Abstract
Objectives
Feline immunodeficiency virus (FIV) is a lentivirus that induces AIDS-like disease in cats. Some of the antiretroviral drugs available to treat patients with HIV type 1 are used to treat FIV-infected cats; however, antiretroviral therapy (ART) is not used in cats as a long-term treatment. In this study, the effects of long-term ART were evaluated in domestic cats treated initially with the nucleoside transcriptase reverse inhibitor (NTRI) zidovudine (AZT) over a period ranging from 5–6 years, followed by a regimen of the NTRI lamivudine (3TC) plus AZT over 3 years.
Methods
Viral load, sequencing of pol (reverse transcriptase [RT]) region and CD4:CD8 lymphocyte ratio were evaluated during and after treatment. Untreated cats were evaluated as a control group.
Results
CD4:CD8 ratios were lower, and uncharacterized resistance mutations were found in the RT region in the group of treated cats. A slight increase in viral load was observed in some cats after discontinuing treatment.
Conclusions and relevance
The data strongly suggest that treated cats were resistant to therapy, and uncharacterized resistance mutations in the RT gene of FIV were selected for by AZT. Few studies have been conducted to evaluate the effect of long-term antiretroviral therapy in cats. To date, resistance mutations have not been described in vivo.
Introduction
Feline immunodeficiency virus (FIV) belongs to the Retroviridae family, and infects domestic and wild cats. 1 In domestic cats, FIV infection is associated with a decline in CD4+ and an increase in CD8+ peripheral blood lymphocyte counts, with consequent inversion of the CD4:CD8 T-cell ratio. 2 The final stage of FIV-induced pathology is characterized by the development of AIDS, 3 similar to the syndrome caused in humans by human immunodeficiency virus type 1 (HIV-1).
Antiretroviral therapy (ART) was designed to control HIV-1 viremia and allow immune reconstitution. 4 Different steps of the HIV-1 replicative cycle are the targets of therapeutic intervention. Inhibitors of viral entry and fusion, reverse transcriptase, integrase and protease were developed specifically to control HIV-1 replication. 5 Combinations of antiretroviral agents directed against at least two distinct targets suppresses viral replication, reducing the HIV-1 viral load (VL) below detection limits and restoring T lymphocyte CD4+ blood counts. Thus, these parameters are commonly used as markers of AIDS progression and are also used to evaluate the effectiveness of therapy in patients infected with HIV-1. 6
Reverse transcriptase (RT) is a viral enzyme that converts the viral single-stranded RNA into a double-stranded proviral DNA that can be integrated into the genome of the host cell. This is a crucial step in the replicative cycle of the virus. 7 Inhibitors of this enzyme are classified as nucleoside or non-nucleoside RT inhibitors, and were the first class of antiretroviral drugs to be developed. 8 Nucleoside RT inhibitors (NRTIs) act as chain terminators, blocking DNA synthesis. Non-nucleoside RT inhibitors (NNRTIs) are directed to a specific binding pocket within the HIV-1 RT, distinct from the catalytic site. 5 Among the NRTIs used, zidovudine (AZT) was the first anti-HIV drug approved by the Food and Drug Administration. 9 Lamivudine (3TC) is another NRTI that works synergistically with AZT. 10 The AZT/3TC combination in patients infected with HIV-1 yields better results in immunologic and virologic parameters than does treatment with either drug alone. 10
Despite the benefits of ARTs, antiviral drug resistance and side effects are critical factors that restrict the success of therapy. 11 Resistance to a determined antiviral agent is identified by the presence of specific mutations in the molecular target of the therapy, detected through genotypic analysis. 12
At present, there is no specific treatment to control FIV. Several ARTs used to control HIV-1 proved to be toxic to cats and some of them do not inhibit FIV, for example protease inhibitors. 13 So far, AZT is the only antiviral recommended by the European Advisory Board on Cat Diseases guidelines to treat cats infected with FIV. 14 AZT and 3TC, alone or in combination, have been used for in vivo and in vitro tests against FIV infection,15–17 and information about clinical use and dosage are varied and contradictory. 15 In natural FIV infection, VL reduction and CD4:CD8 cell recovery were more pronounced with AZT/3TC (25 mg/kg/12 h) than AZT (5 mg/kg/12 h) alone. 18 In experimental infection, AZT/3TC (20 mg/kg of each drug q12h) did not lead to a reduction in VL in peripheral blood mononuclear cells (PBMCs) nor to recovery of CD4:CD8 ratio. 16 In these studies, cats were treated for 1 year and 8 weeks, respectively, and genotypic evaluation of the FIV RT gene was not done in either case.
