First insights into circulating XDR and pre-XDR Mycobacterium tuberculosis in Southern Brazil

Infection, Genetics and Evolution 78 (2020) 104127

Contents lists available at ScienceDirect

Infection, Genetics and Evolution journal homepage: www.elsevier.com/locate/meegid

Research Paper

First insights into circulating XDR and pre-XDR Mycobacterium tuberculosis in Southern Brazil

T

⁎

Richard Steiner Salvatoa,b, , Elis Regina Dalla Costab,c, Ana Júlia Reisd, Sun Hee Schiefelbeinb, Maria Laura Halonb, Regina Bones Barcellosb, Gisela Unise, Cláudia Fontoura Diase, Miguel Viveirosf, Isabel Portugalg, Pedro Eduardo Almeida da Silvad, Afrânio Lineu Kritskic, João Perdigãog, Maria Lucia Rosa Rossettia,b a

Programa de Pós-graduação em Biologia Celular e Molecular, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Rio Grande do Sul, Brazil Centro de Desenvolvimento Científico e Tecnológico (CDCT), Centro Estadual de Vigilância em Saúde, Secretaria Estadual da Saúde do Rio Grande do Sul, Porto Alegre, Rio Grande do Sul, Brazil c Centro de Pesquisa em Tuberculose, Faculdade de Medicina, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil d Medical Microbiology Research Center (NUPEMM), Faculdade de Medicina, Universidade Federal do Rio Grande - FURG, Rio Grande, Rio Grande do Sul, Brazil e Hospital Sanatório Partenon, Porto Alegre, Rio Grande do Sul, Brazil f Unidade de Microbiologia Médica, Global Health and Tropical Medicine, Instituto de Higiene e Medicina Tropical, Universidade Nova de Lisboa, Lisboa, Portugal g iMed.ULisboa - Research Institute for Medicines, Faculdade de Farmácia, Universidade de Lisboa, Lisboa, Portugal b

A R T I C LE I N FO

A B S T R A C T

Keywords: Tuberculosis Multi-drug resistance Extensively drug-resistance Transmission Whole genome sequencing

Drug-resistant tuberculosis (DR-TB) is major problem in the fight against TB. Multidrug resistant (MDR) TB patients have a reduced treatment success rates and for, extensively drug-resistant (XDR) TB the cure rate does not exceed 25% in many countries. To evaluate the pre-XDR-TB and XDR-TB prevalence and transmission in Rio Grande do Sul State, in southern Brazil, we performed a retrospective WGS-based analysis of 87 MDR-TB cases, aiming to identify resistance-conferring mutations and its phylogenetic distinctiveness. Using a five SNP threshold for genomic clustering, 60 strains were genomically linked within 10 clusters, including 14 likely transmission events identified by retrospective conventional epidemiological investigation. Moreover, five likely transmission events involved 17 patients deprived of liberty in the same prison establishment. Mutations associated with isoniazid and rifampicin resistance were identified respectively in 97.70% and 98.85% of MDR M.tb strains, more frequently in katG and rpoB genes. In total, we identified eight (9.19%) pre-XDR and four (4.59%) XDR M.tb strains. Resistance to ofloxacin was observed in seven (8.04%) strains, all of them presenting resistance-conferring mutations. Phenotypic resistance from capreomycin and kanamycin was found in seven (8.04%) and four (4.59%) strains respectively, but no classic mutations associated with resistance to these drugs was identified. The results put in evidence a scenario involving multiple phylogenetically distinctive clades associated with pre-XDR and XDR-TB in the largest state of southern Brazil, while stressing the potential of using WGS to predict anti-TB drug resistance and need to halt MDR-TB transmission in the region.

1. Introduction Worldwide, tuberculosis (TB) is now the tenth leading cause of death. It is estimated that 1.3 million people died from active TB in 2017.(WHO, 2018) One of the major factors contributing for this scenario is the increasing multiple resistance to anti-TB drugs that significantly reduces treatment success. In fact, in contrast with the overall global treatment success rate of 82%, multidrug-resistant (MDR) TB is

associated with only 60% of positive treatment outcomes and for extensively drug-resistant (XDR) TB merely 26% (WHO 2018a; Bastos et al., 2017). In more recent years, whole genome sequencing (WGS) presents an attractive option for the detection and genotypic prediction of drug susceptibility. As such, several low-incidence countries are now moving to phase-out classical culture-based phenotypic drug susceptibility towards WGS-based diagnosis. Nevertheless, in many settings data

⁎

Corresponding author at: Centro de Desenvolvimento Científico e Tecnológico, Centro Estadual de Vigilância em Saúde, Secretaria Estadual da Saúde do Rio Grande do Sul, Av. Ipiranga 5400, CEP 90610-000, Porto Alegre, Rio Grande do Sul, Brazil. E-mail addresses: [email protected], [email protected] (R.S. Salvato), [email protected] (M. Viveiros), iportugal@ff.ulisboa.pt (I. Portugal). https://doi.org/10.1016/j.meegid.2019.104127 Received 27 July 2019; Received in revised form 20 November 2019; Accepted 24 November 2019 Available online 26 November 2019 1567-1348/ © 2019 Elsevier B.V. All rights reserved.

Infection, Genetics and Evolution 78 (2020) 104127

R.S. Salvato, et al.

