INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Blood monocytes and tissue macrophages play a major role in the pathogenesis of cytomegalovirus (CMV) infections by serving as target cells in infected organs, as disseminators of the virus throughout the host, or as sites of CMV latency (reviewed in reference 32). The prominence of this cell lineage as a host for human CMV (HCMV) infection and replication is revealed by the frequent detection of HCMV DNA, RNA, or antigens in progenitor cells of the monocyte/macrophage lineage, in peripheral blood monocytes, or in differentiated macrophages infected in vitro or in vivo with HCMV (8, 18, 22, 25, 29, 31-33, 49, 51, 69). Infected blood monocytes, likely derived from infection of bone marrow progenitor cells, may disseminate HCMV to other permissive cells of the host (32, 65). The relatively nonpermissive monocytes and bone marrow progenitor cells harbor latent HCMV DNA (11, 22, 24), which reactivates upon cellular differentiation (25, 29, 50, 53, 54, 58).

Mouse studies using murine CMV (MCMV) corroborate findings with humans by identifying the macrophage as a major target cell. In mouse bone marrow, MCMV predominates within stromal cells (55), although hematopoietic progenitor cells may also be infected (20, 41). Circulating blood monocytes disseminate MCMV during acute infection, and their subsequent cellular differentiation into mature macrophages favors productive MCMV replication (2, 4, 16, 17, 34, 56, 66). During acute infection of the spleen, virus localizes to areas corresponding to marginal-zone macrophages (56). Since tissue macrophages are in close contact with circulating blood, they are likely one of the first cell types infected by blood-borne virus. Furthermore, macrophages serve as reservoirs of latent MCMV infection (20, 41).

In addition to their role as targets of MCMV infection, macrophages are important in the early inflammatory and innate immune responses to infection. Activated macrophages infiltrate the site of MCMV infection (17), where they may initiate a cascade of antiviral cytokines. Initially, the cytokines alpha/beta interferon (IFN-/), tumor necrosis factor alpha (TNF-), interleukin (IL)-1 alpha, IL-6, and IL-12 are produced in response to early MCMV infection (14, 15, 17, 36, 37, 46, 66, 68). The former two cytokines have direct antiviral activity in that they inhibit MCMV replication (10, 17, 28, 40, 66). In addition, IFN-/ enhances NK-cell-mediated blastogenesis and cytotoxicity, while TNF- augments IFN- produced by NK cells (37). IL-12 also activates NK cells to produce IFN-, which directly inhibits MCMV replication (10, 28, 36-38). These two cytokines, IL-12 and IFN-, may also cooperate in activating cytotoxic T lymphocytes, which are pivotal in the subsequent clearing of MCMV from most target organs (23). The role of macrophages in initiating this cascade of cytokine-mediated antiviral activities has not been determined.

Macrophages therefore appear to have dual, even paradoxical, roles: one as hosts for CMV replication and another as inducers of antiviral cytokines. Growth of HCMV and MCMV in macrophages is somewhat restricted, with virus production being delayed (9) or curtailed (17, 66) compared to infection in fibroblasts. Thus, tissue macrophages may serve as an anatomical and/or functional "filter" in intrinsically or extrinsically protecting other, more-permissive cell types. We hypothesized that the magnitude of virus replication in tissue macrophages dictates the amount of virus accessible to other, more-permissive cells. Concomitantly, infection of macrophages likely initiates the production of antiviral immune responses to guarantee efficient clearance of acute infection, thereby sparing from death the host that serves as a reservoir of latent infection.

With respect to the above hypotheses, we addressed the following questions. (i) Does the ability of MCMV to replicate in macrophages influence virulence and virus titers in target organs? (ii) Does the absence of tissue macrophages enhance or deter virus replication in the spleen and liver, two major target organs for MCMV replication? (iii) What contributions do macrophages make to the production of early, antiviral cytokines? Our approach utilized a deletion mutant of MCMV which was defective for replication in mature macrophages. The highly attenuated phenotype of this mutant virus in vivo indicated that the capacity of MCMV to replicate in differentiated macrophages was a major determinant of virus virulence. However, in the absence of splenic macrophages, growth of the mutant virus in this organ was restored and replication of wild-type (WT) virus was enhanced. The induction of at least two antiviral cytokines, TNF- and IFN-, required the presence of tissue macrophages and was influenced by the capacity of MCMV to replicate in this cell type. Therefore, while tissue macrophages support MCMV replication, they exert a net protective role in the pathogenesis of MCMV infection.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Mice. Six-week-old BALB/cAnN (Harlan Sprague Dawley, Indianapolis, Ind.) or C57BL/6J (Jackson Laboratories, Bar Harbor, Maine) mice were housed in sterile microisolator cages with sterile food, water, and bedding. CB17 severe combined immunodeficient (SCID) mice were bred and housed as previously described (17). Both BALB/cAnN and CB17 mice are of the H-2^d haplotype; C57BL/6 mice are of the H-2^b haplotype.

