Progressive Expansion of an L-Selectin–Negative CD8 Cell with Anti–Feline
Immunodeficiency Virus (FIV) Suppressor Function in the Circulation
of FIV-Infected Cats
Douglas H. Gebhard,aJanet L. Dow, Tedd A. Childers,
Jose I. Alvelo, Mary B. Tompkins,
and Wayne A. F. Tompkins
Department of Microbiology, Pathology, and Parasitology, College
of Veterinary Medicine, North Carolina State University, Raleigh
The acute stage of feline immunodeficiency virus (FIV) infection is characterized by the
appearance of a major CD8 subpopulation with reduced expression of the CD8 b chain
(CD8a+blo). CD8 antiviral activity was subsequently shown to be mediated by the CD8a+blo
phenotype, which is the dominant CD8 phenotype in long-term infected cats. Two- and three-
color flow cytometric analysis demonstrated that the CD8a+blosubset is L-selectin negative
(CD62L?) and has increased expression of CD44, CD49d, and CD18, consistent with an
activation phenotype. The CD8a+bloCD62L?cells but not the CD8a+bhiCD62L+cells dem-
onstrated strong antiviral activity in the FIV acute-infection assay. The progressive expansion
of the CD8a+bloCD62L?effector subset cells in FIV-infected cats parallels that seen in human
immunodeficiency virus (HIV)–infected patients, suggesting that failure in homeostatic mech-
anisms regulating lymphocyte activation or trafficking (or both) may be a consequence of
both HIV and FIV infections.
Feline immunodeficiency virus (FIV), similar to human im-
munodeficiency virus (HIV), induces a CD8 T cell lymphocy-
tosis during the acute stage of infection that persiststhroughout
the asymptomatic stage of disease [1–3]. In both FIV and HIV
infections, CD8 lymphocytosis correlates with the emergence
of CD8 cells that provide antiviral immunity through both cy-
totoxic and virus-suppressive mechanisms [4–8].
CD8 cells that mediate immunity through cytotoxic or sup-
pressor mechanisms can be subdivided by differential expres-
sion of surface markers. Landay et al.  demonstrated that
HLA-DR and CD38 expression on CD8 lymphocytes cor-
relates with high antiviral activity. However, the CD8?
CD38?phenotype is also associated with high viremia, high
provirus burden, and disease progression [10, 11]. In contrast,
long-term nonprogressor patients possess high numbers of
CD8?HLA-DR?CD38?cells . Of interest, the CD8?
HLA-DR?CD38?phenotype that predominates in long-term
nonprogressor patients is noncytotoxic, raising the possibility
that these cells reduce viremia by a suppressive mechanism.
Barker et al.  demonstrated that the predominant HIV sup-
pressor cell in HIV-infected patients was a CD8?CD25?
Received 25 March 1999; revised 13 July 1999; electronically published
8 October 1999.
Grant support: NIH (AI-38177, AI-43858).
aPresent affiliation: Central Research Division, Pfizer,Inc.,EasternPoint
Rd., Bldg. 126, Groton, CT 06340-8020.
Reprints or correspondence: Dr. Wayne A. Tompkins, College of Veter-
inary Medicine, North Carolina State University, 4700 Hillsborough St.,
Raleigh, NC 27606 (email@example.com).
The Journal of Infectious Diseases
? 1999 by the Infectious Diseases Society of America. All rights reserved.
CD28?phenotype and that CD28 costimulation induced up-
regulation of CD25 expression and increased suppressor
Rabin et al.  recently made the interesting observation
that HIV-infected children had a marked decrease in the cir-
culating naive CD8 subset (CD8?CD62L?CD11aloCD45RAhi)
and a reciprocal increase in atypicalCD8Tcells,predominantly
of the CD8?CD62L?CD11ahiCD45RAlophenotype. In a sec-
ond study, Roederer et al.  demonstrated that the well-
described increase in CD8 T cells in adults with HIV infection
was primarily due to a progressive expansion of at least three
CD8 subsets, which were all distinct from CD8 naive cells on
the basis of the expression of high levels of the integrin CD11a.
These CD11ahisubsets, including the CD8?CD62L?CD11ahi
major subpopulation, represented ∼80% of the total CD8 cells
in the circulation of patients with long-term HIV infections.
Others have reported that the increase in CD8 cells inasymp-
tomatic HIV-infected individuals occurs within subsets that are
phenotypically atypical (e.g.,CD28?,CD38?,CD57?,CD11b?,
HLA-DR?, CD11ahi, CD45RO?/RA?, and CD62L?). Recent
studies by Zimmerman et al.  and Hamann et al.  dem-
onstrated that CD8 effector cells could be distinguished from
both naive CD8 and memory CD8 cells on the basis of the loss
of surface L-selectin (CD62L) expression. This suggests that
the major atypical CD8?CD62L?CD11ahipopulations induced
by HIV infection is not a memory phenotype but rather an
effector or activation phenotype. In support of this, a number
of authors have reported that the CD8?CD62L?activation
phenotype, in contrast to the CD8?CD62L?memory pheno-
type, in murine models also had increased expression of the
integrins CD49d and CD11b [16, 18]. This persistent high level
by guest on December 26, 2015
1504Gebhard et al. JID 1999;180 (November)
of cats during acute stage of feline immunodeficiency virus (FIV) in-
fection. Four cats were infected intravenously with FIV-NCSU1, and
blood was obtained at intervals for analysis of CD62L expression on
CD8 T cells by two-color flow cytometry using monoclonal antibodies
(MAb) 117 (anti-CD8b) and Leu-8 (anti-CD62L). Two uninfected cats
served as controls. y axis represents percentage of CD8?CD62L?cells
of total CD8b?cells. Similar percentage of CD8?CD62L?and
CD8?CD62L?cells were obtained when anti–CD8 a chain antibody
(MAb 357) was used. POS, positive; NEG, negative.
Development of CD8?CD62L?cells in peripheral blood
of activated CD8 cells in the circulation is a highly unusual
homeostatic condition: in healthy individuals, naive T cells are
present in high frequency in the circulation, whereas activated
CD8 cellsarevirtually absentexceptatthetimeofpeakimmune
Phenotypic analysis of peripheral blood mononuclear cells
(PBMC) suggests that a similar marked and progressive alter-
ation in the distribution of CD8 subpopulations occurs in FIV-
infected cats. Lehmann etal. and Willettetal.described
in FIV-infected cats an early and sustained expansion of an
atypical CD8 subpopulation with reduced expression of the
CD8 molecule and increased expression of major histocom-
patibility complex (MHC) class II molecules. Shimojima et al.
 subsequently confirmed these observations and demon-
strated that the loss of the CD8 molecule on PBMC from FIV-
infected cats was specific to the CD8 b chain (CD8a?blo).These
authors did not attribute any function to the CD8a?bloMHC
Bucci et al.  also reported a progressive increase of an
atypical CD8a?blophenotype with increased expression of
MHC class II molecules in FIV-infected cats and demonstrated
that this CD8 subpopulationpossessedpotentanti-FIVactivity.