2.3. DNA extraction

regarding the genetic diversity of M.tb or the prevalence of unusual M.tb strains, especially those related with drug resistance, is still lagging behind and, it is increasingly needed to complement the disease control measures (Meehan et al., 2019) Among the global WHO estimated cases of MDR-TB, about 60% occur in the BRICS countries (Brazil, Russia, India, China, and South Africa). Rio Grande do Sul (RS), in southern Brazil, is currently fourth in TB incidence rate among the Brazilians states, with 39.5 cases per 100,000 people.(Brasil. Ministério da Saúde, 2018) About the circulating M.tb strains genetic profile, the predominance of M.tb Lineage 4 in the region is well known, (Perdigão et al., 2019) as shown by data from the CPLP-TB database (available at http://cplp-tb.ff.ulisboa.pt). In addition, previous studies conducted in the state, using classical genotyping techniques demonstrated the high degree of M.tb genotype clustering among MDR-TB,(Esteves et al., 2018; Salvato et al., 2019) indicating possible events of recent transmission. Facing the global need for MDR/XDR-TB surveillance, we undertook a WGS-based analysis of MDR M.tb clinical strains circulating in the southern Brazil between 2013 and 2014, aiming to identify and characterize the panel of mutations conferring resistance to first and secondline anti-TB drugs, as well as, a state-wide MDR-TB transmission analysis, ultimately genotypically and epidemiologically characterizing the pre-XDR and XDR situation in this region.

The genomic DNA of the eighty-seven M.tb samples was extracted from sputum culture in Lowenstein–Jensen solid medium using Cetyltrimethylammonium Bromide (CTAB) method, as described by Van Embden et al. (1993) 2.4. Mycobacterial interspersed repetitive units - variable number of tandem repeats (MIRU-VNTR) 24-locus genotyping MIRU-VNTR 24loci was performed as previously described.(De Beer et al., 2012). Fragment size of the amplicons was analyzed on an ABI 3130xl DNA sequence analyzer (Applied Biosystems, California, USA) and the number of copies of each locus was determined by automated assignment using the Genemapper 3.2.1 software (Applied Biosystems, California, USA). Undefined results were double checked on agarose gels and VNTRs are defined according Supply et al.(Supply et al., 2000) A minimum spanning tree was created with the MIRU-VNTRplus database website. Strains with maximum difference of two loci were defined as MIRU-VNTR clonal complexes and clusters were defined as strains with identical MIRU-VNTR patterns.(Allix-Béguec et al., 2008) 2.5. Whole genome sequencing The eighty-seven clinical strains were subjected to WGS. Briefly, approximately one microgram of genomic M.tb DNA per sample was fragmented using a Q800R2 sonicator (QSonica, Newtown, CT, USA) with the following parameters: 3 min sonication with 15 s pulse on, 15 s pulse off, and 20% amplitude. The fragmented DNA was selected by size to target 600–650 bp by fragment separation using the Agencourt AMPure XP beads (Beckman Coulter, Code A63882). DNA Library preparation was performed using the NEBNext® Ultra™ II DNA Library Prep Kit for Illumina (New England BioLabs, Code E7645L). Adapters and 8 bp index oligos purchased from IDT® (Integrated DNA Technologies, San Diego, CA) based on Kozarewa and Turner (2011), were used in place of those supplied in the NEB preparation kit in a dual-indexing approach.(Stone et al., 2016) Paired-end sequencing (2 × 150 bp) was performed on an Illumina NextSeq machine using either a 300 cycle v2 mid output or high output kit (Illumina, Code FC404-2003 or Code FC-404-2004) using standard Illumina® procedure.

2. Material and methods 2.1. Study population and Mycobacterium tuberculosis strains The M.tb samples were obtained of the State Central Laboratory (LACEN- Rio Grande do Sul), the TB diagnosis reference laboratory of the Rio Grande do Sul State, and were collected from sputum samples from TB patients attending the Hospital Sanatório Partenon (HSP), the state reference centre for TB-DR treatment, between 2013 and 2014. This study included patients who had 18 years or older, with bacteriological confirmation of MDR-TB. Over the study period, the HSP notified a total of 186 patients with MDR-TB in the SITE-TB website (national database for DR-TB cases reported in Brazil), and only the first sample collected from each patient was included in the study. Epidemiological data was retrieved from SITE-TB and SINAN databases (http://portalsinan.saude.gov.br).

2.6. Bioinformatics and data analysis Initial sequence read quality control was performed using FastQC (v0.11.7) (http://www.bioinformatics.babraham.ac.uk/projects/ fastqc/), and repeated after the reads were trimmed to remove adapter sequences and low quality reads using trimmomatic (v0.33) (parameters: LEADING:3 TRAILING:3 SLIDINGWINDOW:4:20 MINLEN:36).(Bolger et al., 2014) Surviving reads were mapped to the M.tb H37Rv reference genome (GenBank accession number: NC_ 000962.3) using BWA-MEM (v0.7.16). Samtools (v1.9)(Li et al., 2009) was used to convert from SAM to BAM format and sorting of mapped sequences. The quality of the resulting BAM file was checked using Qualimap(Cruz et al., 2012) and sambamba (v0.6.8) was used to mark read duplicates.(Tarasov et al., 2015) Variants (SNPs and small INDELs) were called using Samtool tools. Variants were filtered based on the following criteria: mapping quality ≥50, base alignment quality ≥23 and ≤ 2000 reads covering each site. After, we filtered the raw VCF with the following parameters: minimum read depth of 10 to call a variant and 2000 maximum read depth. Variant functional annotation of the VCF files was performed with snpEff (v4.3).(Cingolani et al., 2012) Mutations associated with drug-resistant TB were initially detected using TB-Profiler(Coll et al., 2015) pipeline in command line version (0.3.0), using as input the BAM files and default settings. In addition, mutations were also manually annotated from VCF and visually

