Positively Charged Amphiphilic Chitosan Derivative for the Transscleral Delivery of Rapamycin

Abstract

Purpose.: We explored the potential of an amphiphilic chitosan derivative to facilitate the transscleral delivery of rapamycin, a potential multitherapeutic agent with poor water solubility.

Methods.: The amphiphilic chitosan derivative, O-octanoyl-chitosan-polyethylene glycol (OChiPEG) graft copolymer, was analyzed using Fourier-transform infrared spectroscopy (FT-IR). OChiPEG micelles were prepared via the thin film method and characterized for their size using dynamic light scattering (DLS), zeta potential using laser Doppler velocimetry (LDV), morphology using transmission electron microscopy (TEM), drug entrapment efficiency (EE), and drug loading (DL) efficiency using reversed-phase high performance liquid chromatography (RP-HPLC), critical micelle concentration (CMC) using spectrofluorometry, and thermal properties using differential scanning calorimetry (DSC) and x-ray powder diffraction (XRPD). Scleral permeation and retention of rapamycin from the drug-loaded micelles were determined in porcine sclera clamped in Ussing chambers, using RP-HPLC.

Results.: Conjugation of hydrophilic and hydrophobic groups to chitosan was confirmed using FT-IR. Rapamycin-loaded micelles of particle size 40.6 nm and zeta potential + 6.84 mV were prepared successfully. These carriers exhibited a high EE and DL of 85.6 and 16.3%, respectively, and a CMC of 16.6 μM. OChiPEG micelles showed a high rapamycin scleral retention (14.8 ± 0.81 μg/g) with successful transscleral permeation (5.57 ± 1.04 × 10⁻⁸ cm²·s⁻¹).

Conclusions.: Positively charged OChiPEG micelles loaded with rapamycin were prepared successfully. These showed a high scleral retention and successful permeation of rapamycin, and therefore may be useful for the topical delivery of other hydrophobic agents.

Introduction

Rapamycin currently is under investigation in several clinical trials of anterior and posterior segment eye disease. It is a lipophilic macrocyclic triene antibiotic that originates from Streptomyces hygroscopicus, and was identified initially as an antifungal agent for the treatment of Candida albicans . ¹ It was shown later to have immunosuppressive, cytostatic, and anti-angiogenic properties. ^2–4 Therefore, rapamycin is an attractive therapeutic drug for inflammatory and neovascular ocular conditions. However, its poor water solubility (2.6 μg/mL) and the lack of functional groups to favor ionization in solutions of varying pH render it unsuitable for noninvasive topical administration. ⁵ Buech et al. used liposomes, cyclodextrin solution and microemulsions to attempt to solubilize rapamycin. ⁶ However, these formulations did not show ocular permeation. This may have been due to their large sizes as compared to micelles. 

We hypothesized that polymeric micelles could be used as nanocarriers to overcome these limitations, with the additional benefits of increased bioavailability, prolonged drug release, improved stability, and passive and site-specific targeting. ⁷ Furthermore, micelles are spherical and very small in size (typically less than 50 nm) and can be charged through surface modification. These characteristics can be exploited to facilitate scleral permeation and retention, as the sclera contains negatively charged proteoglycans, is hydrophilic, and has been shown to be more permeable to smaller particles. ^8–10  

We selected the amphiphilic chitosan derivative, O-octanoyl-chitosan-polyethylene glycol (OChiPEG) graft copolymer, to prepare rapamycin-containing micelles. These were chosen as the delivery system for rapamycin, because in addition to the aforementioned benefits they provide several other potential advantages. They contain chitosan, a polycationic biodegradable mucoadhesive polymer that enhances scleral permeation and retention, and may enhance the therapeutic effects of rapamycin as it exhibits antiangiogenic, antioxidant, and wound healing activity. ^11–16 Furthermore, polyethylene glycol (PEG) also is found within this conjugate, and this has been shown to enhance the formulation's stability, bioavailability and accumulation at the target site. ^7,17,18  

