TP53-inducible Glycolysis and Apoptosis Regulator (TIGAR) Metabolically Reprograms Carcinoma and Stromal Cells in Breast Cancer*

A subgroup of breast cancers has several metabolic compartments. The mechanisms by which metabolic compartmentalization develop in tumors are poorly characterized. TP53 inducible glycolysis and apoptosis regulator (TIGAR) is a bisphosphatase that reduces glycolysis and is highly expressed in carcinoma cells in the majority of human breast cancers. Hence we set out to determine the effects of TIGAR expression on breast carcinoma and fibroblast glycolytic phenotype and tumor growth. The overexpression of this bisphosphatase in carcinoma cells induces expression of enzymes and transporters involved in the catabolism of lactate and glutamine. Carcinoma cells overexpressing TIGAR have higher oxygen consumption rates and ATP levels when exposed to glutamine, lactate, or the combination of glutamine and lactate. Coculture of TIGAR overexpressing carcinoma cells and fibroblasts compared with control cocultures induce more pronounced glycolytic differences between carcinoma and fibroblast cells. Carcinoma cells overexpressing TIGAR have reduced glucose uptake and lactate production. Conversely, fibroblasts in coculture with TIGAR overexpressing carcinoma cells induce HIF (hypoxia-inducible factor) activation with increased glucose uptake, increased 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase-3 (PFKFB3), and lactate dehydrogenase-A expression. We also studied the effect of this enzyme on tumor growth. TIGAR overexpression in carcinoma cells increases tumor growth in vivo with increased proliferation rates. However, a catalytically inactive variant of TIGAR did not induce tumor growth. Therefore, TIGAR expression in breast carcinoma cells promotes metabolic compartmentalization and tumor growth with a mitochondrial metabolic phenotype with lactate and glutamine catabolism. Targeting TIGAR warrants consideration as a potential therapy for breast cancer.


A subgroup of breast cancers has several metabolic compartments. The mechanisms by which metabolic compartmentalization develop in tumors are poorly characterized. TP53 inducible glycolysis and apoptosis regulator (TIGAR) is a bisphosphatase that reduces glycolysis and is highly expressed in carcinoma cells in the majority of human breast cancers. Hence we set out to determine the effects of TIGAR expression on breast carcinoma and fibroblast glycolytic phenotype and tumor growth. The overexpression of this bisphosphatase in carcinoma cells induces expression of enzymes and transporters involved in the catabolism of lactate and glutamine. Carcinoma cells overex-
pressing TIGAR have higher oxygen consumption rates and ATP levels when exposed to glutamine, lactate, or the combination of glutamine and lactate. Coculture of TIGAR overexpressing carcinoma cells and fibroblasts compared with control cocultures induce more pronounced glycolytic differences between carcinoma and fibroblast cells. Carcinoma cells overexpressing TIGAR have reduced glucose uptake and lactate production. Conversely, fibroblasts in coculture with TIGAR overexpressing carcinoma cells induce HIF (hypoxia-inducible factor) activation with increased glucose uptake, increased 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase-3 (PFKFB3), and lactate dehydrogenase-A expression. We also studied the effect of this enzyme on tumor growth. TIGAR overexpression in carcinoma cells increases tumor growth in vivo with increased proliferation rates. However, a catalytically inactive variant of TIGAR did not induce tumor growth. Therefore, TIGAR expression in breast carcinoma cells promotes metabolic compartmentalization and tumor growth with a mitochondrial metabolic phenotype with lactate and glutamine catabolism. Targeting TIGAR warrants consideration as a potential therapy for breast cancer.
Multiple metabolic compartments exist in human tumors including breast cancer (1). There is metabolic coupling with lactate transfer between highly glycolytic carcinoma cells and carcinoma cells with reduced glycolysis in models of vulvar and colon cancer (2). High stromal glycolysis with low cancer cell glycolysis have also been described in breast, ovarian, prostate, bladder, head and neck carcinomas, and sarcomas (3).
The interactions between carcinoma cells and fibroblasts in breast cancer play an important role in tumor progression. Fibroblasts have been shown to promote breast cancer tumor growth (4) and metastasis (5). Carcinoma cell invasiveness and resistance to chemotherapy in breast cancer are induced by fibroblasts (6). It is unknown how cancer cell metabolism in breast cancer modulates adjacent fibroblast metabolism.
