Expression pattern of genes associated with tumor microenvironment in prostate cancer

Gerashchenko G.V.1, Grygoruk O.V.2, Kononenko O.A.3, Gryzodub O.P.4, Stakhovsky E.O.3, Kashuba V.I.1

Summary. Aim: To assess relative expression (RE) levels of CAF-, TAM-specific, immune defense-associated genes in prostate tumors and to show correlation of RE with clinical, pathological and molecular characteristics, with the aim to define clinically significant specific alterations in a gene expression pattern. Methods: RE of 23 genes was analyzed by a quantitative polymerase chain reaction in 37 freshly frozen samples of prostate cancer tissues of a different Gleason score (GS) and at various tumor stages, compared with RE in 37 paired conventionally normal prostate tissue (CNT) samples and 20 samples of prostate adenomas. Results: Differences in RE were shown for 11 genes out of 23 studied, when tumor samples were compared with corresponding CNTs. 7 genes, namely ACTA2, CXCL14, CTGF, THY1, FAP, CD163, CCL17 were upregulated in tumors. 4 genes, namely CCR4, NOS2A, MSMB, IL1R1 were downregulated in tumors. 14 genes demonstrated different RE in TNA at different stages: CXCL12, CXCL14, CTGF, FAP, HIF1A, THY1, CCL17, CCL22, CCR4, CD68, CD163, NOS2A, CTLA4, IL1R1. RE changes of 9 genes — CXCL12, CXCL14, HIF1A, CCR4, CCL17, NOS2A, CTLA4, IL1R1, IL2RA — were found in tumors with different GS. Moreover, 9 genes showed differences in RE in TNA, dependently on the presence or absence of the TMPRSS2/ERG fusion and 7 genes showed differences in RE of groups with differential PTEN expression. Significant correlations were calculated between RE of 9 genes in adenocarcinomas and the stage, and GS; also, between RE of 2 genes and the fusion presence; and between RE of 4 genes and PTEN expression. Conclusions: Several gene expression patterns were identified that correlated with the GS, stage and molecular characteristics of tumors, i.e. presence of the TMPRSS2/ERG fusion and alterations in PTEN expression. These expression patterns can be used for molecular profiling of prostate tumors, with the aim to develop personalized medicine approaches. However, the proposed profiling requires a more detailed analysis and a larger cohort of patients with prostate tumor.

Submitted: July 05, 2018.
*Correspondence: E-mail: g.v.gerashchenko@imbg.org.ua
Abbreviations used: A — sample of prostate adenomas; CAF — cancer-associated fibroblasts; CNT/N — conventionally normal prostate tissue; CPC — clinical and pathological characteristics; FDR — false discovery rate; GS — Gleason score, IAG — immune-associated genes; PCR — polymerase chain reaction; RE — relative gene expression; st — stage; T – sample of prostate cancer, adenocarcinoma; TAM — tumor-associated macrophages; TNM — International System of Classification of Tumors, based on tumor-node-metastasis.

It is well known that tumor initiation and progression, including prostate cancer, depend on genetic and epigenetic changes, leading to transcriptomic and proteomic disorders in cells upon malignant transformation [1–3]. Also, many parameters are altered, that is hierarchically regulated at different levels (organism, organs and systems, tissues and cells) [4–7]. All these processes constitute a common phenomenon named as a tumor-host interaction. One of the disorders in the organism that contributes to the appearance of tumors is chronic inflammation [8, 9]. It could be of different etiologies, such as viral, bacterial, parasitic, chemical or radiation-induced and that of unidentified etio­logy [8, 10]. However, regardless of its origin, both, the epithelial cells of the organs and the cells of the stroma (fibroblasts, macrophages, endothelial cells, etc.) are involved. Later, all these cells become the tumor microenvironment [11], changing their properties from tumor-suppressive to tumor-stimulating and supporting [12, 13]. For example, fibroblasts are transformed into cancer-associated fibroblasts (CAF) [14, 15] and macrophages change their phenotype to tumor-associated macrophages (TAM) [16–19]. Tumors are filled with tumor-infiltrating leukocytes which suppress immune defense and reactions [20, 21].

We have to mention, that the tumor growth is affected by antimicrobial proteins, secreted by epithelial cells of the prostate. One of such proteins is a product of the MSMB gene [22]. Furthermore, elements of a major histocompatibility complex have as immunostimulating as immunosuppressive pro­perties. One of the immunosuppressive glycoprotein is HLA-G [23].

Immunohistochemistry, the new generation sequencing and a real-time polymerase chain reaction (PCR) have been used to identify markers, typical for different types of cells in a tumor microenvironment [15, 16, 20]. We selected a number of genes specific for fibroblasts and CAF, macrophages and TAM and various subpopulations of T-lymphocytes, NK cells. Also, markers of inflammation and nonspecific immune reactions were chosen. The levels of relative gene expression (RE) were assessed and putative correlation with clinical and pathological characteristics (CPC) and molecular features was analyzed. Further, we wanted to monitor the specific alterations in an expression pattern of selected genes that are clinically significant, for the possible application in the molecular profiling of prostate cancers.

