Evidence for Human Streptococcus pneumoniae in wild and captive chimpanzees: A potential threat to wild populations

Habituation of wild great apes for tourism and research has had a significant positive effect on the conservation of these species. However, risks associated with such activities have been identified, specifically the transmission of human respiratory viruses to wild great apes, causing high morbidity and, occasionally, mortality. Here, we investigate the source of bacterial-viral co-infections in wild and captive chimpanzee communities in the course of several respiratory disease outbreaks. Molecular analyses showed that human respiratory syncytial viruses (HRSV) and human metapneumoviruses (HMPV) were involved in the etiology of the disease. In addition our analysis provide evidence for coinfection with Streptococcus (S.) pneumoniae. Characterisation of isolates from wild chimpanzees point towards a human origin of these bacteria. Transmission of these bacteria is of concern because – in contrast to HRSV and HMPV - S. pneumoniae can become part of the nasopharyngeal flora, contributing to the severity of respiratory disease progression. Furthermore these bacteria have the potential to spread to other individuals in the community and ultimately into the population. Targeted vaccination programs could be used to vaccinate habituated great apes but also human populations around great ape habitats, bringing health benefits to both humans and wild great apes.

Due to their close genetic relationship, humans and other great apes share susceptibility to a considerable number of pathogens 1,2 with the transmission of human respiratory pathogens to captive apes being a well-documented concern 3-11 . In this context, it had been long suspected that pathogen transmission from humans might account for disease outbreaks observed in wild great apes habituated to human presence for research and tourism 7,[12][13][14] . This suspicion was confirmed when common human respiratory paramyxoviruses (human respiratory syncytial virus (HRSV) and human metapneumovirus (HMPV)) were identified in chimpanzee (Pan troglodytes verus) lung tissue and faeces collected during separate disease outbreaks in Taï National Park (Côte d'Ivoire) 15,16 . Further studies on wild chimpanzees in Tanzania (Pan troglodytes schweinfurthii) and gorillas (Gorilla beringei) in Rwanda and the Central African Republic have demonstrated additional incidences in which human respiratory paramyxoviruses have been transmitted to great apes indicating that these events are neither isolated nor a feature of a particular region (Table 1) [17][18][19] .
In addition to human viruses, the bacterium Streptococcus pneumoniae (or pneumococcus) was also detected during outbreaks where samples of lung tissue had been collected 15,18,20 , leading to the conclusion that it was the combination of respiratory viruses and pneumococci, and not the viral infection alone, that caused the fatalities 15,16 . Given that invasive pneumococcal infections are also a primary cause of severe pneumonia in humans worldwide 21,22 , these bacterial co-infections are a cause for concern since they could lead to increased mortality and persist in the great ape populations.
However the provenance of the pneumococci strains from the great ape cases mentioned above remains uncertain. Sequence analyses have so far only been performed on S. pneumoniae positive chimpanzee samples from Taï National Park and these showed differences to known human sequences 20,23 . Unfortunately, there is a lack of data on regional human pneumococci sequence types and therefore transmission from humans could not be confirmed 20 .
In this article we investigate respiratory disease outbreaks causing mortalities among wild and captive wild-borne great apes in West and western-Central Africa. Again, HRSV and HMPV were detected together with pneumococci and we provide evidence that all were of human origin. These results are important since, in contrast to human respiratory viruses, pneumococci may become an established part of the great ape oropharyngeal flora and could severely impact the outcome of subsequent respiratory infections or, as in humans, become the primary cause of severe pneumonia.

Results and Discussion
Samples were collected in the course of several respiratory disease outbreaks that occurred at two different locations: in the Limbe Wildlife Center (LWC) situated in Cameroon, and in Taï National Park (TNP), Côte d'Ivoire.
