GPCR-mediated PI3K pathway mutations in pediatric and adult thyroid cancer

Abstract

Whole exome sequencing (WES) recently identified frequent mutations in the genes of GPCR-mediated PI3K pathway (LPAR4, PIK3CA, and PTEN) in a Chinese population with papillary thyroid cancers (PTCs). The study found LPAR4 mutations as novel gene mutations in adult population with differentiated thyroid cancer (DTC). Here, we determine the prevalence of somatic mutations in this pathway (LPAR4 (exon 1), PIK3CA (exons 9 and 20) and PTEN (exons 5, 6, 7 and 8) in 323 thyroid samples consisting of 17 multinodular goiters (MNG), 89 pediatric DTCs, 204 adult DTCs, and 13 aggressive thyroid cancers including 10 poorly differentiated (PDTC) and 3 anaplastic thyroid cancer (ATC) from another ethnic population. We found 3.37% and 2.45% (includes Q214H, a novel PTEN mutation) in GPCR-mediated PI3K pathway of pediatric and adult DTCs, respectively. Analyses of 507 DTCs from thyroid Cancer Genome Atlas data (TCGA) revealed a low prevalence of mutations in this pathway (1.18%). In 13 cases with PDTC and ATC, we found no mutation in genes of this pathway. By contrast, analyses of 117 aggressive thyroid cancers (PDTC and ATC) from TCGA showed 13% of mutations in this pathway. Moreover, analyses of 1080 pan-cancer cell lines and 9020 solid tumors of TCGA data revealed high rates of mutations in this pathway (cell lines, 24.8%; tumors, 24.8%). In addition, PIK3CA + PTEN (p = <0.001) and LPAR4 + PIK3CA (p = 0.003) significantly co-occurred. Our study reveals a low prevalence of GPCR-mediated PI3K pathway mutations both in pediatric and adult DTCs corroborating the TCGA data and suggests a significant role of this pathway only in a small portion of DTCs. The high prevalence of mutations in this pathway in other solid malignancies suggests an important role in their pathogenesis making it an attractive target for therapeutic intervention both in a small subset of DTCs and other solid cancers.

Keywords: PIK3CA; PTEN; mutations; oncogene; thyroid cancer.

Figure 1. Identification of PIK3CA mutations in pediatric and adult DTCs.

The illustration shows the sequencing electropherogram of the PIK3CA gene in various samples. The sequence shown on top and bottom indicates the sense and antisense strand, respectively. Altered nucleotide positions and its related amino acid (within bracket) are shown on top of each corresponding sequence. Mutated nucleotides are indicated by arrows. Nucleotide number refers to the position within the coding sequence of the PIK3CA gene, where nucleotide position 1 indicates the first nucleotide of the translation initiation codon. All samples were sequenced in 2 repeated experiments with independent PCR by sense and antisense primers.

Figure 2. Functional domains of p110α and its interactome.

(A) Schematic diagram of p110α protein. Shown are various somatic mutations identified in the PIK3CA gene of pediatric and adult DTCs and located in various corresponding domains of p110α. (B) Structure of p110α with DTCs associated mutations identified in this study. Shows ribbon diagram of native modeled structure of p110α /niSH2 (p85) as a heterodimer and its PDB ID is 2RDO. Mutated amino acid residues in DTCs were plotted in p110α native protein structure using SWISS PDB viewer. All the mutated residues are shown in a sphere shape, p110α is shown in green color and the niSH2 domain of p85α is shown in red color. (C) Interactome shows the interacting network of PIK3CA (p110α). The network nodes represent proteins and the red-colored node indicates the query protein and first shell of interactors. Bond in light blue and purple color shows known interactions from the curated database and experimentally determined, respectively. Bond in green, red and dark blue shows predicted interactions from gene neighborhood, gene-fusions, and gene co-occurrence, respectively. Black indicates co-expression while parrot green and pale blue shows text mining and protein homolog, respectively.

Figure 3. Identification of PTEN mutations, corresponding domains, and interactome of PTEN.

