Download The establishment of cytomegalovirus latency in organs is not linked

Journal of General Virology (1994), 75, 2329--2336. Printedin Great Britain
2329
The establishment of cytomegalovirus latency in organs is not linked to
local virus production during primary infection
Monika Balthesen, Liane Dreher, Pero Lu~in and Matthias J. Reddehase*t
Department of Virology, Institute for Microbiology, University of Ulm, Albert-Einstein-Alice 11, 89081 Ulm, Germany
Recovery from primary cytomegalovirus (CMV) infection is associated with resolution of the productive
infection without clearance of the virus genome from
affected organs. The presence of latent CMV genome in
multiple organs provides the molecular basis for recurrence of CMV within multiple organs, and explains
the diversity in the organ manifestations ofrecrudescent
CMV disease during states of immunodeficiency. As a
part of a unifying concept of multifocal CMV latency
and recurrence, previous work has demonstrated the
importance of primary virus replication for the overall
load of latent CMV in organs and the risk of recurrence.
In the present report, the establishment of CMV latency
was studied in a murine model in which the course of
primary infection in the immunocompromised host after
syngeneic bone marrow transplantation was modulated
Introduction
Lymphohaematopoietic reconstitution of a cytolytic T
cell response after bone marrow transplantation (BMT)
is pivotal for the control of a post-transplantation
primary or recurrent infection of patients with the
human herpesvirus type 5 (HHV-5), human cytomegalovirus (CMV) (Quinnan et al., 1982; Reusser et al., 1991).
Experimental studies in the murine model of CMV
disease (for reviews, see Koszinowski et al., 1990, 1993)
have identified CD8 + T cells as the principal antiviral
effector subset of T lymphocytes that limits virus
multiplication and tissue lesions in organs (Reddehase et
al., 1985, 1987a, b, 1988). Recent clinical trials were
aimed at supplementing an insufficient CD8 + T cell
reconstitution in patients after BMT and HHV-5
infection by adoptive transfer of CMV-specific CD8 + T
cell lines (Riddell et al., 1992).
The establishment of latency after the resolution of
primary infection is a characteristic of herpesvirus
infections (Roizman & Sears, 1987), such as CMV
t Present address: Institute for Virology, Johannes-GutenbergUniversity,Hochhausam Augustusplatz,55101 Mainz,Germany.
0001-2295 © 1994SGM
by a CD8 + T cell immunotherapy. The antiviral CD8 +
effector cells limited virus replication in all organs and
protected the recipients from lethal CMV disease, but
after resolution of the productive infection virus DNA
remained. Interestingly, the copy number of latent virus
DNA in tissue did not quantitatively reflect the preceding virus production in the respective organ. Specifically, in contrast to the case in the lungs and the
salivary glands, virus replication in the spleen was
suppressed by CD8 + T cells to below the limit of
detection; yet, virus DNA was also detected in the
spleen during latency and accordingly, virus recurrence
in the spleen could be induced. These findings demonstrate that the control of virus replication in a
particular organ does not prevent the establishment of
latency in that organ.
infection (for reviews, see Jordan, 1983; Mocarski et al.,
1990). Latent murine CMV DNA has been detected in a
variety of organs that are also the sites of productive
infection and viral pathogenesis, including the spleen,
lungs, salivary glands, heart, liver, brain, kidney and the
adrenal glands (Klotman et al., 1990; Balthesen et al.,
1993; Collins et al., 1993; Reddehase et al., 1994). We
have shown previously that the copy number of latent
virus DNA detected in tissues by PCR correlates with the
risk of virus recurrence in the respective organ (Balthesen
et al., 1993), and that virus replication and dissemination
during primary infection determines the load of latent
CMV (Reddehase et al., 1994).
It is obvious that CD8 + T cells primed during the
physiological immune response do not prevent CMV
latency. However, it is an essentially different question,
so far unanswered for CMV, whether or not a timely
provision of terminally differentiated antiviral CD8 +
effector cells by cell transfer can clear the viral DNA.
Whereas the resolution of productive infection by CD8 +
T cells has been demonstrated for many different viruses
(for a review, see Kaufmann & Reddehase, 1994),
clearance of the viral genome has rarely been analysed.
That CD8 + T cells are principally capable of clearing
virus genome has been documented for the persistent
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 05 May 2017 00:38:43
2330
M. Balthesen and others
Draining popliteal
lymph node
Replicates
bp
BALB/c
- 363
100
No disease
CMV s.c.
/~4,_
J
,~,,
41 ~L~
:-Ot
8 days
Immune response
:~
' CD8
- 363
~'~
CTL-P c
1
......
