Tuberculosis 107 (2017) 73e79

Contents lists available at ScienceDirect

Tuberculosis
journal homepage: http://intl.elsevierhealth.com/journals/tube

Differential expression of long non-coding RNAs in patients with

tuberculosis infection
Jianan He a, b, 1, Qingye Ou c, 1, Chunxiao Liu a, Lei Shi a, Chunzhong Zhao a, Yunqing Xu a,
Siu Kai Kong d, Jacky Fong Chuen Loo d, e, *, Boan Li f, **, Dayong Gu a, b, ***
a
Central Laboratory of Health Quarantine, Shenzhen International Travel Health Care Center, Shenzhen Entry-Exit Inspection and Quarantine Bureau,
Shenzhen, 518033, PR China
b
Shenzhen Academy of Inspection and Quarantine, Shenzhen, 518010, PR China
c
Zhuhai Center for Chronic Disease Control, Zhuhai, 519000, PR China
d
Biochemistry Programme, School of Life Sciences, The Chinese University of Hong Kong, Hong Kong Special Administrative Region
e
Department of Biomedical Engineering, The Chinese University of Hong Kong, Hong Kong Special Administrative Region
f
302 Military Hospital of China, Beijing, 100039, PR China

a r t i c l e i n f o a b s t r a c t

Article history: Tuberculosis (TB) remains a major worldwide health problem and has caused millions of deaths in the
Received 22 January 2017 past few years. Current diagnostic methods, such as sputum smear microscopy and sputum culture, are
Received in revised form time-consuming and cannot prevent the rapid spreading of TB during the diagnostic period. In this
20 August 2017
connection, detecting biomarkers specific to TB at molecular level in plasma of patients will provide a
Accepted 21 August 2017
rapid means for diagnosis. In this study, we first evaluated the differential expression of the long non-
coding RNAs (lncRNAs) in the plasma from patients with TB (TB positive), community acquired pneu-
Keywords:
monia (CAP) and healthy individuals (CG) using lncRNA microarray scanning. It was found that there
Tuberculosis
Long non-coding RNA
were 2116 specific lncRNAs differentially expressed in the TB positive samples (1102 up-regulated and
Microarray 1014 down-regulated), which accounted for 6.96% of total lncRNAs. Twelve differentially expressed
Biomarkers lncRNAs discovered in microarray were subsequently validated by using real-time quantitative PCR (RT-
Plasma qPCR). Two lncRNAs (ENST00000354432 and ENST00000427151) were further validated with more
Tuberculosis samples. These results suggested the expression level of lncRNAs and the two validated
lncRNAs in plasma could be the potential molecular biomarkers for the rapid diagnosis of Tuberculosis.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction
Tuberculosis (TB) is caused by Mycobacterium tuberculosis,
resulting in millions of deaths in 2016, and has become a consid-
erable threat to public health worldwide [1]. More than five mil-
lions cases of active TB infection were reported in the population
aged 15 and above have been reported and the number of newly
reported cases is increasing at the rate of one million per year in
Asia [2]. The improper and inadequate treatment of TB, including
the irrational use of anti-tuberculous drugs, has led to the devel-
opment of multidrug resistance (MDR). One third of the world's
MDR-TB cases are in Asia and the number of MDR-TB cases is still
increasing [1]. Laboratory diagnosis of TB is the key for downstream
TB treatment. Sputum smear microscopy and sputum culture are
the current gold standard for TB diagnosis recommended by the
World Health Organization. Sputum smear microscopy is

Abbreviation: CAP, Community Acquired Pneumonia; CG, Control Group; Lnc RNA, Long non-coding RNA; LTBI, latent tuberculosis infection; MDR, multidrug resistance;
TB, Tuberculosis; RT-qPCR, Real-time quantitative polymerase chain reaction.
* Corresponding author. Department of Biomedical Engineering, The Chinese University of Hong Kong, Hong Kong Special Administrative Region.
** Corresponding author. 302 Military hospital of China, Beijing, 100039, PR China.
*** Corresponding author. Central Laboratory of Health quarantine, Shenzhen International Travel Health Care Center, Shenzhen Entry-exit Inspection and Quarantine
Bureau, Shenzhen, 518033, PR China.
E-mail addresses: jackyfcloo@cuhk.edu.hk (J.F.C. Loo), lba@263.net (B. Li), wanhood@163.com (D. Gu).
1
These authors contributed to the work equally and should be regarded as co-first authors.

