Hussain Habibi1, Amir Atashi2, Saeid Abroun1, Mehrdad Noruzinia3. 1. Department of Hematology, School of Medical Sciences, Tarbiat Modares University, Tehran, Iran. 2. Department of Hematology, School of Medical Sciences, Tarbiat Modares University, Tehran, Iran. Electronic Address: atashia@modares.ac.ir. 3. Department of Medical Genetics, School of Medicine, Tarbiat Modares University, Tehran, Iran.
Beta-thalassemia and sickle cell disease (SCD) are
inherited disorders that are caused by mutations in
beta-globin chain (1, 2), Patients suffering from these
disorders need blood transfusion for survival; however,
iron overload is an important side effect of frequent blood
transfusion leading to liver diseases and heart attack (3,
4). For this reason, researchers have been looking for a
better treatment since long ago.There are several approaches such as hematopoietic stem
cell transplantation (HSCT), gene therapy and utilization
of induced pluripotent stem cells (iPS), employed for
treatment of beta-thalassemia and sickle cell anemia
to ameliorate clinical symptoms of these conditions
and reduce the need for blood transfusion. However,
disadvantages of these approaches such as rare matched
HLA donors, risk of graft versus of disease (GVHD) and
virus vector transmission (5), Beta-thalassemia and SCD
are the most frequent beta-hemoglobinopathies in the
world, and a great number of countries that are affected
by these diseases cannot perform HSCT and gene therapy
easily (6, 7); so, researchers are looking for alternative
therapies with lower risks and cost but higher chance
of success. Typically, these patients have no symptoms
at birth and clinical manifestations appear with HbF
(α2γ2) switching to HbA (a2ß2) six months after birth
(8-10). Scientists have found that high levels of HbF can
ameliorate clinical symptoms in Beta-thalassemia and
SCD patients (11).There are some ß-like thalassemia conditions such as
Hereditary Persistence of Fetal Hemoglobin (HPFH), dßthalassemia
and Corfu anemia that show elevated HbF
and these patients do not have severe anemia and do not
usually need blood transfusion (12-14). For three decades,
scientists have focused to find pharmacological agents
to reactivate Gamma-globin gene after birth (15-19).
However, toxicity associated with these agents and other
issues have restricted their use. Although hydroxyurea
has been approved by the FDA as a HbF-inducing agent,
its usage has been limited because it is not effective for all
SCD patients and was effective only in few ß-thalassemia
patients; also, it has a narrow therapeutic index due to
decreased blood cells (especially neutrophils) count
(20-23). Importantly, in countries with high prevalence
of the mentioned hemoglobinopathies, utilization of
pharmacological agents that can increase HbF in these
patients is more affordable as compared to other methods.Studies have shown that there are several specific
inhibitors for Gamma-globin expression after birth
such as Histone deacetylase ½ (HDAC1/2) and B-cell
lymphoma/leukemia 11a (BCL11a) (24-27). Macari et al.
(28) showed that Simvastatin (SIM) as a BCL11a inhibitor
can induce HbF in CD34+ obtained from peripheral blood
cells. Also, Bates et al. (29) have noted that Romidepsin
(ROM) can increase HbF in cutaneous T-cell lymphoma
patients via inhibition of HDAC1/2. Many studies have
introduced several agents such as BCL11a, HDAC1/2,
KLF1, SOX2, MBD2, DRED, and DNMT that inhibit
Gamma-globin expression after birth (25). Also, SIM
and ROM are reported to be able to inhibit BCL11a and
HDAC1/2, respectively, and were approved by the FDA
for reduction of cholesterol, prevention of cardiovascular
diseases (for Simvastatin) and treatment of cutaneous
T-cell lymphoma (for Romidepsin). In this study, we
evaluated the synergistic effect of ROM and SIM on
induction of fetal hemoglobin in erythroid progenitors
differentiated from cord blood stem cells.