'~
8 days
> CTL e
in culture plus IL-2
A "
- 363
6 Gy
Fig. 1. Sensitivity of the detection of a viral sequence by PCR. A 363 bp
sequence of exon 4 of the ieI gene of murine CMV was amplified from
1, 10 and 100 molecules ofplasmid piE111 in 10 independent replicates
each. The total amount of DNA per sample was adjusted to 3 lag with
certified-negative carrier DNA from mouse spleen. Shown are the
autoradiographs obtained after hybridization of the amplification
products with a radioactively labelled internal oligonucleotide probe.
~ccCD4+C
Adoptive transfer i.v.
°1
.i °'
L
CMV s.c~- Bone marrow
transplantation i.v.
~
CD8
infection of mice with lymphocytic choriomeningitis
virus (Oldstone et al., 1986). By combining the approach
of adoptive CD8 + T cell immunotherapy with the
detection of viral DNA in organs after the resolution of
primary infection we demonstrate here that a dose of
CD8 + T cells that suffices for protection against lethal
disease and that prevents overt virus replication in the
spleen does not preclude latent infection of the spleen.
Methods
Bone marrow transplantation, infection and cytoimmunotherapy.
Recipients as well as donors for syngeneic BMT were specifiedpathogen-free, 8-week-old female mice of the inbred strains BALB/c or
BALB/c-H-2 dm~ (dm2). The dm2 strain is a mutant of BALB/e, in
which a region of the major histocompatibility complex (MHC),
encompassing the gene encoding the MHC class I glycoprotein Ld, is
deleted (Klein et al., 1983). For haematoablative conditioning, BMT
recipients were total-body ),-irradiated with a single dose (6 Gy) from
a caesium-137 source (OB58; Buchler) delivering a dose rate of
0.708 Gy/min. Donor femoral bone marrow cells (BMC) were obtained
as described previously (Mutter et al., 1988). BMC (1 × 106) were
infused into the tail vein of recipients about 6 h after the irradiation.
Infection with 1 × l0 b p.f.u, of murine CMV, strain Smith (ATCC VR194) was performed subcutaneously at the left hind footpad about 2 h
after BMT. In the case of cytoimmunotherapy, 1 × 10~ antiviral T cells
of the CD8 + subset were infused intravenously simultaneously with the
BMC. The CD8 + T cells were prepared as described previously
(Reddehase et al., 1987b). In brief, BALB/c mice were infected at the
left hind footpad and, 8 days later, sensitized T lymphocytes obtained
from the draining popliteal lymph node were propagated in microcultures for a further 8 days in the presence of recombinant interleukin
(IL)-2, but with no viral antigen added. CD4 + T cells were eliminated
from the polyclonal, short term CD4 + and CD8 + T cell line by
treatment with monoclonal antibody (MAb) GK-1.5 and rabbit
complement.
Read out: survival
control of infection
clearance of CMV genome
Fig. 2. Experimental regimen for CD8 ÷ T cell supplementation therapy.
Antiviral T lymphocytes were sensitized by subcutaneous (s.c.) infection
of BALB/c mice with murine CMV. Cells derived from the draining
lymph node were propagated in culture in the presence of IL-2. During
this period, cycling cytolytic T lymphocyte precursors (CTL-P °)
develop into mature effector cells (CTLe). CD4 ÷ T cells were eliminated
from the short term T cell line by treatment with anti-CD4 antibody
and complement (c~CD4+ C). The purified CD8 + T cells were mixed
with BMC and infused intravenously (i.v.) into ),-irradiated (6 Gy)
recipients. The steps in the purification of primed CD8 + T cells were
monitored by two-colour cytofluorometric analysis by using
fluorochrome-conjugated antibodies specific for the marker molecules
CD4 and CD8. The scales show fluorescence intensities. Contour lines
represent cell frequencies in a 70% log mode. Numbers in the plot
corners give the percentage of cells located in the indicated quadrants.
Cytofluorometric analys&. Two-colour cytofluorometric analysis was
performed with a FACScan (Becton Dickinson) using LYSIS (Becton
Dickinson) software for computer analysis. Two-dimensional contour
plots represent 25000 cells. Contour lines show cell frequencies in a
70% log mode. Single-fluorescence histograms represent 10000 cells.
O) CD-phenotyping o f T cells. Lymph node ceils as well as cells of the
T cell line were stained with the directly fluorochrome-conjugated
MAbs rat anti-mouse CD8-fluorescein isothiocyanate (FITC) (clone
53-6.7; Becton Dickinson) and rat anti-mouse CD4-phycoerythrin
(PE) (clone SK3; Becton Dickinson).
(ii) Tracking of donor CD8 + T cells in M H C chimeras. Cells from the
spleen (enriched for T cells using nylon wool), were stained with three
antibodies: mouse MAb anti-L d (clone CL 9011-A; Cedarlane),
FITC-conjugated goat anti-mouse IgG2a (Medac) and MAb rat antimouse CD8 PE (clone 53-6.7; Boehringer Mannheim). CD8 + T cells
were selected analytically by the setting of an electronic window on cells
with positive PE fluorescence.
