Purified Alpha-Latrotoxin from L. tredecimguttatus was obtained from Alomone Labs (Jerusalem, Israel, #LSP-130). Full venom from L. mactans was obtained from Spider Pharm (Yarnell, Arizona, USA, SKU 108V:1L0005) and from Octolab (Veracruz, Mexico). AlamarBlue HS was obtained from Invitrogen (Carlsbad, California, USA, #A50101). All other reagents and solutions were obtained from Roth (Karlsruhe, Germany) or Sigma-Aldrich (St. Louis, Missouri, USA).

The antibody selection was performed as described previously with slight adjustments (46, 47).

For panning in MTP, 4 µg of α-LTX was diluted in phosphate-buffered saline (PBS) and immobilized on High Binding 96 well plates (Corning, Costar) overnight at 4°C. Next day the plates were blocked with 330 µL of 2% Panningblock (1% skim milk powder, 1% BSA in PBS with 0.05% Tween20) for 1 h at room temperature (RT) and then washed three times with ddH₂0-T (ddH₂0 ₊ 0.05% Tween20). Prior to incubating the library on coated antigen, the library (HAL9/10 separately with 1*10¹¹ colony-forming units (cfu) each, or HAL9/10 mixed with 5*10¹⁰ cfu per library) was incubated on 2% panningblock for 1 h at RT, in order to deprive the library of unspecific/sticky phage (preclearance). Afterwards the library-mix was transferred to the coated antigen and incubated for 2 h at RT and subsequently washed 10 times. Bound phages were eluted using trypsin (10 µg/mL) at 37°C for 30 min and used for the next panning round. The phage solution was transferred to a 96 deep well plate (Greiner Bio-One), mixed with 150 µL E. coli TG1 (OD₆₀₀ = 0.5) and incubated at 37°C for 30 min, followed by 30 min incubation with 650 rpm shaking, to facilitate the phage infection of E. coli. 1 mL of 2xYT-GA (1.6% (w/v) Tryptone; 1% (w/v) Yeast extract; 0.5% (w/v) NaCl (pH 7.0), 100 mM D-Glucose, 100 µg/mL ampicillin) was added and incubated for 1 h at 37°C, 650 rpm (OD₆₀₀ = ~0.5). Afterwards, 1*10¹⁰ cfu M13K07 helper phage was added and again incubated 30 min at 37°C followed by 30 min at 37°C and 650 rpm shaking to facilitate phage infection of E. coli. The infected bacteria were pelleted at 3220xg for 10 min, supernatant was discarded and the pellet resuspended in 2xYT-AK (1.6% (w/v) Tryptone; 1% (w/v) Yeast extract; 0.5% (w/v) NaCl (pH 7.0), 100 µg/mL ampicillin, 50 µg/mL kanamycin). Phage amplification was done overnight at 30°C, 650 rpm shaking, and the amplified phage were used for the next panning round. In total, five rounds of panning were performed, while increasing washing stringency (10x round 1, 20x round 2, 30x round 3, 4 and 5) and reducing antigen availability (4 µg round 1, 2 µg round 2, 1 µg round 3, 4 and 5). In the fourth panning round, washing and amount of antigen was kept constant and the fourth and fifth rounds were conducted mainly to increase phage amplification, without increasing the selection pressure. After the fifth panning round, single clones from panning rounds three, four and five were analyzed for production of anti-α-LTX specific scFv by screening ELISA.

For panning in solution and Strep-captured panning, α-LTX was biotinylated using EZ-Link Sulfo-NHS-LC-Biotin kit (Thermo Fisher Scientific) according to manufacturer’s instructions. α-LTX was dialyzed against PBS. Similarly to panning in MTP, in a first step, libraries (1*10¹¹ cfu of HAL9/HAL10 separately) were precleared of unspecific phage by preincubating library on 2% BSA in one well of a MTP for 45 min at RT. Afterwards, a second preincubation step on magnetic Streptavidin beads (Dynabeads M-280 Streptavidin, Invitrogen) was performed in solution for 45 min at RT rotating. The supernatant containing the precleared library was separated using a magnetic stand and 100 ng biotinylated α-LTX added and rotated for 2 h at RT. Bound phage were extracted by adding Streptavidin beads and incubation was done for 30 min at RT under rotation. Unbound phages were washed 10x times with PBS supplemented with 0.05% Tween20 (PBST) by separating beads on a magnetic stand and discarding the supernatant. Afterwards bound phages were eluted and panning continued as described for panning in MTP. In total, three rounds of panning were performed.

