mTOR limits the recruitment ofCD11b�Gr1�Ly6Chigh myeloid-derivedsuppressor cells in protecting againstmurine immunological hepatic injury

Yan Zhang,*,†,1 Yujing Bi,‡,1 Hui Yang,*,†,1 Xi Chen,*,†,1 Huanrong Liu,*,†,1 Yun Lu,*,†

Zhengguo Zhang,*,† Jiongbo Liao,*,† Shan Yang,*,† Yiwei Chu,*,† Ruifu Yang,‡

and Guangwei Liu*,†,2

*Key Laboratory of Medical Molecular Virology of Ministries of Education and Health, Department of Immunology, School ofBasic Medical Sciences, and †Biotherapy Research Center, Fudan University, Shanghai, China; and ‡State Key Laboratory of

Pathogen and Biosecurity, Beijing Institute of Microbiology and Epidemiology, Beijing, China

RECEIVED SEPTEMBER 1, 2013; REVISED JANUARY 15, 2014; ACCEPTED FEBRUARY 2, 2014. DOI: 10.1189/jlb.0913473

ABSTRACT

The mTOR pathway integrates diverse environmental

inputs, including immune signals and metabolic cues, to

direct the innate and adaptive immune responses.

MDSCs are a heterogeneous cell population that plays a

crucial regulatory effect in immune-related diseases.

However, whether mTOR signaling affects the functions

of MDSCs remains largely unknown. Here, we show

that mTOR signaling is a pivotal negative determinant of

MDSC recruitment in IMH disease. In the context of

IMH, inhibition of mTOR with rapamycin in CD11b�Gr1�

MDSCs mediates protection against IMH and serves as

a functional, suppressive immune modulator that re-

sults in increased CD11b�Gr1�Ly6Chigh MDSC recruit-

ment to inflammatory sites. In agreement with this,

mTOR down-regulation promotes CD11b�Gr1�Ly6Chigh

MDSC migration in vitro and in vivo. Mechanistically,

mTOR activity down-regulation in MDSCs induced iNOS

expression and NO production. Pharmacologic inhibi-

tion of iNOS completely eliminated MDSC recruitment.

This study identifies MDSCs as an essential component

for protection against IMH following rapamycin treat-

ment. Rapamycin treatment or mTOR inhibition pro-

motes CD11b�Gr1�Ly6Chigh MDSC recruitment and is

critically required for protection against hepatic injury.

This study further validates the targeting of mTOR sig-

naling as a potential therapeutic approach to IMH-re-

lated diseases. J. Leukoc. Biol. 95: 000–000; 2014.

Introduction

The mTOR is an evolutionarily conserved and ubiquitously ex-

pressed controller of cell growth, proliferation, and survival [1–

5]. Rapamycin [i.e., sirolimus (Rapamune)] is a specific inhibitor

of mTOR and is a commonly used pharmacologic tool for the

study of mTOR biology [6–8]. It binds with the FK-binding pro-

tein 12 complex that binds with high affinity to mTOR, prevent-

ing activation of p70S6 kinase and 4E-BP1, which provide critical

signals that allow for cell proliferation [9–11]. For instance, in

organ transplantation, where rapamycin has been used effectively

to prevent organ allograft rejection, IL-2 stimulation of lympho-

cyte proliferation through p70S6 kinase is blocked [9, 12]. In

addition, rapamycin is approved by the U.S. Food and Drug Ad-

ministration for immunosuppression in transplant patients, can-

cer chemotherapy, and the prevention of local coronary artery-

stent thrombosis [13, 14].

MDSCs are comprised of a heterogeneous cell population that

suppresses T cell proliferation and function, blocks NK cell cyto-

toxicity, and promotes the development of Tregs in tumor-bearing

hosts [15–17]. MDSCs have been detected in the blood of cancer

patients and in the bone marrow, spleen, and peripheral blood

of tumor-bearing mice [18, 19]. Murine MDSCs are defined as

CD11b�Gr1� myeloid cells that suppress T cell proliferation [20,

21]. Although much has been learned about the immunosup-

pressive function of MDSCs in a variety of cancers, less is known

about liver MDSCs, especially with respect to their specific regula-

tory mechanisms in immunological liver disease.

In the present study, we have sought to determine whether

MDSCs are an essential immune component in rapamycin-

mediated protection against CIH and PIH, two typical immu-

nological hepatic injury models. Rapamycin treatment signifi-

cantly diminished pathological hepatic lesions and mortality in

1. These authors contributed equally to this work as cofirst authors.

2. Correspondence: Fudan University, Yixueyuan Rd. 138, Xuhui District,Shanghai, China 200032. E-mail: liugw@fudan.edu.cn

Abbreviations: ALT�alanine aminotransferase, AST�aspartate aminotrans-

ferase, B6�C57BL/6, CIH�Con A-induced hepatitis, IMH�immune-medi-

ated hepatic injury, L-NMMA�L-NG-monomethyl arginine acetate,

MDSC�myeloid-derived suppressor cell, mTOR�mechanistic target of

rapamycin, PCl�picryl chloride, PD-L1�programmed cell death protein 1

ligand, PIH�picryl chloride-induced hepatitis, Treg�regulatory T cell

The online version of this paper, found at www.jleukbio.org, includes

supplemental information.

Article

0741-5400/14/0095-0001 © Society for Leukocyte Biology Volume 95, June 2014 Journal of Leukocyte Biology 1

Epub ahead of print February 25, 2014 - doi:10.1189/jlb.0913473

Copyright 2014 by The Society for Leukocyte Biology.

a mouse hepatic injury model. In addition, it significantly re-

duced inflammatory cell infiltration concomitant with en-

hanced CD11b�Gr1� MDSC accumulation in inflamed liver

tissue. These cells are required for the rapamycin-mediated

protection against immunological hepatic injury. Thus, the

current study shows MDSCs to be an essential immune compo-

nent in rapamycin-mediated protection against liver damage.

MATERIALS AND METHODS

Mice and the hepatic injury modelCD45.1� B6 mice were obtained from the Center of Model Animal Research

at Nanjing University (Nanjing, China). mTORflox/flox and Rosa26-Cre-ERT2

mice were obtained from The Jackson Laboratory (Bar Harbor, ME, USA).

mTORflox/flox were bred with Rosa26-Cre-ERT2 mice and had been back-

crossed to the B6 background for at least eight generations. A Cre-ER fusion

gene was recombined into the ubiquitously expressed Rosa26 locus to gener-

ate mTORflox/floxRosa26-Cre-ERT2 mice (called mTORCreER mice hereafter).

For in vivo tamoxifen treatment, WT or mTORCreER mice were injected i.p.

with 2 mg tamoxifen (T5648; Sigma-Aldrich, St. Louis, MO, USA)/mouse for

4 consecutive days, a dose that does not produce toxicity in mice. B6

(CD45.2�) mice were obtained from the Fudan University Experimental Ani-

mal Center (Shanghai, China). All of the mice were bred and maintained in

specific pathogen-free conditions. Sex-matched littermate mice, 6–8 weeks of

age, were mainly used for the experiments. The experimental protocols were

approved by the Animal Ethics Committee of Fudan University.

CIH and PIH mouse models were used for the hepatic injury studies.

For CIH induction, age- and gender-matched mice received an i.v. injec-

tion of Con A (C2010; Sigma-Aldrich) at a dose of 15 or 25 mg/kg body

weight. A more lethal dose (25 mg/kg) was used to generate survival

curves, and the lower dose (15 mg/kg) allowed assessment of liver inflam-

matory cell infiltration, liver pathology, and other parameters following

Con A challenge. For PIH induction, age- and gender-matched mice were

sensitized twice by painting 0.1 ml 1% PCl (dissolved in ethanol; Nacalai

Tesque, Kyoto, Japan) on the skin of the abdomen with an interval of 5

days (inductive phase) [22]. Five days after the second sensitization, livers

were injected with 10 �l 0.5% PCl (dissolved in olive oil). Mice were exsan-

guinated, and livers were removed under anesthesia in the effector phase

(�20 h after PCI injection). Injured liver samples were fixed in 4% PFA,

embedded in paraffin, and stained by H&E. Serum ALT enzyme activity

was determined using a kit from Cayman Chemical (Ann Arbor, MI, USA).

