Porcine Thymic Grafts Protect Human Thymocytes from HIV‐1–Induced Destruction

900 • JID 2007:196 (15 September) • Hongo et al.

M A J O R A R T I C L E

Porcine Thymic Grafts Protect Human Thymocytesfrom HIV-1–Induced Destruction

David Hongo, Sima Hadidi,a Scott Damrauer,a Valerie Garrigue, Daniel Kraft,a David H. Sachs, Boris Nikolic,a

and Megan SykesBone Marrow Transplantation Section, Transplantation Biology Research Center, Massachusetts General Hospital/Harvard Medical School, Boston

Human immunodeficiency virus type 1 (HIV-1) infection depletes thymocytes and destroys thymic structure.Functional, tolerant human T cells develop in vivo in immunodeficient mice receiving porcine thymus andhuman fetal liver fragments under the kidney capsule. In this model, we evaluated the potential of porcinethymus to protect human thymocytes from the effects of HIV-1. Compared with that observed in control micewith human thymic grafts, porcine thymus attenuated human thymocyte depletion by the CCR5-tropic isolateJR-CSF without preventing thymocyte infection. Porcine thymus protected human thymocytes from infectionand depletion by a CXCR4-tropic HIV-1 isolate without reducing peripheral blood viral loads or T cell infection.Human thymocytes from human but not porcine grafts showed decreased Bcl-2 expression and increasedapoptosis after NL4.3 infection. Thus, porcine thymus protects human thymocytes from the cytopathic effectof HIV-1, suggesting a possible approach to achieving immune restoration in patients with acquired immu-nodeficiency syndrome who have incomplete responses to antiretroviral therapy. The model allows analysisof the mechanisms of HIV-mediated thymic dysfunction.

HIV-1 destroys thymocytes and thymic stromal cells [1,

2]. CD4 T cell deficiency after HIV-1 infection reflects

shortened survival and failure to replenish depleted cells

[3, 4]. Thymic regeneration can contribute to immune

Received 20 December 2006; accepted 26 March 2007; electronically published7 August 2007.

Potential conflicts of interest: none reported.Presented in part: 12th International Congress of Immunology and 4th Annual

Conference of the Federation of Clinical Immunology Societies, Montreal, 18–23July 2004 (abstract 4006); International Symposium on HIV and Emerging InfectiousDiseases, Toulon, France, 3–5 June 2004 (oral presentation OP8.4).

Financial support: National Institutes of Health (National Cooperative DrugDiscovery Group Integrated Preclinical/Clinical Program grant PO1 AI39755).

a Present affiliations: National Center for Cool and Cold Water Aquaculture,North Atlantic Area, Agricultural Research Service, Research, Education, andEconomics, US Department of Agriculture, Kearneysville, West Virginia (S.H.);Department of Surgery, Massachusetts General Hospital (S.D.), and the Renal Unit,Massachusetts General Hospital/Harvard Medical School (B.N.), Boston; De-partment of Pathology, Stanford University School of Medicine, Stanford, Cali-fornia (D.K.).

Reprints or correspondence: Dr. Megan Sykes, Bone Marrow TransplantationSection, Transplantation Biology Research Center, Massachusetts General Hospital/Harvard Medical School, MGH-East Bldg. 149–5102, 13th St., Boston, MA 02129([email protected]).

The Journal of Infectious Diseases 2007; 196:900–10� 2007 by the Infectious Diseases Society of America. All rights reserved.0022-1899/2007/19606-0015$15.00DOI: 10.1086/520516

reconstitution after effective highly active antiretroviral

therapy (HAART) [5]. Although thymus involution oc-

curs with age, the adult thymus remains active late in

life [6, 7]. However, some patients do not achieve ad-

equate immune recovery despite having good antiviral

responses [8], and some develop resistance to antiret-

rovirals [9, 10].

Allogeneic thymic transplants achieved immune res-

toration in children with DiGeorge syndrome [11], and

similar transplants were evaluated in patients with

AIDS, without success, during the pre-HAART era [12].

Here, we explore the possibility that porcine thymic

xenotransplantation might replace the HIV-1–affected

human thymus and restore thymopoiesis. We have pre-

viously shown that porcine thymic grafts can replace

the host thymus in thymectomized T cell–depleted mice

to generate a xenotolerant, functional T cell repertoire

[13–15]. Furthermore, porcine thymic grafts support

the development of polyclonal, functional, and tolerant

human T cells in immunodeficient mice [16]. Because

porcine cells resist infection with HIV-1, we hypothe-

sized that porcine thymic transplantation might permit

immune restoration in the presence of HIV-1 infection.

by guest on September 8, 2016

http://jid.oxfordjournals.org/D

ownloaded from

http://jid.oxfordjournals.org/

Porcine Thymus in HIV-1 Infection • JID 2007:196 (15 September) • 901

We tested this hypothesis in immunodeficient mice receiving

thymic xenografts with human hematopoietic progenitors.

MATERIALS AND METHODS

Mice. NOD/SCID mice (purchased from Jackson Laborato-

ries) were maintained in sterile microisolator cages according

to National Institutes of Health guidelines and underwent

transplantation at 5–7 weeks of age.

Transplantation. Human fetal tissue was obtained from

Advanced Bioscience Resources with institutional approval.

Porcine fetal tissue was harvested from Massachusetts General

Hospital (MGH) miniature swine as described elsewhere [13].

Human (gestation week 17) or porcine (gestation day 66–72)

fetal thymic fragments were grafted with human fetal liver frag-

ments (hereafter, HU/HU and SW/HU, respectively) under the

kidney capsule as described elsewhere [16]. In later experi-

ments, mice also received human fetal liver–derived55 � 10

CD34 cells from the same donor intravenously to enhance

human T cell and non–T cell reconstitution [17, 18]. NOD/

SCID mice received whole-body irradiation (1.5 Gy) in a Mark

I Cs137 irradiator (JL Shepard and Associates) 2 h before

transplantation.

