Reprinted with permission from the Encyclopedia of Virology Plus CD-ROM. (Edited by Robert G. Webster and Allan Granoff.) © 1995 Academic Press Ltd. All Rights Reserved. This article may NOT be copied/reposted in whole or part without express permission from Academic Press.
MARBURG AND EBOLA VIRUSES
Hans-Dieter Klenk, Werner Slenczka and Heinz Feldmann
Institut für Virologie Philipps Universitãt, Marburg
Marburg, Germany
History
Marburg and Ebola viruses both cause severe
hemorrhagic
fevers. Marburg virus was first recognized in laboratory workers in Marburg, Germany, and Belgrade, Yugoslavia, in 1967. These workers had been exposed to tissues and blood from African green monkeys (
Cercopithecus aethiops
) imported from Uganda. There were 25 primary cases and six secondary cases in the outbreak. Seven of the primary cases died. Since then, sporadic, virologically confirmed Marburg disease cases have occurred in Zimbabwe, South Africa and Kenya. Ebola virus first emerged in two major disease outbreaks which occurred almost simultaneously in Zaire and Sudan in 1976. Over 500 cases were reported, with mortality rates of 88% in Zaire and 53% in Sudan. A single case was confirmed by virus isolation in Zaire in 1977 and 1979, Ebola
hemorrhagic
fever occurred again in Sudan at the site that was involved in 1976. Besides these episodes documented by virus isolation, two more fatal and two nonfatal cases have been reported. No association with monkeys could be attributed to any of the Ebola outbreaks. A third filovirus, serologically related to Ebola virus was isolated from cynomolgus monkeys (
Macaca fascicularis
) which originated in the Philippines. According to the evidence available to date, Reston virus is infectious to humans but does not seem to cause serious human disease. Recently a filovirus was isolated from cynomolgus monkeys imported from the Philippines into Italy. The isolated agent is likely to be Reston virus or a filovirus closely related to Reston virus. Human infections were not observed during this outbreak.
Taxonomy and Classification
Marburg, both subtypes of Ebola (Ebola-Zaire and Ebola-Sudan) and Reston viruses are members of a new family of negative-stranded RNA viruses, the
Filoviridae
. The filoviruses are similar in morphology, density and polyacrylamide gel electrophoresis profile, and there is a serological relationship between Ebola and Reston viruses. Originally classified as rhabdoviruses, they appear to be more closely related to paramyxoviruses on the basis of recent genome sequence data. However, filoviruses are sufficiently distinct from the other nonsegmented negative-stranded RNA viruses to warrant taxonomic status as a separate virus family. A classification within the family
Filoviridae
has not yet been proposed. However, based on the data available to date, a separation into two distinct groups of viruses. Marburg-like and Ebola-like viruses, is obvious.
Properties of Virion
Filovirus particles are morphologically similar to rhabdovirus particles but much longer. By
electron microscopy, virions are pleomorphic, appearing as long filamentous, sometimes branched forms, or as 'U'-shaped, 'b'-shaped or circular forms. The particles vary greatly in length (up to 14 000 nm), but have a uniform diameter of about 80 nm. Virions purified by ratezonal gradient centrifugation are bacilliform in outline and show an average length associated with peak infectivity of 665 nm for Marburg and 805 nm for Ebola virus. Except for the difference in length, filoviruses seem to be very similar in morphology. Virions contain a nucleocapsid consisting of a dark, central space (20 nm in diameter) surrounded by a helical capsid (50 nm in diameter) bearing cross-striations with a periodicity of approximately 5 nm. Within the nucleocapsid is an axial channel of 10-15 nm. The helical nucleocapsid is surrounded by a lipoprotein unit membrane envelope derived from the host cell
plasma membrane. Spikes of approximately 7 nm length, spaced at approximately 10 nm intervals are visible on the virion surface.
