Gross cystic disease fluid protein 15 (GCDFP-15) expression in breast cancer subtypes
- Silvia Darb-Esfahani1Email author,
- Gunter von Minckwitz2, 3,
- Carsten Denkert1,
- Beyhan Ataseven4,
- Bernhard Högel5,
- Keyur Mehta2,
- Gabriele Kaltenecker6,
- Thomas Rüdiger7,
- Berit Pfitzner1,
- Kornelia Kittel8,
- Bettina Fiedler9,
- Klaus Baumann10,
- Roland Moll11,
- Manfred Dietel1,
- Holger Eidtmann12,
- Christoph Thomssen13 and
- Sibylle Loibl4
© Darb-Esfahani et al.; licensee BioMed Central Ltd. 2014
Received: 31 October 2013
Accepted: 16 July 2014
Published: 28 July 2014
Gross cystic disease fluid protein 15 (GCDFP-15), which is regulated by the androgen receptor (AR), is a diagnostic marker for mammary differentiation in histopathology. We determined the expression of GCDFP-15 in breast cancer subtypes, its potential prognostic and predictive value, as well as its relationship to AR expression.
602 pre-therapeutic breast cancer core biopsies from the phase III randomized neoadjuvant GeparTrio trial (NCT00544765) were investigated for GCDFP-15 expression by immunohistochemistry. Expression data were correlated with disease-free (DFS) and overall survival (OS) time as well as pathological complete response (pCR) to neoadjuvant chemotherapy.
239 tumors (39.7%) were GCDFP-15 positive. GCDFP-15 expression was positively linked to hormone receptor (HR) and HER2 positive tumor type, while most triple negative carcinomas were negative (p < 0.0001). GCDFP-15 was also strongly correlated to AR expression (p 0.001), and to the so-called molecular apocrine subtype (HR-/AR+, p < 0.0001). Higher rates of GCDFP-15 positivity were seen in tumors of lower grade (<0.0001) and negative nodal status (p = 0.008). GCDFP-15 positive tumors tended to have a more favourable prognosis than GCDFP-15 negative tumors (DFS (p = 0.052) and OS (p = 0.044)), which was not independent from other factors in multivariate analysis. GCDFP-15 expression was not linked to pCR. Histological apocrine differentiation was frequent in molecular apocrine carcinomas (60.7%), and was associated with GCDFP-15 within this group (p = 0.039).
GCDFP-15 expression is higher in tumors with favorable prognostic features. GCDFP-15 expression is further a frequent feature of AR positive tumors and the molecular apocrine subtype. It might have reduced sensitivity as a diagnostic marker for mammary differentiation in triple negative tumors as compared to HR or HER2 positive tumor types.
KeywordsGCDFP-15 Breast cancer Neoadjuvant chemotherapy Apocrine CUP
Gross cystic disease fluid protein 15 (GCDFP-15, syn. prolactin-inducible protein, PIP) is a 15 kDa protein that was originally detected in the cystic fluid from cystic mastopathy . It is not expressed in normal ductal or lobular epithelium but in apocrine metaplasia of the breast . Apart from breast cancer, only very few tumors, such as prostate cancer and carcinomas of the skin appendages express GCDFP-15 . It is therefore highly specific for mammary differentiation in females, and is frequently used as an immunohistochemical marker for the evaluation of a potential mammary origin of metastatic carcinoma of unknown primary site. The expression of GCDFP-15 is regulated by the androgen receptor (AR) , however, little is known about its function. A recent study on gene expression profiles in androgen-stimulated, GDCFP-15 expressing versus GCDFP-15 non-expressing breast cancer cell lines, reported an up-regulation of pro-apoptotic and anti-proliferative genes along with GCDFP-15 . In carcinomas of the breast, GCDFP-15 is also used as a marker for apocrine differentiation [2, 6–9]. Apocrine breast carcinoma is a rare subtype of invasive ductal carcinoma, which is primarily defined by morphological features such as abundant eosinophilic and granular cytoplasm, and shows frequent expression of the androgen receptor (AR) . Some years ago a so-called molecular apocrine subset of breast carcinoma has been defined by gene expression analysis, and was characterized by active AR and weak or absent estrogen receptor (ER) signalling . In this study, all tumors that were assigned to the molecular apocrine group had strong morphological features of apocrine differentiation. The existence of the molecular apocrine subtype has since then been reproduced [12, 13]. However, its clinical impact is conflictive to date.
