Recent data regarding the intratumor heterogeneity in renal cancer have resulted in a heightened awareness in the scientific community about this topic.1 To recap, this study analyzed different regions from 10 different renal cell carcinomas and showed marked variations in the mutation spectrum in these areas. Tumor heterogeneity has also been well documented in breast cancer including preclinical and clinical setting.2–4 Besides the intertumor heterogeneity, the presence of intratumor heterogeneity is enormously significant in this era of targeted therapies. It raises multiple important questions regarding the efficacy of targeted therapies and strategies that would be needed to combat resistance. In this review, we will reexamine the issues related to tumor heterogeneity, both intratumor and intertumor heterogeneity, and provide guidelines on how to deal with these factors in clinical situations, focusing on breast cancer.
EXTENT OF TUMOR HETEROGENEITY?
Layers of Tumor Heterogeneity
In an idealized situation, cancer would be just one condition in which all cells are identical in their genetic content. More importantly, the cells would be susceptible to a single “noxious” agent, the administration of which would result in complete destruction of the entire population and thus a “cure.” Anyone who has ever come in contact with cancer knows that this is far from the truth. Patients present with small and large tumors that progress at different speeds and result in different outcomes after getting the standard therapies. Some cancers are associated with calcifications and can be easily detected by mammographic screening, whereas others are radiologically invisible despite being palpable. Similarly, the presence and extent of nodal involvement at presentation can be completely variable in patients with similarly sized primary tumors. The tumor heterogeneity exists at multiple layers as summarized below (Fig. 1).
The World Health Organization classification recognizes morphologic heterogeneity in breast cancer.5 The most recent version of the classification categorizes breast cancer in 17 different subtypes on the basis of the histologic characteristics (Fig. 2). The most common form of invasive carcinoma forms ductal or glandular structures and is commonly referred to as invasive ductal carcinoma. This designation “ductal” is being increasingly questioned for its appropriateness and has been eliminated from common usage in some countries such as United Kingdom. However, in most parts of the world, including the United States, it still remains in common usage. The second common subtype is referred to as “lobular” carcinoma. This term is equally inappropriate as the tumor does not form lobules. However, it was originally used to differentiate these tumors from the “ductal” subtype. There was also a hypothesis that these tumors possibly arise from lobules. However, there is no evidence of this, and both types of tumors are now believed to arise from terminal ductal lobular units. In addition to the classical ductal and lobular types, a number of distinct histologic patterns of breast cancer are described. Tumors with squamous differentiation and differentiation toward nonepithelial elements are also well described.
Tumor heterogeneity also forms the basis of the grading system. Simply stated, grade is a measure of the similarity (or lack thereof) that the tumor has to the normal epithelium. Thus a low-grade tumor resembles the normal breast epithelium more than a high-grade tumor. The grading system has evolved from the original proposed in the early 1940s. The current system, Nottingham modification of the Scarff-Bloom Richardson, quantifies 3 parameters: gland formation, nuclear pleomorphism, and mitotic activity.6 The impact of grade on prognosis has been documented in multiple studies. Tumors of low histologic grade are associated with good prognosis and those high grade are associated with bad prognosis. This is true even after the patients are stratified by other clinicopathologic parameters such as age, tumor size, and extent of disease. Furthermore, as chemotherapy is based on targeting proliferating cells, high-grade tumors are more likely to respond to these therapies. This is best illustrated in neoadjuvant chemotherapy studies wherein pathologic complete response (pCR) is rarely noted in low-grade tumors.
It is not uncommon to observe foci within the same tumors that have distinct morphologic features. Gland or tubule formation tends to be very heterogenous. Small foci of tubular differentiation are commonly noted in tumors that otherwise have the morphology of classic lobular carcinomas. It is debatable whether these small foci merit reclassification of these tumors as mixed ductal-lobular carcinomas, particularly because, from the oncological standpoint, it does not make any difference. Tumors with a lobular growth pattern are difficult to feel for the surgeons and are therefore more likely to require reexcisions. For this reason, in our practice, we routinely report the presence of a single cell/diffuse (lobular-like) growth pattern within tumors. The presence of higher-grade foci within tumors (as compared with biopsies) is well recognized; this can cause difficulties in reporting of HER2 expression. The ASCO-CAP guidelines recommend reanalysis of the tumors for HER2 expression in such situations.7
The tumor is composed of millions of cells distributed in sheets, chords, and small islands distributed within the tumor stroma. Depending on the proximity of fibrous tissue and blood vessels (or lack thereof), the tumor cells may show varying degrees of hypoxic change. In most cases, this is subtle; however, in extreme cases, this might be seen in the form of geographic necrosis (Fig. 1).
