- Open Access
Basal keratin expression in breast cancer by quantification of mRNA and by immunohistochemistry
© Kordek et al; licensee BioMed Central Ltd. 2010
- Received: 19 December 2009
- Accepted: 28 April 2010
- Published: 28 April 2010
Definitions of basal-like breast cancer phenotype vary, and microarray-based expression profiling analysis remains the gold standard for the identification of these tumors. Immunohistochemical identification of basal-like carcinomas is hindered with a fact, that on microarray level not all of them express basal-type cytokeratin 5/6, 14 and 17. We compared expression of cytokeratin 5, 14 and 17 in 115 patients with operable breast cancer estimated by real-time RT-PCR and immunohistochemistry.
Despite the method of dichotomization and statistical analysis, there were cases with discordant results comparing immunohistochemistry and RT-PCR analysis. For dichotomisation based on quartiles and ROC, 14% of cases were negative on immunohistochemical examination for CK5/6, but presented high CK5 mRNA levels. There were also 48-55% cases, which were CK5/6-immunopositive, but were negative by mRNA examination. Similar discordances were observed for CK14 and CK17.
Basal keratin mRNAs did not correlate with ER mRNA levels, while immunohistochemistry produced significant relationship with ER status.
Our observation suggest that both method may produce different results in a small proportion of cases. Discordance between immunohistochemistry and RT-PCR may confound attempts to establish a simple methods for identification of basal-like tumors.
- Epidermal Growth Factor Receptor
- Myoepithelial Cell
- Epidermal Growth Factor Receptor Expression
- High Molecular Weight Cytokeratins
- Basal Keratin
Heterogeneity of breast cancer at the molecular level was supported by data from cDNA microarrays [1, 2]. Tumors lacking ER form three groups: a basal-like subtype, HER2-positive subtype, and a normal breast-like subtype. Basal-like subtype is characterized by multigenetic signature, usually with high expression of high molecular weight cytokeratins normally expressed in basal myoepithelial cells: keratin 5 (CK5), 14 (CK14) and keratin 17 (CK17) [1, 2]. They usually express vimentin and p-cadherin, and more than 60% of them also express epidermal growth factor receptor (EGFR) [3, 4].
A great interest in basal like-cancers produced attempts to determine basal-like tumors by the use of a much more easier technique such as immunohistochemistry. Unfortunately, both methods -- oligonucleotide microarrays and immunohistochemistry - do not produce identical results. In the study by Nielsen and al., immunohistochemical panel for basal-like cancers was defined as lack of ER and HER2 expression and positivity for CK5/6 or EGFR . Unfortunately, this panel still presented only 76% sensitivity for basal-like tumors derived from a microarray study.
Another attempt to simplify the determination of basal-like tumors was regarding them as synonymous with "triple negative tumors", regarded as lack of ER, PGR and HER2 . But according to comparative studies, as much as 15-54% of basal-like tumors defined on mRNA level, still express at least one of these markers [4, 5, 7–9].
Quantitative real-time RT-PCR technology provides a precise assessment of even small changes in gene expression. In this aspect, real-time RT-PCR is a much more sensitive assay when compared with oligonucleotide microarray and could be considered as a referential method . This raises the question whether microarray-based classification of breast tumors could be reconstructed or even improved by the use of data from the quantification of expression of selected genes assessed by real-time RT-PCR. Recently, there have been published some data supporting this thesis .
In a previous study, we have compared ER expression estimated by RT-PCR and by a routine immunostaining, and have validated which method might be more reliable for the molecular subtyping in relation with basal-type keratins and HER2 genes expression . Both methods produced discordant results in a proportion of cases, and lack of prognostic relevance of ER-mRNA level has been demonstrated, whereas the assessment by immunostaining has been related to clinical outcome. Also expression of basal keratins and HER2 genes significantly differed between ER-positive and ER-negative tumors divided on the basis of immunostaining, but not by mRNA level. Whereas immunostaining results are specific for tumor cells, mRNA for the RT-PCR analysis could originate not only from cancer cells but also from normal breast epithelium, myoepithelial and stromal cells. Furthermore, due to post-transcriptional and post-translational mechanisms, the amount of detected mRNA not always directly reflects protein level.