Cells. Murine NIH 3T3 fibroblasts (ATCC CRL-1658) (American Type Culture Collection, Rockville, Md.) were propagated in Dulbecco's modified essential medium (Mediatech, Herndon, Va.) supplemented with 10% heat-inactivated bovine calf serum (Hyclone Laboratories, Logan, Utah) and 1% L-glutamine (Gibco/BRL, Grand Island, N.Y.). IC-21 cells, a simian virus 40-transformed, C57BL/6 mouse peritoneal macrophage line (30) (ATCC TIB 186), were propagated in RPMI medium (Mediatech) supplemented with 10% heat-inactivated fetal calf serum (Gibco/BRL) and 1% L-glutamine.

Viruses. The parental WT virus used in these studies was the MCMV Smith strain (ATCC VR 194). The generation of mutant MCMV RV6, with open reading frames (ORFs) m137 and m138 deleted, was described previously (2). Mutant MCMV RV10, with ORFs M139, M140, and M141 deleted, with was generated by homologous recombination, using by our previously described methods (2). A pGEM-4Z recombination plasmid containing the first (left) two SstI fragments of MCMV HindIII-J, an e1--glucuronidase (-Glu) cassette (-glucuronidase reporter gene under the control of the MCMV e1 promoter), and the second SstI fragment within HindIII-I was constructed (pJIglu-10) (Fig. 1). The pJIglu-10 recombination plasmid was cotransfected with WT viral DNA to generate an MCMV mutant lacking 4.8 kb corresponding to ORFs M139, M140, and M141 (bases 193984 through 198832) (43). Revertant virus, RV10Rev, was produced by homologous recombination by using infectious, RV10 parental DNA and a recombination plasmid (pDJI) containing WT sequence from base 191764 (within m137) to base 201966 (within m143) (Fig. 1). Transfections were performed by using calcium phosphate precipitation as described previously (2). Virus RV10 was selected based on its blue plaque phenotype and was plaque purified five times. Virus RV10Rev was selected based on its white plaque phenotype as described elsewhere (21). The genotypes of the recombinant viruses were confirmed by Southern blot analyses. A genetic map of the mutant viruses is shown in Fig. 1. All virus stocks were prepared in NIH 3T3 cells and quantitated by standard plaque assay on NIH 3T3 cells. Mock virus preparations were supernatants from uninfected NIH 3T3 cells.

View larger version (25K): [in this window] [in a new window]   FIG. 1.   Construction of the MCMV mutants RV10 and RV10Rev. The open boxes indicate wild-type MCMV sequences within the HindIII J and I fragments. Solid lines denote deleted sequences. Shaded boxes indicate the locations of the e1--glucuronidase cassette. Plasmid pJIglu-10 contains MCMV and e1--glucuronidase sequences in the vector pGem4Z; pDJI contains the denoted MCMV sequences in the vector pcDNA3. The numbers refer to sizes (in kilobases) of the indicated fragments.

Virion purification by gradient centrifugation. In some experiments, gradient-purified virus was used to infect fibroblasts or IC-21 macrophages. MCMV was purified by gradient centrifugation on a 20 to 70% sorbitol gradient according to published protocols (5). Briefly, stock preparations of MCMV grown in NIH 3T3 cells were harvested and cellular debris was removed by centrifugation. Bacitracin (Calbiochem, La Jolla, Calif.) was added to the clarified supernatant to a final concentration of 100 µg/ml. The virus-containing supernatant was layered over a cushion of 20% sorbitol in Tris-buffered saline (TBS) (150 mM Tris, 30 mM NaCl [pH 7.5]) containing bacitracin (100 µg/ml) and then centrifuged at 26,000 × g for 1 h at 18°C. The pelleted virus was resuspended in 2 ml of TBS-bacitracin, layered onto a 20 to 70% sorbitol gradient, and centrifuged at 65,000 × g for 1 h. Low-density virions banding at the 40 to 50% interface or at the 50 to 60% interface were harvested, diluted 1:10 in TBS-bacitracin, and pelleted over a 20% sorbitol cushion for 1 h at 26,000 × g. The virus was resuspended in TBS-bacitracin and frozen, and titers were determined by standard plaque assay.

Southern blot analysis. Southern blots were used to confirm the genomic alterations of each viral mutant. Two million NIH 3T3 cells in 100-mm-diameter tissue culture plates were infected with virus at a multiplicity of 2 PFU/ml. When the cytopathic effect reached 100%, DNA was harvested for Southern blot analysis, which was performed as described previously (2).

