This article has Open Peer Review reports available.
Women with familial risk for breast cancer have an increased frequency of aldehyde dehydrogenase expressing cells in breast ductules
© Isfoss et al.; licensee BioMed Central Ltd. 2013
Received: 22 April 2013
Accepted: 30 October 2013
Published: 4 November 2013
Knowledge is limited regarding the association between stem cells in histologically benign breast tissue and risk factors for breast cancer, and hence we addressed this issue in the present study. Recently, we assessed the histology of benign breast tissue from cancer and non-cancer patients for cells positive for the putative stem cell marker aldehyde dehydrogenase 1 A1 (ALDH), and the findings indicated an association between expression of ALDH and the hormonal factors menopause and hormone therapy. The current investigation examined possible associations between various known clinical and genetic risk factors for breast cancer and cellular expression of ALDH in ductules in benign human breast tissue.
The study included breast surgery patients that were BRCA1/2 mutation carriers without breast cancer (n = 23), had BRCA1/2 (n = 28) or sporadic (n = 21) breast cancer, or required non-cancer-related mammoplasty (n = 34). The distribution and frequency of ALDH-immunolabelled cells were correlated to patient subgroups with different risk factors, using mammoplasty patients as a control group. Statistical analyses comprised linear and logistic regression, Spearman’s rank test, Pearson’s test, and Fisher’s exact test. In two-tailed tests, p < 0.05 was considered significant.
A strong association was found between family history of breast cancer and a high frequency of ALDH+ cells (p = 0.001) at all ductular levels in all groups, regardless of BRCA status, age, parity, or occurrence of cancer. In pre-menopausal non-BRCA cancer patients, the frequency of ALDH+ cells increased with age (p < 0.01) but decreased with increasing parity (p < 0.03). High frequencies of ALDH+ cells were found in the non-basal ductular levels in BRCA1 mutation carriers (p = 0.03), but in the basal ductular level in BRCA2 cancer patients (p = 0.02). Among post-menopausal patients, only on-going hormone replacement therapy was correlated with a high number of ALDH+ cells (p < 0.03).
In histologically normal breast tissue, we found a positive association between the frequency of ductular ALDH+ cells and several breast cancer risk factors, particularly family history of this disease, which supports previous evidence that ALDH plays a role in breast cancer.
A stem-like cancer cell is defined as a cell with the capacity for self-renewal and the ability to generate different cell types that form a tumor . In 2003, the presence of benign stem cells in non-histological breast tissue preparations was identified on the basis of functional cell characteristics [2, 3], and a brief histological description of these cells was published in 2007 . It has been proposed that aldehyde-dehydrogenase-expressing (ALDH+) breast cells include malignant stem-like cell populations that maintain and cause progression of cancer [4, 5]. ALDH catalyzes oxidation and is necessary for retinoic acid synthesis, and it has been suggested that this enzyme is also involved in stem cell preservation and initiation of differentiation . Normal mammary stem cells and stem-like cancer cells were recently shown to express an alternative isoform of ALDH , but the stem cell properties or cancer-related functions of cells that are positive for any isoform still remain unclear.
Several clinical and molecular factors predict the risk of developing breast cancer. For example, the presence of BRCA1 and BRCA2 gene mutations confers a 40–80% lifetime risk , and exogenous hormone exposure is associated with relative risks of 1.5 to 2.0 . Furthermore, non-malignant histological changes such as atypical ductal hyperplasia are associated with a 28% risk of breast cancer .
ALDH+ cells are found in up to 48% of breast cancer tumors and are believed to cause late recurrence, and these cells are also associated with an adverse prognosis and poor outcome after conventional anti-cancer drug treatment [5, 11, 12].
