doi:10.1182/blood-2006-04-017442Prepublished online February 8, 2007;2007 109: 4548-4556   

 Edmund K. WallerMohammad S. Hossain, John D. Roback, Brian P. Pollack, David L. Jaye, Amelia Langston and Chronic GvHD decreases antiviral immune responses in allogeneic BMT

http://bloodjournal.hematologylibrary.org/content/109/10/4548.full.htmlUpdated information and services can be found at:

(1880 articles)Transplantation   � (5019 articles)Immunobiology   �

Articles on similar topics can be found in the following Blood collections

http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#repub_requestsInformation about reproducing this article in parts or in its entirety may be found online at:

http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#reprintsInformation about ordering reprints may be found online at:

http://bloodjournal.hematologylibrary.org/site/subscriptions/index.xhtmlInformation about subscriptions and ASH membership may be found online at:

Copyright 2011 by The American Society of Hematology; all rights reserved.Washington DC 20036.by the American Society of Hematology, 2021 L St, NW, Suite 900, Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly    

For personal use only. by guest on June 2, 2013. bloodjournal.hematologylibrary.orgFrom

TRANSPLANTATION

Chronic GvHD decreases antiviral immune responses in allogeneic BMTMohammad S. Hossain,1 John D. Roback,2 Brian P. Pollack,3 David L. Jaye,2 Amelia Langston,1 and Edmund K. Waller1

1Department of Hematology and Oncology, Division of Stem Cell and Bone Marrow Transplantation, Winship Cancer Institute, Emory University School ofMedicine, Atlanta, GA; 2Department of Pathology and Laboratory Medicine, Emory University School of Medicine, Atlanta, GA; 3Departments of Dermatologyand Pathology, Emory University School of Medicine, Atlanta, GA

Chronic graft-versus-host disease (cGvHD)is associated with functional immunodefi-ciency and an increased risk of opportu-nistic infections in allogeneic bone mar-row transplantation (BMT). We used aparent to F1 model of allogeneic BMT totest the hypothesis that cGvHD leads toimpaired antigen-specific antiviral immu-nity and compared BM transplant recipi-ents with cGvHD to control groups ofallogeneic BM transplant recipients with-out GvHD. Mice with and without cGvHDreceived a nonlethal dose of murine cyto-megalovirus (MCMV) �100 days after

transplantation. Recipients with cGvHDhad more weight loss and higher viralloads in the spleen and liver. MCMV infec-tion led to greater than 25-fold expansionof donor spleen–derived MCMV peptide–specific tetramer-positive CD8� T cells inblood of transplant recipients with andwithout cGvHD, but mice with cGvHD hadfar fewer antigen-specific T cells in periph-eral tissues and secondary lymphoid or-gans. The immunosuppression associ-ated with cGvHD was confirmed byvaccinating transplant recipients with andwithout cGvHD using a recombinant Lis-

teria expressing MCMV early protein (Lm-MCMV). Secondary adoptive transfer oflymphocytes from donor mice with orwithout cGvHD into lymphopenic con-genic recipients showed that cGvHD im-paired tissue-specific homing of antigen-specific T cells. These results indicatethat cGvHD causes an intrinsic immuno-suppression and explain, in part, the func-tional immunodeficiency in allogeneictransplant recipients. (Blood. 2007;109:4548-4556)

© 2007 by The American Society of Hematology

Introduction

Allogeneic bone marrow transplantation (BMT) can cure patientswith relapsed or refractory hematologic malignancies throughestablishment of a functional immune system from a geneticallydissimilar donor with resultant graft-versus-tumor (GvT) effect.Donor T cells are typically responsible for the beneficial GvT effectand protect against opportunistic posttransplantation infections1

but are also responsible for acute and chronic graft-versus-hostdisease (cGvHD).2-5 cGvHD occurs in approximately 60% to 80%of long-term survivors of allogeneic BMT5 and is a major cause ofmorbidity and mortality.6 The development of cGvHD has anegative influence on reconstitution of T-cell function that contrib-utes to severe posttransplantation immunodeficiency.7 The detailedmechanisms responsible for immunodeficiency associated withcGvHD, including the relative paucity of T cells in lymphoidorgans, are still poorly understood.

Previously, several strategies have been described to preventand treat cGvHD complications in allogeneic BMT patients bydepleting or impairing the function of donor T cells.8-10 Thedepletion of T cells from the donor graft reduces the incidence ofboth acute and cGvHD but increases the risk of opportunisticinfection and the rates of posttransplantation relapse. Amongpatients receiving the T-cell depleted (TCD) graft, cytomegalovirus(CMV) pneumonia may develop in up to 50% of BM transplantrecipients with virus reactivation.11 Polymerase chain reaction(PCR)–guided pre-emptive ganciclovir (GCV) therapy reducesviral load and the frequency of CMV disease12 but progressiveCMV disease may develop in patients with severely impairedimmune function. Furthermore, prophylactic and pre-emptive

GCV therapy is associated with myelosuppression, bacterial andfungal infections, as well as late CMV disease.12,13 Therefore, toimprove transplantation outcomes for patients with cGvHD, whoare at risk for posttransplantation CMV and other opportunisticinfections, a better understanding of the relationship among theopportunistic infections, cGvHD, and immunodeficiency is needed.

In this study, we used a murine model of MHC-mismatchedallogeneic BMT coupled with a late posttransplantation CMVinfection to investigate the relationship between cGvHD andantiviral immunity. We established a clinically relevant experimen-tal allogeneic BMT model in which recipients of a sublethal dose ofuntreated donor splenocytes (3 � 106/mouse) developed clinicalsigns compared with control transplant recipients treated with exvivo amotosalen-treated donor splenocytes (10 � 106/mouse) andrecipients of TCD BM alone, which developed no GvHD.14,15 Wefound that recipients with cGvHD had impaired immune responseto murine CMV (MCMV) with higher viral loads, impaired homingof antigen-specific antiviral T cells to target tissues, and fewernumbers of antigen-specific T cells.

Materials and methods

Mice

CB6F1 (C57BL/6 X BALB/c)(H-2b/d, CD45.2/Thy1.2) mice and PepBoy(B6.SJL-PtprcaPep3b/BoyJ)(H-2b, CD45.1/Thy1.2) mice on the C57BL/6background were obtained from Jackson Laboratories (Bar Harbor, ME).BA mice (H-2b, CD45.2/Thy1.1) on the C57BL/6 background were bred atEmory.16 Procedures conformed to National Institutes of Health (NIH)

Submitted April 20, 2006; accepted January 16, 2007. Prepublished online asBlood First Edition Paper, February 8, 2007; DOI 10.1182/blood-2006-04-017442.

The publication costs of this article were defrayed in part by page charge

payment. Therefore, and solely to indicate this fact, this article is herebymarked ‘‘advertisement’’ in accordance with 18 USC section 1734.

© 2007 by The American Society of Hematology

4548 BLOOD, 15 MAY 2007 � VOLUME 109, NUMBER 10

For personal use only. by guest on June 2, 2013. bloodjournal.hematologylibrary.orgFrom

animal care guidelines and were approved by the Emory UniversityInstitutional Animal Care and Use Committee (IACUC).

