Tumor suppressor genes and their underlying interactions in paclitaxel resistance in cancer therapy

To comprehensively collect all of the TSGs related to PTX resistance, we searched the PubMed online database and google web site, followed by an advanced search using the terms “paclitaxel response” or “paclitaxel sensitive” and “drug resistance” or “chemotherapy resistance,” and “cancer” or “carcinoma,” and “tumor suppressor genes” or “negative regulated protein” or “antioncogene” This search identified 22 TSGs including breast cancer 1 (BRCA1), tumor protein p53 (TP53), phosphatase and tension homolog (PTEN), adenomatous polyposis coli (APC), cyclin-dependent kinase inhibitor 1A (CDKN1A), cyclin-dependent kinase inhibitor 2A (CDKN2A), high in normal-1 (HIN-1), ras association domain-containing protein 1 (RASSF1), yes-associated protein 1 (YAP), inhibitor of growth 4 (ING4), polo-like kinase 2 (PLK2), f-box and WD repeat domain containing 7 (FBW7), zinc finger MYND type containing 10 (BLU), leucine zipper tumor suppressor 1 (LZTS1), re-1 silencing transcription factor (REST), fas-associated death domain protein (FADD), programmed cell death 4 (PDCD4), transforming growth factor-β-induced (TGFBI), inhibitor of growth 1 (ING1), bcl-2-associated X protein (Bax), PIN2/TRF1 interacting telomerase inhibitor 1 (PinX1) and one putative tumor suppressor gene, FERM domain-containing protein 6 (hEx), which contributed to PTX-resistance in cancer. The status, regulation manner, pathway and cancer type involved in PTX-resistance have been summarized, as shown in Table 1.

Tumor suppressor BRCA1 is involved in several cellular functions including DNA damage repair, cell cycle checkpoint activation and transcription [11]. Several preclinical studies indicated that BRCA1 might be an important determinant of response to PTX-based chemotherapy. It was shown that reconstitution of exogenous BRCA1 in the BRCA1-mutant HCC1937 breast cancer cell line resulted in enhanced sensitivity to PTX [12]. In accordance, low BRCA1 mRNA expression in ovarian cancer cell lines resulted in decreased and increased apoptotic response to PTX and platinum respectively and PTX-sensitive human brain and neck squamous cell carcinoma (HNSCC) with acquired cisplatin resistance had high expression of BRCA1 [13, 14]. In order to investigate the underlying PTX-resistance mechanisms conferred by loss of BRCA1, Chabalier et al. reduced BRCA1 protein levels by using small interfering RNA (siRNA) in MCF7 breast cancer cells resulted in PTX resistance through premature inactivation of spindle checkpoint [15]. Sung et al. found that BRCA1 knockdown conferred A549 cells resistance to PTX and sensitivity to cisplatin through improving microtubule dynamics which prevented the formation of stable microtubule for caspase-8 accumulation of PTX induced apoptosis [16]. A further study suggested that BRCA1 might represent an important mediator of the PTX stress-response dependent c-Jun N-terminal kinase/stress-activated protein kinase (JNK/SAPK) or p38/mitogen-activated protein kinase (p38/MAPK) pathway [17]. Taken together, these studies provided evidence that BRCA1 mutation or reduced expression could predict the response to PTX–based chemotherapy. BRCA1 deficiency led to increased microtubule dynamics, impaired cell cycle checkpoint and signaling pathway which rendered less sensitivity to PTX-induced apoptosis. Here we consider that BRCA1 may become a molecular marker to predict the PTX resistance.

