Precision medicine driven by cancer systems biology

Abstract

Molecular insights from genome and systems biology are influencing how cancer is diagnosed and treated. We critically evaluate big data challenges in precision medicine. The melanoma research community has identified distinct subtypes involving chronic sun-induced damage and the mitogen-activated protein kinase driver pathway. In addition, despite low mutation burden, non-genomic mitogen-activated protein kinase melanoma drivers are found in membrane receptors, metabolism, or epigenetic signaling with the ability to bypass central mitogen-activated protein kinase molecules and activating a similar program of mitogenic effectors. Mutation hotspots, structural modeling, UV signature, and genomic as well as non-genomic mechanisms of disease initiation and progression are taken into consideration to identify resistance mutations and novel drug targets. A comprehensive precision medicine profile of a malignant melanoma patient illustrates future rational drug targeting strategies. Network analysis emphasizes an important role of epigenetic and metabolic master regulators in oncogenesis. Co-occurrence of driver mutations in signaling, metabolic, and epigenetic factors highlights how cumulative alterations of our genomes and epigenomes progressively lead to uncontrolled cell proliferation. Precision insights have the ability to identify independent molecular pathways suitable for drug targeting. Synergistic treatment combinations of orthogonal modalities including immunotherapy, mitogen-activated protein kinase inhibitors, epigenetic inhibitors, and metabolic inhibitors have the potential to overcome immune evasion, side effects, and drug resistance.

Keywords: ARID1A; ARID2; BRAF; Big data; CSD; CTLA4; Cancer metabolism; Cancer systems biology; Combination therapy; Driver; EZH2; Epigenomics; Immunotherapy; MEK; Melanoma; Neoantigen; Oncometabolite; PD1; PDL1; PRC2; Personalized medicine; Precision medicine; SCNA; SWI/SNF; Subtype.

Fig. 1

Essential components in precision medicine and its application to cancer. The precision medicine infrastructure relies on a fruitful interplay between a a collaborative research team of clinicians and scientists, b personalized data allowing for fast and seamless interpretation, and c targeted pharmacological strategies. Precision disease management comprises targeted, personalized treatment aligned with the patient’s genotype offering confidence to receive/provide care and hope for cure. In malignant melanoma, rational, orthogonal combinations of immune checkpoint inhibitors (IMMUNOi), epigenetic inhibitors (EPIi), metabolic inhibitors (METABi), or inhibitors of specific signaling pathways such as the mitogen-activated protein kinase inhibitors (MAPKi) will provide best standard of care. Balanced and effective data sharing is based on patient consent, secure data exchange, and synergistic, standardized data formats

Fig. 2

Rate and challenges of cancer genomics studies with sequenced tumors. a Sequenced and publically available cancer genomics studies per year. The data contains cancer genomics studies with a total of more than 20,000 whole exome sequenced or whole genome sequenced tumor specimens. Exponential trend lines are based on current approximate slope. Extrapolated data indicated by dashed line. b Exponential increase and trend lines of whole genome and whole exome sequenced specimens per year for melanoma and studies across all tissues (Pan-cancer). c Melanoma specimens with high mutational burden carry numerous somatic alterations and require somatic mutation calls at high frequency. Depending on cancer tissue, cohort size, and mutational burden, the dynamic range of somatic mutation calls can span several orders of magnitude. Average trend lines are shown for different cancers tissues with low (cyan), medium (blue), and high (purple) mutational burden. Despite the fact that the currently available melanoma cohort covers less than a tenth of all cancer tissues, the data demands of somatic mutation calls in melanoma genomics are equally challenging to that of other current cancer genomics studies combined

Fig. 3

Precision medicine profile and rational drug targeting of malignant melanoma. a Whole-slide tumor tissue image of malignant melanoma shows tumor microenvironment and its impact on metabolically and mitotically active melanoma cells. b Multiomics data of tumor specimen is matched with high-throughput functional genomics data of cellular cultures. The personalized precision medicine chart shows deregulation of important signaling and epigenetic molecules across matched data tracks of genomic, epigenomic, transcriptomic, proteomics, and metabolomics platforms. c Comparison of patient data with somatic copy number alterations (SCNAs) of melanoma cohort identifies significant amplifications (red) and d deletions (blue). Detected somatic alterations of BRAF and EZH2 coincide, fall into mutational hotspots, and result in gain-of-function oncogenes. The somatic landscape of ARID1A and ARID2 is characterized by somatic non-sense and missense mutations, which result in loss of function of a tumor suppressor complex involved in chromatin remodeling. e Rewiring of metabolism and metabolic signaling affects melanogenesis and tunes central carbon metabolism to support evasion, proliferation, and survival of malignant melanoma. f Oncometabolites impact the epigenetic machinery by blocking or supplying carbons for histone modifiers. Epigenetic master regulators in cancer control transcriptional activation of other oncogenes or repression of tumor suppressors. g Personalized medicine strategies to overcome treatment resistance of malignant melanoma patients. h Patient profiles of The Cancer Genome Altas (TCGA) reveal co-occurrence of BRAF and EZH2 hyperactivity making combination therapy of mitogen-activated kinase inhibitors (MAPKi) and epigenetic inhibitors (EPIi) viable. As alternative option if immunotherapy (IMMUNOi) fails due to immune evasion and suppression of immune receptors, combination therapy of IMMUNOi and metabolic inhibitors (METABi) is sensible

