INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Feline immunodeficiency virus (FIV) is an exogenous lentivirus of domestic cats which causes the development of a progressive immunodeficiency in its host (37). Infection of young cats with FIV has been a useful model for pediatric AIDS, particularly when organs normally unaccessible during the disease course such as the thymus are studied (16, 22). The thymus has been proposed to be the major site of replication in young hosts during the acute phase of FIV and human immunodeficiency virus (HIV) infection (6, 13, 36). Furthermore, studies have identified thymic dysfunction as a predictor of rapid disease progression and mortality in HIV-infected infants (19, 21). Researchers have used several model systems to characterize thymic dysfunction and to begin addressing mechanisms which may contribute to thymic dysfunction resulting from lentivirus infection (1, 4, 27, 30; reviewed in reference 12). Characteristic thymic lesions caused by lentivirus infections include decreased cellularity of the cortex, loss of demarcation of the cortico-medullary junction, altered proportions of thymocyte subsets, and lymphoid follicular hyperplasia.

Studies of SCID-hu mice with thymic implants have demonstrated a correlation between HIV type 1 replication and pathogenesis (5, 28). Viral replication in the SCID-hu mouse system was shown to be highly correlated to depletion of thymus single-positive CD4⁺ CD8 cells, and it was shown that a minimum level of replication must be achieved before pathogenesis is evident. Attempts to tie virus replication to thymic involution by modulating replication using antiviral therapy during FIV infection of cats have been unsuccessful (10). In fact, thymotropic agents, such as insulin-like growth factor 1, have been shown to ameliorate thymic lesions and cause regeneration of the thymic cortex in FIV-infected cats, with little to no change in rates of viral replication (35). In this study, we have used a genetic approach to modulate viral replication in vivo in order to compare thymic lesions in cats infected with a highly pathogenic FIV molecular clone to lesions in cats infected with a replication-impaired FIV molecular clone.

FIV is classified as a complex retrovirus with three accessory genes, vif, ORF-A/2, and rev, in addition to the structural genes, gag, pro-pol, and env. The accessory genes are important in the regulation and intracellular transport of viral mRNA transcripts and function of the infectivity of the cell-free virion (15, 24a). The ORF-A/2 gene product was demonstrated to transactivate the FIV long terminal repeat (LTR) in vitro (7). Earlier infection studies identified an FIV molecular clone, 34TF10, which contained a premature stop codon in ORF-A/2 and replicated poorly on mitogen stimulated peripheral blood lymphocytes (PBLs) (24, 32). Subsequent experiments demonstrated replacement of the stop codon with a tryptophan codon restored replication of 34TF10 in laboratory T-cell lines and feline lymphocytes (34). However, some researchers have characterized 34TF10 as atypical as a result of isolation from a tissue culture-adapted isolate of FIV and have characterized other domains which contribute to the tropism of 34TF10 (2).

In vivo studies of the role of ORF-A/2 in pathogenesis have used FIV molecular clones isolated from tissue culture adapted strains (29) or molecular clones which cause little thymic injury (15). The first study inoculated cats with the FIV molecular clones 34TF10 and pPPR (29). The authors of this study correlated virus replication in vivo with virus replication in vitro in cultures of feline lymphocytes. Both 34TF10 and pPPR were less capable of suppressing a key prognostic indicator of disease, the peripheral CD4/CD8 ratio. In addition, cats infected with 34TF10 did not consistently seroconvert. The poor pPPR pathological response stands in contrast to results obtained by others (25) and has been partially addressed by studies examining the in vivo dose response to pPPR (14, 30). Thymic injury was not described to a significant degree in this study. The second study used the FIV molecular clone pTM219 and accessory gene deletion mutants (15). ORF-A/2 was determined to be dispensable for viral replication in vivo, and the deletion mutant resulted in a slow antibody response, low viral loads, a less severe reduction of the peripheral blood CD4/CD8 ratio, and only mild histopathological findings. It is worth noting that only one of three animals infected with the wild-type pTM219 clone demonstrated decreased cellularity of the thymic cortex, and no lymphoid follicular hyperplasia within the thymus was reported. This lack of thymic lesions stands in contrast to the FIV-induced changes reported by others (1, 16, 23, 36) and indicates a lack of thymic tropism in this clone or possibly a reduced in vivo pathogenesis for pTM219.

