Feline Immunodeficiency Virus Infection Is Characterized by B7+

Feline Immunodeficiency Virus Infection Is Characterized by B71CTLA41

T Cell Apoptosis

Mary B. Tompkins,1 Marta E. Bull,2 Janet L. Dow,1

Judith M. Ball,3 Ellen W. Collisson,3

Barbara J. Winslow,4 Anagha P. Phadke,3

Thomas W. Vahlenkamp,1 and Wayne A. Tompkins2

1Departments of Microbiology, Pathology, and Parasitology,

and 2Farm Animal Health and Resource Management, College

of Veterinary Medicine, North Carolina State University, Raleigh;3Department of Veterinary Pathobiology, College of Veterinary Medicine,

Texas A & M University, College Station; 4Schering-Plough Animal

Health, San Diego, California

The B7.1 and B7.2 costimulatory molecules on antigen-presenting cells provide second signals

for regulating T cell immune responses via CD28 and cytotoxic T lymphocyte antigen 4 (CTLA4)

on T cells. CD28 signals cell proliferation, whereas CTLA4 signals for anergy or apoptosis, termi-

nating the immune response. Because T cell apoptosis and immunodeficiency is a characteristic of

feline immunodeficiency virus (FIV)– infected cats, it is possible that negative T cell signaling via

B7 and CTLA4 may be favored in these cats. Flow cytometry revealed high percentages of CD8þ

and CD4þ cells expressing B7.1, B7.2, and CTLA4 in lymph nodes of FIV-positive cats and a large

fraction of CTLA4þ T cells coexpressing B7.1 and B7.2. Three-color analysis with anti-B7.1, anti-

B7.2, or anti-CTLA4 and TUNEL (terminal deoxynucleotidyl transferase nick-end-labeling)

analysis revealed that apoptosis was a characteristic of B7.1þB7.2þCTLA4þ T cells. These data

support the hypothesis that lymph node apoptosis and immune deterioration in FIV-infected

cats results from chronic B7.1- and/or B7.2-CTLA4–mediated T-T interactions.

Human immunodeficiency virus (HIV) and feline immunode-

ficiency virus (FIV) are lentiviruses that cause a remarkably sim-

ilar immunopathology [1] characterized by a gradual loss in

T helper cell numbers and function, resulting in immune defi-

ciency and increased susceptibility to secondary pathogens.

Several mechanisms have been proposed to account for virus-

induced T cell depletion, including direct viral cytopathogen-

esis [2], immune hyperactivation and exhaustion [3], immune

suppression mediated by viral and regulatory gene products [4],

and inappropriate immune killing of uninfected cells [5]. On

the basis of observations that T cells from asymptomatic HIV-

infected persons fail to produce interleukin (IL)–2 and prolif-

erate after antigenic stimulation, a number of investigators have

speculated that HIV immunopathogenesis is the result of accel-

erated programmed cell death or apoptosis [6]. This speculation

is supported by reports of abnormally high frequency of apopto-

tic T cells in the lymph nodes (LNs) of HIV-infected persons

[7, 8] and FIV-infected cats [9, 10].

Programmed cell death is a normal process that regulates the

development of the cellular repertoire throughout the life of an

organism and is initiated by engagement of specific cell surface

receptors with their ligands (e.g., FasL with Fas, tumor necro-

sis factor [TNF]–a with TNF-a receptor [TNF-aR], and B7

with cytotoxic T lymphocyte antigen 4 [CTLA4]) that transduce

intracellular signals mediating cell death. Although the Fas and

TNF receptors are highly conserved and induce caspase-me-

diate apoptosis in many tissue types [11, 12], the B7-CTLA4

apoptotic pathway appears to be T lineage–specific and med-

iates antigen-specific clonal deletion of previously activated T

cells via suppression of IL-2 [13, 14]. It is not known whether

the accelerated apoptosis observed in LNs of HIV-infected per-

sons and FIV-infected cats represents enhanced activity of the

Fas–TNF-aR pathway or the B7-CTLA4 pathway or both. In

support of the B7-CTLA4 clonal deletion pathway, progressive

immune deterioration in the asymptomatic stage of HIV and FIV

infections is manifested by the inability of CD4 cells to produce

IL-2 and proliferate in vitro in response to major histocompat-

ibility complex (MHC) class II–restricted recall antigens [6,

15]. This T cell immune hyporesponsiveness is reminiscent

of a state of anergy, as defined by the inability of T cells to pro-

duce IL-2 and proliferate in response to stimulation with com-

petent antigen-presenting cells (APCs) [16].

The B7 family (principally B7.1 [CD80] and B7.2 [CD86]) of

costimulatory molecules is normally found on professional APCs,

and these molecules interact with CD28 and CTLA4 (CD152)

on T cells to provide the necessary second signals for regulating

The Journal of Infectious Diseases 2002;185:1077–93q 2002 by the Infectious Diseases Society of America. All rights reserved.0022-1899/2002/18508-0012$02.00

Presented in part: Fifth International Feline Retrovirus Research Symposium,

Cannes, France, 25–27 May 2000 (pathogenesis 3.1).

The animal use in this study was approved by the University Animal Care

and Use Committee.

Financial support: National Institutes of Health (AI-38177 to M.B.T.; AI-

43858 to W.A.T.; AI-32360 to E.W.C.); Morris Animal Foundation (grant

96FE-09 to E.W.C.).

Reprints or correspondence: Dr. Wayne Tompkins, North Carolina State

University, College of Veterinary Medicine, Dept. of Farm Animal Health and

Resource Management, 4700 Hillsborough St., Raleigh, NC 27606 (Wayne

[email protected]).

Received 27 August 2001; revised 21 November 2001; electronically pub-

lished 1 April 2002

1077

Dow

nloaded from https://academ

ic.oup.com/jid/article/185/8/1077/815715 by guest on 26 D

ecember 2021

the immune response [14, 17]. B7.1 and B7.2 initially engage

CD28 on T cells, ensuring sustained IL-2 secretion, T cell pro-

liferation, and development of an immune response. If the B7-

CD28 signal does not occur after APC–T cell receptor (TCR)

engagement, the T cell becomes unresponsive or anergic to the

antigen and, in the absence of IL-2, undergoes apoptosis after

stimulation by a competent APC [18, 19]. T cells can also be an-

ergized via an active mechanism, whereby B7 molecules en-

gage CTLA4 on activated T cells. Although CD28 is constitu-

tively expressed on CD4þ T cells and transduces a signal for

cell survival and proliferation, CTLA4 is expressed only after

the CD4þ cells become activated (2–3 days after APC-TCR en-

gagement) [20–22], and, after engagement with B7 molecules,

transduces a negative signal to T cells, suppressing IL-2 tran-

scription and further proliferation, resulting in anergy and apopto-

sis [20–23]. Because the binding avidity of B7.1 and B7.2 for

CTLA4 is 20–100 times greater than that for CD28 [14], nega-

tive signaling would dominate on activated T cells expressing

CTLA4, thereby terminating the immune response. Thus,

sequential engagement of CD28 and CTLA4 by B7 molecules

maintains a balance of T cell activation or proliferation and ap-

optosis, a process important in regulating normal immune re-

sponses by clonal deletion of previously activated antigen-

specific T cells.

Although B7.1 and B7.2 receptors are normally restricted to

APCs, such as dendritic cells, activated monocytes, and B cells

[24, 25], numerous studies have demonstrated that these mole-

cules are up-regulated on CD8þ and CD4þ cells activated in vitro,

as well as on a small subset of CD4þ and CD8þ cells in patients

with autoimmune disease or HIV infection [26–31]. A number

of studies have demonstrated that B7þ T cells can function as

APCs [30, 31]. However, data suggest that T-T antigen presen-

tation does not costimulate IL-2 production and proliferation

but primes T cells for anergy and apoptosis after subsequent

antigen presentation by professional APCs [32, 33].

Thus, the accelerated apoptosis observed in LNs of HIV-

infected patients and FIV-infected cats, as well as under other

conditions of chronic immune stimulation such as autoimmune

diseases [26], could result from anergic signaling mediated by

activated T cells expressing B7 molecules. We propose that the

progressive loss of immune function in FIV infection is the re-

sult of chronic interactions between B7þ and CTLA4þ T cells,

leading to T cell anergy and apoptosis.

Materials and Methods

Cats. Specific pathogen–free cats were obtained from Liberty

Laboratories or Cedar River Laboratory; they ranged in age from 6

months to 10 years at the time samples were collected. FIV-infec-

ted cats were inoculated with the NCSU1 isolate of FIV [34–36] in-

travaginally with cell-associated virus, as described by Bucci et al.

[37]. All cats were�6 months old when inoculated with FIV and be-

came antibody-positive, as detected by commercial ELISA (IDEXX),

and provirus-positive, as detected by polymerase chain reaction

(PCR) for gag sequences, 4–8 weeks after infection. At the time

samples were taken, cats had been infected with FIV for a duration

of 4 weeks to 10 years. All FIV-infected cats had inverted CD4þ:

CD8þ ratios in the blood. Cats infected for >8 years had chronic

secondary oral or respiratory infections, and, at necropsy, several

were found to have B cell lymphomas. Control cats also ranged in

age from 6 months to 10 years.

