Immunity

Previews

are leading to the identification of thera-

peutic targets and the potential reposi-

tioning of existing drugs with good safety

profiles that modulate the genetically-

validated pathways active in patients or

those at risk of the disease (Plenge

et al., 2013). What clues do these new re-

sults offers for therapeutic strategies in

autoimmune disease? As the authors

discuss, targeting the ATPase activity of

MDA5 might not be that useful, notwith-

standing the reasonable position that

pharmacological inhibition of an essential

antiviral defense molecule might not be

safe. Nevertheless, this study and others

indicate that targeting more downstream

autoimmune events as a consequence

of activating variants of MDA5, such as

inhibition of the IL-6 receptor or of

TNF-a or depletion of (antigen-specific, if

possible) effector B and T cells and/or

improving immunosuppressive T regula-

tory cell functions, such as CTLA-4 or

IL-10, might provide safe therapeutic

approaches. There are obviously already

successful drugs in use clinically for

some of these targets (Plenge et al.,

2013). Nevertheless, dose and frequency

of administration of the appropriate drug

in humans (results from mice might not

translate) should be determined in rela-

tively small, open-label, statistically-

designed, mechanistic studies or trials to

investigate immunological efficacy before

launching into large and expensive late-

phase trials. The beauty of this new

single-gene model of autoimmune dis-

ease is that it is caused by a single gene

that is also an autoimmune disease gene

in humans and could therefore be useful

to explore therapeutics preclinically. As

discussed recently (Plenge et al., 2013),

knowledge from human genetics and

genomics studies is, and will continue,

transforming medicine.

REFERENCES

Downes, K., Pekalski, M., Angus, K.L., Hardy, M.,Nutland, S., Smyth, D.J., Walker, N.M., Wallace,C., and Todd, J.A. (2010). PLoS ONE 5, 5.

Ferreira, R.C., Hui Guo, H., Coulson, R.M.R.,Smyth, D., Pekalski, M.L., Burren, O.S., Cutler,A.J., Doecke, J.D., Flint, S., McKinney, E.F., et al.

Immunity 40,

(2014). Diabetes 63. Published online February16, 2014. http://dx.doi.org/10.2337/db13-1777.

Funabiki, M., Kato, H., Miyachi, Y., Toki, H.,Motegi, H., Inoue, M., Minowa, O., Yoshida, A.,Deguchi, K., Sato, H., et al. (2014). Immunity 40,this issue, 199–212.

Ivashkiv, L.B., and Donlin, L.T. (2013). Nat. Rev.Immunol. 14, 36–49.

Molineros, J.E., Maiti, A.K., Sun, C., Looger, L.L.,Han, S., Kim-Howard, X., Glenn, S., Adler, A., Kelly,J.A., Niewold, T.B., et al.; BIOLUPUS Network(2013). PLoS Genet. 9, e1003222.

Nejentsev, S., Walker, N., Riches, D., Egholm, M.,and Todd, J.A. (2009). Science 324, 387–389.

Plenge, R.M., Scolnick, E.M., and Altshuler, D.(2013). Nat. Rev. Drug Discov. 12, 581–594.

Robinson, T., Kariuki, S.N., Franek, B.S., Kumabe,M., Kumar, A.A., Badaracco, M., Mikolaitis, R.A.,Guerrero, G., Utset, T.O., Drevlow, B.E., et al.(2011). J. Immunol. 187, 1298–1303.

Smyth, D.J., Cooper, J.D., Bailey, R., Field, S.,Burren, O., Smink, L.J., Guja, C., Ionescu-Tirgoviste, C., Widmer, B., Dunger, D.B., et al.(2006). Nat. Genet. 38, 617–619.

Yoneyama, M., Kikuchi, M., Natsukawa, T.,Shinobu, N., Imaizumi, T., Miyagishi, M., Taira, K.,Akira, S., and Fujita, T. (2004). Nat. Immunol. 5,730–737.

How T Cells Lose Their Touch

Michael L. Dustin1,2,*1Kennedy Institute of Rheumatology, NDORMS, The University of Oxford, Headington, Oxford OX3 7FY, UK2Skirball Institute of Biomolecular Medicine, New York University School of Medicine, New York, NY 10016, USA*Correspondence: michael.dustin@kennedy.ox.ac.ukhttp://dx.doi.org/10.1016/j.immuni.2014.02.001

T cells are among the most sensitive of cells, but in this issue of Immunity, Honda et al. (2014) demonstratethat effector T cells must lose their touch within hours to protect the host from immunopathology.

