The Safety of Alcaligenes Lipid A in a Virus-Induced Immune Disease Model Associated with IgA, Th17 Cells, and Microbiota

TL;DR

This study shows that ALA, a bacterial compound, does not worsen a virus-induced immune disease in mice, despite concerns about its effect on immune responses.

Contribution

The study demonstrates the safety of ALA in a virus-induced immune disease model involving IgA, Th17 cells, and gut microbiota.

Findings

ALA administration did not worsen TMEV-induced neurological disease or viral persistence in mice.

ALA did not enhance IgA or Th17 immune responses in the model.

ALA altered gut microbiota composition, increasing Bacteroidota phylum members like Alistipes and Bacteroides.

Abstract

Lipid A is a component of lipopolysaccharide (LPS) of Gram-negative bacteria. Previously, we demonstrated that synthesized lipid A derived from Alcaligenes faecalis (ALA) could enhance antigen-specific immunoglobulin (Ig) A and T helper (Th) 17 responses, when ALA was co-administered experimentally with an antigen as a vaccine adjuvant. This raised concerns about the safety of the ALA usage, since IgA and Th17 responses have been suggested to play a pathogenic role in several immune-mediated diseases, including multiple sclerosis (MS). We investigated whether ALA administrations could exacerbate an animal model of MS, Theiler’s murine encephalomyelitis virus (TMEV) infection. TMEV-infected SJL/J mice were administered ALA at various time points, and their neurological signs were observed for 7 weeks. We found that ALA administrations did not exacerbate TMEV-induced inflammatory disease or viral persistence in the central nervous system (CNS), clinically or histologically. Furthermore, ALA administrations did not enhance TMEV-specific humoral and cellular responses, including IgA and Th17 responses. On the other hand, principal component analysis (PCA) of the fecal, not the ileal, samples showed significant changes in the microbiota, characterized by increases in the relative abundance of bacteria belonging to the phylum Bacteroidota, including the genera Alistipes and Bacteroides. Therefore, ALA injections could be safe for use in immune-mediated diseases, whose immunopathology has been associated with IgA and Th17 responses.

Figures

Full text

1. Introduction

Lipopolysaccharide (LPS), which is also called endotoxin, is a major glycoconjugate of the outer membrane of Gram-negative bacteria [1,2]. LPS has immunostimulatory and toxic functions [3], depending on the bacterial taxa [4,5]. The glycolipid lipid A, an acylated disaccharide at the terminus of LPS, is the active component of LPS, and is linked to the terminus of the polysaccharide moiety via the bacteria-specific acidic sugar, 3-deoxy-D-manno-octulosonic acid (Kdo) [2]. Since lipid A can induce various immunomodulatory responses, such as cytokine production [6] and lymphocyte activation/migration [7,8], lipid A has been investigated as a novel vaccine adjuvant [9,10]. On the other hand, pathogen-derived LPS and lipid A can cause severe toxicity in humans, including liver damage, fever, and septic shock; these agents were found to be unsuitable as vaccine adjuvants [11,12,13,14,15,16,17]. In one solution to this problem, chemical modification of lipid A from Salmonella minnesota R595 led to the derivative monophosphoryl lipid A (MPL), which has been shown to be a safe and efficacious adjuvant. Currently, only one of the MPL derivatives is used as a vaccine adjuvant; 3-O-deacyl-4′-monophosphoryl lipid A (3D-MPL) was developed by GlaxoSmithKline (GSK, London, UK) and approved as an adjuvant. 3D-MPL has been widely used as a component of the adjuvant system AS04 in the bivalent human papillomavirus (HPV) vaccine (Cervarix^®^, GSK) and hepatitis B virus vaccine (Fendrix^®^, GSK), as well as the adjuvant system AS01 in the varicella-zoster virus vaccine (Shingrix^®^, GSK), respiratory syncytial virus vaccine (Arexvy^®^, GSK), and malaria vaccine (Mosquirix^®^, GSK).

Alcaligenes spp., including A. faecalis, are aerobic, non-fermentative, Gram-negative lymphoid-tissue-resident commensal bacteria and inhabit Peyer’s patches in the small intestine [18,19,20,21,22,23,24,25]. Alcaligenes LPS can induce immunoglobulin (Ig) A and T helper (Th) 17 cells without causing excessive inflammation [26]. IgA is a major component of the mucosal immune system and is essential for protection against microbial infections on the mucosal surfaces, including the respiratory and gastrointestinal tracts. Th17 cells, a subset of T cells, also reside mainly at mucosal sites, secrete various cytokines, including interleukin (IL)-17, and play a crucial role in mucosal immunity [27]. Th17 responses can maintain epithelial barriers, recruit neutrophils, and confer protection against extracellular bacterial and fungal infections [28,29,30,31]. IL-17 has also been reported to regulate the generation and transcytosis of pathogen-specific IgA [32]. Thus, Alcaligenes LPS could be an adjuvant in vaccines against these types of infections; particularly, the induction of pathogen-specific IgA antibodies is unique and useful, since currently available vaccine adjuvants did not induce antigen-specific IgA. We synthesized a novel lipid A derived from A. faecalis (ALA) as a promising vaccine adjuvant [33]. The efficacy of the vaccines containing ALA as an adjuvant has been demonstrated in experimental infections with bacteria, including Haemophilus influenzae type B [34] and Streptococcus pneumoniae [35]. In H. influenzae infection, ALA enhanced antibody production by B cells [34]; in S. pneumoniae infection, ALA induced high pneumococcal-specific IgA and IgG responses as well as Th17 responses [35].