Heterogeneity in the Tumor Microenvironment
Apart from the variability in the tumor cells themselves, tumors might elicit different stromal responses. The amount and quality of the stromal reaction is of prognostic relevance. In the old days, this was quantified as “scirrhous”; tumors with prominent “elastic” fibers were shown to be of good prognosis. More recently, the availability of gene expression microarrays has enabled quantification of the stromal reaction in the form a gene signature. West and colleagues have classified tumors on the basis of the presence of “desmoid-fibromatosis–like reaction” and “solitary fibrous tumor–type reaction.”8,9 The former were associated with better prognosis. Similarly, Finak et al10 performed gene expression analysis on microdissected tumor stroma and adjacent “normal” stroma from patients with breast cancer. They identified a stroma-derived prognostic predictor that outperformed standard clinical prognostic factors in multiple data sets including sets containing HER2+ tumors. Its prognostic power was further improved in some data sets by the addition of other previously published parameters. Importantly, a number of genes related to immune response were identified in the stroma-derived prognostic predictor.
The role of immune cells in solid cancers is not well understood. It is not unusual to see lymphoid infiltrates in foci of regression in melanomas but the rest of the solid tumors are not thought of as having a significant “immune” component. Recent studies have suggested that immune cells play a major role in solid tumors. The evidence was first noted in gene signatures, wherein high expression of these genes was associated with prognosis.11 The tumor stroma in breast cancers is also composed of variable numbers of immune cells. The role of these cells in modulating the behavior of the tumor is far from clear. Lymphocytes are present in all types of tumors. In some of the older studies, the presence has been associated with prognosis, particularly in highly proliferating tumors. More recently, Denkert et al,12 on behalf of the German breast group, analyzed a large series of patients treated by neoadjuvant chemotherapy. In this study of more than a thousand patients, they identified the prognostic significance of “lymphocyte-predominant breast cancer,” the latter defined as >50% of the stroma containing mononuclear inflammatory cells. Loi et al13 analyzed a large series of patients treated in an adjuvant chemotherapy trial for the presence of lymphocytes. The study showed that the lymphoid infiltrate had little prognostic impact in ER+ tumors. However, in the ER− tumors (HER2+ and triple negative breast cancers), the presence of stromal lymphocytic infiltrate connoted a good prognosis in chemotherapy-treated populations.12–15 The data were further confirmed by our group.16 In the joint analysis of 2 Eastern cooperative oncology group trails (E1199 and E2197), we showed the prognostic benefit of tumor-infiltrating lymphocytes in triple-negative breast cancers. Importantly, the degree of benefit identified in all of these studies was almost identical. In addition, the benefit was also observed when the lymphocytic infiltrate was quantified in percentiles rather than as a dichotomous variable (greater or less than 50%). These studies suggest a role for immune therapies in breast cancer (Fig. 3). It is, however, not clear as to which subpopulation of lymphoid cells is associated with good prognosis. A number of studies have looked at the expression of CD4, CD8, and a number of other immune markers and have shown them to be associated with good prognosis. Most recently, Denkert et al17 analyzed a number of proinflammatory and anti-inflammatory signaling molecules in a series of breast cancers. This analysis showed that the presence of these molecules, irrespective the presumed function, was associated with better pCR rates. The expressions of other immune markers, such as CTLA-4, PD1, and its ligand PD1-L, are being analyzed in a number of studies with variable results.18 The issues that are being encountered are the specificity of antibodies, location of the staining, and the cutoff points to be used to determine positivity or negativity. Standardization of the methods of analysis for these markers is urgently needed. In this vein, it is important to note that the interactions between the cancer cells and T lymphocytes are mediated by at least 20 different molecules.19 In addition, the term lymphocytes in these studies is used in the generic manner and does not discriminate between the different types of lymphocytes, macrophages, myeloid-derived suppressor cells, or a host of other cells present in the vicinity of the tumor. There is sufficient literature to document a prognostic impact for each of the cell types in a number of cancers including breast cancer.
The importance of angiogenesis for the progression of tumors is well recognized in cancers. Epithelial cells by definition depend on imbibition of nutrients from nearby capillaries. It is therefore believed that neovascularization is necessary for tumors to grow and for invasion to be sustained.20 Newly formed vessels (neoangiogenesis) are thought to result in poorly formed vessels that are leaking and associated with blind loops. It is difficult to distinguish newly formed blood vessels from preexisting vessels. However, the presence of pericyte coverage around the endothelial cells is usually taken as evidence of well-formed vessels. Studies quantifying neovascularization using a variety of immunohistochemical (IHC) markers have shown that highly vascularized tumors tend to have poor outcomes. Targeted therapies directed at angiogenesis, although not successful in breast cancer,21,22 have been shown to be effective in a number of cancers such as glioblastomas, lung, and colon cancers. Therapies with antiangiogenic drugs have been shown to cause normalization of the vascular network as evidenced by pericytic coverage.23
HETEROGENEITY OF PROTEIN EXPRESSION
Apart from the obvious heterogeneity that can be observed by routine light microscopic examination of tumors, a marked heterogeneity has also been noted in protein expression using a variety of techniques. Hormone receptors are the most common proteins analyzed in breast cancer. The expression of estrogen receptor (ER) has been analyzed using ligand-binding assays and more recently using IHC and reverse transcription polymerase chain reaction.24 The studies confirmed that the expression within tumors is variable, ranging from near-complete absence to extremely high levels. The expression of the receptors can be extremely heterogenous with positive and negative cells arranged in an apparently random manner within the tumor. The expression of ER and PR (progesterone receptor) is a weak prognostic factor in breast cancer but a strong predictive factor.25 The minimal amount of expression required to generate endocrine response has been a matter of dispute for a prolonged period of time. The current ASCO-CAP guidelines26 have suggested that expression in 1% of tumor cells might be sufficient to warrant a therapeutic trial with endocrine agents. The percentage of ER expression has been correlated with likelihood of response to endocrine therapies with tumors having high expression being associated with a better chance of the response.27 The expression of PR is often used as a surrogate of activity of the ER pathway. Patients with tumors that express both ER and PR are more likely to respond to endocrine therapy.25 Of note, not all patients with ER+/PR+ tumor respond to the endocrine therapy. The mechanisms responsible for endocrine resistance, although well studied in animal models and cell lines, are poorly understood and are subject of extensive research.
