Development of generic immunoassay for the detection of a series of aminoglycosides with 6′-OH group for the treatment of genetic diseases in biological samples

Abstract

Over the last two decades, a growing number of scientific evidences highlighted the potential therapeutic value of several structures of aminoglycoside antibiotics (including gentamicin and G418) for the treatment of various genetic diseases caused by nonsense mutations. These findings resulted in a fast evolvement of synthetic derivatives of aminoglycosides which were shown to be more target specific and less toxic than the clinically used antibiotics. The emerging progress in drug design and development has necessitated the urge to develop a fast, easy and accurate procedure for the determination of these potential therapeutic agents in various biologically derived matrices. Here we describe the preparation of a generic polyclonal antibody that was used for the development of homologous and heterologous immunoassays for the detection of a wide range of natural and synthetic aminoglycoside derivatives, highlighted today as potential therapeutic agents for the treatment of various genetic diseases. A common two-ring scaffold, NB82, present in the majority of compounds exhibiting potent biological activity, was used as a generic immunization hapten for the immunization of two rabbits. By using a series of chemical steps, NB82 was selectively conjugated via the N-1 position through glutaric acid linker to a carrier protein. Sensitivity (I₅₀) values for the recognition of three representative compounds NB82, NB84 and NB124 were determined to be 10±3 ng mL⁻¹, 0.5±0.04 μg mL⁻¹ and 1±0.12 μg mL⁻¹, respectively. Limits of detection were determined to be 1±0.3 ng mL⁻¹ for NB82, 20±7 ng mL⁻¹ for NB84 and 15±8 ng mL⁻¹ for NB124. The developed assays were further exploited for the in-vivo monitoring of the therapeutic compounds in mice serum. Serum experimentations exhibited similar detection limits as observed for the PBS calibration experiments, demonstrating no interference with assays sensitivity, with rather high recovery ratios ranging from 92–107% in whole blood samples.

1. Introduction

Aminoglycosides are highly potent antibacterial agents, which are known to exert their deleterious effects on bacterial cells by interfering with the translation process [1,2]. Over the last 20 years, a few natural aminoglycosides have been highlighted as promising therapeutic agents for the treatment of several genetic disorders caused by nonsense mutations [3–5]. These mutations generally lead to the production of truncated, non-functional proteins. In human, these mutations have been linked to nearly 2,000 genetic disorders such as cystic fibrosis (CF), Duchenne muscular dystrophy (DMD), ataxia-telangiectasia, Hurler syndrome, hemophilia A, hemophilia B and Tay-Sachs [6,7]. For many of these diseases there is no effective treatment. Aminoglycoside antibiotics including gentamicin, paromomycin and G418 (Fig. 1) have been shown to suppress premature termination codons and partially restore functional protein production for more than twenty genetic diseases [8]. However, the potential toxicity associated with aminoglycosides has limited their clinical applications in suppression therapy. Over the last few years several synthetic derivatives of aminoglycosides were developed to overcome some of the enhanced toxicity issues. Many of these derivatives have been designed and synthesized in our lab [9–12]. Our most recent lead compounds of NB-series, including NB74, NB84, and NB122-NB125 (Fig. 1) demonstrated a remarkable biological activity altogether with reduced toxicity values in various in-vitro, ex-vivo and in-vivo systems [10–14].

Figure 1.

Chemical structures of standard and semi-synthetic aminoglycoside derivatives used in this study.

Further progress in drug development necessitated a fast and effective methodology for the detection and quantification of the developed lead NB-compounds in various biological derived matrices. Over the last few decades many analytical and bioanalytical assays have been suggested for the qualitative and quantitative analysis of aminoglycosides that are in clinical, veterinary and agricultural use to treat bacterial infections. These methods were applied for the detection of drug residues in a wide range of biological derived matrices such as cell tissues [15], serum [16], milk [17], eggs [15,18], and honey [18]. Chemical methodologies included the application of gas chromatography (GC) [19], thin layer chromatography (TLC) [20], high-performance liquid chromatography (HPLC) [21] and capillary electrophoresis (CE) [22]. Biological methodologies included the development of microbiological assays [23], radiochemical and radioimmunochemical assays (RIA) [24], enzyme linked- and fluoro- immunoassays (ELISA, FIA) [16, 25, 26], nano-sensors and nano-particle based immunoassays [27,28].

Immuno-based methodologies such as ELISA, FIA and RIA have shown to be as sensitive and accurate as the chemical methods [29]. These methods could be easily applied on a wide variety of biologically derived matrices, are simple to perform and analyze, and do not require the use of high cost instrumentations. Immuno-based methodologies are commercially available for the detection of some widely used natural aminoglycosides such as gentamicin, streptomycin, dihydrostreptomycin, neomycin and kanamycins (e.g. MaxSignal® gentamicin ELISA Test Kit, Aminoglycosides enzyme immune-assay-EIA kit for the detection of gentamicin, neomycin, streptomycin and dihydrostreptomycin - Europroxima, and kanamycins ELISA kit - Wanger). These commercially available kits can be exploited for highly sensitive detection of aminoglycosides in various matrices, and some of them are used in hospitals to monitor the serum levels of clinically used aminoglycosides such as gentamicin and amikacin. However, these assays are mostly used for the detection of aminoglycoside antibiotics that are based on a neamine core, containing an amino group at their 6′ position (ring I, Fig. 1). The use of the derived antibodies for the detection of therapeutic derivatives containing other substituents on the neamine core is quite limited due to low cross reactivity values. Recent documentations demonstrated the superiority of 6′-hydroxyl containing aminoglycosides such as paromomycin and G418 over the 6′-amino derivatives in the treatment of various genetic disorders [4]. These derivatives are mostly based on either a paromamine or a 6′-(R)-methyl-paromamine (NB82) 2-ring core (Fig. 1) and are extensively explored for the treatment of various genetic diseases [3, 9–14]. To our knowledge, no immunoassays exist for the detection of aminoglycosides containing a 6′-hydroxyl group.

