 (Fig. 2b). Insulin increased hCAT-1 protein abundance in normal pregnancies was blocked by A_2AAR antagonist, but this antagonist did not alter hCAT-1 protein abundance in the absence of insulin. Similar responses to insulin were obtained for hCAT-1 mRNA expression in both cell types (Fig. 2c, d).

Fig. 2.

Involvement of A₁AR and A_2AAR on GDM and insulin effect on hCAT-1 expression. a Western blot for hCAT-1 protein abundance in HUVECs from normal (Normal) or gestational diabetes mellitus (GDM) pregnancies incubated in the absence (−) or presence (+) insulin (1 nM, 8 h) and/or A₁ adenosine receptor (A₁AR) antagonist (Antag). Lower panel, hCAT-1/β-actin ratio densitometries normalized to 1 in cells from normal pregnancies in the absence of insulin or the antagonist. b hCAT-1 protein abundance as in a for A_2AAR antagonist. c hCAT-1 mRNA expression as in a for A₁AR antagonist. d hCAT-1 mRNA expression as in b for A_2AAR antagonist. e Luciferase (Luc) reporter constructs containing two truncations of SLC7A1 promoter (−1606 and −650 bp from the transcription start point) were transfected in HUVECs from normal or GDM pregnancies, along with Renilla reporter plasmid, and assayed for Firefly and Renilla luciferase activity, respectively. Results depict ratio of Firefly/Renilla luciferase activity. After 36 h of transfection, cells were incubated for further 8 h without (Without insulin) or with (With insulin) insulin (1 nM) in the absence (Control) or presence of A₁AR or A_2AAR antagonists. Cells were also transfected with the empty pGL3-basic vector or pGL3-control vector (SV40 pGL3) as negative or positive controls, respectively (see “Materials and methods”). In a and c, *P < 0.05 vs. corresponding values in Normal. †P < 0.05 vs. values in the absence of insulin or A₁AR antagonist. ‡P < 0.05 vs. all other corresponding values. In b and d, *P < 0.05 vs. corresponding values in Normal. †P < 0.05 vs. all other corresponding values. In e, *P < 0.05 vs. vs. all other values except between themselves in the corresponding promoter constructs. Values are mean ± SEM (n = 38)

SLC7A1 promoter activity

Reporter luciferase activity in cells from normal pregnancies transfected with pGL3–hCAT-1⁻¹⁶⁰⁶ or pGL3–hCAT-1⁻⁶⁵⁰ constructs was lower compared with GDM in the absence of insulin (Fig. 2e). Insulin increased the reporter activity for both constructs in normal pregnancies. Insulin effect was unaltered by the A₁AR antagonist, but blocked by the A_2AAR antagonist in cells transfected with the pGL3–hCAT-1⁻¹⁶⁰⁶, but not with the pGL3–hCAT-1⁻⁶⁵⁰ construct. In the absence of insulin, GDM-increased reporter activity for both constructs was reversed only by the A_2AAR antagonist. However, insulin reversed the GDM increase in the reporter activity for both constructs, an effect that was abolished by the A₁AR antagonist and unaltered by the A_2AAR antagonist.

Adenosine receptors expression

A₁AR protein abundance was higher in cells from GDM compared with normal pregnancies and was unaltered by insulin in both cell types (Fig. 3a). On the contrary, A_2AAR was lower in cells from GDM compared with normal pregnancies, and insulin increased this ARs protein abundance in both cell types. However, A_2BAR or A₃AR protein abundance was unaltered by insulin in cells from normal or GDM pregnancies.

Fig. 3.

