AAV Gene Therapy for MPS IVA with Induction of Immune Tolerance via Oral Administration of Epitope Peptides of N-Acetylgalactosamine-6-sulfate Sulfatase

Abstract

Mucopolysaccharidosis IVA (MPS IVA) is caused by the accumulation of undegraded glycosaminoglycans due to the deficiency of the N-acetylgalactosamine-6-sulfate sulfatase (GALNS) enzyme. MPS IVA manifests as progressive systemic skeletal dysplasia. Gene therapy (GT) is potentially a one-time treatment in which the enzyme is continuously produced, circulated, and delivered to target tissues. However, immune responses to gene products can diminish therapeutic efficacy. We hypothesized that oral delivery of tolerogenic peptides induces immune tolerance to human GALNS (hGALNS) in MPS IVA mice, enhancing therapeutic efficacy. Neonatal mice deficient in mouse GALNS (mGALNS) were treated orally with three T-cell/B-cell epitope peptides or hGALNS protein on alternate days from day 3 after birth to day 20 before intravenous injection with AAV9 vectors encoding human GALNS on day 30. The results are encouraging, with anti-hGALNS antibodies undetectable in the plasma of orally administered peptide groups. hGALNS enzyme activities in plasma and tissues were higher in the orally treated groups than in the non-tolerized control group. Keratan sulfate levels in plasma, liver, and bone were normalized. Complete correction for heart vacuolization was achieved in peptide-treated groups, and partial correction for bone pathology was observed in all GT-treated groups. Overall, oral tolerance induction using immunodominant peptides promises to significantly enhance the efficacy of AAV-GT for MPS IVA.

1. Introduction

Mucopolysaccharidosis IVA (MPS IVA) is an autosomal recessive inherited metabolic disease grouped into lysosomal storage diseases (LSDs) because of the lysosomal accumulation of the undegraded substrate in the deficiency of the N-acetylgalactosamine-6-sulfate sulfatase (GALNS) enzyme. The incidence of MPS IVA is approximately one in 250,000 births [1]. Undegraded glycosaminoglycans (GAGs), including keratan sulfate (KS) and chondroitin-6-sulfate (C6S), accumulate primarily in bone, cartilage, and other connective tissues, leading to growth impairment and progressive systemic skeletal dysplasia, which can cause physical disabilities and result in death due to respiratory failure and heart disease [2]. The outcome of disease onset begins with skeletal symptoms such as a humpback and prominent chest, leading to short stature with a short neck, hypoplasia of the odontoid process, tracheal obstruction, pectus carinatum (restrictive lung disease), kyphoscoliosis, protruding chest, flared rib cage, hip dysplasia, and genu valgum [1,2,3]. Enzyme replacement therapy (ERT) and hematopoietic stem cell transplantation (HSCT) are therapeutic options for MPS IVA. However, ERT has several challenges, including immune responses to the infused enzyme, the requirement for weekly 5–6-h infusions, and prohibitive costs. HSCT has several drawbacks, including difficulty finding suitable donors [4,5], graft rejection, and graft-versus-host disease (GVHD). Moreover, both treatments show limited effects on bone growth and skeletal abnormalities.

To overcome these obstacles, adeno-associated virus gene therapy (AAV-GT) has been proposed as a promising one-time treatment, as the enzyme is expected to be produced continuously at supraphysiological levels from transduced cells [6]. Compared with other viral-derived vectors, AAV vectors exhibit several advantages, including lower immunogenicity, the ability to transduce non-dividing cells, and a reduced likelihood of inducing insertional mutagenesis, as the transgene persists as an episome outside the host genome. Approximately 23 AAV-GT clinical trials in MPS are underway [6]. However, pre-existing anti-AAV immunity may hinder AAV transduction and therapeutic efficacy through neutralizing and non-neutralizing antibodies [7]. Moreover, patients can mount new immune responses to AAV vectors and their gene products, reducing transduction and eliminating transduced cells. Both humoral and cytotoxic immune responses may be elicited by AAV vectors and their gene products. Antibodies against the transgene and the AAV vector capsid can cause hypersensitivity reactions and adverse effects, including glomerulonephritis resulting from immune complex deposition in the kidneys [7]. Similar symptoms were reported with ERT due to antibody production against the infused enzymes [8,9]. Evidence of cellular immune responses includes elevated liver transaminase levels, associated with the loss of transgene expression observed in clinical trials of AAV-GT for hemophilia B, likely due to cytotoxic T-cell-mediated killing of liver cells [10]. Further observations were noted in a second clinical trial in hemophilia B, including increased alanine transaminase (ALT), reduced factor IX expression, and cellular immune responses [10]. Another study suggested that elevated liver transaminase levels in AAV-GT could be managed with immunosuppressive corticosteroids [11]. However, corticosteroids have been associated with side effects and an increased risk of infection [10]. As alternatives to corticosteroids, azathioprine and methotrexate are non-specific immunosuppressive drugs. However, treatment with these drugs can result in severe adverse effects on the functions of bone marrow, gastrointestinal tract, liver, and kidney, and an increased risk of infection [12]. Developing specific immune tolerance against gene products is therefore urgently required to avoid the adverse effects of these non-specific immunosuppressants.

Immune tolerance can be induced prior to treatment by administering an antigen under specific conditions. Oral tolerance refers to unresponsiveness or suppression of specific immune responses to antigens, such as food in the gastrointestinal tract, when antigens are administered in consecutive doses [13]. The first report of oral tolerance dates back to the early 20th century when Alexander Besredka found that guinea pigs that ingested milk became resistant to anaphylaxis following intracerebral injection of milk [14]. Feeding autoantigens, such as collagen and myelin basic protein (MBP), to experimental models of arthritis [15,16] and multiple sclerosis [17,18], respectively, can prevent the development of these diseases. Several therapeutic approaches for inducing oral tolerance have also been reported for the treatment of hemophilia B [19,20]. Using bio-encapsulated Factor (FIX) and intravenous replacement therapy, Xiamoei et al. induced complex IL-10-dependent, antigen-specific systemic immune suppression of pathogenic antibody formation in hemophilia B mice [19].

