CD73 is Required for Efficient Entry of Lymphocytes, PNAS

8/13/2019 CD73 is Required for Efficient Entry of Lymphocytes, PNAS

1/6

CD73 is required for efficient entry of lymphocytesinto the central nervous system during experimentalautoimmune encephalomyelitisJeffrey H. Mills*, Linda F. Thompson, Cynthia Mueller*, Adam T. Waickman*, Sirpa Jalkanen, Jussi Niemela,Laura Airas, and Margaret S. Bynoe*

*Department of Microbiology and Immunology, College of Veterinary Medicine, Cornell University, Ithaca, NY 14853; Immunobiology and Cancer Program,Oklahoma Medical Research Foundation, Oklahoma City, OK 73104; MediCity Research Laboratory and Department of Neurology, University of Turku,20500 Turku, Finland

Edited by Ralph M. Steinman, The Rockefeller University, New York, NY, and approved April 22, 2008 (received for review November 26, 2007)

CD73 is a cell surface enzyme of the purine catabolic pathway that

catalyzes the breakdown of AMP to adenosine. Because of thestrong immunosuppressive and antiinflammatory properties of

adenosine, we predicted that cd73/ mice would develop severeexperimental autoimmune encephalomyelitis (EAE), an animal

model for the central nervous system (CNS) inflammatory disease,

multiple sclerosis. Surprisingly,cd73/mice were resistantto EAE.However, CD4 T cells from cd73/ mice secreted more proinflam-

matory cytokines than wild-type (WT) mice and were able to

induce EAE when transferred into nave cd73/ T cell-deficientrecipients. Therefore, the protection from EAE observed in cd73/

mice was not caused by a deficiency in T cell responsiveness.

Immunohistochemistry showed thatcd73/ mice had fewer infil-trating lymphocytes in their CNS compared with WT mice. Impor-

tantly, susceptibility to EAE could be induced incd73/ mice afterthe transfer of WT CD73CD4 T cells, suggesting that CD73 must

be expressed either on T cells or in the CNS for disease induction.In the search for the source of CD73 in the CNS that might facilitate

lymphocyte migration, immunohistochemistry revealed a lack ofCD73 expression on brain endothelial cells and high expression inthe choroid plexus epithelium which regulates lymphocyte immu-

nosurveillancebetween the blood and cerebrospinal fluid. Becauseblockade of adenosine receptor signaling with the A2aadenosine

receptor-specific antagonist SCH58261 protected WT mice fromEAE induction, we conclude that CD73 expression and adenosine

receptor signaling are required for the efficient entry of lympho-cytes into the CNS during EAE development.

adenosine multiple sclerosis inflammation choroid plexus

Multiple sclerosis (MS) is a chronic debilitating inflamma-tory disease that affects the CNS (1). Patients with MSexperience a progressive loss in neurological function caused byimmune-mediated axonal demyelination in multiple areas of thebrain and spinal cord. Despite years of research, the etiology ofMS is still unknown. Although there is evidence for genetic (2)and environmental components (3), many studies demonstratethat the immune system plays an integral role in the progression

of MS (1). For example, studies using the dominant MS animalmodel, experimental autoimmune encephalomyelitis (EAE),demonstrate that myelin antigen-specific CD4 T cells caninduce CNS inflammation, demyelination, and neurodegenera-tion, resulting in the loss of motor function (paralysis) (4).

For MS or EAE to develop, autoreactive immune cells mustgain entry into the CNS (5). Typically, the anatomical charac-teristics of the brain, such as the specialized endothelial cell tight

junctions that constitute the framework of the blood brainbarrier (BBB), limit the entry of lymphocytes into the CNS (6).

Although the brain is subject to immunosur veillance undernormal physiological conditions (7), disease progression in MSand EAE is commonly associated with increased lymphocyteinfiltration/extravasation (5). In recent years, experimental MS

therapies directed at either inhibiting lymphocyte trafficking tothe CNS or suppressing inflammation have had varying degreesof success in clinical trials (8). Because adenosine has beenshown to regulate leukocyte migration across endothelial bar-riers (9, 10) and the production of inflammatory cytokines (11),

we asked whether CD73 (ecto-5-nucleotidase), a cell surfaceenzyme that catalyzes the formation of extracellular adenosine,has a role in EAE disease progression.

CD73 is a 70-kDa cell surface enzy me expressed on many cell

types including subsets of lymphocytes (12), endothelial cells(13), and epithelial cells (14). During inflammation and theinitiation of primary immune responses, ATP is released fromdamaged target cells into the extracellular environment and isconverted to AMP by CD39 (15). This AMP is converted toextracellular adenosine by the catalytic action of CD73. Extra-cellular adenosine has potent immunosuppressive effects, me-diated through its four G protein-coupled receptors: A1, A2a,

A2b, and A3(16). Although extracellular ATP has been shown tofunction as a proinflammatory, immunostimulatory mediator,inducing immune cells to secrete cytokines such as IL-1,TNF-, IL-12, and IL-18 (17), adenosine inhibits the productionof proinflammatory cytokines and promotes the induction ofIL-10 (17). Thus, adenosine acts as a negative feedback signal tocounteract ATP-mediated immune stimulation, preventing un-controlled inflammation and lessening the collateral damage tohealthy tissues. Recently, it has also been suggested that thegeneration of adenosine by CD73 mediates the immunosuppres-sive ability of regulatory T cells (Tregs) (18, 19).

