Embed Size (px)
Citation preview
MicroRNA Profiling in Experimental
Sepsis-Induced Acute Lung Injury
by
Dun Yuan Zhou
A thesis submitted in conformity with the requirements
for the degree of Master of Science
Graduate Department of Institute of Medical Science
University of Toronto
© Copyright by Dun Yuan Zhou (year of 2014)
ii
Abstract
MicroRNA Profiling in Experimental Sepsis-Induced Acute Lung Injury
Master of Science
Year of Convocation – 2014
Dun Yuan Zhou
Institute of Medical Science
University of Toronto
Introduction: Currently, there are no specific pharmacological treatments for sepsis-induced
acute respiratory distress syndrome (ARDS). And mesenchymal stem cells (MSCs) have shown
reparative potential in both sepsis and ARDS.
Objectives: To determine the role of MSC administration in the modulation of pulmonary host-
responses to sepsis via differential regulation of regulatory microRNAs (miRNAs/miRs).
Methods: MicroRNA and mRNA profiling was performed to identify differential expression.
Quantitative real time polymerase chain reaction (qRT-PCR), trans-endothelial electrical
resistance (TEER) measurements, and luciferase activity assay were used.
Results: MicroRNA expression was examined in Human Pulmonary Microvascular Endothelial
Cells (HPMECs). One miRNA – miR-193b-5p, targets occludin, a tight junction protein associated
with endothelial leakage. A specific regulatory relationship between miR-193b-5p and occludin
iii
was identified. The loss in endothelial integrity was rescued when miR-193b-5p inhibitor was
transfected.
Conclusion: miR-193b-5p is a suppressor of occludin. Studying transcriptional changes allows
identification of therapeutically relevant mediators for ARDS/ALI treatment.
iv
Acknowledgements
It would not have been possible to write this Master’s thesis without the help and support of
the people around me, to only some of whom it is possible to give particular mention here.
Above all, I would like to thank my parents for their unequivocal support and great patience at
all times. As always, my mere expression of thanks cannot possibly suffice.
I would like to extend my gratitude to the many people who helped to bring this research
project to fruition. First, I am most grateful to Professor Claudia dos Santos for giving me the
opportunity of be part of her laboratory, taking part in the Master of Science program at the
Institute of Medical Science. I am so deeply thankful for her help, guidance, professionalism,
and financial support throughout this entire project and through my entire program of study
that I do not have enough words to express my deep and sincere appreciation.
Also, I would also like to thank the experts and members of the Program Advisory Committee
for this research project: Dr. Phil Marsden and Dr. Wolfgang Kuebler. Without their passionate
participation and input, the project would have not been at where it is now.
I would like to thank Dr. Yuexin Shan for her expert opinions on all technical issues in and
around the lab at all times. Her presence and professionalism have made a great impact on the
quality of research that we are able to perform at the institute. Likewise, I extend my special
v
thanks to Dr. Pingzhao Hu for his tremendous help on microarray data analysis, without him
this entire project would not even exist.
Last, but not least, I must thank my friends, Hajera Amatullah, Simon Rozowsky, Paymon Azizi,
Patricia Gali, for their support and encouragement throughout. Thank you for making the past
two years a great experience and an unforgettable memory in my life.
Thank you all! This accomplishment would not have been possible without any of you. Thank
you.
Dun Zhou
Aug 6, 2013
vi
Table of Contents
Abstract ii
Acknowledgements iv
Table of Contents vi
List of Abbreviations ix
List of Tables xiii
List of Figures xiv
Chapter 1: Introduction
1.1 Acute Respiratory Distress Syndrome (ARDS) 1
1.1.1 Definition and Background of ARDS
1.1.2 Endothelial and Epithelial Injury in ARDS
1.1.3 Mechanisms of Increased Permeability
1.1.4 Signaling Mechanisms Increasing Paracellular Permeability
1.1.5 Signaling Mechanisms Restoring Normal Paracellular Permeability
1.1.6 Approach to ARDS Treatment
1.2 Transcriptomic analysis 10
1.2.1 Background
1.2.2 Microarray and Sepsis
1.2.3 Novel Genes, Pathways, and Networks
vii
1.2.4 Microarray Application and ARDS
1.3 Mesenchymal stem cells 19
1.3.1 Background
1.3.2 MSC and ALI/ARDS
1.3.3 MSC Confers Protective Benefits in Animal Models of ALI
1.4 MicroRNA 23
1.4.1 Background
1.4.2 microRNA Processing and Function
1.4.3 microRNA Regulation
1.4.4 Targets of miRNAs
1.4.5 Roles of miRNAs in the Lung
1.4.6 miRNAs as Potential Molecular Therapies and Biomarkers for ARDS
1.5 Endothelial Permeability and Occludin 31
1.5.1 Background
1.5.2 Regulation of Tight Junction Permeability
1.5.3 Occludin as a Membrane Component of the Tight Junction
1.5.4 Occludin Interaction in the Tight Junction
1.5.5 Functional Role of Occludin in the Tight Junction
1.5.6 Disturbance of Barrier Function
Chapter 2: Objectives and Hypotheses 39
Chapter 3: Materials and Methods 44
viii
Chapter 4: Results 56
Chapter 5: Discussion 105
Chapter 6: References 122
ix
List of Abbreviations
3’UTR 3’ untranslated region
5’UTR 5’ untranslated region
Ago2 argonaute 2
ALI acute lung injury
Ang-1 angiopoietin 1
ANOVA analysis of variance
APACHE II acute physiology and chronic health evaluation II
ARDS acute respiratory distress syndrome
ATP adenosine triphosphate
BAL broncoalveolar lavage
CIP calf-intestinal alkaline phosphatase
CLP cecal ligation and perforation
cDNA complementary DNA
CX3CR1 CX3C chemokine receptor 1
DAD diffuse alveolar damage
DGCR8 microprocessor complex subunit DGCR8
DMSO dimethyl sulfoxide
EBM-2 Endothelial Cell Basal Medium MV 2
eIF4E eukaryotic translation initiation factor 4E
eIF4G eukaryotic translation initiation factor 4G
eIF6 eukaryotic initiation factor 6
ELISA enzyme-linked immunosorbent assay
ERK extracellular signal-regulated kinase
FDR false discovery rate
x
FBS fetal bovine serum
FC fold change
FITC fluorescein isothiocyanate
GAPDH glyceraldehyde-3-phosphate dehydrogenase
GuK guanylate kinase
HLA human leukocyte antigen
HPMEC human pulmonary microvascular endothelial cells
ICU intensive care unit
IEJ interendothelial junctions
IgG immunoglobulin G
IL-1 interleukin-1
IL-6 interleukin-6
IL-7 interleukin-7
IL-10 interleukin-10
IL-12 interleukin-12
LIMMA linear models for microarray data
LNA locked nucleic acid
LPS lipopolysaccharide
MDCK Madin-Darby canine kidney epithelial cell
MHC major histocompatibility complex
MMP-8 matrix metalloproteinase-8
MODS multiple organ dysfunction syndrome
mRNA messenger RNA
miR / miRNA microRNA
miRDB miRNA data base
xi
MiRISC miRNA-incorporated RNA-induced silencing complex
MSCs mesenchymal stem cells
MSC-CdM mesenchymal stem cell conditioned medium
MTT 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide
NADH nicotinamide adenine dinucleotide (reduced form)
NF-kB nuclear factor kappa-light-chain-enhancer of activated B cells
OCC2 occludin second extracellular domain peptide
OCLN occludin
PACT protein kinase R activating protein
PBS phosphate buffered saline
PCA principal components analysis
pCO2 partial pressure of carbon dioxide
pO2 partial pressure of oxygen
qRT-PCR quantitative real-time polymerase chain reaction
RISC RNA-induced silencing complex
RNase III ribonuclease
SDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis
SEM standard error of mean
SMAD1 SMAD family member 1
SOFA sequential organ failure assessment
TBS tris-buffered saline
TC tissue culture
TER transepithelial resistance
TEER transendothelial electrical resistance
TGF transforming growth factor
xii
TJ tight junction
TLR2 toll-like receptor 2
TLR3 toll-like receptor 3
TLR4 toll-like receptor 4
TNFa tumour necrosis factor alpha
TNFRSF1B tumour necrosis factor receptor super family 1 B
TNFRSF11B tumour necrosis factor receptor super family 11 B
TRBP TAR RNA binding protein
VAP-33 vesicle-associated membrane protein (VAMP)-associated protein of 33 kDa
VA/Q ratio ventilation/perfusion ratio
VEGF vascular endothelial growth factor
VILI ventilator induced lung injury
WT wild type
ZIP8 zinc transporter ZIP8
ZO-1 tight junction protein 1
ZO-2 tight junction protein 2
ZO-3 tight junction protein 3
xiii
List of Tables
Table 1 reverse transcription and PCR primer sequences for selected miRNAs.
Table 2 targets of miR-193b-5p with differential expression in the mRNA microarray
xiv
List of Figures
Figure 1 principal component analysis of mRNA data between sham and CLP
Figure 2 hierarchical clustering of mRNA with differential expression between sham and CLP
Figure 3 principal component analysis of mRNA data between CLP and MSC
Figure 4 volcano plot showing results of LIMMA analysis of miRNA expression between sham and
CLP
Figure 5 volcano plot showing results of LIMMA analysis of miRNA expression between CLP and MSC
Figure 6 volcano plot showing results of LIMMA analysis of miRNA expression between sham and
MSC
Figure 7 Heat map of occludin mRNA expression in Sham, CLP, and MSC mouse lung tissues.
Figure 8 Raw intensity values for miR-193b-5p by miRNA profiling
Figure 9 Raw intensity values for occludin by microarray
Figure 10 quantitative RT-PCR for miR-193b-5p in mouse lung tissue
Figure 11 quantitative RT-PCR for occludin in mouse lung tissue
Figure 12 external validation of miR-193b-5p expression in lung tissue from separate mice
Figure 13 external validation of OCLN mRNA expression in lung tissue from separate mice
Figure 14 external validation of OCLN protein expression in lung tissue from separate mice
Figure 15 quantitative RT-PCR for miR-193b-5p in HPMEC with TNFa treatment
Figure 16 quantitative RT-PCR for occludin in HPMEC with TNFa treatment
Figure 17 western blot of OCLN in HPMEC with TNFa treatment
Figure 18 quantitative RT-PCR for miR-193b-5p in HPMEC transfected with miR mimic
Figure 19 quantitative RT-PCR for OCLN in HPMEC transfected with miR mimic
Figure 20 quantitative RT-PCR for miR-193b-5p in HPMEC transfected with miR inhibitor
Figure 21 quantitative RT-PCR for OCLN in HPMEC transfected with miR inhibitor
Figure 22 western blot of OCLN in HPMEC transfected with miR inhibitor under TNFa treatment
xv
Figure 23 cell viability assay showing effect of inhibitor and mimic on viability
Figure 24 assessment of paracellular leak in HPMEC transfected with miR inhibitor
Figure 25 assessment of endothelial permeability in HPMEC transfected with miR inhibitor
Figure 26 luciferase activity assay showing miR binding to wild type or mutated OCLN 3’UTR
Figure 27 luciferase activity assay showing miR binding to wild type or mutated OCLN 3’UTR
Figure 28 quantitative RT-PCR for pre-miR-193b in transfected HPMEC
Figure 29 quantitative RT-PCR for pre-miR-193b in inhibitor-transfected HPMEC with TNFa
Figure 30 quantitative RT-PCR for miR-193b-3p in transfected HPMEC
Figure 31 quantitative RT-PCR for miR-193b-3p in inhibitor-transfected HPMEC with TNFa
Figure 32 quantitative RT-PCR analysis of miR-193b-5p in conditioned medium treated HPMECs
Figure 33 quantitative RT-PCR analysis of occludin in conditioned medium treated HPMECs
Figure 34 quantitative RT-PCR analysis of miR-193b-5p in conditioned medium treated mice
Figure 35 quantitative RT-PCR analysis of occludin in conditioned medium treated mice
1
1.1 Acute Respiratory Distress Syndrome (ARDS)
1.1.1 Definition and Background of ARDS
Acute respiratory distress syndrome, is characterized by increased permeability of the alveolar-
capillary barrier, and accumulation of a protein-rich pulmonary edema, and infiltration of
neutrophils, macrophages, and cytokines. The term acute lung injury encompasses a spectrum
of bilateral pulmonary damage (endothelial and epithelial), which can be initiated by numerous
conditions. Clinically, acute lung injury manifests as acute respiratory distress syndrome (ARDS):
(1) acute onset of dyspnea, (2) decreased arterial oxygen pressure (hypoxemia), and (3)
development of bilateral pulmonary infiltrates on the chest radiograph, in the absence of
clinical evidence of primary left-sided heart failure [1, 2]. Histologically, diffuse alveolar damage
is the major feature of ARDS and this is in turn characterized by damage to the alveolar capillary
membrane, inflammatory cell infiltration, exudative pulmonary edema and hyaline membrane
deposition.
The incidence ARDS is variable among studies, it ranges from 5 to 86 cases per 100,000 person-
years [3-5]. Overall, in the United States, ALI has a high incidence of approximately 200,000 per
year. Moreover, mortality remains high ARDS. The 2011 Berlin definition revised the 1994
criteria, aiming to address issues of reality and validity that had emerged over the years [2, 6].
Acute lung injury (ALI) was eliminated as a distinct category and ARDS was re-classified on the
basis of severity of hypoxemia: mild (200 mm Hg < PaO2/FIO2 ≤ 300 mm Hg), moderate (100
mm Hg < PaO2/FIO2 ≤ 200 mm Hg), and severe (PaO2/FIO2 ≤ 100 mm Hg).
2
Patient level meta-analysis of 4188 ARDS patients from four multicentre clinical data sets was
used to evaluate the Berlin definition [7]. Under the Berlin definition, mild, moderate, and
severe stages of ARDS were associated with increased mortality (24%-30%; 29%-34%; and 42%-
48%, respectively; P < .001), as well as lengthened median duration of mechanical ventilation in
survivors (5 days; interquartile [IQR], 2-11; 7 days; IQR, 4-14; and 9 days; IQR, 5-17, respectively;
P < .001).
Clinically, ARDS involves a series of diverse processes that produce widespread alveolar damage.
In an injured alveolar capillary barrier, there will be exudation of fluids, cells and proteinaceous
contents across the membrane, leading to the subsequent alveolar edema to cause refractory
hypoxemia the cardinal physiological manifestation of this syndrome. The American/ European
Consensus Conference in 1994 [8] defined ARDS as: acute onset of hypoxemia, associated with
bilateral infiltrates on chest X-ray (CXR) in the absence of left ventricular failure.
1.1.2 Endothelial and Epithelial Injury in ARDS
ARDS is characterized histologically by diffuse alveolar damage (DAD). This is a final common
pathology caused by a large variety of insults including respiratory infections, sepsis, shock,
aspiration of gastric contents, inhalation of toxic gases, near-drowning, radiation pneumonitis
and many drugs and other chemicals. These conditions are linked by the fact that they can all
injure alveolar epithelial and endothelial cells, thereby producing DAD.
3
About 90 percent of the alveolar surface area is made up of flat type I epithelial cells that are
easily injured, and cuboidal type II cells make up less than 10% of the alveolar surface area. The
main functions of type II cells include surfactant production, ion transport, proliferation and
differentiation to type I cells after injury. In ARDS, epithelial integrity is severely compromised.
Normally, the endothelial barrier is much more permeable than the epithelial barrier [9, 10]. As
a result, epithelial injury can directly contribute to alveolar flooding. Also, the loss of epithelial
integrity and injury to type II cells critically disrupt normal epithelial fluid transport, impairing
the removal of edema fluid from the alveolar space [11, 12]. Moreover, ARDS also leads to
reduced production and turnover of surfactant as type II cells are injured [13].
Under normal conditions, both epithelial and endothelial barriers are intact, whereas in disease
conditions, gaps are formed in the endothelium and or epithelium. Pulmonary edema is
produced as the barrier functions are compromised in the epithelium and endothelium. There
is leakage of water, protein, inflammatory cells, and red blood cells into the interstitium and
alveolar lumen [13].
1.1.3 Mechanisms of Increased Permeability
At the site of gas exchange, the capillary-alveolar barrier acts to prevent air bubble formation in
the blood, and to ensure that blood does not enter the alveolar space. The capillary-alveolar
barrier is permeable to various gases, including oxygen and carbon dioxide. In ARDS
development, destruction of endothelial cells and damaged pulmonary capillary endothelium
are two characteristics [14, 15].
4
In terms of capillary fluid flux in the lung, endothelial activation affects the process in several
ways. In a more general pathway, activated endothelium releases inflammatory mediators that
amplify endothelial injury directly or indirectly by recruiting inflammatory cells into the vascular,
interstitial, and alveolar spaces. [16]. Some mediators such as, angiotensins, bradykinin, and
endothelin, have crucial effects on vascular regulation, and damage in the endothelium may
substantially impair their metabolism. As a result, there may be interstitial fluid flux through
physiologic mechanisms that are normally closely regulated. In addition to these vaso-
regulatory mediators, reactive nitrogen and oxygen species may further deteriorate
microcirculatory dysfunction in ARDS. If these reactive species enhance local blood flow or
increase postcapillary resistance, they may increase capillary pressure and worsen pulmonary
edema [17].
Increased endothelial permeability and reduction of alveolar liquid clearance capacity are two
leading pathogenic mechanisms of pulmonary edema, which is a major complication of acute
lung injury. Understanding of fundamental mechanisms involved in regulation of endothelial
permeability is essential for development of barrier protective therapeutic strategies.
The endothelial barrier function is regulated via two pathways: the transcellular and the
paracellular pathways. The transcellular pathway, also known as transcytosis, is a constitutive
process which transports macromolecules larger than 3nM through the cells via vesicles [18,
19]. Paracellular pathway passively allows transport of solutes ions smaller than 3nM in
5
diameter [19, 20]. The paracellular pathway is regulated by interendothelial junctions (IEJs) that
are responsible for maintaining cell–cell contacts and will be the focus of this thesis.
1.1.4 Signaling Mechanisms Increasing Paracellular Permeability
1. Intracellular Calcium
Intracellular calcium is known to play a role in the regulation of membrane permeability.
Transient receptor potential family (TRP) members are involved in calcium-mediated channels
for permeability [21]. TRPC1 and TRPC4 are the two ion channels with greatest effects on
increasing lung endothelial cellular permeability as activation of calcium entry through its stores.
TRPC1 and TRPC4 channels exert most of their influence on the lung extraalveolar endothelial
barrier, and TRPV4 channels are more involved in lung capillary endothelial barrier function [22,
23]. When the alveolar septal barrier is compromised, it can lead to alveolar flooding and
impaired gas exchange [21, 23].
2. Role of Rho GTPases
One established pathway known to be involved in endothelial contractility is the Rho/ Rho
kinase (ROCK) pathway. The small GTPase RhoA and ROCK are known to regulate cellular
adherence, migration, and proliferation through control of the actin–cytoskeletal assembly and
cell contraction [24, 25] Rho kinases have been identified to have a variety of functions,
including regulatory role in T cell transendothelial migration, prevention of thrombin-induced
intercellular adhesion molecule 1 (ICAM-1) expression and inhibition of nuclear factor NF-kB
activity [26]. Specifically for the endothelium, background Rho kinase activity is fundamental for
6
the maintenance of endothelial barrier integrity [27] When RhoA/ROCK is over-activated, it can
lead to cytoskeletal rearrangement, consequently leading to barrier dysfunction in vascular
endothelium [27, 28] When endothelial integrity is no longer maintained, there is an imbalance
between endothelium-derived relaxing factors (EDRF) and endothelium-derived constricting
factors (EDCF). When endothelial function is dramatically compromised, EDCFs mediate
endothelium-dependent contraction by activating thromboxane–prostanoid (TP) receptors on
vascular SMCs [29, 30], and one such signaling molecule activated by TP receptor in smooth
muscle is Rho kinase [31]. Also, ROCK has been shown to contribute to endothelium-dependent
contraction to Ach and enhanced vasocontractile activity [32]. Therefore, studies that focus on
inhibition of the RhoA/ROCK pathway may have significant clinical implications.
Experiments addressing the mechanisms of endothelial permeability increases in acute lung
injury and ARDS have focused on endothelial responses to agonists such as thrombin or
histamine. Binding of such mediators to their cell-surface receptors elicits a series of signaling
events that culminate in cell rounding and interendothelial gap formation, which represents
disruption of interendothelial junctions [33, 34].
3. Role of Myosin Light Chain Kinases
Normally, the tension between actin-myosin interaction plays a critical role in controlling the
semi-permeable nature of the microvasculature. Myosin light chain kinase (MLCK)
phosphorylates myosin light chain (MLC), which elicits the actin-myosin interaction and reduces
endothelial cell-cell interaction. In ARDS conditions, inflammatory or angiogenic mediators can
7
cause vascular hyperpermeability [35]. MLCK inhibitors have been shown to reverse the initial
increase in endothelial permeability when MLCK is activated. MLCK210 knockout mice have also
been used to associate MLCK with endothelial hyperpermeability. Mediators, including
thrombin, TNF-alpha, and LPS, activate MLCK, and GTPase RhoA (introduced in the previous
section). Ca2+ activation of MLCK and RhoA disrupts junctions [36].
1.1.5 Signaling Mechanisms Restoring Normal Paracellular Permeability
1. Sphingosine Kinases
Sphingolipids can act as intracellular second messengers and extracellular mediators that have
critical roles in cellular processes. Sphingosine-1-phosphate (S1P) has received the greatest
attention, as it is a potent angiogenic factor that elevates lung endothelial cell integrity, and it
has been shown to prevent vascular permeability and alveolar flooding in preclinical animal
models of ALI [37]. S1P is produced from sphingosine by two kinases, SphK1 and SphK2, which
have been identified as possible drug targets for the treatment of ARDS [38]. Another
important protein in S1P response is focal adhesion kinase (FAK) [39]. Endothelial FAK is
stimulated by S1P – tyrosine phosphorylation results in FAK activation and dynamic interactions
with other effector molecules, all of which act to improve endothelial barrier function.
8
2. Role of S1P Lyase
As an endogenous lipid regulator of endothelial permeability, S1P enhances barrier function
through actin and junctional protein rearrangement, leading to cytoskeletal changes [37, 40].
S1P lyase is the enzyme that cleaves S1P to permanently degrade the molecule. Zhao et al. have
demonstrated that inhibition of S1PL could ameliorate ALI in murine models [41]. The Zhao
group also illustrated that intratracheal instillation of LPS to mice exacerbated S1PL expression,
which led to a decreased S1P level in lung tissue, and the resultant enhanced level lung
inflammation and injury.
1.1.6 Approach to ARDS Treatment
Despite the fact that improvement in the supportive care of patients with ARDS has been seen,
its mortality rate of ARDS remains high. Treatment approaches have extended from modulating
inflammation to a careful search for the underlying causes, with particular attention paid to
sepsis or pneumonia. Several novel treatment strategies have been evaluated based on the
field’s increased level of understanding of ARDS pathogenesis.
There has been a large number of pharmacological treatments, including glucocorticoids,
surfactants, inhaled nitric oxide, antioxidants, protease inhibitors, and a variety of other anti-
inflammatory treatments, all of which have been assessed in various Phase II and III clinical
trials for ARDS. However, none of these aforementioned pharmacological therapies has proven
to be effective over the majority of patients [42, 43]. Given the lack of a specific pharmacologic
9
treatment, lung-protective ventilation has continued to be the most widely used supportive
care therapy, and has reduced the mortality of ALI from 40% in 2000 to 25% in 2006 [44].
Statin therapy for ARDS patients has been proven to be effective to a certain extent,
particularly in patients with sepsis-induced ARDS. Also, given the fact that sepsis is the most
harmful cause of ARDS, recent studies have looked at how directly combatting with sepsis may
have significance in treating ARDS, notably by testing a Toll-4 receptor inhibitor in patients with
sepsis [45].