Materials and Methods

Materials

The following materials were used as received and were analytically pure: rapamycin (sirolimus; LC Laboratories, Woburn, MA), acetonitrile (Fisher, Leicestershire, UK), absolute ethanol 99.6% (Haymans Ltd., Essex, UK), chloroform (VWR International, Lutterworth, UK), chitosan oligosaccharide (molecular weight [MW] ≥5 kilodaltons [kDa], degree of deacetylation [DDA] >90%; Kitto Life, Seoul, Korea), succinic anhydride (Merck Schuchardt OHG, Hohenbrunn, Germany), and N,N′-dicyclohexylcarbodiimide (DCC, 99%; Alfa Aesar, Ward Hill, MA). The following materials were obtained from Sigma-Aldrich (St. Louis, MO): octanoyl chloride (99%), pyrene (≥99%), trifluoroacetic acid (TFA, ≥99%), poly(ethylene glycol) 2000 monomethyl ether (mPEG), methane sulfonic acid (CH₃SO₃H), N-hydroxysuccinimide (NHS, 98%), and sodium bicarbonate. The water used throughout all of the experimental procedures was deionized water (HPLC grade; Fisher). 

Preparation of OChiPEG

Preparation of O-Octanoyl-Chitosan.

O-octanoyl-chitosan was prepared as described by Huang et al. ¹⁹ with modifications. Briefly, 5 g of chitosan oligosaccharide ≥5 kDa were dissolved in 45 mL of methane sulfonic acid (MeSO₃H) under continuous stirring (multiposition magnetic stirrer, RO 10 power; Ika-Werke, Staufen, Germany) at room temperature for 1 hour. The resultant solution was kept at −20°C for 2 hours, after which octanoyl chloride was added drop-wise to this solution. The reaction continued under stirring conditions at room temperature overnight and was stopped by the addition of crushed ice. Sodium bicarbonate then was used to neutralize the acid. The excess sodium bicarbonate and unreacted products were removed by dialysis (3.5 kDa molecular weight cut-off; Medicell International Ltd., Liverpool, UK) in water. Finally, the dialyzed mixture was lyophilized to obtain the O-octanoyl-chitosan powder. 

Preparation of OChiPEG.

mPEG was activated initially by an esterification reaction using succinic anhydride, NHS, and DCC as described by Huang et al. ¹⁹ This then was reacted with O-octanoyl-chitosan, and the reaction mixture dialyzed and lyophilized to produce OChiPEG powder. 

Characterization of OChiPEG by Fourier-Transform Infrared Spectroscopy (FT-IR).

The chemical structure of OChiPEG graft copolymers was analyzed using a PerkinElmer Spectrum 100 FT-IR spectrometer (PerkinElmer, Waltham, MA) covering a range of 400 to 4000 cm⁻¹ at a resolution of 4 cm⁻¹. Data were analyzed using the PerkinElmer Spectrum Express software (PerkinElmer). 

Preparation of Empty and Rapamycin-Loaded Micelles.

Micelles were prepared by the modification of the thin-film hydration method described by Wei et al. ²⁰ Briefly, 10 mg of OChiPEG (with or without 2 mg of rapamycin) were dissolved in 8 mL of absolute ethanol and 1 mL of chloroform in a round bottom flask. The solvents then were evaporated using a rotary evaporator (Hei-VAP Advantage Rotary Evaporator; Heidolph, Schwabach, Germany) at 150 revolutions per minute (rpm), 80°C, and under vacuum (KNF Laboport; KNF Neuberger, Freiburg, Germany) to obtain a thin film. The resultant thin film was hydrated with 8 mL of water (preheated to 55°C) and mixed thoroughly on a 55°C water bath (Buchi B480; Buchi Corp., Danderyd, Sweden) for 5 minutes. This mixture then was bath sonicated (Ultrawave bath sonicator; Ultrawave Ltd., Cardiff, UK) for 10 minutes and left to settle at room temperature for a further 10 minutes before filtering it through a sterile 0.22 μm filter (Millex-MP; Millipore, Carrigtwohill, Ireland) to remove un-entrapped rapamycin. This filtrate was used for further analysis. This filtrate also was lyophilized and analyzed. Samples were protected from light throughout the preparation and in all of the analytical procedures. 

Characterization of Micelles

Analysis of Particle Size and Zeta Potential.

The samples were analyzed for particle size and zeta potential (or charge) using the Zetasizer Nano ZS instrument (Malvern Instruments Ltd., Worcestershire, UK). This system uses dynamic light scattering (DLS) to measure the particle size and laser Doppler velocimetry (LDV) to measure the zeta potential. 

Analysis of Morphology.

