A transposase-derived gene required for human brain development

Background:

Transposase Co-option

Transposable elements (TEs) cause disease by disrupting host genome structure
(e.g., interrupting or removing host exons, recombining with each other to cause large deletion/duplication events, and misregulating host gene expression) (2). Host defenses chase and repress TEs in an arms race (3, 4). One interesting end for this arms race is co-option: where the host can use this reorganization and introduction of new regulatory (5) or catalytic (6) elements to diversify function of genes and tissues. PiggyBac Transposable Element Derived 5 (PGBD5) is a DNA transposase expressed in most childhood solid tumors where it mediates oncogenic DNA rearrangements (7). Despite this, it is the most conserved domesticated DNA transposase in vertebrates even though it is predominantly expressed in the nervous system, specifically the brain (8). We often assume conservation implies function. Does PGBD5 have a beneficial function in the developing vertebrate brain?

Developmental Apoptosis + DNA Rearrangements

The development of the immune and nervous systems seems uniquely dependent on DNA damage and two of its outcomes: cell apoptosis and DNA rearrangements (1). The potential function of large-scale apoptosis in the developing nervous system is less clear, but one popular hypothesis is that apoptosis maintains synaptic plasticity – and so it is important for memory, learning, and response to injury (9). While genetic mosaicism has been documented, evidence for DNA rearrangements in vertebrate brain development had not been found prior to this preprint (1). Over 50 years ago, it was proposed that protocadherin gene diversification could organize and structure the vertebrate brain (10). In recent years, DNA breaks have been found in many neuronal genes (9). This preprint shows evidence for locus-specific rearrangements in multiple mouse brain tissues which are dependent on the domesticated transposase Pgbd5. Future work might confirm whether factors like Pgbd5 do diversify cell adhesion receptors or other regulators of synaptic connections which organize the vertebrate brain.

 

Zapater and team analyze the function of PGBD5 in neurodevelopment through:

1) characterizing the effect of Pgbd5’s nuclease activity on brain morphology and behavior in mouse and human

2) confirming whether PGBD5-dependent DNA breaks are repaired and make rearrangements which are sustained to adulthood

 

Key Findings:

1. Reducing the nuclease activity of Pgbd5 impairs neurodevelopment

Zapater and team use three types of mutations in Pgbd5 to study the effect of the nuclease activity of Pgbd5 on neurodevelopment. First, they found two families through GeneMatcher which carry mutations in PGBD5 near a conserved aspartate triad thought to be necessary for the transposase domain’s biochemical activity. These families had epilepsy, non-progressive motor dysfunction, and impaired expressive language acquisition. Their symptoms paired with thinned corpus collosum and decreased cerebellar volume. Next, they compared phenotypes observed in patients to their own Pgbd5^-/-  mice (lacking exon 4 of PGBD5). Pgbd5^-/-  mice similar to the patients had tonic-clonic seizures and reduced cortical and cerebellar volumes. In behavioral tests, Pgbd5^-/-  mice showed increased locomotor activity, reduced anxiety-like behavior, and impaired motor learning. Comparing mouse behavior to patient symptoms, it seems that PGBD5 has some effect on learning and memory which may be linked to plasticity in development. Last, to confirm if these defects were due to the nuclease activity of Pgbd5, the authors use a knockin mouse model Pgbd5^D262A which removes one of the conserved aspartate residues and should reduce Pgbd5 nuclease activity. Strikingly, knockin mice also had tonic-clonic seizures. Based on these three mutations, Pgbd5’s nuclease activity seems necessary for healthy neurodevelopment.

Fig 1. Mutations to PGBD5’s transposase domain cause abnormal brain development in mice

A. Preprint Figure 1H: Schematics of the engineered Pgbd5 locus and breeding strategy to generate Pgbd5^-/- mice. B. Preprint Figure 1J: Brain weights of 30-day old WT, Heterozygous, and Pgbd5^-/- mice, showing significant reduction in brain weights in male and female mice (t-test p = 0.016 and ** p = 3.4E-3, respectively). Pgbd5wt/wt n=5 female and males, Pgbd5wt/- n=12 females and n=7 males, and Pgbd5-/- n=8 females and n=9 males. C. Preprint Figure 2J: Pgbd5 protein sequence schematic indicating the location of the conserved aspartate triad (black) and the exon 2 D262A substitution (red) D. Preprint Figure 2O: Bar plot of seizure activity Pgbd5-/- versus Pgbd5 D262A/D262A littermate mice (χ2-test p = 7.41E-05). n indicates number of mice assayed.