Cell-based approaches are the most promising new therapies, and are receiving the most
attention. In preclinical studies, research in mouse models of ARDS has shown that
mesenchymal stem cells may be of value in treating ALI as well as sepsis. MSCs have been
shown to have immune-modulatory effects in the lung, and a recent study reported that MSCs
decreased pro-inflammatory cytokines and increased in anti-inflammatory cytokines,
particularly IL-10 [46]. Also administration of MSCs was shown to alleviate epithelial
permeability to protein, to reduce pulmonary edema, and to increase alveolar fluid clearance
[47]. Moreover, MSCs have been found to release paracrine factors which confer protective
effect [48]. More preclinical studies are conducted to evaluate the protective mechanisms of
MSCs in ALI treatment
10
1.2 Transcriptomic Analysis
1.2.1 Background
The most fundamental advantage associated with microarray technology is the ability to
simultaneously monitor thousands of mRNA transcripts at the same time, what is also known as
transcriptomics. RNA is reverse transcribed into cDNA, and a labeling molecule is incorporated
for detection. Once labelled, the cDNA targets are applied to a chip/plate that already contains
nucleotide sequences that correspond to specific probes. Then, hybridization occurs between
the probes and targets, and the amount of hybridization will be representative of the
abundance of a specific mRNA. Raw mRNA abundance data are retrieved once the chip/plate
has been scanned.
Data analysis for microarray is a constantly evolving subject. Commonly, it involves data
normalization, statistical comparison across groups by analysis of variance, and correction for
multiple comparisons [49]. The initial data normalization allows comparisons across samples by
reducing technical errors, so that the true biological variations can be acknowledged [50, 51].
Then, by analysis of variance, statistical comparisons across multiple groups of interest can be
performed. This step is critical but can be troublesome in terms of generating consistent lists of
differentially regulated gene. This is largely due to the number of available statistical tests that
could be used, and from the same data set, depending on the test used, a different list of
differentially expressed genes could be produced [52, 53] This corrects for the multiple
comparisons performed in each individual test.
11
Some of the most popular deciphering techniques include hierarchical clustering, principal
component analysis, expression mosaic map, etc. Hierarchical clustering will generate heat
maps which group genes and samples together based on their relatedness of expression.
Principal component analysis groups samples together and attempt to account for their
variances within the group and across different treatment groups. Both of these methods help
interpret and visualize a picture of global gene expression patterns, and can help identify
outliers and disease state expression based on differential analysis [54]. The gene mosaic
approach analyzes the microarray data based on pattern recognition, and this method has been
clinically applied to categorize patients into relevant subclasses [55, 56].
By statistical tests, genes with differential expression can be identified from the microarray
experiment. However, understanding the meaning of such genes may be a challenging process.
To understand about the meaning of genes, data mining is an approach that can reveal
abundant potential interactions and relevance of gene sets under specific conditions. To
designate biological function to genes, both public and proprietary databases have been set up.
Once submitted, the microarray data will be examined to determine whether certain gene lists
are functionally or structurally related based on literature. These databases could be used to
help identify possible biological functions of genes in a given context, and generate P values
based on the evaluation of the likelihood of how a gene is enriched for a particular biological
function. Biological meaning of a single gene also depends on gene networks and interactions
[57]
12
1.2.2 Microarray and Sepsis
In the past two decades, microarray technology has become increasingly popular in studying
trends of gene expression. Researchers have utilized the available microarray technology in
combination with various bioinformatics techniques to a number of complex diseases, including
cancer and inflammation.
Specifically for clinical sepsis, genome-wide analysis allows for the ability to simultaneously and
efficiently measure mRNA abundance of thousands of transcripts from a given tissue source
across different treatments or conditions, and it has inspired an unmatched opportunity to gain
a broader, genome level understanding of complex and heterogeneous clinical syndromes such
as sepsis [57, 58]. Microarray-based transcriptome-level analysis has led to a genome level
understanding of the complexity of disease, identified novel candidate targets and pathways for
potential therapeutic intervention, pinpointed candidate diagnostic biomarkers, enabled
stratification of patients based on their expression of sepsis-related genes [59]. Also, the
genome-level screening approach has been effective on reducing researcher’s bias. Since the
entire genomic library is being closely examined rather than a predetermined set of genes that
could be arbitrarily handpicked by the research, the discovery of potentially meaningful
molecular targets and tools is much broader and more reliable.
Specifically for sepsis, as a disease that displays tremendous complexity in its gene expression,
microarray screening could be an extremely rewarding approach. One group was able to
uncover that a large number of transcripts related to inflammation and innate immunity were
13
substantially up-regulated in response to endotoxin when mononuclear cell-specific RNA was
used in their microarray study [60]. In accordance with the acute phase of ARDS development,
the researchers found the peak transcriptomic response in less than 24h following a single
endotoxin administration. On top of that, the investigators also identified genes that exhibited
differential expression in response to endotoxin exposure that were not previously thought to
have any role in acute inflammation [57, 60]. Using network analysis, a large number of gene
pathways and networks that are related to mitochondrial function, namely, energy production,
showed substantial down-regulation in their expression trend – some of the genes involved in
the electron transport chain, NADH dehydrogenase 1, ATP synthase were significantly knocked
down. As an end result, protein synthesis related genes also had reduced gene expression, such
as ribosomal protein L3, ribosomal protein S8, eukaryotic translation initiation factor. These
findings in the repression of mitochondrial genes have been confirmed in human sepsis patients
[61, 62].
Adaptive immunity has been implicated to have a key role in the development of acute
inflammation as well as ARDS. Not surprisingly, exposure to endotoxin was shown to lead to
suppression of a list of genes directly involved in the proper functioning of adaptive immunity.
Studies that involve human endotoxemia can provide results that aid in the advancement of
knowledge in sepsis at a transcriptomic level. Nonetheless, this model of endotoxemia does not
fully replicate the complex and heterogeneous syndrome seen at the bedside following
infection with live microbes [63].
14
In response to better characterize the global gene expression changes in sepsis, various
investigators have attempted microarray-based studies in critically ill patients with sepsis and
septic shock. Despite the fact that inherent heterogeneity exists in clinical sepsis, several
studies have provided novel insight into the overall genome-level response to sepsis [64-66].
A widespread pattern that has been found in many microarray-based sepsis studies is the
considerable overexpression of genes associated with inflammation and innate immunity.
Microarray data generated from these patients are largely in accordance with what is already
known with sepsis, such as the initial hyperactive inflammatory response, making the
microarray data highly plausible in further explaining clinical sepsis.
Microarray based studies have also been able proven effective in documenting gene expression
changes associated with adaptive immune system and the inability to clear infection [67].
Previously, in a mouse experimental model of sepsis, the administration of an anti-apoptotic
cytokine, IL-7, led to lymphocyte survival and expansion which ultimately led to improved
survival in the animals [68].
In patients with sepsis, early repression of adaptive immunity genes has been reported [66, 69].
And even in children with sepsis, a similar trend of early repression of adaptive immunity-
related genes has been documented several times, such genes include T cell receptors, and
genes that respond to the receptor [64]. Collectively, microarray screening on the genome level
15
has been able to document and support the existing understanding of adaptive immune
dysfunction as an early and prominent feature of clinical sepsis.
1.2.3 Novel Genes, Pathways, and Networks
The ability to closely examine the entire human genome constitutes a novel approach to
discover individual molecules, as well as gene pathways and networks that may play role in
sepsis. Recently, several studies have demonstrated the potential of genome-wide screening of
gene profiles in uncovering novels targets for sepsis treatment. For instance, interleukin-6 was
identified as a critical contributor to myocardial depression in patients with meningococcal
sepsis, using microarray-guided in vitro approaches. The relatively homogenous patient
samples further ensured the result of the study.
As one of pioneer studies that used microarray to identify genes with differential expression
between survivors and non-survivors, the chemokine receptor CX3CR1 was the most highly
expressed in survivors [70, 71]. The researchers conducted in vitro experiments and found that
dys-regulation of CX3CR1 in monocytes contributed to immune paralysis in human sepsis [72,
73]. Therefore, there are considerable potential that lies in the use of the microarray approach
to discover novel pathways through gene expression profiling.
As a beneficial strategy proven in experimental sepsis zinc supplementation has not reached
clinical trials due to safety concerns, despite the fact that decreased plasma zinc concentrations
have been found in patients with sepsis, and that low plasma zinc concentrations correlate with
16
higher illness severity [74]. Since normal zinc homeostasis is essential to immune function, zinc
supplementation has been raised as a potential treatment for sepsis. Interestingly, microarray-
based studies in children with septic shock have reported high levels of ZIP8, a zinc transporter,
expression in non-survivors, relative to survivors [75, 76] . More experiments will be necessary
to fully establish the networks of genes that interact with zinc.
Also, metalloproteinase (MMP)-8 has shown marked increases in its expression in patients with
septic shock, compared to control patients [64, 77]. Moreover, among patients with sepsis, the
ones with more severe sepsis expressed a greater amount of MMP-8; similarly, non-survivors
also showed more MMP-8 expression than survivors. It is known that chemokines and cytokines
contribute to MMP-8 production. Recently, studies have demonstrated that genetically
truncated MMP-8 or direct inhibition of MMP-8 activity may provide significant survival
advantage in an animal experimental model of sepsis [78].
1.2.4 Microarray Application and ARDS
Microarray-based genome screening has remained a new technology to the field of ARDS.
Based on microarray data of gene expression from animal models of lung injury, Hu et al.,
performed a meta-analysis to identify potential injury-specific gene expression signatures that
may help predict the development of lung injury in humans [79]. From the study, two injury-
specific lists of differentially expressed genes were discovered from animal lung injury models.
External data sets and prospective data from animal models of VILI validated the results. By
examining the potential pathways of gene sets, it was identified that a large number of
17
differentially regulated genes are involved VILI-related pathways. Gene expression signatures
identified in animal models of lung injury were used to predict development of primary graft
failure in lung transplant recipients, and achieved greater than 80% accuracy. The data suggests
that microarray analysis of gene expression data allows for the detection of “injury" gene
predictors that can classify lung injury samples and identify patients at risk for clinically relevant
lung injury complications.
Also, another recent study from our own group identified three commons effects of MSC
administration in an experimental animal model of sepsis-induced inflammation and organ
injury. By using a network knowledge-based approach, common effects of MSC in five different
organs were discovered, including, i) attenuation of sepsis-induced mitochondrial-related
functional derangement, ii down-regulation of endotoxin/Toll-like receptor innate immune
proinflammatory transcriptional responses, and iii) coordinated expression of transcriptional
programs implicated in the preservation of endothelial/vascular integrity. Conclusively,
transcriptomic analysis has shown that the beneficial effects of MSC therapy in sepsis are not
limited to a single mediator or pathway, but rather involves a substantial number of biological
networks playing critical roles metabolism and inflammatory response of the cell [80].
In a recent study published by our lab, genome level screening identified transcriptional
responses of target organs to MSC therapy [80]. Rather than individual genes that exhibited
altered levels of expression, a large number of gene sets and gene pathways showed significant
changes in their expression following sepsis and MSC administration. Thus, it is highly plausible
18
that the beneficial effect of MSC therapy is not attributed to single mediators or pathways,
instead, a series of networks and activities could be involved in the biological control of the
host system in response to systemic inflammation. miRNAs are the ideal candidates for studies
that involve large numbers of targets/intermediates, as each miRNA can potentially have
hundreds of target mRNAs, and each mRNA can be regulated by a number of miRNAs under
different conditions [81]. In contrast to a proteomics study, the combination of miRNA and
mRNA microarrays can yield much information on the interaction between miRNA and mRNA
prior to translation after which proteins are produced. The expression signature obtained from
such studies will be much more dynamic than proteomics alone. Thus, the focus of miRNA and
miRNA pairing/interaction will be much more reflective of the changes in expression, as well as
downstream phenotypes, making the approach more representative of the nature of sepsis and
sepsis-induced ARDS.
In summary, these aforementioned molecules came to the sight of sepsis researchers through
the use of microarray-based genome screening. Even though many of these studies demand
further validation experiments to confirm the proposed relationship, the novel approach of
genome-based analysis has yielded a new approach to identify targets that may have potential
therapeutic value in the clinical setting.
19
1.3 Mesenchymal stem cells
1.3.1 Background
Mesenchymal stem cells (MSCs) are multipotent, self-renewing progenitor cells. MSCs can be
isolated from the bone marrow, umbilical blood, placenta, fat, skeletal muscles and tendons
[82-84]. MSCs have a very slow rate of differentiation. And there are three major criteria to
characterize MSCs, (a) adherence to plastic, (b) expression of CD105, CD73, and CD90; lack of
expression of CD45, CD34, CD14, CD11b, CD79α, CD19, HLA II, and (c) ability to differentiate
into osteoblasts, adipocytes, and chondroblasts in vitro [85, 86].
MSCs have the ability to differentiate into any other cell type in the body, and previous studies
have examined how such potential could aid in the repair of various tissues. Recently, however,
MSCs have received much attention due to their potent immunomodulatory effects. A wide
variety of molecules are produced by MSCs, such as hematopoietic factors, chemokines, and
angiogenic factors [87, 88]. It has been found that MSCs are able to modulate the immune
response [89]. One critical characteristic of MSCs is their lack of MHC Class II antigens, allowing
them to evade immune destruction and enabling them to be used in gene therapies [90].
In preclinical trials, MSCs have been used widely and found to have beneficial protective effects;
and the feasibility of pursuing cell based therapies is currently being explored for a variety of
lung diseases.
20
1.3.2 MSC and ALI/ARDS
In the past decade, mesenchymal stem cells have been used in the treatment for experimental
ARDS. MSCs have been shown to possess both anti-inflammatory as well as reparative potential.
Since the repair process and attenuation of inflammatory responses are important for ARDS
treatment, the ability of MSCs to release growth factors, such as keratinocyte growth factor
(KGF), hepatocyte growth factor, to induce repair in the lung parenchyma makes the approach
very promising [79]. In fact, MSC has been shown to repair damaged lung tissue, its effects
include re-epithelialization, replacement of injured vascular endothelium, promotion of
endothelial cell growth, induce generation of fibroblasts and macrophages, and even
angiogenesis [80].
Under disease conditions, following endothelial and epithelial injury, the alveolus contains a
large amount of edema and infiltration of inflammatory cells. MSCs release a number of
paracrine factors to help mediate the effects of the disease, including Ang-1, PGE 2, IL-1ra, TGF-
b, and KGF [78].
In animal models, intrapulmonary MSC administration has been shown to improve survival and
reduce pulmonary edema formation. Moreover, there was a significant increase in lung
function in the animals, accompanied with elevated repair mechanism in the epithelium and
endothelium [47, 91]. Also, when MSC or MSC conditioned medium was administered to cells,
the systemic inflammatory response seen in LPS-induced models of ARDS was mitigated, and a
balance between the release of pro- and anti-inflammatory mediators was established [92].
21
A number of studies have pointed out that pulmonary engraftment of MSC in the lungs is very
limited, the beneficial effects seen in MSC administration rests with paracrine effects, the
ability to secrete soluble factors to modulate immune responses of different diseases [93]. The
effects of paracrine secretion were demonstrated through the use of MSC conditioned medium.
Even though the exact paracrine constituents responsible for the therapeutic effects identified
in cells or animals following MSC conditioned medium treatment remain to be elucidated, there
has been evidence that suggests using the conditioned medium, rich in IL-1, growth factors,
such as VEGF-A, and TGF-b promoted attenuation of inflammatory response and stimulated
endothelial cell migration, contributing to repair of vascular tissue injured in an ischemic muscle
injury model [94].
1.3.3 MSC Confers Protective Benefits in Animal Models of ALI
Some of the most widely used approaches to induce ALI in animal models include, oleic acid,
endotoxin, acid aspiration, hyperoxia, CLP, and bleomycin. MSC administration has been found
beneficial in a number of such models. LPS is the most widely used endotoxin to elicit acute
lung injury in animal models.
In a recent work by Gupta et al., intratracheal MSC administration to mice following
intratracheal LPS administration showed significantly improved survival [47]. Histologically,
there was much less pulmonary edema in tissues; as well as a lowered wet-to-dry ratio. In a
another study, when LPS was administered intratracheally, intravenous MSC infusion produced
22
a significant drop in total BAL and neutrophil counts, 72h after the initial insult [95]. Moreover,
the Mei group also found that there was a significant decrease in inflammatory infiltrates,
interalveolar septal thickening, and interstitial edema, by examining the histology of lung slices.
The same group found that in a CLP model, MSC infusion resulted in decreased BAL fluid cell
count and decreased total protein, and large proteins, ie. albumin. There was also an
attenuated level of lung infiltrates and interstitial edema. Most interesting was that in animals
that received antibiotic therapy, MSC treatment led to a 50% reduction in animal mortality [96].
Also, there was reduced kidney and liver injury, improved glycogen storage and increased
bacterial clearance in the spleen.
23
1.4 MicroRNA
1.4.1 Background
The discovery of miRNAs was made twenty years ago as a major breakthrough. In 1993 the first
miRNA was identified in C.elegans [97], and eight years later, the highly conserved let-7 was
found to be crucial for developmental timing in C.elegans [98].
MicroRNAs are small, endogenous RNAs, approximately 20-22 nucleotides long. miRNAs
regulate gene expression at a post-transcriptional level and have direct effects on gene
expression by inducing mRNA degradation or translation inhibition. Recently, miRNAs have also
been documented to exert global effects indirectly through alteration of methylation and
transcription factor activity [99]. Over the past decade, interest in miRNAs has increased
exponentially, unfolding more of their importance in the regulation of biological processes –
development, proliferation, apoptosis etc. In comparison to the known number of miRNAs in
animals, relatively few miRNAs have been studied in detail, and their biological significance still
remains to be fully understood. In humans, close to half of the genome may be regulated by
miRNAs. Expression levels of miRNAs vary greatly among tissues and it is believed that
dysregulation of miRNAs can contribute to disease pathology [100, 101].
1.4.2 miRNA Processing and Function
Genes for miRNAs can be found in introns or exons of protein-coding genes, or in non-coding
regions of the genome. miRNA genes are much longer than the biologically active form, mature
miRNAs are only produced after a multi-step process [102] Once activated, the transcription is
24
carried out by RNA polymerase II or III, giving rise to a primary miRNA (pri-miR) structure that
contains a hairpin loop. Inside the nucleus, both ends of the pri-miR are cut by the
Drosha/DGCR8 complex, leading to the precursor of mature miR (pre-miR), with an
approximate size of 70–100 nucleotides [103, 104]
Transport of the pre-miR to the cytoplasm is via Exportin 5 where pre-miRNAs are subsequently
cleaved by RNase III Dicer into ~22-nt miRNA duplexes [102]. One strand of the miRNA duplex is
degraded, otherwise known as the "passenger" strand, indicated as miR*; whereas the other
"guide" strand, or simply miR, is incorporated into the RNA-induced silencing complex (RISC)
and serves as a functional, mature miRNA [105]. Selection of the "guide" strand is based on the
base pairing stability of both dsRNA ends [106].The argonaute (Ago) 2 protein, also known as
“slicer,” has the capacity to cleave target mRNAs [107].
Fundamentally, miRNAs are repressors of genic expression at the post-transcriptional level, by
degrading messenger RNA (mRNA) or inhibiting protein translation. Once miR is incorporated
into the RISC (miRISC), through the interaction between the complementary regions of the miR
and mRNA, the complex is able to recognize the target mRNAs. To date, most of the binding
sites between miR and mRNA are at the 3′ end of the non-coding region of the mRNA (3’UTR).
In animals, partial sequence complementarity between the miRNA and its target will guide the
complex to bind to miRNAs target sites, mainly within the 3′-UTR of genes. A “seed region” of 2-
8 nucleotides long exists in all miRNAs, and it is critical for miRNA target selection [102]. The
25
seed region is utilized by mature miRNAs to bind with specificty to miRNA recognition elements
within the 3'UTR of mRNAs.
miRNAs also function to inhibit translation initiation, which requires both the cap and poly(A)
tail structures in eukaryotes [108]. Translation is facilitated by both of these structures. It has
been demonstrated that miRNAs inhibit translation by hindering the assembly of the 80S
ribosomal complex in a cap (m7GpppG)-dependent manner [109]. Partially, this effect could
also be attributed to the recruitment of the anti-association factor eIF6 by RISC, which acts to
impede the assembly of the 80S complex [110]. Moreover, the Ago2 protein within the RISC
complex contains a cap-binding domain that is in direct competition with eIF4E and inhibits
binding of the initiation complex to the m7GpppG cap. Ago2 also has the capacity to compete
with eIF4E for its interaction with eIF4G, which further disrupts assembly of the initiation
complex [111]
Certain mRNA can be a shared target of many miRNAs, and each miR can suppress tens or
hundreds of genes. In addition, there is evidence that suggests that miRNAs can be transferred
from one cell to another, generating an interesting mechanism of intercellular regulation and
communication [112].
Given the fact that miRNAs have an enormous capability to influence the expression of a large
part of the genome, their deregulation also has significant impacts on the homeostasis of the
26
organism. An ever growing number of miRNAs have been shown to be involved in different lung
diseases, thus, it is critical that we gain a better understanding on their pathogenesis.
1.4.3 miRNA Regulation
miRNA regulation and expression is thought to have multiple levels of control. O'Connell et
al proposed three levels of regulation, namely at the stages of "(i) transcription, (ii) processing
and (iii) subcellular localization" [113]. At the level of transcription, induction of miRNA
expression can be modulated by transcription factors in response to inflammatory stimuli and
cellular stresses. During processing of miRNA, dicer inhibition or post-transcriptional
modifications may lead to a dramatic slow-down [114]. When miRNAs localize to stress
granules and p-bodies to alter local transcriptional levels, the exact mechanism for regulation
on miRNA still remains unknown. Epigenetic mechanisms for miRNA expression also exist.
Selective post-transcriptional modification on pre-miR can change a miRNA's targeting capacity.
Adenosine can be converted to inosine in this through RNA-editing [115]. The first such miRNA
was pre-miR-22 that was edited at various sites. Approximately, about 6% of human miRNAs
are subject to post-transcriptional editing [116, 117]. Interestingly, some of the edited sites
have been found within the “seed” region, which would most likely lead to an altered set of
targets for the miRNA.
27
1.4.4 Targets of miRNAs
As mentioned previously, single miRNAs can regulate the expression of many mRNAs, and each
mRNA is also likely to be regulated by several miRNAs simultaneously [118]. Hence, successful
identification of miRNA target genes is challenging and of great importance. A number of
computational algorithms are available to predict miRNA targets [119, 120]. Algorithms take
thermodynamic stability and conservation analysis in to account to maximize specificity when
predicting targets. Recently, many web-based applications have combined prediction programs
with functional databases to better forecast the role and effect of miRNAs [121].
The effect of a particular miRNA during health or disease rests in the function of its target
mRNAs. Several computational target prediction tools have been developed based on distinct
but overlapping algorithms, which are continuously modified and updated as more empirical
validation comes about. Generally, these algorithms rely on base pairing between the “seed
sequence” of the miRNA and the 3′-UTR of its target, on top of evolutionary conservation of the
targeted sequence. Although most studies have focused on functional mRNA-miRNA pairing
that resides in the 3′-UTRs, miRNA target sites can also reside in the coding region of a gene,
the close proximity of the arrangements of miRNAs and their targets may offer regulatory and
organizational advantage [122].
It may be beneficial to search multiple databases for potential targets of the same miRNA.
Nevertheless, not all putative targets are biologically genuine due to spatial or temporal
restrictions. Moreover, binding of the miRNA to the targeting site might be modulated by 3′-
28
UTR cis-acting sequences. Again, a single 3′-UTR may be targeted by multiple miRNA. Taken
together, the level of a mRNA or its translation product is governed collectively by these factors
that may act on its targeting miRNA.
Prediction algorithms identify putative miRNA:target pairing based on structural
complementarity, experimental validation is still necessary. Brennecke et al. [123] empirically
confirmed that nucleotides 2–8 from the 5′-end of a miRNA are the most essential in base
pairing with its target mRNA. The groups also found that the position of the base pairing was a
greater determinant in the functional efficiency of the miRNA than the pairing energy.
