ORIGINAL ARTICLE

Subclinical GvHD in non-irradiated F1 hybrids: severe lymphoid-tissue

GvHD causing prolonged immune dysfunction

B Sprangers1,5, B Van Wijmeersch1,2,5, A Luyckx1,5, X Sagaert3, B Verbinnen4, O Rutgeerts1,C Lenaerts1, T Tousseyn3, B Dubois2, M Waer1 and AD Billiau1

1Laboratory of Experimental Transplantation, University of Leuven, Leuven, Belgium; 2Laboratory of Neuroimmunology, Universityof Leuven, Leuven, Belgium; 3Department of Pathology, University of Leuven, Leuven, Belgium and 4Laboratory of ExperimentalImmunology, University of Leuven, Leuven, Belgium

GvHD is an important complication of allogeneichematopoietic SCT. Parent-in-F1 models are frequentlyused to study GvHD immunobiology; the characteristicsof parent-in-F1 GvHD vary between strain combinationsand induction protocols. Here, we observed that ahigh-dose challenge of non-irradiated B6DBA2F1 andB6SJLF1 recipients with C57BL/6 splenocytes left themajority of recipients clinically healthy, while inducingprogressive high-grade donor T-cell chimerism. Weinvestigated this previously undescribed pattern ofparent-in-F1 T-cell alloreactivity and studied the effectof serial parental splenocyte infusions on epithelial andlymphohematopoietic tissues. The majority of recipientsof 4 weekly splenocyte infusions showed long-termsurvival with gradual establishment of high-grade donorchimerism and without any signs of epithelial-tissueGvHD. A minority of recipients showed BM failure typeof GvHD and, respectively, graft rejection. Moreover,long-term F1 chimeras showed protracted pancytopenia,and in peripheral lymphoid tissues severe lymphopenia andnear-complete eradication of APCs and dysfunction inantigen-presenting capacity in remaining APC. Hemato-poiesis and lymphoid tissue composition recovered onlyafter multilineage donor chimerism had established.In conclusion, we report on a novel type of parent-in-F1hybrid GvHD, where a cumulative high dose of C57BL/6parental splenocytes in non-irradiated F1 mice inducessubclinical but severe hematolymphoid-tissue GvHD,causing prolonged immuno-incompetence.Bone Marrow Transplantation (2011) 46, 586–596;doi:10.1038/bmt.2010.162; published online 5 July 2010Keywords: GvHD; mouse; APC; parental splenocyteinfusion

Introduction

GvHD remains an important complication after allogeneichematopoietic SCT, in particular in the context of donorlymphocyte infusion (DLI) therapy for induction of GvLeffects.1 In recent years, substantial progress has beenmade in the understanding of the immunobiology of DLI-associated GvHD and GvL effects. Mouse studies haveshown that recipient APCs residing in epithelial tissues arecritical targets for alloreactive T cells in the induction ofGvHD,2 and that local inflammation, such as that resultingfrom pretransplant conditioning, is a prerequisite forthe migration of alloreactive T cells into these GvHDtarget tissues.3 Indeed, abatement of conditioning-inducedinflammation is thought to contribute to the observedreduced GvHD risk when DLI is delayed with severalweeks or months after a transplant.4–9 In contrast, the GvLresponse of DLI has been shown in mice to result from theinteraction between donor T cells and recipient APC in thelymphohematopoietic tissues, resulting in a lymphohema-topoietic GvH reaction (LHGvHR).2 In mice, mixedchimerism and a LHGvHR with conversion of mixedto full donor chimerism have been shown essential forGvL responses after DLI.8,9 Moreover in the clinicalsetting, antitumor responses after DLI frequently occur inassociation with conversion of mixed to full donor T-cellchimerism.10,11

Parent-in-F1 models are frequently used to study theimmunobiology of GvHD.12,13 Earlier studies in theC57BL/6-B6DBA2F1 model showed that high-doseC57BL/6 splenocytes induce acute epithelial-tissue GvHDthat is lethal over several months.14 In contrast, high-doseC57BL/6 lymph node (LN) cells induces a BM failuretype of GvHD (BMGvHD), which is rapidly lethal;15–17

BMGvHD is also seen if recipients are irradiated, in whichcase lower quantities of C57BL/6 T cells are required. Thecharacteristics of parent-in-F1 GvHD is equally known tovary between strain combinations, and to depend on strain-specific alloreactive T-cell-precursor frequency.18 We heremade the interesting observation in the C57BL6-B6DBA2F1 model that a high-dose challenge with parentalsplenocytes in non-irradiated recipients caused typicalBMGvHD in only 1/3 of recipient mice. The majority of

Received 15 March 2010; revised 14 May 2010; accepted 23 May 2010;published online 5 July 2010

Correspondence: Dr AD Billiau, Laboratory of Experimental Trans-plantation, University of Leuven, Campus Gasthuisberg, CDG box 811,Herestraat 49, Leuven B-3000, Belgium.E-mail: an.billiau@med.kuleuven.be5These authors contributed equally to this work.

Bone Marrow Transplantation (2011) 46, 586–596& 2011 Macmillan Publishers Limited All rights reserved 0268-3369/11

www.nature.com/bmt

recipient mice remained clinically healthy, while theyshowed progressive evolution towards complete donorT-cell chimerism. A similar evolution was seen in C57BL/6-B6SJLF1 mice. This pattern of parent-in-F1 alloreac-tivity has so far not been reported on, and we documenthere that this represents a form of subclinical GvHD withsevere lymphoid-tissue GvH reactivity causing prolongedimmuno-incompetence. These data may hold implicationsfor DLI strategies in non-myeloablative transplant regi-mens, which aim at inducing strong lymphohematopoieticGvH reactivity in the absence of conditioning-relatedinflammation.

Materials and methods

AnimalsB6DBA2F1 (H-2b/d), B6SJLF1 (H-2b/s) or B6C3HF1 (H-2b/k)female mice were used as recipients and C57BL/6 (H-2b)female mice as donors, and purchased from Janvier(Le Genest Saint Isle, France). Recipients were housed inindividually ventilated cages; during procedures, animalswere kept under laminar flow. Diet consisted of standar-dized pellet chow and UV-decontaminated water. Allexperiments were approved by the local ethical committee.

BM transplantation and parental splenocyte infusionIn the serial parental splenocyte infusion (PSI) model,B6DBA2F1 or B6SJLF1 animals were used as recipientsand given 4 weekly IV injections of 25� 106, respectively,50� 106 C57BL/6 splenocytes. In the allogeneic BM trans-plant (BMT) model, B6SJLF1 mice were given 9.5Gy TBI(linear accelerator 18MEV photons (General Electric,Baden, Germany), dose rate 3.9Gy/min with focus-to-midbody distance 100 cm) and IV infusion of 5� 106

T-cell-depleted (described earlier19) C57BL/6 BM cells.