Determination o f virus titres and induction of i n vivo recurrence. Virus
titres in organ homogenates during acute infection were determined by
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 05 May 2017 00:38:43
Local C M V production and latency
Survival rate
2331
Virus replication in organs
(a)
I
I
[
I
I
l
I
100%
-7
£
50%
-
6
-
5
4
37
0%
2
m
L
(b)
I
I
I
2
3
(weeks)
4
I
I
"-"7"-~--
I
T
I
l
I
l
I
F"-
4
5
~c ~
0
100%
50%
0%
0
1
0
1
2
3
(months)
Time after infection
Fig. 3. Time course of CMV infection and disease. Syngeneic BMT was performed with BALB/c as donor and recipient strain. (a) No
therapy; (b) therapy with 1 × l0 s BALB/c CD8 ÷ T cells. Left panel: survival rates (n = 20). Right panel: virus replication in organs
determined by plaque assay from organ homogenates. Symbols represent virus titres in three individuals per time point. Vertical bars
indicate the range, horizontal bars mark the median values. DL, Detection limit of the standard plaque assay. Cases in which the
respective organ was negative for all three tested individuals are depicted only on the first occasion in the kinetics. The asterisk indicates
that the organ homogenate was retested with the highest sensitivity plaque assay capable of detecting 1 p.f.u, per organ. Symbols: (O),
salivary glands; (©), lungs; (ll), spleen.
a standard 4 day plaque assay on permissive mouse embryo fibroblasts
essentially as described previously (Reddehase et al., 1985). The
sensitivity of detection was increased 20-fold by using centrifugal
infection. Testing of an aliquot of 1% of the homogenate thus defines
the detection limit of this assay as 100 p.f.u, per organ. To ensure the
absence of infectious virus during latent infection, the sensitivity was
increased to 1 p.f.u, per organ by plating all of the homogenate.
Further, the plaque assay was then performed without semi-solid
overlay to allow any secondary plaque formation. A negative score is
assigned to organs if no plaque was detected after 4 days as well as 10
days of the assay. Recurrent infection in vivo was induced by )~irradiation (6 Gy) of latently infected mice. Recurrent virus was
detected 14 days later with the highest sensitivity plaque assay.
Inactivation of murine CMV was done by irradiation with u.v. light of
wavelength 254 nm. Its efficacy was monitored by the highest sensitivity
plaque assay, and was verified by the absence of infection after
inoculation of y-irradiated mice.
hybridization with a 32P-end-labelled internal oligonucleotide probe.
Autoradiographic signal intensities were found to increase linearly only
between 10 and 100 copies of the test sequence (Balthesen et al., 1993).
The sensitivity of detection was determined by the amplification of 100,
10 and 1 molecules of piE111 in 10 independent replicates (Fig. 1). The
signal variance seen with 10 template molecules reflects the statistical
variance given by the Poisson distribution. From the fractions of
negative replicates, the most probable number of detectable templates
is calculated as being four with a 95 % confidence interval of 2 to 8
(maximum likelihood method).
Detection o f vital and cellular D N A sequences. Organs and blood
leukocytes were processed for DNA isolation, and a 363 bp sequence
from exon 4 of the murine CMV immediate early (IE) gene iel as well
as a sequence from the mouse fl-actin gene were amplifed by PCR in
35 cycles as described previously (Balthesen et aL, 1993). Plasmid
piE111, which encompasses the iel gene of murine CMV (Smith
strain), served as the positive control. Amplification products were
analysed by agarose gel electrophoresis, Southern transfer and
Supplementation of B M C with mature, antiviral CD8 + T
effector cells was performed in syngeneic experimental
B M T to modulate the course of a concurrent primary
murine C M V infection. Recipients of B M T and donors
of B M C as well as of CD8 + T cells were B A L B / c mice
( M H C haplotype H-2d). The number of B M C was
adjusted so as to provide full protection against the
Results
Protective and antiviral efficacy of CD8 + T cell
supplementation therapy
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 05 May 2017 00:38:43
2332
M. BaIthesen and others
radiation-mediated haematological aplasia, but to be
insufficient for controlling C M V infection (Mutter et al.,
1988). This strategy was chosen as a model for clinical
trials of a CD8 + T cell supplementation therapy in
patients with a failure in endogenous T cell reconstitution
after B M T (Riddell et at., 1992). The experimental
regimen for B M T and cytoimmunotherapy of C M V
infection is outlined in Fig. 2. The steps in the generation
and purification of polyclonal, antiviral CD8 + T effector
cells were monitored by two-colour cytofluorometric
analysis of the expression of CD4 and CD8 cell surface
molecules.
The CD8 + T cell supplementation therapy was clearly
beneficial in that it significantly reduced mortality
between the first and third week after B M T and infection
(Fig. 3) as well as wasting syndrome morbidity (daily
inspection, not shown). This protection was accompanied by a 100-fold (lungs) to 1000-fold (salivary
glands) reduction of virus titres at the peak time of
mortality, that is at around 2 weeks (Fig. 3). The control
of infection in the spleen is particularly efficient.
