Distribution and kinetics of the Kv1.3-blocking peptide HsTX1[R14A] in experimental rats

The peptide HsTX1[R14A] is a potent and selective blocker of the voltage-gated potassium channel Kv1.3, which is a highly promising target for the treatment of autoimmune diseases and other conditions. In order to assess the biodistribution of this peptide, it was conjugated with NOTA and radiolabelled with copper-64. [64Cu]Cu-NOTA-HsTX1[R14A] was synthesised in high radiochemical purity and yield. The radiotracer was evaluated in vitro and in vivo. The biodistribution and PET studies after intravenous and subcutaneous injections showed similar patterns and kinetics. The hydrophilic peptide was rapidly distributed, showed low accumulation in most of the organs and tissues, and demonstrated high molecular stability in vitro and in vivo. The most prominent accumulation occurred in the epiphyseal plates of trabecular bones. The high stability and bioavailability, low normal-tissue uptake of [64Cu]Cu-NOTA-HsTX1[R14A], and accumulation in regions of up-regulated Kv channels both in vitro and in vivo demonstrate that HsTX1[R14A] represents a valuable lead for conditions treatable by blockade of the voltage-gated potassium channel Kv1.3. The pharmacokinetics shows that both intravenous and subcutaneous applications are viable routes for the delivery of this potent peptide.

Although HsTX1[R14A] adopts a very stable structure that is resistant to proteolysis 14 , an oral route of administration appears unlikely. We have therefore explored buccal [15][16][17] and pulmonary 18 delivery of this peptide, both of which proved to be effective. We are also exploring slow-release formulations (unpublished). An important question that arises in considering the optimal route of administration and frequency of dosing, however, is the lifetime of the peptide in vivo and its tissue distribution. In the case of ShK-186, for example, a 111 In-labelled 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid conjugate was used to assess whole-blood pharmacokinetic parameters as well as peptide absorption, distribution, and excretion. ShK-186 was absorbed slowly from the injection site, resulting in blood concentrations above the Kv1.3 channel-blocking IC 50 value for up to 7 days in monkeys 8 . In delayed-type hypersensitivity, chronic relapsing-remitting experimental autoimmune encephalomyelitis, and pristane-induced arthritis rat models, a single dose of ShK-186 every 2 to 5 days was as effective as daily administration 8 . The slow dissemination of ShK-186 from the injection site and its long residence time on the Kv1.3 channel contribute to its prolonged therapeutic effect in animal models of autoimmune disease. In this study we have used HsTX1[R14A] modified at its N-terminus with a 1,4,7-triazacyclononane-triacetic acid (NOTA) tag for labelling with 64 Cu as an ideal positron emitter 19,20 , enabling positron emission tomography (PET) studies of peptide distribution in rats over a period of days. The results show a long in vivo half-life as a result of slow renal clearance of the peptide. In view of the high potency and selectivity of HsTX1[R14A] for the target channel Kv1.3, and the importance of this channel as a therapeutic target 6,21 , the persistence of this peptide in vivo strengthens the case for its further development as a therapeutic for the treatment of the above-mentioned immune-related diseases.