Adenosine receptors expression. a Western blot for A₁AR (37 kDa), A_2AAR (45 kDa), A_2BAR (37 kDa), and A₃AR (55 kDa) protein abundance in HUVECs from normal (Normal) or gestational diabetes mellitus (GDM) pregnancies incubated in the absence (−) or presence (+) of insulin (1 nM, 8 h). Lower panel, Adenosine receptors (ARs)/β-actin ratio densitometries normalized to 1 in cells from normal pregnancies in the absence of insulin. b Western blot for A₁AR and A_2AAR protein abundance in HUVECs as in A in non-transfected (−) or transfected (+) cells with siRNA against A₁AR (A₁AR-siRNA) or A_2AAR (A_2AAR-siRNA) (see “Materials and methods”). Scramble siRNA (sc-siRNA) was used as control. Lower panel, A₁AR/β-actin or A_2AAR/β-actin ratio densitometries normalized to 1 in cells transfected with sc-siRNA from normal or GDM pregnancies in the absence of insulin. In a, *P < 0.05 vs. corresponding values in Normal, †P < 0.05 vs. corresponding values in the absence of insulin in A_2AAR. In b, *P < 0.05 vs. all other corresponding values in sc-siRNA. Values are mean ± SEM (n = 29)

Adenosine receptors suppression and L-arginine transport

Protein abundance for A₁AR and A_2AAR was abolished in ^KDA₁AR and ^KDA_2AAR cells, respectively, from normal or GDM pregnancies (Fig. 3b). Insulin increase in V_max, V_max/K_m, and v_i for L-arginine transport was unaltered in ^KDA₁AR, but blocked in ^KDA_2AAR cells from normal pregnancies (Table 3). Changes in V_max/K_m caused by insulin in normal or GDM pregnancies were paralleled by similar changes in hCAT-1 protein abundance in ^KDA₁AR (Fig. 4a) or ^KDA_2AAR (Fig. 4b) cells. In addition, insulin effect in hCAT-1 mRNA expression (Fig. 4c) was similar to hCAT-1 protein abundance and activity in both cell types.

Table 3.

Involvement of A₁AR and A_2AAR on L-arginine transport in HUVECs

	V max (pmol/μg protein/min)	K m (μM)	V max/K m (pmol/μg protein/min/(μM))
Normal pregnancies
Without insulin
Control	1.3 ± 0.3	91 ± 18	0.0142 ± 0.0030
KDA1AR	1.1 ± 0.3	99 ± 19	0.0111 ± 0.0026
KDA2AAR	1.1 ± 0.2	97 ± 32	0.0113 ± 0.0028
With insulin
Control	2.6 ± 0.2*	99 ± 24	0.0263 ± 0.0042*
KDA1AR	2.5 ± 0.3*	92 ± 22	0.0272 ± 0.0049*
KDA2AAR	1.2 ± 0.2†	89 ± 27	0.0135 ± 0.0032†
GDM pregnancies
Without insulin
Control	3.9 ± 0.4*	97 ± 25	0.0402 ± 0.0053*
KDA1AR	3.4 ± 0.3*	93 ± 23	0.0366 ± 0.0083*
KDA2AAR	1.4 ± 0.4†	87 ± 35	0.0161 ± 0.0090†
With insulin
Control	1.5 ± 0.4	88 ± 29	0.0170 ± 0.0051
KDA1AR	4.2 ± 0.6*†‡	83 ± 27	0.0506 ± 0.0118*†‡
KDA2AAR	1.1 ± 0.3	97 ± 26	0.0113 ± 0.0031

Data are presented as mean ± SEM (n = 29). L-Arginine transport (0–1000 μM, 1 min, 37 °C) was measured in HUVECs from normal (Normal pregnancies) or gestational diabetes mellitus (GDM pregnancies) pregnancies exposed (8 h) to culture medium without (Without insulin) or with (With insulin) insulin (1 nM). Transport assays were done in cells expressing A₁AR and A_2AAR (Control), cells knockdown for A₁AR (^KDA₁AR) or for A_2AAR (^KDA_2AAR) as described in “Materials and methods.” Maximal velocity (V _max) and apparent Michaelis-Menten constant (K _m) of saturable transport were calculated assuming a single Michaelis-Menten hyperbola. V _max/K _m represents maximal L-arginine transport capacity

*P < 0.05 versus corresponding values in Normal pregnancies Without insulin; †P < 0.05 versus corresponding Control; ‡P < 0.03 versus corresponding values in ^KDA_2AAR

Fig. 4.