In this study, we describe the in vivo properties of an ORF-A/2-deficient mutant of a highly pathogenic FIV molecular clone, JSY3 (22, 38). Our data confirm previous reports of lower in vivo levels of viral replication for ORF-A/2-deficient FIV molecular clones. In addition, we demonstrate that cell-associated viral load measured by lymphocyte coculture varied in a manner which was tissue dependent but independent of ORF-A/2. Unlike previous studies, we report that an ORF-A/2-deficient FIV molecular clone causes significant lesions within the thymus, characterized by the nodular expansion of lymphoid cells in the absence of a thymic epithelial cell network. These structures, termed lymphoid follicles, were not correlated with viral replication within the thymus. However, the expansion of the single-positive CD4 CD8⁺ subpopulation of thymocytes was correlated with viral replication.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cell lines. CrFK cells were cultured in complete minimum essential medium (cMEM) with Earle's salts and L-glutamine supplemented with 100 mM MEM nonessential amino acids and 10% fetal horse serum (Life Technologies, Inc., Gaithersburg, Md.). CrFK cells were incubated at 37°C in 5% CO₂. A feline CD4⁺ cell line, CD4E, was cultured in complete RPMI 1640 medium (cRPMI) supplemented with 10% fetal calf serum, 2 mM L-glutamine, 10 mM HEPES, 0.075% sodium bicarbonate, 2 mM sodium pyruvate, 2-mecaptoethanol, and 100 U of recombinant human interleukin 2 (rIL-2; provided by the NIH AIDS Research and Reference Reagent Program, Rockville, Md.) per ml. Primary cat lymphocytes were cultured in cRPMI with 50 U of rIL-2 per ml. Concanavalin A (ConA)-stimulated lymphocytes were cultured overnight in cRPMI containing 2.5 µg of ConA (Sigma, St. Louis, Mo.) per ml. Nonadherent cells were then used for infection studies. All lymphocytes and CD4E cells were incubated at 37°C in 7% CO₂.

Construction of ORF-A/2-deficient molecular clone. The infectious FIV molecular clone JSY3 was obtained from Wayne A. F. Tompkins (North Carolina State University, Raleigh, N.C.). The JSY3ORF-A/2 molecular clone was constructed from JSY3 by site-specific mutagenesis. Two doublet nucleotide deletions (nucleotides 6013/6014 and 6028/6029 of the JSY3 genome) were introduced in ORF-A/2 by site-specific mutagenesis using degenerate PCR primers. The sequences of all primers are listed in Table 1. Briefly, primer pairs AYM112/AYM114 and AYM111/113 were used to create two amplicons overlapping only at the degenerate primer sequences. Equal molar concentrations of these amplicons were then added to a third PCR using the primer pair AYM111/AYM112. The final PCR product containing the site-specific deletions was then digested with restriction endonucleases BspEI and Bst1107I. The molecular clone JSY3 was digested with the same restriction endonucleases and ligated to the final PCR amplicons. These mutations cause successive frameshifts beginning in codon seven of ORF-A/2 and cause the early termination of the gene product. The mutations were confirmed by sequencing at the University of Florida DNA sequencing core laboratory (Gainesville, Fla.). Molecular cloning and plasmid preparations were conducted using standard techniques (27).

Virus preparation. Cell-free virus was prepared as described previously (22). Briefly, 4 × 10⁵ CrFK cells were plated in cMEM and incubated until growth reached 70 to 80% confluence. Plasmid DNA containing either the JSY3 or JSY3ORF-A/2 provirus was then transfected using the LipofectAMINE reagent as instructed by the manufacturer (Life Technologies, Inc., Gaithersburg, Md.). One day after transfection, CrFK cells were placed in cRPMI with 100 U of rIL-2 per ml and cocultured with 2 × 10⁶ CD4E cells. After 48 h, CD4E cells were removed from coculture and passaged normally. Supernatant was collected, clarified at 800 × g for 10 min, and snap-frozen. Viral titers were calculated using Reed and Muench calculations from fourfold dilutions done in sextuplet on CD4E cells as previously described (17). DNA was extracted from titer cultures for JSY3ORF-A/2, and the deletions were again confirmed by sequencing.

Animals and inoculation. One-day-old kittens (n = 9) were inoculated by intraperitoneal injection of 50% tissue culture infective doses of 10⁴ of JSY3 or JSY3ORF-A/2 cell-free virus preparation. All inocula were adjusted to 200 µl (total volume) with sterile RPMI 1640. All cats were kept at a specific-pathogen-free facility until sacrifice at week 14 except one JSY3-infected cat, 8C13-18. This animal was euthanized at week 10 for humane reasons due to a secondary infection. Data from 8C13-18 were included in analysis of hematological cell counts until week 8. No other data collected from 8C13-18 were used in this study. At the completion of the study, DNA was extracted from thymus tissue for JSY3ORF-A/2-infected cats, and the conservation of the deletions was confirmed in all animals by sequencing.

