Background
Medical imaging can provide morphological, structural, metabolic and functional information of tumors and is an essential part of cancer clinical protocols. In order to accurately detect and characterize tumors, exogenous contrast agents are often used. Positron emission tomography (PET) imaging of
18fluoro-2-deoxy-glucose (
18FDG), an analogue of glucose, is most widely used to characterize cancers based on the high uptake rate and glycolytic activity of tumors compared to healthy tissue [
1‐
5]. Although
18FDG-PET provides valuable functional and metabolic information of cancers, it may be limited by the lack of specificity of
18FDG uptake and the patients exposure to ionizing radiation [
6].
Magnetic resonance imaging (MRI) coupled with administration of gadolinium based contrast agents (GBCAs) provides exquisite contrast between normal and tumorous tissues without exposure to ionizing radiation and helps with clinical decision-making [
7]. However, recent studies have reported the deposition of GBCAs in brain and bone matrix found by MRI and mass spectrometry [
8‐
13]. Further studies are required to evaluate the long-term effects of gadolinium detected in the brain tissues or other organs on normal functioning of the organs. This provides an impetus to explore new MRI contrast agents that are non-toxic and non-metabolized. Ideal contrast agents would also be inexpensive.
Magnetic resonance imaging based on the chemical exchange saturation transfer (CEST) effect has gained widespread attention for its ability to image certain metabolites indirectly at high resolution [
14‐
19]. In CEST, a long, frequency-selective radiofrequency pulse saturates the labile protons of a metabolite solute. The exchange of the saturated magnetization of the solute with the bulk water protons leads to a reduction in the bulk water signal compared to the signal without saturation [
20‐
23]. The CEST method has been shown to provide higher sensitivity than direct observation with traditional proton MR spectroscopy (MRS) and was applied in monitoring the changes in metabolite and macromolecular levels in various human diseases [
15,
24‐
28]. Recently, Glucose and its analogues have been used as CEST contrast agents (glucoCEST) to image cancers in animal models [
17,
29‐
31] and human cancer patients [
32]. However, there are a couple of confounding factors in the glucoCEST contrast because of how readily glucose is metabolized, and care must be taken with regards to differences in its metabolism in healthy and tumorous tissues, glucose’s metabolic products, and the accumulation in the extracellular and extravascular. The extent of these contributions to the glucoCEST contrast is currently unknown [
33]. Nonetheless, the CEST arising in the tumor region following the glucose administration has been aptly labeled as glucoCEST.
In this study, we demonstrate the feasibility of using the popular sweetener sucralose (commercial name “Splenda”) as an MRI contrast agent to detect cancer. Sucralose does not metabolize but accumulates into tumor tissue due to the enhanced permeability and retention effect, and it exhibits CEST contrast through its labile hydroxyl protons. We termed this new method as ‘sucCEST’. The concentration and pH dependence of sucCEST contrast was measured in vitro in solution phantoms, and the sucCEST contrast was evaluated in a rat brain gliosarcoma model and compared with the gadolinium-diethylenetriamine pentaacetic acid (Gd-DTPA) contrast enhanced image. Finally, the application of sucCEST in cancer and other pathological conditions in humans is discussed.
Methods
Phantom preparations
For high-resolution 1H NMR spectra, 200 mM of sucralose (Sigma Aldrich, USA) solution was prepared in phosphate buffered saline (PBS) at pH 7. For imaging, phantoms were prepared in PBS and experiments were performed at 37 °C. To measure the pH dependence of sucCEST, phantoms with 10 mM sucralose concentration in PBS were prepared at a varying pH from 6.6 to 7.4 in step of 0.2 pH unit. The pH was adjusted using 1 N NaOH/HCl. For measuring concentration dependence of sucCEST contrast, phantoms with 2, 4, 6, 8, and 10 mM concentrations of sucralose were prepared in PBS at pH 7.
To obtain the SplendaCEST, we purchased Splenda from local market and prepared 0.1, 0.3 and 0.5% of Splenda solutions at pH 7.
Phantom imaging
High-resolution 1H NMR phantom experiments from 200 mM sucralose solution were performed on a vertical bore Bruker Avance DMX 400 MHz spectrometer (Bruker Corporation, Germany) equipped with a 5 mm PABBI proton probe using TR = 4 s and 128 averages. The proton MRS spectrum was gathered at different temperatures (5, 15, 25, and 37 °C).
The sucCEST imaging of phantom was performed on a 9.4 T, 30 cm horizontal bore magnet (Agilent, USA) interfaced to a Varian console, with a 20-mm volume coil (M2M Imaging, USA) using a custom-programmed GRE readout pulse sequence with a frequency selective continuous wave preparation pulse for saturation. The sequence parameters were as follows: field of view (FOV) = 20 × 20 mm2, slice thickness = 10 mm, flip angle (FA) = 15°, repetition time (TR) = 6.2 ms, echo time (TE) = 2.9 ms, matrix size = 128 × 128. Saturation was applied every 15 s and immediately followed by 128 segment acquisitions before a long delay to allow for T1 recovery. CEST images were collected using variable saturation lengths (1 through 3 s) and saturation pulse amplitudes (B1rms: 2.35, 3.5, 4.7, 5.9, 7, 8.2, 9.4, 10.6, 11.7 µT). For concentration and pH dependent studies, CEST images were collected using 1 s saturation pulse at B1rms of 7 µT for multiple frequencies (−3.6 to +3.6 ppm in 0.2 ppm steps) from bulk water.
