Spatial transcriptomics using light-directed DNA synthesis

Spatial transcriptomics is a powerful technology for detailed molecular characterization of tumor tissue at the transcriptional level. We have developed a novel and cost-effective spatial transcriptomics platform that provides an unbiased view of the transcriptome at single-cell resolution. Its advantages are low-cost glass slides, single-cell resolution, and open chemistry for customized library preparation. For spatial analysis, sections are first stained with hematoxylin and eosin, after which bright-field microscopic imaging is performed. These images are combined with detailed mRNA expression patterns to visualize tumor tissue organization, cellular heterogeneity, and treatment effects.

Microarrays have been instrumental in the genomics revolution to parallelize hybridization-based genetic readouts such as gene expression, gene copy number, and single-nucleotide polymorphisms. While superseded now by massively parallel sequencing, microarrays are still useful, amongst others thanks to their ability to have spatially resolved oligonucleotides printed or synthesized in situ. Unfortunately, most arrays have the reactive 3’ side of the oligonucleotide probe attached to the glass surface, rendering them inutile as a cDNA synthesis primer.

Our collaborator Prof. Somoza (Vienna, Austria) has unique chemical expertise for in situ synthesis of DNA oligonucleotides at high resolution and developed an open-source photolithographic DNA synthesizer for massively parallel synthesis of oligonucleotides from 5’ to 3’, making this type of microarray suitable for spatial-omics applications (1). Our custom-built DNA synthesizer has a printable array size of 14 x 10 mm, consisting of 786,432 squares with single-cell resolution (13.6 x 13.6 μm) and a spacing of only 1 μm between spots (Figure 1). Printing can be done in different patterns: 1:1 (full density), 1:2 (checkerboard) or 1:4. Custom-printed microarrays, using micromirror-guided photolithography, are low-cost alternatives compared to commercial slides.

Designed entirely in-house, the platform is highly versatile, supporting applications such as short-read (Illumina) or long-read (Nanopore) mRNA profiling (transcriptome-wide or targeted) and targeted (methylated) DNA analyses (Figure 2). This fosters collaborations and advancement in precision medicine.

Application 1 – Whole transcriptome spatial analysis: Fresh frozen, O.C.T. embedded tissue is sectioned using a cryotome (Figure 3). Tissue sections are positioned onto the microarrays, stained with hematoxylin and eosin, and imaged using a microscope. After permeabilization, the poly(A) tail of cellular transcripts binds the anchored oligo(dT) part from the capture probe, after which reverse transcription (RT) initiates first-strand copy DNA (cDNA) synthesis. Next, second strand cDNA synthesis is done using Nextera-tailed random primers. After denaturation, the second-strand cDNA is eluted from the slide. Afterwards, indexed amplification, library cleanup, and sequencing are performed. Through data-analysis, the barcode will indicate the original capture location of the transcript on the slide. Combining this information with the hematoxylin-eosin-stained image provides unique genome-wide spatial information. A custom-developed containerized pipeline panoramaseq processes the FASTQ files into a standard h5ad object that bundles per-barcode gene counts with the histology image, enabling downstream analysis with tools such as Squidpy or other popular spatial transcriptomics frameworks.

Application 2 – Targeted transcriptome spatial analysis: To overcome the low detection sensitivity of low-abundant transcripts in a whole-transcriptome approach, a targeted workflow is under development to selectively quantify transcripts of interest. By replacing the second-strand cDNA synthesis with a targeted, limited-cycle pre-amplification PCR, transcripts of interest are enriched. We use a primer pair consisting of a 3’ universal, non-tailed, Nextera-compatible primer and a 5’ target-specific primer (Figure 4). For the latter, a primer design pipeline PeakPrime was developed to accurately capture these transcripts of interest. Specifically, the pipeline leverages prior bulk RNA-seq data to identify peak-coverage regions and design complementary primers in a coverage-informed manner.

The increased sensitivity provided by this approach can be applied to answering more specific research questions at lower cost, for example, regarding drug effects on known pathways or mapping of cell-type-specific expression patterns.

Application 3 – Spatial mutation profiling through full-length transcriptome sequencing: A novel workflow is currently being developed to enable somatic mutation calling within the spatial tissue context. A template-switch oligo (TSO) is included in the RT reaction mix and hybridizes to the non-templated cytosine nucleotides incorporated at the 3’ end of the first-strand cDNA. Following this, the reverse transcriptase continues cDNA synthesis using the TSO as template. As a result, a known sequence is added to the 3’ end of all cDNA molecules, allowing long-range PCR where only full-length cDNA molecules are amplified. In the final step of the library preparation, Oxford Nanopore Technologies (ONT) sequencing adapters are ligated to the ends of the cDNA molecules (Figure 5). The libraries are then sequenced on a MinION or PromethION device. The resulting data is processed through a customized version of the Nextflow pipeline to call somatic mutations (under development).

This transcriptome-based approach allows the identification of a broad range of somatic mutations in situ, without the need for prior knowledge of mutations from (bulk) sequencing data or mutation-specific panels, or for complex probe designs.

References

1. Behr J, Michel T, Giridhar M, Santhosh S, Das A, Sabzalipoor H, et al. An open-source advanced maskless synthesizer for light-directed chemical synthesis of large nucleic acid libraries and microarrays [Internet]. ChemRxiv; 2024 [cited 2025 Nov 25]. Available from: https://chemrxiv.org/engage/chemrxiv/article-details/65ba15e39138d231611ab534