Retrospective epidemiologic and genomic surveillance of arboviruses in 2023 in Brazil reveals high co-circulation of chikungunya and dengue viruses

Abstract

Background

The rapid spread and increase of chikungunya (CHIKV) and dengue (DENV) cases in Brazilian regions in 2023 has raised concerns about the impact of arboviruses on public health. Epidemiological and genomic surveillance was performed to estimate the introduction and spread of CHIKV and DENV in Brazil.

Methods

This study obtained results from the Hermes Pardini (HP), a private medical laboratory, and the Health Department of Minas Gerais state (SES-MG). We investigated the positivity rates of CHIKV and DENV by analyzing the results of 139,457 samples tested for CHIKV (44,029 in 2022 and 95,428 in 2023) and 491,528 samples tested for DENV (163,674 in 2022 and 327,854 in 2023) across the five representative geographical regions of Brazil. Genome sequencing was performed on 80 CHIKV and 153 DENV samples that had been positive for RT-PCR tests.

Results

In our sampling, the data from CHIKV tests indicated that the Northeast region had the highest regional positivity rate in 2022 (58.1%). However, in 2023, the Southeast region recorded the highest positivity rate (40.5%). With regard to DENV, the South region exhibited the highest regional positivity rate in both 2022 (40.8%) and 2023 (22.7%), followed by the Southeast region in both years (34.8% in 2022; 21.4% in 2023). During the first 30 epidemiological weeks of 2023 in the state of Minas Gerais (MG), there was a 5.8-fold increase in CHIKV cases and a 3.5-fold increase in DENV compared to the same period in 2022. Analysis of 151 new DENV-1 and 80 CHIKV genomes revealed the presence of three main clusters of CHIKV and circulation of several DENV lineages in MG. All CHIKV clades are closely related to genomes from previous Brazilian outbreaks in the Northeast, suggesting importation events from this region to MG. We detected the RNA of both viruses in approximately 12.75% of the confirmed positive cases, suggesting an increase of co-infection with DENV and CHIKV during the period of analysis.

Conclusions

These high rates of re-emergence and co-infection with both arboviruses provide useful data for implementing control measures of Aedes vectors and the urgent implementation of public health politics to reduce the numbers of CHIKV and DENV cases in the country.

Background

Arboviruses (arthropod-borne viruses) are a paraphyletic set of viruses, with a significant number of species presenting medical and/or veterinary importance transmitted by mosquitoes, ticks or sandflies. These diseases, such as Zika virus (ZIKV), chikungunya virus (CHIKV), and dengue virus (DENV) are spread by Aedes mosquitoes [1, 2]. In Brazil, during the first nine months of 2023, 1,530,940 and 143,739 cases of DENV and CHIKV have already been confirmed, respectively [3].

CHIKV is an Alphavirus, member of the family Togaviridae, that can be classified into four lineages: West African, Asian, Indian Ocean (IOL), and East/Central/South/African (ECSA). It is an enveloped virus with a single-stranded positive-sense RNA about 11 kb in length. The genomic sequence is composed of two open reading frames (ORFs) located between the 5′ and 3′ untranslated regions (UTRs). The coding sequences encompass genes encoding polyproteins that are processed into different non-structural (nsP1-nsP4) and structural proteins (capsid, E1, E2, E3, and 6 K) [4,5,6,7]. Chikungunya disease is characterized by an acute illness with a variety of symptoms, encompassing fever, headache, polyarthralgia, arthritis, myalgia, and rash.

CHIKV was first isolated from the blood serum of a patient in 1952 in Tanzania [8]. In late 2004, an outbreak initiated in Kenya, spread worldwide when a strain from the ECSA lineage expanded. The virus was introduced in the Americas in 2013. In Brazil, the virus was described in September of 2014 through two independent introduction events: the first one by the Asian lineage in the Amapá state (North region) and the second in the Bahia state (Northeast region) by the ECSA lineage. After the introduction, many cases were reported in the country, and in 2017, the ECSA lineage was the most predominant lineage found in Brazil [9,10,11]. Due to its tropical climate, Brazil has become one of the main reservoirs for CHIKV re-emergence and endemic transmission.

DENV belongs to the Flavivirus genus, a member of the Flaviviridae family. This virus has a single-stranded positive-sense RNA genome, responsible for codifying three structural proteins and seven non-structural proteins in a single ORF. The UTRs at the 5′ and 3′ ends are crucial for viral replication and translation [1]. Four distinct serotypes (DENV-1, DENV-2, DENV-3, and DENV-4) were identified through neutralization assays [12]. Genomic classifications of these serotypes include six genotypes in DENV-1 (1I-VI), six in DENV-2 (2I-VI), five in DENV-3 (3I-V), and five in DENV-4 (4I-V) [13,14,15,16]. Dengue stands as the foremost viral disease transmitted by arthropods with significant public health implications. From being reported in just nine countries during the 1950s, its geographical range has expanded to over 100 countries across the globe [17]. DENV symptoms can vary from a mild flu-like syndrome to the most severe forms of the disease. Dengue is classified in three groups: dengue with or without warning signs (abdominal pain, mucosal bleeding, lethargy, liver enlargement and others) and severe dengue. Severe dengue presents severe plasma leakage, bleeding, or any organ failure [18].