Physical Properties
Virus particles have a molecular weight of approximately 3-6
x 10
8
and a density in potassium tartrate of 1.14 g/cm
3
. Uniform, bacilliform particles have a sedimentation coefficient of 1300-1400S, whereas larger particles have a higher sedimentation coefficient. Virus infectivity is quite stable at room temperature. Inactivation can be further performed by UV and
gamma irradiation, 1% formalin,
beta-propiolactone, and brief exposure to phenolic disinfectants and lipid solvents, like deoxycholate and ether.
Properties of Genome
The genome of filoviruses consists of a molecule of linear, nonsegmented, negative-strand RNA which is noninfectious, not polyadenylated and complementary to viral-specific
messenger RNA. The genome amounts to 1.1% of the total virion weight and the sedimentation coefficient is 46S (0.15 M NaCl, pH 7.4). Filovirus genomes are approximately 19 kb in length and very rich in adenosine and uridine residues. Genomes show a linear gene arrangement in the order 3' untranslated region -- nucleoprotein (NP) -- viral structural protein --
(VP) 35 -- VP40 -- glycoprotein (GP) -- VP30 -- VP24 -- polymerase (L) -- 5' untranslated region. Genes are flanked at their 3' and 5' ends by highly conserved transcriptional start and termination signal sequences, respectively, which all contain the pentamer 3'-UAAUU-5'. They are either separated by intergenic sequences variable in length and nucleotide composition or by gene overlaps which are limited to the conserved transcription signals centered around the common pentanucleotide sequence. Ebola-like viruses show three overlaps that alternate with intergenetic sequences while the Marburg virus genome contains a single overlap at a different position. Extragenic sequences are present at the 3' and 5' ends of filovirus genomes which are complementary at their very extremities. These sequences are comparable to those found in genomes of other nonsegmented negative-strand RNA viruses and are known as leader sequences. However, neither (+) nor (-) leader RNAs have been detected in filovirus-infected cells.
Properties of Viral Proteins
Virions contain at least seven proteins with presumed identical functions for the different viruses. Nonstructural proteins have not yet been identified. The electrophoretic mobility patterns of the structural proteins are characteristic for Marburg-like viruses on the one hand and Ebola-like viruses on the other. Four proteins are associated with the viral ribonucleoprotein complex (the nucleoprotein, the polymerase and the viral structural proteins 30 and 35), the single glycoprotein is inserted in the envelope, and the location of two proteins (viral structural proteins 40 and 24) has not been determined exactly, but they seem to be membrane-associated. The L protein is the largest protein and, like other L proteins of nonsegmented negative-stranded RNA viruses, is the virion-associated RNA-dependent RNA polymerase. Its size, as calculated from the deduced amino acid sequence of the Marburg virus (Musoke strain) L gene, is 267 kD. The glycoprotein (GP) (Marburg virus 170 kD, Ebola virus 125 kD) is an integral membrane protein and forms the surface projections of the virion. It is reasonable to assume that the glycoprotein is the mediator of virus entry into the cell. Functional sites for receptor recognition and binding and perhaps for fusion should be located on this protein. The glycoprotein of Marburg virus is a type I transmembrane protein and inserted in the lipid membrane as a homotrimer. The carbohydrate structures of this highly glycosylated protein account for more than 50% of the molecular weight of the mature protein. They include oligomannosidic and hybrid type N-glycans as well as bi-, tri and tetra-antennary complex species, and high amounts of neutral mucin-type O-glycans. Lectin binding studies on the carbohydrates of Ebola and Reston virus glycoproteins revealed N- and O-linked oligosaccharides with similar structures as found for Marburg virus. In contrast to Marburg virus carbohydrate structures, which totally lack terminally sialic acid residues, these structures are terminally
sialylated. The major nucleoprotein (NP) (Marburg virus 96 kD, Ebola virus 104 kD) and VP30 (Marburg virus 28 kD, Ebola virus 30 kD), which may represent a minor nucleoprotein, seem to be intimately associated with virion RNP. VP35 (Marburg virus 32 kD, Ebola virus 35 kD) is loosely associated with the RNP and seems to be a component of the transcriptase complex analogous to the P protein of paramyxoviruses and the NS (P) protein of rhabdo-viruses. The functions of VP40 (Marburg virus 38 kD, Ebola virus 40 kD) and VP24 (Marburg virus 24 kD, Ebola virus 24 kD) are not known, but they are probably membrane components. VP40 is the most prominent viral structural protein, and its weak association with the ribonucleoprotein complex suggests a role as the matrix protein of filoviruses.
