Virus-like particle vaccine with B-cell epitope from porcine epidemic diarrhea virus (PEDV) incorporated into hepatitis B virus core capsid provides clinical alleviation against PEDV in neonatal piglets through lactogenic immunity.

Virus-like particle vaccine with B-cell epitope from porcine epidemic diarrhea virus (PEDV) incorporated into hepatitis B virus core capsid provides clinical alleviation against PEDV in neonatal piglets through lactogenic immunity

Yi Lu a, Sherrie Clark-Deener b, Frank Gillam a, Connie Lynn Heffron c, Debin Tian c, Harini Sooryanarain c,

Tanya LeRoith c, Jessica Zoghby d, Mallori Henshaw d, Steven Waldrop d, Jeremy Pittman e, Xiang-Jin Meng c,

Chenming Zhang a,⇑ a Department of Biological Systems Engineering, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061, USA b Department of Large Animal Clinical Sciences, Virginia-Maryland College of Veterinary Medicine, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061, USA c Department of Biomedical Sciences and Pathobiology, Virginia-Maryland College of Veterinary Medicine, Virginia Polytechnic Institute and State University, Blacksburg, VA

24061, USA d Doctor of Veterinary Medicine Professional Program, Virginia-Maryland College of Veterinary Medicine, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061, USA e Smithfield Foods, Inc., 434 E Main St., Waverly, VA 23890, USA a r t i c l e i n f o

Article history:

Received 7 March 2020 Received in revised form 25 May 2020

Accepted 4 June 2020 Available online xxxx Keywords:

Virus-like particle (VLP) Hepatitis B virus core antigen (HBcAg)

Porcine epidemic diarrhea virus (PEDV) Epitope Viral Neutralization

Pigs a b s t r a c t Porcine epidemic diarrhea virus (PEDV) has had a negative economic impact on the global swine industry for decades since its first emergence in the 1970s in Europe. In 2013, PEDV emerged for the first time in the United States, causing immense economic losses to the swine industry. Efforts to protect U.S. swine herds from PEDV infection and limit PEDV transmission through vaccination had only limited success so far. Following the previous success in our virus-like particle (VLP) based vaccine in mouse model, in this study we determined the immunogenicity and protective efficacy of a VLP-based vaccine containing

B-cell epitope 748YSNIGVCK755 from the spike protein of PEDV incorporated into the hepatitis B virus core capsid (HBcAg), in a comprehensive pregnant gilt vaccination and piglet challenge model. The results showed that the vaccine was able to induce significantly higher virus neutralization response in gilt milk, and provide alleviation of clinical signs for piglets experimentally infected with PEDV. Piglets from preg- nant gilt that was vaccinated with the VLP vaccine had faster recovery from the clinical disease, less small intestinal lesions, and higher survival rate at 10 days post-challenge (DPC).

2020 Elsevier Ltd. All rights reserved.

1. Introduction Since its first emergence in Europe in the 1970s [1] and its introduction in 2013 into the United States [2], porcine epidemic diarrhea virus (PEDV) has shown to be highly transmittable with worldwide outbreaks and increasing genetic diversity [3], causing immense economic losses to the global swine industry [4]. The cur- rent available vaccines had only limited success against PEDV so far. Modified live-attenuated vaccines (MLVs) tend to have better efficacy, but long-term safety of MLVs remains a concern [5]. Killed virus-based vaccines often offer only partial protection in neonatal piglets [6] with uncertainty in their cross-protection ability against heterologous strains [7].

In order to develop a safe, efficacious vaccine against PEDV, there is a renewed interest in the subunit vaccine approach [5].

Several studies have examined the possibility of using S1 domain from the PEDV spike protein as a potential vaccine, but none was able to provide neonatal piglets with a complete protection from clinical disease [8–10]. The spike protein of PEDV has been the tar- get for vaccine development, because its large structural projec- tions of 18–23 nm [3] is critical for viral attachment and entry during PEDV infection, and a major antigen for eliciting neutraliz- ing antibodies in pigs [11–13]. https://doi.org/10.1016/j.vaccine.2020.06.009

0264-410X/
2020 Elsevier Ltd. All rights reserved.

⇑Corresponding author.

E-mail addresses: lewislu@vt.edu (Y. Lu), sherrie@vt.edu (S. Clark-Deener), fbgillam@vt.edu (F. Gillam), cheffron@vt.edu (C.L. Heffron), debint@vt.edu (D. Tian), harinis@vt.edu (H. Sooryanarain), tleroith@vt.edu (T. LeRoith), jesslz10@vt.edu (J.

Zoghby), mallori5@vt.edu (M.

Henshaw), stevenw3@vt.edu (S.

Waldrop), jpittman@smithfield.com (J. Pittman), xjmeng@vt.edu (X.-J. Meng), cmzhang@ vt.edu (C. Zhang).