Studies by Slamon et al28–30 in the late 1980s have documented the importance of ERBB2 (HER2) gene in breast cancer. The expression of HER2 is predominantly noted in high-grade tumors. HER2 is a poor prognostic factor but is a strong predictor of response to anti-HER2–directed therapies. A number of such therapies are available; these include inhibitors of the tyrosine kinase receptor as well as antibodies that recognize the extracellular domain of the receptor.31–33 New antibody-based strategies have tried to prevent dimerization of the receptor (pertuzumab) or use the antibody to carry a chemotherapeutic agent to the tumors cells (ado-trastuzumab emtasine; TDM1). In clinical practice, HER2 analysis is performed either by using IHC or fluorescence in situ hybridization (Fig. 4). In the vast majority of cases, both assays identify the same cases as being positive, that is, showing overexpression and/or amplification. However, marked heterogeneity in expression/amplification can generate false-positive and/or false-negative results. The ASCO-CAP guidelines recognize the heterogeneity and have recommended repeat testing in a number of different settings.7 More recently, Hanna et al2 have described different patterns of heterogeneity in breast cancer. The heterogeneity of HER2 expression makes it imperative to repeat testing when clinically indicated. For example, it has been documented that up to 10% of metastatic lesions can be HER2+ even when the primary tumor has been shown to be negative for HER2 expression/amplification.34,35
As is well illustrated by the HER2 IHC-fluorescence in situ hybridization concordance story, the expression of proteins, at least in some cases, is not related to DNA amplification events. This is in part due to the complexity of the molecular mechanisms leading to protein synthesis. Briefly, after generation of messenger RNA (mRNA) from the DNA template, the mRNA may be acted upon by a host of different molecules including noncoding RNA (microRNA and long noncoding RNA) and other proteins that influence the RNA processing such as the splicing mechanism. Depending on the actions of these molecules, the mRNA expression will lead to the synthesis of the full-length protein or a smaller protein. Alternatively, the mRNA could be targeted for destruction. Thus the cellular environment has the ability to modulate the size, shape, and function of the protein in addition to the quantity of the protein. This ability to alternatively splice mRNAs is retained in cancer cells, which often use it to their advantage. The classic examples of alternative splicing include alternative splicing of pyruvate kinase resulting in altered cellular metabolism (Warburg effect), which is the basis the positron emission tomography scan.36,37 Alternative splicing of BCLx mRNA can result in expression of a longer isoform that is antiapoptotic as opposed to the shorter form that is proapoptotic.38–40 In the context of breast cancer, alternative splicing of the PR has been shown to generate 2 isoforms, PR-A and PR-B. PR-B is a full-length isoform that is important for growth and differentiation, whereas PR-A is a truncated protein that interferes with the function of PR-B.41 The ratio of PR-B to PR-A is what determines the prognostic impact of PR in breast cancer. Of note, the currently available reagents do not discriminate between PR-A and PR-B. For more detailed information regarding alternative splicing the reader is directed to excellent reviews on this topic.42,43
It has been debated whether all cells within tumors have the capacity to regenerate and give rise to metastasis. It has been postulated that cancers contain a subpopulation of cells, referred to as cancer stem cells (CSCs), which are responsible for recurrence and metastasis. Al-Hajj et al,44 in a seminal paper, showed that cells that are CD44high/CD24low/– have a significantly greater ability to give rise to metastasis in animal models. Our group was among the first to raise concerns about the use of these markers for identifying CSCs.45 We analyzed a number of established cell lines using flow cytometry for the presence of these markers. Surprisingly, CD44+/CD24low/– cells were identified only in cell lines that had a mesenchymal phenotype. None of the luminal cell lines show expression of these markers. More recently, the expression of aldehyde dehydrogenase 1 (ALDH1) was suggested in the marker for CSCs.46 A number of studies have corroborated these findings, whereas some groups, including ours, remain yet to be convinced.47–51 The expression of ALDH1 is not restricted to the epithelial cells within the tumor. Our group did not find a prognostic relevance for ALDH1 expression within tumor cells but documented its role as a prognostic factor when expression within stromal cells was analyzed. Needless to say, the topic is controversial with excellent review articles highlighting the pros and the cons of the concept.52–55
Recently, pluripotent cells that can differentiate toward different lineages have been described even in normal breast tissue.56,57 Roy et al56 identified these cells using cell surface markers associated with repression of p16INK4a/cyclin-dependent kinase inhibitor 2A (CDKN2A). Further analysis of these cells showed that they expressed OCT3/4, SOX2, and NANOG at levels similar to those measured in human embryonic stem cells. Reexpression of these markers has been used to induce fibroblasts to pluripotent stem cells status.58 Our group identified these cells from explants of normal breast tissue obtained from volunteers donating tissue for the research at the Susan G. Komen normal tissue bank.57 The cells identified were from normal nontransformed epithelium, and detailed analysis showed differentiation along multiple lineages including but not limited to melanocytes, neural, chondrocytes, and osteocytes (that is, both ectodermal and mesodermal lineages). Data such as these show the degree of plasticity that exists within both normal and tumor cell populations.