Replication
The mechanism of virus entry into host cells is unknown, but it is reasonable to assume that the glycoprotein as the only known transmembrane protein of the virion particle mediates the adsorption and the penetration process. The genetics of filoviruses are probably similar to those of the other two nonsegmented negative-stranded RNA virus families,
Paramyxoviridae
and
Rhabdoviridae
. Transcription and
replication
take place in the cytoplasm of infected cells. The 3' untranslated region of the genome probably provides the encapsidation site for the nucleoprotein as well as the entry site for the polymerase. Filovirus genomes are transcribed to yield monocistronic subgenomic
messenger RNA
species which are complementary to viral genomic RNA. The 5' ends of the subgenomic RNAs start at the transcription start signal sequence, and the 3' ends carry a poly(A) tail generated by the polymerase at a run of uridine residues located at the 5' ends of all transcription termination signal sequences. Transcription efficiency might be influenced by gene order, formation of secondary structures at the 3' ends of the genes, secondary structure formation within the intergenic sequences, overlapping genes and presence of duplicated termination sites.
Replication
of the genome is mediated by the synthesis of a full-length complementary antigenome ((+) sense) which then serves as the template for the synthesis of progeny negative-strand RNA anticomplementary to the parental template RNA. The complementarity of the genome extremities suggests a single identical encapsidation site on the genome and antigenome and an identical entry site for the polymerase for both the transcription as well as the
replication
mode. The cytoplasm of virus-infected cells contains prominent
inclusion bodies
consisting of viral nucleocapsid. As infection proceeds, they grow and become highly structured. Budding of completed virus particles takes place at
cell membrane
sites which are altered by insertion of the viral glycoprotein and alignment of viral membrane-associated proteins as well as of preformed nucleocapsids.
Geographic Distribution
Marburg and Ebola viruses are indigenous to Africa. The virus responsible for the 1967 outbreak of Marburg disease originated from Uganda, whereas the cases observed subsequently occurred in Zimbabwe, South Africa and Kenya. Serological studies suggest that Ebola or related viruses are endemic in Zaire, Sudan, the Central African Republic, Gabon, Nigeria, Ivory Coast, Liberia, Cameroon and Kenya. The geographic range of Ebola strains may extend to other African countries, for which adequate survey is lacking. In contrast to Marburg and Ebola viruses, Reston virus was isolated from animals originating from the Philippines. Thus, it appears the filoviruses are not confined to Africa.
Host Range and Virus Distribution
The natural reservoirs for human infections with Marburg and both subtypes of Ebola viruses and the natural source of Reston virus are unknown. Human infections with Reston virus have been documented during the 1989 epizootic. Experimental hosts include monkeys for which infection with Marburg and Ebola-Zaire virus are usually lethal, whereas some animals survive Ebola-Sudan virus infection. Reston virus infection of monkeys showed that this filovirus is less pathogenic for primates than Marburg virus and both subtypes of Ebola virus. Guinea pigs show febrile responses 4-10 days after inoculation with Marburg and Ebola viruses. However, none of these viruses kills guinea pigs consistently on primary inoculation. Ebola-Zaire is usually pathogenic for
newborn
mice after i.c. and i.p. inoculation. For growth in cell culture, primary monkey kidney cells and monkey kidney cell lines, primarily Vero cells, are used. Marburg virus grows also in human endothelia maintained as primary cell cultures or as organ cultures.