TP53 is one of the earliest detected tumor suppressor genes and the most frequently mutated gene in carcinoma. More than half of the TP53 mutations found in cancers lead to loss of function. Functional p53 participates in various cellular processes including cell cycle progression, cell motility, aging, apoptosis, genetic instability, DNA repair, anti-angiogenesis and cell metabolism [18]. TP53 gene mutation status has recently been shown to be correlated to PTX-based therapy and prognosis [19‐21]. It was also found that an augmented concentration of intracellular p53 protein sensitized three non-small-cell lung carcinoma (NSCLC) cell lines to PTX [22]. p53 upregulated modulator of apoptosis (PUMA) is an important regulator of apoptosis and is involved in drug resistance [23]. It was demonstrated that PUMA was downregulated in PTX-resistant ovarian cell line SKOV3/PTX, and delivery of p53 into SKOV3/PTX could upregulated the expression of PUMA and restored the apoptotic response to PTX [24]. TP53 hot spot mutation (TP53-m273) increased multidrug resistance protein 1 (MDR1, regulating efflux of PTX and doxorubicin) expression and resistance to PTX [25, 26]. In addition, studies suggested that some regulatory factors depended on p53 related pathway may mediated PTX resistance. For example, the up-regulation of inhibitor of apoptosis-stimulating protein of p53 (iASPP), a p53 suppression factor, has been found to affect PTX sensitivity in ovarian cancer by inhibiting both mitotic catastrophe and apoptosis [27]. Astrin, a protein localized with mitotic spindles at M phase, silencing of astrin triggered a p53-dependent apoptotic pathway and induced Hela cells sensitive to PTX [28].

PTEN, is a negative regulator of the phosphatidylinositol 3-kinase/protein kinase B (PI3k/Akt) signaling pathway. Its dysfunction mutation results in reduced dephosphorylation of phosphatidylinositol 3, 4, 5-triphosphate (PIP3), further increasing cell survival, cell migration, cell size and cell proliferation [29]. Recently, reports mainly concentrate on the role of PTEN in the response of human cancer cells to anti-cancer drugs and in multiple drug resistance (MDR) reversion [30‐33]. Several reports showed that PTEN was involved in PTX resistance. Cyclin B1 plays a key role in G2/M transition. Ou et al. detected suppressing of cyclin B1 protein sensitized esophageal squamous cell carcinoma (ESCC) cells to PTX-induced apoptosis through the PTEN/PI3 k pathway [34]. Overexpression of microRNA 22 (miR-22) reversed PTX-induced cytotoxicity and this function was mediated by the regulation of PTEN levels in TP53 negative colon cell line [35]. Although PTEN is not the primary target of PTX resistance, evidences showed that its regulator can be an important target, such as suppression of cyclin B1, miR-22 or combining with inhibitor of Akt could be an attractive strategy for PTX therapy.

Tumor suppressor gene APC is most commonly mutated and deleted in colorectal cancers, as well as many other epithelial cancers like breast, gastric and lung cancer. The best-known function of the APC protein is the regulation of the Wnt signaling cascade through down-regulation of β-catenin can modulate cell cycle progression, however, APC has many Wnt independent roles, such as microtubule dynamic, cytoskeletal organization and cell adhesion [36, 37]. Since PTX is to interfere with microtubule protein stability, the interaction between APC and PTX has been explored. Monica et al. showed loss of APC in breast cancer cells from mouse mammary tumor virus promoter-polyoma middle T-antigen (MMTV-PyMT) mouse lead to increased expression of MDR1 after treatment with cisplatin and PTX [38]. It has been demonstrated that APC expression is regulated by a microRNA 135a (miR-135a) [39]. So it is not surprising that miR-135a is shown to be involved in PTX resistance by downregulation of APC [40]. Moreover, Ling et al. found APC-deficient cancer cells defect in mitotic spindle checkpoint and in cell–cell adhesion and were more resistant to PTX [41, 42]. Consequently, APC deficiency impairs the PTX sensitivity of cancer cells by interfering with the mitotic spindle checkpoint and decreasing apoptosis.

Loss of cell cycle control promotes tumorigenesis, key regulators of the cell cycle are a family of serine/threonine kinases: cyclin-dependent kinases (CDKs). CDKs act at different stages of the cell cycle and are responsible for the transition from one cell cycle phase to the next [43]. Endogenous cyclin-dependent kinase inhibitors (CKIs) are negative regulators of CDKs [44]. There are two families of CKIs: the INK4 families, consisted of p16, p15, p18 and p19 which can inhibit the complex of cyclin dependent kinase 4/6 (CDK4/6) and cyclin complex activities. And the CIP/KIP families include p21, p27 and p57, regulate border CDKs [45]. Recently, evidences have showed CKIs family members involved in PTX resistance in human cancers. p21, is required to maintain the G2 arrest after DNA damage [46], the level of p21 expression has been known to play an important role in determining sensitivity of tumor cells to PTX [47], and a remarkable induction of p21 in A375P cells after treatment of PTX and apoptosis induction after mitotic arrest with PTX. However, PTX lightly increased the levels of p21 in A375P/Mdr cells, which exhibited strong resistance to PTX [48]. p16, mainly inhibits CDK4 activity, the loss of p16 expression reduced the response of breast cancer cells to PTX by conferring cancer stem cell properties and the tumorsphere formation was not significantly enhanced [49], those results indicated that CKIs affect PTX efficacy mainly through the cell cycle regulation.