Many factors contributed to the burden and dissemination of arboviruses globally, such as population growth and urbanization, the spread of Aedes spp. mosquitoes, a lack of vector control and effective treatment, and increased human mobility [19,20,21,22]. Recent outbreaks of arboviruses around the world highlight the importance of tracing the main routes of dissemination for these viruses. Genomic surveillance has proven to be a reliable tool for studying spread of arboviruses over the years [20, 23,24,25,26,27]. Additionally, understanding the diversity of arbovirus genotypes circulation in the country can aid health services in monitoring the epidemiology of these diseases [23, 24]. Thus, genomic surveillance, linked to monitoring the epidemiology of the diseases, is of paramount importance to investigate the dispersion and main areas of transmission risk. Herein, we performed a genomic and epidemiologic surveillance of CHIKV and DENV in Brazilian territory, focusing on Minas Gerais (MG), which reported the highest number of confirmed cases of CHIKV and DENV during the period of the study.

Methods

Ethics aspects

The study was approved by the Research Ethics Committee (CAAE- 33,202,820.7.1001.5348). The authorization allows access to epidemiological and viral data with exemption to the consent form in samples from regular viral diagnosis.

Epidemiological analyses of CHIKV and DENV in Brazil

The results of 66,150 samples tested for CHIKV and 157,564 samples tested for DENV at the Hermes Pardini Institute (HP) were subjected to analysis (see Additional File 1 Table S1). HP is a private Brazilian medical laboratory with numerous units distributed across all geographic regions of the country. The samples were analyzed at HP between January 2022 and June 2023 across Brazil`s five geographical regions: North, Northeast, South, Southeast, and Midwest. The number of samples contributed by each region varies considerably, given that the laboratory units of HP are more densely concentrated in the Southeast region. Molecular tests for CHIKV and DENV were performed using an in-house assay, ZDC Multiplex RT-PCR Assay (Bio-Rad Laboratories), and the Molecular ZC D-Tipagem Kit (Bio-Manguinhos, FIOCRUZ, Brazil). The antigen test for DENV was performed by ELISA for the NS1 antigen (dengue NS1, Euroimmun Medizinische Labordiagnostika AG) following the manufacturer’s instructions. Arboviruses serological tests were conducted by detecting IgM against CHIKV and DENV (chikungunya IgM or anti-dengue Type 1–4 IgM, Euroimmun Medizinische Labordiagnostika AG/Germany and dengue IgM Bioline, Abbott Laboratories, USA) according to the manufacturer’s instructions. Samples that tested positive on either a molecular or serological test were included in the study and evaluated independently. Samples tested for IgM were not assessed by RT-qPCR or NS1, and vice versa. For each virus, were calculated the positivity index by dividing the number of positive results by the total number of samples analyzed. The national positivity is calculated based on the aggregate of all samples collected across Brazil, whereas the regional positivity rate is based on samples collected solely within that specific region, given the varying contributions of each region in terms of the number of tests conducted. The proportion of positive tests across different age groups and sexes was determined by dividing the number of positive cases within each age and sex group by the total number of positive results.

CHIKV and DENV frequencies and mortality in MG state

Confirmed cases and deaths were obtained from all 853 municipalities in the state of MG during the study period. Confirmed deaths were evaluated by analyzing clinical and epidemiological data from individuals who tested positive for CHIKV or DENV through molecular (re-qPCR), serological tests (ELISA-IgM), or virus isolation. For death cases, histopathological and immunohistochemical exams were performed for corroboration.

The case and death data were obtained from the Health Department of Minas Gerais state (SES-MG). Spatiotemporal heatmaps were constructed using the number of cases in each city of MG per epidemiological week. A statistical model was used to estimate the interpolation and dispersion of arbovirus cases in the state.

DENV serotyping and co-infection investigation

DENV serotyping (1–4) and arboviruses co-infection (DENV, CHIKV, and ZIKV) were investigated by RT-qPCR multiplex panels from 243 positive samples from 2023 outbreak (epidemiological weeks 3 to 17). The molecular ZC D-Tipagem Kit (Bio-Manguinhos, FIOCRUZ, Brazil) is a triplex assay that detects and differentiates the ZIKV, CHIKV, and the four serotypes of DENV (DENV 1–4) by targeting specific viral sequences with different probes and fluorophores. Since the ZC D-Tipagem Kit also detects ZIKV, the samples were screened for this virus as well. The results were analyzed using the Design and Analysis software v.1.5.2 (Thermo Fisher, MA, USA) according to genotyping kit recommendations. CHIKV and DENV co-infections were investigated by the same multiplex panel, and samples with a cycle threshold (Ct) < 38 for a given target were considered positive for the investigated arbovirus (Additional File 1: Table S1).

