Transcriptome analysis of genetically matched human induced pluripotent stem cells disomic or trisomic for chromosome 21

Abstract

Trisomy of chromosome 21, the genetic cause of Down syndrome, has the potential to alter expression of genes on chromosome 21, as well as other locations throughout the genome. These transcriptome changes are likely to underlie the Down syndrome clinical phenotypes. We have employed RNA-seq to undertake an in-depth analysis of transcriptome changes resulting from trisomy of chromosome 21, using induced pluripotent stem cells (iPSCs) derived from a single individual with Down syndrome. These cells were originally derived by Li et al, who genetically targeted chromosome 21 in trisomic iPSCs, allowing selection of disomic sibling iPSC clones. Analyses were conducted on trisomic/disomic cell pairs maintained as iPSCs or differentiated into cortical neuronal cultures. In addition to characterization of gene expression levels, we have also investigated patterns of RNA adenosine-to-inosine editing, alternative splicing, and repetitive element expression, aspects of the transcriptome that have not been significantly characterized in the context of Down syndrome. We identified significant changes in transcript accumulation associated with chromosome 21 trisomy, as well as changes in alternative splicing and repetitive element transcripts. Unexpectedly, the trisomic iPSCs we characterized expressed higher levels of neuronal transcripts than control disomic iPSCs, and readily differentiated into cortical neurons, in contrast to another reported study. Comparison of our transcriptome data with similar studies of trisomic iPSCs suggests that trisomy of chromosome 21 may not intrinsically limit neuronal differentiation, but instead may interfere with the maintenance of pluripotency.

Fig 1. Confirmation of differentiation of iPSCs into cortical neuronal cultures.

Images taken from clones C2 (trisomic) and C3 -D21 (disomic) 40 days after initiation of the differentiation protocol. Fixed cells on chamber slides were probed with antibodies against the marker proteins listed in the left column. Neuronal marker Beta III tubulin is red; DAPI is blue; and Ctip2, TBR1, SLC17A7 (vGLUT1) and GRIK2 are green. Size bar = 20 μ.

Fig 2. Principal component analysis of transcriptome datasets.

Note principal component 1 clearly segregates iPSC from neuronal cultures, while component 2 partitions trisomic from disomic cells.

Fig 3. Trisomic iPSCs overexpress neuronal markers.

A. RT-PCR quantitation of transcripts overexpressed in trisomic iPSCs relative to disomic iPSCs. cDNA was prepared from the same RNA used for the original sequencing libraries, specifically early passage trisomic iPSC culture (C2B) and an early passage disomic clone (C2-4-3). Similar expression levels were observed using RNA from independent trisomic (C2C) and disomic (C2-4-3) clones. B. iPSC colonies (trisomic, C2B; disomic C2-4-4) fixed and stained for MAP2 and OCT4. C. Representative immunoblot from trisomic and disomic iPSCs. Note two MAP2 isoforms are detected. D. Quantitation of MAP2 immunoblots. (* = P<0.02, ** = P< 0.01, ***, P< 0.001, paired T-tests).

Fig 4. Expression of Chr 21 genes in iPSCs and cortical neuronal cultures (trisomic/disomic ratios).

Lines represent individual gene ratios across chromosome 21. Note strong general trend for Chr 21 genes to be over-expressed, although specific genes may be strongly up-or-down regulated depending on cell type. (Black lines indicate 1.5 fold increases or decreases in trisomic cells).

Fig 5. RNA editing levels do not differ between trisomic and disomic cortical neuronal cultures.

SPRINT software [25], which employs clustering of editing sites to distinguish bona fide RNA editing from DNA polymorphisms and sequencing errors, was used to quantify overall RNA editing levels. Shown are percent overall editing (reads containing edits/potential editing sites) calculated from transcriptome data from biological replicates of trisomic (C2) and disomic (C3 Di) cortical neuronal cultures.

Fig 6. Altered splicing pattern of APOO in trisomic cells.

A. JunctionSeq output supporting the skipping of exon 4 in trisomic iPSCs. Black arrows indicate primers designed to amplify alternatively spliced exon; white arrows indicate control exons 6 and 7. B. RT-PCR targeting exon 4 exclusion. Arrow indicates exon exclusion band (band "B") only recovered in trisomic cells. Quantitative RT-PCR was used to show that the included exon is ~ 10 fold enriched in the trisomic samples relative to the unchanged control exons (C2B: C243, 12.7 +/- 1.3 SEM fold, C2B:C244, 9.4 +/- 1.6 SEM fold).

Fig 7. Cluster analysis of comparable studies using isogenic disomic/trisomic cell pairs.

A. Dendrogram representation of relatedness of datasets listed in Table 2. DS1 and DS4, from Weick et al. [13]; clones 1,2,3 from Jiang et al [14]; "twin" from Letourneau et al [15]. B. Section of heat map display showing expression ratios for individual genes calculated for datasets shown in (A.). Note overall similarity between "twin" data (leftmost column 1) and the data from this study (column 2).