Isolation of human CD34 cells. Human fetal liver CD34

cells were isolated as described elsewhere [18] using MACS and

anti-CD34 microbeads (Miltenyi Biotec), yielding 190% CD34

cells.

HIV-1 infection. Infection was induced 8–16 weeks after

grafting. Thymic grafts were injected with supernatant from

HIV-1– or mock-infected peripheral blood mononuclear cells

(PBMCs) via laparotomy under general anesthesia. HIV-1 iso-

lates included JR-CSF (R5 strain; obtained from the Partners

AIDS Research Center), NL4.3 (�4 strain), and JH-580 (dual

tropic) (the latter 2 provided by L. Su, University of North

Carolina, Chapel Hill). JR-CSF and NL4.3 doses were 2000–

4000 TCID50, whereas that for JH-580 was TCID50.57 � 10

Immunohistochemical staining of thymic tissue sections.

Thymic tissue sections were washed, incubated with horse se-

rum and then with mouse anti–human cytokeratin monoclonal

antibody (MAb; AE1/AE3; Dako), washed, incubated with per-

oxidase blocking solution, rinsed, incubated with biotinylated

horse anti–mouse IgG and then with avidin-biotinylated en-

zyme complex, rinsed, and incubated with diaminobenzidine

peroxidase substrate. Sections were counterstained with he-

matoxylin and then rinsed before dehydration in ethanol, fol-

lowed by clearing with xylene.

Flow cytometric analysis. Single-cell suspensions from thy-

mic grafts and spleens were prepared, and PBMCs were isolated

over Ficoll-Histopaque 1077 (BioWhitakker). Splenocytes were

treated with ACK erythrocyte-lysing buffer (BioWhitakker).

Anti-FcRg MAb (2.4G2) or human serum was added before

staining with MAbs. Intracellular staining for p24 and Bcl-2

expression was performed using the Cytofix/Cytoperm Kit (BD

Pharmingen) before staining with MAbs for 30 min at 4�C in

Perm/Wash buffer (1�). Anti–human CD8–allophycocyanin,

anti–human CD4–phycoerythrin (PE), anti–human CD45–

fluorescein isothiocyanate (FITC), rat anti–mouse IgG2a–PE,

Ly.5-PE, anti–human CD3–FITC, anti–human CD14–PE, and

anti–human CD19–PE were purchased from Pharmingen.

Anti–porcine thymocyte/T cell marker CD2 (MSA-4–FITC)

was obtained from D.H.S. Anti-p24 MAb KC57–FITC was pur-

chased from Coulter, and anti-Bcl2–PE was purchased from

BD Immunocytometry Systems. Stained cells from HIV-infec-

tion studies were fixed in 1% paraformaldehyde in PBS before

analysis on a FACSCalibur (BD Biosciences). Apoptotic cells

were detected with annexin V together with 7-amino actino-

mycin D or propidium iodide to exclude dead cells. Gated

lymphocytes (5000–10,000) were analyzed using CellQuest

(version 1.0; BD Biosciences) or FlowJo software (version 6.4.5;

Tree Star).

Determination of plasma viral load. Blood was anticoag-

ulated in EDTA, and plasma was isolated and stored at �80�C.

Viral loads were determined using the reverse-transcription

polymerase chain reaction (PCR) Amplicor HIV Monitor Test

(Roche) at the MGH Microbiology Laboratory.

Measurement of viral DNA. Nested PCR amplification of

thymocyte DNA with gag-specific primers SK100 and SK104

as well as SK38 and SK39 or of human-specific b-actin with

control primers was performed as described elsewhere [19].

PCR products were subjected to agarose gel electrophoresis,

imaged by ethidium bromide staining, transferred to a nylon

membrane, hybridized with either a 32P-labeled HIV-specific

probe (SK19) or b-actin–specific probes [20], and analyzed by

phosphoimager (BioRad Molecular Imager System GS-363).

Viral copy quantification was achieved by comparison to stan-

dard curves consisting of DNA isolated from known numbers

of ACH-2 cells (1 copy of HIV-1 genome per cell) and from

uninfected human lymphocytes. The signal from each sample

was compared with both b-actin and SK19 standard curves,

and the percentage of infected cells was calculated accordingly.

Double-immunofluorescence staining. Thymic grafts were

harvested from HU/HU and SW/HU recipients 5 weeks after

infection. Frozen thymic sections (8-mm thickness) were fixed

in acetone, air-dried, and washed and were then permeabilized

in 0.1% saponin, blocked with 0.5% bovine serum albumin

(Vector Laboratories), and stained with either purified anti–

human CD11c MAb (CBR-p150/4G1, mouse IgG2a; Chemi-

con) for myeloid dendritic cells (DCs) or CD68 MAb (PG-M1,

mouse IgG3; Dako) for macrophages and with either PE-labeled

anti–HIV-1 p24 MAb (KC57-RD1, mouse IgG1; Beckman

Coulter) or IgG isotype–matched antibodies (negative controls)



ownloaded from



before incubation overnight at 4�C in the dark. In preliminary

titrations, we showed that the specific MAbs did not cross-react

on porcine thymus. After being washed, sections were incubated

with biotinylated secondary MAbs (biotinylated rat anti–mouse

IgG2a MAb R12-3 or rat anti–mouse IgG3 MAb R40-82). Sec-

tions were washed, stained with FITC-avidin, washed, and then

mounted in 4′-6-diamidimo-2-phenylindole–containing Vec-

tastain medium (Vector Laboratories). Fluorescence was ex-

amined using a Nikon inverted fluorescence microscope (Nikon

Eclipse TE2000-U). The 2-color fluorescence images are the

computer-merged (Photoshop) view of both FITC and PE

staining.

Statistical analysis. For parametric data, groups were com-

pared using Student’s unpaired t test. For nonparametric data,

the Mann-Whitney U test was used. was considered toP ! .05

indicate statistical significance.

RESULTS

Resistance of human T cells developing in porcine thymic

grafts to the cytopathic effect of the R5-tropic strain JR-CSF.