Although the spleen is a prominent target organ for
C M V after B M T (Fig. 3 a), virus replication at this site
was suppressed by antiviral CD8 + T cells to below the
limit of detection (Fig. 3b). As this limit was 100 p.f.u, in
the standard virus plaque assay, the spleens were then
retested by plating all of the homogenate in the highest
sensitivity plaque assay. No p.f.u, were detected in spleen
homogenates at 1, 2 and 4 weeks after CD8 + T cell
transfer. As the mice are negative for antibodies to C M V
early after BMT, the assay was not influenced by virus
neutralization. After 4 to 5 months, acute infection had
resolved in all tested organs and in both experimental
groups.
Adoptively transferred CD8 + T cells are recruited to a
target site o f viral pathogenesis and persist long term in
the recipients
Since the effect of CD8 + T cell transfer was most
pronounced in the first weeks after infection and was
most prominent in the spleen, we tested whether the
transferred cells were indeed recruited to that site to
perform their function. A chimeric model was used to
distinguish transferred donor CD8 + T cells from recipient
CD8 + T cells regenerated during lymphohaematopoietic
reconstitution. D o n o r cells for the immunotherapy were
again the BALB/c-derived antiviral CD8 + T cells used in
the preceding experiment, whereas syngeneic B M T was
performed in the L d gene deletion mutant B A L B / c - H 2 am~, in which the M H C class I glycoprotein L ~ is not
expressed. The fate of the donor cells within the recipients
was tracked by cytofluorometric detection of L d on the
surface of CD8 + T cells (Fig. 4). It should be noted that
Time p.i. (weeks)
4
2
90%
8
35%
21%
Expression of L~t(log10fluorescence intensity)
Fig. 4. Fate of transferred donor-type CD8÷ T cells in the recipient.
Syngeneic BMT was performed with the La gene deletion mutant
BALB/c-H-2din2 as donor and recipient strain. La-positive CD8+ T
cells for the immunotherapy were derived from BALB/c mice. The
origin of CD8÷ T cells in the chimeras was determined by two-colour
cytofluorometric analysis by using antibodies specificfor CD8 and La.
An electronic window was set to select for CD8+ T cells. The
distribution of recipient (Ld-negative)and donor (La-positive)CD8+ T
cells is shown by single-fluorescencehistograms. Horizontal bars mark
the donor cell fluorescence. Numbers in the upper right corners givethe
percentage of donor-derived cells among the CD8÷ T cells.
antiviral CD8 + T cells of these two mouse strains are
cross-protective (not shown), which shows that protective immunity is not restricted to antigenic peptides
presented by L d (for review, see Koszinowski et al.,
1992). At 2 weeks after the transfer, most of the CD8 + T
cells recovered from the spleen were of donor origin,
whereas endogenous reconstitution of recipient CD8 + T
cells became effective between 2 and 4 weeks posttransfer. Yet, donor cells were replaced only slowly and
remained detectable even after 2 months (Fig. 4). That
transferred CD8 + T cells migrate to target sites and can
persist long term in the recipients is in accordance with
previous reports (Cheever et al., 1986; Reddehase et al.,
1988; Jamieson & Ahmed, 1989; Riddell et al., 1992). As
the peak of mortality is at around 2 weeks, the main
benefit of therapy by donor CD8 + T cells is to bridge the
interim between haematoablative treatment and haematopoietic reconstitution, whereas termination of persistent
infection is likely to be mediated mainly by the recipient's
own cells.
Load of viral genome in blood leukocytes during acute
infection
Acute infection is expected to be accompanied by blood
cell-associated viraemia and P C R - D N A positivity of
blood leukocytes. Viraemia was tested by a limiting
dilution infectious centre assay that was performed by
plating graded numbers of blood leukocytes in 24
replicates in flat-bottom well microcultures with m o n o layers of mouse embryo fibroblasts. The frequency of
leukocytes positive for infectious virus was determined
from the fractions of plaque-negative cultures on the
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 05 May 2017 00:38:43
Local C M V production and latency
Controls
1¢ 1 21
(a)
Blood leukocytes
] 0
1
2
3
4
t
oO
No
therapy
- 363 bp
t
CD8
t-
therapy
(b)
1¢
I
211
2
3
4
5
6
363 bp
I
- 363 bp
(c)
Blood cell D N A
I 1
2
3 4 5
I
-
fl-actin
Fig. 5. Detection of viral D N A in blood leukocytes during acute
infection. (a) A 363 bp sequence from exon 4 of gene iel of routine
CMV was amplified by PCR from blood leukocyte D N A isolated at 1
month p.i. and BMT with or without CD8 + T cell supplementation
therapy. Leukocytes were titrated (10-fold dilutions) from 104 cells to
1 cell (lanes 4 to 0, respectively) before the D N A was isolated.
