In Vivo Monocyte Tropism of Pathogenic Feline Immunodeficiency Viruses

Abstract

Virus-infected monocytes rarely are detected in the bloodstreams of animals or people infected with immunodeficiency-inducing lentiviruses, yet tissue macrophages are thought to be a major reservoir of virus-infected cells in vivo. We have identified feline immunodeficiency virus (FIV) clinical isolates that are pathogenic in cats and readily transmitted vertically. We report here that five of these FIV isolates are highly monocytotropic in vivo. However, while FIV-infected monocytes were numerous in the blood of experimentally infected cats, viral antigen was not detectable in freshly isolated cells. Only after a short-term (at least 12-h) in vitro monocyte culture were FIV antigens detectable (by immunocytochemical analysis or enzyme-linked immunosorbent assay). In vitro experiments suggested that monocyte adherence provided an important trigger for virus antigen expression. In the blood of cats infected with a prototype monocytotropic isolate (FIV subtype B strain 2542), infected monocytes appeared within 2 weeks, correlating with high blood mononuclear-cell-associated viral titers and CD4 cell depletion. By contrast, infected monocytes could not be detected in the blood of cats infected with a less pathogenic FIV strain (FIV subtype A strain Petaluma). We concluded that some strains of FIV are monocytotropic in vivo. Moreover, this property may relate to virus virulence, vertical transmission, and infection of tissue macrophages.

Immunodeficiency-inducing lentiviruses such as human immunodeficiency virus (HIV), simian immunodeficiency virus, and feline immunodeficiency virus (FIV) infect T lymphocytes and macrophages in lymphoid tissues, and FIV-infected macrophages have been found in lymphoid tissues, brain tissue, bone marrow, and the peritoneal cavity (3, 4, 12, 22, 29, 33, 40). In HIV-infected humans and SIV-infected nonhuman primates, tissue macrophages and CD4⁺ T cells are considered the major reservoirs of virus infection (14, 19, 30, 32, 41). In addition, recent evidence suggests that HIV-infected macrophages in sanctuary sites are resistant to even prolonged treatment with potent antiretroviral agents (38). Moreover, the probable involvement of chemokine receptors on macrophages and dendritic cells in HIV infection (1, 5, 7, 11, 16, 19, 22, 26) points to the importance of mononuclear phagocytes in lentivirus host-virus interactions.

While lentivirus-infected macrophages are expected in certain tissues, it remains unclear whether these macrophages are infected in tissues, blood, or bone marrow. For example, while HIV-infected monocytes are considered rare in the blood, macrophage-tropic HIV strains can be recovered by coculture of peripheral blood mononuclear cells (PBMC) with monocyte-derived macrophages (43) and infected macrophages can be demonstrated readily in situ in lymph node, brain, and lung tissues (2, 14, 19, 30, 35). FIV infection of macrophages has been demonstrated in vivo and can be accomplished in vitro (3, 4, 12). However, as with HIV, previous studies have failed to detect FIV infection in circulating monocytes (6, 15). By contrast, two lentiviruses of ungulate species (visna virus and caprine arthritis-encephalitis virus [CAEV]) have been shown to infect circulating monocytes and tissue macrophages in vivo (20, 21, 33, 37).

Our interest in FIV-monocyte relationships arose during the screening of FIV clinical isolates for in vivo pathogenicity, when we detected virus-bearing monocytes in cats infected with five isolates. Here we describe studies of the monocytotropism of one of these isolates (FIV subtype B strain 2542), a strain shown in separate studies to be transmissible mucosally and vertically (34) and to induce high viral RNA titers in plasma and produce an accelerated immunodeficiency syndrome upon serial intravenous passage in vivo (9, 10). We suggest that some strains of FIV are monocytotropic in vivo and that this property may correlate with viral virulence and transmissibility.

MATERIALS AND METHODS

Animals and viruses.

Specific-pathogen-free (SPF) animals were maintained in the Department of Pathology, Colorado State University, in accordance with standards consistent with U.S. Department of Agriculture, National Institutes of Health, and University Animal Care and Use Committee guidelines. The five FIV isolates used were obtained from blood samples of five naturally infected cats with various signs attributable to FIV infection (Table 1). Plasma from each of the proband infected cats was transferred into one SPF cat each. Each recipient SPF cat was monitored for FIV infection by immunoblotting, PBMC coculture, T-cell numbers, and other hematologic parameters. Recipient cats were also screened for antibodies to feline spumivirus.

TABLE 1.

Origins of the monocyte-tropic FIV clinical isolates used in this study^a

FIV isolate	Viral env subtype	Clinical status of source cat	Geographic origin
2542	B	Immune deficiency, chronic diarrhea	San Francisco, Calif., area
2531	A	Immune deficiency, chronic diarrhea	San Francisco, Calif., area
2561	B	Immune deficiency, weight loss	Chicago, Ill., area
2560	B	Chronic infections	San Francisco, Calif., area
2546	A	Immune deficiency, neurologic disease	San Francisco, Calif., area

^aFive FIV isolates were obtained from five naturally infected cats exhibiting various clinical signs of immunosuppression. Plasma from each index cat was used to inoculate groups of four SPF cats per isolate. Inoculated SPF cats were monitored for FIV infection and shown to be negative for feline spumivirus infection by serologic analysis. 