Vaccine xxx (xxxx) xxx Contents lists available at ScienceDirect

Vaccine journal homepage: www.elsevier.com/locate/vaccine

Please cite this article as: Y. Lu, S. Clark-Deener, F. Gillam et al., Virus-like particle vaccine with B-cell epitope from porcine epidemic diarrhea virus (PEDV) incorporated into hepatitis B virus core capsid provides clinical alleviation against PEDV in neonatal piglets through lactogenic immunity, Vaccine, https:// doi.org/10.1016/j.vaccine.2020.06.009

Instead of using the complete domains of the PEDV spike pro- tein as candidate vaccines, we took a unique approach by incorpo- rating only the short previously-identified

B-cell epitope 748YSNIGVCK755 on the spike protein [14] into the hepatitis B virus capsid protein (HBcAg) recombinantly in order to form a virus-like particle (VLP). It is well known that VLPs are more potent vaccine candidates than subunit antigens, since VLPs resemble infectious viruses. HBcAg has been studied extensively for its application as chimeric VLP vaccines in combating infectious diseases [15–17], and more recently, cancers [18,19]. Each HBcAg VLP is composed of 240 monomeric units when the assembled icosahedral capsid has a triangulation number of 4 [20]. By inserting two copies of the 8-amino acid epitope into the HBcAg backbone (Fig. 1A), we can present this PEDV epitope antigen 480 times on just a single

VLP particle, thereby with the potential of eliciting potent humoral immune response against PEDV.

This chimeric VLP vaccine platform has been tested in mice through our previous work, where significantly high virus neutral- izing (VN) antibody titer was induced in immunized mouse sera [21,22]. Two a.a. substitutions with cysteine were implemented in the HBcAg backbone at D29 and R127 for this design (Fig. 1A), which had been proven to improve VLP particle stability [23] and vaccine efficacy [22] due to the formation of additional disulfide linkages.

The newborn piglets are most vulnerable to PEDV infection with almost 100% mortality rate in 1–3-day old piglets [3], because neonatal piglets would not have sufficient time to develop immu- nity against PEDV from direct vaccination, even if an effective vac- cine were available. Therefore, our goal is to stimulate robust

PEDV-specific viral neutralizing antibody response first in preg- nant gilts, then through lactogenic immunity, passively transferred to neonatal piglets and protect them from subsequent PEDV infection.

In this piglet challenge study, we evaluated the efficacy of the

VLP-based vaccine candidate, both in terms of its ability to stimu- late systemic and lactogenic VN responses in pregnant gilts, and more importantly, in piglets against PEDV.

2. Methods 2.1. Plasmid and cells The DNA sequence of the PEDV spike protein B-cell epitope (50-TACTCTAACATCGGTGTTTGCAAA-30) was synthesized by IDT (Coralville, IA), and cloned into the mutated HBcAg backbone including D29C and R127C substitutions on a pET-28a(+) plasmid (Novagen, Madison, WI) using overlap extension PCR. Plasmid containing the vaccine design was then transformed into T7

Express Competent E. coli (NEB, Ipswich, MA). Sanger sequencing was used to verify cell line integrity post transformation and before long term cryo storage at -80 C.

2.2. Vaccine antigen expression E. coli cells were grown in 2.8 L shake flasks with 2YT media and 30 mg/mL kanamycin at 37 C, shaking at 200 rpm, after inoc- ulation of 0.2% (v/v) overnight culture that was grown under the same condition. The culture was induced with 1 mM isopropyl-b- D-thiogalactopyranoside (IPTG) once the OD600 reached 0.6–0.8.

The incubation temperature was then lowered to 28 C for over- night protein expression. Cells were pelleted by centrifugation the next morning before storage at -20 C for further processing.

2.3. Purification Vaccine antigen purification followed a similar procedure as described earlier [22]. Briefly, inclusion bodies were solubilized using solubilization buffer (20 mM Na2HPO4, 50 mM NaCl, 2 M urea, 0.9% sarkosyl) after cell lysis. Solubilized protein was then loaded onto a column packed with diethylaminoethyl (DEAE)

Sepharose Fast Flow anion exchange chromatography resin (GE

Healthcare, Marlborough, MA), followed by an immobilized metal ion affinity chromatography (IMAC) packed with IMAC Sepharose

6 Fast Flow resin (GE Healthcare). NiSO4 (200 mM) was used to charge the IMAC column. IMAC eluate was buffer exchanged into

Fig. 1. Vaccine design and characterizations. (A) Schematic diagram of vaccine design with a MW of 21 kDa. Mutated HBcAg includes D29C and R127C substitutions (mHBcAg) with poly-histidine (6-His) tag attached to the N-terminus that enables purification by immobilized metal affinity chromatography (IMAC). Represented by ‘‘S” in the figure is the B-cell epitope 748YSNIGVCK755 from PEDV spike protein, which is inserted at both the mHBcAg major immunodominant region (MIR) between a. a. 79 and 80, and the N-terminus of the His tag. (B) SDS PAGE of cell lysate sample (lane 2) and purified sample (lane 3). Protein standard is in lane 1. MW unit is kDa. (C) Western blot on lysate sample (lane 2) and purified sample (lane 3) using anti-6-His as the primary antibody for target protein detection. Protein standard is in lane 1. MW unit is kDa. (D)

TEM image of purified vaccine sample stained with 2% uranyl acetate. The image was taken with a maximum magnification of 300,000. Red arrows point to the virus-like particles (VLPs). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

2 Y. Lu et al. / Vaccine xxx (xxxx) xxx Please cite this article as: Y. Lu, S. Clark-Deener, F. Gillam et al., Virus-like particle vaccine with B-cell epitope from porcine epidemic diarrhea virus (PEDV) incorporated into hepatitis B virus core capsid provides clinical alleviation against PEDV in neonatal piglets through lactogenic immunity, Vaccine, https:// doi.org/10.1016/j.vaccine.2020.06.009

1PBS with 0.9% sarkosyl at pH 7.4, followed by stepwise dialysis to gradually decrease sarkosyl concentration to 0% to aid VLP for- mation. Unassembled protein was eliminated using Amicon ultra- centrifugal filter with 100 kDa MWCO (Millipore, Danvers, MA).

