Tumor tissue katanin P60 expression correlates with lymph node metastasis and worse prognosis in patients with breast cancer: A cohort study

Abstract

OBJECTIVE:

To evaluate the correlation of katanin P60 expression with clinicopathological grade and overall survival (OS) in patients with breast cancer (BC).

METHODS:

Three hundred and four operative BC patients were retrospectively reviewed in this cohort study. Tumor tissue sample was acquired from each patient during surgery. Immunofluorescent staining was used to assess katanin P60 expression.

RESULTS:

There were 265 BC patients with katanin P60 negative expression and 75 patients with katanin P60 positive expression. Higher N stage ( $p<$ 0.001) and TNM stage ( $p<$ 0.001) were observed in katanin P60 positive expression group compared to katanin P60 negative expression group. Kaplan-Meier (K-M) curves showed association of katanin P60 positive expression status with shorter OS. Multivariate Cox analysis illustrated katanin P60 positive expression ( $p<$ 0.001) could independently predict worse OS. Subgroups analysis indicated that katanin P60 positive expression correlated with worse OS in patients of TNM stage II ( $p=$ 0.001) and stage III ( $p$ =0.001), while no correlation was found between katanin P60 expression and OS in BC patients with stage I ( $p=$ 0.538). According to molecular subtypes, katanin P60 positive expression was correlated with shorter OS in patients with HER2 ${}^{+}$ HR ${}^{+}$ ( $p<$ 0.001), HER2 ${}^{+}$ HR ${}^{-}$ ( $p<$ 0.001) and HER2 ${}^{-}$ HR ${}^{-}$ ( $p=$ 0.001) status compared to katanin P60 negative expression, while no correlation between katanin P60 expression and OS was observed in patients with HER2 ${}^{-}$ HR ${}^{+}$ ( $p=$ 0.073).

CONCLUSION:

Katanin P60 positive expression was correlated with higher lymph node metastasis and worse OS, and it could be regarded as a novel and convincing biomarker to independently predict the prognosis of BC patients.

1. Introduction

Breast cancer (BC) is considered as one of the most common carcinomas, which has been reported that estimated 1700000 BC new cases and 520000 deaths occur in 2012, occupying about 25% of all cancer cases and 15% of cancer deaths among women worldwide [1, 2]. In the United States, approximately 255180 new cases and 41070 cancer deaths occur in 2017 [3]. In China, approximately 268600 new cases and 69500 cancer deaths occur in 2015 [4]. Despite improvements in BC management that lead to more than 30% decrease in mortality rate from 1989 to 2014, BC is still the second leading cause of cancer-associated mortality among women worldwide and its long-term prognosis is still far more from satisfactory, mainly due to recurrence and complications after surgery [5, 6]. Hence, investigating additional convincing biomarkers for supervising tumor progression and prognosis in BC patients is necessary.

Katanin is a microtubule-severing AAA (ATPases associated with various cellular activities) protein, composed of two subunits, including a 60 kDa ATPase subunit (P60) and an 80 kDa subunit (P80). Among these, katanin P60 is an ATPase-containing protein and a microtubule-severing enzyme, involved in the process of mitosis mediation, meiosis to mitosis transition and neuronal development [7, 8]. Although the role of katanin P60 has been detected in the process of tumorigenesis and tumor progression in some carcinomas, such as ovarian choriocarcinoma, lung cancer and prostate cancer, few study have investigated the effect of katanin P60 in the prognosis of BC [9, 10, 11]. Thus, the aim of this study was to evaluate the correlation of katanin P60 expression with clinicopathological grade and overall survival (OS), if any, in patients with BC.

2. Materials and methods

2.1 Patients

Three hundred and four operative BC patients were retrospectively reviewed from Hospital database from January 2004 to December 2008 in this cohort study. The screening criteria were: (1) Diagnosed with primary BC which was confirmed by clinical evaluation, radiographic examination and pathological demonstrations. (2) TNM stage I-III, and received breast surgery. (3) Formalin-fixed and paraffin embedded tumor tissue were available for immunofluorescent staining assay. Patients who lost follow-up or with missing basic clinicopathological information were excluded. This study was approved by the Ethics Committee of Suzhou Kowloon Hospital, Shanghai Jiao Tong University. Written informed consents or oral agreements were acquired from patients or patients’ family members.

