of May 18, 2018.This information is current as

-Expressing T CellsααReceptor in the Development of CD8

Intrinsic Requirement for the Vitamin D

Danny Bruce and Margherita T. Cantorna

http://www.jimmunol.org/content/186/5/2819doi: 10.4049/jimmunol.1003444January 2011;

2011; 186:2819-2825; Prepublished online 26J Immunol 

MaterialSupplementary

4.DC1http://www.jimmunol.org/content/suppl/2011/01/26/jimmunol.100344

Referenceshttp://www.jimmunol.org/content/186/5/2819.full#ref-list-1

, 15 of which you can access for free at: cites 34 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2011 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on May 18, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on May 18, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

Intrinsic Requirement for the Vitamin D Receptor in theDevelopment of CD8aa-Expressing T Cells

Danny Bruce and Margherita T. Cantorna

Vitamin D and vitamin D receptor (VDR) deficiency results in severe symptoms of experimental inflammatory bowel disease in

several different models. The intraepithelial lymphocytes of the small intestine contain large numbers of CD8aa+ T cells that have

been shown to suppress the immune response to Ags found there. In this study, we determined the role of the VDR in the

development of CD8aa+ T cells. There are fewer total numbers of TCRab+ T cells in the gut of VDR knockout (KO) mice,

and that reduction was largely in the CD8aa+ TCRab+ cells. Conversely TCRgd+ T cells were normal in the VDR KO mice. The

thymic precursors of CD8aa+ TCRab+ cells (triple-positive for CD4, CD8aa, and CD8ab) were reduced and less mature in VDR

KO mice. In addition, VDR KO mice had a higher frequency of the CD8aa+ TCRab+ precursors (double-negative [DN] TCRab+

T cells) in the gut. The proliferation rates of the DN TCRab+ gut T cells were less in the VDR KO compared with those in wild

type. Low proliferation of DN TCRab+ T cells was a result of the very low expression of the IL-15R in this population of cells in

the absence of the VDR. Bone marrow transplantation showed that the defect in VDR KO CD8aa+ TCRab+ cells was cell

intrinsic. Decreased maturation and proliferation of CD8aa+ TCRab+ cells in VDR KO mice results in fewer functional CD8aa+

TCRab+ T cells, which likely explains the increased inflammation in the gastrointestinal tract of VDR KO and vitamin D-deficient

mice. The Journal of Immunology, 2011, 186: 2819–2825.

The human body is composed of ∼100 trillion cells, and 10times that many bacteria reside in the lumen of the in-testine (1). The intestinal epithelial layer not only forms

a physical barrier to protect against invading pathogens but alsocontains a highly specialized immune system. The GALT hasevolved to have effector responses to invading pathogens whilemaintaining tolerance to harmless commensal flora (2). When thebalance between effector and tolerogenic response is lost, in-testinal inflammation can occur like that seen in inflammatorybowel disease (IBD) (2). The intestinal epithelial layer containsintraepithelial lymphocytes (IELs) that are responsible for main-taining intestinal health.The IEL contains several unique cell types including CD8aa+

T cells. Unlike the TCR coreceptor CD8ab, CD8aa does not actas a coreceptor, and T cells that express CD8aa are not MHC Iclass restricted (3, 4). CD8aa has been shown to bind to thenonclassical MHC molecule thymic leukemia (TL) Ag with ahigher affinity than that for MHC class I (5). CD8aa+ TCRab+

IELs are self-reactive but not self-destructive and are believed to

be regulatory T cells that help to maintain tolerance in the gut (6).In addition, CD8aa+ TCRab+ IELs have been shown to suppressintestinal inflammation in the T cell transfer model of IBD (7).The homodimeric form of CD8 can be expressed on both ab andgd T cells in the gut, and expression of CD8aa is IL-15 dependent(8, 9). In addition, IL-15 has been shown to induce maturation andenhance survival and proliferation of both CD8aa+ TCRab+ andCD8aa+ TCRgd+ IELs (9).The intestine can support lymphopoiesis as is evident by the

presence of CD8aa+ IELs in athymic nude mice and in irradiatedneonatal thymectomized mice reconstituted with bone marrow(BM) (4). However, the CD8aa+ IELs in athymic mice are largelyof the TCRgd variety (4, 10). More recent data suggest that thethymus is required for the CD8aa+ TCRab+ IEL (8). TCRgd+

cells diverge from the TCRab+ cells at an early double-negative(DN) stage in the thymus. Like conventional TCRab+ T cells,CD8aa+ TCRab+ IEL progenitors develop from double-positive(DP) thymocytes (8). The DP thymocytes that become CD8aa+

TCRab+ IEL precursors become triple-positive (TP) expressingCD4, CD8ab, and CD8aa (8). The development of these self-reactive T cells requires exposure to self-agonist peptides for se-lection in the thymus like other regulatory T cell populations (4).After surviving agonist selection, CD8aa+ TCRab+ IEL pre-cursors downregulate expression of CD4 and CD8 to become DNTCRab+ thymocytes that express CD5 (8). Unlike conventionalT cells, DN TCRab+ thymocytes egress the thymus and migratedirectly to the intestine (11). Upon entering the IL-15–rich envi-ronment of the intestine, DN TCRab+ cells downregulate CD5and become mature CD8aa+ TCRab+ IELs (8). Even though thegut contains both CD8aa+ TCRab+ and TCRgd+ T cells and theremay be some overlap in function, the two cell types are devel-opmentally distinct.The vitamin D receptor (VDR) is a member of the steroid

hormone family of nuclear receptors (12). The VDR contains aDNA binding domain that is accountable for the high-affinitybinding of the active form of vitamin D (1,25-dihydroxyvitaminD3), for dimerization with retinoid X receptor, and for binding

Department of Veterinary and Biomedical Sciences, Center for Molecular Immunol-ogy and Infectious Disease, The Pennsylvania State University, University Park, PA16802

Received for publication October 18, 2010. Accepted for publication December 23,2010.

