TGF-β/PI3K/AKT/mTOR/NF-kB pathway. Clinicopathological features in prostate cancer

Abstract

Prostate cancer is one of the most common cancers in the male population. The objective of this investigation was to study the relationship of components of transforming growth factor-B (TGF-β)/phosphoinositide-3-kinases (PI3K)/AKT/mammalian target of rapamycin (mTOR)/nuclear factor kappa B (NF-kB) transduction pathway with clinical-pathological markers. By immunohistochemical methods, we determined the expression of several factors [TGF-β, Transforming Growth Factor B Receptor I (TGFBRI), TGFBRII, PI3K, AKT-Ser, AKT-Thr, mTOR, p-mTOR, inhibitor kB kinase (IKK), pIKK, inhibitor kB (IkB), pIkB, NF-kBp50, and NF-kBp65]. To know their relationship with established classical markers (Preoperative serum prostate specific antigen, pathological tumor stage, clinical tumor stage, Gleason score, perineural invasion, node involvement, positive surgical margins, biochemical progression, and survival) and their importance in the prognosis of biochemical progression, Spearman test, survival analysis, Log-rang test, Kaplan–Meier curves, univariate and multivariate Cox proportional Hazard regression analyses were performed. Spearman analysis showed that there was at least one correlation between TGF-β, TGFBRI, PI3K, pAKT-Thr, p-mTOR, NF-kBp50, and classical markers. Cox multivariate analysis between the prognostic variables (pathological tumor stage, Gleason score, and node involvement) and inmunohistochemical parameters confirmed TGFBR1 and PI3K as a prognostic and independent marker of biochemical progression in prostate cancer. Our results suggest that TGFBR1 and PI3K could be used as useful biomarkers for early diagnosis and prognoses for biochemical recurrence in prostate cancer after radical prostatectomy.

Keywords:

• Prostate cancer
• TGF-B
• NF-kB
• AKT
• mTOR
• PI3K

Background

In developed countries, prostate cancer is the most incidental cancer and the third in mortality. Up to 50% of elderly men have small clinically insignificant tumors and the risk of developing prostate cancer, in more developed areas, is around 7.8%, becoming a public health problem [1]. Prostate cancer is related to many factors. Among them, there would be factors as different as cytokines [2], human papillomavirus [3], or even nutrition [4].

In the tumor microenvironment, poor regulation of the transforming growth factor-B (TGF-β) pathway results in the loss of control of cellular functions, allowing cell survival, invasion or metastasis. TGF-B expression may be useful clinic-pathological features for prostate cancer [5]. In previous results published by our group, we suggest a possible role for prostatic expression of TGF-B, as prognostic markers for prostate cancer [6]. Previous work describes the importance of studying the poor regulation of TGF-β and its ability to activate several pathways such as mitogen-activated protein kinases (MAPKs), phosphoinositide-3-kinases (PI3K), serine/threonine protein kinases (AKT), nuclear factor kappa B (NF-kB) among others [7]. Previously the PI3K/AKT/mammalian target of rapamycin (mTOR) pathway deregulation was related with cancer [8]. In previous works, our group [9] suggests PI3K/AKT/mTOR pathway as an important therapeutic target and very viable option as predictive biomarkers of onset, behavior, and progression of PC after prostatectomy.

PI3K are a family of lipid kinases that regulate multiple signaling pathways, physiological functions, and cellular processes [10]. The pathological activation of the PI3K pathway is among the most frequent signaling events associated with cell transformation, cancer, and metastasis [11]. PI3K is a heterodimer consisting of a regulatory subunit (p85) and a catalytic subunit (p110) [12]. The phosphorylation of the p85 subunit leads to a conformational change of said protein that leads to the binding of p110. The PIK3CA gene encoding the PI3K catalytic isoform p110α, the second most frequently mutated oncogene [13]. Active PI3K phosphorylates phosphatidyl inositol 3,4 diphosphate (PIP2) making it the second messenger phosphatidyl inositol 3,4,5 triphosphate (PIP3) and of phosphatase and tensin homolog (PTEN) coding for PIP3 which is among the most mutated tumor suppressor genes [13]. PIP3 downstream leads to activation of the AKT protein [12].

In mammals, AKT family consists of three members AKT1 (PKBα), AKT2 (PKBB), and AKT3 (PKBγ) that can be activated by the same mechanism [14]. After activation, AKT translocates to intracellular compartments where it phosphorylates several protein substrates. The downstream targets of AKT are numerous due to multiple interactions with other substrates [15].

