Under the light: high prevalence of haemoparasites in lizards (Reptilia: Squamata) from Central Amazonia revealed by microscopy

Blood samples from 330 lizards of 19 species were collected to investigate the occurrence of haemoparasites. Samplings were performed in areas of upland (terrafi rme) forest adjacent to Manaus municipality, Amazonas, Brazil. Blood parasites were detected in 220 (66%) lizards of 12 species and comprised four major groups: Apicomplexa (including haemogregarines, piroplasms, and haemosporidians), trypanosomatids, microfi larid nematodes and viral or bacterial organisms. Order Haemosporida had the highest prevalence, with 118 (35%) animals from 11 species. For lizard species, Uranoscodon superciliosus was the most parasitised host, with 103 (87%; n = 118) positive individuals. This species also presented the highest parasite diversity, with the occurrence of six taxa. Despite the diffi culties attributed by many authors regarding the use of morphological characters for taxonomic resolution of haemoparasites, our lowcost approach using light microscopy recorded a high prevalence and diversity of blood parasite taxa in a relatively small number of host species. This report is the fi rst survey of haemoparasites in lizards in the study region. It revealed a high diversity of lizard haemoparasites and highlights the need to understand their impacts on hosts.


INTRODUCTION
The protozoologist Dr. Ralph Lainson (1992) two decades ago in his work on neglected parasites in the Amazonia basin quoted a phrase from P.C.C. Garnham, his former advisor: "There is a serious danger that malarial parasites become extinct." Since that time, very few efforts have been made to contain the threats to the diversity of these parasites and other organisms (Ferrante & Fearnside 2019). In fact, these threats have been aggravated by increased habitat destruction in recent years, particularly in tropical regions (INPE 2019). Extinction, alteration in the abundance or introduction of parasites can have profound impacts on the health of a large number of free-living species (Dobson et al. 2008), because parasites are ecologically involved in important mechanisms that regulate wildlife populations and structure communities (Tompkins & Begon 1999, Thomas et al. 2000. Moreover, they may infl uence their host biological processes, such as sexual selection (Ehman & Scott 2002, Megía-Palma et al. 2018, predation and competition dynamics (Schall 1992, Garcia-Longoria et al. 2015, as well as speciation and extinction processes (Anderson & May 1978, Poulin 1999, Prenter et al. 2004.
Reptiles are hosts for a wide variety of parasites, especially for diverse groups that parasitise blood cells (Davies & Johnston 2000, Telford 2009). These blood parasites may be intra-or extracellular organisms that range from protozoan kinetoplastids (Killick-Kendrick et al. 1986, Telford 1995 and apicomplexan parasites (Levine 1988, O'Donoghue 2017, to microfilarid nematodes (Thoisy et al. 2000, Halla et al. 2014) as well as viral and bacterial inclusions (Telford 2009). Except for the last two pathogens, whose transmission is not yet clear, the other three parasitic taxa share a common feature by using a range of haematophagous invertebrates as the main vectors for transmission between vertebrate hosts (Smallridge & Paperna 1997, Viana et al. 2012, Van As et al. 2015, Fermino et al. 2019. Furthermore, the haemoprotozoans of Phylum Apicomplexa Levine, 1970 are apparently the most studied of all and also represent the taxon with the largest number of species parasitising reptiles (Levine 1988). Only in lizards (Squamata: Sauria), approximately 14 genera were recorded (O'Donoghue 2017); haemogregarines and haemosporidians are the most frequently identified groups (Smith 1996, Perkins 2014. Although Brazil is a megadiverse country and has the third richest reptilian fauna in the world (Costa & Bérnils 2018), approximately 795 species, knowledge about haemoparasite diversity in these hosts consists of mainly a few concentrated studies in the eastern Amazon region (Lainson 1992(Lainson , 2012. These studies recorded a rich haematozoan fauna in lizards and also suggest that the Amazon biome has a great potential for the discovery of new haemoparasitic species in these vertebrates, as 29 (80%) of the 36 known protozoan species in the country occur in this region (Table I). However, these records are limited to a total of 20 lizard species (Table I), which represent 7% (n = 276) of the described Brazilian lizard fauna and 10% (n = 16/152) for the Amazon region (Costa & Bérnils 2018). This small number is probably due to the difficulties in collecting these hosts and also the lack of specialists interested in working with haemoparasites from herpetofauna.