Throughout, the PCR was run with 3 pg of DNA. For cell numbers
~< 105, the amount of D N A was adjusted with certified-negative carrier
D N A from the spleen of an uninfected mouse, for cell numbers > 105
a respective aliquot of the D N A was tested. Arrows mark the
autoradiographic signal that gives the frequency of positive leukocytes.
Lanes ~b, contain all reagents except DNA. Control lanes 1 and 2
contain 10 and 100 copies, respectively, of plasmid piE111 adjusted to
3 gg with carrier DNA. (b) Blood leukocytes were titrated for D N A
isolation and PCR at 1 month after inoculation of mice with a dose of
inactivated virus that corresponds to l0 b p.f.u, o f infectious virus.
Lanes 2 to 6 represent 100 to 106 cells in increasing 10-fold steps. (e) A
sequence of the cellular fl-actin gene was detected by PCR in fractions
of the D N A isolated from l0 s leukocytes from the experiment shown
in (b). Lanes 1 to 5 contain 60 pg to 600 ng of D N A in increasing
10-fold steps with no carrier DNA.
basis of the Poisson distribution by using the maximum
likelihood method of calculation. In the group with no
CD8 + T cell supplementation therapy, the frequency was
213 per 106 (i.e. about 1 per 5000) leukocytes at 4 weeks
(95 % confidence interval: 149 to 306 per 106 leukocytes;
P = 0.62). In contrast, after therapy, the number of
positive leukocytes was below the detection limit.
The presence of viral DNA in blood leukocytes was
tested by PCR amplification of a 363 bp sequence from
2333
exon 4 of the iel gene of murine CMV (Balthesen et al.,
1993). We have shown in a previous report that murine
CMV DNA is associated with all subpopulations of
blood leukocytes, with the highest relative frequency
within Gr-I+CDllb ++ granulocytes (Balthesen et al.,
1994). In Fig. 5(a), the frequency of blood leukocytes
carrying viral DNA was determined at 4 weeks after
BMT and infection. With 1 per 10 leukocytes, this
frequency was much higher than the frequency of cells
containing infectious virus and, surprisingly, there was
no notable difference between the groups. An explanation could be that leukocytes that passively carry
viral DNA are exempt from immune control and are
saturated in both groups of mice. It should be considered
that the high virus production in salivary glands (Fig. 3)
probably does not contribute to viral DNA load in the
blood, because virus produced in acinar glandular
epithelial cells of the salivary glands is secreted via the
salivary duct (Hensen & Strano, 1972; Jonji6 et al.,
1989). That DNA-PCR positivity of blood leukocytes
requires replicative CMV is documented in Fig. 5(b).
Inoculum DNA from the same nominal dose of u.v.inactivated CMV was not detected in up to 106 blood
leukocytes at 4 weeks after transfer. Amplification of a
cellular sequence, namely a sequence of the mouse flactin gene (Balthesen et al., 1993), from the same DNA
preparation assured that the PCR was working correctly
(Fig. 5c).
Cytomegalovirus latency and recurrence in organs
The cellular site and the molecular nature of cytomegalovirus latency are still unknown. We therefore use the
term 'latency' in its operational definition, namely the
absence of detectable infectious virus in the presence of
functional viral genomes capable of reactivation
(Roizman & Sears, 1987). This definition is generally
applied also for CMV latency (for a review see Jordan,
1983). At 1 year, that is about 8 months after the
resolution of acute infection (Fig. 3), the absence of
infectious virus was verified in three-quarter fractions of
salivary gland, lung and spleen tissue of individual
survivors of the CD8 + T cell therapy group by the highest
sensitivity virus plaque assay. However, since putative
virus replication at any other untested site necessarily
escaped detection, we introduced absence of viral DNA
from blood leukocytes as an additional criterion for a
more stringent definition of CMV latency in organs.
As shown in Fig. 5, acute infection is reflected by the
presence of viral DNA in a high proportion of blood
leukocytes, and this fact is used in clinical tests as a
diagnostic marker for acute primary or recurrent CMV
infection in patients (Chou, 1993). In a previous
manuscript it was shown that murine CMV DNA is
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 05 May 2017 00:38:43
2334
M . Balthesen and others
(a)
Controls
I~
1
°
211
I
a
II
b II a
III
b
ID
-- 363 bp
(b)
[ ~b 1
21[
a
b
c
d
e
f
[
e~
II
III
O~
II
III
group analysed at 1 year (Fig. 6a; I to III), CMV DNA
was virtually cleared from the blood. The two samples
tested for each individual represented the DNA from 106
blood leukocytes. The yield of blood leukocytes and
sensitivity of the PCR give an estimate of < 20 copies of
viral DNA in total. Thus, there was no evidence for a
low-level persistent infection at any site, and also a
significant contamination of organ material by CMV
DNA-positive blood leukocytes was practically excluded.
