In Vitro and In Vivo Coinfection and Superinfection Dynamics of Mayaro and Zika Viruses in Mosquito and Vertebrate Backgrounds

The last several decades have seen a dramatic increase in the incidence of arboviral diseases, a trend attributed to increased international trade and transport, climate change, and urban crowding. Increasing landscape fragmentation and human activities are disturbing natural ecosystems, thereby promoting the spillover of new pathogens with epidemic potential into naive geographic regions (1). More than 150 different arboviruses have been reported in the neotropical region of the Americas (2), where they coexist and cocirculate in many areas through sylvatic and/or urban cycles (3, 4). Despite this diversity and cocirculation of multiple arboviruses in the same geographical areas, some pathogens often go unnoticed in both vector and host due to lack of surveillance, overlapping clinical symptoms, and/or the absence of specific diagnostic tests (5). Clinical studies have identified the cocirculation of multiple arboviruses such as dengue (DENV), Zika (ZIKV), Chikungunya (CHIKV), and Mayaro (MAYV) viruses in epidemic areas (6 to 9), but there is only one report of a field-collected mosquito carrying multiple arboviruses (10). However, this may be an artifact of methodology, as most surveillance programs analyze pooled mosquito samples to save time and resources, making it impossible to discern the number of coinfected mosquitoes and their potential contribution to the transmission cycle of those viruses. In contrast, coinfection and cotransmission of certain arboviruses have been demonstrated repeatedly under laboratory conditions (11 to 14).

Mayaro virus (MAYV; Togaviridae, Alphavirus) is an emerging neglected arbovirus thought to be transmitted by Aedes, Anopheles, and Haemagogus mosquitoes (15). Mayaro virus is actively circulating in Latin America as well as sporadically in the Caribbean (16, 17) and is on a trajectory of geographical expansion and increasing incidence, particularly in Brazil and Peru (17 to 19). For these reasons, MAYV was categorized as an emerging threat to public health by the Pan American Health Organization (PAHO) in 2019 (20). Zika virus (ZIKV; Flaviviridae, Flavivirus) is primarily transmitted by Aedes mosquitoes and emerged as a major arboviral public health concern due to its capacity to cross the placenta in pregnant women and severely harm the development of human fetuses (21, 22). The overlapping distributions of these two viruses and the ubiquity of Aedes vectors in Latin America (Fig. 1) both suggest the potential for coinfection of humans, possibly driven by the bites of coinfected mosquitoes. Indeed, there are several reports of MAYV patients coinfected with CHIKV (8, 19), ZIKV (23), and DENV (23, 24), which could potentially lead to double-infected mosquitoes. Aedes aegypti is especially worrying as a potential double-infection vector due to its highly anthropophilic behavior (25) and heightened vectorial capacity of ZIKV (26, 27). To date, there are no empirical data on Ae. aegypti coinfection with MAYV and ZIKV in nature. However, a recent study identified MAYV and DENV-4 (Flaviviridae, Flavivirus) in pools of Ae. aegypti collected in Mato Grosso (Brazil) (28), possibly indicating the presence of double-infected mosquitoes. Considering the vector competence of the species for both ZIKV (26, 27) and MAYV (15, 29), the overlapping distributions of pathogens and vector in Latin America, and the field-collected data above, we hypothesize that Ae. aegypti can sustain MAYV and ZIKV transmission cycles simultaneously.

Simultaneous exposure of a vector to multiple pathogens through a single infection event is referred to as coinfection (CI), while serial infection of the vector with different pathogens during sequential feeding events is referred to as superinfection (SI) (30, 31) (Fig. 2). While both are possible, superinfection may be more likely in nature for mosquito species that blood feed multiple times during a single gonotrophic cycle, show aggressive feeding behavior, and/or seek a second bloodmeal following incomplete feeding, as is the case for both Ae. aegypti and Ae. albopictus (32). There are three potential outcomes when multiple arboviruses co- or superinfect the same vector: (i) an antagonistic interaction that results in partial or total inhibition of one or both viruses, (ii) a synergistic interaction that enhances one or both viruses, or (iii) a neutral coexistence of both viruses without any alteration of their viral fitness (33). Here, we coinfect and superinfect Ae. aegypti mosquitoes with MAYV and ZIKV to study their interactions and consequent effects on vector competence. We also investigate the interaction kinetics of CI and SI in vertebrate and invertebrate cells, and we demonstrate that coinfection occurs at the level of single cells.

Here, we examine MAYV and ZIKV interactions in a mosquito vector as well as in vitro in mosquito, primate, and human cells. We find that adult mosquitoes and all tested cell lines can become infected with MAYV and ZIKV simultaneously. Moreover, the circumstances of infection—i.e., whether the two pathogens grew as single infections, coinfections, or superinfections—had significant effects on viral growth in multiple contexts.