Materials and Methods
CD34+ cells separation and expansion
In this experimental study, umbilical cord blood samples
(n=5) were collected at Sarem Hospital according to
the guidelines of Medical Ethics Committee of Sarem
Research Center and Tarbiat Modares University, Then,
cord blood bags were transferred to the research laboratory
of Sarem Hospital for further analysis. For isolation of
mononuclear isolation cells, we used gradient separation
(Ficoll-Paque plus GE Healthcare), and CD34+ cells
were separated by a MACS procedure (Miltenyi Biotec,
CD34 Micro Bead Kit, Germany). The mean number
of cells in each bag was 1.2×106. Separated cells were
checked for CD34 expression by flowcytometry using
FITC-conjugated anti-CD34. Also, the CD34+ cells were
cultured in expansion medium [StemLine II serum-free
culture medium, Sigma S0192 Containing stem cell factor
(SCF) 100 ng/ml, IL3 1 ng/ml, thrombopoietin (TPO) 100
ng/ml, fms related tyrosine kinase 3 (Flt3) 100 ng/ml] for
4 days. We counted cells on the first and fourth day of
expansion; number of cells of five bags on the first day
was 6 million, which became 18 million after expansion.
Also, we evaluated viability of cells on the 4th day by
trypan blue staining.
MTT assay
We dissolved SIM (Cayman chemical company) and
ROM (AOBIOUS) in dimethyl sulfoxide (DMSO) and
then evaluated the cytotoxicity of SIM and ROM using
MTT assay (Sigma, Germany). Here, 103 CD34+ cells
were cultured and treated with different concentrations
of SIM and ROM into each well in a 96-well microplate
for 48 hours at 37oC with 95% humidity and 5% CO2.
Then, the media was removed slowly (without removing
cells) and 10 µl MTT reagent was added. After 4 hours
of incubation at room temperature in the dark, 50 µl
DMSO was added to solubilize the formazan particles.
Then, optical density of each well was measured at
570 nm. Based on the MTT results, we used 10 µM/
ml SIM and 10 nM/ml Romidepsin, as they showed the
greatest effectiveness with the least cytotoxicity at these
concentrations.
Erythroid differentiation, fetal hemoglobin induction
and colony assay
Cells were cultured in erythroid differentiation
medium [Stem line, Sigma S0192 containing
erythropoietin (EPO) 3U/ml, SCF 100 ng/ml, IL3 5 ng/
ml, transferrin 5 ng/ml] for 7 and 14 days. Erythroid
differentiation was confirmed by flowcytometry
(Thermo Fisher, ABI, Attunetm NxT Flow Cytometer)
following incubation with PE-conjugated CD36
(Invitrogen, Denmark) and FITC-conjugated CD71
(Invitrogen, Denmark) monoclonal antibodies.
According to the results, on the 14th day, 78.5% and
63.3% of the differentiated cells expressed CD36 and
CD71, respectively as erythroid lineage markers. We
used five groups namely, control [that was culture
in erythroid differentiation medium (EDM) only],
ROM [that was treated with EDM+ROM (10 nM/
ml)], SIM [that was treated with EDM+SIM (10 µM/
ml)], Romidepsin/SIM (ROM/SIM, that was treated
with EDM+ROM [10 nM/ml)+SIM (10 µM/ml)], and
sodium butyrate [SB, that was treated with EDM+SB
(100 µM/ml)]. The Changing of the condition medium
for all the groups was performed once a week using
150 µl of fresh medium.It should be noted that we used the SB group
because, it was confirmed that SB can increase HbF,
we used SB group to compare its results with the
other groups. Colony assay evaluation was done using
1×103 CD34+ cells that were vigorously mixed in 3
ml of methylcellulose medium (MethoCult H4230,
Stem Cell Technologies) containing EPO 3 U/ml,
SCF 100 ng/ml, IL3 5 ng/ml, transferrin 5 ng/ml.
Then, the methylcellulose medium was placed into
two 30-mm sterile petri dishes, each containing 1.5
ml of the medium, then, cells were spread slowly
and subsequently placed in an incubator with 95%
humidity and 5% CO2 at 37°C. After 14 days, the
erythroid colonies were scored by a phase-contrast
inversion microscope according standard colony assay
protocol.
RNA extraction and cDNA synthesis
RNA extraction was done by RNX-Plus solution
for total RNA isolation (SinaClon Bioscience, Iran)
and quality control procedure of the isolated RNA was
undertaken with measurement of the absorbance at
260/280 nm by Biophotometer; the isolated RNA had
an optical density between 1.9-2 with double distilled
water used as blank. Afterward, cDNA was produced
by a GeneAll kit (HyperScriptTM Reverse Transcriptase,
South Korea); cDNA synthesis was done in 20 µl volume
containing 3 µl extracted RNA, 1 µl dNTP, 1 µl oligo dT
and 9 µl nuclease-free distilled water, which was heated
to 65°C for 5 minutes and then placed on ice. After
that, 6 µl of RT buffer including 10X RTase reaction
buffer, DTT, HyperScriptTM Reverse Transcriptase and
ZymAllTMRNase inhibitor were added and incubated for
50 minutes at 55°C followed by 5 minutes at 85°C.
