**1. Introduction**

In tumor diagnosis and treatment, superparamagnetic iron oxide (SPIO) nanoparticles play a unique and important role, because they possesses versatile applications for clinical diagnosis and tumor adjunctive therapy [1,2]. For tumor diagnosis, SPIO nanoparticles have been used widely as a contrast agent (CA) in magnetic resonance imaging (MRI), as it could improve the contrast in anatomical imaging to highlight the situation and structure of a tumor by shortening the spin−spin relaxation time (*T*2) of the proton [3]. For tumor therapy, magnetic hyperthermia (MH) is a noninvasive hyperthermia that inhibits tumor growth by the Brownian relaxation and Néel relaxation of SPIO nanoparticles under an alternating magnetic field (AMF) [4]. Furthermore, many studies also found that SPIO nanoparticles could induce apoptosis of tumor cells directly by producing ferrous ions, which can generate toxic reactive oxygen species (ROS) by the Fenton reaction (FR) [5]. Although SPIO nanoparticles present much potential, their effectiveness in tumor diagnosis and therapy still depends on their accumulation in the tumor, which is similar to that of a chemotherapeutic drug. Learning from the progress of stimuli-responsive polymeric nanocarriers for tumor targeting [6,7], the targeted accumulation of SPIO nanoparticles in a tumor could also be improved by constructing stimuli-responsive magnetic nanofluid (MNF) through the self-assembly of SPIO nanoparticles and stimuli-responsive polymers.

**Citation:** Sheng, L.; Zhu, X.; Sun, M.; Lan, Z.; Yang, Y.; Xin, Y.; Li, Y. Tumor Microenvironment-Responsive Magnetic Nanofluid for Enhanced Tumor MRI and Tumor multi-treatments. *Pharmaceuticals* **2023**, *16*, 166. https://doi.org/ 10.3390/ph16020166

Academic Editor: Huijie Zhang

Received: 1 December 2022 Revised: 12 January 2023 Accepted: 18 January 2023 Published: 23 January 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Because tumor angiogenesis and aerobic glycolysis have been recognized as features of most malignant solid tumors, regardless of their tissue origin or genetic background [8], pH-sensitive nanocarriers have attracted tremendous interest in tumor diagnosis and subsequent treatment over the past decades [9,10]. In order to construct pH-sensitive nanocarriers, many types of pH-sensitive polymers have been designed and synthesized, as they possess ionizable basic or acidic residues [11], resulting in varied physicochemical properties (solubility or chain conformation) with a change in the surrounding pH [12]. However, slight pH differences between a tumor stroma (pH = 7.1–6.7) [13] and the physiological condition (pH = 7.35–7.45) is impossible to be recognized by most pH-sensitive deliveries, which usually consist of pH-sensitive polymers and occur during phase transition in endocytic organelles with a lower pH value (pH ≤ 6.0), such as endosomes and lysosomes [9,10]. There are only a few pH-sensitive polymers with dissociation constants (p*K*a) around neutral pH, such as cationic polymers with repeating ionizable tertiary amines [14,15] or anionic polymers with suitable sulfonamide groups [16,17]. Until now, the most successful pH-sensitive polymers used in targeting tumor microenvironment were cationic polymers with ionizable tertiary amine blocks, showing a hydrophobic–hydrophilic phase transition in weakly acidic microenvironments [18,19], which disassemble rapidly in tumor microenvironment, leading to their application in targeted tumor chemotherapy [20] and enhanced tumor fluorescence imaging [19]. Unfortunately, ionized cationic polymer exhibit a potential risk of hemolysis in vivo, as its positive charge could easily damage membranes of red blood cells [21]. Therefore, a sulfonamide-based anionic polymer with suitable p*K*<sup>a</sup> within a tumor microenvironment pH value (7.1–6.7) should be an optimized option to prepare MNF with tumor microenvironment responsiveness.

According to the property of an anionic polymer, its hydrophilic–hydrophobic phase transition can be triggered under a certain pH value, which is usually lower than the p*K*<sup>a</sup> of anionic polymer [11]. As a result, a pH-responsive MNF with a suitable p*K*<sup>a</sup> (7.1–6.7) should lose its colloidal stability in tumor stroma, resulting in the efficient accumulation and retention of SPIO in tumor tissue. Furthermore, tumor tissue had existing amounts of endogenous hydrogen ion (H<sup>+</sup> ) and hydrogen peroxide (H2O2). Due to the phase transition of anionic polymers in tumor stroma, encapsulated SPIO nanoparticles obtained the opportunity to interact with surrounding H<sup>+</sup> , resulting in the release of ferrous ion (Fe2+). Next, the endogenous H2O<sup>2</sup> in the tumor tissue could be decomposed under the catalysis of Fe2+, implying the possibility of FR in the tumor microenvironment. When the phase transition of anionic polymer blocks occurred completely, stranded SPIO nanoparticles could form many aggregations with a large size spontaneously. According to previous studies, a closed packing structure of the multiple SPIO nanoparticles exhibited its attractive effects on improving the negative signal contrast of pathological tissue [22] and enhanced the efficiency of MH [23] simultaneously. Therefore, sulfonamide-based MNF with a pH responsiveness not only possesses a specific advantage in tumor targeting but also shows other potentials for improving the sensitivity of tumor MRIs and enhancing the antitumor efficacy by a combination of MH- and FR-mediated chemodynamic therapy (CDT).

In this study, we designed and fabricated a tumor microenvironment-responsive MNF, which maintained stability in blood vessels and formed aggregations in tumor microenvironment to enhance tumor MH, ROS generation and MRIs simultaneously, as shown in Figure 1.

