Human Infection of Methicillin-Susceptible Staphylococcus aureus CC398: A Review

Staphylococcus aureus (SA) belonging to the clonal complex 398 (CC398) took a special place within the species due to its spread throughout the world. SA CC398 is broadly separated in two subpopulations: livestock-associated methicillin-resistant SA (MRSA) and human-associated methicillin-susceptible SA (MSSA). Here, we reviewed the global epidemiology of SA CC398 in human clinical infections and focused on MSSA CC398. The last common ancestor of SA CC398 was probably a human-adapted prophage φSa3-positive MSSA CC398 strain, but the multiple transmissions between human and animal made its evolution complex. MSSA and MRSA CC398 had different geographical evolutions. Although MSSA was present in several countries all over the world, it was mainly reported in China and in France with a prevalence about 20%. MSSA CC398 was frequently implicated in severe infections such as bloodstream infections, endocarditis, and bone joint infections whereas MRSA CC398 was mainly reported in skin and soft tissue. The spread of the MSSA CC398 clone is worldwide but with a heterogeneous prevalence. The prophage φSa3 played a crucial role in the adaptation to the human niche and in the virulence of MSSA CC398. However, the biological features that allowed the recent spread of this lineage are still far from being fully understood.


Introduction
Staphylococcus aureus (SA) is a leading cause of morbidity worldwide [1]. During the past 20 years, livestock animals, especially pigs, have been identified as a reservoir of methicillin-resistant SA (MRSA) isolates that clustered in clonal complex 398 (CC398). MRSA CC398 has emerged worldwide and has been found to colonize and infect animals and humans [2]. Recent genomic analyses have demonstrated the existence of two main subpopulations within the CC398: an ancestral human-adapted clade with the integrase group 3 prophage (ϕSa3) containing the immune evasion cluster (IEC) genes and the erythromycin-resistant gene erm(T) and an animal-associated clade associated with the absence of ϕSa3, and the acquisition of resistance to tetracycline and methicillin [1,3,4]. Severe infections, such as bloodstream infection (BSI), endocarditis, osteomyelitis and necrotizing pneumonia or mild infections such as skin and soft-tissue infections (SSTI) were reported in both populations of CC398. However, the majority of publications examining SA CC398 to date have focused on livestock-associated MRSA.
In this review, we first discussed the evolution of SA CC398 and its epidemiology in clinical infections in human and then focused on methicillin-susceptible SA (MSSA) CC398. The last common ancestor of Staphylococcus aureus (SA) CC398 was probably a human-adapted immune evasion cluster (IEC)-positive methicillin-susceptible SA (MSSA) CC398 strain, which at a later stage acquired SCCmec, leading to the emergence of human methicillin-resistant SA (MRSA) CC398 strains. At some point, the ancestral MSSA CC398 strain jumped to livestock, which was accompanied by the loss of IEC and the acquisition of tet(M), conferring resistance to tetracycline and later on by the acquisition of SCCmec, conferring resistance to methicillin, leading to the emergence of livestock-associated MRSA CC398. However, Ward et al. [14] supported the existence of distinct humanand livestock-associated clades that emerged at similar times, which was represented by "?*" in the figure. Because some interspecies transmission in both directions was described, a double-arrow was used to represent the different livestock-associated clade subpopulations. These transmissions were probably responsible for the acquisition of the prophage ϕSa3 in the LA clade, which differentiated human (Hu) SA CC398 subpopulations from the adapted Humans (HuA) livestock-associated (LA) SA CC398 subpopulations.
A variety of MRSA CC398 infections has been documented, ranging from relatively minor or localized infections (SSTI) to more serious or invasive infections (including bloodstream infection (BSI), osteomyelitis, pneumonitis, and endocarditis) [24]. Despite the diverse array of infection types, it has been suggested that MRSA CC398 may not cause as much severe diseases as MSSA CC398 [25]. Details on MSSA CC398 studies were presented in the Table 1.
The analysis of the publications that have specified the clinical diagnosis of SA CC398 infections is summarized in Table 2. BSI, endocarditis, and bone joint infections (BJI) were reported more frequently in MSSA CC398 studies, whereas SSTI, respiratory tract and urinary tract infections were reported more frequently in MRSA CC398 studies. However, most countries do not perform MSSA surveillance, and when they do, they focus on invasive isolates, meaning that MSSA SSTIs are likely to be underreported in the literature.
Concerning BJI, a retrospective study in 2008 found that only 1.9% of MSSA were ST398 [31]. Valour et al. found that 68 of the 485 MSSA prosthetic joint infection isolates (14%) belonged to CC398 in a multicenter study in France between 2009 and 2012 [35], with a geographical heterogeneity and a regional prevalence ranging from 3.1 to 23.5%.
In patients with diabetic foot infection (DFI), MSSA ST398 represented 26 of 166 (15.7%) and was the main lineage isolated. Moreover, it was identified exclusively in cases with osteomyelitis [32]. These results were confirmed by a multicenter study of patients with DFI, with a prevalence of 21.7% (35/161) that reached 38.3% in patients with diabetic foot osteomyelitis only [33]. Recently, our group reported an increasing trend of prevalence of MSSA CC398 in BJI over a 8 year period (from 4% in 2010 to 26% in 2017), echoing what we observed in BSIs [30]. Overall, since the incidence of MSSA CC398 steadily increased in the last decade, it became a predominant SA lineage responsible for infections in France.