Evolution
Besides morphological and biochemical similarities, all nonsegmented negative-strand RNA viruses share several features in their mechanisms of transcription and
replication:
(1) similar genome organization, (2) complementarity of the genome extremities, (3) homologous sequences in the 3' untranslated region, (4) conserved transcriptional signals, (5) interruption of genes by intergenic sequences, (6) possession of a virion-associated polymerase, (7) helical nucleocapsid as the functional template for synthesis of replicative and
messenger RNA, (8)
replication
by synthesis of a full-length antigenome, (9) transcription of messenger RNAs by sequential interrupted synthesis from a single promotor, (10) transcription and
replication
in the cytoplasm, and (11) maturation by envelopment of independently assembled nucleocapsids at membrane sites containing inserted viral proteins. These data suggest that all nonsegmented negative-strand RNA viruses are derived from a common progenitor and support the classification of the families
Filoviridae, Paramyxoviridae
and
Rhabdoviridae
in the order
Mononegavirales
. In addition, comparative amino acid sequence analyses of nucleoproteins and polymerase proteins suggest that filoviruses are more closely related to paramyxoviruses than to rhabdoviruses.
Serologic Relationships
Marburg, Ebola and Reston viruses are serologically distinct. The Ebola strains and Reston virus show cross-reactions in broadly specific tests, such as
immunofluorescence
assays, whereas Marburg shows only unusual or no cross-reactivity with Ebola and Reston viruses. Neutralization tests for Marburg and Ebola viruses have not yet been shown sufficiently reliable to enable determination of taxonomic relationships.
Epidemiology
The source of infection in the first Marburg virus outbreak were African green monkeys imported from Uganda to Germany and Yugoslavia, but these animals were not considered to be the reservoir of the virus. With the other cases of Marburg virus infections and with the Ebola virus outbreaks, the sources of infection are unknown. In each Ebola outbreak, the initial patient spread the disease to close family members through intimate contact with the patient. In one case, sexual transmission of Marburg virus has been verified. After hospitalization of an infected person, the disease spread rapidly via contaminated needles and contact with blood. The cynomolgus monkeys affected in the recent Reston virus outbreaks were probably infected in the Philippines. During the 1989 Reston epizootic four animal caretakers with high exposure to the infected monkeys seroconverted, and virus was isolated from one case. None of the four has had symptoms which could be related to a disease caused by filoviruses.
Transmission and Tissue Tropism
The mode of primary infection in any natural setting is unknown with Marburg and Ebola viruses. All secondary cases have been nosocomial or caused by intimate contact with a patient. Transmission occurs usually by contaminated blood samples. One Marburg case was acquired by sexual contact more than 60 days after the original infection. In addition, there is evidence to suggest respiratory spread of infection. Epidemiological data of the 1989 Reston outbreak suggest that droplet or vomit transmission was a major factor in virus spread within quarantine facilities.
Virus is usually recovered from acute-phase sera and has also been found in throat washes, urine, soft tissue effusates, semen and anterior eye fluid, even when the specimens were obtained late in convalescence. It has also been regularly isolated from autoptic material, such as spleen, lymph nodes, liver and kidney but rarely from brain or other nervous tissues.
Pathogenicity
Marburg and Ebola viruses cause severe
hemorrhagic
fever in humans and laboratory primates. According to the evidence present to date. Reston virus may also cause
hemorrhagic
fever in monkeys, but appears to be apothogenic for humans. In man, the Zaire strain of Ebola virus apparently carries the highest mortality when compared with the Sudan strain or Marburg virus, although it is not clear to what extent these differences depend on the mode of transmission.
Clinical Features of Infection
Clinical symptoms are similar with Marburg and Ebola virus infections. Following incubation periods of 4-16 days, onset is sudden, marked by fever, chills, headache, anorexia and myalgia. These signs are soon followed by nausea, vomiting, sore throat, abdominal pain and diarrhea. When first examined, patients are usually overtly ill, dehydrated, apathetic and disoriented. Pharyngeal and conjunctival injections are usual. Most of the patients develop severe
hemorrhagic
manifestations, usually between days 5 and 7. Bleeding is often from multiple sites, with the
gastrointestinal tract,
lungs and gingiva the most commonly involved. Bleeding and oropharyngeal lesions usually herald a fatal outcome. Death occurs between days 7 and 16, usually from shock with or without severe blood loss.
