Development of a dual antigen lateral flow immunoassay for detecting Yersinia pestis.

BACKGROUND: Yersinia pestis is the causative agent of plague, a zoonosis associated with small mammals. Plague is a severe disease, especially in the pneumonic and septicemic forms, where fatality rates approach 100% if left untreated. The bacterium is primarily transmitted via flea bite or through direct contact with an infected host. The 2017 plague outbreak in Madagascar resulted in more than 2,400 cases and was highlighted by an increased number of pneumonic infections. Standard diagnostics for plague include laboratory-based assays such as bacterial culture and serology, which are inadequate for administering immediate patient care for pneumonic and septicemic plague. PRINCIPAL
FINDINGS: The goal of this study was to develop a sensitive rapid plague prototype that can detect all virulent strains of Y. pestis. Monoclonal antibodies (mAbs) were produced against two Y. pestis antigens, low-calcium response V (LcrV) and capsular fraction-1 (F1), and prototype lateral flow immunoassays (LFI) and enzyme-linked immunosorbent assays (ELISA) were constructed. The LFIs developed for the detection of LcrV and F1 had limits of detection (LOD) of roughly 1-2 ng/mL in surrogate clinical samples (antigens spiked into normal human sera). The optimized antigen-capture ELISAs produced LODs of 74 pg/mL for LcrV and 61 pg/mL for F1 when these antigens were spiked into buffer. A dual antigen LFI prototype comprised of two test lines was evaluated for the detection of both antigens in Y. pestis lysates. The dual format was also evaluated for specificity using a small panel of clinical near-neighbors and other Tier 1 bacterial Select Agents.
CONCLUSIONS: LcrV is expressed by all virulent Y. pestis strains, but homologs produced by other Yersinia species can confound assay specificity. F1 is specific to Y. pestis but is not expressed by all virulent strains. Utilizing highly reactive mAbs, a dual-antigen detection (multiplexed) LFI was developed to capitalize on the diagnostic strengths of each target.

Background

Plague is a febrile illness caused by Yersinia pestis, a Gram-negative, nonmotile coccobacillus. The bacterium was responsible for the Black Death, which devastated over a third of Europe’s population between 1347–1353 [1]. The Centers for Disease Control and Prevention classifies Y. pestis as a Tier 1 Select Agent due to its infectious nature, high mortality rates, threat to public health, and potential as a biothreat. The spread of Y. pestis is facilitated by small mammals and insect vectors. The bacterium is transmitted to humans through flea bites, contact with animal excretions, or inhalation of aerosolized droplets. The different routes of infection lead to three forms of plague: bubonic, pneumonic, and septicemic. Bubonic plague is easily diagnosed by the inflammation of lymph nodes resulting in the formation of painful swellings called buboes. Bubonic plague is the least fatal of the three forms, with a 40–70% case fatality rate (CFR) when left untreated; however, bubonic plague may develop into more serious forms of the infection [2]. Pneumonic and septicemic plague present with nonspecific flu-like symptoms, leading to death in as few as three days post-exposure [3]. The CFR for pneumonic and septicemic infections approach 100% when left untreated [2]. Time is a critical factor for treating plague as effective therapeutics must be administered within 20 hours from the onset of symptoms to ensure the best chance for patient survival [4].

Y. pestis is a zoonotic bacterium found in all geographical regions besides Oceania, with Madagascar and the Democratic Republic of Congo being primary hot spots for annual plague outbreaks [5,6]. Madagascar had accounted for 74% of all cases reported to the World Health Organization between 2010 and 2015 [7]. During this period, 200–700 cases were reported annually, mainly in the form of bubonic plague which is rarely transmissible human-to-human [8,9]. During the 2017 plague season, Madagascar reported a total of 2,417 cases of plague with a CFR of 8.6% [10]. Identification of the initial cluster of infections allowed for a proper response to prevent a larger epidemic [11]. This outbreak not only marked an increase in overall cases, but more importantly, an increased percentage of pneumonic infections (70% of total infections) [8,11]. The increase in pneumonic infections may, in part, be attributed to human-to-human transmission that occurred via infectious droplets [12-14]. This along with high mortality rates of pneumonic infections warrant the need for development of additional countermeasures for plague.

The gold standard for diagnosing plague is bacterial culture [15]. Patient serology can be used to indicate if an individual was infected with Y. pestis; however, this method is hindered by its capacity to detect active infections as an antibody response is delayed upon exposure and a positive result may be due to a previous exposure [15-17]. These methods can be time consuming and require specialized laboratories and trained technicians. Advancements in rapid diagnostic tests (RDT) such as LFIs have made the detection of Y. pestis more feasible in low-technology settings and are instrumental in controlling plague outbreaks in endemic regions. A comparative study on various diagnostic methods has deemed an LFI to be an ideal platform for diagnosing plague [18]. LFIs are rapid, membrane-based immunoassays using antibodies linked to gold-nanoparticles for visual detection. Temperature stable reagents and user-friendly protocols make LFIs ideal for resource limited settings. The current LFI used in Madagascar detects the F1 antigen and has a sensitivity range of 25–100% and a specificity range of 59–79% when compared to bacterial culture and PCR [19-23]. Of concern, however, is that F1- strains, mutants that do not express the F1 pilus structure, have been identified in clinical as well as laboratory settings [19,24,25]. In the present study, we describe the initial evaluation of a prototype multiplexed RDT that allows for detection of two Y. pestis antigens.

Y. pestis is a facultative anaerobe equipped with several virulence factors to allow survival within macrophages and in extracellular spaces [26]. Genes encoding virulence factors are dispersed throughout the genome, and are also encoded on three plasmids (pCD1, pPCP1, pMT1) [27]. Current LFIs used to diagnose plague detect F1, a protein encoded by the caf1 gene on pMT1, a plasmid unique to Y. pestis [18,25,28]. The 15.5 kDa F1 antigen polymerizes to form a filament capsule surrounding the bacterium, protecting it from phagocytosis [29]. Expression of F1 is temperature-induced at >33°C [28]. The antigen is known to be shed during infection, making it a viable candidate biomarker for diagnosing plague [30-32]. Though attenuated in its ability to cause bubonic plague, F1- isolates remain highly virulent by the inhalation route leading to pneumonic infections [25,33].

The pCD1 plasmid carries genes encoding a type-III secretion system (T3SS) and its related effector proteins [34]. The pCD1 plasmid is found in clinically relevant neighbors Y. pseudotuberculosis and Y. enterocolitica [35]. The T3SS is associated with a low-calcium response crucial for pathogenicity of Yersinia species [34]. Low-calcium response V (LcrV) is a multifunctional protein serving as the needle tip of the T3SS to translocate Yersinia outer proteins into host cells [36]. LcrV is also translocated during the process and suppresses the host inflammatory response by upregulating interleukin 10 via Toll-like receptor 2 [37]. Mutations in lcrV lead to avirulence in mice due to the inability to translocate effector proteins [38,39]. Though LcrV is displayed on the surface of the bacterium and translocated via the T3SS, the antigen has also been reported to be shed into growth media in vitro [40]. A study conducted using a murine model of infection indicated that LcrV is detectable in serum and bronchoalveolar lavage fluid (BALF) in mice displaying symptoms of bubonic and pneumonic plague [41]. Since LcrV is essential for pathogenicity and appears to be shed during infection, it may serve as an important alternative biomarker for the diagnosis of plague [42-44].

While Y. pestis remains susceptible to many antibiotics, a growing number of resistant strains have been reported; and the ease of transferring resistance via plasmids is well established [45-47]. Furthermore, there is no FDA-approved vaccine for plague. A short incubation period, high mortality rates, and potential for mass infection through aerosolization warrant the development of novel countermeasures for the plague. In this study, prototype immunoassays were developed for the detection of LcrV and F1 for potential diagnosis of plague infections. Hybridoma cell lines producing mAbs against LcrV and F1 antigens were produced. The resulting mAbs were evaluated for binding kinetics and used to develop sensitive immunoassay prototypes. Many mAb pair combinations were evaluated to develop LFIs and ELISAs for the detection of LcrV and F1. Surrogate clinical samples were used to determine the analytical sensitivity or LOD for the LFI prototypes. Pathogenic near neighbors and other Tier 1 bacterial Select Agents were used to begin to evaluate specificity of the LFI. The overall goal is for these prototypes to eventually be validated in a variety of sample matrices collected from plague patients followed by FDA approval.

Methods

Ethics statement

The use of Normal Human Serum has been reviewed by the Institutional Review Board at the University of Nevada, Reno (OHRP #IRB00000215). Normal Human Serum acquired from a commercial source has been classified as exempt human subject research (exemption 4) as i) no specimens will be collected specifically for this study and ii) there are no subject identifiers. As a consequence, the use of clinical samples in this project does not meet the criteria of human subject research as per 45 CFR 46 of the HHS regulations.

The use of laboratory animals in this study was approved by the University of Nevada, Reno Institutional Animal Care and Use Committee (protocol number 00024). All work with animals at the University of Nevada, Reno was performed in conjunction with the Office of Lab Animal Medicine, which adheres to the National Institutes of Health Office of Laboratory Animal Welfare (OLAW) policies and laws (assurance number A3500-01).

Monoclonal antibody (mAb) production

Animals were used in this study to produce reagents (mAbs) for the development of immunoassay assays. All animal work and husbandry were in the animal facility at the University of Nevada, Reno. Twenty female CD1 mice, 6–8 weeks old (Charles River Laboratories, Inc.), were immunized with either recombinant F1/LcrV fusion protein (F1-V), LcrV, or F1 (Biodefense and Emerging Infections Research Resources Repository [BEI Resources], Manassas, VA). Subcutaneous injections of recombinant F1-V, LcrV, or F1 in emulsions of Titermax Gold Liquid Adjuvant (TiterMax, Norcross GA) were performed with subsequent boosts at weeks 4 and 8. Immunizations of recombinant F1 were also performed using Freund’s complete adjuvant (MilliporeSigma, Billerica, MA) via intraperitoneal injection with subsequent boosts using Freund’s incomplete adjuvant (MilliporeSigma) at weeks 4, 8 and 12. The number of animals was selected to account for variability in the immune response, human error, and for any unforeseeable causes (i.e. death of animal). Sera samples were collected via submandibular or retro-orbital survival bleeds and titers were screened by indirect ELISA. A final boost of recombinant protein without adjuvant was administered intravenously three days prior to splenectomy. Euthanasia was performed by CO2 asphyxiation and final bleeds were performed via cardiac sticks. Splenocytes from all mice were harvested and hybridoma cell lines were produced by standard technique using the P3X63Ag8.653 fusion partner [48]. Splenocytes from mice with the highest titers against the target antigen, as assessed by indirect ELISA, were prioritized to produce hybridoma cell lines. Remaining splenocytes were frozen down and stored in liquid nitrogen. Hybridoma cell lines were isolated through cloning by limiting dilutions. Cells were grown in vitro using Dulbecco’s Modified Eagle’s Medium (DMEM) containing fetal bovine serum (FBS) and conditioned media. Supernatant was collected and purified by protein A affinity chromatography yielding >10 mg/L purified mAbs.

