Development of a Nanoscale Chemical Exchange Saturation Transfer Magnetic Resonance Imaging Contrast Agent That Measures pH

Abstract

AcidoCEST MRI can measure the extracellular pH (pHe) of the tumor microenvironment in mouse models of human cancers and in patients who have cancer. However, chemical exchange saturation transfer (CEST) is an insensitive magnetic resonance imaging (MRI) contrast mechanism, requiring a high concentration of small-molecule agent to be delivered to the tumor. Herein, we developed a nanoscale CEST agent that can measure pH using acidoCEST MRI, which may decrease the requirement for high delivery concentrations of agent. We also developed a monomer agent for comparison to the polymer. After optimizing CEST experimental conditions, we determined that the polymer agent could be used during acidoCEST MRI studies at 125-fold and 488-fold lower concentration than the monomer agent and iopamidol, respectively. We also determined that both agents can measure pH with negligible dependence on temperature. However, pH measurements with both agents were dependent on concentration, which may be due to concentration-dependent changes in hydrogen bonding and/or steric hindrance. We performed in vivo acidoCEST MRI studies using the three agents to study a xenograft MDA-MB-231 model of mammary carcinoma. The tumor pHe measurements were 6.33 ± 0.12, 6.70 ± 0.15, and 6.85 ± 0.15 units with iopamidol, the monomer agent, and polymer agent, respectively. The higher pHe measurements with the monomer and polymer agents were attributed to the concentration dependence of these agents. This study demonstrated that nanoscale agents have merit for CEST MRI studies, but consideration should be given to the dependence of CEST contrast on the concentration of these agents.

Graphical Abstract

Tumor acidosis is a common characteristic of many solid tumors. The dysregulated metabolism of tumor cells relies on enhanced glycolysis, known as the Warburg effect, which produces excess lactic acid that is secreted from tumor cells.^1–3 This metabolic process causes the extracellular tumor microenvironment to become acidic, typically with an extracellular pH (pHe) of 6.3–7.0.⁴ A low pHe promotes tumor invasion into adjacent tissues and metastasis to remote tissues and also contributes resistance to chemotherapy and immunotherapy.^5–13 Therefore, measurements of tumor pHe may be used to improve tumor diagnoses and monitor the effect of anticancer treatments that directly or indirectly change tumor glycolysis.¹⁴

We have developed a noninvasive magnetic resonance imaging (MRI) method that measures tumor pHe, based on exogenous chemical exchange saturation transfer (CEST).^15,16 This method selectively saturates the MR signal of a proton on an agent with a radiofrequency pulse, causing the MR signal to disappear (Figure 1a, left). Subsequent chemical exchange of this proton with a proton on water causes the water signal to decrease (Figure 1a, right). We refer to this decrease in water MR signal as a “CEST signal”, which can be detected using conventional MR image acquisition methods. The chemical exchange of many types of exchangeable protons is base-catalyzed, and therefore the CEST signal is dependent on pH. More specifically, we have shown that the ratio of two CEST signals generated by selective saturation of two types of exchangeable protons can be used to measure pH in a concentration-independent manner, a process known as acidoCEST MRI.¹⁷ We have used acidoCEST MRI to perform many preclinical imaging studies with paramagnetic agents^18–20 and diamagnetic agents,^21–35 and we have translated acidoCEST MRI for clinical imaging evaluations.³⁶ Other research programs have developed and used similar acidoCEST MRI methods with paramagnetic^37–42 and diamagnetic agents^43–53 that have two selectively detectable CEST signals or agents that generate two CEST signals by using two saturation powers.^54–59

Figure 1.

CEST MRI of monomer and polymer CEST agents. (a) Schematic of the CEST MRI process. RF saturation is applied that causes a proton to lose its coherent magnetic signal. Then, the saturated proton undergoes chemical exchange from the agent to water, which effectively transfers the saturation to water and lowers the water signal. The lower water signal is measured using standard MRI methods. (b) Chemical structures of studied monomer (left) and polymer agent (right; n ≈ 125) with amide and salicylic acid protons that resonate at 5.0 and 9.2 ppm, respectively.

CEST MRI is an insensitive imaging technique.^15,16 As an advantage, one agent molecule can affect the MR signal of approximately 100 to 1000 water molecules due to the multiplication effect of the chemical exchange rate that is on the order of 100–1000 Hz. However, CEST must suppress the MR signal of ~1 M of water protons to suppress 1.3% of the ~78 M of water protons in biological tissues (assuming that 1.3% is a minimally detectable CEST signal, pure water is 55.5 M and water protons are 111 M, and the water concentration in most tissues is ~70%). Therefore, the CEST agent with a 100–1000 Hz exchange rate must have a minimum concentration of approximately 1–10 mM to affect ~1 M of water protons during CEST MRI studies. Our acidoCEST MRI method uses iopamidol, a clinically approved contrast agent that can be administered at 972 mM concentration to preclinical tumor models and patients who have cancer.¹⁷ However, even this high concentration of a well-tolerated agent can fail to deliver enough agent with adequate penetrance throughout the tumor, causing pHe measurements with acidoCEST MRI to be spatially incomplete.

