Exceedingly small iron oxide nanoparticles as positive MRI contrast agents

Significance

Gadolinium (Gd)-based contrast agents (GBCAs) are currently the mainstream clinical MRI contrast agents. Some GBCAs have shown a long-term toxicity—nephrogenic systemic fibrosis (NSF)—and Gd depositions in the brain. The NSF has triggered a Food and Drug Administration (FDA) black-box warning and a contraindication of some GBCAs. The finding of Gd depositions led to an ongoing FDA investigation to monitor their possible long-term adverse effects. Here, we present T₁-weighted contrast-enhanced MR imaging and angiography using zwitterion-coated exceedingly small superparamagnetic iron oxide nanoparticles (ZES-SPIONs) in mice and rats. Renal clearance and biodistribution results further demonstrate that ZES-SPIONs are qualitatively different from previously reported SPIONs. This work may open up opportunities to develop exceedingly small SPIONs that show effective T₁ contrast as Gd-free alternatives to GBCAs.

Abstract

Medical imaging is routine in the diagnosis and staging of a wide range of medical conditions. In particular, magnetic resonance imaging (MRI) is critical for visualizing soft tissue and organs, with over 60 million MRI procedures performed each year worldwide. About one-third of these procedures are contrast-enhanced MRI, and gadolinium-based contrast agents (GBCAs) are the mainstream MRI contrast agents used in the clinic. GBCAs have shown efficacy and are safe to use with most patients; however, some GBCAs have a small risk of adverse effects, including nephrogenic systemic fibrosis (NSF), the untreatable condition recently linked to gadolinium (Gd) exposure during MRI with contrast. In addition, Gd deposition in the human brain has been reported following contrast, and this is now under investigation by the US Food and Drug Administration (FDA). To address a perceived need for a Gd-free contrast agent with pharmacokinetic and imaging properties comparable to GBCAs, we have designed and developed zwitterion-coated exceedingly small superparamagnetic iron oxide nanoparticles (ZES-SPIONs) consisting of ∼3-nm inorganic cores and ∼1-nm ultrathin hydrophilic shell. These ZES-SPIONs are free of Gd and show a high T₁ contrast power. We demonstrate the potential of ZES-SPIONs in preclinical MRI and magnetic resonance angiography.

MRI signal arises from the excitation of low-energy nuclear spins, which are formed in a permanent magnetic field, by applying radiofrequency pulses followed by the measurement of the spin relaxation processes (i.e., T₁ recovery or T₂ decay) (1, 2). Different chemical environments as well as water concentration result in different signal strengths and therefore provide contrast between fat, tissue, and bones. Paramagnetic compounds can be used to enhance the contrast of MR images by promoting relaxation of water near the compound. MRI contrast agents are classified as either T₁ (i.e., positive) or T₂ (i.e., negative). Radiologists strongly prefer T₁ contrast agents because T₂ contrast shows as darkened areas, which can be difficult to distinguish from internal bleeding, air–tissue boundaries (3), or other susceptibility artifacts, resulting in less accurate patient diagnosis. We also note that progress has been made to generate positive contrast by artificially turning dark and susceptibility-related negative contrast bright (4). Moreover, T₁ relaxation times in animal tissues are generally much longer than their T₂ relaxation times, meaning that the relaxivity of the T₁ contrast agent can be smaller than that for a T₂ agent to obtain the same amount of change in image intensity. All gadolinium (Gd)-based contrast agents (GBCAs) (5–9) used in the clinic today are T₁ contrast agents (10), whereas GBCAs still have a small risk of adverse effects including nephrogenic systemic fibrosis (11, 12) and Gd deposition in the human brain (13–16).