Sample collection. LN cells were obtained by peripheral LN

biopsy, as described by Levy et al. [38]. In brief, cats were anesthe-

tized with intravenous ketamine and diazepam and maintained with

inhalant anesthetic (isoflurane). A popliteal LN was excised, and

butorphanol tartrate was administered to control postoperative dis-

comfort. Single-cell suspensions of LN cells were prepared for sub-

sequent flow cytometric analysis of surface antigen expression and

TUNEL-based analysis of apoptosis by gently passing the tissue

through a steel mesh screen. Blood was collected by jugular veni-

puncture into sodium citrate anticoagulant tubes. Whole blood was

used for flow cytometric analysis of lymphocyte surface antigens

by means of a whole blood lysis technique, as described by David-

son et al. [39] For TUNEL-based flow cytometric analysis of apop-

tosis, peripheral blood mononuclear cells (PBMC) were purified on

Percoll, as described by Tompkins et al. [40].

Antibodies. Feline CD4þ and CD8þ T cells were identified by

use of monoclonal antibodies (MAbs) developed in our laboratory

[41, 42]. A species–cross-reactive human recombinant CTLA4

pro tein fused to the Fc of murine IgG (huCTLA4-Ig) was obtained

from Ancell. A rabbit polyclonal antibody was made against puri-

fied feline CTLA4 peptide [43] linked to keyhole limpet hemocya-

nin (KLH). Purified peptide was glutaraldehyde-linked to KLH by

mixing peptide and KLH at a ratio of 100:1 (in nanomoles). Glutar-

aldehyde was added to a final concentration of 0.4%. After reacting

for 1 h at room temperature, glutaraldehyde was removed by dialy-

sis against PBS. A rabbit was immunized with 200 nmol of cross-

linked peptide in Freund’s complete adjuvant and boosted 3 times

at 4-week intervals with 200 nmol of peptide in Freund’s incom-

plete adjuvant. Specificity of the antibody was confirmed by West-

ern blot analysis. A fluorescein-conjugated donkey anti–rabbit

antibody was obtained from Jackson Laboratory. A biotinylated

MAb specific for feline MHC class II antigen (CAG5-307) was ob-

tained from Custom Monoclonals. A cross-reacting antibody to hu-

man CD14 (Clone TUK4) was obtained from Dako.

Construction of swinepox virus (SPV)/B7–histidine (HIS) expres-

sion vectors. The feline B7.1 and feline B7.2 genes were cloned

from concanavalin A–stimulated feline PBMC, as described else-

where [44]. Truncations of feline B7.1 (nt 1–720) and feline B7.2

(nt 1–748) genes excluding the putative transmembrane regions

were cloned by PCR, with use of gene-specific primer pairs (B7.1

sense primer, 50-TCGAGAATTCGGGTCACGCAGCAAAGTGG-

30; B7.1 antisense primer, 50-CTGAGGATCCGATGTAGACAG-

GTGAGATCTTTTCCA-30; B7.2 sense primer, 50-TCGACAATT-

GGATGGGCATTTGTGACAG-30; and B7.2 antisense primer, 50-

GGATCCTGTTCTGGGTCTTTATCCTTAG-30). Construction of

recombinant SPV expression vectors has been described elsewhere

[44]. In brief, the truncated B7 PCR gene products were cloned down-

stream of a synthetic, late-early pox promoter and fused to a 6X-

HIS tag (Novagen) at the 30 end of the B7 open-reading frame. A

Tompkins et al.1078 JID 2002;185 (15 April)

Dow



ecember 2021

gene cassette containing the Escherichia coli uidA gene driven by

an early synthetic pox promoter was cloned upstream of the B7 cas-

sette. An SPV homology vector was created containing the HindIII

K fragment of the SPV genome [45]. The B7-HIS and uidA gene cas-

settes were inserted into a unique Eco RI restriction site within the

HindIII K SPV genome fragment. Recombinant SPV/B7.1-HIS and

SPV/B7.2-HIS viruses were created by homologous recombination

in embryonic swine kidney (ESK-4) cells and plaque-purified ac-

cording to methods described elsewhere [46]. In brief, ESK-4 cells

were preinfected with wild-type SPV (Kasza strain) at an MOI of

0.01–0.001 for 2 h before transfection with 15 mg of homology plas-

mid with use of LIPOFECTIN reagent (Life Technologies Gibco-

BRL). Infection-transfections were incubated for 4–6 days until

complete cytopathic effect was observed. Cell lysates and super-

natant were harvested and frozen at 270�C. Infection-transfection

stocks were plaque-purified as described elsewhere [47], with mod-

ification, by use of the presence of E. coli b-glucuronidase and sub-

stituting the substrate 5-bromo-6-chloro-3-indolyl-b-d-galactopyr-

anoside (BIOSYNTH).

Confirmation of expression of feline B7.1 and feline B7.2 from

recombinant viruses, SPV/B7.1-HIS and SPV/B7.2-HIS, respec-

tively, was done by standard Western blot methods [48] with use

of polyclonal goat anti–human B7.1 and goat anti–human B7.2 anti-

bodies (R&D Systems).

Production of purified B7 protein. B7-HIS fusion proteins

were purified from clarified supernatants from ESK-4 cells infected

with recombinant SPV/B7-HIS viruses. B7-HIS proteins were puri-

fied by nickel chelate and DEAE ion-exchange chromatography.

Functionality of purified B7.1 and B7.2 proteins was confirmed in

a competitive binding assay (scintillation-proximity assay) for spe-

cific binding to human CTLA-4 and human CD28, with use of io-

dinated human B7.1 protein as competitor [49] (data not shown).

Production of feline B7.1 and B7.2 rabbit polyclonal antibodies.

Polyclonal antibodies to B7.1-HIS and B7.2-HIS purified proteins

were generated in rabbits. First, 200 mg of purified protein was

injected with complete Freund’s adjuvant. Rabbits were boosted

twice, 3 weeks apart, with 100mg of protein with incomplete Freund’s

adjuvant.

Production of feline B7.1 MAbs. BALB/c mice were injected

intraperitoneally with 20 mg of B7.1-HIS mixed with Corixa Ribi

adjuvant on 2 occasions 3 weeks apart. Three weeks after the second

injection, serum was tested by immunostaining of recombinant

SPV-B7 plaques, as described below. The mice were given a final

injection of 10 mg of B7.1-HIS in Ribi adjuvant, and the spleen cell

fusions were done 5 days later. Hybridoma supernatants were ini-

tially screened with use of a feline CD4þ cell line (FCD4E) known

to express high levels of B7.1 and B.2 on their surfaces. B7.1 speci-

ficity of the positive supernatants was determined with the recombi-

nant SPV-B7 plaque assay. One positive hybridoma clone, B7.1.66,

was selected for expansion, and supernatants were collected for

these studies. A fraction of supernatant was purified and conjugated

with phycoerythrin (PE).

B7 antibody specificity. Specificity of B7.1 and B7.2 rabbit

polyclonal antibodies (figure 1) and B7.1.66 MAb hybridoma super-

natant (figure 2A) was confirmed by immunostaining of recombi-

nant SPV plaques expressing full-length B7.1 and B7.2 proteins.

Recombinant SPV expressing full-length B7.1 (SPV/flB7.1) or full-

length B7.2 (SPV/flB7.2) was constructed as described above. Im-

munostaining of recombinant virus plaques was carried out as

described elsewhere [48]. In brief, ESK-4 cells were infected with

SPV/flB7.1 or SPV/flB7.2 overnight and overlaid with nutrient

agarose. After 7–10 days, the agarose was removed and the mono-

layer was fixed with 100% methanol. B7 rabbit antibodies, diluted

1:1000 in 5% dried milk in 1£ TS (50 mM Tris-HCl and 150 mM

NaCl), or hybridoma cell culture supernatant from B7.1 MAb, di-

luted 1:10, were applied to the monolayer, and monolayers were

incubated for 2 h at room temperature. Monolayers were washed

with 1£ TS and reacted with alkaline phosphatase–conjugated sec-

ondary antibody. Alkaline phosphatase substrate, nitroblue tetrazo-

lium–BCIP (200 mg/mL nitroblue tetrazolium and 150 mg/mL

5-bromo-4-chloro-3-indolyl-phosphatase-toluidine salt; Fisher Bio-

tech), diluted in 100 mM Tris (pH 9.6), 100 mM NaCl, and 50 mM

MgCl2, was added, and incubations were done for 5–30 min at

room temperature until plaques turned a purple-black color. Spe-

cificity of the rabbit B7 antibodies and supernatant from B7.1 MAb

was confirmed with SPV/fl-B7.1 and SPV/fl-B7.2 lysates in Western

blots (data not shown).

To further confirm that MAb B7.1.66 bound only B7.1, blocking

studies with the rabbit polyclonal anti-B7 antibodies were done. LN

cells were incubated with either rabbit polyclonal anti-B7.1 or anti-

B7.2, followed by incubation with PE-conjugated MAb B7.1.66,

and were analyzed by flow cytometry. Polyclonal rabbit anti-B7.1

reduced B7.1.66 binding to LN T cells by �80%, whereas there was

no reduction in B7.1.66 binding when cells were pretreated with

rabbit polyclonal anti-B7.2 (figure 2B).

Figure 1. Rabbit polyclonal antibody specificity for feline B7.1 orB7.2. Embryonic swine kidney–4 cells were infected with swinepoxvirus (SPV) expressing full-length B7.1 (SPV/fl-B7.1) (A) or SPV/fl-B7.2 (B), were fixed, and were incubated with polyclonal rabbitanti–feline B7.1 or B7.2. Controls consisted of preimmune rabbitserum. Plaques were developed with alkaline phosphatase–conjugatedsecondary antibody. Note that plaques are present only when appropri-ate polyclonal antibody is added and that there is no cross-reactivity.

