Abstract
Most studies that investigate the prevalence of infections with feline leukemia virus (FeLV) are based on the detection of p27 antigen in blood, but they do not detect proviral DNA to identify the prevalence of regressive FeLV infections. The aim of the present study was to assess the prevalence and status of FeLV infection in cats in Southern Germany. P27 antigen enzyme-linked immunosorbent assay (ELISA), anti-p45 antibody ELISA, DNA polymerase chain reaction (PCR) of blood and RNA PCR of saliva were performed. Nine out of 495 cats were progressively (persistently) infected (1.8%) and six were regressively (latently) infected (1.2%). Cats with regressive infections are defined as cats that have been able to overcome antigenemia but provirus can be detected by PCR. Twenty-two unvaccinated cats likely had abortive infections (regressor cats), testing FeLV antigen- and provirus-negative but anti-p45 antibody-positive. Most of the FeLV-vaccinated cats did not have anti-FeLV antibodies. Both progressive, as well as regressive infections seem to be rare in Germany today.
Introduction
Feline leukemia virus (FeLV) infection is still a very important viral infection in Europe, with infection rates between 2.7% and 15.6% in healthy cats and between 7.6% and 30.4% in sick cats,1,2 and cannot be treated effectively. 3 Usually, veterinary practices use tests that measure FeLV p27 antigen in blood, but these tests do not identify cats that have recovered from transient viremia, as routine antigen tests are negative after the stage of transient viremia. 4 New, more sensitive diagnostic methods [including real-time polymerase chain reaction (PCR)] now make detection of infected, but antigen-negative cats possible. These methods have recently changed the understanding of the pathogenesis and outcome of FeLV infections. Hofmann-Lehmann et al developed a new classification scheme for the different outcomes of FeLV infection (Table 1). 5
Different outcomes of FeLV infection and the number of cats in each
nk = not known as PCR of organ samples was not performed
A significant decrease in the prevalence of infection among cats in many countries has been noticed in recent years. This decrease is caused by testing and elimination, as well as vaccination programs.6,7 A decrease has also been observed in Germany. 8 However, with only few exceptions, FeLV prevalence studies are uniquely based on the detection of FeLV p27 antigen in blood using enzyme-linked immunosorbent assays (ELISA) or similar immunochromatography techniques. However, the pathogenesis of FeLV infection is complex; free antigen can only be detected in some infected animals, usually the ones in a viremic state. Regressively infected cats test negative for antigen and only harbor provirus in their bone marrow cells after overcoming antigenemia.9,10 Thus, antigen testing underestimates the true prevalence of infection. Hofmann-Lehmann et al found that in addition to 6.9% of cats being antigen-positive, provirus-positive (sick and healthy) in Switzerland, 10.0 % of the cat population that tested negative for p27 antigen were positive for proviral DNA in their blood. 11 This result is surprisingly high and raises the question of whether the same situation occurs in other countries. The aim of this study was to survey different FeLV infection outcomes of cats in Southern Germany by using various testing systems (including antigen and antibody tests, as well as PCR) to detect different stages of FeLV infection.
Materials and methods
Animals
Blood samples of 495 randomly selected cats were tested for free FeLV p27 antigen, anti-p45 antibodies and FeLV DNA (provirus) in blood, as well as FeLV RNA (virus) in saliva. From August 2007 to May 2008, all cats that were presented to the Clinic of Small Animal Medicine of the Ludwig Maximilian University of Munich, Germany, were included in the study (366, 73.9%). Only cats whose owners did not permit blood sampling were excluded. Additionally, healthy cats (129, 26.1%) from different animal shelters in Southern Germany were sampled during the same time period. All cats available at the shelters were sampled. The health status of all cats was confirmed by a detailed physical examination performed by the first author (TE). In addition, these cats had no history of clinical signs during the 6 months prior to presentation. Data regarding vaccination status and other information were collected from the owners via a questionnaire. Missing data were obtained by telephone. However, some data could not be obtained, for example, those of most shelter cats.
All animals were classified as ‘indoor’ and ‘outdoor’ cats. Indoor cats were never allowed to go outside, not even in the garden or on a balcony. Outdoor cats were allowed to go outside and were able to roam freely.
ELISA for the detection of free p27 antigen in serum
Free FeLV p27 antigen was detected using a commercial ELISA (SNAP Kombi PlusFeLV/FIV antibody test; Idexx GmbH). The tests were performed according to manufacturer’s instructions.