We used a large and well-characterized cohort of breast cancer patients who underwent anthracycline/taxane-based neoadjuvant chemotherapy (NACT) in the phase III randomized GeparTrio trial (NCT00544765) to investigate the distribution of GCDFP-15 expression in biological subtypes of breast cancer, its potential prognostic and predictive value, as well as its relationship to AR expression. GCDFP-15 expression and biological tumor types were determined by immunohistochemistry in pre-therapeutic breast cancer core biopsies.
Association of GCDFP-15 expression with baseline clinico-pathological parameters
HR (n = 576)
HER2 (n = 576)
Biological tumor types (n = 569)
AR (n = 545)
Age (n = 602)
< 50 years
> = 50 years
Histological type (n = 602)
Grading (n = 601)
cT (n = 591)
cN (n = 583)
Construction of a tissue micro array (TMA) of pre-therapeutic core biopsies has been explained previously . Immunohistochemistry was performed on a Ventana BenchMark XT instrument (Ventana Medical Systems Inc., Tucson, AZ, USA) after pre-treatment with protease using a mouse monoclonal antibody directed against human GCDFP-15 (clone D6, dilution 1:400, Covance, Princeton, NJ, USA). For visualization, the iView DAB detection kit (Ventana Medical Systems Inc.) was used. Stained slides were digitized and evaluated on the computer monitor with support of the TMA Evaluator software (VMScope GmbH, Berlin, Germany) by a board certified pathologist (S. D.-E.). Both staining intensity and the percentage of stained tumor cells were evaluated and combined to an IRS (see previous chapter).
Statistical analysis was performed using SPSS Statistics 19 (IBM Corporation, Somers, NY, USA). In logistic regression analyses, significance of the correlation with pCR was assessed by the Wald test. Survival analyses were performed with the Kaplan-Meier method and univariate log-rank test and with Cox regression analysis for multivariate tests. The association between GCDFP-15 expression and clinico-pathological factors, biological tumor types, and AR expression was analysed by Fisher’s exact test or Pearson’s chi square test, as indicated. All tests were two-sided, and p-values <0.05 were considered as significant.
GCDFP-15 expression pattern in human breast carcinomas
Association with biological tumor types and clinico-pathological factors
GCDFP-15 expression was significantly enriched in tumors with certain biological characteristics. There was a modest increase of GDCFP-15 expression in HR and in HER2 positive tumors (Table 1). GCDFP-15 positivity rate was 43.1% in HR positive tumors and 33.7% in HR negative tumors (p = 0.034), and 48.3% HER2 positive carcinomas expressed GCDFP-15 as opposed to 37.4% HER2 negative tumors (p = 0.035). Consequently, GCDFP-15 was also differentially distributed among biological tumor types, as defined by HR and HER2 status: frequency of positive tumors was significantly higher in luminal subtypes (HR+/HER2-: 42.4%, HR+/HER2: 45.6%) as well in HER2 positive tumors (HR-/HER2+: 55.8%) as opposed to in triple negative breast carcinomas (TNBC, HR-/HER2-: 26.2%, p = 0.001, Table 1). Consistent with GCDFP-15 being a downstream target gene of AR , there was a strong association between GCDFP-15 and AR expression. 51.5% of AR positive tumors were also positive for GCDFP-15, whereas only 24% of AR negative carcinomas showed GCDFP-15 staining (p < 0.0001). As AR is a frequent feature of apocrine tumor differentiation and GCDFP-15 has also been proposed as a marker for apocrine differentiation, we tested GCDFP-15 expression for an association with the so-called molecular apocrine subtype (HR-/AR+), as described by Farmer et al. . Tumor types according to Farmer were grouped as follows: HR + (AR+/-), n = 365), HR-/AR + (n = 56), and HR-/AR- (n = 101). HER2 positivity was more frequent in molecular apocrine carcinomas (42.9% as opposed to 17.1% in HR + (AR+/-) and 15.8% in HR-/AR-, p < 0.0001). There was a significant association between GCDFP-15 and molecular apocrine subtype, with 67.9% GCDFP-15 positive cases in this group (p < 0.0001). 41.6% of HR + (AR+/-) tumors were also positive for GCDFP-15, while the rate of GCDFP-15 positive cases in the subgroup that was completely negative for steroid hormone receptors (HR-/AR-) with 18.8% was lower than the one in triple negative tumors (HR-/HER2-, 26.2%, Table 1). GCDFP-15 was further associated with certain favorable tumor features, such as lower tumor grade (p < 0.0001), and negative nodal status (p = 0.008, Table 1). Using the currently by ASCO/CAP guidelines  recommended cutoff for ER/PR evaluation (<1% stained tumor cells = negative, > = 1% stained tumor cells = positive) we obtained similar results: HR positivity rate in the total study group was 79.5%, GCDFP-15 expression still was associated with HR positivity (p = 0.046) and with molecular apocrine tumor type (p < 0.0001), although the rate of molecular apocrine tumors was now decreased to 6.5%.
