The long view: a selective review of 40 years of Newcastle disease research

Abstract

This review is written for the series celebrating the 40th year since the first issue of Avian Pathology. The aim of the authors was to cover the developments in Newcastle disease (ND) research over the last 40 years that they considered significant. During those 40 years there have been several panzootics of this serious disease in poultry and for the last 30 years there has been a continuing panzootic in domestic pigeons, which has spread to wild birds and poultry. The 40-year period began with worldwide outbreaks of severe ND, which served as an important impetus for ND research work. Although early work was concerned with controlling the disease, specifically by improving and developing new vaccines and vaccine regimens, even prior to the 1970s ND virus was seen as a useful laboratory virus for replication and virulence studies. This review covers the historical developments in the following areas: understanding the molecular basis of virulence; epidemiology and relatedness of different ND strains, both antigenically and genetically; the emergence of virulent strains and their relationship with viruses of low virulence; sequencing and understanding the viral genome and genes; the development of rapid molecular-based diagnostic tests; and the phylogeny and molecular taxonomy of ND virus. The authors suggest areas in which future research could or should be undertaken.

Introduction

This review is written for the series celebrating the 40th year since the first issue of Avian Pathology; it is short, selective and subjective. Throughout those 40 years, Newcastle disease (ND) has been a very serious problem for poultry production in many countries and enormous efforts have been made at controlling the disease and understanding its epidemiology and virology. Clearly there will be many colleagues in that time whose significant and important contributions in these fields we have ignored or not given sufficient weight to; to them we apologize. Equally, when several groups are working in the same area and publishing about the same time it is often difficult to determine many years later who made the first discovery of an important finding; again, if we have made historical mistakes we apologize. Our intention was to produce a review that is readable for the non-specialist and, for those who take an interest in the historical development of disease research, to provide points of reference that will enable deeper evaluation.

The virulent strains of avian paramyxovirus type 1 (APMV-1) cause a serious disease in chickens and other birds known as ND. Outbreaks of ND may be devastating, with flock mortality approaching 100% in fully susceptible chickens. It has often spread rapidly during epizootics in poultry, causing severe economic losses due to disease and, for countries that export poultry or poultry products, losses resulting from trade restraints and embargoes.

ND was first described in 1926 in Newcastle-on-Tyne, England (from where it got its name) and on the island of Java, now part of Indonesia, although there have been some suggestions that there may have been earlier outbreaks (Alexander, 2001). It appears that initially the disease spread rapidly in Asia (Brandly, 1964) but only slowly to the rest of the world, taking between 16 and 40 years to become a true panzootic, with some vanguard epizootics such as that in Newcastle-on-Tyne.

When the first issue of Avian Pathology appeared in 1972 the poultry industries in just about every country in the world were in the throes of, beginning to recover from, or about to experience devastating outbreaks due to a particularly virulent form of ND virus (NDV). This “second panzootic” appears to have emerged in the late 1960s and taken only 4 years to spread across the world, considerably faster than the first. The impact of this panzootic cannot be underestimated; in Great Britain, for example, following 43 outbreaks in 1969 there had been 3328 outbreaks in 1970 and 4217 in 1971 before the widespread use of stringent live vaccination regimens began to bring the spread of disease under control (Alexander, 2001).

Research before the 1970s

Although Doyle had demonstrated clearly and carefully that ND was a new virus disease of poultry (Doyle, 1927), there was considerable scepticism—with many putting forward the view that the disease was just another form of “fowl plague” (now highly pathogenic avian influenza) and early research mainly focused on confirming this or not. Doyle finally published a definitive paper in 1935 (Doyle 1935), which, with the work of Burnet & Ferry (1934), ended the argument. At about this time a “new” disease of poultry, termed “pneumoencephalitis”, which subsequently spread across the USA, was emerging in California. Even though the disease was far less pathogenic than the ND described by Doyle, Beach showed that the causative organism of this disease was also NDV (Beach, 1944). The appearance of two different disease forms, coupled with the later identification of strains of NDV in the USA that produced little or no disease in chickens, opened up a whole new line of research aimed at determining why apparently very similar strains of the same virus should show such differences in virulence. These low virulence NDVs also presented the opportunity of live vaccines following the isolation of what are now probably the two most frequently used animal vaccines: Hitchner B-1 (the origins of which were described in a beautifully written paper by Hitchner, 1975), and the slightly more virulent La Sota (Goldhaft, 1980).

