Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 6, 8896

Generation of Functional Hippocampal Neurons From Self-Organizing Human Embryonic Stem Cell-Derived Dorsomedial Telencephalic Tissue

Affiliations

Generation of Functional Hippocampal Neurons From Self-Organizing Human Embryonic Stem Cell-Derived Dorsomedial Telencephalic Tissue

Hideya Sakaguchi et al. Nat Commun.

Abstract

The developing dorsomedial telencephalon includes the medial pallium, which goes on to form the hippocampus. Generating a reliable source of human hippocampal tissue is an important step for cell-based research into hippocampus-related diseases. Here we show the generation of functional hippocampal granule- and pyramidal-like neurons from self-organizing dorsomedial telencephalic tissue using human embryonic stem cells (hESCs). First, we develop a hESC culture method that utilizes bone morphogenetic protein (BMP) and Wnt signalling to induce choroid plexus, the most dorsomedial portion of the telencephalon. Then, we find that titrating BMP and Wnt exposure allowed the self-organization of medial pallium tissues. Following long-term dissociation culture, these dorsomedial telencephalic tissues give rise to Zbtb20(+)/Prox1(+) granule neurons and Zbtb20(+)/KA1(+) pyramidal neurons, both of which were electrically functional with network formation. Thus, we have developed an in vitro model that recapitulates human hippocampus development, allowing the generation of functional hippocampal granule- and pyramidal-like neurons.

Figures

Figure 1
Figure 1. Generation of choroid plexus-like tissues with pleated structure from hESCs.
(a) Schematic of examined conditions to induce choroid plexus tissues. (b) Immunostaining of neocortical structure induced from hESCs on day 35. (c) Comparison of aggregate formation from Foxg1::Venus knock-in hESCs on day 35. CHIR (GSK3 inhibitor) plus BMP4 treatment (condition 2) induced thin epithelia with many folds with significant attenuation of Foxg1::Venus expression. (d,e) qPCR for genes expressed in dorsomedial telencephalon (***P<0.001). (d) foxg1 significantly attenuated in condition 2 (n=3, unpaired t-test). (e) ttr significantly increased in condition 2 (n=3, unpaired t-test with Welch's correction). (f) Bright-field view of one aggregate cultured in condition 2 on day 42. (gk) The induction of choroid plexus tissues with pleated structure from hESCs. The pleated epithelia are Lmx1a+ (g,h), Otx2+ (i) and TTR+ (h). TTR mainly stained apically (h), and Aquaporin-1 stained mainly in apical parts, with less staining in basal parts (j). ZO-1 stained the surface of epithelia (k). Scale bars, 50 μm (b; 500 μm (c, f); 200 μm (g); 20 μm (hk). Bars in graph, s.e.m. Nuclear counter staining (blue), 4,6-diamidino-2-phenylindole (DAPI).
Figure 2
Figure 2. Transient exposure of dorsalizing factors can induce patterned dorsomedial telencephalic tissues.
(a) Schematic of mouse medial pallium and neighbouring tissues at E12.5–13.5. (b) Schematic of condition to induce medial pallium tissues. (c) Bright-field view of aggregates cultured in condition 3 on day 36. (d,e) By shortening the period (day 18–21), the aggregates expressed Foxg1::Venus with patterning into Foxg1::Venus+ (arrow) and Venus (arrowhead) NE domains. (f) Histogram of percentage of Foxg1::Venus protrusion and Foxg1::Venus+ main body induction. Patterning of Foxg1::Venus+/Venus NE domains were induced in around 70–80% of aggregates (bars in graph, s.e.m.). (gl) The Venus NE domains of aggregates (g) showed TTR+/Lmx1a+ in distal parts (h,i) and Lmx1a+/TTR in proximal parts (jl, box in h). (mo) In hESC-derived NE, Foxg1::Venus+ main bodies expressed Lef1 (m) and Lhx2 (n). (o) On the basal side, the thick cell layer of the aggregate expressed Tbr1 and Lhx5. (p) Foxg1::Venus and Lef1 co-expressed in continuous NE (p,q, box in h), suggesting medial pallium-like NE generation. (r) Schematic of hESC-derived dorsomedial telencephalic tissues. Choroid plexus-, hem- and medial pallium-like regions were continuously generated as seen in vivo. (s) Schematic of method to examine Wnts/BMPs expression in Venus protrusions. The aggregates were cut into Venus protrusions and Venus+ main bodies at day 35–40, and wnts/bmps expression were examined using RT–qPCR. (t) qPCR for genes expressed in Venus protrusion versus Venus+ main body (**P<0.01; ***P<0.001, n=3, unpaired t-test). foxg1 was significantly attenuated in Venus protrusions, and a significant increase in lmx1a/ttr mRNA expression in Venus protrusions was confirmed. wnt 2b/3a/5a and bmp 4/6 significantly increased in Venus protrusions. (u) Schematic summary of conditions examined in Figs 1 and 2. Scale bars, 500 μm (ce); 200 μm (gp). Bars in graph, s.e.m. Nuclear counter staining (blue), 4,6-diamidino-2-phenylindole (DAPI).
Figure 3
Figure 3. Expression of hippocampus marker Zbtb20 in medial pallium-like tissue.
Under the optimized conditions, the aggregate was cultured with less formation of rosette-like NE (a,b). (eh is box in c) Immunostaining on day 61 showed that the medial pallium-like continuous NE, positive for Foxg1::Venus (e) and Lef1 (f) were formed adjacent to the Lmx1a+ choroid plexus-like domain (c,d). (g) The more basal layer of the Lef1+ NE was positive for Nrp2. (h) Zbtb20 was expressed in the Lef1+ NE. (i) Zbtb20 mRNA expression at day 73 showed a significant increase compared with day 18. *P<0.05, n=3, Unpaired t-test with Welch's correction. (j) The medial pallium-like continuous NE, positive for Foxg1::Venus and Lef1 were clearly formed adjacent to the Lmx1a+ choroid plexus-like domain even at day 75. (k,l) Zbtb20 expression in cells beneath the Lef1+ NE was clearly seen at day 75 (l is enlarged box in k). Scale bars, 300 μm (a,b); 200 μm (ch); 100 μm (j,k); 50 μm (l). Bar in graph, s.e.m. Nuclear counter staining (blue), 4,6-diamidino-2-phenylindole (DAPI).
Figure 4
Figure 4. Hippocampal marker-positive pyramidal and granule-like neurons were generated by long-term dissociation culture.
(ac) Immunostaining of dissociated cells at day 197. Foxg1::Venus+ neurons (a) were connected to one another by MAP2+ neurites (b,c). GFAP+ glial-like cells were also detected (d). (e,f) KA1+/Foxg1::Venus+ neurons (CA type) tended to have large cell bodies; in contrast, Prox1+/Foxg1::Venus+ neurons (DG type) mostly had small somata (f is enlarged box in e). Zbtb20 was co-expressed with both KA1+ and Prox1+ neurons (g,h). (i) The percentage of Zbtb20+ cells was about 75% (means: 76.5%, s.e.m.: 1.65) and the percentages of KA1+ (means: 34.4%, s.e.m.: 1.83) and Prox1+ (means: 33.8%, s.e.m.: 2.33) were both about 34%. Countered neurons: Zbtb20+ (n=2764), KA1+ (n=147), Prox1+ (n=188). (j) Cell body size was significantly larger in KA1+ cells. ***P<0.001, paired t-test. Countered neurons: KA1+ (n=56), Prox1+ (n=55). (k,l) CaMKII was expressed in KA1+ (k) and Prox1+ (l) neurons. Scale bars, 200 μm (ac); 50 μm (d); 20 μm (e,gh); 10 μm (f,k,l). Bars in graph, s.e.m. Nuclear counter staining (blue), 4,6-diamidino-2-phenylindole (DAPI).
Figure 5
Figure 5. Functional analysis of dissociated hippocampal-like neurons using electrophysiology and calcium imaging.
(a) A representative cell image for patch-clamp recording. Many neurons showed voltage-dependent Na–K current (b, n=30), action potential following injection of depolarizing currents (c, n=18) and spontaneous excitatory postsynaptic currents (d, n=9, 2.33±0.75 Hz), at day 136 (dissociated at day 83). The percentage of neurons recorded showing Na–K currents, action potentials and synaptic responses were 100% (30 out of 30), 72% (18 out of 25) and 75% (9 out of 12), respectively. (eh) The data set of calcium imaging. (e) Representative image of active neurons (day 143, 8 weeks after dissociation), and their firing pattern shown by trace image of calcium response (f). (g) Pharmacological perturbation by TTX (**P<0.01, n=3, unpaired t-test with Welch's correction). (h) Time-dependent promotion of neuronal activity. Percentage of active neurons significantly increased at 8 weeks after dissociation (***P<0.001, n=3, one-way analysis of variance). (i) Synchronization analysis by cross-correlations of 100 neurons at 8 and 4 weeks after dissociation. Colours from red to blue indicate their cross-correlations from high (synchronized activity) to low. Cross-correlation was higher at 8 weeks after dissociation than 4 weeks. The data are representative of three independent experiments. (j) Histogram of correlation coefficients indicated a strong correlation at 8 weeks after dissociation. The data are representative of three independent experiments. (k) Histogram of average correlation coefficients of one hundred neurons (*P<0.05, n=3, unpaired t-test). Scale bars, 20 μm (a); 100 μm (e). Bars in graph, s.e.m.

Similar articles

See all similar articles

Cited by 54 PubMed Central articles

See all "Cited by" articles

References

    1. Wilson S. W. & Rubenstein J. L. Induction and dorsoventral patterning of the telencephalon. Neuron 28, 641–651 (2000) . - PubMed
    1. Grove E. A. & Tole S. Patterning events and specification signals in the developing hippocampus. Cereb. Cortex 9, 551–561 (1999) . - PubMed
    1. Lee S. M., Tole S., Grove E. A. & McMahon A. P. A local Wnt-3a signal is required for development of the mammalian hippocampus. Development 127, 457–467 (2000) . - PubMed
    1. Emerich D. F., Skinner S. J., Borlongan C. V., Vasconcellos A. V. & Thanos C. G. The choroid plexus in the rise, fall and repair of the brain. Bioessays 27, 262–274 (2005) . - PubMed
    1. Grove E. A., Tole S., Limon J., Yip L. & Ragsdale C. W. The hem of the embryonic cerebral cortex is defined by the expression of multiple Wnt genes and is compromised in Gli3-deficient mice. Development 125, 2315–2325 (1998) . - PubMed

Publication types

MeSH terms

LinkOut - more resources

Feedback