Because of the immunomodulatory and immunosuppressiveproperties of adenosine, we evaluated the role of CD73 in EAE.Based on a report of exacerbated EAE in A1adenosine receptor(AR)-deficient mice (20), we expected that cd73/ mice that areunable to catalyze the production of extracellular adenosine wouldexperience severe EAE. Surprisingly, cd73/ mice were highlyresistant to the induction of EAE. However, CD4 T cells fromcd73/ mice do possess the capacity to generate an immuneresponse against CNS antigens and cause severe EAE when adop-tively transferred into cd73/ T cell-deficient mice. CD73CD4

T cells from WT mice also caused disease when transferred intocd73/ recipients, suggesting that CD73 expression, either onlymphocytes or in the CNS, is required for lymphocyte entry into

Author contributions: J.H.M., S.J., and M.S.B. designed research; J.H.M., C.M., A.T.W., S.J.,

J.N., L.A., and M.S.B. performed research; L.F.T. contributed new reagents/analytic tools;

J.H.M., L.F.T., C.M., A.T.W., S.J., J.N., L.A., and M.S.B. analyzed data; and J.H.M., L.F.T., and

M.S.B. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

To whom correspondence may be addressed. E-mail: [email protected].

This article contains supporting information online at www.pnas.org/cgi/content/full/

0711175105/DCSupplemental .

2008 by The National Academy of Sciences of the USA

www.pnas.orgcgidoi10.1073pnas.0711175105 PNAS July 8, 2008 vol. 105 no. 27 93259330
http://www.pnas.org/cgi/content/full/0711175105/DCSupplementalhttp://www.pnas.org/cgi/content/full/0711175105/DCSupplementalhttp://www.pnas.org/cgi/content/full/0711175105/DCSupplementalhttp://www.pnas.org/cgi/content/full/0711175105/DCSupplementalhttp://www.pnas.org/cgi/content/full/0711175105/DCSupplemental


2/6

the brain during EAE. Because cd73/ mice were protected fromEAE induction by the broad-spectrum AR antagonist caffeine andtheA2aAR-specific antagonistSCH58261, our data suggest that theextracellular adenosine generated by CD73, and not CD73 itself,

regulates the entry of lymphocytes into the CNS during EAE. Ourresults demonstrate a role for CD73 and adenosine in regulatingthedevelopment of EAE.

Results

cd73/ Mice Are Resistant to EAE Induction.To determine whetherCD73 plays a role in controlling inflammation during EAEprogression, cd73/ and WT (cd73/) mice were subjected tothe myelin oligodendrocyte glycoprotein (MOG) EAE-inducingregimen (21, 22). Daily monitoring for EAE development re-

vealed that cd73/ mice c onsistently displayed dramaticallyreduced disease severity compared with their WT counterparts(Fig. 1). By 3 weeks after disease induction,cd73/ mice had anaverage EAE score of only 0.5 (weak tail) compared with 2.0

(limp tail and partial hind limb paralysis) for WT mice (Fig. 1).

CD4 T Cells from cd73/ Mice Respond to MOG Immunization. Weasked whether the resistance ofcd73/ mice to EAE inductioncould be explained by either an enhanced ability of cd73/

lymphocytes to suppress an immune response or an inability ofthese lymphocytes to respond to MOG stimulation. Naturallyoccurring CD4CD25FoxP3 T cells, or Tregs, can regulateactively induced EAE (23). Because Tregs have recently beenshown to express CD73 and some reports suggest that theenzymatic activity of CD73 is needed for Treg function (18, 19),

we asked whether the number and suppressive activity of Tregswere normal in cd73/ mice. As shown in Fig. 2A, there wereno significant differences in the frequencies of CD4FoxP3

Tregs in WT and cd73/ mice, either before or after EAE

induction. Similarly, the percentage of CD4

T cells that ex-pressed CD73 changed only modestly after EAE induction inWT mice (Fig. 2B). Additionally, no significant difference wasobserved in the suppressive capacity of WT and cd73/ Tregsto inhibit MOG antigen-specific CD4 effector T cell prolifer-ation (data not shown). To determine whether cd73/ T cellscan respond when stimulated with MOG peptide, the capacity ofthese cells to proliferate and produce cytokines was assessed.CD4 T cells from MOG-immunized cd73/ and WT micedisplayed similar degrees of in vitroproliferation in response to

varying concentrations of MOG peptide (data not shown).However, CD4 T cells from MOG-immunized cd73/ micesecreted higher levels of IL-17 and IL-1 after in vitro MOGstimulation, compared with those of WT CD73CD4 or

CD73CD4 T cells (Fig. 2C). Elevated levels of IL-17 areassociated with MS (24) and EAE development (25), whereashigh levels of the proinflammatory IL-1 cytokine are a riskfactor for MS (26) and an enhancer of IL-17 production (27). Nodifference in IL-2, IL-4, IL-5, IL-10, IL-13, INF-, and TNF-secretion was observed between W T and cd73/ T cells afterMOG stimulation (Fig. 2C and data not shown). Overall, theresults from these assays suggest that cd73/ T cells can respond

well to MOG immunization.We next asked whether T cells from cd73/ mice possess the

ability to cause EAE. To test this possibility, CD4 T cells fromthe spleen and lymph nodes of MOG immunized cd73/ mice

were evaluated for their abilit y to induce EAE after transfer intotcr/ (cd73/) recipient mice; t cr/ mice lack endogenousT cells and cannot develop EAE on their own (ref. 28 and datanot shown).cd73/tcr/ recipient mice that received CD4 Tcells from cd73/ donors developed markedly more severedisease compared with those that received WT CD4 T cells(Fig. 2D). WT and cd73/ donor CD4 T cells displayed equal

degrees of expansion after transfer into cd73/

tcr/

recipientmice (data not shown). Thus, CD4 T cells from cd73/ miceare not only capable of inducing EAE, but cause more severeEAE than those derived from W T mice when transferred intocd73/tcr/ mice. These results are consistent with in vitroassays in which cd73/ CD4 T cells secreted elevated levels ofIL-17 and IL-1 (which are known to exacerbate EAE) inresponse to MOG stimulation (Fig. 2C), and they suggest thatcd73/ mice are resistant to MOG-induced EAE because of alack of CD73 expression in nonhematopoietic cells (most likelylack of expression in the CNS).

cd73/ Mice Exhibit Little/No Lymphocyte Infiltration into the CNSAfter EAE Induction. EAE is primarily a CD4 T cell-mediateddisease (29); and during EAE progression, lymphocytes must

Fig. 1. cd73/ mice are resistant to EAE. EAE was induced, disease activity

wasmonitoreddaily,and themean EAEscorewas calculated forcd73/ (open

diamonds,n 11) and WT (cd73/) (filled squares,n 13) mice. The results

shown are representative of 11 separate experiments.