2.4. VLP assembly VLP formation was confirmed first with dynamic light scatter- ing (DLS) using Zetasizer Nano (Malvern, Malvern, UK). Subse- quently, transmission electron microscopy (TEM) was used to further characterize VLP particle assembly. Uranyl acetate (2%) was used to negatively stain the analyte before examination with

JEM 1400 (JEOL, Peabody, MA) under 300,000 magnification.

2.5. SDS PAGE Protein purity and target protein molecular weight (MW) were characterized using NuPAGE 4–12% Bis-Tris Protein Gel (Invitro- gen, Carlsbad, CA). Samples were treated at 75 C for 20 min before loading onto the gel. Loaded gel was run at constant voltage of

200 V for 35 min in 1 NuPAGE MES SDS Running Buffer (Invitro- gen). Gel was then stained with SimplyBlue SafeStain (Invitrogen) for 1 h after 3 times 5-minute water wash. Destain took place over- night before imaging with ChemiDoc Imaging System (Bio-Rad,

Hercules, CA). Image Lab (Bio-Rad) was used to confirm product purity and verify target protein’s MW. Precision Plus Protein Stan- dards (Bio-Rad) were used as MW references.

2.6. Western blot analysis Anti-6-His tag antibody was used to probe the target vaccine protein in both cell lysate and purified vaccine samples using Wes- tern blot. Following the SDS PAGE procedure as described with Pre- cision Plus Protein WesternC Standards (Bio-Rad) loaded as the reference, proteins from the gel were then blotted onto a 0.2 mm nitrocellulose membrane (Bio-Rad) using Trans-Blot Turbo Blotting

System (Bio-Rad). The membrane was first washed with TBS and then blocked in the blocking solution (TBS-T with 0.05% Tween

20 and 5% nonfat milk) for 1 h. After washing, the membrane was incubated in 6-His Tag Monoclonal Antibody (Invitrogen) (1:3,000 dilution) for 1 h, then in Goat Anti-Mouse IgG-HRP Conju- gate (Millipore) (1:5,000 dilution) and

Precision Protein StrepTactin-HRP Conjugate (Bio-Rad) (1:10,000 dilution) for 1 h.

Between and after each incubation, the membrane was washed with TBS-T three times, each for 5 min. Signal was then developed with Clarity Max Western ECL Substrate (Bio-Rad) and captured by

ChemiDoc Imaging System (Bio-Rad).

2.7. Endotoxin detection Endotoxin level was determined to be below 1 EU/mL for the purified vaccine using Chromogenic Endotoxin Quant Kit (Thermo

Scientific, Waltham, MA) following manufacturer’s instructions.

2.8. Vaccine formulation Syringe filters with 30 mm diameter and 0.10 mm pore size (Celltreat, Pepperell, MA) were used for sterile filtration of each vaccine dose. Every individual dose was formulated in a 2 mL

PBS mixture for intramuscular injection in pigs. The first 3 injec- tions included squalene-based oil-in-water adjuvant AddaVax (InvivoGen, San Diego, CA) for both groups, 200 mg of VLP for Group

2. The 4th and final injection included 600 mg of VLP for Group 2, and did not include AddaVax for both groups. Formulation for each injection is detailed in Table 1. All injections were prepared the night before administration and stored at 4 C.

2.9. Challenge virus Vero (African green monkey kidney) cells cultured in DMEM (Gibco, Waltham, MA) supplemented with 10% FBS (Gibco) and

1 Antibiotic-Antimycotic (Gibco) were used to propagate the

PEDV 2013 Colorado strain (commercially purchased from the

National Veterinary Services Laboratories, Ames, IA). After virus inoculation, cells were cultured in MEM (Gibco) supplemented with 0.3% tryptose phosphate broth (Gibco), 0.02% yeast extract (Sigma Aldrich, St. Louis, MO), 2 mg/mL trypsin (Gibco), and 1

Antibiotic-Antimycotic (Gibco). After 4 days of incubation at

37 C with 5% CO2, the propagated viruses were collected through

3 cycles of freeze and thaw. Cell lysate along with culture super- natant was stored at -80 C after centrifugation at 3,000g for

10 min at 4 C.

To determine the infectious titer of the propagated virus, 100 mL serially-diluted viral stock (from 10-1 to 10-5) was added into each well on a 96-well plate containing confluent monolayers of Vero cells. Three days after incubating at 37 C with 5% CO2, cells were fixed with methanol. Immunofluorescence assay (IFA) was done with mouse anti- PEDV N IgG (Medgene Labs, Sioux Falls, SD) and anti-mouse IgG-Alexafluor 594 both at a dilution of 1:500.

