Regulatory T Cells in Central Nervous System Injury: A Double-Edged Sword

Acute injury to the CNS evokes cellular and molecular responses that lead to secondary neurodegeneration, a process of sustained neuronal degeneration (1). Accompanying this period of secondary degeneration is a coordinated immune response to the trauma, including chemotaxis of microglia to ATP released from the damaged cells (2) and directed migration of both the innate and adaptive immune cells to the injury site due to chemokine signals (3). The dogma that this infiltration of immune cells into the injury site was a detrimental response has been challenged by the finding that neuronal survival could be improved by boosting T cell activity rather than by its suppression (4–6), although the phenotype of protective T cells after CNS injury, and particularly the role of regulatory T (T_reg) cells in this process, is still a matter of debate (7–10).

Naturally occurring T_reg cells, which express the transcription factor Foxp3 (11–13), have been intensively studied for their ability to suppress adaptive immune responses (14–17). This subset of T cells, which develops with high avidity to self-Ags, is especially important in controlling autoimmunity (18). Therefore, it has been proposed that T_reg cells mediate their actions by attenuating both protective and inflammatory postinjury immune responses, and thus either exacerbating (19) or ameliorating (20) neuronal degeneration. Despite these studies, the exact mechanism of their action in the injured CNS remains unclear.

Recently, the heterogeneity of macrophages has come to light, with two general classes being described as classically or alternatively activated (21). Although classically activated macrophages express high levels of proinflammatory cytokines such as TNF and IL-1β, and exhibit a robust respiratory burst (22), alternatively activated (tissue-building) macrophages express high levels of arginase-1 and several factors that play a role in promoting tissue homeostasis and recovery from insults (23). Several studies have shown the neuroprotective ability of alternatively activated macrophages in CNS injury (24–26), but what leads to, and sustains, this phenotype is unclear in the context of CNS trauma.

In this study, we show that the regulation of the T cell response to CNS injury is taking place in the draining deep cervical lymph nodes (dCLN) rather than at the site of injury. In line with this, surgical resection of the dCLN results in impaired neuronal survival. We show that removal of T_reg cells that leads to exaggerated response of T effector (T_eff) cells is associated with reduction in alternatively activated macrophages at the site of injury and leads to impaired neuronal survival. Exogenous supply of activated T_reg cells, however, results in suppression of neuroprotective IL-4–producing T cells, and consequently also results in suppression of alternatively activated macrophages at the site of injury. Thus, both depletion and addition of T_reg cells are detrimental for neuronal survival after injury through regulation of macrophage phenotype.

Female C57BL/6 (stock 000664) and UBC-GFP (stock 004353) mice were purchased from The Jackson Laboratory (Bar Harbor, ME); DEREG mice were a gift of T. Sparwasser (Institute of Infection Immunology, Twincore, Germany) (16); KN2 mice were a gift of M. Mohrs (Trudeau Institute, Sarnac Lake, NY) (27). All animals were housed in temperature- and humidity-controlled rooms, maintained on a 12-h/12-h light/dark cycle (lights on 7:00 A.M.), and age matched in each experiment. All strains were kept in identical housing conditions. All procedures complied with regulations of the Institutional Animal Care and Use Committee at the University of Virginia.

A total of 200 μg all-trans retinoic acid (ATRA; Fisher) was dissolved in corn oil and injected i.p. every other day starting 3 d before injury. Diphtheria toxin (DTx) was dosed at 40 μg/kg in PBS and was injected into C57BL/6 or DEREG mice 2 d before optic nerve injury and on the day of optic nerve injury. A total of 250 μg anti-CD25 (clone PC-61) was injected into mice 8 d before optic nerve injury.