MATERIALS AND METHODS

Prostate tissue collection. Tissues of prostate cancer (T) and the paired conventionally normal prostate tissues (CNT, or N from a site opposite to cancer) were frozen in liquid nitrogen directly after surgery. All samples were collected at National Cancer Institute (Kyiv, Ukraine). Benign prostate tumors (prostate adenoma samples — A) were collected with the same procedure at the Institute of Urology (Kyiv, Ukraine) after radical prostatectomy. The samples were collected in accordance with the Declaration of Helsinki and the guidelines, issued by the Ethic Committee of the Institute of Urology of the National Academy of Medical Sciences of Ukraine and the National Cancer Institute and the Ethic Committee of the Institute of Molecular Biology and Genetics of the National Academy of Sciences of Ukraine. 37 prostate adenocarcinomas of different Gleason score (GS) and stages, 37 paired CNT and 20 samples of benign prostate tumors (A, adenomas) were studied. Tumors were characterized, according to the International System of Classification of Tumors based on the tumor-node-metastasis (TNM) and the World Health Organization (WHO) criteria classification. CPC of prostate cancer samples were described earlier [24].

Total RNA isolation and cDNA synthesis. 50–70 mg of frozen prostate tissues were mashed to powder in liquid nitrogen. Total RNA was extracted by TRI-reagent (Sigma, USA), according to the manufacturer’s protocol. The total RNA concentration was analyzed by a spectrophotometer (NanoDrop Technologies Inc., USA). The quality of the total RNA was determined in a 1% agarose gel by band intensity of 28S and 18S rRNA (28S/18S ratio). cDNA was synthesized from 1 µg of the total RNA treated with RNase free DNase I (Thermo Fisher Scientific, USA) using RevertAid H Minus M-MuLV Reverse Transcriptase (Thermo Fisher Scientific, USA) according to the manufacturer’s protocol.

Quantitative PCR. Levels of a RE of 23 genes were assessed by quantitative PCR using a Maxima SYBR Green Master mix (Thermo Fisher Scientific, USA) on Bio-Rad CFX96 Real-­Time PCR Detection System (USA) under the following conditions: 95 °C — 10 min, following 40 cycles of 95 °C — 15 s, 60 °C — 30 s, elongation 72 °C — 30 s. Primers for all genes were selected from a qPrimerDepot (https://primerdepot.nci. nih.gov/) database and confirmed, using an https:// http://www.ncbi.nlm.nih.gov/tools/primer-blast/algorithm.

Four reference genes TBP, HPRT, ALAS1 and TUBA1B were used for gene expression normalization [25] in comparison with TBP [26]. Two main mo­dels (2-ΔCt and 2-ΔΔCt methods), described earlier [25, 27] were used for calculation and an analysis of RE levels.

Statistical analysis. The Kolmogorov — Smirnov test was applied to assess the normality of distribution. The Wilcoxon Matched Pairs test was performed for comparison of RE in prostate adenocarcinomas and paired normal tissues samples. RE fold differences in 2-ΔΔCt model were considered increased in cases of changes more than 2 fold (RE fold changes > 2.01) and decreased when changes were more than 2 fold (RE fold changes < 0.49), when gene RE were compared in adenocarcinoma versus CNT. The Fisher exact test was calculated to find out differences between these sample groups [24]. The Benjamini — Hochberg procedure with false discovery rate (FDR) 0.10–0.25 was used under multiple comparisons detection [28]. The Kruskal — Wallis test was used to determine differences between experimental groups. The Dunn — Bonferroni post hoc test for multiple comparisons was performed to determine RE differences between pairs of prostate samples. The Spearman’s rank correlation test was used to find possible correlations between gene RE and CPC of prostate tumors [25].

RESULTS

We have analyzed RE of 23 genes in prostate T, CNT and A. Eight genes are associated with fibroblasts and CAFs, we called them a CAF group. Six genes are associated with pro-inflammatory and immuno-suppressive macrophages (the TAM group). Nine genes RE is associated with immune defense and reactions (the immune-associated genes (IAG) group).

From the beginning, we have found differences in RE in T, compared with CNT (N) samples. Statistically significant differences in RE in the paired T/N samples are shown in Table 1 (FDR = 0.2).

Table 1. Genes with significant RE differences between prostate adenocarcinoma samples and paired CNT (2-ΔΔCt model)
Group Gene RE fold changes
> 2.01 < 0.49
CAF ACTA2 +
CXCL14 +
CTGF +
THY1 +
FAP +
TAM CD163 +
CCR4 +
CCL17 +
NOS2A +
IAG MSMB +
IL1R1 +

Differences in RE were calculated for 11 genes out of 23 in the paired T/N samples. Increased RE (> 2.01) in T in comparison with the paired N (< 0.05) was observed for 7 genes: ACTA2, CXCL14, CTGF, THY1, FAP, CD163, CCL17. Decreased RE (< 0.49) in T in comparison with the paired N (p < 0.05) was observed for 4 genes: CCR4, NOS2A, MSMB, IL1R1.

Next, RE was analyzed in 3 sample groups, namely T, the paired CNT (N) and A. Furthermore, the GS, stage, presence of the TMPRSS2/ERG fusion [27], levels of PTEN expression (low (l) or high (h)) [25, 29] was taken into consideration for grouping of the samples. Summarized data are presented in Table 2.