B
0 and B
1 field maps were also gathered and used to correct the CEST contrast map using the methods described previously [
16,
34]. Briefly, CEST data is acquired from −1 to 1 ppm at step size of 0.2 ppm to find the spatial dependence of the frequency of water. This spatially-dependent frequency shift is then used to correct the sucCEST z-spectra using a quadratic polynomial interpolation.
For B
1 correction, two images were obtained using preparation square pulses with duration (τ) and prescribed flip angles of 30° and 60°. The RF pulse amplitude for a 30° flip angle was used as the reference B
1 or B
1ref. B
1 maps were generated by solving the equation:
$$\frac{{{ \cos }(2\phi)}}{{{ \cos }(\phi)}} = \frac{{S(2\phi)}}{{S(\phi)}}$$
where S (
ϕ) and S (2
ϕ) denote pixel signals in an image with preparation flip angle
ϕ and 2
ϕ respectively. From the flip angle map, a B
1 field map can be obtained using the relation, B
1 =
ϕ*(360τ)
−1. B
1 is then corrected assuming a linear dependence of sucCEST contrast.
Rat tumor model preparation
To validate the sucCEST in vivo, 9L-gliosarcoma rat brain tumor model was used. It is well known that brain tumors disrupt the function of blood–brain barrier (BBB) locally in a nonhomogeneous manner [
35]. The compromised BBB and enhanced permeability of tumor vasculature will enable the nonhomogeneous distribution of the injected sucralose in the tumor region.
To develop intracranial tumors, rat gliosarcoma cells (9L) were used. Syngeneic female Fisher rats (F344/NCR, 4–6 weeks old) weighing 130–150 g were used as described previously [
36]. General anesthesia was induced using 2% isoflurane mixed with 1 l/min oxygen followed by 1–2% isoflurane. A 10 µl suspension of 50,000 9L cells in PBS was injected into the cortex at a depth of 3 mm with a Hamilton syringe and a 30-gauge needle using stereotactic apparatus (3 mm lateral and 3 mm posterior to the bregma). 5 weeks after implantation of tumor cells, the rats were subjected to MRI.
Rat MR imaging
Rats (n = 5) with brain tumors were anesthetized with isoflurane (3% for induction, 1.5% maintenance) and a polyethylene catheter (PE50) was inserted into the tail vein for sucralose injection. Rats were transferred to a 9.4 T horizontal bore small animal MR scanner (Varian, Palo Alto, CA) and placed in a 35-mm diameter commercial quadrature proton volume head coil (m2m Imaging Corp., Cleveland, OH). Rats were kept under anesthesia (1.5% isoflurane in 1 l/min oxygen) and their body temperature maintained at 37 °C with the air generated and blowing through a heater (SA Instruments, Inc., Stony Brook, NY). Respiration and body temperature were continuously monitored using an MR compatible small animal monitoring system (SA Instruments, Inc., Stony Brook, NY).
Fast-spin-echo T2 weighted MRI was performed prior to the CEST experiments to determine the slice positioning of glioma. The parameters for T2 weighted imaging were: TR = 8000 ms, TE = 50 ms, FA = 90°, echo train length = 16, number of slices = 12, slice thickness = 2 mm, FOV = 30 × 30 mm2, matrix size = 128 × 128 and number of averages (NA) = 2. Following the whole brain T2 weighted brain imaging, a single slice 3 mm thick containing the tumorous region was acquired. This led to an in-plane resolution of 0.234 × 0.234 mm2. This same slice was used for all the subsequent sucCEST and Gd-DTPA experiments.
Chemical exchange saturation transfer imaging of rat brain tumor was performed using similar pulse sequence parameters as the phantom imaging experiments except FOV 30 × 30 mm2 B1rms = 2.35 µT, saturation duration = 2 s and T1 delay = 8 s, slice thickness = 3 mm, NA = 4. After baseline imaging, the rats were injected with 2 ml of 200 mM sucralose solution at a rate of 0.2 ml/min through a catheter inserted in a tail vein (for 10 min). Following sucralose administration, CEST imaging was performed every 30 min.
Gadolinium weighted imaging
Following the CEST imaging, a baseline T1-weighted image was acquired using the following parameters: FOV 30 × 30 mm2, TR = 6.22 ms, TE = 2.9 ms, FA = 20°, slice thickness = 3 mm, and NA = 12. Gd-DTPA (100 µl, 287 mg/ml) was injected as a bolus in 5 s through tail vein and another T1-weighted image was acquired to see Gd enhanced signal in the glioma.
