Experimental Mucosal Infection with Molecularly Cloned Feline Immunodeficiency Viruses

Abstract

Four of six specific pathogen-free cats were infected after intravaginal exposure to molecularly cloned lymphotropic but non-Crandell feline kidney (CRFK)-tropic feline immunodeficiency virus strain TM2 and its AP-1 deletion mutant. The sequences of the env V3-to-V5 region which defines the CRFK tropism were unchanged in the infected cats through the infection. These data suggest that the strain was transmitted across the mucosal epithelium without a broadening of cell tropism.

Feline immunodeficiency virus (FIV) is a lentivirus which causes an AIDS-like disease in cats (24). FIV infection in cats is an important animal model for vaccine development and antiviral therapy for human immunodeficiency virus (HIV) infection. Although biting is the principal mode of transmission, FIV can be transmitted experimentally in both a cell-associated and cell-free manner through the vaginal or rectal mucosa (4, 8) and orally via milk (27). Moreover, the virus and infected cells are detected in genital secretions of FIV-infected cats such as semen (13) and vaginal fluids (22), as observed in HIV infection in humans. Experimental transmucosal infection in cats with FIV is a valuable model for understanding the mechanism of vaginal infection of HIV.

Seven transmembrane segment receptors including CCR5 and CXCR4 for chemokines have been shown to be essential, in addition to CD4, for HIV type 1 infection. In contrast to HIV, FIV does not use CD4 for entry and displays a broader cellular tropism including infection of CD8⁺ T lymphocytes, B lymphocytes, macrophages, and astrocytes (3, 6, 7, 24). A few laboratory FIV strains can adapt to infect a feline epithelial cell line (Crandell feline kidney [CRFK] cells), retaining the ability to infect lymphoid cells (17, 29). They have been shown to use CXCR4 alone to infect the CRFK cells (25, 30, 31); however, the receptor(s) used to infect lymphoid cells remains controversial (9, 10, 12, 17, 26). Similar to HIV type 1, FIV has considerable sequence variation in the env gene, and the third to fifth variable regions (V3-V5) of the env gene contain an immunodominant neutralization domain and a determinant of CRFK tropism (15, 17, 29). Chimeric and sequencing analyses of the env gene of the CRFK-tropic viruses suggest that the principal determinant of the CRFK tropism resides in the V3 loop of the surface unit Env protein (29).

In the present study, we examined the amino acid changes in the V3-V5 region of non-CRFK-tropic FIVs through vaginal infection by molecularly cloned viruses. Our results showed that non-CRFK-tropic FIVs were transmitted across the mucosal epithelium without an amino acid change in the region.

CRFK cells (5) were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. MYA-1 cells (19) were grown in RPMI 1640 growth medium supplemented with 10% fetal calf serum, 50 μM 2-mercaptoethanol, 2 μg of polybrene/ml, antibiotics, and 100 U of recombinant human interleukin-2 per ml. These cells were cultured at 37°C in a humidified atmosphere of 5% CO₂ in air.

To prepare virus stocks, pTM219 (18) and pSTM2D2 (20), which are infectious clones of FIV TM2 and its deletion mutant lacking a 31-bp fragment containing one AP-1 binding site and an adjacent AP-4 binding site, respectively, were used. The strain TM2 belongs to clade B (23) and infects and proliferates well in feline lymphocytes but not CRFK cells (15). Ten micrograms of each plasmid was transfected into CRFK cells by the calcium phosphate coprecipitation method. Two days later, the supernatants of transfected cells were collected and filtrated through a 0.45-μm-pore-size Millipore filter. These stock viruses were designated TM2 and ΔAP-1, respectively, and divided into aliquots and stored at −80°C. The titers of these stock viruses were 2.0 × 10³ 50% tissue culture infective dose (TCID₅₀)/ml when titrated on MYA-1 cells as described previously (14).

Nine female specific-pathogen-free cats aged 5 months were purchased from Harlan Sprague Dawley Inc. (Madison, Wis.) and divided into three groups in isolation units. The cats were immobilized by an intramuscular injection with 20 mg of ketamine HCl (Sankyo Inc., Tokyo, Japan)/kg of body weight at the time of inoculation and when peripheral blood samples were taken.

For intravaginal infection, 2.0 × 10⁶ MYA-1 cells were infected with 2.0 × 10³ TCID₅₀ of stock virus and incubated for 12 days. About 30% of the MYA-1 cells were positive when examined by indirect immunofluorescence assay (IFA). We adjusted the antigen-positive cell number to 10⁶ for inoculation and suspended the cells in 20 μl of phosphate-buffered saline. Then, the infected cells were atraumatically administered into the vaginal vault of each immobilized cat using a blunt-ended sterile soft plastic pipette.

The establishment of infection was confirmed by virus isolation (VI) and PCR tests. For the VI test, peripheral blood mononuclear cells (PBMCs) of each cat were separated from heparinized whole blood by Ficoll-Paque (Pharmacia, Uppsala, Sweden) gradient centrifugation. The separated PBMCs were stimulated with 10 μg of concanavalin A/ml for 3 days and cultured in the RPMI 1640 growth medium for 1 week and then cocultured with MYA-1 cells for at least 4 weeks, with monitoring of the FIV antigen-positive cells by IFA. To detect proviral DNA from the PBMCs, total cellular DNA was extracted with a QIAamp Tissue Kit (QIAGEN, Hilden, Germany) according to the manufacturer's instructions. FIV proviral DNA was amplified by seminested PCR with the outer primer pair 5′-GAACTCCTGCAGACCTTGTG-3′ (Pr-1) and 5′-GGATCAAGGGCTACATACTG-3′ (KA-18) and the inner primer pair Pr-1 and 5′-GTCTCTGGTATATCACCTGG-3′ (KA-14), and then the specificity of PCR products was confirmed by Southern blotting as described previously (11, 16).

