Contamination of live attenuated vaccines with an infectious feline endogenous retrovirus (RD-114 virus).

Retroviruses are classified as exogenous and endogenous retroviruses according to the mode of transmission. Endogenous retroviruses (ERVs) are retroviruses which have been integrated into germ-line cells and inherited from parents to offspring. Most ERVs are inactivated by deletions and mutations; however, certain ERVs maintain their infectivity and infect the same host and new hosts as exogenous retroviruses. All domestic cats have infectious ERVs, termed RD-114 virus. Several canine and feline attenuated vaccines are manufactured using RD-114 virus-producing cell lines such as Crandell-Rees feline kidney cells; therefore, it is possible that infectious RD-114 virus contaminates live attenuated vaccines. Recently, Japanese and UK research groups found that several feline and canine vaccines were indeed contaminated with infectious RD-114 virus. This was the first incidence of contamination of 'infectious' ERVs in live attenuated vaccines. RD-114 virus replicates efficiently in canine cell lines and primary cells. Therefore, it is possible that RD-114 virus infects dogs following inoculation with contaminated vaccines and induces proliferative diseases and immune suppression, if it adapts to grow efficiently in dogs. In this review, we summarize the incidence of contamination of RD-114 virus in live attenuated vaccines and potential risks of infection with RD-114 virus in dogs.

Endogenous retroviruses (ERVs)

Retroviruses enter host cells via binding of the envelope proteins with the host receptor(s). After entering the cells, viral RNA is reverse-transcribed into DNA and then the DNA is integrated into the host genome to be a provirus. Viral RNA is transcribed from the provirus and the structural and enzymatic proteins of the virus are synthesized from the transcribed viral RNA.

Retroviruses are classified as exogenous and endogenous retroviruses according to the mode of transmission [14]. Usually, exogenous retroviruses infect somatic cells but not germ-line cells, and they are transmitted horizontally by infection via viral particles. On the other hand, ERVs are retroviruses which have been integrated into germ-line cells. ERVs behave like normal genes and are inherited from parents to offspring as Mendelian’s law. ERVs occupy about 10 % of mammalian genomes and are mostly inactivated by deletions and mutations [14, 15, 17, 27]. However, a limited number of ERVs maintain their infectivity and infect new hosts as exogenous retroviruses [6].

Exogenous retroviruses are classified into seven genera, i.e., alpharetrovirus, betaretrovirus, gammaretrovirus, deltaretrovirus, epsilonretrovirus, spumaretrovirus and lentivirus. ERVs are divided into at least three classes I, II and III [14]. Class I ERV is closely related to exogenous counterparts of gammaretrovirus and epsilonretrovirus. Class II and III ERVs are similar to alpharetrovirus and betaretrovirus, and spumavirus, respectively.

Feline ERVs

All domestic cats have an infectious ERV, termed RD-114 virus [4, 12, 32]. RD-114 viral genomes have not been detected in large felids, such as lions and pumas [4]. RD-114 virus is closely related to baboon endogenous retrovirus (BaEV) in env region, but is distantly related to BaEV in gag-pol region, and is considered to be a recombinant virus between a feline ERV, termed FcEV, in gag-pol region and BaEV in env region [47]. BaEV is also a recombinant virus between a Papio cynocephalus ERV, termed PcEV, and a simian type D virus [19]. The gag-pol regions of both RD-114 virus and BaEV are closely related to gammaretroviruses (class I ERV), and the env region is closely related to betaretroviruses (class II ERV).

