Imaging the DNA Alkylator Melphalan by CEST MRI: An Advanced

Jul 11, 2016 - Ultimately, this method could be used to noninvasively monitor melphalan delivery to brain tumors and also to validate therapeutic resp...
3 downloads 5 Views 3MB Size
Download PDF
Article pubs.acs.org/molecularpharmaceutics

Imaging the DNA Alkylator Melphalan by CEST MRI: An Advanced Approach to Theranostics Ethel J. Ngen,†,‡ Amnon Bar-Shir,†,‡ Anna Jablonska,†,‡ Guanshu Liu,†,§ Xiaolei Song,†,§ Roxana Ansari, Jeff W. M. Bulte,†,‡,§ Miroslaw Janowski,†,‡,∥,⊥ Monica Pearl,#,¥ Piotr Walczak,*,†,‡,¶ and Assaf A. Gilad*,†,‡,§ †

Division of Magnetic Resonance Research, Russell H. Morgan Department of Radiology and Radiological Science, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, United States ‡ Cellular Imaging Section and Vascular Biology Program, Institute for Cellular Engineering, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, United States § F. M. Kirby Research Center for Functional Brain Imaging, Kennedy Krieger Research Institute, Baltimore, Maryland 21205, United States ∥ NeuroRepair Department, Mossakowski Medical Research Centre, PAS, 02106 Warsaw, Poland ⊥ Department of Neurosurgery, Mossakowski Medical Research Centre, PAS, 02106 Warsaw, Poland # Division of Interventional Neuroradiology, Russell H. Morgan Department of Radiology and Radiological Sciences, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, United States ¥ Department of Radiology, Children’s National Medical Center, Washington, D.C. 20010, United States ¶ Department of Radiology, Faculty of Medical Sciences, University of Warmia and Mazury, Olsztyn, Poland S Supporting Information *

ABSTRACT: Brain tumors are among the most lethal types of tumors. Therapeutic response variability and failure in patients have been attributed to several factors, including inadequate drug delivery to tumors due to the blood− brain barrier (BBB). Consequently, drug delivery strategies are being developed for the local and targeted delivery of drugs to brain tumors. These drug delivery strategies could benefit from new approaches to monitor the delivery of drugs to tumors. Here, we evaluated the feasibility of imaging 4[bis(2-chloroethyl)amino]-L-phenylalanine (melphalan), a clinically used DNA alkylating agent, using chemical exchange saturation transfer magnetic resonance imaging (CEST MRI), for theranostic applications. We evaluated the physicochemical parameters that affect melphalan’s CEST contrast and demonstrated the feasibility of imaging the unmodified drug by saturating its exchangeable amine protons. Melphalan generated a CEST signal despite its reactivity in an aqueous milieu. The maximum CEST signal was observed at pH 6.2. This CEST contrast trend was then used to monitor therapeutic responses to melphalan in vitro. Upon cell death, the decrease in cellular pH from ∼7.4 to ∼6.4 caused an amplification of the melphalan CEST signal. This is contrary to what has been reported for other CEST contrast agents used for imaging cell death, where a decrease in the cellular pH following cell death results in a decrease in the CEST signal. Ultimately, this method could be used to noninvasively monitor melphalan delivery to brain tumors and also to validate therapeutic responses to melphalan clinically. KEYWORDS: CEST, MRI, melphalan, theranostics, drug imaging, therapeutic response assessment penetration into the tumor.7−11 Currently, several interventional procedures are being developed to improve drug delivery across the BBB and into brain tumors.12−16 We recently developed an MRI-guided neurointerventional procedure for

1. INTRODUCTION Brain tumors are among the most difficult types of cancers to treat.1,2 Standard therapeutic regimens include total resection surgery, followed by radiation therapy and adjuvant chemotherapy.3 However, in cases where the tumors are considered inoperable, chemotherapy is a major component of the treatment plan.4−6 One of the challenges in delivering chemotherapeutic agents to brain tumors is bypassing the blood−brain/tumor barrier (BBB), which restricts drug © XXXX American Chemical Society

