Translational control of tumor immune escape via the eIF4F-STAT1-PD-L1 axis in melanoma.

Preventing the immune escape of tumor cells by blocking inhibitory checkpoints, such as the interaction between programmed death ligand-1 (PD-L1) and programmed death-1 (PD-1) receptor, is a powerful anticancer approach. However, many patients do not respond to checkpoint blockade. Tumor PD-L1 expression is a potential efficacy biomarker, but the complex mechanisms underlying its regulation are not completely understood. Here, we show that the eukaryotic translation initiation complex, eIF4F, which binds the 5' cap of mRNAs, regulates the surface expression of interferon-γ-induced PD-L1 on cancer cells by regulating translation of the mRNA encoding the signal transducer and activator of transcription 1 (STAT1) transcription factor. eIF4F complex formation correlates with response to immunotherapy in human melanoma. Pharmacological inhibition of eIF4A, the RNA helicase component of eIF4F, elicits powerful antitumor immune-mediated effects via PD-L1 downregulation. Thus, eIF4A inhibitors, in development as anticancer drugs, may also act as cancer immunotherapies.

Regulation of RNA polymerase III transcription during transformation of human IMR90 fibroblasts with defined genetic elements.

RNA polymerase (Pol) III transcribes small untranslated RNAs that are essential for cellular homeostasis and growth. Its activity is regulated by inactivation of tumor suppressor proteins and overexpression of the oncogene c-MYC, but the concerted action of these tumor-promoting factors on Pol III transcription has not yet been assessed. In order to comprehensively analyse the regulation of Pol III transcription during tumorigenesis we employ a model system that relies on the expression of five genetic elements to achieve cellular transformation. Expression of these elements in six distinct transformation intermediate cell lines leads to the inactivation of TP53, RB1, and protein phosphatase 2A, as well as the activation of RAS and the protection of telomeres by TERT, thereby conducting to full tumoral transformation of IMR90 fibroblasts. Transformation is accompanied by moderately enhanced levels of a subset of Pol III-transcribed RNAs (7SK; MRP; H1). In addition, mRNA and/or protein levels of several Pol III subunits and transcription factors are upregulated, including increased protein levels of TFIIIB and TFIIIC subunits, of SNAPC1 and of Pol III subunits. Strikingly, the expression of POLR3G and of SNAPC1 is strongly enhanced during transformation in this cellular transformation model. Collectively, our data indicate that increased expression of several components of the Pol III transcription system accompanied by a 2-fold increase in steady state levels of a subset of Pol III RNAs is sufficient for sustaining tumor formation.

eIF4F is a nexus of resistance to anti-BRAF and anti-MEK cancer therapies.

In BRAF(V600)-mutant tumours, most mechanisms of resistance to drugs that target the BRAF and/or MEK kinases rely on reactivation of the RAS-RAF-MEK-ERK mitogen-activated protein kinase (MAPK) signal transduction pathway, on activation of the alternative, PI(3)K-AKT-mTOR, pathway (which is ERK independent) or on modulation of the caspase-dependent apoptotic cascade. All three pathways converge to regulate the formation of the eIF4F eukaryotic translation initiation complex, which binds to the 7-methylguanylate cap (m(7)G) at the 5' end of messenger RNA, thereby modulating the translation of specific mRNAs. Here we show that the persistent formation of the eIF4F complex, comprising the eIF4E cap-binding protein, the eIF4G scaffolding protein and the eIF4A RNA helicase, is associated with resistance to anti-BRAF, anti-MEK and anti-BRAF plus anti-MEK drug combinations in BRAF(V600)-mutant melanoma, colon and thyroid cancer cell lines. Resistance to treatment and maintenance of eIF4F complex formation is associated with one of three mechanisms: reactivation of MAPK signalling, persistent ERK-independent phosphorylation of the inhibitory eIF4E-binding protein 4EBP1 or increased pro-apoptotic BCL-2-modifying factor (BMF)-dependent degradation of eIF4G. The development of an in situ method to detect the eIF4E-eIF4G interactions shows that eIF4F complex formation is decreased in tumours that respond to anti-BRAF therapy and increased in resistant metastases compared to tumours before treatment. Strikingly, inhibiting the eIF4F complex, either by blocking the eIF4E-eIF4G interaction or by targeting eIF4A, synergizes with inhibiting BRAF(V600) to kill the cancer cells. eIF4F not only appears to be an indicator of both innate and acquired resistance but also is a promising therapeutic target. Combinations of drugs targeting BRAF (and/or MEK) and eIF4F may overcome most of the resistance mechanisms arising in BRAF(V600)-mutant cancers.

