Perfluorocarbon Nanoparticles for Physiological and Molecular Imaging and Therapy

Abstract

Herein we review the use of non-nephrotoxic perfluorocarbon nanoparticles (PFC NP) for noninvasive detection and therapy of kidney diseases, and provide a synopsis of other related literature pertinent to anticipated clinical application. Recent reports indicate that PFC NP allow quantitative mapping of kidney perfusion, and oxygenation after ischemia-reperfusion injury with the use of a novel multi-nuclear ¹H/¹⁹F magnetic resonance imaging (MRI) approach,. Furthermore, when conjugated with targeting ligands, the functionalized PFC NP offer unique and quantitative capabilities for imaging inflammation in the kidney of atherosclerotic ApoE-null mice. Additionally, PFC NP can facilitate drug delivery for treatment of inflammation, thrombosis, and angiogenesis in selected conditions that are comorbidities for to kidney failure. The excellent safety profile of PFC NP with respect to kidney injury positions these nanomedicine approaches as promising diagnostic and therapeutic candidates for treating and following acute and chronic kidney diseases.

A. Introduction

In this brief review, we introduce the topic of nanomedicine approaches for detection and potential therapy of acute and chronic kidney diseases. The prevalence and severe consequences of ischemia, hypoperfusion, and hypoxia on renal function mark these pathological substrates as clear targets for new approaches to ameliorate or preserve kidney function, especially in acute kidney injury where diagnosis is challenging and therapy primarily supportive. Because the kidney is a highly metabolic organ, protracted bouts of ischemia or hypoxia can initiate an injury program leading to extensive cellular necrosis and/or apoptosis, especially under circumstances of mixed acute and chronic disease, or in the face of comorbidities such as diabetes or atherosclerosis or sepsis. Accordingly, sensitive noninvasive methods to detect and monitor intrarenal blood flow and tissue oxygenation could be a boon to emerging therapeutic regimens designed more specifically to interdict the progression of renal dysfunction at its earliest stages.

Unfortunately, quantitative methods for assessing kidney hypoxia are neither well developed nor commonly deployed in the clinic. Traditional blood and urine markers, such GFR or kidney injury molecules, do not provide direct measures of kidney hypoxic injury. Even such metrics as total oxygen consumption based on the difference in arterial and venous pO₂ appears insensitive to kidney hypoxia because less than 10% of delivered oxygen is needed for kidney metabolism.¹ Although contrast-enhanced CT or MRI could provide multiple indexes (e.g. time of peak perfusion, blood volume) for assessing kidney perfusion,^2–4 the increased risk of contrast agent induced kidney failure or nephrogenic systemic fibrosis limits the application of these imaging methods for patients with existing kidney diseases.^5–8

In the last two decades, non-invasive BOLD and arterial spin labeling MRI have been applied to image kidney hypoxia and perfusion, which might obviate the need for exogenous contrast agents.^9–13 The BOLD MRI approach maps blood oxygenation based on the increased ratio of paramagnetic deoxyhemoglobin in the capillary bed of hypoxic tissue that elicits shortening of proton T2* relaxation times.² However, BOLD readouts are sensitive to many factors such as the blood volume, flow rate, magnetic field inhomogeneity, or image "susceptibility" artifacts induced by hemorrhage. Moreover, the BOLD readout may not be precisely correlated with the severity of kidney dysfunction at various stages of chronic kidney disease ¹⁴. The arterial spin labeling approaches have been proposed to elucidate kidney perfusion by quantifying the intra-renal signal derived from magnetically prepared or "labeled" segments of flowing blood.^{15, 16} Although the technique might provide an accurate and reproducible readout of renal perfusion, its low signal-to-noise features are problematic for clinical application. Given these limitations, ultimately it may be the case that multi-parametric MRI with the use of BOLD, spin labeling, and/or other techniques could improve detection and quantification of intrarenal perfusion defects but at the cost of time and expense required to perform the linked studies.^{11, 17, 18}