Flow cytometry. Subpopulations of tissue lymphocytes were analyzed by dual-fluorescence flow cytometry as previously described (23). Briefly, lymphocytes were incubated with anti-feline CD4-biotin (antibody CAT30A-bio) and anti-feline CD8-fluorescein isothiocyanate (antibody CAT357) on ice for 30 min. Cells were then washed with phosphate-buffered saline-2% fetal calf serum, and streptavidin-phycoerythrin was added. Cells were incubated for 15 min on ice, washed once, and resuspended in isotonic 0.25% paraformaldehyde. Data were collected with a FACScan flow cytometer and analyzed using the LYSIS-II program (Becton Dickinson, San Jose, Calif.). Lymphocytes were gated from tissue preparations on the basis of light scatter profiles as previously described (23).

Tissue preparation and virus isolation. Intact tissues from infected animals were collected aseptically after euthanasia. PBLs were enriched by Percoll (Pharmacia, Piscataway, N.J.) discontinuous gradient centrifugation as previously described (33). Thymus and lymph node tissues were weighed, and full-thickness, transversely cut samples were cut into approximately 1-mm cubes. Cubes were then minced and lymphocytes were dissociated from the stroma by gentle pressing with a plunger from a 6-ml syringe. Cell counts were done by dye exclusion. Virus was isolated by cocultivating 10⁵ primary lymphocytes with an equivalent number of CD4E cells for 10 days and measuring Mg²⁺-dependent reverse transcriptase (RT) activity as previously described (18).

Immunohistochemistry. Paraffin-embedded tissues were sectioned at 5 µm, dewaxed in xylene, and rehydrated by passage through a gradient of ethanol solutions. Endogenous peroxidase activity was quenched in a hydrogen peroxide solution, and staining was enhanced by digestion in a 0.1% trypsin solution and microwave pretreatment. Sections were incubated overnight in a humidified chamber at 4°C with either anti-FIV p24 Gag antibody PAK3-2C1 (Custom Monoclonals, West Sacramento, Calif.) or anticytokeratin antibody AE1/ZE3 (Zymed, South San Francisco, Calif.). Staining was performed by the ABC (avidin-biotinylated enzyme complex) technique (Vecta-Stain Elite ABC staining kit; Vector Laboratories Burlingame, Calif.) according to the recommended protocol and visualized with diaminobenzidine chromogen. Sections were counterstained using Harris's hematoxylin. Additional sections were stained with hematoxylin and eosin (HE) for morphologic analysis.

Quantification of proviral load. Genomic DNA was extracted from 25 mg of frozen tissue by using a QIAamp DNA Mini kit (Qiagen, Inc., Valencia, Calif.) as instructed by the manufacturer. DNA concentration and purity were determined by A₂₆₀/A₂₈₀. The DNA was resuspended in water containing 20 µg of glycogen (Roche Molecular Biochemicals, Mannheim, Germany) per ml. Proviral load in thymus tissue was then determined by competitive fragment quantitative PCR and verified by a real-time quantitative PCR assay. Competitive fragments and primer pairs for the NCSU-1 FIV isolate gag and feline glyceraldehyde-3-phosphate dehydrogenase gene (G3PDH) were kindly provided by M. B. Tompkins (North Carolina State University, Raleigh, N.C.) (9, 20, 26). Primer sequences are provided in Table 1. The competitive PCR was done as described previously (26), with some modifications. In brief, 100 ng of genomic DNA was used in a reaction series containing known amounts of competitive fragment DNA. Thermocycler parameters to detect FIV gag were set for 94°C for 1 min, then 40 cycles of 94°C for 30 s, 59°C for 90 s, and 72°C for 2 min, and a final elongation step of 72°C for 10 min. Thermocycler parameters for feline G3PDH were as previously described (26). PCR products were separated on 2% agarose in 1× Tris-acetate-EDTA buffer and stained with ethidium bromide. Fluorescence intensity for each band was measured using an AlphaImager 2200 documentation and analysis system (Alpha Innotech, San Leandro, Calif.), and copies of target DNA were calculated as previously described (26). Proviral copy number was normalized by levels of G3PDH.