For Strep-captured panning, two wells of an MTP were coated with 2 µg/mL Streptavidin and wells blocked with 330 µL 2% Panningblock for 1 h at RT. HAL9/10 were mixed (5*10¹⁰ cfu per library) and preincubated on Streptavidin, while α-LTX was captured on a second Streptavidin well for 1 h at RT. Precleared library mix was transferred to antigen and panning done for 2 h at RT. To reduce selection of Streptavidin-specific phage further, 5 µg Streptavidin was added to the panning well as soluble competition. After washing, panning was continued as described above. In total, five rounds of panning were performed.

Monoclonal, soluble antibody fragments (scFv) were produced in 96 well polypropylene MTPs (U96PP, Greiner Bio-One) as described before (46, 48). In brief, 180 µL of 2xYT-GA were inoculated with single colonies bearing the scFv-expressing phagemids and incubated overnight at 37°C, 800 rpm in a MTP shaker (VorTemp 56™ Shaking Incubator, LabNet, Edison, NJ, USA). Next day, 170 µL 2xYT-GA were inoculated with 10 µL of the overnight culture and grown at 37°C, 800 rpm for 2 h (approx. OD_600nm 0.5). Bacteria were pelleted at 3220xg for 10 min and supernatant was discarded. To induce expression of antibody genes, pellets were resuspended in 180 µL 2xYT-A supplemented with 50 µM of isopropyl-beta-D-1-thiogalactopyranoside (IPTG) and incubated overnight at 30°C, 800 rpm. Next day, bacteria were pelleted at 3220xg, 4°C for 20 min and supernatant was used in screening ELISA.

For the ELISA, 100 ng/well of antigen were coated in PBS overnight at 4°C in High Binding 96 well plates (Corning, Costar). The next day, plates were blocked using 2% milk powder diluted in PBST (MPBST) for 1 h at RT and plates were washed three times using ddH₂O with 0.05% Tween20. 40 µL supernatant, containing the secreted scFv, was mixed with 60 µL 2% MPBST and incubated on the antigen for 2 h at RT. Subsequently, the plates were washed again three times using ddH₂O with 0.05% Tween20. Bound scFv were detected using murine mAb 9E10 (diluted 1:50 in 2% MPBST), which recognizes the C-terminal myc-tag and a secondary goat-anti-mouse serum conjugated with horseradish peroxidase (HRP) (A0168, Sigma-Aldrich, 1:42,000 dilution in 2% MPBST), each incubated for 1 h at RT, followed by three washing steps using ddH₂O with 0.05% Tween20. Bound antibodies were visualized with tetramethylbenzidine (TMB) substrate [20 parts TMB solution A (30 mM Potassium citrate; 1% (w/v) Citric acid (pH 4.1)] and 1 part TMB solution B [10 mM TMB; 10% (v/v) Acetone; 90% (v/v) Ethanol; 80 mMH₂O₂ (30%)] were mixed). After stopping the reaction by addition of 1 N H₂SO₄, absorbance at 450 nm with a 620 nm reference was measured in an ELISA plate reader (Epoch, BioTek). Monoclonal binders were sequenced and analyzed using VBASE2 (www.vbase2.org) (49) and possible glycosylation positions in the complementarity-determining regions (CDRs) were analyzed (50).