The primary and secondary antibody incubations were both done at 37 C for 1 h. Wells with fluorescent foci were considered posi- tive. Infectious viral titer was then calculated using the Reed- Muench method [24]. The virus stock was diluted to a titer of

105 TCID50/2 mL for the piglet challenge study.

2.10. Experimental design for animal study All procedures pertaining to the animal study were approved by

Virginia Tech Institutional Animal Care and Use Committee (IACUC protocol 19-092). Four White/Landrace cross genetic breed preg- nant gilts were obtained from a farm that was negative for both

PEDV and PRRSV (porcine reproductive and respiratory syndrome virus). They were divided into 2 groups as described in Table 1 with 2 gilts in each group. All gilts received the first 3 intramuscu- lar injections which were 2 weeks apart starting at 6 weeks before farrowing. A final booster injection was administered 1 week before farrowing. Blood samples were collected via jugular venipuncture before each injection and one additional time at

3 days before farrowing. All gilts were induced to farrow on the same day. Colostrum samples were collected on the day of farrow- ing. Milk samples were collected on day 3 post-farrowing (DPF).

Additional piglets were euthanized at 3 DPF so that each group had 10 piglets (5 piglets/gilt). The 20 remaining piglets were then challenged orally with PEDV 2013 Colorado stain at 4 DPF at a dose of 105 TCID50/ piglet. Fecal swab materials were collected from each piglet daily for up to 5 days post-challenge (DPC). All animals were monitored daily for clinical signs (activity, body condition, and diarrhea) until 9 DPC. Clinical scores ranging from 1 to 3 were recorded, with 1 being normal, and 3 being severely ill. If a score of

3 was reached for any of the three categories (activity, body condi- tion, and diarrhea), then the animal was euthanized and necropsy was performed. All surviving piglets were euthanized for final necropsy at 10 DPC.

2.11. Virus neutralization (VN) assay PEDV neutralizing activity in pig serum and colostrum/milk was evaluated by South Dakota State University’s Animal Disease

Research & Diagnostic Lab through a fluorescent focus neutraliza- tion assay (FFN) [25]. Briefly, for pig serum, heat inactivated sam- ples were 1:2 serially-diluted with 1:20 as the starting dilution before mixing with PEDV 2013 Colorado strain at a concentration of 100 foci forming units/100 mL for a 1-hour incubation at 37 C.

Y. Lu et al. / Vaccine xxx (xxxx) xxx 3 Please cite this article as: Y. Lu, S. Clark-Deener, F. Gillam et al., Virus-like particle vaccine with B-cell epitope from porcine epidemic diarrhea virus (PEDV) incorporated into hepatitis B virus core capsid provides clinical alleviation against PEDV in neonatal piglets through lactogenic immunity, Vaccine, https:// doi.org/10.1016/j.vaccine.2020.06.009

Then the mixture was added to a plate containing confluent monolayers of Vero cells for a 2-hour initial incubation followed by a 24-hour incubation at 37 C with a wash step in between.

The plate was then fixed with 80% acetone and stained with

FITC-conjugated mAb SD6-29 to probe the infected cells. Virus neutralizing titer was reported as the highest dilution at which at least 90% reduction in the number of fluorescent foci was observed versus assay control. For pig colostrum/milk, 5 mg/mL rennet (Sigma Aldrich) was added to samples before a 30-minute incubation at 37 C. Coagulant was sedimented by centrifuga- tion at 2,000g for 15 min at 4 C. The resulting whey sample was then tested by the same procedure for pig serum as described above.

2.12. Quantitative PCR Piglet fecal swab materials were tested for PEDV RNA load up to

5 DPC by South Dakota State University’s Animal Disease Research

& Diagnostic Lab using a quantitative PCR (qPCR) assay [26].

Briefly, 7 mL of extracted RNA sample from each fecal swab was mixed with 18 mL assay master mix before the initial reverse tran- scription step (15 min at 48 C) and subsequent inactivation step (2 min at 95 C). Thirty-eight amplification cycles (5 s at 95 C fol- lowed by 40 s at 60 C) were then performed with PEDV positive control set at 38 cycles. Cycle threshold (Ct) for each sample was then reported.

2.13. Gross pathology and histology Gross small intestine pathology and colon content were evalu- ated at the time of piglet necropsy. A score between 1 and 3 was assigned. For small intestine gross pathology, 1 is normal, 2 is either thin walled or gas-distended small intestine, while 3 is both thin walled and gas-distended small intestine. For colon fecal con- tent, 1 is solid or pasty feces, 2 is semi watery feces, while 3 is watery feces with no solid content. Small intestine tissue was also collected during necropsy and fixed in formalin for histological analysis. Hematoxylin & eosin (H&E) slides were subsequently pre- pared from fixed and sectioned tissues. Villous length (v) and crypt depth (c) were measured at 10 different sites on each jejunum slide section from each piglet by a board-certified veterinary pathologist who was blinded to different treatment groups. The average v to c ratio was calculated. A lower v to c ratio indicates more severe intestinal lesion while a higher v to c ratio shows better small intestinal health.

2.14. Statistical analysis Prism 6 (GraphPad, San Diego, CA) was used to perform all sta- tistical analysis on data collected from this study. Unpaired t test was used to determine statistically significant differences between groups with a significance level (a) of 0.05.

