Metal Ion Sensing Using ion Chemical Exchange Saturation Transfer (iCEST)-¹⁹F MRI

Abstract

Although metal ions are involved in a myriad of biological processes, a non-invasive means of detecting free metal ions in a deep tissue remains a formidable challenge. We present an approach for specifically sensing the presence of Ca²⁺ in which the amplification strategy of chemical exchange saturation transfer (CEST) is combined with the broad range in chemical shifts found in ¹⁹F NMR to obtain MR images of Ca²⁺. We exploit the chemical shift change (Δω) of ¹⁹F upon binding of Ca²⁺ to the difluoro derivative of [1,2,-bis(o-aminophenoxy) ethane-N,N,-N′,N′, tetra-acetic acid] (5F-BAPTA), by RF labeling at the bound-19F frequency, ω_{[Ca-5F-BAPTA]}, and detecting the label transfer to the free-¹⁹F frequency, ω_5F-BAPTA. Through the substrate binding kinetics we were able to amplify the signal of Ca²⁺ onto free 5F-BAPTA and thus indirectly detect low Ca²⁺ concentrations with high sensitivity.

Metal ions play a crucial role in a myriad of biological processes, and the ability to monitor real-time changes in metal ion levels is essential for understanding a variety of physiological events. Ca²⁺ has garnered interest due to its involvement in many cellular functions and signaling pathways.¹ Currently, imaging dynamic changes in Ca²⁺ levels is restricted to fluorescence-based methodologies,^2,3 which are limited by low tissue penetration and therefore restrict in vivo Ca²⁺ imaging in deep tissues. Recent advances in the field of molecular magnetic resonance imaging (MRI) has lead to the development of new strategies in the design and synthesis of responsive contrast agents for detecting biologically relevant metal ions. Lanthanide-based complexes^4–7 and modified superparamagnetic iron oxide^8,9 nanoparticles have been developed for Ca²⁺ sensing using MRI. 1,2-bis(o-aminophenoxy) ethane-N,N,N′,N′-tetraacetic acid (BAPTA), was proposed by Tsien¹⁰ as a Ca²⁺ indicator, and later, its difluoro-derivative, 5F-BAPTA, showed large ¹⁹F NMR chemical shifts upon chelating divalent cations.¹¹ The high selectivity of the binding of 5F-BAPTA to Ca²⁺ compared to Mg²⁺, and the high resolution in the ¹⁹F-NMR spectra have been exploited for intracellular Ca²⁺ detection in vitro and in vivo.^11–13 However, MR spectroscopy (MRS)-based approaches rely on observation of the ¹⁹F resonance of the Ca-5F-BAPTA complex for Ca²⁺ detection resulting in limited spatial resolution due to sensitivity considerations. One alternative, suggested by Kuchel and co-workers,¹⁴ is the possibility to transfer magnetization between bound Ca²⁺ and free 5F-BAPTA during NMR experiments.

Chemical exchange saturation transfer (CEST) is a widely used MRI contrast mechanism in which a dynamic exchange process between radiofrequency labeled protons and bulk water is exploited for contrast enhancement, and has been used for many applications in molecular and cellular MRI.^15–22 We employ a saturation transfer approach that couples ¹⁹F- and CEST-MRI for sensing the presence of Ca²⁺ or Mg²⁺ through their substrate binding kinetics, which we have termed ion CEST (iCEST). Using RF labeling at the bound ion [Ca-5F-BAPTA] ¹⁹F frequency and detection of label transfer to the free 5F-BAPTA ¹⁹F frequency (0 ppm), we are able to amplify the signal of bound Ca²⁺ by a factor of 100. We demonstrate that the resulting Z-spectra display supreme sensitivity to bound Ca²⁺ over other M²⁺ cations.

Figure 1a illustrates the dynamic exchange process between free 5F-BAPTA and its complex with M²⁺, [M²⁺-5F-BAPTA]. Upon M²⁺ binding, there is a ¹⁹F chemical shift change (Δω) for 5F-BAPTA. If the exchange rate (k_ex) between M²⁺-bound and free 5F-BAPTA is fast on the NMR time scale (Δω≪k_ex), no peak can be resolved as is demonstrated in Figure 1b for Mg²⁺.

Figure 1.

M²⁺ binding 5F-BAPTA. a) Schematic depiction of the dynamic exchange process between free 5F-BAPTA and bound [M²⁺-5F-BAPTA]. b) ¹⁹F NMR spectra (470 MHz) of 5F-BAPTA in the presence of Mg²⁺ (orange), Zn²⁺ (green), or Ca²⁺ (blue).