In contrast, organs of the same three mice harboured
viral DNA with copy numbers of/> 100 (salivary glands
and lungs) or 10 to 100 (spleen) per 3 gg sample of organ
DNA, with almost no variance between six samples
tested, and with only little variance between the
individuals (Fig. 6b). It should be noted that essentially
the same result was obtained for individual mice with no
CD8 ÷ T cell therapy (not shown). However, a quantitative comparison is unwarranted, because the few
survivors after no therapy (Fig. 3 a) represent a highly
selected group with the most efficient endogenous
reconstitution of CD8+ T cell control.
As a 3 lag sample of organ DNA represents the DNA
of about 5 x 105 diploid cells, 100 copies per tested
sample represent a load of approximately 200 viral
genomes per 106 tissue cells. This estimate is close to the
organ load observed in another murine CMV latency
model, namely in latency established after the infection
of newborn mice (Balthesen et al., 1993; Reddehase et
al., 1994). In that model, the viral DNA load in organs
was predictive of the risk of recurrence. In fair
accordance with this previous finding, organ incidences
of recurrence induced by y-irradiation (6 Gy) in 20 mice
at 1 year after BMT, infection and CD8 ÷ T cell therapy
were found to be 12/20 for the lungs, 11/20 for the
salivary glands and 6/20 for the spleen.
Discussion
~
~
~l~ia~
~
~l~i
II
III
Fig. 6. Detection of viral DNA in organs during latency. Presence of
the 363 bp viral test sequence in blood leukocyte and organ DNA was
tested by PCR at 1 year after CD8 ÷ T cell supplementation therapy.
Three mice were tested individually (I, II and III). Lanes a to f contain
replicate 3 pg samples of DNA amplified independently. Control lanes
are as in Fig. 5.
maintained in blood leukocytes for an extended period
after resolution of the acute infection, but is eventually
cleared to below the limit of detection by PCR (Balthesen
et al., 1993). In three individual mice of the therapy
These data demonstrate the capacity of adoptively
transferred donor CD8 + T cells to protect against lethal
CMV disease after BMT in the period of immunological
incompetence before haematological reconstitution
becomes effective. However, although the peak virus
replication was markedly reduced, CD8 + T cells had
difficulties in terminating a low-level persistent infection
in salivary glands and lungs, and they failed to clear the
viral genome from solid tissues. This failure cannot be
explained by a principal limitation of CD8 + T cell
function, as experiments on another virus infection have
shown that adoptively transferred CD8 + T cells can clear
a viral genome from all organs (Oldstone et al., 1986).
Hence, the observed maintenance of the CMV genome
most likely reflects a property of this virus to escape
immune control. Recent work on the regulation of
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 05 May 2017 00:38:43
Local C M V production and latency
murine (Del Val et al., 1992) and human (Gilbert et al.,
1993) CMV antigenic peptide presentation in permissive
cells in vitro has identified molecular mechanisms of
evasion from recognition by CD8 + T cells. However, as
the described evasion is operative through viral genes
expressed during the productive cycle and results in
completion of the productive cycle, it can explain the
persistence of virus production but not latency.
Our data indicate that persistence and latency are
indeed unlinked phenomena, as latency was found to be
established in organs in which persistence occurred,
namely in salivary glands and lungs, but also in the
spleen where productive infection was prevented by the
transfer of CD8 + T cells. A discrepancy between local
virus production during acute infection and the amount
of latent viral DNA in particular organs was also
observed in our previous study, in which we compared
latency established after primary infection at adult age
with that established after neonatal infection (Reddehase
et at., 1994). Specifically, after infection of neonates, the
high and long-lasting virus production in the salivary
glands was not reflected by an accordingly high load of
latent CMV in salivary gland tissue. After infection of
adults, virus production was restricted to the salivary
glands and was undetectable in the lungs. Nonetheless,
during latency, the lungs harboured more viral genomes
than did the salivary glands. A similar dissociation
between local virus production during acute infection
and latency phenotype was mentioned in a review by
Mocarski et al. (1990) for mutants of CMV. Thus,
mutant RM 408 is debilitated for establishing latency in
the spleen, even though it replicates well there. Conversely, this mutant becomes latent in the salivary
glands, even though it is debilitated for salivary gland
growth during the acute phase of the infection.
However, that a certain overall level of virus replication and dissemination during primary infection is
required to reach all target cells for latent infection was
indicated by the marked difference in the organ load of
latent CMV and the risk of recurrence between mice
infected as neonates or as adults (Reddehase et al., 1994).
