Mycobacterium tuberculosis evolution from monoresistance to pre-extensive drug resistance during a prolonged household outbreak

Abstract

Whole genome sequencing (WGS) is sensitive tool for the analysis of tuberculosis transmission and drug-resistance. We used WGS to analyze the Mycobacterium tuberculosis evolution from isoniazid monoresistance to MDR/preXDR during a prolonged household outbreak. The outbreak started with a isoniazid resistant strain (katG S315T mutation) and evolve in two cases to pre-XDR phenotype (with mutations in katG, rpoB, embB, pncA and gyrA genes). Based on WGS data and epidemiological interview we proposed a possible chain of transmission an evolution of the strains.

Similar intra-patient and inter-patient acquisition of variability was observed, making difficult to distinguish reinfection or reactivation. Analysis of WGS data together with epidemiological clinical history are discussed in order to distinguish between prolonged infections or transition from latency to reactivation. Classical interview and clinical history taking should be consider to fully understanding WGS data. With a still low incidence of TB cases, Uruguay could use universal WGS of all isolates to reduce time of diagnosis, detect outbreaks and perform public actions to reduce TB incidence.

1. Introduction

Tuberculosis (TB) stills remains a major public health in low and high burden countries. Whole genome sequencing (WGS) strategies have the potential to revolutionize TB diagnostics and outbreak investigation, plus assist in understanding Mycobacterium tuberculosis (MTB) evolution and pathogenicity [1].

Several studies [2], [3], [4] have demonstrated the utility of WGS for drug-resistance detection and molecular epidemiology studies in high and low incidence countries allowing public health interventions, especially in short-time transmission chains and to distinguish between relapse or reactivation and reinfection [3], [4], [5].

Uruguay has a low to middle TB incidence (29/100,000 inhabitants) with low MDR prevalence (<1% of detected cases) [5]. All the reported cases are cultured in the National Reference Laboratory and a biobank of DNA is available since 2005. This makes the country a good candidate to perform WGS as a routine public health alternative.

As a proof of concept, we performed retrospective WGS analysis of the first preXDR-case in Uruguay, tracing the epidemiological contacts links, and resolving the mutation acquisition from isoniazid mono-resistance to pre-XDR in a 10 year-long household outbreak.

2. Case reports

A father (Patient A, 53 years old) and two daughters (Patient B, 17 years old and patient C, 16 years old at the moment of diagnosis) were diagnosed with pulmonary tuberculosis in 2004. Chest X-ray indicated unilateral (the father) and bilateral (the two daughters) advanced cavitary lesions. Although the three patients start the classic treatment at that time (INH/RIF/PZA), initial resistance to INH was detected by phenotypic susceptibility testing. Patient A suffered a relapse in 2006, with an Isoniazid, Rifampicin and Fluoroquinolon resistant strain. He was then treated with Cycloserine, Kanamycin and p-Aminosalicylic acid. One year later he had persistent disease and grew a pre-XDR isolate with resistance to INH, RIF, PZA and Fluoroquinolones. Patient A experienced treatment failure and died in 2009. Patient B has also suffered a relapse of TB in 2007, with an MDR profile. Patient C, showed clinical cure until 2014, when she presented a relapse of the disease with a pre-XDR strain. Patients B and C were treated with Cycloserine, Kanamycin, Ethionamide, Levofloxacin and Pyrazinamide drugs and were cured.

The last isolate, patient D (female, 32 years old at the moment of diagnosis, not traced as contact in the epidemiological interviews), was diagnosed in 2010 with an MDR strain and was cured using the same second-line drug treatment. Patient D was associated to the outbreak inspecting relationship between MDR strains. All data are summarized in Table 1.

Table 1.

Clinical and Treatment Outcomes of Pulmonary Tuberculosis Patients.

