Mimiviridae, Marseilleviridae, and virophages as emerging human pathogens causing healthcare-associated infections.

AIM: During the last decade it became obvious that viruses belonging to Mimiviridae and Marseilleviridae families (order Megavirales), may be potential causative agents of pneumonia. Thus, we have performed a review of the association of Mimiviridae, Marseilleviridae, and virophages with pneumonia, particularly healthcare-associated pneumonia, and other infections of the respiratory tract. RESULTS AND DISCUSSION: According to the analysis of the published articles, viruses belonging to Mimiviridae family can be potential agents of both community-acquired and healthcare-associated pneumonia. In particular, these viruses may be associated with poor outcome in patients of intensive care units. The exact mechanism of their pathogenicity, however, still remains unclear. The discrepancies between the results obtained by serological and genomic methods could be explained by the high polymorphism of nucleotide sequences of Mimiviridae family representatives. Further investigations on the Mimiviridae pathogenicity and on the determination of Mimiviridae-caused pneumonia risk groups are required. However, the pathogenicity of the viruses belonging to Marseilleviridae family and virophages is unclear up to now.

Introduction

The existence of viruses with extremely large particle and genome sizes has been predicted since the discovery of jumbo bacteriophages in the 1970s and the phycodnaviruses in the early 1980s [1], [2]. An investigation of pneumonia outbreak with no clear causative agent in Bradford, England (1992) by Tim Rowbotham and his team considered water of a cooling tower as a probable origin of the outbreak and amoebae as potential culprits [3] since they were established as Trojan horses due to their ability to harbor multiple agents of human pneumonia that can survive and multiply under the protection from various external physical and chemical agents within the amoebae encyst [4], [5]. This led to an isolation of intra-amoebal pathogenic bacteria, one of which had the appearance of a Gram-positive coccus and was therefore named Bradford coccus [3], [6], [7], [8], [9]. However, it resisted the PCR amplification and 16S ribosomal DNA sequencing. This led to the observation of Bradford coccus using electron microscopy, which revealed that it had a viral icosahedral structure. The viral nature of Bradford coccus was further confirmed by an eclipse phase during its replication cycle [6], and genome sequencing [10]. In 2003, the Bradford coccus was renamed as Acanthamoeba polyphaga Mimivirus (APMV) due to its mimicry of a bacterium by its size and appearance by Gram staining [3]. Raoult et al. [3], [10] reported that APMV genome was the largest among all of viruses (1,181 kb), encoding more than 900 proteins, including those never identified previously in viruses. Overall, the Mimivirus discovery has resulted in a considerable shift in our understanding of the definition, origin, and evolution of viruses [11], [12]. Mimiviruses were classified into a separate group of viruses called nucleocytoplasmic large DNA viruses (NCLDVs) which was first described in 2001 [13] and included the families of Poxviridae, Asfarviridae, Iridoviridae, Ascoviridae, and Phycodnaviridae [14], [15], [16]. In 2007, a first member of the Marseilleviridae family also related to NCLDVs, Acanthamoeba polyphaga marseillevirus (APMaV), was isolated from water collected from a cooling tower in Paris, France, using a method based on Acanthamoeba polyphaga culture [17]. This virus was named in honor of its amoebal host and of the name of the French city, Marseille, where it was discovered [17]. The Marseillevirus was characterized by a 368-kb genome, 457 genes, and a minimum of 49 proteins [17]. Furthermore, the first member of the second branch of Mimiviridae family, Cafeteria roenbergensis virus (CroV), was described in 2010 as an agent infecting marine zooplankton, and its genome consisted of ≈730 kb of double-stranded DNA [18]. Recently, Colson et al. [19], [20] proposed assigning an official taxonomic rank to the NCLDVs as the order Megavirales, due to the large size of the virions and genomes of these viruses, and because of their common ancestral origin [13], [15], [21], [22]. Families and genus belonging to this new tentative order are summarized in Table 1 (Tab. 1), whilst a brief timeline of discoveries in this field is presented in Table 2 (Tab. 2). During the last decade, plenty of other giant virus strains belonging to Mimiviridae and Marseilleviridae families were discovered [23], and their diversity is reflected in Table 3 (Tab. 3) and Table 4 (Tab. 4).