In this issue of Immunity, Honda et al.

(2014) use a delayed-type hypersensitivity

(DTH) model to look at how highly sensi-

tive T cells turn off responses at sites of

inflammation. One of the major problems

that prevented prior studies from being

able to address this issue was asyn-

chrony. The DTH model offered a way to

synchronize the response to antigen and

observe it through the intact ear skin

by two-photon microscopy. Honda et al.

(2014) take advantage of the fact that

recruitment of activated T cells into sites

of inflammation is antigen independent.

Therefore, when mice are injected with a

strong adjuvant and a mixture of keyhole

limpet hemocyanin (KLH) and ovalbumin

peptide (OVA) subcutaneously and then

challenged by intradermal injection of

KLH in the ear pinnae 7 days later, the

activation of a few KLH-specific cells in

the dermis leads to recruitment of more

KLH- and OVA-specific cells over the

next couple of days (Honda et al., 2014).

The inflammation induced by the T cells

leads to ear swelling, and the OVA-spe-

cific T cells migrate within the inflamed

dermis in search of antigen. Intravenous

OVA peptide introduced at this point

rapidly permeates into the inflamed site

and loads onto I-Ab molecules to activate

OVA-specific T cells, leading to arrest of

the T cells within 1 min. Thus, the activa-

tion process can be synchronized and

the kinetics of the effector cell response

to antigen in an inflammatory setting

could be studied in detail.

Honda et al. (2014) first asked whether

the apparent desensitization of the T cells

was due to loss of antigen, other changes

in theAPCs,or achange in theTcells. They

found that antigen was still active by

February 20, 2014 ª2014 Elsevier Inc. 169

Figure 1. PD-L1 Is Required for T Cell DesensitizationFresh effector T cells utilize antigen and adhesion molecules to form a stable synapse in response to few MHC-peptide ligands. Changes in the T cell, includingupregulation of PD-1, coincide with PD-L1-dependent desensitization. PD-L1 may act by binding to PD-1 or CD80 on T cells to attenuate the T cell response.These antigen-dependent kinapses may allow chemokine production that can recruit fresh T cells into the reaction as long as antigen is present.

Immunity

Previews

injecting a second cohort of OVA-specific

effector cells by adoptive transfer, which

then entered the tissue site, stopped in

contact with tissue APCs, and produced

cytokine. Thus, antigen was still present

and the APCs were fully competent.

This suggested that the T cells were

desensitizing.

T cells have remarkable sensitivity to

antigen, so it was of interest to know

what accounted for this loss of touch.

The expression of T cell receptor re-

mained high and essentially constant

throughout the time course. However,

they found that increased PD-1 expres-

sion on antigen-specific T cells correlated

with a physical dissociation from APCs

and that PD-L1 antibodies could delay

the departure of the T cells and sustain

cytokine production (Honda et al., 2014).

A caveat with use of the PD-L1 antibody

is that PD-L1 also binds to CD80. Honda

et al. (2014) show that PD-L1 is expressed

on both the T cells and APCs, but they

don’t evaluate CD80 expression on the

T cells. This interaction has also been

implicated in inhibition of T cell functions

(Butte et al., 2007). Thus, although PD-1

ligation by PD-L1 is the most likely molec-

ular explanation, the experiments per-

formed do not rule out other edges in

the cosignaling network. Blocking CTLA-

4, another candidate for T cell mobiliza-

tion, had a minor stabilizing effect, but

it did not reach statistical significance.

A challenge in this experiment is that

some CTLA-4 antibodies are agonistic

for motility and might simply replace the

go signal provided by CD80 and CD86

170 Immunity 40, February 20, 2014 ª2014 E

with one delivered by the antibody

(Ruocco et al., 2012). It is clear, however,

that the cosignaling network has a sig-

nificant role in acutely controlling T cell

sensitivity to antigen (Figure 1).