HIV-1 JR-CSF, an M-tropic strain that destroys thymus and

thymic epithelial cells in SCID-hu mice [3, 21, 22], was injected

into the grafts of mice with �5% human peripheral blood

lymphocytes 3–4 months after grafting. Six weeks after infec-

tion, grafts infected with heat-inactivated virus maintained nor-

mal thymocyte subsets. Live virus–infected HU/HU grafts

showed profound depletion of CD4/CD8 double-positive thy-

mocytes and of CD4 single-positive thymocytes with increased

total numbers of double-negative thymocytes, compared with

the numbers in mock-infected control mice (figure 1A and 1B).

In contrast, infected mice with SW/HU grafts, which often have

increased numbers of CD4/CD8 double-negative cells that are

porcine CD2 thymocytes [16], maintained human thymocyte

subsets comparable to those in control mice. No loss of CD4

or CD8 single-positive or double-positive thymocytes and no

increase in the number of double-negative thymocytes was ob-

served in infected compared with control SW/HU grafts (figure

1A and 1C). The reduced number of human CD4- and CD8-

negative thymocytes in infected versus uninfected SW/HU re-

cipients may reflect unexplained differences in the numbers of

porcine thymocytes [16] between the groups.

Cytokeratin staining demonstrated a marked loss of archi-

tecture and cytokeratin in infected HU/HU but not in SW/HU

grafts (figure 1D). Thus, porcine thymic grafts were protected

from the destructive effects of HIV-1 infection.

Infected recipients of HU/HU grafts demonstrated a marked

depletion of peripheral blood ( , ) andmean � SE 0.5% � 0.50%

splenic ( , ) ( ) T cells, com-mean � SE 0.3% � 0.005% n p 4

pared with those in mock-infected HU/HU control recipients

( , of PBMCs [ ] andmean � SE 32.2% � 0.050% P p .0003

of splenocytes [ ], respectively) (44% � 1.00% P p .0005 n p

), at 6 weeks. Although SW/HU control recipients had fewer4

peripheral blood human T cells than did HU/HU control

recipients, HIV-1 infection did not affect their numbers in

the peripheral blood ( , vs.mean � SE 6.5% � 3.40% 4.8% �

for control; ) or spleens ( , 5.05%0.005% P p .667 mean � SE

�2.05% vs. for control; ) of SW/HU5.0% � 0.005% P p .98

recipients. Thus, the loss of peripheral blood human T cell pop-

ulations was avoided in HIV-1–infected mice with porcine thymic

grafts.

Productive infection with HIV-1 JR-CSF in human thy-

mocytes developing in porcine thymic grafts. A wide range

of HIV copy numbers (10–10,500 copies/ thymocytes;51 � 10

mean, 6376 copies/ thymocytes; ) was detected at51 � 10 n p 8

6 weeks in HU/HU grafts, with no evidence of infection in

inoculated porcine thymi grafted with porcine fetal liver (SW/

SW grafts) (data not shown). Despite the lack of thymocyte

depletion, infected SW/HU grafts demonstrated a similarly wide

range of viral copy numbers (10–10,000 copies/ thy-51 � 10

mocytes; mean, 2406 copies/ thymocytes; ) to51 � 10 n p 10

that in infected HU/HU grafts.

At 3 weeks after infection, plasma viremia was detected in

1 of 4 infected HU/HU recipients and in 3 of 5 SW/HU re-

cipients. By 6 weeks, 4 of 4 infected HU/HU recipients (range,

988–6110 copies/mL) and 5 of 5 SW/HU recipients (564–

470,000 copies/mL) had measurable plasma HIV RNA, with

the highest level in a SW/HU mouse. Thus, no evidence was

obtained for a delay or failure to infect porcine thymic grafts,

despite the resistance of human thymopoiesis in SW/HU grafts

to the effects of HIV-1 infection.

To address the possibility that porcine thymic grafts simply

delayed rather than prevented thymic disruption by JR-CSF

infection, similar analyses were performed at 8 and 10 weeks

after infection. However, many of the uninfected HU/HU grafts

had lost their cellularity and peripheral blood T cells were min-

imal to undetectable by these times, making it difficult to assess

the effect of HIV-1 infection. Viral loads were minimal to un-

detectable in 4 of 6 infected HU/HU recipients, suggesting that

the loss of human T cells led to a loss of the viral reservoir.

Although only 2 of 6 infected HU/HU grafts harvested at 8 or

10 weeks contained live thymocytes ( and to-6 62.5 � 10 4 � 10

tal), 4 of 6 SW/HU grafts at these time points contained live

thymocytes ( , , , and ),6 6 6 64 � 10 4.1 � 10 4.9 � 10 109.5 � 10

suggesting that they were still protected from the cytopathic

effect of HIV-1 despite measurable viral loads in 5 of the 6

mice. However, as shown in table 1, the absolute number of

human double-positive thymocytes in uninfected SW/HU

grafts at 8 and 10 weeks was significantly higher than that in

infected SW/HU grafts, suggesting that a delayed cytopathic

effect of HIV-1 JR-CSF may occur in the xenogeneic grafts.

No depletion of infected peripheral blood CD4 T cells in

HIV-1 NL4.3–infected SW/HU recipients. CXCR4-using



ownloaded from



Figure 1. Depletion of CD4/CD8 double-positive (DP) and CD4 single-positive (SP) thymocytes in HU/HU grafts (human fetal thymic fragments graftedwith human fetal liver fragments) but not SW/HU grafts (porcine fetal thymic fragments grafted with human fetal liver fragments) infected with theR5-tropic HIV-1 strain JR-CSF. A, Examples of representative individual infected and uninfected HU/HU and SW/HU grafts 6 weeks after infection. Band C, total no. of thymocytes in each subset from groups (np 4) of infected and uninfected HU/HU (B) and SW/HU (C) recipients killedMean � SE6 weeks after infection with HIV-1 JR-CSF. Results are inclusive of 2 separate experiments with similar results. D, Loss of thymic architecture andcytokeratin staining in HU/HU but not SW/HU grafts caused by infection with HIV-1 JR-CSF. Immunoperoxidase staining using an anti-cytokeratinantibody that detects both porcine and human thymic epithelial cells is shown. Results are representative of thymic sections from 2 mice per group,which showed similar results. NS, not significant.