Statistical analysis. Confidence intervals for hematological data from historical, age-matched uninfected control cats were constructed using a two-tailed test with P set at 0.025. Peripheral blood CD4/CD8 ratios and lymphocyte subsets were compared using the unpaired t test of group means. All other comparisons were calculated by the Mann-Whitney U test. The geometric mean was calculated as a measure of central tendency in the log-transformed data for thymic proviral load. The Spearman rank order correlation coefficient (r_s) was performed to determine the monotonic association between percent nonkeratinized area during cytokeratin staining and percent B cells as determined by flow cytometry.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Viral kinetics in ConA-stimulated PBLs and cocultures. We infected mitogen-stimulated PBLs to establish that the mutations introduced in JSY3ORF-A/2 produced an ORF-A/2-deficient phenotype. ORF-A/2-deficient molecular clones of FIV have been previously reported to replicate poorly in cultures of mitogen-stimulated lymphocytes (24, 32). Infection studies of JSY3 and JSY3ORF-A/2 demonstrated a significant reduction of replication in JSY3ORF-A/2-infected lymphocyte cultures (Table 2, Mann-Whitney U test, P < 0.02).

An ORF-A/2-deficient virus yields reduced cell-associated viral load in primary tissues. To estimate the relative cell-associated viral load within the thymus, peripheral blood, and lymph nodes, tissue lymphocytes were cocultured with CD4E cells for 10 days, and the supernatant were assayed for RT activity (9). Thymus, PBL, and lymph node tissue cocultures from JSY3-infected animals resulted in RT activities of (2,343 ± 278) × 10² (mean ± standard error), (923 ± 103) × 10² and (628 ± 112) × 10² cpm per ml, respectively (Fig. 1). Tissue coculture RT assays from JSY3ORF-A/2-infected animals resulted in RT activities of (818 ± 181) × 10², (275 ± 69) × 10², and (121 ± 13) × 10² cpm per ml. RT activity in tissue cocultures from JSY3-infected animals was threefold greater in thymus and PBL cocultures and fivefold greater in lymph node cocultures (Mann-Whitney U test, P < 0.001 for all tissues). A JSY3ORF-A/2-infected cat, C5-1, exhibited amounts of RT activity in coculture supernatant similar to levels obtained in JSY3-infected cocultures. Thymus, PBL, and lymph node tissue cocultures done in quadruplicate from animal C5-1 resulted in RT activities of (2,245 ± 92) × 10², (837 ± 75) × 10², and 125 ± 3 × 10² cpm per ml, respectively. Cell-associated viral load varied in a tissue-dependent manner and was three- to fivefold greater in JSY3 than JSY3ORF-A/2-infected cats.

View larger version (19K): [in this window] [in a new window]   FIG. 1.   Coculture supernatant RT activity. Primary lymph node lymphocytes (LN), PBLs, and thymocytes were aseptically collected and prepared as described in Materials and Methods. Equivalent numbers of primary tissue cells and a laboratory lymphoid cell line, CD4E, were cocultured for 10 days. Clarified supernatants were then assayed for RT activity. Each bar represents the treatment group mean of quadruplicate cultures done for each tissue coculture. Error bars represent the standard error of the mean. A pattern of stepwise levels of replication, highest in cocultured thymocytes, lower in PBLs, and lowest in lymph node tissue, was conserved in both JSY3- and JSY3ORF-A/2-infected cats. JSY3 replicated at higher levels in all tissues examined (Table 2, Mann-Whitney U test, P < 0.01 for all weeks). The greatest difference (fivefold) occurred in lymph node tissue cocultures.

