Cell Cycle Glycolytic cancer associated fibroblasts ...

This article was downloaded by: [116.112.66.102] On: 02 September 2015, At: 05:51 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: 5 Howick Place, London, SW1P 1WG

Cell Cycle Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/kccy20

Glycolytic cancer associated fibroblasts promote breast cancer tumor growth, without a measurable increase in angiogenesis: Evidence for stromal-epithelial metabolic coupling Gemma Migneco, Diana Whitaker-Menezes, Barbara Chiavarina, Remedios Castello-Cros, Stephanos Pavlides, Richard G. Pestell, Alessandro Fatatis, Neal Flomenberg, Aristotelis Tsirigos, Anthony Howell, Ubaldo E. Martinez-Outschoorn, Federica Sotgia & Michael P. Lisanti Published online: 15 Jun 2010.

To cite this article: Gemma Migneco, Diana Whitaker-Menezes, Barbara Chiavarina, Remedios Castello-Cros, Stephanos Pavlides, Richard G. Pestell, Alessandro Fatatis, Neal Flomenberg, Aristotelis Tsirigos, Anthony Howell, Ubaldo E. MartinezOutschoorn, Federica Sotgia & Michael P. Lisanti (2010) Glycolytic cancer associated fibroblasts promote breast cancer tumor growth, without a measurable increase in angiogenesis: Evidence for stromal-epithelial metabolic coupling, Cell Cycle, 9:12, 2412-2422, DOI: 10.4161/cc.9.12.11989 To link to this article: http://dx.doi.org/10.4161/cc.9.12.11989

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Versions of published Taylor & Francis and Routledge Open articles and Taylor & Francis and Routledge Open Select articles posted to institutional or subject repositories or any other third-party website are without warranty from Taylor & Francis of any kind, either expressed or implied, including, but not limited to, warranties of merchantability, fitness for a particular purpose, or non-infringement. Any opinions and views expressed in this article are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor & Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions It is essential that you check the license status of any given Open and Open Select article to confirm conditions of access and use.

Report

Cell Cycle 9:12, 2412-2422; June 15, 2010; © 2010 Landes Bioscience

Glycolytic cancer associated fibroblasts promote breast cancer tumor growth, without a measurable increase in angiogenesis Evidence for stromal-epithelial metabolic coupling

Gemma Migneco,1,2 Diana Whitaker-Menezes,1,2 Barbara Chiavarina,1,2 Remedios Castello-Cros,1,2 Stephanos Pavlides,1,2 Richard G. Pestell,1,2 Alessandro Fatatis,3 Neal Flomenberg,5 Aristotelis Tsirigos,6 Anthony Howell,7 Ubaldo E. Martinez-Outschoorn,1,2,5 Federica Sotgia1,2,7,* and Michael P. Lisanti1,2,5,7,*

Downloaded by [116.112.66.102] at 05:51 02 September 2015

1 Departments of Stem Cell Biology & Regenerative Medicine, and Cancer Biology; 2The Jefferson Stem Cell Biology and Regenerative Medicine Center; 4Department of Pathology, Anatomy & Cell Biology; and 5Department of Medical Oncology; Kimmel Cancer Center; Thomas Jefferson University; Philadelphia, PA USA; 3Departments of Pharmacology & Physiology, and Pathology & Laboratory Medicine; Drexel University College of Medicine; Philadelphia, PA USA; 6Computational Genomics Group; IBM Thomas J. Watson Research Center; Yorktown Heights, NY USA; 7Manchester Breast Centre & Breakthrough Breast Cancer Research Unit; Paterson Institute for Cancer Research; School of Cancer, Enabling Sciences and Technology; Manchester Academic Health Science Centre; University of Manchester; Manchester, UK

Key words: caveolin-1, tumor stroma, glycolytic fibroblasts, cancer associated fibroblasts, aerobic glycolysis, hexokinase, enolase, pyruvate kinase, lactate dehydrogenase, the Warburg effect, H-Ras (G12V)

Previously, we proposed a new model for understanding the Warburg effect in tumorigenesis and metastasis. In this model, the stromal fibroblasts would undergo aerobic glycolysis (a.k.a., the Warburg effect)—producing and secreting increased pyruvate/lactate that could then be used by adjacent epithelial cancer cells as “fuel” for the mitochondrial TCA cycle, oxidative phosphorylation, and ATP production. To test this model more directly, here we used a matched set of metabolically well-characterized immortalized fibroblasts that differ in a single gene. CL3 fibroblasts show a shift towards oxidative metabolism, and have an increased mitochondrial mass. In contrast, CL4 fibroblasts show a shift towards aerobic glycolysis, and have a reduced mitochondrial mass. We validated these differences in CL3 and CL4 fibroblasts by performing an unbiased proteomics analysis, showing the functional upregulation of 4 glycolytic enzymes, namely ENO1, ALDOA, LDHA and TPI1, in CL4 fibroblasts. Many of the proteins that were upregulated in CL4 fibroblasts, as seen by unbiased proteomics, were also transcriptionally upregulated in the stroma of human breast cancers, especially in the patients that were prone to metastasis. Importantly, when CL4 fibroblasts were co-injected with human breast cancer cells (MDA-MB-231) in a xenograft model, tumor growth was dramatically enhanced. CL4 fibroblasts induced a >4-fold increase in tumor mass, and a near 8-fold increase in tumor volume, without any measurable increases in tumor angiogenesis. In parallel, CL3 and CL4 fibroblasts both failed to form tumors when they were injected alone, without epithelial cancer cells. Mechanistically, under co-culture conditions, CL4 glycolytic fibroblasts increased mitochondrial activity in adjacent breast cancer cells (relative to CL3 cells), consistent with the “Reverse Warburg Effect.” Notably, Western blot analysis of CL4 fibroblasts revealed a significant reduction in caveolin-1 (Cav-1) protein levels. In human breast cancer patients, a loss of stromal Cav-1 is associated with an increased risk of early tumor recurrence, metastasis, tamoxifen-resistance, and poor clinical outcome. Thus, loss of stromal Cav-1 may be an effective marker for predicting the “Reverse Warburg Effect” in the stroma of human breast cancer patients. As such, CL4 fibroblasts are a new attractive model for mimicking the “glycolytic phenotype” of cancer-associated fibroblasts. Nutrients derived from glycolytic cancer associated fibroblasts could provide an escape mechanism to confer drug-resistance during anti-angiogenic therapy, by effectively reducing the dependence of cancer cells on a vascular blood supply.