Table 2. Significant RE changes according to Dunn — Bonferroni post hoc test for multiple comparisons in investigated genes
Group Gene T/N paired TNA total TNA/stages TNA/GS TNA/fusion status TNA/PTEN,
l/h
CAF ACTA2 +
CXCL12 + + + + +
CXCL14 + + + + + +
CTGF + + + + +
FAP + + + # + +
HIF1A + +
S100A4
THY1 + + +
TAM CD68 + #
CD163 + + +
CCR4 + + + + + +
CCL17 + + + + +
CCL22 + +
NOS2A + + +
IAG CIAS1
CTLA4 + + + + +
HLA-G
IRF1
IL1RL1
IL1R1 + + + + +
IL2RA +
KLRK
MSMB +
Note: #Presence of RE differences only in TBP normalized genes.

These data show that the greatest number of changes in RE, for 14 out of 23 genes could be found in TNA groups, when the stage of tumor was analyzed as well. The expression pattern of these genes and also another 2 genes (ACTA2 and MSMB), for which RE was changed only in the paired T/N samples is shown on Figure. It should be noted, that for two genes (FAP and CD68) the use of four reference genes and one reference gene for normalization of RE resulted in various fold change differences.

111 Expression pattern of genes associated with tumor microenvironment in prostate cancer
Fig. 1. RE profiles of genes with differences in different stages of disease in prostate adenocarcinomas (carcinomas), CNT and adenomas (ae), f — genes with RE differences in paired T/N

Descriptive statistics of the data on RE in T, N and A samples at the various stages and also significant differences between groups are shown in Table 3. Among genes showing the significant differences in RE (14 genes) eight genes showed the lowest RE in the A group, namely CTGF, CXCL14, FAP, THY1, CD163, CCL17, CCL22, and CTLA4. Increased RE in T at stages 1–2 (p < 0.05) was observed for CTGF, FAP, THY1, CCL22, CTLA4, whereas increased RE in T stages 3–4 (p < 0.05) was found for CD163 and CCL17. CXCL14 was upregulated in all tumors.

