Humoral immune response to equine mesenchymal stem/stromal cells (MSCs): MHC compatibility has more influence than inflammatory priming or chondrogenic differentiation

Abstract

Background

Mesenchymal stem/stromal cells (MSCs) hold great potential for treating various conditions, but their effectiveness and safety in allogeneic therapies can be affected by the recipient’s immune response, among other factors. Allogeneic MSC administration has been associated with immune recognition in different species, including horses. These animals are both veterinary patients and translational models, offering valuable insights into how allogeneic MSCs interact with the immune system in vivo. Factors such as inflammation, chondrogenic differentiation, and major histocompatibility complex (MHC) compatibility between donor and recipient can influence immune responses, potentially affecting the feasibility of repeated MSC administrations. This study evaluates the humoral immune response following repeated administration of chondrogenically differentiated (MSC-chondro), pro-inflammatory primed (MSC-primed), and basal (MSC-naïve) equine MSCs into autologous, MHC-matched, and MHC-mismatched recipients.

Methods

Equine MSC-chondro, MSC-primed or MSC-naïve were embedded into alginate scaffolds and repeatedly implanted subcutaneously into autologous, MHC-matched or MHC-mismatched allogeneic horse recipients. Serum samples collected before and 1, 3, and 6 weeks after each administration were analysed for cytotoxic allo-antibodies against donor MHC. Recipient sera (neat, 1:2 and 1:16 dilution) were tested against donor-derived target cells using two-stage microcytotoxicity assays.

Results

Autologous and MHC-matched recipients did not develop antibodies against the MHC of the donor, regardless of the MSC type administered. In contrast, MHC-mismatched horses developed allo-antibodies after the first administration, with a notable increase following the second administration. All the MSC types triggered a humoral immune response in MHC-mismatched recipients, but MSC-chondro provoked the strongest immune recognition.

Conclusions

Donor-recipient MHC compatibility was more important than inflammatory priming or chondrogenic differentiation of equine MSCs to prevent the development of allo-antibodies. Thus, when donor and recipient are MHC-matched, different types of equine MSCs (MSC-chondro, MSC-primed or MSC-naïve) could be administered without producing a humoral response.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s13287-026-05157-4.

Introduction

Mesenchymal stem/stromal cells (MSCs) have raised interest in the field of regenerative medicine in humans and in several animal species, with a particular focus on horses. Since some diseases share similar underlying mechanisms in both people and animals, research on MSCs can provide valuable insights for both human and veterinary medicine, reinforcing the “One Health, One Medicine” approach [1, 2].

The therapeutic effectiveness of MSCs is primarily attributed to their paracrine activity, with remarkable contribution of their immunomodulatory properties [3, 4]. The discovery and progressive understanding of these properties has led to a major change in the previous paradigm about differentiation-driven therapeutic effects, thus expanding the range for MSC therapeutic applications to inflammatory and immune-mediated conditions [5–7]. Indeed, there is an increasing number of studies aimed at understanding MSC immune-related mechanisms in the equine species, in order to apply them in inflammatory conditions in horses [8, 9]. Furthermore, horses play a significant role in translational medicine, as they suffer from various conditions comparable to human patients, particularly those impacting the musculoskeletal system [10, 11]. Thus, the horse plays a pivotal role in driving translational progress within the regenerative field [2, 10, 12].

To transfer efficiently research into clinical application, the establishment of allogeneic MSC banks derived from healthy donors offers a valuable solution by providing readily available and well-characterized MSCs, thus allowing early-stage disease administration. Additionally, the preservation of specific batches of characterized cells mitigates the variability associated with autologous cell use, facilitating a more homogeneous and expedited therapeutic approach by reducing ex vivo expansion time [6]. Although the use of allogeneic MSCs offers notable advantages, the potential recognition of MSCs by the immune system needs to be clarified. Initially, MSCs were considered immuno-privileged and unlikely to elicit an immune response. However, subsequent investigations have demonstrated that MSCs could be recognized by the immune system of the recipient [13]. Such recognition may have implications for the efficacy and safety of MSC-based therapies, potentially limiting their repeated administration due to the development of humoral-mediated immune memory [14].

The recipient's immune response is influenced by different factors. Of note, donor–recipient compatibility for the major histocompatibility complex (MHC) could influence the recognition of MSCs by the recipient’s immune system. Indeed, the administration of equine allogeneic MSCs into MHC-mismatched recipients can lead to the production of allo-antibodies, while the humoral immune response in MHC-matched recipients is comparable to that of autologous MSCs [14–16]. However, MHC compatibility is not always accounted as a factor when designing studies or clinical trials, which might partially explain certain discrepancies in the outcomes of different studies.

Moreover, and additionally to MHC compatibility, there are other factors that can jointly influence the interaction of the allogeneic MSCs with the immune system. These factors include the expression levels of MHC molecules in the cell surface, and the immunomodulatory capacity of MSCs. Both MHC expression and immune regulatory capacity can vary donor-to-donor [17], but are also influenced by the exposure of cells to an inflammatory environment, as well as upon differentiation of the cells [18, 19]. These factors are particularly relevant as they reflect conditions that MSCs may encounter in vivo following administration at an injury site, and because they have been explored as potential strategies to enhance the therapeutic efficacy of MSC-based treatments.

For instance, exposure of equine MSCs with pro-inflammatory cytokines interferon gamma (IFNγ) and tumour necrosis factor alpha (TNFα) enhanced their regulatory effect in an in vivo model of equine osteoarthritis [20]. However, pro-inflammatory priming can also increase MSC immunogenicity by inducing the expression of MHC molecules [21]. This translated into higher humoral response upon repeated MSC administration in the same model. Nonetheless, partial MHC-compatibility between the donor and the recipient seemed to mitigate the secondary humoral response after re-exposure [18].

