- Email: [email protected]
A novel approach for tuberculosis diagnosis using exosomal DNA and droplet digital PCR
Journal Pre-proof A novel approach for tuberculosis diagnosis using exosomal DNA and droplet digital PCR Sun-Mi Cho, Saeam Shin, Yoonjung Kim, Wonkeun Song, Seong Geun Hong, Seok Hoon Jeong, Myung Seo Kang, Kyung-A. Lee PII:
S1198-743X(19)30611-1
DOI:
https://doi.org/10.1016/j.cmi.2019.11.012
Reference:
CMI 1846
To appear in:
Clinical Microbiology and Infection
Received Date: 26 August 2019 Revised Date:
7 November 2019
Accepted Date: 9 November 2019
Please cite this article as: Cho S-M, Shin S, Kim Y, Song W, Hong SG, Jeong SH, Kang MS, Lee KA, A novel approach for tuberculosis diagnosis using exosomal DNA and droplet digital PCR, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.11.012. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Ltd on behalf of European Society of Clinical Microbiology and Infectious Diseases.
1
Original article
2
A novel approach for tuberculosis diagnosis using exosomal DNA and droplet digital
3
PCR
4 5
Sun-Mi Choa*, Saeam Shinb*, Yoonjung Kimb*, Wonkeun Songc, Seong Geun Honga, Seok
6
Hoon Jeongb, Myung Seo Kanga†, Kyung-A Leeb†
7 8
a
9
Seongnam, Republic of Korea
Department of Laboratory Medicine, CHA Bundang Medical Center, CHA University,
10
b
11
Republic of Korea
12
c
13
Republic of Korea
Department of Laboratory Medicine, Yonsei University College of Medicine, Seoul,
Department of Laboratory Medicine, Hallym University College of Medicine, Seoul,
14 15
*
These authors contributed equally to this work.
16 17
†
18
Kyung-A Lee
19
Department of Laboratory Medicine, Yonsei University College of Medicine, 211, Eonju-ro,
20
Gangnam-gu, Seoul, 06273, Republic of Korea
21
Tel: +82-2-2019-3531
22
Fax: +82-2-2019-4822
23
E-mail: [email protected]
Corresponding authors
24 1
25
Myung Seo Kang
26
Department of Laboratory Medicine, CHA Bundang Medical Center, CHA University, 59,
27
Yatap-ro, Bundang-gu, Seongnam, 13496, Republic of Korea
28
Tel: +82-31-780-5384
29
Fax: +82-31-780-5476
30
E-mail: [email protected]
31
2
32
ABSTRACT
33
Objectives: The rapid diagnosis of tuberculosis (TB) is important for patient treatment and
34
infection control. Current molecular diagnostic techniques for TB have insufficient sensitivity
35
to detect samples with low bacterial loads. The sensitivity of molecular testing depends on
36
not only performance of assay technique but also nucleic acid extraction method. Here, we
37
have presented a novel approach using exosomal DNA (exoDNA) and droplet digital PCR
38
(ddPCR) platform to detect Mycobacterium tuberculosis DNA in clinical samples.
39
Methods: The ddPCR platform targeting IS6110 was evaluated in parallel using total DNA
40
and exoDNA. The clinical performance of ddPCR method was assessed with 190 respiratory
41
samples from patients with suspected pulmonary TB.
42
Results: Compared with mycobacterial culture, sensitivity and specificity of ddPCR were
43
61.5% (95% CI, 44.6%-76.6%) and 98.0% (95% CI, 94.3%-99.6%) using total DNA, and
44
76.9% (95% CI, 60.7%-88.9%) and 98.0% (95% CI, 94.3%-99.6%) using exoDNA
45
respectively. Among 15 culture-positive specimens with low concentrations of target
46
molecules (2~99 positive droplets with exoDNA), only 53.3% (8/15), 46.7% (7/15), and 26.7%
47
(4/15) of cases were detected using ddPCR with total DNA, real-time PCR with exoDNA, or
48
real-time PCR with total DNA.
49
Conclusions: Our platform using ddPCR and exoDNA has the potential to provide sensitive
50
and accurate methodology for TB diagnosis.
3
51
Introduction
52
Tuberculosis (TB), which is caused by Mycobacterium tuberculosis (MTB) infection,
53
is a critical health problem with high infectivity, morbidity and mortality. Mycobacterial
54
culture is time-consuming and acid-fast bacilli (AFB) smear has low sensitivity. Because of
55
the limitations of conventional diagnostic methods, nucleic acid amplification tests (NAATs)
56
have been usefully applied in rapid diagnosis of TB. To improve sensitivity, several
57
platforms adopt sensitive molecular techniques such as real-time polymerase chain reaction
58
(PCR), nested PCR, and isothermal transcription-mediated amplification. However, the
59
sensitivity of NAATs is still insufficient compared with conventional mycobacterial
60
culture [1].
61
Droplet digital PCR (ddPCR) is an emerging technology capable of absolute nucleic
62
acid quantification without using standard curves [2]. ddPCR technology uses a combination
63
of microfluidics and proprietary surfactant chemistries to divide PCR samples into water-in-
64
oil droplets [3]. These droplets support PCR amplification of single template molecules
65
using reagents and workflows similar to those widely used for real-time PCR applications [3].
66
Following PCR, each droplet is analyzed to assign positivity or negativity based on their
67
fluorescence amplitude, after which the concentration of target DNA in the original sample is
68
calculated [3]. These principles offer the advantages of precise and reproducible data without
69
being affected by PCR inhibitors in samples over real-time PCR [4]. Since 20,000 droplets
70
are generated for a sample, the theoretical detection limit of ddPCR is 0.005% (1/20,000) and
71
reported detection limits are 0.01~0.001% [3,5,6]. With low detection limits for ddPCR,
72
many applications have been developed for analyzing pathogens or cancer-derived mutations
73
present at low levels in clinical samples. Accordingly, we sought to evaluate the clinical 4
74
performance of ddPCR in detecting MTB in clinical samples.
75
Exosomes are small membrane vesicles (30–100 nm) secreted by many cell types
76
during physiological and pathological conditions [7]. They contain a variety of molecules
77
derived from the original cell, including DNA, RNA, proteins, and lipids [8]. Recent
78
evidence has indicated that the exosomal nucleic acids are more stable than other forms of
79
nucleic acids, probably by the protective effect of lipid bilayer coating [9]. Exosomes are
80
isolated in most biological fluids, such as serum, urine, and bronchoalveolar lavage (BAL)
81
fluid [10]. Such ubiquitous nature of exosomes in body fluids makes them ideal for use as
82
diagnostic biomarkers. Previous reports indicated that exosomes isolated from the BAL fluid
83
of MTB-infected mice or sera TB patients contain mycobacterial proteins [11].
84 85
In this study, we attempted a novel approach for TB diagnosis targeting exosomal DNA (exoDNA), using ddPCR.
86
87
Methods
88
Clinical specimens
89
The study protocol was approved by the Institutional Review Boards (IRBs) of the
90
participating institutions (Hallym University Kangnam Sacred Heart Hospital, Yonsei
91
University Gangnam Severance Hospital, and Bundang CHA General Hospital). Between
92
August 2018 to October 2019, a total of 174 consecutive subjects undergoing evaluation for
93
pulmonary TB were included in the study. In 174 consecutive samples, mycobacterial
94
culture-positive sample was 13% (n=23). Therefore, we selectively added 16 culture-positive
95
specimens to evaluate the efficiency of exoDNA and ddPCR for detecting MTB. The final 5
96
clinical diagnosis of TB was made by infectious physicians or respiratory physicians based
97
on integration of patients’ medical history, radiologic findings, and laboratory findings
98
(microbiological, molecular and immunological tests). The researchers retrospectively
99
reviewed the participant's medical records, including the final clinical diagnosis. All samples
100
were tested using AFB smear, mycobacterial culture, real-time PCR, and ddPCR. For non-
101
sterile specimens, an equal volume of 4% NaOH was added for emulsification and
102
decontamination. Then the mixture was transferred to a sterile tube and centrifuged for 20
103
minutes at 3,000 g. After performing the AFB smear procedure with an auramine-rhodamine
104
fluorescent stain, it was confirmed by Ziehl-Neelsen staining. For mycobacterial culture, all
105
decontaminated samples were inoculated into a mycobacterial growth indicator tube (MGIT
106
960 system; Becton Dickinson, Sparks, MD) and on 3% Ogawa agar (Shinyang, Seoul,
107
Republic of Korea) and cultured for 6 weeks.
