Genomics, Transcriptomics ISSN 1993-6842 (on-line); ISSN 0233-7657 (print)

Biopolymers and Cell. 2018. Vol. 34. N 2. P 85–96

and Proteomics doi: http://dx.doi.org/10.7124/bc.000973

UDC 577.218+616.65

A role of expression level of reference and investigated genes

in prostate tumors for qPCR analysis
G. V. Gerashchenko1, E. O. Stakhovsky2, L. I. Chashchina1,
O. P. Gryzodub3, V. I. Kashuba1
1 Instituteof Molecular Biology and Genetics, NAS of Ukraine
150, Akademika Zabolotnoho Str., Kyiv, Ukraine, 03680
2 National Cancer Institute,

33/43, Lomonosova Str., Kyiv, Ukraine, 03022

3 State Institution «Institute of Urology of NAMS of Ukraine»

9-a, Yu. Kotsubyns'koho Str., Kyiv, Ukraine, 04053

g.v.gerashchenko@imbg.org.ua

Aim. To determine the expression profiles of a set of cancer-associated genes in prostate tumors,
using various normalization protocols (with 1, 2 and 4 reference genes) and to optimize a
combination of reference genes to calculate the relative expression (RE) of the investigated
genes in prostate cancers. Methods. Relative expression level of 23 genes was analyzed by
quantitative PCR (qPCR) in 37 prostate cancer tissues (T) with different Gleason scores (GL)
and at various stages and compared with 37 corresponding normal prostate tissue (CNT)
samples and with 20 samples of prostate adenomas. Results. Theoretical calculations of the
RE deviation showed no influence of the normalization protocols on the results for both the
reference and the investigated genes. The experimental data that were calculated using a 2-ΔΔCt
showed statistically significant differences in the expression of 17 out of 23 investigated genes,
when the paired T/CNT were compared. RE values calculated using the 2-ΔCt method showed
a high similarity of statistical data in all reference gene groups for tumor-CNT-adenoma groups
(>  82 %). Data grouping by a cancer stage showed 69 %, and by the GL score – 64.5 % of
the data overlapping. Conclusions. All three types of normalization protocols, as expected,
can be used for RE normalization in prostate tumor samples. The usage of either the 2-ΔCt or
2-ΔΔCt models showed no difference in the calculated RE levels for the studied reference genes.
The most important factor was the constitutive expression of the reference genes. Moreover,
the expression levels of the investigated genes, changes in RE values, number of samples in
groups and heterogeneity of gene expression are important parameters for the selection of the
threshold in expression level differences between groups for a reliable data interpretation.
K e y w o r d s: prostate tumors, relative expression, reference genes validation, expression
level, genes expressed at low levels.

© 2018 G. V. Gerashchenko et al.; Published by the Institute of Molecular Biology and Genetics, NAS of Ukraine on behalf
of Biopolymers and Cell. This is an Open Access article distributed under the terms of the Creative Commons Attribution
License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any
medium, provided the original work is properly cited

85
G. V. Gerashchenko, E. O. Stakhovsky, L. I. Chashchina
G. V. Gerashchenko, E. O. Stakhovsky, L. I. Chashchina et al.

Introduction Materials and Methods

A quantitative real-time PCR (qPCR) is a Collection of prostate tissues. The samples
widely used method to assess the gene expres- of cancer tissues and conventional normal tis-
sion in a basic and clinical research [1–3]. sues (CNT, taken from the other prostate lobe
Relative quantification requires the use of a outside of the tumor) were frozen in the liquid
reference gene (or a few reference genes) for nitrogen immediately after surgical resection
normalization of the gene expression. Usually, at the National Cancer Institute (Kyiv,
several housekeeping genes are used for this Ukraine). Benign prostate tumors (prostate
purpose [4]. The main quality of the reference adenoma samples) were collected at the
gene is the constitutive expression under vari­ Institute of Urology (Kyiv, Ukraine) after rad-
ous experimental conditions, and also in path- ical prostatectomy and frozen as described
ological processes and in specific tissues. above. All samples were collected in accor-
It is known that upon carcinogenesis the dance with the Declaration of Helsinki and the
expression of many genes, including some guidelines, issued by an Ethic Committee of
housekeeping genes, altered. This creates prob- the Institute of Urology of National Academy
lems when searching for the reference genes of Medical Sciences of Ukraine and of the
for qPCR normalization, as there are no refer- National Cancer Institute of National Academy
ence genes universal for all types of tumors [5]. of Sciences of Ukraine (NASU), and the Ethic
Such genes must be validated, according to a Committee of the Institute of Molecular biol-
tumor type and experimental conditions. ogy and genetics of NASU. Experimental stud-
Moreover, the features of their expression ies were conducted, using 37 prostate adeno-
should also be considered. This is especially carcinoma samples of different Gleason scores
important for the low-expressed genes, which and at various stages; 37 corresponding con-
are often the subject of research, due to the ventional normal tissue (CNT) samples; and
peculiarity of their functions in physiological 20 samples of adenomas [10, 11]. The tumor
and pathological processes. samples were characterized, according to the
The validation of the reference genes for International System of Classification of
prostate tumors, lymph nodes from patients Tumors, based on the tumor-node-metastasis
with prostate cancer and also prostate cancer (TNM) and the World Health Organization
cell lines resulted in the creation of a set of (WHO) criteria. The clinical characteristics of
genes, namely TBP, HPRT1, ALAS1, TUBA1B, the tumors were described earlier [10, 11].
GAPDH and B2M that are expressed constitu- Total RNA isolation and cDNA synthesis.
tively in prostate cancer and normal tissues, 50–70 mg of frozen prostate tissues were ho-
making them suitable for qPCR normaliza- mogenized to a powder in liquid nitrogen.
tion [6–9]. Total RNA was isolated, using TRI-reagent
In the present work, we used four reference (Sigma-Aldrich, USA). The concentration of
genes (TBP, HPRT1, ALAS1, TUBA1B) in dif- the isolated total RNA was assessed, using a
ferent combinations – from 1 to 4 genes, to spectrophotometer (NanoDrop Technologies
compare the qPCR results after normalization. Inc. USA). The quality of RNA was deter-