Preparation of lymphocytes for adoptive immunotherapy

Donor splenocytes from previously MCMV-infected PepBoy mice wereharvested as previously described and enriched by removing plasticadherent cells.14,15 A dose of 3 � 106/0.2 mL/mouse nonadherent untreatedlive nucleated cells was used to induce cGvHD. Control mice with donorsplenocyte–derived T-cell chimerism but lacking cGvHD received 10 � 106

nonadherent donor splenocytes treated with 2 nM amotosalen (S-59psoralen; Cerus, Concord, CA).15

Irradiation, cell transfer, and MCMV infectionof BM transplant recipients

On day �1, CB6F1 mice received a total of 11 Gy irradiation divided into 2doses, 5.5 Gy 3 hours apart.17 The following day, BM was flushed fromfemora and tibiae of naive BA donor mice and CD3� T cells were depleted(TCD) as described.17 TCD BM cells (5 � 106) were transplanted alone orwith 3 � 106 untreated (�GvHD) and 10 � 106 amotosalen-treated (non-GvHD control) donor splenocytes via tail vein injection into irradiatedCB6F1 recipient mice. Recipient mice were infected with 2.5 � 104

plaque-forming units (PFUs) MCMV intraperitoneally or vehicle between100 and 133 days after transplantation (no major differences were observedin the physical and immunologic parameters among the mice with cGvHDwhen assessed at 100 to 150 days after transplantation). The development ofclinical cGvHD was monitored twice weekly by measuring weight loss andclinical signs of hair loss, ruffled fur, diarrhea, and decreased activity.17

According to our IACUC protocol, moribund mice and animals with greaterthan 25% weight loss were euthanized and considered to have died on theday following euthanasia for analysis of posttransplantation survival.

Lymphocytes isolation from organs of BM transplant recipients

Blood samples were collected in tubes containing 50 �L heparin from thetail vein of transplant recipients with and without GvHD on days 0, 3, 10,22, and 40 after MCMV infection, and red blood cells (RBCs) were lysed.18

Splenocytes, liver-cell suspensions, and thymocytes were harvested andsingle-cell suspensions were prepared from transplant recipients as previ-ously described.18,19

Determination of viral load per liver and spleen

Liver and spleen viral loads were determined as described by Hossain etal.20 Briefly, liver and spleen were aseptically collected from the MCMV-infected recipients, homogenized, and centrifuged. Serial diluted superna-tants were added on a 3T3 confluent monolayer of 24 tissue culture platesand incubated for 90 minutes at 37°C overlayered with 1 mL 2.5%methylcellulose in DMEM. MCMV PFUs were directly counted after 4days under a light microscope (Nikon, Melville, New York) after stainingthe 3T3 confluent monolayer with methylene blue.

Flow cytometry

The origin of CD8� and CD4� T cells in BM transplant recipients wasdetermined by staining with monoclonal antibodies (mAbs) specific fordonor BM (Thy1.1�, CD45.2�, Thy1.2�), donor spleen (Thy1.2�, CD45.2�,CD45.1�), or host T cells (CD45.2�, Thy1.2�) in combination with mAbsto CD8 and CD4 and other surface activation markers (Pharmingen, SanJose, CA) as previously described.17 Anti-MCMV–specific CD8� T cellswere counted using an APC-conjugated HGIRNASFI-H-2Db tetramer(NIAID Tetramer Core Facility, Atlanta, GA).21 Intracellular IFN-�– andTNF-�–producing CD8� T cells were determined by using a Cytofix/Cytoperm kit purchased from BD Pharmingen (San Jose, CA). The cellswere acquired by FACS Aria (Becton Dickinson, San Jose, CA) andanalyzed by FlowJo software (Ashland, OR).

Assessment of cGvHD

Transplant recipients were monitored for weight loss, hair loss, eczematoidrash and dry skin in tail and ears, and motor activity twice weekly for the

first 30 days, and then weekly. Additionally, cGvHD was scored throughhistologic analysis of tissue sections as described.17,22 Three to 5 mice fromeach treatment group were killed at predetermined time points and skin,liver, and intestine were collected in formalin. Chronic GvHD scores wereassigned by trained pathologists unaware of the treatment group (B.P.P. andD.L.J.) using previously reported criteria17,22 and � 40 microscopy offormalin-fixed histologic tissue sections. The microscope was a NikonEclipse E400, with a 20�/0.75 Plan Apo objective lens. The camera was aSPOT Insight 3.2.0, and we acquired images with SPOT Advanced 4.1software. The slides were stained with hematoxylin and eosin, and noimaging medium or solution was used. Scores (0-5) were based on thepresence of lymphocyte infiltration in the liver periportal region, thepresence of apoptotic cells in the intestinal epithelium along with foci ofepithelial ulceration, and the presence of sclerodermoid sclerosis or fibrosisin the skin. The results for each mouse were tabulated and averaged by thefirst author (M.S.S.).

Assessment of the ability of antiviral T cells to home to sitesof MCMV infection

Donor lymphocytes were harvested, using Lympholyte-Mammal (Cedar-lane, Hornby, ON, Canada), from the pooled blood of 2 groups of 5 micewith or without cGvHD: either CB6F1 mice that had previously receivedtransplants of 3 � 106 allogeneic Thy1.2�CD45.1� C57BL/6 splenocytesand 5 � 106 TCD BM (cGvHD� lymphocytes) or C57BL/6 mice that hadpreviously received transplants of 10 � 106 Thy1.2� CD45.1� congenicdonor splenocytes and 5 � 106 TCD BM (cGvHD� lymphocytes). Bothgroups of lymphocyte donors were infected with 2.5 � 104 PFUs MCMVintraperitoneally 10 days before the lymphocytes were collected (day 100after transplantation). Secondary recipients (congenic Thy1.2� CD45.2�

C57BL/6 mice) were irradiated and received transplants of 5 � 106 TCDBM syngenic BM cells in combination with 3 � 106 congenic bloodlymphocytes from either cGvHD� or cGvHD� donors. The secondaryrecipients were then infected with 3 � 103 PFUs MCMV intraperitoneallythe following day. Nucleated cells from peripheral-blood mononuclear cells(PBMCs) and spleen and liver were harvested from the secondary recipientson 3 and 10 days after viral infection, as described previously.19 T cellsderived from the transferred cGvHD� or cGvHD� lymphocytes (Thy1.2�

CD45.2�) and donor lymphocyte–derived MCMV peptide–specific tetramer-positive CD8� T cells per mL blood, gram of liver, or spleen were countedusing flow cytometry.19

Immunization with an MCMV vaccine

Mice were vaccinated intraperitoneally with 1 � 106 CFUs Lm-MCMV, aListeria monocytogenesis that has been rendered nonpathogenic byknockout of bacterial genes associated with virulence23 and engineeredto express the MCMV H-2b–immunodominant peptide HGIRNASFI.21

The vaccine is prepared and supplied by Cerus. Nonvaccinated micewere killed for baseline data; vaccinated mice were killed 7 days aftervaccination. Lymphocytes were harvested from blood, liver, and spleen,and vaccine-induced MCMV-specific CD8� T cells were assayed usingMCMV-HGIRNASFI peptide–specific tetramer and fluorescence-activated cell sorter (FACS) analysis.