In order to prepare the tumor microenvironment-responsive MNF, we synthesized a sulfonamide-based amphiphilic copolymer, polycaprolactone-*b*-poly(sulfadimethoxine acrylamide) (PCL-*b*-pSMA) by reversible addition−fragmentation chain transfer (RAFT) polymerization, according to relevant studies [24,25], which showed p*K*<sup>a</sup> of approximately 7.0. Then, the MNF was fabricated by the simple self-assembly of the PCL-*b*-pSMA and hydrophobic Fe3O<sup>4</sup> nanoparticles. Due to the fact of the pH sensitivity of PCL-*b*-pSMA at a neutral pH value, the MNF in the aqueous phase displayed a similar pH responsiveness at a neutral pH value (≈7.0). According to the pH sensitivity of the MNF on neutral medium, we further investigated its application potentials for MH, FR and MRI for tumor theranos-

tics. The relevant results show that the MNF under neutral conditions exhibited a better performance in enhancing the specific absorption rate (SAR), increasing the generation of hydroxyl radical (•OH) and improving the *T*<sup>2</sup> relaxivity (*r*2), simultaneously, compared to its counterpart in a physiological environment. Based on these advantages, we further studied the effects of the tumor microenvironment-responsive MNF on MH-induced cell death and ROS generation under different strengths of AMF (*H*applied). As the intercellular ROS level and cell mortality rate under MH showed a high degree of correlation, this study suggests that the MNF can stimulate CDT by MH and inhibit cell proliferation efficiently by integrating MH and CDT effectively. Therefore, the tumor microenvironment-responsive MNF possesses a versatile potential for tumor targeting, diagnosis and treatment. *Pharmaceuticals* **2023**, *16*, x FOR PEER REVIEW 3 of 17

**Figure 1.** Schematic illustration of the tumor microenvironment-responsive MNF, which was prepared by self-assembly of Fe3O4 nanoparticles and pH-responsive polymer. According to the phase transition of the pH-responsive micelles under a neutral condition, the tumor microenvironmentresponsive MNF could maintain an individual state and form aggregations in tumor microenvironment to improve tumor MH, ROS generation and MRI simultaneously. **Figure 1.** Schematic illustration of the tumor microenvironment-responsive MNF, which was prepared by self-assembly of Fe3O<sup>4</sup> nanoparticles and pH-responsive polymer. According to the phasetransition of the pH-responsive micelles under a neutral condition, the tumor microenvironmentresponsive MNF could maintain an individual state and form aggregations in tumor microenvironment to improve tumor MH, ROS generation and MRI simultaneously.

#### In order to prepare the tumor microenvironment-responsive MNF, we synthesized a **2. Results and Discussion**

#### sulfonamide-based amphiphilic copolymer, polycaprolactone-*b*-poly(sulfadimethoxine *2.1. Synthesis and Characterization of the pH-Responsive Amphiphilic Copolymer*

acrylamide) (PCL-*b*-pSMA) by reversible addition−fragmentation chain transfer (RAFT) polymerization, according to relevant studies [24,25], which showed p*K*a of approximately 7.0. Then, the MNF was fabricated by the simple self-assembly of the PCL-*b*-pSMA and hydrophobic Fe3O4 nanoparticles. Due to the fact of the pH sensitivity of PCL-*b*-pSMA at a neutral pH value, the MNF in the aqueous phase displayed a similar pH responsiveness at a neutral pH value (≈7.0). According to the pH sensitivity of the MNF on neutral medium, we further investigated its application potentials for MH, FR and MRI for tumor theranostics. The relevant results show that the MNF under neutral conditions exhibited a better performance in enhancing the specific absorption rate (SAR), increasing the generation of hydroxyl radical (•OH) and improving the *T*2 relaxivity (*r*2), simultaneously, compared to its counterpart in a physiological environment. Based on these advantages, As an anionic polymer usually possesses hydrophilicity in neutral and physiological conditions, we selected the polycaprolactone (PCL) segment as the hydrophobic block of the amphiphilic copolymer because of its high biocompatibility. The synthetic route, as shown in Figure S1 (Supplementary Materials), contained ring opening polymerization (ROP), an esterification reaction and RAFT polymerization simultaneously. By the ROP, the PCL was synthesized successfully, which was confirmed by the H proton nuclear magnetic resonance (1H NMR) spectrum, as shown in Figure S2. S-1-dodecyl-S'-(a,a'-dimethyla <sup>00</sup>-acetic acid)trithiocarbonate (DDMAT) was used as the chain transfer agent (CTA) forthe RAFT polymerization. The following product was PCL-DDMAT by the esterification reaction between PCL and DDMAT, which was also confirmed by its structure using the<sup>1</sup>H NMR spectrum, as shown in Figure S3.

we further studied the effects of the tumor microenvironment-responsive MNF on MHinduced cell death and ROS generation under different strengths of AMF (*H*applied). As the intercellular ROS level and cell mortality rate under MH showed a high degree of correlation, this study suggests that the MNF can stimulate CDT by MH and inhibit cell proliferation efficiently by integrating MH and CDT effectively. Therefore, the tumor microenvironment-responsive MNF possesses a versatile potential for tumor targeting, diagnosis and treatment. **2. Results and Discussion**  *2.1. Synthesis and Characterization of the pH-Responsive Amphiphilic Copolymer*  As an anionic polymer usually possesses hydrophilicity in neutral and physiological The final reaction was the RAFT polymerization for the preparation of the pHresponsive amphiphilic copolymer. In this study, we selected sulfadimethoxine acrylamide (SMA) as the pH-sensitive monomer, because the poly(methacryloyl sulfadimethoxine) showed a p*K*<sup>a</sup> at approximately 7.0 in a relevant study [24]. Utilizing DMSO-*d*<sup>6</sup> as a solvent, we characterized the structures of the PCL-*b*-pSMA and SMA, as shown in Figure 2. According to previous studies [24,26,27], all of the characteristic peaks of the SMA were identified and marked in Figure 2A (bottom spectrum). Based on the <sup>1</sup>H NMR results of the SMA (Figure 2A, bottom spectrum) and PCL-DDMAT (Figure S3), we further identified all of the characteristic peaks of the PCL-*b*-pSMA, also shown in Figure 2A (top spectrum). Apparently, because of the polymerization of SMA, the characteristic peaks of the acrylamide in the SMA disappeared; meanwhile, all of the characteristic peaks of the SM broadened. In

conditions, we selected the polycaprolactone (PCL) segment as the hydrophobic block of