86
 A role of expression level of reference and investigated genes in prostate tumors for qPCR analysis

mined by electrophoresis in a 1 % agarose gel experimental groups. The differences between
by band intensity of 28S and 18S rRNA experimental groups (adenocarcinomas, CNT
(28S/18S ratio). 1 µg of the total RNA was and adenomas) were determined by Kruskal-
treated with RNase-free DNase I (Thermo Wallis test with following tests for multiple
Fisher Scientific, USA); cDNA was synthe- comparisons. The Dunn-Bonferoni post-hoc
tized, using RevertAid H-Minus M-MuLV test was performed to determine RE differ-
Reverse Transcriptase (Thermo Fisher ences between pairs of prostate samples under
Scientific, USA). multiple gene comparisons [13]. The Benja­
Quantitative PCR (qPCR). Relative gene mini-Hochberg procedure was used to adjust
expression (RE) levels of 23 genes were as- a false discovery rate (FDR) set at 0.10–0.25,
sessed, using the Bio-Rad CFX96 Real-Time when multiple comparisons were per-
PCR Detection System (USA) with Maxima formed [14].
SYBR Green Master mix (Thermo Fisher
Scientific, USA). The qPCR cycling conditions Results
were as follows: 95°C×10´, (95°C×15´´, 60°C RE of 23 genes, representing markers of can-
×30´´, 72°C×30´´ for 40 cycles). Primers were cer-associated fibroblasts (CAF) (the CAF
selected with the help of a “qPrimerDepot – gene group), tumor-associated macrophages
A quantitative real time PCR primer database” (TAM) (the TAM gene group) and inflamma-
(http://primerdepot.nci.nih.gov) and Primer- tion-associated genes (the INF gene group)
BLAST (https://www.ncbi.nlm.nih.gov/tools/ have been determined. Genes were divided
primer-blast/). also by RE level into three groups: showing a
Four reference genes – TBP, HPRT1, high expression (Ct  < 20 cycles), the moder-
ALAS1 and TUBA1B – were used for nor­mali­ ate expression (Ct  = 20–29 cycles) and the
za­tion of RE levels [4, 7] in different combina- low expression (Ct > 29 cycles).
tions: 1 reference gene (1 ref) – TBP, 2 refer- The reference genes ALAS1 and TUBA1B
ence genes (2 ref) – TBP and HPRT and 4 refe­ showed a high level of expression, whereas
rence genes (4 ref) – TBP, HPRT, ALAS1 and TBP and HPRT were expressed at a moderate
TUBA1B. RE levels were calculated, using two level. TBP demonstrated the lowest expression
common methods (2-ΔCt and 2-ΔΔCt) described level among the references. Only three genes
earlier [10–12]. (ACTA2, MSMB and HLA-G) out of 23 studied
Statistical analysis. The Kolmogorov- demonstrated high RE levels. 10 genes were
Smirnov test was used to analyze the normal- expressed at a moderate level and 10 – at low
ity of distribution. The RE levels in prostate level of expression.
adenocarcinoma and paired CNT were com- A theoretical calculation of a hypothetical
pared, using the Wilcoxon Matched Pairs test. deviation of the RE of reference genes ex-
RE fold differences in 2-ΔΔCt model were con- pressed at high and low levels was developed,
sidered significant, when expression changed taking 0.5 Ct as a hypothetical error. RE of the
more or less, than 2 folds. The Fisher exact studied genes was calculated, using the 2-ΔCt
test was used to monitor differences between method (Table 1).

