Nationwide Molecular Epidemiology of Measles Virus in Japan Between 2008 and 2017

Genotyping evidence that supports the interruption of endemic measles virus (MV) transmission is one of the essential criteria to be verified in achieving measles elimination. In Japan since 2014, MV genotype analyses have been performed for most of the measles cases in prefectural public health institutes nationwide. With this strong molecular epidemiological data, Japan was verified to have eliminated measles in March, 2015. However, even in the postelimination era, sporadic cases and small outbreaks of measles have been detected repeatedly in Japan. This study investigated the nationwide molecular epidemiology of MV between 2008 and 2017. The 891 strains in the total period between 2008 and 2017 belonged to seven genotypes (D5, D4, D9, H1, G3, B3, and D8) and 124 different MV sequence variants, based on the 450-nucleotide sequence region of the N gene (N450). The 311 MV strains in the postelimination era between 2015 and 2017 were classified into 1, 7, 8, and 32 different N450 sequence variants in D9, H1, B3, and D8 genotypes, respectively. Analysis of the detection period of the individual N450 sequence variants showed that the majority of MV strains were detected only for a short period. However, MV strains, MVs/Osaka.JPN/29.15/ [D8] and MVi/Hulu Langat.MYS/26.11/ [D8], which are named strains designated by World Health Organization (WHO), have been detected in many cases over 2 or 3 years between 2015 and 2017. The WHO-named strains have circulated worldwide, causing outbreaks in many countries. Epidemiological investigation revealed repeated importation of these WHO-named strains into Japan. To demonstrate the elimination status (interruption of endemic transmission) in situations with repeated importation of the same strains is challenging. Nevertheless, the detailed sequence analysis of individual MV strains and chronological analysis of these strains provided sufficient evidence to show that Japan has still maintained its measles elimination status in 2017.

Genotyping evidence that supports the interruption of endemic measles virus (MV) transmission is one of the essential criteria to be verified in achieving measles elimination. In Japan since 2014, MV genotype analyses have been performed for most of the measles cases in prefectural public health institutes nationwide. With this strong molecular epidemiological data, Japan was verified to have eliminated measles in March, 2015. However, even in the postelimination era, sporadic cases and small outbreaks of measles have been detected repeatedly in Japan. This study investigated the nationwide molecular epidemiology of MV between 2008 and 2017. The 891 strains in the total period between 2008 and 2017 belonged to seven genotypes (D5, D4, D9, H1, G3, B3, and D8) and 124 different MV sequence variants, based on the 450-nucleotide sequence region of the N gene (N450). The 311 MV strains in the postelimination era between 2015 and 2017 were classified into 1, 7, 8, and 32 different N450 sequence variants in D9, H1, B3, and D8 genotypes, respectively. Analysis of the detection period of the individual N450 sequence variants showed that the majority of MV strains were detected only for a short period. However, MV strains, MVs/Osaka. JPN/29.15/ [D8] and MVi/Hulu Langat. MYS/26.11/ [D8], which are named strains designated by World Health Organization (WHO), have been detected in many cases over 2 or 3 years between 2015 and 2017. The WHO-named strains have circulated worldwide, causing outbreaks in many countries. Epidemiological investigation revealed repeated importation of these WHO-named strains into Japan. To demonstrate the elimination status (interruption of

INTRODUCTION
In 2005, the WHO member states in the Western Pacific Region (WPR) decided to target the elimination of measles by 2012 in this region. Aiming to eliminate measles in Japan, the Ministry of Health, Labour and Welfare, Japan (MHLW) announced the Special Infectious Disease Prevention Guidelines on Measles (MHLW Measles Guideline) and requested mandatory reporting of all measles cases (i.e., case-based surveillance) since January 2008. In November 2010, to strengthen the surveillance, MHLW issued a notice by the Director of Tuberculosis and Infectious Diseases Control Division in MHLW (Director's Notice), requesting local governments to conduct measles virus (MV) testing by PCR for all suspected measles cases as possible. The MHLW Measles Guideline was revised in April 2013, further requesting local governments to conduct MV detection testing by PCR. The detection of MV was critically important to obtain genotyping evidence that supports the interruption of endemic MV transmission, one of the essential criteria to be verified in achieving measles elimination (World Health Organization [WHO] and Regional Office for the Western Pacific [WPRO], 2013). The variable 450-nucleotide sequence region of the MV nucleocapsid gene (N450) has been generally used for genotyping and monitoring individual MV strains (World Health Organization [WHO], 2015). These MV detection and genotyping data were included in the annual report by the National Verification Committee (NVC) for measles elimination in Japan. After the careful review of the NVC annual report, the Regional Verification Commission (RVC) for measles elimination in WPR verified that Japan achieved measles elimination on March 27, 2015. This study focuses on the molecular epidemiology of MV between 2008 and 2017 in Japan.