Sequencing, genome assembly, and classification

Positive samples diagnosed by the HP laboratory for each tested virus or serotype were screened based on their Ct using Molecular ZC D-Tipagem Kit (Bio-Manguinhos, FIOCRUZ, Brazil). Positive samples from 2023 outbreak presenting Ct value ≤ 35 were selected for whole genome sequencing and DNA library preparation using the reagents from the Illumina® COVIDSeq Test Kit, with adaptations [25, 26]. Viral cDNA synthesis was performed following the COVIDSeq Test Kit recommendations, and specific primer sets (pool 1 and pool 2) covering the whole genome of each virus were used for genome amplification (Additional File 2: Table S2), following Artic and CADDE protocols. The final concentration of primers in the reaction was 1.72 μM (final volume of 25 μL) (https://www.caddecentre.org/protocols/) [25, 26]. For each virus, a pool containing all libraries was assembled and quantified using the QIAseq Library Quant Assay Kit (QIAGEN, NW, Germany). Sequencing was carried out on MiSeq System with a v3 cartridge (600 cycles) (Illumina, San Diego, CA, USA).

For genome assembly, a custom pipeline was used to process the data [27]. The tool ViralUnity includes three main steps to obtain the consensus genome. First, adapters, primer sequences, and short and low-quality reads (Phred score < 30) were removed by Trimmomatic v0.39 [28]. The remaining reads were mapped against the reference genome (Genbank accession numbers for CHIKV: KP164568.1, DENV-1: MW208056 and DENV-2: NC_001474) with Bowtie2 [29]. Mapping files were generated and sorted using SAMtools [30]. The consensus genome was generated with BCFtools [26], and low-coverage sites were masked with BEDtools [31]. The code for the described pipeline is available on GitHub (https://github.com/filiperomero2/ViralUnity). Consensus sequences were then classified using the Genome Detective platform (https://www.genomedetective.com/db/ui/login).

Phylogenetic analysis

Two datasets were created to contextualize the new genomes generated in our study: DENV-1 and CHIKV. All CHIKV or DENV-1 genomes with more than 8000 nucleotides, along with metadata related to date and location of sample collection available in NCBI database (https://www.ncbi.nlm.nih.gov/labs/virus/vssi/#/), were downloaded. The DENV-1 sequences were retrieved using the aforementioned criteria on June 27, 2023, and the CHIKV sequences were retrieved on July 3, 2023. The genomes of each dataset were aligned with the newly generated genomes and the reference genome of CHIKV (KP164568.1) or DENV-1 (MW208056) using MAFFT v7.480 [32]. Each alignment was manually checked and trimmed at the ends and in intergenic regions. Maximum-likelihood phylogenetic trees were estimated using the IQ-tree v2.1.2 program [33]. Model-Finder was used to estimate the best-fit substitution model, and the Shimodaira-Hasegawa-like approximate likelihood ratio test (SH-aLRT) was chosen to quantify the node support. A time-calibrated phylogenetic tree was constructed for each dataset (CHIKV and DENV-1) using the TreeTime program v. 0.11.1 [34], and a nextstrain build was created for visualization of each dataset (https://github.com/InstitutoTodosPelaSaude/flexpipe/tree/main).

Results

Epidemiological data of Arboviruses in Brazil

A total of 312 positive results (38.1%) for CHIKV were obtained from 818 RT-qPCR tests conducted on samples from all Brazilian regions (Fig. 1A) over the entire study period (Additional File 1: Table S1). Focusing on the first 6 months of each year, the laboratory reported 122 CHIKV-RT-qPCR positive tests in the first half of 2022, representing a positivity rate of 42.4% for the country as a whole. The Northeast (55.4%) and Midwest (37.3%) exhibited the highest regional positivity rates. In 2023, 152 (43.9%) positive cases were reported in all Brazilian regions, with the Southeast (47.4%) and Northeast (24.2%) demonstrating the highest positivity (Fig. 1B). Over the 2022–2023 period, women exhibited a slightly higher positivity rate (52.9%; n = 165). Children under the age of 12 accounted for 6.4% of cases, those aged 12 to 59 comprised 58.7%, and those over the age of 59 represented 34.9% (Fig. 1C).

Fig. 1

Positivity frequencies and social data for CHIKV diagnostic conducted in Brazil from 1 January 2022 to 30 June 2023. A Brazilian geographic regions. B CHIKV RT-qPCR positivity frequency following the period for each Brazilian geographical region. Each color line corresponds to the geographical region on the Brazil map, and the total frequency is presented as a black curve (Brazil). The North and South regions did not present a statistically significant number of cases that were well distributed throughout the study period. As a result, these regions are not included in the graph. C Age and sex distribution of CHIKV RT-qPCR positivity in Brazil over the 18-month study period. D Serological prevalence (IgM) for CHIKV in Brazil during 2022 and 2023. Each color line corresponds to the geographical region on the Brazil map, and the total frequency is presented as a black curve. E Age and sex distribution for IgM anti-CHIKV serological positive data in the whole period. The graphs were constructed using the ggplot2 package in R software

In serological testing, 65,332 anti-CHIKV IgM tests conducted at HP over the period resulted in 26,861 positives (41.1%) nationwide (Additional File 1: Table S1). In the first half of 2022, Brazil exhibited a positivity rate of 46.3% (n = 8182), with the Northeast (58.1%) and Midwest (37.9%) regions exhibiting the highest rates. In 2023, the Southeast (42.1%) and Northeast (33.1%) had the highest positivity rates, with an overall rate of 39.1% (n = 12,747) for Brazil (Fig. 1D). Throughout the 2022–2023 period, 65.7% of positive cases were women, 4.5% were children under 12, 64.1% were individuals aged 12–59, and 31.4% were those aged 60 or older (Fig. 1E).