(A) Chromatopherograms of the sequences of PTEN mutated samples. Sequences shown on the top and bottom indicate the sense and antisense strand, respectively of a sample. Altered nucleotide positions and its related amino acid (within bracket) are shown on top of each corresponding sequence. Arrows indicate the mutated nucleotides. Nucleotide number refers to the position within the coding nucleotides of the PIK3CA gene, where nucleotide position 1 shows the first nucleotide of the translation initiation codon. (B) Schematic diagram of PTEN protein. Marked-amino acids are various somatic mutations identified in the PTEN gene of pediatric and adult DTCs and located in various corresponding domains of PTEN. (C) A native modeled 3D structure of PTEN with pediatric and adult DTCs-associated mutations identified in this study. Ribbon diagram shows native modeled structure of PTEN protein and its PDB ID is 1D5R. Mutated amino acid residues in thyroid cancer are plotted in PTEN native protein structure using WebGL viewer as explained in materials and methods. All the mutated residues are marked in a sphere shape. (D) Interactome shows the interacting network of PTEN protein. The network nodes represent proteins and the red-colored node indicates the query protein and first shell of interactors. The bond represents protein-protein associations. Color of the bond and the type of associations are as mentioned above in Figure 2D.

Figure 4. Prevalence of GPCR-mediated PI3K pathway mutations in DTCs.

(A) OncoPrint of GPCR-mediated PI3K pathway- Thyroid cancer TCGA. The OncoPrint summarizes genomic alterations in LPAR4, PIK3CA and PTEN genes across the TCGA sample set. Each horizontal row indicates a gene and each vertical column shows a tumor sample. Green bars show nonsynonymous mutations and blue bars shows homozygous deletions. (B) Prevalence of LPAR4, PIK3CA, and PTEN in DTCs of various studies. The bar shows each LPAR4, PIK3CA and PTEN mutations in Sanger sequencing in the present study, and next-generation sequencing studies from China and USA (TCGA). (C) Mutations in the GPCR-mediated PI3K pathway. Bar indicates the combined frequency of mutation in the GPCR-mediated PI3K pathway in DTCs of the present study and other studies. (D) Interactome shows the interacting network of LPAR4 protein. The network nodes represent proteins and the red-colored node indicates the query protein and first shell of interactors. Bond represents protein-protein associations. Bond color and the type of associations are as mentioned above in Figure 3D. (E) Schematic diagram of the GPCR (LPAR4)-mediated PI3K signaling pathway. Growth factor activates the LPAR4 and activated LPAR4 triggers potential downstream signaling members such as PIK3CA, PTEN, and AKT, and the active PI3K pathway leads to growth, cell survival and inhibition of apoptosis.

Figure 5. Prevalence of GPCR-mediated PI3K pathway mutations in PDTC and ATC (aggressive thyroid cancers).

(A) Pie chart of aggressive thyroid cancers (PDTC and ATC). The chart shows the type of tumors, the number of samples in each type and total number of samples studied. (B) The histogram indicates the mutation frequency of GPCR-mediated PI3K pathway genes in thyroid cancers. The bar shows the overall frequency of LPAR4, PIK3CA and PTEN mutations in 6% and 30% of PDTC and ATC patients, respectively. (C–E) Mutation tab. The schematic diagrams indicate GPCR-mediated PI3K members (from top) LPAR4, PIK3CA (p110α) and PTEN protein domains and position of a particular mutation. The length of the line connected between mutation annotation and protein directly correspond to the number of mutated samples. The most frequent mutation is indicated in the diagram. (F) The OncoPrint tab. Tab shows the LPAR4, PIK3CA and PTEN mutations across the PDTC and ATC (MSKCC cohort). Each row represents a particular gene of GPCR-mediated PI3K pathway and each column shows a tumor sample. The green squares plotted on the columns are non-synonymous mutations. (G) Survival curve. The diagram is the Kaplan–Meier plot of overall survival of aggressive thyroid cancer-bearing patients (PDTC and ATC) absence or presence of PIK3CA mutations in blue and red color, respectively.

Figure 6. Mutated gene-mediated interactome of DTC and aggressive thyroid cancers (PDTC and ATC).