The synthesis and breakdown of Fru-2,6-P 2 depends on the bifunctional 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase isoenzymes, products of four genes (PFKFB1-4) that code for the different PFKFB isoenzymes and that have distinct cell expression patterns and display different kinase/bisphosphatase activity ratios and control by different protein kinases (10 -12). In tumor cells, the concentration of Fru-2,6-P 2 is generally elevated due to overexpression and activation of PFKFB3, which has opposite effects of TIGAR (11)(12)(13). Conversely, TIGAR depletion increases glycolytic flux by increasing the activity of PFK-1 and the glycolytic flux by increasing Fru-2,6-P 2 levels (7,14). It is unknown if TIGAR modulates catabolism of other substrates such as lactate and glutamine, which have been shown to be alternate catabolites to glucose for carcinoma cells (1). TIGAR expression has been shown to be inversely associated with glycolysis as TIGAR expression was inversely associated with 2-deoxyglucose uptake on PET scans in subjects with non-small cell lung cancer (15). TIGAR regulates hexokinase 2 (HK2) activity and increases mitochondrial membrane potential, but the effect of TIGAR on mitochondrial metabolism, oxygen consumption rates, and ATP generation is unknown (16). In sum, TIGAR reduces glycolysis, but its effects on the catabolism of other substrates and mitochondrial metabolism is poorly characterized.
TIGAR has been reported to mediate human cancer aggressiveness, although the mechanism is unclear, and its effect in breast cancer is unknown. In addition to reducing glycolysis, TIGAR reduces apoptosis (7). TIGAR is overexpressed in nasopharyngeal carcinoma, and genetic overexpression increases tumor growth with carcinoma cell growth, colony formation, migration, and invasion with NFB activation in carcinoma cells (17). Knockdown of this bisphosphatase induces apoptosis of HepG2 hepatocellular carcinoma cells (18). TIGAR downregulation in HepG2 carcinoma cells reduces the size of hepatocellular carcinoma xenografts (19). TIGAR reduces apoptosis rates of non-small cell lung cancer H-1299 and osteosarcoma U2OS cells (7). TIGAR is also overexpressed in the majority of glioblastomas, protecting cells from starvation-induced cell death by up-regulating respiration and improving cellular redox homeostasis (20). Radiosensitization of glioma cells also occurs with TIGAR knockdown (14). However, another study has shown that TIGAR decreases cell viability in glioblastoma (21). TIGAR is required for proliferation of small intestine cells (22). TIGAR null mice have decreased tumorigenesis and increased survival in a model in which the tumor suppressor adenomatous polyposis coli (APC) is deleted in LGR5 ϩ intestinal stem cells (22). TIGAR is highly expressed in the majority of breast cancers and carcinoma cells in breast cancer frequently have low rates of glycolysis and apoptosis (7,8,24, and 25). However, no reports have described the effect of TIGAR on breast cancer tumor growth in vivo. Also, the role of TIGAR in the low rates of glycolysis and apoptosis with high proliferation observed in human breast cancer is unknown.
Breast tumors have high expression of mitochondrial biogenesis transcription factors and mitochondrial markers of oxidative phosphorylation (OXPHOS) (26) (27). Multiple metabolic compartments exist in human malignancies (28). Metabolic compartmentalization occurs in breast cancer with highly glycolytic stromal cells and cancer cells with high OXPHOS (28). There are no data on the metabolic effects of TIGAR expression in breast cancer and if it plays a role in metabolic compartmentalization or heterogeneity in tumors. We hypothesize that TIGAR expression in carcinoma cells induces aggressive disease with utilization of alternative catabolites to glucose and metabolic compartmentalization.

TIGAR Induced Markers of Lactate and Glutamine Catabolism with Increased NADPH in Carcinoma
Cells-To investigate the effects of TIGAR on markers of lactate and glutamine utilization, we overexpressed TIGAR in breast carcinoma cells. TIGAR overexpression up-regulated moderately monocarboxylate transporter 2 (MCT2) protein and mRNA expression (Fig. 1, A and B). MCT2 is a lactate importer. TIGAR also upregulates lactate dehydrogenase B (LDH-B), which is the enzyme that converts lactate into pyruvate for mitochondrial metabolism (Fig. 1C) and TIGAR up-regulates glutaminase 1 (GLS1) modestly, which is the rate-limiting enzyme in glutamine catabolism and converts glutamine to glutamate (Fig. 1D). Conversely, glutamate ammonia ligase (GLUL), which is the enzyme that mediates glutamine synthesis from glutamate and ammonia, was down-regulated with TIGAR overexpression (Fig. 1E). TIGAR markedly decreased the NADP ϩ /NADPH ratio (Fig. 1, F and G), which is consistent with increasing flux through the pentose phosphate pathway.