Patient (relation). Age at diagnosis	Year of Diagnosis	Initial Diagnosis	X-ray Findings	Initial Treatment	Initial Resistance	Relapse	Resistance Profile During Relapse	Treatment During Relapse	Outcome
A (Father)53 years old	2004	Pulmonary Tuberculosis	Unilateral advanced cavitary lesions	INH/RIF/PZA	INH	2006	Isoniazid, Rifampicin, Fluoroquinolones (pre-XDR in 2007)	Cycloserine, Kanamycin, p-Aminosalicylic acid	Died in 2009
B (Daughter)17 years old	2004	Pulmonary Tuberculosis	Bilateral advanced cavitary lesions	INH/RIF/PZA	INH	2007	MDR	Cycloserine, Kanamycin, Ethionamide, Levofloxacin, Pyrazinamide	Cured
C (Daughter)16 years old	2004	Pulmonary Tuberculosis	Bilateral advanced cavitary lesions	INH/RIF/PZA	INH	2014	pre-XDR	Cycloserine, Kanamycin, Ethionamide, Levofloxacin, Pyrazinamide	Cured
D (Unrelated Female)32 years old	2010	Pulmonary Tuberculosis	Not specified	Cycloserine, Kanamycin, Ethionamide, Levofloxacin, Pyrazinamide	MDR	N/A	N/A	N/A	Cured

3. Material and methods

DNA Extraction. A loopfull of mycobacterial growth on LJ was suspended in 1 ml of distilled water, heated 30 min at 90 °C, frozen and thawed three times, and centrifuged at 13,500 RCF 10 min.

Sequencing. From DNA of clinical isolated strains we performed WGS sequencing with Illumina MiSeq technology (Illumina, USA). Briefly, libraries were made with Nextera XT Library Prep (Illumina, USA). Multiplexed libraries were paired-end sequenced (2x150 cycles).

3.1. Informatics analysis

To perform WGS analysis we use two online software to identify mutation resistance profile, and SNP profiling to typing the strains: TGS-TB [6] and Phyresse [7]. Phylogenetic tree was obtained in TGS-TB and visualized using MEGA software. For Single Nucleotide Variation we filter total SNP detected by TGS-TB and retain only informative position (>=20 Phred-score and > 5X coverage. Mixed bases are considered using a threshold of 5 % of reads showing the alternative nucleotide.

4. Results

We sequenced the whole genomes of eight isolates of the four patients, obtaining > 50X coverage (Supplementary File 1). The molecular WGS typing performed by Phyresse [7] and TGS-TB [6] platforms confirmed the clonality of the outbreak (Fig. 1A); all strain belonged to the S family (Euro-American Linage 4, Fig. 1 A).

Fig. 1.

A. Phylogenetic tree. Complete phylogenetic tree and subtree of the outbreak. Gray Box: resistant mutations observed in each strain. B. Proposed chain of transmission based on SNP distance analysis and multiple initial strain infection hypothesis. Time line showing the proposed events of the outbreak. In Red the most probable events of transmission. Vertical lines indicates number of SNP between isolates (in red: proposed SNP, in gray: SNPs necessary to explain distance between isolates considering initial infection with only one strain).

The resistant mutations profile observed in the samples suggested a first infection with an isoniazid-resistant strain (Patients A, B and C in 2004) (Fig. 1A). Patient A, with poor adherence to treatment was developed a pre-XDR profile in 2007 (Fig. 1A) and died in 2009 after treatment failure. In 2007 patient B, after two negative cultures in 2005 and 2006, showed a reactivation of the disease with an MDR profile (Fig. 1A).

In the case of Patient C, after clinical cure in 2004, the patient presented a reactivation of the disease in 2014, with a preXDR strain (same mutations profile observed in patient A in 2007) (Fig. 1A). Finally, the patient D, showed a mutation profile identical to patient B_2007 (Fig. 1A). The patient D was not detected as a part of the tracing contacts in the epidemiological interview but was included in the outbreak after the observation of the mutational profile considering that these cases represented the first observation of MDR profiles in the country. Although was not considered contact in the interview, the patient lives in the same locality (small town in the middle of the country with 13,000 inhabs.). The proposed probable chain of transmission was reflected in Fig. 1B based on phylogenetic tree, SNV information and resistant mutations profiles.

5. Discussion

In this work we investigated the first two cases of pre-XDR TB in Uruguay. Using WGS we can determine the evolution of the outbreak strain from mono-antibiotic resistance to pre-XDR profile. In a recent review, Nikolayevsky et al. [8] show that a genetic distance between 0 and 5 SNP could be used to indicate a recent transmission episode; strain with > 12 SNPs were not considered in direct transmission. These epidemiological cut-offs were validated in low-incidence and low-drug-resistance settings [8]. According to these parameters, we could demonstrate an outbreak of an initial mono-resistance to INH strain in 2004 (Patient A, B and C with 0–1 SNP apart). However, this threshold could not be used to differentiate re-infection or relapse. In the studied cases, we observed minor populations in all the isolates, suggesting possible multiple strain infection (Supplementary file 1). Based on the phylogenetic tree and the SNV analysis (Fig. 1A and B) we proposed a possible transmission chain (Fig. 1B).

