Thymic Innervation Impairment in Experimental Autoimmune Encephalomyelitis Open Access

The thymus is a primary immune organ, formed by lymphoid and microenvironmental compartments, and is responsible for the generation of T lymphocytes. Intrathymic T-cell development is controlled by the thymic microenvironment, which comprises epithelial cells, dendritic cells, macrophages, and fibroblasts, as well as extracellular matrix protein and components of the nervous system [1‒3]. Thymic innervation is mainly composed by catecholaminergic fibers, which play a role in thymocyte differentiation [4]. Moreover, both lymphoid and epithelial cells in the thymus express receptors for neurotransmitters, and the thymic microenvironment also release some neurotransmitters [2]. Although the presence of the nerve fibers in the thymus is well established since the 1980 decade, little is known about the impact of dysfunctional local nerve fibers during thymic involution.

Although in health conditions the thymus involutes with aging, some conditions may induce a premature involution of the organ. In most cases, especially in acute infectious diseases, thymic involution results mainly in thymocyte death, and to a much lesser extent, from the increased egress of immature thymocytes to the periphery of the immune system [1, 5]. Interestingly, artificial thymus aging results in a failure of negative selection and impairment in the generation of regulatory T cells, as well as potential expansion of autoreactive T cells able to induce chronic inflammation [6]. Therefore, optimal thymic operation is crucial to avoid and control autoimmune responses. Indeed, although the majority of autoimmune diseases present multifactorial etiology, monogenetic autoimmune diseases are often related to alterations in the self-tolerance mechanisms [7]. Actually, it is widely accepted that individuals with autoimmune diseases, in general, fail to proper eliminate autoagressive cells [8]. Interestingly, most autoimmune diseases are not age-related, affecting mainly young adults (20–40 years old).

As regards multiple sclerosis (MS), there is a premature involution of the thymus paralleling disease progression [9]. MS is an autoimmune and neurodegenerative disease of the central nervous system (CNS), being characterized by CNS lesions in the white and grey matters mediated by auto-reactive CD4⁺T cells type 1 and 17 [10, 11]. The lesions result in disruption of the flow of information within the CNS as well as between the CNS and body, which yields a series of symptoms, including motor and sensory loss, pain, fatigue, and cognitive dysfunction [12, 13].

In this scenario, it is conceivable that thymic innervation in MS could be altered, with putative dysregulation of the thymic homeostasis and function. This prompted us to evaluate the thymic innervation in the murine model of MS, namely experimental autoimmune encephalitis (EAE), attempting to correlate putative changes in the thymic innervation with alterations in the lymphoid and microenvironmental cell compartments.

Our results show that during EAE development, the thymus is targeted, with changes in the innervation pattern, beta-adrenergic receptor expression by thymocytes, as well as changes in serotonin contents and synaptophysin release. A deep analysis of thymic cellularity also revealed both thymocyte and epithelial cell impairments, indicating direct connection between CNS integrity and thymus function maintenance.

Eight-weeks-old female C57BL6/J mice and Lewis rats were immunized with 15 μg of MOG_35-55 or 50 μg myelin basic protein (MBP) fragment, respectively, or ovalbumin (OVA), in the presence of Freund’s Complete Adjuvant (CFA) containing 2 mg/mL of Mycobacterium tuberculosis H37RA (Difco, Detroid, MI, USA), as previously described [14‒16]. The disease clinical development was graded on a severity scale of 0–5 using a well-known system: 0 for no disease, 1 for loss of tail tone, 2 for hind limb weakness, 3 for hind limb paralysis, 4 for hind limb paralysis/forelimb paralysis or weakness, and 5 for moribund/death [17]. Mice grouped at MOG onset analysis presented scores of 0 or 1 on the 10th day of immunization and MOG peak group contained mice presenting scores of 2–4 on the 14th day of immunization. Mice that did not develop the disease by the day of study were not included to the study. Rats also presented a graded system from 0 to 5; but the first clinical signal appeared within 11 ± 1 and disappeared by 19 ± 1 days after immunization [15]. Animals were kept in the maintenance room of the animal house in the Institute of Biology, University of Campinas (Campinas, Brazil), under species pathogen-free conditions. Experiments were performed three times, using 3 mice, per group, in each experiment. All protocols were approved by the Institutional Ethical Committees (CIBIO #1/2013, CEUA #3889-1; #4393-1; #3626-1, #3754-1 University of Campinas).