The regulation of AKT is achieved mainly through PTEN, which antagonizes with PI3K. PTEN is a tumor suppressor gene that frequently mutates in different types of cancer. The loss of PTEN leads to the elevation of PI3K and activates the AKT pathway [16]. Therefore, PTEN negatively regulates the AKT pathway, whereas the loss of PTEN leads to hyperactivation of AKT, which induces proliferation and promotes survival by inhibiting apoptosis [17]. Activated AKT can activate the mTOR.

The mTOR serine/threonine kinase functions as two structurally distinct multiprotein complexes, the mTOR 1 complex (mTORC1) and the mTOR 2 complex (mTORC2). mTORC1 and mTORC2 are functionally different kinases with distinct subcellular localization that activate separate paths, but still interconnected with different regulators upstream and downstream of the targets [18]. Activation of mTORC1 regulates and promote cell growth by inducing and inhibiting anabolic and catabolic processes, respectively [18]. MTORC2 responds to growth factors and regulates cell survival and metabolism, as well as the cytoskeleton through the phosphorylation of AKT-Ser [18]. The deregulation of mTOR signaling in cancer occurs through complex mechanisms, including overexpression or activation of oncogenic genes or by or loss of function of tumor suppressor genes, has also been implicated in the mechanisms of resistance to cancer [19].

Both AKT and mTOR are capable of promoting the activation of the inhibitor kB (IkB) kinase complex (IKK) to allow the transduction of NF-kB. IKK family is composed to two catalytic subunits (IKK-a, IKK-B) and a regulatory subunit (IKK-γ) [20]. Among these three subunits, IKK-B promotes NF-kB signaling and the pathogenesis and progression of inflammatory diseases [21]. Under normal conditions NF-kB is retained in the cytoplasm together with IkB (inhibitor factor-kB) forming an IkB/NF-kB complex [22]. For the activation of NF-kB to occur, phosphorylation of IkB by the IKK kinase is necessary, producing the separation of the complex, NF-kB penetrates the nucleus acting as a transcription factor. NF-kB has been considered the central mediator of the inflammatory process and innate immunity, contributing to important intracellular and extracellular processes. Activation of NF-kB is common in carcinomas, where its activation is mainly driven by the inflammatory cytokines present in the tumor microenvironment [23]. NF-kB acts in a specific way activating the survival genes and inflammation promoters [24].

Up to the present time and as far as we have been able to consult in the bibliography, there are no published works that link TGF-B, PI3K/AKT/mTOR pawthay and NK-kB in prostate cancer. Thus, the aim of the present study was to know the relationship established between the TGF-β/PI3K/AKT/mTOR/NF-kB transduction pathway members and the classical markers established. It is possible that some of them are related to biochemical progression and, together with PSA (prostate specific antigen), may serve as a predictive factor for recurrence.

Methods

Study patients

All the procedures followed were examined and approved by the University of Alcalá and Principe de Asturias Hospital Ethics Committees (reference number 2013/003/20130214) and were in accordance with the ethical standards of the Committee for Human Experimentation, with the Helsinki Declaration of 1975 (revised in Tokyo 2004) and the Committee on Publication Ethics (COPE) guidelines. This study was made with the consent of the patients’ relatives or their family in autopsy cases.

Prostates were obtained from: (a) histologically normal prostates (NP) obtained at autopsy (8–10 h after death) from 20 men (aged between 20 and 38 years) without histories of reproductive, endocrine or related diseases; (b) radical prostatectomies from 86 men (aged between 52 and 74 years) with prostate cancer. Prostate cancer was detected by serum PSA screening and rectal examination, and diagnostic was confirmed by histopathological examination of needle biopsy cores. The median age (range) at time of surgery was 66 (52–74) years. Patients were generally scheduled to have a serum PSA measure every 3 months for the first year and every 6 months thereafter. Median follow-up (range) time of the cohort was 75.9 (15.6–159.2) months, being defined as the time between the surgery and the endpoint of the study or the last record. Clinic-pathological features of patients are shown in Table 1 [22].

Table 1. Clinic-pathological features of patients.