The morphology of the particles was observed using a FEI CM 120 Bio Twin transmission electron microscope (TEM; Philips Electron Optics BV, Eindhoven, The Netherlands). Approximately 40 μL of the preparation were placed on a copper grid with a nitrocellulose covering and stained negatively with 1% uranyl acetate. 

Thermal Properties of Micelles.

Differential Scanning Calorimetry (DSC).

DSC (DSC Q2000 module; TA Instruments, New Castle, UK) was used to determine the thermal properties of the samples. Lyophilized preparations were sealed in aluminium hermetic pans (TA Instruments), the lids of which contained a 50 μm pinhole as previously described. ²¹ The powders were analyzed at a heating rate of 20°C/min from 0 to 300°C. They then were cooled down to −20°C and heated back up to 300°C. The pure drug was sealed in a similar pan and analyzed at a heating rate of 10°C/min from 0 to 300°C. It then was cooled down to −20°C and then reheated back up to 300°C. The measurements were done under nitrogen atmosphere at a flow rate of 50 mL/min. 

X-Ray Powder Diffraction Studies.

X-ray diffraction patterns were obtained for the samples using an X-ray diffractometer (Oxford Diffraction Xcalibur novaT X-ray diffractometer; Oxford Diffraction Ltd., Abingdon, UK) which uses Cu Kα radiation. The datasets obtained were processed and scaled using CrysalisPro (Oxford Diffraction Ltd.). The data were collected at room temperature and scanned with a step size of 10°2-theta. 

Determination of the Critical Micelle Concentration Using Spectrofluorometry.

A fluorescence spectrometer (PerkinElmer precisely LS55 luminescence spectrometer; PerkinElmer) was used to measure the critical micelle concentration of the formulation with pyrene as the fluorescence probe. Serial dilutions of the micelles were prepared and mixed thoroughly with pyrene at a concentration of 6 × 10⁻⁷ M. The fluorescent intensity then was measured for each concentration at an emission of 331 nm and an excitation of 334 nm. The emission intensity ratio of I₁/I₃ was plotted against the log of the micelle concentration and the CMC value was taken as the point of intersection from the two tangents drawn to the curve at high and low concentrations. 

Determination of Drug Loading and Drug Entrapment Efficiency.

Reversed-phase high performance liquid chromatography (RP-HPLC) was used to quantify the amount of drug entrapped within the micelles (drug loading [DL] and entrapment efficiency [EE]) and to calculate the drug content that was retained in the sclera and had permeated across it. The DL and EE were calculated using Equations (1) and (2) below: 

 

 

 

 

Scleral Permeation and Retention Studies

Species Selection and Tissue Preparation.

The experimental procedures used in our study adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Porcine eyes were selected as the animal model of choice as previous studies have shown a strong similarity between the scleral thickness, and ultrastructure of porcine and human eyes. ^22,23 Eyes were obtained from an abattoir and used within three hours of slaughter. The eyes were rinsed, and kept in an ice cold mixture of PBS and 1% antibiotic/antimycotic solution during dissection. The sclera was cut into 1 × 2 cm tissue, weighed, and the thickness measured using a digital vernier calliper (PR5638; Toolspot, Tiverton, UK). 

Scleral Permeation and Retention of Drug-Loaded Micelles.

Scleral permeation studies were assessed using a modified Ussing chamber consisting of two hemi-chambers and separated by a 6 mm interchamber aperture. ²⁴ The external surface of the sclera was facing the hemi-chamber that contained 1 mL of the formulation, while the internal surface was facing the hemi-chamber that contained 1 mL of water. The chambers were secured into place by tightening the screws on either end to prevent leakage while maintaining the tissue's integrity, as reported previously. ²⁵ The fluid volume in the hemi-chambers was kept identical throughout the procedure to prevent the build up of hydrostatic pressure. This was followed by placing a 2 × 2 mm micro polytetrafluoroethylene (PTFE)-encased magnetic stirring bar (FisherBrand, Loughborough, UK) into each hemi-chamber, and sealing their tops with pierced insulation tape to prevent evaporative loss and pressure gradient build up across the chambers. The rapamycin concentration was measured using RP-HPLC by diluting the samples with acetonitrile (ACN). 

