PREreview of Decellularisation and characterisation of porcine pleura for lung tissue engineering

Summary

This paper aims to explore the decellularization of porcine pleural membranes (PPM) by developing a reproducible decellularization protocol and assessing its efficiency in successfully removing cells while preserving the structural and functional characteristics of the pleural extracellular matrix. 

The researchers describe a two-step physico-chemical method to decellularize the PPM and then characterize the resulting decellularized PPM (dPPM) using various techniques. Histology, quantitative assays, mechanical testing, and sterility evaluation were performed to assess the properties of dPPM. They indicated that the decellularization process effectively removed the cells from the PPM and that the general mechanical integrity and functionality of the extracellular matrix (ECM) in dPPM remained relatively unaltered, even with detectable disruptions to its biochemical composition and microstructure. And lastly, they find that human mesothelial cells could attach and proliferate successfully on dPPM.

Given this, the authors hypothesize that transplantable acellular pleural scaffolds obtained by this protocol could potentially be used in tissue-specific applications such as pleural repair. This could be used to reduce the incidence, complications and general severity of conditions like persistent air leaks.

Major considerations:

1. In these samples, pleural mesothelial cells (PMCs) are active sources of a variety of connective tissue macromolecules such as collagen¹. PMCs also produce a significant amount of hyaluronic acid which constitutes pericellular matrices coating their microvilli². Hyaluronic acid, although not an sGAG, produces a similar absorption pattern to the sGAG-DMMB complex, being a known contaminant in DMMB assays³. Nucleic acids such as DNA and RNA can also have the same effect³.   This provides possible quantitative confounding factors for the bioquantitative assays as the detected decreases in the measured molecules may not entirely reflect changes in the molecular composition of the ECM. Several methods can be used to attenuate these risks³, but in the case of a complete study such as this, a statement about possible measurement artifacts would be recommended.

2. When examining the mean differences between two groups using unpaired, two-tailed student t-tests, an experimental n of 3 or 5 grants a very low statistical power, even for large effect sizes (for an alpha error probability = 0.05, we can only achieve approximately 12% power for n = 3, and 20% for n = 5, even for a large effect size d of 0.8 – post hoc parameters calculated in GPower v3.1.9.4). Many experiments in this study achieved statistical significance despite this limitation, possibly because of very large effect sizes. But nonetheless, the low sensitivity does not provide us with sufficient confidence to discard the alternative hypothesis in those that did not. So the conclusion that Youngs modulus and ultimate tensile strength of dPPM were not significantly different from native PPM is not very solid. Again, this could be addressed in a statement about the limitations of the current study.

3. Although the curves in Figure 7B present clear evidence of negligible cytotoxicity, the related statistical analysis provided for the assay does not contribute directly to assessing the relevant hypothesis. This sequential comparison of data points only serves to show that the alive and dead cell count curves are different, while a more worthwhile comparison could be done between the respective alive and dead curves of dPPM samples and controls. This could be done using time-series classification models to obtain a single significance value for the similarity of the curves⁴ or simply by applying sequential t-tests as in the present study. The latter approach would probably be most adequate due to the small amount of time points evaluated. Moreover, polynomial regressions could be optionally used to assess quantitative differences in biocompatibility. Additionally, a correction such as Bonferroni could be applied (both to the approach suggested here and to the study as is) to counteract the multiple comparisons problem inherent to sequential testing such as this.

Other comments:

1. When reporting t-tests, it’s important to include whether they were one or two-tailed. It can be reasonably assumed that all performed in this study were two-tailed by the nature of the hypothesis being tested, but even then reporting it is good practice.

2. There’s no legend for the statistical significance level “***” in the “statistical analysis” section. When creating a reference for annotations across your paper, you should include all existing categories.

3. There are no cited sources for the following assertion: “Although known for providing intrinsic compressive strength, reports suggest reductions in sGAG do not have an adverse effect on the gross mechanical characteristics of ECM”. It’s an important element in the author’s argumentation that relies on external results, so citing these reports would be recommended.

4. By examining the sample images in Figure 6A, it would seem that there are some inconsistencies with the infection logs in Figure 6B. This could be due to the low resolution of the images, but they are seemingly supported by the written description in that section. Furthermore, the unmatched record dates between Figure 6A (including images from days 0, 3 and 7) and 6B (including records for days 1, 5 and 7) could be a source of unnecessary confusion. If possible, either reporting the data from the same 3 days or having more data points in the logs, but still including the 3 same days shown in the images, could solve this.

5. In Figure 8, there should be an indication of the day on which the unseeded dPPM negative control pictures were taken. Logically, it should also be from day 15, but including it explicitly would improve clarity.

6. There’s no unit of measurement for the amount of glycine used in the sGAG assay. 

References

1. Rennard SI, Jaurand MC, Bignon J, et al. Role of Pleural Mesothelial Cells in the Production of the Submesothelial Connective Tissue Matrix of Lung. Am Rev Respir Dis. 1984;130(2):267-274. doi:10.1164/arrd.1984.130.2.267

2. Mutsaers SE. The mesothelial cell. The International Journal of Biochemistry & Cell Biology. 2004;36(1):9-16. doi:10.1016/s1357-2725(03)00242-5

3. Zheng C, Levenston M. Fact versus artifact: Avoiding erroneous estimates of sulfated glycosaminoglycan content using the dimethylmethylene blue colorimetric assay for tissue-engineered constructs. eCM. 2015;29:224-236. doi:10.22203/ecm.v029a17

4. Serrà J, Arcos JLl. An empirical evaluation of similarity measures for time series classification. Knowledge-Based Systems. 2014;67:305-314. doi:10.1016/j.knosys.2014.04.035

Competing interests

The author declares that they have no competing interests.