On the other hand, chondrogenic induction of MSCs prior to administration has been explored as a therapeutic approach to stimulate the specific secretion of molecules with potential benefit for joint pathologies [22]. However, chondrogenic differentiation may decrease the immunomodulatory properties of MSCs, decreasing their ability to evade the immune system [19, 23, 24]. Nevertheless, other authors reported that, regardless of the MHC haplotype, the repeated administration of chondrogenic and tenogenic induced MSCs did not result in the production of allo-MSC antibodies [25, 26].

Therefore, the immune response triggered by both MHC-matched and MHC-mismatched allogeneic equine MSCs remains poorly understood, particularly when other factors such as inflammation and differentiation are also present. To gain deeper insight into the in vivo interactions between equine MSCs and the immune system, the goal of this study was to evaluate how MSC priming, chondrogenic differentiation, and MHC compatibility between donor and recipient influence the humoral immune response following both single and repeated administrations.

To do this, equine MSCs were evaluated in three conditions: chondrogenically differentiated (MSC-chondro), pro-inflammatory primed (MSC-primed) or in basal conditions (MSC-naïve). Each cell type was administered twice into autologous and allogeneic MHC-matched/mismatched recipients, using a subcutaneous implantation system of hydrogel-encapsulated cells. At different times after each administration, the production of cytotoxic allo-antibodies targeting the donor's MHC was assessed, to characterize the generation of humoral immune responses in vivo.

Our initial hypothesis was that the second administration of equine MSCs would result in increased immune recognition and higher production of allo-antibodies, with a more pronounced response in MSC-chondro recipients compared to MSC-primed recipients, due to their distinct immunomodulatory properties. Additionally, we hypothesized that MHC-matched recipients would generate allo-antibody levels comparable to those of autologous recipients, while the exposure to MHC-mismatched MSCs would trigger a noticeable allo-antibody production.

Materials and methods

Study design

Mesenchymal stem/stromal cells derived from equine bone marrow were repeatedly administered to horses that were either MHC-matched or MHC-mismatched with the MSC donors, and their humoral immune response was analysed over time. Three donor horses homozygous for specific MHC haplotypes (HapPRE10/HapPRE10, HapPRE11/HapPRE11, and HapMAI04/HapMAI04), were selected [27]. The recipient horses were classified into two groups according to their MHC compatibility status relative to the donors. The MHC-matched group included eight heterozygous horses sharing one haplotype with one of the donors: three HapPRE10, three HapPRE11, and two HapMAI04. The MHC-mismatched group consisted of nine horses carrying haplotypes different from HapPRE10, HapPRE11, and HapMAI04. The rationale for using homozygous individuals to match with various heterozygous individuals was based on the concept of haplo-banks [28].

Within each MHC-matched/mismatched group, recipients were further divided into three subgroups of 2–3 horses, each subgroup receiving one type of MSCs: chondrogenically differentiated (MSC-chondro), pro-inflammatory primed (MSC-primed), or basal (MSC-naïve). Additionally, the three homozygous donor horses also served as autologous recipients for control, each donor receiving its own cells under one of the three conditions (Fig. 1A).

Fig. 1.

Study design. A Groups of recipient horses according to donor–recipient MHC compatibility (autologous horses, which also served as donors; allogeneic MHC-matched recipients and allogeneic MHC-mismatched recipients) and according to the type of MSCs administered (MSC-chondro, MSC-primed and MSC-naïve). The MHC haplotypes corresponding to each animal are shown. B Schematic overview of the study design. Each recipient horse received three subcutaneous scaffolds containing the same MSC type. Serum samples were obtained before administration and at 1, 3, and 6 weeks after implantation. The same procedure was repeated one month later to assess the response to a second exposure to the same MSC type. Created with BioRender

Each horse received three alginate hydrogel scaffolds simultaneously into subcutaneous pockets in the neck. Each scaffold contained 5 × 10⁶ MSCs of the assigned condition (MSC-chondro, MSC-primed, or MSC-naïve), making a total of 15 × 10⁶ MSCs per animal and per administration. The rationale for delivering MSCs embedded into scaffolds, and to place these subcutaneously, was to retain the MSCs at a specific anatomical site to minimize potential interference of cell migration. This system also allowed for scaffold recovery at different time points (1, 3 and 6 weeks post-implantation) for a separate study assessing the cellular [27] and local immune response [29]. One month after removing the last implant, each horse was re-exposed to the same MSC condition (MSC-chondro, MSC-primed, or MSC-naïve; autologous, MHC-matched or MHC-mismatched) on the contralateral side of the neck. The scaffolds were retrieved at the same time intervals than after the first administration. This second administration was conducted to investigate the potential development of immune memory mechanisms after repeated administration of the same MSCs.

The present study focused on evaluating the humoral immune response under the conditions described. For this purpose, serum samples were collected from each horse at specific time intervals. Blood samples were taken before the first implantation (T0) and at weeks 1 (T1), 3 (T2), and 6 (T3) post-implantation. One month later, the same procedure was repeated representing a second MSC administration, with blood samples collected at the same time points: prior to MSC re-exposure (T4), and at weeks 1 (T5), 3 (T6), 6 (T7) after re-exposure. A schematic timeline is illustrated in Fig. 1B.

Animal selection by MHC haplotyping

Twenty healthy horses between 2 and 12 years of age (12 geldings, 7 mares, and 1 stallion) were selected based on their MHC haplotype. None of the animals had previously undergone MSC therapy or received blood transfusions. Their MHC haplotypes were identified through microsatellite analysis based on a validated 10-marker panel corresponding to highly polymorphic regions within the MHC genes [18, 30]. Haplotype assignment in these animals was performed in a previous study from our group, where full information about methodology and haplotypes is openly accessible [27].

