In vivo Efficacy and Pharmacokinetics of Optimized Apidaecin Analogs

Proline-rich antimicrobial peptides (PrAMPs) represent promising alternative therapeutic options for the treatment of multidrug-resistant bacterial infections. PrAMPs are predominantly active against Gram-negative bacteria by inhibiting protein expression via at least two different modes of action, i.e., blocking the ribosomal exit tunnel of 70S ribosomes (oncocin-type binding) or inhibiting the assembly of the 50S ribosomal subunit (apidaecin-type binding). The in vivo efficacy and favorable biodistribution of oncocins confirmed the therapeutic potential of short PrAMPs for the first time, whereas the in vivo evaluation of apidaecins is still limited despite the promising efficacy of apidaecin-analog Api88 in an intraperitoneal murine infection model. Here, the in vivo efficacy of apidaecin-analog Api137 was studied, which rescued all NMRI mice from a lethal intraperitoneal infection with E. coli ATCC 25922 when administered three times intraperitoneal at doses of 0.6 mg/kg starting 1 h after infection. When Api88 and Api137 were administered intravenous or intraperitoneal at doses of 5 and 20 mg/kg, their plasma levels were similarly low (<3 μg/mL) and four-fold lower than for oncocin-analog Onc72. This contradicted earlier expectation based on the very low serum stability of Api88 with a half-life time of only ~5 min compared to ~6 and ~3 h for Api137 and Onc72, respectively. Pharmacokinetic data relying on a sensitive mass spectrometry method utilizing multiple reaction monitoring and isotope-labeled peptides revealed that Api88 and Api137 were present in blood, urine, and kidney, and liver homogenates at similar levels accompanied by the same major metabolites comprising residues 1–16 and 1–17. The pretended discrepancy was solved, when all peptides were incubated in peritoneal lavage. Api137 was rapidly degraded at the C-terminus, while Api88 was rather stable despite releasing the same degradation products. Onc72 was very stable explaining its higher plasma levels compared to Api88 and Api137 after intraperitoneal administration illuminating its good in vivo efficacy. The data indicate that the degradation of therapeutic peptides should be studied in serum and further body fluids. Moreover, the high efficacy in murine infection models and the fast clearance of Api88 and Api137 within ~60 min after intravenous and ~90 min after intraperitoneal injections indicate that their in vivo efficacy relates to the maximal peptide concentration achieved in blood.


Pharmacokinetics and In Vivo Efficacy of Optimized Apidaecin Derivatives
Rico Schmidt, Daniel Knappe, Eszter Ostorházi, and Ralf Hoffmann  Table S1. Sequences, monoisotopic masses, and MRM transitions of apidaecins S5 Table S2. Equipment for LC-MS analysis S6 Table S3. Ion source and MS analyzer settings for ESI-QqLIT-MS (4000 QTRAP ® ) S7 Table S4. MS/MS settings for ESI-QqLIT-MS (4000 QTRAP ® ) S8 Table S5. Ion source and MS analyzer settings for ESI-LTQ-Orbitrap XL TM -MS S9 Table S6. LOD, LOQ, and LDR obtained from calibration without matrix S9 Table S7. Recovery and matrix effects after solid phase extraction S10 Table S8. Precision and accuracy of the RPC-MRM method S11 Table S9. LOD and LOQ obtained from calibration in urine and organ homogenates S12 Table S10. Recovery rates in blood and organ homogenates S13 Table S11. Api88 peptide recoveries calculated for blood, organs, and total animal S14 Table S12. In vivo metabolites in murine kidney homogenates S15 Table S13. Api137 peptide recoveries calculated for blood, organs, and total animal S16 Figure S1. ESI-QqLIT-MS and MS/MS of Api88 and Api137 S17 Figure S2. Influence of pH on signal intensities in MRM of Api88 and Api137 S18 Figure S3. Chromatograms of MRM-analysis of Api88 and Api137 S19 Figure S4. Pharmacokinetics of Api88/Api137 and their metabolites (5 mg/kg, i.p) S20 Figure S5. Exponential fittings of concentration profiles after i.p. administration. S21 Figure S6. Abundance of Api88/Api137 metabolites in blood after i.p. administration S22

Method M1: Multiple reaction monitoring (MRM) transitions for apidaecin peptides
Peptides (1 µg/mL) were dissolved in aqueous acetonitrile (25% or 60% v/v) containing either formate buffer (26 mmol/L, pH 2.6, 3.0, 3.3 or 3.9) or formic acid (0.1 % v/v) and infused (flow rate 5 µL/min, PHD 2600 syringe pump) into the ESI-QqLIT-MS (4000 QTRAP ® ) to optimize buffer composition and MRM. Transitions were selected from the tandem mass spectra (MS/MS) and confirmed on an ESI-QqTOF-MS (QSTAR ® pulsar I) providing higher mass resolutions and mass accuracies than the QTRAP ® . The settings for declustering potential (DP), collision potential (CE), and collision cell exit potential (CXP) were optimized by ramping the respective potentials using the compound optimization tool of the Analyst ® software.
Limits of detection (LOD) and quantification (LOQ) corresponded to the lowest peptide concentrations in water, plasma, or homogenates of brain, liver, and kidney providing a peak height exceeding the noise in the retention time window at least threefold (LOD) or ten times (LOQ). Recovery rates of SPE and matrix effects were determined by the peak areas of spiked S3 plasma samples and organ homogenates relative to equally concentrated peptide solutions in pure water. Concentrations of samples below the LOQ but still detectable were set to the value of the LOD, whereas signals with peak heights below the LOD were set to 0 µg/mL.