108
DNA extraction
109
For analysis of total DNA, 100 µL aliquots of the decontaminated specimens were
110
re-suspended in DNA extraction buffer. Total DNA was extracted using QIAamp DNA mini
111
kit (QIAGEN, Hilden, Germany) following manufacturer’s instructions. For analysis of
112
exoDNA, we isolated exosomal fractions from 1 mL of respiratory specimen using
113
ExoQuick™
114
recommendations (System Biosciences Inc., Mountain View, CA, USA). Submitted
115
specimens were centrifuged at 3000 x g for 15 minutes to remove cells and cell debris.
116
Briefly, 1/4 volume of ExoQuick Solution was added to respiratory specimen, and samples
117
were refrigerated at 4 °C overnight. The mixture was centrifuged at 1500 × g for 30 min, and
118
supernatant was removed by aspiration. Pelleted fractions were re-suspended in nuclease-free
Exosome
Precipitation
Solution
6
according
to
the
manufacturer's
119
water. Subsequently, exosome-derived DNA was extracted using QIAamp DNA mini kits as
120
described above.
121
Real-time PCR
122
Real-time PCR was performed using the PowerCheck MTB/NTM Real-time PCR
123
assay (Kogene Biotech, Seoul, Korea), targeting the MTB-specific IS6110, following
124
manufacturer’s instructions. The 20 µL PCR mix was composed of 10 µL 2X Real-time PCR
125
Master Mix, 5 µL of primer/probe mix, and 5 µL total DNA. The analysis was done with
126
CFX96 real-time PCR detection system (Bio-Rad Laboratories, Hercules, CA, USA).
127
ddPCR
128
Digital PCR reactions were performed with a QX200 Droplet Digital PCR System
129
(Bio-Rad Laboratories, Hercules, CA, USA). The reactions were carried out using the
130
previously described primers and probes, targeting IS6110, and thermal cycling
131
conditions[12]. A no-template control was used in every ddPCR batch. The 20 µL PCR mix
132
was composed of 10 µL Bio-Rad Super mix TaqMan, 1 µL of each amplification
133
primer/probe mix, 5 µL DNA and 4 µL nuclease-free water. Results were analyzed with
134
QuantaSoft v.1.7.2 software (Bio-Rad Laboratories). This provided the number of positive
135
and negative droplets, as well as quantification of IS6110 of MTB, expressed as copies/mL of
136
ddPCR reaction. At least two positive droplets were required for a positive test result of the
137
ddPCR assay [12].
138
Limit of detection of real-time PCR and ddPCR assay
139
The analytical sensitivities of real-time PCR and the ddPCR assay were evaluated
140
using spiked samples with ATCC 25177. The limits of detection (LODs) for both PCR assays 7
141
were measured using five dilutions around LOD. For all concentrations, repetitive tests were
142
performed with 10 replicates in three different runs at each concentration, and LODs were
143
calculated by probit analysis for a 95% positive result.
144
Statistical analysis
145
SPSS version 24.0 software (SPSS Inc., Chicago, IL, USA) and MedCalc version
146
19.1 (MedCalc software, Mariakerke, Belgium) was used for statistical analyses. The chi-
147
squared test was used to analyze tabular data. A correlation study was performed with the
148
Spearman’s rank correlation procedure (r). All p-values were two-sided, and values less than
149
0.05 were considered statistically significant.
150
151
Results
152
Patient characteristics
153
A total of 190 respiratory specimens, including 151 (79.5%) sputa, 31 (16.3%)
154
bronchial washings, and 8 (4.2%) bronchoalveolar lavage (BAL) fluids, were included in this
155
study. The demographic data for 190 subjects are listed in Table 1. Forty four patients (23.2%)
156
were finally diagnosed with pulmonary tuberculosis.
157
Considering mycobacterial culture as a reference method, sensitivity and specificity
158
were 35.9% (95% CI, 21.2%-52.8%) and 98.0% (95% CI, 94.3%-99.6%) for AFB smear
159
(Table S1) Compared with final clinical diagnosis, sensitivity and specificity were 34.1% (95%
160
CI, 20.5%-49.9%) and 98.6% (95% CI, 95.1%-99.8%) for AFB smear and 88.6% (95% CI,
161
75.4%-96.2%) and 100% (95% CI, 97.5%-100.0%) for mycobacterial culture, respectively 8
162
(Table 3).
163
Results of real-time PCR using exosomal DNA (exoDNA) and total DNA
164
Probit analysis of 95% positivity in 10 replicates in three different runs at five
165
concentrations showed MTB had a LOD of 13.8 copies/reaction (Table S2). Real-time PCR
166
using exoDNA and total DNA yielded 24 positive suspects and 21 positive suspects,
167
respectively, and three of each were culture negative (Table 2). All three of these subjects
168
also showed positive results in all other PCR assays, and 1 patient had a positive result by
169
AFB smear. Based on culture results, the sensitivities of the real-time PCR assays with both
170
exoDNA and total DNA were 53.9% (95% CI, 37.2-69.9) and 46.2% (95% CI, 30.1%-
171
62.8%), respectively (Table 2). Compared with final clinical diagnosis, the sensitivities of
172
both exoDNA and total DNA were 54.6% (95% CI, 38.9%–69.6%) and 47.7% (95% CI,
173
32.5%-63.3%), respectively (Table 3). The specificities of exoDNA and total DNA according
174
to culture results were both 98.0% (95% CI, 94.3%-99.6%) and according to the final clinical
175
diagnosis were both 100% (95% CI, 97.5%-100.0%) (Table 2, 3).
176
Results of ddPCR using total DNA
177
We found that the MTB LOD with ddPCR was 2.3 copies/reaction (Table S3). The
178
ddPCR yielded 27 positive suspects, 3 of which were culture negative (Table 2). Of these 3
179
suspects, 2 had real-time PCR positive results; only 1 patient had a positive result by AFB
180
smear. Of the 163 ddPCR negative suspects, 15 were culture positive. Based on culture
181
results, the sensitivity and specificity of the ddPCR assays with total DNA were 61.5% (95%
182
CI, 44.6-76.6) and 98.0% (95% CI, 94.2–99.6), respectively (Table 2). Compared with final
183
clinical diagnosis, the sensitivity and specificity of the ddPCR with total DNA were 61.4% 9
184
(95% CI, 45.5-75.6), and 100% (95% CI, 97.5%–100%), respectively (Table 3).
185
Results of ddPCR using exosomal DNA (exoDNA)
186
The ddPCR with exoDNA yielded 33 positive suspects, 3 of whom were culture
187
negative (Table 2). Of these 3 suspects, all 3 had real-time PCR positive results; only 1
188
patient had a positive result by AFB smear. Of the 157 ddPCR negative suspects, 9 were
189
culture positive. Based on culture results, the sensitivity and specificity of the ddPCR with
190
exoDNA for detecting MTB were 76.9% (95% CI, 60.7-88.9) and 98.0% (95% CI, 90.2–
191
97.6), respectively (Table 2). Compared with final clinical diagnosis, the sensitivity and
192
specificity of the ddPCR with exoDNA were 75.0% (95% CI, 59.7-86.8) and 100% (95% CI,
193
97.5–100), respectively (Table 3).
194
Receptor operating characteristics (ROC) curve analysis was performed to evaluate
195
diagnostic accuracy of ddPCR to predict MTB culture positivity. The area under the ROC
196
curve was 0.80 (95% CI, 0.70-0.89) and 0.88 (95% CI, 0.79-0.96) with the total DNA and
197
exoDNA, respectively (Fig. S1).
198
Measured IS6110 copy numbers of exoDNA using ddPCR were correlated with those
199
of total DNA (p<0.01, r=0.827, Spearman correlation) (Fig. 1A). The Bland-Altman analysis
200
of ddPCR results using total DNA and exoDNA showed the mean difference of 0.2 ± 0.94
201
log10 with 95% Limits of Agreement (LoA) ranging from -1.7 to 2.0 in TB culture-positive
202
samples. (Fig.1B). Notably, in low concentration samples (< 100 positive droplets in
203
exoDNA) with positive culture results, assays with exoDNA showed improved sensitivity
204
over assays with total DNA (Table 4, Fig. S2). The Bland-Altman analysis indicated that the
205
difference of measured TB copies between exoDNA and total DNA was 0.45 ± 0.91 log10 10
206
(LoA: -1.3 to 2.2), and TB copies were 79% higher on average in exoDNA than in total DNA
207
(Fig. S2). And, the detection rates of ddPCR with total DNA, real-time PCR with exoDNA,
208
real-time PCR with total DNA were 53.3%, 46.7%, and 26.7% of cases, respectively,
209
compared with ddPCR with exoDNA (Table 4).