Considering both testing methods, RT-qPCR and serology, the regional CHIKV positivity rates were as follows: 2022—Northeast (58.1%), Midwest (37.9%), Southeast (35.2%), North (30.9%), and South (7.7%); 2023—Southeast (40.5%), Northeast (30.7%), Midwest (26.4%), North (21.3%), and South (16.9%).

A total of 6481 individuals (16.7%; n = 38,770) from all over Brazil (Fig. 2A) tested positive for the NS1 antigen and RT-qPCR for dengue virus during the 2022–2023 period at HP laboratory (Additional File 1: Table S1). In the initial stages of the epidemic in early 2022, there were 4582 positive tests (26.4%). South (47.2%) and Midwest (29.4%) showed the highest regional positivity. In 2023, there were 1790 positive tests nationwide (10.0%), with the Southeast region accounting for 93.1% of these results. However, the highest regional positive rates were presented by North (30.2%) and South (19.1%) (Fig. 2B). Overall, 50.6% of those who tested positive were female. Children under the age of 12 represented 12.5% of positive tests, while those aged 12 and older constituted 72.5%. Elderly individuals (over the age of 59) accounted for 15.0% (Fig. 2C).

Fig. 2

Positivity frequencies and social data for DENV diagnostic conducted in Brazil from 1 January 2022 to 30 June 2023. A Brazilian geographic regions. B DENV RT-qPCR and NS1 positivity frequency following the period for each Brazilian geographical region. Each color line corresponds to the geographical region on the Brazil map, and the total frequency is presented as a black curve (Brazil). C Age and sex distribution of DENV RT-qPCR and NS1 positivity in Brazil over the 18-month study period. D Serological prevalence (IgM) for DENV in Brazil during 2022 and 2023. Each color line corresponds to the geographical region on the Brazil map, and the total frequency is presented as a black curve. E Age and sex distribution for IgM anti-DENV serological positive data in the whole period. The graphs were constructed using the ggplot2 package in R software

Anti-DENV IgM serological tests revealed 33,549 positive results (28.2%) out of 118,794 tests conducted (Additional File 1: Table S1). In the first half of 2022, there were 18,416 positive cases, while in 2023, there were 12,202. The Southeast consistently exhibited the highest regional positivity rates, with 39.5% in early 2022 and 26.5% in 2023 (Fig. 2D). Over the course of the whole period, 56.7% of positive tests were from women, 10.6% were from children under the age of 12, 69.3% were from individuals aged between 12 and 59, and 20.1% were from individuals aged 60 or above (n = 6755) (Fig. 2E).

Considering all testing methods (RT-qPCR, NS1 and IgM), the regional DENV positivity rates were as follows: 2022—South (40.8%), Southeast (34.8%), Midwest (28.0%), Northeast (23.3%), and North (22.8%); 2023—South (22.7%), Southeast (21.4%), Midwest (15.8%), North (14.4%), and Northeast (13.8%).

Arboviruses frequencies and mortality in MG state

According to data obtained from SES-MG, 73,307 confirmed cases of CHIKV and 333,964 cases of DENV were identified during the epidemiological weeks (EW) 1–51 of 2022 and EW 1–30 of 2023. However, the number of confirmed cases of arboviruses in 2023 was approximately 4.5 times higher than in 2022 for the same period (EW 1–30). Specifically, for CHIKV, 10,765 cases and two deaths were recorded in 2022, whereas in 2023, there were 62,542 cases and 33 deaths, representing a 5.8-fold increase in confirmed cases and a 16.5-fold increase for deaths (Fig. 3A). For dengue, 73,648 cases and 65 deaths were reported in 2022, while in 2023, there were 260,316 cases and 165 deaths, showing a 3.5-fold increase in confirmed cases and a 2.5-fold increase in deaths, including cases of severe dengue (Fig. 3B).

Fig. 3

Epidemiological scenario of arboviruses in Minas Gerais state. A, B Chikungunya and dengue number of cases and deaths during the years of 2022 and 2023 in MG, respectively. C Spatiotemporal cases distribution of CHIKV and D DENV in MG. The data was obtained by the Health Department of MG State (SES-MG). Cases of arboviruses were confirmed using serological (IgM/IgG detection) or molecular methods (RT-qPCR). In 2022, SES-MG reported 10,765 cases of CHIKV with only two deaths confirmed, while 62,542 cases and 33 deaths were confirmed in 2023. 73,648 DENV cases and 65 deaths were reported in 2022, while 260,316 cases and 165 deaths were confirmed in 2023

Besides CHIKV had a higher number of cases in 2023, the spread of the virus increased to the central area of the state. In Fig. 3C, during EW 01, the first cases were diagnosed in the city of Governador Valadares (n = 29), which is close to the state of Espírito Santo. In the following EWs, the virus dispersed to the North and Midwest region of the state.