Introduction Recently, we have identified a loss of Cav-1 protein expression as a new stromal fibroblast biomarker for poor clinical outcome in human breast cancer patients.1-3 More specifically, a loss of Cav-1

in cancer-associated fibroblasts predicted early tumor recurrence, lymph node metastasis, and tamoxifen-resistance.2 A loss of stromal Cav-1 was especially valuable in breast cancer patients with lymph node metastasis, where a 11.5-fold stratification of 5-year survival was obtained [Cav-1 (+), 80% recurrence-free survival

*Correspondence to: Federica Sotgia and Michael P. Lisanti; Email: [email protected] and [email protected] Submitted: 03/31/10; Accepted: 04/06/10 Previously published online: www.landesbioscience.com/journals/cc/article/11989 2412

Cell Cycle

Volume 9 Issue 12

report


Report

vs. Cav-1 (-), 7% recurrence-free survival]. Thus, a loss of stromal Cav-1 is a strong predictor of recurrent and metastatic disease progression in breast cancer.3 Similar results were also obtained with a cohort of DCIS patients.4 In this cohort, 100% of DCIS patients lacking stromal Cav-1 underwent DCIS recurrence, and 80% of these patients progressed to invasive ductal carcinoma. As such, a loss of stromal Cav-1 may also be involved in tumor initiation, as well as progression and metastasis.4 Complementary results were also obtained with human prostate cancers, where a loss of stromal Cav-1 was associated with a higher Gleason score (poor prognosis) and prostate cancer progression and metastasis (to lymph nodes and bone).5 To better understand the molecular mechanism(s) underlying the prognostic value of stromal Cav-1, we turned to Cav-1 (-/-) deficient mice as a model system. Unbiased proteomic and transcriptional analysis of Cav-1 (-/-) deficient stromal cells revealed a shift towards aerobic glycolysis (aka, the Warburg effect).6 Then, we validated the expression of glycolytic enzymes in the tumor stroma of human breast cancer patients that lacked Cav-1 protein expression.6 Based on these findings, we proposed a new model to explain the Warburg effect in tumor metabolism. In this new model, termed “The Reverse Warburg Effect,” stromal fibroblasts would undergo aerobic glycolysis and provide adjacent cancer cells with energy-rich nutrients (such as pyruvate and lactate) in a paracrine fashion.6 In essence, these glycolytic cancer-associated fibroblasts would promote tumor growth, by feeding the TCA cycle/oxidative metabolism in epithelial tumor cells. This would represent an example of “Epithelial-Stromal Metabolic Coupling.”6 This model could explain the strong prognostic value of a loss of Cav-1 in the tumor stromal fibroblast compartment. To test this hypothesis further, we devised a new xenograft model system, where nude mice were co-injected with (1) Cav-1 (-/-) deficient fibroblasts, and (2) a human breast cancer cell line, namely MDA-MB-231 cells.7 Under these conditions, Cav-1 (-/-) deficient fibroblasts, which show a proteomic shift towards aerobic glycolysis, significantly promoted both tumor growth (∼2.5-fold) and angiogenesis (∼3-fold). Interestingly, the growthpromoting effects of Cav-1 (-/-) deficient fibroblasts could be prevented by treatment with a combination of well-established glycolysis inhibitors [2-deoxy-glucose (2-DG) and dichloroacetate (DCA)], resulting in a 4.5-fold reduction in tumor mass.7 These results support the idea that chemically-induced metabolic restriction with glycolysis inhibitors is a valid approach to target the Cav-1-deficient population of cancer-associated fibroblasts, and provides additional evidence for the existence of the “Reverse Warburg Effect.”7 Here, we extend these observation to a pair of genetically well defined, but metabolically diverse, human fibroblast cell lines. These two matched fibroblast cell lines were previously immortalized with telomerase, and were engineered to express a series of transforming oncogenes.8-11 Remarkably, one of these fibroblast lines, CL3, shows a shift towards oxidative metabolism, while the other, CL4, shows a shift towards aerobic glycolysis.12 Thus, these two cell lines differ in only one gene mutation (H-Ras (G12V)) and are ideal to further test the validity of the “Reverse

www.landesbioscience.com

Warburg Hypothesis.” Notably, we show here that CL4 fibroblasts have a loss of Cav-1 protein expression and an increase in vimentin expression, consistent with a glycolytic myofibroblastic cell phenotype. We validated the glycolytic phenotype of CL4 fibroblasts by performing an unbiased proteomics analysis, supplemented by Western blot analysis. Most importantly, we show that CL4 glycolytic fibroblasts dramatically promote breast cancer tumor growth, with a four-fold increase in tumor mass, and a near eight-fold increase in tumor volume, as would be predicted by the “Reverse Warburg Hypothesis.” Interestingly, this growth-promoting effect appeared to be independent of tumor angiogenesis. Furthermore, under co-culture conditions, CL4 glycolytic fibroblasts increased mitochondrial activity in adjacent breast cancer cells (relative to CL3 cells), consistent with the “Reverse Warburg Effect.” Thus, CL3/CL4 fibroblasts are a new genetically defined stromal fibroblast model system for studying the tumorpromoting effects of glycolytic cancer associated fibroblasts. Results Origins of CL3 and CL4 cells. BJ5a fibroblasts were originally derived from primary cultures of human foreskin and were first immortalized with telomerase (hTERT), to generate a stable immortalized fibroblast cell line without any oncogenic mutations.8-11 Then, Hahn, Weinberg and colleagues proceeded to serially transform hTERT-BJ5a cells with a variety of well-known oncogenes, such as the SV40 large T (LT) and small T (ST) antigens, as well as activated Ras [H-Ras (G12V)].8-11 This resulted in the generation of 4 matched cell lines, termed CL1-to-CL4. As a result, CL3 cells harbor hTERT, LT and ST, while CL4 cells contain hTERT, LT, ST and H-Ras (G12V) (summarized in Fig. 1). Interestingly, Schreiber and colleagues noted strong differences in metabolism between the CL3 and CL4 fibroblast cell lines, and proceeded to characterize these differences in detail.12 More specifically, they showed that CL4 fibroblasts show a shift towards aerobic glycolysis (aka, the Warburg effect), with increased glucose uptake and increased production of lactate [at least six-fold relative to CL1 (hTERT alone)].12 Moreover, CL4 fibroblasts are 1,000-fold more sensitive to 2-deoxy-glucose, an established glycolysis inhibitor. In contrast, CL3 fibroblasts appear to show a shift towards oxidative phosphorylation (“mitochondrial metabolism”), based on metabolic profiling, inhibitor studies, and transcriptional increases in genes that are important for mitochondrial biogenesis.12 Thus, CL3 and CL4 cells are a matched set of immortalized fibroblasts that differ in a single gene [H-Ras (G12V)], but show major shifts in metabolism, with CL4 fibroblasts showing a “glycolytic phenotype.”12 As a consequence, we restricted our studies to a comparison of the CL3 and CL4 fibroblast cell lines. Unbiased proteomics with validation supports the glycolytic phenotype of CL4 fibroblasts. In order to validate the glycolytic phenotype of CL4 fibroblasts, they were first subjected to an unbiased proteomics analysis. Protein lysates (derived from CL3 and CL4 cells) were labeled and separated by 2D