Evaluation of gene expression using real-time
polymerase chain reaction
The primer sequences used to evaluate the expression
levels of Gamma-globin, BCL11a and genes,
are mentioned in Table 1. The primers were designed using
UCSC and NCBI databases and Gene runner software.
At least one of the designed primers was pair spans an
exon junction to avoid gene amplification on DNA. Also,
cDNA synthesis by the GeneAll kit involved a step to
assure that traces of contaminating DNA were removed.
Then, polymerase chain reaction (PCR) was implemented
in a 20-µl reaction in cap strip at 95°C for 10 minutes
followed by 40 cycles at the denaturation temperature (30
seconds at 95°C), annealing temperature (30 seconds at
60°C) and extension temperature (30 seconds at 70°C).
Each real-time PCR reaction was performed in duplicate.
We used beta actin primers and control group to normalize
our data by real time instrument (Applied biosystem, Step
one, USA) and real time master mix (SYBR, Ampliqon
real time master mix2x, high ROX). The ABI step one
software was used to analyze data, including the cycle
threshold (Ct), amplification plot and melting curve for
each product. Moreover, efficiency of primers for each
gene was evaluated by a standard curve generated using
fourfold dilution series of synthesized cDNAs. Real-time
results analysis was done by 2-ΔΔct method and finally,
statistical analysis for each gene was done by GraphPad
Prism 7.
Immunocytochemistry
Erythroid progenitors differentiated from cord
blood on the 14th day, were collected and washed with
phosphate buffered saline (PBS) three times. Then,
105 cells were suspended in 1 ml of PBS, and cytospin
cells were prepared on slide and fixed with absolute
methanol (Merck, Germany) for 10 minutes and the
slides were completely air-dried. Next, the fixed cells
were permeabilized by 0.1% Triton X-100 at 18-25oC
for 10 minutes. Then, we performed immunostaining
by anti-HbF (BD Pharmingen™, Denmark) conjugated
with fluorescein isothiocyanate (FITC) overnight at 4oC
in the dark. Next, the stained cells were photographed
using a fluorescence microscope (Motic BA410, with
Moticam pro 282, Canada). We used newborn blood as
positive sample for quality-control of staining protocol;
also, we compared HbF induction between the control
group (untreated) and the groups treated with ROM and
Simvastatin.
Results
CD34+ Cells isolation and colony assay
The CD34+ cells were isolated by MACS positive
selection and evaluation of CD34+ cells purity was done
with anti CD34-FITC and flowcytometry. Flowcytometry
analysis showed that 89.4% of the cells isolated from cord
blood expressed CD34 as a HSC marker (Fig .1A). The
viability of isolated cells was 99% as assessed by trypan
blue staining (only 1% of cells were stained by trypan blue
and the rest of them were alive). Also, the result of colony
assay on the 14th day confirmed that the isolated cells can
differentiate into erythroid commitment cell (Fig .1B).
The flowcytometry analysis on the 14th day showed that
the hematopoietic stem cells (HSCs) that were cultured
in the erythroid differentiation medium expressed CD36
(78.5%) and CD71 (63.7%) as erythroid markers; thus,
HSCs isolated from cord blood could differentiate into
erythroid progenitors cells (Fig .1C).
Fig.1
Flowcytometry result for CD34 isolation and erythroid differentiation. A. It was observed that 89.4% of the isolated cells of cord blood with MACS
expressed CD34 as a HSC marker, B. The CD34+ isolated cells that were cultured in MethoCult medium, could differentiate into the erythroid commitment
cells after 14 days (×100), and C. After 14 days 78.5 and 63.7% of CD34+ cells that were cultured in erythroid differentiation medium, expressed CD36 and
CD71, respectively as erythroid markers.
MACS; Magnetic-activated cell sorting and HSC; Hematopoietic stem cell.
