1,25-Dihydroxyvitamin D3 selectively and reversiblyimpairs T helper-cell CNS localizationInna V. Grishkan, Amanda N. Fairchild, Peter A. Calabresi, and Anne R. Gocke1

Department of Neurology, The Johns Hopkins University School of Medicine, Baltimore, MD 21287

Edited by David A. Hafler, Yale University School of Medicine, New Haven, CT, and accepted by the Editorial Board November 11, 2013 (received for reviewApril 1, 2013)

Pharmacologic targeting of T helper (TH) cell trafficking poses anattractive opportunity for amelioration of autoimmune diseasessuch as multiple sclerosis (MS). MS risk is associated with vitaminD deficiency, and its bioactive form, 1,25-dihydroxyvitamin D3[1,25(OH)2D3], has been shown to prevent experimental autoim-mune encephalomyelitis, a mouse model of MS, via an incompletelyunderstood mechanism. Herein, we systematically examined 1,25(OH)2D3 effects on TH cells during their migration from the lymphnodes to the CNS. Our data demonstrate that myelin-reactive THcells are successfully generated in the presence of 1,25(OH)2D3, se-crete proinflammatory cytokines, and do not preferentially differ-entiate into suppressor T cells. These cells are able to leave thelymph node, enter the peripheral circulation, and migrate to the s.c. immunization sites. However, TH cells from 1,25(OH)2D3-treatedmice are unable to enter the CNS parenchyma but are instead main-tained in the periphery. Upon treatment cessation, mice rapidlydevelop experimental autoimmune encephalomyelitis, demon-strating that 1,25(OH)2D3 prevents the disease only temporarilylikely by halting TH cell migration into the CNS.

Multiple sclerosis (MS) is an immune-mediated demyelinatingdisorder of the central nervous system (CNS). Whereas its

etiology is unknown, a number of environmental and geneticfactors contribute to the risk of developing MS (1). Epidemi-ologic studies demonstrate a strong correlation between MSrisk and vitamin D deficiency (2). Higher levels of circulating25-hydroxyvitamin D3 have been recently shown to correlatewith a reduction in the number of new T2 and gadolinium-enhancing lesions (3). Moreover, the bioactive form of vitamin D,1,25-dihydroxyvitamin D3 [1,25(OH)2D3], can completely preventand treat established experimental autoimmune encephalomyelitis(EAE) (4), a murine MS model. Nonetheless, the mechanism of1,25(OH)2D3

’s in vivo action remains incompletely understood.The pathogeneses of MS and EAE require localization of self-

reactive T cells to the CNS. Myelin-specific effector T cells areprimed in the lymph nodes (LNs) and subsequently travel a com-plex route and arrive at the CNS barriers. There, a sequentialinteraction with a multitude of endothelial adhesion molecules andchemokines leads to T helper (TH) cell migration into the CNSparenchyma. Several ligand/receptor pairs have been implicated inthe pathogeneses of MS and EAE (5, 6). Therefore, manipulationof molecules involved in T-cell trafficking poses an attractiveopportunity for therapeutic interventions. Recent studies havehighlighted a role for vitamin D in immune cell migration (7–10), but whether 1,25(OH)2D3 affects trafficking of lympho-cytes in MS and/or EAE has not been investigated.A recent study using conditional deletion of the vitamin D

receptor (VDR) has demonstrated that 1,25(OH)2D3 preventsEAE by acting directly on the encephalitogenic TH cells (11). Wehypothesized that 1,25(OH)2D3 could be inhibiting encephali-togenic T cells via regulation of molecules involved in theirtrafficking into the CNS. Herein, we provide unique evidencethat 1,25(OH)2D3 prevents EAE by modulating the migratoryphenotype of the pathogenic TH cells without affecting theirpriming or effector functions. Our findings further emphasizethat 1,25(OH)2D3 is a reversible immune modulator, as treatmentcessation leads to rapid disease onset. These findings may have

important implications for the use of vitamin D as an adjuncttreatment of immune-mediated diseases such as MS.

Results1,25(OH)2D3 Prevents Accumulation of Inflammatory Infiltrates in theCNS. 1,25(OH)2D3 is known to inhibit EAE induction, but themechanism of disease prevention remains unclear. To investigateits effects on TH cells, 1,25(OH)2D3 or vehicle were administeredto C57BL/6 mice daily by oral gavage, beginning a day beforeimmunization with myelin oligodendrocyte glycoprotein (MOG)35-55. Whereas T lymphocytes infiltrated the CNS and inducedEAE in vehicle-treated animals, no T cells, or CD11b+CD45hi

monocytic infiltrates (Fig. S1), were recovered from the CNS of1,25(OH)2D3-treated mice, all of which remained disease-free(Fig. 1 A and B). Moreover, both vehicle- and 1,25(OH)2D3-treated groups had significantly higher proportions of splenicCD4+T cells compared with the naive control (Fig. 1B). Thus, EAEprevention is accompanied by TH cell expansion in the peripheryand a complete absence of inflammatory infiltrates in the CNS of1,25(OH)2D3-treated animals, leading us to hypothesize that 1,25(OH)2D3 affects the ability of encephalitogenic T cells to enter theCNS without interfering with their activation.

1,25(OH)2D3 Does Not Affect Activation of Pathogenic TH Cells. Wenext examined the effect of 1,25(OH)2D3 on the peripheral in-flammatory response following immunization (Fig. S2). Anequivalent increase in the proportion of the CD44hiCD62Llow

effector and IFNγ+ or IL-17+ TH cells was observed in thedraining lymph nodes (dLNs) and spleens in both vehicle- and1,25(OH)2D3-treated animals, compared with naive controls. Fur-thermore, dLN cells and splenocytes isolated 8 d after immuniza-tion secreted equivalent amounts of IFNγ and IL-17 in response toin vitro restimulation with either MOG 35-55 or anti-CD3/CD28

Significance

Vitamin D plays an important role in regulating the immunesystem in health and disease and may be beneficial for patientswith multiple sclerosis. It prevents CNS autoimmunity in miceby an incompletely understood mechanism. The present studyis a systematic evaluation of the vitamin D effects on T lym-phocytes at each step of their journey to the CNS. The datademonstrate that vitamin D does not affect generation ofpathogenic cells but prevents their presence in the CNS. Unlikecurrent long-acting drugs that impair immune cell trafficking,the effect of vitamin D is quickly reversed after treatmentcessation, which could prove advantageous when immunefunction needs to be reestablished in the setting of infection.

