A multiplex PCR coupled with hybridization by RLB for the characterization of C. burnetii
was designed, allowing for its classification into the previously known 8 GG [15
] and into up to 16 genotypes, depending on adaA
presence/absence. For validation, 15 reference strains characterized in previous studies were used (Additional file 1
: Table S1). All of them fell in the same GGs as previously described, when data was available, or grouped in the same clade as described [8
]. Consequently, an excellent correlation with some previously published schemes and, specifically, with the microarray-based whole genome typing of Beare et al. [15
] was observed: the 4 isolates studied by Beare et al. that were also analyzed in this study (NMI, GG I; Henzerling, GG II; Priscilla, GG IV; and Scurry Q217, GG V) were classified with this method into the same GG as described. Also, the analysis of the results by InfoQuest disclosed a tree whose topology was similar to that of Beare et al. [15
], with the only exception of GG VIII, which in this study grouped together with GGs I, II and III, instead of with GGs IV, V, VI and VII as in the Beare study.
With this methodology, a preliminary characterization of C. burnetii variants circulating in Spain has been performed showing a high variability of this organism in clinical and environmental settings, identifying 7 GG, with the exception of GG V, and 10 different GTs.
In Spain, while a respiratory disease is observed in about 80% of cases reported from the Northern region of the Basque Country [26
], the Southern regions of Andalusia and the Canary Islands report a clear predominance (about 90% of cases) of FID with liver involvement [28
]. This last has also been described in Australia, France, Greece, or Taiwan [35
], among others. Even taking into account the limited size of this study and the constraint of an extrapolation, a strain-associated factor that might explain the different clinical presentations of acute Q fever is hypothesized for our country. The pattern observed in cases of acute Q fever indicates an association between absence of adaA
and FID with liver involvement, produced in this study by adaA
negative strains in both regions (the Southern regions of Andalusia and the Canary Islands), although is not statistically significant in this study (p
0.09) due to low number of samples. Also, another sample of a case of hepatitis from the north (Basque Country) yielded an adaA
negative result as well. The same applies for the 2 reference isolates from hepatitis cases analyzed in this study: F2, a French isolate and SQ217, recovered in the USA from a case of chronic hepatitis, are both adaA
negative as well. In contrast, pneumonia predominates over liver involvement in Northern Spain, being the only case of this clinical form available for the study produced by an adaA
positive strain. No other marker used in this study correlated with the clinical presentation of acute Q fever. Availability of samples from cases of acute Q fever for genotyping is much less frequent than from cases of chronic Q fever, even though acute Q fever is much more prevalent. In this study, 11 samples from acute cases were analyzed, although only one was from a case with respiratory symptoms, reflecting the limited availability of such samples, which may be due to a poor clinical awareness.
From the 10 GTs found in the country, only 5 have been detected in humans and, among them, GT IV- is the most frequently found in acute and chronic cases (75% of cases). This GT has also been found in many mammal species (sheep, goat, wild boar and rats). Whether this could be interpreted as a higher tendency of this GT to cause illness in humans can not be inferred by this study, mainly considering that most of the acute cases (8/11) came from the same area (Gran Canaria Island). In any case, GT IV- is highly prevalent also in our chronic cases that came from 8 distant areas of the country, showing a more intensive circulation of this GT in humans.
The association previously proposed between adaA-
negative strains and chronic disease [19
] has been reproduced here for all chronic cases studied (13 cases), except for a sample from a vascular infection that was adaA
positive. Beare et al. [15
] hypothesized that the association between GG IV and chronic cases (as in 12 out of 13 chronic cases studied here) could be related to the slow growth of isolates from this genotype and, therefore, the induction of a decrease in the immune response. On the other hand, Zhang et al. hypothesized that adaA
positive strains were related to acute cases [19
], as it is the case of the only sample from a patient with acute pneumonia available. However, in our study, acute cases of FID with liver involvement were all produced by adaA
GTs found in humans were also found in sheep, goats, rats, wild boar and ticks. This distribution of GTs suggests that sheep and goats are responsible for the transmission of C. burnetii
to humans in Spain, as in other areas [39
], and exhibit a high variability of GT. However, although in general domestic ruminants are important reservoirs for C. burnetii
and play a relevant role in its transmission to humans, 4 of 24 human samples were found carrying GTs not found in ruminants in this work. A recent Spanish study [40
] has also detected C. burnetii
in roe deer, wild boar, carrion birds and hares. Although there is no data available on the genotyping of these specimens, more studies are needed to characterize the enzootic cycle of C. burnetii
and its GT distribution in wildlife, as well as to ascertain whether other sources could be responsible for the transmission of C. burnetii
to humans. GG VII was only found in ticks (H. lusitanicum, Dermacentor marginatus
and Rhipicephalus sanguineus
) and in 3 cases out of 10 of FID with liver involvement. It is to note that, while reference isolates from ticks belonged mostly to GG II, this GG has not been found in ticks in our study. Although the analyzed tick specimens came from 5 different areas, they were all from Central Spain, which could be biasing this data. Transmission of Q fever by tick bite still remains controversial [41
], and cases of simultaneous or consecutive infections with C. burnetii
and other tick-borne agents have been described [43
]. Whether C. burnetii
can be transmitted by tick bite or not, the detection in ticks of GT VII-, found only in human patients revives this debate. More studies are needed to definitely clarify this question. On the other hand, given that GG VII isolates have not been found in cattle, sheep and goats in this study, we could think of other unknown reservoirs that could be involved as a source of infection of this GG for both ticks and humans. Traditional mammal species on which the tick species analyzed in this study feed on include rabbits (frequent all over Spain) for the immature stages of H. lusitanicum,
which seems to be very important for the maintenance of populations of this tick species, small mammals for those of D. marginatus,
and canids for all stages of R. sanguineus
. Adult H. lusitanicum
and D. marginatus
normally feed on large ungulates. Animals present in the tick study areas included, apart from cattle, high densities of rabbits and other wildlife. It is to note that 40 liver samples from rabbits hunted in Gran Canaria analyzed by PCR were all negative (data not shown), although more studies are needed. Whether some of the above mentioned animals may act as reservoirs for GG VII C. burnetii
remains to be studied.
Interestingly, in 7 cattle samples from 4 distant regions, only GG III was detected. In the study of Arricau-Bouvery [13
] most of the cattle isolates (12/14) analyzed by MLVA also grouped together in a clade that is close but different to the one that include GG I isolates, as in this study. In Beare’s study GG III is also philogenetically close to GG I and both clades appear together in the tree. This GG having never been found in humans in Spain so far lead us to hypothesize that cattle could represent a low risk for Q fever transmission to humans in our country.
One of the added values of the method described here is that it could be applied to any PCR-positive sample carrying at least 10 genome equivalents of the target organism, thus avoiding the need for culturing the organism to obtain data on the global circulation of C. burnetii. The frequent lack of human isolates from outbreaks, which are needed to apply the yet described methods, hamper a correct outbreak study that are necessary to identify the source of infection. This methodology allows the characterization directly from clinical samples avoiding the culture step of this fastidious bacterium, and proves to be valuable identifying so far 10 different GTs circulating in Spain. This method can be performed in any laboratory with basic equipment. It can easily determine relationships among C. burnetii from different origins by using PCR-positive samples, thus helping in the identification of the source of an outbreak in a rapid analysis.