87
G. V. Gerashchenko, E. O. Stakhovsky, L. I. Chashchina et al.

Table 1. Calculation of changes in RE of

investigated (Inv) and reference (Ref) genes, reference groups; 7 of these genes were ex-
expressed at different levels (high (h), pressed at high and moderate levels. Diver­gen­
moderate (m) and low (l)), when the hypothetical ces of RE were observed for 7 genes in
error was 0.5 Ct (e). 10 comparative groups, 6 of which showed
Genes Сt Ref Inv1 low Inv2 high low expression. Thus, the threshold value of
Ct Inv 31 17 matching differences for highly and moder-
RE /Ref h 15 0.000015 0.250 ately expressed genes was set at 25–30 %
RE /Ref he 15.5 0.000022 0.354 (10–11 samples out of 37), whereas for low
RE /Ref m 25 0.016 256.000 expressed genes the value should be no less,
RE /Ref me 25.5 0.022 362.039 than 35 % (more than 13 samples out of 37),
RE /Ref l 32 2.000 32768.000 to avoid possible expression deviations of the
RE /Ref le 32.5 2.828 46340.950 reference genes and minimize the influence of
RE fold changes qPCR reaction inhibitors for PCR analysis of
1.414 1.414
Ref he/h low-expressed genes.
RE fold changes RE values were investigated using the 2-ΔCt
1.414 1.414
Ref me/m method for three sets of the samples:
RE fold changes 1. The TNA set – 3 total sample groups:
1.414 1.414
Ref le/l
Adenocarcinomas (T, n = 37), CNT (N,
Notes: Ref h – high expression of the reference gene, Ref he – n  = 37) and adenomas (A, n = 20);
Ref h with 0.5 Ct error, Ref m – moderate expression of the
reference gene, Ref me – Ref m with 0.5 Ct error, Ref l – low 2. The cancer stage set – 5 groups of sam-
expression of the reference gene, Ref le – Ref with 0.5 Ct error. ples at the various tumor stages: adeno-
carcinomas of stage 1–2 (T1-2, n = 28),
Our calculations showed that the RE devia- adenocarcinomas of stage 3-4 (T3-4,
tion with an error of 0.5Ct for reference genes n = 9), CNT of stage 1–2 (N1-2, n = 28),
was the same (1.414) for all analyzed genes, CNT of stage 3-4 (N3-4, n = 9), adeno-
regardless expression levels of the reference mas (A, n = 20);
genes (Table 1). This data indicates the impor- 3. A set divided by the GL – 7 groups: ad-
tance of the constitutive expression of the enocarcinomas GL < 7 (T < 7, n = 11),
reference gene when comparing RE of the adenocarcinomas GL = 7 (T = 7, n = 9),
analysed and the reference genes. adenocarcinomas GL > 7 (T > 7, n = 17),
The experimental data calculated, using the CNT GL < 7 (N < 7, n = 11), CNT
2 -ΔΔCt model, showed statistical significant dif- GL = 7 (N = 7, n = 9), CNT GL > 7
ferences between the paired T/CNT in one (N > 7, n = 17), adenomas (A, n = 20).
reference group (17 out of 23 investigated Fold changes in RE for genes with statisti-
genes) (Table 2). cally significant differences between sample
A complete match of statistical data was groups (with normalization by 3 reference
observed for all three reference groups for 16 types) and p-values are shown in Table 3A-C.
out of 23 genes. Eleven genes beyond 16 A high similarity was found for all three
showed significant changes of RE in all three reference groups with different types of group-

88
 A role of expression level of reference and investigated genes in prostate tumors for qPCR analysis

Table 2. Numbers of adenocarcinoma samples with changes in RE (2-ddCt model), normalized with the use
of 1, 2 or 4 reference genes
1 reference gene 2 reference genes 4 reference genes
Gene group Genes
> 2.01 < 0.49 > 2.01 < 0.49 > 2.01 < 0.49
ACTA2 9 4 9 3 7& 3
CXCL14 19 3 19 4 17 4
CTGF 12 0 12 2 11 1
HIF1A 5 0 3 0 3 0
CAF
S100A4 3 6 3 5 2 5
THY1 9 3 9 2 7& 1
CXCL12 4 7 5 6 4 6
FAP 12 0 11 1 13 1
CD68 8 4 6 3 5 6
CD163 14 5 12 6 11 5
CCR4 8 9 6 8 5 10
TAM
CCL17 8 6 9 8 10 8
CCL22 10 8 6 7 6 6
NOS2A 7 16 6 13 4 15
MSMB 6 10 5 10 6 9
HLA-G 2 3 3 4 4 2
IRF1_T1 3 6 4 7 3 6
IL1R1_T17 1 11 1 8 1 8
INF CIAS 4 6 4 6 3 5
CTLA4 5 11 8 12 6 7
IL1RL1 2 11 3 8 3 7
IL2RA 8 8 8 7 7 6
KLRK 8 10 8 9 7 4
Notes: statistically significant differences between T/CNT, calculated, using the Fisher exact test with correction on
multiple comparisons, FDR = 0.2 are shown in bold (black and red); in black (bold) – statistically significant differ-
ences, that have a complete match for all groups of reference genes; in red (bold) –divergences of statistical results
between reference groups; & – significant differences with FDR = 0.2; green boxes – highly expressed genes; white
boxes – moderately expressed genes; pink boxes – low expressed genes.