(X4) T-tropic strains of HIV-1 cause rapid depletion of thy-

mocytes in SCID-hu thymus/liver-grafted mice [23]. We com-

pared the effects of an X4 strain, NL4.3, in porcine and human

thymic grafts. SW/HU and HU/HU recipients were prepared

with the addition of intravenous injection of CD34 cells from

the human fetal liver donor to enhance the level of human T

cell recovery, as described elsewhere [17]. Eight to 10 weeks

later, the grafts were injected with 2000 TCID50 of NL4.3 or

with mock-infected culture supernatants. In 2 separate exper-

iments, all infected SW/HU and HU/HU recipients showed very

high plasma viral loads 3, 5, and 7 weeks after infection

(1750,000 copies/mL), whereas virus was undetectable in the

plasma of mock-infected control mice.

By 3 weeks after infection, significant depletion of human

CD3 cells was observed in the blood of infected HU/HU re-

cipients but not of SW/HU recipients (figure 2A), which failed

to show depletion even by 7 weeks after infection (data not

shown). However, peripheral blood T cells in both infected

HU/HU and SW/HU recipients were productively infected,

with abundant p24 expression (figure 2B).



ownloaded from



Table 1. Absolute no. of human CD4/CD8 double-positive (DP)thymocytes at 8 and 10 weeks after infection.

Group

DP cells,median (range),

absolute no. (104)

Viral load,median,

copies/mL

HU/HU mock infected (n p 7) 8.5 (0–18,340) NDHU/HU infected (n p6) 0 (0–125) NDSW/HU mock infected (n p5) 2088 (372–3694) NDSW/HU infected (n p6) 240 (0–8979) 3035

NOTE. For HU/HU (human fetal thymic fragments grafted with human fetalliver fragments) mock infected vs. HU/HU infected, ; for SW/HU (porcineP p .1fetal thymic fragments grafted with human fetal liver fragments) mock infectedvs. SW/HU infected, ; for HU/HU mock infected vs. SW/HU mockP p .04infected, ; and for HU/HU infected vs. SW/HU infected, . ND,P p .04 P p .07not detectable.

Figure 2. Partial depletion of peripheral blood CD4 T cells in HU/HUgrafts (human fetal thymic fragments grafted with human fetal liverfragments) but not in SW/HU grafts (porcine fetal thymic fragmentsgrafted with human fetal liver fragments) caused by HIV-1 NL4.3 eventhough infection is seen in peripheral blood CD4 T cells in both groupsof mice. A, Percentage of peripheral blood human (h) CD3, hCD4, andhCD8 T cells among peripheral blood lymphocytes (PBLs) in the indicatedgroups 3 weeks after infection (for HU/HU NL4.3 infected, ; forn p 3HU/HU mock infected, ; for SW/HU NL4.3 infected, ; forn p 4 n p 5SW/HU mock infected, ; * ). Infected SW/HU recipients alson p 4 P ! .03failed to show depletion at 5 and 7 weeks after infection (data notshown). Data are values. B, Similar percentages of HIV-1mean � SEp24–expressing cells among peripheral blood hCD3 cells of infected HU/HU and SW/HU recipients injected with CD34 cells (P 1 .05, for HU/HUvs. SW/HU). Data are values. Similar results were obtainedmean � SEin 2 of 2 experiments. ND, not detectable; NS, not significant.

Protection of human thymocytes from NL4.3 infection in

SW/HU grafts. When mice were killed 3, 5, and 7 weeks after

infection, infected HU/HU grafts showed marked reductions

in total thymocyte numbers, including all thymocyte subsets

but most markedly in double-positive and CD4 single-positive

subsets, compared with the numbers in mock-infected grafts

(figure 3A). The lower thymocyte numbers in control groups

in this experiment, compared with those shown in figure 1,

reflect the biological variation between experiments seen in this

model. In contrast to HU/HU grafts, infected SW/HU grafts

showed no difference in total thymocyte number or subset

distribution compared with those in mock-infected control

grafts, even by 7 weeks after infection (figure 3A). Thymocytes

from HIV-1–infected HU/HU grafts were productively infected,

as indicated by p24 expression, whereas p24 was not expressed

by thymocytes from infected SW/HU grafts (figure 4B). Similar

results were achieved in 2 experiments.

Protection of human thymocytes from HIV-1 NL4.3–in-

duced apoptosis conferred by porcine thymus. One of the

hallmarks of HIV-1 infection is the induction of apoptosis

among infected thymocytes [2]. Increased levels of apoptosis

were detected among all thymocyte subsets in infected HU/HU

grafts, compared with those in mock-infected control grafts,

most markedly among double-positive thymocytes (figure 4A).

In contrast, infected SW/HU grafts showed no increase in the

number of apoptotic thymocytes, compared with that in un-

infected control grafts (figure 4A). Additionally, levels of ex-

pression of the antiapoptotic protein Bcl-2 were markedly re-

duced, compared with those in control grafts—especially

among CD4 single-positive and double-positive thymocytes—

in infected HU/HU grafts but were not markedly reduced in

infected SW/HU grafts (figure 4B). Thus, disruption of the

normal up-regulation of Bcl-2 that protects immature thy-

mocytes from apoptosis was associated with increased apoptosis

in infected HU/HU but not SW/HU grafts.

Similar expression of CXCR4 among thymocytes in both

infected HU/HU and SW/HU grafts. A failure to express the

coreceptor CXCR4 could not explain the lack of thymocyte

infection in SW/HU recipients, given that high levels of CXCR4

were expressed on almost all double-positive thymocytes and



ownloaded from



Figure 3. A, Human thymocyte nos. after NL4.3 infection. Infection with the X4 virus NL4.3 resulted in reduced nos. of human thymocytes by 3weeks after infection in HU/HU grafts (human fetal thymic fragments grafted with human fetal liver fragments), but SW/HU grafts (porcine fetalthymic fragments grafted with human fetal liver fragments) showed no depletion compared with that in uninfected control grafts even by 7 weeks.The top panel shows the nos. in groups of infected and uninfected SW/HU recipients killed at the indicated time points after infection (for HU/HUNL4.3 infected, ; for HU/HU mock infected, ; for SW/HU NL4.3 infected, ; for SW/HU mock infected, ; * and **n p 3 n p 4 n p 5 n p 4 P ! .03 P !