To address this problem, we proposed that a nanoscale agent may deliver a greater payload of exchangeable protons to the tumor. A variety of nanoscale CEST agents have been developed that use polymers, dendrimers, proteins, and viral particles to generate a strong CEST signal.^60–81 Paramagnetic nanoscale agents have the advantage of very large MR frequencies (also known as chemical shifts) that are well separated from the MR frequency of water (defined as 0 ppm for MRI studies).^60–71 This frequency separation allows the selective saturation of the agent’s MR resonance to be performed with high radiofrequency power, which further improves CEST detection. However, paramagnetic agents often use potentially toxic lanthanide metals or use other metals that have unstable oxidation states especially under in vivo conditions.⁸² Therefore, diamagnetic nanoscale agents that are nonmetallic have advantages for biomedical studies.^72–81 Notably, the initial studies with paramagnetic and diamagnetic nanoscale agents were initially reported in 2003, but reports of paramagnetic nanoscale agents have not appeared since 2013, while studies with diamagnetic nanoscale agents have continued.

Surprisingly, diamagnetic nanoscale CEST agents have not yet been developed for acidoCEST MRI applications that measure tumor pHe. Therefore, we sought to develop such an agent that can potentially improve the comprehensive spatial interrogation of tumor tissues. To develop a nanoscale polymer, we used 4-amidosalicylate as a monomeric unit, because this monomer has two types of exchangeable protons that are known to have high MR frequencies, which facilitates selective saturation and potentially allows for the use of higher saturation powers that can improve detection sensitivity (Figure 1b). We designed this polymer to have >100 exchangeable protons, to investigate if the higher payload of exchangeable protons can improve CEST sensitivity on a permolecule basis. We also investigated acidoCEST MRI of the monomer agent, 4-acrylamidosalicylic acid, which has a similar electronic structure to the polymer agent. We performed detailed acidoCEST MRI studies with chemical samples to evaluate if the monomer and polymer agents can accurately measure pH with an acidoCEST MRI protocol. We then performed a preliminary in vivo study with a tumor model of mammary carcinoma using acidoCEST MRI with the polymer agent, monomer agent, and iopamidol, to evaluate the utility of a nanoscale agent for this biomedical imaging application.

RESULTS AND DISCUSSION

Synthesis and Characterization of Monomer and Polymer Agents.

The monomer 4-acrylamidosalicylic acid was synthesized by following a previously reported method with minor modifications (Scheme S1b).⁸³ The starting material, 4-acrylamidosalicylic acid, was polymerized using 4,4′-azobis(4-cyanovaleric acid) as initiator at 70 °C to afford the linear polymer. The polymer was purified by dialysis using a 10 kDa MW cutoff membrane. We also synthesized a monomer agent, 4-acetamido-2-hydroxybenzoic acid, which has a similar electronic structure to the polymer (Figure 1b, Scheme S1a).

Ultraviolet−visible (UV−vis) spectra further confirmed the presence of the salicylic acid repeating groups in the polymer, based on the characteristic maximum absorption wavelength for salicylic acid at approximately 310 nm (Figure 2a). The resolved repeat units shown in a matrix-assisted laser desorption/ionization−time-of-flight (MALDI-TOF) spectrum were separated by 207.04 g/mol, which confirmed that the analyte was a polymer that exclusively contained the salicylic acid-based monomer (Figure 2b). Gel permeation chromatography (GPC) results showed that the weight-average molecular weight $\left({\overline{M}}_{\text{W}}\right)$ of the polymer agent was 26 000 g/mol with a 1.8 polydispersity index (Figure 2c). Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) of the polymer agent showed a variable morphology and particle size of the polymer with an average particle size of 361 nm and a range from 45 to 710 nm, as expected based on high polydispersity index from GPC (Figure 2d).

Figure 2.

Chemical characterization of the monomer and polymer agents. (a) Normalized UV−vis spectra of the monomer and polymer agents at pH 7.0. (b) MALDI-TOF spectrum of the polymer. (c) GPC trace of the polymer. (d) SEM image (left) and TEM image (right) of the polymer.

AcidoCEST MRI of Chemical Samples.

The CEST properties of the monomer and polymer agents were assessed over a range of pH values at 37.0 °C and 7 T magnetic field strength. The Z spectra (which show the Z-axis magnetization after steady state CEST saturation; Figure 3a,d) were analyzed by fitting a sum of three Lorentzian line shapes to the experimental results (Figure 3b,e). The samples were placed in our customized holder with agarose gel that reduced B₀ and B₁ magnetic field inhomogeneities and greatly improved the quality of our CEST results, as evidenced by the low residuals of the Lorentzian line fitting analysis (Figure S1).

Figure 3.

CEST properties of the monomer and polymer agents. Samples consisted of 20 and 0.16 mM of each agent, respectively, and were analyzed at 37 °C, 4 μT saturation power, and 6 s saturation time. (a) Z-spectrum, (b) CEST Lorentzian spectrum, and (c) CEST−pH graph of the monomer agent and (d−f) similar spectra and graph for the polymer agent showing the pH dependence of the CEST effects for these agents. The pH values ranged from 6.0 to 8.0 in 0.1–0.15 unit increments.