Despite these conflicting results, treatment with AZT and/or 3TC is performed by some veterinarians, with some clinical improvement. However, the prolonged use of ARTs in feline medicine is limited by the high cost and side effects of drugs, so data are scarce. We undertook a descriptive study in which 3TC was added to a regimen of long-term AZT therapy. Cats evaluated in this study were from a previous study. 19 The analysis was conducted from 2006–2010.
Material and methods
Study population
Four cats (aged between 8 and 15 years) naturally infected with FIV were followed for 4 years. The cats lived exclusively indoors after being rescued from the street. The cats were rescued at different times and maintained in a shelter. All the cats were asymptomatic. Serologic testing was conducted at the time of rescue, and the cats were positive for FIV antibodies and negative for the feline leukemia virus (FeLV) p27 antigen. The AZT treatment (5 mg/kg/12 h PO) was initiated soon after diagnosis of FIV infection, and the infected cats were kept isolated from other cats.
Monotherapy with AZT was maintained for a period ranging from 5–6 years, and then changed to AZT (5 mg/kg/12 h PO) plus 3TC (4 mg/kg/12 h PO) (Figure 1). The AZT/3TC therapy was maintained from March 2007 to March 2010. In the last year of AZT monotherapy we started follow-up (Figure 1). Every year, the animals were physically examined and a blood sample was collected to perform the analysis described in Figure 1 and Table 1.
Schematic timeline for the treatment of the antiretroviral therapy (ART) group and the follow-up. Zidovudine (AZT) monotherapy was initiated at different points, as the cats were rescued and feline immunodeficiency virus infection was diagnosed. The follow-up comprised the final period of AZT monotherapy and the AZT/lamivudine (3TC) therapy. Black arrows indicate the evaluation of viral load and sequencing. Gray arrows indicate the evaluation of CD4+ and CD8+ lymphocytes. Light-gray dashed arrow shows the time of death of cats RJ06 and RJ24
Description of analysis among groups
This year evaluations were done twice to the feline immunodeficiency virus (FIV) + antiretroviral therapy (ART) group (before and after therapy)
Genotyping performed on DNA and RNA
Genotyping performed on DNA
Two years after initiation of the AZT/3TC therapy, the CD4:CD8 ratio and hematologic evaluation was done in privately owned healthy cats in order to do a comparison with AZT/3TC-treated cats. They were divided into three groups: ART (four cats treated with AZT/3TC); FIV-NEG (five cats uninfected by FIV/FeLV); and drug-naIve (seven FIV-infected cats naive to ART). The drug-naive group was left untreated.
Genotyping of the RT region of FIV was done in the groups of ART and drug-naive cats yearly. Table 1 describes the groups and the analyses performed.
Statistical analysis
The non-parametric Kruskal–Wallis test was used for statistical analysis to compare differences among groups; Epi Info 3.5.1 (http://wwwn.cdc.gov/epiinfo/) was used. Values of P <0.05 were considered statistically significant.
FIV diagnosis
The diagnosis of FIV was performed in all cats using a test kit (SNAP FIV/FeLV Combo; IDEXX) according to the manufacturer’s instructions. FIV status was confirmed by nested PCR for gag, pol and env. 19
Hematologic evaluation
Complete blood counts were evaluated by an automatic cell counter (Sysmex 1800i), and differentials were determined by microscopic evaluation of blood smears after May-Grünwald Giemsa staining.
Flow cytometry
Absolute CD4+ and CD8+ lymphocytes were quantified by flow cytometry, according to Barlough et al. 20 Whole blood samples were incubated with mouse antifeline CD4 fluorescein isothiocyanate and mouse antifeline CD8 α/β phycoerythrin monoclonal antibodies (AbD Serotec). Acquisition and analyses were performed on FACSCalibur flow cytometer (Becton and Dickinson).
Genomic extraction
Genomic DNA was extracted from 200 µl of whole blood and RNA was extracted from 140 µl of plasma with a QIAamp DNA Blood Mini Kit (QIAGEN) and a QIAamp viral RNA Mini Kit (QIAGEN), respectively, according to the manufacturer’s protocol.