Leukemia challengeFor GvL studies, the P815 cell line (DBA2 mousemastocytoma, ATCC, H2Kd) was used. Cells from frozenstock were maintained in vitro for a limited number ofpassages. For each experiment, the in vivo behavior wasverified by inoculating naive host-type mice. BM chimeraswere challenged IV with 5� 105 leukemia cells 7 days afterthe first PSI.

Animal follow-upClinical monitoring. Animals were inspected daily forweight loss, signs of acute GvHD affecting general conditionand epithelial organs,20 leukemic disease and mortality.

Donor chimerism. Peripheral blood lineage-specific donorchimerism was studied using flowcytometry (FACStarPlus,FACSCanto, BD Biosciences, Erembodegem, Belgium), asdescribed earlier,21 using mAb against CD3; CD45R/B220;CD11c; CD11b; H-2Kd; H-2Ks; H-2Kk; H-2Kb (BDBiosciences, eBioscience, Hatfield, UK).

Histopathology. Moribund animals were killed forhistopathology: following standard fixation procedures,

paraffin-embedded sections were prepared and stainedwith hematoxylin-eosin H&E. Pictures were taken with aZeiss Axioplan-2 microscope (Carl Zeiss AG, Gottingen,Germany) equipped with a cooled charge-coupled devicecamera COHU 4910 (Diagnostic Instruments, Detroit, MI,USA) or a Leica DM LB2 microscope (Leica Microsystem,Groot Bijgaarden, Belgium) equipped with a Nicon HCL3TP camera (Nikon Belux, Brussels, Belgium).

Cell phenotypic studies. Cellular compositions of BM,spleen and LNs were studied by flowcytometry(as described earlier). Total cell counts were enumeratedmanually using Trypan Blue Exclusion. Complete bloodcounts (CBCs) were determined using a Cell Dyn 3500R(Abbott Diagnostics, Ottignies/Louvain-La-Neuve, Bel-gium). The normal range of CBC values was determinedin a group of 5–7 age-matched naive F1 mice. Values inexperimental mice were considered abnormal if lower orhigher than the mean±2 s.d. from this control group.

Induction of experimental autoimmune encephalomyelitisand in vitro stimulation assaysInduction of experimental autoimmune encephalomyelitis.Experimental autoimmune encephalomyelitis (EAE) wasinduced by bilateral hind-footpad injection of myelinoligodendrocyte glycoprotein35–55 (MOG35–55) peptide(Sigma-Aldrich, Bornem, Belgium), heat-killed Mycobac-terium Tuberculosis in CFA (BD Biosciences) and IVinjection of Bordetella Pertussis toxin (Sigma-Aldrich), asdescribed earlier.22

In vitro stimulation assays. EAE-induced chimeras werekilled on day 10. Popliteal LN cells (LNc) were stimulatedwith MOG35–55, as described earlier.22 In brief, LNc wereplated (2� 105 cells/well, quadruplicate) in stimulationmedium (RPMI containing Na pyruvate; NEAA;L-glutamine; 2-ME; FCS) together with 2� 105 Mitomycin-C-treated (Kyowa Hakko Kogyo, Tokyo, Japan) host-typesplenocytes (as APC source). MOG35–55 was added at10 mg/mL, or ConA at 6 mg/mL ConA. After 4-day culture,[3H]thymidine-incorporation was determined (c.p.m.), asdescribed earlier.22 Results are expressed as stimulationindex¼ c.p.m. of stimulated/c.p.m. of non-stimulated cells.

MOG35–55-specific T-cell proliferation assay. B6SJLF1mice were immunized as described and killed on day 10.Popliteal LNc were stimulated for 72 h at 4� 106 cells/mLin stimulation medium containing 20 mg/mL MOG35–55.T cells were subsequently expanded in RPMI containing10% FCS and 20U/mL rIL-2 for 5–7 days. Finally, 4� 104

T cells were re-stimulated with 10 mg/mL MOG35–55 in thepresence of 0.8� 106 Mitomycin-C-treated splenocytes orLNc (from chimeric or naive mice). [3H]Thymidineincorporation was determined as described above.

MLRMLR was performed as described earlier.21 In brief,responder cells were nylon wool-enriched splenocytes ofnaive C57BL/6 mice, and stimulator cells Mitomycin-C-treated LNc or splenocytes from chimeras or naive

P-in-F1 lymphoid-tissue GvHD with prolonged immuno-incompetenceB Sprangers et al

587

Bone Marrow Transplantation

host-type mice. After 5-day culture, [3H]Thymidineincorporation was determined. Results are expressed asc.p.m. or stimulation index.

Statistical analysisThe Mann–Whitney U test was used to estimate the level ofstatistical significance of difference between groups of data,and the log-rank test to test the difference in survivalbetween groups (Po0.05 was considered as evidence forstatistical significance).

Results

Repetitive parental C57BL/6 splenocyte infusionin unconditioned B6DBA2F1 recipients can induceprogressive engraftment without clinical GvHDB6DBA2F1 recipients were given 4 weekly PSIs of 25� 106

C57BL/6 splenocytes. Two patterns of clinical outcome andchimerism evolution were documented. Approximately 1/3of recipients exhibited rapidly progressive donor T-cellchimerism (arbitrarily defined as 440% on day 14),

reaching 490% around day 30 (Figure 1a). This evolutionwas associated with a mortality rate of 79% by day 42(Figure 1b), the animals experienced sudden death withoutsignificant weight loss (Figure 1c) or clinical pathologicalsigns. In a separate experiment, moribund animals inthis early phase were killed: peripheral blood CBC showedpancytopenia (Table 1); BM histology (Figure 2a) showeddilated sinuses, depletion of hematopoiesis with absenterythropoiesis, maturation arrest of the myeloid lineageand dysplastic megakaryocytes, residual dendriticreticulum cell background and presence of lymphocytes.This is consistent with the previously described C57BL/6-B6DBA2F1 form of BM failure type of acute GvHD.15

The majority of recipients, however, showed a gradualincrease of donor T-cell chimerism (defined as o40% onday 14), reaching 490% around day 90 (Figure 1a). Thiswas associated with 84% long-term survival (Figure 1b)and a favorable weight evolution (Figure 1c), and none ofthe animals showed clinical signs of acute epithelial GvHD(not shown). CBC analysis in long-term surviving animalsequally showed pancytopenia, with recovery of cell counts inall lineages after day 61 (Figure 3), suggesting thatlow-grade mortality was equally due to BM failure; indeed,

100

80

60

40

20

0

140

120

100

80

60

40

20

0

100

80

60

40

20

0

100

80

60

40

20

0

F1 rapid (n=37)F1 gradual (n=84)

F1 chimera (n=27)Control F1(n=13)

F1 rapid (n=12)

Control F1 (n=76)F1 gradual (n=74)

F1 rapid (n=37)

*

*

Control F1 (n=76)F1 gradual (n=84)

0 20 40 60 80 100

0 20 40 60 80 100 0 10 20 40 50 60 7030

0 20 40 60 80 100Days after start PSI regimen

Days after start PSI regimen Days after leukemia challenge

Days after start PSI regimen

T c

ell d

on

orc

him

eris

m (

%)

