Aging and neuroinflammation: Changes in immune cell responses, axon integrity, and motor function in a viral model of progressive multiple sclerosis

Abstract Although aggravated multiple sclerosis (MS) disability has been reported in aged patients, the aging impact on immune cells remodeling within the CNS is not well understood. Here, we investigated the influence of aging on immune cells and the neuroinflammatory and neurodegenerative processes that occur in a well‐established viral model of progressive MS. We found an anomalous presence of CD4+ T, CD8+T, B cells, and cells of myeloid lineage in the CNS of old sham mice whereas a blunted cellular innate and adaptive immune response was observed in Theiler's murine encephalomyelitis virus (TMEV) infected old mice. Microglia and macrophages show opposite CNS viral responses regarding cell counts in the old mice. Furthermore, enhanced expression of Programmed Death‐ligand 1 (PD‐L1) was found in microglia isolated from old TMEV‐infected mice and not in isolated CNS macrophages. Immunocytochemical staining of microglial cells confirms the above differences between young and old mice. Age‐related axonal loss integrity in the mouse spinal cord was found in TMEV mice, but a less marked neurodegenerative process was present in old sham mice compared with young sham mice. TMEV and sham old mice also display alterations in innate and adaptive immunity in the spleen compared to the young mice. Our study supports the need of new or adapted pharmacological strategies for MS elderly patients.

Indeed, premature immune aging has been described in some MS patients, perhaps underlying the transition from the RR form to a progressive disease course (Thewissen et al., 2005). In this context, it is also significant that most MS patients are diagnosed between 20 and 40 years of age, and only a few individuals are diagnosed after they have reached 50 years of age (Scalfari et al., 2014). In recent years, interest in understanding the impact of aging on the course of MS has emerged in an attempt to find novel therapies or to adapt current treatments to the chronological age of patients (Vaughn et al., 2019).
Key age-related changes in the CNS are triggered by microglia, yet a diminished efficiency of the adaptive and innate immune cell responses may be contributing to the pathological progression of the disease in elderly MS patients (Bektas et al., 2017;Musella et al., 2018;Rawji et al., 2016). Exaggerated immunosenescence is a concept that has been employed in MS trying to explain the accelerated changes in remodeling immune response during aging (Mandolesi et al., 2015). Although it has been postulated that the lesions that develop in aging patients may do so on a pre-injured CNS, it is not clear whether this may also be the case in experimental models (Musella et al., 2018;Wimmer et al., 2019).
Most experimental studies on aging impact in MS have been performed by modeling the RRMS course following immunization with myelin proteins (Bolton & Smith, 2018), yet it remains unknown what happens in experimental progressive MS. One of the most applicable models to study how age affects the course of progressive MS is the Theiler's murine encephalomyelitis virus-induced demyelinating disease (TMEV-IDD) mode as this is the reference model to investigate viral-mediated mechanisms of primary progressive MS (Olson et al., 2001). Intracerebral infection with TMEV in susceptible mice induces a chronic progressive neurodegenerative disease that is clinically and histopathologically similar to human progressive MS (Denic et al., 2011;Lipton & Dal Canto, 1976). Here, we focused on CNS and spleen cellular immune responses during the neurodegenerative phase of the disease, by assessing axonal damage and motor disability, both well reproduced in the TMEV model and that potentially reflect the pathological changes in the progressive phase of MS. Hence, the aim of our study was to investigate the impact of aging on neuroinflammation particularly, in microglia and macrophage cells that play a critical role in the pathogenesis of MS.
Age-related alterations in the expression of the immune checkpoint PDL-1 by microglia and macrophages together with the inability to mount an adequate cell innate and adaptive responses during TMEV infection may be affecting the neurodegeneration and disability in TMEV-IDD.