). The bottom panel shows the relative reduction in CD4/CD8 double-positive (DP) and CD4 single-positive (SP) thymocytes and the increased.007proportion of double-negative (DN) thymocytes in infected HU/HU but not SW/HU grafts (*** and ^^^ , for infected vs. noninfectedP ! .0004 P ! .0001grafts). Data are values. B, HIV-1 p24 expression among all thymocyte subsets of infected HU/HU grafts but not in any subset in infectedmean � SESW/HU grafts. The top panel shows the mean percentages of p24-expressing cells among the indicated thymocyte subsets of each group 5 weeksafter infection (for HU/HU NL4.3 infected, ; for HU/HU mock infected, ; for SW/HU NL4.3 infected, ; for SW/HU mock infected,n p 3 n p 4 n p 5

). All groups are presented in the bar graph, but lack of detectable p24 led to the absence of visible bars for any except the infected HU/HUn p 4group. The bottom panel shows gated DP thymocytes from 1 representative mouse in each group. Data are values from 1 of 2 similarmean � SEexperiments.

on high percentages of CD4 and CD8 single-positive thymo-

cytes in infected SW/HU grafts (figure 4C). In contrast, ex-

pression of CXCR4 was reduced in these thymocyte subsets in

infected HU/HU grafts (figure 4C).

Infection of human myeloid DCs and macrophages by

NL4.3 in HU/HU but not SW/HU grafts. In situ immuno-

fluorescence staining 5 weeks after infection revealed human

CD11c myeloid DCs (figure 5A, subpanels a, e, i, and m) and

CD68 macrophages (figure 5B, subpanels a, e, i, and m) in both

SW/HU and HU/HU grafts with or without NL4.3 infection.

A significant number of the CD11c DCs in the infected HU/

HU grafts were infected. Colocalization of HIV-1 p24–express-

ing cells with CD11c cells in HU/HU grafts was evidenced by

yellow fluorescence of merged colors (figure 5A, subpanels k

and l). CD68 macrophages and HIV-1 p24 antigen also colo-

calized in the infected HU/HU grafts (figure 5B, subpanels k

and l). In contrast, there was no colocalization of human

CD11c-positive DCs or human CD68-positive macrophages

and HIV-1 p24–expressing cells in SW/HU grafts, and p24

expression was faint or undetectable in infected SW/HU grafts

(figure 5A and 5B).

DISCUSSION

Disruption of thymopoiesis by HIV-1 contributes to HIV-in-

duced immunodeficiency. Polyclonal, functional, and xenotol-



ownloaded from


906

Figure 4. A, Increased apoptosis among infected HU/HU grafts (human fetal thymic fragments grafted with human fetal liver fragments) but notamong SW/HU grafts (porcine fetal thymic fragments grafted with human fetal liver fragments). The left panel shows total thymocytes from individualrepresentative mice in each group. The right panel shows the mean percentage of apoptotic cells in each thymocyte subset for each group at 5 weeksafter infection (for HU/HU NL4.3 infection, ; for HU/HU mock infection, ; for SW/HU NL4.3 infection, ; for SW/HU mock infection,n p 3 n p 4 n p 5

). Data are values from 1 of 2 similar experiments. All groups are presented in the bar graph, but a lack of detectable apoptoticn p 4 mean � SEcells led to the absence of visible bars in any except the infected HU/HU group. B, Loss of Bcl-2 expression among CD4/CD8 double-positive (DP)and CD4 single-positive (SP) thymocytes of infected HU/HU grafts but not SW/HU grafts. The left panel shows the mean percentages of Bcl-2–expressing cells among the indicated thymocyte subsets of each group at 5 weeks. Data are values from a single experiment (***mean � SE P !

). Similar results were obtained from 2 of 2 experiments. The right panel shows gated CD4 SP thymocytes from 1 representative mouse in each.0001group (for HU/HU NL4.3 infected, ; for HU/HU mock infected, ; for SW/HU NL4.3 infected, ; for SW/HU mock infected, ). C,n p 3 n p 4 n p 5 n p 4Reduction of CXCR4 expression in CD4 SP, CD8 SP, and DP thymocyte subsets of infected HU/HU grafts but not in any subset in infected SW/HUgrafts. The left panel shows the mean percentages of CXCR4-expressing cells among the indicated thymocyte subsets of each group at 5 weeks( per group; ** and *** ). Similar results were obtained from 2 of 2 experiments. Data are values from 1 experiment.n p 5 P ! .0064 P ! .0001 mean � SEThe right panel shows CXCR4 expression among gated DP thymocytes from 1 representative mouse in each group. 7-AAD, 7-amino actinomycin D;APC, allophycocyanin; PE, phycoerythrin.



ownloaded from


907

Figure 5. No productive infection with HIV-1 NL4.3 of human myeloid dendritic cells (A) and macrophages (B) in SW/HU (porcine fetal thymicfragments grafted with human fetal liver fragments) thymic grafts. HU/HU (human fetal thymic fragments grafted with human fetal liver fragments)and SW/HU thymus was either mock infected or infected with 2000 TCID50 of HIV-1 NL4.3. Five weeks after intrathymic infection, thymic tissueswere harvested. Sections were costained for either the myeloid DC marker CD11c (fluorescein isothiocyanate [FITC]; green staining; A) or the macrophagemarker CD68 (FITC; green staining; B ) and the HIV-1 p24 antigen (phycoerythrin [PE]; red staining). Colocalization of HIV-1–infected cells (CD68 orCD11c and p24; subpanels c, d, g, h, k, l, o, and p) is illustrated by yellow. Slides were examined with a Nikon immunofluorescence microscope(Nikon Eclipse TE2000-U). Analysis of thymic sections from 2 mice in each group showed similar results. Original magnifications were �20 for allsubpanels except d, h, l, and p and were �40 for d, h, l, and p.



ownloaded from



erant human T cells can develop in porcine thymic grafts im-

planted with human fetal liver tissue under the kidney capsule

of immunodeficient mice [16]. Because porcine source animals

could provide essentially unlimited numbers of genetically de-

fined and specific pathogen-free thymic tissues, we addressed

the hypothesis that porcine thymic grafts might protect devel-

oping human thymocytes from HIV-1 infection. We have dem-

onstrated that the porcine thymus protects human thymo-

poiesis from the effects of an R5 and an X4 HIV-1 isolate.

