dc-hil-expressing myelomonocytic cells are critical ... · to syndecan-4 on effector tcells, we...

DC-HIL-Expressing Myelomonocytic Cells Are CriticalPromoters of Melanoma GrowthJin-Sung Chung1, Kyoichi Tamura1, Ponciano D. Cruz Jr1 and Kiyoshi Ariizumi1

A major barrier to successful cancer immunotherapy is the tumor’s ability to induce T-cell tolerance by exploitinghost regulatory mechanisms. Having discovered the DC-HIL receptor, which inhibits T-cell responses by bindingto syndecan-4 on effector T cells, we posited the DC-HIL/syndecan-4 pathway to have an important role in cancerpromotion. Among DC-HILþ myelomonocytic cells, during growth of implanted mouse melanoma, CD11bþ

Gr1þ cells were the most expanded population and the most potent at suppressing T-cell activation. Deletion ofthe DC-HIL gene or infusion of anti-DC-HIL mAb abrogated these cells’ suppressor function and expansion, andmarkedly diminished melanoma growth and metastasis. IL-1b and IFN-g were elevated in mice bearing melanoma,and concurrent exposure to both cytokines optimally induced DC-HIL expression by tumor-infiltrating CD11bþ

Gr1þ cells. Ligation of DC-HIL transduced phosphorylation of its intracellular immunoreceptor tyrosine-basedactivation motif, which in turn induced intracellular expression of IFN-g and inducible nitric oxide synthase(iNOS), known to mediate T-cell suppression by CD11bþGr1þ cells. Thus, DC-HIL is the critical mediator of thesecells’ suppressor function in melanoma-bearing mice and a potential target for improving melanomaimmunotherapy.

Journal of Investigative Dermatology (2014) 134, 2784–2794; doi:10.1038/jid.2014.254; published online 10 July 2014

INTRODUCTIONDespite recent advances in the treatment of metastaticmelanoma, it remains the most lethal form of skin cancer, inlarge part because of its ability to overcome host antitumorimmunity (Fang et al., 2008). Among immune mechanismsexploited by melanoma are regulatory T cells, type-2 macro-phages, immature dendritic cells (DCs), and CD11bþGr1þ

cells (which overlap with myeloid-derived suppressor cells(MDSCs)) (Serafini et al., 2006). These latter cells are the mostpotent suppressors of T-cell function.

CD11bþGr1þ cells consist of myeloid progenitors(Gabrilovich and Nagaraj, 2009), which in healthy indivi-duals are confined to the bone marrow (BM), where theydifferentiate into DCs, granulocytes, or macrophages, none ofwhich are immunosuppressive. By contrast, in cancer patients,

differentiation is blocked by tumor-derived soluble factors,causing these cells’ exponential proliferation and expansioninto blood and other organs, where they become immuno-suppressive (Diaz-Montero et al., 2009).

CD11bþGr1þ cells were shown to suppress T-cell functionby producing inhibitory soluble factors such as urea/L-ornithine(produced by arginase I) (Rodriguez and Ochoa, 2008), nitricoxide (NO) (Bingisser et al., 1998), and reactive oxygen species(Kusmartsev and Gabrilovich, 2003). Co-inhibitory pathwayssuch as CD80/86 (ligands for CTLA-4) and PD-L1 (for PD-1)were shown to mediate suppressor function (Yang et al., 2006;Liu et al., 2008, 2009), but inconsistently and sometimes in acontradictory manner. Moreover, the molecular underpinningfor these cells’ suppressor function is not fully understood.

We discovered the DC-HIL receptor on antigen-presentingcells (APCs), which binds syndecan-4 (SD-4) on effector/memory (but not naive) T cells and downregulates T-cellreceptor (TCR) activation signals (Chung et al., 2007a, b). Weshowed the DC-HIL/SD-4 pathway to inhibit inflammatoryresponses in mouse models of contact hypersensitivity andgraft-versus-host disease (Chung et al., 2007a; Akiyoshi et al.,2010). Furthermore, DC-HIL is expressed by melanoma cells,and it suppresses T-cell activation via SD-4 on cytotoxicT cells (Tomihari et al., 2010), but had little to no impact onthese cells’ tumor-killing activity.

Our present study heralds a melanoma-promoting role forDC-HIL expressed by myelomonocytic cells bearing theCD11bþGr1þ phenotype. Blocking DC-HIL function viagene deletion or specific antibody (Ab) abrogated these cells’suppression of T-cell function and promotion of melanoma

See related commentary on pg 2675ORIGINAL ARTICLE

1Department of Dermatology, The University of Texas Southwestern MedicalCenter and Dermatology Section (Medical Service), Dallas Veterans AffairsMedical Center, Dallas, Texas, USA

Correspondence: Kiyoshi Ariizumi, Department of Dermatology, TheUniversity of Texas Southwestern Medical Center and Dermatology Section(Medical Service), Dallas Veterans Affairs Medical Center, 5323 Harry HinesBoulevard, Dallas, Texas 75390, USA.E-mail: [email protected]

Received 19 November 2013; revised 18 February 2014; accepted 14 March2014; accepted article preview online 17 June 2014; published online 10 July2014

Abbreviations: Ab, antibody; APC, antigen-presenting cell; BM, bonemarrow; DC, dendritic cell; iNOS, inducible nitric oxide synthase; ITAM,immunoreceptor tyrosine-based activation motif; KO, knocked out; MDSC,myeloid-derived suppressor cell; s.c., subcutaneous; SD-4, syndecan-4;TCR, T-cell receptor; WT, wild-type

2784 Journal of Investigative Dermatology (2014), Volume 134 & 2014 The Society for Investigative Dermatology

http://dx.doi.org/10.1038/jid.2014.254

mailto:[email protected]

growth. Ligated DC-HIL activated its intracellular immuno-receptor tyrosine-based activation motif (ITAM) and inducedthe expression of IFN-g and inducible nitric oxide synthase(iNOS) (producing NO). Our findings indicate that DC-HIL isresponsible for secretion of T-cell–inhibitory factors byCD11bþGr1þ cells, which render them immunosuppressive.

RESULTSDC-HIL gene deletion inhibited the growth of melanoma

We created mice knocked out (KO) for the DC-HIL gene withno gross abnormality nor developmental defects in lymphoidorgans, except for the absence of DC-HIL protein and whoseDCs are twofold greater activators of T cells than wild-type (WT) counterparts (Supplementary Figure S1 and S2online), thereby confirming DC-HIL’s inhibitory effect onT-cell activation.