Author contributions: I.V.G., P.A.C., and A.R.G. designed research; I.V.G. and A.N.F. per-formed research; I.V.G. and A.R.G. analyzed data; and I.V.G., P.A.C., and A.R.G. wrote thepaper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission. D.A.H. is a guest editor invited by the EditorialBoard.1To whom correspondence should be addressed. E-mail: agocke4@jhu.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1306072110/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1306072110 PNAS | December 24, 2013 | vol. 110 | no. 52 | 21101–21106

IMMUNOLO

GY

Dow

nloa

ded

by g

uest

on

Aug

ust 1

, 202

0

antibodies and exhibited equivalent MOG-specific proliferativecapacity. Moreover, 1,25(OH)2D3 did not induce apoptosis ofperipheral TH cells. Although it has been proposed that 1,25(OH)2D3 induces TH cell apoptosis in the CNS (9, 11, 12), ab-sence of inflammatory infiltrates precluded us from examiningpotential CNS-specific apoptosis. Taken together, these findingsindicate that the treatment with 1,25(OH)2D3 does not affectgeneration or activation of MOG-specific TH cells, suggesting analternative mechanism for disease prevention. Similarly, 1,25(OH)2D3 ameliorated established disease and reduced the fre-quency of T cells in the CNS but not in the spleen, suggestingthat 1,25(OH)2D3 might be blocking migration of new waves ofactivated MOG-specific cells into the CNS (Fig. S3).

1,25(OH)2D3 Does Not Induce Suppressor T Cells. Induction of reg-ulatory TH cells (Tregs) has been suggested as a potentialmechanism by which 1,25(OH)2D3 prevents EAE (13, 14). Nosignificant difference in the proportion of the CD25+Foxp3+Tregs was detected between the treatment groups at either EAEday 15 or 36 (Fig. S4A). Furthermore, CD45.2+ 2D2 TH cellswere transferred into CD45.1+ congenic mice before immuni-zation, and examination of Foxp3+ 2D2 TH cells 8 d later re-vealed no difference in the proportion of antigen-specific Tregsbetween the treatment groups (Fig. S4B). An ex vivo suppressionassay demonstrated no significant differences between theeffects of TH cells from vehicle- and 1,25(OH)2D3-treated ani-mals on proliferation of naive responder cells (Fig. S4C). Takentogether, these findings demonstrate that 1,25(OH)2D3 does notincrease the frequency of total orMOG-specific Tregs or augmenttheir suppressive capacity, suggesting a direct effect of 1,25(OH)2D3 on encephalitogenic TH cells as a disease preventionmechanism.

1,25(OH)2D3 Does Not Sequester TH Cells to the Lymph Nodes but DoesSustain Leukocytosis. To evaluate the effect of 1,25(OH)2D3 onexpansion and distribution of MOG-specific TH cells, we trans-ferred CD45.2+ 2D2 TH cells into CD45.1+ mice before im-munization with either MOG 35-55 in Complete Freund’s

Adjuvant (CFA) or CFA only. MOG-immunized recipientswere treated with vehicle, 1,25(OH)2D3, or the immunomod-ulatory drug FTY720 which sequesters lymphocytes in lymphnodes. Expansion of MOG-specific cells in the dLNs wasequivalent in the three treatment groups compared with theCFA-only control (Fig. 2A). Whereas equivalent proportions ofMOG-specific TH cells were observed in the spleen and pe-ripheral circulation of 1,25(OH)2D3- and vehicle-treated mice,FTY720 treatment, as expected (15), led to a significant re-duction of 2D2 cells in these compartments and to lymphope-nia (Fig. 2 A and B). Furthermore, leuko- and lymphocytosiswere observed 8 d after immunization and before disease onsetin both vehicle- and 1,25(OH)2D3-treated mice compared withday 0 (Fig. 2C). After clinical symptoms had developed in thevehicle- but not in 1,25(OH)2D3-treated animals (day 16),blood counts remained elevated in the 1,25(OH)2D3 group butreturned almost to baseline in the vehicle-treated animals, sug-gesting that, whereas encephalitogenic cells migrate into the CNS

Fig. 1. 1,25(OH)2D3 inhibits EAE and prevents inflammatory infiltrates inthe CNS. (A) Clinical scores of vehicle (Veh; n = 8) and 1,25(OH)2D3-treated[1,25(OH)2D3; n = 8] mice immunization with MOG 35-55 in CFA. **P =0.0033 (Mann–Whitney test). Data are representative of six independentexperiments. (B) Mice were killed on EAE day 16. Mononuclear cells from theCNS (Upper) and splenocytes (Lower) were isolated, and the proportion ofCD4+ and CD8+ T cells were examined by FACS analysis. Representative flowplots are shown (Left). The corresponding graphs (Right) are representativeof at least four independent experiments (n ≥ 5 per group per experiment).***P < 0.0001.