Stochastic heterogeneity is a sophisticated way of stating that not all cells in a given population are doing the same thing at any given time. It is quite obvious to any pathologist that all the tumor cells do not exhibit mitotic activity at the same time. These differences in cellular states of cells that are otherwise genetically similar are in contrast to “deterministic heterogeneity,” wherein focal areas of differentiation (or dedifferentiation) are noted within tumors. These differences in cellular activity or differentiation states are evident at the molecular level but can also be manifested at the light microscopic level.
The last decade has seen the development of high-throughput technologies that have enabled assessment of thousands to millions of markers simultaneously. Initial studies were based on analysis of cDNA microarrays, followed later by massively parallel sequencing. These analyses have shown that no two tumors are alike. It is currently believed that cancers arise from a single cell, which may be a differentiated cell or (as discussed previously) a stem cell. Tumors undergo a Darwinian evolution process by which genetic heterogeneity is generated. It is equally possible that cancers arise from transformation of more than one cell, giving rise to additional dimensions of heterogeneity. On the basis of the “survival of the fittest” model, clones of tumor best suited for the environment dominate; these can be further influenced by therapy as well as changes in tissue environment.59–61 Analyses of matched primary and metastatic tumors have shown significant differences with clones barely detectable (or undetectable) in primary tumors predominating at the metastatic sites.62 This highlights the massive degree of heterogeneity that exists within tumors.
Expression patterns of mRNA have been used to classify tumors. A number of classifications have been proposed, of which the intrinsic classification is most commonly used. Perou et al63 divided breast cancers into ER+ subtypes (luminal A and luminal B) and ER− subtypes (basal-like, HER2-enriched, and normal-like carcinomas). The initial designations were based on microarray data but recently have been adapted to the reverse transcription polymerase chain reaction platform and even more recently to the NanoString platform (ProSigna). A number of issues related to this assay have popped up; these are principally related to the utility of the intrinsic classification in clinical practice and identification of the molecular clusters using routine pathology and IHC techniques.64 The assay has been shown to predict outcomes in patients treated with endocrine therapy and, in smaller studies, with chemotherapy.65–67
Lehmann et al68 genomically analyzed a large series of triple-negative breast cancers and identified at least 6 different subtypes. They postulated that, on the basis of the subtype, different oncocytic pathways had been activated and that treatment should be defined by subtypes. To support this claim, they subclassified cell lines on the basis of this schema and were able to show differential responses to drugs based on the subcategory of triple-negative tumors. Curtis et al69 performed expression and copy number analysis in a series of >2000 breast cancers. On the basis of this analysis, they identified 10 different subclasses (termed integrated clusters) of tumors. Of the 10 clusters, triple negative breast cancers/basal-like carcinomas are predominantly located in a single cluster. This raises the question regarding how many different clusters or groups of breast cancer exist. Experts in bioinformatics analysis state that hierarchical clustering is somewhat of a continuous process, and the number of groups or classes identified is a subjective decision.