210
211
Discussion
212
The ddPCR assay using exoDNA had higher sensitivity than ddPCR using total DNA
213
in our study. Therefore, exosomes may serve as a feasible target for detection of pathogen-
214
derived nucleic acid. To our knowledge, this is the first study reporting detection of MTB
215
DNA from exosomes using clinical specimens isolated from TB patients. The detection rate
216
in low concentration samples was significantly higher when using exoDNA than total DNA
217
in PCR-based assays (Table 4). The results could be explained by stability and abundance of
218
exoDNA in clinical samples. The stability is accomplished by the lipid bilayers which protect
219
exosomal nucleic acids from the extracellular environment [13]. The number of exosome
220
reported by the previous study is approximately 1.15 × 1010 in 10 million Tb infected cells,
221
which cultured for three days [14].
222
In this study, three culture negative specimens showed positive results with ddPCR
223
using exoDNA. These cases were confirmed pulmonary tuberculosis based on radiologic
224
findings and/or real-time PCR, and treated with anti-TB medication. Although mycobacterial
225
culture has long been the conventional gold standard test for TB diagnosis, culture results can
226
be negative in mild TB infection [15], latent TB infection and early sub-clinical disease [16].
227
In addition, the decontamination procedure of sputum prior to culture dilutes the sample and 11
228
leads to the loss of viability [17]. Because identification of these groups is important for
229
infection control, ddPCR could provide fast and more accurate results.
230
In this study, the false negative results both in ddPCR and real-time PCR might have
231
been due to IS6110 copy number variation [18]. Factors that affect NAAT sensitivity include
232
the quality of sputum [19] and clinician bias with regard to diagnostic approaches.
233
There are several limitations in this study that could be addressed in future studies.
234
First, we have the potential to study fluctuation in exosome concentration of TB infection.
235
The biogenesis and release of exosomes are affected by various activating signals and cellular
236
stress that reflect changes in the environmental conditions [20]. Previous study has suggested
237
that infection induces exosome secretion and that this is correlated with bacterial burden [14].
238
However, the mechanisms regulating exosome release have not been well elucidated to date.
239
Further investigation is required to determine controlling factors. Second, the pre-analytical
240
phase is an important step with impact on results downstream. In clinical settings, most
241
requested samples are analyzed by the batch as in this study. So, samples are usually frozen
242
and stored until analysis. Exosomes in plasma, urine, and saliva seem stable during a freeze-
243
thaw cycle and storage [21]. However, an optimal protocol is needed specific to the type of
244
specimens and the downstream analysis. Third, altering the isolation methods of exosome
245
from specimen may improve the sensitivity of this assay. Here we used commercial kits
246
showing higher extraction efficiency than ultracentrifugation [22] and better performance
247
than ddPCR using total DNA. Laboratory clinicians and researchers should be aware of the
248
advantages and disadvantages of each method and compare and validate before clinical
249
application of exoDNA. In addition, the results of the current study were obtained from a
250
small sample size. Thus, further studies with larger sample size in multicenter settings are 12
251
required to validate these results.
252
The ddPCR using exoDNA could be helpful in clinical conditions where the bacterial
253
load in the specimen is low such as in certain patients with extrapulmonary TB [23],
254
children [24], and patients with HIV-infection [25]. The increased sensitivity of ddPCR is
255
applicable in other areas, such as a method of detection of small numbers of drug resistant TB
256
or in determining the therapeutic effect of the treatment of TB. Serum or urine derived
257
exosomes have recently emerged as non-invasive diagnostic and prognostic biomarkers in
258
cancer [26]. These biological fluids could be alternative specimens for NAAT. This
259
supposition requires future verification with minimally invasive samples.
260
This is the first study reporting detection of MTB DNA from exosomes using clinical
261
specimens isolated from TB patients in combination with ddPCR assay to date. In conclusion,
262
ddPCR with exoDNA has the potential to provide a highly sensitive and accurate method for
263
diagnosis of MTB infection.
264
265
Acknowledgements
266
We thank the dedicated physicians at each hospital where this multicenter study was
267
conducted (YS Sim and J Lee from Hallym University Kangnam Sacred Heart Hospital, EK
268
Kim and M Kim from CHA Bundang Medical Center, and HJ Park and KH Lee from Yonsei
269
University Gangnam Severance Hospital). This study was supported by a grant of the
270
National Research Foundation of Korea (NRF-2019R1C1C1010916).
271
13
272
Transparency declaration
273
All authors report no competing interests or patents.
274
275
References
276
[1] Boehme CC, Nicol MP, Nabeta P, Michael JS, Gotuzzo E, Tahirli R, et al. Feasibility,
277
diagnostic accuracy, and effectiveness of decentralised use of the Xpert MTB/RIF test for
278
diagnosis of tuberculosis and multidrug resistance: a multicentre implementation study.
279
Lancet 2011;377:1495-505.
280
[2] Pinheiro LB, Coleman VA, Hindson CM, Herrmann J, Hindson BJ, Bhat S, et al.
281
Evaluation of a droplet digital polymerase chain reaction format for DNA copy number
282
quantification. Anal Chem 2012;84:1003-11.
283
[3] Hindson BJ, Ness KD, Masquelier DA, Belgrader P, Heredia NJ, Makarewicz AJ, et al.
284
High-Throughput Droplet Digital PCR System for Absolute Quantitation of DNA Copy
285
Number. Anal Chem 2011;83:8604-10.
286
[4] Dingle TC, Sedlak RH, Cook L, Jerome KR. Tolerance of droplet-digital PCR vs real-
287
time quantitative PCR to inhibitory substances. Clin Chem 2013;59:1670-2.
288
[5] Azuara D, Ginesta MM, Gausachs M, Rodriguez-Moranta F, Fabregat J, Busquets J, et al.
289
Nanofluidic Digital PCR for KRAS Mutation Detection and Quantification in
290
Gastrointestinal Cancer. Clin Chem 2012;58:1332-41.
291
[6] Taly V, Pekin D, Benhaim L, Kotsopoulos SK, Le Corre D, Li X, et al. Multiplex
292
picodroplet digital PCR to detect KRAS mutations in circulating DNA from the plasma of
293
colorectal cancer patients. Clin Chem 2013;59:1722-31. 14
294
[7] Schorey JS, Cheng Y, Singh PP, Smith VL. Exosomes and other extracellular vesicles in
295
host-pathogen interactions. EMBO Rep 2015;16:24-43.
296
[8] Ruivo CF, Adem B, Silva M, Melo SA. The Biology of Cancer Exosomes: Insights and
297
New Perspectives. Cancer Res 2017;77:6480-8.
298
[9] Kahlert C, Melo SA, Protopopov A, Tang J, Seth S, Koch M, et al. Identification of
299
double-stranded genomic DNA spanning all chromosomes with mutated KRAS and p53
300
DNA in the serum exosomes of patients with pancreatic cancer. J Biol Chem 2014;289:3869-
301
75.
302
[10] Yanez-Mo M, Siljander PR, Andreu Z, Zavec AB, Borras FE, Buzas EI, et al. Biological
303
properties of extracellular vesicles and their physiological functions. J Extracell Vesicles
304
2015;4:27066.
305
[11] Kruh-Garcia NA, Wolfe LM, Chaisson LH, Worodria WO, Nahid P, Schorey JS, et al.
306
Detection of Mycobacterium tuberculosis peptides in the exosomes of patients with active
307
and latent M. tuberculosis infection using MRM-MS. PloS one 2014;9:e103811.
308
[12] Ushio R, Yamamoto M, Nakashima K, Watanabe H, Nagai K, Shibata Y, et al. Digital
309
pcr assay detection of circulating mycobacterium tuberculosis DNA in pulmonary
310
tuberculosis patient plasma. Tuberculosis (Edinb) 2016;99:47-53.
311
[13] Taylor DD, Gercel-Taylor C. MicroRNA signatures of tumor-derived exosomes as
312
diagnostic biomarkers of ovarian cancer. Gynecol Oncol 2008;110:13-21.