Indirect ELISA

Microtiter plates (96-wells) were coated with recombinant protein (LcrV or F1) in phosphate buffered saline (PBS) overnight at room temperature. Plates were washed with PBS containing 0.05% Tween-20 (PBS-T) then blocked for 90 minutes at 37°C with PBS-T containing 5% non-fat milk (blocking buffer). Primary antibodies (mouse sera or purified mAbs) were diluted in blocking buffer and serial two-fold dilutions were performed across plates. Primary antibodies were incubated for 90 minutes at room temperature. Plates were washed with blocking buffer then incubated with horseradish peroxidase (HRP) labeled goat anti-mouse IgG antibody (Southern Biotech, Birmingham, AL) for 60 minutes at room temperature. Isolated mAbs were also analyzed using IgG subclass specific (IgG3, IgG1, IgG2a, IgG2b) goat anti-mouse secondary antibodies (Southern Biotech) for further characterization. Plates were washed with PBS-T and incubated with tetramethylbenzidine (TMB) substrate (Kirkegaard & Perry Laboratories, Inc., Gaithersburg, MD) for 30 minutes at room temperature. An equal volume of 1 M phosphoric acid (H3PO4) was used to stop the reaction, and colorimetric data was read at an optical density of 450 nm (OD450).

Bacterial lysate preparation

At biosafety level 2 (BSL2), a glycerol stock of Y. pestis Harbin-35 (BEI Resources) was streaked onto a brain-heart infusion (BHI) agar plate and incubated at 28°C for 48 hours. An individual colony was picked, inoculated into tryptic soy broth (TSB), and grown overnight at 37°C shaking in 5% CO2. Larger cultures were inoculated from the starter culture and grown for 48 hours at 37°C shaking in 5% CO2. Bacterial cells were pelleted by centrifugation and resuspended in PBS. Cells were heat-inactivated at 80°C for 2 hours. Bacterial supernatant was 0.2 μm filtered. Bacterial cell lysate and supernatant were backplated onto BHI agar and incubated at 37°C for at least 72 hours to ensure no viable cells were present.

Additionally, bacterial lysates for Y. pestis KIM D19, Y. pestis A12 Derivative 6, Y. pseudotuberculosis IP2666, Y. enterocolitica WA, Francisella tularensis B38, F. tularensis LVS, Bacillus anthracis Ames-35, Burkholderia pseudomallei K96243, B. pseudomallei 1026B, and B. pseudomallei Bp82 were prepared as per instructed (BEI Resources). Bacterial lysates were heat inactivated and separated by cells and supernatant. B. pseudomallei strains K96243 and 1026B were propagated at biosafety level 3 (BSL3) and confirmed nonviable before removal to (BSL2) by a validated protocol. Y. pestis KIM D19, Y. pestis A12 Derivative 6, Y. pseudotuberculosis IP2666, Y. enterocolitica WA, F. tularensis B38, F. tularensis LVS, B. anthracis Ames-35, and B. pseudomallei Bp82 lysates were propagated at BSL2 as per protocol above. Inactivated bacterial cells were resuspended in PBS to an optimal density at 600 nm (OD600) of 0.5.

Recombinant LcrV and F1 cloning and expression

Genes encoding LcrV and F1 were cloned into Escherichia coli, expressed and purified. The F1 encoding gene (caf1) and lcrV were amplified by polymerase chain reaction (PCR) from the Y. pestis Harbin-35 lysate, using primers shown in S1 Table. The caf1 gene was void of the first 63 base pairs which encodes for a cleaved signal peptide [49]. Each gene was cloned into the pQE-30 Xa Vector (Qiagen, Hilden, Germany) by Gibson Assembly (New England Biolabs (NEB), Ipswich, MA). Plasmids were sequence verified and transformed into E. coli M15 for protein expression. E. coli containing expression plasmids were grown at 37°C to an optimal density at OD600 of 0.6 before inducing protein expression by the addition of isopropyl ß-D-1-thiogalactopyranoside (IPTG) at an end concentration of 1mM. Induced cultures were grown at 37°C for 12–16 hours. Bacterial cell pellets were collected by centrifugation then lysed with BugBuster 10X Protein Extraction Reagent (MilliporeSigma) and sonication. The soluble fraction for both were purified using Protino Ni-TED resin (Macherey-Nagel, Duren, Germany).

Western blot

A standard Western blot procedure was performed using semidry blotting. 6x reducing or non-reducing Laemmli loading buffer was added to bacterial lysate. The reduced sample was then boiled for 10 minutes to denature the proteins. Samples were separated on a 10% sodium dodecyl sulfate (SDS) gel, and proteins were transferred to a nitrocellulose membrane (Bio-Rad Laboratories, Hercules, CA). HRP-labeled LcrV mAbs were used to probe the membrane directly. Unlabeled F1 mAbs were used to probe the membrane, followed by an HRP-labeled goat anti-mouse Ig for detection. Signal was developed using SuperSignal™ West Femto Maximum Sensitivity Substrate (ThermoFisher Scientific, Grand Island, NY) and imaged using a ChemiDoc XRS imaging system (Bio-Rad Laboratories).

LFI

LFIs were initially constructed using Fusion-5 matrix membrane (GE Healthcare, Piscataway, NJ), FF120HP nitrocellulose membrane (GE Healthcare) and C083 cellulose fiber sample pad strips (Millipore Sigma). FF120HP membranes were striped with unlabeled antibodies using the BioDot XYZ platform (BioDot, Irvine, CA). Purified Y. pestis mAbs were striped at 1 mg/mL and served as the test line. Unlabeled goat anti-mouse Ig (Southern Biotech) were striped serving as the control line. All mAbs were conjugated to 40 nm colloidal gold particles (DCN Diagnostics, Carlsbad, CA), blocked with bovine serum albumin and concentrated to an optical density of 10 at 540 nm. Stability of the mAb conjugates were tested by stability in high salt conditions. Colloidal gold conjugated mAbs were spotted roughly 8mm from the top of the Fusion-5 matrix membrane prior to running the assay and served as the detection antibody. Initial screening was done using all possible mAb pairs for LcrV and F1. A single concentration of recombinant protein (100 ng/mL) was used to evaluate mAb pairing. Samples of 40 μL were loaded onto the sample pad then placed in a well containing 150 μL of chase buffer only. Each prototype was run in parallel with chase buffer only, as a negative control. Test line intensity was read promptly after 20 minutes using the ESE Lateral Flow Reader (Qiagen). Top performing pairs were then down selected using a concentration of 1 ng/mL of recombinant protein.

Top LFI prototypes were optimized for testing using human sera. Recombinant protein (LcrV or F1) was spiked into six lots of pooled normal human sera ([NHS], Bioreclamation IVT, Westbury NY & Innovative Research, Novi MI) and evaluated for signal intensity and non-specific background signal. LFI components optimized include conjugate release pads, nitrocellulose membranes, wicking pads, chase buffers, striping concentrations, gold conjugate diluents and sample pre-treatment steps (S2 Table). The final prototypes were constructed using UniSart CN140 nitrocellulose membrane (Sartorius, Germany), CFSP203000 absorbent pad, and conjugate pad grade 8951 (Ahlstrom, Finland). F127 and 10G surfactants were added to the gold diluent for the LcrV and F1 prototypes respectively to help with the release of the conjugate from the conjugate pad. Mouse IgG was added to the human serum samples as a pretreatment step to prevent human anti-mouse antibody (HAMA) interference [50].

Antigen-capture ELISA

The top eight performing mAbs ranked by LFI testing were screened to develop antigen-capture ELISAs. Microtiter plates were coated overnight with 1 μg/mL capture mAb. After blocking, recombinant protein (F1 or LcrV) was diluted into blocking buffer and serial two-fold dilutions were performed across plates. Detection antibody at 0.1 μg/mL was incubated for 60 minutes at room temperature. HRP labeling of mAbs was performed using EZ-Link Plus Activated Peroxidase (ThermoFisher Scientific). Plates were washed with PBS-T and incubated with TMB substrate for 30 minutes at room temperature. An equal volume of 1M H3PO4 was used to stop the reaction, and colorimetric data was read at OD450.

Checkerboard ELISAs were performed using the top two mAb pairs to optimize concentrations of capture and detection mAbs. Optimization was performed using two-fold serial dilutions of either the capture or detection mAbs in independent experiments. The capture mAbs was first optimized by using concentrations ranging from 0.078–10 μg/mL with 1 μg/mL detection mAb. Capture mAb concentrations were chosen based on signal intensity and minimal background. The selected capture mAb concentrations were used to optimize the detection mAbs from a range of 0.0078–1 μg/mL. The optimized ELISA conditions were selected based on the lowest LOD defined as the concentration at two-fold background signal. The top two mAb pairs for each antigen were evaluated to determine the theoretical ELISA limit of detection defined as two-fold background signal.

Surface plasmon resonance (SPR)

SPR experiments were conducted on the Biacore X100 instrument using the His Capture format (GE Healthcare). A CM5 chip surface was prepared using the His Capture Kit as per manufacturer’s recommendation. For each cycle, his-tagged recombinant protein (LcrV or F1) diluted into HBS-EP+ buffer (GE Healthcare) was immobilized onto the anti-His capture surface. LcrV was immobilized at a concentration of 5 μg/mL and F1 was immobilized at a concentration of 1 μg/mL. Antigen capture was performed at 5 μL/second for 60 seconds followed by 120 second stabilization. These conditions were established for optimal kinetic analyses of each mAb. Full kinetic analyses were performed by injecting each purified mAb for 7 cycles at a concentrations range of 0.5–50 μg/mL over the LcrV or F1 surface for 120 seconds followed by a dissociation period of 240 seconds at 30 μL/second. The anti-His capture surface was regenerated between each cycle using 10 mM glycine pH 1.5 (GE Healthcare) for 30 seconds at 10 μL/second. Binding kinetics and affinity were evaluated using a bivalent model on the Biacore X100 Evaluation Software (GE Healthcare).

Results

Hybridoma production and mAb reactivity

In order to develop antigen-capture immunoassays for the detection of LcrV and F1, a large panel of mAbs were isolated and evaluated to determine the optimal conditions for assay performance. Titermax Gold or Freund’s adjuvants were combined with antigens and used for immunization [51,52]. Twenty-two hybridoma cell lines that produce mAbs against Y. pestis LcrV or F1 were isolated (Table 1). MAb reactivity was confirmed by indirect ELISA to recombinant F1 or LcrV proteins. Additional mAbs were isolated from mice immunized with the F1-V fusion protein; however, these mAbs were only reactive to the F1-V fusion protein and not reactive to the individual proteins. Subclass specific secondary antibodies were used to characterize each mAb subclass, and all antibodies were determined to be members of the IgG1, IgG2b, or IgG2a subclass (Table 1). These antibody subclasses are preferred over the IgG3 subclass in the development of antigen-capture immunoassays as the IgG3 subclass can self-associate resulting in increased background signal and potential false positive reactions [53].