Multinuclear ¹H/¹⁹F MRI coupled with perfluorocarbon (PFC) based nanoparticles (NP) represent an alternative readout applicable for imaging kidney hypoperfusion and associated injury (Figure 1).^{19, 20} PFC NP are a class of clinically approved vehicles with no apparent toxic effect on the kidney.^{21, 22} As a blood-pool agent, the linear relationship between the ¹⁹F signal intensity and PFC NP’ concentration allows direct measure of blood volume in tissue.^23–25 In addition, the relationship between the ¹⁹F relaxation time (R₁= 1/T₁) of PFC NP and dissolved O2 that alters ¹⁹F relaxation linearly according to pO₂, has been utilized for in vivo assessment of pO₂ in blood and tissue^26–28. Finally, functionalized PFC NP with the use of selected targeting ligands that can bind to and detect inflammatory markers (e.g. VCAM-1) may also allow early detection of inflammatory kidney diseases. This review offers an overview of diagnostic ¹H/¹⁹F MRI with the use of PFC NP, as well as insights into combined therapeutics with the same ¹⁹F nanosystem carrying drug and gene cargoes.

Figure 1.

Paradigm for physiological and molecular imaging with perfluorocarbon nanoemulsions. Reprinted from Wickline et al.²⁰ with permission.

B. PFC NP For Imaging

PFCs are a group of ¹⁹F-containing compounds derived from hydrocarbons by partial or complete substitution of ¹H with ¹⁹F.²⁹ PFCs are biologically stable, inert, well tolerated, and can carry a significant payload of O2 in dissolved form, which established their original use as blood substitutes. Liquid PFCs also have been employed as an organ preservation media for kidney transplant.^{30 31, 32}

For in vivo applications, PFCs typically are emulsified into a nanoparticle form with a surfactant coating for stabilization and functionalization (Figure 1). A typical PFC NP emulsion contains 40% (v/v) PFOB, 2% (w/v) safflower oil, 2% (w/v) of surfactant commixture, 1.7% (w/v) glycerin, and water balance.³³ Thus the final formulated PFC NP comprises principally a liquid PFC core encapsulated by a lipid monolayer with a diameter of 150–250 nm and a variable surface charge depending on the types of phospholipids and surfactants used to construct it (e.g., −40 to +20 mV).

The use of PFC NP for examining tumor hypoxia has been pursued for some time and recent excellent reviews are available on the subject from pioneers in this area such as Mason.^{26, 34–37} In most of these prior applications, targeting ligands have not been included as direct injections were often required to provide sufficient signal for detection of the ¹⁹F signal in order to generate proof of concept for pO₂ measurements. However, as interest in the broad field of "molecular imaging" was emerging around that time, our group was the first to explore the large signal generating capacity of a nanoparticle when linked to chelated lanthanides to formulate a robust paramagnetic "bright spot" molecular MRI contrast agent.^{38, 39} Because the PFC NP technology was well known and offered a ready substrate for functionalization with its surrounding lipid-surfactant surface layer, it was relatively straightforward to consider adding gadolinium complexes to the surface in large quantities (e,g., 100,000–200,000 Gd per particle), which would provide substantial signal amplification upon localization to molecular binding sites. In addition to the large cargo of Gd, a molecular targeting ligand could be added as well, where both imaging agent and targeting ligand were covalently coupled to a phospholipid anchor that was directly formulated into the particle in a single emulsification step (see specific descriptions below).

Furthermore, this PFC NP construct also was developed as the first reported ultrasound molecular imaging agent due to special features of its composition that render it visible to mechanical waves at clinically relevant imaging frequencies.⁴⁰ Fluorescence and SPECT imaging capabilities also exist for the PFC NP system when it is coupled to contrast moieties appropriate to those modalities.^{41, 42}

As compared with iodinated or Gd contrast agents, PFC NP could be well suited to patients with kidney disease. PFC NP exhibit a good safety profile with minor dose dependent flu-like symptoms or fever in 24 hours after dosing. ^{43, 44} No renal toxicity was reported in animals and human.²² With a nominal diameter of 250 nm, PFC NP are not susceptible to glomerular filtration but rather are removed from the circulation primarily by the reticuloendothelial system and then the PFC component ultimately is vaporized through respiration.²¹ The blood clearance half-life of PFC NP ranges from 3 to 42 hours depending the exact preparation method,⁴¹ thereby providing sufficient time for MRI detection.

