Dynamic glucose enhanced MRI of the placenta in a mouse model of intrauterine inflammation

Highlights

• Feasibility of glucoCEST to access placental function was examined in a mouse model.

• GlucoCEST effect was obtained with an on-resonance variable delay multi-pulse method.

• Time-resolved glucoCEST effect increased after glucose infusion in normal placenta.

• Placental glucoCEST effect reduced acutely after intrauterine inflammation.

• Glucose transporter 1 did not respond to glucose infusion in the injured placentae.

Abstract

Introduction

We investigated the feasibility of dynamic glucose enhanced (DGE) MRI in accessing placental function in a mouse model of intrauterine inflammatory injury (IUI). DGE uses the glucose chemical exchange saturation transfer (glucoCEST) effect to reflect infused d-glucose.

Methods

IUI was induced in pregnant CD1 mice by intrauterine injection of lipopolysaccharide (LPS) on embryonic day 17. In vivo MRI was performed on an 11.7 T scanner at 6 h s after injury, and glucoCEST effect was measured using an on-resonance variable delay multi-pulse (onVDMP) technique. onVDMP acquisition was repeated over a period of 25 min, and d-glucose was infused 5 min after the start. The time-resolved glucoCEST signals were characterized using the normalized signal difference ($\Delta S_N$) between onVDMP-labeled and nonlabeled images.

Results

$\Delta S_N$ in the PBS-exposed placentae (n = 6) showed an initial drop between 1 and 3 min after infusion, followed by a positive peak between 5 and 20 min, the time period expected to be associated with the process of glucose uptake and transport. In the LPS-exposed placentae (n = 10), the positive peak was reduced or even absent, and the corresponding area-under-the-curve (AUC) was significantly lower than that in the controls. Particularly, the AUC maps suggested prominent group differences in the fetal side of the placenta. We also found that glucose transporter 1 in the LPS-exposed placentae did not respond to maternal glucose challenge.

Discussion

DGE-MRI is useful for evaluating placental functions related to glucose utilization. The technique uses a non-toxic biodegradable agent (d-glucose) and thus has a potential for rapid translation to human studies of placental disorders.

Introduction

Glucose, as the primary substrate for energy production and a mediator in many biosynthetic processes [1,2], is crucial to fetal development [3]. The placental capacity to deliver glucose to the fetus is an important indicator of placental health. Placental utilization of glucose has been investigated for many years, for example, malfunction of glucose transporters was found to be a contributor in a range of placental disorders, such as preeclampsia [4,5] and intrauterine growth restriction (IUGR) [6,7]. However, such studies were performed using in vitro approaches in excised placental tissue slices, isolated placental villi, or perfused placenta [8]. Clinical measures are mostly limited to biochemical sampling of the maternal blood, e.g., hormonal factors in a glucose tolerance test [9]. Direct in vivo assessment of glucose utilization in the placenta would be optimal and clinically useful, but such techniques are currently not available.

Magnetic resonance imaging (MRI) has become an increasingly important imaging tool to evaluate placental anatomy and function, in addition to ultrasound [10], especially in complex cases. For example, dynamic contrast enhanced (DCE) MRI [11,12], arterial spin labeling (ASL) [13,14], and intravoxel-incoherent motion (IVIM) MRI [[15], [16], [17], [18]] have been adopted to measure placental perfusion; and T2 and T2* mapping [[19], [20], [21]], and BOLD MRI [22,23] have shown sensitivity to placental oxygenation. Traditionally, the detection of glucose concentration and metabolic rate is performed by MR spectroscopy [[24], [25], [26], [27], [28], [29], [30]], which is a slow and low-resolution technique. Recently, it was demonstrated that glucose in brain and body organs can be monitored with Chemical Exchange Saturation Transfer (CEST) MRI [31,32], as well as relaxometry-based MRI contrasts, such as T1ρ [33,34] and T2-weighted MRI [35]. Briefly, CEST MRI utilizes radiofrequency (RF) irradiation to selectively saturate exchangeable protons in the low-concentration solutes typically only measurable with MRS. This saturation is subsequently transferred to water protons through proton exchange, replaced by non-saturated protons and again saturated. Because of the rapid exchange of these protons, the process is repeated many times within a single MRI detection, resulting in the building up of a reduction in water signal intensity. Thus, CEST MRI can amplify the signal from low-concentration metabolites to be detected via the abundant bulk water protons, in a manner proportional to the concentration of these agents. By tuning the CEST frequency to the hydroxyl protons on d-glucose, called glucoCEST MRI, we and others have shown the detection of the presence of d-glucose or its derivatives using the MRI signal in vivo [[36], [37], [38]]. The measurement of time-resolved changes of glucoCEST MRI signal after a bolus intravenous infusion of glucose, referred to as Dynamic Glucose Enhanced (DGE) MRI, reports the changes in glucose concentration in biological tissues and thus contains information on glucose delivery, transport and metabolic kinetics. DGE MRI using continuous-wave (CW) or pulsed CEST has been used to image the glucose uptake in tumors and other tissues both in animals [[36], [37], [38], [39], [40], [41]] and humans [42]. Feasibility of DGE in the placenta is not yet known.

An On-Resonance Variable Delay Multiple Pulse (onVDMP) technique [43] was recently developed to further enhance the sensitivity of glucoCEST. In contrast to CW-CEST that targets a single off-resonance frequency, the onVDMP method efficiently labels a range of fast-exchanging protons, covering multiple resonance frequencies of the five hydroxyl protons with exchange rates between 3 and 6 kHz. It has been demonstrated that DGE signal changes in the brain detected by onVDMP were about two times higher than those measure by CW-CEST [44]. In this study, we investigated the feasibility of onVDMP-based DGE to monitor glucose kinetics in normal and injured placentae. The study was performed in an established mouse model of intrauterine inflammation (IUI) [45,46], in which placental and fetal injury was induced by excessive maternal production of inflammatory cytokines due to lipopolysaccharide (LPS) administration. This model mimics the most common clinical scenario associated with preterm birth. In this model, we have previously identified acute maternal vascular malperfusion (MVM) with IVIM and immunohistochemistry measurements in the placenta [47], and thus, the model serves as an ideal test bed for potential clinical translation of DGE MRI to assess placental function. In this study, we performed DGE MRI to examine the time-resolved placental response to a glucose challenge in LPS-exposed placentae, in comparison with the controls.