Table 1

Members of the proposed order Megavirales

Table 2

A brief timeline of discoveries regarding the nucleocytoplasmic large DNA viruses (NCLDVs) superfamily and the Megavirales order

Table 3

Known members of the Mimiviridae family

Table 4

Known members of the Marseilleviridae family

For decades, viruses were considered to be a final level of parasitism in living nature. However, in 2008, La Scola et al. [24] described an icosahedral small virus Sputnik (50 nm in size) which was associated with an APMV. Sputnik did not replicate in Acanthamoeba castellanii but demonstrated a rapid growth in the giant virus factory revealed in amoebae co-infected with APMV. The most incredible fact was that Sputnik life cycle was harmful for APMV and led to the production of abortive forms and abnormal capsid assembly of the host virus. The authors suggested that Sputnik belongs to a new family of viruses and classified it as a virophage [24]. In 2011, Fischer and Suttle [25] described a new virophage called Mavirus, and Yau et al. [26] almost simultaneously reported a discovery of Organic Lake Virophage. Eighteen months later, Desnues et al. [27] revealed a new Sputnik strain, which had 99% identity to the original Sputnik virophage genome sequence. A recent comprehensive investigation carried out by Zhou et al. [6] suggested an existence of five new virophages, namely Yellowstone Lake Virophages 1, 2, 3, 4, and Ace Lake Mavirus. Finally, third Sputnik strain was described in 2013 by Gaia et al. [7]. The discovery of virophages revolutionized our understanding of host-parasite interrelations. Features of known virophages are presented in Table 5 (Tab. 5).

Table 5

Features of known virophages

There is a significant lack of knowledge regarding the biology of virophages. Obviously, they have greater importance in living nature then we can suggest at the moment. Virophages perform a control of non-viral host-virus dynamics as the essential regulators of ecological interrelations [26], participate in a transfer of genetic information between various organisms as the mobile genetic elements [25], particularly in an interviral gene transfer [27], may be the pathogens of the multicellular organisms [28], and can possibly be used as a new way to fight with the emerging viral infections. Bacteriophages are being successfully used against bacteria in a clinical practice; presumably, a similar pattern does not seem to be impossible in the case with virophages and human pathogenic viruses. Further research on the phenomenon of virophages will definitely shed light on their role in living nature.

It was suspected that mimiviruses may be potential causative agents of pneumonia due to the setting of its initial discovery and to the involvement of some water-associated amoebae-resistant bacteria, including Legionella pneumophila, in such infections [3], [4]. Experimentally, Mimivirus was found to be capable of inducing pneumonia in mice [29] and infecting macrophages through phagocytosis [30]. A number of case reports showing the pathogenicity of viruses belonging to Mimiviridae and Marseilleviridae families in humans during the last decade (Table 6 (Tab. 6)) [31], [32], [33], [28], [34], [35], [36], [37], [38]). Further, certain reports demonstrated that asymptomatic Marseillevirus infection is not rare in healthy persons and may be transmitted by transfusion [39]. In addition, the results of previous studies identified sequences related to members of families of the order Megavirales in human blood, nasopharyngeal samples, or stools [40], [41], [42], [43], [44], [45], [46].

Table 6

Cases of human infections caused by giant viruses

Thus, we aimed to perform a review of the association of Mimiviridae, Marseilleviridae, and virophages with pneumonia, particularly healthcare-associated pneumonia (HAP), and other infections of the respiratory tract.

Materials and methods

To the best of our knowledge, all relevant articles published before December of 2013 and available in PubMed database were included in this review. The generation of search queries was performed by combination of words placing at certain positions in the structure of the query, and all feasible variants were browsed:

First position: “mimivirus”, “mimiviruses”, “mimiviridae”, “marseillevirus”, “marseilleviruses”, “marseilleviridae”, “giant virus”, “giant viruses”, “nucleocytoplasmic large DNA virus”, “nucleocytoplasmic large DNA viruses”, or “megavirales”.

Second position: “pneumonia”, “nosocomial”, “community-acquired”, “hospital-acquired”, “ventilator-associated”, “healthcare-associated”, or “respiratory”.

Reference lists of all relevant articles were also screened for the papers which could elude from our search. According to the results of the search, we identified 11 relevant articles in which epidemiological investigations were performed.

Results and discussion

Unfortunately, it was not possible to carry out a meta-analysis since due to significant differences in study design, sample types, and methods of detection used in distinct studies. Thus, we performed only the qualitative comparative analysis.