Fig. 2. cd73/ T cells produce elevatedlevels ofIL-1 and IL-17and mediate

EAEsusceptibility whentransferredto cd73/tcr/ mice. (A) CD4andFoxP3

expression was measured on splenocytes from nave mice and from cd73/

andWT mice day13 after EAEinduction.(B) Splenocytes from navemice and

from WT mice day 13 after MOG immunization were analyzed for CD4 and

CD73 cell surface expression by flow cytometry. (C) Sorted cells from immu-

nized WT orcd73/ mice were cultured with 1 104 irradiated splenocytes

and 0 or 10 g/ml MOG peptide. Supernatants were taken at 18 h and run on

a cytokine Bio-plex assay. Results represent the fold change in cytokine levels

between the0 and10 g/mlMOG peptide groups. Samples werepooledfrom

fourmice andare representative of oneof three similarexperiments.(D)CD4

T cells from the spleen and lymph nodes from MOG-immunizedcd73/ mice

(open diamonds, n 5) or WT mice (filled squares, n 5) were adoptively

transferred into T cell-deficient cd73/tcr/ mice. EAE was induced, and

disease progression was monitored daily. Results are representative of two

separate experiments.

9326 www.pnas.orgcgidoi10.1073pnas.0711175105 Millset al.


3/6

first gain access into the CNS to mount their inflammatoryresponse against CNS antigens, resulting in axonal demyelina-tion and paralysis (4). To determine whether CNS lymphocyteinfiltration is observed after EAE induction in cd73/ mice,brain and spinal cord sections were examined for the presenceof CD4 T cells and CD45 cells by immunohistochemistry.cd73/

mice displayed a dramatically lower frequency of CD4

(Fig. 3 DG) and CD45 [supporting information (SI) Fig. S1]lymphocytes in the brain and spinal cord compared with WTmice (Fig. 3 AC and G) at day 13 after MOG immunization.

Additionally, in ly mphocy te-tracking experiments where MOG-specific T cells from 2d2 TCR transgenic mice (30) weretransferred into either WT or cd73/ mice with concomitantEAE induction, the percentage of 2d2 cells in the CNS increasedseveralfold with time in WT recipient mice, but not at all incd73/ recipients (Fig. S2). Overall, these results suggest thatthe observed protection against EAE induction incd73/ miceis associated with considerably reduced CNS lymphocyte infil-tration. Nevertheless, CD4 T cells from MOG-immunizedcd73/ mice possessed the ability to gain access to the CNS

when transferred into cd73/tcr/ mice that were concomi-tantly induced to develop EAE (Fig. 3 Kand L). In fact, earlierand more extensive CNS CD4 lymphocyte infiltration wasobserved in cd73/tcr/ mice that received cd73/ CD4 T

cells (Fig. 3 K

and L

) than in those that received WT CD4

Tcells (Fig. 3 HJ). Therefore, these results demonstrate thatdonor T cells from cd73/ mice have the ability to infiltrate theCNS of cd73/ recipient mice.

CD73 Must Be Expressed Either on Lymphocytes or in the CNS for

Efficient EAE Development.We next asked whether CD73 expres-sion on CD4 T cells can compensate for a lack of CD73expression in the CNS and allow the development of EAE.Therefore, we adoptively transferred CD4 T cells from MOG-immunized WT mice into cd73/ recipients, concomitantlyinduced EAE, and compared disease activity with that ofsimilarly treated WT recipients (Fig. 4A). Although WT recip-ients developed disease after EAE induction as expected,cd73/ recipients also developed prominent EAE with an

average disease score of 1.5 by 3 weeks after disease induction.This result was much higher than the 0.5 average score thatcd73/ mice normally showed at this same time point (Fig. 1).To define further the association of CD4 T cell CD73 expres-sion with EAE susceptibility, sorted CD73CD4 andCD73CD4 T cells from immunized WT mice, or total CD4

(CD73) T cells from immunizedcd73/mice, were transferredinto cd73/ recipients with concomitant EAE induction (Fig.4B). cd73/ mice that received CD73CD4 T cells from WTmice developed EAE with an average score of1.5 at 3 weeksafter induction. Conversely, cd73/ mice that received WT-derived CD73CD4 T cells did not develop significant EAE.

Additionally, CD4 cells from cd73/ donor mice, which havethe ability to cause severe EAE in CD73-expressing tcr/ mice

Fig.3. cd73/ micedisplaylittleor no CNSlymphocyte infiltration afterEAE

induction; donorcd73/ T cells infiltrate the CNS of cd73/tcr/ recipient

miceafter EAEinduction. (AF) Frozentissuesections from WT mice (AC) and

cd73/ mice (DF) day 13 after EAE induction were labeled with a CD4

antibody. (G) Themeannumber of CD4-infiltrating lymphocytes in the brain

andspinal cord wasquantified perfieldin frozentissue sections from WT and

cd73/ mice day 13 after EAE induction. Eight anatomically similar fields per

brain and four fields per spinal cord per mouse were analyzed at 10 mag-

nification(n5 miceper group). Error barsrepresentthe standarderrorof the

mean. (HL) Frozentissuesectionsof hippocampus(H, I,and K) andcerebellum(Jand L) labeled with a CD4 antibody from EAE-induced tcr/ mice that

received CD4 cells from WT (HJ) or cd73/ (KandL) mice at day 12 (K), 18

(HandL), or 22 (IandJ) after EAE induction. Immunoreactivity was detected

with HRP anti-rat Ig plus AEC (red) against a hematoxylin-stained nuclear

background (blue). Arrows indicate sites of lymphocyte infiltration. (Scale

bars: 500m.)