TIGAR Induced OXPHOS and ATP Generation in the Presence of Lactate and Glutamine-To investigate the effects of TIGAR on mitochondrial biogenesis, we studied the expression of peroxisome proliferator-activated receptor ␥ coactivator 1 (PGC1) and nuclear respiratory factor 1 (NRF1). TIGAR overexpression induced these markers of mitochondrial biogenesis ( Fig. 2A).
To investigate the effects of TIGAR and exposure to glutamine and lactate on OXPHOS, we studied oxygen consumption rates (OCR) in control and TIGAR-overexpressing carcinoma cells. Oxygen consumption rates were higher for TIGAR overexpressing carcinoma cells exposed to glutamine, lactate, or glutamine and lactate compared with carcinoma cells expressing the control vector (p Ͻ 0.05) (Fig. 2B). TIGAR overexpressing carcinoma cells exposed to glutamine have a 1.2-fold greater OCR than control carcinoma cells exposed to glutamine (p Ͻ 0.05) and 1.5-fold greater OCR when exposed to lactate or glutamine and lactate than the control carcinoma cells exposed to the same conditions (p Ͻ 0.05). OCRs were not increased by glutamine or lactate in the absence of TIGAR overexpression. TIGAR overexpressing carcinoma cells had lower OCR than control cells (0.8-fold) when cultured with glucose but without glutamine and lactate (p Ͻ 0.05).
Fibroblasts Induced Aggressive T47D Cells with Increased TIGAR Expression, Reduced Expression of Glycolytic Markers, Increased Proliferation, and Decreased Apoptosis-To investigate stromal-epithelial interactions, we used a coculture model system composed of 1) human fibroblasts immortalized with the telomerase catalytic domain (hTERT-BJ1 fibroblasts) and 2) T47D or MCF7 cells, which are well established breast cancer cell lines.
TIGAR and PFKFB3 are markers of glycolysis, although they have opposite effects. TIGAR mRNA expression was increased 1.5-fold in T47D cells in coculture (p Ͻ 0.01) (Fig. 3A) and 1.6-fold in MCF7 cells in coculture (p Ͻ 0.05), whereas PFKFB3 mRNA expression was reduced 1.9-fold in MCF7 cells in coculture (p Ͻ 0.05) (Fig. 3B). No statistically significant difference in PFKFB3 mRNA expression was noted in T47D cells in coculture (Fig. 3A).
Fibroblasts induced TIGAR expression in carcinoma cells (Fig. 3C), and TIGAR induced aggressive carcinoma cells with increased proliferation and reduced apoptosis. Carcinoma cells cocultured with fibroblasts had a 4.7-fold increase in DNA synthesis at the expense of G 0 -G 1 and G 2 -M (p Ͻ 0.01) (Fig. 3D). Carcinoma cells also had a 2-fold decreased apoptosis rate when in coculture with fibroblasts (p Ͻ 0.05) (Fig. 3E).