Control OVA, MOG_35-55 immunized mice, and MBP immunized rats were anesthetized and weighed. Thymuses were then collected, cleaned, and weighed. The thymic index for each animal was calculated using the following equation: thymus/mouse weight (mg/g).

For histological evaluation, thymuses were collected and fixed in a solution 2% paraformaldehyde for 12 h at room temperature. The specimens were submitted to diaphanization with xylene, dehydrated by graded ethanol series, embedded in paraffin, and cut into 5 µm thick sections, which were then stained with Hematoxylin-Eosin (H&E). Histologic changes were evaluated and photographed using an Inverted Microscope Axioshop (Carl Zeiss AG, Germany).

For immunostaining, mice were sacrificed by exsanguination under anesthesia, according to the humanized procedures and approved protocols (listed above). Thymic sections were obtained with a cryostat and fresh frozen sections were thaw-mounted onto slides, fixed with ice-cold acetone, and blocked with BSA (1% w/v). The slides were incubated with reagents that label specific thymic structures, at 1:100 dilution, namely: anti-Ly-51 (130-101-844) antibody, UEA-1 marker (donated by Dr. Clare Blackburn) that, respectively, decorate the cortical and medullary thymic epithelial cell (TEC) network. To observe nerve fibers we applied, separately, the antibodies with specificities to neurofilaments (ab1989), serotonin (ab66047), tyrosine-hydroxylase (TH, ab75875), synaptophysin (ab187259), beta-adrenergic receptors 1 (ab169523), and 2 (ab151727). Those reagents were applied to the specimens overnight at 4°C. Secondary antibodies, produced in goat and specific to Fc portion of specific immunoglobulin of each primary antibody conjugated to fluorescent protein Alexa Fluor 488 and/or Alexa Fluor 594 (Invitrogen) at 1:500 dilution plus DAPI. The sections were washed with PBST (×1 PBS with Tween 20 (1% v/v, Fisher Scientific) between each step and mounted in VectaShield (VectorLabs). Thymic pictures were taken to show both cortical and medullar thymic areas containing the representative protein expression of the entire tissue. All samples were examined in the National Institute of Science and Technology on Photonics Applied to Cell Biology (INFABIC) (State University of Campinas, Brazil) using a Zeiss LSM 780-NLO confocal on an Axio Observer Z.1 microscope (Carl Zeiss AG, Germany).

In order to localize catecholamines within the thymic lobules, we used the sucrose-potassium phosphate-glyoxylic acid (SPG) method as previously described [18]. Briefly, thymus frozen sections were submitted to reaction with SPG solution (10.2 g sucrose, 4.8 g potassium dihydrogen phosphate, and 1.5 g glyoxylic acid into 100 mL of distilled water), and subsequently dried in cool air for 10–15 min. When completely dried, the sections received Light USP Mineral Oil and were placed in a pre-warmed incubator set at 95°C for 2 min. The oil was replaced by a new drop and a cover slip was placed on top of the tissue for further analysis. The entire tissue was scanned and imaged in different magnifications to separate cortical and medullar thymic regions, and at least 3 sections per organ, in a total of up to 5 thymus per group, were examined using a Zeiss LSM 780-NLO confocal on an Axio Observer Z.1 microscope (Carl Zeiss AG, Germany).