Light microscopy is an important tool for diagnosing infections that has crossed centuries and generations of scientists, still being the fastest and most accessible technique for searching parasites (Halla et al. 2014). This is especially true for studies adopting horizontal approaches that aim to estimate parasitism in poorly known groups. In this sense, we sought to investigate using light microscopy the presence and diversity of haemoparasites in lizards from Central Amazonia.

Study area
The study was conducted in four upland (terrafirme) forest sites in Brazilian Central Amazonia, all located in the State of Amazonas, Brazil ( Figure  1). The first study area was the Federal University of Amazonas forest fragment campus (UFAM; 3º4'34"S, 59º57'30"W), located in the eastern part of the city of Manaus. The three remaining study areas were located, respectively, 38 km (UFAM Experimental Farm; 2°38'57.6"S, 60°3'11"W), 80 km (Biological Dynamics of Forest Fragments Project [BDFFP]; 2°25'S, 59°50'W), and 160 km (Agrovila Rio Pardo; 1°48'S, 60°19' W) north of Manaus. These sampling regions present a mean annual temperature of approximately 26°C with relative air humidity over 80% (Araujo et al. 2002). The yearly precipitation is over 2,000 mm and mostly concentrated in a rainy season that usually occurs from December to May (Marques-Filho et al. 1981). The vegetation of the sampling sites is mainly composed of a mosaic of upland Amazonian rainforest, which varies from primary and secondary forests to open areas. The average elevation is 40-160 m above sea level (Laurance et al. 2011). Some of these landscapes are relatively undisturbed (Deichmann et al.   (2018). The blood samples were obtained by tail or cardiac puncture using a sterile insulin syringe (Samour et al. 1984). A portion of collected blood was used to make smears, which were fixed with absolute methanol and stained with 10% Giemsa. The other portion was applied to a filter paper for molecular analyses. Lizards were released within 24 h of capture, but in the case of cardiac puncture, the blood was collected after euthanasia (via injection of 2% lidocaine). Specimens were preserved in 10% formalin and deposited in the Zoological Collections of the National Institute of Amazonian Research (INPA) and UFAM in Manaus, Brazil. Lizard sampling and access to the genetic data were authorised by the Brazilian Ministry of the Environment (SISBIO n° 53851-4 and SISGEN AA6199D, respectively). All procedures were approved by the ethics committee on animal use from Universidade Federal do Amazonas (protocol number 012/2016).

Microscopic analyses
Blood smears were examined for up to 20 min under a Leica DM4B microscope (Leica Microsystems, Heerbrugg, Switzerland) at × 400 and × 1000 total magnification. The slides with parasites were carefully examined and images were captured with an attached Leica DMC4500 digital camera and processed with LAS V4.8 (Leica Microsystems Suiza Limited 2015). Morphometric measurements were taken with this same system. However, they will not be presented here in this work, as they are part of ongoing taxonomic studies. Haematozoan parasites were taxonomically identified by comparing their morphologies to the descriptions from the guides of Telford (2009) and Lainson (2012), besides original description articles. Additionally, to confirm the identification of some haemosporidian species, we compared our material with that of the collection of Dr. Ralph Lainson, deposited at the Evandro Chagas Institute (IEC) in Belém, Brazil.
Among the positive lizards, Tropiduridae and Teiidae were the families that showed the highest prevalence, with 86% (n = 112/130) and 66% (n = 90/135) positive animals, respectively. With regard to lizard species, Uranoscodon superciliosus Linnaeus, 1758 stood out for presenting a high prevalence, with 87% (n = 103/118) of infected individuals, and also because it was the species with the greatest diversity of parasites, with the occurrence of six different taxa: Haemohormidiidae, Plasmodiidae, Garniidae, Trypanosomatidae, microfilarial worms and unidentified inclusions.
Parasites of phylum Apicomplexa (Table  III) were found in all infected lizard species; 14 species from five families were identified. Two morphotypes of the genus Hepatozoon  (Table III). Notably, U. superciliosus had the highest number of parasite occurrences, with 32 (27%; n = 118) positive specimens.
Haemosporidian parasites presented the highest prevalence, with 35% (n = 118/330) animals infected and, except for Alopoglossus angulatus Linnaeus, 1758, all positive host species were parasitised by malaria. Based on blood stage morphology, 13 species from two families, Plasmodiidae and Garniidae, were identified (Table III; Figure 2d-j). It is important to note that despite some authors (e.g., Levine
Extracellular parasites of the family Trypanosomatidae (Table II) were found in 83 U. superciliosus (70%; n = 118) and one P. umbra (8%; n = 12): each tropidurid species had one Trypanosoma morphotype. The trypanosome of U. superciliosus had an elongated body and diffuse nucleus (Figure 2k), while the observed P. umbra had a rounded shape and compact nucleus (Figure 2l). Microfilarial worms (Nematoda) occurred in five lizard species (Table II), with higher prevalence in A. ameiva with 37% (n = 27/72) positive specimens. These blood parasites exhibited highly variable sizes and shapes (Figure 2m-n) and were very similar to the genus Piratuba. However, accurate diagnoses of filarial worms is mainly based on morphological features of adult worms. Thus, identification of this group in the present study remains indeterminate.