Cell purification and cultureFollowing cardiac perfusion with PBS, the liver was aseptically removed and

mechanically disrupted between sterile-frosted microscope slides, as de-

scribed previously [23, 24]. Liver CD11b� cells were isolated using anti-

CD11b magnetic beads and positive-selection columns (Miltenyi Biotec,

Bergisch Gladbach, Germany). Gr1� cells were isolated using an anti-

Gr1-PE mAb (RB6-8C5; eBioscience, San Diego, CA, USA) and positive im-

munomagnetic separation using a selection kit (Stemcell Technologies,

Vancouver, BC, Canada) or sorting on a FACSAria II (Becton Dickinson,

San Diego, CA, USA). Flow cytometry verified that all of the isolated cells

yielded a �90% pure population.

mAb and flow cytometryFor flow cytometry method analysis of cell-surface markers, cells were

stained with antibodies in PBS containing 0.1% (wt/vol) BSA and 0.1%

NaN3. The following antibodies were obtained from eBioscience: anti-

CD11b (M1/70), anti-F4/80 (BM8), anti-Gr-1 (RB6-8C5), anti-CD4

(GK1.5), anti-CD11c (N418), anti-NK1.1 (Ly-55), anti-PD-L1 (B7-H1), anti-

CD86 (GL1), and anti-CCR7 (4B12). The following antibodies were ob-

tained from BD Biosciences (San Jose, CA, USA): anti-CD45 (TU116), anti-

CD8 (53-6.7), anti-CD115 (AFS98), and anti-CXCR2 (242216). The follow-

ing antibodies were obtained from Miltenyi Biotec: anti-CD3 (145-2C11)

and anti-CD19 (6D5). The following antibodies were obtained from

BioLegend (San Diego, CA, USA): anti-CD206 (068C2) and anti-CD124

(1015F8). Anti-CCR2 (475301) was obtained from R&D Systems (Minneap-

olis, MN, USA). Flow cytometry data were acquired on a FACSCalibur (Bec-

ton Dickinson) or a Beckman Coulter Epics XL benchtop flow cytometer

(Beckman Coulter, Brea, CA, USA), and data were analyzed with FlowJo

Software (TreeStar, Ashland, OR, USA). Cell numbers of various popula-

tions were calculated by multiplying the percentage of interested cells with

the total cell number.

Quantitative RT-PCRRNA was extracted with an RNeasy kit (Qiagen, Dusseldorf, Germany), and

cDNA was synthesized using SuperScript III RT (Invitrogen, Carlsbad, CA,

USA). An ABI 7900 real-time PCR system was used for quantitative PCR, with

primer and probe sets obtained from Applied Biosystems (Carlsbad, CA, USA;

Table 1). Results were analyzed using SDS 2.1 software. The expression of

each target gene is presented as the “fold change” relative to that of control

samples (2-�� comparative threshold), as described previously [2, 25].

In vitro functional assay of CD11b�Gr1� cellsThe suppressive function of MDSCs was assessed by determining their abil-

ity to inhibit T cell activation, as described previously [26, 27]. Purified T

cells were plated at a density of 2 � 105 cells/well in 1 �g/mL anti-CD3

antibody-coated plates in the presence of 1 �g/mL anti-CD28 antibody.

Isolated MDSCs were added to the wells at different ratios. Cell prolifera-

tion was subsequently determined 72 h later, after incubation with 3H-thy-

midine for the last 16 h of culture.

Cellular apoptosis assayApoptosis of MDSCs in mouse peripheral blood or liver cells was assayed by

flow cytometry using Annexin V staining with a gating strategy aligned to

analyzing CD11b�Gr1� cells, as described previously [28, 29].

Transwell assaysChemotaxis was evaluated according to the procedure described [28, 30]. A

24-well microchemotaxis plate (Corning, Corning, NY, US) was used, in which

TABLE 1. Primer Sequences Used for Real-Time PCR Assays

Genes Forward primer Reverse primer

Arg I 5=-CCAGAAGAATGGAAGAGTCAGTGT-3= 5=-GCAGATATGCAGGGAGTCACC-3=

CCL3 5=-TTCTCTGTACCATGACACTCTGC-3= 5=-CGTGGAATCTTCCGGCTGTAG-3=

CXCL1 5=-ACTGCACCCAAACCGAAGTC-3= 5=-TGGGGACACCTTTTAGCATCTT-3=

CXCL2 5=-CCAACCACCAGGCTACAGG-3= 5=-GCGTCACACTCAAGCTCTG-3=

iNOS 5=-CACCAAGCTGAACTTGAGCG-3= 5=-CGTGGCTTTGGGCTCCTC-3=

GAPDH 5=-GACTTCAACAGCAACTCCCAC-3= 5=-TCCACCACCCTGTTGCTGTA-3=

Arg, Arginase.

2 Journal of Leukocyte Biology Volume 95, June 2014 www.jleukbio.org

the chambers were separated by a polycarbonate membrane. DMEM, contain-

ing 0.01% BSA or CXCL1 (30 nM), was placed in the lower chamber. Purified

PBS or rapamycin-treated liver CD11b�Gr1� cell suspensions (5�105 cells in

500 �l) were placed in the upper chamber. Chambers were incubated at 37°C

with 5% CO2 for 2 h. The results are expressed as the mean number of

MDSCs/well and are representative of triplicate measurements. In some

assays, cells were pretreated with inhibitors for 30 min, as indicated.

NO production assayAfter incubating equal volumes of culture supernatant (100 �l) with a Greiss

reagent (Sigma-Aldrich), the absorbance at 550 nm was measured using a mi-

croplate reader (Bio-Rad, Hercules, CA, USA), as described before [26, 31].

Arginase assayArginase activity assay was performed as described previously [32, 33]. Briefly,

cells were lysed in 0.1% Triton X-100. Tris-HCl was added to 12.5 mM and

MnCl2 was added to 1 mM final concentration. Arginase was activated by heat-

ing for 10 min at 56°C, and L-arginine substrate was added to a 250-mM final

concentration. Reactions were incubated at 37°C for 30 min and stopped by

the addition of H2SO4/H3PO4. After the addition of a-isonitrosopropiophe-

none and heating for 30 min at 95°C, urea production was measured by absor-

bance at 540 nm and normalized to the total protein content.

Statistical analysesAll data are presented as the mean � sd. Student’s unpaired t-test was applied

for comparison of means and was used to compare differences among groups.

Comparison of the survival curves was performed using the Log Rank (Mantel-

Cox) test. P (� value) 0.05 was considered to be statistically significant.

RESULTS

Rapamycin protects against immunological hepatic injuryWe first investigated the effects of rapamycin on CIH mortal-

ity. As shown in Fig. 1A, the daily treatment with the indicated

doses of rapamycin significantly reduced mortality compared

with the control CIH group. A dose of 3 mg/kg rapamycin was

used in most of the following experiments as a result of its op-

timal level of protection against CIH in the mice. The kinetics

of liver damage were examined further in mice that received

15 mg/kg Con A (Fig. 1B–D). The serum ALT and AST activ-

ity levels (Fig. 1B) were significantly lower in the rapamycin-

treated group than the control group. Also, the rapamycin-

treated mice displayed significantly fewer hepatocytic necrosis

than did their control counterparts (Fig. 1C). The necrotic

area in the liver was smaller in the rapamycin group than the

control group (Fig. 1C).

Consistent with this, the effects of rapamycin on PIH were

studied. Rapamycin-treated mice displayed a significantly

smaller hepatocytic necrotic area than did their control coun-

terparts (Fig. 1D). The serum ALT and AST activity of the ef-

fector phase were both significantly lower in the rapamycin-

treated group than the control group (Fig. 1E).

mTOR inhibition with rapamycin treatment inducedmore CD11b�Gr1� cells during acute hepatic injuryWe next analyzed the liver inflammatory cell infiltration in

acute CIH with PBS or rapamycin treatment. Single-cell sus-

pensions from CIH liver tissue of rapamycin-treated and con-

trol mice were isolated and analyzed by flow cytometry. As

shown in Fig. 2A, there was a significantly lower number of

CD19� B cells, CD4� T cells, CD8� T cells, and NK1.1� cells

in the liver tissue at 20 h after the injection of Con A in the

rapamycin group compared with the control group. Con-

versely, a significantly higher number of myeloid CD11b�Gr1�

cells but not F4/80� macrophages, CD115� monocytes, or

CD11c� DC myeloid-derived cells were found in the rapamy-

0

20

40

60

80

100

0 20 40 60 80 100

PBS Rapa (1.5 mg/kg) Rapa (3 mg/kg)

Su

rviv

al (%

)

Time (h)

0

400

800

1200

1600

CIH; PBS

CIH; Rapa

0

20

40

60

80

100

CIH; PBS

CIH; Rap

ALT

(U

/ml)

AS

T (

U/m

l)

0

10

20

30

40

50

60

70

% o

f n

ecro

sis

are

a

ALT

(U

/ml)

AS

T (

U/m

l)

