In this issue of the Journal, Couroucli et al. report the results of a technique based on the polymerase chain reaction (PCR) to screen tracheal aspirates from infants for nucleic acids specific for several infectious agents (1). Comparison of samples from patients who developed bronchopulmonary dysplasia (BPD) to controls showed no difference for nucleic acid markers for most organisms including some that were previously suspected of being associated with BPD. Interestingly, they did find that a target sequence of the adenoviral capsid gene was present in increased frequency in BPD patients. Although this finding raises a suspicion that the adenovirus may be involved in the pathogenesis of BPD, much more information is required to establish its exact role. To prove that adenoviral infection was present in these infants, culture is required to prove viral replication, electron microscopy to prove assembly of complete viral particles, in situ hybridization and/or immunohistochemistry to establish the type of cells infected, and a host response to the adenovirus must be demonstrated. Obviously, this is too much to expect from a preliminary study designed to screen for a large number of infectious agents and the authors should be congratulated for introducing an hypothesis suggesting that adenoviral infection may be involved in the pathogenesis of BPD. Indeed, their findings should provide the stimulation required for further investigation of this problem in many laboratories.

The adenovirus is a double-stranded DNA virus that was first isolated from primary cell cultures of human adenoids (2) and airway secretions during an epidemic of respiratory disease in military recruits (3). The adenovirus family is known to consist of more than 40 individual serotypes that are responsible for a wide range of infections in mammalian and avian hosts. The discovery that human adenovirus 12 produced malignant tumors in newborn hamsters sparked an early and sustained interest in the molecular biology of the adenovirus and this viral particle was among the first to have its genome sequenced (4). Studies of more than 100 individual isolates show that the genome has the same general features with two sites of replication, five early transcription units, two delayed late units, and one major late unit that is processed to produce five different families of mRNA. Adenoviral replication is dependent on transcriptional activation, which is induced by the protein produced by one of the early transcription units referred to as the E1A gene (5). The E1A protein has the ability to bind to specific proteins in the host transcriptional machinery and dramatically increases the rate of transcription of both host and viral genes (6).

Adenoviral DNA has been shown to persist in adenoids (7), peripheral blood lymphocytes (8), and adult lung tissue (9). During this form of latent infection, viral DNA remains in the infected tissue, in which it can generate individual viral proteins but not a complete virus. In adult human lungs, the adenoviral E1A gene has been demonstrated in airway epithelial cells (9), in which it produces E1A protein (10). An animal model of this form of latent infection is associated with a residual bronchiolitis (11), and these animals show an excessive inflammatory response to a single dose of cigarette smoke (12). Transfection of airway epithelial cells with E1A results in an excessive production of IL-8 (13) and ICAM-1 (14) following challenge. Collectively, these data have been used to support the concept that latent adenoviral infection amplifies cigarette smoke-induced lung inflammation and provides a partial explanation for the epidemiologic fact that everyone who smokes develops airways inflammation whereas only 10–15% of heavy smokers develop airways obstruction (15). Whether latent adenoviral infection is also able to up-regulate the inflammatory response that underlies the pathogenesis of BPD remains to be determined.

Couroucli et al. 's report establishes that a portion of the gene for the hexon protein in the viral coat is present in excess in tracheal aspirates from children with BPD and suggest the hypothesis that it may have a pathogenetic role. As the hexon gene has no transactivating properties, it will be important to determine whether transactivating genes such as E1A and its protein are also present in the lungs of BPD patients. Studies based on animal models and/or transfected cell lines might then be used to determine whether the relationship between adenoviral DNA in tracheal aspirates and the presence of BPD in these patients is causal or coincidental.