This article is mentioned in:

Introduction

Peroxiredoxins (Prdxs) are a family of small proteins that catalyze the reduction of peroxides using their conserved Cys residues as catalytical centers (1). Six Prdx isoforms have been found in mammalian cells, but they are non-redundant antioxidant proteins (2). The six isoforms of human Prdxs are located on chromosomes 1, 4, 8, 19 and X, with both PrdxIII and V located on chromosome 19 (3). The regulation of Prdxs has been investigated in various types of cancer. The expression of Prdxs, especially III, IV and V, has been found to be increased in breast malignancy, suggesting the induction of the expression of Prdxs as a response to the increased production of reactive oxygen species (ROS) in carcinoma tissues (4,5). Moreover, some members of Prdxs are thought to be cancer cell biomarkers (3). The knockdown of members of the Prdx family was previously shown to lead to clear distortion of cell signaling and tumor formation (6–9).

The aim of this study was to validate a previous finding according to which PrdxV constitutes a tumor marker of the breast in Sudanese patients (unpublished data), by investigating the expression pattern of a panel of Prdxs family members in Sudanese and Chinese patients.

Materials and methods

Patients

A panel of 106 Sudanese breast cancer patients (91 tumors and 79 normal breast tissues, of which 59 were tumor control pairs) were included in this case-control hospital-based study (Table I). The tissue samples were obtained from the El-Zahrawi Medical Center in Khartoum (Sudan) and the Institute of Endemic Diseases (IEND; University of Khartoum, Khartoum, Sudan) Tumor Bank. This study was approved by the Ethics Committee of the IEND and all patients provided informed consent. The samples were preserved in the form of formalin-fixed paraffin-embedded (FFPE) tissue blocks.

Table I Patient clinicopathological characteristics of the Sudanese patients.

Table I

Patient clinicopathological characteristics of the Sudanese patients.

Clinicopathological characteristics	No. (%)
Sample no.	106 (100)
Histological type
Ductal carcinoma in situ (DCIS)	2 (2)
Invasive ductal carcinoma (NOS)	60 (57)
Mixed ductal carcinoma (DCIS+NOS)	18 (27)
Lobular carcinoma	2 (2)
Others (e.g., papillary and medullary)	13 (12)
Mixed types	5 (5)
Missing data	5 (5)
Total	105
Mean age (years)	46.24 (24–79)
Gender
Female	101 (95.3)
Male	2 (1.9)
Missed data	2 (1.9)
Ethnicity
Afro-Asiatic	41 (39)
Nilo-Saharan	19 (18)
Niger-Kordofanian	1 (1)
Unknown	44 (42)
Nodal metastasis
Presence	37 (35.2)
Absence	19 (18.1)
Unknown	46 (46.7)
Tumor size (cm)
from 0 to <2	6 (5.7)
from 1 to ≥2	70 (66.04)
Unknown	30 (28.3)

[i] DCIS, ductal carcinoma in situ; NOS, invasive carcinoma.

Chinese breast cancer patients

A panel of 31 paired (tumors and controls) tissue samples from Chinese breast cancer patients (invasive ductal carcinoma) in the form of tissue arrays (Shanghai Outdo Biotech Co., Ltd., Shangai, China) were also included in the study.

Immunohistochemistry (IHC)

Both Sudanese and Chinese breast cancer tissue sections and controls were examined immunohistochemically, for the following Prdxs antibodies: PrdxI, V and VI (Table II). IHC was performed using the 2-step plus Poly-HRP anti-mouse/rabbit IgG detection system (biotin-free, anti-mouse/rabbit multivalent) kit (Golden Bridge International, Everett, WA, USA).

Table II Antibodies used in this study.

Table II

Antibodies used in this study.

Protein	Clone	Source	Dilution	Origin
PrdxV	(FL-214)	Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA)	1:500	Rabbit polyclonal
PrdxI	(ab59538)	Abcam (Cambridge, MA, USA)	1:1,000	Rabbit polyclonal
PrdxVI	(ab59543)	Abcam	1:1,000	Rabbit polyclonal
PARP	9542	Cell Signaling Technology, Inc. (Beverly, MA, USA)	1:1,000	Rabbit polyclonal
C-Myc	9E10	BD Biosciences (Franklin Lakes, NJ, USA)	1:1,000	Mouse polyclonal

[i] PARP, poly (ADP-ribose) polymerase.

