key: cord-0005907-5lwv4aob
authors: Roelke-Parker, Melody E.; Munson, Linda; Packer, Craig; Kock, Richard; Cleaveland, Sarah; Carpenter, Margaret; O'Brien, Stephen J.; Pospischil, Andreas; Hofmann-Lehmann, Regina; Lutz, Hans; Mwamengele, George L. M.; Mgasa, M. N.; Machange, G. A.; Summers, Brian A.; Appel, Max J. G.
title: A canine distemper virus epidemic in Serengeti lions (Panthera leo)
date: 1996
journal: Nature
DOI: 10.1038/379441a0
sha: 7bfae66271ae1863c4da0006b9c7c40475cdbf31
doc_id: 5907
cord_uid: 5lwv4aob

CANINE distemper virus (CDV) is thought to have caused several fatal epidemics in canids within the Serengeti–Mara ecosystem of East Africa, affecting silver-backed jackals (Canis mesomelas) and bat-eared foxes (Otocyon megalotis) in 1978 (ref. 1), and African wild dogs (Lycaon pictus) in 1991 (refs 2, 3). The large, closely monitored Serengeti lion population(4,5) was not affected in these epidemics. However, an epidemic caused by a morbillivirus closely related to CDV emerged abruptly in the lion population of the Serengeti National Park, Tanzania, in early 1994, resulting in fatal neurological disease characterized by grand mal seizures and myoclonus; the lions that died had encephalitis and pneumonia. Here we report the identification of CDV from these lions, and the close phylogenetic relationship between CDV isolates from lions and domestic dogs. By August 1994, 85% of the Serengeti lion population had anti-CDV antibodies, and the epidemic spread north to lions in the Maasai Mara National reserve, Kenya, and uncounted hyaenas, bat-eared foxes, and leopards were also affected.

CANINE distemper virus (CDV) is thought to have caused several fatal epidemics in canids within the Serengeti-Mara ecosystem of East Africa, affecting silver-backed jackals (Canis mesomelas) and bat-eared foxes (Otocyon mega/otis) in 1978 (ref. 1) , and African wild dogs (Lycaon pictus) in 1991 (refs 2, 3) . The large, closely monitored Serengeti lion population 4 ' 5 was not affected in these epidemics. However, an epidemic caused by a morbillivirus closely related to CDV emerged abruptly in the lion population of the Serengeti National Park, Tanzania, in early 1994, resulting in fatal neurological disease characterized by grand mal seizures and myoclonus; the lions that died had encephalitis and pneumonia. Here we report the identification of CDV from these lions, and the close phylogenetic relationship between CDV isolates from lions and domestic dogs. By August 1994, 85% of the Serengeti lion population had anti-CDV antibodies, and the epidemic spread north to lions in the Maasai Mara National reserve, Kenya, and uncounted hyaenas, bat-eared foxes, and leopards were also affected.

In early 1994, six lions in the Serengeti National Park, Tanzania were observed with grand mal seizures, and three other lions developed facial and forelimb myoclonus (recurrent twitching). Additional lions were noted to be disoriented, ataxic and profoundly depressed. Between January and March 1994, 11 lion carcasses were found, representing a dramatic increase in mortality from previous years and indicating that a serious epidemic was emerging.