In the case of DENV (Fig. 3D), a completely different scenario was observed. In 2022 and 2023, the first cases occurred in cities across different regions of the state and quickly spread throughout the territory. Although DENV distribution in MG was homogeneous, the Midwest and Triângulo Mineiro regions had a higher incidence of cases. These areas border states in the Brazilian Midwest region, which are endemic for different DENV serotypes.

Genomic surveillance of CHIKV and DENV

To elucidate the circulating genotypes/serotypes of CHIKV and DENV, 243 RT-qPCR confirmed DENV cases were genotyped/sequenced by Molecular ZC D-Tipagem Kit and whole genome sequencing. Both strategies yielded the same results. The majority of the DENV samples were classified as serotype 1 (n = 239; 94.5%), while four samples were classified as serotype 2 (1.6%) (see Additional File 1: Table S1). The four samples classified as DENV-2 came from the states of MG (one sample) and São Paulo (three samples) (Southeastregion). A total of 110 samples were classified as CHIKV according to Molecular ZC D-Tipagem Kit.

Moreover, we evaluated the possibility of coinfection with Arboviruses in the samples using Molecular ZC D-Tipagem Kit. Out of 243 DENV-confirmed investigated samples, 31 were also positive for CHIKV. Although this multiplex assay could also detect ZIKV, we did not find any positive sample for this arbovirus. Our sampling strategy results suggest a coinfection rate of 12.75% (DENV and CHIKV) (Additional File 1: Table S1). The analysis of coinfection cases revealed that the median Ct values were lower for DENV detection than for CHIKV, indicating a higher DENV viral load in patients with coinfections (Additional File 3: Fig. S1). The p-value (p < 0.001) indicated a statistically difference between the two groups, confirming a higher DENV viral load in the coinfected patients with CHIKV.

The whole genome of CHIKV and DENV-positive samples with Ct values lower than 35 were sequenced. We generated 151 DENV-1, two DENV-2, and 80 CHIKV new consensus genomes with coverage over 70%. The average genome coverage was 92% for DENV and 89.5% for CHIKV. The sequencing metrics for each sample are available at Additional File 1:Table S1.

Phylodynamics of CHIKV in MG state

The CHIKV genomes generated in the study clustered with the ECSA-American lineage with high statistical support (SH-aLRT = 100; Additional File 4: Fig. S2). We also constructed a dated phylogeny, using only genomes from ECSA-Americas clade and Brazilian sequences (n = 348 and 80 new genomes). The analysis corroborated that the ECSA clade was introduced into Brazil in 2014 in the Northeast region (Bahia state) (mean: 27 August, CI: 07 July to 28 August) (Fig. 4A, B). The generated phylogeny exhibited a strong temporal signal (R² = 0.62), indicating good reliability and confidence in the molecular clock analysis.

Fig. 4

Timetree phylogeographic reconstruction for the CHIKV ECSA clade in Brazil. A Timetree molecular clock phylogeny annotated according to the probable ancestral location. Each color represents a different Brazilian state, as described in the Brazilian map. Tip shapes indicate genomes generated in our study. Two mutations in non-structural and structural proteins are shown in the main MG clades. Clade 2 is represented by the presence of the non-structural mutation nsP4: P110S, while clade 3 is represented by the mutation in the polyprotein E1: I288T. In total, 1711 public genomes and 80 new genomes were used in our phylogeny. Number of genomes used in our phylogeny classified by state: Alagoas (AL—9), Amazonas (AM—5), Bahia (BA—27), Ceará (CE—64), Maranhão (MA—1), Mato Grosso do Sul (MS—2), Mato Grosso (MT—35), Minas Gerais (MG—86), Pará (PA—5), Paraíba (PR—2). Pernambuco (PE—7), Rio de Janeiro (RJ—91), Rio Grande do Norte (RN—11), Rio Grande do Sul (RS—1), Roraima (RR—11), São Paulo (SP—51), Sergipe (SE—2), Tocantins (TO—43). B Root-to-tip distance plot, showing a clock-like signal (r = 0.82)

The new genomes generated grouped into three major clades: clade 1, which included MG genomes (n = 13) grouping with a genome from the Northeast region (Aracaju -Sergipe state); clade 2, with MG genomes (n = 36) clustered with genomes from the Southern region (Novo Hamburgo and Rio Grande do Sul state); and clade 3, consisting only of MG samples (n = 14) (Fig. 4A). No new genomes generated in this study clustered with public sequences available in databases, suggesting CHIKV evolution in the recent outbreak. However, clades 2 and 3 share a common ancestor with sequences from the most recent genomic surveillance study conducted in the Brazilian Northeast region between the years 2018 and 2022 [35]. The MG clades resulted from an introduction from the Southeast region (São Paulo state) in mid-2019 (mean: 16 May, CI: 02 January 2019 to 02 May 2020). The time to the most recent common ancestor (tMRCA) for clade 1 was estimated to be in 2022 (mean: 10 February, CI: 10 January to 21 April 2022). Clade 2 and clade 3 were also dated to 2022 (mean: 23 August, CI: 25 May to 25 October 2022; and mean: 30 September, CI: 27 July to 23 November 2022, respectively) (Additional File 5).

