Genetically Divergent Strains of Feline Immunodeficiency Virus from the Domestic Cat (Felis catus) and the African Lion (Panthera leo) Share Usage of CD134 and CXCR4 as Entry Receptors

Abstract

The env open reading frames of African lion (Panthera leo) lentivirus (feline immunodeficiency virus [FIV_Ple]) subtypes B and E from geographically distinct regions of Africa suggest two distinct ancestries, with FIV_Ple-E sharing a common ancestor with the domestic cat (Felis catus) lentivirus (FIV_Fca). Here we demonstrate that FIV_Ple-E and FIV_Fca share the use of CD134 (OX40) and CXCR4 as a primary receptor and coreceptor, respectively, and that both lion CD134 and CXCR4 are functional receptors for FIV_Ple-E. The shared usage of CD134 and CXCR4 by FIV_Fca and FIV_Ple-E may have implications for in vivo cell tropism and the pathogenicity of the E subtype among free-ranging lion populations.

Lentiviruses are widespread pathogens of the Felidae, infecting both domestic and nondomestic felids (5, 6, 10, 21, 40, 41). In domestic cats (Felis catus), feline immunodeficiency virus (FIV_Fca) is a significant cause of disease, with infection resulting in a variable immunodeficiency syndrome characterized by recurrent gingivitis-stomatitis, cachexia, wasting, neuropathology, and an increased incidence of tumor development (1, 8, 9, 22, 26, 27, 37, 38, 52). The association between lentiviral infection and disease development in nondomestic felids is less certain, although African lions (Panthera leo) infected with the lion lentivirus FIV_Ple show a dramatic decline in CD4⁺ T lymphocytes (7, 30) and an expansion of an activated CD8⁺ lymphocyte subpopulation expressing low levels of the CD8αβ heterodimer, analogous to observations of FIV (35, 45) and human immunodeficiency virus (HIV) (33) infections.

FIV targets activated CD4⁺ T lymphocytes by utilizing CD134 (OX40) as a primary attachment receptor (34) and CXCR4 as a coreceptor (43, 47, 51). Thus, receptor utilization is likely to be a significant contributory factor to the CD4⁺ T lymphopenia observed in FIV-infected cats. Little is known about the receptor usage of the lentiviruses of nondomestic felids; however, the results of previous studies have indicated that the puma (Puma concolor) lentivirus FIV_Pco-1695 and subtype B lion lentivirus FIV_Ple-458 use alternative attachment receptors from FIV_Fca (36). If neither the lion nor the puma lentiviruses use CD134 or CXCR4 as viral receptors, how then do they induce a depletion of CD4⁺ lymphocytes?

Proviral genome sequence analysis of the subtype B and E lion lentiviruses indicated that the FIV_Ple-B env was most closely related to the virus of the Asiatic Pallas’ cat (Felis manul) (4), whereas the FIV_Ple-E env was more similar to that of the domestic cat virus FIV_Fca (25) than to FIV_Ple-B. These data raise the possibility of either an ancient recombinatorial event between FIV strains in the wild followed by a substantial period of divergence or a more-recent recombinatorial event with an as-yet-uncharacterized but highly divergent FIV species from lions or another lentivirus-infected African species. Here, we ask whether the phylogenetic relationship between FIV_Ple-E and the domestic cat virus FIV_Fca is reflected in the receptor usage of the two viruses.

Peripheral blood mononuclear cells were isolated from 15 seropositive lions from the Moremi game reserve in the Okavango Delta region of Botswana, and one animal (Sangre) yielded replication-competent lentivirus. Sangre was an 8-year-old male lion in good health at the time of sampling. The FIV_Ple-E-Sangre env gene was cloned, and its predicted amino acid sequence displayed 87% identity with the FIV_Ple-1027 isolate of FIV_Ple-E (25) across the SU-TM coding sequence and 54% identity with FIV_Fca-GL8 (46). FIV_Ple-E and FIV_Fca HIV(FIV) luciferase pseudotypes were prepared (34) and their receptor usages compared (Fig. 1A). Both FIV_Ple-E and FIV_Fca-GL8 infected cells expressing feline, but not human, CD134. Neither FIV_Ple-E nor FIV_Fca-GL8 utilized feline cysteine-rich domain 1 (CRD-1) against human CD134 (48, 50) chimera, suggesting that the FIV_Ple-E Env interacts with CD134 in a manner similar to FIV_Fca-GL8 (48).

FIG. 1.

Receptor utilization by FIV_Fca and FIV_Ple-E. HIV-luciferase pseudotypes were prepared bearing the Envs of either FIV_Fca or FIV_Ple-E and used to infect MCC cells stably transduced with retroviral vectors bearing feline (domestic cat) CD134 (fCD134), feline CD134 CRD1/human CD134 chimera (fCRD1/hCD134), human CD134 (hCD134), or vector only (control) (A); NP2 cells transduced with retroviral vectors bearing lion CD134, lion CXCR4, feline CD134, or feline CXCR4 in combination (B); or lion CD134, lion CXCR4, feline CD134, or feline CXCR4 alone or lion CXCR4 in combination with lion or feline CD134 (C). Cells were infected with pseudotypes bearing FIV_Ple-E or FIV_Fca-GL8 (A); FIV_Fca-PPR, GL8, 1419, TM2, or CPG41 (50) (B); or FIV_Ple-E or FIV_Fca-GL8 (C). Histograms are representative of the results of at least two independent experiments and display means ± standard errors (n = 3).