The safety of 3D-MPL has been well-documented in health and disease conditions clinically, including various immune-mediated diseases; for example, bivalent HPV vaccination was not associated with the development of autoimmune diseases and neurological diseases, including Guillain–Barré syndrome and multiple sclerosis (MS) [36,37]. On the other hand, although the safety of ALA has also been confirmed experimentally when ALA was used as a vaccine adjuvant, the unique immunomodulatory characteristics of ALA raised concerns about the usage of ALA in immune-mediated diseases, whose immunopathology was associated with IgA or Th17 responses. Potential pathogenic roles of IgA have been reported in various diseases, including coeliac disease [38] and inflammatory bowel disease (IBD) [39]. IgA nephropathy is the most common glomerulonephritis that can progress to renal failure [40]. The pathology of IgA nephropathy was characterized by IgA deposition in the glomerular mesangium of the kidneys; both non-specific IgA antibodies and tissue-specific IgA autoantibodies have been reported to be pathogenic [41,42]. Th17 cells have also been involved in the pathogenesis of various autoimmune and inflammatory diseases, including psoriasis, IBD, and rheumatoid arthritis (RA) [28,43]. IL-17 inhibitors have been used clinically to treat immune-mediated inflammatory diseases, including psoriatic arthritis [44].

MS is an inflammatory demyelinating disease of the central nervous system (CNS) that affects about 2 million people worldwide [45,46]. The etiology of MS has been proposed to have both immune-mediated and environmental components, particularly virus infection [47]. Viral etiology of MS has been supported clinically and experimentally [48,49]. Theiler’s murine encephalomyelitis virus (TMEV) is a non-enveloped, positive-sense single-stranded RNA virus that belongs to the family Picornaviridae [50]. TMEV infection in mice has been used as an animal model of MS, since TMEV infection can induce an inflammatory demyelinating disease in the CNS, similar to human MS [51,52,53]. In the TMEV model, CNS-infiltrating immune effectors and viral persistence in the macrophages and glial cells in the white matter have been shown to play a pathogenic role in the induction of CNS lesions [54]. Similar to human MS, pathogenic roles of various immune effectors, including IgG antibodies, CD4^+^ and CD8^+^ T cells, and macrophages, have been reported in the TMEV model. More recently, among the immune effectors, IgA and Th17 cells have been proposed as potential effectors in the pathophysiology of both MS and the TMEV model [55,56,57,58]. In MS, although approved vaccines were considered overall safe and not associated with disease activities [59,60], the risk of disease exacerbation has been reported clinically and experimentally, depending on the type of vaccines [61,62]. Thus, for approval of novel vaccine/adjuvant candidates, the safety of the candidates needs to be tested, particularly when their immunomodulatory effects are unique.

In this study, we aimed to investigate whether ALA injections could be safe in the TMEV model. Since ALA administrations can change mucosal IgA responses, leading to changes in the gut microbiota that has been associated with pathophysiology of MS and the TMEV model [55,63], we also examined the effect of ALA injections on the gut microbiota in the TMEV model. We administered TMEV-infected mice with ALA at various time points, intranasally or subcutaneously. We monitored neurological signs and body weight changes for 7 weeks, and examined neuropathology, viral persistence, anti-viral IgG isotype and IgA titers, anti-viral lymphoproliferation, cytokines, and the gut microbiota. We observed no detrimental changes in clinical signs, neuropathology, or immune responses in any mice. Thus, we demonstrated the safety of ALA usage in the TMEV model. On the other hand, although we anticipated that intranasal, but not subcutaneous, ALA administrations would influence the gut microbiota compositions, both intranasal and subcutaneous ALA administrations changed the overall microbiota profile in fecal samples with increases in the phylum Bacteroidota.

2. Materials and Methods

2.1. Animal Experiments

We purchased 4- to 5-week-old SJL/J female mice (JAX^®^ Mice Strain) from the Jackson Laboratory Japan (Yokohama, Japan). Mice were maintained under specific pathogen-free conditions in the animal care facility at Kindai University Faculty of Medicine (Osaka, Japan). The Institutional Animal Care and Use Committee of Kindai University Faculty of Medicine approved all experimental procedures performed according to the outlined criteria by the National Institutes of Health (NIH) [64].

We inoculated 6-week-old mice intracerebrally with 2 × 10^5^ plaque-forming units (PFUs) of the Daniels (DA) strain of TMEV on day 0 [53]. ALA was chemically synthesized, as described previously [33], reconstituted in dimethyl sulfoxide (DMSO) at a concentration of 1 mg/mL, and stored at −20 °C. The ALA solution was prepared in sterile phosphate-buffered saline (PBS) at a concentration of 1 µg/15 µL and sonicated for 10 s. We conducted two independent experiments (Expts): mice were divided into five and four groups in Expts. 1 and 2, respectively (five mice/group) (Table 1).

In Expt. 1, the control group was intranasally injected with vehicle (15 µL of PBS containing 1 µL of DMSO) on days −1, 6, 13, 20, 27, 34, and 41. Three groups of mice were injected intranasally with 15 µL of the ALA solution (1 µg/mouse) on days −1 and 6 (acute group), on days 20, 27, 34, and 41 (chronic group), or days −1, 6, 13, 20, 27, 34, and 41 (whole group). One group of mice was subcutaneously injected with 200 µL of the ALA solution (1 µg/mouse) on days −1, 6, 13, 20, 27, 34, and 41 (subcutaneous group). For the microbiota analyses, we used a group of naïve mice with no TMEV or ALA injection (n = 5). In Exp. 2, three groups of mice were intranasally injected with 15 µL of the ALA solution (1 µg/mouse) on days −1 and 6 (acute group), days −14 and −7 (prophylactic group), or days −14, −7, −1, and 6 (prophylactic + acute group). The control group received 15 µL of vehicle on days −14, −7, −1, and 6.

2.2. Neurological Signs Scored by Righting Reflex

To determine the effects of ALA on TMEV-infected mice, we monitored their neurological signs and body weight changes daily for 7 weeks. The neurological signs were evaluated by scoring the impairment of the righting reflexes; the proximal end of the mouse’s tail was grasped and twisted to the right and then left sides. The impaired righting reflexes were scored as follows: 0, no sign, a healthy mouse resisted being turned over; 1, the mouse was flipped onto its back but immediately righted itself on one side; 1.5, the mouse was flipped onto its back but immediately righted itself on both sides; 2, the mouse righted itself in 1 to 5 s; 3, righting took more than 5 s; and 4, the mouse could not right itself [65,66].