B7þCTLA4þ T Cell Apoptosis and FIV InfectionJID 2002;185 (15 April) 1079

Dow



ecember 2021

Flow cytometric analysis of lymphocyte surface antigens.

Three-color flow cytometric analysis was used to determine the

percentages of CD4þ, CD8þ, and B7þ cells in whole blood and LN

samples with use of fluorescein isothiocyanate–anti–feline CD4

(MAb CAT30A), allophycocyanin–anti–feline CD8 (MAb 117),

and PE–huCTLA4-Ig (Ancell) for B7. When rabbit anti–feline

B7.1 and anti-B7.2 antibodies were used, cells were stained first

with either of these antibodies, followed by a donkey PE–anti-rab-

bit antibody, and finally with the anti-CD4 and anti-CD8 directly

conjugated antibodies. Negative controls consisted of irrelevant iso-

type-matched MAbs or preimmunization rabbit serum. All reac-

tions were incubated at 4�C. Care was taken to perform all flow cy-

tometry as soon after collecting samples from the cats as possible, to

avoid any culture-induced changes in expression of costimulatory

molecules. Data were acquired on a FACSCaliber flow cytometer

(Becton Dickinson), with use of a helium-neon laser as the second

excitation source for the allophycocyanin-stained samples. For all

samples, data from at least 15,000 cells were acquired and stored

list-mode fashion for subsequent analysis. Gated data then were

generated for fluorescent analysis of lymphocytes, as defined by for-

ward and side scatter.

Three-color flow cytometric–based TUNEL assay for identifica-

tion of apoptotic cells. T cell apoptosis in LNs and PBMC was

determined by TUNEL assay (Boehringer Mannheim). Cells were

Figure 2. Monoclonal antibody (MAb) B7.1.66 specificity for feline B7.1. A, Plaque assay using swinepox virus (SPV) expressing full-lengthB7.1 (SPV/fl-B7.1)–infected or SPV/fl-B7.2–infected embryonic swine kidney–4 monolayers immunostained with B7.1.66. B, Blocking ofMAb B7.1.66 binding by rabbit polyclonal anti-B7.1 but not anti-B7.2. Lymph node cells were incubated first with either rabbit polyclonal anti-B7.1 or -B7.2 followed by MAb B7.1.66 and then fluorescein isothiocyanate (FITC)–conjugated anti-mouse antibody (dashed line). Controlsconsist of staining with rabbit anti-B7.1 or -B7.2 alone (heavy line) or incubating with preimmune rabbit serum before MAb B7.1.66 (lightline). Rabbit anti-B7.1 reduced MAb B7.1.66 binding by �80%, whereas rabbit anti-B7.2 did not interfere with MAb binding.


Dow



ecember 2021

first stained with rabbit anti-B7.1 or anti-B7.2 antibody, followed

by the donkey PE–anti-rabbit antibody. Cells were then stained

with either biotinylated anti-CD4 or anti-CD8 antibodies and devel-

oped with streptavidin-conjugated allophycocyanin. The cells were

fixed in 4% paraformaldehyde and permeabilized with Tween 20

and sodium citrate buffer. After permeabilization, the cells were in-

cubated with terminal deoxynucleotidyl transferase and fluorescein-

conjugated dUTP for 20 min at 37�C for nick-end-labeling. After

washing, cells were analyzed on a FACSCaliber flow cytometer.

Statistical analysis. The Mann-Whitney U test (t test– like

for nonparametric data) was used to compare age-matched FIV-

infected cats with control cats for different parameters (e.g., B7

expression). For comparisons of parameters for different times of

infection, Kruskal-Wallis (analysis of variance for nonparametric

data) was used to determine that at least 1 group differed. Pairwise

comparisons were obtained with the Mann-Whitney U test. Differ-

ences were considered to be significant at P , :05.

Results

huCTLA4-Ig fusion protein binds to T cells from LNs of FIV-

infected cats. B7 costimulatory molecules (B7.1 and B7.2)

normally are not expressed on large numbers of T cells but are

up-regulated on T cells activated in vitro (and in vivo under some

conditions of chronic immune stimulation) [14, 29, 30]. To de-

termine whether FIV infection, which is manifested as a chronic

low-level viremia [50], induces the expression of B7 on LN T

cells, we analyzed by 3-color flow cytometry the binding of PE-

conjugated huCTLA4-Ig on CD8þ and CD4þ cells from cats in-

fected with FIV for 4 weeks to 10 years and from age-matched

control cats. Figure 3A is a representative flow cytometry histo-

gram demonstrating increased binding of the huCTLA4-Ig fusion

protein to CD8þ and CD4þ LN cells from an FIV-infected cat,

compared with an age-matched healthy cat. Analysis of data from

a large number of cats indicated a progressive increase with time

of infection in the percentage of LN T cells, in particular CD8þ

cells, that bind the huCTLA4-Ig fusion protein (figure 3B). In

cats infected for ,6 months (n ¼ 9), the percentage of CD8þ

cells binding the huCTLA4-Ig fusion protein ranged from 6%

to 26%, which is significantly (P ¼ :002) increased over bind-

ing to LN CD8þ cells from age-matched controls (n ¼ 11),

which ranged from 0.4% to 10%. By 1–5 years after infection

(n ¼ 30), the percentage of CD8þ cells binding huCTLA4-Ig

ranged from 6.8% to 52%, again significantly (P ¼ :0006) in-

creased over binding to LN CD8þ cells of age-matched controls

(n ¼ 8; 2.6%–16.5%). The majority of LN CD8þ cells from cats

infected long term (8–10 years) that were displaying signs of

AIDS (n ¼ 5) bound huCTLA4-Ig on their surface (62%–95%),

which was significantly higher (P ¼ :02) than huCTLA4-Ig

binding to CD8þ cells from LNs of the 4 age-matched controls

(9%–18%). There was also a trend toward increased binding

of huCTLA4-Ig on CD4þ cells in FIV-infected cats that was sig-

nificantly different from control cats during the acute stage in-

fection (P ¼ :0008) and long-term infection (P ¼ :02) but not

the asymptomatic stage (1–5 years after infection; figure 3B).

In comparing acute with asymptomatic and asymptomatic with

long-term infections, the increased binding of huCTLA4-Ig was

also significant with time after infection for both CD8þ (P , :05)

and CD4þ (P , :01) cells. Some of this increase may be age-

related, because there was a increase in binding of huCTLA4-Ig

over time in healthy cats. However, it was significant only from

5 years to 10 years for CD8þ cells (P ¼ :02) and from 6 months

to 5 years for CD4þ cells (P ¼ :01). Because B7.1 and B7.2 are the

only receptors for CTLA4 [51], we conclude that the huCTLA4-Ig

is binding to B7.1 and/or B7.2 on feline T cells and that these co-

stimulatory receptors are increasingly up-regulated on LN T cells

throughout the course of FIV infection. This progressive up-regu-

lation of B7 molecules over time is inversely correlated with LN

CD4þ:CD8þ ratios (figure 3C) and peripheral blood CD4þ and

CD8þ cell counts (figure 3D).

Expression of B7.1 and B7.2 is increased on T cells from FIV-

infected cats. Although huCTLA4-Ig binding indicated that B7

was expressed at a high frequency on CD8þ and CD4þ LN cells

from FIV-positive cats, we were not able to determine whether

one or both B7 molecules was up-regulated. However, rabbit poly-

clonal antibodies specific for cloned feline B7.1 and B7.2 became

available to us during the course of this study, which allowed for

specific measurements of B7.1 and B7.2 on LN T cells. Figure 4

demonstrates increased binding of both rabbit anti-B7.1 and anti-

B7.2 on LN CD8þ and CD4þ T cells from an FIV-infected cat,

compared with an age-matched uninfected cat. Specificity of anti-

B7.1 and anti-B7.2 binding was demonstrated by failure of the

appropriate rabbit preimmune serum to bind either CD4þ or

CD8þ cells from FIV-infected cats (figure 4A) and the fact that

pretreatment of cells with either anti-B7.1 or anti-B7.2 blocked

binding of huCTLA4-Ig (figure 4B).

Simultaneous analysis of staining by huCTLA4-Ig fusion pro-

tein, rabbit anti-B7.1, and rabbit anti-B7.2 on LN T cells from 9

cats infected with FIV (1–10 years after infection) and 10 age-

matched controls revealed that all 3 reagents bound a greater per-

centage of CD8þ cells from the FIV-infected cats than from the

control cats (figure 5A). The difference between FIV-infected and

healthy cats in binding of anti-B7.1, anti-B7.2, and huCTLA4-

Ig to CD4þ cells was less striking than that to CD8þ cells, because

normal CD4þ cells tend to have a higher background expression

of B7.1 than do CD8þ cells. Figure 5B compares the binding of

anti-B7.1 and anti-B7.2 to LN CD8þ and CD4þ cells from a large

number of FIV-infected cats (3 months to 10 years after infec-

tion) and age-matched controls. Again, there was a significantly

greater expression of B7.1 (P ¼ :0001) and B7.2 (P ¼ :0001)

on CD8þ cells from FIV-infected cats, compared with controls.