GDM and insulin effect on hCAT-1 expression in A₁AR and A_2AAR knockdown cells. a Western blot for hCAT-1 protein abundance in HUVECs in the absence (−) or presence (+) of insulin (1 nM, 8 h) in non-transfected (−) or transfected (+) cells with siRNA against A₁AR (^KDA₁AR). Lower panel, hCAT-1/β-actin ratio densitometries normalized to 1 in cells transfected with sc-siRNA from normal or GDM pregnancies in the absence of insulin. b Western blot for hCAT-1 protein abundance with siRNA against A_2AAR (^KDA_2AAR) as in A. c hCAT-1 mRNA expression in ^KDA₁AR or ^KDA_2AAR cells as in A. d Luciferase (Luc) reporter construct pGL3-hCAT-1⁻¹⁶⁰⁶ of SLC7A1 promoter transfected in ^KDA₁AR or ^KDA_2AAR cells, along with Renilla reporter plasmid. After 36 h of transfection, cells were incubated without (−) or with (+) insulin (1 nM, 8 h) (see “Materials and methods”). e Luciferase (Luc) reporter construct SLC7A1 promoter transfected in ^KDA₁AR or ^KDA_2AAR cells as in d. In a, *P < 0.05 vs. all other values except between themselves, †P < 0.05 vs. all other corresponding values in GDM. In b, *P < 0.05 vs. all other values. †P < 0.05 vs. all other values except between themselves. In c–e, *P < 0.05 vs. all other values in Normal except between themselves. †P < 0.05 vs. all other values in GDM except between themselves. ‡P < 0.05 vs. corresponding values in Normal. Values are mean ± SEM (n = 29)

Insulin increase in pGL3–hCAT-1⁻¹⁶⁰⁶ construct reporter activity was abolished in ^KDA_2AAR, but unaltered in ^KDA₁AR cells from normal pregnancies (Fig. 4d). However, insulin increase in pGL3–hCAT-1⁻⁶⁵⁰ construct reporter activity was unaltered in ^KDA_2AAR or ^KDA₁AR cells from normal pregnancies. Insulin reversed GDM-increase in pGL3–hCAT-1⁻¹⁶⁰⁶ or pGL3–hCAT-1⁻⁶⁵⁰ constructs reporter activity, an effect blocked in ^KDA₁AR, but unaltered in ^KDA_2AAR cells. However, in the absence of insulin GDM-increased constructs, reporter activity was unaltered in ^KDA_2AAR or ^KDA₁AR cells.

Discussion

This study shows that GDM-associated increase in L-arginine transport is mediated by hCAT-1 in HUVECs and is reversed by insulin to values in cells from normal pregnancies via a mechanism requiring expression and activity of A₁AR. This effect of insulin includes reestablishment of hCAT-1 maximal transport capacity (V_max/K_m), and protein abundance and mRNA expression. Insulin restored SLC7A1 (for hCAT-1) gene expression, likely due to activation of a promoter region located between −650 bp from the transcription start point of this gene. In the absence of insulin, GDM-increased hCAT-1 expression and activity depends on A_2AAR expression. Thus, GDM effect on hCAT-1 expression and activity results from differential activation of A₁AR in the presence, but A_2AAR in the absence of insulin in HUVECs.

L-Arginine transport in GDM

L-Arginine transport is mediated by more than one transport system in HUVECs, from where the very-high affinity system y⁺L (K_m ~1 μM), and the system y⁺ family members hCAT-1 (K_m ~100 μM) and hCAT-2B (K_m ~250 μM) [6, 14, 15, 18] play crucial roles. In HUVECs from GDM, the overall L-arginine transport was semisaturable with a non-saturable L-arginine transport component (K_D) detected from L-arginine concentrations >250 μM, suggesting that L-arginine transport was likely mediated by at least two membrane transport mechanisms in this cell type. The possibility that hCAT-1 mediates L-arginine transport is likely since the apparent K_m value for saturable transport is within the range of this transport system in HUVECs [6, 18] and other cell types [14, 15, 19, 20]. In addition, L-arginine transport was trans-stimulated by ~7-fold only when a concentration of L-arginine (100 μM in this study) was close to the K_m for hCAT-1. Since trans-stimulation of hCAT-1-mediated L-arginine transport by L-lysine is known to be higher (~5–10 fold) than hCAT-2B-mediated transport (~2–3 fold) [14, 15, 19–21], involvement of hCAT-2B transport activity in HUVECs from GDM pregnancies is unlikely.