Results
Peptide characterisation. HsTX1[R14A] was synthesised as described previously 14 but with a NOTA tag coupled to the N-terminus via an aminoethyloxyethyloxyacetyl (AeeA) linker, as shown in Fig. 1. NOTA-HsTX1[R14A] folded rapidly to a single major product, resulting in the typical pattern of a major earlier-eluting peak by RP-HPLC followed by later-eluting misfolded species and side-products ( Supplementary Fig. S1). When tested against the voltage-gated potassium channel Kv1.3 expressed in L929 mouse fibroblast cells, the tagged peptide had an IC 50 of 68 ± 12 pM ( Supplementary Fig. S2), which was close enough to that of HsTX1[R14A] (IC 50 45 ± 3 pM) to confirm that the tagged peptide was an excellent mimic of the parent peptide for the purpose of this study.
Radiolabelling. The peptide HsTX1[R14A] with a NOTA tag was efficiently labelled with 64 64 Cu II complex formed was stable in the presence of 0.1 M aqueous EDTA solution for at least 24 h. These results verify that NOTA is an appropriate bifunctional chelating agent for 64 Cu-labelling of peptides 22,23 , and the corresponding 64   a delay in the 64 Cu activity accumulation in several tissues. After 60 min this effect was significant in the spleen, pancreas, kidneys, heart and the remaining body (Supplementary Table S1). However, due to the fast diffusion from the injection site into the circulation, no significant difference in the whole blood concentration between the two injection types could be detected after 1 h. After 4 h, the spleen and femur represent the only organs where the Scientific RepoRts | 7: 3756 | DOI:10.1038/s41598-017-03998-x 64 Cu activity concentration is significantly different between i.v. and s.c. injection. At later time points (1-2 days after injection), this difference had disappeared (Supplementary Tables S1 and S2).
The analysis of the blood after i.v. injection (blood activity concentration after s.c. injection is too low for analysis) showed that the distribution of [ 64 Cu]Cu-NOTA-HsTX1[R14A] in the different blood compartments ( Fig. 3) did not change during the time of detection (0-4 h). It was located primarily in the blood plasma (82 ± 4% of the whole blood 64 Cu activity) with 75 ± 3% in the plasma water and only 5 ± 3% bound to plasma proteins.
The PET data (Figs 4 and 5) show that the biodistribution of the radiotracer after 5 h was very similar in both animals, independent of the site of injection. The peptide was mainly eliminated renally and in parallel the liver uptake was relatively low. No lymph nodes were detectable in the images. In the late frames of the PET studies (Figs 4f and 5f), but also in the biodistribution (Tables S1 and S2), activity was visible in the skeleton. In the extractive biodistribution experiments of selected organs and tissues, increased activity was detected in the femur. As this localisation of the radiotracer in the bone was unexpected, an additional study was carried out in which the femurs were extracted 4 h after injection, frozen and cut. The sections were studied by autoradiography. In Fig. 6, a typical autoradiogram, section and image of the surface of a femur are presented. The detailed autoradiographic study of the activity distribution showed that [ 64 Cu]Cu-NOTA-HsTX1[R14A] was located primarily in the growth plate, the region of increased metabolism. This bone activity accumulation pattern and kinetics were, however, different from directly bone-seeking agents like 99m Tc-MDP or 18 F − or 68 Ga-bisphosphonates 24,25 .
The kinetic analysis of the dynamic standard uptake ratio (tissue to blood ratios) PET data of the kidneys, liver, and bone marrow (epiphyseal plates) showed linear tissue to blood time activity curves for the standardised uptake ratio SUR ( Supplementary Fig. S5). This is most likely the result of the radiotracer accumulation in these organs and tissues according to a two-compartment model with irreversible binding, like it is typical for [ 18 F] fluoro-deoxy-D-glucose accumulation in many tissues 26,27 . The elimination organs kidney and liver trapped the radiolabelled peptide. The shape of the bone marrow SUR curve could be an indication that cells in the bone marrow bind and retain the [ 64 Cu]Cu-NOTA-HsTX1[R14A] that explains the clearly visible accumulation of the radiotracer in the region of the epiphyseal plate of the bone marrow (Fig. 6).
The time course of the [ 64 Cu]Cu-NOTA-HsTX1[R14A] is shown in Fig. 7a and b. The PET shows that after 5 h the activity concentration in the heart, skeleton, and kidneys was the same for both injection sites.
The remaining activity in the body (Fig. 7c) decreased to 48 ± 3%ID (i.v.) and 45 ± 3%ID (s.c.). The biological half-lives of the total [ 64  Elimination (Fig. 7c,d, Supplementary Table S1) was predominantly renal (kidney and urine), with maximum after 4 h of 65 ± 2% ID, while in the same time only 12 ± 3% ID (i.v.) and 11.4 ± 4%ID (s.c.) were found in the liver and intestine. The maximum of the hepatobiliary elimination was observed at one day after injection with 31 ± 3% ID (i.v.) and 30 ± 3% ID (s.c.), although at this time point the renal elimination was also larger, with 48 ± 3% ID (i.v.) and 51 ± 2% ID (s.c.). . The largest amounts of activity were accumulated in the kidneys and urine. The analysis showed that the radiotracer was metabolised in the kidneys and excreted into the urine but there was no reuptake into the blood. The 64 Cu activity in the kidneys and in the urine decreased to 31% and 15% of the original compound during the period of investigation (4 h), respectively. The chemical forms of the radioactive