Despite the apparent utility of nanostructures carrying large payloads of Gd for T₁ weighted MRI, it was also clear that the fluorine core of the PFC NP might provide an alternative approach for magnetic resonance detection. This could represent a unique imaging signature because there is no detectable ¹⁹F in the body at typical clinical field strengths. For imaging there would be no background signal as is the case for other contrast agents that rely on their effect on proton relaxation to create contrast.⁴⁵ In the case of ¹⁹F imaging, quantification of the amount of PFC NP in the imaging field of view is possible by directly assessing the spectral energy contained within the ¹⁹F readout, which is proportional to the ¹⁹F concentration, and thus the number of PFC NP containing them.^46–49 Thus, the non-targeted PFC NP agents have been developed for blood pool imaging and perfusion,^{46–48, 50} cellular (stem cell, dendritic cell, etc.) labeling and tracking,^{49, 51–55} cellular and tissue uptake for imaging inflammation and allograft rejection, among others. ^{56, 57}

Elucidation of Kidney Function and Injury with PFC NP

Hypoxic/ischemic injury is a common pathology in acute and chronic kidney diseases.^{35, 58–60} Some contributing factors include: interstitial fibrosis that increases the diffusion distance of O2 from capillaries to tubular epithelial cells,⁶¹ vessel rarefraction that reduces the capillary density,⁶² arterial stenosis that can reduce the overall amount of blood delivered to the kidney,⁶³ and acute thrombosis that blocks regional blood flow. Hypoxic/ischemic injury may develop regionally, e.g. in the cortex or medulla,⁶¹ but nevertheless still affect the global renal function. This regional manifestation of the disease process can pose a challenge to the precise elucidation of the disease process and stage when only global kidney functional readouts are available such as GFR or blood/urine markers.⁶⁴ We therefore have proposed that non-invasive regional mapping of kidney oxygenation could be beneficial for understanding the disease process and for clinical management that may be stage and molecular pathway specific.

Quantitative ¹⁹F MRI using PFCs or PFC NP has been extensively exploited for in vivo imaging of hypoxia.^{29, 35} PFCs can carry a high payload of O₂ that exhibit a linear dissociation curve, which renders them potential reporters for local pO₂ such that at a given temperature, the partial pressure of dissolved paramagnetic O₂ in PFCs is linearly correlated with the ¹⁹F relaxation rate.⁶⁵ The exchange of O₂ between PFCs and surrounding media occurs through free diffusion operating on a millisecond timescale.⁶⁶ It is straightforward to generate a calibration curve specifying the relationship between ¹⁹F R₁ and pO₂ at a given field strength in order to read out the pO₂ to provide a non-invasive measure of tissue oxygenation.⁶⁷ In solid tumor where tissue temperature is heterogeneous, hexafluorobenzene and perfluoropolyether that exhibit minimal sensitivity of ¹⁹F R1 to temperature changes between 30 and 42°C are typically used for ¹⁹F MRI oximetry.⁶⁸ The precision of ¹⁹F MRI oximetry may be within 1–3 mmHg and the absolute readout of pO₂ is comparable to that measured with fiberoptic,⁶⁹ and near-infrared spectroscopy^{70, 71} in tumor tissues.

