Polymorphisms in MMP-14 and MMP-2 genes and ovarian cancer survival

Abstract

BACKGROUND:

Functional polymorphisms in matrix metalloproteinases can increase or decrease the risk of cancer. This study focused on ovarian cancer and investigated how polymorphisms in the coding region of MMP-14 and the promoter region of MMP-2 are related to clinical characteristics including survival.

METHODS:

In 144 patients with ovarian tumours from a Caucasian population, polymorphisms of MMP-14 ($+$7096 and $+$6767) and MMP-2 ($-$735 and $-$1306) were analysed. These results were then correlated to the immunohistochemical expression of MMP-14 and MMP-2 and clinical characteristics.

RESULTS:

In these patients, the MMP-14 $+$7096 polymorphism showed only TT genotype, in sharp contrast to the described MAF (minimal allele frequency) C of 27%. The MMP-14 $+$6767 G$>$A polymorphism was found to have a hazard ratio of 2.09 (CI 1.00–4.35, $p$ 0.046) for recurrence-free survival in advanced-stage patients. However, this significance disappeared after Bonferroni correction for multiple testing. No other correlations between MMP-14 and MMP-2 polymorphisms, immunohistochemistry and clinical characteristics were found, except between the MMP-2 $-$1306 polymorphism and differentiation grade, with a Spearman correlation coefficient of $-$0.19, $p$ 0.064.

CONCLUSIONS:

In ovarian cancer, the MMP-14 $+$6767 G$>$A polymorphism in the coding region seems to improve recurrence-free survival with a hazard ratio of 2.09 (CI 1.00–4.35, $p$ 0.046). However, as this significance disappeared after correction for multiple testing, there is a need for further research on the functional effect of this change in the MMP-14 gene with larger patient sample sizes.

1. Background

In cancer, various matrix metalloproteinases (MMPs) play an important role in progression and metastasis [1]. MMPs are highly conserved proteases, which apart from their activity in pathophysiological processes such as cancer, inflammation and wound healing are also important in physiological processes such as reproduction [2]. In general, cancer prognosis is worse if MMPs are found to be overexpressed in primary and metastatic tumours [3, 4].

In this study, we focused on MMP-14, a membrane-bound collagenase and the main activator of MMP-2, and MMP-2 itself, a gelatinase. A number of studies in ovarian cancer indicate that MMP-14 and MMP-2 expression as determined with immunohistochemistry are independent prognosticators, whereas other studies report that their predictive value is limited [5, 6]. Molecular MMP-14 and MMP-2 mRNA data have revealed decreased survival for advanced-stage patients expressing high levels of mRNA [3, 7].

In 2001, the functional impact of polymorphisms in the promoter region of the MMP-2 gene was described. In particular, the C$>$T transition at $-$1306 (rs 243865) displayed strikingly lower promoter activity, with the T allele causing diminished transcription of the MMP-2 gene by disrupting an Sp1-type promoter site (CCACC box) [8]. Studies on the impact of this MMP2 $-$1306 C$>$T polymorphism showed that it is significantly associated with a reduced risk of cancer [9]. In contrast, the $-$735 C$>$T (rs 2285053) polymorphism increases cancer risk in Caucasians [10].

There are conflicting research results on polymorphisms in the MMP-14 gene, with some researchers reporting increased or reduced cancer risk, and others reporting poor clinical outcomes [11, 12, 13] Variants such as $+$7096 CC (rs2236307) increase the risk of cervical neoplasia but are not correlated with the clinicopathological characteristics of cervical cancer [11]. For MMP-14 polymorphisms and oral squamous cell carcinoma (OSCC), the $+$7096 TC/CC genotype reduces the risk and $+$6767 GA/AA (rs1042704) causes poor clinical outcomes [12]. Haplotypes $+$6767 G and $+$7096 T increase OSCC risk. MMP-14 $+$6767 G/A and $+$7096 C/C individuals show a reduced risk of hepatocellular cancer (HCC). The $+$7096 TC/CC genotype increases the risk of hepatitis C and cirrhosis in HCC patients. MMP-14 $+$6767 G and $+$7096 T frequency is higher in patients with hepatocellular cancer than in controls [13].