The last of the four major groups, inclusions of uncertain nature (Figure 2o-p), were detected in erythrocytes of five lizard species and showed little morphological variation. They consisted of a large spherical shape with a rarely darker stained margin. These vacuoles resemble rickettsial parasites recorded for other reptilian hosts, although without ultrastructural study it was not possible to confirm this identification.

DISCUSSION
We observed a high prevalence of blood parasites among lizards from Central Amazonia: more than half of the sampled individuals and species were infected. We also demonstrated that lizards are the hosts for a wide variety of haemoparasites. Indeed, we observed great parasite richness in a small number of host species and in a limited sampling area. This finding reinforces that the neotropical region holds a rich haemoparasite fauna, as shown by studies conducted in other localities across the Amazon Basin (Renjifo et al. 1952, Telford 1970, 1973, 1980, Ayala et al. 1973, Lainson 1992, Thoisy et al. 2000, Matta et al. 2018. Furthermore, it is important to note that we sampled lizard species with diversified microhabitat use, ranging from terrestrial (e.g., A. ameiva), semi-aquatic (e.g., Neusticurus bicarinatus), scansorial (e.g., P. umbra) to arboreal (e.g., U. superciliosus) (Vitt et al. 2008). This environmental diversity may imply determinant characteristics for the composition of the haemoparasite assemblages found in these lizards because different species of vectors, including mosquitoes, sandflies and

Host species (N) Parasites n infected (% infected)
Alopoglossidae Alopoglossus angulatus ( (27) ticks, are likely distributed along the gradient occupied by these hosts. Parasite and host checklists are crucial in expanding our knowledge on species distribution. Nonetheless, surveys and descriptive studies of haemoparasite species on lizards conducted in the Amazonian biome have decreased considerably in recent years. In Brazil, it has been 10 years since a haematozoan species from a lizard has been described (Lainson et al. 2010). Most access to this haemoparasitic diversity in the Neotropics departs from the classical approach by using light microscopy to investigate prevalence and parasite identity. The exclusive use of morphological attributes for the diagnosis of species has been strongly criticised as unreliable because molecular tools have advanced in solving taxonomic problems (Pineda-Catalan et al. 2013, Perkins 2014. In fact, molecular biology constitutes a modern and acurate tool in parasitology, but its use still faces financial and technical limitations -i.e. difficulties in developing protocols and molecular markers-, especially in megadiverse and developing countries such as Brazil (Perkins et al. 2011, Morand 2018. Nevertheless, in comprehensive multi-species approaches like ours, whose main objective is not to solve systematic and phylogenetic questions, observations of blood smears under a microscope still prove to be a feasible method to access the prevalence and parasite diversity hidden in these hosts, despite some taxonomic limitations. Among the 12 infected lizard species, haemoparasites were recorded for the first time in three of them: Arthrosaura reticulata O'Shaughnessy, 1881, Norops planiceps Troschel, 1848 and A. angulatus. However, we did not find blood parasites in nine lizard taxa (Table II), even though parasites have already recorded in some of these hosts in other localities (Table   I). This discrepancy may simply reflect unequal sampling efforts. The methods we used for capture were diversified and effective for certain hosts groups, such as Teiidae and Tropiduridae, but are limited for many lizard species, mainly those that access subterranean microhabitats (Faria et al. 2019). Indeed, Teiidae, Tropiduridae and a lizard species, U. superciliosus, were the hosts with highest parasite prevalence. However, with the myriad known problems in obtaining samples (Perkins et al. 2011)-financial, technical and logistical difficulties in accessing remote areas-and the need to move forward on other parasitology research fronts, such as vectors and life cycle, landscape and epizootiology studies, those abundant taxa may be an interesting choice to be included in ecological parasitic systems as model organisms. Additionally, for many reasons lizards are considered model organisms (Huey et al. 1983, Camargo et al. 2010, as they respond very well when testing ecological and evolutionary hypotheses (Schall 1996).
