Pharmacokinetic Characterization of Kalata B1 and Related Therapeutics Built on the Cyclotide Scaffold

Accepted Manuscript

Pharmacokinetic characterization of kalata B1 and related therapeutics built on the cyclotide scaffold

Aaron G. Poth, Yen-Hua Huang, Thao T. Le, Meng-Wei Kan, David J. Craik

PII: S0378-5173(19)30354-0 DOI: https://doi.org/10.1016/j.ijpharm.2019.05.001 Reference: IJP 18331

To appear in: International Journal of Pharmaceutics

Received Date: 31 January 2019 Revised Date: 24 April 2019 Accepted Date: 3 May 2019

Please cite this article as: A.G. Poth, Y-H. Huang, T.T. Le, M-W. Kan, D.J. Craik, Pharmacokinetic characterization of kalata B1 and related therapeutics built on the cyclotide scaffold, International Journal of Pharmaceutics (2019), doi: https://doi.org/10.1016/j.ijpharm.2019.05.001

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Pharmacokinetic characterization of kalata B1 and related therapeutics built on the cyclotide scaffold

Aaron G. Potha, Yen-Hua Huanga, Thao T. Lea,b, Meng-Wei Kana, David J. Craika*

aInstitute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland 4072, Australia

Email addresses:

Aaron G. Poth: [email protected] Yen-Hua Huang: [email protected] Thao T. Leb: [email protected] Meng-Wei Kan: [email protected] David J. Craik*: [email protected] bPresent address: School of Science, Edith Cowan University, Perth, WA, 6027, Australia

*Corresponding author’s address:

Professor David J. Craik

Institute for Molecular Bioscience

The University of Queensland

Brisbane, QLD, 4072

Australia

Phone: +61 7 3346 2019

Fax: +61 7 3346 2101

Declarations of interest: none

Abstract:

Oral activity has been described for cyclotide-containing traditional medicines, and demonstrated for reengineered cyclotides bearing grafted therapeutic epitopes, highlighting their potential for translation to the clinic. Here we report preclinical pharmacokinetic parameters for the prototypic cyclotide kalata B1 (kB1) and two orally active grafted analogues, ckb-KAL and ckb-KIN, to provide the first in vivo dose-exposure metrics for cyclotides. Native and grafted kB1 molecules exhibited multiple compartment kinetics and measurable but limited oral bioavailability of similar magnitude to several orally administered peptide drugs in the clinic. Cyclotides are mostly associated with the central compartment, and display small (0.07–0.1 L kg-1 for kB1 and ckb-KIN) to moderate (1 L kg-1 for ckb-

KAL) steady state volumes of distribution. This study provides new data essential to the evaluation of cyclotides as therapeutics, validating them as a viable drug design scaffold with tunable pharmacokinetic properties.

Keywords: Cyclotide; oral bioavailability; drug scaffold; pharmacokinetics; grafting; knottin

Chemical compounds studied in this article: kalata B1 (PubChem CID: 52945815)

1. Introduction

With the enormous chemical space and extended surface areas accessible by their natural sidechain variations

(Khazanov and Carlson, 2013), protein drugs (biologics) can display exquisite selectivity and potency for their targets, typically endowing them with large therapeutic windows compared to traditional small molecule (<500 Da) drugs (Khazanov and Carlson, 2013). However, due to their size (>5000 Da) and physicochemical properties, biologics are often hampered by low membrane permeability, poor oral absorption, and high susceptibility to proteolysis, leading to short biological half-lives.

Cyclotides are a family of plant-derived peptides featuring a cyclic backbone cross-braced by a cystine knot in which a trio of disulfide bonds link through one another to stabilize the peptide core (Craik et al., 1999). The traditional uterotonic medicine ‘kalata-kalata’ from which cyclotides were first identified (Craik et al., 1999) is prepared by boiling the leaves of the African herb Oldenlandia affinis, and so reports of its oral bioactivity in ethnomedicinal settings suggest oral and heat stability of the active ingredients (Gran, 1970; Nworu et al., 2017). Subsequent studies on cyclotides isolated from O. affinis have shown them to be impervious to degradation by enzymes in vitro, high temperatures, or low pH environments (Colgrave and Craik, 2004).

From the time the ultra-stable cyclic cystine knot (CCK) motif of the cyclotide kalata B1 (kB1) from O. affinis was characterized (Colgrave and Craik, 2004; Saether et al., 1995), considerable effort has been invested in evaluating this motif as a scaffold for drug design. In particular, the ‘grafting’ of bioactive peptide epitopes into the stabilized

CCK has been widely explored. The CCK scaffold is especially suitable for this application and has been used to stabilize a range of peptide epitopes with potential applications in the treatment of cancer (Aboye et al., 2016; Chan et al., 2015; Dsouza et al., 2015; Getz et al., 2013; Huang et al., 2015; Quimbar et al., 2013), obesity (Eliasen et al.,

2012), pain (Wong et al., 2012), cardiovascular disease (Getz et al., 2011), multiple sclerosis (Thell et al., 2016) and inflammation (Thongyoo et al., 2009). Acyclic knottins also feature a stable cystine knot motif (Colgrave and Craik,

2004; Werle et al., 2006) and several of them have been investigated for their capacity to increase the stability of contrast agents for tumor imaging (Kimura et al., 2012; Moore et al., 2013) or to modulate therapeutic targets (Krause et al., 2007). 3

Recently, several grafted knottins and cyclotides have been shown to exhibit oral activity and have been evaluated from in vitro concentration-response or in vivo dose-response perspectives. Wong et al. (Wong et al., 2012) reported the oral activity of two engineered cyclotides bearing therapeutic grafts (ckb-KAL and ckb-KIN) targeting bradykinin

B1 receptors in a mouse model of inflammatory pain. These grafted cyclotides inhibited pain pathways in peripheral sensory ganglia, indicating their systemic distribution (Wong et al., 2012). In another study, the cyclotide MCoTI-I was shown to undergo cellular uptake in vitro and interact with intracellular targets, thus opening up new avenues for peptide therapeutics (Contreras et al., 2011). In terms of biodistribution in rats for intravenously administered

MCoTI-II, the intact peptide was primarily detected in the serum and kidneys indicating renal clearance, although no uptake into brain tissues was observed (Wang et al., 2016). Furthermore, fluorescently tagged [T20K] kB1 was observed to associate particularly with the kidneys following oral administration to rats, and detectable amounts of unlabeled [T20K] kB1 were observed in plasma (although not quantitated) (Thell et al., 2016). In combination, these studies indicate promise in linking in vitro dose-effect with in vivo dose-responses, but a critical part of evaluating the clinical translatability of any therapeutic is gaining an understanding of its dose-exposure metrics.

Preclinical pharmacokinetic analyses of investigational therapeutics remain a prerequisite for progression to first-in- human clinical studies, whereby allometric scaling allows reasonable prediction of the primary clinical pharmacokinetic parameters (Caldwell et al., 2004). A key metric for assessing the clinical translatability of cyclotides is oral bioavailability (F%) as drugs with a low F% tend to exhibit higher variability in drug exposure among recipients, leading to higher risks of administering sub-therapeutic or toxic doses, particularly when administering drugs with a narrow therapeutic index (Hellriegel et al., 1996). Notwithstanding the variable predictive correlation between human and animal drug bioavailability (Musther et al., 2014), at present there remains a dearth of information regarding the in vivo pharmacokinetics of cyclotides, which have been only minimally explored quantitatively in animal studies (Colgrave et al., 2005; Melander et al., 2016).

