Cytotoxic CD8⁺ T cell ablation enhances the capacity of regulatory T cells to delay viral elimination in Theiler's murine encephalomyelitis

Abstract

Theiler's murine encephalomyelitis (TME) of susceptible mouse strains is a commonly used infectious animal model for multiple sclerosis. The study aim was to test the hypothesis whether cytotoxic T cell responses account for the limited impact of regulatory T cells on antiviral immunity in TME virus‐induced demyelinating disease (TMEV‐IDD) resistant C57BL/6 mice. TME virus‐infected C57BL/6 mice were treated with (i) interleukin‐2/‐anti‐interleukin‐2‐antibody‐complexes to expand regulatory T cells (“Treg‐expansion”), (ii) anti‐CD8‐antibodies to deplete cytotoxic T cells (“CD8‐depletion”) or (iii) with a combination of Treg‐expansion and CD8‐depletion (“combined treatment”) prior to infection. Results showed that “combined treatment”, but neither sole “Treg‐expansion” nor “CD8‐depletion,” leads to sustained hippocampal infection and virus spread to the spinal cord in C57BL/6 mice. Prolonged infection reduces myelin basic protein expression in the spinal cord together with increased accumulation of β‐amyloid precursor protein in axons, characteristic of myelin loss and axonal damage, respectively. Chronic spinal cord infection upon “combined treatment” was also associated with increased T and B cell recruitment, accumulation of CD107b⁺ microglia/macrophages and enhanced mRNA expression of interleukin (IL)‐1α, IL‐10 and tumor necrosis factor α. In conclusion, data revealed that the suppressive capacity of Treg on viral elimination is efficiently boosted by CD8‐depletion, which renders C57BL/6 mice susceptible to develop chronic neuroinfection and TMEV‐IDD.

INTRODUCTION

Multiple sclerosis (MS), one of the most frequent central nervous system (CNS) diseases in young adults, is a chronic demyelinating disease of unknown etiology and possibly multifactorial causes. Based on the generation of myelin‐specific immune responses, MS is regarded as an autoimmune disease 3, 64, presumably triggered by virus infection 30, 61. Because of the clinical and pathological similarities, Theiler's murine encephalomyelitis (TME) represents a commonly used infectious animal model for the chronic‐progressive form of MS in humans 14, 57, 59. Theiler's murine encephalomyelitis virus (TMEV) is a single‐stranded RNA virus of the genus Cardiovirus (family Picornaviridae). Infection of susceptible SJL mice with the low neurovirulent BeAn‐strain causes an acute transient polioencephalitis 44, followed by virus persistence with immune‐mediated myelin loss in the spinal cord (TMEV‐induced demyelinating disease [TMEV‐IDD]) 27, 44, 52, 58, 84, 85. In contrast, TMEV‐IDD‐resistant C57BL/6 mice eliminate the virus from the CNS by specific cellular immunity 60, especially by effective CD8‐mediated cytotoxicity directed against viral epitopes of the VP2 protein 5, 29. The central role of early cytotoxicity has been demonstrated by studies in mice lacking CD8⁺ T cells, which develop TMEV‐persistence and demyelinating disease despite the resistant C57BL/6 background 62, 71, 72. Furthermore, transfer of virus‐specific CD8⁺ T cells to susceptible SJL mice results in virus clearance and prevention of demyelinating disease 21, 56.

Regulatory T cells (Treg), characterized by the expression of the transcription factor forkhead box P3 (Foxp3), play a key role in the maintenance of immunological tolerance and can prevent immunopathology 75, 76. However, in viral diseases Treg can exhibit dual functions with beneficial effects by reducing immune‐mediated tissue damage and detrimental effects because of their immunosuppressive properties, causing disease exacerbation and viral persistence 10, 17, 42, 45, 46, 47, 69, 70, 83, 94. Therefore, a detailed knowledge about antiviral immunity, immune‐mediated damage and viral replication kinetics is required for the use of therapeutic Treg‐modulating strategies in CNS disorders.

Treg contribute to impaired virus‐specific responses in the Friend retrovirus mouse model 100 and experimental herpesvirus infection 17, 82 of mice by inhibiting virus‐specific CD8⁺ T cells. Similarly, an imbalance between Treg and CD8⁺ T cell responses during the acute infection phase is supposed to contribute to virus persistence in TMEV‐IDD‐susceptible SJL mice 70. Accordingly, rapid expansion of Treg has been demonstrated in the brain of susceptible SJL mice but not in resistant C57BL/6 mice following TMEV‐infection 24, 70. Moreover, functional Treg‐inactivation by anti‐CD25‐antibodies prior to infection results in enhanced virus‐specific immunity, reduced viral load and delayed disease progression, while adoptive transfer of ex vivo‐expanded Treg leads to disease exacerbation in TMEV‐infected SJL mice 50, 70. However, these Treg‐modulating strategies have failed to influence the disease course in C57BL/6 mice, demonstrating the complexity of antiviral immunity in TMEV‐IDD‐resistant mouse strains. In addition, in a previous study we did not observe any effect on virus clearance in TMEV‐infected transgenic DEREG mice (BAC‐transgenic Foxp3 reporter mice on C57BL/6 background) following ablation of Foxp3⁺ Treg during the acute disease phase, despite strong effector T cell responses 67.

The study aim was to test the hypothesis that strong cytotoxic T cell responses account for the limited impact of Treg on antiviral immunity in TMEV‐IDD‐resistant mice. For this purpose, the effects of in vivo Treg‐expansion, in vivo CD8‐depletion and a combination of both treatments prior to TMEV infection were compared. Results revealed that the suppressive capacity of in vivo‐expanded Treg on viral elimination is efficiently boosted by CD8‐depletion, which leads to sustained hippocampal infection and increased virus spread to the spinal cord.

MATERIALS AND METHODS

Experimental design

Female C57BL/6JOlaHsd and SJL/JCrHsd mice were purchased from Harlan. At the age of 5 weeks, mice were inoculated in the right cerebral hemisphere with 1.14 × 10⁵ plaque forming units of the BeAn strain of TMEV diluted in 20 µL Dulbecco's modified Eagle Medium (PAA Laboratories) with 2% fetal calf serum (GE Healthcare Laboratories) and 50 µg/kg gentamicin (Sigma‐Aldrich). TMEV was kindly provided by Prof. Dr. H. L. Lipton (Department of Neurology, Northwestern University Medical School, Chicago). For intracerebral injection, mice were anesthetized with medetomidine (0.5 mg/kg, Dormitor, Orion Pharma) and ketamine (100 mg/kg, Ketamine 10%, WDT eG).

A clinical examination of the animals was performed weekly and symptoms were assessed using a three‐tiered score including (a) general appearance (0 = normal posture, smooth and shiny hair, 1 = normal posture, shaggy and dull hair, 2 = slightly hunched back, shaggy and dull hair, 3 = severely hunched back, unkempt appearance, incontinence), (b) behavior (0 = alert and curious, 1 = very quiet, mildly reduced spontaneous movement, no reduction of induced movement, 2 = moderately reduced spontaneous movement, slightly reduced induced movement, 3 = no or minimal spontaneous or induced movement), (c) gait abnormalities (0 = normal gait, 1 = mild ataxia with inconsistent waddling gait, 2 = moderate ataxia with consistent waddling or stiff gait and paddling tail, 3 = severe ataxia with stiff or sliding gait and reduced righting response, and 4 = severe ataxia and paresis of the hind legs), as described previously 88. Animals were euthanized prior to the experimental endpoint, if they reached a score of 3 in the categories “general appearance” and/or “behavior,” a score of 4 in the category “gait abnormalities” or if they lost ≥20% of body weight for animal welfare purposes.

72 C57BL/6 mice in groups of 4–7 mice were euthanized at 3, 14 and 42 days post infection (dpi) with an overdose of medetomidine (2 mg/kg) and ketamine (400 mg/kg) following anesthesia as described above. In addition, 5 infected SJL mice, representing a TMEV‐IDD susceptible mouse strain, were euthanized at 42 dpi for comparison with spinal cord inflammatory responses induced in treated C57BL/6 mice. For characterizing changes in peripheral lymphoid organs under steady state conditions, 48 non‐infected C57BL/6 mice were euthanized at 3, 7 and 14 days of treatment. Animals were perfused via the left ventricle of the heart with phosphate buffered saline (PBS). Blood samples were taken before perfusion and collected for flow cytometric analysis after adding 50 international units of heparin (Heparin‐Natrium‐5000‐ratiopharm® GmbH). Brain, spinal cord, spleen, thymus and cervical lymph nodes were removed immediately. The cerebrum was cut transversally at the level of optic chiasm and the spinal cord was divided into three parts (cervical, thoracic and lumbar). The caudal part of the cerebrum and one half of each spinal cord segment was fixed in 10% formalin for 24 hours and subsequently embedded in paraffin wax. The rostral part of the cerebrum and parts of each spinal cord segment were snap frozen and stored at −80°C until further use for RNA isolation.