The aim of the present study was to generate a generic antibody that will be able to recognize a wide range of potential therapeutic members containing a 6′-hydroxyl group. For this purpose, the 2-ring scaffold NB82 designed in our lab, was chosen to serve as an immunogen for the development of a generic antibody. The resulting antibody was shown to cross react with a series of standard and synthetic derivatives of aminoglycosides that shared structural similarity with the immunization scaffold using a homologous ELISA. The antibody was further used for the development of 3 highly sensitive heterologous ELISAs for the detection of selected potential therapeutic agents. The assays were shown to have high recovery from spiked blood and serum samples and have been applied for in-vivo monitoring of the synthetic derivatives in mice serum. The approach of a generic immunoassay developed in this study can be applied for the development of similar assays for the detection of other classes of drug candidates.

2. Materials and methods

2.1. Materials

Paromomycin, neomycin, geneticin (G418), gentamicin and kanamycin were purchased from Sigma as sulfate salts. All the synthetic derivatives including neamine, paromamine, NB82, NB74, NB84, NB122, NB123, NB124 and NB125 were prepared as described previously by Baasov and coworkers [9–12] and were used as their sulfate salts. All other chemicals and biochemicals, unless otherwise stated, were obtained from Merck or Sigma. Rabbits immunization has been performed in Sigma-Aldrich (Rehovot, Israel) following the standard immunization protocols.

2.2. General techniques

¹HNMR spectra was recorded on a Bruker Avance^™ 500 spectrometer at 500 MHz. Chemical shifts reported (in ppm) were relative to the internal standard Me₄Si (δ = 0.0) with CDCl₃ as the solvent. ¹³CNMR spectra was recorded on a Bruker Avance^™ 500 spectrometer at 125.8 MHz, and the chemical shifts reported (in ppm) were relative to the residual solvent signal for CDCl₃ (δ = 77.00). Mass spectra analysis was obtained on a Bruker Daltonix Apex 3 mass spectrometer under electrospray ionization (ESI). Reactions were monitored by TLC on Silica Gel 60 F254 (0.25 mm, Merck), and spots were visualized by charring with a yellow solution containing (NH₄)Mo₇O₂₄•4H₂O (120 g) and (NH₄)₂Ce(NO₃)₆ (5 g) in 10% H₂SO₄ (800 mL). Flash column chromatography was performed on Silica Gel 60 (70–230 mesh).

Data analysis

The sensitivity (I₅₀) and limit of detection (I₂₀) values were obtained from fitting concentration vs. % binding curves to a log-logit plot using the Origin software, version 8.5 (Mirocal Software, North Hampton, MA, USA). Correlation coefficient between the assays linear range (I₂₀-I₈₀) was ≥0.98 in all standard curves. Data precision was calculated according to the National Committee of Clinical Laboratory Services guideline [30].

2.3 Preparation of 6′-(R)-methyl-2′,3-diazido-1-N-[5-oxopentanoic acid]-paromamine (compound B, Fig. 2)

Figure 2.

Synthetic scheme for the preparation of immunization\coating conjugates.

A mixture of compound A [10] (500 mg, 1.2 mmol) and glutaric anhydride (0.29 mg, 2.5 mmol) in 20 mL anhydrous methanol was stirred under argon for 6 h at room temperature. The reaction was monitored by TLC (EtOAc/MeOH 60:40) which indicated completion after 6 h. The reaction mixture was evaporated under vacuum to dryness and purified by flash chromatography, resulting in 400 mg (62% yield) of pure compound B (Fig. 2). The purity of the product was determined by NMR and MALDI-TOF Mass Spectra analysis. ¹HNMR (CDCl₃, 500 MHz): ‘Ring I’ δH 1.26 (d, 3H, J=6.0 Hz, CH3), 3.10 (dd, 1H, J1=4.2 Hz, J2=10.5 Hz, H-2′), 3.38 (t, 1H, J=9.7 Hz, H-4′), 3.92–3.98 (m, 2H, H-5′ and H3′), 4.04 (m, 1H, H-6′), 5.74 (d, 1H, J=3.1 Hz, H-1′); ‘Ring II’ δH 1.43 (ddd, 1H, J1= J2= J3=12.5 Hz, H-2ax), 2.20 (td, 1H, J1=4.5, J2=12.5 Hz, H-2eq), 3.26 (t, 1H, J=9.3 Hz, H-4), 3.52 (m, 3H, H-3, H-5, H-6), 3.78 (m, 1H, H-1). Linker peaks appeared in the spectrum as follows: δH 1.92 (t, 2H, J=7.2 Hz), 2.29 (m, 2H), 2.36 (t, 2H, J=7.5 Hz). ¹³CNMR(CDCl₃, 125.8 MHz): δC 18.0, 22.2, 33.9, 34.2, 36.2, 50.4, 61.4, 64.7, 69.3, 72.3, 74.2, 75.2, 76.2, 78.7, 80.3, 98.7 (C-1′), 175.5 (C=O), 177.1 (C=O). MS: Calculated mass for C₁₈H₃₀N₇O₁₀ ([M+H]+) m/z: 504.20; measured m/z: 504.20.