The Hippo signaling pathway, which regulates cell proliferation and apoptosis, is a highly conserved signaling pathway first discovered in Drosophila cells. It also exists in mammals and controls organ size, cell proliferation and apoptosis. The main function of Hippo signaling pathway is to phosphorylate transcriptional co-activator PDZ-binding motif (TAZ) and YAP, preventing them from entering the nucleus and promoting gene transcription which induces cell proliferation, metastasis and invasion [50]. Recent discoveries have identified the Hippo signaling pathway as a new target for cancer chemotherapy resistance [51]. For example, hEx, one of the Hippo upstream signal input factors, a putative tumor suppressor gene, overexpression of hEx dramatically inhibited breast cancer cell proliferation and sensitivity to PTX [8]. RASSF1A, a member of the RASSF1 family, is a downstream regulator of Hippo, there are approximately 50 % of ovarian tumors harbor hypermethylation of RASSF1 [52], investigations shown that overexpression of RASSFIA could increase stabilization of microtubules then restore PTX sensitivity [53]. YAP, a nuclear effector of Hippo, has been shown exist in many pathways except in Hippo, it is critical for DNA damage in breast cancer cells as well as in certain types of neuronal apoptosis [54]. It acted as a tumor suppressor in breast cancer and its silencing could induce normal breast epithelia more resistance to PTX effect on cell cycle but not apoptosis [55].

In addition to the TSGs mentioned above, abnormity of ING4, Bax, HIN-1, PLK2, FBW7, LZTS1, BLU, TGFBI, REST, FADD, PDCD4, ING1 and PinX1 have also been found to mediate PTX resistance in some experiments. The protein level of ING4 was sharply decreased in PTX-resistance lung cancer cells. In contrast, overexpression of ING4 protein could induce apoptosis and G2/M arrest by decreasing B-cell CLL/lymphoma 2 (Bcl-2)/Bax ratio then reversed PTX-resistance [56]. Bax is a proapoptotic Bcl-2 family member that plays a key role in induction of mitochondrial dependent apoptosis. Study found there was an increase of Bcl-2/Bax ratio in PTX-resistant breast cancer cell lines, high ratio reduced the PTX-induced apoptosis in breast cancer and ovarian cancer cells [57]. Hypermethylation downregulated the expression of HIN-1 and weakened the sensitivity to PTX through the PI3k/Akt pathway [9]. Hypermethylation of PLK2 reduced ovarian cancer cells sensitivity to PTX, accompanied by reduced G2-M arrest and apoptosis [58]. Low protein expression of LZTS1 showed little response in patients who received PTX-based chemotherapy in ovarian carcinoma and breast cancer patients, it was a worse prognosis patients outcome [59, 60], previous study by generating LZTS1 knockout mice, detected accelerate mitotic progression resistance to PTX-induced M phase arrest by decreasing CDK1 activity [61], indicated cell cycle distribution may be involved in the above two human cancer. Ovarian and colon cancer cells which harbored mutant FBW7 were more resistant to PTX, functional FBW7 is required to degrade myeloid cell leukemia 1 expression by a ubiquitin ligase SKP1–cullin-1–F-box complex that contains FBW7 [62]. Ovarian cancer patients with methylated BLU had significantly shorter progression free survival, in vitro, BLU could decrease the Bcl-2/Bax ratio in ovarian cancer cells when encountered with PTX [63, 64]. TGFBI acts as a tumor suppressor in lung cancer [65]. Irigoyen et al. identified a strong association between elevated TGFBI expression and the response to chemotherapy, TGFBI mediated the susceptibility of NSCLC cells to PTX and this may be the result of direct TGFBI induction of cell apoptosis through the binding of its proteolytic fragments to the β3 integrin, the same phenomenon was proven in ovarian cancer [66, 67]. REST directly regulates Akt2, loss of REST leads to a de-regulation of Akt phosphorylation [68]. Tubulin beta 3 class III (TUBB3) was a biomarker of the resistance of chemotherapies [69]. Gao et al. found REST might suppress the expression of TUBB3 to sensitize ovarian cancer cells to PTX by activating the PI3k/Akt pathway [70]. Phosphorylation of FADD affected both upstream and downstream of the JNK/SAPK pathway, which was critical for sensitivity to PTX-induced apoptosis [71‐73], and PDCD4 mediated PTX sensitivity through interacting mitotic exit regulation axis, upregulation of microRNA 182 (miR-182) accelerated cell cycle process and enhanced chemo-resistance of ovarian cancer cells to PTX through negatively regulating PDCD4 [74, 75]. P33^ING1, one of the ING1 gene products, could enhance PTX-induced apoptosis in human osteosarcoma U2OS cells by p53-dependent pathway and its target genes p21 and Bax were increased [76]. In cervical squamous cell carcinomas, the expression of PinX1 in patients was significantly associated with the response of the combination of PTX and cisplatin chemotherapy. In vitro, knockdown of PinX1 could dramatically enhance PTX effects, whereas the augment of PinX1 levels substantially enhanced the G2 phase cells through influencing spindle assembly checkpoint [77].

