Epidemiology of Human Metapneumovirus

Abstract

Since the discovery of human metapneumovirus (hMPV) in 2001, the virus has been identified worldwide. hMPV is a common respiratory pathogen, particularly in infants and young children. The virus is associated with both upper and lower respiratory tract infections and may be a trigger for asthma. At least two major genotypes of hMPV circulate during community outbreaks. Whether these genotypes represent distinct serotypes remains controversial. The major challenges faced by the medical and scientific communities are the understanding of the pathogenesis of hMPV disease and the development of a safe and effective vaccine to protect against infection and disease caused by this newly recognized respiratory virus.

INTRODUCTION

Respiratory tract infections are a leading cause of morbidity and mortality worldwide. For children <5 years old, respiratory tract infections are ranked as the second leading cause of death regardless of geographical area (82). Acute respiratory tract infections are the most common illness regardless of age or gender (80). Pneumonia, influenza, and influenza-like illnesses (ILI) are the sixth leading cause of death in the United States, accounting for >45,000 deaths annually (96). Each year in the United States, respiratory tract infections account for 4 million ambulatory care visits and >500,000 hospitalizations, at an annual cost of $14.6 billion (107). In The Netherlands, the overall incidence of acute respiratory tract infection is 5,445 per 10,000 person-years (124).

Although the clinical manifestations of respiratory tract disease are easily recognized, the etiological agent responsible for disease is often not identified. For community-acquired pneumonia, a microbiological diagnosis can be made in <50% of cases (83, 102, 140). In the pediatric population, respiratory syncytial virus (RSV), parainfluenza viruses, and influenza virus are known as the major causes of bronchiolitis and lower respiratory tract infections (LRTI). However, in a third of these cases of LRTI in children, an infectious agent cannot be identified (27, 135). In nearly half of upper respiratory illnesses (URI) in children, an infectious cause cannot be determined (86). The etiology of a majority of LRTI is thought to be viral (39), yet in only 40% of cases can a viral agent be identified, even with use of state-of-the-art genomic amplification methods (70). These observations suggest that previously unknown pathogens may be circulating and may be responsible for a substantial proportion of respiratory tract disease.

DISCOVERY

In 2001, van den Hoogen et al. reported the discovery of a novel virus from children with respiratory tract illness in The Netherlands (120). This agent was detected in respiratory secretions of 28 children that were collected during a 20-year period. This virus was distinct from common respiratory viruses, as immunological assays using virus-specific antibodies and PCR-based methods using virus genome-specific primers failed to identify this agent. The genetic characterization of this agent remained a mystery until the tools of molecular biology were applied to identify portions of the genomic sequence. Using a technique known as randomly primed PCR, these Dutch researchers were able to obtain genomic sequence of this novel pathogen. Based on limited sequence data, this virus appeared to be closely related to the avian pneumovirus, a member of the Metapneumovirus genus, and it was called human metapneumovirus (hMPV) (38, 120). Several lines of evidence suggested that hMPV was a common human respiratory pathogen. Seven of 68 respiratory specimens (10%), collected in the winter of 2000, that were screened tested positive for hMPV by reverse transcription-PCR (RT-PCR). Serological studies, based on both immunofluorescence and viral neutralization assays, revealed that by the age of 5 nearly all individuals had evidence of hMPV infection. Furthermore, screening of banked human sera showed that hMPV, or a virus closely related to it, had been circulating for at least 50 years, suggesting that the virus did not recently “jump” to the human population from an animal reservoir such as birds (120).

CLASSIFICATION AND GENOMIC STRUCTURE

The complete sequence of the hMPV genome has been elucidated (9, 49, 119). As the preliminary sequence data for the hMPV genome suggested, the virus is a member of the Metapneumovirus genus, a branch of the family Paramyxoviridae, and is genetically similar to, though distinct from, the avian pneumovirus (previously called turkey rhinotracheitis virus). The family Paramyxoviridae is composed of two subfamilies, one of which is Pneumovirinae, which includes both the Pneumovirus genus and the Metapneumovirus genus (Fig. 1) (38).

FIG. 1.

Human pathogens in the family Paramyxoviridae (data from reference 38).