The results of the first investigation on the association of APMV with pneumonia were published in 2005 by La Scola et al. [47]. The researchers collected serum samples from 376 Canadian patients with community-acquired pneumonia (CAP, 121 ambulatory and 255 hospitalized) and 511 healthy control subjects. Microimmunofluorescence assay revealed that patients with CAP were mimivirus-positive significantly more frequently compared to controls (9.66% vs. 2.30%, P<0.01) (Table 7 (Tab. 7)). Furthermore, the authors identified hospitalization from a nursing home and rehospitalization after discharge as independent risk factors of mimivirus-associated CAP (P<0.05), possibly due to poor efficacy of standard antimicrobial agents against viruses. In addition, immunoelectron microscopy revealed that antibodies of APMV-positive patients specifically recognized mature APMV particles whilst antibody fixation was not found in serum samples from APMV-negative patients. The authors suggested APMV as a particularly hazardous etiologic agent of pneumonia acquired in institutions [47], [48]. In addition, the investigators recruited a second study sample consisted of 26 French patients with intensive care unit (ICU)-acquired pneumonia and 50 healthy controls, showing that APMV can be found in patients with ICU-acquired pneumonia and multiple populations [47]. Moreover, serological positivity to APMV was associated with a higher risk of ICU-acquired pneumonia (19.2% in cases vs. 0.0% in controls, P<0.01) [47] (Table 7 (Tab. 7)). In the third sample (32 French patients with ICU-acquired pneumonia, 21 intubated controls in ICU without ICU-acquired pneumonia), APMV DNA was detected in bronchoalveolar lavage specimen from one of the patients with ICU-acquired pneumonia but in none of the controls, confirming that APMV may reach the respiratory tract of these patients [47] (Table 7 (Tab. 7)). So, La Scola et al. [47] were very first who proposed that APMV should be tested as a possible novel human pathogen, particularly in pneumonia patients.

Table 7

Studies on the association of the giant viruses with infections of the respiratory tract

The second study devoted to this issue was carried out by Berger et al. [49]. This investigation included 157 French ICU patients with 210 episodes of pneumonia (120 episodes of healthcare-associated pneumonia (HAP), 62 episodes of CAP, and 28 episodes of mixed pneumonia (CAP complicated with HAP) [49]. Using the similar microimmunofluorescence approach, the authors showed that 15/210 episodes (7.1%) in 14/157 patients (8.9%) were associated with APMV [49]. In addition, laboratory investigations for amoeba-associated microorganisms (AAMs) revealed 59/210 (28.1%) diagnoses in 40/157 (19.0%) patients [49]. Seroconversion among patients with ventilator-associated pneumonia (VAP) was observed in 13/51 episodes (25.5%), whereas seroconversion among patients with CAP was detected in 2/8 episodes (25.0%) that may demonstrate the possible pathogenic role of APMV in both VAP and CAP [49] (Table 7 (Tab. 7)). However, there was no control group in this study; therefore the prevalence of APMV in a general population remains unknown [49]. In the same year, Arden et al. [50] and Larcher et al. [51] failed to find any APMV-positive individuals among 315 Australian patients with suspected acute respiratory tract infections and 214 Austrian patients with pediatric CAP, respectively; nonetheless, the authors used predominantly nasopharyngeal aspirates as a study material and real-time [50] or nested suicide [51] PCR instead of microimmunofluorescence as a method of APMV detection (Table 7 (Tab. 7)). Similar results were further obtained by Dare et al. [52] who were unable to detect APMV using real-time PCR in 496 patients with pneumonia (249 patients with CAP, 71 patient with HAP, 87 bone marrow transplant recipients with pneumonia, and 89 lung transplant recipients with pneumonia) from 9 epidemiologically varied pneumonia patient populations (Thailand, Canada, USA) (Table 7 (Tab. 7)). The authors suggested that seropositivity may reflect chronic exposure to APMV antigen rather than active infection, and the potential for nonspecific cross-reactions with the serologic assays used may have inflated the true prevalence of APMV colonization/infection [52]. Nevertheless, most of the specimens examined in this study were obtained from the upper respiratory tract but not from the lower respiratory tract where the only reported APMV PCR-positive sample was identified [47]. Distinct composition of the studies' populations could also play a role in the differences between the studies [52]. However, in 2009 Vincent et al. [53] found 59/300 (19.6%) French patients with suspected VAP to be APMV-positive, detected by microimmunofluorescence (Table 7 (Tab. 7)). In addition, APMV-positive patients had longer duration of mechanical ventilation and ICU stay with median excesses of 7 days and 10 days, respectively, so a positive serology for mimivirus was associated with a poorer outcome in mechanically ventilated ICU patients [53].