2.3. Neuropathology and Immunohistochemistry

Seven weeks post-infection (p.i.), we killed mice with isoflurane (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan), collected blood from the heart, and perfused mice with PBS, followed by a 4% paraformaldehyde (PFA, FUJIFILM Wako Pure Chemical Corporation) solution in PBS. The spinal cord was harvested, divided into 10 to 13 transversal segments, and embedded in paraffin. Then, 4 μm thick sections were made using the HM 325 Rotary Microtome (Thermo Fisher Scientific., Waltham, MA, USA) and stained with Luxol fast blue (Solvent blue 38; MP Biomedicals, Irvine, CA, USA) for myelin visualization. We evaluated neuropathology as described previously [66,67]. The spinal cord section was divided into four quadrants: the ventral, the dorsal, and two lateral funiculi. Any quadrant containing meningitis, perivascular cuffing (inflammation), or demyelination was given a score of 1 in that pathological class. The total number of positive quadrants for each pathological class was determined, then divided by the total number of quadrants present on the slide, and multiplied by 100 to give the percentage involvement for each pathological class. An overall pathology score was also determined by calculating the percentage of involvement, assigning a positive score if any pathology was observed in the quadrant [68]. We visualized TMEV antigens and IgA by immunohistochemistry using an anti-TMEV antibody [66,69] and a rat anti-mouse IgA monoclonal antibody (1:2000 dilution, clone 11-44-2, Beckman Coulter, Brea, CA, USA). The antibody/antigen complexes were visualized using a secondary antibody Histofine MAX-PO kit (Nichirei Biosciences, Tokyo, Japan) or anti-rat IgG-peroxidase (Nichirei Bioscience) with a 3,3′-diaminobenzidine (DAB) substrate solution (FUJIFILM Wako Pure Chemical Corporation). We divided the spinal cord sections into four quadrants: the ventral, dorsal, and two lateral funiculi and counted the numbers of TMEV antigen^+^ and IgA^+^ cells in each quadrant under a light microscope using a 20× objective lens [66].

2.4. Anti-TMEV Antibody Enzyme-Linked Immunosorbent Assay (ELISA)

To quantify anti-TMEV antibodies, we collected blood from the heart and feces from the rectum and anal canal 7 weeks p.i. For serum collection, we centrifuged the blood at 2775× g at 4 °C for 20 min. To extract IgA from the fecal samples, we dissolved 10 mg of feces in 100 μL of PBS containing a proteinase inhibitor cocktail tablet (1 tablet/25 mL PBS, cOmplete™ EDTA-free, Roche Diagnostics, Mannheim, Germany). The fecal pellets were broken with a toothpick, vortexed for 1 min to form a uniform solution, and then centrifuged at 17,344× g for 10 min at 4 °C. The supernatants were collected from the homogenate, diluted four times with PBS containing the proteinase inhibitor, and centrifuged at 17,344× g at 4 °C for 10 min. We collected the supernatants and used them for the fecal IgA analysis. We quantified anti-TMEV antibody levels by an enzyme-linked immunosorbent assay (ELISA), as described previously [55]. Ninety-six-well flat-bottom Nunc-Immuno plates (Thermo Fisher Scientific) were coated with 50 μL/well of TMEV antigens or anti-mouse IgA antibody (20 ng/well) and incubated at 4 °C overnight. After blocking with an assay diluent composed of 10% fetal bovine serum (FBS, Sigma-Aldrich Japan, Tokyo, Japan) and 0.2% Tween 20 (FUJIFILM Wako Pure Chemical Corporation) in PBS, we added serum samples diluted with the assay diluent (2^7^ to 2^28^) or the fecal supernatant samples without dilution to the plates and incubated the plates for 75 min at room temperature (RT). After washing with a buffer containing 0.2% Tween 20 in PBS, a horseradish peroxidase-conjugated anti-mouse IgA secondary antibody (2000-fold dilution, Thermo Fisher Scientific) was added to the plates and incubated for 75 min at RT. After washing the plate with the washing buffer, we detected the immunoreactive complexes using the BD OptEIATM TMB Substrate Reagent Set (BD Biosciences, San Jose, CA, USA), according to the manufacturer’s instruction, and stopped the reaction with a 2N sulfuric acid solution (H_2_SO_4_, FUJIFILM Wako Pure Chemical Corporation). The absorbances were measured at 450 nm using the Synergy H1 Hybrid Multi-Mode Microplate Reader (Agilent Technologies, Santa Clara, CA, USA). Serum anti-TMEV antibody titers were determined as the highest reciprocal of the dilution that had an absorbance higher than the mean plus two standard deviations of naïve serum samples at 2^7^-fold dilution [66]. The mean absorbances of naïve serum samples at 2^7^-fold dilution were as follows: anti-TMEV IgG1, 0.072; anti-TMEV IgG2c, 0.062; and anti-TMEV IgA, 0.058. Fecal anti-TMEV IgA titers were shown as the absorbances of undiluted fecal samples; the mean absorbance of naïve fecal samples was 0.075.

2.5. Lymphoproliferative Assays

To examine TMEV-specific cellular immune responses, we harvested the spleen from all groups of mice 7 weeks p.i. To make a single-cell suspension, we mashed the spleen on a metal mesh with 50 μm pores, using the plunger of a 5 mL syringe, and isolated splenic mononuclear cells (MNCs), using Histopaque^®^-1083 (Sigma-Aldrich Japan). The MNCs were cultured in RPMI-1640 medium (Sigma-Aldrich Japan) supplemented with 10% FBS, 2 mM L-glutamine (Sigma-Aldrich Japan), and 50 mM β-mercaptoethanol (FUJIFILM Wako Pure Chemical Corporation) at 2 × 10^5^ cells/well in a 96-well plate (Sumitomo Bakelite, Tokyo, Japan). We incubated the MNCs with or without TMEV at a multiplicity of infection (MOI) of 5 at 37 °C with 5% CO_2_ for 5 days. To quantify the levels of lymphoproliferative responses to TMEV, we added 3 μL/well of a Cell Counting Kit-8 (CCK-8) solution (Dojindo Laboratories, Kumamoto, Japan) and incubated for the last 24 h. We conducted the cell culture in triplicate and measured the absorbance at 450 nm using the Synergy H1 Hybrid Multi-Mode Microplate Reader. The results were expressed as stimulation indexes: (mean absorbance of wells stimulated with TMEV)/(mean absorbance of wells without stimulation) [66].