RSV, a pneumovirus, is the most closely related human pathogen to hMPV, and the genomes of the viruses have similar features as well as distinct differences. The negative-strand RNA genomes of both viruses contain open reading frames (ORFs) that encode three putative viral envelope glycoproteins, the F (fusion), G (attachment), and SH (short hydrophobic) proteins, though the order of these genes differs in the two viruses (23, 119) (Fig. 2). Of these three genes, the F gene is the most highly conserved (33% identity on the amino acid level) between hMPV and RSV (119). hMPV F contains several features that are conserved among the F proteins of paramyxoviruses, which includes a characteristic distribution of cysteine residues, a putative cleavage site (a distinguishing feature of a fusion proteins), fusion domains, and anchor sequences (119).

FIG. 2.

Genomic maps of the Pneumovirinae. Genomic maps of the negative-sense, single-stranded RNA genomes of hMPV and RSV are displayed in the 3′-to-5′ orientation. In hMPV, the F and M2 genes are 3′ to the SH and G genes, whereas in RSV, the order of these genes is reversed. The RSV genome encodes two nonstructural proteins, NS-1 and NS-2 (shaded), that are not present in the hMPV genome. The M2 genes of both viruses carry two ORFs (M2-1 and M2-2) (not shown). The M2 and L genes overlap in the RSV genome (triangle). The L gene of each virus, encoding the viral RNA-dependent RNA polymerase, comprises two-thirds of the viral genome and is shortened for figure clarity. The genomes are not drawn to scale.

The amino acid identity between the G and SH proteins of hMPV and RSV could not be determined because the predicted amino acid sequences of the corresponding proteins of the two viruses could not be aligned (119). The putative G protein of hMPV is considerably smaller than the G protein of RSV (236 versus 299 amino acids). Unlike that in RSV, the region of the hMPV genome that encodes the putative G protein also carries an ORF immediately downstream of the G gene (in the same reading frame) and ORFs in the two other reading frames. It is unknown whether these accessory ORFs are expressed as separate proteins or are transcribed as part of the G protein through some RNA-editing event (119). The predicted amino acid sequences of the G gene of both hMPV and RSV indicate that these genes encode anchored type II glycoproteins (119, 127). The intracellular amino-terminal cytoplasmic domain of the G protein of both hMPV and RSV is relatively short (∼30 amino acids) and is adjacent to the hydrophobic transmembrane domain. RSV G is heavily glycosylated with both O- and N-linked sugars (24), and the predicted amino acid sequence suggests that the same is true for hMPV. Among isolates of RSV, the G gene is the most variable of all RSV genes, and this appears to be the case for the G gene of hMPV (55). The reasons for the genetic variability of the G gene are not clear, but it may have to do with evasion of the host's immune system by the virus. The SH gene of hMPV encodes a protein (180 amino acids) that is substantially longer than the corresponding protein of RSV (64 amino acids). The function of the SH proteins of both viruses remains unknown.

Other than the gene order, the major differences between the genomes of hMPV and RSV are the lack of two nonstructural genes, NS-1 and NS-2, at the 3′ end of the hMPV genome (Fig. 2). Because of the gradient of transcription along the RSV genome, NS-1 and NS-2 are transcribed in the greatest molar amounts of all the genes the of the RSV genome. These RSV genes encode an anti-interferon activity (15). The significance of the lack of NS-1 and NS-2 in the pathogenesis of hMPV is unknown.

Reverse genetics, developed over the last decade, allows for the recovery of recombinant negative-strand RNA viruses from cloned DNA (101). Manipulation of the hMPV genome by using the reverse genetics systems has begun to yield clues about the features of the viral genome (10, 49). In the relatively short time since the recovery of hMPV from cloned DNA, it is now clear that several hMPV genes are not required for viral propagation in cell culture. Like for RSV, both the G and SH genes can be deleted from the viral genome without eliminating the ability of the virus to replicate in vitro or in vivo (8, 58). The replication of hMPV lacking SH in the African green monkey model is comparable to that of wild-type virus, whereas the replication of hMPV lacking G is reduced sixfold compared to that of wild-type virus (8). These findings suggests that the F proteins of both RSV and hMPV contain an attachment and fusion activity and raise questions about whether the G and SH proteins play an essential role in interactions with the infected cell or immune system of the human host that have not yet been appreciated.