Most of the parasites found in this study belong to phylum Apicomplexa. Indeed, all host species had some representative of this group. One of them was the genus Hepatozoon, relatively common parasite in reptiles and, despite the great diversity of lizards sampled in this study, was found exclusively infecting A. ameiva. Hepatozoon ameivae was described by Carini & Rudolph 1912 in A. ameiva in the State of Minas Gerais and later recorded in the municipality of São João da Barra, State of Rio de Janeiro, both in southeastern Brazil (Carini & Rudolph 1912, Sabagh et al. 2015. Lainson et al. (2003) also probably recorded H. ameivae in lizards from the municipality of Capanema, State of Pará, northern Brazil. This parasite has an outstanding feature: its gametocytes are found in the erythrocyte nucleus, a relatively uncommon developmental pattern in the Apicomplexa that can lead to severe distortion and even lysis of the infected cell nucleus (Telford 2009). It is important to note that H. ameivae found here was morphologically and molecularly characterized, and the analysis of its phylogenetic position clearly showed that this parasite belongs to the genus Hepatozoon (Picelli et al. 2020).
Our results showed a relatively low prevalence for Sauroplasma-like and we thought that positive lizard species were not previously recorded for piroplasms (Table III). Sauroplasma infections are common in lizards, even though there are only three species described for these hosts: Sauroplasma thomasi du Toit, 1938, Sauroplasma zonurum Pienaar, 1962, and Sauroplasma boreale Svahn, 1976(Telford 2009, Halla et al. 2014). In Brazil, these parasites were recently recorded in the freshwater turtle Podocnemis expansa (Picelli et al. 2016). Morphologically, they are small (2.5-4 µm) vacuole-shaped intraerythrocytic parasites with chromatin granules associated (Halla et al. 2014, Picelli et al. 2016. These morphological features mislead many authors to identify Sauroplasmalike inclusions as Chelonoplasma, Nuttalia or Pirhemocyton (Bardi et al 2019). They can also be overlooked as artefacts or bacterial and viral infections (Telford 2009). Parasitologists always pay attention to this conflicting taxonomic situation, but no molecular data is yet known for this genus.
H a e m o s p o r i d i a n w e re t h e m o s t predominant and richest taxon detected on lizards, mainly from Plasmodiidae parasites. It is well known that malaria parasites are widely distributed geographically, ubiquitous for most lizard families and are morphologically diverse, with over 100 species reported to infect reptiles (Schall 1996, Telford 2009). In the Eastern Brazilian Amazonia, 21 species of lizard malaria are known, and 13 (61%) of them were found in our research. For some of these (Garnia cf. uranoscodoni Lainson et al. 1975, Garnia cf. multiformes Lainson et al. 1975, Garnia cf. utingensis Lainson et al. 1971, this finding is the first occurrence record away from their type localities. Recently, Matta et al. (2018) reported the presence of Plasmodium kentropyxi Lainson et al., 2001 andPlasmodium carmelinoi Lainson et al., 2010 in Teiidae lizards, at a low prevalence, in the Colombia Orinoco-Amazon basin. The difference between our findings is that here P. cf. kentropyxi was found at a relatively high prevalence only in its type host, Kentropyx calcarata Spix, 1825. Another interesting species seen in our study is Plasmodium cf. tropiduri Aragão & Neiva, 1909. It was the only haemosporidian species found in two different host species: K. calcarata and Copeoglossum nigropunctatum Spix, 1825. This haemoparasite was one of the world's first reptilian malaria parasites, described by Aragão & Neiva (1909) in the lizard Tropidurus torquatus Wied-Neuwied, 1820. Since then, it was observed across many lizard families and can be considered one of the most widespread saurian malaria species in South America (Telford 2009). In fact, most of the haemosporidians present here were previously reported in other Amazonian locations (Telford 2009, Matta et al. 2018, evidence that malaria species rediscovered here may be widely distributed throughout the biome. Phylogenetic and phylogeographic studies that involve samples from different Amazonian localities may provide insights regarding the diversification and evolution of this group.
Parasites of the genus Trypanosoma were restricted to two lizard species and at a low prevalence when compared to protozoans of the phylum Apicomplexa, which parasitised more than half of the captured lizards. However, these flagellates were found in several U. superciliosus.