Nucleotide Sequencing of the 450-Nucleotide Region in the N Gene
According to the MHLW Measles Guideline under the Infectious Diseases Control Law in Japan, clinical samples (including throat swab, blood, and urine samples) were collected from measles-suspected cases and sent to PHIs to detect MV. The criteria for measles-suspected cases were appearance of two or three clinical signs of fever, rash, and either cough, coryza, or conjunctivitis. Samples, which were positive for MV Abbreviations: MV, measles virus; MHLW, Ministry of Health, Labour and Welfare, Japan; PCR, polymerase chain reaction; PHIs, public health institutes; WHO, World Health Organization. RNA by reverse-transcription PCR (RT-PCR), were subjected to nucleotide sequencing of the WHO-recommended 450nucleotide region in the N gene (N450) (nucleotide positions 1233-1682). The region encodes the carboxyl-terminal 150 amino acids of the nucleoprotein. Briefly, the extracted viral RNAs were reverse-transcribed into cDNAs using a commercially available reverse transcription kit and random hexamer primers. The nucleotide region containing the N450 region was amplified by nested PCR. The first PCR primer set was pMvGTf1m (5 -CGRTCTTACTTYGATCCRGC-3 ) and pMvGTr1 (5 -TTATAACAATGATGGAGG-3 ), and the nested PCR primer set was pMvGTf2m (5 -AGAYTAGGRCARGAGATGGT-3 ) and pMvGTr2 (5 -GAGGGTAGGCGGATGTTGTT-3 ). After purification of the fragment, the nucleotide sequence was determined by a fluorescent dye-terminator cycle sequencing method using the primers, pMvGTf2m or pMvGTr2.

Phylogenetic Analysis
Phylogenetic analysis was conducted using the MEGA program (version 7.0.6). In addition to the MV strains detected in Japan, WHO reference MV strains, the WHO-named strains, and MV strains detected outside of Japan (overseas MV strains) and having the identical sequences to the MV strains detected in Japan were included in this analysis (Supplementary Table 1). The list of MV strains used in this study and DDBJ/EMBL/GenBank accession numbers of these strains are shown in Supplementary  Table 1. The overseas MV strains having the identical sequences to the MV strains detected in Japan were searched by using WHO Measles Nucleotide Surveillance Database (MeaNS 1 ) (WHO and PHE). Phylogenetic trees were constructed by the maximumlikelihood method using the Tamura-Nei model. The reliability of the tree at each branch node was assessed by the bootstrap method with 1,000 replicates. The genotype of MV strains was determined based on the phylogenetic tree topology constructed with the WHO reference strains.

The Global Situation Towards Measles Elimination
In 2012, the Global Vaccine Action Plan (GVAP) was endorsed by the World Health Assembly (World Health Organization [WHO], 2013a). GVAP set a goal to eliminate measles and rubella in at least five WHO regions by 2020 by strengthening immunization programs (World Health Organization [WHO], 2013a). As a result of the global effort on immunization, the number of reported measles cases has declined greatly, resulting in a 80% decrease in measles deaths between 2000 and 2017. The region of the America has declared measles elimination in this region in 2016. Thirty-seven countries (70%) in European region have been also verified as having achieved measles elimination in 2017. In the South-East Asia Region, two countries (Bhutan and Maldives) have achieved measles elimination in 2017. On the other hand, no countries has achieved measles elimination in the Eastern Mediterranean and Africa regions. Japan is a member state in WPR.  Dabbagh et al., 2018).
As just described, we have made great progress toward measles elimination by improving vaccination coverages. However, considerable variations in vaccination coverages exist among regions, countries, and areas, causing measles outbreaks continuously somewhere in the world. Even a small reduction of vaccination rate due to vaccine hesitancy may result in outbreaks of measles (Lo and Hotez, 2017).