Recently, fast ¹⁹F MRI techniques like FREDOM (Fluorocarbon Relaxometry using Echo planar imaging for Dynamic Oxygen Mapping) and BESR (Blood Enhanced Saturation Recovery) have been developed to facilitate in vivo mapping of pO₂,^72–74 which should be applicable for kidney imaging and pO₂ measurements. Our approach using the BESR method mirrors some existing methods for defining proton relaxivity in vivo but takes advantage of flowing blood to quantize the ¹⁹F signal that is moving rather than stationary, which facilitates correlation of the detected signal with blood volume, blood flow, and regional pO₂.⁷⁴ Figure 2 shows the effect on the MRI signal intensity for ¹⁹F in circulating PFC NP as it is affected by blood pO₂. The presaturated (darker cardiac chambers at left side of panels) ¹⁹F spins in the heart recover (i.e., brighten) more quickly as the signal is measured over 2 seconds under conditions of hyperoxia (lower panels) than of normoxia (upper panels). Moreover, the signal in the RV recovers more slowly than that in the LV due to the inherently lower pO₂ in the RV chamber under all conditions. The quantification of LV and RV pO₂ matches well with the expected values under each condition. These types of measurements in the kidney should now be possible in vivo in actual patients due to the recent implementation of similar approaches on clinical 3T MRI scanners that can acquire simultaneous ¹H and ¹⁹F signals for mapping PFC NP concentrations, regional pO₂,⁷⁵ arterio-venous differences in pO₂⁷⁴ or blood volume fraction,⁷⁶, therefore to derive global renal oxygen consumption.

Figure 2.

Left¹⁹F BESR images of the mouse heart under hyperoxia and normoxia illustrating signal due to 19F PFC NP that is responsive to blood pO2. Right: Measured blood pO₂ based on measured ¹⁹F T1 of PFC NP showed pO₂ difference between LV and RV, and between normoxic and hyperoxic conditions for each ventricle. Reprinted from Hu et al. ⁷⁴ with permission.

Moving on to the kidney itself, we have employed an acute kidney injury model and applied ¹⁹F MRI in conjunction with PFC NP to assess ischemic injury and the consequences of regional hypoperfusion.⁷⁷ The objective here is to develop a multi-nuclear ¹H/¹⁹F MRI approach for functional renal imaging. We first demonstrated the feasibility of ¹⁹F MR mapping of renal blood volume and pO₂ in healthy mouse kidneys (Figure 3. Left panel). Subsequently, a unilateral renal ischemia-reperfusion model of AKI⁷⁸ to demonstrate that transient ischemia damages endothelial and tubular cells causing sustained medullary microvessel non-perfusion⁷⁹ and tubular ischemia.^80–82 The result of this study showed that ¹⁹F MRI sensitively detected decreased renal blood volume in the cortical-medullary (CM) junction, where ¹H BOLD MRI also detected increased T2* reflecting reduced susceptibility effect from blood flow. (Figure 3. Right panel). The follow-up histology analysis confirmed a reduced density of perfused blood vessels in the CM junction of injured kidneys. The intrarenal blood pO₂, however, remained unchanged, which may attributed to the capability of kidney to auto-regulate perfusion to maintain oxygenation despite the reduction of blood flow.^{1, 83}

Figure 3.

Representative ¹H BOLD T2*-weighted image, ¹⁹F spin density weighted image and ¹⁹F T1-mapping derived pO₂ map in the left control (A–C) and right injured (D–F) kidneys of the same mouse. The inserted panel in (F) shows ¹⁹F MRI detected pO₂ map of another mouse kidney during ischemia. White arrow points to the CM junction of the injured kidney where abnormally increased ¹H T2* and decreased ¹⁹F signal intensity were detected. Reprinted from Hu et al.⁷⁷ with permission.

Our previous work have demonstrated the feasibility of in vivo ¹⁹F MRI angiography in rabbits on a 1.5T Philips clinical MR system.⁵⁰ Given the high blood volume in the kidney and the availability of 3T clinical MR consoles, ¹⁹F MRI assessment of kidney blood volume and pO₂ should be achievable in clinical settings.

Potential Molecular Imaging Applications For Targeted PFC NP

The panoply of molecular imaging applications for kidney disease that can be addressed with the use of targeted contrast agents is extensive. Here we review some indications for targeted PFC NP that are directly related to or associated with important comorbid conditions that apply to progressive renal disease.

