Co-Circulation of Chikungunya and Multiple DENV Serotypes and Genotypes, Western Indonesia 2015–2016

Dengue is a mosquito-borne disease of public health concern affecting tropical and subtropical countries, including Indonesia. Although studies on dengue epidemiology have been undertaken in Indonesia, data are lacking in many areas of the country. The aim of this study was to determine dengue virus (DENV) and chikungunya virus (CHIKV) molecular epidemiology in western regions of the Indonesian archipelago. A one-year prospective study was conducted in Aceh and Jambi in 2015 and 2016, respectively, where patients with dengue-like illness were enrolled. Of 205 patients recruited, 29 and 27 were confirmed with dengue in Aceh and Jambi, respectively, and three from Jambi were confirmed with chikungunya. DENV-1 was the predominant serotype identified in Aceh while DENV-2 was predominant in Jambi. All DENV-1 and DENV-2 from both regions were classified as Genotype I and Cosmopolitan genotype, respectively, and all DENV-3 viruses from Jambi were Genotype I. Some viruses, in particular DENV-1, displayed a distinct lineage distribution, where two DENV-1 lineages from Aceh were more closely related to viruses from China instead of Jambi highlighting the role of travel and flight patterns on DENV transmission in the region. DENV-2 from both Aceh and Jambi and DENV-3 from Jambi were all closely related to Indonesian local strains. All three CHIKV belonged to Asian genotype and clustered closely with Indonesian CHIKV strains including those previously circulating in Jambi in 2015, confirming continuous and sustainable transmission of CHIKV in the region. The study results emphasize the importance of continuous epidemiological surveillance of arboviruses in Indonesia and simultaneous testing for CHIKV among dengue-suspected patients.


Introduction
Dengue, caused by infection with any of four dengue virus (DENV-1 to DENV-4) serotypes, is a major international public health concern and is associated with significant morbidity, mortality, and economic losses [1][2][3]. Nearly 60% of the Indonesian population (277 million) live in areas where DENV activity is known. Currently, Indonesia reports the highest average number of dengue cases in Asia [4,5]. The incidence rate (IR) of dengue hemorrhagic fever (DHF) has had a cyclical pattern, with peaks occurring approximately conditions: (a) supportive serology on a single serum sample and (b) occurrence at the same location and time as confirmed cases of DF [32].
Demographic data, clinical manifestations and laboratory information from all probable dengue patients were collected. On the first visit day, approximately 3-5 mL of venous blood was collected, and serum samples were sent to Eijkman Institute for Molecular Biology in Jakarta and stored at −80 • C before use.

NS1 EIA and DENV-Serotyping
Serum samples were initially tested for the presence of DENV NSI antigen with Platelia Dengue NS1 Antigen EIA kit (Bio-Rad, Marnes-la-Coquette, France). To attempt DENV isolation, NS1-positive sera were inoculated and cultured on to Vero cells (African Green Monkey kidney epithelium, ATCC CCL-81) maintained in Dulbecco's Modified Eagle Media (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 5% heatinactivated fetal bovine serum, 1% L-glutamine and 1% penicillin/streptomycin (Life Technologies, Carlsbad, CA, USA). Cultures were incubated at 37 • C with 5% CO 2 for 7 days and blind passaged for two rounds to increase the isolation success rate. NS1-negative samples were also inoculated onto Vero cells to attempt isolation of other viruses. All NS1-positive samples were subjected to DENV serotyping PCR using an in-house designed RT-PCR targeting the NS5 gene region (PathWest Laboratory, Nedlands, WA, Australia).