This work was supported by the National Institutes of Health/National Institute ofDiabetes and Digestive and Kidney Diseases (DK070781) and by the National Centerfor Complementary and Alternative Medicine and the Office of Dietary Supplements(AT005378).

Address correspondence and reprint requests to Dr. Margherita T. Cantorna, Depart-ment of Veterinary and Biomedical Sciences, Center for Molecular Immunology andInfectious Disease, The Pennsylvania State University, 115 Henning Building, Uni-versity Park, PA 16802. E-mail address: mxc69@psu.edu

The online version of this article contains supplemental material.

Abbreviations used in this article: BM, bone marrow; DN, double-negative; DP,double-positive; IBD, inflammatory bowel disease; IEL, intraepithelial lymphocyte;KO, knockout; MFI, mean fluorescence intensity; TL, thymic leukemia; TP, triple-positive; VDR, vitamin D receptor; WT, wild-type.

Copyright� 2011 by The American Association of Immunologists, Inc. 0022-1767/11/$16.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1003444

by guest on May 18, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

other transcription factors (12). The hetrodimeric complex ofVDR and retinoid X receptor binds to vitamin D response ele-ments and regulates transcription of the target genes (12).Vitamin D is an important modulator of the immune system.

Signaling through the VDR has been shown to suppress multiplemodels of Th1- and Th17-driven autoimmune diseases includingIBD (13). Vitamin D can affect T cell function as well as thedevelopment of specific T cell populations. In vitro, supplemen-tation with 1,25-dihydroxyvitamin D3 limits secretion of IFN-gby CD4 T cells and promotes IL-5 and IL-10, which favors Th2responses over Th1 responses (14, 15). In addition, VDR knockout(KO) TCRab+ cells show an impaired ability to migrate to theintestine when adoptively transferred to Rag KO mice (16). VDR-deficient mice have normal numbers of conventional CD4 andCD8 T cells in the peripheral lymphoid organs. VDR KO micehave increased proportions of Th1 cells, reduced Th2 responses,and fewer iNKT cells and CD8aa+ TCRab+ T cells comparedwith those of wild-type (WT) mice (16–18). CD4 T cells fromVDR KO mice overproduce IFN-g and proliferate twice as muchin MLRs (13). VDR KO and vitamin D-deficient WT mice havea significant reduction in the number of CD4+ IELs that coexpressCD8aa (16).We show here that intrinsic defects occur during the develop-

ment of VDR KO CD8aa+ TCRab+ T cells that result in theimpaired development of these cells in the IEL. VDR KO micehave normal numbers of CD8aa+ TCRgd+ IELs. There is a signi-ficant difference in the percentages and total numbers of T cellsexpressing TCRab in the VDR KO IELs compared with thoseof WT IELs. WT BM can reconstitute VDR KO IELs to normallevels, but VDR KO CD8aa+ TCRab+ IELs fail to developnormally in a WT host. The number of TP thymocytes and thefrequency of maturing TCRb+ TP thymocytes are significantlyreduced in neonatal VDR KO mice. The less mature VDR KO DNTCRab+ IELs are more prevalent in the IELs, fail to becomeIL-15R+, and do not mature and proliferate in response to IL-15.Our results suggest that the VDR is an important factor in the de-velopment and maturation of CD8aa+ TCRab+ IELs.

Materials and MethodsMice

Age- and sex-matched VDR KO and WT C57BL/6 mice were produced atthe Pennsylvania State University (University Park, PA). VDR KO and WTmice were housed in the same room and under the exact same housingconditions. For embryonic thymocytes, timed breedings were performed.Breeding pairs were caged together in the evening, and the following mor-ning females were inspected for seminal plugs. Females with plugs wereseparated to establish embryonic day 0 and monitored for weight gain as anindicator of pregnancy. Experimental procedures received approval fromthe Office of Research Protection Institutional Animal Care and Use Com-mittee at the Pennsylvania State University.

T cell isolation and cell culture

For IELs, the small intestine was removed and flushed with HBSS (Sigma-Aldrich, St. Louis, MO) containing 5% FBS, and the Peyer’s patches wereremoved. The intestine was cut into 0.5-cm pieces. The pieces were in-cubated twice in media containing 0.15 mg/ml DTT (Sigma-Aldrich) andstirred at 37˚C for 20 min. Supernatants were collected, and the IELswere collected at the interface of 40/80% Percoll gradients (Sigma-Aldrich). Thymocytes were prepared with 70-mm nylon strainer andsuspended in RPMI 1640 supplemented with 10% FCS (Thermo-FisherScientific, Rockford, IL). For in vitro stimulation, thymocytes weresuspended at 1.0 3 106 cells/ml in RPMI 1640 supplemented with 10%FCS (Thermo-Fisher Scientific), 2 mM L-glutamine, 100 U/ml penicillinand 100 mg/ml streptomycin, and 5 mM 2-mercaptoethanol (Invitrogen,Carlsbad, CA) and stimulated with 0.5 ng/ml anti-CD3 Ab (BD Bio-sciences, San Jose, CA) and 100 ng/ml recombinant IL-15 (R&D Sys-tems, Minneapolis, MN).

Immunofluorescence

Cells were stained according to standard procedures and analyzed on aFC500 benchtop cytometer (Beckman Coulter, Brea, CA). The followingAbs were used: ECD anti-CD4 (Southern Biotech, Birmingham, AL), PEanti-CD8b, FITC anti-CD122, PE anti-CD25, FITC anti-TCRgd, PE-Cy5anti-TCRb, FITC anti-NK1.1, PE anti-CD5, FITC CD45.1 or CD45.2, andPE-Cy7 anti-CD8a and appropriate isotype controls including FITC IgG1,l (A110-1), PE-Cy7 IgG2a, k (R35-95), PE IgG2b, k (A95-1), FITCIgG2a, k (G155-178), or PE-Cy5 IgG2, l (Ha 4/8) (BD Biosciences).Isotype controls were used to set appropriate gating. CD8aa was detectedwith PE-labeled TL-tetramer (T3b) (8). The tetramers were a gift from Dr.Hilde Cheroutre (La Jolla Institute for Allergy and Immunology, La Jolla,CA) and the National Institutes of Health Tetramer Core Facility (Atlanta,GA). Data were evaluated with WinMDI 2.9 software (Scripps Institute, LaJolla, CA).