In this study, we explored the possibility that the CD8a?blo
anti-FIV phenotype is the counterpart of the CD8?CD
62L?CD11ahiCD45RAloeffector cell found in the circulationof
pediatric and adult HIV patients [14, 15]. We analyzed by flow
cytometry the expression of various activation markers onCD8
cells from cats at different stages of FIV infection.
Materials and Methods
Animals and infection.
pathogen–free (SPF) cats obtained from Liberty Laboratories(Lib-
erty Corners, NJ) and ranged in age from 6 months to 9 years at
the time samples were collected. FIV-infected cats were inoculated
with the NCSU1isolate of FIV, which has been described in detail
elsewhere [1, 24, 25], either intravenously with cell-freevirusculture
supernatant ( TCID50) as described by Davidson et al. 6?10
or intravaginally with cell-associated virus as described by Bucci
et al. . All cats were ∼6 months of age when inoculated with
FIV and became antibody positive as determined by ELISA (ID-
EXX, Westbrook, ME) and provirus positive as determined by
polymerase chain reaction between 4 and 8 weeks after infection.
As reported by Bucci et al. , intravaginally and intravenously
infected cats behaved identically with regard to provirus load in
PBMC, seroconversion, and lymphocyte phenotype changes and
therefore are not separated in the data analysis presented herein.
Control cats ranged in age from 1 to 6 years: 6 cats were 1–2 years
old, 5 were 2–3 years old, and 6 were 3–6 years old.
Whole blood was obtained by jugular ve-
nipuncture into tubes with sodium citrate anticoagulant prior to
and at intervals after FIV infection and used for complete blood
cell counts and flow cytometry. Peripheral lymph nodes were ob-
tained from 12 control cats and 12 FIV-infected cats 24 weeks after
infection, as described by Levy et al. . In brief, cats were an-
esthetized with intravenous ketamine and diazepam and main-
tained with inhalant anesthetic (isoflurane). Either a prescapular
or popliteal lymph node was excised, and butorphanol tartratewas
administered to control postoperative discomfort. Single-cell sus-
pensions of lymph node cells were prepared for subsequent flow
cytometric analysis by gently passing thetissuethroughasteelmesh
We identified feline T cells using monoclonal an-
tibodies (MAbs; developed in our laboratory) to CD4, CD8, and
pan T cell epitopes [23, 28]. MHC class II molecules were detected
with MAb 42.3H2 raised against feline MHC class II antigens .
Two anti-CD8 MAbs were used: MAb 3.357 , which recognizes
the CD8 a chain , and MAb 117, which recognizes the CD8 b
chain [23, 30]. CD62L was identified by use of a cross-reactiveanti-
human MAb (Leu8; Becton Dickinson, San Jose, CA). Other anti-
human CD MAbs found to be cross-reactivewithfelinelymphocyte
surface activation markers were used: anti-CD44 (Caltag Labo-
ratories, Burlingane, CA), anti-CD49d (VLA-4; Becton Dickin-
son), and anti-CD18 (Caltag). A number of other anti-human CD
The cats used in this study were specific
by guest on December 26, 2015
JID 1999;180 (November)CD8?CD62L?Cells in FIV-Infected Cats1505
throughout course of experimental feline immunodeficiencyvirus(FIV)
infection. Peripheral blood mononuclear cells from cats infected with
FIV for varying lengths of time were analyzed by two-color flow cy-
tometry for expression of CD62L on CD8 T cells. Each box and
whisker plot represent 5th and 95th percentiles (whiskers), 25th and
75th percentiles (box), and median (middle line). ?, values for indi-
vidual cats: seronegative,(1–2 year,
n = 17
year, ); seropositive, (6 months after infection,
n = 6
n = 63
years after infection,; and ?7 years after infection,
n = 23
was significant change in expression of CD62L on CD8 cells among 4
groups of cats ( ).
P ! .001
Increase in percentage of CD8?CD62L?cells continues
; 2–3 year,; 3–6
n = 6
n = 5
n = 31
n = 9). There
MAbs, including CD11a, CD11b, CD45RA, and CD45RO (all
from Caltag), were nonreactive with feline lymphocytes.
Two-color flow cytometric analysis as de-
scribed by Davidson et al.  was used to determine the presence
of CD4, CD8, and B cell populations in both whole blood samples
and single cell suspensions of lymph nodes. CD8 cells were also
monitored for the development of distinct subpopulations as de-
termined on the basis of fluorescence intensity of staining by an-
tibodies specific for the CD8 a chain (MAb 357) or the CD8 b
chain (MAb 117), as described by Bucci et al. . With the ex-
ception of the cells from the long-term (?7 years) FIV-infected
cats, the percentage of CD8?cells that stained with MAb 117 and
the percentage that stained with MAb357 were not significantly
different. Thus, the percentages of CD8bloand CD8bhicells were
calculated on the basis of total CD8b?cells as determined by stain-
ing with MAb 117. Data were acquired by use of FACSCalibur or
FACScan flow cytometer (Becton Dickinson). The FACSCalibur,
with a helium-neon laser as the second excitation source, was used
for samples stained with antigen-presenting cells. For all samples,
data were obtained from at least 15,000 cells and stored in list-
mode fashion for subsequent analysis. Gated data were then gen-
erated for fluorescent analysis oflymphocytesasdefinedbyforward
and side scatter. Care was taken to perform all flow cytometric
analyses as soon as possible after cat blood was obtained to avoid
any loss of surface CD62L as a result of activation of the protease
responsible for cleaving CD62L from the surface .
Anti-FIV activity of CD8a?bloCD62L?and CD8a?bhiCD62L?
FIV suppressor activity of CD8 cells in FIV-infected cats
was assessed in the FIV acute-infection assay as described by Bucci
et al. . Percoll-separated PBMC were purified into CD
8a?bhiCD62?and CD8a?bloCD62?subpopulations (?98% purity)
by flow cytometry, using a Cytomotion MoFlo Sorter (Flow Cy-
tometry Facility, University of North Carolina, Chapel Hill, NC).