The reverse genetics system has also generated insights into the control of gene expression. The M2 gene of hMPV carries two overlapping ORFs, designated M2-1 and M2-2, and both are expressed during viral infection (17). Similar to the case for RSV, the M2-2 gene appears to control the switch of the viral RNA polymerase from viral gene transcription to viral genome replication. During infection with a recombinant hMPV lacking M2-2 or with a recombinant RSV lacking M2-2, there is accumulation of viral mRNA compared with viral genome template (7, 17). These findings may have implications for vaccine development. A recombinant hMPV lacking the M2-2 gene was found to be attenuated in viral replication but retained adequate gene expression capabilities to induce protective immunity (see “VACCINE CANDIDATES AND ANIMAL MODELS” below) (17).

CLINICAL MANIFESTATIONS OF hMPV INFECTION

Lower Respiratory Tract Disease

Evidence from many studies has demonstrated that hMPV is responsible for a substantial proportion of LRTI in infants and young children and is second only to RSV as a cause of bronchiolitis in early childhood (13, 22, 40, 65, 81, 106, 123, 125, 129). Williams et al. screened respiratory specimens from >2,000 children <5 years old, collected during a 20-year period at Vanderbilt University Medical Center, for hMPV (129). These children were followed as part of the Vanderbilt University Vaccine Clinic center in Nashville. In 20% of cases of LRTI (characterized by wheezing, rales, tachypnea, and dyspnea) which could not be attributed to a culturable respiratory virus, hMPV was detected. The incidence of hMPV-associated LRTI in young children varies with geographical location and time of year. The incidence estimates range from 5 to 15% in most studies (6, 13, 34, 79, 93, 122, 123, 129), but higher rates have been reported in other studies (28, 75).

The clinical manifestations of hMPV infection in young children are indistinguishable from the clinical manifestations of RSV infection. Features of hMPV infection included tachypnea, fever, cough, hypoxia, and changes on chest radiographs such as infiltrates, hyperinflation, and peribronchial cuffing (6, 13, 34, 79, 93, 122, 123, 129). Asymptomatic infection with hMPV in young children appears to be uncommon. In the study by Williams et al., only 1 of 86 asymptomatic children (1.2%) tested positive for hMPV, indicating a causal role of hMPV in LRTI (129).

LRTI associated with hMPV in infants and young children is a frequent cause of hospitalization. hMPV likely accounts for ∼10% of LRTI hospitalizations and is second to RSV as a cause of LRTI requiring hospitalization (6, 13, 16, 33, 34, 40, 81, 84, 89, 123, 129). The most common diagnoses in hospitalized children who test positive for hMPV are bronchiolitis, pneumonia, and bronchitis. There are no manifestations of infection that are specific for hMPV-associated disease. In the limited number of population-based studies performed, children <2 years old appear to be the most likely to be hospitalized due to hMPV-associated LRTI (81, 84). hMPV has been associated with severe LRTI requiring intensive care, in one case requiring extracorporeal membrane oxygenation (118), though data from some studies suggest that, in general, the severity of disease associated with hMPV may be less than that observed with RSV (13, 125). Although the mean age for hMPV-associated LRTI is slight greater than the mean age for RSV-associated LRTI in some studies (81), but not in others (129), it is now clear that infants and young children are the groups in the pediatric population that are most susceptible to severe disease caused by either RSV or hMPV. Effective therapies and vaccine strategies will need to target this patient population. Risk factors for severe RSV disease, such as a history of premature birth, underlying heart or lung disease, and a compromised immune system, are likely risk factors for severe hMPV disease (12, 34, 90, 123).