In situ hybridization (ISH)

To design the PrdxV probe, PCR was performed using the modified primers (10,11) by the addition of the EcoRI and BamHI restriction sites (italics) (Invitrogen, Carlsbad, CA, USA): Prdx5, F, 5′-CGGAATTCATGGCCCCAATCAAGGTGGGAGAT-3′ and R, 5′-CGGGATCCCAGAGCTGTGAGATGATATTGG-3′.

The purified Prd xV gene was cloned using pcDNA™3.1/myc-His(-)B MCS plasmid (Invitrogen). Plasmids were prepared using standard methods, as described in a previous study (12). RNA probes were labeled with digoxigenin (DIG) using the DIG RNA labeling kit according to the manufacturer’s instructions (SP6/T7; Roche Diagnostics, Indianapolis, IN, USA). The optimal concentration of 100 pg/μl was chosen using DIG wash and block buffer according to the manufacturer’s instructions (Roche Diagnostics). ISH was performed as described by Breitschopf et al(13). The labeled antisense RNA probe was diluted (100 pg/μl) in hybridization buffer (12) and a labeled sense RNA probe was used as the control. The sections were incubated with alkaline-phosphatase-conjugated anti-DIG antibody, incubated with NBT/BCIP color reagent (Roche Diagnostics) overnight, and mounted with a water-soluble mounting medium.

Statistical analysis

The slides were independently examined by two experienced observers who were blinded to the initial results of the other observer. Immunoreactivity was allocated a score based on the percentage of positive tumor cells over total tumor cells ranging from 0 to 100% and on staining intensity (0, negative; 1, weak; 2, moderate; and 3, strong). Prdxs immunostaining scores were as follows: 0, negative or weak staining; and 1, moderate or strong staining. Clinicopathological parameters for the Sudanese patients were sub-classified as described in Table I. Descriptive statistics were calculated using the SPSS 11.5 statistical program. P<0.05 (two-sided test) was considered to indicate a statistically significant difference.

Results

Immunohistochemistry

The PrdxV level of expression in samples from Sudanese patients (a panel of 77 tumor and 68 control samples of which 51 were paired samples) was notably low compared to previously published studies (4,5). Only 9/77 (11.7%) breast cancer tissue samples were immunoreactive for the PrdxV antibody, whereas 88.3% of the samples were negative (Fig. 1). In the control samples, 29/68 (42.6%) were positive for the PrdxV antibody (Fig. 2), indicating a significant difference between tumors and non-malignant controls (P=0.000) (Figs. 3 and 4; Table III). In Chinese samples, PrdxV was found to be predominantly overexpressed in both tumor and control samples with 24/30 (80%) tumor samples, and 26/31 (83.9%) control samples being immunoreactive for the PrdxV antibody (Table III).

Figure 1 PrdxV expression in breast carcinoma. (a) Strong positive tumor (IHC); (b) same focus (H&E); (c) stained positive tumor (IHC); (d) same focus (H&E); (e) negative tumor (IHC); (f) same sample (H&E). IHC, immunohistochemistry; H&E, hematoxylin and eosin.

Figure 2 PrdxV expression in normal breast tissue (control). (a) Strong positive control (IHC); (b) same focus (H&E). IHC, immunohistochemistry; H&E, hematoxylin and eosin.

Figure 3 PrdxV protein scores are higher in controls compared to tumors (P=0.0008).

Figure 4 Difference in PrdxV expression between tumors and controls (P<0001).

Table III Percentage of Prdx family expression in Sudanese and Chinese breast cancer patients.

Table III

Percentage of Prdx family expression in Sudanese and Chinese breast cancer patients.