Fig. 4. Adoptively transferred CD73 T cells from WT mice can confer EAE

susceptibility to cd73/ mice. (A) CD4 T cells from the spleen and lymph

nodes of MOG-immunizedWT micewere enrichedand adoptively transferred

intoWT (filled squares,n5)or cd73/ (open diamonds,n5)mice followed

by concomitant EAE induction. Results are shown from one of two indepen-

dent experiments. (B) T cells from the spleen and lymph nodes of immunized

WT and cd73/ mice were sorted based on CD4 and CD73 expression and

adoptively transferred into cd73/ micefollowedby concomitant EAE induc-

tion (n 5 in each group). Filled squares represent donor cells from WT mice

that express CD73;opensquares representdonor cells from WT mice that lack

CD73 expression; open diamonds represent donor cells from cd73/ mice. (C

and D) Frozentissuesectionsof theCNS choroid plexusfromnaveWT(C Left)

andcd73/ (C Right) mice and WT mice day 12 after EAE induction ( D) were

stained with a CD73- (C) orCD45- (D) specific antibody. Immunoreactivity was

detected with HRP anti-rat Ig plus AEC (red) against a hematoxylin-stained

nuclear background (blue). Brackets indicate CD73 staining. Arrows indicate

CD45 lymphocyte staining. (Scale bars: 500 m.)

Millset al. PNAS July 8, 2008 vol. 105 no. 27 9327
http://www.pnas.org/cgi/data/0711175105/DCSupplemental/Supplemental_PDF#nameddest=SF1http://www.pnas.org/cgi/data/0711175105/DCSupplemental/Supplemental_PDF#nameddest=SF1http://www.pnas.org/cgi/data/0711175105/DCSupplemental/Supplemental_PDF#nameddest=SF1http://www.pnas.org/cgi/data/0711175105/DCSupplemental/Supplemental_PDF#nameddest=SF2http://www.pnas.org/cgi/data/0711175105/DCSupplemental/Supplemental_PDF#nameddest=SF2http://www.pnas.org/cgi/data/0711175105/DCSupplemental/Supplemental_PDF#nameddest=SF1


4/6

(Fig. 2D), were also incapable of potentiating EAE in recipientcd73/mice (Fig. 4B). Therefore, although CD73 expression onT cells can partially compensate for a lack of CD73 expressionin nonhematopoietic cells, EAE is most efficiently induced whenCD73 is expressed in both compartments.

The identity of the CD73-expressing nonhematopoietic cellsthat promote the development of EAE is not known. Weconsidered vascular endothelial cells at the BBB as likely can-didates because many types of endothelial cells express CD73

(13). However, immunohistochemistry revealed that mousebrain endothelial cells are CD73 (data not shown). Duringthese experiments, we observed that CD73 is, however, highlyexpressed in the brain on the choroid plexus (Fig. 4C), which isan entry point into the CNS for lymphocytes during EAEprogression (4). Fig. 4Dshows infiltrating lymphocytes in asso-ciation with the choroid plexus of WT mice 12 days after EAEinduction. Minimal CD73 staining was also observed on sub-meningeal regions of the spinal cord (data not shown). Takentogether, our results suggest that CD73 expression, whether onT cells or in the CNS (perhaps on the choroid plexus), isnecessary for efficient EAE development.

AR Antagonists Protect Mice Against EAE Induction. Because CD73

catalyzes the breakdown of AMP to adenosine and ARs areexpressed in the CNS (20, 31), we next determined whether ARsignaling is important during EAE progression. WT andcd73/

mice were treated with the broad-spectrum AR antagonistcaffeine (32) at 0.6 g/liter in their drinking water, which corre-sponds to an approximate dose of 4.0 mg per mouse of caffeineper day (33), 1 day before and throughout the duration of theEAE experiment (Fig. 5A). WT mice that received caffeine weredramatically protected against EAE development (Fig. 5A),comparable with published results (20). As expected, cd73/

mice that received caffeine did not develop EAE (Fig. 5A).Because CD73 is highly expressed on the choroid plexus (Fig.4C), we next determined whether ARs are also expressed on thechoroid plexus. Using the Z310 murine choroid plexus cell line

(34), only mRNA for the A1and A2aAR subtypes were detectedby quantitative PCR (qPCR) (Fig. 5B). Because A1AR/ mice

have been shown to develop severe EAE after disease induction(20), we asked whether treatment of WT mice with SCH58261(35), an AR antagonist specificfor theA2a subtype, could protectagainst EAE development. WT mice were given 1 mg/kgSCH58261 in DMSO or DMSO alone both i.p. and s.c. (for atotal of 2 mg/kg) 1 day before EAE induction and daily for 30days throughout the course of the experiment (Fig. 5C). WTmice that received SCH58261 were dramatically protectedagainst EAE development compared with WT mice that re-ceived DMSO alone (Fig. 5C). Additionally, WT mice givenSCH58261 displayed a significantly lower frequency of CD4

lymphocytes in the brain and spinal cord compared with DMSO-treated WT mice at day 15 after EAE induction (Fig. 5D).Because studies have shown that adhesion molecules (such asICAM-1, VCAM-1, and MadCAM-1)on thechoroid plexus playa role in the pathogenesis of EAE (36), we determined whetherSCH58261 treatment affected adhesion molecule expression onthe choroid plexus after EAE induction. Comparison of thechoroid plexus from DMSO- and SCH58261-treated WT miceshows that A2a AR blockade prevented the up-regulation ofICAM-1 that normally occurs during EAE progression (Fig. S3).Based on these results, we conclude that the inability ofcd73/

mice to catalyze the generation of extracellular adenosine ex-plains their failure to develop EAE efficiently after MOGimmunization and that CD73 expression and A2a AR signalingat the choroid plexus are requirements for EAE progression.