The two subsets were separated on the basis of relative mean index
ate–conjugated MAb 117 (anti-CD8 b chain) and phycoerythrin-
conjugated MAb Leu 8 (anti-CD62L). Subpopulations were co-
cultured at an effector : target ratio of 2 : 1 with
cells infected 18 h earlier with FIV at an MOI of 0.1 as described
by Bucci et al. . As a control, FIV-infected FCD4E cells were
cultured in the absence of CD8 effector cells. Different dilutions
of culture supernatants were assayed for FIV p26 production after
5 days of coculture by ELISA.
The nonparametric Kruskal-Wallis test
 was used to compare the expression of surface markers on
CD4 or CD8 cell populations.
P ! .05
The paired t test was used to assess significance of difference be-
tween means in the CD8 FIV suppressor assay.
was considered significant.
FIV infectionresults inthe rapid
We previously reported on
CD8a?bloCD62L?T cell subset.
the expansion of an atypical CD8 subset with a marked re-
duction of the b chain (CD8a?blo) and increased expression of
MHC class II molecules in the peripheral blood during acute
FIV infection . Rabin et al.  and Roederer et al. 
reporteda markeddecreasein naiveCD8Tcellsandanincrease
in atypical (memory/effector) CD8 cells lacking surface ex-
pression of CD62L in HIV-infected children and adults. To
determine whether the CD8 cells that expand during early
acute-stage FIV infectionreflectchangesinthememory/effector
or naive peripheral T cell pool, we did two-colorflowcytometry
on samples from 4 cats during the first 4 weeks after FIV in-
fection. CD62L expression distinguishes naive and memory T
cells from effector T cells [16, 17]; therefore, MAbs to CD62L
(Leu-8) and the CD8 b (MAb 117) and CD8 a (MAb 357)
chains were added in combination to evaluate the expression
of CD62L on CD8 cells during acute-stage infection.
Four 6-month-old SPF cats were infected intravenously with
TCID50of FIV NCSU1as described elsewhere . All6?10
cats were seropositive for FIV antigen as determinedbyELISA,
and PBMC were positive for FIV provirus as determined by
PCR for gag sequences by 4 weeks after infection (data not
shown). As shown in figure 1, there was a progressive and
parallel increase in the percentage of CD8b?CD62L?cells with
time after infection in the 4 FIV-infected cats compared with
2 age-matched controls. We evaluated 3 of the 4 cats at 8 and
52 weeks after infection, and in all cases, the percentage of
CD8b?CD62L?cells remained high (figure 1), suggesting that
this phenotype is present in high numbers throughout the
course of infection.ThepercentageofCD62L?cellsthatstained
by guest on December 26, 2015
1506 Gebhard et al.JID 1999;180 (November)
cells from healthy cat (A) and cat infected with feline immunodefeiciency virus for 24 weeks (B). Note loss of CD62L on CD8 cells is restricted
to CD8a?blocells. Nos. in upper and lower right quadrants represent percentage of positive cells.
Dot plot demonstrating, by two-color flow cytometric analysis, CD62L and CD8 b chain expression on peripheral blood mononuclear
with the CD8 a chain MAb (MAb 357) was similar to the
percentage of CD8b?CD62L?cells (data not shown), suggest-
ing that we were not missing significant numbers of CD8 cells
by using the b chain MAb.
To further substantiate that the increase in CD8b?CD62L?
cells continues throughout the asymptomatic and late phase of
FIV infection, we analyzed a large number of healthy cats and
cats infected with FIV for between 6 months and 9 years. As
shown in figure 2, there is a progressive and significant (P !
) increase in the proportion of CD8b?CD62L?cells .001
throughout the course of infection. In the case of cats infected
for ?7 years, the meanpercentageoftheCD8b?CD62L?subset
was 180% of the total CD8b?T cells.
Examination of the absolute CD8b?cell numbers in FIV-
infected cats revealed that the increase in CD8b?cells was due
to an increase in CD8b?CD62L?cells (from
infected cats to at 6 months after infection) and a 440?215
decrease in the number of CD8b?CD62L?naive cells (from
in uninfected cats to641?420
infection). The increase in the CD8 activation phenotype and
loss of the CD8 naive phenotype is particularly prominent dur-
ing the acute-stage infection, consistent with the change in the
relative proportions of these two phenotypes at this stage of
infection. Although the control cats ranged in age from 1 to 6
years, there was no significant difference in proportions of
CD8b?CD62L?among them (figure 2), suggesting that the in-
crease in this phenotype in FIV-positive cats was not an effect
As shown in figure 1, the increase in a CD8b?CD62L?subset
parallels the expansion of the CD8a?blosubset previously de-
in un- 119?132
at 6 months after398?304
scribed by Bucci et al. , suggesting that they may be the
same cell population. In support of this, the two-color flow
analysis scatter plot of CD62L expression on CD8 cells from
an FIV-infected cat revealed that the loss of CD62L is pre-
dominately restricted to the CD8 subset with reduced expres-
sion of the CD8 b chain (figure 3B). Figure 3 clearly shows
that there are greater numbers of CD8a?blocells in the PBMC
of the FIV-infected cat (13.3%, figure 3B) than in the control
cat (1.6%, figure 3A).
Analysis of the MIF of the CD8 b chain on CD62L?cells
also suggests that CD8b?CD62L?cells are of the CD8blophe-
notype. Figure 4 shows data for 12 cats at 24 weeks after FIV
infection expressed as the MIF for CD4 or CD8 staining on
either the CD62L?or CD62L?populations. Note that while
CD8 cells clearly fall into twosignificantlydifferentpopulations
on the basis of expression of the CD8 b chain and CD62L
( ), CD8bhiCD62L?, and CD8bloCD62L?, the expres-
P ! .001
sion of the CD4 molecule does not appear to change at this
stage of FIV infection. The MIF for CD4 is similar for the
CD62L?and CD62L?T cellsubset(
of CD62L expression on CD4 cells in FIV-infected cats dem-
onstrates a trend toward an increase in the percentage of
CD4?CD62L?cells with time after infection such that
CD4?CD62L?cells represent 180% of total CD4 cells in the
long-term (?7 years after infection) FIV-infected cats (P !