Received: February 16, 2016 Revised: June 20, 2016 Accepted: July 11, 2016

A

DOI: 10.1021/acs.molpharmaceut.6b00130 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics

Figure 1. (a) Mechanism of CEST MRI contrast generation by melphalan. Saturated melphalan amine protons are highlighted in red. The biologically active site of melphalan is highlighted in blue (the Cl− leaves for DNA alkylation by the activated carbocation to occur). ksw is the exchange rate constant of the labile melphalan protons with exchangeable protons of endogenous biological molecules represented by bulk water protons.

prodrugs to their active components by CEST MRI.32 Here, we evaluated the feasibility of directly imaging melphalan by CEST MRI without additional chemical modification of the drug. We also evaluated the feasibility of monitoring therapeutic responses to melphalan with CEST MRI. Melphalan is a DNA alkylating agent, used clinically in the intra-arterial treatment of retinoblastomas and currently in phase I clinical trials for the treatment of diffuse pontine gliomas (NCT01688401) and medulloblastomas.33,34 Although brain tumor chemotherapeutic regimens vary, DNA alkylating agents are usually an integral component of the treatment plans.35−38 Melphalan has two exchangeable amine protons that resonate at approximately 2.5 ppm relative to water protons, and which can be saturated with an appropriate radiofrequency pulse for CEST MRI (Figure 1).39 However, DNA alkylating agents such as melphalan are very reactive, and this accounts for their high therapeutic effectiveness (cytotoxicity).38 Melphalan induces its cytotoxic effects by cross-linking DNA strands via alkylation of, mainly, the deoxyguanosine residues.38,40 This usually occurs via substitution nucleophilic reaction mechanisms.38,41 However, in an aqueous milieu, melphalan can undergo reactions with other nucleophiles, such as water (hydrolysis), and subsequently generate hydrochloric acid (HCl), which could, in turn, affect the pH of the milieu.42−44 Chemical instability of DNA alkylating agents is a major clinical concern.45 The clinical formulation of melphalan has been reported to readily undergo hydrolysis.46 However, the rate of hydrolysis and subsequent pH changes to the milieu are all dependent on a number of physicochemical parameters.44,47−49 All these physicochemical parameters could affect melphalan’s CEST MRI contrast generation. Thus, in this study, we evaluated the feasibility of imaging melphalan with CEST MRI despite its chemical instability, and

targeting selected regions of the brain and transiently opening the BBB in a predictable fashion.17,18 Image-guided, localized BBB opening is an important step toward the more effective treatment of brain tumors; however, the success of targeted drug delivery ultimately is dependent on the drug reaching the tumor. In this context, noninvasive imaging methods that facilitate the direct detection of injected drugs, and subsequently permit the visualization of the drug delivery process, are of utmost importance.19−24 Also critical is the development of methods to monitor early tumor responses to therapy, following drug delivery.22,25−28 Current methods to monitor tumor responses to therapy are based on measuring either the changes in tumor volume or the changes in tumor metabolism.22,25−28 However, tumor shrinkage and treatment-induced metabolic changes usually require time to occur. In addition, most currently available methods to monitor tumor responses to therapy do not directly report on the drug delivery process. Consequently, developing methods capable of monitoring early tumor responses to therapy after drug delivery would be beneficial in validating the efficiency of the drug delivery process. Ultimately, these methods could enable the personalization of therapeutic regimens. Magnetic resonance imaging (MRI) offers several advantages for imaging drug delivery and tumor responses to therapy.22−27 Chemical exchange saturation transfer magnetic resonance imaging (CEST MRI) is particularly promising for imaging a variety of organic molecules by saturating specific exchangeable protons on the molecules of interest, with defined radiofrequency pulses.29,30 However, CEST MRI is dependent on a number of physicochemical parameters especially the pH of the milieu.22,31 Previously, we demonstrated the feasibility of imaging the enzymatic conversion of chemotherapeutic B

