Single (19)F probe for simultaneous detection of multiple metal ions using miCEST MRI.

The local presence and concentration of metal ions in biological systems has been extensively studied ex vivo using fluorescent dyes. However, the detection of multiple metal ions in vivo remains a major challenge. We present a magnetic resonance imaging (MRI)-based method for noninvasive detection of specific ions that may be coexisting, using the tetrafluorinated derivative of the BAPTA (TF-BAPTA) chelate as a (19)F chelate analogue of existing optical dyes. Taking advantage of the difference in the ion-specific (19)F nuclear magnetic resonance (NMR) chemical shift offset (Δω) values between the ion-bound and free TF-BAPTA, we exploited the dynamic exchange between ion-bound and free TF-BAPTA to obtain MRI contrast with multi-ion chemical exchange saturation transfer (miCEST). We demonstrate that TF-BAPTA as a prototype single (19)F probe can be used to separately visualize mixed Zn(2+) and Fe(2+) ions in a specific and simultaneous fashion, without interference from potential competitive ions.

A major challenge in the biomedical sciences is to monitor, characterize, quantify, and understand the multiplexity of biological events in vivo. Advanced imaging methodologies are being developed to visualize multiple biological changes simultaneously within the same anatomical frame. One strategy is the use of multimodal imaging approaches, where more than one imaging methodology is used to obtain information from multiple targets.[1−6] However, the complexity of coregistering the obtained information into an accurate spatial representation calls for probing multiple targets using a single imaging approach.

Metal ions play a pivotal role in nearly all biological processes, and deviation from normal levels is often associated with disease onset and progression.[7] Today, our knowledge of the role of metal ions in biology is mostly based on the use of optical dyes,[8] originally developed by Roger Tsien.[9,10] Although optical dyes have made an enormous contribution to an understanding of the role of metal ions in biological systems, the optical signal from fluorescent dyes limits their applications to in vitro studies or monitoring of surface phenomena in vivo using superficially injected dyes.[11] To overcome these limitations, magnetic resonance imaging (MRI) has been explored as a whole-body, noninvasive imaging technique to sense changes in metal ion levels in vivo.[12−14] However, currently available probes are designed to alter the T1 and T2 proton relaxation rates upon binding to the metal ion of interest,[15−17] where interpreting images and quantifying local metal levels may be difficult as this approach is not specific: changes in T1 or T2 may result from other sources, while the background contrast without the presence of metals is often unknown. The specific chemical shifts (Δω) of nuclear magnetic resonance (NMR)-detectable nuclei (e.g., 19F or 129Xe) in a synthetic probe upon metal ion binding provide ultimate specificity with regard to the ion of interest.[18−21] Unfortunately, NMR spectroscopy-based approaches do not provide spatial information on the location of the investigated ion and rely on the identification and integration of a specific NMR peak that may fall below a detectable signal to noise ratio (SNR).

Chemical exchange saturation transfer (CEST) imaging[22−26] is an MRI contrast mechanism that enables the detection of low concentration solutes via the transfer of their magnetization to the bulk (high concentration) nuclei, from which the MRI signal is derived. Using 1H as the bulk nucleus, CEST MRI has been used in a wide range of applications,[22−26] including simultaneous imaging of different probes, based on their different Δω values.[27,28] In our previously suggested approach, which we termed ion CEST (iCEST),[29] a combination of 19F MRI and CEST was used to spatially monitor Ca2+ with high specificity, capitalizing on the dynamic exchange between the ion-bound and free 19F chelate, and the shift in the Δω of 19F upon ion binding. By using 5,5′,6,6′-tetrafluoro-BAPTA (TF-BAPTA) as the 19F iCEST probe (Figure 1a), we demonstrate here that Zn2+ and Fe2+ ions can be detected specifically and simultaneously.

Figure 1

(a) Chemical structure of TF-BAPTA showing the 19F atom substituents on the 5 (purple) and 6 (green) positions. (b) 19F NMR spectrum (470 MHz) of 5 mM TF-BAPTA (20 mM Hepes buffer, pH = 7.2) in the presence of 0.5 mM Zn2+ or Fe2+.