2.6. Cytokine ELISAs

To determine whether ALA injections could alter cytokine production, we cultured splenic MNCs at 8 × 10^6^ cells/well in 6-well plates and stimulated them with TMEV at an MOI of 1 or with concanavalin A (ConA) at 5 μg/mL (Sigma-Aldrich Japan). We incubated the plates at 37 °C with 5% CO_2_ for 2 days. The culture supernatants were collected and stored at −80 °C until examined. The concentrations of IL-17A (BioLegend., San Diego, CA, USA), interferon (IFN)-γ (BD Biosciences), IL-4 (BD Biosciences), and IL-10 (BD Biosciences) were quantified in duplicate wells, using ELISA kits, according to the manufacturers’ instructions. The lower detection limits were as follows: IL-17A, 15.6 pg/mL; IFN-γ, 32.5 pg/mL; IL-4, 15.6 pg/mL; and IL-10, 31.3 pg/mL. We did not detect any cytokines from MNCs under unstimulated conditions.

2.7. Microbiome Analyses

DNA was isolated from the ileal contents and feces, using the QIAamp^®^ Fast DNA Stool Mini Kit (QIAGEN, Germantown, MD, USA), according to the manufacturer’s instruction [55]. Bacterial DNA was amplified by PCR, using a primer set for V3 and V4 regions of bacterial 16S rRNA. The primer sequences were as follows: forward primer = 5′-TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG-3′; reverse primer = 5′-GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGTATCTAATCC-3′. The bacterial DNA library was processed and sequenced using the Illumina MiSeq^®^ System (Illumina, San Diego, CA, USA). Fastq files were processed and visualized by QIIME 2. Using QIIME 2, we also conducted alpha diversity analyses of the gut microbiota: the Faith’s phylogenetic diversity for richness analysis, Pielou’s evenness index for evenness analysis, and Shannon index for a combination of richness and evenness. A taxonomic bar plot was drawn, using OriginPro 2025 (OriginLab Corporation, Northampton, MA, USA). We conducted a principal component analysis (PCA), using an R program “prcomp,” and drew a graph of PCA with 80% confidence interval for each group, using R version 4.5.2 and the packages “ggplot2” and “dplyr.” As a functional analysis, we conducted a predictive metagenome analysis, using a PICRUSt2 program, on QIIME 2 [70]. Fastq data and processed data have been deposited to the Sequence Read Archive (SRA) at NCBI (Bioproject accession no. PRJNA1297922; http://www.ncbi.nlm.nih.gov/bioproject/1297922, accessed on 21 January 2026).

2.8. Anti-ALA Antibody ELISA and Pattern Matching

We quantified serum antibody levels against ALA using an anti-glycolipid antibody ELISA with slight modifications [71]. We diluted 40 μL of the stock ALA solution (in DMSO, at 1 mg/mL) with 10 mL of 100% ethanol (final ALA concentration was 4 μg/mL). Ninety-six-well flat-bottom Nunc-Immuno plates were coated with 50 μL/well of the ALA solution (200 ng ALA/well), air-dried in a biological safety cabinet class II (VH-1302BH-2A2, Nippon Medical & Chemical Instruments Co., Ltd., Tennoji-ku, Osaka, Japan) for 40 min at RT, and stored overnight at 4 °C. In control wells, we added 50 μL of 0.4% DMSO in ethanol without ALA. After blocking with 80 μL of 1% bovine serum albumin (BSA) (ICPbio Ltd., Henderson, Auckland, New Zealand) in PBS at RT for 30 min, we added 50 μL of serum samples diluted at 1:64 in blocking solution to the wells and incubated for 90 min at RT. Plates were washed with 300 μL of 0.1% BSA in PBS three times, and incubated with a peroxidase-conjugated anti-mouse F(ab’)2 antibody (1:5000 dilution; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) for 90 min at RT. After washing the plate with 0.1% BSA in PBS three times, we detected the immunoreactive complexes using the TMB Substrate Reagent Set (BD Biosciences) and measured the absorbance at 450 nm (Abs 450 nm) by subtracting the absorbance of the control wells.

To examine the associations between the anti-ALA antibody levels and microbiome data, we conducted pattern matching, using R. We compared anti-ALA antibody levels (Abs 450 nm) versus the relative abundances of bacteria using the phylum and species level references [72]. Pearson correlation coefficient (r) and p value were calculated using linear fit functions in OriginPro 2025. Correlation analyses were performed only for bacterial taxa with fewer zero values (<50%) per mouse group (at least 3 samples per group had the relative abundance values of the bacterial taxa of interest) [73,74]. Correlations with p < 0.05 were considered statistically significant. We used the interpretation for the correlation by Mukaka (2012) [75] as follows: 0.7 to 1.0 (−0.7 to −1.0), high positive (negative) correlation; 0.5 to 0.7 (−0.5 to −0.7), moderate positive (negative) correlation; 0.3 to 0.5 (−0.3 to −0.5), low positive (negative) correlation; and 0 to 0.3 (0 to −0.3), negligible correlation [76].

2.9. Statistical Analyses

Using OriginPro 2025, the Kruskal–Wallis test with Dunn’s post hoc test and analysis of variance (ANOVA) with Fisher’s post hoc LSD test were conducted for nonparametric and parametric data, respectively. p < 0.05 was considered a significant difference among the groups.