Oral immunotherapy (OIT) has emerged as a practical approach in various autoimmune and allergic diseases due to its long-lasting antigen-specific effects [21]. OIT has been demonstrated to prevent and treat various disease conditions, including rheumatoid arthritis, atherosclerosis, colitis, diabetes, allergic asthma, and antigen-specific food allergy [22]. In addition, OIT clinical trials have been conducted for a broad range of diseases such as autoimmune thyroid disease [23], multiple sclerosis [24,25], rheumatoid arthritis [26,27], uveitis [28,29], egg allergy [30,31], cow milk protein allergy [32,33,34], dust mite allergy [33,35] and peanut allergy [36,37].

The successes of these studies suggest that OIT is a promising approach for inducing specific immune tolerance to the human GALNS (hGALNS) gene product in AAV-GT for MPS IVA. In a previous study, Sosa et al. identified key hGALNS epitopes that bind to mouse MHC II molecules and induce oral tolerance in MPS IVA models and assessed their impact on the efficacy of enzyme replacement therapy (ERT). The hGALNS OIT approach resulted in undetectable hGALNS-specific immunoglobulin G (IgG) antibodies in the mice after ERT, thereby improving treatment efficacy [38]. The main objective of this study is to evaluate the effect of hGALNS OIT on AAV-GT in MPS IVA mouse models. Compared with ERT, AAV-GT may elicit different immune responses to hGALNS, as the antigen is continuously produced at supraphysiological levels in vivo rather than via intravenous administration of weekly 4–5-h injections with a short half-life of 40 min.

In this study, we aimed to enhance the therapeutic efficacy of AAV-GT by inducing oral immune tolerance to the gene product before AAV administration in MPS IVA mouse models. To this end, specific hGALNS proteins or hGALNS-epitope peptides were administered to neonatal mice to achieve efficient OIT [39]. We then evaluated the therapeutic efficacy of AAV-GT in bone and heart lesions after oral immunotolerance induction.

2. Results

2.1. Anti-hGALNS Antibodies in Plasma

In this study, we orally administered three immunodominant peptides of hGALNS (

Table 1. Peptide sequences chosen for the epitopes of hGALNS [38].

Peptide No.	hGALNS Region	Structural Domain	Sequence	Type of Epitope
C4	135–154	Loop	PNCHFGPYDNKARPNIPVYR	B/T
E8	200–219	β-sheet/loop	FFLYWAVDATHAPVYASKPF	T
I10	447–466	α-helix/loop	QQHQEALVPAQPQLNVTNWA	T

) or purified hGALNS protein (Elosulfase Alfa) to newborn mGALNS knockout (Galns^-/-) mice, henceforth referred to as MPS IVA mice, followed by intravenous injection of the AAV9 vector encoding human GALNS (CAG-AAV9-hGALNS) (Figure 1). Untreated MPS IVA and wild-type mice were used as controls.

To evaluate the immune response to the hGALNS protein produced after AAV9 GT, an ELISA was performed to assess hGALNS-specific antibody levels in plasma collected biweekly through 32 weeks post-AAV9 GT. We did not detect total IgG antibodies against hGALNS in the low- and high-dose [2.5 μg/g body weight (BW) and 5 μg/g BW] peptide groups throughout this period (Figure 2A). However, both low- and high-dose hGALNS protein groups showed elevated anti-hGALNS IgG antibodies starting 2 weeks post-AAV injection, peaking at 16 weeks, and then gradually decreasing until 32 weeks. A similar result was observed for the PBS + AAV9 group. As expected, no antibody was detected in the untreated wild-type and MPS IVA groups. These results suggested that immune tolerance was induced in groups administered orally with immunodominant hGALNS peptides. Antibody levels trended inversely with plasma enzyme activities (Figure 2A,B). The peptide groups, especially the 5 µg/g BW group, exhibited the highest enzyme activity, with no detectable anti-hGALNS IgG antibodies. In contrast, the hGALNS protein and PBS + AAV9 groups exhibited higher antibody levels and lower enzyme activity until 24 weeks. However, at 32 weeks, the hGALNS protein and PBS + AAV9 groups showed increased enzyme activity as antibody levels decreased over time.

Overall, these findings indicated that oral administration of hGALNS-epitope peptides, but not hGALNS protein, induced complete oral immunotolerance regarding antibody production.

2.2. T Cells Producing the Immunosuppressive Cytokine IL-10

Mouse spleens were collected at the 32-week experimental endpoint, when the overall impact of AAV-GT following oral tolerance induction was assessed. Immunosuppressive cytokine IL-10 is associated with oral immune tolerance in gut-associated lymphoid tissue (GALT) [40]. To determine the number of IL-10-secreting T cells, mouse spleens were collected at the experimental endpoint of 32 weeks, when the overall impacts of AAV-GT following oral tolerance induction were assessed. Splenocytes were isolated and used in an IL-10 ELISPOT assay after stimulation with hGALNS peptides. Our results showed that both the high- and low-dose peptide groups had the highest number of IL-10 spots, which were significantly higher than those in the PBS + AAV9 and hGALNS protein groups (Figure 2C). We also determined the number of T cells producing the pro-inflammatory cytokine IFN-γ. We observed that the number of IFN-γ spots in the PBS group was significantly higher than in the untreated group (Figure 2D). These data indicated that the cell-mediated immune responses in the peptide-treated mice correlated with the humoral immune response.

2.3. hGALNS Enzyme Activity in MPS IVA Mice

hGALNS enzyme activities were measured in plasma and tissues collected after autopsy at 32 weeks. Plasma enzyme activity increased at 2 weeks post-AAV-GT in all treated groups. The high-dose peptide group (5 µg/g BW) (Figure 2B) consistently showed the highest enzyme activity over the weeks until autopsy, a more than 4-fold increase compared to the level in untreated wild-type control mice (22.68 nmol/h/mL). The low-dose peptide group (2.5 µg/g BW) (Figure 2B) exhibited stable enzyme activity until 28 weeks, increasing 4-fold at 32 weeks at autopsy. The hGALNS protein groups (2.5 µg/g and 5 µg/g BW) also exhibited supraphysiological enzyme levels. However, these levels were significantly lower than those in the high-dose peptide group. The low-dose hGALNS protein group showed a 5-fold increase in enzyme activity compared to the wild-type group at 32 weeks. The PBS + AAV9 group exhibited the lowest enzymatic activity among the treated groups. However, an increase was observed after 30 weeks, bringing it to the level of the high-dose hGALNS protein group.

To determine the cross-correction of hGALNS deficiency, we evaluated hGALNS activities in various tissues, including the liver, heart, lung, bone, muscle, kidney, and trachea, at 32 weeks (Figure 3).