Table 3. Descriptive statistics of the RE data in T, N, A sample groups at the various of disease stages and significant differences between pairs of groups
Gene Group, stage Median, r. u. Minimum, r. u. Maximum, r. u. 25.000th 75.000th Pairs with
differences*
ACTA2 T 1–2 st 89.909 20.917 412.01 62.79 129.333
T 3–4 st 96.673 35.072 281.768 76.795 156.101
N 1–2 st 86.332 29.923 203.147 65.074 125.702
N 3–4 st 55.296 26.263 154.637 37.942 115.354
A 92.329 56.143 299.884 74.621 132.95
CTGF T 1–2 st 10.691 5.431 65.433 8.107 16.302 T 1–2 st/A
T 3–4 st 14.209 2.994 24.378 7.745 17.114 T 1–2 st/N 3–4 st
N 1–2 st 8.358 2.699 29.541 6.176 15.361
N 3–4 st 5.196 2.628 14.929 3.964 6.724
A 4.583 1.803 28.079 3.287 10.298
CXCL12 T 1–2 st 0.124 0.029 0.334 0.082 0.166 T 1–2 st/A
T 3–4 st 0.141 0.03 0.329 0.089 0.216 T 3–4 st/A
N 1–2 st 0.13 0.025 0.527 0.102 0.213 N 1–2 st/A
N 3–4 st 0.112 0.036 0.315 0.086 0.163 N 3–4 st/A
A 0.312 0.11 0.651 0.249 0.417
CXCL14 T 1–2 st 2.304 0.199 22.365 1.14 5.517 T 1–2 st/A
T 3–4 st 2.753 0.887 63.428 0.989 22.299 T 3–4 st/A
N 1–2 st 0.865 0.141 3.715 0.573 1.762 N 3–4 st/A
N 3–4 st 2.188 0.111 49.347 1.123 3.796
A 0.394 0.143 4.779 0.293 0.485
HIF1A T 1–2 st 1.965 0.977 5.145 1.531 2.675 T 1–2 st/T 3–4 st
T 3–4 st 1.023 0.649 2.36 0.921 1.157 T 1–2 st/N 3–4 st
N 1–2 st 1.866 0.005 3.662 1.446 2.206 T 3–4 st/A
N 3–4 st 0.92 0.598 1.991 0.786 1.041 N 1–2 st/N 3–4 st
A 2.068 0.965 4.799 1.645 2.947 N 3–4 st/A
FAP T 1–2 st 0.176 0.013 0.575 0.11 0.238
T 3–4 st 0.142 0.05 0.647 0.073 0.358 T 1–2 st/A
N 1–2 st 0.109 0.012 0.402 0.054 0.212
N 3–4 st 0.091 0.039 0.408 0.067 0.182
A 0.091 0.019 1.279 0.039 0.132
S100A4 T 1–2 st 21.185 2.556 59.247 11.746 29.515
T 3–4 st 50.537 8.466 258.211 11.063 94.751
N 1–2 st 21.286 3.141 81.063 12.579 32.854
N 3–4 st 32.631 5.37 162.32 13.918 81.289
A 27.764 8.39 76.886 16.446 36.073
THY1 T 1–2 st 1.633 0.643 5.119 1.061 2.721 T 1–2 st/A
T 3–4 st 1.469 0.535 5.83 1.351 1.83
N 1–2 st 1.515 0.403 3.351 0.878 2.213
N 3–4 st 1.564 0.34 2.158 1.476 2.03
A 0.973 0.25 3.938 0.681 1.532
CD68 T 1–2 st 2.205 0.885 18.661 1.486 4.37 T 1–2 st/T 3–4 st#
T 3–4 st 0.803 0.134 10.319 0.551 1.625 T 1–2 st/N 3–4 st#
N 1–2 st 2.358 0.669 8.172 1.641 3.704
N 3–4 st 1.651 0.169 6.998 0.539 3.124
A 2.731 0.708 22.498 1.471 4.231
CD163 T 1–2 st 2.478 0.725 14.614 1.515 4.938 T 1–2 st/T 3–4 st
T 3–4 st 35.354 1.166 142.803 11.618 61.782 T 1–2 st/N 3–4 st
N 1–2 st 1.985 0.689 9.936 1.248 3.617 T 3–4 st/A, N 3–4 st/A
N 3–4 st 27.173 1.037 93.815 16.979 64.857 T 3–4 st/N 1–2 st
A 2.019 0.771 23.007 1.156 4.488 N 1–2 st/N 3–4 st
CCL17 T 1–2 st 0.01 0.001 0.067 0.004 0.017 T 1–2 st/N 3–4 st
T 3–4 st 0.054 0.006 0.236 0.009 0.116 T 3–4 st/A
N 1–2 st 0.006 0.001 0.07 0.004 0.011 T 3–4 st/N 1–2 st
N 3–4 st 0.084 0.011 0.336 0.056 0.132 N 1–2 st/N 3–4 st
A 0.006 0.001 0.024 0.004 0.009 N 3–4 st/A
CCL22 T 1–2 st 0.2 0.01 0.642 0.117 0.292
T 3–4 st 0.074 0.003 0.344 0.048 0.082 T 1–2 st/T 3–4 st
N 1–2 st 0.158 0.013 0.462 0.104 0.21 T 1–2 st/A
N 3–4 st 0.163 0.002 0.57 0.097 0.246 T 3–4 st/N 1–2 st
A 0.086 0.002 0.335 0.056 0.134
CCR4 T 1–2 st 0.156 0.017 0.654 0.085 0.272 T 1–2 st/T 3–4 st
T 3–4 st 0.079 0.002 0.142 0.016 0.111 T 3–4 st/A
N 1–2 st 0.164 0.027 0.592 0.1 0.263 T 3–4 st/N 1–2 st
N 3–4 st 0.137 0.001 0.523 0.077 0.201 N 3–4 st/A
A 0.232 0.094 3.138 0.191 0.405
NOS2A T 1–2 st 0.007 0 0.257 0.003 0.017
T 3–4 st 0.002 0 0.007 0 0.006 T 3–4 st/N 1–2 st
N 1–2 st 0.01 0.001 0.096 0.004 0.03
N 3–4 st 0.002 0.001 0.022 0.001 0.006
A 0.006 0 0.103 0.004 0.01
CIAS T 1–2 st 0.153 0.051 0.47 0.124 0.202
T 3–4 st 0.127 0.034 0.363 0.087 0.238
N 1–2 st 0.182 0.046 0.603 0.094 0.271
N 3–4 st 0.139 0.074 0.427 0.118 0.196
A 0.127 0.052 0.593 0.089 0.221
CTLA4 T 1–2 st 0.087 0.007 0.425 0.052 0.184 T 1–2 st/A
T 3–4 st 0.07 0.028 0.239 0.06 0.117 N 1–2 st/A
N 1–2 st 0.097 0.026 0.365 0.067 0.165 N 3–4 st/A
N 3–4 st 0.12 0.051 0.349 0.061 0.142
A 0.039 0.006 0.219 0.03 0.071
Table 3 (breakover)
Gene Group, stage Median, r. u. Minimum, r. u. Maximum, r. u. 25.000th 75.000th Pairs with differences*
HLA-G T 1–2 st 137.874 25.081 244.89 88.761 188.734
T 3–4 st 162.055 97.866 448.585 125.594 245.191
N 1–2 st 119.608 31.088 356.051 70.546 184.076
N 3–4 st 141.729 79.736 497.198 119.216 390.41
A 104.086 53.32 230.095 78.528 147.66
IL1R1 T 1–2 st 2.142 0.14 8.884 1.705 4.047 T 3–4 st/A
T 3–4 st 1.01 0.456 6.783 0.845 1.375 N 3–4 st/A
N 1–2 st 2.639 0.958 8.55 2.129 5.261
N 3–4 st 1.397 0.294 8.454 0.501 2.004
A 3.733 0.708 8.572 2.88 4.871
IL1RL1 T 1–2 st 0.132 0.05 0.494 0.09 0.2
T 3–4 st 0.163 0.026 0.37 0.065 0.255
N 1–2 st 0.15 0.07 0.842 0.12 0.226
N 3–4 st 0.21 0.009 0.335 0.171 0.289
A 0.211 0.058 0.701 0.156 0.294
IL2RA T 1–2 st 0.107 0.019 0.533 0.08 0.134
T 3–4 st 0.08 0.02 0.156 0.055 0.137
N 1–2 st 0.092 0.026 0.237 0.072 0.127
N 3–4 st 0.076 0.035 0.201 0.059 0.125
A 0.072 0.025 0.451 0.036 0.102
IRF1 T 1–2 st 1.121 0.386 153.203 0.778 1.648
T 3–4 st 0.767 0.688 2.926 0.736 0.813
N 1–2 st 0.947 0.325 3.361 0.686 1.414
N 3–4 st 1.055 0.576 2.59 0.835 1.895
A 0.94 0.486 5.527 0.696 1.073
KLRK T 1–2 st 0.004 0 0.027 0.002 0.006
T 3–4 st 0.002 0 0.053 0 0.008
N 1–2 st 0.004 0 0.04 0.001 0.009
N 3–4 st 0.006 0 0.022 0.003 0.008
A 0.008 0 0.032 0.005 0.016
MSMB T 1–2 st 384.275 3.563 1858.97 176.88 715.69
T 3–4 st 102.783 3.661 4887.53 56.79 240.35
N 1–2 st 564.827 19.582 3196.3 307.6 1121.02
N 3–4 st 75.143 9.688 4813.28 18.51 181.68
A 670.962 1.793 4367.28 178.17 1933.82
Note: *Dunn — Bonferroni post hoc method for multiple comparisons with FDR = 0.2 (p < 0.05); #presence of RE differences only in TBP normalized genes; st — stage; 25.000th — 25th percentile, 75.000th — 75th percentile.