DENV E Genes Amplification and DENV E Gene Sequencing
DENV RNA was extracted from culture supernatant using the spin protocol of the QIAamp Viral RNA Mini kit (Qiagen, Hilden, Germany). Complementary DNA was synthesized using the SuperScript III First-Strand Synthesis System for RT-PCR following the manufacturers' protocol (Invitrogen-Life Technologies, Carlsbad, CA, USA). Two pairs of specific primers for each serotype were used concurrently to amplify the E gene, the primer sequences for which have been published elsewhere [33]. The amplifications were run using a PCR master mix GeneAmp 10× PCR buffer II and Taq DNA polymerase Gene sequencing was conducted using the Sanger sequencing approach [34]. Prior to sequencing, the excess oligonucleotide primers and dNTPs from PCR product were removed using Illustra ExoStar (GE Healthcare Life Sciences, Buckinghamshire, UK). The BigDye Terminator v3.1 Cycle Sequencing kit and BigDye Terminator 5× Sequencing Buffer (both from Applied Biosystem, Foster City, CA, USA) were used for the DNA sequencing reaction. The samples were re-cleaned using the DyeEx 2.0 Spin kit (Qiagen GmbH, Hilden, Germany) and a capillary electrophoresis was performed on a 16-capilary genetic analyzer (3130xl Genetic Analyzer, Applied Biosystems, Foster City, CA, USA).

CHIKV Whole Genome Sequencing
Whole CHIKV genomes were determined by Illumina sequencing. Vero cell virus culture supernatant was harvested and then concentrated using Millipore centrifugal filter unit (100,000 kDa) for 20 min at 40,000× g (Merck Millipore Ltd., County Cork, Ireland). Total RNA was extracted from 150-200 µL of concentrated virus using the Roche High Pure RNA Isolation kit as per manufacturers' instructions (Roche Diagnostic GmbH, Mannheim, Germany). DNA libraries were prepared with the Illumina TruSeq Stranded mRNA kit (Illumina Inc., San Diego, CA, USA). The DNA library was validated with a High Sensitivity DNA kit on the Agilent 2100 Bioanalyzer (both from Agilent Technologies, Waldbronn, Germany). The concentration of the libraries was normalized to 4 nM using HT1 buffer and pooled together. The final concentration of the pooled DNA library was then re-examined using the Qubit dsDNA Assay kit (Invitrogen-Life Technologies, Carlsbad, CA, USA). The pooled DNA library was sequenced using a MiSeq Reagent Micro Kit, v2 (Illumina Inc., San Diego, CA, USA). The high throughput sequence reads from WGS were paired and used to construct contiguous sequences, de novo, using CLC Genomics Workbench 10.0 (Qiagen, Aarhus, Denmark) after quality assessment within FastQC Version 0.11 (Babraham Bioinformatics, Cambridge, UK).

Phylogenetic Analysis
For each virus, derived sequences (DENV E gene and CHIKV complete genomes) together with available sequences from GenBank were aligned separately using Multiple Alignment using Fast Fourier Transform (MAFFT) v.7.309 [35] as implemented in Geneious v.10.1.3 [36]. To provide the relationship between Indonesian viruses and viruses isolated from other parts of world, an estimation of the maximum likelihood-based phylogenetic tree was performed for each virus using RAxML v.7.2.8 [37,38]. GTR with gamma substitution model (GTR+Γ) and a rapid bootstrap procedure (100 replicates) were employed. Resultant phylogenies were visualized using FigTree v1.4.3 and bootstrap of at least 80% was used as an indicator for strong support for grouping of viruses into distinct lineages.

Patient Characteristics
In total, 205 dengue-suspected patients were recruited (63 from Aceh and 142 from Jambi), of which 56 were confirmed as dengue based on NS1 EIA (29/63 from Aceh and 27/142 from Jambi). Among the 56 confirmed dengue cases, the median age was 21.3 years with the majority aged between 21-30 years (33.9%), followed by the 11-20 years age group (26.8%) ( Table 1). There was an equal number of reported male and female cases. Complete clinical histories were available for 43 patients from both regions. The most common reported clinical symptoms were fever, nausea or vomiting, headache, and aches and pains (myalgia or arthralgia). Dengue with warning signs based on the WHO criteria [39] was observed in 12 (27.9%) patients in whom abdominal pain or tenderness was the most frequent sign (20.9%) followed by mucosal bleeding (9.3%). Of those with a warning sign, the vast majority (11,91.6%) were from Aceh, most from the Zainoel Abidin Hospital (8, 66.6%). This is unsurprising, given the Zainoel Abidin Hospital is a provincial referral hospital where most of the severe cases were referred. No severe dengue was identified in this study.