 

2.  PGBD5 causes DNA rearrangements during cortical development

As neurons exit the cell cycle and specify their cell fate, they accumulate dsDNA breaks and activate end-joining repair. To determine how Pgbd5’s activity is affecting neurodevelopment, the authors focused on the window of peak cortical development in mice (E14.5) and measured the frequency of DNA damage and repair in post-mitotic neurons from WT and Pgbd5 ^-/-  mice. PGBD5 ^-/- mice had significantly fewer post-mitotic neurons with yH2AX foci at E14.5 than WT mice. This reduction was not observed in mitotic cells nor at E12.5, so PGBD5-dependent breaks seem specific to this stage of development. Next, the authors ask whether PGBD5-dependent breaks require repair and could make sustained rearrangements. They compare break repair in Pgbd5 ^-/- , Xrcc5 ^-/- , and double knockout Xrcc5^-/-  Pgbd5 ^-/-  mice to test whether Xrcc5’s repair mechanism (also used in RAG-mediated DNA rearrangements) can work on Pgbd5-mediated breaks. Pgbd5 ^-/-   mice had fewer post-mitotic neurons with unrepaired breaks at E14.5 than WT mice. Xrcc5^-/-  mice had more unrepaired breaks (more yH2AX foci) causing increased apoptosis during neurogenesis, but this was not specific to post-mitotic neurons. Xrcc5^-/-  Pgbd5 ^-/-  mice showed reduced DNA damage only in post-mitotic neurons. Together, these results point to PGBD5-dependent breaks in post-mitotic neurons being repaired by Xrcc5. But are these rearrangements or is this just repair? The authors sampled multiple brain regions from WT and knockout mice for whole genome sequencing to find DNA rearrangements by identifying recurring double-stranded breakpoints or complete overlap regions. WT adult mice had significantly more recurrent DNA rearrangements in multiple brain regions compared to PGBD5 ^-/-  mice. These breakpoints also had a distinct genomic distribution. This trend, combined with previously published evidence of Pgbd5 nuclease activity, indicated Pgbd5 nuclease activity is likely locus-specific.

 

 

Fig 2. PGBD5-dependent DNA breaks undergo NHEJ repair and form rearrangements sustained into adulthood.
A. Formed from preprint Figures 3A and 4A: Immunofluorescence micrographs of Pgbd5wt/wt , Pgbd5-/-, Xrcc5-/-;Pgbd5wt/wt , and double knockout Xrcc5-/-;Pgbd5-/-  14.5-day old litter mate embryos. DAPI (blue) stains nuclei. γH2AX (white) indicates double-stranded DNA breaks, and Tuj1 (red) stains differentiated post-mitotic neurons. B. Preprint Figure 5E: One histogram of variant junction split and discordant reads used by Delly2 to determine structural variants. Here shown is validated locus chr13:82838396-82839269.

 

Quick Sum Up and Implications:

 

Mutations in PGBD5 thought to reduce its nuclease activity cause neurodevelopmental defects in mouse and human. PGBD5-dependent breaks create sustained rearrangements in multiple brain tissues of adult mice. The authors suggest that:

 

1)     PGBD5-dependent DNA breaks may be an underlying mechanism behind developmental apoptosis.

2)    PGBD5-dependent DNA rearrangements of target genes may promote neuronal diversification and shape organization of the brain as hypothesized decades ago.

 

Through both of these mechanisms, and seen in the phenotypes of impaired motor learning and dramatic changes to synaptic strength (namely epilepsy), PGBD5-mediated breaks and rearrangements may also be linked to synaptic plasticity.

 

 

Why I chose this preprint:

Nearly half of all cells produced in mammalian neurogenesis die via developmental apoptosis. The evolution of a developmental process that requires such large-scale programmed cell death is incredible in the light of units of selection. Selection would have to act on genes and cells (likely negatively) first before they join to become tissues and organs like the brain, and before any anticipated function for organization in the brain can benefit the organism. I love stories where systems are built despite, or even because of conflicts at lower levels of selection. Similarly, TEs and their host ride the line between friend and foe, and I love studying how cooperation and complexity evolve in systems like this. I think Pgbd5 and factors like it can tell a similar story, by making a mechanistic connection between developmental apoptosis, DNA damage and rearrangements, and synaptic plasticity.

