202
facilitate spontaneous self-organisation and potentially improve morphogenesis
in vitro. In a protocol using fully defined APEL media and an initial 7 days of PS
and IM induction, cultures were dissociated to single cells, aggregated and cultured
using floating filter cultures as is standard for the ex vivo culture of mouse embry-
onic kidney (Saxén and Lehtonen 1987 ). This approach resulted in the formation of
complex multicellular kidney organoids that, after 11 days in organoid culture, con-
tain not only patterning and segmenting nephrons but also a contiguous collecting
duct epithelium. Large numbers of nephrons are present within each organoid with
evidence of progression through S-shaped body to capillary loop stages of pattern-
ing, including the eventual formation of UMOD+ loops of Henle. Glomeruli are
evident and contain early podocytes with tightly interdigitated foot processes. While
the majority of these represent early nephrons and hence are avascular, the organ-
oids also contain an extensive network of endothelial cells, themselves adjacent to
early PDGFRA+ pericytes and mesangial cells. The endothelium is positive for
SOX17, KDR and PECAM and in some instances can be seen within glomeruli,
suggesting the initial formation of glomerular capillaries. In addition, the epithelial
structures, nephrons and collecting ducts, are surrounded by a stromal population
expressing MEIS1+, a reported marker of the renal interstitium (Brunskill and Potter
2012 ). The presence of both an endothelial population and a renal interstitial popu-
lation is consistent with these cell types arising from the MM or the IM, although
they may also arise from a closely associated paraxial or lateral plate population.
The presence of more than eight identifiable kidney cell types, arranged in an
appropriate morphological arrangement, strongly suggests the successful recreation
of a developing human kidney from pluripotent stem cells (Fig. 11.3). However, the
identity of these cell types has been based on what we know from mouse. Indeed,
there are few genes that are specific to only one organ and many genes expressed
during metanephros and expressed in mesonephros and gonad, as expected based on
their common IM origin. To address whether these organoids represented the kid-
ney, RNAseq expression profiling was performed in total iPSC-derived kidney
organoids and compared in an unbiased fashion to the transcriptional profiles of >20
human foetal tissue types using a previously defined algorithm, KeyGenes (Roost
et al. 2015 ). Using this unbiased analysis, hPSC-derived organoids after 11 days of
culture most closely matched trimester 1 human kidney. Of note, early after generat-
ing the aggregate, these cultures showed a closer alignment to indeterminate gonad,
the only other derivative of IM analysed. An analysis of the temporal gene expression
changes across organoid culture shows a significant increase in markers of nephron-
specific cell types, including podocytes and proximal tubule cells, with time.
Electron microscopic analyses showed distinct evidence for proximal and distal
tubules and avascular glomeruli. Initially, LTL+ proximal tubular segments were
EpCAM+ but CDH1-, as has been reported during development (Combes et al.
2015 ). However, with time, LTL + CDH1+ epithelial segments formed. To investi-
gate whether this proximal tubular epithelium was sufficiently mature to selectively
respond to nephrotoxic injury, organoids were treated with cisplatin to induce apop-
tosis. Maturing proximal tubular segments showed a specific and dose-responsive
caspase 3 activation, suggesting that they were affected by this nephrotoxicant
(Takasato et al. 2015 ).
M.H. Little et al.