Recent studies have focused on documentation of mutations in breast cancer. The Cancer Genome Atlas project has analyzed >1000 tumors using multiple “-omics” technologies.70 It is documented that apart from high-frequency mutations in TP53 and PIK3CA, most mutations are relatively infrequent (<1%). These somewhat pessimistic results have diminished the hopes of having mutation-based therapies in primary breast cancer. In addition, the data also confirmed the presence of multiple mutations in any given tumor. ER− tumors harbor a significantly higher number of mutations compared with ER+ tumors.70 It is difficult to determine which of these mutations are driving the oncogenic process (driver mutations) and which are incidental to the disturbed DNA replication processes (passenger mutations) present within cancers. Recent studies seem to suggest that the designation of driver and passenger mutations is somewhat arbitrary but, more importantly, is contextual in nature. Passenger mutations seem to be important for adapting to the stresses induced by hypoxia and other factors involved in the metastatic process. Deep sequencing analysis of primary tumors and metastatic lesions has revealed significant similarities and differences between these two. A number of mutations that are present in the metastatic tumors are undetectable in the primary lesions. The evolution of clones and their role in the development of recurrences/metastasis is a subject of active research. It is currently not known whether a metastatic lesion arises from the primary tumor or from other metastatic lesions. Targeted therapies have clearly documented evolution or “takeover” of the tumor by clones that are resistant to the therapeutic agents. Furthermore, these clones may be undetectable in the primary tumor.
There has been an intense focus on repair mechanisms associated with DNA damage. The roles of BRCA1 and BRCA2 genes in breast cancer are well documented. It appears that these tumors are particularly vulnerable to agents that target the DNA pathway. Therapeutic agents, such as inhibitors of the PARP (poly-ADP ribose polymerase) pathway, have been shown to be effective for these mutations.71,72 Clinical trials have failed to demonstrate efficacy in patients with triple-negative tumors. A number of other mutations have been recently associated with hereditary breast cancers. It is currently not known whether these patients might respond better with any specific therapies.
The mutations in TP53 and PIK3CA are by far the most common in breast cancer. Therapeutic targeting of p53 has proven to be difficult. A number of agents directed against PI3 kinase are in development. These agents seem to have significant activity in cell line and animal models, but efficacy in humans remains to be proven. It was believed that the presence of this mutation would sensitize the cancers to mTOR inhibitors such as Everolimus. However, the BOLERO-2 clinical trial using this agent along with endocrine therapy showed benefit in all patients irrespective of the PI3 kinase mutations status.73
Given the plethora of low-frequency mutations identified in The Cancer Genome Atlas, some investigators have taken an alternative approach that focuses on the type of mutation (eg, C>T) rather than the gene in which the mutation occurs.74 On the basis of the type of mutations, at least 21 different mutational signatures have been identified. The mechanisms leading to these mutations are far from clear. However, the APOBEC enzymes (1 and 3) have been casually implicated in multiple cancers including breast cancer.75,76 Therapeutic studies focused on these enzymes are being actively developed.
Implications of Tumor Heterogeneity
Pathologists have recognized tumor heterogeneity for a prolonged period of time, using it to classify tumors into different subtypes and to grade tumors. It should therefore be no surprise to them that heterogeneity exists at the molecular level. A number of different approaches have been used to deal with the heterogeneity. The common approach has been “bean counting.” It is well known that expression of biomarkers in tumors is heterogenous, and this heterogeneity is documented in the form of percentage of cells expressing that particular marker. The most common examples of these include analysis of expression of ER and PR. The data are reported as a continuous variable describing both the intensity and percentage of staining. Another approach has been to count the number of cells and then dichotomize the data into positive or negative categories. This approach has also been used for reporting ER and PR expression data. Until recently, a 10% cutoff was used for ER. The ASCO-CAP guidelines for hormone receptor expression have recommended that expression >1% should be regarded as positive. A variation of the same theme is noted in the reporting of HER2. The ASCO-CAP guidelines recognize broad categories namely positive, negative, and equivocal. The current guidelines categorize tumors exhibiting strong membranous staining of HER in >10% of the cells as positive. A further variation on the same theme is noted in molecular classifications of breast cancer. Here a clustering algorithm is used to collate cases into broad categories such as luminal A and luminal B. Lastly, the most common approach has been to ignore the heterogeneity. This approach has been widely used. Of note, until recently, all breast cancers were more or less treated in a similar manner. Similarly, all patients with ER+ or HER2+ tumors are treated similarly despite the different degrees of positivity.
In conclusion, tumor heterogeneity exists at all levels—clinical, histopathologic, and molecular (Fig. 5). It is undoubtedly important in the treatment of cancer. However, we currently lack the understanding and the tools to effectively combat the heterogeneity. The current tools essentially consist of recognizing the heterogeneity and documenting it in details so that one can develop effective strategies in the future.
1. Gerlinger M, Rowan AJ, Horswell S, et al.. Intratumor heterogeneity and branched evolution revealed by multiregion sequencing. N Engl J Med. 2012;366:883–892.
2. Hanna WM, Ruschoff J, Bilous M, et al.. HER2 in situ hybridization in breast cancer
: clinical implications of polysomy 17 and genetic heterogeneity. Mod Pathol. 2014;27:4–18.
3. Shipitsin M, Campbell LL, Argani P, et al.. Molecular definition of breast tumor heterogeneity
4. Metzger-Filho O, Tutt A, de Azambuja E, et al.. Dissecting the heterogeneity of triple-negative breast cancer
. J Clin Oncol. 2012;30:1879–1887.