	Sudanese patients				Chinese patients
Antigen	Informative cases, n	Positive cases n (%)	Informative control, n	Positive cases n (%)	Informative cases, n	Positive cases n (%)	Informative control, n	Positive cases n (%)
PrdxI (protein)	88	65 (83)	62	52 (87.1)	30	21 (70)	27	21 (77.8)
PrdxV (protein)	77	9 (11.7)a	68	29 (42.6)	30	24 (80)	31	26 (83.9)
PrdxV (mRNA)	69	45 (56.2)	50	41 (87.2)	31	29 (93.5)	31	30 (96.8)
PrdxVI (protein)	43	20 (46.5)	37	21 (56.8)	30	17 (56.7)	25	17 (68)

a Bold, significantly different.

Unlike the Sudanese samples, the difference between PrdxV expression in tumors and controls in Chinese breast cancer patients was found to be insignificant (P=0.749).

Similarly no correlation was detected between PrdxV expression and the available pathological parameters, such as lymphatic invasion (P=1.000), tumor size (P=1.000) and grade (P=1.000). In addition, there was no correlation with other demographic parameters, such as age (P=0.412), gender (P=0.108), ethnicity (P=0.682) and patients’ geographical origin (P=0.686) (Table IV).

Table IV Correlations between PrdxV pathological characteristics and other studied Prdxs (Sudanese samples).

Table IV

Correlations between PrdxV pathological characteristics and other studied Prdxs (Sudanese samples).

Characteristics	Negative n (%)	Positive n (%)	P-value
Age (years)
≤46	43 (65.2)	3 (42.9)	0.412 (F)
>46	23 (34.8)	4 (57.1)
Total	66 (100)	7 (100)
Gender
Female	74 (100)	8 (88.9)	0.108 (F)
Male	0 (0)	1 (11.1)
Total	74 (100)	9 (100)
Lymphatic invasion
Negative	13 (33.3)	2 (40)	1.000 (F)
Positive	26 (66.7)	3 (60)
Total	39 (100)	5 (100)
Tumor size (cm)
from 0 to <2	5 (9.4)	0 (0)	1.000 (F)
from 1to ≥2	48 (90.6)	7 (100)
Total	53 (100)	7 (100)
Grade
Low	1 (2.4)	0 (0)	1.000 (F)
High	40 (97.6)	6 (100)
Total	41 (100)	6 (100)
Ethnicity
Nilo-Saharan	15 (35.7)	1 (20)	0.682 (F)
Afro-Asiatic	26 (61.9)	4 (80)
Niger-Kordofanian	1 (2.4)	0 (0)
Total	42 (100)	5 (100)
Patients’ geographical origin
North	25 (56.8)	4 (80)	0.686 (F)
East	1 (2.3)	0 (0)
West	11 (25)	0 (0)
Center	5 (11.4)	1 (20)
South	2 (4.5)	0 (0)
Breast cancer type
NOS	41 (56.2)	6 (75)	0.403 (F)
DCIS	0 (0)	0 (0)
LCIS	1 (1.4)	0 (0)
NOS+DCIS	14 (19.2)	2 (25)
Other types	12 (16.4)	0 (0)
Mixed types	5 (6.8)	0 (0)
Total	73 (100)	8 (100)
ISH
Negative	22 (95.7)	38 (86.4)	0.000 (Chi)
Positive	1 (4.3)	6 (13.6)
Total	23 (100)	44 (100)
PrdxI
Negative	16 (23.2)	0 (0)	0.000 (Chi)
Positive	53 (76.8)	5 (100)
Total	69 (100)	5 (100)
PrdxVI
Negative	20 (55.6)	1 (33.3)	0.000 (Chi)
Positive	16 (44.4)	2 (66.7)
Total	36 (100)	3 (100)

[i] F, Fisher’s exact test; Chi, Chi-square test; NOS, invasive carcinoma; DCIS, ductal carcinoma in situ; LCIS, lobular carcinoma in situ; ISH, in situ hybridization; PrdxI, peroxiredoxin I; PrdxVI, peroxiredoxin VI.

Expression of PrdxI and VI protein was also examined immunohistochemically in Sudanese and Chinese tissue samples and was found to be overexpressed in both tumors and controls in Sudanese and Chinese breast tissue samples (Tables III and VI, respectively). The difference between PrdxI and VI protein expression in tumors and controls in Sudanese and Chinese samples was statistically examined and found to be insignificant. No significant correlation was observed between the studied clinicopathological parameters and PrdxI and VI protein expression in both Sudanese and Chinese breast carcinomas (Tables V and VI, respectively).