For neuron morphology analysis, thymuses were removed as cited above and processed according to the protocol described before [19]. Briefly, thymus samples were incubated in Golgi-Cox solution (Stock solution: Potassium dichromate, Mercuric Chloride, and Potassium chromate diluted 5% w/v in distilled water), in the dark and at room temperature for 24 h; when it was moved to a new solution and kept in the same conditions for 10 days. The samples were then transferred from the Golgi Cox solution to a flask containing protectant solution (300 g sucrose, 10 g polyvinylpyrrolidone, 300 mL ethylene glycol into 500 mL 0.1 m phosphate buffer) for 24 h, in the dark, at 4°C. This step was repeated once. Finally, the sections were dehydrated in a protocol as follow: distilled water, 50% ethanol, 3:1 ammonia solution, distilled water, 5% sodium thiosulfate, distilled water, 70/95/100% ethanol, and xylol. After that, the samples were mounted in glycerol/mounting media. At least 3 sections per organ, in a total of up to 5 thymus per group, were examined and imaged using a Zeiss LSM 780-NLO confocal on an Axio Observer Z.1 microscope (Carl Zeiss AG, Germany) to demonstrate a representative portion of both cortical and medullar thymic areas.

After sacrificing the animals, the thymuses and spleen were removed and processed for characterization of both lymphoid and microenvironmental compartments by flow cytometry. Up to five different samples of thymocytes or spleen lymphocytes per group were acquired by meshing the organs, counted, and then diluted in PBS/2% BSA 2 mm EDTA and incubated with specific antibodies from BD Bioscience^®: for thymocyte subpopulation: rat anti-mouse CD25 (PE #553075), CD44 (fluorescein isothiocyanate [FITC] #553133), CD3 (PECy7 #552774), CD4 (PERCPCy5 #550954), and CD8 (APC #553035). For apoptosis analysis, the same cell suspensions were incubated with Annexin V binding buffer (556,454) plus 5 μL of Annexin-V (FITC #556419) and 10 μL Propidium Iodide (556463) after the incubation of CD3 (APC #553066), CD4 (PECy7 #563933), and CD8 (PERCPCy5.5 #567597. Another cell suspension was acquired to evaluate T regulatory cells, for that cell suspensions were stained with CD3 (APC #553066), CD4 (PECy7 #563933), and CD8 (PERCPCy5.5 #567597), washed and fixed with BD Pharmingen™ Mouse Foxp3 Fixation Buffer (51-9006124) for 30 min at 4°C in the dark, the fixative was removed by centrifugation and the pellet was suspended and incubated in 200 μL of freshly and warmed (37°C) BD Pharmingen™ Mouse Foxp3 Permeabilization Buffer (51-9006125) 30 min at 4°C in the dark. Cells were washed and suspended in BD Pharmingen™ Stain Buffer and incubated with Foxp3 for 30 min at room temperature, washed twice and suspended in stain buffer to proceed with flow cytometry. For BAR’s analysis, thymocytes received CD3 (APC #553066) and beta-adrenergic receptors 1 (ab169523) and 2 (ab151727) antibodies and incubated for 30 min at 4°C in the dark, washed and incubated with secondary antibody Goat Anti-Rabbit IgG (FITC) (ab6717) for 30 min at 4°C in the dark when was washed, suspended, and analyzed.

For TEC enrichment, thymuses were submitted to an enzymatic dissociation in a solution of 0.25 mg/mL Collagenase IV (Roche Life Science #11088858001) plus 0.1 mg/mL DNase I (Roche Life Science #4716728001) for 3 steps of 20 min at 37°C with rotation. The second and third cell suspensions were pooled together; the first was discarded, and submitted by a CD45⁺ depletion using CD45 MicroBeads (Miltenyi Biotec^® # 130-052-301). The negative cell suspensions were then incubated with anti-UEA-1 (donated by Dr. Clare Blackburn), Ly-51 (BD Bioscience^® #553160), EPCAM1 (ThermoFisher^® #25-5791-80), and MHC-2 (eBioscience^® 17-5321) diluted 1:100 in PBS 2% (w/v) BSA 2 mm EDTA. All analysis were performed at FACS Verse Cytometer (BD Biosciences, USA). The gating strategy for each flow cytometry analysis is presented as a supplementary figure (for all online suppl. material, see https://doi.org/10.1159/000535859). Fluorescence minus one sample was used to determine positivity for specific stain, regarding the detection of Foxp3, BAR’s, and Annexin-V.