Following in vivo coinfection with MAYV, we consistently found reduced ZIKV infection rates. This finding is similar to previous work, where Rückert et al. (12) demonstrated a reduction in ZIKV IR when coinfecting with CHIKV, an Alphavirus belonging to the same antigenic complex as MAYV (Semliki Forest virus complex). Similarly, Muturi et al. (31) demonstrated that replication of DENV-4 was suppressed by Sindbis virus (Togaviridae, Alphavirus) in coinfected Ae. albopictus. In contrast to ZIKV, we did not observe any variation in the IR of MAYV during coinfection with ZIKV. Our data for MAYV are consistent with previous studies of Flavivirus/Alphavirus coinfection, which showed the IR and DIR of CHIKV were not influenced by the presence of ZIKV or DENV-2 (12, 35). Taken together, these data suggest that in coinfected mosquitoes, the IR of flaviviruses are negatively affected by the presence of an Alphavirus. However, this effect is not reciprocal, as the success of an alphavirus was not determined by the presence of a flavivirus in our study. We did not find any statistically significant differences in the TEs or respective saliva viral titers of MAYV or ZIKV between single and coinfected groups. Consistent with previous studies of Chikungunya/ZIKV coinfection (12, 13), our data suggest that the simultaneous intake of both viruses may negatively impact their capacity to infect the midgut and subsequently disseminate through the mosquito body, but the capacity for salivary transmission remains unaffected.

We show that Ae. aegypti can become infected and transmit both pathogens following sequential exposure (i.e., superinfection), regardless of the infection order. In general, our data suggest that vectors are more permissive to double infection from sequential exposure, rather than from simultaneous exposure. Although we analyzed a relatively small number of animals, our data nevertheless demonstrate the capacity of the vector to cotransmit both pathogens in its saliva. Transmission of both pathogens was slow, occurring only 14 days after the second infection event. This implies that, based on our SI model, the vector may need to survive for approximately 26 to 28 days in natural conditions to cotransmit MAYV and ZIKV. Field observation and laboratory experiments indicate Ae. aegypti can survive even longer (36 to 38), but because its daily survival rate decreases sharply over time, the probability of double transmission may also drop precipitously (39). Additional experiments—e.g., to assess the mosquito’s extrinsic incubation period or time points between 7 and 14 days after the second infection (i.e., 9 and 12 dpi)—are needed to understand the minimum time required for cotransmission.

Interestingly, we observed an increase of MAYV IR in previously ZIKV-infected mosquitoes (ZM group). This result indicates a positive effect of a previous ZIKV infection on MAYV’s ability to stably infect the midgut. The underlying molecular mechanism remains unclear. However, one possibility is that the previous ZIKV infection could have altered the mosquito immune response and thereby created a more suitable environment for MAYV replication, similar to what has been hypothesized for CHIKV-ZIKV superinfection (14). Previous studies demonstrated that a subsequent meal with uninfected blood can increase the DIR in infected mosquitoes and potentially enhance their TE (40 to 42). This effect may potentially arise from physical expansion of the midgut following a bloodmeal, which enlarges pores in the basal lamina and allows viral particles to pass through (40, 43). To investigate whether an increase in DIR in our experiment could be influenced by the bloodmeal itself and not by the pathogen, we included control groups fed with uninfected bloodmeals before (U-) or after (-U) viral challenges (UM, UZ, MU, and ZU). Our results did not show an effect of uninfected bloodmeals on DIR, even though our assay used a longer interval between the two bloodmeals (7 days) than the previous study reporting an effect (3 days) (42). This temporal gap was likely sufficient for the basal lamina to revert to normal conformation, which may limit the window during which a second bloodmeal can affect the DIR. Further experiments, including different temporal gaps between bloodmeals and a more in-depth structural analysis of the basal lamina, are needed to investigate this hypothesis.

Our experimental model assumes that the vector becomes coinfected or superinfected with a similar titer of both viruses, a scenario that may not reflect actual field conditions, where the possible combinations of viruses and their respective titers may be incredibly variable. Relevant studies have found variable effects when infection parameters were altered. Muturi et al. (31) showed that varying the virus titer ratio between DENV-4 and SINV affected the replication of one or both viruses in an in vitro SI study. Those data suggest that the relative amounts of different pathogens could influence the outcome of CI or SI experiments—a parameter we did not test. Additional studies would be needed to evaluate the impact of different titer combinations in CI and SI double-infection routes.