Due to high similarities in miRNA sequences, usually, a large number of putative miRNA binding
sites, as well as mRNA targets are produced by computational algorithms [124]. Thus,
experimental validation in biological system remains fundamental and essential to complete
the understanding of target prediction [121, 125]; antagomir studies or immunoprecipitation of
Ago-bound mRNAs have been specifically developed for miRNA-mRNA studies. Thus, using
highly specific monoclonal antibodies against members of the Ago protein family, Ago-bound
mRNAs can be co-immunoprecipitated [126]
1.4.5 Role of miRNAs in the Lung
In the lung, there is a very specific miRNA profile that shows highly conservation across
mammalian species [127]. Nonetheless, the role of miRNAs in physiological and pathological
conditions in the lung remains to be elucidated.
29
In normal lung development, a number of miRNAs have been shown to play a role. The miR-17-
92 cluster has been proposed to regulate lung development. Its expression is high in the
embryo, and steadily diminishes through development into adulthood [128]. It has been
demonstrated in mice deficient in the miR-17-92 cluster, such animals were not viable and
showed lung hypoplasia/ventricular septal defects. Moreover, B cells were much more inhibited
in the absence of the miRNA cluster [128, 129]. Interestingly, when the miRNA cluster was
overexpressed in murine models, terminal air sacs were absent in the lung, instead, highly
proliferative, undifferentiated pulmonary epithelium developed. As for miR-26a, its expression
has been associated with bronchial and alveolar epithelial cells in murine lung [130].
Transcription factor SMAD1 is one of the targets of miR-26a, making the miRNA a critical
regulator in lung development and pulmonary vascular remodelling [131]. Also, miR-155 is
crucially involved in the differentiation of naive T-cells into Th1 and Th2 cells [132, 133].
Animals that lacked miR-155 were immune-deficient and exhibited increased lung remodelling,
higher bronchoalveolar leukocytes and impaired T- and B-cell responses to inflammatory
stimuli [132]. As a key factor for normal granulocyte development in the lung, mutant mice for
miR-223 spontaneously developed neutrophilic lung inflammation with tissue destruction after
exposure to endotoxin.
1.4.6 miRNAs as Potential Molecular Therapies and Biomarkers in ARDS
Increasing amounts of evidence have shown that changes in miRNA expression are associated
with immune response, inflammation pathways, and the pathogenesis of inflammatory lung
30
disease including ARDS, miRNAs have great prospects to become a group of potential novel
therapeutic agents. The key function of miRNA regulation is through altering the abundance of
copy numbers of mRNAs [134]. Even with a moderate change in expression of a target gene,
the downstream effects could be potentially amplified, which may influence various
physiological pathways. For example, miRNAs that are responsible for down-regulating genes in
the inflammation pathway could be used to avoid over-expression of cytokines to repress acute
inflammatory response in patients. Moreover, miRNAs involed in ARDS pathogenesis could be
knocked down based on sequence complementarity [135].
microRNA expression is a dynamic process that changes in response to the ever-changing
cellular environment. Therefore, miRNAs could act as biomarkers to pinpoint
pathophysiological processes in specific disease states or development stages. With the
expected advances in understanding the relationships between miRNAs and ALI/ARDS through
genome-wide miRNA profiling, more miRNA biomarkers associated with immune response,
inflammation pathways, disease development, and disease severity could be evaluated as a
novel tool in the diagnosis and monitor for inflammatory lung disease.
31
1.5 Endothelial Permeability and Occludin
1.5.1 Background
In the endothelial and epithelial cell lining, the tight junction (TJ) forms a continuous,
circumferential belt separating apical and basolateral plasma membrane domains, constitutes
the principal barrier to passive movement of fluid, electrolytes, macromolecules and cells
through the paracellular pathway, and works as a barrier within the intercellular space and as a
fence within the plasma membrane. The paracellular pathway is defined as the space between
the cells and comprises both the TJ and the lateral intercellular space [136, 137]
Recently, the physical composition of TJ transmembrane molecules has received a lot of
attention, which has yielded much knowledge to the structure and function of TJs. Freeze-
fracture electron microscopy has been utilized to analyze the morphology of TJs – a set of
continuous, trans-membranous particle strands on the inner leaflet of the plasma membrane
with complementary vacant grooves on the outer leaflet [138]. Depending on the cell type, the
number and complexity of ramification of the TJ network varies, leading to differences in
permeability barrier function between different tissues. The permeability barrier formed by
polarized epithelial and endothelial cells is of fundamental importance in cell biology. [138, 139].
1.5.2 Regulation of Tight Junction Permeability
As a physical barrier, TJ control physiological fluxes. A wide range of growth factors, cytokines,
drugs, and hormones regulate the TJ and its barrier function. Some of the most well
documented regulators that enhance the barrier function include, glucocorticoid
32
hydrocortisone, prolactin, and unsaturated fatty acids, all of which increase the expression of
occludin in endothelial and epithelial cells [139]. It has also been found that astrocyte-
conditioned media is capable of elevating barrier properties and the expression of ZO-1 in
retinal and brain-derived endothelial cells through a mechanism that remains unknown [140].
On the other hand, cytokines and several growth factors have been demonstrated to increase
permeability and mediate discontinuous cell border staining of ZO-1 and/or occludin in
endothelial and epithelial cells, such as tumor necrosis factor-α, interferon-γ, interleukin-1β,
transforming growth factor-α, and platelet-derived growth factor [141, 142]. Additionally,
hepatocyte growth factor, VEGF, and histamine increase permeability as well through a similar
mechanism that leads to reduced expression of occludin or ZO-1 in vascular endothelial cells
[143]. Clearly, a diverse group of extracellular factors can stimulate changes in intracellular
signalling cascades which affect tight junction permeability. To date, the exact mechanism that
links signaling pathways to enhanced permeability in TJ are yet to be elucidated.
1.5.3 Occludin as a Membrane Component of the Tight Junction
Occludin was the first identified transmembrane component localized in the TJ. Its discovery
was made in chicken and has a molecular weight of 65 kDa [144]. Later on, occludin was found
in mammals as well [145]. In the protein, its transmembrane domains cross the membrane four
times, and the protein has a short cytoplasmic N-terminus and a much longer C-terminal
cytoplasmic domain. The protein consists of two extracellular sections – an N-terminal cytosolic
domain, and one at the C-terminus [144]. Occludin also has three cytoplasmic domains: one
intracellular short turn, a small amino terminal domain and a long carboxyl terminal region.
33
Occludin protein has shown slightly variable molecular weights in different animals. However,
in all species, occludin resolves as multiple bands on SDS–PAGE gel with the lowest being the
most abundant. The higher bands on the gel correspond to phosphorylation states of occludin –
bands of higher molecular weight are in the more phosphorylated state. In the extracellular
loops, a large number of tyrosine residues can be located; more significantly, more than half of
the residues are tyrosine and glycine in the first extracellular loop. Previously, studies have
shown that occludin enters a hyperphosphorylated state as the tight junction assembles itself.
Also, the main form of occludin that is found in the tight junction is hyperphosphorylated, and
the hyperphosphorylation leads to a substantial detergent insolubility [146, 147].
1.5.4 Occludin’s Interaction in the Tight Junction
In the TJ, occludin is associated with a number of structural proteins and functional proteins. It
has been reported previously, that occludin’s cytoplasmic domain, the C-terminus was able
bind to the GuK region of ZO-1 in vitro, strongly suggesting a possible physical link between
occludin to the actin cytoskeleton [147, 148]. Also, in parallel in vitro studies, ZO-2 and ZO-3
have been identified to bind to the C-terminus of occludin [149] Moreover, occludin has been
demonstrated to interact with a gap junction protein, connexin-32; and the co-localization
between the two proteins have been proven to exist at tight junctions. Kojima et al also showed
that exogenous expression of connexin-32 in mouse hepatocytes raised the expression of
occludin and increased the number of tight junction strands [150]. Aside from interactions with
34
other junctional proteins, occludin molecules also homodimerize via a C-terminal coiled-coil
domain [151].
Besides interactions between occludin and structural proteins, there has been established
interactions between occludin and a number of regulatory proteins at tight junctions. With the
use of yeast two-hybrid screening, c-Yes, connexin-26, protein kinase C, and p85 were shown to
bind to the C-terminus of occludin. These protein-protein interaction relationships were
identified using only a 27 amino acid sequence of the C-terminus of occludin that gives rise to a
putative coiled-coiled domain. However, the functional significance of these interactions has
yet to be elucidated [151]. For p85, as the regulatory subunit of phosphatidylinositol 3-kinase,
its interaction with occludin may lead to amplified downstream effects. Also, it has been
proposed that as a nonreceptor tyrosine kinase, the association between c-Yes and occludin in
an in vivo setting may potentially be responsible for tyrosine phosphorylation of occludin [152].
Moreover, occludin’s C-terminus has been found to bind to VAP-33. VAP-33 co-localized with
occludin both at the tight junction and in an intracellular vesicle pool. And when overexpressed,
VAP-33 led to a changed pattern of endogenous occludin localization. Lapierre and colleagues
proposed that VAP-33 either regulates vesicle targeting to tight junctions or the localization of
occludin at tight junctions [153]. At the other end of the protein, occludin may interact with
other molecules via its N-terminal cytoplasmic region. The E3 ubiquitin-protein ligase, Itch, was
found to associate with the N-terminus of occludin in vitro and in vivo, a possible ubiquitination
mechanism in occludin regulation. Taken together, occludin interacts with a wide array of
cellular proteins and may be regulated by diverse cellular signaling pathways.
35
1.5.5 Functional Role of Occludin in the Tight Junction
Functionally, occludin has been suggested to play a significant role in the regulation of tight
junction barrier function, and there is a large amount of evidence in support of it. In MDCK cells,
expression of chicken occludin substantially increased TER and increased the number of tight
junction strands compared to untreated cells [154] [155]. In contrast, when the second
extracellular loop of occludin (OCC2) was targeted by synthetic peptides, significantly decreased
TER and increased flux of several paracellular tracers in confluent monolayers of
an xenopus kidney epithelial cell line were documented [156]. Similarly, with the aid of TER
measurement, synthetic peptides homologous to regions of the first extracellular loop of
occludin were able to impede junction re-sealing following calcium depletion and re-addition
[157]. These aforementioned experiments indicate that the extracellular loops of occludin
exhibit adhesive properties and are critical in the formation of paracellular barriers.
In fact, on top of playing a critical role in regulating paracellular barriers, occludin may have a
vital role in the formation of paracellular channels. Overexpression of occludin led to increased
TER [154]. Overexpression of an occludin construct containing a C-terminal deletion also led to
increased TER, associated with an even higher amount of paracellular flux than overexpression
of the wild-type occludin. Interestingly, cells expressing occludin mutants with short deletions
within the extracellular loops were able to form electrically tight monolayers, without increase
paracellular permeability [155]. These lines of evidence have highlighted that intact
36
extracellular loops are essential for occludin to form paracellular channels, and the C-terminus
of occludin may act as a vital regulatory spot for flux through paracellular channels.
The functional role of the C-terminal cytoplasmic portion of occludin has received intense
research and been partly elucidated. In Madin-Darby bovine kidney cells, expression of C-
terminal deletion occludin mutants identified that the distal half of the C-terminus, or amino
acid residues 358–504 were critical and necessary for proper localization of occludin to tight
junctions [158]. The same amino acids were necessary for the interaction between occludin and
ZO-1. These studies demonstrated that ZO-1 binding was required for occludin localization at
tight junctions; however, binding to ZO-1 was not sufficient for the localization of occludin to
tight junctions, indicating that other factors may also be involved [158]. Moreover, a different
group also reported that the proximal region of the C-terminal tail to the plasma membrane
lacking the ZO-1 binding region was not sufficient to target occludin to tight junctions [159]. It
has been demonstrated that the binding of the C-terminus of occludin to ZO-1, is required for
proper localization of occludin at tight junctions, and for targeting occludin to the basolateral
membrane [160].
Cellular adhesion has been shown to partly depend on occludin expression. Through cross-
species transfection studies, Furuse et al have demonstrated that chicken occludin was able to
promote cellular adhesion in insect cells [161]. Due to a lack of ZO-1 expression in insect cells,
although occludin did not localize to cell borders, occludin was detected in multilamellar
cytoplasmic structures that resembled tight junctions. In a parallel study, occludin was
37
introduced to several kinds of occludin-deficient fibroblasts, to establish its role in cell–cell
adhesion. In fibroblast cell lines that expressed adherens-like junctions and ZO-1 at sites of cell–
cell contact, exogenous occludin was able to localize to cell–cell contact surfaces and led to
cell–cell adhesion [162]. In contrary, as identified by the same group of researchers, in a
fibroblast cell line that lacked adherens-like cell junctions or localization of ZO-1 at cell–cell
contacts, exogenous occludin was not able to localize to cell–cell contacts or confer cell–cell
adhesion. This evidence suggests that ZO-1 was required to recruit occludin to cell junctions. It
was also found that the Cell–cell adhesion in fibroblasts could be eliminated when cells were
treated with synthetic peptides corresponding to the first extracellular loop of occludin,
stressing the key role played by this region of occludin in cell adhesion [162]. Given its ability to
promote cell adhesion, occludin may likely play critical roles in the formation of paracellular
barriers.
1.5.6 Disturbance of the Barrier Function
Because tight junctions are located at the apical most areas of basolateral spaces between
epithelial, or endothelial cells, they are involved in a wide variety of pathological conditions
where the normal physiological regulation of passage of ions, molecules, and inflammatory cells
may be compromised. It is particularly clear that tight junctions are targets of pathogens such
as bacteria, viruses, allergens, etc.
Cytokine production can be identified in nearly all kinds of pathological conditions. For instance,
the proinflammatory cytokines TNF-α and interferon-γ downregulate the expression of occludin,
38
causing dysfunction of tight junctions [163]. Also, both factors can cause redistribution of JAM
in endothelial cells, which is also another key player in maintaining tight junction barrier
function [164]. In certain inflammatory diseases, such as the inflammatory bowel disease, there
are increases in intestinal permeability, likely due to reduction in occludin expression [165, 166].
Generally, inflammation is always accompanied by an increase in vascular permeability, which
can be partly attributed to the increased expression of vascular endothelial growth factor
(VEGF), which primarily affects the barrier function of tight junctions by both phosphorylation
and down-regulation of occludin [167, 168]
39
Objectives and Hypotheses
In sepsis treatment, we do not have a complete understanding of the molecular responses to
injury, the development of immunosuppressive complications, the limitations of delivery
systems to specific organs, and the role of timing and combination therapy34-37. And due to our
lack of knowledge in the molecular mechanisms of sepsis, a considerable number of clinical
trials have failed16,31-33. Recently, MSC administration has been shown to have immune-
regulatory and beneficial effects in a number of animal models of lung diseases, however, the
exact mechanism by which the beneficial effects are conferred by MSCs remains unknown.
General Objective of the Project
Despite the fact that there are no specific pharmacological treatments for ARDS, post-
translational gene regulatory processes offer the hope for new interventions. In this project, we
used an informed target discovery approach on genome-wide analysis of functional
transcriptional regulation in response to MSC treatment of sepsis, aiming to identify
molecules/pathways that play a role in MSC-conferred protection from sepsis-induced mortality
and ARDS, and to better understand the protective mechanisms of benefit by MSC
administration.
With the use of microarray-based genome-wide analysis, we aim to determine the role of MSC
administration in the modulation of pulmonary host-responses to sepsis via differential
transcription of regulatory microRNAs.
40
By analyzing the transcriptional response to polymicrobial sepsis, the molecular phenotype
associated with the disease can be defined. In our experimental model of sepsis, the cecal
ligation puncture model, mice were fluid resuscitated and antibiotic treated. MSC
administration in sepsis-induced mice was shown to reduce mortality and increase bacterial
clearance. In order to better understand the exact mechanism by which MSCs confer their
beneficial effects, we used microarray to examine the global gene expression, and identify
expression signatures that may potentially be therapeutically relevant for ARDS.
In order to address the overall hypothesis, we established the following aims to examine
individually.
Aim 1: Identification of miRNAs that are differentially expressed in mouse septic lungs after
treatment with MSCs
We hypothesized that by combining miRNA and mRNA profiles from murine model of CLP-
induced ARDS treated with either vehicle or MSCs will provide novel insight into the regulation
of gene expression during sepsis.
First, to validate our in-silico hypothesis described above, we screened for differentially
expressed miRNAs in the lung. miRNA profiling was performed using the Exiqon miRCURY LNA™
microRNA Arrays. To gain insight into miRNAs involved in the regulation of the transcriptional
response to polymicrobial sepsis, specifically in the case of sepsis-induced ARDS, we performed
organ-specific analyses to examine expression profiles of miRNAs. Two main comparisons were
41
considered: 1) Sham vs. CLP+placebo (CLP); and 2) CLP+placebo (CLP) vs. CLP+MSC (MSC).
Results were normalized using Lowess normalization. Determination of differential miRNA
expression was performed in R/bioconductor using the LIMMA package. miRNAs with a false
discovery rate (FDR) of ≤ 0.05 were considered differentially expressed.
Aim 2: Selection of candidate miRNAs for “functional validation” based on mRNA:miRNA
profiling
Recently, our group successfully used genome-wide analyses to identify 3 prominent effects of
MSC administration on sepsis-induced MODS in target organs. These effects are 1)
reconstituted transcription of mitochondrial related genes; 2) down-regulated innate immune
pro-inflammatory transcriptional responses; and 3) coordinated expression of transcriptional
programs implicated in the preservation of endothelial/vascular integrity.
Microarray analysis was used to identify expression profiles associated with common pathways
of MSC-mediated protection and repair from sepsis-induced ARDS in the lung. Similar to the
miRNA microarray, two main comparisons were considered: 1) Sham vs. CLP+placebo (CLP);
and 2) CLP+placebo (CLP) vs. CLP+MSC (MSC). Results were normalized using Lowess
normalization. Determination of differential miRNA expression was performed in
R/bioconductor using the LIMMA package. Differentially expressed genes were defined as those
having a False Discovery Rate (FDR)-corrected p-value <0.05 (adjusted p-value, corrected for
multiple comparisons).
42
To generate a list of putative target genes, we used a prediction algorithm, miRDB, for miRNAs
that displayed the most substantial expression changes. This list was then merged against the
list of differentially expressed genes as determined using the two comparisons: 1) Sham vs.
CLP+placebo (CLP); and 2) CLP+placebo (CLP) vs. CLP+MSC (MSC).
Based on the merged list, miRNAs that have differentially regulated target genes that are
relevant to the ARDS would be selected for functional validation. Putative target genes were
screened to determine whether changes in their expression would elicit a corresponding
phenotype associated with ARDS.
Aim 3: Determination of molecular regulation of candidate miRNAs and their putative targets
in vitro and in vivo
Based on our bioinformatics data, we were able to determine that miR-193b-5p was
differentially up-regulated in mice subjected to CLP surgeries, and was also responsive to MSC
administration following the CLP insult. With mirdb.org, an online prediction algorithm, we
were able to identify occludin, a tight junction protein found in the endothelium, as a putative
target gene of miR-193b-5p.
We hypothesized that through manipulation of miR-193b-5p expression, endothelial/vascular
integrity can be maintained in response to TNFa in the HPMEC cell model.
43
Thus, the potential association between miR-193b-5p and occludin made it possible for miR-
193b-5p to be a critical regulator of the endothelial integrity. Moreover, the mechanism of our
miRNA of interest may be able to explain, in part, the beneficial effects that MSCs exert on the
lung in sepsis-induced acute lung injury.
To validate our in silico findings, we selected HPMEC cells as our in vitro model. HPMEC cells
were treated with TNFa to simulate an inflammatory state. Through overexpression and
knockdown of miR-193b-5p, with or without TNFa stimulation, permeability measurements,
and luciferase reporter assay, we aimed to establish a regulatory relationship between miR-
193b-5p and occludin.
44
Materials and Methods
RNA Isolation
Total RNA was isolated from lung tissues of mice treated with Sham, CLP, and CLP + MSC, for
microarray, as well as for qRT-PCR studies. For tissue samples, 1mL of Trizol (Life Technologies,
Rockville, MD) was added for every 1000mg of tissue. For in vitro cell experiments, once the
treatment was completed, medium was removed, and 0.5 ml of TRIzol reagent was added to
each well in a 6-well plate. Cells were placed at 4°C, and total RNA extraction was performed.
Chloroform was added to the samples at a 1:5 ratio with the volume of Trizol. Samples were
centrifuged at 12,000g for 25min to separate RNA from DNA and proteins. Supernatant of each
sample was transferred into a new eppendorf tube, and 100% isopropanol was used to first
wash the RNA at 12,000g for 25min. The supernatant was removed and 95% ethanol was added
to the pellet to complete the second wash at 7,500g for 5min. The final pellet product was air-
dried then re-suspended in 20ul of DEPC RNase free water. RNA quality was ensured by
spectrophotometric analysis (OD260/280).
mRNA Microarray
Total RNA extracted from mouse lung tissues was isolated using Trizol and purified using Qiagen
was hybridized to the Illumina Mouse WG 6v1.1 expression bead array as per manufacturer’s
specifications (n=5 animals/group). Probe based analysis was performed in the R-project for
Bioconductor. Variance-stabilization transformation was used to refine normalization.
45
Principal component analysis was performed in Partek (St. Louis, Missouri). For clustering, Log2-
transformed normalized expression data were gene-centered (mean). Hierarchical clustering
was performed in JMP and in Partek using a correlation matrix and Ward linkage.
microRNA Microarray
microRNA profiling was performed using the Exiqon miRCURY LNA™ microRNA Arrays
(http://www.exiqon.com/microrna-microarray-analysis-microrna-array). The RNA samples
isolated from mouse lung tissues were labeled using the miRCURY LNA™ microRNA Hi-Power
Labeling Kit. First, 5’-phosphates were removed from the microRNA termini using Calf Intestinal
Alkaline Phosphatase (CIP). Second, a fluorescent label was attached enzymatically to the 3’-
end of the microRNAs in the total RNA sample. This was followed by an enzyme inactivation
step after which the sample was ready for hybridization to the microRNA Array (6th gen) can
measure the abundance of 1081 possible features/miRNAs. Results were normalized and
determination of differential miRNA expression was performed in R/bioconductor using the
LIMMA package. miRNAs with a FDR of ≤ 0.05 were considered differentially expressed. Target
genes were predicted for differentially expressed miRNAs according to the algorithm of mirdb
and then compared to changes in genome-wide expression levels.
Before the preprocessing, we kept only the mice-specific probes on the arrays, which includes
4324 probes. Each miRNA has printed 4 times on the arrays so we have total 1081 miRNAs. We
took the median of the within-array normalized intensities of the 4 probes for a given miRNA.
46
To gain insight into miRNAs involved in the regulation of the transcriptional response to
polymicrobial sepsis, we analyzed the 3’UTR region of MSC-responsive genes in the lung to
identify enrichment for putative miRNAs.
Microarray Data Analysis and Normalization
Differential gene expression in the lung was determined by computing empirical Bayes
moderated t-statistics with the LIMMA package. Two main comparisons were considered: 1)
Sham vs. CLP+placebo (CLP); and 2) CLP+placebo (CLP) vs. CLP+MSC (MSC). Differentially
expressed genes were defined as those having a False Discovery Rate (FDR)-corrected p-value
<0.05 (adjusted p-value, corrected for multiple comparisons).
The LIMMA R package was used for preprocessing and differential analysis. For preprocessing,
we performed background correction using “normexp” option and set the offset to be 50. This
method adjusts the foreground adaptively for the background intensities and results in strictly
positive adjusted intensities. The use of an offset damps the variation of the log-ratios for very
low intensities spots towards zero [169]. Within-array normalization was performed using Loess
normalization approach which assumes that the bulk of the probes on the array are not
differentially expressed. It doesn't assume that that there are equal numbers of up and down
regulated genes or that differential expression is symmetric about zero [170]. Between-array
normalization was carried out using the Quantile normalization approach. The differential
analysis is based on an Empirical Bayes approach [170]. The false discovery rate was controlled
using Benjamini-Hochberg approach [171].