BM transplantation

Donor bone marrow cells were isolated and transferred into sublethallyirradiated CD45 allele mismatched recipients. Mice were allowed 6 wk torecover, and reconstitution of BM was evaluated in the blood by flowcytometry and staining for donor CD45 expression.

BrdU incorporation assay

Two-week-old mice were injected i.p. with 50 ml 25 mg/ml BrdU dissolvedin PBS at days 0, 2, and 4 of the experiment. The mice were sacrificed onday 7, and the IELs were analyzed for BrdU incorporation using a BrdUflow kit (BD Pharmingen) according to the manufacturer’s instructions.PE-IgG1, k (MOPC-)21 was used as isotype control (BD Pharmingen).

Statistics

Bar graphs are represented as mean 6 SEM. Data was analyzed by two-tailed unpaired t test using Prism 5.0 statistical software (GraphPadSoftware, La Jolla, CA). The p values #0.05 were considered statisticallysignificant.

ResultsGastrointestinal TCRab T cells are reduced in the absence ofvitamin D signaling

Equal total numbers of cells were isolated from WT and VDR KOIELs (Supplemental Fig. 1A). The frequency of innate immunecells including dendritic cells (WT, 42 6 4%; VDR KO, 30 64%), macrophage (WT, 6 6 2%; VDR KO, 11 6 2%), and NKcells (WT, 2.3 6 0.3%; VDR KO, 3.2 6 0.6%) were the same inthe IELs of WT and VDR KO mice. The frequency of both CD4+

and CD8ab+ T cells were not different in VDR KO and WT mice(16). Staining for the TCRg-chain showed that WT and VDR KOmice had similar frequencies of gd T cells in the IELs (Fig 1A).Staining for TCRb showed significantly fewer TCRab+ T cells inVDR KO mice (Fig. 1A). More than 40% of the IELs in WT miceexpressed the TCRab compared with only 25% of the VDR KOIELs (Fig. 1A). CD8aa can be expressed on many different celltypes but primarily on gd and ab T cells (6). The majority of theCD8aa+ IELs are gd T cells in WT mice (Fig. 1B). In the VDRKO IELs, the frequencies of CD8aa+ TCRgd+ (VDR KO, 67%;WT, 54%) and of CD8aa+ non-T cells (VDR KO, 12%; WT, 5%)were higher than those in the WT IELs (Fig. 1B). However, thetotal numbers of CD8aa+ TCRgd+ and CD8aa+ non-T cells inWT and VDR KO IELs were the same (Supplemental Fig. 1B,1D). Conversely, the frequency and total cell number of CD8aa+

TCRb+ T cells in WT IELs were higher than the VDR KO values(WT, 39%; VDR KO, 17%; Fig. 1B, Supplemental Fig. 1C). In theabsence of the VDR, fewer TCRab+ and fewer CD8aa-express-ing TCRab+ T cells were present in the gut.

Intrinsic defect in VDR KO cell CD8aa+ TCRab+ T cells

Reciprocal BM chimeras were produced using VDR KO and WTmice. Reconstitution of WT–WT (donor–recipient), VDR KO–WT, and WT–VDR KO mice was .80% complete in the blood of

2820 VITAMIN D AND CD8aa T CELLS

by guest on May 18, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

all mice tested (Fig. 1C). Importantly, VDR KO mice werereconstituted with the same efficiency as WT mice (Fig. 1C). TheIELs of WT mice reconstituted with WT BM (WT–WT) had 43%of the CD8aa+ TCRab+ T cells of donor origin (Fig. 1D, 1E).VDR KO recipients of WT BM (VDR KO–WT) had slightlylower reconstitution with WT donor cells, but the values were notsignificantly different from those of WT–WT (Fig. 1D, 1E). FewerCD8aa+ TCRab+ T cells of VDR KO origin were recovered fromWT recipients (VDR KO–WT), and the difference was signifi-cantly different from both the WT–WT and WT–VDR KO chi-meras (Fig. 1D, 1E). The data suggest that VDR KO BM has a cellintrinsic defect in the generation of CD8aa+ TCRab+ T cells thatreside in the IELs.

Early CD8aa+ TCRab+ thymic precursors in the VDR KOmice

CD8aa+ T cell precursors are first detectable in the thymus atembryonic day 16 of fetal development (8). Expression of CD8aaalong with CD8ab and CD4 results in a TP cell type (8). The TPthymocytes in WT mice make up .40% of the thymus at em-bryonic day 16, peak at embryonic day 17 (58%), and are only10% by birth (day 1, Fig. 2A). Similar frequencies of TP thy-mocytes with the same kinetics were found in the VDR KO thy-mus (Fig. 2A). Frequencies of DP thymocytes were also not dif-ferent in the fetal thymus of VDR KO and WT mice.From birth to 3 wk of age, WT mice maintained the percentage

of TP thymocytes (∼10%), whereas the percentage of VDR KOTP thymocytes significantly decreased (Fig. 2B). The frequencyof TP cells in the thymus then declined to that found in the adultthymus by 6 wk (∼5%), and at 6 wk there were equal numbers ofTP cells in WT and VDR KO thymi (Fig. 2B). The frequenciesof DP thymocytes are not different in WT and VDR KO miceregardless of age (Fig. 2C). Expression of the TCR is a step in thematuration of the CD8aa+ T cell precursors (8). There were moreWT TP cells that express the TCRb receptor than VDR KO TPcells that express the TCRb receptor at 2 and 3 wk of age. Thekinetics for the appearance of TCRb+ TP thymocytes was thesame in VDR KO and WT mice, but the WT mice had a higherfrequency of TCRb+ CD8aa precursors (Fig. 2C). By 6 wk whenthe frequency of TP cells was low and not different between WTand VDR KO mice, the expression of TCRb on TP cells was alsonot different (Fig. 2D).