Rather than summarize research leading up to the 1970s, it is probably wisest and simplest to refer the reader to three publications. The first is the proceedings of a symposium held in Madison, Wisconsin in 1963 (Hanson, 1964). Bob Hanson managed to get together representatives of nearly every group working on ND or NDV, and the proceedings of the “Madison Symposium” became a written summary of all that was known of ND and the causative virus at that time and the research work that was continuing. The second publication was an incredibly detailed review written by John Lancaster (1966) covering all aspects of ND over the period 1926 to 1964, including epidemiology, diagnosis, control and vaccines and vaccination. The third was the chapter “Newcastle Disease” in the sixth edition of Diseases of Poultry published in 1972 (Hanson, 1972). This chapter was written by R. P. “Bob” Hanson, who had first started working with ND in the late 1940s and had become the pre-eminent scientist on the veterinary aspects of ND; this is reflected in this excellent review of the subject up to the beginning of the 1970s.

During the 1960s virologists, not necessarily concerned with the veterinary aspects of ND, had shown an interest in NDV. It was a convenient virus for laboratory work as it affected humans only mildly, if at all, and was relatively easy to grow, but its close similarity to human pathogens such as mumps, measles and influenza viruses meant it could be used as a model virus and particularly, because of the different virulence of different strains, as a model for virus virulence. So at the time of the second ND panzootic and the first publication of Avian Pathology, in various parts of the world in addition to those interested in ND as a disease to be controlled there were others pursuing more academic lines of research investigating virus structure, replication, cytopathogenicity and virulence.

The 1970s

Research work undertaken in the 1970s was greatly influenced by the panzootic with which the decade had begun. More money became available for academic research on the virus and there was also a need to understand the disease, especially the way it spread, and to develop methods to bring it under control.

In Great Britain and much of Europe the second panzootic appeared to have been brought under control by the use of live vaccines, although the methods of mass application used, often involving aerosol application, frequently resulted in severe reactions including mortalities. Clearly, if vaccination was to be used preventively, methods of administration and vaccine regimens would need to be developed that reduced the levels of reaction to live vaccines to a minimum. Similarly, prior to the panzootic, inactivated vaccines had been largely aqueous based, at best using aluminium hydroxide as an adjuvant. These vaccines had been found wanting and the need for better inactivated vaccines resulted in the development of oil emulsion inactivated vaccines during the 1970s. Much of the type of developmental work on NDV vaccines carried out during the 1970s is described in detail in a book entitled Newcastle Disease Vaccines: Their Production and Use by Allan et al. (1978).

The second panzootic also led to numerous studies to determine whether there was a reservoir of virus in wild birds. The finding of significance was that there were reservoirs of NDVs in migratory waterfowl species that were of low virulence for chickens and apparently unrelated to known virulent viruses (reviewed by Hanson, 1976).

Advances in understanding the structure and replication of NDV also progressed during the 1970s. One of the important findings in this period was the demonstration by Scheid & Choppin (1973) that, unlike influenza viruses, the haemagglutination and neuraminidase activities of paramyxoviruses, including NDVs, were both associated with the same glycoprotein. Scheid and Choppin also suggested that the second, smaller, structural glycoprotein was responsible for cell fusion. The other important observation, made by several groups at about the same time, was that this smaller glycoprotein underwent post-translational cleavage; we now know that this is crucial for virus infectivity and plays a principal role in virus virulence.

Perhaps the most significant work on NDV to take place in the 1970s was that on understanding the molecular basis of virulence, primarily carried out by R. Rott, H.-D. Klenk and their co-workers. It was shown that the smaller glycoprotein was responsible for cell fusion and, since it enabled fusion of the virus and cell membrane so that the RNA could enter the cell, was essential for virus infectivity. However, it was known that NDV particles are produced during replication, with this fusion (F) protein as an inactive precursor glycoprotein, F0; this has to be cleaved to F1 and F2 polypeptides, which remain bound by disulphide bonds, for the protein to be active and the virus particles to be infectious. Further, it was shown that this post-translation cleavage is mediated by host cell proteases, that trypsin is capable of cleaving F0 for all NDV strains and that in vitro treatment of non-infectious virus with trypsin will restore infectivity (Nagai et al., 1976a, b). That the cleavability of the fusion protein was related to virus virulence in vivo was first established by showing that viruses virulent for chickens could replicate in a wide range of cell types in vitro with or without added trypsin, whereas strains of low virulence could replicate only when trypsin was added (Rott, 1979).