CEST MRI measurements revealed that both agents show CEST signals at 5.0 and 9.2 ppm, which is consistent with CEST signals of similar agents (Figure 3a,b).^84–89 The CEST signal at 5.0 ppm in the polymer agent is not as strong as in the monomer agent (Figure 3a,d). This lower contrast may be a result of steric blocking between water molecules and the amide in the polymer due to the proximity of the amide to the polymer backbone, relative to the water−amide interactions of the monomer agent and the water−salicylic acid interactions in the polymer and monomer. Steric blocking can reduce the chemical exchange rate (k_ex), resulting in a lower CEST signal.⁹⁰ A similar result has been previously observed with other polymeric CEST agents.⁷⁵ Furthermore, intramolecular hydrogen bonding within the polymer may also lead to a slower k_ex, leading to a lower CEST signal. Indeed, hydrogen bonding within a salicylic acid moiety is known to reduce k_ex from this group relative to k_ex of hydroxyl and carboxylic acid groups, showing that hydrogen bonding affects CEST.⁸⁴ Although steric hindrance and hydrogen bonding are well-known molecular characteristics that affect chemical exchange rates, other characteristics and conditions may also affect the chemical exchange rate of the polymer relative to the monomer agent.

The Z-spectra (Figure 3a,d) and the CEST Lorentzian spectra (Figure 3b,e) show the pH-dependence of the CEST effects for both agents. Interestingly, the monomer and polymer agents exhibited different pH dependencies (Figure 3c,f). In general, the chemical exchange processes of the salicylic acid and amide are known to be base-catalyzed. A faster k_ex causes greater CEST signal, until k_ex becomes too fast for the CEST process and less CEST can be generated (Figure 1a). The CEST signal for the salicylic acid of the monomer agent (Figure 3c, filled circles) follows this classic description, because k_ex of salicylic acid is moderate at low pH, reaches a faster rate that is ideal for generating CEST at pH ~7.1 (at 37.0 °C and 7 T magnetic field strength), and an even faster k_ex at pH > 7.1 generates less CEST signal. For comparison, k_ex for the amide of the monomer agent is slow at low pH and becomes faster with increasing pH, but does not reach an ideal k_ex value for the greatest CEST signal until a pH value > 7.45, causing the CEST−pH relationship to be monotonic for the amide at pH < 7.45 (Figure 3c, open circles). The salicylic acid of the polymer agent has a similar CEST−pH relationship to the amide of the monomer agent (Figure 3f, filled circles), indicating that the k_ex for the polymeric salicylic acid is slow at low pH (relative to the monomeric salicylic acid) and becomes faster with increasing pH, but does not reach an ideal k_ex value for the greatest CEST signal until a pH value > 7.5. Finally, the amide of the polymer agent has a low CEST signal at all pH values, indicating a very slow k_ex (especially relative to the monomeric amide) that does not appreciably increase with increasing k_ex values (Figure 3f, open circles). The slower k_ex of the polymer agent relative to the monomer agent may be due to steric hindrance, increased hydrogen bonding, or other characteristics that are not yet known, as described above.⁸⁴

Optimization of CEST Saturation Conditions.

We optimized the saturation power for acidoCEST MRI with the monomer and polymer agent (Figure S2a,b). The Hanes−Woolf quantification of exchange rate using varying saturation power (HW-QUESP) method was used to determine that % CEST at 5.0 and 9.2 ppm reached >80% and >65% of the maximum CEST signal at 4.0 μT saturation power for the monomer and polymer agents, respectively.⁹¹ A saturation power of 4 μT has been used as a safe power level for in vivo imaging,¹⁶ so this power was selected as a power level for subsequent experiments. The reverse linear quantification of exchange rate using varying saturation time (RL-QUEST) method was used to evaluate the effect of saturation time, which exhibited that % CEST at 5.0 and 9.2 ppm reached >98% of the maximum CEST signal at 6 s saturation time for both monomer and polymer agents (Figure S2c,d).⁹² Therefore, we used a 6 s saturation time for our subsequent studies. Importantly, these analyses were performed with monomer and polymer agent concentrations of 20 mM and 160 μM, respectively, showing that the polymer can generate detectable CEST signals with a 125-fold reduction in concentration relative to the monomer.

Measuring pH with acidoCEST MRI.

We investigated the ability of the monomer and polymer agents to measure pH via CEST MRI, using our imaging method known as acidoCEST MRI.¹⁷ Samples ranging from pH 6.4 to 7.5 were imaged with optimized saturation conditions, and the results were used to calibrate a log₁₀ ratio of the agent’s CEST effects with pH. Both calibrations were linearly correlated with pH within this physiological pH range with an R² value of 0.97 and 0.98 for monomer and polymer agents, respectively (Figure 4a,b). The calibrations showed opposite trends due to the different regimes of k_ex for the faster-exchanging monomer agent vs the slower-exchanging polymer agent, as described above and shown in Figure 3c,f. These calibrations were used to calculate the pH for each pixel within acidoCEST MR images of samples that consisted of each agent over a range of pH values (Figure 4c,d).