All procedures involving animals were approved by the Ethical Advisory Committee for Animal Experimentation of the University of Zaragoza (PI-15/16). Animal care and experimental procedures were conducted in accordance with the Spanish regulations for animal protection (RD118/2021), in compliance with European Union Directive 2010/63. Throughout the study, horses were housed in outdoor paddocks at the Animal Research Service facilities of the University of Zaragoza and had ad libitum access to water and grass hay. The work has been reported in line with the ARRIVE guidelines 2.0.

Preparation of MSCs: MSC-chondro, MSC-primed and MSC-naïve

Equine bone marrow-derived MSCs were obtained and characterized following previously established protocols [21] as part of a prior study by our group [27, 31]. Briefly, MSCs (n = 3) were isolated from bone marrow and expanded at a density of 5,000 cells/cm² in standard MSC culture medium consisting of low-glucose Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% foetal bovine serum (FBS; Cat. no. F2442, Lot. 20K284), 2 mM L-glutamine, 0.1 mg/mL streptomycin, and 100 U/mL penicillin (all from Sigma–Aldrich), at 37 °C with 5% CO₂. Cells were used between passages two and four.

Scaffold preparation followed the methodology described by our group [29]. In brief, cells were detached using 0.25% trypsin–EDTA (Sigma–Aldrich), resuspended at a concentration of 5 × 10⁶ MSCs/mL in phosphate-buffered saline (PBS; Gibco, Thermo Fisher), and mixed in a 1:1 ratio with a 3% alginate solution (ultrapure low-viscosity, 67% guluronate; UPLVG, NovaMatrix, FMC Corporation). This yielded a final alginate concentration of 1.5% in a total volume of 2 mL per scaffold, each one containing 5 × 10⁶ MSCs, as previously described [32]. Scaffolds were polymerized by incubation with 102 mM CaCl₂ (Sigma–Aldrich) for 30 min at 37 °C and 5% CO₂, following established protocols [33–35].

For MSC-naïve and MSC-primed conditions, scaffolds were maintained in standard MSC culture medium for 48 h. In addition, MSC-primed scaffolds were stimulated by adding 5 ng/mL of TNFα and IFNγ (R&D Systems) to the culture medium during 12 h before in vivo implantation, based on previous reports [20].

To establish the MSC-chondro condition, chondrogenesis was induced in a three-dimensional alginate matrix system after encapsulation [19, 33]. After alginate polymerization as described above, scaffolds were washed three times in PBS and once with incomplete chondrogeneic differentiation medium (ICM), consisting of high-glucose DMEM supplemented with 10% FBS, 2 mM L-glutamine, 0.1 mg/mL streptomycin, 100 U/mL penicillin, 1% insulin, transferrin and selenium supplements solution (ITS + premix, Beckton Dickinson), 1 mM sodium pyruvate (Gibco), 100 nM dexamethasone, 50 μg/mL ascorbic acid-2-phosphate and 40 μg/mL proline (all from Sigma–Aldrich). Subsequently, two millilitres of complete chondrogenic medium (CCM), consisting of ICM supplemented with 10 ng/mL transforming growth factor beta-3 (TGFβ-3; PeproTech), were added per scaffold. Cultures were incubated at 37 °C and 5% CO₂ for 21 days, with the medium replaced every other day.

To remove potential xenogeneic antigens from the FBS that could influence the immune response, all the scaffolds underwent a depletion process for 24 h prior to in vivo implantation. Specifically, scaffolds were washed twice with PBS, and the corresponding culture medium was replaced with fresh medium supplemented with 10% sterile-filtered autologous serum from each recipient to substitute the FBS [15]. Immediately prior to implantation, the scaffolds were washed three more times with PBS. All procedures related to the construction of the scaffolds, their chondrogenic characterisation, and the optimisation of their in vivo administration have been previously described in detail by our research group [29].

In vivo MSC administration and serum collection

Autologous, MHC-matched, and MHC-mismatched recipients underwent subcutaneous implantation of three alginate scaffolds containing the same type of MSCs (either MSC-chondro, MSC-primed or MSC-naïve) in the neck region. Scaffolds were placed along the neck using a subcutaneous pocket technique [36]. The procedure was carried out with sedation using detomidine (0.01 mg/kg, Sedaquick, Fatro) and butorphanol (0.02 mg/kg, Torbugesic, Pfizer), together with local anaesthesia (lidocaine 5%, Braun). To implant each scaffold, a 2 cm longitudinal incision was made, with a 10 cm gap between each incision. The subcutaneous tissue was carefully dissected distally to the incision site to create a 2.5 × 2.5 cm pocket, where the scaffold was placed. The surgical incisions were subsequently sutured in two planes, with the subcutaneous tissue closed using 2/0 USP polyglyconate sutures and the skin closed with surgical staples. This administration approach was selected to enable the evaluation of implant-associated responses in a parallel study [29]. Horses were clinically monitored throughout the procedure and received a single intravenous dose of flunixin meglumine (1.1 mg/kg, Niglumine, Calier) immediately before surgery, followed by intramuscular procaine penicillin (15 mg/kg once daily for 3 days, Procapen, Livisto) during the postoperative period. Each one of the three scaffolds was retrieved after 1, 3 and 6 weeks, respectively, for analysis of the local immune response [29]. One month after recovering the last scaffold, the procedure was repeated on the other side of the neck.

To assess the humoral immune response, sequential collection of sera from all recipients was conducted prior to each MSC administration, and prior to the extraction of each scaffold, following the time-points described above in the Study Design. Ten mL of blood were collected from the jugular vein into tubes without anticoagulant. The tubes were left in vertical position for 24 h to allow the clot formation and serum separation. Subsequently, the blood collection tubes were centrifuged at 1,200 × g for 15 min and approximately 2–4 mL of serum per animal were recovered and stored at − 20 °C.