Discussion
HsTX1[R14A] is a potent Kv1.3 channel blocker with excellent selectivity over other potassium channels, and is therefore a potential therapeutic for the treatment of a variety of immune-related diseases associated with chronic inflammation, including systemic or organ-specific autoimmune diseases 13 . An important prerequisite for a therapeutic application is knowledge of its pharmacokinetic and pharmacodynamic properties in vivo. ShK-186, for example, was released slowly from the s.c. injection site, resulting in blood concentrations above those required for effective Kv1.3 blockade for up to 7 days in monkeys 8 . As a consequence, single doses of ShK-186 every 2-5 days were as effective as daily administration in delayed-type hypersensitivity, chronic relapsing-remitting experimental autoimmune encephalomyelitis, and pristane-induced arthritis rat models 8 . As the potency of HsTX1[R14A] is similar to that of ShK-186 and the selectivity for Kv1.3 over all other ion channels tested is significantly higher, it was important to assess the biodistribution and elimination of this peptide in animals. We have undertaken these studies in rats as there is a considerable body of evidence indicating that Kv1.3 plays a key role in regulating the membrane potential of T EM cells in this species like in humans 6,28 .
A 64 Cu-labelled conjugate of HsTX1[R14A] was used to investigate the pharmacokinetic profile. This peptide appears to be very stable in vivo, as anticipated from its structure, which is stabilised by four disulfide bridges, and confirmed experimentally in proteolysis assays 14 . Independent of the route of administration, it was rapidly eliminated from the circulation via the kidneys. The primarily location of the peptide in the arterial blood was the plasma, and there were no indications of significant plasma protein binding. The uptake in erythrocytes did not exceed the normal water space distribution in the red blood cells.
The relatively high accumulation in the epiphysis could be interpreted as binding to activated chondrocyte subpopulations 29 in the growth plate, a specialised bone marrow cell fraction. Postnatal growth of the long bones occurs through the stimulation of chondrocyte proliferation at the epiphyseal growth plates. The epiphyseal activity is regulated by the growth hormone and other endocrinal factors including insulin-like growth factor 1 (IGF-1). IGF-1-induced proliferation of cells was inhibited by both potassium channel blockers and inhibitors of PI3-kinase. IGF-1 through PI3-kinase, PDK1 and SGK1 up-regulates Kv channels, an effect required for the proliferative action of the growth factor 30,31 . This could be a mechanism of increased accumulation of the The highly stabile, hydrophilic, radiolabelled peptide [ 64 Cu]Cu-NOTA-HsTX1[R14A] showed a general biodistribution pattern that is typical for small, metabolically inert pharmaceuticals. The blood clearance was relatively slow, caused by filtration in the kidneys glomeruli. However, the peptide was also trapped in the kidney cortex. The hydrophilic character of the molecule was also reflected in its negligible brain uptake, which makes neurological effects improbable. The low protein binding, minor uptake by macrophages of the lungs or liver and insignificant binding to any scavenger receptors in the blood vessel surface are prerequisites for a high bioavailability of the peptide. Moreover, the selectivity and the low normal-tissue uptake of [ 64 Cu] Cu-NOTA-HsTX1[R14A] provide the potential to image regions with increased potassium channel expression.

Conclusions
The high stability and bioavailability, low normal-tissue uptake of [ 64