| Aging diminishes CD4 + and CD8 + T-cell immune responses in the CNS of TMEV-infected mice
The course of TMEV-IDD infection in SJL/J mice indicates that by 12 wpi, animals develop motor deficits and CNS neuroimmune alterations as we previously described (Mecha et al., 2013).
However, it is unknown how aging affects the adaptive immune responses of TMEV-infected mice. In order to investigate this, we infected 8-week-old female mice with TMEV ( Figure 1a) and 12 or 40 weeks later, we analyzed the T-cell populations following the gating strategy shown in Figure 1b. Note that the mice of the SJL/J strain experience physiological age-related events before other mouse strains like C57BL/6 or BALB (Murray et al., 1998), with SJL/J mice having a shorter lifespan. TMEV infection increased the frequency of CD4 + T cells in the CNS at 12 (p < 0.05) but not at 40 wpi ( Figure 1c). This surprising result is partially due to the fact that aging increases the proportion of CD4 + T cells (p < 0.001 vs. 12 wpi Sham and p < 0.01 vs. 12 wpi TMEV; Figure 1c), being significantly higher in sham mice than in TMEV-infected mice (p < 0.01 vs. Sham; Figure 1d). Therefore, the relative change in the proportion of CD4 + T cells induced by TMEV was significantly different between young and old animals (p < 0.05 vs 12 wpi TMEV; Figure 1e). Aging significantly diminished the naïve and central memory T cells, while there were more effector memory CD4 + T cells in old sham mice (p < 0.05 vs 12 young sham mice;  Figure 1n). However, in the case of old mice, TMEV infection did not significantly modify these F I G U R E 1 Aging impairs the CD4 + and CD8 + T-cell immune responses in the CNS of Theiler's murine encephalomyelitis virus (TMEV)infected mice. (a) Scheme of the experimental groups, where 8-week-old mice were infected with 2 × 10 6 pfu of Theiler's virus and followed for 12 or 40 weeks. (b) Representative FACs plots of the gating strategy used to identify the different T-cell populations. (c) Quantification of the CD3 + CD4 + T cells in CNS. (d, e). Fold change of relative percentage of CD4 + T cells induced by aging and TMEV, respectively (f-h) Quantification of the percentage of naïve, central memory and effector memory CD4 + T cells, together with the change induced by aging (i, k, m) or TMEV infection (j, l, n), respectively. (o) Quantification of CD3 + CD8 + T cells in CNS. (p, q) Fold change of the relative percentage of CD3 + CD8 + T cells induced by aging or TMEV infection, respectively. The data are presented as the mean ±SEM: *p < 0.05 vs. sham at the same age; **p < 0.01 vs. sham at the same age; # p < 0.05 vs. the same experimental group at 12 wpi; ## p < 0.01 vs. the same experimental group at 12 wpi; ### p < 0.001 vs. the same experimental group at 12 wpi; + p < 0.05 vs. 12 wpi; ++ p < 0.01 vs. 12 wpi; $ p < 0.05 vs.sham; $$ p < 0.01 vs. sham; $$$ p < 0.001 vs. sham (n = 6 sham 12 wpi; n = 6 TMEV 12 wpi; n = 5 sham 40 wpi; n = 5 TMEV 40 wpi) immune populations, except for a tendency to increase the population of naïve cells (Figure 1j,l,n). Collectively, our results not only demonstrate the anomalous presence of CD4 + T cells in old sham mice but also show a deficient age-related CD4 + T cells response in the CNS of mice subjected to TMEV infection as a model of progressive MS. CD8 + T cells showed a similar profile to CD4 + T cells. As such, CD8 + T cells were increased in response to TMEV in the young mice only (p < 0.01 vs. 12 wpi sham: Figure 1o). Again, old sham mice have an elevated frequency of CD8 + T cells (p < 0.001 vs. 12 wpi sham: Figure 1o,p); however, TMEV-infected mice did not show this increase (Figure 1p). Notably, there was a weaker relative fold change in the CD8 + T-cell response in the CNS of TMEV 40 wpi mice (p < 0.01 vs. TMEV 12 wpi: Figure 1q), which may lead to a deficient control of inflammation in the aged brain. As such, the age of the mice clearly affects the CNS cellular adaptive responses in TMEV-IDD mice.