2.2 Study flow

Five hundred and thirty-five primary BC patients who received surgery between January 2004 to December 2008 were screened, in 72 cases, we were not able to obtain the formalin-fixed and paraffin embedded tumor tissue for immunofluorescent staining while in 123 cases could not be followed because of lost follow-up or lacking of basic clinicopathological data. Thus, the remaining 340 patients were included in the present study.

2.3 Data collection and follow-ups

Age, gender, histologic grade, T stage, N stage and status of ER, PR and HER2 were collected. TNM stage was assessed following the 6th edition of the American Joint Committee on Cancer (AJCC) cancer staging manual. ER, PR or HER2 positivity status was determined by routine immunohistochemistry (IHC) assay in our hospital. ER, PR or HER2 positivity status were defined as the sample with stained cells above 10%. HR positivity status was defined as any one of ER positivity status and PR positivity status, while HR negativity status was defined as both ER negativity status and PR negativity status. OS was calculated from the duration spanning/between time of surgery to date of death occurring due to any cause. The median follow-up duration was 115.00 months (1/4 to 3/4 quarter: 96.00–136.75 months) and the last follow up date was April 2017.

Table 1

Baseline characteristics of 340 analyzed breast cancer patients

Parameter	Patients (N $=$ 340)		Katanin p60 negative (N $=$ 265)		Katanin p60 positive (N $=$ 75)		$p$ value
Age							0.240
$⩽$ 50 years (n/%)	161	(47.4)	121	(75.2)	40	(24.8)
$>$ 50 years (n/%)	179	(52.6)	144	(80.4)	35	(19.6)
Histologic grade							0.450
I (n/%)	69	(20.3)	50	(72.5)	19	(27.5)
II (n/%)	254	(74.7)	202	(79.5)	52	(20.5)
III (n/%)	17	(5.0)	13	(76.5)	4	(23.5)
T stage							0.140
I (n/%)	96	(28.2)	75	(78.1)	21	(21.9)
II (n/%)	217	(63.8)	173	(79.7)	44	(20.3)
III (n/%)	27	(7.9)	17	(63.0)	10	(37.0)
N stage							$<$ 0.001
0 (n/%)	141	(41.5)	130	(92.2)	11	(7.8)
I (n/%)	106	(31.2)	86	(81.1)	20	(18.9)
II (n/%)	76	(22.4)	46	(60.5)	30	(39.5)
III (n/%)	17	(5.0)	3	(17.6)	14	(82.4)
TNM stage							$<$ 0.001
I (n/%)	38	(11.2)	32	(84.2)	6	(15.8)
II (n/%)	203	(59.7)	179	(88.2)	24	(11.8)
III (n/%)	99	(29.1)	54	(54.5)	45	(45.5)
Subtypes							0.476
HER2+HR+ (n/%)	66	(19.4)	55	(83.3)	11	(16.7)
HER2+HR- (n/%)	45	(13.2)	35	(77.8)	10	(22.2)
HER2-HR+ (n/%)	169	(49.7)	132	(78.1)	37	(21.9)
HER2-HR- (n/%)	60	(17.6)	43	(71.7)	17	(28.3)

Comparison was detected by Chi-square test. $p<$ 0.05 was considered significant. HER2, human epidermal growth factor receptor 2; HR, hormone receptor.

2.4 Sample collection and immunofluorescent staining

Tumor tissue sample was acquired from each patient during breast cancer surgery, subsequently fixed in formalin and stored at 4 ${}^{\circ }$ C. The section was embedded in paraffin. After cutting into slices, the slice was placed at 65 ${}^{\circ }$ C for 3 h, and then deparaffinating was performed by xylene at room temperature. After dehydrating, the piece was washed with phosphate buffered saline (PBS) of pH 7.3. Subsequently, the slide was incubated overnight at 4 ${}^{\circ }$ C with rabbit anti-KATNA1 monoclonal antibody (Abcam, USA) used as primary antibody followed by washing three times with PBS. Following incubation at 37 ${}^{\circ }$ C for 20 minutes with secondary antibody (Anti-rabbit IgG (H $+$ L), F (ab’) 2 Fragment (Alexa Fluor ${}^{\mathrm{\circledR }}$ 488 Conjugate) #4412), the section was counterstained with Hoechst 33342. An Olympus BX63 microscope and cellSens Dimension (Version1.8) software (Olympus, Japan) were used to obtain images of immunofluorescent staining. The intensity of katanin P60 staining was assessed using semiquantitative method of Pavelic et al. [12]. The scoring criteria used for staining strength were as follows: 0 (no staining); 100 (weak staining); 200 (modest staining); and 300 (strong staining). After choosing five high-power fields of the most uniformly stained area, 100 cells were counted within each field and the mean number of positive cells was calculated for the rate of positive cells. Finally, the score of staining strength was multiplied with the percentage of positive cells, and the final result was classified into two groups: negative (final value $⩽$ 100, positive (final value $>$ 100).