The first 3 wk of life are a critical time for the appearance ofCD8aa+ TCRab+ T cells in the IELs (4). WT IELs contained20% TCRab+ T cells by 1 wk of age, and that percentage grad-ually increased as the mouse matured to the adult levels of 30%(Fig. 2E). CD8aa+ TCRab+ IELs can be found in low frequenciesat 1 and 2 wk in WT mice, and a significant increase occurs be-tween 2 and 3 wk of age to when .40% of the TCRab+ T cellsexpress CD8aa (Fig. 2F). VDR KO TCRab+ IELs fail to appearin the intestine at the same rate as WT T cells, and this defect can

FIGURE 1. Intrinsic defect in VDR KO mice. The T cell subsets in the IELs of WT and VDR KO mice were characterized by flow cytometry. A, The

percentage of gd and ab TCR+ IELs in WT and VDR KO mice, n = 3 per group, and one representative experiment of two. B, The percentage of the

CD8aa IELs from WT and VDR KO mice that are TCRgd1, TCRab1, and TCR2, n = 3 per group, and one representative experiment of two. C, The

percentage of donor-derived cells in the blood of recipient mice (donor→recipient), n = 5 per group. D, Percentage of donor-derived CD8aa+ TCRab+

IELs in the transplant recipients, n = 5 per group. Values are mean6 SEM of three to five mice per group. Data represent one of two experiments. E, Gating

strategy to identify donor-specific CD45 and TCRb+ expressing cells (top left). Dot plot of one recipient of each of the three transplants shown in C and D.

FIGURE 2. The development of CD8aa+ TCRab+ IELs and thymic

precursors in fetal and neonatal VDR KO mice. The percentages of

CD8aa+ TCRab+ thymic precursors and CD8aa+ TCRab+ IELs were

evaluated in fetal and neonatal WT and VDR KO mice. A, The percentage

of the thymus that is TP (CD4+/CD8ab+/CD8aa+) in fetal (embryonic day

[e] 16, e17, and e18) and newborn (day 1 [d1]) WT and VDR KO mice, n =

4 per group. B, The percentages of TP thymocytes in 1-, 2-, 3-, and 6-wk-

old WT and VDR KO mice, n = 3 per group. C, The frequency of DP

thymocytes of WT and VDR KO neonatal mice, n = 3 per group. D, The

percentage of maturing (TCRb+) TP cells in WT and VDR KO mice, n = 3

per group. E, The percentage of the total IELs of WT and VDR KO mice

that express TCRab at 1, 2, 3, and 6 wk of age, n = 3 per group. F, The

percentage of WT and VDR KO CD8aa+ TCRab+ IELs in neonatal mice,

n = 3 per group. These data represent one of two independent experiments.

Values are mean 6 SEM. *p , 0.05 (VDR KO values are significantly

different from WT values).

The Journal of Immunology 2821

by guest on May 18, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

be seen as early as the first week of life (Fig. 2E). The frequencyof VDR KO CD8aa+ TCRab+ T cells in the gut is the same asthat of WT at 1 and 2 wk of age (Fig. 2F). By 3 wk, there arelower frequencies of VDR KO CD8aa+ TCRab+ IELs thanWT CD8aa+ TCRab+ IELs, and the numbers of the CD8aa+

TCRab+ fail to increase further as the mice age (Fig. 2F).

Reduced proliferation of TCRab+ IELs in the absence of theVDR

At the time of CD8aa+ TCRab+ appearance in the IELs (between2 and 3 wk of age in WT mice; Fig. 2F), mice were treated withBrdU to measure the in vivo proliferation of the cells. Total BrdUincorporation of the IELs of WT mice showed that 34% of thecells had proliferated (Fig. 3A). Only 22% of the VDR KO IELsincorporated BrdU over the same time period, which was signif-icantly less than that of WT IELs (Fig. 3A). To identify whatpopulation of IELs was proliferating at a slower rate in the VDRKO mice, the cells were stained for TCRb, CD8aa, and BrdU.Incorporation of BrdU by the TCRb2 IELs (gd T cells, NK cells,etc.) was similar in the VDR KO and WT mice (Fig. 3B). In-corporation of BrdU in the CD8aa+ TCRb2 IELs (largely theCD8aa+ TCRgd+ cells; Fig. 1B) was higher but not significantlyhigher in VDR KO (32%) than that in WT (24%) mice (Fig. 3C).The frequency of proliferating TCRab+ T cells in WT IELs wastwice as high as that of VDR KO IELs (Fig. 3D). Different TCRb+

subsets proliferated at different rates. In WT mice, the frequencyof proliferating DN TCRb+ T cells was significantly greater com-pared with that of CD8aa+ TCRab+ T cells (Fig. 3D). The fre-quency of proliferating DN TCRab+ IELs from VDR KO micewas not different than that of CD8aa+ TCRab+ T cells (Fig. 3D).There was a significant decrease in the frequency of proliferat-ing DN TCRab+ and CD8aa+ TCRab+ IELs in VDR KO micecompared with that in WT mice (Fig. 3D). Only the VDR KOCD8aa+ TCRb+ IELs and their precursors had impaired homeo-static proliferation.

