TICKS PARASITIZING THE SPUR-THIGHED TORTOISE (TESTUDO GRAECA) POPULATION OF TUNISIA Open Access

From an ecological point of view, it is known that ectoparasitism affects host population dynamics (Brown and Brown 1986, 1992) and should be considered in wildlife management and conservation programs. Relationships between tortoises and tick parasitism are modulated by different factors such as sex, age, and host density (Tiar et al. 2016; Segura et al. 2019). Compared with the adjacent countries of Algeria and Morocco, little is known about host association between ticks and the spur-thighed tortoise (Testudo graeca) in Tunisia.

Tortoises retrieved from illegal Mediterranean trade have been reported to be highly infested with Hyalomma aegyptium (Brianti et al. 2010; Gharbi et al. 2015). The impact of international trade in reptiles remains largely underestimated despite the known associated risk of introduction of their pathogens and parasites to new areas and the subsequent emergence of tick-borne diseases (Burridge and Simmons 2003; Rosen and Smith 2010; Schmidt et al. 2013).

High infection prevalence of H. aegyptium with Crimean–Congo hemorrhagic fever virus (CCHFv) collected from spur-thighed tortoises captured in Algeria (28.6%, Kautman et al. 2016) and in Syria (30%, Široký et al. 2014) has been reported. The role of this reptilian host is central for pathogens with vertical transmission in vectors, as CCHFv has been shown to be transmitted transovarially in other tick species such as Hyalomma marginatum (Zeller et al. 1994). Therefore, it is of major ecoepidemiological importance to assess tick parasitism of the spur-thighed tortoise in Tunisia.

The spur-thighed tortoise is a protected species in Tunisia and this study was performed with the permission of the General Directory of Forest at the Tunisian Department of Agriculture (authorization no. 5 following the request made by Marwell Wildlife no. 1441 on 13 March 2017).

Tunisia covers a wide climatic range, from the Mediterranean climate with its rainy winter in the north to the Saharan climate in the south. The northern part of the country is separated from the south by the Tunisian Ridge. The latter is a range of hills that runs from NE to SW for some 220 km, marking the climatic boundary between the Mediterranean north and the dry steppe of central Tunisia. Between the northern slopes of the Tunisian Ridge and the chains of hills bounding it on the south are extensive plateaus called the High Tell. The Sahara is separated from the central steppe land by a series of salted areas called chotts.

Spur-thighed tortoises are abundant in northern Tunisia (Blanc 1978). Wild tortoises from different sites, located in humid and subhumid biogeographic areas of northern Tunisia (Fig. 1), were sampled on foot transects conducted in their natural habitats during March–May 2017. Captured tortoises were identified on site using identification keys (Ernest and Barbour 1989; Highfield 1990): the endemic spur-thighed tortoise has distinctive morphologic traits that allow quick species taxonomic identification. The age of tortoises was estimated by counting the scute rings annuli (Castanet and Cheylan 1979). Each tortoise was examined thoroughly for the presence of ticks and released at the point of capture. The number of ticks and their attachment sites were recorded for each tortoise. All ticks attached to each infested tortoise were collected and placed in a snap-cap vial. A subsample of ticks collected from a random sample of captured tortoises was identified using the identification keys of Hoogstraal (1956). The remaining collected ticks were not identified and stored at –80 C for examination later for the presence of CCHFv (Fares et al. 2019).

Field-based tortoise searches were carried out in independent 1-ha quadrats made of homogeneous habitat types (Fig. 1) within the expected distribution range of spur-thighed tortoises. The vegetation in each quadrat was characterized by the density of both the upper (trees higher than 2 m) and medium (bushes lower than 2 m) canopy cover, assessed on ordinal scale as either absent, scattered, or dense. Farmland (olive plantations and fields) was recorded as separate categories. The habitat classification resulted in eight habitat categories (Fig. 2) that represent the features expected to match the tortoises' biological requirements (Anadon et al. 2006).

