Human mutations in integrator complex subunits link transcriptome integrity to brain development

Abstract

Integrator is an RNA polymerase II (RNAPII)-associated complex that was recently identified to have a broad role in both RNA processing and transcription regulation. Importantly, its role in human development and disease is so far largely unexplored. Here, we provide evidence that biallelic Integrator Complex Subunit 1 (INTS1) and Subunit 8 (INTS8) gene mutations are associated with rare recessive human neurodevelopmental syndromes. Three unrelated individuals of Dutch ancestry showed the same homozygous truncating INTS1 mutation. Three siblings harboured compound heterozygous INTS8 mutations. Shared features by these six individuals are severe neurodevelopmental delay and a distinctive appearance. The INTS8 family in addition presented with neuronal migration defects (periventricular nodular heterotopia). We show that the first INTS8 mutation, a nine base-pair deletion, leads to a protein that disrupts INT complex stability, while the second missense mutation introduces an alternative splice site leading to an unstable messenger. Cells from patients with INTS8 mutations show increased levels of unprocessed UsnRNA, compatible with the INT function in the 3'-end maturation of UsnRNA, and display significant disruptions in gene expression and RNA processing. Finally, the introduction of the INTS8 deletion mutation in P19 cells using genome editing alters gene expression throughout the course of retinoic acid-induced neural differentiation. Altogether, our results confirm the essential role of Integrator to transcriptome integrity and point to the requirement of the Integrator complex in human brain development.

Fig 1. Biallelic INTS8 mutations in a family with a severe neurodevelopmental syndrome.

(A-C) Magnetic resonance imaging (MRI) of affected individual III-2 showing cerebellar hypoplasia (A,C, arrow), reduced volume of the pons and brainstem and periventricular nodular heterotopia (B, arrows) versus (D-F) normal MRI from unaffected individual. (G) Pedigree of the extended family; filled symbols represent affected individuals. Below each individual the INTS8 alleles (wt = wild type) are shown. (H) Schematic of INTS8 including the four tetratricopeptide (TPR) motifs (blue blocks), the patient mutations and in the lower panel the conservation of the affected amino acids residues throughout evolution. (I-L) Electropherograms from Sanger sequencing of INTS8 wild type and mutant alleles.

Fig 2. Characterization of the INTS8 mutations.

(A) qRT-PCR on fibroblast-derived RNA of the patients (III-2, III-3, III-4), their unaffected sibling (III-1), and two age-matched control cell lines (C1, C2), normalized for GAPDH expression. Expression of the c.893A>G allele vs. wild type was measured using a primer located at the c.2917-2925del locus (INTS8 non-ΔEVL allele). (B) Schematic overview of INTS8 genomic and protein sequence. The c.893A>G mutation (in red) is located at the 5’ end of the exon8 of the transcript variant 3 that contains a premature stop codon (PTC). (C) Schematic of the GFP-minigene reporter construct used to evaluate the effect of the c.893A>G mutation on INTS8 exon 8 splicing pattern. Size of the corresponding amplicons is indicated on the left (D) RT-PCR analysis of RNA isolated from HeLa (lanes 1–3) or HEK293T cells (lanes 4–6) transfected with the GFP-minigene constructs. The empty reporter (GFP) is used as a control. (E) Western blot analysis of flag-affinity eluates from HEK293T stable lines expressing 3xFlag-tagged INTS8 wild type (WT) or INTS8ΔEVL. (F) Western blots on total cell extracts from patient and control primary fibroblasts. (G) qRT-PCR showing normalized expression of misprocessed U1, U2 and U4 snRNAs in total RNA extracted from patient III-2 and III-4 fibroblasts compared to two controls. All pairwise comparisons between patient and control UsnRNA levels are significant (at least p<0.05, Student’s T-test) to the exception of III-2 and C1 for UsnRNA U1 (p<0.06).

Fig 3. Dysregulated transcriptome in patient skin fibroblasts.

(A, B) qRT-PCR validation of gene expression variation in patient cells for two illustrative examples, NPTX1 and OSR2 mRNAs. (C) Correlation analysis of differential gene expression data from exon arrays (X axis) and RNA-seq (Y axis). (D) Pie chart representing the different types of alternative splicing events detected in patient cells vs control in RNA-seq data (n = 215, p<0.01, 292 total events). (E, F) Experimental verification by RT-PCR of the splicing changes associated with INTS8 mutations for two illustrative examples ADAM15 (E) and ATL3 (F) mRNAs.

Fig 4. Effect of INTS8ΔEVL mutation on P19 cell neuronal differentiation.

(A) Two P19 clonal cell lines bearing a homozygous INTS8ΔEVL mutation were generated by CRISPR/Cas9 mediated genome editing using two different guide RNAs. After genomic DNA extraction, the region surrounding the INTS8ΔEVL mutation is amplified by PCR and the corresponding DNA digested with NheI to detect homologous recombination or mock digested (Unc = uncut). The P19 parental cell line is used as a control. (B) The protein expression of different INT subunits is monitored by Western Blot in total cellular extracts of INTS8ΔEVL mutant P19 cell lines. The P19 parental cell line is used as a control. Tubulin serves as a loading control. (C) Expression of neuronal differentiation markers during RA-induced differentiation of P19 cells. Wild-type and INTS8ΔEVL P19 cell lines are treated with RA and let to differentiate for 8 days. Cells are harvested at the indicated time points after the initiation of the differentiation protocol (D0 = day zero, D2 = day two, D4 = day4, D8 = day8) and RNA was extracted and reverse transcribed. Gene expression is determined by qRT-PCR (n = 3, +/- SEM). GAPDH expression is used as a normalizer.