Statistical analyses

Student t test was used to compare mean values for parameters ofmice. Differences were considered significant when P values below .05were obtained.

Results

The effects of MCMV infection on transplant recipientswith cGvHD

Our overall goal was to investigate the mechanism of immunedeficiency observed in allogeneic BM transplant recipients with

CHRONIC GvHD AND ANTIVIRAL IMMUNITY 4549BLOOD, 15 MAY 2007 � VOLUME 109, NUMBER 10

For personal use only. by guest on June 2, 2013. bloodjournal.hematologylibrary.orgFrom

cGvHD in the context of CMV infection. We established anexperimental mouse allogeneic BMT model of parent to F1transplant recipients in which lethally irradiated CB6F1 recipientsreceived transplants of 3 � 106 allogeneic splenocytes from MCMV-immunized donors along with 5 � 106 TCD BM. These micedeveloped cGvHD that has a similar clinical presentation to thatwhich develops in BMT patients more than 100 days aftertransplantation. Greater than 80% of mice with this form of cGvHDtypically survived for more than 100 days after transplantation;mice with severe cGvHD were euthanized per IACUC guide-lines.14,15 While the severity of cGvHD in mice surviving to day�100 was generally not life threatening, these mice nonethelesshad developed hair loss and eczematoid rash and dry skin in tail andears with decreased general activity compared with control recipi-ents of amotosalen-treated splenocytes or TCD BM alone that didnot develop GvHD. Mice with cGvHD surviving to day �100 hada mean weight of 20.3 � 3.1 g (baseline 27.5 � 2.1 g) comparedwith recipients of TCD BM alone (day �100 weight 26.6 � 1.2 g;baseline weight 28.1 � 1.2 g) and control mice without GvHD thathad received amotosalen-treated donor splenocytes (day �100weight 30.0 � 1.7 g; baseline weight 29.6 � 1.9 g). These differ-ences were highly significant, comparing the day �100 weightsbetween mice with cGvHD to recipients of TCD BM alone or BMplus treated splenocytes (P � .001).

To investigate the antiviral immune responses in recipients withor without cGvHD, mice were infected with a sublethal dose

(5 � 104 PFU/mouse) of MCMV between day 100 and day 133after transplantation. While all recipients challenged with this doseof MCMV survived until 40 days after infection, recipients withcGvHD suffered greater weight loss than the recipients withoutcGvHD (Figure 1A). To test the effects of CMV infection on thepathophysiology of cGvHD, transplant recipients with cGvHDwere killed on the day of MCMV infection and recipients withoutcGvHD were killed 10 days later. The sum of average cGvHDscores obtained from all 3 organs of recipients with cGvHD wassignificantly higher (P � .05) in both uninfected and MCMV-infected mice compared with the recipients without cGvHD(Figure 1B). The presence of sclerodermoid sclerosis and fibrosiswas found only in the skin of recipients with cGvHD and not inrecipients without cGvHD (Figure 1C). Comparing cGvHD histo-logic scores before and 10 days after MCMV infection, there wasno significant exacerbation of cGvHD induced by MCMV infec-tion (Figure 1B).

To determine whether the presence of cGvHD interfered withantiviral immunity, we next determined the viral loads in the liverand spleen of transplant recipients following intraperitoneal infec-tion with a lethal dose of MCMV. As reported previously, viralloads in the liver and spleen peak 3 days after MCMV infection innormal C57BL/6 mice20 and in the irradiated recipients of alloge-neic BM transplants (our unpublished data, October 2004). Tissuehomogenates of liver and spleen were prepared on day 3 afterinfection. Viral loads were 15-fold higher in the spleen and 13-fold

Figure 1. Evidence of cGvHD and antiviral immune response in recipients with untreated splenocytes. Irradiated CB6F1 recipients of 3 � 106 untreated donorsplenocytes (cGvHD), 10 � 106 amotosalen-treated donor splenocytes (no-GvHD–treated splenocyte recipients), or TCD BM alone (no GvHD) were infected with MCMV(5 � 104 PFU intraperitoneally) on 100� days after transplantation. (A) Percentage weight loss was calculated from the initial weight measured on day 0 of transplantation.Here, day 0 represents the percentage weight loss measured on day 133 after transplantation (�, cGvHD, splenocytes recipients; f, no GvHD; and Œ, TCD BM). (B)Recipients with cGvHD, no-GvHD–treated splenocyte recipients, and TCD BM were killed on day 100 after transplantation (day 0 after MCMV infection) and day 10 after5 � 104 PFU MCMV infection. Skin, liver, and small intestine were collected in formalin solution and cGvHD scores were assessed from the histologic tissue sections of skin,liver, and small intestines. The average mean value and standard deviation of cGvHD scores obtained from the 3 organs of recipients with cGvHD (�), no GvHD (f), and TCDBM (Œ) on days 0 and 10 after MCMV infection. (C) The presence of sclerodermoid sclerosis and fibrosis (yellow arrows) is shown in the histologic slides of skin tissue sectionscollected from the recipients with cGvHD, no GvHD, and TCD BM under uninfected conditions (day 100 after transplantation) and day 10 after MCMV infection (day 110 aftertransplantation). The data are representative of 1 of 2 similar experiments and 4 to 5 mice were used per group in each time point. (D) Recipients with cGvHD, no GvHD, andTCD BM alone were infected with MCMV (5 � 104 PFU intraperitoneally) at 100 days after transplantation. Liver and spleen were aseptically collected on day 3 after infectionand viral load per organs was determined by spreading the tissue homogenates on a 3T3-cell confluent monolayer as described in “Materials and methods.” Mean viral PFUand standard deviations obtained from each of the infected organs of 5 to 6 mice per group were used (*P � .001, Student t test).

4550 HOSSAIN et al BLOOD, 15 MAY 2007 � VOLUME 109, NUMBER 10

For personal use only. by guest on June 2, 2013. bloodjournal.hematologylibrary.orgFrom

higher in the liver among recipients with cGvHD compared withrecipients of treated splenocytes without cGvHD (Figure 1D;P � .001). We also found significantly higher viral loads in thelungs of recipients with cGvHD compared with recipients withoutGvHD 10 days after MCMV infection (data not shown). Basedupon the increased viral loads among mice with cGvHD, we thenexplored potential mechanisms for the relative lack of in vivoantiviral immunity including (1) the lack of antigen-specific Tcells; (2) the lack of organ-specific T-cell expansions after MCMVinfection; (3) the lack of functional antiviral T cells; and (4) thelack of tissue-specific homing of T cells.

Delayed MCMV infection resulted in extensive peripheralexpansion of donor T cells in the blood of recipientswith cGvHD

We tested whether increased MCMV viral loads in mice withcGvHD were due to lack of peripheral expansion of antigen-specific donor T cells. Following MCMV infection, antiviral T cellsin the blood of mice were enumerated using MCMV peptide–specific class I H-2b tetramer. Surprisingly, more than 50-foldexpansion of lymphocyte-gated donor spleen–derived T cellsoccurred in the blood of recipients with cGvHD and over 25-foldexpansion occurred in mice without GvHD that received trans-plants 10 days after MCMV infection followed by a contractionphase during which T-cell populations in blood returned to baselinelevels by day 40 after infection (Figure 2A). Accordingly, theabsolute number of donor spleen–derived T cells per mL of bloodincreased significantly (P � .005) 10 days after MCMV infection,in recipients both with and without cGvHD compared with theuninfected mice. Of note, the number of donor spleen–derived Tcells in blood of mice with cGvHD was significantly higher(P � .05) compared with recipients without cGvHD 10 days afterinfection (Figure 2B). However, in both types of recipient, donorT-cell levels returned to similar baselines by day 40 after MCMVinfection (Figure 2B).