The liver enzyme level in all treated groups, including the PBS + AAV9 group, was significantly higher than in the wild-type group (Figure 3A). The liver enzyme activity was 15–20-fold higher in both peptide groups than in wild-type mice. In both the hGALNS protein and PBS + AAV9-treated groups, liver enzyme activities increased by 5–8-fold (Figure 3A). However, there was a significant difference between the PBS + AAV9 group and the peptide- or hGALNS protein-treated groups.

In the heart, both low- and high-dose peptide groups displayed significantly higher enzyme activities than PBS + AAV9, hGALNS (at both low and high doses), and wild-type controls (Figure 3B). In both peptide-treated groups, enzyme levels were 3-fold higher than in the wild-type group, which is a statistically significant difference. In contrast, there were no significant differences between the PBS + AAV9 and wild-type groups, or the hGALNS protein groups.

In bones, there were no significant differences in the enzyme activity between the PBS + AAV9 and wild-type groups. However, the low- and high-dose peptide groups showed 20- and 10-fold increases in enzyme levels, respectively, compared with the hGALNS protein group (Figure 3C).

There were no significant differences in lung enzyme activity between the treatment groups (Figure 3D). However, there was a 2.5-fold increase in enzyme activity in the PBS + AAV9 group compared with the wild-type group. The PBS + AAV9 group showed 4- to 5-fold higher muscle enzyme activity than the wild-type group, a significant difference (Figure 3E). The low- and high-dose hGALNS protein groups had significantly greater enzymatic activity in muscle than the PBS + AAV9 group (Figure 3E). A significant difference in muscle enzyme activity was also found between the hGALNS protein and peptide groups. Compared to the wild type, none of the treated groups displayed significant enzyme activity in the trachea (Figure 3F).

2.4. KS Levels in the Blood and Tissues

We measured the primary storage material, mono-sulfated KS, in the plasma at 16 and 32 weeks of age, as well as in the liver and bone of MPS IVA mice at 32 weeks (the time of autopsy) (Figure 4). Plasma and liver mono-sulfated KS levels were significantly higher in the untreated MPS IVA mice than in all the treated groups (Figure 4A–C). All treated groups had normal plasma KS levels at 16 and 32 weeks (Figure 4B). No significant difference was observed among the groups in bone KS levels (Figure 4D).

2.5. Micro-Computed Tomography Analysis of the Femur

Bone trabecular and cortical structures in the femur were assessed using micro-computed tomography (micro-CT) to evaluate morphometric differences between treated and untreated MPS IVA groups and the wild-type group (Figure 5). Untreated MPS IVA mice, compared to wild-type mice, showed an increase in bone volume (BV), percent bone volume (BV/TV), bone mineral density (BMD), and trabecular number (Tb.N), while untreated mice showed a decrease in trabecular separation (Tb.Sp), highlighting the disease impact on bone architecture. Among the treated groups, only the low-dose peptide group normalized these parameters among treated groups (Figure 5). Both the low- and high-dose hGALNS protein groups, as well as the high-dose peptide group, exhibited elevated BV, BV/TV, BMD, and Tb.N and a reduced Tb.Sp, which significantly differed from those of the wild-type group in terms of bone morphology. The PBS + AAV9 group normalized BV, BV/TV, and BMD, whereas this group did not normalize Tb.N or Tb.Sp, which significantly differed from the wild-type group. None of the treated groups showed significant differences relative to the untreated group in other bone parameters, including total volume of interest (VOI), trabecular thickness, degree of anisotropy, cortical thickness, total area, bone area, and medullary area (Supplementary Figure S1). Overall, the low-dose peptide group showed the greatest effect on bone morphology, whereas the other treatment groups did not normalize all parameters.

2.6. Pathology

To assess improvement in pathological features, heart and knee-joint lesions were sectioned with 0.5 µm and stained with toluidine blue after the autopsy at 32 weeks of age. The pathological improvement was then assessed [1].

2.6.1. Bone Pathology

Pathological scoring was performed using light microscopy to assess vacuole formation in the growth plate, articular cartilage, meniscus, and ligaments. The vacuolization of chondrocytes in the growth plate and articular cartilage areas of the femur and tibia, along with ligament and meniscus, was evaluated in treated groups, compared with untreated MPS IVA and wild-type groups. The PBS + AAV9 group did not show a significant difference from the untreated group in vacuolization in the growth plate and articular cartilage of the tibia and femur, nor for ligament and meniscus (Figure 6B). Similarly, no significant difference in bone vacuolization was observed between the low- and high-dose hGALNS protein groups and the untreated group. Although both low- and high-dose hGALNS protein groups showed reduced pathological scores for the ligament and meniscus, these differences were not statistically significant compared to the PBS + AAV9 or untreated groups. In contrast, a significant reduction in vacuole number was observed in the femoral articular cartilage of both peptide-treated groups relative to the untreated group (Figure 6B). The low-dose peptide group showed a significantly lower vacuolization score in the ligament than the untreated group (Figure 6B).

Untreated MPS IVA mice displayed a disorganized columnar structure in the growth plate and articular cartilage of the tibia and femur (Figure 6A–E). Among the treated groups, the PBS + AAV9 group exhibited the highest (worst) pathological scores for column structure in the tibial growth plate and articular cartilage (Figure 6C). Notably, the low-dose peptide group showed significant improvement in the column structure of the femoral articular cartilage compared to the untreated group (Figure 6C), and partial improvement was also observed across other treated groups (Figure 6A,C). Combining oral hGALNS peptide administration with AAV9-GT resulted in a more significant improvement in bone pathology than AAV9-GT alone. To assess whether hGALNS expression reduces intracellular storage material, chondrocyte size was measured in the growth plates of the tibia and femur (Figure 6D,E). In the tibial growth plate, the chondrocyte size in the hGALNS protein and PBS + AAV9 groups remained unchanged relative to the untreated group (Figure 6D). In contrast, both peptide-treated groups exhibited chondrocyte sizes normalized to wild-type levels. In the femoral growth plate, the PBS + AAV9 group showed a significant difference in size compared to the wild-type and high-dose peptide groups (Figure 6E). No significant differences were observed between the hGALNS protein and PBS + AAV9 groups. The high-dose peptide group exhibited chondrocyte sizes indistinguishable from those of the wild-type group, although they differed significantly from those of the hGALNS protein and PBS + AAV9 groups. The low-dose peptide group also showed a significant reduction in chondrocyte size compared to the untreated group (Figure 6E).