To inject intrathymically Dil Stain (1,1“-Dioctadecyl-3,3,3',3'-Tetramethylindocarbocyanine Perchlorate) (ThermoFisher) or FITC, mice were anesthetized with 100–200 mg/kg body weight of ketamine and 5–16 mg/kg body weight for xylazine, injected intraperitoneally. A small transverse incision was then made in the skin, perpendicular to the sternum and the needle was positioned in the center of the first intercostal space with angle about 30° relative to the sternum. Twenty microliters of Dil (0.1 mg/mL) or FITC (1 mg/mL) were injected into one of the lobes of the thymus. The skin lesion was closed using surgical glue. Up to 5 mice were sacrificed at the peak of disease (MOG peak) being up to 7 or 4 days, respectively, after the intrathymic injection. The brain was collected to cryotomy at the midbrain region and once in cryosections, Dil stain was visualized with a Zeiss LSM 780-NLO confocal, or an Axio Observer Z.1 microscope (Carl Zeiss AG, Germany). Quantification of positive areas of at least 3 immunofluorescence images per mice was performed using the plug in “set measurement” on Image J software [20]. Spleens were collected and processed accordingly to flow cytometry analysis as described above.

Statistical analysis for MOG-onset/peak versus OVA samples from each experiment was compared with an Anova (when MOG onset group was present) or Student’s t test (when samples were control and MOG peak) to determine significance of the results, which were plotted on Prism 5 version 8.01 (GraphPad Software, Inc., La Jolla, CA, USA) with p < 0.05 considered statistically significant. Statistical values are expressed as mean ± SEM of at least 5 mice/group.

Since OVA-treated versus untreated C57BL6/J mice revealed no differences in all thymic parameters evaluated, results are presented comparing MOG versus OVA immunized animals.

We first evaluated the thymic innervation by analyzing the staining for nervous system, using the Golgi-Cox staining, which remains a major tool to detect neurons processes including spines [19]. The thymus from MOG immunized mice (at peak of disease) presented a complex alteration in the nerve fibers of the thymic microenvironment, which is filled by thicker nerve fibers through the capsular, subcapsular, and cortex areas. In contrast, thymuses from OVA treated mice presented less density and thinner fibers in those corresponding areas. Yet, in the medullary area, we observed that MOG thymuses presented thicker nerve fibers with varicosities, which were very rare in the thymus from OVA mice (Fig. 1a).

Overall, the Golgi-Cox staining revealed that at the peak of disease, MOG thymus has an alteration in the distribution pattern of nerve ramification throughout the thymic lobules, suggesting a deep commitment of the intrathymic innervation during the inflammation peak in EAE. We also found that neurofilament-containing network was present around thymic cells, in both medulla and cortex of thymic lobules. Again, there was a difference in the profile when comparing OVA-immunized with MOG-immunized mice: the nerve fibers were denser in the capsular, subcapsular, and central portion of the organ in MOG-treated animals (Fig. 1b), thus confirming the Golgi-Cox observations. Yet, staining for the catecholaminergic fibers revealed a higher density of those fibers throughout the thymic microenvironment of MOG-treated in both cortical and medullary areas, as compared to the OVA thymus (Fig. 1c). Overall, we showed that the distribution pattern of the thymic innervation from MOG-treated mice is altered. Accordingly, it seemed conceivable that the sympathetic nervous system function is also modified, based in the increase of nerve varicosities.

Since we observed various alterations of nerve fibers in MOG thymuses, we immunostained thymic sections for detection of TH as indicator of sympathetic activity. Accordingly, TH immunoreactivity in thymus can be interpreted as a catecholaminergic marker [21]. The TH positivity in MOG peak thymus was larger in distribution and reactivity through the thymic microenvironment compared to OVA thymuses (Fig. 2). Such an increase suggests (i) a compensatory increase in the dopamine signaling between thymic cells and (ii) a decompensated TH activity due the disturbance in the neurofilaments.