The 1980s

Large-scale nucleotide sequencing and phylogenetic techniques now used to compare strains of viruses, especially for epidemiological purposes, were not available for most of the 1980s, but there was a need for diagnosticians and epidemiologists to know whether strains and isolates of NDV were related or distinct from others. Russell & Alexander (1983) addressed this problem by producing mouse monoclonal antibodies (mAbs) to NDV strain Ulster 2C. Using a panel of these mAbs they showed that isolates sharing the ability to react or not with the same mAbs shared epizootiological and other properties, and this approach enabled distinction between different epizootic viruses and between vaccine and field strains.

Throughout the 1980s the veterinary ND world was dominated by another panzootic of ND (Alexander, 2001), which occurred not in poultry but in pigeons, primarily racing and other domesticated pigeons, although there was spread to feral pigeons and to poultry. Fortuitously, the NDV responsible showed some antigenic variation, particularly a unique reaction pattern with the mAbs of Russell and Alexander, so that its spread around the world could be easily detected and monitored (Alexander et al., 1985).

During the 1980s from about 1986, newly developed molecular biological techniques, especially nucleotide sequencing, were applied to NDV to enable estimations of genome size and gene order (Millar & Emmerson, 1988).The estimated genome size of 15,126 nucleotides was close to but slightly lower than the now known sizes of 15,186, 15,192 and 15,198 for different strains of NDV. The genome was considered to consist of six genes each coding for a structural protein, and the genome organization was very similar to the current model (Figure 1)—although it was known that there was a second overlapping open reading frame within the polymerase (P) gene, suggesting there may be a seventh protein (see below).

Figure 1.  NDV genome. Updated from Millar & Emerson (1988). The order of the genes is shown in the middle row, the size of the genes (in nucleotides) is shown in the bottom row and the number of nucleotides making up intergenic and leader regions is shown in the top row. Gene P has been shown to have two overlapping open reading frames and also codes for the V protein (Steward et al., 1993). Where a size range is shown these are the values recorded for different NDVs.

Nucleotide sequencing also allowed the deduction of the amino acids at the cleavage site of the F0 protein and, even when only a small number of viruses had been sequenced, it appeared that viruses of low virulence usually had the amino acid sequence ¹¹³ K/RQG/ER*L¹¹⁷ (where * marks the cleavage site) but the sequence was ¹¹³ RQK/RR*F¹¹⁷ in virulent viruses (Rott & Klenk, 1988). The presence of the additional basic amino acid suggested that cleavage could be effected by a protease or proteases present in a wide range of host tissues and organs; but for viruses of low virulence, cleavage could occur only with proteases recognizing a single arginine (i.e. trypsin-like enzymes). This suggested that viruses of low virulence would be restricted in infected hosts to the sites where they are able to replicate (i.e. those with trypsin-like enzymes, such as the respiratory and intestinal tracts), whereas virulent viruses would be able to replicate and cause damage in a range of tissues and organs, resulting in fatal systemic infections. This theory has now been substantiated.

As with the previous decades, a convenient book detailing the then up-to-date knowledge on all aspects of ND was published in 1988 (Alexander, 1988).

The 1990s

One important advance during the 1990s was the testing of the theory of the dependence of NDV virulence on the F0 cleavage site put forward by Rott and Klenk. Several publications, such as that by Collins et al. (1993), reported the testing of a large number of NDV isolates and strains, and confirmed that all those virulent in in vivo tests in chickens had, as a minimum, deduced basic amino acids at positions 113, 115 and 116 with phenylalanine at position 117, while all viruses of low virulence had basic amino acids only at positions 113 and 116 with leucine at position 117. These confirmatory studies were important in establishing the molecular basis of virulence ahead of the adoption of a molecular-based definition of ND by the World Organisation for Animal Health (OIE) at the end of the decade.