The study was conducted in accordance with German law for animal protection and with the European Communities Council Directive 86/609/EEC for the protection of animals used for experimental purposes and in accordance with the German Animal Welfare Act. All animal experiments were approved and authorized by the Local Institutional Animal Care and Research Advisory committee and permitted by the local authorities (Regierungspräsidium, Hannover, Germany, permission numbers: 33.9‐42502‐04‐11/0538 and 33.12‐42502‐04‐13/1138).

Expansion of regulatory T cells and depletion of CD8⁺ T cells

To expand regulatory T cells (Treg‐expansion) in vivo, animals were treated with interleukin‐2/anti‐interleukin‐2 immune complexes (IL‐2C) as described elsewhere 6, 97. IL‐2C consist of 10 µg of anti‐IL‐2 antibodies (eBiosciences, clone JES6‐1A12) mixed with 2 µg murine recombinant IL‐2 protein (eBiosciences, catalog number 34‐8021), incubated at 37°C for 30 minutes before injection. IL‐2C were injected intraperitoneally (i.p.) on three consecutive days (day ‐3, ‐2 and ‐1) prior to TMEV‐infection on day 0 (group Treg↑TMEV).

For depletion of CD8⁺ T cells (CD8‐depletion), animals received 500 µg anti‐CD8 antibodies 12 on three consecutive days (day ‐3, ‐2 and ‐1) i.p. prior to TMEV‐infection at day 0 (group CD8↓TMEV). For simultaneous Treg‐expansion and CD8‐depletion (combined treatment), IL‐2C and anti‐CD8‐antibodies were injected with an interval of 6 hours on the same day (group Treg↑CD8↓TMEV). Control animals received PBS injections i.p. at the same time points (group TMEV).

According to the same application scheme, non‐infected C57BL/6 mice were either treated with PBS (control), IL‐2C (group Treg↑) or IL‐2C and anti‐CD8‐antibodies (group Treg↑CD8↓) to investigate the effect of treatment under steady state (non‐infectious) conditions.

Histology

Formalin‐fixed, paraffin‐embedded cerebral and spinal cord tissue was sectioned at 2 µm thickness and stained with hematoxylin and eosin (HE). The cerebrum was sectioned coronally at the level of bregma −1.46 to −1.82 65. Myelitis was determined by evaluating the spinal cord segments: cervical, thoracic and lumbar. A semiquantitative scoring system based on the degree of perivascular infiltration (0 = no changes, 1 = one layer, 2 = two to three layers, 3 = more than three layers of inflammatory cells) and hypercellularity (0 = no change, 1 = 1–25 cells, 2 = 26–50 cells, 3 = >50 cells) was applied as described previously 18, 19, 24. The total myelitis score was determined by adding the scores of all three segments for each animal.

Immunohistochemistry

For immunohistochemistry, antibodies detecting viral protein (TMEV), Foxp3 (Treg), CD3 (T cells), CD107b (microglia/macrophages), CD45R (B cells), β‐amyloid precursor protein (β‐APP; damaged axons) and myelin basic protein (MBP) were used on paraffin‐embedded sections. Antibodies detecting CD4⁺ and CD8⁺ T cells in the spinal cord were used on frozen tissue sections. Antibody details are summarized in Table 1. As negative controls, primary antibodies were replaced by specific isotype control or specific serum, respectively. For visualization, a peroxidase‐conjugated avidin‐biotin complex (Vector Laboratories) and 3.3‐diaminobenzidine‐tetrachloride in 0.1 M imidazole (Sigma‐Aldrich) was used. Subsequently, sections were counterstained with Mayer's hematoxylin (Merck). The absolute numbers of labeled cells (TMEV‐infected cells, Foxp3⁺ Treg, CD3⁺ T cells, CD107b⁺ microglia/macrophages and CD45R⁺ B cells) were counted in coronal sections (bregma −1.46 to −1.82) of the hippocampus. The area of hippocampus was measured on digitalized images using the analySIS® 3.2 software (SOFT Imaging System) and the number of labeled cells per area [cell number/mm²] was calculated. Densitometric analysis was also used to quantify the percentage of CD45R‐positive area in spleen sections stained by immunohistochemistry. In all three segments of the spinal cord the total number of labeled cells (TMEV, Foxp3, CD3, CD4, CD8 CD107b, CD45R) was determined. Similarly, β‐APP‐positive axons were enumerated in entire cross sections of spinal cord segments. The amount of myelin loss in the spinal cord was quantified by manually outlining and measuring the MBP‐negative area [µm²] using the analySIS® 3.2 software. To detect small, randomly distributed foci of myelin loss (MBP) and axonal damage (β‐APP), four cross sections per spinal cord segment (cervical, thoracic, lumbar) were evaluated in each animal.

Table 1.

Antibodies used for immunohistochemistry.

Antigen	Company; product number; clone	Blocking serum	Pre treatment	Dilution	Secondary antibody	Reference
TMEV	–	Goat	–	1:2000	Goat anti‐rabbit	( 40)
Foxp3	eBioscience; 14‐5773; Clone: FJK‐16s	Rabbit	Microwave, 20 minutes, citrate buffer	1:20	Rabbit anti‐rat	( 24)
CD3	Dako/Agilent Technologies; A0452	Goat	Microwave, 20 minutes, citrate buffer	1:1000	Goat anti‐rabbit	( 24)
CD4	BD BioScience; L3T4, Clone; RM4‐5	Rabbit	–	1:2000	Rabbit anti‐rat	( 18)
CD8	BD BioScience; Ly‐3.2; Clone: 53‐5.8	Rabbit	–	1:1000	Rabbit anti‐rat	( 18)
CD107b	AbD serotec; MCA 2293B; M3/84	*	Microwave, 20 minutes, citrate buffer	1:200	*	( 24)
CD45R	BD Biosciences; 553085; RA3‐6B2	*	Microwave, 20 minutes, citrate buffer	1:1000	*	( 24)
β‐APP	Chemicon; MAB348; 22C11	Mouse	Microwave, 20 minutes, citrate buffer	1:2000	Goat anti‐mouse	( 39)
MBP	Merck/Millipore; AB980;	Goat	–	1:500	Goat anti‐rabbit	( 25)

TMEV = Theiler's murine encephalomyelitis virus, Foxp3 = Forkhead box P3, β‐APP = beta‐amyloid precursor protein, MBP = myelin basic protein.

*Biotinylated antibody, no blocking serum and secondary antibody used.

Flow cytometry

Spleen samples were dissolved to single cell suspension in PBS containing 0.2 % bovine serum albumin (PBS/BSA; Gibco) using a 100 µm sieve and centrifuged (330 × g, 10 minutes, room temperature). Erythrocytes were lysed by adding 1 mL ammonium‐chloride‐potassium lysing buffer (Gibco) for 2 minutes at room temperature. Cells were washed, resuspended and filtered (30 µm) in PBS/BSA. The cell number was determined and 2 × 10⁶ cells were used of each sample. For erythrocyte lysis of blood samples, 75 µL of heparin‐blood were mixed with 2 mL lysing buffer and incubated for 4 minutes at room temperature. Lysis reaction was stopped by adding 3 mL RPMI 1640 medium with L‐glutamine (Gibco) with 10% of fetal calf serum (Biochrom AG) and 50 U/mL Penicillin‐streptomycin (Gibco). Samples were centrifuged (330 × g, 8 minutes, 4°C), supernatants were discarded and cells were washed in 3 mL PBS/BSA (Gibco). Blood and spleen cells were stained as follows: For FcγR blockade, 100 µL of Fcy block medium (BioXcell, 2.4G2, 1:100, diluted in PBS with 0.02% sodium azide) were added to each tube, incubated for 5 minutes at 4°C and washed in PBS/BSA. Live/dead staining was performed using the LIVE/DEAD Fixable Blue Dead Cell Stain Kit (Invitrogen, 1:500, diluted in PBS). About 100 µL were added to each tube, samples were incubated for 30 minutes at 4°C and washed in PBS/BSA. For cell surface staining, a fourfold concentrated antibody mix detecting helper T cells (CD4) and cytotoxic T cells (CD8) was added. For antibody details see Table 2. Samples were incubated for 15 minutes at 4°C in the dark. Subsequently, cells were washed in 3 mL PBS/BSA. For Foxp3 staining the Foxp3 Transcription Factor Staining Buffer Set (eBioscience) was used. Cells were re‐suspended in 1 mL fixation/permeabilization concentrate (fixation permeabilization concentrate/fixation permeabilization diluent 1:3) and incubated for 30 minutes at 4°C in the dark. After centrifugation (300 × g, 8 minutes, 4°C) cells were re‐suspended in 600 µL PBS/BSA and kept at 4°C in the dark until further processing. Cells were washed twice with 2 mL of 1× working solution of the permeabilization buffer and supernatants were discarded. About 50 µL permeabilization buffer, including rat IgG ([40 µg/mL], Jackson Immuno Research) were added and samples were incubated for 15 minutes at 4°C in the dark. Subsequently, without washing, 100 µL permeabilization buffer and the anti‐Foxp3 antibody were added and incubated for 30 minutes at 4°C in the dark. Afterward, samples were washed in permeabilization buffer and 400 µL of permeabilization buffer were added. Samples were incubated for 5 minutes at room temperature in the dark and centrifuged (1400 rpm, 4 minutes, 4°C). Finally, cells were re‐suspended in PBS/BSA, filtered and stored at 4°C in the dark until use. For flow cytometry, a LSRII SORP cytometer (BD Biosciences) was used and data were analyzed using FlowJo software version 9.6.4 (Tree Star).