In addition, we observed an increase in the beta-adrenergic receptor 1b (BAR’s 1) expression in MOG peak thymus, with a denser concentration in the thymic capsular and cortical microenvironmental areas (Fig. 3a1), followed by a reduction in the percentage of BAR’s 1⁺ CD3⁺ thymocytes compared to the profile seen in the OVA thymus (Fig. 3a2). The expression of the beta-adrenergic receptor 2b (BAR’s 2) was reduced in the thymic microenvironment of MOG peak mice (Fig. 3b), whereas the percentage of BAR’s 2⁺ CD3⁺ thymocytes was increased compared to OVA thymus (Fig. 3a2). The alterations seen in terms of BAR’s 1 and 2 expression supports the hypothesis that EAE development to peak phase compromises the intrathymic sympathetic nervous component, not only by changing the localization and morphology of the nerve fibers processes but also interfering in the dopaminergic production and reuptake by thymocytes.

We further evaluated serotonin expression that revealed an increased and sparse labeling in MOG peak thymus presenting a much denser staining, when compared to OVA thymus (Fig. 4a). Moreover, synaptophysin immunostaining revealed the presence of large varicosities, and increase in the positive staining in the thymic microenvironment from MOG mice, whereas the OVA thymus presented lesser and restrict positive staining (Fig. 4b).

Based on the disturbance in the thymic innervation of MOG peak thymus and in the recently reported thymic atrophy during the EAE model [22], we hypothesized that these changes would also be compromising the thymic microenvironmental compartment, eliciting a dysfunctional T-cell development along with the clinical evolution of EAE.

To evaluate thymic alterations during EAE evolution, we immunized C57BL6/J mice with MOG_35-55/CFA and quantified thymic relative weight during onset (10 d.a.i +/− 1) and in the peak (14 d.a.i +/− 1) of the disease. Our results demonstrated that thymuses from MOG animals presented a significant reduction of the thymus in comparison with OVA-immunized animals (Fig. 5a–e). In the mouse model (EAE mice), we found that the thymic involution was already present at the disease onset; being even more prominent at the peak of the disease compared both with 0 and onset time points (Fig. 5b). Moreover, as seen in Figure 5c, mice with clinical signs of EAE (score >0) presented significantly lower thymic relative weight, in comparison with immunized animals without clinical signs (score = 0), suggesting that the thymic involution is correlated with the course of the neuroinflammation occurring in these animals.

We further evaluated the same parameters during the evolution of EAE induced in Lewis rats (EAE rats). Differently from EAE in mice, EAE in rats is a classical monophasic and spontaneous limited disease. In addition, EAE in Lewis rat is induced by a different autoantigen (MBP) [14, 15]. We used this model to exclude a possible role of the clinical course, the specific autoantigen, or the animal strain in the thymus events seen in mice. Again, our results demonstrated a significant involution of the thymus in monophasic EAE (Fig. 5d, e). Histological studies conducted in mice confirmed that the thymic atrophy was accompanied by morphological changes, characterized by the loss of cortex-medullary boundaries (Fig. 5f). This was followed by changes in the pattern of thymocyte subpopulations: as a disease stage-dependent manner, the relative numbers of CD4⁻CD8⁻, CD4⁻CD8⁺, and CD4⁺CD8^- thymocyte subsets were increased, whereas the CD4⁺CD8⁺ subpopulation was decreased in the MOG groups compared to OVA (Fig. 5g). Importantly, when we evaluated the absolute cell numbers of the various thymocyte subsets, we found a significant reduction in all subsets in MOG thymuses, when compared to OVA counterparts (Fig. 5h), thus correlating with the thymic atrophy and consequent thymocyte subpopulations loss.