Indirect ELISA for the detection of FeLV anti-p45 antibodies
FeLV anti-p45 antibodies were detected in serum using an indirect ELISA described previously. 12 Antibody concentrations ≤25% of the positive control were defined as negative. Antibody concentrations >25% of the positive control were defined as positive. This cut-off was chosen because specific pathogen-free (SPF) cats can have antibody concentrations of up to 25% (upper limit defined by mean + 3 standard deviations) (H Lutz, unpublished data) of the positive control in this ELISA. These results are most likely caused by unspecific reactions or antibodies against endogenous FeLV (enFeLV). EnFeLV is assumed to derive from ancient retrovirus infections with originally rodent retroviruses.13,14
DNA PCR in whole blood
DNA was extracted from whole blood using the MagNA Pure LC Total Nucleic Acid Isolation Kit (Roche Diagnostics AG). Phosphate buffered saline (PBS; Sigma) (100 µl) was added to the 1.5 ml microcentrifuge tubes containing 100 µl whole blood. Lysis buffer (300 µl) was added. The tubes were vortexed briefly and centrifuged. The whole volume was used for extraction of total nucleic acids by MagNA Pure LC Instrument (Roche Diagnostics AG) according to the manufacturer’s recommendations. DNA was detected in a TaqMan− fluorogenic real-time PCR detecting a U3 region of the long terminal repeats (LTR) using an ABI Prism 7700 sequence detection system (Applied Biosystems). PCR reactions were performed using primers (Table 2) as described. 15
Primers and probe used for TaqMan real-time PCR 15
Sequences of primers and probe are given in 5–3 orientation
bp = base pairs
RNA PCR in saliva
Saliva samples were collected using sterile cotton wool swabs (Copan Innovation). Swabs were rubbed gently under the tongue and on the cheek and stored at -80°C. Total nucleic acids were extracted from saliva swabs according to Gomes-Keller et al 16 using the MagNA Pure LC Total Nucleic Acid Isolation Kit. PBS (200 µl) was added to the 1.5 ml microcentrifuge tubes containing the swabs, and samples were vortexed briefly. Samples were incubated at 42°C for 10 min. After incubation, microcentrifuge tubes were centrifuged at 8000 x g for 1 min; swabs were inverted and re-centrifuged at 8000 x g for 1 min. Swabs were then discarded and 140 µl of the eluate was added to 300 µl lysis buffer. The entire volume of 440 µl was used for extraction of total nucleic acids by MagNA Pure LC Instrument according to the manufacturer’s recommendations.
For the RNA PCR, samples of 10 cats were pooled. 17 If the pooled samples were positive, all 10 cats were retested separately. In addition, all provirus-positive cats were tested separately. FeLV RNA was detected by TaqMan fluorogenic real-time RT-PCR assay with an ABI Prism 7700 sequence detection system using primers (Table 2) as described. 15
Statistical methods
The χ2 test was used to calculate a statistical significance between antigen-negative and antigen-positive cats, as well as between PCR-negative and PCR-positive cats. As this was a no case-control study, but animals were randomly selected and examined for the presence of antibodies, relative risks (RR) were calculated. Continuous variables were tested for normal distribution. Neither age nor concentration of antibodies were distributed normally. Therefore, in order to evaluate a difference concerning these variables, a Kruskal Wallis test was used. All significant factors were included in a logistic regression model with backwards selection. P values of <0.05 were considered significant.
Results
Nine out of 495 cats tested positive for free FeLV p27 antigen (1.8%) (Table 3). The presence of proviral DNA in whole blood was tested by PCR in all samples, of which 15 (3.0%) had positive results. All FeLV p27 antigen-positive cats tested positive for FeLV provirus. Six out of 486 antigen-negative cats were provirus-positive (1.2%). The saliva of 385 cats was tested by RNA PCR (Table 4). Four cats had positive results (1.0%); each of these cats was also p27 antigen-positive.
p27 antigen ELISA results in 495 cats tested
n = numbers of cats in every group, P = P-value
RNA PCR and anti-p45 antibody ELISA results of antigen-negative/provirus-negative, antigen-negative/provirus-positive, antigen-positive/provirus-negative and antigen-positive/provirus-positive cats, and positive results in total
n = number of cats
Of the 476 cats tested for anti-p45 antibodies, 78 had antibodies (16.4%) (Table 4). Fifteen cats were known to have been vaccinated (3.2%) and 23 were not vaccinated (4.8%). The other 398 cats had no antibodies (83.6%) (Table 4). Of the 462 antigen-negative, provirus-negative cats tested for anti-p45 antibodies, 75 cats showed positive results (16.2%). The difference in antibody presence between p27 antigen-negative, provirus-negative cats and p27 antigen-negative, provirus-positive cats was significant [P = 0.035; RR = 3.70; 95% confidence interval (CI): 1.75–7.79]. Median antibody concentrations in p27 antigen-positive, provirus-positive cats were lower than in p27 antigen-negative, provirus-negative cats (Table 4), and concentrations of the antigen-negative, provirus-negative ones were lower than those of cats that were p27 antigen-negative, provirus-positive. This difference was not statistically significant (P = 0.181).