Cell Cycle

2413


DIGE (two-dimensional difference gel electrophoresis). Twodimensional separation of CL3 and CL4 cell lysates yielded at least 75 protein spots which were differentially expressed. We used mass spec/protein micro-sequencing to determine the molecular identity of 26 protein spots that were specifically upregulated in CL4 fibroblasts. See Figure 2 for representative 2D gel analysis. Interestingly, six of the protein spots were identified as glycolytic enzymes or related proteins (ALDOA, PBEF, ENO1, TPI1 and LDHA) (summarized in Table 1). Moreover, six other proteins spots were identified as protein chaperones, or proteins associated with oxidative or cell stress (CFL1, SOD2, NACA, TBCA, HSP90 and TPT1). Notably, proteins associated with metastasis (NME1 and CLIC1) and translational control (EEF2 and EEF1B2) were also upregulated in CL4 fibroblasts. To further validate the glycolytic phenotype of CL4 fibroblasts, they were subjected to Western blot analysis with a panel of 8 antibodies directed against glycolytic enzymes. Figure 3A and B shows that numerous glycolytic enzymes were indeed upregulated at the protein level, including HK1/2, ENO1/2, PKM2 and LDHA/B/C, consistent with our unbiased proteomics analysis. Since we have previously shown that Cav-1 (-/-) deficient stromal fibroblasts6 also show a proteomic shift towards aerobic glycolysis,6 accompanied by the onset of a myofibroblastic phenotype,6 we next examined the status of Cav-1 and vimentin by Western blot analysis. Figure 4 shows that CL4 fibroblasts undergo a dramatic reduction in Cav-1 protein expression and have increased vimentin expression. Thus, CL4 fibroblasts may pheno-copy the effects of a Cav-1 deficiency, as they show (i) a loss of Cav-1 expression, (ii) an increase in vimentin (a myofibroblast marker), and (iii) an increase in numerous glycolytic enzymes under normoxic conditions (aerobic glycolysis, aka, the “Warburg effect”). CL4 glycolytic fibroblasts promote the growth of aggressive breast cancer tumors. We have previously postulated that a shift towards aerobic glycolysis in cancer-associated fibroblasts can promote tumor growth in a paracrine fashion, by transfering energy-rich metabolites (such as pyruvate and lactate) from “glycolytic” fibroblasts to “oxidative” cancer cells.6 We have termed this novel idea “The Reverse Warburg Effect.”6 Thus, CL3 and CL4 cells represent an ideal model to further test the “The Reverse Warburg Hypothesis,” as they are a pair of genetically-defined matched fibroblast cell lines that differ in a single gene mutation, but show dramatic changes in cell metabolism. In this regard, the glycolytic phenotype of CL4 fibroblasts would be predicted to promote tumor growth, in comparison with CL3 fibroblasts. To test our hypothesis in the context of a xenograft model, we next co-injected CL3 or CL4 fibroblasts with a highly aggressive human breast cancer cell line (MDA-MB-231) in nude mice. MDA-MB-231 cells have been previously classified as basal-like triple-negative breast cancer cells, based on their epithelial marker status. After 2 weeks, tumors were harvested and subjected to further analysis. Figure 5 shows that co-injection with CL4 fibroblasts increased both the mass (>4-fold) and volume (nearly 8-fold)

2414

Figure 1. Schematic description of CL3 and CL4 fibroblasts. BJ5a fibroblasts were originally derived from primary cultures of human foreskin and were first immortalized with telomerase (hTERT), to generate a stable immortalized fibroblast cell line without any other oncogenic mutations. Then, Hahn, Weinberg and colleagues proceeded to transform hTERT-BJ5a cells with a series of well-known oncogenes, such as the SV40 large T (LT) and small T (ST) antigens, as well as activated Ras [H-Ras (G12)]. This resulted in the generation of 4 matched cell lines, termed CL1-to-CL4. As a result, CL3 cells harbor hTERT, LT and ST, while CL4 cells contain hTERT, LT, ST and H-Ras (G12V). Interestingly, CL4 cells show a shift towards aerobic glycolysis (a.k.a., the Warburg effect). Moreover, CL4 cells are 1,000-fold more sensitive to 2-deoxy-glucose, an established glycolysis inhibitor. In contrast, CL3 cells appear to show a shift towards oxidative phosphorylation (“mitochondrial metabolism”). hTERT (human telomerase); LT (SV40 Large T antigen); ST (SV40 Small T antigen).