Superparamagnetic iron oxide nanoparticles (SPIONs) are single-domain magnetic iron oxide particles with hydrodynamic diameters (HDs) ranging from single nanometers to >100 nm (17–19). SPIONs can be monodisperse and coated by biologically compatible ligands, are chemically and biologically stable, and are generally nontoxic in vivo (20). However, commercially available SPION contrast agents are composed of polydisperse inorganic cores with large HD, ranging from ∼16 to ∼200 nm. Generally, large SPIONs function as T₂ contrast agents, whereas small SPIONs have limited T₂ activity and therefore are potential T₁ contrast agents. In addition, due to their large HD, existing SPIONs (21) prevent efficient renal clearance after i.v. administration, greatly differing from GBCA elimination pathways. As a result, large HD SPIONs predominately accumulate in the body (22) and can cause a persistent negative contrast over several weeks or months, which prevents repeated imaging studies and limits the clinical management of patients. Furthermore, current SPION formulations are almost quantitatively metabolized and absorbed into the iron pool with the potential of clinical side effects from iron overload (22). We note that the development and production of existing SPION-based MRI contrast agents, including Resovist, Feridex, Combidex, Supravist, Clariscan (21), and Gastromark (23, 24), has largely ceased. Hence, there is a perceived need for developing Gd-free SPIONs that are exceedingly small for T₁-weighted MRI and as a potential substitute for GBCAs. In contrast to previous large HD SPIONs, and to address the perceived need for a Gd-free positive contrast agent, we have designed zwitterion-coated exceedingly small superparamagnetic iron oxide nanoparticles (ZES-SPIONs) with two considerations: (i) optimizing T₁ contrast and (ii) enabling renal clearance. Optimizing T₁ contrast (25) for SPIONs translates into minimizing the ratio r₂/r₁ while preserving an absolute r₁ value similar to GBCAs, where r₂ and r₁ are the transverse and longitudinal relaxivities, respectively (26). We achieve this by using the less magnetic maghemite (Fe₂O₃) structure, which shows a lower magnetization at clinical strength (1.5 T) applied magnetic fields (27, 28), as opposed to the magnetite (Fe₃O₄) structure (29), and this translates into a correspondingly low T₂ effect (30, 31). To enable renal clearance, ZES-SPIONs must have a HD below ∼5.5 nm and show minimal nonspecific interactions (i.e., minimal biofouling) in vivo (32), which renders ZES-SPIONs as ideal candidates for future targeted imaging applications. Most existing SPIONs described for biomedical applications (21) have an HD greater than 7 nm. Hence, there is a perceived need for developing Gd-free SPIONs that are exceedingly small for T₁-weighted MRI and enhanced renal clearance as potential substitutes for GBCAs. Here, we show that ZES-SPIONs as T₁ contrast agents are indeed qualitatively different from existing SPIONs (21) and indeed approach the performance properties of GBCAs.

Our SPIONs were synthesized from the thermal decomposition of Fe(oleate)₃ in the presence of oleic acid and in solvent mixtures rationally designed for tuning their boiling point. This is followed by oxidation with trimethylamine N-oxide (Fig. 1A) (33). The oxidation step ensures particles with a maghemite structure. Syntheses of SPIONs with a range of ∼2.5- to ∼7.0-nm inorganic core diameters were produced with narrow size distribution by adjusting the solvent mixture boiling point, using mixtures of 1-tetradecene (TDE), 1-hexadecene (HDE), and 1-octadecene (ODE), keeping both the concentration of precursors and growth time constant (Fig. S1) (34, 35). Fig. 1B shows a high-resolution transmission electron microscopy (HR-TEM) image of ∼3.0-nm SPIONs. Relaxivity results at 7 T show that this size provides good T₁-weighted MRI signal (Table S1). We characterized the magnetic behavior of ∼3.0-nm SPIONs coated in native ligands at room temperature (298K) using a superconducting quantum interference device (SQUID) and compared it to the commercially available and US Food and Drug Administration (FDA)-approved ferumoxytol (Feraheme), which consists of 3- to 10-nm mixed magnetite/maghemite cores [confirmed by X-ray powder diffraction (XRD) in Fig. S2]. In Fig. 1C, the superparamagnetism of our ∼3-nm SPIONs at room temperature is confirmed by the absence of a hysteresis loops near zero field. Fig. 1C also shows that Gd-DTPA (Magnevist) has a small magnetization (M), desired for T₁-weighted MRI. In the range of clinical magnetic field strengths (1.5–3 T), Fig. 1C shows that the M of magnetite-containing ferumoxytol is 94 emu/g [Fe], whereas the M of our ∼3-nm SPIONs is only 35 emu/g [Fe]. Assuming the magnetic volume/field gradient and water-SPION interaction are largely unchanged, this low M would yield a reduced r₂ relaxivity and an enhanced T₁-weighted MRI signal, a strategy that has been successfully used for doped ferritin-based SPIONs (36).

Fig. 1.

(A) Rationally designed synthetic route of ZDS-coated SPIONs, (B) HR-TEM images of SPIONs with ∼3.0-nm inorganic core diameter, and (C) SQUID curves of 3-nm ES-SPIONs, ferumoxytol (Feraheme), as well as Gd-DTPA (Magnevist).

Fig. S1.

(A–D) TEM images of SPIONs with ∼7.0-, 5.5-, 3.0-, and 2.5-nm core diameters, respectively.