CXCL12, HIF1A, CD68, CCR4, NOS2A, and IL1R showed the highest RE in the A group 1. The significantly decreased RE (more than two-fold) in T sta­ges 3–4 (p < 0.05) was demonstrated for five genes, namely HIF1A, CD68, CCR4, NOS2A and IL1R1. Only CXCL12 was downregulated in all tumors. Significant changes in RE (up- and downregulation) in the CNT group of patients with tumors of the various stages was also found for several of these genes (Table 4). For example, decreased RE of HIF1A and IL1R1 was detected in N compared with A in patients with T sta­ges 3–4. The CXCL12 was downregulated in all normal tissues. Increased RE in the CNT group was found for four genes. The CXCL14, CD163 and CCL17 genes were upregulated in T stages 3–4. One gene (CTLA4) was upregulated in all tumors.

Table 4. RE differences between pairs in TNA groups with TMPRSS2/ERG fusion status (4A) and different PTEN RE (4B)
A TNA/Fusion status B TNA/PTEN RE (h/l)
Gene Pairs p-value* Gene Pairs p-value*
THY1 T F–/A 0.0497 THY1 T h/A l 0.0431
CXCL12 T F–/A 0.0178 CXCL12 T l/A h 0.0000
T F+/A 0.0000 T h/A h 0.0406
N F–/A 0.0281 N l/A h 0.0000
N F+/A 0.0001 T l/N h 0.0358
CXCL14 T F–/A 0.0000 CXCL14 T l/A h 0.0000
T F+/A 0.0001 T h/A h 0.0023
N F–/A 0.0021 N l/A h 0.0303
T F–/N F+ 0.0256 T l/A l 0.0123
CTGF T F–/A 0.0211 CTGF T l/A h 0.0294
T F+/A 0.0153 T h/A h 0.0028
FAP T F–/A 0.0280 N h/A h 0.0340
T F–/N F+ 0.0419 T h/N l 0.0189
CTLA4 N F–/A 0.0010 FAP T h/A l 0.0374
CD163 N F–/NF+ 0.0309 CTLA4 N l/A h 0.0437
CCR4 T F–/A 0.0195 CCR4 T l/A h 0.0037
N F+/A 0.0455 IL1R1 T l/A h 0.0240
CCL17 N F–/A 0.0225
CCL22 N F–/A 0.0134
Note: *Dunn — Bonferroni post hoc method for multiple comparisons with FDR = 0.2, F+ and F– groups with presence (+) and absence (–) of TMPRSS2/ERG fusion, h/l — groups with high (h) and low (l) PTEN RE.

Nine genes in groups with different GS showed similar RE changes as genes in groups with different stages (Table 2). IL2RA was altered only in groups with different GS. Pairs with RE variability in the TNA group and depending on the presence/absence of the TMPRSS2/ERG fusion and also of PTEN RE are shown in Table 4. We have found 10 genes in TNA groups with the different TMPRSS2/ERG fusion status, showing significant differences in RE (p < 0.05). The most interesting are genes, which were differently expressed in T with and without the TMPRSS2/ERG fusion. Three genes — THY1, FAP and CCR4 — showed alterations only in T without the fusion, compared with the A group. Four genes (CTLA4, CCL17, CCL22) were expressed differently in the CNT group without the fusion, compared with the A group. One gene was (CD163) expressed at various levels in the fusion negative compared with the fusion-positive CNT groups.

Eight genes expressed at the differential levels in TNA groups (p < 0.05), when PTEN mRNA was analyzed. The most interesting among them are THY1 and FAP that were expressed at significantly different RE levels in adenocarcinomas with high PTEN (T h) and in adenomas with high PTEN RE (A h). CCR4 and IL1R1 showed differences in adenocarcinomas with low PTEN RE (T l) and adenomas with high PTEN RE (A h).