To investigate the epidemic, tissue and serum samples were obtained from 23lions that died or were killed in a moribund state, 13 live lions with obvious signs of disease, and 72 apparently healthy, anaesthetized lions. Sera from 111 healthy lions sampled between 1984 and 1994 (refs 6-8) were also analysed to track previous exposure of this population to viruses. Tissues from 19 ##To whom correspondence should be addressed. dead lions, examined by histopathology, had either encephalitis, interstitial pneumonia, and/or lymphocytic depletion in lymph nodes and spleen ( Fig. 1 ; Table 1 ). Rare multinucleated syncytia and intranuclear and/or intracytoplasmic viral inclusions characteristic of morbilliviral infection 9 were also found in these lions (Table 1) . Because these lesions were seen in zoo cats in the United States in the 1991 and 1992 canine distemper epidemics 10 , we used monoclonal 11 and polyclonal CDV antibodies to confirm that CDV nucleocapsid antigens were present in affected tissues. We then tested all available lion sera for neutralizing antibody titres to CDV 10 • 12 , and found that 63 of the 72 apparently healthy lions and 8 of the 11 sick lions sampled in 1994 had CDV titres (Fig. 2a) . We isolated CDV from the cerebrospinal fluid of one lion cub with grand mal seizures that subsequently died with CDV encephalitis 13 • The monoclonal antibody-binding pattern of this virus was compared to binding patterns of viruses isolated from: a bat-eared fox (0. megalotus), a spotted hyaena (Crocuta crocuta), and a domestic dog that died during the 1994 Serengeti epidemic; with CDV isolated from a lion during the 1992 California epidemic; with CDV isolated from a raccoon (Procyon lotor) and fox (Vulpes vulpes); and with a virulent CDV (A75- 17, 1975 ) and two strains of attenuated CDV from domestic dogs (Rockborn 1958 and Onderstepoort 1948) . Monoclonal antibodies were donated by C. Orvell (Huddinge, Sweden), and were directed against viral nucleoprotein (N), polymerase (P), fusion glycoprotein (F), and haemagglutinin glycoprotein (H). Lymphocytes infected with the Serengeti lion virus bound the same group of CDV monoclonal antibodies as cells infected with these other viruses, suggesting that the lion morbillivirus was CDV. To characterize further the lion morbillivirus, the genomic sequences of the CDV P gene, a conserved region of the virus, were amplified from buffy-coat

FIG. 1 Characteristic histopathological changes and intralesional immunoreactive viral nucleocapsid antigens in tissues of African lions infected with canine distemper virus. a, Haematoxylin-and eosin-stained section of the brain from a lion with CDV encephalitis (x350). The dentate gyrus in the hippocampus has neuronal loss, dense gliosis and occasional multinucleated giant cells (arrows). b, Immunohistochemical identification by monoclonal antibody N3.991 against COY-nucleocapsid protein in a similar area to a. Several neurons contain CDV antigen in the cytoplasm (some indicated by arrows) ( x 350). c, Immunohistochemical identification ofCDV nucleocapsid in a parahippocampal gyrus ( x 350) . Many neurons are strongly positive for CDV antigen in perikarya and neurites. d, Immunohistochemical identification of CDV nucleocapsid proteins in lung with interstitial pneumonia, characterized by type II pneumocyte hyperplasia and alveolar septal thickening ( x 350). METHODS. Tissues were fixed in 10% formalin, embedded in paraffin, sectioned at 5-7 ~m and stained with haematoxylin and eosin. Immunohistochemical procedures were applied to sections of formalin-fixed, paraffin-embedded tissues that had typical CDV lesions on stained sections to demonstrate intralesional viral antigens 10 . Paraffin-embedded tissues had the paraffin removed, were treated to remove endogenous peroxidase, then incubated with a mouse monoclonal antibody to a CDV-N protein (MAb N3.991) 11 , or with rabbit primary polyclonal antibody raised with the Rockborn strain of CDV. A commercial avidin-biotin kit was used to identify sites of monoclonal antibody binding to tissues, and a commercial peroxidaseanti peroxidase kit was used for polyclonal antibodies, then tissue sections were counterstained with Gill's haematoxylin, dehydrated and mounted with Permount. Negative controls were duplicate sections stained using a monoclonal antibody for influenza virus replacing CDV monoclonal antibodies, and positive controls were brain sections from a confirmed case of CDV in a domestic dog.

lymphocytes of two lions with neurological signs. Phylogenetic relationships between the P-gene sequences of the two RT-PCRderived lion CDV and the P-gene sequences of other morbilliviruses were then examined. These analyses ( Fig. 3 ) indicated that Serengeti lion CDV virus is closely related to the Onderstepoort strain of canine distemper virus isolated from a domestic dog in South Africa. The statistics of the epidemic are given in Table 1 .