Fig. 4

Precision medicine profile and rational drug targeting of malignant melanoma. a Rewiring of metabolism and metabolic signaling affects melanogenesis and tunes central carbon metabolism to support evasion, proliferation, and survival of malignant melanoma. b Oncometabolites impact the epigenetic machinery by blocking or supplying carbons for histone modifiers. Epigenetic master regulators in cancer control transcriptional activation of other oncogenes or repression of tumor suppressors. c Personalized medicine strategies to overcome treatment resistance of malignant melanoma patients. d Patient profiles of The Cancer Genome Atlas (TCGA) reveal co-occurrence of BRAF and EZH2 hyperactivity making combination therapy of mitogen-activated kinase inhibitors (MAPKi) and epigenetic inhibitors (EPIi) viable. As alternative option, if immunotherapy (IMMUNOi) fails due to immune evasion and suppression of immune receptors, combination therapy of IMMUNOi and metabolic inhibitors (METABi) is sensible

Fig. 5

Melanoma subtypes and UV signature of mutagenesis in melanoma drivers. a Melanoma subtypes are divided by (I) genomic and (II) non-genomic activation of the mitogen-activated protein kinase (MAPK) pathway. Melanoma with genomic MAPK activation contain (IA) non-chronic sun-induced damage (non-CSD) melanomas with BRAF(V600E) mutation (50%) and (IB) chronic sun-induced damage (CSD) melanomas with mutations of central genes of the MAPK pathway including KIT, NRAS, NF1, BRAF(non-V600E), MAP2K, or MAPK (∼30%). Both non-CSD and CSD share subtype (I) and have genomic activation of the MAPK pathway in common (∼80%). Any melanoma without genomic activation of MAPK elements is defined as subtype (II) (∼20%). Subtype (II) melanomas with low mutational burden are enriched in driver genes of (IIA) membrane receptors and/or G protein signaling, (IIB) epigenetic regulators, and/or (IIC) metabolic regulators. Subtype (II) frequently share effector activation of MAPK signaling by non-genomic mechanisms. b UV signature of nucleotide transitions in melanoma is determined by exome-wide genome sequencing. Percentage of driver mutations caused by UVA (C>A) or UVB (C>T) are plotted vs the exome-wide sample average percentage. Example genes a–e provide representative examples of melanoma drivers: i BRAF, ii RAC1, iii STK11, iv NRAS, and v CDKN2A. The non-CSD driver gene BRAF contains a high fraction of non-UV mutations deviating from exome-wide sample average; CSD drivers contain nucleotide transversions by cyclobutane pyrimidine dimers (CPDs) responsible for the initiation of the predominant C>T UV melanoma signature mutations. Tumor suppressor genes with high mutational burden and co-occurrence with CSD genes can closely reflect the exome-wide sample average of UV-related base transitions. Oncogenic drivers are subject to molecular evolution selecting for specific base transitions, which can cause deviation of the exome-wide sample average. Structural representation of non-CSD and CSD mutations of melanoma driver BRAF. Residue V600 in the center of the activator loop of BRAF is highlighted in purple. The comprehensive somatic molecular landscape of BRAF in melanoma is illustrated in Figs. 3 and 4. CSD hotspots include the activator loop in pink and the ATP-binding site in yellow

Fig. 6

Pathway analysis of genomic alterations in malignant melanoma. a Oncogenic alterations of MAPK pathway is illustrated by somatic mutations, SCNAs, and transcriptional alterations (from inner box to outer box; see legend). Genomic activation and inactivation are shown in red and blue, respectively. Genomic co-occurrence and mutual exclusivity points to multiple parallel signals within pathway. b MAPK pathway. c PI3K pathway. d WNT and G protein signaling. e Cell cycle, senescence, and apoptotic signaling. Signaling maps of melanoma and melanogenesis highlight predominant oncogenes and tumor suppressors, in red and blue, respectively

Fig. 7

Molecular signatures and genomic alterations in malignant melanoma. Oncogenic alterations of a MAPK pathway; b BRAF at the mutational level (MUT), the protein level (PROT), the level of somatic copy number alterations, and at the gene expression level; c tumor suppressors PTEN and TP53 correlating with total mutation burden and age at diagnosis; d the PI3K/PTEN/AKT axis; e cell cycle inhibitors and tumor suppressors CDKN2A, CDKN2B, and RB1 correlating with total mutation burden and fraction of the genome altered; and f the E2F family of cell cycle-related transcription factors. Somatic mutations are marked as dots. SCNAs are highlighted as red and blue bars for amplification and loss, respectively. Significant deviation of mRNA gene expression by more than two standard deviations is marked with transparent box for each skin cutaneous melanoma (SKCM) patient of The Cancer Genome Atlas (TCGA) cohort