Summarily, the pre-transcriptional (epigenetic/genetic), transcriptional and post-transcriptional changes of TSGs contribute to PTX resistance in cancer, which may lead to new treatment methods to overcome drug resistance. Actually, the reversion of epigenetic changes of DNA and gene transfer skills (gene therapy) has already proved to be effective in reversing PTX resistance. For example, 5-aza-2-deoxycytidine (a demethylation agent) reversed the sensitivity to PTX treatment in breast cancer and ovarian cancer cells [9, 78]. Exogenous increased levels of p53 significantly improved the sensitivity of PTX providing a basis for gene therapy [22]. For another example, the CDK4/6 inhibitor was shown to downregulate p16/cyclin D1/CDK4/(retinoblastoma protein)Rb signaling pathway and enhance the cytotoxicity of PTX for KRAS mutation-positive lung adenocarcinoma cells [79]. These results encouraged further studies on TSGs associated with PTX resistance in cancers. Moreover, drug resistance was rarely induced by single gene, it was almost caused by two or more genes. For example, tumor suppressor p33^ING1 markedly increased PTX-induced growth inhibition and apoptosis in TP53-wild cells, but not in TP53-mutant cells [76]. Twenty two genes were involved in the regulation of PTX resistance in cancers through certain pathways, particularly through cell cycle and apoptosis (Table 1).