Wei

gh

t (%

of

star

tin

g w

eig

ht)

Su

rviv

al (

%)

Su

rviv

al (

%)

ba

dc

Figure 1 Evolution of donor T-cell chimerism, weight, survival and survival after leukemia challenge in the C57BL/6-B6DBA2F1 chimeras. B6DBA2F1recipients given parental splenocyte infusions (PSI) were stratified according to the pattern of donor T-cell chimerism evolution. (a) Evolution of donorT-cell chimerism: animals showing X40% T-cell donor chimerism on day 14 after the first infusion were assigned to the ‘F1 rapid’-group (n¼ 37) and thoseshowing o40% T-cell donor chimerism were assigned to the ‘F1 gradual’-group (n¼ 84). (b) Kaplan–Meier survival curves showing the survival of the ‘F1rapid’- and the ‘F1 gradual’-group, as well as that of a group of naive host-type mice that were included in each experiment as controls (control F1). Resultsin panels a and b are from 16 similarly designed experiments. *Po0.05 between ‘F1 rapid’ and ‘F1 gradual’ (log-rank). (c) Weight evolution (expressed asmean±s.d. percentage of individual starting weights) in the ‘F1 rapid’-group (n¼ 12, long-term survivors n¼ 4), the ‘F1 gradual’-group (n¼ 74), and agroup of host-type control animals (n¼ 76). Results are from 12 similarly designed experiments. (d) B6DBA2F1 mice given repetitive C57BL/6 infusion(F1 chimera), and age-matched naive B6DBA2F1 mice not given PSI (control F1) were challenged with 5� 105 P815 leukemia cells on day 7. Kaplan–Meiersurvival curve is shown of a total of 27 F1 chimera and 13 control F1 animals, from four identically designed experiments, which included simultaneoustesting of all three groups. *Po0.05 for the difference in survival between groups (log-rank).

P-in-F1 lymphoid-tissue GvHD with prolonged immuno-incompetenceB Sprangers et al

588

Bone Marrow Transplantation

BM histology of moribund mice was similar to that of mori-bund animals in the early mortality group (as in Figure 2a).

These data indicate that strong LHGvH reactivity occurswith repetitive PSI. As this type of alloreactivity isexploited by DLI strategies to obtain GvL effects, in anext set of experiments we investigated whether repetitivePSI would allow for a GvL response. One week after

the first PSI, chimeric B6DBA2F1 recipients were chal-lenged with 5� 105 P815 cells. PSI-treated animalsshowed a significant survival benefit over naive host-typeanimals, with long-term survival of 37% (Figure 1d).Histopathology in moribund animals confirmed tumorinfiltration in LN, liver, spleen, salivary gland, muscle andBM (not shown).

Table 1 Peripheral blood CBC values

Day of analysis White blood cell count(10e3/mL)

Red blood cell count(10e6/mL)

Hemoglobin(g/100mL)

Platelet count(10e3/mL)

1 (wD20) D6 3 8.46 14.88 5256D13 1.34 7.4 11.12 378

2 (wD16) D6 3.7 9.6 14.62 1200D13 1.19 6.46 9.82 728

D16a 0.25 4.87 6.97 34.2

3 (wD7) D6 3.46 12.44 18.44 14804 (wD15) D6 2.1 9.36 14.06 1172

D13 1.19 5.8 8.7 632

5 (wD20) D6 4.04 9.96 14.7 1122

D13 0.92 6.68 9.94 310

D20a 1.4 5.39 7.75 33.6

Reference values (range)b 3.29–9.65 8.1–10.16 10.52–16.98 1128–1538

Abbreviation: CBC¼ complete blood count.aMoribund.bCalculated as detailed in Materials and methods.Bold font indicates cytopenia.wKilled at.

a b

c d

Figure 2 Histological findings and CBC analysis in F1 chimeras. (a–d) Histological findings in F1 chimeras. H&E stain of a BM section of ‘F1 rapid’B6DBA2F1 chimera demonstrating dilated bone sinuses, a depletion of hematopoiesis with absent erythropoiesis, maturation arrest of the myeloid lineageand dysplastic megakaryocytes, residual dendritic reticulum cell background and presence of lymphocytes (magnification � 20). H&E staining of LN of anaive B6DBA2F1 mouse (b) showing the normal architecture with abundance of lymphocytes, a C57BL/6-B6SLJF1 chimera at day 60 (c) showing aneffaced architecture of the lymph tissue that is depleted of lymphoid cells but shows an abnormal presence of myeloid precursor cells identifiable by theirkidney- to donut-shaped nuclei and granular cytoplasm (detail shown in inset), and a C57BL/6-B6DBA2F1 chimera at day 60 (d) showing an effacedarchitecture of the lymph tissue depleted of lymphoid cells without an associated expansion of myeloid precursor cells. Pictures were taken with a ZeissAxioplan 2 microscope (Carl Zeiss AG, Gottingen, Germany) equipped with a cooled charge-coupled device camera COHU 4910 (Diagnostic Instruments,Detroit, MI, USA). Original magnifications � 40.

P-in-F1 lymphoid-tissue GvHD with prolonged immuno-incompetenceB Sprangers et al

589

Bone Marrow Transplantation

Thus, in unconditioned F1 recipients, repetitive PSIcan establish strong LHGvHR with a gradual evolutionto high-grade donor chimerism, and with a strongGvL response. Despite the high cumulative T-cell dose,this occurs without acute epithelial GvHD. However,although rapid institution of donor T-cell chimerism isassociated with rapidly lethal BMGvHD, the gradualcourse is equally accompanied by prolonged pancytopenia,indicating the development of subclinical BMGvHD that issevere but that holds a significantly lower mortality risk.

To determine whether the additional infusion of donorstem cells would alter the characteristics of the GvHR, wedetermined the effect of serial PSI in F1 recipients where

donor BM was coinfused with the parental splenocyteinoculums. F1 recipients were challenged with serialPSIs only, or with serial PSIs and 5� 106 T-cell-depletedC57BL/6 BM cells at the first or the third PSI. CBC weresequentially determined but showed that coinfusion of BMcells influenced neither the severity nor the duration ofpancytopenia (Figures 3b and c).