Distribution of DENV Genotypes and Lineages in Aceh and Jambi
A total of eight DENV sequences from Aceh were successfully sequenced of which, five were typed as DENV-1 and three as DENV-2 ( Table 2). All characterized DENV-1 viruses were genotyped as Genotype I and all DENV-2 viruses belonged to the Cosmopolitan genotype. In Jambi, 14 E genes were sequenced (2 DENV-1, 8 DENV-2 and 4 DENV-3). Similar to what was observed in Aceh, all DENV-1 and DENV-2 viruses grouped within the Genotype I and Cosmopolitan genotypes, respectively. Phylogenetic analysis revealed that all DENV-3 belonged to Genotype I ( Table 2). All DENV-1 isolated from Aceh and Jambi belong to DENV-1, Genotype I ( Figure 2). Viruses from Aceh fell into four phylogenetically distinct lineages (Lineage 1-4) and clustered with sequences from the Southeast Asia region. These lineages could be further differentiated into two major groups based on the origin of the viruses: Indonesian local lineages (Lineage 2 and 4) and introduced lineages (Lineage 1 and 3) (

DENV-2
The three DENV-2 viruses isolated from Aceh and eight from Jambi all belonged to the Cosmopolitan genotype and were further differentiated into two phylogenetically distinct lineages (Lineage 1 and 2) (Figure 3). Lineage 1 consisted of viruses exclusively from Jambi while Lineage 2 consisted of viruses from both regions.  Viruses within Lineage 1 (D2/IDN_Jambi_025/2016 and D2/IDN_Jambi_032/2016) clustered together with viruses isolated from other Indonesian regions Bali, Jakarta, Sukabumi, and from other countries including Papua New Guinea (PNG) (Figure 3). Interestingly, while these viruses were positioned in close relation to the most predominant DENV-2, isolated during the 2014-2015 dengue outbreak in Jambi (Figure 3), the bootstrap support between these two groups of viruses was 83% (well above the threshold of 80%) and therefore could classified into two different lineages.
Lineage 2 included all three DENV-2 sampled from Aceh and six viruses from Jambi. The lineage also included viruses isolated during the 2014-2015 dengue outbreaks in Jambi as well as other viruses from Sumatra and Java such as Purwokerto. This suggests that viruses within this lineage are distributed from the westernmost part of Sumatra Island to Java Island. Interestingly, viruses of this lineage were also isolated from neighboring countries including Singapore and Malaysia.

DENV-3
Four DENV-3 viruses were isolated from Jambi and identified as Genotype I ( Figure 4). All these viruses grouped into one lineage together with Indonesian strains that were isolated from various locations including Bali, Surabaya, Purwokerto, and Jakarta. This lineage also included the only virus that was isolated (KU529754) during the 2015 dengue outbreak in Jambi [23]. This suggests there was little change in circulating DENV-3 in Jambi between 2015 and 2016. This lineage also contained viruses isolated in Malaysia and Singapore. Four DENV-3 viruses were isolated from Jambi and identified as Genotype I ( Figure  4). All these viruses grouped into one lineage together with Indonesian strains that were isolated from various locations including Bali, Surabaya, Purwokerto, and Jakarta. This lineage also included the only virus that was isolated (KU529754) during the 2015 dengue outbreak in Jambi [23]. This suggests there was little change in circulating DENV-3 in Jambi between 2015 and 2016. This lineage also contained viruses isolated in Malaysia and Singapore.

Currently Circulating CHIKV in Aceh and Jambi
Of 149 NS1-negative sera (34 from Aceh and 115 from Jambi), three were positive for CHIKV as confirmed by RT-PCR (Table 3). Phylogenetic analysis based on full or partial genome sequences indicated these viruses were closely related to viruses that circulated in Jambi in the previous year (2015) ( Figure 5). The viruses were also closely related to an

Currently Circulating CHIKV in Aceh and Jambi
Of 149 NS1-negative sera (34 from Aceh and 115 from Jambi), three were positive for CHIKV as confirmed by RT-PCR (Table 3). Phylogenetic analysis based on full or partial genome sequences indicated these viruses were closely related to viruses that circulated in Jambi in the previous year (2015) ( Figure 5). The viruses were also closely related to an Indonesian virus imported to Japan in 2015 (LC259091).