2.5 Statistics

Statistical analysis was performed by SPSS 21.0 software (IBM, USA). Data was mainly presented as mean value $\pm$ standard deviation or count (percentage). Comparison between two groups was determined by Chi-square test. Kaplan-Meier (K-M) curves and log-rank test was used to evaluate OS. Univariate Cox proportional hazard regression was performed to assess baseline clinicopathological factors predicting OS, all factors with $p$ value below 0.1 in univariate Cox model were further analysed by multivariate Cox proportional hazard regression model. $p$ value $<$ 0.05 was considered significant.

3. Results

3.1 Baseline characteristics

One hundred and sixty-one (47.4%) patients were age $⩽$ 50 years, and 179 (52.6%) patients were age $>$ 50 years (Table 1). The numbers of patients with histological grade I, grade II and grade III were 69 (20.3%), 254 (74.7%) and 17 (5.0%) respectively. There were 38 (11.2%), 203 (59.7%) and 99 (29.1%) patients in the TNM stage I, stage II and stage III. As to molecular subtypes, the numbers of patients with HER2 ${}^{+}$ HR ${}^{+}$ , HER2 ${}^{+}$ HR ${}^{-}$ , HER2 ${}^{-}$ HR ${}^{+}$ and HER2 ${}^{-}$ HR ${}^{-}$ were 66 (19.4%), 45 (13.2%), 169 (49.7%) and 60 (17.6%) respectively. Other baseline characteristics of patients were shown in Table 1.

Figure 1.

Katanin P60 expression. Immunofluorescent staining results revealed that katanin P60 was mainly expressed on cytoplasm. (A) katanin P60 negative expression; (B) katanin P60 positive expression.

3.2 Katanin P60 expression

Immunofluorescent staining assay revealed that katanin P60 was mainly expressed in cytoplasm (Fig. 1A and B). Subsequently, a total of 340 BC patients were divided into two subgroups: patients with katanin P60 negative expression (N $=$ 265) and patients with katanin P60 positive expression (N $=$ 75). Higher N stage ( $p<$ 0.001) and TNM stage ( $p<$ 0.001) were observed in katanin P60 positive expression group compared to katanin P60 negative expression group, while no difference was observed between two groups in other baseline features, including age, histologic grade, T stage and molecular subtypes (Table 1).

3.3 Correlation of katanin P60 expression with OS

K-M curves and log-rank test were performed to assess the correlation of katanin P60 expression with OS in BC patients. Results showed association of katanin P60 positive expression with shorter OS as compared to the katanin P60 negative expression (Fig. 2).

Figure 2.

Correlation between katanin P60 expression and OS in BC patients. BC patients with katanin P60 positive expression had shorter OS compared to those with katanin P60 negative expression. Kaplan-Meier curves and log-rank test were used to compare OS according to katanin P60 status in total BC patients. $p<$ 0.05 was considered significant.

Figure 3.

Correlation between katanin P60 expression and OS in BC patients with TMN stage I, II and III subgroups. Correlations of katanin P60 expression with OS in patients in stage I (A), stage II (B) and stage III (C) was evaluated. Katanin P60 positive expression was correlated with worse OS in stage II and stage III subgroups, while no difference in the correlation between katanin P60 expression and OS in stage I subgroup. Kaplan-Meier curves and log-rank test were used to analyse the correlation of katanin P60 expression with OS in each subgroup. $p<$ 0.05 was considered significant.

Figure 4.

