Computational assessment of SKA1 as a potential cancer biomarker

Introduction

Mitosis, which is a critical process in cell cycle, includes separation of replicated chromosomes into daughter cells. Proper spindle formation, depolymerization and connection of chromosomes to spindle microtubules are crucial for chromosome segregation in mitosis [1]. Mutations and aberrant expression of proteins, which function in microtubule dynamics and stability, have been associated with tumorigenesis, metastasis and response of cells to chemotherapeutic agents [2], [3]. In addition, some of those proteins and several centromere and kinetochore proteins have been proposed as markers for different cancers [4]. Involvement and prognostic value of these proteins in various cancers emphasize the importance of exploring additional proteins, which function in similar mechanisms, as cancer markers.

Chromosome segregation during cell division is achieved by the proper attachment of kinetochores, protein complexes on centromeres of the chromosomes to spindle microtubules [5], [6]. During microtubule polymerization and depolymerization at mitosis, kinetochores stay attached to microtubules and chromosomes are pulled apart from the kinetochore regions [1], [7]. In addition kinetochores regulate the order of mitotic events [7].

In humans, SKA complex, which consists of SKA1 (also known as C18orf24), SKA2 (also known as FAM33A) and SKA3 (also known as RAMA1 and C13orf3), is essential for the formation of proper kinetochore–microtubule attachment and timely mitotic progression [8], [9]. The microtubule binding activity of SKA1 is required for the interaction of kinetochores and microtubules, and kinetochore movement [10]. Depletion of SKA1 protein or inhibition of SKA1 microtubule binding activity causes a mitotic delay in human cells [9], [10].

In different cancers, SKA1 expression has been shown to affect cell proliferation, cell invasion, apoptosis and cell cycle progression of cancer cells [11], [12], [13], [14], [15]. In addition to cell proliferation and invasion, association of SKA1 expression with metastasis and cancer progression has also been reported in non-small cell lung carcinoma cells [15].

Here, the occurrence of SKA1 mutations, CNVs and aberrant expression, as upregulation or downregulation, have been shown in several cancers and the possibility of SKA1 as a biomarker for predicting patient prognosis has been investigated by using bioinformatics approaches. Furthermore, in order to detect the deleterious mutations for SKA1 protein function and structure, missense mutations were selected and deleterious mutations have been predicted by using in silico prediction methods. The localization of amino acids, which carry the predicted deleterious mutations, was shown on 3D SKA1 structure.

Materials and methods

Results

The effect of missense mutations on SKA1 protein

The effect of 32 missense mutations on SKA1 protein was assessed by using PredictSNP1.0, which classifies the effect of missense mutations as neutral or deleterious. Among the 32 missense mutations, 15 mutations were predicted as deleterious by PredictSNP1.0 (Table 1).

Table 1:

The prediction and expected accuracy results of 15 missense SKA1 mutations predicted as deleterious by PredictSNP.

AA Mutation	PredictSNP
p.R27I	Deleterious	0.61
p.L70F	Deleterious	0.61
p.D83Y	Deleterious	0.65
p.E144G	Deleterious	0.76
p.F145L	Deleterious	0.51
p.P149L	Deleterious	0.87
p.S185Y	Deleterious	0.72
p.S188C	Deleterious	0.55
p.S188F	Deleterious	0.61
p.S188Y	Deleterious	0.61
p.R208C	Deleterious	0.65
p.R239W	Deleterious	0.76
p.R245Q	Deleterious	0.79
p.G246E	Deleterious	0.87
p.R251C	Deleterious	0.87

The localization of 15 missense mutations, which were predicted as deleterious by PredictSNP, was then investigated on the crystal structure of SKA1, which is a 255 amino acid protein. While, the N terminal region of SKA1 (residues 1–130), which includes coiled-coil region, is sufficient for the formation of SKA complex and to interact with SKA2 and SKA3 [8], the C terminal region of SKA1 (residues 133–255) is the main microtubule-binding region for the SKA complex [1]. Of the 32 missense mutations, which were given to PredictSNP, 12 missense mutations were scattered between 1 and 130 amino acids and 20 missense mutations were between residues 133–255. After prediction of deleterious mutations, only three mutations; R27I, L70F and D83Y, in the N terminal region were predicted as deleterious (Figure 1A, B), while 12 of the deleterious mutations were located in the microtubule-binding region of SKA1 (Figure 2A, B).

Figure 1:

Localization of 3 predicted deleterious mutations on N terminal region of SKA1.