Selected scFv fragments were produced in mammalian cell culture in Expi293F cells. For this purpose, scFv gene fragments were subcloned into feasible expression vectors. In a first step, scFv fragments were subcloned into pCSE2.7-hFc-IgG (48) using NcoI/NotI (NEB). Additionally, expression vectors for production of hIgG1 and hFab antibodies were constructed. Here, VH and VL were cloned separately into pCSEHh1c-XP (heavy chain) and pCSL3hλ-XP/pCSL3hκ (light chain lambda/kappa) for IgG (51) and into pCSE2.5-Hc-hFab.2-XP (heavy chain) and pCSLCl-hFab.2-XP/pCSLCk-hFab.2-XP (light chain lambda/kappa) for production of hFab respectively via Golden Gate Assembly using Esp3I restriction enzyme (New England Biolabs, Frankfurt, Germany). Expi293F cells were cultured in GIBCO FreeStyle F17 expression media (Thermo Fisher Scientific) supplemented with 0.1% Pluronic F68 and 8 mM L-Glutamine (PAN Biotech) at 37°C, 110 rpm and 5% CO₂. Transfection was carried out at cell densities between 1.5 – 2*10⁶ cells/mL and over 90% viability. DNA: PEI complexes were formed with 1 µg DNA/mL transfection volume (1:1 ratio of vector for IgG or Fab production) and 5 µg/mL transfection volume 40 kDa polyethylenimine (PEI) (Polysciences). Here, DNA and PEI were first diluted separately in supplemented F17 media in 5% transfection volume, then mixed thoroughly and incubated for 25 min at RT before adding into the cells. 48 h post transfection cells were fed by adding the same volume of HyClone SFM4Transfx-293 media (GE Healthcare) supplemented with 8 mM L-Glutamine. Additionally, cells were supplemented with HyClone Boost 6 (GE Healthcare) adding 10% of culture volume. One week post transfection, culture supernatant was harvested by centrifugation at 1500xg for 15 min and purified using Protein-A purification.

Purification was done as described previously (52). Small-scale production was purified with 24 well filter plates with 0.5 mL resin. Larger scale productions (above 15 mL) were purified with 1 mL Mab-Select SuRe or HiTrap Fibro PrismA columns (Cytiva) on Äkta Go or Äkta Pure (Cytiva) or using the Profinia System (BIO-RAD). His-tagged proteins were purified using HisTrap FF Crude columns (Cytiva). All purifications were done according to manufacturer’s instructions.

Samples were diluted in PBS supplemented with only Laemmli-buffer (non-reducing) or containing beta-mercaptoethanol (reducing) and boiled for 10 min at 56°C (non-reducing) or 95°C (reducing). Samples were cooled down and 1–2 µg of protein was separated with 12% SDS-PAGE at 180 V for 45–75 mins. Electrophoresis of venom samples was done by using two different separation gels, as black widow spider venom contains very small as well as larger proteins. Small proteins were separated in the lower 12% gel, while the large latrotoxins were separated in the upper 8% gel. SDS-PAGE was stained with Coomassie Brilliant Blue and destained with 10% acetic acid.

For Immunoblot analysis, separated proteins were transferred onto a polyvinylidene fluoride (PVDF) membrane using the TransBlot Turbo transfer device (BIO-RAD). Afterwards membranes were blocked with 2% MPBST for 1 h at RT. Primary antibodies were diluted to 10 µg/mL in MPBST and incubated for 1 h at RT. Binding of antibodies to venom proteins was detected using goat-anti-human IgG (Fc-specific, A0170 Sigma-Aldrich) conjugated to horse radish peroxidase (HRP) in a final dilution of 1:70,000. As positive control, polyclonal anti- α-Latrotoxin produced in rabbit (L1913, Sigma-Aldrich) diluted to 1:50 was used and binding detected using donkey anti-rabbit IgG (711–035-152, Jackson ImmunoResearch) conjugated to HRP in a final dilution of 1:20,000.

For titration ELISA, L. tredecimguttatus α-LTX or L. mactans venom was immobilized with 1 µg/mL or 2 µg/mL diluted in PBS, respectively in High binding 96 well MTP (Corning) at 4°C overnight. Next day, plates were blocked using 2% MPBST for 1 h at RT and plates washed subsequently using ddH₂0-Tween. Antibodies were titrated from 100 nM to 0.001 nM for α-LTX titration ELISA and from 310 nM to 0.0031 nM for venom titration ELISA and incubated for 1 h at RT and plates washed again. Detection of binding was done using HRP conjugated goat-anti-human IgG (Fc-specific, A0170, Sigma-Aldrich, final dilution 1:70,000) and plates were washed subsequently. TMB was used as substrate and ELISA developed for 15 min before reaction was stopped using 1 N H₂SO₄. Plates were measured using at 450 nm with 620 nm as reference wavelength.