of MDA-MB-231 tumor xenografts, relative to CL3 fibroblasts. Importantly, under these conditions, CL3 and CL4 fibroblasts both failed to form any tumors when injected alone in nude mice (Table 2). Thus, CL4 fibroblasts are pro-tumorgenic in the context of MDA-MB-231 cells, but did not form any tumors on their own in this short time frame. CL4 glycolytic fibroblasts do not promote increased tumor angiogenesis, as compared with CL3 fibroblasts. One possible explanation for the dramatic tumor-promoting effects of CL4 fibroblasts is that they somehow stimulate tumor angiogenesis. To test this hypothesis directly, we used CD31 immuno-staining to quantify the extent of tumor angiogenesis in the MDA-MB-231 xenografts that were grown in the presence of either CL3 or CL4 fibroblasts. First, since the MDA-MB-231 breast cancer cells are GFPtagged, we verified that the bulk of the tumor mass was indeed composed of breast cancer cells (Fig. 6). Interestingly, further quantitative analysis of CD31-stained frozen sections derived from these breast cancer tumors revealed no differences in vessel density (# of vessels per mm2) (Fig. 7). Thus, an increase in tumor angiogenesis cannot explain the tumor-promoting properties of CL4 fibroblasts. Other mechanism(s), such as “The Reverse Warburg Effect,” may be contributing energy-rich nutrients that can directly stimulate tumor cell growth, without the need for increased angiogenesis. CL4 glycolytic fibroblasts increase mitochondrial activity in adjacent breast cancer cells, consistent with the “Reverse Warburg Effect.” To better understand the tumor-promoting

Cell Cycle

Volume 9 Issue 12


Figure 2. Proteomic analysis of CL3 and CL4 fibroblast cell lysates. Protein extracts from CL3 (green) and CL4 (red) fibroblasts were first labeled with Cy2 and Cy3, respectively. Then, equal amounts of these extracts were mixed and subjected to 2D gel electrophoresis [IEF (pH 3–10) and SDS-PAGE]. After separation, image analysis was used to identify spots that showed upregulation or downregulation. The image shows an overlay of the CL3 and CL4 cells in Image Quant TL (GE Healthcare). Approximately 75 spots showed significant changes (at least ∼1.5-fold). The top 26 spots showing the most significant changes were subjected to spot-picking and protein identification by mass spectrometry analysis (MALDI/TOF).

activity of CL4 fibroblasts, we first assessed their mitochondrial status, relative to CL3 fibroblasts. Interestingly, by Western blot analysis using a panel of anti-mitochondrial antibody probes, we observed reductions in key enzymes/subunits associated with the mitochondrial TCA cycle, such as pyruvate dehydrogenase (PDH), and oxidative phosphorylation in CL4 fibroblasts (Fig. 8). Similar results were obtained by immuno-fluorescence analysis of CL4 fibroblasts, showing reductions in both mitochondrial mass (using a mitochondrial marker protein) and mitochondrial activity (using MitoTracker) (Fig. 9). Thus, these data further validate the observation that CL4 fibroblasts show a shift towards aerobic glycolysis, and are consistent with the results of Schreiber and colleagues, who also observed a reduction in MitoTracker activity in CL4 fibroblasts.12 Next, to determine the effects of CL4 fibroblasts on MDA-MB-231 cells, we prepared co-cultures. For comparison purposes, MDA-MB-231 cells were also co-cultured with


CL3 fibroblasts in parallel, under identical conditions. We then assessed the mitochondrial activity of these co-cultures using MitoTracker. Figure 10 shows that in CL3 co-cultures, there is a dramatic reduction in mitochondrial activity, especially in MDA-MB-231 cells (marked by GFP). In contrast, in the CL4 co-cultures, there is a significant increase in mitochondrial activity, predominantly in MDA-MB-231 cells. Thus, it appears that CL4 fibroblasts may stimulate tumor growth by promoting mitochondrial oxidative metabolism in adjacent breast cancer cells. As such, these observations provide direct morphological and functional support for the existence of the “Reverse Warburg Effect.” Intersection of the proteomic profiles of CL4/CL3 fibroblasts with transcriptional profiling data from the stromal compartment of human breast cancers: associations with lymph-node metastasis. In order to determine if our results may have relevance to human breast cancers, we compared the list of proteins upregulated by proteomic analysis in CL4 fibroblasts

Cell Cycle

2415

Table 1. Proteomic analysis of hTERT CL4/CL3 fibroblasts Fold change (CL4/CL3)

Accession number

Protein spot number

fructose 1,6-bisphosphate aldolase A (ALDOA)

2.7

gi|4557976

45

nicotinamide phosphoribosyltransferase (PBEF/NAMPT/visfatin)

2.5

gi|5031977

32

enolase, alpha (ENO1)

2.4

gi|203282367

37

enolase, alpha (ENO1)

2.3

gi|62896593

36

triosephosphate isomerase 1 (TPI1)

2.3

gi|4507645

57

lactate dehydrogenase A (LDHA)

2.0

gi|13786847

50

3.0

gi|5031635

66

Glycolytic Enzymes and Related Proteins

Chaperones and Proteins Associated with Cell Stress


cofilin 1 (non-muscle) (CFL1) manganese superoxide dismutatase (SOD2)

2.8

gi|194708958

62

nascent-polypeptide-associated complex alpha (NACA)

2.5

gi|62897731

46

tubulin-specific chaperone A (TBCA)

2.1

gi|4759212

70

heat shock 90 kDa protein 1, alpha (HSP90AA1)

2.0

gi|62914009

9

tumor protein, translationally-controlled 1 (TPT1)

2.0

gi|55662177

58

non-metastatic cells 1 (NME1)

2.3

gi|38045913

65

nuclear chloride channel (CLIC1)

2.2

gi|4588526

55

nuclear chloride channel (CLIC1)

2.0

gi|4588526

54

eukaryotic translation elongation factor 2 (EEF2)

4.0

gi|4503483

6

eukaryotic translation elongation factor 2 (EEF2)

3.4

gi|4503483

5

eukaryotic translation elongation factor 1 beta2 (EEF1B2)

2.3

gi|4503477

53

3.4

gi|119620394

75

nucleolin (NCL)

3.4

gi|55956788

8

nucleolin (NCL)

2.6

gi|189306

7

Metastasis Associated Proteins

Translation Associated Proteins

Other hypothetical protein FLJ13305 (FAM161A)

fatty acid binding protein 5 (psoriasis-associated) (FABP5)

2.8

gi|4557581

72

ATP-binding cassette, sub-family E, member 1 (ABCE1)

2.5

gi|108773782

18

Rho GDP dissociation inhibitor (GDI) alpha (ARHGDIA)

2.3

gi|4757768

56

leucine aminopeptidase 3 (LAP3)

2.0

gi|37588925

31

coactosin-like 1 (COTL1)