Both clades 1 and 3 still have non-synonymous defining mutations. The mutations in clade 1 occurs in a structural protein (E1: I288T), while the clade 3 mutation occurs in a non-structural protein (nsP4: P110S). All three clades were found circulating in the metropolitan region of Belo Horizonte (state capital) (Additional File 6: Fig. S3). However, in the northern part of the state, only clades 1 (without non-synonymous mutations) and 3 (harboring I288T mutation) were observed, while in the area close to the South and the Triângulo Mineiro, exclusively clade 2 was identified (P110S mutation) (Additional File 6: Fig. S3). The genomes harboring these mutations were collected during February to April 2023.

Phylodynamics of DENV-1 in MG state

The maximum-likelihood phylogenetic analysis confirmed that all genomes sequenced in our study belong to the DENV-1 genotype V with high statistical support (SH-aLRT = 100; Additional File 6: Fig. S3). This result corroborates with ZC-D serotyping diagnostic for DENV, which classified the sequenced samples as serotype 1. All new DENV-1 genomes clustered with different strains of genotype V with high statistical support (SH-aLRT = 100; Additional File 7: Fig. S4 and Additional File 8).

The dated phylogeny of the DENV-1 genomes indicated that genotype V was introduced in the country in 1984 in the Southeast region (Rio de Janeiro state—RJ) (mean: 16 March, CI: 29 July 1983 to 04 November 1984) (Fig. 5A). The phylogenetic analysis showed a strong correlation between substitution rates and time (R² = 0.94) (Fig. 5B).

Fig. 5

Timetree phylogeographic reconstruction for the DENV-1 genotype V with Brazilian genomes. A Timetree molecular clock phylogeny annotated according to the probable ancestral location. Each color represents a different Brazilian state, as described in the Brazilian map. Tip shapes indicate genomes generated in our study. The clades are showed in the figure. In total, 1328 publicly available genomes and 151 new genomes were used to construct the phylogeny. Number of genomes used in our phylogeny classified by state: Alagoas (AL—3), Amazonas (AM—10), Amapá (AP—12), Bahia (BA—19), Distrito Federal (DF—16), Espírito Santo (ES—2), Goiás (GO—36), Minas Gerais (MG—161), Mato Grosso (MT—47), Mato Grosso do Sul (MS—32), Pará (PA—2), Paraiba (PB—39), Pernambuco (PE—26), Rio de Janeiro (RJ—27), Rondônia (RO—11). Santa Catarina (SC—7), Sergipe (SE—1), São Paulo (SP—316). B Root-to-tip distance plot, showing a clock-like signal (r = 0.94)

Several clades with MG samples were observed in the phylogeny. The three main clades are as follows: clade 1, containing most of the new genomes, which groups with sequences from all Brazilian geographic regions with a probable origin in 2018 (mean: 14 April, CI: 23 October 2017 to 19 March 2019); clade 2, which includes MG genomes grouping with genomes from the South, Southeast, Midwest, and North regions, with a probable origin in 2019 (mean: 15 January, CI: 11 August 2018 to 28 April 2019); and clade 3, which includes MG genomes grouping with genomes from the South and Southeast regions, originating in 2017 (mean: 18 December, CI: 31 July to 22 May 2018). These results demonstrate the circulation of several lineages of genotype V in MG, with clade 1 being the main determinant of the 2023 epidemic.

Discussion

In this study, we carried out genomic and epidemiological surveillance of two arboviruses (CHIKV and DENV) in Brazil and specifically in the state of MG. The study was performed in the years of 2022 and 2023 according to the epidemiological data of test results evaluated by the HP related to the whole Brazilian territory and Health Department of the MG state. More than 630,000 diagnostic tests (molecular and serological) were performed in the period. According to our results, we observed a particular increase in positivity for CHIKV in 2023 compared to 2022. For the MG state, it was observed an increase for CHIKV and DENV. The Northeast had the highest positivity rate for CHIKV in 2022 (55.4% molecular; 58.1% serological), whereas in 2023, the Southeast showed the highest positivity rate (47.4% molecular; 42.1% serological). Molecular and antigen tests of DENV showed in the same period of both years the highest regional positivity in the South region in 2022 (47.2%) and in the North region in 2023 (30.2%). Southeast presented the highest regional positivity for serological DENV tests in both years (39.5% in 2022; 26.5% in 2023). Comparatively, the epidemiological bulletins of the Brazilian Ministry of Health indicated that the Northeast and Southeast regions had high positivity rates for both viruses. For example, only in the nine first EWs of 2023, an increase of 109.6% in CHIKV and 47.7% in DENV cases was reported in the Brazilian territory. The states of MG (Southeast region) and Bahia (BA–Northeast region) had the highest positivity coefficients in their respective geographical region (DENV: MG—313.9; BA—52.7; CHIKV: MG—107.2; BA—20.5) [3, 36]. MG and BA states border each other, suggesting a circulation of arboviruses between their territories.