The sclera was removed from the chamber and examined to ensure that it remained intact at the end of the experiment. It then was cut into the 6 mm “exposed” and “surrounding” tissues, and placed into separate vials. The tissues were rinsed briefly with water to remove any superficial drug residue. This was followed immediately by the addition of ACN and homogenization at 24,000 rpm using a probe homogenizer (Ultra Turrax T25 basic; Ika-Werke) for 10 minutes. The tissues then were left to settle, in the dark, before filtering through a 0.45 μm acetonitrile-resistant filter (Phenex-RCSyringe Filters; Phenomenex, Cheshire, UK). The filtrate was analyzed using RP-HPLC. 

Calculation of Sclera Permeation, Retention, and Diffusion Coefficient.

The drug concentrations that permeated across and had retained in the sclera were calculated from a predetermined calibration graph using linear regression analysis. Units of ng/mL that were retained in the scleral tissue were converted to ng/g based on tissue weight. 

The diffusion coefficient (D)(cm²·s⁻¹) was calculated using Fick's first law: 

 

where R = diffusion flux (nanomoles·s⁻¹), L = scleral thickness (cm), A = surface area of interchamber aperture (cm²), and C = concentration gradient (as the concentration in the receiver chamber typically was less than 1% of that in the donor chamber, hence the concentration in the donor chamber, in nanomoles/cm³, was taken as the concentration gradient). 

The relationship between the diffusion coefficient (D) and permeability coefficient (P) is given by: 

 

where K = partition coefficient (5.77) ²⁶ and L = scleral thickness (cm). 

Experiments were conducted at 21°C. The diffusion coefficient (and permeability coefficient) was adjusted to body temperature (37°C) using Einstein's relationship: 

 

where D ₂₀ and D ₃₇ = diffusion coefficient at 20°C and 37°C (cm²·s⁻¹), T ₂₀ and T ₃₇ = absolute temperature in Kelvin at 20°C and 37°C, and μ₂₀ and μ₃₇ = viscosity of water at 20°C (0.9804 × 10⁻³ Pa·s at 21°C) and 37°C (0.692 × 10⁻³ Pa·s). 

Statistical Analysis

The values were calculated as an average of three separate repeats and presented as mean ± SD. Comparison of two groups was done using paired Student's t-test. A P value less than 0.05 was considered significant. 

Results

Characterization of OChiPEG by FT-IR

The FT-IR spectrum of OChiPEG (Fig. 1B) shows an absorption peak at ∼1740 cm⁻¹, which corresponds to octanoyl groups and three peaks at ∼840, 950, and 2880 cm⁻¹, which represent the mPEG segment of the chemical structure. A broad peak with reduced relative intensity was observed at 3200 to 3500 cm⁻¹, and corresponds to the hydroxyl and amino groups of chitosan structure (Fig. 1A). Thus, chitosan was grafted successfully with the octanoyl and mPEG groups, and is consistent with the report of Huang et al. ¹⁹  

Figure 1. 

 

FT-IR spectra for OChiPEG and separate excipients.

Characterization of Rapamycin-Loaded Micelles

Analysis of Particle Size and Zeta Potential.

The particle size of the empty and rapamycin-loaded micelles was 37.41 and 40.6 nm, respectively (see Table). Lyophilization did not alter the particle size significantly for the drug-loaded micelles (P = 0.13). However, there was a significant difference after lyophilization for the empty micelles (P = 0.0025). The PDI values are suggestive of a mono-disperse solution. This was consistent with Figures 2 and 3, which show a uniform particle size distribution. It also was shown by the clarity of the solutions upon macroscopic investigation of the dispersed samples. The zeta potential for the drug-entrapped micelles was + 6.84 ± 0.29 mV. 

Figure 2. 

 

Histogram of rapamycin-loaded OChiPEG particle size.

Figure 3. 

 

TEM images for drug-loaded OChiPEG micelles. (A) Freshly prepared solution and (B) lyophilized and redispersed preparation.

Table.

 

Average Particle Size and Polydispersity Index of Empty and Drug-Loaded Micelles (n = 3 ± SD)

OChiPEG Micelles	Size (nm) ± SD	PDI ± SD
Empty	37.41 ± 0.09	0.41 ± 0.01
Empty (FD)	45.21 ± 0.71	0.52 ± 0.01
+ rapamycin	40.60 ± 3.49	0.50 ± 0.04
+ rapamycin (FD)	36.44 ± 2.41	0.33 ± 0.06

 Empty, unloaded micelles; FD, freeze dried (lyophilized); PDI, polydisperisty index.