Fig. 2. 1,25(OH)2D3 treatment does not sequester TH cells to the lymphnodes but does sustain leukocytosis. (A) A total of 5 × 106 CD4+ T cells permouse isolated from unimmunized CD45.2+ 2D2 mice were transferred i.v.into CD45.1+ congenic mice. The recipients were immunized the followingday with either MOG 35-55 in CFA or CFA only. Draining LNs, spleens, andperipheral blood were isolated on day 8, and expansion of CD45.2+ 2D2 cellswas analyzed by FACS. Transferred 2D2 cells as percent of total cells aregraphed. (B) Representative flow plots for circulating leukocytes are shown.(C) Absolute counts of white blood cells (WBC; Left) and lymphocytes (Right)were obtained on EAE days 0, 8, 16, and 22 and are shown overlaid withthe clinical scores. * denotes significant differences between the treatmentgroups within the same time point. @, #, and $ denote significant differencecompared with day 0 (@P < 0.05, #P < 0.01, $P < 0.001). (D) Splenocytes anddLN cells isolated on EAE day 15 were restimulated with 10 μg/mL MOG 35-55 or 0.5 μg/mL soluble anti-CD3/CD28 for 72 h. Secreted IFNγ and IL-17 werequantified by ELISA. All data are representative of at least two independentexperiments (n = 5 mice per group per time point).

21102 | www.pnas.org/cgi/doi/10.1073/pnas.1306072110 Grishkan et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 1

, 202

0

in the vehicle group, the cells remain in the circulation in the1,25(OH)2D3-treated mice. During the chronic stage of disease(day 22), blood counts returned to baseline in both groups, il-lustrating the contraction phase of the immune response. In vitrorestimulation of splenocytes and LN cells isolated at the peak ofdisease revealed that splenocytes from the 1,25(OH)2D3-treatedanimals secreted significantly higher levels of IFNγ and IL-17than the cells from the vehicle group, indicating paucity of MOG-reactive TH cells in the spleens of vehicle- but not 1,25(OH)2D3-treated mice (Fig. 2D). Cells from all groups demonstrated anappropriate polyclonal response to anti-CD3/CD28 stimulation.These findings indicate that MOG-specific TH cells migrate intothe CNS in vehicle-treated animals to cause EAE but remain inthe periphery in the 1,25(OH)2D3-treated animals, suggestingthat 1,25(OH)2D3 prevents circulating pathogenic TH cells fromentering the CNS.

MOG-Specific TH Cells Migrate to the Immunization Sites in 1,25(OH)2D3-Treated Mice. Next, we used in vivo bioluminescent im-aging to further evaluate the effect of 1,25(OH)2D3 on locationof encephalitogenic TH cells. TH cells from unimmunized 2D2-Luc mice (obtained by crossing pan-luciferase and 2D2 mice)were transferred into the B6(Cg)-Tyrc-2J/J strain of mice a daybefore immunization. Increased signal was observed in the dLNs3 d postimmunization, after which a dramatic increase in thesignal was detected at the immunization sites, suggesting that2D2-Luc TH cells expand within the dLNs and then migrate toand accumulate in the nonlymphoid injection sites (Fig. 3). Aftersignal intensity at the injection site peaked around day 9, withminimal signal observed elsewhere, the signal gradually de-creased to baseline. Because significant signal was not observedoutside the dLNs and sites of immunization, the decline in signalafter day 10 can be best explained by cell loss during the con-traction phase of the immune response rather than the re-distribution of 2D2-Luc cells to other tissues such as the CNS.The signal intensity approached the preimmunization levels bythe time the vehicle-treated animals developed clinically appar-ent disease (day 14), suggesting that the endogenous Luc-nega-tive MOG-reactive TH cells were the cells causing the disease.Importantly, no difference in the signal intensity and distributionwas observed between the vehicle- and 1,25(OH)2D3-treatedanimals, indicating that 1,25(OH)2D3 does not affect expansionof MOG-specific TH cells or their migration to the site of im-munization. These data further confirm that 1,25(OH)2D3 doesnot affect generation of MOG-specific TH cells, which are notconfined to the lymphoid organs but are able to enter non-lymphoid injection sites.

1,25(OH)2D3-Treated Mice Have Intact Blood–CNS Barrier. The ap-parent inhibition of transmigration of TH cells into the CNSparenchyma could occur at the level of the blood–brain barrier(BBB). Evaluation of Evans Blue dye extravasation at the peakof EAE revealed an intact BBB of the 1,25(OH)2D3-treatedmice, whereas the paralyzed vehicle-treated mice exhibited sig-nificant loss of BBB integrity (Fig. S5A). Moreover, IgG extra-vasated into the CNS parenchyma in the vehicle- but not 1,25(OH)2D3-treated mice (Fig. S5B). Although these findings sug-gest a potential direct effect of 1,25(OH)2D3 on the BBB en-dothelium, an intrinsic inability of encephalitogenic T cells toenter the CNS would also leave the BBB intact. Hence, thesefindings maintain that 1,25(OH)2D3 prevents TH cells’ migrationinto the CNS and argue against the possibility that the patho-genic TH cells transiently enter the CNS parenchyma and rapidlyrecirculate through cervical lymphatics or undergo apoptosis inthe CNS, as this would undermine BBB integrity.

T Cells Do Not Cross BBB or Blood–Leptomeningeal Barrier Endotheliumin 1,25(OH)2D3-Treated Mice. To migrate into the CNS paren-chyma, encephalitogenic TH cells must cross the BBB or blood–leptomeningeal barrier (BLMB) and the glia limitans (16). Cellsequestration within the perivascular or leptomeningeal space

prevents clinical disease (17, 18). Before the onset of the clinicalsymptoms, inflammatory infiltrates were contained within theleptomeningeal space in approximately half of the vehicle-treatedmice, whereas no infiltrates were detected in any of 1,25(OH)2D3-treated or naive mice (Fig. S6A). At the peak of disease, infiltrateswere observed within the spinal cord parenchyma of all vehicle-treated controls, whereas no meningeal or parenchymal infiltrateswere detected in the 1,25(OH)2D3-treated or naive mice (Fig. S6B).Moreover, vehicle- but not 1,25(OH)2D3-treated mice exhibitedabundant perivascular cuffing within the spinal cord at the peak ofdisease (Fig. S6C). T-cell infiltration was further examined by CD3staining (Fig. S7), which paralleled the histology findings. Thesefindings indicate that the encephalitogenic T cells are not trappedin the perivascular or leptomeningeal space, suggesting that 1,25(OH)2D3 maintains TH cells in the circulation.