In contrast to the MMP-2 polymorphisms, these polymorphisms in the MMP-14 gene are located in the coding region, where MMP-14 $+$7096 results in a synonymous change (p.Gly285Gly) and MMP-14 $+$6767 results in a change from aspartate to asparagine (p.Asp273Asn). This amino acid change may result in an electrically driven increase or decrease in enzymatic activity.

Most of these studies were performed among Asian populations on types of cancer that are often caused by infection, such as cervical and hepatocellular cancer, and cancer types in which smoking is a major risk factor, such as head-and-neck cancer and lung cancer.

Little information is available on MMP-14 and MMP-2 polymorphisms in other cancer types and among Caucasian populations. In several studies on breast cancer, MMP-2 polymorphisms were found to influence risk and clinical characteristics [14, 15]. In a Dutch study on COPD, the MMP-2 $-$1306 TT polymorphism was related to an accelerated decline in lung function [16].

To our knowledge, there has been no previous study on these polymorphisms in ovarian cancer, which is not caused by infection, smoking or alcohol abuse.

In the present study, based on a regional Dutch cohort with five-year follow-up, polymorphisms in the coding region of MMP-14 $+$7096 and $+$6767 and the promoter region of MMP-2 $-$735 and $-$1306 were analysed in tumour tissue and then correlated to immunohistochemical expression of MMP-14 and MMP-2, clinical characteristics and survival.

2. Patients and methods

2.1. Clinical data

From the Regional Oncological Gynaecological Registry (ROGY) in the southern Netherlands, all ovarian tumours diagnosed between 1 January 2007 and 31 December 2008 were selected. After excluding metastases from other primary tumours, 148 patients remained, all diagnosed with a primary ovarian tumour at Elisabeth-TweeSteden Hospital in Tilburg, Amphia Hospital in Breda, Catharina Hospital in Eindhoven or Jeroen Bosch Hospital in Den Bosch, all in the Netherlands. These patients were followed until 1 June 2013. In the case of four patients, no tumour tissue was found in the archived paraffin-embedded material, leaving 144 patients for analysis.

All 144 patients underwent a laparotomy with staging or debulking when indicated by clinical stage and frozen-section results. Debulking surgery was found to be complete if no residual tumour tissue was found at the completion of surgery. In those patients with FIGO stage Ia or Ib ovarian cancer with differentiation grade I, no adjuvant therapy was given. All the other patients received six courses of adjuvant platinum-based chemotherapy.

Data collection from the ROGY registry was prospe-ctive. Additional clinical and follow-up data were extracted from the patients’ medical records. The FIGO stage (I to IV) was categorized according to World Health Organization (WHO) criteria. Serum levels of CA-125 were determined preoperatively.

2.2. Histopathological data

All histopathological results and previous slides were reviewed. Histology and differentiation grade were categorized according to World Health Organization (WHO) criteria, the grade being assigned according to the observer’s impression of both architectural and cytological features. The clear-cell histological subtype was excluded from this assignment, as these carcinomas are poorly differentiated by definition (Grade III) [17].

2.3. DNA extraction and analysis

From paraffin-embedded tissue, collected at the time of surgery or biopsy, 10 $\mu$m slides were prepared. DNA was isolated with TET-lysis buffer (10 mmol/L Tris-HCl, pH 8.5; 1 mmol/L EDTA, pH 8; 0.1% Tween-20) containing 5% Chelex-100 (Bio-Rad, Hercules, CA, USA). Protein was digested by adding 20 $\mu$L of proteinase K to each sample and incubation at 56${}^{\circ }$C for 48 hours. Fresh proteinase K (10 $\mu$L) was added after 24 hours. Next, heat inactivation at 95${}^{\circ }$C for ten minutes was applied. The samples were centrifuged for ten minutes at 16,000 g at room temperature, and the DNA concentration of the supernatant was measured using a Nanodrop ND-1000 (NanoDrop Technologies, Wilmington, DE) before storage at 4${}^{\circ }$C. We used a tetra-primer amplification refractory mutation system (ARMS) PCR to analyse the polymorphisms [18]. The PCR primers and conditions are listed in Supplementary Table 1. PCR reactions took place in a total volume of 25 $\mu$L, containing 30 ng of template DNA, 10 pmol of each forward primer and 1 pmol of each reverse primer. The mixes were incubated for two minutes at 95${}^{\circ }$C, followed by 35 cycles of one-minute denaturation, one-minute annealing (63${}^{\circ }$C) and one-minute extension (72${}^{\circ }$C).