Table 2.

Antibodies used for flow cytometry.

Antibody	Company	Clone
αCD4 PercpCy5.5	eBioscience	RM‐4‐5
αCD8 APC	BioLegend	53‐6.7
Foxp3 PE	eBioscience	FJK‐16S

Foxp3 = Forkhead box P3.

Ribonucleic acid isolation and reverse transcription

Snap frozen brain and spinal cord tissues were cut and ribonucleic acid (RNA) was isolated from 10 to 40 mg tissue of each brain/spinal cord using an Omni's PCR Tissue Homogenizing Kit (Süd‐Laborbedarf GmbH), QIAzol^® lysis reagent and RNeasy^® Mini Kit (Qiagen). RNA sample quality and amount of RNA was measured by applying a NanoDrop 1000 spectrophotometer (Thermo Fischer Scientific). Equal amounts of RNA were subsequently transcribed into cDNA with the Omniscript^TM RT Kit (Qiagen), RNaseOUT™, Recombinant Ribonuclease Inhibitor (life technologies) and random primers (Promega). For reverse transcription, the following thermocycler program was used: 10 minutes at 25°C, followed by 1 hour at 37°C and at the end 5 minutes at 93°C.

Polymerase chain reaction

Reverse transcription quantitative polymerase chain reaction (RT‐qPCR) for the quantification of viral RNA (TMEV) and the three reference genes glyceraldehyde 3‐phosphate dehydrogenase (GAPDH), β‐actin and hypoxanthine‐guanine phosphoribosyltransferase (HPRT) in brain and spinal cord tissues was performed using the Mx3005P™ Multiplex Quantitative PCR System (Agilent Technologies) and Brilliant III SYBR^® Green Mastermix (Agilent Technologies). In addition, Foxp3 and cytokine mRNA expression (interleukin [IL]‐1α, IL‐2, IL‐10, IL‐23, tumor necrosis factor [TNF]‐α, transforming growth factor [TGF]‐β1) were analyzed in spinal cord and brain tissues. Primer sequences for IL‐23 were created by using the NCBI Primer‐Blast‐Tool. All other sequences were taken from the literature 23, 24, 25, 88, 89 For primer details see Table 3. For quantification, 10‐fold serial dilution standards ranging from 10⁸ to 10² copies/µL were prepared. The normalization factor for correction of experimental variations was calculated from the three reference genes with geNorm software version 3.4 93. Specificity of each reaction was controlled by melting curve analysis.

Table 3.

Primer sequences used for polymerase chain reaction.

Gene	Primer	sequence 5′ → 3′	Length	Reference
GAPDH	Forward	GAG GCC GGT GCT GAG TAT GT	288 bp	( 89)
	Reverse	GGT GGC AGT GAT GGC ATG GA
HPRT	Forward	GGA CCT CTC GAA GTG TTG GA	169 bp	( 89)
	Reverse	TTG CGC TCA TCT TAG GCT TT
β‐actin	Forward	GGC TAC AGC TTC ACC ACC AC	233 bp	( 89)
	Reverse	ATG CCA CAG GAT TCC ATA CC
TMEV	Forward	GAC TAA TCA GAG GAA CGT CAG C	129 bp	( 88, 91)
	Reverse	GTG AAG AGC GGC AAG TGA GA
Foxp3	Forward	TTC TCA CAA CCA GGC CAC TTG	88 bp	( 23)
	Reverse	CCC AGG AAA GAC AGC AAC CTT
IL‐1α	Forward	AAG CAA CGG GAA GAT TCT GA	179 bp	( 25)
	Reverse	TGA CAA ACT TCT GCC TGA CG
IL‐2	Forward	GCA GGA TGG AGA ATT ACA GGA	183 bp	( 25)
	Reverse	TGA AAT TCT CAG CAT CTT CCA A
IL‐10	Forward	CCA AGC CTT ATC GGA AAT GA	162 bp	( 25)
	Reverse	TTT TCA CAG GGG AGA AAT CG
IL‐23	Forward	CCT CCA GCC AGA GGA TCA CCC C	145 bp	–
	Reverse	GTG GGC AAA GAC CCG GGC AG
TNF‐α	Forward	GCC TCT TCT CAT TCC TGC TT	203 bp	( 25)
	Reverse	CAC TTG GTG GTT TGC TAC GA
TGF‐β1	Forward	TTG CTT CAG CTC CAC AGA GA	183 bp	( 25)
	Reverse	TGG TTG TAG AGG GCA AGG AC

GAPDH = glyceraldehyde 3‐phosphate dehydrogenase, HPRT = hypoxanthine‐guanine phosphoribosyltransferase, TMEV = Theiler's murine encephalomyelitis virus, Foxp3 = Forkhead box P3, IL = Interleukin, TNF‐α = tumor necrosis factor‐α, TGF‐β1 = transforming growth factor‐β1.

Statistical analyses

For statistical analysis of non‐normal distributed data, multiple Mann–Whitney U‐tests were performed to determine any difference between treatment groups. A P‐value of ≤0.05 was considered as being statistically significant. Analyses were performed using SPSS for windows (SPSS Inc.). Graphs were created with GraphPad Prism^® (GraphPad Software).

RESULTS

Expansion of regulatory T cells maintains virus infection in the brain following depletion of CD8⁺ T cells in C57BL/6 mice

To elucidate the effect of Treg‐cytotoxic T cell interplay on antiviral immunity in TMEV‐IDD‐resistant C57BL/6 mice, Foxp3⁺ Treg were expanded by IL‐2C application (Treg‐expansion) with and without simultaneous antibody‐mediated depletion of CD8⁺ T cells. Anti‐CD8‐antibody (αCD8)‐treated mice and combined treated mice (groups CD8↓TMEV and Treg↑CD8↓TMEV) exhibited significantly elevated virus RNA levels compared to untreated controls (group TMEV) at 3 dpi (Figure 1A). Strikingly, Treg‐expansion together with CD8‐depletion (group Treg↑CD8↓TMEV) delayed viral elimination, showing significantly increased virus RNA concentrations in the brain compared to all other groups at 14 dpi (Figure 1A, Supporting Information Table S1). Subsequently, at 42 dpi, virus RNA decreased in combined treated mice, but was still detectable and significantly higher compared to groups Treg↑TMEV and CD8↓TMEV (Figure 1A). Although increased virus RNA levels were also found at 14 dpi following CD8‐depletion (group CD8↓TMEV) with significantly more TMEV RNA copies than found in Treg‐expanded mice (group Treg↑TMEV), values did not reach the level of significance compared to untreated controls (group TMEV) (Figure 1A, Supporting Information Table S1). As expected, untreated, TMEV‐infected C57BL/6 mice (group TMEV) nearly cleared the virus after the acute polioencephalitis phase at 42 dpi, as demonstrated by RT‐qPCR (Figure 1A). Similarly, Treg‐expansion alone (group Treg↑TMEV) did not alter the amount of viral RNA in the brain compared to group TMEV mice and virus elimination was found at 42 dpi, indicative of sustained effective antiviral immunity (Figure 1A).

Figure 1.

Theiler's murine encephalomyelitis virus (TMEV) infection in the brain is prolonged after concurrent regulatory T cell expansion and CD8+ T cell depletion. A. Quantification of TMEV RNA copy numbers by reverse transcriptase quantitative polymerase chain reaction. B. Quantification of TMEV‐infected cells by immunohistochemistry. C. Representative image showing the distribution of infected cells (brown signal) in the hippocampus of a combined treated mouse at 14 days post infection (dpi). Note preferential infection of the pyramidal layer (CA2 region; arrow). D. Higher magnification of (C) showing infected cells with neuronal morphology. (A, B) Box plots display median and quartiles with minimum and maximum values. Significant differences (P ≤ 0.05; Mann–Whitney U test) are labeled: ∗ significant difference to group TMEV; # significant difference to group Treg↑TMEV; ○ significant difference to group CD8↓TMEV. TMEV = untreated, infected mice; Treg↑TMEV = regulatory T cell expanded, infected mice; CD8↓TMEV = CD8⁺ T cell depleted, infected mice; Treg↑CD8↓TMEV = combined treated, infected mice. (C, D) Immunohistochemistry, scale bars: 200 µm (C) and 20 µm (D).