Discussion

The goal of this work was to evaluate the role of CD73 in EAE,an animal model for MS. Because CD73 catalyzes the formationof extracellular adenosine that is usually immunosuppressive(17) and A1AR/ mice exhibit severe EAE (20), we predictedthat cd73/ mice would also develop severe EAE. However,cd73/ mice were highly resistant to EAE induction, a surpris-ing finding considering the plethora of studies demonstratingthat cd73/mice aremore prone to inflammation. For example,cd73/ mice are more susceptible to bleomycin-induced lunginjury (37) and are more prone to vascular inflammation andneointima formation (38). Consistent with these reports, weshowed that cd73/ T cells produced higher levels of theEAE-associated proinflammatory c ytokines IL-1 and IL-17after MOG stimulation. Furthermore, the adoptive transfer ofcd73/ T cells to tcr/ mice, which lack T cells but express

CD73 throughout their periphery, resulted in severe CNS in-flammation after MOG immunization, consistent with a role foradenosine as an antiinflammatory mediator. It is interesting tonote that IFN- treatment, one of the most effective therapiesfor MS, has been shown to up-regulate CD73 expression onendothelial cells both in vitroand in vivo(39). Thus, although themechanism by which IFN- benefits MS patients is incompletelyunderstood, increased production of adenosine accompanied byits known antiinflammatory and neuroprotective effects couldbe a factor.

Consistent with their resistance to EAE induction, cd73/

mice had a lower frequency of cells infiltrating the CNS duringEAE compared with WT mice. This finding was also unexpectedbecause CD73-generated adenosine has been shown to restrict

Fig. 5. AR blockade protects mice from EAE development. (A) EAE was

induced, disease activity was monitored daily, and the mean EAE score was

calculated in WT (squares)and cd73/ (diamonds) micegiven either drinking

water (filled symbols) alone or drinking water supplemented with 0.6 g/ml

broad-spectrum AR antagonist caffeine (open symbols). Results are from one

experiment (n

5 mice per group). (B) AR mRNA expression levels relative tothe GAPDH housekeeping gene in the Z310 murine choroid plexus cell line.

Samples were run in triplicate; error bars represent the standard error of the

mean.(C) Mice were treated with theA2a ARantagonist SCH58261 at 2 mg/kg

(1 mg/kg s.c. and 1 mg/kg i.p.) in 45% DMSO (filled squares, n 4 mice per

group) or 45% DMSOalone (open squares, n 5 mice pergroup)1 daybefore

anddaily up today 30 after EAEinduction.These results arerepresentative of

two experiments. (D) The mean number of CD4-infiltrating lymphocytes in

thebrainand spinal cordquantifiedper fieldin frozen tissue sectionsfrom day

15 after EAEinduction in SCH58261- andDMSO-treatedmice areshown. Eight

anatomicallysimilar fields perbrainand four fields perspinal cord permouse

were analyzed at 10 magnification (n 4 mice). Error bars represent the

standard error of the mean.

http://www.pnas.org/cgi/data/0711175105/DCSupplemental/Supplemental_PDF#nameddest=SF3http://www.pnas.org/cgi/data/0711175105/DCSupplemental/Supplemental_PDF#nameddest=SF3


5/6

the migration of neutrophils across the vascular endotheliumduring hypoxia and of lymphocytes across high endothelial

venules of draining lymph nodes (9, 10). Our data, in contrast,suggest that CD73 and the extracellular adenosine generated byCD73 are needed for the efficient passage of pathogenic T cellsinto the CNS. Therefore, the role that CD73 and adenosine playin CNS lymphocyte infiltration during EAE is more profoundthan their role in modulation of neuroinflammation.

It is important to know where CD73 must be expressed for T cell

migration into the CNS. CD73 is found on subsets of T cells (12)and on some epithelial (14) and endothelial cells (13). Our dataclearly demonstrate that although cd73/ T cells respond well toMOG immunization, they cannot enter the CNS unless CD73 isexpressed in nonhematopoietic tissues (i.e., cd73/tcr/ micethat develop EAE after adoptive transfer of CD4 T cells fromcd73/ mice). A lack of CD73 on nonhematopoietic cells can alsobe compensated for, in part, by CD73 expression on T cells (i.e.,cd73/ mice become susceptible to EAE when CD73, but notCD73, CD4 T cells are adoptively transferred). Although weconsidered BBB endothelial cells as a relevant source of CD73 inthe CNS because CD73 is expressed on human brain endothelialcells (39), immunohistochemistry revealed that mouse brain endo-thelial cells are CD73. However, CD73 was found to be highlyexpressed on choroid plexus epithelial cells, which form the barrier

between the blood and the cerebrospinal fluid (CSF) and have arole in regulating lymphocyte immunosurveillance in the CNS (40).The choroid plexus has also been suggested to be theentry pointforT cells during the initiation of EAE progression (4). Although weacknowledge the well documented role of lymphocytebrain en-dothelial cell interactions via VLA-4/VCAM-1 binding in bothEAE and MS (41), perhaps lymphocyte trafficking across theendothelial BBB is more important for disease maintenance andprogression than for disease initiation, at least in EAE.