On the contrary, qRT-PCR may produce similar results of ER expression as obtained by immunohistochemistry, but only with a very stringent and quality controlled systems . Besides, immunohistochemistry still remains the gold standard for estimation of ER status in breast cancer.
Although, as stated by Reis-Filho and Tutt, "from a scientific perspective, microarray-based expression profiling analysis remains the gold standard for the identification of basal breast cancers", stringent analysis of profiles discloses that in basal-like cases there is low expression of basal cytokeratins in a few cases [2–4]. Similarly, in some luminal-type tumors there are cases with high expression of CK5 or CK14 [2, 3]. As mRNA for basal-type cytokeratins may originate from myoepithelial cells forming normal breast tissue intermixed with cancer cell, or the number of cancer cells even presenting these cytokeratins may be to sparse -- in both situations false results may be obtained.
The aim of this retrospective study was to compare basal-cell-type cytokeratin expression estimated by real-time RT-PCR and by a routine immunostaining.
Tumor specimens and study patients
Number of patients
Number of patients
CK5/6 status (IHC)
CK14 status (IHC)*
CK17 status (IHC)
Immunohistochemistry and scoring
Paraffin embedded sections were routinely processed. Slides for immunostaining for ER (Dako), CK14 and CK17 (both Novocastra) were pretreated with citrate buffer in a microwave oven. CK5/6 antibody from Dako was applied following autoclaving with high pH buffer. Antibody dilutions were as follows: ER - 1:35, CK5/6 - 1:100, CK14 -- 1:20, CK17 - 1:40. All following procedures were done according to standard protocols with EnVision+ System HRP (Dako). ER nuclear staining scoring was done using the method described by McCarty et al. . Tumors were considered as being positive for ER if Histo-score was above 100. The results of basal keratin membranous staining were classified as follows: negative - no staining seen in invasive cancer cells, positive -- weak or strong staining seen in invasive cancer cells. HER2 expression was examined with the commercially available Herceptest kit from Dako and score +3 denoted HER2-positive tumors.
Real-time RT-PCR analysis
Real-time RT-PCR primers and reaction conditions
Gene primers (5'-3')
Annealing temperature ( ° C)
Detection temperature ( ° C)
PCR product size (base pairs)
Beta-2-microglobulin ( B2M )
H3 histone, family 3A ( H3F3A )
Ribosomal phosphoprotein ( RPLP0 )
Ribosomal protein S17 ( RPS17 )
Estrogen receptor 1 ( ER )
Keratin 5 (CK5)
Keratin 17 ( CK17 )
Keratin 14 ( CK14 )
All reactions were made in triplicate. Detection of PCR products was performed with SYBR™ green I using qPCR Core kit for SYBR™ green I (Eurogentec, Belgium). Expression levels of target genes were normalized using four housekeeping genes: B2 M, H3F3A, RPLP0, and RPS17. Relative gene expression was calculated with the use of the mathematical model described by Pfaffl.
Mann-Whitney U test was employed to evaluate significance of differences in mRNA level between groups. Dichotomized values of mRNA level were compared with immunohistochemistry using the matched pairs Liddell's exact test and Scott's π test. Data were analyzed with respect to sensitivity and specificity derived from the receiver operating characteristic (ROC) and immunohistochemistry was regarded as a referential test. Kendall's rank correlation (τ) was used to test the strength of an association between expression of genes. Pearson's χ2 test or Fisher's exact test were used to test for contingency between dichotomized values of basal keratin expression (negative and positive) and values of other histopathological parameters. All results were considered statistically significant when two-sided p was less than 0.05.
mRNA of respective basal keratin genes depending on their status assessed by immunohistochemistry
Status by IHC
Mean ± SD
206.67 ± 727.20
424.48 ± 731.51
209.45 ± 684.34
1480.20 ± 2958.21
0.69 ± 2.47
3.11 ± 5.49
The comparisons between dichotomized values of CK5-mRNA level and CK5/6 immunohistochemical status demonstrated, that despite the method of dichotomization and statistical analysis, there were cases with discordant results comparing immunohistochemistry and RT-PCR analysis. For two methods of dichotomisation (quartiles and based on ROC; the ROC curve analysis was performed assuming that immunostaining was a reference test), there were still 48-55% cases, which were CK5/6-immunopositive, but were negative by mRNA examination. Similarly, 14% of cases were negative on immunohistochemical examination, but presented high mRNA levels. Similar discordances were observed for CK14 and CK17.