(A) Ten copies of SKA1₁₋₉₁, full length SKA2 and SKA3₁₋₁₀₁ form an oligomeric assembly (PDB: 4AJ5, [5]). The N terminal SKA1 chains and the three residues, which carry predicted deleterious mutations (ball and stick presentation), were displayed on SKA core complex. (B) The three predicted deleterious mutations (ball and stick presentation) were shown on the SKA1 chain

Figure 2:

Localization of 12 predicted deleterious mutations on microtubule-binding region of SKA1.

(A) The residues, which carry predicted deleterious mutations (stick presentation) were shown on the chain A (right) and B (left) on the 3D model of microtubule binding domain of SKA1 (PDB: 4C9Y, [1]). (B) The mutated residues were shown in stick presentation on chain B.

Survival analysis in cancer patients

SKA1 is represented with a single probe id, which is 217640_x_at, in the Kaplan-Meier plotter mRNA gene chip section (on the date of April 2019). According to Kaplan-Meier plotter analysis, breast, lung and gastric cancer patients with high SKA1 expression showed poor overall survival compared to patients with low SKA1 expression, while SKA1 expression showed no association with overall survival of ovarian cancer patients (Figure 3 and Table 3). The results also showed that high expression of SKA1 was significantly associated with a relapse free survival (RFS), distant metastasis free survival (DMFS) and post progression survival (PPS) in breast cancer patients (Table 3). High SKA1 expression also significantly associated with PPS and first progression (FP) in lung cancer patients and with FP in gastric cancer patients (Table 3).

Figure 3:

Correlation of SKA1 expression with overall survival.

Kaplan-Meier survival plots showing overall survival in (A) breast, (B) ovarian, (C) lung and (D) gastric cancers.

Table 3:

Correlation of SKA1 with survival outcomes in cancer patients.

Cancer Type	Survival outcome	Number of patients	Cutoff value	HR	95% CI	Log rank p-value
Breast Cancer	OS	1402	42	1.6	1.23–2.09	a4.3E-04
	RFS	3951	50	1.53	1.36–1.72	a5.5E-13
	DMFS	1746	48	1.79	1.43–2.23	a2.2E-07
	PPS	414	44	1.42	1.06–1.91	a1.9E-02
Ovarian Cancer	PFS	1435	94	1.08	0.95–1.23	2.4E-01
	OS	1656	74	1.12	0.99–1.28	7.9E-02
	PPS	782	59	1.19	0.99–1.43	6.1E-02
Lung Cancer	FP	982	44	2.08	1.7–2.53	a1.4E-13
	OS	1926	31	2.53	2.11–3.03	a<1E-16
	PPS	344	43	1.81	1.38–2.37	a1.1E-05
Gastric Cancer	FP	359	106	1.67	1.31–2.14	a3.6E-05
	OS	593	116	1.4	1.14–1.73	a1.5E-03

1. HR, hazard ratio; CI, confidence interval; OS, overall survival; RFS, relapse free survival; DMFS, distant metastasis free survival; PPS, post progression survival; PFS, progression-free survival; FP, first progression. ^ap<0.05.

Discussion

SKA1, as an essential member of SKA complex, is an important part of cell division, carries the major microtubule binding part of the SKA complex and is required for proper function of the complex [1]. SKA1 is critical for the formation of the robust interactions between kinetochores and microtubules, proper kinetochore movement and proper mitotic progression [1], [9], [10]. In addition to SKA2 and SKA3, SKA1 interacts with other proteins, such as PP1 [7], Aurora B and Hec1 [38]. As a necessary part of mitosis, SKA1 expression affects cell proliferation, apoptosis and cell cycle progression of cancer cells [11], [12], [13], [14], which highlights the potential of SKA1 as a biomarker for human cancers. From this point of view, in this study genomic and trancriptomic alterations of SKA1 have been investigated in a wide range of cancers. Genomic alterations, particularly missense mutations, which cause production of an altered amino acid in the protein, were selected and the deleterious missense mutations for SKA1 structure and function were predicted. The prognostic and predictive potential of SKA1 expression have been searched in different cancer types and subtypes.

