Spatial isoform analysis of the neonatal mouse heart reveals transcript-level patterns masked in gene-level analyses

Spatial transcriptomics has advanced the study of gene expression in tissues, but most current approaches rely on 3′-end sequencing and provide limited information on alternative splicing (AS). Spatially resolved isoform analysis can be achieved by combining microtissue sampling with Smart-seq2 RNA sequencing. Using 100-µm microtissues obtained from a single neonatal mouse heart (postnatal day 1; P01), we identified spatially variable transcript isoforms that are not readily detectable by conventional gene-level analyses. Although gene- and transcript-level clustering produced broadly similar spatial patterns, several genes exhibited isoform-level variation. For example, transcripts of Tnni1 and Ckb suggested spatial variation that was less structured than that of Pdlim5 , while Pdlim5 isoforms displayed distinct spatial distributions near the left ventricle. Comparison with postnatal day 7 (P07) hearts provided developmental context, suggesting that these spatial isoform patterns may undergo further remodeling during postnatal maturation. These results demonstrate the technical feasibility of spatial isoform analysis using a Smart-seq2-based workflow and provide a proof-of-principle framework for investigating spatial regulation of transcript isoforms during tissue development. Representative transcript-level observations were independently supported by isoform-specific qPCR. Although the biological findings require validation in independent samples, this approach establishes a practical strategy for transcript-level spatial analysis beyond conventional gene-level studies.