In vivo antiviral efficacy of LCTG-002, a pooled, purified human milk secretory IgA product, against SARS-CoV-2 in a murine model of COVID-19

ABSTRACT

Immunoglobulin A (IgA) is the most abundant antibody (Ab) in human mucosae, with secretory form (sIgA) being dominant and uniquely stable. sIgA is challenging to produce recombinantly but is naturally found in human milk, which could be considered a global resource for this biologic, justifying its development as a mucosal therapeutic. Presently, SARS-CoV-2 was utilized as a model mucosal pathogen, and methods were developed to efficiently extract human milk sIgA from donors who were naïve to SARS-CoV-2 or had recovered from infection that elicited high-titer anti-SARS-CoV-2 Spike sIgA in their milk (pooled to make LCTG-002). Mass spectrometry determined that proteins with a relative abundance of 1% or greater were all associated with sIgA. Western blot demonstrated that all batches consisted predominantly of sIgA. Compared to control IgA, LCTG-002 demonstrated significantly higher Spike binding (mean endpoint of 0.87 versus 5.87). LCTG-002 was capable of blocking the Spike receptor-binding domain – angiotensin-converting enzyme 2 (ACE2) interaction with significantly greater potency compared to control (mean LCTG-002 IC50 154ug/mL versus 50% inhibition not achieved for control), and exhibited significant neutralization activity against Spike-pseudotyped virus infection (mean LCTG-002 IC50 49.8ug/mL versus 114.5ug/mL for control). LCTG-002 was tested for its capacity to reduce viral lung burden in K18+hACE2 transgenic mice inoculated with SARS-CoV-2. LCTG-002 significantly reduced SARS-CoV-2 titers compared to control when administered at 0.25 mg/day or 1 mg/day, with a maximum TCID50 reduction of 4.9 logs. This innovative study demonstrates that LCTG-002 is highly pure and efficacious in vivo, supporting further development of milk-derived, polyclonal sIgA therapeutics.

KEYWORDS:

• Secretory IgA
• human milk
• mucosal infection
• SARS-CoV-2
• COVID-19

Introduction

Immunoglobulin A (IgA) is one of 5 human antibody (Ab) subclasses. Healthy humans naturally produce IgA in response to infections. Secretory IgA (sIgA) is the predominant Ab class found in human mucosal compartments (e.g. the respiratory tract, gastrointestinal and genitourinary tracts, and the oral/nasal cavity) and is distinguished from monomeric IgA by the scaffolding of IgA monomers into an end-to-end formation (typically dimeric) via an intermediary J-chain, which is further wrapped in Secretory Component (SC) as it is secreted into mucosae. SC confers protection to sIgA, stabilizing it in relatively harsh mucosal environments that can otherwise degrade biomolecules via low pH, proteases, ciliary activity, and mucus entrapment¹. This protective SC-conferred mucosal stability does not occur for IgG, which is the predominant Ab class found in the blood, or any other monomeric Ig class. sIgA primarily acts in the respiratory tract or other mucosae by recognizing pathogens (i.e., viruses, bacteria, fungi, and parasites) in an antigen-specific manner and neutralizing their activity, preventing their attachment and entry into target cells through a process called immune exclusion.² The neutralization effect has been shown to be significantly more potent using sIgA compared to monomeric IgA, likely due in part to its enhanced durability and increased avidity as a polymer.^3,4 Unique among human Ab classes, sIgA can also neutralize intracellular viruses by interfering with their replication and/or assembly.⁵ sIgA also mediates various non-neutralizing functions via its Fc domain via interactions with Fc receptors (FcRs).^6–9 Furthermore, the Fc region of sIgA is a potent activator of the alternative pathway of complement.¹⁰ The abundant glycans on sIgA also activate the lectin pathway of complement and nonspecifically contribute to pathogen clearance.^3,4,10

Respiratory pathogens must access the nostrils and/or mouth to infect airway cells, replicate, and subsequently penetrate the nasal cavity, deeper airways, and lungs. Therefore, by (a) presenting an sIgA barrier and (b) physically entrapping pathogens and particles in mucus, the nasal cavity’s mucosal layer provides two key layers of early defense against respiratory infections.

SARS-CoV-2, the causative agent of COVID-19, infects human airway cells via the receptor-binding domain (RBD) of its Spike protein binding to human receptor angiotensin-converting enzyme 2 (ACE2).¹¹ COVID-19 infection provokes mucosal immunity including the production of protective, SARS-CoV-2-specific IgA as confirmed in COVID-19 patient cohorts in multiple countries. In COVID-hospitalized patients in the UK, SARS-CoV-2-specific nasal IgA concentrations increased over baseline as early as 14 days after infection and remained elevated for up to 9 months following infection.¹² In a study of 338 triple-vaccinated Swedish health-care workers, high levels of wild-type spike-specific mucosal IgA were correlated with protection against subsequent omicron breakthrough infection.¹³ In the same vein, clonally expressed dimeric IgA from COVID-convalesced US patients were more effective at RBD binding and SARS-CoV-2 neutralization compared to monomeric IgA.¹⁴

