Download Relationship Between MRI Features and miRNA Gene Expression

Relationship between MRI features and miRNA gene expression in patients with glioblastoma
multiforme
LI Wenbin1, CHEN Huiyuan1, ZHANG Wei2, YAN Wei2, SHI Rui1, LI Shaowu3 and JIANG Tao2
1
Cancer Center, Beijing Shijitan Hospital, Capital Medical University, Beijing, 100038, China,
Department of Neurosurgery, 3Department of Radiology, Beijing Tiantan Hospital, Capital Medical
University, Beijing, 100050, China
2
Correspondence to: Dr. JIANG Tao, Department of Neurosurgery, Beijing Tiantan Hospital, Capital
Medical University, Beijing, 100050, China (Fax: 86-10-67118050. Email: [email protected])
This work was supported by grants from Chinese National Key Project of Science and Technology
Supporting Programs (No. 2007BAI05B08), and the National Natural Science Foundation of China (No.
30772238).
Key words：glioblastoma; magnetic resonance imaging; gene expression; microRNAs
Abstract
BACKGROUND: Magnetic resonance imaging (MRI) is commonly utilized as the part of the diagnostic
workup for the clinical diagnosis of glioblastoma multiforme (GBM), further guiding the clinical
treatment of this aggressive cancer. Recent research has shown that micro RNAs (miRNAs) may act as
oncogenes, or in some cases, tumor suppressor genes that in turn may reflect the genotypic features of
GBM.
METHODS: In order to identify the relationship between the radiographic findings of MRI with those
identified changes in miRNA gene expression of GBM, we reviewed the MRI images of GBM patients
and compared them to the identified miRNA expression profiles utilizing microarray analysis of paired
GBM tumor samples. We chose five MRI imaging features: 1. contrast tumor enhanced/necrosis ratio, 2.
contrast tumor enhanced/T2 ratio, 3. multiple lesions, 4. hemorrhage and 5. necrotic volume. The
relationship between these five imaging features and miRNA expression was studied using Significance
1
Analysis of Microarrays (SAM) analysis.
RESULTS: We found that the expression of miRNA’s hsa-miR-892b, hsa-miR-892a, hsa-miR-888 was
inversely correlated with a enhanced/necrosis ratio ≥ 1. The miRNA’s hsa-miR-95, hsa-miR-498 and
hsa-miR-1300 were associated with a contrast tumor enhanced/T2 ratio ≥ 1. The miRNA’s hsa-miR-612，
hsa-miR-524-3 and hsa-miR-1282 were associated with multiple lesions identified on MRI and the
expression of miR-221 was associated with hemorrhage in GBM. The expression of miR-let-7, including
miR-let-7f, miR-let-7i, miR-let-7f-1*, were down-regulated in the hemorrhage group. The gene
expression of of miRNA’s hsa-miR-140-5p, hsa-miR-30e and hsa-miR-301a was relatively low when
compared with the larger necrotic volume group as identified by MRI.
CONCLUSION: The miRNA gene expression profiles correlate with several select MRI features of
patients with GBM. Further analysis of key imaging features of MRI with correlation with miRNA gene
expression patterns may help to guide treatment decisions based upon these unique correlative profiles of
GBM.
Introduction
Glioblastoma multiforme (GBM) is the most common primary brain tumor in adults. Magnetic resonance
imaging (MRI) is a commonly utilized imaging modality with a high sensitivity for defining tumor size
and spacial configuration of GBM, often helpful in the subsequent planning of clinical treatment.
However, the specific MRI modalities identified with GBM are quite variable, even when comparing
nearly identical histologic GBM specimens. This suggests that MRI may be useful in differentiating
various sub-types of GBM based upon specific MRI features, possibly correlating with certain features of
GBM tumor biology.
2
The recent development of cutting-edge technologies now allows for the functional genomics of cancer
cells, with gene microarray analysis previously the mainstay of research into the cancer genome.
However, microarray analysis can be technically challenging and further burdened with complex
comparative statistical analyses. The IlluminaTM human miRNA gene expression 12-sample universal
BeadChip (Illumina, USA) contains 1,146 assays, with the ability to detect a wide range of known
miRNA’s, as described by the Sanger Institute’s miRBase release 12.0.
