Worldwide transmission and infection risk of mosquito vectors of West Nile, St. Louis encephalitis, Usutu and Japanese encephalitis viruses: a systematic review

Field approach

Our searches uncovered 301 papers reporting field studies. After screening the titles abstracts, and full texts, we kept 130 articles for the analysis (Supplementary Fig. 1), from which we obtained 1342 observations regarding 57 Cx. mosquito species from 28 countries and 135 localities (Fig. 1A). Of these 1342 observations, 733 (54.61%) were classified as high quality, (i.e., the number of individuals tested was specified) (Supplementary Tables 1 and 2). The best represented countries were the USA (64.7%, number of observations = 869), Italy (9.3%, n = 125), and Iran (2.9%, n = 39). Based on mosquito field surveillance and individuals testing positive, we concluded that JES is distributed mainly in the Nearctic, Palearctic and Oriental regions (Fig. 1A).

West Nile virus

WNV was detected mainly in the USA (76.5%, number of observations = 826), Italy (4.9%, n = 53) and Iran (3.6%, n = 39) (Fig. 1A). We also recorded 23 species (57%, 41 species) interacting with this virus (Supplementary Tables 1 and 3). Cx. quinquefasciatus became naturally infected in North America [Infection Frequency (IF) = 2.33] (Table 1). We recorded WNV interacting with Cx. tritaeniorhynchus in Asia (IF = 1.02), with Cx. pipiens in Europe (IF = 1.74) and with Cx. antennatus, Cx. neavei, Cx. perexiguus, Cx. perfuscus, Cx. poicilipes, Cx. quinquefasciatus and Cx. tritaeniorhynchus in Africa (IF = 1) (Supplementary Table 3).

The highest infection rates were found in North America in Cx. restuans [Standardized minimum infection rate (SMIR) = 56.01], and in Africa and Europe in Cx. pipiens (SMIR = 20.45 and 29.25, respectively). No positive SMIR values were reported in Asian mosquitoes, and Oceanic mosquitoes were not sampled for this virus (Fig. 2A and Supplementary Table 3). The highest infection risk or potential was recorded in species from the USA, such as Cx. restuans (Infection Risk (IR) = 69.50), Cx. pipiens (IR = 55) and Cx. tarsalis (IR = 52.16) (Fig. 3A and Supplementary Table 3). Finally, WNV lineage 1 was detected in Algeria, Turkey, Portugal, Mexico, Tunisia, Iran, Spain and Italy, lineage 2 in Italy, Bulgaria, Greece and the Czech Republic, and lineage 5 in India (Supplementary Table 2).

Japanese encephalitis virus

JEV was detected mainly in Taiwan (21.6%, n = 24), Korea (18%, n = 20) and Australia (15.3%, n = 17) (Fig. 1A). We found 23 mosquito species interacting with JEV: Cx. vishnui was the one most frequently found to be positive (IF = 1.20), followed by Cx. tritaeniorhynchus (IF = 1.17), Cx. pipiens and Cx. annulus (IF = 0.98) in Asia, while the most susceptible species in Oceania were Cx. sitiens and Cx. gelidus (IF = 0.71) (Supplementary Table 3).

The highest SMIR values were recorded in Asia in Cx. rubithoracis (62.38), Cx. annulus (47.68) and Cx. tritaeniorhynchus (28.16) (Fig. 2A and Supplementary Table 3), and Cx. fuscocephala had the highest estimated natural IR (Fig. 3A, Supplementary Table 3). Three genotypes were recorded: genotype I (strain VNKT/479/2007, VNKT/486/2007, and JEV Ishikawa12), genotype III (Tibet-Culex-JEV1-5), and genotype V (K12YJ1174). These were isolated in China, Vietnam, and Japan (Genotype I), Italy, China (Genotype III), and Korea (Genotype V) (Supplementary Table 2).

Usutu virus

Field studies on USUV have been conducted in Europe, Africa, and Asia, most of them in Italy (66.6%, n = 72), Czechoslovakia (11.1%, n = 12) and Slovakia (7.4%, n = 8) (Fig. 1A). Six species were reported to be susceptible to natural infection. Cx. perexiguus had the highest IF and SMIR (1.30) (Supplementary Table 3). In Africa, Cx. antennatus (IF = 1), and in Asia Cx. pipiens (IF = 1) were the most likely to be positive, while Cx. pipiens had the highest IR (5.19) (Fig. 3A, Supplementary Table 3). The recorded strains were USU181_09/USU090-10/USU173_09 (Italy) and USU/Croatia/Zagreb-102/2018 (Italy).

St. Louis encephalitis virus

The field studies on SLEV focused on North America (97.7%, n = 43) and Brazil (2.2%, n = 1). Three species were recorded interacting with this virus. Cx. erraticus had the highest IF (2.06), SMIR (2.06) and IR, followed by Cx. quinquefasciatus (North America) (IF = 0.73, SMIR = 1.97) (Fig. 2A, Supplementary Table 3).

