論文の公開元へStreptomycin and ethambutol resistance associated mutations of multidrug-resistant Mycobacterium tuberculosis clinical isolates in Lusaka, Zambia
概要
Tuberculosis (TB) is an infectious disease caused by the bacillus Mycobacterium tuberculosis (Mtb). It has remained a health concern since its declaration as a public health problem in 1993 [1]. Although global incidences of TB have been reducing in past years, in 2020, nearly 9.9 million people developed TB with most cases being reported in the WHO regions of South-East Asia (43%), Africa (25%), and Western Pacific (18%) (figure 1) [2]. It is among the leading causes of mortality worldwide. An estimated 1.3 million people died from TB in 2020 alone, an increase from 1.2 million reported in 2019 among HIV-negative people. In HIV-positive patients, mortality increased from 209 000 in 2019 to 214 000 in 2020 [2].
Mtb the causative agent of TB, is a gram-positive bacillus belonging to the phylum Actinobacteria. Members of this phylum have a high G+C content in their genome, exhibit pleomorphic morphologies, and can be found freely living in wide-ranging ecological environments including soil, water, plants, animals, and human [3]. The Mycobacterium genera of Actinobacteria constitute important human pathogens. Members of this genus are characterized by a complex cell wall and a waxy cell envelope. The cell wall structure can be divided into 3 major components: a peptidoglycan, arabinogalactan, and mycolic acid layers [4]. The peptidoglycan layer functions to provide shape and rigidity to the cell, protecting against turgor pressure, and anchoring other outer cell components [3]. Arabinogalactan a polysaccharide, links mycolic acids to the peptidoglycan layer and is important for the cell growth [5]. A distinctive feature of the members of the genus Mycobacterium is the presence of mycolic acids, unique fatty acids that make up the mycobacteria cell envelope and responsible for the characteristic acid fastness. Mycolic acids are essential for the cell viability and virulence [5]. The mycolic acids make an intercalating complex with other lipids to form a hydrophobic outer membrane highly impermeable to many compounds and enabling it to survive in stressful environmental conditions such as drug presence [6]. Mtb and Mycobacterium leprae causing tuberculosis and leprosy, respectively, are the most important obligate human pathogens of the genus Mycobacterium.
Infection by Mtb is by inhalation of droplet nuclei containing live bacteria. It primarily causes pulmonary TB, the most infectious form of TB. Upon inhalation, the bacteria are phagocytosed by alveolar macrophages, forming a phagosome where they persist by escaping the macrophage mechanism to clear intracellular pathogens such as preventing phagosome and lysosome fusion and resistance to reactive nitrogen species [4][7].
Infection by Mtb has 3 possible outcomes including 1) clearance by host immune system, 2) containment as latent infection, and 3) progression to active disease. In latent TB infection (LTBI), the host immune response restricts Mtb growth and maintains the bacteria in a state of dormancy surrounded by monocytes, dendritic cells, lymphocytes, and fibroblasts [8]. This lesion containing bacteria also known as a granuloma or tubercle, is the hallmark of TB. The granuloma prevents bacteria from spreading throughout the tissue and other body organs, and live bacteria can persist in the lesion for decades. An estimated 21% to 25% of the global population has LTBI [9]. People with LTBI are asymptomatic but, when the immune system weakens, they can develop symptomatic TB, thus LTBI presents a large reservoir for TB disease. Approximately 10% of people with LTBI have a risk of developing active TB at some point in their life [10]. Active TB is characterized by persistent cough with bloody phlegm, weight loss, night sweat, fatigue, and fever. For clarity, in this thesis, active TB will be referred to as TB.
The immune system suppression predisposes to TB development by compromising the integrity of the granuloma causing it to rupture and release live bacteria. Factors such as diabetes, malnutrition, smoking, alcohol, age, immunosuppressive therapy, cancer, and HIV contribute to immune system suppression and increase the risk of developing TB [11–13]. In addition to increasing the risk of TB reactivation from LTBI, HIV can also increase progression to TB disease after initial infection or reinfection [14].