Increased frequency but decreased numbers of CD8aa+

TCRab+ IEL precursors in VDR KO mice

CD8aa+ TCRab+ IELs mature from DN TCRab+ IELs (8). Inthe WT IELs, there was a minor population of less mature DNTCRab+ T cells that had not upregulated CD8aa+ in the IELs(Fig. 4A). VDR KO mice had a higher frequency of DN TCRab+

IELs than that in WT mice (Fig. 4A). Because VDR KO mice havefewer TCRab+ cells, the total number of immature T cells wasless in the VDR KO IELs than that in the WT IELs (Fig. 4B, 4C).The increased frequency of DN TCRab+ IELs in the VDR KOIELs was not due to an increase in NKT cells because the per-centage of TCRab+/NK1.1+ expressing cells was also less in theVDR KO mice than that in the WT mice (Fig. 4D). Immature DNTCRab+ IELs first express CD5 that is later downregulated as the

FIGURE 3. Reduced proliferation of VDR KO

CD8aa+ TCRab+ IEL precursors. Proliferation

was evaluated by BrdU incorporation of the IEL

subsets of WTand VDR KO mice between 2 and 3

wk of age. A, BrdU incorporation of the total IELs

from WT and VDR KO mice at 3 wk of age, n = 3

per group. *p , 0.01. B, Total BrdU incorporation

of TCRab2 IELs from 3-wk-old WT and VDR

KO mice, n = 3 per group. C, BrdU incorporation

in the CD8aa+ and TCRb2 cells, n = 3 per group.

D, The frequency of proliferating TCRb+, DN

TCRab+, and CD8aa+ TCRab+ IELs in 3-wk-old

WT and VDR KO mice (each gated on TCRb+

cells first), n = 3 per group. One representative

histogram or dot plot (A–D) is shown from two

independent experiments. The shaded histograms

are isotype staining, whereas open histograms are

BrdU staining. Values are mean 6 SEM of the

MFI or percentages. *p , 0.05 (VDR KO values

are significantly different from WT values).

FIGURE 4. Immature CD8aa+ TCRab+ pre-

cursors in the VDR KO IELs. CD8aa+ TCRab+

IEL precursors in the intestine of WT and VDR KO

mice were evaluated by flow cytometry. A, The

frequency of DN (CD42/CD82) TCRab+ IELs in

WT and VDR KO mice, n = 4 per group. B, The

total number of TCRab+ IELs isolated from WT

and VDR KO mice, n = 4 per group. C, The total

number of DN TCRab+ IELs in WT and VDR KO

mice, n = 4 per group. D, The percentage of

NKT cells or DN TCRab+ and NK1.1+ cells in the

IELs, n = 4 per group. E and F, The frequency (E)

and total numbers (F) of CD5+ DN TCRab+ cells

in the IELs of WT and VDR KO mice, n = 4 per

group. These data represent one of two experi-

ments. Values are mean 6 SEM.

2822 VITAMIN D AND CD8aa T CELLS

by guest on May 18, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

cell matures and expresses CD8aa (8). Of the DN cells in theIELs of WT mice, 25% express CD5 (Fig. 4E). More than 42% ofthe DN IELs in the VDR KO expressed CD5, which was signif-icantly more than that of the WT DN cells (Fig. 4E). As a result ofthe reduced numbers of TCRab+ cells, the total number of CD5+

DN TCRab+ IELs in VDR KO mice was significantly less thanthat in the WT mice (Fig. 4F). VDR KO mice had a higher fre-quency of immature DN TCRab+ IELs than that of WT mice.

IL-15 unresponsiveness of VDR KO CD8aa precursors in theVDR KO mice

IL-15 is required to induce the maturation of CD8aa+ TCRab+

IELs (8, 19). In addition, IL-15 is required for CD8aa+ TCRgd+

IELs, CD44highCD8+ memory T cells, and NK cells (20). CD8aa+

TCRgd+ IEL numbers were not different in WT and VDR KOmice (Fig. 1, Supplemental Fig. 1B). Of the CD8+ T cells in thespleen, 49 6 3% WT and 45 6 6% VDR KO were of the memoryphenotype or CD44high. Similarly, the frequencies of NK cells(NK1.1+ and CD32) were similar in the spleens of VDR KO andWT mice (4–5%). Of the four cell types that are IL-15 dependent,only the CD8aa+ TCRab+ IEL is affected by VDR deficiency.Expression of the IL-15R was measured on T cells in the thymus

and IELs by measuring CD122 on the CD252 T cells. The meanfluorescence intensity (MFI) of the IL-15R on DN T cells in thethymi of WT mice was low (Fig. 5A). The level of IL-15R ex-pressed on WT and VDR KO DN T cells in the thymus wassimilar (Fig. 5A). Expression of the IL-15R in all IELs of WT andVDR KO mice was also low and similar (Fig. 5B). The MFI of IL-15R was slightly higher on VDR KO (MFI 6.9) CD8aa+ TCRab+

IELs than that on WT CD8aa+ TCRab+ IELs (MFI 5.9), but thedifference did not reach significance (Fig. 5C; p = 0.07). The MFIof IL-15R in the CD8aa+ TCRb2 cells (largely gd T cells; Fig.1B) was ∼4.0 in both the WT and VDR KO IELs (Fig. 5D). Thelevel of expression of IL-15R on WT DN TCRab+ IELs wassignificantly higher (MFI 7.0) than that on VDR KO DN TCRab+

IELs (MFI 4.1, p = 0.01). The frequency of IL-15R expressionwas low on all CD8aa+ cells in the IELs of either VDR KO orWT mice: CD8aa+ TCRab+ (2–3%; Fig. 5E), CD8aa+ TCRb2

(1.8–2.3%), and CD8aa+ TCRgd (0.9–1.2%). Only 2% of WTand 3% of VDR KO CD8aa+ TCRab+ IELs express high levelsof the IL-15R (Fig. 5E). Conversely, 37% of DN TCRb+ cellsexpressed the IL-15R in the WT IELs (Fig. 5F). VDR KO DNTCRb+ IELs failed to express the IL-15R (0.6%; Fig. 5F). Ex-ogenously delivered IL-15 induces the expression of CD8aa inactivated thymocytes (8). Twenty percent of the cells expressedCD8aa when WT thymocytes were activated and cultured withIL-15 (Fig. 5G). Ten percent or half as many of the similarlytreated VDR KO thymocytes upregulated CD8aa in response toIL-15 (Fig. 5G). The data point to defects in DN cell expression ofthe IL-15R and intrinsic defects in the response to exogenous IL-15 in the absence of the VDR.