Mice were anesthetized, and the skull was exposed and immobilized in a stereotactic device. Holes were drilled in the skull above the superior colliculus (bilaterally 2.9 mm caudal to bregma and 0.5 mm lateral to midline). A quantity amounting to 1 μL 4% Fluoro-gold was injected 2 mm below the meningeal surface at a rate of 0.5 μl/min using a Hamilton syringe and an automatic injector. The dye was allowed to diffuse into the tissue for 1 min before the syringe was removed. The scalp was then sutured closed, and the mice were allowed to recover on warming pads at 37°C before returning them to their cage.

Mice were subjected to an optic nerve injury 3 d after stereotactic surgery. Briefly, mice were anesthetized with a 1:1:8 mixture of ketamine:xylazine:saline. An incision was made in the connective tissue above the sclera. The venous sinus around the optic nerve was retracted to expose the optic nerve, and the nerve was crushed using N5 self-closing forceps 2 mm behind the globe for 3 s. The mice were then allowed to recover at 37°C on a warming pad before returning to their cages.

Mice were enucleated, and the cornea removed at the corenal limbus. The lens and the underlying vitreous were removed with forceps. The retina was separated from the sclera and pigment epithelium. Four cuts were made toward the optic disc, and the retina was mounted on nitrocellulose paper and fixed in 4% paraformaldehyde overnight. Pictures of all four quadrants of the retina were taken at equal distances from the optic disc of the retinas using an Olympus IX-71 microscope. The pictures were then counted by a blinded observer to determine the number of retinal ganglion cells (RGCs).

Mice were anesthetized with a 1:1:8 mixture of ketamine:xylazine:saline. A 10-mm incision was made midline above the trachea. Salivary glands and sternocleidomastoid muscles were retracted bilaterally to expose the dCLN. dCLN were removed using Dumont forceps, and the skin was sutured closed. Mice that received sham surgery had their dCLN exposed, and then the skin was sutured. The mice were allowed to recover on a 37°C warming pad before returning to their cages. Mice were allowed to recover from surgery for at least 2 wk before optic nerve injury.

C57BL/6 mice underwent split-dose lethal irradiation (350 rad, then 950 rad 48 h later). Bone marrow cells were isolated by flushing out the femur and tibia of UBC-GFP mice. A total of 1 × 10⁷ bone marrow cells was injected into the irradiated mice 3 h after the last irradiation. Mice were allowed to reconstitute for 6 wk before being used in experiments.

For T_eff cultures, total lymph nodes were dissected, and a single-cell suspension was made by mashing through a 70-μm mesh. A total of 3 × 10⁶ cells/ml was incubated in T cell culture media supplemented with 1 μg/ml anti-CD3 (clone 145-2C11; American Type Culture Collection stock CRL-1975; Ab grown and isolated by UVA lymphocyte culture center) and 1 μg/ml anti-CD28 (clone 37.51; Bioxcell, stock BE0015-1). For T_reg cells in vitro (iT_reg) cultures, the media was supplemented with 10 nM ATRA (Fisher), 5 ng/ml TGF-β (PeproTech), and 250 U/ml IL-2 (R&D Systems). The cultures were maintained for 5 d before CD4⁺ T cells were isolated using magnetic bead separation (Miltenyi Biotec) and injected i.v. into C57BL/6J mice.

Axillary and inguinal lymph nodes or dCLN were isolated and mashed through a 70-μm strainer in PBS containing 1% BSA and 2 mM EDTA. The following Abs were used, and are all from eBioscience, unless otherwise noted: Foxp3-Alexa 488, CD4-PerCP Cy5.5, TCR-β allophycocyanin eFluor780, CD45-allophycocyanin, CD8-eFluor 450, CD25-PE (BD Bioscience), IL-4 PE, IFN-γ allophycocyanin, and CD69-PE Cy7. For intranuclear staining, the cells were fixed overnight in Foxp3 Fix/Perm buffer (eBioscience) before incubating with Foxp3 Ab in FACS buffer containing 0.3% saponin (Fisher). For intracellular staining, mice were injected with 3 μg brefeldin A 5 h before they were sacrificed. The cells were stained for extracellular markers before they were fixed in the fixation buffer (eBioscience) for 30 min, then permeabilized and stained for intracellular Ags in FACS buffer containing 0.3% saponin.