; figure 5)..001
Expression of adhesion and integrin molecules is increased on
the CD8a?bloCD62L?subset of FIV-infected cats.
loss of CD62L is restricted to an activation phenotype [16, 17],
the expansion of the CD8a?bloCD62L?subset in the peripheral
P = .92
by guest on December 26, 2015
JID 1999;180 (November) CD8?CD62L?Cells in FIV-Infected Cats 1507
expression on CD62L-positive and -negative cells in peripheral blood
mononuclear cells of feline immunodeficiencyvirus(FIV)–infectedcats.
Blood was obtained from 12 cats 24 weeks after FIV infection and
analyzed by two-color flow cytometry for expression of CD62L on
CD4 and CD8 (monoclonal antibody [MAb] 117) cells. MIFofstaining
by MAb to CD4 and to CD8 b chain on CD62L-positive and -negative
cells was determined for each cat and plotted as box and whisker plot.
Each box and whisker plot represents 5th and 95th percentiles (whisk-
ers), 25th and 75th percentiles (box), and median (middle line). Ex-
pression of CD8b is reduced on CD62L?cells (
of CD4 is not ( ).
P = .92
Mean index of fluorescence (MIF) of CD8b and CD4
) but expression
P ! .001
immunodeficiency virus (FIV) infection. Peripheralbloodmononuclear
cells from cats infected with FIV for various lengths of time were
analyzed by two-color flow cytometry for expression of CD62L on
CD4 cells. ?, individual cat values: seronegative,
; 2–3 year,; 3–6 year,
n = 6
n = 5
n = 6
after infection,; 1–3 years after infection,
n = 31
after infection, . There was significant change in expression of
n = 9
CD62L on CD4 cells among 4 groups of cats (
Increase in percentage of CD4?CD62L?cells duringfeline
; 6 months
; and 7 years
n = 17
n = 63
n = 23
P ! .001
blood suggests that FIV infection results in a generalized CD8
T cell activation. In addition to the loss of CD62L, activated
CD8 cells are characterized by increased surface expression of
the adhesion molecules and integrins, including CD44, CD49d,
CD11a, CD11b, and CD18 and MHC class II antigens .
Although MHC class II antigens are constitutively expressed
on feline CD4 and CD8 T cells , Willett et al.  and Bucci
et al.  reported that MHC class II was up-regulated on
CD8a?blocells in FIV-infected cats.
To further characterize the CD8a?bloCD62L?cell subset, a
number of MAbs produced to human T cell activation markers
were tested on PBMC from FIV-infected cats. Of the MAbs
tested, only MAb against human CD44, CD49d, and CD18
reacted with the homologues on feline PBMC. Two-color flow
cytometry was used to examine the expression of integrins on
CD8a?bloCD62L?cells in the peripheral blood of 12 FIV-in-
fected cats 24 weeks after infection. Histogram analysis of the
CD44 expression on CD8a?bloand CD8a?bhisubsets revealed
that CD8a?bloT cells express increased levels of CD44. Al-
though the increase in CD44 was restricted to the CD8a?blo
phenotype, the difference in CD44 expression was not sufficient
in all cases to clearly distinguish a CD44hifrom a CD44losubset.
Therefore, the MIF of the entire CD44-positive populationwas
used to describe changes in expression. As shown in figure 6,
CD44 expression is greater on the CD8blothan the CD8bhi
subset (). Similar two-color flow cytometric analysis of
P ! .001
CD49d and CD18 coexpression on CD8 cellsfromFIV-infected
cats demonstrated that expression of both integrins is also in-
creased on CD8a?bloCD62L?cells (
FIV infection is characterized by the development of a domi-
nant population of CD8 cells in the circulation that express
an activation phenotype (CD8a?bloCD62L?CD44hiCD49dhi
Expansion of the CD8a?bloCD62L?subsetismorepronounced
in the peripheral blood than in the lymph nodes.
typic changes that occur in the blood during infection are
thought to reflect an antigen-driven T cell immune activation
in lymph nodes; therefore, to evaluate CD62L expression on
CD4 and CD8 cells, we analyzed in parallel two-color flow
cytometry of peripheral blood and lymph nodes from 12 cats
that had been infected with FIV for 24 weeks. As shown in
figure 7A, the median percentage of CD8a?bloCD62L?cells of
total CD8 cells in the PBMC of FIV-negative cats was 17%,
whereas the median for CD8a?bloCD62L?cells in FIV-positive
cats was 66% of the total CD8 cells. In contrast, the median
percentage of CD8a?bloCD62L?cells in the lymph nodes was
high for both groups of cats: 66% and 50% for FIV-positive
and -negative cats, respectively.
The trend was similar for CD4?CD62L?cells in the two
tissue compartments (figure 7B), although the increase in the
percentage of CD4?CD62L?cells in the peripheral blood of
FIV-positive cats (median, 31%) compared with negative cats
(median, 18%) was not as great as the increase in CD
8a?bloCD62L?cells. Thus, the marked expansion of CD62L-
negative cells in FIV-infected cats appears to be principally a
; figure 6). Thus,
P ! .05
by guest on December 26, 2015
1508 Gebhard et al.JID 1999;180 (November)
from peripheral blood mononuclear cells (PBMC) of feline immuno-
deficiency virus (FIV)–infected cats. PBMC from cats infected with
FIV for 24 weeks were analyzed by two-color flow cytometry for ex-
pression of CD44 (), CD49d (
n = 12
cells. Mean index of fluorescence of each integrin on CD8bhiand
CD8blocells was determined for each cat and plotted as box and
whisker plot. There was significant difference in expression of all 3
integrins ( ) on CD8a?blocells compared with that on CD8a?bhi
P ! .05
Integrin and adhesion molecule expression on CD8 cells
), and CD18 () on CD8
n = 4
n = 5
peripheral blood mononuclear cells (PBMC) vs. lymph nodes of cats.
PBMC and lymph node cells were collected from 12 healthy cats and
12 feline immunodeficiency virus–positive cats 24 weeks after infection
and analyzed for CD62L expression on CD8 (A) and CD4 (B) cells
by two-color flow cytometry. Data are expressed as 5 of total CD8 or
CD4 cells. NEG, negative; POS, positive.
Percentage of CD8 and CD4 cells that were CD62L?in
characteristic of CD8?cells and is a characteristic of the blood
but not the lymph nodes, particularly during the early asymp-
tomatic stage of infection. These data also show that there was
a higher percentage of CD62L-negative cells inthelymphnodes
than in the blood of control SPF cats.