2.0

gi|21624607

71

Figure 3. Validating the expression of glycolytic enzymes in CL4 fibroblasts. Cell lysates were prepared from CL3 and CL4 fibroblasts, and subjected to SDS-PAGE and Western blot analysis with antibodies directed against glycolytic enzymes. Note that numerous glycolytic enzymes were indeed upregulated at the protein level, including HK1/2, ENO1/2 (A), PKM2 and LDHA/B/C (B), consistent with our proteomics analysis. Immunoblotting for b-tubulin is also shown as a control for equal protein loading. HK, hexokinase; ENO, enolase; PK, pyruvate kinase; LDH, lactate dehydrogenase.

with the transcriptional profiles of breast cancer tumor stroma that were isolated by laser-capture microdissection.13 For this purpose, comparisons were made with three different tumor stromal gene data sets.14

2416

The “tumor stroma” list consists of genes that are upregulated in tumor stroma, relative to normal breast stroma.14 The “recurrence-prone” list consists of tumor stromal genes that are upregulated in breast cancer patients with recurrence, relative to patients that did not undergo recurrence.14 The “metastasisprone” list consists of tumor stromal genes that are upregulated in breast cancer patients with lymph-node metastasis, relative to patients that did not show metastasis.14

Cell Cycle

Volume 9 Issue 12

Table 2. Tumor incidence: CL3 and CL4 fibroblasts alone do not form tumors


Figure 4. CL4 glycolytic fibroblasts show reduced Cav-1 levels and increased vimentin. Cell lysates were prepared from CL3 and CL4 fibroblasts, and subjected to SDS-PAGE and Western blot analysis with antibodies directed against Cav-1 and vimentin. Note that Cav-1 protein levels were downregulated, while vimentin levels were increased, consistent with a shift towards a more “myofibroblastic” phenotype. Immunoblotting for beta-tubulin is also shown as a control for equal protein loading.

Cell line(s) injected

Tumor incidence (10 injections)

CL3 Fibroblasts Alone

0/10; 0%

CL4 Fibroblasts Alone

0/10; 0%

MDA-MB-231 + CL3

10/10; 100%

MDA-MB-231 + CL4

10/10; 100%

Table 3 shows that 10 of the 26 proteins that we identified as upregulated by proteomics in CL4 fibroblasts were also transcriptionally upregulated in the tumor stroma of human breast cancer patients. Importantly, 7 of the 10 proteins that were upregulated in CL4 fibroblasts were associated with the “metastasis-prone” stromal gene list. Thus, our findings may also have importance for understanding how “The Reverse Warburg Effect” may promote cancer cell metastasis. Discussion

Figure 5. CL4 glycolytic fibroblasts dramatically promote tumor growth. hTERT-immortalized fibroblasts (CL3 vs. CL4) were co-injected with a human breast cancer cell line (MDA-MB-231 cells) in the flanks of nude mice. After 2 weeks, the tumors that formed were harvested and subjected to a detailed analysis. Note that co-injection with CL4 fibroblasts promoted the growth of MDA-MB-213 derived tumors by 4-to-8-fold. However, CL3 and CL4 fibroblasts injected alone did not form tumors in this time frame. (A) tumor weight; (B) tumor volume.


Here, we show that a bonafide glycolytic fibroblast cell line is sufficient to promote tumor growth. CL3 cells are an oxidative fibroblast cell line, while CL4 fibroblasts are a glycolytic cell line; however, they differ in the expression of only a single mutated oncogene, namely H-Ras (G12V). Co-injection of CL4 fibroblasts with MDA-MB-231 cells is sufficient to increase tumor mass (>4-fold) and tumor volume (nearly 8-fold), relative to CL3 fibroblasts, without a measurable increase in angiogenesis. These observations directly support the “Reverse Warburg Hypothesis” or “Stromal-Epithelial Metabolic Coupling,” which states that stromal fibroblasts can undergo myofibroblastic differentiation and produce/secrete energy-rich metabolites (such as pyruvate and lactate via aerobic glycolysis) that can then be used by adjacent epithelial cancer cells to fuel their TCA cycle (oxidative phosphorylation), resulting in increased ATP production in cancer cells.6,7 Thus, CL3/CL4 fibroblasts are a new model system to study “The Reverse Warburg Effect” in tumor metabolism. Similar results were recently obtained independently by another laboratory, however, they dared not challenge the validity of the “conventional” Warburg Effect.15 Instead, they attributed their results to increased CXCL12-CXCR4 signaling.15 Of course, ultimately, it may be a combination of paracrine growth factor signaling and glycolytic stromal metabolism that drives tumor growth, and it may be difficult to completely separate or uncouple these two phenomena, which occur concurrently during tumor initiation and progression. More specifically, Taichman, Wang and colleagues noted that a glycolytic enzyme (phosphoglycerate kinase-1; PGK1) was increased in the tumor stromal compartment (cancer-associated fibroblasts; CAFs) of human prostate cancers.15 Their results were obtained by laser-capture micro-dissection and cDNA microarray analysis of CAFs in situ. They went on to show that recombinant expression of PGK1 in normal fibroblasts was sufficient to induce the myofibroblast phenotype, with co-expression of smooth

Cell Cycle

2417


Figure 6. GFP-tagged tumor epithelial cells constitute the bulk of the MDA-MB-231 tumor masses. MDA-MB-231 (GFP+) tumor xenografts were used to prepare frozen sections (6 µm thickness). To visualize the tumor vasculature, these sections were then immuno-labeled with rat anti-mouse CD31 IgG (1:100 dilution; BD Biosciences) (red fluorescence) followed by AlexaFluor 594 donkey anti-rat antibody (1:1,000 dilution; Invitrogen Molecular Probes, Carlsbad, CA). The DAPI nuclear counterstain (blue fluorescence) and GFP signal present in MDA-MB-231 tumor cells (green fluorescence) are also shown. Bar = 50 microns. Note that virtually identical results were obtained with MDA-MB-231 cells that were co-injected with either CL3 or CL4 fibroblasts.

Figure 7. Quantitation of tumor angiogensis in CL3- and CL4-containing breast cancer tumors. CD31-positive vessels were enumerated in 4–6 fields in the central area of each MDA-MB-231 tumor using a 20x objective lens and an ocular grid (0.25 mm2 per field). However, no significant differences in tumor angiogenesis were observed for MDA-MB-231 tumors grown in the presence of either CL3or CL4-fibroblasts. n.s., not significant.