2.2. Phenotypic drug susceptibility testing (DST) DST for the first-line anti-TB drugs; rifampicin (RIF) (1.0 mg/L), isoniazid (INH) (0.1 mg/L), ethambutol (EMB) (5.0 mg/L) and streptomycin (STR) (1.0 mg/L), was carried out using the liquid BACTEC™ MGIT™ 960 SIRE Kit for the BACTEC Mycobacteria Growth Indicator Tube 960 (MGIT 960) system (Becton Dickinson Diagnostic Systems, Sparks, MD). From a total of 186 MDR-TB samples, 87 were available for second-line DST, since it was a retrospective study, the rest of the samples were not viable or available perform DST and DNA extraction. The determination of the second line susceptibility profile was performed using the resazurin microtiter assay (REMA) as previously described.(Palomino, 2002; Martin and Palomino, 2009). We evaluated the M.tb strains susceptibility to ofloxacin (OFX), kanamycin (KAN) and capreomycin (CAP), with concentration breakpoints of 4 mg/L, 5 mg/L and 5 mg/L, respectively. The MDR M.tb strains with additional resistance to at least one fluoroquinolone (FQ) or one second-line injectable drug (CAP or KAN) were considered pre-XDR, strains resistant to both groups were considered XDR. The REMA method was also used to determine the MIC for INH and RIF in strains that showed a disagreement between the WGS-based resistance prediction and DST, using as cut-off concentrations of ≤0,25 mg/L and ≤ 0,5 mg/L to INH and RIF respectively. 2

Infection, Genetics and Evolution 78 (2020) 104127

R.S. Salvato, et al.

checked in the BAM files. Phylogenetic analysis was performed using snippy pipeline v3.1 (https://github.com/tseemann/snippy) for variant calling and alignment of all core-genomes. SNP positions within PE/ PPE genes or other repetitive regions associated with low mappability scores were removed from the final core-genome alignment, which was composed of 10,670 sites. A maximum-likelihood phylogenetic tree was generated using the software RAxML (v8.10.12),(Stamatakis, 2014) applying the generalized time reversible (GTR) model and 1000 bootstrap replicates, and the resulting tree was rooted using M. canettii (Genbank accession number: NC_019950.1). In silico spoligotyping profiles were obtained using SpoTyping (v2.0).(Xia et al., 2016) Genomic pairwise distances were obtained by snp-dists tool (https:// github.com/tseemann/snp-dists) and a minimum spanning tree generated using Phyloviz Online (http://online2.phyloviz.net/) and the implemented goeBURST algorithm.(Francisco et al., 2009) A threshold of ≤5 SNPs was used to define genomic clusters(Rosłanowski et al., 2012) among the core-genome alignment.

3.3. Multi-drug resistance and recent transmission: A state-wide perspective Next, to identify genomic clusters associated with ongoing recent transmission events, a five SNP threshold was used to define genomic clusters. The phylogenetic inference resulted in the identification of 60 cases genomically linked distributed across 10 clusters (Table 1). Among these genomically linked cases, 41 were linked through classical epidemiological investigation (Fig. 2). Retrospective analysis of epidemiological data revealed eight epidemiological links (epi-links) among genomic clusters in which patients lived in nearby addresses. It was also identified five possible transmission events, involving 17 patients that spent time in the same prison. The mean SNP difference between the clustered strains was of 1.32 SNPs and most epidemiologically linked strains had a distancing up to 3 SNPs. MIRU-VNTR analysis showed nine clonally complexes and one additional cluster with two strains (Supplementary Fig. 1), eight of these clonally complexes were also identified by WGS. The unidentified complex and cluster had distances > 12 SNPs between the strains and had no epi-link identified. WGS analysis identified two strains with genetic distance of 2 SNPs that were not grouped by MIRU-VNTR and showed also a genomic cluster with two strains with identical coregenome, identified as house contact, which was grouped into two different clonally complexes in MIRU-VNTR clustering. Interestingly, MIRU-VNTR did not group several strains clustered by WGS with confirmed epidemiological links, including clusters with identical coregenome. The most interesting was a cluster of six strains that had the same core-genome, five of which from patients with confirmed epilinks.

2.7. Sequence data Mycobacterium tuberculosis genome data were deposited in the NCBI BioProject PRJNA 535343, see supplementary materials. 2.8. Ethical approval This study was approved by the Research Ethics Committee of the Fundação Estadual de Produção e Pesquisa em Saúde (FEPPS/RS), protocol number 1.587.621 CAAE: 18269313.0.0000.5320.

3.4. Drug resistance mutations

3. Results

From the 87 MDR M.tb strains, nine (10.34%) were also resistant to EMB, and 11 (12.64%) to STR. Among these, two strains were resistant to the four first-line drugs tested. Mutations associated with INH resistance were identified in 85 (97.70%) of MDR-TB samples, being the Ser315Thr mutation in the katG gene the most frequent - 89.41% (76/ 85). Mutations conferring resistance to RIF were identified in 86 (98.85%) isolates, with the mutation Ser450Leu in the rpoB gene the most frequent - 74.41% (64/86). An atypical insertion of 12 nucleotides at the 435 codon in the rpoB gene, previously reported by our group, (Esteves et al., 2018) was found in nine (10.34%) RIF-resistant strains. The discordant results between the WGS-based prediction and the phenotypic result for INH and RIF were confirmed with MIC results for these drugs. Concerning EMB resistance, among the nine EMB-resistant strains, eight (88.88%) bore the embB Met306Val mutation. However, 25 (32.05%) of the 78 EMB susceptible strains also carried EMB relatedresistance mutations such as Met306Val present in 19 (24.35%) of the EMB susceptible strains. Among the 11 STR resistant strains, six (54.54%) of them there were mutations related to resistance (Table 2). Regarding resistance to second-line anti-TB drugs, OFX resistance was initially observed in six strains, and all of them agreed with WGSbased prediction with mutation in codons 88, 90 and 94 of the gyrA gene (Table 3). One susceptible strain by DST carried a gyrA mutation (Asp94Asn) associated with OFX resistance, however, review of medical records showed that this patient had a new DST performed three months after presenting resistance to OFX, thus, we also considered this strain as OFX-resistant, with OFX resistant strains amounting to 8.04% among the 87 MDR-TB strains. In relation to CAP and KAN, seven (8.04%) and four (4.59%) strains showed resistance, respectively. However, we did not identify mutations previously associated with resistance to these drugs in these strains, although the review of medical records showed that they were non-responsive to the second-line regimen including aminoglycosides and were considered to be XDR-TB cases and treated as such.