Figure 4.

CEST MRI of the monomer and polymer agents can measure pH. CEST properties of 20 mM monomer agent and 0.16 mM polymer agent were measured at 37 °C, 4 μT saturation power, and 6 s saturation time. A ratio based on CEST at 9.2 and 5.0 ppm is linearly correlated with pH within the physiological pH range (6.3 and 7.5 units) for the (a) monomer agent and (c) the polymer agent. These calibrations were used to convert % CEST signal to pH values for each pixel within the MR images of samples that contained (b) monomer agent at pH values of 6.6, 6.75, 6.9, 7.0, 7.15, 7.3, and 7.45 units and (d) polymer agent at pH values of 6.45, 6.6, 6.75, 6.9, 7.0, 7.15, 7.3, and 7.45 units.

Evaluation of Sample Conditions on acidoCEST MRI.

To examine how temperature affects pH measurements with acidoCEST MRI, CEST spectra were collected for the monomer and polymer agents between 25 and 49 °C and analyzed to measure % CEST values (Figure S3). For the monomer agent, the CEST signal increased with increasing temperature for CEST generated from the amide and salicylic acid, as expected. However, the pH measurement had a negligible dependence on temperature (ΔpH = 0.0089 unit per °C; R² = 0.89), because our ratiometric analysis method largely negates the effect of temperature. For the polymer agent, there was no significant change in CEST signal with increasing temperature for CEST generated from the amide and salicylic acid, and therefore the pH measurement had a negligible dependence on temperature (ΔpH = 0.0097 units per °C; R² = 0.84). This result was an improvement relative to the minor yet non-negligible temperature dependence of acidoCEST MRI measurements performed with a paramagnetic CEST agent.^20,93

We investigated how the concentration of each agent may affect pH measurements. CEST spectra were collected for the monomer agent between 1.25 and 40 mM and for the polymer agent between 9.5 and 300 μM (Figures S4, S5). The results indicate that each of the CEST signals was dependent on concentration, as expected. Unfortunately, the pH measurements with the monomer agent were dependent on the concentration of the agent, despite using a ratiometric approach that should negate a concentration dependence. The monomer only measured pH values with some accuracy at high pH and at high concentration greater than10 mM (Figure S4g,h). The monomer overestimated pH at high pH values when low concentrations of ≤10 mM were used and at all concentrations for low pH values (Figure S4e,f). Similarly, the pH measurements with the polymer agent also showed a dependence on the concentration of the agent. The polymer measured pH with some accuracy at high concentrations of ≥20 μM (Figure S5). However, low concentrations of polymer overestimated low pH values and underestimated high pH values. The extracellular tumor microenvironment is expected to have low pH, and intravenous injections of highly concentrated solutions often deliver only low concentrations of agent to tumors, indicating that the monomer and polymer agents may overestimate in vivo tumor pHe. This dependency of the pH measurement on the concentration of each agent suggests that the level of intermolecular hydrogen bonding between agents may be concentration-dependent, because hydrogen bonding affects CEST signal amplitudes.⁸⁴

Cytotoxicity of the CEST Agents.

We used a standard 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assay to analyze the cytotoxicity of each agent against MDA-MB-231 cells after 24 h of exposure to concentrations ranging from 0.125 to 62.5 mM for the monomer agent and 1–500 μM for the polymer agent (which equates to 0.125–62.5 mM of monomeric units in the polymer, based on an average of 125 units per polymer). The results indicate that cell viability was >95% at concentrations as high as 1.25 mM and 10 μM for the monomer and polymer, respectively. The cell viability was >77% and >71% for the highest concentrations that were tested (Figure 5). These results demonstrate that the agents should be biocompatible for in vivo applications. Although more thorough studies will need to be performed to evaluate the safety of each agent, this initial evaluation was sufficient to justify our subsequent in vivo acidoCEST MRI studies with each agent.

Figure 5.

In Vitro cytotoxicity of the monomer and polymer agents. Viabilities of MDA-MB-231 cells were evaluated after incubation with (a) the monomer agent and (b) the polymer agent.

In Vivo acidoCEST MRI.