A fourth locus for autosomal dominant hypercholesterolemia maps at 16q22.1.

Autosomal dominant hypercholesterolemia (ADH) is characterized by isolated increase in plasmatic low-density lipoprotein (LDL) cholesterol levels associated with high risk of premature cardiovascular disease. Mutations in LDLR, APOB, and PCSK9 genes have been shown to cause ADH. We now report further genetic heterogeneity of ADH through the study of a large French family in which the involvement of these three genes was excluded. A genome-wide scan mapped the disease-causing gene, named HCHOLA4, at 16q22.1 in a 7.89-Mb interval containing 154 genes with a maximum LOD score of 3.9. To reduce the linked region, we genotyped 18 smaller non-LDLR/non-APOB/non-PCSK9-ADH families at the HCHOLA4 locus. Six families did not exclude linkage to the locus, but none allowed reduction of the disease interval. The 154 regional genes were sorted according to the function of the encoded protein and tissue expression profiles, and 57 genes were analyzed through sequencing of their coding region and close flanking intronic parts. No disease-causing mutation was identified in these families, particularly in the LCAT gene. Finally, our results also show the existence of other ADH genes as nine families were neither linked to LDLR, APOB, and PCSK9 genes nor to the new HCHOLA4 locus.

Novel mutations of the PCSK9 gene cause variable phenotype of autosomal dominant hypercholesterolemia.

Autosomal dominant hypercholesterolemia (ADH) is a frequent (1/500) monogenic inherited disorder characterized by isolated elevation of LDL leading to premature cardiovascular disease. ADH is known to result from mutations at two main loci: LDLR (encoding the low density lipoprotein receptor), and APOB (encoding apolipoprotein B100), its natural ligand. We previously demonstrated that ADH is also caused by mutations of the PCSK9 (proprotein convertase subtilisin/kexin type 9) gene that encodes Narc-1 (neural apoptosis-regulated convertase 1). However, the role of this novel disease locus as a cause of hypercholesterolemia remains unclear. In the present study, we analysed the PCSK9 coding region and intronic junctions in 130 adult or pediatric patients with ADH, previously found as being non LDLR/non APOB mutation carriers. Four novel heterozygous missense variations were found: c.654A>T (p.R218S), c.1070G>A (p.R357H), c.1405C>T (p.R469W), and c.1327G>A (p.A443T). All mutations were absent in 340 normolipidemic controls. Except for the A443T, all mutations are nonconservative and modify a highly conserved residue. Segregation with hypercholesterolemia is incomplete in one pedigree. Type and severity of hyperlipidemia and of cardiovascular disease could vary among subjects from the same family. Finally, the proband carrying the R357H mutation exhibited very high plasma cholesterol during pregnancy, whereas the proband carrying the p.R469W mutation exhibited a severe phenotype of hypercholesterolemia in combination with a LDLR mutation resulting from a frameshift at residue F382 (1209delC). These observations suggest that variations in PCSK9 are a rare cause of non LDLR/non APOB ADH (approximately 2.3%) and that additional environmental or genetic factors may contribute to the phenotype caused by PCSK9 missense mutations in humans.

NARC-1/PCSK9 and its natural mutants: zymogen cleavage and effects on the low density lipoprotein (LDL) receptor and LDL cholesterol.

