Preimmunization of SPF Mice With Capsid Virus‐like Particle Has no Effect on CD8 ⁺ Immune Memory and Skin Inflammation in an Ovalbumin‐Induced Dermatitis Model

ABSTRACT

Specific pathogen‐free (SPF) laboratory mice have in recent years been criticized for their limited bench to bedside translational value, specifically within immunological and inflammatory research, due to their immature immune system caused by restricted pathogen exposure. Pathogenic reintroduction would result in the spread of infections, compromising staff safety and animal welfare and a risk of reducing reproducibility of in vivo studies. Previous studies have shown that preimmunizing SPF mice with inactivated murine pathogens efficiently induces increased levels of CD8⁺ effector memory T cells and higher inflammatory responses in a skin inflammation model. However, the method is laborious and carries infection risks if inactivation is improper. Therefore, a simpler and more standardized method is preferred to improve reproducibility and safety. We hypothesized that preimmunizing mice with an AP205 capsid virus‐like particle (cVLP) would be a safe, easy, and effective method to stimulate memory T cells in SPF mice and increase the inflammatory response in an ovalbumin‐induced (OVA) dermatitis model similar to the preimmunization with inactivated pathogens. The preimmunized mice seroconverted to the cVLP antigens, but the preimmunization did not induce higher levels of effector memory T cells compared to vehicle treated mice, nor did it affect the inflammatory response in the skin inflammation model. In conclusion, preimmunization with cVLP was not sufficient to induce a cellular immune response in the mice.

Study Highlights

• What is the current knowledge on the topic?

  • ○
    SPF mice have limited immunological memory due to restricted exposure to pathogenic stimuli, making their translational value for modeling certain human diseases questionable.
• What question did this study address?

  • ○
    Can the adaptive immune response of SPF mice be preconditioned using a synthetic, noninfectious cVLP‐based immunization strategy to better mimic the immune status of adult humans and thereby increase their translational relevance?
• What does this study add to our knowledge?

  • ○
    Preimmunization of SPF mice with two doses of a cVLP‐based vaccine is insufficient to activate enhanced memory T cell responses and did not impact the inflammatory response in an ovalbumin‐induced skin inflammation model. Activation of immunological memory is complex.
• How might this change clinical pharmacology or translational science?

  • ○
    Stimulation of memory T cells is expected to create inflammatory mouse models that translate better to adult humans, but using this synthetic cVLP preimmunization is insufficient for this purpose.

1. Introduction

Laboratory mice have long been used in experimental in vivo models mimicking human diseases with the purpose of strengthening disease biology understanding, immunological responses and/or testing new drug candidates in the preclinical phase of drug development. Systematic breeding and specific pathogen‐free (SPF) housing have reduced animal mortality and experimental variation. However, many findings in murine models do not translate well to humans, limiting progression of preclinical research to continue into clinical development [1, 2, 3]. One reason may be the ultra‐hygienic SPF housing, which prevents exposure to pathogens such as viruses, parasites, and various bacteria [3]. Pathogenic exposure, early‐life infections, and microbiota play a critical role in developing and shaping the immune system in mice as well as humans [4, 5]. When exposed to intracellular antigens, the immune system is activated, leading to the recognition and subsequent eradication of the infections. Dendritic cells or other antigen‐presenting cells present antigens to naïve CD8⁺ T cells, which differentiate into cytotoxic CD8⁺ T cells. After pathogen eradication, most effector cytotoxic T cells perish, while some become long‐lived memory CD8⁺ T cells, ready to respond to future encounters with the same antigen [6, 7]. This memory CD8⁺ T cell pool develops due to pathogenic exposure and provides lifelong protection [8]. While SPF housing protects mice from infections, it also limits immune maturation, resulting in an immunological profile that resembles that of neonatal rather than adult humans [4, 9].

To address this, it has been attempted to create “dirty” mice with elevated immunological memory through various methods. One method is cohousing of laboratory mice with pet‐shop mice, which induced elevated levels of memory CD8⁺ T cells and changes in blood cell gene expression, reflecting adult human immune gene signatures more closely, but resulting in higher mortality amongst cohoused SPF mice due to uncontrolled infections [10]. Another is feralization of SPF mice by supplying farmyard materials to cages. This promotes immune similarity to adult humans by inducing a mature immunophenotype characterized by increased expression of NK and T cell maturation markers and an elevated IFN‐γ response to stimulation, which has been associated with protective effects against colorectal carcinogenesis [11]. Also, “wildlings” have been developed by transplanting laboratory mouse embryos to wild mice, preserving offspring inbred genetics, but restoring natural microbiota and pathogens, thus creating more comparable immune responses to adult humans [12, 13]. Despite the obvious translational advantages of using dirty mouse models, they also result in the reintroduction of pathogens, posing risks of uncontrolled infections, higher mortality, compromised staff safety, increased costs, and study variation, limiting research reproducibility and feasibility. Thus, SPF mice remain essential in most research, necessitating compliance with SPF housing regulations.