5. Lakhani SR. International Agency for Research on Cancer
, World Health Organization. WHO Classification of Tumours of the Breast
, 4th ed. Lyon: International Agency for Research on Cancer
6. Rakha EA, Reis-Filho JS, Baehner F, et al.. Breast cancer
prognostic classification in the molecular era: the role of histological grade. Breast Cancer
7. Wolff AC, Hammond ME, Hicks DG, et al.. Recommendations for human epidermal growth factor receptor 2 testing in breast cancer
: American Society of Clinical Oncology/College of American Pathologists clinical practice guideline update. J Clin Oncol. 2013;31:3997–4013.
8. Beck AH, Espinosa I, Gilks CB, et al.. The fibromatosis signature defines a robust stromal response in breast
carcinoma. Lab Invest. 2008;88:591–601.
9. West RB, Nuyten DS, Subramanian S, et al.. Determination of stromal signatures in breast
carcinoma. PLoS Biol. 2005;3:e187.
10. Finak G, Bertos N, Pepin F, et al.. Stromal gene expression predicts clinical outcome in breast cancer
. Nat Med. 2008;14:518–527.
11. Alistar A, Chou JW, Nagalla S, et al.. Dual roles for immune metagenes in breast cancer
prognosis and therapy prediction. Genome Med. 2014;6:80.
12. Denkert C, Loibl S, Noske A, et al.. Tumor-associated lymphocytes as an independent predictor of response to neoadjuvant chemotherapy in breast cancer
. J Clin Oncol. 2010;28:105–113.
13. Loi S, Sirtaine N, Piette F, et al.. Prognostic and predictive value of tumor-infiltrating lymphocytes in a phase III randomized adjuvant breast cancer
trial in node-positive breast cancer
comparing the addition of docetaxel to doxorubicin with doxorubicin-based chemotherapy: BIG 02-98. J Clin Oncol. 2013;31:860–867.
14. Loi S, Michiels S, Salgado R, et al.. Tumor infiltrating lymphocytes are prognostic in triple negative breast cancer
and predictive for trastuzumab benefit in early breast cancer
: results from the FinHER trial. Ann Oncol. 2014;25:1544–1550.
15. Issa-Nummer Y, Darb-Esfahani S, Loibl S, et al.. Prospective validation of immunological infiltrate for prediction of response to neoadjuvant chemotherapy in HER2-negative breast cancer
--a substudy of the neoadjuvant GeparQuinto trial. PLoS One. 2013;8:e79775.
16. Adams S, Gray RJ, Demaria S, et al.. Prognostic value of tumor-infiltrating lymphocytes in triple-negative breast
cancers from two phase III randomized adjuvant breast cancer
trials: ECOG 2197 and ECOG 1199. J Clin Oncol. 2014;32:2959–2966.
17. Denkert C, von Minckwitz G, Brase JC, et al.. Tumor-infiltrating lymphocytes and response to neoadjuvant chemotherapy with or without carboplatin in human epidermal growth factor receptor 2-positive and triple-negative primary breast
cancers. J Clin Oncol. 2015;33:983–991.
18. Postow MA, Callahan MK, Wolchok JD. Immune checkpoint blockade in cancer
therapy. J Clin Oncol. 2015. [Epub ahead of print].
19. Pardoll DM. The blockade of immune checkpoints in cancer
immunotherapy. Nat Rev Cancer
20. Folkman J. Fighting cancer
by attacking its blood supply. Sci Am. 1996;275:150–154.
21. Montero AJ, Escobar M, Lopes G, et al.. Bevacizumab in the treatment of metastatic breast cancer
: friend or foe? Curr Oncol Rep. 2012;14:1–11.
22. Hayes DF, Miller K, Sledge G. Angiogenesis as targeted breast cancer
. 2007;16(suppl 2):S17–S19.
23. Tong RT, Boucher Y, Kozin SV, et al.. Vascular normalization by vascular endothelial growth factor receptor 2 blockade induces a pressure gradient across the vasculature and improves drug penetration in tumors. Cancer
24. Badve SS, Baehner FL, Gray RP, et al.. 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;26:2473–2481.
25. Early Breast Cancer
Trialists' Collaborative Group, Davies C, Godwin J, et al.. Relevance of breast cancer
hormone receptors and other factors to the efficacy of adjuvant tamoxifen: patient-level meta-analysis of randomised trials. Lancet. 2011;378:771–784.
26. Hammond ME, Hayes DF, Dowsett M, et al.. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for immunohistochemical testing of estrogen and progesterone receptors in breast cancer
(unabridged version). Arch Pathol Lab Med. 2010;134:e48–e72.
27. Allred DC, Harvey JM, Berardo M, et al.. Prognostic and predictive factors in breast cancer
by immunohistochemical analysis. Mod Pathol. 1998;11:155–168.