Antiviral activity of CD8a?bloCD62L?cells.
 reported that CD8a?blobut not CD8a?bhicells from FIV-
infected cats showed strong antiviral activity in an in vitro
FIV acute-infection assay. In this experiment, we confirmed
that the anti-FIV activity could be attributed to the
CD8a?bloCD62L?CD44hiCD49hiCD18hiphenotype in FIV-
infected cats. Purified CD8a?bhiCD62L?and CD8a?bhi
CD62L?subsets were sorted from 6 FIV-infected cats by dual
antibody flow cytometry sorting with MAb 117 (anti-CD8 b
chain) and Leu 8 (anti-CD62L). The purified CD8 subsets
were analyzed for anti-FIV activity in the acute-infection as-
say . CD8blobut not CD8bhicells from all 6 cats showed
anti-FIV suppressor activity in vitro, as indicated by reduced
p26 levels. Figure 8 illustrates the strong antiviral activity of
highly purified (198%) CD8a?blo62L?CD44hiCD49dhiCD18hi
cells from 2 representative FIV-infected cats. In contrast,
CD8a?bloCD62L?naive cells failed to inhibit FIV replication
in the same assay. Thus, the CD8a?bloCD62L?cells fromFIV-
infected cats possess strong antiviral activity indicative of an
activation phenotype. Cocultures of uninfected FCD4E cells
with either purified CD8bloor CD8bhicells from FIV-infected
cats did not yield any measurable p26 after 5 days of culture
(data not shown).
The CD8a?bloCD62L?anti-FIV cells do not express CD57.
Bucci et al.
Barker et al.  reported that anti-HIV–suppressing activity
could be enhanced by depleting CD8 cells expressing CD57,
suggesting that the suppressor cells were CD57 negative. In
contrast, CD8 lymphokine-activated killer cells in both FIV-
and HIV-infected subjects express CD57 [33, 34]. To explore
the relationship between CD57-expressing CD8 cells and the
anti-FIV–suppressor activity of CD8 cells, we performed two-
color flow cytometry analysis of PBMC from FIV-infected and
control cats, using MAb to the CD8 b chain and CD57. Figure
9 clearly shows that CD57?is expressed on a fraction of
CD8a?bhicells but not on CD8a?blocells, supporting the ar-
gument that the CD8a?bloCD62L?anti-FIV cells are FIV sup-
pressor cells and not lymphokine-activated killer cells. Table 1
summarizes the surfacephenotypedistinguishinganti-FIVCD8
by guest on December 26, 2015
JID 1999;180 (November)CD8?CD62L?Cells in FIV-Infected Cats1509
CD8a?bloCD62L?cells from FIV-infected cats. Peripheral blood mon-
onuclear cells from 2 cats infected with FIV were sorted into
CD8a?bhiCD62L?and CD8a?bloCD62L?populations (?98% pure)
and assayed for anti-FIV suppressor activity in FIV acute-infection
assay. Purified CD8 subsets were added to acutely FIV-infectedFCD4E
cells at effector : target ratio of 1 : 2, and supernatant (diluted 1 : 5, 1
: 25, and 1 : 125) was assayed for p26 by ELISA after 5 days of culture.
Target controls consisted of FIV-infected FCD4E cells cultured alone.
A, cat infected with FIV for 24 weeks with 35% CD8a?bloCD62L?
cells, B, cat infectedwithFIV for 30 weekswith46%CD8a?bloCD62L?
cells. Bars represent mean and SD of triplicate samples. p26 levels in
CD8a?bloCD62L?cell cultures were significantly different (
than levels in CD8a?bhiCD62L?cell and control cultures by paired t
Anti–feline immunodeficiency virus (FIV) activity of
P ! .005
cells from naive CD8 cells in the circulation of FIV-infected
Presence of CD8a?b?CD62L?cells in asymptomatic FIV-in-
In addition to expansion of a CD8a?blosubset
of cells during the asymptomatic stage of FIV infection, a sub-
population of CD8 cells completely lacking the CD8 b chain
(CD8a?b?) emerges in the blood after FIV infection, particu-
larly late in infection (data not shown). While this observation
has been reported by Shimojima et al. , there was no state-
ment regarding the expression of activation markers on the
CD8a?b?cell. To address this question, we performed three-
color flow cytometry on PBMC from 4 cats with long-termFIV
infections, using MAbs 117 (CD8b), 357 (CD8a), and Leu-8
(CD62L), as described in Materials and Methods. Figure 10A
shows a representative two-color dot plot of the expression of
CD8a and CD8b on PBMC from a long-term (?7 years) FIV-
infected cat, indicating the presence of three subsets of CD8 T
cells: CD8a?b?(box a), CD8a?blo(box b), and CD8a?bhi(box
c). Similar two-color flow analysis did not reveal a significant
number of CD8a?b?cells in the blood of normal SPF cats.
Also, a large number of cats with shorter term infections(figure
2) was analyzed for CD8a and CD8b expression and did not
demonstrate a significant fraction of CD8a?b?cells, suggesting
that the complete loss of b chain expression on CD8 cells is a
characteristic of long-term infected cats.
To determine whether CD8a?b?cells express CD62L, we
gated and analyzed the CD8a?b?population for theexpression
of CD62L. Figure 10B is a histogram showing the expression
of CD62L on CD8a?bhi, CD8a?blo, and CD8a?b?. These data
clearly demonstrate that CD62L expression is restricted to the
naive CD8a?bhisubset in the periphery, suggesting that both
CD8a?bloand CD8a?b?are activation phenotypes.
Similar to the numerous reports on HIV infection [2, 3], the
total CD8 cell count in the peripheral blood of FIV-infected
cats usually increases early after infection and remains high
throughout the long asymptomatic stage of infection, falling
only late in the disease . In both lentivirus infections, the
CD8 lymphocytosis is characterized by expansion of pheno-
typically atypical subpopulations of CD8 cells and a corre-
sponding decrease in the normal naive subpopulation. In the
case of FIV infection, Bucci et al. [7, 23] reported that the CD8
lymphocytosis was manifested by a rapid and sustained ex-
pansion of a T cell subset with a marked reduction in the
expression of the CD8 b chain (CD8a?blo). In the case of HIV
infection, Schmitz et al.  used a MAb specific for theCD8ab
heterodimer to demonstrate a reduction in the MIF of the CD8
receptor complex on PBMC of HIV-seropositive patients. A
similar reduction in the CD8 MIF was not seen when we used
MAb to the CD8 a chain, suggesting a selective reduction in
b chain expression.