http://dx.doi.org/10.1016/j.tube.2017.08.007
1472-9792/© 2017 Elsevier Ltd. All rights reserved.
74 J. He et al. / Tuberculosis 107 (2017) 73e79
convenient and inexpensive, but is limited by its inherent low
specificity (14e47%) and low sensitivity (at least 103 bacteria cfu/
mL sample) that cause difficulty in early TB diagnosis. On the other
hand, the sputum culture of M. tuberculosis can effectively distin-
guish the dead from the live bacteria, but it is labor-intensive and
requires laboratory skilled workers. It is also time-consuming (2e8
weeks) as a result of the long doubling time of the M. tuberculosis
[3]. These hinders its application for an early diagnosis.
Recently, TB diagnosis based on the molecular biomarker has
been developed and such detection could be a solution for rapid
diagnosis of TB with high sensitivity and specificity [4]. Long non-
coding RNAs (lncRNAs) are a subset of non-coding RNAs with a
size longer than 200 nucleotides, accounting for about 80% of non-
coding RNAs [5]. Many studies have shown that lncRNAs partici-
pated in the regulation of many biological processes, such as gene
expression at the transcriptional, post-transcriptional and epige-
netic level [6,7]. Also, recent studies have demonstrated that the
dysregulated expression of lncRNA in various tissues is associated
with a variety of diseases [8e10]. For example, Hui et al. reported a
large number of lncRNAs were differentially expressed in patients
with chronic obstructive pulmonary [11]. These findings suggest
that lncRNA can serve as a promising biomarker for disease
detection. However, whether this approach can be applied to the
diagnosis of TB remains to be investigated. Therefore, we aim at
identifying the lncRNAs in the plasma of hosts in response to TB by
lncRNA microarray analyses using the corresponding samples from
patients with community acquire pneumonia (CAP) infection and
health individuals as control group (CG). The finding from the
screening of lncRNA biomarkers is expected to be useful for the
development of next generation method for TB diagnosis.
2. Materials and methods
2.1. Plasma samples of TB, CAP and CG
The use of human samples in this study has been ethically
approved by the hospital ethics committee. Written informed
consent was obtained from healthy individuals as CG and patients
before sample and data collection. Plasma and sputum samples
from patients diagnosed as TB positive (active pulmonary TB) by
sputum smear microscopy and sputum culture in Zhuhai Center for
Chronic Disease Control during the period from October, 2014 to
November, 2015 were collected, while plasma samples from pa-
tients with CAP and healthy subjects during the same period were
collected respectively as the differential diagnosis group and CG.
Regarding the CAP patients, they shared similar symptoms similar
to TB infection (Include a cough with pus-like phlegm, fever and
chills, loss of appetite, shortness of breath) during hospitalization. It
was later confirmed that the CAP patients did not have pulmonary
TB by standard TB tests including sputum smear and TB culture. No
other disease was found in these CAP patients and they had
recovered eventually. Meanwhile, The Control group samples were
selected from those who had no clinical symptoms and normal
physical examination, and specific tests for pulmonary TB, hepatitis
B and tumor markers were negative. The CAP and healthy CG were
all latent tuberculosis infection (LTBI) negative. Each group con-
sisted of 53 samples and 3 samples were used for lncRNA micro-
array analysis and the other 50 samples for real-time quantitative
PCR (RT-qPCR).