For epitope binning, MRU44–4-A1 was used as a capture antibody in Fab format and High binding 96 well MTP (Corning) coated with 2 µg/mL diluted in PBS overnight at 4°C. Next day, plates were blocked using 330 µL 2% MPBST and incubated 1 h at RT and afterwards plates washed using ddH₂0-Tween. L. tredecimguttatus α-LTX was diluted to 1 µg/mL in 2% MPBST and incubated on capture antibody for 1 h at RT and plates subsequently washed again. Antibodies were titrated from 10 µg/mL to 0.1 µg/mL in 2% MPBST and incubated on α-LTX for 1 h at RT. After plates were washed, detection of bound IgG was done using HRP conjugated goat-anti-human IgG (Fc-specific, A0170, Sigma-Aldrich, final dilution 1:70,000) and plates washed a final time. TMB was used as a substrate and ELISA was developed for 15 min before reaction was stopped using 1 N H₂SO₄. Plates were measured using an ELISA reader (Epoch, BioTek) at 450 nm with 620 nm as reference wavelength.

For screening of α-LTX neutralization and selection of lead antibodies, a functional assay was established using the neurosecretory cell line PC-12 and alamarBlue cell viability reagent. PC-12 cells were grown in RPMI1640 medium (Capricorn Scientific) supplemented with 5% fetal bovine serum (Merck) and 10% donor horse serum (Biochrom) at 37°C and 5% CO₂. In a first step, α-LTX dependent activity on PC-12 cells inducing pathogenic effects was validated. Different assay parameters were tested (number of cells per well, time of intoxication, time of alamarBlue development, α-LTX concentration) (data not shown). Finally, a protocol was established intoxicating 20,000 cells/well with α-LTX starting at a concentration of 50 nM titrated down to 0.05 nM. Intoxication was done for 15 h at 37°C and 5% CO₂. All media were spiked with varying CaCl₂ concentrations (0, 1, 5 and 10 mM) and after intoxication, 10% (v/v) of alamarBlue was added to the culture volume. Assay was developed for 6–8 h at 37°C and 5% CO₂. α-LTX-induced pathogenic effects were determined by measuring relative fluorescent units (RFU) at 595 nm with 555 nm excitation wavelength using a Tecan Spark multimode plate reader. Mock intoxication controls, treating the cells with medium without containing any α-LTX and α-LTX controls without adding any antibody were included.

For screening of α-LTX neutralization, a mean α-LTX concentration of 4.2 nM was chosen. All media were spiked with 10 mM CaCl₂. Antibodies were prepared in 50-fold molar excess to α-LTX with 210 nM per antibody and titrated to 0.067 nM and preincubated with toxin for 1 h at RT. Afterwards 20,000 PC-12 cells/well were intoxicated with preincubation mix for 15 h at 37°C, 5% CO₂. Assay was developed by adding 10% (v/v) alamarBlue to the cells and incubating for 6–8 h at 37°C, 5% CO₂. Cell viability was determined by measuring RFU at 595 nm with 555 nm excitation using a multimode plate reader (Spark, Tecan). All antibodies were tested in duplicates. Due to the multitude of antibodies, screening was done in three individual experiments. STE90-C11 was included as unrelated isotype control. MRU44–4-A1 was included in each experiment as internal positive control. Furthermore, controls intoxicating the cells without any antibody and mock intoxicating the cells with media not containing any α-LTX were included.