Confirming RT‐qPCR results, highest numbers of TMEV‐infected cells were found in combined treated mice (group Treg↑CD8↓TMEV) at 14 dpi (Figure 1B). Notably, an increased infection of the hippocampal pyramidal cell layer was noticed in Treg‐expanded mice (group Treg↑TMEV) compared to untreated infected mice (group TMEV) at 14 dpi.

Immunohistochemistry and histology revealed that TMEV‐infection and inflammatory responses were largely restricted to the hippocampus as described before 51 (Figures 1C,D and 2A–C). Encephalitis was characterized by perivascular infiltration of mononuclear cells and hypercellularity of the neuroparenchyma (Figure 2A–C).

Figure 2.

Theiler's murine encephalomyelitis virus (TMEV)‐induced hippocampal inflammation. A. Representative image of a combined treated mouse showing marked hippocampal inflammation with perivascular infiltrates (arrows) at 14 days post infection (dpi). B. Higher magnification of (A) showing perivascular cuffs composed of mononuclear cells. C. Higher magnification of (A) showing prominent hypercellularity within the pyramidal layer associated with neuronal loss. D. Note lack of overt inflammation because of virus elimination in an infected mouse without treatment at 14 dpi. (A–D) Hematoxylin and eosin staining. Scale bars: 200 µm (A, D) and 50 µm (B, C).

Taken together, concurrent antibody‐mediated depletion of CD8⁺ T cells potentiates the ability of Treg‐expansion to reduce antiviral responses in C57BL/6 mice during the early TME phase. Spatiotemporal analyses revealed that delayed viral elimination in combined treated mice is associated with a higher TMEV load in the hippocampus.

CD8‐depletion leads to an enhanced recruitment of Foxp3⁺ regulatory T cells to the infected hippocampus

Immunohistochemistry was performed to characterize dynamic changes of cellular responses and quantify accumulating immune cells in the hippocampus associated with Treg‐ and/or CD8‐modulation. In all treatment groups, acute hippocampal infection (3 dpi) was dominated by CD3⁺ T cells with significantly increased T cell numbers following Treg‐expansion alone (group Treg↑TMEV; Figure 3A–C, Supporting Information Table S1). Similarly, significantly enhanced infiltration of Foxp3⁺ Treg was found in Treg‐expanded mice at 3 dpi (Figure 3D). Remarkably, despite reduced CD3⁺ T cell infiltration in both αCD8‐treated groups (groups CD8↓TMEV and Treg↑CD8↓TMEV), CD8‐depletion combined with Treg‐expansion significantly increased the amount of infiltrating Foxp3⁺ Treg in the hippocampus compared to controls at 3 dpi. Associated with this, a significantly increased Foxp3 mRNA expression was found in combined treated mice compared to all other groups (Figure 3E). Probably attributed to termination of prolonged hippocampal inflammation, significantly elevated numbers of Foxp3⁺ Treg together with elevated Foxp3 mRNA levels were found in the hippocampus of both αCD8‐treated groups (group CD8↓TMEV and Treg↑CD8↓TMEV; Figure 3D–F, Supporting Information Table S1) also at 42 dpi.

Figure 3.

Phenotyping of immune cell subsets in the hippocampus. A–C. Quantification of CD3⁺ T cells, D–F. Foxp3⁺ regulatory T cells, G–I. CD45R⁺ B cells and J–L. CD107b⁺ microglia/macrophages in the hippocampus by immunohistochemistry. (E) Quantification of Foxp3 mRNA in the cerebrum by reverse transcriptase quantitative polymerase chain reaction. Representative images of a combined treated mouse showing the distribution of (B, C) CD3⁺ T cells, (F) Foxp3⁺ regulatory T cells, (H, I) CD45R⁺ B cells, and (K, L) CD107⁺ microglia/macrophages at 14 days post infection (dpi). Immuno‐labeled cells (brown signal) are indicated by arrows. (A, D, E, G, J) Box plots display median and quartiles with minimum and maximum values. Significant differences (P ≤ 0.05; Mann–Whitney U test) are labeled: ∗ significant difference to group TMEV; # significant difference to group Treg↑TMEV; ○ significant difference to group CD8↓TMEV. TMEV = untreated, infected mice; Treg↑TMEV = regulatory T cell expanded, infected mice; CD8↓TMEV = CD8⁺ T cell depleted, infected mice; Treg↑CD8↓TMEV = combined treated, infected mice. Immunohistochemistry (B, C, F, H, I, K, L). Scale bars: 200 µm (B, H, K) and 20 µm (C, F, I, L).

In contrast to T cells, CD45⁺ B cell infiltration was significantly reduced in the hippocampus following Treg‐expansion (group Treg↑TMEV) compared to untreated mice (group TMEV) at 3 dpi. Sustained brain infection in combined treated mice (group Treg↑CD8↓TMEV) led to significantly increased CD45R⁺ B cell numbers at 14 dpi. Increased numbers of CD45R⁺ B cells, significantly different from group Treg↑TMEV, were also found in αCD8‐treated (group CD8↓TMEV) and combined treated mice (group Treg↑CD8↓TMEV) at 42 dpi (Figure 3G–I, Supporting Information Table S1).

CD8‐depletion (group CD8↓TMEV) significantly decreased the amount of CD107b⁺ cells during acute TMEV infection (3 dpi). In parallel with virus load kinetics, CD107b⁺ microglia/macrophage accumulation peaked at 14 dpi in combined treated mice (group Treg↑CD8↓TMEV) with significantly increased values compared to all other groups (Figure 3J–L, Supporting Information Table S1).

In summary, data confirmed that in vivo Treg‐expansion by IL‐2C treatment transiently increases the CNS recruitment of Foxp3⁺ Treg during acute TMEV‐infection, which is markedly enhanced by CD8‐depletion. Phenotypical analyses also revealed that prolonged hippocampal infection in combined treated C57BL/6 mice is associated with enhanced CD107b⁺ microglia/macrophage responses and CD45R⁺ B cells infiltration.

Inflammatory responses in the brain were further quantified by cytokine expression analyses. RT‐qPCR revealed a significantly increased transcription of the pro‐inflammatory cytokines IL‐1α and TNF‐α in both αCD8‐treated groups (group CD8↓TMEV and Treg↑CD8↓TMEV) at 14 dpi with highest values found in combined treated mice (Figure 4A,B). Similarly, IL‐2 mRNA levels were significantly elevated in combined treated mice at 14 dpi (Figure 4C). IL‐23 mRNA levels were transiently decreased in αCD8‐treated mice (group CD8↓TMEV) at 3 dpi and increased in Treg‐expanded mice (group Treg↑TMEV) at 14 dpi. Notably, a significantly increased IL‐23 transcription was measured in both αCD8‐treated groups (group CD8↓TMEV and Treg↑CD8↓TMEV) at 42 dpi (Figure 4D), indicative of prolonged inflammatory responses in the brain. Anti‐inflammatory IL‐10 and TGF‐β mRNA expression was highest in αCD8‐treated mice with Treg‐expansion (group Treg↑CD8↓TMEV) at 14 dpi (Figure 4E,F). Together with sustained elevated Foxp3 mRNA expression (Figure 3E), prolonged brain infection was also associated with elevated IL‐10 mRNA expression in combined treated mice at 42 dpi (Figure 4E), characteristic of counter regulatory attempts to reduce neuroinflammation.

Figure 4.

Cytokine expression in the cerebrum. Quantification of A. interleukin (IL)‐1α, B. tumor necrosis factor (TNF)‐α, C. IL‐2, D. IL‐23, E. IL‐10 and F. transforming growth factor‐β mRNA in the cerebrum by reverse transcriptase quantitative polymerase chain reaction. Note prominent pro‐ and anti‐inflammatory cytokine mRNA expression at 14 dpi and prolonged IL‐10 transcription at 42 dpi in combined treated mice. Box plots display median and quartiles with minimum and maximum values. Significant differences (P ≤ 0.05; Mann–Whitney U test) are labeled: ∗ significant difference to group TMEV; # significant difference to group Treg↑TMEV; TMEV = untreated, infected mice; Treg↑TMEV = regulatory T cell expanded, infected mice; CD8↓TMEV = CD8⁺ T cell depleted, infected mice; Treg↑CD8↓TMEV = combined treated, infected mice. dpi = days post infection.

Prolonged viral encephalitis promotes virus spread to spinal cord and leukomyelitis in C57BL/6 mice

The inability of susceptible SJL mice to clear the virus during the acute encephalitis phase of TMEV‐infection is a key step in developing chronic spinal cord infection and TMEV‐IDD 63. Therefore, the question was raised, whether delayed virus elimination from the brain following combined treatment leads to chronic spinal cord infection in resistant C57BL/6 mice.