Molecular techniques being developed during the 1990s allowed comparisons of the genetic relatedness of different NDVs. Seal et al. (1995) used reverse transcription-polymerase chain reaction and nucleotide sequencing to generate sequences of a relatively small portion of the F gene of a number of NDV isolates and compared them phylogenetically. An initial study by Ballagi-Pordany et al. (1996), using restriction enzyme analysis, identified six distinct genetic lineages (I to VI). This was preliminary to a much larger study by the same laboratory using similar techniques as Seal et al., which established seven genetic lineages (I to VII; Lomniczi et al., 1998). This approach for detecting genetic variation was to prove extremely useful for epidemiologists and diagnosticians, especially since using partial sequencing of the F gene that included the cleavage site region would allow determination of the virulence of the NDV isolate.

Understanding this molecular basis for virulence and the ability to sequence new isolates also resulted during the 1990s in finding a possible solution for questions that had puzzled those working in the field for many years; particularly what was the origin of ND and why had it apparently suddenly appeared in 1926? Hanson had specifically addressed this in publications in the 1970s and considered there were three possibilities: the virus had always been present in poultry but had gone unrecognized until 1926, possibly as a result of changes in the way poultry were reared; the virulent virus was present in another species from which it had spread to poultry; or the virus emerged as a result of mutation. Hanson and just about every contemporary ND virologist favoured the second possibility, fuelled by the fact that the second panzootic virus had been introduced into and spread in California by captive caged birds and the knowledge that psittacine species showed some resistance to virulent NDVs. Thus it became something of a dogma that virulent NDVs had reservoirs in psittacine species in the tropics, occasionally spreading to poultry, despite the lack of direct evidence that this was the case.

Two outbreaks in poultry in the 1990s—one restricted, the other extensive—yielded viruses that suggest the third possibility may be more likely for the emergence of virulent viruses. The first of these related to the examination of a virus responsible for two ND outbreaks in Ireland in 1990. This virus proved to be antigenically and genetically very different from all other virulent viruses, but was genetically close to a group of viruses of low virulence that had been primarily isolated from waterfowl (now known to be Class 1 viruses; see below). These findings led to the suggestion that the virulent virus may have emerged from a virus of low virulence, although this would have required four point mutations at the cleavage site (Table 1). The second outbreaks occurred in Australia in 1998 and gave far better evidence that the virulent virus had emerged as a result of mutation from a virus of low virulence. Historically, Australia had been free of ND for over 60 years until a series of outbreaks began in 1998. It was shown that the virus responsible would have required only two nucleotide changes at the cleavage site motif if it had arisen from a virus of low virulence known to be present in wild birds in Australia and occasionally infecting poultry (Table 1). Perhaps even more conclusive were the isolations of NDVs that appeared to be intermediate in terms of nucleotide/amino acid changes at the F0 cleavage site (Gould et al., 2001). The inference from these two studies was that virulent viruses may well emerge from viruses of low virulence, such as those known to be present in wild bird reservoirs, by mutations in the F gene coding for the F0 cleavage site. If this was how virulent NDVs emerged it could explain the sudden appearance of the disease in 1926 and the subsequent emergence of distinguishable virulent viruses. It could also be argued that such mutations of low to high virulence may have occurred prior to 1926, as suggested by some authors (Alexander, 2001), but resulted in only limited outbreaks.

Table 1.  Nucleotide/amino acid sequences at the F0 cleavage site of virulent NDVs isolated from poultry in Ireland and Australia compared with genetically closely related virus of low virulence.

Virus	Virulence	Nucleotide/amino acid sequence at F0 cleavage site
MC110	Low	GAA CGG CAG GAG CGT CTG
		^112ERQER*L^117
Ireland/34/90	High	A AA CGG CAG A AG CGT T T T
		^112 KRQKR*F ^117
Australian isolates
1154/98	Low	GGA AGG AGA CAG GGG CGT CTT
		^111GRRQGR*L^117
1249/98	High	GGA AGG AGA CAG A GG CGT T TT
		^111GRRQRR*F ^117

Steward et al. (1993) showed that the P gene of NDV has two overlapping open reading frames and is edited during replication to produce three mRNA species, termed P (unedited), V and W. Later work has shown the V protein to have some importance in the replication of NDV (see below), whereas the function of W is unknown.