(ROP), an esterification reaction and RAFT polymerization simultaneously. By the ROP, the PCL was synthesized successfully, which was confirmed by the H proton nuclear magnetic resonance (1H NMR) spectrum, as shown in Figure S2. S-1-dodecyl-S'-(a,a'-dimethyla"-acetic acid)trithiocarbonate (DDMAT) was used as the chain transfer agent (CTA) for the RAFT polymerization. The following product was PCL-DDMAT by the esterification reaction between PCL and DDMAT, which was also confirmed by its structure using the

1H NMR spectrum, as shown in Figure S3.

addition to the characteristic peaks of SM, the characteristic peaks of PCL can also observed in Figure 2A (top spectrum), which confirms the successful polymerization of SMA as a product of the PCL-DDMAT. can also observed in Figure 2A (top spectrum), which confirms the successful polymerization of SMA as a product of the PCL-DDMAT.

The final reaction was the RAFT polymerization for the preparation of the pH-responsive amphiphilic copolymer. In this study, we selected sulfadimethoxine acrylamide (SMA) as the pH-sensitive monomer, because the poly(methacryloyl sulfadimethoxine) showed a p*K*a at approximately 7.0 in a relevant study [24]. Utilizing DMSO-*d*6 as a solvent, we characterized the structures of the PCL-*b*-pSMA and SMA, as shown in Figure 2. According to previous studies [24,26,27], all of the characteristic peaks of the SMA were identified and marked in Figure 2A (bottom spectrum). Based on the 1H NMR results of the SMA (Figure 2A, bottom spectrum) and PCL-DDMAT (Figure S3), we further identified all of the characteristic peaks of the PCL-*b*-pSMA, also shown in Figure 2A (top spectrum). Apparently, because of the polymerization of SMA, the characteristic peaks of the acrylamide in the SMA disappeared; meanwhile, all of the characteristic peaks of the SM broadened. In addition to the characteristic peaks of SM, the characteristic peaks of PCL

*Pharmaceuticals* **2023**, *16*, x FOR PEER REVIEW 4 of 17

**Figure 2.** Structure and pH responsive of the PCL-*b*-pSMA: (**A**) 1H NMR spectrum of PCL-*b*-pSMA, marking all characteristic peaks; (**B**) pH-dependent transmittance curve and diameter variation of PCL-*b*-pSMA, showing a p*K*a at 7.0. **Figure 2.** Structure and pH responsive of the PCL-*b*-pSMA: (**A**) <sup>1</sup>H NMR spectrum of PCL-*b*-pSMA, marking all characteristic peaks; (**B**) pH-dependent transmittance curve and diameter variation of PCL-*b*-pSMA, showing a p*K*a at 7.0.

After researching the structure of the PCL-*b*-pSMA, we characterized the p*K*a of the PCL-*b*-pSMA, as shown in Figure 2B. According to relevant studies [28,29], the cloud point method was utilized to determine the p*K*a of the PCL-*b*-pSMA, as shown in Figure 2B (red line), by quantifying the turbidity of the PCL-*b*-pSMA under different pH buffers at 500 nm. We prepared the micelles of the PCL-*b*-pSMA in an alkaline solution first; then, we observed its light transmittance using a UV-Vis spectrophotometer under a decreasing pH from 9.13 to 5.18. Apparently, the transmittance of the PCL-*b*-pSMA micelles with a high concentration (≈3 mg mL−1) was influenced by the environmental pH value. In the alkaline environment, the PCL-*b*-pSMA micelles showed almost 100% transmittance. When the pH value decreased to 7.5, its transmittance decreased slightly. However, the turbidity increased sharply with the decrease of the pH from 7.27 to 6.88, and the corresponding light transmittance decreased from 87.7% to 29.3%. When the pH value decreased further (≤6.59), the PCL-*b*-pSMA formed obvious sediment, and the corresponding light transmittance was near 0%. As PCL is a typical hydrophobic polymer, the high After researching the structure of the PCL-*b*-pSMA, we characterized the p*K*<sup>a</sup> of the PCL-*b*-pSMA, as shown in Figure 2B. According to relevant studies [28,29], the cloud point method was utilized to determine the p*K*<sup>a</sup> of the PCL-*b*-pSMA, as shown in Figure 2B (red line), by quantifying the turbidity of the PCL-*b*-pSMA under different pH buffers at 500 nm. We prepared the micelles of the PCL-*b*-pSMA in an alkaline solution first; then, we observed its light transmittance using a UV-Vis spectrophotometer under a decreasing pH from 9.13 to 5.18. Apparently, the transmittance of the PCL-*b*-pSMA micelles with a high concentration (≈3 mg mL−<sup>1</sup> ) was influenced by the environmental pH value. In the alkaline environment, the PCL-*b*-pSMA micelles showed almost 100% transmittance. When the pH value decreased to 7.5, its transmittance decreased slightly. However, the turbidity increased sharply with the decrease of the pH from 7.27 to 6.88, and the corresponding light transmittance decreased from 87.7% to 29.3%. When the pH value decreased further (≤6.59), the PCL-*b*-pSMA formed obvious sediment, and the corresponding light transmittance was near 0%. As PCL is a typical hydrophobic polymer, the high transmittance of the PCL-*b*-pSMA depends on the hydrophilicity of the pSMA in an alkaline solution. When the solution pH value downregulated from a weak alkaline to faintly acid, the increasing turbidity of the PCL-*b*-pSMA micelles indicated the rapid phase transition of the pSMA from hydrophilicity to hydrophobicity.