3. Results

3.1. Effects of ALA on Clinical Signs in TMEV Infection

Intracerebral inoculation of SJL/J mice with TMEV has been known to cause a biphasic disease. Infected mice develop acute polioencephalomyelitis during the acute phase, 1 week p.i., and progressive inflammatory demyelination during the chronic phase, 1 month p.i. To determine whether ALA could affect TMEV infection, we inoculated mice with TMEV on day 0 and administered ALA to mice at various time points, as shown in Table 1. Since ALA has been used experimentally as an adjuvant for intranasal vaccinations, we administered ALA intranasally. In Expt. 1, ALA was administered at the early time points on days −1 and 6 (acute group), late time points on days 20, 27, 34, and 41 (chronic group), or weekly (seven time points) on days −1, 6, 13, 20, 27, 34, and 41 (whole group). Another group of TMEV-infected mice was administered ALA subcutaneously seven times weekly to assess the effect of the ALA administration route. Control TMEV-infected mice received intranasal vehicle injections weekly. We monitored the neurological signs and body weights for 7 weeks. Regardless of the ALA administration schedule, all groups had mild impairment of righting reflexes with weight loss during the acute phase, and recovered completely. Then, 1 month p.i., all mice started to develop chronic progressive neurological signs. We found no significant differences in neurological signs or body weight changes among mice receiving ALA and control mice (Figure 1).

In Expt. 2, we examined whether ALA administrations prior to TMEV infection could affect the TMEV-induced disease, as shown in Table 1. One prophylactic group received ALA on days −14 and −7 (prophylactic group); another prophylactic group received ALA on days −14, −7, −1, and 6 (prophylactic + acute group). For comparison, one group of mice received ALA on days −1 and 6 (acute group), and the control TMEV-infected mice received vehicle (control group). All mice received ALA or vehicle intranasally. We found that all groups had similar neurological scores and body weight changes during the acute and chronic phases (Supplementary Figure S1). Thus, regardless of the administration schedule, ALA administrations had no effects on neurological signs or body weight changes of TMEV-infected mice.

3.2. Effects of ALA on Neuropathology in TMEV Infection

Since TMEV infection has been known to induce inflammatory demyelination in the white matter of the spinal cord, 1 month p.i. (chronic phase) [77], we investigated whether ALA administrations could affect neuropathology in TMEV-infected mice. For myelin visualization, we stained spinal cord sections with Luxol fast blue and compared the severities of neuropathology, using a spinal cord pathology scoring system among the groups receiving ALA and the control group. Regardless of the ALA administration schedules, all groups receiving ALA had similar distributions and severities of meningitis, perivascular cuffing (inflammation), and demyelination in the white matter of the spinal cord, comparable to the control group (Figure 2A,B and Figure S2A,B). In all groups, demyelination was often accompanied by perivascular cuffing composed of MNCs, and was observed more in the lateral and ventral funiculi than in the dorsal funiculus, particularly in the thoracic segments. We also quantified the neuropathology scores and found no statistical differences in the levels of meningitis, inflammation, or demyelination in the spinal cord (Figure 2C and Figure S2C).

3.3. Viral Persistence and IgA Deposition in the Spinal Cord Following TMEV Infection

TMEV can persistently infect the spinal cord during the chronic phase; viral persistence has been known to be essential for causing inflammatory demyelination [78]. We investigated whether ALA administrations could alter the numbers and localizations of viral antigen^+^ cells among the groups. In all groups, viral antigen^+^ cells were observed around inflammatory demyelinating lesions in the white matter of the spinal cord (Figure 3A,B and Figure S3A,B). Viral antigen^+^ cells were detected mainly in the lateral and ventral funiculi; there were no differences in the numbers and localizations of the viral antigen^+^ cells among the groups (Figure 3C and Figure S3C).

IgA is known to play a major role in mucosal immunity, protecting mucosal surfaces from pathogens [79]. Previously, we demonstrated IgA^+^ cells and IgA deposition around inflammatory demyelinating lesions of the spinal cord during the chronic phase of TMEV infection, suggesting a role of IgA in pathophysiology [55]. To determine whether ALA administrations could alter the levels and distributions of IgA^+^ cells and IgA deposition, we performed immunohistochemistry using an antibody against mouse IgA. In all groups, we detected similar numbers and distributions of IgA^+^ cells in the meninges, perivascular spaces, and the parenchyma of inflammatory demyelinating lesions, which were accompanied by parenchymal IgA deposition (Figure 3D,E and Figure S3D,E). There were no statistical differences in the numbers of IgA^+^ cells among the groups receiving ALA and the control group receiving vehicle (Figure 3F and Figure S3F).

3.4. Anti-TMEV Antibody Isotype and Lymphoproliferative Responses

To determine whether ALA administrations could alter anti-TMEV humoral and cellular responses, we harvested sera, feces, and MNCs from TMEV-infected mice receiving ALA or vehicle 7 weeks p.i. Using ELISAs, we titrated serum anti-TMEV IgG1, IgG2c, and IgA antibodies. In all groups, substantial levels of anti-TMEV IgG1, IgG2c, and IgA were detected. Although anti-TMEV IgG1 titers were slightly lower in the ALA-injected groups than in the control group, anti-TMEV IgG2c levels were similar among the groups (Figure 4A,B), resulting slightly higher in the IgG2c/IgG1 ratios, which reflected Th1/Th2 balance, in the ALA-injected groups than in the control group (Figure 4C). Since ALA has been reported to enhance antigen-specific IgA production when ALA was given as a vaccine adjuvant [35], we titrated anti-TMEV IgA in both serum and fecal samples. We found no significant differences in serum or fecal IgA titers among the ALA-injected groups and the control groups (Figure 4D,E). In TMEV infection, TMEV-specific lymphoproliferative responses, reflecting anti-viral T-cell responses, have been shown to play either a protective role in viral clearance or a detrimental role in inflammatory demyelination [80]. Using splenic MNCs from TMEV-infected mice, we analyzed anti-TMEV lymphoproliferative responses. Regardless of the different ALA administration schedules, we found no significant differences among the groups receiving ALA and the control group (Figure 4F and Figure S4).