ing of analyzed samples (> 82 % – TNA group, in 3 sample groups of TNA (17.65 %) with RE
69 % – Cancer stage group, 64.5 % – GL fold changes less than 1.7 times.
group).10 out of 23 genes in the TNA sample Another grouping type (by tumor stages)
groups showed significant differences in RE (Table 3B) demonstrated significant differ-
in 17 pairs (Table 3A). No similarity was ob- ences in RE for 14 genes in 45 pairs of sample
served for the 3 reference group normalization groups. No similarity in the 3 reference group

89
G. V. Gerashchenko, E. O. Stakhovsky, L. I. Chashchina et al.

Table 3. Differences in the fold changes and p-values of RE differences between pairs of groups,
calculated by the Dunn-Bonferroni post hoc method for multiple comparisons, normalized to various
reference genes in prostate tumors, grouped by TNA (A), stages (B), Gleason score (C).
A.
Gene Fold changes p-value
Gene Pairs of groups
group 1 ref 2 ref 4 ref 1 ref 2 ref 4 ref
T/A 7.80 6.57 6.03 0.000 0.000 0.000
CXCL14 T/N 3.26 2.32 2.27 0.011 0.019 0.025
N/A 2.39 2.83 2.66 0.005 0.002 0.003
T/A 2.06 2.43 2.51 0.001 0.000 0.001
CTGF
CAF T/N 1.58 1.51 1.50 0.036 0.041 0.055&
THY1 T/A 1.87 1.79 1.71 0.017 0.006 0.011
T/A 0.35 0.39 0.45 0.000 0.000 0.000
CXCL12
N/A 0.38 0.40 0.41 0.001 0.000 0.000
FAP T/A 1.63 1.78 1.91 0.049 0.024 0.015
CD163 T/A 2.14 1.68 1.39 0.045 0.129 0.250
T/A 0.57 0.56 0.54 0.037 0.009 0.002
CCR4
TAM N/A 0.78 0.71 0.70 0.149 0.054& 0.040
T/A 2.12 1.99 2.09 0.004 0.009 0.015
CCL17
N/A 1.77 1.71 1.51 0.016 0.038 0.065
IL1R1 T/A 0.69 0.52 0.51 0.031 0.023 0.005
INF T/A 2.40 2.13 2.16 0.043 0.023 0.016
CTLA4
N/A 2.72 3.12 2.61 0.001 0.002 0.003

B.
Gene Fold changes p-value
Gene Pairs of groups
group 1 ref 2 ref 4 ref 1 ref 2 ref 4 ref
T1-2/A 6.48 6.4 5.84 0 0 0
T3-4/A 17.82 7.66 6.98 0 0 0
CXCL14
N3-4/A 6.09 6.27 5.55 0.008 0.004 0.004
T1-2/N1-2 3.56 2.75 2.66 0.036 0.06 0.089
T1-2/A 2.08 2.48 2.33 0.001 0.001 0.005
CAF CTGF
T1-2/N3-4 3.31 2.22 2.06 0.001 0.006 0.028
T1-2/T3-4 2.47 2.83 1.92 0.001 0.003 0.008
T1-2/N3-4 2.65 3.03 2.14 0 0.001 0.001
HIF1A
T3-4/A 0.43 0.4 0.49 0.012 0.012 0.01
N1-2/N3-4 2.02 2.49 2.03 0.03 0.026 0.012

90
 A role of expression level of reference and investigated genes in prostate tumors for qPCR analysis

continued Table 3B
Gene Fold changes p-value
Gene Pairs of groups
group 1 ref 2 ref 4 ref 1 ref 2 ref 4 ref
HIF1A N3-4/A 0.4 0.38 0.44 0.005 0.004 0.001
THY1 T1-2/A 1.69 2.28 1.8 0.026 0.013 0.041
T1-2/A 0.46 0.41 0.41 0.002 0.001 0
CAF T3-4/A 0.41 0.41 0.47 0.008 0.028 0.034
CXCL12
N1-2/A 0.54 0.45 0.43 0.022 0.011 0.001
N3-4/A 0.41 0.34 0.37 0.004 0.007 0.002
FAP T1-2/A 1.32 1.85 1.94 0.051& 0.043 0.057&