Antibody affinity of MRU44–4-A1 to α-LTX was measured in different antibody formats using Bio-layer interferometry (BLI) with the Octet qKe (Fortebio/Sartorius GmbH, Göttingen, Germany). IgG and scFv-Fc were measured using ProA biosensors, while Fab format was measured using FAB2G sensor. Baseline measurement was performed in assay buffer (TBS supplemented with 1% BSA and 0.05% Tween20) for 60 s. Subsequently, antibodies were loaded onto the sensors for 180 s in assay buffer with 10 µg/mL of antibody. Measurements were done in presence or absence of 10 mM CaCl₂ to induce oligomerization of α-LTX. After establishing a stable baseline of loaded antibodies, association of α-LTX was measured for 300 s in dilution series from 158 nM to 0.5 nM. A control without α-LTX and a control with unloaded sensor was included. Following association, dissociation of α-LTX was measured for 600 s by transferring the sensors to respective assay buffer (in presence or absence of 10 mM CaCl₂). Analysis was done by subtracting recorded baseline measurements (0 nM α-LTX) and modeling of binding kinetics was done using a global 1:1 binding model.

Lyophilization was done as described in (53). For buffer formulation, 10 mM sodium phosphate buffer (pH 7.4) supplemented with 2% Trehalose was used and antibodies concentration set to >1 mg/mL. IgG were lyophilized as 250, 500 or 1000 µL vials in 11 mm crimp-glass vials. Vials were loosely closed with lyophilization caps to ensure water evaporation and prefrozen at -80°C. Lyophilization was performed using Alpha2–4 LSCplus (Martin Christ) and shelf precooled to -80°C. Primary drying was done at -50°C and vacuum of 0.1 mbar for 40 min before temperature was raised to -35°C, while vacuum was kept constant for in total 37 h. For final drying of samples, temperature was increased to 20°C and vacuum lowered to 0.08 mbar for in total 36 h. Subsequently chamber was flooded with nitrogen and lyophilization caps closed tightly with height-adjustable table and vials finally closed with metal caps. Reconstitution was done using ddH₂O using lyophilized volume and antibodies quality controlled in size exclusion chromatography and in vitro neutralization assay.

Neocortical neuronal cultures from P0 to P1 mice were prepared similarly as previously described (54). Briefly, mice were decapitated and cerebral cortices were removed, dissected and enzymatically digested with Papain (Sigma) in the presence of DNAse (Sigma), followed by mechanical dissociation and centrifugation through a cushion of 4% bovine serum albumin (Sigma). These steps were completed using Hibernate medium (ThermoFisher). Cells were then plated onto Poly-L-Lysine (Sigma) coated coverslips in 24 well plates. For each coverslip, 35,000 cells were allowed to settle in a 40 μL drop for about 30 min and then each well was filled with 500 μL growth medium: NeurobasalA/B27 (Invitrogen) supplemented with GlutaMax (0.25%, Invitrogen), Glutamine (0.25–0.5 mM, Sigma), Penicillin/Streptomycin (1:100, ThermoFisher) and heat-inactivated fetal calf serum (10%, Sigma). Medium was partially exchanged on day 3 (800 μL) and day 14 (500 μL) with fresh maintenance medium: BrainPhys (StemCell), B27 (2%, Invitrogen), GlutaMax (0.25%, Invitrogen), Penicillin/Streptomycin (1%, ThermoFisher). Cultures were maintained for up to 3 weeks at 37°C and 5% CO₂.

Postsynaptic voltage-clamp recordings were performed in mouse cortical neurons [days in vitro (DIV) 14–17] using a HEKA EPC10 amplifier (HEKA Elektronik, Lambrecht/Pfalz, Germany). Pipette solution for voltage clamp recordings contained (in mM): 130 CsMeSO₃, 10 TEA-Cl, 10 HEPES, 5 EGTA, 3 Mg-ATP, 0.3 Na-GTP, 5 Na-Phosphocreatin, 3 QX314-Cl, pH adjusted with CsOH to 7.32 and osmolarity adjusted by sucrose to 300 mOsm. Series resistance (Rs) was on average 9.4 ± 0.3 MΩ (n=32, mean ± SEM) and was compensated to a remaining Rs of 4.7 ± 0.2 MΩ. Pipettes were pulled from borosilicate glass (Science Products, Hofheim, Germany) with a DMZ Universal Electrode (Zeitz Instruments, Martinsried, Germany) with resistance of 3 - 4 MΩ.