Three years later, Costa et al. [54] investigated the prevalence of APMV in bronchoalveolar lavage (BAL) specimens from 30 ventilated compared to 39 nonventilated Italian patients using real-time PCR but could not detect any signs of APMV colonization/infection among this population (Table 7 (Tab. 7)). Then, Lysholm et al. [41] investigated nasopharyngeal aspirates from 210 Sweden patients with lower respiratory tract infections through metagenomic sequencing approach but detected APMV only in 2/210 (0.95%) cases. Furthermore, Vanspauwen et al. [55] collected 220 sputum and serum samples from 109 Dutch patients with chronic obstructive pulmonary disease (COPD) and analyzed them using the microimmunofluorescence assay and real-time PCR; however, only 3/210 (2.7%) patients were APMV-seropositive and none of the patients were positive for APMV DNA (Table 7 (Tab. 7)). The authors suggested that the low seropositivity might be explained by either a low abundance of the virus or the presence of the virus may depend on the spatiotemporal regional variation as in the case with Legionella-caused infections [55], [56]. Negative PCR results could be explained by a number of reasons [55]. The viral load of the samples may have been below the detection limit of the PCR (12 copies/reaction), or a polymorphism in the area of amplification could occur [55]. A wide genomic diversity of Mimiviridae family may support the latter suggestion [19], [57], [55], [58], so patients could be positive for other viruses of the Mimiviridae family [55]. In addition, sputum is possibly not an appropriate study material compared to BAL sample [55]. In the second study performed by this research group with 260 bronchoalveolar lavage fluid samples from 214 Dutch patients suspected for VAP, no APMV DNA was detected in all of the samples using real-time PCR [59] (Table 7 (Tab. 7)). Finally, in the recent study of Bousbia et al. [60], seroconversion to APMV was observed in 14/71 (19.7%) French pneumonia patients (ICU) with paired serum samples (Table 7 (Tab. 7)). The authors also observed an elevation of the antibody response (both IgG and IgM) to APMV antigens in the convalescent-phase sera in comparison with the admission sera (16/41 (39.0%) patients with HAP and 4/29 (13.8%) controls (ICU patients without pneumonia), respectively, P=0.02) [60].

There are several explanations for the inconsistency observed between the results of the above-mentioned studies. The most evident explanation is that Mimiviridae family DNA detection could possibly have been hampered in studies that used genomic methods of detection. PCR primers used in these studies targeted only APMV genome, whilst a number of APMV relatives exhibiting considerable genetic diversity have been described (Table 3 (Tab. 3)). Recently developed modern real-time PCR systems targeting giant viruses and their virophages are able to accurately detect all or most of the members of the currently delineated lineages of giant viruses infecting Acanthamoebae as well as the virophages [61]; hence, this obstacle may be overcome in the near future. Other reasons for the disparities may include differences in prevalence of giant viruses in distinct populations along with spatiotemporal regional variation as well as the use of inappropriate study material (for instance, samples from upper respiratory tract instead of bronchoalveolar lavage (BAL) specimens). In some cases, giant viruses possibly could not be identified due to the too small viral load for the PCR detection in some cases, insufficient amount of specimens. In addition, APMV was shown to contain up to 23 proteins that may cause immune response resulting in a production of antibodies [33], so proteins of other Mimiviridae family members may cross-react with those of APMV [71]. In this case, molecular detection of APMV could be negative [62]. Exposure to APMV, since it is an AAM, most likely occurs from environmental sources such as contaminated hospital water supplies [52], which is the case in ICU patients. The high rate of seroconversion in pneumonia patients in certain studies [47], [49], [53] suggests that they may have had a contact with APMV or a cross-reacting agent which could possibly be other member of Mimiviridae family.

Conclusion

Viruses belonging to Mimiviridae family can be potential agents of both CAP and HAP. In particular, these viruses may be associated with poorer outcome in ICU patients. The exact mechanism of their pathogenicity, however, still remains unclear. The discrepancies between the results obtained by serological and genomic methods could be explained by the high polymorphism of nucleotide sequences of Mimiviridae family representatives. Further investigations on the Mimiviridae pathogenicity and on the determination of Mimiviridae-caused pneumonia risk groups are required. However, pathogenicity of the viruses belonging to Marseilleviridae family is unclear.

Notes

Competing interests

The authors declare that they have no competing interests.

Funding

There was no funding for this paper.