It has been previously demonstrated that different 19F-BAPTA derivatives have different Δω values in their respective 19F NMR spectra and variable Kd properties for various metal ions.[19] Because of the fact that free TF-BAPTA exchanges too rapidly (kex ≈ 10,000 s–1) with Ca2+-bound TF-BAPTA[30] to be useful for generating iCEST contrast, we hypothesized that it could be used for the detection of other metal ions. Figure 1b shows the 19F NMR spectrum of TF-BAPTA in the presence of either Zn2+ or Fe2+ (10:1 molar ratio). The Δω of the 19F atoms at 5- (purple) and 6-positions (green) are shifted downfield and upfield, respectively, in the presence of either of the ions, with a larger effect for the paramagnetic Fe2+. One of the potential drawbacks of 19F MRI using 5F-BAPTA is a possible line broadening of the bulk signal of the free ligand in vivo in live tissue,[19] such as seen when high amounts of Mg2+ are added (Figure S1, Supporting Information (SI)). As a result, images with reduced SNR may be experienced and smaller observed Δω values may not be sufficient for selective saturation of poorly shifted nuclei without direct bulk saturation. However, as also previously demonstrated,[30] the fast exchange between Ca2+ and TF-BAPTA broadens the peak that is related to the 5-positioned 19F atom and does not affect the NMR characteristics of the 6-positioned 19F atom (Figure S2, SI). Additionally, a high Mg2+ concentration does not affect the NMR properties of 5F and 6F atoms of TF-BAPTA (Figure S2, SI), making the latter a suitable 19F MRI probe in a biological setup. The 6-positioned 19F atom for the signal of the bulk (in 19F-CEST experiments) is thus preferable since it does not broaden due to ion exchange. The two frequencies that are observed in the 19F NMR spectrum of TF-BAPTA require the center frequency offset (O1) to be placed at the resonance of one of these frequencies when performing 19F MRI. Therefore, all 19F MRI experiments in this study were performed with O1 set at the frequency of the 6-positioned 19F, while the signal from the 5-positioned 19F was suppressed using a spectrally selective excitation pulse and spoiler gradient.

Figure 2b,c shows the 1H and 19F MR images of seven tubes containing 10 mM TF-BAPTA and 200 μM added ion, without any observable changes in 1H or 19F MR contrast. However, the 19F iCEST images show a clear differential MR contrast between the samples containing Zn2+ (Figure 2d) and Fe2+ (Figure 2e), for a saturation pulse applied at Δω = −2.8 and −18 ppm, respectively. These Δω values were chosen from the 19F NMR spectra, using the offset values of TF-BAPTA upon the addition of Zn2+ or Fe2+, respectively (see Figure 1b). Figure 2f clearly shows that both ions can be simultaneously visualized using TF-BAPTA as a single iCEST probe. Figure 3 shows the corresponding 19F iCEST spectra for samples containing either Zn2+ (Figure 3a) or Fe2+ (Figure 3b). The dynamic 19F exchange between TF-BAPTA and [M2+-TF-FBAPTA] results in an iCEST effect for both ions, at Δω = −2.8 ppm for [Zn2+-TF-BAPTA] and at Δω = −18 ppm for [Fe2+-TF-BAPTA], respectively. Using Bloch simulations (Figure 3a,b), the exchange rate (kex) between free and bound TF-BAPTA is estimated to be ∼20 s–1 for both ions. This kex is rather low, and much higher CEST contrast may be obtained for 19F chelates with higher kex values. Despite this slow exchange, we were still able to detect 10% CEST contrast for a 200 μM ion concentration with the sensitivity from a 10 mM signal strength. The use of 19F based CEST enables a reduction in the concentration of the 19F iCEST probe to a biological relevant molar ratio (probe: ion), a feat that is not possible with 1H CEST, which is based on water. Additionally, 19F enables “hot spot” tracer detection without an endogenous background signal,[31] contrary to 1H CEST, which suffers from a large nonspecific endogenous background signal. This may further reduce the 19F probe concentration to below 10 mM, alleviating potential toxicity effects from Ca2+ buffering. Although TF-BAPTA did not show a significant buffering effect for intracellular Ca2+,[30] further studies are needed prior to its use in vivo. Importantly, when balanced salt solutions containing physiological levels of other ions (1.3 mM Ca2+, 0.9 mM Mg2+, 5.9 mM K+, and 143 mM Na+) and glucose (6 mM) were used, the iCEST effect from Zn2+ was not affected (Figure S3, SI). This is a great advantage for the use of TF-BAPTA as an iCEST probe compared to 5F-BAPTA, which exchanges much faster with other metal ions, causing broadening of the bulk signal in the 19F NMR spectrum, limiting its applications.