In Europe
In Europe, MSSA CC398 represented 2.1% of the 565 MSSA isolates with data of spa typing from invasive infections (mainly BSI) collected in 26 countries during the period 2006-2007 [25]. These isolates were retrieved in eight different countries which suggested the spread of this clone in Europe. However, in a retrospective study from Germany, the prevalence of MSSA CC398 isolates from 2006 to 2012 remained low (0.14% among isolates from infection in humans) [38]. As in a large collection of 610 consecutive isolates of SA BSI episodes in a multicenter study in The Netherlands that only retrieved two MSSA CC398 (0.3%) isolates during the period 2010-2011 [47]. Rijnders et al. also found a limited prevalence of CC398 among SA isolates in intensive care units (2 of 936, 0.2%) in this country [46]. In Belgium, among 212 MSSA isolates collected during a national survey in 2008, five belonged to CC398 and were retrieved from patients located in different regions [27]. In southern Europe, in a collection of 465 MSSA isolates from infection and colonization collected in Portugal over a 19 year period (1992-2011), 14 (3%) late isolates found during the period 2010-2011, were MSSA ST398-t571, indicating the recent emergence of this lineage in this country [49]. In Greece, the first case of an MSSA CC398 infection was described in 2011 in a BSI [40]. Then, Sarrou et al. reported, among a large collection of 492 clinical MSSA isolates from infected humans during the period 2012-2013, 2.6% of MSSA CC398 (13/492 isolates) [42].

In America
MSSA CC398 infections are also observed in America. In South America, MSSA ST398 spa type t571 has been described in Colombia as causing infection in a woman without livestock association and in Ecuador from a patient with BSI [69,72]. From 70 consecutively BSI and non-consecutive isolates from different clinical infections, only one (1.4%) isolate from BSI was MSSA CC398 [72]. Recently, Jewel et al. reported a case of severe disseminated infection caused by MSSA CC398 in an immunocompetent man, shortly arrived in the United Kingdom from Colombia [70]. Another case was reported in an oncology patient in Sao Paulo, Brazil, with the first Cfr-producing MSSA ST398 from a patient with a fatal pneumonia [68]. In Dominican Republic and Martinique, CC398 prevalence reached, respectively, 7.8% (7/90) and 10.4% (9/87) among clinical MSSA [71].
Finally, MSSA CC398 was retrieved all over the world but with different prevalence depending on the country. France and China were countries where this clone has been most studied and where it is most frequently isolated from human infections. However, most countries do not perform MSSA surveillance, and the disease burden of MSSA CC398 at the national and international level is therefore largely unknown.

BJI
BJI related to CC398 included principally prosthetic joint infection (PJI) and diabetic foot osteomyelitis (DFO) [33]. MSSA CC398 was the dominant clone in patients with DFO, accounting for 17.9% among all SA BJI and 38% among all SA DFO [30,33]. The frequency of CC398 SA PJI isolates varied from 1.8 to 18.1% of all SA PJI [30,31,35]. Moreover, Senneville et al. showed that SA CC398 isolates were significantly more frequent in osteomyelitis than in SSTI patients with diabetic foot infection, suggesting a particular tropism of this clone for bone [33]. However, CC398 had a low potential for osteoblast invasion in ex vivo models, suggesting unidentified virulence factors [35].

Respiratory Tract Infections
MRSA CC398 seems to be significant in respiratory tract infections (RTI) [24]. It could be related to the mode of transmission, with close contact with animals. In contrast, fewer cases of MSSA CC398 RTI have been reported [7,8,16,19,20,27,29,37,39,50,55,61,62,68,74]. Hence, a French study of 89 clinical isolates of MSSA CC398 reported that 12% were isolated from pneumonia, 22% from BSI and 29% from SSTI [19]. However, this collection included isolates voluntarily sent by French microbiology laboratories for further characterization because of a particular antibiotic-resistant phenotype or unusual infections, or in the setting of targeted studies. Thus, the collection does not represent the entire MSSA CC398 population in France. Similarly, pneumonia represented 14% of all MSSA CC398 infections (excepted BSI) in New York [8], and MSSA CC398 represented 4.9% of all SA respiratory tract infections in China [61]. Moreover, Song et al. reported that 10.3% of SA pneumonia in pediatric inpatients was due to MSSA CC398 [62]. RTI was also reported as the portal of entry of BSI in several studies in which ≤ 1/4 of the MSSA CC398 BSI isolates were of pulmonary origin [7,20,50]. Two fatal cases of pneumonia have been reported in France and Brazil [29,68].