Falkenberg et al. [14] increased effector memory CD8⁺ T cell levels in BALB/c and C57BL/6 mice by preimmunizing with a mixture of inactivated viruses and Mycoplasma with a potent adjuvant elevating CD8⁺ memory T cell levels in both strains, with more than 80% of the mice reaching levels comparable to pet shop mice. This method was further explored in an ovalbumin‐induced (OVA) skin inflammation model. Here, preimmunization with inactivated pathogens resulted in increased skin inflammatory responses, but also higher interindividual variation [14]. In a diet‐induced obesity model, preimmunized mice displayed similar obese phenotypes as the nonimmunized mice, but increased effect size after treating with liraglutide resulting in increased statistical power [15]. Preimmunization has potential as it avoids active infections and can be implemented in SPF facilities. However, preimmunization with a complex pathogen mixture still poses infection risks if inactivation is insufficient. Commercial use would require thorough batch controls, cause delays, and make it a demanding and expensive alternative.

We therefore aimed for a simpler, standardized stimulus to activate the adaptive immune response using preimmunization that should be able to meet four criteria: 1) no infection risk, 2) easy industrial implementation, 3) elevated levels of memory CD8⁺ T cells, and 4) minimal interindividual variation.

Capsid virus‐like particles (cVLPs) are highly immunogenic and can induce high levels of antibodies even without an extrinsic adjuvant [16, 17]. Therefore, we proposed testing cVLP immunization in C57BL/6 (B6) mice to stimulate adaptive memory by elevating splenic effector memory CD8⁺ T cells and local inflammatory responses in a skin inflammation model. cVLPs mimic the structure, size, and symmetry of viruses without containing viral genome or replicase, making them safe even for immunodeficient individuals [18]. This technology is used in several vaccines, with the HPV vaccine being a successful example [18, 19, 20]. cVLPs elicit both humoral and cellular immune responses, making them suitable for stimulating immunological memory in laboratory mice [18, 19]. While most cVLP research focuses on B‐cell response and antibody production, relatively few studies have examined the ability of cVLPs to stimulate effector memory CD8⁺ T cell pools [19, 20, 21]. Large‐scale, low‐cost cVLP production is available [19], making cVLP immunization a feasible and affordable approach for industrial SPF laboratory facilities.

We hypothesized that immunizing SPF B6 mice with synthetic AP205 cVLPs would elevate the level of effector memory CD8⁺ T cells to levels comparable to those in pet shop mice, which were used as positive controls, and reduce interindividual variation in an OVA‐induced skin inflammation model.

2. Materials and Methods

2.1. Animal Ethics

This animal study was planned following the PREPARE guidelines [22] and reported according to the ARRIVE guidelines [23]. Animal experiments were approved by the Animal Experiment Inspectorate under the Ministry of Food, Fishery, and Agriculture in Denmark and assigned the license number 2022‐15‐0201‐01251. Animal studies were performed according to the Danish Act on Animal Experimentation (LBK nr. 1107 af 01/07/2022) and the EU directive on protecting animals used for scientific purposes (2010/63/EU). Following international guidelines [9], health monitoring discovered no signs of listed infections.

2.2. Animals

Animal group sizes were determined based on power analysis and data from previous studies.

A total of 56 C57BL/6NTac (B6) laboratory mice, 5 weeks of age, 28 females and 28 males, were randomly selected from Taconic Biosciences (Lille Skendsved, Denmark). All B6 mice were earmarked upon arrival and randomly allocated into groups of immunization, vehicle, immunization and OVA, vehicle and OVA with equal division of males and females in groups. Mice were housed in open‐type 1284 L Eurostandard II L cages (Techniplast; Varese, Italy). Immunized and vehicle treated mice were cohoused at the University of Copenhagen's AAALAC‐accredited facility, a barrier‐protected experimental rodent facility within the Faculty of Health and Medical Sciences, Copenhagen, Denmark at a mean temperature of 21.8°C, mean relative humidity of 39.2%, and light/dark cycle of 12 h/12 h.

The cages were bedded with aspen material (Tapvei; Kiili Harjumaa, Estonia), paper nesting material (Enviro‐Dri; Cleveland, USA), a handling tube (Scanbur; Karlslunde, Denmark), a cardboard igloo/house, a play tube, and a wooden chew block (LBS Biotechnology, Hookwood, UK).

All mice were fed Altromin 1324 (Altromin, Lage, Germany) and tap water ad libitum.

Pet shop mice were chosen for positive control as murine infections are known to be widespread in pet shop colonies. On the day of euthanasia, seven females and seven males, adult pet shop mice aged 10–12 weeks or older, were obtained from Bonnie Dyrecenter (Copenhagen NV, Denmark) and transported in a transport cage with aspen bedding, nesting material, HydroGel, and Altromin 1324 until procedures a few hours later.

2.3. Immunization Study

cVLP assembled from the major coat protein of the Acinetobacter phage, AP205 (Gene ID: 956335), genetically fused to SpyCatcher (Genbank: OK422508.1) was kindly provided by AdaptVac (Frederiksberg, Denmark).

On the day of immunization, the cVLP suspension was thawed at room temperature for 10 min, diluted with sterile Phosphate‐Buffered Saline (PBS) to a concentration of 0.04 g/L, and kept on ice during transport to the animal facility.