Next, we cloned the lion homologues of CXCR4 and CD134 and expressed both molecules in human NP2 cells. Lion and feline CD134 were coexpressed with either feline or lion CXCR4 by retroviral transduction, and the cells were challenged with FIV_Fca pseudotypes. Each of the four receptor/coreceptor combinations rendered the cells permissive for infection with HIV(FIV_Fca) pseudotypes bearing diverse Envs (Fig. 1B) with a similar efficiency, confirming that lion CD134 and CXCR4 were functional receptors for FIV_Fca. Stable NP2-derived lines were then generated expressing lion CD134, lion CXCR4, feline CD134, and feline CXCR4. The cells expressing lion CXCR4 were then transduced again with vectors carrying feline or lion CD134 and selected by immunomagnetic separation (MACS; Miltenyi Biosciences). The cells were then infected with FIV_Ple-E and FIV_Fca-GL8 pseudotypes, and viral entry was assessed (Fig. 1C). The coexpression of lion CXCR4 in conjunction with feline or lion CD134 (but not CXCR4 expression alone) rendered the cells permissive to infection with both FIV_Ple-E and FIV_Fca-GL8. Thus, FIV_Ple-E infection requires the coexpression of both CD134 and CXCR4, and lion CD134 and CXCR4 are functional primary receptors and coreceptors, respectively. Next, eukaryotic expression vectors bearing the FIV_Ple-E and FIV_Fca-GL8 Envs were transfected directly into AH927 cells stably expressing feline CXCR4 (42) (AH927-FX4P) and feline CD134. Transfection of either the Ple-E (Fig. 2B) or GL8 (Fig. 2D) env into AH927-FX4P-CD134 cells resulted in syncytium formation, while the results for transfected AH927-FXP-Control cells (Fig. 2A and C, respectively) did not differ significantly from those for mock-transfected cells.

FIG. 2.

Syncytium formation in FIV_Ple-E env-transfected cells. AH927 cells expressing feline CXCR4 (A, C) or feline CXCR4 plus feline CD134 (B, D) were transfected with eukaryotic expression vectors bearing FIV_Ple-E (A, B) or FIV_Fca-GL8 (C, D) env. Monolayers were fixed and stained at 48 h posttransfection, and photographed representative images are shown.

FIV_Fca strains vary in their sensitivity to the inhibition of viral entry by soluble CD134L (49) and the CXCR4 antagonist AMD3100 (49). The sensitivity of FIV_Ple-E, FIV_Fca-GL8, and FIV_Fca-B2542 (14) to soluble feline CD134L and AMD3100 was assessed with MYA-1 (24) cells. AMD3100 inhibited infection with all three viral pseudotypes in a dose-dependent manner (Fig. 3A and C), with ≥95% inhibition of infection at 400 ng/ml antagonist. In contrast, while GL8 pseudotypes resisted inhibition by soluble CD134L at all but the highest concentration (50 μg/ml) (Fig. 3B and D), infection with the B2542 and FIV_Ple-E pseudotypes was markedly reduced at 0.4 μg/ml CD134L (∼70% inhibition at 0.4 μg/ml) (Fig. 3D).

FIG. 3.

Sensitivity of FIV_Ple-E to AMD3100 and soluble CD134L. MYA-1 T cells were infected with HIV(FIV)-luciferase pseudotypes bearing Envs from FIV_Ple-E or FIV_Fca-GL8 and B2542 in the presence of increasing concentrations of CXCR4 antagonist AMD3100 (A, C) or soluble Fc-TNC-CD134L (sFc-TNC-CD134L) (B, D). (A, B) Luciferase activity at 72 h postinfection. Each point represents the mean (n = 3) ± standard error. CPM, counts per minute. (C, D) Percent infection relative to mean infectivity of control with no antagonist. TNC, tenascin.

FIV replication in canine CLL cells is CD134 dependent (50); we therefore assessed the growth of FIV_Ple-E in CLL cells, asking whether CD134 expression was essential for viral growth. FIV_Ple-E was compared with FIV_Fca-GL8, a strain of FIV for which productive infection is CD134 dependent, and FIV_Ple-458 (6), a subtype of FIV_Ple known to infect independently of CD134 and CXCR4 expression (36, 44). CD134 expression rendered CLL cells permissive for productive infection with FIV_Fca-GL8 and FIV_Ple-E, while control cells were resistant to infection with either virus (Fig. 4). In contrast, both CLL and CLL-CD134 were susceptible to infection with FIV_Ple-458, confirming that infection with FIV_Ple-458 is CD134 independent.

FIG. 4.