The highest estimated IR of JES was for Cx. pipiens (Europe), which can be naturally infected with WNV and USUV, followed by Cx. quinquefasciatus (North America), which can be infected with WNV and SLEV (Fig. 3A).

Experimental approach

Experimental studies were reported in 481 articles. After screening the titles, abstracts, and full texts, as well as opportunistic records, 95 articles remained for the analysis (Supplementary Fig. 2). From these we obtained 189 high quality observations of the TE of JES in 11 countries, 40 localities, and 21 species (Fig. 1B, Supplementary Table 1). The USA was the best represented country (54.4%, n = 103), followed by Germany (13.2%, n = 25) and Australia (12.6%, n = 24). There was, however, a notable lack of information on the vector competence of Cx. mosquitos for JES in many regions of the world, such as Central and South America, and Africa (Fig. 1B).

The most common means of infection was oral (94.8%, 395 observations), while the rest were intrathoracic. Intrathoracic infection bypasses the midgut barrier so is not considered natural infection. We therefore carried out the subsequent analyses using only the data on oral infection (Supplementary Table 4).

We used a generalised linear model (GLM) for the statistical analysis, which was conducted only on the WNV dataset (strain NY99), the only one with sufficient observations for the purpose (n = 63). We did not find a significant effect of viral titre, temperature, or days post infection on TE. However, more data with a wide range of values is necessary to confirm these observations. On the other hand, we found that the Extrinsic Incubation Period (as DPI) was shorter at higher temperatures (Fig. 4 and Supplementary Table 5).

West Nile virus

Mosquito populations from many locations on all continents have been studied for their vector competence for this virus, particularly in the USA (60.3% of observations, n = 96), Germany (15.7%, n = 25) and Australia (6.9%) (Fig. 1B). Our bibliographic research revealed 21 species of Cx. with the ability to transmit WNV under laboratory conditions (Supplementary Table 6). Cx. pipiens (North America) and Cx. tarsalis were the most frequently studied species and were the most efficient in transmitting the virus (Transmission Frequency (TF) = 2.33) (Table1). Cx. quinquefasciatus had the highest TF (1.70) in Africa, Cx. modestus in Europe (TF = 1.32), and Cx. annulirostris and Cx. quinquefasciatus in Oceania (TF = 1.48) (Supplementary Table 6).

Concerning Standardized Transmission Rates (STE) estimates, Cx. quinquefasciatus had the highest values in the USA (STE = 1.63), Cx. pipiens in Europe (0.90), Cx. tritaeniorhynchus in Asia (1.8), Cx. neavei in Africa (0.17) and Cx. annulirostris in Oceania (2.45) (Fig. 2B, Supplementary Table 6). We found 20 different strains of WNV tested. The TE of the various WNV strains vary considerably, but lineage 1 was more efficient than lineage 2. There were also more studies on the lineage 1 strains (n = 11), which exhibited high variation (Fig. 5).

Japanese encephalitis virus

JEV has been studied mainly in mosquito populations from France (45%, n = 20) and Australia (34%, n = 15), but also the United Kingdom, India, Taiwan, New Zealand, and the USA (Fig. 1B). Six mosquito species are capable of transmitting JEV. Cx. pipiens (Europe) had the highest TF (1.85), while Cx. gelidus had high values of STE (1.73) (Fig. 3B and Supplementary Table 6).

St. Louis encephalitis virus

Vector competence for SLEV has been studied in two countries: the USA (93.3%, n = 42) and Argentina (6.6%, n = 3), and 7 mosquito species have been investigated. Cx. nigripalpus was the most efficient in transmitting the virus (TF = 1.60), while Cx. pipiens had the highest STE (0.68) (Figs. 2B, 3B and Supplementary Table 6).

Usutu virus

Studies have also been conducted on the Usutu virus in mosquito populations in the USA (28.57%, n = 4), the United Kingdom (42.8%, n = 6) and Senegal (25%, n = 4), in particular on Cx. neavei, Cx. pipiens and Cx. quinquefasciatus (TF = 1). Cx. neavei also had the highest STE (0.79) (Fig. 3B).

We found reports of JES transmission under laboratory conditions in 22 Cx. species, and natural infections in 32 species (55.1% of the total sample) in the field. Cx. pipiens complex (biotypes quinquefasciatus, pipiens, molestus and pallens) was the most common vector accounting for 36.9% (n = 660) of the experimental observations and 25.7% (n = 1342) of the field observations. With both approaches, WNV was the most common flavivirus, accounting for 80.4% of the field observations and 86.7% of the experimental data (Fig. 1A,B). Only WNV, therefore, had enough observations to make comparison between the experimental and field data possible. We were able to compare 16 mosquito species and found a high positive correlation between TF and IF (R = 0.57, p = 0.02) (Fig. 7).

In summary, we found that the species with the highest infection-transmission risk (IRT) for WNV was Cx. restuans, for USUV it was Cx. pipiens (Europe), for SLEV Cx. quinquefasciatus (North America), and for JEV Cx. gelidus (Oceania) (Fig. 6 and Supplementary Tables 2 and 6).

Source: Ecology - nature.com