Correlation analysis of RE related to CPC and molecular characteristics of tumors. Using the Spearman’s rank correlation (rs) test, a number of correlations was suggested between the gene expression pattern and CPC, the TMPRSS2/ERG fusion status and PTEN levels in prostate adenocarcinomas (Table 5). RE of the nine genes, namely CXCL14, HIF1A, S100A4, CCL17, CCR4, NOS2A, CTLA4, IL2RA and MSMB correlated with GS. RE of the nine genes (HIF1A, S100A4, CD68, CD163, CCL17, CCL22, CCR4, NOS2A, IL1R1) showed significant correlations with the stage. Six genes out of nine belong to the TAM group. CD163 RE demonstrated the maximal positive correlation with the stage (rs = 0.615). RE of the THY1 gene correlated significantly with an age. RE of CXCL14, IL1RL1 and IRF1 correlated with PSA levels. RE of ACTA2 and CXCL12 negatively correlated with the TMPRSS2/ERG fusion status. These both genes belong to the CAF group. RE of other genes from the CAF group, namely ACTA2, CXCL12, FAP and S100A4 positively correlated with PTEN levels.

Table 5. Spearman rank order correlations (rs) of CPC and molecular characteristics with genes RE
Gene group CAF TAM
CPC, molecular characteristics ACTA2 CTGF CXCL12 CXCL14 HIF1A FAP S100A4 THY1 CD68 CD163 CCL17 CCL22 CCR4 NOS2A
GS –0.072 –0.133 0.008 0.360 –0.403 –0.021 0.353 –0.056 –0.217 0.224 0.437 –0.304 –0.492 –0.352
Stage 0.158 0.070 0.151 0.181 –0.538 –0.024 0.336 0.012 –0.369 0.615 0.435 –0.398 –0.436 –0.407
Age 0.180 0.248 0.153 0.148 0.204 0.278 0.027 0.328 0.136 0.193 0.092 0.004 0.193 0.034
PSA, ng/ml –0.018 0.008 0.069 0.479 –0.233 0.098 0.141 0.225 –0.097 0.230 0.324 –0.174 –0.214 –0.169
Fusion status –0.406 –0.126 –0.415 0.155 0.090 –0.123 –0.176 –0.157 –0.106 –0.143 0.051 0.129 0.090 –0.019
PTEN RE 0.447 0.285 0.529 0.002 0.146 0.439 0.431 0.205 0.234 0.290 –0.007 0.160 0.180 0.183
Gene group IAG
CPC, molecular characteristics CIAS CTLA4 HLA-G IL1R1 IL1RL1 IL2RA IRF1 KLRK MSMB
GS –0.247 –0.382 –0.076 –0.272 –0.057 –0.449 –0.316 –0.154 –0.437
Stage –0.082 –0.148 0.246 –0.406 0.033 –0.155 –0.303 –0.111 –0.315
Age 0.213 0.172 –0.191 –0.084 0.020 0.023 0.151 0.294 –0.180
PSA, ng/ml –0.119 –0.117 0.128 –0.321 –0.410 –0.166 –0.417 –0.190 –0.114
Fusion status –0.114 –0.082 –0.042 0.130 –0.017 0.095 0.066 0.113 0.098
PTEN RE 0.103 0.024 0.192 0.133 0.109 0.026 0.096 0.050 –0.126

Note: p < 0.05 (red); p < 0.01 (red bold italic); p < 0.001 (red bold).

We have to mention, that several genes show correlation in their expression levels. For example, in the T group, expression of CD163 positively correlated with S100A4 levels, as did FAP and THY1, CXCL12 and IL2RA, and also CTLA4 and IL2RA. The correlation indexes were high — rs = |0.633–0.712|, and correlation was p < 0.0001. The largest number of correlations out of 23 investigated genes (p < 0.01) demonstrated CD68, CCR4 and CCL22, i.e. each of these genes showed 8 correlations. For NOS2A, seven dependences were found, for HIF1A — six and for CTLA4, CIAS, IL2RA — five.

DISCUSSION

A surgical material of prostate cancer is very hete­rogeneous, as a rule [30]. Tumor samples usually show a different structure with various sets of cells (i.e. cancerous cells, normal epithelial cells, stromal and immune cells, etc.) [31, 32]. Therefore, a study on specific tumor-host interactions is a very difficult task. To solve this problem, at least, partially, specific markers of different cell types were selected and gene RE was assessed in benign and malignant prostate tumors. We aimed to find the clinically significant alterations in the gene expression pattern. We found significant changes in RE only in a proportion of the investigated genes. Thus, 6 out of 8 genes of the CAF group showed differences in RE in the TNA group, containing tumors at various stages. Of note, the known CAF marker ACTA2 [33] showed RE changes only in the group of the paired T/N samples.

The CXCL14, CTGF, FAP and THY1 also diffe­rentially expressed in tumors and CNT (N), and they showed alterations in RE in the TNA group, including tumors of the various stages. These four genes that are upregulated in tumors encode proteins secreted by CAFs. These proteins play an important role in paracrine interactions in the tumor microenvironment, enhancing tumor progression and metastasizing [34]. Moreover, the FAP protein exhibits immune-suppressive properties [35]. RE of HIF1A and S100A4 (also the CAF genes) correlated with the stage of tumors. Only CAF genes demonstrated correlation of RE with molecular characteristics of prostate tumors, namely the TMPRSS2/ERG fusion status and PTEN levels. Probably, this indicates the closest and selective type of interaction of the tumor and fibroblasts, as matrix-regulating elements of the prostate stroma [13, 15, 33].