Real time polymerase chain reaction primer sequences
Relative Gamma-globin gene expression
Evaluation of Gamma-globin gene expression by real
time PCR showed that SIM and ROM treatment led to
1.7-fold increase in Gamma-globin gene expression
compared to untreated group, on the 7th and 14th day.
However, when we used SIM and ROM together (SIM 10
µM/ml and ROM 10 nM/ml), 3-fold increment in gamma
gene mRNA was observed compared to untreated group,
on the 7th and 14th day (Fig .2). These findings indicated
that SIM and ROM can increase Gamma-globin gene
expression synergistically (P<0.05). Also, no significant
differences were observed in gamma expression between
the 7th and 14th days; thus, we could have finished our study
on the 7th day. However, to compare the results obtained
on the 7th day with those of the 14th day, we continued the
experiment until the 14th day. In this study, we used SB as
a drug which was shown to induce Gamma-globin gene
induction, and compared its results with those of ROM
and SIM treatment. Our results showed that ROM is more
marked upregulation of Gamma-globin gene expression
compared to SB.
Fig.2
SIM and ROM can increase Gamma-globin gene expression, but the combination of these drugs synergistically induced Gamma-globin gene
expression. Also, ROM/SIM can induce Gamma-globin to higher levels compared to SB, as a confirmed HbF inducer.
*; P<0.05, ROM; Romidepsin, SIM; Simvastatin, SB; Sodium butyrate, and HbF; Fetal hemoglobin.
Relative BCL11a gene expression
Evaluation of BCL11a gene expression by real time
step one software showed that ROM no significantly
inhibited BCL11a (mean: 0.6-fold higher than the control
group), whereas SIM treatment led to a significant
inhibition of BCL11a mRNA transcription (mean: 0.065fold
higher than that of the control group, P<0.05). Also,
consistent with our study, Macari et al. (28) reported that
SIM can inhibit BCL11a gene expression. In addition,
our results showed that the combination of ROM and
SIM significantly downregulated BCL11a compared to
untreated group (P<0.05, Fig .3).
Fig.3
Results of the 7th and 14th day showed that only SIM and ROM/SIM can significantly inhibit BCL11a gene transcription, as compared to the control
group, while ROM and SB did not show significant inhibition of BCL11a.
*; P<0.05, ROM; Romidepsin, SIM; Simvastatin, and SB; Sodium butyrate.
Flowcytometry result for CD34 isolation and erythroid differentiation. A. It was observed that 89.4% of the isolated cells of cord blood with MACS
expressed CD34 as a HSC marker, B. The CD34+ isolated cells that were cultured in MethoCult medium, could differentiate into the erythroid commitment
cells after 14 days (×100), and C. After 14 days 78.5 and 63.7% of CD34+ cells that were cultured in erythroid differentiation medium, expressed CD36 and
CD71, respectively as erythroid markers.MACS; Magnetic-activated cell sorting and HSC; Hematopoietic stem cell.SIM and ROM can increase Gamma-globin gene expression, but the combination of these drugs synergistically induced Gamma-globin gene
expression. Also, ROM/SIM can induce Gamma-globin to higher levels compared to SB, as a confirmed HbF inducer.
*; P<0.05, ROM; Romidepsin, SIM; Simvastatin, SB; Sodium butyrate, and HbF; Fetal hemoglobin.Results of the 7th and 14th day showed that only SIM and ROM/SIM can significantly inhibit BCL11a gene transcription, as compared to the control
group, while ROM and SB did not show significant inhibition of BCL11a.*; P<0.05, ROM; Romidepsin, SIM; Simvastatin, and SB; Sodium butyrate.
Romidepsin and Simvastatin effects on HDAC1 and
HDAC2 expression
Relative quantitative real time PCR was done for the
CD34+ cells that had been treated with ROM, SIM and
ROM/SIM using the primers mentioned in Table 1 for
HDAC1/2. Our results showed that HDAC1 expression
was significantly downregulated by ROM and ROM/SIM
(P<0.05), but not by SIM alone. In addition, results of the
quantitative real time PCR showed that neither ROM, SIM
nor ROM/SIM had significant effects on the expression of
HDAC2. It seems that the effect of the mentioned drugs
on Gamma-globin gene expression only was mediated by
their effects on HDAC1 (Fig .4) and BCL11a inhibition
(Fig .3).