DiscussionWe show here that there is a T cell intrinsic defect in VDR KOCD8aa T cell development that is TCRab T cell specific. In VDRKO mice, TP thymocyte development was normal during fetaldevelopment and in the adult thymus, but there are fewer TPthymocytes during the first 3 wk of life. VDR KO DN precursorsfailed to mature and proliferate under the influence of IL-15 due tothe low expression of IL-15R. These results identify the VDR asan important regulator of homeostasis, development, maintenance,and proliferation of CD8aa+ TCRab+ cells in the IEL.The GALT can support extrathymic development of CD8aa+

T cells as evident by the discovery that T cells can be found in the

IEL compartment of athymic nude mice (10). The majority of theT cells found in athymic mice are CD8aa+ TCRgd+ T cells, butthe numbers were 4-fold less than those of normal euthymic mice,and very few CD8aa+ TCRab+ IELs existed in athymic mice(10). These discoveries indicated that the thymus is more impor-tant for the development of CD8aa+ TCRab+ than CD8aa+

TCRgd+ IELs. Under normal physiological conditions, TCRgd+

T cell development occurs both thymically and extrathymicallyduring fetal development and continues through adulthood (21).The discovery that TP cells are thymic precursors of CD8aa+

TCRab+ IELs supports a thymic requirement for these cells (8).The ability of CD8aa+ TCRgd+ T cells to develop normally inVDR KO mice suggests that signals specific for CD8aa+ TCRab+

FIGURE 5. Immature CD8aa precursors express low levels of the IL-

15R and respond poorly to IL-15 in the absence of the VDR. A–D, Cell

surface expression of the IL-15R (CD122+ CD252) on (A) DN TCRab+

thymocytes, (B) TCRb+ IELs, (C) CD8aa+ TCRab+ IELs, and (D)

CD8aa+ TCRb2 IELs, n = 3 per group. Values are mean MFI 6 SEM.

p , 0.05 (values for VDR KO are significantly different from the WT

values). E and F, The frequency of (E) CD8aa+ TCRab+ IELs and (F) DN

TCRab+ IELs that express the IL-15R, n = 3 per group. These data are one

representative dot plot from two experiments. Values are mean 6 SEM.

There is a significant difference between WT and VDR KO DN TCRab+

IELs: *p , 0.001. IL-15 induces CD8aa expression on WT and VDR KO

thymocytes in vitro. G, The percentage of WT and VDR KO T cells that

are CD8a and TL-tetramer+ after IL-15 treatment. One representative dot

plot is shown, and gates were set from negative stain and isotype control.

Cultures were performed in triplicate. Data represent one of three

experiments. Values represent mean 6 SEM percentage for VDR KO and

WT cultures. *p , 0.05.

The Journal of Immunology 2823

by guest on May 18, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

T cell development such as thymic selection may be VDR de-pendent.The development of both CD8aa+ TCRab+ and CD8aa+

TCRgd+ IELs is severely impaired in the absence of IL-15, IL-15Ra, or IL-15/IL-2Rb (9). In addition, NK cells and memoryCD8+ T cells fail to develop in the absence of IL-15 and/or the IL-15 receptors (9). IL-15 is produced and trans-presented via theIL-15Ra by the epithelial cells of the intestine (9). Because VDRKO mice had normal numbers of CD8aa+ TCRgd+ IELs and WTCD8aa+ TCRab+ IELs developed normally in a VDR KO host,expression and presentation of IL-15 by intestinal epithelial cellsmust be adequate for induction of these IL-15–dependent celltypes. Furthermore, the selective defect in only the CD8aa+

TCRab+ IEL suggests that there are not global VDR-mediatedeffects on the IL-15 pathway. We found that VDR KO DNTCRab+ T cells failed to become IL-15R+ and proliferated lessthan WT DN T cells in the IELs. The severe reduction in the levelof proliferation in VDR KO DN TCRab+ reflects the loss of IL-15signaling in only this cell type. In addition, thymocytes from VDRKO mice expressed normal amounts of the IL-15R but failed torespond to exogenous IL-15 as well as the WT thymocytes toupregulate CD8aa expression. These data indicate that in theVDR KO mice, CD8aa precursors also have a defect in IL-15Rsignaling. The maturation of CD5+ DN TCRab+ T cells requiresIL-15, and VDR KO IELs contain a higher frequency of CD5+ DNTCRab+ T cells (8). The increased percentage of VDR KO DNTCRab+ T cells that remain in an immature CD5+ state alsoreflects a lack of IL-15 signaling and likely contributed to theoverall reduction of the mature CD8aa+ TCRab+ T cells in theIELs of VDR KO mice. The data suggest that expression ofthe VDR and other transcription factors found only in the precur-sor of the CD8aa+ TCRab+ are required for IL-15R signaling andto upregulate IL-15R expression that results in survival, maturation,and CD8aa+ expression in the IELs. Other IL-15–dependent cellsdo not use vitamin D to regulate IL-15R expression.iNKT cell development is also impaired in VDR KO mice (18).