T1-2/T3-4 4.51 2.96 2.75 0.056& 0.021 0.082

CD68 T1-2/N3-4 4.01 3.96 1.34 0.04 0.033 1
T3-4/N1-2 0.22 0.34 0.34 0.111 0.048 0.166
T1-2/T3-4 0.07 0.08 0.07 0.045 0.042 0.016
T1-2/N3-4 0.1 0.11 0.09 0.114 0.083 0.05
T3-4/A 17.8 17.13 17.51 0.002 0.005 0.005
CD163
T3-4/N1-2 12.78 15.26 17.81 0.006 0.005 0.003
N1-2/N3-4 0.11 0.1 0.07 0.019 0.011 0.01
N3-4/A 12.84 11.42 13.46 0.006 0.011 0.017
T1-2/T3-4 2.55 2.08 1.99 0.06 0.049 0.062
T3-4/A 0.26 0.34 0.34 0.002 0 0
CCR4
TAM T3-4/N1-2 0.31 0.45 0.48 0.04 0.027 0.037
N3-4/A 0.55 0.56 0.59 0.105 0.034 0.203
T1-2/N3-4 0.2 0.16 0.12 0.023 0.014 0.009
T3-4/A 8.49 8.59 9.52 0.006 0.005 0.004
CCL17 T3-4/N1-2 7.12 7.41 8.82 0.113 0.05 0.023
N1-2/N3-4 0.11 0.1 0.07 0.001 0.001 0
N3-4/A 10.41 11.05 14.72 0 0 0
T1-2/T3-4 3 3.15 2.7 0.004 0.006 0.044
CCL22 T1-2/A 1.93 2.19 2.32 0.012 0.025 0.039
T3-4/N1-2 0.41 0.39 0.47 0.015 0.032 0.156
T3-4/N1-2 0.16 0.23 0.16 0.014 0.013 0.008
NOS2A
N1-2/N3-4 5.12 5.73 4.63 0.039 0.047 0.125

91
G. V. Gerashchenko, E. O. Stakhovsky, L. I. Chashchina et al.

continued Table 3B
Gene Fold changes p-value
Gene Pairs of groups
group 1 ref 2 ref 4 ref 1 ref 2 ref 4 ref
T3-4/A 0.32 0.26 0.27 0.086 0.039 0.007
IL1R1
N3-4/A 0.51 0.54 0.37 0.237 0.178 0.014
INF T1-2/A 2.33 2.32 2.25 0.127 0.047 0.047
CTLA4 N1-2/A 2.65 2.85 2.49 0.021 0.016 0.022
N3-4/A 3.78 3.81 3.09 0.028 0.077 0.113

C.

 Gene Fold changes p-value

Gene Pairs of groups
group 1 ref 2 ref 4 ref 1 ref 2 ref 4 ref

T < 7/A 4.47 4.86 3.74 0.022 0.016 0.058

T = 7/A 7.91 6.57 6.27 0 0 0
CXCL14
T > 7/A 8.61 8.07 7.56 0 0 0
N > 7/A 4.22 4.8 4.43 0.017 0.007 0.008
T < 7/A 3.12 4.07 3.13 0.041 0.022 0.098
CTGF T = 7/A 3.19 2.87 2.57 0.005 0.011 0.027
T = 7/N > 7 2.57 2 1.75 0.019 0.099 0.176
CAF T = 7/T > 7 2.45 2.36 2.14 0.001 0.002 0.004
HIF1A
T = 7N > 7 1.97 2.03 2.09 0.001 0.004 0.006
THY1 T < 7/A 1.96 2.59 1.78 0.078 0.026 0.098
T < 7/A 0.56 0.56 0.48 0.222 0.176 0.009
T = 7/A 0.31 0.38 0.29 0.118 0.073 0.003
CXCL12 T > 7/A 0.33 0.37 0.43 0.001 0.001 0.001
N = 7/A 0.32 0.36 0.35 0.15 0.146 0.035
N > 7/A 0.34 0.35 0.37 0.004 0.003 0.001
T < 7/T > 7 3 3.24 3.29 0.027 0.02 0.025
CCR4 T > 7/A 0.42 0.39 0.34 0.005 0.001 0
N > 7/A 0.55 0.57 0.66 0.103 0.033 0.04
TAM
T > 7/A 3.18 2.93 3.39 0.002 0.003 0.004
CCL17
N > 7/A 2.98 3.12 2.79 0.002 0.005 0.01
CCL22 T < 7/A 2.35 2.46 2.58 0.037 0.045 0.051