The LWC is a wildlife rehabilitation project in which rescued great apes are kept in naturalistic enclosures. Here, a total of five distinct outbreaks of respiratory disease occurred over a period of two years (2006/2007), in three separate chimpanzee groups (see Table 1): group 1 consisted of 22 individuals (age >5years), group 2 of 14 individuals between two and five years of age and group 3 consisted of four individuals of less than two years of age. The chimpanzees showed varying clinical symptoms (coughing, serous to mucopurulent nasal discharge, lethargy, inappetence) over the course of the outbreaks. Symptoms indicative of lower respiratory tract infection, such as dyspnoea and wheezing sounds, were noted on auscultation of the lungs of 12 individuals. Duration of the outbreaks ranged from several days up to nine weeks. Morbidity ranged from 23 to 100% (with 100% in three of the five outbreaks; see Table 1). During these outbreaks pharyngeal swabs from six chimpanzees were taken under anesthesia. During one outbreak an infant chimpanzee (referred to as LWC infant) died, a necropsy was performed and tissue samples were taken. Throughout the outbreaks, animals with severe symptoms were placed on antiobiotic therapy (see Supplementary Information).
In TNP, wild chimpanzees are habituated to the presence of human observers and individually known as a result of a project focusing on wild chimpanzee behaviour 24 . Here, three chimpanzee communities (so called North, East and South community) are under observation. In November 2009 the South community was affected by a respiratory disease outbreak. Morbidity and mortality were high: 32/37 chimpanzees developed severe respiratory symptoms (including high frequency of heavy cough, lethargy, nasal discharge). Four adults (>15 years; hereinafter referred to as TNP adult 1-4) and two infants (0-5 years, referred to as TNP infant 1 and 2) were found shortly after death and necropsies were performed. In addition, for one chimpanzee (adult) only one arm and the head was found due to scavenging. Four infants were found dead but necropsies were not performed and three further adults disappeared in the course of the outbreak and are likely victims of the outbreak as well. In total likely 8 adults and 6 infant chimpanzees died in this outbreak. Twelve out of the 37 individuals who showed severe symptoms of respiratory disease were treated with Extencilline (Benzathine benzylpenicillin, Sanofi-Aventis France), a long acting antibiotic administered by remote injection. Of these nine survived.
Histopathology. Histopathological examination of lung tissue samples from TNP and LWC chimpanzees revealed a severe purulent bronchopneumonia as cause of disease. The bronchoalveolar tissue was diffusely infiltrated with neutrophilic granulocytes and macrophages. Furthermore, all lung tissue samples collected during the TNP outbreak showed multinucleated cell syncytia, pointing towards an HRSV infection (Fig. 1). Additionally, gram-positive cocci could be demonstrated within the altered tissue giving evidence for a secondary bacterial infection. In addition, the lung tissue of one animal (TNP adult 3) revealed signs of a lung mite infection (likely Pneumonyssus simicola). The lung tissue from the LWC infant showed a multifocal to diffuse purulent bronchopneumonia with marginal alveolar histiocytosis. The periodic acid schiff (PAS) reaction was negative for all samples excluding a fungal infection.

Molecular analysis.
All samples collected from the TNP and LWC outbreaks were analysed molecularly by screening PCR and tested positive for either HMPV-or HRSV-specific genes as well as for the lytA gene harboured by S. pneumoniae (see Table 1). Phylogenetic analyses of generated paramyxovirus sequences revealed that both HRSV and HMPV sequences (Genbank accession no. JX489496-JX489498) belonged to recent pandemic waves (see Supplementary Figs 1-3). Informing a relaxed molecular clock model with tip dates 25 we estimated that the LWC chimpanzee-infecting HMPV diverged from its closest human-infecting relatives in 2004 (95% highest posterior density: 2003-2004). The date of divergence between the HRSV-A strain from TNP chimpanzees and humans was 2007 (95% highest posterior density: 2006-2008) and for HRSV-B strains infecting LWC chimpanzees and humans 2007 (95% highest posterior density: 2007) These results mirror findings from previous outbreaks [15][16][17][18][19] and emphasize the high incidence of anthropozoonotic transmission of human respiratory paramyxoviruses.
