Small ruminant lentiviruses (SRLV) are a group of retroviruses causing a persistent multisystemic and fatal disease in sheep and goats. This viral continuum includes the formerly known Maedi-Visna virus (MVV) described in sheep and the caprine arthritis encephalitis virus (CAEV) described in goats [1
]. SRLVs are nowadays classified into five genotypes, from A to E [2
]. Genotype A and B strains, which are sometimes referred to as MVV-like and CAEV-like, respectively, are widely distributed, while genotype C, D, and E strains were only found in more restricted areas [5
]. Previous phylogenetic studies have demonstrated that these viruses are capable to cross the species barrier between sheep and goats [2
SRLV infections are mostly transmitted at an early age when kids suckle milk and colostrum from their infected mother. Less efficient SRLVs can also disseminate by air when animals are housed in close contact in intensive production systems [11
]. Once SRLVs are transmitted and they cross the mucosal surface, they target monocytes and macrophages where they integrate as provirus in the host cell genome. Through the circulation, stem cells or precursors cells from the bone marrow get infected by circulating macrophages allowing the establishment of a life-long infection in the animal [1
After an incubation period of several years, up to 30% of the SRLV infected animals will start to suffer from degenerative lesions in joints, mammary glands, and lungs [14
]. Besides the impact on animal welfare, SRLV cause considerable economic losses in the small ruminant industry due to early culling, reduced milk production, and animal movement restrictions [3
No vaccines or therapies are currently available and the control of SRLV mostly relies on strategies to prevent the introduction of the virus in the herd or transmission from infected mothers to her newborns. These preventive actions rely on good diagnostic methods that ideally should have perfect sensitivity, specificity, and allow for an early detection of SRLV positive animals [17
]. In routine diagnostics, the most frequently used assays are the agar gel immunodiffusion (AGID) and the enzyme-linked immunosorbent assays (ELISA). AGID is considered to be highly specific and reproducible, but it has a relative low sensitivity [18
]. ELISA, on the other hand, is known to be cheaper, can be automated, works for a variety of diagnostic matrices (milk, serum, semen), and allows quantitative interpretations. These advantages make that ELISAs are used as the method of choice for monitoring and large surveillance programs [17
]. Despite the good performance of ELISA tests, none of them allows for detecting all infected animals. Many indirect and competitive ELISAs have been developed using antigens of one or multiple strains, but new strains continue to emerge and thereby reduce the detection range of individual kits. In addition, the slow seroconversion of newly infected animals [22
], the tendency for fluctuating antibody response during the first month post-infection [23
] and individual differences between sheep and goats in the development of the immune response against SRLV antigens [23
] make that negative results in ELISA, or in other serological tests, sometimes do not reflect the correct infection status of an animal [21
]. Besides issues with imperfect sensitivity, also false positive results in ELISA cause major problems in control programs, leading to the incorrect suspension of the SRLV free status and necessitating time and money to perform confirmatory testing.
In addition to serological tests, real-time PCR (qPCR) assays have been developed to detect the presence of viral nucleic acid in tissue and blood. qPCR can detect SRLV infection prior to the antibody response and in theory allows for detecting the latent provirus during the entire lifespan of the animal [21
]. Although considered to be highly specific, the sensitivity of this technique can be hampered by the high genetic heterogeneity between strains and the low proviral load existing in infected animals [26
]. To improve sensitivity, new qPCRs have been developed over the years using new primers and probes to include more strains from various geographical areas [20
In Belgium, a voluntary control program has been implemented that offers the possibility to obtain a SRLV-free certification. To receive the SRLV free status, farmers regularly have to show the negative status of their flocks. The current testing protocol is based on an initial screening where each animal is tested in one ELISA (Elitest, Hyphen Biomed, Neuville-sur-Oise, France) that was shown to have good test characteristics in previous studies. ELISA positive samples are then tested by a combination of two AGIDs (Maeditect kit, Apha Scientific, Addlestone, Surrey, United Kingdom; AGID CAEV P28, Idexx, Westbrook, ME, USA) for confirmation [28
]. When the ELISA result is not confirmed by the AGID, a new serum and whole blood sample need to be collected from the suspected animal, which are then tested in ELISA, AGID, and qPCR before a final decision on the infection status of the animal is taken [4
]. In 2014, 508 sheep farmers and 27 goat farmers participated to the voluntary program, and although it functions well, regular problems are experienced [4
]. These mostly relate to some unexpected positive ELISA results in flocks that are certified SRLV-free for multiple years and are suspected to be false positive reactions. These positive results, together with the temporal suspension of the certificate and the time and costs associated with the collection of extra samples for confirmatory testing, causes frustration among participants, making it that farmers regularly drop out of the program.
