In patients with diabetes mellitus, TB treatment outcomes are poorer. Most parameters, when measured, reflect the slower bacteriological conversion from positivity to negativity and higher risks of disease relapse and mortality, as well as a greater propensity to develop drug-resistant TB. Aside from the well-known immunological dysfunction inherent to patients with diabetes mellitus, oxidative stress is likely a major underlying mechanism adversely impacting their TB treatment outcomes. Mycobacterium tuberculosis persisters, formed as a result of the core dormancy response to stress, possibly play a central role in this hypothesis. This hypothetical model also underscores the paramount importance of programmatic management of TB and diabetes mellitus, in collaboration, to improve the outcomes of patients with both diseases. The validity of these ideas could be further ascertained by laboratory and clinical research.

About 15% of patients with TB globally are estimated to have diabetes mellitus and this proportion will likely increase in the coming decades.1 Diabetes mellitus increases the risk of TB and worsens TB treatment outcomes, including slower bacteriological conversion, lower cure, higher relapse and mortality and possibly escalated drug resistance. Reciprocally, TB also hampers the optimal management of diabetes mellitus.1

It has been shown that the function of neutrophils, macrophages, dendritic cells and natural killer cells, as well as some other components of innate immunity and adaptive/acquired immunity, are significantly compromised by metabolic alterations in diabetes mellitus.2 Thus, immune dysfunction as a result of diabetes may play an important role in allowing the reactivation of TB from endogenous latent infection and increasing the host’s susceptibility to exogenous reinfection.2

In recent years, diseases that involve chronic inflammation and altered metabolism have been shown to be linked pathogenetically to oxidative stress.3 Complications of these diseases often result in significant morbidity and mortality. The higher mortality of TB in patients with diabetes mellitus may also be closely associated with complications of the metabolic disease.4 In diabetes mellitus, endogenous predisposition and exogenous factors may lead to development of inappropriate oxidative stress (Figure 1). Endogenous oxidative stress is likely related to hyperglycaemia and advanced glycation end-products.5 Exogenous oxidative stress factors principally include tobacco smoke, indoor air pollution, chemicals and radiation. We hypothesize that oxidative stress is directly responsible for the poorer TB treatment outcomes in patients with diabetes mellitus, largely through inducing formation of Mycobacterium tuberculosis persisters (Figure 1).

Figure 1

Oxidative stress and TB treatment outcomes in patients with diabetes mellitus.

Figure 1

Oxidative stress and TB treatment outcomes in patients with diabetes mellitus.

Persister phenotype in bacteria is well known as a population phenomenon. The characteristics of bacterial persistence generally include dormancy, multidrug tolerance and stochasticism.6 Environmental cues that induce formation of bacterial persisters are diverse, with antibiotic use, starvation, extreme changes in pH, temperature and oxygen status, as well as host immunity, being commonly listed as stress conditions.7

Recently there have been a number of reports regarding stronger evidence of oxidative stress in stimulating the formation of bacterial persisters. Paraquat-generated oxidative stress was found to induce the SoxRS and OxyR regulons and to markedly increase persister cell formation in Escherichia coli, especially after fluoroquinolone challenge.8 Oxidative stress generated by pyocyanin was shown to increase the production of catalase and superoxide dismutase in Acinetobacter baumannii, together with a significant rise in the number of bacterial persisters.9 Similarly, oxidative stress, through bile administration, was shown to induce persister formation in Salmonella Typhi, after amikacin/carbenicillin exposure.10 Furthermore, the role of the quorum-sensing mechanism(s) related to oxidative stress-induced persister formation was shown.9,10,A. baumannii has one quorum-sensing system, involving the transcriptional regulator AbaR, which forms a complex with the lactone encoded by the auto-inducer synthase gene (abaI). AbaI and AbaR have significant homology with LuxI and LuxR, a family of cell density-responsive transcriptional regulators first reported in Vibrio fischeri. By using A. baumannii M2 and its quorum-sensing mutant M2 (abaI::Km), a lower production of anti-oxidative enzymes in the mutant alongside its greater pyocyanin sensitivity, as compared with the WT strain, was observed.9 Similar results were obtained in another report regarding Salmonella Typhi.10