47
CLP Model of Sepsis
C57 mice were anesthetized with 200 mg/kg ketamine (Ketalean, 100 mg/ml; Bimeda-MTC
Animal Health Inc., Cambridge, ON, Canada) and 10 mg/kg xylazine hydrochloride (Rompun, 20
mg/ml; Bayer Inc., Toronto, ON, Canada) via intraperitoneal injection. The mice were shaven
and cleaned with 70% ethanol. The cecum was brought outside of the abdomen and was
ligated 1 cm from the apex with 3–0 silk suture and double punctured with an 18-gauge needle.
The abdomen was closed with a 4–0 nylon suture. For sham surgery, where the cecum was
exteriorized but not ligated or punctured, was performed in control animals. Animals were fluid
resuscitated with 50 ml/kg saline injected subcutaneously. Twenty-four hours after initial
procedures, mice were killed to collect tissues for analysis.
MSC Administration
Mice that had undergone CLP procedures were given either saline or MSCs 6 hours after initial
surgeries. Animals were anesthetized by xylazine and ketamine via intraperitoneal injection. By
blunt dissection, the jugular vein was exposed and a small canula (PE10) was placed into the
vein. Saline or 2.5 × 105 male MSCs were slowly injected 6 hours after sham operation or CLP.
When the injection was completed, the vein was ligated, and the incision was closed using
nylon suture.
Quantification of miRNA and mRNA by qRT-PCR, Primer Design
All the quantitative realtime polymerase chain reaction (PCR) was performed with ViiA PCR
system (Applied Biosystems, Foster City, CA, USA) in triplicate. Total RNA was extracted from
48
HPMEC cells or mouse lung tissues with Trizol. For quantification of miRNAs, 500 ng of total
RNA was reverse transcribed to cDNA following a miRNA-specific approach [172] using
Superscript II RT Kit (Invitrogen) with miR-191 as loading control. miR-191 was previously
reported to be the most suitable gene for endogenous control due to the lowest stability values,
and the least amount of intergroup variation in its expression level [173].
For quantification of mRNAs, 500 ng of total RNA was reverse-transcribed using Superscript II
RT Kit (Invitrogen) and PCR was performed with SYBR Master Mix (Life Technologies, Rockville,
MD) with GAPDH as loading control. GAPDH was examined in our mouse lung tissue samples
and cell samples, and showed consistent expression levels across various treatment groups in
our model. All the real-time PCR analyses were performed in triplicate.
For mature miRNAs, miRNA-specific primers are listed in Table 1 below. Primers used in the
reverse transcription step were designed so that the last 8bp would be reverse complementary
to the last 8bp of the mature miRNA sequence, offering specificity in the binding; and the
preceding sequence of the RT primer would form a stem loop which helps maintain the stability
of structure. For the PCR primers, the forward primer was designed to be complementary to
the mature miRNA sequence that lied before the 8bp reverse complementary to the RT primer.
And the reverse primers were designed to contain a sequence that was entirely the same as a
segment of the RT primer.
49
For immature precursor forms of miRNAs, the RT primer and the reverse PCR primer were the
same. Also, the forward and reverse PCR primers were designed to largely overlap with the
mature miRNA sequences in the precursor step loop structure. The forward primer shared a
substantial overlap with the 5p mature sequence, the reverse with the 3p mature sequence.
miR RT PCR Forward PCR Reverse
193b-3p
CTCAACTGGTGTCGTGGAGTCGGCAATTCAGAGCGGGAC
ACACTCCAGCTGGGAACTGGCCCTCAAA
TGG TGT CGT GGA GTC G
193b-5p
CTC AAC TGG TGT CGT GGA GTC GGC AAT TCA GTC ATC TCG
ACA CTC CAG CTG GGC GGG GTT TTG AGG G
TGG TGT CGT GGA GTC G
323a-5p
CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAG GCGAACGC
ACACTCCAGCTGGGAGGTGGTCCGTGGC
TGG TGT CGT GGA GTC G
423-5p
CTC AAC TGG TGT CGT GGA GTC GGC AAT TCA GAA AGT CTC
ACA CTC CAG CTG GGT GAG GGG CAG AGA GC
TGG TGT CGT GGA GTC G
762 CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGGCTCTGTC
ACACTCCAGCTGGGGCTGGGGCCGGGACAGAGC
TGG TGT CGT GGA GTC G
883a-5p
CTC AAC TGG TGT CGT GGA GTC GGC AAT TCA GGT AAC TGC
ACA CTC CAG CTG GGT GCT GAG AGA AGT A
TGG TGT CGT GGA GTC G
191-5p
CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGCAGCTGCT
ACACTCCAGCTGGGCAACGGAATCCCAAA
TGG TGT CGT GGA GTC G
pre-193b
ACCCCAAAAGCGGGACTTTGAGGGCCAGTT
TCTCAGAATCGGGGTTTTGAGGGCGAGATGA
ACCCCAAAAGCGGGACTTTGAGGGCCAGTT
Table 1. miRNA-specific primers for cDNA synthesis and PCR.
50
General Cell Culture
HPMEC cells were selected as our in vitro model to validate the in silico findings. Lung
endothelial injury is one of the key characteristics of ARDS. And, recent studies have shown that
MSCs secrete paracrine factors that modulate immune responses and alter the responses of the
endothelium to injury. These two pieces of knowledge make the choice of HPMEC as our in
vitro model highly suitable to imitate the insult and rescue what would otherwise be observed
when sepsis was inflicted in mice. HPMEC cells were obtained from Promocell (Heidelberg,
Germany). Cells were cultured in Endothelial Cell Basal Medium MV 2 (EBM-2, C-22221;
PromoCell) supplemented with 10% (vol/vol) fetal bovine serum (FBS) (Lonza), hEGF,
hydrocortisone, gentamicin, amphotericin-B, VEGF, hFGF-B, R3-IGF-1, and ascorbic acid. Cell
cultures were incubated at 37°C in a humidified atmosphere containing 5% CO2. All
experiments were conducted with cells younger than passage six.
Transfection – Gain and Loss of Function
Our miRNA of interest, miR-193b-5p has complete sequence homology between mouse and
human. In the transfection experiments, the reagents used, including anti-hsa-miR-193b-5p
miScript miRNA Inhibitor (Cat# MIN0004767), hsa-miRNA inhibitor negative control (Cat#
1027271), Syn-hsa-miR-193b-5p miScript, hsa-miRNA Mimic (Cat# MSY0004767) and AllStars
siRNA Negative Control (Cat # 1027280) were purchased from Qiagen. Transfection
experiments were carried out in triplicate as per manufacturer’s instructions using HiPerFect
Transfection Reagent (HPF, QIAGEN Cat# 301705). 10^6 Cells were seeded in 6-well plates and
allowed to grow to >80% confluency prior to any transfection. A final concentration of 50nM for
51
miR-193b-5p inhibitor was used, and 5nM for miR-193b-5p mimic transfection. In a 6-well plate
(Falcon, BD Biosciences), 18uL of miR-193b-5p inhibitor was diluted into 100uL of EBM-2 along
with 18uL of HiPerFect transfection reagent. The transfection mixture was incubated for 10 min
at room temperature (RT) before being added to 500uL of serum free culture medium in each
well. Five hours into the transfection process, 500uL of additional serum-containing culture
medium was added to each well. Twenty four hours later, the cells were treated with
recombinant TNFα (10ng/ml) (Life Technologies, Rockville, MD) for an additional 24 hours
before RNA or protein was collected.
Transfection – Co-transfection of Luciferase Reporter Construct and miRNA Mimic/Inhibitor
For co-transfection of luciferase reporter constructs with miRNA mimic, inhibitor, or non-
targeting control, cells were in a 96-well white TC plate in 100μl total volume of antibiotic free
media. The transfection experiments were conducted as per manufacturer’s manual
(Switchgeargenomics, Menlo Park, CA). Two mixtures were added to each well. In mixture 1,
100ng of luciferase reporter construct, and 50nM (final concentration) of miR-193b-5p inhibitor
or 5nM (final concentration) of miR-193b-5p mimic were used, and serum free media topped
up the volume to 10uL. In mixture 2, 0.5uL of HiPerFect transfection reagent was added to
9.5uL of serum free media. The mixtures were incubated at room temperature for 10 minutes
before added to each well.
52
Luciferase Reporter Assays
Whether miR-193b-5p would directly inhibit the activity of a luciferase reporter construct,
pLightSwitch_3UTR (SwitchGearGenomics), constructs that contained the human OCLN-3’UTR
and constructs that contained the control 3’UTR from a non-target gene were compared
against each other. HPMEC cells (2*10^4 per well) were seeded onto 96-well plates. The
following morning, cells were transiently transfected using HiPerFect reagent (Qiagen, Venlo,
Netherlands) with luciferase reporter that contained either (a)the complementary seed
sequence for miR-193b-5p in the 3’UTR of occludin, or (b) a mutated version of the 3’UTR
sequence of occludin where the complementary seed sequence had been removed. Prior to
reading the plate, 100uL of LightSwitch Assay Reagents were added to each well and incubated
in a dark room at room temperature for 30min. For co-transfection experiments with miR-
193b-5p mimic, the plate was read with a spectrophotometer (Thermoscientic, Waltham, MA)
24 hours post- transfection; for co-transfection experiments with miR-193b-5p inhibitor, the
reading took place 48 hours post-transfection.
The size of the reporter prior to 3’UTR insertion is approximately 3910bp long. After the
addition of the OCLN 3’UTR, the total length of the reporter construct will reach a length less
than 5200bp. RenSP expression is responsible for the chemo-luminescence, and the 3’UTR will
be found upstream to it.
53
Western Blot
Tissue and cell lysates were analyzed by SDS-PAGE using 10% polyacrylamide gels (Biorad,
Hercules, CA). Proteins were transferred to nitrocellulose membranes, blocked for 1 hour, and
probed overnight with the primary antibody (1:1000 for occludin) at 4°C. After washing in TBS
with Tween 20, blots were incubated with horseradish peroxidase–conjugated secondary
antibodies for 1 hour, washed, and then visualized. For protein quantification, films were
scanned and analyzed using Image J (National Institutes of Health). The image of the gel was
converted to grey scale, and a rectangular box was drawn around the first lane and applied to
the other lanes in the image. The density of each of the bands was then converted to a peak,
and the “straight line” tool was employed to close the peak to generate an enclosed area
representative of the band’s density. The “wand” tool was used to determine the area of the
peak. The integrated intensity was normalized to the amount of protein loaded.
Cell Viability Assay
The effect of hsa-193b-5p mimic or inhibitor on cellular viability in the presence or absence of
TNFα was determined using MTT assay (Sigma, St. Louis, MO). Cells were seeded at a density of
10^4 cells/100 µl in a 96-well plate. After treatments of transfection and TNFa insult, cells were
incubated overnight. After washing twice with PBS, cells were given MTT (3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyltetrazoliumbromide) solution (100 ul/well of 5 mg/ml
solution in PBS) for 3 h. This assay was used to assess cell viability by examining the activity of
cellular enzymes that reduced the yellow tetrazolium dye, MTT, to the insoluble formazan,
giving rise to the observed purple color in living cells. Next, 100 µl/well DMSO (Sigma, St. Louis,
54
MO) was added to dissolve the converted purple formazan, and the absorbance of formazan
was measured at 570 nm using an ELISA reader (BioTek, Winooski, VT, USA). Absorbance values
were normalized to media control.
Transwell Assays of HPMECs
HPMECs, seeded on 0.4 µm-pore polyester transwells (Costar, Tewksbury, MA) that were
coated with Attachment Factor (Invitrogen), were grown until confluency 2–3 days later.
Confluency of the monolayer was determined using transendothelial electrical resistance (TEER)
and microscopic observations. FITC-dextran (70kDa) and TEER were the two independent
approaches employed to measure the changes in endothelial monolayer integrity.
Following TNFα and/or transfection of the miRNA inhibitor/mimic, FITC-dextran was added to
the top chamber of the transwell. Then, transwells were incubated at room temperature for 3
hours, then the media in the top and bottom portions of the transwells were collected and
transferred into a 96-well plate for reading, using spectrophotometer (Thermoscientic,
Waltham, MA). Dextran permeability readings were then measured and compared to the
control baseline.
TEER was measured using the Endohm-12 (WPI, Florida). When the monolayer of HPMECs was
determined to be near confluency, cells were transfected with miR-193b-5p inhibitor/mimic,
followed by treatment of TNFα (10ng/mL) for 24 hours. TEER was measured at different time
55
points following the treatment, 0h, 3h, 5h, and 24h. TEER readings were then measured and
compared to baseline.
Statistics
All experiments were performed at least 3 times unless otherwise stated. Data are expressed as
mean. Statistical values were defined using two-way ANOVAs as appropriate. A p value <0.05
was considered significant.
56
Results
In order to identify the regulatory relationship between septic stimuli and tissue function in the
lung, we undertook two microarray profiling studies, (one for mRNA, the other for miRNA),
using total RNA extracted from murine lungs that had been subjected to Sham, CLP surgeries
for 24h and treated with either placebo or MSCs.
Effect of MSC Treatment on Global Gene Expression Profiles of Lungs in Experimental Sepsis
A total of 45282 genes from 5 different animals per group were profiled using Illumina Mouse
WG 6v1.1 expression bead arrays. Three electronic experiments were conducted based on the
microarray data. First, all of the probes in the lung were processed with zero filtering, and we
obtained a global expression signature of the 45282 genes. Second, probes that showed
differential expression from the LIMMA analysis (adj p <0.05), in the CLP vs. Sham comparison
were plotted. Lastly, probes that showed differential expression from the LIMMA analysis (adj p
<0.05), in the CLP vs. MSC comparison.
Principal component analysis (PCA) is a mathematical procedure where orthogonal
transformation is employed convert a set of observations of possibly correlated variables into
linearly uncorrelated variables. PCA is a principal axis rotation of the original variables that
preserves the variation in the data. In PCA, the data points are rotated around their mean in
order to align with the principal components. Therefore, the total variance of the original
variables is equal to the total variance of the principal components. The eigenvectors and
eigenvalues define the rotation and variation. Eigenvalues are the variances of the principal
57
components. The eigenvectors are the direction cosines of the new axes, PCs, relative to the old,
original variables, thus they define the rotations of the original axes.
PCA was used to account for the variances among samples, and normalized unfiltered genes
revealed considerable differences in apparent gene expression in the lung tissues profiled. We
decided to first determine the genes that were responsive to the CLP insult. Between sham and
CLP, the first three components explain 74.7% of the variability in the data, and a total of 11451
probes were significantly different between the two treatment groups (Figure 1). As for
between CLP and MSC, a total of 4752 probes were different, and the first three components
were able to explain 79.0% of the variability in the data (Figure 3). PCA was used to determine
the molecular responses to CLP treatment based on gene expression profiles. As shown in
hierarchical clustering plots, MSC treatment restored transcriptional responses to an expression
profile intermediate to both the sham and CLP profiles in the lungs (Figure 2).
58
Figure 1. Principal component analysis (PCA) of the mRNA data that characterizes the trends
of global gene expression between Sham and cecal ligation puncture (CLP). As exhibited by
the gene profiles of Sham (green), cecal ligation perforation (CLP, red), and cecal ligation
perforation followed by mesenchymal stem cell treatment (MSC, blue), each mouse lung
sample is represented by one spot, or the eigenvalue vector, and each color represents a
different type of sample. PCA finds the 3 most important components from all the genes and
uses them to create the 3d view. The percentage of the variance in genetic distance explained
by each PC is shown on the axis, and a total of 74.7% could be explained by the first 3 axes.
59
Figure 2. hierarchical clustering using two-channel microarray data of genes that displayed
differential expression between Sham and CLP (adj P < 0.05). The rows represent individual
genes, and the mouse samples are shown as columns (1 column per sample). The color
represents the expression level of the gene. By convention, relative to the global change in
gene expression, red represents up-regulation, while green down-regulation. The dendrograms
constitute a quanlitative means of assessing the distance between individual genes and mouse
lung samples. The figure was generated by Partek V6.6.
60
Figure 3. Principal component analysis (PCA) of the mRNA data that characterizes the trends
of global gene expression between cecal ligation puncture (CLP) and cecal ligation perforation
followed by mesenchymal stem cell treatment (MSC). As exhibited by the gene profiles of
Sham (green), CLP (red), and MSC (blue), each dot represents a mouse lung sample and each
color represents a different type of sample. The percentage of the variance in genetic distance
explained by each PC is shown on the axis, and a total of 79% could be explained by the first 3
axes. CLP shows the greatest ellipse, while sham shows the smallest cluster. The ellipses
enclose data points of the same class. It is drawn at a distance beyond which the probability of
a point belonging to the class is low, and can be thought of as a class boundary. Along the first
principal component, CLP shows the greatest variance, and the MSC ellipse lies in between the
Sham and CLP.
61
Effect of CLP and MSC Treatment on miRNA Expression Profiles of Lungs in Experimental
Sepsis
A total of 1081 miRNAs from 5 animals per group were profiled using Exiqon bead arrays.
Volcano plots were generated based on fold change and adjusted p value from LIMMA analysis
for individual miRNAs. Between sham and CLP, as shown in Figure 4, up- or down-regulation
was observed depending on the miRNA, and the orange line represents a cut-off that selects
only the top 100 miRNAs. miR-193b-5p was observed to be up-regulated. Similarly, between
CLP and MSC, up- or down-regulation was observed depending on the miRNA, and the orange
line represents a cut-off that selects only the top 100 miRNAs (Figure 5), and miR-193b-5p
showed down-regulation. In order to study the role of miRNAs in ALI, we intended to look for
the ones that were differentially expressed following CLP surgeries, but also showed significant
responses to MSC treatment. Therefore, the expression of the miRNA expression following MSC
treatment would be on a similar level compared to that of the Sham group (Figure 6). As shown
in Figures 4 and 5, miR-193b-5p shows differential expression following CLP insult, as well as in
response to MSC treatment. Moreover, miR-193b-5p exhibits a complete homology between its
human and mouse sequences, making it highly relevant for future clinical studies.
62
Figure 4. Volcano plot of statistical significance against fold change between Sham, and CLP.
Each dot corresponds to one miRNA. The log fold change is plotted on the x-axis and the
negative log10 p-value is plotted on the y-axis. Note that not all genes with a fold change of 1.5
or more have significant adjusted p values (the adjusted P values are shown on the vertical axis
of the volcano plot), conversely, not all the genes with significant P values satisfy the fold
change cut-off. False discovery rate (FDR) was used as a cut-off, the corresponding adjusted P
value was 0.05
microRNA Expression Microarray
-0.3 -0.2 -0.1 -0.0 0.1 0.2
1
2
3
4
5
Log Fold Change Sham vs. CLP
-Lo
g1
0 (
P-v
alu
e)
Sh
am
vs.
CL
P
miR-193b-5p .
63
Figure 5. Volcano plot of statistical significance against fold change between CLP and MSC.
Each dot corresponds to one miRNA. The log fold change is plotted on the x-axis and the
negative log10 p-value is plotted on the y-axis. Note that not all genes with a fold change of 1.5
or more have significant adjusted p values (the adjusted P values are shown on the vertical axis
of the volcano plot), conversely, not all the genes with significant P values satisfy the fold
change cut-off. False discovery rate (FDR) was used as a cut-off, the corresponding adjusted P
value was 0.05
microRNA Expression Microarray
-0.2 -0.1 0.0 0.1 0.2
0.5
1.0
1.5
2.0
2.5
Log Fold Change MSC vs. CLP
-Lo
g1
0 (
P-v
alu
e)
MS
C v
s.
CL
P
miR-193b-5p .
64
Figure 6. Volcano plot of statistical significance against fold change between Sham vs. MSC.
Each dot corresponds to one miRNA. The log fold change is plotted on the x-axis and the
negative log10 p-value is plotted on the y-axis. As shown in the plot, miR-193b-5p did not
exhibit an expression that was statistically significant or past the fold change off, displaying a
similar expression pattern between Sham and MSC. Note that not all genes with a fold change
of 1.5 or more have significant adjusted p values (the adjusted P values are shown on the
vertical axis of the volcano plot), conversely, not all the genes with significant P values satisfy
the fold change cut-off. False discovery rate (FDR) was used as a cut-off, the corresponding
adjusted P value was 0.05
microRNA Expression Microarray
-0.4 -0.3 -0.2 -0.1 -0.0 0.1 0.2
1
2
3
4
5
Log Fold Change MSC vs. Sham
-Lo
g1
0 (
P-v
alu
e)
MS
C v
s.
Sh
am
65
Functional Prediction of a Differentially Expressed miRNA to MSC Treatment of Experimental
Sepsis
To develop a functional role of the identified hsa-miR-193b-5p, we used prediction algorithms
to determine putative candidate target genes that our miRNA of interest could potentially
regulate. After entering the name of the miRNA into the mirdb.org database, the putative
targets were prompted. The prediction algorithm used the seed region of the miRNA, 2nd to the
8th base, as a criterion of complementarity between the miRNA and mRNA. Out of the 268
putative targets of hsa-miR-193b-5p predicted by mirdb.org, 23 of them were in fact also
documented in our list of mRNAs in the lung that showed differential expression, one of such
targets was occludin, with a target score of 58.
To further validate the prediction, we joined our list of differentially expressed mRNAs in
response to MSC treatment of experimental sepsis against the list putative targets of hsa-miR-
193b-5p generated by mirdb.org. Occludin was also confirmed as one of the top predicted
target genes of hsa-miR-193b-5p, with a fold change of 1.50, and adjusted p value of 0.034
(Table 2). With the use of hierarchical clustering, it was obvious that MSC treatment returned
occludin’s expression profile to one that lies intermediate to both sham and CLP (Figure 7).
66
Figure 7. Heat map of occludin mRNA expression in Sham, CLP, and MSC mouse lung tissues.
Hierarchical clustering depicts the relatedness of the 15 samples, and the mouse samples are
shown as columns (1 column per sample). The color represents the expression level of occludin
mRNA. Red represents high expression, while green low expression. The dendrogram
constitutes a quanlitative means of assessing the similarity between individual samples. As
indicated by the dendrogram, the lung tissues that were given MSC treatment display closer
relatedness to the Sham lung samples, in terms of occludin expression. The figure was
generated by Partek V6.6.
67
Table 2. Targets of miR-193b-5p with differential expression in the mRNA microarray.
Occludin exhibits a fold change of 1.4990, adj p < 0.05. This shows occludin is potentially
strongly regulated by miR-193b-5p. Some of the top regulated mRNAs are also listed here.
68
miR-193b-5p Expression is Responsive to both CLP and MSC and Correlates with Occludin
Expression
By screening the miRNAs, we focused on the top 100 miRNAs that showed differential
expression following CLP and were also responsive to MSC treatment. Out of the top 100
miRNAs, we examined the ones that showed 100% sequence homology between mouse and
human. And miR-193b-5p was one such miRNA. From the in silico data analyses, expression of
miR-193b-5p and its putative target gene occludin are shown (Figures 8, 9), when miR-193b-5p
expression level was stimulated by CLP, occludin mRNA level was significantly down-regulated.
On the other hand, when miR-193b-5p expression was rescued towards one similar to that of
sham’s, occludin showed a significant attenuation in its expression.
69
Figure 8. Gene expression of miR-193b-5p by normalized raw intensity values across Sham,
CLP and MSC. As exhibited by the gene profiles of Sham (green), CLP (red), and MSC (blue),
miR-193b-5p expression was elevated by CLP, and suppressed in CLP + MSC samples. Results
were normalized using Lowess normalization. Determination of differential miRNA expression
was performed in R/bioconductor using mainly the LIMMA package. The values plotted are
means SEMs of N=4-5 independent experiments for each treatment group.
miR-193b-5p expressionin mouse lung tissue
Sham
CLP
MSC-0.10
-0.05
0.00
0.05
0.10
0.15
0.20
No
rmalized
raw
in
ten
sit
y
70
Figure 9. Gene expression of occludin by normalized raw intensity values across Sham, CLP
and MSC. Results were normalized using Lowess normalization. Determination of differential
miRNA expression was performed in R/bioconductor using mainly the LIMMA package. As
exhibited by the gene profiles of Sham (green), CLP (red), and MSC (blue), occludin expression
was significantly down-regulated by CLP (p<0.05), and alleviated in CLP + MSC samples (p<0.05).