Bioinformatics analysis has been widely used in nature and life sciences, and it is a feasible and valuable method for gene/protein function prediction. Numerous networks of molecular interactions have made it possible to study gene/protein function using online databases [80]. GeneMANIA is a web-based interface for prediction of gene/protein function on the basis of multiple networks derived from different proteomics and genomic data, and it is fast enough to predict gene/protein function with a significant accuracy rate [81]. The protein interactions of the 22 genes were analyzed using GeneMANIA. The co-localization, co-expression, pathway, shared protein domains and genetic, physical and predictive interactions of the 22 TSGs were shown in Figs. 1, 2, 3, 4, 5, 6, 7 (Genes/proteins are depicted as colored circles and experimentally detected relationships between genes/proteins as connecting lines. Black circles are the 22 TSGs, gray circles are other genes/protein related to the 22 TSGs). In detail, BRCA1 has similar expression level with CDKN2A, participates in the same pathway with TP53, and has physical interactions with CDKN2A and TP53. TP53 has similar expression level with FADD, PDCD4, CDKN1A and Bax, participates in the same pathway with BRCA1, APC, CDKN1A, PTEN and Bax, and has physical interactions with BRCA1, ING1, CDKN1A, ING4, Bax and CDKN2A. PTEN has similar expression level with PDCD4, FBW7, TP53, PLK2, APC and TGFBI, participates in the same pathway with TP53, and has physical interactions with TP53, FBW7. APC has co-localization with FBW7 and the similar expression level with FBW7, hEx, PTEN and REST, and participates in the same pathway with TP53. CDKN1A has the similar expression level with TGFBI, ING1, FADD, TP53, CDKN2A and PLK2, and the same protein domain and physical interaction with TP53. CDKN2A has the similar expression level with CDKN1A, CDKN2A, and physical interactions with ING1, BRCA1, TP53, ING4. hEx has the similar expression level with FBW7, ING1, RASSF1, PLK2, REST, PDCD4 and TGFBI. RASSF1 has the similar expression level with BLU, ING4 and TGFBI. YAP has the similar expression level with TGFBI, PinX1, ING1, ING4 and FADD. ING4 has the similar expression level with YAP, RASSF1, FADD, PDCD4, Bax and PLK2, the same protein domain with ING1 and physical interactions with TP53 and CDKN2A. HIN-1 has the similar expression level with FADD, REST, LZTS1 and BLU. PLK2 has the same protein location with TGFBI, and the similar expression level with TGFBI, PTEN, LZTS1, ING4, CDKN1A, ING1, hEx and FBW7. LZTS1 has the similar expression level with PLK2, FBW7 and HIN-1. FBW7 has the same protein location with APC, interacts with REST and PTEN, and has the similar expression level with PLK2, APC, TGFBI, LZTS1, PDCD4, CDKN1A, PTEN, REST and Bax. BLU has the similar expression level with RASSF1, PDCD4, HIN-1 and REST. TGFBI has the similar expression level with FBW7, hEx, Bax, PTEN, CDKN1A, YAP and RASSF1. REST has the similar expression with BLU, APC, PinX1, FBW7, FADD and HIN-1. FADD has the similar expression of ING1, PinX1, CDKN1A, ING4, HIN-1, REST, YAP, Bax and TP53. PDCD4 has the similar expression with FBW7, TP53, ING4, BLU and PTEN. Bax has the similar expression with ING4, PinX1, TP53, CDKN1A, ING1, FBW7, FADD and TGFBI, and participates in the same pathway with TP53. PinX1 has the similar expression level with YAP, REST, Bax and FADD. ING1 shares the same protein domain with ING4 and interacts with TP53 and CDKN2A. The cross-interaction of the 22 TSGs demonstrates that these genes may contribute to PTX resistance as a group.

The molecular function of the 22 TSGs may be predicted by using GeneMANIA network. By analysis, there were four summarized functions to be possibly close to chemotherapy resistance (Table 2). The cell cycle-related function covered 12 of the 22 TSGs and additional 10 genes. The cell apoptosis-related function covered eight TSGs and additional seven genes. The protein ubiquitination-related function covered six TSGs and additional five genes. The cell growth-related function covered seven TSGs and additional four genes. Since the cell cycle-related function was annotated with the highest false discovery rate (FDR), the regulation function of TSGs on cell cycle was considered more closely related to chemotherapy resistance than other three pathway functions.

In the network, 20 other related genes/proteins that mediate the relationship of the 22 TSGs. They have interactions with the 22 candidate genes, and participate in the similar biological process which the 22 genes are involved (Table 2), which indicates that they may be related to PTX or other drug resistance in cancers. Some of them had been studied in drug resistance. For example, tumor protein 63 (TP63), a TP53 family protein, which expressed a variety of isoforms. DeltaNp63alpha belonged to the members of the N-terminally truncated (DeltaN) p63 subfamily, it can trigger anti-apoptotic related pathway result to chemo-resistance in hepatocellular carcinoma [82]. Ovarian cancer cell line with acquired resistance to carboplatin revealed low levels of the gene cyclin-dependent kinase inhibitor 1C (CDKN1C), and demethylation agent can reverse the silencing of CDKN1C and increased the apoptotic response to carboplatin [83]. Bcl-2, is specifically considered as an important anti-apoptotic protein and classified as an oncogene, the expression of Bcl-2 can affect PTX induced apoptosis [84]. The Bcl-2/Bax ratio has shown a significant increase in PTX-resistance cancer cells [56, 57]. Hairy/enhancer-of-split related with YRPW motif protein 1 (HEY1), cyclin-dependent kinase inhibitor 1B (CDKN1B) and fas cell surface death receptor (Fas) acted as a part of signaling pathways involved in drug resistance [85]. Death-associated protein 6 (DAXX) has been shown to regulate PTX-sensitivity in tumor [86]. CDK4 and CDK2 are not only involved in the formation of tumor resistance but also the inhibitors of cyclin-dependent kinase which has been the anti-cancer agent used in clinic [87].