The pHe values of eight mice with a xenograft tumor of MDA-MB-231 breast cancer were scanned with acidoCEST MRI using iopamidol, monomer agent, and polymer agent (Figure 6). Anatomical images of each mouse were used to localize the tumor (left column of Figure 6 and Figures S6–S8), and then we acquired CEST images before and after injection of each agent. We injected 7.8 mmol/kg of iopamidol, which we have used in our previous acidoCEST MRI studies.^17,21–35 For the monomer and polymer agent, we injected 2.0 and 0.016 mmol/kg, respectively. The polymer agent was injected at 125-fold and 488-fold lower concentrations than the monomer agent and iopamidol, respectively, which further tested our objective of measuring tumor pHe with the delivery of a lower concentration of agent. Using a rudimentary assumption that the agent is equally distributed throughout the body, and assuming an average tissue density of 1 g/mL, these injections equate to 2.0 and 0.016 mM of the monomer and polymer agents, respectively. These roughly estimated concentrations are comparable to the concentrations that resulted in >95% cell viability in our cytotoxicity tests (Figure 5), which justified injections at these concentrations. Furthermore, the solubility of the monomer agent was estimated to be approximately 300 mM. We injected 200 μL of 250 mM monomer agent, which is below this solubility limit. The parametric maps of % CEST signal amplitude at 5.0 ppm (second column of Figure 6 and Figures S6–S8) and 9.2 ppm (third column of Figure 6 and Figures S6–S8) indicate sufficient detection of both CEST effects throughout most of the tumor that was imaged. The pHe maps (fourth column of Figure 6 and Figures S6–S8) show colorful voxels where two significant CEST signals were measured and converted to pHe values.

Figure 6.

In vivo acidoCEST MRI. Images of a MDA-MB-231 tumor model following intravenous injection of 200 μL of (a) 976 mM iopamidol, (b) 250 mM monomer agent, and (c) 2.0 mM polymer agent (equivalent to 250 mM in monomer units). Anatomical images show the location of the tumor (left column). The parametric maps of % CEST signal amplitude at 5.0 ppm (second-left column) and 9.2 ppm (third-left column) demonstrate sufficient detection of both CEST effects throughout the tumor. The pHe maps (right column) show colorful pixels where two significant CEST signals (second-left and third-left columns) were measured and converted to pHe values.

The average tumor pHe of the tumor model measured with acidoCEST MRI using iopamidol was 6.33 ± 0.12 units (Figure 7). The consistent tumor pHe measured among the mice resulted in a low standard deviation, suggesting that acidoCEST MRI with iopamidol can perform precise evaluations of tumor acidosis. The average tumor pHe was measured to be 6.70 ± 0.15 and 6.85 ± 0.15 units with the monomer agent and polymer agent, respectively. Therefore, acidoCEST MRI overestimated tumor pHe with the monomer and polymer agents relative to measurements made with iopamidol. This result is consistent with our studies of chemical solutions that evaluated the concentration dependence of pH measurements, which indicated that the monomer and polymer agents may overestimate tumor pHe if both the pHe and agent concentrations in tumors were low.

Figure 7.

Results of sequential in vivo acidoCEST MRI studies. (a) AcidoCEST MRI was performed with iopamidol, the monomer agent, and the polymer agent. Each bar is labeled with the mouse number. (b) Comparison of average tumor pH using each agent (n = 4 for each bar).

The penetrance of a nanoscale agent into tumor tissue is typically lower than the penetrance of small-molecule agents due to the relative permeabilities of tumor vasculature for agents of different sizes. When designing our study, we hypothesized that the lower penetrance of a polymer agent may be overcome by the higher payload of exchangeable protons provided by the polymer. However, as shown in Figures 6, SS6, S7, and S8, the polymer agent had lower penetrance into the tumor region, resulting in lower spatial coverage of tumor pHe that was not overcome by the higher payload for adequate detection with CEST MRI. This result reinforces the importance of considering tissue penetrance when designing studies with nanosized imaging agents.

The tumor pHe of two mice were analyzed with three sequential acidoCEST MRI scans that used each agent, while six mice were analyzed with only one acidoCEST MRI scan. The sequential scans produced similar average tumor pHe measurements for the mice that were sequentially scanned or scanned only one time (Figure 7). This result was expected, because iopamidol and the monomer agent are small molecules that should clear from in vivo tissues within 1 day.⁹⁴ Therefore, acidoCEST MRI may be performed with sequential scans to monitor changes in tumor acidosis during tumor progression and during tumor response to treatment.

CONCLUSIONS

We have developed a small-molecule monomer agent and a nanoscale polymer agent for pH measurements with acidoCEST MRI. Optimized MRI conditions can detect CEST signals with a polymer agent that was 125-fold and 488-fold less concentrated than the monomer agent and iopamidol, respectively, demonstrating an advantage of the nanoscale CEST agent. A ratiometric approach for correlating CEST with pH showed excellent linear relationships, although these correlations had different trends for the monomer and polymer, presumably due to hydrogen bonding and/or steric hindrance within the polymer. AcidoCEST MRI with the monomer and polymer agents can measure pH with a negligible dependence on temperature, but showed a dependence on concentration that caused pH to be overestimated in tumors with low pHe and typically low agent uptake. Indeed, in vivo acidoCEST MRI with both agents overestimated tumor pHe relative to similar measurements made with iopamidol. To summarize, our polymer agent demonstrated many of the attributes needed for improved acidoCEST MRI with a lower concentration of agent, but future developments of nanoscale polymer agents for acidoCEST MRI should focus on CEST agents that potentially avoid hydrogen bonding and steric hindrance.