Table VI Correlations of PrdxVI with pathological characteristics (Sudanese samples).

Table VI

Correlations of PrdxVI with pathological characteristics (Sudanese samples).

Characteristics	Negative n (%)	Positive n (%)	P-value
Age (years)
≤46	43 (65.2)	3 (42.9)	0.412 (F)
>46	23 (34.8)	4 (57.1)
Total	66 (100)	7 (100)
Gender
Female	74 (100)	8 (88.9)	0.108 (F)
Male	0 (0)	1 (11.1)
Total	74 (100)	9 (100)
Lymphatic invasion
Negative	13 (33.3)	2 (40)	1.000 (F)
Positive	26 (66.7)	3 (60)
Total	39 (100)	5 (100)
Tumor size (cm)
from 0 to <2	5 (9.4)	0 (0)	1.000 (F)
from 1 to ≥2	48 (90.6)	7 (100)
Total	53 (100)	7 (100)
Grade
Low	1 (2.4)	0 (0)	1.000 (F)
High	40 (97.6)	6 (100)
Total	41 (100)	6 (100)
Breast cancer type
NOS	13 (56.5)	10(52.6)	0.056 (F)
DCIS	0 (0)	0 (0)
LCIS	0 (0)	0 (0)
NOS+DCIS	5 (21.7)	4 (21.1)
Other types	3 (13)	4 (21.1)
Mixed types	2 (8.7)	1 (5.3)
Total	23(100)	19 (100)
Ethnicity
Nilo-Saharan	15 (35.7)	1 (20)	0.682 (F)
Afro-Asiatic	26 (61.9)	4 (80)
Niger-Kordofanian	1 (2.4)	0 (0)
Total	42 (100)	5 (100)
Patients’ geographical origin
North	25 (56.8)	4 (80)	0.960 (F)
East	1 (2.3)	0 (0)
West	11 (25)	0 (0)
Center	5 (11.4)	1 (20)
South	2 (4.5)	0 (0)
Total	44 (100)	5 (100)

[i] F, Fisher’s exact test; NOS, invasive carcinoma; DCIS, ductal carcinoma in situ; LCIS, lobular carcinoma in situ.

Table V Correlations of PrdxI with pathological characteristics (Sudanese samples).

Table V

Correlations of PrdxI with pathological characteristics (Sudanese samples).

Characteristics	Negative n (%)	Positive n (%)	P-value
Age (years)
≤46	43 (65.2)	3 (42.9)	0.412 (F)
>46	23 (34.8)	4 (57.1)
Total	66 (100)	7 (100)
Gender
Female	74 (100)	8 (88.9)	0.108 (F)
Male	0 (0)	1 (11.1)
Total	74 (100)	9 (100)
Lymphatic invasion
Negative	13 (33.3)	2 (40)	1.000 (F)
Positive	26 (66.7)	3 (60)
Total	39 (100)	5 (100)
Tumor size (cm)
from 0 to <2	5 (9.4)	0 (0)	1.000 (F)
from 1 to ≥2	48 (90.6)	7 (100)
Total	53 (100)	7 (100)
Grade
Low	1 (2.4)	0 (0)	1.000 (F)
High	40 (97.6)	6 (100)
Total	41 (100)	6 (100)
Breast cancer type
NOS	8 (50)	37(59.7)	0.056 (F)
DCIS	0 (0)	0 (0)
LCIS	0 (0)	1 (1.6)
NOS+DCIS	2 (12.5)	14 (22.6)
Other types	6 (37.5)	5 (8.1)
Mixed types	0 (6.8)	5 (8.1)
Total	16 (100)	62 (100)
Ethnicity
Nilo-Saharan	15 (35.7)	1 (20)	0.682 (F)
Afro-Asiatic	26 (61.9)	4 (80)
Niger-Kordofanian	1 (2.4)	0 (0)
Total	42 (100)	5 (100)
Patients’ geographical origin
North	25 (56.8)	4 (80)	0.686 (F)
East	1 (2.3)	0 (0)
West	11 (25)	0 (0)
Center	5 (11.4)	1 (20)
South	2 (4.5)	0 (0)
Total	44 (100)	5 (100)

[i] F, Fisher’s exact test; NOS, invasive carcinoma; DCIS, ductal carcinoma in situ; LCIS, lobular carcinoma in situ.