ORF-A/2 deficiency delays the reduction of the peripheral blood CD4/CD8 ratio. Biweekly hematological profiles were obtained to assess the ability of an ORF-A/2-deficient molecular clone to reduce the peripheral blood CD4/CD8 ratio of infected cats. JSY3 maintains the in vivo biological properties of its parental FIV isolate, NCSU-1, including the inversion of the peripheral blood CD4/CD8 ratio (22, 38). PBL profiles were analyzed by flow cytometry in both groups of infected cats. CD4/CD8 ratios and PBL counts were plotted against 95% confidence intervals constructed from data obtained from historical, age-matched uninfected control animals (Fig. 2). Data points falling outside of the 95% confidence area of Fig. 2 are significantly different from historical, age-matched uninfected control cats (P < 0.05). JSY3-infected animals demonstrated a pronounced reduction of the CD4/CD8 ratio by week 4 postinfection (p.i.) (Fig. 2A). JSY3ORF-A/2-infected animals did not demonstrate a reduction of the CD4/CD8 ratio, compared to a 95% confidence interval from control animals, until week 10 p.i. By week 4 p.i., CD4/CD8 ratios of JSY3-infected animals were significantly lower than those of JSY3ORF-A/2-infected animals (unpaired t test of group means, P = 0.03). Lower ratios were maintained in JSY3-infected animals during weeks 8 and 14 p.i. (P = 0.02 and P = 0.05, respectively). Using flow cytometry and total cell counts, differences in the CD4/CD8 ratio were attributed to a strong early trend toward a decrease in CD4⁺ lymphocytes in JSY3 versus JSY3ORF-A/2-infected animals (Fig. 2B, weeks 2 and 4 p.i., both P = 0.06). Both JSY3 and JSY3ORF-A/2-infected cats failed to exhibit a peak CD4⁺ PBL count that has been reported in uninfected, age-matched control cats at approximately week 4 of age (3). In addition, a later increase in CD8⁺ lymphocytes in JSY3-infected cats was not observed in JSY3ORF-A/2-infected cats (Fig. 2c, week 14 p.i., P = 0.04). Infection with an ORF-A/2-deficient molecular clone of FIV, compared to the wild-type clone, led to a slower reduction of the peripheral CD4/CD8 ratio primarily due to a less pronounced loss of CD4⁺ lymphocytes early in the acute phase of infection.

View larger version (32K): [in this window] [in a new window]   FIG. 2.   Biweekly hematological data. (A) Mean CD4/CD8 ratios of JSY3 ()- and JSY3ORF-A/2 ()-infected cats plotted against a 95% confidence interval constructed from uninfected, age-matched controls. Data points falling outside of the 95% confidence area (grey) are significantly different from historical, age-matched uninfected control cats (P < 0.05). Ratios were calculated from percent CD4+ and CD8+ cells as determined by flow cytometry. Asterisks denote weeks in which JSY3-infected cats maintained significantly lower CD4/CD8 ratios than JSY3ORF-A/2-infected cats (P  0.05). (B) Absolute counts of CD4+ cells plotted against 95% confidence interval from uninfected, age-matched controls. Cell counts were calculated using percent CD4+ cells determined by flow cytometry, complete blood counts, and differential. , strong trend in lower CD4+ cell counts in JSY3-infected cats (P = 0.06). (C) Absolute counts of CD8+ cells plotted against 95% confidence interval from uninfected, age-matched controls. Cell counts were calculated as in (B) for percent CD8+ cells. , significant increase of CD8+ cells in JSY3-infected cats (P = 0.04). Data are the means of treatment groups, and error bars represent the standard error of the mean.

JSY3ORF-A/2 infection resulted in reduced thymic proviral load. Analysis of viral DNA by quantitative PCR was conducted to compare thymic proviral load in JYS3 and JSY3ORF-A/2-infected cats. DNA was extracted from infected thymuses, and FIV gag target was quantified using sequence-specific primers. The level of proviral load was highest in JSY3-infected cats, except for animal C5-1 (Fig. 3). Animal C5-1 maintained a proviral load similar to that of JSY3-infected cats. Proviral load was significantly reduced in JSY3ORF-A/2-infected cats (Mann-Whitney U test, P = 0.05). Proviral load was approximately a log and a half lower in cats infected with JSY3ORF-A/2, using the geometric mean for the measure of central tendency of log-transformed data (60,400 copies in JSY3 thymuses versus 1,100 copies in JSY3ORF-A/2-infected thymuses). Infection with an ORF-A/2-deficient FIV molecular clone resulted in reduced thymic proviral load.

View larger version (13K): [in this window] [in a new window]   FIG. 3.   Thymic proviral load. weeks p.i. Cells isolated from cats 14 weeks p.i. were lysed, nucleic acids were purified, and quantitative PCR was performed with primer pairs specific for FIV gag and feline G3PDH. Proviral copy number was standardized to G3PDH and expressed as copies per 100 ng of genomic DNA. The geometric mean was determined for proviral copy number in JSY3- and JSY3ORF-A/2-infected cats and marked as a transecting bar on the graph. Proviral load was significantly reduced in JSY3ORF-A/2-infected cats (Mann-Whitney U test, P = 0.05). JSY3 proviral load was approximately a log and a half greater than JSY3ORF-A/2 proviral load. The highest proviral load in the JSY3ORF-A/2-infected cats was animal C5-1.