	Median (range)
Age (years)	66.0 (52–74)
Preoperative serum PSA (ng/ml)	11.35 (1,4–233)
	% (n)		% (n)
Preoperative serum PSA		Gleason score
<10 ng/ ml	36.0 (31)	≤6	81.4 (70)
≥10 ng/ ml	64.0 (55)	>6	18.6 (16)
Clinical Tumor stage		Pathological Tumor stage
I	54.7 (47)	II	65.1 (56)
II	45.3 (39)	III	31.4 (27)
		IV	3.5 (3)
Perineural invasión (yes)	26.7 (23)	Node involvement (yes)	4.7 (4)
Positive surgical margins	38.4 (33)	Biochemical progression (yes)	33.7 (29)
Survival (yes)	73.3 (63)

Reagents

Total serum PSA was measured by the AxSYM system (Abbott, IL, USA). All primary antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The primary antibodies used were: rabbit anti-human TGF-β, TGFBRI (Transforming Growth Factor B Receptor I), TGFBRII (Transforming Growth Factor B Receptor II), PI3K, pAKT-Ser, pAKT-Thr, mTOR, p-mTOR, IKK, and pIKK; mouse anti-human IkB, pIkB, NF-kB p50, and NF-kB p65. For immunohistochemical analysis antibodies were diluted at NF-kB p50 (used at 1:500), TGF-β (used at 1:200), TGFBRII and PI3K (used at 1:100), mTOR, p-mTOR and IkB (1:75), TGFBRI, pAKT-Ser, pAKT-Thr, pIkB and NF-kB p65 (1:50), pIKK (1:25), in TBS (Tris buffered saline). Biotin-conjugated antibodies (DAKO, Barcelona, Spain) and vectastain ABC (avidin–biotin complex) kit (Vector Labs, Barcelona, Spain) were at a 1:500 dilutions. Immnohistochemical study has been performed as shown in previous publications [9,25].

Immunohistochemical analysis and scoring

Immediately after surgery prostate tissues were fixed, dehydrated, and embedded in paraffin, and 5 µm thick sections were processed following the avidin–biotin complex (ABC) method as described previously [26].

Specificity controls for immunohistochemistry were performed as previously published [2]. Briefly, for negative controls tissues were incubated with blocking peptides or preimmune serum (Santa Cruz Biotechnology). As external positive controls, histologic sections of human lymph nodes, skin or thymus were incubated with the same antibodies.

Immunostaining was assessed by two independent pathologists, blinded for the endpoint, in five randomly selected fields per section and six sections per patient. First, patients were stratified as positive (those showing staining in more than 5% of the surface of the corresponding compartment) or negative. Moreover, a score combining both intensity and percentage of immunostained surface was assigned to each sample, and accordingly patients were stratified as having “negative,” “positive low,” or “positive high” immunostaining [25].

Statistical analysis

The endpoint of the study was time to biochemical progression, defined as the time between definitive therapy to the first of at least two consecutive elevations in the total serum PSA level above 0.2 ng/ml. Established prognostic variables included in the study were preoperative serum PSA, pathological and clinical T stages (2010 AJCC/UICC TNM classification [27], postoperative Gleason score (2014 International Society of Urological Pathology (ISUP) [28]), lymph node involvement and surgical margins status. To evaluate the association between clinicopathological (clinical Tumor stage, pathological Tumor stage, preoperative serum PSA and Gleason score) and immunohistochemical variables Spearman test was performed. Log-rank test and Kaplan–Meier analyses were used for survival comparisons. To explore the correlation of the studied immunohistochemical parameters and established prognostic variables with biochemical progression, univariate and multivariate Cox proportional hazard regression analyses were performed. All statistical analyses were performed using the SPSS 22.0 software (SPSS Inc. Chicago, IL). p values <0.05 were considered as significant.

Results

Expression of components TGF-β/PI3K/AKT/mTOR/NF-kB pathway

In previous works, we have studied the expression of some of these proteins by immunohistochemistry [6,9,29].

In normal prostate, immunoreaction was found for TGFBRII, mTOR, p-mTOR, IKK, pAKT-Ser, pIKK, IkB, and pIkB. Immunostaining was weak for TGFBRII, mTOR, p-mTOR, and IKK. No immunoreaction was found to TGF-β, TGFBRI, PI3K, NF-kB p50, and NF-kB p65 (Table 2).

Figure 1. TGF-β (A, B), TGFBRI (C, D), TGFBRII (E, F), PI3K (G, H), p-AKT-Ser (I, J) and p-AKT- Thr (K, L) in normal prostate (A, E, G) and prostate cancer (B, C, D, F, J, K, L). Scale bars: 25 μm (B, C, D, E, F, I, J, K, L), 20 μm (G, H), and 15 μm (A).