The toxin–antitoxin systems are of great importance in controlling the survival of bacteria. More than 30 years ago, the hipA7 allele in E. coli K-12 was found to be associated with a greater propensity to induce persister formation. Subsequently, mutagenesis studies have shown a marked effect on persistence, with deletion of tisB, mqsR or hipA operons.11 Association of MazEF and RelBE with bacterial persistence has also been found.12 Among the toxin–antitoxin type II system families, HipA-HipB is highly recognized for its role in bacterial persistence.11 In M. tuberculosis, there are many more toxin–antitoxin systems than in other bacteria, suggesting the probably greater capacity of tubercle bacillus to enter into dormancy and persistence under conditions of stress. Using Mycobacterium smegmatis, the toxins of the largest family of the systems, VapBC, were found to act by inhibiting translation via mRNA cleavage. The toxin–antitoxin systems, Rv2009-2010 and Rv1955-1956, were induced during hypoxia. These loci share the same genomic island with dosT and fdxA, belonging to the dormancy regulon. During macrophage infection, Rv1560-1561 and Rv0549c-0550c were found as induced systems.13 The RelBE module in the toxin–antitoxin systems has the notable role of inducing M. tuberculosis persistence when confronted with antibiotic stress.14 There has been a similar report regarding the up-regulation of three important relMtb toxin genes, and the corresponding antitoxins under states of oxidative stress.15 Alongside the Dos regulon, the definitive roles of mycobacterial genes pertaining to stringent response, SOS, DNA repair and protection, energy production, efflux activity and trans-translation, in the context of oxidative stress and other forms of stress, would benefit from further data accumulation.16–18 A recent study of high-persister mutants of M. tuberculosis has revealed the involvement of a wide array of candidate genes related to lipid biosynthesis and carbon metabolism, as well as transcriptional regulators, similar to the findings in some other bacteria.12,19

In addition to a high overall load of M. tuberculosis, an elevated proportion of persisters in the bacillary population may delay sputum bacteriological conversion and lower the probability of cure of TB. In diabetes mellitus a higher reactivation (relapse) rate of TB, especially in elderly subjects, has been observed,1,20 probably in association with increased formation of mycobacterial persisters (Figure 1). Immune dysfunction and oxidative stress are significant issues in ageing.21,22

Bacterial persisters are noted for their phenotypic tolerance to antimicrobials.8,12 Increased activity of efflux pumps is probably largely responsible for the phenotypic tolerance of M. tuberculosis persisters to anti-TB drugs.23 In diabetes mellitus, reduced exposure to rifampicin could occur in a proportion of patients.24,25 The underlying mechanisms are not fully known, but lowered anti-TB drug bioavailability, when present, may constitute a pharmacokinetic–pharmacodynamic scenario to further induce drug efflux pumps of M. tuberculosis persisters.26 It is possible that even a small effect may have dramatic consequences, as the currently recommended dosing of rifampicin in standard TB treatment regimens may be suboptimal.27

Phenotypic tolerance is conceivably not dichotomous with genetic resistance.28,29 The former may permit or even facilitate the development of the latter, especially with a high mycobacterial burden and host immune dysfunction. As clinical drug resistance results from the selection of spontaneously emerging drug-resistant genetic mutants (Figure 1), largely in the face of poor physician prescription, poor patient adherence and poor drug quality/supply, phenotypic tolerance possibly provides the necessary but not sufficient condition for evolution towards clinically relevant drug resistance. Well-functioning TB control programmes can probably address most of these issues. Disparate findings regarding drug resistance in M. tuberculosis strains in diabetes mellitus cohort studies in different countries and geographical areas might partly reflect the diversity in performance of their TB control programmes.4,30

Regarding adverse reactions due to anti-TB drugs, isoniazid-induced neurotoxicity and hepatotoxicity31,32 and clofazimine-induced cardiotoxicity33 were recently shown in animal studies to be associated with oxidative stress. Complications of diabetes mellitus, putatively also driven by oxidative stress,3 may aggravate these therapeutically induced adverse reactions.

The oxidative stress mechanism may appear too simple to fully explain the impact of diabetes mellitus on the outcomes of TB treatment. It is likely that other molecular mechanisms, especially nitrosative stress and reductive stress, play additional and essential roles. Consequently, participation of the redox homeostasis in a broad perspective is biologically plausible.34 Furthermore, oxidative stress can interact with immunological responses.2 These caveats notwithstanding, the present hypothesis can serve as a platform to explore the interaction of TB, diabetes mellitus, smoking and ageing, and underscores the paramount importance of programmatic management of both TB and diabetes mellitus in collaboration.1 The putative mechanisms impacting TB treatment outcomes, depicted in the hypothesis, may help to shed light on the directions and strategies worthy of exploration in the research agenda for currently rampant epidemics of TB and diabetes mellitus. Oxidative stress induces a number of derangements in cell physiology, including lipid peroxidation, protein oxidation and change in levels of antioxidant substrates and enzymes.3 These alterations potentially constitute a basis for testing the hypothesis in both experimental and clinical settings. Furthermore, studies on the association of oxidative stress and poor glycaemic control in diabetic patients are informative,3 as would be also those regarding the association of oxidative stress and TB pathogenesis, cure, relapse and mortality in animal models and patient populations. M. tuberculosis persister assays have undergone substantial advancement.35 Well-designed epidemiological and clinical cohort studies to better inform the impact of performance of TB control programmes, with respect to drug-resistant TB scenarios in diabetes mellitus, would further complement testing the validity of the current hypothesis.