2.2. LncRNA extraction and preparation of cRNA for microarray
The total RNA, including lncRNA, was extracted using TRIzol®
Reagent (Life Technologies) and RNeasy Mini Kit (Qiagen) according
to the manufacturer's protocol. Samples of isolated RNA were
stored at
70  C until use. The synthesis and the labelling of cRNA
were performed with Quick Amp Labeling Kit, One-Color (Agilent).
Briefly, the total RNA was amplified and transcribed into fluores-
cent cRNA with the random priming method. Fluorescence labeling
of each sample was performed using 1 mg of RNA. The labeled
cRNAs were then purified with RNeasy Mini Kit (Qiagen) according
to the manufacturer's protocol.
2.3. Analysis of quality of lncRNA and cDNA for microarray
Prior to microarray detection, the quality of RNA extraction from
plasma samples, in terms of the quantity and the purity, were
assessed by the optical density ratio of OD260/OD280 using spec-
trometry (Nanodrop-ND 1000). The result of spectrophotometer
showed that the ratios of total RNAs all lay between 1.6 and 1.8,
indicating a good purity of RNA. In addition, the total quantity of
RNA extracted from different samples was 0.5 mg, which was
sufficient for downstream labelling, for all samples. The purification
quality and the activity of cRNA were also validated using spec-
trometry (NanoDrop ND-1000). The activity of labeled RNA (spe-
cific activity) was calculated using the following formula: specific
activity ¼ (pmol dyes per mg cRNA) ¼ [(pmol per mL dye)/(mg per mL
cRNA)]. Hybridization was performed using Agilent Gene Expres-
sion Hybridization Kit (Agilent) according to the manufacturer's
protocol. The activity results on downstream labeling were calcu-
lated. Monochromatic labeling was performed using cy3 dye, the
total quantity of labeled RNA was 1.65 mg and the specific activity
was 9.0 pmol Cy3 per mg cRNA for all samples (Supplementary
Table 1). Based on these output, the fluorescence labeling of RNA
was found to be successful and these were latter used in down-
stream microarray hybridization and analysis.
2.4. Microarray image acquisition and analysis
Arraystar Human LncRNA Microarray V3.0 with 8  60 k format,
which simultaneously detected 30586 lncRNAs and 26109 protein
coding transcripts, was used for the microarray analysis of lncRNA.
Microarray scanning was performed using Agilent Microarray
Scanner (Agilent p/n G2565BA). The microarray raw data and im-
ages were acquired using Agilent Feature Extraction software
(v11.0.1.1). Quantile normalization and data processing of raw data
were performed using GeneSpring GX v12.1 software (Agilent
Technologies) according to the software instruction. The micro-
array data have also been deposited to GEO (GEO accession num-
ber: GSE101805).
2.5. Real-time quantitative PCR for lncRNA validation
RT-qPCR was performed using ABI7500 Real-time PCR machine
(Life Technologies) to validate the expression pattern of lncRNA
from microarray analysis. The sequence of primers targeting at 12
selected lncRNAs, synthesized from Takara, were shown in Table 1.
GADPH was selected as the internal reference gene to normalize the
expression level of each lncRNA in individual sample. The thermal
cycling conditions were set as the following: the initial activation of
Taq polymerase at 95  C for 10 min, 40 cycles of PCR amplification at
95  C for 10 s, annealing/elongation at 60  C for 6 s and then at 72  C
for 30 s. The relative expression of lncRNA was evaluated by the
2
DDCt formula, in which DCt ¼ Ct lncRNA - Ct GAPDH RNA,
DDCt ¼ DCtpatient e DCtcontrol.
2.6. Statistical analysis
The statistical analysis of results was performed using IBM SPSS