| B-cell responses to long-term TMEV infection are limited in aging mice
The involvement of the B-cell population in MS pathogenesis is undeniable (Sospedra, 2018) and clear shifts in B cells in the elderly have been described, suggesting that age-related B-cell changes contribute to immunosenescence (Hao et al., 2011).

| Aging impact microglia and macrophages in an opposite way in the CNS of TMEV-infected mice
As CNS sensors microglia play a crucial role in health and in pathological conditions, including MS (Colonna & Butovsky, 2017). Microglial activity depends on different variables, which include age, neuropathological status, disease stage, and environmental factors (Wolf et al., 2017). In this study, we assessed the percentage of homeostatic microglia by FACs, labeling them for the markers CD45 and P2RY12 (see gating strategy in Figure 3a). Because the expression of P2RY12 can be modified by the activation state of microglia and then could be confused as macrophages, we also followed the gating strategy that try to distinguish microglia and macrophages by CD45 and CD11b markers (the population with lower CD45 expression corresponds to microglia and the population with higher CD45 expression refers to macrophages). The results observed gating on CD45 + P2RY12 + cells and CD45 low CD11b + cells were similar (data not shown). The proportion of homeostatic microglia increased 12 weeks after TMEV infection compared to that found in sham mice (p < 0.05: Figure 3b). By contrast, the proportion of CD45 + P2RY12 + cells dropped in old TMEV 40 wpi mice (p < 0.01 vs. 12wpi TMEV: Figure 3b). Importantly, old sham mice display a proportion of microglia similar to that observed in young TMEVinfected mice, suggesting the existence of a basal age-related CNS inflammation in the older sham mice at least, in terms of the microglial cells count. This scenario leads to a smaller relative fold change in microglial frequency induced by TMEV in the older mice (p < 0.05 vs 12 wpi TMEV mice: Figure 3B). More interestingly, the proportion of microglia expressing the immune inhibitory checkpoint protein PDL-1 (CD274) was augmented in both young and old TMEV mice in comparison with their corresponding sham mice (p < 0.05 and p < 0.01, respectively:

| Age-related axonal damage in the mouse spinal cord
As axon damage in the spinal cord is a well-established hallmark of TMEV-IDD (Ure & Rodriguez, 2000), next we evaluated the integrity of axons in transverse sections of the cervical spinal cord of young and old mice by immunolabeling with the NF-H antibody ( Figure 4a). In agreement with our previous work (Feliú et al., 2017), 12 weeks after intracranial viral inoculation there is weaker NF-H immunostaining in the ventral spinal cord at cervical level, which reflects increased axonal damage (p < 0.001; Figure 4b). The loss of axon integrity was higher in old TMEV mice (p < 0.05 vs. young TMEV Figure 4b). Surprisingly, this weaker NF-H labeling was also found in the spinal cord of old sham mice (p < 0.001 vs. young sham), although it was more marked in their counterpart mice infected with the virus.

| Motor disability in young and old mice infected with Theiler's virus
Given the changes in innate and adaptive immunity between young (12 wpi) and elderly (40 wpi) TMEV-infected mice, we further as- Representative flow cytometry plots of the gating strategy. (b) Quantification of the percentage of B220 + CD19 + B cells in the CNS; (c, d) Fold change of relative percentage of B220 + CD19 + cells aging-or TMEV-induced, respectively (e) Quantification of the proportion of CD5 + CD1d high cells gated on B220 + CD19 + B cells. (f, g) Fold change of relative percentage of CD5 + CD1d high cells induced by aging or TMEV, respectively The data are expressed as the mean ± SEM (n = 5-6 mice for each group): **p < 0.01 vs. sham at the same age; ***p < 0.001 vs. sham at the same age; ### p < 0.001 vs. the same experimental group at 12 wpi; $ p < 0.05 vs. sham (n = 6 sham 12 wpi; n = 6 TMEV 12 wpi; n = 5 sham 40 wpi; n = 5 TMEV 40 wpi) +p < 0.05 vs. 12 wpi and +++p<0.001 vs. 12wpi Although habituation to the cage (lack of novelty) may influence the results obtained, we cannot rule out that sham mice become less active as they age at least in deambulatory activity (p < 0.05 one trial 20 weeks old vs. 48 weeks old, Figure 5D).

