An ATF3-CreERT2 Knock-In Mouse for Axotomy-Induced Genetic Editing: Proof of Principle

Axotomy does not induce ATF3 in Schwann cells. A, Cryosections from injured DRG (inset) and distal sciatic nerve from the same mouse processed for ATF3 immunohistochemistry (Novus NBP1-85816). B, Punctate staining in the nerve proved to be non-specific fluorescence of leukocytes (note non-nuclear signal in the absence of primary antibody). C, In intact sciatic nerves, cells morphologically identical to Remak cells had at some point undergone recombination. D, Following injury, their numbers increased. E, This was attributable to their proliferation in the injured nerve (as opposed to ATF3 induction and subsequent recombination). C′, D′, Magnification of areas outlined in C and D demonstrate the spindle shaped morphology characteristic of Remak cells.

Loss of ATF3 function delays functional recovery following sciatic nerve crush. A, B, The rate of functional recovery (nocifensive reflex withdrawal to a toe pinch and presence of any grasping ability) was reduced in mice lacking both wild-type ATF3 alleles (log rank Mantel–Cox test). Haploinsufficiency was also suggested by the statistically-significant trend from wild-type to homozygous knock-in (log rank Mantel–Cox test), n = 7, n = 9, and n = 8 for ATF3^+/+, ATF3^+/cre, and ATF3^cre/cre, respectively. C, Representative EMG traces from ipsilateral and contralateral sides of an ATF3^+/+ mouse 28 d after sciatic nerve crush, and composite traces from seven ATF3^+/+ mice and seven ATF3^cre/cre mice. D, EMG thresholds did not differ between genotypes (paired t test). E, F, While absolute latencies and amplitudes did not differ between genotypes (paired t test), their ipsilateral/contralateral ratios (correcting for mouse size) indicated reduced conduction velocity (E) and extent of reinnervation (F) in mice lacking one or both wild-type ATF3 alleles (unpaired t test), n = 6, n = 7, and n = 6 for ATF3^+/+, ATF3^+/cre, and ATF3^cre/cre, respectively.

Loss of ATF3 function modestly reduces axonal regeneration following sciatic nerve crush. A, B, There was no difference in axonal regeneration 2 d following injury between ATF3^+/+, ATF3^+/cre, and ATF3^cre/cre mice (n = 7, n = 6, and n = 6, respectively, groups were compared with a one-way ANOVA on cumulative densities). C, D, ATF3^cre/cre mice exhibited significantly diminished axonal regeneration 3 d following injury 2–4 mm distal to the injury compared to ATF3^+/+ mice (n = 5 for both groups, one-way ANOVA followed by Dunnett’s multiple comparison test). The hemizygous group (n = 4) tested positive as a significant intermediary between both control and ATF3 null groups (post hoc test for trend). Dotted line indicates distal border of crush site, 500 μm from the edge of the block. Scale bars: 500 μm.

Axotomy-induced PTEN deletion in sensory and motoneurons. A, B, Axotomy results in significant loss of PTEN expression in the DRG. Arrows in A indicate small PTEN-positive DRG neurons. The PTEN antibody results in high background staining in all animals, to which PTEN immunoreactivity is reduced in axotomized DRGs of ATF3^+/cre:PTEN^fl/fl mice (B; one-way ANOVA followed by Dunnett’s multiple comparison test, n = 5 for both groups). C, PTEN immunoreactivity is weak in all motoneurons, rendering difficult confirmation of axotomy-induced knock-down. However, ventral root (VR; large arrow) axons close to the ventral root exit zone were intensely immunopositive, single arrows indicate axons that were both tdtomato and PTEN positive, whereas double arrows indicate recombination without PTEN immunoreactivity. In sections of ventral roots, we were able to demonstrate a significant decrease in PTEN immunoreactivity in tdtomato-positive axons (D; Kolmogorov–Smirnov goodness of fit test), n = 4 and n = 3 for ATF3^+/cre:PTEN^+/+ and ATF3^+/cre:PTEN^fl/fl, respectively. Scale bars: 50 μm (A), 100 μm (C), 10 μm (E).

Axotomy-induced PTEN deletion and improves functional recovery following sciatic nerve crush. A, B, While there was no difference in recovery of reflex nociception or grasping in mice with floxed PTEN alleles (log rank Mantel–Cox test, n = 14 and n = 9 for ATF3^+/cre:PTEN^+/+ and ATF3^+/cre:PTEN^fl/fl, respectively), EMG responses (C–G) indicated enhanced recovery of neuromuscular function. C, D, Representative EMG traces from ipsilateral and contralateral sides of an ATF3^+/cre:PTEN^+/+ mouse 28 d after sciatic nerve crush (C) and composite traces from eight ATF3^+/cre:PTEN^+/+ mice and six ATF3^cre/cre:PTEN^fl/fl mice. E–G, Ipsilateral/contralateral ratios of EMG thresholds (E), peak EMG latencies (F), and maximum CMAP amplitudes (G) all indicated more complete muscle reinnervation 28 d after injury; n = 8 and n = 6 for ATF3^+/cre:PTEN^+/+ and ATF3^+/cre:PTEN^fl/fl, respectively, averages were compared using paired t tests and ipsilateral/contralateral ratios with unpaired t tests.

Axotomy-induced PTEN deletion and anatomic regeneration following sciatic nerve crush. A, B, Axonal regeneration 2 d subsequent to sciatic nerve crush (dotted line) was significantly enhanced in ATF3^+/cre:PTEN^fl/fl mice. In C, bar graphs represent the cumulative density of SCG10 immunoreactivity from 0 μm (the distal extent of the crush site) to 2000 μm; n = 6 for both groups and cumulative densities were compared using an unpaired t test. Scale bars: 500 μm.

Axotomy-induced recombination one week following spinal hemisection: spinal cord. A, Two examples of the injury site from separate animals in longitudinal section. Insets show consistently-recombining ipsilaterally-projecting neurons near Clarke’s column below the injury. B, Cartoon of transverse section of the hemisected spinal cord illustrating relative positions of positionally and morphologically distinguishable recombined neurons (colored dots correspond to examples in C–F). C, Examples of recombination after injury in cervical (top left), lumbar (two examples middle and bottom-left), and thoracic (top right). The most consistent findings were small ipsilaterally-projecting neurons in the thoracic cord (C’), and large neurons contralateral to injury from just lateral to area X to the ventral gray matter in the lumbar cord (C’’). Arrows in C’, C’’ indicate midline-crossing axons. Arrow pointing to tdtomato⁺ axons in dorsal cervical white matter indicates probable rubrospinal (RST) and/or raphespinal tracts. Arrows pointing to reporter-positive axons in ventral white matter indicate probable reticulospinal (RtST) and vestibulospinal (VST), and an unknown descending projection (?) tracts rostral to the injury, and spinothalamic (STT) and possible dorsal spinocerebellar [DSCT(?)] below the injury (thoracic and lumbar sections). D, Neurons in Clarke’s column ipsilateral to the hemisection. E, F, Large and small, respectively, putative spinothalamic tract neurons contralateral to the hemisection. Arrows in E, F indicate commissural axons; cc, central canal.