The number of tortoises screened was determined using the Kerijcie and Morgan formula (Krejcie and Morgan 1970) with the assumption that the proportion of H. aegyptium will not be substantially different from 50% (this actually provides the maximum sample size) of the ticks collected and allowing a subsequent reliable survey of pathogen infection incidence within the remaining parasites (confidence interval=95%; margin of error=1%). Consequently, the ticks of 13% (18/134) of the infested tortoises were identified, representing about 10% (n=120) of the total collected.

To describe the tick infestation of tortoises, we estimated three parasitological indicators: infestation prevalence (%) = number of infested tortoises × 100/total number of tortoises; infestation intensity = number of ticks/number of infested tortoises; and relative abundance = number of ticks/total number of tortoises.

The variation of these indicators was then analyzed against environmental (extrinsic) factors (bioclimatic stage, type of habitat) or demographic (intrinsic) factors (tortoises' sex) by using generalized linear model (GLM) and Kruskall-Wallis rank sum test. Principal components analysis was performed to assess the tick distribution with respect to the different parts of the tortoise: right anterior limb (RAL), right posterior limb (RPL), left anterior limb (LAL), left posterior limb (LPL), neck, tail, plastron, and carapace. The correlation of these indices with height and tortoises' sizes as indicated by the straight-line carapace length in millimeters, weight, and age was also investigated by using linear regression (LR). All statistics were performed using the R software for statistical computing version 3.2.1 (R Development Core Team 2017). A P value <0.05 indicates a significant difference.

A total of 147 tortoises (63 males, 72 females, 12 juveniles) were captured and identified as the spur-thighed tortoise. Of those, 134 were infested with at least one Hyalomma sp., tick, yielding an infestation prevalence of 91.2%. A total of 1,174 adult ticks was collected, from which 120 were taken from 18 randomly selected tortoises and used for tick identification. Of the 120 sampled ticks, representing 10% of the total collected, all were identified as H. aegyptium; there were 88 males and 32 females. Among this subsample of collected ticks, the mean number was 5±2.8 male ticks and 2±2 female ticks per tortoise. The sex ratio of ticks was 2.8±1.8, indicating a male bias.

Across habitat types, significantly more tortoises were encountered in sparse forests with dense shrub (GLM, with Poisson error function: z=3.55; P<0.001) and sparse shrub habitats (z=2.56; P=0.010). The infestation prevalence of ticks on tortoises ranged from 83% in the forests with sparse shrubs to 100% for farmland and dense scrublands. The analysis of variance showed that there was no significant difference (Kruskal-Wallis χ²=5.1207, df=6, P=0.528) between the different habitat types encountered in the study areas.

Infestation prevalence observed among male, female, and juvenile tortoises of unknown sex was 95% (63/147), 96% (72/147), and 42% (12/147), respectively. The number of infested juveniles was fewer compared with adult males and females (LR: F_2,144=15.28; β=–62.50; P<0.001) and there was no significant difference between infestation prevalence of males and females (adults only, GLM: F_1,133=0.64; P=0.420).

The overall infestation intensity and abundance was 8.5 and 7.8, respectively. Detailed analysis of the intensity of infestation, according to the tick attachment sites on the tortoise, was conducted. The majority of ticks (89%) was attached to the posterior parts, namely hind limbs and tail (Fig. 3). Results showed that the preferred sites for adult tick attachment were the posterior limbs, which had an average infestation intensity of about three ticks per limb.

Juveniles had the lowest levels of tick infestation compared with adult males and females (GLM: F_2,144=9.37; β=–6.65; P<0.001) and there was no significant difference between relative abundance of ticks within males and females (adults only, GLM: F_1,133=0.13; P=0.720). Furthermore, the intensity of infestation did not differ significantly between male and female tortoises (adults only, GLM: F_1,102=0.20; P=0.650), but a correlation between the infestation intensity and the estimated age of the tortoises was observed (LR: F_1,107=14.89; β=0.4572; R²=0.122; P<0.001). This is confirmed by the correlation between the intensity of infestation and the weight in grams (LR: F_1,106=6.52; β=0.0053; R²=0.058; P=0.012), and even more significantly with tortoise' size (LR: F_1,107=14.11; β=0.0915; R²=0.117; P<0.001). The increased dimension of the carapace posterior aperture, namely the post-anal gap, which indicates the size of the possible ticks' attachment site on the tortoise's skin, was also associated with a larger number of ticks per tortoise (LR: F_1,100=8.91; β=0.3059; R²=0.082; P=0.004). On the contrary, there is no correlation between the infestation intensity and the carapace aperture in the front measured by the frontal width of the marginal scutes (LR: F_1,100=2.29; β=0.0149; R²=0.003; P=0.590).