Statistics 22 software. The comparison of the mean between two
 J. He et al. / Tuberculosis 107 (2017) 73e79
Table 1
The sequences of specific lncRNA primers used in this study.
Gene name Primers 50 e 30
(F¼Forward; R ¼ Reverse)
GAPDH(HUMAN) F: GGGAAACTGTGGCGTGAT
R: GAGTGGGTGTCGCTGTTGA
ENST00000354432 F: GCTTGGGGAAAATGAACCGC
R: ACCCATTAAGCAACCCAGCC
uc004cov.4 F: ACACTTATCCCCACCTTGGC
R: AGGAAGTATGTGCCTGCGTT
NR_044997 F: ACTCAGGACACCTCCCGTTG
R: CGCCTATACGCCTCGTTTGA
ENST00000427151 F: CTGGCAGGCTGAATACAAGA
R: GGGACCTAACACAAAACTCGC
ENST00000442037 F: AGAAGCGGGTCTGTTTCTTTACT
R: CGATGGTGTCTTTGATGTTGG
ENST00000546607 F: CAAAAGTTCACACCTGGGAT
R: GGTTGCTGAGGAGTAAGGC
ENST00000428188 F: GCTGTGCTAATGTTTGGGGC
R: CGAATAGTGTGGAAAATCCTGC
TCONS_00019972 F: TTCCTGAACCACTTCCACCCT
R: GTCTGCTTTCTCATCCCATCTT
TCONS_00004316 F: ATCTGAAGGGGAATGGGC
R: GGCAAGTTTTTGTTCCTCCAC
ENST00000568137 F: GCTTACTGCTCGGTGTCTCC
R: GGTTCATCTCCAACTCAGGC
ENST00000584688 F: CCCTCCACCATCACCCAAT
R: GGCAAACCTCTGCTCACCAC
TCONS_00019584 F: GTGTCCTCCCAAGGTTTCG
R: AGCAGATTCCCCAGAGTCAG
samples was performed using t-test, and the comparison of the
mean among multiple samples was performed using the analysis of
variance. A p-value < 0.05 was considered statistically significant.
c2 test was used for enumeration data. Wilcoxon test was used for
rank data.
Fig. 1. Volcano plot to show the lncRNA gene variation between TB, CAP and CG samples. The horizontal axis represents fold changes (log2 conversion) and the longitudinal axis
represents the p value (log10 conversion). The signal values were distributed in the corresponding area after the data normalization. The squares in red are the lncRNA distribution
using a filtering criteria of fold change >2 and p < 0.05. (A) Tuberculosis and CG. (B) CAP and CG. (For interpretation of the references to colour in this figure legend, the reader is
referred to the web version of this article.)
75
3. Results and discussion
Amplicon 3.1. Apparent analysis of microarray results on lncRNA pattern in TB
Size (bp) samples
299
The comparison between the tuberculosis group and CG was
79 shown in a volcano plot, in Fig. 1, where the horizontal axis rep-
resents fold changes (log2 conversion) and the longitudinal axis
77
represents the p value (log10 conversion), and the signal values
99 were distributed in the corresponding areas after normalization.
Fold change >2 and p < 0.05 were used as the filtering criteria. It
87
can be seen that the lncRNAs in the red area were both significantly
259
and differentially expressed (Fig. 1).
87
3.2. Differential expression analysis to select lncRNA biomarkers for
205
tuberculosis diagnosis
174
Hierarchical cluster analysis is able to group similarly expressed
75 lncRNAs based on the different expression quantity of each group,
and the samples of the same cluster may possess similar biological
280
properties. Therefore, the difference in lncRNA expression profile
146 can be observed easily and the homogeneity of samples within each
group can be tested at the same time. Hierarchical clustering was
95 performed to show distinguishable lncRNA expression patterns
among different samples. The homogeneity of expression within
each group was grouped and shown in Fig. 2. Firstly, the expression
quantity of lncRNA was compared between TB group and CG, and
the lncRNA was considered to be differentially expressed when fold
change >2 and p < 0.05. With these criteria, 2870 lncRNAs were
differentially expressed in TB (1534 up-regulated and 1336 down-
regulated), which accounted for 9.39% of total lncRNAs tested.
Among the up-regulated lncRNAs, 6 lncRNAs had a fold change
76 J. He et al. / Tuberculosis 107 (2017) 73e79