Northern blot analysis. Northern blot analyses of viral RNA transcripts from RV10-, RV10Rev-, or WT virus-infected cells were performed essentially as described previously (2). For analysis of transcripts from the HindIII-J and -I regions, NIH 3T3 cells were infected at a multiplicity of 2 PFU/cell for 24 h. For analysis of immediate-early RNAs transcribed from HindIII-L or HindIII-I, 2 × 10⁶ NIH 3T3 fibroblasts or IC-21 macrophages were infected at a multiplicity of 1 PFU/cell with virions banding at the 40 to 50% or 50 to 60% sorbitol interface. After 2 h of adsorption and penetration, virus inocula were removed, medium was added, and infection was allowed to proceed for an additional z h. Total RNA was harvested by using the RNeasy kit (Qiagen, Chatsworth, Calif.) according to the manufacturer's instructions. Samples containing a total of 5 µg of RNA, extracted from equal numbers of cells, were loaded in each lane of a formaldehyde- agarose gel. The loading of equal amounts of RNA was confirmed by ethidium bromide staining of 18S and 28S rRNA. RNAs were first hybridized with the MCMV HindIII L fragment to detect ie1, ie2, and ie3 transcripts. The blot was then stripped by boiling for 10 min in 0.1% sodium dodecyl sulfate and hybridized with an SstI fragment from HindIII I (bases 198832 to 201744) to probe for expression of the immediate-early genes m142 and m143. Probes were labeled with digoxigenin by random priming for detection by chemiluminescence.

In vitro growth of MCMV mutants in fibroblasts and macrophages. Mutant and WT MCMV were compared for their abilities to replicate in fibroblast and macrophage cell lines and in primary, LPS-induced peritoneal macrophages. Approximately 10⁶ NIH 3T3 fibroblasts or IC-21 macrophages in T-25 flasks or 10⁵ primary macrophages in 6-well plates or T-25 flasks were infected at a multiplicity of 0.1 PFU/cell (cell lines) or 0.1 to 0.5 PFU/cell (primary macrophages). At the indicated times postinfection, both intracellular virus and extracellular virus were harvested collectively and their titers were determined on NIH 3T3 cells. Growth of each virus in the three cell types was quantitated at least twice.

Southern blot analysis of cell-associated virus. The mutant virus RV10 was compared to RV10Rev for its ability to bind to and/or penetrate IC-21 macrophages. A standard assay for herpesvirus penetration into cells (27) was modified to account for the fact that MCMV produces no cytopathic effects in IC-21 macrophages and consequently does not form plaques in this cell type. Therefore, in order to detect virus bound to or penetrated into macrophages, Southern blotting was used to detect viral genomic DNA. Initially, the validity of this method was tested by using infection of NIH 3T3 fibroblasts to compare results of the standard penetration assay, which assesses plaque formation, with those of Southern blotting for viral DNA. For the former assay, the published procedure was used (27). Briefly, 10⁶ NIH 3T3 cells in 100-mm-diameter dishes were infected with WT virus at a multiplicity of 0.0005 PFU/cell (to yield 200 PFU/dish) at room temperature for 1 h. The virus inocula were then removed, the monolayers were washed with phosphate-buffered saline (PBS), complete medium was added to each dish, and the temperature of the dishes was shifted to 37°C for the indicated periods to allow attached virus to penetrate. The monolayers of infected cells were treated with acid-glycine saline (0.8% NaCl, 0.038% KCl, 0.01% MgCl₂, 0.01% CaCl₂, 0.7% glycine [pH 3]) for 1 min at 0, 15, 30, 60, 90, or 120 min after shifting to 37°C in order to inactivate and/or remove low-affinity-bound virus that was attached but had not yet penetrated. Infected monolayers were then overlaid with semisolid medium to determine the number of plaques formed in each treatment group. The percentage of bound virus which penetrated was calculated as follows: (number of plaques formed on acid-glycine-washed monolayers)/(number of plaques formed on unwashed monolayers) × 100.

In vivo growth of MCMV mutants. Recombinant and WT viruses were compared for their abilities to replicate in mouse spleen and liver tissues. Mice received 3 × 10⁵ PFU of tissue culture-passaged virus intravenously (i.v.) in the tail vein, and organs were harvested from three individual mice as 20% (wt/vol) homogenates on each of days 1, 2, and 3 postinfection and titered as described previously (2). Blood was also collected, in heparinized tubes, and serial 10-fold dilutions of whole blood in tissue culture media were used for virus titration. These experiments were performed at least twice.