ERVs and pathogenicity

Generally, ERVs do not induce diseases in the original hosts. However, there are several incidences where ERVs exhibit pathogenicity; for example, ERVs of AKR mice induce lymphoma in the host [31]. Recently, it was found that replication activity of mouse ERVs was resurrected in Rag−/− deficient mice, which had no mature B- and T- lymphocytes, and in mice deficient in Toll-like receptors 3, 7 and 9 (TLR3, TLR7 and TLR9) triple deficient mice (TLR3−/−, TLR7−/−, TLR9−/− deficient mice) [52, 53]. In addition, activated ERVs in TLR3−/−, TLR7−/−, TLR9−/− deficient mice induced lymphoma in the host [53]. These reports suggest that infectious ERVs can exhibit oncogenicity in the host. Moreover, certain ERVs infect a new host and exhibit pathogenicity; for example, gibbon ape leukemia virus (GALV) that induces lymphoma in gibbons is considered to originate from an ERV of Asian rodents (Mus caroli, Mus cervicolor and Vandeleuria oleracea) [3, 18]. Recently, it was reported that an endogenous koala retrovirus (KoRV) is biologically active and may be associated with neoplastic diseases and immune suppression in koalas [39, 41–43]. The origin of KoRV is still unknown at present, although partial retroviral sequences closely related to KoRV and GaLV were found in an Australian rodent (http://espace.library.uq.edu.au/view/UQ:244963) and a bat (Megaderma lyra) [9]. Intriguingly, KoRV could transmit to rats by experimental infection and induced fibrosarcoma [10, 11].

Contamination of ERVs in human vaccines

Mice, pigs, cats, and chickens have infectious ERVs [6, 34]. In previous studies, it was reported that MMR vaccines (measles, mumps and rubella vaccines) and yellow fever vaccines that were propagated in chicken embryos were contaminated with endogenous avian leukosis viruses (ALVs) and endogenous avian retroviruses (EAVs), which originate from chicken embryonic fibroblast substrates [16, 45]. It is unknown whether contaminated endogenous ALV and EAV are infectious ERVs, because these studies only detected the viral RNA, proteins and reverse transcriptase (RT) activities in vaccines using RT-polymerase chain reaction (PCR), immunoblotting and the RT assay, respectively (16, 45). Nevertheless, no evidence was found that these ERVs had infected humans by vaccination [16, 45].

Contamination of animal vaccines with ‘infectious’ RD-114 virus

Many live attenuated vaccines for animals are manufactured using feline cell lines which may produce infectious RD-114 virus (Table 1). Therefore, it is possible that infectious RD-114 virus contaminates these vaccines [24]. We developed RT-PCR and realtime RT-PCR to detect RD-114 viral RNA and the LacZ marker rescue assay to detect infectious RD-114 virus in vaccines [37, 49]. When we examined feline live attenuated vaccines purchased in Japan (Vaccines F/g1, F/h2 and F/d3) for the presence of infectious RD-114 virus by the LacZ marker rescue assay, a vaccine manufactured by one company (Vaccine F/g1) was contaminated with infectious RD-114 virus (Table 2) [25]. The Japanese regulatory authority, National Veterinary Assay Laboratory (NVAL), also confirmed this finding independently [30]. In addition, we also confirmed that three products of ‘canine’ live attenuated vaccines purchased in Japan (Vaccines C/a1, C/a2 and C/b3), manufactured using ‘feline’ cell lines, were contaminated with infectious RD-114 virus (Table 2) [25]. The titers of RD-114 viruses in the contaminated vaccines were 1,800, 1,000 and 1.8 50 % tissue culture infective dose (TCID50)/dose, respectively (Table 2). Copy numbers of RD-114 viral RNA were also estimated by real-time RT-PCR. We found that 4.5×107, 9.7×107 and 8.3×106 copy number/dose of RD-114 viral RNAs were present in Vaccines C/a1, C/a2 and C/b3, respectively (Table 2) (unpublished data). Another research group in the University of Glasgow also confirmed that feline and canine live attenuated vaccines purchased in the United Kingdom were contaminated with infectious RD-114 virus using immunoblot analysis and RT assay [25]. In addition, we found that two canine live attenuated vaccines (Vaccines C/f8 and C/f9) produced using ‘non-feline’ cell lines (Table 1) were contaminated with infectious RD-114 virus (Table 2) [49]. The infectious titers of RD-114 virus in contaminated vaccines were 180 and 10,000 TCID50/dose, respectively and the copy numbers of RD-114 viral RNAs were 2.1×108 and 5.0×108 copies/dose respectively (Table 2) [49]. The NVAL also confirmed these findings independently [29].