There has been an increasing number of studies that have shown miRNA acting as oncogenes or as tumor
suppressor genes, based upon whether they function as key modulators in pathways involved in cell
proliferation and differentiation. Investigating the relationship between features seen on MRI and miRNA
gene expression may help to further our understanding of the growth characteristics of GBM, possibly
sub-classifying various GBM’s based upon these unique characteristics. Additionally, it may benefit the
healthcare system by creating a convenient and low-cost, genetically-based diagnostic method, with the
potential for a variety of broad-based clinical applications.
Material and Methods
Patients
A total of 49 patients diagnosed with GBM were selected from the Tiantan Hospital between January
2006 and July 2008. The study group was comprised of 33 males and 15 females, with the range of age
from 17 to 70 years (average age, 45 years old). All of the patients were diagnosed with GBM, World
Health Organization (WHO) Grade IV. All GBM diagnoses were confirmed by at least 2
neuropathologists at our institute.
Treatment options
3
All 49 patients underwent an operation, with 33/49 (67%) receiving adjuvant radiotherapy and
chemotherapy (comprehensive treatment group), 9/49 (18%) receiving radiotherapy only (radiotherapy
group) and 3/49 (6%) receiving chemotherapy only (chemotherapy group). There were 4/49 (8%) cases
that were treated with surgery alone (surgery group). The treatment regimen for radiation consisted of a
total of 50-60 Gy, 2 Gy per day for a duration of approximately 5-6 weeks. For chemotherapy, patients
received temozolomide, 75 mg/m2 of body surface area daily (during radiotherapy if concomitant),
followed by 5 days of treatment every month for 6 months, with the dosing at 150-200 mg/m2 of total
body surface area. This study was approved by the Ethics Committee of Beijing Tiantan Hospital, Capital
Medical University.
Patient outcome
The overall survival for each patient was calculated from the date of surgery to the date of their death. If
the patient was still alive, we used the date of their last follow-up visit. Survival assessment was last
performed in September 2011, with 45/49 (92%) patients dead of disease and an average survival of 79.4
weeks (19.8 months). A total of 4/49 (8%) patients were still alive with an average survival of 194.3
weeks (48.5 months). The overall average survival for all 49 patients was 88.8 weeks (22.2 months).
Patients were further categorized into the four groups as described above: 1. comprehensive treatment
group (surgery followed by chemotherapy and radiotherapy), 2. chemotherapy group (surgery plus
chemotherapy), 3. radiotherapy group (surgery plus radiotherapy); 4. surgery alone group.
The primary end point of the study was overall survival (OS), defined as the period from the date of
surgery until death from any cause, or until the last follow-up visit of September 2011. The statistical
software, SPSS 19.0(SPSS, USA), was used to estimate survival probabilities between the treatment
options. To assess statistical significance for the effect of each feature on survival, the log-rank test was
4
used for data analysis of patients who received surgery, radiotherapy and chemotherapy. For all analyses,
a value of P <0.05 was considered statistically significant.
MRI features
A total of 43 patients with MRI images were available for review. All MRI studies were performed using
a 1.5T MR scanner, with T1- and T2-weighted non-enhanced and contrast-enhanced images obtained with
three dimensions (axial, sagittal, and coronal). Each patient with MRI images underwent a detailed and
thorough comparison of pre-operative and post-operative MRI features. The selected images were then
photographed utilizing a SONY cybershotTM digital camera (SONY, Japan). The pictures were stored on a
CD-ROM and hard disk storage with the size of 2272x1704. MRI scans were reviewed by two
experienced neuroradiologists who were blinded to the pathological and miRNA expression results.
Five imaging features were selected for this study. The imaging definitions are as follows:
1) Contrast tumor enhanced /necrosis ratio: “1” on behalf of the ratio ≥1, “2” on behalf of the ratio＜ 1
2) Contrast tumor enhanced/T2 ratio: “1” on behalf of the ratio ≥1, “2” on behalf of the ratio＜ 1
3) Multiple lesions: “1” for multiple lesions, “2” for single lesion
4) Hemorrhage: “1” for cases with hemorrhage, “2” for cases without hemorrhage
5) Necrotic volume: “1” for cases smaller than median necrotic volume, “2” for cases larger than median
necrotic volume.