There was no difference in the age distribution between the p27 antigen-negative, provirus-negative cats, the p27 antigen-negative, provirus-positive cats, and the p27 antigen-positive, provirus-positive cats (P = 0.089).
Outdoor access was allowed for 236 out of 495 cats (47.7%). Of these 236 cats, six cats were p27 antigen- positive, provirus-positive (2.5%) and four cats were p27 antigen-negative, provirus-positive (1.7%). Of the 215 strictly indoor cats (43.4%), one was p27 antigen-positive, provirus-positive only (0.5%); another one was p27 antigen-negative, provirus-positive (0.5%). The difference in antigen and provirus status between cats living indoors and those living outdoors was significant (P = 0.029; RR = 4.56; 95% CI: 1.01–20.56). Of the 495 cats, 383 had contact with other cats (77.4%). Eight of these were p27 antigen-positive, provirus-positive (2.1%). Six cats were p27 antigen-negative, provirus-positive (1.6%). Only 74 cats lived with no known contact with other cats (14.9%). None of them was p27 antigen-positive or provirus-positive. The difference in antigen and provirus status between cats that had contact with other cats and those that did not was not significant (P = 0.081).
The vaccination status of 288 cats was known. Of these, 42 cats were vaccinated against FeLV (14.6%). Four of the 246 cats without FeLV vaccination history were antigen-positive, provirus-positive (1.6%); another four cats were antigen-negative, provirus-positive (1.6%). Only one FeLV-vaccinated cat was antigen-positive, provirus-positive (2.4%). This difference between FeLV vaccinated and unvaccinated cats was not significant (P = 0.613). Only 15 (35.7%) of the vaccinated cats had antibodies, but this percentage was significantly higher than the percentage in the unvaccinated group (9.8%; P <0.001; RR = 3.63; 95% CI: 2.07–6.37).
In the logistic regression with the factors access to outdoors, presence of antibodies and originating from the clinic patient population versus from shelter, only the latter factor remained significant (P = 0.023). Animals presented to the clinic had a lower risk of being provirus positive [odds ratio (OR) = 0.245; 95% CI: 0.073–0.822].
Discussion
The standard diagnostic method for FeLV infection consists of tests that detect free p27 antigen in blood by ELISA or immunochromatography. 18 These tests can easily be performed in veterinary practice and are highly reliable to detect antigenemia. 19 The prevalence of FeLV antigenemia in this study was 1.8%. These cats were also provirus-positive and likely have what is now called ‘progressive infection.’ The prevalence of FeLV antigenemia is lower than reported in previous studies in Germany,6,8,20,21 but this is expected owing to the decrease of FeLV antigen-positive cats worldwide. Widespread testing and elimination programs, especially in shelters and breeding facilities, as well as vaccination programs, contribute to this decline.6,22,23
There is a significantly higher risk of FeLV antigenemia in this study for ‘outdoor’ cats than for those who live only indoors. Nevertheless, there were two cats in this study that were FeLV-infected and, according to the owners’ information, never had access to outdoors. One of these cats was a young pure-bred cat from a breeder who had at least one other FeLV-infected cat, which explains that particular positive result. The source of FeLV infection in the other cat is unknown. It could not be ascertained whether this cat had been with the owners from birth, so infection could have perhaps taken place before the cat became an ‘indoor’ cat.
In addition to the antigen-positive cats, FeLV provirus was detected in six antigen-negative cats by DNA PCR. Therefore, the percentage of ‘regressively-infected’ cats was 1.2%. This percentage is much lower than in a similar study in Switzerland where the prevalence of FeLV antigen-negative, provirus-positive cats was up to 10%. 11 The difference may be explained by the time span between the Swiss study and the present one. The Swiss study was performed in 1999 and 2000; in the meantime, FeLV prevalence may have further decreased. Alternatively, the FeLV infection rate in Switzerland may, indeed, be much higher than in Germany. Every antigen-positive cat was also provirus-positive, so there were no evidently false-positive antigen ELISA results.
As expected, the excretion of RNA in saliva was neither discovered in cats that were p27 antigen-negative, provirus-negative, nor in those that were p27 antigen-negative, provirus-positive. The latter ones harbor the provirus in their bone marrow and in blood cells released from the bone marrow; however, they do not carry the replicating virus and are therefore not a source of infection for other cats. 6 One cat was antigen-positive and provirus-positive, but tested negatively for RNA in saliva. It is very likely that this cat also sheds RNA in saliva but potentially only intermittently, which could explain the negative result. The negative result could also be potentially caused by inappropriate sample collecting or handling. However, this is very unlikely because all samples were taken and processed by the same person (TE). Theoretically, it could be possible that all samples were processed incorrectly. However, this is unlikely because all of the other antigen-positive cats tested for RNA in saliva were correctly found to be shedding the virus. Measuring RNA in saliva would be an easy, non-invasive method with very easy sampling technique compared with blood sampling and could be performed by owners and breeders themselves. 16 However, the fact that one cat remained undetected questions the sensitivity of the method. Although Gomes-Keller et al reported that they were able to detect FeLV RNA in saliva in 100% of experimentally- and naturally-infected antigenemic cats,16,17 using the same PCR method, these results cannot be confirmed completely in the present study. One possibility for the different results may be the different infection time. Gomes-Keller et al sampled experimentally-infected cats in a very early stage of infection. It is likely that the amount of virus shedding is higher (and, thus, more easily detectable) than the amount shed by a chronically infected cat.
Pooling the eluates of the buccal swabs certainly causes a dilution effect. Consequently, one would expect a lower sensitivity of this method compared with a separate analysis of any eluate. However, it could be shown that even after pooling the eluates of up to 30 cats, RNA in the saliva of one RNA-shedding cat could be detected by PCR. 17
Antibodies were detected in 78 cats (16.4%). All antigen-positive, provirus-positive cats had no antibodies. This is expected in persistently antigenemic cats because of the missing immune response.24–26 Antigen-negative, provirus-positive cats had, on average, significantly higher antibody concentrations (49.0%) than antigen-negative, provirus-negative or antigen-positive, provirus-positive cats. However, unexpectedly, two of these antigen-negative, provirus-positive cats had no antibodies. This could be potentially explained by a strong cellular and missing humoral immune response in these cats that also is able to keep the FeLV infection controlled. 27
There was a significant difference between vaccinated and unvaccinated cats concerning the number of cats with antibodies. More vaccinated cats had antibodies (35.7%); however, 64.3% of vaccinated cats had undetectable antibodies. Compared with other studies, detecting positive antibody titers in the majority of vaccinated cats,5,28 this result is surprising. The vaccines used today do not generally induce neutralizing antibodies, 29 and it is known that the antibody titer does not correlate with protection against FeLV and that a cat with a low antibody titer can be protected as well as a cat with a high titer. 30 Other studies have shown that some cats vaccinated against FeLV do not develop antibody titers before coming into contact with FeLV. 31 Another explanation for FeLV-vaccinated cats without antibodies could be the time of measuring antibodies. In many studies, antibodies are determined a few weeks after vaccination5,28 but, potentially, cats in this study may have been vaccinated a longer time ago. Additionally, the information of the owner or the vaccination record could have been incorrect in some cases.
The difference in number of FeLV-antigenemic animals between vaccinated (n = 42, 14.6%) and not vaccinated (n = 246, 85.4%) cats was not statistically significant. This is unexpected but could be explained by the low number of FeLV-infected cats in both groups. One of the vaccinated cats was progressively infected (antigen-positive, provirus-positive) which could be caused by an infection that had existed before the vaccine was applied or by the fact that the vaccination does not protect 100% of cats even if the vaccination is administered in accordance with the recommended vaccination schedule. 32 According to the owner it is unknown if the cat was tested before vaccination. Furthermore, it was shown that vaccination does not prevent provirus integration and viral replication; 25 therefore, a ‘reactivated’ latent infection could be another explanation for the positive result in this cat.
In 22 (9.2%) p27 antigen-negative, provirus-negative cats that had definitively not been vaccinated against FeLV, antibodies were detected. It is likely that these cats had contact with FeLV but could overcome the infection before the bone marrow became infected. According to a recent study by Hofmann-Lehmann et al, in every cat that has contact with FeLV, the provirus can be detected by PCR. 25 As the proviral load may be near the detection limit of the provirus PCR, these cats can be PCR-negative intermittently but may turn positive again later on. 5 The 22 antibody-positive but antigen- and provirus-negative cats of this study were not retested. Thus, it cannot be excluded that they will turn PCR-positive in the future. However, it is more likely that these cats, indeed, had an abortive infection.
A limitation of this study is the large number of cats in one of the three groups (antigen-negative, provirus-negative cats) and the very small number of cats in the other groups (antigen-negative, provirus-positive cats; antigen-positive, provirus-positive cats). As a result of this, a lot of testing revealed no statistical significance. Another limitation is that evaluations of anamnestic data are based on what the owners completed in the questionnaire. In addition, the data of the questionnaire only allows a statement about the housing conditions at the time of presentation and does not take into account if a cat had previous outdoor access.
Footnotes
Acknowledgements
We thank all owners who gave their support for sample collection. We are very thankful that the first author (TE) could perform all the tests by using the facilities at the Center for Clinical Studies at the Vetsuisse Faculty of the University of Zurich. We thank Dr Marina Meli for helping with every issue that arised.
Funding
We thank Idexx for sponsoring the FeLV/FIV Snap Tests used in this study.
Conflict of interest
The authors do not have any potential conflicts of interest to declare.