Table 2. Number (n) and percentages of patients (%) showing positive immunohistochemical reactions to all the antibodies studied in normal prostate (NP) and prostatic carcinoma (PC); and average optical densities (O.D.) of immunostainings in positive patients (epithelium and stroma).

	n (%)	NP O.D.	n (%)	PC O.D.	P Low n (%)	P High n (%)
TGF-B	–	–	49 (57.0)	19.80 ± 1.14	23 (26.7)	26 (30.2)
TGFBRI	–	–	54 (62.8)	18.73 ± 2.42	31 (36.0)	23 (26.7)
TGFBRII	8 (40.0)	14.88 ± 0.9^a	73 (84.9)	21.18 ± 2.22^b	21 (24.4)	52 (60.5)
PI3K	–	–	82 (95.3)	6.29 ± 1.9	6 (7.0)	76 (84.4)
pAKT-Ser	20 (100)	33.62 ± 1.8^a	78 (90.7)	41.0 ± 1.6^b	12 (14.0)	66 (76.7)
pAKT-Thr	20 (120)	21.5 ± 1.6^a	69 (80.2)	28.5 ± 1.95^b	2 (2.3)	66 (76.7)
mTOR	17 (85.0)	15.77 ± 2.1	96.5 (83)	18.45 ± 3.65	17 (19.8)	66 (76.7)
p-mTOR	20 (100)	15.75 ± 1.7^a	83 (96.5)	22.05 ± 1.55^b	11 (12.8)	72 (83.7)
IKK	20 (100)	19.86 ± 3.65^a	63 (73.3)	46.82 ± 3.95^b	10 (11.6)	53 (61.6)
pIKK	25 (29.1)	27.28 ± 3.85^a	69 (80.2)	43.54 ± 4.02^b	7 (8.1)	64 (72.1)
IkB	12 (60)	30.67 ± 5.05^a	46 (53.5)	48.35 ± 5.2^b	7 (8.1)	39 (45.3)
pIkB	8 (40.0)	27.91 ± 4.2^a	67 (77.9)	50.88 ± 5.1^b	4 (4.7)	63 (73.3)
NF-kB p50	–	–	72 (83.7)	31.50 ± 2.65	7 (8.1)	65 (75.6)
NF-kB p65	–	–	71 (82.6)	32.00 ± 3.11	4 (4.7)	67 (77.9)

Patients positive for immunohistochemistry were stratified according to the intensity of immunostaining in positive low (P Low) and positive high (P High).

Average optical densities were evaluated only in patients showing positive immunoreactions. Values denoted by different superscripts are significantly different from each other. Those values that do not have superscript means that they are not statistically different from each other. Statistical analysis refers to each antibody separately. Significance was determined by the Fisher exact test at p > 0.05.

In prostate cancer, all these factors (except mTOR) are expressed increasing its expression with respect to normal prostate (Figure 1; Table 2).

Correlation between clinic-pathological features of patients and immunohistochemical variables

In this study, no significant correlation was observed between the expression of TGFBRI, TGFBRII, pAKT-Ser, mTOR, IKK, pIKK, IkB, pIkB, and NF-kBp65 with the established classical markers (Table 3). The rest of the proteins studied had an inverse correlation with some of the classic markers except for the preoperative PSA and the Gleason score.

Table 3. Correlation expression of studied proteins and clinic-pathological features.

	TGF-B		TGFBRI		TGFBRII		PI3K		pAKT-Ser		pAKT-Thr		mTOR
	r	p	r	p	r	p	r	p	r	p	r	p	r	p
Preoperative serum PSA	−0.65	0.550	−0.077	0.482	0.021	0.846	0.064	0.557	0.093	0.394	−0.008	0.943	−0.011	0.922
Pathological Tumor stage	−0.064	0.560	0.076	0.489	0.111	0.310	−0.102	0.352	−0.118	0.279	0.067	0.539	0.011	0.922
Clinical Tumor stage	0.037	0.737	0.121	0.266	0.058	0.593	−0.132	0.227	−0.030	0.784	−0.017	0.876	−0.209	0.054
Gleason score	−0.007	0.949	−0.003	0.979	0.035	0.750	−0.036	0.740	−0.053	0.630	0.012	0.911	0.091	0.405
Perineural invasion	−0.006	0.960	−0.024	0.826	−0.038	0.726	−0.116	0.287	−0.078	0.476	−0.294	0.006*	0.115	0.292
Node involvement	0.080	0.462	−0.058	0.593	−0.061	0.577	0.049	0.656	0.071	0.518	0.110	0.315	0.042	0.701
Positive surgical margins	0.010	0.930	−0.036	0.744	0.066	0.546	−0.280	0.009*	0.088	0.420	−0.029	0.794	0.020	0.857
Biochemical progression	−0.175	0.107	0.193	0.075	0.026	0.810	−0.310	0.004*	−0.110	0.312	−0.078	0.474	0.136	0.213
Survival	−0.218	0.044*	−0.078	0.473	−0.038	0.726	−0.366	0.001*	−0.078	0.476	−0.096	0.380	0.115	0.292