Preparation of target cells for microcytotoxicity assays

To perform the microcytotoxicity assays, passage 2–4 cryopreserved MSCs from the selected donors were thawed and cultured in standard MSC culture medium at 37 °C and 5% CO₂ for one week. Four days prior to conducting the microcytotoxicity assays, the FBS in the medium (10%) was substituted with a combination of 5% FBS and 5% horse serum (Pan–Biotech, Cat. no. P30-0701) tested for different equine diseases, free of xenogeneic antigens and previously heat-inactivated. After 48 h, the medium was replaced again by fresh medium containing 10% equine serum for additional 48 h. Subsequently, MSCs were detached with 0.25% trypsin–EDTA (Sigma–Aldrich) and washed with PBS. The cell concentration was adjusted to 2 × 10⁶ cells/mL in PBS to perform the microcytotoxicity assays immediately.

Microcytotoxicity assays

The microcytotoxicity dye exclusion assay was conducted using equine MSCs as target cells to detect cytotoxic allo-antibodies against the MHC of the administered MSCs following established protocols [16–18]. The MSCs derived from each donor were tested against sera (neat, diluted 1:2 and 1:16) collected at the 8 designated time points from the different recipient horses, as described above under the Study Design.

To establish a positive control, it was used equine MSCs and sera from a previous study of the group, whereby cell death exceeding 80% was observed [18, 20]. The commercial horse serum described in the above section was used neat as a negative control. Another negative control was established where only MSCs but not sera were used. These controls enabled determining that the cell death observed under experimental conditions could be attributed to a cytotoxic reaction mediated by the presence of allo-antibodies in recipients’ sera.

Thawed sera from all horses were centrifuged at 12,000 × g for 10 min at 4 °C and each serum sample was then evaluated undiluted (neat) as well as in serial dilutions of 1:2 and 1:16, using commercial horse serum as the diluent. Subsequently, 1 μL of corresponding serum was incubated with 1 μL of the MSCs suspension (total of 2,000 MSCs/well) in Terasaki plates (One Lambda) under heavy mineral oil (Sigma–Aldrich). After 30 min RT incubation, 5 μL of rabbit complement (One Lambda) were added and plates were incubated 1 h at RT. Each well was stained with 2 μL of 5% Eosin Y dye (Sigma–Aldrich) for 5 min and fixed with 5 μL of 10% formalin (Sigma–Aldrich, pH 7.2–7.4). The plates were kept at 4 °C for a minimum of 2 h to allow the cells to settle at the bottom of the wells for microscopic evaluation. Two technical replicates were performed for each condition (recipient, time point, and serum dilution), as well as for each control type (positive, negative and only target cells).

Within a timeframe of 12 h, three authors (LB, AC, and EB) evaluated the percentage of dead cells using an inverted phase contrast microscope (Nikon Eclipse TS100) at a magnification of 10x. Live cells were identified by their refringent and rounded appearance, while dead cells exhibited a dark, flat, and irregular morphology. Supplementary Material 1 includes representative images illustrating the morphology of live and dead cell populations. The cytotoxicity produced under each condition was assessed using the scoring system previously described [15, 16]: grade 1 (< 10% dead cells), grade 2 (10–19% dead cells), grade 4 (20–49% dead cells), grade 6 (50–80% dead cells), grade 8 (81–100% dead cells). This evaluation system considers that cytotoxicity scores equal to or greater than 6 (indicating over 50% of cell death) in neat sera are considered positive for allo-antibodies, while scores equal to or lower than 2 (indicating less than 20% of cell death) in neat serum were considered as no relevant presence of allo-antibodies in the serum. The average cytotoxicity score of technical replicates was calculated and used for subsequent statistical analyses.

Statistical analysis

Statistical analyses were performed using IBM SPSS Statistics 26.0 software (SPSS, Inc). Interrater reliability was assessed using Cronbach’s alpha for each target cell type within each recipient group (MSCs-chondro MHC-matched and autologous, α = 0.948; MSC-primed MHC-matched and autologous, α = 0.971; MSC-naïve MHC-matched and autologous, α = 0.956; MSC-chondro MHC-mismatched, α = 0.923; MSC-primed MHC-mismatched, α = 0.919; MSC-naïve MHC-mismatched, α = 0.920). Provided the high interrater agreement, mean scores of the three observers were calculated and used for subsequent statistical analysis, as in previous studies [18].

In the statistical analysis, mean cytotoxic scores above 6 and below 2 were considered indicative of significant and non-significant antibody production, respectively, within each recipient group. The three homozygous donor horses, which served as autologous controls, were excluded from analytical statistics because each group consisted of only one animal. Statistical analyses were conducted to evaluate score differences according to the study variables. These variables included “group” (2 categories: MHC-matched and MHC-mismatched), “cell type” (3 categories: MSC-chondro, MSC-primed and MSC-naïve), “time” (8 categories: T0, T1, T2, T3, T4, T5, T6, T7) and “score” (3 categories: neat serum, 1:2 diluted serum, and 1:16 diluted serum). Cytotoxicity score was used as the dependent variable, whereas the remaining variables were considered as factors to evaluate: (1) differences between recipient groups for each target cell type at each time-point, (2) differences between types of target cells at each time within each group of recipients and (3) differences over time within each group of recipients for each cell type. For the different statistical analyses, non-parametric tests were used after assessing normality using the Shapiro–Wilk test. The Mann–Whitney U test was used for comparisons between two groups (1). For variables with more than two groups, Friedman or Kruskal–Wallis tests followed by Dunn's test as a post hoc procedure were used for related (3) or independent (2) samples, respectively, with p-values adjusted using the Bonferroni correction to control for multiple comparisons. The level of significance was established at p < 0.05 for all statistical analyses. Graphs were generated using GraphPad Prism 10.4.