The cellular site(s) of CMV latency in the various
organs has not yet been identified. Analysis is made
difficult by the low copy number of latent genomes in
tissues, which predicts for the lungs a maximum
frequency of one latently infected cell per 5000 tissue cells
(Balthesen et al., 1993). For the spleen, the sinusoidal
lining cell of the stroma was proposed as a candidate,
because this cell type supported virus replication in acute
infection and because latent virus was detected in a
stromal fraction of the spleen (Mercer et al., 1988;
Pomeroy et at., 1991). In contrast, the observed dissociation between local virus production and latency
could indicate that the latently infected cells are not
2335
stationary tissue cells, but are immigrants derived from
an organ site at which virus replication had occurred in
the acute phase of the infection. An alternative, and
probably more straightforward, explanation is that cells
which account for the gross virus production during
acute CMV infection are distinct from the cells in which
latency is established, as is the case for herpes simplex
virus (Roizman & Sears, 1987). The fact that productive
CMV infection is cytocidal and causes histological
lesions in affected organs further argues against an
establishment of latency in productively infected cells.
We propose that in its basic principles fl-herpesvirus
latency is related to ~-herpesvirus latency.
This work was supported by grants to M.J.R. by the Deutsche
Forschungsgemeinschaft, Sonderforschungsbereich 322, and by the
Bundesministerium fiir Forschung und Technologie, project
01KE8817/0.
References
BALTHESEN,M., MESSERLE,M. & REDDEHASE,M. J. (1993). Lungs are
a major organ site of cytomegalovirus latency and recurrence.
Journal of Virology 67, 5360-5366.
BALTHESEN, M., SUSA, M., LUC1N, P. & REDDEHASE,M.J. (1994).
Cytomegalovirus DNA detected in blood leukocytes after resolution
of productive infection does not originate from latently infected
hematopoietic stem cells in the bone marrow. Croatian Medical
Journal 35, 19-25.
CHEEVER, M.A., BRITZMAN THOMPSON, D., KLARNET, J.P. &
GREENBERG,P. D. (1986). Antigen-driven long term-cultured T cells
proliferate in vivo, distribute widely, mediate specific tumor therapy,
and persist long-term as functional memory cells. Journal of
Experimental Medicine 163, 1100-1112.
CHOU, S. (1993). Molecular diagnostic techniques for CMV infection.
Excerpta Medica International Congress Series 1032, 183 194.
COLLINS,T., POMEROY,C. & JORDAN,M. C. (1993). Detection of latent
cytomegalovirus DNA in diverse organs of mice. Journal oflnfectious
Diseases 168, 725 729.
DEL VAL, M., HENGEL,H., HACKER,H., HARTLAUB,U., RUPPERT, T.,
LUON, P. & KoszINOWSIO,U. H. (1992). Cytomegalovirus prevents
antigen presentation by blocking the transport of peptide-loaded
major histocompatibility complex class I molecules into the medialGolgi compartment. Journal of Experimental Medicine 176, 729-738.
GILBERT, M. J., RIDDELL,S. R., Lt, C.-R. & GREENBERG,P. D. (1993).
Selective interference with class I major histocompatibility complex
presentation of the major immediate-early protein following infection with human cytomegalovirus. Journal of Virology 67,
3461 3469.
HENSON,D. • STRANO,A. J. (1972). Mouse cytomegalovirus: necrosis
of infected and morphologically normal submaxillary gland acinar
cells during termination of chronic infection. American Journal of
Pathology 68, 183 202.
JAMIESON, B.D. & ArtMED, R. (1989). T cell memory. Long-term
persistence of virus-specific cytotoxic T cells. Journal of Experimental
Medicine 169, 1993 2005.
JONJIC, S., MUTTER, W., WETLAND, F., REDDEHASE, M.J. &
KOSZINOWSKI, U. H. (1989). Site-restricted persistent cytomegalovirus infection after selective long-term depletion of CD4-positive T
lymphocytes. Journal of Experimental Medicine 169, 1199-1212.
JORDAN, M. C. (1983). Latent infection and the elusive cytomegalovirus. Reviews of Infectious Diseases 5, 205-215.
KAUFMANN, S. H. E. t~£ REDDEHASE,M. J. (1994). T cell subsets and
defense against bacteria and viruses. In Handbook of T and B
lymphocytes, pp. 237-267. Edited by E. C. Snow. San Diego:
Academic Press.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 05 May 2017 00:38:43
2336
M . Balthesen and others
KLEIN, J., FIGUEROA,F. & DAVID,C. S. (1983). H-2 haplotypes, genes
and antigens: second listing. Immunogenetics 17, 553-596.
KLOTMAN, M.E., HENRY, S.C., GREENE, R.C., BRAZY, P.C.,
KLOTMAN, P.E. 8¢ HAMILTON, J.H. (1990). Detection of mouse
cytomegalovirus nucleic acid in infected mice. Journal of Infectious
Diseases' 161,220-225.
KOSZINOWSKI, U.H., DEE VAL, M. t~ REDDEHASE, M.J. (1990).
Cellular and molecular basis of the protective immune response to
cytomegalovirus infection. Current Topics in Microbiology and
Immunology 154, 189-220.