Morphological features of molecular apocrine carcinomas
Prognostic impact of GCDFP-15 expression
Univariate survival analysis
Mean survival, years (SE)
Mean survival, years (SE)
Biological tumor types
< 50 years
> = 50 years
Predictive value of GCDFP-15 expression
Univariate logistic regression: association with pCR
Biological tumor types
< 50 years
> = 50 years
We investigated the expression of GCDFP-15 in a large and well-characterized clinical trial cohort of breast carcinomas treated with NACT with a special emphasis on its distribution in breast cancer subtypes and its prognostic impact. We found that GCDFP-15 was increased in HR positive as well as in HER2 positive subtypes as compared to TNBC (HR-/HER2-). GCDFP-15 expression was not predictive for response to NACT. Although GCDFP-15 was a favorable prognostic factor for DFS and OS in univariate analysis, this impact was not independent from other factors and not evident within breast cancers subtypes. GCDFP-15 was furthermore strongly associated with AR and therefore enriched in the so-called molecular apocrine breast cancer subtype.
Although widely used as a diagnostic marker for breast carcinoma in pathology, the prognostic value of GCDFP-15 has not been systematically evaluated to date. Pagani (1994), in a small case series of 33 breast cancers found evidence of a longer relapse-free survival in patients with tumors positive for GCDFP15 gene expression . Fritzsche et al. (2007) reported GCDFP-15 as a positive prognostic factor in a cohort of 165 carcinomas . However, in this study the prognostic impact of GCDFP-15 expression was not investigated for each biological tumor type separately. Due to the very different molecular biology of those breast cancer subtypes, biomarkers may have quite varying prognostic implication within the subtypes. We show here that the prognostic impact of GCDFP-15 is most likely a bystander effect of its association with other factors, such as HR expression, nodal stage, and tumor grade. Our study thereby confirms previous findings of an association between GCDFP-15 expression and features of good-prognosis tumors [7, 22, 26], and it might be speculated that GCDFP-15 parallels the expression of its regulatory factor AR, which is also linked to favorable prognostic clinico-pathological features, as we showed previously . We further found that GCDFP-15 is differentially expressed in breast cancer subtypes and is enriched in luminal and HER2 positive carcinomas, while being relatively sparse in TNBC. Similarly, Huo et al. (2013) reported a rather low percentage of GCDFP-15 positives in primary (14%) and metastatic TNBC (21%) . Lewis et al. (2011) found even higher rates of GCDFP-15 expression than us in luminal (65-71%) and in HER2 positive carcinomas (64%), and found only one out 33 TNBC (basal-like and unclassified triple negative tumors) to be positive for GCDFP-15, however, their cohort being relatively small, might have underestimated the frequency of GCDFP-15 positivity in TNBC . Taken together, these data warrant care if GCDFP-15 is used as a diagnostic marker for mammary differentiation of metastases of a cancer of unknown primary (CUP) because a significant proportion of breast cancers, particularly TNBC might be negative. An extended panel of immunohistochemical markers for mammary differentiation should be used to increase sensitivity. We show an enrichment of GCDFP15 expression in HER2 positive tumors and a strong association with AR expression, and are therefore in line with previous reports [7, 22, 27]. Not surprisingly, we further found GCDFP-15 to be elevated in the so-called molecular apocrine carcinomas that are defined by AR