	p-mTOR		IKK		pIKK		IkB		pIkB		NF-kBp50		NF-kBp65
	r	p	r	p	r	p	r	p	r	p	r	p	r	p
Preoperative serum PSA	−0.143	0.190	−0.016	0.884	0.053	0.628	0.077	0.482	−0.108	0.323	0.063	0.567	−0.026	0.812
Pathological Tumor stage	−0.290	0.007*	−0.142	0.193	−0.130	0.233	0.081	0.457	−0.084	0.441	−0.295	0.006*	−0.036	0.745
Clinical Tumor stage	−0.209	0.054	0.075	0.490	0.042	0.704	0.053	0.626	−0.022	0.843	−0.041	0.706	0.111	0.309
Gleason score	−0.072	0.510	0.019	0.863	−0.138	0.206	0.146	0.179	0.111	0.311	−0.032	0.770	−0.016	0.880
Perineural invasion	0.115	0.292	0.068	0.532	0.036	0.742	−0.121	0.266	0.005	0.962	−0.374	0.000*	0.070	0.522
Node involvement	−0.259	0.016*	0.133	0.221	0.110	0.315	0.095	0.383	0.118	0.281	−0.052	0.633	−0.044	0.688
Positive surgical margins	−0.111	0.311	0.099	0.366	0.152	0.164	−0.127	0.244	−0.041	0.709	−0.300	0.005*	0.111	0.311
Biochemical progression	0.002	0.989	−0.014	0.901	0.107	0.327	0.123	0.260	0.024	0.825	−0.085	0.435	0.069	0.530
Survival	−0.028	0.796	0.068	0.532	0.036	0.742	−0.069	0.530	−0.058	0.595	−0.232	0.032*	−0.068	0.531

* Significance values p < 0.05.

The negative expression of NF-kBp50 correlated with pathological stage T, perineural invasion, surgical margins, and survival. The expression of PI3K was inversely related to surgical margins, biochemical progression, and survival.

The negative expression of p-mTOR was correlated with clinical stage T and node involvement. Increased expression of TGF-β was correlated with low survival.

Correlation between established classical markers and intensity of expression

Significant correlation between established classical markers and intensity of expression of TGF-β, TGFBRII, pAKT-Ser, mTOR, pIKK, IkB, pIkB, and NF-kBp65 was not observed (Table 4). In addition, the classic PSA preoperative markers, clinical T stage and Gleason score did not correlate significantly with any of the proteins studied.

Table 4. Correlation between immunohistochemical store of studied proteins and clinic-pathological features.

	TGF-B		TGFBRI		TGFBRII		PI3K		pAKT-Ser		pAKT-Thr		mTOR
	r	p	r	p	r	p	r	p	r	p	r	p	r	p
Preoperative serum PSA	−0.083	0.445	−0.098	0.369	−0.086	0.431	−0.039	0.724	0.004	0.971	−0.016	0.884	0.098	0.372
Pathological Tumor stage	−0.066	0.544	0.060	0.580	0.107	0.327	−0.060	0.586	−0.143	0.190	0.014	0.902	0.013	0.904
Clinical Tumor stage	−0.045	0.679	0.203	0.061	0.112	0.306	0.027	0.805	−0.101	0.356	−0.038	0.725	0.036	0.742
Gleason score	0.001	0.995	0.023	0.833	0.077	0.483	−0.104	0.341	−0.090	0.411	0.046	0.675	−0.010	0.928
Perineural invasion	−0.002	0.983	0.010	0.926	−0.051	0.641	0.044	0.689	0.063	0.564	−0.288	0.007*	0.032	0.767
Node involvement	0.098	0.372	0.021	0.846	−0.056	0.608	0.080	0.464	0.121	0.269	0.121	0.266	−0.005	0.967
Positive surgical margins	0.032	0.770	−0.046	0.673	−0.020	0.856	−0.246	0.022*	−0.003	0.981	−0.016	0.886	−0.179	0.099
Biochemical progression	−0.132	0.226	0.275	0.010*	−0.056	0.607	−0.289	0.007*	−0.083	0.446	−0.070	0.521	0.167	0.124
Survival	−0.194	0.074	−0.086	0.430	−0.095	0.383	−0.365	0.001*	0.007	0.948	−0.098	0.369	0.092	0.399