After 6 days of acclimatization, at Day 0, body weight of all individuals was measured for health monitoring. Each mouse was injected subcutaneously (SC) according to their allocated group near the base of the tail with 50 μL of either PBS or cVLP suspension with a 1 mL syringe and 29G needle (KRUUSE; Langeskov, Denmark). Post‐injection, mice were fed sunflower seeds, mealworms, and dried corn (Brogaarden; Lynge, Denmark) to distract them from the stressful procedures. The injections were repeated 14 days later (Figure 1). Additional weighing was performed on study days 7 and 21 for health monitoring after injections.

FIGURE 1.

Study timeline showing main experimental procedures. Day 6: Arrival of 5‐weeks‐old C57BL/6NTac (B6) male and female mice, earmarking, randomized group allocation, six days acclimatization (n = 84). Day 0 and 14: SC injection of cVLP vaccine or vehicle (n = 42, n = 42). Day 28: Euthanasia and validation of vaccine (n = 28 immunized, n = 28 vehicle). Flow cytometry was attempted on all animals euthanized at Day 28, but due to technical issues on the machine, data was lost from vehicle 2 and immunized 2 groups. Vehicle 1 and immunized 1 groups flow data was analyzed as described in methods, and no issues were detected. Day 27: Sensitization IP with OVA/alum to validate immunization in an OVA/alum skin inflammation model (n = 14 immunized, n = 14 vehicle). Day 37: ID challenge right ear with OVA. Day 38: Euthanasia and sampling (n = 14 immunized, n = 14 vehicle). Pet shop mice, males, and females (n = 7, n = 7) were euthanized the same day of their arrival, day 38. Illustration created in Biorender.com. Alum, Alhydrogel; cVLP, capsid virus‐like particle; ID, intradermal; IP, intraperitoneal; OVA, ovalbumin; SC, subcutaneous.

At study Day 28, half of the animals (n = 14 immunized and n = 14 vehicle) were anesthetized by SC injection with a 1:1 blend of midazolam (5 mg/mL midazolam; B. Braun; Melsungen, Germany) and Hypnorm (0.315 μg/mL fentanyl citrate, 10 mg/mL fluanisone; Skanderborg Apotek; Skanderborg, Denmark) and dosed according to body weight (0.1 mL/10 g body weight). Blood was sampled from the periorbital plexus, after which mice were euthanized by cervical dislocation (Figure 1). Post‐mortem, the spleen was harvested, placed in ice‐cold PBS, and kept cool until preparation for flow cytometry the same day. Blood was left at room temperature for an hour of clotting, then cooled and centrifuged within 4 h (8000 rpm, 10 min, 4°C; Sigma model 1–14, Buch & Holm; Denmark) and serum was transferred to 1.4 mL noncoated sterile tubes (Micronic; Lelystad, Netherland) and stored at −80°C until analysis.

2.4. OVA/Alum Skin Inflammation Study

On study Day 27 (Figure 1), the remaining mice were divided into two groups: 14 immunized mice (7 females and 7 males) and 14 vehicle mice (7 females and 7 males). These mice were then sensitized with an intraperitoneal injection of 200 μL OVA/alum solution containing 0.05 mg/mL ovalbumin (Albumin A5503; Sigma‐Aldrich, St. Louis, Missouri, USA) and 1% aluminum hydroxide (Alhydrogel Adjuvant 2%; InvivoGen; Toulouse, France).

Mice were offered snacks and had access to DietGelRecovery (ClearH₂O; Westbrook, ME, USA) 3 days prior and 3 days postinjection to maintain body weight, which was obtained two consecutive days following sensitization.

At study Day 37, mice were anesthetized with isoflurane/O₂ (Baxter; Søborg, Denmark) and injected intradermally (ID) on the dorsal side of their right ear with 10 μL ovalbumin (OVA Endofit; InvivoGen; Toulouse, France) using an insulin syringe (KRUUSE; Langeskov, Denmark).

Twenty‐four hours later, all mice were anesthetized with 1:1 midazolam and Hypnorm, blood was sampled from the periorbital plexus, and mice were euthanized by cervical dislocation (Figure 1). An 8 mm biopsy from the center of the right ear was collected with Biopsy pouches (KRUUSE; Langeskov, Denmark), and the weight of each ear biopsy was registered. The biopsies were cut into quarters and placed in 0.5 mL Precellys standard tubes containing 1.4 mm ceramic (zirconium oxide) beads (Ref: P000933‐LYSK0‐A; Precellys Lysing Kit; Bertin Instruments; Linhamn, Sweden) followed by snap‐freezing in dry ice and stored at −80°C until tissue homogenization and protein analysis.

2.5. Flow Cytometry

Freshly excised spleens were kept in ice‐cold PBS until squeezed through a 70 μm cell strainer (Corning; Durham, NC, USA) over a 50 mL Falcon tube using the back of a plunger from a 2 mL syringe (KRUUSE, Langeskov, Denmark) to create a single‐cell suspension. Samples were centrifuged (8000g, 5 min, 4°C), resuspended in FACS buffer (2% Fatal bovine serum (FBS) in PBS), and transferred to a sterile 96‐well plate for counting of total leukocytes (limited to cell size 4‐12 μm) on a TC20 Automated Cell Counter (Bio‐Rad) and the total cell count per organ was calculated.