Performance of MV Detection and Genotype Analysis
As described above, Japan started case-based surveillance of measles from 2008 in accordance with the MHLW Measles Guideline. At that time, laboratory testing was not mandatory to confirm a clinical diagnosis of measles, but many cases were laboratory-confirmed by detecting MV-specific serum IgM. In Japan, serum specimens are sent to commercial laboratories for detection of measles-specific IgM by ELISA. By contrast, the percentage of measles cases confirmed by PCR was only 0.7-4.7% between 2008 and 2010 ( Table 1). Clinical specimens must be sent to PHIs for virus detection by RT-PCR. After the Director's Notice was issued in November 2010, the ratio of measles cases with virus detection data provided by PCR increased from 23.2 to 28.3% between 2011 and 2013. This ratio further increased from 71.4 to 97.8% between 2014 and 2017 after the revised MHLW Measles Guideline was put into force in April 2013 ( Table 1). The ratio of the cases with virus detection data provided by PCR was 84.5% (387 out of 458 cases) in 2014. When MV is detected by RT-PCR, PHIs (or NIID) then conduct genotype analysis. The great majority of cases, which were positive for RT-PCR, had the genotyping data for MV, which have provided strongly supportive evidence for interruption of endemic MV transmission (measles elimination) in Japan by 2015. The total number of measles cases in 2015 was only 35 (Table 1), which was a record low in the history of Japan. However, several MV epidemics triggered by imported measles cases were recorded in 2016 and 2017, which resulted in 165 and 185 MV cases in Japan. To assess the measles elimination status in Japan, MV genotype data were obtained from most of the cases in 2016 and 2017 (140 and 181 cases out of 165 and 185 cases in 2016 and 2017, respectively) ( Table 1).
Many genotype D9 strains were detected in Japan between 2009 and 2015 ( Table 1; Aoki et al., 2009). They were detected every year. Between 2010 and 2011, large D9 outbreaks have occurred in the Philippines (Centeno et al., 2015;Takashima et al., 2015), and many cases in Japan had travel history to the Philippines ( Centeno et al., 2015). The sequence analysis of MV strains has confirmed that certain D9 strains in Japan have the identical N450 nucleotide sequences to the Philippine D9 strains (Centeno et al., 2015). The D9 strains were first detected in the Philippines in 2007 and have been predominant until 2012 in the Philippines (Fuji et al., 2011;Centeno et al., 2015;Takashima et al., 2015). Outbreaks between 2010 and 2011 in the Philippines were presumably caused by several D9 strains with different N450 sequences (Centeno et al., 2015). D9 virus was also exported from the Philippines to Taiwan and Australia in 2010 and 2011, respectively (Jayamaha et al., 2012;Cheng et al., 2013 (Rosewell et al., 2012;Cheng et al., 2013;Ho et al., 2014), where endemic transmission of MV has been eliminated or well controlled. In 2011, South Korea has also experienced imported D9 cases, although the origin was unclear (Park et al., 2013). In Japan, many genotype B3 strains were detected between 2013 and 2014, while only small numbers of genotype B3 strains were detected after 2015 (Table 1). In total, 260 genotype B3 strains were detected between 2013 and 2017 ( Table 1). In 2014, large measles outbreaks have occurred in the Philippines (Takashima et al., 2015) and many measles cases in 2014 in Japan had travel history to the Philippines (Takahashi et al., 2014;Minagawa et al., 2015;Miyoshi et al., 2015 Yagi et al., 2015;Miyoshi et al., 2017;Kaida et al., 2018). Imported cases from the Philippines were also reported in Korea (Park et al., 2017) and many European countries (Nic Lochlainn et al., 2016;Santibanez et al., 2017).
In recent years, genotype D8 has been one of the major genotypes detected in Japan (Table 1). In 2017, the great majority of MV strains detected in Japan were genotype D8 strains ( Table 1). In total, 324 genotype D8 strains were detected between 2009 and 2017. In 2014, a large outbreak of the D8 genotype strain occurred in Vietnam (Pham et al., 2014)

Detailed Analysis of Each Genotype
One or two reference strains have been set for each genotype. However, using only the genotype classification and reference strains is currently insufficient for monitoring the global transmission of MV. Considering this situation, WHO has set a certain number of "named strains" in addition to the reference strains. These named strains represent epidemiologically significant MV strains that circulate in multiple countries (World Health Organization [WHO], 2015).