The values plotted are means SEMs of N=5 independent experiments for each treatment
group.
Occludin mRNA expressionin mouse lung tissue
Sham
CLP
MSC
0
2
4
6
8 * *
No
rmalized
raw
in
ten
sit
y
71
To confirm this finding that we extracted from microarray analyses, it was necessary to quantify
the genes by qRT-PCR. In the internal validation experiment, using the same RNA samples in the
microarray experiment, it was clear that miR-193b-5p was significantly up-regulated by more
than 3 fold, and down-regulated to a level similar to sham following MSC treatment (Figure 10).
In line with the expression level from the microarray data, occludin mRNA expression in mouse
lung tissue was regulated in a fashion complementary to miR-193b-5p expression trend (Figure
11).
In an external separate set of experimental septic mice, we noted the same correlation
between miR-193b-5p and occludin 24h after the initial CLP surgeries (Figures 12, 13). However,
by 48h, the surge in miR-193b-5p, and the decrease in occludin had subsided to pre-surgery
levels (Figures 12, 13). To monitor protein level change, western blotting was used and we
detected a significant reduction in occludin in the CLP mice (Figure 14).
72
Figure 10. Quantitative RT-PCR analysis of miR-193b-5p. RNA samples from the same lung
tissues for the microarray were used. The expression of miR-193b-5p was corrected for
expression of the control gene, miR-191. It was noted that CLP led to significantly up-regulated
expression of miR-193b-5p (P<0.05), and MSC reversed the expression (P<0.05). The values
plotted are means SEMs of N=4-6 independent experiments for each treatment group.
miR-193b-5p expression in mouse lung tissue
Sham
CLP
MSC
0
1
2
3
4
5
**
Fo
ld
diffe
ren
ce
no
rmalized
to
miR
-191
73
Figure 11. Quantitative RT-PCR analysis of occludin. RNA samples from the same lung tissues
for the microarray were used. The expression of occludin was corrected for expression of the
control gene, GAPDH. It was noted that CLP led to significantly down-regulated expression of
occludin (P<0.05), and MSC rescued the expression (P<0.05). The values plotted are means
SEMs of N=4-6 independent experiments for each treatment group.
Occludin mRNA expressionin mouse lung tissue
Sham
CLP
MSC
0.0
0.5
1.0
1.5
**
Fo
ld d
iffe
ren
ce
no
rmalized
GA
PD
H
74
Figure 12. Quantitative RT-PCR analysis of miR-193b-5p. We aimed to confirm our in silico
findings using a different set of mice. RNA samples were extracted from lung tissues of this
external group of mice that underwent CLP surgeries. The expression of miR-193b-5p was
corrected for expression of the control gene, miR-191. It was noted that CLP (24h) led to
significantly up-regulated expression of miR-193b-5p (P<0.05), but CLP (48h) showed a subsided
expression (P<0.05). The values plotted are means SEMs of N=4-6 independent experiments
for each treatment group.
75
Figure 13. Quantitative RT-PCR analysis of occludin. RNA samples from an entirely different
set of mouse lung tissues were used. The expression of occludin was corrected for expression of
the control gene, GAPDH. It was noted that CLP (24h) led to significantly down-regulated
expression of occludin (P<0.05), but CLP (48h) showed a recovered expression (P<0.05). The
values plotted are means SEMs of N=4-6 independent experiments for each treatment group.
76
Figure 14. Down-regulation of occludin protein in mouse lung tissue. Lysates from tissue
samples were analyzed by SDS-PAGE using 10% polyacrylamide gels. The Western blot is a
representative image for 3 independent experiments. For densitometry, films were scanned
and analyzed using Image J, and the integrated intensity was normalized to the control protein
– GAPDH. Occludin expression was significantly reduced 24h after the initial CLP surgery.
OCLN
GAPDH
24h Sham 24h CLP
77
miR-193b-5p Expression is Up-Regulated by TNFa and Correlated to Loss of Occludin in
HPMECs
HPMEC cells were selected to be the in vitro model to further validate our findings from the in
silico analyses for the following reasons. First, lung endothelial injury is one of the key
characteristics of ARDS. Also, recent studies have shown that MSCs secrete paracrine factors
that modulate immune responses and alter the responses of the endothelium to injury. These
two pieces of knowledge make the choice of HPMEC as our in vitro model highly suitable to
imitate the insult and rescue that would be similarly observed when CLP and MSCs were
applied in mice. When the cells were treated with TNFa for 24h, there was a significant and
pronounced rise in miR-193b-5p expression (Figure 15). As predicted by previous experiments,
occludin mRNA expression was significantly down-regulated (Figure 16). This was also
confirmed using western blotting, and we noted a 63% reduction in occludin protein (Figure 17).
78
Figure 15. Quantitative RT-PCR analysis of miR-193b-5p. Sub-confluent proliferating HPMEC
cells were treated with TNFa at 10 ng/ml for T=24 hr, in serum-free culture medium, and then
harvested for quantitative RT-PCR assays (see Methods section). The expression of miR-193b-
5p was corrected for expression of the control gene, miR-191. TNFa treatment for 24h led to a
significantly overexpression of miR-193b-5p (P<0.05). The values plotted are means SEMs of
N=3-4 independent experiments for each treatment group.
79
Figure 16. Quantitative RT-PCR analysis of occludin. Sub-confluent proliferating HPMEC cells
were treated with TNFa at 10 ng/ml for T=24 hr, in serum-free culture medium, and then
harvested for quantitative RT-PCR assays (see Methods section). The expression of occludin was
corrected for expression of the control gene, GAPDH. TNFa treatment for 24h led to a
significantly suppressed expression of occludin (P<0.05). The values plotted are means SEMs
of N=3-4 independent experiments for each treatment group.
80
Figure 17. A) Western blot analysis of HPMEC cells treated with TNFa. Cells were treated with
TNFa (10ng/mL, for 24h). Lysates were analyzed by SDS-PAGE using 10% polyacrylamide gels. B)
For densitometry, films were scanned and analyzed using Image J, and the integrated intensity
was normalized to the control protein – GAPDH. Occludin expression was significantly reduced
24h after the initial TNFa administration.
OCLN
Control TNFa 24h
GAPDH
OCLN protein expression quantificationin HPMEC
contr
ol
TNF
0.0
0.5
1.0
1.5
*
Rela
tive E
xp
ressio
n
81
Overexpression and Knockdown of miR-193b-5p Correlates with Occludin Expression
Microarray analysis identified a putative regulatory relationship between miR-193b-5p and
occludin in the lung. Through confirmation by qRT-PCR in a greater number of animals, we
suspected that the regulatory relationship may in fact exist. To test this, gain- and loss-of-
function experiments were carried out. When the mimic of miR-193b-5p was transfected into
HPMEC cells, we noted a significant change in miR-193b-5p expression compared to control,
greater than 2000 fold. Cells treated with TNFa 24h following the initial transfection did not
have any significant statistical difference compared to cells transfected with the mimic alone
(Figure 18). Expectedly, for occludin, cells that were treated with TNFa, mimic only, or TNFa +
mimic, all showed significant decrease in occludin mRNA levels (p<0.001). However, the
resulting occludin expression levels were not statistically significant between TNFa, mimic only,
or TNFa + mimic, while TNFa + mimic treated cells had the lowest level of occludin (Figure 19).
Transfection of the inhibitor alone led to no statistically significant change in miR-193b-5p
expression level (Figure 21), and a slight increase in occludin mRNA expression (Figure 21).
When the inhibitor of miR-193b-5p was transfected into HPMEC cells, the effect of TNFa
stimulation on miR-193b-5p was much attenuated (p<0.01) (Figure 20). And this transfection
reduced the loss of occludin due to TNFa significantly (p<0.01) (Figure 21).
Using western blotting, we found that after transfection of miR-193b-5p inhibitor into HPMEC
cells, in the presence of TNFa, occludin protein expression was also rescued significantly
compared to a level that was almost entirely knocked out (Figure 22).
82
Figure 18. Quantitative RT-PCR analysis of miR-193b-5p. Sub-confluent proliferating HPMEC
cells were transfected with miR-193b-5p mimic, in serum-free culture medium, one group was
also treated with TNFa at 10 ng/ml for 24 hr. Cells were then harvested for quantitative RT-PCR
assays. The expression of miRNA was corrected to expression of the control gene, miR-191. The
mimic of miR-193b-5p was able to be detected following transfection, showing a significant
difference compared to the control group (P<0.001), and mimic transfection in combination
with TNFa stimulation led to significant overexpression of miR-193b-5p compared to the group
that only received TNFa(P<0.001). The values plotted are means SEMs of N=3-4 independent
experiments for each treatment group.
miR-193b-5p expression followingmiR-193b-5p mimic transfection in HPMEC
Contr
ol
TNF
Mim
ic o
nly
Mim
ic +
TNF
0
1000
2000
3000 ***
***F
old
diffe
ren
ce
no
rmalized
to
miR
-191
83
Figure 19. Quantitative RT-PCR analysis of occludin. Sub-confluent proliferating HPMEC cells
were transfected with miR-193b-5p mimic in serum-free culture medium, one group was also
treated with TNFa at 10 ng/ml for 24 hr. Cells were then harvested for quantitative RT-PCR
assays. The expression of miRNA was corrected to expression of the control gene, GAPDH. Cells
that were overexpressed with miR-193b-5p displayed significant down-regulation in occludin
expression compared to the control group (P<0.001), and mimic transfection in combination
with TNFa stimulation was also associated with significant reduction in occludin (P<0.001). The
values plotted are means SEMs of N=3-4 independent experiments for each treatment group.
Occludin mRNA expressionfollowing miR-193b-5p mimic
transfection in HPMEC
Contr
ol
TNFa
Mim
ic o
nly
Mim
ic +
TNFa
0.0
0.5
1.0
1.5***
***
***F
old
d
iffe
ren
ce
no
rmalized
to
GA
PD
H
84
Figure 20. Quantitative RT-PCR analysis of miR-193b-5p. Sub-confluent proliferating HPMEC
cells were transfected with miR-193b-5p inhibitor in serum-free culture medium, one group
was also treated with TNFa at 10 ng/ml for 24 hr. Cells were then harvested for quantitative RT-
PCR assays. The expression of miRNA was corrected for expression of the control gene, miR-
191. Cells transfected with the miRNA inhibitor did not show significant differences in miR-
193b-5p expression compared to the control group, however, the inhibitor was able to lower
miR-193b-5p expression in response to TNFa stimulation (P<0.01) that would otherwise be
observed when TNFa was given to the cells alone. The values plotted are means SEMs of N=3-
4 independent experiments for each treatment group.
miR-193b-5p expressionfollowing miR-193b-5p inhibitor
transfection in HPMEC
Contr
ol
TNFa
only
Inhib
itor O
nly
inhib
itor +
TNF
0
1
2
3**
Fo
ld
diffe
ren
ce
no
rmalized
to
miR
-191
85
Figure 21. Quantitative RT-PCR analysis of occludin. Sub-confluent proliferating HPMEC cells
were transfected with miR-193b-5p inhibitor in serum-free culture medium, one group was also
treated with TNFa at 10 ng/ml for 24 hr. Cells were then harvested for quantitative RT-PCR
assays. The expression of miRNA was corrected for expression of the control gene, GAPDH.
Cells transfected with the miRNA inhibitor did not show significant differences in occludin
expression compared to the control group, however, to a large degree, the inhibitor prevented
the decreased level of occludin expression in the response to TNFa stimulation (P<0.01). The
values plotted are means SEMs of N=3-4 independent experiments for each treatment group.
Occludin mRNA expressionfollowing miR-193b-5p inhibitor
transfection in HPMEC
Contr
ol
TNFa
only
Inhib
itor O
nly
inhib
itor +
TNFa
0.0
0.5
1.0
1.5*****
Fo
ld
diffe
ren
ce
no
rmalized
to
GA
PD
H
86
Figure 22. A) Western blot analysis of occludin expression HPMECs after miR-193b-5p
knockdown. Cells were transfected with either miR-193b-5p inhibitor or scrambled control,
allowed to grow for 24h, then treated with TNFa (10ng/mL, for 24h). Lysates were analyzed by
SDS-PAGE using 10% polyacrylamide gels. B) For densitometry, films were scanned and
analyzed using Image J, and the integrated intensity was normalized to the control protein –
GAPDH. Occludin expression was significantly reduced 24h after the initial TNFa administration,
but rescued in cells transfected with miR-193b-5p.
87
Effect of miR-193b-5p Inhibitor on HPMEC Cell Permeability
Gain-and loss-of-function experiments described previously have suggested a strong correlation
in expression between miR-193b-5p and its putative target occludin. To determine whether the
manipulation of miR-193b-5p levels led to any functional effects, we performed permeability
measurements across the monolayer of HPMEC cells.
The first experiment that we did was to examine whether there would be altered cellular
viability among various treatment conditions. Apoptosis would be a confounding factor in
evaluating permeability across a cell monolayer as increasing gaps would exist due to increased
cell loss. With the use of a spectrophotometer, we found no statistically significant differences
among cells that underwent different treatment conditions (Figure 23).
FITC-dextran was used first to measure monolayer permeability. As shown in Figure 24, at 24h
after the initial transfection of the miR-193b-5p inhibitor, there were not significant differences
between the groups. However, when TNFa was administered to the cells accordingly, non-
transfected cells displayed marked increases in cellular leakage, as indicated by the amount of
FITC-dextran measured. miR-193b-5p inhibitor transfection was able to reduce the leakage to
less than half compared to cells that did not receive a knock down in miR-193b-5p (p<0.05).
We also used transendothelial electrical resistance as a marker for paracelluar and transcellular
leakage. HPMEC cells that were given TNFa only and cells given TNFa plus the scambled control
showed the greatest loss in their monolayer electrical resistance at 24h, there was no statistical
88
significant difference between the two groups. When cells were transfected with the miR-193b-
5p inhibitor alone, there was a slight increase in electrical resistance at the 24h mark compared
to the control group. Most importantly, transfection of miR-193b-5p inhibitor followed by TNFa
stimulation displayed a significant difference compared to the scrambled control plus TNFa
group (p<0.05). Also, there was no statistically significant difference between the inhibitor plus
TNFa and control cells (Figure 25).
89
Figure 23. Effect of transfection on HPMEC cell viability. HPMEC cells were transfected with
respective mimic, inhibitor, or scrambled control, followed by TNFa treatment (10ng/mL for
24h) if indicated. The viability of the cells was assessed by MTT assay (see Methods). Results are
expressed as enrichment factor relative to DMSO-treated control. Each experiment was done at
least twice in triplicate and representative data from a single experiment are shown. Columns,
mean (n=3); there was no statistical significance different (P<0.05) between the indicated
groups by paired t-test.
Cell Viability
Contr
ol
TNF M
IM INH
Scr
amble
d
MIM
+ T
NF
INH +
TNF
Scr
amble
d + T
NF
0.0
0.5
1.0
1.5
OD
in
ten
sit
y
(no
rmalized
to
co
ntr
ol)
90
Figure 24. Effect of miR-193b-5p inhibitor on permeability. HPMEC cells were seeded in
transwells and allowed to grow to near confluency prior to transfection (2-3 days). Following
transfection, cells were given TNFa at 10ng/mL for 24h. FITC-dextran was given to all cells, and
measured in both control cells and transfected cells. The inhibitor of the miRNA was able to
maintain a similar level of FITC-dextran that leaked across the cell monolayer compared to the
control at 24h, and showed significant improvement compared to the negative control treated
with TNFa (P<0.05). The values plotted are means SEMs of N=3-4 independent experiments
for each treatment group.
Effects of miR-193b-5pinhibitor transfection
on HPMEC permeability
Contr
ol
Inhib
itor
Scr
amble
d Contr
ol
Contr
ol
Contr
ol + T
NF
Inhib
itor +
TNF
Scr
amble
d Contr
ol + T
NF
0
1
2
3
4
* *F
old
Diffe
ren
ce f
rom
Co
ntr
ol
(FIT
C-D
extr
an
)
91
Effects of miR-193b-5p inhibitoron HPMEC permeability
0 3 5 24
0.0
0.5
1.0
1.5Control
TNFa only
miR-193b-5p Inhibitor only
miR-193b-5p Inhibitor + TNFa
Negative Control + TNFa
Time (h)
*
miR-193b-5p mimic only
Tra
ns-e
nd
oth
elial
Ele
ctr
ical R
esis
tan
ce
(no
rmalized
to
co
ntr
ol)
Figure 25. Effect of miR-193b-5p inhibitor on permeability. HPMEC cells were seeded in
transwells and allowed to grow to near confluency prior to transfection (2-3 days). Following
transfection, cells were given TNFa at 10ng/mL for 24h. TEER was measured at 4 different time
points as indicated in the figure. The inhibitor of the miRNA was able to maintain a similar level
of TEER compared to the control at 24h, and showed significant improvement compared to the
scrambled control (P<0.05). The values plotted are means SEMs of N=3-4 independent
experiments for each treatment group. Also, to assess whether a tight monolayer had formed,
we repeated measured the TEER prior to any transfection or treatment, a confluent monolayer
displayed a resistance of 320-350 ohms cm2.
92
miR-193b-5p Knockdown Inhibits Loss in Occludin
The ability to limit the extent of pulmonary edema is key in treating ARDS. So far, we have
hypothesized down-regulated miR-193b-5p activity as a potential player in conferring beneficial
effects in the lung to reduce cellular leakage. In the 3’UTR region of the occludin gene, a binding
sequence for the 7-mer seed sequence in miR-193b-5p is identified there. When occludin-
3’UTR-luciferase reporter and miR-193b-5p mimic were co-transfected into HPMEC cells,
increased miR-193b-5p led to decreased level of chemo-luminescence, indicating a decreased
level of luciferase expression due to the occludin-3’UTR contained in the reporter (Figure 26).
On the other hand, when transient co-transfection of an occludin-3’UTR-luciferase reporter and
miR-193b-5p inhibitor was performed on HPMEC cells, there was no significant loss in chemo-
luminescence of luciferase, even in the presence of TNFa. More importantly, mutating the
binding site for the seed sequence in the 3’UTR of occludin in the plasmid prevented miR-193b-
5p from suppressing luciferase expression in HPMEC cells (Figure 27).
93
Figure 26. miR-193b-5p activity in HPMEC cells expressing occludin: HPMEC cells were
transiently transfected with luciferase construct plasmids that contained a wildtype 3’UTR or a
mutated 3’UTR of occludin, and the mimic of miR-193b-5p for 24hr. Cells were harvested and
luciferase activity was determined using the Luciferase Assay (Switchgear genomics) per the
manufacturer’s protocol. The two controls used in the experiment were, ACTB that stands for
beta actin was used as the housekeeping gene, and R01 was used as a random sequence
control, it contains non-conserved, non-genic, and non-repetitive human genomic
fragments. The treatment of miR-193b-5p mimic (24h) led to a significant reduction in
luciferase activity (P<0.05). This effect was not observed in cells expressing mutated plasmids.
Effect of miR-193b-5pmimic transfection on
luciferase luminescence
NC
MIM N
CM
IM NC
MIM N
CM
IM
0
5
10
15
20ACTB
R01
OCLN MUT
OCLN
*
Ch
em
o-l
um
inescen
ce (
RL
U)
94
Figure 27. miR-193b-5p activity in HPMEC cells expressing occludin: HPMEC cells were
transiently transfected with luciferase construct plasmids that contained a wildtype 3’UTR or a
mutated 3’UTR of occludin, and the inhibitor of miR-193b-5p for 48hr, followed by a 24h TNFa
treatment. Cells were harvested and luciferase activity was determined using the Luciferase
Assay (Switchgear genomics) per the manufacturer’s protocol. The two controls used in the
experiment were, ACTB that stands for beta actin was used as the housekeeping gene, and R01
was used as a random sequence control, it contains non-conserved, non-genic, and non-
repetitive human genomic fragments. The treatment of TNFa (24h) led to a significant
reduction in luciferase activity (P<0.05), and the inhibitor of miR-193b-5p was able to rescue it
(P<0.05). This effect was not observed in cells expressing mutated plasmids.
Effect of miR-193b-5pinhibitor transfection and TNFa stimulus
on luciferase luminescence
NCIN
H
TNFa
only
INH +
TNFa N
CIN
H
TNFa
only
INH +
TNFa N
CIN
H
TNFa
only
INH +
TNFa N
CIN
H
TNFa
only
INH +
TNFa
0
10
20
30ACTB
R01
OCLN MUT
OCLN* *
Ch
em
o-l
um
inescen
ce (
RL
U)
95
Precursor and Complementary Mature Forms of miR-193b-5p Are Affected by Transfection of
miR-193b-5p Mimic and Inhibitor
When the mimic of miR-193b-5p was transfected into HPMEC cells, both pre- miR-193b and
miR-193b-3p showed significantly increased levels of expression, while the transfection of miR-
193b-5p inhibitor did not lead to statistically significant changes in either one of the two
(Figures 28, 30).
Similar to the changes in miR-193b-5p expression level in response miR-193b-5p inhibitor
transfection, we noted substantially diminished expression in both pre- miR-193b and miR-
193b-3p. Moreover, miR-193b-5p inhibitor transfection in the presence of TNFa maintained a
level of miR-193b-3p expression nearly unchanged compared to the control cells in response to
TNFa stimuli (Figures 29, 31).
96
Figure 28. Quantitative RT-PCR analysis of pre-miR-193b. Sub-confluent proliferating HPMEC
cells were transfected with miR-193b-5p mimic or inhibitor in serum-free culture medium,
allowed to grow for 24h, and then harvested for quantitative RT-PCR assays (see Methods
section). The expression of miRNA was corrected for expression of the control gene, pre-miR-
191. Cells transfected with the miRNA inhibitor did not show significant differences in pre-miR-
193b expression compared to the control group, however, a significant up-regulation in pre-
miR-193b was observed in cells treated with TNFa and cells transfection with miR-193b-5p
mimic (P<0.05). The values plotted are means SEMs of N=3-4 independent experiments for
each treatment group.
pre-miR-193b expression followingmiR-193b-5p mimic/inhibitor
transfection in HPMEC
CTR
L
TNF
Mim
ic O
nly
Inhib
itor o
nly
0.0
0.5
1.0
1.5
2.0*
*F
old
diffe
ren
ce
no
rmalized
to
pre
-miR
-191
97
Figure 29. Quantitative RT-PCR analysis of pre-miR-193b. Sub-confluent proliferating HPMEC
cells were transfected with miR-193b-5p mimic or inhibitor in serum-free culture medium,
allowed to grow for 24h, given TNFa (10ng/mL) for another 24h, and then harvested for
quantitative RT-PCR assays (see Methods section). The expression of miRNA was corrected for
expression of the control gene, pre-miR-191. Unlike the mature form of miR-193b-5p, cells
transfected with the miRNA inhibitor did not return pre-miR-193b to a level similar to the
control, and showed no significant differences in pre-miR-193b expression compared to the
TNFa treated group. The values plotted are means SEMs of N=3-4 independent experiments
for each treatment group.
pre-miR-193b expressionfollowing miR-193b-5p inhibitor
transfection in HPMEC
Contr
ol
TNFa
Inhib
itor +
TNFa
Scr
amble
d + T
NFa
0
1
2
3
4
5
Fo
ld d
iffe
ren
ce
no
rmalized
to
pre
-miR
-191
98
Figure 30. Quantitative RT-PCR analysis of miR-193b-3p. Sub-confluent proliferating HPMEC
cells were transfected with miR-193b-5p mimic or inhibitor in serum-free culture medium,
allowed to grow for 24h, and then harvested for quantitative RT-PCR assays (see Methods
section). The expression of miRNA was corrected for expression of the control gene, miR-191.