Approximately 1,000,000 leukocytes were stained for extra‐ and intracellular markers with fluorochrome‐conjugated antibodies (Invitrogen; Waltham, MA, USA or Biolegend; San Diego, CA, USA) (Table 1 and Figure S1). Extracellular staining antibodies were mixed in dark Eppendorf tubes diluted in eBioscience Flow Cytometry Staining Buffer (Invitrogen; Waltham, MA, USA) and added to each sample (Table 1). Intracellular staining antibodies were mixed in dark Eppendorf tubes and diluted in eBioscience Permeabilization Buffer (Invitrogen; Waltham, MA, USA). Staining was performed over five cycles involving staining, incubation (10–20 min), centrifugation (3 min, 500 RCF, 4°C), and pouring of supernatant. Staining rounds were performed in this order: Live/dead staining, Fc‐blocking, extracellular panel staining, fixation in fix buffer (1:4 Fixation/Permeabilization Concentrate and 3:4 eBioscience Fixation/Perm Diluent [Invitrogen; Waltham, MA, USA]), and intracellular panel staining. The plate was kept cool for analysis the next day at Core Facility for Flow Cytometry and Single Cell Analysis, Faculty of Health and Medical Sciences, University of Copenhagen using a BD LSR Fortessa III (BD Bioscience, BD Biosciences, cat. no.: 647177; San Jose, CA, USA). Data were compensated, extracted and exported using BD FACSDivaSoftware (BD Bioscience) (Figure S1).

TABLE 1.

Summary of the fluorochrome‐conjugated antibodies for detection of extra‐ and intracellular biomarkers for identifying T cell subsets on BD LSR Fortessa III.

Cell surface protein	Antibody (Ab)	Fluorochrome	Clone	Isotope	Commercial concentration	Mixing dilution (Ab:Buffer)	Production company
Fc‐receptor block (CD16/CD32)	Antimouse CD16/32		93	Rat IgG2a, γ	0.5 mg/mL	1:100 FACS buffer*	Biolegend
CD44	Antihuman/−mouse CD44	FITC	IM7		0.5 mg/mL	1:200 Staining buffer**	Invitrogen (ebioscience)
TCRab	Antimouse TCR beta chain	PE‐Cy7	H57‐597	Armenian hamster	0.2 mg/mL	1:200 Staining buffer**	Invitrogen (ebioscience)
CD4	Antimouse CD4	BV510	GK1.5	Rat IgG2b, κ	0.2 mg/mL	1:200 Staining buffer**	Biolegend
CD8a	Antimouse CD8a	BV421	53–6.7	Rat IgG2a, κ	0.05 mg/mL	1:200 Staining buffer**	Biolegend
CD62L	Antimouse CD62L	BV650	MEL‐14	Rat IgG2a, κ	0.2 mg/mL	1:200 Staining buffer**	Biolegend
FoxP3	Antimouse/−rat foxP3	APC	FJK‐16s		0.2 mg/mL	1:100 Perm buffer***	Invitrogen (ebioscience)
Gata3	Antihuman/−mouse Gata‐3	PE	TWAJ		100 tests (5 μL (0.06 μg)/test)	1:50 Perm buffer***	Invitrogen (ebioscience)
Live/dead		APC‐Cy7				1:2000 PBS	Invitrogen (ebioscience)

Note: Asterixis indicate the buffer each antibody has been diluted in *FACS Buffer (PBS + 2% Fetal bovine serum), **eBioscience Flow Cytometry Staining Buffer (Invitrogen; Waltham, MA, USA), ***Permeabilization Buffer 10X (Invitrogen; Waltham, MA, USA) diluted 1:10 with MiliQ water.

2.6. cVLP‐Specific IgG in Serum

Enzyme‐linked immunosorbent‐assay (ELISA) analysis was performed in the laboratories of AdaptVac using their ELISA protocol to measure the level of cVLP‐specific antibodies in mouse serum. The cVLP‐specific antibody level was measured as area under the curve (AUC), providing a qualitative answer for antibody response and indicating the level of antibodies based on the endpoint titter.

Briefly, 96‐well ELISA plates (Nunc MaxiSorp, Invitrogen) were coated overnight at 4°C with 0.1 μg/well of SpyC‐AP205. Plates were blocked with 0.5% skimmed milk in PBS, pH 7.4 (i.e., blocking buffer). Mouse serum diluted in blocking buffer in a threefold dilution range starting from 1:50 was added to the plate and incubated for 1 h at room temperature. One row on each plate was left blank to quality check the background absorbance, and one row had serum previously assessed as a positive control. Plates were washed three times in PBS, pH 7.4 in between steps. Total cVLP‐specific IgG was detected using Horseradish peroxidase (HRP)‐conjugated goat antimouse IgG (Invitrogen) diluted 1:1000 in blocking buffer and incubated for 1 h at room temperature. Plates were prepared with TMB X‐tra (Kem‐en‐Tec), and the reaction was stopped using 0.2 M H₂SO₄. The absorbance was measured at 450 nm using a BioSan HiPo Mpp‐96 microplate reader (BioSan, Riga, Latvia).