To assess the potential of ALDH as a predictive marker for subsequent development of breast cancer, it is necessary to define the normal ranges of frequencies and distribution of ALDH+ cells in histologically benign breast tissue in women with and without breast cancer. Such data are limited at present, although a small case-control study showed elevated levels of ALDH+ cells in epithelium and stroma of patients who later developed cancer , and an investigation of African women revealed a higher frequency of ALDH+ cells in breast cancer tissue compared to benign breast tissue . It is also essential to determine whether the frequency or distribution of ALDH+ putative stem cells in histologically normal breast tissue is related to risk factors for breast cancer. We recently described in detail the distribution of ALDH+ cells in terminal duct lobular units (TDLUs) and stroma in benign breast tissues . The correlations with risk factors that were observed in the small group of patients assessed in that investigation suggested that density and distribution of ALDH+ cells are associated with menopausal state and hormone replacement therapy. Therefore, our aim in the present study was to examine ALDH expression in ductular cells in a larger patient group and to elucidate the relationship between such expression and various risk factors for breast cancer, including BRCA1/2 mutation status, familial breast cancer history, hormone intake, parity, and age at menarche.
The patients considered for inclusion in this study were female patients who had been treated with breast surgery in Skåne County and met the criteria for one of the following groups:
Group A (n = 30): breast cancer patients without BRCA1/2 mutation; surgery during the period 1999–2006.
Group B (n = 19): breast cancer patients with BRCA1 mutation; surgery during the period 1984–2009.
Group C (n = 16): breast cancer patients with BRCA2 mutation; surgery during the period 1984–2009.
Group D (n = 13): BRCA1 mutation carriers without breast cancer; prophylactic mastectomy during the period 1996–2010.
Group E (n = 13): BRCA2 mutation carriers without breast cancer; prophylactic mastectomy during the period 1996–2010.
Group F (n = 35): patients without breast cancer or BRCA1/2 mutations; mammoplasty during the period 1993–1994.
Patient groups in the study
Breast cancer and mutation status
Number of patients
Age in years, median (range)
Hormone use at time of surgery
1 or 2
3 or 4
Prophylactic mastectomy, BRCA1
Prophylactic mastectomy, BRCA2
Mammoplasty, non-cancer, non-BRCA
Data on the following clinical parameters were available for the majority of the patients: indication for surgery, age at time of surgery, current or previous use and total duration of use of oral contraceptives and or hormone replacement therapy (HRT), age at menarche, number of live childbirths (parity), family history of breast cancer (1st or 2nd degree relative), and BRCA1/2 mutation status. Table 1 presents information on the age, hormonal exposure, and reproductive status of the women included in the study.
Sixty-seven of the women were pre-menopausal, 35 were post-menopausal, and data were missing for four. The median age at onset of menopause was 49 years (range, 35–55 years). The numbers of live births among the women were as follows: para 0 in 32, para 1 in 11, para 2 in 28, para 3 in 25, and para 4 in four. The median age at menarche was 13 years (range, 10–18 years; data missing for 32 women). Data regarding exogenous hormone treatment status at the time of surgery were as follows: 70 women had no such treatment, 15 used oral contraceptives, seven received HRT, and six had a progestin intrauterine device. Sixty-five women had used oral contraceptives at sometime during their lives (median duration, 7.1 years), and 11 had received HRT at some point (median duration, 5.0 years). Clinical data and breast tissue samples representing 25 of the 28 patients in our previous study  were included in the present analysis.
Histology and immunohistochemistry
If two tissue blocks from the same patient were of similar superior quality, both (from the same or both breasts) were analyzed in the same manner for quality assurance purposes. New sections were obtained from all relevant paraffin tissue blocks, and consecutive sections were H&E stained and used for immunohistochemistry, as described previously in detail . Briefly, after antigen retrieval, the sections were incubated with mouse monoclonal antibodies directed against human ALDH1 A1 (aa 7–128, N-term, diluted 1:100; Becton, Dickinson and Company, Franklin Lakes, NJ, USA). Detection of immunoreactive sites was performed using a system based on horseradish peroxidase (HRP) and di-aminobenzidine (DAPI) labelling of primary antibodies raised in mice (Envision+ System-HRP, Dako, Glostrup, Denmark). Sequential incubations were performed with secondary antibodies and HRP-conjugated antibodies. Following the peroxidase reaction in DAPI and H2O2, the sections were rinsed and counterstained with hematoxylin, and then dehydrated and coverslipped in Mountex medium (Histolab Products AB, Göteborg, Sweden). To confirm the specificity of ALDH labelling, slides that were not incubated with the primary antibodies were included in each immunohistochemistry run.