CEST image processing
First the acquired CEST weighted images were corrected for the B
0 inhomogeneity and used to generate sucCEST [magnetization transfer ratio asymmetry, (
MTR
asym)] contrast map using Eq. [
1].
$$MTR_{asym} \;(\% ) = 100 \times \left( {\frac{{S_{ - ve} - S_{ + ve} }}{{S_{0} }}} \right)$$
(1)
where
S
−ve, S
+ve, and
S
0 are the B
0 corrected MR signals at −1, 1 and 20 ppm, respectively. The CEST contrast map was further corrected for B
1 inhomogeneity and overlayed onto anatomical proton image as false colors. Regions of interests (ROIs) were manually drawn on tumor and normal appearing brain regions. All image processing and data analysis were performed using software routines written in MATLAB (R2015b) as described in details elsewhere [
16,
24].
Discussion
In the current study, we evaluated the use of the inexpensive, non-caloric sweetener sucralose as an MRI contrast agent based on chemical exchange to image tumors in vivo. In the normal rat brain, no change in the sucCEST contrast was observed following intravenous injection of sucralose, suggesting that sucralose does not cross the BBB and therefore can be used to image BBB disruption. Increased sucCEST contrast was observed in the tumor region, which is presumably due to the accumulation of sucralose in the extravascular extracellular space (EES) of the tumor. The brain tumor compromises the BBB and allows sucralose to enter the tumor EES. SucCEST sensitivity of 1.1% per mM sucralose translates to a ~1000-fold higher sensitivity than the direct detection with MRS, enabling the detection of millimolar concentrations.
Recently,
d-glucose has been used as a CEST MRI contrast agent to image cancers (glucoCEST) [
17,
29]. However, the interpretation of glucoCEST results might be intricate as the
d-glucose is readily metabolized by both tumors and healthy tissue. Glucose analogues such as 2-deoxy-
d-glucose (2-DG) and 2-fluoro-deoxy-
d-glucose (FDG) were also shown to have higher CEST effect compared to glucose [
31,
38]. This may be due to rapid conversion of glucose into lactate by the tumors [
39] whereas 2-DG and FDG are not metabolized. As tumors are highly glycolytic, the injected glucose or pyruvate rapidly metabolize into lactate. Using the recently developed LATEST method to measure CEST contrast from lactate [
18], it may be possible to map the glycolytic behavior of tumors as well as probe the kinetics of LDH activity in tumor. Another glucose derivative, 3-
O-methyl-glucose (3-OMG) has recently been used as an MRI contrast agent to image cancer in orthotopic xenograft of a mammary adenocarcinoma model [
40]. 3-OMG is taken up rapidly and preferentially by the tumor cells and stored. This contrasts with 2-DG and FDG, which undergo phosphorylation. Studies have shown that 3-OMG diffuses into normal brain tissue [
30], though, limiting its use in the brain tumor imaging.
Unlike other CEST methods, sucCEST selectively highlighted the tumor. As sucralose is not metabolized in the body, the tumor sucCEST kinetics may be governed by the wash-in/wash-out of sucralose. Although we demonstrated the sucCEST in the brain tumor model, the method can potentially be useful to image other types of tumors and to monitor anti-tumor drug efficacy.
Sucralose phantom studies showed the highest CEST effect for saturation parameters of 7 µT and 3 s duration in vitro. However, in vivo T2 of water in the brain is much shorter (~40 ms) [
41]. Short T2 leads to a large direct saturation effect when using relatively large B1 amplitudes for saturation and can obscure the desired CEST contrast [
42]. Hence, we optimized the saturation B1 amplitude and duration separately in vivo and found that 2.35 µT and 2 s RF saturation pulse gave reliable 1 ppm MTR
asym in the brain.
The acute and sub chronic toxicity effect of oral administration of sucralose has been evaluated previously in animals, and no sucralose-related adverse effects were observed following the dietary administration of sucralose in mice (16 g/kg), rats (10 g/kg) and dogs (900 mg/kg/day) [
43]. Another study in human volunteers showed no adverse effect of acute or chronic oral dosage of sucralose [
44]. These studies established that sucralose is non-toxic following high acute oral administration. While other studies reported no toxic effect of intravenous administration of sucralose at lower dosage (20 mg/kg) in mice and rats [
45,
46], we are not aware of any published toxicity result at the intravenous administration dosage (500 mg/kg) used in the present study. In this preliminary study, we did not observe any adverse effect of intravenously injected sucralose in normal or tumor-bearing rats. However, immediately following the start of the infusion of sucralose, the respiration rate of the rats was found to increase by ~20 breath/min before getting back to the baseline rate in ~2 to 3 min. For the future studies, the optimum concentration and the rate of injected sucralose may be explored.
We suggest that sucCEST with intravenous infusion of sucralose can serve as a diagnostic and therapeutic response monitoring tool in preclinical studies of tumors. While it is possible to use this method to study cancer patients at ultrahigh fields, more detailed toxicity studies of intravenously administered sucralose are required before undertaking such studies.
Authors’ contributions
PB, MH contributed to conception and design of the study, performed experiments, analyzed data and wrote the manuscript. FM, NEW and MDS contributed to manuscript writing and editing, HH provided the technical support and helped with manuscript editing. RR provided conception and overall experimental design and contributed to manuscript writing and editing. All authors read and approved the final manuscript.