Bone marrow was isolated from wild-type mice and cultured on untreated petri dishes in DMEM/F12 containing 10 ng/ml M-CSF (eBioscience), 10% FCS, l-glutamine, and pen-strep (Invitrogen). The media was changed every 3 d, and macrophages were used after 8 d in vitro. The day before the macrophages were used, they were replated on tissue culture–treated 24-well plates. CD4⁺ T cells from injured or uninjured deep cervical or skin draining lymph nodes (SDLN) were isolated using magnetic bead separation (Miltenyi Biotec) and incubated at 1 × 10⁶ cells/well in complete macrophage media. Twenty-four hours after addition of T cells, the macrophages were washed five times with PBS to remove the nonadherent T cells, and RNA was isolated from the macrophages.

For arginase-1 staining, mice were perfused transcardially with ice-cold PBS containing 4 U/ml heparin and then with 4% paraformaldehyde. Eyes were enucleated and frozen on dry ice in OCT. The 10-μm sections were cut on a Lyca cryostat and mounted on gelatin-coated slides. Sections were then stained for arginase-1 (Santa Cruz Biotechnology; clone V20), CD68 (BioLegend; clone FA11), Iba1 (Biocare Medical; polyclonal), and GFP (Abcam; polyclonal). For CD4 and CD11b staining, mice were perfused transcardially with ice-cold PBS containing 4 U/ml heparin. Eyes were enucleated and frozen on dry ice in OCT. The 10-μm sections were cut on a Lyca cryostat and mounted on gelatin-coated slides. Slides were postfixed in 3:1 acetone:ethanol at 4° before staining with the flowing Abs: CD4-FITC (eBioscience; clone GK-1.5), CD11b (BioLegend; clone M1/70), and Foxp3-biotin (eBioscience; clone FKJ-16s). For Foxp3 and CD4 costaining, CD4 was detected with an anti-fluorescein secondary Ab (Life Technologies) and FoxP3 was detected with Alexfluor 594–conjugated streptavidin (Jackson Immunochemical).

To determine where the immune response to CNS injury was occurring, we first examined CNS-draining dCLN as compared with SDLN (axillary and inguinal) for T cell activation and proliferation upon CNS injury. We found an increase in the number and percentage of CD4⁺ T cells and a concurrent reduction in the percentage of CD8⁺ T cells in CNS-draining dCLN (Fig. 1A–C). No change in the number or percentage of CD4⁺ T cells was observed in the SDLN (Fig. 1D–F). When the induced CD4⁺ T cells were examined for subpopulation (T_reg versus T_eff), both activated T_eff (CD4⁺CD25⁺Foxp3⁻) and T_reg (CD4⁺Foxp3⁺) cells were increased in the dCLN after the injury (Fig. 1G–I), but not in the SDLN (Fig. 1J–L). To determine whether the immune response in the dCLN was playing an important role in the response to CNS injury, we used an optic nerve injury model, in which RGCs are prelabeled with the neuronal tracer Fluoro-gold and then the optic nerve is injured and the number of surviving RGCs in the retina is quantified (Fig. 1M). This injury leads to a decrease in the number of RGCs in mice that underwent the dCLN removal from those that received a sham surgery (Fig. 1N), whereas their contralateral uninjured retinas did not display a loss of RGCs (Fig. 1O).