The increased V_max/K_m for saturable transport in cells from GDM was paralleled by similar changes in hCAT-1 protein abundance. Thus, increased transport could result from higher bioavailability of this type of membrane transporters in HUVECs from GDM pregnancies. This phenomenon could be due to higher SLC7A1 expression since the promoter activity for the region from −1606 bp to the transcription start point for this gene and the hCAT-1 mRNA expression were increased in a similar proportion (~2.6-fold) to the increase detected for V_max/K_m and hCAT-1 protein abundance in cells from GDM pregnancies.

Adenosine receptors involvement on insulin effect

We previously reported that insulin increased the V_max/K_m for L-arginine transport and hCAT-1 expression in HUVECs from normal pregnancies [6]. In the present study, insulin reverses the GDM-associated increase in V_max/K_m and hCAT-1 expression in HUVECs. This apparent contradictory effect of insulin on L-arginine transport agrees with previous reports showing a similar dual effect of this hormone on membrane transport mechanisms, including the human equilibrative nucleoside transporters 1 (hENT1) and hENT2 for adenosine uptake in HUVECs [8, 12] and human placental microvascular endothelial cells (hPMECs) [22] from normal or GDM pregnancies. Thus, a dual effect of insulin is not restricted to L-arginine transport via hCATs. Interestingly, insulin did not alter the K_D in cells from normal or GDM pregnancies, suggesting that a parallel plasma membrane transport mechanism to hCAT-1 is either unaltered or minimally altered reaching a negligible influence on overall transport by this hormone. The latter is supported by previous findings showing that insulin did not alter the protein abundance of hCAT-2A/B (i.e., low affinity, high capacity hCAT-2A and high affinity, low capacity hCAT-2B) in HUVECs from normal pregnancies [6]. Additionally, it is reported that insulin increases CAT-1, but not CAT-2B mRNA level and CAT-1-mediated L-arginine transport in rat neonatal ventricular myocites [23]. Thus, a selective phenomenon regarding modulation of hCAT1 and hCAT-2B by insulin is likely in human placenta endothelium.

Protein abundance for A₁AR is higher, but A_2AAR is lower in cells from GDM compared with normal pregnancies in the absence of insulin, suggesting that these two ARs subtypes play a role in HUVECs under this pathological condition. Since GDM-increased V_max/K_m for transport was blocked by an A_2AAR antagonist in the absence of insulin, this ARs subtype could be acting as hCAT-1 activator. Extracellular level of adenosine in primary cultures of HUVECs [4, 24], hPMECs [22], and in the umbilical vein blood [6, 8, 22, 25] from GDM pregnancies is higher (~1 μM) than in normal pregnancies (~0.4 μM). Since A_2AAR affinity for adenosine (~0.3 μM) [10, 26] is within this range of the adenosine concentration detected in GDM, an activated state of these receptors is expected. However, we cannot rule out the possibility that the increase in the extracellular concentration of adenosine could lead to adenosine-reduced expression of A_2AAR in HUVECs from GDM. Additionally, as a potential limitation of our study, it is feasible that adenosine release from HUVECs after 1 min of incubation (time used for transport assay measurements), even in the presence of adenosine deaminase 1, was enough to saturate A₁AR and A_2AR and cause a basal increase in L-arginine transport in this cell type. Since HUVECs from GDM pregnancies are shown to express hENT1 and hENT2 accounting for equilibrative adenosine transport [8, 12], and because hENT1 expression and activity is downregulated in HUVECs from GDM [8, 12], we speculate on the possibility that a potential efflux of adenosine via hENT2 from HUVECs is likely. Even when the affinity of A₁AR and A₃AR for adenosine are ~0.1 and ~0.3 μM, respectively [10], agonists or antagonists for these ARs subtypes did not alter GDM-effect on transport, suggesting that their involvement in this phenomenon is unlikely.