MATERIALS AND METHODS

Dry solvents were purchased and used without further purification or drying treatment. ACS grade solvents for chromatography, Na₂SO₄, phosphate-buffered saline (PBS; pH 7), and dialysis cassettes (10K molecular weight cutoff) were purchased from Fisher Scientific (part of Thermo Fisher Scientific, Waltham, MA, USA). Sodium 4-aminosalicylate and acryloyl chloride were purchased from Alfa Aesar (Tewksbury, MA, USA). 4,4′-Azobis(4-cyanovaleric acid) and acetic anhydride were purchased from Sigma-Aldrich (St. Louis, MO, USA). NMR spectroscopy was performed with a 600 MHz NMR spectrometer (Bruker Biospin, Corp., Billerica, MA, USA). LC-MS studies were conducted using a Waters Acquity ultra performance liquid chromatography (UPLC) MS system with an Acquity UPLC BEH C-18 1.7 μm, 2.1 mm × 50 mm column, UV detection between 200 and 400 nm (2996 PDA detector), and a TQ detector mass spectrometer with electrospray ionization (ESI).

Chemical Synthesis of 4-Acetamido-2-hydroxybenzoic acid.

4-Aminosalicylic acid (1.53 g, 10.01 mmol) was dissolved in anhydrous acetone (30 mL) under argon. Then the solution was cooled using an ice bath. Acetic anhydride (1 mL, 10.58 mmol) was added while stirring, and the reaction mixture was allowed to stir for 24 h. The organic solvent was evaporated under vacuum, and a solid residue was washed with water and filtered. The resulting compound was resuspended in a small amount of water and lyophilized to yield the dry final compound (white powder, 1.46 g, 75%). ¹H NMR (600 MHz, DMSO-d₆) δ (ppm): δ 13.58 (s, 1H), 11.39 (s, 1H), 10.19 (s, 1H), 7.71 (d, J = 8.7 Hz, 1H), 7.35 (d, J = 2.0 Hz, 1H), 7.04 (dd, J = 8.6, 2.1 Hz, 1H), 2.07 (s, 3H). LRMS-ESI (m/z): [M + H]⁺ calcd for [C₉H₉NO₄], 196.05; found, 196.01

Chemical Synthesis of 4-Acrylamidosalicylic Acid.

Sodium 4-aminosalicylate (4 g, 0.018 mol) was dissolved in ice-cold distilled water (50 mL) and stirred for 1 h at 5 °C. Then acryloyl chloride (2.35 mL; 0.029 mol) was added as two separate additions (1.5 and 0.85 mL), and the solution was stirred for 1 h after each addition at 5 °C. A white precipitate formed. The precipitate was filtered and extensively washed with distilled water (200 mL). Then the product was lyophilized to obtain dry 4-acrylamidosalicylic acid (white powder, 3.12 g, 78%). ¹H NMR (600 MHz, DMSO-d₆) δ: 13.70 (s, 1H), 11.35 (s, 1H), 10.38 (s, 1H), 7.73 (d, J = 8.7 Hz, 1H), 7.50–7.39 (m, 1H), 7.12 (dd, J = 8.7, 2.1 Hz, 1H), 6.44 (dd, J = 17.0, 10.2 Hz, 1H), 6.29 (dd, J = 17.0, 2.0 Hz, 1H), 5.81 (dd, J = 9.9, 1.7 Hz, 1H). LRMS-ESI (m/z): [M + H]⁺ calcd for [C₁₀H₉NO₄], 208.05; found, 208.06

Chemical Synthesis of the Polymer Agent.

4-Acrylamidosalicylic acid (1.0 g, 0.0048 mol) and 4,4′-azobis(4-cyanovaleric acid) (0.05 g, 0.18 mmol, 5% w/w) were dissolved in DMF (25 mL) under argon and stirred at 70 °C for 24 h. Then the reaction mixture was diluted with water and a white precipitate formed. Then the reaction mixture was dialyzed with a 10 000 molecular weight cutoff filter against water and lyophilized to obtain the final polymer. This final polymer agent was analyzed by GPC and MALDI-TOF.

Gel Permeation Chromatography.

The molecular weight of the polymer agent was determined using an Agilent high-performance liquid chromatography system with a UV detector and a Phenomenex SEC-S-3000 column (300 × 7.8 mm). PBS was used as an eluent with 0.5 mL/min flow rate. Sodium polystyrenesulfonate standards were employed for instrument calibration.

MALDI-TOF.

The MALDI-TOF spectrum of the polymer agent was recorded with a Bruker Autoflex Speed MALDI-TOF spectrophotometer, using 2,5-dihydroxybenzoic acid (DHB) as a matrix. The matrix and polymer were dissolved in a solvent system of 50:50 water:acetonitrile with 0.1% TFA. Samples were loaded onto a target plate and mixed on the target with 1 μL of supernatant of saturated matrix solution. All MALDI spectra were acquired in linear mode, which operated over a mass range from 500 to 75 000 Da. The number of shots and laser power were adjusted to improve spectrum quality.

UV−Vis Spectroscopy.