We applied principal components analysis to the infestation rate of the different body parts of the tortoise and showed that the number of ticks in the RPL, LPL, neck, and tail were actually associated with age, which in turn was moderately but significantly correlated with weight and height. Similarly, the numbers of ticks in the RAL, plastron, and LAL were associated together. Finally, the number of ticks on the carapace was independent of the rest of the variables (RAL, RPL, LAL, LPL, neck, tail, and plastron; Fig. 4).

The altitude was not correlated with the relative tick abundance (LR: F_1,117=1.30; β=0.0787; R²=–0.011; P=0.813), nor was it with the individual intensity of infestation (LR: F_1,96=0.05; β=0.8130; R²=0.001; P=0.256). Infestation intensity, however, varied significantly according to habitat type: the tortoises found in dense shrub habitats were the most infested by ticks (GLM, with Poisson error function: z=2.77; P<0.006), with an infestation intensity of 13.63±6.35 (Fig. 5).

Adult H. aegyptium was the only tick species identified among the subsample collected during the study period. Even if the species identification was completed on only 10% of collected ticks, the confidence interval and margin of error related to the sample size were satisfactory; all 134 parasitized spur-thighed tortoises that we captured were almost solely infested by adults of H. aegyptium. However, a small percentage of other tick species among the remaining collected ticks was not formally identified and so was not included in our study. Among a subsample of collected ticks from spur-thighed tortoises captured in Morocco, 96.5% were identified as H. aegyptium and the remaining 4.4% of ticks were identified as H. marginatum, Hyalomma excavatum, and Hyalomma scupense (Segura et al. 2019). Several studies corroborated our results. Custom-seized spur-thighed tortoises from Tunisia were infested only by adult H. aegyptium (Brianti et al. 2010; Gharbi et al. 2015). In Iran, 117 ticks collected from 14 infested spur-thighed tortoises were identified as adult H. aegyptium (Tavasoli et al. 2007). In Syria, 245 ticks collected from 38 infested spur-thighed tortoises were identified as adult H. aegyptium (Široký et al. 2014). In southwest Bulgaria and in Algeria, spur-thighed tortoises were found infested with all life stages of H. aegyptium, but the sampling occurs later in the year (Široký et al. 2006; Tiar et al. 2016). In northwestern Morocco, adult H. aegyptium are the predominant tick species (95.6%) among a subsample of ticks (52%) collected from spur-thighed tortoises captured in cork oak (Quercus suber) forest (Segura et al. 2019). Nymphs were not found, most probably because they are considered to be more common during late summer and autumn (Robbins et al. 1998).

The infestation prevalence was 91%, varying from 82% to 100% by capture location. We report a higher prevalence of infestation of tortoises by H. aegyptium than did Gharbi et al. (2015) on confiscated Tunisian tortoises (66%), Tiar et al. (2016) in Algeria (from 9% to 98%), and Tavasoli et al. (2007) in Iran (44%). We did not find any significant difference in the infestation prevalence within the humid and subhumid biogeographical areas of Tunisia. Other authors reported a lower prevalence of infestation of tortoises in the Saharan Atlas of Algeria (Tiar et al. 2016) and in an arid area located in southern Morocco (Široký et al. 2009). Whereas some studies indicated low humidity as a possible limiting factor for the distribution of H. aegyptium (Široký et al. 2009), Tiar et al. (2016) did not find any significant relationship between the climatic parameters investigated and the intensity and prevalence of infestations.

Despite the fact that tortoise age estimation using scute rings has been shown to have significant limitations (Wilson et al. 2003; Attum et al. 2011; Rodriguez-Caro et al. 2015), a correlation between the infestation intensity and the estimated age of tortoises was observed. Similar results were reported by Segura et al. (2019). No significant difference was observed among infestation prevalence in male and female spur-thighed tortoises as has been reported by others (Široký et al. 2006; Brianti et al. 2010; Gharbi et al. 2015; Tiar et al. 2016).