Ten months after the initial recipient cats became FIV coculture positive, each FIV isolate was passaged by inoculation of 5.0 ml of whole blood into six groups of four SPF cats each and infection in these newly inoculated cats was monitored as described above. In total, 24 SPF cats were infected with three different FIV subtype (clade) B isolates (2542, 2560, and 2561) and two different FIV subtype AB isolates (2531 and 2546). At 12 and 16 weeks postinoculation (p.i.), the level of monocyte infection was assessed. Blood from an FIV subtype B strain 2542-infected cat with the highest level of monocyte infection was transferred to four additional SPF cats which had been used in an early pathogenesis study (see below).

Two isolates of subtype FIV A Petaluma (36) (kindly provided by Niels Pedersen, University of California, Davis) were also studied in vivo. One FIV subtype A Petaluma isolate maintained by continuous in vivo passage at the University California, Davis, was passaged into two SPF cats by intravenous inoculation of 5.0 ml of blood. A second version of FIV subtype A Petaluma that had been maintained in Crandell feline kidney (CrFK) cells was inoculated into three SPF cats by intraperitoneal injection of 10⁶ FIV-infected CrFK cells. Virus stocks for in vitro inoculation experiments were prepared from supernatants of CrFK cells persistently infected with either FIV subtype A strain Petaluma or FIV subtype B strain 2542 (Table 1) and adjusted to contain 100 μg of p26 Gag antigen per ml.

Feline monocyte culture.

PBMC were recovered from blood by Ficoll gradient centrifugation and resuspended at 2 × 10⁶/ml of medium. A 200-ml volume of the PBMC suspension was then added to wells of an eight-well chamber culture slide (Nunclon, Naperville, Ill.) or to wells of a 96-well plate that had been precoated with 10 μg of affinity-purified feline immunoglobulin G (IgG) (Kirkegaard & Perry Laboratories, Gaithersburg, Md.) per ml. Cells were allowed to adhere for 1 h at 37°C, and nonadherent cells were removed by vigorous pipetting with phosphate-buffered saline (PBS). The wells were then refilled with monocyte culture medium (Dulbecco’s modified Eagle medium with a high glucose concentration [Sigma Chemical Co., St. Louis, Mo.], 5% heat-inactivated fetal bovine serum plus 5% heat-inactivated calf serum [Hyclone, Logan, Utah], 1% penicillin plus streptomycin, 5 mM 2-mercaptoethanol, 10 mg of polymyxin B per ml). The medium was supplemented with 20 ng of human recombinant interleukin-4 (rIL-4; R&D Systems, Minneapolis, Minn.) per ml, which promoted maximal survival and differentiation of feline monocytes in vitro. Monocytes were derived from bone marrow in the same manner as described for blood.

Characterization of cultured feline monocytes.

The phenotype and purity of in vitro-cultured feline monocytes/macrophages were assessed in several ways. Nonspecific esterase activity was determined by using a commercial assay kit (Sigma Chemical Co.). Expression of low-density lipoprotein receptors was assessed by incubation of cultured monocytes for 4 h with 20 ml of DiI-labeled acetylated low-density lipoprotein (Biomedical Technologies, Stoughton, Mass.) per ml, followed by fixation in paraformaldehyde and epifluorescence microscopy. Phagocytic ability was assessed by incubation of live cells for 30 min with a 10-mg/ml suspension of opsonized zymosan (Sigma Chemical Co.), followed by three washes in PBS, fixation in methanol, and Giemsa counterstaining. Expression of the ligand for the lectin Ricinus communis agglutinin I (RCA-I; Vector Laboratories, Carpinteria, Calif.) was assessed by incubation of fixed monocytes with biotinylated RCA-I, followed by incubation with streptavidin-horseradish peroxidase and addition of aminoethylcarbazole as a substrate. Purity of cultures was assessed by microscopic examination and by flow cytometry to detect contaminating lymphocytes as described previously (40). Three-day-cultured monocytes were detached by 30 min of incubation with 12 mM lidocaine in PBS at 37°C and then washed and immunostained with fluorescein isothiocyanate (FITC)-conjugated monoclonal antibodies (MAbs) to feline CD4 and CD8 and analyzed on a EPICS C flow cytometer (Becton Dickinson) (6, 8).

Immunocytochemical and fluorescence detection of FIV antigens in infected monocytes.