B16 melanoma subcutaneously inoculated into WT versusDC-HIL� /� mice grew aggressively in the former, but itsgrowth was markedly inhibited in the latter (Figure 1a).Having reported DC-HIL expression by B16 cells to promotetumor growth by suppressing T-cell activation (Tomihari et al.,2010), we probed this influence in DC-HIL� /� mice, usingB16 cells knocked down for DC-HIL (KD-B16). KD-B16melanoma growth was slower than parental B16 melanoma,and its growth was markedly inhibited in KO mice (Figure 1b).We then examined lung metastasis (Figure 1c and d) followingi.v. (intravenous) injection of B16 cells: KO mice hadmarkedly lighter lungs, less metastatic foci, less melanin

content per lung, and less melanin per metastatic focus. Thus,melanoma growth was supported by tumor-associated DC-HILand by host-derived DC-HIL.

DC-HIL is expressed by CD11bþGr1þ cells in mice bearingmelanoma

We next addressed which DC-HIL-expressing host cellspromote melanoma growth. As DC-HIL is expressed bymyelomonocytic cells (Chung et al., 2009), we identified themost expanded DC-HILþ myeloid population from spleen ofmice with melanoma (Figure 2a), which turned out to beCD11bþGr1þ cells, representing about 10% of splenocytes(vs. CD11cþ DC or F4/80þ macrophages, which were about5% each). Forty percent of splenic CD11bþGr1þ cellsexpressed DC-HIL, whereas 20% and 66–73% of these cellswere, respectively, PD-L1þ and CD80/CD86þ (Figure 2b).For CD11bþGr1þ cells from BM, blood, and the B16 tumorsite, DC-HIL expression ranged from 27 to 61% (Supplemen-tary Figure S3a–c online). By contrast, CD11bþGr1þ cellsfrom tumor-free mice did not express DC-HIL (Figure 2c).Thus, melanoma induced DC-HIL expression of someCD11bþGr1þ cells in many organs.

Given DC-HIL’s T-cell–inhibitory function, we comparedthe three subsets of DC-HIL-expressing myelomonocytic cellsfor suppressor function (Figure 2d). Each was similarly assayedfor suppressor activity to carboxyfluorescein succinimidylester-labeled T cells activated by anti-CD3/CD28 Ab. F4/80þ

macrophages or CD11cþ DCs did not inhibit T-cell prolifera-tion even at the highest cell ratio (1:1), whereas CD11bþ

Gr1þ cells showed inhibition dose dependently. We nextexamined their T cell–stimulatory capacity by mixing CD8þ

T cells from pmel-1 transgenic mice (in which all CD8þ

T cells express the same TCR specific for gp100 Ag) (Overwijket al., 2003) with each of the three DC-HILþ subsets pulsedwith gp100 peptide. CD11cþ DC or F4/80þ macrophagesinduced strong T-cell proliferation at cell ratios of 1:0.25 orhigher (Figure 2e), whereas CD11bþGr1þ cells exhibitedsimilar stimulatory capacity at 1:0.25, while inhibiting T-cellproliferation completely. Thus, CD11bþGr1þ cells in micewith melanoma were the predominant DC-HIL-expressingpopulation and the most potent suppressors of T-cellproliferation.

DC-HIL mediated the T cell –suppressive function ofCD11bþGr1þ cells

CD11bþGr1þ cells from melanoma-bearing mice werecocultured with pmel-1 splenocytes and gp100 peptide at a1:1 cell ratio; anti-DC-HIL mAb or control IgG was added toblock DC-HIL on CD11bþGr1þ cells (Figure 3a). The mAb(but not control) restored pmel-1 T-cell activation dosedependently and completely, but it had no effect on T-cellactivation without CD11bþGr1þ cells. Anti-CD80, anti-CD86, or anti-PD-L1 Ab had no significant effect on CD11bþ

Gr1þ cells’ suppressor activity (Figure 3b). Thus, among theco-inhibitors tested, DC-HIL was responsible solely for thesuppressor activity of CD11bþGr1þ cells.

We next compared the T-cell–suppressor capacity ofDC-HILþ versus DC-HILnegCD11bþGr1þ cells. Undepleted

6

8B16 melanoma *

5DC-HIL-KD B16 *

4

KO

WT2

3

4

WT

0 6 8 10 13 15 170

2 KO

20

1 KO

Day after inoculation

Tum

or v

ol (

cm3 )

0

Tum

or v

ol (

cm3 )

1


Lungs of mice inoculated:

WT

KO

0.9Mel/focus (μg)

2.0Melanin (mg)

3# Foci (x103)

0.4

0.3

Weight (g)

0.6

1.0

1.5

*2

0.2 *

0.30.5

*

1

*0.1

0KOWT

000

5 7 9 12 14 16 19 21 23

Figure 1. Growth and metastasis of B16 melanoma are suppressed in DC-

HIL� /� mice. (a) Tumor volume of B16 cells implanted into wild-type (WT) or

DC-HIL knocked-out (KO) mice (n¼ 5). (b) Tumor volume after implanting

DC-HIL-KD-B16 cells (n¼ 5). Lung metastasis (c) at 19 days after B16 cells

were injected i.v. into WT or KO mice (n¼ 10); lung weight, number of

metastatic foci, melanin content/lung, and melanin content/focus plotted (d).

Representative data from three separate experiments, *Po0.01 versus WT.

J-S Chung et al.DC-HIL on Myeloid Suppressor Cells

www.jidonline.org 2785

http://www.jidonline.org

CD11bþGr1þ cells inhibited gp100-triggered T-cell activa-tion, whereas the DC-HIL-depleted CD11bþGr1þ fractionshowed no inhibition at all doses examined (Figure 3c).In vivo suppressor ability of DC-HILþ cells was assessed byinjecting mice with pmel-1 CD8þ T cells, followed byinfusion of undepleted or DC-HIL-depleted CD11bþGr1þ

cells and by gp100 vaccination. Ten days later, mice infusedwith CD8þ T cells but without CD11bþGr1þ cells generatedmany activated (IFN-gþ ) T cells in lymph node (Figure 3d),whereas co-infusion of undepleted CD11bþGr1þ cells led tofewer activated T cells, and co-infusion of DC-HIL-depletedCD11bþGr1þ cells prevented suppression. An experimentusing DC-HIL� /�CD11bþGr1þ cells showed similar results

(Supplementary Figure S4 online). Thus, DC-HILþ CD11bþ

Gr1þ cells were responsible for suppressor activity.We next co-injected undepleted or DC-HIL-depleted

CD11bþGr1þ cells with B16 cells subcutaneously (s.c.) intonaive mice. A week later, similarly treated CD11bþGr1þ

cells alone were infused i.v. (Figure 3e). Melanoma in miceco-injected with undepleted suppressor cells grew markedlylarger than cohorts infused with B16 cells alone, whereastumors in mice treated with DC-HIL-depleted CD11bþGr1þ

cells were similar to those given B16 alone. This outcome forDC-HIL depletion was not observed using Rag2 KO mice(Supplementary Figure S5 online), suggesting that T cells wereinvolved. Experiments using DC-HIL-deficient CD11bþGr1þ

612.7%

UntreatedTumor-free

T cellsalone

F4/80

CD11cMelanoma

CGr1 *

53% 49 51

1:0.25 1:0.5

0

CGr1 from tumor-free mice

5.4%

52 54 50

DC

-HIL

CD

11c

Gr1

17 11 6.58 * *

*F4/80

CD11cC

Gr1

4

6CGr1

2

No0

0.25

Cell ratios

CGr1 from spleen of melanoma mice

3.2% 39 26

CD

80

PD

-L1

DC

-HIL

CD

86

66 73

IgG

Gr-1

Treated

Cell ratio

IgG

F4/

80

1 :1

3 H -

TdR

(x

104

cpm

)

10.5

% DC-HIL+ leukocytes/spleen2015105

Figure 2. Melanoma induces DC-HIL expression by most potent CD11bþGr1þ suppressors. (a) Splenocytes from B16 melanoma-bearing or tumor-free mice

(n¼ 3) were assayed for % DC-HILþ cells in three myeloid populations. (b, c) CD11bþGr1þ (CGr1) cells isolated from mice with (b) or without (c) melanoma

were examined for the expression of Gr1 and co-inhibitory receptors (%). (d) These myeloid cells (increasing cell ratios) were cocultured with carboxyfluorescein

succinimidyl ester (CFSE)-labeled T cells activated by anti-CD3/CD28 Ab. T-cell proliferation (%) was determined by FACS (histograms). (e) Purified myeloid cells

were examined for T cell–stimulatory capacity. Different numbers of myeloid cells were pulsed with gp100 Ag and added to a culture of CD8þ pmel-1 T cells.

Culture of T cells with Ag served as control (No). T-cell proliferation was measured by 3H-thymidine (TdR) incorporation. *Po0.01.


2786 Journal of Investigative Dermatology (2014), Volume 134

3

4*

10

DC-HIL+

2 *68

* * *

0

1 *

24

Spl

CGr1

++++++ + + +

+++++ + +

+ + +No

0–– –

– – – –Ab 50 25

IgG IgG αDC-HILαDC-HIL

1212 25 Cell ratio

3PBS3

* 2

*

αDC

-HIL

αPD

-L1

2

αCD

80αC

D86

*Rat

IgG

1 *1

Spl 00

Day after inoculation1

0

Spl+CGr1

Mice injected: CGr1 from:

WTKOPBSNo CGr1/CD8T

0 20 40 60 80

*

4

Gr-1

3WT

2

0

1

0 0.12 0.25 0.5 1 0.12 0.25 0.5 1

CGr1/Spl

3No CGr1WT-CGr1

*

2KO-CGr1 *

*1


00

KO

3 H-T

dR (

x104

cpm

)3 H

-TdR

(x

104

cpm

)3 H

-T

dR (

x104

cpm

)

3 H-T

dR (

x104

cpm

)

5025 5025 50

10.50.25

DC-HILnegCGr1

Rab

IgG

μg ml –150 2512

DC-HIL+CG1/CD8TDC-HILnegCGr1/CD8T

# IFN-γ+ cells/2 x 105 LN

DC

-HIL

1715131197

DC-HILnegCGr1DC-HIL+CGr1

Tum

or v

ol (

cm3 )

Tum

or v

olum

e (c

m3 )

CD11bneg/Spl

17151310

Figure 3. DC-HIL mediates T cell–suppressor activity of CD11bþGr1þ (CGr1) cells. (a, b) CGr1 cells from melanoma-bearing mice were cocultured with pmel-1

splenocytes (Spl), gp100 Ag, and anti-DC-HIL mAb (a), anti-co-inhibitor antibody (Ab), or control IgG (b). 3H-TdR incorporation was measured. (c) Undepleted

(DC-HILþ ) or DC-HIL-depleted (DC-HILneg) CGr1 (Supplementary method) were assayed for the suppression of pmel-1 splenocyte proliferation triggered by Ag

(increasing ratios). (d) Mice (n¼ 5) injected with pmel-1 CD8þ T cells and DC-HILþCGr1 or DC-HILnegCGr1. Ten days after giving gp100, IFN-g-producing cells

in lymph node (LN) were counted. (e) Tumor growth following co-injection of DC-HILþ or DC-HILneg CGr1 cells with B16 cells s.c. into naive mice (n¼5). Using

similar methods, DC-HIL� /�CGr1 cells were compared with DC-HILþ /þ counterparts for DC-HIL expression by FACS (f), T cell–suppressing ability (g), and

tumor-promoting ability (CD11bneg cells as control) (h). *Po0.01.




cells from KO mice did not suppress T-cell activation orpromote melanoma progression (Figure 3f–h). Thus, DC-HILþ

CD11bþGr1þ cells were critical suppressors of T cells andpromoters of melanoma growth.

IFN-c and NO-mediated T cell–suppressive activity ofCD11bþGr1þ cells

We addressed the contribution of soluble factors to T-cellsuppression by adding specific inhibitors to cocultures ofpmel-1 splenocytes and CD11bþGr1þ cells. NeutralizingAb to TGF-b (Filipazzi et al., 2007) or IL-10 (Wang et al.,2010) had no effect on suppressor activity, whereas anti-IFN-gAb blocked suppression completely (Figure 4a). Chemical

inhibitors for arginase and indoleamine restored T-cellactivation marginally, and catalase and superoxide dismutasefor neutralizing reactive oxygen species had no effect(Figure 4b). By contrast, inhibitors of all NOS molecules orthose specific to NOS-2 blocked suppressor function substan-tially or completely. Thus, IFN-g and NOS-2 independentlymediated suppressor activity of CD11bþGr1þ cells inducedby melanoma.

T-cell expression of SD-4 is required for CD11bþGr1þ cells’suppressor activity

As the inhibitory function of DC-HIL on APC requires ligationof SD-4 on T cells, a similar mechanism was likely for

36

2 *4

*

*

1 Rat

IgG

αTG

F-β

1αI

L-10

αIF

N-γ

* 2

Ctr

l IgG

αDC

-HIL

DC

-HIL

-Fc

*

3 H-T

dR (

x104

cpm

)3 H

-TdR

(x1

04 cp

m)

3 H-T

dR (

x104

cpm

)

3 H-T

dR (

x104

cpm

)

20

Spl+CGr1 Spl+CGr1

Spl cell /CGr1

20 μg ml–10

Spl0

Spl 0 6 12 25

3NOS2*

50 μg ml–1

WT SD-4 KO6*

2

s 4* *

1

NO

S

Arg

Indo

lC

-RO

SS

-RO

S

* 2

0Spl No Inhibitors 0 0.125 0.25 0.5 1

Spl+CGr1

0

20

25αDC-HILCtrl

10

1510 min10 0 1 4

IFN

-γ m

RN

A

5 IgG

αDC

-HIL

0Day 1 Day 2

4*

p-Tyr

DC-HIL

2

340

50 Day 1Day 2

60

80

1

IgG

αDC

-HIL

20

30* 40

IgG

αDC

-HIL

*NO

S m

RN

A

0

NOS-1

NOS-2

NOS-3

Fol

d in

crea

se

IL-10TGFβ

0

10 20

NO

(μM

)

Unt

reat

ed

TNFαIFNγ 0

Figure 4. Cross-linked DC-HIL on CD11bþGr1þ cells induces tyrosine phosphorylation and IFN-c/iNOS expression. (a–c) CD11bþGr1þ (CGr1) cells

cocultured with pmel-1 splenocytes (1:1 ratio) with inhibitors, including (a) anti-cytokine antibody (Ab), (b) 5 mM L-NG-monomethyl-arginine citrate (NOSs),

0.5 mM N6-(1-iminoethyl)-L-lysine (NOS-2), 1 mM N-hydroxyl-nor-arginine (Arg), 0.2mM 1-methyl-tryptophan (indol), 1,000 U ml�1 catalase (C-ROS),

200 U ml�1 superoxide dismutase (S-ROS), and (c) anti-DC-HIL mAb or DC-HIL-Fc. 3H-TdR uptake was measured. (d) CGr1 cells cocultured with SD-4þ /þ or

SD-4� /� pmel-1 splenocytes. (e–i) At varying times after cross-linking with Ab, CGr1 cells were assayed for tyrosine phosphorylation (p-Tyr) on DC-HIL protein

(e); cytokine messenger RNA (mRNA) and secretion (f, g); mRNA of NOS genes (h); or NO production (i). Data (mean±SD, n¼ 3) are shown as fold increase

relative to control. *Po0.01.



CD11bþGr1þ cells. DC-HIL-Fc, which interferes withDC-HIL/SD-4 binding (Chung et al., 2007b), restored T-cellactivation dose dependently, whereas control Ig did not(Figure 4c). For SD-4 specificity, we compared the effects ofCD11bþGr1þ cells on SD-4þ /þ versus SD-4� /� pmel-1splenocytes with gp100 Ag (Figure 4d). Without CD11bþ

Gr1þ cells, splenocytes from both mice exhibited similarT-cell proliferation, whereas their presence inhibited SD-4þ /þ

(but not SD-4� /� ) T-cell activation. Thus, melanoma-inducedCD11bþGr1þ cells inhibited T-cell activation via binding toSD-4.

Cross-linked DC-HIL induced tyrosine phosphorylation andexpression of IFN-c and NOS-2 (iNOS) by CD11bþGr1þ cells

As DC-HIL has an intracellular ITAM (Shikano et al., 2001),we posited that ligation of DC-HIL would activate CD11bþ

Gr1þ cells’ suppressor function via ITAM. Cross-linkingDC-HIL with specific Ab induced tyrosine phosphorylationon the receptor (Figure 4e), as shown using transfectionexperiments (Chung et al., 2009).

Does activated DC-HIL induce IFN-g and iNOS in CD11bþ

Gr1þ cells? IFN-g messenger RNA in CD11bþGr1þ cells wasmarkedly increased a day after DC-HIL cross-linking, followedby a quick fall (Figure 4f). IFN-g was the predominant cytokinesecreted, followed by TNF-a, but neither IL-10 nor TGF-bwere secreted (Figure 4g). Cross-linked DC-HIL also inducediNOS expression fourfold greater than control (Figure 4h).Among NOS species, iNOS was most abundantly expressed,consistent with its known selective inducibility by IFN-g(Kamijo et al., 1994) and supported by increased NOafter DC-HIL cross-linking (Figure 4i). These outcomes werenot seen in DC-HIL� /�CD11bþGr1þ cells (SupplementaryFigure S6 online).

Anti-DC-HIL mAb treatment of melanoma-bearing micesuppressed tumor growth and reduced circulatingCD11bþGr1þ cells

We next assessed the effects of the mAb on melanoma growthand frequency of CD11bþGr1þ cells in blood. Because thesecells begin to accumulate in spleen 6 days after tumorinoculation (when tumors reach B0.1 cm3) (Cheng et al.,2008), we injected anti-DC-HIL mAb intraperitonealy thenand every other day, for six treatments. In mice treated withcontrol IgG, melanoma grew aggressively, in proportion to thefrequency of blood CD11bþGr1þ cells. The mAb markedlysuppressed subsequent melanoma growth (Figure 5a) and pre-vented expansion of CD11bþGr1þ cells in blood (Figure 5band c): the latter effect was supported by the inability of KOmice to expand CD11bþGr1þ cells (Supplementary Figure S7online). It also significantly enhanced the IFN-g response byT cells from mice with melanoma (Figure 5d). Blockedexpansion may be due to reduced tumor size with lesssecretion of relevant soluble factors.

Inhibited CD11bþGr1þ cell function accounted for thebeneficial effects of anti-DC-HIL mAb on melanoma

Because DC-HIL is expressed by B16 cells, APC, and CD11bþ

Gr1þ cells, we compared their respective contributions via

effects of anti-DC-HIL mAb. For DC-HIL on melanoma itself,we implanted DC-HIL-knocked-down B16 cells (KD-B16) intoWT mice, while injecting the mAb as before (Figure 5e). In thisassay, in which CD11bþGr1þ cells and APCs were both DC-HILþ , KD-B16 tumor hardly grew in mice treated with anti-DC-HIL mAb. Thus, DC-HIL on melanoma was not critical tothe overall effect of anti-DC-HIL mAb. When DC-HILþ DCswere infused into DC-HIL� /� mice bearing DC-HILþ B16tumor, mAb treatment had no significant effect on the APCfunction (Supplementary Figure S8 online). Finally, we gavethe mAb to DC-HIL� /� mice inoculated with B16 cells andCD11bþGr1þ cells (both DC-HILþ ), and it blocked tumorpromotion (Figure 5f), indicating that its beneficial effects aredue primarily to neutralizing CD11bþGr1þ cells’ function.

