Molecular Magnetic Resonance Imaging with Gd(III)-Based Contrast Agents: Challenges and Key Advances

Abstract

In an era of personalized medicine, the clinical community has become increasingly focused on understanding diseases at the cellular and molecular levels. Magnetic resonance imaging (MRI) is a powerful imaging modality for acquiring anatomical and functional information. However, it has limited applications in the field of molecular imaging due to its low sensitivity. To expand the capability of MRI to encompass molecular imaging applications, we introduced bioresponsive Gd(III)-based magnetic resonance contrast agents (GBCAs) in 1997. Since that time, many research groups across the globe have developed new examples of bioresponsive GBCAs. These contrast agents have shown great promise for visualizing several biochemical processes, such as gene expression, neuronal signaling, and hormone secretion. They are designed to be conditionally retained, or activated, in vivo in response to specific biochemical events of interest. As a result, an observed MR signal change can serve as a read-out for molecular events. A significant challenge for these probes is how to utilize them for noninvasive diagnostic and theranostic applications. This Perspective focuses on the design strategies that underlie bioresponsive probes, and describes the key advances made in recent years that are facilitating their application in vivo and ultimately in clinical translation. While the field of bioresponsive agents is embryonic, it is clear that many solutions to the experimental and clinical radiologic problems of today will be overcome by the probes of tomorrow.

Figure 1.

Chemical structures of Gd(III)-DTPA, Gd(III)-DOTA and Gd(III)-DO3A. Different R groups (red) can be inserted to create bioresponsive GBCAs.

Figure 2.

Key parameters that affect inner-sphere relaxivity include hydration number (q), mean residence time of the bound water (τ_m), and rotational correlation time (τ_R).

Figure 3.

Effect of rotational correlation time on r₁ as a function of field for a Gd(III) complex with a water residency time 100 ns. τ_R = 0.1 ns (···), 1.0 ns (---), 10 ns (—). At low field (1.5 T), long τ_R (10 ns) gives the highest r₁. However, at high field, the intermediate τ_R (1 ns) gives the highest r₁. Reproduced with permission from ref 13. Copyright 2009 John Wiley and Sons.

Figure 4.

Depiction for targeted GBCAs. Upon binding the target biomarker, the GBCAs exhibit enhanced retention in vivo as well as increased r₁ due to RIME effect.

Figure 5.

(A) Chemical structure and targeting mechanism of GdOA. GdOA has an oxyamine group that can reversibly bind to allysine present in oxidized collagen. GdOX is a control probe that does not interact with allysine. (B) Coronal MR images (9.4 T) of normal and bleomycin injured mice injected with GdOA or GdOX. Bleomycin injured mice injected with GdOA (2nd) clearly showed an enhanced MR signal in the lung compared to normal mice injected with GdOA (1st) or injured mice injected with the control probe GdOX (3rd). In the last image, β-aminopropionitrile (BAPN) which inhibits allysine production was used in bleoinjured mice, and little MR signal enhancement was seen when mice were injected with GdOA. Reproduced with permission from ref 19. Copyright 2017 John Wiley and Sons.

Figure 6.

(A) Chemical structures of PSMA targeting GBCAs, Gd1, Gd2, Gd3. The R group is the targeting ligand for PSMA. (B) T₁-weighted MR signal enhancement map (9.4 T) in PC3 PIP (PSMA+) and PC3 flu (PSMA-) tumors superimposed upon T₂-weighted images after the mice were injected with Gd3. (C) T₁ time courses calculated for the entire volume of each tumor during 1–1600 min post injection. Gd3 was retained in the PSMA+ PIP tumor in the first 3hrs but was quickly washed out from PSMA-flu tumor. Reproduced with permission from ref 20. Copyright 2017 John Wiley and Sons.

Figure 7.

Chemical structures of HaloTag targeting GBCAs of different linker length. 2CHTGd has both optimal binding and maximal relaxivity increase upon binding.

Figure 8.

Pictorial depiction for activatable GBCAs that can be used to image enzyme activities. Upon cleavage of the capping ligand, the inner-sphere relaxivity is increased as water access to the metal center is resumed.

Figure 9.

(A) Chemical structures of β-galactosidase-responsive GBCAs: Egad, α-EGadMe and β-EGadMe. (B) MRI (11.7 T) detection of β-galactosidase mRNA expression in living X. laevis embryos injected with EgadMe at the two-cell stage. The embryo on the right was also injected with β-gal mRNA, resulting in higher MR signal in certain regions. s, somite; b, brachial arches; e, eye; c, cement gland. Scale bar = 1 mm. Reproduced with permission from ref 27a. Copyright 2000 Springer Nature.

Figure 10.

Activatable GBCA for detecting esterase activities. Esterase unmasks the carboxylate groups which repels the carbonate binding to the Gd(III) metal center.

Figure 11.

Pictorial depiction for activatable GBCAs for sensing metal ions. Upon binding of the target metal ion, the capping ligand is lifted leading to increased q and hence higher inner-sphere relaxivity.

Figure 12.

Chemical structure of the Ca(II)-sensitive GBCA, DOPTA–Gd. The BAPTA core is responsible for binding with Ca(II) of physiologically relevant concentration, leading to increased q and hence higher r₁.

Figure 13.