transmittance of the PCL-*b*-pSMA depends on the hydrophilicity of the pSMA in an alkaline solution. When the solution pH value downregulated from a weak alkaline to faintly acid, the increasing turbidity of the PCL-*b*-pSMA micelles indicated the rapid phase transition of the pSMA from hydrophilicity to hydrophobicity. The phase transition of the pSMA not only decreased the transmittance of the PCL*b*-pSMA micelles at the macro level but also reduced their stability at the micro level, which could be observed by dynamic light scattering (DLS), which was assessed under a low concentration of the PCL-*b*-pSMA micelles (0.3 mg mL−1). As shown in Figure 2B (gray line), the DLS result exhibited the effect of the pH value on the PCL-*b*-pSMA micelle's hydrated diameter. According to the result, the hydrated diameter of the PCL-*b*-pSMA micelles in an alkaline solution decreased slightly from 64.7 (pH ≈ 9.21) to 62.9 nm (pH ≈ The phase transition of the pSMA not only decreased the transmittance of the PCL-*b*pSMA micelles at the macro level but also reduced their stability at the micro level, which could be observed by dynamic light scattering (DLS), which was assessed under a low concentration of the PCL-*b*-pSMA micelles (0.3 mg mL−<sup>1</sup> ). As shown in Figure 2B (gray line), the DLS result exhibited the effect of the pH value on the PCL-*b*-pSMA micelle's hydrated diameter. According to the result, the hydrated diameter of the PCL-*b*-pSMA micelles in an alkaline solution decreased slightly from 64.7 (pH ≈ 9.21) to 62.9 nm (pH ≈ 7.43) with the decreasing pH value. When the surrounding pH value decreased to 7.05, the diameter of the PCL-*b*-pSMA micelles increased sharply to 198.9 nm. With the pH decreasing further (pH ≈ 6.85), the particle size of the PCL-*b*-pSMA micelles increased to 379.4 nm, indicating the agglomeration of the PCL-*b*-pSMA micelles. Finally, the diameter of these agglomerated micelles reached almost 600 nm in an acidic environment, which corresponded to the very low transmittance of its counterpart, with 10 times the concentration. According to the results shown in Figure 2B, we estimated the p*K*<sup>a</sup> value of the PCL-*b*-pSMA at 7.0, because the PCL-*b*-pSMA micelles presented 50% transmittance at that pH value.

Furthermore, we observed the PCL-*b*-pSMA micelles under a physiological and neutral pH value directly by transmission electron microscopy (TEM), as shown in Figure 3A,B. By the negative staining of phosphotungstic acid, many bright spheres could easily be

that pH value.

observed, which represented the inner core of the PCL-*b*-pSMA micelles. Apparently, the PCL-*b*-pSMA micelles in the buffer with a physiological pH value displayed high colloidal stability, because their inner cores (bright spheres) could be separated from each other by a hydrophilic shell composed of pSMA. On the contrary, in the neutral buffer (pH ≈ 7.05), the PCL-*b*-pSMA micelles did not display a larger inner core, but also collected together spontaneously, as shown in Figure 3B, indicating that the colloidal stability of the PCL-*b*-pSMA micelles was broken. be observed, which represented the inner core of the PCL-*b*-pSMA micelles. Apparently, the PCL-*b*-pSMA micelles in the buffer with a physiological pH value displayed high colloidal stability, because their inner cores (bright spheres) could be separated from each other by a hydrophilic shell composed of pSMA. On the contrary, in the neutral buffer (pH ≈ 7.05), the PCL-*b*-pSMA micelles did not display a larger inner core, but also collected together spontaneously, as shown in Figure 3B, indicating that the colloidal stability of the PCL-*b*-pSMA micelles was broken.

Furthermore, we observed the PCL-*b*-pSMA micelles under a physiological and neutral pH value directly by transmission electron microscopy (TEM), as shown in Figure 3A,B. By the negative staining of phosphotungstic acid, many bright spheres could easily

7.43) with the decreasing pH value. When the surrounding pH value decreased to 7.05, the diameter of the PCL-*b*-pSMA micelles increased sharply to 198.9 nm. With the pH decreasing further (pH ≈ 6.85), the particle size of the PCL-*b*-pSMA micelles increased to 379.4 nm, indicating the agglomeration of the PCL-*b*-pSMA micelles. Finally, the diameter of these agglomerated micelles reached almost 600 nm in an acidic environment, which corresponded to the very low transmittance of its counterpart, with 10 times the concentration. According to the results shown in Figure 2B, we estimated the p*K*a value of the PCL-*b*-pSMA at 7.0, because the PCL-*b*-pSMA micelles presented 50% transmittance at

*Pharmaceuticals* **2023**, *16*, x FOR PEER REVIEW 5 of 17

**Figure 3.** Morphologies (TEM) and particle size distributions (DLS) of PCL-*b*-pSMA micelles in the aqueous phase (0.3 mg mL<sup>−</sup>1) with different pH values: (**A**) TEM result of PCL-*b*-pSMA micelles in the buffer with a physiological pH value (pH = 7.43); (**B**) TEM result of PCL-*b*-pSMA micelles in neutral buffer (pH = 7.05); (**C**) DLS results of the samples in (**A**,**B**). **Figure 3.** Morphologies (TEM) and particle size distributions (DLS) of PCL-*b*-pSMA micelles in the aqueous phase (0.3 mg mL−<sup>1</sup> ) with different pH values: (**A**) TEM result of PCL-*b*-pSMA micelles in the buffer with a physiological pH value (pH = 7.43); (**B**) TEM result of PCL-*b*-pSMA micelles in neutral buffer (pH = 7.05); (**C**) DLS results of the samples in (**A**,**B**).