Infection with JSY3ORF-A/2 results in a reduction of viral p24 Gag-expressing thymocytes. To compare productively infected thymocytes in JSY3 and JSY3ORF-A/2-infected cats, we examined paraffin-embedded thymus sections by immunohistochemistry for expression of viral p24 Gag protein. Virus was detected in the thymus of all cats by immunohistochemistry for FIV p24 Gag (Fig. 4A). Consistent with previous observations of mRNA expression, p24-positive cells were distributed evenly throughout the thymus but excluded from lymphoid follicles (20, 23). Sections were scored on a semiquantitative scale for number of p24-positive cells per field (Fig. 4B). Two JSY3-infected cats showed high numbers of p24-positive cells in the thymus (5+, over 20 cells per field), while one showed moderate levels (2+, 6 to 10 cells per field). In contrast, JSY3ORF-A/2-infected cats mostly exhibited low numbers of p24-positive cells (1+, 1 to 5 cells per field). One JSY3ORF-A/2-infected animal, C5-1, exhibited high numbers of p24-positive cells in the thymus (5+, over 20 cells per field). Analysis of viral p24 Gag by immunohistochemistry demonstrated a greater number of p24-positive cells in thymus tissue from JSY3-infected cats.

Thymic lymphoid follicular hyperplasia is not reduced during JSY3ORF-A/2 infection. HE-stained sections of thymuses from all cats showed a decrease in cellularity of the thymic cortex, loss of demarcation of the cortico-medullary junction, and lymphoid follicular hyperplasia (Fig. 4A). Disruption of the thymic epithelium was seen in all cats by cytokeratin immunohistochemistry (Fig. 4A). Thymic areas which did not stain for cytokeratin have been correlated previously with lymphoid follicle formation (20, 23). The percent nonkeratinized area for each cat was quantified using the NIH Image software to determine the range of lesions associated with JSY3 and JSY3ORF-A/2 infection (Fig. 4B). The percent nonkeratinized area correlated with percent B cells in thymus tissue as determined by flow cytometry for immunoglobulin G (IgG)-bearing cells which were pan-T negative (pan-T IgG⁺ cells) (r_s = 0.95). All experimentally infected cats had greater than 5% nonkeratinized area per section, with no apparent differences between JSY3 or JSY3ORF-A/2 (Fig. 4B). Uninfected, age-matched control cats had less than 3% nonkeratinized area. A similar comparison using percent B cells (pan-T IgG⁺) in thymus tissue as determined by flow cytometry also fails to demonstrates a difference (data not shown). The thymic architecture was compromised to a similar degree in both JSY3 and JSY3ORF-A/2-infected cats, as demonstrated by cytokeratin immunohistochemistry and flow cytometry.

View larger version (108K): [in this window] [in a new window]   FIG. 4.   Histopathology and immunohistochemistry of thymic tissue. (A) Thymus sections from JSY3- and JSY3ORF-A/2-infected animals were compared by HE staining (magnification, ×60), immunohistochemistry for cytokeratin (magnification, ×120), and immunohistochemistry for FIV p24 Gag (magnification, ×260). The thymuses of both groups contained distinctive areas of lymphoid follicular hyperplasia (F), as determined by HE staining. A thymic lobule with normal morphology is located immediately right of the follicle in the JSY3-infected cat (N). Large areas of nonkeratin staining have previously been described and correlate with areas of follicular hyperplasia (F). FIV p24 Gag-positive cells were apparent in all cats and distributed evenly throughout the thymus but excluded from lymphoid follicles. (B) The percent nonkeratinized area for each animal was quantified as described in Materials and Methods. The mean percent nonkeratinized area was plotted for each cat. Greater variability was seen in JSY3 than JSY3ORF-A/2-infected cats, but otherwise no differences were evident. The percent nonkeratinized area was correlated to percent B cells in thymus tissue as determined by flow cytometry (rs = 0.95), and uninfected, age-matched control cats had less than 3% nonkeratinized area. FIV p24 Gag-positive cells per field were scored on a semiquantitative scale as described in Materials and Methods (magnification of ×40 used for scoring). The most frequent score per animal was used as data points. JSY3ORF-A/2 cats displayed a lower level of viral p24 Gag-positive cells compared to JSY3-infected cats. The highest semiquantitative score in the JSY3ORF-A/2-infected cats was animal C5-1.