Serum samples obtained from these cats were serially diluted with phosphate-buffered saline starting from a 50-fold dilution in 2-fold steps, and antibodies against TM2-infected MYA-1 cells in the samples were examined by IFA as described previously (14). Titers were expressed as the inverse of the highest dilution that gave a positive result.

The PBMCs of the infected cats were isolated at 4 weeks postinfection (wpi). For sequencing analyses, total cellular DNA was extracted from the PBMCs by using a QIAamp blood kit (QIAGEN) and the envelope gene sequences encoding the V3-V5 region were amplified with the primer pair HV3f and HV5r (23). The amplified fragments were directly subjected to sequencing analysis, as described previously (23).

Cats 304, 305, and 306 were inoculated with MYA-1 cells infected with TM2, and Cats 307, 308, and 309 were inoculated with MYA-1 cells infected with ΔAP-1, whereas Cats 301, 302, and 303 were inoculated with uninfected MYA-1 cells as controls (Table 1). These cats were kept anesthetized until the inoculum had become semidry and were fitted with plastic Elizabethan collars for 24 h after the inoculation to prevent infection via the oral route. By 4 wpi, two cats per group inoculated with each virus were found positive by VI and PCR tests (Table 1). The other five cats including two inoculated with each virus remained negative throughout the experiment. Serially diluted sera were examined for the anti-FIV antibody by IFA (Table 2). From the serum samples of the four cats with positive VI and PCR tests, antibodies were detected from 3 and 5 wpi in cats infected with TM2 and ΔAP-1, respectively. No antibody response was detected in cats that tested negative. From these data, we conclude that the four cats were certainly infected with the non-CRFK tropic FIVs.

TABLE 1.

Detection of proviral DNA and isolation of FIV from PBMCs

Cat no.	Inoculum	Resulta of proviral DNA detection/virus isolation at wpi:
		1	2	3	4	5	6	7	8
304	TM2	−/−	−/−	−/NT	+/+	+/NT	+/+	NT/NT	+/+
305	TM2	+/−	+/−	+/NT	+/+	+/NT	+/+	NT/NT	+/+
306	TM2	−/−	−/−	−/NT	−/−	−/NT	−/−	−/NT	−/−
307	ΔAP-1	−/−	−/−	−/NT	+/+	+/NT	+/+	NT/NT	+/+
308	ΔAP-1	−/−	−/−	−/NT	+/+	+/NT	+/+	NT/NT	+/+
309	ΔAP-1	−/−	−/−	−/NT	−/−	−/NT	−/−	−/NT	−/−

^a+, positive; −, negative; NT, not tested.

TABLE 2.

Time course of appearance of anti-FIV antibody in serum samples

Cat no.	Titera of anti-FIV antibody at wpi:
	1	2	3	4	5	6	7	8
304	<50	<50	50	100	100	200	NT	800
305	<50	<50	100	200	400	400	NT	3200
306	<50	<50	<50	<50	<50	<50	NT	<50
307	<50	<50	<50	<50	50	100	NT	800
308	<50	<50	<50	<50	50	200	NT	800
309	<50	<50	<50	<50	<50	<50	NT	<50

^aExpressed as the inverse of the highest plasma dilution determined by IFA.

To examine the amino acid changes of the viruses through the vaginal infection, we tried to amplify the 642-bp nucleotides covering the env V3-V5 region from FIV-infected cat at the early stage (4 wpi) of infection by PCR. We succeeded in the amplification and sequencing from samples from Cats 304, 305, and 308. We compared the sequences with the sequence of the original infectious clone, pTM219. That from Cat 304 showed one nucleotide substitution which resulted in an amino acid change in the V3 region, while no nucleotide change was observed in the sequence from either Cat 305 or Cat 308 (Fig. 1).

FIG. 1.

Comparison of predicted amino acid sequences of the V3-V5 region from vaginally infected animals with the original sequence. Identical amino acids are indicated by dots. The amino acid at position 407, which defines CRFK tropism, is indicated by an arrow.

It was reported that with certain uncloned FIV strains it is difficult to infect cats via the vaginal route (1). This finding suggests that FIV strains which do infect via the vaginal route have a particular amino acid sequence, possibly in the env V3-V5 region. The molecularly cloned FIV strain TM2 infected cats via the vagina efficiently. The availability of the molecularly cloned FIVs that infect cats via the vagina may provide a clue as to the viral determinant. Verschoor et al. (29) reported that an E-to-K change at Env position 407 was responsible for the broadening of cell tropism to infect CRFK cells. In two of three cats examined, the sequences spanning the V3-V5 region were unchanged. The amino acid at position 407 was also unchanged in the other cat. In addition, when the reisolated viruses were inoculated onto CRFK cells, no significant reverse transcriptase activity in the cell supernatants was observed 16 days after inoculation (data not shown). These results suggest that the non-CRFK-tropic FIVs can be transmitted across the intact mucosal epithelium without a broadening of cell tropism. In HIV infection, viral particles were observed to cross an epithelial cell barrier by transcytosis in vitro (2). In simian immunodeficiency virus infection, lymphocytes and Langerhans dendritic cells in the lamina propria of the vaginal mucosa are suggested to be the first target (28). Recently, Obert and Hoover (21) reported that FIV crosses mucous membranes within hours after exposure and is rapidly transported via dendritic and T cells to systemic lymphoid tissues. The present study seems to support their findings and indicate that the first cells targeted in vaginal infections of FIV are not epithelial cells but dendritic and/or T cells.