The tissue sections selected for ALDH immunolabeling were assessed regarding the following features: tissue fragment size, epithelial atypia or hyperplasia, invasive carcinoma, and total number of TDLUs. The tissue areas available in each chosen block ranged in size from 48 to 621 mm2with a median value of 254 mm2. Areas displaying ductular or lobular hyperplasia, atypia, carcinoma in situ (four cases), or invasive carcinoma (10 cases) were excluded from further analysis. To evaluate the distribution and number of immunolabelled ALDH+ cells, 50 TDLUs were examined, guided by a grid, at evenly spaced intervals in each tissue section; if fewer than 50 TDLUs were present in a section, all the TDLUs were evaluated. The number of TDLUs found to be available for analysis of ALDH immunoreactivity in each patient ranged from 13 to 50 (mean, 45 TDLUs). Only ductules (i.e., no ducts or stroma) were evaluated. In our previous study , we described ALDH+ cells as being in a basal, luminal, or intermediate location. For each patient in the present study, the number of TDLU cross sections with ductular ALDH+ cells was recorded for each ductular level (luminal, intermediate, or basal) in relation to the total number of TDLU cross sections examined.
ALDH immunoreactive cells that were labelled either strongly or weakly were considered to be positive for ALDH. Furthermore, cells found in luminal and intermediate ductular locations were defined as non-basal (luminal plus intermediate) in some analyses, as indicated in the text. This was done because determination of cell location by ductular level is less accurate when ductules lack microscopically discernible lumina.
Blind testing of ALDH+ cell frequency in duplicate tissue samples from 13 patients yielded correlation coefficients of 0.45 (Pearson) and 0.53 (Spearman), indicating similarity between independent observations (rs = 0.65, p = 0.01). These results confirmed the robustness of the analysis and demonstrated similarity between different areas of breast tissue in the same patient.
The number of TDLUs with detectable ALDH+ cells and basal and non-basal localization of those cells were analyzed by linear regression using occurrence of ALDH+ cells in the mammoplasty group as the independent variable. Logistic regression was used to analyze differences between risk factor groups. Correlation between ALDH findings for duplicate samples from the same patient was analyzed using Spearman’s rank test and Pearson’s test. Two-tailed tests were performed to assess the level of statistical significance, and results with p < 0.05 were considered significant. SPSS 18 software (IBM, Armonk, NY, USA) was used for all statistical analyses.
The study was approved by the Research Ethics Committee for Southern Sweden (approval 11-92, 349-00).
Results and discussion
Division of patient inclusion groups A–F into analysis groups
Patients in inclusion groups A–F were divided into relevant groups on the basis of genetic and hormonal status to enable various analyses of ALDH+ cells.
Distribution of ALDH+ cells
ALDH+ cells in the six patient groups
ALDH+ cells were present in the epithelium at some ductular level in 90% of the subjects (19/21) in Group A (breast carcinoma with no BRCA mutations) and in a similar fraction (94%, 16/17) of the patients in Group B (breast carcinoma with BRCA1 mutation), but in a smaller fraction (73%, 8/11) of those in Group C (breast carcinoma with BRCA2 mutation). Also, ALDH+ ductular cells were detected in 100% (12/12) of the patients in group D (prophylactic mastectomy due to a BRCA1 mutation) and 100% (11/11) of those in Group E (prophylactic mastectomy due to a BRCA2 mutation). In Group F (mammoplasty patients with neither cancer nor BRCA1/2 mutations), ALDH+ ductular cells were found in 76% (26/34) of the patients.
There was no statistically significant difference in the frequency of ALDH+ cells among the patient groups (p = 0.09), although a higher frequency was noted for patient Groups A–E combined compared to the mammoplasty group (p = 0.06, Fisher’s exact test). This finding, together with previous evidence that ALDH+ cells in different levels of the ductular epithelium are differentially associated with risk factors for breast cancer,  prompted a detailed analysis of patient subgroups regarding the occurrence of ALDH+ cells in the various levels of the ductular epithelium (see below).