3.5. Cytokine Productions in TMEV Infection

To determine whether ALA injections could alter the production of Th-cell-related cytokines, we quantified concentrations of IL-4, IL-10, IL-17, and IFN-γ, using MNC culture supernatants stimulated with TMEV or a mitogen, ConA, by ELISAs. In TMEV stimulation, we detected higher levels of IFN-γ production in the chronic and subcutaneous groups than in the control group (p < 0.05, ANOVA); IL-17 was not detectable in any groups (Figure 5A and Figure S5A). The amounts of IL-4 and IL-10 were similar among the groups (Figure 5B and Figure S5B). In ConA stimulation, we detected similar levels of IL-17, IFN-γ, IL-4, and IL-10 concentrations among the groups (Figure 5C,D and Figure S5C,D).

3.6. Gut Microbiota Changes by ALA Injections in TMEV Infection

Since LPS has been shown to affect the gut microbiota composition [81], we determined the effects of ALA injections on the gut microbiota using 16S rRNA sequencing of fecal samples from TMEV-infected mice. First, we compared the overall microbiome patterns among the samples by PCA of fecal microbiome data. As we reported previously [55], PCA did not distinguish between the naïve uninfected versus control TMEV-infected groups, indicating that TMEV infection alone did not cause overall microbiota changes. In contrast, in Expt. 1, PCA clearly separated all ALA-injected groups from the naïve and control TMEV-infected groups on the principal component (PC) 1 axis (Figure 6A,B). The proportion of variance indicated that 68% of the variance among the samples was explained by the distributions on the PC1 axis (Figure 6C). The factor loading for PC1 indicated that the genus Alistipes contributed to the PC1 separation strongly (Figure 6D). Similarly, in Expt. 2, PCA separated the control group significantly from two groups receiving ALA during the acute phase (acute group and prophylactic + acute group) on the PC2 axis, although PCA did not separate the control group from the group receiving ALA prior to TMEV infection (prophylactic group) (Supplementary Figure S6).

Next, we compared alpha diversity indexes among the groups. We used the Faith’s phylogenetic diversity (PD) index for the comparison of richness (number of species), Pielou’s evenness index for the comparison of evenness, and Shannon index for the comparison of the combination of richness and evenness. We found no differences in the microbial diversities between the naïve and control TMEV-infected groups. On the other hand, ALA injections during the acute phase (acute group) decreased the microbial diversities with significant differences in Faith’s PD and Shannon indexes, compared with the control group in Expt. 1 (Figure 7). We found a similar difference in the Faith’s PD index between the control and acute groups in Expt. 2 (Supplementary Figure S7).

We also compared the relative abundances of bacterial compositions in feces among the groups (Supplementary Tables S1–S4). Using a phylum-level reference, we found that the phyla Bacillota and Bacteroidota were dominant in all groups (Figure 8A and Figure S8A). Since the ratio between the phyla Bacillota and Bacteroidota has been associated with various diseases [81,82], we calculated the ratios using the fecal samples. Compared with the control TMEV-infected group, the ratios of all ALA-injected groups were lower with a significant difference between the control and subcutaneous groups (Supplementary Figure S9A). Using a species-level reference, we compared the naïve group with the control TMEV-infected group. We detected the genus Paramuribaculum only in the naïve group, indicating that TMEV infection altered the fecal microbiota, as we described previously [55]. Compared with the control group, we found an increase in the genus Alistipes in all ALA-injected mice in Expt. 1 (Figure 8B). In Expt. 2, although the group receiving ALA before TMEV infection (prophylactic group) had no significant changes compared with the control group, two groups receiving ALA during the acute phase (acute group and prophylactic + acute group) had an increase in some bacterial taxa, including Bacteroides acidifaciens, but not the genus Alistipes (Supplementary Figure S8B). Since we did not detect the genus Alistipes in any fecal samples in Expt. 2, B. acidifaciens appeared to increase as an alternative bacterium in place of the genus Alistipes; both genera Alistipes and B. acidifaciens belong to the phylum Bacteroidota.

To explore potential pathways that could work in the fecal microbiota community, we conducted a predictive metagenome analysis, using a PICRUSt2 program (Supplementary Table S5). We found a significant increase in fecal bacteria associated with three LPS- and lipid A-related pathways in all four TMEV-infected groups receiving ALA than in the control TMEV-infected group: lipid IV_A_ biosynthesis [83], CMP-3-deoxy-D-manno-octulosonate biosynthesis, and Kdo transfer to lipid IV_A_ III (Chlamydia) [84] (Table 2 and Table S5), suggesting that ALA injections could increase the abundance of fecal bacteria-producing LPS [85].

We also determined the changes in microbiota following ALA injections in the ileal content of TMEV-infected mice (Supplementary Tables S6 and S7). We compared the overall microbiome patterns by PCA among the groups (Supplementary Figure S10A,B). Unlike the fecal samples, PCA did not separate the ALA-injected groups from the naïve and control TMEV-infected groups. We examined the alpha diversities and found no statistical differences in the Faith’s PD, Pielou’s evenness, or Shannon indexes among the groups (Supplementary Figure S10C–E). In the relative abundances of the gut microbiota, we also found no statistical differences among the ALA-injected and control TMEV-infected groups at either the phylum or species levels (Supplementary Figure S10F,G). On the other hand, when we compared the relative abundances of the gut microbiota between the naïve and control TMEV-infected groups, we found significant increases in the relative abundances of the phyla Bacillota (p < 0.01, Kruskal–Wallis test with Dunn’s post hoc test) and Actinomycetota (p < 0.05, Kruskal–Wallis test with Dunn’s post hoc test) at the phylum level, and decreases in bacteria belonging to the family Muribaculaceae, including the genus CAG-873, in the control TMEV-infected group.