Monomer and polymer agents were dissolved in 0.01 mM PBS, and UV spectral scans were taken between 230 and 400 nm using the Synergy H4 microplate reader (BioTek Instruments, Inc., Winooski, VT, USA).

SEM.

The polymer agent was dissolved in 0.1 M PBS, and a drop was placed onto a 12 mm diameter glass coverslip and allowed to dry overnight. The dissolved polymer agent and the dried polymer agent were both mounted onto double-stick carbon tabs (Ted Pella. Inc., Redding, CA, USA), which have been previously mounted onto a glass microscope slide. The samples were then coated under vacuum using a Balzer MED 010 evaporator (Technotrade International, Manchester, NH, USA) with platinum alloy for a thickness of 25 nm, then immediately flash carbon coated under vacuum. The samples were transferred to a desiccator for storage. Samples were examined/imaged in a JSM-5910 scanning electron microscope (JEOL, USA, Inc., Peabody, MA, USA) at an accelerating voltage of 15 kV. These results were analyzed to determine particle size.

TEM.

The polymer agent was dissolved in 0.1 M PBS, and a drop was placed onto 100 mesh Formvar/carbon-coated copper grids treated with poly l-lysine for approximately 1 h. The sample was blotted from the grid with filter paper and was allowed to dry. The sample was then examined in a JEM 1010 transmission electron microscope (JEOL, USA, Inc.) at an accelerating voltage of 80 kV. Digital images were obtained using the AMT Imaging System (Advanced Microscopy Techniques Corp., Danvers, MA, USA).

Cytotoxicity Assay.

The cytotoxicity evaluations of the monomer and polymer agents were performed using an MTT cell growth assay kit (EMD Millipore Chemicon). Approximately 1 × 10⁵ cells mL⁻¹ of MDA-MB-231 mammary carcinoma cells in their exponential growth phase were seeded in a flat-bottomed 96-well polystyrene-coated plate and were incubated for 24 h at 37 °C in a 5% CO₂ incubator. Then media was removed and 100 μL of media was added to the plate that included a range of concentrations (0.125, 0.625, 1.25, 6.25, 12.5, and 62.5 mM) of the monomer or the “monomeric unit concentration” of the included polymer. After 24 h of incubation at 37 °C, the media was removed and 10 μL of MTT reagent was added to each well containing fresh media and incubated at 37 °C for 4 h. Afterward, formazan crystals that formed in each well were dissolved in isopropanol with 100 μL of 0.04 N HCl and plates were read at 570 nm in a microplate reader (SYNERGY H4 reader, Bio Tek Instruments, Inc., VT, USA). Wells without cells, or only with cells, or with cells and MTT reagent were used as control samples.

AcidoCEST MRI of Chemical Samples.

The monomer and polymer agents were prepared at 1.25–40 mM and 95–300 μM concentrations in 10 mM PBS, respectively. The pH was adjusted using a benchtop pH meter and very small volumes of HCl and NaOH that did not significantly affect agent concentrations (Mettler Toledo Seven compact S220 pH meter, Columbus, OH, USA). Samples were placed in a customized holder (Figure S1a,b) that maintained the samples at 37.0 ± 0.2 °C or at 25.0, 29.0, 34.0, 42.0, 45.0, or 49.0 °C during temperature-dependent studies. Our customized sample holder was filled with 4% agarose to reduce B₁ and B₀ inhomogeneities. We incubated our samples and holder at the desired temperature for 4–18 h prior to each MRI study. We used warm air and an automated temperature feedback system to maintain temperature in the MRI magnet (SA Instruments, Stony Brook, NY, USA). We validated our sample temperatures by using the same holder to analyze samples of ethylene glycol, using localized MR spectroscopy, which is an ideal method for measuring the temperature of chemical solutions without disturbing the solutions in the MRI magnet.⁹⁵

MRI studies were performed with a Bruker Biospec MRI scanner operating at 7 T (300 MHz) magnetic field strength with a 72 mm volume transceiver coil (Bruker Biospin, Inc., Billerica, MA, USA). To identify the location of the samples in the magnet, an initial set of images was acquired using a multislice spin echo MRI protocol with the following parameters: 750 ms repetition time (TR); 10 ms echo time (TE); 180° excitation flip angle (FA); coronal image orientation; 64 × 64 matrix; 5.12 × 5.12 cm field-of-view (FOV); 0.8 × 0.8 mm in-plane spatial resolution; 1 mm slice thickness (SL); 40 slices; 1 average; 48 s total scan time. We used a CEST-FISP (fast imaging with steady-state free precession) MRI acquisition protocol with the following parameters: 4.9832 ms TR; 2.2916 ms TE; 30° FA; centric encoding; coronal image orientation; 256 × 256 matrix; 5.12 × 5.12 cm FOV; 0.2 × 0.2 mm in-plane spatial resolution; 1 mm SL; 1 slice; and 1 average.⁹⁶ We acquired CEST images with selective saturation applied at 4 μT for 6 s at saturation frequencies in 0.1 ppm increments from −13 to 13 ppm, for a total scan time of 28.0 min. We also collected images at 0.5 to 6.0 μT saturation powers (at 6 s saturation time) and at 1–8 s saturation times (at 4 μT saturation power).