3.1. Multi-drug resistant tuberculosis in Rio Grande do Sul, Brazil Between 2013 and 2014 a total of 186 MDR-TB patients were notified in Rio Grande do Sul State representing 1.51% of all TB cases notified in this period, and 87 of them were available for further molecular studies (Supplementary Table 1). Briefly, among the MDR-TB cases studied, 65 (74.71%) patients were male and 52 (59.77%) aged between 26 and 45 years. All patients had pulmonary TB, although one patient also presented miliary TB and two had pleural TB. Among all patients, 15 were considered primary MDR-TB cases, since no previous treatment of these individuals was found in any database. Besides that, 29 (33.33%) individuals were tobacco smokers, 30 (34.48%) were illicit drug users, 23 (26.44%) alcoholic and 29 (33.33%) were infected with HIV/AIDS. Regarding the anti-TB treatment that was being administered at the time of sample collection, favourable and unfavourable outcomes were identified in 43 (49.43%) and 44 (50.57%) patients, respectively. 3.2. Mycobacterium tuberculosis genetic diversity associated with multidrug resistant tuberculosis in Rio Grande do Sul The eighty-seven MDR M.tb strains were classified in 17 distinct spoligotypes, comprising seven spoligotype families, 14 sub-families and 17 different spoligo-international types (SIT). The LAM family was the most frequent (60.91% of the strains), followed by T (22.98%) and PINI2 (9.19%). The main subfamilies were LAM5 (20.68%), LAM9 (14.94%) and T1 (14.94%). In addition, most of the strains belonged to SIT 93 (20.68%), SIT 42 (10.34%) and SIT 53 (9.19%). When we used SNP-based typing, using the SNP barcode proposed by Coll et al. (2014),(Coll et al., 2014) all 87 strains were classified as Lineage 4 (Euro-American) and four different sublineages were detected, of which the most prevalent sublineage were LAM (78.16%) and Haarlem (13.79%) (Fig. 1). 3

Infection, Genetics and Evolution 78 (2020) 104127

R.S. Salvato, et al.

Fig. 1. Maximum likelihood tree constructed from SNP alignment of the 87 M.tb strains. Colored ranges represent the SNP-based M.tb sublineages according to Coll et al. (2014) SNP barcode. The spoligotype is represented together with the identification of the strains.

patient profile of the 12 pre-XDR/XDR-TB patients (Supplementary Table 2), we observed that 41.7% (5/12) were prisoners, 58.3% (7/12) were illicit drug users and all of them had TB in the past. All XDR and four pre-XDR patients were in a MDR-TB treatment using second-line drugs at least three months, but all patients only presented negative bacilloscopy and negative culture for M.tb after initiate treatment for XDR-TB. Besides that, four XDR-TB cases were identified and two of them, despite having negative culture after treatment for XDR-TB, had death as outcome.

Table 1 WGS-based genomic clusters among the 87 MDR-TB strains and identified epilinks. Genomic cluster

Strains

Epi-links

Strains in epi-links

Community epilinks

Prison epilinks

1 2 3 4 5 6 7 8 9 10 Total

13 9 9 8 4 4 7 2 2 2 60

3 2 2 1 1 2 1 1 0 1 14

9 4 5 6 3 4 6 2 0 2 41

2 0 1 0 1 2 0 1 0 1 8

0 2 1 1 0 0 1 0 0 0 5

4. Discussion Rio Grande do Sul is the largest state among the three states of southern Brazil region and has the third leading incidence TB rate of the country.(Brasil. Ministério da Saúde, 2018) Herein, we have evaluated the prevalence of pre-XDR and XDR-TB strains and the genomic epidemiology of MDR-TB at a state-wide level. The predominance of strains belonging to the LAM (78.16%) and Haarlem (13.79%) sublineages observed in our study is in accordance with previous studies performed in Rio Grande do Sul.(Esteves et al., 2018; Salvato et al., 2019) Using a five SNPs cut-off to determine

3.5. Pre/extensively drug-resistant tuberculosis According DST results, we identified eight (9.19%) pre-XDR M.tb strains and four (4.59%) XDR M.tb strains. When we analyzed the 4

Infection, Genetics and Evolution 78 (2020) 104127

R.S. Salvato, et al.

Fig. 2. Minimum spanning tree from 87 M.tb core-genome alignment showing the 10 genomic clusters identified. The strains contained in the green dashed highlights are from patients with community epi-link found. In the red dashed highlights are the strains from patients who spent some time in the same prison. When strains are grouped at the same point means that there is no difference SNP, and strains with an asterisk represent household contacts. #One of the patients had no epi-link. In the map, is represented geographic distribution of the 87 MDR-TB strains in the Rio Grande do Sul State, in the zoomed image is represented the Porto Alegre metropolitan region. The colors represent the genomic cluster to which the strain belongs. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).

