Aberrant Gene Expression Yet Undiminished Retinal Ganglion Cell Genesis in iPSC-Derived Models of Optic Nerve Hypoplasia

Major Points and Findings:

This was a laboratory study. No child received any cells and no treatment was tested. Researchers at Children’s Hospital Los Angeles took blood from three children with optic nerve hypoplasia (ONH), reprogrammed the blood cells into induced pluripotent stem cells (iPSCs), and grew them into small pieces of retina in a dish (“retinal organoids”). The aim was to watch how the retinal ganglion cells, the cells whose fibres form the optic nerve, develop when they carry a patient’s own genes. The work is the first human disease model of ONH. We found no other recent stem cell research paper that deals specifically with ONH.

Clinical evidence on stem cell treatment for ONH is very limited. The only controlled study of such treatment in children with ONH that we could locate (Fink, Garcia-Filion and Borchert, Journal of AAPOS, 2013) compared two treated children with two matched untreated children. According to its abstract, it found no evidence that the treatment improved visual acuity, and vision improved slightly in children whether or not they were treated. We could not obtain the full text of that paper, so we have not summarised it. The study summarised here does not alter that evidence. It helps explain what probably goes wrong in ONH, which is relevant when judging what any cell treatment could realistically do.

Aim:

To test a long-standing assumption, that ONH happens because the developing eye fails to make enough retinal ganglion cells (RGCs), and to look for genes whose activity differs in RGCs grown from children with ONH.

Methods:

Donors: The donors were three unrelated children with moderate to severe ONH in both eyes and three healthy, unrelated children with normal vision, all aged 8 to 17 years. The best visual acuities of the children with ONH were 8/200, light perception only, and 20/80. All three had growth hormone deficiency and developmental delay. In addition, two had hypothyroidism, one also had adrenal insufficiency and diabetes insipidus, two had a history of seizures, and one had cerebral palsy with cortical malformations on MRI. None had autism.

Cells: iPSC lines were made commercially from blood mononuclear cells using an episomal method and checked for pluripotency markers and a normal karyotype. Two to three separate clones were used per child (17 cell lines in total) so that a quirk of one line would not mislead.

Several published protocols for growing retina failed with these lines. The team modified an existing method by adding “dual SMAD” inhibition for the first 6 days and WNT inhibition for the first 9 days. After that, more than 90% of starting cell aggregates from every line formed retinal tissue.

Measurements:

• RGCs were identified with a BRN3 antibody and photoreceptors with CRX, and counted by flow cytometry between day 20 and day 90 of culture.
• For gene activity, BRN3-positive cells were sorted at day 40 to 43, when RGCs are most abundant. Three independent preparations per child were sequenced, each pooling at least 60 organoids.
• Differences were called with two statistical packages (DESeq2 and edgeR), and only genes flagged by both were treated as reliable. As a check, the six children were also shuffled into every possible random three-against-three grouping to see how many differences arise by chance.

Results:

RGC production was normal. RGCs peaked at 7.9% of cells on day 43 in ONH organoids and 7.1% on day 41 in controls. After the peak, RGC numbers fell at the same rate in both groups, as expected in a dish where the cells have no brain target. In the sorted samples, ONH preparations contained 12.8% (plus or minus 2.7%) RGCs against 11.7% (plus or minus 1.7%) in controls (n = 9 each, p = 0.30). The authors state that these experiments gave no support to the idea that an ONH genetic background reduces RGC production.

Gene activity did differ between the groups, although modestly:

• 70 genes differed between ONH and control RGCs by both methods at the 5% level (40 lower, 30 higher); 106 at the 10% level. Most changes were small.
• The ONH versus control grouping produced roughly 4.2 times the proportion of doubly confirmed genes seen in the random groupings, at least 2.4 standard deviations above the mean, so the signal is unlikely to be pure noise.
• No clear biological theme emerged. An apparent enrichment for cell adhesion was driven by one gene family (protocadherins) and vanished when that family was counted once.
• 24% of the differing genes (25 of 106) sat together in eight small chromosome regions, which points towards faults in shared regulatory “switches”.
• Candidates with the most disease-relevant features were RAPGEF4 and DMD. Many candidates act in axon growth, synapse formation or cell survival.
• None of the 111 candidate genes appeared among the 45 genes previously reported to cause an ONH-like picture. The authors point out that those earlier reports mostly describe unusual syndromes. Only 13 of 108 such patients (12%) had hypopituitarism, against about 80% in typical ONH.
• 22 of the candidates carried rare variants in a separate exome study of ONH families. Overlap with autism genes was no greater than chance.

Conclusions:

The authors conclude that ONH is probably not a failure to make RGCs. The cells appear to be born in normal numbers and then lost, most plausibly through excess cell death at the stage when their fibres should be growing to the brain and forming connections. They present their gene list as candidates for future work and do not claim that these genes cause ONH.

The limitations they list are the small sample of three patients and three controls, the study of only one early stage of RGC development, and the fact that organoids lack blood vessels, neighbouring tissues and a brain target. The real fault may also lie outside the RGCs, in supporting astrocytes or in the brain cells the optic nerve connects to.

Background Information:

ONH is described here as the leading congenital cause of permanent blindness in the United States and Europe, with an incidence of about 1 to 2 per 10,000. Pituitary or neurological problems, including autism in about 25%, accompany it in up to 80% of children. Consistent risk factors are young maternal age and first pregnancy. Familial cases are rare, and trio exome sequencing of 34 families found no strong single-gene cause.

The authors call ONH a “non-progressive congenital neurological defect.” The optic nerve fibres were lost before birth. Nothing in this paper suggests that infused stem cells could rebuild those fibres or reconnect the eye to the brain, and the paper does not test any treatment. For families considering cell therapy, its practical value lies in setting realistic expectations. The paper also shows that the hormonal and developmental problems are part of the same condition and need their own follow-up.