3.7. Correlations of Anti-ALA Antibodies and Fecal Bacterial Abundance

LPS and lipid A have been shown to be antigenic [86,87]; we examined whether ALA injections induced anti-ALA antibody production. We quantified serum anti-ALA antibodies by ELISA and compared antibody levels among the groups. Although anti-ALA antibodies were detectable in all groups, including the naïve and control TMEV-infected groups, only the S.C. group had significantly higher anti-ALA antibodies than the other groups (Figure 9A). Then, using pattern-matching and correlation analyses, we determined the association between anti-ALA antibody levels and the relative abundance of fecal bacteria. Using a phylum-level reference, we found significant high positive and negative correlations of anti-ALA antibody levels with the relative abundance of the phylum Bacteroidota (r = 0.71, p < 0.05) and the phylum Bacillota (r = −0.71, p < 0.05), respectively (Figure 9, Supplementary Table S8). Using a species-level reference, we found high positive correlations in three bacterial taxa: Duncaniella sp001689575 (r = 0.82, p < 0.01); Duncaniella muris (r = 0.80, p < 0.01); and Moraxellaceae abundance (r = 0.76, p < 0.01), and high negative correlations in four bacterial taxa: Lachnospiraceae (r = −0.9, p < 0.001); Ruminiclostridium siraeum (r = −0.83, p < 0.01); Clostridia 258483 (r = −0.78, p < 0.01); and RUG13077 (r = −0.75, p < 0.05) (Supplementary Table S9).

4. Discussion

Previously, Wang et al. [88] demonstrated that experimental ALA injections in naïve mice as a vaccine adjuvant did not cause severe side effects. In this study, we aimed to investigate the safety of ALA in the TMEV model of MS, in which IgA and Th17 cells have been shown to play a potential pathogenic role. We used SJL/J mice, which have been most widely used in the TMEV model, since almost 100% of TMEV-infected mice developed a robust, full-blown, inflammatory demyelinating disease with similar clinical and pathological severities among the infected mice. On the other hand, to detect small increases in encephalitogenic activity or changes in disease susceptibility, a suboptimal dose of TMEV infection [89,90] or TMEV-resistant mouse strains, including C57BL/6 and BALB/c mice [91], have been used, respectively. We did not use these approaches since IgA and Th17 responses have not been characterized in these experimental settings.

High levels of virus persistence and TMEV-specific immune responses have been associated with exacerbation of TMEV-induced inflammatory demyelinating diseases by direct viral infection of myelin-forming oligodendrocytes (viral pathology) and immune-mediated tissue damage (immunopathology), respectively [66,92]. In the TMEV model, although anti-viral immune responses can contribute to viral clearance, uncontrolled anti-viral immune responses have been shown to exacerbate inflammatory demyelination [55,56]. A lack of exacerbation in the TMEV model could be due to ALA injections having no effect on overall TMEV-specific immune responses. In this study, regarding humoral responses, we found no increase in anti-TMEV antibody titers in any ALA-injected groups. Regarding cellular immune responses, we also found no differences in the levels of TMEV-specific lymphoproliferation (T-cell responses) among the groups. Thus, ALA injections did not affect overall anti-viral antibody or T-cell responses, which most likely resulted in no change in viral pathology or immunopathology, although ALA injections slightly changed TMEV-specific Th1/Th2 responses (see below).

On the other hand, TMEV-specific IFN-γ production was higher in the ALA-injected groups than in the control group. The higher TMEV-specific IFN-γ production in the ALA-injected groups was consistent with slightly higher anti-TMEV IgG2c/IgG1 ratios, reflecting Th1/Th2 balance, in the ALA-injected groups than in the control group. Since Sun et al. [93] previously reported that ALA-treated dendritic cells could polarize Th1 differentiation, ALA may contribute to Th1 differentiation depending on the condition. In the current study, we also stimulated splenic MNCs with the mitogen ConA to assess whether ALA injections could affect systemic Th-cytokine profiles. We found no differences in mitogen-induced cytokine profiles among the groups, indicating that ALA injections did not alter Th-cell activation or differentiation systemically.

Communications between the immune system and gut microbiota have been shown to influence the components of each other. ALA was derived from a component of the resident gut microbiota and had immunomodulatory effects. Thus, we investigated the gut microbiota profiles in ALA-injected mice and found the changes in the microbiota. Previously, we demonstrated that an adjuvant injection alone could alter the gut microbiota in experimental mice [94]. We injected C57BL/6 mice with two adjuvants: complete Freund’s adjuvant (CFA) subcutaneously and pertussis toxin (PT) intraperitoneally; we found the changes in the gut microbiota in the ileum and feces. Compared with naïve mice, the overall microbial profiles visualized by PCA demonstrated changes in the microbial composition in feces, but not in the ileum, of adjuvant-injected mice, where the genera Lachnospiraceae NK4A136 group and Alistipes contributed to the microbial changes. These findings were similar to our current study, where ALA injections altered the overall microbiota profiles in feces, but not in the ileum, where Alistipes contributed to the alteration (Figure 6).

Associations between the gut microbiomes and disease activities have been reported to differ among the distinct anatomical sites [95,96]. In the experimental autoimmune encephalomyelitis (EAE) model of MS, the overall microbiota compositions in the ileum, but not in feces, have been correlated with clinical and histological severities [63]. In the EAE model, the CFA-PT adjuvant injection alone reduced the overall microbial differences between the fecal and ileal microbiota [94]. On the other hand, our current TMEV model study showed that PCA separated the fecal and ileal microbiota clearly as distinct populations in all groups, including the naïve, control, and ALA-injected groups (Supplementary Figure S11). A lack of changes in the ileal microbiota, but not the fecal microbiota, in our current study may also explain no differences in disease activities. The above differences between different adjuvant injections (i.e., CFA-PT versus ALA) may be due to several differences in experimental settings, such as the route of adjuvant administrations (intranasal, subcutaneous, and intraperitoneal), mouse species (C57BL/6 versus SJL/J), and animal models (EAE versus TMEV).