To determine whether T cells from the injured dCLN displayed a different phenotype after CNS injury than the SDLN, we used flow cytometry to analyze the intracellular cytokines produced in the lymph nodes after injury. T cells from the dCLN displayed higher levels of IL-4 after optic nerve injury than those from the SDLN (Fig. 2A, 2B). To examine whether this phenotype is induced by the injury, we used KN2 reporter mice that express human CD2 (hCD2) when IL-4 is being translated. In these mice, hCD2 expression is seen in CD44⁺ memory T cells (Fig. 2C), and this hCD2 expression is induced after injury in the dCLN after injury, but not in the SDLN (Fig. 2D). To determine whether T cells induced after optic nerve injury in the draining lymph nodes are capable of supporting alternative activation of macrophages, we isolated T cells from injured and uninjured dCLN and SDLN and cocultured them with a pure population of bone marrow–derived macrophages. T cells from the dCLN of optic nerve–injured mice were able to support an alternative activation phenotype of bone marrow macrophages in vitro, whereas T cells obtained from SDLN of injured mice or from dCLN of uninjured mice were unable to promote this alternatively activated (tissue-building) phenotype of macrophages (Fig. 2E). This suggests that the injury indeed induces T_eff cells in the draining dCLN that are capable of promoting a neuroprotective macrophage phenotype.

Because we observed that T cells in the draining lymph node were able to promote an alternative activation of macrophages, we addressed a possibility that T cells are controlling the phenotype of the infiltrating monocytes/macrophages. Arginase-1–expressing macrophages (M2 type) have been previously described to support neuronal survival after CNS injury (24–26, 28). Indeed, using immunohistochemistry of injured optic nerves, we demonstrate that arginase-1 expression is induced in the injury site (Fig. 2F), whereas there is no detectable arginase-1 staining in the uninjured optic nerves (Supplemental Fig. 1A). We established GFP⇒C57BL/6 bone marrow chimeric mice (29), whose peripheral immune system is replaced by the GFP⁺ bone marrow, but that have a significant number of GFP⁻ microglia in the optic nerve (Supplemental Fig. 1B). Therefore, despite the issues inherent with bone marrow transplantation following irradiation (30), the large majority of the engrafted Iba1-positive cells in the uninjured CNS of chimeric mice are of non-GFP origin. It is, therefore, conceivable to assume that an increase of GFP⁺ cells after injury results primarily from their peripheral recruitment of monocyte-derived macrophages, although, as stated above, the procedure has its limitation. In the chimeric mice, most of the GFP⁺Iba1⁺ cells in the site of the injury were arginase-1 positive, suggesting that at least the majority of the infiltrating cells were highly skewed after injury. However, significantly fewer of the radio-resistant GFP⁻Iba1⁺ microglia were arginase-1 positive, suggesting that macrophages infiltrating from the periphery are the primary source of alternatively activated myeloid cells (Fig. 2G, 2H). This preferential skew of infiltrating myeloid cells suggests that their phenotype switch took place in the periphery prior to infiltration rather than in the CNS parenchyma. These results are in line with previous findings, suggesting that monocytes with an alternatively activated phenotype are arriving from a periphery through a unique path into the injured CNS (25).

The contribution of different subsets of T cells to neuronal survival after CNS injury has been intensively studied (4, 5, 19, 31, 32), yet their role in this postinjury neuronal survival remains controversial (8, 10, 33). Because T_reg cells are known to exert asymmetric control of T cell responses in nonpathological situations (34), we tested the hypothesis that T_reg cells were responding to injury in the draining lymph nodes, where they controlled the phenotype of T_eff cells. We used DEREG mice (16), which express the DTx receptor under the Foxp3 promoter to assess the effect of T_reg depletion on neuronal survival. Treatment of these mice with 40 μg/kg DTx 2 d before injury completely eliminates T_reg cells in the bloodstream (Supplemental Fig. 2A). Seven days after injury, the DEREG mice treated with DTx still displayed decreased numbers of T_reg cells in their CNS-draining dCLN (Fig. 3A) and an increase in the number of activated T_eff cells in the dCLN and SDLN (Fig. 3B, Supplemental Fig. 2B).