Upper Respiratory Tract Infection

Like other common human respiratory viruses, hMPV is also associated with URI. While the definition of URI varies among studies, hMPV may be responsible for 5 to 15% of cases of URI in children (129, 133). In a recent study, Williams et al. screened respiratory specimens collected over a 20-year period from children <5 years old with URI, characterized by coryza, conjunctivitis, pharyngitis, otitis media (OM), or stomatitis. These children did not have a prior virological diagnosis. Of the >2,000 specimens screened by real-time RT-PCR, the percentage of URIs attributed to hMPV was 1 to 5%. This varied from year to year. Overall, the percentage of URIs associated with hMPV was lower than that observed with influenza virus, parainfluenza viruses, adenovirus, and RSV (133).

Respiratory viruses have been implicated in the pathogenesis of OM (48, 103). Therefore, it was not surprising that hMPV has been identified in children with acute OM (109, 116, 132). In one study, 50% of the children with hMPV-associated URI were diagnosed with otitis media (133). In another study, one-third of children with hMPV-associated LRTI were diagnosed with concomitant acute OM (129). hMPV was detected in the nasal washes of 6% of children who presented with acute OM (132), suggesting that hMPV may be responsible for a small but significant percentage of cases of acute OM. The detection of RSV in the middle ear fluid of children with OM suggests a role of the virus in the pathogenesis of OM (48). Of eight children with OM whose nasal wash specimens tested positive for hMPV, only one had middle ear fluid that also tested positive for hMPV (132). Similar findings were observed elsewhere (116). These findings suggest that the inflammatory reaction to the virus, leading to incomplete or complete blockage of the eustachian tubes and subsequent invasion by bacteria, is a likely mechanism for virus-induced OM.

Wheezing and Asthma

A causative role of respiratory viruses in the initiation and progression of asthma remains a controversial issue. However, it is fairly well established that common respiratory viruses, such as rhinoviruses, can cause wheezing in some children and that infection with these viruses may induce exacerbations of reactive airways disease or asthma (56). hMPV was detected in 8% of children who presented to a Finnish hospital with exacerbation of wheezing. These hMPV-infected children also had increased levels of interleukin-8 in respiratory secretion (54). Several studies have reported the association of hMPV and asthma exacerbations (123, 126, 129, 131), while this association was not observed in another study (98). In addition, wheezing is a common symptom observed in children with hMPV-associated LRTI (33, 34, 89, 123, 126, 128, 129). hMPV has also been associated with acute asthma exacerbations requiring hospitalizations in adults (128). Investigation of the inflammatory mediators and cytokines induced by hMPV during infection should shed light on the mechanisms of hMPV-induced wheezing. Whether these potential mechanisms of wheezing are specific for hMPV or a more general response to a wide variety of respiratory pathogens remains to be determined.

hMPV in the Adult Population

hMPV has been associated with ILI, bronchitis, pneumonia, and exacerbations of both asthma and chronic obstructive pulmonary disease (COPD) in adults (37). The overall rates of hMPV disease in adults are likely lower than those observed in children. hMPV-associated disease occurs in adults of all ages (36). Falsey et al. detected hMPV in 3.4% of adults with respiratory tract illness. None of the 158 controls, who had no respiratory complaints, tested positive for hMPV (35). In previously healthy adults with acute respiratory infections screened during a 3-month period in San Francisco, Calif., hMPV was infrequently detected, though this may be due to the limited period of time of the study (70). In a study of patients who presented to general practitioners with ILI, 2.2% of individuals who tested negative for both RSV and influenza virus tested positive for hMPV (114). The major risk factor for hMPV-associated respiratory tract disease in adults without COPD is advanced age and underlying cardiopulmonary disease (12, 36, 47, 50). Dyspnea was more likely in elderly adults than in young adults infected with hMPV (36). The role of hMPV in exacerbations of COPD is less clear. hMPV has been detected in individuals with acute exacerbations of COPD (47, 76, 100). However, the duration of viral shedding in individuals with COPD is not established, and therefore a causal role may be difficult to define. Further investigations are needed to define the role of hMPV in COPD.