To improve the testing protocol of our national control program, we decided to implement an extensive comparative analysis of commercially available tests for SRLV detection in sheep and goat. Our analysis includes recently developed tests that were not thoroughly investigated before and uses a large number of samples from different backgrounds. Besides serological tests, such as AGID and ELISAs, also two qPCRs were included and the possibility to use blood clots as an alternative to PBMCs for molecular SRLV detection was evaluated. Good results with the latter matrix would allow for performing both serological and molecular SRLV detection on one single blood sample, thereby reducing the time and costs to obtain final results. The present study will not only be helpful for the Belgian national authorities in the fight against SRLV, but also for other countries that are willing to implement or improve their control programs for SRLV infections.
Due to economic consequences of SRLV infections and the absence of vaccines, multiple countries, including Belgium, have implemented voluntary or obligatory control programs [4
]. These control programs are essential to detect and eliminate SRLV infected animals from flocks and their success mostly relies on good diagnostic tools [17
]. The applied testing protocols differ between countries and use (combinations of) commercial or in-house produced tests. In Belgium, an indirect ELISA (Elitest MVV/CAEV, Hyphen Biomed, Neuville-sur-Oise, France) is used as a first screening tool, and positive ELISA results need to be confirmed in AGID. If the positive ELISA result is not confirmed in AGID, a new serum and whole blood sample is collected from the suspected animal and tested in ELISA, AGID, and qPCR before a final infection status is accorded to the suspected animal [4
]. This protocol has shown good performance but reoccurrence of SRLV infection in certified flocks indicates that a small percentage of infected animals remain undetected. On the other hand, occasional false positive results during the first line ELISA screening cause time consuming confirmatory analyses and frustration among participants. Therefore, we performed an extensive comparison of diagnostic tests for SRLV detection in the Belgian sheep and goat population in order to potentially identify new diagnostic strategies for SRLV detection. We evaluated five ELISAs, two AGIDs, and one in-house qPCRs detecting genotype A and B strains.
In line with previous results, individual AGID tests showed to be highly specific but they had a reduced sensitivity [20
]. Especially, the AGID based on CAEV p28 protein performed rather poorly. Besides fluctuation in antibody response over time [23
] and differences in the development of humoral response between individual animals [23
], the most probable explanation for this observation is the fact that antibodies against the capsid arise early in the infection and tend to decline later in the infection [36
] On the contrary, antibodies against the envelope glycoproteins, such as gp135, which are used in the Maeditect kit, are present at a later stage of the infection [25
]. Since the samples from the naturally infected animals (panel 1) were collected from animals of at least one year old, the positive field samples most probably did not originate from recently infected animals, making that it was to be expected that more animals would react positive with the Maeditect kit than with the AGID CAEV P28 kit. Furthermore, it is interesting to observe that more goat samples reacted positive with the AGID kit based on MVV antigen than with the kit using the CAEV antigen. This confirms the serological cross-reactivity induced by SRLV infection in sheep and goats [19
], and potentially indicates that some CAEV strains present in Belgium originate and have evolved from an ancient Maedi-Visna virus that was successfully transmitted from a sheep to a goat [39
]. An important outcome of this study was the finding that combining the results of both AGID tests resulted in 100% sensitivity and specificity of SRLV detection in both sheep and goats. Although AGID tests remain too labor intensive to serve as first line screening tests, the combination of both AGIDs seems to constitute an appropriate confirmatory test. It will be interesting to analyze in the future whether the antigens of both tests could be combined to produce one AGID test with perfect test characteristics.
When looking at the ELISAs, an overall good performance was observed. The highest sensitivities (>96%) in both sheep and goat samples were found with kits that use multiple antigens (Elitest MVV/CAEV, ID screen®
MVV/CAEV indirect, Eradikit™ SRLV screening test). These kits also showed good specificity (>95%). Boshoff et al. reported before that the combined use of envelop and capsid antigens resulted in the best ELISA results [41
] and the Elitest MVV/CAEV kit was also found as one of the best performing test in a previous comparative study [42
]. The highest specificity was found with the MVV/CAEV p28 Ab screening test that only uses one capsid antigen, which seems to strongly decrease the sensitivity of the test.
Both AGID and ELISA tests were capable to detect seroconversion early (i.e., three weeks post infection) after experimental infection. This might indicate that viral replication and the induction of the humoral immune response after intratracheal infection differs from what occurs after natural transmission. More recent studies however also found that lentivirus infections were characterized by rapid seroconversion [23
]. It is important to consider that only three experimentally infected animals were used in this study and that they were furthermore infected with different SRLV strains and different infectious doses, making that no generalized conclusions on viral replication and SRLV induced immunity can be drawn from this experiment.