3. Results 3.1. VLP vaccine production and characterization

The DNA sequence of the vaccine construct (Fig. 1A) was veri- fied by Sanger sequencing. An expected protein of 21 kDa was revealed at the correct MW on both SDS PAGE (Fig. 1B) and Wes- tern blot (Fig. 1C). The product purity was above 93% for the puri- fied vaccine. This included only the monomer band at 21 kDa. If the dimer band at 42 kDa had been included as well, the product purity would have been even higher. The VLP particle assembly was visu- alized by TEM (Fig. 1D). Before administration to animals, the endotoxin level for the purified vaccine was measured and found to be below 1 EU/mL.

3.2. Virus neutralization The viral naturalization (VN) titer from gilt sera remained below the detection limit (<1:20 dilution) 4 days after the 4th and final injection for all groups (data not shown). However, VN titer grad- ually increased post-farrowing in gilt colostrum and milk for the

Vaccine group (Group 2). With a mean VN titer of 120 three days post farrowing, it was significantly higher (p  0.05) than the

PBS control (Fig. 2).

3.3. Clinical evaluations Once the piglets were challenged with PEDV, they were evalu- ated daily up to 9 DPC across 3 clinical categories (activity, body condition, and diarrhea) with a score from 1 to 3, 1 being normal, and 3 being severely ill (Fig. 3). For activity, the clinical sign started to develop for both groups at 2 DPC. The Vaccine group (Group 2)

Fig. 2. Virus neutralizing (VN) antibody titers against PEDV 2013 Colorado strain.

Gilt colostrum samples collected on the day of farrowing (0 DPF) and milk samples collected 3 days post-farrowing (3 DPF) were tested. Antibody titer was calculated as the highest dilution at which 90% or greater reduction in the number of fluorescent foci was observed versus assay control. Statistical significance is evaluated using unpaired t test with right tailed hypothesis testing. Error bars indicate ± SD for each group on each day. p values are represented as *p  0.05,

**p  0.01, ***p  0.001, ****p  0.0001.

Table 1 Immunization schedule for pregnant gilts.

Group Weeks prior to farrowing 6 weeksa 4 weeksa 2 weeksa

1 weekb 1 PBS PBS PBS PBS 2 Vaccine Vaccine Vaccine

Vaccine a All injections in both groups were formulated with 1 mL of AddaVax. 200 mg VLP vaccine was included in Group 2. b All injections in both groups were formulated without AddaVax. 600 mg VLP vaccine was included in Group 2.

4 Y. Lu et al. / Vaccine xxx (xxxx) xxx Please cite this article as: Y. Lu, S. Clark-Deener, F. Gillam et al., Virus-like particle vaccine with B-cell epitope from porcine epidemic diarrhea virus (PEDV) incorporated into hepatitis B virus core capsid provides clinical alleviation against PEDV in neonatal piglets through lactogenic immunity, Vaccine, https:// doi.org/10.1016/j.vaccine.2020.06.009 had significantly lower (p  0.05) score when compared to the PBS control at 8 DPC, indicating improved activity and alertness in the vaccinated animals at 8 DPC (Fig. 3A). Clinical signs for body con- dition (spinous processes and hook bone visibility) gradually developed through the initial 5 DPC for both groups. The Vaccine group showed significant improvement (p  0.05) at 8 DPC and 9

DPC when compared to the PBS control (Fig. 3B). Both the PBS con- trol and the Vaccine group started to developed semi-watery diar- rhea at 1 DPC. Piglets in the Vaccine group started to improve in fecal health at 5 DPC, while the PBS control did not start the recov- ery process until 7 DPC (Fig. 3C). Statistical significance was not evaluated for diarrhea scores since fecal scoring could not be car- ried out due to the lack of defecation at the time of evaluation for various groups on different days. However, with continuous improvement in average diarrhea scores starting at 5 DPC, the Vac- cine group had numerically lower mean scores than the PBS con- trol starting at 6 DPC, indicating a faster recovery.

No statistical significance was detected between the PBS control and the Vaccine group for viral RNA loads at 1, 3, and 5 DPC due to the large intra-group variability (Fig. 4). At 3 DPC, the Vaccine group had a numerically higher mean Ct value than the PBS control.

The survival rate at 10 DPC for the PBS control was 30%, while the Vaccine group had a survival rate of 60% (Fig. 5).

3.4. Gross pathology and histology Although no statistical significance was detected between the

PBS control and the Vaccine group for either small intestine gross pathology or colon fecal content, the PBS control group exhibited numerically higher average score in both categories when com- pared to the Vaccine group (Fig. 6).

For histological analysis, different degrees of lesion were visual- ized in representative microscopy images of jejunum H&E slides from both groups (Fig. 7A). The Vaccine group had a significantly higher (p  0.05) jejunum villous length to crypt depth ratio when compared to the PBS control (Fig. 7B), signifying a less severe small intestinal lesion for piglets in the Vaccine group.

4. Discussion Subunit vaccine has its appeal in terms of safety due to the lack of viral genetic materials, but, like all the other alternative vaccine approaches, the issue of vaccine efficacy is critically important and still needs to be addressed. Our unique approach in this study was to take advantage of HBcAg’s polymeric viral capsid structure for repetitive PEDV-specific antigen presentation, in turn, eliciting robust immunogenicity in the host.