In the case of the R5 isolate, the porcine thymus protected

against the cytopathic effect of HIV-1 without preventing thy-

mocyte infection. Consistent with the findings of previous stud-

ies [24–26], HIV-1 JR-CSF depleted double-positive and single-

positive populations in HU/HU grafts within 6 weeks, and an

increase in the number of double-negative cells occurred. In

contrast, no such changes occurred in SW/HU mice. Moreover,

thymic cellularity and structure were maintained in infected

SW/HU mice, whereas they were obliterated in infected HU/

HU mice. The absence of provirus in thymocytes from SW/

SW mice indicates that the pig is resistant to HIV-1 infection.

Although the mechanisms of the protective effect of the porcine

thymus against the cytopathic effect of R5 HIV-1 infection are

unknown, we hypothesize that the porcine thymic microen-

vironment may resist destruction by HIV-1 [3]. Additionally,

species incompatibilities in stromally derived cytokines and

other indirect mechanisms involved in HIV-induced cytopath-

icity [27] may render thymocytes resistant to destruction in

porcine thymic grafts. Because the natural life span of the thy-

mic graft was limited, it was difficult to obtain meaningful data

later than 6 weeks after infection. Nevertheless, the data ob-

tained in SW/HU grafts at 8 and 10 weeks after infection in-

dicate ongoing human thymopoiesis at these time points but

do not rule out a gradual loss of thymopoiesis due to HIV

infection.

Thymopoiesis in SW/HU mice was also protected from the

disruptive effects of infection with an X4 strain, NL4.3, but by

a different mechanism than that involved after infection with

the R5 strain. Human thymocytes, macrophages, and DCs,

which were abundant in SW/HU grafts, avoided productive

infection by NL4.3, whereas all 3 cell types were infected in

HU/HU grafts. Infection of thymic macrophages and DCs may

lead to thymocyte infection and influence T cell maturation,

given their involvement in negative and positive selection [28].

Type I interferon production by thymic pre-DCs after HIV-1

infection may impair the development of functional T cells

[29]. The lack of infection of human thymic DCs and mac-

rophages, despite their presence in large numbers in infected

SW/HU grafts, suggests that thymic stroma–derived (i.e., non-

hematopoietic) factors promoting HIV-1 infection (e.g., spe-

cies-specific cytokines) in HU/HU grafts may be lacking in SW/

HU grafts. Some cytokines have been suggested to promote

HIV-1 infection or replication in thymocytes [30, 31], and some

(interleukin [IL]–7, tumor necrosis factor–a, IL-1, and IL-6) have

been shown to promote thymocyte infection in vitro [32, 33].

HIV-1–induced apoptosis of T cells is associated with in-

creased expression of apoptosis-inducing caspases and reduced

expression of survival factors such as Bcl-2 [34, 35]. Increased

apoptosis in association with reduced Bcl-2 expression, espe-

cially by the CD4 single-positive and double-positive subsets,

was observed in human thymocytes in NL4.3-infected HU/HU

but not SW/HU grafts. The results in infected HU/HU thy-

mocytes are consistent with results in a macaque simian im-

munodeficiency virus model [36]. Although our in vivo thymic

infection studies revealed decreased Bcl-2 expression among all

thymocyte subsets in HU/HU grafts, in vitro studies have sug-

gested Bcl-2 down-regulation only among double-positive cells,

with Bcl-2 up-regulation among CD4 single-positive thymo-

cytes after HIV-1 infection [37]. These differences may be ex-

plained by a contribution from the intact thymic microenvi-

ronment to indirect cytopathicity in the presence of HIV-1

infection.

Protection from X4 virus infection in porcine thymic grafts

was not due to failure to express the CXCR4 coreceptor, which

showed a normal expression pattern [19, 38] in human thy-

mocytes of SW/HU grafts, with high levels on double-positive

and on CD4 and CD8 single-positive thymocytes. Reduced

CXCR4 expression among these subsets in HU/HU grafts prob-

ably reflected selective destruction of CXCR4high thymocytes

and, hence, enrichment for those lacking CXCR4.

NL4.3-infected SW/HU mice were productively infected, as

demonstrated by p24 expression in peripheral blood T cells and

high plasma viral load. It is possible that human thymocytes

developing in porcine thymic grafts do not become infected

until they leave the thymus and enter the peripheral lymphoid

tissues, where a peripheral HIV-1 reservoir exists. This reservoir

may reside in the monocyte/macrophage and DC lineages, as

high levels of reconstitution with these human cells occur in

this humanized mouse model [18]. Monocytes, macrophages,

and DCs are important in initial M-tropic HIV-1 infection

because they express CD4 and the CCR5 coreceptor needed for

fusion and entry [39]. Infected monocytes may maintain a long-

lasting HIV-1 reservoir as macrophages or migratory DCs [40,

41] supporting high levels of viral replication [26]. However,

it is unclear how such a reservoir could be established by an

X4 virus, and we favor the hypothesis that mature peripheral

blood T cells are infected in SW/HU mice because of capture

of HIV-1 by molecules such as DC-SIGN (DC-specific inter-

cellular adhesion molecule–3–grabbing nonintegrin) [42] on

uninfected DCs in the periphery. Further studies are needed to

address this possibility. Moreover, it is unclear how HIV-1 en-

ters the periphery in these animals; leakage of virus during the

intrathymic injection process or emigration of an unidentified



ownloaded from



infected cell population from the thymus to the periphery are

possibilities.