Analysis of Morphology.

Figure 3A shows nano-sized OChiPEG micelles in solution. This also was the case for the re-dispersed lyophilized OChiPEG micelles (Fig. 3B). 

Thermal Properties of Micelles.

Differential Scanning Calorimetry.

Unmodified chitosan had two weight-loss phases. The first (Fig. 4) was a broad endothermic peak at 100°C, attributed to the presence of water in the sample, as reported previously. ^27,28 A second endothermic peak also is expected at 320°C and is a result of monomer dehydration, glycoside bond cleavage, and decomposition of the acetyl and deacetylated units as shown in a previous study. ²⁹ The two endothermic peaks at 52°C and 50°C correspond to the melting temperatures of mPEG and OChiPEG, respectively. ¹⁹ Rapamycin had a bifurcated endothermic peak in the region of 190 to 200°C (Fig. 5). This corresponds to its melting temperature and may indicate the presence of two crystalline polymorphic forms. ³⁰ These two peaks also were observed in the RP-HPLC chromatogram for rapamycin and in previous studies. ³¹ The thermograms of the synthesized, empty, rapamycin-containing, and physical mixture of OChiPEG all had an endothermic peak at 50°C owing to the presence of OChiPEG. The bifurcated endothermic peak for rapamycin was absent in all of the obtained thermograms except for that of the physical mixture, suggesting a possible interaction between the drug and the micelles. 

Figure 4. 

 

Differential scanning calorimetry thermograms for: mPEG (green line), OChiPEG polymer (blue line), and chitosan oligosaccharide ≥5 kDa (red line).

Figure 5. 

 

Differential scanning calorimetry thermograms for: rapamycin (green line), empty OChiPEG micelles (dark blue line), OChiPEG polymer (red line), rapamycin-loaded OChiPEG micelles (pink line), and the physical mixture of OChiPEG and rapamycin (light blue line).

X-Ray Powder Diffraction Studies.

The XRPD diffractogram for the synthesized OChiPEG polymer and micelles showed two sharp intense peaks at 2θ = 19.3 and 23.3 (Fig. 6), corresponding to mPEG and a smaller peak at 2θ = 20, which is attributed to chitosan. Rapamycin showed strong crystalline characteristics through the presence of sharp peaks at 2θ = 4.94, 7.21, 10.2, 12.2, 14.5, 20.4, and 21.8. These results are consistent with the DSC findings in our studies and with previous studies. ^32–35 Upon the addition of rapamycin to the OChiPEG micelles, the characteristic peaks for the drug disappears. This highlights an important decline in the crystallinity of the drug. 

Figure 6. 

 

Crystallograms for the synthesized OChiPEG polymer, prepared micelles, and individual excipients.

Determination of the Critical Micelle Concentration Using Spectrofluorometry.

The OChiPEG micelles were found to have a CMC of 16.6 μM at room temperature. 

Determination of DL and EE.

The OChiPEG micelles had a DL of 16.3 ± 0.06% and an EE of 85.6 ± 0.29%. 

Scleral Permeability and Retention Studies.

The mean retained rapamycin content from the OChiPEG micellar system was 15.3 ± 3.46 μg/g, and the diffusion and permeability coefficients at 37°C were 8.53 × 10⁻⁹ ± 2.17 × 10⁻⁹, and 5.18 × 10⁻⁷ ± 9.70 × 10⁻⁸ cm²·s⁻¹, respectively. 

Discussion

Rapamycin has considerable therapeutic potential for a range of important eye diseases due to its multiple beneficial modes of action, but like many lipophilic drugs, its physicochemical properties render it unsuitable for topical delivery. OChiPEG nanocarriers were synthesized to encapsulate rapamycin. This design was selected to promote scleral drug retention and permeation via the actions of chitosan and the stability-enhancing effects of the PEG component. 

The grafting of the octanoyl and PEG groups to the backbone of chitosan oligosaccharide to produce the OChiPEG graft copolymers was confirmed using FT-IR. Thermal analysis of the synthesized polymer showed that the grafting of these groups to the chitosan backbone led to a significant reduction in their diffractogram peaks. This may indicate that this conjugation suppresses the crystallization of both of these polymers, which also was noted in other studies. ^36,37 This amphiphilic polymer was used to prepare rapamycin-loaded micelles. Thermal analysis of these micelles did not show the peaks relating to the drug, but only those corresponding to the polymer. This highlights an important decline in the crystallinity of the drug, which may be attributed to the formation of intermolecular interactions between the drug and the hydrophobic core of the micelles. 