Patient A’s original and relapse isolates were five SNPs apart (Fig. 1B), suggesting an evolution from mono-resistance to MDR/preXDR TB (Fig. 1A). When analyzing the evolution of these isolates, two possible explanations could be considered:

• 1.The acquisition of MDR resistance following inadequate treatment, where the initial INH-resistant strain developed additional resistance due to ineffective therapy.
• 2.A multiple strain infection, where Patient A was initially infected with both and INH-resistant strain and a low population of fluoroquinolone-resistant (FQ) strain. Over time, the FQ-resistant strain could have dominated the infection after INH-resistant strain was eliminated by RIF and PZA treatment. These scenarios could explain the observed SNP variation. Comparing only the first isolate of patient A in 2004 with the last isolate in 2007 reveals a distance of 6 SNP, reaching the threshold to consider two strain not epidemiological related. However, examining the SNPs observed in 2006, particularly at positions 4,269,308 (a reversion event never observed in M. tuberculosis) and 7582 (a mutation in the same position, very rare), the presence of multiple strains provides a more plausible explanation of the outbreak. To accurately distinguish between relapse and re-infection, the complete series of events should be analyzed by integrating WGS data with classical epidemiological interviews. Although, the intra-patient variability was previously observed in prolonged TB cases [9], for Patient A, the most likely explanation is a multiple strain infection rather than a host microevolution pattern, considering the three sequenced genomes (Fig. 1B).

In the case of Patient B, according to the phylogenetic tree (Fig. 1A) and the observed SNV, a re-infection with the prevalent strain in patient A isolated in 2006 was the proposed chain of transmission, instead a relapse with the original strain (Fig. 1B). However, we cannot entirely exclude the possibility that the 2007 relapse was due to the reactivation of a minor population of the original strain present in 2004.

In the case of Patient C, who had good adherence to treatment and achieved clinical cure after the first episode, the first and second isolates, taken 10 years apart, show an eight SNP difference (representing 0.8 SNPs per genome per year). This mutation rate falls within the expected range for normal MTB evolution and would typically indicate a relapse of the infection. However, when considering the phylogenetic tree (Fig. 1A) and the specific SNVs observed, it is more likely that Patient C experienced a re-infection with the strain from Patient A, rather than a relapse. This conclusion is supported by the close genetic relationship between the 2014 isolate from Patient C and the strain from Patient A, which shows only a one SNV difference between them, suggesting a transmission event between 2007 and 2009.

Finally, Patient D, showed a profile of mutations indicating an infection with the strain of patient B in 2007 (0 SNVs apart).

Although effort to standardize WGS data interpretation [1], [8], [10] were made and online analysis tools were developed [6], [7], the cases reported here proved the need of more accurate parameters to distinguish between relapse and reinfection, and infections with multiple MTB populations (Supplementary file 1). The use of WGS in Uruguay could be a great opportunity to reduce time of diagnosis, detect MDR/XDR-TB outbreaks and perform public actions to attempt the goal of STOP TB initiative. WGS data should be analyzed together with classical epidemiological interview to better understand outbreaks and transmission profiles.

6. Financial support

This work was supported by PEDECIBA, Uruguay and FOCEM − Fondo para la Convergencia Estructural del Mercosur (COF 03/11). G.G. and C.R. are researchers from PEDECIBA, and the Sistema Nacional de Investigadores (ANII), Montevideo Uruguay. Financial Support: This work was supported by PEDECIBA, Uruguay and FOCEM − Fondo para la Convergencia Estructural del Mercosur (COF 03/11).

CRediT authorship contribution statement

G. Greif: Writing – review & editing, Writing – original draft, Investigation, Funding acquisition, Formal analysis. C. Coithino: Writing – review & editing, Investigation. M.N. Bentancor: Methodology, Investigation, Conceptualization. C. Robello: Writing – review & editing, Resources, Investigation, Funding acquisition.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Gonzalo Greif reports financial support and equipment, drugs, or supplies were provided by Fondo para la Convergencia Estructural del Mercosur. Gonzalo Greif reports financial support and equipment, drugs, or supplies were provided by Programme for the Development of Basic Sciences. Gonzalo Greif reports financial support was provided by ANII. Carlos Robello reports financial support was provided by ANII. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: