In this study, we employed a pan-viral microarray assay, the Virochip, to identify a novel adenovirus associated with a fulminant pneumonia outbreak in a colony of New World titi monkeys. Despite the absence of an animal model, which precludes a strict fulfillment of Koch's postulates, there are several lines of evidence implicating this novel adenovirus, TMAdV, as the cause of the outbreak. First, conventional testing for other pathogens, including other viruses by Virochip, was negative, and affected monkeys did not respond to empiric therapy with antibiotics or antivirals (ribavirin and oseltamivir in anecdotal use are not effective against adenoviral infections) 
. Second, the clinical presentation of pneumonia and hepatitis is consistent with the known spectrum of disease associated with adenoviral infections. Third, TMAdV sequence was recovered by PCR in various body fluids and tissues from affected monkeys, including blood, respiratory secretions, and lung/liver tissue (). Fourth, the finding of intranuclear inclusions in diseased tissues, as well as direct visualization of adenoviral-like particles (TMAdV) in lung alveoli by electron microscopy ( to ), support a primary role for TMAdV in the pathogenesis of tissue injury in affected monkeys. Finally, there was a significant neutralizing Ab response in surviving animals, with 2 monkeys having titers undetectable prior to the outbreak but rising to >1
512 at convalescence ().
Although TMAdV retains the core genomic features common to all adenoviruses (), phylogenetic analysis clearly places TMAdV within a separate branch, with no closely related neighbors ( and S1
). A phylogenetic distance of >10% combined with the lack of cross-neutralization defines TMAdV as a new species 
. Since emerging adenovirus strains such as HAdV-14 and HAdV-D22/H8 (otherwise known as HAdV-D53) are known to arise from recombination events among related ancestral strains 
, we performed bootscanning analysis to look for such events in TMAdV. The bootscanning analysis, however, failed to show evidence of recombination, likely because closely related and/or ancestral strains to TMAdV have not yet been identified.
Entry of adenoviruses into cells involves an initial attachment of the fiber knob to the cell receptor, followed by internalization via a secondary interaction of the penton base with αv
. The presence of an RGD motif in the TMAdV penton base implies that the virus uses αv
integrins for internalization 
. However, the high sequence divergence in the fiber protein (), as well as the absence of fiber motifs conserved among adenoviruses that bind CAR 
(coxsackievirus-adenovirus receptor) or CD46 
(data not shown), suggest that neither of these two human adenoviral receptors may be the attachment receptor for TMAdV. Further studies will be necessary to identify the preferred cellular attachment and internalization receptors for TMAdV.
Despite its isolation from affected titi monkeys, we were unable to propagate TMAdV in both established (BSC-1) and primary (PMK) monkey kidney cells (). The virus, however, grew efficiently in a human A549 lung adenocarcinoma cell line. One explanation for this finding is that TMAdV may be unable to productively infect cells derived from Old World monkeys (e.g. rhesus and African green monkeys). An alternative possibility is that successful propagation of TMAdV may depend on infection of a specific host cell type, such as A549 lung, and not BSC-1 or PMK kidney cells. Nevertheless, after 10 passages in human A549 cells, the fully adapted strain of TMAdV exhibits an extended host range with the ability to productively infect both monkey and human cells. This observation implies that TMAdV possesses an inherent capacity to cross the species barrier and infect both humans and nonhuman primates. Efforts to identify host range and cell tropism of TMAdV, as well as the specific sequence changes responsible for adaptation to growth in cell culture, are currently underway.
The virulence of TMAdV in healthy and apparently immunocompetent titi monkeys (83% case fatality rate) is highly unusual for infections by adenovirus. In humans, deaths due to adenovirus infections or outbreaks are generally low (up to 18% for pneumonia associated with HAdV-14 
). Furthermore, severe infections from human adenoviruses are more commonly associated with older age, immunosuppression, and chronic underlying conditions such as kidney failure 
. Young, healthy individuals are in general much less likely to succumb to adenoviral-related illness. The severity of TMAdV-related illness in affected titi monkeys suggests that this species of monkey may not be the natural host for the virus. The failure to detect fecal shedding of TMAdV in convalescent or asymptomatic animals also suggests that the virus does not normally infect titi monkeys ().