Based on the TEM results, we investigated these TEM samples again by digital photos and DLS again; all are shown in Figure 3C. It was clear that the PCL-*b*-pSMA micelles in neutral solution looked similar to a milk solution; meanwhile, its counterpart in the physiological buffer with the same concentration showed excellent transmittance. Corresponding to these photos, the PCL-*b*-pSMA micelles in the physiological buffer in the neutral solution presented a smaller diameter (62.9 nm) and a narrower particle size distribution (PDI = 0.141) than their counterparts in the neutral solution, which explains the high transmittance of the former and the low transmittance of the latter. Based on the TEM results, we investigated these TEM samples again by digital photos and DLS again; all are shown in Figure 3C. It was clear that the PCL-*b*-pSMA micelles in neutral solution looked similar to a milk solution; meanwhile, its counterpart in the physiological buffer with the same concentration showed excellent transmittance. Corresponding to these photos, the PCL-*b*-pSMA micelles in the physiological buffer in the neutral solution presented a smaller diameter (62.9 nm) and a narrower particle size distribution (PDI = 0.141) than their counterparts in the neutral solution, which explains the high transmittance of the former and the low transmittance of the latter.

Due to the fact of the results of the transmittance, DLS, TEM and digital photos, we prepared an anionic pH-responsive amphiphilic copolymer with a p*K*a of approximately Due to the fact of the results of the transmittance, DLS, TEM and digital photos, we prepared an anionic pH-responsive amphiphilic copolymer with a p*K*<sup>a</sup> of approximately 7.0, which should be suitable for the preparation of a tumor microenvironment-responsive MNF.

#### *2.2. Characterization of the Tumor Microenvironment -Responsive MNF*

In order to prepare the tumor microenvironment-responsive MNF, we firstly synthesized hydrophobic Fe3O<sup>4</sup> nanoparticles at approximately 8 nm. After, the MNF was prepared by the simple self-assembly between the amphiphilic PCL-*b*-pSMA and hydrophobic Fe3O<sup>4</sup> nanoparticles. It was clear that the Fe3O<sup>4</sup> nanoparticles in hexane (0.1 mg mL−<sup>1</sup> ) presented monodispersity with a uniform particle size, as shown in Figure 4A, due to the fact of their high hydrophobicity. In order to improve the water dispersibility of the hydrophobic Fe3O<sup>4</sup> nanoparticles, the polymeric micelles provided a powerful platform to load the hydrophobic molecules into their hydrophobic cores. Therefore, we prepared the MNF by the self-assembly of the hydrophobic Fe3O<sup>4</sup> nanoparticles and the amphiphilic

MNF.

PCL-*b*-pSMA, with a corresponding weight ratio of 3/7. The obtained pH-responsive MNF could be dispersed in an alkaline buffer (pH ≈ 9.0) directly. We further observed the morphology of the MNF at a certain concentration (0.5 mg mL−<sup>1</sup> ), which is shown in Figure 4B. Apparently, in the aqueous phase, the hydrophobic Fe3O<sup>4</sup> nanoparticles formed round packed clusters unlike the monodispersed nanoparticles in hexane (Figure 4A), which confirmed the successful fabrication of micellar MNF by encapsulation of the PCL-*b*-pSMA. could be dispersed in an alkaline buffer (pH ≈ 9.0) directly. We further observed the morphology of the MNF at a certain concentration (0.5 mg mL−1), which is shown in Figure 4B. Apparently, in the aqueous phase, the hydrophobic Fe3O4 nanoparticles formed round packed clusters unlike the monodispersed nanoparticles in hexane (Figure 4A), which confirmed the successful fabrication of micellar MNF by encapsulation of the PCL-*b*pSMA.

7.0, which should be suitable for the preparation of a tumor microenvironment-responsive

In order to prepare the tumor microenvironment-responsive MNF, we firstly synthesized hydrophobic Fe3O4 nanoparticles at approximately 8 nm. After, the MNF was prepared by the simple self-assembly between the amphiphilic PCL-*b*-pSMA and hydrophobic Fe3O4 nanoparticles. It was clear that the Fe3O4 nanoparticles in hexane (0.1 mg mL−1) presented monodispersity with a uniform particle size, as shown in Figure 4A, due to the fact of their high hydrophobicity. In order to improve the water dispersibility of the hydrophobic Fe3O4 nanoparticles, the polymeric micelles provided a powerful platform to load the hydrophobic molecules into their hydrophobic cores. Therefore, we prepared the MNF by the self-assembly of the hydrophobic Fe3O4 nanoparticles and the amphiphilic PCL-*b*-pSMA, with a corresponding weight ratio of 3/7. The obtained pH-responsive MNF

*Pharmaceuticals* **2023**, *16*, x FOR PEER REVIEW 6 of 17

*2.2. Characterization of the Tumor Microenvironment -Responsive MNF* 

**Figure 4.** Characterization of the Fe3O4 nanoparticles and MNF, including the morphologies (TEM), particle size distributions (DLS) and magnetic properties (magnetic hysteresis loops): (**A**) TEM result of the hydrophobic Fe3O4 nanoparticles in hexane (0.1 mg mL<sup>−</sup>1); (**B**) TEM result of the MNFs in alkaline buffer (pH ≈ 9.0, 0.5 mg mL−1); (**C**) DLS results of the samples in (**A**,**B**); (**D**) magnetic properties of the desiccative Fe3O4 nanoparticles (60.4 emu g<sup>−</sup>1) and MNF (17.9 emu g−1). **Figure 4.** Characterization of the Fe3O<sup>4</sup> nanoparticles and MNF, including the morphologies (TEM), particle size distributions (DLS) and magnetic properties (magnetic hysteresis loops): (**A**) TEM result of the hydrophobic Fe3O<sup>4</sup> nanoparticles in hexane (0.1 mg mL−<sup>1</sup> ); (**B**) TEM result of the MNFs in alkaline buffer (pH <sup>≈</sup> 9.0, 0.5 mg mL−<sup>1</sup> ); (**C**) DLS results of the samples in (**A**,**B**); (**D**) magnetic properties of the desiccative Fe3O<sup>4</sup> nanoparticles (60.4 emu g−<sup>1</sup> ) and MNF (17.9 emu g−<sup>1</sup> ).