2.7. IgE in Serum

The concentration of nonspecific serum IgE was quantified as a primary readout for the OVA/alum model to indicate an acute allergic reaction [14]. This measurement was selected to evaluate the general allergic response in immunized SPF mice subjected to OVA/alum allergenic stimulus compared to baseline IgE levels in pet shop mice. IgE was measured by ELISA using the Mouse IgE ELISA Kit (Cat.no.: E99‐115; Bethyl Laboratories Inc.; Montgomery, TX, USA). The procedures were performed following the accompanying manual from Bethyl Laboratories. The ELISA kit was placed at room temperature for approx. 30 min. for the reagents to reach room temperature. All samples were thawed and diluted 1:20 as recommended by the manufacturer. For the plate washing procedures, an automated plate washer (Agilent BioTek Washer Dispenser; BioTek Instruments Inc.; Winooski, VT, USA) was used, running three washing cycles each time. The optical density was read at 450 nm within 30 min after adding the stopping solution using an Epoch Microplate Spectrophotometer, and data were analyzed with the Gen5 V. 3.08 Microplate Reader and Image Software (BioTek Instruments Inc.; Winooski, VT, USA).

2.8. Protein Analysis

All samples of ear tissues were standardized by size, 8 mm in diameter, as biopsies were performed with Biopsy pouches (KRUUSE; Langeskov, Denmark).

Before protein detection with mesoscale, ear biopsies were homogenized using the Precellys Lysing Kit (Ref: P000933‐LYSK0‐A; Bertin Instruments; Linhamn, Sweden). Lysing buffer was mixed in a 15 mL falcon tube with 9 mL Milli‐Q water, 1 mL 10X Cell Signaling Lysis Buffer (#9803), 1 protease inhibitor cocktail tablet (Ref: 11873580001; Roche Diagnostics GmbH), 100 μL halt phosphatase inhibitor cocktail (Ref: 8159680747; Thermo Scientific), and 100 μL 100 mM vanadate. 150 μL of lysing buffer was added to each Precellys tube containing the sample. All samples were homogenized twice using a Precellys 24 Machine (6800 rpm 2 × 30 s, 30 s break). Samples were cooled on ice for 20 min post homogenization and centrifuged (15,000 rpm, 15 min, 4°C), then the supernatant was transferred to 1.4 mL noncoated sterile tubes (Micronic; Lelystad, Netherland) and stored at −80°C until analysis.

For protein analysis, the U‐PLEX Custom Biomarker Group 1 Mouse Assay (Cat.no.: K15069M‐1; Meso Scale Discovery LLC.; Rockville, MD, USA) was used for testing the concentration of the following 10 inflammatory proteins: Eotaxin, IL‐1β, IL‐4, IL‐5, IL‐13, IL‐17A, KC/GRO, MMP‐9 (total), TARC, TNF‐α. One day before analysis, the plate was coated with the capture antibodies according to the manual and incubated in the fridge (4°C–8°C) overnight. The next day, the samples were thawed at room temperature and kept cool on ice when not handled. The following steps were performed according to the instructions provided by Meso Scale Discovery LLC. The plate was read with a MESO QuickPlex SQ 120 (Model no.: 1300; Meso Scale Discovery LLC.; Rockville, MD, USA), and data was analyzed with Discovery Workbench 0.4 software.

2.9. Statistics

Data was analyzed with GraphPad Prism version 10.0 for Mac OS (GraphPad Software; Boston, MA, USA). Results with a p‐value of 0.05 or lower (p ≤ 0.05) were deemed significant. All data were tested for normal distribution with a Shapiro–Wilk normality test and an F‐test (two groups) or a Brown Forsythe test (> 2 groups) for equal variance.

Flow cytometry data fulfilling normal distribution and equal variance criteria were analyzed with one‐way ANOVA and Tukey's multiple comparisons. Data with normal distribution, but unequal variance was analyzed using Welch ANOVA and Dunnett's multiple comparisons. Data not fulfilling normal distribution nor equal variance criteria were analyzed with Kruskal–Walli's test and Dunn's multiple comparisons.

Inflammatory protein data not fulfilling normal distribution and/or equal variance criteria were ranked before analyzing with two‐way ANOVA and Šídák's multiple comparisons.

cVLP‐specific antibody data were plotted in a graph with an x‐axis of log (1/con), creating a sigmoid curve for each sample. Area under the curve (AUC) was calculated for each sigmoid graph, and a three‐way ANOVA with subsequent Šídák's multiple comparisons was performed. Grubbs' method (α = 0.05) was applied to identify outliers in the AUC data.

IgE data were ranked and tested with a two‐way ANOVA and Šídák's multiple comparisons.

3. Results

3.1. Immune Response to Preimmunization With cVLP

3.1.1. cVLP Preimmunization Did Not Induce Higher Levels of CD8 ⁺ Effector Memory T Cells

The effect of preimmunizing B6 SPF mice with cVLP injections was assessed and the numbers and percentages of various T cell subsets from the spleen were compared with vehicle treated B6 SPF mice and pet shop mice.

None of the measured CD8⁺ T cell populations differed between preimmunized or vehicle treated mice (Figure 2A–H). On the other hand, pet shop mice differed from the SPF mice: pet shop mice had a lower percentage of CD8⁺ T cells (Figure 2B); lower numbers and percentages of CD8⁺ central memory T cells (Figure 2G,H); and higher numbers and percentages of CD8⁺ effector memory T cells (Figure 2E,F) compared to SPF mice.