Our previous studies in mice during the development of this

VLP-based vaccine demonstrated a strong virus neutralizing (VN) ability in vitro by mouse sera [21,22]. The high VN titer from

Fig. 3. Clinical signs monitored up to 9 days post-challenge (DPC). Each piglet was monitored daily for clinical signs in terms of activity, body condition, and diarrhea, then given a score from 1 to 3. (A) Activity scores. 1, normal, bright, and alert. 2, dull, depressed, and lethargic. 3, recumbent, unresponsive. (B) Body condition scores. 1, undetectable spinous processes and hook bones. 2, spinous processes and hook bones were slightly felt. 3, spinous processes and hook bones were easily felt and visible.

Statistical significance is tested using unpaired t test for each day. Error bars indicate ± SD for each group on each day. p values are represented as *p  0.05, **p  0.01,

***p  0.001, ****p  0.0001. (C) Diarrhea scores. 1, normal to pasty feces. 2, semi-liquid diarrhea with some solid content. 3, liquid diarrhea with no solid content. Due to missing scores from different groups on different days since it could not be judged if piglet did not defecate at the time of scoring, statistical significance was not evaluated for diarrhea scores.

Fig. 4. Viral RNA load in piglet feces. Fecal swab materials collected from all surviving piglets from 1, 3, and 5 days post-challenge (DPC) were tested for PEDV viral RNA load through qPCR. RNA load is reported as cycle threshold (Ct). Statistical significance is tested using unpaired t test for each day. Error bars indicate ± SD for each group on each day.

Fig. 5. Piglet survival. 10 days post-challenge (DPC) was the end of the study and the day of final necropsy.

Y. Lu et al. / Vaccine xxx (xxxx) xxx 5 Please cite this article as: Y. Lu, S. Clark-Deener, F. Gillam et al., Virus-like particle vaccine with B-cell epitope from porcine epidemic diarrhea virus (PEDV) incorporated into hepatitis B virus core capsid provides clinical alleviation against PEDV in neonatal piglets through lactogenic immunity, Vaccine, https:// doi.org/10.1016/j.vaccine.2020.06.009 potential candidates (mean VN titer around 370 two weeks after

3rd injection) provided a strong justification for this pig study.

When the initially planned 3-dose immunization regime of pregnant gilts did not produce detectable PEDV neutralizing anti- body response in sera for the Vaccine group, we then administered a 4th injection to all the gilts with higher amount of VLP vaccine (600 mg) for each gilt in the test group (Group 2) 1 week before farrowing (Table 1). We removed AddaVax in the formulation for the 4th injection to eliminate any potential interference of the adjuvant on the VLP structural integrity. At 3 DPF, which was

10 days after the 4th injection, the Vaccine group (Group 2) had a mean VN titer of 120 in pig milk, which is significantly higher (p  0.05) than the PBS control (Fig. 2). The continuation of rela- tively flat VN response up till the day of farrowing (0 DPF), comb- ing with the observation of steady decline of VN titer in sow milk post-farrowing in previous subunit PEDV vaccine study [10], led us to believe that the 4th injection with higher vaccine dose had an effect in boosting the VN response. Both live and killed virus- based vaccines had exhibited differences in VN response between

20-week old pigs and 8-week old pigs when young pigs were used as a surrogate model in another PEDV vaccine study [6]. Dosage dependency was one of the most likely causes. The vaccine dose to body weight ratio difference between BALB/C mice and pregnant gilts could provide an explanation for the large differences in VN response when compared to our previous mouse studies, not ignoring species specific characteristics.

Our initial vaccine dosage design of 200 mg per injection was largely comparable with other subunit PEDV vaccine studies in sows [8–10]. However, considering there are 2 repeats of short spike protein epitope 748YSNIGVCK755 on every HBcAg monomer (21 kD), we expected more efficient PEDV-specific antigen delivery with repetitive epitope presentation on every single VLP.

The fact that the VN response in milk peaked at 3 DPF provided useful insight on the timing of sow immunization regime. Having the final booster dose close to farrowing could sustain a higher antibody titer in milk and last for longer to enhance lactogenic immunity in nursing piglets.

The higher VN titer from the Vaccine group increased clinical protection of neonatal piglets from PEDV. In terms of clinical relief, piglets from the Vaccine group experienced faster recovery (p  0.05) at 8 DPC for activity, and at both 8 & 9 DPC for body con- dition when compared to the PBS control (Fig. 3A-B). Albeit the lack of statistical analysis for the overall 9-day diarrhea scores, the Vaccine group had more prompt clinical improvement in diar- rhea starting at 5 DPC and better fecal health towards the end of the challenge study (Fig. 3C). In terms of survival, piglets in the

Vaccine group had an increased 10 DPC survival rate of 60% when compared to the rate of 30% from the PBS control (Fig. 5).

For viral shedding quantification, it was difficult to control the variability within each group largely due to both the sensitivity of the qPCR assay and inconsistency of the amount of fecal swab

Fig. 6. Gross pathology evaluation for piglets. Small intestine gross pathology and colon fecal content were scored from 1 to 3 at the time of necropsy after either euthanasia or natural death. (A) Small intestine gross pathology scores. 1, normal. 2, either thin walled or gas-distended small intestine. 3, both thin walled and gas-distended small intestine. (B) Colon fecal content scores. 1, solid or pasty feces. 2, semi-watery feces. 3, watery feces with no solid content. Statistical significance is tested using unpaired t test. Error bars indicate ± SD for each group.