In the original study of grafted cyclotides ckb-KAL and ckb-KIN, the authors suggested that the variability in absolute oral activity between them could be due to differences in their oral bioavailability, potentiated by the presence of two charged amino acids and a proline in the grafted loop- the presence of which might favor membrane

4 interaction (Wong et al., 2012). Thus, the aim of the present work was to investigate the pharmacokinetic parameters of the prototypic cyclotide kB1 alongside grafted cyclotides ckb-KAL and ckb-KIN (Figure 1) to uncover quantitative bases for their differences in activity that could be used to improve the future design of grafted cyclotides. In this study, we compared the plasma protein binding, blood component distribution, distribution and elimination kinetics, and oral bioavailability of cyclotides and evaluate their strengths, weaknesses, addressable knowledge gaps, and overall potential for translation as clinical therapeutics.

2. Materials and methods

2.1 Chemicals and peptides

All chemicals were of analytical grade and sourced from Sigma-Aldrich. The native cyclotide (kB1) was extracted and purified from O. affinis as previously described (Plan et al., 2007). Grafted kalata peptides ckb-KAL and ckb-

KIN were synthesized via solid phase peptide synthesis using established methods (Daly et al., 1999a; Daly et al.,

1999b; Huang et al., 2010; Rosengren et al., 2003). The purities of each of the peptides was >95% via HPLC-UV analysis.

2.2 H chemical shift analysis

The structures of kalata B1, ckb-KIN and ckb-KAL were analyzed using nuclear magnetic resonance (NMR) spectroscopy. Peptides were solubilized in H2O/D2O (9:1, v/v) at a concentration of 1 mM. The one- and two- dimensional spectra (TOCSY and NOESY) were recorded on a Bruker Avance-600 MHz NMR spectrometer and the mixing time for the two-dimensional spectra was 80 and 200 ms at 298 K, respectively. The spectra were referenced to 2,2-dimethyl-2-silapentane-5-sulfonic acid at 0 ppm and assigned using Sparky.

2.3 Quantitative determination of cyclopeptides

Peptides in the in vitro assays and pharmacokinetic plasma samples were quantified via ultra-performance liquid chromatography (UPLC) electrospray ionization (ESI)-tandem mass spectrometry (MS/MS) in multiple reaction monitoring (MRM) mode. Ionization conditions were optimized using purified analyte peptides dissolved in 20%

5 acetonitrile (v/v) with 0.1% formic acid and delivered via a Harvard syringe pump to a TurboV ionspray source- equipped 4000 QTRAP LC-MS/MS instrument (SCIEX; Framingham, Mass., USA). Collision energy and declustering potential voltages were optimized following assessment of the effects of automatic ramping at several ionspray and fragmentation voltages and source temperatures upon signal-to-noise for quantifier and qualifier transitions, summarized in Table 1 in reference (Poth et al., 2019)).

A 10 L sample of the peptide to be quantified was introduced to a Phenomenex Kinetex C18 UPLC column (100 x

2.1 mm, 1.7m, 100 Å) using a UPLC system (Shimadzu) interfaced with a 4000 QTRAP LC-MS/MS instrument.

The flow rate was set at 0.4 mL min-1, and the analytes eluted using a linear acetonitrile gradient. Multiple transitions, including a quantifier ion and two qualifier ions, were monitored for each of the cyclopeptide analytes and an internal standard peptide (cVc1.1). Quantitative analysis of the peptides in each sample was conducted using the Quantitation

Wizard function within Analyst 1.5 (SCIEX) software, normalizing to internal standard peptide signals.

2.4 In vitro plasma protein binding

Plasma protein binding was evaluated for candidate peptides in rat and human plasma using ultracentrifugation as described previously (Srikanth et al., 2013). Frozen rat or human plasma was thawed and spiked with peptide analyte to a concentration of 1 M (~3000 ng mL-1). Triplicate samples of spiked plasma (200 μL) for each peptide were ultracentrifuged at 37°C for 2 h at 436 000 g. To measure free drug concentrations, plasma water samples were collected from respective menisci of centrifuge tubes and deconvoluted via 3:1 (v/v) acetonitrile precipitation and benchtop centrifugation at 17 000 g for 20 min to remove proteins. Initial drug concentration in spiked plasma

(representing free and bound drug) was monitored by precipitation of spiked plasma that had been incubated at 37°C for 2 h. Peptide analyte concentrations were determined via LC-MS or LC-MRM as appropriate. Plasma protein binding (PPB) was calculated using the equation:

x PPB = (1) y where x is the average concentration of analyte quantified in plasma water of spiked plasma, and y is the average concentration of analyte quantified in spiked whole plasma. 6

2.5 Peptide distribution in blood

Peptide distribution was determined in human and rat blood via measurements of blood-to-plasma and erythrocyte- to-plasma ratios. In triplicate pairs, peptide stock solution or an equivalent volume of PBS was spiked into whole human blood to a final concentration of 620−1000 ng mL-1 and incubated at 37°C. At several time points (5, 25, 45,

90 and 135 min), 200 L aliquots of spiked blood were removed from each tube. Following centrifugation at 1500 g for 5 min, the plasma fraction generated from kB1-spiked blood was retained and mixed with the control erythrocytes generated following equivalent centrifugation of PBS-spiked blood. The PBS control plasma was reciprocally combined with the kB1-spiked erythrocytes. For direct comparison, kB1-spiked whole blood was also centrifuged as above, then redistributed following gentle vortexing. All tubes were freeze-thawed three times, bath- sonicated for 5 min, and ice-cold acetonitrile added at a 1:3 ratio (v/v) to facilitate protein precipitation. Samples were centrifuged at 17 000 g for 20 min at 4°C, and supernatants analyzed directly via UPLC-MRM analysis on a

4000 QTRAP LC/MS/MS system.

2.6 Lipophilicity (LogD7.4 shake-flask) assay

Prior to assay, 1-octanol was combined with PBS (pH 7.4) and stirred overnight to achieve equilibrium in reciprocal saturation. In triplicate, PBS-saturated octanol and octanol-saturated PBS were added 1:1 to dried test compounds and sonicated briefly to ensure dissolution. The analytes were allowed to equilibrate between the organic and aqueous phases via a combination of benchtop mixing and intermittent application of microwave energy based on established methods (Song et al., 2014). Once at equilibrium, the sample tubes were centrifuged at 17 000 g for 5 min to allow collection from the separated organic and aqueous layers. Samples were diluted with 1% formic acid and DMSO at a sufficient ratio to completely solve the respective phases, and the relative quantities of analyte determined within the samples via LC-MRM and post-run comparison of analyte peak areas. The LogD7.4 of each peptide was calculated using the following equation:

Coct logD7.4 = log ( ) (ퟐ) Caq

7 where Coct = concentration of drug in the octanolic phase, and Caq = concentration of drug in the aqueous phase.

2.7 Pharmacokinetic experiments

Pharmacokinetic parameters pertaining to stability and bioavailability were determined in male Wistar rats. Rats aged 6−8 weeks (260−330 g body weight) were provided with ad libitum access to solid pellet food and water for an acclimatization period of 7 days following inter-institute transport, and housed individually. Cyclic peptides dissolved in 0.9% PBS were passed through a syringe filter (0.45 micron, Sarstedt) and quantified by UV absorbance prior to administration by oral gavage (per os) or intravenous injection into the lateral tail vein over 10 seconds.

Group sizes were n=3 per route, except that in the orally dosed ckb-KIN group there was n=2 at the 180 and 240 min time points due to file corruption (unrecoverable data). At each time point, ~200 L of blood was collected from the tail vein of each rat in heparinized tubes and kept on ice (0°C) until further bioanalysis. Blood samples were collected

30 min prior to drug administration via either route, and then at 5, 15, 30, 60, 90, 150, 300 and 480 min after IV administration. Blood collections were at 30, 60, 90, 120, 180, 240, 360 and 480 min after oral administration. Blood samples were centrifuged at 1500 g for 10 min to pellet red blood cells (RBCs), and the plasma was used immediately or retained at -20°C until further bioanalysis.