313
[14] Singh PP, Smith VL, Karakousis PC, Schorey JS. Exosomes isolated from mycobacteria-
314
infected mice or cultured macrophages can recruit and activate immune cells in vitro and in
315
vivo. J Immunol 2012;189:777-85.
316
[15] Levy H, Feldman C, Sacho H, van der Meulen H, Kallenbach J, Koornhof H. A
317
reevaluation of sputum microscopy and culture in the diagnosis of pulmonary tuberculosis. 15
318
Chest 1989;95:1193-7.
319
[16] Salgame P, Geadas C, Collins L, Jones-López E, Ellner JJ. Latent tuberculosis infection–
320
revisiting and revising concepts. Tuberculosis (Edinb) 2015;95:373-84.
321
[17] Pheiffer C, Carroll N, Beyers N, Donald P, Duncan K, Uys P, et al. Time to detection of
322
Mycobacterium tuberculosis in BACTEC systems as a viable alternative to colony counting.
323
Int J Tuberc Lung Dis 2008;12:792-8.
324
[18] Sankar S, Kuppanan S, Balakrishnan B, Nandagopal B. Analysis of sequence diversity
325
among IS6110 sequence of Mycobacterium tuberculosis: possible implications for PCR based
326
detection. Bioinformation 2011;6:283-5.
327
[19] Pan X, Yang S, Deighton MA, Qu Y, Hong L, Su F. A comprehensive evaluation of
328
Xpert MTB/RIF assay with bronchoalveolar lavage fluid as a single test or combined with
329
conventional assays for diagnosis of pulmonary tuberculosis in china: a two-center
330
prospective study. Front Microbiol 2018;9:444.
331
[20] Zhang X, Yuan X, Shi H, Wu L, Qian H, Xu W. Exosomes in cancer: small particle, big
332
player. J Hematol Oncol 2015;8:83.
333
[21] Yuana Y, Boing AN, Grootemaat AE, van der Pol E, Hau CM, Cizmar P, et al. Handling
334
and storage of human body fluids for analysis of extracellular vesicles. J Extracell Vesicles
335
2015;4:29260.
336
[22] Li P, Kaslan M, Lee SH, Yao J, Gao Z. Progress in Exosome Isolation Techniques.
337
Theranostics 2017;7:789-804.
338
[23] Denkinger CM, Schumacher SG, Boehme CC, Dendukuri N, Pai M, Steingart KR. Xpert
339
MTB/RIF assay for the diagnosis of extrapulmonary tuberculosis: a systematic review and
340
meta-analysis. Eur Respir J 2014;44:435-46. 16
341
[24] Perez-Velez CM, Marais BJ. Tuberculosis in children. N Engl J Med. 2012;367:348-61.
342
[25] Getahun H, Harrington M, O'Brien R, Nunn P. Diagnosis of smear-negative pulmonary
343
tuberculosis in people with HIV infection or AIDS in resource-constrained settings:
344
informing urgent policy changes. Lancet 2007;369:2042-9.
345
[26] Huang T, Deng C-X. Current progresses of exosomes as cancer diagnostic and
346
prognostic biomarkers. Int J Biol Sci 2019;15:1.
347
17
1
Table 1. The demographic data for 190 subjects Pulmonary tuberculosis
Other diagnosis
(n=44)
(n=146)
Male
21 (47.7)
84 (57.5)
Female
23 (52.3)
62 (42.5)
60 (53-75)
68 (54-79)
7 (15.9)
25 (17.1)
Sputum
32 (72.7)
119 (81.5)
Bronchial washing
7 (15.9)
24 (16.4)
BAL
5 (11.4)
3 (2.1)
Characteristics
Sex (%)
Age, median (IQR) Previous tuberculosis history (%) Referred respiratory specimens (%)
IQR, interquartile range; BAL, bronchoalveolar lavage
2
Table 2. Analytical performance of assays for detection of Mycobacterium tuberculosis depending on culture results Samples
3
Assays
Culture positive Culture negative Sensitivity (95% CI) Specificity (95% CI) PCR + PCR - PCR + PCR qPCR with total DNA 18 21 3 148 46.2% (30.1%-62.8%) 98.0% (94.3%-99.6%) qPCR with exoDNA 21 18 3 148 53.9% (37.2%-69.9%) 98.0% (94.3%-99.6%) All ddPCR with total DNA 24 15 3 148 61.5% (44.6%-76.6%) 98.0% (94.3%-99.6%) ddPCR with exoDNA 30 9 3 148 76.9% (60.7%-88.9%) 98.0% (94.3%-99.6%) qPCR with total DNA 12 2 1 2 85.7% (57.2%-98.2%) 66.7% (9.4%-99.2%) Smear qPCR with exoDNA 13 1 1 2 92.9% (66.1%-99.8%) 66.7% (9.4%-99.2%) positive ddPCR with total DNA 13 1 1 2 92.9% (66.1%-99.8%) 66.7% (9.4%-99.2%) (n=17) ddPCR with exoDNA 14 0 1 2 100% (76.8%-100.0%) 66.7% (9.4%-99.2%) qPCR with total DNA 6 19 2 146 24.0% (9.4%-45.1%) 98.7% (95.3%-99.8%) Smear qPCR with exoDNA 8 17 2 146 32.0% (15.0%-53.5%) 98.7% (95.3%-99.8%) negative ddPCR with total DNA 11 14 2 146 44.0% (24.4%-65.1%) 98.7% (95.3%-99.8%) (n=173) ddPCR with exoDNA 16 9 2 146 64.0% (42.5%-82.0%) 98.7% (95.3%-99.8%) +, positive; -, negative; CI, confidence intervals; qPCR, real-time PCR; ddPCR, droplet digital PCR; exoDNA, exosomal DNA
4
Table 3. Clinical performance of assays for detection of Mycobacterium tuberculosis depending on confirmed diagnosis Method AFB smear qPCR with total DNA qPCR with exoDNA ddPCR with total DNA ddPCR with exoDNA
Results Pos Neg Pos Neg Pos Neg Pos Neg Pos
PTB (n=44) 15 29 21 23 24 20 27 17 33
Non-TB (n=146) 2 144 0 146 0 146 0 146 0
Neg
11
146
Sensitivity (95% CI)
Specificity (95% CI)
34.1% (20.5%-49.9%) 98.6% (95.1%-99.8%)
Accuracy (95% CI) 83.7% (77.7%-88.6%)
47.7% (32.5%-63.3%) 100% (97.5%-100.0%) 87.9% (82.4%-92.2%) 54.6% (38.9%-69.6%) 100% (97.5%-100.0%) 89.5% (84.2%-93.5%) 61.4% (45.5%-75.6%) 100% (97.5%-100.0%) 91.1% (86.1%-94.7%) 75.0% (59.7%-86.8%) 100% (97.5%-100.0%) 94.2% (89.9%-97.1%)
5
Pos 39 0 88.6% (75.4%-96.2%) 100% (97.5%-100.0%) 97.4% (94.0%-99.1%) Neg 5 146 PTB, pulmonary tuberculosis; Non-TB, patients diagnosed other than tuberculosis; Pos, positive; Neg, negative; CI, confidence intervals;
6
qPCR, real-time PCR; ddPCR, droplet digital PCR; exoDNA, exosomal DNA.
Culture
7
8
9
Table 4. Results of both ddPCR and real-time PCR results using two sources of DNA in low concentration samples ddPCR with qPCR with Mycobacterial exoDNA total DNA culture (No. of positive (No. of positive exoDNA total DNA droplets) droplets) Pos (2) Neg (1) Neg Neg MTB Pos (3) Neg (1) Neg Neg MTB Pos (3) Pos (9) Neg Neg MTB Pos (4) Pos (24) Neg Pos MTB Pos (4) Pos (2) Neg Neg MTB Pos (4) Neg (0) Neg Neg MTB Pos (7) Pos (2) Neg Neg MTB Pos (8) Neg (1) Neg Neg MTB Pos (15) Neg (1) Pos Neg MTB Pos (20) Neg (0) Pos Neg MTB Pos (52) Neg (0) Pos Neg MTB Pos (69) Pos (6) Pos Pos MTB Pos (93) Pos (143) Pos Pos MTB Pos (94) Pos (817) Pos Pos MTB Pos (99) Pos (7) Pos Neg MTB Positive rate (%) 53.3 46.7 26.7 100 Pos, positive; Neg, negative; qPCR, real-time PCR; ddPCR, droplet digital PCR; exoDNA, exosomal DNA. 10
11 12
S1198-743X(19)30611-1
DOI:
https://doi.org/10.1016/j.cmi.2019.11.012
Reference:
CMI 1846
To appear in:
Clinical Microbiology and Infection
Received Date: 26 August 2019 Revised Date:
7 November 2019
Accepted Date: 9 November 2019
Please cite this article as: Cho S-M, Shin S, Kim Y, Song W, Hong SG, Jeong SH, Kang MS, Lee KA, A novel approach for tuberculosis diagnosis using exosomal DNA and droplet digital PCR, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.11.012. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Ltd on behalf of European Society of Clinical Microbiology and Infectious Diseases.