FIGURE 2.

cVLP preimmunization did not induce higher levels of CD8⁺ effector memory T cells. B6 mice received two SC injections with 50 μL cVLP preimmunization or vehicle with 14‐day intervals at equal time points. (A) CD8⁺ T cell numbers, (B) CD8⁺ T cell percentage, (C) naïve CD8⁺ T cell numbers, (D) naïve CD8⁺ T cell percentage, (E) effector memory CD8⁺ T cell numbers, (F) effector memory CD8⁺ T cell percentage, (G) central memory CD8⁺ T cell numbers, (H) central memory CD8⁺ T cell percentage. Means were compared with one‐way ANOVA and Tukey's multiple comparison or Welch's ANOVA and Dunnett's multiple comparisons or Kruskal–Wallis and Dunn's multiple comparison depending on normality and variance. Pet shop n = 7 females and n = 7 males, B6 preimmunized n = 7 females and n = 7 males, B6 vehicle n = 7 females and n = 7 males. Bar top represents mean with SD, asterisks illustrate significance between groups with p values of *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. cVLP, capsid virus‐like particle; SC, subcutaneous.

3.1.2. cVLP Preimmunization Did Not Induce Higher Levels of CD4 ⁺ T Cell Subsets, and Pet Shop Mice Had More Regulatory T Cells and Fewer Type 2 Helper Cells Compared to SPF Mice

As with CD8⁺ T cells, none of the measured CD4⁺ T cell populations differed between preimmunized or vehicle treated mice (Figure 3A–H). While pet shop mice had lower percentages of CD8⁺ T cells, they oppositely had higher percentages of overall CD4⁺ T cells (Figure 3B). Additionally, pet shop mice exhibited higher numbers and percentages of naïve‐ and effector memory CD4⁺ T cells (Figure 3C–F), but lower percentage of CD4⁺ central memory T cells (Figure 3H) compared to SPF mice. Finally, pet shop mice had significantly higher percentages of Treg cells (Figure 4B), but lower numbers and percentages of Th2 cells (Figure 4C,D) compared to SPF mice.

FIGURE 3.

Numbers and percentages of CD4⁺ T cell subsets from mouse spleen of female and male B6 mice 10 weeks of age and pet shop mice. B6 mice received two SC injections with 50 μL cVLP preimmunization or vehicle with 14‐day intervals at equal time points. (A) CD4⁺ T cell numbers, (B) CD4⁺ T cell percentage, (C) naïve CD4⁺ T cell numbers, (D) naïve CD4⁺ T cell percentage, (E) effector memory CD4⁺ T cell numbers, (F) effector memory CD4⁺ T cell percentage, (G) central memory CD4⁺ T cell numbers, (H) central memory CD4⁺ T cell percentage. Means were compared with one‐way ANOVA and Tukey's multiple comparison or Welch's ANOVA and Dunnett's multiple comparisons or Kruskal–Wallis and Dunn's multiple comparison depending on normality and variance. Pet shop n = 7 females and n = 7 males, B6 preimmunized n = 7 females and n = 7 males, B6 vehicle n = 7 females and n = 7 males. Bar top represents mean with SD, asterisks illustrate significance between groups with p values of *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. cVLP, capsid virus‐like particle; SC, subcutaneous.

FIGURE 4.

Numbers and percentages of Treg and Th2 cells from mouse spleen of female and male B6 mice 10 weeks of age and pet shop mice. B6 mice received two SC injections with 50 μL cVLP preimmunization or vehicle with 14‐day intervals at equal time points. (A) Treg numbers, (B) Treg percentages, (C) Th2 numbers, (D) Th2 percentages. Means were compared with Welch's ANOVA and Dunnett's multiple comparisons or Kruskal–Wallis and Dunn's multiple comparisons depending on normality and variance. Pet shop n = 7 females and n = 7 males, B6 preimmunized n = 7 females and n = 7 males, B6 vehicle n = 7 females and n = 7 males. Bar top represents mean with SD, asterisks illustrate significance between groups with p values of *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. cVLP, capsid virus‐like particle; SC, subcutaneous.

3.1.3. Preimmunized Mice Seroconverted to the cVLP Antigen, Whereas Cohoused Vehicle Treated Mice Remained Seronegative

cVLP‐specific IgG levels in the sera of mice were analyzed by ELISA before and after being subjected to the OVA‐induced skin inflammation. Three datapoints were identified as outliers using Grubbs' method (α = 0.05) and were excluded from the dataset.

All preimmunized mice had seroconverted and showed significantly higher levels of cVLP‐specific IgG antibodies compared to the cohoused vehicle‐treated mice (Figure 5A). OVA‐treated mice and vehicle‐treated mice did not differ in cVLP antibody response.