Highly significant, moderate, positive correlations between mRNA levels of CK5 and CK14 (τ = 0.40, 95%CI 0.29-0.51, p < 0,001), between CK5 and CK17 (τ = 0.51, 95%CI 0.40-0.62, p < 0,001), and between CK14 and CK17 (τ = 0.36, 95%CI 0.25-0.47, p < 0,001) were observed.
Relations between basal keratins expression and ER status assessed by immunohistochemistry
Relationship between ER and basal keratin status assessed by immunohistochemistry
Basal keratin status
ER status (number of patients)
CK5/6 and CK14 and CK17 negative
CK5/6 or CK14 or CK17 positive
Basal-like breast cancers recently have raised a great interest not only regarding clinical differences, but also in relation with new therapeutic possibilities. The vast majority of BRCA1 mutation-related breast tumors represent basal-like subtype. Moreover, Turner et al. have recently reported the high prevalence of BRCA1 downregulation in sporadic basal-like breast cancer . There are some promising data that platinum-based chemotherapy may be more effective in patients with BRCA-1 germline mutations or in "triple-negative" breast cancer [16, 17]. These observations may emphasize the importance of an easy and simple determination of basal-like phenotype.
A microarray analysis is a very elegant and sophisticated method, but for individual genes it is equivalent to estimation of mRNA level by the use of RT-PCR. Both methods have one important weakness -- the assessment of gene expression is based on total mRNA presented in the examined tissue, not only in cancer cells - and this weakness may produce false results in a proportion of cases. In our study, in a comparison of immunohistochemistry and RT-PCR, regardless of the method of dichotomization and statistical analysis used, there were cases with discordant results. For each cytokeratin, there were cases which were regarded as being positive by one method, and negative by the other one. Fourteen percent of cases were negative for CK5/6 as assessed by an immunohistochemical examination, but presented high CK5 mRNA levels. Similar discordances were also observed for CK14 and CK17. This observation suggests that in some cases high levels of basal keratin mRNA may originate not from cancer cells but possibly also from preexisting normal myoepithelial cells. Furthermore, due to the post-transcriptional and post-translational mechanisms, the amount of detected mRNA not always directly reflects protein level. The same applies to oligonucleotide microarrays.
Similar differences were observed in an opposite direction - some cases which were positive by immunohistochemistry were regarded as being negative by real-time RT-PCR. For CK5/6, there is a theoretical possibility that cells may express only CK6 and not CK5, but the same observation was made for CK14 and CK17. Possibly, the amount of immunopositive cancer cells in the sample was too small to give positive results by RT-PCR when mRNA levels were dichotomized.
Moreover, for both types of discordances, it may be one universal explanation: because of the heteregeneity of the tumor, tissue examined by immunohistochemistry was not exactly the same tissue which was examined by real-time RT-PCR.
We have found that basal keratin mRNA does not inversely correlate with ER mRNA level. This is an interesting observation, as in the published studies with the use of microarray technology such correlation is clear [1–3]. But when our samples were divided regarding basal keratin status on the basis of immunohistochemistry results, we observed significant relationship with ER status, estimated both by RT-PCR and by immunohistochemistry. It shows that immunohistochemistry may be a better method than RT-PCR in rendering a biological difference of basal-like tumors.
Studies that were conducted to establish which immunohistochemical markers were helpful for the best definition of basal-like tumors gave different results [18–22]. Rakha et al. suggested that only expression of basal-type cytokeratins (CK5/6 and CK14) should be included in such definition . In their study, no other marker was related with worse prognosis. More recently, some authors have claimed that EGFR expression should be added to the panel, and even in the absence of basal-cytokeratins, ER- and HER2-negative tumors presenting EGFR should be regarded as basal-type ones [5, 20, 21]. Nielsen at al. determined that 13 of 21 basal-type cancers from microarray study were CK5/6-positive by immunohistochemistry, 12 of them were EGFR-positive, and 6 of them were c-KIT-positive . However, these authors regarded as a positive case even the weakest reaction. They also found that EGFR-positivity was correlated with basal-type gene expression and was related with worse survival; the same applied to CK5/6-positive tumors. This observation is encouraging but it is still questionable that EGFR-positive tumors should be named as "basal-type". Fulford et al. found a good correlation with clinical outcome when as the "basal-like" tumors were only regarded the cases with the presence of keratin 14 .