2.4 Preparation of immunization conjugate (compound E, Fig. 2)

A mixture of compound B (100 mg, 0.2 mmol), N-hydroxysuccinimide (NHS) (23 mg, 0.2 mmol) and N′, N′-dicyclohexylcarbodiimide (DCC) (41.2 mg, 0.2 mmol) in 10 mL dimethylformamide (DMF) was stirred for 4 h at room temperature followed by a 12 h incubation at 4 °C. The mixture was then centrifuged for 15 min at 2,500× g at room temperature. 1 mL of the supernatant containing 10 mg of the compound C were added to 10 mg KLH dissolved in 4 mL of 0.13 M NaHCO₃ at pH 9.2. The reaction was incubated 12 h at 4 °C, followed by size exclusion product purification for 25 min at 3,000× g at room temperature using Centricon 30 (Amicon, Millipore, Billerica, MA) to afford compound D. The concentrated product D (Fig. 2) was washed 3 times with 4 mL of NaHCO₃ solution. The final volume was adjusted to 4 mL by adding NaHCO₃ solution. The reduction of azido-groups was performed by adding 100 μL trimethylphosphine (PMe₃) solution (1 M solution in tetrahydrofuran - THF) and the mixture was allowed to react 12 h at 4 °C. The final product E (referred as a compound E, Fig. 2) was further purified using size exclusion purification as described above, followed by 2 washes with 5 mL NaHCO₃ solution. The final volume was adjusted to 5 mL by adding double distilled water (DDW) and was stored at −20°C pending to use.

2.5 Immunization and Antiserum production

The immunization conjugate (compound E from the above section, 0.5 mL) were emulsified with either Complete Freund’s adjuvant (1st immunization), or incomplete adjuvant (2nd-4th boosts) prior to immunization. The emulsion was injected subcutaneously to two rabbits 8244 and 8245 (~1 mg conjugate were injected per rabbit at each time point). Bleeds were collected after each boost and were tested for anti-antigen activity using the checkerboard assay (Fig. S1). The 3rd bleed from Rabbit 8245 has been shown to have the highest anti-antigen activity, and was therefore used for further experimentations (Fig. S1).

2.6 Preparation of NB82-OVA coating antigen

NB82-OVA conjugates were prepared similarly to immunization conjugate, as described above, except that 5 mg ovalbumin (OVA), that were dissolved in 2.5 mL NaHCO3 solution, were conjugated to compound B at molar ratios of 1:1, 1:2, 1:5 and 1:10. The final volume of the resulted NB82-OVA conjugate was adjusted to 1 mL by adding 0.13 M NaHCO₃, and the pH of the solution was adjusted to 9.6 by adding dilute solution (0.01 M) of NaOH. Protein content has been determined to be 5 mg mL⁻¹ using the Bradford assay (BioRad protein assay, BioRad, Hercules, CA) and the resulting conjugates antigen activity was evaluated towards 3rd bleed antiserum from Rabbits 8244 and 8245 using the checkerboard assay (data not shown). The 1:1 NB82-OVA conjugate crossed with 3rd bleed antiserum derived from Rabbit 8245 has been shown to have the highest sensitivity, and was therefore used for further experimentations (Fig. S2). The conjugates were stored in aliquots at −20°C.

2.7 Preparation of NB82/84/124-GTA-OVA coating antigens

To prepare coating antigens for heterologous-cross reactivity assays we used glutaraldehyde (GTA) coupling method as described by Chen at al. [31]. In brief, 20 mg OVA was dissolved in 10 mL of 0.01 M phosphate-buffered saline (PBS) at pH 6.5 (the pH adjustment was performed prior to the addition of OVA by adding dilute solution, 0.01 M, of HCl). 20 mg of each compound (NB82/84/124 sulfate) was added to the solutions and stirred for a few min at room temperature 1.5 mL of freshly prepared 1% GTA solution was added in a drop-wise manner using a 1 mL syringe equipped with a 0.5 mm × 16 mm needle. The reaction mixture was allowed to be stirred at room temperature for 15 min, of which yellow color indicated the imine formation. Reduction of imines to amines has been performed by adding 115 mg sodium borohydride, followed by 12 h incubation at 4 °C. The final product has been purified using Centricon 30 for 30 min at 3,000× g, followed by 2 washes with 1 mL 0.01 M PBS pH 7.2. The final volume of conjugates has been adjusted to 1 mL using PBS pH 7.2, and their protein content has been determined to be 17.5–19 mg mL⁻¹ using the Bradford methodology. Conjugates antigen reactivity was tested against 3rd bleed antiserum from Rabbits 8244 and 8245 using the checkerboard assay (data not shown). The conjugates were stored in aliquots at −20°C pending to use.

