Snap frozen biopsies represent a vast resource for clinical research as they have superior molecular integrity for performing the full array of downstream nucleic acid analysis compared to formalin-fixed tissue. However, obtaining 3D structural information from frozen tissue prior to analysis is challenging. In contrast, formalin-fixed tissue is the standard for obtaining 2D structural information. Formalin-fixed tissue is less compatible with nucleic acid extraction and analysis, and tissue distortion through histologic processing still occurs, which is particularly relevant to 3D analysis of complex, soft tissues, such as myocyte arrangement in the myocardium. A procedure that can combine imaging of (snap frozen) tissue, in its near native state, and be followed by nucleic acid analysis could therefore be clinically relevant for combined in-depth characterisation and targeting of disease.
In this study we demonstrate the feasibility of synchrotron-based cryo-X-PCI of frozen myocardial biopsies prepared under differing conditions. We describe, for the first time, the combination of 3D structural analysis from synchrotron-based cryo-X-PCI with assessment of nucleic acid analysis, focussing particularly on DNA and RNA integrity post-cryo-X-PCI.
As a marker of high-level morphologic preservation, we demonstrate that quantification of myocyte orientation (myomapping) in frozen myocardial biopsies is feasible through structure tensor analysis and that HA, IA and FA morphologic parameters could be assessed. Myomapping showed a gradual change in HA, from positive HA in the endocardium to negative HA in the epicardium and is consistent with previously reported studies which have quantified myocyte orientation in whole heart, in both animal models and human samples45,46,50. With biopsies, there is an additional layer of complexity in that the orientation of the biopsy is not always known. In order to mitigate this, datasets were carefully oriented with respect to epicardium and endocardium before performing myomapping. We also compared cryo-X-PCI with 2D traditional histology through cryo-sectioning and H&E staining and confirmed similar appearances before and after imaging. The increased spacing between myocyte aggregates found in our frozen biopsies closely resemble that reported in frozen tendon tissue imaged by cryogenic contrast-enhanced micro-CT39 suggesting that this feature is inherent to imaging tissues in their frozen state. Ice crystal artefacts were present in some biopsies, including in the control samples which is not ideal. Despite the presence of freezing artefacts, quantification of orientation of myocyte aggregates was still feasible in cryo-X-PCI samples as indicated by the myomapping results. Our study was designed to replicate the freezing process that is commonly used in clinical practice which typically involves rapid collection of biopsies in a cryogenic tube and snap freezing them in liquid nitrogen. However, other techniques for snap freezing samples could be investigated in the future.
Importantly, our study also shows that DNA and RNA can be successfully extracted from cardiac biopsies following cryo-X-PCI. Fresh frozen samples had optimal recovery of DNA as indicated by the DIN values greater than 6 for all extracted DNA samples followed by lower DIN values for biopsies preserved in 4% PFA and 10% formalin. Sequence-specific (Mybpc3 exon 28) PCR was successful for all fresh frozen and 4% PFA-fixed samples which all showed the 330 bp PCR product band. The 10% F samples had the lowest integrity as indicated by the low DIN values and only one faint Mybpc3 exon 28 PCR product band could be observed. Most likely this was due to extensive crosslinking from 10% formaldehyde fixation and is in agreement with our preliminary study25.
Low RIN values from TapeStation and Nanodrop measurements for the extracted RNA from biopsies were observed. Nonetheless, we optimised RT-PCR conditions to account for the small biopsy size and low RNA concentrations and showed that RNA could still be reverse transcribed to cDNA. The cDNA concentrations were very low and had to be re-amplified for the bands to show up clearly on the gel, but they were still viable for gene-specific PCR amplification as shown by the success of Mybpc3 exon 28 PCR.
We also compared DIN and RIN values for cryo-X-PCI biopsies with control biopsies (those that did not undergo cryo-X-PCI). There were no significant differences between the mean DIN and RIN values for fresh frozen samples that had cryo-X-PCI (mean DIN = 7.0, mean RIN = 4.4) compared to the fresh frozen samples that did not undergo cryo-X-PCI (mean DIN = 7.2, mean RIN = 4.0). We found no significant differences in nucleic acid quantity and quality, suggesting again that the small size of the biopsy was the limiting factor instead of cryo-X-PCI itself. Therefore, overall, if the goal is to have high nucleic acid integrity for downstream genomics and transcriptomics applications, such as sequencing, it would be best to avoid fixing samples, and simply prepare biopsies with optimal preservation of DNA and RNA, followed by freezing. On the other hand, if the aim is to have superior myocardial morphological detail but not perform subsequent omics, then fixing the biopsies in formaldehyde and processing them into FFPE blocks would be sufficient.
To conclude, cryo-X-PCI can provide non-destructive 3D assessment of myocardial morphology and can be combined with nucleic acid analysis of frozen biopsies. Synchrotron-based cryo-X-PCI does not appear to affect DNA or RNA integrity for downstream genomics and transcriptomics applications, and we recommend fresh frozen sample preparation for optimal results in both morphology and nucleic acid quality. Although we only tested cryo-X-PCI in mouse myocardial biopsies, this technique has the potential to be integrated into the clinical setting as a technique to image frozen biopsies from a range of diseases. We foresee clinical potential of cryo-X-PCI to examine further resources of snap frozen material (both cardiac and non-cardiac soft tissue) to provide 3D virtual histopathology and correlation between structural information and genetics.
Limitations
The current study has limitations. Snap freezing of samples was designed to replicate clinical practice but was not optimised to limit use of animal tissues and beamtime. Further studies are required to assess if this can be minimised. Our study focussed on DNA and RNA integrity as a marker of nucleic acid analysis preservation and protein integrity was not assessed. However, we have previously shown that X-PCI does not inherently affect protein epitopes as seen via immunohistochemistry25. Our study was limited by availability of synchrotron beamtime which is both competitive and expensive and our samples were limited to mouse myocardial biopsies which were imaged to show proof-of-concept before use of valuable human myocardial biopsies.