Fig. 2. Cluster analysis between TB, CAP and CG. lncRNA profile on microarray was performed and the results were shown using cluster analysis to reveal the fold change of up-
regulated and down-regulated lncRNAs for a comparison between (A) TB and CG, (B) CAP and CG.

above 500 and 31 lncRNAs had a fold change between 100 and 500. Table 2
Among the down-regulated lncRNAs, 3 lncRNAs had a fold change Up-regulated specific lncRNAs differentially expressed in TB.

above 500 and 11 lncRNAs had a fold change between 100 and 500. Sequence name Probe Name FDR Fold Change
The most up-regulated one was ENST00000494340 (fold change: ENST00000354432 ASHGA5P032248 0.002753082 601
1203), followed by TCONS_00006916 (fold change: 776) and the TCONS_00014296 ASHGA5P057197 0.002073347 537
most down-regulated one was ENST00000563895 (fold change: uc004cov.4 ASHGA5P046074 0.00207334 210
1720), followed by NR_034120 (fold change: 929). Next, the TCONS_00001220 ASHGA5P044622 0.0121836 206
NR_044997 ASHGA5P052689 0.000611876 198
expression quantity of lncRNA between CAP and CG were
uc002tfi.3 ASHGA5P015326 0.014508807 151
compared. 1259 lncRNAs were found differentially expressed in ENST00000442037 ASHGA5P017469 0.002109831 139
total (640 up-regulated and 619 down-regulated) (fold change>2 ENST00000560602 ASHGA5P022669 0.002109831 125
and p < 0.05). The most up-regulated lncRNA was ENST00000452466 ASHGA5P026213 0.00331794 106
uc001qeg.1 ASHGA5P026908 0.00088196 102
TCONS_00013456 (fold change: 1438), followed by
ENST00000483236 ASHGA5P019420 0.006217351 101
ENST00000433012 (fold change: 1231). The most down-regulated ENST00000512284 ASHGA5P020365 0.00449272 100
lncRNA was ENST00000563895 (fold change: 1457), followed by NR_027391 ASHGA5P021819 0.044565006 93
ENST00000414686 (fold change: 735). ENST00000425176 ASHGA5P015956 0.08513433 89
Intersection analysis of TB and CAP was also performed. There NR_047671 ASHGA5P045272 0.021229735 87
ENST00000420143 ASHGA5P015422 0.062913217 82
were 754 lncRNAs differentially expressed in both TB and CAP
NR_029380 ASHGA5P054569 0.085646182 80
groups (432 up-regulated and 322 down-regulated). With the ENST00000448001 ASHGA5P038422 0.005942831 79
exclusion of the lncRNAs differentially expressed in both two TCONS_00019631 ASHGA5P042895 0.026453551 77
groups, there were 2116 specific lncRNAs differentially expressed in ENST00000373604 ASHGA5P041923 0.04012896 71
TB group (1102 up-regulated and 1014 down-regulated), which ENST00000584722 ASHGA5P033730 0.00856447 71
ENST00000427151 ASHGA5P016132 0.002108628 69
constituted 6.96% of the total lncRNAs. With this data processing TCONS_00016358 ASHGA5P036233 0.01626739 68
work, the top 25-fold changes of up- and down-regulation specific uc021rro.1 ASHGA5P029473 0.003292917 68
lncRNAs differentially expressed in TB were summarized in Tables 2 ENST00000546607 ASHGA5P021993 0.000611876 63
and 3 respectively with FDR <0.05. Venn diagram was also used to Up regulation of specific lncRNA probe related to their sequence were listed with the
summarize the finding (Fig. 3). FDR larger than 0.05 and the top 25-fold changes.
 J. He et al. / Tuberculosis 107 (2017) 73e79 77

3.3. Real-time PCR validation on potential lncRNA biomarkers

To further validate the diagnostic utility of the potential lncRNA

biomarkers from micorarray screening for TB diagnosis, we
analyzed six up-regulated lncRNA and six down-regulated lncRNAs
highlighted in Tables 2 and 3 by RT-qPCR, from three plasma
samples of each TB infection, CAP and CG groups. It was shown that
five of the selected lncRNAs (ENST00000354432, uc004cov.4,
ENST00000427151, ENST00000584688, TCONS_00019584) were

Table 3
Down-regulated specific lncRNAs differentially expressed in TB.