Summarizing, we have demonstrated a discordance between real-time RT-PCR and immunohistochemistry in assessing basal-type cytokeratin status. This observation gives another difficulty in establishing an easy and simple method of identification of tumors that have a basal-like signature in microarray analysis.
The study was supported by a research grant from the Medical University of Lodz, Poland (No. 502-11-592; 502-11-744; 503-1034-3), and grants from the Polish National Committee of Scientific Research (KBN, Warsaw, Poland; No. 2 P05E 099 28 and No. 2 P05A 015 29).
- Perou CM, Jeffrey SS, Rijn van de M, Rees CA, Eisen MB, Ross DT, Pergamenschikov A, Williams CF, Zhu SX, Lee JC, Lashkari D, Shalon D, Brown PO, Botstein D: Distinctive gene expression patterns in human mammary epithelial cells and breast cancers. Proc Natl Acad Sci USA. 1999, 96: 9212-9217. 10.1073/pnas.96.16.9212.View ArticleGoogle Scholar
- Perou CM, Sorlie T, Eisen MB, Rijn van de M, Jeffrey SS, Rees CA, Pollack JR, Ross DT, Johnsen H, Akslen LA, Fluge O, Pergamenschikov A, Williams C, Zhu SX, Lønning PE, Børresen-Dale AL, Brown PO, Botstein D: Molecular portraits of human breast tumours. Nature. 2000, 406: 747-752. 10.1038/35021093.View ArticleGoogle Scholar
- Sorlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnsen H, Hastie T, Eisen MB, Rijn van de M, Jeffrey SS, Thorsen T, Quist H, Matese JC, Brown PO, Botstein D, Eystein Lønning P, Børresen-Dale AL: Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci USA. 2001, 98: 10869-10874. 10.1073/pnas.191367098.View ArticleGoogle Scholar
- Reis-Filho JS, Tutt ANJ: Triple negative tumours: a critical review. Histopathology. 2008, 52: 108-118. 10.1111/j.1365-2559.2008.03046.x.View ArticleGoogle Scholar
- Nielsen TO, Hsu FD, Jensen K, Cheang M, Karaca G, Hu Z, Hernandez-Boussard T, Livasy C, Cowan D, Dressler L, Akslen LA, Ragaz J, Gown AM, Gilks CB, Rijn van de M, Perou CM: Immunohistochemical and clinical characterization of the basal-like subtype of invasive breast carcinoma. Clin Cancer Res. 2004, 10: 5367-5374. 10.1158/1078-0432.CCR-04-0220.View ArticleGoogle Scholar
- Dent R, Trudeau M, Pritchard KI, Hanna WM, Kahn HK, Sawka CA, Lickley LA, Rawlinson E, Sun P, Narod SA: Triple-negative breast cancer: clinical features and patterns of recurrence. Clin CancerRes. 2007, 13: 4429-34. 10.1158/1078-0432.CCR-06-3045.View ArticleGoogle Scholar
- Calza S, Hall P, Auer G, Bjöhle J, Klaar S, Kronenwett U, Liu ET, Miller L, Ploner A, Smeds J, Bergh J, Pawitan Y: Intrinsic molecular signature of breast cancer in a population-based cohort of 412 patients. Breast Cancer Res. 2006, 8: R34-10.1186/bcr1517.View ArticleGoogle Scholar
- Sotiriou C, Neo SY, McShane LM, Korn EL, Long PM, Jazaeri A, Martiat P, Fox SB, Harris AL, Liu ET: Breast cancer classification and prognosis based on gene expression profiles from a population-based study. Proc Natl Acad Sci USA. 2003, 100: 10393-10398. 10.1073/pnas.1732912100.View ArticleGoogle Scholar
- Jumppanen M, Gruvberger-Saal S, Kauraniemi P, Tanner M, Bendahl PO, Lundin M, Krogh M, Kataja P, Borg , Fernö M, Isola J: Basal-like phenotype is not associated with patients survival in estrogen-receptor-negative breast cancers. Breast Cancer Res. 2007, 9: R16-10.1186/bcr1649.View ArticleGoogle Scholar