Donor spleen– and BM–derived T cells in the bloodof recipients with and without cGvHD had differentialexpression of CD44 surface markers

Previous reports claimed that reactivity of T cells depends on thesurface expression of CD62L24 and CD44.25 As most of theantiviral effector T cells undergo apoptosis during the contractionphase after 10 days after infection, the persistence of activated(effector or allo-reactive) T cells and MCMV-specific memory Tcells was measured at day 40 after MCMV infection based upon theexpression of CD62L and CD44 surface antigens on donor spleen–and BM-derived CD4� and CD8� T cells. The expression ofCD62L at baseline and at day 40 after infection was similar amongCD4� and CD8� T cells of either donor spleen or donor BM originsin recipients with and without GvHD (data not shown). Inrecipients with cGvHD, few donor spleen–derived and a majorityof the BM-derived CD8� and CD4� T cells were CD44lo on day 0of viral infection (133 days after transplantation), whereas donorspleen–derived CD8� and CD4� T cells in transplant recipientswithout GvHD were almost all CD44hi. Unlike the donor spleen–derived T cells, donor BM–derived CD8� and CD4� T cells weremostly CD44med, a pattern consistent with the naive T-cell pheno-types in recipients both with and without cGvHD (Figure 2C toppanels). At 40 days following MCMV infection, following thecontraction phase of the immune response, all donor spleen–derived CD8� and CD4� T cells were CD44hi (memory phenotype)in recipients both with and without cGvHD. Of note, recipientswith cGvHD had a fraction of donor BM–derived CD44lo CD8� Tcells, a type of T cell that may be responsible for cGvHD (Figure2C, bottom panels).

Anti-CMV tetramer-positive CD8� T cells increased toequivalent degrees in the blood of recipientswith and without cGvHD

To investigate overall T-cell immune reconstitution and antiviral Tcells in the peripheral blood of BM transplant recipients with

Figure 2. Extensive expansion of periph-eral donor T cells in recipients with cGvHDafter viral infection. Recipients with cGvHDand no GvHD were bled on 133 days aftertransplantation (day 0 prior to infection) andmice were infected with 5 � 104 PFUsMCMV intraperitoneally. Panel A representsthe kinetics of peripheral expansion of donorspleen–derived (Thy1.2� CD45.1�), host-derived (Thy1.2� CD45.2�) T cells, as wellas a mixture of BM-derived (Thy1.1� CD45.2�)T cells and host-derived non-T cells. The FACSdata shown here represent the same mouse ineach group that is representative of 4 to 5 micein each group. Panel B represents the totalnumber of donor spleen–derived T cells per mLof blood. Data from an average of 4 to 5 mice pergroup are shown for each time point (*P � .001for uninfected data vs infected data; *P � .05 forcGvHD vs no GvHD on day 10 after infection).Panel C represents the surface expression ofCD44 antigen on the donor spleen– and BM-derived CD8� and CD4� T cells that weremeasured prior to MCMV infection (133 daysafter transplantation) as shown in top 2 panelsand 40 days after MCMV infection as shown inbottom 2 panels from the same mouse of eachgroup. Dotted histogram plot represents theisotype control mAb used for CD44. The dataare representative from 1 mouse of each groupout of 4 to 5 mice used per group.

CHRONIC GvHD AND ANTIVIRAL IMMUNITY 4551BLOOD, 15 MAY 2007 � VOLUME 109, NUMBER 10

For personal use only. by guest on June 2, 2013. bloodjournal.hematologylibrary.orgFrom

cGvHD, we next determined the absolute numbers of CD4� T andCD8� T cells as well as MCMV peptide–specific tetramer-positiveCD8� T cells in the blood following MCMV infection. Interest-ingly, CD8� T cells of donor origin significantly expanded inresponse to MCMV infection in recipients both with and withoutcGvHD, but no differences were found in the number of host-derived T cells or donor CD4� T cells between the infected anduninfected recipients (Table 1). Similarly, the absolute number ofdonor spleen– and BM-derived MCMV peptide–specific tetramer-positive CD8� T cells per mL of blood was significantly increasedin recipients both with and without cGvHD 10 days after MCMVinfection compared with the uninfected mice. Moreover, expansionof donor spleen–derived MCMV peptide–specific tetramer-positive CD8� T cells predominated over BM-derived antiviral Tcells in recipients with cGvHD, whereas expansion of donorBM–derived MCMV peptide–specific tetramer-positive CD8� Tcells predominated over donor spleen–derived tetramer-positive Tcells in recipients without cGvHD (Table 1). Thus, MCMVinfection resulted in significant expansion of CD8� T cells as wellas MCMV peptide tetramer–specific CD8� T cells in the blood oftransplant recipients with and without cGvHD, whereas there waslimited expansion of the CD4� T cells.

Atrophy of primary and secondary lymphoid organs in cGvHD

The spleens of mice with cGvHD (day �100 after transplantation)were atrophic, and splenocytes isolated from these organs failed toproliferate in response to MCMV infection in contrast to briskproliferation of splenic T cells in response to MCMV infection inrecipients without cGvHD (Figure 3A). The thymus was atrophicin recipients with cGvHD, and the number of thymocytes de-creased significantly in all mice after MCMV infection comparedwith uninfected mice (Figure 3B). On the other hand, no differ-ences were found in the number of lymphocytes isolated from theliver of recipients with and without cGvHD both before and afterMCMV infection (Figure 3C). Accordingly, the numbers of donorspleen– and BM-derived T-cell populations in the spleen andthymus of recipients with cGvHD were significantly less than in the

recipients without GvHD (P � .001; Figure 3D-E,G-H). Of note,donor spleen– and BM-derived T cells were significantly increasedin the spleen and decreased in the thymus on day 10 after MCMVinfection in recipients without GvHD compared with the uninfected

Table 1. T-cell populations in the blood of recipients with cGvHD, no GvHD, and TCD BM