Although, as discussed above, the nucleotide sequence of the whole genome of NDV could be inferred from the partial sequences of different strains in the 1980s, it was not until the late 1990s that the whole sequence of the genome of a single strain was determined (Phillips et al., 1998). It is a measure of the advances in this technology that what proved so difficult just over 10 years ago is now commonplace and the full genome sequences of numerous strains and isolates have been obtained and published almost as a matter of routine. At about the same time, research workers in the Netherlands not only determined the whole genome sequence of strain La Sota (de Leeuw & Peeters, 1999) but were able to assemble a full-length cDNA of this virus (Peeters et al., 1999) from which infectious virus could be rescued, thus allowing reverse genetics techniques to be applied to NDV. In this early study they were able to demonstrate that changing the F0 cleavage site of strain La Sota to that of a typical high virulence virus greatly increased its virulence in vivo. This confirmed the molecular basis of virulence and the potential for virulent viruses to emerge from those of low virulence by mutation.

The 2000s

With the ongoing improvement and availability of sequencing techniques during the decade, the 2000s were dominated by a range of sequencing, phylogenetics and molecular epidemiological papers and reports investigating and defining the genetic variation of APMV-1 strains, building on the initial pioneering studies of the 1990s. Early studies by Seal et al. (2000) used phylogenetic approaches to assess the relationships within the Paramyxoviridae according to the nucleotide sequence of the Matrix protein gene, concluding based on these analyses that NDV should be assigned to its own genus (now Avulavirus genus).

Aldous et al. (2003) studied the partial nucleotide sequences of 338 isolates of NDV representing a range of viruses of different phenotypes and temporal, geographical and host origins. They divided the isolates into six broadly distinct genetic groups (lineages 1 to 6), most of which could be divided into several sub-lineages. Generally the lineages or sub-lineages corresponded to the genetic lineages described earlier (I to VIII), but lineage 6 represented a new group. Aldous et al. (2003) stressed that while the NDVs placed in genetic lineages 1 to 5 (or I to VIII) were genetically close, viruses that were placed in lineage 6 were genetically very different from all the other NDV isolates.

Czegledi et al. (2006) investigated the evolution of APMV-1 strains genetically and concluded that viruses could be divided into Class I viruses (corresponding to lineage 6 of Aldous et al., 2003) with a genome size of 15,198 nucleotides and Class II viruses with a genome size of 15,186, representing the ancestors of all the other lineages, which by this time in their system had risen to 10 (lineages I to X). They further suggested that a major branch in the Class II tree had occurred, resulting in a clade of viruses with a genome size of 15,192 (lineages V to X). Kim et al. (2007) examined a number of Class 1 ND isolates and suggested there were nine distinguishable lineages. Many other studies have looked at genetic relationships that exist within specific geographical and/or chronological limits. The data produced in all these studies provide insights and observations into the progression of ND panzootics, evolution, interactions, relationships between viruses in wild birds and those responsible for outbreaks in poultry and emerging strains or lineages of NDV.

The technological advances in nucleotide sequencing and the amount of data that became available during this decade added to the acceptance of a molecular basis for confirming virulent virus by the OIE, and has led to a move to develop reliable, robust and rapid molecular tests for the detection and differentiation of ND isolates. One of the earlier studies in this area was by Aldous et al. (2001), in which a panel of TaqMan probes was applied to enable the detection and differentiation of many variable strains of APMV-1. The outbreaks of ND in California in 2002 and 2003 provided the impetus for the swift development and application of high-throughput testing methods for detection of NDV, and Wise et al. (2004) reported the development and use of a real-time reverse transcription-polymerase chain reaction test. The use of the real-time reverse transcription-polymerase chain reaction dramatically improved the speed at which a positive diagnosis of NDV could be completed; however, the classical problems associated with molecular detection remain, particularly the considerable genetic variation that occurs between APMV-1 strains and isolates and the potential for mixed infections with high and low virulence viruses. Some molecular tools have been developed to discriminate between isolates of high and low virulence but no one assay has yet been developed that is able to sensitively detect all lineages of NDV.