DOI: 10.1021/acs.molpharmaceut.6b00130 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics

where melphalan is most soluble (pH 0.0), the rate of melphalan hydrolysis and its effects on pH changes were studied over three different pH ranges at which CEST MRI could be performed: mildly acidic (pH 6.7−5.8); neutral (pH 7.7−6.9); and mildly basic (pH 8.7−7.9). The specific pH ranges were chosen for the following reasons: (1) pH range 6.7−5.8 was chosen since the melphalan clinical formulation is prepared at pH 6.5. (2) pH range 7.7−6.8 was chosen to study melphalan’s hydrolysis close to physiological pH 7.4. (3) pH range 8.7−7.8 was chosen to study melphalan’s hydrolysis above the physiological pH. 2.3. Evaluation of the CEST Contrast of Melphalan as a Function of pH. Melphalan solutions (10 mM) were prepared in 10 mM PBS, and the pH of the solutions was adjusted to pH 6.7, 7.7, and 8.7 respectively, as described above. 80 μL of the respective solutions were then transferred to 1 mm microcapillaries, which were sealed, placed in a holder, and imaged by CEST MRI.50,51 Glutamate solutions (10 mM) prepared in 10 mM PBS at pH 7.4 were used as positive CEST MRI controls because glutamate possesses amine protons similar to melphalan and has been shown to generate a CEST contrast.52 To investigate how the CEST signal generated by melphalan changed as a function of hydrolysis-induced pH changes, melphalan samples were imaged sequentially at three time points following sample preparation, and the pH of the samples was measured over the scan period. All CEST MR images were acquired on an 11.7 T vertical Bruker Avance system equipped with a 20 mm radiofrequency coil, at 37 °C, using the high-throughput MRI method for screening CEST agents (Figure S1), and previously described.53,54 A modified RARE sequence, including a magnetization transfer module with the following parameters, was used to acquire CEST-weighted images: echo time (TE) = 5 ms; effective echo time (ETE) = 20 ms; repetition time (TR) = 6000 ms; RARE factor =16; slice thickness (ST) = 1 mm; number of slices (NSL) = 1; field of view (FOV) = 17 × 17 mm; matrix size (MS) = 128 × 64; resolution = 0.13 × 0.26 mm2; number of averages (NA) = 2; saturation power (B1) = 4.6 μT; saturation module time (tsat) = 4000 ms; saturation pulse bandwidth = 0.43 Hz; initial saturation transfer offset (ω) = 5 ppm; final saturation transfer offset (−ω) = −5 ppm; saturation offset step (Δω) = 0.2 ppm; pulse shape = hermite; encoding order = linear; and acquisition time = 62.4 min. The B0 inhomogeneity of all the respective CEST images was mapped and corrected using a water saturation shift reference (WASSR) method modified with Lorentzian analysis, as previously described.53,55 The same parameters as those used for the CEST image acquisition were used for the WASSR acquisition, except for the following modifications: TR = 1500 ms; tsat = 500 ms; saturation pulse bandwidth = 2.56 Hz; B1 = 0.5 μT; ω = 1.5 ppm; −ω = 1.5 ppm; Δω = 0.1 ppm; and acquisition time = 6.1 min. The acquired data was next processed using custom-written MATLAB (Mathworks) scripts, with the signal-to-noise ratio set at 30. CEST spectra (Z-spectra) were plotted from the mean signal intensities of individual sample ROIs, after B0 inhomogeneity mapping and correction. The CEST contrast was next computed by calculating the magnetization transfer asymmetry (MTRasym = 100 × (S−Δω − SΔω)/S0) at the respective frequency offsets, as previously described.53 2.4. Lentivirus Production and Mammalian Cell Transduction. Human embryonic kidney 293 cells (HEK293FT) were cultured in Dulbecco’s modified Eagle’s

determined the physicochemical parameters that affect melphalan’s CEST contrast generation both in MRI phantoms and in vitro in cancer cell culture preparations. We also evaluated the feasibility of monitoring early therapeutic responses to melphalan, by exploiting the physicochemical properties that affect melphalan’s CEST signal. Ultimately, this could provide a clinically translatable method with which to monitor, in real time, the biodistribution of melphalan, without compromising its therapeutic effectiveness.32