Infection in the Immunocompromised Host

hMPV can cause prolonged and serious infections in the immunocompromised host. Several case reports and small case series have described the detection of hMPV in immunocompromised individuals with severe lung disease (18, 34, 67, 78, 90, 130). In one study of lung transplant recipients, 9 of 25 individuals who were screened during a 1-year period tested positive for hMPV at least one time. Several fatalities were reported. It is difficult to determine the clinical features of hMPV disease in these individuals, as many of them had concurrent bacterial or fungal infections (67). In a prospective 4-year study of adults with hematological malignancies and respiratory tract infection, 9% of episodes of respiratory tract infections tested positive for hMPV. Sixteen of the 22 individuals who tested positive for hMPV (73%) were hematopoietic stem cell transplant recipients. Twenty hMPV-positive individuals had URI symptoms, but eight of these subsequently progressed to LRTI. Three of the patients with hMPV-associated LRTI died, two of whom had concomitant bacterial infection. hMPV was the only pathogen identified in one patient who died, and death occurred within a month after the onset of LRTI (130). hMPV was reported in an immunocompromised child, who had two hMPV infections with two different strains of the virus during a 10-month period (90). hMPV has been detected in children with human immunodeficiency virus (HIV) infection, though because of the small numbers of children in the study, it is difficult to determine whether hMPV-associated disease was more or less severe in HIV-infected children than in non-HIV-infected children (73).

Coinfection with hMPV and Other Respiratory Pathogens

Because the seasonal distributions of hMPV and RSV, two common pathogens, overlap, the potential for dual infection exists. Indeed, several studies have found a coinfection rate of <10% (34, 125, 129, 138). However, Greensill et al. reported that 70% of RSV-infected children who required intensive care in Liverpool, United Kingdom, were coinfected with hMPV, suggesting that the disease caused by RSV may be augmented by a concurrent hMPV infection, particularly in otherwise healthy children (44). In another study from the United Kingdom, hMPV and RSV coinfection conferred an increased risk of admission to the pediatric intensive care unit (112). However, population-based and case-control studies of hospitalized children have found that hMPV-RSV coinfections are uncommon. Of 668 hospital admissions screened by Mullins et al., none of the children tested positive for both hMPV and RSV (81). Lazar et al., using a case-control method, screened children with severe RSV disease (cases, defined as children who required admission to a pediatric intensive care unit) and children with mild RSV disease (controls) during a period in which hMPV was circulating in the community. Overall, none of the cases and none of the controls tested positive for hMPV (68).

In 2002, a new respiratory illness, severe acute respiratory syndrome (SARS), emerged in southeast Asia and was eventually attributed to a novel coronavirus (SARS coronavirus) (29, 62). In some locations in China and Canada, though not in others, a substantial proportion of individuals with SARS were also infected with hMPV (20, 21, 63, 97). It is unclear whether dual infection resulted in greater severity of illness. The significance of this dual infection is not known.

Bacterial pneumonia is frequent complication of influenza virus infection but not RSV infection (45). The current data suggest that bacterial infection complicating hMPV infection is uncommon. However, the proper studies to investigate this issue have not been conducted.

Other Potential Manifestations of hMPV Disease

The current understanding of hMPV is that during infection both the virus and the disease caused by the virus are limited to the respiratory tract. This may be due, in part, to the relative lack of published studies that have specific screened for hMPV in other organ systems or bodily fluids. In a case report on a 14-month-old, previously healthy child who died of encephalitis, hMPV RNA was detected by RT-PCR in both the brain and lung tissues, though hMPV antigens were not detected in either tissue by immune staining (111). In a 1-year study of hMPV in children with respiratory tract disease in Japan, 1 child of 29 children who tested positive for hMPV presented with encephalitis. hMPV was detected in respiratory secretions, though cerebrospinal fluid from this patient was apparently not tested for hMPV (57). These findings suggest that hMPV may cause disseminated infection in some individuals. Replication of RSV is also thought to be limited to the respiratory tract, though the detection of RSV sequences in blood has been reported (139). These findings have not been substantiated and remain controversial. Whether hMPV spreads beyond the respiratory system during infection remains to be determined.