Recent studies have also confirmed the role of host genetic factors to susceptibility to TB infection as well as the development of disease. Some studies have pointed to a genetic variant on chromosome 11p13 associated with resistance to TB. A single nucleotide polymorphism (SNP) rs2057178 near WT1 on this chromosome had a protective association against TB in study populations from Ghana (OR 0.77, p value = 2.63x10-9), Gambia (OR 0.80, p value = 4.87x10-4), Russia (OR 0.91, p value = 0.02), South Africa (OR 0.62, p value = 2.7x10-6), and Morocco (OR 0.78, p value = 0.43) [15–17]. A genome-wide association study conducted on the Icelandic population found 2 HLA class II variants that affected the risk of pulmonary TB. The SNP rs557011 [T] appeared to increase the odds of developing the disease (OR 1.25, p value = 5.8x10-12) while rs9271378[G] protected against disease development (OR 1.25, p value = 5.8x10-12), possibly by affecting antigen presentation to T cells [18]. Host genetic susceptibility to TB was also exemplified in the Lübeck disaster in which 251 children were accidentally inoculated with Bacille Calmette-Guérin (BCG) vaccine contaminated with variable concentrations of pathogenic Mtb. Out of the 251 inoculated children, 228 developed different forms of TB. Seventy-seven deaths were recorded within a year of exposure, and 72 of these were attributed to TB. Remarkably, 17 (6.8%) children had no clinical signs of TB while 68% of children who had developed symptomatic TB, spontaneously cured stressing the role of host genetics in TB disease resistance and severity [19]. The observed variability in TB outcome could also be attributed to different dosage of Mtb administered to the children suggesting that bacteria dosages at exposure also influences the outcome of infection. Undoubtedly, host genetic variability and exposure dosage can influence the risk of developing TB.
Mtb has intrinsic resistance to many antibiotics due to the highly impermeable lipid-rich cell envelope and presence of drug inactivating enzymes such as beta-lactamases. Nonetheless, TB is a treatable disease. The recommended first-line anti-TB drug combination of rifampicin (RIF), isoniazid (INH), ethambutol (EMB), and pyrazinamide (PZA) for the treatment of drug-susceptible TB has approximately 85% treatment success [20]. The emergence of drug resistance to these drugs is a public health concern globally, particularly rifampicin resistance (RR) and multi-drug resistant TB (MDR-TB) which is resistance to the two most potent anti-TB drugs RIF and INH [21]. For both forms of drug resistance (RR and MDR), second-line drugs are used for treatment. History of previous TB treatment, poor quality drugs, inconsistent supply of TB drugs, delayed diagnosis, accessibility of health care facilities, and patient-related factors such as non-adherence to treatment, social- economic factors increase the likelihood of drug resistance emergence [18]. Human immunodeficiency virus (HIV) co-infection has also been documented to increase the likelihood of developing drug resistance due to drug malabsorption, leading to suboptimal drug plasma concentration [22]. The emergence of MDR/RR-TB is problematic compared to drug-susceptible TB because treatment of this form of TB is highly costly, uses toxic drugs producing adverse side effects; in addition, the patient has to endure a high pill burden for a prolonged treatment period [23]. Moreover, MDR-TB treatment often has poor outcome with only 57% global success rate [20]. MDR/RR-TB emergence is thus a threat to the global efforts to end TB.
Drug resistance in Mtb is mediated by the acquisition of chromosomal mutations in genes that encode proteins targeted by the drugs [6]. These mutations can cause drug target structural alteration such as RNA polymerase mutations that cause RIF resistance [24], upregulation of drug target proteins as is the case with NADH-dependent enoyl-acyl carrier protein mutations conferring resistance to INH [25], and overexpression of efflux pumps such as mutations in Rv0678 leading to upregulation of the MmpL5 efflux pump and causing resistance to bedaquiline and clofazimine [26]. Some drugs are prodrugs and require enzymatic activation to exert their antimicrobial properties. Mutations in genes encoding enzymes responsible for activating the drugs abrogate drug activation. An example of prodrugs is PZA and INH [6].
It has been recognized that an interplay of bacterial determinants influences the evolution of drug resistance-conferring mutations and their maintenance. The biological cost of mutations on the organism’s fitness is an important predictor of the acquisition of drug resistant confering mutations. The degree of impact on fitness depends on the specific mutation, the environment, and genetic background of the bacteria. Forexample, among RIF resistance confering mutations, rpoB S450L is a least fitness cost in Mtb [27]. Addittionally, the fitness cost of a mutation also differs in different growth conditions [28]. Clearly, drug resistance mutations may affect the fitness of an organism.