0

200

400

600

800

0

20

40

60

80

CIH; PBS

CIH; Rapa

400 X

400 X

A B C D

***

*** ******

******

0

20

40

60

80

% o

f n

ecro

sis

are

a

**

EFigure 1. mTOR inhibitor rapamycin protects against immunological hepatic injury. (A) Analysis of the effectof various doses of rapamycin treatment on survival in a mouse model of hepatic injury. B6 mice (n�10)were i.p.-injected with PBS or rapamycin (Rapa; 1.5 or 3 mg/kg body weight) daily, starting at 6 h before theinjection of a lethal dose of Con A (25 mg/kg). (B) Rapamycin (Rapa or Rap; 3 mg/kg) was administratedto B6 mice (n�5) daily, starting at 6 h before the low-dose Con A injection (CIH; 15 mg/kg), and the serumALT and AST levels were assessed 20 h after the injection. (C) In a separate group of the experiment in B,liver tissue was taken 40 h after the Con A injection (15 mg/kg) and analyzed by H&E staining. (Left) A rep-resentative picture is presented with the original scale bar of 50 �m; (right) the percentage of the necroticarea was quantitated by ImageJ software (version 2.1.4.7; U.S. National Institutes of Health, Bethesda, MD,USA). For each experiment, three random sections/liver were taken from three livers/group. (D) The effectof rapamycin treatment on liver histological injury at 20 h following PCl injection (10 �l 0.5% PCl in oliveoil) was assessed by histologic analysis (H&E staining). The percentage of necrotic area was quantitated us-

ing ImageJ software. For each experiment, three random sections/liver were taken from three livers/group. (E) In a similar experiment to D,the serum ALT and AST levels were assessed by ELISA at 20 h (the effector phase) after PCl injection (n�3). The data are representative ofthree (B and C) and two (D and E) independent experiments (mean�sd). **P 0.01, and ***P 0.001 for comparisons made between theindicated groups.

Zhang et al. mTOR inhibits MDSC migration

www.jleukbio.org Volume 95, June 2014 Journal of Leukocyte Biology 3

cin group compared with the control (Fig. 2A and B). Interest-

ingly, rapamycin treatment in myeloid cells significantly pro-

moted myeloid CD11b�Gr1� cell recruitment in a time-depen-

dent manner (Fig. 2C), whereas mTOR was down-regulated

significantly (Fig. 2D). These data suggest collectively that

CD11b�Gr1� cells are involved in rapamycin mTOR signaling

inhibition, thus affording protection against hepatic injury.

Rapamycin treatment promotes CD11b�Gr1� MDSCactivityWe then examined whether these CD11b�Gr1� cells that ac-

cumulated following rapamycin treatment function as immune

suppressors, as described for MDSCs [16, 17, 34]. Compared

with cells isolated from the control group, the CD11b�Gr1�

cells isolated from rapamycin-treated CIH liver (Fig. 3A) and

0

5

10

15

20

25

CIH; PBS

CIH; Rapa

B cells

Ce

lls (

x1

05/g

)

0

2

4

6

8

10

12

CIH; PBS

CIH; Rapa

CD4+T cells

0

1

2

3

CIH; PBS

CIH; Rapa

CD8+T cells

0

1

2

3

4

5

6

CIH; PBS

CIH; Rapa

NK1.1+ cells

0

5

10

15

20

25

30

CIH; PBS

CIH; Rapa

F4/80+ cells

0

1

2

CIH; PBS

CIH; Rapa

CD11c+ cells

CD

11

b+G

r1+

ce

lls (

x1

06/g

)

Time (h)

0

1

2

3

4

5

6

0 24 48 96

CIH; PBS

CIH; Rapa

0

10

20

30

40

50

60

CIH; PBS

CIH; Rapa

% o

f C

D11

b+G

r1+

ce

lls

A B0.5 25.5

4.6

CIH; PBS

CIH; Rapa

0.9 42

5.2

CD11b

C

******

** ***

***

Disotype

CIH; PBSCIH; Rapa

p-S6

Ce

ll#

*

***

**

Figure 2. Treatment with rapamycin promotes CD11b�Gr1� myeloid cell recruitmentin CIH mice. (A and B) B6 mice were injected with PBS or rapamycin (3 mg/kg) daily, start-ing from 6 h before the injection of Con A (15 mg/kg). At 20 h after the Con A injection, theabsolute cell number of CD19� B cells, CD4� T cells, CD8� T cells, NK1.1� cells, F4/80� mac-rophages, and CD11c� cells in the liver tissue from CIH mice (n�4–5) was determined byflow cytometry (A). (B) The CD11b�Gr1� cells in the liver were determined by flow cytometryand presented in a representative FACS plot (left) and the absolute numbers of CD11b�Gr1�

cells (n�4; right). (C) The number of CD11b�Gr1� cells in liver tissue at the indicated time-points in the CIH mice (n�4). (D) Liver MDSCs isolated from the rapamycin and PBS-treatedCIH mice. MDSCs were analyzed for the phosphorylation of p70S6 (p-S6); representative re-sults are shown. The data are representative of three (A, C, and D) and four (B) independentexperiments (mean�sd). *P 0.05, **P 0.01, and ***P 0.001 for the comparisons madebetween the indicated groups.

A

0.1 21.9

10.4

0.1 9.5

8.7

Gr1

Ly6

C

CIH; PBS

CIH; Rapa

C

B

% o

f p

ositiv

e c

ells

***

CIH; PBS CIH; RapaGate on CD11b+Gr1+ cells

CD206

PDL1

CD86

CD124

CCR2

CCR7

MF

I

**

***

**

0

50

100

150

200

250

CD

20

6

PD

L1

CD

12

4

CD

80

CD

86

CD

23

CD

54

CX

CR

4

CC

R2

CC

R7

CIH; PBS CIH; Rapa

0

10

20

30

40

50

60

70

80

1:1 1:2 1:4

CIH; PBS

CIH; Rapa

Su

pp

ressio

n (

%)

MDSC: T

**

*

0

5

10

15

20

25

Gr1+Ly6Chi cells

Gr1+Ly6Ghi cells

CIH; PBS CIH; Rapa

Figure 3. CD11b�Gr1� MDSC phenotype changes inCIH. (A–C) B6 mice were injected with PBS or rapa-mycin daily, starting at 6 h before the administration ofthe Con A (15 mg/kg) i.v. injection. At 40 h after ConA injection, the liver CD11b�Gr1� MDSC cells wereisolated. (A) Liver CD11b�Gr1� cell-suppressive activitywas analyzed in vitro, as described in Materials andMethods. T cells were stimulated with anti-CD3/CD28antibodies in the presence of liver MDSCs that wereisolated from Con A-injected PBS control or rapamy-cin-treated mice. T cell proliferation was determinedwith 3H-thymidine incorporation (n�3). (B) The ex-pression patterns of CD206, PD-L1, CD124, CD80,CD86, CD23, CD54, CXCR4, CCR2, and CCR7 on theliver CD11b�Gr1� cells were analyzed with flow cytom-etry. The results of the indicated cell populations arepresented as a representative FACS histogram (upper)and a bar graph (lower; n�3). MFI, Mean fluorescenceintensity. (C) Hepatic mononuclear cells from PBS-treated control or rapamycin-treated CIH mice(n�6–8) were stained with anti-Gr1, anti-Ly6C, andanti-Ly6G, followed by flow cytometry analysis, and thepercentages of Gr1�Ly6Chigh- and Gr1�Ly6Ghigh-stained cells are summarized. The data are representa-tive of three independent experiments (mean�sd).*P 0.05, **P 0.01, and ***P 0.001 for the com-parisons made between the indicated groups.

4 Journal of Leukocyte Biology Volume 95, June 2014 www.jleukbio.org

spleen (data not shown) displayed a significantly enhanced

immune-suppressive effect on CD4� T cell proliferation than

that seen in controls in vitro. Additionally, liver CD11b�Gr1�

cells, following rapamycin treatment, displayed a phenotype of

MDSC, CD206high, PD-L1high, and CD86low but with identical

levels of CD124, CD23, CD80, CD54, CXCR4, CCR2, and

CCR7 (Fig. 3B). Myeloid CD11b�Gr1� cells consist of two ma-

jor subsets: CD11b�Gr1� Ly6GhighLy6Clow granulocytic cells

and CD11b�Gr1�Ly6GlowLy6Chigh monocytic cells [24, 35].

We found that the percentage of CD11b�Gr1�Ly6Glow

Ly6Chigh myeloid cells and not the percentage of CD11b�

Gr1�Ly6GhighLy6Clow myeloid cells in the liver was signifi-

cantly higher in the rapamycin-treated CIH mice than the con-

trol mice (Fig. 3C). Moreover the phenotype alterations in the

MDSCs were only found for the CD11b�Gr1�Ly6ChighLy6Glow

myeloid cells, not the CD11b�Gr1�Ly6GhighLy6Clow myeloid

cells (Supplemental Fig. 1), indicating that the phenotype

alterations are associated with the different myeloid cell

subsets. Thus, these observations of CD11b�Gr1� cells, es-

pecially the CD11b�Gr1�Ly6Chigh cells, exhibited a suppres-

sive cell phenotype and mediated the protective effects

against hepatic injury.