In addition to TEM, we also used DLS to characterize the diameters of the Fe3O4 nanoparticles and MNF, as shown in Figure 4C. Apparently, the DLS result confirmed the Fe3O4 nanoparticles with a uniform particle size of approximately 8 nm, again, because of their low particle size distribution (PDI = 0.125). In addition, the DLS result also confirmed that the MNF possessed a larger particle size (145.2 nm) and a wider particle size distribution (PDI = 0.145) than those of the Fe3O4 nanoparticles, which corresponded to the TEM results in Figure 4A,B. Furthermore, the magnetic hysteresis curves of the Fe3O4 nanoparticles and MNF were characterized, as shown in Figure 4D. The Fe3O4 nanoparticles presented obvious superparamagnetism, corresponding to their diameters at approximately 8 nm. Consequently, the MNF also showed superparamagnetism without significant In addition to TEM, we also used DLS to characterize the diameters of the Fe3O<sup>4</sup> nanoparticles and MNF, as shown in Figure 4C. Apparently, the DLS result confirmed the Fe3O<sup>4</sup> nanoparticles with a uniform particle size of approximately 8 nm, again, because of their low particle size distribution (PDI = 0.125). In addition, the DLS result also confirmed that the MNF possessed a larger particle size (145.2 nm) and a wider particle size distribution (PDI = 0.145) than those of the Fe3O<sup>4</sup> nanoparticles, which corresponded to the TEM results in Figure 4A,B. Furthermore, the magnetic hysteresis curves of the Fe3O<sup>4</sup> nanoparticles and MNF were characterized, as shown in Figure 4D. The Fe3O<sup>4</sup> nanoparticles presented obvious superparamagnetism, corresponding to their diameters at approximately 8 nm. Consequently, the MNF also showed superparamagnetism without significant remanent magnetization. Meanwhile, the Fe3O<sup>4</sup> nanoparticles exhibited a high saturation magnetism (*M*<sup>s</sup> = 60.4 emu g−<sup>1</sup> ); on the contrary, the MNF showed a relatively low high saturation magnetism (*M*<sup>s</sup> = 17.9 emu g−<sup>1</sup> ). To understand the phenomenon, we measured the Fe3O<sup>4</sup> content in the MNF by thermogravimetric analysis (TGA). As shown in Figure S4, the content of the Fe3O<sup>4</sup> nanoparticles was 29.3 wt%, which was close to its feed ration. Therefore, it was reasonable that the saturation magnetism of the MNF was 30% of that of the Fe3O<sup>4</sup> nanoparticles. According to these results, the MNF was prepared successfully by loading the hydrophobic Fe3O<sup>4</sup> nanoparticles into the inner core of the polymeric micelles.

In order to verify the pH responsiveness of the MNF, we also studied the macroscopic states and microscopic morphologies of the MNF in different buffers with the same concentration (0.5 mg mL−<sup>1</sup> ), as shown in Figure 5.

celles.

centration (0.5 mg mL−1), as shown in Figure 5.

**Figure 5.** The macroscopic states (digital photo) and microscopic morphologies (TEM) of the tumor microenvironment-responsive MNF in different buffers (0.5 mg mL<sup>−</sup>1): (**A**) digital photos of the MNF for 2 and 30 min; (**B**) TEM results of the MNF for 2 min. **Figure 5.** The macroscopic states (digital photo) and microscopic morphologies (TEM) of the tumor microenvironment-responsive MNF in different buffers (0.5 mg mL−<sup>1</sup> ): (**A**) digital photos of the MNF for 2 and 30 min; (**B**) TEM results of the MNF for 2 min.

remanent magnetization. Meanwhile, the Fe3O4 nanoparticles exhibited a high saturation magnetism (*M*s = 60.4 emu g−1); on the contrary, the MNF showed a relatively low high saturation magnetism (*M*s = 17.9 emu g−1). To understand the phenomenon, we measured the Fe3O4 content in the MNF by thermogravimetric analysis (TGA). As shown in Figure S4, the content of the Fe3O4 nanoparticles was 29.3 wt%, which was close to its feed ration. Therefore, it was reasonable that the saturation magnetism of the MNF was 30% of that of the Fe3O4 nanoparticles. According to these results, the MNF was prepared successfully by loading the hydrophobic Fe3O4 nanoparticles into the inner core of the polymeric mi-

In order to verify the pH responsiveness of the MNF, we also studied the macroscopic states and microscopic morphologies of the MNF in different buffers with the same con-