Table S1.

r₁ and r₂ relaxivity measurements of a series of iron oxide nanoparticles at 7 T

Nanoparticles	Core, nm	r1, s−1⋅mM−1	r2, s−1⋅mM−1	r2/r1
Feraheme	∼3–10	3.1	68	22
Our SPIONs	∼7	3.4	60	18
Our SPIONs	∼5.5	2.9	51	18
Our SPIONs	∼3	1.5	17	11
Our SPIONs	∼2.5	1.3	17	13

Fig. S2.

XRD of ferumoxytol.

After synthesizing the above inorganic core SPIONs, the native hydrophobic oleic acid coating was exchanged with a zwitterionic dopamine sulfonate (ZDS) that we have previously developed and described (35, 37). Our prior studies have demonstrated that ZDS-coated SPIONs show small HDs, low nonspecific interactions in vitro with cells and in vivo in mice, offering the opportunity for specific labeling, and stability with respect to time, pH, and salinity. Moreover, ZDS-coated SPIONs can retain the magnetic properties of as-synthesized hydrophobic SPIONs, making them useful for MRI. The biocompatibility of ZDS was also confirmed by other research groups on several types of nanoparticles (10, 38, 39). Fig. S3 shows the result of using HPLC with a size exclusion column to determine the HD of ∼3-nm core particles coated with ZDS (ZES-SPIONs), giving an average HD of 4.7 nm.

Fig. S3.

(A) GFC curve of ZDS-coated SPIONs with 3-nm inorganic core diameters (ZES-SPIONs) and (B) the calibration curve of HD vs. retention time by using protein standards.

To quantitatively evaluate (40) the T₁ contrast power of ZES-SPIONs and differentiate it with previously reported agents, we compare the r₁ and r₂ values of ZES-SPIONs and other types of T₁ contrast agents (3, 30, 41). The data for ferumoxytol (42, 43), very-small iron oxide nanoparticles (VSOPs) (44), as well as Gd-DTPA (Magnevist) (24, 45) are either measured or taken from the published literature. As shown in Fig. 2, at 1.5 T, which is the field strength of most clinical MRI scanners, ZES-SPIONs with a 3-nm inorganic core and a 4.7-nm HD have an r₁ of 5.2 s⁻¹⋅mM⁻¹ and r₂ of 10.5 s⁻¹⋅mM⁻¹, leading to an r₂/r₁ of 2.0, which is three times smaller than the r₂/r₁ of the commercially available SPION-based ferumoxytol. The r₂/r₁ of ZES-SPIONs is slightly lower and therefore slightly better than that of VSOP (44), a state-of-the-art SPION-based T₁ contrast agent. The r₂/r₁ of our ZES-SPIONs is more preferable than that of other SPION-based MRI contrast agents including Feraheme, Resovist, Feridex, Combidex, Supravist, Clariscan, and others (3, 8, 9, 21, 46, 47). Furthermore, as shown in Fig. 2, the r₂/r₁ of ZES-SPIONs at 1.5 T is within a factor of 2 to that of gadolinium-based chelates (GBCAs) such as Gd-DTPA. Crucially, not only do our ZES-SPIONs show a more promising r₂/r₁ than previous SPIONs, but they are also smaller and qualitatively different; unlike previous SPION-based T₁ contrast agents, the HD of ZES-SPIONs falls below the effective, expected renal clearance cutoff (i.e., the glomerular filtering threshold) of ∼5.5 nm as detailed below (32). The HD of ZES-SPIONs is 4.7 nm, which is approximately two times smaller than the HD of VSOP, and approximately six times smaller than the HD of ferumoxytol, respectively (Fig. 2).

Fig. 2.

HD and T₁ contrast power of different contrast agents (ferumoxytol, Feraheme; Gd-DTPA, Magnevist; VSOP, very small iron oxide nanoparticles; ZES-SPIONs, zwitterion-coated exceedingly small iron oxide nanoparticles).