FIGURE 5.

cVLP specific IgG and total IgE in serum from 12‐week‐old female and male B6 mice. (A) AUC of cVLP‐specific IgG from ELISA analysis. Comparison (ranked three‐way ANOVA) of antibody response in mice immunized with cVLP (n = 7 females, n = 7 males) or vehicle treatment with PBS (n = 7 females, n = 7 males), and the influence of being subjected to OVA/alum‐induced skin inflammation on the right ear. (B) IgE antibody concentration (ng/mL) was measured with an ELISA kit after OVA/alum sensitization and challenge in all four groups (n = 28) and in pet shop mice not induced with OVA/alum. Comparison (Kruskal–Wallis and Dunn's multiple comparison) between groups. Pet shop n = 7 females and n = 7 males, B6 preimmunized n = 7 females and n = 7 males, B6 vehicle n = 7 females and n = 7 males. Bar top represents mean with SD, asterisks illustrate significance between groups with p values of *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Alum, aluminum hydroxide; AUC, area under the curve; cVLP, capsid virus‐like particle; ELISA, enzyme‐linked immunosorbent assay; ID, intradermal; Ig, immunoglobulin; IP, intraperitoneal; OVA, ovalbumin; SC, subcutaneous.

3.2. Response to cVLP Preimmunization in the Skin Inflammation Model

3.2.1. IgE Levels in Serum Did Not Differ Between Preimmunized or Vehicle Treated Mice Subjected to the OVA/Alum Skin Inflammation Model

IgE antibody levels were measured in serum of mice subjected to the OVA/alum skin inflammation model and compared to baseline levels from pet shop mice. IgE concentrations did not differ significantly between preimmunized and vehicle treated mice, neither when compared to pet shop levels (Figure 5B).

3.2.2. cVLP Preimmunization Did Not Affect the Inflammatory Response in a Mouse Skin Inflammation Model

Relevant inflammatory proteins, including cytokines and chemokines, were measured in ear tissue from both preimmunized and vehicle‐treated mice with OVA‐induced skin inflammation. The levels of inflammatory proteins from the challenged ear did not differ between preimmunized and vehicle‐treated mice (Figure 6). The Th2 cytokine IL‐13 showed a significantly higher level in preimmunized male mice (Figure 6C). However, no other significant differences were shown.

FIGURE 6.

cVLP preimmunization did not affect the inflammatory response in a mouse model of skin inflammation. Inflammatory protein concentration (pg/mL) in the ears of preimmunized or vehicle treated mice subjected to OVA/alum skin inflammation. (A) IL‐4, (B) IL‐5, (C) IL‐13, (D) MMP‐9, (E) Eotaxin, (F) TARC, (G) IL‐1β, (H) KC/GRO, (I) TNF‐α, and (J) IL‐17A. Proteins from homogenized ear tissue from 12‐week‐old female and male B6 mice (n = 14 females, n = 14 males) exposed to a skin inflammation model induced by IP sensibilization and intradermal ID ear injection of OVA/alum. Comparison (two‐way ANOVA) of mice immunized with cVLP (n = 7 females, n = 7 males) or vehicle (n = 7 females, n = 7 males). Data not fulfilling normal distribution and/or equal variance criteria were ranked before analysis. Bar top represents mean with SD, *p < 0.05. Alum, aluminum hydroxide; cVLP, capsid virus‐like particles; ID, intradermal; IP, intraperitoneal; OVA, ovalbumin.

4. Discussion

In this study, we investigated a safe, noninfectious, and simpler method to stimulate the adaptive immune system of SPF mice through preimmunization with cVLP. Despite clear seroconversion in all cVLP preimmunized mice, there was no effect on any T cell subsets investigated compared to vehicle groups.

Effector memory CD8⁺ T cell activation typically requires intracellular infection, leading to the presentation of antigenic peptides via MHC Class I molecules on the cell surface. However, synthetic cVLPs lack the ability to cause intracellular infections. Instead, dendritic cells capture cVLPs and process their surface‐displayed antigenic epitopes for presentation via MHC Class II molecules, thereby promoting CD4⁺ T cell activation and subsequent B cell‐mediated antibody production [6, 24, 25]. This mechanism likely explains the absence of CD8⁺ T cell activation observed in our study, as the MHC class I pathway was not effectively engaged. Nevertheless, cVLP stimulation was also ineffective in activating CD4⁺ T cells in this experiment. This result may stem from insufficient uptake or processing of the cVLPs by dendritic cells, suboptimal adjuvant activity, or the absence of strong MHC Class II‐binding epitopes.

Falkenberg et al. [14] used an inactivated pathogen mixture containing five viruses and one bacterium, including remnants of MHC allogenic cells used for virus cultivation. These cells have shown to efficiently stimulate activated T cell responses [26]. The mixture also included Mycoplasma bacteria, which contain far more immunogenic epitopes than viruses [27]. This complex combination of whole viruses, intracellular bacteria, and cells thus has more diverse pathogenic epitopes compared to our simple cVLP antigen structure. Identifying the specific epitope profiles involved in CD8⁺ T cell activation could simplify the method and limit infection risks but requires comprehensive profiling of MHC‐bound pathogenic epitopes [28].