2.8 Bioanalytical sample preparation

Rat plasma was deconvoluted using a solvent-first precipitation method. Briefly, 100 μL plasma samples were added to tubes already containing 300 μL acetonitrile (resultant ratio 1:3 [v/v]) spiked with an internal standard peptide

(Vc1.1), vortex-mixed for 30 s and then left on ice (0°C) for 20 min to induce precipitation of proteins. Precipitated proteins were pelleted by centrifugation at 17 000 g for 20 min, and supernatants were transferred to sample vials and stored at 4°C until further analysis via LC-MRM.

2.9 Pharmacokinetic analysis

Pharmacokinetic parameters for the peptide analytes were determined following deconvolution of natural log- transformed concentration vs. time data, fitting open two-compartment models. Using the method of residuals (curve stripping), the terminal phase (biological) half-life  and the projected concentration at time zero (C0) following 8 bolus intravenous (IV) administration were determined by back-extrapolation of log-transformed concentration vs. time plots during the observed terminal phase. Figure 2A illustrates a block diagram of flows and kinetics for the open two compartment model, and Figure 2B illustrates analysis of an example two compartment pharmacokinetic dataset using the method of residuals. Equations used in calculation of the various pharmacokinetic parameters are detailed in the text and Table 2 in reference (Poth et al., 2019). All statistical and graphical analyses of the concentration vs. time data used in the above calculations with method of residuals data were performed using

GraphPad Prism v6.0.

3. Results

3.1 Ha chemical shift analysis of kB1 and analogs

We confirmed that the overall structure of ckb-KIN and ckb-KAL is the same as their parent peptide kB1 using Hα chemical shift analysis, as detailed in Figure 3. The Hα chemical shifts of the grafted peptides follow the same trend as that of kB1, demonstrating that their overall folding is identical to the parent peptide, and that they have the same disulfide connectivity. As expected, the main differences in chemical shifts were observed in the grafted loop (loop

6) and the adjacent regions.

3.2 Lipophilicity of native and grafted kB1 cyclotides

The lipophilicity of native kB1 and grafted cyclotides was assessed using the shake-flask method. Overall, the experimental lipophilicity of the tested cyclotides was uniformly low, with logD7.4 values of -0.64, -0.55 and -1.08 for kB1, ckb-KIN and ckb-KAL respectively. Interestingly, the low partitioning of cyclotides to the octanol layer is at odds with their strong retention on C18 chromatographic media. LogD7.4 values less than 0 typically correlate with lower plasma protein binding, ineffective transfer across the lipid core of membranes resulting in lower cell permeability, and lower clearance in liver microsomes (Rand et al., 2012). In this context, the tested cyclotides occupy positions on a logK / logD7.4 plot consistent with (anionic) acidic drugs, despite their disparate calculated pIs

(Table 1). The similarity in their experimental logD7.4 values might therefore be significantly influenced by their common CCK motif.

Table 1. Physicochemical parameters for natural and grafted cyclotides.

a b c d Peptide MW (M0 Da) % aromatic Net charge Calculated pI % DEKR % AGS kalata B1 2890.15 3.4 0 5.93 6.90 20.69 ckb-KIN 2928.18 3.3 -1 3.85 3.33 23.33 ckb-KAL 3132.28 6.5 +1 7.73 9.68 22.58

a Calculated monoisotopic uncharged mass M0. b Calculation inclusive of disulfide bonds and head-to-tail cyclization.

c Percentage of amino acids comprising aspartic acid, glutamic acid, lysine and arginine. d Percentage of amino acids comprising alanine, glycine and serine.

3.3 Blood distribution and plasma protein binding of cyclotides

Knowledge of the blood distribution properties of potential therapeutics is crucial for determining free drug exposure in the plasma and by extension at their pharmacological targets (Hinderling, 1997; Smith et al., 2010; Yu et al.,

2005). To validate the rat pharmacokinetic model for evaluating the performance of cyclotide-based therapeutics in humans, we measured the plasma protein binding and blood distribution of kB1 in human and rat plasma following established comparative methods (Damre et al., 2011; Hinderling, 1997; Yu et al., 2005). Blood-to-plasma (Kb/p), blood-to-erythrocyte (Kb/e), and erythrocyte-to-plasma (Ke/p) distribution ratios for kB1 were similar between rat and human vascular samples (Figure 4). The Kb/p ratio was 1.05 ± 0.28 (n=3) for kB1 in human vascular fluid, and

1.46 ± 0.22 in rat vascular fluid, while the Kb/e ratios indicated that approximately 22% of the peptide was associated 10 with human erythrocytes, and a similar proportion associated with rat erythrocytes. Although the calculated plasma clearance can overestimate blood clearance for compounds with Kb/p significantly greater than 1, due to the modest

Kb/p values for kB1 and the increased complexity of the whole blood matrix, quantitative bioanalyses in support of pharmacokinetic modeling were hereafter conducted using plasma.

Plasma protein binding of the cyclotides was measured to estimate the free drug concentration present within pharmacokinetic rat plasma samples, which is representative of the peptide available to permeate and interact with targets in the therapeutic biophase (Smith et al., 2010). Rat plasma was a reasonable model of human plasma in terms of the plasma protein binding; kB1 was highly bound in both human (4% free) and rat (8% free) plasma, whereas ckb-KAL and ckb-KIN displayed only moderate binding to plasma protein (Figure 5).

3.4 Pharmacokinetic profiles of native and grafted kB1 cyclotides

Pharmacokinetic parameters were determined for natural and grafted cyclotides in juvenile male Wistar rats as detailed in Table 3 in reference (Poth et al., 2019). The concentration vs. time plots for intravenously administered cyclotides best fit open two-compartment models (Figure 6), and the primary pharmacokinetic parameters for each peptide were calculated using the method of residuals (curve stripping).

All the tested cyclotides demonstrated biological (terminal phase) half-lives in the range 1.25−3.2 hours, with kB1 having a biological (terminal phase) half-life of 192 minutes, longer than that recently reported (Melander et al.,

2016). The calculated volumes of distribution for the central compartment (Vc) spanned a least an order of magnitude across the native and grafted kB1 cyclotides (kB1, 42.4 mL kg-1; ckb-KIN, 114.22 mL kg-1; ckb-KAL, 1034.45 mL

-1 -1 kg ), and likewise for the volume of distribution at steady state (Vdss) (kB1, 71.14 mL kg ; ckb-KIN, 136.00 mL

-1 -1 -1 kg ; ckb-KAL, 1053.60 mL kg ). Whereas both kB1 and ckb-KIN exhibited low Vdss (~0.07 to 0.1 L kg ) consistent with a distribution confined within the total blood volume, ckb-KAL had a moderate distribution of ~1 L kg-1, consistent with affinity for tissues beyond the plasma volume and potentially encompassing interstitial fluids. Despite the known strong affinity of kB1 for phosphatidylethanolamine (PE)-containing membranes such as those

11 surrounding RBCs (PE represents 18% of the lipids present in the erythrocyte plasma membrane) (Henriques et al.,

-1 2012), the Vdss of kB1 was very low (0.07 L kg ) and approximated the total blood volume.

The ratios of the determined volumes of distribution- Vc, Vdss, Vd and Vdextrap for each cyclotide are given in Table

4 in reference (Poth et al., 2019). All cyclotides exhibited a low Vc/Vdextrap coefficient, consistent with the polyexponential decay and pronounced distribution phases observed following their IV administration (Figure 6).

Ckb-KIN exhibited the largest Vdextrap/Vd coefficient at 18.46, which represents the fractional error in total clearance one would encounter by incorrectly assuming a monoexponential fit versus the chosen polyexponential fit to describe its disposition. The values of (Vdss/Vc)-1 for kB1 and ckb-KIN ranged from 0.68 to 0.19, and very low (0.02) for ckb-KAL, indicating that the bulk of the administered dose for the grafted peptides remained within the central compartment. In the case of ckb-KAL, this suggested <1% of the dose resided within the peripheral compartments at steady state.

