Although 1000–1500 lung transplants per year are performed in the United States, a significant shortage of suitable donor lungs and the drawbacks of lung transplantation, including an approximate 50% 5-year mortality and need for lifelong immunosuppression, demonstrate a critical need for new approaches.25
Unlike cadaveric organs such as kidneys, which have been successfully utilized for many years, transplantation of cadaveric lungs has not yet been feasible. Ex vivo
re-cellularization of de-cellularized three-dimensional whole lung scaffolds to produce functional lung tissue suitable for transplantation has recently been investigated as an alternative approach that would allow use of cadaveric lungs.10–14
Initial promising results have been obtained in rodent models and de-cellularized rat lungs, re-cellularized with a mixture of endothelial cells, fetal lung homogenates, and tumorigenic lung epithelial cells, have been implanted and maintained for up to 2 weeks.12,13,26
However, different techniques have been utilized to de-cellularize whole lungs including physical methods (multiple freeze thaws) as well as use of different detergents.10–14
We have previously provided detailed characterization of de-cellularized mouse lungs14
utilizing a Triton/SDC-based method originally described by Lwebuga-Mukasa et al
and modified by Price et al
Other groups have described SDS or CHAPS-based protocols.11–13
It is not clear which, if any, of these might produce an optimal de-cellularized lung scaffold. Further it is not clear what an “optimal” de-cellularized lung scaffold will be.2
Considerations include appropriate maintenance of gross and microscopic structure, maintenance of key ECM proteins, and the ability to support appropriate binding and growth of a range of cells that might potentially be utilized for re-cellularization.2,5–7
Whether different residual contents of intracellular and other proteins or whether differences in mechanical properties of the de-cellularized lungs will affect re-cellularization and generation of functional lung tissue has not yet been clarified. Further, the potential immunogenicity of the de-cellularized scaffolds is a critical consideration for clinical transplantation of re-cellularized human lung scaffolds.
To address whether re-cellularization would be affected by different de-cellularization approaches, we compared three different detergent-based lung de-cellularization protocols and found significant differences in histologic appearance, gelatinase activation, content and distribution of ECM proteins, and a range of intracellular proteins. To best assess the specific effects of the different detergents utilized, the SDS- and CHAPS-based de-cellularization protocols were otherwise standardized to the timeline of the Triton/SDC approach we have previously utilized based on the method of Price et al.10,14
Additional steps in the de-cellularization approaches, including use of hypertonic saline, EDTA, DNAse, and FBS, were utilized as had been done in previous reports of lung de-cellularization.10–14
Notably, 0.1% Triton alone resulted in increased proteolytic activity at a much lower concentration than what is normally used to activate MMPs for gelatin zymography (e.g., 2.5% Triton-X 100 at 37°C).8
These results highlight that the different protocols will result in different de-cellularized lungs and notably that detergents will differentially activate proteolytic activities early during the de-cellularization process that may critically affect the remaining ECM matrix. A recent comparison of SDS- versus CHAPS-based lung de-cellularization protocols similarly found differential retention of ECM proteins (collagen, elastin, and glycosaminoglycans) with more marked loss of collagen and elastin in SDS de-cellularized lungs.27
Notably, mechanical testing of lung strips demonstrated better retention of tensile strength in lungs de-cellularized with CHAPS as compared with SDS, possibly related to better retention of collagen and elastin. In contrast, we found that airways resistance and whole lung elastance were relatively similar with the use of CHAPS and SDS although elastance was decreased with use of Triton/SDC. The consistently lower than physiologic levels of elastance () and negative values for resistance () in the CHAPS and SDS lungs were likely due to a significantly greater degree of leak during the delivery of oscillatory perturbations rather than being obviously attributable to differences in any given ECM proteins. In contrast, despite a few lungs in the Triton/SDC group exhibiting abnormally low elastance and resistance values, the majority of the Triton/SDC lungs exhibited elevated elastance values and above-zero resistance values that were reproducible over time, likely representing a loss of surfactant function in the absence of significant leak. Likewise, values for the COD (“goodness of fit”), derived when fitting the Zrs raw data to the viscoelastic lung model, were significantly higher in the Triton lungs (), suggesting much more linear pressure response and significantly less leak in these lungs. Taken together, these data suggest that use of Triton/SDC as utilized is significantly less disruptive to the integrity of the de-cellularized matrix, and thus more conducive to following lung mechanical changes over time, both during de-cellularization and re-cellularization.
However, despite differences in ECM and other protein content and differences in whole lung elastance, initial binding and short-term (2 weeks) proliferation of two different cell types, a stromal progenitor cell (MSC) and a mouse lung epithelial cell line (C10), appeared comparable when cells were intratracheally inoculated into lungs de-cellularized with the different protocols and lung slices grown at normoxia under static (i.e., nonventilated or perfused) conditions. We have previously found that MSCs inoculated into and similarly cultured in Triton/SDC-derived de-cellularized lungs thrived for at least up to 1 month.14
These results suggest that despite significant differences in the composition of lungs produced with the different detergent-based protocols, inoculated cells are able to recognize appropriate ligands that will allow binding and initial proliferation. Notably the C10 cells appear to comparably re-epithelialize medium-sized and small airways in each of the protocols utilized.
One additional point of comparison is that use of allogeneic re-cellularization with MSCs (C57Bl/6 MSCs inoculated into de-cellularized BALB/c mouse lungs) in the current studies resulted in similar initial patterns of cell localization as compared with the syngeneic re-cellularization (C57Bl/6 MSCs inoculated into de-cellularized C57Bl/6 mouse lungs) utilized in our previous studies.14
C10 cells are of BALB/c origin21
and we have not yet compared syngeneic versus allogeneic inoculation of C10 cells into de-cellularized C57Bl/6 lungs. Further comparison of syngeneic versus allogeneic re-cellularization approaches may provide clues about the critical factors in the de-cellularized lungs that direct re-cellularization as well as the degree of acceptable genetic mismatch compatibility.
These results suggest that development of techniques for production of an “optimum” de-cellularized scaffold is more dependent on re-cellularization potential rather than de-cellularization methodology. However, despite similarities in initial re-cellularization between the three detergent-based protocols assessed, there are still many questions to be answered, including long-term growth and differentiation of inoculated cells, reconstitution of lung functions including gas exchange and appropriate lung mechanics, and also appropriate antiinflammatory and immune functions of the lung.28
Further, there are other permutations yet to be clarified for de-cellularization approaches as some published protocols have utilized only vascular perfusion with de-cellularizing agents and also shorter incubation times with the different detergents.12,13,26,27
As such, the optimal de-cellularization protocol is still to be determined.