As observed in figure 5A, the binding of anti-B7.1 to FIV-infected

and control LN CD4þ cells was less striking and was not signifi-

cantly different. Although the increase in B7.2 expression on

CD4þ cells from the FIV-infected cats was not as dramatic as

that on the CD8þ cells, it was still significant (P ¼ :04), compared

with the control cats (figure 5B). These data collectively demon-


Dow



ecember 2021

Figure 3. Species–cross-reactive human recombinant cytotoxic T lymphocyte antigen 4 (CTLA4) protein fused to Fc of murine IgG (huC-TLA4-Ig) binds to feline lymph node (LN) T cells. Three-color flow cytometry with monoclonal allophycocyanin-conjugated anti–felineCD8, fluorescein-conjugated anti–feline CD4, and phycoerythrin (PE)–conjugated huCTLA4-Ig fusion protein was done on LN cells fromcats experimentally infected with feline immunodeficiency virus (FIV) and age-matched controls. A, Histogram demonstrating staining of huC-TLA-Ig fusion protein on CD8þ and CD4þ cells from a cat infected with FIV for 5 years (heavy line), compared with an age-matched control(light line). Higher percentage of CD8þ and CD4þ cells from FIV-infected cat bound fusion protein. B, Box-whisker plots demonstrating the in-crease in the percentage of T cells stained with huCTLA4-Ig fusion protein through course of FIV infection, compared with age-matched con-trols. Each box-whisker plot represents 5th and 95th percentiles (whiskers), 25th and 75th percentiles (box), and median (horizontal line dividingbox). Symbols indicate values for individual cats: percentage of huCTLA4-Ig–positive CD8þ or CD4þ cells of total CD8þ or CD4þ cells. Therewas a significant increase in the percentage of huCTLA4-Ig–positive cells in FIV-infected cats, compared with normal controls, for acute (,6months; CD8þ and CD4þ cells), asymptomatic (1–5 years; CD8þ cells), and long-term (8–10 years; CD8þ and CD4þ cells) infections. C, Box-whisker plot demonstrating lymph node CD4þ:CD8þ ratios of cats shown in panel B. There was an inverse correlation between CD4þ :CD8þ

ratios and huCTLA4-Ig binding: as the percentage of huCTLA4-Ig–positive CD8þ and CD4þ cells increased over time, CD4þ :CD8þ ratios de-creased. Ratios were significantly decreased, compared with normal cats, in asymptomatic and long-term infections. D, Peripheral blood CD4þ

and CD8þ lymphocyte numbers of FIV-infected cats shown in panel B. Similar to CD4þ:CD8þ ratios, there was an inverse correlation betweenhuCTLA4-Ig binding on lymph node T cells and peripheral blood T cell numbers. Data show mean þ SD of n cats. M, months; neg, negative;pos, positive; y, years.

1082

Dow



ecember 2021

Figure 4. Rabbit antibodies to cloned feline B7.1 and B7.2 bind feline lymph node (LN) T cells and block binding of species–cross-reactivehuman recombinant cytotoxic T lymphocyte antigen 4 (CTLA4) protein fused to Fc of murine IgG (huCTLA4-Ig). A, Three-color flow cytometrywas done with monoclonal anti–feline CD8, anti–feline CD4, and rabbit polyclonal antibodies to feline B7.1 or B7.2. Cells were stained with rabbitpolyclonal antibodies, followed by phycoerythrin (PE)–conjugated donkey anti-rabbit antibody, then with anti-CD8 and anti-CD4 antibodies, asdescribed in the legend to figure 3. There was a greater percentage of B7.1þ and B7.2þ CD8þ and CD4þ cells in LN cells of feline immunodefi-ciency virus (FIV)–infected cats (5 years after infection; heavy line) than from age-matched control cats (light line). B, Blocking of CTLA4-Igbinding by rabbit polyclonal anti-B7.1 and anti-B7.2. LN cells from FIV-infected cats (5 years after infection) were incubated with monoclonalantibodies to feline CD8 and CD4 as described in the legend to figure 3, followed by rabbit anti-B7.1 or -B7.2, followed by PE-conjugated huC-TLA4-Ig fusion protein (dashed line). Controls consisted of staining with huCTLA4-Ig fusion protein without pretreatment with rabbit serum(heavy line) or incubating with preimmune rabbit serum followed by CTLA4-Ig fusion protein (light line).

1083

Dow



ecember 2021

strate that both B7.1 and B7.2 are up-regulated on CD8þ and

CD4þ LN cells from FIV-infected cats and that the huCTLA4-

Ig fusion protein binds both feline B7.1 and B7.2.

To confirm that these anti-B7 antibodies were actually bind-

ing to CD4þ and CD8þ cells and not other cell types, such as

monocytes or macrophages, we performed 2-color flow cyto-

metric analysis with anti-CD14 (expressed on feline mono-

cytes and macrophages) and either anti-CD4 or anti-CD8 (by

use of an anti-CD8 b-chain MAb). We found no dual staining of

CD14 on either CD4þ or CD8þ LN cells (data not shown), indi-

cating that we were indeed detecting B7 on T cells. These data

also indicate that feline monocytes and macrophages do not ex-

press CD4, as has been reported for human monocytes [52]. The

observation that B7.1 and B7.2 were coexpressed with the CD8

b chain eliminated the possibility that we were detecting a sub-

set of dendritic cells that are CD8a/a positive [53].

In addition to surface expression of B7.1 and B7.2 costimula-

tory molecules, human and mouse T cells activated in vitro coex-

press MHC class II molecules [30, 31]. To determine whether

this was also true for B7þ T cells from FIV-infected cats, we per-

formed 2- and 3-color flow cytometry to colocalize MHC class II

molecules with B7.1 and/or B7.2 on CD4þ and CD8þ LN T cells.

Interestingly, the majority of T cells (>85%) from FIV-infected

as well as control cats expressed class II antigens on their surface

whether or not they were B7þ (data not shown). Others have re-

ported previously that the majority of T cells in the blood of FIV-

infected as well as naive control cats constitutively express MHC

class II antigen on their surface [54–57]. Thus, class II expression

Figure 5. A, Comparison of staining of lymph node (LN) cells by species–cross-reactive human recombinant cytotoxic T lymphocyte antigen4 (CTLA4) protein fused to Fc of murine IgG (huCTLA4-Ig), rabbit anti-B7.1, and rabbit anti-B7.2. LN cells from feline immunodeficiencyvirus (FIV)–infected cats (1–10 years after infection) and from age-matched controls were simultaneously analyzed for binding of huC-TLA4-Ig fusion protein, as described in the legend to figure 3, and for expression of B7.1 and B7.2, as described in the legend to figure 4. B,Analysis of large number of FIV-infected cats (3 months to 10 years after infection) and age-matched controls for expression of B7.1 andB7.2 on LN CD8þ and CD4þ cells. Box-whisker plots are as described in the legend to figure 3. Values represent the percentage of huC-TLA4-Igþ, B7.1þ, or B7.2þ cells of total CD8þ or CD4þ cells. Neg, negative; pos, positive.


Dow



ecember 2021

is not useful in distinguishing naive from activated T cells in

LNs of FIV-infected or control cats.

B7.1 and B7.2 expression on peripheral blood T cells from

FIV-infected cats. Bucci et al. [42] and Gebhard et al. [58] re-

cently described a subset of CD8þ cells expressing an activation

phenotype (CD8bloCD62LnegCD49dhiCD44hiCD18hi) in the blood

of FIV-infected cats. Similar to the B7.1 and B7.2 activation

phenotypes in LNs, the CD8blo phenotype in blood increased in

number as the FIV infection progressed. To determine whether

the B7.1þ and B7.2þ T cells were also present in the blood, and

whether they were associated with the previously described

CD8blo activation phenotype, 2-color flow cytometry was done

on PBMC from chronically FIV-infected cats with anti–CD8 b

chain and anti-B7.1 or anti-B7.2. B7.1þ peripheral blood CD8þ

cells from FIV-infected cats were significantly (P ¼ :0001) grea-

ter in number than in control cats (figure 6A) and were similar

to findings in LNs (figure 5A), whereas B7.2þCD8þ cells were

low in the blood of FIV-positive cats and were not significantly

different from control cats (figure 6A). Two-color flow cytometric

analysis with anti–CD8 b chain revealed that B7.1þCD8þ cells

in the blood were almost exclusively CD8blo (figure 6B), indica-

tive of an activation phenotype. Analysis of B7.1 and B7.2 recep-

tors on CD4þ cells from PBMC of FIV-positive and control cats

revealed a pattern similar to that observed on CD8þ cells (data

not shown). Thus, although CD8þB7.2þ and CD4þB7.2þ T cells

in FIV-positive cats are largely restricted to LNs, the B7.1þ T

cells have a significant presence in blood as well as in LNs, and

the B7.1þCD8þ cells in blood have an activation phenotype, as

indicated by decreased expression of the CD8 b chain [42, 58].

B7.1- and B7.2-expressing LN T cells undergo spontaneous

apoptosis. The LNs of FIV-infected cats and HIV-infected per-

sons are characterized by a high frequency of spontaneous T cell

apoptosis that correlates with the state of LN cell activation [7–

10]. Because an unusually large number of CD8þ and CD4þ cells

in the LNs of FIV-infected cats express B7.1 and B7.2, indicative

of an activation phenotype, we hypothesized that the apoptosis

seen in FIV-positive LNs is associated with B7-expressing cells.