Preserved thymopoiesis may explain the lack of depletion of

peripheral blood human T cells in NL4.3-infected SW/HU

mice, despite high viral loads, because T cells continually pro-

duced in the porcine thymus may replace peripheral blood T

cells that are destroyed. The contrasting rapid depletion of pe-

ripheral blood T cells in infected HU/HU mice may reflect the

failure of the disrupted thymopoietic process in human thymic

grafts to replace T cells destroyed in the periphery.

Thus, human thymocytes developing in porcine thymic grafts

can be protected from the cytopathic effect of HIV-1 by 2

different mechanisms, one in which thymocytes and antigen-

presenting cells resist infection and another in which thymo-

cytes are infected but are protected from the cytopathic effect.

In studies using a very high dose of the highly thymotropic

dual-tropic HIV-1 isolate JH 580, no significant protection of

human thymocytes was observed in porcine thymic grafts (data

not shown), indicating that the protective effect of xenogeneic

thymic tissue can be overridden by infection with a virus strain

that is exceptionally cytopathic to thymocytes [43]. Further

analyses of this model should provide insights into the mech-

anisms of thymocyte infection and destruction by HIV-1 and

may suggest ways of promoting immune restoration in HIV-

1–infected persons.

Acknowledgments

We thank Drs. Christian LeGuern and Paul Johnson for helpful discus-sions and critical reading of the manuscript. We also thank Dr. AkiraShimizu for performing immunohistochemical analysis, Mr. Timothy Ho-gan for technical assistance, Mr. Orlando Moreno for animal husbandry,Dr. Sharilyn Stanley for assistance with the studies, Ms. Kelly Walsh forexpert assistance with the manuscript, and Mr. Scott Arn for procurementof pigs for the study.

References

1. McCune JM. The dynamics of CD4+ T-cell depletion in HIV disease.Nature 2001; 410:974–9.

2. Su L, Kaneshima H, Bonyhadi M, et al. HIV-1-induced thymocytedepletion is associated with indirect cytopathogenicity and infectionof progenitor cells in vivo. Immunity 1995; 2:25–36.

3. Stanley SK, McCune JM, Kaneshima H, et al. Human immunodefi-ciency virus infection of the human thymus and disruption of thethymic microenvironment in the SCID-hu mouse. J Exp Med 1993;178:1151–63.

4. Bonyhadi ML, Rabin L, Salimi S, et al. HIV induces thymus depletionin vivo. Nature 1993; 363:728–32.

5. Hardy G, Worrell S, Hayes P, et al. Evidence of thymic reconstitutionafter highly active antiretroviral therapy in HIV-1 infection. HIV Med2004; 5:67–73.

6. Jamieson BD, Douek DC, Killian S, et al. Generation of functionalthymocytes in the human adult. Immunity 1999; 10:569–75.

7. Poulin JF, Viswanathan MN, Harris JM, et al. Direct evidence forthymic function in adult humans. J Exp Med 1999; 190:479–86.

8. Pakker NG, Kroon ED, Roos MT, et al. Immune restoration does not

invariably occur following long-term HIV-1 suppression during anti-retroviral therapy. INCAS Study Group. AIDS 1999; 13:203–12.

9. Hirsch M, Steigbigel R, Staszewski S, et al. A randomized, controlledtrial of indinavir, zidovudine, and lamivudine in adults with advancedhuman immunodeficiency virus type 1 infection and prior antiretro-viral therapy. J Infect Dis 1999; 180:659–65.

10. Nettles RE, Kieffer TL, Simmons RP, et al. Genotypic resistance inHIV-1-infected patients with persistently detectable low-level viremiawhile receiving highly active antiretroviral therapy. Clin Infect Dis 2004;39:1030–7.

11. Markert ML, Alexieff MJ, Li J, et al. Postnatal thymus transplantationwith immunosuppression as treatment for DiGeorge syndrome. Blood2004; 104:2574–81.

12. Dwyer JM, Wood CC, McNamara J, Kinder B. Transplantation of thy-mic tissue into patients with AIDS: an attempt to reconstitute theimmune system. Arch Intern Med 1987; 147:513–7.

13. Lee LA, Gritsch HA, Sergio JJ, et al. Specific tolerance across a dis-cordant xenogeneic transplantation barrier. Proc Natl Acad Sci USA1994; 91:10864–7.

14. Zhao Y, Fishman JA, Sergio JJ, et al. Immune restoration by fetal pigthymus grafts in T cell-depleted, thymectomized mice. J Immunol1997; 158:1641–9.

15. Zhao Y, Swenson K, Sergio JJ, Arn JS, Sachs DH, Sykes M. Skin grafttolerance across a discordant xenogeneic barrier. Nat Med 1996; 2:1211–6.

16. Nikolic B, Gardner JP, Scadden DT, Arn JS, Sachs DH, Sykes M. Normaldevelopment in porcine thymus grafts and specific tolerance of humanT cells to porcine donor MHC. J Immunol 1999; 162:3402–7.

17. Lan P, Wang L, Diouf B, et al. Induction of human T-cell tolerance toporcine xenoantigens through mixed hematopoietic chimerism. Blood2004; 103:3964–9.

18. Lan P, Tonomura N, Shimizu A, Wang S, Yang YG. Reconstitution ofa functional human immune system in immunodeficient mice throughcombined human fetal thymus/liver and CD34+ cell transplantation.Blood 2006; 108:487–92.

19. Ou CY, Kwok S, Mitchell SW, et al. DNA amplification for directdetection of HIV-1 in DNA of peripheral blood mononuclear cells.Science 1988; 239:295–7.

20. Shen H, Cheng T, Preffer FI, et al. Intrinsic human immunodeficiencyvirus type 1 resistance of hematopoietic stem cells despite coreceptorexpression. J Virol 1999; 73:728–37.