2. EXPERIMENTAL SECTION All solvents and reagents were used as obtained from Sigma− Aldrich and Thermo Fisher Scientific, unless otherwise stated. All NMR spectra were acquired using a Bruker AVANCE 500 MHz spectrometer, at ambient temperature. Deuterated NMR solvents with internal standards were used for all NMR measurements. All CEST MRI data were acquired on a vertical 11.7 T Bruker Avance scanner, equipped with a 20 mm RF coil, at 37 °C. All pH measurements were performed using a Mettler Toledo SevenCompact pH meter. 2.1. Evaluation of the Stability of Melphalan by NMR Spectroscopy. Two solutions of 4-[bis(2-chloroethyl)amino]L-phenylalanine (melphalan) were prepared by dissolving 2.5 mg of melphalan in 500 μL of a 1:20 mixture of deuterium chloride (DCl) and deuterium oxide (D2O). The final concentration of melphalan in solution 1 was adjusted to 10 mM using D2O, and the pH of solution 1 was measured as prepared. The pH of solution 2 was adjusted to pH 7.4 using high concentrations of DCl and sodium deuteroxide (NaOD). No pH corrections were made for the deuterium or temperature effects. High concentrations of DCl and NaOD were used to minimize the changes in the sample concentration. The final concentration of melphalan in the solution was then adjusted to 10 mM using D2O. These two pH values were chosen to understand the hydrolysis rate of melphalan in acidic and neutral conditions. This was done because the rate of melphalan hydrolysis has been previously reported to be dependent on both the pH of the milieu and the concentration of melphalan.41,48,51 The solutions were then filtered using cotton filters, and 300 μL of the respective solutions were next transferred to 5 mm NMR tubes. The NMR spectra of the solutions were then serially acquired over 24 h, at 30 min intervals. All NMR spectral changes were analyzed using the Advanced Chemistry Development, Inc. (ACD), 1D NMR processor, on the ACDLABS 12.0 software. 2.2. Evaluation of the Rate of pH Changes Induced by Melphalan Hydrolysis. Three solutions of melphalan were prepared by dissolving 2.0 mg of melphalan in 100 μL of a 1:20 mixture of concentrated hydrochloric acid (HCl) and deionized water (H2O). The mixtures were then added to 400 μL of 10 mM phosphate-buffered saline (PBS), and the pH of the solutions was adjusted to 6.7, 7.7, and 8.7, respectively. The final concentration of melphalan in the solutions was then adjusted to 10 mM using PBS, and the final pH of the solutions was again measured. The rate of pH change in the respective solutions was next monitored over 6 h, at 30 min intervals. Since the rate of melphalan hydrolysis has been previously reported to be both concentration and pH dependent, at the melphalan concentration needed for CEST MRI detection, the rate of melphalan hydrolysis and its effects on the pH of the milieu could differ from previous reports.41,48,51 Thus, since CEST MRI cannot be performed in purely acidic solutions, C

DOI: 10.1021/acs.molpharmaceut.6b00130 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics

Figure 2. (a) NMR spectra of a 10 mM melphalan solution prepared at pH 7.4. NMR spectra were collected sequentially over 12 h, and at ambient temperature. (b) Melphalan-induced pH changes through substitution nucleophilic reaction mechanisms (hydrolysis). (c) Rates of pH changes of 10 mM melphalan solutions prepared at pH 8.7, pH 7.7, and pH 6.7, and measured at ambient temperature. The error bars represent the standard deviation from the experimental mean (n = 3). The dashed lines connect the experimental points for better visualization.