The serial PSI regimen induces a disruption of the cellularcomposition of lymphohematopoietic organsThe pattern of parent-in-F1 alloreactivity that allowsprogressive donor chimerism and GvL with good overall

8.0

7.0

6.0

5.0

4.0

3.0

2.0

1.0

0.0

8.0

7.0

6.0

5.0

4.0

3.0

2.0

1.0

0.0

8.0

7.0

6.0

5.0

4.0

3.0

2.0

1.0

0.0

Wh

ite

blo

od

cel

ls (

103 /µ

l)

D61D77

D35D42

0 20 40 60 80 100 7 29 54 73 13 2328 61 83

13 23 28 61 83

13 23 28 61 83

0 20 40 60 80 100 7 29 54 73

0 20 40 60 80 100 7 29 54 73

12.0

10.0

6.0

4.0

2.0

0.0

8.0

12.0

10.0

6.0

4.0

2.0

0.0

8.0

12.0

10.0

6.0

4.0

2.0

0.0

8.0

Red

blo

od

cel

ls (

106 /µ

l)P

late

lets

(10

3 /µl)

1600

1400

1200

1000

800

600

400

200

0

1600

1400

1200

1000

800

600

400

200

0

1600

1400

1200

1000

800

600

400

200

0

% CD3 donor chimerism Days after first PSI Days after first PSI

Figure 3 Longitudinal analysis of peripheral blood counts in long-term surviving B6DBA2F1 chimeras, and relation with level of peripheral blood donorT-cell chimerism and effect of BM coinfusion. (a) Analysis of peripheral CBC and donor T-cell chimerism in C57BL/6-B6DBA2F1 chimeras given parentalsplenocyte infusions, at four different time points: D35(x), D42(,), D61(J), D77(E). Dotted lines indicate the reference range of CBC values as determinedin age-matched naive F1 mice (see Materials and methods). The data from 1 out of 4 independent experiments is shown (n¼ 13 at D35 and D42, n¼ 12 atD61 and 77). (b and c) Evolution of peripheral complete blood counts in long-term surviving B6DBA2F1 chimeras with (E) or without (J) coinfusion ofdonor BM cells on day 0 (¼ first PSI) (b) and on day 14 (¼ third infusion) (c). Dotted lines indicate the reference range of CBC values as determined in age-matched naive F1 mice (see Materials and methods). In both b and c, the results of 1 of 2 representative experiments are shown, n¼ 7 per group.

P-in-F1 lymphoid-tissue GvHD with prolonged immuno-incompetenceB Sprangers et al

590

Bone Marrow Transplantation

tolerance and long-term survival has previously notbeen reported. The effects on peripheral blood CBC areindicative of subclinical GvHD targeting BM function. Wefurther explored the extent of this subclinical GvHDby evaluating the cellular composition and histologicalappearance of lymphoid and hematopoietic tissues, andepithelial GvHD target tissues. Groups of threeB6DBA2F1 mice receiving serial splenocyte infusions werekilled at selected time points, as indicated in Figure 4. Noneof the animals showed histological signs of acute epithelialGvHD, or chronic SLE-like GvHD (for example immune-complex glomerulonephritis) previously reported in F1hybrids23 (not shown). Total cell counts and proportions ofT, B and myeloid cells were determined on spleen, BM andLN (Figures 4a–c). Results were expressed relative to themean value obtained in a group of three naive B6DBA2F1mice that were analyzed at each time point. The chimerasshowed a transient expansion of total cell numbers,followed by a progressive decline in spleen and LN cell

counts beyond day 84, and a temporary decrease in BM cellcounts with normalization by day 84 (Figure 4a). Analysisof the cell subpopulations in spleens and LN (Figures 4band c) showed a 2-week phase of lymphoproliferation and amyeloid cell expansion, followed by a progressive declinein these cell lineages.

Lymphoproliferation probably is a result of alloreactiveT-cell proliferation in BM and lymphoid organs. Theexpanding CD11bþ population showed a CD11bþF4/80�GR1hi phenotype, and a morphology typical of myeloidprogenitor cells (not shown), similar to those previouslydescribed in BM chimeras after myeloablation.9,21

We postulate therefore that this peripheral expansionof myeloid progenitor cells occurs as a consequenceof the BM depression caused by alloreactive T cells. Inconclusion, in addition to causing subclinical BMGvHD,the serial PSI regimen induces a profound disruptionof the cellular composition of lymphohematopoieticorgans.

600

500

400

300

200

100

0

Tota

l cel

l co

un

t (%

, rel

ativ

e to

nai

ve B

6DB

A2F

1)

Tota

l cel

l co

un

t (%

, rel

ativ

e to

nai

ve B

6DB

A2F

1)

Tota

l cel

l co

un

t (%

, rel

ativ

e to

nai

ve B

6DB

A2F

1)

Tota

l cel

l co

un

t (%

, rel

ativ

e to

nai

ve B

6SJL

F1)

SpleenLymph nodeBone marrow

0 20 40 60 80 100

Days after start PSI regimen

0 20 40 60 80 100

Days after start PSI regimen

0 20 40 60 80 100

Days after start PSI regimen

0 20 40 60 80 100

Days after start PSI regimen0 20 40 60 80 100

Days after start PSI regimen

0 20 40 60 80 100

Days after start PSI regimen

400

350

300

250

200

150

100

50

0

10000

1000

100

10

1

Tota

l cel

l co

un

t (%

, rel

ativ

e to

nai

ve B

6SJL

F1)

10000

1000

100

10

1

CD11b+ cellsCD3+ cellsB220+ cells

CD11b+ cellsCD3+ and B220+ cells

SpleenLymph nodeBone marrow

10000

1000

100

10

1

0.1 Tota

l cel

l co

un

t (%

, rel

ativ

e to

nai

ve B

6SJL

F1)

10000

1000

100

10

CD11b+ cellsCD3+ cellsB220+ cellsCD11c+ cells

CD11b+ cellsCD3+ cellsB220+ cellsCD11c+ cells

Figure 4 Evolution of total cell counts and cellular composition in spleen, BM and LN of C57BL/6-B6DBA2F1 chimeras and C57BL/6-B6SJL2F1chimeras. Groups of three B6DBA2F1 chimera (a–c) or B6SJL2F1 mice (d–f) were killed at selected time points. Total cell enumeration was performed onspleen, BM and LN (a, d) and the proportions of lymphocytes (CD3þT cells, B220þB cells), CD11bþ myeloid cells and/or CD11cþ dendritic cells weredetermined in spleen (b, e) and LN (c, f). Each test included the analysis of a group of three naive host-type mice and results are expressed relative to themean value obtained in this group. Data represent mean %±s.e.m. of individual experimental mice.

P-in-F1 lymphoid-tissue GvHD with prolonged immuno-incompetenceB Sprangers et al

591

Bone Marrow Transplantation

The cellular disruption of the lymphoid organs after serialPSIs includes an APC defectImmune function requires a functional APC compartment,and host-type APC have a critical role in the developmentof GvL and GvHD.2,8,24,25 We specifically investigated theevolution of host-type APC in the C57BL/6-B6DBA2F1model, but also in the similar C57BL/6-B6SJLF1combination. In the latter model, we also recently showedthat 60% of non-irradiated B6SJLF1 recipients develophigh-grade donor chimerism while remaining clinicallyhealthy, whereas 40% of them rejected the parentalsplenocyte grafts.22 We also showed that such C57BL/6-B6SJLF1 chimeras become resistant to the development ofMOG35–55-induced EAE.22 Induction of EAE by footpadinjection of MOG peptide/CFA relies on the presentationof MOG peptide by APC in the draining popliteal LN. Wehypothesized that—in view of the lymphohematopoieticchanges seen in C57BL/6-B6DBA2F1 chimeras, whichincluded a profound decline in CD11cþ cells (Figure 4c)—the resistance to EAE induction in long-term C57BL/6-B6SJLF1 chimeras relates to disappearance of APC indraining LN. Both C57BL/6-B6SJLF1 and C57BL/6-B6DBA2F1 models were used to study the evolutionof APC in more detail.