Correlation between katanin P60 expression and OS in BC patients with different molecular subtypes. Correlations between katanin P60 expression and OS in HER2+HR+ patients (A), HER2+HR- patients (B), HER2-HR+ patients (C) and HER2-HR- patients (D) were assessed. Katanin P60 positive expression was correlated with shorter OS in HER2+HR+, HER2+HR- and HER2-HR- subgroups, while no difference in the correlation between katanin P60 expression and OS was observed in HER2-HR+ subgroup. Kaplan-Meier curves and log-rank test were performed to evaluate the correlation of katanin P60 expression with OS in each subgroup. $p<$ 0.05 was considered significant.

Table 2

Univariate Cox analysis of baseline factors predicting OS

	Univariate Cox proportional
	hazard regression
	$p$ value	HR	95% CI
			Lower	Higher
Katanin p60 positive	$<$ 0.001	3.121	2.109	4.618
Age above 50 years	0.149	1.330	0.903	1.959
Histologic grade	$<$ 0.001	2.257	1.482	3.436
T stage	0.053	1.411	0.995	2.000
N stage	0.001	1.396	1.144	1.702
TNM stage	$<$ 0.001	1.804	1.307	2.489
HER2 positive	0.631	0.902	0.593	1.372
HR positive	$<$ 0.001	0.418	0.284	0.616

Univariate Cox proportional hazard regression was performed to analyze the baseline factors influencing OS. $p<$ 0.05 was considered significant. HER2, human epidermal growth factor receptor 2; HR, hormone receptor.

Table 3

Multivariate Cox analysis of baseline factors predicting OS

	Multivariate Cox proportional
	hazard regression
	$p$ value	HR	95% CI
			Lower	Higher
Katanin p60 positive	$<$ 0.001	3.569	2.300	5.537
Histologic grade	$<$ 0.001	2.560	1.660	3.950
T stage	0.771	1.067	0.689	1.652
N stage	0.182	1.135	0.866	1.488
TNM stage	0.080	1.882	0.928	3.817
HR positive	$<$ 0.001	0.404	0.273	0.598

Multivariate Cox proportional hazard regression was performed to analyzed baseline factors with a $p$ value below 0.1 in univariate Cox analysis. $p<$ 0.05 was considered significant. HR, hormone receptor.

3.4 Analysis of baseline factors affecting OS in total BC patients

Univariate Cox analysis was used to analyse baseline factors affecting OS. As presented in Table 2, katanin P60 positive expression ( $p<$ 0.001), higher histologic grade ( $p<$ 0.001), N stage ( $p=$ 0.001) and TNM stage ( $p<$ 0.001) were correlated with shorter OS, while HR positive ( $p<$ 0.001) was associated with better OS in BC patients. All factors with a $p$ value below 0.1 in univariate Cox analysis were further evaluated by multivariate Cox proportional hazard regression, which illustrated that katanin P60 positive expression ( $p<$ 0.001) and higher histologic grade ( $p<$ 0.001) could independently predict worse OS, while HR positive was an independent factor for prolonged OS in BC patients (Table 3).

3.5 Correlation of katanin P60 expression with OS in BC patients with TMN stage I, II and III subgroups

According to the TNM stage, all patients were classified into three subgroups: group with patients in stage I, group with patients in stage II and group with patients in stage III. K-M curves indicated that katanin P60 positive expression was correlated with worse OS in stage II ( $p=$ 0.001, Fig. 3B) and stage III ( $p=$ 0.001, Fig. 3C) subgroups compared to katanin P60 negative expression, while no correlation between katanin P60 expression and OS in BC patients with stage I ( $p=$ 0.538, Fig. 3A) was found.

3.6 Correlation of katanin P60 expression with OS in BC patients with different molecular subtypes

Subsequently, we divided all patients into four subgroups according to molecular subtypes, including patients with HER2 ${}^{+}$ HR ${}^{+}$ , patients with HER2 ${}^{+}$ HR ${}^{-}$ , patients with HER2 ${}^{-}$ HR ${}^{+}$ and patients with HER2 ${}^{-}$ HR ${}^{-}$ . As revealed in Fig. 4, compared to katanin P60 negative expression, katanin P60 positive expression was correlated with shorter OS in patients with HER2 ${}^{+}$ HR ${}^{+}$ ( $p<$ 0.001, Fig. 4A), patients with HER2 ${}^{+}$ HR ${}^{-}$ ( $p<$ 0.001, Fig. 4B) and patients with HER2 ${}^{-}$ HR ${}^{-}$ ( $p=$ 0.001, Fig. 4D) subgroups, while no correlation between katanin P60 expression and OS was observed in patients with HER2 ${}^{-}$ HR ${}^{+}$ ( $p=$ 0.073, Fig. 4C).