Cultured monocytes from FIV-infected and uninfected control cats were immunostained by using high-titered anti-FIV serum from an SPF cat infected with FIV subtype A strain Petaluma for 2 years as described previously for detection of in vitro-infected cells (12). The FIV-immune serum recognized all major FIV structural proteins, as determined by immunoprecipitation (39). Control serum was obtained from an age-matched, FIV-negative, SPF cat. Monocytes on chamber slides were fixed for 2 min with a 1:1 solution of acetone-methanol at −20°C, rinsed in PBS, and then incubated with a 0.2% solution of hydrogen peroxide in methanol for 5 min to block endogenous peroxidase activity. After being rinsed in PBS, cells were incubated with a 1:400 dilution of FIV antiserum in PBS with 1% bovine serum albumin and 5% normal goat serum at room temperature for 40 min. After being rinsed in PBS, cells were incubated for 25 min with biotin-conjugated goat anti-cat IgG (Kirkegaard & Perry Laboratories) diluted 1:400 in immunofluorescence assay buffer. After being rinsed in PBS, cells were incubated with either streptavidin-peroxidase (Zymed Laboratories, San Francisco, Calif.) diluted 1:500 in PBS for 15 min. For light microscopy, the substrate aminoethylcarbazole (Sigma Chemical Co.) was added for 12 min. After washing, cells were counterstained for 5 min with hematoxylin and photographed. Negative controls included (i) monocytes from FIV-infected cats reacted with nonimmune cat serum, (ii) monocytes incubated only with the secondary antibody, and (iii) monocytes from uninfected SPF cats and SPF cats infected with feline leukemia virus (FeLV) and feline infectious peritonitis virus incubated with the FIV immune serum.

An indirect immunofluorescence assay was also used to localize FIV antigens in cultured monocytes as described previously for detection of FIV antigens in cultured neural cells (12). Briefly, adherent monocytes at various times in culture were fixed and then reacted with FIV-immune or control cat serum and then with FITC-conjugated, affinity-purified goat anti-cat IgG (Kirkegaard & Perry). Slides were examined by epifluorescence microscopy and photographed. A MAb (51G1.1) (13) also was used to detect FIV p26 by immunofluorescence in cultured monocytes by using a biotinylated goat anti-mouse IgG, followed by streptavidin-FITC (Boehringer Mannheim, Indianapolis, Ind.). Negative controls included an isotype-matched MAb to FeLV p26, omission of the primary antibody, and monocytes from uninfected cats. A fibroblastoid cell line (LNC) persistently infected with FIV subtype B-2542 served as a positive control.

PCR detection of FIV in monocytes.

Monocytes were obtained from peripheral blood by adherence to tissue culture plastic for 1 h, followed by extensive washing (as described above). The adherent cells were then lysed and digested in 500 μl of a solution of 10 mM Tris, 50 mM KCl, 100 μg of gelatin per ml, 0.45% Nonidet P-40, 0.45% Tween 20, and 20 μg of proteinase K per ml for 1 h at 50°C. A 1.0-μl aliquot of this solution was then subject to 35 cycles of PCR, using FIV env primers, as described previously (45). After the PCR, samples were assayed by agarose gel electrophoresis for a DNA band of the appropriate size. Negative controls included monocyte samples from uninfected cats, and a plasmid containing the FIV subtype A strain Petaluma env sequence was included as a positive control.

Quantitation of monocyte infection and screening for monocytotropic FIV isolates.

To quantitate the level of monocyte infection in a given animal, the percentage of infected monocytes was determined by immunocytochemical analysis of monocytes cultured for 72 h in chamber slides. Slides were examined by light microscopy, and at least 200 cells per field were evaluated. The number of FIV-positive cells was divided by the total number of cells counted (FIV positive plus FIV negative) to determine the percentage of infected monocytes. A positive control (monocytes from an FIV-infected cat with known persistent high levels of monocyte infection) and a negative control (monocytes from an uninfected cat) were included with each monocyte quantitation experiment.

Enzyme-linked immunosorbent assay (ELISA) for FIV antigens in monocytes.

PBMC (10⁶) were added to triplicate wells of a 24-well plate, allowed to adhere for 1 h, washed free of nonadherent cells, and cultured in 300 ml of monocyte medium. At various culture time points, the concentration of p26 was determined in cell supernatants and lysates (lysis in 100 ml of Tris-EDTA–2% fetal bovine serum–2% bovine serum albumin–2% Tween 20) as previously described (13). Positive controls included FIV-infected CrFK cells lysed in a similar manner.

Adherence-independent monocyte culture and cytokine stimulation.