Genes of the TAM group showed altered RE in the TNA group, including tumors at different stages. Some of genes demonstrated RE changes in the TNA group, when GS and the TMPRSS2/ERG fusion status varied. All of the 6 TAM genes correlated with tumor stage. We may speculate on the clinical significance of the selected markers of the TAM group [10, 16, 19]. Genes CD68 and NOS2A are the markers of normal, non-cancerous macrophages [16], whereas CD163 is the TAM marker [18]. CCR4, CCL17 and CCL22 are associated with transition of the macrophage into TAM, i.e. these genes play a role in tumor progression [17].

The smallest number of genes with altered RE is found in the IAG group. Only two genes (CTLA4 and IL1R1) showed changes in RE in the TNA group, including the tumors of different stages. RE of IL1R1and MSMB was changed in the paired T/N group. It is known that inhibition of CTLA4 expression is very important in immune checkpoint therapy of prostate cancer, due to the fact that CTLA4 transmits the immune inhibitory signals to T cells [36]. An IL1R1 signaling suppresses proliferation of mammary tumor cells, early in tumorigenesis [37]. We found the significant decrease of IL1R1 RE in prostate cancer at the advanced stages of tumor development.

As mentioned earlier, the histological analysis indicated focal manifestations of leukocyte infiltration in the prostate tumors [31]. Therefore, expression of leukocyte genes could vary quite much within even one sample [21, 32]. Based on this, it was important for us to assess the gene expression in the various immune cells. Also, the immune score indexes could have their own independent diagnostic and prognostic values [38]. More suitable markers of prostate carcinogenesis should be studied, to reveal clinically significant alterations of immune response.

The observed high dispersion of the RE values did not make it possible to find the real difference between the groups of samples in all genes. In addition, the lack of the adequate control (a healthy donor’s prostate with the relevant age and nationality) also affected experimental results. Moreover, the changes in expression of the studied genes in adenomas and CNT might have a specific character, showing the effect of hypertrophy and hyperplasia in adenomas and the influence of malignant tumors on prostate CNT, especially at the late stages of tumor development [31, 32]. This assumption is confirmed by the results of the present study. We got the different number of genes with RE changes in the group of paired T/N (11 genes) and in adenomas, taking into account the tumor stage (14 genes). Of note, ACTA2 and MSMB showed RE changes only in the paired samples. Unexpectedly, ACTA2 was highly upregulated in adenomas. Probably, this is due to the specific expression of various ACTA2 isoforms upon tumor development and the specific expression of ACTA2 in different cell types, like smooth muscle cells [33, 34].

Earlier, we have shown that, based on RE of prostate-specific genes and EMT-associated genes, samples with the same clinical parameters could be grouped into specific RE subtypes [24, 25]. To confirm this, more experiments should be done on the larger patient cohort.

CONCLUSIONS

Several gene expression patterns were identified that correlated with the GS, stage and molecular characteristics of tumors, i.e. presence of the TMPRSS2/ERG fusion and alterations in PTEN expression. These expression patterns can be used for molecular profiling of prostate tumors, with the aim to develop personalized medicine approaches. However, the proposed profiling requires a more detailed analysis and a larger cohort of patients with prostate tumor.