2.8 NB82 indirect competitive ELISA

The assay developed was an indirect competitive ELISA in which tested samples or standard solutions were allowed to compete with an immobilized coating antigen for antiserum binding. The assay has developed by using microtiter 96 well plates (F96 Maxisorp, Nunc Immuno Plate, Roskilde, Denmark). Wells were coated with 100 μL NB82-OVA (1:1) conjugate, diluted 1:8,000 (62.5 ng conjugate per well) in 0.5 M carbonate buffer, pH 9.6 (CB). After an overnight incubation at 4°C, wells were washed 3 times with PBST (defined as PBS pH 7.2 containing 0.3% [v/v] Tween-20). 50 μL of unknown samples or standard solutions (in PBS, mice serum or Fetal bovine serum - FBS, Biologycal Industries, Israel) were added to the wells, followed by the addition of 50 μL of antiserum (Rabbit 8245) diluted 1:80,000 in PBST. The standard samples were comprised of 12 serial dilutions of NB82, ranging from 10-0.0049 μg mL⁻¹ dissolved in PBS or serum. Plates were incubated overnight at 4°C, and washed 3 times with PBST. 100 μL secondary antibody (2nd Ab) conjugated to horseradish peroxidase (HRP) (anti-rabbit HRP conjugated, Sigma), diluted 1:20,000 in PBST were added to each well. Plates were incubated for 2 h at room temperature, rinsed with PBST, and tested for HRP activity by the addition of 100 μL substrate solution - 3,3′,5,5′-tetramethyl benzidine (TMB substrate chromagen, Dako, Glostrup, Denmark). The reaction was stopped after 20 min by adding 50 μL of 4 N sulfuric acid, and the absorbance was measured with ELISA reader (Spectra Max M2, Molecular devices, Sunnyvale, CA) at 450 nm. In all experiments a 5 mg mL⁻¹ OVA solution in PBS has been used as a negative control to evaluate and normalize the unspecific binding of antisera. Unspecific binding values obtained were almost negligible in all experiments. Wells coated with coating antigen that were not introduced with a free antigen sample were used as a positive control and determined the 100% binding of antisera.

2.9 Cross reactivity experiments

Cross reactivity values of antisera with various aminoglycosides were determined using the heterologous indirect ELISA as described above by adding the subjected compounds (instead of NB82) at 24 serial dilutions ranging from 10 mg mL⁻¹ to 1.19 ng mL⁻¹, except for NB84, of which serial dilution ranged from 1 mg mL⁻¹ to 0.119 ng mL⁻¹. All aminoglycosides tested were at their sulfated form and were dissolved and diluted in PBS or serum. The standard solutions of the compounds NB82, NB84 and NB124 were shown to be stable in PBS and standard serum (FBS) solutions that were frozen (−20°C) for a few months and demonstrated similar values using at least 5 repeated freezing/thawing cycles.

2.10 NB82/84/124 indirect competitive ELISA

In order to enhance antiserum sensitivity towards hetero-compounds that share similar epitops with the derivate of which antiserum Antibodies has been produced against, an indirect heterologous-assay has been developed. The assay has been performed similarly to the indirect assay for NB82 as described in section 2.8, except for the coating antigens of which for the NB82 heterologous-test, NB82-GTA-OVA conjugate diluted 1:128,000 in CB was used as the coating antigen; for the NB84 heterologous-test, NB84-GTA-OVA conjugate diluted 1:64,000 in CB was used as the coating antigen; and for the NB124 heterologous-test, NB124-GTA-OVA conjugate diluted 1:16,000 in CB was used as the coating antigen. Solutions that contained 17.5–19 mg mL⁻¹ OVA in PBS were used for coating negative control wells. Cross reactivity experiments have been performed as indicated in section 2.9.

2.11 Determination of NB compounds level in mice serum

Determination of NB compounds serum levels included an intraperitoneal (IP) injection of NB84/NB124 to mice and collection of blood samples at various time points. The protocol included the IP injection of 28.5 μg/gr body weight of the relevant NB compound to 7 male mice (C57/BL, 2.5–3.5 months old), followed by collection of fresh blood samples (~200 μL each) at 7 time points (5, 10, 20, 30, 40, 50, 60 min). Blood was collected to sterile tubes containing 20 μL of 50 mM EDTA pH 8.0. Serum was further extracted by centrifugation at 2,000× g at 4°C for 10 min. NB compounds content in serum samples was determined by using the pre-developed microtiter assay, as described above, when compared to a linearized calibration curve transformed to a log-logit plot using the Origin software, version 8.5. Each sample was tested in duplicates at five dilutions. All in-vivo samples were tested immediately after serum preparation. Serum recovery ratios were determined by spiking experiments of which whole blood samples derived from untreated mice were spiked with known amounts of the relevant aminoglycosidic derivatives. Blood samples were incubated for 1 h at 4°C followed by serum preparation by centrifugation as described above.

3. Results and Discussion

3.1 Preparation of immunization conjugate

The main goal of the present study was to develop an immunochemical assay for monitoring lead compounds of NB-series in biologically derived samples. The specific aims included the production of a generic polyclonal antiserum that will be able to recognize a wide variety of NB compounds (Fig 1.) that have previously shown to exhibit a great potential for the treatment of various nonsense mutations related genetic diseases [9–14].

NB82, a semi-synthetic 2-ring fragment derived from G418, is a common scaffold, appearing in nearly all compounds of NB-series (Fig. 1). We assumed that by choosing NB82 as hapten for immunogen preparation we could obtain polyclonal antibodies that will be able to recognize a diverse selection of these compounds. Similar methodology was previously described by Van Amerongen and coworkers [32] by using neamine as an immunogenic hapten for the generation of polyclonal antiserum capable of recognizing some natural aminoglycosides such as neomycin, kanamycin A and gentamicin.