Synthesis of NOTA-functionalised HsTX1[R14A].
The primary assembly of HsTX1[R14A] was completed on a Protein Technologies Prelude synthesizer using an Fmoc-tBu strategy. All eight Cys residues were side-chain protected with the Trt group. All couplings were mediated with diisopropyl carbodiimide/HOBT. Upon completion of the primary chain, an additional AeeA linker was coupled to the N-terminus and subsequently the final Fmoc group was removed and the NOTA(tBu) was coupled as an HOBT-ester. This coupling was monitored via the Kaiser test and judged to be complete after overnight coupling. The peptide was simultaneously cleaved from the resin and deprotected using a TFA-based acidolytic cleavage cocktail with cationic scavengers. The crude peptide was isolated by precipitation in ice-cold ether following filtration through a fritted glass funnel. The crude peptide was subsequently dissolved in 50% aqueous acetic acid and diluted into H 2 O containing 0.1 mM reduced and oxidised glutathione. The pH was adjusted to 7.8 and the solution was slowly stirred for 18 h. The folded peptide was purified by preparative RP-HPLC using a gradient of 5-25% acetonitrile into H 2 O   Film Co. Ltd., Japan). The image data were recorded and expressed as digital photo-stimulated luminescence (PSL). Analysis was performed with the AIDA 2.11 program (Raytest, Germany). The whole body cryostat sections were scanned for histologic-autoradiography comparison.
PET scans. The procedures are described in detail elsewhere 24,34,35 . Rats were anesthetised using 9% ± 1% desflurane in 30% oxygen and placed on a heat mat. The animals were kept warm under anesthesia until the end of the scan with a total duration of 4 h. The anesthetised animals were localised in a prone position in the axial direction of the scanner. A needle catheter was installed in a lateral tail vein or subcutaneously in the neck for injection using a syringe pump. PET studies were performed with the dedicated small animal PETs NanoPET/CT (Mediso, Budapest, Hungary) and microPET (R) P4 (Siemens Medical Solutions, Erlangen, Germany). Transmission correction was performed with computer tomography attenuation or transmission scans of 10 min using a 57 Co point source that were performed before tracer application. Data were acquired over 300 min. Simultaneous with the start of data acquisition, was the infusion of approximately 20 MBq [ 64 in saline with a duration of 5 min initialised. The PET images were iteratively reconstructed by a 3-dimensional ordered-subset expectation maximisation algorithm (3D OSEM/MAP) with transmission correction and with voxel size of 0.050 × 0.050 × 0.050 cm. No additional corrections were made of partial-volume effects and recovery. Three-dimensional regions of interest (ROI) were determined for subsequent data analysis. The standardised uptake values (SUV, g/mL) and standardised uptake ratios (SUR, as ratio of the SUVs of the tissue of interest and the blood SUV, derived from a region over the caudal arteria abdominalis and vena cava) were used to quantify the activity uptake and kinetics.
Metabolite analysis. The analysis of [ 64 Cu]Cu-NOTA-HsTX1[R14A] degradation was analysed in one rat after single intravenous injection. The gas-anaesthetised (8% desflurane, 30% oxygen, air) rat was prepared on a heating bed for the blood sampling. A needle catheter was inserted in a tail vein and fixed for the injection of the radiotracer and supplementation of the withdrawn volume by E-153 (electrolyte infusion solution). A small incision on the right hind limb was made and a small superficial artery was exposed. A thin tube (polyethylene, 0.3 mm inner and 0.6 mm outer diameter) was inserted, moved to the femoral artery and fixed. Arterial blood samples were collected at 1, 3, 5, 10, 20, 30, 120 min p.i of the [ 64 Cu]Cu-NOTA-HsTX1[R14A]. After the end of experiment the anaesthetised animals were euthanised by infusion of KCl solution. For the differentiation of the activity distribution in the blood compartments (Fig. 3b) were blood samples centrifuged (3 min at 14,000 × g) to obtain the plasma (plasma total and blood cells in the precipitate). One kidney was homogenised with PBS (30% w/v) and then centrifuged as described for the blood samples. Urine was treated in analogy. All the supernatants from blood, kidney and urine samples were mixed with equal volume of MeCN to precipitate the protein fraction. The samples were again centrifuged (3 min at 14,000 × g), resulting in the precipitate the protein bound fraction (plasma proteins) and the supernatants (e.g. plasma water fraction) were analysed by radio-TLC with RP18 aluminium foil, MeCN/0.1% TFA mixed with H 2 O/0.1% TFA (v/v 1:1). The TLC plates were measured by radioluminography as described above. Aliquots of the whole blood samples and of all supernatants were measured for the 64 Cu activity concentration in a well counter.

Statistical analysis.
Statistical analyses were carried out with GraphPad Prism version 6 (GraphPad Software, San Diego California USA, www.graphpad.com). The data are expressed as mean ± SD or SEM when indicated. The unmatched biodistribution data at different time points after injection are expressed as % ID (activity amounts) and SUV (activity concentration) were submitted to one-way analysis of variance (ANOVA) followed by Tukeys post-hoc analysis for all data sets to compare the effect of the i.v. and s.c. injection on the means between the groups. Values of p < 0.05 were considered statistically significant and indicated by an asterisk (*), p < 0.01 (**), p < 0.001 (***), p < 0.0001 (****).