Table 1

Monoclonal antibody (mAb) library against Y. pestis LcrV and F1 antigens.

Antigen	mAb	Immunization	Subclass
LcrV	2B2	F1-V	IgG2a
	4E8	LcrV	IgG2a
	5D3	LcrV	IgG1
	6E5	LcrV	IgG1
	6F10	LcrV	IgG1
	8F3	LcrV	IgG1
	8F7	LcrV	IgG2a
	8F10	LcrV	IgG1
F1	3A2	F1-V	IgG2a
	3F2	F1	IgG1
	4E5	F1	IgG2a
	4F12	F1	IgG1
	5E10	F1	IgG2a
	9B7	F1	IgG1
	10D9	F1	IgG1
	10E3	F1	IgG1
	11B8	F1	IgG2a
	11C7	F1	IgG1
	12B6	F1	IgG2a
	12E10	F1	IgG2a
	12F5	F1	IgG2b
	15C4	F1	IgG2a

Western blot analysis probing Y. pestis Harbin-35 lysate determined mAb reactivity against the two Y. pestis proteins. High density bands were detected at the expected molecular weight of roughly 40 kDa indicating reactivity with monomeric LcrV protein (Fig 1A) [54]. Reactivity is observed at higher molecular weights, indicating that both monomeric and multimeric forms are produced in bacterial culture [54]. MAbs 4E8 and 5D3 had limited reactivity to the non-reduced protein as well as the multimeric forms suggesting reactivity to linear epitopes that are more available in reducing conditions (Fig 1B). MAbs isolated against F1 showed differing levels of reactivity to the monomeric F1 antigen in the reduced Western blot (Fig 2A). The non-reduced Western blot analysis showed reactivity of the F1 mAbs against the assembled F1 capsule which resulted in an expected laddering pattern (Fig 2B) [55]. Reducing conditions disrupt disulfide bonds as heat disrupts other protein-protein interactions; therefore these conditions disassemble the F1 capsule’s native polymeric structure. MAb 3A2 was isolated from a mouse immunized with the F1-V fusion protein and showed preferential reactivity to the monomeric structure and had low reactivity to the assembled capsule. Inversely, mAbs 9B7 and 12F5 showed limited reactivity to the reduced bacterial lysate and were more reactive against higher molecular weight F1 multimers present in the non-reduced lysate (Fig 2B). Reactivity of mAbs 9B7 and 12F5 may suggest binding only occurs to a specific epitope formed in the F1 capsule between at least three or more subunits (>46.5 kDa).

Fig 1

Western blot analysis of anti-LcrV monoclonal antibodies (mAbs) against Yersinia pestis Harbin-35 lysate.

Horseradish peroxidase (HRP) conjugated LcrV mAbs (1 μg/mL) were used to probe (A) reduced and (B) non-reduced Y. pestis Harbin-35 bacterial lysate (1.5x106 cells/lane) by direct Western blot.

Fig 2

Western blot analysis of anti-F1 monoclonal antibodies (mAbs) against Yersinia pestis Harbin-35 lysate.

Anti-F1 mAbs (1 μg/mL) were used to probe (A) reduced and (B) non-reduced Y. pestis Harbin-35 bacterial lysate (1.5x106 cells/lane) by indirect Western blot. HRP-conjugated goat anti-mouse Ig was used for detection of F1 mAb binding.

Binding affinity and kinetics by surface plasmon resonance (SPR)

Characterization of the mAbs in each library included analysis of binding kinetics by SPR. Experiments were performed in triplicate and evaluated using a bivalent binding model. The association rate (ka), dissociation rate (kd), and equilibrium dissociation constant (KD; KD = kd/ka) are reported in Table 2. LcrV mAbs displayed a narrow range of equilibrium dissociation constants (0.3–4.5 nM). F1 mAbs had a larger range of 0.002–250 nM. Interestingly, mAbs 2B2 and 3A2 were isolated from mice immunized with the F1-V fusion protein. 2B2 displayed a high affinity to recombinant LcrV (KD = 2.6 nM); however, 3A2 displayed the poorest affinity to recombinant F1 (KD = 250 nM).

Table 2

Affinity and kinetics analysis of Y. pestis mAbs by surface plasmon resonance.

Antigen	mAb	ka x 103 (M-1s-1)	kd x 10−3 (s-1)	KD (nM)
LcrV	2B2	52	0.14	2.6
	4E8	210	0.78	3.7
	5D3	65	0.22	3.4
	6E5	120	0.54	4.5
	6F10	140	0.58	4.1
	8F3	93	0.17	1.9
	8F7	100	0.18	1.8
	8F10	250	0.074	0.3
F1	3A2	5.6	1.4	250
	3F2	65	0.48	7.4
	4E5	250	0.39	1.6
	4F12	91	1.8	19
	5E10	230	0.12	0.5
	9B7	42	0.45	11
	10D9	36	0.048	1.3
	10E3	100	0.054	0.5
	11B8	170	0.17	1.0
	11C7	180	0.91	5.2
	12B6	110	0.00026	0.002
	12E10	83	0.23	2.8
	12F5	170	13	79
	15C4	210	0.13	0.6

LFI development and optimization

The library of mAbs were evaluated as capture and detection components for the development of LFI prototypes. MAb pairs were quantitatively ranked based on the signal intensity at the test line when analyzing a sample containing 100 ng/mL of recombinant protein (LcrV or F1) spiked into chase buffer minus nonspecific signal when chase buffer alone was analyzed (S3 and S4 Tables). Though LFIs are generally evaluated by visual detection, the use of Qiagen ESE Lateral Flow Reader provided a standardized measure to minimize human variability and error. Visual LOD is estimated to have an intensity between 15–30 mm*mV among various LFIs developed in our laboratory. The mAb pairs were then down selected for detection of 1 ng/mL recombinant protein (Table 3). The top four performing pairs (capture/detection) were evaluated for detection of LcrV in bacterial lysate (S1 Fig). MAb pair 8F10/2B2 was highly reactive to bacterial lysate. This pair, however, consistently had nonspecific reactivity with control, prompting the evaluation of the next top performing pair (8F10/6F10). Further optimization is warranted to minimize nonspecific reactivity as these results together suggest 8F10/2B2 is more analytically sensitive when detecting native protein despite 8F10/6F10 having similar results with recombinant protein (Table 3). The top pair for F1 detection was 11C7/3F2 when evaluated with 1 ng/mL recombinant F1 protein (Table 3). S5 Table shows additional F1 LFI prototypes that displayed reactivity in chase buffer alone.

Table 3

Assay sensitivity of the top 4 mAb pairs (LcrV or F1) by lateral flow immunoassay.

Antigen	Capture mAb	Detection mAb	OD (mm*mV) 1 ng/mL antigen	OD (mm*mV) Chase Buffer only
LcrV	8F10	6F10	56	0
	8F10	2B2	53	0
	8F10	6E5	51	0
	8F7	6F10	26	0
F1	11C7	3F2	115	0
	11C7	15C4	93	0
	4E5	3F2	89	0
	10D9	3F2	57	0

Testing complex matrices such as human sera often lead to assay signal loss and nonspecific binding described as matrix effects [56]. To account for interference of the matrix on downstream applications, the top LFI prototypes were further optimized for assaying human serum. The LOD was defined as the minimum concentration of recombinant protein at which a positive signal is observed in spiked pools of normal human sera (NHS). In this study, a positive signal was defined as an intensity reading greater than 20 mm*mV using the Qiagen ESE Lateral Flow Reader. Due to sample variability, difference in assay performance was observed between each lot of pooled human serum. Six different pools of NHS were used to optimize each LFI prototype for nonspecific background signal. Then the pools were spiked with recombinant protein and used to determine the LOD of the LFI prototypes to account for possible matrix effects on signal intensity. The LOD of the 8F10/6F10 LFI was estimated to be 2 ng/mL when recombinant LcrV was spiked into NHS (Lot NHS207) (Fig 3A). The LOD of the 11C7/3F2 LFI was estimated to be 1 ng/mL when recombinant F1 was spiked into NHS (lot NHS207) (Fig 3B). Despite high analytical sensitivity to each antigen, differences of assay signal were observed between the six pools of NHS (S2 and S3 Figs). The aggregate LOD was determined as the lowest concentration which consistently resulted in positive LFIs among the pools of NHS. The LcrV and F1 prototypes produced LODs of roughly 1 ng/mL. Detection of both antigens in pooled NHS #28614 appeared slightly reduced (S2 Fig) and illustrates the variation in signal intensity that can occur when testing within different lots of pooled human serum.

Fig 3

Sensitivity of Y. pestis lateral flow immunoassays (LFI) using recombinant LcrV and F1.

LFI prototypes were tested with recombinant (A) LcrV and (B) F1 serial diluted into pooled normal human serum ranging from 0.25 to 256 ng/mL. Assay signal was evaluated and quantitated by optical density using a Qiagen ESE reader. Intensity ≥ 20 mm*mV scores as positive.

Generation of a multiplexed LcrV/F1 assay

To detect potentially all pathogenic isolates of Y. pestis, a multiplexed LFI was developed. The prototype was constructed with two test lines, one specific for LcrV (8F10/6F10) and one for F1 (11C7/3F2). Initial specificity testing of the dual LFI prototype was performed using bacterial lysates from Select Agent exempt strains of Y. pestis, clinically relevant near neighbors, and other Tier 1 bacterial Select Agents (Fig 4). Cell lysates were adjusted to an OD600 of 0.5. Y. pestis Harbin-35 and KIM19 strains were positive for LcrV and F1. Y. pestis A12 Derivative 6, an LcrV- strain, was positive only for F1. Y. pseudotuberculosis IP2666 lysate was positive for LcrV, but Y. enterocolitica was negative for LcrV. High levels of homology attribute to some degree of mAb cross-reactivity to LcrV of near neighbors [42,57]. Additionally, all other Tier 1 bacterial Select Agents were negative for both LcrV and F1.

Fig 4

Specificity testing of dual Yersinia pestis lateral flow immunoassay (LFI) against clinically relevant bacterial panel.

The dual LFI prototype containing test lines specific for LcrV (8F10/6F10) and F1 (11C7/3F2) was tested against a panel of bacterial lysates. Bacterial lysate (50 μL at OD600 = 0.5) was applied onto the conjugate pad and chased with buffer. LFIs were imaged after 20 minutes.