CD8aa+ TCRab+ T cells have been called “gut NKT cells” be-cause of their expression of NK receptors (6). Like NKT cells,CD8aa+ TCRab+ T cells express NK receptors including mem-bers of the Ly49 family, CD94 and NKR-PI, and use the invariantsignaling component FcεRIg as a part of their CD3 complex (6,22). Both cells have a memory phenotype and rapidly producecytokines upon stimulation (6, 22). In addition, both cells candevelop in the absence of MHC class I or class II but do requireb2-microglobulin (6, 11, 22). An interesting similarity betweenCD8aa+ TCRab+ T cells and NKT cells is their developmentaldependence on expression of the NF-kB family transcriptionfactor RelB (6). RelB deficiency is associated with autoimmunitythat is believed to be linked to dysfunctional T cell selection in thethymus, which allows the survival of conventional T cells thatexpress high-affinity receptors that would normally be deletedduring the selection process (23, 24). The reduced numbers ofCD8aa+ TCRab+ and NKT cells and the survival of conventionalT cells with self-reactive TCR suggests that RelB plays a role inthymic function during agonist selection (6). Notably, RelB istranscriptionally regulated by vitamin D and has been shown tohave VDR response elements in its promoter in both mice andhumans (25). It would be interesting to determine what the role ofRelB is in the VDR-mediated development of CD8aa+ TCRab+

cells.The process of agonist selection allows T cells carrying “for-

bidden repertoires” to survive the selection process but reprogramsthe cell to be regulatory in nature, carrying a self-reactive TCRwhile maintaining a nondestructive function (6, 10). Both

CD8aa+ TCRab+ and NKT cells undergo agonist selection in thethymus as DP cells (6, 22). Each cell type then diverges alonga separate developmental pathway. NKT cells that survive selec-tion downregulate CD8 or CD4 and CD8, rapidly expand, andupregulate expression of CD44 and NK1.1 (22). At the DP stage,CD8aa+ TCRab+ precursors transition to a TP stage in a pre-TCR signal dependent manner (8). The TP thymocytes completeTCR rearrangement and undergo agonist selection (8). Aftersurviving selection, TP cells downregulate all CD4 and CD8 be-coming DN TCRab+ T cells that egress the thymus and continuetheir maturation in the gut under the influence of IL-15 (8). Thesimilarities in the selective requirement for the VDR in CD8aa+

TCRab+ and NKT cells suggest a common mechanism for reg-ulation. Further investigation is needed to determine the role ofvitamin D signaling in the thymus during agonist selection.Vitamin D and signaling through the VDR are important for

regulating intestinal health. VDR deficiency increases the severityof inflammation in several mouse models of intestinal disease (13).The prevalence of Crohn’s disease and ulcerative colitis is higherin northern versus southern climates and urban versus rural areas(26, 27). These are also factors that correlate with vitamin D de-ficiency. Several studies have reported deficiencies in adult andpediatric patients with IBD, and vitamin D deficiency is commoneven when the patient is in remission (28, 29). VDR KO and vi-tamin D-deficient mice are highly susceptible to multiple differentmodels of experimental IBD (13, 30, 31). VDR KO mice do notdevelop overt inflammation, but there is microscopic evidence ofincreased TNF-a and IFN-g in the gastrointestinal tract of theVDR KO mouse, and as it ages the amounts and variety ofcytokines increased [IL-12 was detected (32)]. The data suggestthat IBD is not a vitamin D-deficiency disease but rather vitaminD may be an environmental factor that regulates inflammation inthe gut. Genome-wide association studies have shown that poly-morphisms within the VDR gene increase susceptibility to IBD(33, 34). It seems likely that reduced signaling through the VDRwould limit the development of CD8aa+ TCRab+ T cells in thehuman gut.We have shown that there is an intrinsic and specific defect in

the development of VDR KO CD8aa+ TCRab+ T cells. TP pre-cursors in the thymus are reduced in neonatal VDR KO mice. Im-mature DN TCRab+ T cells fail to upregulate the IL-15R, pro-liferate less, and maintain a more immature phenotype. CD8aa+

TCRab+ T cells are important regulators of intestinal immuneresponses, and their failure to mature in VDR KO mice likelycontributes to the increased susceptibility of these mice to intestinalinflammation.

AcknowledgmentsWe thank Dr. Hilde Cheroutre and the National Institutes of Health Tetramer

Core Facility for the TL-tetramers used in these experiments, Jing Chen for

technical assistance, and the members of the Center for Molecular Immu-

nology and Infectious Diseases for lively discussion.

DisclosuresThe authors have no financial conflicts of interest.

References1. Guarner, F., and J. R. Malagelada. 2003. Gut flora in health and disease. Lancet

361: 512–519.2. Xavier, R. J., and D. K. Podolsky. 2007. Unravelling the pathogenesis of in-

flammatory bowel disease. Nature 448: 427–434.3. Cheroutre, H., and F. Lambolez. 2008. Doubting the TCR coreceptor function of

CD8alphaalpha. Immunity 28: 149–159.4. Lambolez, F., M. Kronenberg, and H. Cheroutre. 2007. Thymic differentiation of

TCR alpha beta(+) CD8 alpha alpha(+) IELs. Immunol. Rev. 215: 178–188.

2824 VITAMIN D AND CD8aa T CELLS

by guest on May 18, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

5. Leishman, A. J., O. V. Naidenko, A. Attinger, F. Koning, C. J. Lena, Y. Xiong,H. C. Chang, E. Reinherz, M. Kronenberg, and H. Cheroutre. 2001. T cellresponses modulated through interaction between CD8alphaalpha and the non-classical MHC class I molecule, TL. Science 294: 1936–1939.

6. Cheroutre, H. 2004. Starting at the beginning: new perspectives on the biology ofmucosal T cells. Annu. Rev. Immunol. 22: 217–246.

7. Poussier, P., T. Ning, D. Banerjee, and M. Julius. 2002. A unique subset of self-specific intraintestinal T cells maintains gut integrity. J. Exp. Med. 195: 1491–1497.

8. Gangadharan, D., F. Lambolez, A. Attinger, Y. Wang-Zhu, B. A. Sullivan, andH. Cheroutre. 2006. Identification of pre- and postselection TCRalphabeta+intraepithelial lymphocyte precursors in the thymus. Immunity 25: 631–641.

9. Ma, L. J., L. F. Acero, T. Zal, and K. S. Schluns. 2009. Trans-presentation of IL-15 by intestinal epithelial cells drives development of CD8alphaalpha IELs. J.Immunol. 183: 1044–1054.

10. Lin, T., G. Matsuzaki, H. Kenai, and K. Nomoto. 1995. Extrathymic and thymicorigin of murine IEL: are most IEL in euthymic mice derived from the thymus?Immunol. Cell Biol. 73: 469–473.