Our next challenge was to choose a proper conjugation strategy for the preparation of immunization conjugate. Some of our lead compounds, like NB84, contain an (S)-4-amino-2-hydroxybutanoyl (AHB) moiety at their N-1 position (Fig. 1). Our earlier studies have demonstrated that the addition of AHB group selectively at N-1 position significantly enhances the readthrough activity of the resulted derivatives in comparison to that of the parent structure without the AHB, while its toxicity profile was not significantly affected [9–12]. Taking into account that the antibody raised against a hapten linked to the carrier protein through N-1 position would have similar affinity to both compounds with and without AHB, further substantiated our strategy for linking NB82 to a carrier protein selectively at N-1 position.

In order to selectively conjugate the NB82 scaffold to a carrier protein via the N-1 position, we first performed a selective azidation of NB82 at 2′ and 3 positions to yield 2′,3-diazido-NB82 (compound A, Fig. 2). Treatment of compound A with glutaric anhydride afforded compound B in 62% isolated yield. The carboxylic acid function of compound B at its N-1 side-chain was activated by reaction with N-hydroxysuccinimide in the presence of DCC to yield compound C. Conjugation of the NHS-activated compound C with the carrier protein KLH was followed by the conversion of two azido groups to the corresponding amines using Staudinger reaction (Me₃P, pH 9.2) to yield the immunization conjugate NB82-KLH (compound E, Fig. 2). The resulting conjugate was used as an immunization agent and was injected to two rabbits.

3.2 Development of an indirect competitive ELISA

The two rabbits immunized with the NB82-KLH conjugate showed high antibody serum titers, as indicated in checkerboard homologous-assays performed by using the NB82-OVA conjugate as a coating antigen (Figs. S1, S2). These experiments were further used to determine the optimal concentrations of coating antigen (NB82-OVA, 1:1), rabbit antiserum (8245) and secondary antibody to be 1:8,000, 1:80,000 and 1:20,000, respectively. These conditions were used to determine the I₅₀ value and assay’s detection limit (I₂₀) of NB82 to be 50±6 ng mL⁻¹ and 8±1 ng mL⁻¹ (n=10), respectively (Fig. 3, Table 1).

Figure 3.

Standard curve of NB82 (Homologous assay). NB82-OVA conjugate dilution 1:8,000; Antisera dilution (Rabbit 8245) 1:80,000. I₅₀ 50±6 ng mL⁻¹; Limits of detection: 8±1 ng mL⁻¹. Each value represents the mean±S.E.M. (Standard error mean) of 10 independent repeats performed in duplicates.

Table 1.

Concentrations at 50% inhibition (I₅₀) and limits of detection (I₂₀) of various aminoglycosides using the homologous ELISA with NB82-OVA as a coating antigen^a.

Compound	I50 (μg/mL)	Detection limit (I20) (μg/mL)
NB82	0.05±0.006	0.008±0.001
G418	0.4±0.04	0.06±0.02
NB125	1.25±0.14	0.3±0.03
Paromamine	3.0±0.34	0.6±0.1
NB84	3.1±0.75	0.7±0.1
NB74	3.5±0.15	0.7±0.1
NB124	12±1.1	3±0.7
Gentamicin	130±8.99	20±3
NB123	250±22.6	40±1
Paromomycin	325±56.6	50±2
NB122	400±34.2	80±4
Neomycin	700±71.7	90±4
Kanamycin A	1000±144	120±11
Neamine	2000±251	150±14

^aNB82-OVA conjugate dilution 1:8,000 was used as coating antigen; Antisera dilution (Rabbit 8245) 1:80,000. Each value represents the mean±S.E.M. of at least 3 experiments performed in duplicates.

In order to evaluate the efficiency of the assay for the detection of various aminoglycosides, 13 representative members were chosen to be detected for antiserum cross-reactivity via homologous-cross-reactivity assays, using the NB82-OVA conjugate as a coating antigen. Five of them including G418, gentamicin, paromomycin, neomycin and kanamycin A are commercial aminoglycosides (Fig. 1). Six compounds including NB74, NB84, NB122, NB123, NB124 and NB125, are semi-synthetic agents systematically developed in our laboratory for the treatment of genetic diseases caused by premature stop codon mutations [10–12]. Three additional compounds selected in this study included 2-ring structures as common scaffolds: neamine, a common scaffold of neomycin B and kanamycin A; paromamine, a common scaffold of paromomycin, NB122 and NB123; NB82, a common scaffold of G418, NB74, NB84, NB124 and NB125. As can be seen from the data in Fig. 4, antiserum antibodies were able to recognize all tested members at various sensitivity ranges, where NB82, as expected, demonstrated the lowest I₅₀ value, thus indicating highest antibody affinity (Table 1).

Figure 4.

Homologous assay cross reactivity. Antiserum cross reactivity of natural and synthetic representative aminoglycoside derivatives: NB82 ( ), geneticin (G418) ( ), NB125 ( ), paromamine (□), NB84 ( ), NB74 ( ), NB124 (△), gentamicin (○), NB123 ( ), paromomycin (+), NB122 ( ), neomycin ( ), kanamycin A ( ), and neamine ( ). NB82-OVA was used as coating antigen (dilution 1:8,000); Antiserum dilution (Rabbit 8245) 1:80,000. Each value represents the mean of at least 3 experiments performed in duplicates. Mean I₅₀ values ±S.E.M. are summarized in Table 1.