EAE Inhibition by 1,25(OH)2D3 Is Reversible. Next, we sought to eval-uate whether 1,25(OH)2D3 permanently alters TH cell phenotypeand ability to cause EAE. Animals developed EAE within ∼7 d oftreatment cessation, whereas mice maintained on 1,25(OH)2D3remained disease-free (Fig. 4A). Equivalent proportions ofsplenic TH cells in vehicle-treated and withdrawal mice wereCD44hiCD62Llow and produced IFNγ and IL-17 (Fig. S8 A and B).In addition, the immune response has contracted by the timewithdrawal mice developed the disease, as indicated by returnof leukocyte and lymphocyte counts to baseline (Fig. S8C) and

Fig. 3. MOG-specific TH cells migrate to the immunization sites in 1,25(OH)2D3-treated mice; 2D2-Luc CD4+ T cells were transferred into B6(Cg)-Tyrc-2J/J mice a day before the immunization. Treatment with vehicle or 1,25(OH)2D3 was initiated on the day of transfer. The mice were subjected tobioluminescent imaging at indicated days following EAE induction. (A)Representative images at various times after immunization are shown (n = 5per group per experiment). (B) Bioluminescent emission (p/s) was analyzedover time for the whole body. No statistically significant difference wasdetected between the groups at any of the examined time points. (C)Clinical scores of vehicle- and 1,25(OH)2D3-treated mice at various times afterimmunization (n = 5 mice per group). (D) The mice were subjected to bio-luminescent imaging on EAE day 3 (d3, Upper) and day 6 (d6, Lower). De-fined anatomical regions are indicated.

Grishkan et al. PNAS | December 24, 2013 | vol. 110 | no. 52 | 21103

IMMUNOLO

GY

Dow

nloa

ded

by g

uest

on

Aug

ust 1

, 202

0

undetectable IFNγ and IL-17 secretion in response to in vitrorestimulation with MOG. Interestingly, whereas the proportionsof TH cells in the CNS were comparable between the two groups,the proportion of CD8+ T cells was significantly lower in thewithdrawal animals (Fig. 4B), suggesting an altered mode ofdisease induction following 1,25(OH)2D3 withdrawal. Thesefindings demonstrate that although 1,25(OH)2D3 prevents EAE,this effect is not permanent, suggesting that pathogenic TH cellsare present but unable to enter the CNS to cause disease duringthe treatment course. Moreover, the prompt reversal of 1,25(OH)2D3 effects upon treatment cessation illustrates that, similarto other nuclear hormones, it signals rapidly and reversibly andsuggests that it affects a cellular target with a rapid turnover.

1,25(OH)2D3 Reversibly Reduces the Expression of the ChemokineReceptor CXCR3 on TH cells. Finally, we evaluated whether 1,25(OH)2D3 can modulate the expression of molecules involved inCNS transendothelial migration of TH cells. A significant re-duction in the proportion of splenic CXCR3+ TH cells in 1,25(OH)2D3-treated animals was observed, equivalent to the pro-portion in unimmunized naive animals (Fig. 5A). However, nochange in the expression of α4-integrin (CD49d) or the che-mokine receptor CCR6, receptors involved in TH cell CNSmigration (19, 20), was detected on splenic TH cells of vehicle-or 1,25(OH)2D3-treated animals at the examined time point (Fig.S9). Importantly, after cessation of the 1,25(OH)2D3 treatment,the proportion of CXCR3+ TH cells reached that of the vehicle-treated animals (Fig. 5B). These findings suggest that 1,25(OH)2D3 may prevent TH cell extravasation across the BBBendothelium by regulating chemokine receptor expression, andthat the effect of 1,25(OH)2D3 is easily reversible, as CXCR3expression is rapidly turned over in the absence of treatment.

DiscussionTo induce an organ-specific autoimmune response, self-reactiveT cells must migrate to their target tissues, a process that isregulated by chemokines and their cognate receptors, selectins,integrins, and metalloproteinases. Pharmacologic targeting of THcell migration poses an attractive opportunity for amelioration ofimmune-mediated diseases such as MS. The data presentedherein provide unique evidence that 1,25(OH)2D3 reversiblyprevents EAE by modulating encephalitogenic TH-cell CNS lo-calization. Specifically, we demonstrate that TH cells from 1,25(OH)2D3-treated EAE mice become activated in response toimmunization, secrete proinflammatory cytokines, and do notdifferentiate into suppressor T cells. These cells are able to leavethe lymph nodes, enter the peripheral circulation, and arrive atthe s.c. immunization sites. However, TH cells from 1,25(OH)2D3-treated mice do not migrate into the CNS parenchyma, but areinstead maintained in the periphery and display down-regulationof the homing receptor CXCR3. Moreover, treatment cessationleads to rapid disease onset, demonstrating that 1,25(OH)2D3

halts TH-cell CNS invasion only temporarily. Our findings em-phasize the direct and reversible effects of 1,25(OH)2D3 onautoreactive TH cells in prevention of autoimmune conditions.1,25(OH)2D3 has been shown to prevent and reverse EAE (4,