EPIDEMIOLOGY

Geographical and Seasonal Distribution and Molecular Epidemiology of hMPV

hMPV has a worldwide distribution and has been identified on every continent (2, 12, 14, 26, 28, 30, 33, 40, 41, 43, 60, 71, 75, 85, 87, 89, 99, 104, 120). In temperate climates, hMPV circulates predominately in the late winter and spring, and the peak of activity at any given location often coincides with or follows the peak of RSV activity (1, 13, 16, 22, 34, 57, 61, 75, 81, 84, 88, 89, 105) (Fig. 3). In many communities, hMPV has been detected throughout the year, albeit at lower levels during the late spring, summer, and fall (34, 81, 88, 123, 129). Based on genomic sequencing and phylogenetic analysis, there are two major genotypes of hMPV, designated A and B (9, 22, 34, 93, 123, 129). These analyses are based on sequencing of the N, M, F, G, or L gene, and the genotype groupings are concordant regardless of which gene is studied. Whether these two genotypes represent distinct serogroups remains controversial (see “SEROLOGICAL PROPERTIES OF hMPV GENOTYPES” below). Each genotype appears to have at least two distinct subgroups (5, 14, 72, 122).

FIG. 3.

Distribution of hMPV and other respiratory viruses in New Haven, Connecticut. (Top panel) Weekly distribution of hMPV genotypes from 1 November 2001 to 31 October 2002. Genotype assignments are indicated. Every other week is marked for figure clarity. There are a total of 54 hMPV-positive samples. (Bottom panel) Monthly distribution of hMPV, RSV, influenza A and B viruses, and human parainfluenza viruses 1 to 3. The percentage of total positive specimens for each virus during the 1-year period is indicated. (Reprinted from reference 34 with permission. © 2004 by the Infectious Diseases Society of America. All rights reserved.)

The genomic organization of the two viral genotypes is identical (9). The major differences between the A and B genotypes are nucleotide polymorphisms, and the G and SH proteins contain the highest concentration of these polymorphisms (9). Like the G gene of RSV (94), the G gene of hMPV displays significant strain-to-strain variability (9, 52, 92). This nucleotide variability results in significant amino acid variability, insertions that retain the reading frame, and the use of alternate transcriptional termination codons (4). Most of the amino acid variability is observed in the putative extracellular domain of the protein. Overall, there is 32 to 37% amino acid identity of the G protein between the A and B genotypes of hMPV (1, 4, 92). This is analogous to that observed with G genes of RSV subgroups A and B (77, 94, 115).

Phylogenetic analysis of strains of hMPV reveals that the epidemiology of hMPV is complex and dynamic. Unlike influenza virus, where two or three strains spread across the globe each year, outbreaks of hMPV appear to be a local phenomenon. Strains of hMPV differ from community to community, and strains identified in one location may be quite similar to strains identified in other locations in different years. For example, the prototype strain identified in The Netherlands is genetically similar to strains identified in Australia; New Haven, Conn.; and Quebec, Canada in different years (34) (Fig. 4). Based on F gene sequences, viruses isolated in Australia (2001), France (2000 and 2002), Canada (1999, 2000, 2001, and 2002), Israel (2002), and The Netherlands (2001) were closely related, with few polymorphisms in the F gene (14). In any given year, viruses of both genotypes (and both subgroups in each genotype) can circulate (1, 14, 19, 34, 57, 93, 121). In St. Louis, Missouri, the predominant genotype of hMPV switched in consecutive years from genotype A to genotype B (1). Similar phenomena have been observed elsewhere (42). In the study from St. Louis, the severity of illness associated with the two genotypes did not differ. The observed epidemiological features of hMPV are similar to those of RSV, where viruses of both A and B subgroups cocirculate each year and the predominant strains vary from location to location and from year to year (94).

FIG. 4.

Phylogenetic analysis of hMPV. Sequences of the hMPV F genes from isolates originating from Connecticut, The Netherlands, Australia, and Canada were used to construct a phylogenetic tree. Bootstrap values are displayed at major branch points. A representative set of Connecticut sequences is displayed. Genotypes are labeled group A and group B as originally proposed in reference 121. NL/1/00 is a prototype group A strain, and CAN 98-75 is a prototype B strain. The dashed line separates group A and B viruses. (Modified [to reflect the current terminology] from reference 34 with permission. © 2004 by the Infectious Diseases Society of America. All rights reserved.)