1
Original article
2
A novel approach for tuberculosis diagnosis using exosomal DNA and droplet digital
3
PCR
4 5
Sun-Mi Choa*, Saeam Shinb*, Yoonjung Kimb*, Wonkeun Songc, Seong Geun Honga, Seok
6
Hoon Jeongb, Myung Seo Kanga†, Kyung-A Leeb†
7 8
a
9
Seongnam, Republic of Korea
Department of Laboratory Medicine, CHA Bundang Medical Center, CHA University,
10
b
11
Republic of Korea
12
c
13
Republic of Korea
Department of Laboratory Medicine, Yonsei University College of Medicine, Seoul,
Department of Laboratory Medicine, Hallym University College of Medicine, Seoul,
14 15
*
These authors contributed equally to this work.
16 17
†
18
Kyung-A Lee
19
Department of Laboratory Medicine, Yonsei University College of Medicine, 211, Eonju-ro,
20
Gangnam-gu, Seoul, 06273, Republic of Korea
21
Tel: +82-2-2019-3531
22
Fax: +82-2-2019-4822
23
E-mail: [email protected]
Corresponding authors
24 1
25
Myung Seo Kang
26
Department of Laboratory Medicine, CHA Bundang Medical Center, CHA University, 59,
27
Yatap-ro, Bundang-gu, Seongnam, 13496, Republic of Korea
28
Tel: +82-31-780-5384
29
Fax: +82-31-780-5476
30
E-mail: [email protected]
31
2
32
ABSTRACT
33
Objectives: The rapid diagnosis of tuberculosis (TB) is important for patient treatment and
34
infection control. Current molecular diagnostic techniques for TB have insufficient sensitivity
35
to detect samples with low bacterial loads. The sensitivity of molecular testing depends on
36
not only performance of assay technique but also nucleic acid extraction method. Here, we
37
have presented a novel approach using exosomal DNA (exoDNA) and droplet digital PCR
38
(ddPCR) platform to detect Mycobacterium tuberculosis DNA in clinical samples.
39
Methods: The ddPCR platform targeting IS6110 was evaluated in parallel using total DNA
40
and exoDNA. The clinical performance of ddPCR method was assessed with 190 respiratory
41
samples from patients with suspected pulmonary TB.
42
Results: Compared with mycobacterial culture, sensitivity and specificity of ddPCR were
43
61.5% (95% CI, 44.6%-76.6%) and 98.0% (95% CI, 94.3%-99.6%) using total DNA, and
44
76.9% (95% CI, 60.7%-88.9%) and 98.0% (95% CI, 94.3%-99.6%) using exoDNA
45
respectively. Among 15 culture-positive specimens with low concentrations of target
46
molecules (2~99 positive droplets with exoDNA), only 53.3% (8/15), 46.7% (7/15), and 26.7%
47
(4/15) of cases were detected using ddPCR with total DNA, real-time PCR with exoDNA, or
48
real-time PCR with total DNA.
49
Conclusions: Our platform using ddPCR and exoDNA has the potential to provide sensitive
50
and accurate methodology for TB diagnosis.
3
51
Introduction
52
Tuberculosis (TB), which is caused by Mycobacterium tuberculosis (MTB) infection,
53
is a critical health problem with high infectivity, morbidity and mortality. Mycobacterial
54
culture is time-consuming and acid-fast bacilli (AFB) smear has low sensitivity. Because of
55
the limitations of conventional diagnostic methods, nucleic acid amplification tests (NAATs)
56
have been usefully applied in rapid diagnosis of TB. To improve sensitivity, several
57
platforms adopt sensitive molecular techniques such as real-time polymerase chain reaction
58
(PCR), nested PCR, and isothermal transcription-mediated amplification. However, the
59
sensitivity of NAATs is still insufficient compared with conventional mycobacterial
60
culture [1].
61
Droplet digital PCR (ddPCR) is an emerging technology capable of absolute nucleic
62
acid quantification without using standard curves [2]. ddPCR technology uses a combination
63
of microfluidics and proprietary surfactant chemistries to divide PCR samples into water-in-
64
oil droplets [3]. These droplets support PCR amplification of single template molecules
65
using reagents and workflows similar to those widely used for real-time PCR applications [3].
66
Following PCR, each droplet is analyzed to assign positivity or negativity based on their
67
fluorescence amplitude, after which the concentration of target DNA in the original sample is
68
calculated [3]. These principles offer the advantages of precise and reproducible data without
69
being affected by PCR inhibitors in samples over real-time PCR [4]. Since 20,000 droplets
70
are generated for a sample, the theoretical detection limit of ddPCR is 0.005% (1/20,000) and
71
reported detection limits are 0.01~0.001% [3,5,6]. With low detection limits for ddPCR,
72
many applications have been developed for analyzing pathogens or cancer-derived mutations
73
present at low levels in clinical samples. Accordingly, we sought to evaluate the clinical 4
74
performance of ddPCR in detecting MTB in clinical samples.
75
Exosomes are small membrane vesicles (30–100 nm) secreted by many cell types
76
during physiological and pathological conditions [7]. They contain a variety of molecules
77
derived from the original cell, including DNA, RNA, proteins, and lipids [8]. Recent
78
evidence has indicated that the exosomal nucleic acids are more stable than other forms of
79
nucleic acids, probably by the protective effect of lipid bilayer coating [9]. Exosomes are
80
isolated in most biological fluids, such as serum, urine, and bronchoalveolar lavage (BAL)
81
fluid [10]. Such ubiquitous nature of exosomes in body fluids makes them ideal for use as
82
diagnostic biomarkers. Previous reports indicated that exosomes isolated from the BAL fluid
83
of MTB-infected mice or sera TB patients contain mycobacterial proteins [11].
84 85
In this study, we attempted a novel approach for TB diagnosis targeting exosomal DNA (exoDNA), using ddPCR.
86
87
Methods
88
Clinical specimens
89
The study protocol was approved by the Institutional Review Boards (IRBs) of the
90
participating institutions (Hallym University Kangnam Sacred Heart Hospital, Yonsei
91
University Gangnam Severance Hospital, and Bundang CHA General Hospital). Between
92
August 2018 to October 2019, a total of 174 consecutive subjects undergoing evaluation for
93
pulmonary TB were included in the study. In 174 consecutive samples, mycobacterial
94
culture-positive sample was 13% (n=23). Therefore, we selectively added 16 culture-positive
95
specimens to evaluate the efficiency of exoDNA and ddPCR for detecting MTB. The final 5
96
clinical diagnosis of TB was made by infectious physicians or respiratory physicians based
97
on integration of patients’ medical history, radiologic findings, and laboratory findings
98
(microbiological, molecular and immunological tests). The researchers retrospectively
99
reviewed the participant's medical records, including the final clinical diagnosis. All samples
100
were tested using AFB smear, mycobacterial culture, real-time PCR, and ddPCR. For non-
101
sterile specimens, an equal volume of 4% NaOH was added for emulsification and
102
decontamination. Then the mixture was transferred to a sterile tube and centrifuged for 20
103
minutes at 3,000 g. After performing the AFB smear procedure with an auramine-rhodamine
104
fluorescent stain, it was confirmed by Ziehl-Neelsen staining. For mycobacterial culture, all
105
decontaminated samples were inoculated into a mycobacterial growth indicator tube (MGIT
106
960 system; Becton Dickinson, Sparks, MD) and on 3% Ogawa agar (Shinyang, Seoul,
107
Republic of Korea) and cultured for 6 weeks.