expression in the absence of HR expression . In our study histological apocrine differentiation was found in 60.7% of molecular apocrine carcinomas; additionally, GCDFP-15 within the molecular apocrine subgroup was associated with histological signs of apocrine differentiation, which suggests that ER/PR, AR, and GCDFP-15 expression are helpful markers to confirm apocrine differentiation in morphologically conspicuous cases. On the other hand, 39.3% of HR-/AR + carcinomas did not show apocrine morphology in our cohort, which indicates that molecularly and morphologically defined apocrine groups overlap only partly. The clinical significance of the molecular apocrine subtype is not clear to date and remains to be determined as proposed by the current WHO Classification of Tumors of the Breast  (2012), similarly conflictive data exist regarding the prognostic impact of histologically defined apocrine subtype (reviewed by Vranic et al. ). Our study does not point to a particular prognosis or therapy response of HR-/AR + carcinomas, as pCR rate and survival times were quite similar to those in HR- tumors without AR expression. Interestingly, HER2 expression seems to interact with AR in HR negative tumors in prognostic terms, as in our previous study in the same cohort, AR positivity was a positive prognostic factor only in the molecular subgroup of triple negative breast cancer (as defined by ER/PR/HER2 negativity).
A limitation of our study is the reduced sample size in the HER2 positive tumor types (HR+/HER2+: n = 68, HR-/HER2+: n = 43), which might hamper the detection of a differential expression of GCDFP-15 or a prognostic impact of GCDFP-15 expression within those tumor types. Furthermore, relatively short follow-up times indicate that survival analysis should be interpreted cautiously (the GeparTrio study not being powered for survival as a primary end point). An additional limitation might be that we used a TMA constructed out of core biopsies, which in some cases contained only few tumor cells and which together with the focal expression pattern of GCDFP-15 might result in a reduced sensitivity for detection of GCDFP-15 expression in some tumors. However, the rate of GCDFP-15 expression in our study group (39.7%) was in the range reported in the literature [21, 26]. Only 602 out of 2.357 patients in the original GeparTrio studies could be included for this project; however this is still the largest study on GCDFP-15 expression to date.
GCDFP-15 is expressed in all major biological breast cancer subtypes, and may be particularly useful as a diagnostic marker for mammary differentiation in HR and HER2 positive tumors, while there is reduced sensitivity in the triple negative subset. Due to its strong link to AR expression it may also be a marker for the so-called molecular apocrine subtype. GCDFP-15 is linked to clinico-pathological factors that indicate a better patient outcome, but is by itself no independent prognostic factor and is not predictive of response to anthracycline/taxane-based NACT.
We thank Mrs. Petra Wachs for her excellent technical assistance. This project was supported by institutional funding.
- Haagensen DE, Mazoujian G, Holder WD, Kister SJ, Wells SA: Evaluation of a breast cyst fluid protein detectable in the plasma of breast carcinoma patients. Ann Surg. 1977, 185: 279-285.PubMed CentralView ArticlePubMed
- Mazoujian G, Pinkus GS, Davis S, Haagensen DE: Immunohistochemistry of a gross cystic disease fluid protein (GCDFP-15) of the breast. A marker of apocrine epithelium and breast carcinomas with apocrine features. Am J Pathol. 1983, 110: 105-112.PubMed CentralPubMed