In 2004, a variant of hMPV, isolated from a 6 1/2-year-old girl with an acute exacerbation of asthma, was found to be genetically distinct from viruses of the four lineages of hMPV (110). This strain was detected only with primers that targeted the N gene. Whether this strain represents a distinct genotype of hMPV needs to be confirmed with further sequence analysis. Diagnostic genomic amplification methods may have to be modified if there is widespread circulation of this variant. A recent study found that viruses in the A genotype may be more divergent than previously appreciated (51).

Seroepidemiology

Infection with hMPV appears to be common in childhood. Several methods have been used to detect hMPV-specific antibodies in serum, including immunofluorescence using hMPV-infected cells as the antigen (31, 120, 134), enzyme-linked immunosorbent assays using recombinant hMPV proteins (46, 53, 69), and virus neutralization assays (120). Enzyme-linked immunosorbent assays based on recombinant hMPV F or recombinant hMPV N protein are both sensitive and specific for the detection of hMPV-specific antibodies (46, 53, 69). Independent of methodology, several studies have demonstrated that hMPV infection occurs early in childhood (31, 69, 120, 134). By the age of 5 years, >90% of individuals screened have evidence of hMPV infection. The seroprevalence of hMPV-specific antibody in adults is nearly 100% (69, 120). The seroprevalence of hMPV-specific antibody in infants <3 months of age is >90%, indicating that maternally derived antibodies are present in young children (69) (Fig. 5). Whether this hMPV-specific antibody protects against infection or lessens the severity of illness remains to be determined.

FIG. 5.

Seroepidemiology of hMPV. The percentage of hMPV seropositivity and the number of serum samples (n) for each age group are indicated. Error bars represent 95% confidence intervals. (Reprinted from reference 69.)

SEROLOGICAL PROPERTIES OF hMPV GENOTYPES

The observation that the greatest genetic diversity between the two genotypes of hMPV occurs in the structural genes (G > SH ≫ F) suggested that the two genotypes may represent distinct serotypes. To address this issue, animal models have been used. Skiadopoulos et al. used hMPV strains CAN98-75 and CAN97-83, which represent the two genotypes of hMPV, in rodent and nonhuman primate models. Infection with a virus of one hMPV genotype protected animals against infection with a virus of the heterologous genotype. Protection was defined as the inhibition of hMPV replication in nasal turbinates or lungs in infected animals that were previously infected with a strain of hMPV (113). Reciprocal cross-neutralization assays with postinfection serum demonstrated that each strain induced a high level of neutralizing antibody to homologous and heterologous strains. Furthermore, immunization with a recombinant human parainfluenza virus expressing the F protein of CAN97-83 induced antibodies that neutralized viruses from both lineages and protected animals from challenge with either hMPV strain. Taken together, these data suggest that cross-protective immunity occurs after infection with a virus of one genotype and that the F protein is the major antigenic determinant on the virus. Therefore, the two genotypes may not represent distinct serotypes. This is analogous to the case for RSV, in which infection with a virus of either the A or B subgroup elicits an antibody response that is specific for viruses of both the homologous and heteologous subgroups (23). However, van den Hoogen et al., using a ferret model, demonstrated that serum obtained from an animal infected with a representative virus of one genotype could not neutralize a virus of the heterologous genotype in vitro, suggesting that the two genotypes were, in fact, distinct serotypes (121). The antigenicity of hMPV in the natural human host remains to be determined. Does hMPV infection in humans elicit a protective antibody- or cell-mediated response to viruses of the homologous group or the heterologous group? Clearly, the answer to this question will have significant implications for vaccine development.

DIAGNOSIS

hMPV grows poorly in cell culture. The replication of hMPV in vitro is restricted to a limited number of cell lines and requires supplementation of the medium with trypsin for propagation (120). These properties may explain why the virus was not identified until recently. Nonetheless, the sensitivity of cell culture techniques for the detection of hMPV in secretions from the respiratory tract is poor. Currently, RT-PCR is the most common method used to detect hMPV. Several genes of hMPV have been primer targets for genomic amplification. Cote et al. reported that primers that bind regions of the N and L genes are highly sensitive for the detection of hMPV strains of both genotypes (25). Real-time RT-PCR is another powerful method to detect hMPV and other respiratory viruses. Several genes have been targeted for the real-time assays, and certain primers can detect viruses of both subgroups of both genotypes (64, 72, 74, 108).