Although cVLP preimmunization induced a potent specific IgG response, it did not affect inflammatory biomarkers in the skin inflammation model, which is in line with the lack of cellular immune effects, as these were positively associated in the study by Falkenberg et al. [14]. Interestingly, pet shop mice had fewer Th2 cells and more Treg cells compared to SPF mice, in support of the hygiene hypothesis [29]. Th2 cells are key in allergic disorders, producing cytokines like IL‐4, IL‐5, and IL‐13, which recruit granulocytes causing pruritus and barrier disruption [30], whereas Treg cells establish tolerance against commensal bacteria and other microbes, preventing allergic reactions [31, 32]. Frequent microbial exposure and infections stimulate IL‐12 production, which favors Th1 differentiation over Th2 responses and thereby limits atopic inflammation [32]. Consequently, the pet‐shop mice would be expected to exhibit reduced allergic response in the Ova/alum model, similar to what has been observed in mice transplanted with wild mouse‐derived gut microbiota, which showed attenuated inflammatory responses to house dust mite‐induced allergic airway inflammation compared to SPF microbiome‐transplanted mice [33]. Conversely, allergic Th2 responses were not reduced in wildlings [12], highlighting the unresolved question regarding the specific factors responsible for allergy protection as proposed by the hygiene hypothesis.

For this study, pet shop mice were used as positive controls due to their high stocking density, poor ventilation, and limited protection from infectious agents, creating an environment for continuous reinfections and demonstrating immune alterations in “dirty” living conditions. However, their extreme housing conditions may not represent natural human bacterial exposure. Therefore, effector memory CD8⁺ T cell levels in feral mice and wildlings may be more comparable to adult humans and could serve as better positive controls for future research [3, 10].

Other vaccines are known to induce long‐term memory beyond antibody response, but many are live attenuated vaccines, unsuitable for SPF facilities due to infection risks [34]. Administering vaccines meant for one species to another may prevent infection while stimulating immune activation, as seen with the smallpox vaccine developed from cowpox virus [35]. Future studies should investigate whether vaccines known to induce long‐lived cellular immunological memory in humans can elicit similar responses in SPF mice, preferably synthetically produced or inactivated in a safe controlled manner.

5. Limitations

Due to practical limitations, the cVLP‐specific IgG serology data were obtained from mice in vehicle 2 and immunized 2 groups, while the mice used for flow cytometry were from vehicle 1 and immunized 1 groups. However, mice from the above mentioned groups were matched by strain, age, sex, group sizes, and housing conditions. The same cVLP batch was used for preimmunization of all animals, ensuring consistency with the results presented in this paper.

Another limitation of this study concerns the detection of regulatory T cells (T‐regs). Typically, T‐regs are identified using a combination of fluorochrome‐conjugated antibodies targeting CD3, CD4, CD25, and FoxP3 (CD3⁺/CD4⁺/CD25⁺/FoxP3⁺). However, our method lacks the CD25⁺ marker, which may result in the inclusion of other T‐cell subtypes in our T‐reg population. This could potentially obscure the true levels of T‐regs in our results.

6. Conclusion

We conclude that preimmunization with AP205 cVLP was not sufficient to induce cellular immune activation in SPF mice and cannot be used to improve the comparability of inflammatory skin responses to human inflammatory reactions.

Author Contributions

K.S.K. wrote the manuscript. A.K.H., C.H.F.H., J.K., C.B., and I.W.H. designed the research. K.S.K. and I.W.H. performed the research. K.S.K. and I.W.H. analyzed the data. A.F.S.B., L.G., C.B., and J.K. contributed with regents/analytical tools.

Funding

The Innovation Fund Denmark funded this study (grant no. 039‐00056B) based on a collaboration between the University of Copenhagen and LEO Pharma A/S as part of an industrial PhD project.

Conflicts of Interest

I.W.H. was employed at Leo Pharma and received funding from the Innovation Fund Denmark for the PhD project. C.B. was employed at Leo Pharma. L.G. was employed by Adaptvac, who provided the vaccines. A.F.S.B. has collaborated with the pharmaceutical industry and is CSO of AdaptVac and cofounder of NextGen Vaccines supplying the cVLP vaccine for this study. C.H.F.H. has collaborated with the pharmaceutical industry and received funding as described on https://ivh.ku.dk/english/employees/?pure=en/persons/306048. A.K.H. has collaborated with and received funding from the pharmaceutical industry and declares to be owner of a diabetes related patent described on https://ivh.ku.dk/english/employees/?pure=en/persons/107126. K.S.K. and J.K. declares no competing interest for this work.

Supporting information

Figure S1: Flow cytometry gating strategy. Lymphocytes (forward scatter vs. side scatter); single cells; live cells (live/dead staining); T cells (TCRb); CD4⁺ and CD8⁺ T cells. For each CD4⁺ and CD8⁺ T cell population, effector memory T cells were selected with CD44⁺ CD62L⁻, naïve T cells with CD44⁻ CD62L⁺ and central memory T cells with CD44⁺ CD62L⁺. Treg cells were selected with FoxP3 expression and Th2 cells were selected with gata3 expression. Both Tregs and Th2 cells were gated on CD4⁺ T cells using fluorescence‐minus‐one staining techniques.

Figure S2: Body weight date from all mice over time. Female and male mice from immunized Group 1 and 2 are pooled (blue graph), female and male mice from placebo Group 1 and 2 are pooled (red graph). Female and male pet shop mice are pooled (green graph). Body weight of pet shop mice was recorded on the day of arrival/euthanasia. Numbers on the X‐axis represent the days of intervention. Datapoints represents group mean with SD. OVA, ovalbumin.