21); however, the mechanism of its in vivo action has not beenfully elucidated and may differ for prevention and treatmentparadigms. A recent study using conditional deletion of VDRdemonstrated that 1,25(OH)2D3 prevents EAE by acting directlyon the encephalitogenic TH cells (11). 1,25(OH)2D3 has beenshown to inhibit T-cell proliferation and IL-17 and IFNγ pro-duction in in vitro studies (7, 22), and it has been suggested that1,25(OH)2D3 prevents EAE by inhibiting the differentiation andfunction of the pathogenic TH1 and TH17 cells (7, 14). However,in our prevention paradigm, 1,25(OH)2D3 did not lower theproportion of CD44hiCD62Llow effector or IL-17– and IFNγ-producing TH cells in the peripheral lymphoid organs or reducethe profound leukocytosis elicited by immunization, uniquelydemonstrating, to our knowledge, that the dramatic systemicimmune reaction in EAE is unaffected by 1,25(OH)2D3. Furtherexamination of MOG-reactive TH cells confirmed that 1,25(OH)2D3 does not inhibit generation of a myelin-specific re-sponse. The discrepancies between our and previous findingscould be related to dosing, time, or mode of 1,25(OH)2D3 ad-ministration. Importantly, our findings indicate that 1,25(OH)2D3acts as a selective immunomodulator, thus highlighting its po-tential advantages over broadly immunosuppressive agents thatheighten infection susceptibility.EAE in IL-10/IL-10R–deficient mice has been shown to be

unresponsive to 1,25(OH)2D3, indicating potential immunoreg-ulatory mechanisms of its action (13). However, more recently,Hayes and coworkers (11) have demonstrated that TH cellsexpressing the transcription factor Foxp3 have virtually no VDRtranscripts, whereas Foxp3– effector cells exhibit abundant ex-pression. In agreement with these observations, we detected nochanges in the proportion of total or MOG-specific Tregs or inthe suppressive capacity of TH cells from 1,25(OH)2D3-treatedanimals. Although it is possible that 1,25(OH)2D3 exerted un-detected immunoregulatory effects, one would expect this to belong lasting. Tregs have been shown to lose regulatory activityand acquire a proinflammatory phenotype under certain con-ditions (23); however, the majority of Tregs are considered to bephenotypically stable (24). Thus, the rapid EAE onset observed

Fig. 4. EAE inhibition by 1,25(OH)2D3 is reversible. Immunized mice weretreated with 1,25(OH)2D3 for the duration of the experiment or up to day 17postimmunization (arrow). (A) Clinical scores of mice treated with vehicle(Veh; n = 5), 1,25(OH)2D3 for the entirety of the study [1,25(OH)2D3; n = 5], orwith 1,25(OH)2D3 up to day 21 (withdrawal and W/d; n = 5) are shown.Representative data from four independent experiments are shown. (B) OnEAE day 29, the proportions of CD4+ and CD8+ T cells in the inflammatoryinfiltrates in the brain were determined by FACS analysis.

Fig. 5. 1,25(OH)2D3 reversibly reduces CD4+ T-cell CXCR3 expression. (A) OnEAE day 15, difference in the proportion of splenic CXCR3+ TH cells from 1,25(OH)2D3- (D3) and vehicle-tread (Veh) mice was evaluated by flow cytometricanalysis (**P = 0.0068). (B) Treatment with 1,25(OH)2D3 was stopped on EAEday 17, and the animals were killed on day 29, when the withdrawal animalsreached peak disease severity. The difference in the proportion of splenicCXCR3+ TH cells from vehicle-treated (Veh) and the withdrawal (W/d) groupswas evaluated. The representative flow plots are shown (Left). The corre-sponding graphs (Right) are representative of at least two independentexperiments (n = 5 per group per experiment).

21104 | www.pnas.org/cgi/doi/10.1073/pnas.1306072110 Grishkan et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 1

, 202

0

upon 1,25(OH)2D3 withdrawal suggests a lack of Treg inductionand instead supports the direct effect of 1,25(OH)2D3 on thepathogenic TH cells. This intriguing phenomenon, which has notbeen reported in the literature, raises a question regarding thefate of encephalitogenic T cells during the treatment. For diseaseto develop within a week of treatment cessation, activatedpathogenic T cells must already be present in the periphery,awaiting a chance to reach the CNS. The onset of clinicalsymptoms coincided with a reduction in circulating leukocytesand MOG-specific TH cells in vehicle-treated mice, whereas inthe 1,25(OH)2D3-treated animals, leukocytes remained in theperiphery and responded strongly to in vitro restimulation withMOG 35-55. These findings indicate that the pathogenic TH cellsare present in the peripheral circulation but are unable to enterthe CNS while exposed to 1,25(OH)2D3. The nature and locationof antigen presentation that would be required for TH cell reac-tivation upon 1,25(OH)2D3 withdrawal remains to be elucidated.Of note, rebound effect has been previously shown only fortreatment of established disease (21), where TH cells havealready caused damage in the CNS. In this scenario, 1,25(OH)2D3withdrawal would allow TH cells in the CNS to reinitiate theireffector functions, which is different from the prevention para-digm, where TH cells do not enter the CNS until the drugwithdrawal.The complex route of encephalitogenic TH cells from the LNs

to the CNS is only partially defined. With the use of in vivobioluminescent imaging, we demonstrate that migration to thes.c. immunization sites is an important step in TH cell journeyto the CNS and is unaffected by 1,25(OH)2D3. However, thefunctional significance of this localization for EAE inductionremains to be elucidated. Interestingly, no bioluminescent signalwas detected in the CNS in either treatment group, and the totalbody signal intensity approached the preimmunization levels bythe time clinical symptoms were apparent, indicating that the2D2-Luc cell population contracted by the onset of disease. Thisfinding suggests that the endogenous TH cells were the EAE-inducing cells. Low avidity for the MHC–MOG complex couldallow 2D2 cells to be outcompeted by the endogenous MOG-specific TH cells over time, leading to their contracture beforethe onset of disease (25). Alternatively, this imaging modalitymight not be capable of detecting the few 2D2-Luc cells that doarrive in the CNS. Overall, the imaging data established that 1,25(OH)2D3 does not affect the ability of the MOG-specific TH cellsto migrate to the immunization sites, whereas FACS and im-munohistochemical approaches demonstrate that TH cells arenot sequestered in the perivascular/leptomeningeal space, sug-gesting that 1,25(OH)2D3 prevents T cells from crossing the CNSendothelium and thus maintains them in the periphery. Takentogether, these findings emphasize a potential difference be-tween migration of TH cells to the s.c. injection sites versus theCNS, indicating that 1,25(OH)2D3 might be selectively modu-lating the CNS-tropic TH cell migration.Although much progress has been made in understanding T-