Rapamycin treatment recruits CD11b�Gr1� MDSCduring hepatic injuryThe increased frequency of the infiltrating CD11b�Gr1� cells

might have occurred because of a decrease in cell death, an

increase of the differentiation of CD11b�Gr1� cells from pro-

genitors, or an enhanced recruitment into the liver. However,

rapamycin treatment significantly altered the cell death kinet-

ics of liver CD11b�Gr1� cells (Fig. 4A). In addition, both the

percentage (Fig. 4A) and the absolute cell number (data not

shown) of BrdU�CD11b�Gr1� cells were identical in the rapa-

mycin-treated and control groups, suggesting that there were

no defects in cell proliferation following rapamycin treatment.

However, the expression of CXCL1 and CXCL2, two chemo-

kines that are known to regulate CD11b�Gr1� cell migration,

was up-regulated significantly following rapamycin treatment

(Fig. 4B and C). In addition, the expression of CXCR2, the

receptor of the chemokines CXCL1 and CXCL2, was up-regu-

lated in the CD11b�Gr1� cells following rapamycin treatment

(Fig. 4D). Finally, anti-CXCR2 antibody injection blocked the

CD11b�Gr1� MDSC infiltration and consequently, resulted in

severe mortality, pathological liver damage, and increased ALT

and AST levels (Fig. 4E and Supplemental Fig. 2A and B).

In accord with this, a significantly higher frequency and

number of myeloid CD11b�Gr1� cells were observed in PIH

than control mice (Supplemental Fig. 3A). In addition, the

expression of CXCR2 in the rapamycin-treated CD11b�Gr1�

MDSCs was significantly higher than the control (Supplemen-

tal Fig. 3B). Finally, the blocking of CD11b�Gr1� MDSC infil-

tration with an anti-CXCR2 antibody injection resulted in an

increased ALT level (Supplemental Fig. 3C). Collectively, these

observations show that rapamycin promoted the recruitment

of CD11b�Gr1� MDSCs in mice with inflammatory hepatic

injury.

Rapamycin treatment promotes myeloidCD11b�Gr1�Ly6Chigh MDSC migration duringhepatic injuryNext, we investigated the MDSC migration capacity in vitro.

Rapamycin-treated total CD11b�Gr1� cells, especially

CD11b�Gr1�Ly6ChighLy6Glow cells and not CD11b�Gr1�

Ly6GhighLy6Clow cells, showed enhanced migration capacity

***

0

50

100

150

200

250

300

CIH; PBS

CIH; Rapa

CX

CR

2 e

xp

ressio

n (

MF

I)

0

200

400

600

800

1000

1200

1400

1600

CIH; PBS

CIH; Rapa

0

20

40

60

80

100

CIH; PBS

CIH; Rapa

CIH + IgG CIH + anti-CXCR2 mAb

ALT

(U

/ml)

AS

T (

U/m

l)

Se

rum

(n

g/m

l)

0

1

2

3

4

5

CXCL1 CXCL2

CIH; PBS

CIH; Rapa

mR

NA

(re

lative

exp

ressio

n)

B C D

E

0

2

4

6

8

10

12

14

16

18

WBC Liver

CIH; PBS CIH; Rapa

% o

f A

nn

exin

V+C

D11

b+G

r1+

ce

lls

0

1

2

3

4

5

6

7

8

9

10

WBC Liver

% o

f B

rdU

+C

D11

b+G

r1+

ce

lls

A

*

****

***

*

*

****

0

20

40

60

80

100

CXCL1 CXCL2 CCL3

CIH; PBS CIH; Rapa

Figure 4. Treatment with rapamycin increases the liver expression of chemokines that mediateCD11b�Gr1� myeloid cell recruitment. (A–D) B6 mice were injected with PBS or rapamycin (3mg/kg) daily, starting from 6 h before the injection of Con A (15 mg/kg). (A) Cell death (left) orcell proliferation (right) of liver CD11b�Gr1� cells isolated from control and rapamycin-treatedCIH mice 20 h after Con A injection (15 mg/kg) were determined by Annexin-V staining or BrdUincorporation, respectively. BrdU was i.v.-injected 28 h before the Con A injection (n�4). WBC,White blood cell. The concentration of CXCL1, CXCL2, and CCL3 in the serum (B) or the mRNAlevel of CXCL1 and CXCL2 in the liver CD11b�Gr1� cells (C) was determined by ELISA (n�3) orquantitative PCR (n�4). (D) The CXCR2 expression in liver CD11b�Gr1� cells was determined byintracellular staining followed by flow cytometry analysis (n�3). (E) The indicated CIH mice wereadministered 50 �g anti-CXCR2 mAb (242216; R&D Systems) or IgG isotype control (54447; R&DSystems) via i.v. injection. At 20 h after Con A injection, the serum ALT and AST levels were as-

sessed by ELISA (n�4–6). The data are representative of two independent experiments (mean�sd). *P 0.05, **P 0.01, and ***P 0.001 forthe comparisons made between the indicated groups.

Zhang et al. mTOR inhibits MDSC migration

www.jleukbio.org Volume 95, June 2014 Journal of Leukocyte Biology 5

compared with PBS-treated control in an in vitro transwell cell

migration assay (Fig. 5A and B). Consistent with these results,

a well-established in vivo model was used to study MDSC mi-

gration, to determine the altered migration of rapamycin-

treated MDSCs. A significantly higher number of recruited

CD11b�Gr1� cells (Fig. 5C), especially CD11b�Gr1�

Ly6ChighLy6Glow cells, not CD11b�Gr1�Ly6GhighLy6Clowcells

(Fig. 5D), with a higher level of CXCR2 expression (data not

shown) were observed in the inflammatory liver tissue of rapa-

mycin-treated CIH mice compared with the control group.

These results suggest that mTOR inhibition by rapamycin

treatment promotes CD11b�Gr1�Ly6Chigh MDSC migration in

inflammation.

Rapamycin induces CD11b�Gr1� MDSC recruitmentvia NO productionNO production has been suggested as a critical component

mediating the immunosuppressive activity of MDSC [18, 24,

36]. We thus measured the NO levels following Con A injec-

tion. In agreement, the average NO level but not the arginase

level was markedly higher in the rapamycin-treated CIH com-

pared with control mice (Fig. 6A and B). In addition, the ex-

pression level of iNOS, a NO-producing metabolic enzyme

[37], was induced significantly (Fig. 6C), whereas the expres-

sion level of arginase, a metabolic enzyme sharing the same

substrate of iNOS, was identical in CD11b�Gr1� MDSCs fol-

lowing rapamycin treatment in the CIH mice (data not

shown). To determine whether NO production is an essential

component in MDSC recruitment-mediated protection against

hepatic injury, we applied L-NMMA, an inhibitor of iNOS, to

the in vivo functional assay. L-NMMA significantly reduced NO

production (data not shown) and effectively reduced the re-

cruitment of MDSC (Fig. 6D), the ALT level (Fig. 6E), and the

recruitment of Ly6Chigh MDSCs (Fig. 6F) in rapamycin-treated

CIH mice compared with the control group. However, the

CXCR2 expression in the MDSCs was not affected by the L-

NMMA treatment in the CIH mice (Supplemental Fig. 2C).

Taken together, these data suggest that NO is required for

rapamycin-induced MDSC recruitment in protecting against

immunological hepatic injury.

mTOR signaling negatively controls the recruitmentof CD11b�Gr1� Ly6Chigh MDSC in immunologicalhepatic injuriesTo distinguish whether it is a result of MDSC or whether other

cell (such as hepatocytes) mTOR signaling mediated the re-

cruitment of CD11b�Gr1�Ly6Chigh MDSCs in immunological

hepatic injuries, we performed cell-adoptive transfer experi-

ments in mice. We treated CD45.2� B6 recipient mice with

PBS or rapamycin (3 mg/kg) for 3 days. Such treatment re-

sulted in efficient deletion of mTOR in CD11b�Gr1� cells and

other cells (such as hepatocytes) in recipient mice (data not

shown). Sorted donor liver CD11b�Gr1� cells from CIH

CD45.1� B6 mice were transferred into PBS- or rapamycin-

treated recipient mice, respectively. Then, recipient mice were

injected with Con A (15 mg/kg) for CIH induction at the in-

***

0

1

2

3

4

5

6

7

8*

***

N.S.