2418

Cell Cycle

muscle actin, vimentin and CXCL12.15 Finally, co-injection of PGK1-expressing normal fibroblasts promoted prostate cancer tumor growth. Thus, expression of this single glycolytic enzyme (PGK1) was sufficient to induce myofibroblast differentiation, and promote tumorigenesis.15 However, they failed to recognize the potential metabolic implications of their findings. Instead they stated that: “One possible role of PGK1 in stroma may have the ability to increase metabolism and cell cycling, thus eliciting a CAF phenotype. This suggests that a “Warburg effect” may also be associated with tumor stromal cells; however, further studies are needed.”15 Similarly, we have previously shown that Cav-1 (-/-) deficient stromal fibroblasts spontaneously undergo myofibroblastic differentiation,16 express numerous myofibroblast markers (smooth muscle actin, vimentin, calponin, collagen I, SPARC), and upregulate eight glycolytic enzymes (including phosphoglycerate kinase-1; Pgk1; 2.4-fold), as well as overexpress chemokines (Cxcl12, Cxcl9, Cxcl1 and Ccl5) and growth factors (Vegf, Pdgf, Hgf) (up to 8-fold higher for Cxcl9).6 Interestingly, Cav-1 (-/-) deficient fibroblasts also promote the growth of MDA-MB-231 breast cancer tumors in a xenograft model.7 Thus, the onset of aerobic glycolysis and the secretion of pro-inflammatory cytokines/chemokines may be simply two related and general features of the myofibroblast differentiation program that synergistically promote tumor growth.6,7 In summary, here we show that Ras-activation in stromal fibroblasts is associated with a metabolic/proteomic shift towards aerobic glycolysis and that these fibroblasts have the ability to dramatically promote tumor growth, independently of tumor angiogenesis. This is consistent with recent studies showing the Ras-transformed fibroblasts have defects in mitochondrial metabolism, leading to (1) inhibition of mitochondrial complex I, and (2) a block in oxidative phosphorylation.17-19

Volume 9 Issue 12


Figure 9. CL4 fibroblasts show both reductions in mitochondrial mass and mitochondrial activity. The status of mitochondria in CL3 and CL4 fibroblasts was also evaluated morphologically. Left: Mitochondria were visualized with a monoclonal antibody directed against the surface of intact mitochondria (Red). Right: Mitochondrial activity was visualized with the fluorescent probe, MitoTracker (red). Note the reductions in both mitochondrial mass (left) and mitochondrial activity (right) in CL4 fibroblasts. DAPI was used to counter-stain nuclei (blue). Importantly, images were acquired using identical exposure settings. Original magnification, 63x.

Figure 8. The expression of key mitochondrial enzymes/subunits is reduced in CL4 fibroblasts. Lysates were prepared from CL3 and CL4 fibroblasts and subjected to Western blot analysis with antibodies directed against (A) pyruvate dehydrogenase (PDH) components, (B) subunits involved in oxidative phosphorylation, and (C) proteins involved in maintaining mitochondrial integrity. In part A, two exposures are shown for the PDH subunit E1-alpha, so that the E3-binding protein could also be visualized. IM, inner membrane; OM, outer membrane; IMS, intermembrane space.


Figure 10. CL4 fibroblasts increase mitochondrial activity in adjacent breast cancer cells. Mitochondrial activity was monitored in MDA-MB231-[GFP(+)] cells that were co-cultured with either CL3 or CL4 fibroblasts. Functional mitochondria were visualized by MitoTracker staining (red). MDA-MB-231-[GFP(+)] cells were visualized by auto-fluorescence (green). DAPI was used to stain nuclei (blue). Note that mitochondrial activity is greatly increased in MDA-MB-231 cells co-cultured with CL4 fibroblasts, as compared to MDA-MB-231 cells co-cultured with CL3 fibroblasts. Importantly, images were acquired using identical exposure settings. Original magnification, 63x.

Cell Cycle

2419


Materials and Methods

Table 3. Intersection of CL4/CL3 proteomics with transcriptional profiling of human breast cancer tumor stroma