To assess the in vivo stability of our ZES-SPION formulation, we designed and performed a biodistribution experiment using ⁵⁹Fe radioisotope-labeled ZES-SPIONs. As shown in Fig. 3A, ES-SPIONs were first labeled with ⁵⁹Fe (48) and then ligand exchanged with ZDS. The resulting radioactive ZES-SPIONs were analyzed by gel filtration chromatography (GFC) equipped with a size exclusion column, which showed the simultaneous elution of high radioactivity and ZES-SPIONs, indicating the successful incorporation of ⁵⁹Fe. The ⁵⁹Fe-labeled ZES-SPIONs were i.v. injected into mice. After 4 and 24 h, the urine and feces were collected and separated, and the ⁵⁹Fe concentration was measured by a radioactivity counter. As reported in Fig. 3B, 65 ± 1.4% of the injected ⁵⁹Fe-labeled ZES-SPIONs was renally cleared (most of it within 4 h). Organs and blood were collected 24 h postinjection, and the ⁵⁹Fe activity was measured to determine the biodistribution. Fig. 3B shows that, 24 h after injection, 13 ± 0.70% of the injected ⁵⁹Fe remained in the liver. However, the combination of blood, spleen, kidney, lung, gastrointestinal tract, heart, and tail had around 12%. The remaining carcass had around 13% of the injected ⁵⁹Fe. These results suggest that the majority of ZES-SPIONs are cleared through the renal route.

Fig. 3.

(A) Schematic of mouse biodistribution study using ⁵⁹Fe-labeled ZES-SPIONs, and (B) the percentages of ⁵⁹Fe-labeled ZES-SPIONs that stay in urine, blood, and organs, compared with the total amount of ⁵⁹Fe-labeled ZES-SPIONs injected (GIT, gastrointestinal tract). After 24 h following injection, 12.5 ± 0.7% of the injected ⁵⁹Fe-labeled ZES-SPIONs remained in the liver. The combination of blood, spleen, kidney, lung, gastrointestinal tract, heart, carcass, and tail had less than 25% of the injected ⁵⁹Fe-labeled ZES-SPIONs.

To further visualize renal clearance, our ZES-SPIONs were i.v. injected into mice and their urine was collected at different time points. Fig. 4K shows normal urine color before injection. At 30 min and 1.5 h postinjection, the urine becomes dark brown, which is characteristic of ZES-SPIONs elution. After 2.5 h postinjection, the urine returns to its normal yellow color, indicating that the particle excretion happens mostly within the first 2.5 h, which is in agreement with the radioactive biodistribution data. Moreover, we performed relaxivity measurements on the ZES-SPIONs that were renally cleared (Table S2). The results show that the ZES-SPIONs in urine have an r₂/r₁ of 1.9, very close to the r₂/r₁ of ZES-SPIONs in PBS before injection. Together, these results indicate that ZES-SPIONs clear through the kidney efficiently as particles and their MR contrast power remains largely unchanged.

Fig. 4.

(A–J) T₁-weighted MR images of a mouse injected with ZDS-coated exceedingly small SPIONs (ZES-SPIONs) at 7 T. Time points underneath each image: the time after ZES-SPIONs injection: (A–E) one sagittal slice showing the heart (red arrow), the vena cava (green arrow), and the bladder (yellow arrow), and (F–J) another sagittal slice showing the kidney (blue arrow); (K) urine samples from mice taken at different time points after injection showing renal clearance of ZES-SPIONs in vivo. Before the injection of ZES-SPIONs, the heart (red arrow), the vena cava (green arrow), and the bladder (yellow arrow) do not show appreciable positive contrast. After injection, the heart and the vena cava display high positive contrast immediately after injection. At 8 min after injection, the bladder displays some positive contrast, indicating an excretion of urine containing ZES-SPIONs. With the increase of time postinjection, Fig. 4 D and E shows that the positive contrast region of the bladder increases, suggesting an accumulation of urine with ZES-SPIONs.

Table S2.

r₁ and r₂ relaxivity measurements of ZES-SPIONs (3-nm core and 4.7-nm HD) at 7 T

Nanoparticles	Batch no.	r1, s−1⋅mM−1	r2, s−1⋅mM−1	r2/r1
ZES-SPIONs in PBS	①	5.0	9.9	2.0
ZES-SPIONs in PBS	②	4.1	7.8	1.9
ZES-SPIONs in PBS	③	6.6	13.8	2.1
Average of ①, ②, and ③		5.2	10.5	2.0
ZES-SPIONs in urine	③	4.7	9.1	1.9

Next, we demonstrate the preclinical potential of our ZES-SPIONs as positive contrast agent for MR imaging (49, 50) in an animal model. We used a MRI scanner to image mice injected with ZES-SPIONs at a concentration of 0.2 mmol [Fe]/kg, comparable to the concentrations of GBCAs (∼0.1–0.25 mmol [metal]/kg) administered in the clinic. The T₁-weighted MR images of one of the mice are shown in Fig. 4 and in Movies S1–S3. Fig. 4A shows a sagittal slice before the injection of ZES-SPIONs. Before injection the heart (red arrow), the vena cava (green arrow) and the bladder (yellow arrow) do not show appreciable positive contrast, as expected. Fig. 4B shows that the heart and the vena cava display strong positive contrast immediately after injection. Fig. 4C shows that, at 8 min after injection, the bladder displays some positive contrast, indicating an excretion of ZES-SPIONs. With the increase of time postinjection, Fig. 4 D and E show that the positive contrast in the bladder increases, suggesting an accumulation of ZES-SPIONs (see Movies S1–S3 and Fig. S4 for additional rat and mouse MRI data).