While these clinical studies highlight the importance of dimeric sIgA in protective responses to COVID-19, they also underscore the consequences of Ab deficiency in the context of patients with common variable immune deficiency (CVID) which is characterized by deficiencies in IgA and IgG, and in some cases, IgM as well. Indeed, COVID-19 vaccines tested in a Spanish CVID cohort were found to be less efficient at inducing COVID Spike-specific serum Abs as compared to healthy controls,¹⁵ and a study in a Swiss CVID cohort found a similar outcome.¹⁶ While these studies measured serum and not mucosal samples, they emphasize the overall deficiency in CVID patients’ generation of SARS-CoV-2-neutralizing Ab responses to vaccines and underscore the potential value of sIgA replacement for these patients. sIgA is the predominant Ab class in human milk, at a concentration of over 2 g/L in the first 2 years of lactation,^17,18 making human milk the only naturally abundant source of polyclonal human sIgA. The therapeutic potential of human milk sIgA against SARS-CoV-2 was previously tested by demonstrating that milk from COVID-recovered donors contained IgA specific for the receptor-binding domain (RBD) of SARS-CoV-2 Spike protein, and the IgA was in the secretory form.¹⁹ Furthermore, human milk samples contain persistent Spike-specific sIgA titers for 12 months or more after infection.^20,21 Therefore, a medical preparation of polyclonal human sIgA, delivered directly to the nasal cavity, is a promising avenue for pre- and post-exposure prophylaxis against COVID-19.

The present study sought to characterize LCTG-002, a therapeutic candidate consisting of purified polyclonal sIgA extracted from pooled human milk according to a proprietary, validated and patented production method. This milk was obtained from lactating people previously infected with SARS-CoV-2, but not vaccinated against COVID-19, who had been shown to exhibit high titers of Spike-specific milk sIgA.²⁰ LCTG-002 and control sIgA batches were characterized for batch production consistency, purity, and stability, then assessed in vitro to demonstrate binding to SARS-CoV-2 Spike protein and neutralization of Spike binding to ACE2 receptor. Finally, LCTG-002 was delivered to the airways of ACE2 transgenic mice before and after inoculation with wild-type SARS-CoV-2 to ascertain in vivo antiviral efficacy in a prophylactic dosing regimen.

Materials and methods

Selection of human milk donors and milk procurement

Milk samples used for this study were obtained from study participants who either (1) had a SARS-CoV-2 infection confirmed by an FDA-approved COVID-19 PCR test 3–8 weeks prior to the initial milk sample collection or (2) had no history of SARS-CoV-2 infection (control samples) as described previously.²⁰

LCTG-002 purification and preparation

Frozen human milk was thawed, clarified by centrifugation at 13,000 rpm for 60 min followed by removal of the fat layer, slowly combined with ammonium sulfate under the stir bar to reach a saturation of 65%, then removed from the stir plate and stored at 4°C overnight. The resulting precipitate was centrifuged at 7,600×g for 45 min at 4°C, the supernatant was discarded, and the pellet collected and resuspended in approximately 400 mL of 1× PBS over a stir plate for 15 min. The resuspended pellet was then packed into dialysis membrane tubing (Spectra/Por molecular membrane tubing; 3.5 kDa), and the tubes were clipped and dialyzed in 10 L of 1X PBS overnight at 4°C. The dialyzed sample was centrifuged at 10,000×g for 30 min, the supernatant was collected and filtered through a 0.45um filter unit, then was loaded onto an affinity column that had been preloaded with 10 ml of CaptureSelect IgA Affinity Matrix resin, washed with five column volumes (CV) MilliQ water, and equilibrated with 5CV 1× PBS. The supernatant loading was followed by a 1× PBS wash until baseline absorbance was observed on an ÄKTA Pure Protein Purification System UV meter. IgA was then eluted with 0.1 M Glycine (pH 3.0). The fraction of interest was collected from the immediate start of a peak in UV to its decay and was immediately treated with 100ul of neutralization buffer (1 M Tris, pH 8.0) for every 1 mL of elution collected. Pooled elution fractions were dialyzed overnight in 10 L of 1× PBS using 20 mL dialysis cartridges (Sigma Aldrich Pur-A-Lyzer Mega 3500 Dialysis Kit). Finally, the dialyzed material was pooled and passed through a 0.2 micro filter aseptically under a biosafety cabinet into a sterile endotoxin-free container and then aliquoted. The final concentration of the purified IgA was measured via A280 and divided by the extinction coefficient (E0.1% of secretory IgA = 1.158) to calculate the IgA concentration in mg/mL.

LCTG-002 protein imaging

A normalized 2ug sample of LCTG-002 was mixed with an equivalent volume of 2X laemmli loading buffer containing BME. 1ug LCTG-002 samples, as well as a 8uL sample of Precision Plus Protein Unstained Standard by BioRad, were loaded onto a Mini-PROTEAN TGX Stain-Free Gel by BioRad and run with 1X Tris/Glycine SDS Buffer at 200 V for 35 min. The gel was stained with GelCode Blue Safe Protein Stain for 1 h at room temperature with gentle mixing, the stain was poured off, the gel was unstained with distilled water overnight at room temperature with gentle mixing, and gel was imaged with the Coomassie-stained program from Syngene’s G BoxMini. The image was analyzed with ImageJ to determine relative purity as follows: image was cropped to remove any white background regions, and background was subtracted starting with a rolling ball radius of 100 and adjusted as needed for lightness/darkness. Gel lanes of interest were selected, and the peak regions were selected and separated from other peaks, while keeping note of major peaks predicted to be near the corresponding molecular weight to the target protein. The area of the chosen peaks were determined via ImageJ software, and the data were transferred onto a spreadsheet to calculate relative purity percentage: ((Sum of peaks of interest)/(sum of all peaks))*100.