Viral load
Plasma VL was evaluated through quantitative PCR. Quantification of FIV RNA copies was based on a standard curve prepared with the plasmid pTM219. This plasmid contains the FIV TM2 genome (subtype B) cloned in the bacterial plasmid pUC119. 21
RNA was extracted from feline plasma samples, as described previously, and reverse transcribed into complementary DNA using a High Capacity cDNA Archive Kit (Applied Biosystems), according to the manufacturer’s instructions. Real-time cycling conditions were 95°Cfor 10 mins, followed by 45 cycles of 95°C for 15 s and60°C for 1 min, and a final elongation step at 72°C for5 mins. Primers directed to gag region (forward5′ TGATCGTACCCATCCTCCTGA 3′ and reverse5′ CTGGGCTCTGCTTGTTGTTCT 3′; 106 base pair amplicon) were designed from the sequence of TM2, a Japanese isolate of FIV.
Nested PCR amplification and sequencing
Sequencing of the pol (RT) region was carried out in DNA and RNA samples. The nested PCR primers and conditions were as described in Martins et al. 19 Sequences were aligned with FIV reference sequences and HIV-1 sequence isolate HXB2 available at the GeneBank (www.ncbi.nlm.nih.gov/) using the ClustalW tool.
Results
All cats evaluated were asymptomatic at the beginning of the study. During the 4 years of evaluation (Figure 1) one cat remained asymptomatic (RJ03). At the end of study one cat (RJ05) presented with cachexia, two cats (RJ06 and RJ24) presented with renal disease and chronic diarrhea, and one of them (RJ06) showed neurologic alterations (ataxia and head tilt) (Table 2); neutrophilia and normocytic/normochromic anemia were hematologic alterations associated with both cases. As AZT and 3TC are metabolized by the kidneys,22,23 ART was suspended, owing to renal disease (evaluated through high blood levels of creatinine and blood urea nitrogen) in February 2010. Owing to clinical decompensation, 4 months after therapy was withdrawn, two cats (RJ06 and RJ24) were euthanized (Figure 1 and Table 2).
Relationship between clinical conditions, therapy, viral loads and hematologic values during follow-up
Numbers in bold indicate changes after termination of therapy
HCT = hematocrit; MCV = mean cell volume; AZT = zidovudine; 3TC = lamivudine; ND = not determined
The first analysis of hematologic and CD4:CD8 parameters was performed in April 2009, around 2 years after the start of the AZT + 3TC therapy (Table 2). Although red blood cell (RBC) parameters remained within normal values, significant differences were found in RBC (P = 0.0095), hemoglobin (P = 0.0464) and mean cell volume (P = 0.0277) values among the three groups studied (Table 3). Paired statistical evaluation was carried out between the following groups: ART × FIV-NEG; ART × drug-naIve and drug-naIve × FIV-NEG. In these evaluations we observed that some hematologic parameters were statistically significant when comparing the ART group with the FIV-NEG and DRUG-NAIVE groups. No statistical significance was found in the analysis between DRUG-NAIVE × FIV-NEG (Table 3). Anisocytosis and polychromasia were observed in RBCs from cats receiving ART. After drugs were withdrawn, RBC values improved for all cats, albeit only mildly in cats with renal disease (RJ06 and RJ24) (Table 2).
Hematologic parameters (mean ± SD) among groups
Numerical data in bold indicates P <0.05
Analyses performed 2 years after zidovudine + lamivudine therapy (2009)
RBCs = red blood cells; Hgb = hemoglobin; HCT = hematocrit; MCHC = mean cell hemoglobin concentration; MCH = mean cell hemoglobin; MCV = mean cell volume; ART = treated with antiretroviral therapy (ART); FIV-NEG = uninfected by feline immunodeficiency virus (FIV)/feline leukemia virus; DRUG-NAIVE = FIV infected cats naive to ART
Although total lymphocytes, CD4+ and CD8+ cells, and CD4:CD8 ratios were lower in the ART group (Table 2), only relative CD4+ lymphocyte counts were significantly different among the three groups (P = 0.0072) (Table 3); there was no statistically significant difference in total lymphocytes, CD4+ and CD8+ cells, or CD4:CD8 ratios.
The first assessment of VL was made after an average period of 4 years of AZT monotherapy. Statistical analyses were done after changing AZT monotherapy to AZT + 3TC therapy and after treatment interruption, and no statistical differences were observed (P = 1.0000 and P = 0.2482, respectively). Twenty days after 3TC was added to AZT, a decrease of 1 log (factor of 10) in VL was observed in one cat (RJ06) but it was not sustained during the time of the study. Four months after drug interruption, a slight rise in VL was seen for 3/4 cats, but the VL and CD4+ T-lymphocyte counts were not correlated with clinical manifestations (Table 2).