Fig. S4.

T₁-weighted MRI in vivo in (A) rat and (B) mouse (heart, blue oval; kidney, yellow oval; bladder, red oval) injected with ZES-SPIONs.

Fig. 4 F–J show a different sagittal slice at the same time points. Fig. 4F shows the absence of contrast before injection for the kidney (blue arrow). Immediately after injection, the kidney displays strong positive contrast enhancement. With increasing time postinjection, Fig. 4 G–J demonstrates continuing positive contrast enhancement in the kidney (Fig. S5A). These results suggest that our ZES-SPIONs can clear through the kidney at a rate that is consistent with the timescales used in multiphase dynamic imaging. Previously reported SPIONs have all have HDs that are larger than 5.5 nm, whereas the ZES-SPIONs here have a qualitatively smaller HD of 4.7 nm, which is smaller than the glomerular filtering threshold.

Fig. S5.

(A) The volume of urine containing contrast in the bladder of a mouse injected with ZES-SPIONs vs. time measured by MRI and (B) the decay of the log of relative MRI signal (T₂*) in mice blood after particles injection.

We further demonstrate the future preclinical potential of our ZES-SPIONs in T₁-weighted magnetic resonance angiography (MRA), an important clinical use of T₁-weighted contrast (51). Fig. 5 A–C shows MRA of a mouse injected with ZES-SPIONs at different time points. Due to the strong T₁ contrast and a circulation time that is sufficiently long, the blood vessels in this MRA study can be imaged with a spatial resolution of ∼0.2 mm. Fig. 5 A–C shows strong positive contrast of the heart and blood vessels, which fades in time while the signal from the bladder increases, again consistent with a renal excretion pathway. The application of ZES-SPIONs as a blood pool agent can be further described by comparing signal-to-noise ratios (SNRs) in the vascular system and the less perfused tissues. Fig. 5 D–H shows different spatial perspectives of the MRA extracted from a three-dimensional (3D) scan (see Movies S4 and S5 for the full 3D scan). Here, contrast-to-noise ratio (CNR) between blood in the vena cava and muscle tissue increases from 4 (before injection) to 55 (after injection). The initial SNRs were 13 for blood and 9 for muscle, respectively; and after ZES-SPION administration, the SNRs were 65 and 10, respectively, showing a relative signal enhancement of fivefold in blood and essentially no pronounced signal enhancement in muscle. We note here also that the longer circulation times (the half-life in blood is ∼19 min, which was measured by monitoring the T2* signal over time as shown in Fig. S5B) of ZES-SPIONs relative to the very rapid clearance of some Gd-based contrast agents such as Gd-DTPA (∼2 min blood half-life) may provide an advantage in angiography. Moreover, the blood half-life of ZES-SPIONs is comparable to that of Gadofosveset [Ablavar; ∼23 min (52)], a type of GBCA used for angiography. In addition, our future efforts will be geared toward the development of targeted formulations based on the presented ZES-SPIONs, such as by using thiol-maleimide conjugation strategies.

Fig. 5.

(A–C) T₁-weighted magnetic resonance angiography (MRA) of a mouse injected with ZDS-coated exceedingly small SPIONs (ZES-SPIONs) at 7 T. The time points beneath each image are the time after injection of the ZES-SPIONs. (D–H) Five different perspectives of the MRA, which are extracted from the 3D scan, at the 4-min mark. A clear positive contrast of the heart and blood vessels are seen, and this positive contrast fades in time while the signal from the bladder increases, consistent with a renal excretion pathway.

In summary, our results demonstrate that ZES-SPIONs have a small enough HD to show kidney clearance, and that their T₁ contrast power is high enough that these particles can be used for MRA and conventional positive MRI contrast. This is an example of a class of Gd-free MRI contrast agents that could be used for MRA in a way that is similar to the GBCAs. Furthermore, unlike existing SPION-based MRI contrast agents, which exhibit prolonged contrast and a potential for iron overload, the pharmacokinetic properties of ZES-SPIONs are such that long-term contrast changes may be avoided, and the iron dose that remains in the body can be kept in a safe range. This material system can be the basis for developing positive Gd-free MRI contrast agents as alternatives to GBCAs in the clinic.