Atherosclerosis

The role of atherosclerosis in predisposing to acute and chronic renal failure under many conditions is clear.^{60, 84} Moreover, the inflammatory milieu in both early and advanced plaques can be addressed with ¹⁹F imaging in several ways. We have shown recently that non-targeted PFC NP can diffuse passively into the core of intimal plaques where vascular barrier disruption has occurred over time as a consequence of endothelial damage, apoptosis, and cellular sloughing.⁵⁶ Indeed, in the later stages of the disease process in fat fed rabbits for example, massive infiltration of nanoparticles into the plaque extending up to the medial boundary can occur in vivo, and be detected with ¹⁹F MRI, indicating the ready access of blood coagulation elements to a hypercoagulable inflamed plaque milieu, predisposing to acute and/or chronic thrombosis (Figure 4). Similar passive penetration into human carotid endarterectomy samples excised for clinical indications in symptomatic patients mimicked the experimental detection of endothelial barrier disruption. These data suggest that vascular endothelial disruption in acute kidney injury might similarly be amenable to quantification with ¹⁹F MRI, as we showed above for nanoparticles passively penetrating the renal interstitium in ischemic injury.

Figure 4.

MRI imaging and quantification of NP signal in cholesterol plaques (A) Left: 3D saggital rendering of nanoparticle signals from aorta of 12 mo cholesterol fed rabbit after 12 hours circulation time in vivo.¹⁹F MR image registering nanoparticle fluorine cores (color) overlaid on ¹H MR image (gray) of aorta show intimal (I) location of particles trapped in thickened plaque (color) atop the medial (M) layer (gray). Particle concentration per voxel is coded in nM (scale bar). Right: Close up of intimal layer. (B) ¹H, ¹⁹F and overlay MR transverse images of aortic rings with nanoparticles trapped in intima of thickened plaque from 9 mo cholesterol fed rabbit after 2 hour circulation in vivo. Black artifacts are small air bubbles. Note lack of ¹⁹F signal from more normal adjacent tissue sections. (C) Comparison of normalized CE NP number to the endothelial surface area between 3 month diet and >7 month diet rabbit aorta samples showing 10 fold greater accumulation in older plaques. Reprinted from Zhang et al. ⁵⁶ with permission.

Alternatively, angiogenesis imaging is possible with a number of modalities including MRI, PET, SPECT, and ultrasound.^85–87 The role of plaque neovasculature that is driven by tissue/macrophage hypoxia is known to be required for early plaque growth and predisposes to late plaque rupture and subsequent thrombosis and acute vessel occlusion.^88–92 For example, in experimental atherosclerosis in fat fed rabbits, the neovasculature emerging from the inflamed adventitial vasa vasorum can be detected and quantified by MRI with Gd-conjugated NP targeted to upregulated endothelial integrins (Figure 5) at clinical field strengths.^{39, 93–95} The same agents have been used for tumor angiogenesis imaging, which also have been detected in vivo with ¹⁹F MRI.⁹⁶

Figure 5.

Top Panel: MRI of neovasculature in rabbit plaque with α_vβ₃-targeted nanoparticles. Significantly lower angiogenesis one week after treatment with targeted fumagillin particles. Rabbits treated with non-targeted fumagillin nanoparticles show high levels of angiogenesis one week post treatment, similar to particles without fumagillin. Bottom Panel: Angiogenesis in rabbit atherosclerotic plaque. (A) Histological section of abdominal aorta from control animal demonstrating intimal plaque and strong staining for α_vβ₃-integrin (LM-609) at the media-adventia and media-intima interfaces. Angiogenic vessels (arrows) are observed in the vasa vasorum (B), media (C) and intimal plaque (D). Reprinted from Winter et al. ¹²⁰ with permission.