Later in 1994, CDV-infected lions were identified in the northern and western areas of the Serengeti National Park and in the Maasai Mara National Reserve in Kenya (Fig. 2b) . Of 54 apparently healthy lions sampled from October 1994 to March 1995 in the Mara and neighbouring areas of Kenya, 23 had high serum titres of CDV antibodies, suggesting recent exposure.

During 1994,39 CDV-associated lion deaths were documented, but this is probably an underestimation of true mortality statistics because most of the Serengeti lion population outside the Seronera study area is not under close observation. The overall lion population in the Serengeti ecosystem, estimated at 3,000 before the outbreak, is now estimated at 2,000. In 1994 and 1995, CDV-related deaths were also confirmed in seven spotted hyaenas (C. crocuta) by pathology and virology. Because CDV is historically not pathogenic in lions or hyaenas 9 · 1 \ the emerging biotype of CDV has apparently extended its host range. During the epidemic, CDVinduced disease also was confirmed in two bat-eared foxes, and CDV-like neurological disease was observed in a common jackal (Canis aureus) and two silver-backed jackals, indicating that the Serengeti CDV biotype conserves its pathogenicity for canids.

The magnitude of this epidemic may be explained in part by the lion population being immunologically naive to CDV when it was introduced in 1994. All hut one of the 34 lions sampled between 1990 and 1993 were seronegative, and the seropositive lion was Fig. 2 a, Temporal patterns of mortality and seroconversion in the longterm study population of lions in the Serengeti National Park. Top, number of lions resident in the long-terrn study area, measured monthly frorn August 197 4 to February 1995 (refs 4, 5) . Middle, prevalence of anticanine distemper virus (CDV) antibodies 10 in Serengeti lions from 1990 to 1994. The single seropositive lion in 1993 was sampled in December. Bottom, age-specific mortality (proportion dying per year) in female lions in the study population. The solid line represents the combined mortality rate from 1966 to 1993; the broken line represents mortality in 1994. Male mortality was comparable to female mortality, but was not included here because male disappearances could also be caused by normal, nondisease-related dispersals 4 · 5 • b, Chronology and geographic distribution of CDV disease in the Serengeti ecosystem. The months in boxes are dates in 1994 when lions were first observed with neurological disease. The proportions of CDV-seropositive domestic dogs in regions surrounding the Serengeti National Park are represented by pie charts for 1992 to 1994 (n, number of dogs tested). The domestic dog that was confirmed to die from canine distemper disease in September 1994 was located in the Ngorongoro Conservation Area. Endemic infection was confirmed in the Serengeti district where CDV seropositivity was detected in pups (<12 months old) in each year of the study (1992) (1993) (1994) . This district also has one ofthe highest dog population densities. In the Loliondo and Ngorongoro areas, seropositivity was detected in pups in 1994 (3/7), but not in 1992 (0/6) or 1993 (0/17). METHODS. Sera were tested against the Onderstepoort strain of CDV adapted to Vero cells as described previouslyi 2 • Sera were also tested against CDV isolated from a lion during the California epidemic (A92-27/ 20). Log titres of 1.0 or greater were considered positive. b sampled in December of 1993 (Fig. 2a) . Although this abrupt seroconversion indicates a recent exposure of these lions to CDV, analysis of sera from 77 lions in the region from 1984 to 1989 disclosed that 22 lions had antibodies to CDV (none of these previously tested lions was alive during the 1994 epidemic), indicating that lions in the Serengeti ecosystem had previously encountered CDV without an increase in disease-related mortality. To determine if the high lion mortality during the 1994 CDV epidemic was due to co-infection with another viral pathogen, we compared the prevalences of serum antibody titres to feline immunodeficiency virus (FeiV) 15 ,., FeCV, P > 0.14) fails to support a role for these viruses as necessary cofactors in CDV morbidity. CDV infections in equatorial African wildlife usually occur as periodic epidemics because environmental factors limit viral persistence outside susceptible carnivore hosts, and these hosts usually succumb or rid themselves ofvirus 9 • CDV persists in dense populations of domestic dogs because pups provide a constant reserve of susceptible hosts. The source of CDV in the Serengeti epidemic was probably the domestic dogs of the local villages adjacent to the National Park. CDV seroprevalance increased in Oligonucleotide primers were synthesized based on conserved regions of the CDV phosphoprotein (P) gene as described previously 18 . PCR product fragments were cloned and sequenced, and the sequences aligned with P-gene sequences from other morbilliviruses. Derived P-gene sequences from two lions were 99% identical. A phylogenetic analysis using cladistic, phenetic and maximum-likelihood methods revealed a close relationship between the lion morbillivirus and the Onderstepoort strain of CDV, the sequences having 95% nucleotide identity. PAUP maximum parsimony tree 19 (left), with branch lengths equal to the number of nucleotide substitutions. Numbers shown on branches are 10 • 12 . Buffy-coat isolation for viral cultures and molecular analyses were derived from 20 ml blood taken from anaesthetized lions.