TIGAR in Carcinoma Cells Induced a Glycolytic Phenotype in Fibroblasts-Next we studied the effect of TIGAR overexpression in T47D cells on fibroblast metabolism. Fibroblasts cocultured with T47D cells overexpressing TIGAR had a 1.2-fold increase in 2-NBDG uptake compared with fibroblasts in control cocultures (p Ͻ 0.05) (Fig. 5A). Glucose uptake in fibro-blasts was measured by quantifying green fluorescence of the 2-deoxyglucose compound 2-NBDG. 2-Deoxyglucose and 2-NBDG uptake is an additional measure of glycolysis with lower uptake indicating reduced glycolysis. PFKFB3 and LDH-A expression increased in fibroblasts cocultured with TIGAR overexpressing carcinoma cells (Fig. 5, B and C). PFKFB3 and LDH-A are markers of glycolysis. LDH-B expression is reduced with coculture with TIGAR overexpressing carcinoma cells (Fig. 5C). LDH-B has opposite effects to LDH-A, converting lactate to pyruvate. We next assessed the effect of TIGAR on activation of hypoxia-inducible factor (HIF) in fibroblasts by using a HIF1A luciferase reporter as HIF1A is one of the main glycolytic transcription factors. NIH3T3 fibroblasts stably transfected with a HIF1A luciferase reporter were cultured with control T47D cells or TIGAR-overexpressing T47D cells. Activation of HIF1A is increased by 1.6-fold (p Ͻ 0.05) in fibroblasts cocultured with T47D cells overexpressing TIGAR cells in 0.5% O 2 hypoxia compared with control coculture conditions (Fig. 5D). Note that there was no significant change in HIF1A activation in normoxia between control coculture and FIGURE 2. Effect of TIGAR on mitochondria. T47D cells overexpressing TIGAR and control cells were cultured and lysed and subjected to immunoblot analysis for PGC1 and NRF1 (A). B, oxygen consumption was measured in control T47D cells (blue bar) and with TIGAR overexpression (red bar); relative concentration is shown. Cells were cultured with control media, control with 2 mM glutamine, control with 10 mM lactate, or control with 2 mM glutamine and 10 mM lactate. C, T47D and MCF7 cells overexpressing TIGAR and control cells were cultured and lysed and subjected to immunoblot analysis for MITONEET and TOMM20. D, MCF7 cells overexpressing HA-TOMM20 were cultured and lysed and subjected to immunoblot analysis for TOMM20, TIGAR, and actin. E, ATP levels measured by quinacrine fluorescence intensity in TIGAR overexpressing and control MCF7 cells. Cell culture conditions as described above in B. F, ATP levels in TIGAR overexpressing and control MCF7 cells with or without 10 M 3PO. n.s., not significant. coculture with T47D cells overexpressing TIGAR. Fibroblasts exposed to carcinoma cells have reduced TIGAR expression, but overexpression of TIGAR in fibroblasts led to 1.2-fold reduced glucose uptake compared with control fibroblasts (Fig.  5E). In sum, TIGAR overexpression in carcinoma cells induced a glycolytic phenotype in fibroblasts.
TIGAR Induced Tamoxifen Resistance but Higher Sensitivity to Mitochondrial Modulators-T47D cells were cultured alone or with fibroblasts and exposed to tamoxifen. TIGAR overexpression in carcinoma cells exposed to tamoxifen induced apoptosis resistance with a 1.2-fold reduction in homotypic culture and a 1.4-fold reduction in coculture with fibroblasts (p Ͻ 0.05) (Fig. 6A). In contrast, TIGAR-overexpressing carcinoma cells were more sensitive to apoptosis by mitochondrial modulators. Metformin, which is an OXPHOS complex I inhibitor, doxycycline, which inhibits mitochondrial translation, and ABT-199, which inhibits BCL2 as a single agent or in combination induced higher rates of apoptosis in TIGAR-overexpressing T47D cells than controls (Fig. 6, B-E). Exposure to lactate in MCF7 cells led to increased TIGAR and BCL2 expression with reduced MCT4 expression (Fig. 6F). Exposure to tamoxifen in MCF7 cells decreased expression of TIGAR, BCL-XL, BCL2, MITONEET, and labile subunits of OXPHOS (Fig. 6, G-H).
MDA-MB-231 cells overexpressing TIGAR with three mutations (H11A/E102A/H198A) that render it catalytically inactive as a Fru-2,6-P 2 bisphosphatase (TIGAR triple mutant) were generated. The triple mutant TIGAR as expected did not reduce Fru-2,6-P 2 levels and did not increase tumor growth compared with control carcinoma cells (Fig. 8B). Hence, a catalytically active TIGAR is required to promote tumor growth.

Discussion
The current study delineates the role of TIGAR in OXPHOS and glycolytic metabolic reprogramming in breast cancer. This bisphosphatase enzyme, which reduces Fru-2,6-P 2 levels, inhibits glycolysis and increases pentose phosphate pathway flux with increased NADP ϩ to NADPH production. We have discovered that TIGAR promotes the growth of breast-invasive ductal carcinoma in vivo and with utilization of lactate and glutamine as substrates, mitochondrial OXPHOS metabolism, and ATP generation in cancer cells (Figs. 2 and 8C). Increased ATP levels in TIGAR-overexpressing cells upon exposure to 3PO suggests that TIGAR provides metabolic flexibility with changes in Fru-2,6-P 2 levels. We also demonstrate that TIGAR expression in carcinoma cells induces reciprocal metabolic changes in fibroblasts (Figs. 5 and 8D).