Both P. umbra and U. superciliosus already had trypanosomatids recorded by Walliker (1965) and Lainson et al. (1975), respectively. The first author provided a poor morphological description of Trypanosoma superciliosae Walliker, 1965 without reporting their prevalence in U. superciliosus from the municipality of Codajás, Amazonas State. Interestingly, Lainson et al. (1975) mentioned that they searched, in Pará state, for this parasite in a large number of U. superciliosus individuals but were unsuccessful. Nevertheless, the same authors described Trypanosoma plicae Lainson et al., 1975 in P. umbra. Besides these species, there is only one other species described for this genus on Brazilian lizards: Trypanosoma rudolphi, recorded just once in C. nigropunctatum (Carini & Rudolph 1912). This low species richness is probably due to the lack of studies conducted on these parasites in Brazilian lizards. Indeed, trypanosome species have been reported worldwide in lizards more than in any other reptilian group (Fermino et al. 2019). Although trypanosomes have a unique stage of their life cycle by circulating in the blood of reptiles, trypomastigote forms exhibit high polymorphism and plasticity (Spodareva et al. 2018). Therefore, it is not possible to confirm that we found the same species described for those hosts, even with some morphological similarities.
Our data revealed a low prevalence of microfilaria, which are larval stages from nematodes of the superfamily Filaroidea. These vector-borne parasite larvae are commonly found in the peripheral blood of vertebrates and here, except for U. superciliosus, all lizard species that we found positive for these parasites already had records for adult worms from many Onchocercidae species in other locations (Ávila & Silva 2010, Macedo et al. 2017. For U. superciliosus, the occurrence of microfilariae has been vaguely reported in eastern Amazonia and these studies did not provide morphological characterisation of these nematodes (Lainson et al. 1975). In reptiles, Oswaldofilariinae, a onchocercid subfamily, stands out as the main filarid group that parasitise these hosts. Some genera that infected lizards include Oswaldofilaria, Piratuboides and Piratuba (Pereira et al. 2010). Adult worms from this taxon are recognised by the long distance between the head and vulva, and a series of other characters are used for species identification (Pereira et al. 2010). Given that there is scarce information on their larval morphology and we did not collect data related to the adult phase of these helminths, we are unable to advance the identification of this group in this study.
One of the most intriguing findings of our work was the intraerythrocytic inclusions of an uncertain nature. These vacuole-like inclusions appeared at a low prevalence and resembled some bacterial infections, caused by Rickettsia, and also to the viruses of the Lizard Erythrocytic Virus (LEV) group, such as Pirhemocyton (Telford & Jacobson 1993, Telford 2009). In fact, pirhemocytonosis are commonly found in lizards, mainly green iguanas (Iguana iguana), as white square vacuole-like cytoplasmic inclusions (Harr et al. 2001, Halla et al. 2014). Viral or bacterial infections have been reported in many amphibians and reptiles across the world and some of them can cause diseases in these hosts (Davies & Johnston 2000, Ariel 2011). However, these organisms are poorly studied and their diagnosis can be complex because it involves several approaches, including electron microscopy, serological surveys and molecular tools (Ariel 2011). Unfortunately, our knowledge about these inclusions and its occurrence throughout the Amazonian biome is very limited, and therefore we were unable to deepen in their identification.
Parasites commonly co-occur in the same host (Vaumourin et al. 2015, Galen et al. 2019, and we detected a high prevalence of this interaction. Indeed, we observed the co-occurrence of very distinct groups of haemoparasites in terms of life cycles, evolutionary history and in the exploitation of their hosts. The presence of an infracommunity in a host may be the result of a random occurrence of these parasites or a consequence modulated by the existence of a previous infection (Vaumourin et al. 2015, Hernandes-Córdoba & Braga 2019. Meanwhile, there are several challenges to understanding these interactions. Most previous studies ignored them, and only recently has the importance of such multiparasitism been recognised (Vaumourin et al. 2015). For lizards, parasitic ecological systems are frequently based on the one-on-one interactions and focus mainly on ecology of coccidian or malarial parasitism (Schall 1996, Amo et al. 2005, Hernandes-Córdoba & Braga 2019, Megía-Palma et al. 2020. From our perspective, there is still a long and curious path to explore until we can better understand haemoparasites and their lizard hosts.
This study is the first multi-species haemoparasite survey performed on lizard assemblages in Central Amazonia. We also present the most complete and updated list of haematozoan species described for these hosts in this region. Furthermore, our lowcost investigation using light microscopy demonstrates that Central Amazonia has a high prevalence and significant diversity with potential for new records of haemoparasites, especially malaria species. These findings might support future taxonomic characterisation of the parasites reported here, as well as further studies in parasite ecology and evolution. At last, our work emphasizes the importance of screening parasites in wildlife animals to allow a better understanding of the biodiversity of this biome. ÁVILA