Sequence name Probe Name FDR Fold Change

ENST00000428188 ASHGA5P035194 0.001705837 360

TCONS_00019972 ASHGA5P057632 0.003191115 176
uc001oou.3 ASHGA5P019712 0.002798086 164 Fig. 4. Comparisons of the lncRNA levels in the plasma from the CG, CAP and pa-
ENST00000464125 ASHGA5P038845 0.002640599 127 tients with TB. Validation on six up- and six down-regulated lncRNAs highlighted in
TCONS_00003870 ASHGA5P034099 0.018693985 91 Tables 2 and 3 in TB samples as a comparison to both CAP and healthy CG samples by
TCONS_00018420 ASHGA5P044461 0.000650101 85 RT-qPCR. * representing p < 0.05, N/A representing p > 0.05.
TCONS_00004316 ASHGA5P035558 0.002077581 77
ENST00000441700 ASHGA5P032773 0.011062316 75
TCONS_00018641 ASHGA5P057539 0.012317204 60
TCONS_00021223 ASHGA5P041173 0.006033566 55 differential expressed in tuberculosis samples (Fig. 4). To further
ENST00000439891 ASHGA5P042240 0.005658173 53 confirm this result, we increased the sample size of plasma from TB,
NR_051961 ASHGA5P017666 0.02533649 49 CAP and CG groups (n ¼ 50) to validate the above five lncRNAs.
NR_027074 ASHGA5P025694 0.000539493 48 ENST00000354432 (t ¼ 16.95, p < 0.05) and ENST00000427151
ENST00000568137 ASHGA5P031287 0.002978503 47
(t ¼ 7.56, p < 0.05) were validated with significant up-regulation in
TCONS_00009862 ASHGA5P056851 0.007868581 43
NR_024146 ASHGA5P025643 0.011758935 36 the TB samples compared to that of both CAP and CG groups (Fig. 5).
ENST00000584688 ASHGA5P033041 0.001162053 35 When cut-off values were set as 0.05 and 0.03 for ENST000
NR_024376 ASHGA5P045410 0.027885603 32 00354432 and ENST00000427151 respectively (dotted square in
ENST00000553496 ASHGA5P022257 0.030334685 30
Fig. 5), 82% of the TB patients with expression levels larger than the
NR_036546 ASHGA5P026967 0.090097234 28
ENST00000423402 ASHGA5P050511 0.008049285 28 cut-off in these two lncRNA markers would be identified for further
NR_029394 ASHGA5P015662 0.014995538 25 investigation, while none from patients with CAP and healthy CG.
TCONS_00019584 ASHGA5P043281 0.008168996 24 We have therefore suggested that these two lncRNA can be used as
NR_033883 ASHGA5P045483 0.023916238 24 the biomarkers in plasma for TB diagnosis. In our suggestion, two
ENST00000397112 ASHGA5P055588 0.006385391 20
are better than one because this can reduce the error from differ-
Down regulation of specific lncRNA probe related to their sequence were listed with ences in individual expression.
the FDR larger than 0.05 and the top 25-fold changes.
LncRNAs shared similar structure characteristics with mRNA but
have no open reading frame or high-density termination codon.
They have been considered as the by-products in genome tran-
scription. Findings from recent studies that many lncRNAs devel-
oped structural characteristics, such as poly A tail after splicing, and
were found with conserved secondary structures in sub-cellular
localizations. Interestingly, these fragments have stable nucleic
acid splicing sites, a low base replacement rate, self-deletion rate,
foreign fragment insertion rate, and a space specific expression
mode. These features suggest that the differential expression of
lncRNA may impose biological significance [12,13]. To date, most of
the lncRNA studies focused on cancer development, cardiovascular
diseases, neuro- and immuno-diseases [14e16]. These two lncRNAs
upregulated in TB patients, where ENST00000427151 is an anti-
sense transcript targeting at CD81 antigen. It is also known as CD81
antisense RNA 1 (CD81-AS1), and ENST00000354432 is an anti-
sense transcript targeting at the sodium/glucose cotransporter 5
isoform 1 (SLC5A10). The former is related to TB pathogenesis, and
CD81 has been reported to reduce the threshold for B cell activation
[17,18]. The latter prevents SLC5A10 transcription, which is highly
expressed in the kidney, and its malfunction is related to renal
dysfunction, which happens in TB infection [19,20]. Our study is
therefore a pioneer work to investigate the relationship between
lncRNAs and infectious diseases by using clinical TB samples to
Fig. 3. Venn diagram on the summary of TB, CAP and CG. 2870 lncRNAs were show the importance of lncRNAs as the biomarkers for the diag-
differentially expressed in TB (1534 up-regulated and 1336 down-regulated). 1259 nosis of the disease. In this report, we provide an alternative means
lncRNAs were found differentially expressed in CAP compared to CG (640 up-regulated
of TB diagnosis by screening lncRNA biomarkers, other than TB
and 619 down-regulated). 754 lncRNAs differentially expressed in both TB and CAP
groups (432 up-regulated and 322 down-regulated).
mRNA. The higher mutation rate of the TB coding gene happened to
78 J. He et al. / Tuberculosis 107 (2017) 73e79