Macrophage depletion. Replication of recombinants and WT viruses was compared in intact and macrophage-depleted spleens and livers of BALB/cAnN mice. Anesthetized animals were injected i.v. with a total volume of 0.5 ml containing 300 µl of PBS and 200 µl of multilamellar liposomes encapsulating dichloromethylene-bisphosphonate (Cl₂MBP) or, as a control, with 0.5 ml of PBS alone. The Cl₂MBP was a kind gift from Boehringer GmbH (Mannheim, Germany). When liposomes encapsulating Cl₂MBP (L-Cl₂MBP) are administered i.v., they are readily phagocytized by liver macrophages (Kupffer cells) and splenic red pulp macrophages, marginal-zone macrophages, marginal metallophilic macrophages, and marginal-zone dendritic cells (26, 35, 63, 64). Lysosomal enzymes present specifically within these cell populations degrade the liposomes and release the Cl₂MBP, leading to apoptosis of the cells (62). Peripheral blood monocytes and other tissue macrophages (including peritoneal cells) are spared from L-Cl₂MBP-induced death when the liposomes are administered i.v. For this study, use of PBS as a control for L-Cl₂MBP treatment was preferred over use of empty liposomes, which may alter macrophage functions, including antiviral activities, in otherwise intact cells (63). The L-Cl₂MBP method of tissue macrophage depletion efficiently and specifically depletes tissues of macrophages and macrophage-like marginal dendritic cells without toxicity for other cell types (61, 63).

Cytokine analysis. Levels of cytokines produced by splenocytes in response to WT or mutant virus infections in mice with intact spleens or macrophage-depleted spleens were compared. Mice were injected i.v. with L-Cl₂MBP or PBS and 48 h later were infected i.v. with tissue culture-passaged virus as described above. Spleens were harvested from three individual mice in each treatment group at 36 h after infection, the time at which early cytokine production in response to MCMV infection peaks (46). Suspensions containing 5 × 10⁶ total splenocytes per ml were prepared in RPMI 1640 containing 1% low-endotoxin serum (Hyclone) according to published procedures (37). After 24 h of culture, cell-free supernatants from each individual spleen were pooled and frozen. Levels of TNF-, IL-12 p40, IL-12 p70, IFN-, and IL-6 in the splenocyte-conditioned media were quantitated by sandwich enzyme-linked immunosorbent assay (ELISA) as previously described (6). Student's t test was performed to evaluate statistical differences in cytokine values between mice treated with L-Cl₂MBP or PBS and between mice infected with WT or mutant MCMV.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction and in vitro growth characteristics of RV10 and RV10Rev. We previously described a mutant MCMV, RV7, in which ORFs m137 through M141, contained within the HindIII-J and -I regions of the viral genome, are deleted and which grows poorly in the differentiated macrophage cell line IC-21 (2). Another mutant virus, RV6, with m137 and m138 deleted, grows like WT virus in IC-21 macrophages (1), suggesting that an ORF(s) within the M139 to M141 region contributes to growth of MCMV in this cell type. Therefore, a mutant virus (RV10) with ORFs M139, M140, and M141 deleted was generated to assess the function of these ORFs in macrophage-specific growth and viral pathogenesis. A marker-rescued revertant of RV10, RV10Rev, was also generated by homologous recombination of WT sequences back into RV10. Southern blotting of viral DNA digested with HindIII and BamHI confirmed the genotypes of RV10 and RV10Rev (data not shown). Northern blot analyses of RNA isolated from RV10-infected fibroblasts during the late phase of replication revealed the expression of RNA transcripts of WT size from genes proximal to the deleted region, although RV10 m138 RNA was larger than the RNA of WT virus. Transcripts from the HindIII-J and -I region in RV10Rev-infected cells were identical to those in WT virus-infected cells (data not shown).

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Growth of RV10 and WT MCMV in fibroblasts and macrophages in vitro. Monolayers of NIH 3T3 cells (A) or IC-21 macrophages (B) were infected at a multiplicity of 0.1 PFU/cell. Cell-free and cell-associated viruses were collectively titerate at the indicated times postinfection by plaque assay on NIH 3T3 cells. Data shown are representative of three individual experiments.

Viral gene expression in RV10-infected macrophages. It was of interest to identify the stage of the MCMV replication cycle at which growth of RV10 in macrophages was curtailed. The growth rate of RV10 in IC-21 macrophages was similar to that previously reported for the mutant virus RV7, which lacks ORFs m137 through M141 (2). The defect in growth of RV7 was identified as a block in expression of the major immediate-early genes ie1, ie2, and ie3 (2). Therefore, immediate-early gene expression in RV10-infected IC-21 macrophages was assessed and compared to that in RV10Rev-infected cells. For these experiments, highly purified infectious virions obtained by centrifugation through a sorbitol density gradient were used for infection of fibroblasts and IC-21 macrophages. Experiments were performed by using each of the two bands (40 to 50% interface and 50 to 60% interface) which contained most of the infectious virus and which likely contained a majority of single-capsid, enveloped virus (3). Equal amounts of RNA from equal numbers of infected cells were analyzed by Northern blotting for expression of immediate-early genes. Expression from two immediate-early gene regions was assessedthe major immediate-early gene region in HindIII-L (ie1, ie2, and ie3) and immediate-early genes m142 and m143 in HindIII-I (12). These two immediate-early gene regions have different kinetics of expression and are therefore likely to be independently regulated (12).