Similar to the brain, prolonged TMEV‐infection determined by RT‐qPCR was found in the spinal cord of combined treated mice. Group Treg↑CD8↓TMEV animals displayed significantly more viral copy numbers compared to untreated control mice (group TMEV) and Treg‐expanded mice (group Treg↑TMEV) at 14 dpi. At 42 dpi, TMEV RNA load was significantly higher in group Treg↑CD8↓TMEV animals compared to all other groups (Figure 5A, Supporting Information Table S1). Virus RNA was also detected at 14 and 42 dpi in C57BL/6 mice following anti‐CD8‐antibody treatment alone (group CD8↓TMEV), however, values did not reach the level of significance compared to untreated controls or Treg‐expanded animals (Figure 5A, Supporting Information Table S1). Similar to viral RNA, TMEV antigen was observed in group CD8↓TMEV and group Treg↑CD8↓TMEV animals at 14 and 42 dpi. However, only combined treated mice (group Treg↑CD8↓TMEV) showed significantly increased TMEV‐infected cell numbers compared to group TMEV and Treg↑TMEV (Figure 5B, Supporting Information Table S1).

Figure 5.

Concurrent regulatory T cell expansion and CD8+ T cell depletion increases Theiler's murine encephalomyelitis virus (TMEV) infection in the spinal cord. A. Quantification of TMEV RNA in the spinal cord by reverse transcriptase quantitative polymerase chain reaction. B. Quantification of TMEV‐infected cells in the spinal cord by immunohistochemistry. C. Representative image of a combined treated mouse showing infected cells (brown signal, arrow) in the ventrolateral region of the spinal cord at 14 days post infection (dpi). D. Higher magnification of (C). (A, B) Box plots display median and quartiles with minimum and maximum values. Significant differences (P ≤ 0.05; Mann–Whitney U test) are labeled: ∗ significant difference to group TMEV; # significant difference to group Treg↑TMEV; ○ significant difference to group CD8↓TMEV. TMEV = untreated, infected mice; Treg↑TMEV = regulatory T cell expanded, infected mice; CD8↓TMEV = CD8⁺ T cell depleted, infected mice; Treg↑CD8↓TMEV = combined treated, infected mice. (C, D) Immunohistochemistry. Scale bars: 200 µm (C) and 20 µm (D).

Immunohistochemistry showed the presence of TMEV protein in spinal cord grey and white matter at 14 dpi (Figure 5C,D), while a preferential infection of the spinal cord white matter was found at 42 dpi. Since a cell tropism switch of TMEV from grey matter neurons to white matter cells is a prerequisite for demyelination in TMEV‐IDD‐susceptible mice 41 spinal cord pathology including integrity of myelin sheaths and axons was evaluated. While histological scoring revealed significant myelitis in αCD8‐treated animals compared to group TMEV and group Treg↑TMEV at 42 dpi, combined treatment (group Treg↑CD8↓TMEV) caused an even stronger inflammatory response at 14 and 42 dpi, with significantly increased scores compared to group TMEV and Treg↑TMEV at 14 dpi and all other groups at 42 dpi (Figure 6A,B and Supporting Information Table S1). Multiple inflammatory foci with significantly reduced MBP‐staining, indicative of myelin loss, were found in the spinal cord white matter predominantly of both αCD8‐treated groups at 42 dpi (Figure 6C,D). Significant axonal damage characterized by axonal swelling with β‐APP accumulation was found within white matter foci only in combined treated mice (group Treg↑CD8↓TMEV), indicating detrimental effects on axonal integrity triggered by prolonged virus infection (Figure 6E,F, Supporting Information Table S1).

Figure 6.

Enhanced Theiler's murine encephalomyelitis virus (TMEV) infection in the spinal cord causes demyelinating disease in C57BL/6 mice. A. Histological scoring of inflammatory responses in the spinal cord (myelitis score). B. Representative image of a combined treated mouse depicts inflammatory infiltrates (arrows) and vacuolization within spinal cord at 14 days post infection (dpi). Insert = higher magnification of inflamed area. C. Densiometric analysis for quantification of demyelinated areas based on myelin basic protein (MBP) expression. D. Representative image of a combined treated mouse showing reduced MBP expression (arrows) in the lateral aspect of the spinal cord at 42 dpi, indicative of myelin loss. Insert = higher magnification of demyelinated area. E. Quantification of β‐amyloid precursor protein (β‐APP)‐expressing axons by immunohistochemistry. F. Representative image of a combined treated mouse showing β‐APP⁺ axons in the spinal cord at 42 dpi, indicative of axonal damage. (A, C, E) Box plots display median and quartiles with minimum and maximum values. Significant differences (P ≤ 0.05; Mann–Whitney U test) are labeled: ∗ significant difference to group TMEV; # significant difference to group Treg↑TMEV; ○ significant difference to group CD8↓TMEV. TMEV = untreated, infected mice; Treg↑TMEV = regulatory T cell expanded, infected mice; CD8↓TMEV = CD8⁺ T cell depleted, infected mice; Treg↑CD8↓TMEV = combined treated, infected mice. (B) Hematoxylin and eosin staining. (D, F) Immunohistochemistry. Scale bars: 200 µm (B, D), 50 µm (B, D inserts) and 15 µm (F).

Weekly clinical examination revealed no symptoms in groups TMEV, Treg↑TMEV and CD8↓TMEV at any investigated time point. Interestingly, three out of seven combined treated animals (group Treg↑CD8↓TMEV) showed mildly to moderately elevated clinical scores (total scores of 1 to 2) starting at 21 dpi. However, statistical analyses revealed only a statistical tendency between treatment groups (P > 0.06, Supporting Information Table S1), suggestive of marginal effects of neuropathology upon clinical manifestation in combined treated mice.

Collectively, Treg‐expansion together with CD8‐depletion (combined treatment) facilitates TMEV spread and chronic infection of the spinal cord in C57BL/6 mice. Viral leukomyelitis is associated with myelin loss and axonopathy as observed in TMEV‐IDD‐susceptible mouse strains.

Chronic virus infection in the spinal cord stimulates leukocyte recruitment and cytokine expression of C57BL/6 mice

Phenotypical changes of the immune cell composition and cytokine expression in the spinal cord associated with prolonged TMEV‐infection were analyzed by immunohistochemistry and RT‐qPCR (Figures 7A–O and 8A–C).

Figure 7.

Increased Theiler's murine encephalomyelitis virus (TMEV) infection triggers inflammatory responses in the spinal cord of C57BL/6 mice. Analyses of A–C. CD3⁺ T cells, D. CD4⁺ T cells, E, F. CD8⁺ T cells, G–I. CD107b⁺ microglia/macrophages, J–L. CD45R⁺ B cells and M–O. Foxp3⁺ regulatory T cells in the spinal cord. (A) Quantification of CD3⁺ T cells by immunohistochemistry. (B) Representative image of a combined treated mouse showing CD3⁺ T cell infiltrates (brown signal, arrows) in the spinal cord white matter at 42 days post infection (dpi). (C) Higher magnification of (B). (D) Quantification of CD4⁺ T cells and (E) CD8⁺ T cells by immunohistochemistry. (F) Representative image of a mouse treated with anti‐CD8‐antibodies showing CD8⁺ T cell infiltrates (brown signal) in the spinal cord white matter at 42 dpi. (G) Quantification of CD107b⁺ microglia/macrophages by immunohistochemistry. (H) Representative image of a combined treated mouse showing accumulation of CD107⁺ microglia/macrophages (brown signal, arrow) in the spinal cord white matter at 42 dpi. (I) Higher magnification of (H). (J) Quantification of CD45R⁺ B cells by immunohistochemistry. (K) Representative image of a combined treated mouse showing CD45R⁺ B cell infiltrates (brown signal, arrow) in the spinal cord at 42 dpi. (L) Higher magnification of (K). (M) Quantification of Foxp3⁺ regulatory T cells by immunohistochemistry. (N) Quantification of Foxp3 mRNA expression by reverse transcriptase quantitative polymerase chain reaction. (O) Representative image of a combined treated mouse showing Foxp3⁺ regulatory T cells in the spinal cord at 42 dpi. (A, D, E, G, J, M, N) Box plots display median and quartiles with minimum and maximum values. Significant differences (P ≤ 0.05; Mann–Whitney U test) are labeled: ∗ significant difference to group TMEV; # significant difference to group Treg↑TMEV; ○ significant difference to group CD8↓TMEV. TMEV = untreated, infected mice; Treg↑TMEV = regulatory T cell expanded, infected mice; CD8↓TMEV = CD8⁺ T cell depleted, infected mice; Treg↑CD8↓TMEV = combined treated, infected mice. (B, C, F, H, I, K, L, O) Immunohistochemistry. Scale bars: 200 µm (B, H, K) and 20 µm (C, F, I, L, O).