In MG, an increase of 4.5 times in the number of cases of these arboviruses was observed. The 20 first EWs of 2023 present a greater number of cases, period assigned as summer in Brazil, in which temperatures are generally high and with higher rain precipitation rates [37]. High temperatures favor the growth and development of the vector, as described previously [38, 39]. One of the metrics used by Brazilian health agencies is to carry out an entomological survey in cities to identify the presence of the vector A. aegypti and A. albopictus in the urban region. The first report carried out by the health department of MG (February 2023) indicated out that nearly 40% of the municipalities are at risk of arboviruses transmission, and 41% are on alert due to the presence of the vector in different regions [40]. The presence of the vector in different areas may increase the circulation of arboviruses. During our study period, we observed an increase in the number of cases and deaths, which could be related to the high transmissibility and evolution of the virus. The state of Bahia was one of the first locations where CHIKV was introduced to the national territory [9]. The proximity of borders between the two states may facilitate the importation of CHIKV cases into MG. In 2023, the first CHIKV cases were diagnosed in the cities of Januária (n = 979) and Jaíba (n = 276), in the Northern region of MG, close to the state of Bahia (Northeastern of Brazil), and in the following EWs, the incidence of cases was concentrated in this region [41]. In the state of Bahia occurred one of the first introductory events of CHIKV in the national territory [9]. The proximity of borders between the two states may allow the importation of CHIKV cases into MG. The data evaluated in our study demonstrate that the incidence of CHIKV cases in the Northern region of MG is higher than in other regions and may be due to its proximity to the state of Bahia. Moreover, the high positivity rate for CHIKV may also be attributed to the increase in the number of molecular diagnoses and the effectiveness of tests used to detect CHIKV in the state. The number of CHIKV cases in other regions of the country may be higher. However, many diagnoses are reported as dengue probably due to the similarity of clinical symptoms and the insufficient capacity of the laboratory to diagnose all suspected individuals. MG state has a surveillance structure that allows molecular tests to be carried out systematically for differentiated diagnosis of arboviruses, allowing cases to be classified correctly [42].

We performed the whole genome sequencing of 80 CHIKV new genomes and phylogenetic analysis. The analysis suggests that CHIKV ECSA was first introduced in MG in the middle of 2015 through an event of importation from the Northeast region. This resulted in an increase in the incidence of cases in the state in March 2016 [43]. Events of importation of CHIKV in MG and other states have already been described in other studies since the first description of the virus in the country, in September 2014. In this year, different events of importation and local transmission of two different genotypes were identified, CHIKV-Asian in the North region and ECSA in the Northeast region of Brazil [9]. All sequenced genomes were classified as ECSA genotype. The dated phylogenetic analysis showed the presence of three main CHIKV clusters. In both groups, it is possible to corroborate that CHIKV is related to circulation events between MG and the Northeast region of the country. One of the clusters grouped with sequences from the Northeast region, in which there was a description of cases of virus recurrence in regions with few numbers of cases in previous epidemics [35], suggesting again that CHIKV is not a virus with established transmission in some Brazilian states; however, since CHIKV ECSA-American lineage was introduced in Northeastern region of Brazil (Bahia state) [9], different groups were identified circulating in different regions in the following years, with dispersion to the North, Southeast, and Midwest regions. The new CHIKV genomes from 2023 epidemics grouped in three clusters based on non-synonymous mutations in the nsP4: P110S (clade 3) or E1: I288T (clade 1) genes. These mutations differ from previous genomic surveillance analyses carried out in the state of Rio de Janeiro (Southeastern Brazil) in the year 2022, in which the NS1: D513G and nsP4: A481D mutations were described [44]. Moreover, none of the new genomes assembled showed mutations in the E1:A226V and E2:L210Q genes, associated with higher infectivity of the virus in the vector and transmission to humans [44]. The mutations found in CHIKV genomes sequenced in our study, although not yet identified previously, may indicate an adaptive evolution event and impact the fitness and transmissibility of the virus. Nevertheless, the grouping of these two clusters that present mutations may be due to the limited number of publicly available CHIKV genomes. Clade 2 has the same common ancestor with clade 3 but did not present any non-synonymous mutations.

The geographical distribution of three clades in MG was analyzed. Clade 2 was found in both the State Capital and the North of the state. This suggests that clade 2 may be the primary cause of the increase in the number of cases in the North of MG. Moreover, clade 2 was also identified in the capital city, co-circulating with the other two identified clades (clades 1 and 3). These results suggest that MG is a hotspot for CHIKV transmission in the country and clades 1 and 3 co circulation might be responsible for the increasing cases in the MG state. The increase in the number of cases may be related to events of introduction in the Northeast region. However, it is worth noting that as of 2023, no Brazilian CHIKV genome has been made available in public databases. This limitation may have affected the analysis of mutations and import data.