Table 1

List of cell lines for live attenuated vaccines purchased in Japan

Virus	Vaccine IDa
	C/a1	C/a2	C/b3	C/c4	C/c5	C/d6	C/e7	C/f8	C/f9
Canine distemper virus	Ab	A	S	S	Unknown	S	C	A	S
Canine adenovirus type 2	P	P	C	C	Unknown	n.a.	C	C	C
Canine parvovirus	A	A	C	F (CRFK)	Unknown	F	C	M	M
Canine parainfluenza virus	F (CRFK)	F (CRFK)	C	C	Unknown	n.a.	C	n.a.	S
Canine coronavirus	F (CRFK)	F (CRFK)	F	n.a.	Unknown	n.a.	F	n.a.	n.a.

This information is indicated in each product catalog and on theNVAL homepage (http://www.nval.go.jp/asp/asp_dbDR_idx.asp)

A, avian cell line; C, canine cell line; F, feline cell line; M, mink cell line; P, porcine cell line; S, simian cell line; n.a., not applicable

aEach vaccine sampled was assigned an anonymized code (F/1, C/1, etc.). The first letter before the slash indicates the species (i.e., F/ for cats and C/ for dogs). The lower case letter after the slash indicates the manufacturer. The number indicates the specific type of vaccine

bSpecies of the cell line used for vaccine production

Table 2

Contamination of live attenuated vaccines purchased in Japan with RD-114 virus

Vaccine IDa	RNA copy number/doseb	TCID50/dosec
C/a1	4.5×107	1,800
C/a2	2.8-9.7×107	1,000
C/b3	8.3×106	1.8
C/c4	BGL	n.t.
C/c5	BGL	n.t.
C/d6	BGL	n.t.
C/e7	BGL	n.t.
C/f8	2.1×108	180
C/f9	5.0×108	10,000
F/g1	BGL	1.8
F/h2	BGL	n.t.
F/d3	BGL	n.t.

BGL, background level; n.t., not tested

aCodes used to anonymize the vaccines used. Details of codes are described in Table 1

bCopy numbers of RD-114 viral RNA were measured by real-time RT-PCR

cInfectious titers of RD-114 virus were measured by LacZ marker rescue assay and expressed as 50 % tissue culture infectious dose (TCID50)

Possible contamination routes of RD-114 virus in live attenuated vaccines

Several feline cell lines such as CRFK cells, MCC cells and FER cells produce infectious RD-114 virus [2, 7, 33, 38, 48]. Therefore, if the vaccine strains of feline and canine viruses are propagated in RD-114 virus-producing feline cell lines, RD-114 virus contaminates live attenuated vaccines (Fig. 1). Moreover, RD-114 virus infects and proliferates efficiently in human, canine and mink cell lines [2, 36, 49, 51]. Therefore, when seed stock viruses are contaminated with infectious RD-114 virus and the vaccines are produced using RD-114 virus-permissive cell lines, RD-114 virus may contaminate live attenuated vaccines, irrespective of the species origin of the cell lines (Fig. 1). Actually, as mentioned above, two canine live attenuated vaccines (Vaccines C/f8 and C/f9) produced using ‘non-feline’ cell lines were contaminated with infectious RD-114 virus (Table 2) [49]. These vaccines contained an attenuated canine parvovirus type 2 (CPV-2) (Table 1) and many CPV-2s have been attenuated using CRFK cells [26]. When we examined CPV-2 stock viruses in an assay laboratory of a Japanese vaccine company for the presence of infectious RD-114 virus, seven out of eighteen CPV-2 vaccine stocks were contaminated with infectious RD-114 virus [50].

Fig. 1

Possible contamination routes of infectious RD-114 virus in live attenuated vaccines. Infectious RD-114 virus may contaminate vaccines manufactured using RD-114 virus-producing cells. Several canine parvoviruses are isolated and attenuated using CRFK cells producing infectious RD-114 virus. Therefore, infectious RD-114 virus may contaminate canine live attenuated vaccines produced using ‘non-feline’ cell lines when seed stock viruses were contaminated with infectious RD-114 virus