Fig. 5. Comparisons of the lncRNA levels in the plasma from the healthy CG, CAP and patients with TB infection. RT-qPCR validation on two target lncRNAs by RT-qPCR (A)
ENST00000354432 and (B) ENST00000427151 showing up-regulation of fold change > 2.0 in Fig. 4. * representing p < 0.05, N/A representing p > 0.05. These two lncRNAs were
found significantly up-regulated in the TB samples, compared to both CAP and healthy CG. The relationship of these two lncRNAs biomarkers expression in (C) TB, (D) CAP and (E)
CG shows the effectiveness of a double-use of two biomarkers for TB diagnosis, with the dotted square as cut-off value (0.05 and 0.03 for ENST00000354432 and ENST00000427151
respectively) to distinguish between TB and both CAP and healthy CG.

hinder the molecular-based detection based on the base-pairing future.

and the recognition, such as the commonly used method PCR,
and reports have shown the profiling of the expression of the host
non-coding RNA, such as miRNA, could provide a solution to this
situation [21,22].
In fact, TB is a serious public health issue and its infection rate is
still on the rise, in spite of the recent improvement in the diagnostic
and the therapeutic interventions. Also, CAP and TB shared similar
onset symptoms, and the chest X-ray manifestation of early TB is
similar to that of CAP. The commonly used clinical practice to
differentiate these two diseases is to see whether the lesion sub-
sides after two-week anti-infection treatment. Yet, this approach
undoubtedly delays the treatment in patients. In view of this, we
investigated the lncRNAs participating in TB. To eliminate those
differentially expressed lncRNAs induced by common pulmonary
lesions, we have examined the plasma samples from patients with
CAP for differential diagnosis and identified two specific lncRNAs
for TB diagnosis. Yet, there are some limitations in our study. For
example, other infections from non-TB infectious agents and the
immune status of patients may affect the lncRNA expression pat-
terns. In order to minimize this effect, more samples should be used
in the microarray analysis to further validate the lncRNA expression
patterns. Also, using more samples for RT-qPCR could minimize the
individual differences so that the highly specific lncRNAs bio-
markers for TB can be found. Secondly, the functional studies of the
lncRNA dysregulation in the TB patients are still elusive. More in-
formation is needed in this regard to establish and understand
what the functional differences are among distinct lncRNAs.
Nevertheless, the investigation on the lncRNAs related to TB in this
study provides insight into TB diagnosis at the molecular level, and
the lncRNAs biomarkers (ENST00000354432 and ENST00000
427151) are expected to be useful for TB diagnosis in the coming
Conflict of interests
The authors declare that there is no conflict of interests
regarding the publication of this paper.
Acknowledgement
This research is supported by National Key Research and
Development Plan of China (No. 2016YFF0203203, 2016YFC12
01404), Guangdong Science and Technology Foundation (2016A0
20247001, 20160223, 2016A020219005), National Natural Science
Foundation of China (No. K16026), Shenzhen Science and Tech-
nology Foundation (No. JCYJ20140419151618022, No. CXZZ20150
504163004339), Science and Technology Foundation of Shenzhen
Enter-exit Inspection and Quarantine Bureau (No. SZ2014101).
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.tube.2017.08.007.
References
[1] Organization, W. H. Tuberculosis http://www.who.int/mediacentre/factsheets