- Rajeevan MS, Ranamukhaarachchi DG, Vernon SD, Unger ER: Use of real-time quantitative PCR to validate the results of cDNA array and differential display PCR technologies. Methods. 2001, 25: 443-451. 10.1006/meth.2001.1266.View ArticleGoogle Scholar
- Perreard L, Fan C, Quackenbush JF, Mullins M, Gauthier NP, Nelson E: Classification and risk stratification of invasive breast carcinomas using a real-time quantitative RT-PCR assay. Breast Cancer Res. 2006, 8: R23-10.1186/bcr1399.View ArticleGoogle Scholar
- Potemski P, Pluciennik E, Bednarek AK, Kusinska R, Kubiak R, Kordek R: Evaluation of oestrogen receptor expression in breast cancer by quantification of mRNA. Histopathology. 2007, 51: 829-36. 10.1111/j.1365-2559.2007.02886.x.View ArticleGoogle Scholar
- Badve SS, Baehner FL, Gray RP, Childs BH, Maddala T, Liu ML, Rowley SC, Shak S, Perez EA, Shulman LJ, Martino S, Davidson NE, Sledge GW, Goldstein LJ, Sparano JA: Estrogen- and progesterone-receptor status in ECOG 2197: comparison of immunohistochemistry by local and central laboratories and quantitative reverse transcription polymerase chain reaction by central laboratory. J Clin Oncol. 2008, 26: 2473-81. 10.1200/JCO.2007.13.6424.View ArticleGoogle Scholar
- McCarty KS, Miller LS, Cox EB, Konrath J, McCarty KS: Estrogen receptor analyses. Correlation of biochemical and immunohistochemical methods using monoclonal antireceptor antibodies. Arch Pathol Lab Med. 1985, 109: 716-21.Google Scholar
- Turner NC, Reis-Filho JS, Russell AM, Springall RJ, Ryder K, Steele D, Savage K, Gillett CE, Schmitt FC, Ashworth A, Tutt AN: BRCA1 dysfunction in sporadic basal-like breast cancer. Oncogene. 2007, 24: 2126-32. 10.1038/sj.onc.1210014.View ArticleGoogle Scholar
- Byrsky T, Huzarsky T, Dent R, Gronwald J, Zuziak D, Cybulski C, Kladny J, Gorski B, Lubinski J, Narod SA: Response to neoadjuwant therapy with cisplatin in BRCA1- positive breast cancer patients. Breast Cancer Res Treat. 2008, 115: 359-63. 10.1007/s10549-008-0128-9.View ArticleGoogle Scholar
- Sirohi B, Ardnedos M, Popat S, Ashley S, Nerurkar A, Walsh G, Johnston S, Smith IE: Platinum-based chemotherapy in triple negative breast cancer. Ann Oncol. 2008, 19: 1975-6.Google Scholar
- Bertucci F, Finetti P, Cervera N, Esterni B, Hermitte F, Viens P, Birnbaum D: How basal are triple-negative breast cancers?. Int J Cancer. 2008, 123: 236-40. 10.1002/ijc.23518.View ArticleGoogle Scholar
- Cheang MC, Voduc D, Bajdik C, Leung S, McKinney S, Chia SK, Perou CM, Nielsen TO: Basal-like breast cancer defined by five biomarkers has superior prognostic value than triple-negative phenotype. Clin Cancer Res. 2008, 4: 1368-76. 10.1158/1078-0432.CCR-07-1658.View ArticleGoogle Scholar
- Rakha EA, El-Sayed ME, Green AR, Lee AH, Robertson JF, Ellis IO: Prognostic markers in triple-negative breast cancer. Cancer. 2007, 109: 25-32. 10.1002/cncr.22381.View ArticleGoogle Scholar
- Tischkowitz M, Brunet JS, Begin LR, Huntsman DG, Cheang MC, Akslen LA, Nielsen TO, Foulkes WD: Use of immunohistochemical markers can refine prognosis in triple negative breast cancer. BMC Cancer. 2007, 7: 134-10.1186/1471-2407-7-134.View ArticleGoogle Scholar
- Fulford LG, Reis-Filho JS, Ryder K, Jones Ch, Gillet ChE, Hansby A, Easton D, Lakhani SR: Basal-like grade invasive ductal carcinoma of the breast: patterns of metastasis and long term survival. Breast Cancer Res. 2007, 9: R4-10.1186/bcr1636.View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.