The affinities of serum antibodies towards the structurally related aminoglycosides containing a secondary (R)-6′-OH on ring I (NB82, G418, NB74, NB84, NB124, NB125) were higher than those obtained for similar compounds containing either a primary 6′-OH (NB122, NB123, paromomycin, paromamine) or a 6′-NH₂ (neomycin, neamine, gentamicin, kanamycin A). These results correlate well with our expectations that the resulting antibodies will show higher specificity towards compounds that share greater similarity with the parent compound used for immunization. Using NB82-OVA as the coating agent the antiserum indicated a rather high cross reactivity value of 12.5% for G418, with I₅₀ of 400 ng mL⁻¹ (Fig. 1, Table 1). These values can be attributed to the high number of similar epitops G418 shares with NB82, as can be seen in Fig. 1.

Cross reactivity values of the structurally related synthetic derivatives NB74, NB84, NB124 and NB125 where in the range of 0.4–4%, with I₅₀ values of 1.25–12 μg mL⁻¹ (Table 1). These compounds are derived from G418; therefore they all contain the NB82 scaffold. The observed 1- to 2-orders of magnitude gaps in cross reactivity values (when compared to G418) can be explained by the difference in their substitution pattern regarding the 2-deoxy-streptamine (2-DOS) ring (ring II, Fig. 1). NB compounds are 4,5 di-substituted 2-DOS derivatives, whereas G418 is a 4,6 di-substituted 2-DOS member. Since immunization conjugate has been linked via the N-1 position, it is reasonable to hypothesize that the epitops detected by the various antibody populations are located in the 1st sugar ring (ring I) and the glycosidic bond relative positions of the 2-DOS ring (ring II); as far as possible from the N-1 position. Following this line of thought, compounds of NB-series substituted with ring III at C5 position will enhance greater steric interference with the antibody than that of G418 in which ring III is substituted at C6 position. According to this rationale, compounds that differ only in the substitution pattern at N-1 position should have similar affinities to the antibody. Indeed, as can be seen from the data in Table 1, NB74 and NB84 that differ at their N-1 position demonstrated similar cross reactivity values of 3.5±0.15 μg mL⁻¹ and 3.1±0.75 μg mL⁻¹, respectively. Cross reactivity values for compounds that did not contain NB82 as a main scaffold were shown to be less than 0.04%, with I₅₀ values of 130–2000 μg mL⁻¹ (NB122, NB123, paromomycin, neomycin, gentamicin, kanamycin A and neamine, Fig. 4, Table 1).

Interestingly, the antibody populations generated were able to differentiate between the diastereomeric compounds that differ only by the configuration at a single chiral carbon center. Among the series of compounds tested in this study we had two such pairs of compounds: NB122 (5″-S) vs NB123 (5″-R), and NB124 (5″-S) vs NB125 (5″-R). The I₅₀ values of NB125 (5″-R) were 1 order of magnitude lower than those obtained for its 5″-diastereomer, NB124 (5″-S) (Table 1). Similarly, the values obtained for NB123 (5″-R) were twice lower than those obtained for NB122 (5″-S). These results demonstrate a stereo-selective preference towards the 5″-R isomers. An opposite tendency was observed for these compounds regarding their readthtough activity and citotoxicity values [11]. These observations could be attributed to the different pattern of interactions between the 5″-groups and the antibody binding pocket.

3.3 Development of heterologous-assays

The cross reactivity values obtained in the homologous assay demonstrated the antibodies ability to bind a wide range of synthetic and natural derivatives. However, the enhanced affinity towards the homologous coating conjugate (NB82-OVA) used in these assays resulted in a 1–5 orders of magnitude gap between the affinities to NB82 and the other tested compounds. Previous studies have demonstrated that using heterologous conjugates as coating antigens can significantly enhance the sensitivity of such assays towards different antigens [32]. In this case, a heterologous coating conjugate will have a lower affinity to the antibody, and consequently, the free hapten and or similar structures will be able to displace the antibodies more sensitively. Therefore, in attempts to enhance the sensitivity of our assay, we developed heterologus ELISAs towards 2 leading NB compounds, NB84 and NB124, which have shown the greatest activity in ex-vivo and in-vivo experimentations [10–14]. Three heterologous-coating conjugates have been prepared using the GTA conjugation method [31]: NB82-GTA-OVA (control), NB84-GTA-OVA and NB124-GTA-OVA. Conjugate preparation was followed by the development of 3 indirect heterologous-assays to test cross reactivity values of 3 representative semi-synthetic members. The obtained results indicated a major improvement in heterologous-antigen recognition, where I₅₀ values for NB82, NB84 and NB124 were reduced from 50±6 ng mL⁻¹ to 10±3 ng mL⁻¹, from 3.1±0.75 μg mL⁻¹ to 0.5±0.04 μg mL⁻¹ and from 12±1 μg mL⁻¹ to 1±0.12 μg mL⁻¹, respectively (Table 2). Detection limits, while using the heterologous-assays were 0.001±0.0003 μg mL⁻¹ for NB82, 0.02±0.007 μg mL⁻¹ for NB84 and 0.015±0.008 μg mL⁻¹ for NB124 (Table 2).

Table 2.