Microarray
Tissue samples from the same 43 cases whose MRI images were available were procured at the time of
the initial operation and immediately snap frozen in liquid nitrogen upon removal of the tumor. After
assessment of the total percentage of tumor cells within the specimen, only samples with greater than
80% tumor cells were selected. Total RNA was extracted from tissue samples using mirVana miRNA
5
Isolation kit (Ambion, USA). The quality and quantity of RNAs were tested by electrophoresis and
spectrophotometer, respectively. Briefly, one microgram of total RNA from each sample were hybridized
on Illumina Human MiRNA Expression 12-sample Universal BeadChipsTM, according to the
manufacturer’s instructions. Data were processed by Significance Analysis of Microarrays (SAM) to
determine significant differences in miRNA expression.
Results
Correlation between overall survival and treatment options
The average survival (AS) and median survival (MS) of the patients in each group (comprehensive group,
radiotherapy group, chemotherapy group and surgery group) are shown in Table 1. The longest survival
was 259 weeks with the shortest survival of 34 weeks in the comprehensive treatment group. The 1-year,
2-year and 3-year survival rates were 90% (30/33), 42% (14/33) and 15% (5/33), respectively, in the
comprehensive treatment group. The longest and shortest survival was 118 and 81 weeks, respectively, in
the chemotherapy group, 138 weeks and 19 weeks in radiotherapy group and 55 weeks and 2 weeks in the
surgery alone group.
The overall survival (OS) for all patients within the different treatment groups are shown in Figure 1. We
compared the OS between the comprehensive treatment group and radiotherapy group (log rank χ2
3.711，P=0.054), the comprehensive group and surgery alone group (log rank χ2 25.835，P <0.001), the
radiotherapy and chemotherapy group (log rank χ2 0.353, P= 0.552), the radiotherapy and surgery alone
group (log rank χ2 1.418, P=0.234), the chemotherapy and surgery alone group (log rank χ2 5.629,
P=0.018). From these data, it is clear that the differences in OS between the comprehensive and surgery
6
alone groups, as well as the differences between the chemotherapy and surgery alone group were
statistically significant. The noted differences in OS between other groups did not show statistical
significance, with further research necessary in order to make any clear conlcusions as to the optimal
treatment regimen for GBM patients. Although the overall number of patients in this study are small,
there is a marked difference in the OS for those treated in the comprehensive treatment group.
Relationship between miRNA expression and MRI features
Relationship between miRNA expression and contrast tumor enhanced /necrosis ratio
Figure 2 shows the five imaging features of case NO. 31. The miRNA’s, Hsa-miR-892b, hsa-miR-892a
and hsa-miR-888 showed low levels of expression in the enhanced/necrosis ratio ≥ 1 group and high
levels of expression in the enhanced/necrosis ratio <1 group (Figure 3A). The SAM analysis indicates a
statistically significant difference (false discovery rate % = 0)
Relationship between miRNA expression and contrast tumor enhanced/T2 ratio
The miRNA’s, Hsa-miR-95, hsa-miR-498, hsa-miR-1300 ， hsa-miR-801:9.1, hsa-miR-594:9.1,
hsa-miR-1291, hsa-miR-1244, hsa-miR-1307 and hsa-miR-335* all showed high expression levels in the
contrast tumor enhanced/T2 ratio ≥ 1 group and low expression patterns in the contrast tumor
enhanced/T2 ratio <1 group. On the other hand, hsa-miR-548 showed low expression in the contrast
tumor enhanced/T2 ratio ≥ 1 group and high expression in the contrast tumor enhanced/T2 ratio < 1 group
(Figure 3B). SAM analysis of microarrays again showed statistical significance, with a false discovery
rate (%) = 7.226107226.
Relationship between miRNA expression and multiple lesions
The miRNAs, Hsa-miR-612，hsa-miR-524-3, hsa-miR-1282 and hsa-miR-1178 showed high expression
patterns in the multiple lesions group and low expression in the single lesion group. Conversely,
7
hsa-miR-20a, hsa-miR-17 showed low miRNA gene expression levels in the multiple lesions group and
high expression in the single lesion group (Figure 3C). SAM analysis of microarrays indicated a statistical
significance, with a false discovery rate (%) = 12.08236208.