Results

The presence of cytotoxic allo-antibodies developed against different types of equine MSCs (MSC-chondro, MSC-primed and MSC-naïve) after their repeated administration into three types of recipients (autologous, MHC-matched and MHC-mismatched) was assessed by microcytotoxicity assays. All the negative controls remained below score 2, and baseline values at T0 were also consistently lower than 2 in both MHC-matched and MHC-mismatched recipients, confirming the absence of pre-existing antibodies in all the recipients prior to MSC administration. Autologous and allogeneic MHC-matched recipients did not develop relevant levels of cytotoxic antibodies against the MHC of the MSC donor, regardless of the type of MSCs received. In these recipients, all the analysed sera yielded scores ranging from 1 to 2 across all the dilutions tested. On the other hand, sera from MHC-mismatched recipients showed progressively higher cytotoxic scores over time against all the three MSC types. In this group, scores generally ranged from 1 to 4.5 across time points and remained below 6 under all the tested conditions, including recipient group, target cell type, and serum dilution (neat, 1:2, and 1:16).

Horses receiving MHC-mismatched MSCs exhibited a consistently higher production of allo-antibodies compared to MHC-matched horses across all the tested time-points

When comparing cytotoxic scores between MHC-matched and MHC-mismatched groups at each time-point, a consistent pattern was observed: regardless of the MSC type received, autologous and allogeneic MHC-matched recipients showed similarly low cytotoxic scores with average values below 2, while MHC-mismatched recipients showed markedly higher scores (Fig. 2).

Fig. 2.

Evolution of cytotoxic scores along the time comparing the three groups of recipients (autologous, MHC-matched and MHC-mismatched) grouped per cell type received (MSC-chondro, MSC-primed, MSC-naïve). Mean ± S.E.M. of cytotoxic scores (Y axis) assigned to neat sera (top row; A, B, C), 1:2 diluted sera (middle row; D, E, F) and 1:16 diluted sera (bottom row; G, H, I) from recipients of MSC-chondro (green bars, left column; A, D, G), MSC-primed (orange bars, middle column; B, E, H) and MSC-naïve (blue bars, right column, C, F, I) along the time (X axis; T0, pre-administration of corresponding MSCs; T1, 1 week after first MSCs administration; T2, 3 weeks after first MSCs administration, T3, 6 weeks after first MSC administration, T4, just before the second MSC administration; T5, 1 week after second MSCs administration; T6, 3 weeks after second MSCs administration, T7, 6 weeks after second MSCs administration). At each time-point, the three groups of recipients are presented: autologous (light coloured), MHC-matched (middle coloured) and MHC-mismatched (dark coloured). Asterisks (*) point out statistically significant differences (* p < 0.05) between groups of recipients at a specific time-point

Notably, in recipients of MSC-chondro, significantly higher production of allo-antibodies was detected in MHC-mismatched recipients over MHC-matched recipients across all the time-points and dilutions analysed (p < 0.05 in all time-points; Fig. 2A, D and G).

In the case of MSC-primed recipients, sera from MHC-mismatched recipients also displayed higher cytotoxic scores compared to sera from MHC-matched recipients at all the analysed time-points. However, none of these differences reached statistical significance, although all the analysed comparisons yielded p-values below 0.1 (Fig. 2B, E and H).

MHC-mismatched recipients of MSC-naïve showed higher cytotoxic scores as well. Compared to MHC-matched MSC-naïve recipients, the production of allo-antibodies was significantly higher from T1 onwards in both neat sera (p < 0.05; Fig. 2C) and 1:2 diluted sera (p < 0.05; Fig. 2F). At the 1:16 dilution, cytotoxic scores remained significantly different between MHC-matched and MHC-mismatched recipients from T2 onwards (p < 0.05; 2I).

MHC-mismatched administration of MSC-chondro produced higher levels of allo-antibodies than MSC-naïve and MSC-primed

Depending on the type of MSCs administered, MHC-mismatched recipients showed higher cytotoxic scores at specific time points. Overall, recipients of MSC-chondro tended to show higher antibody production compared with those that had received either MSC-primed or MSC-naïve (Fig. 3). On the other hand, MHC-matched recipients showed scores ranging between 1 and 2 in all cases and no significant differences were found depending on the type of MSCs received (Fig. 2).

Fig. 3.

Evolution of cytotoxic scores along the time in MHC-mismatched recipients, comparing the different MSC types administered. Mean ± S.E.M. of cytotoxic scores (Y axis) assigned to neat (A), 1:2 diluted (B), and 1:16 diluted sera (C) obtained from recipients of MSC-chondro (green bars), MSC-primed (orange bars) and MSC-naïve (blue bars) along the time (X axis; T0, pre-administration of corresponding MSCs; T1, 1 week after first MSC administration; T2, 3 weeks after first MSC administration; T3, 6 weeks after first MSC administration; T4, just before the second MSC administration; T5, 1 week after second MSC administration; T6, 3 weeks after second MSC administration; T7, 6 weeks after second MSC administration). Significant differences between a particular time-point over either T0, T1, T2, T3 or T6 are respectively represented by asterisk (*), circumflex (^), hashes (#), cross ( +) and hyphen (–) above the corresponding bar (*, #, ^, + , –, p < 0.05; **, ^^, p < 0.01; ***, p < 0.001). Significant differences between different target cells at a specific time-point are represented by a squared line with an asterisk (* p < 0.05)

In MHC-mismatched recipients, significantly higher cytotoxic scores were observed at one week after the first administration (T1) of MSC-chondro compared to MSC-naïve, both in 1:2 diluted sera (p < 0.05; Fig. 3B) and in 1:16 diluted sera (p < 0.05; Fig. 3C). Furthermore, one week after the second administration (T5), higher cytotoxic scores were also observed for MSC-chondro over MSC-naïve recipients in neat sera (Fig. 3A), 1:2 diluted sera (Fig. 3B), and 1:16 diluted sera (Fig. 3C), as well as at T7 in neat sera (Fig. 3A) and in 1:2 diluted sera (Fig. 3B) (p < 0.05 in all cases). In addition, MSC-chondro induced higher allo-antibody levels than MSC-primed at T5 in neat sera (p < 0.05; Fig. 3A). MSC-primed produced higher cytotoxic scores compared to MSC-naïve, but significant differences were only punctually found at T2 and T7 in 1:2 diluted sera (p < 0.05 in both cases; Fig. 3B).