ALDH+ cells in patients with a family history of breast cancer but no BRCA1/2 mutations
Median percentage of TDLUs containing ALDH+ cells (range)
Luminal ductular level
Intermediate ductular level
Basal ductular level
Pre-menopausal with family history of breast cancer, including patients with BRCA1 or BRCA2 mutation (n = 37)
Pre-menopausal with BRCA1 mutation (n = 18)
Pre-menopausal with BRCA2 mutation (n = 13)
Pre-menopausal without family history of breast cancer (n = 25)
Post-menopausal receiving HRT at the time of surgery (n = 7)
Post-menopausal not receiving HRT at the time of surgery (n = 24)
Nocancer, underwent mammoplasty (n = 34)
ALDH+ cells in patients with a BRCA1 or BRCA2 mutation
Compared to breast tissue from mammoplasty patients, such tissue from pre-menopausal women who underwent preventive mastectomy due to BRCA1 carrier status contained significantly larger numbers of ALDH+ cells at non-basal ductular levels (p = 0.03; see Figure 2B). A positive association of borderline significance (p = 0.06) was noted for ALDH+ cells located at the basal ductular level in the BRCA1 mutation carriers. No difference in the occurrence or distribution of ALDH+ cells in breast tissue was observed between women with breast cancer and BRCA1 gene mutations and those who underwent mammoplasty. Larger numbers of ALDH+ cells at non-basal ductular levels were discerned in tissue from pre-menopausal women with breast cancer and BRCA2 gene mutations than in tissue from mammoplasty patients (p = 0.02; Figure 2C). The post-menopausal BRCA1 and BRCA2 patient subgroups showed no significant differences compared to the mammoplasty patients. Table 2 gives further information regarding these results.
It should also be mentioned that a higher frequency of ALDH+ cells in ductules was found in patients with BRCA1 mutations than in those with BRCA2 mutations. This may reflect the dissimilarities in oncogenesis between different types of BRCA mutations that have been described by other investigators .
ALDH+ cells in relation to age and menarche
ALDH+ cells in relation to parity
ALDH+ cells in relation to HRT and oral contraceptives
In contrast, neither the total duration of contraceptive pill usage nor the duration of HRT was associated with the frequency of ALDH+ cells, which suggests that exogenous hormones have only a transient effect on the occurrence of ALDH+ cells in benign breast tissue, and is consistent with previously reported data on sex hormone therapies suggesting that ongoing or recent intake of HRT increases the risk of breast cancer .
ALDH+ cells in relation to cancer versus non-cancer
The present findings support the hypothesis that ALDH+ breast cell populations participate in oncogenesis. We found several associations between high frequencies of ALDH+ cells in benign breast tissue and clinical risk factors for cancer. Also, statistical analysis of subgroups demonstrated significant correlations between ALDH+ cells and family history of breast cancer, parity, and HRT exposure (see Table 2). Larger studies are needed to assess other risk factors for breast cancer in this context.
The patients evaluated in the present study were treated between 1984 and 2010, and the surgical techniques and oncological treatment protocols being applied during that period were subject to changes. However, the potential impact of those modifications on the results of our investigation can be assumed to be negligible for the following reasons: (1) patients who received neoadjuvant therapy were excluded; (2) possible influence of variation in antigenicity related to heterogeneous tissue fixation was reduced to a minimum by using ALDH positivity in stromal cells as internal control, and by considering all cells exhibiting weak or strong immunoreactivity as positive for expression of ALDH.
Our study demonstrated associations between ALDH+ cell frequencies in benign breast tissue and risk factors for breast cancer, but we found no significant relationship between ALDH+ cell frequencies in benign tissue and the co-presence of cancer. This finding might be explained by the influence of temporal and spatial factors of a developing cancer; more specifically, that the cancer is established after the occurrence of a risk factor, and the tumor can overgrow ALDH+ cells.
In our previous study conducted using the same methods , a higher frequency of ALDH+ ductular cells was found in mammoplasty patients than in patients with established breast cancer. In contrast, in the present investigation, data on mammoplasty patients were used to allow comparison of non-breast cancer and breast cancer patients, the latter group including many individuals with a familial risk of breast cancer. However, inasmuch as increased cell mass in a breast predisposes to cancer [22, 23], it would be most suitable to obtain control tissue from patients without breast disease, not from those with breast hyperplasia.
The present study demonstrated a positive association between the frequency of ALDH+ ductular cells in benign female breast tissue and several well-known risk factors for breast cancer. These findings suggest that ALDH cell positivity in microscopically normal breast tissue should be further evaluated as a potential marker for the risk of breast cancer, especially in patients with familial breast cancer with or without BRCA1/2 mutations.