- Wick MR, Lillemoe TJ, Copland GT, Swanson PE, Manivel JC, Kiang DT: Gross cystic disease fluid protein-15 as a marker for breast cancer: immunohistochemical analysis of 690 human neoplasms and comparison with alpha-lactalbumin. Hum Pathol. 1989, 20: 281-287.View ArticlePubMed
- Loos S, Schulz KD, Hackenberg R: Regulation of GCDFP-15 expression in human mammary cancer cells. Int J Mol Med. 1999, 4: 135-140.PubMed
- Debily MA, Marhomy SE, Boulanger V, Eveno E, Mariage-Samson R, Camarca A, Auffray C, Piatier-Tonneau D, Imbeaud S: A functional and regulatory network associated with PIP expression in human breast cancer. PLoS One. 2009, 4: e4696-PubMed CentralView ArticlePubMed
- Vranic S, Schmitt F, Sapino A, Costa JL, Reddy S, Castro M, Gatalica Z: Apocrine carcinoma of the breast: a comprehensive review. Histol Histopathol. 2013, 28: 1393-409.PubMed
- Honma N, Takubo K, Akiyama F, Sawabe M, Arai T, Younes M, Kasumi F, Sakamoto G: Expression of GCDFP-15 and AR decreases in larger or node-positive apocrine carcinomas of the breast. Histopathology. 2005, 47: 195-201.View ArticlePubMed
- Honma N, Takubo K, Arai T, Younes M, Kasumi F, Akiyama F, Sakamoto G: Comparative study of monoclonal antibody B72.3 and gross cystic disease fluid protein-15 as markers of apocrine carcinoma of the breast. APMIS. 2006, 114: 712-719.View ArticlePubMed
- Eusebi V, Millis RR, Cattani MG, Bussolati G, Azzopardi JG: Apocrine carcinoma of the breast. A morphologic and immunocytochemical study. Am J Pathol. 1986, 123: 532-541.PubMed CentralPubMed
- Vranic S, Tawfik O, Palazzo J, Bilalovic N, Eyzaguirre E, Lee LM, Adegboyega P, Hagenkord J, Gatalica Z: EGFR and HER-2/neu expression in invasive apocrine carcinoma of the breast. Mod Pathol. 2010, 23: 644-653.View ArticlePubMed
- Farmer P, Bonnefoi H, Becette V, Tubiana-Hulin M, Fumoleau P, Larsimont D, Macgrogan G, Bergh J, Cameron D, Goldstein D, Duss S, Nicoulaz AL, Brisken C, Fiche M, Delorenzi M, Iggo R: Identification of molecular apocrine breast tumours by microarray analysis. Oncogene. 2005, 24: 4660-4671.View ArticlePubMed
- Doane AS, Danso M, Lal P, Donaton M, Zhang L, Hudis C, Gerald WL: An estrogen receptor-negative breast cancer subset characterized by a hormonally regulated transcriptional program and response to androgen. Oncogene. 2006, 25: 3994-4008.View ArticlePubMed
- Sanga S, Broom BM, Cristini V, Edgerton ME: Gene expression meta-analysis supports existence of molecular apocrine breast cancer with a role for androgen receptor and implies interactions with ErbB family. BMC Med Genomics. 2009, 11: 59-View Article
- von Minckwitz G, Blohmer JU, Raab G, Löhr A, Gerber B, Heinrich G, Eidtmann H, Kaufmann M, Hilfrich J, Jackisch C, Zuna I, Costa SD: In vivo chemosensitivity-adapted preoperative chemotherapy in patients with early-stage breast cancer: the GEPARTRIO pilot study. Ann Oncol. 2005, 16: 56-63.View ArticlePubMed
- von Minckwitz G, Kümmel S, Vogel P, Hanusch C, Eidtmann H, Hilfrich J, Hilfrich J, Gerber B, Huober J, Costa SD, Jackisch C, Loibl S, Mehta K, Kaufmann M: Intensified neoadjuvant chemotherapy in early-responding breast cancer: phase III randomized GeparTrio study. J Natl Cancer Inst. 2008, 100: 552-562.View ArticlePubMed
- von Minckwitz G, Kümmel S, Vogel P, Hanusch C, Eidtmann H, Hilfrich J, Hilfrich J, Gerber B, Huober J, Costa SD, Jackisch C, Loibl S, Mehta K, Kaufmann M: Neoadjuvant vinorelbine-capecitabine versus docetaxel-doxorubicin-cyclophosphamide in early nonresponsive breast cancer: phase III randomized GeparTrio trial. J Natl Cancer Inst. 2008, 100: 542-551.View ArticlePubMed