In situ hybridization

The mRNA expression level was examined only for PrdxV using ISH to verify whether the difference in expression between tumors and controls is at the protein level only or present at the mRNA level as well. In Sudanese breast tissue samples, a total of 69 tumor samples and 50 controls were examined, of which 45/69 (65.2%) tumor samples (Fig. 5) and 41/50 controls (87.2%) (Fig. 6) were positive. The difference in expression levels between tumors and controls was found to be significant (P= 0.044) (Table III).

Figure 5 PrdxV expression in breast carcinoma. (a) Strong positive tumor (ISH); (b) same focus (H&E); (c) medium positive tumor (ISH); (d) same focus (H&E); (e) negative tumor (ISH), (f) same focus (H&E). ISH, in situ hybridization; H&E, hematoxylin and eosin.

Figure 6 PrdxV expression in breast normal tissue (control). (a) Faint positive control (ISH); (b) same focus (H&E). ISH, in situ hybridization; H&E, hematoxylin and eosin.

In Chinese samples, the mRNA level of PrdxV correlations expression was also studied, using ISH. Tumor samples (29/31) (93.5%) were positive, as were 30/31 (96.8%) control samples (Table III). Unlike Sudanese samples, no significant difference was found between tumor and control samples (P=1.000).

Discussion

In the present study, a panel of Sudanese breast cancer tissue samples and healthy controls were investigated. Tissue sections were examined immunohistochemically to assess the PrdxV protein level of expression. The PrdxV mRNA expression was also evaluated using ISH. The bold score of PrdxV expression is representative of the finding of a preliminary proteomic study suggesting that PrdxV is differentially expressed in Sudanese breast cancer tissues as compared to healthy controls (unpublished data) (Table III). Various Prdx family members (PrdxI and VI) were also examined to ensure the specificity of PrdxV, mainly as a tumor marker among other family members. Additionally, to determine whether the PrdxV mode of expression is universal or limited to Sudanese breast cancer patients, the same experimental design was implemented in a panel of Chinese breast carcinoma samples and controls, and the two populations were compared.

Of the various Prdx family members, PrdxV was the only one that was significantly downregulated in tumor samples obtained from Sudanese breast cancer patients. At the protein level, only a few tumor samples were immunoreactive for PrdxV, only 9/77 (11.7%) were positive (P<0.0001), while 29/68 (47%) of the controls were immunoreactive (P=0.225). Based on these results, it appears that the PrdxV protein is not abundantly in present Sudanese neoplastic breast tissues, likely due to the fact that PrdxV may have a different role in these cells. This finding is contradictory to previous findings by Karihtala et al(5), where 79.8% of the studied cases were positive for PrdxV antibody, suggesting that a larger set of controls (n=68) was investigated in this study compared to that by Karihtala et al(5) where only three controls were examined immunohistochemically. The finding that PrdxV is downregulated in Sudanese breast cancer samples is also inconsistent with results obtained in a panel of Chinese breast carcinoma samples and controls, where PrdxV protein was found to be overexpressed in both tumors (80%) and controls (83.9%). This finding is also inconsistent with various Prdx family members examined (PrdxI and VI) in both Sudanese and Chinese tumor tissues and controls (Table III), suggesting the expression discrepancy to be restricted to the PrdxV protein of Sudanese patients only.