	p-mTOR		IKK		pIKK		IkB		pIkB		NF-kBp50		NF-kBp65
	r	p	r	p	r	p	r	p	r	p	r	p	r	p
Preoperative serum PSA	−0.078	0.478	−0.002	0.988	0.028	0.768	0.092	0.400	0.028	0.798	0.127	0.245	−0.094	0.390
Pathological Tumor stage	−0.125	0.253	−0.235	0.030*	−0.143	0.190	0.097	0.373	−0.143	0.190	−0.293	0.006*	−0.013	0.903
Clinical Tumor stage	−0.178	0.100	0.059	0.590	0.005	0.961	0. 600	0.585	0.005	0.961	0.060	0.582	0.190	0.081
Gleason score	−0.192	0.076	0.013	0.909	−0.168	0.123	0.100	0.360	−0.168	0.123	−0.012	0.912	−0.031	0.778
Perineural invasion	−0.008	0.940	0.017	0.880	0.028	0.796	−0.103	0.347	0.028	0.796	−0.250	0.021*	0.070	0.520
Node involvement	−0.215	0.047*	0.170	0.118	0.136	0.212	0.118	0.280	0.136	0.212	−0.013	0.903	−0.022	0.844
Positive surgical margins	−0.048	0.660	0.127	0.243	0.088	0.419	−0.115	0.290	0.088	0.419	−0.291	0.007*	0.083	0.446
Biochemical progression	−0.146	0.180	0.010	0.999	0.117	0.282	0.111	0.307	0.117	0.282	−0.061	0.575	0.034	0.759
Survival	−0.088	0.419	0.017	0.880	0.075	0.494	−0.075	0.493	0.075	0.949	−0.271	0.012*	−0.062	0.574

* Significance values p < 0.05.

The decrease in the intensity of NF-kBp50 expression was significantly correlated with the pathological stage T, perineural invasion, surgical margins and increased survival. The intensity of PI3K expression was inversely correlated with surgical margins, biochemical progression and survival.

Inverse correlation was found between IKK and pathological stage T; between pAKT-Thr and the perineural invasion; and, between p-mTOR and node involvement. Only the intensity of TGFBRI expression correlated directly with the biochemical progression.

Correlation of immunohistochemical variables with time to biochemical progression

We evaluated the relationship between expressions of components of the TGF-β/PI3K/AKT/mTOR/NF-kB pathway and time to biochemical progression using the Kaplan–Meier method combined with the log-rank test (Figure 2), showed that patients positive for PI3K expression had longer biochemical progression times than negative ones. When we studied the intensity of PI3K expression, we observed that high positives had progression times twice as fast as low positives.

Figure 2. Significant Kaplan–Meier analyses of immunohistochemical expression for time to biochemical progression. Vertical tick marks represent censored observations. Statistical significance was evaluated by long-rank test (p values).

The intensity of TGFBRI expression showed that negative patients had biochemical progression times longer than positive ones and that low positives had times three times shorter than high positives.

Univariate analysis for to biochemical progression of expression of components TGF-β/PI3K/AKT/mTOR/NF-kB pathway

The univariate Cox proportional hazards regression model (Table 5) confirmed the results obtained for PI3K expression. In the case of intensity, it was confirmed for TGFBRI and PI3K.

Table 5. Univariate Cox proportional hazard regression analysis for biochemical progression according to either expression immunostaining (expression) and immunostaining score (intensity).