Corresponding to a p*K*a value of the PCL-*b*-pSMA, the MNF could disperse very well in the aqueous phase with a physiological pH value (pH ≈ 7.43), forming a transparent brown liquid in the macroscopic state. Furthermore, we assessed the long-term stability of the MNF (1~13 d) under a physiological pH value (≈7.43), as shown in Figure S5, which confirmed its high colloidal stability under physiological conditions. At the same time, the MNF solution with a neutral pH value (pH ≈ 7.02) became a dark-brown liquid at the macroscopic level. With the further decrease in the pH value, the MNF lost its colloidal stability in the solution with weak acidity (pH ≈ 6.77), as shown in Figure 5A; many sediments could be observed after adjusting the pH to 6.77 for 2 min. With the prolonged observation time of 30 min, in physiological and neutral conditions, the MNF maintained its original states; however, the MNF in a weak acidic environment dropped completely to the bottom of the container. In addition to the differences at the macroscopic level, the MNF solutions with corresponding pH values for 2 min also showed quite different microscopic morphologies, as shown in Figure 5B. The MNF in the aqueous phase with pH = 7.43 exhibited a similar morphology as the TEM result in Figure 4B, indicating its colloidal stability under the physiological conditions again. However, in a neutral environment, several micellar MNF formed a large Fe3O4 cluster, as shown in Figure 5B, resulting in the low transmittance of the sample. When the surrounding pH value was downregulated to weak acidity (pH ≈ 6.77), many Fe3O4 nanoparticles piled up chaotically, also shown in Figure 5B, because the surrounding pH value was lower than the p*K*a of the PCL-*b*-pSMA. In addition to the macroscopic states and microscopic morphologies, we also quantified the diameters (Figure S6) and zeta potentials (Figure S7) of the MNF in solutions with Corresponding to a p*K*<sup>a</sup> value of the PCL-*b*-pSMA, the MNF could disperse very well in the aqueous phase with a physiological pH value (pH ≈ 7.43), forming a transparent brown liquid in the macroscopic state. Furthermore, we assessed the long-term stability of the MNF (1~13 d) under a physiological pH value (≈7.43), as shown in Figure S5, which confirmed its high colloidal stability under physiological conditions. At the same time, the MNF solution with a neutral pH value (pH ≈ 7.02) became a dark-brown liquid at the macroscopic level. With the further decrease in the pH value, the MNF lost its colloidal stability in the solution with weak acidity (pH ≈ 6.77), as shown in Figure 5A; many sediments could be observed after adjusting the pH to 6.77 for 2 min. With the prolonged observation time of 30 min, in physiological and neutral conditions, the MNF maintained its original states; however, the MNF in a weak acidic environment dropped completely to the bottom of the container. In addition to the differences at the macroscopic level, the MNF solutions with corresponding pH values for 2 min also showed quite different microscopic morphologies, as shown in Figure 5B. The MNF in the aqueous phase with pH = 7.43 exhibited a similar morphology as the TEM result in Figure 4B, indicating its colloidal stability under the physiological conditions again. However, in a neutral environment, several micellar MNF formed a large Fe3O<sup>4</sup> cluster, as shown in Figure 5B, resulting in the low transmittance of the sample. When the surrounding pH value was downregulated to weak acidity (pH ≈ 6.77), many Fe3O<sup>4</sup> nanoparticles piled up chaotically, also shown in Figure 5B, because the surrounding pH value was lower than the p*K*<sup>a</sup> of the PCL-*b*-pSMA. In addition to the macroscopic states and microscopic morphologies, we also quantified the diameters (Figure S6) and zeta potentials (Figure S7) of the MNF in solutions with corresponding pH values, which are shown in Figures S6 and S7. Apparently, in a neutral and weak acid solution, the MNF not only displayed larger diameters but also showed higher surface potentials (negative charge) compared to their counterparts in a physiological environment. According to these results, the MNF exhibited a similar pH responsiveness with a neutral pH value, which can be ascribed to the p*K*<sup>a</sup> of the PCL-*b*-pSMA. Considering a tumor microenvironment pH value (7.1~6.7) [13], the tumor microenvironment-responsive MNF should possess a potential for targeted accumulation in a tumor stroma.

#### *2.3. Advantages of Tumor Microenvironment-responsive MNF for Tumor Treatment and Diagnosis 2.3. Advantages of Tumor Microenvironment-responsive MNF for Tumor Treatment and Diagnosis*

corresponding pH values, which are shown in Figures S6 and S7. Apparently, in a neutral and weak acid solution, the MNF not only displayed larger diameters but also showed higher surface potentials (negative charge) compared to their counterparts in a physiological environment. According to these results, the MNF exhibited a similar pH responsiveness with a neutral pH value, which can be ascribed to the p*K*a of the PCL-*b*-pSMA. Considering a tumor microenvironment pH value (7.1~6.7) [13], the tumor microenvironmentresponsive MNF should possess a potential for targeted accumulation in a tumor stroma.

Due to the high magnetic particle content, we studied the magnetocaloric effects of the MNF under different conditions, and the corresponding heating curves are shown in Figure 6A. Due to the high magnetic particle content, we studied the magnetocaloric effects of the MNF under different conditions, and the corresponding heating curves are shown in Figure 6A.

*Pharmaceuticals* **2023**, *16*, x FOR PEER REVIEW 8 of 17

**Figure 6.** The effect of the tumor microenvironmental pH value on the tumor microenvironmentresponsive MNF's applications, including (**A**) enhanced MH with a concentration of Fe at 100 μg mL−1; (**B**) increased ROS generation with the concentration of the MNF at 250 μg mL−1; (**C**,**D**) improved *T*2 relaxivity of the MRI with varied Fe concentrations from 1.36 to 0.0425 mM. **Figure 6.** The effect of the tumor microenvironmental pH value on the tumor microenvironmentresponsive MNF's applications, including (**A**) enhanced MH with a concentration of Fe at 100 µg mL−<sup>1</sup> ; (**B**) increased ROS generation with the concentration of the MNF at 250 µg mL−<sup>1</sup> ; (**C**,**D**) improved *T*<sup>2</sup> relaxivity of the MRI with varied Fe concentrations from 1.36 to 0.0425 mM.