First, we confirmed that long-term surviving C57BL/6-B6SJLF1 chimeras exhibit a similar disruption of thelymphohematopoietic system (Figure 4d): spleen andLN equally showed a consecutive expansion and depletionof myeloid and lymphoid elements, and, in particular,a marked reduction in total CD11cþ cell counts (Figures 4eand f). Of note, a difference in the kinetics was seen in thetwo models, with the myeloid expansion showing a moreprotracted course in the B6SJLF1 mice (Figure 4). Histo-logy of lymphoid organs on day 60 showed disruptionof the normal architecture of LN (Figure 2c) and spleen(not shown): LN were largely depleted of lymphocytes,and, in accordance with flowcytometry data, in B6SJLF1mice they showed a prominent presence of myeloidprogenitor cells.

Next, LNc of these PSI chimeras were used in in vitroassays to document the functional presence or absence ofAPC. LNc derived from C57BL/6-B6SJLF1 chimeras(on day 56 after the first splenocyte infusion) andfrom naive B6SJLF1 mice were used as APC source in aMOG35–55-specific B6SJLF1 T-cell proliferation assay or asstimulators in allogeneic MLR. In contrast to naiveB6SJLF1 LNc, LNc from C57BL/6-B6SJLF1 chimerassupported neither the MOG-specific proliferative T-cellresponse (Figure 5a), nor the MLR response of C57BL/6responder cells (Figure 5b). In the latter setting, coculturewith T-cell-depleted B6SJLF1 splenocytes (as an additionalsource of host-type APC) restored the MLR reaction(Figure 5b). From the flowcytometric and these functionaldata, we conclude that the disruption of lymphoid organs inB6SJLF1 chimeras includes a quantitative defect in host-typeAPC. However, LN of C57BL/6-B6SJLF1 chimeras con-tain a considerable proportion of myeloid progenitor cells(Figures 2c and 4f), which may act as suppressor cells andaccount for the lack of proliferative T-cell responses in vitro.21

Therefore, similar MLR experiments were performedusing cells from B6DBA2F1 chimeras, where myeloid

progenitor cells disappear more rapidly (Figures 2d and4c). Spleen cells and LN cells from chimeras challengedwith PSIs only, or challenged with PSIs together with BMcells were used as stimulator cells for C57BL/6 T cellsin MLR. In this particular experiment, day 73 chimerasshowed persistent cytopenia in LN and spleens (notshown), but importantly, the frequencies of CD11bþ andCD11cþ cells had normalized relative to naive F1 mice,whereas host chimerism in CD11cþ cells was still 480%(not shown). We found that spleen and LN cells from thesechimeras were poor stimulators for naive C57BL/6responder cells (Figures 5c and d). Although purifiedAPC were not tested owing to insufficient cell yield fromthe severely cytopenic lymphoid organs, the data suggestthat—in addition to a quantitative defect in APC num-bers—long-term PSI chimeras may also develop a qualita-tive dysfunction of residual APC.

Long-term surviving PSIs chimeras do not exhibit generalT-cell hyporesponsivenessIn view of the profound lymphopenia in long-term F1chimeras induced by PSIs, we examined T-cell function. Wepreviously showed that, in contrast to EAE-inducedB6SJLF1 control mice, splenocytes from C57BL/6-B6SJLF1 chimeras fail to exhibit an in vitro MOG35–55-specific proliferative response.22 Here, we confirmed thatLNc from such chimeras (taken on day 56 after the start ofPSIs, 14 days after MOG challenge) also fail to generate aMOG-specific proliferative response, but exhibit a vigorousresponse to ConA (Figure 5e). Next, we tested splenocytesfrom C57BL/6-B6SJLF1 chimeras (taken on day 56) inMLR with splenocytes from naive B6SJLF1 or BALB/c.Splenocytes from naive B6SJLF1 and C57BL/6-B6SJLF1radiation chimeras (alloBMT/F1, see Materials and meth-ods) were tested as positive and negative controls,respectively. The MLR response against third party wascomparable in all groups, whereas only C57BL/6-B6SJLF1splenocyte infusion chimeras showed a significant anti-B6SJLF1 response (Figure 5f). These data indicate thatT cells from PSI chimeras readily respond to ConA,allogeneic and third-party targets, and that—despite severelymphopenia—the disruptive effect on the lymphoid organsdoes not include a state of general T-cell hyporesponsiveness.

Lymphohematopoietic recovery in long-term PSI chimerasparallels establishment of multilineage donor chimerismFinally, we illustrated that CBC in B6DBA2F1 PSIchimeras evolved in parallel with the level of donorT-cell chimerism (Figure 3c), as most animals showedrecovery of cell counts on day 77 together with near-complete donor T-cell chimerism. This suggestedthat recovered lymphohematopoiesis was derived fromdonor-type stem cells. In LN of long-term (day 88)B6DBA2F1 PSI chimeras, we documented that high-gradedonor chimerism had established in not only the T-celllineage (94%±0.7 s.e.) but also in B cells (90%±1.3 s.e.)and dendritic cells (65%±3.2 s.e.). Similar results wereobtained in spleen cells, with—in addition—high-gradedonor chimerism in the CD11bþ lineage (80% ±1.6 s.e.)(data from one out of two representative experiments).

P-in-F1 lymphoid-tissue GvHD with prolonged immuno-incompetenceB Sprangers et al

592

Bone Marrow Transplantation

Discussion

We here report on a previously undescribed pattern ofGvH reactivity in the well-known C57BL/6-B6DBA2F1parent-in-F1 hybrid mouse model. We found thatserial infusions of parental splenocytes, establishing a

cumulative high dose of alloreactive lymphocytes, can leadto a strong LHGvH response resulting in a formof subclinical GvHD with profound perturbation oflymphohematopoietic tissues, including a depletion ofperipheral APCs with associated prolonged immuneincompetence.