Similar to the brain lesions, CD3⁺ T cells dominated the inflammatory response in the spinal cord of all treatment groups (Figure 6A–C). In αCD8‐treated (group CD8↓TMEV) and combined treated mice (group Treg↑CD8↓TMEV), significantly increased numbers of CD3⁺ T cells, CD4⁺ T cells and CD107b⁺ microglia/macrophages compared to untreated controls (group TMEV) were found at 42 dpi (Figure 7A–I). Strikingly, CD8⁺ T cells were only elevated in αCD8‐treated (group CD8↓TMEV) mice at 42 dpi, suggestive of impaired CNS‐infiltration of cytotoxic T cells in combined treated mice (Figure 7E,F, Supporting Information Table S1). CD45R⁺ B cell numbers were significantly increased only in combined treated mice at 42 dpi (Figure 7J–L, Supporting Information Table S1). In comparison to TMEV‐IDD susceptible SJL mice, significantly lower numbers of CD3⁺ T cells, CD45R⁺ B cells and CD107b⁺ microglia/macrophages were found in combined treated mice (group Treg↑CD8↓TMEV) at 42 dpi (Supporting Information Figure S1A–C, Supporting Information Table S2).

Inflammatory responses in group CD8↓TMEV were accompanied by significant infiltration of Foxp3⁺ Treg at 42 dpi (Figure 7M, Supporting Information Table S1). Suggestive of increased Treg responses, significantly elevated numbers of Foxp3⁺ Treg together with a raised Foxp3 mRNA expression at 14 and 42 dpi was only found in combined treated mice (group Treg↑CD8↓TMEV; Figure 7M,N; Supporting Information Table S1). Numbers of Foxp3⁺ cells in combined treated mice were not significantly different from those observed in TMEV‐IDD susceptible SJL mice (Supporting Information Figure S1D; Supporting Information Table S2).

Inflammatory responses in the spinal cord were further studied by cytokine expression analyses. RT‐qPCR showed a significantly increased transcription of the pro‐inflammatory cytokines IL‐1α and TNF‐α at 14 and 42 dpi in combined treated mice (group Treg↑CD8↓TMEV) compared to untreated controls and Treg‐expanded animals (group TMEV and Treg↑TMEV, Figure 8A,B; Supporting Information Table 1). Additionally, TNF‐α mRNA levels were significantly higher in combined treated animals than in αCD8‐treated animals at 42 dpi (Figure 8B; Supporting Information Table S1). The anti‐inflammatory cytokine IL‐10 was significantly elevated at 14 and 42 dpi in group Treg↑CD8↓TMEV compared to untreated controls (group TMEV; Figure 8C, Supporting Information Table S1). Other investigated cytokines (IL‐2, IL‐23 and TGF‐β) were not differentially expressed at both time points (Supporting Information Table S1).

Figure 8.

Increased cytokine expression in Theiler's murine encephalomyelitis virus (TMEV)‐infected spinal cord following concurrent regulatory T cell expansion and CD8⁺ T cells depletion. Quantification of A. interleukin (IL)‐1α, B. tumor necrosis factor (TNF)‐α and C. IL‐10 mRNA in the spinal cord by reverse transcriptase quantitative polymerase chain reaction. Box plots display median and quartiles with minimum and maximum values. Significant differences (P ≤ 0.05; Mann–Whitney U test) are labeled: ∗ significant difference to group TMEV; # significant difference to group Treg↑TMEV; TMEV = untreated, infected mice; Treg↑TMEV = regulatory T cell expanded, infected mice; CD8↓TMEV = CD8⁺ T cell depleted, infected mice; Treg↑CD8↓TMEV = combined treated, infected mice. dpi = days post infection.

The presented data indicate that chronic TMEV‐infection in combined treated C57BL/6 mice stimulates innate immune responses mediated by microglia/macrophages and enhances lymphocyte recruitment to the spinal cord.

Early expansion of regulatory T cells impedes recovery of peripheral CD8⁺ T cell population during chronic virus infection

To elucidate whether peripheral immune responses are associated with delayed viral elimination in combined treated C57BL/6 mice, the dynamics of CD4⁺Foxp3⁺ Treg and CD8⁺ T cell populations in blood and spleen during the disease course were analyzed by flow cytometry.

IL‐2C application successfully elevated the frequency of circulating CD4⁺Foxp3⁺ Treg in blood and spleen samples at 3 dpi with and without concurrent CD8‐depletion (group Treg↑TMEV and Treg↑CD8↓TMEV; Figure 9A,C, Supporting Information Table S1). Phenotypic analyses revealed that antibody‐mediated CD8‐depletion significantly decreased the frequency of CD8⁺ T cells in the blood and spleen of TMEV‐infected C57BL/6 mice at 3 and 14 dpi (groups CD8↓TMEV and Treg↑CD8↓TMEV; Figure 9B,D, Supporting Information Table S1). Circulating CD8⁺ T cell numbers rebounded at 42 dpi in group CD8↓TMEV, resulting in frequencies similar to untreated control mice (group TMEV). Notably, blood CD8⁺ T cells were still significantly reduced in combined treated mice (group Treg↑CD8↓TMEV) compared to all other groups at 42 dpi, indicating a delayed recovery of the peripheral CD8⁺ T cell population following simultaneous Treg‐expansion (Figure 9D, Supporting Information Table S1). Similarly, significantly reduced CD8⁺ T cell frequencies compared to all other groups were found in the spleen of combined treated mice (group Treg↑CD8↓TMEV). Accordingly, the ratio of CD4⁺Foxp3⁺ Treg to CD8⁺ T cells (CD4⁺Foxp3⁺/CD8⁺ cell ratio) was increased in group CD8↓TMEV only until 14 dpi while it stayed elevated until 42 dpi in blood and spleen samples of combined treated mice (group Treg↑CD8↓TMEV) with significantly higher values compared to all other treatment groups (Figure 9E, Supporting Information Table S1). Interestingly, sole Treg‐expansion (group Treg↑TMEV) was also able to decrease the frequency of circulating CD8⁺ T cells compared to untreated mice (group TMEV) at 14 and 42 dpi in blood samples and at 42 dpi in spleen tissue, respectively. Moreover, significantly increased CD4⁺Foxp3⁺/CD8⁺ cell ratio values were detected in blood and spleen at 3 and 42 dpi of Treg‐expanded mice (group Treg↑TMEV).

Figure 9.

Flow cytometric analyses of T cell subsets in the blood and spleen of Theiler's murine encephalomyelitis virus (TMEV)‐infected C57BL/6 mice. A, B. Representative dot plots from an infected mouse (TMEV), regulatory T cell expanded, infected mouse (Treg↑TMEV), CD8⁺ T cell depleted, infected mouse (CD8↓TMEV), and combined treated, infected mouse (Treg↑CD8↓TMEV) at 3 days post infection (dpi) showing (A) CD4⁺Foxp3⁺ T cells and (B) CD8⁺ T cells. Note the (A) fivefold increase of CD4⁺Foxp3⁺ regulatory T cells after interleukin‐2 complex treatment (groups Treg↑TMEV and Treg↑CD8↓TMEV) and (B) marked reduction of CD8⁺ T cells following antibody‐mediated CD8‐depletion (groups CD8↓TMEV and Treg↑CD8↓TMEV). C. Percentage of CD4⁺Foxp3⁺ T cells in blood and (C′) spleen samples. D. Percentage of CD8⁺ T cells in blood and (D′) spleen samples. E. Ratio of CD4⁺Foxp3⁺ regulatory T cells to CD8⁺ T cells in blood and (E′) spleen samples. Note significantly increased values in combined treated mice. (C, D, E) Box plots display median and quartiles with minimum and maximum values. Significant differences (P ≤ 0.05; Mann–Whitney U test) are labeled: ∗ significant difference to group TMEV; # significant difference to group Treg↑TMEV; ○ significant difference to group CD8↓TMEV. TMEV = untreated, infected mice; Treg↑TMEV = regulatory T cell expanded, infected mice; CD8↓TMEV = CD8⁺ T cell depleted, infected mice; Treg↑CD8↓TMEV = combined treated, infected mice.

In order to determine the kinetics of peripheral T cells not expressing Foxp3 (non‐Treg), CD4⁺ and CD8⁺ T cells were gated on Foxp3⁻ cells. Data revealed that circulating CD4⁺ Foxp3⁻ T cells were significantly reduced in group Treg↑CD8↓TMEV mice compared to group TMEV mice (control) at 3 and 14 dpi (Supporting Information Figure S2A). Similarly, CD4⁺ Foxp3⁻ T cells were significantly decreased in spleens of group Treg↑TMEV animals at all time points, indicating a primary increase of Foxp3⁺ Treg following IL‐2C treatment (Supporting Information Figure S2B). At 42 dpi, a significant increase of splenic CD4⁺Foxp3⁻ T cells was detected in αCD8‐treated mice with and without Treg‐expansion (groups CD8↓TMEV and Treg↑CD8↓TMEV), suggestive of CD4⁺ effector T cell expansion associated with prolonged CNS infection. Except from a mild but significant increase of circulating CD8⁺ Foxp3⁻ T cells in Treg‐expanded infected mice (group Treg↑TMEV) compared to controls (group TMEV), dynamic changes of CD8⁺ Foxp3⁻ T cells in blood and spleen samples (Supporting Information Figure S2C,D) were similar to those found for entire CD8⁺ T cell populations (Figure 9D), showing only a minor effect of treatment on regulatory CD8⁺Foxp3⁺ T cells.