View larger version (38K): [in this window] [in a new window]   FIG. 3.   Expression of RV10 and RV10Rev immediate-early genes in fibroblasts and macrophages. NIH 3T3 fibroblasts or IC-21 macrophages were mock infected (M) or were infected with RV10Rev (Rev) or RV10 at a multiplicity of 1 PFU/cell. The virus inocula were purified virions banding at the 40 to 50% sorbitol interface of a 20 to 70% sorbitol gradient. Five micrograms of total RNA harvested at 4 h after infection was run on a formaldehyde gel for Northern blot analyses. The loading of equal amounts of RNA was confirmed by ethidium bromide staining of the 18S and 28S rRNA. (A) Analysis of the major immediate-early gene region. The Northern blot was hybridized to a probe corresponding to the HindIII L region of MCMV in order to detect RNA from ie1 (2.7 kb), ie2 (1.7 kb), or ie3 (2.7 kb). (B) Analysis of the m142 and m143 immediate-early gene region. The Northern blot shown in panel A was stripped and hybridized with a probe corresponding to m142 and m143 (bases 198832 to 201744 of the MCMV genome) within HindIII-I. The faint band visible between the m142 and m143 transcripts represents incompletely stripped ie1/ie3 signal.

Attachment and penetration of RV10 during infection of macrophages. The above-presented results indicated that expression of at least two MCMV immediate-early gene regions in infected IC-21 macrophages was compromised by deletion of the M139 to M141 region. Because of these findings and the fact that M139, M140, and M141 are expressed at early times after infection (12), we considered the possibility that these gene products influence entry of MCMV into macrophages. It is possible that specific viral gene products are required for efficient receptor-mediated attachment to and/or penetration into a phagocytic cell. Unfortunately, standard procedures to assess the rate of viral penetration could not be used because MCMV does not form plaques in IC-21 cells. Therefore, we compared, by Southern blotting, the amounts of macrophage-associated RV10 DNA and RV10Rev DNA, representing attached and/or penetrated virus.

View larger version (35K): [in this window] [in a new window]   FIG. 4.   Wild-type MCMV and RV10Rev attachment to or penetration into NIH 3T3 fibroblasts and IC-21 macrophages. (A) Penetration assay. NIH 3T3 cells were infected with WT MCMV at room temperature for 1 h, were shifted to 37°C, and at the indicated times after the temperature shift were washed with acid-glycine buffer. Monolayers were then overlaid with semisolid complete medium and incubated for 5 days to quantitate the number of plaques which formed in each treatment group. The percentage of bound virus which penetrated was calculated as follows: (the number of plaques formed with acid-glycine washing)/(the number of plaques formed without acid-glycine washing) × 100. (B and C) Viral DNA stably attached to or penetrated into fibroblasts and macrophages. NIH 3T3 fibroblasts (B) or IC-21 macrophages (C) were infected as described above at room temperature or at 4°C with RV10Rev (4 PFU/cell). At the indicated times after the shift to 37°C, monolayers were washed with acid-glycine buffer and returned for a total of 4 h of incubation in complete medium. Cells were then harvested for extraction of viral DNA and Southern blotting. The DNA was hybridized with a probe corresponding to HindIII-L. The C denotes control cells that were not washed with acid-glycine but were incubated for a total of 4 h after the shift to 37°C.

View larger version (62K): [in this window] [in a new window]   FIG. 5.   RV10 attachment to or penetration into NIH 3T3 fibroblasts and IC-21 macrophages. Experiments were performed essentially as described in the legend for Fig. 4B. In this experiment, monolayers of NIH 3T3 fibroblasts or IC-21 macrophages were infected with RV10 or RV10Rev (Rev), were shifted to 37°C, and at 120 min after the shift were washed with acid-glycine buffer. Infected cells were incubated for a total of 4 h in complete medium. Digestion of RV10 DNA with HindIII and BamHI yields bands of 2.8, 2.6, and 1.1 kb which hybridize to the HindIII-I probe. RV10Rev DNA digested with the same enzymes yields bands of 4.1, 2.8, and 2.6 kb which hybridize to this probe.

Virulence of RV10 in SCID mice. If macrophages are a prominent target cell of MCMV infection in vivo, then a mutant virus that replicates poorly in macrophages, such as RV10, is likely to be attenuated for virulence in vivo. This phenotype would not be evident in normal mice, such as BALB/c mice, because in these animals tissue culture-passaged virus is highly attenuated. Stocks of virulent RV10 cannot be obtained because replication of this mutant virus was not detectable in mouse salivary glands (52), as previously shown for mutant virus RV7 (2). Therefore, we compared the virulence of RV10 and RV10Rev in SCID mice. Because these mice are devoid of mature, functional T and B lymphocytes, they are exquisitely sensitive to MCMV infection, succumbing to as few as 3 PFU of tissue culture-passaged virus (42). Thus, these mice provide minimal control of MCMV replication within macrophages and other target cell types. The numbers and function of tissue macrophages, which serve as potential target cells, are normal in these otherwise immunodeficient mice (1).

View larger version (17K): [in this window] [in a new window]   FIG. 6.   Lethality of RV10 and WT MCMV for SCID mice. (A and B) Mice (four or five per group) were infected i.p. with 104 PFU of the indicated virus and observed daily for mortality. The experiments were performed twice.

Growth of RV10 in spleen and liver tissue. The above-presented data revealed a correlation between the virulence of MCMV and the ability of the virus to replicate in macrophages. However, RV10 may be replication defective for several cell types within target organs, and in sum, these deficiencies contributed to its high degree of attenuation in vivo. To test this hypothesis, we assessed the ability of RV10 to replicate in two major target organs of BALB/c mice, the spleen and liver. RV10 and RV10Rev were injected i.v. to directly administer virus to these two target organs. The spleen contains a dense network of macrophages in direct contact with the circulation. The liver also contains an abundance of macrophages, but they are diffusely distributed throughout the sinusoids. In addition, hepatocytes are permissive for MCMV replication (39, 44).

View larger version (16K): [in this window] [in a new window]   FIG. 7.   Growth of RV10 and WT MCMV in spleen and liver tissues. BALB/c mice were inoculated i.v. with 3 × 105 PFU of the indicated virus. Spleen (A) and liver (B) tissues were harvested from three individual mice at the indicated times postinfection. Virus titers in 20% (wt/vol) tissue homogenates were determined by plaque assay on NIH 3T3 cells. Data points are the average values for three individual mice, and error bars represent standard deviations.

Growth of RV10 in macrophage-depleted tissues. These data suggested that the ability of MCMV to replicate in macrophages may directly influence the level of virus replication in the macrophage-dense environment of the spleen. Is growth of RV10 in macrophages the limitation to replication in the spleen, or is RV10 replication defective for other cell types in this organ as well? In order to demonstrate that splenic macrophages were the limitation to efficient RV10 replication in this target organ, growth of RV10 in intact spleens was compared to that in spleens specifically depleted of tissue macrophages. Mice were depleted of splenic macrophages by i.v. injection of L-Cl₂MBP prior to MCMV infection with RV10 or RV10Rev. This method is highly specific for depletion of macrophages and macrophage-like dendritic cells, is nontoxic and, when administration is i.v., is selective for spleen and liver macrophages exclusively (see Materials and Methods). Virus titers in macrophage-depleted spleens and livers were quantitated on days 1 to 3 postinfection and compared to those in control mice with intact organs. As a control for L-Cl₂MBP administration, PBS, rather than empty liposomes, was chosen to ensure that control mice were normal with respect to macrophage numbers and, importantly, with respect to function as well (63). Histological examination of spleen tissue collected from WT virus-infected mice 3 days postinfection confirmed that L-Cl₂MBP treatment was effective in eliminating 80 to 90% of the acid phosphatase-positive marginal zone and red pulp macrophages and in maintaining depletion of these cells even after virus infection (Fig. 8).

View larger version (46K): [in this window] [in a new window]   FIG. 8.   Depletion of splenic macrophages by L-Cl2MBP treatment. Frozen tissue sections of spleens from mice injected i.v. with L-Cl2MBP (A) or PBS (B) and infected with 3 × 105 PFU of WT MCMV were stained for the macrophage-specific marker acid phosphatase. Mice received virus 48 h after L-Cl2MBP or PBS treatment, and spleens were harvested at 3 days postinfection. Acid phosphatase-positive cells in the red pulp, marginal zone, and white pulp stain bright red.

View larger version (14K): [in this window] [in a new window]   FIG. 9.   Growth of RV10 and WT MCMV in macrophage-depleted and intact tissues. Mice were injected i.v. with either L-Cl2MBP or PBS and 48 h later were infected with 3 × 105 PFU of the indicated virus. Spleens (A) and livers (B) were harvested from four or five individual mice at the indicated times postinfection. Virus titers in 20% (wt/vol) homogenates were determined by plaque assay on NIH 3T3 cells. Data points are the average values for individual mice, and error bars represent standard deviations. Rev designates infection with RV10Rev, the suffix -P designates treatment with PBS, and the suffix -L designates treatment with L-Cl2MBP.