KOSZINOWSKI, U.H., REDDEHASE, M.J. 8L DEE VAL, M. (1992).
Principles of cytomegalovirns antigen presentation in vitro and in
vivo. Seminars in Immunology 4, 71 79.
KOSZINOWSKI,U. H., REDDEHASE,M. J. & JONJ1C,S. (1993). The role of
T-lymphocyte subsets in the control of cytomegalovirus infection. In
Viruses and the Cellular Immune Response, pp. 429-445. Edited by
D B. Thomas. New York: Marcel Dekker.
MERCER,J. A., WILEY,C. A. & SPECTOR,D. H. (1988). Pathogenesis of
murine cytomegalovirus infection: identification of infected cells in
the spleen during acute and latent infections. Journal of Virology 62,
98%997.
MOCARSKI,E. S., ABENES,G. B., MANNING,W. C., SAMBUCETTI,L. C.
& CHERRINGTON, J.M. (1990). Molecular genetic analysis of
cytomegalovirus gene regulation in growth, persistence, and latency.
Current Topics in Microbiology and Immunology 154, 47-74.
MUTTER, W., REDDEHASE,M.J., BUSCH, F.W., BOHRING~H.-J. &
KOSZINOWSKI,U. H. (1988). Failure in generating hemopoietic stem
cells is the primary cause of death from cytomegalovirus disease in
the immunocompromised host. Journal of Experimental Medicine
167, 1645 1658.
OLDSTONE, M. B. A., BLOUNT,P. & SOUTHERN,P. J. (1986). Cytoimmunotherapy for persistent virus infection reveals a unique clearance
pattern from the central nervous system. Nature, London 321,
239 243.
POMEROY,C., HILLEREN,P. J. & JORDAN,M. C. (1991). Latent murine
cytomegalovirus DNA in splenic stromal cells of mice. Journal of
Virology 65, 3330-3334.
QUINNAN, G. V., KIRMANI, N., ROOK, A. H., MANISCHEWITZ, J. F.,
JACKSON, L., MORESCHI,G., SANTOS,G. W., SARAL, R. t~¢ BURNS,
W. H. (1982). Cytotoxic T cells in cytomegalovirus infection. HLArestricted T-lymphocyte and non T-lymphocyte cytotoxic responses
correlate with recovery from cytomegalovirus infection in bonemarrow-transplant recipients. New England Journal of Medicine 307,
7-13.
REDDEHASE,M. J., WEILAND,F., MUNCH, K., JONJIC,S., L/.)SKE,A. &
KOSZINOWSKI, U.H. (1985). Interstitial murine cytomegalovirus
pneumonia after irradiation: characterization of cells that limit viral
replication during established infection of the lungs. Journal of
Virology 55, 264-273.
REDDEHASE, M.J., MUTTER, W. & KOSZINOWSKI,U.H. (1987a). In
vivo application of recombinant interleukin 2 in the immunotherapy
of established cytomegalovirus infection. Journal of Experimental
Medicine 165, 650-656.
REDDEHASE, M.J., MUTTER, W., MONCH, K., BLTHRING, H.-J. &
KOSZINOWSKI,U. H. (1987b). CD8-positive T lymphocytes specific
for murine cytomegalovirus immediate-early antigens mediate
protective immunity. Journal of Virology 61, 3102-3108.
REDDEHASE, M.J., JONJI~, S., WE/LAND, F., MUTTER, W.
KOSZlNOWSKI, U.H. (1988). Adoptive immunotherapy of murine
cytomegalovirus adrenalitis in the immunocompromised host: CD4
helper-independent antiviral function of CD8-positive memory T
lymphocytes derived from latently infected donors. Journal of
Virology 62, 1061 1065.
REDDEHASE,M. J., BALTHESEN,M., RAPP, M., JONJIC, S., PAVI(~, I. 8¢
KOSZlNOWSKI, U.H. (1994). The conditions of primary infection
define the load of latent viral genome in organs and the risk of
recurrent cytomegalovirus disease. Journal of Experimental Medicine
179, 185 193.
REUSSER,P., RIDDELL, S. R., MEYERS,J. & GREENBERG,P. D. (1991).
Cytotoxic T lymphocyte response to cytomegalovirus after human
allogeneic bone marrow transplantation: pattern of recovery and
correlation with cytomegalovirus infection and disease. Blood 78,
1373-1380.
RIDDELL, S. R., WATANABE,K., GOODRICH,J., LI, C.-R., AGHA, M. &
GREENBERG,P. D. (1992). Restoration of viral immunity in immunodeficient humans by adoptive transfer of T cell clones. Science 257,
238 241.
ROIZMAN,B. & SEARS,A. E. (1987). An inquiry into the mechanisms of
herpes simplex virus latency. Annual Review of Microbiology 41,
543 571.
(Received 7 December 1993; Accepted 24 March 1994)
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 05 May 2017 00:38:43