Like the mature form of miR-193b-5p, cells transfected with the miRNA mimic displayed an
overexpression of miR-193b-3p (P<0.05), while the inhibitor of miR-193b-5p did not have a
significant decrease on the 3p expression. The values plotted are means SEMs of N=3-4
independent experiments for each treatment group.
miR-193b-3p expression followingmiR-193b-5p inhibitor
transfection in HPMEC
Contr
ol
TNF
Mim
ic O
nly
Inhib
itor O
nly
0.0
0.5
1.0
1.5
2.0
2.5 *
Fo
ld d
iffe
ren
ce
no
rmalized
to
miR
-191
99
Figure 31. Quantitative RT-PCR analysis of miR-193b-3p. Sub-confluent proliferating HPMEC
cells were transfected with miR-193b-5p scrambled control or inhibitor in serum-free culture
medium, allowed to grow for 24h, given TNFa (10ng/mL) for another 24h, and then harvested
for quantitative RT-PCR assays (see Methods section). The expression of miRNA was corrected
for expression of the control gene, miR-191. Like the mature form of miR-193b-5p, cells
transfected with the mir-193b-5p inhibitor showed a markedly reduced level of miR-193b-3p
under TNFa stimulation (P<0.05), while the scrambled control of miR-193b-5p did not have a
significant decrease on the 3p expression. The values plotted are means SEMs of N=3-4
independent experiments for each treatment group.
miR-193b-3p expressionfollowing miR-193b-5p inhibitor
transfection in HPMEC
Contr
ol
TNF
Inhib
itor +
TNF
Scr
amble
d + T
NF
0.0
0.5
1.0
1.5
2.0
2.5 *
Fo
ld d
iffe
ren
ce
no
rmalized
to
miR
-191
100
Wildtype MSC Conditioned Medium (MSC-CdM) Altered miR-193b-5p Expression
With the established regulatory relationship between miR-193b-5p and occludin, we then
aimed to determine whether the use of wildtype MSC conditioned medium would be able to
elicit the change brought about by the inhibitor. As shown in Figure 32, MSC-CdM did not
induce a change in miR-193b-5p expression compared to the control medium when there was
no external stimuli. However, in the presence of TNFa, the MSC-CdM group maintained a
similar level of miR-193b-5p expression, whereas the control group exhibited a significantly
increased amount of miR-193b-5p. On the other hand, occludin expression was significantly
downregulated in the control group following TNFa stimulation, but the loss in occludin was
significantly reduced when the cells were immerse in MSC-CdM (Figure 33). In animals that
received control medium after CLP, miR-193b-5p expression was significantly increased; while
MSC-CdM was able to attenuate much of the hike in miR-193b-5p expression following CLP
(Figure 34). Similar to the HPMECs, animals injected with MSC-CdM showed a much more mild
reduction in occludin following CLP insult (Figure 35).
101
Figure 32. Quantitative RT-PCR analysis of miR-193b-5p in conditioned medium treated
HPMECs. HPMECs were treated with control medium or wildtype MSC-CdM, with or without
TNFa (10ng/mL) stimulus for 24h, and then harvested for quantitative RT-PCR assays (see
Methods section). Cells stimulated with TNFa in the presence of the control medium showed a
marked increase in miR-193b-5p expression, while cells exposed to MSC-CdM showed a level of
miR-193b-5p similar to that of control’s. Expression of miRNA was corrected for expression of
the control gene, miR-191. The values plotted are means SEMs of N=3-4 independent
experiments for each treatment group.
miR-193b-5p expression in HPMECfollowing TNFa treatment in conditioned medium
Ctrl C
M
Ctrl M
SC C
dM
TNFa
CM
TNFa
MSC
CdM
0.0
0.5
1.0
1.5
2.0
2.5
*F
old
diffe
ren
ce
no
rmalized
to
miR
-191
102
Figure 33. Quantitative RT-PCR analysis of occludin in conditioned medium treated HPMECs.
HPMECs were treated with control medium or wildtype MSC-CdM, with or without TNFa
(10ng/mL) stimulus for 24h, and then harvested for quantitative RT-PCR assays (see Methods
section). Cells stimulated with TNFa in the presence of the control medium showed a marked
decrease in occludin expression, while cells exposed to MSC-CdM exhibited an attenuated level
of loss of occludin mRNA expression. Expression of miRNA was corrected for expression of the
control gene, GAPDH. The values plotted are means SEMs of N=3-4 independent experiments
for each treatment group.
OCLN mRNA expression in HPMECfollowing TNFa stimulusin conditioned medium
Ctrl C
M
Ctrl M
SC C
dM
TNFa
Ctrl C
M
TNF M
SC C
dM
0.0
0.5
1.0
1.5
* *F
old
diffe
ren
ce
no
rmalized
GA
PD
H
103
Figure 34. Quantitative RT-PCR analysis of miR-193b-5p in conditioned medium treated mice.
Wildtype C57 6J mice were administered control medium or wildtype MSC-CdM, following
sham or CLP surgery, for 24h, and then lung tissues were harvested for quantitative RT-PCR
assays (see Methods section). CLP-operated mice showed a marked increase in miR-193b-5p
expression when control medium was given. Sham-operated mice displayed no significant
difference in miR-193b-5p expression when injected with either control or conditioned medium.
CLP-operated mice that received MSC-CdM showed a noticeable decrease in miR-193b-5p
expression. Expression of miRNA was corrected for expression of the control gene, miR-191.
The values plotted are means SEMs of N=2 independent experiments for each treatment
group.
miR-193b-5p expression followingMSC-CdM administration
in wildtype mice
Sham
+ C
M
CLP
+ C
M
Sham
+ C
dM
CLP
+ C
dM
0
1
2
3
4
Fo
ld d
iffe
ren
ce
no
rmalized
to
miR
-191
104
Figure 35. Quantitative RT-PCR analysis of occludin in conditioned medium treated mice.
Wildtype C57 6J mice were administered control medium or wildtype MSC-CdM, following
sham or CLP surgery, for 24h, and then lung tissues were harvested for quantitative RT-PCR
assays (see Methods section). CLP-operated mice showed a marked decrease in occludin
expression when control medium was given. Sham-operated mice displayed no significant
difference in occludin expression when injected with either control or conditioned medium.
CLP-operated mice that received MSC-CdM showed an evidently increased level of occludin
compared to mice that were given control medium. Expression of miRNA was corrected for
expression of the control gene, GAPDH. The values plotted are means SEMs of N=2
independent experiments for each treatment group.
OCLN expression followingMSC-CdM administration
in wildtype mice
Sham
+ C
M
CLP
+ C
M
Sham
+ C
dM
CLP
+ C
dM
0.0
0.5
1.0
1.5F
old
diffe
ren
ce
no
rmalized
to
miR
-191
OCLN mRNA expression in HPMECfollowing TNFa stimulusin conditioned medium
Ctrl C
M
Ctrl M
SC C
dM
TNFa
Ctrl C
M
TNF M
SC C
dM
0.0
0.5
1.0
1.5
* *
Fo
ld d
iffe
ren
ce
no
rmalized
GA
PD
H
105
Discussion
In the past decade, miRNAs have been demonstrated to play key regulatory roles in
inflammatory and immune responses, including development and differentiation of B and T
cells, proliferation of monocytes and neutrophils, antibody production, and the release of
inflammatory mediators [174-176]. Also, miRNAs have been implicated in the pathogenesis of
various inflammatory lung diseases, making them potential candidates to contribute to the
pathogenesis of ARDS as well [81, 177, 178].
The main molecular effects that miRNAs exert are represented by gene expression variation.
These alterations in the amount of gene copies produced by miRNAs are usually moderate, but
the consequences potentially could affect numerous target genes [178, 179], which in turn may
influence various physiological pathways in normal and disease states. miRNA can be used to
inhibit mRNA production and function of disease-related genes. For instance, miRNAs that keep
the expression of inflammation pathway genes in control could be used to avoid overexpression
of cytokines, thereby to limit or suppress acute inflammatory response in patients. In addition,
miRNAs that directly participate in the pathogenesis of ALI/ARDS could be down-regulated by
antisense oligonucleotides, taking advantage of the sequence structure of these small RNA
molecules [180].
miRNAs and Sepsis
Also, miRNAs can be detected in a variety of sources, including tissue, blood, and body fluids.
Despite their short lengths, they are reasonably stable which increases their potential as clinical
106
biomarkers [178]. Recently, miRNAs were identified in serum and plasma as biomarkers for
diagnosing and monitoring several diseases including cancer, cardiovascular diseases, and
rheumatic diseases [181]. In septic patients, circulating miR-146a and miR-223 were shown to
be significantly reduced compared with healthy controls [182]. miR-146a and miR-223 have
been demonstrated by Wang et al to serve as potential biomarkers with high specificity and
sensitivity for sepsis. In contrast, Vasilescu and colleagues reported that the expression of miR-
150 correlated with the aggressiveness of sepsis; therefore, they suggested that miR-150 could
be a plasma prognostic marker in patients with sepsis [183]. Similarly, the expression of miR-
147 has been proven critical for endotoxin-induced tolerance [184]. miR-147 has been shown to
become induced in LPS-treated murine peritoneal macrophages, and in LPS-exposed murine
lungs via the stimulation of multiple TLRs. Such TLRs include TLR2, TLR3, and TLR4, all of which
enable inflammatory cells to recognize invading microbial pathogens. Also, activated TLR3, TLR4,
and TLR9 induced the expression of miR-148 and miR-152, which inhibited the production of
pro-inflammatory mediators, including IL-12, IL-6, and TNFa [185]. Moreover, as a predicted
suppressor for NF-kB, miR-9 was induced in human polymorphonuclear neutrophil and
monocytes following TLR4 activation [186]. Therefore, a negative feedback loop exists in which
TLR stimulation induces miRNAs such as miR-147, miR-9, miR-148, and miR-152 to prevent
excessive inflammatory responses, and such a mechanism may contribute to immune
homeostasis and immune regulation.
107
Microarray, miRNAs, and ARDS
With the use of high-throughput miRNA profiling, and quantitative real time polymerase chain
reaction, more miRNAs associated with immune response, inflammation pathways, disease
development, and disease severity could be evaluated as a novel tool in diagnosis monitoring,
or as therapy in inflammatory lung diseases . More specifically, paired miRNA and mRNA
microarrays to could be used identify the ones that exhibit differential expression in
contributing to immune responses in the context of experimental sepsis-induced ARDS.
Recently, in a microarray-based study, miR-181b has been demonstrated to regulate NFkB-
mediated endothelial cell activation and vascular inflammation in response to pro-
inflammatory stimuli and that rescue of miR-181b expression could provide a new target for
anti-inflammatory therapy and critical illness [187]. In addition, the entire list of putative target
genes of miR-181b contained 6 biological signalling pathways associated with NF-kB activation.
In another microarray-based study, through the screening, miR-127 was identified to be
involved in lung inflammation in ARDS. During lung injury, miR-127 was predicted to be down-
regulated. By targeting IgG FcγRI, miR-127 successfully reduces cytokine release by
macrophages [188]. The researchers were able to translate their in silico findings to their in
vitro and in vivo models, and validated the anti-inflammatory effects of miR-127 in their IgG
immune complex model.
miR-193b on Endothelial Integrity
Previously, functional studies on miR-193b have largely focused on its role as a potential tumor
suppressor [189-191]. miR-193b was demonstrated to be a negative regulator of uPA in breast
108
cancerous tissues [190]. In metastatic breast cancer tissues, miR-193b was downregulated,
which in turn upregulated uPA expression and contributed to the development of breast cancer.
Also, miR-193b has been proven to down-regulate myeloid cell leukemia sequence 1 (Mcl-1) in
melanoma cells [189]. From miR microarray profiling, miR-193b expression was significantly
lower in malignant melanoma than in normal tissues. In a survey of melanoma samples, the
level of Mcl-1 is inversely correlated with the level of miR-193b and that down-regulation of
miR-193b in vivo could be an early event in melanoma progression. Moreover, miR-193b has
previously been suggested to be epigenetically regulated in lymph node metastatic cell lines
originating from colon, skin and head and neck cancers [192].
miR-193b has not been investigated in depth in the field of sepsis or ARDS. Only very recently,
in a clinical publication, miR-193b was reported to be a biomarker for septic patients [193].
Wang and colleagues reported a significant overexpression in miR-193b in human serum in the
non-survivors, and generated a composite score based on the expression of six miRNAs that
had differential expression in patients with septic shock. The group discovered that the
predictive value of miR-193b for sepsis mortality was better than SOFA scores and APACHE II
scores, both of which are established composites that integrate numerous sepsis indicators.
They concluded that a combination of miR-15a, miR-16, miR-193b, miR-483-5p, SOFA scores,
APACHE II scores, would have a much better predictive value for sepsis mortality. In fact, all
four of these miRNAs suggested by the Wang group were identified in our miRNA microarray
analysis.
109
Endothelial cells perform multiple critical functions in vascular homeostasis, including
controlling leukocyte trafficking, regulating vessel wall permeability, and maintaining blood
fluidity. Pulmonary edema is a hallmark of ARDS. Herein, we provide evidence that miR-193b-
5p is dynamically regulated in response to sepsis and TNFa, and functions to suppress the
expression of a critical protein found in the tight junction, occludin. We have clearly
demonstrated the ability of alterations that inflammatory stimuli have on the function of tight
junctions. Here, we significantly expand this theme by demonstrating that manipulation of the
expression of miR-193b-5p in an acute inflammatory response could alleviate the loss of
permeability barrier function by rescuing occludin protein expression. In support, using
complementary gain-and loss-of-function approaches, we demonstrated that knockdown of
miR-193b-5p inhibits the degradation of occludin in HPMEC cells. Furthermore, overexpression
of miR-193b-5p through mimic transfection was able to significantly suppress the activity of
luciferase reporter plasmids that contained 3’UTR of occludin. Taken together, these studies
identify miR-193b-5p as a novel regulator of vascular permeability.
However, it is important to note that individual miRs regulate multiple mRNAs, and likewise,
each mRNA is often targeted by multiple miRs. Thus, it is likely that the collective actions of a
number of miRs acting on a host of target mRNAs ultimately contribute to the pro-
inflammatory effects of TNFa. Therefore, future studies need to map the combined actions of
miR networks regulated by CLP/TNFa stimulation and to determine how these events may have
an additive effect on endothelial integrity.
110
Interestingly, as predicted by the miRDB algorithm, we noted that the inhibition of some of the
other putative targets of miR-193-5p may in fact also have a role in anti-inflammatory
mechanisms and endothelial integrity. TNFRSF11B, and TNFRSF1B, both of which belong to the
super family of tumor necrosis factor signalling, which has key roles in the development and
regulation of the immune system [mirdb.org]. Members of the TNF receptor superfamily
contain cysteine-rich domains in their extracellular portion and bind to ligands of the
TNFSF. The TNFRSF has key functions in the development and regulation of the immune system
and knockout mice of different family members usually have some form of immunopathology
[194, 195]. Moreover, TNFSF ligands frequently activate NF-κB and loss of this NF-κB response is
likely to contribute to immune dysfunction. Furthermore, loss of NF-κB signalling, for example
in inhibitor of nuclear factor κ-B kinase (IKK) knockouts, mimics some of the effects seen in
TNFSF knockouts [196, 197]. Thus, the up-regulation of miR-193b-5p brought about by MSC
administration in our mouse model could potentially act to diminish the pro-inflammatory
effects caused by TNFa signalling.
Limitations on Microarray Studies
Some of the limitations of microarray studies based on genome-level analyses include the fact
that such methods only generate data extracted from a particular point in time, at a steady
state condition [49, 198]. This may in fact overlook the dynamic changes that take place
continuously in cells or tissues. In our validation experiment by qRT-PCR, as shown in Figure 12,
we identified an initial surge in the expression of miR-193b-5p at 24h. However, this
overexpression of the miRNA waned by 48h when we examined again. This subsiding trend
111
observed in miR-193b-5p expression could likely be attributed to a survivor bias, as miR-193b-
5p has been shown to play a role in detrimental to endothelial integrity. On the other hand, if
the true expression of miR-193b-5p did see a decrease from 24h to 48h, this would have been
an overlooked phenomenon.
One other limitation to microarray studies is that there is no direct information about protein
abundance, or post-translational modifications, such as phosphorylation or glycation. To
address this weakness, a proteomic screening process could be used in parallel to the
microarray screen [191]. When the transcriptome level data are examined in conjunction with
the proteome level data, first, the end result of the miRNA down regulation can be directly
determined, second, a combinatorial effect that takes into account knock-down of multiple
genes can better address the broader impact of miR-193b-5p inhibitor on other gene pathways
and networks that may assist in its function of preserving endothelial integrity.
Also, it is common to use RNA samples that are relatively homogenous and closely represent
the disease condition, and RNA samples that come from a single cell type are widely used in
microarray studies [69, 198]. Even though this approach provides a much more homogenous
source of RNA, important and biologically relevant gene expression profiles could potentially be
overlooked from this experimental setup. In our study, total RNA was extracted from whole
lung tissues, making the expression signatures were representative to treat the organ as one
entity.
112
Other Mechanisms by Which MSCs May Act on Occludin to Confer Protection
Another possible mechanism by which MSCs confer protective benefits to improve endothelial
integrity in sepsis or sepsis-induced ARDS is through the interaction angiopoietin-1 and occludin.
Increased permeability of the endothelium and epithelium, and decreased alveolar fluid
clearance, may both contribute to the formation of edema in the alveoli. At the site of gas
exchange, the alveolar epithelium normally forms a tighter barrier than the endothelium, and
its loss of integrity in ALI is of great significance [199, 200]. Commonly, to measure the degree
of lung endothelial and epithelial permeability, BAL albumin and protein are used as markers
[95]. Several groups of researchers have documented beneficial effects of MSC administration
on improving endothelial integrity.
One possible mechanism involves angiopoietin-1 (Ang-1). Mei and colleagues found that
intratracheal administration of LPS led to increased BAL albumin, total protein, and IgM, and
this was attenuated by intravenous MSC following the injury [95]. Previously, Ang-1 has been
shown to reduce permeability and promote endothelial cell survival [201, 202] Two groups
were able to determine that MSCs overexpressing Ang-1 led to significant reduction in BAL
protein, and albumin. It was proposed that Ang-1 was able to diminish the influx of
inflammatory cells and plasma protein leakage by acting on the vascular endothelium. Two
groups were able to determine that MSCs overexpressing Ang-1 led to significant reduction in
BAL protein, and albumin. It was proposed that Ang-1 was able to diminish the influx of
inflammatory cells and plasma protein leakage by acting on the vascular endothelium [203,
113
204]. These findings suggest that MSCs, acting in part through angiopoietin, improve the critical
barrier function of the alveolar epithelium in ALI.
A previous study using rat brain capillary cells discovered that pericyte-derived angiopoietin-1
multimeric complex induces occludin gene expression. The up-regulation observed in occludin
expression was significantly inhibited by angiopoietin-1-neutralizing antibody [205]. In addition,
immunoprecipitation and western blot analyses confirmed that multimeric angiopoietin-1
induced occludin mRNA in endothelial cells. Also, Yu and colleagues reported that angiopoietin-
1 ameliorates the expressions of ZO-1, occludin, VE-cadherin, and PKCα signaling after focal
cerebral ischemia/reperfusion in rats [206]. Ang-1 injection significantly decreased blood brain
barrier permeability, and increased the protein expressions of ZO-1, occludin, and VE-cadherin.
Similar findings on Ang-1 and occludin interaction have also emerged from retinal
microvascular endothelial cells [207]. Taken together, these studies have demonstrated that
MSCs may act on Ang-1 - occludin interaction to help decrease the permeability of endothelial
cells in response to TNFa or polymicrobial insult.
Dicer and miRNA Processing
While we were able to establish a direct regulatory relationship by miR-193b-5p on occludin
mRNA expression, which also led to a change in its protein levels in the tight junction, it was
quite interesting for us to find that miR-193b-3p and the precursor were both responsive to the
overexpression and knock down of the 5p mature form. As shown in Figure 28, when miR-193b-
5p was overexpressed through transfection, we noted a significant increase in the expression
114
level of the precursor. And when the inhibitor of miR193b-5p was used along with TNFa
stimulation, the increase in precursor copy number was dramatically reduced (Figure 29).
Likewise, similar trends were observed in miR-193b-3p expression, as miR-193b-5p was
manipulated through transfection (Figures 30, 31). The data suggest that there could in fact be
a negative feedback pathway in the miRNA processing machinery that regulates both the
precursor and mature forms of the miRNA, and the Dicer protein could have significant roles in
such process.
Given our findings in this current study, it would be interesting to test the hypothesis that other
cellular stresses or stimuli also lead to changes in the miRNA processing machinery. Indeed, it
has been reported that Dicer expression is down-regulated by multiple stresses, such as
hypoxia [208], reactive oxygen species, and many others [114]. Importantly, the biological
significance of Dicer’s down regulation has not been understood in these cases. Thus, the
functional changes associated with changes in dicer expression under stress need to be
addressed in future studies. As produced in our current study, measurements of precursor as
well as mature miRNA species may be a potentially useful indicator.
As established in the literature, biogenesis of miRNAs is a multistep process, and this process
critically depends on Dicer for proper functioning. A specific set of proteins governs the
formation and maturation of miRNAs from long, primary transcripts, to medium-length
precursor forms, to the final stage of ~22 basepair mature miRNAs. Sequentially, the long initial
miRNA transcript that contains a 7-methylguanosine cap and a poly-(A) tail, and more than one
115
stable step loop structures, encounters the RNAse III enzyme Drosha, in complex with
DeGiorgio Critical Region 8 together with several other cofactors binds and gets cleaved. The
step loop regions of the pri-miRNA then give rise to the precursor miRNA [209, 210]. At a length
of less than 100 basepairs, the precursor is home to the 5′ or 3′ arm of the mature forms of
mIRNAs. The two arms are highly conserved, and this sequence complementarity found in the
opposing arms is key for the recognition by exportin 5 for miRNAs to be transported to the
cytoplasm for their final functions [211]. The pre-miRNA stem-loop is processed into a dsRNA
duplex containing only the 5p and 3p miRNA sequences by another RNAse III enzyme, Dicer
[212]. Usually, the more stable strand of the two processed mature miRNAs will be loaded into
an argonaute (Ago) protein for target binding/degradation. Argonaute proteins mediate the
effector function of the miRNAs through inhibiting translation of the target mRNA or, in the
case of Ago2, by directly degrading the mRNA transcript [213].
Based on the essential role that Dicer plays in the miRNA processing, we investigated its
potential function in our experimental sepsis model. In fact, we have also found that
mesenchymal stem cell conditioned medium (MSC CdM) has been able to change Dicer and
miR-193b-5p expression in a complementary pattern for miRNA function. From our
unpublished data, we noted that MSC CdM administered to HPMEC cells that also received
TNFa stimulation had significant impacts on the tight junction protein occludin and the miRNA
processing machinery. For one thing, MSC CdM was able to alleviate the significant drop in
occludin expression following CLP. Also, a number of proteins involved in the processing
machinery, dicer, ago2, and exportin 5 all exhibited MSC CdM dependent expression -- MSC
116
CdM was able to suppress Dicer, Ago2, and exportin 5 expressions in response to TNFa, at both
the mRNA level, as well as the protein level. We hypothesized that Dicer could act as a critical
point in the entire miRNA regulatory mechanism.
It has been demonstrated that, at the catalytic core of the pre-to-mature miRNA processing
complex, Dicer can also serve as an important site for regulation of miRNA biogenesis, which is
in full support of our experimental findings. In fact, sequence association factors that bind to
the pre-miRNA step loop have been proven involved in the promotion of miRNA maturation
[214, 215]. Moreover, transactivating response RNA binding protein (TRBP) has been identified
as an important component of the Dicer complex [216]. From HEK293 cells, with the use of
gain- and loss-of-fucntion experiments, and microarray analyses, another group has
demonstrated that phosphorylatable TRBP is critical for the expression of a variety of miRNAs,
including miR-17, miR-20a and miR-92a, that were induced upon phosphorylation. TRBP
proteins are phosphorylated by extracellular signal-regulated kinase, ERK1/2. The ERK pathway
is a fundamental modulating component of various cellular pathways, thus, TRBP recruitment
to the Dicer complex may act as a regulatory switch activated by ERK signaling to promote
expression of miRNAs [217, 218].