Antigen-capture ELISA development and optimization

The top eight mAbs ranked by LFI were used to develop a quantitative antigen-capture ELISAs for the detection of Y. pestis antigens. ELISA mAb pairs were initially evaluated using 1 μg/mL capture mAb and 0.1 μg/mL detection mAb. To compare the performance of each mAb pair, antigen concentrations at five-fold background signal were determined from a linear regression plot generated from a two-fold serial dilution of each antigen (S4 Fig). This value (five-fold background) was selected to account for possible background signal caused by the mAb pairing. The top two performing mAb pairs for each antigen were optimized for capture and detection conditions by checkerboard ELISAs (Table 4). The theoretical ELISA LODs (defined as two-fold background signal) for each assay were determined using recombinant protein spiked into buffer and reported in Table 4 and S5 Fig. LcrV was detected at 74 pg/mL by 6E5/8F10 and F1 was detected at 61 pg/mL by 11B8/11C7.

Table 4

Limit of detection of enzyme-linked immunosorbent assays using recombinant LcrV and F1 antigens in buffer.

Antigen	Capture mAb	[Capture mAb] (μg/mL)	Detection mAb	[Detection mAb] (μg/mL)	LOD* (pg/mL)
LcrV	6E5	2.5	8F10	0.13	74 ± 7.9
	6F10	2.5	8F10	0.063	75 ± 4.0
F1	10D9	2.5	11C7	0.13	100 ± 41
	11B8	1.3	11C7	0.13	61 ± 2.0

*LOD defined as the concentration at two-fold background signal

Discussion

The quick progression of plague infections warrants the need for sensitive, specific, and rapid diagnostic tools. Though most infections lead to bubonic plague, the more severe forms of the infection present with nonspecific clinical symptoms that can be challenging to distinguish from many other diseases. Commercially available LFIs have been developed for the detection of the F1 antigen and is used for patient screening, however no product is currently FDA approved to diagnose plague [18,20]. F1 encapsulates the bacteria and is shed; however, F1- isolates have been identified and are fully virulent making the biomarker inadequate for diagnosing all plague infections [25,33]. To combat the potential of widespread infections, an RDT capable of detecting all pathogenic strains of Y. pestis is imperative. Through the isolation of a library of mAbs reactive to F1, as well as LcrV, we have developed assays for the detection of various pathogenic Yersinia species which may prove to be useful on multiple fronts.

The success of Y. pestis as a pathogen may be attributed to its genetic diversity. Widespread pandemics have been traced to three global biovars of Y. pestis (antiqua, medievalis, and orientalis). The lcrV gene is conserved among the biovars virulent in humans [54]. Polymorphisms in lcrV have been observed in the biovar microtus [58]. Some polymorphisms occur in critical regions impacting LcrV to form multimers, suggesting the attenuation of certain biovars in humans [54]. In addition to polymorphisms in lcrV, various isoforms of the F1 antigen exist due to point mutations [59,60]. The NT1 isoform (Ala48 Phe117) is most common and is expressed by the three global strains responsible for human infections [59]. Moreover, the adaptive immune response elicited by the NT1 isoform are cross-reactive to the NT2 (Ser48 Phe117) and NT3 (Ala48 Val117) isoforms suggesting these amino acid substitutions may not affect the binding of the described mAbs [61]. Y. pestis Harbin-35 and KIM D19 strains are included within biovar medievalis and maintain the three plasmids associated with virulence (pCD1, pPCP1, pMT1), but are attenuated due to the deletion of the pgm locus, which encodes a segment of the chromosome that includes a pathogenicity island [62]. The pgm locus encodes for various iron-regulated proteins necessary for bubonic and pneumonic plague infections [63]. Unlike most laboratory strains, Harbin-35 and KIM D19 maintain the pCD1 plasmid and LcrV expression. The A12 D6 strain is included within the biovar orientalis and is lacking the entire pCD1 plasmid. mAbs isolated in this study show reactivity to Y. pestis Harbin-35, KIM D19, and A12 D6 proteins which suggest mAb binding may be against all human pathogenic strains of Y. pestis (Figs 1–2 & 4). Further characterization of the mAb panel should be performed to confirm reactivity among all pathogenic Y. pestis strains.

Binding kinetics derived from SPR demonstrate that most of the mAbs produced in this study possess high affinity to their targets (Table 3). In general, higher affinity mAbs are preferred in immunoassays to achieve optimal analytical sensitivity which should result in improved clinical sensitivity; however, the mAb pairs which performed well in LFI and ELISA formats did not necessarily have the highest affinity or superior kinetics. Initial screening of LcrV mAbs by ELISA indicated that the pairing of high affinity mAbs 8F7 and 8F10 resulted in LODs 10 to 20-fold less sensitive than 8F10 and either 6E5 or 6F10, mAbs with lower degrees of affinity (Table 2; S4 Fig). Likewise, the mAbs used in the optimized F1 ELISAs (10D9, 11B8, and 11C7) were ranked among the middle for binding affinity (KD). Additionally, mAb 12B6 had binding affinity over 2 logs greater (KD = 0.002 nM) than the other mAbs in the panel but did not perform well in the antigen-capture format. Overall, the mAbs used to develop the sandwich assays all displayed dissociation constants less than 10 nM. These results exemplify that in addition to having high affinity, mAbs used in antigen-capture immunoassays require pairing synergy.

Interestingly, the top performing mAb pairs in the LFI format were not the same as the ELISA. Evaluation of LFI and ELISA mAb pairs for Ebola glycoprotein and F. tularensis in our laboratory have also shown differences between the two formats [64,65]. Though both LFIs and ELISAs are antigen-capture immunoassays, there are notable differences in the assay formats. ELISAs reach binding equilibrium with long incubation periods, minimize nonspecific binding with several wash steps, and are more sensitive due to enzymatic signal amplification. However, the ELISA format is less accessible in low-technology settings such as in the field or in developing countries due to laboratory equipment requirements and storage of temperature-sensitive reagents. In contrast, LFIs utilize capillary flow to apply reagents systematically without intermediate washing steps, are self-contained within a single dipstick, and produce results in less than 20 minutes by the colorimetric sensing of gold-nanoparticle aggregates, optimal for visual detection by the human eye.

The RDT immunoassay prototypes developed in this study had analytical sensitivities within appropriate levels for assaying serum. Levels of F1 in patient serum can range from 4–50,000 ng/mL [66]. Concentrations of soluble LcrV in human samples have yet to be determined, but a mouse model of infection indicates serum concentrations of LcrV are between 6–26.5 ng/mL 48 hours post-inhalational exposure [41]. To determine the diagnostic potential of LcrV and F1, further studies will need to be conducted within the 48 hours of exposure as this window is crucial for administering treatment for patient survival The ELISAs developed in this study would be useful in quantifying LcrV and F1 in plague patient samples to further validate each diagnostic target. Additionally, the ELISA format may also be important in outbreak settings to allow for high throughput screening of patient samples. Furthermore, the top mAb pair for detecting LcrV by LFI was evaluated in a vertical flow immunoassay (VFI) format. The mini VFI prototype can detect as low as 25 pg/mL LcrV [67].

Y. pestis can invade the bloodstream and be detected within days of exposure, making blood a common test matrix among the three forms of plague. The LFI prototypes were optimized for serum as it is similar to whole blood and can cause matrix effects, or interference by components present in the clinical sample at the test line. Matrix effects were evident in testing both the LcrV and F1 LFIs as pools of normal human serum resulted in varying signal intensity. The LOD results for both the LcrV and F1 prototypes roughly 1 ng/mL. Additionally, the optimized LFI prototypes were able to analyze up to 50 μL of neat human sera with a minimal degree of pre-treatment. The tested assay protocol includes a sample preparation step in which mouse IgG was added to prevent HAMA interference. However, mouse IgG could be integrated into the sample pad prior to finalization of the LFI along with separation pads to allow for whole blood testing. Furthermore, the LFI prototypes described should be evaluated using other clinical matrices such as sputum or pus as these may be more clinically relevant for pneumonic and bubonic infections, respectively.

Plague remains a modern threat to public health, and LFIs are ideal tools for detecting and limiting the spread of infections. In addition to the risk of naturally occurring plague, Y. pestis has been used as a biological weapon. The use of Y. pestis in medieval siege warfare has been well documented, and in modern times, plague-infected fleas were disbursed by the Japanese army in China during World War II [68]. Several countries, including the United States of America and the former Soviet Union, have investigated the utility of Y. pestis as a bioweapon [9]. Further investigation of Y. pestis as a biological weapon is no longer permit under a treaty signed at the Biological Weapons Convention in 1972. Nonetheless, continued development of countermeasures against Y. pestis are warranted in defense of nefarious actions. Of concern is the intentional release of an F1- strain as a bioterrorism tactic, as the F1 antigen is primarily regarded as a diagnostic and vaccine target [69-74]. The multiplexing of assays detecting LcrV and F1 increases the diagnostic ability of a Y. pestis RDT, as it may be capable of detecting all pathogenic Y. pestis strains including those found to be F1 deficient. The dual assay shows high specificity to Y. pestis without cross-reacting with other Tier 1 Select Agents with similar symptoms. It was anticipated that the LcrV/F1 prototype would be positive for LcrV Y. pseudotuberculosis as the homology is 97% compared to Y. pestis [75]. In cases where plague is suspected, a positive result for either F1 or LcrV should prompt immediate treatment [76]. More specificity testing needs to be performed using these prototypes with an increased panel of microbes known to present with similar symptoms. The prompt detection of all pathogenic Y. pestis is critical for minimizing casualties during naturally occurring outbreak or a potential bioterrorism act. Further evaluation on clinical samples collected from plague patients is needed to fully validate this multiplexed LFI.

While the immunoassay prototypes developed in this study were designed with patient samples in mind, a secondary use would be to evaluate animal populations and vectors for the presence of Y. pestis. As a zoonotic disease, plague surveillance requires the monitoring of natural animal reservoirs and associated vectors. Current methods of surveying wild populations include serological testing of sentinel animals, such as coyotes, which prey on smaller mammals [77]. Though coyotes do not present with symptoms, they do elicit an immune response to Y. pestis antigens [78]. The downside of serological surveillance is the persistence of antibodies long after exposure, meaning that a positive response may not be indicative of an active infection. Though expression of LcrV and F1 are regulated by temperature, leaky expression of LcrV at 26°C has been observed [79]. This leaky expression should be evaluated as a mean to detect Y. pestis in vectors such as fleas. Screening of small mammals and vectors could be conducted using an LFI in the field to gain more data regarding the prevalence of Y. pestis in these reservoir populations.