B cells in spleens were quantified by immunohistochemistry. Densiometric analyses revealed an increase of CD45R⁺ B cells in αCD8‐treated mice with and without Treg‐expansion (groups CD8↓TMEV and Treg↑CD8↓TMEV), suggestive of enhanced humoral immune responses triggered by increased CNS infection (Supporting Information Figure S3).

Taken together, analyses demonstrate the ability of a transiently expanded Treg‐compartment to induce long‐lasting disturbances of the CD8⁺ T cell compartment, which potentially promote a chronic CNS infection in combined treated, TMEV infected C57BL/6 mice.

Depletion of CD8⁺ T cells potentiates regulatory T cell expansion in peripheral lymphoid organs

Above data indicate that CD8‐depletion leads to increased Foxp3⁺ Treg responses in the CNS of TMEV‐infected mice. To determine whether αCD8‐application enhances IL‐2C‐mediated Treg‐expansion in peripheral lymphoid organs, the primary treatment effect was investigated under non‐infectious (steady state) conditions (Figure 10A–D).

Figure 10.

Flow cytometric analyses of T cell subsets in the blood of non‐infected C57BL/6 mice (steady state conditions). A, B. Representative dot plots from an untreated mouse (control), regulatory T cell expanded mouse (Treg↑), and combined treated mouse (Treg↑CD8↓TMEV) at day 3 after treatment, showing (A) CD8⁺ T cells and (B) CD4⁺Foxp3⁺ T cells. Note (A) marked reduction of CD8⁺ T cells following antibody‐mediated CD8‐depletion (group Treg↑CD8↓) and (B) four‐ to fivefold increase of CD4⁺Foxp3⁺ regulatory T cells after Treg‐expansion (groups Treg↑ and Treg↑CD8↓). C. Percentage of CD8⁺ T cells and D. CD4⁺Foxp3⁺ T cells in blood samples. Box plots display median and quartiles with minimum and maximum values. Significant differences (P ≤ 0.05; Mann–Whitney U test) are labeled: ∗ significant difference to control group; # significant difference to group Treg↑; control = untreated mice; Treg↑ = regulatory T cell expanded mice; Treg↑CD8↓ = combined treated mice.

Reduced frequencies of circulating CD8⁺ T cells were found at 3, 7, and 14 days after αCD8‐application (group Treg↑CD8↓, Figure 10A,C, Supporting Information Table S3). As described before 6, 97, IL‐2C treatment rapidly increases the yield of CD4⁺Foxp⁺ Treg in the blood at day 3 (group Treg↑, Figure 10B,D). Strikingly, CD8‐depletion significantly enhances the frequency of CD4⁺Foxp⁺ Treg at 3 dpi (group Treg↑CD8↓). Data show the ability of concurrent CD8‐depletion to transiently enhance peripheral Treg‐expansion induced by IL‐2C treatment in C57BL/6 mice.

In addition, immunohistochemistry performed at day 3 revealed a significant increase of Foxp3⁺ cells in lymphoid organs in combined treated mice (group Treg↑CD8↓) compared to control animals (spleen, thymus, and cervical lymph node) and animals treated only with IL‐2C (group Treg↑; spleen and cervical lymph node), indicating generalized Treg‐expansion in the peripheral immune system (Supporting Information Table S3; Supporting Information Figure S4).

DISCUSSION

CD8‐mediated cytotoxicity critically controls virus infection and is crucial for disease resistance in the TME model 4, 62. Results of the present study reveal that early Treg‐expansion maintains virus infection in the brain in combination with CD8‐depletion, thereby promoting virus spread to the spinal cord and leukomyelitis in TMEV‐IDD‐resistant C57BL/6 mice.

Besides eliminating TMEV directly 1, 21, 70, CD8⁺ T cell responses seem to reduce the ability of Treg to inhibit antiviral immunity in this MS model. In support of this notion, strong antigenic stimulation of cytotoxic CD8⁺ T cells has been shown to limit the suppressive capacity of Treg in an OT‐I‐mediated, oligodendrocyte directed ex vivo mouse model 22. In TMEV‐IDD‐susceptible SJL mice, adoptive transfer of Treg during the acute TME phase leads to increased viral replication, decreased CNS leukocyte recruitment and clinical deterioration 50, 70, while functional inactivation of Treg leads to an increase of TMEV‐specific IFN‐γ producing CD8⁺ T cells 70. However, in TMEV‐IDD‐resistant C57BL/6 mice, adoptive transfer (“gain of function” approach) or genetic ablation and antibody‐mediated inactivation of Treg (“loss of function” approaches) have failed to influence viral load in previous studies 50, 67, 70. In agreement with these reports, the present study showed that in vivo Treg‐expansion delayed viral elimination only if CD8⁺ cytotoxic T cells were depleted simultaneously.

Simultaneous CD8‐depletion and Treg‐expansion (combined treatment) caused prolonged infection of pyramidal neurons of the hippocampus of C57BL/6 mice. Associated cellular responses and pro‐inflammatory cytokine (IL‐1α, IL‐2, TNF‐α) expression in the brain might initiate antiviral immunity but probably also hippocampal damage 11, 13, 35. A variety of viruses, such as Borna disease virus, measles virus, neuroadapted Sindbis virus, herpes virus, rabies virus and human immunodeficiency virus have the ability to infect and damage hippocampal neurons, leading to disturbances in the excitatory network and memory system 8, 9, 34. Notably, also picornaviruses (eg, Coxsackievirus A21) are currently discussed to cause subclinical brain infection and hippocampal damage in human beings 9, 48. TMEV‐infection of C57BL/6 mice has become a valuable model to investigate virus‐induced hippocampal damage and infection‐induced epilepsy 43, 78, 79, 80. Hippocampal infection is also associated with impaired cognitive ability, anxiety‐like behavior and memory impairment of TMEV‐infected C57BL/6 mice 9, 51, 92. Recently, it has been shown by Broer et al that the TMEV‐BeAn substrain used in the present experiment did not elicit epileptic seizures in C57BL/6 mice 7. Similarly, we did not observe overt seizures in the present study. However, targeted diagnostic methods such as video/EEG monitoring and behavioral tests (eg, Morris water maze) are needed to detect subtle clinical changes and fully discover the functional relevance especially of prolonged hippocampal infection in C57BL/6 mice in future studies.

Our data reveal that Treg‐expansion is associated with an accelerated recruitment of Foxp3⁺ T cells to the hippocampus together with an increased Foxp3 mRNA expression during the acute TME phase, which is potentiated by simultaneous CD8‐depletion (group Treg↑CD8↓TMEV). Explanations for this effect include an increased peripheral expansion of CD4⁺Foxp3⁺ Treg following CD8‐depletion, as observed in blood and lymphoid organs, as well as an enhanced Treg migration to the brain probably by an altered chemokine milieu in the hippocampus and/or increased expression of inflammatory homing receptors on Treg 98. For instance, CCR5⁺ Treg accumulate preferentially in the brain of Japanese encephalitis virus‐infected mice 33 and CCR6 expression enhances the CNS influx of Treg in experimental autoimmune encephalomyelitis 38, 96. Increased Treg responses in the CNS have the potential to alter the local environment, in which antiviral immunity is reduced by the release of inhibitory cytokines, like IL‐10 and TGF‐β, as demonstrated by RT‐qPCR. Furthermore, Treg‐mediated cytokine consumption (IL‐2 deprivation), disturbed antigen presenting cell function, and apoptosis of effector T cells might contribute to diminished protective immunity 31, 36, 53, 66, 95.

