The effect of different scaffolds on the differentiation of a
single cell type: 2D and 3D scaffolds on neural progenitor cells
differentiation.
A longstanding difference between in vitro and in vivocell growth conditions is the dimensionality difference between the two.In vitro experiments have traditionally used two-dimensional cell
monolayers (discounting recent organoid developments), while cells in
live organisms develop in a three-dimensional setting. The
dimensionality of a cell’s microenvironment can have profound effects on
the function and differentiation of said cell (Willerth &
Sakiyama-Elbert, 2019). Transcriptomic responses can be documented using
high-throughput techniques like microarray and RNA sequencing.
In 2011, Lai et al used 2D and 3D porous polystyrene scaffolds to assess
and compare the effect of dimensionality on neural progenitor cell (NPC)
differentiation (Lai et al., 2011). NPCs were maintained in neural basal
media supplemented with penicillin, L-glutamine, recombinant human
leukemia inhibitory factor, basic fibroblast growth factor (bFGF) and
B-27 (Lai et al., 2011). Later on, bFGF was removed from above mentioned
factors to start differentiation process 25. NPCs
cultured in the presence of the 2D substrate showed significant
transcriptomic remodeling with 890 DEGs, with 439 of these genes
up-regulated (FC>2 and p-value <0.05). The
differentiated NPCs were clearly segregated from undifferentiated cells
(Fig. 1a). The most statistically significant upregulated genes were
those related to the central nervous system (CNS), including nervous
system development, axon guidance, and axonogenesis. (Fig. 1b, Table
S1). The same type of analysis with the same criteria (FC>2
and p-value <0.05) for cells grown in a 3D scaffold led
us to find 1082 DEGs, most of which were upregulated. The differentiated
cells in presence of the 3D substrate also were separated from
undifferentiated cells (Fig1a). The analysis of biological processes
similarly showed neuronal differentiation related terms as the most
significantly upregulated terms (Fig. 1c, Table S2). We then compared
the list of DEGs between each scaffold dimension to elucidate the
differences between cultures. The results of this comparison revealed a
high level of transcriptome similarity between culturing neural
progenitor cells in 2D and 3D scaffolds (Fig. 1d). 52% of DEGs in the
2D scaffold dataset were differentially expressed with the same
expression pattern in the presence of the 3D scaffold. As expected, 480
common DEGs found in both 2D and 3D scaffold data sets were involved in
nervous system related biological processes and functions, like neuron
differentiation (FDR q-val = 0.0046) (Fig. 1e and 1f, Table S3).
Despite similarities, the expression of many genes was different between
2D and 3D scaffolds, with many more upregulated genes in the 3D
culturing condition. Because tissue cells grow in a 3D structure, we
hypothesized that neuronal cells differentiated on the 3D scaffold would
be more similar to their in vivo counterparts. Therefore, we
expected that these up-regulated genes in the 3D culturing condition
would be associated with nervous system development. We performed
ontology analysis for the most upregulated genes that were exclusive to
the 2D biomaterial and we found that the most affected terms were
related to nervous system development and functions, including synapse
assembly, chemical synaptic transmission, and anterograde trans-synaptic
signaling (Fig. 1g). We performed the same analysis for the most
upregulated genes exclusive to the 3D biomaterial, and unexpectedly
found a high number of terms associated with inflammatory response (Fig.
1h). The 2D most upregulated terms lacked any that were related to
immune function. Overall, the transcriptomic profile of neural
progenitor cells cultured in 2D and 3D scaffolds revealed that both
induce common gene expression related to nervous system development. In
addition to these genes, however, the 3D scaffold induces more
biological processes that are not related to nervous system development,
including inflammatory responses. Therefore, in contrast to our
hypothesis, it seems culturing of NPCs on 2D porous polystyrene produce
neurons with more mature cell characteristics according to their
transcriptome profile, however, further experimental studies are needed
to confirm our result in more detail.
Dimensionality is an important contributor to stem cell differentiation
and relatedness to in vivo counterparts. However, there are other
modalities of biomaterials that can influence the success of
differentiation. Here, we investigate the role of degradable vs
non-degradable biomaterials as one influencer. Gjorevski et al. used
either degradable or non-degradable enzymatically crosslinked
polyethylene glycol (PEG) during epithelial organoid formation from
intestinal stem cells (ISCs) (Gjorevski et al., 2016). Remodeling is
important for repair and morphogenesis, so they hypothesized that the
degradable matrix would improve organoid formation. Interestingly, their
results indicated that the use of degradable PEG resulted in an abnormal
shape of colonies and fewer organoids formed. The transcriptomic profile
of ISCs demonstrated that genes related to stress and inflammation were
significantly enriched in ISCs cultured in the presence of degradable
PEG compared to non-degradable PEG. We re-analyzed their RNA-seq data to
further investigate the possible mechanism of the poor organoid
formation from ISCs cultured in the degradable PEG.
Comparing ISC differentiation in the presence of degradable and
non-degradable PEGs showed 170 DEGs (FC>2 andp-value <0.05). Among these, 115 genes were upregulated
in degradable PEGs, and the majority were associated with an
inflammatory response and a positive regulation of leukocyte chemotaxis
in comparison with non-degradable PEGs (Fig. 2a). To find the main
regulators of the inflammatory response related genes in ISCs
differentiation, we generated a gene regulatory network containing 33
genes (nodes) and 53 interactions between them (edges) (Fig. 2b).
Analysis of this constructed gene regulatory network revealed ATF3 and
KLF6 as the main regulators of the inflammatory response (Fig. 2c). In
addition, we constructed another network using a list of differentially
expressed TFs, called protein-protein interaction network (Fig. 2d). In
the constructed protein-protein interaction network, ATF3 and KLF6 are
found to interact with NFKBIA that subsequently interacts with TNF.
Interestingly, it was previously shown that KLF6 expression dramatically
increased in both intestinal tissue and myeloid cells in inflammatory
bowel disease (Goodman et al., 2016). So, it seems that recruitment of
transcription factors ATF3 and KLF6 could be a possible mechanism for
activation inflammatory cascaded in abnormal organoid formation.