Images were processed using ParaVision v5.1, and CEST MR image analyses were performed with Matlab 2018b (The MathWorks, Inc., Natick, MA, USA). To measure CEST signal amplitudes of the monomer and polymer, each CEST spectrum from a region of interest (ROI) was analyzed by fitting the spectrum to a sum of three Lorentzian line shapes to account for two CEST signals (5.0 and 9.2 ppm) and direct saturation of water.⁹⁷ The same procedure was used to measure CEST signal amplitudes of iopamidol, using a sum of five Lorentzian line shapes the account for two CEST signals from amide groups (4.2 and 5.6 ppm), two hydroxyl groups (0.8 and 1.8 ppm), and direct saturation of water. The center, width, and amplitude of each Lorentzian line were allowed to change in order to obtain an optimal fit.

in Vivo AcidoCEST MRI Studies.

All in vivo mouse studies were conducted with the approval of the Institutional Animal Care and Use Committee of the MD Anderson Cancer Center in accordance with applicable guidelines and regulations. A subcutaneous model of MDA-MB-231 mammary carcinoma was developed by injecting 10⁶ tumor cells in the right rear flank of eight 6–8-week-old female mice. Tumors were allowed to grow for 5–6 weeks to a diameter of 3–6 mm before imaging studies. Two mice were analyzed with acidoCEST MRI three times during sequential scans performed on day 0, 2, and 4 (where day 0 indicates the initial scan for the mouse). These sequential scans used iopamidol on day 0, the monomer agent on day 2, and the polymer agent on day 4. Six additional mice were analyzed with acidoCEST MRI only one time, using only iopamidol for two mice, the monomer agent for two mice, and the polymer agent for two mice.

To prepare for each imaging scan, a mouse was anesthetized with 1.5–2.0% isoflurane in 100% oxygen carrier gas. A tail vein was catheterized with a 27 G needle. The mouse was then placed in a customized cradle, and respiratory and rectal temperature leads were positioned to monitor respiration and core body temperature throughout the imaging scan (SA Instruments Inc.). Body temperature was maintained at 37.0 ± 0.2 °C using warm air. To localize the tumor within an image, we performed anatomical MRI scans with each mouse using the multislice spin− echo MRI protocol with the following parameters: 1000 ms TR; 5.42 ms TE; 90° FA; axial image orientation; 128 × 128 matrix; 3.84 × 3. 84 cm FOV; 0.3 × 0.3 mm in-plane spatial resolution; 1 mm SL; 30 slices; 1 average; 2:08 min total scan time.

In vivo acidoCEST MRI was performed using a CEST-FISP protocol with the following parameters: 3.310 ms TR; 1.655 ms TE; 15° FA; centric encoding; axial image orientation; 128 × 128 matrix; 3.84 × 3.84 cm FOV; 0.3 × 0.3 mm in-plane spatial resolution; 1 mm SL; 1 slice; 1 average. Four image sets were initially acquired to produce CEST spectra for each image pixel prior to intravenous administration of the contrast agent. Then, 200 μL of 976 mM iopamidol, 250 mM monomer agent, or 2 mM polymer agent (equivalent to 250 mM in monomer units) were injected intravenously. These injection concentrations equate to 7.8, 2.0, and 0.016 mmol/kg of mouse body weight, respectively. The injection line was connected to an infusion pump that infused 400 μL/h of agent during the next 30 min. Six sets of CEST MR images were acquired immediately after injection to produce six postinjection CEST spectra for each imaging pixel. We acquired CEST images with selective saturation applied at 3 μT for 6 s at saturation frequencies of 14.0 ppm (which was subsequently discarded during analysis); 13.0 to 3.0 ppm in increments of 0.4 ppm; 2.5 to −2.0 ppm in increments of 0.5 ppm; and −3.0 to −12.0 ppm in increments of 3 ppm for a total of 40 saturation frequencies that required a total scan time of 4:23 min.

CEST MR images were analyzed using our previously reported methods.^23–25 Images were averaged that were acquired before injection of agent and at the same saturation frequency. Postinjection images at the same saturation frequency were also averaged. Each averaged image was processed by Gaussian spatial smoothing to improve signal-to-noise. These processed preinjection images were subtracted from the processed postinjection images to remove endogenous CEST signals that are the same before and after injection. The resulting CEST spectra for each pixel of the monomer and polymer were fit with the sum of three Lorentzian line shapes to measure the amplitude of CEST signals at 5.0 and 9.2 ppm and to account for the direct saturation of water at 0 ppm. The same procedure was used to measure CEST signal amplitudes of iopamidol, using a sum of five Lorentzian line shapes to account for two CEST signals from amide groups (4.2 and 5.6 ppm), two hydroxyl groups (0.8 and 1.8 ppm), and direct saturation of water. The pH was estimated based on a log₁₀ ratio of the two CEST signals (Figure 4 for the monomer and polymer agents and ref 17 for iopamidol).