/fs104/en/.
[2] Cheon SA, Cho HH, Kim J, Lee J, Kim HJ, Park TJ. Recent tuberculosis diagnosis
toward the end TB strategy. J Microbiol Methods 2016;123:51e61.
[3] Norbis L, Alagna R, Tortoli E, Codecasa LR, Migliori GB, Cirillo DM. Challenges
and perspectives in the diagnosis of extrapulmonary tuberculosis. Expert Rev
Anti-infective Ther 2014;12:633e47.
[4] Loo JFC, Kwok HC, Leung CCH, Wu SY, Law ILG, Cheung YK, et al. Sample-to-
answer on molecular diagnosis of bacterial infection using integrated lab-on-
 J. He et al. / Tuberculosis 107 (2017) 73e79
a-disc. Biosens Bioelectron 2016. http://dx.doi.org/10.1016/jbios2016.09.001.
[5] Costa FF. Non-coding RNAs: new players in eukaryotic biology. Gene
2005;357:83e94.
[6] Dinger ME, Amaral PP, Mercer TR, Pang KC, Bruce SJ, Gardiner BB, et al. Long
noncoding RNAs in mouse embryonic stem cell pluripotency and differenti-
ation. Genome Res 2008;18:1433e45.
[7] Mercer TR, Dinger ME, Sunkin SM, Mehler MF, Mattick JS. Specific expression
of long noncoding RNAs in the mouse brain. PNAS 2008;105:716e21.
[8] Yin Z, Guan D, Fan Q, Su J, Zheng W, Ma W, et al. lncRNA expression signatures
in response to enterovirus 71 infection. Biochem Biophys Res Commun
2013;430:629e33.
[9] Su YJ, Yu J, Huang YQ, Yang J. Circulating long noncoding RNA as a potential
target for prostate cancer. Int J Mol Sci 2015;16:13322e38.
[10] Kim J, Kim KM, Ji HN, Yoon JH, Abdelmohsen K, Gorospe M. Long noncoding
RNAs in diseases of aging. Biochim Biophys Acta 2016;1859:209e21.
[11] Bi H, Zhou J, Wu D, Gao W, Li L, Yu L, et al. Microarray analysis of long non-
coding RNAs in COPD lung tissue. Inflamm Res 2015;64:119e26.
[12] Guttman M, Rinn JL. Modular regulatory principles of large non-coding RNAs.
Nature 2012;482:339e46.
[13] Eades G, Zhang YS, Li QL, Xia JX, Yao Y, Zhou Q. Long non-coding RNAs in stem
cells and cancer. World J Clin Oncol 2014;5:134e41.
[14] Li J, Chen Z, Tian L, Zhou C, He MY, Gao Y, et al. LncRNA profile study reveals a
three-lncRNA signature associated with the survival of patients with oeso-
phageal squamous cell carcinoma. Gut 2014;63:1700e10.
79
[15] Fang K, Han BW, Chen ZH, Lin KY, Zeng CW, Li XJ, et al. A distinct set of long
non-coding RNAs in childhood MLL-rearranged acute lymphoblastic leuke-
mia: biology and epigenetic target. Hum Mol Genet 2014;23:3278e88.
[16] Yuan SX, Yang F, Yang Y, Tao QF, Zhang J, Huang G, et al. Long noncoding RNA
associated with microvascular invasion in hepatocellular carcinoma promotes
angiogenesis and serves as a predictor for hepatocellular carcinoma patients'
poor recurrence-free survival after hepatectomy. Hepatology 2012;56:
2231e41.
[17] Levy S, Todd SC, Maecker HT. CD81 (TAPA-1): a molecule involved in signal
transduction and cell adhesion in the immune system. Annu Rev Immunol
1998;16:89e109.
[18] Raju B, Hoshino Y, Belitskaya-Levy I, Dawson R, Ress S, Gold JA, et al. Gene
expression profiles of bronchoalveolar cells in pulmonary TB. Tuberculosis
2008;88:39e51.
[19] NCBI SLC5A10 solute carrier family 5 member 10 [Homo sapiens (human)]
https://www.ncbi.nlm.nih.gov/gene/125206.
[20] Abbara A, Davidson RN. Etiology and management of genitourinary tuber-
culosis. Nat Rev Urol 2011;8:678e88.
[21] Loo JFC, Wang SS, Peng F, He JA, He L, Guo YC, et al. A non-PCR SPR platform
using RNase H to detect MicroRNA 29a-3p from throat swabs of human
subjects with influenza a virus H1N1 infection. Analyst 2015;140:4566e75.
[22] Szewczyk B, Bienkowska-Szewczyk K, Krol E. Introduc- tion to molecular
biology of influenza a viruses. Acta Biochim Pol 2014;61:397e401.
́ ́