To address this question, monoclonal anti-CD4 and anti-CD8

antibodies, the polyclonal rabbit anti-B7.1 and anti-B7.2 anti-

bodies, and a fluorescein-conjugated dUTP TUNEL assay were

used in 3-color flow cytometric analysis of LN T cells from FIV-

infected and control cats. Similar to findings of other researchers

[9, 10], we found that freshly isolated LN CD4þ and CD8þ cells

from FIV-infected cats tended to have a higher frequency of spon-

taneous apoptosis than did cells from uninfected cats (figure 7A).

In addition, T cells from FIV-infected cats cultured in vitro for 24

h showed increased apoptosis, compared with T cells from control

cats (figure 7A). Figure 7B represents a typical 3-color flow cyto-

metric analysis of LN CD4þ TUNEL-positive cells from an FIV-

infected cat and the B7.1 expression status. These data clearly

show that apoptosis was largely restricted to B7.1þ cells, as op-

posed to the B7.12 cells. Further examination of this relationship

in a number of FIV-infected and control cats revealed that apo-

ptosis was primarily a characteristic of B7.1þ and B7.2þ CD4þ

and CD8þ LN cells (figure 7C). Preliminary TUNEL analysis

of B cells revealed increased frequency of B cell apoptosis in

the LNs of FIV-infected cats that also was associated with B7.1

and/or B7.2 expression (data not shown). Interestingly, apoptosis

Figure 6. B7.1 and B7.2 expressed on CD8þ cells in peripheral blood. Three-color flow cytometric analysis was done as described in the leg-end to figure 3 on feline peripheral blood mononuclear cells (PBMC). A, Expression of B7.1 and B7.2 on peripheral blood CD8þ cells fromchronically feline immunodeficiency virus (FIV)–infected (1–5 years after infection) and age-matched control cats. Box-whisker plots are asdescribed in the legend to figure 3. Values represent the percentage of B7.1þ or B7.2þ CD8þ cells of total CD8þ cells. There was a significantincrease in B7.1þCD8þ but not B7.2þCD8þ PBMC. B, Dot plot demonstrating B7.1 and CD8 b chain expression on PBMC from a cat infectedwith FIV for 3 years. B7.1 expression was largely confined to the CD8blo population, which we have previously demonstrated to be activationphenotype. Neg, negative; pos, positive.


Dow



ecember 2021

Figure 7. Apoptosis of lymph node (LN) cells and B7 expression. Three-color analysis was done for CD4 or CD8 expression and B7.1 or B7.2expression; apoptosis was measured by TUNEL assay. Cells were stained with rabbit anti-B7.1 or anti-B7.2 and phycoerythrin-conjugated anti-rabbit, followed by biotinylated anti-CD4 or anti-CD8, and developed with streptavidin-allophycocyanin. After permeabilization, cells werestained with fluorescein-dUTP. A, Analysis of total lymphocyte population for TUNEL-positive cells. TUNEL-positive cells were increasedin feline immunodeficiency virus (FIV)–infected cats (1–5 years after infection), compared with age-matched controls, whether cells were ana-lyzed immediately after harvesting LN or after 24 h of culture. Each bar represents 1 cat. B, Analysis of association of B7.1 expression andTUNEL-positive CD4þ cells from the LN of an FIV-infected cat. Histogram demonstrates TUNEL-positive CD4þ cells defined by their B7.1expression in dot plot above it. R2 (heavy line) represents CD4þB7.1þ cells; R3 (light line) represents CD4þB7.12 cells. TUNEL-positivecells are predominantly CD4þB7.1þ. C, Relationship of B7.1 and B7.2 expression and apoptosis on CD4þ and CD8þ cells from FIV-infectedand age-matched control cats. Bars represent mean þ SE of 3 control and 5 FIV-infected cats (1–5 years after infection).

Dow



ecember 2021

was also a feature of the relatively small numbers of B7.1þ and

B7.2þ CD4þ and CD8þ cells (figure 7C), as well as B cells, from

control cats, suggesting that B7 molecules expressed on activated

lymphocytes may be involved with the normal process of T and

B cell turnover in LN germinal centers.

CTLA4 is up-regulated on the surfaces of CD8 þ and CD4 þ T

cells from LNs of FIV-infected cats. Most studies with in vitro–

activated T cells examining B7 molecules suggest that T-T anti-

gen presentation leads not to costimulation but to anergy and

apoptosis via preferential engagement of CTLA4 [32, 59]. Be-

cause B7.1 and B7.2 are not known to transduce a proliferation

or anergy signal, as does CTLA4, one would predict that the B7þ

TUNEL-positive cells in LNs of FIV-infected cats should also be

positive for CTLA4. Measurement of cell surface–associated

CTLA4 in the absence of in vitro stimulation has been problematic,

because most of the protein is sequestered intracytoplasmically,

even in activated T cells, and the cell surface CTLA4 is rapidly

internalized [60]. However, analysis of freshly explanted, unstim-

ulated LN cells and PBMC by 2-color flow cytometry with a

rabbit polyclonal anti-CTLA4 revealed significant levels of

CTLA4 on the surface of a fraction of CD4þ and CD8þ cells

from FIV-infected cats. As illustrated by the representative

flow cytometry histogram in figure 8A, a significant number of

CD8þ cells and CD4þ cells from the LN of an FIV-infected

cat expressed surface CTLA4, whereas only a few cells from a

healthy cat expressed CTLA4. Analysis of a number of FIV-

infected cats with asymptomatic infection (1–5 years after infec-

tion) and age-matched control cats revealed a significant (P ,

:01) increase in CTLA4þCD4þ and CTLA4þCD8þ cells in both

blood and LNs of FIV-infected cats, compared with controls

(figure 8B).

Coexpression of CTLA4 and B7 molecules on LN T cells from

FIV-positive cats. Because the above documentation of B7.1þ,

B7.2þ, and CTLA4þ T cells was derived with rabbit polyclonal

antibodies for all 3 receptors, the data do not address the ques-

tion of cellular colocalization of CTLA4 with B7.1 and/or B7.2.

However, during the course of this study, we developed a MAb

to feline B7.1 (MAb B7.1.66; figure 2) to explore the question

of coexpression of B7.1 with B7.2 and CTLA4 on LN T cells

from FIV-positive cats. The dot plots and histograms in figure 9

illustrate the 3-color flow cytometric analysis for coexpression

of B7.1 and either B7.2 or CTLA4 on CD8þ and CD4þ LN T

Figure 8. Expression of cytotoxic T lymphocyte antigen 4 (CTLA4) on lymphocytes from feline immunodeficiency virus (FIV)–infectedcats. Three-color analysis was done as described in the legend to figure 3 with monoclonal anti–feline CD8 and anti–feline CD4 antibodiesand rabbit polyclonal anti–feline CTLA4 antibody. A, Histogram demonstrating increased expression of CTLA4 on lymph node T cells froma cat infected with FIV for 5 years, compared with an age-matched control. B, Box-whisker plots demonstrating an increase in the percentageof CTLA4þ T cells in lymph nodes (LNs) and blood of FIV-infected cats (1–5 years after infection), compared with age-matched controls.Box-whisker plots are as described in the legend to figure 3. Values represent the percentage of CTLA4þCD8þ or CTLA4þCD4þ cells oftotal CD8þ or CD4þ cells. There was a significant increase in the percentage of CTLA4þ T cells in both the blood and LNs of FIV-infectedcats. Neg, negative; PE, phycoerythrin; pos, positive.


Dow



ecember 2021

Figure 9. Coexpression of B7.1 on B7.2þ or cytotoxic T lymphocyte antigen 4–positive (CTLA4þ) feline lymph node T cells. Three-coloranalysis was done for CD4 or CD8 expression, B7.2 or CTLA4 expression, and B7.1 expression. Cells were stained with rabbit anti-B7.2 oranti-CTLA4 and fluorescein-conjugated anti-rabbit, followed by phycoerythrin-conjugated monoclonal anti–feline B7.1 and allophycocyanin-conjugated anti–feline CD4 or anti–feline CD8 antibodies. B7.2þCD8þ or B7.2þCD4þ cells (A) or CTLA4þCD8þ or CTLA4þCD4þ cells (B)from a feline immunodeficiency virus (FIV)–infected cat were gated (R4) and analyzed for B7.1 expression. Histograms demonstrate the per-centage of B7.12 (M1) and B7.1þ (M2) gated cells. C, Expression of B7.1 on B7.2þ or CTLA4þCD4þ or CTLA4þCD8þ lymph node cells.Bars represent mean þ SD of 4 cats infected with FIV for 3–5 years.

Dow



ecember 2021

cells from a single FIV-positive cat: 74.2% of CD8þB7.2þ

cells and 75.5% of CD4þB7.2þ cells were also B7.1þ (figure

9A), and 71.2% of CD8þCTLA4þ cells and 67.7% of CD4þ-

CTLA4þ cells were B7.1þ (figure 9B). Figure 9C illustrates

the data from 4 FIV-positive cats and clearly shows that the

majority of CTLA4þCD8þ and CTLA4þCD4þ cells from LNs

of FIV-positive cats coexpressed B7.1. Furthermore, because

the majority of B7.2þ T cells coexpress B7.1, it can be predicted

that a large fraction of B7.2þ cells coexpress CTLA4. Collec-

tively these data suggest that a significant proportion of acti-

vated CD8þ and CD4þ T cells in LNs of FIV-positive cats coex-

press B7.1, B7.2, and their ligand, CTLA4.

