210
establish a structural or molecular phenotype. Freedman et al. ( 2013 ) identified
lower ciliary expression of polycystin-2 in embryoid bodies and differentiated hepa-
toblasts in patients with PKD1 mutations, which corrected with transfected overex-
pression of wild-type PKD1. The same group later introduced biallelic PKD1 or
PKD2 mutations (a genotype associated with embryonic lethality) into a normal
hPSC line using CRISPR/Cas9 and differentiated them according to their epiblast
spheroid sandwich culture protocol, observing the development of sizeable, translu-
cent cysts arising from tubules expressing proximal tubular markers in 6% of organ-
oids (Freedman et al. 2015 ). This report represents a reasonable proof of principle,
but to date there exists no evidence of a patient-derived iPSC kidney organoid dis-
ease model.
While disease modelling using this approach seems promising, there are limita-
tions to this approach. Due to immaturity and lack of a circulation, organoids cannot
model inflammatory disease. As they do not have an accurate overall organ struc-
ture, they are also limited in their ability to model congenital structural renal dis-
eases such as hypoplasia or duplex kidney. The avascular glomeruli lack a
well-developed GBM, currently limiting the study of GBM disorders (e.g. Alport
syndrome). Recent work demonstrating the incorporation of murine vasculature
into subcapsular transplants of human iPSC-derived glomeruli indicates this imped-
iment is likely to be overcome for disease modelling purposes (Sharmin et al. 2016 ).
Furthermore, current organoid nephrons are immature with their capacity to mature
further using existing culture methods unclear (Little and Takasato 2015 ; Takasato
et al. 2015 ). Thus, organoids have not yet been proven suitable to study diseases
without a foetal or infantile phenotype. Conversely, genomic examination of devel-
oping renal tissue prior to the development of a phenotype may control for differen-
tial gene expression occurring secondary to phenotype rather than the result of the
primary genomic imbalance. In contrast, modelling a potential disease causing gene
in mouse has the advantage that the organ formed is anatomically correct, can
mature into a functional organ and can also reflect the effect of systemic changes on
the organ as a result of the mutation. However, the mouse does not model the poten-
tially mutant allele in the context of the patient’s genome. Balancing these advan-
tages and limitations, it is likely that both mouse and organoid techniques will be
necessary to facilitate informative disease modelling research until such time that
differentiation protocols are optimised to generate more mature and complex iPSC-
derived tissues.
11.5 Longer-Term Directions and Challenges for Directed
Differentiation to the Kidney
The initial isolation of human pluripotent stem cells was seen as opening up the
prospect of regenerative medicine. The directed differentiation of human pluripo-
tent stem cells to endpoints such as dopaminergic neuron is now providing
M.H. Little et al.