Discussion
We characterized DENV and CHIKV circulating in Aceh and Jambi, two provinces in western Indonesia, in 2015-2016. This study was the first virological investigation of dengue and chikungunya in Aceh, the western-most region of Indonesia. In Jambi, where an assessment of DENV and CHIKV molecular epidemiology had been conducted in 2014-2015 [23,31], this present follow-up study provides updated and more detailed information on transmission dynamics of both DENV and CHIKV in the region.

Distinct DENV Lineages Circulate in Aceh and Jambi
The molecular epidemiology of DENV in Aceh (2015) and Jambi (2016) was investigated. Although both regions are within Sumatra Island, distinct DENV lineage distributions were observed. Of four DENV-1 Genotype I lineages characterized in Aceh, two (Lineage 1 and Lineage 3) were related to viruses from China and only one (Lineage 2) was closely related and shared a common ancestor with viruses from Jambi ( Figure 2). However, the short length E gene sequence fragments from some DENV Lineage 1 and 3 viruses (227 bp), may not have conferred robust phylogenetic reconstruction. Nevertheless, there are some reasons to support the close relationship of viruses between Aceh and China. Aceh Province is located next to North Sumatra Province where Chinese is the fourth most populous ethnicity, most living in Medan, the capital city North Sumatra, with a high population mobility. There are frequent flights between Medan and cities in China via Kuala Lumpur (Malaysia) and Singapore ( Figure 6). In addition, there are direct flights between Aceh's capital, Banda Aceh, and Penang in Malaysia. This transport pattern supports virus movement between Aceh in Indonesia, and China. However, no E gene sequence are available from North Sumatra; the only available sequences are of four NS1 gene sequences (2 DENV-2, 1 DENV-3 and 1 DENV-4) [40]. A brief analysis using Dengue Genographic Viewer [41] indicated, that except for DENV-3, all viruses isolated from North Sumatra Province are closely related with viruses isolated in China.
In contrast, all DENV-1 isolated from Jambi in 2016, grouped within Lineage 5, were closely related to viruses that circulated in Jambi in prior years (2014-2015) ( Figure 2). All viruses within Lineage 5 were associated with large-scale dengue outbreaks in Jambi in 2014-2015 in which dengue shock syndrome occurred in 2% of 107 confirmed dengue patients [23]. Interestingly, no DENV-1 isolated from Aceh belong to this lineage indicating distinct DENV-1 viruses circulated in Aceh in 2015, which were distinct to viruses circulating in Jambi during 2014-2015. There are no direct flights between Aceh and Jambi, and flights between Medan and Kuala Lumpur in neighboring Malaysia are far more frequent than between Medan and Jambi, within Indonesia. This pattern of population movements supports our data regarding distinct DENV circulation patterns between geographically distinct, yet proximal, areas in Indonesia.
In contrast to the observed DENV-1 epidemiology, DENV-2 and DENV-3 are more homogeneous between Aceh and Jambi. All DENV-2 from Aceh and Jambi grouped into the Cosmopolitan genotype, the most common genotype in Indonesia and neighboring countries. All DENV-2 viruses clustered with locally sampled Indonesian viruses. The same pattern was observed for DENV-3: all isolated viruses from Jambi belonged to Genotype I and clustered with Indonesian viruses including those that circulated during the 2014-2015 dengue outbreaks in Jambi. In contrast to the observed DENV-1 epidemiology, DENV-2 and DENV-3 are more homogeneous between Aceh and Jambi. All DENV-2 from Aceh and Jambi grouped into the Cosmopolitan genotype, the most common genotype in Indonesia and neighboring countries. All DENV-2 viruses clustered with locally sampled Indonesian viruses. The same pattern was observed for DENV-3: all isolated viruses from Jambi belonged to Genotype I and clustered with Indonesian viruses including those that circulated during the 2014-2015 dengue outbreaks in Jambi.