Tra

nsm

itte

d c

ell

ratio

(re

lative

)

0

0.2

0.4

0.6

0.8

1None

CXCL1

Rap

CXCL1+Rap

Tra

nsm

itte

d c

ells

(x1

05)

0

1

2

3

4

5

6

7

PBS MDSC; PBS

MDSC; Rapa

Re

cru

ite

d d

on

or

ce

lls (

x1

04)

0

1

2

3

4

5

6

Ly6Chi Ly6Ghi

PBS

Rapa

Re

cru

ite

d d

on

or

ce

lls (

x1

04)

A B C

D

48.2

50.2

0.871.75

CD45.1

CF

SE

Before transfer

6 h after transfer

***

Figure 5. Treatment with rapamycin promotes the migration of CD11b�Gr1�Ly6Chigh MDSCs in vitro and invivo. (A and B) Liver MDSC in vitro transwell migration was determined. Liver CD11b�Gr1� MDSCs (A) andliver Gr1�Ly6GhighLy6Clow or Gr1�Ly6chighLy6Glow MDSCs (B) isolated from CIH mice, respectively, and pre-treated by PBS or rapamycin (100 uM) for 4 h before in vitro MDSC transwell migration analysis was per-formed (n�4). (C and D) Liver MDSC in vivo migration was determined. Sorted PBS-treated liver MDSCs(CD11b�Gr1�CD45.2�) were labeled with CFSE and mixed with rapamycin-treated liver MDSCs (CD11b�

Gr1�CD45.1�) at a ratio of 1:1. These cells were i.v.-injected to syngeneic host CD45.2� mice for in vivo migra-tion assays. (C) More rapamycin-treated CD45.1�CD11b�Gr1� MDSCs were detected in the liver tissue at 20 hafter Con A (15 mg/kg) injection than PBS-treated control cells in the same host. A representative FACS plot(left) and the recruited donor MDSC absolute cell number (right; n�4). (D) More rapamycin-treatedCD45.1�Gr1�Ly6ChighLy6Glow MDSCs were detected in the liver tissue at 20 h after Con A (15 mg/kg) injec-tion than PBS-treated control cells in the same host. The recruited donor MDSC absolute cell number was cal-culated (n�4). The data are representative of three (A and B) and two (C and D) independent experiments(mean�sd). *P 0.05, and ***P 0.001 for the comparisons made between the indicated groups; N.S., notsignificant.

6 Journal of Leukocyte Biology Volume 95, June 2014 www.jleukbio.org

dicated points compared with PBS-treated recipient control,

and the rapamycin-treated recipients exhibited similarly re-

cruited donor-derived CD11b�Gr1� MDSC cells in inflamma-

tory liver tissue (Fig. 7A). Consistent with this, we treated WT

and mTORCreER mice with tamoxifen for 4 days. Such treat-

ment resulted in efficient deletion of mTOR in CD11b�Gr1�

cells and other cells (such as hepatocytes) from mTORCreER

mice (Supplemental Fig. 4). Sorted donor liver CD11b�Gr1�

cells from CIH WT CD45.1� mice were transferred into WT or

mTORCreER recipient mice, respectively. Then, recipient mice

were injected with Con A for CIH induction, and at 6 h after

Con A injection, compared with WT recipient control, the

mTORCreER recipient showed similar recruited donor-derived

CD11b�Gr1� MDSC cells in inflammatory liver tissue (Fig.

7B). Thus, these data show that recipient tissues (such as

hepatocytes) that are mTOR-deficient had no alteration in the

donor MDSC recruitment. However, when the sorted donor

CD11b�Gr1� cells from PBS- and rapamycin-treated CIH mice

were transferred into the identical CD45.1� B6 recipient mice,

the rapamycin-treated MDSC transfer groups contained more

donor-derived MDSC compared with the PBS-treated MDSC

transfer control groups in recipient CIH mouse inflammatory

liver tissue (Fig. 7C). In agreement, when we transferred the

sorted donor CD11b�Gr1� cells from WT and mTORCreER

CIH mice into the identical CD45.1� B6 recipient mice, the

mTORCreER MDSC transfer groups contained more donor-de-

rived MDSC recruitment compared with WT control transfer

groups in recipient mouse inflammatory liver tissue (Fig. 7D).

These data suggest that mTOR deficiency in MDSCs is impor-

tant for their recruitment in immunological liver-injury dis-

eases. Furthermore, donor rapamycin-treated mouse

CD11b�Gr1�Ly6ChighLy6Glow cell-adoptive transfer groups but

not CD11b�Gr1�Ly6GhighLy6Clow cell transfer groups exhib-

ited greater donor cell recruitment in recipient CIH mouse

inflammatory liver tissue compared with PBS-treated control

donor cells (Fig. 7E). Thus, the results suggest collectively that

mTOR signaling intrinsically controls the recruitment of

CD11b�Gr1�Ly6Chigh MDSCs in protecting against immuno-

logical hepatic injury.

DISCUSSION

Although emerging evidence indicates that mTOR signaling

negatively regulates immune-mediated liver injury [38, 39], the

underlying mechanism remains unclear. Here, we show that

MDSCs are essential for rapamycin protection against CIH and

PIH, two typical immunological heptic injury models. Mecha-

nistic insight was provided by showing the modulatory effects

of mTOR signaling in MDSCs. With the targeting of the

mTOR signaling pathway, rapamycin potentiates MDSC re-

cruitment for protection against hepatic injuries. This is con-

sistent with the previously established role of MDSCs in limit-

ing immunological hepatic inflammatory injury [40]. These

findings demonstrate a previously unknown feature of MDSC

function in liver homeostasis, i.e., the recruitment of MDSCs,

0

1

2

3

4

5

6

CIH

; P

BS

CIH

; L

-NM

MA

CIH

; R

ap

a

CIH

; R

ap

a; L

-NM

MA

***

0

200

400

600

800

1000

CIH

; P

BS

CIH

; L

-NM

MA

CIH

; R

ap

a

CIH

; R

ap

a; L

-NM

MA

0

1

2

3

4

5

6

7

CIH

; R

ap

a

CIH

; R

ap

a; L

- NM

MA

0

10

20

30

40

CIH; PBS

CIH; Rapa

NO

(u

Mo

l/L

)

CD

11

b+G

r1+

ce

lls (

x1

06/g

)

ALT

(U

/ml)

A D E***

***

***

**

Re

cru

ite

d d

on

or

ce

lls (

x1

04)F

***

0

10

20

30

40

50

60

70

80

90

CIH; PBS

CIH; Rapa

iNO

S/C

D11

b+G

r1+

ce

lls (

%) **C

0

10

20

30

40

50

60

CIH; PBS

CIH; Rapa

Arg

ina

se

(x1

02

mU

)

B

Figure 6. Rapamycin treatment promotes CD11b�Gr1�Ly6Chigh MDSC recruitment through NO production.(A–C) The CIH in the indicated groups was induced by Con A injection (15 mg/kg). All of the mice werekilled 20 h later, and liver MDSCs were isolated from the PBS or rapamycin-treated CIH mice, followed by stim-ulation with LPS (10 ng/mL) or IL-4 (1000 U/mL). The NO (A) and arginase (B) levels were determined(n�3). (C) The percentage of iNOS in the liver CD11b�Gr1� cells isolated from the indicated CIH mice(n�3). (D and E) B6 mice were treated by PBS, rapamycin (3 mg/kg), or PBS or rapamycin in combinationwith L-NMMA (M7033; 80 mg/kg, gavage daily; Sigma-Aldrich), starting at 6 h before the low-dose Con A (15mg/kg) injection. All of the mice were killed 20 h after the Con A injection, and then, the absolute cell num-ber of the CD11b�Gr1� cells in the liver tissue (D) and serum ALT level (E; n�4) was determined. (F) Sortedrapamycin and Rapa � L-NMMA (80 mg/kg, gavage daily)-treated liver MDSCs (CD11b�Gr1�CD45.1� cells)were labeled with CFSE and PKH (named for their discoverer Paul Karl Horan), respectively, and then i.v. in-jected into syngeneic host CD45.2� mice for in vivo migration assays. The blocking of NO production with L-

NMMA significantly diminished the rapamycin treatment-recruited CD45.1�CD11b�Gr1�Ly6Chigh MDSC cell number (n�3). The data arerepresentative of three (A, B, D, and E) and two (C and F) independent experiments (mean�sd). **P 0.01, and ***P 0.001 for thecomparisons made between the indicated groups.

Zhang et al. mTOR inhibits MDSC migration

www.jleukbio.org Volume 95, June 2014 Journal of Leukocyte Biology 7

which represents a novel mechanism of rapamycin-mediated

protection against immunological hepatic injuries by targeting

the mTOR signaling pathway.