RT sequencing was performed yearly from DNA samples obtained from PBMCs of infected cats. In a previous study, 19 we found three uncharacterized resistance mutations in positions related to drug resistance in HIV-1 RT in some cats from the ART group (K64R, K69R and L73I). These mutations were not found in the drug-naive group or in reference sequences. 19
In this study, these uncharacterized resistance mutations were also evaluated in plasma-derived viral RNA, the viral population directly exposed to ART selection pressure. In two cats (RJ05 and RJ06), we found concordant profiles to the ones obtained from PBMC DNA, although in some analyses a mixed population (wild type/mutant) was observed (Table 4). For one of the cats (RJ24), mutations K64R + L73I were found in DNA from PBMCs, but from plasma samples, wild type population was seen in codon 64 (Table 4). However, 4 months after treatment interruption, only wild type virus was detected on codon 73. In FIV from drug-naive cats, no mutations in the RT region were seen during follow-up (data not shown). Taken together, our results suggest that, similarly to HIV, drugs do select for variant forms of the virus that are not necessarily the fittest forms.
Unusual mutations in peripheral blood mononuclear cell DNA and plasma RNA in the antiretroviral treatment group during follow-up
AZT = zidovudine; 3TC = lamivudine
Discussion
AZT monotherapy was the first therapy to control HIV-1 infection in the late 1980s and the first drug evaluated in FIV treatment.15,24 However, long-term use of AZT is not common in feline medicine owing to the development of anemia shortly after the start of treatment. 16 In our ART group, statistical analysis showed the influence of ART in some hematological parameters (Table 3). RBC count was significantly lower compared with other groups, while hematocrit did remain constant (Table 3). The rise in RBCs after treatment interruption shows the direct effect of AZT in this parameter (Table 2). However, owing to chronic renal disease, this rise was slight in cats RJ06 and RJ24 (Table 2). This is associated with anemia due to reduced production of erythropoietin, a renal hormone that controls bone marrow production of RBCs. 25 Polychromasia indicates that bone marrow is still able to respond to AZT’s toxicity. A similar result was observed in a study where symptomatic FIV-infected cats were treated with AZT (10 mg/kg/day) for 2 years. 17 Neutropenia, another side effect related to AZT in humans, 26 was not seen in our group of cats (Table 3) but it was reported in cats treated with higher doses of AZT (100–150 mg/kg/day) than we used (10 mg/kg/day). 16
Drug resistance is another factor that restricts prolonged monotherapy for the treatment of HIV-1 infection. 7 The absence of proofreading activity by RT in association with a high mutation rate allows errors to be introduced during DNA synthesis, leading to genetic variation within retroviral populations.27,28 Hence, ART selects for a viral population with resistance mutations in HIV-1 after a short period of treatment. 29 FIV resistance to NRTIs has been evaluated in vitro but pathways to acquiring resistance, as well as the types of mutations obtained, are unknown.30–34 The FIV RT gene has two molecular signatures responsible for high-level resistance to multiple NRTIs in HIV-1 RT: M41L and an amino acid deletion at codon 67 (d 67),followed by T69G.19,35,36 These molecular signatures may represent a natural genetic barrier to some NRTIs.
Resistance mutations in the FIV RT gene were not seen in a study evaluating the efficacy of AZT in vivo, but the cats were treated for only 4 weeks. 35 We found uncharacterized resistance mutations K64R and K69R in two cats and a novel mutation, L73I, in one cat in the first genotypic RT evaluation after >4 years of AZT monotherapy. These amino acid positions are located in the fingers subdomain of HIV-1 RT, which surround the incoming nucleotide. 37 In a previous study, 19 these mutations were not seen in 18 untreated cats, just as in reference sequences. They are in positions related to the HIV-1 resistance mutations K65R, K70R and L74I. 19 In HIV-1, mutation K65R is selected for with 3TC, 38 tenofovir, abacavir (ABC), dideoxycytidine and didanosine (ddI) treatments. 39 K70R is the first mutation selected for by long-term AZT monotherapy in HIV-1, and is associated with high levels of AZT and d4T resistance. 29 The L74I mutation is associated with resistance to ABC and ddI, 40 and with combined therapy AZT + ABC. 41 One cat (RJ24) had mutations K64R + L73I, but these mutations were not seen together in the circulating virus (RNA), only in the integrated proviral DNA (Table 4). As seen for HIV-1, the selection for K65R and L74V/I on the same genome is uncommon, 42 with 65R >K reversion in the double mutant K65R + L74V virus. Like HIV-1 double mutants, it seems that K64R + L73I mutations are incompatible in FIV.