92
 A role of expression level of reference and investigated genes in prostate tumors for qPCR analysis

continued Table 3C

 Gene Fold changes p-value

Gene Pairs of groups
group 1 ref 2 ref 4 ref 1 ref 2 ref 4 ref

T > 7/N = 7 0.1 0.13 0.15 0.004 0.001 0.002

TAM NOS2A
N = 7/N > 7 7.42 6.42 4.44 0.052& 0.027 0.045
T > 7/A 0.39 0.41 0.37 0.059 0.027 0.006
IL1R1
T > 7/N = 7 0.26 0.23 0.27 0.015 0.018 0.032
T < 7/A 4.1 4.8 4.5 0.012 0.002 0.002
T < 7/T > 7 2.25 2.81 2.68 0.125 0.031 0.031
INF CTLA4
T > 7/N = 7 0.32 0.44 0.4 0.042 0.065 0.065
N = 7/A 5.62 3.92 4.18 0.004 0.006 0.006
T < 7/A 2.18 2.44 1.89 0.02 0.003 0.008
IL2RA
T < 7/T > 7 0.2 0.12 0.09 0.075 0.006 0.061
Notes: – significant differences with FDR = 0.2; red p-value; – p < 0.05 is considered as statistically significant;
&

p-value 0.000 – p < 0.001; white boxes – moderately expressed genes; pink boxes – low expressed genes

normalization was observed for 14 pairs of primers, non-specific products and loss in the
sample groups (31 %) with RE changes less activity of Tag-polymerase [15–17]. All these
than 3–4 folds. factors inadvertently impact the efficiency of
Prostate cancers grouped by GL (Table 3C) PCR, thus, resulting in erroneous RE levels.
showed significant changes in RE for 12 genes This, in turn, leads to difficulties in assessment
out of 23, for 31 pairs of samples. No similar- of the low expressed genes, regardless of the
ity in the 3 reference group normalization was optimization of qPCR conditions. Especially,
observed for 11 sample groups (35.5 %) with this is important if the reference genes are
changes in RE less than 5 fold. expressed at low levels. So, the low expressed
genes should not be chosen as the reference.
Discussion Other parameters that impact RE are the va­
Performed hypothetical calculations indicate lues of fold changes and a proportion of the
that the expression of both, reference and ana- samples where the expression of a certain gene
lyzed genes does not influence the deviation changed significantly. High heterogeneity of
(variation) in obtained RE, if the 2-ΔCt method gene expression in prostate cancer samples [18]
was used. This confirms the need for constitu- makes this impact more complicated. Note­
tive expression of reference genes in all ana- worthy, in the cases, when fold change is high,
lyzed samples [5, 6]. Some cautions concern the expression levels of the reference do not
the low expressed genes, for example, during influence the calculated values, as shown by our
PCR analysis the PCR inhibitors may increase. results and literature data [7, 13]. When we
By PCR inhibitors we mean formed dimers of compared the changes lower than 2-fold or in a

93
G. V. Gerashchenko, E. O. Stakhovsky, L. I. Chashchina et al.

proportion of samples below 30 % of all studied, levels of investigated and reference genes have
even if differences in RE were statistically sig- no impact regardless usage of the 2-ΔCt and
nificant, we could get both, false positive and 2-ΔΔCt models; the constitutive expression of
false negative results, namely differences could reference genes is the important parameter.
appear where they are not present, groups over- Thus, the values of expression of the analysed
lapped, etc. This impact became more evident, genes, as well as RE value changes, the num-
when the low expressing genes were analysed, ber of samples in groups and high heterogene-
using both methods, the 2-ΔCt and 2-ΔΔCt. ity of gene expression are important parame-
The next important factor of the statistical ters for choosing the threshold level differ-
analysis is the number of samples in a group ences between the groups of samples for reli-
[19]. This is supported by the data presented in able data interpretation.
this article. For example, the largest number of
samples in groups (20 to 37 grouped samples References
(TNA group)) produced the lowest proportion 1. Sanders R, Mason DJ, Foy CA, Huggett JF. Con-
of inconsistences of statistical results for all siderations for accurate gene expression measure-
reference groups. Additionally, this amount of ment by reverse transcription quantitative PCR when
samples in groups demonstrated the highest rate analyzing clinical samples. Anal Bioanal Chem.
of matching results (82 %) and the lowest 2014;406(26):6471–83.
2. Chen J, Zhao Z, Chen Y, Zhang J, Yan L, Zheng X,
threshold of fold changes (1.7 times) to observe
Liao M, Cao W. Development and application of a
the statistically significant differences between SYBR green real-time PCR for detection of the
the analysed groups for all of reference genes. emerging avian leukosis virus subgroup K. Poult
The type of grouping is no less important, Sci. 2018;97(7):2568–2574.
than the number of samples in groups. Obviously, 3. Shi W, Wang Y, Ren X, Gao S, Hua X, Guo M,
the gene expression pattern correlates with the Tang L, Xu Y, Ren T, Li Y, Liu M. EvaGreen-based
clinical and pathological characteristics, thus real-time PCR assay for sensitive detection of sal-
monid alphavirus. Mol Cell Probes. 2018;39:7–13.
providing the possibility to define the genes with 4. Sharan RN, Vaiphei ST, Nongrum S, Keppen J,
altered expression at a given stage of the disease Ksoo M. Consensus reference gene(s) for gene ex-
(HIF1A, CD68, CCL22, NOS2A1), or related to pression studies in human cancers: end of the tunnel
a specific GL score (HIF1A, CCL22, NOS2A, visible? Cell Oncol (Dordr). 2015;38(6):419–31.
IL2RA1). Noteworthy, in the TNA group, that 5. Dundas J, Ling M. Reference genes for measuring
contained tissues, collected at the different stag- mRNA expression. Theory Biosci. 2012;131(4):215–23.
6. Ohl F, Jung M, Xu C, Stephan C, Rabien A, Bur-
es of disease or tumors attributed with various
khardt M, Nitsche A, Kristiansen G, Loening SA,
GL score, the expression changes were nullified, Radonić A, Jung K. Gene expression studies in
due to a high RE deviation. prostate cancer tissue: which reference gene should
be selected for normalization? J Mol Med (Berl).
Conclusions 2005;83(12):1014–24.
All three types of reference genes can be used 7. Souza AF, Brum IS, Neto BS, Berger M, Branchi-
for normalization of RE for prostate tumor ni G. Reference gene for primary culture of prostate
cancer cells. Mol Biol Rep. 2013;40(4):2955–62.
samples. The differences in the expression