With the targeting of mTOR signaling, rapamycin has been

considered a potent agent with anti-inflammatory properties.

However, studies on the involvement of mTOR in the regula-

tion of innate inflammation have been contradictory in recent

reports [41–43]. Rapamycin exhibited an anti-inflammatory

effect on in vitro-generated DCs, whereas it enhanced IL-23

and IL-12 proinflammatory cytokine production in in vitro-

generated macrophages via enhancing NF-�B and blocking

STAT3 activity [41, 44]. The activation of monocytes and DCs

with LPS in the presence of rapamycin leads to an increase in

IL-12 production and a decrease in IL-10 production. These

findings suggest that mTOR inhibits proinflammatory gene

expression in these cells. Alternatively, rapamycin may block

the maturation of bone marrow-derived DCs [13, 45, 46]. Such

cells display decreased MHC and costimulatory molecules and

actually promote the induction of anergic and Tregs. Interest-

ingly, in spite of the ability of the rapamycin-matured DCs to

promote tolerance, they also produce increased IL-12 upon

stimulation with LPS. These findings indicate, as is the case for

different innate immune cell subsets, that the outcome of

mTOR activation (and inhibition) is complex and cell type-

dependent. However, the precise effects of mTOR signaling

from MDSCs remain unclear. In the present study, we show

that mTOR deficiency, induced by rapamycin treatment, po-

tentiates the recruitment of hepatic MDSCs in a murine immu-

nological hepatitis model. Furthermore, the recruitment of

MDSCs is critically required for this protection against immu-

nological hepatic injury.

In mice, Con A administration or chronic hepatitis B virus

infection results in a profound liver infiltration of immunosup-

pressive MDSCs [40, 47]. This may limit inflammation follow-

ing Con A treatment or hepatitis virus challenge. In humans,

relatively recent work [48] has shown that peripheral blood

monocytes from patients with autoimmune immune hepatitis

are increased. Compared with circulating monocytes from

healthy controls, the number of circulating monocytes from

patients is much higher with frequent correlation with in-

creased ALT level in the blood. However, this study did not

examine whether blood (or liver) monocytes from patients

with immunological hepatic diseases are capable of inhibiting

T cell proliferation in vitro. More importantly, whether and

how myeloid cells are involved in inflammatory and/or immu-

nological processes in the liver remain unanswered. In the

present study, we showed that treatment with rapamycin ap-

peared to alter the infiltration of MDSCs. Additionally, inhibi-

0

2

4

6

8

10

6 40

PBS

Rapa

Re

cru

ite

d d

on

or

ce

lls (

x1

04/g

)

RecipientA

Time (h)

Re

cru

ite

d d

on

or

ce

lls (

x1

04/g

)

RecipientB

0

2

4

WT

mTORCreER

0

2

4

6

8

10

12

14

16

6 40

PBS

Rapa

Donor

Re

cru

ite

d d

on

or

ce

lls (

x1

04/g

)

C

Time (h)

***

***

0

2

4

6

8

10

12

14

PB

S;L

y6

C

Ra

pa

;Ly6

C

PB

S;L

y6

G

Ra

pa

;Ly6

G

Re

cru

ite

d d

on

or

ce

lls (

x1

04/g

)

E**

N.S.

0

2

4

6

WT

mTORCreER

Donor

Re

cru

ite

d d

on

or

ce

lls (

x1

04/g

)

**

D

Figure 7. mTOR signaling from CD11b�Gr1� Ly6Chigh MDSCs controls cell recruitment in immunological hepatic injuries. (A) Donor CD45.1�

B6 mouse CIH was induced by Con A (15 mg/kg) i.v. injection, and at 40 h after Con A injection, 1 � 106 CD11b�Gr1� MDSCs were isolatedfrom the liver tissue of the CIH mice and transferred into PBS-treated or rapamycin (3 mg/kg)-treated CD45.2� B6 recipient mice via i.v. injec-tion. At 6 h after cell transfer, all of the groups were injected with Con A (15 mg/kg), and then at the indicated time-points, the liver donor-de-rived CD11b�Gr1� cells were analyzed with flow cytometry, and the absolute cell number was calculated. (B) WT or mTORCreER mice were treatedwith tamoxifen for 4 days. CIH was induced by Con A i.v. injection in donor WT CD45.1� B6 mice. At 40 h after Con A injection, 1 � 106

CD11b�Gr1� MDSCs were isolated from the liver tissue of donor CIH mice and transferred into WT or mTORCreER recipient mice via i.v. injec-tion. After cell transfer, all of the groups were injected with Con A. At 6 h, the liver donor-derived CD11b�Gr1� cells were analyzed with flow cy-tometry, and the absolute cell number was calculated. (C) Similarly, CIH was induced by Con A in PBS-treated or rapamycin-treated donorCD45.2� B6 mice. At 40 h after Con A injection, 1 � 106 CD11b�Gr1� MDSCs were isolated from the liver tissue of the CIH mice and transferredinto WT CD45.1� B6 recipient mice via i.v. injection. At 6 h after cell transfer, recipient mice were injected with Con A, and at the indicated time-points, the liver donor-derived CD11b�Gr1� cells were analyzed, and absolute cell number was calculated. (D) CIH was induced by Con A in WTor mTORCreER mice. At 40 h after Con A injection, CD11b�Gr1� MDSCs were isolated from the liver of the donor CIH mice and transferred intoWT CD45.1� B6 recipient mice. After cell transfer, recipient mice were injected with Con A, and at 6 h, the liver donor-derived CD11b�Gr1� cellswere analyzed with flow cytometry, and the absolute cell number was calculated. (E) Furthermore, 1 � 106 CD11b�Gr1�Ly6Chigh MDSCs andCD11b�Gr1�Ly6Ghigh MDSCs were isolated from the liver tissue of PBS- or rapamycin-treated donor CIH mice and transferred into CD45.1� B6recipient mice, respectively. After cell transfer, recipient mice were injected with Con A. At 40 h, the liver donor-derived MDSCs were analyzed,and absolute cell number was calculated. The data are shown as mean � sd, n � 4–6, from one of two (A–D) and three (E) independent experi-ments. **P 0.01, and ***P 0.001 compared with the indicated groups.

8 Journal of Leukocyte Biology Volume 95, June 2014 www.jleukbio.org

tions of mTOR signaling with rapamycin treatment up-regulate

the expression of the chemokines CXCL1 and CXCL2 that

mediate CD11b�Gr1� MDSC recruitment during hepatic in-

jury. Blocking the recruitment of MDSCs by the administration

of an anti-CXCR2 mAb accelerated immunological hepatic in-

jury. Thus, our study suggests that the recruitment of rapamy-

cin-induced CD11b�Gr1� MDSCs is at least partially responsi-

ble for the amelioration of immunological hepatic injury.

The chemoattractant factors released by resident cells, in-

cluding the CXC chemokines and inflammatory mediators

[49, 50], recruit MDSC to the site of challenge during the ini-

tiation of infection and in the course of the inflammatory re-

sponse. In our acute hepatic injury model, serum and liver

MDSCs from rapamycin-treated CIH mice expressed higher

levels of CXCL1 and CXCL2 compared with control mice.

Compared with controls, a significantly higher level of CXCR2

(the receptor of CXCL1 and CXCL2) expression was observed

in the rapamycin-treated CD11b�Gr1� MDSCs. Importantly,

the blockage of CXCR2 with an anti-CXCR2 mAb significantly

reversed the pathological injury and MDSC recruitment in the

rapamycin-treated mice compared with control. It is believed

that inflammatory stimuli trigger CXCL1/2 expression in he-

patic tissues, which in turn, serves as a potent chemoattractant

for CD11b�Gr1� MDSC migration into the liver by recogniz-

ing CXCR2, resulting in a protective effect against immunolog-

ical hepatic injury. Furthermore, rapamycin treatment signifi-

cantly promoted the apoptosis of liver CD11b�Gr1� MDSCs in

CIH mice but did not significantly alter proliferation. This fur-

ther provides evidence that MDSC recruitment is an important

reason for the increased number of CD11b�Gr1� cells in the

inflammatory liver tissue. This also suggests that there proba-

bly exists a self-limiting type of negative-feedback modulation

of MDSCs in the local immunological inflammation that oc-

curs in IMH. Taken together, the data suggest that the CXC1/

2–CXCR2-dependent recruitment of CD11b�Gr1� MDSCs is

one major reason for the MDSC increase at inflammatory sites

in rapamycin-treated CIH mice.

MDSCs are a heterogenic population, comprised of mono-

cytic MDSCs expressing CD11b�Gr1�Ly6Chigh and granulo-

cytic MDSCs with a CD11b�Gr1� Ly6Clow phenotype [40].