Table 1.

Patient characteristics

	Borderline	Early-stage	Advanced-stage
	($n=$ 51)	($n=$ 23)	($n=$ 74)
Age (range)	53 (17–81)	59 (24–81)	64 (30–84)
Histological subtype
Serous low grade	20	10	12
Serous high grade			44
Mucinous low grade	30	3	4
Mucinous moderate grade		1	1
Mucinous high grade
Seromucinous low grade		2
Endometrioid low grade		1	2
Endometrioid high grade			1
Clear cell		6	2
Adenocarcinoma unspecified	1*	0	7
Mixed		0	2
Tumortype	Not applicable
I		20	21
II		3	53
CA 125	29 (4–780)	61 (7–3222)	615 (9–37500)
FIGO stage
IA-IC	47	23
IIB-C	4		14
III			44
IV			16
Cytoreduction (only advanced stage disease)
Complete	4		62
Incomplete			12
Number of deceased patients (ovarian tumor related)	4 (2)	4 (3)	53 (50)
Overall Survival (months)	62 (10–78)	64 (28–75)	40 (0–76)

*Brenner tumor.

2.4. Immunohistochemistry

Immunohistochemistry was performed as previously described from the same paraffin-embedded blocks [2]. Sections (3 $\mu$m) were deparaffinised in xylene and rehydrated with graded alcohol. Each slide included positive controls for MMPs, consisting of placental tissue. Each run included negative controls without a primary antibody. Endogenous peroxidase was blocked with 3% H${}_2$O${}_2$ and 5% normal goat serum. After each incubation step, the slides were washed twice with PBS. As the primary antibody for MMP-2, a monoclonal antibody was used (clone A-Gel vc2, Thermo Scientific) [19]. The slides were incubated overnight in a 1:10 dilution at 4${}^{\circ }$C. As the primary antibody for MMP-14, a polyclonal antibody (Thermo Scientific) [20] was used in a 1:20 dilution for 60 minutes at room temperature. After washing with PBS, the slides were incubated with secondary antibody, poly-HRP-GAM/R/R IgG Powervision (Immunologic, Duiven, the Netherlands) for 60 minutes at room temperature. Staining was done with concentrated diaminobenzidine solution in substrate buffer (20 $\mu$L/mL) (Immunologic, Duiven, the Netherlands) for five minutes. Slides were counterstained with hematoxylin. The slides were scored according to Kamat [5], where 0 $=$ no expression or  5% of cells positive, 1 $=$ weak expression (1–2 points), 2 $=$ moderate expression (3–4 points) and 3 $=$ strong expression (5–6 points). The overall score (OS) was dichotomized into 0 for no/weak expression (1–2 points) and 1 for moderate-to-strong expression (3–6 points). Part of the immunohistochemical data of the ovarian cancer patients were presented before [21].

2.5. Statistical analysis

Descriptive statistics were used to describe patient characteristics. For polymorphisms and immunohistochemistry, Spearman correlation coefficients were determined. For the advanced-stage patients, recurrence-free and overall survival were analysed with Cox-regression and Kaplan-Meier curves were plotted. All statistical analyses were performed using the Statistical Package for Social Sciences 23.0 (SPSS Inc., Chicago, IL, USA).

3. Results

Characteristics of the patients included in this study are summarised in Table 1. This cohort is a representative sample of the patients with an ovarian tumour from our region. The follow-up was at least five years. Figure 1 shows an example of immunohistochemical staining with conventional hematoxylin-eosin staining, and MMP-14 and MMP-2 immunostaining.

Figure 1.

Example of immunohistochemical staining of a serous borderline ovarian tumour with A $=$ hematoxylin-eosin staining, B $=$ MMP-14 immunostaining and C $=$ MMP-2 immunostaining.

3.1. MMP-14 and MMP-2 polymorphisms

Polymorphisms were identified by using two inner and two outer primers for this tetra-primer ARMS PCR. In a homozygous sample, one of the primers will result in a shorter specific product and the other product will not bind. The outer primer will result in a longer product. In a heterozygous sample, both inner primers will result in a product, so eventually three bands will be visible on the gel for this tumour sample. The inner primers for the two alleles are designed to be placed at different distances from the polymorphism in order to result in products of different length. Thus, the position of the bands on the gel indicate that patient’s specific polymorphism. Table 2 shows the frequencies of the SNPs for the different polymorphisms. The result of the MMP-14 $+$7096 polymorphism was surprising, because all tested samples showed TT diplotype. Therefore, this polymorphism was not statistically analysed. Table 3 shows the different polymorphisms with their described MAF (minimal allele frequency) and the results of this study.