94
 A role of expression level of reference and investigated genes in prostate tumors for qPCR analysis

8. Zhao H, Ma TF, Lin J, Liu LL, Sun WJ, Guo LX, 19. Carlisle AJ, Prabhu VV, Elkahloun A, Hudson J,
Wang SQ, Otecko NO, Zhang YP. Identification of Trent JM, Linehan WM, Williams ED, Emmert-
valid reference genes for mRNA and microRNA Buck MR, Liotta LA, Munson PJ, Krizman DB.
normalisation in prostate cancer cell lines. Sci Rep. Development of a prostate cDNA microarray and
2018;8(1):1949. statistical gene expression analysis package. Mol
9. Tsaur I, Renninger M, Hennenlotter J, Opper- Carcinog. 2000;28(1):12-22.
mann E, Munz M, Kuehs U, Stenzl A, Schilling D.
Reliable housekeeping gene combination for quan- Роль рівнів експресії референсних
titative PCR of lymph nodes in patients with prostate та досліджуваних генів при раку передміхурової
cancer. Anticancer Res. 2013;33(12):5243–8. залози у кПЛР аналізі
10. Gerashchenko GV, Mankovska OS, Dmitriev AA, Г. В. Геращенко, Е. О. Стаховський, Л. І . Чащина,
Mevs LV, Rosenberg EE, Pikul MV, Marynychenko MV, О. П. Гризодуб, В. І. Кашуба
Gry-zodub OP, Stakhovsky EO, Kashuba VI. Epithe-
lial-mesenchymal transition related gene expression in Мета. Визначити профілі експресії пухлино-асоційова-
prostate tumours. Biopolym Cell. 2017; 33(5): 335–55. них генів у пухлинах передміхурової залози з викорис-
11. Mevs LV, Gerashchenko GV, Rosenberg EE, Pi- танням різних протоколів нормалізації (з одно-, дво- та
kul MV, Marynychenko MV, Gryzodub OP, Vozia­ чотириреференсними генами АБО з одним, двома та
nov SO, Stak-hovsky EO, Kashuba VI. Detection of чотирма референсними генами) та оптимізувати комбі-
prostate specific ETS fusion transcripts in cancer нації референсних генів для розрахунку відносної екс-
samples. Biopolym Cell. 2017; 33(4): 256–67. пресії (ВЕ) у раку передміхурової залози. Методи.
12. ivak KJ, Schmittgen TD. Analysis of relative gene Кількісною ПЛР (кПЛР) проаналізовано ВЕ 23 генів у
expression data using real-time quantitative PCR 37 зразках тканин передміхурової залози (Т) з різними
and the 2(-Delta Delta C(T)) Method. Methods. показниками Глісона та різними стадіями пухлин у по-
2001;25(4):402–8. рівнянні з 37 умовно-нормальними зразками тканини
13. Schmidt U, Fuessel S, Koch R, Baretton GB, передміхурової залози (УНТ) та 20 зразками аденом
Lohse A, Tomasetti S, Unversucht S, Froehner M, передміхурової залози. Результати. Теоретичні розра-
Wirth MP, Meye A. Quantitative multi-gene expres- хунки відхилення ВЕ не підтвердили впливу значень
sion profiling of primary prostate cancer. Prostate. рівнів експресії на цей параметр ані у ВЕ референсного
2006;66(14):1521–34. гена, ані у ВЕ досліджуваних генів. Експермиментальні
14. Benjamini Y, Hochberg Y. Controlling the false dis- дані, які були отримані, з використанням 2-ΔΔCt моделі,
covery rate: a practical and powerful approach to показали статистичні значущі відмінності у експресії
multiple testing. J R Stat Soc. 1995; 57(1): 289–300. 17 з 23 досліджуваних генів, при порівнянні парних T/
15. Opel KL, Chung D, McCord BR. A study of PCR УНТ. Показники ВЕ, розраховані з використанням мо-
inhibition mechanisms using real time PCR. J Fo- делі 2-ΔCt, показали високий рівень співпадіння статис-
rensic Sci. 2010;55(1):25–33. тичних даних у всіх групах референтних генів для груп
16. Pionzio AM, McCord BR. The effect of internal аденокарциноми-УНТ‑аденоми (понад 82 %). Слід за-
control sequence and length on the response to PCR значити, у 69 % випадків, а за показниками Глісона –
inhibition in real-time PCR quantitation. Forensic у 64,5 %. Висновки. Всі три типи референсних генів,
Sci Int Genet. 2014;9:55–60. як і було передбачено, можуть бути використані для
17. Schrader C, Schielke A, Ellerbroek L, Johne R. PCR нормалізації ВЕ у зразках пухлини передміхурової за-
inhibitors – occurrence, properties and removal. лози. Використання моделей 2-ΔCt або 2-ΔΔCt не має
J Appl Microbiol. 2012;113(5):1014–26. впливу на рівень ВЕ для референсних генів.
18. Cancer Genome Atlas Research Network. The mo- Найважливішим фактором була їх стабільна експресія.
lecular taxonomy of primary prostate cancer. Cell. Важливими параметрами для вибору порогу відміннос-
2015;163(4):1011–25. тей рівнів експресії між групами з метою правильної