Genotype D9
In total, 108 genotype D9 strains were detected between 2009 and 2015. These strains were divided into as many as 19 different sequence variants (D9 seq1 to D9 seq19 [D9 seq1-19] variants). A phylogenetic analysis was performed using a dataset consisted of the D9 seq1-19 variants, WHO D9 reference strain (MVi/Victoria.AUS/12.99 [D9]), and two WHO-named strains. In addition, 11 overseas D9 strains with the identical sequences to one or other D9 seq1-19 variants were included ( Figure 1B). The tree was constructed together with genotype D4 strains. Unlike the genotype D4 strains, the majority of the genotype D9 strains had no matched named strain with the identical N450 sequence. Exceptionally, D9 seq12 variant had an identical N450 sequence to the MVs/Bristol.GBR/13.05/ (D9-Bristol)-named strain. Between 2011 and 2012, D9-Bristol strain was detected in many cases in Malaysia and Turkey, or from individuals with travel history to Malaysia and Indonesia (data from MeaNS; WHO and PHE). In Japan, the D9 seq12 (D9-Bristol) variant was detected from eight cases (one, three, and four in 2011, 2012, and 2013, respectively  . In Indonesia, the genotype D9 and G3 strains were endemically circulated (D8 was newly detected in 2014) (Subangkit et al., 2017). In Malaysia, both genotype D9 and D8 strains were predominantly detected between 2013 and 2016 (Hagan et al., 2018). Between 2013 and 2015, D9-Yamanashi strain was also detected in Australia, Belarus, Germany, and United States from individuals with travel history to Indonesia or Malaysia (WHO and PHE).
In 2014, D8-gr.1a, D8-gr.1b, and D8-gr.2b MV strains were detected in Japan. D8-gr.1a, D8-gr.1b, and D8-gr.2b contain three, five, and four WHO-named strains, respectively, suggesting that MV strains in these sub-groups circulated globally. The D8 seq6 variant in D8-gr.1b was one of the major strains detected in Japan between 2011 and 2014. The D8 seq6 variant had the identical N450 sequence to the MVi/Villupuram.IND/03.07/ (D8-Villupuram) named strain. The D8-Villupuram strain has been detected worldwide between 2010 and 2015 (WHO and PHE). In 2015, only 35 measles cases were detected in Japan, and 12 measles cases were caused by genotype D8 strains in the D8-gr.1b and D8-gr.2b (Figure 4 and Table 1). In 2016, the D8-gr.2a MV strains were detected for the first time in Japan together with the D8-gr.2b MV strains. No D8gr.1 MV strains were detected in 2016, suggesting interruption of the transmission of D8-gr.1 MV strains. In 2017, the D8-gr.1b MV strains were again detected together with D8-gr.2a and D8gr.2b MV strains. In 2017, the D8-gr.1c MV strains were detected for the first time in Japan from a single case.