To obtain in vitro-differentiated monocytes/macrophages under nonadherent culture conditions, PBMC were cultured on a feline fibroblast monolayer. Under these conditions, most of the nonmonocytic cells died over the 3- to 4-day culture period while the monocytes survived in suspension and assumed morphologic and phenotypic characteristics of mature macrophages, including increased size and expression of nonspecific esterase activity. After 3 days in culture, the nonadherent monocytes were removed from the stromal cell monolayer and replated for 1 h onto uncoated tissue culture plastic, to which they rapidly adhered. The adherent cells were fixed, and the percentage of virus antigen-positive monocytes was determined by immunocytochemical analysis. Several different cytokines and test substances were added to nonadherent monocyte cultures to evaluate the effect on virus antigen expression. The test substances were added for the 3-day culture period and included 10 ng of human rIL-4 (R&D Systems) per ml, 10 ng of recombinant human tumor necrosis factor alpha (TNF-α; R&D Systems) per ml, 100 U of human rIL-6 (Boehringer Mannheim) per ml, 10 ng of recombinant human granulocyte-macrophage colony-stimulating factor (GM-CSF) (Immunex, Seattle, Wash.) per ml, 10 mg of lipopolysaccharide (LPS; Sigma) per ml, and 10 nM phorbol myristate acetate (PMA; Sigma). The cross-species activity of the four recombinant cytokines had been determined previously in bioassays done with feline leukocytes (data not shown).

In vivo monocytotropic viral pathogenesis study.

The early response to viral infection with monocytotropic FIV subtype B strain 2542 was assessed in four 12-week-old SPF cats (2883, 2884, 2887, and 2889) which were inoculated intravenously with 5 ml of blood from a clinically ill SPF cat infected with FIV subtype B strain 2542. Four age-matched control cats were inoculated with 5 ml of whole blood from an uninfected SPF cat. These eight cats were then evaluated weekly for 12 weeks and then twice monthly thereafter to determine PBMC-associated virus titers, levels of monocyte infection, T-cell numbers, and signs of clinical illness. Clinical signs of FIV infection included lymphadenopathy, weight loss, diarrhea, gingivitis, and wasting, as reported previously (8, 10).

PBMC-associated virus titration by coculture.

PBMC-associated virus was titrated by coculture of PBMC from infected cats with mitogen-pulsed PBMC from naive cats. Titer were expressed as numbers of tissue culture-infective doses per 10⁶ input PBMC (28). Briefly, 10-fold serial dilutions of PBMC from infected or control cats were added to triplicate wells of 96-well plates. To each well were then added 5 × 10⁵ PBMC from FIV-naive SPF cats prestimulated for 3 days with concanavalin A. Cultures were maintained with periodic medium changes for 1 month. FIV replication was detected weekly by capture of p26. The viral titer was expressed as the lowest cell dilution (100% tissue culture-infective dose) that gave three of three positive wells.

Hematologic and clinical monitoring.

CD4⁺ and CD8⁺ lymphocytes in PBMC were quantitated flow cytometrically as described previously (6, 8). Clinical signs were assessed by weekly physical examination of each cat as described previously (8, 10).

RESULTS

In vitro survival of feline monocytes is favored by IL-4.

It has been difficult to maintain viable feline monocytes in vitro. Survival and differentiation of feline monocytes were enhanced by human rIL-4 (10 to 20 ng/ml), IgG-coated wells, and medium containing 5% newborn calf serum (in addition to 5% fetal calf serum). Under these conditions, feline monocytes were obtained at high purity (Fig. 1) and maintained in culture for up to 2 weeks. After 72 h in culture, monocytes were uniformly positive for nonspecific esterase, RCA-I binding, phagocytosis of zymosan, and uptake of acetylated low-density lipoprotein (Fig. 1). The cultures contained <1% contaminating T cells as determined by flow cytometry. Thus, adherent feline blood mononuclear cells cultured in human rIL-4 consisted of a nearly pure population of monocytes/macrophages.

FIG. 1.

Enhancement of in vitro culture of feline monocytes by human rIL-4. Feline monocytes were obtained from PBMC by 1 h of adherence to plastic chamber slides that were precoated with purified feline IgG. After being washed to remove nonadherent cells, the monocytes were cultured in monocyte medium supplemented with human rIL-4 at 20 ng/ml (see Materials and Methods). Over the next 3 to 4 days of culture, the monocytes increased in size and assumed phenotypic characteristics of macrophages. Cells were fixed and stained for detection of nonspecific esterase activity, which revealed an essentially pure population of monocyte-derived macrophages (magnification, ×200).

FIV is expressed ex vivo by in vivo-infected monocytes.

An immunofluorescence assay using either serum from an FIV-infected cat (Fig. 2a and b) or an FIVp26 Gag-specific MAb (Fig. 2c) detected strong intracytoplasmic FIV antigen expression in short-term-cultured but not freshly isolated monocytes from cats infected with monocytotropic FIV subtype B strain 2542 and four other viral isolates (Table 1). Neither the FIV immune serum nor the Gag-specific MAb stained monocytes cultured from FIV-negative cats or from cats infected with either FeLV or feline infectious peritonitis virus (data not shown).

FIG. 2.