Concentrations at 50% inhibition (I₅₀) and limits of detection (I₂₀) of 3 representative members (NB82, NB84 and NB124) using the heterologous ELISAs^b.

Compound	Coating antigen	I50 (μg/mL)	Detection limit (I20) (μg/mL)
NB82	NB82-OVA	0.05±0.006	0.008±0.001
	NB82-GTA-OVA	0.05±0.009	0.007±0.001
	NB84-GTA-OVA	0.04±0.006	0.007±0.001
	NB124-GTA-OVA	0.01±0.003	0.001±0.0003
NB84	NB82-OVA	3.1±0.75	0.7±0.1
	NB82-GTA-OVA	6±0.3	0.5±0.1
	NB84-GTA-OVA	0.5±0.04	0.02±0.007
	NB124-GTA-OVA	0.8±0.03	0.03±0.009
NB124	NB82-OVA	12±1.1	3±0.7
	NB82-GTA-OVA	7±0.5	0.5±0.06
	NB84-GTA-OVA	2±0.1	0.03±0.009
	NB124-GTA-OVA	1±0.12	0.015±0.008

^bNB82-OVA conjugate dilution was 1:8,000; NB82-GTA-OVA conjugate dilution was 1:128,000; NB84-GTA-OVA conjugate dilution was 1:64,000; NB124-GTA-OVA conjugate dilution was 1:16,000; Antisera dilution (Rabbit 8245) 1:80,000. Each value represents the mean±S.E.M. of at least 3 experiments performed in duplicates.

3.4 Determination of selected NB compounds serum levels in mice

Our next challenge was to use the pre-developed assay to monitor aminoglycosides content in mice serum in-vivo. For this purpose, we initially determined the antibodies affinities and detection limits to aminoglycosides in serum. We prepared 24 serial dilutions of 3 representative members NB82, NB84 and NB124 in PBS, FBS and mice serum and quantified them by using our pre-developed ELISA. In these experiments, NB82 was used as a positive control to test serum interference with ELISA protocol, whereas NB84 and NB124 were chosen due to their high potential as suppression therapy candidates resulting from previous results [10–14] and preliminary data in various in-vitro, ex-vivo and in-vivo models. The measured I₅₀ and I₂₀ values for all three compounds tested (NB82, NB84 and NB124) in mice serum were similar to those observed in PBS solutions (Table 3). Next, we determined the recovery of these derivatives from whole mice blood samples by spiking 250 μL blood samples with 10, 25 or 50 μg mL⁻¹ of the relevant compound, followed by serum extraction. Recovery ratios were determined to be 97–102% for NB82, 92–105% for NB84 and 100–107% for NB124 (Table 4). Assays imprecision (coefficients of variation percentage) was determined lower than 6% for all tested samples.

Table 3.

Concentrations at 50% inhibition (I₅₀) and limits of detection (I₂₀) of 3 representative members (NB82, NB84 and NB124) in mice serum^c.

Compound	Matrix	I50 (μg/mL)	Detection limit (I20) (μg/mL)
NB82	PBS	0.01±0.003	0.001±0.0003
	Serum	0.008±0.001	0.001±0.0002
NB84	PBS	0.5±0.04	0.02±0.007
	Serum	0.4±0.1	0.02±0.009
NB124	PBS	1±0.12	0.015±0.008
	Serum	0.8±0.06	0.01±0.005

^cFor NB82 and NB124 detection: NB124-GTA-OVA conjugate dilution 1:16,000 was used as coating antigen; For NB84 detection: NB84-GTA-OVA conjugate dilution 1:64,000 was used as coating antigen; Antisera dilution for all experiments was 1:80,000 (Rabbit 8245). Each value represents the mean±S.E.M. of at least 3 experiments performed in duplicates.

Table 4.

Recovery of NB82, NB84 and NB124 from whole blood samples ^d.

Compound	Spiked concentration (μg/mL)	Detected concentration (μg/mL)	% Precision	% Recovery
NB82	10	10.0±0.0
	25	24.2±0.5	96	97
	50	50.8±0.8	97	102
NB84	10	9.9±0.2	96	99
	25	26.3±0.3
	50	46.1±1.5	94	92
NB124	10	10.3±0.1	98	103
	25	26.7±0.7	95	107
	50	50.0±0.1	99	100

^dEach value represents the mean±S.E.M. of at least 3 experiments performed in duplicates.

To monitor serum levels of NB84 and NB124 in-vivo, we carried out IP injections at therapeutic doses (28.5 mg/kg body weight) to mice, and collected blood samples at various time points post injection. Similar doses of NB84 (30.0 mg/kg body weight) were previously shown to enhance therapeutic effect in-vivo in Idua-W392X model mice for MPS I-H disease without causing any detectable side effects [13]. ELISA measurements of aminoglycosides content in serum samples indicated that the serum concentrations of NB84 and NB124 reached the peak levels of 121.6 μg mL⁻¹ and 140 μg mL⁻¹, respectively, within 10 min post injection, and then declined to ~10 μg mL⁻¹ and 28 μg mL⁻¹ by 60 min after injection (Fig. 5). Remarkably similar kinetics were previously reported for the aminoglycosides gentamicin and amikacin detected in mice serum [25,26]. The obtained results also correspond to the alpha phase in the 3-compartment pharmacokinetic model of aminoglycosides serum availability, of which, a fast deterioration is observed in the first 60 min after injection [33].