Fig.4
Effect of ROM, SIM and ROM/SIM on CD34+ cells showed that ROM and ROM/SIM can only downregulate HADC1, but not HDAC2. Seemingly,
Gamma-globin upregulation is related to HDAC1 and BCL11a downregulation not HDAC2.
*; P<0.05, ROM; Romidepsin, SIM; Simvastatin, SB; Sodium butyrate.
ROM, SIM and ROM/SIM increased HbF in the
treated cells
HbF was evaluated using FITC-conjugated anti-F
and fluorescence microscopy. In our study, we used
two controls as follows: i. Newborn blood was used as
positive control of fluorescence staining and ii. Untreated
group was used to compare its results in terms of HbF
production with those of the treated groups (ROM, SIM,
and ROM/SIM). Results of fluorescence staining showed
that both ROM and SIM can increase HbF in erythroid
progenitors differentiated form cord hematopoietic stem
cells. These findings are consistent with results of Makala
et al. (30) for ROM and Macari et al. (28) for SIM.
However, the results revealed greater HbF production,
when ROM/SIM (ROM 10 nM/ml and SIM 10 µM/
ml) were added to erythroid differentiation medium, as
compared to results obtained from using ROM and SIM
alone (Fig .5A). These results suggest that ROM and SIM
increase HbF production synergistically; however, results
of gene expression by real time analysis in our study are
more evident. Real time results showed that ROM/SIM
significantly downregulated BCL11a and HDAC1 while
caused upregulation of HbF expression (2.91-fold higher
than the untreated group) on the 7th and (3.09-fold higher
than the untreated group) 14th days (Fig .5B).
Fig.5
Immunofluorescence staining for HbF production and Real time evaluation for BCL11a, HDAC and Gamma-globin genes expression. A. Results of
immunofluorescence staining showed that ROM and SIM can induce HbF production in erythroid progenitors differentiated form cord hematopoietic stem
cells: 1) Newborn blood was used as positive control for immunofluorescence staining, 2) Untreated group, 3) ROM group, 4) SIM group, 5) ROM/SIM
group (×100) and B. CD34+ cells were treated with ROM, SIM and ROM/SIM for 7 and 14 days. Results of gene expression showed that ROM and SIM can
induce Gamma-globin gene expression and downregulate BCL11a and HDAC1 genes expression on the 7th and 14th day, but the combination of ROM and
SIM can induce Gamma-globin gene transcription synergistically by downregulation of BCL11a and HDAC1 genes on the 7th and 14th day. We also found
that ROM and SIM had no effect on HDAC2 gene expression.
Effect of ROM, SIM and ROM/SIM on CD34+ cells showed that ROM and ROM/SIM can only downregulate HADC1, but not HDAC2. Seemingly,
Gamma-globin upregulation is related to HDAC1 and BCL11a downregulation not HDAC2.
*; P<0.05, ROM; Romidepsin, SIM; Simvastatin, SB; Sodium butyrate.Immunofluorescence staining for HbF production and Real time evaluation for BCL11a, HDAC and Gamma-globin genes expression. A. Results of
immunofluorescence staining showed that ROM and SIM can induce HbF production in erythroid progenitors differentiated form cord hematopoietic stem
cells: 1) Newborn blood was used as positive control for immunofluorescence staining, 2) Untreated group, 3) ROM group, 4) SIM group, 5) ROM/SIM
group (×100) and B. CD34+ cells were treated with ROM, SIM and ROM/SIM for 7 and 14 days. Results of gene expression showed that ROM and SIM can
induce Gamma-globin gene expression and downregulate BCL11a and HDAC1 genes expression on the 7th and 14th day, but the combination of ROM and
SIM can induce Gamma-globin gene transcription synergistically by downregulation of BCL11a and HDAC1 genes on the 7th and 14th day. We also found
that ROM and SIM had no effect on HDAC2 gene expression.*; P<0.05, HbF; Fetal hemoglobin, ROM; Romidepsin, SIM; Simvastatin, and SB; Sodium butyrate.
Discussion
ß-thal and SCD are the most frequent hemoglobinopathies
in the world. Almost more than 80% of patients with
ß-thal and SCD are born in non-industrial and developing
countries, while these countries do not have adequate
facilities for prenatal screening, diagnosis, treatment and
proper management of these patients (31, 32).Traditional therapies such as frequent blood transfusion
can lead to iron overload, liver disease, heart attack,
risk of virus transmission and alloimmunization (33).