Identification of in vivo FIV-infected monocytes by in vitro culture and immunofluorescence assay. Monocytes obtained from a cat infected with the FIV isotype B isolate 2542 were cultured in vitro for 3 days, fixed, and then immunostained for detection of FIV antigens. Cells were reacted first with FIV-immune cat serum (a), nonimmune cat serum (b), an anti-FIV p26 MAb (c), or an irrelevant isotype-matched MAb (d). The cells were then incubated with the appropriate FITC-conjugated secondary antibodies as described in Materials and Methods. Numerous FIV-positive monocytes were detected by staining with anti-FIV serum (a) or the anti-p26 MAb (c), whereas staining was minimal with nonimmune cat serum (b) or the irrelevant MAb (d). Monocytes from noninfected cats also did not stain with FIV-immune cat serum (data not shown) (magnification, ×200).

FIV antigens were not be detectable in cells fixed immediately after in vitro adherence but were detectable by 12 to 24 h ex vivo (Fig. 3). FIV DNA, however, was detected by single-round and nested PCRs in monocytes from three of three cats immediately after in vitro adherence, indicating the pre-existence of the viral DNA in circulating monocytes (data not shown). The percentage of FIV Gag-bearing monocytes was maximal at 24 h postculture and decreased thereafter (Fig. 3). Thus, viral infection did not appear to spread horizontally during in vitro culture.

FIG. 3.

Kinetics of FIV antigen expression in in vivo FIV-infected monocytes during in vitro culture as assessed by immunocytochemical analysis. Monocytes were obtained from an FIV-infected cat by adherence to plastic for 1 h and then fixed either immediately after adherence (a) or after 24 h (b), 48 h (c), or 72 h (d) in culture. The monocytes were then immunostained to detect FIV antigens by using an immunoperoxidase technique as described in Materials and Methods. Expression of FIV antigens was not detected in monocytes immediately after adherence but then appeared and increased with time. However, the percentage of antigen-expressing monocytes decreased over a 3-day culture period, indicating that the infection did not spread to uninfected cells. (magnification, ×200).

Naive feline monocytes are resistant to FIV infection in vitro.

Despite extensive washing to remove nonadherent cells, it was important to exclude the possibility that contaminating virus-infected lymphocytes or other cells served as a source for monocyte infection after in vitro culture. We therefore determined whether naive monocytes were susceptible to FIV infection in vitro. Fresh supernatants from mitogen-stimulated PBMC from infected cats that had high levels of virus infection or from CrFK cells infected with culture-adapted FIV subtype B strain 2542 were added to cultures of blood monocytes. The number of viral antigen-positive monocytes never exceeded ≥1 to 2%. Thus, cultured monocytes from FIV-naive cats proved highly refractory to in vitro infection with FIV stocks prepared from either lymphocytotropic or macrophage-tropic FIV isolates. Moreover, the number of antigen-positive monocytes decreased with time in culture, indicating absence of horizontal spread in the cultures.

These results, taken together with those described above, suggested that FIV replication was rapidly upregulated by in vitro adherence and culture of provirus-bearing monocytes.

FIV infection in blood and bone marrow monocytes is correlated.

The presence of FIV-infected monocytes in blood coupled with inherent monocyte resistance to in vitro infection suggested that monocyte precursors become infected in the bone marrow. Therefore, fresh bone marrow mononuclear cells and PBMC from four FIV subtype B strain 2542-infected cats were cultured by the same methods. By 72 h of culture, >90% of the marrow-adherent cells expressed phenotypic and biochemical markers similar to those of blood monocytes (esterase positivity, low-density lipoprotein uptake, major histocompatibility complex class II positivity, and zymosan phagocytosis). At this time, equivalent numbers of bone marrow- and blood-derived monocytes/macrophages expressed FIV antigens, as detected by immunocytochemical analysis (Fig. 4 and 5).

FIG. 4.

Infection of bone marrow monocytes/macrophages with FIV in vivo. Bone marrow mononuclear cells were obtained by aspiration from the humerus of an FIV-infected cat. Bone marrow monocytes/macrophages were enriched by adherence for 1 h, washed, and then cultured for 3 days in monocyte medium. After the cells were fixed, FIV antigen expression was detected by immunocytochemical analysis. Immediately after adherence (a), FIV antigens could not be detected in bone marrow-derived monocytes. However, strong FIV expression could be detected in numerous bone marrow-derived monocytes/macrophages after 72 h in culture (b).

FIG. 5.

Similar levels of monocyte infection in blood and bone marrow. Monocytes were obtained from blood (black bars) and bone marrow (light bars) from four different cats 10 weeks after inoculation with in vivo-passaged FIV subtype B strain 2542. Monocytes were cultured for 72 h in monocyte medium, and then the percentage of FIV-infected monocytes in each culture was quantitated by immunohistochemical analysis. Similar results were obtained in one additional experiment using blood and bone marrow specimens from the same four cats.

The above-described results were consistent with the tenet that monocyte precursors become infected in bone marrow. However, it was also possible that monocytes became infected by exposure to virus in plasma. Therefore, monocytes were cultured in plasma containing a high titer of infectious FIV subtype B strain 2542. FIV-positive monocytes could not be detected after up to 7 days of culture (data not shown), suggesting that infection by virus in plasma is a relatively inefficient mechanism to explain the presence of FIV-bearing monocytes in the circulation.