Although the exact origin of TMAdV remains unclear, we can speculate on several possibilities. One possibility is that a cross-species “jump” from captive macaques to a susceptible colony of titi monkeys precipitated the outbreak. As there have been no new introductions of monkeys into the closed colony for the past 2 years, this conjecture relies on asymptomatic infection and transmission of TMAdV in the captive rhesus/cynomolgus macaque population at the CNPRC. CNPRC personnel who visited macaque areas would occasionally enter titi rooms with no change in personal protective equipment, thus providing a potential route of transmission for the virus. In addition, specific antibodies were detected in 1 of 36 (2.8%) asymptomatic rhesus macaques housed in the same building (), indicating that TMAdV has the capacity to infect this species of Old World monkey. Notably, the closest identified phylogenetic relative to TMAdV among the complete genomic sequences available in GenBank is a rhesus monkey adenovirus, SAdV-3 (; ). Furthermore, serological evidence for cross-species adenoviral transmission events between different nonhuman primate species has been previously reported in the literature 
Although we failed to detect TMAdV in rodent droppings found near titi monkey cages (), it is still possible that the virus arose from an unknown animal reservoir. In this regard, the high sequence divergence of TMAdV relative to the known human/simian adenoviruses (), and comparable sequence similarity in the polymerase gene to a porcine adenovirus ( and S1
) are striking. The four-week interval between the index case and the second case appears overly long given a typical incubation period for adenovirus infections of no more than 1 week 
. This may be explained by our finding of a high rate of subclinical infection by TMAdV in asymptomatic titi monkeys (35%), but may also be due to separate introductions of TMAdV into the colony from an as-yet unidentified reservoir.
Our study data also support the potential for cross-species transmission of TMAdV between monkeys and humans. The researcher's fever, cough, and pleuritic symptoms (“burning sensation in the lungs”) are consistent with the development of a prolonged viral respiratory illness. Interestingly, pleurisy has been specifically reported in association with certain human adenovirus infections 
. The clinical presentation, time of illness concurrent with the onset of the outbreak, and presence of neutralizing Abs in convalescent serum all strongly point to primary infection of the researcher by TMAdV. The detection of weakly neutralizing Abs (1
8) in a serum sample from a sick family member of the researcher also suggests that TMAdV may be capable of human-to-human transmission. The decreased levels of neutralizing Abs to TMAdV in the researcher (1
32) and a family member (1
8) relative to those in infected titi monkeys (up to >1
512) are consistent with a recent study showing much higher levels of neutralizing antibodies in chimpanzees than in humans with adenovirus infections, possibly due to more robust adenovirus-specific T-cell responses in humans than in monkeys 
Several lines of evidence support the contention that the direction of TMAdV transmission was zoonotic (monkeys to humans) rather than anthroponotic (humans to monkeys). First, the closest known relative to TMAdV in GenBank is SAdV-3, an Old World monkey adenovirus (; ). Second, our results show that PCR assays for human adenoviruses in common use are capable of detecting TMAdV. Although sequencing of PCR amplicons for human adenoviruses is not performed routinely in diagnostic virology, TMAdV would presumably have been detected previously in large-scale studies of hexon sequencing of Ad field isolates if it were circulating in the community 
. Finally, the available sequence data in GenBank is heavily biased towards human adenoviruses, and much less is known about the potential diversity of the simian adenoviruses. We also cannot formally exclude the possibility that the outbreak arose from anthroponotic transmission. In our study, 2 of 81, or 2.5% of random adult blood donors exhibited borderline titers of neutralizing antibody to TMAdV, indicating either a low prevalence of TMAdV in the human population or cross-reactivity to a related virus (although no evidence of cross-reactivity was found with HAdV serotypes 1 through 35). Future large-scale studies of TMAdV seroepidemiology will be needed to better understand transmission of TMAdV between monkeys and humans. Nevertheless, our discovery of TMAdV, a novel adenovirus with the capacity to cross species barriers, highlights the need to monitor adenoviruses closely for outbreak or even pandemic potential.