Based on the good performance of the ELISA tests, they remain the best choice for the first line screening. For the Belgian situation, where sheep and goat samples are screened using the same protocol and using the same tests, an ELISA that is highly sensitive and specific in both species is preferred. Therefore, the use of the currently used Elitest MVV/CAEV kit seems justified and the ID screen®
MVV/CAEV indirect test could be a valid alternative for the first line screening. As indicated above, testing the positive samples found in the first line ELISA screening with the combined AGID test would allow to confirm the positive status of suspected animals or identify potential false positive ELISA results. Alternatively, it could be envisioned to implement a second ELISA as a confirmation test, since this would allow for obtaining confirmatory results in a shorter time span. It is normally advised to use a highly specific confirmatory test when accredited flocks are monitored [25
]. Based on our results, the MVV/CAEV p28 Ab screening test kit would thus be most suitable for that purpose. However, if we evaluate this protocol (Elitest MVV/CAEV followed by MVV/CAEV p28 Ab screening test for confirmation of positive results) in a theoretical and a posteriori analysis with our field samples, this does not seem a good option, since too many positive samples from the first screening would be classified as (false) negative (six sheep and one goat) due to the low sensitivity of the ELISA MVV/CAEV p28 Ab screening test kit, which would allow the virus to keep on circulating in the flocks. A better option seems the combination of the Elitest MVV/CAEV kit followed by the ID screen®
MVV/CAEV indirect kit as confirmatory test. This sequence of tests allowed for correctly identifying the infection status of all animals in an a posteriori analysis.
As PCRs theoretically should be able to detect SRLV infections before seroconversion and during the entire life span of infected animals [20
], we also evaluated the usefulness of this technique. Early detection before seroconversion was confirmed in samples from our experimental infection, but the level of proviral DNA was generally low and only intermittently detected in one out of three inoculated animals. De Andres et al. already described that the low proviral load observed after seroconversion makes qPCR less sensitive [20
]. Our qPCR results on field samples indicate that genotype A and B strains circulate in Belgium and that co-infection with both genotypes occurs. Importantly, our qPCR showed to be highly specific in both sheep and goat samples, but had a suboptimal sensitivity. This is probably related to the high genetic heterogeneity between strains circulating in Belgium and/or the low proviral load in the collected samples [21
A major disadvantage of the current use of the qPCR as confirmatory test is that a second blood sample needs to be collected from animals that are suspected to be SRLV infected in order to prepare leucocyte pellets, making that time and money is lost before a final status can be determined. We therefore examined whether blood clots, which contain monocytes/macrophages and are available in the tube used for serum sampling, could not be an alternative to leucocyte pellets. This however showed not to be a valid option due to a high loss in sensitivity. This is probably due to a combination of the inhibiting effect of hemoglobin that is present in DNA extracts on the Taq polymerase [46
] and the further dilution of infected monocyte/macrophages within the blood clot harboring mostly red blood cells as compared to the amount present in leucocyte pellets. Interestingly, we observed that the relative sensitivity of SRLV detection in blood clots compared to leucocyte pellets was significantly lower in goats compared to sheep, but we cannot provide clear explanation for this observation at this moment.
In conclusion, a first screening with the Elitest MVV/CAEV ELISA kit, combined with both AGID tests and/or the ID screen®
MVV/CAEV indirect ELISA kit as a confirmation test seems to be a valid testing protocol for SRLV monitoring and certification in Belgium. It would allow for determining the correct infection status of participating animals and final responses would be obtained faster and cheaper when compared to the current protocol that still includes qPCR as a confirmation test. Although some differences have been observed between genotypes and subtypes circulating in different countries, genotype A and B strains have always been predominant [8
], making that these results will also be helpful for other countries thinking of implementing SRLV control programs.
Despite the good results obtained with this proposed protocol in our study, it remains to be seen if it will actually allow to completely eradicate the disease if that would be the goal of an implemented control program. Other diagnostic protocols that succeeded in reducing SRLV prevalence have been confronted with problems once the seroprevalence becomes low at the end of an eradication program [35
]. At that moment, other aspects like e.g., low level circulation of genotypes that are not efficiently detected by implemented tests [47
], the presence of some individual non-responders that remain undetected and continuously reintroduce the disease [33
], reintroduction of SRLV via illegal animal movements, via infected wildlife or via infected sheep in the goat population or vice versa if only eradication in one species is envisioned, etc. make that even with a well validated diagnostic protocol, it becomes difficult to identify the last remaining infected animals. This means that authorities have to continue to invest to prevent reintroduction and to identify and prevent the spread of less prevalent genotypes.