- von Minckwitz G, Blohmer JU, Costa SD, Denkert C, Eidtmann H, Eiermann W, Gerber B, Hanusch C, Hilfrich J, Huober J, Jackisch C, Kaufmann M, Kümmel S, Paepke S, Schneeweiss A, Untch M, Zahm DM, Mehta K, Loibl S: Response-Guided Neoadjuvant Chemotherapy for Breast Cancer. J Clin Oncol. 2013, 31: 3623-3630.View ArticlePubMed
- Wolff AC, Hammond ME, Schwartz JN, Hagerty KL, Allred DC, Cote RJ, Dowsett M, Fitzgibbons PL, Hanna WM, Langer A, McShane LM, Paik S, Pegram MD, Perez EA, Press MF, Rhodes A, Sturgeon C, Taube SE, Tubbs R, Vance GH, van de Vijver M, Wheeler TM, Hayes DF: American Society of Clinical Oncology; College of American Pathologists. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. J Clin Oncol. 2007, 25: 118-145.View ArticlePubMed
- Hammond ME, Hayes DF, Dowsett M, Allred DC, Hagerty KL, Badve S, Fitzgibbons PL, Francis G, Goldstein NS, Hayes M, Hicks DG, Lester S, Love R, Mangu PB, McShane L, Miller K, Osborne CK, Paik S, Perlmutter J, Rhodes A, Sasano H, Schwartz JN, Sweep FC, Taube S, Torlakovic EE, Valenstein P, Viale G, Visscher D, Wheeler T, Williams RB, Wittliff JL, Wolff AC: American Society of Clinical Oncology/College Of American Pathologists guideline recommendations for immunohistochemical testing of estrogen and progesterone receptors in breast cancer. J Clin Oncol. 2010, 28: 2784-2795.PubMed CentralView ArticlePubMed
- Loibl S, Müller BM, von Minckwitz G, Schwabe M, Roller M, Darb-Esfahani S, Ataseven B, du Bois A, Fissler-Eckhoff A, Gerber B, Kulmer U, Alles JU, Mehta K, Denkert C: Androgen receptor expression in primary breast cancer and its predictive and prognostic value in patients treated with neoadjuvant chemotherapy. Breast Cancer Res Treat. 2011, 130: 477-487.View ArticlePubMed
- Huo L, Zhang J, Gilcrease MZ, Gong Y, Wu Y, Zhang H, Resetkova E, Hunt KK, Deavers MT: Gross cystic disease fluid protein-15 and mammaglobin A expression determined by immunohistochemistry is of limited utility in triple-negative breast cancer. Histopathology. 2013, 62: 267-274.View ArticlePubMed
- Lewis GH, Subhawong AP, Nassar H, Vang R, Illei PB, Park BH, Argani P: Relationship between molecular subtype of invasive breast carcinoma and expression of gross cystic disease fluid protein 15 and mammaglobin. Am J Clin Pathol. 2011, 135: 587-591.PubMed CentralView ArticlePubMed
- Weigelt B, Horlings HM, Kreike B, Hayes MM, Hauptmann M, Wessels LF, de Jong D, Van de Vijver MJ, Van't Veer LJ, Peterse JL: Refinement of breast cancer classification by molecular characterization of histological special types. J Pathol. 2008, 216: 141-150.View ArticlePubMed
- Pagani A, Sapino A, Eusebi V, Bergnolo P, Bussolati G: PIP/GCDFP-15 gene expression and apocrine differentiation in carcinomas of the breast. Virchows Arch. 1994, 425: 459-465.View ArticlePubMed
- Fritzsche FR, Thomas A, Winzer KJ, Beyer B, Dankof A, Bellach J, Dahl E, Dietel M, Kristiansen G: Co-expression and prognostic value of gross cystic disease fluid protein 15 and mammaglobin in primary breast cancer. Histol Histopathol. 2007, 22: 1221-1230.PubMed
- Luo MH, Huang YH, Ni YB, Tsang JY, Chan SK, Shao MM, Tse GM: Expression of mammaglobin and gross cystic disease fluid protein-15 in breast carcinomas. Hum Pathol. 2013, 44: 1241-1250.View ArticlePubMed
- Hall RE, Clements JA, Birrell SN, Tilley WD: Prostate-specific antigen and gross cystic disease fluid protein-15 are co-expressed in androgen receptor-positive breast tumours. Br J Cancer. 1998, 78: 360-365.PubMed CentralView ArticlePubMed
- Lakhani S, Ellis IO, Tan PH, van de Vijver MJ: WHO Classification of Tumors of the Breast. 2012, Lyon: IARC Press
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2407/14/546/prepub
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 credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.