DC-HIL was not required for growth of EL-4 lymphoma and LL2lung carcinoma

Does DC-HIL contribute to the growth of other malignancies?EL-4 or LL2 cells implanted s.c. into DC-HIL� /� or WT miceshowed no significant difference in growth (Figures 6a,P40.05). Neither tumor was associated with significantDC-HIL expression on CD11bþGr1þ cells (Figure 6b). Usingthe carboxyfluorescein succinimidyl ester-dilution assay(Figure 6c), CD11bþGr1þ cells from spleen of mice withLL2 or EL-4 tumor showed some suppressor activity that wassignificantly weaker than those from melanoma. Thus, DC-HILdid not contribute to the growth of EL-4 and LL2 malignancies.

Inducible DC-HIL expression by CD11bþGr1þ cells requiredIL-1b and IFN-cTo identify inducers of DC-HIL expression on CD11bþGr1þ

cells, we examined soluble factors that are known to activatethese cells from lineage marker–negative (Linneg) BM cellscultured in GM-CSF/IL-4 (Cheng et al., 2008). GM-CSF/IL-4treatment alone generated 80% CD11bþGr1þ cells, withoutsignificant DC-HIL expression. IL-1b, IFN-g, and TNF-a eachinduced DC-HIL expression by 20–40% (Figure 6d). Com-bined IL-1b and IFN-g treatment amplified DC-HIL expressionto 60%. TNF-a enhanced the effects of IL-1b and of IFN-g,while suppressing CD11bþGr1þ cell proliferation. Thus,IL-1b and IFN-g synergistically triggered DC-HIL expressionby CD11bþGr1þ cells.

Could differential secretion of cytokines account for thedistinctive behavior of B16 melanoma versus other tumors?Because B16 cells in vitro did not secrete IL-1b or IFN-g(Supplementary Figure S9 online), we assayed blood cytokinesfrom mice bearing B16, LL2, or EL-4 tumor (Figure 6e). SerumIL-1b and IFN-g were elevated significantly in mice withmelanoma, but not in mice with other tumors, tumor-freemice, or KO mice with melanoma. No other cytokines testedwere expressed preferentially by mice with melanoma,thereby linking melanoma to IL-1b/IFN-g secretion by hostcells.

Do IL-1b and IFN-g induce CD11bþGr1þ cells to expressDC-HIL? We injected both cytokines into LL2 tumors on miceand examined tumor-infiltrating CD11bþGr1þ cells for DC-HIL expression (Figure 6f). Injection of either cytokine inducedmarginal expression by LL2-infiltrating CD11bþGr1þ cells.




Co-injection of both cytokines enhanced DC-HIL expression,and repeated concurrent injections of both IL-1b and IFN-g(but not just either cytokine or phosphate-buffered saline)promoted LL2 tumor growth (Figure 6g). Thus, melanomastromal cells may secrete IL-1b and IFN-g that activate the DC-HIL pathway.

DISCUSSIONThe detrimental effect of an expanding population ofT-cell suppressors in cancer patients is well established, butthe mechanism underlying these cells’ profound immuno-suppression remains unclear. We now show that melanomais associated with an autocrine loop in which theT cell–inhibitory receptor DC-HIL induced on CD11bþGr1þ

myelomonocytic cells promotes melanoma growth. Binding ofDC-HIL on CD11bþGr1þ cells to SD-4 on effector T cellsinactivates the latter, while also transducing ITAM signals inthe former that nudge the IFN-g/iNOS axis into secreting NO.Although this mechanism did not apply to models of lungcarcinoma and lymphoma, it is premature to consider itdistinct for melanoma, as studies involving other malignanciesare ongoing.

Researchers have used the term MDSCs for cells coexpres-sing CD11b and Gr1 and exerting T cell–inhibitory activity.We have avoided the MDSC label because of phenotypic andfunctional heterogeneity among these cells. In fact, we sortedmelanoma-CD11bþGr1þ cells into Ly6ChighGr1low, Ly6C-medGr1hgih, Ly6CmedGr1low, and Ly6ClowGr1low subsets, with

40Mice/IgG

Ab injection

8

30Mice/αDC-HIL

*

*

*

6 Mice/αDC-HIL

Mice/IgG

204

102 %

of C

Gr1

in P

BM

C


00


Tum

or v

olum

e (c

m3 )

00

Day 18Day 0 Day 14Day 1015 IgG

IgG 32126.83.0%

10

αDC-HIL

4.41.60.93.15

CD

11b

αDC

-HIL

IFN

-γ+ c

ells

/1 x

104

0

Gr-1 Spleen

*

*

DLN

Ab injection Ab injection

4Mice/KD-B16/IgG

2

No CGr13 Mice/KD-B16/αDC-HIL *

*

*

*

*

*

1.5

CGr1/IgG

CGr1/αDC-HIL

2 1

0.5

0

1

Tum

or v

olum

e (c

m3 )

Tum

or v

olum

e (c

m3 )

1


00

Day after inoculation3 5 7 10 12 14 17 19 21

6 8 10 13 15 17 20

10 13 15 17 20

10 14 18

Figure 5. Infusion of anti-DC-HIL mAb suppresses melanoma growth and expansion of CD11bþGr1þ (CGr1) cells. (a) Tumor volume 6 days after implanting

B16 cells into wild-type (WT) mice (n¼ 7), and mice injected with anti-DC-HIL mAb or control IgG on indicated days (closed arrows). On specific days (gray

arrows in a), blood taken from mouse, CGr1 cells counted from peripheral blood mononuclear cells (PBMCs) by FACS (b), and data summarized (c). (d) A day after

three injections, IFN-g-secreting cells in spleen or lymph node (LN) in each mouse (n¼ 3) counted as number per 1� 104 cells. (e) Tumor volume on mice treated

with the two antibodies (Abs) 11 days after implanting KD-B16 cells (n¼ 7). (f) Tumor volume on DC-HIL� /� mice treated with Ab 7 days after co-injection of B16

and CGr1 cells. *Po0.001.



3LL2 **

4

**2

KOWT

2

3

KO

WT

EL-4

11

0

Tum

or v

ol (

cm3 )

0



EL-4 LL2

3.56.8%

Gr1

100

100

101

101

102

102

103

103

10

10100

100

101

101

102

102

102

103

10

10 100101102103 10 100101102103 10 100101102103 10 100101102103 10

DC

-HIL

PBS IL-1β IL-1β/IFNγIFNγ

5% 10%11% 42%

DC

-HIL

Gr1

T cellalone

91%

Cell ratios

CGr1: 8.92941%

1:0.25 1:0.5 1 :1

B16

4

3

PBS

**

*IL-1β

2*

IFN-γIL-1β/IFNγ

Tum

or v

ol (

cm3 )

0

1

1 6 11 13 15 184 8Day after inoculation

EL-4

657486%

76% 58 40

LL2

No

IL-6

IL-1β *

IL-12

IL-10

IFNγ

IL-13

*PGE2

TGFβ

TNFα *

IFNγ/TNFα

IL-1β/

IFNγ/TNFα

IL-1β/TNFα

IFNγ/IL-1β

VEGF

0 20 40 60 80 0 20 40 60 80 100% CGr1 % DC-HIL+ CGr1

IL-17A6

4

0

2ng m

l–1

ng m

l–1

TGFβ0.8

0

0.4

TNFα

pg m

l–1

pg m

l–1

pg m

l–1

600

200

400

0

* IFNγ

IL-1β

200

100

0

*N

one

B16 LL

2

EL-

4

0.4

0.2

0

Non

e

B16

WT KO

Figure 6. IFN-c and IL-1b induce DC-HIL expression by CD11bþGr1þ (CGr1) cells. (a) Growth of EL-4 or LL2 tumor in wild-type (WT) or DC-HIL� /� mice

(n¼ 5, **P40.05). CGr1 cells from those mice were examined for Gr1 versus DC-HIL expression (b) and suppression of carboxyfluorescein succinimidyl ester

(CFSE)-labeled T cells activated by anti-CD3/CD28 (c). (d) Linneg BM cells cultured with cytokines were assayed for % CGr1 or DC-HILþ cells among CD11cþ /

CD19þ -depleted or total CGr1 cells, respectively (mean±SD, n¼ 3). (e) Blood from WT or DC-HIL knocked-out (KO) mice (n¼ 5) with or without (None) tumor

measured for cytokines. (f) DC-HIL versus Gr1 expression by CGr1 cells within LL2 tumor on mice one day after the second intratumoral injection of cytokines. (g)

Tumor volume after cytokines injected intratumorally (arrows) into LL2 tumor-bearing mice. *Po0.01.




Ly6ChighGr1low monocytes as highest expressors of DC-HILand sole suppressors of T cells (Supplementary Figure S3d and eonline).

Because CD80, CD86, and PD-L1 deliver negative signalsto T cells through corresponding receptors (Egen et al., 2002;Keir et al., 2007), each might mediate CD11bþGr1þ cells’effects in melanoma-bearing mice. Indeed, these cellsexpressed CD80, CD86, and PD-L1 at levels similar orgreater than DC-HIL, but only DC-HIL was responsible forsuppressor activity. This distinct DC-HIL property may be dueits transduction of bidirectional signals: binding to SD-4activates a protein tyrosine phosphatase CD148 in T cells(Chung et al., 2011), whereas tyrosine phosphorylation of itsITAM motif in CD11bþGr1þ cells leads to IFN-g and iNOSexpression. Our preliminary results suggest that ligated DC-HIL on CD11bþGr1þ cells activates Syk kinase, a majorsignal mediator for ITAM (Lowell, 2011). Unlike DC-HIL,CD80/CD86 and PD-L1 lack tyrosine-based signal motifs,restricting them to unidirectional signaling via respectiveTCRs (Carreno and Collins, 2002; Sharpe et al., 2007).

CD11bþGr1þ cells differ from APCs in the mechanism(s)for suppressing T-cell function. The latter appears to befocused on the downregulation of TCR signals, whereas theformer may use the same pathway plus other avenues. Thus,classic co-inhibitors may not be as critical for CD11bþGr1þ

cells’ suppressor activity.IFN-g might have simultaneously opposing effects on

cancers. Its well-known antitumor benefits (Kerkar et al.,2011; Takeda et al., 2011) are mediated via activation andproliferation of T cells and natural killer cells, whereas itsprotumor effects are transmitted via MDSCs. We showed thatIFN-g’s deleterious protumor effects arise not only by inducingDC-HIL expression on MDSCs but also by MDSCs expressingthe cytokine to mediate MDSCs’ T-cell–suppressor function.These data are consistent with MDSCs from IFN-g� /� micewith colon carcinoma that lost their suppressor activity (Gallinaet al., 2006). Indeed, we showed IFN-g� /� CD11bþGr1þ

cells to be poor tumor promoters, similar to DC-HIL� /�

CD11bþGr1þ cells (Supplementary Figure S10 online).MDSC’s phenotypic and functional plasticity to microenvir-

onmental conditions may apply to different cancers in whichdivergent soluble factors produced by the tumors or theirstromal cells drive MDSC differentiation toward particularsubsets (Condamine and Gabrilovich, 2011). IL-1b-over-expressing breast cancer honed MDSC differentiation towardthe Gr1highLy6Cneg phenotype with highest ability to suppressnatural killer cell activity (Elkabets et al., 2010), consistentwith our DC-HIL-expressing fraction among CD11bþGr1þ

cells responsible for melanoma-induced immunosuppression.The DC-HIL phenotype may be nurtured by IL-1b and IFN-gsecreted by melanoma stromal cells, as elicited by ourexperimental administration of both cytokines into a non-melanoma LL2 carcinoma. Highly elevated IL-1b and IFN-gexpression was reported in skin and lymph node bearingspontaneous melanoma in a transgenic mouse model (Meyeret al., 2011).

High DC-HIL expression on blood CD14þHLA-DRno/low

cells (the equivalent of mouse CD11bþGr1þ cells) in

melanoma patients and the restoration of suppressed T-cellfunction in those patients using anti-DC-HIL mAb (seeaccompanying submitted manuscript) provide further rationalefor exploring DC-HIL-expressing myelomonocytic suppressorcells as a target for improving melanoma immunotherapy.

MATERIALS AND METHODSReagents, animal, cell culture, and leukocyte preparations

Antibodies were purchased from eBioscience (San Diego, CA) and

Upstate Biotechnology (Lake Placid, NY); recombinant cytokines

were purchased from Pepro Tech (Rocky Hill, NJ); and chemical

inhibitors were purchased from Sigma-Aldrich (St Louis, MO). We

generated 1E4 rat anti-mouse DC-HIL (for immunoblotting) and

UTX103 rabbit anti-mouse DC-HIL (for flow cytometric analysis

and functional blocking) (Chung et al., 2009). DC-HIL-Fc fusion

protein was produced (Chung et al., 2007b). C57BL/6 mice and pmel-

1 TCR transgenic mice were purchased from Harlan (Indianapolis, IN)

and Jackson Laboratory (West Grove, PA), respectively. The DC-

HIL� /� mice were generated by the deletion of exons 2 and 3

of the gene, resulting in depletion specifically of the DC-HIL protein

(Supplementary Figure S1 online). SD-4� /� mice were obtained

from Dr Kojima (Nagoya University) (Ishiguro et al., 2000), and

SD-4� /� pmel-1 mice were produced by breeding between these

strains.

All tumor cell lines were purchased from the American Type

Culture Collection (Manassas, VA). CGr1 and/or CD11bneg cells

were isolated from spleen, blood, or tumor-infiltrating cells of

mice (2–3 weeks after implanting with tumor cells) (Chung et al.,

2014). Depletion of DC-HILþ cells from the CGr1 preparation

was performed using UTX103 mAb-coated magnetic beads, and

BM-DC, splenic CD11cþ , or F4/80þ cells were prepared (Chung

et al., 2014).

Tumor growth assays

Tumor cells (5� 105) were injected s.c. into the right shaved flank of

mice. Tumor volume was measured (Tomihari et al., 2010). Eight

days after implanting LL2 cells, phosphate-buffered saline or IL-1band/or IFN-g (50 ng per tumor) was injected into the tumor four times

every 2 days. CD11bþGr1þ (CGr1) cells were purified from the

tumor the day after the second injection. For lung metastasis, B16

cells (1� 106) were injected into mice via the lateral tail veil. Lungs

were harvested 18 or 19 days post injection, and their weight,

metastatic foci, and melanin content were determined (Tomihari

et al., 2010).

T-cell assays

Myeloid cells purified from splenocytes of tumor-bearing mice

(2–3 weeks after implanting s.c.) were cocultured with WT-T cells

or pmel-1 splenocytes (1� 105 cells per well) at varying cell ratios

with anti-CD3/CD28 Ab or gp100 peptide (each 1mg ml� 1), respec-

tively, for 3 days. Inhibitors were added to splenocytes/CGr1 cells

(1:1 ratio). CGr1 cells were also cocultured with SD-4þ /þ or

SD-4� /� pmel-1 splenocytes. For T cell–stimulatory assays, myeloid

cells were pulsed with 1mg ml� 1 hgp100 peptide and cocultured

with pmel-1 CD8þ T cells at increasing cell ratios. After 2 days of

culture, T-cell activation was measured by carboxyfluorescein succi-

nimidyl ester-dilution assay or 3H-thymidine uptake (Chung et al.,

2007b).



Cross-linking and soluble factor assaysCGr1 cells (5� 106) were cross-linked with anti-DC-HIL mAb and

assayed for tyrosine phosphorylation (Chung et al., 2009). CGr1

cells were also cultured in 96-well plates (2� 105 cells per well)

precoated with anti-DC-HIL mAb or control IgG (10mg ml� 1).

After 1 or 2 days of culture, the culture supernatant and cell pellets

were collected, respectively, for cytokine secretion using ELISA and

whole-cell extracts for measuring NO2 production using Griess

method (De Santo et al., 2005) or total RNA for measuring IFN-gand NOS messenger RNA using quantitative reverse-transcriptase–

PCR.

Antibody treatment

On day 6 or 11 after inoculating 5� 105 B16 or KD-B16 cells s.c.,

mice were sorted to two groups with similar tumor volume

(B0.1 cm3) and injected intraperitonealy with Ab (200mg per mouse)

every other day for —five to six injections. At varying times, blood

was taken from tails of five representative mice per group, depleted of

red blood cells, pulsed with OVA257–264 peptide (1mg ml� 1), and

examined by FACS for % CGr1 in total H-2Kbþ cells.

Generation of CGr1 cells

Linneg BM cells isolated using the Lineage cell depletion kit (Miltenyi,

Auburn, CA) were cultured (5� 105 cells per well) for 5 days with

GM-CSF/IL-4 (each 10 ng ml� 1) plus cytokines (10–20mng ml� 1).

Cultured cells were depleted of CD11cþ /CD19þ cells (10–15% in

total cells) and examined for % CGr1 cells and DC-HIL expression.

Adoptive transfer

For T-cell suppression, mice were injected i.v. with pmel-1 CD8þ

T cells (1� 106 cells per mouse); 3 days later, they were injected i.v.

with undepleted or DC-HIL-depleted CGr1 (1� 106 cells per mouse)

and vaccinated with Freund’s adjuvant/gp100 peptide (1 mg ml� 1);

and 10 days later lymph node cells were restimulated by culturing for

2 days with gp100 peptide in enzyme-linked immunospot wells, and

IFN-g-secreting cells were counted. For melanoma promotion,

2� 106 CGr1 and 2� 105 B16 cells were co-injected s.c. into mice.

A week later, mice were injected again with CGr1 cells. Some mice

were injected intraperitonealy with Ab (200mg per mouse) starting on

day 7, every other day, for five injections.

Serum cytokines

Two weeks post implantation of tumor cells (5� 105 cells per mouse),

sera were taken and albumin was removed by passing through the

Swell Blue albumin removal spin column (Pierce, Rockford, IL) before

ELISA assays for cytokines.

Statistical analysis

Statistical analyses were performed using two-sided Student’s t-test,

with P (o0.01) considered significant. All data shown are represen-

tative of at least two independent experiments.

CONFLICT OF INTERESTThe authors declare no conflict of interest.

ACKNOWLEDGMENTSWe thank Irene Dougherty and Megan Randolph for technical and adminis-trative assistance, respectively. This research was supported by the NationalInstitutes of Health grant (AI064927-05).

SUPPLEMENTARY MATERIAL

Supplementary material is linked to the online version of the paper at http://www.nature.com/jid

REFERENCES

Akiyoshi H, Chung JS, Tomihari M et al. (2010) Depleting syndecan-4þT lymphocytes using toxin-bearing dendritic cell-associated heparansulfate proteoglycan-dependent integrin ligand: a new opportunity fortreating activated T cell-driven disease. J Immunol 184:3554–61

Bingisser RM, Tilbrook PA, Holt PG et al. (1998) Macrophage-derived nitricoxide regulates T cell activation via reversible disruption of the Jak3/STAT5 signaling pathway. J Immunol 160:5729–34

Carreno BM, Collins M (2002) The B7 family of ligands and its receptors:new pathways for costimulation and inhibition of immune responses.Annu Rev Immunol 20:29–53

Cheng P, Corzo CA, Luetteke N et al. (2008) Inhibition of dendritic celldifferentiation and accumulation of myeloid-derived suppressor cells incancer is regulated by S100A9 protein. J Exp Med 205:2235–49

Chung JS, Cruz PD Jr, Ariizumi K (2011) Inhibition of T-cell activation bysyndecan-4 is mediated by CD148 through protein tyrosine phosphataseactivity. Eur J Immunol 41:1794–9

Chung JS, Dougherty I, Cruz PD Jr et al. (2007a) Syndecan-4 mediates thecoinhibitory function of DC-HIL on T cell activation. J Immunol 179:5778–84

Chung JS, Sato K, Dougherty II et al. (2007b) DC-HIL is a negative regulator ofT lymphocyte activation. Blood 109:4320–7

Chung JS, Tamura K, Akiyoshi H et al. (2014) The DC-HIL/syndecan-4 pathwayregulates autoimmune responses through myeloid-derived suppressorcells. J Immunol 192:2576–84

Chung JS, Yudate T, Tomihari M et al. (2009) Binding of DC-HIL todermatophytic fungi induces tyrosine phosphorylation and potentiatesantigen presenting cell function. J Immunol 183:5190–8

Condamine T, Gabrilovich DI (2011) Molecular mechanisms regulatingmyeloid-derived suppressor cell differentiation and function. TrendsImmunol 32:19–25

De Santo C, Serafini P, Marigo I et al. (2005) Nitroaspirin corrects immunedysfunction in tumor-bearing hosts and promotes tumor eradication bycancer vaccination. Proc Natl Acad Sci USA 102:4185–90

Diaz-Montero CM, Salem ML, Nishimura MI et al. (2009) Increased circulatingmyeloid-derived suppressor cells correlate with clinical cancer stage,metastatic tumor burden, and doxorubicin-cyclophosphamide chemo-therapy. Cancer Immunol Immunother 58:49–59

Egen JG, Kuhns MS, Allison JP (2002) CTLA-4: new insights into its biologicalfunction and use in tumor immunotherapy. Nat Immunol 3:611–8

Elkabets M, Ribeiro VS, Dinarello CA et al. (2010) IL-1beta regulates a novelmyeloid-derived suppressor cell subset that impairs NK cell developmentand function. Eur J Immunol 40:3347–57

Fang L, Lonsdorf AS, Hwang ST (2008) Immunotherapy for advancedmelanoma. J Invest Dermatol 128:2596–605

Filipazzi P, Valenti R, Huber V et al. (2007) Identification of a new subset ofmyeloid suppressor cells in peripheral blood of melanoma patients withmodulation by a granulocyte-macrophage colony-stimulation factor-based antitumor vaccine. J Clin Oncol 25:2546–53

Gabrilovich DI, Nagaraj S (2009) Myeloid-derived suppressor cells as regula-tors of the immune system. Nat Rev Immunol 9:162–74

Gallina G, Dolcetti L, Serafini P et al. (2006) Tumors induce a subset ofinflammatory monocytes with immunosuppressive activity on CD8þT cells. J Clin Invest 116:2777–90

Ishiguro K, Kadomatsu K, Kojima T et al. (2000) Syndecan-4 deficiency impairsfocal adhesion formation only under restricted conditions. J Biol Chem275:5249–52

Kamijo R, Harada H, Matsuyama T et al. (1994) Requirement for transcriptionfactor IRF-1 in NO synthase induction in macrophages. Science 263:1612–5

Keir ME, Francisco LM, Sharpe AH (2007) PD-1 and its ligands in T-cellimmunity. Curr Opin Immunol 19:309–14



http://www.nature.com/jid

http://www.nature.com/jid


Kerkar SP, Goldszmid RS, Muranski P et al. (2011) IL-12 triggers a program-matic change in dysfunctional myeloid-derived cells within mousetumors. J Clin Invest 121:4746–57

Kusmartsev S, Gabrilovich DI (2003) Inhibition of myeloid cell differentiationin cancer: the role of reactive oxygen species. J Leukoc Biol 74:186–96

Liu Y, Yu Y, Yang S et al. (2009) Regulation of arginase I activity andexpression by both PD-1 and CTLA-4 on the myeloid-derived suppressorcells. Cancer Immunol Immunother 58:687–97

Liu Y, Zeng B, Zhang Z et al. (2008) B7-H1 on myeloid-derived suppressorcells in immune suppression by a mouse model of ovarian cancer. ClinImmunol 129:471–81

Lowell CA (2011) Src-family and Syk kinases in activating and inhibitorypathways in innate immune cells: signaling cross talk. Cold Spring HarbPerspect Biol 3, piia002352

Meyer C, Sevko A, Ramacher M et al. (2011) Chronic inflammation promotesmyeloid-derived suppressor cell activation blocking antitumor immunityin transgenic mouse melanoma model. Proc Natl Acad Sci USA108:17111–6

Overwijk WW, Theoret MR, Finkelstein SE et al. (2003) Tumor regression andautoimmunity after reversal of a functionally tolerant state of self-reactiveCD8þ T cells. J Exp Med 198:569–80

Rodriguez PC, Ochoa AC (2008) Arginine regulation by myeloid derivedsuppressor cells and tolerance in cancer: mechanisms and therapeuticperspectives. Immunol Rev 222:180–91

Serafini P, Borrello I, Bronte V (2006) Myeloid suppressor cells in cancer:recruitment, phenotype, properties, and mechanisms of immune suppres-sion. Semin Cancer Biol 16:53–65

Sharpe AH, Wherry EJ, Ahmed R et al. (2007) The function of programmed celldeath 1 and its ligands in regulating autoimmunity and infection. NatImmunol 8:239–45

Shikano S, Bonkobara M, Zukas PK et al. (2001) Molecular cloningof a dendritic cell-associated transmembrane protein, DC-HIL,that promotes RGD-dependent adhesion of endothelial cells throughrecognition of heparan sulfate proteoglycans. J Biol Chem 276:8125–34

Takeda K, Nakayama M, Sakaki M et al. (2011) IFN-g production by lung NKcells is critical for the natural resistance to pulmonary metastasis of B16melanoma in mice. J Leukoc Biol 90:777–85

Tomihari M, Chung JS, Akiyoshi H et al. (2010) DC-HIL/glycoprotein Nmbpromotes growth of melanoma in mice by inhibiting the activation oftumor-reactive T cells. Cancer Res 70:5778–87

Wang L, Liu JQ, Talebian F et al. (2010) Tumor expression of CD200inhibits IL-10 production by tumor-associated myeloid cells andprevents tumor immune evasion of CTL therapy. Eur J Immunol 40:2569–79

Yang R, Cai Z, Zhang Y et al. (2006) CD80 in immune suppression by mouseovarian carcinoma-associated Gr-1þCD11bþ myeloid cells. CancerRes 66:6807–15



dc-hil-expressing myelomonocytic cells are critical ... · to syndecan-4 on effector tcells, we...

Documents