| Aging increases Treg cells and effector memory CD4 + and CD8 + T cells in the spleen of TMEV-infected mice
The peripheral immune cell changes were modest in the TMEV-IDD model in accordance with our previous studies . In the spleen, there were no changes in the percentage of CD4 + T cells between sham and TMEV mice irrespective of their age ( Figure 6b). However, there was an age-related increase in the percentage of Tregs in both old sham and TMEV mice (p < 0.001 vs. 12 wpi sham or 12 wpi TMEV mice: Figure 6b). An increase in the levels of Tregs suggests that old SJL/J mice may exhibit an exacerbated suppression of the peripheral inflammatory responses compromising CD4 + T cells response to pathogens. Furthermore, the percentage of naïve CD4 + T cells was reduced in sham and TMEV old mice in comparison with their corresponding younger mice (p < 0.001 vs. 12 wpi sham or 12 wpi TMEV mice: Figure 6a).

| DISCUSS ION
Although Nevertheless, we cannot infer the relevance of Bregs reduction in the specific neurodegenerative process of TMEV-IDD as it was also observed under non-pathological condition (old sham mice).
The neuroinflammatory component of MS and its animal models preferentially involves microglia/macrophage cells associated with cytotoxic mechanisms, but also with endogenous tissue repair (Liu et al., 2018;Mecha et al., 2016). Although the purinergic receptor P2RY12 labels homeostatic microglia as its expression is downregulated when microglia is activated (Zrzavy et al., 2017) several studies suggest that P2RY12 also identified other populations of microglia not only homeostatic microglia (Walker et al., 2020