The production of allo-antibodies increases after the second administration into MHC-mismatched horses

Throughout all the evaluated time-points, a progressive increase in cytotoxic scores was observed in both neat and 1:2 diluted sera from MHC-mismatched horses, with the most pronounced increase occurring after the second administration. Although the scores in 1:16 diluted sera followed the same trend, no significant changes were detected along the time-points analysed (Fig. 3). Conversely, recipients of either MHC-matched or autologous MSCs did not show changes along the time and presented scores up to 2 in all cases (Fig. 2).

Scores in neat sera did not significantly change at any time-point after the first administration of any MSC type into MHC-mismatched horses, although an increasing trend was noticed. In contrast, following the second administration, significant differences were detected over the earliest time-points (T0 and T1). In particular, and regardless of the type of MSCs received, the scores from the MHC-mismatched recipients after the second administration showed a significant increase over T0 at: T5 (MSC-primed and MSC-naïve, p < 0.05; MSC-chondro, p < 0.01), T6 (MSC-chondro and MSC-primed, p < 0.05; MSC-naïve, p < 0.01), and T7 (MSC-chondro and MSC-primed, p < 0.01; MSC-naïve, p < 0.05) (Fig. 3A). Furthermore, the MHC-mismatched horses that received either MSC-chondro or MSC-naïve also showed a significant increase in allo-antibody production after the second administration compared to one week after the first one (T1): at T5 (p < 0.05 in both cases), T6 (MSC-chondro, p < 0.05; MSC-naïve, p < 0.01), and T7 (p < 0.05 in both cases). Although a similar trend was observed in MHC-mismatched horses receiving MSC-primed, the cytotoxicity increase was only statistically significant at T7 compared to T1 (p < 0.05) (Fig. 3A).

Consistent patterns were observed in the serial dilutions of sera across the three groups of MHC-mismatched recipients, albeit with lower average scores. Overall, and as observed in the neat sera, 1:2 diluted sera from MHC-mismatched recipients after the first exposure showed a gradual increase in cytotoxic scores across all time points (Fig. 3B). This increase was statistically significant in recipients of MSC-naïve (at T3 and at T4 compared to T0, p < 0.05) and in recipients of MSC-primed (at T4 compared to both T0 and T1, p < 0.05), while recipients of MSC-chondro did not show significant changes in allo-antibody production after the first administration. Following the second administration, scores in 1:2 diluted sera from all the MHC-mismatched recipients increased compared to T0, which was statistically significant at: T5 (MSC-chondro, p < 0.001; MSC-primed, p < 0.05; MSC-naïve, p < 0.01), T6 (MSC-chondro, p < 0.05; MSC-primed, p < 0.01; MSC-naïve, p < 0.001) and T7 (MSC-chondro, p < 0.01; MSC-primed, p < 0.05). Moreover, the second administration of all the three types of MSCs into MHC-mismatched recipients induced higher levels of allo-antibodies than the first administration. In particular, MSC-chondro recipients showed a significant increase in cytotoxic scores at one week after the second administration (T5) compared to the first administration (T1, T2, and T3; p < 0.05 in all cases) (Fig. 3B). Recipients of MSC-primed showed significantly higher scores at T5, T6 and T7 compared to T1 (T5 and T7, p < 0.05; T6, p < 0.01) (Fig. 3B), while horses that received MSC-naïve increased the scores at T6 over both T1 and T2 (p < 0.05 in both cases). Of note, cytotoxic scores decreased in 1:2 diluted sera from recipients of MSC-naïve at the final time-point over the previous time point (T7 over T6, p < 0.05), with T7 scores showing similar levels to those obtained prior to the second administration (T4) (Fig. 3B).

Discussion

This is the first work integrating the study of both MHC compatibility and MSC condition (MSC-naïve, MSC-primed, and MSC-chondro) within the same in vivo experimental framework. This combinatorial approach allows advancing the current knowledge on humoral responses against MSCs in the horse, which so far has focused on the study of separate factors. By directly comparing different MSC types across MHC compatibility settings, our results show that MHC compatibility was the main determinant of humoral immune memory, regardless of the MSC type. The development of a humoral immune response was restricted to MHC-mismatched recipients, while it was not observed in MHC-matched or autologous recipients of any MSC type. Of note, the choice of the MSC type may influence the magnitude of the response in an MHC-mismatched context, with MSC-chondro raising higher antibody levels. Furthermore, the progressive increase in cytotoxic allo-antibodies, peaking after the second administration, confirms the development of both primary and secondary humoral responses in all MHC-mismatched scenarios, with potential implications for the repeated use of allogeneic MSCs of any type.

Limitations of the study and particular considerations

First, the sample size is limited due to the challenges inherent in working with large animal species such as horses [10], and particularly with MHC-matched individuals. This limitation was particularly evident in the MSC-primed MHC-matched group (n = 2), where the high variability of MHC haplotypes [30] prevented us from identifying additional MHC-matched animals. This lower sample size in comparison to the other allogeneic groups might have hindered the detection of further significant differences within this group, although its trends aligned with those observed in the rest of allogeneic MHC-matched groups.

Second, the autologous group consisted of only three horses (one per MSC type), as they were involved solely for control purposes as previously described [15, 16].

