18F-fluorodeoxy-glucose positron emission tomography marks MYC-overexpressing human basal-like breast cancers

Abstract

In contrast to normal cells, cancer cells avidly take up glucose and metabolize it to lactate even when oxygen is abundant, a phenomenon referred to as the Warburg effect. This fundamental alteration in glucose metabolism in cancer cells enables their specific detection by positron emission tomography (PET) following i.v. injection of the glucose analogue (18)F-fluorodeoxy-glucose ((18)FDG). However, this useful imaging technique is limited by the fact that not all cancers avidly take up FDG. To identify molecular determinants of (18)FDG retention, we interrogated the transcriptomes of human-cancer cell lines and primary tumors for metabolic pathways associated with (18)FDG radiotracer uptake. From ninety-five metabolic pathways that were interrogated, the glycolysis, and several glycolysis-related pathways (pentose phosphate, carbon fixation, aminoacyl-tRNA biosynthesis, one-carbon-pool by folate) showed the greatest transcriptional enrichment. This "FDG signature" predicted FDG uptake in breast cancer cell lines and overlapped with established gene expression signatures for the "basal-like" breast cancer subtype and MYC-induced tumorigenesis in mice. Human breast cancers with nuclear MYC staining and high RNA expression of MYC target genes showed high (18)FDG-PET uptake (P < 0.005). Presence of the FDG signature was similarly associated with MYC gene copy gain, increased MYC transcript levels, and elevated expression of metabolic MYC target genes in a human breast cancer genomic dataset. Together, our findings link clinical observations of glucose uptake with a pathologic and molecular subtype of human breast cancer. Furthermore, they suggest related approaches to derive molecular determinants of radiotracer retention for other PET-imaging probes.

Figure 1. Deriving an “FDG-uptake” metabolic gene expression signature

(A) Schematic of our experimental approach. (B) 18F-FDG retention in 16 cancer cell lines, including prostate cancer (CaP), glioblastoma (GBM), and melanoma (MEL). Error bars indicate standard error. (C) Clinical datasets. (Left) Tumor FDG-uptake in 18 breast cancer patients. SUV, Standardized Uptake Value. (Right) FDG-PET scan and Brain Magnetic Resonance Imaging (MRI) in a patient with anaplastic astrocytoma. (Area 1 FDG-high, Area 2 FDG-low). Cross-hairs are centered at sites of biopsy. (D) ‘High FDG uptake’ samples show transcriptional enrichment for glycolysis and glycolysis-branch metabolic pathways. Shown are average rank-based GSEA results (see Table S1 for the results of all KEGG metabolic pathways).

Figure 2. FDG signature score predicts in-vitro FDG-uptake

(A) RNA levels of core glycolysis enzymes (red) versus all genes (gray) in high FDG-PET (n=5, y-axis) versus low FDG-PET (n=6, x-axis) breast carcinomas. Tumors included in the analysis are denoted with asterisks in Table 1. (B) Average rank of Glycolysis/Gluconeogenesis, Carbon Fixation, and Pentose Phosphate Pathway enzyme members across all sample sets (see also Fig. S1). Low rank numbers represent high expression in the “FDG high” samples. Enzymes in green promote glycolysis, while those in red promote gluconeogenesis. The core glycolysis pathway was included as a point of reference. Enzyme rankings for the primary breast tumor-based FDG signature alone are shown in Fig. S1C. The schematic below shows the glycolysis/ gluconeogenesis pathway (and related pathways) as annotated by KEGG. Enzyme names are in italics. (C) Correlation between the observed (x-axis) and predicted (y-axis) FDG-uptake (see text and Fig, S1C for details) in seven breast cancer cell lines. The correlation (r=0.92) was statistically significant (sample label permutation p-value = 0.03). Cell lines from low to high FDG uptake are: HCC1500, BT474, ZR7530, ZR751, HCC70, UACC812, MCF7. (D) Correlations between observed FDG uptake and ‘FDG signature score’ predictions, using either only genes from the three glycolysis-related metabolic pathways which comprise the “FDG signature” (as in panel C; gly+cf+pp) or ‘lists of the top n genes’ with the highest differential expression between FDG-high and FDG-low breast cancers. Error bars: standard error.

Figure 3. FDG signature overlaps with “basal-like” breast cancer subtype

(A) Overlap between the FDG uptake signature and signatures for intrinsic breast cancer subtypes (16). Using the rank-rank hypergeometric overlap (RRHO) approach (Fig. S2), genes were ranked by their degree of correlation with FDG-PET SUV values across the tumors (n=18) or their degree of differential expression between the indicated subclasses to define the rank-based signature. The Spearman rank correlation coefficient (ρ) between signatures was calculated. All cases, except the ERBB2 case, had significant correlation based on sample permutation-based statistical analysis (p-value < 0.0001). (B) FDG-signature score preferentially identifies basal-like breast tumors. (Left) Schematic of experimental approach. (Right) Rank-ordered distribution of intrinsic human breast cancer subtypes relative to their predicted FDG signature score. See also Table S4. (C) Elevated genomic instability of human breast cancers with high ‘FDG signature score’ (top, n=18 high, 18 low) or high FDG-PET uptake (bottom, FDG-PET SUV>10 [n=5] compared to SUV < 5 [n=5]). Shown are cGH profiles with regions of copy-number gain (loss) shown as shades of red (blue). The graphs on the right show a quantification of gene copy number alterations. The higher absolute number of transitions per chromosome in the lower plot (versus upper plot) is due to the higher resolution of the cGH platform.

Figure 4. MYC activation in FDG-high primary human breast cancers

(A) Rank of MSigDB gene sets when analyzed for enrichment in the FDG-uptake signature. NES: normalized enrichment score. Enrichment results for all 1822 MSigDB C2 gene sets are listed in Table S5. (B) MYC IHC-positive breast cancers have higher FDG-PET SUV values. (Left) Representative MYC-IHC images (40x). Sample numbers and FDG-PET SUVs (in parenthesis) are shown above each panel. (Right) Distribution of FDG-PET SUVs in MYC IHC positive versus negative breast carcinomas. (C) RNA levels of MYC and MYC target genes in tumors with negative (left) versus positive (right) nuclear MYC staining by IHC. Red = high; p-values represent t-test analysis of the MYC IHC negative vs positive samples. (D) Breast Cancers with high FDG signature score are more likely to have elevated c-myc gene dosage (MYC DNA), MYC transcript level (MYC RNA), and MYC target gene expression. Tumors are ordered by FDG signature score (low to high); p-values represent a t-test analysis of the top 40 versus bottom 40 scoring FDG signature samples. Similar p-values were obtained for top/bottom 18 samples, or for the correlation of copy number or expression values with the FDG signature score.