Materials and Methods

Chemicals were obtained from Sigma-Aldrich and used as received unless specified. Air-sensitive materials were handled under dry nitrogen atmosphere with oxygen levels <0.2 ppm in an Omni-Lab VAC glove box. All solvents were purchased from EMD Biosciences and spectrophotometric grade. TEM images and energy-dispersive X-ray spectroscopy (EDX) of the SPIONs were obtained with a JEOL 200CX electron microscope operated at 120 kV and a JEOL 2010 electron microscope operated at 200 kV. ζ-Potential measurements were performed on a Malvern Instruments Nano-ZS90. Magnetization measurements were performed on a Quantum Design MPMS-XL SQUID. Mice for urine study were acquired from Charles River Laboratories International, and they were housed in the Division of Comparative Medicine at Massachusetts Institute of Technology in an Association for Assessment and Accreditation of Laboratory Animal Care-accredited facility. All mice were studied in accordance with approved institutional protocols at Massachusetts Institute of Technology and University Medical Center Hamburg–Eppendorf. Mice and rats were anesthetized by either i.p. injection of ketamine and xylazine or inhalation of isoflurane and oxygen gas. Additional experimental details about synthesis, characterizations, relaxivity measurements, and MR imaging are provided in Supporting Information.

Synthesis of SPIONs

Using ∼3-nm SPIONs as an example, 0.900 g of Fe(oleate)₃ and 190 µL of oleic acid were added into a mixture of 4.0 mL of 1-tetradecene and 1.0 mL of 1-hexadecene at room temperature (RT). After 30-min degassing at 100 °C, the temperature of reaction mixture was rapidly increased to a final temperature of 270 °C, and then kept constant for 1 h. Then the reaction mixture was allowed to cool to RT upon the completion of the reaction, followed by the addition of 0.113 g of (CH₃)₃NO oxidizing agent (33). The reaction temperature was then increased to 130 °C and held constant for 1 h to fully oxidize the particles. The reaction mixture was then allowed to cool to RT. Approximately 40 mL of acetone was added, and the resulting mixture was centrifuged. The supernatant was then discarded and the SPION pellet was dispersed and stored in hexane. Energy-dispersive X-ray spectroscopy (EDX) results show that ∼3-nm SPIONs have 18.33% Fe and 81.67% O (by weight), where the excess amount of oxygen should be from organic ligands.

Organic Ligands and Ligand Exchange

Zwitterionic dopamine sulfonate (ZDS) (M_r: ∼300 g/mol) was synthesized and ZDS-coated SPIONs were prepared according to the protocol in the supporting information in a previously published manuscript (37). The ζ-potential of ZDS-coated ∼3-nm SPIONs was measured to be −5.02 ± 0.20 mV, which is almost neutral.

Gel Filtration Chromatography and Hydrodynamic Diameter Measurements

Gel filtration chromatography (GFC) measurements with a size-exclusion column were done by using the protocols in our prior publications (35, 37). After being filtered by Whatman 0.02-µm filter, ZDS-coated ∼3.0-nm core SPIONs [zwitterionic dopamine sulfonate-coated exceedingly small SPIONs (ZES-SPIONs)] in PBS (pH 7.4) have a narrow size distribution with a retention time peak of 35.2 min, as shown in Fig. S3A. In Fig. S3B, after calibrating the size-exclusion column with protein standards (Bio-Rad; catalog #151-1901) containing thyroglobulin [P1, hydrodynamic diameter (HD) = 18.8 nm], γ-globulin (P2, HD = 11.1 nm), ovalbumin (P3, HD = 6.1 nm), myoglobin (P4, HD = 3.8 nm), and vitamin B₁₂ (P5, HD = 1.67 nm), it can be found that the relationship between retention time (in minutes) and HD (in nanometers) is as follows: HD = −80.6291 + 11.79692*time − 0.43275*time² + 0.00473*time³. Therefore, a retention time peak of 35.2 min corresponds to an averaged HD of 4.7 nm. These results suggest that the ZDS ligand only contributes ∼1 nm to the overall radius, and the size change induced by the ZDS ligand and their stability are consistent with our prior study (35, 37); in particular, these ZDS-coated SPIONs can have a HD smaller than the ∼5.5-nm threshold for efficient renal clearance (32).