Immunofluorescence, using virus-specific antibodies, is a rapid method used to detect respiratory viruses and is commonly used in diagnostic laboratories. hMPV-specific antibodies have been developed for immunofluorescence assays, though this method may not be as sensitive as RT-PCR for the detection of hMPV (32, 66, 91). hMPV-specific antibodies for immunofluorescence are available from commercial sources.

VACCINE CANDIDATES AND ANIMAL MODELS

The development of a safe and effective vaccine to protect against hMPV is a reasonable goal. Several promising vaccine candidates have been tested in animal models. A live recombinant human parainfluenza virus that contains the hMPV F gene has been shown to induce hMPV-specific antibodies and to protect experimental animals from hMPV challenge (113). A chimeric bovine/human parainfluenza virus 3 expressing the hMPV F elicits neutralizing antibodies against both parainfluenza virus and hMPV (117). Recombinant hMPVs lacking the G, SH, or M2 gene or both G and SH are attenuated in cell culture, though they retain immunogenicity and elicit protective immunity in experimental animals (8, 11). Antibodies specific for the hMPV F protein are likely induced by these viruses. However, the results of these animal challenge studies should be interpreted cautiously. There are many limitations of the small-animal model for testing potential hMPV vaccines, not the least of which is that the pathogen is highly host restricted.

As stated earlier, hMPV lacking the M2-2 gene is attenuated, though immunogenic, in experimental animals (17). The approach of using live, recombinant, gene-deleted hMPVs as vaccines is based on experience with similar recombinant viruses for the prevention of RSV infection. Live recombinant RSV vaccine candidates containing attenuating mutations and lacking the SH gene are currently in clinical trials (59). An RSV recombinant lacking the G and SH genes has also been studied in human volunteers (58). The preliminary experience with these recombinant RSV vaccines is that they are either overattenuated or underattenuated. Therefore, the safety and efficacy of a recombinant hMPV vaccine may be difficult to predict. Chimeric hMPV/avian pneumovirus recombinant viruses in which the N or P gene of hMPV was replaced with the corresponding gene from avian pneumovirus are attenuated in African green monkeys (the P gene chimera is 10- to 1,000-fold more attenuated than the N chimera), though both are as immunogenic as wild-type hMPV (95). This represents another strategy for the development of recombinant attenuated live hMPV vaccines.

ANTIVIRALS

Other than influenza virus, antiviral therapy for respiratory viruses has not shown tremendous potential. The effectiveness of ribavirin therapy for RSV is limited, at best, and remains a controversial issue (3). The relative ineffectiveness of antivirals for RSV may be due, in part, to the inflammatory reaction elicited in the host, which in all likelihood would not be squelched by antiviral compounds. It is unknown what role the host's inflammatory response plays during hMPV disease. Nonetheless, antiviral compounds have been tested with hMPV. The antiviral activity of ribavirin to inhibit the replication of hMPV is equivalent to that observed with RSV (136). Other compounds, such as NMSO3, a sulfated sialyl lipid that has been shown to have potent antiviral activity against RSV in tissue culture cells, have been shown to have anti-hMPV activity in vitro (137). It is likely that an hMPV-neutralizing monoclonal antibody for prophylaxis of high-risk infants (similar to the anti-RSV F humanized monoclonal antibody currently used for prevention of severe RSV disease) will be developed and tested. The progress towards an effective antiviral strategy for hMPV is currently limited by the scant data on pathogenesis of the virus in the natural host.

CONCLUSIONS

Since the discovery of hMPV in 2001, the virus has been identified worldwide. hMPV is a common respiratory pathogen, particularly in infants and young children. The virus is associated with both upper and lower respiratory tract infections and may be a trigger for asthma. At least two major genotype of hMPV circulate during community outbreaks. Whether these genotypes represent distinct serotypes remains controversial. The major challenges faced by the medical and scientific communities are the understanding of the pathogenesis of hMPV disease and the development of a safe and effective vaccine to protect against infection and disease caused by this newly recognized respiratory virus.