We then evaluated the progression of early thymocyte differentiation within the CD4⁻CD8⁻ (DN) compartment. This stage can be further subdivided in at least four subsets, based on the differential expression of CD25 and CD44, being: DN1 (CD25⁺CD44⁻), DN2 (CD25⁺CD44⁺), DN3 (CD25⁻CD44⁺), and DN4 (CD25⁻CD44⁻); the last evolving to CD4⁺CD8⁺ (DP) thymocytes. We noticed that in the MOG group, accompanying neurofilament increase during thymic atrophy, there was a gradual reduction in the relative numbers of DN1, DN2, and DN3 subsets and an increase of total DN numbers, from the onset to the peak of symptoms. By contrast, the relative numbers of DN4 cells did remain similar in both MOG onset and MOG peak groups (Fig. 5i). Nonetheless, when the absolute numbers of each DN subpopulation were evaluated, we found in the MOG group an increase in DN3 thymocytes compared to OVA, and a reduction in DN1, DN2 and DN4 subpopulations (Fig. 5j). It seems conceivable that developing thymocytes in the EAE group have an accelerated differentiation from DN1 to DN3/4 stage, but then slowed down to progress properly toward the DP differentiation stage, contributing to the observed DP decrease during the EAE progress.

We further investigated whether the alterations seen in thymocyte subpopulations could be due an increase of thymocyte death (Fig. 5k). In fact, thymuses from MOG mice at the peak of disease presented higher percentages of Annexin-V⁺ thymocytes in all subpopulations; whereas the numbers of cells Annexin-V⁺ cells were reduced, accompanying the reduced number of total thymocytes on MOG-peak mouse thymuses.

We also observed a higher percentage and a reduced total number of natural T regulatory cells (nTreg/CD3⁺CD4⁺CD8⁻CD25⁺Foxp3⁺) in thymus from MOG peak mice, compared to OVA controls (Fig. 5l). In the periphery, both relative and absolute number of T regulatory cells (Treg/CD3⁺CD4⁺CD8⁻Foxp3⁺) was reduced in the spleen of MOG peak mice, as compared to OVA treated animals (Fig. 5m). Altogether, the results presented in Figure 5 demonstrate that the imbalance of developing thymocytes with reduction of regulatory T cells in the organ and in the periphery of the immune system correlates with the neuroinflammation present in the CNS of EAE mice.

Since the thymic microenvironment controls intrathymic T-cell development and mature T-cell export, and considering the thymocyte depletion seen in EAE animals, we hypothesized that disturbances in the thymic microenvironmental cell compartment could also occur. We approached this issue by analyzing morphology and cellularity of TEC subpopulations. We found alterations in TEC network through specific labeling that discriminate between cortical (Ly-51⁺) and medullary (UEA-1⁺) TEC subsets. Control samples presented TEC distributed as previous demonstrated in the literature (Fig. 6a). Differently, thymuses from MOG mice presented a tiny area positive to medullary TEC stain and higher density of cortical TEC staining (Fig. 6b).

We further investigated whether the microenvironmental alterations observed in MOG thymuses could be somehow related to the exit of immature thymocytes to the periphery of the immune system. To test this hypothesis, we injected FITC intrathymically, just before the animals presented EAE clinical signs (day 9 post-MOG inoculation) and evaluated the presence of FITC⁺ DN and FITC⁺DP T cells 4 days later (MOG peak) (Fig. 6c). When we analyzed the relative and absolute numbers of FITC⁺ T cells in the spleen, we observed that MOG-treated mice revealed an increase in absolute numbers of FITC⁺ T cells within CD4/CD8-defined T lymphocyte subsets (Fig. 6c). The results revealed that the neuroinflammation conditions coincides with a transformation on the thymic microenvironment, which in turn contributes to alterations of thymocyte differentiation, affecting the DN to DP transition and including a premature T-cell export from the organ.

Lastly, to verify whether the intrathymic innervation functionally maintains the retrograde signaling to the CNS, we injected a retrograde tracer into the MOG-onset thymus and analyzed the re-capitation in the CNS (Fig. 7a). Interestingly, MOG mouse brain reuptake of the thymus-injected tracer was decreased, as compared to OVA treated group (Fig. 7b). This result demonstrates that, in addition to the intrathymic neuro-compartment derangement, there is an impairment in terms of signaling back to CNS, suggesting a continuous loop of dysfunction in the CNS-thymus bidirectional signaling paralleling the neuroinflammation.