The largest differences in ALDH+ cell frequency between risk groups were observed in the luminal and intermediate (non-basal) ductular locations. Our data indicate that ALDH-expressing cells may play a yet undefined role in the development of female breast cancer, and they also suggest that immunohistochemical analysis of ALDH can aid prediction of the risk of breast cancer. It is conceivable that having a high number of ALDH+ stem cells promotes a high number of progenitor cells, and that the latter are susceptible to oncogenesis. The current results underline the need for further research to assess the previously proposed stem and/or progenitor qualities of ALDH+ cells in normal breast tissue and their potential role in carcinogenesis.
This study was supported by grants from the Swedish Medical Research Council, the Swedish Cancer Society, and local funds at the University Hospital and Region Skane. We acknowledge the support of European Research Council Advanced Grant ERC-2011-294576. We extend many thanks to Cristina Ciornei and Lena Tran at the pathology laboratory of the university hospital in Lund for excellent technical assistance.
- Clarke MF, Dick JE, Dirks PB, Eaves CJ, Jamieson CH, Jones DL, Visvader J, Weissman IL, Wahl GM: Cancer stem cells—perspectives on current status and future directions: AACR workshop on cancer stem cells. Cancer Res. 2006, 66: 9339-9344. 10.1158/0008-5472.CAN-06-3126.View ArticlePubMedGoogle Scholar
- Alvi AJ, Clayton H, Joshi C, Enver T, Ashworth A, Vivanco M, Dale TC, Smalley MJ: Functional and molecular characterisation of mammary side population cells. Breast Cancer Res. 2003, 5: R1-R8.View ArticlePubMedGoogle Scholar
- Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF: Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci. 2003, 100: 3983-3988. 10.1073/pnas.0530291100.View ArticlePubMedPubMed CentralGoogle Scholar
- Ginestier C, Hur MH, Charafe-Jauffret E, Monville F, Dutcher J, Brown M, Jacquemier J, Viens P, Kleer CG, Liu S, Schott A, Hayes D, Birnbaum D, Wicha MS, Dontu G: ALDH1 is a marker of normal and malignant human mammary stem cells and a predictor of poor clinical outcome. Cell Stem Cell. 2007, 15: 555-567.View ArticleGoogle Scholar
- Tanei T, Morimoto K, Shimazu K, Kim SJ, Tanji Y, Taguchi T, Tamaki Y, Noguchi S: Association of breast cancer stem cells identified by aldehyde dehydrogenase 1 expression with resistance to sequential paclitaxel and epirubicin-based chemotherapy for breast cancers. Clin Cancer Res. 2009, 15: 4234-4241. 10.1158/1078-0432.CCR-08-1479.View ArticlePubMedGoogle Scholar
- Ma I, Allan AL: The role of human aldehyde dehydrogenase in normal and cancer stem cells. Stem Cell Rev. 2011, 7: 292-306. 10.1007/s12015-010-9208-4.View ArticlePubMedGoogle Scholar
- Eirew P, Kannan N, Knapp DJ, Valliant F, Emerman JT, Lindeman GJ, Visvader JE, Eaves CJ: Aldehyde dehydrogenase activity is a biomarker of primitive normal human mammary luminal cells. Stem Cells. 2012, 30: 344-348. 10.1002/stem.1001.View ArticlePubMedGoogle Scholar
- Fackenthal JD, Olopade OI: Breast cancer risk associated with BRCA1 and BRCA2 in diverse populations. Nat Rev Cancer. 2007, 7: 937-948. 10.1038/nrc2054.View ArticlePubMedGoogle Scholar
- Beral V, Reeve G, Bull D, Green J, Million Women Study Collaborators: Breast cancer risk in relation to the interval between menopause and starting hormone therapy. J Natl Cancer Inst. 2011, 103: 296-305. 10.1093/jnci/djq527.View ArticlePubMedPubMed CentralGoogle Scholar
- Sinn HP, Elsawaf Z, Helmchen B, Aulmann S: Early breast cancer precursor lesions: lessons learned from molecular and clinical studies. Breast Care. 2010, 5: 218-226. 10.1159/000319624.View ArticlePubMedPubMed CentralGoogle Scholar