Recordings of the holding current (I_hold) and spontaneous miniature excitatory postsynaptic currents (mEPSCs) were performed at a holding potential of -70 mV in an extracellular recording solution containing (in mM): 145 NaCl, 2.5 KCl, 1.2 MgCl₂, 2 CaCl₂, 10 HEPES, 10 glucose, pH adjusted by NaOH to 7.4. To isolate mEPSCs the extracellular solution was supplemented by 1 µM tetrodotoxin (TTX) to block action potential induced synaptic transmission and 10 µM SR95531 to block GABAA receptors. The voltage was not corrected for a calculated junction potential of 9.6 mV. Recordings were performed at RT.

After establishing a stable baseline recording period (4 minutes), preincubated extracellular solution supplemented by either 1 nM α-LTX alone or 1 nM α-LTX and 10 nM of the respective antibody was washed-in through the perfusion system and I_hold and mEPSCs were monitored up to 20 minutes after wash-in. Spontaneous mEPSC were detected with the waveform template matching algorithm of the NeuroMatic plug-in (Version 3) for Igor Pro (WaveMetrics, Lake Oswego, OR, USA; Version 9). For each cell, the mEPSC frequency was calculated for every minute and normalized to the mean mEPSCs frequency before wash in of α-LTX in the presence of the respective antibody. The onset time of the toxic effect was set as the time when the mEPSC frequency exceeded 4xSD of the mean mEPSC frequency before wash-in.

Experimenter was blind to treatment and blinding was performed by an independent investigator. Unblinding of the experimental groups was performed after analysis and disclosure of the data by the experimenter.

The differences between the effect of α-LTX in the presence of the antibodies were tested by one-way ANOVA followed by Bonferroni’s post-hoc comparisons tests. Calculations of statistical tests were performed with jamovi (https://www.jamovi.org).

In order to develop therapeutic antibodies against α-LTX, phage display panning of naïve libraries HAL9/10 was performed on L. tredecimguttatus α-LTX. Various panning approaches were employed in order to find the most efficient strategy ( Table 1 ).

As the available assays which so far used in α-LTX research were not providing the necessary combination of high-throughput compatibility, safety and sensitivity, a new assay was developed to screen the antibodies for α-LTX neutralization, α-LTX induces Ca²⁺-dependent, cytotoxic effects on the neurosecretory PC-12 cell line (55). AlamarBlue is a non-toxic, cell viability reagent, which can be used to monitor oxidation in metabolically active cells, and thereby can be used quantitatively for cell viability assays. In a first step to establish a novel α-LTX neutralization assay, the concentration-dependent activity of α-LTX on PC-12 cells using alamarBlue was verified. In preliminary experiments, the number of cells per well, Ca²⁺ concentration, time of intoxication and time of assay development were optimized (data not shown) to measure cytotoxic effects of α-LTX on PC-12 cells in a toxin- and calcium concentration-dependent manner ( Figure 1 ). A mean α-LTX concentration of 4.2 nM was chosen to ensure a 50-fold molar excess of antibody to toxin in the highest antibody concentration and to work in the lower end of the linear range of α-LTX activity curve. Always 20,000 cells per well were intoxicated for 15 h at 37°C and 5% CO2. All media were spiked with 10 mM CaCl₂ and alamarBlue developed for 6–8 h at 37°C and 5% CO₂ before RFU were determined in the Tecan Spark multimode plate reader at 555 nm excitation and 595 nm emission.

The alamarBlue/PC12 cell-based assay allowed to detect α-LTX neutralization by an antibody when it provided cell viability comparable to non-intoxicated control. An unrelated anti-SARS-CoV-2 IgG (STE90-C11) (52) was used as negative control. 45 out of 53 tested antibodies showed neutralization of α-LTX with varying efficacies, while 35 reached or exceeded viability levels of mock intoxication control ( Figure 2 ). MRU44–4-A1 showed substantially better neutralization potency than all other tested antibodies and was in the following included in each experiment as an internal positive-control. The half maximal inhibitory concentrations (IC₅₀) were determined to rank antibodies for their neutralization efficacy ( Figure 3B ). The 14 antibodies exhibiting the lowest IC₅₀ (values ranging from 0.3 to 24 nM respectively) were selected for further development (highlighted in color in Figure 2 ).