Expression	p value	Hazard ratio (95% CI)	Intensity	p value	Hazard ratio (95% CI)
TGFB	0.110	0.551 (0.265–1.146)	TGFB	0.231	0.760 (0.486–1.191)
TGFBRI	0.077	2.156 (0.920–5.051)	TGFBRI	0.011	1.841 (1.151–2.947)*
TGFBRII	0.698	1.232 (0.429–3.542)	TGFBRII	0799	0.941 (0.589–1.504)
PI3K	0.000	0.102 (0.031–0.328)*	PI3K	0.000	0.347 (0.195–0.616)*
pAKT-Ser	0.275	0.555 (0.193–1.598)	pAKT-Ser	0.340	0.776 (0.461–1.306)
pAKT-Thr	0.571	0.782 (0.334–1.831)	pAKT-Thr	0.637	0.903 (0.593–1.376)
mTOR	0.452	21.48 (0.007–63729)	mTOR	0.120	2.162 (0.818–5.712)
p-mTOR	0.858	0.833 (0.113–6.129)	p-mTOR	0.247	0.681 (0.356–1.305)
IKK	0.843	0.921 (0.408–2.079)	IKK	0.854	0.962 (0.636–1.455)
pIKK	0.451	1.501 (0.522–4.314)	pIKK	0.417	1.237 (0.740–2.069)
IkB	0.213	1.611 (0.761–3.412)	IkB	0.264	1.243 (0.849–1.820)
pIkB	0.881	1.071 (0.437–2.631)	pIkB	0.779	1.067 (0.678–1.680)
NF-kB p50	0.329	0.639 (0.260–1.571	NF-kB p50	0.370	0.813 (0.517–1.279)
NF-kB p65	0.601	1.326 (0.461–3.811)	NF-kB p65	0.744	1.087 (0.660–1.790)

* Significance values p < 0.05.

Multivariate cox proportional hazard regression analysis for time to biochemical progression of components TGF-β/PI3K/AKT/mTOR/NF-kB pathway

The most commonly used classical markers as prognostic markers of biochemical relapse, after prostatectomy are PSA levels, pathological and clinical T stages, Gleason score, perineural invasion, lymph node involvement, surgical margin, and overall survival. These classic markers were studied in this cohort of patients using the Kaplan–Meier survival curves. Pathological T stage, Gleason score, and lymph nodes were able to predict biochemical progression, but not preoperative serum PSA, perineural invasion, positive surgical margins and survival.

When analyzing multivariate Cox, PI3K expression remained an independent prognostic factor for the effects of Gleason score and involvement of lymph nodes (Table 6). The intensity of TGFBRI expression was maintained as an independent prognostic factor after the adjustment of the pathological stage T and the Gleason score and intensity of IP3K with node involvement.

Table 6. Multivariate Cox proportional hazard regression analysis for biochemical progression for immunohistochemical expression.

	p value	Hazard ratio (95% CI)
Pathological Tumor stage	0.013	2.083 (1.166–3.721)
Gleason score	0.007	2.990 (1.340–6.669)
TGFBRI intensity	0.013*	1.843 (1.138–2.985)*
Gleason score	0.056	2.215 (0.980–5.006)
Node involvenment	0.002	6.322 (1.996–20.02)
PI3K expression	0.000*	0.085 (0.025–2.85)*
Node involvenment	0.000	7.963 (2.584–24.538)
PI3K intensity	0.000*	0.079 (0.024–0.265)*

* Significance values p < 0.05.

CI: confidence interval.

Discussion

In the last decades, the researchers direct their efforts in understanding different research pathways activated by growth factors and that could improve the cancer prediction. In previous studies of our group suggest that TGF-β was associated with biochemical progression [6]. Our group also suggests the effect of PI3K/AKT/mTOR pathway as an important therapeutic target in PC after prostatectomy [9]. Since TGF-B has been activated the PI3K/AKT/mTOR pathway and NF-kB, we try to describe the relationship of this whole route in the progression of prostate cancer.

Elevated concentrations of TGF-β, both in serum and in plasma, have been associated with advanced stage cancer, metastasis and poor clinical outcomes in breast, prostate, pancreatic, and renal cancer [30]. In addition, the presence of TGF-β in the semen is probably related to the tumor stage [31]. In the present study, TGF-β was related to survival. When evaluating the receptors only the expression TGFBRI was related to the biochemical progression, these results were similar to others described in previous works of our group when relating the immunostaining of TGFBRI in the stroma with the biochemical progression [6].