Figure 5.

Serum levels of NB84 and NB124 following a single dose IP injection in male mice (C57/BL, 2.5–3.5 months old), Each point represents the aminoglycoside concentration (NB84 - black parallelogram, NB124 - white triangle) in a single blood sample taken from different mice at various time points following the injection of 28 μg/gr body weight NB compound. The concentrations were determined using the pre-developed heterologous-immunoassays. Each sample was prepared at 5 different concentrations and tested in duplicates.

Due to the relatively high toxicity of aminoglycosides, monitoring the serum levels of these compounds in treated patients is highly important. According to the Physicians Desk Reference (PDR) [34], the recommended intravenous (IV) doses of gentamicin and amikacin for antibacterial treatment in patients are 6–7.5 mg/kg/day and 15 mg/kg/day, respectively. The maximum peak serum levels allowed for these compounds are 12 μg mL⁻¹ and 35 μg mL⁻¹. However, in a recent clinical trial, performed in France, testing the potential use of gentamicin for the treatment of CF patients carrying various nonsense mutations, higher gentamicin peak serum levels of 20–40 μg mL⁻¹ were allowed [35]. These values are 2- to 4-folds higher than those approved for antibacterial treatment. No toxic effects were reported using the declared doses during the 15 days trial, of which gentamicin was administrated once daily at 10 mg/kg body weight. Our previous studies have indicated that all the compounds of NB-series used in this study lack antibacterial activity and exhibit significantly lower toxicity than that of gentamicin [9–14]. However, whether the observed peak levels of 121.6 μg mL⁻¹ and 140 μg mL⁻¹, upon the administration of ~30 mg/kg body weight of NB 84 and NB 124, are within the safety doses or not for the administration in humans, is not clear yet and requires further investigation. Nevertheless, the sensitivity and versatility of the generic immunoassay developed in this study are well adequate for the detection of NB84 and NB124; lowest detection limits of 20 ng mL⁻¹ and 15 ng mL⁻¹, respectively that are far below the peak levels of these compounds observed in mice serum.

4. Summary and Conclusions

Using a common two-ring scaffold, NB82, as an immunization hapten, we were able to develop a generic antibody that was further used for the development of highly sensitive ELISAs that were used for the detection of a wide range of natural and synthetic derivatives. The unique conjugation method used in this study enabled us to selectively conjugate the NB82 scaffold to a carrier protein via the N-1 position; therefore enabling to partially control the epitops selection, directing the recognition towards a wide variety of therapeutic agents. Using heterologous-coating antigens we were able to monitor the assays sensitivity towards 3 selected representative members: NB82, NB84 and NB124 (Fig. 1). Sensitivity (I₅₀) values for the recognition of these representatives were determined to be 10±3 ng mL⁻¹, 0.5±0.04 μg mL⁻¹ and 1±0.12 μg mL⁻¹, respectively (Table 2). Limits of detection were determined to be 1±0.3 ng mL⁻¹ for NB82, 20±7 ng mL⁻¹ for NB84 and 15±8 ng mL⁻¹ for NB124 (Table 2). These values were similar to the previously reported sensitivities for the detection of natural aminoglycosides in various analytical and bioanalytical manners [29].

The developed methodology has been successfully applied for monitoring selected aminoglycosides serum levels in-vivo. Spiking experimentations demonstrated relatively high and consistent recovery ratios for all tested compounds. In-vivo serum kinetics were shown to be similar for the 2 representative compounds tested and correlated well with pre-existing information regarding the serum levels of standard aminoglycosides in-vivo [25,26]. The assays developed in this study can further be exploited for the detection of NB compounds and other aminoglycosides containing 6′-OH group at ring I in various biological derived samples. Similar methods can be applied for the generation of new and improved scaffolds for the development of a wide variety of bioassays.

Highlights.

• A broad-specific ELISA for the detection of aminoglycosides with 6′-OH was developed.

• These6′-OH aminoglycosides are best candidates for treatment of genetic diseases.

• A two-ring scaffold was selectively conjugated to a carrier protein at N-1 position.

• A generic antibody allowed highly sensitive monitoring of therapeutic agents in vivo.

• The assay will aid in further development of therapeutic compounds in clinic.

Abbreviations list

2-DOS

2-deoxy-streptamine

AHB

(S)-4-amino-2-hydroxybutanoyl

CB

carbonate buffer

CF

cystic fibrosis

CE

capillary electrophoresis

DCC

N′,N′-dicyclohexylcarbodiimide

DDW

double distilled water

DMD

Duchenne muscular dystrophy

DMF

dimethylformamide

ELISA

Enzyme-Linked Immunosorbent Assay

FBS

Fetal bovine serum

FIA

fluoroimmunoassays

GAG

glycosaminoglycan

GC

gas chromatography

GTA

glutaraldehyde

HPLC

high-performance liquid chromatography

HRP

horseradish peroxidase

TLC

thin layer chromatography

KLH

Keyhole Limpet Hemocyanin

MPS I-H

mucopolysaccharidosis type-I-Hurler

NHS

N-hydroxysuccinimide

OVA

ovalbumin

PBS

phosphate-buffered saline

PBST

PBS containing Tween-20

PMe₃

trimethylphosphine

RIA

radioimmunochemical assays

S.E.M

Standard error mean

THF

Tetrahydrofuran

TMB

3,3′,5,5′-tetramethyl benzidine