To test the effect of T_reg cell depletion on CNS injury, we again used the optic nerve crush injury model. As expected from previous studies (35), DEREG mice treated with DTx, and thus depleted of T_reg cells, showed a decrease in the number of surviving RGCs 7 d after injury, as compared with wild-type mice treated with DTx (Fig. 3C). We examined the contralateral retina of injured mice (Supplemental Fig. 2C) and histological sections of uninjured mice treated with DTx (Supplemental Fig. 2D), which did not display any loss of RGCs or immune cell infiltrate, suggesting DTx by itself did not have destructive effects on uninjured CNS tissue. Furthermore, there was no difference in neuronal survival in C57BL/6 mice treated with saline or DTx (Supplemental Fig. 2E), confirming that DTx treatment was not causing nonspecific effects at the dose that we are using. To further establish the role of depletion of T_reg cells, we used an anti-CD25 Ab, which depletes T_reg cells (that express high levels of CD25) (36). In mice that were T_reg depleted, there was a decrease in the number of CD25⁺Foxp3⁺ Treg cells in the dCLN even 7 d after optic nerve injury and a corresponding decrease in neuronal survival (Supplemental Fig. 2F, 2G), consistent with the results seen in DEREG mice.

Although no change in overall numbers of CD4⁺ T cells (Fig. 3D) or CD11b⁺ myeloid cells (Fig. 3E) at the site of injury was found, the phenotype of accumulated macrophages was altered in DEREG mice treated with DTx. A significant decrease in arginase-1–expressing CD68 (a marker of activated macrophages) cells was evident (Fig. 3F, 3G), suggesting a decrease in alternatively activated macrophages (M2 type) after injury in Treg-depleted mice. No difference in the total amount of CD68⁺ area was detected (Supplemental Fig. 3A). To confirm the histological observations, we also examined the injured tissue by PCR. The mRNA expression of arg1, the gene for arginase-1, was reduced in DEREG mice, confirming the decrease in alternatively activated macrophages after T_reg depletion (Fig. 3H).

A complete depletion of T_reg cells using DEREG mice resulted in impaired outcome of CNS injury in our optic nerve crush injury model (Fig. 3C). However, the question still remains whether increased activity of T_reg cells would conversely offer a benefit after CNS injury. First, we tested the physiological outcome of T_reg manipulation via potentiation of T_reg-suppressive function by treating mice with ATRA, which induces differentiation of T_reg cells (37), stabilizes the T_reg phenotype (38), and makes T_reg cells more suppressive (38). As expected, treatment of mice with ATRA increased the T_reg population in the dCLN after injury (Fig. 4A), but surprisingly not in the SDLN (Supplemental Fig. 4A), and resulted in a decrease of activated T_eff cells in the dCLN and SDLN (Fig. 4B, Supplemental Fig. 4B). Interestingly, and in line with some reports (19) but contrary to other previous findings (7), mice treated with ATRA exhibited decreased neuronal survival compared with vehicle-treated mice (Fig. 4C), suggesting that induction of highly suppressive T_reg cells limits the protective T_eff responses. To rule out a possible in vivo effect of ATRA on cells other than T cells, we differentiated iT_reg using ATRA and TGFβ (Supplemental Fig. 4C). Injection of iT_reg cells into CNS-injured mice also resulted in an increase in T_reg cells (Fig. 4D) and attenuation of their activated T_eff response to injury in the dCLN (Fig. 4E), but no change in the number of T_reg cells and T_eff cells in the SDLN (data not shown) and a reduction in neuronal survival (Fig. 4F). Mice treated with T_eff cells (that were activated without TGF-β and ATRA, and that contained only ∼3% Foxp3⁺ T_reg compared with ∼85% in T_reg-designated culture conditions) did not show any change in neuronal survival (Fig. 4F), possibly due to the large number of T_eff cells already present in wild-type mice.

Interestingly, 7 d after T_reg injection, only very few of the injected cells were found in the dCLN, despite a substantial increase in overall T_reg numbers (Supplemental Fig. 4D). These results suggest that the injected T_reg cells induce endogenous T cell differentiation, possibly through their high levels of secreted and membrane-bound TGF-β (39, 40).