Missense mutations are particularly important in cancer research, since missense mutations may alter the physicochemical properties of amino acids or protein-protein interactions [39]. Therefore, it is important to evaluate the effect of missense mutations on protein structure and function, and to define the pathogenicity of missense mutations. In this study, the effect of 32 missense mutations on SKA1 structure and function was assessed and of the 32 mutations, 15 missense mutations were identified as deleterious (Table 1). Localization of predicted mutations were searched on protein structure, since pathogenic missense mutations may accumulate in certain critical sites of proteins. Accordingly, most of the deleterious mutations were located in the microtubule-binding region (Figure 2) and only three mutations were located in the N terminal region, including coiled-coiled region of SKA1 (Figure 1). The N terminal region of human SKA1 (residues 1–130) functions in the formation of SKA complex [8]. One of the three mutations, R27I, which were predicted as deleterious in this study, lied in region of SKA1 that forms the short bundle, while L70F and D83Y located in region that forms the long bundle of SKA complex (Figure 1B). The short and long bundles consist of three parallel helices of SKA1, SKA2 and SKA3. R27 is particularly important for short bundle, since it is one of the interaction sites, which stabilize the coiled-coil structure of the SKA complex [5]. Therefore, mutation on this site might affect the interaction of SKA1 with the complex.

The C terminal region of the SKA1 provides the main microtubule binding properties of SKA complex [5], [10]. The region including amino acids 133–255 is the microtubule-binding site (MTBD) of the SKA1 and the principal microtubule-binding component of the SKA complex; in addition, deletion of the C terminal region of SKA1 (residues 92–255) completely eliminates the microtubule binding activity of the complex [1], [5]. In this study, most of the predicted deleterious mutations were found to be located in MTBD of SKA1 (Figure 2). Therefore, 12 missense mutations in this region might be more relevant in cancer formation or progression. One critical site might be S185, one of the conserved Aurora B phosphorylation sites. Aurora B negatively regulates SKA complex localization at kinetochores [38]. According to results presented here S185Y has been predicted as deleterious (Table 1). Particularly, another mutation in R245 to R245A (arginine to alanine), one of the residues predicted to have a deleterious mutation in this study, has previously been reported to be critical for microtubule binding; this mutation abolished microtubule-binding activity and caused a mitotic delay in human cells [1], [10]. However, prediction results presented here proposed that the mutation of arginine to glutamine on this site (R245Q) was also deleterious (Table 1).

Previously, it has been suggested that over expression of SKA1 is more relevant with the oncogenic function of SKA1, since SKA1 genomic alterations observed at low frequency in cancers [40]. From this point of view, the expression level of SKA1 mRNA has been investigated in cancers. According to Oncomine analysis results presented here, a significant increase in SKA1 mRNA expression has been observed in datasets from several cancers (Table 2). Increased expression of SKA1 in has been shown in cancers, such as gastric cancer [12], hepatocellular carcinoma [11] and non-small cell lung carcinoma [15]. Accordingly, different types of kidney, colorectal, breast and lung cancers showed high expression of SKA1 (Table 2). However, in two datasets, Finak Breast [26] and Lee Brain [30], SKA1 expression has been found to be decreased. The high SKA1 expression has been observed in datasets of invasive ductal and lobular carcinomas (Table 2), while SKA1 expression was lower in breast tumor stroma (Table 2). Therefore, the expression difference of SKA1 in breast cancer datasets may be related with the breast part of the carcinoma.

High SKA1 expression may contribute to cell proliferation and invasion of cancer cells, since the knock down of SKA1 expression resulted in cell cycle arrest and a significant decrease in cell proliferation in different cancers [11], [12], [13], [14]. In addition to cell proliferation and invasion, SKA1 expression also affected metastasis in non-small cell lung carcinoma cells [15]. SKA 1 expression may also be required for cell survival, since SKA1 inhibition caused apoptosis in human oral adenosquamous carcinoma cell line [14] and hepatocellular carcinoma cells [11].

In this study, high expression of SKA1 has been shown to be significantly associated with overall survival, RFS, DMFS and PPS in breast cancer patients (Figure 3 and Table 3), with overall survival, FP and PPS in lung cancer patients and with overall survival and FP in gastric cancer patients (Figure 3 and Table 3). SKA1 expression did not show any significant correlation with survival of ovarian cancer patients (Figure 3 and Table 3). Consistently, altered SKA1 expression has not been detected in ovarian cancer datasets in Oncomine analysis. The findings suggest that high SKA1 expression might be a prognostic marker for breast, lung and gastric cancers.

In conclusion, mutations in the microtubule-binding domain of SKA1 might be particularly important for cancer research, since most of the predicted deleterious mutations scattered in this region. High SKA1 expression has been observed in several cancers and correlated with poor overall survival rate in breast, lung and gastric cancers. Therefore SKA1 might be predictive and prognostic marker for those cancers. However, it must be noted that further follow up experimental studies, such as sequencing for mutations and RT-qPCR for gene expression, are needed to test the potential of SKA1 as biomarker in cancers.