None of the resistance mutations previously related with the RT gene were found in our samples in either PBMC DNA or plasma RNA.30,32,33 However, our results suggest that selective drug pressure allowed the virus to accumulate mutations at codons 64, 69 and 73 over time. Although the samples were not analyzed before the start of therapy, these mutations were not seen in samples from the drug-naive group. 19
For HIV-1, factors such as incorrect use of ART and suboptimal drug metabolism are related to treatment failure due to the emergence of viral strains resistant to the drugs. 43 In our analysis, persistent viremia during follow-up suggests treatment failure. However, therapeutic success may have occurred at the beginning of therapy, as treatment was being carried out for several years before the first VL evaluation. Cats demonstrated a higher VL 20 days after the introduction of AZT+ 3TC therapy, except cat RJ06. This sample had uncharacterized resistance mutation K69R since the beginning of follow-up (Table 4). A related mutation in HIV-1, K70R, is selected with AZT. 29 Thus, the decrease of almost 1 log in VL observed only in this sample after introduction of 3TC can be explained by sensitivity to 3TC. Nevertheless, this sample showed a progressive increase in VL since 2009, with concomitant neurologic symptoms (Table 2), a condition potentially related to FIV. 18
Four months after treatment interruption, a slight increase in VL was observed for 3/4 cats, two cats were euthanized because of worsening clinical conditions (RJ06 and RJ24) and the other cat (RJ05) presented with cachexia (Table 2). Among the three samples with uncharacterized resistance mutations, one (RJ24) presented the wild viral population in plasma RNA (Table 4). Data on clinical evolution and sequence profile may demonstrate that the wild type viral population is fitter than the mutant population. Viral fitness of mutant and wild type FIV was demonstrated in experimental infection, in which the wild type Petaluma FIV strain was more virulent than the AZT-resistant strain. 20 Nevertheless, this resistant strain was not genotypically characterized.
It is important to emphasize that, as with HIV, FIV infection has a slow progression. 3 The length of infection of the cats in this study was not determined. The strain was previously characterized as belonging to subtype B, 19 which appears to be a less pathogenic strain than subtype A. 44 Therefore, it is difficult to state if clinical deterioration can be associated with the natural course of the infection or if the treatment was controlling the virus.
Beyond VL, CD4+ T lymphocyte recovery is an indicator of therapeutic success in HIV-1. 4 In our study, we found no correlation between CD4+ levels with clinical progression and ART therapy. Asymptomatic cats in both treated and naive groups presented low CD4+levels (between 100 and 200 cells/μl).
The CD4 molecule is the primary receptor binding to HIV-1,45,46 and FIV uses the CD134 molecule as primary receptor binding, which is expressed most abundantly on activated T-lymphocyte surfaces. 47 The presence of high titers of antibodies to CD134 was correlated with disease progression in cats. 48 This can be a good progression marker for FIV infection.
FIV was described 3 years after HIV-1. Since then, relatively little has been studied with respect to specific treatment. Further studies are necessary to better understand how antiretroviral drugs are metabolized in cats, which dosages are more effective with lower side effects, and which would be the best markers of progression and therapeutic success.
Conclusions
The data generated here, together with data previously evaluated, 19 suggest that cats receiving ART for prolonged periods are prone to therapeutic failure, just like humans infected with HIV-1, 12 and that uncharacterized resistance mutations may be related to AZT resistance.
GenBank accession numbers
Accession numbers for sequences generated in this work and submitted to GenBank: RJ44 – KC632575; RJ49 – KC632576.
Footnotes
Acknowledgements
We thank Ana Lucia Moraes Giannini for language corrections and critical comments; Orlando Ferreira for statistical analysis; Ricardo Gran for hematologic analysis; Lidia Theodoro Boullosa and Diana Mariani for assistance with sequencing; Alexandre Aleixo Rocha and Maria José Maia de Miranda for allowing us to collect blood from cats; and Takayuki Miyazawa for donating the plasmid pTM219.
Conflict of interest
The authors do not have any potential conflicts of interest to declare.
Funding
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.