Cell type

Days after MCMV infection

cGVHD No cGVHD TCD BM

0 10 0 10 0 10

T cells*

Donor spleen

CD8 0.1 � 0.2 12.4 � 0.1‡ 0.1 � 0.01 2.8 � 1.6‡ — —

CD4 0.03 � 0.0 0.2 � 0.1 0.01 � 0.02 0.3 � 0.02 — —

Donor BM

CD8 0.3 � 0.1 1.7 � 0.8‡ 0.3 � 0.1 2.7 � 1.4‡ 0.2 � 0.1 4.2 � 2.1‡

CD4 0.5 � 0.2 0.7 � 0.3 0.7 � 0.2 0.6 � 0.2 0.5 � 0.2 0.7 � 0.1

Host

CD8 0.003 � 0.004 0.03 � 0.02 0.05 � 0.04 0.3 � 0.4 0.08 � 0.05 0.7 � 0.3

CD4 0.002 � 0.003 0.01 � 0.003 0.08 � 0.07 0.06 � 0.05 0.1 � 0.07 0.1 � 0.03

MCMV tetramer-positive CD8� T cells†

Donor spleen 0.1 � 0.2 38.6 � 33.1§ 0.3 � 0.4 9.3 � 5.5§

Donor BM � 0.1 13.6 � 8.5§ � 0.1 32.1 � 9.2§ 9.6 � 7.3 50.7 � 28.5§

Host � 0.1 6.2 � 6.2 � 0.1 11.3 � 4.5§ � 0.1 9.8 � 6.0§

— indicates not applicable.*Recipients of untreated DLI (cGvHD), amotosalen-treated (no GvHD), and TCD BM (TCD BM only) were bled on 133 days after transplantation (day 0), infected with 2.5 �

104 PFUs MCMV, and bled on day 10 after infection. The CD4� and CD8� T cells (� 106) per mL of blood were measured by FACS analysis.†The absolute number of MCMV peptide-specific tetramer-positive CD8� T cells (� 103) per mL of blood was measured by FACS analysis.‡P � .05 (Student t test) compared with uninfected condition. Four to 5 mice were used per group. The data represent 1 of 2 similar experiments.§P � .05 (Student t test) compared with the uninfected conditions.

Figure 3. Splenic and thymic functions were severely suppressed in recipientswith cGvHD. Spleen, thymus, and liver were collected from the recipients with andwithout cGvHD and TCD BM alone prior to MCMV infection (day 100 aftertransplantation) and day 10 after MCMV infection (day 110 after transplantation).Lymphocytes were isolated and absolute numbers of nucleated cells per organ weremeasured by counting the live cells under fluorescence microscopy. Panels A, B, andC represent the absolute number of nucleated cells per spleen, thymus, and liver,respectively (p, recipients with cGvHD; f, recipients with no GvHD; and �, recipientsof TCD BM). Panels D, E, and F represent the absolute number of donor spleen Tcells per spleen, thymus, and liver, respectively (p, recipients with cGvHD; f,recipients with no GvHD). Panels G, H, and I represent the absolute number of donorBM–derived T cells per spleen, thymus, and liver, respectively (p, mice with cGvHD;f, recipients with no GvHD) that were calculated from the data obtained from theFACS analysis. Data represent an average of 5 to 6 mice were used per group(*P � .05 [Student t test] is for the significant decrease compared with the cellnumbers in the spleen and thymus of recipients with no GvHD and TCD BM alone).

4552 HOSSAIN et al BLOOD, 15 MAY 2007 � VOLUME 109, NUMBER 10

For personal use only. by guest on June 2, 2013. bloodjournal.hematologylibrary.orgFrom

mice (Figure 3D-E,G-H). Again, there were no differences in thenumbers of donor spleen– and BM-derived T cells in liversbetween the recipients with and without cGvHD in both MCMV-infected and uninfected mice (Figure 3F,I).

Mice with cGvHD lack functional antiviral T cells in the spleenand liver

MCMV peptide–specific tetramer-positive donor antiviral CD8� Tcells increased considerably in blood 11 days after MCMVinfection in recipients with cGvHD, but they failed to control viralreplication in the liver and spleen (Table 1; Figure 1D). Toinvestigate the possible causes of the failure of antiviral activities inthese target organs, we estimated organ-specific antiviral T cellsusing tetramer staining on day 10 after viral infection. Significantlylower frequencies of both donor spleen– and BM-derived MCMV-specific antiviral CD8� T cells were found in the spleen and liver ofrecipients with cGvHD compared with the recipients withoutcGvHD (Figure 4A,D).

To investigate the functional status of anti-CMV T cells, westimulated lymphocytes harvested from the spleen and liver in vitrowith the HGIRNASFI MCMV peptide21 (same peptide used as intetramer) for 4 hours in presence of Brefeldin A and measured thefrequency of intracellular IFN-�– and TNF-�–producing CD8� Tcells by multicolor FACS analysis. Significantly lower levels ofboth donor spleen– and BM-derived IFN-�� and TNF-�� CD8� Tcells were detected in the spleen and liver of recipients withcGvHD compared with recipients without GvHD (Figure 4B,E,and 4C,F). The reduced antiviral T-cell response in recipients withcGvHD could be due to an immunosuppressive effect of higherviral loads in the infected organs or the results of cGvHD itself. Toclarify this issue we used a nonpathogenic, highly immunogeneicrecombinant Listeria vaccine (Lm-MCMV) that expresses the earlyMCMV peptide to immunize allogeneic transplant recipients withcGvHD, syngenic transplant recipients with no GvHD, and normal

C57BL/6 mice. Vaccinated mice had normal weight gain andactivity, with no signs of illness. We measured the frequency ofantiviral T cells in the blood, liver, and spleen on 0 and 7 days aftervaccination. Significantly lower frequencies of MCMV peptide–specific tetramer-positive CD8� T cells were found on day 7 afterinoculation in organs collected from mice with cGvHD comparedwith syngenic transplant recipients with no GvHD or controlnormal mice (Figure 5A). The absolute numbers of antiviral T cellswere comparable between syngenic recipients and control micethat did not receive transplants and were markedly reduced in the inblood (P � .05), liver (P � .05), and spleen (P � .001) of micewith cGvHD (Figure 5B).

T cells from mice with cGvHD lack homing ability

Following MCMV infection, the number of antigen-specific antivi-ral CD8� T cells in the liver of mice with cGvHD was markedlyreduced compared with the recipients without GvHD, althoughthere were similar numbers of tetramer-positive CD8� T cells in theblood (Table 1). The lack of antiviral T cells in specific tissuescould be due to the defects of T-cell homing or a failure ofantigen-specific T cells to expand in infected tissues in mice withcGvHD. To address this issue, we next determined the ability ofantiviral CD8� T cells to home to the spleen and liver bytransferring antigen-specific T cells from donor mice with orwithout cGvHD to syngenic recipients without cGvHD. Threemillion blood lymphocytes were prepared from cGvHD� CB6F1and cGvHD� C57BL/6 mice containing 9210 and 25 900 CMV

Figure 4. Liver and spleen antiviral T-cell responses were significantly de-creased in recipients with cGvHD. Lymphocytes harvested from the spleen andliver of recipients with and without cGvHD were stained with MCMV peptide–specificMHC class I tetramer, and tetramer-positive CD8� T cells were measured by FACSanalysis. Panels A and D represent the absolute number of donor spleen– (f), donorBM– (�), and host-derived (p) MCMV peptide–specific (HGIRNASFI-H2b) tetramer-positive CD8� T cells detected from the spleen and liver, respectively. Panels B, Cand E, F represent the total frequency of donor spleen– and donor BM–derivedintracellular IFN-�– and TNF-�–producing CD8� T cells per spleen and liver,respectively, on day 10 after MCMV infection as measured after 4 hours in vitrostimulation of total lymphocytes with HGIRNASFI MCMV peptide (�, cGvHD; f, noGvHD). *P � .05 and **P � .01 (Student t test) are for the decreased absolutenumbers of MCMV peptide–specific tetramer-positive IFN-�� and TNF-�� CD8� Tcells per organ in recipients with cGvHD compared with no GvHD.