Neovasculature expressing heterodimeric molecular epitopes such as α_vβ₃ integrins also has been detected sensitively with ultrasound imaging in experimental models.^97–99 In a genetic model of squamous cell carcinoma, extremely limited extents of inflammatory neovasculature could be demonstrated in the mouse pinna with high resolution ultrasound imaging systems and special signal processing algorithms that harbor quantitative potential based on the "information content" of the data stream.^98–100 Both pathological (e.g., atherosclerosis, cancer, inflammation) and tissue reparative processes (e.g., recovery from ischemic injury) invoke significant angiogenesis that might be candidates for renal imaging,

An array of other inflammatory molecules involved in atherosclerosis might be amenable to noninvasive molecular imaging. For example, we have used ¹⁹F MRI to detect the upregulation of renal vascular VCAM-1 in mice fed a Western diet for 35 weeks (Figure 6).¹⁰¹ Because VCAM and other related adhesion molecules are responsible for trafficking of immune effector cells such as monocytes and neutrophils to areas of damage and inflammation, assessment of the potential for cellular recruitment in kidney injury might be useful disease staging and optimization of therapeutic approaches. Ultrasound approaches with targeted microbubbles containing a gaseous perfluorocarbon also have been demonstrated for renal VCAM imaging,¹⁰² although there is lesser potential for quantification.

Figure 6.

The ApoE^−/− kidney is characterised by significant up-regulation of VCAM-1 expression and a marked infiltration of macrophages into the glomeruli. A representative ¹H MR image (A) showing structural detail, ¹⁹F MR image (B), and a composite ¹H/¹⁹F image (C) showing VCAM-1-targeted nanoparticle accumulation in atherosclerotic ApoE^−/− (top) and wild-type control (bottom) kidneys, imaged at 12T. Voxels with a signal intensity greater than twice the standard deviation were defined as positive for ¹⁹F to produce a distribution map of ¹⁹F throughout each kidney, which corresponds to nanoparticle binding via VCAM targeting. Data represent mean (n=6±SD). *represents p<0.05. Reprinted from Southworth et al.¹⁰¹ with permission.

Thrombosis

Fibrin detection in formed clots also has been demonstrated in vivo with ultrasound at clinical frequencies after binding of PFC NP targeted to selected fibrin epitopes, again with a monoclonal antibody.^{38, 40, 103} Because fibrin clots are barely echogenic and therefore difficult to detect with clinical ultrasound, this could be useful for renal diseases that are prone to thrombosis such as hemolytic uremic syndrome, AKI, sepsis, atherosclerosis, etc. Additional examples of fibrin imaging have been reported with the use of MRI by detection of either the 1H or ¹⁹F signatures of PFC NP (Figure 7) ^{47, 104–106} As compared with standard ultrasound molecular imaging, the MRI readouts for ¹⁹F are quantitative and can be expressed in molar concentrations of bound nanoparticles or fluorine atoms (see Figure 7) permitting longitudinal determination of the local concentration of targeted molecular epitopes.

Figure 7.

Fibrin targeted PFC NP binding to human carotid endarterectomy specimen in vitro. (A) Optical image of a 5 mm cross-section of a human carotid endarterectomy sample. This section showed moderate lumenal narrowing as well as several atherosclerotic lesions and areas of calcification. (B) A ¹H image acquired at 4.7 T at the same location shows signal enhancement due to the presence of gadolinium on the targeted particles (red arrows). (C) A ¹⁹F projection image acquired at 4.7 T through the entire carotid artery sample shows high signal in the same areas due to nanoparticles bound to fibrin. (D) ¹H image in B with a false color overlay of the quantified nanoparticle concentration in the carotid as derived from the ¹⁹F image. Reprinted from Morawski et al.³³ with permission.

The presence of thrombin itself in newly forming clots can be delineated with thrombin targeted PFC NP imaged with MRI.¹⁰⁷ Thrombin is the key regulator of clot formation because it converts fibrinogen to fibrin locally after being activated on an acute clotting surface by factor Xa. Nanoparticles bearing a small, modified peptide, PPACK (phenylalanine-proline-arginine-chloromethylketone), can covalently couple to the active pocket of thrombin molecules within forming clots and can be imaged dynamically and quantified with ¹⁹F MRI (Figure 8). Another advantage of this approach concerns the role of thrombin as a signaling molecule operating through protease activated receptors (e.g., PAR-1) that participate in signaling events in inflammation, atherosclerosis, and thrombosis, potentially conferring the capability to noninvasively delineate certain molecular regulatory pathways that might attend selected renal pathologies.