Relationship between miRNA expression and Hemorrhage
Hsa-miR-34c-5o, hsa-miR-452*:9.1, hsa-miR-520e, hsa-miR-615-5o, Hsa-miR-455-5p, hsa-miR-944,
hsa-miR-375,
Hsa-miR-15a*,
hsa-miR-559,
hsa-miR-202*:9.1,
Hsa-miR-1184,
hsa-miR-374b,
hsa-miR-34a, Hsa-miR-517a, hsa-miR-519c-3p, Hsa-miR-518c*, hsa-miR-933, hsa-miR-502-5p,
Hsa-miR-187*, hsa-miR-1182, hsa-miR-887, Hsa-miR-130a, hsa-miR-431, hsa-miR-214* showed high
expression in hemorrhage group and low expression in non-hemorrhage group. Hsa-miR-92a,
hsa-miR-594:9.1, hsa-let-7f, hsa-miR-454*，hsa-miR-221, hsa-miR-30e*, hsa-let-7i, hsa-miR-28-5o,
hsa-miR-28-3p, hsa-miR-940, hsa-miR-664, hsa-miR-1201, hsa-miR-432, hsa-miR-421, hsa-miR-98,
hsa-miR-342-5o, hsa-miR-628-3o, hsa-miR-628-5p, hsa-miR-26b*, hsa-miR-151-3p, hsa-miR-128,
hsa-let-7f-1*, hsa-miR-488*, hsa-miR-1271 showed low expression in hemorrhage group and high
expression in non-hemorrhage group (Figure 3D). SAM analysis of microarray indicates a statistically
significant difference, with a false discovery rate (%) = 10.18181818.
Relationship between miRNA expression and Necrotic volume
Hsa-miR-140-5p, hsa-miR-30e, hsa-miR-301a, Hsa-miR-17-5p:9.1, hsa-miR-20a, hsa-miR-18a showed
low expression in larger necrotic volume group and high expression in smaller necrotic volume group
(Figure 3E). SAM analysis again indicates statistical significance, with a false discovery rate (%) =
18.14296814.
Discussion
8
Currently, the most common treatment regimen for the treatment of GBM is an initial attempt at surgical
resection, followed by adjuvant radiotherapy given concomitantly with chemotherapy (1). Most of the
patients with GBM will recur within 2 years of their initial operation, ultimately dying of their disease
with a,median OS ranging from 9.9–14.6 months (2-4). In the comprehensive treatment group, the longest
OS was 259 weeks (60.5 months) with an average OS for the group of 100 weeks (23 months). The
median OS was 81 weeks (19 months), with 1-year and 2-year survival of 90% and 42%, respectively,
with a 3-year OS of 15.2%. Compared to historical controls, the patients in this study on average had a
longer 2-year and OS. We also identified statistically significant differences several of the treatment
groups in this study. Although there were no statistical significant differences noted among the
comprehensive group and chemotherapy group or radiotherapy group, patients with a longer OS were all
within the comprehensive group.
Several lines of research have shown that miRNAs can function as either oncogenes or in some cases,
tumor suppressor genes. Bioinformatics has helped us to better understand the role of miRNA’s,
predicting that the human genome may encode over 1,000 miRNAs, many of these with the capacity to
regulate nearly 1/3 of all known messenger RNA’s (mRNAs) (5, 6). The examination of miRNA’s in
GBM are still quite premature and is rapidly evolving, but GBM-specific miRNA expression profiles
have provided a new tool in our current understanding and treatment of this aggressive disease. Ciafre et
al. examined the global expression levels of 245 miRNAs in 10 GBM cell lines by microarray (7). They
identified a set of brain-enriched miRNAs, specifically miR-128, miR-181a, miR-181b, and miR-181c,
that were down-regulated in GBM, further indicating that they may act as tumor suppressor genes. They
also showed that the expression of 9 miRNAs, especially of miR-221, were significantly up-regulated in
malignant glioma. A second study has shown that miR-221 and miR-222 are both up-regulated during
9
serum stimulation of quiescent human cells (8).
In the present study, we found that expression of miR-221 was correlated with hemorrhage in
glioblastoma. Jennifer et al. found markedly elevated miR-21 levels in human glioblastoma tumor tissues,
early-passage glioblastoma cultures, and in 6 established glioblastoma cell lines (A172, U87, U373,
LN229, LN428, and LN308). They further implicate miR-21 in the activation of MMP and promotion of
tumor invasion by inhibiting the expression of RECK and TIMP3 (9). Further studies showed that
specific inhibition of miR-21 with antisense oligonucleotides leads to elevated levels of RECK and
TIMP3 and therefore reduces MMP activities in vitro and in a human model of gliomas in nude mice (10).