Insulin reverses GDM-associated increase in L-arginine transport most likely via a mechanism that requires A₁AR since an antagonist for A₁AR, but not other ARs antagonists, blocked insulin effect. Thus, a change in the phenotype in HUVECs from GDM caused by insulin regarding these receptor subtypes is plausible. The latter could result from a change in the metabolic state of this type of cells. Supporting this possibility is the recently proposed change caused by insulin from a preferential mitogenic phenotype (i.e., activated p44/44^mapk/activated Akt ratio >1) to a preferential metabolic phenotype (i.e., activated p44/44^mapk/activated Akt ratio <1) in HUVECs from GDM pregnancies [8]. Since A₁AR activation leads to p44/42^mapk and Akt-mediated signaling, but A_2AAR activation leads to p44/42^mapk signaling [10, 26], a preferential A₁AR-associated, Akt-dependent metabolic phenotype could be caused by insulin in HUVECs from GDM. This could be a phenomenon potentiated in cells from GDM pregnancies since a lower A_2AAR protein abundance is detected in this cell type, thus further reducing the possibility of an A_2AAR-associated, p44/42^mapk-dependent mitogenic phenotype. Interestingly, insulin increases the A_2AAR protein abundance in both cell types, which could be interpreted as a more preferential mitogenic phenotype. However, considering that this hormone did not alter A₁AR protein abundance in cells from normal pregnancies, but increased its protein abundance in GDM, a predominant A_2AAR-associated, normal-like phenotype (A_2AAR/A₁AR ratio ~1.7) induced by insulin is seen in cells from normal pregnancies. Since A_2AAR/A₁AR ratio was ~1.3 in cells from GDM, a minor role of A_2AAR with a major role of A₁AR in the response to insulin in HUVECs is likely.

Since suppression of A₁AR expression (^KDA₁AR cells), but not in ^KDA_2AAR cells, abolished the changes caused by insulin in GDM, A₁AR role on insulin effects is directly supported. Interestingly, in the absence of insulin, the GDM effect on hCAT-1 expression and activity was reversed only in ^KDA_2AAR cells, complementing the results obtained with the use of an A_2AAR antagonist in cells expressing this ARs subtype, or when an A_2AAR agonist was used in ^KDA_2AAR cells. Insulin effect on L-arginine transport in GDM depends on an A₁AR-dependent activation of SLC7A1 expression at the promoter region −650 bp from the transcription start point. Previous studies show that this region of the SLC7A1 promoter contains at least four consensus sites between −177 and −105 bp from the transcription start point for the general transcription factor specific protein 1 (Sp1), and that Sp1 binding to these sites is increased by insulin in cells from normal pregnancies [18]. Thus, Sp1 as well as other transcription factors associated with the control of expression of membrane transporters families, such as C/EBP homologous protein 10 (CHOP) binding to SLC7A1 in C6 rat glioma cells [27], could be responsible of insulin modulation of hCAT-1 expression in cells from GDM pregnancies. There is no information addressing a role for Sp1 as a cell signaling mechanism triggered by activation of ARs [10, 11, 26]. Thus, we do not rule out the possibility that biological effects of insulin could require adenosine acting via A₁AR as a helper or positive regulator via Sp1 activation in HUVECs from GDM pregnancies. Since umbilical blood insulin level in GDM pregnancies is higher (~1.9-fold) compared with normal pregnancies, and because a larger concentration of insulin was required to reverse GDM increase in L-arginine transport in vitro, the observed increase in the plasma insulin at the fetal circulation may be not enough to restore GDM-associated alterations in the human fetoplacental circulation. The latter could also result from fetal insulin resistance as suggested by an elevated HOMA-IR value (~2.2-fold versus normal pregnancies) detected in this group of newborns, agreeing with previous reports in GDM [8, 22, 28].