When the k_ex is sufficiently slow at the field strength used, a well-defined peak is observed for the [M²⁺-5F-BAPTA] resonance as is shown for Zn²⁺ (green, Δω≫k_ex) and Ca²⁺ (blue, Δω>k_ex). As was previously reported, the observed Δω’s are typical and unique for each ion that is complexed by 5F-BAPTA and ranges from a few ppm in the cases of Ca²⁺, Zn²⁺, Ba²⁺, Sr²⁺, Cd²⁺, Pb²⁺ and others to tens of ppm upon binding of Fe²⁺, Co²⁺ and Ni^2+.11,23 The dissociation constant (K_d) of [M²⁺-5F-BAPTA] is different for each M²⁺, and as a result so is the k_ex for the process in Figure 1a.^24,38 The Zn²⁺-5F-BAPTA peak (Figure 1b, green, 4.1 ppm) is sharper than that of Ca²⁺-5F-BAPTA (Figure 1b, blue, 6.2 ppm), which is correlated with their reported differences in K_d.^23,38 Note that increasing the temperature from 25°C to 37°C (Figure S3) or the addition of high concentrations of fast exchanging ions such as K⁺ and Mg²⁺ (Figure S4) lead to an upfield shift of the free 5F-BAPTA resonance at the ¹⁹F-NMR spectrum.

The ¹⁹F-iCEST properties of 5F-BAPTA in the presence of Ca²⁺ (slow-to-intermediate k_ex), Zn²⁺ (very slow k_ex) and Mg²⁺ (fast k_ex) were determined on a 16.4 T MRI scanner and are summarized in Figure 2 for two different pH values, i.e. 7.2 (Figure 2a–c) and 6.4 (Figure 2d–f). A pronounce saturation transfer contrast was detected in the Ca²⁺ containing solutions (Figure 2a,d) but not in the Zn²⁺ or Mg²⁺ containing solutions (Figure 2b,e or Figure 2c,f, respectively). Importantly, a broad asymmetry is observed at very high fractional Mg²⁺ concentrations (Figure S5b, χ_{(5F-BAPTA/Mg)}=50:1), which peaks at ~1.8 ppm, a frequency much lower than Ca²⁺ (Figure S5a). For faster ion exchange processes between free 5F-BAPTA and bound M²⁺-5F-BAPTA, such as the exchange observed for Mg²⁺ (Figure S5b), other CEST imaging methods, such as frequency-labeled exchange (FLEX), may be considered to improve the detection of these ions.^36–37 Interestingly, the Δω between [Ca-5F-BAPTA] and free 5F-BAPTA was found to be dependent on pH (Figures 2, 3, S1, S2 and S6 and Table S1), but the k_ex between [Ca-5F-BAPTA] and 5F-BAPTA was preserved for all examined pH values as determined by Bloch simulations (190±10 s⁻¹, Figures 2 and S1).²⁵

Figure 2.

iCEST characteristics: ¹⁹F-iCEST Z-spectra of solutions containing 10 mM of 5F-BAPTA and 50 μM of M²⁺ (M²⁺: Ca²⁺, blue; Zn²⁺, green; or Mg2+, orange) in 40 mM Hepes buffer with the pH of the solutions adjusted to 7.2 (a–c) or 6.4 (d–f). Dots represent the raw experimental data. For Ca²⁺, lines represent Bloch simulations (two pool model) and arrows point to the frequency of the [Ca²⁺-5F-BAPTA] complex.

Figure 3.

Imaging Ca²⁺ with iCEST. ¹H-MRI, ¹⁹F-MRI, and iCEST (Δω=6.2 or 5.0 ppm) of M²⁺ solutions with pH values of 7.2 or 6.4. Each tube contains 10 mM of 5F-BAPTA and 50 μM of M²⁺. Small water tubes (shown on ¹H-MRI) were included to determine the orientation of the samples.

These results are in a good agreement with a previous report showing that the binding of Ca²⁺ was unaffected at pH 6–8¹¹ using ¹⁹F-MRS. ¹⁹F-NMR spectra collected with an internal reference revealed that upon pH change, the frequency of the free 5F-BAPTA shifts but not the frequency of bound M²⁺-5F-BAPTA (Figure S2). The T₂ values of 5F-BAPTA are also sensitive to pH as can be seen by the broadening in the Z-spectra (Figures 2, S1 and Table S1). The T₂-value changes seem to be dependent on 5F-BAPTA protonation and not k_ex-dependent based on the observation that the same Z-spectra line widths were found for solutions containing Mg²⁺ (Δω≪k_ex) and Zn²⁺ (Δω≫k_ex). Figure 3 shows MR images of the samples that have been used in this study, i.e., 10 mM of 5F-BAPTA and 50 μM of M²⁺. As expected no difference in MR contrast was observed between the samples when using conventional ¹H-MRI or ¹⁹F-MRI.