In Vitro MR Relaxivity Measurements at 7 T

The r₁ and r₂ relaxivities of our ZDS-coated SPIONs were measured on a 7-T MRI scanner to evaluate their performance as T₁ contrast agents (40). Samples were linearly diluted in PBS (pH 7.4) to span concentrations. SPION dilutions were then added to the center of a 384-well microtiter plate (BD Falcon). Remaining wells were filled with PBS to reduce susceptibility-based image artifacts and improve shimming efficiency. Relaxivity data were collected at ambient temperatures on a Bruker Avance III 7T MRI scanner with a 20-cm bore and using a Bruker Biospin MRI coil. A 2-mm horizontal slice was imaged through the center of the filled wells with a spin echo pulse sequence using a 10-ms echo time (T_E) and a 73-ms to 3-s repetition time (T_R = 73, 116, 186, 298, 477, 763, 1,220, 1,953, and 3,125 ms) to collect T₁ relaxation times. Meanwhile, T_E = 10 ms, 30 echos, and T_R = 3,125 ms were used to collect T₂ relaxation time. Thirty time points of 2D Carr–Purcell–Meiboom–Gill (CPMG) echo images (T_E = 10–300 ms with 10-ms interval) were collected as 256 × 256-point matrices. All data were subsequently analyzed using custom Matlab (Mathworks) routines. The echo time vs. integrated intensity from imaged wells was fit to an exponential to obtain transverse (R₂) and longitudinal (R₁) relaxation rates for each sample. The reported relaxivities, r₁ and r₂, are computed from the slope of the relationship between relaxation rate and iron concentration for each sample dilution. As shown in Table S1, the r₂/r₁ ratio of our ZDS-coated SPIONs can be tuned between 11 and 18 at 7 T, and the r₂/r₁ ratio of Feraheme is 22 at 7 T.

MR Relaxivity Measurements at 1.5 T and Mouse Urine Study

The relaxivity measurements were performed on a 1.5-T (clinical magnetic field strength) bench-top Bruker minispec relaxometer, and the results were averaged based on three batches of ZES-SPIONs that were individually synthesized using the same protocol. After relaxivity measurements, ZES-SPIONs were digested by hydrochloric acid and their iron concentrations were determined by bathophenanthroline assay according to the protocols in our previous publication (3). As shown in Table S2, batch ① (e.g., used for MRI and MR angiography): r₁ = 5.0 mM⁻¹⋅s⁻¹, r₂ = 9.9 mM⁻¹⋅s⁻¹, r₂/r₁ = 2.0; batch ② (e.g., used for ⁵⁹Fe-labeled biodistribution): r₁ = 4.1 mM⁻¹⋅s⁻¹, r₂ = 7.8 mM⁻¹⋅s⁻¹, r₂/r₁ = 1.9; batch ③ (e.g., used for mouse urine study): r₁ = 6.6 mM⁻¹⋅s⁻¹, r₂ = 13.8 mM⁻¹⋅s⁻¹, r₂/r₁ = 2.1. Therefore, on average, the relaxivities of ZES-SPIONs at 1.5 T in PBS are as follows: r₁ = 5.2 mM⁻¹⋅s⁻¹, r₂ = 10.5 mM⁻¹⋅s⁻¹, r₂/r₁ = 2.0. Moreover, in a mouse urine study, ZES-SPIONs (batch ③) were i.v. injected into four mice separately, and their urine was collected after 2.5 h. The relaxivities of ZES-SPIONs in urine (after renal clearance) were again measured on a 1.5-T bench-top Bruker minispec relaxometer, and their iron concentrations were also determined, yielding the results: r₁ = 4.7 mM⁻¹⋅s⁻¹, r₂ = 9.1 mM⁻¹⋅s⁻¹, r₂/r₁ = 1.9. During the iron determination process of urine samples, we note that the acidified mouse urine generally forms white precipitates and has a brown color of supernatants, upon the addition of hydrochloric acid. So the as-determined iron concentrations of urine samples may include some systematic errors, and therefore we use the r₂/r₁ ratio, which is independent of iron concentration results, for comparison.