For the native gel western blot analysis, 0.5ug for the IgA blot and 5ug for the SC blot of LCTG-002 was loaded on a 7.5% pre-casted Gel Mini Protein gel TGX (BioRad) in native gel buffer. Gels were transferred by iBlot2 to a PVDF membrane and blocked with 5% nonfat milk for 1 h. Blotting was performed using 1:10000 anti-IgA HRP (EMD Milipore) or 1:3000 anti-SC HRP (Nordic Mubio). A standard ECL substrate was used to develop (Biorad), and blot was visualized on a BioRad Imaging System.

Mass spectrometry

Samples were analyzed as described previously.²² In brief, 10ug samples of LCTG-002 (three replicates per group) were run ~1 cm into an SDS-PAGE. Gels were stained with Coomassie blue and gel pieces excised. Proteins were digested with in-gel trypsin and resulting peptides excised and analyzed on a Thermo Scientific Eclipse mass spectrometer. Data were analyzed using Maxquant and Perseus to provide an FDR-controlled method to compare differences in protein expression between the two groups. Absolute protein amounts were determined using the label-free iBAQ component of Maxquant.

Spike binding assays

ELISA was performed to measure Spike binding as previously described, using titrated LCTG-002 and control IgA at a starting concentration of 100ug/mL.²⁰ Purified pooled IgA samples were tested in separate assays measuring IgA and secretory Abs, in which the secondary Ab used for the latter measurement was specific for free and bound SC. Half-area 96-well plates were coated with the full trimeric recombinant Spike protein.²⁰ Plates were incubated at 4°C overnight, washed in 0.1% Tween 20/PBS (PBS-T), and blocked in PBS-T/3% goat serum/0.5% milk powder for 1 h at room temperature. Samples were titrated in 1% bovine serum albumin (BSA)/PBS and added to the plate. After 2 h of incubation at room temperature, plates were washed and incubated for 1 h at room temperature with horseradish peroxidase-conjugated goat anti-human-IgA (Fisher) or goat anti-human-secretory component (MuBio) diluted in 1% BSA/PBS. Plates were developed with 3,3′,5,5′-Tetramethylbenzidine (TMB) reagent followed by 2N hydrochloric acid (HCl) and read at 450 nm on a BioTek Powerwave HT plate reader. Endpoint titers were calculated by fitting to 4-parameter non-linear regression curves of log-transformed data using a conventional cutoff value of OD = 0.2.

An ELISA-based proxy neutralization assay that detects Abs specific for the RBD of Spike that can prevent RBD-ACE2 binding was also performed, using titrated LCTG-002 and control IgA at a starting concentration of 242ug/mL, following manufacturer’s protocol (GenScript). Plates were read by Powerwave plate reader, and the percent inhibition was calculated according to the positive and negative controls from the kit as [1-(mean sample wells OD/mean control wells OD)] x 100. IC50 was calculated from 4-parameter regression of log-transformed curves using GraphPad Prism. If 50% neutralization was not achieved, samples were assigned an IC50 of 500 ug/mL. Significance was determined using unpaired Student’s t-tests.

Pseudovirus neutralization assay

A replication-competent EGFP-reporter vesicular stomatitis virus (VSV) system, rcVSV-CoV2-S, which encodes SARS-CoV-2 Spike from Wuhan-1 strain in place of VSV-G, coupled with a clonal VERO-ACE2-TMPRSS2 cell line optimized for highly efficient Spike-mediated infection, and was used to measure neutralization capacity of LCTG-002. This assay was performed essentially as described previously.²³ Briefly, 2 × 10⁴ cells per well were seeded in 96-well plates 1-day prior to use. Virus stocks were pre-mixed with 3-fold serially diluted purified IgA (200 ug/ml to 0.823 ug/ml) in DMEM + 2% FBS for 30 min at room temperature before transferring the virus/IgA mix onto target cells. At 48 h post-infection, GFP intensities were measured on a Cytation 3 imaging Reader (BioTek). Percent neutralization was calculated after background (cells only) subtraction as [1-(mean sample wells GFP/mean virus+cell control wells GFP)] x 100. IC50 was calculated as above. Significance was determined using unpaired Student’s t-tests.

Mouse studies

The K18+hACE2 mouse model (Jackson Laboratory) was chosen based on internal data and literature indicating that SARS-CoV-2 productively replicates in this mouse model expressing human ACE2.²⁴ All animal experiments were performed under Hackensack Meridian Health Center for Discovery and Innovation (CDI) IACUC approval and the animals were housed at the CDI Research Animal Facility which is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALACi). All mice were housed in individual ventilated caging (IVC) units at CDI and handled under sterile conditions for a minimum of 72 h for acclimation prior to any study start. For necropsy, mice were humanely euthanized by cervical dislocation, and nasal washes were collected. Lungs were aseptically collected and weighed. Half of the lungs were placed in 2 ml screw cap tubes and stored at −80°C. The other half were weighed, placed in a GentleMACS Octo dissociator tube containing 2.5 ml of DMEM + 2% FBS + 1× AB/AM and homogenized (modified from²⁵) 0.475 ml was transferred to a screw cap tube, and 0.025 ml of Proteinase added, and then heat inactivated for at least 60 min at 65C for qRT-PCR. 0.5 ml of untreated supernatant was collected for PFU and TCID50 assays. 0.1 ml of the nasal washes were used for these assays, which were described previously.²⁶