The critical micelle concentration for the preparation was 16.6 μM at room temperature. This value was similar to that of Kwon et al., ³⁸ but up to 1700-fold higher than that of Lu et al. ³⁹ for rapamycin-containing polymeric micelles. There are several possible explanations for this finding. The positive charge of the chitosan component of the amphiphile may increase the electrostatic repulsion and cause steric hindrance, shifting the CMC to a higher concentration. This was alluded to in recent studies. ^40,41  

Upon macroscopic inspection of the micellar solution, the samples were visibly clear with no signs of aggregates. The particle size of the micelles containing rapamycin was below 50 nm. This size allows for an ideal compromise where the particles are small enough to permeate across the sclera, while being large enough to enhance bioavailability, as well as reduce clearance and immune-mediated attack, the latter effect also is reduced by particle PEGylation. ^42,43 The micellar size obtained in our study was smaller than that demonstrated in recent studies, which had prepared rapamycin-loaded micelles with sizes in the range of 44 to 107 nm. ^39,44  

The OChiPEG micelles were positively charged. This was attributed to the chitosan component of the micelles, and allows for strong electrostatic interaction between the micelles and the negatively charged proteoglycan matrix of the sclera. ⁹ This resulted in a higher scleral retention of the micelles as compared to the diluted rapamycin solution prepared by Cooper et al. ⁴⁵ The drug permeation of this solution was higher than the OChiPEG micelles (8.82 × 10⁻⁶ cm²·s⁻¹ and 9.56 × 10⁻⁶ cm²·s⁻¹ of rapamycin solution as compared to 5.18 × 10⁻⁷ ± 9.70 × 10⁻⁸ cm²·s⁻¹ of micellar solution). This may have been because of permeability enhancing effects of dimethyl sulfoxide (DMSO) and methanol, which were used to solubilize the drug in the preparation. ⁴⁶ However, the use of organic solvents in ocular formulations is undesirable as they can cause irritation and toxicity. ⁴⁷ Furthermore, a solubilized and diluted formulation does not have a defined sustained-release system. The higher scleral retention observed with the OChiPEG micellar system, on the other hand, will prolong the ocular residence time of the drug. This prolonged exposure is expected to enhance transscleral permeation and slow elimination rates, thereby improving drug bioavailability as compared to conventional eye drops. The chitosan portion also is expected to enhance further ocular permeation if assessed in vivo, as studies have shown that the chitosan exhibits ocular penetration-enhancing effects via transcellular or paracellular pathways with no toxic effects. ^48,49 Furthermore, positively charged particles have shown a higher rate of paracellular diffusion compared to negatively charged particles. ⁵⁰ In addition, the OChiPEG micelles showed high drug loading and entrapment efficiencies, thus exposing the sclera to a high concentration of drug per unit volume. 

Conclusions

Rapamycin has multiple therapeutic activities and, hence, the potential to treat many ocular conditions. However, given its poor water solubility and the lack of functional groups to favor ionization in solutions of varying pH, there are a number of limitations to overcome when delivering this drug via a noninvasive topically administered formulation. Micelles were prepared with the primary aim of solubilizing this drug, prolonging its ocular residence time and enhancing its permeation across the ocular surface. These objectives were met by preparing OChiPEG micelles with high drug entrapment efficiency and scleral retention. The high rapamycin retention and permeation observed with these nanocarriers may be attributed to the retention and permeability-enhancing properties of chitosan, as well as the hydrophilic exterior and small size of these nanocarriers. Therefore, OChiPEG micelles have the potential to be used as vehicles for the delivery of water insoluble drugs for the treatment of anterior or posterior segment diseases. 

Acknowledgments

David McCarthy assisted with TEM and John Frost provided technical assistance with the Ussing chambers. 

Footnotes

 Supported by Kings College Hospital NHS Foundation Trust, and by NeoVista, Novartis, and Oraya (TLJ) for unrelated projects. The authors alone are responsible for the content and writing of this paper.

Footnotes

 Disclosure: N. Elsaid, None; T.L. Jackson, None; M. Gunic, None; S. Somavarapu, None