Antiviral cytokine response to RV10 and RV10Rev infection of the spleen. An explanation for the data presented above is that splenic macrophages may be an initial target cell for MCMV replication and that in this cell type in vivo, production of WT virus may be efficient yet limited while replication of RV10 is severely curtailed. In the absence of such a macrophage filter, MCMV may have direct access to more-permissive cell types that have not yet been identified. However, there is an alternative explanation for the finding that MCMV titers were higher in macrophage-depleted spleens. Macrophages are likely a major source of antiviral cytokines and regulators of NK cell activity early in MCMV infection (36, 37, 46). The antiviral activity of cytokines and NK cells acting upon growth-retarded RV10 may significantly inhibit mutant virus production, conferring the attenuated phenotype in vivo. In the absence of macrophages and thus the antiviral activity, perhaps RV10 has the opportunity to replicate to substantial levels in other types of cells that normally support only low levels of RV10 replication. If this hypothesis is correct, then RV10 infection of intact spleens should induce levels of antiviral cytokines comparable to those induced by WT viruses in this organ. In the absence of macrophages, RV10 should induce significantly lower levels of cytokines.

View larger version (37K): [in this window] [in a new window]   FIG. 10.   Cytokine levels produced by splenocytes in response to RV10 or WT MCMV infection of normal or macrophage-depleted mice. Mice were administered L-Cl2MBP (L) or PBS (P) i.v., and 48 h later, they were inoculated i.v. with mock virus preparation or with RV10Rev or RV10 (3 × 105 PFU). Thirty-six hours later, spleens were harvested from three individual mice, and suspensions containing 5 × 106 splenocytes/ml were incubated overnight. After 24 h, supernatants were collected and cytokines were quantitated by sandwich ELISA assays. IL-12 data are for IL-12 p40. Values represent the mean titers and error bars denote standard deviations for three mice. A single asterisk denotes a significant difference (P  0.05) between L-Cl2MBP- and PBS-treated mice infected with RV10Rev. A double asterisk denotes a significant difference (P  0.05) between RV10- and RV10Rev-infected mice treated with PBS. The data are representative of two separate experiments.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Results of in vitro studies using mutant virus RV10 suggest that the products of MCMV M139, M140, and/or M141 function to regulate growth of this virus in macrophages, a biologically important target cell. Mutant RV10 is impaired in the early stages of MCMV replication in macrophages, i.e., at or preceding the time of viral immediate-early gene expression. The levels of expression of at least four immediate-early genes from two different regions of the MCMV genome, HindIII-L and HindIII-I, were significantly lower in RV10-infected macrophages than in RV10Rev-infected cells. The results of experiments performed to semiquantitatively assess the amount of virus stably bound to or penetrated into macrophages indicate that the M139, M140, and/or M141 gene product likely functions in efficient MCMV attachment to or penetration into macrophages. Products of M139, M140, or M141 apparently do not confer resistance to the antiviral effects of IFN-/, which is produced early after infection of macrophages, because neutralization of these type I interferons does not restore growth of RV10 in IC-21 macrophages (52). Current studies aim to (i) identify the gene(s) which bestows efficient replication of MCMV in macrophages, (ii) characterize the gene product(s), and (iii) assess the function of the gene product(s) in specific stages of MCMV entry into macrophages or as a transcriptional transactivator of immediate-early genes.

Through the use of RV10 and L-Cl₂MBP, several interesting yet paradoxical findings concerning the role of tissue macrophages in MCMV pathogenesis were obtained. The fact that RV10 was defective for growth in macrophages in vitro and in spleen tissue in vivo supports the notion that efficient replication of MCMV in macrophages is required for abundant growth in at least this target organ. This cell type provided the limitation to RV10 replication in the spleen, as RV10 grew to nearly WT levels in macrophage-depleted spleens. The data on MCMV replication in L-Cl₂MBP-treated compared to PBS-treated spleens, however, clearly demonstrate that splenic marginal-zone macrophages, marginal metallophilic macrophages, red pulp macrophages, and marginal-zone dendritic cells are not required for replication of MCMV in the spleen. In contrast, these cells provide a net protective effect.

The fact that efficient replication of MCMV in macrophages appeared to be a prerequisite for virus replication in the spleen even though macrophage and dendritic cell depletion enhanced virus growth seems paradoxical. However, this may be explained by the concept of macrophages/dendritic cells serving as a protective barrier or filter. In the spleen, macrophages are likely one of the first cell types exposed to blood-borne virus. The abilities of macrophages to (i) destroy, through phagocytosis, a percentage of extracellular virions, (ii) produce antiviral cytokines, and (iii) regulate growth of HCMV (8, 9, 25) or MCMV (17, 66) which enters macrophages via receptor-mediated events collectively may confer their protective function. Consequently, the ability of MCMV to replicate efficiently in this cell type and subsequently seed neighboring, perhaps more-permissive cell types may ultimately determine virus titers in the spleen.