Our finding that TIGAR promotes growth of breast-invasive ductal carcinoma is relevant to human disease because high TIGAR expression is found in the majority of patients with these tumors (8). A catalytically active TIGAR is required to promote tumor growth. However, other activities of TIGAR may also influence the biological properties of tumor cells.
Drugs targeting TIGAR's bisphosphatase activity may be beneficial in the management of breast-invasive ductal carcinoma. Inhibitors of complex I of OXPHOS such as metformin, of mitochondrial translation such as doxycycline, and BCL2 such as ABT-199 might have higher affinity in TIGAR-overexpressing tumors and will need to be studied further. Conversely, TIGAR overexpression induces resistance to the antiestrogen drug tamoxifen in vitro, and future studies will need to determine whether TIGAR expression predicts responsiveness to FIGURE 4. Effect of TIGAR on glycolysis and MCT1 in carcinoma cells. A, glucose uptake in TIGAR overexpressing carcinoma cells in coculture. RFP-tagged fibroblasts were cocultured with T47D cells either overexpressing TIGAR or control T47D cells. NBDG uptake in T47D cells was measured by flow cytometry. B, lactate production in TIGAR overexpressing carcinoma cells. Lactate production in T47D cells was measured and normalized to cell number. C, lactate production in HA-TOMM20-overexpressing carcinoma cells. Lactate production in MCF7 cells was measured and normalized to cell number. D, MCF7-overexpressing TIGAR and control T47D cells were cultured and lysed and subjected to immunoblot analysis for MCT1. MCF7 with TIGAR down-regulation and control cells using CRISPR-Cas9 were cultured and lysed and subjected to immunoblot for TIGAR (E) and MCT1 (F). tamoxifen in women with hormone receptor-positive breast cancer. Resveratrol has been shown to reduce TIGAR expression, and research into drugs targeting TIGAR is being conducted (29).
TIGAR is regulated by TP53, TP73, SP1, and AKT and indirectly by MYC, c-Met, and miR-144 (7, 30 -35). It is important to note that although TP53 up-regulates TIGAR expression experimentally, its expression in human breast cancer is inversely correlated to the expression of TP53 (8). Future work will need to determine the main drivers of TIGAR expression in breast cancer.
TIGAR is an enzyme that reduces glycolysis (7). TIGAR overexpression in carcinoma cells reduced glycolysis, as expected in the current study, and increased OXPHOS in these cells. Aggressive cancer cells with low glycolysis and high pentose phosphate activity have been described (36). High OXPHOS is also associated with aggressive cancer (26,37). The current study demonstrates that TIGAR induces reciprocal metabolic changes in adjacent fibroblasts with enhanced glycolysis compared with the metabolic effects in carcinoma cells. The increased fibroblast glycolysis in the current study occurred in the context of increased PFKFB3 and LDH-A expression with HIF1A activation. HIF1A is one of the main glycolytic transcription factors, and it regulates the expression of PFKFB3 and LDH-A. TIGAR in carcinoma cells also led to reduced LDH-B and reduced TIGAR expression in fibroblasts which increase glycolysis. It has been shown that glycolytic fibroblasts increase breast cancer aggressiveness and tumor growth (38) (39). In sum, the current study demonstrates that TIGAR induces aggressive breast cancer with metabolic reprogramming of the tumor microenvironment with low glycolysis in carcinoma cells and opposite changes in stromal cells. Future studies will need to determine the contribution of the metabolic state of carcinoma cells and stromal cells to breast cancer tumor growth.
The current study demonstrates that TIGAR expression in carcinoma cells reduces TIGAR expression in wild-type TP53 fibroblasts. TIGAR is regulated by TP53 (7). Additional factors that regulate TIGAR expression in carcinoma cells are likely to modulate TIGAR expression in fibroblasts. For example, FIGURE 5. Effect of carcinoma TIGAR expression on fibroblast glycolysis. A, glucose uptake in fibroblasts cocultured with TIGAR-overexpressing carcinoma cells. RFP-tagged fibroblasts were cocultured with T47D cells either overexpressing TIGAR or control T47D cells. NBDG uptake was measured by flow cytometry. B and C, fibroblasts with a GFP tag were cocultured with T47D cells either overexpressing TIGAR or control vectors. Fibroblasts were then lysed and subjected to immunoblot analysis for PFKFB3 (B) and LDHA and LDHB (C). HIF1A luciferase reporter in fibroblasts in coculture with carcinoma cells overexpressing TIGAR or control vector. Coculture was performed in normoxia (21% O 2 ) or hypoxia (0.5% O 2 ) (D). E, glucose uptake in fibroblasts overexpressing TIGAR. Fluorescent 2-deoxyglucose uptake was measured by flow cytometry using 2-NBDG.