28. Slamon DJ, Clark GM. Amplification of c-erbB-2 and aggressive human breast
tumors? Science. 1988;240:1795–1798.
29. Slamon DJ, Clark GM, Wong SG, et al.. Human breast cancer
: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science. 1987;235:177–182.
30. Slamon DJ, Godolphin W, Jones LA, et al.. Studies of the HER-2/neu proto-oncogene in human breast
and ovarian cancer
. Science. 1989;244:707–712.
31. Blackwell KL, Burstein HJ, Storniolo AM, et al.. Overall survival benefit with lapatinib in combination with trastuzumab for patients with human epidermal growth factor receptor 2-positive metastatic breast cancer
: final results from the EGF104900 Study. J Clin Oncol. 2012;30:2585–2592.
32. Swain SM, Baselga J, Kim SB, et al.. Pertuzumab, trastuzumab, and docetaxel in HER2-positive metastatic breast cancer
. N Engl J Med. 2015;372:724–734.
33. Verma S, Miles D, Gianni L, et al.. Trastuzumab emtansine for HER2-positive advanced breast cancer
. N Engl J Med. 2012;367:1783–1791.
34. Liedtke C, Broglio K, Moulder S, et al.. Prognostic impact of discordance between triple-receptor measurements in primary and recurrent breast cancer
. Ann Oncol. 2009;20:1953–1958.
35. Niikura N, Liu J, Hayashi N, et al.. Loss of human epidermal growth factor receptor 2 (HER2) expression in metastatic sites of HER2-overexpressing primary breast
tumors. J Clin Oncol. 2012;30:593–599.
36. David CJ, Chen M, Assanah M, et al.. HnRNP proteins controlled by c-Myc deregulate pyruvate kinase mRNA splicing in cancer
. Nature. 2010;463:364–368.
37. Zhan C, Yan L, Wang L, et al.. Isoform switch of pyruvate kinase m1 indeed occurs but not to pyruvate kinase m2 in human tumorigenesis. PLoS One. 2015;10:e0118663.
38. Bielli P, Bordi M, Di Biasio V, et al.. Regulation of BCL-X splicing reveals a role for the polypyrimidine tract binding protein (PTBP1/hnRNP I) in alternative 5’ splice site selection. Nucleic Acids Res. 2014;42:12070–12081.
39. Wang Y, Chen D, Qian H, et al.. The splicing factor RBM4 controls apoptosis, proliferation, and migration to suppress tumor progression. Cancer
40. Moore MJ, Wang Q, Kennedy CJ, et al.. An alternative splicing network links cell-cycle control to apoptosis. Cell. 2010;142:625–636.
41. Bamberger AM, Milde-Langosch K, Schulte HM, et al.. Progesterone receptor isoforms, PR-B and PR-A, in breast cancer
: correlations with clinicopathologic tumor parameters and expression of AP-1 factors. Horm Res. 2000;54:32–37.
42. Chen J, Weiss WA. Alternative splicing in cancer
: implications for biology and therapy. Oncogene. 2015;34:1–14.
43. Oltean S, Bates DO. Hallmarks of alternative splicing in cancer
. Oncogene. 2014;33:5311–5318.
44. Al-Hajj M, Wicha MS, Benito-Hernandez A, et al.. Prospective identification of tumorigenic breast cancer
cells. Proc Natl Acad Sci USA. 2003;100:3983–3988.
45. Sheridan C, Kishimoto H, Fuchs RK, et al.. CD44+/CD24- breast cancer
cells exhibit enhanced invasive properties: an early step necessary for metastasis. Breast Cancer
46. Ginestier C, Hur MH, Charafe-Jauffret E, et al.. ALDH1 is a marker of normal and malignant human mammary stem cells and a predictor of poor clinical outcome. Cell Stem Cell. 2007;1:555–567.
47. Neumeister V, Rimm D. Is ALDH1 a good method for definition of breast cancer
stem cells? Breast Cancer
Res Treat. 2010;123:109–111.
48. Tan EY, Thike AA, et al.Breast
Surgical Team at Outram. ALDH1 expression is enriched in breast
cancers arising in young women but does not predict outcome. Br J Cancer
49. Zhong Y, Lin Y, Shen S, et al.. Expression of ALDH1 in breast
invasive ductal carcinoma: an independent predictor of early tumor relapse. Cancer
Cell Int. 2013;13:60.
50. Zhong Y, Shen S, Zhou Y, et al.. ALDH1 is a better clinical indicator for relapse of invasive ductal breast cancer
than the CD44+/CD24- phenotype. Med Oncol. 2014;31:864.
51. Resetkova E, Reis-Filho JS, Jain RK, et al.. Prognostic impact of ALDH1 in breast cancer
: a story of stem cells and tumor microenvironment. Breast Cancer
Res Treat. 2010;123:97–108.
52. Azizi E, Wicha MS. Point: cancer
stem cells—the evidence accumulates. Clin Chem. 2013;59:205–207.
