Defective lamin A-Rb signaling in Hutchinson-Gilford Progeria Syndrome and reversal by farnesyltransferase inhibition

Abstract

Hutchinson-Gilford Progeria Syndrome (HGPS) is a rare premature aging disorder caused by a de novo heterozygous point mutation G608G (GGC>GGT) within exon 11 of LMNA gene encoding A-type nuclear lamins. This mutation elicits an internal deletion of 50 amino acids in the carboxyl-terminus of prelamin A. The truncated protein, progerin, retains a farnesylated cysteine at its carboxyl terminus, a modification involved in HGPS pathogenesis. Inhibition of protein farnesylation has been shown to improve abnormal nuclear morphology and phenotype in cellular and animal models of HGPS. We analyzed global gene expression changes in fibroblasts from human subjects with HGPS and found that a lamin A-Rb signaling network is a major defective regulatory axis. Treatment of fibroblasts with a protein farnesyltransferase inhibitor reversed the gene expression defects. Our study identifies Rb as a key factor in HGPS pathogenesis and suggests that its modulation could ameliorate premature aging and possibly complications of physiological aging.

Figure 1. Genome-wide expression profiling of HGPS and control fibroblast cultures.

(A) Microarray plot profiles indicate changes in gene expression in control, HGPS, FTI-treated control and FTI-treated HGPS fibroblasts. Each continuous line corresponds to the normalized intensity value of an individual probe set. Line colors denote the intensity of the signal (red: strong and blue: low signal). Probes that satisfied a greater or less than two-fold cutoff and statistically significant difference of p <0.01 are displayed. (B) Pie chart indicates the predicted subcellular localization of proteins encoded by the 352 genes differentially expressed in HGPS. The list of differentially expressed genes in HGPS versus control cells was analyzed using Ingenuity Pathway Analysis (IPA) and encoded proteins assigned a subcellular localization based on information contained in the Ingenuity Knowledge Base. (C) Genes differently expressed in HGPS (352 genes) were assigned to diverse cellular functions using the “Functional Analysis” tool of IPA software (www.ingenuity.com). Columns represent groups of genes associated with specific cellular functions (x-axis). The significant genes were compared to IPA database and ranked according to a p-values generated with Fisher extract test. P-values less than 0.05 indicates a statistically significant, non-random association between a set of significant genes and a set of all genes related to a given function in IPA database. The ratio (y-axis) represents the number of genes from the dataset that map to the pathway divided by the number of all known genes ascribed to the pathway. The yellow line represents the threshold of p<0.05. (D) As described above, the 352 genes were assigned to diverse physiological systems according to IPA.

Figure 2. The lamin A-Rb network.

The main network (center left) shows downstream interactions between lamin A/C and the 352 differentially expressed genes in HGPS fibroblasts. The network was built using MetaCore analyses, which finds known interactions between gene products. (A) The network divides into several distinct regions, the main region being the regulatory and signaling network. (B) The detailed view of the main region (top right) shows that the only gene differentially expressed in HGPS directly downstream of lamin A/C (layer 1) is that encoding Rb (layer 2). In turn, most of the immediate downstream partners of Rb included in this dataset are p107, NCOA2, SP1, and ATF-2 and E2F1 (layer 3). Of the remaining genes that occur downstream of layer 3, nearly half of the 280 genes interact with only one or more of these transfactors associated to the main network (center left); these gene products are placed above the centre. From layer 4 and 5, most of the genes can be connected at least to one entity according to GeneGO. In the HGPS dataset, 124 genes that had no known interactions in GeneGO are not shown. From the center regulatory and signaling network (zoom left, panel B)), several groups of genes segregate into six subnetworks, based on mutual interactions: a group denoted DNA repair (bottom left, panel (C)), motility (bottom right, panel (D)), and three circles of genes regulated by E2F1, ATF-2 and SP1, respectively. Symbols associating the genes with functions are indicated. Genes labeled with blue circles were downregulated and red circles were upregulated. Higher magnifications of panels A to D are provided in Figures S1, S2, S3 and S4.

Figure 3. FTI inhibition of prelamin A and progerin farnesylation and FTI-induced gene expression changes in fibroblasts from normal subjects.