In addition to use in a diagnostic assay, exogenous antibodies have shown to be effective in protecting from and treating plague infections when administered pre- and post-exposure. In a mouse model of pneumonic plague, F1 specific IgG mAbs were able to confer protection when administered prophylactically [80]. Additionally, three human mAbs (one against F1, two against LcrV) isolated by naïve human phage displayed Fab libraries demonstrate some protection for bubonic plague in mice [81]. LcrV and F1 mAb cocktails do have synergy when administered together [82]. LcrV subunits form pentamers at the tip of the T3SS at the cell surface [36,83]. LcrV in this pentameric structure is associated with immunosuppression properties and have shown to elicit a more protective response in vaccine studies [37,84]. Testing the therapeutic potential of this large panel of Y. pestis mAbs in an animal model of infection may result in the development of additional treatment options for plague patients.

Monoclonal antibodies, specific to LcrV and F1, were used to develop antigen-capture LFIs and ELISAs with high analytical sensitivity. The LFI prototypes resulted in LODs of roughly 1 ng/mL for both LcrV and F1 when assaying antigen spiked into normal serum. The ELISAs were able to achieve an analytical sensitivity in the range of 61–74 pg/mL, at this point testing was only preformed in assay buffer. The detection of the F1 antigen is widely used to diagnose plague infections, however mutant strains of Y. pestis lacking the F1 antigen have been identified. Initial inclusivity/exclusivity studies of the multiplexed LFI demonstrates that the inclusion of a second antigen, LcrV, should improve the performance of the RDT. The LcrV antigen is crucial for virulence and may be used as an alternative marker of plague. Further evaluation of LcrV is warranted and can be accomplished using the tools developed in this study. In addition, side-by-side analyses of the assays developed here with other plague assays is a logical next step in order to rank performance. Finally, the immunoassays developed hopefully will be diagnostically useful and there could be potential of the mAbs being further developed into therapeutics, thereby assisting in the control of future plague outbreaks.

Top four LcrV LFI prototypes tested with (A) PBS or (B) Y. pestis Harbin-35 lysate for the detection of native antigen. Test lines were sprayed at 1 mg/mL and 5 uL of gold conjugated mAb (OD540 = 10) was applied.

(PDF)

Click here for additional data file.

LFI prototypes (8F10-capture/6F10-detection) were tested with recombinant LcrV in six pools of normal human serum.

Each panel represent a different lot of pool serum from (A-C) Bioreclamation IVT or (D-F) Innovative Resources. Lot numbers are provided for each panel. Assay signal was evaluated and quantitated by optical density using a Qiagen ESE reader. Intensity ≥ 20 mm*mV scores as positive.

(PDF)

Click here for additional data file.

F1 prototypes (11C7-capture/3F2-detection) were tested with recombinant F1 in six pools of normal human serum.

Each panel represent a different lot of pool serum from (A-C) Bioreclamation IVT or (D-F) Innovative Resources. Lot numbers are provided for each panel. Assay signal was evaluated and quantitated by optical density using a Qiagen ESE reader. Intensity ≥ 20 mm*mV scores as positive.

(PDF)

Click here for additional data file.

Preliminary screen to identify top performing antigen-capture ELISA mAb pairs.

Values shown are the concentrations of (A) recombinant LcrV and (B) recombinant F1 in ng/ml at five times background for each mAb pair. The values represent the mean of two independent ELISAs (each performed in biological triplicates).

(PDF)

Click here for additional data file.

Antigen-capture ELISAs were performed to determine the limits of detection (LOD) for recombinant (A & B) LcrV and (C & D) F1. LODs were calculated using the linear regression of the optimized ELISA conditions to determine the concentration of recombinant protein in ng/ml at two-fold background. The values represent means of three independent ELISAs (each performed in biological triplicates).

(PDF)

Click here for additional data file.

Primers for cloning LcrV and F1 genes from Y. pestis Harbin-35 into the pQe-30 Xa vector.

(PDF)

Click here for additional data file.

Summary of lateral flow immunoassay components evaluated.

(PDF)

Click here for additional data file.

Preliminary assay sensitivities of top mAb pairs by LFI for LcrV at 100 ng/mL.

(PDF)

Click here for additional data file.

Preliminary assay sensitivities of top mAb pairs by LFI for F1 at 100 ng/mL.

(PDF)

Click here for additional data file.

Preliminary assay sensitivities of top mAb pairs by LFI for F1 at 1 ng/mL.

(PDF)

Click here for additional data file.

8 Nov 2021

Dear Dr. AuCoin,

Thank you very much for submitting your manuscript "Development of a dual antigen lateral flow immunoassay for detecting Yersinia pestis" for consideration at PLOS Neglected Tropical Diseases. As with all papers reviewed by the journal, your manuscript was reviewed by members of the editorial board and by several independent reviewers. In light of the reviews (below this email), we would like to invite the resubmission of a significantly-revised version that takes into account the reviewers' comments.

Editor comments:

LFI for detecting Yersinia pestis is often used in the field or at local health centres without any laboratory. Withdrawing and centrifuging blood are often not possible. Therefore you also should test other matrices such as sputum and pus collection.

Early detection of Yersinia pestis in serum is interresting. Could you give results on the monitoring of F1 and LcrV detection in samples of infected patients (or infected animals)?

We cannot make any decision about publication until we have seen the revised manuscript and your response to the reviewers' comments. Your revised manuscript is also likely to be sent to reviewers for further evaluation.

When you are ready to resubmit, please upload the following:

[1] A letter containing a detailed list of your responses to the review comments and a description of the changes you have made in the manuscript. Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

[2] Two versions of the revised manuscript: one with either highlights or tracked changes denoting where the text has been changed; the other a clean version (uploaded as the manuscript file).

Important additional instructions are given below your reviewer comments.

Please prepare and submit your revised manuscript within 60 days. If you anticipate any delay, please let us know the expected resubmission date by replying to this email. Please note that revised manuscripts received after the 60-day due date may require evaluation and peer review similar to newly submitted manuscripts.

Thank you again for your submission. We hope that our editorial process has been constructive so far, and we welcome your feedback at any time. Please don't hesitate to contact us if you have any questions or comments.

Sincerely,

Anne-Sophie Le Guern

Guest Editor

PLOS Neglected Tropical Diseases

Javier Pizarro-Cerda

Deputy Editor

PLOS Neglected Tropical Diseases

***********************

Editor comments:

LFI for detecting Yersinia pestis is often used in the field or at local health centres without any laboratory. Withdrawing and centrifuging blood are often not possible. Therefore you also should test other matrices such as sputum and pus collection.

Early detection of Yersinia pestis in serum is interresting. Could you give results on the monitoring of F1 and LcrV detection in samples of infected patients (or infected animals)?

Reviewer's Responses to Questions

Key Review Criteria Required for Acceptance?

As you describe the new analyses required for acceptance, please consider the following:

Methods

-Are the objectives of the study clearly articulated with a clear testable hypothesis stated?

-Is the study design appropriate to address the stated objectives?

-Is the population clearly described and appropriate for the hypothesis being tested?

-Is the sample size sufficient to ensure adequate power to address the hypothesis being tested?

-Were correct statistical analysis used to support conclusions?

-Are there concerns about ethical or regulatory requirements being met?

Reviewer #1: Partly. see file attached

Reviewer #2: The objective of this study was to develop immunoassays for the detection of dual antigens (LcrV and F1) for the potential diagnoses of plague infections. The objective and study design was clearly articulated and the study design was appropriate to address the objective. No concerns over ethical or regulatory requirements were noted.

Reviewer #3: The objectives of the study is clear : to develop a dual Lateral Flow Immunochromatographic Assay (LFIA) for detecting Yersinia pestis in patients who suffer from plague infection by using 2 antigens F1 and LcrV. The novelty seems to be in the use of 2 antigens in a same test in order to limit the risk of false negative. Effectively the choice to add a second target LcrV is relevant as F1 is specific to Yersinia pestis but not expressed by all virulent strains. However LcrV is expressed by all virulent Yersinia pestis strains but homologs produced by others Yersinia species can hinder assay specificity. In this condition it would have been good to guide the final user in the interpretation of the test. What to do if both test line are positive, only F1, only LcrV or none of them ? in a context a patient seems to suffer from plague. What is the place of such test in the patient managment algorithm ?

The authors performed a huge development of mabs against FI and LcrV with the objective to develop Immunoassays LFIA and Enzyme-Linked ImmunoSorbent Assay (ELISA) and mainly an LFIA for detecting Yersinia pestis by targeting two different antigen F1 and LCRV. The characterisation of the selected hybridoma/monoclonal antibodies (mabs) was performed in the rule of the art with the determination of the subclass in order also to eliminate the IgG3 subclass, the realisation of western blot and the determination of the mabs affinity. It could have been useful to have information concerning the hybridoma culture as it seems that mabs are not produced via mousse ascite. Are they produced in vitro in serum free medium or not ? what about the condition of culture ? and what about the yield of mabs production (mg/ml) before and after purification ? This information is key for the future of the test and to be able to evaluate the feasibility to produce it at large scale. The information that selected hybridoma are good or bad producers of mabs is key to switch from research and developement to industrilization and production. All of this is important to be sure that such mabs could impact positively the patient care. One way to accelerate their transfer to production is also the adaptation of the hybridoma to serum free medium as it is more or less mandatory for industrials partners due to the fact selection of feotal calf serum or donor calf serum induce batch selection and is also costlty and time consuming.

The authors also performed an important screening of the best pairs of mabs for LFIA and ELISA development known to be fastidious and intensive work. However it would have been necessary to have a better understanding of the intended use of the test. It is not clearly described and this is reflected by the implementation of the development and format of test developed by the authors. Moreover, it would have been a benefit for the development of the dual antigen Yersinia test to describe a comprehensive target product profile (TPP). The TPP needs to provide details on the minimum and optimal performance and operational characteristics of the diagnostic tests to be developed. Researchers, developers, and manufacturers use TPPs to ensure that R&D activities are focused on relevant products and designed for the contexts and needs of end-users. There is no TPP described on this study. This TPP include among others the choice of the patient sample nature on which the test need to be developped. In this case the authors proposed the serum but this choice is not so much argued except the fact Yersinia pestis can quickly invade the bloodstream and be detected within days of exposure and consequently make serum an ideal matrix to assay bubonic, pneumonic and septicemic plague. In this case why do they not target whole blood ? moreover why not target sputum or saliva ? Does sera is not too late as we understood that treatment need to be delivered within around 24H00 after exposure. What about the possibility to use plasma (in this case the impact of the anticoagulant need to be explored). Serum seems to be too much restrictive except if the authors can justify this choice apart from the ease of access to develop. Serum need to be used during the devopment of the test but you need to evolve to whole blood. We do not feel that is was an objective for the authors. Serum need to be explored but in this study it seems that the developement of the test is too much deconnected from whole blood. Finally the final formulation must work with whole blood, serum and plasma.

There are also in this study no comparison with commercialized tests that would have been mandatory as they tried to demonstrate an improvment via their test vs the existing ones.