SSTI
SSTI seems to be the predominant infection linked to MRSA CC398. However, this overrepresentation could be biased due to the specific surveillance of MRSA infections in several countries, which lacks for MSSA. Indeed, in a French study of 105 clinical isolates of MSSA CC398, SSTI was the most represented infection (35% of cases), followed by BSI with 26.5% of cases [19]. However, all specimens in this study were voluntarily sent to French National Center for Staphylococci, because of a particular antibiotic-resistant phenotype or unusual infections, which make it impossible to generalize these results. For patients with BSI, a SSTI was the suspected portal of entry in 6.3-11% of the cases, in line with other clonal complexes (CCs) [7,17,20]. The proportions of CC398 among MSSA and MRSA isolated from SSTIs in China were similar (12-17%) [58]. Moreover, MSSA CC398 was present in 1/26 (3.8%) SA isolates colonizing the skin of children with atopic dermatitis [79]. Finally, the predominance of MRSA CC398 in SSTI could be due to a publication bias. However, we suggest that the mode of transmission, with close contact with animals, plays a role in the clinical type of MRSA infection.

Community-Acquired or Healthcare-Associated Infection (HAI)
In the study of Bonnet et al., MSSA CC398 BSI were associated with an intravascular device in 56.3% of the cases vs. 31.6% in non-CC398 MSSA BSI [20]. Similarly, Valentin-Domelier et al. showed that BSI related to surgical site infection (SSI) or an intravascular device were found preponderant among CC398 BSI cases (14/18, 78%) [7]. In a retrospective study, MSSA CC398 BSI were more frequent in patients with SSI (22.4 vs. 11.2%) and another study found that patients with MSSA CC398 infections were more likely hospitalized in the past 6 months, suggesting a nosocomial transmission (41 vs. 22%, p < 0.01) [8,17]. Moreover, we found that HAI was associated with CC398 MSSA BSI in multivariate analysis, compared to MSSA BSI with other CCs [17]. These data suggested that this clone was possibly related to HAI. However, existing literature found evidence for MSSA CC398 as both a community-and hospital-associated pathogen [8,9,18,37]. Moreover, MSSA CC398 has been described as colonizer of healthy humans, in addition to causing infections [80]. Prevalence studies should be carried out to test whether patients in a hospital setting are more colonized than a healthy population for MSSA CC398.

Comorbidities
Like most SA infections, many case reports of patients with MSSA CC398 described patients with predisposing conditions (e.g., diabetes, advanced age, immunosuppressive treatment). However, some studies showed different characteristics from patients with MSSA CC398 and non-CC398. Indeed, it has been reported that BSIs due to MSSA CC398 more frequently affected fragile patients with a history of neurological disease [17]. Similarly, Uhlemann et al. found that a higher proportion of alcohol abuse and cirrhosis were associated with MSSA CC398 infections and suggested a poorer health status of patients [8]. In the same way, patients with PJI and DFO due to MSSA CC398 had more comorbidities than those infected by MSSA non-CC398 [30]. In the study of Bonnet et al., patients with BSI due to MSSA CC398 more often had diabetes mellitus (31.3 vs. 25%) and immunosuppression (31.3 vs. 28.3%), but the small size of the patient cohort did not allow to reach the significance threshold [20].

Outcome
As mentioned above, MSSA CC398 causes a wide range of infections, including BSIs. Many studies have reported MSSA CC398 deaths in France. A lethal necrotizing pneumonia caused by Panton-Valentine leukocidin (PVL) producing MSSA CC398 was described in a previously healthy 14 year-old girl [29]. Another patient with a cancer died from a pneumonia with BSI in Brazil without further details of his infection [68]. Two additional deaths were reported, in a 59 year-old male and an 80 year-old male, among 17 patients which had a BSI with MSSA CC398 (11.8%) [7]. Moreover, in a monocentric retrospective study, BSIs due to MSSA CC398 were associated with a high risk of 30 days-mortality than other CC in a multivariate analysis (41.8 vs. 27.6% OR = 2.44) [17]. Ulhemann et al. described a trend to more frequent episodes of secondary invasive disease with CC398 but no difference in the proportion of a fatal outcome (6.3 vs. 4.2% p = 0.5) [8]. However, in this study, BSIs were excluded. Conversely, MSSA CC398 PJIs were significantly associated with a lower biological inflammatory syndrome (p = 0.035) and lower treatment failure rates (0 vs. 37.3%, p = 0.032) than other CCs [35]. These conflicting results may indicate yet unidentified virulence features of MSSA CC398. Alternatively, it may also reflect the underlying comorbidities or an immunosuppressed state of the affected host more than the bacterial virulence itself, as suggested above.