Therefore, both sexes of spur-thighed tortoises appeared to be equally parasitized. However, Robbins et al. (1998) evidenced a higher parasitism in male tortoises by testing differences in mean number of male ticks only, because they are less overdispersed (or clumped, as in our results). They hypothesized that it could be explained by the larger home range of male Testudo (Longepierre et al. 2001).

Corresponding to the regional assessment of the parasitism of free-ranging spur-thighed tortoises previously done in Morocco (Široký et al. 2006; Segura et al. 2019) and Algeria (Tiar et al. 2016), we observed a strong correlation between spur-thighed tortoise body size and infestation in Tunisia. These results are strengthened by studies of the tick load of tortoises retrieved from illegal trade in Tunisia (Gharbi et al. 2015).

The intensity of infestation varied significantly among size and attachment sites of spur-thighed tortoises. Similar results were reported from custom-seized tortoises in Tunisia (Gharbi et al. 2015), Algeria (Tiar et al. 2016), Italy (Brianti et al. 2010), and the Balkans (Široký et al. 2006). The infestation intensity differed significantly between habitat types; the highest intensity was observed in forests with dense shrubs. The overall infestation intensity of 8.5 reported in our study was extremely high compared with those reported from other countries. The infestation intensities reported from Algeria, Jordan, the Balkans, Russia, Italy, Tunisia, and Morocco varied from 1.7 to 9.4 (Tiar et al. 2016), 0.2 to 5.9 (Petney and Al-Yaman 1985), 1.3 (Široký et al. 2006), 5.2 (Robbins et al. 1998), 3.9 on tortoises imported from North Africa (Brianti et al. 2010), 4.3 on tortoises seized by customs (Gharbi et al. 2015), and 6.67 (Segura et al. 2019).

Hyalomma aegyptium has a three-host life cycle where larvae and nymphs infest a wide range of hosts including lizards, birds, small mammals (Apanaskevich 2004; Kolonin 2004), cattle (Bos taurus; Aydn 2000), and humans (Vatansever et al. 2008; Bursali et al. 2010; Kar et al. 2013). However, adult H. aegyptium is host-specific to tortoises, particularly the spur-thighed tortoise (Hoogstraal and Kaiser 1960; Apanaskevich 2004; Široký et al. 2006; Tavasoli et al. 2007; Kalmàr et al. 2015). Our results add more evidence that adult H. aegyptium is the main tick species of spur-thighed tortoises. Taking into account that the spur-thighed tortoise is commonly used as a garden pet in Tunisia in addition to being the object of thriving illegal trade between North Africa and Europe, this species should be considered in wildlife management and conservation programs (Blanc 1978).

The tick H. aegyptium is known to be the vector of the hemogregarine Hemolivia mauritanica (Široký et al. 2009). This tick species has been found to be infected with other human and animal pathogens such as Anaplasama phagocytophilum, Ehrlichia canis, Coxiella burnetii (Paşitu et al. 2012), and Rickettsia aeschlimanii (Bitam et al. 2009). Borrelia turcica isolated from H. aegyptium collected from spur-thighed tortoises in northern Turkey (Ece et al. 2004) was shown to be transstadially transmitted during molting from nymph to adult stages, suggesting the vectorial capacity of this tick species (Kalmàr et al. 2015). Recently, CCHFv has been detected in H. aegyptium collected from spur-thighed tortoises captured in Algeria (Kautman et al. 2016) and in Syria (Široký et al. 2014). We previously showed that no CCHFv was detected in the remaining collected ticks from captured tortoises from northern Tunisia (Fares et al. 2019). Thus, further studies are needed to assess the impact of the high infestation intensity in spur-thighed tortoises by H. aegyptium on the potential transmission of zoonotic pathogens in the Mediterranean Basin.

This work was supported in part by the British charity Marwell Wildlife (registered in England and Wales [275433]) and by a grant from the US Civilian Research and Development Foundation grant OISE-16-62883-1 to E.Z.