In Vivo MRI of Rat (Fig. S4A) and Mouse (Fig. S4B)

MRI studies were accomplished using small-animal MRI systems at 7 T (Bruker) with a mouse body coil and at 9.4 T (Bruker) with a rat body coil. The dosage was 0.2 mmol [Fe]/kg animal. Mice and rats were anesthetized by the inhalation of 98% oxygen and 2% isoflurane (500 mL/min). The respiration rate was maintained at 30 beats per min by adjusting the concentration of isoflurane, and mice and rats were kept warm at ∼37 °C in the experiments. Respiration gating on MRI was enabled by the use of a pressure pad (SA Instrument). On the 7-T scanner, sagittal 3D spoiled gradient echo sequence (FLASH) was used: field of view (FOV), 50 mm × 29.7 mm; matrix, 256 × 152; number of slices, 10; slice thickness, 220 µm; T_R = 6 ms; T_E = 2 ms; flip angle, 30°; number of average, 1; dynamics, 40; resulting in a scan time of 30 min by a temporal resolution of 46 s per 3D volume and a voxel size 195 × 195 × 220 µm³. For scanning at 9.4 T, the following parameters were used: multi slice multi echo sequence (MSME); FOV, 120 mm × 120 mm; matrix, 256 × 256; spatial resolution, 469 × 469 µm; number of slices, 12; slice thickness, 2 mm; T_R = 298 ms; T_E = 10.7 ms; flip angle, 166.5°; number of averages, 2; scan time per image, 2 min, 32.6 s. Eight scans were collected and were spaced 5 min apart for the total experiment time of ∼1 h. By using MRI scanners, the T₁-weighted MR images of rat and mouse injected with ZES-SPIONs are shown in Fig. S4. It can be seen in Fig. S4A that ZES-SPIONs first circulate in the blood vessels and that, beginning at 10 min postinjection, ZES-SPIONs begin to accumulate in the bladder (red oval) showing strong positive contrast, indicating a renal clearance and a high T₁ contrast. A similar phenomenon was observed in mouse with a 7-T MRI scanner: in Fig. S4B, at 6 min postinjection, ZES-SPIONs are observed in the heart (blue oval) and kidney (yellow oval), showing positive contrast. At 12 min postinjection, ZES-SPIONs begin to accumulate in the bladder (red oval). After the ZES-SPIONs are cleared from heart and kidney, a second injection of ZES-SPIONs was introduced. At 5 min post-second injection, ZES-SPIONs are again observed in the heart (blue oval) and kidney (yellow oval).

SNR was measured by the mean signal intensity of a region of interest divided by the SD in air. CNR is the difference of two SNR values. Relative signal enhancement is the signal intensity change due to the injection divided by the signal intensity before the injection. In Fig. 4 angiography data, the brightness in the images was adjusted by using the auto window function of ImageJ by putting a region of interest over the aorta descendants and the surrounding tissue: baseline images were averaged and used to be subtracted from the images to produce background-free images. In the next step, maximum intensity projections under various view angles were produced.

Quantitative Analysis of Mouse Urine with Contrast in the Bladder

A quantitative analysis was performed on the MRI scan in Fig. 4 A–J. For a specific time point, the area of urine (region of interest with positive contrast) was first calculated for each sagittal slice and then multiplied by the thickness of one sagittal slice (220 μm); the volume of urine with contrast is given by the sum of the results of all sagittal slices. Fig. S5A shows that iron oxide nanoparticles begin to accumulate in the bladder at 5 min, increasing until the last time point at 23 min. The volume of urine with contrast is expected to decrease upon the excretion of urine at later time points, as shown in Fig. 4K. Fig. S5B reports that the blood half-life of ZES-SPIONs in mice is ∼19 min.

Quantitative Biodistribution of ES-SPIONs by ⁵⁹Fe Labeling

Incorporation of ⁵⁹Fe into the PEG-coated ES-SPION (intermediate ligand during ligand exchange) was performed according to previously published methods (48). ⁵⁹Fe incorporation into the ES-SPION inorganic core was confirmed by GFC, which showed radioactivity solely in the fraction containing ES-SPIONs. Five mice received tail vein injection of 0.02 mmol [Fe]/kg of radiolabeled ES-SPIONs followed by whole-body and cage radioactivity measurement at 4 and 24 h, during which time mice were awake. After 24 h, mice were euthanized in accordance with approved institutional protocols at University Medical Center Hamburg-Eppendorf, and organs were harvested for individual organ radioactivity measurements.

XRD of Ferumoxytol

XRD measurement was performed on ferumoxytol (>2-y shelf life) by using a PANalytical X’Pert Pro platform. Its composition was analyzed by PANalytical HighScore Plus software, yielding the result of 68% magnetite and 32% maghemite. This result suggests that the majority of ferumoxytol is magnetite, which remained stable in air over an extended period.