Potential risks of infection with RD-114 virus in dogs

RD-114 virus efficiently infects canine cell lines and primary cells [36, 51]. It is important to identify viral receptor(s) in predicting viral tropisms and pathogenicity. In human cell lines, it is found that the receptor for RD-114 virus is a sodium-dependent neutral amino acid transporter, termed ASCT [35, 40]. Humans have two types of ASCT molecules, termed ASCT1 and ASCT2 [1, 46]. The homology between ASCT1 and ASCT2 is about 57 % [1, 46]. Both human ASCT1 and ASCT2 function as RD-114 receptor, but the virus utilizes ASCT2 more efficiently than ASCT1 [20]. In humans, ASCT1 is ubiquitously expressed in tissues [1], whereas the expression of ASCT2 is limited in various tissues, and the expression level of ASCT2 also varies among tissues [13, 46]. Recently, we identified canine ASCT1 and ASCT2 as RD-114 virus receptors [50]. RD-114 may infect a variety of tissues in dogs, although the distribution of ASCT1 and ASCT2 in dogs is unknown at present. Actually, in a previous study, X-linked severe combined immunodeficiency was corrected in dogs by intravenous injection of concentrated RD-114-pseudtyped retrovirus vector encoding the interleukin-2 receptor γ chain [44]. These data indicate that RD-114 virus can infect bone marrow cells in dogs. However, Narushima and coworkers [28] at NVAL reported that RD-114 proviral DNA was not found in dogs inoculated with RD-114 virus subcutaneously. Unfortunately, they only investigated RD-114 provirus in quite limited tissues (lymph nodes, spleen and bone marrow) and peripheral blood, and the sensitivity of the one-step PCR to detect RD-114 proviral DNA was obscure. Recently, we also investigated whether RD-114 virus infects dogs by experimental infection. When four dogs were inoculated with high titer of RD-114 virus, RD-114 proviral DNA was detected in blood cells, mesenteric lymph nodes, spleens and testes (Yoshikawa et al., unpublished). In addition, anti-RD-114 antibodies and neutralizing antibodies were detected in the inoculated dogs (Yoshikawa et al., unpublished).

In human cells, human ASCT2 is a functional receptor for pathogenic retroviruses, such as simian retrovirus (SRV) 1, 2, 3, 4 and 5, avian reticuloendotheriosis virus, and duck spleen necrosis virus [35]. It is known that SRV-1, SRV-2 and SRV-3 induce a fatal immunodeficiency in some macaque species (Celebes and rhesus macaques) [21-23]. In a previous study, we confirmed that both SRV-2 and SRV-3 can utilize canine ASCT2 as a receptor ([51], unpublished data). Intriguingly, RD-114 virus has an immunosuppressive domain in transmembrane envelope protein and the amino acid sequence (LQNRRGLDLLTAEQGGI) of the domain is identical with that of SRV-3 [5, 8]. Therefore, RD-114 virus may induce immunosuppression as well as proliferative diseases such as leukemia/lymphoma, if it adapts to replicate efficiently in dogs.

Concluding remarks

Japanese and UK research groups and the NVAL confirmed that several feline and canine vaccines were contaminated with infectious RD-114 virus. This was the first incidence of contamination of ‘infectious’ ERVs in live attenuated vaccines. Quite importantly, RD-114 virus grew efficiently in cells of dogs which are vaccinees. In future studies, it is necessary to examine the expression profiles of canine ASCT1 and ASCT2 in dogs and determine the principal target of RD-114 virus by experimental infection with high doses of RD-114 virus. Because RD-114 virus does not have any oncogenes, RD-114 virus does not induce acute/subacute proliferative diseases such as fibrosarcoma in dogs. Therefore, after experimental infection of dogs with RD-114 virus, it is necessary to monitor infected dogs for a long period. Even if RD-114 virus does not proliferate in dogs after experimental infection, we cannot dismiss the risk of infection with RD-114 virus in dogs completely. Canine attenuated vaccines are inoculated in several million dogs per year around the world, and RD-114 virus may mutate and acquire more infectivity/productivity in dogs. Therefore, although the risks posed by RD-114 virus are still unclear at present, it is desirable to develop the means to produce RD-114 virus-free vaccines and exclude RD-114 virus-contaminated vaccines.

Below is the link to the electronic supplementary material.

Supplementary material 1 (PPT 877 kb)