TUNEL-positive CD4þ and CD8þ LN T cells are CTLA4þ.

Because the majority of TUNEL-positive T cells reside in the

B7.1þ and B7.2þ fractions and because the majority of B7þ T

cells are CTLA4þ, it is likely that the B7þCTLA4þ cells are also

TUNEL-positive. To confirm this, we performed 3-color flow cy-

tometric analysis on LN CD4þ and CD8þ T cells. Figure 10A illus-

trates a typical 3-color analysis demonstrating that a large fraction

of CD4þCTLA4þ cells from an FIV-infected cat are TUNEL-

positive, whereas only a minor fraction of CD4þCTLA42 cells

are TUNEL-positive. Analysis of a number of FIV-infected and

control cats indicated that the CTLA4þCD4þ or CTLA4þCD8þ

cells from either LN or blood were the predominant TUNEL-

positive populations (figure 10B). Similar to the relationship be-

tween apoptosis and B7 expression, the small numbers of CTLA4þ

T cells in healthy cats were also TUNEL-positive. These data, in

conjunction with the data shown in figure 9, strongly argue that

TUNEL-positive cells in LNs of FIV-infected cats that are un-

dergoing apoptosis express both B7 molecules (B7.1 and B7.2)

and CTLA4 on their surface, supporting a model for bidirec-

tional T-T anergy and apoptosis.

Discussion

Progressive immune deterioration, a hallmark of HIV and FIV

infections, is manifested by the inability of CD4þ cells to produce

IL-2 and to proliferate in vitro in response to MHC class II–re-

stricted recall antigens and mitogens [6, 15], reminiscent of a

state of anergy [16]. Because activated anergic T cells frequently

progress to apoptosis in the absence of IL-2 [23], one would pre-

dict a high frequency of T cell apoptosis in lymphoid tissues of

individuals with these infections. Recent phenotypic analysis of

LN cells and PBMC from asymptomatic HIV-infected persons

and FIV-infected cats have provided strong evidence for elevated

levels of spontaneous T cell apoptosis [7–10]. The present study

demonstrated that T cell apoptosis in LNs of FIV-infected cats is

a characteristic of CD4þ and CD8þ cells expressing the B7.1 and

B7.2 costimulatory molecules and, at least in some cases, their

receptor, CTLA4. Because ligation of T cell CTLA4 with either

B7.1 or B7.2 transduces a signal to arrest T cells in G0/G1 of the

cell cycle, a prelude to anergy and apoptosis, we propose that

aberrant B7 expression on T cells in lymphoid tissues from FIV-

infected cats promotes inappropriate B7- and CTLA4-mediated

T cell ligations, resulting in anergy and apoptosis, and ultimately

the development of AIDS.

Flow cytometric analysis with the huCTLA4-Ig fusion protein,

as well as polyclonal antibodies specific for feline B7.1 and B7.2

costimulatory molecules, revealed a marked increase in the per-

centage of CD8þ cells expressing B7 molecules in LNs of FIV-

infected cats, compared with control cats. The percentages of

CD8þB7.1þ and CD8þB7.2þ cells increased progressively with

time after infection, such that the majority of CD8þ cells in LNs

of cats infected with FIV long term and with symptoms of AIDS

expressed B7 molecules. Increasing percentages of LN CD4þ

cells expressing B7.1 and B7.2 molecules were evident as the

FIV infection progressed.

Expression of B7.1 and B7.2 costimulatory molecules on a

large percentage of T cells is highly unusual, because, in a physio-

logic setting, they are normally restricted to APCs and provide the

necessary second costimulatory signals to T cells for regulating

the immune response [14, 17]. A number of recent studies indicate

that B7 molecules may be up-regulated on T cells activated in

vitro and on a subset of CD4þ and CD8þ T cells in patients with

autoimmune disease [26–29] or HIV infection [30, 31, 61].

Wolthers et al. [61] reported increased B7.1 and B7.2 expression

on in vitro–stimulated CD4þ and CD8þ T cells from HIV-in-

fected persons, compared with that on cells from control sub-

jects. Wyss-Coray et al. [30] reported expression of B7 protein

and mRNA on human CD8þ cells after antigen activation.

These authors also reported that B7 was present on a subset of

CD3þ cells from HIV-infected persons, whereas CD3þ cells

from healthy subjects did not express B7. Examination of a lim-

ited number of HIV-infected patients (n ¼ 12) revealed that

0.5%–20% of CD3þ T cells in PBMC expressed B7 molecules,

whereas none of the CD3þ cells of PBMC from healthy patients

(n ¼ 5) expressed B7. Kochli et al. [31] also reported that a

small subset (2%–5%) of both CD4þ and CD8þ cells in the

blood of HIV-infected persons expressed B7.1 and B7.2 on

their surfaces. We also observed an unusually high fraction of

B7-expressing CD4þ and CD8þ T cells in the blood of FIV-posi-

tive cats. However, our studies indicate that, although significant

numbers of B7.1þCD4þ and B7.1þCD8þ cells were present in

both LNs and blood of FIV-infected cats, B7.2þ T cells were,

for the most part, restricted to LNs. The discrepancy between

the percentage of B7.2þ T cells in LNs and blood could be ex-

plained by a preferential spontaneous apoptosis of B7.2þ T

cells rather than B7.1þ T cells in LNs. Our studies indicate that

all B7.2þ cells in LNs coexpress B7.1, whereas a significant frac-

tion of the B7.1þ cells are negative for B7.2. Thus, it is possible

that LN B7.1þB7.2þ T cells are more susceptible to apoptosis

and die before they can enter the blood. Studies have documen-

ted that spontaneous T cell apoptosis occurs at a much higher

frequency in LNs than in blood of HIV-infected patients [7, 62].

The significance of up-regulation of B7.1 on feline T cells must

be interpreted with caution. There are relatively large numbers of


Dow



ecember 2021

CD4þB7.1þ cells in LNs and blood of healthy cats, making it dif-

ficult to determine what effect FIV infection has on this phenotype.

A relative high frequency of CD4þ cells constitutively express-

ing B7.1 also appears to be a feature of human lymphoid tissue.

Kuiper et al. [63] reported that�30% of CD4þ cells in freshly ex-

planted, nonstimulated human PBMC constitutively expressed

B7.1 on their surface. It is possible that B7.1 gene expression is

not as tightly regulated on CD4þ cells as is B7.2, and the increased

expression of the CD4þB7.1þ phenotype is an artifact of the in

vitro manipulation of the lymphoid tissues. Alternatively, the rela-

tively high frequency of CD4þB7.1þ cells in LNs and blood of

healthy cats could reflect the fact that APC B7.1 is transferred to

the surface of CD4þ cells after APC-TCR engagement [64].

This B7 transfer appears to be fairly selective for B7.1 and CD4þ

T cells [64] and thus could explain the high background num-

bers of CD4þB7.1þ cells in the blood and LNs of both control

and FIV-infected cats. Additional studies must be necessary to

resolve this, including LN immunohistochemistry and in situ

mRNA analysis.

B7.1 and B7.2 expressed on APCs regulate immune responses

by engaging either of 2 TCRs, CD28 and CTLA4. CD28 isconsti-

tutively expressed on most T cells and transduces a signal for cyto-

kine (IL-2)production and cellproliferation, thus initiating the im-

mune response [14, 17]. CTLA4 is expressed only 2–3 days after

Tcellshavebeenactivated [20–22,60]and transducesa signal ac-

tivating the apoptosis pathway, thus effectively terminating the

immune response [20–23]. Our data demonstrate that a signifi-

cant number of CD4þ and CD8þ T cells in LNs of FIV-infected

cats constitutively express CTLA4 on their surfaces. The percen-

tages of CD4þ and CD8þ cells expressing CTLA4 on their sur-

faces increased with time after FIV infection but, in general,

were always lower than the percentages of B7.1þ and B7.2þ T

Figure 10. Apoptosis of CTLA4þ T cells. Three-color analysis was done for CD4, CD8, and CTLA4 expression, and apoptosis was measuredby TUNEL assay, as described in the legend to figure 7. A, Analysis of association of cytotoxic T lymphocyte antigen 4 (CTLA4) expression andTUNEL-positive CD4þ cells from the lymph node (LN) of a feline immunodeficiency virus (FIV)–infected cat. Histogram demonstratesTUNEL-positive CD4þ cells defined by their CTLA4 expression in the dot plot above it. R2 (heavy line) represents CD4þCTLA4þ cells; R3(light line) represents CD4þCTLA42 cells. TUNEL-positive cells are predominantly CD4þCTLA4þ. B, Relationship of CTLA4 expressionand apoptosis on T cells in the blood and LNs of 3 FIV-infected (1–5 years after infection) cats and 2 age-matched controls.


Dow



ecember 2021

cells. In addition, Steiner et al. [65] recently reported increased

numbers of CTLA4þCD4þ (6.4%) but not CTLA4þCD8þ PBMC

in HIV-infected patients, compared with control subjects (3.8%).

Coexpression analysis of B7.1 and/or B7.2 and CTLA4 on LN T

cells from FIV-infected cats revealed that a significant fraction

of CD4þ and CD8þ cells expressed both of the B7 molecules and

CTLA4. Vervilghen et al. [29] reported that a fraction of CD4þ

and CD8þ T cells in synovial fluid of patients with rheumatoid ar-

thritis expressed B7.1 and CTLA4 on their surface but did not de-

termine if the 2 receptors were coexpressed on the same cell. To

the best of our knowledge, this is the first report of freshly ex-

planted, nonstimulated T cells coexpressing the B7 costimulatory

molecules and their ligand CTLA4.