Aberrant miRNA expression profiles may also have potential diagnostic and prognostic value in different
malignancies. Takamizawa et al. first reported significantly shorter survival after potentially curative
resection of patients with non-small cell lung cancer with reduced miR-let-7 expression (11).
In the current study, we found that expression of miR-let-7 (including miR-let-7f, miR-let-7i,
miR-let-7f-1*) were down-regulated in the hemorrhage group of GBM patients. Furthermore, it has been
shown that high miR-21 expression is associated with poor survival in patients with colon and breast
cancer (12, 13). Additonally, others have shown that miRNAs are present in human plasma in a
remarkably stable form, therefore miRNAs in serum or plasma have the potential to serve as diagnostic
markers for GBM (14). Studies of miRNAs in GBM may further lead to more accurate tumor
classification, drug and biomarker discovery, drug efficacy testing, and personalized treatment (15-18).
However, we were unable to find a significant relationship between miRNA expression and survival of
GBM patients, likely due to the small sample size of this study.
MRI has been used in the clinical diagnosis of GBM’s to serve as a guide for clinical treatment for more
than 20 years. It is the most commonly used means of non-invasive diagnostic imaging. Our previous
10
studies have shown that MRI features were correlated with molecular pathology and survival of GBM
patients (19, 20). In the present study, we reported for the first time the relationships between MRI
imaging features and expression of miRNAs in GBM. Analyzing the MRI features may often be
correlative with the miRNA expression profile, possibly further opening the genetic features of GBM and
therefore, improved treatment options.
References
1. Stupp R, Mason WP, Van Den Bent MJ, Weller M, Fisher B, Taphoorn MJ, et al: Radiotherapy plus
concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 352:987-996, 2005. PMID:
15758009
2. Laws ER, Parney IF, Huang W, Anderson F, Morris AM, Asher A, et al: Survival following surgery
and prognostic factors for recently diagnosed malignant glioma: data from the Glioma Outcomes
Project. J Neurosurg 99:467-73, 2003. PMID: 12959431
3. Stupp R, Hegi ME, Mason WP, van den Bent MJ, Taphoorn MJ, Janzer RC, et al: Effects of
radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in
glioblastoma in a randomised phase III study: 5-year analysis of the EORTC-NCIC trial. Lancet Oncol
10:459-66, 2009. PMID: 19269895
4. Helseth R, Helseth E, Johannesen TB, Langberg CW, Lote K, Rønning P, et al: Overall survival,
prognostic factors, and repeated surgery in a consecutive series of 516 patients with glioblastoma
multiforme. Acta Neurol Scand 122: 159-167, 2010. PMID: 20298491
5. Sam G J, Harpreet K S, Stijn V D, Enright AJ: miRBase: tools for microRNA genomics. Nucleic Acids
11
Res 36: 154-158, 2008. PMID: 17991681
6. Lim L P, Glasner M E, Yekta S, Burge CB, Bartel DP: Vertebrate microRNA genes. Science
299(5612): 1540, 2003. PMID: 12624257
7. Ciafre S A, Galardi S, Mangiola A, , Ferracin M, Liu CG, Sabatino G, et al: Extensive modulation of a
set of microRNAs in primary glioblastoma. Biochem Biophys Res 334(4): 1351-1358, 2005. PMID:
16039986
8. Ricardo M, Sayyed K Z, Liu C G, Stein JL, van Wijnen AJ, Croce CM, et al: MicroRNAs 221 and 222
bypass quiescence and compromise cell survival. Cancer Res 68(8): 2773-2780, 2008. PMID:
18413744
9. Chan J A, Krichevsky A M, Kosik K S: MicroRNA-21 is an antiapoptotic factor in human
glioblastoma cells. Cancer Res 65(14): 6029-6033, 2005. PMID: 16024602
10. Gabriely G, Wurdinger T, Kesari S, Esau CC, Burchard J, Linsley PS,et al: MicroRNA 21 promotes
glioma invasion by targeting matrix metalloproteinase regulators. Mol Cell Biol 28(17): 5369-5380,
2008. PMID: 18591254
11. Takamizawa J, Konishi H, Yanagisawa K, Tomida S, Osada H, Endoh H, et al: Reduced expression of
the let-7 microRNAs in human lung cancers in association with shortened postoperative survival.