The CNS-immune cross talk can be seen at work through different ways: (i) in one hand the hypothalamic-pituitary-adrenal axis and (ii) the sympathetic fibers of the autonomic nervous system [23], and on the other hand through lymphocyte circulation and cytokine production by immune cell within CNS [24]. From the brain, nerve fibers arise, innervating primary and secondary lymphoid organs during the embryonic stage; the sympathetic fibers remain during the whole life. In this respect, the CNS input is essential in the development and maintenance of the immune system [1, 2, 25]; the thymus being included. It seems thus conceivable that disturbances in the CNS, as neuroinflammation, seen in several neurodegenerative diseases, could affect the immune axis of neuroimmune biological loops.

Since EAE is an inflammatory process of the CNS, it is plausible to propose that the neuroinflammation evoked by the autoagressive lesion commits the intrathymic sympathetic nervous system signaling. Herein, we showed that the adrenergic innervation in the thymus from MOG-immunized mice is compromised and deficient in maintaining the correct communication to CNS during the peak of the disease, when we observed an increase in the density of sympathetic nerve fibers with varicosities followed by alteration in the pattern of expression of neurotransmitters and corresponding receptors. Such disturbance, correlated with the progressive thymic atrophy, indicates that the thymic shrinkage seen during the progressive CNS inflammation is followed by a rearrangement in the intrathymic sympathetic fibers meshwork, with increase in neuro fibers and neurotransmitter release density. Because the neurotransmitters also present a paracrine function, changes in the lymphoid and no-lymphoid intrathymic cell number during the organ atrophy alters the availability of neurotransmitters such as noradrenaline [26], evoking alteration also in secondary lymphoid organs; as shown by the reduction in the number of splenic T regulatory cells.

It has been previously reported that during EAE peak and chronic stages, there is a thymic atrophy with lymphocyte loss and increase in the percentage of natural regulatory T cells [22, 27, 28]. We showed that the thymic disturbance starts together with the first clinical symptoms of EAE and that thymocyte loss positively correlates with the increase of EAE clinical signs, e.g., disease score. These results reinforce the notion that the inflammatory process of the CNS directly affects the thymic sympathetic innervation and, as consequence, thymocyte maturation ratio. Yet, SNS plays a role in the course of EAE as it acts enhancing the pro-inflammatory Th response at the EAE earlier stages, damping Treg (shown here), and them increase the anti-inflammatory process of a chronic stage by elevating the percentage of Treg [29]. We also showed herein that early stages of thymocyte differentiation are compromised during the EAE development, seen by alterations at the DN subsets. An altered thymic microenvironment is not capable to attract new thymocyte progenitors (ETP) [30], so, to contribute to the correct maturation flow, evoking an impairment at the DN3/4 transition into the DP stage. This DP subpopulation in MOG-immunized mice decrease is also associated with increased apoptosis, and early exportation of both DP and DN immature T cells to the spleen. Alteration of immature thymocyte development together with DP thymocyte exportation to the periphery of the immune system was shown during thymic atrophy seen in acute infectious diseases [5] but, to our knowledge, this is the first time that those events are linked to intrathymic neuroimmune interaction changes.

We also observed a decrease of the total number of regulatory T cells in both thymus and spleen. Catecholamines are known to reduce the expression of IL-10 and TGF-β by Tregs, thus stimulating a pro-inflammatory response led by effector T lymphocytes [31]. In this respect, the SNS and microenvironmental alterations observed during the EAE peak correlates with an alteration of intrathymic T-cell development and export of regulatory T cells, as well as abnormal export of immature thymocytes, potentially contributing to the development of autoimmune reactions, as for example, EAE progression. Of note, the changes observed in the distribution pattern of cortical and medullary TECs (seen at the MOG-peak thymus) correlates with nerve fiber density and neurotransmitter alterations. Because intrathymic nerve fibers are mainly present in the thymic cortical area [2], and we observed a disturbance in DN subpopulation on MOG peak thymus, we suggest that the rearrangement of the cortical/medullary intrathymic compartments was related to the nerve fibers changes, which evoked an alteration in the early development process of thymocytes.