The discovery of autosomal dominant hypercholesterolemic patients with mutations in the PCSK9 gene, encoding the proprotein convertase NARC-1, resulting in the missense mutations suggested a role in low density lipoprotein (LDL) metabolism. We show that the endoplasmic reticulum-localized proNARC-1 to NARC-1 zymogen conversion is Ca2+-independent and that within the zymogen autocatalytic processing site SSVFAQ [downward arrow]SIP Val at P4 and Pro at P3' are critical. The S127R and D374Y mutations result in approximately 50-60% and > or =98% decrease in zymogen processing, respectively. In contrast, the double [D374Y + N157K], F216L, and R218S natural mutants resulted in normal zymogen processing. The cell surface LDL receptor (LDLR) levels are reduced by 35% in lymphoblasts of S127R patients. The LDLR levels are also reduced in stable HepG2 cells overexpressing NARC-1 or its natural mutant S127R, and this reduction is abrogated in the presence of 5 mm ammonium chloride, suggesting that overexpression of NARC-1 increases the turnover rate of the LDLR. Adenoviral expression of wild type human NARC-1 in mice resulted in a maximal approximately 9-fold increase in circulating LDL cholesterol, while in LDLR-/- mice a delayed approximately 2-fold increase in LDL cholesterol was observed. In conclusion, NARC-1 seems to affect both the level of LDLR and that of circulating apoB-containing lipoproteins in an LDLR-dependent and -independent fashion.

Heterozygous TGFBR2 mutations in Marfan syndrome.

Marfan syndrome is an extracellular matrix disorder with cardinal manifestations in the eye, skeleton and cardiovascular systems associated with defects in the gene encoding fibrillin (FBN1) at 15q21.1 (ref. 1). A second type of the disorder (Marfan syndrome type 2; OMIM 154705) is associated with a second locus, MFS2, at 3p25-p24.2 in a large French family (family MS1). Identification of a 3p24.1 chromosomal breakpoint disrupting the gene encoding TGF-beta receptor 2 (TGFBR2) in a Japanese individual with Marfan syndrome led us to consider TGFBR2 as the gene underlying association with Marfan syndrome at the MSF2 locus. The mutation 1524G-->A in TGFBR2 (causing the synonymous amino acid substitution Q508Q) resulted in abnormal splicing and segregated with MFS2 in family MS1. We identified three other missense mutations in four unrelated probands, which led to loss of function of TGF-beta signaling activity on extracellular matrix formation. These results show that heterozygous mutations in TGFBR2, a putative tumor-suppressor gene implicated in several malignancies, are also associated with inherited connective-tissue disorders.

Mutations in PCSK9 cause autosomal dominant hypercholesterolemia.

Autosomal dominant hypercholesterolemia (ADH; OMIM144400), a risk factor for coronary heart disease, is characterized by an increase in low-density lipoprotein cholesterol levels that is associated with mutations in the genes LDLR (encoding low-density lipoprotein receptor) or APOB (encoding apolipoprotein B). We mapped a third locus associated with ADH, HCHOLA3 at 1p32, and now report two mutations in the gene PCSK9 (encoding proprotein convertase subtilisin/kexin type 9) that cause ADH. PCSK9 encodes NARC-1 (neural apoptosis regulated convertase), a newly identified human subtilase that is highly expressed in the liver and contributes to cholesterol homeostasis.

The UMD-LDLR database: additions to the software and 490 new entries to the database.

Mutations in the LDL receptor gene (LDLR) cause familial hypercholesterolemia (FH), one of the most frequent hereditary dominant disorders. The protein defect was identified in 1973, the gene was localized by in situ hybridization in 1985, and since, a growing number of mutations have been reported. The UMD-LDLR database is customized software that has been developed to list all mutations, and also to provide means to analyze them at the nucleotide and protein levels. The database has been recently modified to fulfill the recommendations of the Nomenclature Working Group for human gene mutations. However, in the current version, both the nomenclature and usual LDLR gene mutation names are reported since the latter are more commonly used. The software has also been modified to accommodate the splicing mutations and alleles that carry two nucleotide variations. The current version of UMD-LDLR contains 840 entries, of which 490 are new entries. Point mutations account for 90% of all mutations in the LDLR gene; the remaining are mostly major rearrangements, due to the presence of Alu sequences. Three new routines have been implemented in the software, thus giving users access to 13 sorting tools. In addition to the database, a Web site containing information about polymorphisms, major rearrangements, and promoter mutations is available. Both are accessible to the scientific community (www.umd.necker.fr) and should help groups working on LDLR to check their mutations and identify new ones, and greatly facilitate the understanding of functional classes/genotype relationships and of genotype/phenotype correlations.