cell trafficking to the CNS (6, 16), the molecular signature thatpermits myelin-reactive TH cells to breach the CNS barriers andinfiltrate the parenchyma remains to be fully defined. Chemo-kine receptor expression profiles differ by TH cell subtype and, incombination with selectins and integrins, define their final des-tination (26, 27). Chemokine receptors are dynamic G protein-coupled receptors, and their expression on TH cells is regulatedby T cell receptor (TCR) stimulation, cytokines, vitamins, andother mediators present at the time of priming (28). 1,25(OH)2D3has been previously shown to modulate the expression of severalchemokine receptors on both mouse and human TH cells in vitro(7, 10). However, the effects of 1,25(OH)2D3 on in vivo expressionof chemokine receptors and adhesion molecules known to playa role in EAE pathogenesis have not been examined. A significantand reversible reduction in CXCR3 expression was detected onperipheral TH cells in 1,25(OH)2D3-treated mice. Although therole of CXCR3 and its ligands in migration of TH cells into theCNS remains controversial (29), the majority of TH cells found in

MS active demyelinating plaques are CXCR3+ (30, 31), andblockade of CXCR3 with antibodies or other antagonists preventsTH1 cells from entering the CNS and subsequently amelioratesEAE (32–34). Moreover, a recent study by Hayes and coworkers(35) demonstrated a requirement for a functional IFNγ gene for1,25(OH)2D3 to prevent the accumulation of pathogenic TH cellsin the CNS, suggesting that IFNγ-producing CXCR3+ cells may bea major target of vitamin D therapy. Because naturally occurringIFNγ-producing TH1 cells are almost exclusively CXCR3 positive,and CXCR3 has been demonstrated to be essential for the es-tablishment of the TH1 amplification loop, it is possible that 1,25(OH)2D3-mediated down-regulation of CXCR3 could at leastpartially explain its ability to prevent EAE.Whereas no changes in α4-integrin expression were detected

in 1,25(OH)2D3-treated mice, the integrin conformation was notexamined. G protein-coupled chemokine receptor signalingleads to integrin activation; hence, changes in chemokine re-ceptor expression could potentially impair interactions betweenT-cell integrins and their endothelial ligands, inhibiting stableadhesion of T cells to the CNS endothelium. Furthermore, al-though 1,25(OH)2D3 has been shown to reduce expression ofCCR6 and to inhibit trafficking of TH17 cells toward the che-mokine CCL20 in vitro (7), the role of CCR6 in EAE inductionis controversial. CCR6 deletion has been shown to prevent TH17cell CNS entry through the choroid plexus, thus inhibiting EAE(19), but contradictory studies have undermined this observation(36, 37). Whereas the molecular signature that defines enceph-alitogenic TH cells is not fully characterized, our findings suggestthat the presence of 1,25(OH)2D3 during the priming of MOG-specific TH cells can alter their expression of chemokine recep-tors without affecting differentiation, thus potentially reversiblymodulating TH cell trafficking phenotype and ameliorating EAE.The reversal of 1,25(OH)2D3 effects could prove advantageouswhen immune function needs to be reestablished during an in-fection. However, this rapid rebound effect raises concern aboutthe use of vitamin D for MS treatment. If vitamin D supple-mentation sequesters the pathogenic T cells in the periphery,treatment cessation could potentially release these cells into theCNS and lead to a clinical relapse. Overall, the present studyexpands the understanding of the mechanism by which 1,25(OH)2D3 prevents EAE and emphasizes the importance of un-derstanding the molecular mechanisms underlying the extrava-sation of encephalitogenic T cells into the CNS, as well as itsprevention, in advancing the treatment for diseases caused byCNS inflammation such as MS.

MethodsAdditional methods can be found in SI Methods.

Mice. C57BL/6 and CD45.1+ congenic mice were purchased from the NationalCancer Institute. B6(Cg)-Tyrc-2J/J mice and 2D2 mice (TCR-transgenic specificfor MOG 35-55) were purchased from The Jackson Laboratory. Luc micewere a kind gift from Robert Negrin (Stanford University, Stanford, CA). Lucmice expressing MOG-specific TCR (2D2-Luc) were generated and bred in ourfacility. All mice were maintained in a federally approved animal facility atJohns Hopkins University in accordance with the Institutional Animal Careand Use Committee. Female mice of 8–12 wk of age were used in all ofthe experiments.

Induction of EAE and Treatment. MOG EAE was induced and monitored aspreviously described (38). For adoptive transfer of 2D2 TH cells, 5 × 106

splenic TH cells isolated from unimmunized CD45.2+ 2D2 mice were trans-ferred i.v. into CD45.1+ congenic recipients a day before immunization. 1,25(OH)2D3 (Tocris Bioscience) was dissolved in 100% (vol/vol) ethanol and storedat −80 °C in light-proof containers. FTY720 (Cayman Chemical) was dissolvedin 100% (vol/vol) ethanol and stored in −80 °C. 1,25(OH)2D3 (1.5 μg/kg mouseweight/d), FTY720 (1.5 mg/kg mouse weight/d), and vehicle control (1% eth-anol in water) were administered daily by oral gavage in 100 μL H2O.

Cell Isolation and Culture. Single-cell suspensions were made by passingspleens and lymph nodes through a 70-μm nylon cell strainer (BD Bio-sciences). Splenic TH cells were isolated using EasySep Mouse CD4+ T-cell

Grishkan et al. PNAS | December 24, 2013 | vol. 110 | no. 52 | 21105

IMMUNOLO

GY

Dow

nloa

ded

by g

uest

on

Aug

ust 1

, 202

0

Enrichment kit (Stemcell Technologies), following the manufacturer’s pro-tocol. Mononuclear cells were isolated from the CNS using a 70–30% Percollstep gradient as previously described (39). Cells were cultured in RPMI 1640supplemented with 10% vol/vol FBS, 100 μg/mL penicillin and streptomycin,0.5 μM 2-mercaptoethanol, 10 mM Hepes buffer, 1 mM sodium pyruvate,and MEM Nonessential amino acids.