Schmitz et al.  reported that CD8 cells with reduced ex-
pression of the CD8 b chain were of the activation phenotype,
as manifested by expression of HLA-DR and CD38 surface
markers. Although expression of MHC class II molecules is
not as a reliable marker for feline T cell activation as it is for
human T cells because of low-level constitutive expression on
naive T cells , a number of studies have described increased
expression of MHC class II on CD8a?bloT cells inFIV-infected
cats [21, 23, 36].
In this study, we use the differential CD8 b chain expression
by guest on December 26, 2015
1510Gebhard et al. JID 1999;180 (November)
(FIV) effector cells from naive CD8 cells in FIV-infected cats by pattern
of cell surface marker expression.
Differentiation of CD8 anti–feline immunodeficiency virus
Surface marker density
Anti-FIV CD8 cellsNaive CD8 cells
CD8 a chain
CD8 b chain
MHC class II
aCD57 is expressed on a subpopulation of naive CD8 cells.
MHC, major histocompatibility complex.
cells from control cat (A) and cat infected with feline immunodeficiency virus for 24 weeks (B). Note that CD57 was not expressed on CD8a?blo
effector phenotype but was expressed on subpopulation of CD8a?bhicells.
Dot plot demonstrating two-color flow cytometry analysis of CD57 and CD8 b chain expression on peripheral blood mononuclear
to further define the surface markers, tissue distribution, and
anti-FIV activity of the different CD8 subsets in acute and
asymptomatic FIV-infected cats. Data reported herein indicate
that FIV infection is unusual in that it is characterized by a
profound and sustained loss of the naive CD8a?bhiCD
62L?CD44loCD49dloCD18lophenotype and a concurrent ex-
pansion of an atypical CD8a?bloCD62L?CD44hiCD49dhi
CD18hiphenotype in the circulation. Loss of CD62L and in-
crease in the adhesion molecule (CD44) and integrins (CD49d
and CD18) on the surface of CD8 cells is indicative of T cell
activation [16, 17], suggesting that naive CD8 cells are largely
replaced by activated CD8 cells during the course of FIV in-
fection. In addition, this study demonstrates that within this
CD8-activated cell population resides a phenotype with strong
anti-FIV suppressor activity.
Rapid expansion of a CD8a?bloCD62L?activation pheno-
type and loss of the naive cells is not unique to FIV infection,
as shown by Rabin et al.  and Roederer et al. , who
reported a similar decrease in naive CD8 T cell subsets in HIV-
infected children and adults, respectively. Similar to studies de-
scribed herein, Rabin et al.  and Roederer et al.  used
multiparameter flow cytometry to demonstrate a marked de-
crease in the naive (CD11aloCD45RAhiCD62L?)subsetinHIV-
infected children and a concurrent increase in an atypical
CD11ahiCD45RAloCD62L?subset. Others have described the
expansion of atypical CD8 cells in HIV-infected patients char-
acterized by up-regulation of integrins and loss of CD62L .
Recent studies by Zimmerman et al.  and Hamann et al.
 used multiparameter flow cytometry to document pheno-
typic distinctions between naive, effector, and memory CD8 T
cells in mice and humans. Zimmerman et al.  demonstrated
in the lymphocytic choriomeningitis virus mouse model that
effector CTLs, in contrast to naive cells, have strongly down-
regulated the lymph node homing receptor CD62L, while ex-
pression of the CD44 adhesion molecule and the integrins
CD11a/CD18, CD11b, and CD49d is enhanced. Memory cells,
on the other hand, were mostly positive for CD62L, had no
CD11b on their surface, and had an intermediate level of
Using a similar flow cytometric approach, Hamann et al. 
described three distinct populations within the circulating hu-
man CD8 subset. In addition to the naive CD45RA?
CD62L?integrinlopopulation, a CD8?subset with direct cy-
totoxicity for virus-infected targets (CD8?CD45RA?CD62L?
CD11ahiCD11bhiCD18hiCD49dhi) and a subset with antigen-
CD11bhiCD18hiCD49dhi) were identified. It is clear from these
studies that the absence of the lymph node homing receptor
CD62L on the cell surface is, perhaps, the most useful marker
by guest on December 26, 2015
JID 1999;180 (November)CD8?CD62L?Cells in FIV-Infected Cats1511
antibodies (MAbs) to CD8a (MAb 117), CD8b (MAb 375), and CD62L was performed on peripheral blood mononuclear cells from cat infected
with feline immunodeficiency virus for ?7 years. A, Two-color display of expression of CD8a and CD8b that defines CD8 subsets: box a, CD8a?b?;
box b, CD8a?blo; box c, CD8a?bhi. B, Histogram demonstrating expression of CD62L on 3 CD8 subsets defined in (A). CD62L is not expressed
on CD8a?b?(a, dark thick line) or CD8a?blo(b, light dotted line) subsets but is expressed on CD8a?bhi(c, light solid line) subset.
CD62L is expressed on CD8a?bhisubset but not on CD8a?bloor CD8a?b?subsets. Three-color flow cytometry using monoclonal
for distinguishing CD8 effector cells from both naive and mem-
ory CD8 cells. These observations thus support the argument
that the CD8?CD62L?cells that dominate in the circulation
of HIV- and FIV-infected subjects are activated cells.
In the case of FIV, this argument is supported by the ob-
servation that the CD8a?bloCD62L?CD44hiCD49dhiCD18hiac-
tivation phenotype, but not the CD8 naive phenotype, pos-
sessed strong anti-FIV activity in the acute-infection assay
without prior stimulation. Barker et al.  reported that CD8
cells with strong HIV suppressor activity could bedistinguished
from CD8 nonsuppressor cells by the absence of the CD57
molecule on their surface. We also observed that the
CD8a?bloCD62L?cells were CD57?, whereas a subpopulation
of CD8a?bhiCD62L?were CD57?, suggesting that the former
cells are FIV suppressor cells.
The sustained elevation of CD8 effector cells in the circu-
lation of HIV- and FIV-infected subjects is highly unusual: ki-
netic studies in mice show that CD8 effector T cells are found
only relatively early (7–10 days) after virus infectionanddecline
thereafter [16, 18]. In contrast, CD8 memory cells become
prominent at a later time (∼8 weeks) and recirculateformonths.
The reason for the discrepancy in the kinetics of CD8 effector
T cells in these two lentivirus infections and murine nonlenti-
virus infection models is not clear, but it is possible that per-
sistent virus or stimulatory cytokines may play a role in sus-
taining the high levels of CD8 effector cells. Low-level viremia
and viral antigen, which could possibly maintain CD8 cells in
a state of activation, can be detected in plasma and PBMC
during the asymptomatic stage of FIV and HIV infections [38,
39]. In this regard, Schmitz et al.  reportedthatHIV-infected
patients with reduced CD8ab heterodimer expression showed
an increase CD8ab MIF after highly active antiretroviral ther-
apy designed to reduce virus burden. This response is consistent
with the suggestion that the CD8a?blophenotype is chronically
activated by persistent virus replication. Additional studies will
be required to resolve this issue.