The next issue is how CD73 facilitates the migration of T cellsinto the CNS. Earlier work showed that lymphocyte CD73 canpromote the binding of human lymphocytes to endothelial cellsin an LFA-1-dependent fashion (42). We do not believe that thisresult is the function of CD73 in EAE, however, becauseCD73-deficient T cells can enter the CNS and cause severe

disease in cd73/

tcr/

mice (Fig. 2D). Alternatively, CD73can function as an enzyme to produce extracellular adenosine, aligand for cell surface ARs. It is this latter function that webelieve is relevant for the current work given that AR blockade

with caffeine or SCH58261 can protect mice from EAE. Al-though the broad-spectrum AR antagonist caffeine also inhibitscertain phosphodiesterases (43), its modulation of EAE pro-gression is most likely through its effect on AR signaling (20).This notion is supported by the fact that SCH58261 also preventsEAE progression by specifically inhibiting A2a AR signaling.Because CD73 and the A1and A2a AR subtypes are expressedon the choroid plexus, extracellular adenosine produced byCD73 at the choroid plexus can signal in an autocrine fashion.Because the A1 and A2a ARs are functionally antagonistic toeach other and have different affinities for adenosine (44), the

extracellular concentration of adenosine determines how a cellexpressing the A1and A2areceptors will respond, thus creatinga mechanistic switch whereby low concentrations of adenosineactivate the A1subtype and higher concentrations stimulate the

A2a subtype (45). In the CNS, there is evidence to suggest thatthis A1A2a interaction is important in mediating neuroinflam-mation, where A1 signaling is protective and A2a signalingpromotes inflammation. For example, mice that lack the A1ARdevelop severe EAE after disease induction (20), whereas micethat are given an A2a antagonist are completely protected againstEAE (Fig. 5C). Additionally, mice that lack the A2areceptor areprotected from brain injury induced by transient focal ischemia(46). Therefore, CD73-generated adenosine signaling at thechoroid plexus appears to play a very important role in modu-

lating inflammation in the CNS. This adenosine signaling mostlikely regulates the expression of adhesion molecules at thechoroid plexus. Studies have shown that the up-regulation ofthe adhesion molecules ICAM-1, VCAM-1, and MadCAM-1 atthe choroid plexus is associated with EAE progression (36).Because mice treated with the A2aAR antagonist SCH58261 donot experience increased choroid plexus ICAM-1 expression(Fig. S3), as normally occurs after EAE induction (36), ourresults suggest that A2a AR signaling increases ICAM-1 during

EAE progression.In summary, our data show that CD73 plays a critical role inthe progression of EAE. Mice that lack CD73 are protected fromthe degenerative symptoms and CNS inflammation that areassociated with EAE induction. Our work demonstrates a re-quirement for CD73 expression and AR signaling for the effi-cient entry of lymphocytes into the CNS during EAE. The datapresented here may mark the first steps of a journey that willleadto new therapies for MS and other neuroinflammatory diseases.

Materials and MethodsMice.cd73/ mice have been described in ref. 9 and have been backcrossed to

C57BL/6 for14 generations.cd73/micehave no overt susceptibilityto infection

andappear normalbased on thesizeand cellular composition of their lymphoid

organs and their T and B cell responses in in vivoandin vitroassays (9). C57BL/6

andtcr/ mice on the C57BL/6 background were purchased from The Jackson

Laboratories. Mice were bred and housed under specific pathogen-free condi-

tions at Cornell University or the University of Turku. For AR blockade experi-

ments, mice were given drinking water supplemented with 0.6 g/liter caffeine

(Sigma) or 2 mg/kg SCH58261 (1 mg/kg s.c. and 1 mg/kg i.p.) in DMSO (45% by

volume in PBS) or 45% DMSO alone starting 1 day before EAE induction and

continuing throughout the experiment. All procedures performed on micewere

approved by the relevant animal review committee.

EAE Induction and Scoring.EAE was induced as described in ref. 22. Briefly, a

1:1emulsion of MOG3555 peptide (3 mg/mlin PBS) (Invitrogen)and complete

Freunds adjuvant (CFA; Sigma) was injected s.c. (50 l) into each flank.

Pertussis toxin (PTX, 20 ng) (Biological Laboratories) was given intravenously

(200 l in PBS) at the time of immunization and again 2 days later. Mice were

scored daily for EAE based on disease symptom severity: 0 no disease,

0.51weak/limp tail, 2 limp tail and partial hind limb paralysis, 3 total

hindlimb paralysis,4 both hind limb andforelimb paralysis, 5death. Mice

with a score of 4 were euthanized.

T Cell Preparations and Adoptive Transfer. Mice were primed with MOG3555peptide in CFA without PTX. After 1 week, lymphocytes were harvested from

spleenand lymph nodes andincubated with ACKbuffer (0.15 M NH4Cl, 1 mM

KHCO3, 0.1mM EDTA,pH 7.3) tolyseredbloodcells.Cells were incubatedwith

antibodies to CD8 (TIB-105), IAb,d,v,pr (212.A1), FcR (2.4-G2), B220 (TIB-164),

NK1.1(HB191), andthenBioMag goat anti-mouse IgG, IgM, andgoatanti-rat

IgG (Qiagen). After negative magnetic enrichment, CD4 cells were used

either directly or further sorted into specific subpopulations. For sorting, cells

were stained with antibodies to CD4 (RM4-5) and CD73 (TY/23), and in some

experiments CD25 (PC61), and then isolated by using a FACSAria (BD Bio-

sciences). After sorting, purity was routinely 99%. For adoptive transfer,

total CD4 or sorted T cells were washed and resuspended in sterile PBS.

Recipient mice received 2.5 106 cells i.v. in 200 l of sterile PBS.