Productive infection with FIV_Ple-E is CD134 dependent. Canine CLL cells (CLL) or CLL cells stably transduced with retroviral vectors bearing feline (domestic cat) CD134 or vector only (CLL+CD134) were infected with MYA-1 T-cell-grown FIV_Fca-GL8, FIV_Ple-E, or FIV_Ple-B, and viral replication was quantified by nonisotopic reverse transcriptase (RT) assay (colorimetric assay; absorbance at 405 nm).

In this study, we demonstrate that FIV_Ple-E-Sangre shares a common receptor and coreceptor with the lentivirus of the domestic cat. There have been conflicting reports regarding the pathogenicity of lentiviruses in nondomestic felids, and the observation that two distinct biological phenotypes can be discerned between subtypes B and E of FIV_Ple may indicate distinct pathogenicities in vivo. The virus-receptor interaction is a critical determinant of viral cell tropism and cytopathicity, and FIV_Fca targets activated CD4⁺ T cells primarily due to the restricted expression of the primary receptor CD134 and coreceptor CXCR4 in this T-cell compartment (12, 34, 49). Accordingly, a decline in circulating CD4⁺ T lymphocytes is associated with FIV infection of the domestic cat (1). The shared usage of CD134 and CXCR4 between FIV_Fca and FIV_Ple-E would be consistent with a decline in CD4⁺ T lymphocytes following FIV_Ple-E infection of lions. Previous studies have observed that both captive (7) and free-ranging (7, 30) lion populations show depletions of CD4⁺ T lymphocytes consistent with selective targeting of this T-cell subset by the virus. A study of free-ranging lions (30) examined blood samples from animals in Tanzania and South Africa, where FIV_Ple subtypes A, B, and C have been described previously (6). As FIV_Ple-B infection appears to be independent of CD134 expression, future studies should address whether usage of CD134 as receptor by FIV_Ple-E facilitates more efficient targeting of activated CD4⁺ T lymphocytes and a more marked immunopathology among free-ranging lions in countries such as Botswana, where the subtype E virus has been described (41). Similarly, FIV_Ple-B may have acquired usage of a distinct cell surface molecule that is expressed on CD4⁺ T lymphocytes as a means of targeting this subpopulation. Alternatively, CD4⁺ lymphocyte targeting may not solely be the result of restricted receptor expression; rather, it may be the sum of a sequence of events: the trapping of virions on dendritic cells in the lymph node through DC-SIGN interactions (HIV, simian immunodeficiency virus, and FIV_Fca) (3, 13, 18-20, 28, 29), an interaction with a primary receptor found on CD4⁺ T cells (CD4 for HIV/simian immunodeficiency virus [11, 23, 32] and CD134 for FIV [34]), and a subsequent interaction with seven-transmembrane domain family molecules, such as CCR5 or CXCR4 (2, 15-17, 31, 39, 47, 51). Thus, in order to target CD4⁺ T cells, the primary receptor for the virus may not necessarily be restricted in expression to CD4⁺ T cells; specificity may result from restricted expression of the viral coreceptor or from selective trafficking of CD4⁺ T cells through the dendritic-cell-rich regions of lymphoid tissues where virions are trapped. Comparative studies between the felid and the primate lentiviruses will go some way to addressing the relative significance of the virus-receptor interaction to the depletion of CD4⁺ T lymphocytes.

FIV_Fca strains vary in their sensitivity to antagonism by soluble CD134L (49). When infection with FIV_Ple-E was compared with that of FIV_Fca strains GL8 (46) and B2542 (14), FIV_Ple-E infection was found to be sensitive to modulation by CD134L, suggesting that the interaction between FIV_Ple-E Env and CD134 is of similar affinity to that of B2542. However, unlike B2542, FIV_Ple-E Env-bearing pseudotypes required determinants in CRD-1 and -2 of CD134 for infection, similar to GL8 and CPG41 (50). The nature of the interaction between Env and CD134 may correlate with the propensity of the virus to interact directly with CXCR4 (reviewed in reference 44). FIV_Ple-E is the first virus we have identified to date that is sensitive to CD134L and yet interacts with CD134 in a manner similar to “early” isolates such as GL8 and CPG41, indicating that the two properties are not mutually exclusive. Future studies should investigate whether endogenous lion CD134L plays a role in controlling viral replication and ameliorating viral pathogenicity.

FIV_Ple-E and FIV_Ple-B differ in their capacities to replicate in feline cell lines (36, 44), their sensitivity to receptor antagonists (36, 44), and their requirement for ectopic expression of CD134 for productive infection. Here, we provide further compelling evidence that FIV_Ple-B utilizes a distinct receptor(s) for infection, replicating efficiently in CD134-negative cells. If the differences in receptor usage between FIV_Ple-E and FIV_Ple-B affect cell tropism and pathogenicity in vivo, this will be of significance to the management of both free-ranging endangered felids and breeding populations in captivity.

Nucleotide sequence accession numbers.

The sequences of lion CD134, CXCR4, and FIV_Ple-E-Sangre have been deposited in GenBank under accession numbers EU915482, EU915483, and EU915484.