95
G. V. Gerashchenko, E. O. Stakhovsky, L. I. Chashchina et al.

інтерпретації даних є рівні експресії досліджуваних Экспериментальные данные, полученые с использо-

генів, величина зміни значень ВЕ, розмір вибірки та ванием 2-ΔΔCt модели, показали статистически значи-
висока гетерогенність експресії. мые различия экспрессии у 17 из 23 исследованных
К л ю ч о в і с л о в а: пухлини передміхурової залози, генов при сравнении парных Т/УНТ. ОЭ, рассчитанные
відносна експресія генів, валідація референсних генів, с использованием модели 2-ΔCt, показали высокий
різні рівні експресії, низькоекспресовані гени. уровень совпадений статистических данных во всех
группах референсных генов для групп аденокарцино-
Роль уровней экспрессии референсных мы-УНТ-аденомы (более 82 %). Следует отметить, что
и исследуемых генов при раке простаты при разделении по стадиям совпадение статистических
в кПЦР анализе данных наблюдалось в 69 % случаев, а по показателю
Глисона – в 64,5 %. Выводы. Все три типа референс-
А. В. Геращенко, Э. А. Стаховский, Л. И. Чащина,
А. П. Гризодуб, В. И. Кашуба ных генов, как и ожидалось, могут быть использованы
для нормализации ОЭ в образцах опухолей простаты.
Цель. Определить профили экспрессии ряда опу- Использование моделей 2-ΔCt или 2-ΔΔСt не показало
холь-ассоциированных генов в опухолях предстатель- влияния различий в уровнях ОЭ для референсных
ной железы, используя различные протоколы норма- генов. Наиболее важным фактором была их стабильная
лизации (одним, двумя и четырьмя референсными экспрессия. При выборе порога уровней экспрессии
генами) и оптимизировать комбинацию этих генов для между группами с целью правильной интерпретации
рассчета относительной экспрессии (ОЭ) исследуемых данных важными параметрами являются уровни экс-
генов при раке предстательной железы. Методы.
прессии исследуемых генов, величина изменения зна-
Количественной ПЦР (кПЦР) було проанализировано
чений ОЭ, размер выборки и высокая гетерогенность
ОЭ 23 генов в 37 образцах рака предстательной желе-
экспрессии.
зы (Т) с различными показателем Глисона и на разных
стадиях, в сравнении с 37 условно-нормальными об- К л ю ч е в ы е с л о в а: опухоли предстательной
разцами ткани простаты (УНТ) и 20 образцами аденом железы, относительная экспрессия генов, валидация
предстательной железы. Результаты. Теоретические референсных генов, различные уровни экспрессии,
расчеты отклонения ОЭ не подтвердили влияния ве- низкоэкспрессированные гены.
личины уровней экспрессии на этот параметр ни в ОЭ
референсного гена, ни в ОЭ исследуемых генов. Received 01.02.2018

96