The data presented here indicate that the beginning of neuroinflammation on the CNS evoked by MOG injection, changes the pattern of sympathetic nerve fiber distribution and communication, which in turn may contribute for the disturbance seen in the thymic microenvironmental and lymphoid compartments, resulting in atrophy of the organ. An increase in TH⁺ cells was also observed in thymuses from mice under the onset of arthritis model, followed by an increase during the chronic phase [32]. Nevertheless, while in the arthritis model TH increase was accompanied by loss of sympathetic nerve fibers, we show herein that during EAE peak phase occurs an increase of nerve fibers density concomitant with intrathymic TH⁺ rise.

Lastly, at the peak of disease, the intrathymic injection of tracers for retrograde experiments revealed that fewer areas in the brain from MOG-immunized mice were positive to the NiI dye in comparison with OVA-immunized controls, indicating defective signaling coming from the thymus. Accordingly, impaired cross-talk between the immune system and the CNS seems to be the major trigger for the thymus involution and malfunction. In turn, the loss of thymus functionality may influence directly the control of the autoagressive response. Sympathetic nervous system dysfunction present in individuals with schizophrenia also seems to change the peripheral immune system through neuro-immune interactions, possibly due to a dysfunctional immune tolerance [33].

Altogether, our results demonstrate that the brain during an inflammatory autoimmune reaction can evoke intrathymic neuro-alterations, which compromises the correct development of intrathymic and maturation of peripheral T cells. In turn, thymus fails to properly communicate with the nervous system, contributing for the disease progression, as summarized in Figure 8.

The aim of this study was to mainly characterize the intrathymic nerve fibers in the thymus from mice at onset and peak phase of EAE, thymic conditions during chronic phase were not evaluated in this work. Importantly, the disease clinical evolution correlation with thymic atrophy raises new questions to be further investigated, with new functional experiments to prove the correlation between SNS and the changes in the thymus and spleen subpopulations. We demonstrate that they are occurring at the same time, as it is shown by the literature, indicating that the SNS does play a role in T-cell development and maturation during the inflammatory process of EAE.

We appreciate the valuable technical assistance with the animals provided by Mr. Marcos C. Meneghetti. We are indebted to National Institute of Science and Technology on Photonics Applied to Cell Biology (INFABIC) at the State University of Campinas.

Animals were kept in the maintenance room of the animal house in the Institute of Biology, University of Campinas (Campinas, Brazil), under species pathogen-free conditions and all experimental protocols were approved by the Institutional Ethical Committees (CIBIO #1/2013, CEUA #3889-1; #4393-1; #3626-1, #3754-1 University of Campinas).

The authors declare no conflict of interest to disclose.

The authors are pleased to acknowledge financial support by the state-granting agency (FAPESP; #15/10107-2, 2019/06372-3 and19/16,116-4). WS is funded by CNPq, CAPES, and FAPERJ (Brazil) and the MercoSur Fund for Structural Convergence (FOCEM). Participation of WS is within the framework of the Brazilian National Institute of Science and Technology on Neuroimmunomodulation, the Rio de Janeiro Research Network on Neuroinflammation, and the INOVA-IOC neuroimmunomodulation network in the Oswaldo Cruz Foundation (Brazil).

Carolina Francelin carried out the experiments. Carolina Francelin, Alexandre Borin, Jessica Funari, and Fernando Pradella contributed to sample preparation and analysis. Carolina Francelin, Leonilda M.B. Santos, Wilson Savino, and Alessandro S. Farias contributed to the interpretation of the results. Carolina Francelin, Wilson Savino, and Alessandro S. Farias took the lead in writing the manuscript. All authors reviewed and provided critical feedback to the final version of this manuscript.

Present address of C.F.: School of Medicine - Department of Ophthalmology and Visual Science, the University of Alabama at Birmingham, Birmingham, AL, USA.

We declare hereby that all data acquired and analyzed during the period of the current study are included in this article and more detailed data can be provided by the correspondent author upon reasonable request.