2.6.2. Heart Pathology

The hGALNS protein and PBS + AAV9 groups showed complete correction of vacuolization in heart tissue, except for one mouse in each group that still exhibited vacuoles in the base, valve, and muscle compared to the untreated group. In comparison, both peptide-treated groups demonstrated full correction of heart pathology across all regions assessed, including the base, valve, and muscle (Figure 7A,B). No significant differences were observed among the treated and wild-type groups.

2.7. Adverse Effect

2.7.1. Liver Toxicity

To evaluate the liver toxicity of AAV9 GT, two biomarkers, AST and ALT liver enzymes, were measured in plasma samples from 32-week-old mice at autopsy. No significant difference was observed in AST levels among the treated, wild-type, and untreated groups, which ranged within the normal range (46–221 μ/mL) (Figure 8A). For ALT, the PBS + AAV9 group showed the highest level compared to both peptide groups as well as, the low-dose hGALNS protein group, and the untreated MPS IVA group. However, all groups were within the normal range (22–133 μ/mL) (Figure 8B).

2.7.2. Body Weight (BM)

BW was measured biweekly after AAV9 injection, starting at 4 weeks and continuing for 32 weeks (Supplementary Figure S2). There was no significant difference between the wild-type and untreated MPS IVA groups. BW increased from 8 to 32 weeks in both high- and low-dose peptide groups. However, both high- and low-dose hGALNS protein groups exhibited a reduction in BW after 8 weeks, which persisted until 32 weeks and significantly differed from both the peptide and wild-type groups (Supplementary Figure S2.2). Compared to the untreated group, the PBS + AAV9 group did not exhibit a reduction in BW. It is possible that administering the whole enzyme to neonatal mice triggered an alternative mechanism, such as immune activation or metabolic stress, which may have contributed to the observed reduction in BW.

2.8. Biodistribution of the AAV Genome

Liver samples were used to measure the vector copy number per diploid cell using digital PCR (dPCR) [1]. The vector copy number was not detected in the wild-type control and untreated MPS IVA groups (Supplementary Figure S2.1). All treated groups showed high vector copy numbers. The highest vector copy number was detected in the PBS + AAV9 control group, followed by the high-dose hGALNS protein group. The high-dose peptide group showed the lowest vector copy number among all the treated groups. Both low- and high-dose peptide groups showed significantly lower vector copy numbers than the PBS + AAV9 control group. There was a significant difference between the low-dose hGALNS protein and PBS + AAV9 control groups, followed by AAV-GT. The vector copy number of the high-dose hGALNS protein group was significantly higher than that of the high-dose peptide group. No significant difference was observed between the high-dose hGALNS protein and the PBS + AAV9 groups. Therefore, there was no correlation between the vector copy number and enzyme activity (Supplementary Figure S3).

3. Discussion

In this study, we demonstrated that oral administration of immunodominant peptides derived from the hGALNS enzyme induces immune tolerance to hGALNS produced via AAV-GT, thereby improving therapeutic efficacy. Our results showed that the highest circulating enzyme levels were detected in mice treated with peptides at 2.5 µg/g and 5 µg/g. These enzyme levels inversely correlated with anti-hGALNS antibody levels in plasma, as neither group showed detectable antibodies. In contrast, mice treated orally with hGALNS proteins or with PBS + AAV9 developed significant plasma IgG antibodies beginning at 2 weeks post-treatment, with peak levels at 16 weeks. Enzyme levels in these groups were significantly lower than in peptide-treated groups. Similar results were observed with hGALNS infusion for ERT-induced immune tolerance [38]. Although the anti-hGALNS antibody decreased after 18 weeks post-AAV-GT, the amount of hGALNS produced by the transgene was lower than in the peptide groups, indicating that circulating antibodies neutralized hGALNS and were sufficient to clear hGALNS secreted from the transduced cells. By 18 weeks, antibody levels began to decline in the hGALNS-treated groups due to suppression of B and helper T cells by naturally induced regulatory T cells. We expected to observe higher levels of the anti-inflammatory cytokine IL-10 in the peptide-treated groups compared to the non-immunized groups. Because the hGALNS protein-treated and PBS + AAV9 groups were not immunized, their IL-10 levels remained significantly lower than those of the immunized peptide groups, even after antibody levels declined at 32 weeks. By this point, levels of the pro-inflammatory cytokines had normalized, and we anticipated a baseline IFN-γ expression across all treated and untreated groups. This immunomodulatory effect increased the systemic availability of the hGALNS enzyme. We evaluated enzyme activity levels across tissues in treated and untreated groups. Liver, heart, and bone tissue exhibited significantly higher enzyme activity in the peptide groups than in the hGALNS protein and PBS + AAV9 groups. Other tissues, including the lungs, muscles, and trachea, exhibited significantly higher enzyme activity in the hGALNS protein groups. This may reflect direct correction via AAV-GT in these tissues, given the ubiquitous promoter used. These results are consistent with those of Sawamoto et al., who used the ubiquitous CAG promoter with the AAV8 hGALNS vector to treat MPS IVA mice [1]. The supraphysiological enzyme activity post-gene therapy will reduce patients’ need for infusions considerably and could also be of real clinical interest. A small fraction of a normal enzyme, such as 5–10% of wild-type levels, is required for real substrate clearance in lysosomal storage diseases, including MPS IVA. In addition, AAV-mediated gene transfer is often associated with sustained transgene expression, as strong promoters are used. Increased circulating enzyme concentration in systemic diseases will enhance enzyme delivery to peripheral tissues via the mannose-6-phosphate receptor pathway. This may lead to better correction of skeletal and cardiac abnormalities, which are typically more difficult to treat. Furthermore, the increases in liver enzymatic activity are probably linked to AAV tropism and vector accumulation in the liver. Systemically delivered AAV vectors, such as AAV9, are highly hepatotropic. A high level of transgene expression in hepatocytes means supraphysiological release of the enzyme into the blood circulation. In this way, the enzyme can be cross-corrected to the systemic circulation via mannose-6-phosphate receptor-mediated uptake. Moreover, hepatic antigen presentation may underlie systemic tolerance or partial immune modulation. Once an enzyme is made intracellularly within hepatocytes, circulating antibodies cannot neutralize the intracellular pool, which explains why liver activity remains high even in the setting of systemic humoral responses. Thus, increased hepatic enzyme activity is anticipated and is mechanistically compatible with AAV biodistribution. Furthermore, both peptide-treated groups exhibited lower liver vector copy number levels than the GALNS- or PBS-treated groups. A possible explanation for this could be that pre-exposure of peptides affected systemic immune activation, which could modify the hepatic immune microenvironment and impact the AAV vector’s persistence. The liver contains substantial numbers of resident macrophages (Kupffer cells), antigen-presenting cells, and tolerogenic immune communities that can mediate vector uptake, processing, and clearance. Such programmed-environment, peptide-mediated immune changes can influence innate sensing pathways, antigen presentation, or cytotoxic T cell priming, thereby impacting liver transduction efficiency or vector genome stability. While lung and skeletal muscle VCNs were not quantified, higher GALNS enzyme activity in these tissues indicates relatively greater vector persistence and/or transgene expression. Tissue differences in immune surveillance intensity, antigen display capacity, and local inflammatory tone may thus contribute to differential AAV biodistribution and functional enzyme output between organs. More studies investigating innate immune activation, capsid-specific T cell responses, or tissue-resident immune populations will be needed to disentangle the mechanisms underlying these findings.