Flow Cytometry. Cell suspensionswere stained with fluorochrome-conjugatedantibodies for CD4 (RM4-5), CD73 (TY/23), or FoxP3 (FJK-16s). Intracellular

FoxP3 staining was carried out according to the manufacturers instructions

(eBioscience). Stained cells were acquired on a FACSCalibur (BD Biosciences).

Data were analyzed with FlowJo software (Tree Star).

T CellCytokineAssay. Sorted T cells fromMOG-immunized micewere cultured

in the presence of irradiated C57BL/6 splenocytes with 0 or 10 g/ml MOG

peptide. Supernatants were collected at 18 h and analyzed by using the

Bio-plexcytokine (Bio-Rad)assayfor IL-2, IL-4, IL-5,IL-10,IL-13,IL-17,IL-1,and

TNF-.

Immunohistochemistry. Anesthetized mice were perfused with PBS, and

brains, spleens, and spinal cords were isolated and snap frozen in Tissue

Tek-OCT medium. Five-micrometer sections (brains in a sagittal orientation)

Millset al. PNAS July 8, 2008 vol. 105 no. 27 9329
http://www.pnas.org/cgi/data/0711175105/DCSupplemental/Supplemental_PDF#nameddest=SF3http://www.pnas.org/cgi/data/0711175105/DCSupplemental/Supplemental_PDF#nameddest=SF3


6/6

were affixed to Supefrost/Plus slides (Fisher), fixed in acetone, and stored at

80C.For immunostaining, slides werethawedand treated with0.03% H2O2in PBS to block endogenous peroxidase, blocked with casein (Vector Labora-

tories) in normal goat serum (Zymed), and then incubated with anti-CD45

(YW62.3),anti-CD4(RM4-5),or anti-ICAM-1 (3E2). Slides wereincubated with

biotinylatedgoat anti-ratIg (JacksonImmunoResearch)and streptavidinHRP

(Zymed) and developed with an AEC (Red) substrate kit (Zymed) and a hema-

toxylin counterstain.Coverslips weremountedwith Fluoromount-G and pho-

tographed under light (Zeiss).

Real-Time qPCR.By using TRIzol (Invitrogen), RNA was isolated from the Z310

choroid plexus cell line (34). cDNA was synthesized by using a Reverse-iT kit

(ABGene). Primers (available upon request) specific for ARs were used to

determine geneexpressionlevelsand standardizedto theGAPDH housekeep-

ing gene levels by using a SYBR Green kit (ABGene) run on an ABI 7500

real-time PCR system. Melt curve analyses were performed to measure the

specificity for each qPCR product.

ACKNOWLEDGMENTS. We thank Dr. Wei Zheng (School of Health Sciences,Purdue University, WestLafayette,IN) forhis generousgift of theZ310 murinechoroid plexus cellline. We alsothank Benjamin D. Solomon for proofreadingand Christophe Viret and Michelle Joachims for critical review of this manu-script. L.F.T. holds the Putnam City Schools Distinguished Chair in CancerResearch. This study was supported by National Institutes of Health Grants AI57854 (to M.S.B.) and AI 18220 (to L.F.T.) and by grants from the FinnishAcademy and the Sigrid Juselius Foundation (to S.J.).

1. Keegan BM,et al.(2002) Multiple sclerosis. Annu Rev Med 53:285302.

2. NiinoM, etal. (2007) Recent advancesin genetic analysisof multiplesclerosis: Genetic

associations and therapeutic implications.Exp Rev Neurother 7:11751188.

3. Ascherio A,et al.(2007) Environmental risk factors for multiple sclerosis. II. Noninfec-

tious factors.Ann Neurol61:504513.

4. Brown DA, et al. (2007) Time course and distribution of inflammatory and neurode-

generative events suggest structural bases for the pathogenesis of experimental

autoimmune encephalomyelitis.J Comp Neurol502:236260.

5. Hemmer B,et al. (2004) Multiple sclerosis: A coordinated immune attack across the

blood brain barrier. Curr Neurovasc Res 1:141150.

6. Ribatti D,et al.(2006) Development of the bloodbrain barrier: A historical point of

view. Anat Rec B New Anat289:38.

7. Carrithers MD, et al. (2002) Role of genetic background in P selectin-dependent

immune surveillance of the central nervous system. J Neuroimmunol 129:5157.

8. Hemmer B,et al. (2006) Immunopathogenesis and immunotherapy of multiple scle-rosis.Nat Clin Pract Neurol 2:201211.

9. Thompson LF, et al. (2004) Crucial role for ecto-5-nucleotidase (CD73) in vascular

leakage during hypoxia.J Exp Med 200:13951405.

10. Thompson LF, et al. (2008) Regulation of leukocyte migration across endothelial

barriers by ecto-5-nucleotidase-generated adenosine. Nucleosides Nucleotides Nu-

cleic Acids, in press.

11. Lokshin A, et al. (2006) Adenosine-mediated inhibition of the cytotoxic activity and

cytokine production by activated natural killer cells. Cancer Res66:77587765.

12. Yamashita Y,et al. (1998) CD73 expression and fyn-dependent signaling on murine

lymphocytes.Eur J Immunol 28:29812990.

13. Koszalka P, et al. (2004) Targeted disruption of CD73/ecto-5-nucleotidase alters

thromboregulation and augments vascular inflammatory response. Circ Res 95:814

821.

14. Strohmeier GR,et al. (1997) Surface expression, polarization, and functional signifi-

cance of CD73 in human intestinal epithelia. J Clin Invest99:25882601.

15. Borsellino G,et al. (2007) Expression of ectonucleotidase CD39 by FoxP3 Treg cells:

Hydrolysis of extracellular ATP and immune suppression. Blood 110:12251232.

16. Fredholm BB,et al.

(2001) International Union of Pharmacology. XXV. Nomenclatureand classification of adenosine receptors. Pharmacol Rev 53:527552.