3.5 Oral bioavailability and oral exposure

Grafted cyclotides ckb-KAL and ckb-KIN are bradykinin B1 receptor antagonists, and have been reported to elicit oral activity in a mouse model of inflammatory pain (Wong et al., 2012), with nociception observed 30 to 45 minutes after intraperitoneal injection or oral dosage (Wong et al., 2012). Consistent with this time response, we found that peak plasma concentrations of grafted cyclotides (Cmax of 44 nM and 18 nM for ckb-KAL and ckb-KIN, respectively)

-1 were reached 30-90 minutes (Tmax) after oral gavage in rats (at 8 and 15 mg kg for ckb-KAL and ckb-KIN, respectively). Taking plasma protein binding into account, peak free drug concentrations in rat plasma are approximately 3 nM and 2 nM at 30 minutes post dose for ckb-KAL and ckb-KIN, similar to the previously reported in vitro test levels (~10 nM) required for complete block of bradykinin B1 receptors (Wong et al., 2012). We found that ckb-KAL exhibited increased oral bioavailability over ckb-KIN (0.43% vs. 0.07%) but with a shorter terminal phase half-life (~78 min vs. ~147 min).

4. Discussion

The lipophilicity of a compound is a key physicochemical attribute that influences its passive membrane permeation, solubility, plasma protein binding and ultimately, its distribution and elimination. A convenient relative measure of lipophilicity is calculated with a logD7.4 measurement, which quantifies the relative solubilities of a compound

(partition coefficient) between reciprocally saturated aqueous phase at pH 7.4 (mimicking blood) and 1-octanol, imitating the lipidic core of phospholipid bilayers. A previous analysis of cyclic peptides with LogP in the range 0 to -1 found them to uniformly exhibit oral bioavailabilities of 10% or less, suggesting that lipophilicity is an important contributor to oral bioavailability for these molecules (Nielsen et al., 2017). This is consistent with predictions of the requisite property space for passive permeability (Pye et al., 2017) that suggest increases in lipophilicity are required with increasing molecular size, putting an operational ceiling on the size of membrane permeable molecules at approximately 1500 Å3 (1200 Da). Cyclotides, at 2800 to 3900 Da for the naturally occurring sequences so far characterized, clearly violate this rule and so by this logic are unlikely to possess passive permeability or be orally bioavailable. Notwithstanding this, it is postulated that for peptide drugs, membrane permeation might be more reliably predicted from their hydrogen bonding potentials rather than lipid solubility (Burton et al., 1992).

A high value for a compound’s relative association with erythrocytes over plasma (Ke/p) often correlates with hemotoxicity, and thus the slightly (albeit not significantly) higher ratios observed in rat suggest that kB1 might be more toxic to rat RBCs than to human. However, overall, the similar distribution coefficients between rat and human blood components suggest that the rat is an appropriate model organism for cyclotide pharmacokinetic studies.

One caveat of the ultracentrifugation method for determining plasma protein binding is that the pH typically increases from initial physiological levels (pH 7.4) in the absence of CO2 buffering, which potentially causes higher binding to plasma proteins (Kochansky et al., 2008). Indeed, the fraction of kB1 unbound in rat plasma determined in previously reported equilibrium dialysis was higher (25%) (Melander et al., 2016), suggesting the plasma protein binding values determined in our experiments might be slightly overestimated. However, whether the equilibrium dialysis method used in the previous study (Melander et al., 2016) was set up to minimize pH change through incubation in a CO2-rich atmosphere was not reported. In addition, we observed extensive binding of the investigated 13 peptides to equilibrium dialysis membranes, which limits assay reliability and prompted our use of the ultracentrifugation method, which is particularly suited to analyses of compounds with high nonspecific binding as it eliminates problems arising from free membrane effects (Damre et al., 2011). The extent of binding to equilibrium dialysis membranes in the previous study (Melander et al., 2016) was not reported, but probably also contributes to the observed difference in plasma protein binding for kB1 in our study.

There are a plethora of approaches that can be used to model pharmacokinetic data with the aim of extrapolating and predicting the ADME performance of a drug from one species to another, ranging from complex whole-body 21- compartment physiologically based pharmacokinetics (PBPK) models (Jain et al., 1981) to simplified all-purpose

PBPK models (Cascone et al., 2018; Di Muria et al., 2010) through to 1, 2 or 3 compartmental models, the selection of which is made in balance with model complexity, ethical and practical considerations (Stass, 2006) and data quality. The open 2-compartmental model was selected for the current analysis on the basis of its simplicity and best fit with the pharmacokinetic data (see Table 4 in (Poth et al., 2019)).

In vitro plasma stability assays previously indicated extensive and equivalent resistance to degradation for ckb-KAL, ckb-KIN and a backbone linearized variant kb-KIN (Wong et al., 2012), but it is not yet well-established whether a cyclic backbone is an important or necessary feature for preserving in vivo stability. While the in vivo terminal half- lives of the tested candidates varied almost 3-fold (78/147/192 minutes), there was a substantial (~15-fold) variation in the steady state volume of distribution, ranging from 0.071 to 1.05 L kg-1 (Table 3 in reference (Poth et al., 2019)), indicating a significant range in clearance for this compound class. A limitation of our study arises from the sparse sampling during the later periods of the in vivo pharmacokinetic experiments, which could reduce the accuracy of the determination of terminal half-lives. The biological (terminal phase) half-life determined for kB1 was longer than that recently reported (Melander et al., 2016), and this is most likely due to sampling at later time points after administration and better capturing the terminal phase. Notwithstanding these sources of uncertainty in half-life determination, when scaled via allometry (Caldwell et al., 2004), the half-life values for the cyclotides studied here correspond to half-lives of approximately 5–12 hours in humans, similar to many marketed peptide drugs and venom peptides (Sanhajariya et al., 2018) and detailed in Tables 5 & 6 in reference (Poth et al., 2019). These half-life values

14 are also similar to those estimated for knottins designed for applications in diagnostic imaging (0.4 h (Miao et al.,

2009) and 2.6 h (Moore et al., 2013) in rats). On the surface, these results suggest that backbone cyclization may provide only an incremental advantage in terms of prolonging in vivo stability, but a limitation of the knottin studies is that quantification was based upon either a fluorescence label or a radioactivity measurement, neither of which was shown to directly correlate with the abundance of the intact molecule, and might instead reflect all associated metabolites that retain the label.

Volumes of distribution in humans for small molecule drugs typically range from 0.1 to 500 L kg-1, with molecules exhibiting a large Vd considered to have increased affinity for tissue components away from the plasma in the central compartment. Despite their low logD7.4 values some cyclotides exhibit high affinities for lipid membranes, and display characteristic late elution in RP-HPLC separations (Gruber et al., 2008). Among the endogenous phospholipids, native cyclotides including kB1 are known only to have significant interactions with the zwitterionic headgroup of PE (Henriques et al., 2012). Extensive nonspecific affinity for membranes and plasma proteins can however result in reduced therapeutic windows and toxicity (such as phospholipidosis) where high total plasma drug concentrations are required in order to maintain efficacious unbound plasma drug concentration at the therapeutic biophase (Valko et al., 2017). The low Vc of kB1 however, is congruent with its high affinity for plasma proteins and for PE-containing membranes (such as the erythrocyte outer membrane), both of which confine the majority of kB1 to within the total blood volume. The demonstration that single amino acid substitutions (e.g. [E7D] kB1) can modulate PE membrane affinity (Henriques et al., 2011) suggests careful modulation of membrane binding will be may be a particularly important factor in cyclotide-based drug design.