References

1. Saraon P, Jarvi K, Diamandis EP. Molecular alterations during progression of prostate cancer to androgen independence. Clin Chem 2011; 57: 1366–75.
2. Khemlina G, Ikeda S, Kurzrock R. Molecular landscape of prostate cancer: implications for current clinical trials. Cancer Treat Rev 2015; 41: 761–6.
3. Ruggero K, Farran-Matas S, Martinez-Tebar A, et al. Epigenetic regulation in prostate cancer progression. Curr Mol Biol Rep 2018; 4: 101–15.
4. Zahalka AH, Arnal-Estapé A, Maryanovich M, et al. Adrenergic nerves activate an angio-metabolic switch in prostate cancer. Science 2017; 358: 321–6.
5. Kluth LA, Shariat SF, Kratzik C, et al. The hypothalamic-pituitary-gonadal axis and prostate cancer: implications for androgen deprivation therapy. World J Urol 2014; 32: 669–76.
6. Solomon I, Voiculescu VM, Caruntu C, et al. Neuroendocrine factors and head and neck squamous cell carcinoma: an affair to remember. Dis Markers 2018; 2018: 9787–831.
7. Bryant G, Wang L, Mulholland DJ. Overcoming oncogenic mediated tumor immunity in prostate cancer. Int J Mol Sci 2017; 18: pii: E1542.
8. Todoric J, Antonucci L, Karin M. Targeting inflammation in cancer prevention and therapy. Cancer Prev Res (Phila) 2016; 9: 895–905.
9. Kwon OJ, Zhang L, Ittmann MM, et al. Prostatic inflammation enhances basal-to-luminal differentiation and accelerates initiation of prostate cancer with a basal cell origin. Proc Natl Acad Sci U S A 2014; 111: E592–600.
10. Strasner A, Karin M. Immune infiltration and prostate cancer. Front Oncol 2015; 5: 128.
11. Puhr M, De Marzo A, Isaacs W, et al. Inflammation, microbiota, and prostate cancer. Eur Urol Focu 2016; 2: 374–82.
12. Sfanos KS, Yegnasubramanian S, Nelson WG, et al. The inflammatory microenvironment and microbiome in prostate cancer development. Nat Rev Urol 2018; 15: 11–24.
13. Lorion R, Bladou F, Spatz A, et al. Prostate cancer microenvironment: Its structure, functions and therapeutic applications. Prog Urol 2016; 26: 464–76.
14. Eder T, Weber A, Neuwirt H, et al. Cancer-associated fibroblasts modify the response of prostate cancer cells to androgen and anti-androgens in three-dimensional spheroid culture. Int J Mol Sci 2016; 17. pii: E1458.
15. Attieh Y, Vignjevic DM. The hallmarks of CAFs in cancer invasion. Eur J Cell Biol 2016; 95: 493–502.
16. Hu W, Qian Y, Yu F, et al. Alternatively activated macrophages are associated with metastasis and poor prognosis in prostate adenocarcinoma. Oncol Lett 2015; 10: 1390–6.
17. Maolake A, Izumi K, Shigehara K, et al. Tumor-associated macrophages promote prostate cancer migration through activation of the CCL22-CCR4 axis. Oncotarget 2017; 8: 9739–51.
18. Parisi L, Gini E, Baci D, et al. Macrophage polarization in chronic inflammatory diseases: killers or builders? J Immunol Res 2018; 2018: 8917804.
19. Mantovani A, Locati M. Tumor-associated macrophages as a paradigm of macrophage plasticity, diversity, and polarization: lessons and open questions. Arterioscler Thromb Vasc Biol 2013; 33: 1478–83.
20. Danaher P, Warren S, Dennis L, et al. Gene expression markers of tumor infiltrating leukocytes. J Immunother Cancer 2017; 5: 18. DOI 10.1186/s40425-017-0215-8.
21. Rådestad E, Egevad L, Jorns C, et al. Characterization of infiltrating lymphocytes in human benign and malignant prostate tissue. Oncotarget 2017; 8: 60257–69.
22. Sutcliffe S, De Marzo AM, Sfanos KS, et al. MSMB variation and prostate cancer risk: clues towards a possible fungal etiology. Prostate 2014; 74: 569–78.
23. Langat DK, Sue Platt J, Tawfik O, et al. Differential expression of human leukocyte antigen-G (HLA-G) messenger RNAs and proteins in normal human prostate and prostatic adenocarcinoma. J Reprod Immunol 2006; 71: 75–86.
24. Gerashchenko GV, Mevs LV, Chashchina LI, et al. Expression of steroid and peptide hormone receptors, metabolic enzymes and EMT-related genes in prostate tumors in relation to the presence of the TMPRSS2/ERG fusion. Exp Oncol 2018; 40: 1–8.
25. Gerashchenko G, Mankovska O, Dmitriev A, et al. Epithelial-mesenchymal transition related gene expression in prostate tumours. Biopolym Cell 2017; 33: 335–55.
26. Rosenberg E, Gerashchenko G, Hryshchenko N, et al. Expression of cancer-associated genes in prostate tumors. Exp Oncol 2017; 39: 131–7.
27. Mevs L, Gerashchenko G, Rosenberg E, et al. Detection of prostate specific ETS fusion transcripts in cancer samples. Biopolym Cell 2017; 33: 256–67.
28. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Royal Statist Society 1995; 57: 289–300.
29. Gerashchenko GV, Grygoruk OV, Mevs LV, et al. PTEN-dependent changes of gene expressions in human prostate tumors. Factors Experim Evol Organisms 2018; 22: in press.
30. Tolkach Y, Kristiansen G. The heterogeneity of prostate cancer: a practical approach. Pathobiology 2018; 85: 108–16.
31. Cheng L, MacLennan GT, Lopez-Beltran A, et al. Anatomic, morphologic and genetic heterogeneity of prostate cancer: implications for clinical practice. Expert Rev Anticancer Ther 2012; 12: 1371–4.
32. Runa F, Hamalian S, Meade K, et al. Tumor microenvironment heterogeneity: challenges and opportunities. Curr Mol Biol Rep 2017; 3: 218–29.
33. Yang Z, Peng YC, Gopalan A, et al. Stromal hedgehog signaling maintains smooth muscle and hampers micro-invasive prostate cancer. Dis Model Mech 2017; 10: 39–52.
34. Öhlund D, Elyada E, Tuveson D. Fibroblast heterogeneity in the cancer wound. J Exp Med 2014; 211: 1503–23.
35. Yang X, Lin Y, Shi Y, et al. FAP promotes immunosuppression by cancer-associated fibroblasts in the tumor microenvironment via STAT3-CCL2 signaling. Cancer Res 2016; 76: 4124–35.
36. Graff JN, Puri S, Bifulco CB, et al. Sustained complete response to CTLA-4 blockade in a patient with metastatic, castration-resistant prostate cancer. Cancer Immunol Res 2014; 2: 399–403.
37. Dagenais M, Dupaul-Chicoine J, Douglas T, et al. The Interleukin (IL)-1R1 pathway is a critical negative regulator of PyMT-mediated mammary tumorigenesis and pulmonary metastasis. Oncoimmunology 2017; 6: e1287247.
38. Galon J, Pagès F, Marincola FM, et al. The immune score as a new possible approach for the classification of cancer. J Transl Med 2012; 10: 1–20.

No Comments » Add comments
Leave a comment

ERROR: si-captcha.php plugin says GD image support not detected in PHP!

Contact your web host and ask them why GD image support is not enabled for PHP.

ERROR: si-captcha.php plugin says imagepng function not detected in PHP!

Contact your web host and ask them why imagepng function is not enabled for PHP.