Under the same concentration of Fe (100 μg mL−1), the MNF could increase the surrounding temperature rapidly in solutions with different pH values; however, the fastest heating rate and highest heating temperature occurred in neutral conditions simultaneously. Additionally, the initial heating rate and final heating temperature of the MNF in a weak acid environment were also higher than its counterpart in physiological conditions. Therefore, the ranking of the MNF's SAR in the different buffers is neutral condition > weak acid > physiological condition, which is also shown in Figure 6A. This is an interesting phenomenon, because environmental factors of MH are rarely reported. According to the mechanism of MH, the monodispersed SPIO nanoparticles with a small diameter (≤10 nm), as used in this study, induced a magnetocaloric effect under AMF mostly by Néel relaxation, which is associated with magnetic moment [4]. Therefore, many previous studies have focused on material factors for enhancement of the magnetic moment of SPIO, including composition [30], shape [31], and structure [32]. However, in the past decade, many relevant studies also found that the micellar MNF with a large diameter exhibited high efficiency in producing thermal energy by Néel relaxation [23,33], because the effective magnetic moment of the SPIO could be improved by increasing the diameter of the Under the same concentration of Fe (100 µg mL−<sup>1</sup> ), the MNF could increase the surrounding temperature rapidly in solutions with different pH values; however, the fastest heating rate and highest heating temperature occurred in neutral conditions simultaneously. Additionally, the initial heating rate and final heating temperature of the MNF in a weak acid environment were also higher than its counterpart in physiological conditions. Therefore, the ranking of the MNF's SAR in the different buffers is neutral condition > weak acid > physiological condition, which is also shown in Figure 6A. This is an interesting phenomenon, because environmental factors of MH are rarely reported. According to the mechanism of MH, the monodispersed SPIO nanoparticles with a small diameter (≤10 nm), as used in this study, induced a magnetocaloric effect under AMF mostly by Néel relaxation, which is associated with magnetic moment [4]. Therefore, many previous studies have focused on material factors for enhancement of the magnetic moment of SPIO, including composition [30], shape [31], and structure [32]. However, in the past decade, many relevant studies also found that the micellar MNF with a large diameter exhibited high efficiency in producing thermal energy by Néel relaxation [23,33], because the effective magnetic moment of the SPIO could be improved by increasing the diameter of the magnetic nanocluster. In this study, due to the results shown in Figure 5B and Figure S6, the tumor microenvironment-responsive MNF in neutral and weak acid conditions displayed a larger diameter by the aggregation of multi-magnetic micelles, compared to its counterpart in a physiological environment. Consequently, the magnetocaloric effect of the MNF could be enhanced specifically in tumor stroma, which should benefit the application of tumor microenvironment-responsive MNF for clinical MH. However, it is worth noting that the tumor microenvironment-responsive MNF showed the highest SAR value under neutral conditions. This should be attributed to the Brownian relaxation of the MNF under AMF, because the agglomerated MNF could maintain dispersibility in the

neutral solution, as shown in Figure 5, leading to their free rotation under AMF. Therefore, the tumor microenvironment-responsive MNFs could generate thermal energy efficiently in the neutral solution by integrating Brownian relaxation and Néel relaxation. On the contrary, in a weak acid environment, the MNF could not undergo Brownian motion easily, because it formed precipitates rapidly. Consequently, the MH of the MNF in the weak acid solution could be induced by Néel relaxation alone.

In addition to MH, the neutral condition could also improve the catalytic performance of the MNF to generate •OH, as shown in Figure 6B. It is well known that Fe2+ can decompose low toxic H2O<sup>2</sup> to form high toxic •OH by FR efficiently [34,35], inducing the application of the SPIO nanoparticles for CDT in many previous studies [36]. Therefore, in this study, we evaluated the •OH-generating ability of the MNF in solutions with different pH values using methylene blue (MB) as a detection probe, since the MB could be decomposed by •OH. Obviously, H2O<sup>2</sup> could not break the structure of the MB, as the UV spectrum of the MB alone (a) coincided with that of the mixture of MB and H2O<sup>2</sup> (b). However, by incubating with H2O<sup>2</sup> and the MNF (c~e) simultaneously, the UV absorbance of the MB declined, indicating its degradation. Moreover, the surrounding pH value influenced the degradation of the MB significantly, also shown in Figure 6B, in which a neutral environment (d) was the best condition for the catalytic performance of the tumor microenvironment-responsive MNF. This was an unexpected result, because most previous studies showed that an acidic environment (pH ≤ 6.0) could improve the efficiency of FR [34,37]. However, in these studies, the corresponding iron-based nanocatalysts maintained a high colloidal stability in a neutral environment, which reduced the interaction between the naked SPIO nanoparticles and H2O2. In our study, due to the phase transition of the PCL-*b*-pSMA, the interactions between H2O<sup>2</sup> and the naked SPIO increased. More importantly, as shown in Figure 5, the agglomerated MNF could maintain the dispersed state in the aqueous phase for long time in a neutral environment; meanwhile, its counterpart under weak acidic conditions precipitated rapidly. This phenomenon indicated that the tumor microenvironment-responsive MNF presented a state of transition in a neutral environment, which increased the probability of an interfacial reaction between the Fe3O<sup>4</sup> nanoparticles and H2O<sup>2</sup> dramatically. Considering a neutral or weak acidic microenvironment of a tumor stroma, the tumor microenvironment-responsive MNF possesses a unique advantage for tumor microenvironment-responsive CDT.

The agglomerated MNF in a neutral environment not only presented an excellent catalytic performance but also enhanced the *T*<sup>2</sup> relaxivity in the MRI, which is shown in Figure 6C,D. For a given Fe concentration, the *T*<sup>2</sup> imaging of the MNF under neutral conditions (pH ≈ 6.99) was significantly darker than its counterpart in a physiological environment, as shown in Figure 6C. As the *T*<sup>2</sup> imaging of the MRI represented a negative signal contrast, the darker imaging MNF in neutral conditions indicated the enhancement of the *T*<sup>2</sup> imaging. Based on the results in Figure 6C, we quantified the *T*<sup>2</sup> relaxivities of the MNF under different conditions, as shown in Figure 6D. The MNF in neutral conditions possessed a steeper slope (*r*<sup>2</sup> = 271.6 mM−<sup>1</sup> s −1 ) compared to its counterpart in physiological conditions (*r*<sup>2</sup> = 164.1 mM−<sup>1</sup> s −1 ). The results of the *T*<sup>2</sup> imaging enhancement correspond to many relevant studies [22,38–40]; all of them confirmed that the *T*<sup>2</sup> relaxation rate of the MRI could be enhanced by forming SPIO clusters and increasing the diameter of these clusters. Although the MNF was a micellar cluster of SPIO nanoparticles in a physiological environment, as shown in Figure 5, the particle size of the SPIO clusters could be increased further in neutral conditions. Therefore, the *T*<sup>2</sup> imaging of the MNF could be enhanced by a neutral environment, indicating its application for tumor detection by MRI.