When comparing cyclotide clearance with rat hepatic and renal blood flow (Boxenbaum, 1980), all the examined cyclotides could be characterized as low extraction (restrictively cleared) drugs with E <0.20. Thus, their clearance might be expected to be largely independent of changes in the clearance organ’s blood flow and more sensitive to changes in hepatic and renal function (intrinsic clearance). A calculation of the maximum drug efficiencies (Valko et al., 2017) for the cyclotides analysed in this study results in values >20%, similar to marketed peptide drugs (Table

5 in reference (Poth et al., 2019)), indicating the relatively high proportion of dosed compound available unbound at the therapeutic biophase for this compound class, and low likelihood of off-target effects and toxicity.

While degradation via hepatic elimination processes in known for several peptide drugs and hormones including insulin, glucagon and EGF, the primary route for elimination of peptide and protein therapeutics so far appears primarily to occur in the kidneys via a range of complex processes including glomerular filtration, peritubular extraction and hydrolysis by brush border enzymes. Nonspecific proteolysis is also a major route for their degradation, as proteolytic enzymes are ubiquitous throughout the body and not restricted to the kidney, liver and gastrointestinal tissues (Tang et al., 2004). Indeed, the routes of elimination for fluorescently tagged [T20K]kB1

(Thell et al., 2016) as well as MCoTI-II (Wang et al., 2016) have been reported to be mostly via the kidney, but it remains unresolved whether cyclotides are mainly digested or excreted intact, as cyclotide metabolites have yet to be observed in vitro, in vivo or in planta.

Drugs with high affinities for PS tend to be basic and are known to have relative tissue-to-plasma ratios that closely mirror the PS content in those tissues, with the highest levels of PS found in lung, spleen and kidney (Murakami and

Yumoto, 2011), whereas those with minimal interaction exhibit an even distribution across tissues at steady state.

Previous studies addressing the spatial disposition of labeled cyclotides MCoTI-II (Wang et al., 2016) and

[T20K]kB1 demonstrated their high affinity for the kidney; however, their distributions were measured soon after administration and were thus unlikely to have been measured at steady state, assuming a multiple compartmental kinetic model is valid for those molecules. Excluding the relatively high concentrations observed in kidney and serum, probably influenced by the timing of measurements prior to equilibrium, an even distribution of MCoTI-II was observed in IV-dosed rats across the other major organs including lung (highest tissue PS levels) and heart

(lowest perfused-tissue PS levels), consistent with studies confirming its inability to cross cell membranes by direct permeation (Greenwood et al., 2007). Alternatively, with its increased basic character, [T20K]kB1 is capable of direct membrane translocation in model membranes (Henriques et al., 2015) and is orally available (Thell et al.,

2016). However, it is not clear from the published data whether the tissue distribution of [T20K]kB1 at steady state reflects that of tissue PS levels.

The bioavailability of a drug has been shown to have an inverse correlation with variability in drug exposure in large patient cohorts (Hellriegel et al., 1996), and is thus a crucial parameter in assessing the safety of drugs with a narrow therapeutic index. Even for endogenous peptides such as insulin, large inter-individual variations in clearance and exposure are observed. Therefore, it is important to evaluate the oral bioavailability and exposure of any potential therapeutic peptide. Marketed peptide drugs commonly exhibit a combination of poor permeability through gut mucosa and poor stability in the gastrointestinal (GI) tract, resulting in low bioavailability as summarized in Table 5 in reference (Poth et al., 2019). However, several studies have demonstrated the utility of knottins and cyclotides in crossing selected membranes in ex vivo models. For instance, Werle et al. (Werle et al., 2007 ; Werle et al., 2006) demonstrated that fluorophore-conjugated knottins could cross isolated rat gut membranes with significant apparent

-6 -1 -6 -1 permeability (Papp) of up to 7.4 x 10 cm s compared with FITC-bacitracin control (4.6 x 10 cm s ). In addition, recent studies described active transport of a cystine-knotted peptide from tomato across polarized intestinal epithelium in Caco-2 assays (Treggiari et al., 2017), and another linear cystine knot peptide, EETI-II, was efficiently transported into mammalian cells via active endocytic pathways (Gao et al., 2016). Oral activity was also reported for the linear knottin OAIP-1 in the cotton bollworm Helicoverpa armigera (Hardy et al., 2013), and several cyclotides from the plant Clitoria ternatea provide further examples of gut permeable cystine-knotted peptides (Poth et al., 2011). For kB1, its potential to modulate intracellular targets through internalization pathways by targeting specific phospholipids was also recently demonstrated (Henriques et al., 2015).

In the original study of ckb-KAL and ckb-KIN (Wong et al., 2012), significant pharmacodynamic differences between ckb-KAL and ckb-KIN were reported, with 42% and 28% reductions in writhing response respectively, following 10 mg kg-1 oral dosing. This reported difference in efficacy between orally dosed ckb-KAL and ckb-KIN is probably influenced by differences in their oral bioavailabilities and thus free drug exposure. In the current study, ckb-KAL showed increased oral bioavailability over ckb-KIN (0.43% vs. 0.07%), which is partially offset by its shorter terminal phase half-life (78 min vs. 147 min) and differences in plasma protein binding (35% and 21% for ckb-KAL and ckb-KIN, respectively). However, when the half-life values are allometrically scaled for humans (5–

9 hours), they are equal to or longer than many currently prescribed small molecule analgesics. Taken together, our

17 results show that despite having low oral bioavailabilities, grafted cyclotide exposure in orally dosed rats is consistent with levels able block bradykinin B1 receptors in vitro, and also consistent with the levels demonstrated to potentiate oral analgesic activity in vivo.

In some of the earliest literature on native kB1, in vivo uterotonic activity was observed in rats and rabbits after intraperitoneal administration of 6 M kB1 (dose rate 0.24 mg kg-1) (Gran, 1973). Administration of kB1 to isolated uterine strips was also found to elicit contractions at the same concentration (Gran, 1973). In the same study, the majority of animals receiving intravenous kB1 exhibited no uterine contractile activity at doses up to 0.4 mg kg-1, and no activity was observed in animals receiving intragastric administration (40 mg kg-1) of either purified kB1 or dose-equivalent crude extracts of O. affinis, demonstrating the poor permeability of kB1 and any other potential

‘orally active’ uterotonic component through gut mucosal membranes (Gran, 1973). Indeed, in the initial in vivo study, administration of kB1 at levels equivalent to that found in human-sized doses of the traditional medicine O. affinis extract were unable to elicit any measurable uterotonic effect in rabbits (Gran, 1973), which suggests a potential alternative route of administration for the traditional medicine (Gran et al., 2000). Interestingly, the Cmax of kB1 in the plasma of orally dosed rats (15 mg kg-1) is ~60 nM, which is orders of magnitude lower than the dose required to induce uterine contractions in rats following direct intraperitoneal injection. From our determination of the unbound fraction of kB1 in plasma, and also from the literature (Melander et al., 2016), this would correspond to free drug concentrations of 14–44 ng mL-1 (~5–15 nM). Notwithstanding this, kalata B7, also from O. affinis, is reported as having double digit micromolar potency at oxytocin and vasopressin receptors, and possibly contributes to the utero-stimulant effects of O. affinis traditional medicines (Koehbach et al., 2013).

It will be important in future studies to determine the unbound fraction of cyclotides at their site of action in vivo and to link these values with those determined in plasma to enable refinement of physiologically-based pharmacokinetic models, where free drug concentrations in plasma are in equilibrium with unbound drug at the therapeutic biophase harboring the drug target.

5. Conclusions

Here we determined the pharmacokinetic parameters for native kB1 and two grafted cyclotides (bradykinin antagonists ckb-KAL and ckb-KIN) in orally and intravenously dosed rats, and provide comparative pharmacokinetic parameters for natural and grafted cyclotides. The study shows that along with their exceptional in vitro plasma stability, the in vivo biological half-lives of the native and grafted cyclotides are comparable to peptide drugs in clinical use and consistent with findings for disulfide-rich venom peptides. The tested cyclotides exhibited high drug efficiency comparable with marketed peptide drugs, suggestive of a low likelihood of off-target effects and toxicity, and therefore good prospects for successful progression towards therapeutics where nanomolar potency can be maintained.