PD-1 is under intense scrutiny as

a potential therapeutic target. PD-1

signaling has differential effects on T cell

responses, with high amounts of PD-1

required to inhibit chemokine production,

intermediate levels to inhibit interferon

production, and low amounts needed

to inhibit proliferation in a model system

(Wei et al., 2013). Although Honda et al.

(2014) did not investigate chemokine

production by the CD4+ effector T cells

in this study, T-cell-derived chemokines

have been shown to be important in

driving immunopathology in other sys-

tems. In an LCMV meningitis model,

CD8+ T cells displayed rapid motility in

the meninges and produced chemokines

that recruit myelomonocytic cells, leading

to fatal seizures because of proximity

to the central nervous system (Kim et al.,

2009). It is possible the T cells are not

fully desensitized, but the transient

‘‘kinapses’’ between the motile T cells

and APCs may lead to production of

chemokines that propagate inflammation.

Honda et al. (2014) conclude that PD-1

overcomes the stop signal, which in turn

sets limits on T cell delivery of cytokine

to a single APC. Results from Fife et al.

(2009) in vivo and Yokosuka et al. (2012)

in vitro concur that PD-1 engagement

overrides stop signals. In contrast, results

from Zinselmeyer et al. (2013) show that

PD-1 interactions in chronic LCMV infec-

lsevier Inc.

tion contribute to more stable interactions

of antigen-specific T cells with APCs in

the spleen. The control of acute LCMV

infection doesn’t involve stable interac-

tions between effector T cells and in-

fected targets. Chronic LCMV infection,

in contrast, led to substantial T cell decel-

eration and arrest, which was directly

shown to be dependent on PD-1 and

PD-L1 based on blocking antibodies.

The potential of PD-1 and PD-L1 in

enhancing arrest was confirmed in vitro

via freshly isolated LCMV-specific CD8+

T cells, despite the inhibition of early

TCR signaling. The mechanism by which

early TCR signaling is dissociated from

effects on T cell polarity to maintain a sta-

ble inhibitory synapse is not known. As

discussed by Honda et al. (2014), these

early data sets start to expose the huge

regulatory complexity in balancing anti-

gen, adhesion, chemokines, and diverse

cosignaling in vivo.

Honda et al. (2014) analyze a full cycle

of antigen-dependent arrest and resto-

ration of motility at the effector stage

in vivo, demonstrating a faster cycle time

than previously revealed for priming by

Mempel et al. (2004). This relatively short

period of arrest in response to a bolus in-

jection of antigen could even be captured

in a single continuousmovie (�8 hr). IFN-g

is produced during the transient arrest

and probably is being concentrated in

immunological synapses targeting partic-

ular APCs although this remains to be

demonstrated in vivo (Huse et al., 2006).

Honda et al. (2014) have taken a bold

step forward in opening the analysis of

Immunity

Previews

effector T cell desensitization at the sin-

gle-cell level. Along with single-cell anal-

ysis of other parameters, this study sets

a path to a richer picture of the immune

response that should lead to better design

of vaccines and interventions against

immunopathology.

REFERENCES

Butte, M.J., Keir, M.E., Phamduy, T.B., Sharpe,A.H., and Freeman, G.J. (2007). Immunity 27,111–122.

Fife, B.T., Pauken, K.E., Eagar, T.N., Obu, T., Wu,J., Tang, Q., Azuma, M., Krummel, M.F., and Blue-stone, J.A. (2009). Nat. Immunol. 10, 1185–1192.

Honda, T., Egen, J.G., Lammermann, T., Kasten-muller, W., Torabi-Parizi, P., and Germain, R.N.(2014). Immunity 40, this issue, 235–247.

Huse,M., Lillemeier, B.F., Kuhns,M.S., Chen, D.S.,and Davis, M.M. (2006). Nat. Immunol. 7, 247–255.

Kim, J.V., Kang, S.S., Dustin, M.L., and McGavern,D.B. (2009). Nature 457, 191–195.

Mempel, T.R., Henrickson, S.E., and Von Andrian,U.H. (2004). Nature 427, 154–159.

Immunity 40,

Ruocco,M.G., Pilones, K.A., Kawashima, N., Cam-mer, M., Huang, J., Babb, J.S., Liu, M., Formenti,S.C., Dustin, M.L., and Demaria, S. (2012).J. Clin. Invest. 122, 3718–3730.