Tumor Materials. Antibodies for Western blotGene symbol Recurrence-prone stroma Metastasis-prone stroma stroma ting were obtained from the following ABCE1 commercial sources: anti-caveolin-1 (mAb 2297; BD Biosciences, cat#610407); antiALDOA 1.69E-03 enolase-1 (ENO1; Cell Signaling Tech, ARHGDIA 9.77E-03 cat#3810); anti-enolase-2 (ENO2; Cell CFL1 1.49E-02 Signaling Tech, cat #9536); anti-hexoCLIC1 3.61E-02 kinase-1 (HK1; Cell Signaling Tech, COTL1 7.61E-26 5.31E-03 cat#2804), anti-hexokinase-2 (HK2; EEF2 4.41E-02 Cell Signaling, cat#2106), anti-LDHA EEF1B2 3.59E-02 (Cell Signaling, cat#2012); anti-LDHB (Sigma, cat#AV48210); anti-LDHC ENO1 (Sigma, cat#SAB1100051); anti-PKM2 FABP5 (Proteintech, cat#15822-1-AP/SR); FAM161A anti-beta-tubulin (Sigma, cat#T4026); GORASP2 and vimentin (R28; Cell Signaling, HSP90AA1 1.63E-02 cat#3932). MDA-MB-231 breast cancer LAP3 cells were obtained from ATCC, and were LDHA engineered to express GFP. 3.67E-10 Ldhal6b Animal studies. All animals were housed and maintained in a pathogenNACA free environment/barrier facility at the NCL Kimmel Cancer Center at Thomas NME1 Jefferson University under National PBEF/NAMPT/visfatin Institutes of Health (NIH) guidelines. SOD2 1.45E-11 Mice were kept on a 12-hour light/dark TBCA cycle with ad libitum access to chow and TPI1 water. Approval for all animal protocols used for this study was reviewed and TPT1 approved by the Institutional Animal Care Proteins that were transcriptionally upregulated in laser-capture micro-dissected human breast and Use Committee (IACUC). Briefly, a cancer tumor stroma are shown in bold. Although LDHA was not found to be transcriptionally mixture of tumor cells {MDA-MB-231 upregulated, its close relative LDHAL6B was transcriptionally increased in tumor stroma. [GFP(+)]; 1 x 106 cells} plus CL3 or CL4 fibroblasts (3 x 105 cells) in 100 µl of sterImmunofluorescence microscopy. Frozen tissue sections ile PBS were co-injected into the flanks of athymic NCr nude mice (NCRNU; Taconic Farms; 6–8 weeks of age). Mice were (6 µm) were fixed in 2% paraformaldehyde in PBS for 10 min then sacrificed at 2 weeks post-injection; tumors were excised to and washed with PBS. The sections were blocked with 1% BSA determine their weight and size using calipers. Tumor volume in PBS and incubated with rat anti-mouse CD31 antibody at a was calculated using the formula (X 2Y)/2, where X is the width dilution of 1:100 for 1 hr at room temperature, followed by incubation with the secondary antibody donkey anti-rat Alexa Fluor and Y is the length. Quantitation of tumor angiogenesis. Immunohistochemical 594 (1:1,000, Invitrogen), for 30 minutes at room temperature. staining for CD31 was performed on frozen tumor sections using Finally, sections were washed with PBS and nuclei were stained a 3-step biotin-streptavidin-horseradish peroxidase method. with the DNA-intercalating dye 4,6-diamidino-2-phenylindole Frozen tissue sections (6 µm) were fixed in 2% paraformaldehyde (DAPI) (Invitrogen) and mounted with Prolong Gold anti-fade in PBS for 10 min and washed with PBS. After blocking with reagent (Invitrogen). GFP(+) tumor cells and their surrounding 10% rabbit serum the sections were incubated overnight at 4°C vasculature were visualized using this approach. Proteomic analysis. 2D DIGE (two-dimensional differwith rat anti-mouse CD31 antibody (BS Biosciences) at a dilution of 1:200, followed by biotinylated rabbit anti-rat IgG (1:200) ence gel electrophoresis)20 and mass spectrometry protein idenantibody and streptavidin-HRP. Immunoreactivity was revealed tification were run by Applied Biomics (Hayward, CA). Image with 3,3' diaminobenzidine. For quantitation of vessels, CD31- scans were carried out immediately following the SDS-PAGE positive vessels were enumerated in 4–6 fields within the central using Typhoon TRIO (Amersham BioSciences) following the area of each tumor using a 20x objective lens and an ocular grid protocols provided. The scanned images were then analyzed (0.25 mm2 per field). The total numbers of vessel per unit area was by Image QuantTL software (GE-Healthcare), and then subcalculated using Image J and the data was represented graphically. jected to in-gel analysis and cross-gel analysis using DeCyder

2420

Cell Cycle

Volume 9 Issue 12


software version 6.5 (GE-Healthcare). The ratio of protein differential expression was obtained from in-gel DeCyder software analysis. The selected spots were picked by an Ettan Spot Picker (GE-Healthcare) following the DeCyder software analysis and spot picking design. The selected protein spots were subjected to in-gel trypsin digestion, peptides extraction, desalting and followed by MALDI-TOF/TOF (Applied Biosystems) analysis to determine the protein identity. Immunoblot analysis. Cell lysates were prepared from CL3 and CL4 fibroblasts to validate our results from proteomics analysis by western blot analysis. Briefly, after rinsing with PBS, cells were collected by cell scraping with lysis buffer (10 mM Tris, pH 7.5, 150 mM NaCl, 1% Triton X-100, and 60 mM n-octyl-glucoside), containing protease inhibitors (Boehringer Mannheim, Indianapolis, IN). Samples were incubated on a rotating platform at 4°C and were then centrifuged at 12,000 xg for 10 minutes (at 4°C) to remove insoluble debris. Protein concentrations were analyzed using the BCA reagent (Pierce, Rockford, IL). Samples were then separated by SDS-PAGE (10% acrylamide) and transferred to nitrocellulose. All subsequent wash buffers contained 10 mM Tris, pH 8.0, 150 mM NaCl, 0.05% Tween 20, which was supplemented with 5% nonfat dry milk (Carnation) for the blocking solution, and 1% bovine serum albumin for the antibody diluent. Horseradish peroxidase-conjugated secondary antibodies were used to visualize bound primary antibodies with the Supersignal chemiluminescence substrate (Pierce, Rockford, IL). Epithelial-Fibroblast co-cultures. CL3 or CL4 fibroblasts and MDA-MB-231-[GFP(+)] cells were plated onto glass coverslips in 12-well plates in 1-ml of complete media. The total number of cells per well was 1 x 105. Experiments were performed at a 5:1 fibroblast-to-MDA-MB-231 cell-ratio. As controls, monocultures of fibroblasts and MDA-MB-231 cells were seeded in parallel. The day after, the media was changed to DMEM containing 10% NuSerum (a low protein alternative to FBS; BD Biosciences) and supplemented with antibiotics (Pen-Strep). Evaluating mitochondrial status. To functionally stain mitochondria, the rosamine-based MitoTracker Orange CMTMRos (Invitrogen; cat#M7510) was employed. Lyophilized MitoTracker was dissolved in DMSO to prepare a 1 mM stock solution, immediately before its use. Then, the stock solution was diluted in serum-free DMEM to a final concentration of 25 nM. Cells were incubated with the pre-warmed MitoTracker staining solution References 1. Mercier I, Casimiro MC, Wang C, Rosenberg AL, Quong J, Allen KG, et al. Human breast cancer-associated fibroblasts (CAFs) show caveolin-1 downregulation and RB tumor suppressor functional inactivation: Implications for the response to hormonal therapy. Cancer Biol Ther 2008; 7:1212-25. 2. Witkiewicz AK, Dasgupta A, Sotgia F, Mercier I, Pestell RG, Sabel M, et al. An absence of stromal Caveolin-1 expression predicts early tumor recurrence and poor clinical outcome in human breast cancers. Am J Pathol 2009; 174:2023-34. 3. Witkiewicz AK, Casimiro MC, Dasgupta A, Mercier I, Wang C, Bonuccelli G, et al. Towards a new “stromalbased” classification system for human breast cancer prognosis and therapy. Cell Cycle 2009; 8:1654-8.