Preclinical pharmacokinetic evaluation is a crucial step in the progression of therapeutic candidates for use in a clinical setting, and our findings suggest that these unique molecules possess a pedigree for drug design applications that warrants further investigation, particularly in terms of their unlabeled in vivo biodistribution, affinity for biomembranes, and oral bioavailability.

Figure captions:

Figure 1. Grafting of bioactive epitopes into loop 6 of kalata B1. Loop 6 resides within the dashed line; the amino acids circled in blue are substituted in place of the original loop 6 sequence for ckb-KAL, and those circled in red replace the normal loop 6 sequence in ckb-KIN. Yellow bars denote the disulfide bond connectivity.

Figure 2. The open two compartment pharmacokinetic model. The block diagram in panel A outlines each system flow, and panel B illustrates the method of residuals for determination of pharmacokinetic parameters.

Figure 3. Alpha hydrogen chemical shift comparison between native kalata B1 and grafted peptides ckb-KAL and ckb-KIN. The majority of structural perturbation corresponded with amino acids in loop 6 of the aligned sequences

(kalata B1 = black text, line and circles; ckb-KIN = pink text, line and squares; ckb-KAL blue text, line and triangles).

Sequences for grafted peptides ckb-KAL and ckb-KIN are identical to kB1 apart from loop 6 as indicated. Disulfide bond connectivities are indicated beneath the region of shared homologous sequence, with the Cys1-4 bond indicated with a dot-dashed line, the Cys2-5 bond indicated with a dashed line, and the Cys3-6 bond indicated with a dotted line. Peptide backbone cyclization is indicated with a solid line.

Figure 4. Distribution of kalata B1 (kB1) in human and rat blood components. Concentrations of kB1 were determined in blood components after 135 minutes of incubation. Ke/p denotes the erythrocyte-to-plasma partition coefficient, Kb/p denotes the whole blood-to-plasma partition coefficient, and Kb/e denotes the whole blood-to- erythrocyte partition coefficient; mean and standard deviation (in parentheses) appear above each column. The low

Ke/p and high Kb/e values indicate modest association of kB1 with erythrocytes, and the near equivalence of blood and plasma values (Kb/p near 1) indicates that plasma retains the majority of the analyte observed in whole blood.

Figure 5. Plasma protein binding of native kB1 and grafted mutants in human and rat plasma. Relative peptide concentrations were determined in plasma water vs. whole plasma via UPLC-MS following 4 hours of ultracentrifugation at 37ºC. The mean percentage of unbound drug in plasma is indicated above each bar, and the error bars represent standard deviation.

Figure 6. Plasma concentration-time profiles of natural and cyclotides in male Wistar rats. The concentrations of drug in plasma following intravenous administration (IV) are indicated with blue squares, and red circles indicate per os (PO) administration in panels A, C and E for kalata B1, ckb-KIN and ckb-KAL, respectively. Data are plotted as means with standard error. Natural log-transformed concentration-time profiles in panels B, D and F illustrate cyclotide elimination as fitted to a two-compartment pharmacokinetic model. Group sizes were n=3 per route, except in the orally dosed ckb-KIN group there was n=2 at the 180 and 240 min time points due to file corruption

(unrecoverable data).

Acknowledgements: We thank Alun Jones for expert mass spectrometry advice and access to the Institute for

Molecular Bioscience Molecular and Cellular Proteomics Facility. We thank Philip Sunderland, Phillip Walsh, and

Olivier Cheneval for synthesizing peptides used in the study, and UQBR staff for assistance with animal studies. We thank Julia Bates, PhD, for editorial assistance.

Funding: These studies were supported by a grant from the Australian Research Council (DP150100443) and the

Clive and Vera Ramaciotti Foundation. D. Craik is an ARC Laureate Fellow (FL150100146).

Competing financial interests: The authors declare no competing financial interests, and that the funding sources had no involvement in the study design; collection, analysis and interpretation of data; writing of the manuscript; or decision to submit for publication.

Ethics statement: This work was carried out in accordance with the Australian code of practice for the care and use of animals for scientific purposes (National Health and Medical Research Council, 7th edition 2004), after approval by the University of Queensland Animal Ethics committee.

References:

Aboye, T., Meeks, C., Majumder, S., Shekhtman, A., Rodgers, K., Camarero, J., 2016. Design of a MCoTI-based cyclotide with angiotensin (1-7)-like activity. Molecules 21, 152. Boxenbaum, H., 1980. Interspecies variation in liver weight, hepatic blood flow, and antipyrine intrinsic clearance: Extrapolation of data to benzodiazepines and phenytoin. J. Pharmacokinet. Biopharm. 8, 165-176. Burton, P.S., Conradi, R.A., Hilgers, A.R., Ho, N.F.H., Maggiora, L.L., 1992. The relationship between peptide structure and transport across epithelial cell monolayers. J. Control. Release 19, 87-97. Caldwell, G.W., Masucci, J.A., Yan, Z., Hageman, W., 2004. Allometric scaling of pharmacokinetic parameters in drug discovery: Can human CL, Vss and t1/2 be predicted from in-vivo rat data? Eur. J. Drug Metab. Pharmacokinet. 29, 133-143. Cascone, S., Lamberti, G., Piazza, O., Abbiati, R.A., Manca, D., 2018. A physiologically-based model to predict individual pharmacokinetics and pharmacodynamics of remifentanil. Eur. J. Pharm. Sci. 111, 20-28. Chan, Lai Y., Craik, David J., Daly, Norelle L., 2015. Cyclic thrombospondin-1 mimetics: grafting of a thrombospondin sequence into circular disulfide-rich frameworks to inhibit endothelial cell migration. Bioscience Rep. 35, e00270. Colgrave, M.L., Craik, D.J., 2004. Thermal, chemical, and enzymatic stability of the cyclotide kalata B1: The importance of the cyclic cystine knot. Biochemistry 43, 5965-5975. Colgrave, M.L., Jones, A., Craik, D.J., 2005. Peptide quantification by matrix-assisted laser desorption ionisation time-of-flight mass spectrometry: Investigations of the cyclotide kalata B1 in biological fluids. J. Chromatogr. 1091, 187-193. Contreras, J., Elnagar, A.Y.O., Hamm-Alvarez, S.F., Camarero, J.A., 2011. Cellular uptake of cyclotide MCoTI-I follows multiple endocytic pathways. J. Control. Release 155, 134-143. Craik, D.J., Daly, N.L., Bond, T., Waine, C., 1999. Plant cyclotides: A unique family of cyclic and knotted proteins that defines the cyclic cystine knot structural motif. J. Mol. Biol. 294, 1327-1336. Daly, N.L., Koltay, A., Gustafson, K.R., Boyd, M.R., Casas-Finet, J.R., Craik, D.J., 1999a. Solution structure by NMR of circulin A: a macrocyclic knotted peptide having anti-HIV activity. J. Mol. Biol. 285, 333-345. Daly, N.L., Love, S., Alewood, P.F., Craik, D.J., 1999b. Chemical synthesis and folding pathways of large cyclic polypeptides: Studies of the cystine knot polypeptide kalata B1. Biochemistry 38, 10606-10614. Damre, A.A., Iyer, K.R., Lyubimov, A.V., 2011. The significance and determination of plasma protein binding, in: Lyubimov, A.V. (Ed.), Encyclopedia of Drug Metabolism and Interactions. John Wiley & Sons, Inc., pp. 1-18. Di Muria, M., Lamberti, G., Titomanlio, G., 2010. Physiologically Based Pharmacokinetics: A Simple, All Purpose Model. Industrial & Engineering Chemistry Research 49, 2969-2978. Dsouza, C., Henriques, S.T., Wang, C.K., Cheneval, O., Chan, L.Y., Bokil, N.J., Sweet, M.J., Craik, D.J., 2015. Using the MCoTI-II cyclotide scaffold to design a stable cyclic peptide antagonist of SET, a protein overexpressed in human cancer. Biochemistry 55, 396-405. Eliasen, R., Daly, N.L., Wulff, B.S., Andresen, T.L., Conde-Frieboes, K.W., Craik, D.J., 2012. Design, synthesis, structural and functional characterization of novel melanocortin agonists based on the cyclotide kalata B1. J. Biol. Chem 287, 40493-40501. Gao, X., Stanger, K., Kaluarachchi, H., Maurer, T., Ciepla, P., Chalouni, C., Franke, Y., Hannoush, R.N., 2016. Cellular uptake of a cystine-knot peptide and modulation of its intracellular trafficking. Sci. Rep. 6, 35179. Getz, J.A., Cheneval, O., Craik, D.J., Daugherty, P.S., 2013. Design of a cyclotide antagonist of neuropilin-1 and - 2 that potently inhibits endothelial cell migration. ACS Chem. Biol. 8, 1147-1154. Getz, J.A., Rice, J.J., Daugherty, P.S., 2011. Protease-resistant peptide ligands from a knottin scaffold library. ACS Chem. Biol. 6, 837-844. Gran, L., 1970. An oxytocic principle found in Oldenlandia affinis DC. An indigenous, Congolese drug “Kalata- Kalata” used to accelerate the delivery. Medd. Norsk. Farm. Selskap 32, 173-180. Gran, L., 1973. On the effect of a polypeptide isolated from "Kalata-Kalata" (Oldenlandia affinis DC) on the oestrogen dominated uterus. Acta Pharmacol. et Toxicol. 33, 400-408.