TIGAR expression is also regulated by the transcription factors MYC and SP1 (31,33,40,41). TIGAR is also regulated by NFB and PHD1 (42). Hence, MYC and SP1 may regulate TIGAR expression in breast cancer-associated fibroblasts. Also, TP53 mutations in breast cancer-associated fibroblasts may alter TIGAR expression as TP53 mutations and deletions alter TIGAR expression in carcinoma cells (7,22,40). Carcinoma cells favor the selection of fibroblasts lacking wild-type TP53 (43). TP53 mutations have been demonstrated in breast cancerassociated fibroblasts (44). In addition, fibroblasts from patients with Li-Fraumeni syndrome have an abnormal TP53, as the syndrome is due to germ line abnormalities in TP53. In conclusion, high glycolysis in breast cancer-associated fibroblasts is induced by high TIGAR expression in carcinoma cells. The role of additional modulators of glycolysis such as MYC, SP1, and abnormal TP53 in breast cancer-associated fibroblasts needs to be studied.
This study reveals that TIGAR can induce tumor metabolic compartmentalization with highly glycolytic fibroblasts and carcinoma cells with reduced glycolysis and high OXPHOS. The effects of TIGAR on glycolysis, OXPHOS, and ATP are similar to the changes in marker expression by Western blot in this study. HIF1A activation in fibroblasts was studied to determine the potential mechanisms by which TIGAR overexpression in carcinoma cells induces metabolic asymmetry with increased fibroblast glycolysis. Glycolysis is induced upon HIF1A activation (10) (45,46). The current study demonstrates that HIF1A is activated in fibroblasts when TIGAR is overexpressed in carcinoma cells. It has been previously shown that HIF1A activation in breast cancer-associated fibroblasts promotes tumor growth (47). The contribution of fibroblast HIF1A activation to increased glycolysis, increased PFKFB3 expression, and tumor growth will need to be determined in future studies. mitochondrial markers in carcinoma cells exposed to lactate. MCF7 cells were cultured with DMEM medium with 10% FBS overnight. Then cells were incubated with and without 10 mM lactate for 48 h. Cells were then lysed and subjected to immunoblot analysis for TIGAR, MCT4, and BCL2. Actin was used as loading control. G and H, expression of markers of mitochondrial function in carcinoma cells exposed to tamoxifen. MCF7 cells were cultured with DMEM medium with 10% FBS overnight. Then cells were incubated with and without 10 M tamoxifen for 48 h. Cells were then lysed and subjected to immunoblot analysis for TIGAR, BCL-XL, BCL2, MITONEET (G), and OXPHOS labile subunits (H). GAPDH was used as loading control. I, MCF7 cells overexpressing MITONEET or control MCF7 cells were cultured alone or in coculture with GFP-tagged fibroblasts for 2 days. Then, cells were incubated with and without 10 M tamoxifen for 48 h. Apoptosis rate was measured as in A.
In sum, the current study provides mechanistic information on how TIGAR regulates fibroblast-carcinoma cell metabolic interactions in breast cancer. It has long been recognized that fibroblasts in proximity to cancer cells are pro-tumorigenic (4,48). Most studies have focused on how fibroblast-secreted factors lead to a fibroblast-activated state that promotes tumor growth (4,49). We set out to determine how the metabolic state of carcinoma cells modulates cancer aggressiveness and the metabolic phenotype of fibroblasts. TIGAR expression in breast carcinoma cells drives a mitochondrial OXPHOS phenotype in these carcinoma cells with a glycolytic phenotype in fibroblasts with increased tumor growth in vivo. Further work will need to determine the precise contribution of each of these factors to cancer aggressiveness.