5

4

3

2

1

0

Sti

mu

lati

on

ind

ex

F1 chimera B6SJLF1control

80000

70000

60000

50000

40000

30000

20000

10000

0

Sti

mu

lati

on

ind

ex

F1 chimera B6SJLF1control

F1 chimera B6SJLF1control

40

30

20

10

0

Sti

mu

lati

on

ind

ex

Sti

mu

lati

on

ind

exF1 chimera BM/F1

chimeraNaive F1 F1 chimera BM/F1

chimeraNaive F1

30

20

10

0

100

10

1

Sti

mu

lati

on

ind

ex

F1 chimera (n=4)B6SJLF1 EAE control (n=4)

MOG conA

14000

12000

10000

8000

6000

4000

2000

0

Co

un

ts p

er m

inu

te

PSI/F1 chimera (n=11)

AlloBMT/F1 chimera (n =8)

B6SJLF1 control (n=3)

**

*

B6SJLF1 BALB/c

LNc LNc +host APC

Figure 5 In vitro assays documenting APC and T-cell function in F1 chimeras. (a) MOG35–55-specific T-cell proliferation assay using Mitomycin-treatedLNc of B6SJLF1 chimeras as source of APC. MOG-specific T cells, obtained from EAE-induced control B6SJLF1 mice were stimulated with MOG35–55 inthe presence of LNc derived from C57BL/6-B6SJLF1 chimeras on day 56 (‘F1 chimera’, n¼ 6) or from naive host-type B6SJLF1 mice (‘B6SJLF1 control’,n¼ 2); scatterplot shows individually tested animals and the mean value, and represent data from two independent experiments. (b) MLR response ofC57BL/6 naive splenocytes stimulated with LNc taken from C57BL/6-B6SJLF1 chimeras on day 56 (‘F1 chimera’, n¼ 5) or from naive B6SJLF1 mice(‘B6SJLF1 control’, n¼ 2), either alone (left), or with additional LNc from naive B6SJLF1 mice as a source of host-type APC (right); scatterplot showsindividually tested animals and the mean value, and 1 of 2 representative experiments is shown. (c) MLR response of C57BL/6 naive splenocytes stimulatedwith splenocytes from C57BL/6-B6DBA2F1 chimeras (‘F1 chimera’, n¼ 4 samples, each representing cells of one (one sample) or two mice (threesamples)), splenocytes from C57BL/6-B6DBA2F1 chimeras with BM coinfusion at the first PSI (‘BM/F1 chimera’, n¼ 3 samples, each representing cellsfrom one mouse (one sample) or two mice (two samples)), or splenocytes from naive B6DBA2F1 mice (‘naive F1’, n¼ 6), all taken on day 73. Scatterplotshows individually tested samples and the mean value, and represent data from one experiment. (d) MLR response of C57BL/6 naive splenocytes stimulatedwith LNc from C57BL/6-B6DBA2F1 chimeras (‘F1 chimera’, n¼ 1 sample, representing cells from two mice), splenocytes from C57BL/6-B6DBA2F1chimeras with BM coinfusion at first PSI (‘BM/F1 chimera’, n¼ 2 samples, each representing cells from three mice) or splenocytes from naive B6DBA2F1mice (‘naive F1’, n¼ 4), all taken on day 73. Scatterplot shows individually tested samples and the mean value, and represents data from one experiment.(e) Proliferative response of LNc from C57BL/6-B6SJLF1 chimeras (n¼ 4, harvested on day 56 (¼ day 14 after EAE induction) and from naive B6SJLF1control animals (n¼ 4, harvested 14 days after EAE induction) in response to MOG35–55 or conA. Bars represent mean stimulation index±s.e.m., *Po0.05(Mann–Whitney U-test). (f) MLR response of spleen cells from C57BL/6-B6SJLF1 PSI chimeras (‘PSI/F1 chimera’, n¼ 11) on day 56, naive B6SJLF1control animals (‘B6SJLF1 control’, n¼ 3) or C57BL/6-B6SJLF1 radiation chimeras (‘alloBMT/F1’, n¼ 8) on day 56, in response to B6SJLF1 host-typeor BALB/c third-party spleen cells. Bars represent mean c.p.m.±s.e.m., **Po0.01 (Mann–Whitney U test).

P-in-F1 lymphoid-tissue GvHD with prolonged immuno-incompetenceB Sprangers et al

593

Bone Marrow Transplantation

The C57BL/6-B6DBA2F1 model is considered a well-established model GvHD,12,13 but GvHD characteristicsvary according to the induction protocol. Reportedly, high-dose C57BL/6 splenocytes induce acute epithelial GvHDthat is lethal over several months.14 In contrast, high-doseC57BL/6 LN cells induces a BMGvHD, which is rapidlylethal;15–17 BMGvHD is also seen if recipients areirradiated, in which case lower quantities of C57BL/6T cells are required. Here, we found that challenge ofB6DBA2F1 with C57BL/6 splenocytes gave rise to twodistinct outcome patterns. Only 1/3 of recipients developedlethal BMGvHD; the majority exhibited progressiveinstitution of high-grade donor T-cell chimerism with goodclinical tolerance, long-term survival and no histologicalevidence of acute epithelial-tissue GvHD. However,subclinically these animals showed protracted pancytope-nia, and in lymphoid tissues prolonged lymphopenia anda near-complete eradication of APCs. This pattern oflymphohematopoietic GvHD has previously not beenreported.

Our recipient mice were housed in semi-SPF conditions;hence, variability in intestinal colonization and inflamma-tion may be one of the factors determining to which extentparental splenocyte inocula produce alloactivated cytotoxicT cells26,27 and produce either rapidly lethal or subclinicalGvHD. In this respect, Murai et al.28 showed that C57BL/6splenocyte inoculation in unirradiated B6DBA2F1 resultedin a rapid and preferential accumulation of donor CD8þ Tcells in Peyer’s patches before initiation of GvHD. Weassume that a reduced inflammatory-prone intestinalenvironment protected some mice from aggressiveBMGvHD; however, the resistance of these animals tocumulative high-dose PSI, while developing high-gradedonor chimerism was striking. The data indicate that thisrelates to a delicate balance between donor T cells and hostAPC in a non-inflammatory environment.

The pattern of lymphohematopoietic subclinical GvHDcould be reproducibly obtained in both B6DBA2F1 andB6SJLF1 hybrids challenged with C57BL/6 parental cells,with a slight difference only in the kinetics of the cellularchanges. Both models showed a severe quantitativereduction in dendritic cells in spleens and LN, and ex vivoMLR and MOG-specific T-cell assays showed that thestimulatory capacity of whole spleen and LN cells ofB6SJLF1 mice was defective. Further, the observation thatparental T cells failed to proliferate when stimulated withB6D2F1 chimeric LN or spleen cells in which relativenumbers of CD11cþ cells (that were still largely hostderived) had normalized relative to naive F1 mice, suggeststhat lymphohematopoietic alloreactivity may also affect thefunction of residual APC. Together, the data indicate that aquantitative defect, and possibly also a qualitative defect ofAPC may underlie the resistance of these chimeras to theGvHD-inducing capacity of repeated splenocyte inoculaand their resistance to EAE induction.22

In the B6SJLF1 model, where myeloid cell expansion ismore protracted, an extrinsic regulatory role of these cellson the observed in vitro T-cell proliferative responsescannot be excluded. The expansion of myeloid progenitorcells in these chimeras is reminiscent of the myeloid-derivedsuppressor cells described in association with myelosup-

pression, GvHD, tumors and chronic infections21,29 wherethey have been implicated in tumor-specific and non-specific T-cell dysfunction.30–32 Our earlier data,21 and morerecently those of others,33 suggest that these cells may havea role in controlling GvH reactivity. Whether theycontribute in this model to confining GvH reactivity tothe lymphohematopoietic system is the subject of ongoinginvestigations.