As we know, ERK has been well established to have important roles in the process of
inflammation. It has been reported that ERK is involved in controlling endothelial NF-kB
activation and inflammatory responses. In human endothelial cells, vascular endothelial growth
factor (VEGF) induces NF-kB-dependent transcription of cell adhesion molecules and monocyte
117
adhesion. Consistently, inhibition of ERK significantly increased nuclear translocation of NF-kB
induced by VEGF, whereas overexpression of ERK resulted in the loss of these responses to
VEGF. Using two protein kinase C inhibitors has demonstrated that VEGF downregulates its
negative regulatory signal ERK through protein kinase C that lies downstream. Strikingly,
elevation of ERK in endothelial cells markedly inhibited NF-kB activation as well as monocyte
adhesion induced by IL-1b and TNFa. The data collectively suggest that ERK serves as an anti-
inflammatory signal that suppresses expression of NF-kB-dependent inflammatory genes in
endothelial cells.
Therefore, under an overall inflammatory response, miRNA levels could potentially be altered
by Dicer expression, which in turn may be regulated by an ERK-NF-kB interaction/feedback
[218]. The use of MSC-CdM may exert its effect on this interaction through its secreted
paracrine factors, altering ERK’s levels, producing a much less inflammatory state in the
endothelial cells, and a reduced amount of Dicer available for miRNA processing. Thus,
measuring the existence of ERK activity in vascular endothelial cells may be useful for predicting
the feasibility and potency of inflammatory reactions in the vasculature, and could act as an
indicator for miR-193b-5p expression/activity.
Moreover, another component of the Dicer complex, the protein kinase R activating protein
(PACT) could also have a critical role in the activation of pre-miRNA processing. Despite being a
non-essential component in the process of miRNA maturation, inhibition of PACT has been
demonstrated to decrease expression levels of a number of mature miRNAs [219]. Interestingly,
118
knocking down any components of the Dicer complex inhibits the levels of both pre-miRNAs
and mature miRNAs [220]. Again, this evidence reinforces our hypothesis of an existing
feedback loop by which mature miRNAs could act to modulate the levels of their pre-miRNAs.
In C. elegans, the feedback mechanism was found to affect for let-7 expression where the
argonaute protein can bind a conserved region in the 3’ end of pri-let-7 to promote its
processing [220].
However, similar to the Drosha microprocessor complex, the Dicer complex may still be
bypassed in miRNA biogenesis in many cases. From mice that were deficient in Dicer,
sequencing analyses identified significant enrichment of one specific mature miRNA: miR-451.
Pre-miR-451 has a unique, highly conserved structure that allows it to associate with Ago2,
which contains RNase activity. Loading of pre-miR-451 into Ago2 results in slicing of pre-miR-
451 and generates a 30 nt intermediate miR-451, which is then further trimmed to become a
mature miRNA [221].
Other Potential Microarray Approaches
As discussed in the Methods section, our microarrays were performed using bead-based
technologies. Despite being the most widely used approach in the field, in the near future, new
or adapted platforms that address newly recognized challenges in miRNA detection are
expected to be developed.
119
One such challenge in miRNA studies is the detection of miRNA from sources that yield only low
amounts of miRNA, for instance, plasma or serum; or where miRNA may be difficult to extract,
including formalin-fixed tissue blocks. Even though in our experiment, none of the two sources
of miRNAs was in fact a difficulty for us, the lung still represents a tremendously heterogeneous
tissue with many different cell types that are capable of producing a large variety of miRNAs, a
great number of which are only present at extremely low copy numbers. To address the
difficulty in RNA extraction, methods that eliminate the need of RNA extraction have received
increasing amount of attention. HTG Molecular Diagnostics has recently released the
quantitative nuclease protection technology system (http://www.htgmolecular.com/proof-for-ffpe-
samples-drug-discovery-news/). In this technique, RNA and DNA oligonucleotides complementary
to the miRNAs hybridize to their targets. S1 nuclease is added to the sample to remove non-
protected RNA and excess oligonucleotides. Once the RNA is removed, a specific DNA
oligonucleotide library is generated. The oligonucleotides are then bound to specific
oligonucleotide linkers on the array and are detected with a horseradish peroxidase probe that
produces measurable luminescence in the presence of substrate. Studies involving miRNAs
extracted from formalin-fixed tissues have successfully used this technique to detect miRNA
expression levels. Using this approach may potentially provide us with an even more precisely
documented global miRNA expression profile in the lung, in ARDS.
Future Outlook
To further validate our experimental findings, both in silico and in vitro, we aim to translate
these results into an in vivo model to establish biological significance of the modulation of miR-
120
193b-5p. The same kind of wildtype C576J mice would be used as the ones in the microarray.
LPS will be administered intratracheally, and the inhibitor of miR-193b-5p will be injected
through the tail vein 6h after LPS has been given. We will assess membrane permeability in the
lung by the extravasation of Evans blue dye. Histologically, the permeability and occludin
expression will also be examined. Moreover, we will use wet-to-dry ratio as a factor to help
determine the overall impact of miR-193b-5p inhibitor on the entire organ.
In addition, miR-193b-5p expression will be investigated in mice that are subjected to CLP
surgeries and given MSC conditioned medium as treatment. We will further analyze the impact
that MSC-CdM administration can have on the miRNA processing machinery, miR-193b-5p, 3p,
precursor, and tight junction proteins. This in vivo experiment may help corroborate our
established findings in the in silico and in vitro experiments, and elucidate the changes that
MSC-CdM may have on the global gene expression profile much higher upstream due to its
potential impact and regulation on the miRNA processing machinery since thousands of genes’
expression could be resultantly changed.
Currently, we have human lung tissues from patients that died from ARDS. We are very
interested in finding out the genetic profiles exhibited in these septic patients compared to
normal ICU controls. As demonstrated in a previous study in ICU septic patients, miR-193b has
been found to exhibit changes in the serum, the findings obtained here will serve as the first
critical piece of evidence to complement the protective beneficial effects that MSCs confer in
experimental models of sepsis [193].
121
Summary
Collectively, understanding how miRNAs regulate lung inflammation may represent an
attractive way to modulate ARDS-associated symptoms. In summary, we have identified miR-
193b-5p as a MSC-responsive miRNA that regulates the expression of a key tight junction
protein, occludin, involved in the endothelial integrity in response to inflammation in vitro and
in vivo. These findings support the possibility that miR-193b-5p may serve as an important
miRNA involved in inflammation by controlling essential aspects of endothelial cell homeostasis
in the lung under physiologic and/or pathologic conditions. These studies also revealed that
strategies that keep miR-193b-5p expression in check, eg.the inhibitor of miR-193b-5p, as
potential therapies for the treatment against ARDS. Our current study provides the initial
framework to identify and characterize the role of other potential miRNAs as candidate
therapeutic tools. Taken together, the combination of high-throughput microarrays and
bioinformatics analyses constitutes a valid approach for molecular target identification to limit
acute inflammatory disease states.
122
References 1. Thille, A.W., et al., Comparison of the Berlin definition for acute respiratory distress syndrome
with autopsy. Am J Respir Crit Care Med, 2013. 187(7): p. 761-7. 2. Hernu, R., et al., An attempt to validate the modification of the American-European consensus
definition of acute lung injury/acute respiratory distress syndrome by the Berlin definition in a university hospital. Intensive Care Med, 2013. 39(12): p. 2161-70.
3. Mutlu, G.M. and G.R. Budinger, Incidence and outcomes of acute lung injury. N Engl J Med, 2006. 354(4): p. 416-7; author reply 416-7.
4. Goss, C.H., et al., Incidence of acute lung injury in the United States. Crit Care Med, 2003. 31(6): p. 1607-11.
5. Rubenfeld, G.D., et al., Incidence and outcomes of acute lung injury. N Engl J Med, 2005. 353(16): p. 1685-93.
6. Treggiari, M.M., et al., Effect of acute lung injury and acute respiratory distress syndrome on outcome in critically ill trauma patients. Crit Care Med, 2004. 32(2): p. 327-31.
7. Angus, D.C., The acute respiratory distress syndrome: what's in a name? JAMA, 2012. 307(23): p. 2542-4.
8. Navarrete-Navarro, P., et al., Acute respiratory distress syndrome among trauma patients: trends in ICU mortality, risk factors, complications and resource utilization. Intensive Care Med, 2001. 27(7): p. 1133-40.
9. Pittet, J.F., et al., Biological markers of acute lung injury: prognostic and pathogenetic significance. Am J Respir Crit Care Med, 1997. 155(4): p. 1187-205.
10. Kallet, R.H., et al., Exacerbation of acute pulmonary edema during assisted mechanical ventilation using a low-tidal volume, lung-protective ventilator strategy. Chest, 1999. 116(6): p. 1826-32.
11. Modelska, K., et al., Acid-induced lung injury. Protective effect of anti-interleukin-8 pretreatment on alveolar epithelial barrier function in rabbits. Am J Respir Crit Care Med, 1999. 160(5 Pt 1): p. 1450-6.
12. Sznajder, J.I., Strategies to increase alveolar epithelial fluid removal in the injured lung. Am J Respir Crit Care Med, 1999. 160(5 Pt 1): p. 1441-2.
13. Ware, L.B. and M.A. Matthay, The acute respiratory distress syndrome. N Engl J Med, 2000. 342(18): p. 1334-49.
14. Geiser, T., Mechanisms of alveolar epithelial repair in acute lung injury--a translational approach. Swiss Med Wkly, 2003. 133(43-44): p. 586-90.
15. Luh, S.P. and C.H. Chiang, Acute lung injury/acute respiratory distress syndrome (ALI/ARDS): the mechanism, present strategies and future perspectives of therapies. J Zhejiang Univ Sci B, 2007. 8(1): p. 60-9.
16. Siflinger-Birnboim, A. and A. Johnson, Protein kinase C modulates pulmonary endothelial permeability: a paradigm for acute lung injury. Am J Physiol Lung Cell Mol Physiol, 2003. 284(3): p. L435-51.
17. Kotsovolis, G. and K. Kallaras, The role of endothelium and endogenous vasoactive substances in sepsis. Hippokratia, 2010. 14(2): p. 88-93.
18. Milici, A.J., et al., Transcytosis of albumin in capillary endothelium. J Cell Biol, 1987. 105(6 Pt 1): p. 2603-12.
19. Predescu, D. and G.E. Palade, Plasmalemmal vesicles represent the large pore system of continuous microvascular endothelium. Am J Physiol, 1993. 265(2 Pt 2): p. H725-33.
123
20. Minshall, R.D. and A.B. Malik, Transport across the endothelium: regulation of endothelial permeability. Handb Exp Pharmacol, 2006(176 Pt 1): p. 107-44.
21. Townsley, M.I., J.A. King, and D.F. Alvarez, Ca2+ channels and pulmonary endothelial permeability: insights from study of intact lung and chronic pulmonary hypertension. Microcirculation, 2006. 13(8): p. 725-39.
22. Cioffi, D.L., et al., TRPing on the lung endothelium: calcium channels that regulate barrier function. Antioxid Redox Signal, 2009. 11(4): p. 765-76.
23. Ahmmed, G.U. and A.B. Malik, Functional role of TRPC channels in the regulation of endothelial permeability. Pflugers Arch, 2005. 451(1): p. 131-42.
24. Riento, K. and A.J. Ridley, Rocks: multifunctional kinases in cell behaviour. Nat Rev Mol Cell Biol, 2003. 4(6): p. 446-56.
25. Yao, L., et al., The role of RhoA/Rho kinase pathway in endothelial dysfunction. J Cardiovasc Dis Res, 2010. 1(4): p. 165-70.
26. Anwar, K.N., et al., RhoA/Rho-associated kinase pathway selectively regulates thrombin-induced intercellular adhesion molecule-1 expression in endothelial cells via activation of I kappa B kinase beta and phosphorylation of RelA/p65. J Immunol, 2004. 173(11): p. 6965-72.
27. van Nieuw Amerongen, G.P., et al., Involvement of Rho kinase in endothelial barrier maintenance. Arterioscler Thromb Vasc Biol, 2007. 27(11): p. 2332-9.
28. Miyazaki, T., K. Honda, and H. Ohata, m-Calpain antagonizes RhoA overactivation and endothelial barrier dysfunction under disturbed shear conditions. Cardiovasc Res, 2010. 85(3): p. 530-41.
29. Yang, D., et al., Endothelium-dependent contractions to acetylcholine, ATP and the calcium ionophore A 23187 in aortas from spontaneously hypertensive and normotensive rats. Fundam Clin Pharmacol, 2004. 18(3): p. 321-6.
30. Furchgott, R.F. and P.M. Vanhoutte, Endothelium-derived relaxing and contracting factors. FASEB J, 1989. 3(9): p. 2007-18.
31. Somlyo, A.P. and A.V. Somlyo, Signal transduction by G-proteins, rho-kinase and protein phosphatase to smooth muscle and non-muscle myosin II. J Physiol, 2000. 522 Pt 2: p. 177-85.
32. Masumoto, A., et al., Possible involvement of Rho-kinase in the pathogenesis of hypertension in humans. Hypertension, 2001. 38(6): p. 1307-10.
33. Garcia, J.G., et al., Thrombin-induced increase in albumin permeability across the endothelium. J Cell Physiol, 1986. 128(1): p. 96-104.
34. Mehta, D. and A.B. Malik, Signaling mechanisms regulating endothelial permeability. Physiol Rev, 2006. 86(1): p. 279-367.
35. Shen, Q., et al., Myosin light chain kinase in microvascular endothelial barrier function. Cardiovasc Res, 2010. 87(2): p. 272-80.
36. Vandenbroucke, E., et al., Regulation of endothelial junctional permeability. Ann N Y Acad Sci, 2008. 1123: p. 134-45.
37. Natarajan, V., et al., Sphingosine-1-phosphate, FTY720, and sphingosine-1-phosphate receptors in the pathobiology of acute lung injury. Am J Respir Cell Mol Biol, 2013. 49(1): p. 6-17.
38. Yang, Y. and S. Uhlig, The role of sphingolipids in respiratory disease. Ther Adv Respir Dis, 2011. 5(5): p. 325-44.
39. Belvitch, P. and S.M. Dudek, Role of FAK in S1P-regulated endothelial permeability. Microvasc Res, 2012. 83(1): p. 22-30.
40. Maeda, Y., et al., Sphingosine 1-phosphate (S1P) lyase in thymic perivascular spaces promotes egress of mature thymocytes via up-regulation of S1P receptor 1. Int Immunol, 2013.
41. Zhao, Y., et al., Protection of LPS-induced murine acute lung injury by sphingosine-1-phosphate lyase suppression. Am J Respir Cell Mol Biol, 2011. 45(2): p. 426-35.
124
42. Jain, R. and A. DalNogare, Pharmacological therapy for acute respiratory distress syndrome. Mayo Clin Proc, 2006. 81(2): p. 205-12.
43. Calfee, C.S. and M.A. Matthay, Nonventilatory treatments for acute lung injury and ARDS. Chest, 2007. 131(3): p. 913-20.
44. Donahoe, M., Acute respiratory distress syndrome: A clinical review. Pulm Circ, 2011. 1(2): p. 192-211.
45. Leon, C.G., et al., Discovery and development of toll-like receptor 4 (TLR4) antagonists: a new paradigm for treating sepsis and other diseases. Pharm Res, 2008. 25(8): p. 1751-61.
46. Zhang, B., et al., Mesenchymal stem cell secretes immunologically active exosomes. Stem Cells Dev, 2013.
47. Gupta, N., et al., Intrapulmonary delivery of bone marrow-derived mesenchymal stem cells improves survival and attenuates endotoxin-induced acute lung injury in mice. J Immunol, 2007. 179(3): p. 1855-63.
48. Maron-Gutierrez, T., et al., Cell-based therapies for the acute respiratory distress syndrome. Curr Opin Crit Care, 2014. 20(1): p. 122-31.
49. Kim, K., S.O. Zakharkin, and D.B. Allison, Expectations, validity, and reality in gene expression profiling. J Clin Epidemiol, 2010. 63(9): p. 950-9.
50. Ozsolak, F. and P.M. Milos, RNA sequencing: advances, challenges and opportunities. Nat Rev Genet, 2011. 12(2): p. 87-98.
51. Malone, J.H. and B. Oliver, Microarrays, deep sequencing and the true measure of the transcriptome. BMC Biol, 2011. 9: p. 34.
52. Allison, D.B., et al., Microarray data analysis: from disarray to consolidation and consensus. Nat Rev Genet, 2006. 7(1): p. 55-65.
53. Yang, M.C., et al., Microarray experimental design: power and sample size considerations. Physiol Genomics, 2003. 16(1): p. 24-8.
54. Makretsov, N.A., et al., Hierarchical clustering analysis of tissue microarray immunostaining data identifies prognostically significant groups of breast carcinoma. Clin Cancer Res, 2004. 10(18 Pt 1): p. 6143-51.
55. Guo, Y., et al., Towards a holistic, yet gene-centered analysis of gene expression profiles: a case study of human lung cancers. J Biomed Biotechnol, 2006. 2006(5): p. 69141.
56. Patel, A., et al., Validation of a targeted DNA microarray for the clinical evaluation of recurrent abnormalities in chronic lymphocytic leukemia. Am J Hematol, 2008. 83(7): p. 540-6.
57. Calvano, S.E., et al., A network-based analysis of systemic inflammation in humans. Nature, 2005. 437(7061): p. 1032-7.
58. Feezor, R.J., et al., Functional genomics and gene expression profiling in sepsis: beyond class prediction. Clin Infect Dis, 2005. 41 Suppl 7: p. S427-35.
59. Prabhakar, U., et al., Correlation of protein and gene expression profiles of inflammatory proteins after endotoxin challenge in human subjects. DNA Cell Biol, 2005. 24(7): p. 410-31.
60. Talwar, S., et al., Gene expression profiles of peripheral blood leukocytes after endotoxin challenge in humans. Physiol Genomics, 2006. 25(2): p. 203-15.
61. Tang, B.M., et al., The use of gene-expression profiling to identify candidate genes in human sepsis. Am J Respir Crit Care Med, 2007. 176(7): p. 676-84.
62. Johnson, S.B., et al., Gene expression profiles differentiate between sterile SIRS and early sepsis. Ann Surg, 2007. 245(4): p. 611-21.
63. Lowry, S.F., Human endotoxemia: a model for mechanistic insight and therapeutic targeting. Shock, 2005. 24 Suppl 1: p. 94-100.
64. Cvijanovich, N., et al., Validating the genomic signature of pediatric septic shock. Physiol Genomics, 2008. 34(1): p. 127-34.
125
65. Wong, H.R., et al., Genomic expression profiling across the pediatric systemic inflammatory response syndrome, sepsis, and septic shock spectrum. Crit Care Med, 2009. 37(5): p. 1558-66.
66. Tang, B.M., et al., Gene-expression profiling of peripheral blood mononuclear cells in sepsis. Crit Care Med, 2009. 37(3): p. 882-8.
67. Hotchkiss, R.S. and I.E. Karl, The pathophysiology and treatment of sepsis. N Engl J Med, 2003. 348(2): p. 138-50.
68. Kasten, K.R., et al., Interleukin-7 (IL-7) treatment accelerates neutrophil recruitment through gamma delta T-cell IL-17 production in a murine model of sepsis. Infect Immun, 2010. 78(11): p. 4714-22.
69. Tang, B.M., et al., Gene-expression profiling of gram-positive and gram-negative sepsis in critically ill patients. Crit Care Med, 2008. 36(4): p. 1125-8.
70. Payen, D., et al., Gene profiling in human blood leucocytes during recovery from septic shock. Intensive Care Med, 2008. 34(8): p. 1371-6.
71. Prucha, M., et al., Expression profiling: toward an application in sepsis diagnostics. Shock, 2004. 22(1): p. 29-33.
72. Pachot, A., et al., Decreased expression of the fractalkine receptor CX3CR1 on circulating monocytes as new feature of sepsis-induced immunosuppression. J Immunol, 2008. 180(9): p. 6421-9.
73. Ramilo, O., et al., Gene expression patterns in blood leukocytes discriminate patients with acute infections. Blood, 2007. 109(5): p. 2066-77.
74. Knoell, D.L., et al., Zinc deficiency increases organ damage and mortality in a murine model of polymicrobial sepsis. Crit Care Med, 2009. 37(4): p. 1380-8.
75. Wheeler, D.S., H.R. Wong, and B. Zingarelli, Pediatric Sepsis - Part I: "Children are not small adults!". Open Inflamm J, 2011. 4: p. 4-15.
76. Wong, H.R., et al., Genome-level expression profiles in pediatric septic shock indicate a role for altered zinc homeostasis in poor outcome. Physiol Genomics, 2007. 30(2): p. 146-55.
77. Wong, H.R., et al., Leukocyte subset-derived genomewide expression profiles in pediatric septic shock. Pediatr Crit Care Med, 2010. 11(3): p. 349-55.
78. Lauhio, A., et al., Serum MMP-8, -9 and TIMP-1 in sepsis: high serum levels of MMP-8 and TIMP-1 are associated with fatal outcome in a multicentre, prospective cohort study. Hypothetical impact of tetracyclines. Pharmacol Res, 2011. 64(6): p. 590-4.
79. Hu, P., et al., Microarray meta-analysis identifies acute lung injury biomarkers in donor lungs that predict development of primary graft failure in recipients. PLoS One, 2012. 7(10): p. e45506.
80. dos Santos, C.C., et al., Network analysis of transcriptional responses induced by mesenchymal stem cell treatment of experimental sepsis. Am J Pathol, 2012. 181(5): p. 1681-92.
81. Bartel, D.P., MicroRNAs: genomics, biogenesis, mechanism, and function. Cell, 2004. 116(2): p. 281-97.
82. Bi, Y., et al., Identification of tendon stem/progenitor cells and the role of the extracellular matrix in their niche. Nat Med, 2007. 13(10): p. 1219-27.
83. Bieback, K. and H. Kluter, Mesenchymal stromal cells from umbilical cord blood. Curr Stem Cell Res Ther, 2007. 2(4): p. 310-23.
84. Crisan, M., et al., A perivascular origin for mesenchymal stem cells in multiple human organs. Cell Stem Cell, 2008. 3(3): p. 301-13.
85. Dominici, M., et al., Minimal criteria for defining multipotent mesenchymal stromal cells. The International Society for Cellular Therapy position statement. Cytotherapy, 2006. 8(4): p. 315-7.
86. Gotts, J.E. and M.A. Matthay, Mesenchymal stem cells and acute lung injury. Crit Care Clin, 2011. 27(3): p. 719-33.
126
87. Ries, C., et al., MMP-2, MT1-MMP, and TIMP-2 are essential for the invasive capacity of human mesenchymal stem cells: differential regulation by inflammatory cytokines. Blood, 2007. 109(9): p. 4055-63.
88. Kode, J.A., et al., Mesenchymal stem cells: immunobiology and role in immunomodulation and tissue regeneration. Cytotherapy, 2009. 11(4): p. 377-91.
89. Ghannam, S., et al., Immunosuppression by mesenchymal stem cells: mechanisms and clinical applications. Stem Cell Res Ther, 2010. 1(1): p. 2.
90. Le Blanc, K., et al., Treatment of severe acute graft-versus-host disease with third party haploidentical mesenchymal stem cells. Lancet, 2004. 363(9419): p. 1439-41.
91. Prota, L.F., et al., Bone marrow mononuclear cell therapy led to alveolar-capillary membrane repair, improving lung mechanics in endotoxin-induced acute lung injury. Cell Transplant, 2010. 19(8): p. 965-71.
92. Lee, J.W., et al., Allogeneic human mesenchymal stem cells for treatment of E. coli endotoxin-induced acute lung injury in the ex vivo perfused human lung. Proc Natl Acad Sci U S A, 2009. 106(38): p. 16357-62.
93. Araujo, I.M., et al., Bone marrow-derived mononuclear cell therapy in experimental pulmonary and extrapulmonary acute lung injury. Crit Care Med, 2010. 38(8): p. 1733-41.