Additionally, we analyzed pathological and biochemical improvements associated with the disease. Liver, bone, and plasma KS levels were reduced in all AAV9-GT groups, irrespective of immunization status. We did not find a significant difference in KS levels between immunized and non-immunized mice. These results were confirmed by scoring the degree of vacuolization in chondrocytes, which showed a reduction in vacuoles in pathology (Figure 6A–C). Compared with the untreated group, complete vacuole clearance was observed in the heart in both the tolerized and non-tolerized treated groups. However, a greater decrease in vacuoles was observed in the tolerized groups in the femoral articular cartilage. In addition, the femoral articular cartilage in both peptide groups showed significant improvement in bone column structure.

In a comparative study of bone lesions across MPS types I, IIIA, IVA, and VII using micro-CT imaging, MPS VII exhibited the most pronounced skeletal abnormalities. Both plain radiographs and micro-CT measurements of bone mineral density (BMD) confirmed increased bone mass in MPS VII mice compared with wild-type controls. In contrast, MPS IVA mice showed no detectable trabecular or cortical bone abnormalities on micro-CT scans compared with wild-type controls [41]. Similarly, our MPS IVA mouse model did not exhibit typical skeletal dysplasia or abnormal bone growth. However, previous studies have reported trends suggesting subtle changes in MPS IVA mice relative to wild-type mice, including increased total bone volume, greater bone volume, a higher percentage of bone volume, reduced trabecular separation, and increased bone area [42]. Our current micro-CT analysis confirms similar trends in trabecular and cortical bone parameters. Interestingly, the low-dose peptide group showed significantly improved Tb.Sp compared with the untreated groups, whereas improvements were not observed with AAV-GT alone. Additionally, the low-dose peptide groups showed normalized BV, BV/TV, Tb.N, and BMD, comparable to those of the wild-type group. These findings suggest that suppressing antibody responses against the hGALNS enzyme may enhance its availability across tissues, including bone. However, in both the low- and high-dose hGALNS protein groups, as well as the high-dose peptide group, these parameters were not normalized, although the underlying mechanism remains unclear. We hypothesize that bone proliferation may have been induced in these groups, suggesting a need for future research to elucidate the interactions among therapeutic effects, the immune response, and the bone microenvironment. In future studies, we will increase the sample size to confirm the impact of bone morphology. The MPS IVA rat model, which exhibits severe skeletal phenotypes, will also provide a more comprehensive understanding of bone structural improvement [43]. Oral tolerance is a normal immune response in the intestines. It refers to the non-responsiveness or suppression of specific immune responses to antigens, such as those found in the gastrointestinal tract, when antigens are administered in consecutive doses [21]. Various antigens are regularly encountered by the intestinal immune system, including dietary proteins and components of commensal bacteria. Notably, the gut immune system has evolved tightly regulated mechanisms to suppress unwanted inflammatory responses while protecting against pathogenic organisms [44]. Intestinal dendritic cells (DCs), present in both the lamina propria and mesenteric lymph nodes (MLNs), play a crucial role in the development of oral tolerance by taking up and presenting orally derived antigens in a tolerogenic manner to T cells [22]. The induction of partial immune tolerance to eggs and milk in children has been reported in clinical trials [45,46]. In an egg-allergy clinical trial, one-third of children-maintained tolerance after 3 months of food withdrawal. Similarly, clinical trials for oral immune tolerance induction in 5- to 17-year-old children with severe milk allergies demonstrated that 36% of the treated subjects could consume cow’s milk and dairy products after 1 year, and 50% of the subjects achieved partial tolerance compared with the control group, which had allergic symptoms [46]. Oral tolerance has also been extensively studied to prevent autoimmune diseases. In an oral immunotherapy clinical trial, the daily administration of 7.5 mg insulin capsules improved metabolic parameters in children and adolescents with islet cell autoantibodies and enhanced oral glucose tolerance test responses [47].