Figure 5. Immunosuppression associated with cGvHD responsible for thedeficit antiviral T-cell responses. Allogeneic transplant recipient with cGvHD,syngenic transplant with no GvHD, and normal C57BL/6 mice were intraperitoneallyvaccinated with 1 � 106 CFUs Lm-MCMV. Lymphocytes were harvested from theblood, liver, and spleen collected on days 0 and 7 after vaccination, and MCMVpeptide–specific tetramer-positive CD8� T cells were measured by FACS analysis.(A) Representative dot plots of donor spleen–derived tetramer-positive CD8� T cellsin the blood, liver, and spleen of recipients with cGvHD and no GvHD andCD3�-gated populations of normal C57BL/6 mice. (B) Absolute number of tetramer-positive CD8� T cells per organ measured from the recipients with cGvHD, no GvHD,and normal C57BL/6 mice. *P � .05 in blood and liver and *P � .001 in spleen(Student t test) are for the decreased absolute numbers of MCMV peptide–specifictetramer-positive CD8� T cells per organ in recipients with cGvHD compared withwithout GvHD or normal mice. Three to 5 mice were used per group in each timepoint.

CHRONIC GvHD AND ANTIVIRAL IMMUNITY 4553BLOOD, 15 MAY 2007 � VOLUME 109, NUMBER 10

For personal use only. by guest on June 2, 2013. bloodjournal.hematologylibrary.orgFrom

tetramer–positive CD8� T cells, respectively. The tetramer-positiveCD8� T cells were derived, in both cases, from previouslytransplanted Thy1.2� CD45.1� immune splenocytes. To induceperipheral lymphopenia and facilitate expansion of adoptivelytransferred T cells, Thy 1.2� CD45.2� secondary recipients wereirradiated and received TCD BM. To study the homing ability andfunctional antiviral activity of the adoptively transferred MCMVtetramer–positive CD8� T cells, we infected the secondary recipi-ents with MCMV the day following secondary transfer. Surpris-ingly, while equivalent numbers of Thy 1.2� CD45.2� tetramer-positive CD8� T cells were present in the blood of secondarytransplant recipients on day 3 after transplantation, greater than20-fold fewer tetramer-positive CD8� T cells had homed to theliver and greater than 10-fold fewer tetramer-positive CD8� T cellshad homed to spleen in mice that received PBMCs obtained fromthe cGvHD� donor mice compared with recipients of PBMCs fromcGvHD� donor mice (Figure 6A). Accounting for the 2-fold

difference in the absolute numbers of antigen-specific T cellstransferred from cGvHD� and cGvHD� donors, there still weresignificantly more antigen-specific T cells in peripheral tissuesamong secondary recipients in the former group (10-fold differ-ence in the liver, and 5-fold difference in the spleen). Peripheralexpansion of antigen-specific T cells from cGvHD� donors did notappear to be significantly impaired, as the difference in absolutenumbers of antigen-specific T cells in the liver and spleen ofrecipients of cGvHD� donor lymphocytes versus cGvHD� lympho-cytes on day 10 after transfer had decreased to 3-fold and 7-fold,respectively (Figure 6B).

Discussion

In this study, we have investigated cGvHD-induced immunodefi-ciency using a posttransplantation CMV infection in a murine model ofallogeneic BMT. Mice that developed cGvHD after transplantation with3 � 106 untreated allogeneic donor splenocytes were compared withcontrol allogeneic transplant recipients lacking cGvHD or mice thatreceived TCD BM alone.14,15 We challenged BM transplant recipientswith a sublethal dose of MCMV on days 100 to 133 aftertransplantation in order to study immune responses to a clinicallyrelevant opportunistic infection. We found several significantdifferences in the antiviral T-cell responses between recipients withand without cGvHD. First, although both groups showed signifi-cant increases in the number of donor spleen–derived T cells in theblood 10 days after MCMV infection, mice with cGvHD failed tocontrol viral replication in the spleen and liver. Secondly, splenicand thymic T-cell numbers were significantly reduced in mice withcGvHD compared with mice without cGvHD. Thirdly, signifi-cantly lower numbers of MCMV peptide–specific tetramer-positiveantiviral CD8� T cells were found in spleen and liver on day 10after infection in recipients with cGvHD compared with micewithout cGvHD. Vaccination with Lm-MCMV confirmed the lackof immune response in mice with cGvHD, and secondary transferexperiments documented a defect in homing and tissue expansionof lymphocytes from mice with cGvHD (Figure 6).

The mechanisms of cGvHD-induced immunosuppression re-main to be fully defined but could include deficits in antigenpresentation, homing, or expansion of antigen-specific donor Tcells in MCMV-infected tissue. The results of this study suggestthat cGvHD impairs primary immune responses as well as homingof T cells to lymphoid organs and sites of infection. Significantlyhigher viral loads on day 3 after MCMV infection in the livers andspleens of recipients with cGvHD compared with mice withoutcGvHD (Figure 1D) suggest a lack of early innate antiviral immuneresponses in recipients with cGvHD. Significantly lower levels ofMCMV peptide–specific tetramer-positive IFN-�� and TNF-��

CD8� T cells in the spleen and liver on 10 days after viral infectionin mice with cGvHD suggest the lack of adaptive, antigen-specificantiviral T-cell responses compared with mice without GvHD(Figure 4). In recipients with cGvHD, higher numbers of MCMVpeptide–specific tetramer-positive CD8� T cells were found in theblood on day 10 after infection with MCMV (Figure 2; Table 1), butthe migration of these T cells to the liver and spleen were significantlyimpaired compared with the recipients without cGvHD (Figure 4). Thelack of antiviral T cells in infected tissues may be due to defects inhoming or chemokine receptor expression in antigen-specific T cellsin the presence of cGvHD. Notably, the relative decrease in CD44expression on T cells from mice with cGvHD may impair therolling phase of lymphocyte adhesion26,27 and interfere with

Figure 6. Homing defects of donor T cells to sites of CMV infection. Bloodlymphocytes were harvested from 2 groups of donors: cGvHD� CB6F1 mice thatpreviously received transplants of C57Bl/6 donor cells, and cGvHD� C57BL/6recipients of syngenic transplants. Both groups of recipients were infected with2.5 � 104 PFUs MCMV intraperitoneally on day 100 after transplantation, andPBMCs were harvested from the pooled blood of 5 infected mice on day 10 after viralinfection as described in “Materials and methods.” Three million PBMCs from micewith or without cGvHD were adoptively transferred to 2 groups of irradiated (11 Gy)C57BL/6 mice that were then infected with 3 � 103 PFUs MCMV intraperitoneally thenext day. Nucleated cells were harvested on days 3 and 10 after infection from theblood and spleen (as described in “Materials and methods”) from the infectedsecondary recipients. The numbers of donor-derived lymphocytes per gram of livertissue are shown. Donor-derived T cells and MCMV peptide–specific tetramer-positive CD8� T cells per mL blood or gram liver or spleen were counted by usingmulticolor FACS analysis. (A) MCMV peptide–specific tetramer-positive CD8� T cellsper organ measured from the recipients of lymphocytes harvested from cGvHD�

donors (�) and from recipients of lymphocytes harvested from no-GvHD donors (f)on day 3 after MCMV infection. (B) MCMV peptide–specific tetramer-positive CD8� Tcells per organ measured on day 10 after MCMV infection. Values represent anaverage of data from 4 to 5 mice per group (*P � .05 and **P � .005; Student t test).(C) Schematic diagram for the inhibitory effects of cGvHD on the migration andhoming of donor T cells to different lymphoid and nonlymphoid organs. #cGvHDinterferes with homing of T-cell progenitors to the thymus. ##cGvHD interferes withhoming of donor antiviral T cells from blood to lymphoid organs. ###cGvHD interfereswith homing of antigen-specific T cells to target organs infected by CMV (eg, lung andliver).