108
DNA extraction
109
For analysis of total DNA, 100 µL aliquots of the decontaminated specimens were
110
re-suspended in DNA extraction buffer. Total DNA was extracted using QIAamp DNA mini
111
kit (QIAGEN, Hilden, Germany) following manufacturer’s instructions. For analysis of
112
exoDNA, we isolated exosomal fractions from 1 mL of respiratory specimen using
113
ExoQuick™
114
recommendations (System Biosciences Inc., Mountain View, CA, USA). Submitted
115
specimens were centrifuged at 3000 x g for 15 minutes to remove cells and cell debris.
116
Briefly, 1/4 volume of ExoQuick Solution was added to respiratory specimen, and samples
117
were refrigerated at 4 °C overnight. The mixture was centrifuged at 1500 × g for 30 min, and
118
supernatant was removed by aspiration. Pelleted fractions were re-suspended in nuclease-free
Exosome
Precipitation
Solution
6
according
to
the
manufacturer's
119
water. Subsequently, exosome-derived DNA was extracted using QIAamp DNA mini kits as
120
described above.
121
Real-time PCR
122
Real-time PCR was performed using the PowerCheck MTB/NTM Real-time PCR
123
assay (Kogene Biotech, Seoul, Korea), targeting the MTB-specific IS6110, following
124
manufacturer’s instructions. The 20 µL PCR mix was composed of 10 µL 2X Real-time PCR
125
Master Mix, 5 µL of primer/probe mix, and 5 µL total DNA. The analysis was done with
126
CFX96 real-time PCR detection system (Bio-Rad Laboratories, Hercules, CA, USA).
127
ddPCR
128
Digital PCR reactions were performed with a QX200 Droplet Digital PCR System
129
(Bio-Rad Laboratories, Hercules, CA, USA). The reactions were carried out using the
130
previously described primers and probes, targeting IS6110, and thermal cycling
131
conditions[12]. A no-template control was used in every ddPCR batch. The 20 µL PCR mix
132
was composed of 10 µL Bio-Rad Super mix TaqMan, 1 µL of each amplification
133
primer/probe mix, 5 µL DNA and 4 µL nuclease-free water. Results were analyzed with
134
QuantaSoft v.1.7.2 software (Bio-Rad Laboratories). This provided the number of positive
135
and negative droplets, as well as quantification of IS6110 of MTB, expressed as copies/mL of
136
ddPCR reaction. At least two positive droplets were required for a positive test result of the
137
ddPCR assay [12].
138
Limit of detection of real-time PCR and ddPCR assay
139
The analytical sensitivities of real-time PCR and the ddPCR assay were evaluated
140
using spiked samples with ATCC 25177. The limits of detection (LODs) for both PCR assays 7
141
were measured using five dilutions around LOD. For all concentrations, repetitive tests were
142
performed with 10 replicates in three different runs at each concentration, and LODs were
143
calculated by probit analysis for a 95% positive result.
144
Statistical analysis
145
SPSS version 24.0 software (SPSS Inc., Chicago, IL, USA) and MedCalc version
146
19.1 (MedCalc software, Mariakerke, Belgium) was used for statistical analyses. The chi-
147
squared test was used to analyze tabular data. A correlation study was performed with the
148
Spearman’s rank correlation procedure (r). All p-values were two-sided, and values less than
149
0.05 were considered statistically significant.
150
151
Results
152
Patient characteristics
153
A total of 190 respiratory specimens, including 151 (79.5%) sputa, 31 (16.3%)
154
bronchial washings, and 8 (4.2%) bronchoalveolar lavage (BAL) fluids, were included in this
155
study. The demographic data for 190 subjects are listed in Table 1. Forty four patients (23.2%)
156
were finally diagnosed with pulmonary tuberculosis.
157
Considering mycobacterial culture as a reference method, sensitivity and specificity
158
were 35.9% (95% CI, 21.2%-52.8%) and 98.0% (95% CI, 94.3%-99.6%) for AFB smear
159
(Table S1) Compared with final clinical diagnosis, sensitivity and specificity were 34.1% (95%
160
CI, 20.5%-49.9%) and 98.6% (95% CI, 95.1%-99.8%) for AFB smear and 88.6% (95% CI,
161
75.4%-96.2%) and 100% (95% CI, 97.5%-100.0%) for mycobacterial culture, respectively 8
162
(Table 3).
163
Results of real-time PCR using exosomal DNA (exoDNA) and total DNA
164
Probit analysis of 95% positivity in 10 replicates in three different runs at five
165
concentrations showed MTB had a LOD of 13.8 copies/reaction (Table S2). Real-time PCR
166
using exoDNA and total DNA yielded 24 positive suspects and 21 positive suspects,
167
respectively, and three of each were culture negative (Table 2). All three of these subjects
168
also showed positive results in all other PCR assays, and 1 patient had a positive result by
169
AFB smear. Based on culture results, the sensitivities of the real-time PCR assays with both
170
exoDNA and total DNA were 53.9% (95% CI, 37.2-69.9) and 46.2% (95% CI, 30.1%-
171
62.8%), respectively (Table 2). Compared with final clinical diagnosis, the sensitivities of
172
both exoDNA and total DNA were 54.6% (95% CI, 38.9%–69.6%) and 47.7% (95% CI,
173
32.5%-63.3%), respectively (Table 3). The specificities of exoDNA and total DNA according
174
to culture results were both 98.0% (95% CI, 94.3%-99.6%) and according to the final clinical
175
diagnosis were both 100% (95% CI, 97.5%-100.0%) (Table 2, 3).
176
Results of ddPCR using total DNA
177
We found that the MTB LOD with ddPCR was 2.3 copies/reaction (Table S3). The
178
ddPCR yielded 27 positive suspects, 3 of which were culture negative (Table 2). Of these 3
179
suspects, 2 had real-time PCR positive results; only 1 patient had a positive result by AFB
180
smear. Of the 163 ddPCR negative suspects, 15 were culture positive. Based on culture
181
results, the sensitivity and specificity of the ddPCR assays with total DNA were 61.5% (95%
182
CI, 44.6-76.6) and 98.0% (95% CI, 94.2–99.6), respectively (Table 2). Compared with final
183
clinical diagnosis, the sensitivity and specificity of the ddPCR with total DNA were 61.4% 9
184
(95% CI, 45.5-75.6), and 100% (95% CI, 97.5%–100%), respectively (Table 3).
185
Results of ddPCR using exosomal DNA (exoDNA)
186
The ddPCR with exoDNA yielded 33 positive suspects, 3 of whom were culture
187
negative (Table 2). Of these 3 suspects, all 3 had real-time PCR positive results; only 1
188
patient had a positive result by AFB smear. Of the 157 ddPCR negative suspects, 9 were
189
culture positive. Based on culture results, the sensitivity and specificity of the ddPCR with
190
exoDNA for detecting MTB were 76.9% (95% CI, 60.7-88.9) and 98.0% (95% CI, 90.2–
191
97.6), respectively (Table 2). Compared with final clinical diagnosis, the sensitivity and
192
specificity of the ddPCR with exoDNA were 75.0% (95% CI, 59.7-86.8) and 100% (95% CI,
193
97.5–100), respectively (Table 3).
194
Receptor operating characteristics (ROC) curve analysis was performed to evaluate
195
diagnostic accuracy of ddPCR to predict MTB culture positivity. The area under the ROC
196
curve was 0.80 (95% CI, 0.70-0.89) and 0.88 (95% CI, 0.79-0.96) with the total DNA and
197
exoDNA, respectively (Fig. S1).
198
Measured IS6110 copy numbers of exoDNA using ddPCR were correlated with those
199
of total DNA (p<0.01, r=0.827, Spearman correlation) (Fig. 1A). The Bland-Altman analysis
200
of ddPCR results using total DNA and exoDNA showed the mean difference of 0.2 ± 0.94
201
log10 with 95% Limits of Agreement (LoA) ranging from -1.7 to 2.0 in TB culture-positive
202
samples. (Fig.1B). Notably, in low concentration samples (< 100 positive droplets in
203
exoDNA) with positive culture results, assays with exoDNA showed improved sensitivity
204
over assays with total DNA (Table 4, Fig. S2). The Bland-Altman analysis indicated that the
205
difference of measured TB copies between exoDNA and total DNA was 0.45 ± 0.91 log10 10
206
(LoA: -1.3 to 2.2), and TB copies were 79% higher on average in exoDNA than in total DNA
207
(Fig. S2). And, the detection rates of ddPCR with total DNA, real-time PCR with exoDNA,
208
real-time PCR with total DNA were 53.3%, 46.7%, and 26.7% of cases, respectively,
209
compared with ddPCR with exoDNA (Table 4).
