|Home | About | Journals | Submit | Contact Us | Français|
Samuel I. Stupp: 0000-0002-5491-7442
Thomas J. Meade: 0000-0001-6202-1155
Misregulation of extracellular Ca2+ can indicate bone-related pathologies. New, noninvasive tools are required to image Ca2+ fluxes and fluorine magnetic resonance imaging (19F-MRI) is uniquely suited to this challenge. Here, we present three, highly fluorinated peptide amphiphiles that self-assemble into nanoribbons in buffered saline and demonstrate these nanostructures can be programmed to change 19F-NMR signal intensity as a function of Ca2+ concentration. We determined these nanostructures show significant reduction in 19F-NMR signal as nanoribbon width increases in response to Ca2+, corresponding to 19F-MR image intensity reduction. Thus, these peptide amphiphiles can be used to quantitatively image biologically relevant Ca2+ concentrations.
Abnormal calcium concentrations have been associated with a variety of pathologies, including Paget's disease and osteoporosis.1 Osteoporosis is characterized by pathological reduction in bone density and greatly increases risk of severe bone fractures and hospitalizations.2 Current methods of measuring bone density use ionizing radiation and can only identify advanced stages of osteoporosis.3 Ca2+ fluctuations from 2–10 mM are linked to bone resorption.4 Direct observation is difficult and new tools are needed to image Ca2+ flux in real time.
Magnetic resonance imaging (MRI) allows noninvasive measurement of biological phenomena without the use of ionizing radiation and produces excellent anatomical contrast with high spatial resolution in soft tissues.5 Typically, paramagnetic MR probes produce signal indirectly via 1H signal enhancement of endogenous water by accelerating magnetic 1H relaxation.6 Despite advances in probe design, relaxation-based MR probes are limited by native tissue background.7 We have therefore been interested in investigating alternative methods to develop directly detected MR probes. For example, the absence of fluorine-containing organic molecules in biological systems has encouraged the development of 19F-probes. 19F offers zero-background signal and thus has the potential for quantitative imaging.8–10 While 19F-MRI has low sensitivity on a per-atom basis (~0.5 mM),7,11 the signal can be enhanced by a high local concentration of magnetically equivalent fluorine atoms. This has been accomplished with the use of emulsions or micelles.7,12–15
The tendency of highly fluorinated molecules to aggregate due to their hydrophobicity can be exploited to create biologically responsive agents. Aggregated organofluorine moieties with slow molecular dynamics on the MR time scale (~0.1 s) have been reported to be MR silent due to rapid T2 relaxation.16 The tunable properties of self-assembling systems can be used to design fluorinated agents that can switch between different aggregation states with distinct MR signals in response to a stimulus, such as enzymatic cleavage or pH.17,18 Generally, these responsive probes are designed to switch between assembled and unassembled states.16–20 In this work, we have used self-assembling fluorinated peptide amphiphiles (F-PAs) to create nanoscale agents that can switch between “light” and “dark” states without completely disrupting assemblies.
PAs that self-assemble into high aspect ratio nanofibers were first reported by Stupp et al.21 These molecules contain short peptide sequences that include a domain with the propensity to form β-sheets, charged residues, and a single hydrophobic moiety covalently grafted to one terminus. These sequences can be programmed to exhibit a diverse variety of one-dimensional morphologies and dimensions.20–22 Additionally, these bio-compatible PAs have been shown to be highly functional in imaging applications.23–25 Designing PA molecular imaging probes with variable morphology shows promise for sensing biochemical events. Previously, we reported on a series of F-PAs that self-assembled into various morphologies.26 Some of these agents exhibit changes in morphology, but not necessarily secondary structure, and consequently 19F-MR signal as a function of pH.23 In this work, we examine the self-assembly behavior and the 19F-MR response of three F-PAs to titration with Ca2+ with the long-term objective of sensing Ca2+ in vivo. A series of three F-PAs bearing the β-sheet-forming sequence V2A2 were synthesized incorporating K2, E2, or E3 charged groups for solubility (Figure 1).
The V2A2 sequence is known to promote β-sheet formation and the assembly of 1D nanostructures.27 Glutamic acid was chosen for its ability to chelate Ca2+ and cross-link negatively charged peptide nanofibers.28 A positively charged, lysine bearing sequence was selected as a control without affinity for Ca2+. The peptides were prepared by solid-phase peptide synthesis and tridecafluoroheptanoyl chloride was condensed to the terminal amine of each peptide.26 The resulting fluorinated peptide amphiphiles (F-PAs) were cleaved from the resin and purified by reverse-phase high-pressure liquid chromatography (HPLC). Purity was assessed by analytical HPLC-MS (Figures S1 and S2). Unless otherwise indicated, solutions and gels were prepared for analysis by dissolving each F-PA at 2 mM concentration in 100 mM NaCl and 30 mM Tris buffer adjusted to pH 7.4. The 2 mM peptide concentration had previously been optimized for similar F-PAs, such systems are employed as implants and persist in vivo over long periods without diminishing signal.26,29 Solutions were allowed to equilibrate for at least 12 h.29 CaCl2 was added from concentrated stock solutions to minimize changes in the total solution volume during titrations. In the case of NMR experiments, 10% D2O was included as the lock solvent and trifluoroethanol (TFE) at 3 mM was used as an internal standard. Samples were allowed to equilibrate for at least 5 min between CaCl2 additions.
Characterization of F-PA supramolecular structures by cryogenic transmission electron microscopy (Cryo-TEM) revealed that the morphology was dependent on Ca2+ (Figure 2) in negatively charged C7E2 and C7E3, but not in positively charged C7K2. Persistence length is a comparative metric of fiber rigidity in soft matter assemblies such as F-PAs.30 Notably, C7E2 formed fibers with partial ribbon-like character before the addition of CaCl2. C7E3 formed wider, more rigid nanoribbons than C7E2, likely arising from intermolecular interactions from the additional glutamic acid. At 6 mM CaCl2, it was observed that flat, ribbon nanostructures were the dominant morphology while at 30 mM Ca2+ the solution became a turbid gel. Cryo-TEM indicated that the gelled nanoribbons did not have a regular morphology. When compared across the same concentration range, C7E3 did not exhibit the same low-Ca2+ transition as C7E2, but maintained ribbon morphology at all CaCl2 concentrations tested. In general, higher concentrations of Ca2+ produced shorter, more bundled ribbons (Figure 2). C7K2 morphology did not change with CaCl2 addition (Figure S4).
Variation in fiber width was quantified for each condition based on Cryo-TEM measurements (Figure 3). The width of C7E2 ribbons increased significantly (p < 0.001) upon the addition of CaCl2, from 16 ± 4 nm to 22 ± 5 nm with at least 100 independent measurements. This increase in width corresponds to the same calcium window where NMR signal decreases substantially (Figure 4B). A small effect was observed for C7E3, where ribbon width increased slightly from 39 ± 10 nm to 45 ± 9 nm (p < 0.001). At 30 mM, further changes in the width did not occur in C7E3 nanostructures, and the helicity of C7E2 became irregular, making an accurate measurement by TEM challenging. Structural evidence for β-sheet character in all F-PAs is provided by solid-state FT-IR spectra featuring a carbonyl band at 1630 cm−1 (Figure S5). Furthermore, we observed changes in the circular dichroism spectra upon the addition of Ca2+ in the C7E2 and C7E3 sequences (Figure S6). These changes upon calcium binding are consistent with our recent observation by Overhauser dynamic nuclear polarization relaxometry that the water dynamics around amino acids of peptide amphiphiles change significantly upon the addition of calcium.31
Nuclear magnetic resonance spectroscopy (NMR) and magnetic resonance imaging (MRI) measure the same physical phenomenon. Because NMR is typically orders of magnitude more sensitive, we first quantified the response of our probes to Ca2+ using NMR spectroscopy. 1D 19F-NMR spectra for each amphiphile were obtained using an Agilent DD2 500 MHz spectrometer with HFX probe tuned to 470 mHz, and collected over a range of concentrations. Figure 4A follows the trifluoromethyl peak of the perfluorinated heptanoyl tail and integration of the perfluoromethyl group (at about −81 ppm) was quantified as a measurement of T2 quenching (Figure 4B).
It was observed that C7E3 showed an NMR response to Ca2+ only at high concentrations (>10 mM). Additionally, the slope of the C7E3 signal as a function of Ca2+ concentration was much less negative than the corresponding slope for C7E2 (Figure 4). C7E2 displayed a sharp, order of magnitude signal change, as the nanostructures transition from narrow to wider nanoribbons in response to Ca2+ over the physiologically relevant range (2 to 6 mM). This range corresponds to where the C7E2 nanostructure increases in ribbon width and rigidity. No change in C7K2 signal integration was observed as a function of Ca2+ (Figure S7), indicating that Ca2+ chelation by glutamic acid residues is responsible for the observed effects in C7E2 and C7E3. The NMR signal of each F-PA appears to be stable over time when incubated with Ca2+ and analyzed over 24 h (Figure S10).
When CaCl2 is titrated into a solution of F-PA and centrifuged, the Ca2+ concentrations in the supernatant do not vary for C7K2, but are significantly reduced in C7E2 and C7E3 as quantified by inductively coupled plasma-optical emission spectroscopy (ICP-OES). This supports Ca2+ chelation by C7E2 and C7E3 but not C7K2 (Figure S11). Interestingly, when F-PA is titrated with MgCl2, no change in NMR signal is observed by any of the F-PAs, suggesting selectivity for Ca2+ versus Mg2+ (Figures S11 and S12). This supports previous studies that found Ca2+ binds cooperatively, rather than relying on purely electrostatic interactions.32
19F-MR imaging of the F-PAs was performed using a 19F surface coil at 9.4 T. A steady-state free procession (SSFP) sequence was utilized, including a sinc pulse (1.5 ms in length) as an excitation pulse to excite the desired 19F signal (ca. −76 to −80 ppm).19,20 An internal standard of TFE (3 mM) was added to enable signal comparison between different scans. Two mM C7E2 and C7E3 were each imaged after the addition of 0, 1.7, 3.4, 5.1, and 25.5 mM CaCl2 (Figure 5).
The results of these imaging experiments paralleled the NMR results in Figure 4. It should be noted that voxel size was increased in the case of 19F for detection. Imaging parameters and phantom preparation are described in the Supporting Information. Specifically, the 19F-MR signal of C7E2 was observed to fall off rapidly to background signal levels by 5.1 mM Ca2+. The C7E3 signal did not visibly change until Ca2+ concentrations increased to a much higher value (25.5 mM). These results verify that the response in C7E2 signal over the physiological Ca2+ range (2–6 mM) can be detected a MR imaging experiment. The MR signals of each sample were normalized using the external TFE sample and are compared in Figure S13. Signal decay upon Ca2+ addition is analogous to the responses observed (Figure 4), where a precipitous change in MR signal occurs with ~5 mM Ca2+ for the C7E2 sequence and signal is retained to high concentrations of Ca2+ for the C7E3 sequence.
New supramolecular nanostructures have been reported that are capable of measuring calcium ion concentrations using 19F-MRI. These nanostructures are formed by fluorinated peptide amphiphiles that undergo width changes in response to Ca2+ and thereby modulate the intensity of their fluorine signals. As the width of the F-PA nanostructure increases, the 19F-MR signal decreases. In particular, C7E2 nanostructures were shown to respond strongly to Ca2+ between 2 and 6 mM, a range encompassing physiological extracellular calcium concentration fluctuations. We hypothesize this difference in Ca2+ response arises from the nanostructural differences between the F-PAs, namely small nanostructure width results in higher Ca2+ response. In the case of C7E2 the transition from a thin ribbon to a wide ribbon is highly correlated with a decrease in the 19F-MR signal. The C7E2 nanostructures were also found to exhibit excellent Ca2+ sensitivity, which disappears by the addition of an additional glutamic acid residue to the peptide sequence in C7E3. We believe these results demonstrate how subtle changes in PA chemical structure have a significant impact on the intensity of 19F-NMR signals, thus making the nanostructures useful as 19F-MR imaging agents.
We gratefully acknowledge funding from NIH NHLBI: P01HL10879, NIH NICDR: 5R01DE015920, and NIH NIBIB: R01EB005866. This research was also supported by a Catalyst Award from the Louis A. Simpson and Kimberly K. Querrey Center for Regenerative Nanomedicine at Northwestern University. Compound purification and characterization was performed at Northwestern's SQI Peptide Synthesis Core, Integrated Molecular Structure Education and Research Center (IMSERC), Bioimaging Facility (BIF) and Keck Biophysics Facility. The U.S. Army Research Office, the U.S. Army Medical Research and Materiel Command, and Northwestern University provided funding to develop the Peptide Synthesis facility and ongoing support is being received from the Soft and Hybrid Nanotechnology Experimental (SHyNE) Resource (NSF ECCS-1542205). We also thank Prof. Liam Palmer helpful discussions and Mark Seniw for graphics. F-MRI was conducted at the Advanced Imaging Research Center, UT-Southwestern.
Supporting Information: The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.7b07828.
Detailed synthetic procedures, purification, and NMR characterization of PAs, as well as MRI parameters (PDF)
Notes: The authors declare no competing financial interest.