Third, the subcutaneous administration of MSCs embedded within hydrogels does not fully mirror clinical application but was selected to facilitate immune cell access —as other sites such as joints or eyes are less accessible—, ensure cell immobilization at a defined anatomical site, and allow scaffold retrieval for further analyses in a separate study [29]. Nevertheless, the subcutaneous implantation of scaffolds may be relevant for certain clinical scenarios, such as soft tissue surgeries or wound healing [36]. Furthermore, the use of scaffold-based delivery systems, which are already applied in tissues such as cartilage and tendon [33–35], highlights the need to consider the potential role of scaffolds as physical barriers that may modulate immune recognition and cell–host interactions in vivo [29].

Fourth, it was not possible to use all the MSC types as target cells in the microcytotoxicity assays. While MSC-naïve and MSC-primed cells have previously been used as target cells in similar assays [18], the use of MSC-chondro posed technical limitations: the high morphological heterogeneity and extracellular matrix deposition associated with chondrogenic differentiation prevented accurate cell counting and reliable discrimination between live and dead cells [37, 38]. Therefore, MSCs from the same donor and passage as those administered in vivo were used as target cells, but in the form of MSC-naïve in all cases. While we cannot disregard the possibility that MSC-primed and MSC-chondro might have provided slightly different outcomes than MSC-naïve, previous studies have reported comparable allo-antibody detection using different cell types—either MSCs or donor-derived peripheral blood leukocytes- as target cells [16].

Different factors have different influence in the humoral immune response following MSC administration: the role of MHC compatibility and MSC type

According to our results, the humoral immune response against equine MSCs in vivo is mostly influenced by the MHC compatibility. This finding is consistent with previous studies in horses focusing on this single factor, which reported a lack of humoral immune responses after repeated administrations of autologous MSCs [14, 39] as well as of allogeneic but MHC-matched MSCs [14–16]. These studies also assessed the humoral response developed in MHC-mismatched horses, and in all cases, allo-antibodies were generated, regardless of whether the horses received one or two administrations of MSCs. In our study, which also included the comparison of different MSC types, similar observations were made in all MHC-mismatched recipients in spite of the MSC type administered; however, horses receiving MSC-chondro showed the most pronounced and rapid antibody increase upon re-exposure. This finding is consistent with our initial hypothesis that MSC-primed cells would elicit a weaker humoral response than MSC-chondro, likely because of the activation of the immunomodulatory profile of licensed MSCs.

Previous investigations assessed the humoral immune response in horses following repeated allogeneic administration of differentiated MSCs, in particular, tenogenic primed MSCs (tpMSCs) and chondrogenic induced MSCs (ciMSCs)—both types originating from peripheral blood-derived MSCs [25, 26]. However, the MHC compatibility was not accounted in such investigations. These studies reported no presence of antibodies against ciMSCs or tpMSCs, indicating the absence of humoral response development after multiple injections. Of note, in these studies, antibody detection was performed using a flow cytometric crossmatch assay, which primarily detects antibodies against MSCs rather than donor-specific anti-MHC antibodies. This methodological difference—along with variations in MSC sources, differentiation protocols, and the specific time-points assessed—may account for the discrepancies between our results and other reports.

In a previous study from our group where the same types of MSCs were administered and the systemic cellular response was assessed [27], the second administration of MSC-chondro induced a stronger B-cell response than the first exposure. Interestingly, this response was detected in both MHC-matched and mismatched recipients of MSC-chondro, but to a lesser extent in the former. That observation would not align with the lack of antibody formation in MHC-matched recipients observed in the current study. However, since the microcytotoxicity assay used here detects antibodies directed against the MHC, it could be hypothesized that chondrogenically differentiated equine MSCs might induce immune memory through other mechanisms, even in the absence of foreign MHC molecules. Nevertheless, MSC-chondro produced a higher B cell proliferative response in MHC-mismatched recipients, and this response was stronger than in recipients of MSC-primed [27], thus correlating with the pattern also observed in the current study.

Previous work in horses has shown that MSC-primed cells accelerate primary and secondary humoral responses when administered to MHC-mismatched recipients [18]. Consistent with these findings, our results confirm this effect and further extend previous observations by demonstrating that MSC-primed cells do not elicit a humoral response when administered to MHC-matched recipients.

It has been proposed that the increased expression of MHC molecules on MSC-primed cells may contribute to an accelerated humoral response in MHC-mismatched recipients [18]. However, it should also be noted that the expression of MHC by MSCs can vary in vivo after administration. Our group recently reported that MHC-II was upregulated in vivo in equine MSC-chondro, MSC-primed and MSC-naïve, regardless of being administered into autologous, MHC-matched, or mismatched animals [29]. Nevertheless, in the current study, neither autologous nor MHC-matched animals developed antibodies in spite of the type of MSCs received, reinforcing the notion that MHC compatibility can attenuate immune recognition despite varying expression of MHC antigens.

Previous studies have shown that inflammation-induced changes in MSC MHC expression do not necessarily translate into different humoral responses, as similar allo-antibody levels were observed after repeated administration of allogeneic MSCs in both healthy and inflamed joints [39]. Moreover, allogeneic equine MSCs have been reported to induce allo-antibody production even in the absence of MHC class II expression [15]. The fact that MSCs are capable of inducing allo-antibody production even if expressing low levels of MHC underscores the importance of our study, which simultaneously examined different types of MSC—potentially displaying distinct MHC expression profiles—while directly comparing matched and mismatched settings.

Regarding MSC-naïve, our findings demonstrate that MHC-mismatched recipients develop allo-antibodies, in agreement with previous reports in horses [14–16]. Moreover, consistent with Rowland et al. (2021), our data suggest that MSC-naïve elicited humoral responses of limited magnitude and duration that did not persist in the long term. These findings not only confirm previous evidence but also support the soundness of novel findings in other MSC types.

Monitoring the temporal progression of the humoral response and evaluating the effect of each administration provides valuable insight into the onset, peak, and resolution of antibody production, thereby enhancing our understanding of immune response dynamics and the safety of multiple administration of a variety of MSC types in diverse MHC contexts. Our results point out at the development of immune memory that may restrict the repeated administration of MSCs of any type in MHC-mismatched recipients: while humoral responses were already observed after the first administration, the response after the second administration became more pronounced, consistent with the development of a secondary humoral immune reaction.