210
211
Discussion
212
The ddPCR assay using exoDNA had higher sensitivity than ddPCR using total DNA
213
in our study. Therefore, exosomes may serve as a feasible target for detection of pathogen-
214
derived nucleic acid. To our knowledge, this is the first study reporting detection of MTB
215
DNA from exosomes using clinical specimens isolated from TB patients. The detection rate
216
in low concentration samples was significantly higher when using exoDNA than total DNA
217
in PCR-based assays (Table 4). The results could be explained by stability and abundance of
218
exoDNA in clinical samples. The stability is accomplished by the lipid bilayers which protect
219
exosomal nucleic acids from the extracellular environment [13]. The number of exosome
220
reported by the previous study is approximately 1.15 × 1010 in 10 million Tb infected cells,
221
which cultured for three days [14].
222
In this study, three culture negative specimens showed positive results with ddPCR
223
using exoDNA. These cases were confirmed pulmonary tuberculosis based on radiologic
224
findings and/or real-time PCR, and treated with anti-TB medication. Although mycobacterial
225
culture has long been the conventional gold standard test for TB diagnosis, culture results can
226
be negative in mild TB infection [15], latent TB infection and early sub-clinical disease [16].
227
In addition, the decontamination procedure of sputum prior to culture dilutes the sample and 11
228
leads to the loss of viability [17]. Because identification of these groups is important for
229
infection control, ddPCR could provide fast and more accurate results.
230
In this study, the false negative results both in ddPCR and real-time PCR might have
231
been due to IS6110 copy number variation [18]. Factors that affect NAAT sensitivity include
232
the quality of sputum [19] and clinician bias with regard to diagnostic approaches.
233
There are several limitations in this study that could be addressed in future studies.
234
First, we have the potential to study fluctuation in exosome concentration of TB infection.
235
The biogenesis and release of exosomes are affected by various activating signals and cellular
236
stress that reflect changes in the environmental conditions [20]. Previous study has suggested
237
that infection induces exosome secretion and that this is correlated with bacterial burden [14].
238
However, the mechanisms regulating exosome release have not been well elucidated to date.
239
Further investigation is required to determine controlling factors. Second, the pre-analytical
240
phase is an important step with impact on results downstream. In clinical settings, most
241
requested samples are analyzed by the batch as in this study. So, samples are usually frozen
242
and stored until analysis. Exosomes in plasma, urine, and saliva seem stable during a freeze-
243
thaw cycle and storage [21]. However, an optimal protocol is needed specific to the type of
244
specimens and the downstream analysis. Third, altering the isolation methods of exosome
245
from specimen may improve the sensitivity of this assay. Here we used commercial kits
246
showing higher extraction efficiency than ultracentrifugation [22] and better performance
247
than ddPCR using total DNA. Laboratory clinicians and researchers should be aware of the
248
advantages and disadvantages of each method and compare and validate before clinical
249
application of exoDNA. In addition, the results of the current study were obtained from a
250
small sample size. Thus, further studies with larger sample size in multicenter settings are 12
251
required to validate these results.
252
The ddPCR using exoDNA could be helpful in clinical conditions where the bacterial
253
load in the specimen is low such as in certain patients with extrapulmonary TB [23],
254
children [24], and patients with HIV-infection [25]. The increased sensitivity of ddPCR is
255
applicable in other areas, such as a method of detection of small numbers of drug resistant TB
256
or in determining the therapeutic effect of the treatment of TB. Serum or urine derived
257
exosomes have recently emerged as non-invasive diagnostic and prognostic biomarkers in
258
cancer [26]. These biological fluids could be alternative specimens for NAAT. This
259
supposition requires future verification with minimally invasive samples.
260
This is the first study reporting detection of MTB DNA from exosomes using clinical
261
specimens isolated from TB patients in combination with ddPCR assay to date. In conclusion,
262
ddPCR with exoDNA has the potential to provide a highly sensitive and accurate method for
263
diagnosis of MTB infection.
264
265
Acknowledgements
266
We thank the dedicated physicians at each hospital where this multicenter study was
267
conducted (YS Sim and J Lee from Hallym University Kangnam Sacred Heart Hospital, EK
268
Kim and M Kim from CHA Bundang Medical Center, and HJ Park and KH Lee from Yonsei
269
University Gangnam Severance Hospital). This study was supported by a grant of the
270
National Research Foundation of Korea (NRF-2019R1C1C1010916).
271
13
272
Transparency declaration
273
All authors report no competing interests or patents.
274
275
References
276
[1] Boehme CC, Nicol MP, Nabeta P, Michael JS, Gotuzzo E, Tahirli R, et al. Feasibility,
277
diagnostic accuracy, and effectiveness of decentralised use of the Xpert MTB/RIF test for
278
diagnosis of tuberculosis and multidrug resistance: a multicentre implementation study.
279
Lancet 2011;377:1495-505.
280
[2] Pinheiro LB, Coleman VA, Hindson CM, Herrmann J, Hindson BJ, Bhat S, et al.
281
Evaluation of a droplet digital polymerase chain reaction format for DNA copy number
282
quantification. Anal Chem 2012;84:1003-11.
283
[3] Hindson BJ, Ness KD, Masquelier DA, Belgrader P, Heredia NJ, Makarewicz AJ, et al.
284
High-Throughput Droplet Digital PCR System for Absolute Quantitation of DNA Copy
285
Number. Anal Chem 2011;83:8604-10.
286
[4] Dingle TC, Sedlak RH, Cook L, Jerome KR. Tolerance of droplet-digital PCR vs real-
287
time quantitative PCR to inhibitory substances. Clin Chem 2013;59:1670-2.
288
[5] Azuara D, Ginesta MM, Gausachs M, Rodriguez-Moranta F, Fabregat J, Busquets J, et al.
289
Nanofluidic Digital PCR for KRAS Mutation Detection and Quantification in
290
Gastrointestinal Cancer. Clin Chem 2012;58:1332-41.
291
[6] Taly V, Pekin D, Benhaim L, Kotsopoulos SK, Le Corre D, Li X, et al. Multiplex
292
picodroplet digital PCR to detect KRAS mutations in circulating DNA from the plasma of
293
colorectal cancer patients. Clin Chem 2013;59:1722-31. 14
294
[7] Schorey JS, Cheng Y, Singh PP, Smith VL. Exosomes and other extracellular vesicles in
295
host-pathogen interactions. EMBO Rep 2015;16:24-43.
296
[8] Ruivo CF, Adem B, Silva M, Melo SA. The Biology of Cancer Exosomes: Insights and
297
New Perspectives. Cancer Res 2017;77:6480-8.
298
[9] Kahlert C, Melo SA, Protopopov A, Tang J, Seth S, Koch M, et al. Identification of
299
double-stranded genomic DNA spanning all chromosomes with mutated KRAS and p53
300
DNA in the serum exosomes of patients with pancreatic cancer. J Biol Chem 2014;289:3869-
301
75.
302
[10] Yanez-Mo M, Siljander PR, Andreu Z, Zavec AB, Borras FE, Buzas EI, et al. Biological
303
properties of extracellular vesicles and their physiological functions. J Extracell Vesicles
304
2015;4:27066.
305
[11] Kruh-Garcia NA, Wolfe LM, Chaisson LH, Worodria WO, Nahid P, Schorey JS, et al.
306
Detection of Mycobacterium tuberculosis peptides in the exosomes of patients with active
307
and latent M. tuberculosis infection using MRM-MS. PloS one 2014;9:e103811.
308
[12] Ushio R, Yamamoto M, Nakashima K, Watanabe H, Nagai K, Shibata Y, et al. Digital
309
pcr assay detection of circulating mycobacterium tuberculosis DNA in pulmonary
310
tuberculosis patient plasma. Tuberculosis (Edinb) 2016;99:47-53.
311
[13] Taylor DD, Gercel-Taylor C. MicroRNA signatures of tumor-derived exosomes as
312
diagnostic biomarkers of ovarian cancer. Gynecol Oncol 2008;110:13-21.