t Methods are described in Fig. 3 legend. these dogs between 1991 and 1993, preceding the lion epidemic (Fig. 2b ) , and CDV encephalitis was confirmed by histopathology in a domestic dog in the Ngorongo Crater region in 1994. The precise route of CDV transmission to the lions is unclear, because direct dog-to-lion contact is unlikely for most of the Serengeti ecosystem. A more probable route is by spotted hyaenas, which range among human dwellings and travel long distances within the Park 17 • Nomadic lions could also contribute to viral dissemination. High densities of these susceptible carnivores at kill sites would then provide an ideal environment for CDV amplification and transmission.

Most of the lion deaths in the Serengeti National Park occurred between January and September 1994, and mortality rates have subsequently returned to previous levels (Fig. 2a) . Although this CDV epidemic claimed approximately 30% of the Serengeti and Mara lions, the impact on other carnivore species is unknown. Less dense populations of endangered species, such as cheetahs or wild dogs, are a clear cause for concern if exposed to a virulent pathogen such as this putative new biotype of CDV. The Serengeti is surrounded by approximately 30,000 domestic dogs, most of which are not vaccinated against canid pathogens (including CDV), representing a large reservoir for carnivore diseases. The CDV epidemic clearly emphasizes the need for continued disease surveillance to monitor infectious diseases in valuable wildlife resources, and for initiating vaccination programs for domestic animals in contact with wildlife. D KIDNEY stones (nephrolithiasis), which affect 12% of males and 5% of females in the western world, are familial in 45% of patients 1 • 2 and are most commonly associated with hyper-calciuria\ Three disorders of hypercalciuric nephrolithiasis (Dent's disease\ X-linked recessive nephrolithiasis (XRN) 4 , and X-linked recessive hypophosphataemic rickets (XLRH) 5 ) have been mapped to Xp11.22 (refs 5-7) . A microdeletion 6 in one Dent's disease kindred allowed the identification of a candidate gene, CLCNS (refs 8,9) which encodes a putative renal chloride channel. Here we report the investigation of 11 kindreds with these renal tubular disorders for CLCNS abnormalities; this identified three nonsense, four missense and two donor splice site ttTo whom correspondence should be addressed.