for 10 minutes at 37°C. Then, cells were washed in PBS and fixed with 2% PFA. MitoTracker stains live mitochondria with an intact membrane potential and is preserved upon fixation. It provides a qualitative measurement of mitochondrial membrane potential. An antibody directed against the intact surface of mitochondria (Millipore; cat#MAB1273; clone 113-1) was also used to visualize mitochondria by immuno-cytochemistry. Western blot analysis was also performed using cocktails of antimitochondrial antibodies for detecting (1) subunits of complexes involved in oxidative phosphorylation (cat#MS601), (2) PDH subunits (cat#MSP02) and (3) mitochondrial membrane integrity (cat#MS620), all obtained commercially from Mitosciences, Inc. Statistical analysis. Statistical significance was examined using the Student t-test. Values of p < 0.05 were considered significant. Values were expressed as means +/- SEM. Acknowledgements

M.P.L. and his laboratory were supported by grants from the NIH/NCI (R01-CA-080250; R01-CA-098779; R01-CA-120876; R01-AR-055660), and the Susan G. Komen Breast Cancer Foundation. F.S. was supported by grants from the W.W. Smith Charitable Trust, the Breast Cancer Alliance (BCA), and a Research Scholar Grant from the American Cancer Society (ACS). R.G.P. was supported by grants from the NIH/ NCI (R01-CA-70896, R01-CA-75503, R01-CA-86072 and R01-CA-107382) and the Dr. Ralph and Marian C. Falk Medical Research Trust. The Kimmel Cancer Center was supported by the NIH/NCI Cancer Center Core grant P30-CA-56036 (to R.G.P.). Funds were also contributed by the Margaret Q. Landenberger Research Foundation (to M.P.L.). This project is funded, in part, under a grant with the Pennsylvania Department of Health (to M.P.L.). The Department specifically disclaims responsibility for any analyses, interpretations or conclusions. This work was also supported, in part, by a Centre grant in Manchester from Breakthrough Breast Cancer in the U.K. (to A.H.) and an Advanced ERC Grant from the European Research Council. We are greatly indebted to Dr. William C. Hahn (Harvard Medical School, Dana-Farber Cancer Institute, Broad Institute of Harvard/MIT) for generously providing CL3 and CL4 fibroblasts for the experiments described herein, and for critical review of the manuscript.

4. Witkiewicz AK, Dasgupta A, Nguyen KH, Liu C, Kovatich AJ, Schwartz GF, et al. Stromal caveolin-1 levels predict early DCIS progression to invasive breast cancer. Cancer Biol Ther 2009; 8:1167-75. 5. Di Vizio D, Morello M, Sotgia F, Pestell RG, Freeman MR, Lisanti MP. An absence of stromal Caveolin-1 is associated with advanced prostate cancer, metastatic disease and epithelial Akt activation. Cell Cycle 2009; 8:2420-4. 6. Pavlides S, Whitaker-Menezes D, Castello-Cros R, Flomenberg N, Witkiewicz AK, Frank PG, et al. The reverse Warburg effect: aerobic glycolysis in cancer associated fibroblasts and the tumor stroma. Cell Cycle 2009; 8:3984-4001.

Cell Cycle

7. Bonuccelli G, Whitaker-Menezes D, Castello-Cros R, Pavlides S, Pestell RG, Fatatis A, et al. The Reverse Warburg Effect: Glycolysis inhibitors prevent the tumor promoting effects of Caveolin-1 deficient cancer associated fibroblasts. Cell Cycle 2010; 9:1960-71. 8. Hahn WC, Counter CM, Lundberg AS, Beijersbergen RL, Brooks MW, Weinberg RA. Creation of human tumour cells with defined genetic elements. Nature 1999; 400:464-8. 9. Hahn WC, Weinberg RA. Modelling the molecular circuitry of cancer. Nat Rev Cancer 2002; 2:331-41. 10. Hahn WC, Weinberg RA. Rules for making human tumor cells. N Engl J Med 2002; 347:1593-603. 11. Dolma S, Lessnick SL, Hahn WC, Stockwell BR. Identification of genotype-selective antitumor agents using synthetic lethal chemical screening in engineered human tumor cells. Cancer Cell 2003; 3:285-96.

2421


12. Ramanathan A, Wang C, Schreiber SL. Perturbational profiling of a cell-line model of tumorigenesis by using metabolic measurements. Proc Natl Acad Sci USA 2005; 102:5992-7. 13. Finak G, Bertos N, Pepin F, Sadekova S, Souleimanova M, Zhao H, et al. Stromal gene expression predicts clinical outcome in breast cancer. Nat Med 2008; 14:518-27. 14. Pavlides S, Tsirigos A, Vera I, Flomenberg N, Frank PG, Casimiro MC, et al. Transcriptional evidence for the “Reverse Warburg Effect” in human breast cancer tumor stroma and metastasis: Similarities with oxidative stress, inflammation, Alzheimer’s disease, and “neuron-glia metabolic coupling.” Aging 2010; 2:185-99. 15. Wang J, Ying G, Jung Y, Lu J, Zhu J, Pienta KJ, Taichman RS. Characterization of phosphoglycerate kinase-1 expression of stromal cells derived from tumor microenvironment in prostate cancer progression. Cancer Res 2010; 70:471-80. 16. Sotgia F, Del Galdo F, Casimiro MC, Bonuccelli G, Mercier I, Whitaker-Menezes D, et al. Caveolin-1-/null mammary stromal fibroblasts share characteristics with human breast cancer-associated fibroblasts. Am J Pathol 2009; 174:746-61.

2422

17. Baracca A, Chiaradonna F, Sgarbi G, Solaini G, Alberghina L, Lenaz G. Mitochondrial Complex I decrease is responsible for bioenergetic dysfunction in K-ras transformed cells. Biochim Biophys Acta 2010; 1797:314-23. 18. de Groof AJ, te Lindert MM, van Dommelen MM, Wu M, Willemse M, Smift AL, et al. Increased OXPHOS activity precedes rise in glycolytic rate in H-RasV12/ E1A transformed fibroblasts that develop a Warburg phenotype. Mol Cancer 2009; 8:54. 19. Chiaradonna F, Gaglio D, Vanoni M, Alberghina L. Expression of transforming K-Ras oncogene affects mitochondrial function and morphology in mouse fibroblasts. Biochim Biophys Acta 2006; 1757:1338-56. 20. Marouga R, David S, Hawkins E. The development of the DIGE system: 2D fluorescence difference gel analysis technology. Anal Bioanal Chem 2005; 382:669-78.

Cell Cycle

Volume 9 Issue 12