SARS-CoV-2 inoculum

SARS-CoV-2 virus strain USA-WA1/2020 (P4, Catalog#: NR-53873, Lot#:70039812, 9.2X105 TCID50/ml (Vero E6) Lot#:70039812) was obtained from BEI Resources. SARS-CoV-2 was gently thawed on ice and used to directly inoculate mice. Mice were anesthetized by the inhalation of vaporized isoflurane and intranasally infected with 50 μl of the virus stock to a final infection dose of 4.6 × 10⁴ TCID50.

Administration of test materials to mice

On each day of dosing, LCTG-002 was thawed to room temp and 0.05 ml of LCTG-002 or PBS control was dosed into the mouse nares (0.025 ml per naris) intranasally in the morning and 0.05 ml intratracheally in the evening as outlined in Table 1. Mice were anesthetized by inhalation of vaporized isoflurane for LCTG-002 administration. For Nirmatrelvir administration, a 300 mg pill was pulverized using a sterile mortar and pestle then suspended in sterile water. Nirmatrelvir suspension was orally administered at 250 mg/kg in 10 ml/kg using a tuberculin syringe and a 20 g popper feeding needle.

Table 1. Experimental design for K18+hACE2 mouse studies.

Group	Animals	Test Article	Dose/Route	Frequency	Duration of Infection
1	3 males + 3 females	0.25 mg control IgA	0.125 mg in 50uL (25uL per naris) IN (am); 0.125 mg in 50uL IT (pm)	24 h prior to inoculation*, then daily to day 2 (4 doses total)	3 days
2		0.25 mg LCTG-002
3		1 mg LCTG-002	0.5 mg in 50uL (25uL per naris) IN (am); 0.5 mg in 50uL IT (pm)
4		Nirmatrelvir	250 mg/kg in 10 ml/kg volume PO
5		PBS	50uL (25uL per naris) IN (am); 50uL IT (pm)	24 h prior to, and on the day of inoculation* (2 doses total)	1 day
6		0.25 mg LCTG-002	0.125 mg in 50uL (25uL per naris) IN (am); 0.125 mg in 50uL IT (pm)

*Mice were inoculated with 4.6 × 10⁴ TCID50 SARS-CoV-2.

IN: intranasal; IT: intratracheal.

PO: oral administration.

RNA extraction and RT-PCR

Total RNA isolation of Proteinase K and heat inactivated lung homogenate supernatants was performed using the Qiagen Qiacube HT automated mid- to high-throughput nucleic acid purification instrument with the QIAamp 96 Virus QIAcube HT Kit. For SARS CoV-2, qRT-PCR was performed on the samples using the E gene (from Charité/Berlin (WHO) protocol primer and probe panel 6) and RNase P gene (CDC kit). The primers and probe sequences are listed in Supplemental Table 1.

Results

Characterization of pooled, purified milk IgA

IgA was extracted from pooled human milk samples from either pre-pandemic, COVID-naive lactating people as a negative control (control IgA) or COVID-recovered lactating people (LCTG-002) based on methods disclosed in patent US 11,124,560. Samples were run on SDS-PAGE, stained with Coomassie blue, and imaged. A pattern of three protein bands was consistently visible in both batches (Figure 1) corresponding from top to bottom to the known molecular weights of SC, heavy alpha chain, and light chain. Two additional batches were prepared from pooled human milk and imaged and demonstrated identical protein bands (data not shown). Protein purity was assessed across all four batches by SDS-PAGE imaging and ranged from 93.2% to 99.2% across batches (data not shown). Samples from the control IgA and LCTG-002 batches were also evaluated by western blot to compare sIgA content. The blot was immunostained for human IgA, and the signal was identical in both batches (Figure 2, left panel). A separate blot was immunostained for human SC and the signal was also identical in both batches (Figure 2, right panel), confirming that both batches had similar human IgA content which was predominantly in the secretory (sIgA) form. The upper smear in the left panel is indicative of differently glycosylated sIgA isoforms. The lower band near 130 kDa likely represents IgA monomers (left panel); monomers are not bound by SC, which explains the absence of a corresponding band in the right panel.

Figure 1. Imaging and purity assessment of human ab extracts.

sIgA was purified from pooled human milk samples from donors that were either COVID-naive (Control IgA) or COVID-recovered (LCTG-002), aliquots were run on SDS-PAGE, then stained with Coomassie blue to visualize sIgA protein bands (right lanes) relative to a protein marker (left lane).

Figure 2. Western blot analysis of human milk-derived IgA.

Aliquots of pooled, purified IgA from human milk obtained from COVID-naive (Control) or COVID-recovered (LCTG-002) individuals were assessed by native gel western blots on a 7.5% pre-casted Gel Mini Protein gel TGX (BioRad) in native gel buffer. Gels were transferred by iBlot2 to a PVDF membrane and blocked with 5% nonfat milk for 1 h. Blotting was performed using 1:10000 anti-IgA HRP (EMD Milipore) or 1:3000 anti-SC HRP (Nordic Mubio). A standard ECL substrate was used to develop (Biorad), and blot was visualized on a BioRad Imaging System.