Based on these achievements, we investigated the induction of immune tolerance against hGALNS produced in AAV-GT for MPS IVA. Here, we employed a similar methodology to induce oral tolerance to the hGALNS enzyme, the gene product, in AAV-GT-treated MPS IVA mice for the first time. We used two different doses, low and high, for both peptides and hGALNS protein, as the dosage is a crucial factor in inducing oral immune tolerance in rheumatoid arthritis [26,27]. Previous studies on rheumatoid arthritis treatment have also shown that different antigen doses are used. The effects of four different doses of orally administered bovine collagen II in patients with rheumatoid arthritis were evaluated by Barnett et al. A significant improvement in clinical parameters was observed in the group receiving the lowest dose (20 g/day) after 24 weeks of treatment [26]. It has been proposed that low-dose antigens induce oral tolerance through a regulatory T cell-mediated mechanism, suppressing both helper and cytotoxic T cells. In contrast, high-dose antigens induce apoptosis or anergy of specific T cells [21]. In rheumatoid arthritis, patients who received 25 mg of the DNAJP1 peptide for 6 months showed decreased TNF-α-producing T cells and increased IL-10-producing Treg cells, leading to subsequent improvement in disease severity [27]. Patients with cross-reactive immunological material (CRIM)-negative disease developed higher immune responses than those with CRIM-positive disease [12]. The immune response is crucial in determining treatment outcomes and patient safety across all therapeutic approaches. Previous studies have shown a significant increase in hGALNS-specific IgG antibodies 12 weeks after AAV8- and AAV9-GT in MPS IVA mice [48]. Both humoral and cytotoxic immune responses to gene products and AAV vectors are determinant factors for effective treatment [49]. CD4+ and CD8+ T cells recognize specific peptides derived from the hGALNS enzyme on the cell surface, presented by MHC class II and MHC class I molecules, respectively. Then, CD4+ T cells activate B cells to produce antibodies specific to the AAV vector and its encoded gene product, whereas CD8+ T cells directly target the transduced cells. We were unable to induce humoral immune tolerance in groups treated with the hGALNS protein. Several factors may have contributed to the hGALNS enzymes inability to induce oral tolerance. Firstly, although the same dosage of peptides and the whole hGALNS protein in µg was administered, the molecular weight of tolerance-inducing epitope peptides is much smaller than that of the hGALNS protein; therefore, the number of tolerance-inducing peptide molecules is much higher in the peptides administered than that in the hGALNS protein on a per µg basis. hGALNS is a large protein (> 500 amino acids), so when given as a full-length protein, only a small fraction of the whole mass corresponds to any individual immunodominant CD4⁺ T cell epitope. On the other hand, when short synthetic peptides are administered at the same overall mass, approximately 100% of the administered material corresponds to the immunogenic epitope sequence. This allowed the total administered antigen mass to remain constant, while the therapeutically effective epitope dose varied dramatically between groups. The peptide groups presumably received a substantially greater molar concentration of the relevant T cell epitopes than did the full-length GALNS group. Antigen dose is a vital determinant of immune regulatory mechanisms within the context of oral tolerance. These therapies are associated with high doses that induce clonal deletion or anergy, whereas lower doses might favor the induction of regulatory T cells or fail to promote tolerance when TCR engagement is inadequate. Thus, the considerable divergence in immune responses between peptide and full-length protein groups may only partially reflect differences in effective epitope density, rather than intrinsic differences in tolerogenic capacity. Furthermore, intact protein has to be processed prior to presenting on MHC class II molecules, which has differential properties in epitope generation, processing efficiency, and competition among epitopes. Short peptides, in contrast, can be loaded on MHC class II molecules directly without the need for intracellular processing would facilitate more efficient and targeted T cell-mediated engagement. Collectively, differences in epitope concentration, requirements for antigen processing, and effective TCR-stimulation strength could have contributed to the dissimilar immune effects on the efficacy of peptide-based versus full-length GALNS oral therapy.

Moreover, orally delivered hGALNS may be randomly digested by gastric enzymes, generating peptides that cannot be bound and presented by MHC class II to T cells. To overcome protein digestion, a recent publication on hemophilia A reported the use of a therapeutic antigen fused to the transmucosal carrier, the cholera toxin subunit B, for oral administration [50]. They reported minimal anti-FVIII IgG1 antibodies after oral immune tolerance induction followed by AAV-GT. It was also noted that, using H6P-loaded targeting nanoparticles, H6P was directly delivered to gut DCs of non-obese diabetic (NOD) mice via oral vaccination [51]. Compared with the full hGALNS protein, the peptides we used are short and lack aromatic hydrophobic residues; thus, they may be resistant to pepsin and have a greater chance of binding to the MHC class II molecule for presentation to T cells. In conclusion, future research may involve testing larger animal models to evaluate the translatability and efficacy of these findings. This approach could further elucidate therapeutic potential and refine our strategies for clinical application.

A practical approach to immune tolerance is urgently needed to suppress the elevation of anti-AAV and anti-transgene antibodies. In oral tolerance, immune responses to specific antigens are inhibited by the simple oral administration of the antigen. Thus, compared to drug-based immunosuppression, it is non-toxic, antigen-specific, and potentially long-lasting [52,53]. To prevent or reverse adverse drug reactions against therapeutic proteins, oral tolerance can be achieved through a non-invasive, antigen-specific approach that utilizes natural immune regulatory pathways. Despite the outstanding advantages of oral immune tolerance, it has some limitations. More detailed studies in larger animal models should be conducted to determine whether this treatment is effective in humans. Antigen dose selection, administration schedule, and patient responsiveness to treatment remain critical concerns.

While the underlying mechanisms of oral tolerance remain unclear, knowledge of the mechanisms and the cell types involved in oral tolerance has increased over the past decade. Successful translation to humans requires identifying appropriate patient populations, validating biomarkers to assess immunologic effects, and considering combination therapies with adjuvants to increase T-reg induction. Moreover, human MHC molecules are individual-specific, so designing distinct epitope peptides for OIT may be required for each patient. To overcome this problem, an alternative research strategy will be necessary to develop a peptide drug suitable for all patients. A deeper understanding of the mechanisms underlying the induction and loss of oral tolerance will facilitate the development of new treatments for genetic diseases.

Conclusion: Our innovative approach is designed to enhance the therapeutic efficacy of GT by targeting the immune response against the AAV-GT product. This peptide-based oral immunotherapy has the potential to enhance the effectiveness of GT, mitigate the adverse effects of immunosuppression, and improve treatment outcomes.

4. Materials and Methods

4.1. MPS IVA Mouse Model

The MPS IVA knockout mouse (MKC, Galns^-/- KO) on the C57BL/6 background was used in this study (Jackson Laboratory, Bar Harbor, ME, USA) [54]. The Galns^-/- KO mouse was developed by targeted deletion of exon 2 of hGALNS [54,55], resulting in no enzyme activity in blood and tissues. Moreover, storage materials (GAGs) were observed in vacuoles of reticuloendothelial Kupffer cells, heart valves, muscles, chondrocytes, growth plates, and articular cartilage [1]. Wild-type mice were used as a control, as they produce physiological levels of the enzyme required for GAG metabolism. All animal care and experiments were reviewed and approved by the Institutional Animal Care and Use Committee of the Nemours Children’s Health (Protocol Number: RSP20-12483-003).