ELISA. Spleen anddLN cellswere cultured in complete RPMImedium1640 (5× 106

cells per well in 24-well plates) in the presence of 0 or 10 μg/mL MOG 35-55, or0.5 μg/mL anti-CD3 and anti-CD28 antibodies (BD Biosciences). Supernatantswere collected after 72 h, and the levels of IFNγ and IL-17 were quantified usingELISA kits according to manufacturer’s instructions (BioLegend).

Flow Cytometry. Surface antigens were stained with anti-CD4 (BD Biosciences;RM4-5), anti-CD8 (BD Biosciences; 53-6.7), anti-CD45.2 (BD Biosciences; 104),and anti-CXCR3 (eBioscience; CXCR3-173). Flow cytometric analyses wereperformed on a FACSCalibur instrument (BD Biosciences) and the results wereanalyzed using FlowJo software (TreeStar).

Complete Blood Counts.Absolute counts of white blood cells and lymphocytesin peripheral blood were obtained using Drew Scientific Hemavet 950FS(Drew Scientific).

In Vivo Bioluminescent Imaging. B6(Cg)-Tyrc-2J/J mice were used as recipientsbecause white coats minimize light absorption. Treatment with vehicle or

1,25(OH)2D3 was started on the day of transfer of 2D2-Luc TH cells, and EAEwas induced the following day. Mice were anesthetized with isoflurane, andimages were obtained 8 min after i.p. injections of D-luciferin (150 mg/kgbody weight; Regis Technologies) with a CCD camera cooled to −180 °C,using the IVIS Spectrum Imaging system (Xenogen) with the field of view setat subject height (1.2 cm). The photographic images were obtained usinga 0.2-s exposure, 8f/stop, and small binning. The bioluminescent imageswere obtained using 300-s exposures, 1f/stop, medium binning, and an openfilter. The bioluminescent and gray-scale images were overlaid using LivingImage software (Xenogen). Regions of interest were drawn around entireanimals, and the total flux (p/s) per region of interest was obtained.

Statistical Analysis. Statistical analysis was conducted using GraphPad Prismsoftware (GraphPad). Mann–Whitney test was used to compare data fromnoninterval scales. Two-tailed Student t test was used to analyze normallydistributed data. Results were considered significant if the P value was <0.05:*P < 0.05, **P < 0.01, ***P < 0.001; NS, no significant difference. All errorbars represent SEM.

ACKNOWLEDGMENTS. We thank Drs. J. Bream, J. Powell, and K. Whartenbyfor helpful discussion. This work was supported by National Institutes ofHealth Grant R01 NS041435 (to P.A.C.), the Kenneth and Claudia SilvermanFamily Foundation, and an American Medical Association Foundation SeedGrant (to I.V.G.).

1. Zuvich RL, McCauley JL, Pericak-Vance MA, Haines JL (2009) Genetics and patho-genesis of multiple sclerosis. Semin Immunol 21(6):328–333.

2. Munger KL, Levin LI, Hollis BW, Howard NS, Ascherio A (2006) Serum 25-hydroxy-vitamin D levels and risk of multiple sclerosis. JAMA 296(23):2832–2838.

3. Mowry EM, et al. (2012) Vitamin D status predicts new brain magnetic resonanceimaging activity in multiple sclerosis. Ann Neurol 72(2):234–240.

4. Lemire JM, Archer DC (1991) 1,25-dihydroxyvitamin D3 prevents the in vivo inductionof murine experimental autoimmune encephalomyelitis. J Clin Invest 87(3):1103–1107.

5. Engelhardt B, Ransohoff RM (2005) The ins and outs of T-lymphocyte trafficking tothe CNS: Anatomical sites and molecular mechanisms. Trends Immunol 26(9):485–495.

6. Engelhardt B (2008) Immune cell entry into the central nervous system: Involvementof adhesion molecules and chemokines. J Neurol Sci 274(1–2):23–26.

7. Chang JH, Cha HR, Lee DS, Seo KY, Kweon MN (2010) 1,25-Dihydroxyvitamin D3 in-hibits the differentiation and migration of T(H)17 cells to protect against experi-mental autoimmune encephalomyelitis. PLoS ONE 5(9):e12925.

8. Gysemans CA, et al. (2005) 1,25-Dihydroxyvitamin D3 modulates expression of che-mokines and cytokines in pancreatic islets: Implications for prevention of diabetes innonobese diabetic mice. Endocrinology 146(4):1956–1964.

9. Pedersen LB, Nashold FE, Spach KM, Hayes CE (2007) 1,25-dihydroxyvitamin D3 re-verses experimental autoimmune encephalomyelitis by inhibiting chemokine syn-thesis and monocyte trafficking. J Neurosci Res 85(11):2480–2490.

10. Sigmundsdottir H, et al. (2007) DCs metabolize sunlight-induced vitamin D3 to ‘pro-gram’ T cell attraction to the epidermal chemokine CCL27. Nat Immunol 8(3):285–293.

11. Mayne CG, Spanier JA, Relland LM, Williams CB, Hayes CE (2011) 1,25-Dihydroxy-vitamin D3 acts directly on the T lymphocyte vitamin D receptor to inhibit experimentalautoimmune encephalomyelitis. Eur J Immunol 41(3):822–832.

12. Spach KM, et al. (2004) Gene expression analysis suggests that 1,25-dihydroxyvitaminD3 reverses experimental autoimmune encephalomyelitis by stimulating inflammatorycell apoptosis. Physiol Genomics 18(2):141–151.

13. Spach KM, Nashold FE, Dittel BN, Hayes CE (2006) IL-10 signaling is essential for 1,25-dihydroxyvitamin D3-mediated inhibition of experimental autoimmune encephalo-myelitis. J Immunol 177(9):6030–6037.