Experimental Procedures
Materials-Materials were obtained as follows: NADP/NA-DPH quantification colorimetric kit (BioVision K347-100  or empty vector control were injected into the mammary fat pad of nude mice. Tumor volume and weight were measured after resection at 4 weeks post-injection. C, TOMM20 expression. T47D tumors overexpressing TIGAR or empty vector control tumor sections were stained by immunohistochemistry for TOMM20. Original magnification is 40ϫ. D, Fru-2,6-P 2 levels in T47D-TIGAR or control cells. E, tumor growth. T47D-TIGAR or control cells were injected into the mammary fat pad of nude mice. Tumor volume and weight were measured after resection at 4 weeks post-injection. F, proliferation rates. Mitotic figures in T47D tumors were quantified per high power field. G, Fru-2,6-P 2 levels. MCF7 cells overexpressing TIGAR or empty vector control were cultured, and Fru-2,6-P 2 levels were measured. H, tumor growth. MCF7-TIGAR or control cells were injected into the mammary fat pad of nude mice. Tumor volume and weight were measured after resection at 4 weeks post-injection. I, proliferation rates. Mitotic figures in MCF7 tumors were quantified per high power field. DECEMBER 16, 2016 • VOLUME 291 • NUMBER 51 mixture (MS601, Abcam), BCL2 (ab7973 and ab182858, Abcam for immune blot and immunohistochemistry, respectively), and BCL-XL (CS 2764, Cell Signaling).

TIGAR Reprograms Breast Cancer Mitochondrial Metabolism
Cell Culture-The human breast carcinoma cell lines (T47D, MCF7, and MDA-MB-231) were purchased from ATCC. Human skin fibroblasts immortalized with human telomerase reverse transcriptase catalytic domain (hTERT-BJ1) were purchased from Clontech, and clones were generated with green fluorescent protein (GFP) or red fluorescent protein (RFP) overexpression. Cells were passaged for fewer than 6 months after resuscitation. Cells were cultured in media containing 5 mM glucose and 1 mM pyruvate, 10% fetal bovine serum, 100 units/ml penicillin, and 100 units/ml streptomycin in a 37 ºC, 5% CO 2 incubator. RPMI 1640 was the base medium for T47D cells and DMEM for MCF7 and MDA-MB-231 as recommended by ATCC. To study the effect of lactate and glutamine, control conditions were compared with those with 2 mM glutamine or 10 mM lactate or 2 mM glutamine plus 10 mM lactate.
Co-culture System-Fibroblasts and carcinoma cells were cocultured as previously described (50). Briefly, cells were seeded at a 3:1 fibroblast to carcinoma cell ratio, and the total number of cells per well was 1 ϫ 10 5 . As controls, cultures of fibroblasts and carcinoma cells alone were plated in parallel with the same number of cells of a given type as in coculture (7.5 ϫ 10 4 fibroblasts and 2.5 ϫ 10 4 carcinoma cells).
Messenger RNA Quantification-Total RNA was extracted using the RNeasy Mini Kit(Qiagen). Complementary DNA (cDNA) was randomly primed from 2.0 g of total RNA using the SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen). Real-time PCR was performed in triplicate with a 1:3 dilution of cDNA using the Power SYBR Green PCR system. Relative expression levels were determined using the comparative ⌬⌬Ct method. All mRNA quantification data were normalized to ␤-actin.
Measurement of NADP/NADPH Ratio-Plated cells were washed and centrifuged at 2000 rpm for 5 min. Extraction buffer was admixed with two freeze and thaw cycles. NADP cycling enzyme was added as well as the NADPH developer, and measurement was at A 450 nm .
OCR Assessment-A Seahorse Bioscience XF96 Extracellular Flux Analyzer in a 24-well format was used. T47D cells were preseeded in the XF24 24-well plate with 100 l of RPMI 1640 media with 10% fetal bovine serum, 100 units/ml penicillin, 100 units/ml streptomycin, 1 mM pyruvate and additional 2 mM glutamine and/or 10 mM lactate in a 37 ºC, 5% CO 2 incubator for 2 h. After 2 h, an additional 200 l of growth medium was added for a total volume of 300 l per well. On the second day, cells were incubated in non-buffered RPMI 1640 media for 2 h. Measurements were obtained under basal conditions. OCR was normalized to total protein content. Statistical significance was examined using Student's t test. Values of p Ͻ 0.05 were considered significant.
ATP Assay-Intracellular ATP levels were measured using the ATP-sensitive fluorochrome quinacrine. Briefly, cells were incubated with 20 M quinacrine dihydrochloride at 37 ºC for 1 h, and green fluorescence intensity was measured by flow cytometry as previously described (51).