In order to gain new insights on the bacterial community present in the lung samples collected during the outbreak in TNP, we performed 16S rRNA PCR. Characterization of the sequences using a BLAST search in GenBank indicated the presence of Streptococcus species in all samples analysed. In addition, for one individual (TNP adult 3) Pasteurella (P.) multocida was also among the predominant BLAST hits. P. multocida is a SCIEnTIfIC REPORtS | 7:14581 | DOI:10.1038/s41598-017-14769-z commensal in the upper respiratory tracts of many domestic and wild animal species worldwide and coinfection with P. multocida and S. pneumoniae among TNP chimpanzees has been reported previously 16,20,26 . In addition, for the samples from TNP adult 1, the BLAST search revealed Clostridium sp. Since most species of the genus Clostridium are saprophytic, we conclude that this finding was due to the tissue samples being partly decayed.
Molecular analysis of the S. pneumoniae is considered separately in the next section.
Phenotypic and molecular characterisation of S. pneumoniae. Isolation of pneumococci was achieved for TNP adult 1 and 2 (isolate IMT 113 and 115) from whom lung tissue samples had been stored adequately in Skim Milk-Tryptone-Glucose-Glycerol (STGG). Antibiotic susceptibility testing (Table 2) showed low resistance: isolates were resistant against sulfamethoxazole/trimethoprim (TMP-SMX) only. Further characterisation by Multi Locus Sequence Typing (MLST, see Table 3) indicated that alleles from isolates termed IMT 113 and 115 were identical and matched known alleles from human strains. Their allelic profile revealed 6 matches to various published serotype 19 A strains from Kenya and the Gambia (>100 isolates) but harbour a variant ddl allele. Therefore a new sequence type (ST 8485) was assigned (see Table 2). Serotyping of the TNP samples by PCR also revealed capsular type 19 A. Lung samples from the LWC infant had not been stored adequately for cultivation. Therefore, a full DNA extract of this sample was used for further characterization. Hence, MLST results are based on "virtual clones" (see methods). For the LWC sample all seven alleles matched known alleles from human strains and indeed, allelic profiles were identical to a serotype 9 V strain isolated from a human of African origin (see Table 2). DNA from the LWC sample was tested for capsular type 9 V 27 but tested negative. In addition we tried another 9 V/9 A primer combination described by Carvalho et al. 28 but again the DNA tested negative. Unfortunately due to lack of DNA material it was not possible to test for further serotypes.
In summary, the MLST profiles of the strains described here showed high similarities or were even identical to human isolates, especially in Africa. In addition, results from the antibiotic susceptibility testing ( Table 2) showed that the isolates from TNP chimpanzees were resistant to TMP-SMX. Resistance to TMP-SMX, which has been used widely in developing countries, is observed frequently and globally in human strains of S. pneumoniae 29 . The fact that this drug has never been used in the chimpanzees of TNP supports the assumption that the S. pneumoniae strain found originated in humans.
S. pneumoniae infection, especially in combination with respiratory viruses, can lead to invasive lower respiratory disease with fatal outcomes. Furthermore, and in contrast to respiratory virus infections, which are usually temporary, S. pneumonia infection in humans can result in an asymptomatic colonization of the nasopharynx. Additionally this bacterium can spread horizontally and remain silent within a population 30 . In fact, pneumococcal disease in humans will not occur without preceding nasopharyngeal colonization 30 , hence an increased rate of colonization within a population will increase the incidence of invasive pneumococcal disease. To investigate

Conclusion
Here we provide the first evidence that S. pneumoniae can be transmitted and that the origin of this pathogen might be human. S. pneumoniae has the potential to circulate over long periods of time in a given population (a finding supported by our data on the chimpanzee XENA) and may cause severe disease once an individual is weakened by further primary viral infections or through other circumstances 30,32 . Given that adolescent female chimpanzees migrate to other communities and that other inter-troop encounters occur, the reported findings could have serious implications, not only in respect to the clinical outcome of individuals infected, but also for populations of wild chimpanzees 33 . In addition, the risk of ex-captive chimps transmitting human S. pneumoniae into wild populations is a serious concern for reintroduction programs.