Table 2.

Results of polymorphisms of MMP-14 and MMP-2

	Borderline ($n=$ 47)	Early-stage ($n=$ 23)	Advanced-stage ($n=$ 74)	$p$ value${}^{\ast }$
MMP-14 $+$7096				n.c.
CCs	0	0	0
CT	0	0	0
TT	31	18	40
MMP-14 $+$6767				n.s.
GG	21	12	33
GA	11	3	17
AA	0	2	0
MMP-2 –735				n.s.
CC	34	15	47
CT	7	5	15
TT	0	1	0
MMP-2 –1306				n.s.
CC	19	13	37
CT	9	3	11
TT	2	0	2

${}^{\ast }p$ value based on chi-square between groups, n.s. $=$ not significant, n.c. $=$ not computed.

Table 3.

Polymorphism description, minimal allele frequency (MAF) per population and MAF this study

Polymorphism	MAF per population	MAF this study
rs2236307	1000G American 0.17	CC $=$ 0
MMP-14 $+$7096	1000G European 0.20	CT $=$ 0
Gly285Gly	1000G East Asian 0.36	TT $=$ 99
MAF overall 0.27 (C)	1000G South Asian 0.25	0/198 $=$ 0
rs1042704	1000G American 0.13	GG $=$ 66
MMP-14 $+$6767	1000G European 0.20	GA $=$ 31
Asp273Asn	1000G East Asian 0.03	AA $=$ 2
MAF overall 0.11 (A)	1000G South Asian 0.23	35/198 $=$ 0.18
rs2285053	1000G American 0.18	CC $=$ 96
MMP-2 –735	1000G European 0.11	CT $=$ 27
promotor region	1000G East Asian 0.24	TT $=$ 1
MAF overall 0.13 (T)	1000G South Asian 0.12	29/219 $=$ 0.13
rs243865	1000G American 0.24	CC $=$ 69
MMP-2 –1306	1000G European 0.26	CT $=$ 23
Promotor region	1000G East Asian 0.10	TT $=$ 4
MAF overall 0.14 (T)	1000G South Asian 0.12	31/192 $=$ 0.16

MAF $=$ Minimal Allele Frequency.

4. Correlation between polymorphisms, IHC and clinical data

Spearman correlation coefficients did not show correlations between MMP-14 and MMP-2 polymorphisms and clinicopathological characteristics, such as subtype, stage, result of debulking surgery, platinum sensitivity and hepatic and lymphogenic metastasis. The only parameter that showed a trend for the MMP-2 $-$1306 polymorphism was differentiation grade with a Spearman correlation coefficient of $-$0.19, $p$ 0.064.

4.1. Survival analysis

Table 4 shows the results of the Cox regression analysis of the recurrence-free and overall survival data in advanced-stage patients for the different alleles. In general, polymorphisms can be analysed in a dominant model (CT+TT/CC or GA+AA/GG) or in a co-dominant model (CT/CC or GA/GG). Because the frequencies in our sample of TT for the MMP-14 and MMP-2 polymorphisms and AA for the MMP-14 $+$6767 are very low (0–2, see Table 2), the results of the two models are similar. For recurrence-free survival, MMP-14 $+$6767 G $>$ A showed significance with a hazard ratio of 2.09 (CI 1.00–4.35, $p$ 0.049) for the GG allele. However, this significance disappeared after Bonferroni correction for multiple testing ($p$ 0.05/6 tests $=$$p$ 0.008). Figure 2 shows the Kaplan-Meier curves for recurrence-free and overall survival for MMP-14 $+$6767.

Table 4.