4.2. Oral Administration with hGALNS Protein and Epitope Peptides of hGALNS

Immunodominant epitopes are distinct components of antigenic regions that are unequivocally recognized by the immune system. A previous study identified highly immunogenic regions of the hGALNS enzyme in C57BL/6 (H2-IAb) using algorithms to predict B- and T-cell epitopes [38]. The study aimed to create a detailed profile of the most immunogenic regions in the GLANS enzyme to predict B- and T-cell epitopes. Ten 20-mer peptides with immunogenic regions in the hGALNS enzyme were evaluated, and the C4, E8, and I10 peptides were found to be the most immunogenic (Table 1). Male mice were orally administered with hGALNS enzyme or a mixture of the three peptides within 72 h of birth every alternate day for 20 days at the doses of 2.5 µg/g BW or 5 µg/g BW. The mixture of peptides contains equal amounts of C4, I10, and E8 peptides (Table 1). The experimental design is shown in Figure 1B, with seven experimental groups (n = 3–6): (1) MPS IVA mice orally administered with hGALNS protein (Elosulfase alfa) 2.5 µg/g BW; (2) MPS IVA mice orally administered with hGALNS protein (Elosulfase alfa) 5 µg/g BW; (3) MPS IVA mice orally administered with peptides 2.5 µg/g BW; (4) MPS IVA mice orally administered with peptides 5 µg/g BW; (5) MPS IVA mice orally administered with PBS; (6) untreated wild-type mice (C57BL/6); and (7) untreated MPS IVA mice.

4.3. AAV Gene Therapy for Mouse Models

As shown in Figure 1A, the AAV expression cassette was created with a codon-optimized hGALNS coding sequence downstream of a ubiquitous Cytomegalovirus Enhancer/Chicken β-Actin (CAG) promoter that consists of a modified chicken β-actin promoter linked to a cytomegalovirus early enhancer element. The hGALNS sequence is followed by the Woodchuck hepatitis virus Posttranscriptional Regulatory Element (WRPE) and a bovine growth hormone polyadenylation signal. The viral genome was packaged into AAV9 vectors by Vector Builder and named pAAV [Exp]-CAG> {hGALNSco}: WPRE (Figure 1A). Moreover, 5 × 10¹³ Genome copies (GC)/kg AAV9 vectors were given via a tail vein to the mice at 4 weeks of age (Figure 1B) in the previously mentioned groups (1), (2), (3), (4), and (5).

4.4. Anti-hGALNS Antibodies by ELISA

As reported previously, ELISA (Enzyme-Linked Immunosorbent Assay) was used to quantify anti-hGALNS antibodies in plasma samples [56]. A 96-well microtiter plate was coated with 2 μg/mL of the purified hGALNS enzyme (Elosulfase alfa) in 15 mM Na₂CO₃, 35 mM NaHCO₃, and 0.02% NaN₃ (pH 9.6) overnight. The plate was blocked with 3% bovine serum albumin in PBS (pH 7.2) for 1 h at room temperature after being washed three times with Tris-buffered saline with Tween 20 (TBS-T) (10 mM Tris [pH 7.5], 150 mM NaCl, 0.05% Tween 20). After two additional TBS-T washes, 100-fold-diluted mouse plasma was added to each well, and the wells were incubated for an additional 2.5 h at 37 °C. Peroxidase-conjugated goat anti-mouse immunoglobulin G (IgG) (Thermo Fisher Scientific) was added to the plate after it had been washed four times with TBS-T at a 1:1000 dilution. The plate was then incubated for an additional hour at room temperature. Next, 100 μL of peroxidase substrate (ABTS solution, Invitrogen, Carlsbad, CA, USA) was added to each well after three TBS-T washes, followed by two TBS washes (10 mM Tris, pH 7.5, 150 mM NaCl). The plate was incubated at room temperature for 30 min. After stopping the reaction with 1% SDS, the absorbance at 410 nm was measured in optical density units using a PerkinElmer Victor X4 plate reader (PerkinElmer, Waltham, MA, USA). The standard curve was created using a repeated dilution of known monoclonal anti-mouse IgG concentrations added to the hGALNS-coated plate. The optical density was converted to μg IgG per milliliter of plasma using linear regression.

4.5. Detection of Pro- and Anti-Inflammatory Cytokines Using ELISPOT

The ELISPOT assay was used to determine the cytokines (IL-10 and IFN-γ) secreted by activated immune cells, according to the manufacturer’s protocol (ImmunoSpot, Cleveland, OH). The ELISPOT assay was performed in 96-well plates with nitrocellulose membranes coated with specific primary antibodies for IL-10 and IFN-γ. Single-cell suspensions of splenocytes (1 × 10⁵ cells/well), combined with an appropriate stimulus of 10 μg/mL of cells (peptide mixture Table 1), were added to the wells of the ELISPOT plate after blocking for 24 to 48 h. After removing the cells (1 × 10⁵ cells/well) and washing, Biotinylated secondary detection antibodies, enzyme-labeled streptavidin conjugates, and chromogenic substrate solutions were added stepwise to form colored spots. The number of spots formed was enumerated automatically using the IIMMUNOSPOT reader.

4.6. hGALNS Enzyme Activity Assay

hGALNS enzyme activities in plasma and various tissues were determined as described previously [56]. A Bead Mill Homogenizer (OMNI International, Kennesaw, GA, USA) was used to homogenize frozen tissues in a homogenization solution containing 25 mmol/L Tris-HCl (pH 7.2) and 1 mmol/L phenylmethylsulphonyl fluoride. After transferring the homogenates to microcentrifuge tubes, they were centrifuged at 4 °C for 30 min at 12,000 rpm. The supernatants were then collected. Next, 9 μL of the hGALNS substrate, 22 mM 4-methylumbelliferyl-β-galactopyranoside-6-sulfate (Research Products International, Mount Prospect, IL, USA) in 0.1 M NaCl/0.1 M sodium acetate (pH 4.3), was catalyzed by 2 μL of tissue lysate or plasma at 37 °C for 16 h. After this first incubation, 2 µ of 10 mg/mL Aspergillus oryzae β-galactosidase (Sigma-Aldrich, St. Louis, MO, USA) in 0.1 M NaCl/0.1 M sodium acetate (pH 4.3) was added to break down the enzymatic reaction product 4-methylumbelliferyl-β-D-galactosidase (MUG). The mixture was then incubated for an additional two hours at 37 °C to release 4-methylumbelliferone (4-MU) as a fluorescent product. After stopping the reaction with a 1 M glycine-NaOH (pH 10.5) solution, a PerkinElmer Victor X4 plate reader (PerkinElmer, Waltham, MA, USA) was used to read the fluorescence signal of 4-MU at 366 nm excitation and 450 nm emission. Nanomoles of 4-methylumbelliferone produced per hr. per microliter of plasma or milligram of tissue protein were used to describe the activities. The total protein concentration was measured using a Bicinchoninic Acid (BCA) protein assay kit (Thermo Fisher Scientific, Waltham, MA, USA).