Figure 8.

In mice receiving treatment with control nanoparticles (n=3) or PPACK nanoparticles (n=3), both carotid arteries were excised following induction of occlusive thrombi in the right carotid artery. (A) ¹⁹F MRI at 11.7T exhibited coregistration of ¹⁹F signal from PPACK nanoparticles with ¹H images depicting the occlusive clot in the artery. ¹⁹F MRS was used to quantify retention of nanoparticles in the injured right carotid artery (RA) and the unharmed left carotid artery (LA) for the two tested nanoparticle treatments. (B) Retained are represented in particles mean ± standard error. Reprinted from Myerson et al.¹²¹ with permission.

Ultrasound molecular imaging with targeted liquid PFC NP has been reported to detect other constituents of thromboembolic disease. For example, Tissue Factor (TF), which is the proximate cause of thrombosis during plaque rupture and is also associated with microangiopathic renal thrombosis in AKI and other kidney diseases, is detectable in vascular segments exposed to balloon injury that incites inflammation and upregulation of TF in the media layer (Figure 9).^108–111 In this case, the selective binding and deposition of PFC NP are targeted to TF with the use of a specific conjugated monoclonal antibody fragment.

Figure 9.

High-frequency intravascular ultrasonic images of carotid arteries exposed to tissue factor–targeted or control emulsion nanoparticles after angioplasty. (A) Bright acoustic enhancement (arrows) in a portion of the vessel wall consistent with nonuniform overstretch injury with the balloon. (B) Lack of acoustic enhancement in the vessel wall after control contrast. Similar results were observed in each of 3 animals. Reprinted from Lanza et al.¹⁰⁸ with permission.

Cellular imaging with PFC NP labeled immune, stem, and vascular precursor cells represents another opportunity for addressing inflammatory pathologies or regenerative approaches for therapy. A variety of cell types are capable of being isolated, subjected to PFC NP labeling by endocytosis in vivo, and then reinjected either locally into tissues or systemically where they can be tracked by ¹⁹F MRI.^{49, 51, 52, 55, 112–114} Figure 10 shows an example of two distinct PFC labels in endothelial precursor cells that contribute to tumor angiogenesis, and illustrates the opportunity for tracking multiple cell labels simultaneously. In other example, Ahrens and collaborators have mapped acute allograft rejection with ¹⁹F labeled immune cells.⁵⁷ Such strategies might exhibit clear significance for kidney transplantation or for defining immunologically driven kidney pathologies.

Figure 10.

(A) The different ¹⁹F spectrospic features between PFOB and CE allows selective detection of PFOB or CE NP labeled cells. (B) To determine the feasibility of detecting cells at specific tissue sites after local delivery, one million cells (2×10⁴/µL) loaded with CE or PFOB NP were injected into mouse thigh skeletal muscle. At 11.7T, the PFC labeled cells were imaged rapidly in approximately 4 minutes by selectively tuning to PFOB or CE. (C) Overlaying the fluorine images atop a conventional matched proton image reveals the ¹⁹F PFOB and CE signals are located in the left and right mouse thigh, respectively. Reprinted from Partlow et al. ⁴⁹ with permission.

A limitation for molecular ¹⁹F MRI is the low signal from targeted/localized PFC NP. To reduce imaging time in clinical settings, Gd-DTPA conjugated PFC NP maybe used for fast molecular ¹H MRI to detect targeted/localized PFC NP, while a low resolution ¹⁹F MRI or MRS can be used to quantify the amount of PFC NP. Since Gd-DTPA is associated with increased frequency of acute adverse reactions, the safety profile and optimal dosage of Gd-DTPA loaded PFC NP in patients with kidney injury needs to be evaluated.