Fig. 7. Small intestine lesions. Jejunum from each piglet was fixed in formalin and hematoxylin & eosin (H&E) slide was subsequently prepared. (A) Microscopy images of one representative H&E slide of sectioned jejunum from each group. (B)

Villous length (v) and crypt depth (c) were measured in micrometer (mm) at 10 different sites on each jejunum slide section from each piglet, then the average v to c ratio was calculated. Statistical significance is tested using unpaired t test. Error bars indicate ± SD for each group. p values are represented as *p  0.05, **p  0.01,

***p  0.001, ****p  0.0001.

6 Y. Lu et al. / Vaccine xxx (xxxx) xxx Please cite this article as: Y. Lu, S. Clark-Deener, F. Gillam et al., Virus-like particle vaccine with B-cell epitope from porcine epidemic diarrhea virus (PEDV) incorporated into hepatitis B virus core capsid provides clinical alleviation against PEDV in neonatal piglets through lactogenic immunity, Vaccine, https:// doi.org/10.1016/j.vaccine.2020.06.009 material collected from each individual animal. However, the numerically higher mean Ct values from the Vaccine group at 3

DPC provided corroboration for a statistically significant event (Fig. 4). The higher VN titer (Fig. 2) from the Vaccine group likely contributed to the lower viral RNA load at 3 DPC when both groups developed some degree of diarrhea.

One of the major small intestinal tissue lesion markers is villous length to crypt depth ratio, with higher ratio associated with healthier tissue. This is largely due to the fact that villous epithelial cells in the small intestine are PEDV’s main infection target [3,27].

The higher VN titer from the Vaccine group likely was able to block

PEDV’s infection in the small intestine more effectively than the

PBS control, as a result, significantly less severe (p  0.05) villi atro- phy in the jejunum (Fig. 7). At the same time, the Vaccine group had numerically lower mean gross pathology scores with better health indications in both small intestine gross pathology and colon fecal content when compared to the PBS control (Fig. 6).

Due to space limitation at our large animal research facility, we were not able to include an additional test group of 2 gilts to eval- uate the commercial killed PEDV vaccine. However, we were able to leverage knowledge from a previous subunit vaccine study that looked at effectiveness of the commercial vaccine [10]. The gilts from our Vaccine group had similar level of VN antibody in colos- trum/milk when compared with the sow immunized with the commercial vaccine in the previous study, which utilized the same

VN assay from South Dakota State University’s Animal Disease

Research & Diagnostic Lab. Additionally, our VLP vaccine was able to reduce small intestine lesions in piglets to a comparable extent.

With the developmental nature of our study, an effective dosage design is yet to be refined. We do believe with further develop- ment, in addition to its unmatched safety, our VLP vaccine has promising potential in terms of further clinical alleviation against

PEDV.

During this vaccine study, we also evaluated the effect of Che- mokine ligand 28 (CCL28) as a potential adjuvant for its ability to enhance lymphocyte migration to the mammary gland [28,29] and in turn boost antibody transfer from gilt to piglet. However, our results showed that CCL28 did not help generate a higher VN titer in the colostrum/milk (data not shown).

In summary, our VLP-based vaccine was able to stimulate sig- nificantly higher PEDV neutralizing response in gilt milk, as a result, provided neonatal piglets clinical alleviation with improve- ment in morbidity, small intestinal lesion, and survival. This pre- liminary pig study paved the way for additional development in terms of optimal vaccine dosage design and adjuvant selection in the near future.

Declaration of Competing Interest The authors declare that they have no known competing finan- cial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement We want to thank Smithfield Foods, Inc. for funding this project, also Dr. Terry Coffey and his team at Smithfield for insightful dis- cussions. We would also like to thank Jill Stafford, Heather Dob- bins, Cassandra Fields, Caitlin Swecker, Eli Hall, and Gabbie

Schmitt from Teaching & Research Animal Care Support Service (TRACSS) at Virginia Tech, Jamie Stewart from the Department of

Large Animal Clinical Sciences at Virginia-Maryland College of

Veterinary Medicine, Amy Rizzo from University Veterinarian &

Animal Resources at Virginia Tech, and Kyle Saylor from the

Department of Biological Systems Engineering at Virginia Tech for their help with the pig study. We would also want to acknowledge our appreciation for

Kathy Lowe at Virginia- Maryland College of Veterinary Medicine for her help with TEM,

Aaron Singrey, Travis Clement, and Eric Nelson at South Dakota

State University’s Animal Disease Research & Diagnostic Lab for their help with serology and molecular diagnostic tests.