Acknowledgements

Li Ka Shing Institute of Health Sciences is gratefully acknowledged for providing technical support to this research.

Transparency declarations

None to declare.

References

1
Lonnroth
K
Roglic
G
Harries
AD.
Improving tuberculosis prevention and care through addressing the global diabetes epidemic: from evidence to policy and practice
.
Lancet Diabetes Endocrinol
 
2014
;
2
:
730
9
.
2
Hodgson
K
Morris
J
Bridson
T
et al.  .
Immunological mechanisms contributing to the double burden of diabetes and intracellular bacterial infections
.
Immunology
 
2015
;
144
:
171
85
.
3
Tiwari
BK
Pandey
KB
Abidi
AB
et al.  .
Markers of oxidative stress during diabetes mellitus
.
J Biomark
 
2013
;
2013
:
378790
.
4
Chiang
CY
Bai
KJ
Lin
HH
et al.  .
The influence of diabetes, glycemic control and diabetes-related comorbidities on pulmonary tuberculosis
.
PLoS One
 
2015
;
10
:
e0121698
.
5
Yamagishi
S
Nakamura
N
Suematsu
M
et al.  .
Advanced glycation end products: a molecular target for vascular complications in diabetes
.
Mol Med
 
2015
;
21
Suppl 1:
S32
40
.
6
Scherrer
R
Moyed
HS.
Conditional impairment of cell division and altered lethality in hipA mutants of Escherichia coli K-12
.
J Bacteriol
 
1988
;
170
:
3321
6
.
7
Wang
X
Wood
TK.
Toxin–antitoxin systems influence biofilm and persister cell formation and the general stress response
.
Appl Environ Microbiol
 
2011
;
77
:
5577
83
.
8
Wu
Y
Vulic
M
Keren
I
et al.  .
Role of oxidative stress in persister tolerance
.
Antimicrob Agents Chemother
 
2012
;
56
:
4922
6
.
9
Bhargava
N
Sharma
P
Capalash
N.
Pyocyanin stimulates quorum sensing-mediated tolerance to oxidative stress and increases persister cell populations in Acinetobacter baumannii
.
Infect Immun
 
2014
;
82
:
3417
25
.
10
Walawalkar
YD
Vaidya
Y
Nayak
V.
Response of Salmonella Typhi to bile-generated oxidative stress: implication of quorum sensing and persister cell populations
.
Pathog Dis
 
2016
;
74
:
ftw090.
11
Wen
Y
Behiels
E
Devreese
B.
Toxin–antitoxin systems: their role in persistence, biofilm formation, and pathogenicity
.
Pathog Dis
 
2014
;
70
:
240
9
.
12
Leung
V
Levesque
CM.
A stress-inducible quorum-sensing peptide mediates the formation of persister cells with noninherited multidrug tolerance
.
J Bacteriol
 
2012
;
194
:
2265
74
.
13
Ramage
HR
Connolly
LE
Cox
JS.
Comprehensive functional analysis of Mycobacterium tuberculosis toxin–antitoxin systems: implications for pathogenesis, stress responses, and evolution
.
PLoS Genet
 
2009
;
5
:
e1000767.
14
Singh
R
Barry
CEIII
Boshoff
HI.
The three RelE homologs of Mycobacterium tuberculosis have individual, drug-specific effects on bacterial antibiotic tolerance
.
J Bacteriol
 
2010
;
192
:
1279
91
.
15
Korch
SB
Malhotra
V
Contreras
H
et al.  .
The Mycobacterium tuberculosis relBE toxin:antitoxin genes are stress-responsive modules that regulate growth through translation inhibition
.
J Microbiol
 
2015
;
53
:
783
95
. Erratum in: J Microbiol 2015; 53: 875.
16
Keren
I
Minami
S
Rubin
E
et al.  .
Characterization and transcriptome analysis of Mycobacterium tuberculosis persisters
.
mBio
 