FIV production by monocytes is predominantly intracellular.

To localize FIV production in in vivo-infected monocytes, blood monocyte cultures were established from cats infected with monocytotropic or nonmonocytotropic FIVs (Table 1) and p26 in culture supernatants versus cells was analyzed from 1 to 72 h. Intracellular p26 was detected at 12 h, and its level peaked at 24 h and declined thereafter (Fig. 6). By contrast, p26 was not detectable (or was rarely barely detectable) in supernatants throughout the 72 h course of study (Fig. 6). The concentration of cell-associated p26 detected by antigen capture ELISA and the percentage of p26⁺ monocytes detected by immunocytochemical analysis were closely correlated. Intracellular p26 and FIV⁺ monocytes declined concurrently with time in culture. While extracellular p26 concentrations were extremely low, some extracellular infectious virus was produced since inoculation of naive feline lymphoblasts with monocyte supernatants resulted in productive infection (data not shown). Thus, most ex vivo FIV antigen expression remained intracellular.

FIG. 6.

FIV p26 Gag expression by in vivo-infected monocytes upon in vitro culture. A p26 (CA) capture ELISA was used to quantitate FIV expression by monocytes obtained from a cat infected with FIV subtype B strain 2542. Monocyte cultures were established from PBMC (as described in Materials and Methods) by using triplicate wells of a 24-well plate. At various time points during culture, supernatants were harvested from triplicate wells and the adherent cells in each well were then lysed in 1.0 ml of buffer containing 0.1% Triton X-100. The p26 concentrations in supernatants (□) and lysates (●) were determined by ELISA, and the mean (± the standard error) p26 concentration was plotted versus time in culture. Intracellular expression of p26 by in vivo-infected monocytes was first detectable at 12 h in culture, increased by 24 h in culture, and decreased thereafter, whereas p26 antigen was virtually undetectable in supernatants.

Monocyte adherence triggers FIV expression.

Studies with the CAEV and visna virus systems have suggested that monocyte maturation into macrophages provides a trigger for virus antigen expression in vivo (20, 33). To address this issue in the FIV system, we exploited the observation that feline monocytes can be maintained under nonadherent conditions yet still undergo phenotypic maturation into macrophages if cultured on a feline stromal cell feeder layer. PBMC culture for 3 days on fibroblast underlayers resulted in the death of most lymphocytes, leaving an almost pure population of monocyte-derived macrophages in suspension, as assessed by phenotypic and histochemical criteria. These macrophages could be further purified by replating on uncoated tissue culture plastic, to which they readily adhered. Use of this system allowed us to compare the relative effects of adherence versus maturation on virus antigen expression in in vivo-infected monocytes.

Compared with adherent monocytes (Fig. 3), few nonadherent monocytes from FIV subtype B strain 2542-infected cats expressed detectable FIV antigens (Fig. 7A). However, culture of these nonadherent monocytes/macrophages in the presence of 10 nM PMA induced high-level FIV antigen expression nearly equivalent to that induced by adherence alone (Fig. 7B). By contrast, addition of neither IL-4, IL-6, TNF-α, GM-CSF, nor LPS induced FIV antigen expression (Fig. 7C). To examine the possibility that the fibroblast feeder layer inhibited virus antigen expression in cultured monocytes, we added supernatants from fibroblast cultures prior to the adherence of in vivo-infected monocytes. No blockage of virus antigen expression was detectable (data not shown). Thus, these results suggest that adherence of monocytes in tissues can provide an in vivo trigger for FIV antigen expression.

FIG. 7.

Triggering of FIV expression in in vivo-infected monocytes by adherence. To evaluate the influence of adherence on virus antigen expression by in vivo-infected monocytes, PBMC were obtained from the blood of an FIV-infected cat and cultured either under adherent conditions (adherence to tissue culture plastic) or under nonadherent conditions (culture on a feline fibroblast monolayer) for 3 days. During 3 days of culture on a fibroblast monolayer, monocytes remained fully viable but nonadherent and could be obtained by gentle washing and purified by 1 h of adherence to tissue culture plastic and fixed. The percentage of FIV-positive monocytes after culture under adherent or nonadherent conditions was then determined by immunocytochemical analysis. Virus antigen expression by nonadherent monocytes/macrophages was minimal (a), compared to that by monocytes cultured continuously under adherent conditions (Fig. 3). However, when the nonadherent monocytes/macrophages were cultured for 3 days in the presence of 10 nM PMA, virus expression was strongly upregulated (b). Other cytokines and macrophage stimulants were also evaluated for the ability to upregulate virus antigen expression by nonadherent monocytes/macrophages, including LPS at 10 μg/ml, IL-4 at 20 ng/ml, GM-CSF at 10 ng/ml, and TNF-α at 10 ng/ml (c). Only culture in the presence of 10 nM PMA stimulated virus expression by nonadherent monocytes/macrophages (c), although the level of expression was still less than that in monocytes cultured under continuously adherent conditions. Similar results were obtained in one additional experiment.

FIV monocytotropism is maintained in in vivo passage.

Our five clinical FIV isolates and the prototype Petaluma strain (obtained from Neils C. Pedersen [36]) were evaluated for their relative monocytotropism by in vivo passage into groups of four or five SPF cats each. At 12 and 16 weeks p.i., the frequency of monocyte infection was quantitated in each cat (Fig. 8). Virus-bearing monocytes were detected in cats infected with each of the five clinical isolates and were especially frequent in cats infected with two virus strains (FIV-B-2542 and FIV-B-2531) (Fig. 8). In contrast, infected monocytes were not detected in any of the FIV subtype A strain Petaluma-infected cats, although virus infection was detected by PBMC culture and seroconversion (data not shown).

FIG. 8.

Identification of monocytotropic FIV strains by in vivo passage. Five groups of SPF cats (four per group) were each inoculated intravenously with 5.0 ml of whole blood obtained from an SPF cat that had been infected previously with one of five different clinical FIV isolates. Five additional SPF cats were inoculated with FIV subtype A strain Petaluma (provided by N. Pedersen, University of California, Davis). At 10 and again at 16 weeks p.i., the frequency of FIV infection in monocytes from each cat was evaluated by short-term culture and immunocytochemical analysis and the mean percentage of positive cells (± the standard error) was plotted. Two isolates (2531 and 2542) were identified as the most monocytotropic in vivo.

Four 12-week-old SPF cats were inoculated intravenously with blood from one FIV subtype B strain 2542-infected cat with a high level of monocyte infection (Fig. 8). FIV-bearing monocytes first were detected at 2 weeks p.i. in all of the cats (mean percentage positive, 6.3%) (Fig. 9). The frequency of monocyte infection peaked at 17.2% at 3 weeks p.i., declined gradually to 7.2% at 6 weeks p.i., and varied considerably thereafter but never reached 0 over up to 6 months of observation. These data suggested that the kinetics of monocyte infection in vivo paralleled that in monocyte cultures in vitro, i.e., an acute-phase maximum followed by a gradual decline but not elimination.

FIG. 9.

Kinetics of FIV subtype B strain 2542 monocyte infection after experimental infection. Four 8-week-old SPF cats were inoculated with whole blood from a cat infected with monocytotropic FIV subtype B strain 2542, four age-matched control cats were inoculated with blood from an uninfected control animal, and the level of monocyte infection was monitored over a 7-month period. The percentage of infected monocytes was determined as described in Materials and Methods, and the value for each cat was plotted versus time p.i. (cats: 2883 [●], 2884 [●], 2887 [○], and 2889 [▵]). In cats inoculated with blood from the FIV subtype B strain 2542-infected cat, peak monocyte infection occurred between 14 and 50 days p.i. The level of monocyte infection declined thereafter, remained intermittently detectable through 4 months, and then appeared to recrudesce at 7 months p.i. The cat with the highest level of monocyte infection (▵) was euthanized at 9 months p.i. due to progressive weight loss.

PBMC-associated viral titers in cats infected with FIV subtype B strain 2542 increased rapidly after inoculation, remained high for 3 months, and declined slowly thereafter (Fig. 10a). CD4⁺ T cells progressively declined in all FIV subtype B strain 2542-infected cats (Fig. 10b), and CD8⁺ T cells failed to manifest the normal age-associated increase (Fig. 10c). All FIV subtype B strain 2542-infected cats developed persistent lymphadenopathy, chronic diarrhea, and progressive weight loss (or failure to gain weight) and were euthanized at 9 to 10 months p.i. By contrast, in cats inoculated with FIV subtype A strain Petaluma, viral titers were substantially lower and neither a CD4 decline nor clinical disease occurred. Thus, infection with monocytotropic FIV subtype B strain 2542 induced a clinical immunodeficiency syndrome at <1 year p.i.

FIG. 10.

Viral replication and CD4 and CD8 T-cell kinetics in cats infected with FIV subtype B strain 2542. SPF cats (four per group) were inoculated with 5 ml of blood from a cat infected with FIV subtype B strain 2542 or sham inoculated with 5 ml of blood from an uninfected cat. The PBMC-associated virus titer was determined by coculture every 2 weeks. The mean viral titer (number of tissue culture-infective doses [TCID] per 10⁶ PBMC ± the standard error) was plotted versus time p.i. (a). Also determined were the numbers of CD4⁺ (b) and CD8⁺ (c) cells. Symbols: ●, FIV-infected cats; ○, age-matched controls.

DISCUSSION

High-level persistent monocyte infection by an immunodeficiency-inducing lentivirus appears to be a novel finding. While macrophages are a major reservoir for HIV infection (14, 19, 21, 35), the number of HIV-infected monocytes appears to be quite low in humans, even in patients with high viral titers in PBMC (2, 30, 31). By contrast, Quiros et al. (40) have reported high levels of PCR positivity in monocytes from HIV-infected patients. Our results obtained with a group of predominantly subtype B FIV clinical isolates suggest that, in addition to lymphocytes (6, 15, 45a), blood and bone marrow monocytes are a second major FIV reservoir in vivo.