To investigate whether PrdxV gene expression loss, assessed by IHC, occurs at the protein level or at an earlier stage, such as at the transcription level, the mRNA mode of expression was investigated using ISH and was found to be overexpressed in both tumors and controls, with 56.2% of the studied tumors being immunoreactive to PrdxV antibody (P=0.011), and 87.2% of controls being positive (P<0.0001). PrdxV mRNA expression was also found to be significantly different between tumors and controls (P=0.044). In the Chinese cases, PrdxV mRNA was found to be overexpressed in both tumors (93.5%) and controls (96.8%), with no significant difference being identified in the expression between tumors and controls. ISH results assessing PrdxV mRNA expression levels in this study were in accordance with previous findings (14) assessing mRNA expression levels for all members of the Prdx and thioredoxin (Trx) families. PrdxV mRNA expression levels were found to be upregulated in breast cancer compared to normal breast tissues (14), and were in accordance with the immunohistochemical findings of PrdxV in Sudanese samples regarding the difference between tumors and controls. However, the difference was more apparent and statistically significant at the protein level, suggesting a post-translational modification. The regulation of Prdx activity occurs at the gene expression level and by post-translational protein modification and has received considerable attention (15–18). An example of gene expression loss and its effect on tumorigenesis is the loss of E-cadherin expression in lobular carcinomas of the breast (19–21).

The PrdxV protein expression loss in Sudanese breast cancer patients may have two consequences of antagonistic nature during tumorigenesis. The first one is the expected poor prognosis of tumors due to loss of PrdxV function which protects tissues from harmful ROS, which are usually considered to have carcinogenic potential and promote invasiveness (22,23). Prdx isoforms are non-redundant antioxidant proteins (2), since previous studies have shown that the knockdown of members of the Prdx family led to the distortion of cell signaling and tumor formation (6,7,9). Therefore, the down-regulation of PrdxV in Sudanese breast cancer patients is expected to yield cells more prone to oxidative stress and its accompanying DNA damage, leading to a poorer prognosis of Sudanese breast cancer patients. The second consequence of the PrdxV protein expression loss in Sudanese breast cancer patients is correlated with the absence of the PrdxV anti-apoptotic function, which is expected to induce breast cancer tissue sensitivity to chemo- and radiotherapies. PrdxV has been previously shown to have a protective role against oxidative stress and to lead to apoptosis (24–26).

Clinicopathological data were obtained for most of the studied cases, correlations were examined comparing the Prxds examined in the present study to each other, to available demographic data including age, gender and ethnicity, as well as to the available pathological parameters such as histological types, nodal metastasis and tumor size, which are also known prognostic markers of breast cancer. No statistically significant correlation was identified, whereas Karihtala et al(5) detected significant associations between PrdxV overexpression and larger tumor size, lymph node metastases, and poor differentiation of tumors. The finding in this study may not constitute a conclusion regarding the correlation between PrdxV expression and pathological parameters, since those parameters are not normally distributed, as most breast cancer patients present in Sudan, present with large tumors, high grades, and positive metastatic nodes in the majority of cases. This observation has been previously noted in Sudanese (27,28), as well as in African breast cancer patients (29).

The most important finding in this study is the marked downregulation of PrdxV in Sudanese breast cancer patients, suggesting that the molecule be regarded as a tumor marker. Numerous markers were previously identified with only a limited number of these markers being accepted for routine clinical use, such as HER2, ER and PR (30). Although this study suggests that PrdxV be regarded as a tumor marker of population specificity, it should be taken into consideration that breast cancer in individual patients may differ widely from one another in natural history and response to treatment (31). Population differences in breast cancer characteristics were previously described in terms of pathological parameters (27,31,32), and in terms of the identification of mutations, intronic variant sequences and unclassified variants, as well as variated copy numbers, reporting a role in breast cancer susceptibility that remains to be clarified. For instance, mutations in the predisposition genes BRCA1 and BRCA2, known to be associated with hereditary breast cancer, were studied in different populations such as Sudanese (33,34), Tunisian (35), Indian (36) and Slovak (37), where population-specific genetic disorders were observed. Similarly, the Mspl polymorphism in the 3′-non-coding region of the CYP1A1 gene was associated with breast cancer in African-American but not in Caucasian women (38).

In conclusion, the downregulation of PrdxV in Sudanese breast cancer patients is suggested as a tumor marker of population specificity. However, additional studies are needed to investigate the applicability of PrdxV as a tumor marker in Sudanese breast cancer patients complementing, but not replacing, the traditional diagnostic and prognostic markers. Additionally, further investigation is required to determine ways of incorporating this suggested marker into routine radio- and/or chemotherapy.

Abbreviations:

Acknowledgements

This study was funded by the Third World Organization for Women in Science (TWOWS) and the Swedish International Development Cooperation Agency (SIDA).

References