Our results demonstrated that CD11b�Gr1�Ly6Chigh MDSCs

comprise the most recruited population in rapamycin-treated

CIH mice. Moreover, rapamycin was able to promote the mi-

gration of CD11b�Gr1�Ly6ChighLy6Glow MDSCs efficiently but

not CD11b�Gr1�Ly6GhighLy6Clow MDSCs. In addition, it is

suggested that the NO pathway is responsible for this effect.

Rapamycin-treated CD11b�Gr1� MDSCs produce a profound

quantity of NO. The blocking of NO production by the admin-

istration of L-NMMA, a inhibitor of iNOS, effectively pre-

cluded the recruitment of CD11b�Gr1� MDSCs, especially

CD11b�Gr1�Ly6Chigh cells, and accelerated the development

of immunological hepatic injuries, indicating that NO is one

of the major pathways by which mTOR inhibition in

CD11b�Gr1� MDSCs promotes cell recruitments in CIH mice.

To summarize, mTOR signaling is here, for the first time,

shown to be an important pathway that negatively regulates

the migration and recruitment of immunosuppressive CD11b�

Gr1�Ly6Chigh MDSCs. Additional insight into the control of

MDSCs by rapamycin treatment will not only illuminate the

fundamental molecular principles of MDSC regulation but

also will facilitate the clinical application of modulators of the

MDSCs in treating immune disorders.

AUTHORSHIP

Y.Z. and Y.B. designed and conducted the experiments with

cells and mice and analyzed data. H.Y. conducted the experi-

ments with cells. Y.B. and H.Y. contributed to writing the man-

uscript. X.C. and H.L. analyzed the histological data. Y.L., Z.Z.,

J.L., S.Y., Y.C., and R.Y. contributed to the discussion of the

manuscript. G.L. developed the concept, designed and con-

ducted experiments with cells and mice, analyzed data, wrote

the manuscript, and provided overall direction.

ACKNOWLEDGMENTS

The authors’ research is supported by grants from the Na-

tional Natural Science Foundation for General Programs of

China (31171407 and 81273201), Key Basic Research Project

of the Science and Technology Commission of Shanghai Mu-

nicipality (12JC1400900), Innovation Program of Shanghai

Municipal Education Commission (14ZZ009), and Excellent

Youth Foundation of Chinese Academy of Sciences (KSCX2-

EW- Q-7) to G.L.

DISCLOSURES

The authors declare no competing financial interests.

REFERENCES

1. Chi, H. (2012) Regulation and function of mTOR signalling in T cellfate decisions. Nat. Rev. Immunol. 12, 325–338.

2. Liu, G., Burns, S., Huang, G., Boyd, K., Proia, R. L., Flavell, R. A., Chi,H. (2009) The receptor S1P1 overrides regulatory T cell-mediated im-mune suppression through Akt-mTOR. Nat. Immunol. 10, 769–777.

3. Bi, Y., Liu, G., Yang, R. (2011) mTOR regulates T-cell differentiationand activation in immunity and autoimmunity. Crit. Rev. Eukaryot. GeneExpr. 21, 313–322.

4. Delgoffe, G. M., Kole, T. P., Zheng, Y., Zarek, P. E., Matthews, K. L.,Xiao, B., Worley, P. F., Kozma, S. C., Powell, J. D. (2009) The mTORkinase differentially regulates effector and regulatory T cell lineage com-mitment. Immunity 30, 832–844.

5. Zoncu, R., Efeyan, A., Sabatini, D. M. (2011) mTOR: from growth signalintegration to cancer, diabetes and ageing. Nat. Rev. Mol. Cell. Biol. 12,21–35.

6. Baker, A. K., Wang, R., Mackman, N., Luyendyk, J. P. (2009) Rapamycinenhances LPS induction of tissue factor and tumor necrosis factor-� ex-pression in macrophages by reducing IL-10 expression. Mol. Immunol.46, 2249–2255.

7. Thomson, A. W., Turnquist, H. R., Raimondi, G. (2009) Immunoregula-tory functions of mTOR inhibition. Nat. Rev. Immunol. 9, 324–337.

8. Wullschleger, S., Loewith, R., Hall, M. N. (2006) TOR signaling ingrowth and metabolism. Cell 124, 471–484.

9. Dumont, F. J., Staruch, M. J., Koprak, S. L., Melino, M. R., Sigal, N. H. (1990)Distinct mechanisms of suppression of murine T cell activation by the relatedmacrolides FK-506 and rapamycin. J. Immunol. 144, 251–258.

10. Yang, H., Wang, X., Zhang, Y., Liu, H., Liao, J., Shao, K., Chu, Y., Liu,G. (2014) Modulation of TSC-mTOR signaling on immune cells in im-munity and autoimmunity. J. Cell. Physiol. 229, 17–26.

11. Chen, H., Zhang, L., Zhang, H., Xiao, Y., Shao, L., Li, H., Yin, H.,Wang, R., Liu, G., Corley, D., Yang, Z., Zhao, Y. (2013) Disruption ofTSC1/2 signaling complex reveals a checkpoint governing thymicCD4�CD25�Foxp3� regulatory T-cell development in mice. FASEB J.27, 3979–3990.

12. Sehgal, S. N. (2003) Sirolimus: its discovery, biological properties, andmechanism of action. Transplant. Proc. 35, 7S–14S.

Zhang et al. mTOR inhibits MDSC migration

www.jleukbio.org Volume 95, June 2014 Journal of Leukocyte Biology 9

13. Delgoffe, G. M., Powell, J. D. (2009) mTOR: taking cues from the im-mune microenvironment. Immunology 127, 459–465.

14. Havla, J., Kumpfel, T., Hohlfeld, R. (2012) [2012: update on diagnosis andtreatment of multiple sclerosis]. Dtsch. Med. Wochenschr. 137, 894–899.

15. Ray, P., Arora, M., Poe, S. L., Ray, A. (2012) Lung myeloid-derived sup-pressor cells and regulation of inflammation. Immunol. Res. 50, 153–158.

16. Bunt, S. K., Yang, L., Sinha, P., Clements, V. K., Leips, J., Ostrand-Rosenberg, S. (2007) Reduced inflammation in the tumor microenvi-ronment delays the accumulation of myeloid-derived suppressor cellsand limits tumor progression. Cancer Res. 67, 10019–10026.

17. Ostrand-Rosenberg, S., Sinha, P. (2009) Myeloid-derived suppressorcells: linking inflammation and cancer. J. Immunol. 182, 4499–4506.

18. Jayaraman, P., Parikh, F., Lopez-Rivera, E., Hailemichael, Y., Clark, A.,Ma, G., Cannan, D., Ramacher, M., Kato, M., Overwijk, W. W., Chen,S. H., Umansky, V. Y., Sikora, A. G. (2010) Tumor-expressed induciblenitric oxide synthase controls induction of functional myeloid-derivedsuppressor cells through modulation of vascular endothelial growth fac-tor release. J. Immunol. 188, 5365–5376.

19. Van Rompaey, N., Le Moine, A. (2010) Myeloid-derived suppressor cells:characterization and expansion in models of endotoxemia and trans-plantation. Methods Mol. Biol. 677, 169–180.

20. Liu, G., Bi, Y., Shen, B., Yang, H., Zhang, Y., Wang, X., Liu, H., Lu, Y.,Liao, J., Chen, X., Chu, Y. (2014) SIRT1 limits the function and fate ofmyeloid-derived suppressor cells in tumors by orchestrating HIF1�-de-pendent glycolysis. Cancer Res. 74, 727–737.

21. Ma, J., Pan, P. Y., Eisenstein, S., Divino, C. M., Lowell, C. A., Takai, T.,Shu-Hsia, C. (2011) Paired immunoglobin-like receptor B regulates thesuppressive function and fate of myeloid-derived suppressor cells. Immu-nity 25, 385–395.

22. Jiang, J., Zhou, C., Xu, Q. (2003) Alleviating effects of si-ni-san, a tradi-tional Chinese prescription, on experimental liver injury and its mecha-nisms. Biol. Pharm. Bull. 26, 1089–1094.

23. Schumann, J., Prockl, J., Kiemer, A. K., Vollmar, A. M., Bang, R., Tiegs,G. (2003) Silibinin protects mice from T cell-dependent liver injury. J.Hepatol. 39, 333–340.

24. Cripps, J. G., Wang, J., Maria, A., Blumenthal, I., Gorham, J. D. (2010) Type 1T helper cells induce the accumulation of myeloid-derived suppressor cells inthe inflamed Tgfb1 knockout mouse liver. Hepatology 52, 1350–1359.