Median recurrence-free and overall survival in months (minimum and maximum) per polymorphism and Cox regression analysis for recurrence-free and overall survival

	Recurrence free survival					Overall survival
	Months	HR	95.0% CI for HR		Significance	Months	HR	95.0% CI for HR		Significance
	(min-max)		Lower	Upper		(min-max)		Lower	Upper
MMP-14 $+$7096		Not					Not
CC		calculated					calculated
CT
TT	19 (2–65)					36 (9–76)
MMP-14 $+$6767		2.09	1	4.35	0.049		1.67	0.85	3.26	0.135
GG	21 (2–68)					43 (2–76)
GA	15 (0–42)					28 (0–75)
AA
MMP-2 –735		0.86	0.39	1.89	0.702		1.16	0.57	2.36	0.678
CC	19 (0–68)					38 (0–76)
CT	19 (0–50)					25 (2–75)
TT
MMP-2 –1306		0.81	0.37	1.76	0.596		0.76	0.34	1.66	0.484
CC	18 (0–65)					38 (0–75)
CT	21 (9–68)					33 (15–76)
TT	51 (55–61)					59 (57–61)

HR $=$ Hazard ratio, CI $=$ Confidence Interval, MMP $=$ Matrix MetalloProteinase.

Figure 2.

Kaplan-Meier curves for recurrence-free and overall survival for MMP-14 $+$6767 haplotypes.

5. Discussion

Most previous studies on MMP-14 and MMP-2 polymorphisms were conducted among Asian populations and showed that such polymorphisms influence the susceptibility to, and clinical characteristics of, types of cancer including cervical cancer, oral squamous cell carcinoma, head and neck cancer, lung cancer, prostate cancer and hepatocellular carcinoma [9, 11, 12, 13, 22, 23]. These polymorphisms are known to reduce or increase MMP expression, depending on the haplotype [8].

In this study, which was the first to examine the influence of MMP-14 and MMP-2 polymorphisms in a cohort of Caucasian women with ovarian tumours, it was found that MMP-14 $+$6767 GG influences recurrence-free survival with a hazard ratio of 2.09 (CI 1.00–4.35) with respect to GA. However, this significance disappeared after correction for multiple testing.

For MMP-2 polymorphisms in the promoter region $-$735 and $-$1306, where an effect on the amount of produced protein can be expected, we found no correlation with immunohistochemical expression of the protein nor with clinical characteristics nor with survival. However, we did observe a correlation between the MMP-2 $-$1306 polymorphism and differentiation grade, with a Spearman correlation coefficient of $-$0.19 ($p$ 0.064), which can be considered as a moderate effect. Hereby, MMP-2 $-$1306 CC is associated with a better differentiation grade than CT or TT polymorphism. For this polymorphism, a reduced risk of other cancer types, including breast, prostate, lung and gastro-intestinal cancers, is described [9, 14, 23, 24, 25]. The lack of significance in our study may be attributable to insufficient statistical power due to the small patient sample size. Furthermore, ovarian cancer is different in nature to most other cancers that have been investigated for these polymorphisms, as it is not related to habits such as smoking and alcohol abuse. In contrast to our findings, a reduced risk has been found for breast and prostate cancer [9, 23], which are in the same spectrum as ovarian cancer. Also, with the exception of the studies in breast and prostate cancer, most studies were done among Asian rather than Caucasian populations.

Because all patients showed TT genotype, MMP-14 $+$7096 polymorphism could not be tested for significance. The minor allele frequency of C is described to be 27% overall and 20% in Europeans in the 1000 Genome database. All other studies on MMP-14 polymorphisms in disease were done among Asian populations, so the importance of this finding in a Caucasian population is unclear [12, 13]. Unfortunately, the sample size of our study is rather small and therefore may lack the power tot test this findings against other genotypes for this polymorphism.

The MMP-14 $+$6767 polymorphism is located in the coding region of the gene, resulting in a change from aspartate to asparagine at position 273. Due to this change, a negative charge in the catalytic domain disappears, which may influence the enzyme activity of MMP-14 for its substrates; for example, through a loss of saltbridge formation or a loss of metalion binding. The definitive mode of actin for this is yet unknown, but changes like these may contribute to differences in enzyme-activity.

As the patients’ tumour tissue was examined in this study, we were unable to discern between germline and somatic mutations. With polymorphisms in promoter regions and a correlation with susceptibility, germline mutations are more probable. With polymorphisms in coding regions, both somatic and germline mutations may be important. More research on this aspect is needed. There is also a need for studies in a larger patient sample size, because with the current sample size no definitive conclusions can be drawn. Currently, in ovarian cancer there is insufficient knowledge of the functional effect of these polymorphisms in the MMP-14 gene.