Expansion of single-positive CD4 CD8⁺ thymocytes is reduced during JSY3ORF-A/2 infection. Thymocyte subsets were analyzed by flow cytometry for differences between JSY3 and JSY3ORF-A/2-infected cats. FIV infection causes characteristic alterations of the thymocyte subsets (23, 36). Different expression patterns of CD4 and CD8 on thymocytes delineate four well-characterized subpopulations. In FIV-infected cats, compared to uninfected control cats, the proportion of CD4⁺ CD8⁺ thymocytes was reduced in conjunction with increases in the proportions of the CD4 CD8 and CD4 CD8⁺ thymocyte subpopulations. Thymocyte profiles from representative JSY3 and JSY3ORF-A/2-infected cats are shown as dot plots in Fig. 5A with the profile of an uninfected, age-matched control for reference. Thymocytes were gated on the basis of light scatter profiles as previously described (23). Comparison of the thymocyte subpopulations in JSY3 and JSY3ORF-A/2-infected cats reveals a less pronounced reduction in JSY3ORF-A/2-infected cats of the proportion of CD4⁺CD8⁺ thymocytes (Fig. 5B). The proportion of CD4⁺ CD8⁺ thymocytes in JSY3-infected cats was 63%, compared to 68% for JSY3ORF-A/2-infected cats. In addition, a comparison of the proportions of single-positive CD4 CD8⁺ thymocytes demonstrated a reduction in the characteristic expansion of this subpopulation in JSY3ORF-A/2-infected cats compared to JSY3-infected cats (Fig. 5B). The proportion of CD4 CD8⁺ thymocytes in JSY3ORF-A/2-infected cats was 13%, compared to 22% for JSY3-infected cats (unpaired t test, P = 0.025). Therefore, infection with an ORF-A/2-deficient virus results in a reduction of the characteristic expansion of the single-positive CD4 CD8⁺ thymocyte subpopulation during FIV infection.

View larger version (65K): [in this window] [in a new window]   FIG. 5.   Alterations of thymocyte subpopulations. Preparations of total thymocytes were stained simultaneously with anti-feline antibodies CAT30A-bio and CAT357. (A) Dot plots generated by flow cytometry show an increasing population of CD4 CD8+ cells in JSY3ORF-A/2 and then JSY3-infected animals. (B) The mean thymocyte subpopulation percentages were plotted on a bar representing total thymocytes (JSY3, n = 3; JSY3ORF-A/2, n = 5). Thymocyte subpopulations are identified on the far right as DN, DP, SP CD4, and SP CD8 for CD4 CD8, CD4+ CD8+, CD4+ CD8, and CD4 CD8+ subpopulations, respectively. Uninfected control, JSY3ORF-A/2-infected and JSY3-infected cats demonstrate a stepwise expansion of the CD4 CD8+ thymocyte subpopulation. *, significantly different from JSY3-infected cats (P = 0.025).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this study, we describe the histopathological, phenotypic, and virologic examination of thymic tissue from cats infected with a highly pathogenic molecular clone of FIV, JSY3, and an ORF-A/2-deficient mutant of JSY3. Levels of in vivo viral replication, as measured by lymphocyte coculture, viral p24 Gag immunohistochemistry of thymic tissue, and proviral load, were higher for JSY3-infected cats than for JSY3ORF-A/2-infected cats (Table 2; Fig. 1, 3, and 4B).

The ORF-A/2-deficient mutant of JSY3 used in this study, JSY3ORF-A/2, was unable to produce high levels of RT activity in a laboratory T-cell line and feline PBLs as has been previously demonstrated for other FIV molecular clones (Table 2) (24, 34). JSY3ORF-A/2 was able to infect cats, cause a delayed reduction of the peripheral blood CD4/CD8 ratio, and cause significant thymic lesions. Viral coculture assays demonstrated that JSY3ORF-A/2 was replicating in lymph node, peripheral blood, and thymus tissues (Fig. 1). This result showed that the presence of an intact ORF-A/2 gene and primary tissue type cocultured affected viral replication, as measured by the level of RT activity. Although the basal transcriptional level of the FIV LTR has not been measured in primary tissue directly, some researchers have speculated that the basal transcriptional level is quite promiscuous due to the presence of binding sites for multiple cellular transcription factors in the viral LTR (31). Our data suggest that viral replication is increased three- to fivefold in the presence of ORF-A/2. Importantly, this result also suggests that tissue and/or cellular micro environment affects viral replication independent of ORF-A/2.