4554 HOSSAIN et al BLOOD, 15 MAY 2007 � VOLUME 109, NUMBER 10

For personal use only. by guest on June 2, 2013. bloodjournal.hematologylibrary.orgFrom

lymphocyte–endothelial cell (EC) adhesion.28 The lower levels ofCD44 expression may also impair signaling through tyrosinekinases, indirectly inhibiting homing and activation.29

Since mice with cGvHD had higher viral loads followingMCMV infection, it was possible that the presence of high levels ofMCMV in the blood and tissues had a direct or indirect inhibitoryeffect on primary immune responses and/or T-cell homing to targettissues of MCMV. Alternatively, the lower antiviral T-cell responsein recipients with cGvHD could be due to an intrinsic immunosup-pression associated with the presence of cGvHD. We specificallyaddressed this issue by vaccinating mice with and without cGvHDusing a nonviral Lm-MCMV vaccine. Marked anti-MCMV im-mune responses were observed 7 days after vaccination in normalmice and in syngenic transplant recipients without GvHD, whereastetramer-positive antiviral T cells were nearly undetectable in theblood and organs of mice with cGvHD (Figure 5). Thus, weconclude that the deficit of antiviral T cells in recipients withcGvHD is due, in large part, to the immunosuppressive effects ofcGvHD rather than an effect of the virus.

We specifically tested the effect of cGvHD on homing of T cellsby transferring blood lymphocytes from MCMV-infected donorswith and without cGvHD to irradiated naive recipient C57BL/6mice. Although 3 � 106 PBMCs obtained from cGvHD� donorscontained about half of the amount of donor spleen–derived T cellsthan the 3 � 106 PBMCs obtained from the GvHD� donors, therewas an average of more than 20-fold more donor spleen–derivedtetramer-positive antiviral CD8� T cells in the liver and more than10-fold more tetramer-positive T cells in the spleen of secondaryrecipients that received cGvHD� lymphocytes compared withsecondary transplant recipients that received cGvHD� lympho-cytes (Figure 6A; P � .005). The expansion of antigen-specificCD8� T cells in the liver and spleen, between day 3 and day 10,was similar among recipients of cGvHD� and cGvHD� lympho-cytes, although the absolute numbers of antigen-specific T cellswas higher in the latter group (Figure 6B; P � .05). Taken together,these data suggest that cGvHD acts at multiple levels to impairantigen-specific immunity including (a) an effect on the generationof new T cells from BM progenitors through the thymus; and (b)retarding the homing of antiviral T cells to the secondary lymphoidorgans and target tissues infected with MCMV (Figure 6C).

This study may shed some light on the relationship betweencGvHD and CMV infections in clinical BMT. While cGvHD andCMV infections are often seen together in allogeneic BM trans-plant recipients, the causal relationship between these 2 conditionsis not clear. Soderberg et al30 reported that CMV infection intransplantation patients was associated with the development ofautoantibodies to CD13, a marker expressed on granulocytes,monocytes, fibroblasts, endothelial cells, epithelial cells, smoothmuscle cells, and certain neuronal cells. Through analysis of theimmunoreactivity of serum samples on skin biopsies they showedthat most patients with CMV disease and cGvHD also had

CD13-specific antibodies that may contribute to the cutaneousmanifestations of cGvHD as well as CMV-associated granulocyto-penia in allogeneic BM transplant recipients.30 While Via et al31

found that MCMV infection induced synergistic effects on acuteGvHD, previous reports using experimental murine BMT modelsystems have not described the effects of MCMV infection oncGvHD. The current study used a clinico-pathologic scoringsystem for cGvHD and found that MCMV infection did notexacerbate existing cGvHD when mice received sublethal injec-tions of virus (Figure 1).

In conclusion, the functional deficiency of antiviral immunityobserved in mice with cGvHD is consistent with the clinical reportsthat document an increased incidence of opportunistic viral infec-tions in patients with cGvHD.32 Using a murine allogeneic BMTmodel, we found that peripheral expansions of donor T cellsoccurred in the blood in response to MCMV infection, but these Tcells failed to efficiently home to MCMV-infected tissues and thusfailed to mount effective antiviral immune responses in mice withcGvHD. In contrast, allogeneic transplant recipients without cGvHDsuccessfully developed brisk and effective antiviral immune re-sponses. These findings indicate that studying functional cellularimmune responses of patients undergoing allogeneic BMT mayrequire measurement of tissue-localized effectors in addition tomonitoring antigen-specific cells in the blood. The failure ofeffective tissue-specific antiviral immune responses in transplantrecipients with cGvHD indicates the need for adoptive celltherapies that do not cause GvHD.15,18 The impaired immuneresponses to the MCMV-Listeria vaccine in mice with cGvHDsuggest that intrinsic immunosuppression associated with clinicalcGvHD rather than immunosuppressive drugs limits the efficacy ofvaccination in this setting.

Acknowledgment

The authors wish to thank Cerus for providing amotosalen (S-59Psoralen.HCl) and Listeria-MCMV vaccine.

This work was supported by National Institutes of Health (NIH)grant HL70997 (J.D.R.) and by NIH grant CA-74364 (E.K.W.).

Authorship

Conflict-of-interest disclosure: J.D.R. receives research grant sup-port from Cerus Corporation, a company whose potential productwas studied in the present work. J.D.R. and E.K.W. have a USpatent pending related to the work described in the present study.The remaining authors declare no competing financial interests.

Correspondence: Edmund K. Waller, Department of Hematol-ogy and Oncology, Winship Cancer Institute, Emory University,1365 Clifton Road, Building C, WCI 4th Floor, Room # C4032,Atlanta, GA 30322; e-mail: ewaller@emory.edu.

References

1. Xystrakis E, Bernard I, Dejean AS, Alsaati T,Druet P, Saoudi A. Alloreactive CD4 T lympho-cytes responsible for acute and chronic graft-versus-host disease are contained within theCD45RChigh but not the CD45RClow subset. EurJ Immunol. 2004;34:408-417.

2. Antin JH, Ferrara JL. Cytokine dysregulation andacute graft-versus-host disease. Blood. 1992;80:2964-2968.

3. Ferrara JL, Deeg HJ. Graft-versus-host disease.N Engl J Med. 1991;324:667-674.

4. Ferrara JL. Cellular and cytokine effectors ofacute graft versus host disease. Int J Hematol.2002;76(Suppl 1):195-198.