6. Conclusion

This study on polymorphisms in the coding region of MMP-14 and the promoter region of the MMP-2 gene in ovarian tumours is the first study in a Caucasian population. A hazard ratio of 2.09 (CI 1.00–4.35, $p$ 0.049) was found for recurrence-free survival for the MMP-14 $+$6767 GG with respect to GA haplotype. However, after correction for multiple testing this significance disappeared. For MMP-14 $+$7096, only TT haplotype was detected, which is surprising considering the described MAF of 27% for C. No correlations with immunohistochemical results or clinical characteristics were found, except between the MMP-2 $-$1306 polymorphism and differentiation grade, with a Spearman correlation coefficient of $-$0.19 ($p$ 0.064).

Supplementary data

Supplementary Table 1
PCR conditions
Polymorphisms	Primer sequence	Tm	Amplicon size
MMP-14 $+$7096	Forward inner primer (T allele):	80${}^{\circ }$C	177 bp (T allele)
	ATGCCCCTCGTGTTTTCTGCCCCAGGGTGT
	Reverse inner primer (C allele):	80${}^{\circ }$C	102 bp (C allele)
	GGCATCTTGGTGGGGAACCCTGACGCG
	Forward outer primer:	80${}^{\circ }$C	222 bp (from two outer primers)
	TGCCCATCTGTCTGTCCTTCCGTCCCCG
	Reverse outer primer:	80${}^{\circ }$C
	AACATCTCCCCTCGGAGCATGGCCACGG
MMP-14 $+$6767	Forward inner primer (G allele):	77${}^{\circ }$C	190 bp (G allele)
	GGATGGACACGGAGAATTTTGTGCTGCACG
	Reverse inner primer (A allele):	81${}^{\circ }$C	105 bp (A allele)
	GTTGCTGGATGCCCCGGCGGTCATCCTT
	Forward outer primer:	79${}^{\circ }$C	237 bp (from two outer primers)
	TCGAGCATTCCAGTGACCCCTCGGCCAT
	Reverse outer primer:	79${}^{\circ }$C
	GATAGAGGCAGGGGCAGGTTGGCAGGGG
MMP-2 –735	Forward inner primer (C allele):	77${}^{\circ }$C	146 bp (C allele)
	TATCTCATCCTGTGACCGAGAATGCGGCCC
	Reverse inner primer (T allele):	80${}^{\circ }$C	243 bp (T allele)
	ACCTGCTGGGCTGCACTCCCAGGCGA
	Forward outer primer:	78${}^{\circ }$C	333 bp (from two outer primers)
	TGCCATGGCACTGGTGGGTGCTTCCTTT
	Reverse outer primer:	78${}^{\circ }$C
	GGAAATAGAGCAGTCAGGGGCCCCGCGT
MMP-2 –1306	Forward inner primer (T allele):	77${}^{\circ }$C	161 bp (T allele)
	ATTCCCCATATTCCCCACCCAGCACGCT
	Reverse inner primer (C allele):	68${}^{\circ }$C	185 bp (C allele)
	TGAGCTGAGACCTGAAGAGCTAAAGAGTTG
	Forward outer primer:	72${}^{\circ }$C	288 bp (from two outer primers)
	GTCCTTACTGACCCCTCCAGCTCCATCC
	Reverse outer primer:	72${}^{\circ }$C
	ATAAGCTGTGATGGGGAACTCCAGCAGG

Conflict of interest

The authors declare that they have no competing interests.

List of abbreviations

ARMS	Amplification Refractory Mutation System
DNA	DeoxyriboNucleic Acid
FIGO	Federation Internationale de Gynaecologie et Obstetrie
HCC	Hepatic Cell Carcinoma
MMP	Matrix Metalloproteinase
mRNA	messenger RiboNucleic Acid
OS	Overall Score
OSCC	Oral Squamous Cell Carcinoma
PBS	Phosphate-Buffered Saline
PCR	Polymerase Chain Reaction
ROGY	Regional Oncologic Gynaecology Registry
SPSS	Statistical Packager for the Social Sciences
TET-lysis buffer	Tris-HCl, EDTA, Tween 20-lysis buffer
WHO	World Health Organisation