17. Bours MJ, et al. (2006) Adenosine 5-triphosphate and adenosine as endogenous

signaling molecules in immunity and inflammation.Pharmacol Ther 112:358404.

18. Kobie JJ,et al.(2006) T regulatory and primed uncommitted CD4 T cells express CD73,

which suppresses effector CD4 T cells by converting 5-adenosine monophosphate to

adenosine.J Immunol 177:67806786.

19. Deaglio S,et al.(2007) Adenosine generation catalyzed by CD39 and CD73 expressed

on regulatory T cells mediates immune suppression. J Exp Med 204:12571265.

20. Tsutsui S,et al.(2004) A1 adenosine receptor up-regulation and activation attenuates

neuroinflammation and demyelination in a model of multiple sclerosis. J Neurosci

24:15211529.

21. Swanborg RH (1995) Experimental autoimmune encephalomyelitis in rodents as a

model for human demyelinating disease. Clin Immunol Immunopathol77:413.

22. Bynoe MS, et al. (2003) Epicutaneous immunization with autoantigenic peptides

induces T suppressor cells that prevent experimental allergic encephalomyelitis. Im-

munity 19:317328.

23. Kohm AP,et al.(2002) Cutting edge: CD4CD25 regulatory T cells suppress antigen-

specific autoreactive immune responses and central nervous system inflammation

during active experimental autoimmune encephalomyelitis. J Immunol 169:47124716.

24. Matusevicius D,et al.(1999) Interleukin-17 mRNA expression in blood and CSF mono-

nuclear cells is augmented in multiple sclerosis. Mult Scler 5:101104.

25. Komiyama Y, et al. (2006) IL-17 plays an important role in the development of

experimental autoimmune encephalomyelitis.J Immunol177:566573.

26. deJongBA, etal. (2002) Production of IL-1 and IL-1R as riskfactorsfor susceptibility

and progression of relapse-onset multiple sclerosis. J Neuroimmunol 126:172179.

27. Sutton C, et al. (2006) A crucial role for interleukin (IL)-1 in the induction of IL-17-

producing T cells that mediate autoimmune encephalomyelitis.J Exp Med203:1685

1691.

28. Elliott JI, et al. (1996) Mice lacking T cells are resistant to the induction of

experimental autoimmune encephalomyelitis.J Neuroimmunol70:139144.

29. MonteroE, etal. (2004)Regulation of experimentalautoimmune encephalomyelitis by

CD4, CD25, and CD8 T cells: Analysis using depleting antibodies. J Autoimmun

23:17.

30. Bettelli E, et al. (2003) Myelin oligodendrocyte glycoprotein-specific T cell receptor

transgenicmice developspontaneous autoimmuneoptic neuritis.J Exp Med197:10731081.

31. Rosi S, etal. (2003) Theinfluenceof brain inflammation uponneuronaladenosine A2B

receptors.J Neurochem86:220227.

32. DallIgna OP,et al.(2004) Caffeine as a neuroprotective adenosine receptor antago-

nist.Ann Pharmacother 38:717718.

33. JohanssonB, etal. (1997)A 1 andA2A adenosinereceptorsand A1 mRNAinmousebrain:

Effect of long-term caffeine treatment. Brain Res 762:153164.

34. Zheng W, et al. (2002) Establishment and characterization of an immortalized Z310

choroidal epithelial cell line from murine choroid plexus. Brain Res 958:371380.

35. Melani A,et al.(2006) The selective A2Areceptor antagonist SCH58261 protects from

neurological deficit, brain damage, and activation of p38 MAPK in rat focal cerebral

ischemia.Brain Res10731074:470480.

36. Engelhardt B, etal. (2001) Involvement of thechoroidplexusin centralnervoussystem

inflammation.Microsc Res Tech 52:112129.

37. VolmerJB, etal. (2006) Ecto-5-nucleotidase(CD73)-mediated adenosineproduction is

tissue-protective in a model of bleomycin-induced lung injury.J Immunol176:4449

4458.

38. Zernecke A,et al. (2006) CD73/ecto-5-nucleotidase protects against vascular inflam-

mation and neointima formation. Circulation113:21202127.

39. AirasL, etal. (2007)Mechanismof actionof IFN- inthe treatment ofmultiplesclerosis:

A special reference to CD73 and adenosine. Ann NY Acad Sci1110:641648.

40. Steffen BJ,et al.(1996) ICAM-1, VCAM-1, and MAdCAM-1 are expressed on choroid

plexus epithelium but not endothelium and mediate binding of lymphocytesin vitro.

Am J Pathol148:18191838.

41. Rice GP,et al. (2005) Anti-4 integrin therapy for multiple sclerosis: Mechanisms and

rationale.Neurology64:13361342.

42. Airas L,et al. (2000) CD73 engagement promotes lymphocyte binding to endothelial

cells via a lymphocyte function-associated antigen-1-dependent mechanism.J Immu-

nol 165:54115417.

43. Choi OH,et al.(1988) Caffeine and theophylline analogues: Correlation of behavioral

effects with activity as adenosine receptor antagonists and as phosphodiesterase

inhibitors.Life Sci43:387398.

44. Quarta D,et al.(2004) Opposite modulatory roles for adenosine A1and A2Areceptors

on glutamate and dopamine release in the shell of the nucleus accumbens: Effects of

chronic caffeine exposure. J Neurochem 88:11511158.

45. Ciruela F,et al.(2006) Presynaptic control of striatal glutamatergic neurotransmissionby adenosine A1A2Areceptor heteromers. J Neurosci26:20802087.

46. ChenJF, etal. (1999)A2A adenosinereceptor deficiencyattenuates braininjury induced

by transient focal ischemia in mice. J Neurosci 19:91929200.


Documents

CD73 is Required for Efficient Entry of Lymphocytes, PNAS