Mortensen et al., 2016). The genomic clusters not identified by the MIRU-VNTR analysis, demonstrate the importance and potential of the WGS in the understanding of M.tb transmission in this setting.(BjornMortensen et al., 2016). An interesting finding revealed by our WGS analysis was the

genomic clusters,(Bjorn-Mortensen et al., 2016; Walker et al., 2014) 10 genomic clusters involving strains from 60 patients were identified. Most epidemiologically linked strains had a distancing of up to 3 SNPs in this study, in agreement with the proposed by Roetzer et al.(Roetzer et al., 2013) and also reported by Bjorn-Mortensen et al.(Bjorn5

Infection, Genetics and Evolution 78 (2020) 104127

R.S. Salvato, et al.

Table 2 Frequency of mutations associated with first-line anti-TB drugs resistance. Resistant strains

Susceptible strains

51 (58.62) 25 (28.74) 3 (3.45) 2 (2.30) 2 (2.30) 1 (1.15) 1 (1.15) 1 (1.15) 1 (1.15)

0 0 0 0 0 0 0 0 0

(0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0)

49 (56.32) 13 (14.94) 9 (10.34) 5 (5.75) 2 (2.30) 2 (2.30) 1 (1.15) 1 (1.15) 1 (1.15) 1 (1.15) 1 (1.15) 1 (1.15) 1 (1.15)

0 0 0 0 0 0 0 0 0 0 0 0 0

(0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0)

8 0 0 0 0 0 0 0 1

(9.20) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (1.15)

19 (21.84) 2 (2.30) 1 (1.15) 1 (1.15) 1 (1.15) 1 (1.15) 1 (1.15) 1 (1.15) 51 (58.62)

5 2 2 2

(5.75) (2.30) (2.30) (2.30)

73 (83.91) 3 (3.45) 0 (0.0) 0 (0.0)

Mutations

INH

katG Ser315Thr Ser315Thr

OFLO

KANb CAPb

Mutation

80 (91.95) 2 (2.29) 1 (1.14) 1 (1.14) 1 (1.14) 1 (1.14) 1 (1.14) 83 (95.40) 4 (4.59) 80 (91.95) 7 (8.04)

Susceptible Resistant Resistant Resistant Resistant Resistant Susceptiblec Susceptible Resistant Susceptible Resistant

Wt gyrA gyrA gyrA gyrA gyrA gyrA wt wt wt wt

rpoC

embB Met306Val

embA

Asp735Asn Leu527Val

−11C > A

rpsL

Lys43Arg

Contrariwise to what happens in the general population, there is an increasing trend concerning the incidence of TB in prison establishments nationwide, with current TB incidence rates in prison inmates being 28 times higher when compared to the overall population (Brasil. Ministério da Saúde, 2018). This problem affects not only the population deprived of liberty, but also acts as an important reservoir that feeds the dynamics of TB transmission throughout the population (Sacchi et al., 2015). This fact shows that TB control measures within prisons are therefore urgently needed to reverse this scenario. The large number of MDR-TB strains in genomic clusters, together with the number of epi-links identified, reveals the high rate of recent transmission in the region, similar to the scenario observed in another region of Brazil with a high incidence of TB (Peres et al., 2018), and demonstrate that interrupting MDR-TB strains transmission in the southern Brazil is an urgent need for the disease control. Regarding drug resistance, WGS-based analysis was able to correctly predict 97.70% and 98.85% of INH and RIF resistant strains, in accordance with similar values of previously published evidence on other settings (Zignol et al., 2018; WHO, 2018b). High sensitivity and specificity rates, have been observed in several studies that used WGS to detect INH and RIF resistance, the two main first-line drugs (AllixBéguec et al., 2018). In fact, the knowledge about the molecular basis of drug resistance has increased significantly in recent years, enabling this high accuracy of drug resistance prediction in M.tb by molecular methods related to INH and RIF resistance (especially the latest) (AllixBéguec et al., 2018). A worldwide uncommon insertion of 12 nucleotides to the 435 codon in the rpoB gene, described so far only in the Rio

Table 3 Frequency of mutations associated with second-line anti-TB drugs resistance.

b

rpoB Ser450Leu Ser450Leu/Ala286Val 435D > 435DQNNP His445Asp His445Tyr Ser431Thr/Asp435Tyr 433FMD > F Ser450Leu Ser450Leu Ser450Trp His445Cys Asp435Val No mutations

Asp1024Asn Gln497Lys Gly406Asp Met306Ile Met306Ile/Gly406Asp Met306Val/Gly406Ala No mutations rrs No mutations 492C > T 517C > T

STR

DSTa

−15C > T -15C > T

−48 G > A −57C > T

EMB

Number of strains (%)

ahpC

No mutations Ser315Asn Asn138Ser Asp94Gly

RIF

Drug

inhA-promoter

(Asp94Tyr) (Asp94Asn) (Ala90Val/Asp94Asn) (Asp94Gly) (Gly88Cys) (Asp94Asn)

a

DST: Drug Susceptibility test; OFX (Ofloxacyn); KAN (Kanamycin); CAP (Capreomycin). Strains with breakpoints ≥4 mg/L, ≥5 mg/L and ≥ 5 mg/L, for Ofloxacin, Kanamycin and Capreomycin, respectively, were considered resistants; c New DST performed three months after, presenting resistance to OFX, we considered this strain OFX-resistant. b

identification of a large cluster of seven strains. The latter comprised six strains showing an identical core-genome, five of which from patients that spent time deprived of liberty in the same prison establishment; in addition, four other clusters had prisoners involved. The imprisoned population represents 0.3% of the Brazilian population,(Macedo et al., 2017) however, in 2017 approximately 10.5% of the TB cases in Brazil occurred among prisoners.(Brasil. Ministério da Saúde, 2018). 6

Infection, Genetics and Evolution 78 (2020) 104127

R.S. Salvato, et al.

5. Conclusions

Grande do Sul state,(Esteves et al., 2018) was present in 10.34% of the RIF-resistant strains. In silico study with MIC and growth curve assays that has been carried out by our group (unpublished results), and indicates that this insertion, with a duplication of four amino acids, decreases the binding efficiency of RIF and RNA polymerase, accounting for the phenotypic resistance detected. Still among the first-line drugs, we found in this study sensitivity and specificity values of 88% and 65% for EMB and 54% and 96% for STR, respectively. These values showed high variability in different regions of the world, (Zignol et al., 2018) turning a challenge the prediction of resistance to these drugs by molecular methods. However, in addition to this variation, an important fact already reported by several studies, is the high number of false susceptibilities that occurs in the MGIT960 system for certain drugs, such as EMB, whose criticalconcentrations tested in the suystem might be too high for the epidemiological and clinical cut-offs that are becoming more and more robust for anti-TB drugs.(Yakrus et al., 2016; Banu et al., 2014; Al-Mutairi et al., 2018). These studies also stress the importance of new cut-offs/ critical concentrations determination for the EMB susceptibility test in the MGIT960 system, in order to improve the detection of low-level resistance. In this study we observed 22 strains with mutation at codon 306 of the embB gene that were considered susceptible by the MGIT 960 system, however, one of the limitations of our study was the impossibility of performing the MIC determination in these samples, because of the unavailability of viable strains for the assays. Second-line drug resistance, according to DST, was identified in 8.04%, 8.04% and 4.59% to OFX, CAP and KAN, respectively. The WGS results were able to predict all strains resistant to OFX, and one initially susceptible strain that showed a resistance related mutation, three months later was considered resistant in a new DST. In this way, we can consider a sensitivity and specificity of 100% to OFX prediction. All mutations related to OFX resistance present in the analyzed strains were considered high-confidence mutations, according to previous studies that standardized a method for interpreting the association between mutations and phenotypic drug resistance.(WHO, 2018b; Miotto et al., 2017). The clear correlation between the gyrA gene mutations and FQs resistance, indicates that WGS may play an important role in the resistance detection and be a viable alternative to DST.(Zignol et al., 2018; Miotto et al., 2017). The seven strains resistant to CAP and four strains resistant to KAN by DST, no mutation previously related to resistance to these drugs was found in these strains. Worldwide, related-resistance mutations to KAN e CAP,(Zignol et al., 2018; Miotto et al., 2017), presents sensitivity values ranging from 33% to 100%, depending on the region, demonstrating the geographical variability and diversity of M.tb. Our data show that previously described resistance-associated mutations are still not able to correctly predict the resistance to these drugs in our setting, highlighting the importance of a close epidemiological surveillance of the M.tb strains circulating in each Brazilian State, especially those emerging towards XDR-TB, in order to adjust the TB program responses in control, prevention and treatment more adequate for each scenario. (Dheda et al., 2017). The prevalence of pre-XDR/XDR-TB in this study was of 13.79% among the MDR cases, lower rate than that found in a study conducted in another Brazilian state.(Gallo et al., 2018). The XDR-TB rate, considering a patient who initially had a DST sensitive to OFX, but that a few months later a new DST presented resistance and was thus considered resistant, was 4.59%, similar to that found in other BRICS countries such as South Africa and India.(WHO, 2018). On the other hand, we found 9.19% of pre-XDR-TB strains, which also poses a greater risk of treatment failure and progression to XDR-TB. Although the XDR-TB rate herein reported is below the global XDR-TB rate, (WHO, 2018), there is potential for the emergence of additional XDRTB strains, and we therefore stress the importance of MDR-TB surveillance in the region, as well as the adoption of directed measures to halt the ongoing transmission that is here reported.

Our results showed the scenario of pre-XDR and XDR-TB in the largest state of southern Brazil region, as well as the potential of using WGS to predict resistance to anti-TB drugs and the dynamics of M.tb transmission. The high rate of MDR-TB strains recent transmission in our study show a worry reality in the Rio Grande do Sul State, and demonstrate the need for measures to stop the high MDR-TB strains transmission in this state. In addition, our findings also showed that a better understanding of the molecular basis of drug resistance in this region, especially aminoglycosides, is necessary to enable the use of molecular methods for the screening of pre-XDR and XDR-TB cases among MDR-TB patients, as recommended by WHO. Authors contributions Maria Lucia Rosa Rossetti, Pedro Eduardo Almeida da Silva and Elis Regina Dalla Costa designed the study, reviewed and edited the manuscript, Maria Laura and Regina Bones Barcellos performed the cultivation and DNA isolation, Ana Júlia Reis performed DST testing and analyzed the data, Sun Hee Schiefelbein performed MIRU-VNTR and analyzed the data, Richard Steiner Salvato performed the analyses of WGS data and draft the first version of the manuscript, Gisela Unis and Claudia Fontoura Dias conducted the clinical study, João Perdigão contributed to the phylogenetic analyses and drafted the article providing critical intellectual content, Afrânio Lineu Kritski, Miguel Viveiros and Isabel Portugal helped in the interpretation of data, providing critical intellectual content and contributing to drafting the manuscript. All authors provided key edits to the manuscript, read and approved the final version of the manuscript. Funding This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES), (grant number: 001), and supported by Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul (FAPERGS), (grant number: FAPERGS/ MS/CNPq/SESRS n. 03/2017 – PPSUS). Acknowledgements We are grateful to TGen, C-Path and ReSeqTB for supporting whole genome sequencing and to Brazilian Network of Tuberculosis Research (REDE-TB) for enabling this partnership. The authors would like to thank Centro de Desenvolvimento Científico e Tecnológico (CDCT)/ CEVS/SES/RS for the support and infrastructure. Appendix A. Supplementary data Supplementary data associated with this article can be found in the online version, at https://doi.org/10.1016/j.meegid.2019.104127. These data include the Google maps of the most important areas described in this article. References Allix-Béguec, C., Harmsen, D., Weniger, T., Supply, P., Niemann, S., 2008. Evaluation and strategy for use of MIRU-VNTRplus, a multifunctional database for online analysis of genotyping data and phylogenetic identification of mycobacterium tuberculosis complex isolates. J. Clin. Microbiol. 46, 2692–2699. Allix-Béguec, C., Arandjelovic, I., Bi, L., Beckert, P., Bonnet, M., Bradley, P., et al., 2018. Prediction of susceptibility to first-line tuberculosis drugs by dna sequencing. N. Engl. J. Med. 379, 1403–1415. Al-Mutairi, N.M., Ahmad, S., Mokaddas, E., 2018. Molecular screening versus phenotypic susceptibility testing of multidrug-resistant Mycobacterium tuberculosis isolates for streptomycin and ethambutol. Microb. Drug Resist. 24, 923–931. Available at. https://www.liebertpub.com/doi/10.1089/mdr.2017.0294 (Accessed June 24, 2019).

7

Infection, Genetics and Evolution 78 (2020) 104127

R.S. Salvato, et al.

Miotto, P., Tessema, B., Tagliani, E., et al., 2017. A standardised method for interpreting the association between mutations and phenotypic drug resistance in mycobacterium tuberculosis. Eur. Respir. J. 50 Available at. https://doi.org/10.1183/13993003. 01354-2017. Palomino, J.C., 2002. My Personal/Work/Leisure/Learning Environment, Ed-Techie. Http://NogoodreasonTypepadCoUk/No_Good_Reason/2007/12/My-PersonalWorHtml. 46. pp. 2720–2722. Perdigão, J., Silva, C., Diniz, J., et al., 2019. Clonal expansion across the seas as seen through CPLP-TB database: a joint effort in cataloguing mycobacterium tuberculosis genetic diversity in Portuguese-speaking countries. Infect. Genet. Evol. 72, 44–58. Peres, R.L., Vinhas, S.A., Ribeiro, F.K.C., et al., 2018. Risk factors associated with cluster size of mycobacterium tuberculosis (Mtb) of different RFLP lineages in Brazil. BMC Infect. Dis. 18, 1–10. Roetzer, A., Diel, R., Kohl, T.A., et al., 2013. Whole genome sequencing versus traditional genotyping for investigation of a mycobacterium tuberculosis outbreak: a longitudinal molecular epidemiological study. PLoS Med. 10. Rosłanowski, A., Shelah, S., Spinas, O., 2012. Nonproper products. Bull. Lond. Math. Soc. 44, 299–310. Available at: https://www.thelancet.com/journals/laninf/article/ PIIS1473-3099(12)70277-3/fulltext. Sacchi, F.P.C., Praca, R.M., Tatara, M.B., et al., 2015. Prisons as reservoir for community transmission of tuberculosis, Brazil. Emerg. Infect. Dis. 21, 452–455. Salvato RS, Schiefelbein S, Barcellos RB, et al. Molecular characterisation of multidrugresistant Mycobacterium tuberculosis isolates from a high-burden tuberculosis state in Brazil. Epidemiol. Infect. 2019; 147: e216. (Available at): https://www.cambridge. org/core/product/identifier/S0950268819001006/type/journal_article. Accessed June 24, 2019. Stamatakis, A., 2014. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30, 1312–1313. Stone, N.E., Sidak-Loftis, L.C., Sahl, J.W., et al., 2016. More than 50% of clostridium difficile isolates from pet dogs in Flagstaff, USA, carry toxigenic genotypes. PLoS One 11, 1–21. Supply, P., Mazars, E., Lesjean, S., Vincent, V., Gicquel, B., Locht, C., 2000. Variable human minisatellite-like regions in the mycobacterium tuberculosis genome. Mol. Microbiol. 36, 762–771. Tarasov, A., Vilella, A.J., Cuppen, E., Nijman, I.J., Prins, P., 2015. Genome analysis Sambamba: Fast Processing of NGS Alignment Formats. 31. pp. 2032–2034. Van Embden, J.D.A., Cave, M.D., Crawford, J.T., et al., 1993. Strain identification of mycobacterium tuberculosis by DNA fingerprinting: recommendations for a standardized methodology. J. Clin. Microbiol. 31, 406–409. Walker, T.M., Lalor, M.K., Broda, A., et al., 2014. Assessment of mycobacterium tuberculosis transmission in Oxfordshire, UK, 2007-12, with whole pathogen genome sequences: an observational study. Lancet Respir. Med. 2, 285–292. WHO, 2018a. Global Tuberculosis Report. WHO, 2018b. The use of next-generation sequencing technologies for the detection of mutations associated with drug resistance in Mycobacterium tuberculosis complex: technical guide. Xia, E., Teo, Y.-Y., RT-H, O., 2016. SpoTyping: fast and accurate in silico mycobacterium spoligotyping from sequence reads. Genome Med 8, 19 Available at. http:// genomemedicine.com/content/8/1/19. Yakrus MA, Driscoll J, McAlister A, et al. Molecular and growth-based drug susceptibility testing of Mycobacterium tuberculosis complex for ethambutol resistance in the United States. Tuberc Res Treat 2016; 2016: 1–5. Available at: http://www.hindawi.com/ journals/trt/2016/3404860/. (Accessed June 24, 2019). Zignol, M., Cabibbe, A.M., Dean, A.S., et al., 2018. Genetic sequencing for surveillance of drug resistance in tuberculosis in highly endemic countries: a multi-country population-based surveillance study. Lancet Infect. Dis. 18, 675–683.

Banu, S., Rahman, S.M.M., Khan, M.S.R., et al., 2014. Discordance across several methods for drug susceptibility testing of drug-resistant mycobacterium tuberculosis isolates in a single laboratory. J. Clin. Microbiol. 52, 156–163. Available at. http://www. ncbi.nlm.nih.gov/pubmed/24172155 (Accessed June 24, 2019). Bastos, M.L., Lan, Z., Menzies, D., 2017. An updated systematic review and meta-analysis for treatment of multidrug-resistant tuberculosis. Eur. Respir. J. 49 Available at. https://doi.org/10.1183/13993003.00803-2016. Bjorn-Mortensen, K., Soborg, B., Koch, A., et al., 2016. Tracing mycobacterium tuberculosis transmission by whole genome sequencing in a high incidence setting: a retrospective population-based study in East Greenland. Sci. Rep. 6, 1–8. (Available at). https://doi.org/10.1038/srep33180. Bolger, A.M., Lohse, M., analysis, U.B.G., 2014. Trimmomatic: A Flexible Trimmer for Illumina Sequence Data. 30. pp. 2114–2120. Brasil. Ministério da Saúde, 2018. Implantação do Plano Nacional pelo Fim da Tuberculose como Problema de Saúde Pública no Brasil: primeiros passos rumo ao alcance das metas. Bol Epidemiológico 49 (18) 11. (Available at). http:// portalarquivos2.saude.gov.br/images/pdf/2018/marco/26/2018-009.pdf. Cingolani, P., Patel, V.M., Coon, M., et al., 2012. Using Drosophila melanogaster as a model for genotoxic chemical mutational studies with a new program, SnpSift. Front. Genet. 3, 35. Coll, F., McNerney, R., Guerra-Assunção, J.A., et al., 2014. A robust SNP barcode for typing mycobacterium tuberculosis complex strains. Nat. Commun. 5, 4–8. Coll, F., Mcnerney, R., Preston, M.D., et al., 2015. Rapid determination of anti-tuberculosis drug resistance from whole-genome sequences. Genome Med 7, 1–10. Cruz, L.M., Garcı, F., Okonechnikov, K., et al., 2012. Qualimap: Evaluating NextGeneration Sequencing Alignment Data. 28. pp. 2678–2679. De Beer, J.L., Kremer, K., Ködmön, C., et al., 2012. First worldwide proficiency study on variable-number tandem-repeat typing of mycobacterium tuberculosis complex strains. J. Clin. Microbiol. 50, 662–669. Dheda, K., Gumbo, T., Maartens, G., et al., 2017. The epidemiology, pathogenesis, transmission, diagnosis, and management of multidrug-resistant, extensively drugresistant, and incurable tuberculosis. Lancet Respir. Med. 5, 291–360. Available at. http://www.ncbi.nlm.nih.gov/pubmed/28344011 (Accessed July 9, 2019). Esteves, L.S., Dalla Costa, E.R., Vasconcellos, S.E.G., et al., 2018. Genetic diversity of mycobacterium tuberculosis isoniazid monoresistant and multidrug-resistant in Rio Grande do Sul, a tuberculosis high-burden state in Brazil. Tuberculosis 110, 36–43. Francisco, A.P., Bugalho, M., Ramirez, M., Carriço, J.A., 2009. Global optimal eBURST analysis of multilocus typing data using a graphic matroid approach. BMC Bioinformatics 10, 152. Available at. http://www.biomedcentral.com/1471-2105/ 10/152 (Accessed July 9, 2019). Gallo, J.F., Pinhata, J.M.W., Simonsen, V., Galesi, V.M.N., Ferrazoli, L., Oliveira, R.S., 2018. Prevalence, associated factors, outcomes and transmission of extensively drugresistant tuberculosis among multidrug-resistant tuberculosis patients in São Paulo, Brazil: a cross-sectional study. Clin. Microbiol. Infect. 24, 889–895. Available at. https://doi.org/10.1016/j.cmi.2017.11.015. Li, H., Handsaker, B., Wysoker, A., et al., 2009. The Sequence Alignment / Map format and SAMtools. 25. pp. 2078–2079. Macedo, L.R., Maciel, E.L.N., Struchiner, C.J., Macedo, L.R., Maciel, E.L.N., Struchiner, C.J., 2017. Tuberculose na população privada de Liberdade do Brasil, 2007-2013*. Epidemiol e Serviços Saúde 26, 783–794. Available at: http://revista.iec.gov.br/ template_doi_ess.php?doi=10.5123/S1679-49742017000400783&scielo=S223796222017000400783. Martin, A., Palomino, J.C., 2009. Resazurin Microtiter Assay (REMA): Procedure Manual. Meehan, C.J., Goig, G.A., Kohl, T.A., et al., 2019. Whole genome sequencing of Mycobacterium tuberculosis: current standards and open issues. Nat. Rev. Microbiol. 17 (9), 533–545. 1. Available at. http://www.nature.com/articles/s41579-019-0214-5 (Accessed June 24, 2019).

8