One would predict that the presence of activated LN T cells

expressing B7 and/or CTLA4 would result in frequent T-T in-

teractions mediated by B7.1-CTLA4 or B7.2-CTLA4 ligation

that would transduce a signal for anergy and apoptosis. Three-

color flow cytometric analysis of LN T cells with use of the

TUNEL assay for frequency of single-strand DNA breakage and

apoptosis supported this prediction. Spontaneous apoptosis of

CD4þ and CD8þ cells from LNs of FIV-infected cats tended to

occur at a higher frequency than in control cats and was largely

restricted to T cells expressing B7.1 or B7.2 and CTLA4. In addi-

tion, LN CD4þ and CD8þ cells from FIV-infected cats, as report-

ed with HIV-infected patients [7, 8], showed increased apoptosis

when cultured in vitro, even in the absence of mitogenic simula-

tion. These apoptotic cells were also characterized by expression

of B7.1 and/or B7.2 and CTLA4. Interestingly, spontaneous ap-

optosis also occurred with LN T cells from control cats, although

at a lower frequency than cells from FIV-infected cats. Control T

cell apoptosis was also associated with T cells expressing B7.1

and/or B7.2 and CTLA4. These observations raise the possibility

that B7þ T cells may mediate the normal process of activation-

induced cell death that regulates immune responses in germinal

centers. Thus, it is possible that the T cell immune dysfunction

associated with HIV and FIV infections represents viral conscrip-

tion of a normal T cell regulatory process.

How the chronic and progressive state of T cell activation is

maintained in LNs of FIV-infected cats is not known but some

how must be related to chronic antigenic stimulation. Because

the virus burden is generally considered to be low in cats with

asymptomatic FIV infection, and because apoptosis occurs in

CD4þ, CD8þ, and B cells, the mechanism must be indirect. Al-

though we did not address the question of viremia in this study,

Dean et al. [50] examined the FIV provirus burden in lympho-

cyte subsets and reported that ,0.1% of CD4þ and CD8þ cells

in PBMC and LNs from cats with chronic asymptomatic FIV in-

fection contained FIV provirus. Thus, it is extremely unlikely that

direct FIV infection could account for the large fraction of B7þ

cells or the apoptotic T cells in LNs of FIV-infected cats. In sup-

port of this, the number of lymphocytes from HIV-seropositive

blood that undergo spontaneous apoptosis when cultured in vitro

far exceeds the number of cells infected with HIV [7, 66]. Also,

histochemical analysis of LNs from HIV-infected patients and

simian immunodeficiency virus–infected macaques has revealed

that most apoptotic cells are not infected with virus but lie in close

proximity to infected cells or dendritic cells bearing trapped vi-

rions [8]. Consistent with these observations, the degree of T cell

apoptosis in LNs from HIV-infected patients does not correlate

with either the stage of disease or virus burden but does strongly

correlate with the degree of activation of LN cells [7].

The most straightforward interpretation of our data supports a

model of chronic LN T cell anergy and apoptosis resulting from

interactions between T cells expressing B7.1 or B7.2 and their

negative signaling ligand CTLA4. Because both B7.1 and B7.2

bind CTLA4 with 20–100-fold higher avidity than they bind

CD28 [14], T-T interactions would favor T cell anergy and apo-

ptosis rather than T cell costimulation. Whatever the mechanisms

involved in chronic up-regulation of B7.1 or B7.2 and CTLA4

on LN T cells from FIV-infected cats, and the associated apopto-

sis, it will be important to develop in vitro T-T coculture systems

capable of analyzing B7.1-CTLA4 and B7.2-CTLA4 interac-

tions. Such a coculture model will be necessary to investigate the

role of virus, as well as the possible MHC restriction and antigen

specificity of the hypothesized T-T induced anergy and apopto-

sis. These studies should provide significant insights into cellular

and molecular mechanisms leading to immunodeficiency in FIV-

infected cats and, by inference, HIV-infected patients.

Acknowledgments

We thank Debra Anderson for excellent technical assistance

and Maria Correa for statistical analysis.

References

1. Bendinelli M, Pistello M, Matteucci D, et al. Small animal model of AIDS

and the feline immunodeficiency virus. Adv Exp Med Biol 1993;335:

189–202.

2. Levy JA. Pathogenesis of human immunodeficiency virus infection. Mi-

crobiol Rev 1993;57:183–289.

3. Ascher MS, Sheppard HW. AIDS as immune system activation. II. The pan-

ergic imnesia hypothesis. J Acquir Immune Defic Syndr 1990;3:177–91.

4. Golding H, Shearer GM, Hillman K, et al. Common epitope in human

immunodeficiency virus (HIV) I–GP41 and HLA class II elicits immuno-

suppressive autoantibodies capable of contributing to immune dysfunc-

tion in HIV I–infected individuals. J Clin Invest 1989;83:1430–5.

5. Siliciano R, Lawton T, Knall C, et al. Analysis of host-virus interactions in

AIDS with anti-gp120 T cell clones: effect of HIV sequence variation and

a mechanism for CD4þ cell depletion. Cell 1988;54:561–75.

6. Miedema F. Immunological abnormalities in the natural history of HIV in-

fection: mechanisms and clinical relevance. Immunodefic Rev 1992;3:

173–93.

7. Muro-Cacho C, Pantaleo G, Fauci A. Analysis of apoptosis in lymph

nodes of HIV-infected persons: intensity of apoptosis correlates with

the general state of activation of the lymphoid tissue and not with stage

of disease or viral burden. J Immunol 1995;154:5555–66.

8. Finkel T, Tudor-Williams G, Banda N, et al. Apoptosis occurs predomi-


Dow



ecember 2021

nantly in bystander cells and not in productively infected cells of HIV- and

SIV-infected lymph nodes. Nat Med 1995;1:129–34.

9. Guiot A, Rigal D, Chappuis G. Spontaneous programmed cell death (PCD)

process of lymphocytes of FIV-infected cats: cellular targets and modu-

lation. Vet Immunol Immunopathol 1997;58:93–106.

10. Sarli G, Della Salda L, Zaccaro L, Bendinelli M, Piedimonte G, Marcato P.

Apoptotic fraction in lymphoid tissue of FIV-infected SPF cats. Vet

Immunol Immunopathol 1998;64:33–44.

11. Budihardjo I, Oliver H, Lutter M, Luo X, Wang X. Biochemical pathways

of caspase activation during apoptosis. Annu Rev Cell Dev Biol 1999;

15:269–90.

12. Wallach D, Varfolomeev E, Malinin N, Goltsev Y, Kovalenko A, Boldin

M. Tumor necrosis factor receptor and Fas signaling mechanisms. Annu

Rev Immunol 1999;17:331–67.

13. Boise L, Noel P, Thompson C. CD28 and apoptosis. Curr Opin Immunol

1995;7:620–5.

14. Greenfield E, Nguyen K, Kuchroo V. CD28/B7 costimulation: a review.

Crit Rev Immunol 1998;18:389–418.

15. Lawrence CE, Callanan JJ, Jarrett O. Decreased mitogen responsiveness

and elevated tumor necrosis factor production in cats shortly after feline

immunodeficiency virus infection. Vet Immunol Immunopathol 1992;

35:51–9.

16. Schwartz RH. T cell clonal anergy. Curr Opin Immunol 1997;9:351–7.

17. Harris N, Ronchese F. The role of B7 costimulation in T-cell immunity.

Immunol Cell Biol 1999;77:304–11.

18. Gimmi C, Freeman G, Gribben J, Gray G, Nadler L. Human T-cell clonal

anergy is induced by antigen presentation in the absence of B7 costimu-

lation. Proc Natl Acad Sci USA 1993;90:6586–90.

19. Boussiotis V, Gribben J, Freeman G, Nadler L. Blockade of the CD28 co-

stimulatory pathway: a means to induce tolerance. Curr Opin Immunol

1994;6:797–807.

20. Bluestone JA. Is CTLA-4 a master switch for peripheral T cell tolerance?

J Immunol 1997;158:1989–93.

21. Karandikar NH, Vanderlugt CT, Walunas TL, Miller SD, Bluestone JA.

CTLA-4: a negative regulator of autoimmune disease. J Exp Med 1996;

184:783–8.

22. Walunas T, Lenschow DJ, Bakker CY, et al. CTLA-4 can function as a

negative regulator of T cell activation. Immunity 1994;1:405–13.

23. Scheipers P, Reiser H. Fas-independent death of activated CD4þ T lym-

phocytes induced by CTLA-4 crosslinking. Proc Natl Acad Sci USA

1998;95:10083–8.

24. Gimmi C, Freeman G, Gribben J, et al. B-cell surface antigen B7 provides

a costimulatory signal that induces T cells to proliferate and secrete in-

terleukin 2. Proc Natl Acad Sci USA 1991;88:6575–9.

25. Freedman AS, Freeman GJ, Rhynhart K, Nadler LM. Selective induction

of B7/BB-1 on interferon-gamma stimulated monocytes: a potential

mechanism for amplification of T cell activation through the CD28

pathway. Cell Immunol 1991;137:429–37.