21. Namikawa R, Kaneshima H, Lieberman M, Weissman IL, McCune JM.Infection of the SCID-hu mouse by HIV-1. Science 1988; 242:1684–6.

22. Jamieson BD, Pang S, Aldrovandi GM, Zha J, Zack JA. In vivo path-ogenic properties of two clonal human immunodeficiency virus type1 isolates. J Virol 1995; 69:6259–64.

23. Kaneshima H, Su L, Bonyhadi ML, Connor RI, Ho DD, McCune JM.Rapid-high, syncytium-inducing isolates of human immunodeficiencyvirus type 1 induce cytopathicity in the human thymus of the SCID-hu mouse. J Virol 1994; 68:8188–92.

24. Berkowitz RD, Beckerman KP, Schall TJ, McCune JM. CXCR4 andCCR5 expression delineates targets for HIV-1 disruption of T cell dif-ferentiation. J Immunol 1998; 161:3702–10.

25. Uittenbogaart CH, Anisman DJ, Jamieson BD, et al. Differential tro-pism of HIV-1 isolates for distinct thymocyte subsets in vitro. AIDS1996; 10:F9–16.

26. Pedroza-Martins L, Gurney KB, Torbett BE, Uittenbogaart CH. Dif-ferential tropism and replication kinetics of human immunodeficiencyvirus type 1 isolates in thymocytes: coreceptor expression allows viralentry, but productive infection of distinct subsets is determined at thepostentry level. J Virol 1998; 72:9441–52.

27. Kovalev G, Duus K, Wang L, et al. Induction of MHC class I expressionon immature thymocytes in HIV-1-infected SCID-hu Thy/Liv mice:evidence of indirect mechanisms. J Immunol 1999; 162:7555–62.

28. McCune JM. Thymic function in HIV-1 disease. Semin Immunol 1997;9:397–404.

29. Keir ME, Stoddart CA, Linquist-Stepps V, Moreno ME, McCune JM.



ownloaded from



IFN-alpha secretion by type 2 predendritic cells up-regulates MHCclass I in the HIV-1-infected thymus. J Immunol 2002; 168:325–31.

30. Schnittman SM, Singer KH, Greenhouse JJ, et al. Thymic microen-vironment induces HIV expression: physiologic secretion of IL-6 bythymic epithelial cells up-regulates virus expression in chronically in-fected cells. J Immunol 1991; 147:2553–8.

31. Koka PS, Brooks DG, Razai A, Kitchen CM, Zack JA. HIV type 1infection alters cytokine mRNA expression in thymus. AIDS Res HumRetroviruses 2003; 19:1–12.

32. Chene L, Nugeyre MT, Guillemard E, Moulian N, Barre-Sinoussi F,Israel N. Thymocyte-thymic epithelial cell interaction leads to high-level replication of human immunodeficiency virus exclusively in ma-ture CD4+ CD8� CD3+ thymocytes: a critical role for tumor necrosisfactor and interleukin-7. J Virol 1999; 73:7533–42.

33. Rothe M, Chene L, Nugeyre MT, Braun J, Barre-Sinoussi F, Israel N.Contact with thymic epithelial cells as a prerequisite for cytokine-enhanced human immunodeficiency virus type 1 replication in thy-mocytes. J Virol 1998; 72:5852–61.

34. Korant BD, Strack P, Frey MW, Rizzo CJ. A cellular anti-apoptosisprotein is cleaved by the HIV-1 protease. Adv Exp Med Biol 1998;436:27–9.

35. Rasola A, Gramaglia D, Boccaccio C, Comoglio PM. Apoptosis en-hancement by the HIV-1 Nef protein. J Immunol 2001; 166:81–8.

36. Rosenzweig M, Connole M, Forand-Barabasz A, Tremblay MP, JohnsonRP, Lackner AA. Mechanisms associated with thymocyte apoptosis in-duced by simian immunodeficiency virus. J Immunol 2000; 165:3461–8.

37. Guillemard E, Nugeyre MT, Chene L, et al. Interleukin-7 and infectionitself by human immunodeficiency virus 1 favor virus persistence inmature CD4+CD8�CD3+ thymocytes through sustained induction ofBcl-2. Blood 2001; 98:2166–74.

38. Zhang L, He T, Talal A, Wang G, Frankel SS, Ho DD. In vivo distri-bution of the human immunodeficiency virus/simian immunodefi-ciency virus coreceptors: CXCR4, CCR3, and CCR5. J Virol 1998; 72:5035–45.

39. Kacani L, Frank I, Spruth M, et al. Dendritic cells transmit humanimmunodeficiency virus type 1 to monocytes and monocyte-derivedmacrophages. J Virol 1998; 72:6671–7.

40. Crowe S, Zhu T, Muller WA. The contribution of monocyte infectionand trafficking to viral persistence, and maintenance of the viral res-ervoir in HIV infection. J Leukoc Biol 2003; 74:635–41.

41. Meltzer MS, Nakamura M, Hansen BD, Turpin JA, Kalter DC, Gen-delman HE. Macrophages as susceptible targets for HIV infection, per-sistent viral reservoirs in tissue, and key immunoregulatory cells thatcontrol levels of virus replication and extent of disease. AIDS Res HumRetroviruses 1990; 6:967–71.

42. Geijtenbeek TB, Kwon DS, Torensma R, et al. DC-SIGN, a dendriticcell-specific HIV-1-binding protein that enhances trans-infection of Tcells. Cell 2000; 100:587–97.

43. Meissner EG, Coffield VM, Su L. Thymic pathogenicity of an HIV-1envelope is associated with increased CXCR4 binding efficiency andV5-gp41-dependent activity, but not V1/V2-associated CD4 bindingefficiency and viral entry. Virology 2005; 336:184–97.



ownloaded from


Porcine Thymic Grafts Protect Human Thymocytes from HIV‐1–Induced Destruction

Documents

Transcript of Porcine Thymic Grafts Protect Human Thymocytes from HIV‐1–Induced Destruction