In most experiments, the inference of in vivo fitness cost of mutations on Mtb is done by in vitro comparison of the growth rate of mutant and susceptible strains. This method fails to capture other factors that may have an effect on mutation acquisition such as the host immune system and genetics, nutrient availability, and co-morbidities [28][29]. An alternative way to measure fitness is by quantifying resistant variant frequency in a population which correlates well with in vivo fitness.
The most frequently observed mutations in clinical isolates are considered to have reduced in vivo fitness cost. For example, a genome-wide analysis found that the rpoB mutation S450L previously reported by in vitro experiments as the least costly mutation was the most frequent mutation among 6,465 MDR-Mtb clinical isolates at 64.2% [27,30]. An allelic exchange experiment revealed that rpsL mutation K42R in Mycobacterium smegmatis had no fitness cost and the corresponding mutation in Mtb K43R accounted for 42.2% of STR resistance in MDR-TB clinical isolates [30,31]. Mutations with higher fitness costs, however, may also be found in reasonable frequency due to the occurrence of compensatory mutations that restores fitness to the bacteria [6]. The mutations with high fitness costs can thus persist in the population and transmit easily causing outbreaks of MDR-TB. For example, an outbreak of MDR-TB in South Africa and Eswatini was caused by strains having a rare rpoB mutation I491F whose fitness cost in some strains was alleviated by the acquisition of a rpoC mutation E1033A leading to increased transmissibility of the strains [32]. This interaction between a drug resistance-conferring mutation and a compensatory mutation is an example of epistasis which is defined as a non-additive interaction between two genes/alleles that produces an effect on a phenotype such as fitness [6].
The genomic differences of Mtb strains is another important bacterial determinant of drug resistance as well as several other observed phenotypes of the organism such as transmissibility and virulence. The strain’s genomic SNPs, insertions and deletions (indels), large genomic deletions or duplications, and mobile and repeat elements produce genetic diversity which characterize genetic background of a strain [6]. The genetic background of a strain may influence the impact of a drug resistance mutation on fitness. Variable fitness was observed among 5 clinical isolates having the same S450L amino acid substitution in rpoB, underlining the influence strain genetic background may have on an organism’s fitness upon acquisition of resistance-conferring mutations [27]. Moreover, the level of drug resistance conferred by a particular mutation can differ depending on the strain’s genetic background [33].
Based on genomic differences, Mtb can be divided into Lineages 1 to 4, and Lineage 7, whereas Lineages 5 and 6 are Mycobacterium africanum [34]. These Lineages show differences in geographic distribution and varied associations to drug resistance. Lineage 1 also known as Indo-Oceanic or EAI, is dominant around the Indian Ocean, while Lineage 3 known as Central Asian Strain (CAS) is predominantly found in East Africa and Central and South-Asia, and Lineage 7 is found exclusively in Ethiopia [35][36]. On the other hand, Lineages 2 (also referred to as East-Asian or Beijing strains) is gradually expanding globally while Lineage 4 (referred to as Euro-American) is globally distributed. The Beijing genotype of Lineage 2 has been shown to have a high propensity to acquire drug resistance due to a higher mutation rate, and based on isolate clustering, has been associated with increased transmission of drug-resistant strains compared to Lineage 1 [37–39]. Genetic background can influence the pathway of drug resistance as shown by Lineage 1 having higher odds of acquiring inhA mutations -15C/T for INH resistance compared to other lineages (OR 6.4, p value = 0.002) [33]. In addition, MDR-TB outbreaks have been attributed to different sublineages in different geographical areas. MDR-TB outbreaks in Vietnam and Thailand were attributed to the Beijing genotype [38,39], in South Africa and Eswatini to S genotype of Lineage 4 [32], in South Africa’s KwaZulu Natal region to LAM4 [40], in Argentina to sublineage 4.1.2.1 of Lineage 4 [41], and in India to CAS1_Delhi [42]. Taken together, bacterial determinants such as fitness cost, compensatory mutations, and genomic variability play a key role in driving drug resistance mutation acquisition.