Perhaps the methodology employed in this study (immunocytochemical analysis combined with monocyte culture in a low level of IL-4 [20 ng/ml]) favored detection of virus antigen expression in monocytes versus previous studies employing either no culture or shorter culture times and/or different cytokine supplementation (15, 30, 31, 40). For example, our experience indicated that feline monocytes cultured in cytokines other than IL-4 expressed lower levels of FIV antigens than did monocytes cultured in IL-4 (data not shown). In addition, the FIV isolates studied were all from cats with clinical symptoms of immunodeficiency and were selected for high replicative capacity in PBMC culture (9, 10) (Fig. 10). The viral strain may also be a major factor. English et al. (15) detected FIV provirus in uncultured monocyte-enriched PBMC in only 1 of 10 asymptomatic cats infected with FIV strain NCSU1 (clade A). Thus, monocytotropism of FIV isolates may vary with the viral strain, its replicative capacity, and/or its virulence in vivo.

We cannot exclude the possibility that the monocyte cultures established contained cells with dendritic-cell (DC) differentiation. The studies reported here were conducted before current knowledge of cytokine-driven DC differentiation in other species was obtained and before we had a suitable reagent to distinguish putative DCs from monocytes (CD1a MAb from Peter Moore, University of California, Davis). Nevertheless, by using the current paradigm of cytokine-driven in vitro monocyte selection-differentiation to DCs (33a, 40b), we would not expect efficient selection for DC differentiation in the studies reported here due to (i) an insufficient concentration of human rIL-4 (20 ng/ml) versus the 100+ ng/ml we now know is needed for efficient in vitro feline DC culture (recent unpublished data), (ii) an insufficient culture period versus the ≥6 days in culture required for substantial DC differentiation (consistent with protocols developed with human and mouse DC culture), and (iii) the absence of GM-CSF in the culture system. However, use of human rIL-4 at 20 ng/ml favored the in vitro survival of feline monocytes. Thus, while we cannot prove or disprove that infected DCs were present in the monocytes cultured, their relative numbers and contribution to virus antigen-bearing cells is expected to be very small.

Most HIV transmission is mediated by macrophage-versus-lymphocyte-tropic viral strains (1, 5, 23, 26, 29, 42, 43). Infection of macrophages and monocytes by HIV appears to employ the chemokine receptor CCR5 (1, 5, 7, 16, 27, 44). Monocyte infection has also been linked to microglial cell tropism (12, 25, 27, 37). FIV subtype B strain 2542 is readily transmitted by several routes, including prenatal and postnatal mother-to-offspring transmission; exposure of vaginal, rectal, or oral mucous membranes; and intravenous inoculation, in which rapid serial passage of acute-phase plasma can rapidly induce clinical immunodeficiency (9, 10, 34). We do not have evidence that the monocyte/macrophage tropism of subtype B isolate 2542 is linked to a propensity to cross the blood-brain/neuroendothelial, placental, or mammary barrier, although these pathways of virus spread have been documented for this virus isolate (12, 34, 40a). While culture-adapted clade A FIV isolates have been shown to employ feline CXCR4 for cell entry (38a, 46), the identity of a probable CC chemokine receptor(s) used by primary FIV isolates to initiate mucosal and macrophage infection has yet to be revealed.

Circulating monocytes from our FIV-infected cats did not express detectable virus antigens immediately ex vivo, although viral expression was upregulated within hours of adherence and culture in vitro (Fig. 3 and 6). These findings suggest that either a state of very low-level virus expression exists in vivo, as has been reported for monocytes infected with ovine visna virus and CAEV (20, 25, 33), or that viral transcription is silent and upregulated after appropriate stimuli and/or removal from host immune factors, as has been shown for myelomonocytic cells latently infected with HIV in vitro (17, 18, 24). Studies with PMA-stimulated monocytes infected with either visna virus or CAEV demonstrated that maturation of monocytes into macrophages provided the necessary stimulus for viral replication and that only mature macrophages were capable of supporting productive viral infection (20, 33). Our studies suggest that adherence, as well as other stimuli, such as those provided by PMA (but not cytokines or LPS), may be equally effective in triggering virus expression in in vivo-infected monocytes (Fig. 7). Perhaps virus expression by monocytes in vivo is triggered by adherence to a substrate (e.g., extracellular matrix), as well as by maturation into tissue macrophages.

Infected mononuclear phagocytes have been implicated as a major source of persistent infection in HIV infections resistant to highly effective combination antiviral therapies (38). We report here that some strains of FIV are monocytotropic in vivo to a degree not documented previously and raise the possibility that the level of monocyte tropism correlates with virus virulence. The high level of FIV monocyte infection in vivo may prove useful for studies of lentivirus transmission, transcriptional control, and antiviral therapeutics.