Cancer Res 64(11): 3753-3756, 2004. PMID: 15172979
12. Schetter A J, Leung S Y, Sohn J J, Zanetti KA, Bowman ED, Yanaihara N, et al: MicroRNA
expression profiles associated with prognosis and therapeutic outcome in colon adenocarcinoma.
JAMA 299: 425-436, 2008. PMID: 18230780
13. Yan L X, Huang X F, Shao Q, Huang MY, Deng L, Wu QL, et al: MicroRNA miR-21 overexpression
in human breast cancer is associated with advanced clinical stage, lymph node metastasis and patient
12
poor prognosis. RNA 14(11): 2348-2360, 2008. PMID: 18812439
14. Mitchell P S, Parkin R K, Kroh E M, Fritz BR, Wyman SK, Pogosova-Agadjanyan EL, et al:
Circulating microRNAs as stable blood-based markers for cancer detection. Proc Natl Acad Sci USA
105: 10513-10518, 2008. PMID: 18663219
15. Shai R M, Reichardt J K, Chen TC: Pharmacogenomics of brain cancer and personalized medicine in
malignant gliomas. Future Oncol 4(4): 525-534, 2008. PMID: 18684063
16. Li G, Zhang Z, Tu Y, Jin T, Liang H, Cui G, et al: Correlation of microRNA-372 upregulation with
poor prognosis in human glioma. Diagn Pathol 8;8:1, 2013. PMID: 23298385
17. Koshkin PA, Chistiakov DA, Chekhonin VP: Role of microRNAs in Mechanisms of Glioblastoma
Resistance to Radio- and Chemotherapy. Biochemistry 78(4): 325-334, 2013. PMID: 23590435
18. Nikaki A, Piperi C, Papavassiliou AG. Role of microRNAs in gliomagenesis: targeting miRNAs in
glioblastoma multiforme therapy. Expert Opin Investig Drugs. 21(10):1475-88, 2012. PMID:
22809292
19. Li WB, Tang K, Zhang W, Yan W, You G, Li SW, et al: Relationship between magnetic resonance
imaging and molecular pathology in patients with glioblastoma multiforme. Chin Med J (Engl) 124
(17): 2589-2592, 2011. PMID: 22040408
20. Li WB, Tang K, Chen Q, Li S, Qiu XG, Li SW, et al: MRI manifestions correlate with survival of
glioblastoma multiforme patients. Cancer Bio Med 9: 120-123, 2012. PMID: 23691466
13
Tab 1. Treatment options and Survival
treatment
cases
AS* (weeks)
MS† (weeks)
Surg+Radio+Chemo
33
100.3
81.6
Surg+Radio
9
64.4
48.3
Surg+Chemo
3
102
106.6
Surg+only
4
39
36.1
*average survival, † median survival
Figure 1. Treatment options and patient survival
14
1) Figure 2. Five imaging features of case NO. 31: 1) Contrast tumor enhanced /necrosis ratio≥1; 2)
Contrast tumor enhanced/T2 ratio＜ 1; 3) Single lesion; 4) No hemorrhage; 5) Larger than median
necrotic volum
Figure 3. Relationship between miRNA expression and MRI features. (A) Relationship between miRNA
expression and contrast tumor enhanced /necrosis ratio. “1” on behalf of the ratio ≥1, “2” on behalf of the
ratio＜ 1.
15
(B). Relationship between miRNA expression and contrast tumor enhanced/T2 ratio. “1” on behalf of the
ratio ≥1, “2” on behalf of the ratio＜ 1.
(C). Relationship between miRNA expression and multiple lesions. “1” for multiple lesions, “2” for
single lesion.
16
(D). Relationship between miRNA expression and Hemorrhage. “1” for cases with hemorrhage, “2” for
cases without hemorrhage.
17
(E). Relationship between miRNA expression and Necrotic volume. “1” for cases smaller than median
necrotic volume, “2” for cases larger than median necrotic volume. The expression value for each miRNA
is indicated by color intensity, with red representing high expression and green representing low
expression.
18