Figure 2

iCEST maps. (a) The orientation of the samples in the phantom containing 10 mM TF-BAPTA and 200 μM ion (pH = 7.2). (b) 1H MRI, (c) 19F MRI, (d) iCEST (Δω = −2.8 ppm) overlaid on 19F MRI, (e) iCEST (Δω = −18 ppm) overlaid on 19F MRI, and (f) both iCEST results (Δω = −2.8 ppm, Δω = −18 ppm) overlaid on 19F MRI.

Figure 3

19F iCEST spectra for samples containing 10 mM TF-BAPTA and 200 μM Zn2+ (a) and Fe2+ (b). Circles represent experimental signal; solid lines represent Bloch simulations (two-pool model).

The unique and different Δω value of the exchangeable moiety is one of the most exceptional characteristics of iCEST compared to other MRI sensors. This feature gives CEST sensors an artificial color designation, by which they can be tagged in a singular specific frequency, much like fluorescent dyes. For 1H CEST, this has been exploited for “multi-color” MRI of live cells[28] and in vivo.[27] Here (Figure 4) we investigated whether Zn2+ and Fe2+ could be distinguished from each other when mixed together and with other ions. When a saturation pulse was applied at the resonance of the Zn2+-TF-BAPTA complex (i.e., Δω = −2.8 ppm), only the tubes that contained Zn2+ ions generated an observable iCEST contrast. The contrast did not change when competing ions, such as Ca2+, Mg2+, or Fe2+, were included in the sample solution. Similarly, when the saturation pulse was applied at Δω = −18 ppm (Δω of Fe2+-TF-BAPTA), only the samples that included Fe2+ generated iCEST contrast, without interference from the other coexisting ions Ca2+, Mg2+, or Zn2+. Notably, when both Zn2+ and Fe2+ ions were mixed with TF-BAPTA (center tube in Figures 4a–d), the iCEST contrast could be obtained at both Δω values of Zn2+-TF-BAPTA (−2.8 ppm) and Fe2+-TF-BAPTA (−18 ppm). The unique ability to detect two different ions using a single imaging probe (TF-BAPTA) is clearly reflected in the two distinctive peaks that were obtained in the iCEST spectra (Figure 4e). These experimental results were further supported by Bloch simulations using a three-pool model (Figure 4e).

Figure 4

Simultaneous detection of multiple metal ions. (a) 19F MRI and the orientation of the tubes in the phantom containing 10 mM of TF-BAPTA and 200 μM of mixed ions, (b) iCEST (Δω = −2.8 ppm), and (c) iCEST (Δω = −18 ppm) overlaid on 19F MRI. (d) Merged iCEST image highlights (orange–purple scale) the shared iCEST contrast voxels shown in panels b and c. (e) 19F iCEST spectra. Circles represent experimental signal; lines represent Bloch simulations (three-pool model).

The capability of detecting 19F probes at sub millimolar concentrations,[32] the high sensitivity of the 19F NMR spectrum Δω values to changes in the chemical environment,[33] together with the frequency being specific of these Δωs for certain metal ions, should inspire further development of novel responsive contrast agents for iCEST MRI. One strategy that allows a local increase of the 19F probe concentration and eliminating the need of systemic administration is to coencapsulate the imaging probe with the transplanted target cells.[39] For example, such an approach may be used for the detection of transplanted β cells that release Zn2+ upon the release of insulin.[12] By adding 19F atoms to the two 6-positions of 5F-BAPTA (which previously allowed the detection of only Ca2+ using iCEST[29,34]) it became possible to detect both Zn2+ and Fe2+. Adding one 19F atom to the BAPTA backbone dramatically changes the binding properties of TF-BAPTA.[35] At the same time, the added 19F atom induces kex values that allow the detection of Zn2+ and Fe2+ with 19F iCEST MRI. Although other 1H MRI probe can be used to detect Zn2+[12,36−38] with a potential higher sensitivity as compared to 19F probes, the specificity of iCEST to simultaneously detect different coexisting ions using the same sensor represents a new concept for the rational design of novel MRI probes. While BAPTA derivatives are widely used for the fluorescent detection of metal homeostasis in vitro, the possibility to probe metals in vivo noninvasively with MRI would have profound implications for the biological sciences.