11. Cheroutre, H., and F. Lambolez. 2008. The thymus chapter in the life of gut-specific intra epithelial lymphocytes. Curr. Opin. Immunol. 20: 185–191.

12. Evans, R. M. 1988. The steroid and thyroid hormone receptor superfamily.Science 240: 889–895.

13. Froicu, M., V. Weaver, T. A. Wynn, M. A. McDowell, J. E. Welsh, andM. T. Cantorna. 2003. A crucial role for the vitamin D receptor in experimentalinflammatory bowel diseases. Mol. Endocrinol. 17: 2386–2392.

14. Staeva-Vieira, T. P., and L. P. Freedman. 2002. 1,25-dihydroxyvitamin D3inhibits IFN-gamma and IL-4 levels during in vitro polarization of primarymurine CD4+ T cells. J. Immunol. 168: 1181–1189.

15. Arnson, Y., H. Amital, and Y. Shoenfeld. 2007. Vitamin D and autoimmunity:new aetiological and therapeutic considerations. Ann. Rheum. Dis. 66: 1137–1142.

16. Yu, S., D. Bruce, M. Froicu, V. Weaver, and M. T. Cantorna. 2008. Failure ofT cell homing, reduced CD4/CD8alphaalpha intraepithelial lymphocytes, andinflammation in the gut of vitamin D receptor KO mice. Proc. Natl. Acad. Sci.USA 105: 20834–20839.

17. Wittke, A., V. Weaver, B. D. Mahon, A. August, and M. T. Cantorna. 2004.Vitamin D receptor-deficient mice fail to develop experimental allergic asthma.J. Immunol. 173: 3432–3436.

18. Yu, S., and M. T. Cantorna. 2008. The vitamin D receptor is required foriNKT cell development. Proc. Natl. Acad. Sci. USA 105: 5207–5212.

19. Yu, Q., C. Tang, S. Xun, T. Yajima, K. Takeda, and Y. Yoshikai. 2006. MyD88-dependent signaling for IL-15 production plays an important role in maintenanceof CD8 alpha alpha TCR alpha beta and TCR gamma delta intestinalintraepithelial lymphocytes. J. Immunol. 176: 6180–6185.

20. Kennedy, M. K., M. Glaccum, S. N. Brown, E. A. Butz, J. L. Viney, M. Embers,N. Matsuki, K. Charrier, L. Sedger, C. R. Willis, et al. 2000. Reversible defectsin natural killer and memory CD8 T cell lineages in interleukin 15-deficientmice. J. Exp. Med. 191: 771–780.

21. Hayday, A., and D. Gibbons. 2008. Brokering the peace: the origin of intestinalT cells. Mucosal Immunol. 1: 172–174.

22. Bendelac, A., P. B. Savage, and L. Teyton. 2007. The biology of NKT cells.Annu. Rev. Immunol. 25: 297–336.

23. DeKoning, J., L. DiMolfetto, C. Reilly, Q. Wei, W. L. Havran, and D. Lo. 1997.Thymic cortical epithelium is sufficient for the development of mature T cells inrelB-deficient mice. J. Immunol. 158: 2558–2566.

24. Valero, R., M. L. Baron, S. Guerin, S. Beliard, H. Lelouard, B. Kahn-Perles,B. Vialettes, C. Nguyen, J. Imbert, and P. Naquet. 2002. A defective NF-kappaB/RelB pathway in autoimmune-prone New Zealand black mice is associatedwith inefficient expansion of thymocyte and dendritic cells. J. Immunol. 169:185–192.

25. Griffin, M. D., X. Dong, and R. Kumar. 2007. Vitamin D receptor-mediatedsuppression of RelB in antigen presenting cells: a paradigm for ligand-augmented negative transcriptional regulation. Arch. Biochem. Biophys. 460:218–226.

26. Braus, N. A., and D. E. Elliott. 2009. Advances in the pathogenesis and treat-ment of IBD. Clin. Immunol. 132: 1–9.

27. Koloski, N. A., L. Bret, and G. Radford-Smith. 2008. Hygiene hypothesis ininflammatory bowel disease: a critical review of the literature. World J. Gas-troenterol. 14: 165–173.

28. Pappa, H. M., R. J. Grand, and C. M. Gordon. 2006. Report on the vitamin Dstatus of adult and pediatric patients with inflammatory bowel disease and itssignificance for bone health and disease. Inflamm. Bowel Dis. 12: 1162–1174.

29. Katz, S. 2006. Osteoporosis in patients with inflammatory bowel disease: riskfactors, prevention, and treatment. Rev. Gastroenterol. Disord. 6: 63–71.

30. Cantorna, M. T., C. Munsick, C. Bemiss, and B. D. Mahon. 2000. 1,25-Dihy-droxycholecalciferol prevents and ameliorates symptoms of experimental murineinflammatory bowel disease. J. Nutr. 130: 2648–2652.

31. Froicu, M., and M. T. Cantorna. 2007. Vitamin D and the vitamin D receptor arecritical for control of the innate immune response to colonic injury. BMCImmunol. 8: 5.

32. Froicu, M., Y. Zhu, and M. T. Cantorna. 2006. Vitamin D receptor is required tocontrol gastrointestinal immunity in IL-10 knockout mice. Immunology 117:310–318.

33. Simmons, J. D., C. Mullighan, K. I. Welsh, and D. P. Jewell. 2000. Vitamin Dreceptor gene polymorphism: association with Crohn’s disease susceptibility.Gut 47: 211–214.

34. Dresner-Pollak, R., Z. Ackerman, R. Eliakim, A. Karban, Y. Chowers, andH. H. Fidder. 2004. The BsmI vitamin D receptor gene polymorphism is associatedwith ulcerative colitis in Jewish Ashkenazi patients. Genet. Test. 8: 417–420.

The Journal of Immunology 2825

by guest on May 18, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from