 

Future Directions/Questions for the Authors:

 

Q1. How do you think Pgbd5-dependent breaks are biased to post-mitotic neurons?
Or does this more reflect a difference in repair in post-mitotic neurons?

 

Q2. You mention that the choice to use bulk WGS to identify rearrangements was partly because of the limits on detection of larger variants produced by rearrangements and the requirements of DNA amplification. Is it worth analyzing trends in these breakpoints from the bulk data or looking into expression of alternative transcripts between WT and KO? Have you already seen any functional trends between break sites targeted by Pgbd5 (from either the WGS , ChIP data, or DEG from the knockin model)?

 

Q3. On this note: Simi et al. posted a preprint (11) shortly after yours where they find higher expression of Pgbd5 at E12 and E18, and show that Pgbd5 seems to gate differentiation and migration of neocortical neurons. Combined with your E14.5 phenotypes, I wondered if either:

• • This reflects waves where Pgbd5-dependent breaks enable migration through the diversification of target genes.
  • This reflects waves where apoptosis gates differentiation of these neurons.

What do you make of these findings?

 

 

Cited:

1. L. J. Zapater, E. Rodriguez-Fos, M. Planas-Felix, S. Lewis, D. Cameron, P. Demarest, A. Nabila, J. Zhao, P. Bergin, C. Reed, M. Yamada, A. Pagnozzi, C. Nava, E. Bourel-Ponchel, D. E. Neilson, A. Dursun, R. K. Özgül, H. T. Akar, N. D. Socci, M. Hayes, R. Rabadan, D. Torrents, M. C. Kruer, M. Toth, A. Kentsis, A transposase-derived gene required for human brain development (2023), p. 2023.04.28.538770, , doi:10.1101/2023.04.28.538770.
2. L. M. Payer, K. H. Burns, Transposable elements in human genetic disease. Nat Rev Genet. 20, 760–772 (2019).
3. A. A. Aravin, G. J. Hannon, J. Brennecke, The Piwi-piRNA pathway provides an adaptive defense in the transposon arms race. Science 318:761–64 (2007).
4. F. M. Jacobs, D. Greenberg, N. Nguyen, M. Haeussler, A. D. Ewing et al., An evolutionary arms race between KRAB zinc-finger genes ZNF91/93 and SVA/L1 retrotransposons. Nature 516:242–45 (2014).
5. V. Sundaram, Y. Cheng, Z. Ma, D. Li, X. Xing, P. Edge, M. P. Snyder, T. Wang, Widespread contribution of transposable elements to the innovation of gene regulatory networks. Genome Res. 24, 1963–1976 (2014).
6. R. L. Cosby, J. Judd, R. Zhang, A. Zhong, N. Garry, E. J. Pritham, C. Feschotte, Recurrent evolution of vertebrate transcription factors by transposase capture. Science. 371, 797 (2021).
7. A. G. Henssen et al., PGBD5 promotes site-specific oncogenic mutations in human 45 tumors. Nat Genet 49, 1005-1014 (2017).
8. A. G. Henssen, E. Henaff, E. Jiang, A. R. Eisenberg, J. R. Carson, C. M. Villasante, M. Ray, E. Still, M. Burns, J. Gandara, C. Feschotte, C. E. Mason, A. Kentsis, Genomic DNA transposition induced by human PGBD5. eLife. 4, e10565.
9. C. P. Gilman, M. P. Mattson, Do apoptotic mechanisms regulate synaptic plasticity and growth-cone motility? Neuromol Med. 2, 197–214 (2002).
10. W. R. Gray, W. J. Dreyer, L. Hood, Mechanism of antibody synthesis: size differences between mouse kappa chains. Science 155, 465-467 (1967).
11. A. Simi, F. Ansaloni, D. Damiani, A. Codino, D. Mangoni, P. Lau, D. Vozzi, L. Pandolfini, R. Sanges, S. Gustincich, The Pgbd5 DNA transposase is required for mouse cerebral cortex development through DNA double-strand breaks formation (2023), p. 2023.05.09.539730, , doi:10.1101/2023.05.09.539730.

 

doi: https://doi.org/10.1242/prelights.35009

Read preprint (No Ratings Yet)