To study Mtb genetic diversity, several molecular methods have been developed, including insertion sequence 6110 (IS6110) restriction fragment length polymorphism (RFLP), spacer oligonucleotide typing (spoligotyping), mycobacterial interspersed repetitive-units-variable number tandem-repeat (MIRU-VNTR) typing, large sequence polymorphisms (LSP), and SNP analysis. The IS6110 RFLP method is based on different copy numbers of the IS6110 and the position of insertions in the genome of Mtb isolates to generate a strain-specific fingerprint [43]. Spoligotyping detects the presence or absence of 43 variable spacers in the direct repeat (DR) locus and can differentiate Mtb to sublineage level such as CAS1_Kili, LAM, T, X, Beijing, etc [44]. Unlike spoligotyping which targets one locus, MIRU-VNTR targets multiple loci containing a variable number of tandem repeats and is thus more discriminatory than spoligotyping [45]. LSP is based on the detection of regions of difference due to deletions in the genome of Mtb strains while SNP analysis identifies strain-specific mutations in isolates compared to the reference strains and both can classify Mtb into Lineages and sublineages [46,47]. Genotyping of Mtb is important in the epidemiological surveillance of TB, outbreak investigation, and the identification of transmission clusters in different geographical settings.
As already discussed, drug resistance mutation frequencies are influenced by bacterial genotypes. The distribution of these genotypes is different depending on geographical area. Therefore, regional setting studies including drug resistance mutations and genotypes are relevant to inform specific control strategies to curb antibiotic resistance and contribute to the development of rapid diagnostic tools.
Zambia has seen a downward trend of TB incidences in the past decade (figure 2), yet still, it is regarded as a high TB burden country [2]. An estimated 59, 000 people became ill with TB in 2020 and only 68% of these were started on treatment [2]. The approximately 30% with delayed treatment initiation can become the source of infection in the population. Factors such as death, untraceable patients due to incorrect or change of address, socioeconomic factors, administrative factors, long diagnostic result turnaround time, and health care provider behavior and attitude; may contribute to delayed treatment initiation [48]. In addition to delayed treatment, approximately 40% of TB cases in Zambia do not have access to microbiological testing [49]. The lack of diagnosis of these TB patients may be due to difficulties to access health care and TB facilities, delay to seek health care, and unsuspected TB [49][50]. The undiagnosed and untreated TB patients perpetuate TB transmission and could contribute to the high TB burden in Zambia.
Several methods are used for confirmation of a TB diagnosis in Zambia including smear microscopy, lateral flow urine lipoarabinomannan assay (LF-LAM), Xpert MTB/RIF assay, line probe assay (LPA), and culture. Smear microscopy involves demonstrating by sputum smear staining the presence of acid-fast bacilli (AFB). This technique is widely used in TB screening in high TB burden countries, but it has moderate sensitivity [51]. The LF-LAM assay uses urine for diagnosis of TB and is used in patients who cannot produce sputum such as HIV patients [52]. Isolation of Mtb by culture is more sensitive than smear microscopy but, in Zambia, it is limited to all retreatment cases, suspected treatment failure, and drug resistance surveillance [53]. The purpose of culturing for these classes of patients is to perform drug susceptibility testing (DST) to guide treatment, however, this method’s turnaround time is several weeks to months. In 2018, the molecular tools Xpert MTB/RIF assay and LPA were formally adopted as a first-line tool for the diagnosis of TB in Zambia as well as for DST to first and second-line drugs, respectively. These tools are more rapid and sensitive compared to culture and have improved drug resistance case detection in the country [51][54].
Zambia has seen an increasing trend of MDR/RR-TB in recent years and as a result, was included on the 2021 updated list of high MDR-TB burden countries in the world [55]. Figure 3 shows a gradual increase of laboratory-confirmed MDR-TB over a period of 12 years in Zambia from 2000 to 2011 [56]. It can be said, therefore, that MDR-TB in Zambia has been silently increasing for the past 2 decades. The adoption of molecular-based tools such as Xpert MTB/RIF and second-line DST tool LPA could be one reason for the observed recent steep rise in the number of laboratory-confirmed new MDR/RR-TB cases [54]. The rollout of Xpert MTB/RIF in Zambia started in 2017, and that year saw a significant increase of laboratory-confirmed MDR/RR-TB from 180 in 2016 to 546 (figure 4).