25. Liu, G., Hu, X., Sun, B., Yang, T., Shi, J., Zhang, L., Zhao, Y. (2013)Phosphatase Wip1 negatively regulates neutrophil development throughp38 MAPK-STAT1. Blood 121, 519–529.

26. Wu, T., Sun, C., Chen, Z., Zhen, Y., Peng, J., Qi, Z., Yang, X., Zhao, Y.(2012) Smad3-deficient CD11b(�)Gr1(�) myeloid-derived suppressorcells prevent allograft rejection via the nitric oxide pathway. J. Immunol.189, 4989–5000.

27. Corzo, C. A., Condamine, T., Lu, L., Cotter, M. J., Youn, J. I., Cheng, P.,Cho, H. I., Celis, E., Quiceno, D. G., Padhya, T., McCaffrey, T. V., Mc-Caffrey, J. C., Gabrilovich, D. I. (2012) HIF-1� regulates function anddifferentiation of myeloid-derived suppressor cells in the tumor mi-croenvironment. J. Exp. Med. 207, 2439–2453.

28. Liu, G., Bi, Y., Wang, R., Shen, B., Zhang, Y., Yang, H., Wang, X., Liu,H., Lu, Y., Han, F. (2013) Kinase AKT1 negatively controls neutrophilrecruitment and function in mice. J. Immunol. 191, 2680–2690.

29. Bi, Y., Zhou, J., Yang, H., Wang, X., Zhang, X., Wang, Q., Wu, X., Han, Y.,Song, Y., Tan, Y., Du, Z., Yang, H., Zhou, D., Cui, Y., Zhou, L., Yan, Y., Zhang,P., Guo, Z., Wang, X., Liu, G., Yang, R. (2014) IL-17A produced by neutrophilsprotects against pneumonic plague through orchestrating IFN-�-activated mac-rophage programming. J. Immunol. 192, 704–713.

30. Blomgran, R., Ernst, J. D. (2013) Lung neutrophils facilitate activationof naive antigen-specific CD4� T cells during Mycobacterium tuberculosisinfection. J. Immunol. 186, 7110–7119.

31. Sun, B., Hu, X., Liu, G., Ma, B., Xu, Y., Yang, T., Shi, J., Yang, F., Li, H.,Zhang, L., Zhao, Y. (2014) Phosphatase Wip1 negatively regulates neu-trophil migration and inflammation. J. Immunol. 192, 1184–1195.

32. Liu, L., Cuthbertson, F. (2012) Early bilateral cystoid macular oedemasecondary to fingolimod in multiple sclerosis. Case Report Med. 12,134636–134643.

33. Bunt, S. K., Yang, L., Sinha, P., Clements, V. K., Leips, J., Ostrand-Rosenberg, S. (2007) Reduced inflammation in the tumor microenvi-ronment delays the accumulation of myeloid-derived suppressor cellsand limits tumor progression. Cancer Res. 15, 10019–10026.

34. Kodumudi, K. N., Weber, A., Sarnaik, A. A., Pilon-Thomas, S. (2011)Blockade of myeloid-derived suppressor cells after induction of lym-phopenia improves adoptive T cell therapy in a murine model of mela-noma. J. Immunol. 189, 5147–5154.

35. Rosborough, B. R., Castellaneta, A., Natarajan, S., Thomson, A. W.,Turnquist, H. R. (2013) Histone deacetylase inhibition facilitates GM-CSF-mediated expansion of myeloid-derived suppressor cells in vitro andin vivo. J. Leukoc. Biol. 91, 701–709.

36. Sander, L. E., Sackett, S. D., Dierssen, U., Beraza, N., Linke, R. P., Muller, M.,Blander, J. M., Tacke, F., Trautwein, C. (2009) Hepatic acute-phase proteinscontrol innate immune responses during infection by promoting myeloid-de-rived suppressor cell function. J. Exp. Med. 207, 1453–1464.

37. Mundy-Bosse, B. L., Lesinski, G. B., Jaime-Ramirez, A. C., Benninger, K.,Khan, M., Kuppusamy, P., Guenterberg, K., Kondadasula, S. V., Chaud-hury, A. R., La Perle, K. M., Kreiner, M., Young, G., Guttridge, D. C.,Carson, W. E. (2013) Myeloid-derived suppressor cell inhibition of theIFN response in tumor-bearing mice. Cancer Res. 71, 5101–5110.

38. Van Mourik, I. D., Kelly, D. A. (2001) Immunosuppressive drugs in pae-diatric liver transplantation. Paediatr. Drugs 3, 43–60.

39. Barshes, N. R., Goodpastor, S. E., Goss, J. A. (2004) Pharmacologic im-munosuppression. Front. Biosci. 9, 411–420.

40. Chen, S., Akbar, S. M., Abe, M., Hiasa, Y., Onji, M. (2011) Immunosup-pressive functions of hepatic myeloid-derived suppressor cells of normalmice and in a murine model of chronic hepatitis B virus. Clin. Exp. Im-munol. 166, 134–142.

41. Weichhart, T., Costantino, G., Poglitsch, M., Rosner, M., Zeyda, M., Stuhlmeier,K. M., Kolbe, T., Stulnig, T. M., Horl, W. H., Hengstschlager, M., Muller, M.,Saemann, M. D. (2008) The TSC-mTOR signaling pathway regulates the in-nate inflammatory response. Immunity 29, 565–577.

42. Mcinturff, A. M., Cody, M. J., Elliott, E. A., Glenn, J. W., Rowley, J. W.,Rondina, M. T., Yost, C. C. (2012) Mammalian target of rapamycin reg-ulates neutrophil extracellular trap formation via induction of hypoxia-inducible factor 1 �. Blood 120, 3118–3125.

43. Weichhart, T., Haidinger, M., Katholnig, K., Kopecky, C., Poglitsch, M.,Lassnig, C., Rosner, M., Zlabinger, G. J., Hengstschlager, M., Muller, M.,Horl, W. H., Saemann, M. D. (2011) Inhibition of mTOR blocks theanti-inflammatory effects of glucocorticoids in myeloid immune cells.Blood 117, 4273–4283.

44. Schmitz, F., Heit, A., Dreher, S., Eisenacher, K., Mages, J., Haas, T.,Krug, A., Janssen, K. P., Kirschning, C. J., Wagner, H. (2008) Mamma-lian target of rapamycin (mTOR) orchestrates the defense program ofinnate immune cells. Eur. J. Immunol. 38, 2981–2992.

45. Hackstein, H., Taner, T., Zahorchak, A. F., Morelli, A. E., Logar, A. J.,Gessner, A., Thomson, A. W. (2003) Rapamycin inhibits IL-4-induceddendritic cell maturation in vitro and dendritic cell mobilization andfunction in vivo. Blood 101, 4457–4463.

46. Turnquist, H. R., Raimondi, G., Zahorchak, A. F., Fischer, R. T., Wang,Z., Thomson, A. W. (2007) Rapamycin-conditioned dendritic cells arepoor stimulators of allogeneic CD4� T cells, but enrich for antigen-spe-cific Foxp3� T regulatory cells and promote organ transplant tolerance.J. Immunol. 178, 7018–7031.

47. Hegde, V.l., Nagarkatti, P. S., Nagarkatti, M. (2011) Role of myeloid-derived suppressor cells in amelioration of experimental autoimmunehepatitis following activation of TRPV1 receptors by cannabidiol. PLoSOne 6, e18281.

48. Longhi, M. S., Mitry, R. R., Samyn, M., Scalori, A., Hussain, M. J., Qua-glia, A., Mieli-Vergani, G., Ma, Y., Vergani, D. (2009) Vigorous activationof monocytes in juvenile autoimmune liver disease escapes the controlof regulatory T-cells. Hepatology 50, 130–142.

49. Jamieson, T., Clarke, M., Steele, C. W., Samuel, M. S., Neumann, J.,Jung, A., Huels, D., Olson, M. F., Das, S., Nibbs, R. J., Sansom, O. J.(2012) Inhibition of CXCR2 profoundly suppresses inflammation-drivenand spontaneous tumorigenesis. J. Clin. Invest. 122, 3127–3144.

50. Mei, J., Liu, Y., Dai, N., Hoffmann, C., Hudock, K. M., Zhang, P., Gut-tentag, S. H., Kolls, J. K., Oliver, P. M., Bushman, F. D., Worthen, G. S.(2012) Cxcr2 and Cxcl5 regulate the IL-17/G-CSF axis and neutrophilhomeostasis in mice. J. Clin. Invest. 122, 974–986.

KEY WORDS:

cell migration � inflammatory cell infiltration � immune defense

� autoimmune � inflammation

10 Journal of Leukocyte Biology Volume 95, June 2014 www.jleukbio.org