Scientists offer alternative therapeutic approaches such as
HSCT, gene therapy and iPS usage, to ameliorate clinical
symptoms and reduce need for blood transfusion. But,
these approaches are hardly available in non-industrial
countries. Thus, in the last three decades, researchers
have tried to present therapeutic approaches with lower
risk and cost and easily available (15, 18, 34-36).HbF inducing drugs are the best approach to ameliorate
clinical symptoms of ß-thal and SCD. Atashi et al.
(18) confirmed that SCF and tumor growth factor-beta
(TGF-ß) can induce HbF in CD133+ cells. Also, these
transcription factors have synergistic effects for HbF
induction; however, SCF and TGF-ß are not approved by
the FDA and there is major concern that their long-term
usage may be carcinogenic, in addition to the fact that SCF
and TGF-ß usage are not cost effective. Ahmadvand et al.
(37) showed that thalidomide at 100 µM concentration,
can induce HbF in CD133+ cells differentiated into
erythroid lineage 1.5- fold higher than the control group.
Our results showed that SIM (10 µM) and ROM (10 nm)
can induce HbF production 3.09-fold higher than the
control group. Kukreja et al. (38) introduced some natural
agents that can induce HbF. However, there is still need
for novel agents that safely induce HbF and decrease
need for blood transfusion. Constantoulakis et al. (39)
showed SB and 5-azacytidine have synergistic effect for
HbF induction; However, there are concerns over the
carcinogenic potential of 5-azacytidine. Also, SB and its
derivatives have disadvantages such as requirement of
high doses (15-20 g/day), short half-life, and unpleasant
smell (40).Currently, hydroxyurea is the only drug approved
by the FDA for these patients though it possesses side
effects such as undesirable long-term carcinogenesis
and blood suppression and being not effective for all
patients. For these reasons, efforts are continuing to
introduce better drugs. Since inhibition of HbF occurs in
multiple pathways after birth, we decided to evaluate the
effect of the combination of two FDA-approved drugs
with different mechanism for HbF induction (10). Our
results showed that ROM as a HDAC inhibitor and SIM
as a BCL11a inhibitor can considerably induce HbF in
hematopoietic stem cells. The combination of SIM and
ROM caused simultaneous downregulation of BCL11a
and HDAC1 but significantly increased HbF expression.
Similarly, Elizabeth et al. showed that combination of
SIM and t-butylhydroquinone increases Gamma-globin
expression 3.2-fold higher than the control group.
Currently, SIM is using reduction of cholesterol and
prevention of cardiovascular diseases. Also, no serious
side effect has been reported following long-period
usage of Simvastatin, yet. Moreover, ROM is using for
cutaneous T-cell lymphoma treatment and both drugs are
approved by the FDA (29). In addition, our results showed
that the combination of these drugs increases HbF. Thus,
we suggest their concurrent use for HbF induction as a
therapeutic approach in ß-thal and SCD patients.
Conclusion
HbF inducing is the best approach for treatment of
patients with ß-thal and SCD, and hydroxyurea is the only
FDA-approved drug for HbF induction, but it cannot be
used for all of ß-thal patients and it has side effects such as
suppression of blood counts. Results of our study showed
that the combination of ROM and SIM simultaneously
caused downregulation of HDAC1 and BCL11a while
induced Gamma-globin gene expression. These drugs
are FDA-approved and thus, can be used together to
ameliorate clinical symptoms in ß-thal and SCD patients.
We hope ROM and SIM combination therapy may lead
to promising results in ß-thal and SCD patients with least
side effects and reduce need for blood transfusion.
Table 1
Real time polymerase chain reaction primer sequences
Authors: Jeffrey D Lebensburger; Tamara I Pestina; Russell E Ware; Kelli L Boyd; Derek A Persons Journal: Haematologica Date: 2010-04-07 Impact factor: 9.941
Authors: Susan P Perrine; Serguei A Castaneda; David H K Chui; Douglas V Faller; Ronald J Berenson; Noppadol Siritanaratku; Suthat Fucharoen Journal: Ann N Y Acad Sci Date: 2010-08 Impact factor: 5.691
Fig.1
Fig.2
Fig.3
Fig.4
Fig.5