Despite the improved MDR-TB detection, there are still significant gaps in DST and prompt treatment initiation for MDR/RR-TB in Zambia. In 2018 for example, there were 3100 estimated incidences of MDR/RR-TB. In that year, 627 laboratory-confirmed MDR/RR-TB cases were reported, 506 were started on MDR/RR-TB treatment, and yet still, only 150 were tested for second-line drug resistance (figure 4) [21]. That means a significant proportion of MDR/RR-TB patients remain untested against second-line drugs and are treated empirically without DST.
Further complicating the management of MDR/RR-TB in Zambia, is the fact that only a few drugs used in the treatment of MDR/RR-TB are covered by the adopted rapid molecular tools. Currently, the rapid molecular test LPA is used to detect Rif, INH, fluoroquinolones, and an aminoglycoside amikacin (AMK), while phenotypic DST is used for EMB, streptomycin (STR), clofazimine (CFZ), bedaquiline (BDQ), and linezolid (LZD). Because of the limited DST coverage and the number of drugs tested out of those used in MDR/RR-TB treatment, there may be patients on inadequate treatment. Inadequate treatment is a two-edged sword. On one hand, it leads to patients remaining infectious for long periods of time and thereby becoming the sources of transmission, and on the other hand, predisposes to the emergence and amplification of drug resistance and poor treatment outcome for the patient [57].
Using DNA sequencing and genotyping methods, recent studies have revealed frequent mutations that lead to resistance to INH and RIF in Zambia. Analysis of those mutations revealed clustering of MDR-TB isolates and an association between MDR-TB and CAS1_Kili and LAM1 genotypes [58]. They have thus shown that the rise of MDR-TB cases in Zambia is being driven by the emergence and transmission rather than cross border transmission of MDR-Mtb strains [59][60].
Therefore, more effort is needed to hasten the progress that Zambia has been making to control MDR/RR-TB. To reverse the trend of increasing MDR/RR-TB cases, it is important to recognize that early resistance detection and promptly initiating treatment guided by susceptibility profiles is key. Thus, scaling up coverage of testing is critical. Mutation analysis by DNA Sequencing is proving useful in the management of TB patients as well as in surveillance of drug resistance. It has a short results turnaround time and does not require specialized facilities compared to the conventional phenotypic DST methods which requires Biosafety level 3 and is thus limited to a few facilities in Zambia. Additionally, sequencing can be used to determine resistance patterns to a wide range of drugs, explore clustering, and transmission of Mtb strains. Information about frequency of dug resistant conferring mutations can also inform the development of easier to use and fast diagnostic test methods.
The aim of this thesis was to characterize mutations conferring resistance to STR and EMB, two drugs used in management of MDR-TB and evaluate the utility of DNA sequencing for DST of MDR- TB isolates in Zambia. Additionally, the correlation between drug resistance-conferring mutations and Mtb genotypes was determined using the spoligotyping genotyping method. This method detects the variability in the Direct Repeat Locus. This locus has 36 base pair repeat sequences interrupted by non-repetitive sequences of 34 to 41 base pairs (figure 5) [44].
This thesis has two chapters. Chapter I describes mutations in rpsL, rrs, and gidB genes known to confer resistance to STR and discusses their correlation with the circulating genotypes in Zambia. In chapter II, mutations in embB conferring resistance to EMB are described. Additionally, I discuss the disparity of resistance conferring mutations and phenotypic EMB DST. Using publicly available data on TB profiler website, an interesting phenomenon of possible genotypic background influence on EMB drug resistance trajectory is discussed in chapter II.
Figures
Figure 1: Estimated TB incidence rates, 2020 [2]
Figure 2: Trend in estimated TB incidence rate in Zambia compared with notifications of new and relapse cases, 2000-2020. The blue line shows incidence rates (shaded area shows uncertainty intervals) while the solid black line indicates notification of new and relapse cases for comparison with estimated total incidence rate. The horizontal dashed line show the 2020 milestone of the end TB strategy [2]. (Accessed 5th November 2021)
Figure 3: Number of MDR-TB diagnosed by year from 2000 to 2011 in Zambia [56]
Figure 4: Laboratory confirmed MDR/RR-TB and notified cases in Zambia [21,61–63]
Figure 5: Spoligotyping principle (modified from Kamerbeek et al., 1997) [44]
この論文で使われている画像