measurements, the cells were incubated with 10 μL of a 150 μg/mL luciferin solution (PerkinElmer) for 10 min prior to the measurements. 2.6. Imaging Melphalan in Vitro with CEST MRI. Luciferase-transduced Daoy cells were plated in T150 flasks (3 × 106 cells/flask) and cultured in improved MEM (Ritcher’s modified) without serum for 18 h. The cells were then treated with 100 μM melphalan for 6 h. Following this, the cells were rinsed thrice with PBS and harvested, and 6 × 106 cells were counted using a Cellometer (Nexcelom Bioscience). The cells were then pelleted in 100 μL Eppendorf tubes, the pH of the supernatant was measured, and the cell pellets were imaged with both CEST MRI and BLI. All CEST MR images were acquired on an 11.7 T vertical Bruker Avance system, as described in section 2.3, with the following exceptions: slice thickness (ST) = 2 mm; field of view (FOV) = 20 × 20 mm; resolution = 0.16 × 0.31 mm2; initial magnetization transfer offset (ω) = 10 ppm; and final magnetization transfer offset (−ω) = −10 ppm. Data processing and B0 inhomogeneity corrections were performed as described in section 2.3. BLI was acquired after the MRI scans. The cells were incubated with 10 μL of a 150 μg/mL luciferin solution for 10 min, and were then imaged. 2.7. Imaging Therapeutic Responses to Melphalan. Luciferase-transduced Daoy cells were treated with melphalan, as described in section 2.6. Following this, the cells were rinsed thrice with PBS and harvested, and 6 × 106 cells were counted. The cells were then pelleted in 100 μL Eppendorf tubes and stored at room temperature, and the pH of the supernatant was monitored until it decreased to 6.4. The cell pellets were then imaged with both CEST MRI and BLI, as described in section 2.6. 2.8. Immunocytochemistry of Melphalan-Treated Cells. Nontransduced Daoy cells were plated in eight-welled chamber slides (1 × 105 cells/well) and cultured in improved MEM (Ritcher’s modified) without serum for 24 h. After 18 h,

medium (DMEM) (Invitrogen) containing 10% fetal bovine serum (FBS, Gibco) and 1% penicillin/streptomycin (Sigma). The cells were then seeded in T75 flasks at 80% confluency and maintained at 37 °C and 5% CO2 for 12 h. An infectious viral supernatant was next generated by transiently cotransfecting the lentiviral vector (pLenti-6-Luc2) containing the gene of interest and the packaging vectors into the HEK293T cells, using Lipofectamine 2000. The viral supernatant was collected 48 h post transfection and concentrated using an Amicon Ultra15 100 kDa cutoff filter (Millipore, Billerica, MA). Following this, human medulloblastoma cells (Daoy cells) were transduced for 48 h with the lentiviral vector, using 0.1% Polybrene (Sigma-Aldrich), in improved (Ritcher’s modified) minimum essential medium (MEM), without antibiotics. To evaluate the correlation between the number of transduced Daoy cells and the bioluminescence imaging (BLI) signal, luciferase-expressing Daoy cells were seeded in 24-well plates at varying cell densities. One hour later, 500 μL of a 150 μg/mL luciferin solution (PerkinElmer) was added to each well, and the plate was read 10 min after luciferin addition. All BLI experiments were performed using a Xenogen IVIS 200 optical imaging system (PerkinElmer Inc.), and the images were processed using the Xenogen Living Imaging 4.2 software. The BLI radiance from the samples was used to determine cell viability. 2.5. Determination of the Cytotoxicity of Melphalan in Daoy Cells. To determine the optimum melphalan treatment conditions needed to induce ≥80% of cell death (IC80) in Daoy cells and to determine the time frame over which this occurred, luciferase-transduced Daoy cells were plated in 96-well plates (1 × 104/well) and cultured in improved MEM (Ritcher’s modified) without serum, for 18 h.56 The cells were next treated with varying concentrations of melphalan ranging from 0 to 100 μM for either 3 or 6 h. Following this, the cells were rinsed thrice with PBS and the viability of the cells was measured 1 and 24 h post treatment, using both the MTT assay (Sigma-Aldrich) and BLI. For BLI D

DOI: 10.1021/acs.molpharmaceut.6b00130 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics

Figure 3. Kinetics of CEST MRI contrast generation by 10 mM melphalan solutions prepared at pH 8.7, 7.7, and 6.7, and measured at 37 °C. Zspectra of melphalan solutions: (a) 1 h; (b) 2 h; and (c) 3 h post sample preparation. MTRasym plots of melphalan solutions: (d) 1 h; (e) 2 h; and (f) 3 h post sample preparation. MTRasym maps of melphalan solutions: (g) 1 h; (h) 2 h; (i) 3 h post sample preparation.

100 μM melphalan was added, for 6 h. Next, the cells were rinsed thrice with PBS, fixed in acetone for 10 min at room temperature, air-dried for 10 min at room temperature, and stored overnight at −20 °C. The frozen chamber slides were next brought to room temperature, rinsed thrice in PBS for 5 min each time, and then treated with 100 mM NaOH for 5 min. Next, the cells were rinsed thrice in PBS for 5 min. Melphalan−DNA adducts were then stained using an antimelphalan mouse monoclonal antibody AMP4/42 (Abcam, ab97554). The antibody was prepared at a 1:100 dilution in PBS containing 1% (w/v) bovine serum albumin and 10% (v/ v) horse serum. Following this, the cells were incubated at 4 °C for 30 min with an Alexa Fluor 594 rabbit anti-mouse (Invitrogen) secondary antibody, prepared at a 1:500 dilution in PBS containing 1% (w/v) bovine serum albumin and 10% (v/v) horse serum. The cells were next rinsed thrice with PBS (5 min each time) and counterstained with DAPI. Images were then acquired with a Zeiss microscope equipped with the Zen Pro software. 2.9. Statistical Analyses. Data are presented as the mean ± standard deviation of at least three independent experiments, except when otherwise stated. Statistical comparisons were made using a two-sided Scheffé multiple comparison test, based on the Student’s t test, and considered statistically significant at P < 0.05, except when otherwise stated.

NMR spectral changes were detected over the entire observation period (Figure S2). However, at pH = 7.4, some changes were observed. This included a slight shifting of the peaks (Figure 2a and Figure S3), which suggested a change in the pH of the solution over time, presumably due to melphalan hydrolysis (Figure 2b). Based on the peak integrations, trace changes (13%) were observed over the first 3 h (Figure S4). These spectral changes could also be reflective of trace structural changes. 3.2. Melphalan Hydrolysis Induces Changes to the pH of the Milieu. In all three melphalan solutions evaluated, the pH of the solutions decreased as a function of time (Figure 2c). Overall, the rate of pH change induced by melphalan’s hydrolysis was a function of the initial pH of the milieu. Whereas the rate of pH change was fastest (2.4%/h) in an alkaline milieu (pH = 8.7), it was slowest (0.6%/h) in an acidic milieu (pH 6.7) within the first 6 h of sample preparation. An intermediate rate of pH change (1.3%/h) was observed at a pH close to physiological pH (7.7). This trend is in accordance with previous reports and suggested that melphalan could be imaged with CEST MRI at a pH near 6.7, since the pH of the solutions would change by 6.5) was attributed to the previously described and well-established phenomenon of intermediate to fast exchange mediated chemical shift averaging observed with other CEST contrast agents.52 3.4. Melphalan Induces Significant Cytotoxic Effects in Daoy Cells. Luciferase-transduced cells showed a linear correlation between the BLI radiance and the cell number (Figure S10). All cytotoxic effects of melphalan were observed 24 h post treatment (Figure 4). No significant cytotoxic effects were observed 1 h post treatment by either BLI (Figure 4a,b) or the MTT assay (Figure 4c). Overall, the maximum cytotoxic effects were observed in the cells treated for 6 h, compared to those treated for 3 h (Figure 4a−c). However, significant cell death (P < 0.05) was observed 24 h post treatment in all cells treated with melphalan at 100 μM for either 3 or 6 h. No significant cell death was observed 1 h following treatment at this dose as determined by both the MTT assay and BLI. This most likely can be attributed to the doubling time of the cells, as melphalan interferes with DNA replication. The Doay cells have a doubling time of approximately 18 h, thus at the 24 h post treatment time point, the effects of the drug are most visible. Overall, the maximum cell death (≥80%) was observed 24 h post treatment, in cells treated with 100 μM melphalan for 6 h (Figure 4a−c). In addition, cells treated with melphalan for 6 h, and stained with a melphalan antibody, stained positive for melphalan, compared to the nontreated controls (Figure 4d). The subcellular localization patterns showed distribution throughout the melphalan-treated cells, as opposed to specific localization in the nuclei as expected. Furthermore, both the cytoplasm and nuclei of the treated cells displayed a shriveled morphology, compared to the nontreated cells. These findings suggested that melphalan−DNA adducts could be released into the cytoplasm at the early stages of melphalan-induced cell death. 3.5. The CEST Contrast Generated by Melphalan Can Be Detected in Vitro Following Drug Administration. In melphalan-treated cells, imaged 1 h post treatment, a significant increase (P < 0.05) in the CEST contrast was observed at 2.5 ppm from the water signal, compared to the nontreated cells (Figure 5a−c,g and Figure S11). In addition, a broadening of the melphalan CEST signal peak to a higher frequency offset (∼3.3 ppm) in melphalan-treated cells compared to that

4. DISCUSSION In this study, we evaluated the feasibility of directly imaging melphalan by saturating its exchangeable amine protons at approximately 2.5 ppm from the water proton signal, for CEST MRI. Given melphalan’s propensity to undergo hydrolysis in an aqueous milieu, which could affect the pH and resultant CEST contrast generation, we first assessed melphalan hydrolysisinduced changes to solutions. The NMR experiments were conducted in a deuterated solvent (deuterium oxide) as is generally performed in organic chemistry, as opposed to phosphate buffered saline (PBS) or 900 mM saline for the following reasons: (1) In deuterated solvents such as deuterium oxide, the residual water (H2O) proton peak at δ = 4.75 ppm is very narrow due to the low concentration of water protons compared to deuterium. This permits the clear visualization of all the melphalan proton peaks and enables the assessment of what happens to the full melphalan NMR spectrum over time. Following the assessment of the full NMR spectrum, the specific spectral regions affected can then be analyzed more closely. At high water proton concentrations generally found in PBS and saline, the water proton peak at δ = 4.75 ppm is extremely broad, as previously demonstrated.58 Thus, if the NMR spectrum of melphalan was acquired in PBS or saline at the concentrations at which melphalan is soluble in aqueous milieu at pH 7.4, the water proton peak would interfere with some of the melphalan proton peaks and prevent the proper evaluation of the full melphalan NMR spectrum over time. This phenomenon would be exacerbated by the relatively low maximum solubility of melphalan in aqueous milieu at pH 7.4. In more water soluble compounds this phenomenon, although present, would be less G

DOI: 10.1021/acs.molpharmaceut.6b00130 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics

Melphalan generated a CEST contrast that was dependent on the pH of the milieu. The maximum CEST contrast generated by melphalan at ∼2.5 ppm from the water protons was observed at a pH close to 6.2. Since the CEST contrast of amine protons has been reported to occur via a base-catalyzed reaction mechanism, this could account for the optimum exchange rates and subsequent maximum CEST signal (at ∼2.5 ppm from the water protons) observed at pH 6.2. In this study, melphalan’s CEST contrast was evaluated at a concentration of 10 mM, for the following reasons: Since this concentration falls above the detection limit for CEST MRI, if a contrast was not detected at this concentration, it could be attributed to either the experimental conditions (ionic content, ionic strength, or pH of the solutions) or to molecular structure of the agent. At lower concentration ranges (μM), it would have been more difficult to attribute a lack of contrast, if any, with absolute certitude, to either the experimental conditions, the concentration range tested, or the molecular structure of the agent. Although we initially intended to carry out a concentration dependent study after detecting the initial CEST contrast at 10 mM, we opted against this due to the dynamic contrast generated by melphalan in solution as a result of hydrolysis. Although the melphalan solution concentrations (10 mM) tested for CEST MRI are too high for conventional systemic chemotherapy, CEST MRI would most suitably complement interventional neuroradiology drug delivery techniques, which permit the delivery of high drug concentrations to brain tumors. Clinically, melphalan is reconstituted with an acidified diluent at a concentration of 16 mM, due to solubility issues in aqueous milieu.46 For systemic intravenous administration, this preparation is then diluted to 1.4 mM and administered as a single infusion over 15 to 20 min at a dose of 16 mg/m2, to avoid myelosuppression.59 Some reports using positron emission tomography (PET) for quantification suggest that