NATURE · VOL 379 · 1 FEBRUARY 1996

mutations, together with one intragenic deletion and one microdeletion encompassing the entire gene. Heterologous expression of wild-type CLCNS in Xenopus oocytes yielded outwardly rectifYing chloride currents, which were either abolished or markedly reduced by the mutations. The common aetiology for Dent's disease, XRN and XLRH indicates that CLCNS may be involved in other renal tubular disorders associated with kidney stones. Analysis of CLCN5 reverse transcriptase-polymerase chain reaction (RT-PCR) products encompassing the entire 2238-bp coding sequence 9 from probands of 11 kindreds with Dent's disease, XRN and XLRH, revealed different CLCN5 mutations (Table 1 and Fig. 1 ). Each mutation was confirmed and demonstrated to cosegregate with the disease by using genomic DNA together with the appropriate PCR primers and restriction enzymes, or by sequence-specific oligonucleotide (SSO) probe analysis (Table 1 and Fig. 2 ). In addition, the absence of these CLCN5 abnormalities in 110 alleles from 69 (28 males and 41 females) unrelated, normal individuals established that they were not common polymorphisms. CLCN5 belongs to a family of voltage-gated chloride-channel genes ( CLCNJ to CLCN5 and CLCN-Ka and Kb), which encode proteins (CLC-1 to CLC-5, CLC-Ka and CLC-Kb) that have about 12 transmembrane domains 10 • These chloride channels are important for the control of membrane excitability, transepithelial transport, and possibly regulation of cell volume 10 • However, mutations have been identified previously in only CLCNJ, which is expressed in muscle 11 and is associated with myotonia congenita 12 -14 • Thus, to assess further the functions of CLC-5 and its mutations, we performed heterologous expression studies in Xenopus oocytes. Expression of the human wild-type (WT) CLC-5 reproducibly yielded strongly outwardly rectifying, essentially time-independent currents (Fig. 3) . Ion substitution experiments indicated that these were carried by anions, with a chloride > iodide conductance sequence (Fig. 3b,c) , as is the case with other chloride channels (CLC-0, CLC-1 and CLC-2) of this familyl 2 • 15 • 16 • However, these human CLC-5 currents, which were indistinguishable from those of rat CLC-5 (ref. 17) , differed from the others in being strongly outwardly rectifying and in being observed only at potentials more positive than +10m V. Although positive potentials of even +40 m V have been observed in apical membranes of some actively transporting epithelia 18 , we are not aware of renal cells where these voltages would be reached in vivo. CLC channels are known to function as multimeric complexes, which are most likely to be tetramers 12 , and it seems possible that CLC-5 forms heterooligomers with as yet unknown subunits in situ that may render the channels open at a more physiological voltage. Our functional expression of CLC-5 provides a valuable means of investigating this further and in assessing the functional effects of the CLC-5 mutations.

Expression of the four missense and three nonsense mutations (Table 1) , and the in-frame deletion of the predicted transmembrane domain D2 (Fig. 1) , abolished the CLC-5 currents or reduced them (S244L and S520P) to levels where they were difficult to distinguish from endogenous oocyte currents (Fig.  3d,f) . As a control, we also expressed more conservative changes at codons 244 and 520; S244T, S520T, S244A and S520A all elicited chloride currents that were similar to the WT (data not shown). Thus, the mutations found in the hypercalciuric nephrolithiasis pedigrees grossly and specifically affect CLC-5 function, thereby strongly suggesting a causal role in the disease. Dent's disease, which is characterized by low-molecular-weight proteinuria (LMWP), hypercalciuria, nephrocalcinosis, nephrolithiasis, rickets and eventual renal failure 3 · 6 , has phenotypic similarities to XRN 4 · 7 and XLRH 5 (Fig. 2) . However, there are important differences, as rickets is absent in XRN, and nephrocalcinosis and moderate renal failure are more notable in XLRH. A correlation between these phenotypic differences, the different mutations (Table 1 and Fig. 1 ) and the resulting abnormal chloride currents (Fig. 3) could not be established. Thus, Dent's disease was found to be associated with the mutations W279X,

The Tertiaty Record of Rodents in North America

Evolutionaty Relationships among Rodents. A Multidisciplinaty Analysis

An lnsightinto the HistotyofUfe and of the Earth

Virus Infections in Carnivores

Virus Infections of Carnivores

Phylogenetic Analysis Using Parsimony (PAUP) (Illinois National History SuNey

Phylogeny Inference Package (PHYLIP), Version

East and H. Hoeffer for providing two hyaena carcasses. This work was supported in part by the Messerli Foundation of Zurich

Hammersmith Hospital

Italy 11 L'Hopital de Montreal pour Enfants, 2300 rue Tupper, A717 Montreal, Quebec H3HIP3, Canada #Department of Paediatric Nephrology, Guy's Hospital, StThomas's Street, London SE1 9RT, UK 1? Department of Nephrology, The Middlesex Hospital