Because T_reg cells are exerting a negative effect on the outcome to CNS injury, we sought to determine whether T_reg cells were also gaining access to the site of injury. Upon injury, there is the influx of T_eff cells to the CNS parenchyma (Fig. 5A). Despite being able to visualize T_reg cells using Foxp3 immunolabeling in the spleen (Fig. 5B) and in spinal cords that had been injected directly with T_reg cells (Fig. 5C), we did not see T_reg cells in the parenchyma of the injured optic nerve in animals after exogenous i.v. injection of T_reg cells (Fig. 5D) or in injured wild-type mice, wild-type mice injected with T_eff cells, DEREG mice treated with DTx, and wild-type mice treated with DTx (data not shown). Furthermore, there was no difference in the number of CD4⁺ T_eff cells at the injury site of T_reg-treated mice (Fig. 5E). Therefore, it seems unlikely that acute manipulation of T_reg cells in our experimental paradigms is affecting neuronal survival through the migration of T_eff cells into the site of the injury. Next, to determine whether addition of T_reg cells could affect monocyte migration to the injured CNS, we quantified the number of CD11b⁺ cells accumulating at the site of injury. As with CD4⁺ T cells, there was no change in the number of CD11b⁺ cells that migrated to the site of injury in T_reg-treated mice (Fig. 5F, 5G), further suggesting that T_reg cells are not controlling immune cell migration in this injury model.

To determine whether T cell–derived cytokines are affected by T_reg manipulation after CNS injury, we examined the mRNA expression of genes from the optic nerve of injured mice treated with either T_eff or T_reg cells. Mice treated with T_reg cells display a dramatic decrease in the amount of IL-4 mRNA compared with mice treated with T_eff cells (Fig. 6A), suggesting a change in the Th2 response to damage at the injury site with T_reg treatment. To determine whether these changes are having effects downstream on myeloid cells after CNS injury, we examined markers of myeloid skewing in the injured optic nerves of mice treated with T_reg cells. Indeed, T_reg-treated mice displayed a decrease in the mRNA expression of alternatively activated macrophage markers arg1 and il10, whereas there was no change in the classical activation markers nos2 and tnf (Fig. 6B–E). To further demonstrate that there was a loss of alternative activation of macrophages in T_reg-injected mice, we examined colocalization between CD68, a marker of activated myeloid cells, and arginase-1 by immunofluorescence. Although T_eff-injected mice displayed marked expression of arginase-1 in the CD68⁺ fraction 7 d after injury, injection of T_reg cells led to a decrease in arginase-1 expression by myeloid cells (Fig. 6F, 6G), further demonstrating that both addition and deletion of T_reg cells are detrimental in CNS trauma through their effects on the innate immune response at the site of injury.

There has been an ongoing debate about the role of T_reg cells in CNS injury and in neurodegenerative conditions. We show in this work that complete depletion of T_reg cells using DEREG mice, a manipulation that has been shown to lead to development of numerous organ-specific autoimmune diseases (16), leads to increased neurodegeneration after CNS trauma. However, the same deleterious effect on neuronal survival can be seen when T_reg numbers are increased either by pharmacological compounds or by exogenous supply of T_reg cells. Changes in T_reg numbers correlate with the phenotype of macrophages populating the injury site, with alternatively activated macrophages being spontaneously induced by the injury, yet inhibited by T_reg cell addition or overruled by an exaggerated immune response as a result of T_reg cell depletion.