Habituation effect (d)
phenotype of aged microglia labeled with Iba-1, the fact of observing fewer cells in TMEV old mice maintaining the same percentage of Iba-1 occupied area would suggest a microglia shift toward a big cell body and thick branches, but we cannot discard that macrophages can contribute to the Iba-1 occupied area as this marker also recognizes macrophages. This shift in microglia may unveil their diminished capacity to mount a normal response to the axon loss that took place in TMEV-IDD potentially contributing to an accelerated neurodegenerative process. Reduced migration and phagocytosis ability are characteristics of aged microglia that in the context of MS delays remyelination as a consequence of the impaired clearance of myelin debris (Neumann et al., 2009). We guess that microglia and in general, myeloid cells reprograming is context and time dependent and so, a strict definition of microglia markers may result incorrect.
In the present study, we observed that macrophages behave distinctly to microglia in the aging CNS. First, contrary to microglia macrophage frequency did not differ between young and old sham or TMEV mice. Second, the immune checkpoint PDL-1 expressed by microglia and macrophages show opposite responses. Although the potential contribution of increased microglia expressing PDL-1 to the age-associated neurodegeneration in TMEV-IDD remains unclear, our findings reinforce the interest of PDL-1 in the CNS compartmentalized immune responses in our model and in progressive MS (Bar-Or & Antel, 2016). A previous study showed induced expression of PDL-1 by microglia and macrophages within the CNS, yet not in the periphery shortly after TMEV inoculation (Jin et al., 2013). However, we found increased PDL-1 expression in spleen macrophages of old control and TMEV mice that is, long time after virus administration. The finding of diminished spleen proportion of Ly6C high CD45 + CD11b + monocytes in both control and TMEV mice may uncover a reduced ability of innate immune response against pathogens in the elderly. Besides, our ex vivo functional studies show that spleen purified macrophages from TMEV old mice produce less IL-1β in response to LPS than purified young cells in agreement with F I G U R E 6 Changes in the peripheral T-cell response associated with aging. A spleen cell suspension was stained to analyze the different CD4 + and CD8 + T-cell subpopulations. (a) Quantification of the percentage of CD3 + CD4 + T cells along with the subpopulations of Treg (Foxp3 + CD25 + cells), Naïve (CD44 low CD62L high ), central memory (CD44 high CD62L high ), and effector memory (CD44 high CD62L low ) CD4 + T cells. (b) Quantification of the proportion of CD3 + CD8 + T cells along with the subpopulations of Naïve (CD44 low CD62L high ), central memory (CD44 high CD62L high ), and effector memory (CD44 high CD62L low ) CD8 + T cells. The data are expressed as the mean ± SEM:*p < 0.05 vs. sham at the same age; #p < 0.05 vs. the same experimental group at 12 wpi; ##p < 0.01 vs. the same experimental group at 12 wpi; ###p < 0.001 vs. the same experimental group at 12 wpi (n = 6 sham 12 wpi; n = 10 TMEV 12 wpi; n = 7 sham 40 wpi; n = 7 TMEV 40 wpi). TMEV, Theiler's murine encephalomyelitis virus that observed in microglial cells (Njie et al., 2012). GM-CSF is a cytokine that exists at low levels at steady state but increases drastically during infection and inflammation (Shiomi & Usui, 2015). Then, it is difficult to explain the increased GM-CSF production only by ex vivo splenocytes from old sham mice stimulated with PMA/ionomycin. It is possible that GM-CSF may be contributing to the increased migration of leukocytes into the CNS that we found elevated in old control mice. Our study has unveiled the impact of aging in cell immune  (McCarthy et al., 2012). Susceptible SJL/L mice lacked an innate immune population, the natural killer dendritic cell, which play a critical role in early CNS virus clearance as occurs in resistent strain of mice (Chastain et al., 2015). The mice were inoculated in the right brain F I G U R E 7 The peripheral innate immune response was dampened by aging. (a) Quantification of the percentage of CD45 + CD11b + cells in the spleen. (b, c) Quantification of the percentage of Ly6c high or CD274 + cells, respectively, gated on CD45 + CD11b + monocyte cells. (d) CD11b + monocyte cells isolated using Miltenyi beads were stimulated with LPS and IL-1β production was measured by ELISA. (e) Splenocyte cell suspension was stimulated as indicated in the Materials and Methods, and GM-CSF production was measured by ELISA on the supernatant. The data are expressed as the mean ± SEM (n = 5 mice for each group): *p < 0.05 vs. sham at the same age; **p < 0.01 vs. sham at the same age; ***p < 0.001 vs. sham at the same age; # p < 0.05 vs. the same experimental group at 12 wpi; ## p < 0.01 vs. the same experimental group at 12 wpi; ### p < 0.001 vs. the same experimental group at 12 wpi hemisphere with 2 × 10 6 plaque-forming units (pfu) of the Daniel (DA) strain of TMEV, a picornavirus positive-sense single-stranded RNA, diluted in 30 μl DMEM supplemented with 10% of FCS, as described previously (Arévalo-Martín et al., 2003). Sham mice received 30 μl of DMEM supplemented with 10% of FCS. Two time lines of experimental groups were defined: (i) 8-week-old infected mice sacrificed 12 weeks post-infection (wpi: sham n = 7, TMEV n = 7) and (ii) 8-week-old infected mice sacrificed 40 wpi (sham n = 11, TMEV=15).
Note that the SJL/J strain of mice undergoes physiological aging earlier than other strains of mice, such as C57BL/6 or BALBc (Liu et al., 2014). Specifically, SJL/J mice have a shorter lifespan, with males living 16 months on average and females only 14 months. General health conditions (weight and clinical score) were evaluated periodically every week from 10th to 40th week pi. Clinical scores (Murray et al., 1998) were assigned based on the general appearance and activity of the mice: score 1, mice with a waddling gait; Score 2, mice adopting a more severe waddling gait; score 3, mice that had lost their righting ability, in conjunction with hind limb spasticity; score 4, mice suffering paralysis of their hind limbs; and score 5, mice that were moribund.