Therapeutics with PFC NP

A brief mention of the therapeutic utility of PFC NP is warranted, as both imaging and therapy can be combined to good effect with this system due to the ability to quantify drug deposition based on ¹⁹F spectroscopy. In other words, by knowing how much drug or gene product is present per nanoparticle and measuring the concentration of NP by ¹⁹F MRI as per reported methodologies,^{47, 48, 115} one can estimate the dose of drug deposited at any given time in an imaging voxel. A selected listing of some of the agents that have been used in various diseases in conjunction with therapeutic PFC NP is shown in Table 1. Moreover, it is worth noting that most of the formulations have been molecularly targeted to enhance local deposition and retention in the selected tissues with the use of preselected targeting ligands that are prepared in a one step process along with the therapeutic agent. However, we have demonstrated that it is feasible to prepare a base imaging and/or therapeutic agent and choose the targeting ligand later for "postformulation targeting" with a membrane inserting peptide cargo carrier,^116–118 which allows great flexibility in selection of targets based on changing conditions as diseases progress through different stages. This feature is expected to facilitate more personalized approaches to nanomedicine in the future.

Table 1.

The application of therapeutic PFC NP in various diseases

Therapeutic Agent	Application	Molecular Target	Reference
PPACK (phenyalanine-proline-arginine-chloromethylketone	anti-thrombosis	thrombin	121
Bivalirudin	anti-thrombosis	thrombin	107
Urokinase	thrombolysis	fibrin	122
Fumagillin	anti-angiogenesis in atherosclerosis	Endothelial aminopeptidase	120, 123
Fumagillin	anti-angiogenesis in cancer	Endothelial aminopeptidase	95, 124, 125
Fumagillin	anti-angiogenesis in arthritis	Endothelial aminopeptidase	126–128
Rapamycin	Restenosis prevention, anti-inflammatory	mTOR	129
Rapamycin	Muscular dystrophy	Autophagy enhancer (via mTOR)	130
Paclitaxel/Doxorubicin	Vascular inflammation, smooth muscle cell inhibition	Tissue Factor	131
Anti NFκB peptides (NBD)	Inflammation, cancer	NFκB (p65)	132
Melittin peptides	cancer	Cell lipid membrane disruptor	116, 133
siRNA	various	mRNA	134

Limitations and Challenges

For imaging renal perfusion, ¹⁹F MRI requires longer imaging time (minutes to tens of minutes) than ¹H MRI techniques like ASL and DCE enhanced MRI (takes seconds to a few minutes). The use of PFC NP as a contrast agent will also incur extra cost. However, the unique advantage of ¹⁹F MRI is its capability to directly mapping intrarenal blood volume since the 19F signal intensity is linearly correlated with the quantity of PFC NP in each voxel. When a ¹⁹F inversion-recovery gradient echo sequence was employed to measuring pO₂, the ¹⁹F signal from PFC NP are contributed by both artery and veins. ⁷⁴ Alternatively, the signal from fast moving PFC NP in arterial blood flow can be suppressed by employing spin echo acquisitions. ⁷⁷ The detected ¹⁹F signal are primary contributed by PFC NP in capillaries and veins. Similar to the limitation of BOLD MRI, the limitation of ¹⁹F MRI is that it detects blood pO₂. It is yet unknown how accurately blood pO₂ reflects tissue pO₂.¹¹⁹ Finally, translating PFC NP to clinical settings is limited by the low sensitivity of ¹⁹F signal on the “low field” (e.g. 3T) clinical MR scanner, which may require prolonged imaging time (tens of minutes) to achieve satisfactory signal to noise ratio. This will pose practical problems to image AKI patients who are in critical condition and require intensive care. The motion artifact from bowel movement and respiration during image acquisition should be also corrected for image reconstruction. ⁹⁶

Clinical Summary.

• Perfluorocarbon nanoparticles (PFC NP) are a non-nephrotoxic molecular imaging agent for noninvasive quantitative assessment of kidney function and injury.

• Functionalized PFC NP allow site-targeted drug delivery as a novel therapeutic modality for treating acute and chronic kidney diseases.