Interestingly, Wang et al.34 recently showed in a similarC57BL/6-B6DBA2F1 model involving recipient irradia-tion that IFN-g promotes the allospecific LHGvHresponse, but attenuates the potential of alloreactive Tcells to mediate epithelial-tissue GvHD. Irradiation isknown to influence the activation/migration of allospecificT cells,3 and the outcome of parent-in-F1 GvHD.35

Assuming that also in our irradiation-free model, IFN-ghas a critical role in LHGvH, blockade of this pathwaywould potentially reduce the risk for severe immuno-incompetence. However, such a limitation of LHGvHresponses could equally limit GvL effects, which, as inthe current model, occur in association with LHGvHreactivity.

Coinfusion of donor BM with parental splenocytes,either at the first challenge in naive mice or at the thirdchallenge when cytopenia (and therefore space in therecipient marrow) was already established, influencedneither the severity nor the duration of pancytopenia inour model, and flow cytometric studies and in vitro assaysshowed that the defect in APC numbers and APC functionwas equally profound in these animals. We further showedthat in irradiated B6C3HF1 mice, the coinfusion of C57BL/6 spleen cells (as a source of hematopoietic stem cells) withpurified C57BL/6 T cells can rescue mice from early deathand converts BMGvHD into acute epithelial-tissue GvHD(not shown); this effect was not seen when F1 BM wascoinfused, indicating that there was no innocent bystanderdestruction of donor-type hematopoietic stem cells. Thishad been earlier demonstrated by Rozendaal et al.36 in theDBA/2-B6D2F1 model. In contrast, Chen et al.16,17 haveshown in the C57BL/6-CByB6F1 model that the BM ofF1 mice with BMGvHD not only fails to form stromalfeeder layers in vitro, but also contains T cells capable ofdestroying fresh donor-type stem cells as well, indicatingthat in this model innocent bystander destruction of bothstromal and hematopoietic progenitor cells is taking place.In contrast to these studies, the model we present here doesnot include irradiation, and inflammation may influence thedegree to which bystander destruction is taking place. Thedata, however, argue against bystander destruction ofdonor-type stem cells, as the eventual lymphomyeloidreconstitution is predominantly donor derived; in contrast,the data suggest that repeated challenges with PSIs targetthe BM stroma, which fails to support effective myelo-hematopoiesis until T-cell alloreactivity has subsided.These data reveal a delicate balance between donor T cell,donor stem cell and BM stromal function.

Although this experimental model of unidirectionalT-cell alloreactivity presents some limitations with regardto extrapolations to the clinical situation, the data may berelevant to the development of clinical DLI strategies. Theyindicate that a challenge with non-tolerant T cells in an

P-in-F1 lymphoid-tissue GvHD with prolonged immuno-incompetenceB Sprangers et al

594

Bone Marrow Transplantation

inflammation-free tolerant recipient can lead to severesubclinical lymphoid-tissue GvHD causing prolongedimmuno-incompetence. The haploidentical pattern ofT-cell alloreactivity is not entirely representative of theclinical situation where most patients receive an MHC-matched graft. Haploidentical DLI in humans is expectedto carry a high risk of causing epithelial GvHD, althoughour data in this study indicate that this may depend on adelicate balance. Of note, there is a growing interest inhaploidentical stem cell transplantation,11 and both delayedgraft function and delayed immune recovery have beenreported as complications after MHC-mismatched BMtransplantation,37 supporting the potential relevance ofour data.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgements

This work was supported by grants from the FWO Vlaanderen,the Stichting tegen Kanker and the WOMS (Scientific ResearchMultiple Sclerosis). BS and BVW are doctoral fellows and ADBis a postdoctoral fellow of the FWO Vlaanderen. AL and BV aredoctoral fellows of IWT. BD is supported by the UniversityResearch Council of the University of Leuven, Belgium.

References

1 Ferrara JL, Levine JE, Reddy P, Holler E. Graft-versus-hostdisease. Lancet 2009; 373: 1550–1561.

2 Reddy P, Maeda Y, Liu C, Krijanovski OI, Korngold R,Ferrara JL. A crucial role for antigen-presenting cells andalloantigen expression in graft-versus-leukemia responses.Nat Med 2005; 11: 1244–1249.

3 Chakraverty R, Cote D, Buchli J, Cotter P, Hsu R, Zhao Get al. An inflammatory checkpoint regulates recruitmentof graft-versus-host reactive T cells to peripheral tissues.J Exp Med 2006; 203: 2021–2031.

4 McSweeney PA, Niederwieser D, Shizuru JA, Sandmaier BM,Molina AJ, Maloney DG et al. Hematopoietic cell transplant-ation in older patients with hematologic malignancies: repla-cing high-dose cytotoxic therapy with graft-versus-tumoreffects. Blood 2001; 97: 3390–3400.

5 Mackinnon S, Papadopoulos EB, Carabasi MH, Reich L,Collins NH, Boulad F et al. Adoptive immunotherapyevaluating escalating doses of donor leukocytes for relapse ofchronic myeloid leukemia after bone marrow transplantation:separation of graft-versus-leukemia responses from graft-versus-host disease. Blood 1995; 86: 1261–1268.

6 de Lima M, Bonamino M, Vasconcelos Z, Colares M,Diamond H, Zalcberg I et al. Prophylactic donor lymphocyteinfusions after moderately ablative chemotherapy and stem celltransplantation for hematological malignancies: high remissionrate among poor prognosis patients at the expense of graft-versus-host disease. Bone Marrow Transplant 2001; 27: 73–78.

7 Badros A, Barlogie B, Morris C, Desikan R, Martin SR,Munshi N et al. High response rate in refractory and poor-riskmultiple myeloma after allotransplantation using a nonmye-

loablative conditioning regimen and donor lymphocyteinfusions. Blood 2001; 97: 2574–2579.

8 Mapara MY, Kim YM, Wang SP, Bronson R, Sachs DH,Sykes M. Donor lymphocyte infusions mediate superior graft-versus-leukemia effects in mixed compared to fully allogeneicchimeras: a critical role for host antigen-presenting cells. Blood2002; 100: 1903–1909.

9 Billiau AD, Fevery S, Rutgeerts O, Landuyt W, Waer M.Crucial role of timing of donor lymphocyte infusion ingenerating dissociated graft-versus-host and graft-versus-leu-kemia responses in mice receiving allogeneic bone marrowtransplants. Blood 2002; 100: 1894–1902.

10 Dey BR, McAfee S, Colby C, Sackstein R, Saidman S, TarbellN et al. Impact of prophylactic donor leukocyte infusions onmixed chimerism, graft-versus-host disease, and antitumorresponse in patients with advanced hematologic malignanciestreated with nonmyeloablative conditioning and allogeneicbone marrow transplantation. Biol Blood Marrow Transplant2003; 9: 320–329.

11 Koh LP, Chao N. Haploidentical hematopoietic cell trans-plantation. Bone Marrow Transplant 2008; 42(Suppl 1):S60–S63.

12 Rus V, Svetic A, Nguyen P, Gause WC, Via CS. Kinetics ofTh1 and Th2 cytokine production during the early course ofacute and chronic murine graft-versus-host disease. Regulatoryrole of donor CD8+ T cells. J Immunol 1995; 155: 2396–2406.