| Evaluation of motor function
Theiler's murine encephalomyelitis virus-induced demyelinating disease is a model of chronic progressive MS in which virus inoculation is followed by a latent period of approximately 12 weeks until the clinical signs and motor deficits appear. The screening of locomotor activity was performed in an activity cage coupled to a Digiscan Analyser (Activity Monitor System; Omnitech Electronics Inc.).
Horizontal (HACTV) and Vertical (VACTV) activity were evaluated through the total number of horizontal and vertical sensor beam interruptions in a 10 min session. HACTV and VACTV activities were measured once a week after infection (wpi) from 10 to 40 wpi.

| Sample collection
The mice were anesthetized with pentobarbital (Doletal, 50 mg/kg body weight), and their spleen was removed, kept in cold Roswell Park Memorial Institute (RPMI), and processed for flow cytometry.
After transcardial perfusion with saline (0.9% NaCl), the brain and spinal cord were removed and processed for flow cytometry as indicated below. A little segment of the cervical spinal cord from at least 3 mice per experimental group was reserved for immunohistochemical analysis.  Mowiol. Based on our previous experience with the antibodies used in the present study, the specificity of the staining was confirmed by omitting the primary antibody and not by a matching isotype control.

| Microscopy and image analysis
Immunofluorescence images were acquired on a Leica TCS SP5 confocal microscope and immunohistochemical staining was assessed with a Zeiss Axiocam high-resolution digital color camera. Individual optical sections were examined by analyzing five to six sections (ventral spinal cord area) from at least three to five animals per group.
Staining was quantified using Image J software (NIH), maintaining the threshold intensity constant to compare the experimental and control images obtained within the experiments.

| Monocyte and Splenocyte ex vivo experiments
Spleen monocytes were isolated using a commercial cell isolation kit following the manufacturer's instructions (Miltenyi; Biotec Inc.).
Cells from each animal were plated (10 6 cells/well) and stimulated for 5 h with LPS (1 μg/ml). The cell medium was then collected, centrifuged, and the supernatant frozen. Splenocytes from each mice were plated (10 6 cells/well) and cultured in the medium RPMI supplemented with immobilized hamster anti-mouse CD3ε (10 μg/ml: BD biosciences) and hamster anti-mouse CD28 (2 μg/ml: BD biosciences). Two days later, the cells were cultured for another three days in RPMI plus IL-2 (20 ng/ml) and IL-4 (50 ng/ml: Preprotech), and then stimulated for 5 h with PMA (20 ng/ml) /ionomycin (1 μg/ ml). The cell medium was collected centrifuged and the supernatant frozen.

| ELISAs
The production of IL-1β by monocytes and of GM-CSF (granulocytemacrophage colony-stimulating factor) by splenocytes from each mice was measured by solid-phase sandwich ELISA using Quantikine kits (R&D systems Inc.), following the manufacturer's instructions.
The assay's sensitivity was 1.8 pg/ml for GM-CSF and 2.31 pg/ml for IL-1β.

| Statistical analysis
The data were expressed as the mean ± SEM, and they were analyzed using GraphPath Prism5 Software. One-way ANOVA followed by the Bonferroni post hoc test was used to determine the statistical significance. For non-parametric analyses, a Kruskal-Wallis and Mann-Whitney U test were applied. Multiple data of motor activity obtained by the repeated measurement from one subject were analyzed by ANOVA for repeated measures. p < 0.05 were consider significant.

ACK N OWLED G M ENTS
We thank Laura Ramos and Gema Atienza for their excellent technical support. We are also grateful to Dr. Moses Rodríguez (Department of Immunology and Neurology, Mayo Clinic/Foundation, Rochester, MN, USA) for the generous gift of the Theiler's virus DA strain.
Finally, we thank the personnel of the Instituto Cajal-Microscopy Unit for their brilliant technical assistance. This work was supported by grants from the Ministerio de Economía y Competitividad (MINECO SAF2016-76449-R) and the Red Española de Esclerosis Múltiple (REEM: RD 16/0015/0001 and RD16/0015/0021), sponsored by the Fondo de Investigación Sanitaria (FIS). We also thank Esclerosis Múltiple España for financing the collaborative project EMMP2017.

CO N FLI C T S O F I NTE R E S T
The author(s) have no potential conflicts of interest. All authors approved the final version of the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data available on request from the authors.