In summary, GDM is associated with increased L-arginine transport in HUVECs, which results from higher hCAT-1 expression and activity. Insulin reverses GDM-associated effects on transport requiring activation and higher expression of A₁AR in this cell type. In the absence of insulin, GDM effect on hCAT-1 expression and activity is dependent on A_2AAR expression and activation (see Fig. 5). A simple interpretation of our results is that insulin increases adenosine extracellular concentration in cells from normal pregnancies leading to predominant activation of A_2AAR to mediate insulin increase in L-arginine transport. However, cells from GDM pregnancy show induction of A₁AR with a predominant response resulting from activation of this type of receptor to mediate insulin decrease in L-arginine transport. Thus, a differential modulatory effect of insulin is played by adenosine receptors in endothelial cells from the macrovasculature in the human placenta from GDM. Since this GDM-associated fetal hyperinsulinemia is a condition that seems not to be enough to restore umbilical vein endothelial function, and because a state of less sensitivity to insulin in GDM fetoplacental vasculature is feasible, a higher concentration of insulin could be required to actually see a beneficial effect of this hormone. Insulin therapy is a protocol applied in pregnant women with GDM that are not responsive to diet and/or exercise to normalize their glycemia. However, this approach associates with a risk of the offspring to develop adulthood diseases such as diabetes mellitus type 2 [29, 30] and of the mother to course with GDM in a future pregnancy [31]. Thus, we emphasize the need of a therapeutical approach considering the potential metabolic modulation of circulating adenosine level and/or adenosine receptors activation/inactivation complementing insulin therapy protocol in pregnant women with GDM to restore fetoplacental endothelial dysfunction for the benefit of the mother and the newborn [3].

Fig. 5.

Proposed involvement of A₁AR on insulin-reversal of GDM increased L-arginine transport in human umbilical vein endothelial cells. In the presence of an extracellular level of insulin similar to that detected in the human umbilical vein blood in gestational diabetes mellitus (GDM) (0.07 nM insulin (see Table 1)), the human cationic amino acid transporter 1 (hCAT-1) protein abundance and availability at the plasma membrane of human umbilical vein endothelial cells (HUVECs) result in an increase () in transport of L-arginine (orange arrow) compared with cells from normal pregnancies. This phenomenon is maintained by a higher expression of the SLC7A1 gene for hCAT-1 leading to higher hCAT-1 mRNA expression and protein abundance. GDM effect is caused by adenosine activation of A_2AAR, but not A₁AR leading to increased activity of transcription factors (TFs ?) on the −650 bp from the transcription start point fragment of the SLC7A1 promoter under this environmental condition. Insulin activation of insulin receptors (IRs) does not contributes to the modulation of expression and activity of hCAT-1 in this cell type. Higher extracellular insulin concentration (1 nM) reverses () GDM-associated changes in hCAT-1 expression (SLC7A1 promoter activity, mRNA and protein abundance) and activity (narrow orange arrow) to values in cells from normal pregnancies. Insulin biological effect requires activation of A₁AR, but not A_2AAR, with this subtype of adenosine receptors acting as helper or positive regulator of insulin biological actions in HUVECs from GDM pregnancies

Abbreviations

GDM

Gestational diabetes mellitus

NO

Nitric oxide

eNOS

Endothelial NO synthase

HUVECs

Human umbilical vein endothelial cells

hPMECs

Human placental microvascular endothelial cells

ARs

Adenosine receptors

A₁AR

A₁ adenosine receptors

A_2AAR

A_2A adenosine receptors

A_2BAR

A_2B adenosine receptors

A₃AR

A₃ adenosine receptors

hCAT-1

Human cationic amino acid transporter 1

hENT1

Human equilibrative nucleoside transporters 1

siRNA

Short interference RNAs

^KDA₁AR

A₁AR knockdown cells

^KDA_2AAR

A_2AAR knockdown cells

CHOP

C/EBP homologous protein 10.

Compliance with ethical standards

Funding

This works was supported by Fondo Nacional de Desarrollo Científico y Tecnológico (FONDECYT 1150377, 1150344, and 11150083), Chile.

Conflict of interest

The authors declare that they have no competing interests.