During the 2000s, much work has been done on studying the effect that virus components other than the fusion protein may have on virus virulence employing reverse genetics techniques (reviewed in detail by Dortmans et al., 2011). In particular, the V protein was shown to be an alpha-interferon antagonist and to have some influence on virus virulence (Huang et al., 2003). Other work has suggested that interactions between the HN and F proteins may modulate virulence. Evidence is emerging that virulence is a multigenic trait with all genes contributing through enhancement or reduction protein function. However, a multiple basic amino acid cleavage site sequence of the F0 protein remains the principal determinant for the large clinical difference between strains of high and low virulence in susceptible hosts.

Reverse genetics techniques have also been used by numerous researchers to exploit the ability of NDV to tolerate the inclusion of foreign genes and act as a vaccine vector (Nakaya et al., 2001), especially as a dual-purpose vaccine. Specifically, an avian influenza H5 gene has been incorporated into NDV and protection against both diseases has been demonstrated (Veits et al., 2006). A similar recombinant virus based on NDV vaccine strain La Sota and expressing the Asian lineage H5 HA gene was produced in the People's Republic of China (Ge et al., 2007), where it has been licensed and used widely.

Future research

The last 40 years have shown advances that were almost unbelievable when the first Avian Pathology issue was published. We have no idea what new technologies and approaches are waiting to be developed, but even with our existing tools and knowledge there are several obvious areas that require future development.

The relative ease in which full genome sequences are now able to be produced through the widespread use and reduction in cost of next-generation sequencing technologies are already having an effect. In the 10 years after the first full genome was sequenced, well over 150 full genomes (166 to date) have been published and submitted to online databases. The development of bioinformatics tools to investigate nucleotide data is an area of rapid progress that will undoubtedly shape NDV research in the next 10 years.

Comprehensive analysis by Miller et al. (2009) suggests that the systems available for the nomenclature of NDV would benefit from review, agreement and adoption of the most appropriate. Criteria for identifying genetic lineages, sub-lineages and clades should be established and defined to enable the identification of consistent and accurate grouping of viruses for epidemiological and diagnostic purposes.

Possible recombination of different NDVs that may play a role in the evolution of the virus has been reported by some authors, while others have urged caution in the interpretation of such results (Afonso, 2008). Clearly if recombination occurred it would have both academic and important practical implications; further work in this area is imperative.

Another area that needs evaluating and expanding is understanding the emergence of virulent virus. The suggestion that viruses of low virulence may mutate to high virulence is particularly attractive since it allows answers to several questions that had puzzled workers in the field for many years (see above). But this also raises several further questions: Is this the mechanism responsible for all high virulence viruses? How often does it happen and why? If such mutations occur, do they take place in the wild bird reservoir or after the low virulence viruses have infected poultry? Finally, since billions of doses of live vaccines employing different strains of NDV are used every year, what is the likelihood of these mutating to virulence? Is there a need for screening new strains for their ability to mutate, and how would this be done?

Early phylogenetic work (Lomniczi et al., 1998) suggested that genetic lineages of NDV may still cause outbreaks many years after their apparent emergence. Where are these viruses maintained? It is possible that virulent NDVs may become endemic in vaccinated birds in which they replicate but cause no clinical signs until vaccination is stopped or is unsuccessful due to vaccination failure or immune suppression as a result of infections with other agents, as suggested by some (Alexander, 2011). Similarly, in many countries throughout Asia and Africa, ND remains endemic in commercial poultry despite intensive vaccination programmes that have been applied for decades. Is this because the vaccines are inadequate, or because they are not applied correctly? To what extent do the viruses circulating in the vaccinated populations where ND is endemic show genetic or antigenic variation?

The ND panzootic in pigeons has now continued for over 30 years with little sign of it being controlled and eradicated (Alexander, 2011). Clearly, where there are attempts at control, the methods employed do not seem to be working; if this situation, with its continued threat to poultry, is to be resolved, new methods—preferably resulting in eradication—will have to be designed and evaluated.

Finally, but possibly most importantly, in developing countries throughout Asia and Africa most chickens are reared as village or scavenging birds, which are an extremely important food source. Annual losses of 60% or more are not uncommon and most are attributed to ND (Spradbrow, 1992); there seems an urgent need to tackle this problem.

As befitting such an important disease, there has been an enormous amount of knowledge of ND and the causative virus obtained and deduced over the last 40 years. ND has been recognized for over 85 years, but despite all the research work and attention it has received, it still has a worldwide influence on poultry production either as a continuing disease problem or as a constant threat.