4. Discussion

In our study, we found that 1) katanin P60 positive expression was associated with higher N stage and TNM stage. 2) K-M curves disclosed that katanin P60 positive expression was correlated with worse OS compared to katanin P60 negative expression, and multivariate Cox analysis identified that katanin P60 positive expression could independently predict shorter OS in BC patients.

Katanin P60, one of subunits of katanin heterodimer, is encoded by KATNA1, which could promote ATPase hydrolysis, as well as cut and depolymerize the microtubule [13]. Recent data reveals that katanin P60 is mainly presented in the basal cells in normal human prostate glands, while, in prostatic adenocarcinomas, its expression is primary found in the prostate cancer cells [11]. Katanin P60 cuts off microtubule, shortening them and increasing their overall number to affect cell proliferation and cell migration, thereby acculating tumor metastasis [14, 15, 16, 17]. It also stimulates a G2/M arrest and polyploidy cells accumulation, affecting cell mitosis and tumor progression of cervical cancer [18]. Pervious study in patients with prostate cancer illustrates that katanin P60 up-expression could increase cell motility to promote tumor metastasis [11, 19]. Another interesting study identifies that katanin P60 could be regarded as a key gene related to progression and metastasis of choriocarcinoma [9]. In line with this study, we found that katanin P60 positive expression was correlated with higher N stage and TNM stage. This can be attributed to the fact that katanin P60 regulates microtubule dynamics, affects cell mitosis and meiosis, induces cell migration and thus contributes to tumor metastasis [14, 15, 17]. In terms of prognostic value, although there are few reports exploring the role of katanin P60 in the cancer prognosis, the present study investigated its effect on the prognosis of BC patients. Our results revealed that katanin P60 positive expression was correlated with worse OS compared to its negative expression, and it served as an independent factor for shorter OS in BC patients. The possible reasons are that katanin P60 could induce microtubule severing and target several cells or genes to mediate cell mitosis, promoting cell proliferation and cell migration, thereby accelerating tumor metastasis and decreasing OS in BC patients [15, 16, 17, 19].

According to TNM stage, a total of these 340 BC patients were subsequently divided into three groups including patients in stage I, patients in stage II and patients in stage III. We found the correlation between katanin P60 positive expression and shorter OS in stage II and stage III subgroups, while there was no correlation between katanin P60 and OS in BC patients with stage I. The possible reasons might be that evaluated expression of katanin P60 is accompanied with diseases progression, thus, compared to BC patients in stage II and stage III, the effect of katanin P60 expression on the prognosis of BC patients in stage I might be not obvious [11].

In the current study, we further classified total BC patients into four groups based on molecular subtypes, containing HER2 ${}^{+}$ HR ${}^{+}$ , HER2 ${}^{+}$ HR ${}^{-}$ , HER2 ${}^{-}$ HR ${}^{-}$ , and HER2 ${}^{-}$ HR ${}^{+}$ subgroups [12, 20]. We observed that katanin P60 positive expression was associated with worse OS in HER2 ${}^{+}$ HR ${}^{+}$ , HER2 ${}^{+}$ HR ${}^{-}$ and HER2 ${}^{-}$ HR ${}^{-}$ subgroups, while no difference was shown between katanin P60 expression and OS in HER2 ${}^{-}$ HR ${}^{+}$ subgroup. The detailed mechanism is still not clear, while the possible causes are that these BC patients with HER2 ${}^{-}$ HR ${}^{+}$ may have less lymph node metastasis, which is related to katanin expression, thus, the effect of katanin P60 expression is not obvious on the progression in HER2 ${}^{-}$ HR ${}^{+}$ subgroup.

Some limitations still existed in our study. Firstly, BC patients in stage I to III were retrospectively reviewed, while the role of katanin P60 expression in the prognosis of patients in stage IV was not investigated. Thus, further study enrolled BC patients in stage IV is greatly needed. Secondly, possible mechanisms of katanin P60 related to lymph node metastasis in BC were not explored. More work study is necessary to further investigate the detailed mechanisms of katanin P60 in BC.

In conclusion, katanin P60 positive expression was correlated with higher lymph node metastasis and worse OS, and it could be regarded as a novel and convincing biomarker to independently predict the prognosis of BC patients.