References [1] Antas M, Wozniakowski G. Current status of porcine epidemic diarrhoea (PED) in European Pigs. J Vet Res 2019;63(4):465–70. [2] Kim Y et al. Trypsin-independent porcine epidemic diarrhea virus US strain with altered virus entry mechanism. BMC Vet Res 2017;13(1):356. [3] Lee C. Porcine epidemic diarrhea virus: An emerging and re-emerging epizootic swine virus. Virol J 2015;12:193. [4] Paarlberg P. Updated Estimated Economic Welfare Impacts of Porcine

Epidemic Diarrhea Virus (PEDV). AgEcon SEARCH; 2014. [5] Gerdts V, Zakhartchouk A. Vaccines for porcine epidemic diarrhea virus and other swine coronaviruses. Vet Microbiol 2017;206:45–51. [6] Crawford K et al. Status of vaccines for porcine epidemic diarrhea virus in the

United States and Canada. Virus Res 2016;226:108–16. [7] Li W et al. New variants of porcine epidemic diarrhea virus, China, 2011. Emerg

Infect Dis 2012;18(8):1350–3. [8] Oh J et al. Immunogenicity and protective efficacy of recombinant S1 domain of the porcine epidemic diarrhea virus spike protein. Arch Virol 2014;159 (11):2977–87. [9] Makadiya N et al. S1 domain of the porcine epidemic diarrhea virus spike protein as a vaccine antigen. Virol J 2016;13:57. [10] Subramaniam S et al. Vaccination of sows with a dendritic cell-targeted porcine epidemic diarrhea virus S1 protein-based candidate vaccine reduced viral shedding but exacerbated gross pathological lesions in suckling neonatal piglets. J Gen Virol 2018;99(2):230–9. [11] Park JE, Cruz DJ, Shin HJ. Receptor-bound porcine epidemic diarrhea virus spike protein cleaved by trypsin induces membrane fusion. Arch Virol

2011;156(10):1749–56. [12] Wicht O et al. Proteolytic activation of the porcine epidemic diarrhea coronavirus spike fusion protein by trypsin in cell culture. J Virol 2014;88 (14):7952–61. [13] Li C et al. Cell attachment domains of the porcine epidemic diarrhea virus spike protein are key targets of neutralizing antibodies. J Virol 2017;91(12). [14] Sun D et al. Identification of two novel B cell epitopes on porcine epidemic diarrhea virus spike protein. Vet Microbiol 2008;131(1–2):73–81. [15] Zhang YL et al. Enhanced immunogenicity of modified hepatitis B virus core particle fused with multiepitopes of foot-and-mouth disease virus. Scand J

Immunol 2007;65(4):320–8. [16] Ludwig C, Wagner R. Virus-like particles-universal molecular toolboxes. Curr

Opin Biotechnol 2007;18(6):537–45. [17] Dhanasooraj D, Kumar RA, Mundayoor S. Vaccine delivery system for tuberculosis based on nano-sized hepatitis B virus core protein particles. Int

J Nanomed 2013;8:835–43. [18] Kazaks A et al. Melanoma vaccine candidates from chimeric hepatitis B core virus-like particles carrying a tumor-associated MAGE-3 epitope. Biotechnol J

2008;3(11):1429–36. [19] Mohsen MO et al. Virus-like particles for vaccination against cancer. Wiley

Interdiscip Rev Nanomed Nanobiotechnol 2020;12(1):e1579. [20] Selzer L, Katen SP, Zlotnick A. The hepatitis B virus core protein intradimer interface modulates capsid assembly and stability. Biochemistry 2014;53 (34):5496–504. [21] Gillam F, Zhang J, Zhang C. Hepatitis B core antigen based novel vaccine against porcine epidemic diarrhea virus. J Virol Methods 2018;253:61–9. [22] Gillam F, Zhang C. Epitope selection and their placement for increased virus neutralization in a novel vaccination strategy for porcine epidemic diarrhea virus utilizing the

Hepatitis B virus core antigen.

Vaccine 2018;36 (30):4507–16. [23] Lu Y et al. Assessing sequence plasticity of a virus-like nanoparticle by evolution toward a versatile scaffold for vaccines and drug delivery. Proc Natl

Acad Sci U S A 2015;112(40):12360–5. [24] Reed LJ, Muench H. A simple method of estimating fifty per cent endpoints12.

Am. J. Epidemiol. 1938;27(3):493–7. [25] Okda F et al. Development of an indirect ELISA, blocking ELISA, fluorescent microsphere immunoassay and fluorescent focus neutralization assay for serologic evaluation of exposure to North American strains of Porcine

Epidemic Diarrhea Virus. BMC Vet Res 2015;11:180. [26] Dee S et al. Modeling the transboundary risk of feed ingredients contaminated with porcine epidemic diarrhea virus. BMC Vet Res 2016;12:51. [27] Jung K, Saif LJ. Goblet cell depletion in small intestinal villous and crypt epithelium of conventional nursing and weaned pigs infected with porcine epidemic diarrhea virus. Res Vet Sci 2017;110:12–5. [28] Hernandez-Ruiz M, Zlotnik A. Mucosal Chemokines. J Interferon Cytokine Res

2017;37(2):62–70. [29] Niimi K et al. Development of immune and microbial environments is independently regulated in the mammary gland. Mucosal Immunol 2018;11 (3):643–53.

Y. Lu et al. / Vaccine xxx (xxxx) xxx 7 Please cite this article as: Y. Lu, S. Clark-Deener, F. Gillam et al., Virus-like particle vaccine with B-cell epitope from porcine epidemic diarrhea virus (PEDV) incorporated into hepatitis B virus core capsid provides clinical alleviation against PEDV in neonatal piglets through lactogenic immunity, Vaccine, https:// doi.org/10.1016/j.vaccine.2020.06.009