However, contrary to the Mg²⁺- or Zn²⁺-containing samples, which did not generate iCEST contrast at this concentration, a large iCEST contrast was detected for the Ca²⁺ containing sample when a saturation pulse (B₁=3.6 μT/2000 ms) was applied at the appropriate frequency offset of the [Ca²⁺-5F-BAPTA] complex, i.e., Δω=6.2 ppm (pH=7.2) and Δω=5.0 ppm (pH=6.4). Figure S6 shows the dependence of Δω on pH, with Δω ranging from 2.1 ppm to 6.2 ppm for pH values of 5.6 to 7.2. In addition, iCEST images were acquired for solutions containing mixtures of Ca²⁺ and Mg²⁺ (50 μM Ca²⁺, 200 μM Mg²⁺) and Ca²⁺ and Zn²⁺ (50μM Ca²⁺, 50 μM Zn²⁺) with 10 mM BAPTA at pH 7.2. The iCEST contrast produced by the Ca²⁺ was still significant (~22%) at Δω=6 ppm for all mixtures (Figure S5). Although high Mg²⁺ concentrations generate iCEST contrast at Δω=1.8 ppm (Figure S5a–b) the larger Δω, smaller k_ex of [Ca-5F-BAPTA] and its much higher iCEST contrast makes this approach better for Ca²⁺ sensing (Figure S5b, amplification factor = ×10 for Mg²⁺, ×100 for Ca²⁺).

To evaluate the sensitivity of our suggested approach we examined the iCEST contrast at different ratios of Ca²⁺ to 5F-BAPTA (χ_Ca, Figures 4a and S7). As clearly shown in Figure S7, Ca²⁺ is easily detected with iCEST MRI at χ_Ca=1:1000, where ~11% contrast is observed in the Z-spectrum for this phantom. The same amplification was obtained when 0.5 mM 5F-BAPTA was used to detect 500 nM Ca²⁺ (Figure 4b), showing the potential of iCEST to sense low Ca²⁺ concentrations.

Figure 4.

Ca²⁺ sensing using iCEST. a) χ_Ca vs. MTR plot. b) Detection of 500 nM Ca²⁺ in the presence of 0.5 mM of 5F-BAPTA. Inset depicts ¹⁹F-MRI of the sample with an overlaid iCEST image. Lines represent Bloch simulations. Error bars represent the inter-voxel standard deviations.

In this study, we show for the first time that spatial information of Ca²⁺ and Mg²⁺ levels can be obtained using amplification of the sensitivity by iCEST with 5F-BAPTA as the ion indicator. One advantage of using 5F-BAPTA as an MRI responsive agent for detecting metal ions over ¹H- MRI²⁶ or ¹²⁹Xe-MRI²⁷ based probes is that no attachment of a contrast enhancer is required. The ¹⁹F atoms serve on the chelates as the responsive group as well as contrast generators. Hyperpolarized ¹²⁹Xe-CEST (hyperCEST)^28–30 was the first example of non-¹H CEST-MR imaging, although on a gas bubbled into solution instead of solute such as BAPTA. Earlier heteronuclear NMR experiments using magnetization transfer protocols have allowed detection of exchange between two pools of nuclear spins in MRS studies.^14,31,32

Our study shows the potential of exploiting the iCEST concept using ¹⁹F-MRI, as 1:2000 concentration ratios are amplified to 1:20 changes in ¹⁹F signal (Figures 4a and S7), i.e., an amplification factor of ~100 for a k_ex of 190 s⁻¹. In addition to the advantages of using ¹⁹F-MRI, i.e., high γ, 100% natural isotopic abundance, and negligible amount of ¹⁹F in soft tissues, ^33,34,39 the large range of ¹⁹F chemical shifts (about 20 times that of ¹H³⁵) and the sensitivity of ¹⁹F Δω to the details of the local environment is an ad- vantage for iCEST-based applications. One obstacle of the iCEST approach would be the detectability level of the free ¹⁹F-agent. This could be surmounted by collecting high-resolution ¹H-MR images, which provide spatial information, and reducing the resolution for iCEST to allow localized detectability of the ¹⁹F-based agent with improved SNR³⁹ (see SI for detectability discussion). Using paramagnetic ¹H CEST probes⁷ for detecting Ca²⁺ will allow better spatial resolution and higher SNR compared to iCEST, but also will have a worse sensitivity for detecting low Ca²⁺ concentrations. In the iCEST approach, the signal from low concentration solutes [Ca²⁺-5F-BAPTA] is amplified through saturation transfer onto the signal of the high concentration free 5F-BAPTA. Since this contrast is dependent on χ_Ca, lower concentrations of Ca²⁺ can be detected through simply reducing the free 5F-BAPTA concentrations, when these concentrations are NMR detect-able.³⁹ This is an advantage of the iCEST approach, since this feature is not available for ¹H CEST, which is based on water. Finally, the unique Δω found for each [M²⁺-5F-BAPTA] and the diversity of the obtained k_ex may be exploited for multi-ion MR imaging approaches in which each ion generates iCEST contrast with an identifiable amplitude and Δω. This concept was shown for different exchangeable protons in ¹H-CEST and termed multi-color imaging.¹⁷

In conclusion, we have developed a new approach for sensing metal ions with spatial information using MRI, in which the amplification strategy of CEST is combined with the Δω specificity in ¹⁹F frequency. The outlined principles can be further extended for designing new iCEST agents to detect other ions.