Development of robust CD8‐mediated cytotoxicity directed against dominant TMEV epitopes early after infection accounts for virus clearance and protection from demyelinating disease in resistant mouse strains 20, 21. Genetic depletion of CD8⁺ T cells has been shown to reduce antiviral immunity, which leads to TMEV persistence, prolonged inflammation and demyelination in the spinal cord of mice with a C57BL/6 background 21, 62, 71, 72. Genetic β₂‐microglobulin deficiency, which results in disturbed MHC class I‐restricted CD8⁺ T cell responses, also predisposes resistant mouse strains to develop TMEV persistence and myelin loss 71, 72. Moreover, similar to our findings, antibody‐mediated depletion of CD8⁺ T cells prior to TMEV infection has been shown to diminish viral clearance causing an increased severity of demyelinating disease in susceptible SJL mice and the presence of small demyelinating lesions in resistant C57BL/10SNJ mice, respectively 4, 73. Strikingly, additional Treg‐expansion significantly increases spinal cord infection and inflammation in αCD8‐treated mice, indicative of enhanced virus spread within the CNS. Myelitis and myelin loss in combined treated C57BL/6 mice are associated with a preferential infection of the spinal cord white matter, resembling initial white matter lesions in SJL mice 26, 63, 87. The switch of virus cell tropism from neurons to oligodendrocytes and microglia/macrophages in the white matter is a prerequisite for TMEV‐IDD in susceptible mouse strains 41. Although compared to white matter lesions in TMEV‐IDD‐susceptible SJL mice 25 relatively small‐sized myelin lesions and mild cellular infiltrates were found in the spinal cord, our data revealed that chronic infection renders C57BL/6 mice susceptible to develop demyelinating leukomyelitis and axonal damage. The initiation of axon degeneration and axonal transport defects in TME is supposed to be virus‐induced and might represent an active self‐destructive response to limit further virus spread along spinal cord axons in susceptible mice. Accordingly, axonal damage due to prolonged virus infection might have contributed to myelin loss (inside‐out hypothesis) in the present study. Similar to MS patients, axonal damage in infected mice is regarded to account for functional disability and has the ability to trigger immune‐mediated myelin damage 39, 84, 86, 87. Phenotyping of spinal cord infiltrates together with cytokine analyses also argues for the presence of CD4⁺ T cell‐mediated immunopathology in combined treated C57BL/6 mice. Moreover, release of myelinotoxic factors by activated CD107b⁺ microglia/macrophages (bystander effects) following virus infection might have led to myelin and axonal damage 15.

Similar to spinal cord lesions in SJL mice following TMEV‐infection 90, increased infiltration of CD3⁺ T cells, CD45R⁺ B cells, and CD107b⁺ microglia/macrophages were found predominantly in the spinal cord white matter of combined treated animals at late disease stage (42 dpi); however, the amount of recruited immune cells was comparatively lower. Prolonged leukocyte infiltration was associated with an increased expression of the pro‐inflammatory cytokines IL‐1α and TNF‐α. IL‐1α is upregulated by microglia/macrophages in a variety of CNS disorders and initiates protective responses against certain viral infections, including TMEV‐infection. However, excessive IL‐1 signaling has been shown to induce pathogenic Th17 responses, rendering C57BL/6 mice susceptible to TMEV‐IDD 32, 68. Likewise, dual—probably disease phase dependent—functions are described also for TNF‐α in TME with protective antiviral effects on the one hand and promoting Th1‐mediated immunopathology on the other 28, 63, 74, 77. Increased Foxp3⁺ Treg recruitment together with enhanced IL‐10 levels in the spinal cord of combined treated mice is suggestive of compensatory attempts to sustain CNS homeostasis and prevent collateral tissue damage, as observed in different autoimmune and infectious diseases 2, 37. Noteworthy, Treg have been shown to increase IL‐10 production and reduce inflammatory demyelination during chronic disease in TMEV‐infected SJL mice 50, 70.

Similar to the present findings, a protective role of CD8⁺ T cells because of their contribution to viral clearance has been described in SJL mice infected with the BeAn strain, determined by antibody‐mediated CD8‐depletion during early and late TME 4. A recovering CD8⁺ T cell population following antibody treatment might re‐establish protective antiviral immunity, which can explain the rather mild spinal cord inflammation in combined treated C57BL/6 mice compared to fully susceptible SJL mice in the chronic phase (42 dpi). Strikingly, Treg‐expansion reduces the frequency of circulating CD8⁺ T cells in C57BL/6 mice (group Treg↑TMEV) and delays the recovery of peripheral CD8⁺ T cells and their recruitment to the spinal cord in combined treated mice (group Treg↑CD8↓TMEV). Treg have the ability to reversibly suppress cytotoxic T cell expansion, for example, via TGF‐β signaling and IL‐2 deprivation 54, 55, 66. In infectious diseases, Treg directly inhibit virus‐specific effector CD8⁺ T cells as demonstrated in friend retrovirus‐infected C57BL/6 mice 100. Moreover, depletion of Treg in DEREG mice combined with antibody‐mediated blockade of inhibitory receptors (PD‐1 ligand, TIM‐3) reactivates exhausted CD8⁺ T cells and efficiently reduces the chronic load of friend retrovirus 16. Thus it can be speculated that strengthened Treg‐responses following IL‐2C treatment (Treg‐expansion) cause delayed recovery of peripheral CD8⁺ T cell population and impaired development of TMEV‐specific effector T cell responses in αCD8‐treated C57BL/6 mice. Referring to this, early rapid expansion of Foxp3⁺ cells in SJL mice has been shown to induce a disturbed ratio of CD8⁺ effector T cells to Treg, which is supposed to cause diminished TMEV‐specific immune responses and virus persistence in TME‐IDD susceptible mouse strains 70.

In conclusion, results revealed synergistic effects of Treg‐expansion and cytotoxic T cell‐ablation to delay viral elimination in C57BL/6 mice, which partly mimics the immunological situation and neuropathology in TMEV‐IDD‐susceptible mouse strains. Reduced antiviral immunity can exacerbate CNS infection, potentiating the risk of immune‐mediated myelin damage via bystander demyelination and epitope spreading 49. Therefore, novel Treg‐based approaches to treat diseases with an infectious etiology or those that develop in parallel with viral infection are at increased risk of disease exacerbation, particularly in individuals with inherited or acquired CD8‐deficiencies and impaired protective cytotoxicity 81, 99.

Supporting information

Additional Supporting Information may be found in the online version of this article at the publisher's web‐site:

Table S1. Results of statistical analyses (P‐values)—TMEV‐infected animals.

Table S2. Results of statistical analyses (P‐values)—Comparison between combined treated, TMEV‐infected C57BL/6 mice (group Treg↑CD8↓TMEV) and TMEV‐infected SJL mice (group SJL TMEV).

Table S3. Results of statistical analyses (P‐values)—non‐infected mice.

Figure S1. Comparison of inflammatory cell numbers in the spinal cord of SJL mice and combined treated C57BL/6 mice following Theiler's murine encephalomyelitis virus (TMEV) infection. (A) Quantification of CD3⁺ T cells, (B) CD45R⁺ B cells, (C) CD107b⁺ microglia/macrophages, and (D) Foxp3⁺ regulatory T cells in the spinal cord by immunohistochemistry. Note significantly lower numbers of T cells, B cells and microglia/macrophages in combined treated C57BL/6 mice (group Treg↑CD8↓TMEV) compared to susceptible SJL mice. Box plots display median and quartiles with minimum and maximum values. Significant differences (P ≤ 0.05; Mann–Whitney U test) are labeled with an asterisk, ns = not significant; Treg↑CD8↓TMEV = combined treated, infected C57BL/6 mice; SJL TMEV = infected SJL mice.

Figure S2. Flow cytometric analyses of Foxp3⁻ T cell subsets in the blood and spleen of Theiler's murine encephalomyelitis virus (TMEV)‐infected C57BL/6 mice. (A,B) Percentage of CD4⁺ cells gated on Foxp3⁻ cells in (A) blood and (B) spleen samples. (C,D) Percentage of CD8⁺ cells gated on Foxp3⁻ cells in (C) blood and (D) spleen samples. Box plots display median and quartiles with minimum and maximum values. Significant differences (P ≤ 0.05; Mann–Whitney U test) are labeled: ∗ significant difference to group TMEV; # significant difference to group Treg↑TMEV; ○ significant difference to group CD8↓TMEV. TMEV = untreated, infected mice; Treg↑TMEV = regulatory T cell expanded, infected mice; CD8↓TMEV = CD8⁺ T cell depleted, infected mice; Treg↑CD8↓TMEV = combined treated, infected mice.

Figure S3. Quantification of CD45R⁺ B cells in the spleen of Theiler's murine encephalomyelitis virus (TMEV)‐infected C57BL/6 mice. (A) Percentage of area in the spleen stained by CD45R‐specific immunohistochemistry. Box plots display median and quartiles with minimum and maximum values. Significant differences (P ≤ 0.05; Mann–Whitney U test) are labeled: ∗ significant difference to group TMEV; TMEV = untreated, infected mice; Treg↑TMEV = regulatory T cell expanded, infected mice; CD8↓TMEV = CD8⁺ T cell depleted, infected mice; Treg↑CD8↓TMEV = combined treated, infected mice.

Figure S4. Quantification of Foxp3⁺ regulatory T cells in lymphoid organs of non‐infected C57BL/6 mice (steady state conditions). Analysis of Foxp3⁺ regulatory T cells in the (A‐C) spleen, (D‐F) thymus and (G‐I) cervical lymph node. Representative images of (B,E,H) an untreated control mouse and (C,F,I) combined treated mouse (group Treg↑CD8↓) at day 3. Note marked increase of labeled cells (brown signal) in all organs in combined treated animal. (A,D,G). Box plots display median and quartiles with minimum and maximum values. Significant differences (P ≤ 0.05; Mann–Whitney U test) are labeled: ∗ significant difference to control group; # significant difference to group Treg↑; control = untreated mice; Treg↑ = regulatory T cell expanded mice; Treg↑CD8↓ = combined treated mice.