Gran, L., Sandberg, F., Sletten, K., 2000. Oldenlandia affinis (R&S) DC: A plant containing uteroactive peptides used in African traditional medicine. J. Ethnopharmacol. 70, 197-203. Greenwood, K.P., Daly, N.L., Brown, D.L., Stow, J.L., Craik, D.J., 2007. The cyclic cystine knot miniprotein MCoTI-II is internalized into cells by macropinocytosis. Int. J. Biochem. Cell Biol. 39, 2252-2264. Gruber, C.W., Elliot, A., G., Ireland, D.C., Delprete, P.G., Dessein, S., Göransson, U., Trabi, M., Wang, C.K., Kinghorn, A.B., Robbrecht, E., Craik, D.J., 2008. Distribution and evolution of circular miniproteins in flowering plants. Plant Cell 20, 2471-2483. Hardy, M.C., Daly, N.L., Mobli, M., Morales, R.A.V., King, G.F., 2013. Isolation of an orally active insecticidal toxin from the venom of an Australian tarantula. PLoS One 8, e73136. Hellriegel, E.T., Bjornsson, T.D., Hauck, W.W., 1996. Interpatient variability in bioavailability is related to the extent of absorption: Implications for bioavailability and bioequivalence studies. Clin. Pharmacol. Ther. 60, 601- 607. Henriques, S.T., Huang, Y.-H., Castanho, M.A.R.B., Bagatolli, L.A., Sonza, S., Tachedjian, G., Daly, N.L., Craik, D.J., 2012. Phosphatidylethanolamine binding is a conserved feature of cyclotide-membrane interactions. J. Biol. Chem 287, 33629-33643. Henriques, S.T., Huang, Y.-H., Chaousis, S., Sani, M.-A., Poth, A.G., Separovic, F., Craik, D.J., 2015. The prototypic cyclotide kalata B1 has a unique mechanism of entering cells. Chem. Biol. 22, 1087-1097. Henriques, S.T., Huang, Y.-H., Rosengren, K.J., Franquelim, H.G., Carvalho, F.A., Johnson, A., Sonza, S., Tachedjian, G., Castanho, M.A.R.B., Daly, N.L., Craik, D.J., 2011. Decoding the membrane activity of the cyclotide kalata B1: The importance of phosphatidylethanolamine phospholipids and lipid organization on hemolytic and anti-HIV activities. J. Biol. Chem 286, 24231-24241. Hinderling, P.H., 1997. Red blood cells: a neglected compartment in pharmacokinetics and pharmacodynamics. Pharmacol. Rev. 49, 279-295. Huang, Y.-H., Colgrave, M.L., Clark, R.J., Kotze, A.C., Craik, D.J., 2010. Lysine-scanning mutagenesis reveals an amendable face of the cyclotide kalata B1 for the optimisation of nematocidal activity. J. Biol. Chem 285, 10797- 10805. Huang, Y.-H., Henriques, S.T., Wang, C.K., Thorstholm, L., Daly, N.L., Kaas, Q., Craik, D.J., 2015. Design of substrate-based BCR-ABL kinase inhibitors using the cyclotide scaffold. Sci. Rep. 5, 12974. Jain, R.K., Gerlowski, L.E., Weissbrod, J.M., Wang, J., Pierson, R.N., 1981. Kinetics of uptake, distribution, and excretion of zinc in rats. Ann. Biomed. Eng. 9, 347-361. Khazanov, N.A., Carlson, H.A., 2013. Exploring the composition of protein-ligand binding sites on a large scale. PLoS Comp. Biol. 9, e1003321. Kimura, R.H., Teed, R., Hackel, B.J., Pysz, M.A., Chuang, C.Z., Sathirachinda, A., Willmann, J.K., Gambhir, S.S., 2012. Pharmacokinetically stabilized cystine knot peptides that bind alpha-v-beta-6 integrin with single-digit nanomolar affinities for detection of pancreatic cancer. Clin. Cancer. Res. 18, 839-849. Kochansky, C.J., McMasters, D.R., Lu, P., Koeplinger, K.A., Kerr, H.H., Shou, M., Korzekwa, K.R., 2008. Impact of pH on plasma protein binding in equilibrium dialysis. Mol. Pharm. 5, 438-448. Koehbach, J., O’Brien, M., Muttenthaler, M., Miazzo, M., Akcan, M., Elliott, A.G., Daly, N.L., Harvey, P.J., Arrowsmith, S., Gunasekera, S., Smith, T.J., Wray, S., Göransson, U., Dawson, P.E., Craik, D.J., Freissmuth, M., Gruber, C.W., 2013. Oxytocic plant cyclotides as templates for peptide G protein-coupled receptor ligand design. Proc. Natl. Acad. Sci. U. S. A. 110, 21183-21188. Krause, S., Schmoldt, H.U., Wentzel, A., Ballmaier, M., Friedrich, K., Kolmar, H., 2007. Grafting of thrombopoietin-mimetic peptides into cystine knot miniproteins yields high-affinity thrombopoietin antagonists and agonists. FEBS J. 274, 86-95. Melander, E., Eriksson, C., Jansson, B., Goransson, U., Hammarlund-Udenaes, M., 2016. Improved method for quantitative analysis of the cyclotide kalata B1 in plasma and brain homogenate. Biopolymers 106, 910-916. Miao, Z., Ren, G., Liu, H., Kimura, R.H., Jiang, L., Cochran, J.R., Gambhir, S.S., Cheng, Z., 2009. An engineered knottin peptide labeled with 18F for PET imaging of integrin expression. Bioconj. Chem. 20, 2342-2347.