94. Kinnaird, T., et al., Marrow-derived stromal cells express genes encoding a broad spectrum of arteriogenic cytokines and promote in vitro and in vivo arteriogenesis through paracrine mechanisms. Circ Res, 2004. 94(5): p. 678-85.
95. Mei, S.H., et al., Prevention of LPS-induced acute lung injury in mice by mesenchymal stem cells overexpressing angiopoietin 1. PLoS Med, 2007. 4(9): p. e269.
96. Mei, S.H., et al., Mesenchymal stem cells reduce inflammation while enhancing bacterial clearance and improving survival in sepsis. Am J Respir Crit Care Med, 2010. 182(8): p. 1047-57.
97. Lee, R.C., R.L. Feinbaum, and V. Ambros, The C. elegans heterochronic gene lin-4 encodes small RNAs with antisense complementarity to lin-14. Cell, 1993. 75(5): p. 843-54.
98. Lee, R.C. and V. Ambros, An extensive class of small RNAs in Caenorhabditis elegans. Science, 2001. 294(5543): p. 862-4.
99. Fabbri, M., et al., MicroRNA-29 family reverts aberrant methylation in lung cancer by targeting DNA methyltransferases 3A and 3B. Proc Natl Acad Sci U S A, 2007. 104(40): p. 15805-10.
100. Vasudevan, S., Y. Tong, and J.A. Steitz, Switching from repression to activation: microRNAs can up-regulate translation. Science, 2007. 318(5858): p. 1931-4.
101. Grosshans, H. and W. Filipowicz, Molecular biology: the expanding world of small RNAs. Nature, 2008. 451(7177): p. 414-6.
102. Kim, V.N., MicroRNA biogenesis: coordinated cropping and dicing. Nat Rev Mol Cell Biol, 2005. 6(5): p. 376-85.
103. Han, J., et al., The Drosha-DGCR8 complex in primary microRNA processing. Genes Dev, 2004. 18(24): p. 3016-27.
104. Lee, Y., et al., MicroRNA genes are transcribed by RNA polymerase II. EMBO J, 2004. 23(20): p. 4051-60.
105. Sonkoly, E. and A. Pivarcsi, Advances in microRNAs: implications for immunity and inflammatory diseases. J Cell Mol Med, 2009. 13(1): p. 24-38.
106. Schwarz, D.S., et al., Asymmetry in the assembly of the RNAi enzyme complex. Cell, 2003. 115(2): p. 199-208.
107. Meister, G., et al., Human Argonaute2 mediates RNA cleavage targeted by miRNAs and siRNAs. Mol Cell, 2004. 15(2): p. 185-97.
127
108. Humphreys, D.T., et al., MicroRNAs control translation initiation by inhibiting eukaryotic initiation factor 4E/cap and poly(A) tail function. Proc Natl Acad Sci U S A, 2005. 102(47): p. 16961-6.
109. Mathonnet, G., et al., MicroRNA inhibition of translation initiation in vitro by targeting the cap-binding complex eIF4F. Science, 2007. 317(5845): p. 1764-7.
110. Chendrimada, T.P., et al., MicroRNA silencing through RISC recruitment of eIF6. Nature, 2007. 447(7146): p. 823-8.
111. Iwasaki, S., T. Kawamata, and Y. Tomari, Drosophila argonaute1 and argonaute2 employ distinct mechanisms for translational repression. Mol Cell, 2009. 34(1): p. 58-67.
112. Chen, X., et al., Secreted microRNAs: a new form of intercellular communication. Trends Cell Biol, 2012. 22(3): p. 125-32.
113. O'Connell, R.M., et al., Physiological and pathological roles for microRNAs in the immune system. Nat Rev Immunol, 2010. 10(2): p. 111-22.
114. Wiesen, J.L. and T.B. Tomasi, Dicer is regulated by cellular stresses and interferons. Mol Immunol, 2009. 46(6): p. 1222-8.
115. Maas, S. and A. Rich, Changing genetic information through RNA editing. Bioessays, 2000. 22(9): p. 790-802.
116. Luciano, D.J., et al., RNA editing of a miRNA precursor. RNA, 2004. 10(8): p. 1174-7. 117. Blow, M.J., et al., RNA editing of human microRNAs. Genome Biol, 2006. 7(4): p. R27. 118. Lim, L.P., et al., Microarray analysis shows that some microRNAs downregulate large numbers of
target mRNAs. Nature, 2005. 433(7027): p. 769-73. 119. Lewis, B.P., et al., Prediction of mammalian microRNA targets. Cell, 2003. 115(7): p. 787-98. 120. Kiriakidou, M., et al., A combined computational-experimental approach predicts human
microRNA targets. Genes Dev, 2004. 18(10): p. 1165-78. 121. Min, H. and S. Yoon, Got target? Computational methods for microRNA target prediction and
their extension. Exp Mol Med, 2010. 42(4): p. 233-44. 122. Lewis, B.P., C.B. Burge, and D.P. Bartel, Conserved seed pairing, often flanked by adenosines,
indicates that thousands of human genes are microRNA targets. Cell, 2005. 120(1): p. 15-20. 123. Brennecke, J., et al., Principles of microRNA-target recognition. PLoS Biol, 2005. 3(3): p. e85. 124. Kuhn, D.E., et al., Experimental validation of miRNA targets. Methods, 2008. 44(1): p. 47-54. 125. Roubelakis, M.G., et al., Human microRNA target analysis and gene ontology clustering by
GOmir, a novel stand-alone application. BMC Bioinformatics, 2009. 10 Suppl 6: p. S20. 126. Beitzinger, M., et al., Identification of human microRNA targets from isolated argonaute protein
complexes. RNA Biol, 2007. 4(2): p. 76-84. 127. Williams, A.E., et al., microRNA expression in the aging mouse lung. BMC Genomics, 2007. 8: p.
172. 128. Lu, Y., et al., Transgenic over-expression of the microRNA miR-17-92 cluster promotes
proliferation and inhibits differentiation of lung epithelial progenitor cells. Dev Biol, 2007. 310(2): p. 442-53.
129. Ventura, A., et al., Targeted deletion reveals essential and overlapping functions of the miR-17 through 92 family of miRNA clusters. Cell, 2008. 132(5): p. 875-86.
130. Moschos, S.A., et al., Expression profiling in vivo demonstrates rapid changes in lung microRNA levels following lipopolysaccharide-induced inflammation but not in the anti-inflammatory action of glucocorticoids. BMC Genomics, 2007. 8: p. 240.
131. Chen, C., et al., Smad1 expression and function during mouse embryonic lung branching morphogenesis. Am J Physiol Lung Cell Mol Physiol, 2005. 288(6): p. L1033-9.
132. Rodriguez, A., et al., Requirement of bic/microRNA-155 for normal immune function. Science, 2007. 316(5824): p. 608-11.
128
133. Banerjee, A., et al., Micro-RNA-155 inhibits IFN-gamma signaling in CD4+ T cells. Eur J Immunol, 2010. 40(1): p. 225-31.
134. Arvey, A., et al., Target mRNA abundance dilutes microRNA and siRNA activity. Mol Syst Biol, 2010. 6: p. 363.
135. Spizzo, R., et al., RNA inhibition, microRNAs, and new therapeutic agents for cancer treatment. Clin Lymphoma Myeloma, 2009. 9 Suppl 3: p. S313-8.
136. Balkovetz, D.F., Claudins at the gate: determinants of renal epithelial tight junction paracellular permeability. Am J Physiol Renal Physiol, 2006. 290(3): p. F572-9.
137. Anderson, J.M. and C.M. Van Itallie, Tight junctions and the molecular basis for regulation of paracellular permeability. Am J Physiol, 1995. 269(4 Pt 1): p. G467-75.
138. Forster, C., Tight junctions and the modulation of barrier function in disease. Histochem Cell Biol, 2008. 130(1): p. 55-70.
139. Harhaj, N.S. and D.A. Antonetti, Regulation of tight junctions and loss of barrier function in pathophysiology. Int J Biochem Cell Biol, 2004. 36(7): p. 1206-37.
140. Gardner, T.W., et al., Astrocytes increase barrier properties and ZO-1 expression in retinal vascular endothelial cells. Invest Ophthalmol Vis Sci, 1997. 38(11): p. 2423-7.
141. Bolton, S.J., D.C. Anthony, and V.H. Perry, Loss of the tight junction proteins occludin and zonula occludens-1 from cerebral vascular endothelium during neutrophil-induced blood-brain barrier breakdown in vivo. Neuroscience, 1998. 86(4): p. 1245-57.
142. Harhaj, N.S., A.J. Barber, and D.A. Antonetti, Platelet-derived growth factor mediates tight junction redistribution and increases permeability in MDCK cells. J Cell Physiol, 2002. 193(3): p. 349-64.
143. Antonetti, D.A., et al., Vascular permeability in experimental diabetes is associated with reduced endothelial occludin content: vascular endothelial growth factor decreases occludin in retinal endothelial cells. Penn State Retina Research Group. Diabetes, 1998. 47(12): p. 1953-9.
144. Furuse, M., et al., Occludin: a novel integral membrane protein localizing at tight junctions. J Cell Biol, 1993. 123(6 Pt 2): p. 1777-88.
145. Sakakibara, A., et al., Possible involvement of phosphorylation of occludin in tight junction formation. J Cell Biol, 1997. 137(6): p. 1393-401.
146. Forster, C., et al., Occludin as direct target for glucocorticoid-induced improvement of blood-brain barrier properties in a murine in vitro system. J Physiol, 2005. 565(Pt 2): p. 475-86.
147. Bazzoni, G., et al., Interaction of junctional adhesion molecule with the tight junction components ZO-1, cingulin, and occludin. J Biol Chem, 2000. 275(27): p. 20520-6.
148. Fanning, A.S., et al., The tight junction protein ZO-1 establishes a link between the transmembrane protein occludin and the actin cytoskeleton. J Biol Chem, 1998. 273(45): p. 29745-53.
149. Itoh, M., K. Morita, and S. Tsukita, Characterization of ZO-2 as a MAGUK family member associated with tight as well as adherens junctions with a binding affinity to occludin and alpha catenin. J Biol Chem, 1999. 274(9): p. 5981-6.
150. Kojima, T., et al., Induction of tight junctions in human connexin 32 (hCx32)-transfected mouse hepatocytes: connexin 32 interacts with occludin. Biochem Biophys Res Commun, 1999. 266(1): p. 222-9.
151. Nusrat, A., et al., The coiled-coil domain of occludin can act to organize structural and functional elements of the epithelial tight junction. J Biol Chem, 2000. 275(38): p. 29816-22.
152. Chen, Y.H., et al., Nonreceptor tyrosine kinase c-Yes interacts with occludin during tight junction formation in canine kidney epithelial cells. Mol Biol Cell, 2002. 13(4): p. 1227-37.
153. Lapierre, L.A., et al., VAP-33 localizes to both an intracellular vesicle population and with occludin at the tight junction. J Cell Sci, 1999. 112 ( Pt 21): p. 3723-32.
129
154. Balda, M.S., J.M. Anderson, and K. Matter, The SH3 domain of the tight junction protein ZO-1 binds to a serine protein kinase that phosphorylates a region C-terminal to this domain. FEBS Lett, 1996. 399(3): p. 326-32.
155. Balda, M.S., et al., Multiple domains of occludin are involved in the regulation of paracellular permeability. J Cell Biochem, 2000. 78(1): p. 85-96.
156. Wong, V. and B.M. Gumbiner, A synthetic peptide corresponding to the extracellular domain of occludin perturbs the tight junction permeability barrier. J Cell Biol, 1997. 136(2): p. 399-409.
157. Lacaz-Vieira, F. and M.M. Jaeger, Protein kinase inhibitors and the dynamics of tight junction opening and closing in A6 cell monolayers. J Membr Biol, 2001. 184(2): p. 185-96.
158. Furuse, M., et al., Direct association of occludin with ZO-1 and its possible involvement in the localization of occludin at tight junctions. J Cell Biol, 1994. 127(6 Pt 1): p. 1617-26.
159. Mitic, L.L., et al., Connexin-occludin chimeras containing the ZO-binding domain of occludin localize at MDCK tight junctions and NRK cell contacts. J Cell Biol, 1999. 146(3): p. 683-93.
160. Matter, K. and M.S. Balda, Biogenesis of tight junctions: the C-terminal domain of occludin mediates basolateral targeting. J Cell Sci, 1998. 111 ( Pt 4): p. 511-9.
161. Furuse, M., et al., Overexpression of occludin, a tight junction-associated integral membrane protein, induces the formation of intracellular multilamellar bodies bearing tight junction-like structures. J Cell Sci, 1996. 109 ( Pt 2): p. 429-35.
162. Van Itallie, C.M. and J.M. Anderson, Occludin confers adhesiveness when expressed in fibroblasts. J Cell Sci, 1997. 110 ( Pt 9): p. 1113-21.
163. Baker, O.J., et al., Proinflammatory cytokines tumor necrosis factor-alpha and interferon-gamma alter tight junction structure and function in the rat parotid gland Par-C10 cell line. Am J Physiol Cell Physiol, 2008. 295(5): p. C1191-201.
164. Coyne, C.B., et al., Regulation of airway tight junctions by proinflammatory cytokines. Mol Biol Cell, 2002. 13(9): p. 3218-34.
165. Mankertz, J., et al., Expression from the human occludin promoter is affected by tumor necrosis factor alpha and interferon gamma. J Cell Sci, 2000. 113 ( Pt 11): p. 2085-90.
166. Kucharzik, T., et al., Neutrophil transmigration in inflammatory bowel disease is associated with differential expression of epithelial intercellular junction proteins. Am J Pathol, 2001. 159(6): p. 2001-9.
167. Murakami, T., E.A. Felinski, and D.A. Antonetti, Occludin phosphorylation and ubiquitination regulate tight junction trafficking and vascular endothelial growth factor-induced permeability. J Biol Chem, 2009. 284(31): p. 21036-46.
168. Antonetti, D.A., et al., Vascular endothelial growth factor induces rapid phosphorylation of tight junction proteins occludin and zonula occluden 1. A potential mechanism for vascular permeability in diabetic retinopathy and tumors. J Biol Chem, 1999. 274(33): p. 23463-7.
169. Ritchie, M.E., et al., A comparison of background correction methods for two-colour microarrays. Bioinformatics, 2007. 23(20): p. 2700-7.
170. Smyth, G.K. and T. Speed, Normalization of cDNA microarray data. Methods, 2003. 31(4): p. 265-73.
171. Klipper-Aurbach, Y., et al., Mathematical formulae for the prediction of the residual beta cell function during the first two years of disease in children and adolescents with insulin-dependent diabetes mellitus. Med Hypotheses, 1995. 45(5): p. 486-90.
172. Chen, C., et al., Real-time quantification of microRNAs by stem-loop RT-PCR. Nucleic Acids Res, 2005. 33(20): p. e179.
173. Peltier, H.J. and G.J. Latham, Normalization of microRNA expression levels in quantitative RT-PCR assays: identification of suitable reference RNA targets in normal and cancerous human solid tissues. RNA, 2008. 14(5): p. 844-52.
130
174. Sheedy, F.J., et al., Negative regulation of TLR4 via targeting of the proinflammatory tumor suppressor PDCD4 by the microRNA miR-21. Nat Immunol, 2010. 11(2): p. 141-7.
175. Taganov, K.D., et al., NF-kappaB-dependent induction of microRNA miR-146, an inhibitor targeted to signaling proteins of innate immune responses. Proc Natl Acad Sci U S A, 2006. 103(33): p. 12481-6.
176. Ma, F., et al., MicroRNA-466l upregulates IL-10 expression in TLR-triggered macrophages by antagonizing RNA-binding protein tristetraprolin-mediated IL-10 mRNA degradation. J Immunol, 2010. 184(11): p. 6053-9.
177. Nana-Sinkam, S.P., et al., Integrating the MicroRNome into the study of lung disease. Am J Respir Crit Care Med, 2009. 179(1): p. 4-10.
178. Tomankova, T., M. Petrek, and E. Kriegova, Involvement of microRNAs in physiological and pathological processes in the lung. Respir Res, 2010. 11: p. 159.
179. Small, E.M., R.J. Frost, and E.N. Olson, MicroRNAs add a new dimension to cardiovascular disease. Circulation, 2010. 121(8): p. 1022-32.
180. Zhang, W., et al., Gene set enrichment analyses revealed differences in gene expression patterns between males and females. In Silico Biol, 2009. 9(3): p. 55-63.
181. Zhou, T., J.G. Garcia, and W. Zhang, Integrating microRNAs into a system biology approach to acute lung injury. Transl Res, 2011. 157(4): p. 180-90.
182. Wang, J.F., et al., Serum miR-146a and miR-223 as potential new biomarkers for sepsis. Biochem Biophys Res Commun, 2010. 394(1): p. 184-8.
183. Vasilescu, C., et al., MicroRNA fingerprints identify miR-150 as a plasma prognostic marker in patients with sepsis. PLoS One, 2009. 4(10): p. e7405.
184. Nahid, M.A., et al., miR-146a is critical for endotoxin-induced tolerance: IMPLICATION IN INNATE IMMUNITY. J Biol Chem, 2009. 284(50): p. 34590-9.
185. Liu, X., et al., MicroRNA-148/152 impair innate response and antigen presentation of TLR-triggered dendritic cells by targeting CaMKIIalpha. J Immunol, 2010. 185(12): p. 7244-51.
186. Dimmeler, S. and A.M. Zeiher, Circulating microRNAs: novel biomarkers for cardiovascular diseases? Eur Heart J, 2010. 31(22): p. 2705-7.
187. Sun, X., et al., MicroRNA-181b regulates NF-kappaB-mediated vascular inflammation. J Clin Invest, 2012. 122(6): p. 1973-90.
188. Xie, T., et al., MicroRNA-127 inhibits lung inflammation by targeting IgG Fcgamma receptor I. J Immunol, 2012. 188(5): p. 2437-44.
189. Chen, J., et al., miR-193b Regulates Mcl-1 in Melanoma. Am J Pathol, 2011. 179(5): p. 2162-8. 190. Xie, C., et al., CFTR suppresses tumor progression through miR-193b targeting urokinase
plasminogen activator (uPA) in prostate cancer. Oncogene, 2013. 32(18): p. 2282-91, 2291 e1-7. 191. Leivonen, S.K., et al., Identification of miR-193b targets in breast cancer cells and systems
biological analysis of their functional impact. Mol Cell Proteomics, 2011. 10(7): p. M110 005322. 192. Rauhala, H.E., et al., miR-193b is an epigenetically regulated putative tumor suppressor in
prostate cancer. Int J Cancer, 2010. 127(6): p. 1363-72. 193. Wang, H.J., et al., Four serum microRNAs identified as diagnostic biomarkers of sepsis. J Trauma
Acute Care Surg, 2012. 73(4): p. 850-4. 194. Abdalla, S.A., et al., Molecular cloning and characterization of chicken tumor necrosis factor
(TNF)-superfamily ligands, CD30L and TNF-related apoptosis inducing ligand (TRAIL). J Vet Med Sci, 2004. 66(6): p. 643-50.
195. Xie, P., et al., Roles of TRAF molecules in B lymphocyte function. Cytokine Growth Factor Rev, 2008. 19(3-4): p. 199-207.
196. Chen, L.F. and W.C. Greene, Shaping the nuclear action of NF-kappaB. Nat Rev Mol Cell Biol, 2004. 5(5): p. 392-401.
131
197. Mordmuller, B., et al., Lymphotoxin and lipopolysaccharide induce NF-kappaB-p52 generation by a co-translational mechanism. EMBO Rep, 2003. 4(1): p. 82-7.
198. Tang, B.M., S.J. Huang, and A.S. McLean, Genome-wide transcription profiling of human sepsis: a systematic review. Crit Care, 2010. 14(6): p. R237.
199. Matthay, M.A. and J.P. Wiener-Kronish, Intact epithelial barrier function is critical for the resolution of alveolar edema in humans. Am Rev Respir Dis, 1990. 142(6 Pt 1): p. 1250-7.
200. Lee, J.W., et al., Acute lung injury edema fluid decreases net fluid transport across human alveolar epithelial type II cells. J Biol Chem, 2007. 282(33): p. 24109-19.
201. Papapetropoulos, A., et al., Direct actions of angiopoietin-1 on human endothelium: evidence for network stabilization, cell survival, and interaction with other angiogenic growth factors. Lab Invest, 1999. 79(2): p. 213-23.
202. Thurston, G., et al., Leakage-resistant blood vessels in mice transgenically overexpressing angiopoietin-1. Science, 1999. 286(5449): p. 2511-4.
203. Thurston, G., et al., Angiopoietin-1 protects the adult vasculature against plasma leakage. Nat Med, 2000. 6(4): p. 460-3.
204. Xu, J., et al., Mesenchymal stem cell-based angiopoietin-1 gene therapy for acute lung injury induced by lipopolysaccharide in mice. J Pathol, 2008. 214(4): p. 472-81.
205. Hori, S., et al., A pericyte-derived angiopoietin-1 multimeric complex induces occludin gene expression in brain capillary endothelial cells through Tie-2 activation in vitro. J Neurochem, 2004. 89(2): p. 503-13.
206. Yu, H., et al., Recombinant human angiopoietin-1 ameliorates the expressions of ZO-1, occludin, VE-cadherin, and PKCalpha signaling after focal cerebral ischemia/reperfusion in rats. J Mol Neurosci, 2012. 46(1): p. 236-47.
207. Wang, Y.L., et al., Strengthening tight junctions of retinal microvascular endothelial cells by pericytes under normoxia and hypoxia involving angiopoietin-1 signal way. Eye (Lond), 2007. 21(12): p. 1501-10.
208. Ho, J.J., et al., Functional importance of Dicer protein in the adaptive cellular response to hypoxia. J Biol Chem, 2012. 287(34): p. 29003-20.
209. Lee, Y., et al., The nuclear RNase III Drosha initiates microRNA processing. Nature, 2003. 425(6956): p. 415-9.
210. Kim, V.N., J. Han, and M.C. Siomi, Biogenesis of small RNAs in animals. Nat Rev Mol Cell Biol, 2009. 10(2): p. 126-39.
211. Blahna, M.T. and A. Hata, Regulation of miRNA biogenesis as an integrated component of growth factor signaling. Curr Opin Cell Biol, 2013. 25(2): p. 233-40.
212. Ketting, R.F., et al., Dicer functions in RNA interference and in synthesis of small RNA involved in developmental timing in C. elegans. Genes Dev, 2001. 15(20): p. 2654-9.
213. Gregory, R.I., et al., Human RISC couples microRNA biogenesis and posttranscriptional gene silencing. Cell, 2005. 123(4): p. 631-40.
214. Yamagata, K., et al., Maturation of microRNA is hormonally regulated by a nuclear receptor. Mol Cell, 2009. 36(2): p. 340-7.
215. Zhang, X., et al., The ATM kinase induces microRNA biogenesis in the DNA damage response. Mol Cell, 2011. 41(4): p. 371-83.
216. Chendrimada, T.P., et al., TRBP recruits the Dicer complex to Ago2 for microRNA processing and gene silencing. Nature, 2005. 436(7051): p. 740-4.
217. Paroo, Z., et al., Phosphorylation of the human microRNA-generating complex mediates MAPK/Erk signaling. Cell, 2009. 139(1): p. 112-22.
218. Kolch, W., Coordinating ERK/MAPK signalling through scaffolds and inhibitors. Nat Rev Mol Cell Biol, 2005. 6(11): p. 827-37.
132
219. Lee, Y., et al., The role of PACT in the RNA silencing pathway. EMBO J, 2006. 25(3): p. 522-32. 220. Koscianska, E., J. Starega-Roslan, and W.J. Krzyzosiak, The role of Dicer protein partners in the
processing of microRNA precursors. PLoS One, 2011. 6(12): p. e28548. 221. Cheloufi, S., et al., A dicer-independent miRNA biogenesis pathway that requires Ago catalysis.
Nature, 2010. 465(7298): p. 584-9.