It seems that the authors improve the LoD in analytical condition on recombinant antigen but to definitly proof it would have been better to perform an experimental comparaison instead of a bibliographic one and not only by using recombinant but also on strains and sample from patient. In addition, if you do not carry out an analytical comparison on the same recombinant antigens and on a same format of test and same implementation of it, it is difficult to conclude on the performance improvement. Finally, the most important evaluation is that on patient samples because the improvement on recombinant antigen is not necessarily predictive of the improvement on patient samples.

It lacks also the development of an algorithm to vizualise how this test could be used vs the gold standard and others technics whatever you are in the field or in a labs. As an example is it a screening test that need to be confirmed or to use it purely for diagnosis ? you will not develop the test in the same way and the raw material adjustement will not be the same as you need to know if sensitivity is prefer over specificity or if both need to be at their maximum/ideal value near 100%. We do not kow what is acceptable and we have no information on the minimal performance. This is important as it allows during the development phases to adjust the quantity of each raw materials in order to get the best balance between sensitivity and specificity depending you or more on a screening or diagnosis mode. All of this is a prerequisite to develop a test and not sufficiently highlighted. It is also important depending of the prevalance to have an objective concerning the positive predictive value and/or negative predictive value to be achieved that is fonction of sensitivity, specificity and the geographical prevalence. The authors did not develop also such important topic.

The choice to add a second target is relevant as F1 is specific to Yersinia pestis but not expressed by all virulent strains. However LcrV is expressed by all virulent Y.Pestis strains but homologs produced by other yersinia species can hinder assay specificity. This oblige to guide the final user in the interpretation of the test and what they need to do in case for example you have only positivity for LcrV that could be induced by non specific species. A quick guide could have been suggested by the authors.

Again, the authors developed on serum instead of on whole blood independently of the fact that ideally the final test could be used not only on whole blood or capillary blood from fingers but also on plasma and sera. We understood through the article that the test need to be used on the field near patient but as we have no TPP and no algorithm of it implementation it is important that the authors justify such choice. If whole blood is targeted the component of the strip and/or composition need to be adapted to such matrix in order not to allow red blood cell to go through the membrane by using mabs against red blood cells or a porosity that keep them in the sample pad. The serum can be used but in this case its volume and way of implementation need to be think with a final use with whole blood that seems not to be the case as they used a chase buffer, dipstick format and so on. The choice of the authors to use a dipstick format and the use of a chase buffer is not aligned with the final use of such antigen test. The use of both is usually of what is expected or more adapted for IgG/IgM detection as we usually used small volume of whole blood (around 10µl) or serum (around 5µL) that could fit with dipstick format and the necessity in this case to use a chase buffer as you absolutly need it to allow the migration of the patient sample. The choice to use low volume sample in this condition is one way to avoid hook or prozone effect. For antigen detection you can use more volume in order also to improve the sensitivity with the limit of drops of whole blood you can get from the tip of the fingers. But again if you are able to get sample by performing venipuncture volume is not limitant and the less number of step you have the better it is. Consequently chase buffer is usuallly not use for antigen detection. It is the reason why TPP is important as such information are included into it as well as information concerning the final implementation of the test. The casette format is also more appropriate to the field as you do not need to use a tube and above, all nobody can influence the migration of the sample into the strip during the time of migration as usually people takes the strips for observation and take out and in the strip inside the tube during its migration. To come back to the use of a chase buffer and its impact. For antigen detection we usually do not use a chase buffer, in this experimentation 150µL of chase buffer are used, as the volume of serum is not sufficient to allow its migration from the sample pad to the adsorbent pad. The volume of serum used in this case 40 µl plus 150 µl of chase buffer will impact also the level of the interference that could be reduced by using this chase buffer in a context also where the authors used pool of serum that mask potential high interference from one serum of the pool. The choice to use a chase buffer is not without consequence on the choice of the components of the test and in particular that of the nitrocellulose membrane and on the entire formulation of the test.

It would have been better that the authors developed the test by using a casette format. They can start the developement in dispstick but with the appropriate volume of sera and no chase buffer but in all cases a cassette need to be used to continue the development in order to be as soon as possible near the final format of test that will be used on the field. The integration of a strip in a cassette is not without consequences on the choice of components, their thicknesses… as there are key contact at the interface of the strip and the cassette which will condition the quality and speed of migration of the sample, of the conjugate in the whole test and consequently the sensitivity and specificity.

The determination of analytical performance could have benefited of the reader used to read the strips. This opportunity was not seized and could have shed important light on the variability of the tests. The creation of a pool in this case could have been more relevant for multiplying the number of strips per point of concentration in order to determine a Limit of Blank (LoB), a limit of detection (LoD) as conventionally implemented for the determination of the performance of a quantitative method. We understand here that this is a development of a qualitative test but the use of the reader would have been decisive to determine an LoD rather than to estimate its variability through different pools which from a biological point of view have no reality included for non specific binding or cross reactivity study. It would have been a criteria of differentiation compared to others. To provide a standard method for determining LoB, LoD and LoQ, Clinical and Laboratory Standards Institute (CLSI) has published the guideline EP17 and protocols for determination of LoD and LoQ. LoB,LoD and LoQ are terms used to describe the smallest concentration that can be reliably measured by an analytical procedure. The LoB is the highest apparent analyte concentration expected to be found when replicates of a blank sample containing no analyte are tested. LoB = meanblank + 1.645(SDblank). To establish it, the sample type to use is a sample containing no analyte, e.g. zero level calibrator and to perform 60 replicates and 20 for verification. The sample characteristics are negative or very low concentration sample that is commutable with patient specimens. The LoD is the lowest analyte concentration likely to be reliably distinguished from the LoB and at which detection is feasible. LoD is determined by utilising both the measured LoB and test replicates of a sample known to contain a low concentration of analyte. LoD = LoB + 1.645(SD low concentration sample). The samples containing low concentration of analyte, e.g. dilutions of lowest concentration calibrator again 60 replicates to establish and 20 to verify. The samples charcateristics are low concentration samples, such as a dilution of the lowest non-negative assay kit calibrator or patient specimen matrix containing a weighed out amount of analyte, commutable with patient specimens. LoQ is the lowest concentration at which the analyte can not only be reliably detected but at which some predefined goals for bias and imprecision are met. The LoQ may be equivalent to the LoD or it could be at a much higher concentration but not so much relevant to dertermine in this cwork as the final test will be a qualitative one. For a statistical point of view this would have been of great interest whatever it is for the LoD determination of the ELISA or LFIA.

The use of serum plus a chase buffer and the fact that the authors made pool of serum do not allow to well evaluate the risk to get false positive. The pools of serum have no clinical reality in particular to study non specific binding or interference due to human anti mousse antibody or others factors that induce it. It would have been more relevant to go through more individuals sera to measure the risk of false positive due to particular composition with the presence of human anti-mouse antibody, rhumatoid factor, anti-nuclear antibodies or auto antibodies. Moreover we have no information concerning the serum used and the fact they could have been heat inactivated or pre-treated, conditions that induce usually less interferences. To study that, it is better to use specific pannel known to interfere with test developement and test them on the best pairs of antibodies that give the expected sensitivity and to go more through individual sera to estimate the risk of false positive before to add any treatment of the test. It would have been also important to supply sera on others pathologies as people that could suffer from Yersinia pestis have been multiinfected. Sera from countries in which the test will be deployed is better as sera from such patient are more complex that those from developed countries and interferences could be higher.

Information is lacking on the effort made by the authors to set up the conjugate (determination of pH, load of mabs at the surface of the gold nanoparticles) Did they used the salt method ? but also to set up the nitrocellulose membrane (pH and load of mabs per test line), quantity of conjugated per strip… It is difficult to assess if these results could be improved not only for a sensitivity point of view but also to reduce the risk of false positive due to interference. In development we usually test multiple optical density of conjugate vs different quantity of mabs on nitrocellulose in order to determine the optimum for a sensitivity and specificity point of view. Finally do the authors work at fixed concentration or do they perform adjustement, if not it is one way to improve the test.

It seems that the authors did not try to test more complex combinaison by using 3 or 4 mabs that could sometimes improve the performance of the test. It’s something that need to be investigated even if however it is better for a product managment better to have a minimum of different mabs.

The tests are read at 20 minutes. How this time to result has been determined? is this the optimum ? A kinetics could have been useful to ensure that the signal on the test line does not change over time in this case after 20 minutes. It could have been good to have the reading result by eye with one or two operators in order to establish the LoD as the final test will not be read by the reader. This could also be more discussed.

The positioning of the test lines has not been discussed first F1 and LcrV in second position before the control line, is it an optimum ? what about results obtained by a reverse positioning ? is there any rational to put first F1 and LcrV in second position ? The positionning of the test line could also influence the LoD of each target.

The ELISA could be also very usefull first to perform the verfification performance of the LFIA test and also in a laboratory environnment in addition to the PCR or alone for laboratory not equiped to perform PCR for many reason. The ELISA has not been evalauted in serum but in PBS, how the authors justify this choice ? it would have been better to compare the LoD of the LFIA vs ELISA on a same standard range with F1 an LcrV spiked in a same matrix in this case serum. Moreover did the authors tried to mix F1 and LcrV in a same sample in order to determine if it impacts the LoD of each in this condition.

The bacterial strains were used from the point of view of detection and specificity and not from the point of view of LoD which would have strengthened the analytical study performed by spyking the serum with recombinant F1 and LcrV. It would have been usefull to determine the number of bacteria that the test can detect in a context we have no more data in this study based on samples patient analysis. It seems also that LcrV level in patients is not so well defined …so determination of the smallest quantity of bacteria detectable by the test is important.

Overall, the authors made a huge work to generate mabs against F1 and LcrV in order to detect Yersinia pestis strains by LFIA or ELISA. The developemnt plan is not enough point of care oriented as whole blood was not enough in mind whatever it is for the component used for the strip, or the format like dipstick format. The problem is not that the authors performed part of the development by using sera but more that they used a chase buffer. Part of the analytical performance could be done by using sera but not in the conditions determined by the authors. The determination of the analytical sensitivity need to be performed again in an appropriate format of test and with more strips for each concentration of the standard range in serum and in whole blood by spiking F1 and LcrV but also need to be established by using a standrad range of different concentration of bacteria. The concentration of ng/ml of F1 or LcrV is required but not sufficient they need also to perform a standard range with different bacteria species and to determine the lower quantity of bacteria the test is able to detect. However In this study we have no true study of non specific binding, crossreactivity as the use of pool do not allow to study such topic. For a statical point of view the authors did not use the opportunity of the reader. We have also no comparison with commercialized test. Finnaly the way in which the mabs and the formulation of the prototypes could be valued was not addressed by the authors in the discussion and this is an important point in determining whether all this work can benefit to the patients quickly. The conditions of access to the mabs and the conditions of availability of the tests were not discussed. Accessibility is key for users and patients from developing countries because too often tests are too much expensive.