In the fecal microbiota, using a phylum-level reference, we found that the gut microbiota was predominantly composed of the phyla Bacillota and Bacteroidota (Figure 8). Although the ratio of the phyla Bacillota/Bacteroidota has been associated with the pathogenesis of several diseases, including IBD and MS [82,97,98,99], we found no difference between the naïve and the control TMEV-infected groups. Using a species-level reference, we found that ALA injections increased several bacterial taxa in the phylum Bacteroidota, including the genus Alistipes and B. acidifaciens. The genus Alistipes is anaerobic Gram-negative bacteria detected in the healthy gastrointestinal tract [100]. Several reports have suggested that Alistipes may play a protective role in various diseases, such as IBD [101,102,103,104,105,106,107], as well as a pathogenic role in other diseases [108], including MS. Schwerdtfeger et al. [109] reported that Alistipes might be associated with disease progression in MS. On the other hand, in the EAE model of MS, Alistipes was found to be negatively correlated with disease severity [63,110]. B. acidifaciens is also anaerobic Gram-negative bacteria [111], and has been reported to play a protective role in various diseases, including IBD [112,113,114,115]. In the EAE model, although Miller et al. [116] suggested a pathogenic role of B. acidifaciens, we previously demonstrated the potential protective role of B. acidifaciens in the feces and ileal contents [63]. Although the genera Alistipes and B. acidifaciens have been shown to affect pathophysiology in various diseases, our current study found that the changes in both bacterial genera following ALA injections were not associated with the clinical or histological outcomes of the TMEV model.

In the predictive metagenome analysis using PICRUSt2, we found significantly increased abundances of the bacterial species associated with three LPS- and lipid A-related pathways [83,84] in all ALA-injected groups (Table 2). These pathways included the components of the superpathway of Kdo_2_-lipid A biosynthesis [2,117]. Since Lipid A is synthesized by Gram-negative bacteria, ALA injections may increase Gram-negative bacteria. The microbiota changes induced by ALA injections may influence the induction of preventive anti-microbial immune responses following vaccinations, when ALA is co-administered with the microbial antigen of interest as a vaccine adjuvant.

Since ALA injections could induce anti-ALA antibody production, we compared anti-ALA antibody levels among the mouse groups. We found that only the S.C. group had significantly higher anti-ALA antibodies than the other groups. Since we detected similar levels of anti-ALA antibodies in the naïve group and control TMEV-infected group, and the TMEV-infected group receiving intranasal ALA injections, these anti-ALA antibodies seemed to be natural antibodies, but not induced by TMEV infection or intranasal ALA injection. Using pattern matching, we found high positive and negative correlations between anti-ALA antibody levels and the relative abundance of several bacterial taxa (Figure 9 and Supplementary Tables S8 and S9). Among the bacterial taxa that showed a high positive correlation, at the phylum level, we found only the Gram-negative phylum Bacteroidota. Using a species-level reference, we found three bacterial taxa: (i) Duncaniella sp001689575 and (ii) Duncaniella muris were Gram-negative, belonging to the family Muribaculaceae (previously known as S24-7) in the phylum Bacteroidota [118], and (iii) the family Moraxellaceae was also Gram-negative bacteria belonging to the phylum Pseudomonadota (formerly Proteobacteria) [119]. On the other hand, among the bacterial taxa showing a high negative correlation, we found only the phylum Bacillota (formerly Firmicutes) with a high correlation at the phylum level. Using a species-level reference, we found that all three bacterial taxa were Gram-positive, belonging to the class Clostridia in the phylum Bacillota [120]. Here, all positively correlated bacterial taxa were Gram-negative bacteria, and all negatively correlated bacterial taxa were Gram-positive bacteria.

Although anti-lipid A antibody specificities have been shown to depend on phosphates, acylation patterns, and saccharides, several anti-lipid A antibodies have been shown to be polyreactive and cross-react with different lipid A and other molecules, including DNA [86,121]. In the current experiment, we also quantified serum anti-ALA antibodies in C57BL/6 mice injected with a varicella-zoster virus vaccine (Shingrix®) containing an adjuvant AS01 [comprising 3-O-deacyl-4′-monophosphoryl lipid A (3D-MPL) and QS-21] or with AS01 alone [122]. We found higher anti-ALA antibody levels in AS01-injected mice than in control: anti-ALA antibody (Abs 450 nm): AS01-injected mice, 0.67 ± 0.23, and control mice, 0.22 ± 0.10, suggesting the presence of antibody epitopes commonly recognized by both anti-3D-MPL antibodies and anti-ALA antibodies. This could explain why the relative abundance of several bacterial taxa was positively or negatively correlated with anti-ALA antibodies; changes in free ALA-like lipid A and LPS, or bacteria sharing similar lipid A moieties, by anti-ALA antibodies may affect the relative abundance of the gut microbiome; the relative abundance of some Gram-positive bacteria might also be affected by the changes in some Gram-negative bacteria.

Several other factors could contribute to the microbial changes. We found the changes in microbiota by both intranasal and subcutaneous ALA administration routes. Intranasal ALA administration could influence the nasal microbiota and mucosal immunity by ALA-mediated activation of bacteria-specific IgA-producing cells, which can alter the bacterial compositions. By communication between distinct mucosal compartments, mucosal sites have been suggested to function together as a system-wide organ [123]. Thus, the change in the nasal microbiota/immunity could lead to changes in the overall gut microbiota and mucosal immunity. On the other hand, we anticipated that the subcutaneous ALA administrations would affect systemic, but not mucosal, immunity. In this study, however, the subcutaneous ALA administrations resulted in similar changes in the fecal microbiota to those of the intranasal administration. Furthermore, we observed substantial changes in the ileal microbiota by the subcutaneous administrations, but not by the nasal administrations. The gut microbiota has been shown to affect both mucosal and systemic immunities; the mucosal and systemic immunities can also affect each other [124]. Thus, distinct communications between the gut microbiota, mucosal and systemic immunities in the ileum versus the rectum/anal canal may explain the different ALA’s effects on the ileal and fecal microbiota. In addition, comparing the intranasal route with the subcutaneous route, several key steps can influence the effects of ALA, including the adsorption, metabolism, and clearance of ALA via the lymphatic system, the bloodstream, the liver, and possibly the enterohepatic circulation, excretion into the bile, and entering the intestine [125,126,127].

In summary, we demonstrated that ALA administrations did not exacerbate the TMEV model, whose pathophysiology has been associated with IgA and Th17 responses. Although ALA administrations did not affect TMEV-specific humoral or cellular responses, including IgA and Th17 responses, we found significant changes in the fecal microbiota in ALA-injected groups. The microbiota changes induced by ALA administrations may influence the induction of preventive anti-microbial immune responses following vaccinations, when ALA is co-administered with the microbial antigen of interest as a vaccine adjuvant.