The selected lead candidates were also tested in ELISA to determine the half maximal effective concentration (EC₅₀) of antigen binding ( Figures 3A, B ). The strong neutralization efficacy of MRU44–4-A1 to α-LTX was in accordance with the sub-nanomolar EC₅₀ of 0.09 nM measured by ELISA. In general, all lead candidates showed strong binding of α-LTX with EC₅₀ ranging from 0.09 nM to 1.85 nM with a mean EC₅₀ of 0.5 nM.

To test the neutralization efficacy of the antibodies onto an α-LTX-induced increase in neurotransmitter release (56), spontaneous miniature excitatory postsynaptic currents (mEPSCs) were measured in cultured murine cortical neurons by means of whole-cell patch clamp recordings. The mEPSC frequency was calculated every minute before and during bath application of a toxin-antibody mixture consisting of α-LTX and MRU44–4-A1, MRU44–4-A2, MRU44–4-A10, the isotype control STE90-C11 and a combination of MRU44–4-A1 + MRU44–4-A2, respectively ( Figure 4A ). The mEPSC frequency was normalized to the mean mEPSC frequency before wash-in (addition of toxin:antibody mixture) and averaged across all cells for every toxin-antibody mixture, respectively. The onset time of the α-LTX-induced pathogenic effect was defined as the time point in which the measured mEPSC frequency exceeds 4xSD of the mean mEPSC frequency before wash-in of the toxin-antibody mixture ( Figure 4B ). The onset thus reflects the delay until α-LTX-induced neurotransmitter release.

The isotype control STE90-C11 showed an α-LTX-induced increase in neurotransmitter release immediately after wash-in of the toxin-control antibody mixture, confirming the fast pathogenic effect of α-LTX. In comparison, MRU44–4-A1 significantly delayed the onset of α-LTX effect up to 18 min after wash in. In three out of 9 single cell measurements, 4xSD was not reached during recording time and thereby no onset of the α-LTX effect could be determined. In these experiments, the onset time was set to maximum measurement time. MRU44–4-A2 and MRU44–4-A10 slightly delayed the onset of pathogenic α-LTX effects. The strong neutralization efficacy of MRU44–4-A1 shown in the alamarBlue based neutralization assay could be verified in cultured neurons showing a clear deceleration of the α-LTX-induced increase in neurotransmitter release. The combination of MRU44–4-A1 and MRU44–4-A2 did not improve neutralization over levels of MRU44–4-A1 (10 nM) used alone.

In general, MRU44–4-A2 and MRU44–4-A10 did not show a strong neutralization efficacy in electrophysiological recordings. To be noted, only a 10-fold molar excess of antibody to toxin was used in comparison to a 50-fold molar excess in the alamarBlue neutralization assay.

Using oligoclonal mixtures of antibodies can enhance neutralization efficacy of the individual antibodies by additive or synergistic effects and mimic the advantages of a polyclonal antibody serum, while still ensuring a defined product (48). To find antibodies which are binding to different epitopes and can be used in combination with MRU44–4-A1 in an oligoclonal antibody cocktail, this antibody was produced as human Fab. MRU44–4-A1 was chosen as lead candidate, as this antibody showed the highest affinity and neutralization efficacy (see Figures 2 , 3 ). Consequently, MRU44–4-A1 was used as capture antibody in the Fab format and 14 IgGs were titrated over captured α-LTX to analyze epitope compatibility ( Figure 5 ). MRU44–4-A2 and MRU72–2-C12 were the only antibodies binding to α-LTX in the presence of MRU44–4-A1.

Binding affinity in scFv-Fc, IgG and Fab format of MRU44–4-A1 to α-LTX was measured using biolayer-interferometry (BLI) on monomeric/dimeric toxin, as well as tetrameric toxin, by using 2 mM ethylenediaminetetraacetic acid (EDTA) or 10 mM CaCl_2, respectively. This was done to elucidate if antibodies have a preference in binding the inactive (dimeric) or active form (tetrameric) oligomeric state of the toxin. Here, the dissociation constant (K_D) was measured between 1.8 nM to 12.2 nM ( Table 2 ).