Chronological Analysis of MV Strains
On March 27, 2015, the RVC for measles elimination in WPR verified that Japan achieved measles elimination. However, when epidemiological or laboratory evidence indicates the presence of a chain of transmission of an MV strain that persists for >12 months in a defined geographical area, it is considered that reestablishment of endemic transmission has occurred in the area where measles had been eliminated previously (World Health Organization [WHO], 2013b). To understand the situation of MV transmission in Japan in the postelimination era, the detection pattern of each MV sequence variant was analyzed in chronological order. The 891 MV strains detected between 2008 and 2017 were classified into seven genotypes (D5, D4, D9, H1, G3, B3, and D8) and 124 sequence variants (D5 seq1-7, D4 seq1-6, D9 seq1-19, H1 seq1-16, G3 seq1-2, B3 seq1-23, and D8 seq1-51). Among them, the strains detected in the postelimination era (between 2015 and 2017) were classified into 48 MV sequence variants (1, 7, 8, and 32 sequence variants in the D9, H1, B3, and D8 genotypes, respectively). Among the 48 MV sequence variants, most variants were detected in periods shorter than 5 weeks (Supplementary Tables 2-5). Four MV sequence variants (D9 seq15, H1 seq4, H1 seq6, D8 seq38, and D8 seq34) were detected in a period of 5-10 weeks. Only two MV sequence variants, D8 seq8 and D8 seq17, were detected for periods longer than 10 weeks (Supplementary Table 5). More importantly, these variants (D8 seq8 and D8 seq17) were detected for a consecutive 2 and 3 years, respectively (Figure 5 and Supplementary Table 5). The line graphs in Figure 5 show the number of reported measles cases by week (data from Infectious Agents Surveillance Report, Infectious Disease Surveillance Center, NIID, Japan), and the colored horizontal bars in Figure 5 show the genotype data of MV strains. The ratio of measles cases with genotype information was low in 2013, but most measles cases after 2014 included MV genotype data. It was clear that the genotype D9, H1, and B3 strains were detected only intermittently with long (>4 weeks) nondetection periods. However, the genotype D8 strains were  Five lineages (D8-gr.1a, D8-gr.1b, D8-gr.1c, D8-gr.2a, and D8-gr.2b) tentatively proposed in this study are indicated.
detected constantly even during the postelimination era between 2015 and 2017 ( Figure 5A).
When the D8 seq1-51 variants were assessed individually, most of them were detected only sporadically (Supplementary Table 5). Certain strains caused small outbreaks, but were detected only for a short period. Only two sequence variants (D8 seq8 [D8-HL] and D8 seq17 [D8-Osaka]) were detected many times over 2 or 3 years, respectively ( Figure 5B and Supplementary Table 5). Figure 5B shows the detection data of genotype D8 strains by week. The D8 seq8 (D8-HL) variants were detected in the 19th, 38th, 40th, and 46th weeks in 2016 and the 2nd, 3rd, and consecutive 7 weeks between the 10th and 16th weeks. Since each sporadic case or small outbreak was separated by a long period (more than 4 weeks), the transmission of D8 seq8 (D8-HL) variants has not persisted for >12 months in Japan. In 2015, the D8 seq17 (D8-Osaka) variants were only sporadically detected in the 27th, 51th, and 53th weeks ( Figure 5B). Between 2016 and 2017, the D8 seq17 (D8-Osaka) variants were detected for a long period of time (from the 19th week in 2016 until the 20th week in 2017). During this period, in total, D8 seq17 variant (D8-Osaka) was detected in 85 MV cases, including those in 11 small outbreaks (n = 2-16). However, the detection period was apparently separated into two periods. The former period started from 19th week in 2016 and ended at 39th week in the same year, while the latter period started from the 48th week in 2016 and ended at the 20th week in 2017. Between these two periods, there was an 8-week interval when no D8 seq17 (D8-Osaka) variant was detected ( Figure 5B and Supplementary Table 5). In addition, the geographic distribution of the measles cases in the two periods was separated. The measles cases in the former period were detected in three prefectures in the Kansai region, a western region of Japan, and the measles cases in the latter period were detected in two prefectures in the Kanto region, in an eastern region of Japan. D8 seq17 (D8-Osaka) variant was not detected after the 21th week in 2017 (as of December 15, 2018). These data suggested that transmission of D8 seq17 (D8-Osaka) variant did not persist for >12 months in Japan.

CONCLUSION
In this study, we analyzed the molecular epidemiology of MV strains between 2008 and 2017 in Japan. The genotyping evidence that supports the interruption of endemic MV transmission is one of the essential criteria to be verified as achieving measles elimination. Thus, MV detection testing by RT-PCR and genotype analysis has been performed on most of the measles cases in PHIs nationwide in Japan. Following these efforts, Japan had strong molecular epidemiological evidence that supports the interruption of endemic MV transmission in Japan. Accordingly, Japan was successfully verified to have achieved measles elimination in March 2015. This effort to detect MV by PCR and genotyping has continued during the elimination era in Japan. The molecular epidemiological data of MV in certain prefectures or cities were reported in previous studies (Aoki et al., 2009;Kurata et al., 2009;Nagano et al., 2011;Tanaka et al., 2011;Watanabe et al., 2012;Miyoshi et al., , 2015Miyoshi et al., , 2017Nadaoka et al., 2014;Minagawa et al., 2015;Yagi et al., 2015;Kaida et al., 2018;Komabayashi et al., 2018), but a nationwide analysis was conducted in this study for the first time. The data demonstrated that many different MV strains were detected in Japan. Although certain genotypes have been detected almost every year, the detailed sequence analysis of individual MV strains revealed multiple importations and disappearance of great numbers of different MV strains. Epidemiological investigations have also detected overseas travel histories in many measles cases. Especially, WHO-named strains were detected repeatedly in recent years, because these named strains have circulated in many countries worldwide. Demonstrating the elimination status in the situation with repeated importation of the same strains is challenging. Nevertheless, the detailed sequence data of individual MV strains and chronological analysis of these strains provided sufficient evidence showing that Japan has still maintained its measles elimination status.

DATA AVAILABILITY
The datasets generated for this study can be found in GenBank, All accession numbers are listed in Supplementary Table 1.

ETHICS STATEMENT
The Infectious Diseases Control Law in Japan defines measles as a notifiable infectious disease and the MHLW Measles Guideline under the Infectious Diseases Control Law requests mandatory reporting of all measles cases. The Guideline also requests for physicians and local governments to conduct MV detection test by PCR and genotype analysis for all measles suspected cases in principle. The Ethics Committee of the NIID agreed to analyze and publish the MV epidemiological data in academic journals (No. 975).