In contrast to previous studies using ORF-A/2-deficient molecular clones (15, 29), thymuses from animals infected either with JSY3 or JSY3ORF-A/2 demonstrated lesions characteristic of FIV infection. One of the most striking lesions in the thymus of FIV-infected cats is the formation of lymphoid follicles and germinal centers, which represent atypical areas of B-lymphocyte proliferation (1, 4, 8). These follicles are located within the perivascular spaces, areas outside the epithelial network that can be demonstrated histologically by a lack of cytokeratin staining (Fig. 4A) (10, 20, 23). In the present study, JSY3 and JSY3ORF-A/2 produced this lesion with similar incidence and severity despite a marked difference in the level of productive infection. Similarly, Hayes et al. (10) reported that follicles within the perivascular spaces persist even after virus replication is reduced by antiviral therapy. It is likely, therefore, that other mechanisms, such as altered cytokine secretion or unique cell trafficking, may contribute to formation and persistence of thymic lymphoid follicles. A mantle of CD4 T lymphocytes surrounds the B lymphocytes within the perivascular spaces and may produce IL-10 in addition to serving as a virus reservoir (20). Because IL-10 can regulate human B-cell growth and differentiation, it is possible that FIV may initiate a cytokine cascade that promotes the formation of lymphoid follicles, rendering a mechanism of thymus injury that is indirectly linked with infection. Such a mechanism would support the theory that cytokine expression within the perivascular spaces during HIV infection may disrupt the thymic cytokine microenvironment and thereby reduce thymopoiesis (11, 12).

Differences were evident in the proportions of thymocyte subpopulations between JSY3 and JSY3ORF-A/2-infected animals (Fig. 5). The percentage of CD4 CD8⁺ thymocytes was significantly greater in JSY3-infected than JSY3ORF-A/2-infected cats. The origins of this single-positive CD4 CD8⁺ population has not been firmly established. It is either a population retained within the thymus after completion of thymopoiesis or mature cells infiltrating from the periphery. It is possible that this expanding CD4 CD8⁺ population is the direct source of the increased proviral load in JSY3-infected cats. However, the literature contains conflicting reports as to the impact of infiltrating lymphocytes on thymic viral load. One group presented data which suggest that increased viral load in mature, peripheral lymphocytes (CD1^lo) predicates an increase in viral load in immature thymocytes (CD1^hi) (36). However, a more recent study suggests that viral load is barely detectable within CD8⁺ lymphocytes isolated from peripheral lymph nodes and not the major source of viral load within the thymus (20). The differences are likely attributable to the virus isolate used and antibodies used for lymphocyte sorting. Most likely, the present study is comparable to that of Liang et al (20) because the two studies used the same virus and antibodies. Therefore, we feel it probable the expansion of the CD4 CD8⁺ subpopulation is incidental to the level of virus replication within the thymus.

Importantly, the JSY3ORF-A/2-infected cat C5-1 produced a virus isolate which was an exception to the replication-impaired ORF-A/2-deficient phenotype. Virus isolated from C5-1 was able to replicate to high levels in cultures of thymocytes and ConA-stimulated PBLs from an unrelated donor cat (Table 2). Sequencing of ORF-A/2 from virus isolated from animal C5-1 confirmed conservation of the deletions introduced during the construction of JSY3ORF-A/2. We feel that a complementary change in the viral genome has allowed an escape from the replication-restricted phenotype of ORF-A/2 deficiency. A more thorough verification and characterization of this isolate is under way. However, phenotypic reversion of JSY3ORF-A/2 in this animal demonstrates the likely unsuitability of ORF-A/2 deletion mutants as potential vaccines.

In conclusion, we have described thymic lesions in young cats infected with either a highly pathogenic FIV molecular clone, JSY3, or an ORF-A/2-deficient FIV molecular clone, JSY3ORF-A/2. Although viral load and replication were reduced in the JSY3ORF-A/2-infected cats, the presence of areas of lymphoid tissue lacking a thymic epithelial cell network within the thymus remained unchanged. This suggests that lymphoid follicle formation does not correlate with thymic viral replication and perhaps is mediated by an independent, indirect mechanism. Alternatively, lymphoid follicles may form within the thymus only after a certain threshold level of viral replication is reached. In contrast, the proportion of CD4 CD8⁺ thymocytes did correlate with viral replication. Although the origin of these cells is unknown, the expanding CD4 CD8⁺ subpopulation may at least be partially responsible for the altered cytokine profiles of FIV-infected thymuses, as has been shown for gamma interferon expression (20, 22). Thus, viral replication may mediate some cellular events and contribute to thymic dysfunction.