--------------------

Results

-Does the analysis presented match the analysis plan?

-Are the results clearly and completely presented?

-Are the figures (Tables, Images) of sufficient quality for clarity?

Reviewer #1: partly. see file attached

Reviewer #2: The results are clearly and completely presented. The tables and figures are of sufficient quality for clarity.

Reviewer #3: The results correspond well to the plan presented and the quality of the tables, photos and figures are relatively of good quality. However some points need to be discussed.

Table. 1, the objective of the double F1-V immunization has not been sufficiently described in the context of human diagnostic use, nor does the low number of hybridomas selected under these conditions. The F1-V immunization seems to have brought little mab with the exception of 2B2 recognizing LcrV and 3A2 recognizing F1 but which were not kept in the formulation described as being the best performing pairs. Moreover, 2B2 and 3A2 are specifics for LcrV and F1 respectively, but does the authors perform experimental study to verify it.

We have no particular remark regarding the nature of the experiments carried out to generate or characterize the mabs. However, Fig. 1A/B and Fig. 2A/2B, the authors could have discussed the fact that for the detection of the F1 protein the intensity of the bands in reducing condition is among the weakest for the antibodies selected in the best Pair 3F2 and 11 C7 which tends, in contrast to 3A2, to a lower recognition of the monomeric forms. Does The authors have an explanation for this and what could be the impact ? This is not the case for the best pairs determined with regard to the capture of LcrV for the 2B2 and the 6F10 with the exception may be for the 8F10 whose signal intensity seems to be weaker than that of the 2B2 and 6F10 in reducing condition or not. Moreover, it would have been relevant to carry more western blots with others strains than Yersinia Pestis harbin-35.

Table. 2, with regard to the dissociation constants, it is often observed that the antibodies of better affinities are not necessarily the antibodies that are found in the best combinations of antibodies in capture and in revelation as could be observed by the authors. It is in all cases important to benefit from this characterization upstream of a first screening while remaining careful not to rule out too quickly antibodies whose affinity could be assessed as low on the basis of this characterization

Page 17 we understand that the best performing pair (capture/detection) against native LcrV was 8F10/2B2 but that due to consistent nonspecific reactivity the authors switched on 8F10 / 6F10. They refers to supplemental figure 1 where the result in the photo are almost negative for the couple 8F10/6F10 on the lysate of Y pestis Harbin-35 with better results for 8F10 / 2B2 couple. Contrasting also with results in Table 3 that show equal sensitivity for 8F10/6 F10 compared to 8F10/B2. Does the authors could explain this ? We can not measure also the effort done by the authors try to reduce interferences with 8F10/2B2 pairs that seems to be more sensitive vs 8F10/6F10. In this context it would be important to know if the authors made any adjustments of these conjugates vs the antibody load on the nitrocellulose membrane? This work could have allowed to probably keep more combination to be tested on the samples of patients in a context where this combination 8F10 / 2B2 does not seem to interfere with PBS and where the use of a serum pool to assess non-specificity is irrelevant because in practice it is carried out on a large number of individual serum, all coming from or known to interfere.

Fig.3 the quality of the photos is good but the reading with people who are not rapid test developer would have been appreciated vs the measurement with the reader in order to have a better idea of the LoD that can be determined by a neophyte in the field. Developers will see as well as a reader, which may not be the case with uneducated people and more representative of users in the field.

With regard to Fig.4, it would have been preferable to present the expected results for each of the strains vs. the results obtained. How does The authors explain the negative result for LcrV for the strain Y pestis A12 derivative 6 described to be a strain expressing LcrV ? it does not appear that the discussion provides an explicit answer to this question. Moreover, is the negative result for F1 for strain IP2666 is expected ? if not what is the explanation? If the expected positive results are negative is it linked to a lack of sensitivity at this concentration ? what about the result with a higher quantity of bacteria? For the enterocolitica strain, it is specified that the result is negative for LcrV, nor does it seem much more positive for F1. Can the authors provide more details? It seems that the results obtained for the other strains are as expected, ie negative.

ELISA development could perhaps have been benefited from antibody screening less dependent on that carried out in LFIA, in particular for the detection of the F1 antigen. The 6E5/8F10 combination allowing the detection of LcrV is one of the combinations identified in LFIA, without however, that we can clearly understand why it was not retained in LFIA nor indeed the efforts made to maintain it, as others, besides vs 8F10 / 2B2 or 6F10. The authors mentioned page 19 an LOD of 61 pg/ml obtained by an 11B8/3F2 pair in ELISA and refers to table. 4 in which this combination is not mentioned because only 11B8/11C7 appear. Could The authors clarify this? as in supplementary fig 4 11B8/3F2 or the reverse does not seem to be in the best combination. If this is an error ? that allows us to identify that this combination has not been tested in LFIA and could be interesting to test. Even though the combinations are different, they have at least one antibody in common as the 8F10 is in capture in LFIA whereas it is in revelation in ELISA idem with regard to the detection of F1 for the 11C7. In any case, it is not surprising that the combinations of antibodies were different, sometimes they are even further apart or have a different number of antibodies.

Why the authors did not determine the LoD of the ELISA in serum just like for the LFIA ? This induces a non-negligible bias and rather in favor of the ELISA. Also why the strains have not been tested in ELISA. Here again, a determination of the smallest quantity of bacteria detected in ELISA and LFIA would have been a good complement to this study by being closer to patient conditions. The development of ELISA is as well as important as it could have a positive impact on the field. It allows as mentionned in the discussion an hightroughput implementation in laboratory in order to diagnose a lot of patient by using automation fluidhandling device. It could be of interest during a pandemia.

--------------------

Conclusions

-Are the conclusions supported by the data presented?

-Are the limitations of analysis clearly described?

-Do the authors discuss how these data can be helpful to advance our understanding of the topic under study?

-Is public health relevance addressed?

Reviewer #1: partly. see file attached

Reviewer #2: The conclusions are supported by the data presented. The public health relevance was adequately addressed.

Reviewer #3: The conclusions are supported by the data presented but all the limitations of analysis are not clearly described as some of them are due to the methodology excepted the absence of data on sample from patients clearly mentionned by the authors. They mentionned also that ELISA was only tested in spiked PBS. It seems that they improve the LoD compare to commercially available test but we have no comparaison in this study with commercialized one. it could have been interesting to have a comparison of the prototypes vs the tests marketed.

The authors clearly describe the benefit of such test in different context like surveillance in order to prevent outbreaks, to have tools to detect quickly Y.pestis in a context of bioweapon, to evaluate animal populations and vectors for the presence of Y.pestis and the fact they envisage also a potential use of such mabs for protecting from plague infection and for a therapeutic purpose with however no information of their capacity to neutralize the bacteria in this article. We estimate that at this stage the prototypes developed need more development including in analytic condition and before to be tested on a huge volume of samples from patients. Sure that the mabs seems to have a huge potential for such development. It will be a pity not to continue such developement. For a public health point of view the authors need to develop more how they envisage to continue the development and establish the clinical performance of such prototypes. The less sample from patients they will test during the development the more prototypes with different combinaisons of mabs they will need to test to determine the clinical performance. Selection of mabs pairs only by using recombinant or strains is not suficient to predict the performance on sample from patients. It is the reason why it is better to develop the test near the patient whatever the country. Beyond this we need to understand how they envisage a valorization with a product accessible to countries in development with a uniform performance from batch to batch. The fact they wish to publish is good to inform the feasibility to get such dual antigen test but it is not sufficient as the most interresting is how the authors will make it possible on the field to better manage plague outbreak. The establismhent of such plan is of interest.

The limitation of this study is not related only to the lack of results on sample from patients but also to the method used to develop these first prototypes. Furthermore, we can not visualize the effort made by the team to adjust the quantity of mabs in detection or in revelation, which would certainly make it possible to have more possible combinations to test on patient samples. Furthermore, a sensitivity and specificity study cannot be carried out without an adjustment phase. This adjustment phase balances the test to achieve the expected results on these two criteria. It seems that the authors have worked with equal amounts of conjugate and antibody on the membrane which will de facto rule out possible combinations. It seems that they made a choice of combinaison with no study on interferences appearance with individuals serum but more on sensitivity. The two are too much deconnected and the choice of the best combinaisons need to be performed not only on sensitivity criteria but also on specificity. The overall performance of the test is finnaly also influenced by the type, volume of sample and the type of implementation with chase buffer or not in a dispstick or casette format and as the team. The all forms an almost inseparable set of parts. As soon as you modify a parameter the whole performance must be re-evaluated. We do not perceive in this development the parameters which influence more the performance of test. We need to better understand in what context this test will be used. In addition, the development of the LFIA test must be reviewed taking into account that the matrix chosen to allow its use as close as possible to the patient need to be a whole blood matrix and that the test format need to be a cassette format for a better convenience and safety both for the operator and to guarantee the correct performance of the test. This does not prevent that the final test must be able to be used not only in whole blood but also in serum and plasma and that the desired performance of the analytical and clinical test must be achieved in each of these matrices. This means that in the context of this study, the performance established in serum by the team must therefore again be the subject of a study under conditions of implementation of the test more appropriate to the field. This new implementation may impact the work of selecting mabs to determine the best pairs which in any case have not been sufficiently explored from a repeatability point of view for both sensitivity and specificity. In all cases, the sample volumes must be compatible for each of the matrices with an implementation of the test without a chase buffer. Part of the evaluation could be done in serum and then confirm via spiking in whole blood but with volumes allowing these matrices to migrate without chase buffer with as a prerequisite for whole blood to block the red bood cells in the sample pad by playing on the porosity of this component or with the addition of mabs against red blood cells. This LFIA test should be further evaluated with individual sera to better determine the risk of interference, but also in whole blood with adifferent anticoagulants for plasma investigation.

It is globally difficult to evaluate the ways of optimizing this test because we do not have access to all the experiments, in particular if there has been a great effort to carry out on the adjustments of the conjugates and the quantities of mabs on the nitrocellulose membrane. We do not have more information if a sample volume study was generated and if it strongly impacted the sensitivity/specificity balance. Depending on the end-use context of the test, the sample volume can be adapted and participate in improving the performance of the test, hence the importance of testing several sample volumes both for the sensitivity study and specificity. One of the limitations of the study recognized by the authors is the absence of results on well characterized samples from patients and it is all the most important to verify that the feasibility of assaying LcrV in patients remains a question because the concentration levels during infection do not seem to be fully understood in humans. This is not the case for the dosage of F1 in patients where the concentrations are described between 4 and 40,000 ng / ml. A study in mice shows concentration levels of 6 to 26 ng / ml. If it turns out that the concentrations in humans of LcrV are of this order of magnitude, the development of the ELISA in serum and plasma is important to characterize the samples and as back up in case the LFIA test lacks of sensitivity. From a public health point of view, it is important that the authors mention how this test will be accessible for those who need it.