Although analysis was limited and restricted to one time
point, the marked alteration in distribution of naive and ef-
fector CD8 cells seen in the blood of FIV-infected cats does
not appear to be a feature of the lymph nodes, particularly at
the early, asymptomatic stage of infection (24 week after in-
from the lymph nodes as rapidly as they are formed. This is
supported by the observation that lymph nodes do show small
increases in both CD8?CD62L?and CD4?CD62L?cells in
FIV-infected cats, compared with controls. Thus, the CD
8a?bloCD62L?cells could originate in the lymph nodes and
enter the circulation via the efferent lymphatics. Because they
lack L-selectin, the cells could not reenter the lymph nodes and
would thus be maintained in high numbers in the circulation.
Of interest, even normal SPF cats have higher percentages
of CD62L?CD4 and CD8 cells in their peripheral lymph nodes
than in their circulation. This could reflect effects of the specific
microenvironment of the lymph nodes. Picker et al.  re-
ported a differential site-specific regulation of CD62L on T cells
in secondary lymphoid tissue during the virgin–to–memory cell
transition. For example, they reported that in the peripheral
lymph node, T cells maintain a consistent high level of CD62L
expressionduring the virgin–to–memory cell transition,
whereas in the appendix, memory T cells down-regulateCD62L
on their surface. These authors demonstrated that a number
of cytokines, including interleukin-2 and -6 and transforming
by guest on December 26, 2015
1512 Gebhard et al. JID 1999;180 (November)
growth factor–b, can regulate T cell CD62L expression in vitro,
and they speculated that the site-specific factors regulating
CD62L expression in secondary lymphoid tissue were these or
other cytokines. It is also possible that this CD62L-negative
effector phenotype originates in the thymus rather than in the
lymph node, and because it lacks CD62L, it could not emigrate
from the circulation into the lymph node, which could also
result in sustained high levels in recirculation. Studies are in
progress to address these questions.
As reported for HIV-infected patients [14, 15], FIV also ap-
pears to induce a CD62L-negative CD4 cell subset, although
at a slower rate. Similar to the CD8a?bloCD62?phenotype,
CD4?CD62L?cells are a major fraction of the circulating CD4
cells only in long-term (?7 years) FIV-infected cats. Whatever
the cause, the loss of naive CD8 cytotoxic T lymphocyte pre-
cursors, and to a lesser extent naive CD4 T cells, may have
important consequences for the development of immune re-
sponses, particularly in late-stage infections when both CD4
and CD8 naive phenotypes are severely depleted. A progressive
depletion of naive T cells would result in a progressive inability
to mount immune responses to novel antigen and to the rapidly
In addition to the potential long-term detrimental effects of
loss of naive CD8 cells on immune responsiveness to new an-
tigens, high numbers of CD62L?integrinhi-activated CD8 and
CD4 T cells in the circulation could have important and more
immediate pathologic consequences with respect to lymphocyte
trafficking and inflammation or immunodeficiency (or both).
If, because of loss of CD62L, which is a necessary homing
receptor for high endothelialvenulesofperipherallymphnodes,
these activated T cells could not enter lymph nodes but would
concentrate in the blood (as we have observed), then the high
levels of surface integrins may promote a rapid and selective
emigration of large numbers of activated T cells from the cir-
culation into sites of inflammation. For example, we have re-
cently observed a marked and rapid influx of CD8bloCD62L?
cells into the lungs of FIV-infected cats challenged with the
opportunistic pathogen Toxoplasma gondii that was not seen
in cats infected only withT. gondii(unpublisheddata).Aninflux
of CD8bloCD62L?cells into the lung correlated with active
replication of T. gondii tachyzoites in alveolar macrophagesand
development of severe interstitial pneumonia. Thus, it is pos-
sible that these cells may exert an immunosuppressive effect on
the pulmonary immune system.
In light of this profound numerical alteration in naive and
effector CD8 subsets in HIV- and FIV-infected subjects, it will
be important to further define the tissue trafficking and func-
tional properties of the CD8bloCD62L?effector cells to deter-
mine what role they may play in immunity or immunodefi-
ciency. Also, understanding how the ratio of the two CD8
phenotypes changes with disease progression and in response
to antiretroviral therapy may prove very useful in monitoring
the health of patients and their response to therapy.
We thank Deborah Anderson for excellent technical assistance and
Jay Levine for help with the statistical analysis.
1. English RV, Nelson P, Johnson CM, Nasisse M, Tompkins WA, Tompkins
MB. Development of clinical disease in cats experimentally infected with
feline immunodeficiency virus. J Infect Dis 1994;170:543–52.
2. Prince HE, Jensen ER. HIV-related alterations in CD8 cell subsets defined
by in vitro survival characteristics. Cell Immunol 1991;134:276–86.
3. Prince HE, Jensen ER. Three-color cytofluorometric analysis of CD8 cell
subsets in HIV-1 infection. J Acquir Immune Defic Syndr1991;4:1227–32.
4. Borrow P, Lewicki H, Hahn BH, Shaw GM, Oldstone MB. Virus-specific
CD8?cytotoxic T-lymphocyte activity associated with control of viremia
in primary human immunodeficiency virus type 1 infection. J Virol
5. Levy J, Mackewicz C, Barker E. Controlling HIV pathogenesis: the role of
the noncytotoxic anti-HIV response of CD8?cells. Immunol Today
6. Mackewicz CE, Yang LC, Lifson JD, Levy JA. Non-cytolytic CD8 T-cell
anti-HIV responses in primary HIV-1 infection. Lancet 1994;344:1671–3.
7. Bucci J, English R, Jordan H, Childers T, Tompkins M, Tompkins W. Mu-
cosally transmitted feline immunodeficiency virus induces a CD8?anti-
viral response that correlates with reduction of cell-associated virus. J
Infect Dis 1998;177:18–25.
8. Jeng CR, English RV, Childers T, Tompkins MB, Tompkins WAF. Evidence
for CD8?antiviral activity in cats infected with feline immunodeficiency
virus. J Virol 1996;70:2474–80.