Method development and validation
Apidaecin peptides were selectively and sensitively quantified based on the methods established for oncocins Onc72 and Onc112 (Schmidt R. et al JAC 2016). MRM based on triply charged precursor ions ( Figure S1A and C) and a neutral loss of two dimethyl amine molecules cleaved from the N-terminal guanidine group as most intense fragment ions ( Figure S1B and D). To avoid misinterpretation of possible cross talks, an additional MRM transition with an m/z above the m/z selected in quadrupole 1 was used (Table S4). MS settings were optimized in order to obtain the highest sensitivities (Table S4) and the pH of the eluents was elevated to 3.0 to increase the intensity of the signals corresponding to the triply protonated precursor ions ( Figure S2). Interfering sample components, such as proteins, salts, and lipids, were removed by solid phase extraction (OASIS HLB plates) before the samples were analyzed by RP-HPLC-ESI-QqLIT-MS (flow rate: 0.2 mL/min, column temperature: 55°C, linear gradient from 2.7 to 20.7% acetonitrile containing 26 mmol/L formate buffer (pH 3.0) in five minutes) with an overall duration of 20 min. Polypropylene tubes were used for all handling steps, as the highly basic peptides adsorbed strongly to glass surfaces resulting in a hundred fold reduced sensitivity.
The method provided instrumental LODs of 14 and 28 ng/mL for Api88 and Api137, respectively, and only slightly higher LOQs (Table S6). The linear dynamic range spanned S4 more than two orders of magnitude up to 10 µg/mL. When all steps of the analytics were considered, the LODs increased for Api137 and even more for Api88 (Table S9), most likely through losses during SPE, which had recoveries between 35 and 56% for all four peptides, with higher recoveries for low concentrations (Table S7). The successful removal of isobaric components was confirmed by the absence of signals after injection of plasma (untreated animals) after SPE. The weak matrix effects on the signal intensities (<10%) with enhancing effects at low peptide concentrations further confirmed this. The precisions and accuracies met the requirements very well (Table S8) indicating that the developed method allowed a reliable and sensitive quantification of apidaecin peptides in biological samples. However, the LODs and LOQs increased in organ homogenates, especially for liver, and in urine samples (Table   S9).   Table S6. Analytical parameters calculated from a serial dilutions series of Api88 and Api137 analyzed by the optimized MRM method. Peptide solutions (10 µL) were prepared in water without any biological matrix.

Peptide LOD LOQ LDR R²
Api88 14 ng/mL 15 ng/mL 7*10 2 0.9911 Api137 28 ng/mL 29 ng/mL 3*10 2 0.9988 LOD, LOQ, and LDR denote limit of detection, limit of quantification, and linear dynamic range, respectively. S13 Table S10. Peptide amounts recovered in mice samples after intraperitoneal administration of Api88. Quantities were corrected by the recovery rates determined for plasma and each organ homogenate (Table S12). Absolute quantities were calculated by assuming a total plasma volume of 25 mL/kg mouse weight and organ volumes determined in mL corresponded to the measured organ weights in g. Percentages of mean injected peptide amounts recovered in respective samples are provided in brackets.   Table S11. Peptide amounts recovered in mice samples after intraperitoneal administration of Api137. Quantities were corrected by the recovery rates determined for plasma and each organ homogenate (Table S12). Absolute quantities were calculated by assuming a total plasma volume of 25 mL/kg mouse weight and organ volumes determined in mL corresponded to the measured organ weights in g. Percentages of mean injected peptide amounts recovered in respective samples are provided in brackets.   Table S12. Recovery rates of Api88 and Api137 determined in plasma and organ homogenates (n = 3) collected from untreated NMRI mice. Solutions of Api88 (blood: 10 µL, 25 and 100 µg/mL, brain: 500 µL, 0.2 or 1 µg/mL, liver: 500 µL, 2 or 5 µg/mL, kidneys: 500 µL, 20 or 60 µg/mL) and Api137 (blood: 10 µL, 2.5 and 10 µg/mL, brain: 500 µL, 0.2 or 2 µg/mL, liver: 500 µL, 1 or 2.5 µg/mL, kidneys: 500 µL, 20 or 60 µg/mL) were added prior to homogenization and centrifugation. Peptide concentrations were calculated based on the assumption that peptides were equally distributed in plasma (one third of the 2.5 mL blood volume of mice) and present only in the supernatant of the homogenates, i.e. neglecting peptide contents remaining in the pellets. The total peptide amount in mice was calculated from the recovery rates shown in bold.