A nice article could be generated with the determination of an analytical and clinical performance carried out according to the rules of the art in ELISA and LFIA with in addition for each of these formats the development of an IgM detection test and IgG since the teams have the recombinant antigens. The development of a combo LFIA dual antigen test plus a strip for an IgG/IgM detection could also be a good way to catch cases not detected via the antigen test but via IgM level and in any case very useful for carrying out epidemiological studies in the field in order to best map the epidemics.

Beyond the development of the LFIA test, the development of the ELISA is not sufficiently highlighted, probably again due to lack of TPP. We have no analytical performance in serum either via the recombinants or with the bacterial strains either from a sensitivity or specificity point of view. The TPP need to be described to guide the development of the LFIA test and the ELISA test which could probably be more used in confirmation when PCR is not available or when breaks in raw materials happen as it was the case during the sars cov 2 pandemia.

--------------------

Editorial and Data Presentation Modifications?

Use this section for editorial suggestions as well as relatively minor modifications of existing data that would enhance clarity. If the only modifications needed are minor and/or editorial, you may wish to recommend “Minor Revision” or “Accept”.

Reviewer #1: (No Response)

Reviewer #2: Just some minor editorial suggestions:

Line 38: Full stop omitted after the Y for Y. pestis

Line 106: Define F1 minus strain at first use.

Line 117: Expression of F1 is temperature-induced at 33C and above.

Line 171: The word female should not be a capital. “Twenty Female CD1 mice..”.

Line 179: The word “or” should be “for”. “to account or variability”

Line 222: BSL2 should be BSL3. “biosafety level 3 (BSL2)”

Line 224: “Francisella tularensis” should be abbreviated to “F. tularensis”. First use of full scientific name was in line 218.

Line 225: “Bacillus anthracis” should be abbreviated to “B. anthracis”. First use of full scientific name was in line 219.

Line 246: Write out SDS.

Line 262: “ten” should be in numerals. Numbers from ten and above should be written as numerals.

Line 266: “40uL”. Don’t start a sentence with numerals.

Line 294: “2” should be written in words. Numbers from zero to nine should be written in words.

Line 349: “3” should be written in words. As above.

Line 603-809: Check reference list. Full names of authors not listed for a number of publications i.e. references 17 (line 639), 20 (line 648), 21 (line 651), 33 (line 685), 47 (line 725), 54 (line 745), 58 (line 756), 59 (line 760), 62 (line 768), 65 (line 775), 66 (line 778), etc.

Lines 205, 227, 238. 259, 260, 262, 263, 399, 432: There should be a space between the number and the unit of measure e.g. line 205 - 450 nm, line 259 - 1 mg/mL etc.

Reviewer #3: (No Response)

--------------------

Summary and General Comments

Use this section to provide overall comments, discuss strengths/weaknesses of the study, novelty, significance, general execution and scholarship. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. If requesting major revision, please articulate the new experiments that are needed.

Reviewer #1: (No Response)

Reviewer #2: The manuscript describes the development of a dual target (F1 and LcrV) lateral flow immunoassay for detecting Yersinia pestis. There are publications describing the development of single target (F1 antigen) RDTs, but not a dual target LFI. To my knowledge, there is only one commercial RDT targeting dual antigens (F1 and V), but it is not approved for human use.

The LFI developed in this study was only tested against simulated serum samples. As mentioned by the authors, further evaluation of the multiplexed LFI is needed on clinical samples collected from plague patients. The paper is well written and adds to the body of knowledge on the topic.

Reviewer #3: The development of the LFIA and ELISA tests is not far enough and results obtained too much preliminary. The methodology used to develop the tests whatever it is the LFIA or ELISA should be reviewed and should start with writing a TPP in order to define the ideal test needed on the field and/or in the laboratory (cf previous sections for more details).

This work is not completed enough to allow a third party to perform a proper valuation of these prototypes on samples from patients. The authors need to continue their effort in order to accelerate the provision of a sufficiently developed prototype, adapted to the field with a well defined analytical performance (sensitivity and specificity) on recombinants and strains in order to use the precious and rare samples from patients on the best prototypes to verify its clinical performance. Moreover the authors will have to ensure very quickly the performance of the prototypes under development on samples from patients and it seems crucial for the detection of LcrV. It is even preferable to develop on samples from patient but probably the constraints to develop test for plague on such samples are high. The valuation of this work will go through the provision of mabs and the formulation of the prototypes developed and this point is not sufficiently addressed in the discussion. The team need also to describe the collaboration they need to build in order to establish the clinical performance of their prototypes and its added value vs commercialized available tests. Finally in case they are successful they need to establish an access plan in order that patients can benefit of such test that can save their life by being diagnose earlier and therefore treat earlier.

--------------------

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

Figure Files:

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org.

Data Requirements:

Please note that, as a condition of publication, PLOS' data policy requires that you make available all data used to draw the conclusions outlined in your manuscript. Data must be deposited in an appropriate repository, included within the body of the manuscript, or uploaded as supporting information. This includes all numerical values that were used to generate graphs, histograms etc.. For an example see here: http://www.plosbiology.org/article/info%3Adoi%2F10.1371%2Fjournal.pbio.1001908#s5.

Reproducibility:

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Submitted filename: Comments.docx

Click here for additional data file.

31 Jan 2022

Submitted filename: 20220124_ResponsestoReviewers.docx

Click here for additional data file.

28 Feb 2022

Dear Dr. AuCoin,

We are pleased to inform you that your manuscript 'Development of a dual antigen lateral flow immunoassay for detecting Yersinia pestis' has been provisionally accepted for publication in PLOS Neglected Tropical Diseases.

Before your manuscript can be formally accepted you will need to complete some formatting changes, which you will receive in a follow up email. A member of our team will be in touch with a set of requests.

Please note that your manuscript will not be scheduled for publication until you have made the required changes, so a swift response is appreciated.

IMPORTANT: The editorial review process is now complete. PLOS will only permit corrections to spelling, formatting or significant scientific errors from this point onwards. Requests for major changes, or any which affect the scientific understanding of your work, will cause delays to the publication date of your manuscript.

Should you, your institution's press office or the journal office choose to press release your paper, you will automatically be opted out of early publication. We ask that you notify us now if you or your institution is planning to press release the article. All press must be co-ordinated with PLOS.

Thank you again for supporting Open Access publishing; we are looking forward to publishing your work in PLOS Neglected Tropical Diseases.

Best regards,

Anne-Sophie Le Guern

Guest Editor

PLOS Neglected Tropical Diseases

Javier Pizarro-Cerda

Deputy Editor

PLOS Neglected Tropical Diseases

***********************************************************

Reviewer's Responses to Questions

Key Review Criteria Required for Acceptance?

As you describe the new analyses required for acceptance, please consider the following:

Methods

-Are the objectives of the study clearly articulated with a clear testable hypothesis stated?

-Is the study design appropriate to address the stated objectives?

-Is the population clearly described and appropriate for the hypothesis being tested?

-Is the sample size sufficient to ensure adequate power to address the hypothesis being tested?

-Were correct statistical analysis used to support conclusions?

-Are there concerns about ethical or regulatory requirements being met?

Reviewer #1: -Are the objectives of the study clearly articulated with a clear testable hypothesis stated?

-Is the study design appropriate to address the stated objectives?

more or less.

The authors have not addressed the main criticism : they did not use the antigens in their native state for the evaluation of their dual LFIA. They have used bacterial lysates thermally inactivated (80°C, 2h) or recombinant proteins. Sensitivity of the tests is thus based on proteins that do not reflect the native proteins. Evaluation of the performance of a test (and particularly the sensitivity) is a major point when this type of tests is developed, and must be as close as possible of the natural native antigen.

**********

Results

-Does the analysis presented match the analysis plan?

-Are the results clearly and completely presented?

-Are the figures (Tables, Images) of sufficient quality for clarity?

Reviewer #1: see above

**********

Conclusions

-Are the conclusions supported by the data presented?

-Are the limitations of analysis clearly described?

-Do the authors discuss how these data can be helpful to advance our understanding of the topic under study?

-Is public health relevance addressed?

Reviewer #1: see above

The

**********

Editorial and Data Presentation Modifications?

Use this section for editorial suggestions as well as relatively minor modifications of existing data that would enhance clarity. If the only modifications needed are minor and/or editorial, you may wish to recommend “Minor Revision” or “Accept”.

Reviewer #1: (No Response)

**********

Summary and General Comments

Use this section to provide overall comments, discuss strengths/weaknesses of the study, novelty, significance, general execution and scholarship. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. If requesting major revision, please articulate the new experiments that are needed.

Reviewer #1: The main criticism was not addressed by the authors. The article is based on the performance evaluation of a test that could be used for human clinical field diagnosis. The sensitivity of the test is evaluated with antigens that do not reflect the reality of antigens that would be present in real life, nor are there any clinical samples.

The authors argued that they could not use culture supernatants with the excuse that the culture supernatants do not contain enough antigen. I know (because I have already done it) that the bacteria Y. pestis after growing can be centrifuged, resuspended in a small volume of medium or PBS to concentrate them. After pipetting of the bacterial pellet back and forth several times, the F1 and LcrV antigens (that are not tightly attached to the bacteria) are found soluble in the supernatant and can be harvested after centrifugation (they are concentrated in the supernatant (which can then be filtered for safety).

**********

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

18 Mar 2022

Dear Dr. AuCoin,

We are delighted to inform you that your manuscript, "Development of a dual antigen lateral flow immunoassay for detecting Yersinia pestis," has been formally accepted for publication in PLOS Neglected Tropical Diseases.

We have now passed your article onto the PLOS Production Department who will complete the rest of the publication process. All authors will receive a confirmation email upon publication.

The corresponding author will soon be receiving a typeset proof for review, to ensure errors have not been introduced during production. Please review the PDF proof of your manuscript carefully, as this is the last chance to correct any scientific or type-setting errors. Please note that major changes, or those which affect the scientific understanding of the work, will likely cause delays to the publication date of your manuscript. Note: Proofs for Front Matter articles (Editorial, Viewpoint, Symposium, Review, etc...) are generated on a different schedule and may not be made available as quickly.

Soon after your final files are uploaded, the early version of your manuscript will be published online unless you opted out of this process. The date of the early version will be your article's publication date. The final article will be published to the same URL, and all versions of the paper will be accessible to readers.

Thank you again for supporting open-access publishing; we are looking forward to publishing your work in PLOS Neglected Tropical Diseases.

Best regards,

Shaden Kamhawi

co-Editor-in-Chief

PLOS Neglected Tropical Diseases

Paul Brindley

co-Editor-in-Chief

PLOS Neglected Tropical Diseases