26. Georgescu L, Vakkalanka R, Elkon K, Crow M. Interleukin-10 promotes

activation-induced cell death of SLE lymphocytes mediated by Fas

ligand. J Clin Invest 1997;100:2622–33.

27. Folzenlogen D, Hofer MF, Leung DY, Freed JH, Newell MK. Analysis of

CD80 and CD86 expression on peripheral blood B lymphocytes reveals

increased expression of CD86 in lupus patients. Clin Immunol Immuno-

pathol 1997;83:199–204.

28. Ranheim E, Kipps T. Elevated expression of CD80 (B7/BB1) and other

accessory molecules on synovial fluid mononuclear cell subsets in rheu-

matoid arthritis. Arthritis Rheum 1994;37:1637–46.

29. Vervilghen J, Lovis R, De Boer M, et al. Expression of functional B7 and

CTLA4 on rheumatoid synovial T cells. J Immunol 1994;153:1378–85.

30. Wyss-Coray T, Mauri-Hellweg D, Baumann K, Bettens F, Grunow R,

Pichler W. The B7 adhesion molecule is expressed on activated human

T cells: functional involvement in T-T cell interactions. Eur J Immunol

1993;23:2175–80.

31. Kochli C, Wendland T, Frutig K, Grunow R, Merlin S, Pichler W. CD80

and CD86 costimulatory molecules on circulating T cells of HIV in-

fected individuals. Immunol Lett 1999;65:197–201.

32. Greenfield E, Howard E, Paradis T, et al. B7.2 expressed by T cells does

not induce CD28-mediated costimulatory activity but retains CTLA4

binding: implications for induction of antitumor immunity to T cell tu-

mors. J Immunol 1997;158:2025–34.

33. Lombardi G, Sidhu S, Dodi T, Batchelor R, Lechler R. Failure of corre-

lation between B7 expression and activation of interleukin-2–secreting

T cells. Eur J Immunol 1994;24:523–30.

34. English RV, Johnson CM, Gebhard DH, Tompkins MB. In vivo lymphocyte

tropism of feline immunodeficiency virus. J Virol 1993;67:5175–86.

35. English RV, Nelson P, Johnson CM, Nasisse M, Tompkins WAF, Tompkins

MB. Development of clinical disease in cats experimentally infected

with feline immunodeficiency virus. J Infect Dis 1994;170:543–52.

36. Tompkins MB, Nelson PD, English RV, Novotney C. Early events in the

immunopathogenesis of feline retrovirus. J Am Vet Med Assoc 1991;

199:1311–5.

37. Bucci J, English R, Jordan H, Childers TA, Tompkins MB, Tompkins

WAF. Mucosally transmitted feline immunodeficiency virus induces a

CD8þ antiviral response that correlates with reduction of cell-associated

virus. J Infect Dis 1998;177:18–25.

38. Levy JK, Ritchey JW, Rottman JB, et al. Elevated interleukin-10–to–inter-

leukin-12 ratio in feline immunodeficiency virus–infected cats predicts

loss of type 1 immunity to Toxoplasma gondii. J Infect Dis 1998;178:

503–11.

39. Davidson MG, Rottman J, English RV, Lappin MR, Tompkins MB. Feline

immunodeficiency virus predisposes cats to acute generalized toxoplas-

mosis. Am J Pathol 1993;143:1486–97.

40. Tompkins MB, Ogilvie GK, Franklin RA, Kelley KW, Tompkins WA.

Induction of IL-2 and lymphokine activated killer cells in the cat. Vet

Immunol Immunopathol 1987;16:1–10.

41. Tompkins MB, Gebhard DH, Bingham HR, Hamilton MJ, Davis WC,

Tompkins WAF. Characterization of monoclonal antibodies to feline

T lymphocytes and their use in the analysis of lymphocyte tissue distri-

bution in the cat. Vet Immunol Immunopathol 1990;26:305–17.

42. Bucci J, Gebhard D, Childers TA, English R, Tompkins MB, Tompkins

WAF. The CD8þ phenotype mediating antiviral activity in feline immu-

nodeficiency virus–infected cats is characterized by reduced surface

expression of the CD8 b chain. J Infect Dis 1998;178:968–77.

43. Choi I, Hash S, Collisson E. Molecular cloning and expression of feline

CD28 and CTLA-4 cDNA. Vet Immunol Immunopathol 2000;76:45–59.

44. Choi I, Hash S, Winslow B, Collisson E. Sequence analyses of feline

B7 costimulatory molecules. Vet Immunol Immunopathol 2000;73:

219–31.

45. Massung R, Moyer R. The molecular biology of swinepox virus. I. A

characterization of the viral DNA. Virology 1991;180:347–54.

46. Mackett M, Smith G. Vaccinia virus expression vectors. J Gen Virol

1986;67:2067–82.

47. Chakrabarti S, Brechling K, Moss B. Vaccinia virus expression vector: co-

expression of beta-galactosidase provides visual screening of recombi-

nant virus plaques. Mol Cell Biol 1985;5:3403–9.

48. Broder C, Earl P. Recombinant vaccinia viruses: design, generation, and

isolation. Mol Biotechnol 1999;13:223–45.

49. Jenh C, Zhang M, Wiekowski M, et al. Development of a CD28 receptor

binding–based screen and identification of a biologically active inhibi-

tor. Anal Biochem 1998;256:47–55.

50. Dean GA, Reubel GH, Boore PF, Pedersen NC. Proviral burden and infec-

tion kinetics of feline immunodeficiency virus in lymphocyte subsets of

blood and lymph node. J Virol 1996;70:5165–9.

51. Borriello F, Sethna M, Boyd S, et al. B7-1 and B7-2 have overlapping, criti-


Dow



ecember 2021

cal roles in immunoglobulin class switching and germinal center for-

mation. Immunity 1997;6:303–13.

52. Hume D, Allan W, Hogan P, Doe W. Immunohistochemical character-

isation of macrophages in human liver and gastrointestinal tract: expres-

sion of CD4, HLA-DR, OKM1, and the mature macrophage marker 25F9

in normal and diseased tissue. J Leukoc Biol 1987;42:474–84.

53. Maldonado-Lopez R, De Smedt T, Michel P, et al. CD8aþ and CD8a2

subclasses of dendritic cells direct the development of distinct T helper

cells in vivo. J Exp Med 1999;189:587–92.

54. Byrne K, Kim H, Chew B, Reinhart G, Hayek M. A standardized gating

technique for the generation of flow cytometry data for normal canine

and normal feline blood lymphocytes. Vet Immunol Immunopathol

2000;73:167–82.

55. Neefjes J, Hensen E, de Kroon TI, Ploegh HL. A biochemical characteriza-

tion of feline MHC products: unusually high expression of class II anti-

gens on peripheral blood lymphocytes. Immunogenetics 1986;23:341–7.

56. Ohno K, Watari T, Goitsuka R, Tsujimoto H, Hasegawa A. Altered sur-

face antigen expression on peripheral blood mononuclear cells in cats

infected with feline immunodeficiency virus. J Vet Med Sci 1992;54:

517–22.

57. Rideout B, Moore P, Pedersen N. Persistent upregulation of MHC class II

antigen expression on T-lymphocytes from cats experimentally infected

with feline immunodeficiency virus. Vet Immunol Immunopathol 1992;

35:71–81.

58. Gebhard DH, Dow JL, Childers TA, Alvelo JI, Tompkins MB, Tompkins

WAF. Progressive expansion of an L-selectin–negative CD8 cell with

anti– feline immunodeficiency virus (FIV) suppressor function in the

circulation of FIV-infected cats. J Infect Dis 1999;180:1503–13.

59. Lombardi G, Hargreaves R, Sidhu S, et al. Antigen presentation by T cells

inhibits IL-2 production and induces IL-4 release due to altered cognate

signals. J Immunol 1996;156:2769–75.

60. Alegre M, Noel P, Eisfelder B, et al. Regulation of surface and intracellu-

lar expression of CTLA4 on mouse T cells. J Immunol 1996;157:4762–

70.

61. Wolthers K, Otto S, Lens S, et al. Increased expression of CD80, CD86,

and CD70 on T cells from HIV-infected individuals upon activation in

vitro: regulation by CD4þ T cells. Eur J Immunol 1996;26:1700–6.

62. Gougeon ML, Laurent-Crawford AG, Hovanessian AG, Montagnier

L. Direct and indirect mechanisms mediating apoptosis during HIV in-

fection: contribution to in vivo CD4 T cell depletion. Semin Immunol

1993;5:187–94.

63. Kuiper H, Brouwer M, Linsley P, van Lier RA. Activated T cells can induce

high levels of CTLA-4 expression on B cells. J Immunol 1995;155:

1776–83.

64. Sabzevari H, Kantor J, Jaigirdar A, et al. Acquisition of CD80 (B7-1) by T

cells. J Immunol 2001;166:2505–13.

65. Steiner K, Waase I, Rau T, Dietrich M, Fleischer B, Broker B. Enhanced

expression of CTLA-4 (CD152) on CD4þ T cells in HIV infection. Clin

Exp Immunol 1999;115:451–7.

66. Lewis DE, Tang DSN, Adu-Oppong A, Schober W, Rodgers JR. Anergy

and apoptosis in CD8þ cells from HIV-infected persons. J Immunol

1994;153:412–20.


Dow



ecember 2021

Documents

Feline Immunodeficiency Virus Infection Is Characterized by B7+