These results underline the importance of hygiene measures, as established in TNP and described in the new IUCN guidelines, in limiting the risk of pathogen transmission between humans and great apes 34 . Moreover these results underscore the importance of developing non-invasive sample collection methods to assess the risk of the spread of such pathogens to wild chimpanzee and other great ape communities. We note that, characterisation of S. pneumoniae present in human communities in contact to or neighbouring wild great ape populations are needed to allow further conclusions regarding naturally occurring, and introduced, infections. Given the availability of vaccines that provide protection against the serotypes most frequently associated with pneumococcal disease in humans, characterisation of the distribution of serotypes would be of great interest. Such data will inform the design of vaccination campaigns targeting human-habituated great apes but also human populations in contact with wild great apes, providing enormous benefits to both human and wild animal health.

Material and Methods
Sampling. Necropsies on chimpanzees were conducted under high safety standards and precautions such as protection suits, gloves, and face masks were used to avoid contamination of samples with human pathogens. At LWC pharyngeal swabs were collected from six animals under sedation. All samples were preserved in RNAlater (LWC) or liquid nitrogen (TNP) and shipped to Germany for detailed analyses. No experiments on live animals were performed for this study, samples were collected from animals found dead or sedated for treatment necessary to save their lives, and therefore no animal ethics permission is required according to the national laws of Ivory Coast and Cameroon.
Histology. Tissue samples from five of the six deceased chimpanzees from TNP on which necropsies had been performed, and from the deceased chimpanzee from LWC were submitted to the German Primate Centre (Göttingen, Germany) for histopathological examination. All samples were fixed in 4% formalin, embedded in paraffin, cut in 2-4 µm tissue sections and stained with hematoxylin and eosin for microscopic examination. Tissue was also stained with Giemsa and Gram stain and to visualize fungal infection furthermore PAS reaction was performed.   PCR screening for respiratory pathogens and phylogenetic analysis. DNA or cDNA respectively were screened for influenza A&B, HMPV, HRSV, Adeno-, Picornavirus and S. pneumoniae using PCR assays as described previously [35][36][37][38][39][40] . For phylogenetic analysis, the G gene of HMPV or HRSV was amplified using published PCR assays 17,41 PCR products were purified using ExoSAP (USB Europe GmBH, Staufen, Germany) and sequenced subsequently using the ABI Big Dye Termination Kit (Applied Biosystems, Weiterstadt, Germany). Phylogenetic analyses of the paramyxovirus sequences were performed in a Bayesian framework. First, data sets were assembled that comprised one exemplary sequence per outbreak (-year) generated in this study (see supplementary information for details) and human HMPV or HRSV G gene sequences extracted from publicly availably full genomes with known isolation dates (retrieved from NCBI). Both data sets were aligned using MUSCLE 42     Amplicons from the TNP isolates were purified directly using ExoSAP (USB Europe GmBH, Staufen, Germany) and sequenced using the ABI Big Dye Termination Kit (Applied Biosystems, Weiterstadt, Germany). For the sample from LWC (DNA from lung tissue instead of a pure culture) MLST results were generated based on "virtual clones": to minimize the possibility that two or multiple pneumococcus strains were involved, each of the seven amplicons was cloned into a pCR ™ 2.1-TOPO ® vector (Invitrogen, Karlsruhe, Germany). For every cloning reaction, ten clones were picked and colony PCR was performed using M13 primers. PCR products were purified and sequenced as described above. For every allele, sequences were checked for consistency using Genious version 7.1.4 (http://www.geneious.com) 51 .
Sequences obtained were entered in the MLST database and allelic numbers were assigned (all alleles were already present in the database). Since the TNP isolates showed a new allelic combination i.e. differed in one locus (ddl), a new ST was assigned by the curator (Table 2).
Serotyping PCR. Serotyping was done by PCR according to Pai et al. 27 . However, due to a shortage of DNA material, we did not follow the sequential multiplex PCR assay described, but rather chose serotype specific primers according to the serotype information given for human strains in the MLST database that matched our clones the most. DNA from TNP bacterial isolates were tested using the primers specific for capsule type 19 A. DNA from the LWC lung sample was tested for serotype 9 V but tested negative. In addition a 9 V/9 A primer combination described by Carvalho et al. 28 was tried however the DNA tested negative. Further serotype testing was not possible due to a lack of DNA. Data availability. The datasets generated and analysed during the current study are available from the corresponding author on request.
Accessions code. Sequences data on the HRSV and HMPV strains are available in Genbank under the accession number JX489496 -JX489498.