4.7. Glycosaminoglycan Quantification in Plasma and Tissues

Extraction of GAG from different tissues (liver and humerus) was done using the previously described protocol [57]. Acetone was utilized as a defatting solvent to homogenize frozen tissue samples., the homogenate was centrifuged at 4 °C for 30 min, and the pellet was then dried by vacuum centrifugation. It was then suspended in 0.5 M NaOH and incubated at 50 °C for 2 h To extract GAG chains from core proteins. The solution was neutralized with 1 M HCl, then added with 3 M NaCl. The supernatant was collected after centrifugation of the nucleotides and pH adjustment to below 1.0 with an additional 1 M HCl. After further centrifugation, the precipitated proteins were recovered from the supernatant, and 1 M NaOH was added to neutralize the supernatant. After adding two volumes 1.3% potassium acetate and centrifuging for 30 min at 4 °C, the crude GAGS was precipitated from the solution. Following centrifugation, the GAGs were dried and washed with 80% ethanol before being reconstituted in 50 mM Tris-HCl, pH 7.0, and stored at −20 °C. As previously described, liquid chromatography–tandem mass spectrometry (LC-MS/MS) was employed to quantify KS in plasma and tissues [57].

4.8. Micro-CT

Micro-CT analysis was performed on the femur to characterize trabecular and cortical structure and morphometry, as previously described [56]. Briefly, femurs were collected at autopsy from 32-week-old MPS IVA and wild-type mice and fixed in 100% ethanol. For scanning the samples, a SkyScan 1276 micro-CT system (Bruker, Billerica, MA, USA) was used after wrapping them in gauze soaked in 0.9% NaCl.

4.9. Pathological Assessment by Toluidine Blue Staining

Toluidine blue was used to stain heart and knee joint samples from mice, as described previously [1]. Heart and knee joints were collected from 32-week-old treated and untreated MPS IVA and wild-type mice at autopsy in 10% formalin. Light microscopy was used to evaluate lysosomal storage in vacuoles in chondrocytes and cardiovascular tissues after fixing with 2% paraformaldehyde, 4% glutaraldehyde, and toluidine blue in 0.5-µm-thick sections. Pathological slides of knee joints and hearts were scored for vacuolization and column structure in a double-blinded manner to minimize bias. A score from 0 (no visible vacuolization) to 3 (obvious vacuolization) was given to each slide for column structure and vacuolization in the growth plate and articular cartilage [1]. Similar vacuolization scores were assigned to the heart muscle, base, and valve. Statistical analysis was performed on all heart and knee-joint scores to determine the reduction in storage materials.

4.10. Toxicity Analysis

We measured alanine transaminase (EALT-100, BioAssay Systems, Hayward, CA, USA) and aspartate transaminase (EASTR-100, BioAssay Systems) in plasma from all treated and untreated mice at 16 weeks of age, following the manufacturer’s instructions, to assess AAV-GT toxicity. All analyses were conducted in 96-well plates and read at the wavelength specified by the supplier using a PerkinElmer Victor X4 plate reader (PerkinElmer, Waltham, MA, USA).

4.11. Plasma and Tissue Collection

From 4 to 32 weeks, blood was drawn biweekly from all mouse groups using a 22G needle to puncture the superficial temporal vein into EDTA-containing tubes (BD Microtainer). After centrifugation for 10 min at 10,000 rpm, the plasma was extracted and stored at −20 °C. Following RBC lysis with 0.2% NaCl (w/v), WBCs were extracted from the red blood cells. At 32 weeks of age, all mice were euthanized in a CO₂ chamber and perfused with 20 mL of 0.9% normal saline. For further tests to measure KS and hGALNS activity, the liver, kidney, lung, spleen, heart, trachea, femur, tibia, and knee joint were collected on dry ice and stored at −80 °C. In addition, other tissues, including bone and heart, were collected and fixed in 10% neutral-buffered formalin for histopathological analysis. The tibia was stored in 100% ethanol for the micro-CT assay.

4.12. AAV Vector Genome Biodistribution

An AAV vector containing expression cassettes for the human hGALNS transgene, driven by the CAG promoter, was used in this study (Supplementary Figure S1). The number of vector copies in the liver was quantified using digital PCR, as previously described [1]. Briefly, liver samples were used to extract DNA using the Gentra Puregene kit according to the manufacturer’s instructions (QIAGEN, Germantown, MD). Genomic DNA from frozen liver samples of mice was analyzed by digital PCR (dPCR) using specific primers and probes: 5′-AGAAGCCCTGAGCAGAATCA-3′, 5′-CAGTTCATCACGGCCCAATTAC-3′, and 5′-AGCAGCACCAAGAGGCTCTGGTTC-3′.

4.13. Statistical Analysis

All statistical analyses were conducted using GraphPad Prism 9.50 (GraphPad, San Diego, CA, USA). We conducted Shapiro–Wilk and Kolmogorov–Smirnov tests to determine whether the data followed a normal distribution. A mean standard error was calculated for quantitative data with a normal distribution, while a median (95% confidence interval) was calculated for quantitative data without a normal distribution. To analyze more than two groups under the assumption of normality, one-way analysis of variance (ANOVA) with Tukey’s post hoc test or Kruskal–Wallis test with Dunn’s multiple comparison tests was used [1].

5. Conclusions

The results of our study support the concept of combining antigen-specific oral tolerance with AAV-GT to prevent unwanted immune responses against therapeutic transgene products. Inducing oral immune tolerance, followed by AAV-GT administration, provides a sustained enzyme supply, thereby improving bone pathology in MPS IVA mice without elevating anti-hGALNS antibodies. Early diagnosis of MPS IVA is achievable through newborn screening programs in various countries [58]. Therefore, this oral treatment can be applied to affected newborns. A novel approach to in vivo GT using oral immunogenic peptides will provide a new therapeutic paradigm for MPS IVA patients. It may also have a wide range of applications in other MPS types, lysosomal storage diseases, and genetic diseases.