There was no clear trend in enhanced affinity for either monomeric/dimeric or tetrameric toxin. This indicates that the epitope bound by MRU44–4-A1 is accessible in monomeric and multimeric form of α-LTX and is not masked by α-LTX oligomerization. The sensorgrams of the BLI measurements are shown in Supplementary Figure S1 .

Different antibody formats of the same antibody can influence affinity of antibody-antigen interaction and thereby also influence neutralization efficacy. Furthermore, bivalent antibody formats (scFv-Fc and IgG) in contrast to monovalent formats (Fab) can lead to cross-linking of antigens and thereby cause aggregation/agglutination of the antibody-antigen complexes. The influence of antibody format on the neutralization efficacy was investigated for MRU44–4-A1 and MRU44–4-A10 in the alamarBlue/PC12 cell-based assay ( Figure 6 ). Aggregation does not seem to play a role in neutralization, indicated by matching neutralization efficacy and calculated IC₅₀ of 2xFab and IgG. In case of MRU44–4-A1, IC₅₀ of Fab and IgG was nearly twofold higher than in scFv-Fc. For MRU44–4-A10, scFv-Fc performed considerably better and switch from scFv to IgG/Fab induced a loss of affinity.

To increase shelf-life of the lead antibody candidates, they were lyophilized according to Schneider et al. (53). 8 antibodies having the highest neutralization efficacy or performing best in titration ELISA or being compatible in combination with MRU44–4-A1 were selected for further development. Samples of antibodies were stored at 4°C and -80°C. Following lyophilization and reconstitution in PBS, size exclusion chromatography (SEC) was used to check for denaturation or aggregation ( Figure 7A ). MRU44–4-A2 and MRU72–2-E12 showed a small fraction of higher molecular weight peak in SEC. All other reconstituted lyophilized antibodies were monomeric. Further, all antibodies tested after lyophilization showed a comparable neutralization efficacy to samples stored at 4°C and -80°C ( Figure 7B ).

Because of the low incidence of black widow spider envenomation, clinical development of a therapeutic antibody is only reasonable if it neutralizes various Latrodectus species. Therefore, the aim was to generate antibodies that would cross-neutralize among Latrodectus species, which was hypothesized to be likely due to the high sequence identity of Latrotoxins (57). The cross-reactivity of the lead antibodies against α-LTX of the southern black widow (Latrodectus mactans) was assessed using whole venom, as purified α-LTX from L. mactans was not available.

We tested L. mactans venom of two different manufacturers. Here, it became evident, that venom or toxin quality is highly variable and can influence quality of results. Two different lots of each manufacturer were tested and toxicity as well as EC₅₀ were matching. The venom of Octolab (VeraCruz, Mexico) was found to be roughly 200x less potent than L. mactans venom of Spider Pharm (Yarnell, USA) ( Supplementary Figure S3 ). Furthermore, L. tredecimguttatus α-LTX was found to be temperature-sensitive and after 3 months of storage at -20°C, nearly no activity could be measured in alamarBlue viability assay (data not shown).

For analysis of antibody cross-neutralization between L. mactans venom and L. tredecimguttatus α-LTX, in a first step the potency of L. mactans venom was determined in alamarBlue based cell assay ( Figure 8A ). Despite whole venom being a mixture of different proteins, it was still more potent than purified L. tredecimguttatus α-LTX, which was observable in EC₅₀ and total signal (lower RFU for saturation of pathogenic effects). After comparing potency of L. mactans whole venom to L. tredecimguttatus α-LTX, all lead IgGs were tested for their neutralization efficacy against L. mactans whole venom ( Figure 8B ). Only 2 out of 14 showed venom neutralization of L. mactans. MRU44–4-A1 did not neutralize L. mactans venom nor show binding in immunoblotting ( Supplementary Figure S2 ). These findings were confirmed by ELISA ( Figure 8C ). MRU44–4-A2 and MRU72–2-C12 showed stronger binding of whole venom compared to all other antibodies tested. The IC50 and EC50 values are given in Figure 8D . MRU44–4-A4 and MRU44–5-D9 displayed very faint binding of L. mactans venom in ELISA.