9. Landay A, Mackewicz C, Levy J. An activated CD8?T cell phenotype
correlates with anti-HIV activity and asymptomatic clinical status. Clin
Immunol Immunopathol 1993;69:106–16.
10. Ferbas J, Kaplan A, Hausner M, et al. Virus burden in long-term survivors
of human immunodeficiency virus (HIV) infection is a determinant of
anti-HIV CD8?lymphocyte activity. J Infect Dis 1995;172:329–39.
11. Giorgi J, Liu Z, Hultin L, Cumberland W, Hennessey K, Detels R. Elevated
levels of CD38?CD8?T cells in HIV infection add to the prognostic value
of low CD4?T cell levels: results of 6 years of follow up. J Aquir Immun
Defic Syn 1993;6:904–12.
12. Giorgi J, Ho H, Hirji K, et al. CD8?lymphocyte activation at human im-
munodeficiency virus type 1 seroconversion: development of HLA-DR?
CD38?CD8?cells is associated with subsequent stable CD4?cell levels.
The Multicenter AIDS Cohort Study Group. J Infect Dis 1994;170:
cell suppression of HIV replication. J Immunol 1997;159:5123–31.
14. Rabin RL, Roederer M, Maldonado Y, Petru A, Herzenberg LA,Herzenberg
LA. Altered representation of naive and memory CD8 T cell subsets in
HIV-infected children. J Clin Invest 1995;95:2054–60.
15. Roederer M, Dubs JG, Anderson MT, Raju PA, Herzenberg LA, Herzenberg
LA. CD8 naive T cell counts decrease progressivelyinHIV-infectedadults.
J Clin Invest 1995;95:2061–6.
16. Zimmerman C, Brduscha-Riem K, Blaser C, Zinkernagel R, Pricher H. Vi-
sualization, characterization, and turnover of CD8?memory cellsinvirus-
infected hosts. J Exp Med 1996;183:1367–75.
17. Hamann D, Baars P, Rep M, et al. Phenotypic and functional separation of
memory and effector human CD8?T cells. J Exp Med 1997;186:1407–18.
by guest on December 26, 2015
JID 1999;180 (November)CD8?CD62L?Cells in FIV-Infected Cats1513
18. Tripp R, Hou S, Doherty P. Temporal loss of the activated L-selectin–low
phenotype for virus-specific CD8?memory T cells. J Immunol 1995;154:
19. Picker L, Treer J, Ferguson-Darnell B, Collins P, Buck D, Terstappen L.
Control of lymphocyte recirculation in man. I. Differential regulation of
the peripheral lymph node homing receptor L-selection on T cells during
the virgin to memory cell transition. J Immunol 1993;150:1105–21.
20. Lehmann R, von Beust B, Niederer E, et al. Immunization-induced decrease
of the CD4?:CD8?ratio in cats experimentally infected with feline im-
munodeficiency virus. Vet Immunol Immunopathol 1992;35:199–214.
21. Willett BJ, Hosie MJ, Callanan JJ, Neil JC, Jarrett O. Infection with feline
immunodeficiency virus is followed by the rapid expansion of a CD8?
lymphocyte subset. Immunology 1993;78:1–6.
22. Shimojima M, Miyazawa T, Kohmoto M, et al. Expansion of CD8a?b?cells
in cats infected with feline immunodeficiency virus. J Gen Virol 1998;79:
23. Bucci J, Gebhard D, Childers T, English R, Tompkins M, Tompkins W. The
CD8?phenotype mediating antiviral activity in FIV-infected cats is char-
acterized by reduced surface expression of the CD8 b chain. J Infect Dis
24. English RV, Johnson CM, Gebhard DH, Tompkins MB. In vivo lymphocyte
tropism of feline immunodeficiency virus. J Virol 1993;67:5175–86.
25. Tompkins MB, Nelson PD, English RV, Novotney C. Early events in the
immunopathogenesis of feline retrovirus. J Am Vet Med Assoc 1991;199:
26. 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.
27. Levy J, Ritchey J, Rottman J, et al. Elevated interleukin-10–to–interleukin-
12 ratio in feline immunodeficiency virus–infected cats predicts loss of
type 1 immunity to Toxoplasma gondii. J Infect Dis 1998;178:503–11.
28. 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 tissuedistribution
in the cat. Vet Immunol Immunopathol 1990;26:305–17.
29. Rideout B, Moore P, Pedersen N. Distribution of MHC class II antigens in
feline tissues and peripheral blood. Tissue Antigens 1990;36:221–7.
30. Shimojima M, Pecoraro M, Maeda K, Tohya Y, Miyazawa T, Mikami T.
Characterization of anti-feline CD8 monoclonal antibodies. Vet Immunol
31. Chao CC, Jensen R, Dailey MO. Mechanisms of L-selectin regulation by
activated T cells. J Immunol 1997;159:1686–94.
32. Noether G. Introduction to statistics: the nonparametric way. New York:
33. Gryllis C, Wainberg MA, Bentwich Z, Gornitsky M, Brenner BG. Increased
LAK activity against HIV-infected cell lines in HIV-1?individuals. Clin
Exp Immunol 1992;89:356–61.
34. Zhao Y, Gebhard D, English R, Sellon R, Tompkins M, Tompkins W. En-
hanced expression of novel CD57?CD8?LAK cells from cats infected
with feline immunodeficiency virus. J Leukoc Biol 1995;58:423–31.
35. Schmitz J, Forman M, Lifton M, et al.ExpressionoftheCD8ab-heterodimer
on CD8(?) T lymphocytes in peripheral blood lymphocytes of human
immunodeficiency virus?and human immunodeficiency virus?individu-
als. Blood 1998;92:198–206.
36. Lehmann R, Franchini M, Aubert A, Wolfensberger C, Cronier J, Lutz H.
Vaccination of cats experimentally infected with feline immunodeficiency
virus, using a recombinant feline leukemia virus vaccine. J Am Vet Med
37. Rosenberg Y, Anderson A, Pabst R. HIV-induced decline in blood CD4/CD8
ratios: viral killing or altered lymphocyte trafficking? Immunol Today
38. Bendinelli M, Pistello M, Lombardi S, et al. Feline immunodeficiency virus:
an interesting model for AIDS studies and an important cat pathogen.
Clin Microbiol Rev 1995;8:87–112.
39. Fauci A, Schnittman S, Poli G, Koenig S, Pantaleo G. Immunopathogenic
mechanisms of human immunodeficiency virus (HIV) infection. Ann In-
tern Med 1991;114:678–93.
by guest on December 26, 2015