53. Badve S, Nakshatri H. Breast
stem cells-beyond semantics. Lancet Oncol. 2012;13:e43–e48.
54. Gokmen-Polar Y, Badve SS. Counterpoint: cancer
stem cells: nonbelievers’ viewpoint. Clin Chem. 2013;59:208–210.
55. Gokmen-Polar Y, Nakshatri H, Badve S. Biomarkers for breast cancer
stem cells: the challenges ahead. Biomark Med. 2011;5:661–671.
56. Roy S, Gascard P, Dumont N, et al.. Rare somatic cells from human breast
tissue exhibit extensive lineage plasticity. Proc Natl Acad Sci USA. 2013;110:4598–4603.
57. Sauder CA, Koziel JE, Choi M, et al.. Phenotypic plasticity in normal breast
derived epithelial cells. BMC Cell Biol. 2014;15:20.
58. Takahashi K, Tanabe K, Ohnuki M, et al.. Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell. 2007;131:861–872.
59. Voss MH, Hakimi AA, Pham CG, et al.. Tumor genetic analyses of patients with metastatic renal cell carcinoma and extended benefit from mTOR inhibitor therapy. Clin Cancer
60. Misale S, Yaeger R, Hobor S, et al.. Emergence of KRAS mutations and acquired resistance to anti-EGFR therapy in colorectal cancer
. Nature. 2012;486:532–536.
61. Diaz LA Jr, Williams RT, Wu J, et al.. The molecular evolution of acquired resistance to targeted EGFR blockade in colorectal cancers. Nature. 2012;486:537–540.
62. Shah SP, Morin RD, Khattra J, et al.. Mutational evolution in a lobular breast
tumour profiled at single nucleotide resolution. Nature. 2009;461:809–813.
63. Perou CM, Sorlie T, Eisen MB, et al.. Molecular portraits of human breast
tumours. Nature. 2000;406:747–752.
64. Badve S, Dabbs DJ, Schnitt SJ, et al.. Basal-like and triple-negative breast
cancers: a critical review with an emphasis on the implications for pathologists and oncologists. Mod Pathol. 2011;24:157–167.
65. Gnant M, Filipits M, Greil R, et al.. Predicting distant recurrence in receptor-positive breast cancer
patients with limited clinicopathological risk: using the PAM50 Risk of Recurrence score in 1478 postmenopausal patients of the ABCSG-8 trial treated with adjuvant endocrine therapy alone. Ann Oncol. 2014;25:339–345.
66. Liu S, Chapman JA, Burnell MJ, et al.. Prognostic and predictive investigation of PAM50 intrinsic subtypes in the NCIC CTG MA.21 phase III chemotherapy trial. Breast Cancer
Res Treat. 2015;149:439–448.
67. Sestak I, Cuzick J, Dowsett M, et al.. Prediction of late distant recurrence after 5 years of endocrine treatment: a combined analysis of patients from the Austrian Breast
and Colorectal Cancer
Study Group 8 and arimidex, tamoxifen alone or in combination randomized trials using the PAM50 Risk Of Recurrence score. J Clin Oncol. 2015;33:916–922.
68. Lehmann BD, Bauer JA, Chen X, et al.. Identification of human triple-negative breast cancer
subtypes and preclinical models for selection of targeted therapies. J Clin Invest. 2011;121:2750–2767.
69. Curtis C, Shah SP, Chin SF, et al.. The genomic and transcriptomic architecture of 2,000 breast
tumours reveals novel subgroups. Nature. 2012;486:346–352.
Genome Atlas N. Comprehensive molecular portraits of human breast
tumours. Nature. 2012;490:61–70.
71. Turner NC, Lord CJ, Iorns E, et al.. A synthetic lethal siRNA screen identifying genes mediating sensitivity to a PARP inhibitor. EMBO J. 2008;27:1368–1377.
72. Fong PC, Boss DS, Yap TA, et al.. Inhibition of poly(ADP-ribose) polymerase in tumors from BRCA mutation carriers. N Engl J Med. 2009;361:123–134.
73. Beck JT, Hortobagyi GN, Campone M, et al.. Everolimus plus exemestane as first-line therapy in HR(+), HER2(-) advanced breast cancer
in BOLERO-2. Breast Cancer
Res Treat. 2014;143:459–467.
74. Alexandrov LB, Nik-Zainal S, Wedge DC, et al.. Signatures of mutational processes in human cancer
. Nature. 2013;500:415–421.
75. Cescon DW, Haibe-Kains B, Mak TW. APOBEC3B expression in breast cancer
reflects cellular proliferation, while a deletion polymorphism is associated with immune activation. Proc Natl Acad Sci USA. 2015;112:2841–2846.
76. Nik-Zainal S, Wedge DC, Alexandrov LB, et al.. Association of a germline copy number polymorphism of APOBEC3A and APOBEC3B with burden of putative APOBEC-dependent mutations in breast cancer
. Nat Genet. 2014;46:487–491.