313
[14] Singh PP, Smith VL, Karakousis PC, Schorey JS. Exosomes isolated from mycobacteria-
314
infected mice or cultured macrophages can recruit and activate immune cells in vitro and in
315
vivo. J Immunol 2012;189:777-85.
316
[15] Levy H, Feldman C, Sacho H, van der Meulen H, Kallenbach J, Koornhof H. A
317
reevaluation of sputum microscopy and culture in the diagnosis of pulmonary tuberculosis. 15
318
Chest 1989;95:1193-7.
319
[16] Salgame P, Geadas C, Collins L, Jones-López E, Ellner JJ. Latent tuberculosis infection–
320
revisiting and revising concepts. Tuberculosis (Edinb) 2015;95:373-84.
321
[17] Pheiffer C, Carroll N, Beyers N, Donald P, Duncan K, Uys P, et al. Time to detection of
322
Mycobacterium tuberculosis in BACTEC systems as a viable alternative to colony counting.
323
Int J Tuberc Lung Dis 2008;12:792-8.
324
[18] Sankar S, Kuppanan S, Balakrishnan B, Nandagopal B. Analysis of sequence diversity
325
among IS6110 sequence of Mycobacterium tuberculosis: possible implications for PCR based
326
detection. Bioinformation 2011;6:283-5.
327
[19] Pan X, Yang S, Deighton MA, Qu Y, Hong L, Su F. A comprehensive evaluation of
328
Xpert MTB/RIF assay with bronchoalveolar lavage fluid as a single test or combined with
329
conventional assays for diagnosis of pulmonary tuberculosis in china: a two-center
330
prospective study. Front Microbiol 2018;9:444.
331
[20] Zhang X, Yuan X, Shi H, Wu L, Qian H, Xu W. Exosomes in cancer: small particle, big
332
player. J Hematol Oncol 2015;8:83.
333
[21] Yuana Y, Boing AN, Grootemaat AE, van der Pol E, Hau CM, Cizmar P, et al. Handling
334
and storage of human body fluids for analysis of extracellular vesicles. J Extracell Vesicles
335
2015;4:29260.
336
[22] Li P, Kaslan M, Lee SH, Yao J, Gao Z. Progress in Exosome Isolation Techniques.
337
Theranostics 2017;7:789-804.
338
[23] Denkinger CM, Schumacher SG, Boehme CC, Dendukuri N, Pai M, Steingart KR. Xpert
339
MTB/RIF assay for the diagnosis of extrapulmonary tuberculosis: a systematic review and
340
meta-analysis. Eur Respir J 2014;44:435-46. 16
341
[24] Perez-Velez CM, Marais BJ. Tuberculosis in children. N Engl J Med. 2012;367:348-61.
342
[25] Getahun H, Harrington M, O'Brien R, Nunn P. Diagnosis of smear-negative pulmonary
343
tuberculosis in people with HIV infection or AIDS in resource-constrained settings:
344
informing urgent policy changes. Lancet 2007;369:2042-9.
345
[26] Huang T, Deng C-X. Current progresses of exosomes as cancer diagnostic and
346
prognostic biomarkers. Int J Biol Sci 2019;15:1.
347
17
1
Table 1. The demographic data for 190 subjects Pulmonary tuberculosis
Other diagnosis
(n=44)
(n=146)
Male
21 (47.7)
84 (57.5)
Female
23 (52.3)
62 (42.5)
60 (53-75)
68 (54-79)
7 (15.9)
25 (17.1)
Sputum
32 (72.7)
119 (81.5)
Bronchial washing
7 (15.9)
24 (16.4)
BAL
5 (11.4)
3 (2.1)
Characteristics
Sex (%)
Age, median (IQR) Previous tuberculosis history (%) Referred respiratory specimens (%)
IQR, interquartile range; BAL, bronchoalveolar lavage
2
Table 2. Analytical performance of assays for detection of Mycobacterium tuberculosis depending on culture results Samples
3
Assays
Culture positive Culture negative Sensitivity (95% CI) Specificity (95% CI) PCR + PCR - PCR + PCR qPCR with total DNA 18 21 3 148 46.2% (30.1%-62.8%) 98.0% (94.3%-99.6%) qPCR with exoDNA 21 18 3 148 53.9% (37.2%-69.9%) 98.0% (94.3%-99.6%) All ddPCR with total DNA 24 15 3 148 61.5% (44.6%-76.6%) 98.0% (94.3%-99.6%) ddPCR with exoDNA 30 9 3 148 76.9% (60.7%-88.9%) 98.0% (94.3%-99.6%) qPCR with total DNA 12 2 1 2 85.7% (57.2%-98.2%) 66.7% (9.4%-99.2%) Smear qPCR with exoDNA 13 1 1 2 92.9% (66.1%-99.8%) 66.7% (9.4%-99.2%) positive ddPCR with total DNA 13 1 1 2 92.9% (66.1%-99.8%) 66.7% (9.4%-99.2%) (n=17) ddPCR with exoDNA 14 0 1 2 100% (76.8%-100.0%) 66.7% (9.4%-99.2%) qPCR with total DNA 6 19 2 146 24.0% (9.4%-45.1%) 98.7% (95.3%-99.8%) Smear qPCR with exoDNA 8 17 2 146 32.0% (15.0%-53.5%) 98.7% (95.3%-99.8%) negative ddPCR with total DNA 11 14 2 146 44.0% (24.4%-65.1%) 98.7% (95.3%-99.8%) (n=173) ddPCR with exoDNA 16 9 2 146 64.0% (42.5%-82.0%) 98.7% (95.3%-99.8%) +, positive; -, negative; CI, confidence intervals; qPCR, real-time PCR; ddPCR, droplet digital PCR; exoDNA, exosomal DNA
4
Table 3. Clinical performance of assays for detection of Mycobacterium tuberculosis depending on confirmed diagnosis Method AFB smear qPCR with total DNA qPCR with exoDNA ddPCR with total DNA ddPCR with exoDNA
Results Pos Neg Pos Neg Pos Neg Pos Neg Pos
PTB (n=44) 15 29 21 23 24 20 27 17 33
Non-TB (n=146) 2 144 0 146 0 146 0 146 0
Neg
11
146
Sensitivity (95% CI)
Specificity (95% CI)
34.1% (20.5%-49.9%) 98.6% (95.1%-99.8%)
Accuracy (95% CI) 83.7% (77.7%-88.6%)
47.7% (32.5%-63.3%) 100% (97.5%-100.0%) 87.9% (82.4%-92.2%) 54.6% (38.9%-69.6%) 100% (97.5%-100.0%) 89.5% (84.2%-93.5%) 61.4% (45.5%-75.6%) 100% (97.5%-100.0%) 91.1% (86.1%-94.7%) 75.0% (59.7%-86.8%) 100% (97.5%-100.0%) 94.2% (89.9%-97.1%)
5
Pos 39 0 88.6% (75.4%-96.2%) 100% (97.5%-100.0%) 97.4% (94.0%-99.1%) Neg 5 146 PTB, pulmonary tuberculosis; Non-TB, patients diagnosed other than tuberculosis; Pos, positive; Neg, negative; CI, confidence intervals;
6
qPCR, real-time PCR; ddPCR, droplet digital PCR; exoDNA, exosomal DNA.
Culture
7
8
9
Table 4. Results of both ddPCR and real-time PCR results using two sources of DNA in low concentration samples ddPCR with qPCR with Mycobacterial exoDNA total DNA culture (No. of positive (No. of positive exoDNA total DNA droplets) droplets) Pos (2) Neg (1) Neg Neg MTB Pos (3) Neg (1) Neg Neg MTB Pos (3) Pos (9) Neg Neg MTB Pos (4) Pos (24) Neg Pos MTB Pos (4) Pos (2) Neg Neg MTB Pos (4) Neg (0) Neg Neg MTB Pos (7) Pos (2) Neg Neg MTB Pos (8) Neg (1) Neg Neg MTB Pos (15) Neg (1) Pos Neg MTB Pos (20) Neg (0) Pos Neg MTB Pos (52) Neg (0) Pos Neg MTB Pos (69) Pos (6) Pos Pos MTB Pos (93) Pos (143) Pos Pos MTB Pos (94) Pos (817) Pos Pos MTB Pos (99) Pos (7) Pos Neg MTB Positive rate (%) 53.3 46.7 26.7 100 Pos, positive; Neg, negative; qPCR, real-time PCR; ddPCR, droplet digital PCR; exoDNA, exosomal DNA. 10
11 12