2011
;
2
:
e00100
11
.
17
de Keijzer
J
Mulder
A
de Beer
J
et al.  .
Mechanisms of phenotypic rifampicin tolerance in Mycobacterium tuberculosis Beijing genotype strain B0/W148 revealed by proteomics
.
J Proteome Res
 
2016
;
15
:
1194
204
.
18
Zhang
Y.
Persisters, persistent infections and the Yin-Yang model
.
Emerg Microbes Infect
 
2014
;
3
:
e3.
19
Torrey
HL
Keren
I
Via
LE
et al.  .
High persister mutants in Mycobacterium tuberculosis
.
PLoS One
 
2016
;
11
:
e0155127
.
20
Leung
CC
Lam
TH
Chan
WM
et al.  .
Diabetic control and risk of tuberculosis: a cohort study
.
Am J Epidemiol
 
2008
;
167
:
1486
94
.
21
Bandaranayake
T
Shaw
AC.
Host resistance and immune aging
.
Clin Geriatr Med
 
2016
;
32
:
415
32
.
22
Edrey
YH
Salmon
AB.
Revisiting an age-old question regarding oxidative stress
.
Free Radic Biol Med
 
2014
;
71
:
368
78
.
23
Ramon-Garcia
S
Martin
C
Thompson
CJ
et al.  .
Role of the Mycobacterium tuberculosis P55 efflux pump in intrinsic drug resistance, oxidative stress responses and growth
.
Antimicrob Agents Chemother
 
2009
;
53
:
3675
82
.
24
Nijland
HM
Ruslami
R
Stalenhoef
JE
et al.  .
Exposure to rifampicin is strongly reduced in patients with tuberculosis and type 2 diabetes
.
Clin Infect Dis
 
2006
;
43
:
848
54
.
25
Babalik
A
Ulus
IH
Bakirci
N
et al.  .
Plasma concentrations of isoniazid and rifampin are decreased in adult pulmonary tuberculosis patients with diabetes mellitus
.
Antimicrob Agents Chemother
 
2013
;
57
:
5740
2
.
26
Pasipanodya
JG
Gumbo
T.
A new evolutionary and pharmacokinetic-pharmacodynamic scenario for rapid emergence of resistance to single and multiple anti-tuberculosis drugs
.
Curr Opin Pharmacol
 
2011
;
11
:
457
63
.
27
Boeree
MJ
Diacon
AH
Dawson
R
et al.  .
on behalf of the PanACEA Consortium
.
A dose-ranging trial to optimize the dose of rifampin in the treatment of tuberculosis
.
Am J Respir Crit Care Med
 
2015
;
191
:
1058
65
.
28
Aldridge
BB
Keren
I
Fortune
SM.
The spectrum of drug susceptibility in mycobacteria
.
Microbiol Spectr
 
2014
;
2
: MGM2-0031-2013.
29
den Hertog
AL
Menting
S
van Soolingen
D
et al.  .
Mycobacterium tuberculosis Beijing genotype resistance to transient rifampin exposure
.
Emerg Infect Dis
 
2014
;
20
:
1932
3
.
30
Mehta
S
Yu
EA
Ahamed
SF
et al.  .
Rifampin resistance and diabetes mellitus in a cross-sectional study of adult patients in rural South India
.
BMC Infect Dis
 
2015
;
15
:
451
.
31
Ahadpour
M
Eskandari
MR
Mashayekhi
V
et al.  .
Mitochondrial oxidative stress and dysfunction induced by isoniazid: study on isolated rat liver and brain mitochondria
.
Drug Chem Toxicol
 
2016
;
39
:
224
32
.
32
Hassan
HM
Guo
H
Yousef
BA
et al.  .
Role of inflammatory and oxidative stress, cytochrome P450 2E1, and bile acid disturbance in rat liver injury induced by isoniazid and lipopolysaccharide cotreatment
.
Antimicrob Agents Chemother
 
2016
;
60
:
5285
93
.
33
Chan
JY
Leung
RK
Lee
SM
et al.  . Clofazimine-induced cardiotoxicity in zebrafish model. In: Abstracts of the Fifty-fifth Interscience Conference on Antimicrobial Agents and Chemotherapy, San Diego, CA, USA, 2015. Abstract 1469. American Society for Microbiology, Washington, DC, USA.
34
Kumar
A
Farhana
A
Guidry
L
et al.  .
Redox homeostasis in mycobacteria: the key to tuberculosis control?
Expert Rev Mol Med
 
2011
;
13
:
e39
.
35
Rowe
SE
Conlon
BP
Keren
I
et al.  .
Persisters: methods for isolation and identifying contributing factors—a review
.
Methods Mol Biol
 
2016
;
1333
:
17
28
.