To quantitatively assess relative protein abundance, mass spectrometry was performed on the control IgA and LCTG-002 batches, and data were analyzed using the search engine MaxQuant with label-free quantification to generate iBAQ values. Proteins with a relative abundance of 1.0% or greater were plotted in descending order. These proteins were all associated with sIgA, specifically: immunoglobulins, SC (which is a cleavage product of Polymeric immunoglobulin receptor (PIGR)), and J chain (JCHAIN) (Figure 3, Panel A). None of the proteins exhibited statistically significant differences between batches, establishing that protein composition was comparable.

Figure 3. Mass spectrometry characterization of ab preparations.

Relative protein abundance in Control IgA and LCTG-002 batches was calculated from mass spectrometry data by dividing each protein’s individual iBAQ value by the total iBAQ sample value. An abundance of 1.0% was used as the lower threshold. Averages and standard deviations were calculated from triplicate samples and plotted (Panel A); note that the error bars were plotted but are too small to be visible at this scale. Peptides from Control IgA and LCTG-002 batches were also plotted in order of descending spectral count (Y-axis) with a spectral count of 50 used as an arbitrary lower threshold; averages and standard deviations were calculated from triplicate samples and plotted (Panel B).

An additional mass spec-based assessment of individual protein composition between control IgA and LCTG-002 batches was performed by using X!Tandem semi-quantitative peptide analysis of protein spectral counts. These are plotted in descending order (Figure 3, Panel B). Twelve out of 16 proteins were associated with sIgA and these proteins demonstrated similar spectral counts between batches.

Taken together, milk IgA was consistent across control IgA and LCTG-002 batches despite differing in various categories: (1) number of pooled donors per batch, (2) date of milk collection, (3) identity of the donors, (4) COVID-19 infection status, and (5) date of processing. This establishes that the purification methods are reproducible and the batches are highly pure and consistent with respect to protein composition.

In vitro assessment of pooled, purified milk IgA function

The binding specificity of LCTG-002 against SARS-CoV-2 Spike was assessed by ELISA, and binding curves were used to determine endpoint titers. Using a secondary Ab specific for human IgA, Spike-specific IgA binding endpoint was found to be significantly greater for LCTG-002 compared to control IgA (mean endpoint of 0.87 versus 5.87; p = .0133; Figure 4). Using a secondary Ab specific for human SC to confirm that the Spike-specific human IgA was largely in secretory form (sIgA), it was similarly found that LCTG-002 exhibited significantly higher Spike-specific binding compared to control IgA (mean endpoint of 1.82 versus 6.90; p = .0032; Figure 4).

Figure 4. LCTG-002 binding to SARS-CoV-2 Spike protein.

LCTG-002 (blue bars) or Control IgA (purple bars) were tested at 100 ug/mL for binding to Spike protein via ELISA with secondary Ab detecting human IgA (a) or human Secretory Component (b) which were quantified by OD450 (Y-axis). (c) Endpoint titers were calculated by standard methods. Samples were tested in triplicate, and error bars denote standard error of the mean (SEM).

To further probe the biological characteristics of LCTG-002, an ELISA-based proxy neutralization assay was used to measure LCTG-002’s ability to compete with the interaction between ACE2 receptor and the Spike RBD. IC50 was calculated from competition-binding curves, and it was shown that the inhibition activity of RBD-ACE2 binding by LCTG-002 was significantly greater compared to control (mean IC50 of 154ug/mL for LCTG-002; 50% inhibition not achieved for control; p = .0002; Figure 5a,b).

Figure 5. In-vitro neutralization function of LCTG-002.

(a) LCTG-002 inhibition of Spike-RBD interaction with hACE2. LCTG-002 (blue circles) or Control IgA (purple triangles) were tested at a range of dilutions (X-axis) for their ability to compete with SARS-CoV-2 Spike protein RBD interaction with hACE2 receptor (Y-axis). (b) IC50 values were calculated from 4-parameter regression of log-transformed data. (c) An EGFP-reporter vesicular stomatitis virus (VSV)-based neutralization assay was used to measure the neutralization capacity of LCTG-002. (d) IC50 values were calculated as above. Samples were tested in duplicate, and error bars denote SEM.

Additionally, a SARS-CoV-2 pseudovirus cell-based neutralization assay was also performed using VERO-ACE2-TMPRSS2 cells and a Spike-pseudotyped VSV as described previously.²³ Titrated LCTG-002 and control IgA were used to generate neutralization curves, and IC50 was calculated. It was found that LCTG-002 exhibited significantly greater neutralization activity compared to control (mean IC50 of 49.8ug/mL versus 114.5ug/mL; p = .0067; Figure 5c,d).

Taken together, these data demonstrate that LCTG-002 extracted from COVID-19-recovered milk donors is predominantly in secretory form (sIgA), exhibits significantly greater SARS-CoV-2 Spike binding specificity as compared to control IgA, and blocks the SARS-CoV-2 Spike mediating cell entry via the ACE2 receptor.

In vivo assessment of pooled, purified milk IgA efficacy against SARS-CoV-2

In order to test the potential antiviral efficacy of LCTG-002 against SARS-CoV-2 infection, K18+hACE2 transgenic mice were selected for use, based on literature indicating that SARS-CoV-2 productively replicates in this mouse model that expresses human ACE2 and models clinical manifestations of COVID-19 including severe lung disease, thrombosis, and vasculitis.²⁴ All mice were bright, alert, and responsive to manual restraint throughout the study. Mean percent weight loss was not more than 10% for any study group (Sup. Figure S1).

PBS, control IgA, or LCTG-002 were administered both intranasally and intratracheally to model IgA deposition across the nasal cavity, airways, and lungs, which are targets of SARS-CoV-2 infection in humans. Administration occurred 1 day before, the day of, and each day after intranasal inoculation with an infectious dose of 4.6 × 10⁴ TCID50 SARS-CoV-2, as outlined in Table 1. A positive control group of mice received Nirmatrelvir, the SARS-CoV-2 protease inhibitor component of Paxlovid, by the oral route, as this compound has been demonstrated previously to inhibit SARS-CoV-2 replication in the K18+hACE2 transgenic mouse model.^27,28

Viral titers were measured at the end of the study in the lungs 72-h post inoculation via TCID50, plaque forming units (PFU), and qRT-PCR assay. Mice receiving 250 ug/day of 2.5 mg/mL LCTG-002 exhibited statistically significant reductions in viral titer as measured by all three assays compared to mice receiving 250 ug/day of control IgA (p = .02–p < .0001; Figure 6, Panel A). Mice receiving 1 mg/day of 10 mg/mL LCTG-002 exhibited even greater reductions in viral titer across all three assays (p < .0001), notably exhibiting a 4.9-log reduction in TCID50. Compared to control, this group (1 mg/day of LCTG-002) exhibited significantly less weight loss (6.6% versus 0% mean loss by day 3; p = .019), as did the Nirmatrelvir control treatment (6.6% versus 0.6% mean loss by day 3; p = .033) (Sup. Figure 1). Nirmatrelvir treatment caused significant reductions in viral titer by TCID50 and PFU assays but not qRT-PCR, as similarly observed in prior unpublished studies using the same dose and route.

Figure 6. In vivo antiviral efficacy of LCTG-002 against SARS-CoV-2.

Mouse lungs were collected and homogenized 72 h (a) or 24 h (b) after inoculation with 4.6 × 10⁴ TCID50 SARS-CoV-2, following a regimen of PBS, Control IgA, LCTG-002, or Nirmatrelvir (see Table 1) to measure viral burden via TCID50, PFU, and qRT-PCR assays. Viral titer is plotted in log scale; P values are denoted by asterisks as follows: *<.05, **<.01, ***<.001, ****<.0001.

An additional mouse study was performed measuring viral titers in nasal washes and lung homogenates 24 h after SARS-CoV-2 inoculation to determine whether prophylactic administration of LCTG-002 inhibited nasal and/or lung replication of SARS-CoV-2 in an earlier stage of infection. Compared to mice receiving PBS, mice receiving 250ug/day LCTG-002 exhibited statistically significant reductions in viral titers in lung across TCID50, PFU, and qRT-PCR assays (p = .005–p < .0001; Figure 6, Panel B). However, no significant reductions were observed in the nasal washes of mice receiving 250ug/day LCTG-002 compared to PBS treated mice (data not shown). No difference in weight loss was observed (Sup. Figure S1).

Discussion

LCTG-002, consisting of sIgA extracted from pooled human milk donated by lactating people that had recovered from COVID-19, was tested for activity against SARS-CoV-2 using in vitro and in vivo assays. The data presented herein support our hypothesis that pooled, polyclonal human milk sIgA is capable of both recognizing and neutralizing SARS-CoV-2. LCTG-002 exhibits specificity for SARS-CoV-2 Spike (Figures 4–6), consistent with the aforementioned studies in COVID-19 patient cohorts in the UK,¹² Sweden,¹³ and the US.¹⁴ This also establishes a proof of concept for the activity of pooled human milk sIgA against a multitude of other mucosal pathogens, as confirmed through our unpublished studies testing milk sIgA activity against various additional respiratory and enteric pathogens. The polyclonal, pooled nature of human sIgA in LCTG-002 is a critical aspect of translational relevance because polyclonal Ab comprises activity against a broad array of epitopes from numerous pathogens compared to monoclonal Ab. Conceptually, this should also result in a greater likelihood of cross-reactivity against emerging variants compared to monoclonal Ab.

Current recombinant Ab programs do not capture either (a) the polyclonality or (b) the complex attributes of secretory IgA. Recombinant sIgA production requires the co-expression of four transcriptional units encoding the heavy chain, light chain, j chain, and SC, and though there are several methods to produce sIgA, yields are relatively low due to its polymeric nature and likelihood that various undesirable forms are likely.^29–32 Recombinant sIgA has been produced experimentally using bacterial, fungal, insect, plant, mammalian, and transgenic expression systems.^29,33–35 However, due to the heavy and complex glycosylation of sIgA, systems that improve yield fail to produce the proper glycan patters, introducing immunogenicity concerns as well as significant differences in sIgA function.^34,36,37 Additional challenges include capture material selection and characterization of multimers and aggregates as recently reviewed.³⁸ The many difficulties in recombinant production of secretory IgA underscore the practical and translational value in extracting naturally occurring human milk IgA which is abundant, polyclonal, and primarily in the secretory IgA format.

Of great significance in the present study is our landmark in vivo therapeutic efficacy testing of LCTG-002 against SARS-CoV-2 infection. LCTG-002 administered daily to K18+hACE2 transgenic mice both intranasally and intratracheally to model deposition across the respiratory tract prior to and after SARS-CoV-2 infection demonstrated significant antiviral efficacy (Figure 6). LCTG-002 dosed at 250ug/day was effective in the lungs at two time points (24 h and 72 h post inoculation), suggesting it is capable of durable suppression of SARS-CoV-2 replication. As expected, the higher dose (1 mg/day) was more effective than the lower dose at suppressing lung viral titers (Figure 6).

IgA purity and composition were consistent across control IgA and LCTG-002 batches despite various differences in donors, dates of collection, COVID-19 infection status, etc. This underscores that while each donor and each milk sample are immunologically unique, the general parameters of purified milk Ab preparations such as concentration and composition are similar across donors as measured in Figures 1–3. As expected from a previous study,²⁰ the IgA was predominantly in secretory (sIgA) form, as determined in Figures 2, 4 and 6. The consistency in IgA purity across batches ensures that analysis of LCTG-002 activity was not impacted by contaminants, and that activity against SARS-CoV-2 was driven by IgA and not other immunomodulatory human milk components. This study demonstrated the activity of SARS-CoV-2-specific sIgA Abs across different assays but did not measure their relative titer within or across batches. This is because a naturally sourced, pooled polyclonal Ab preparation would not be completely identical from batch to batch, as is the case with intravenous immunoglobulin (IVIG). IVIG is an approved product that replaces immunoglobulins in patients with hypogammaglobulinemia (such as CVID) based on a total immunoglobulin replacement dose and not on pathogen-specific Ab titers.³⁹ As it relates to SARS-CoV-2, the utility of the proposed approach is in automatically capturing the natural evolution of Abs in response to the emergence of COVID-19 variants as demonstrated by activity of recently collected human milk sIgA against newer variants (manuscript in preparation); this beneficial phenomenon would equally apply to other mutating pathogens such as influenza.

Convalescent plasma containing polyclonal Abs from COVID-19-recovered donors was employed as a therapeutic approach for patients hospitalized with COVID-19, particularly during the pre-vaccine pandemic period; notably, plasma contains mostly monomeric IgG. Convalescent plasma treatment was granted Emergency Use Authorization (EUA) by FDA on August 23, 2020. However, convalescent plasma from patients infected with SARS-CoV-2 had diminished capacity for neutralizing COVID variant B.1.351 relative to wild-type virus.⁴⁰ Similarly, plasma samples from another cohort of convalesced COVID-19 patients showed a substantial diminishment in neutralization against Omicron as compared to wild-type virus.⁴¹ These outcomes reflect the antigenic drift evident in SARS-CoV-2 variants, especially in the receptor-binding site (RBS).⁴² For this reason, the EUA was revised to “limit authorization to the use of COVID-19 convalescent plasma with high titers of anti-SARS-CoV-2 Abs for the treatment of COVID-19 in patients with immunosuppressive disease or receiving immunosuppressive treatment in either the outpatient or inpatient setting.”⁴³ However, the clinical benefit of convalescent plasma in this immunosuppressed cohort remains inconclusive, and systemic infusion therapy of primarily IgG may continue to demonstrate only limited efficacy against an evolving mucosal infection such as COVID-19, notwithstanding the requirement for specialized personnel for administration of infusion/injection IgG products. Additionally, it is evident that unlike milk sIgA, specific IgG Ab titers in blood against SARS- CoV-2 wane rapidly over time, limiting the potential donor pool for convalescent plasma in general.⁴⁴ Similarly, a series of injected or infused COVID-19 monoclonal Ab products (bamlanivimab plus etesevimab, casirivimab plus imdevimab, sotrovimab, bebtelovimab, and tixagevimab plus cilgavimab (Evusheld)) were developed against earlier strains of SARS-CoV-2 but have had their Emergency Use Authorizations (EUA) withdrawn because they were not expected to provide significant protection against emerging Omicron variants.⁴³

The aforementioned approaches leverage IgG Abs as therapeutics for systemic delivery, which is an inefficient method to deliver Abs to the airways in patients with respiratory viral infections such as COVID-19 due to the poor transportation of macromolecules across the lung epithelium.⁴⁵ This highlights the inherent limitation of recent recombinant monoclonal IgG Ab development programs and underscores the clinical potential of an immuno-adaptive source of evolving SARS-CoV-2-specific IgA, such as human milk, for direct deposition of IgA into the nasal cavity.

Currently, there are no approved products based primarily on human sIgA, despite its favorable mucosal stability profile. A hurdle to the development of recombinant sIgA is the requirement for macromolecular assembly involving multimeric IgA, J-chain, and Secretory Component, which introduces much more complexity than recombinant monomeric IgG development programs, as described earlier.³⁸ The challenges associated with recombinant sIgA emphasize the potential benefit of developing naturally occurring, pooled, polyclonal sIgA from human milk which is a globally and consistently available resource.

The clinical applicability of polyclonal Abs derived from pooled human material is supported by the well-tolerated usage of IVIG³⁸ which was approved for intravenous delivery in the US in 1980. Of particular relevance, an IVIG development candidate, IgAbulin, was a plasma-derived preparation consisting mainly of IgA (albeit primarily in the monomeric form) that was delivered intranasally to children and demonstrated reduction in rhinitis.⁴⁶ However, an important distinction with IVIG is in the procurement costs, supply shortages, and invasiveness to donors associated with plasma products, which are circumvented by our proposed milk-derived polyclonal Ab biologic designed for intranasal administration.

To support the clinical development of LCTG-002, the authors have established the following elements to meet Good Manufacturing Practices (GMP) standards: a human milk supply network, a Chemistry, Manufacturing, and Controls (CMC) strategy to ensure batch quality and consistency that leverages methods approved for IVIG manufacturing, and a milk pooling schematic. Based on the in vivo efficacy results presented herein, future studies will assess nasal formulation strategies for LCTG-002 which will inform nonclinical toxicology studies to be performed in support of clinical trials. The clinical trial design may contemplate pre- and post-exposure prophylaxis against COVID-19 in adults and adolescents with moderate-to-severe immune compromise due to a medical condition or immunosuppressive medications.

sIgA delivered to the nasal cavity and/or lungs would not be expected to bypass the respiratory epithelium into the systemic circulation, instead favoring local deposition.⁴⁵ In support of this, a SARS-CoV-2-neutralizing mAb delivered intranasally into mice resulted in a thirtyfold lower concentration of Ab in serum and higher lung lavage levels as compared to the same dose delivered intravenously.⁴⁷ Our preliminary analysis of human sIgA bioavailability in rodents is consistent with these observations and supports a daily dosing regimen of up to twice daily to maintain therapeutic levels (manuscript in preparation). In the context of respiratory mucosal infections like COVID-19, deposition and retention of a therapeutic such as LCTG-002 in the local mucosa is highly desirable.

This study establishes a promising proof-of-concept for LCTG-002 efficacy as a pre- and post-exposure prophylactic treatment for COVID-19, yet it highlights translational limitations. The ratio of infectious particles of SARS-CoV-2 per animal used in this study (4.6 × 10⁴ TCID50 per mouse) does not translate to the minimum infectious dose per human, which may number only in the hundreds of viral particles.⁴⁸ This is to be expected, given that mice are not the primary host for SARS-CoV-2 and a relatively high inoculum is therefore needed to model viral replication and pathogenesis. Consequently, LCTG-002 doses administered to mice in this study are not reflective of anticipated human dosing. Furthermore, clearance and biodistribution of airway-delivered exogenous polyclonal Ab in mice, which are obligate nose breathers, are not strongly representative of clearance in humans. Nevertheless, it is encouraging that daily administration of LCTG-002 to mice via intranasal and intratracheal routes was highly efficacious and well tolerated, and future animal studies will assess LCTG-002 efficacy using various dosing and route regimens.

As Spike-specific sIgA titers are detectable in human milk for 12 months or more after SARS-CoV-2 infection,^20,21 with COVID-19 vaccination prior to or following infection boosting this response, human milk represents a sustainable and durable source of protective Ab against COVID-19 with an enormous potential donor pool globally. Unlike monomeric Ab, sIgA is highly stable and resistant to degradation in all relatively harsh mucosal compartments and fluids.^49,50 This study establishes a novel proof of concept for airway administration of human milk-derived, polyclonal, pooled sIgA as a well-tolerated and repeatable prophylactic antiviral treatment, with broader applicability beyond SARS-CoV-2.

Author contributions

VM designed experiments, contributed to data analysis, and wrote the manuscript draft. RM designed experiments, contributed to data analysis, and revised the manuscript. NA, VS, AF, CD, and XY performed experiments and contributed to data analysis. SP and DP designed experiments and contributed to data analysis. RP designed experiments, oversaw data collection, provided and oversaw analysis of milk samples, contributed to data analysis, and revised the manuscript. All authors gave final approval of the manuscript.

Acknowledgments

We acknowledge the support of the Bioexpression and Fermentation Facility (BFF) in the Department of Biochemistry and Molecular Biology at the University of Georgia for protein purification and characterization. We acknowledge the support of the Biological Mass Spectrometry Facility at the Center for Advanced Biotechnology and Medicine at Rutgers, The State University of New Jersey, for mass spectrometry of IgA samples. Finally, we acknowledge the following personnel at Hackensack Meridian Health Center for Discovery and Innovation for their contributions to mouse efficacy studies: Enriko Dolgov MD, Alberto Rojas-Triana MS, Camila Mendez Romeo DVM, Taylor Tillery BS, and Kira Goldgirsh MS. As always, we thank our study participants for their generous milk donations. This work was supported by the NIH/NIAID (R01AI158214); the Icahn School of Medicine at Mount Sinai (Distinguished Scholar Award) and the Catalyst Research & Development Voucher Pilot Program administered by the New Jersey Commission on Science, Innovation and Technology, and Lactiga US, Inc.

Supplemental data

Supplemental data for this article can be accessed on the publisher’s website at https://doi.org/10.1080/21645515.2024.2303226

Disclosure statement

VM and RM are shareholders in Lactiga. VM and RM are authors on a patent filing related to the murine studies described. RP is a paid scientific advisor for Lactiga.

Additional information

Funding

The work was supported by the Icahn School of Medicine at Mount Sinai National Institutes of Health New Jersey Commission on Science, Innovation and Technology Lactiga.