Pathology and Histopathology
Marburg and Ebola viruses cause similar pathological changes in man. The most striking lesions are found in liver, spleen and kidney. These lesions are characterized by focal hepatic necrosis with little inflammatory response and by follicular necrosis of lymph nodes and spleen. In late stages of the disease,
hemorrhage
occurs in the
gastrointestinal tract,
pleural, pericardial and peritoneal spaces and into the renal tubules with deposition of fibrin. Abnormalities in
coagulation
parameters include fibrin split products and prolonged prothrombin and partial thromboplastin times, suggesting that disseminated intravascular
coagulation
is a terminal event. There is usually also profound
leukopenia
in association with secondary bacteremia. Reston virus causes similar pathological changes in monkeys as described for human infections with Marburg and Ebola viruses. In Reston-infected animals it was clearly demonstrated that virus
replication
was extensive in fixed tissue macrophages, interstitial fibroblasts of many organs, circulating macrophages and monocytes, and less frequently in vascular endothelial cells, hepatocytes, adrenal cortical cells and renal tubular epithelium. Macrophages and fibroblasts seem to be the first and prefered site of
replication
by this filovirus.
Immune Response
Humoral immune response to Marburg and Ebola viruses can be detected as early as 10-14 days after infection. Antibodies are directed primarily against the surface glycoproteins. Owing to the unreliability of neutralization tests, little can be said about their protective effects. Little is known also about cell-mediated immune response to these viruses.
Prevention and Control
Chemotherapeutic or immunization strategies are not available to date to prevent filovirus infections by pre-exposure or post-exposure prophylaxis. After vaccination of animals with virus antigens or inactivated whole virus vaccines there was no protection against challenge with live virus. No antiviral drug has been shown to be effective, either, even under
in vitro
conditions.
A specific treatment for filovirus
hemorrhagic
fever is therefore not available, but knowledge of the expected clinical course can anticipate medical complications including disseminated intravascular
coagulation,
shock, encephalomyelitis, cerebral
edema, kidney failure, superinfection hypoxia and hypotension. Patients have to be isolated and clinical personnel to be protected. Human
interferon,
human convalescent plasma and anticoagulation therapy have been used, but the success is still controversial.
Diagnosis
Because Marburg and Ebola virus are highly virulent, special precautions need to be taken when collecting specimens. Isolation of virus from acute-phase serum of appropriate cell cultures (Vero cells, for Reston virus and Ebola-Sudan virus MA-104 cells (from a fetal rhesus monkey kidney cell line) are more sensitive) is the most reliable method of diagnosis. Suitable tissues for virus isolation are also liver, spleen, lymph nodes, kidney and heart obtained at autopsy. During
viremia, particles can usually be visualized by
electron microscopy. Serum
antibody
titers are determined by indirect fluorescent
antibody
immunofluorescence (
(IFA)
and
ELISA,
whereas the specific reactivity of positive sera should be confirmed by either radioimmune precipitation or Western blot analysis. Serum from patients with suspected cases should be inactivated by
gamma irradiation before handling. Neutralization tests are totally unreliable for filoviruses.
Future Perspectives
Experimental work on Marburg and Ebola viruses has been greatly impeded in the past by the high pathogenicity of these agents. With the advent of
recombinant DNA technology,
however, we are beginning now to understand the molecular structure of these viruses and will soon understand the details of virus
replication
and virus-host interactions. Use of new strains of lower pathogenicity, such as the Reston virus, will contribute to reach these goals. These new approaches will also allow refinement of the diagnostic tools to permit more accurate virus identification in the field and to better understand frequency, natural reservoirs and transmission modes of these viruses.
See also:
Measles virus ; Rhabdoviruses.
Further Reading
Jahrling, PB (1991) Filoviruses and arenaviruses. In: Balows A (ed.)
Manual of Clinical Microbiology
. Washington, DC: American Society for Microbiology.
Martini GA and Siegert R (eds) (1971)
Marburg Virus Disease
. New York: Springer Verlag.
Pattyn SR (ed.) (1978)
Ebola Virus Hemorrhagic Fever
. Amsterdam: Elsevier/North-Holland Biomedical Press.