13 Murphy WJ. Revisiting graft-versus-host disease models ofautoimmunity: new insights in immune regulatory processes.J Clin Invest 2000; 106: 745–747.

14 Tschetter JR, Mozes E, Shearer GM. Progression fromacute to chronic disease in a murine parent-into-F1 model ofgraft-versus-host disease. J Immunol 2000; 165: 5987–5994.

15 Bloom ML, Wolk AG, Simon-Stoos KL, Bard JS, Chen J,Young NS. A mouse model of lymphocyte infusion-inducedbone marrow failure. Exp Hematol 2004; 32: 1163–1172.

16 Chen J, Lipovsky K, Ellison FM, Calado RT, Young NS.Bystander destruction of hematopoietic progenitor and stemcells in a mouse model of infusion-induced bone marrowfailure. Blood 2004; 104: 1671–1678.

17 Chen J, Brandt JS, Ellison FM, Calado RT, Young NS.Defective stromal cell function in a mouse model of infusion-induced bone marrow failure. Exp Hematol 2005; 33: 901–908.

18 Via CS, Sharrow SO, Shearer GM. Role of cytotoxic Tlymphocytes in the prevention of lupus-like disease occurringin a murine model of graft-vs-host disease. J Immunol 1987;139: 1840–1849.

19 Billiau AD, Sefrioui H, Overbergh L, Rutgeerts O, Goebels J,Mathieu C et al. Transforming growth factor-beta inhibitslymphokine activated killer cytotoxicity of bone marrow cells:implications for the graft-versus-leukemia effect in irradiationallogeneic bone marrow chimeras. Transplantation 2001; 71:292–299.

20 Fevery S, Billiau AD, Sprangers B, Rutgeerts O, Lenaerts C,Goebels J et al. CTLA-4 blockade in murine bone marrowchimeras induces a host-derived antileukemic effect withoutgraft-versus-host disease. Leukemia 2007; 21: 1451–1459.

21 Billiau AD, Fevery S, Rutgeerts O, Landuyt W, Waer M.Transient expansion of Mac1+Ly6-G+Ly6-C+ early mye-loid cells with suppressor activity in spleens of murineradiation marrow chimeras: possible implications for thegraft-versus-host and graft-versus-leukemia reactivity of donorlymphocyte infusions. Blood 2003; 102: 740–748.

22 Van Wijmeersch B, Sprangers B, Rutgeerts O, Lenaerts C,Landuyt W, Waer M et al. Allogeneic bone marrowtransplantation in models of experimental autoimmuneencephalomyelitis: evidence for a graft-versus-autoimmunityeffect. Biol Blood Marrow Transplant 2007; 13: 627–637.

P-in-F1 lymphoid-tissue GvHD with prolonged immuno-incompetenceB Sprangers et al

595

Bone Marrow Transplantation

23 Rolink AG, Pals ST, Gleichmann E. Allosuppressor andallohelper T cells in acute and chronic graft-vs-host disease.II. F1 recipients carrying mutations at H-2K and/or I-A.J Exp Med 1983; 157: 755–771.

24 Shlomchik WD, Couzens MS, Tang CB, McNiff J, Robert ME,Liu J et al. Prevention of graft versus host disease by inacti-vation of host antigen-presenting cells. Science 1999; 285:412–415.

25 Merad M, Hoffmann P, Ranheim E, Slaymaker S, Manz MG,Lira SA et al. Depletion of host Langerhans cells beforetransplantation of donor alloreactive T cells prevents skingraft-versus-host disease. Nat Med 2004; 10: 510–517.

26 Vossen JM, Heidt PJ, van den BH, Gerritsen EJ, Hermans J,Dooren LJ. Prevention of infection and graft-versus-hostdisease by suppression of intestinal microflora in childrentreated with allogeneic bone marrow transplantation.Eur J Clin Microbiol Infect Dis 1990; 9: 14–23.

27 Beelen DW, Elmaagacli A, Muller KD, Hirche H, SchaeferUW. Influence of intestinal bacterial decontamination usingmetronidazole and ciprofloxacin or ciprofloxacin alone on thedevelopment of acute graft-versus-host disease after marrowtransplantation in patients with hematologic malignancies:final results and long-term follow-up of an open-prospectiverandomized trial. Blood 1999; 93: 3267–3275.

28 Murai M, Yoneyama H, Ezaki T, Suematsu M, Terashima Y,Harada A et al. Peyer’s patch is the essential site in initiatingmurine acute and lethal graft-versus-host reaction.Nat Immunol 2003; 4: 154–160.

29 Bronte V, Zanovello P. Regulation of immune responsesby L-arginine metabolism. Nat Rev Immunol 2005; 5:641–654.

30 Dolcetti L, Marigo I, Mantelli B, Peranzoni E, Zanovello P,Bronte V. Myeloid-derived suppressor cell role in tumor-related inflammation. Cancer Lett 2008; 267: 216–225.

31 Kusmartsev S, Nagaraj S, Gabrilovich DI. Tumor-associatedCD8+ T cell tolerance induced by bone marrow-derivedimmature myeloid cells. J Immunol 2005; 175: 4583–4592.

32 Watanabe S, Deguchi K, Zheng R, Tamai H, Wang LX,Cohen PA et al. Tumor-induced CD11b+Gr-1+ myeloid cellssuppress T cell sensitization in tumor-draining lymph nodes.J Immunol 2008; 181: 3291–3300.

33 Paraiso KH, Ghansah T, Costello A, Engelman RW, KerrWG. Induced SHIP deficiency expands myeloid regulatorycells and abrogates graft-versus-host disease. J Immunol 2007;178: 2893–2900.

34 Wang H, Asavaroengchai W, Yeap BY, Wang MG, Wang S,Sykes M et al. Paradoxical effects of IFN-gamma in graft-versus-host disease reflect promotion of lymphohematopoieticgraft-versus-host reactions and inhibition of epithelial tissueinjury. Blood 2009; 113: 3612–3619.

35 Fowler DH, Breglio J, Nagel G, Eckhaus MA, Gress RE.Allospecific CD8+ Tc1 and Tc2 populations in graft-versus-leukemia effect and graft-versus-host disease. J Immunol 1996;157: 4811–4821.

36 Rozendaal L, Pals ST, Melief CJ, Gleichmann E. Protectionfrom lethal graft-vs-host disease by donor stem cell repopula-tion. Eur J Immunol 1992; 22: 575–579.

37 Ball LM, Lankester AC, Bredius RG, Fibbe WE, van Tol MJ,Egeler RM. Graft dysfunction and delayed immune recon-stitution following haploidentical peripheral blood hemato-poietic stem cell transplantation. Bone Marrow Transplant2005; 35(Suppl 1): S35–S38.

P-in-F1 lymphoid-tissue GvHD with prolonged immuno-incompetenceB Sprangers et al

596

Bone Marrow Transplantation