The cellular and molecular basis of MAYV and ZIKV coinfection and superinfection dynamics remains unclear. Our in vitro observations demonstrate that dual infection by these viruses within individual cells of both invertebrates and vertebrates is possible. However, we noticed several cases of viral interference. Specifically, we observed inhibition of ZIKV replication in mammalian cells during CI and MZ superinfection. At the same time, partial inhibition was recorded in ZM superinfection. Interestingly, when the SI experiment in mammalian cells featured a 2-h incubation period between sequential infections (instead of 12 h), ZIKV was able to replicate in MAYV-infected cells, though it could not do so following 12 h of incubation, highlighting that the effect is time dependent. Interference of arbovirus replication by another arbovirus is influenced by several principal parameters: advantage time of the first virus, order of exposure of the viral pathogens, and difference in MOI (44). Different hypotheses have been proposed to explain viral interference mechanisms, including competition for replication sites and cellular substrate resources (44), transacting proteases induced by the first virus, and superinfection exclusion (45, 46). The latter occurs when a cell infected by a virus becomes refractory to subsequent infection by the same or a closely related virus (referred to as a homologous virus) (45). Superinfection exclusion can occur at different steps of the replication cycle, as demonstrated by several studies using Semliki Forest virus (Togaviridae, Alphavirus) and CHIKV, including binding and internalization, viral genome transcription, and protein translation (47 to 50). In a recent study, Boussier et al. (50) showed that CHIKV excludes influenza A virus (Orthomyxoviridae, Alphainfluenzavirus) replication in vitro, and demonstrated the capacity of alphaviruses to interfere with the replication cycle of different viral families. Several in vivo studies also suggest that sequential infection of Alphavirus and Flavivirus results in coinfection, not in exclusion (13, 51, 52), although none of them have examined MAYV in conjunction with a flavivirus.

We observed coinfected cells using immunofluorescence microscopy, demonstrating that a single cell can be simultaneously infected with both viruses (Fig. 5A and B, third row). We observed a cytopathic effect in mammalian cells caused by MAYV after the first 24 to 36 h postinfection (Fig. S2), similar to other alphaviruses (13, 45, 53). The loss of structural functionality and the partial destruction of cells may partially prevent infection by ZIKV and/or its subsequent replication, especially when the interval between infections is longer (e.g., SI with a 12-h initial incubation time). The replication cycle of alphaviruses is relatively short, and progeny virions can be detected 4 to 6 h postinfection (54 to 56). Conversely, 8 to 12 h are generally needed to detect progeny virions of flaviviruses (57). This disparity in viral kinetics could have influenced the capacity of ZIKV to replicate in the presence of MAYV, which rapidly induces a severe structural and morphological change to the infected cells. Superinfection with a shorter incubation time (2 h) likely limited the impact of MAYV-induced cell damage, which may underlie the smaller effect of MAYV upon ZIKV in the 2-h experiments. However, across all tested experimental conditions, we found MAYV inhibits ZIKV replication to some degree, suggesting viral interference at a certain replication stage. We can therefore speculate that, when introduced in vertebrate cells before or together with ZIKV, MAYV takes over the cellular transcription and/or translational machinery, leaving minimal resources for ZIKV replication. Additional studies are required to understand the molecular dynamics of CI and SI with MAYV and ZIKV in both invertebrate and vertebrate models.

Coinfection and superinfection are poorly studied in mosquito vectors. Despite evidence of circulation of multiple pathogens in the same geographical areas, no data are currently available about prevalence of coinfected vectors in the field. The implementation of precise sampling strategies for field-collected mosquitoes would be fundamental to investigate the prevalence of multi-infected vectors, to understand the importance of CI and SI in virus transmission cycles and on mosquito fitness (if any), as well as an in-depth description of the interactions of pathogens at the cellular and molecular level. Future work could also examine the effects of CI and SI on other vector-borne pathogens, such as Plasmodium or microfilariae nematodes.

We thank the personnel of Eva J. Pell ABSL-3 Laboratory for Advanced Biological Research for their help and technical support, Craig Cameron (University of North Carolina) and Elizabeth McGraw (Pennsylvania State University) for cell lines, and Sage McKeand (Pennsylvania State University) for Fig. 2 design. The following reagents were obtained through BEI Resources, NIAID, NIH, as part of the WRCEVA program: Mayaro Virus, BeAr505411, NR-49910; Zika virus, MR766, NR-50065, Monoclonal Anti-Chikungunya Virus E2 Envelope Glycoprotein, Clone CHK-48 (produced in vitro), NR-44002, monoclonal anti-Flavivirus group antigen, Clone D1-4G2-4-15, NR-50327.

This research was funded by NIH grants R21AI128918, R01AI116636, and R01AI150251, USDA Hatch funds (accession number 1010032; project number PEN04608), and funds from the Dorothy Foehr Huck and J. Lloyd Huck endowment to J.L.R. C.A.H. was supported in part by an NSF Graduate Research Fellowship Program award (ID 2018258101). D.K. was supported in part by a National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT; 2019R1G1A1100559).