5. Ratanatharathorn V, Ayash L, Lazarus HM, Fu J,Uberti JP. Chronic graft-versus-host disease:clinical manifestation and therapy. Bone MarrowTransplant. 2001;28:121-129.

6. Wingard JR, Piantadosi S, Vogelsang GB, et al.Predictors of death from chronic graft-versus-hostdisease after bone marrow transplantation.Blood. 1989;74:1428-1435.

7. Haddad E, Landais P, Friedrich W, et al. Long-term immune reconstitution and outcome afterHLA-nonidentical T-cell-depleted bone marrowtransplantation for severe combined immunodefi-ciency: a European retrospective study of 116patients. Blood. 1998;91:3646-3653.

CHRONIC GvHD AND ANTIVIRAL IMMUNITY 4555BLOOD, 15 MAY 2007 � VOLUME 109, NUMBER 10

For personal use only. by guest on June 2, 2013. bloodjournal.hematologylibrary.orgFrom

8. Butturini A, Gale RP. T cell depletion in bone mar-row transplantation for leukemia: current resultsand future directions. Bone Marrow Transplant.1988;3:185-192.

9. Butturini A, Gale RP. Graft versus leukemia. Im-munol Res. 1992;11:24-33.

10. Barrett AJ. Mechanisms of the graft-versus-leuke-mia reaction. Stem Cells. 1997;15:248-258.

11. Meyers JD, Flournoy N, Thomas ED. Risk factorsfor cytomegalovirus infection after human marrowtransplantation. J Infect Dis. 1986;153:478-488.

12. Einsele H, Hebart H, Kauffmann-Schneider C, etal. Risk factors for treatment failures in patientsreceiving PCR-based preemptive therapy forCMV infection. Bone Marrow Transplant. 2000;25:757-763.

13. Nguyen Q, Champlin R, Giralt S, et al. Late cyto-megalovirus pneumonia in adult allogeneic bloodand marrow transplant recipients. Clin Infect Dis.1999;28:618-623.

14. Hossain MS, Roback JD, Lezhava L, Hillyer CD,Waller EK. Amotosalen-treated donor T cellshave polyclonal antigen-specific long-term func-tion without graft-versus-host disease after allo-geneic bone marrow transplantation. Biol BloodMarrow Transplant. 2005;11:169-180.

15. Roback JD, Hossain MS, Lezhava L, et al. Alloge-neic T cells treated with amotosalen prevent le-thal cytomegalovirus disease without producinggraft-versus-host disease following bone marrowtransplantation. J Immunol. 2003;171:6023-6031.

16. Lieberman M, Hansteen GA, McCune JM, ScottML, White JH, Weissman IL. Indirect induction ofradiation lymphomas in mice: evidence for anovel, transmissible leukemogen. J Exp Med.1987;166:1883-1893.

17. Waller EK, Ship AM, Mittelstaedt S, et al. Irradi-ated donor leukocytes promote engraftment of

allogeneic bone marrow in major histocompatibil-ity complex mismatched recipients without caus-ing graft-versus-host disease. Blood. 1999;94:3222-3233.

18. Giver CR MR, Mittelstaedt S, et al. Ex vivo flu-darabine exposure inhibits graft-versus-host ac-tivity of allogeneic T cells while preserving graft-versus-leukemia effects. Biol Blood MarrowTransplant. 2003;9:616-632.

19. Hossain MS, Takimoto H, Ninomiya T, et al. Char-acterization of CD4- CD8- CD3� T-cell receptor-alphabeta� T cells in murine cytomegalovirusinfection. Immunology. 2000;101:19-29.

20. Hossain MS, Takimoto H, Hamano S, et al. Pro-tective effects of hochu-ekki-to, a Chinese tradi-tional herbal medicine against murine cytomega-lovirus infection. Immunopharmacology. 1999;41:169-181.

21. Gold MC, Munks MW, Wagner M, KoszinowskiUH, Hill AB, Fling SP. The murine cytomegalovi-rus immunomodulatory gene m152 prevents rec-ognition of infected cells by M45-specific CTL butdoes not alter the immunodominance of the M45-specific CD8 T cell response in vivo. J Immunol.2002;169:359-365.

22. Cooke KR, Hill GR, Crawford JM, et al. Tumornecrosis factor- alpha production to lipopolysac-charide stimulation by donor cells predicts theseverity of experimental acute graft-versus-hostdisease. J Clin Invest. 1998;102:1882-1891.

23. Brockstedt DG, Giedlin MA, Leong ML, et al. Lis-teria-based cancer vaccines that segregate im-munogenicity from toxicity. Proc Natl Acad SciU S A. 2004;101:13832-13837.

24. Foster AE, Marangolo M, Sartor MM, et al. Hu-man CD62L- memory T cells are less responsiveto alloantigen stimulation than CD62L� naive Tcells: potential for adoptive immunotherapy.Blood. 2004;104:2403-2409.

25. Seder RA, Ahmed R. Similarities and differencesin CD4� and CD8� effector and memory T cellgeneration. Nat Immunol. 2003;4:835-842.

26. Milinkovic M, Antin JH, Hergrueter CA, UnderhillCB, Sackstein R. CD44-hyaluronic acid interac-tions mediate shear-resistant binding of lympho-cytes to dermal endothelium in acute cutaneousGVHD. Blood. 2004;103:740-742.

27. Lesley J, He Q, Miyake K, Hamann A, Hyman R,Kincade PW. Requirements for hyaluronic acidbinding by CD44: a role for the cytoplasmic do-main and activation by antibody. J Exp Med.1992;175:257-266.

28. DeGrendele HC, Estess P, Picker LJ, SiegelmanMH. CD44 and its ligand hyaluronate mediaterolling under physiologic flow: a novel lympho-cyte-endothelial cell primary adhesion pathway. JExp Med. 1996;183:1119-1130.

29. Taher TE, Smit L, Griffioen AW, Schilder-Tol EJ,Borst J, Pals ST. Signaling through CD44 is medi-ated by tyrosine kinases: association with p56lckin T lymphocytes. J Biol Chem. 1996;271:2863-2867.

30. Soderberg C, Larsson S, Rozell BL, Sumitran-Karuppan S, Ljungman P, Moller E. Cytomegalo-virus-induced CD13-specific autoimmunity: apossible cause of chronic graft-vs-host disease.Transplantation. 1996;61:600-609.

31. Via CS, Shanley JD, Shearer GM. Synergisticeffect of murine cytomegalovirus on the inductionof acute graft-vs-host disease involving MHCclass I differences only: analysis of in vitro T cellfunction. J Immunol. 1990;145:3283-3289.

32. Kulkarni S, Powles R, Treleaven J, et al. Chronicgraft versus host disease is associated with long-term risk for pneumococcal infections in recipi-ents of bone marrow transplants. Blood. 2000;95:3683-3686.

4556 HOSSAIN et al BLOOD, 15 MAY 2007 � VOLUME 109, NUMBER 10

For personal use only. by guest on June 2, 2013. bloodjournal.hematologylibrary.orgFrom