(A) Western blot analysis of protein extracts of fibroblasts from individuals with HGPS and controls that were treated or untreated with FTI (1.5 µM lonafarnib daily for three days). Blots were probed with anti-lamin A/C (LA/C), anti-prelamin A (preLA), anti-actin, and anti HDJ-2 antibodies. Increase in levels of prelamin A and non-farnesylated HDJ-2 with FTI treatment are indicated. (B) 65 genes were differentially expressed in FTI-treated control cells. The signaling network was built upon functional association between protein farnesyltransferase (FTase), the enzymatic activity of which is inhibited by FTI, lamin A/C and Rb even though expression of all three transcripts remained unchanged with FTI treatment. Of note, lamin A is a known substrate for FTase. Downstream FTase interactions between all the 65 genes differentially expressed in FTI-treated control cells have been incorporated based on their interactions according to MetaCore analysis. The MetaCore analysis identified two transfactors, E2F1 and c-Myc, and one transfactor regulator, REP-1 that permit linking all those 65 genes into this single network. Symbols associating the genes functions are indicated. Genes labeled with blue circles were downregulated and by red circles were upregulated.

Figure 4. Genome-wide expression profiling of FTI-treated and untreated fibroblasts from subjects with HGPS.

(A) Venn diagram comparison of microarray datasets of controls versus control-FTI, or HGPS and/or HGPS-FTI. 65 genes were differentially expressed in fibroblasts from control subjects after FTI treatment compared to no treatment (Control+FTI; red); 352 in fibroblasts from subjects with HGPS compared to those from control subjects (HGPS; green); and 804 in fibroblasts from subjects with HGPS after FTI treatment compared to untreated controls (HGPS+FTI; blue). Genes with altered expression common between these different datasets are shown as areas of overlap with different colors in the Venn diagram with numbers of common genes indicated. (B) Pie chart indicates the subcellular localization of the encoded proteins of the differentially expressed genes between control and HGPS-FTI datasets according to information contained in the Ingenuity Knowledge Base. (C) Genes differentially expressed between controls versus HGPS-FTI datasets were assigned to diverse cellular functions according to IPA and (D) were associated to canonical pathways according to IPA. The significant genes were compared to IPA database and ranked according to p-values generated with Fisher extract test. P-values less than 0.005 indicate a statistically significant, non-random association between a set of significant genes and a set of all genes related to a given function in IPA database. The ratio (y-axis) represents the number of all known genes ascribed to the pathway. The Yellow line represents the threshold of p<0.05. (E) Comparison of gene expression alterations between FTI-treated control cells (yellow) with FTI-treated fibroblasts from subjects with HGPS (red). Using criteria of corrected p-value from unpaired t-test <0.01 and two-fold change in expression, only one gene, PIK3CB was upregulated in FTI treated cells from subjects with HGPS compared to cells from normal subjects treated with FTI. Therefore, 99% of the genes in FTI-treated cells from subjects with HGPS were expressed at the same levels as in FTI-treated normal fibroblast.

Figure 5. Real-time RT-PCR validation of microarray results for selected genes.

(A) Validation of a set of genes identified in microarray analysis comparing fibroblast from subjects with HGPS to controls. The mean value of expression for indicated genes measured using real time PCR (red; p<0.05), and microarray (blue; p<0.01) are shown. (B) Validation of a set of genes identified in microarray analysis comparing fibroblasts from control subjects with and without FTI treatment. (C) Validation of a set of genes identified in microarray analysis comparing fibroblasts from subjects with HGPS with and without FTI treatment. (D) Relative expression of LMNA (L, red) and RB1 (R, blue) in human skin derived from 16 unaffected individuals of various age groups was analyzed by quantitative real-time RT-PCR. A relative quantification method was used and the data from the 38 to 55-year-old age group were calibrated to a relative quantity of 1. All values were normalized to GAPDH (Materials and Methods).

Figure 6. Detection of lamin A/C and Rb in cells from subjects with HGPS and in normal human skin biopsies.

(A) Left panel: Western blot of proteins extracted from control fibroblasts (Control-1 [GM02036A], Control-2 [GM03349C]) and fibroblasts from subjects with HGPS (HGADFN127, HGADFN155, HGADFN164, HGADFN188) sequentially probed with anti-lamin A/C, anti- lamin A, anti-progerin (progerin) and anti-actin antibodies. Right panel: Western blot of protein extracts from skin of individuals at different ages as indicated (49 to 88 Year old, Y denote year). Blots were sequentially probed with antibodies cited above. (B) In situ detection of lamin A/C and Rb in human skin sections. Skin sections from 93-year-old and 38-year-old subjects were probed with antibodies against lamin A/C (red) and Rb (green) and signal overlap is shown at right (Lamin A/C/Rb). (C) Skin sections as in panel (B) were probed with antibodies against lamin A (green) and progerin (red) and DNA was stained with 4′,6-diamidino-2-phenylindole (blue); the merged signals are indicated. Scale bar, 20 µM.