Pictorial depiction for activation-binding mechanism. These GBCAs are first conditionally activated which then acquired the ability to bind to proteins such as serum albumin, leading to increased r₁ and retention.

Figure 14.

(A) Chemical structures of Zn(II)-sensitive GBCAs that upon binding with Zn(II) acquire the ability to interact with HSA. GdL2 has lower affinity towards Zn(II) than GdL1. (B) Pictorial depiction of the activation-binding mechanism of GdL1 and GdL2. (C) 3D T₁-weighted MRI (9.4 T) of mouse pancreas pre- and postdelivery of GdL1 or GdL2 plus saline or glucose. Glucose stimulates the release of Zn(II) from β-cells in the pancreas, whereas saline serves as a vehicle control. (D) average MR from 0 to 28 min for saline-treated mice (n = 4) and for glucose-treated animals. Although GdL1 gave higher signal under glucose stimulation, GdL2 had lower signal background when only saline was injected. This was attributed to the lower affinity of GdL2 towards background Zn(II) in the tissue. Bars represent the standard error of the mean; *p-value < 0.05, **p-value < 0.01 Reproduced with permission from ref 18f. Copyright 2018 American Chemical Society.

Figure 15.

Pictorial depiction for activatable GBCAs that based on responsive self-assemble or aggregation. Note that these agents usually are better retained post-activation, as larger particles tend to have a slower clearance rate.

Figure 16.

(A) Chemical structures of MPO-sensitive GBCAs, MPO-Gd. In the presence of MPO, the 5-HT moiety is oxidized to form radicals, leading to oligomerization of the MPO-Gd as well as linking to surrounding proteins. (B) In vivo MRI (7 T) of MPO activity 2 days after myocardial infarction was induced. Strong and persistent MR signal enhancement was seen in the infarct zone (denoted by the yellow arrow) in mice injected with MPO-Gd, but not with Gd(III)-DTPA (C) The signal-to-noise ratio (SNR) showed higher values and stronger retention after injection of MPO-Gd compared to Gd(III)-DTPA. Reproduced with permission from ref 35a. Copyright 2008 Wolters Kluwer Health, Inc.

Figure 17.

Caspase-3/7 responsive GBCA, C-SNAM, is activated by cleavage of the DEVD peptide by caspase-3/7 and disulfide exchange with intracellular GSH. It then undergoes a biocompatible intramolecular cyclization to form a rigid and hydrophobic macrocycle. Reproduced with permission from ref 36b. Copyright 2015 American Chemical Society.

Figure 18.

(A) GGT-responsive GBCA, 1, is activated by cleavage of the glutamate by GGT and disulfide exchange with intracellular GSH. It then dimerizes to form a rigid and hydrophobic macrocycle, whichself-assembles into nanoparticles (B) T₂-weighted coronal MRI (9.4 T) of HeLa tumor bearing mice intravenously injected with 1 (top row), DON for 0.5 h and then 0.08 mmol/kg (middle row), and 0.08 mmol/kg Gd(III)-DTPA (bottom row) at 0 h and 2.5 hrs. (C) Normalized time course tumor-to-muscle (T/M) contrast ratios of T₂ values in panel B. Significant T₂ reduction was observed with Hela-bearing mice 2.5 hrs post injection with 1. Each error bar represents the standard deviation of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001; analyzed by Student’s t-test. Reproduced with permission from ref 37. Copyright 2019 American Chemical Society.

Figure 19.

(A) Manganese-based intracellular calcium sensor (ManICS1-AM) is cell permeable. When the agent enters cells, the AM esters are cleaved by intracellular esterases. This liberates the “off-state”, which can interact with intracellular Ca²⁺ and turns on the MR signal. (B) 1 μL K⁺ infusion causes T₁-weighted MR signal (9.4 T) to increase in the presence of pre-delivered ManICS1-AM (left) but not MnL1F (right), a calcium-insensitive control. Average peak signal change across multiple animals (n = 5) is indicated by the color scale superimposed on a high resolution T₁-weighted image of a representative rat. Scale bar = 3 mm. (C) Region of interest analysis shows the time course of signal changes observed during K⁺ or Na⁺ in the presence of ManICS1 (red and cyan, respectively), and during K⁺ stimulation (vertical gray bar) in the presence of calcium-insensitive MnL1F (blue). Reproduced from ref 42. Distributed under the terms of the Creative Commons CC BY license.

Figure 20.

(A) Fe-based redox-sensitive MRCA Fe(II)-Pyc3A has low r₁ due to diamagnetic Fe(II). In oxidative environment the Fe(II) is oxidized to Fe(III) that is paramagnetic. As a result, a large r₁increase was achieved. (B) T₁-weighted 2D axial images (4.7 T) of saline and caerulein/LPS treated mice recorded prior to and 6 min after injection of 0.2 mmol/kg Fe(II)-PyC3A. Organs are labeled as follows: P = pancreas, Sp = spleen, K = kidney, M = muscle, St = stomach, B = bowel. Note that the pancreas and neighboring kidney are virtually isointense prior to probe injection (top two images). After injection of Fe(II)-PyC3A to saline treated mice the pancreas and kidney remain isointense (bottom left), but that the pancreas is strongly and selectively enhanced after injection of Fe(II)-PyC3A to caerulein/LPS treated mice (bottom right). Reproduced with permission from ref 48e. Copyright 2019 American Chemical Society.