Wei, F., Zhong, S., Ma, Z., Kong, H., Medvec, A.,Ahmed, R., Freeman, G.J., Krogsgaard, M., andRiley, J.L. (2013). Proc. Natl. Acad. Sci. USA 110,E2480–E2489.

Yokosuka, T., Takamatsu, M., Kobayashi-Ima-nishi, W., Hashimoto-Tane, A., Azuma, M., andSaito, T. (2012). J. Exp. Med. 209, 1201–1217.

Zinselmeyer, B.H., Heydari, S., Sacristan, C.,Nayak, D., Cammer, M., Herz, J., Cheng, X., Davis,S.J., Dustin, M.L., and McGavern, D.B. (2013).J. Exp. Med. 210, 757–774.

Intestinal Macrophages and DCsClose the Gap on Tolerance

Guy Shakhar1,* and Masha Kolesnikov11Department of Immunology, The Weizmann Institute of Science, Rehovot 76100, Israel*Correspondence: shakhar@weizmann.ac.ilhttp://dx.doi.org/10.1016/j.immuni.2014.01.008

CD103+ dendritic cells (DCs) must acquire soluble food antigens from the gut lumen to induce oral tolerance.In this issue of Immunity, Mazzini et al. (2014), report that CX3CR1

+ macrophages capture such antigen andtransfer it to the DCs by a route involving gap junctions.

To maintain tolerance toward proteins

from food, mononuclear phagocytes in

the gut mucosa, including the lamina

propria, must sample antigens from the

intestinal lumen and deliver them to lymph

nodes for presentation to T cells.

The lamina propria of the small intes-

tine in mice contains several populations

of mononuclear phagocytes (Figure 1).

Prominent among those are migratory

dendritic cells (DCs)—marked by the

expression of the integrins CD11c,

CD11b and CD103, and macrophages—

marked by the chemokine receptor

CX3CR1 and F4/80 (Varol et al., 2009).

Intestinal CD103+ DCs are short-lived

cells destined to migrate into the drain-

ing mesentery lymph nodes (LNs). Upon

arriving there they prime or tolerize

T cells, depending on the inflammatory

context (Laffont et al., 2010). Thus

CD103+ DCs are thought to be critical for

maintaining T cell tolerance and for elicit-

ing immune response against gut patho-

gens (Schulz et al., 2009). Themore abun-

dant CX3CR1+macrophages are relatively

long-lived phagocytes that do not nor-

mally migrate to LNs and are not as effi-

cient at presenting antigen (Ag) to T cells.

Their immunological role is debated.

When sampling particulate antigen, DCs

can manage on their own. In response

to Salmonella, for instance, DCs enter the

epithelium, send dedicated dendrites to

phagocytose the bacteria, and then pro-

cess their antigens and carry them to the

draining lymphnodes forpresentation (Far-

ache et al., 2013). In contrast, the DCs are

inefficient samplers of soluble antigens,

although their presentation is essential for

maintaining tolerance toward food anti-

gens. Surprisingly, several studies have

shown that CX3CR1+macrophages gather

this sort of antigen much more efficiently

than DCs (Schulz et al., 2009; Farache

et al., 2013). A major mechanism the mac-

rophages use is sending dendrites, depen-

dent on CX3CR1 signaling, to the gut

lumen. Antigen may also directly flow into

the lamina propria through goblet cells to

be collected there by the macrophages

(McDole et al., 2012).

The complementing specialties of the

two cell populations, one in uptake of

antigen and the other in its presentation,

raises an interesting possibility—could

CX3CR1+ macrophages and CD103+ DCs

be collaborating? Could macrophages be

in charge of capturing the Ag and deliv-

ering it to DCs that would in turn carry it

to the lymph node for presentation?

In an elaborate study, the group led by

Maria Rescigno now shows that this is

indeed the case (Mazzini et al., 2014).

Their research follows the extent, mecha-

nism, and immunological consequences

of antigen transfer betweenmacrophages

and DCs in the intestinal lamina propria of

the mouse small intestine.

The study started by verifying, by using

histology and flow cytometry, that solu-

ble ovalbumin was taken up efficiently

by CX3CR1+ macrophages, but not by

CD103+ DCs. Interestingly, uptake by

macrophages was largely limited to the

February 20, 2014 ª2014 Elsevier Inc. 171