(A) Mutational landscapes of DTC. The thin column shows the top 188 mutated genes in DTC and extracted after analyses of TCGA data of DTC with a gene mutational prevalence cutoff of ≥ 0.70% in cBioPortal. The projected wide column indicates the top 50 mutated genes in DTC. (B) Mutational landscapes of aggressive thyroid cancers (PDTC and ATC). The thin column shows 186 genes which include all the mutated genes in PDTC and ATC. These genes were extracted as explained in Figure 6A. The projected wide column indicates the top 50 mutated genes in aggressive thyroid cancers (PDTC and ATC). (C and D) The interaction network (C and D) show the mutated gene-mediated interactome of DTC and aggressive thyroid cancers (PDTC and ATC), respectively. The interactome of DTC and aggressive thyroid cancers (PDTC and ATC) was constructed by STRING v10 after extracting the full set of mutated genes indicated in the mutational landscape in (A and B), respectively.

Figure 7. Prevalence of somatic mutations in genes of GPCR-mediated PI3K pathway in the pan-human cancer cell lines.

(A) Schematic diagram of LPAR4 protein. The diagram shows the domains and the positions of somatic mutations. The length of the line connecting the indicated mutation to the protein represents the number of samples which is positive for the mutation. The labeled mutation in the diagram represents the most recurrent mutation. (B) The frequency of LPAR4 mutation. Bars indicate the frequency of LPAR4 gene mutations across the human cancer cell lines. Only positive and a few negative cases are shown in the figure. (C) Schematic diagram of p110α protein (PIK3CA). The diagram shows various domains and the positions of somatic mutations as stated above in Figure 7A. (D) The frequency of PIK3CA (p110α) mutation. Bars indicate the frequency of somatic mutations of the PIK3CA gene across the human cancer cell lines as explained in Figure 7B. (E) Schematic diagram of PTEN protein. The diagram shows the domains and the positions of somatic mutations as stated above in Figure 7A. (F) The frequency of PTEN mutation. Bars indicate the frequency of somatic mutations of the PTEN gene across the human cancer cell lines as explained in Figure 7B. (G) OncoPrint of GPCR-mediated PI3K pathway of pan-human cancer cell lines. The OncoPrint bar summarizes genomic alterations in LPAR4, PIK3CA and PTEN genes across the TCGA sample set of 1080 human pan-cancer cell lines. Each horizontal row indicates a gene and each vertical column shows a tumor sample. Green bars show nonsynonymous mutations, blue bars show homozygous deletions, black shows truncating mutations, no color indicates the absence of mutation and gap shows not sequenced. (H) Frequency of GPCR-mediated PI3K pathway genes (combined) across the type of human cancer cell lines. Highest and the lowest frequency of these pathway mutations were observed in colorectal adenocarcinoma and melanoma, respectively.

Figure 8. Somatic mutations of the GPCR-mediated PI3K pathway in pan-human cancers (solid tumors) and their association with patient survival.

(A) Schematic diagram of LPAR4 protein. The diagram shows the domain and the positions of somatic mutations. The length of the line connecting the indicated mutation to the protein represents the number of samples which is positive for the mutation. The labeled mutation in the diagram represents the most recurrent mutation. (B) The frequency of LPAR4 mutation in pan-human cancers (solid). The histogram shows the somatic mutation frequencies of the LPAR4 gene mutations across cross cancer studies. (C) Schematic diagram of p110α protein (PIK3CA). The diagram shows various domains and the positions of somatic mutations identified in pan-human cancer (solid) as stated above in Figure 8A. (D) The frequency of PIK3CA mutation in pan-human cancers (solid). The histogram shows the somatic mutation frequencies of the PIK3CA gene cancer studies. (E) Schematic diagram of PTEN protein. The diagram shows various domains and the positions of somatic mutations identified in pan-human cancer (solid) as stated above in Figure 8A. (F) The frequency of PTEN mutation in pan-human cancers (solid). The histogram shows the somatic mutation frequencies of the PTEN gene across cancer studies. (G) OncoPrint of GPCR-mediated PI3K pathway of pan-human cancer. The OncoPrint bar summarizes genomic alterations in LPAR4, PIK3CA and PTEN genes across the TCGA sample set of 9020 pan-cancer samples. Each horizontal row indicates a gene and each vertical column shows a tumor sample. Green bars show nonsynonymous mutations, blue bars show homozygous deletions, black shows truncating mutations, no color indicates the absence of mutation and gap shows not sequenced. (H) Frequency of GPCR-mediated PI3K pathway gene mutations (combined) across the type of pan-human cancer. The highest frequency of these pathway mutations was observed in uterine endometrioid carcinoma and various types of breast carcinoma.