Apoptosis Assessment-Apoptosis in culture was quantified by flow cytometry using PI and annexin-V-APC as previously described (50).
Proliferation Assessment-For DNA content and proliferation analyses, cells were incubated with EdU (Click-iT EdU Flow Cytometry Assay kits) for 1 h. Then cells were stained with propidium iodide and anti-EdU-APC antibody. Cells were then analyzed by flow cytometry for nascent DNA synthesis (EdU incorporation) and ploidy assessment (PI). Proliferating cells in the xenograft tumors were identified on the basis of mitotic figures. Cells were counted in all fields within the central area of each tumor excluding regions with stromal elements using a 20ϫ objective lens and an ocular grid (0.25 mm 2 per field). The total numbers of mitotic figures per unit area was calculated, and the data were represented graphically. Two Cell Populations Sorting by Flow Cytometry-Fibroblasts with an RFP tag were cultured alone or in coculture with carcinoma cells for 4 days. Two cell populations were separated by flow cytometry through sorting with RFP-positive and RFPnegative cells. The presence of RFP in fibroblasts and lack of color in carcinoma cells allows us to separate these two cell populations in the flow cytometry cell sorter. After cell sorting, cells were then lysed and subjected to immunoblot analysis.
Measurement of Lactate Production-10 ϫ 10 4 of cells were plated in each well of a 12-well plate with cell culture medium, and on the 2nd day the medium was changed to RPMI 1640 with 2% FBS. After 24 h, cell media were collected and lactate production was analyzed as described (47).
Luciferase Activity-NIH3T3 fibroblasts stably transfected with a HIF1A luciferase reporter (RC0017, Panomics) were seeded in co-culture with carcinoma control or carcinoma TIGAR carcinoma cells. As controls, fibroblasts were plated in mono-culture using the same cell numbers as in the corresponding co-cultures. Luciferase activity was measured as previously described (52).
Immunohistochemistry-Sections were deparaffinized, rehydrated through graded ethanol, and washed in PBS. Antigen retrieval was performed in 0.01 M citrate buffer, pH 6.0, for 10 min using a pressure cooker followed by blocking for endogenous peroxidase in 3% H 2 O 2 for 15 min. An avidin-biotin kit (Biocare Medical) was used to block endogenous biotin. For BCL2 rabbit antibody (ab 182858, Abcam), sections were blocked with 10% goat serum and incubated with antibody for 1 h at room temperature. After washing in PBS, the sections were incubated with biotinylated goat anti-rabbit IgG (Vector Laboratories) for 30 min and then avidin-horseradish peroxidase complex (Vectastain Elite ABC kit, Vector) for 30 min. Antibody reactivity was detected using liquid DAB substrate chromagen (Dako). For TOMM20 the Vector M.O.M. Basic kit was used followed by ABC and DAB detection. Quantitative analysis of immunohistochemistry was performed employing Aperio software as previously described (54). Increasing intensity using Aperio was quantified as 0, 1ϩ, 2ϩ and 3ϩ.
Animal Studies-To evaluate the in vivo tumor-promoting effects of TIGAR, mammary gland injections were performed on female athymic NCr nude mice (NCRNU; Taconic Farms; at 6 weeks of age). All animals were maintained in a pathogen-free environment/barrier facility at the Kimmel Cancer Center at Thomas Jefferson University. Experiments were performed according to National Institutes of Health guidelines. The Institutional Animal Care and Use Committee (IACUC) approved all animal protocols. MDA-MB-231, MCF7, and T47D cells were injected into the mammary fat pads of nude mice. Cancer cells (1 million cells) and (0.3 million) hTERT-BJ1 fibroblasts were resuspended in 100 l of sterile PBS just before injection into the mammary fat pad. Oophorectomy and estrogen supplementation with 17␤-estradiol pellet placement (0.72 mg/pellet) were performed before injection of the estrogen receptor-positive cell lines MCF7 and T47D as previously described (23). After 4 weeks post-injection, tumors were excised to determine their weight and size. Volumes were calculated using the formula V ϭ (X 2 Y)/2, where V is the tumor volume, x is the length of the short axis, and y is the length of the long axis using electronic calipers. Statistical significance was examined using the Mann-Whitney U test. Values of p Ͻ 0.05 were considered significant.