14. Joshi S, et al. (2011) 1,25-dihydroxyvitamin D(3) ameliorates Th17 autoimmunity viatranscriptional modulation of interleukin-17A. Mol Cell Biol 31(17):3653–3669.

15. Brinkmann V, Pinschewer D, Chiba K, Feng L (2000) FTY720: A novel transplantationdrug that modulates lymphocyte traffic rather than activation. Trends Pharmacol Sci21(2):49–52.

16. Engelhardt B, Ransohoff RM (2012) Capture, crawl, cross: The T cell code to breach theblood-brain barriers. Trends Immunol 33(12):579–589.

17. Agrawal S, et al. (2006) Dystroglycan is selectively cleaved at the parenchymal base-ment membrane at sites of leukocyte extravasation in experimental autoimmuneencephalomyelitis. J Exp Med 203(4):1007–1019.

18. Toft-Hansen H, et al. (2006) Metalloproteinases control brain inflammation inducedby pertussis toxin in mice overexpressing the chemokine CCL2 in the central nervoussystem. J Immunol 177(10):7242–7249.

19. Reboldi A, et al. (2009) C-C chemokine receptor 6-regulated entry of TH-17 cells intothe CNS through the choroid plexus is required for the initiation of EAE. Nat Immunol10(5):514–523.

20. Yednock TA, et al. (1992) Prevention of experimental autoimmune encephalomyelitisby antibodies against alpha 4 beta 1 integrin. Nature 356(6364):63–66.

21. Cantorna MT, Hayes CE, DeLuca HF (1996) 1,25-Dihydroxyvitamin D3 reversibly blocksthe progression of relapsing encephalomyelitis, a model of multiple sclerosis. ProcNatl Acad Sci USA 93(15):7861–7864.

22. Cippitelli M, Santoni A (1998) Vitamin D3: A transcriptional modulator of the in-terferon-gamma gene. Eur J Immunol 28(10):3017–3030.

23. Zhou X, Bailey-Bucktrout S, Jeker LT, Bluestone JA (2009) Plasticity of CD4(+) FoxP3(+)T cells. Curr Opin Immunol 21(3):281–285.

24. Bailey-Bucktrout SL, Bluestone JA (2011) Regulatory T cells: Stability revisited. TrendsImmunol 32(7):301–306.

25. Chewning JH, Dugger KJ, Chaudhuri TR, Zinn KR, Weaver CT (2009) Bioluminescence-based visualization of CD4 T cell dynamics using a T lineage-specific luciferasetransgenic model. BMC Immunol 10:44.

26. Foxman EF, Campbell JJ, Butcher EC (1997) Multistep navigation and the combina-torial control of leukocyte chemotaxis. J Cell Biol 139(5):1349–1360.

27. Ransohoff RM (2009) Chemokines and chemokine receptors: Standing at the cross-roads of immunobiology and neurobiology. Immunity 31(5):711–721.

28. Sigmundsdottir H, Butcher EC (2008) Environmental cues, dendritic cells and theprogramming of tissue-selective lymphocyte trafficking. Nat Immunol 9(9):981–987.

29. Müller M, Carter S, Hofer MJ, Campbell IL (2010) Review: The chemokine receptorCXCR3 and its ligands CXCL9, CXCL10 and CXCL11 in neuroimmunity—a tale ofconflict and conundrum. Neuropathol Appl Neurobiol 36(5):368–387.

30. Balashov KE, Rottman JB, Weiner HL, Hancock WW (1999) CCR5(+) and CXCR3(+)T cells are increased in multiple sclerosis and their ligands MIP-1alpha and IP-10 areexpressed in demyelinating brain lesions. Proc Natl Acad Sci USA 96(12):6873–6878.

31. Simpson JE, Newcombe J, Cuzner ML, Woodroofe MN (2000) Expression of the in-terferon-gamma-inducible chemokines IP-10 and Mig and their receptor, CXCR3, inmultiple sclerosis lesions. Neuropathol Appl Neurobiol 26(2):133–142.

32. Kohler RE, et al. (2008) Antagonism of the chemokine receptors CXCR3 and CXCR4reduces the pathology of experimental autoimmune encephalomyelitis. Brain Pathol18(4):504–516.

33. Sporici R, Issekutz TB (2010) CXCR3 blockade inhibits T-cell migration into the CNSduring EAE and prevents development of adoptively transferred, but not activelyinduced, disease. Eur J Immunol 40(10):2751–2761.

34. Ni J, et al. (2009) The chemokine receptor antagonist, TAK-779, decreased experi-mental autoimmune encephalomyelitis by reducing inflammatory cell migration intothe central nervous system, without affecting T cell function. Br J Pharmacol 158(8):2046–2056.

35. Spanier JA, Nashold FE, Olson JK, Hayes CE (2012) The Ifng gene is essential for Vdrgene expression and vitamin D₃-mediated reduction of the pathogenic T cell burdenin the central nervous system in experimental autoimmune encephalomyelitis,a multiple sclerosis model. J Immunol 189(6):3188–3197.

36. Elhofy A, Depaolo RW, Lira SA, Lukacs NW, Karpus WJ (2009) Mice deficient for CCR6fail to control chronic experimental autoimmune encephalomyelitis. J Neuroimmunol213(1–2):91–99.

37. Yamazaki T, et al. (2008) CCR6 regulates the migration of inflammatory and regu-latory T cells. J Immunol 181(12):8391–8401.

38. Gocke AR, et al. (2012) Kv1.3 deletion biases T cells toward an immunoregulatoryphenotype and renders mice resistant to autoimmune encephalomyelitis. J Immunol188(12):5877–5886.

39. Lebson L, et al. (2010) Cutting edge: The transcription factor Kruppel-like factor 4regulates the differentiation of Th17 cells independently of RORγt. J Immunol185(12):7161–7164.

21106 | www.pnas.org/cgi/doi/10.1073/pnas.1306072110 Grishkan et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 1

, 202

0