Antibiotic Resistance
Whereas MRSA CC398 were frequently multiresistant to antibiotic, including tetracycline, secondary to the wide use of this antibiotic in the pig industry, MSSA CC398 was more often only reported as resistant to erythromycin [24].
Uhlemann et al. reported that MSSA CC398 isolates retrieved in humans in New York (USA) were almost always resistant to erythromycin (97%) and clindamycin (97%) [8]. This resistance was very mainly associated with the presence of ermT gene [19]. Interestingly, ermT was absent in MRSA CC398 from animal origin [81].
Tetracycline resistance was rarely described in MSSA isolates but was frequent in MRSA CC398 due to the presence of tet(M). Indeed, only four MSSA CC398 studies showed tetracycline resistance with the presence of tet(M) gene [28,31,40,42]. However, the absence of WGS of these isolates did not distinguish LA and Hu clades.

Virulence Factors
SA CC398 were initially distinguished by their peculiar resistance to digestion by SmaI, the restriction enzyme most frequently used for PFGE typing [82]. Investigations into the molecular background of MSSA CC398 showed that, most well known staphylococcal virulence genes such as enterotoxins, toxic-shock syndrome toxin or leukocidins are most of the time, lacking [1,35]. However, the 30 day all-cause mortality was higher for patients with CC398 MSSA BSI than for a control group with non-CC398 MSSA BSI [17]. This lineage may represent a more virulent CC398 subtype, as suggested by its higher prevalence in BSI. Genome analysis of CC398 isolates showed that the mobile genetic elements (MGEs) were specific for each population and enabled the differentiation of strains responsible for asymptomatic colonization to invasive infections [12]. Most of the bacterial factors involved in infection severity are MGEs that belong to the accessory genome. The prophage ϕSa3 contains the immune evasion gene cluster (IEC, which facilitates the escape from human immune response) genes sak (coding for a staphylokinase), chp (coding for a chemotaxis inhibitory protein; CHIPS), scn (coding for a complement inhibitor; SCIN), sea (coding for enterotoxin; SEA) and sep (coding for enterotoxin; SEP). However, only a small proportion of the MRSA CC398 isolates retrieved from infected patients contains the ϕSa3-associated IEC [1,7,26,42,66,[83][84][85][86][87]. This suggests that although IEC is not a prerequisite for infections in humans, it plays a crucial role in the adaptation to the human niche. Moreover, the presence as well as the structure of ϕSa3 seems to be directly associated with virulence, even though the different genetic backgrounds of the ϕSa3-positive and ϕSa3-negative isolates was not take into account in this study [88]. Interestingly, human SA CC398 had a higher capacity for binding to human cytokeratin-10 than LA SA CC398, mediated by the clumping factor B (ClfB) and very likely important for nasal colonization of humans [9]. Recently, Laumay and colleagues showed that the introduction of bacteriophages from the genomes of human-adapted SA to that of a naive SA animal colonizer increased the transcription of clfA (encoding the clumping factor A precursor) and fnbA (encoding the fibronectin binding protein), which in turn increased the bacterial virulence in a rat model of infectious endocarditis [89]. This study confirmed the role of MGE in the virulence of SA CC398.
Other virulence factors have been described in CC398 isolates. The presence of genes encoding hemolysins and intracellular adhesion proteins, cap5, together with three genes encoding MSCRAMM (bbp, clfA, clfB) were reported in the genomes of isolates responsible for DFI [33]. Moreover, Liu et al. found that CC398 isolates harbored exfoliatin genes (eta, etb) more frequently than other CC isolates [21]. Enterotoxin and hemolysins genes were also described in two Chinese studies [21,22].

Conclusions
The spread of the MSSA CC398 ancestral clone, and/or the transmission of a human-adapted clone from animals, played a role in the emergence of SA CC398. Prophages have an important role in bacterial virulence and evolution by carrying genes coding for novel characteristics for their host. However, unanswered questions remain on the epidemiology and clinical relevance of the MSSA CC398 sublineage. Moreover, the biological features that allow the recent spread of this lineage in clinical settings of some countries (e.g., China, France) are still far from being fully understood.