In previous works of our group [9] as well as in this investigation, the expression PI3K was significantly inversely related to surgical margins, biochemical progression, and survival. In the tumor, PI3K has been linked to proliferation, survival, migration, and invasion [32,33]. In this work, the expression of AKT-Thr was inversely related to the perineural invasion, the same results were obtained in previous works [9]. The sustained activation of AKT allows the increase of cell survival, due to the interruption of apoptosis mediated by TGF-β [34]. Increases in AKT signaling have been correlated with a poor clinical outcome in prostate cancer [35]. This is because the collaboration of the autocrine signaling TGF-β and the activated PI3K/AKT pathway plays a key role in cancer progression, causing the change in TGF-β/SMAD (Small Mothers Against Decapentaplegic) signaling from its tumor suppressor to its mode tumor promoter [36]. Also, in prostate cancer cells, oncogenic ETS/AP-1 (erythroblastosis virus transcription factor/activator protein-1 transcription factor) binding sequences transcription factors require AKT signaling to activate a gene expression program [37].

We found an inverse and significant relationship between p-mTOR with pathological stage T and involvement of lymph nodes. In previous studies, we had not found a relation of mTOR with any classical marker and p-mTOR only with pathological stage T and involvement of lymph nodes. The specific deletion of mTOR tissue in the mouse prostate inhibits tumor formation driven by the loss of PTEN without disrupting normal prostate tissue [38]. In addition, mTOR catalytic inhibitors can achieve antileukemic effects in doses that preserve the function of normal bone marrow and peripheral lymphocytes [39]. Aberrant activation of mTOR signaling has been detected in chemo, radius and/or hormone resistant to cord, breast cancer, pancreatic cancer and prostate cancer cells [40,41]. In models of CaP cells resistant to castration, mTOR was able to regulate transcriptionally the metabolic gene programs in the absence of androgens, highlighting a novel mechanism of castration resistance to maintain cellular metabolism even without a functional RA [42].

In previous studies of our group, as in this study, we found increased expression of IKK, pIKK, IkB, pIkB, NF-kBp50, and NF-kBp65 in prostate cancer samples with respect to normal prostates [43]. At the same time, we found a significant negative relationship between IKK and pathological stage T; between NF-kBp50 and stage T pathological, perineural invasion, surgical margins, and survival [43]. Numerous studies link IKK and its target NF-kB with inflammatory diseases and its hypothetical role as a tumor promoter in an environment of chronic inflammation [44,45]. Inhibition of IKK may represent a strategy to combat late-stage prostate cancer, at least in part, by regulating the inflammatory environment, while IKKB may be more relevant to the early stages of the disease [46]. Other authors have been more specific in pointing out that the IKKα kinase is a key regulator of oncogenesis and driver of the metastatic process [47,48]. The mechanism by which IKKα affects the progression of cancer is not yet completely clears [49]. Some investigators indicate that in many cancers including prostate cancer, IKK/NF-kB signaling may be impaired or active due to the alteration of other oncogenic pathways such as AKT or the loss of PTEN tumor suppressor function [50,51].

The inhibition of NF-kB has been studied in castration-resistant cancer cells (PC3 and DU45). In these cells, the effect of simvastatin together with an inhibitor of NF-kB (CAPE) mediates apoptosis by decreasing growth cellular [52,53]. Inhibition of NF-kB in castration-resistant (PC3, 22RV1) and androgen-sensitive prostate cancer cells (LNCaP) decreases the expression of androgen receptors through ubiquitin-mediated proteosome degradation [54].

Conclusions

Our results confirm the activation of the TGF-β/PI3K/AKT/mTOR/NF-kB transduction pathway in prostate cancer and that some of its components such as TGF-β, PI3K, AKT-Thr, mTOR, and NF-kBp50 are related with established clinical-pathological markers. TGFBR1 and PI3K could be used as useful biomarkers for early diagnosis and prognoses for biochemical recurrence in prostate cancer after radical prostatectomy.

Ethics approval and consent to participate

All the procedures were examined and approved by the Joint committee of the University of Alcalá and Principe de Asturias Hospital Ethics Committees (PI13/1801; 2013/003/20130214) and were in accordance with the ethical standards of the Committee for Human Experimentation, with the Helsinki Declaration of 1975 (revised in Tokyo 2004) and the Committee on Publication Ethics guidelines. This study was performed with the written consent of the patients or their relatives. All pathological, clinical or personal data were anonymized and separated from any personal identifiers.

Data availability

All data generated or analyzed during this study are included in this published article.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by grants from the “Instituto de Salud Carlos III” [PI13/01801] of Spain.