Our data indicate that macrophages contribute directly or indirectly to production of two cytokines, TNF- and IFN-, both of which inhibit MCMV replication (17, 28). The latter cytokine is produced early in MCMV infection by activated NK cells (38). Indeed, independent studies have shown that L-Cl₂MBP treatment of C57BL/6 mice prior to MCMV infection eliminated NK cell infiltration in the spleen, as assessed by quantitation of NK1.1⁺ and/or DX5⁺ CD3 splenocytes and by killing of NK-sensitive YAC-1 cells (45). This is likely due to a dependence of NK cell infiltration, differentiation, or activation on cytokines or chemokines produced by macrophages or marginal-zone dendritic cells (47, 48). For example, Kupffer cells are required for the differentiation of peripheral blood NK cells into highly activated, hepatic NK cells (60).

Comparisons between the cytokine profiles of macrophage-depleted spleens and intact spleens from mice infected with RV10 indicate that differences in cytokine production cannot explain the restored growth of RV10 in macrophage-depleted animals. There were no significant differences between the two treatment groups for RV10-induced IL-6, IL-12, IFN-, or TNF- levels. Although we cannot rule out the possibility that the levels of other antiviral cytokines induced by RV10 but not quantitated in this study were diminished by macrophage depletion, we conclude that the elimination of macrophages as a target cell for RV10 had the greatest influence on RV10 replication in the spleen.

The ability of MCMV to replicate in the spleen appeared to influence the overall degree of virulence, an observation previously reported (19). In this study, a direct correlation between the ability of MCMV to replicate in the spleen and lethality in SCID mice was noted. RV6, which replicated in the spleen but replicated poorly compared to WT virus or RV10Rev, killed SCID mice with delayed kinetics. We have found that other mutant MCMV which replicate in the spleen, but only poorly compared to WT virus, kill SCID mice with delayed kinetics (13). RV10, which failed to replicate to detectable levels in the spleen, did not kill SCID mice within 90 to 100 days. It is not feasible to repeat the lethality studies using L-Cl₂MBP-treated SCID mice, as most all macrophage subpopulations in the spleen would be restored after L-Cl₂MBP treatment by the expected time of MCMV-induced death (62), and effects of repeated L-Cl₂MBP injections have not been assessed in this immunocompromised host.

Macrophages had a minor role in control of MCMV replication in the liver compared to that of replication in the spleen. The mutant virus RV10, which failed to replicate to detectable levels in the spleen, replicated to WT levels in the liver. Depletion of Kupffer cells enhanced levels of MCMV replication in the liver, but to a lesser extent than in the spleen. The liver contains large numbers of tissue macrophages, diffusely distributed throughout the sinusoids, in contrast to the highly organized, dense network of splenic macrophages. Aside from this anatomical difference, the quality or function of tissue macrophages residing in the liver and spleen may differ. For example, the peaks of the titers of type I interferons are delayed and the titers are significantly lower in the livers of MCMV-infected BALB/c mice compared to those in the spleen (67). Other antiviral immune effector cells, such as NK cells, function divergently in these two organs (57). Furthermore, there is evidence that in the liver, hepatocytes are the preferred target cell for MCMV replication (39, 44). Collectively, these differences between the liver and the spleen may explain the tissue-specific discrepancies in RV10 replication and in the effect of L-Cl₂MBP treatment on MCMV replication in these tissues.

The results of these studies support the hypothesis that macrophages play a pivotal role in the pathogenesis of MCMV infections. Initially, these cells may serve as target cells in at least the spleen, and the capacity of the virus to replicate efficiently in macrophages influences, in part, the magnitude of MCMV replication in that organ. While efficient replication in macrophages may be a necessary (although not sufficient) requirement for robust virus replication in this important target organ, MCMV replication in this cell type leads to controlled growth of the virus, likely due collectively to the growth rate of the virus in this cell type and the early induction of antiviral cytokines. These two conditions may provide a net protective role for at least splenic macrophages and may favor the eventual establishment of MCMV latency in macrophages. It is likely that complex virus-macrophage interactions which are beneficial to both the virus and the host exist, ensuring persistence of CMV within the host.

	ADDENDUM

A recent study using L-Cl₂MBP-treated mice also revealed a protective role for tissue macrophages during the early stages of acute MCMV infection (S. Hamano, H. Yoshida, H. Takimoto, K. Sonoda, K. Osada, X. He, Y. Minamishima, G. Kimura, and K. Nomoto, Microbiol. Immunol. 42:607-616, 1998).