Moore, S.J., Hayden Gephart, M.G., Bergen, J.M., Su, Y.S., Rayburn, H., Scott, M.P., Cochran, J.R., 2013. Engineered knottin peptide enables noninvasive optical imaging of intracranial medulloblastoma. Proc. Natl. Acad. Sci. U. S. A. 110, 14598-14603. Murakami, T., Yumoto, R., 2011. Role of phosphatidylserine binding in tissue distribution of amine-containing basic compounds. Expert Opin. Drug Metab. Toxicol. 7, 353-364. Musther, H., Olivares-Morales, A., Hatley, O.J.D., Liu, B., Rostami Hodjegan, A., 2014. Animal versus human oral drug bioavailability: Do they correlate? Eur. J. Pharm. Sci. 57, 280-291. Nielsen, D.S., Shepherd, N.E., Xu, W., Lucke, A.J., Stoermer, M.J., Fairlie, D.P., 2017. Orally absorbed cyclic peptides. Chem. Rev. 117, 8094-8128. Nworu, C.S., Ejikeme, T.I., Ezike, A.C., Ndu, O., Akunne, T.C., Onyeto, C.A., Okpalanduka, P., Akah, P.A., 2017. Anti-plasmodial and anti-inflammatory activities of cyclotide-rich extract and fraction of Oldenlandia affinis (R. & S.) D.C. (Rubiaceae). Afr. Health Sci. 17, 827-843. Plan, M.R., Göransson, U., Clark, R.J., Daly, N.L., Colgrave, M.L., Craik, D.J., 2007. The cyclotide fingerprint in Oldenlandia affinis: elucidation of chemically modified, linear and novel macrocyclic peptides. ChemBioChem 8, 1001-1011. Poth, A.G., Colgrave, M.L., Lyons, R.E., Daly, N.L., Craik, D.J., 2011. Discovery of an unusual biosynthetic origin for circular proteins in legumes. Proc. Natl. Acad. Sci. U. S. A. 108, 10127-10132. Poth, A.G., Huang, Y.-H., Le, T.T., Kan, M.-W., Craik, D.J., 2019. Pharmacokinetic analysis of kalata B1 and related grafted analgesics built on the cyclic cystine knot scaffold. Data in Brief, Submitted. Pye, C.R., Hewitt, W.M., Schwochert, J., Haddad, T.D., Townsend, C.E., Etienne, L., Lao, Y., Limberakis, C., Furukawa, A., Mathiowetz, A.M., Price, D.A., Liras, S., Lokey, R.S., 2017. Nonclassical size dependence of permeation defines bounds for passive adsorption of large drug molecules. J. Med. Chem. 60, 1665-1672. Quimbar, P., Malik, U., Sommerhoff, C.P., Kaas, Q., Chan, L.Y., Huang, Y.-H., Grundhuber, M., Dunse, K., Craik, D.J., Anderson, M.A., Daly, N.L., 2013. High-affinity cyclic peptide matriptase inhibitors. J. Biol. Chem 288, 13885-13896. Rand, A.C., Leung, S.S.F., Eng, H., Rotter, C.J., Sharma, R., Kalgutkar, A.S., Zhang, Y., Varma, M.V., Farley, K.A., Khunte, B., Limberakis, C., Price, D.A., Liras, S., Mathiowetz, A.M., Jacobson, M.P., Lokey, R.S., 2012. Optimizing PK properties of cyclic peptides: the effect of side chain substitutions on permeability and clearance. MedChemComm 3, 1282-1289. Rosengren, K.J., Daly, N.L., Plan, M.R., Waine, C., Craik, D.J., 2003. Twists, knots, and rings in proteins. Structural definition of the cyclotide framework. J. Biol. Chem 278, 8606-8616. Saether, O., Craik, D.J., Campbell, I.D., Sletten, K., Juul, J., Norman, D.G., 1995. Elucidation of the primary and three-dimensional structure of the uterotonic polypeptide kalata B1. Biochemistry 34, 4147-4158. Sanhajariya, S., Duffull, S., Isbister, G., 2018. Pharmacokinetics of snake venom. Toxins 10, 73. Smith, D.A., Di, L., Kerns, E.H., 2010. The effect of plasma protein binding on in vivo efficacy: misconceptions in drug discovery. Nat. Rev. Drug Discov. 9, 929-939. Song, Y., Wu, L., Lu, C., Li, N., Hu, M., Wang, Z., 2014. Microwave-assisted liquid-liquid microextraction based on solidification of ionic liquid for the determination of sulfonamides in environmental water samples. J. Sep. Sci. 37, 3533-3538. Srikanth, C.H., Chaira, T., Sampathi, S., V. B, S., Bambal, R.B., 2013. Correlation of in vitro and in vivo plasma protein binding using ultracentrifugation and UPLC-tandem mass spectrometry. Analyst 138, 6106-6116. Stass, H., 2006. Determination of minimal sampling time points for reliable pharmacokinetic evaluation of recombinant factor VIII – an exploratory population pharmacokinetic analysis in paediatric patients suffering from severe haemophilia. Haemophilia 12, 50-55. Tang, L., Persky, A.M., Hochhaus, G., Meibohm, B., 2004. Pharmacokinetic aspects of biotechnology products. J. Pharm. Sci. 93, 2184-2204. Thell, K., Hellinger, R., Sahin, E., Michenthaler, P., Gold-Binder, M., Haider, T., Kuttke, M., Liutkevičiūtė, Z., Göransson, U., Gründemann, C., Schabbauer, G., Gruber, C.W., 2016. Oral activity of a nature-derived cyclic peptide for the treatment of multiple sclerosis. Proc. Natl. Acad. Sci. U. S. A. 113, 3960-3965.

Thongyoo, P., Bonomelli, C., Leatherbarrow, R.J., Tate, E.W., 2009. Potent inhibitors of -tryptase and human leukocyte elastase based on the MCoTI-II scaffold. J. Med. Chem. 52, 6197-6200. Treggiari, D., Zoccatelli, G., Chignola, R., Molesini, B., Minuz, P., Pandolfini, T., 2017. Tomato cystine-knot miniproteins possessing anti-angiogenic activity exhibit in vitro gastrointestinal stability, intestinal absorption and resistance to food industrial processing. Food Chem. 221, 1346-1353. Valko, K., Teague, S., Pidgeon, C., 2017. In vitro membrane binding and protein binding (IAM MB/PB technology) to estimate in vivo distribution: Applications in early drug discovery. 5, 14. Wang, C.K., Stalmans, S., De Spiegeleer, B., Craik, D.J., 2016. Biodistribution of the cyclotide MCoTI-II, a cyclic disulfide-rich peptide drug scaffold. J. Peptide Sci. 22, 305-310. Werle, M., Kafedjiiski, K., Kolmar, H., Bernkop-Schnürch, A., 2007. Evaluation and improvement of the properties of the novel cystine-knot microprotein McoEeTI for oral administration. Int. J. Pharmaceut. 332, 72-79. Werle, M., Schmitz, T., Huang, H.-L., Wentzel, A., Kolmar, H., Bernkop-Schnuerch, A., 2006. The potential of cystine-knot microproteins as novel pharmacophoric scaffolds in oral peptide drug delivery. J. Drug Target. 14, 137-146. Wong, C.T.T., Rowlands, D.K., Wong, C.-H., Lo, T.W.C., Nguyen, G.K.T., Li, H.-Y., Tam, J.P., 2012. Orally active peptidic bradykinin B1 receptor antagonists engineered from a cyclotide scaffold for inflammatory pain treatment. Angew. Chem. Int. Ed. 51, 5620-5624. Yu, S.X., Li, S., Yang, H., Lee, F., Wu, J.T., Qian, M.G., 2005. A novel liquid chromatography/tandem mass spectrometry based depletion method for measuring red blood cell partitioning of pharmaceutical compounds in drug discovery. Rapid Commun. Mass Spectrom. 19, 250-254.

Declaration of interests

☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Pharmacokinetic characterization of kalata B1 and related therapeutics built on the cyclotide scaffold

Aaron G. Potha, Yen-Hua Huanga, Thao T. Lea, Meng-Wei Kana, David J. Craika*

aInstitute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland 4072, Australia

CRediT author statement:

Aaron Poth: Conceptualization, Methodology, Data curation, Investigation, Writing- Original draft.

Yen-Hua Huang: Conceptualization, Methodology, Writing- Reviewing and Editing.

Thao Le: Methodology.

Meng-Wei Kan: Visualization.

David Craik: Conceptualization, Writing- Reviewing and Editing, Supervision.