Continuous Co-Circulation of CHIKV and DENV in Jambi
Although multiple chikungunya outbreaks have been reported in Indonesia [25], CHIKV molecular epidemiology studies on Sumatra Island are limited. In the present study based in Aceh (2015) and Jambi (2016), CHIKV was identified in Jambi but not in Aceh. Phylogenetic analysis of CHIKV whole genomes showed that all three Jambi viruses were of the Asian genotype and clustered with Indonesian CHIKV viruses sampled over a wide geographical range including as Sumatra, Java, and Bali, including viruses that circulated in Jambi in 2015 ( Figure 5).
This study confirmed continuous co-circulation of DENV and CHIKV in Jambi as reported previously [31]. Together, these data provide insight that DENV and CHIKV cocirculation likely occurs in other areas of Indonesia, a pattern that has been reported by others [29,42]. Continuous and sustainable transmission of CHIKV in Jambi could add the burden of dengue management in particular owing to the absence of confirmatory laboratory testing for CHIKV in most hospitals in Indonesia. Our findings emphasize the importance of laboratory testing for CHIKV among dengue-suspected patients.

Continuous Co-Circulation of CHIKV and DENV in Jambi
Although multiple chikungunya outbreaks have been reported in Indonesia [25], CHIKV molecular epidemiology studies on Sumatra Island are limited. In the present study based in Aceh (2015) and Jambi (2016), CHIKV was identified in Jambi but not in Aceh. Phylogenetic analysis of CHIKV whole genomes showed that all three Jambi viruses were of the Asian genotype and clustered with Indonesian CHIKV viruses sampled over a wide geographical range including as Sumatra, Java, and Bali, including viruses that circulated in Jambi in 2015 ( Figure 5).
This study confirmed continuous co-circulation of DENV and CHIKV in Jambi as reported previously [31]. Together, these data provide insight that DENV and CHIKV co-circulation likely occurs in other areas of Indonesia, a pattern that has been reported by others [29,42]. Continuous and sustainable transmission of CHIKV in Jambi could add the burden of dengue management in particular owing to the absence of confirmatory laboratory testing for CHIKV in most hospitals in Indonesia. Our findings emphasize the importance of laboratory testing for CHIKV among dengue-suspected patients.
The lack of confirmatory laboratory testing for CHIKV and the similarity of dengue and chikungunya clinical presentations may contribute to under-estimation of chikungunya incidence, disease and mortality. No chikungunya-associated deaths have been recorded in Indonesia since the disease was officially recognized by Indonesian Ministry of Health (MoH) in 1973 [43].
Since 2004, chikungunya has been included in the MOH National Diseases Surveillance Program, and since 2005, reported in the Annual Report of the Indonesia Health Profile, the formal annual report of notifiable diseases from MoH Indonesia [43]. In the National Guideline of Prevention and Control of Chikungunya from MoH [44], there are no clear criteria for assessing or reporting chikungunya-related deaths that should be included in the surveillance system. In addition, as chikungunya shares some clinical signs with dengue and not all reported dengue cases in Indonesia are confirmed by laboratory testing, there is a possibility that chikungunya-related deaths may be wrongly attributed to dengue as a consequence of the prioritization of dengue surveillance. Chikungunya is not considered a priority disease in Indonesia, there is poor surveillance and the disease is underreported. For example, chikungunya was excluded from the 2016 MoH Annual Report of the Indonesia Health Profile [45] and was included in the 2017 Annual Report [46]. Our study confirms that CHIKV circulated in Indonesia during 2016.
It would be beneficial if MoH would institute a clear reporting system for chikungunya and provide clear criteria for chikungunya-associated deaths in the national guidelines. In addition, during disease outbreaks active surveillance may be required to reduce underreporting of chikungunya deaths in the country.

Conclusions
Our phylogenetic analysis showed that all DENV-1 and DENV-2 viruses from Aceh and Jambi belonged to the Genotype I and Cosmopolitan genotypes, respectively and further, that all DENV-3 viruses from Jambi were of the Genotype I. Some viruses, particularly DENV-1, displayed distinct lineage distribution, where two distinct DENV-1 lineages from Aceh contained viruses more closely related to viruses from China, than to Jambi isolates. In contrast, DENV-2 from both regions and DENV-3 from Jambi were all closely related to local Indonesian viruses. Phylogenetic analyses of CHIKV genome sequences from Jambi showed that all viruses belonged to the Asian genotype. These viruses clustered closely with Indonesian CHIKV viruses including those previously circulating in Jambi in 2015-confirming continuous and sustainable transmission of CHIKV in the region. These data emphasize the importance of simultaneous testing for CHIKV among dengue-suspect patients.