In contrast to CHIKV, cases of DENV were not concentrated in just one region of the state. DENV cases were reported in all MG regional health units (RHUs during both years (2022–2023)). This difference between the concentration of CHIKV in the North region of the state (near the Northeast region of the country) and the distribution of DENV throughout the territory is an indication that the dengue virus already has a homogeneous and continuous transmission in the state, which does not occur with CHIKV, which depends on import cases from other states, for example, the state of Bahia. In our study, two serotypes of dengue were found: DENV-1 and DENV-2. All genomes of serotype DENV-1 were classified as genotype V. Phylogenetic analysis demonstrated that the generated genomes are distributed in different clades, which corroborates the hypothesis that the dengue virus has already established transmission in MG. DENV-1 genotype V is the most prevalent in circulation in the Americas, and it was introduced in Brazil in 1990 in the Northeast region. Since then, the circulation of different lineages has been identified [5, 45, 46]. Our phylogeny confirms that the V genotype was introduced in the mid-1990s, and different lineage circulation clusters have been present over the years. Moreover, the main cluster in our phylogeny suggests that MG played an important role in exporting events to other states in the Southeast, Northeast, South, and Midwest regions of the country.

MG is a state that borders two other geographical regions of Brazil (Northeast and Midwest) and has one of the main road networks in the country. Additionally, the climate is favorable for the growth and dispersion of the vector Aedes spp. in the state, increasing its transmission in different regions. In addition to the samples genotyped as DENV-1, we identified four DENV-2 samples from the southeastern region of the country. Of these four samples, only two were sequenced and classified as cosmopolitan genotype (DENV-2-GII). DENV-2 can be classified into five genotypes. The DENV-2-GII genotype is one of the most widespread strains in the world, being described in Asian, African, and American continents [47]. The introduction of DENV-2-GII in the country occurred in 2020 in the Midwest and North regions [48], with the first cases being described in 2021 in those regions [49, 50]. Studies have shown that this genotype is currently in circulation in all geographical regions of Brazil, despite being recent. In 2022, the number of positive cases of dengue was higher, suggesting a rapid dispersion of the genotype throughout the national territory. Therefore, it is important to increase DENV-2 surveillance studies in the country to monitor the dynamics of virus dispersion and evolution and estimate the potential impact of its circulation on the number of cases and disease severity [48].

Our data also indicates that more individuals tested for DENV in 2023 compared to 2022. However, there was no increase in positive cases. This suggests that some individuals who tested negative for DENV may have been infected with CHIKV, as both viruses have similar symptoms, and it is possible that the individuals who were tested were those presenting symptoms.

Given this scenario, we thought that perhaps there is a co-infection phenomenon in Brazilian epidemics that is not being identified by labs and health authorities. This is because most tests performed target only one virus. We investigated it in our sampling, and we found a co-infection rate of 12.75%. Arbovirus co-infection events can occur because the vectors that transmit the viruses can be infected with and transmit multiple arboviruses simultaneously [51]. Co-infection cases have already been described in Brazil and other countries [52, 53]. During infection, four types of responses can be triggered, including enhancement or inhibition of both viruses, competition, and no effect on viral load [54]. Our analysis revealed a difference in the Cts values of the co-infected samples. It was observed that DENV had a higher viral load compared to CHIKV, indicating that one of the hypotheses mentioned earlier may have occurred. Additionally, the multiplex kit used to perform co-infection tests in our study employs different probes and regions for DENV and CHIKV, and the detection of these two viruses is separated in different PCR reactions, ruling out the hypothesis that a best PCR performance to detect DENV could explain the higher viral load of this virus in the co-infections analysis. This eliminates the possibility of primer competition and differing efficiency between the two PCR reactions, suggesting that maybe there is an interaction of both viruses in human host that should be better explored. Indeed, the clinical outcome and severity should be monitored by the health agencies in those patients presenting co-infections cases.

Our study has some limitations. Although we analyzed data from serological and molecular tests in the country, it was not possible to evaluate all states due to the lack of representative tests derived from HP laboratory. The epidemiological data derived from SES-MG, despite being evaluated in this study, the samples were not sequenced, and possible correlations between clinical status (morbidity and mortality) and lineage circulation could not be carried out. Furthermore, although we identified in phylogenetic analysis clades with nonsynonymous mutations for CHIKV, we did not explore the evolution of these lineages due to the lack of clinical information from patients and prediction of protein structure. However, despite the limitations described above, the results indicate that the incidence and mortality of both viruses increased between the years in MG, with a high incidence of cases in regions where arboviruses were reported in high frequency in previous epidemics in Brazil [41, 43].

Conclusions

The findings of this study offer valuable insights into two arboviruses in Brazil, particularly in the state of Minas Gerais, which had one of the highest case rates in 2023. In total, we generated 80 and 151 new genomes of CHIKV and DENV-1, respectively. CHIKV genomes were classified as belonging to the ECSA-American lineage. Three main clades were identified in the state of MG, with two clades presenting mutations in structural and non-structural proteins. The majority of genomes identified circulating in MG were classified as DENV-1 genotype V, with three main clades observed. We estimate a co-infection rate of nearly 13% in the studied sample. The results are beneficial for public health agencies in their efforts to reduce transmission and the number of cases as well as for future research on the epidemiology and distribution of CHIKV and DENV arboviruses in the country.