- Reya T, Morrison SJ, Clarke MF, Weissmann IL: Stem cells, cancer, and cancer stem cells. Nature. 2001, 414: 105-111. 10.1038/35102167.View ArticlePubMedGoogle Scholar
- Nalwoga H, Arnes JB, Wabinga H, Akslen LA: Expression of aldehyde dehydrogenase (ALDH1) is associated with basal-like markers and features of aggressive tumours in African breast cancer. Br J Cancer. 2010, 102: 369-375. 10.1038/sj.bjc.6605488.View ArticlePubMedGoogle Scholar
- Kunju LP, Cookingham C, Toy KA, Chen W, Sabel MS, Kleer CG: EZH2 and ALDH-1 mark breast epithelium at risk for breast cancer developement. Mod Pathol. 2011, 24: 786-793. 10.1038/modpathol.2011.8.View ArticlePubMedPubMed CentralGoogle Scholar
- Schwartz T, Stark A, Pang J, Awuah B, Kleer CG, Quayson S, Kingman S, Aitpillah F, Jiagge E, Oppong JK, Osei-Bonsu E, Martin I, Yan X, Toy K, Adjei E, Wicha M, Newman LA: Expression of aldehyde dehydrogenase 1 as a marker of mammary stem cells in benign and malignant breast lesions of Ghanaian women. Cancer. 2013, 119: 488-494. 10.1002/cncr.27737.View ArticlePubMedGoogle Scholar
- Isfoss BL, Holmqvist B, Alm P, Olsson H: Distribution of aldehyde dehydrogenase 1-positive stem cells in benign mammary tissue from women with and without breast cancer. Histopathology. 2012, 60: 617-633. 10.1111/j.1365-2559.2011.04109.x.View ArticlePubMedGoogle Scholar
- Ashworth A, Weber BL, Domchek SM: Inherited genetic factors and breast cancer. Diseases of the breast. Edited by: Harris JR, Morrow M, Lippman ME, Osborne CK. 2010, Philadelphia: Wolters Kluwer, 209-220.Google Scholar
- Schonfeld SJ, Pfeiffer RM, Lacey JV, Berrington de Gonzalez A, Doody MM, Greenlee RT, Park Y, Schairer C, Schatzkin A, Sigurdson AJ, Hartge P, Visvanathan K: Hormone-related risk factors and postmenopausal breast cancer among nulliparous versus parous women: an aggregated study. Am J Epidemiol. 2011, 173: 509-517. 10.1093/aje/kwq404.View ArticlePubMedPubMed CentralGoogle Scholar
- Collaborative Group on Hormonal Factors in Breast Cancer: Breast cancer and hormone replacement therapy. Collaborative reanalysis of data from 51 epidemiological studies of 52 705 women with breast cancer and 108 411 women without breast cancer. Lancet. 1997, 350: 1047-1059.View ArticleGoogle Scholar
- Olsson HL, Ingvar C, Bladström A: Hormone replacement therapy containing progestins and given continuously increases breast carcinoma risk in Sweden. Cancer. 2003, 97: 1387-1392. 10.1002/cncr.11205.View ArticlePubMedGoogle Scholar
- Beral V, Million Women Study Collaborators: Breast cancer and hormone-replacement therapy in the Million Women Study. Lancet. 2003, 362: 419-427.View ArticlePubMedGoogle Scholar
- Anderson GL, Judd HL, Kaunitz AM, Barad DH, Beresford SA, Pettinger M, Liu J, McNeeley SG, Lopez AM, Women’s Health Initiative Investigators: Effects of estrogen plus progestin on gynecologic cancers and associated diagnostic procedures: the women’s health initiative randomized trial. JAMA. 2003, 290: 1739-1748. 10.1001/jama.290.13.1739.View ArticlePubMedGoogle Scholar
- Albanes D, Winick M: Are cell number and cell proliferation risk factors for cancer?. J Natl Cancer Inst. 1988, 80: 772-774. 10.1093/jnci/80.10.772.View ArticlePubMedGoogle Scholar
- Olsson H: Risk for malignant tumors after oral contraceptive use: is it related to organ size while taking the pill?. Med Oncol Tumor Pharmacother. 1990, 7: 61-64.PubMedGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1472-6890/13/28/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 cited.