Early work in models of stroke and Parkinson’s disease showed that depletion of T_reg cells led to increased neurodegeneration, whereas increases in T_reg cell numbers and function improved disease outcome (10, 41). More recent work, using the same manipulations, has shown that T_reg cells play a detrimental role after CNS injuries (7), supporting the hypothesis that they are suppressing a beneficial autoimmune response (19). There are several factors that may have contributed to these disparate findings. There are technical challenges with the current T_reg depletion strategies that have hindered interpretation of depletion studies, such as targeting of activated effector cells with anti-CD25 treatment (42, 43). Furthermore, several studies have shown that T_reg cells have the potential to downregulate Foxp3 and become effector cells, especially when placed in lymphopenic or inflammatory conditions (38, 44, 45), probably due to heterogeneity in the fate commitment of the T_reg cell population (46), complicating transfer experiments into mice with abnormal adaptive immune systems. However, the conditions that drive this switch from T_reg to T_eff, and relevance of these models in vivo, is still a matter of debate (47). Although DTx treatment may have indirect effects, overall the DEREG mouse offers a unique model for T_reg deletion in vivo by both avoiding targeting CD25, indirectly affecting activated T_eff, and avoiding cell transfer, which inevitably changes the cell phenotype. Therefore, although imperfect, DEREG model presents the best currently available model for T_reg depletion.

Our work further addresses the question of how T_reg cells are affecting the outcome from CNS injury. Although we show that T_reg cells have profound effects on neuronal survival from injury, they are not found at the site of the injury, but are rather enriched in the draining lymph node. T_reg cells are known to exert asymmetric control of T cell responses in nonpathological situations (34), raising the possibility that these T_reg cells are exerting their action on the phenotype of T_eff cells in the draining lymph node. The T_eff cells, in turn, direct the phenotype of the infiltrating innate immune cells. Previous works demonstrated that precursors for alternatively activated macrophages are recruited to the injured CNS through a unique path of the choroid plexus (25) and are coming predetermined to differentiate into alternatively activated macrophages.

Several studies have shown that alternative activation of macrophages is a beneficial response to CNS injury (25, 28). Tissue-building macrophages produce growth factors, such as insulin-like growth factor 1, vascular endothelial growth factor, TGF-β, and factors that remodel the extracellular matrix, such as matrix metalloproteinases and resistin-like molecule alpha and promote a tissue-building phenotype in injured tissue (21). T cells, and specifically Th2 effector cells, produce several cytokines, such as IL-4 for example, that can induce alternatively activated macrophages (48). Our results suggest that T cells induced in the CNS-draining dCLN control the phenotype of the infiltrating monocytes, and future studies need to concentrate on better understanding the molecular interaction between T cells and myeloid cells that results in myeloid cells of a particular phenotype to migrate to the site of injury. There potentially could be additional mediators between T cell response and the phenotype of macrophages recruited to the site of injury. It is also unknown whether T cells recirculate between the injured CNS site and the draining dCLN. If such a recirculation occurs, T cell–macrophage interaction could potentially take place in three different sites—the site of injury, the lymph nodes, and blood. Although we show in this work that removal of dCLN is detrimental for neuronal survival, we have not addressed whether their removal after the injury has been inflicted could affect the phenotype of macrophages and, subsequently, neuronal survival.

Our results support the notion that a spontaneous immune response after CNS trauma is beneficial and is tightly regulated by T_reg cells (19). Elimination of T_reg leads to an excessive immune response that is detrimental for injured tissue. However, injection of T_reg cells or potentiation of their suppressive function inhibits a spontaneous immune response to injury and also results in impaired neuronal survival. Further works should be aimed at understanding the divergent properties of T_reg cells that lead to this dichotomous response to injury and finding the compounds that could alleviate T_reg function, yet preserve the beneficial nature/phenotype of T_eff cells. Without better understanding of T_eff/T_reg interactions after CNS injury, therapies for CNS injuries that primarily target the T_reg compartment should be taken with extreme caution, as alteration of T_reg may result in impaired outcome of CNS trauma.

We thank Shirley Smith for editing the manuscript. We thank the members of the Kipnis laboratory for valuable comments during multiple discussions of this work.

The authors have no financial conflicts of interest.