Previous studies have shown that allo-antibody responses against MHC-mismatched MSCs typically develop within the first two weeks after administration [14, 16, 18, 39] and peak around 2–3 weeks [15]. Consistent with these reports, our study observed a progressive increase in cytotoxic allo-antibodies after both administrations in all MHC-mismatched recipients and showed that this pattern is produced by all the MSC types assessed. Nevertheless, the humoral response recorded in this study seemed to be slower and of lower magnitude compared to previous studies. As discussed in the limitations section, this might be partly explained by the use of MSCs embedded into alginate scaffolds, which could have limited and/or delayed MSC recognition by the host immune system [34], warranting further investigation into how scaffolds may affect immune recognition of therapeutic cells. In addition, differences in humoral responses compared with previous studies may also be influenced by haplotype-dependent variability or unknown partial compatibility between specific MHC haplotypes. Rowland et al. 2021 reported reduced allo-antibody responses following repeated administration of MHC-mismatched MSCs in recipients carrying the A2 haplotype. Although inter-individual variability cannot be excluded, these observations suggest that certain MHC haplotypes may attenuate immune targeting even in mismatched settings. Further studies are needed to test this hypothesis, but the findings of this study underscore the importance of conducting a comprehensive immune monitoring in MSC-based therapies.

Conclusions

This study highlights the importance of MHC compatibility between the donor and the recipient to mitigate the development of a humoral immune response when administering MSCs. This factor could potentially compromise the safety and the effectiveness of repeated allogeneic MSC therapies. Importantly, different types of equine MSCs (MSC-chondro, MSC-primed or MSC-naïve) can be administered without producing a humoral response, as long as donor and recipient are genetically matched for the MHC. However, if MHC-mismatched cells are to be used, MSC-chondro might raise higher levels of allo-antibodies when repeatedly administered. Future in vivo investigations are warranted to elucidate the clinical implications of these humoral responses for the efficacy and safety of allogeneic therapies. Moreover, further research is needed to identify strategies that may help attenuating alloantibody development and improving the survival of allogeneic MSCs in vivo when administered into MHC-mismatched individuals.

Abbreviations

ciMSCs

Chondrogenic induced MSCs

DMEM

Dulbecco’s Modified Eagle’s Medium

|FBS

Fetal bovine serum

IFNγ

Interferon gamma

MHC

Major histocompatibility complex

MSC

Mesenchymal stem cell

PBL

Peripheral blood lymphocytes

PBS

Phosphate buffered saline

TGF-β3

Transforming growth factor beta-3

tpMSCs

Tenogenic primed MSCs

TNFα

Tumor necrosis factor alpha

Author contributions

All authors (AC, AR, EB, BS, AV, SF, IF, PZ, FJV, LB and CR) substantially contributed to the conception and design of the study. AC, LB, EB and IF were responsible for acquisition of data. AC, LB, FJV, EB and CR were responsible for analysis and interpretation of data. AR, AV, SF, BS, EB, IF and FJV provided technical support. LB, AR, FJV, PZ and CR provided important conceptual guidance. CR, PZ and LB were responsible for obtaining funding. AC drafted the article and all authors (AR, EB, BS, AV, SF, IF, PZ, FJV, LB and CR) revised the manuscript for important intellectual content. All authors approved the final submitted version of the article and agree to be accountable for all aspects of the work in ensuring that accuracy or integrity of any part of the work are appropriately investigated and resolved. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by Ministerio de Industria, Economia y Competitividad, Spain (Grant number AGL2017-84411-P, MINECO/AEI/FEDER, UE); Ministerio de Ciencia e Innovación, Spain (Grant Numbers PID2020-116352GB-I00 / AEI /10.13039/501100011033 and PDC2021-121047-I00 / AEI / 10.13039/501100011033) and Gobierno de Aragón (Grupo de Investigación A19_23R, LAGENBIO). This work was also funded by the Complementary Plans of the Recovery, Transformation and Resilience Plan (PRTR) within the Biotechnology Applied to Health frame, co-financed by the Ministerio de Ciencia, Innovación y Universidades, NextGenerationEU, and Gobierno de Aragón, through the Instituto de Investigación Sanitaria de Aragón (IIS Aragón). This research was additionally funded by the Instituto Agroalimentario de Aragón (IA2) through the Strategic Research Line “Medicina regenerativa veterinaria: abordaje One Health para mejorar la salud de animales y personas”. AC was supported by a PhD fellowship from the Gobierno de Aragón and co-funded by European Social Fund (ESF). BS was supported by a PhD fellowship from the Gobierno de Aragón. EB contract is part of the grant PRE2021-097156, funded by MCIN/AEI/10.13039/501100011033 and ESF + . LB was supported by the international talent attraction programme Manuel López, funded by the Programa Propio de Política Científica of the Universidad de Zaragoza.

Data availability

The raw experimental data that support the findings of this study are available in Zenodo with the identifier (https://doi.org/10.5281/zenodo.19350208).

Declarations

Ethics approval and consent to participate

All the procedures involving animals were carried out under the Project License PI 15/16 approved by the Advisory Ethics Committee for Animal Research from the University of Zaragoza (title: “Optimización del uso de MSCs alogénicas en el tratamiento de patologías articulares equinas: equilibrio inmunomodulación-inmunogenicidad” [“Optimizing the use of allogenic MSCs for treating equine joint pathologies: immunomodulation-immunogenicity balance”]; approval date: 22 June 2021). The care and use of animals were performed accordingly with the Spanish Policy for Animal Protection RD118/2021, which meets the European Union Directive 2010/63. The work has been reported in line with the ARRIVE guidelines 2.0.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.