Brain-derived neurotrophic factor modulates GAP-43 but not T alpha1 expression in injured retinal ganglion cells of adult rats.

The administration of neurotrophins affects neuronal survival and growth, but less is known about their ability to modify the expression of growth associated genes following injury to CNS neurons. Here we characterize the effect of brain-derived neurotrophic factor (BDNF) on mRNA levels for T alpha1 alpha-tubulin, and for GAP-43, two genes whose expression levels in retinal ganglion cells (RGC) tend to correlate with growth. We first determined that most adult rat RGCs can retrogradely transport BDNF by injecting 125I-BDNF into RGC target sites in vivo. We then used quantitative in situ hybridization to characterize the effect of axotomy, or axotomy and BDNF administration on mRNA levels for GAP-43 and T alpha1. Axotomy alone resulted in a general decrease in T alpha1 alpha-tubulin mRNA levels by 2 weeks, and elicited an increase in GAP-43 mRNA levels in an average of 30% of surviving RGCs. The intravitreal administration of a single dose of BDNF (5 microg) to axotomized RGCs on the day of injury did not affect T alpha1 alpha-tubulin mRNA levels, but was followed by a moderate (approximately 80%), and short-lasting enhancement of GAP-43 mRNA levels in most RGCs during the first week after axotomy. No significant increase in GAP-43 mRNA levels was observed when BDNF was injected into the uninjured eye. We conclude that BDNF specifically enhances GAP-43 but not T alpha1 mRNA levels in injured RGCs. Because BDNF is known to stimulate branch length of injured RGCs, we suggest that changes in the expression of GAP-43, but not T alpha1 tubulin, correlate with branching of injured neurons as opposed to long distance regrowth.

Marked increase in beta-tubulin mRNA expression during regeneration of axotomized retinal ganglion cells in adult mammals.

Changes in gene expression were investigated in axotomized CNS neurons under conditions that inhibit or permit regrowth of their damaged axons. Levels of mRNA encoding beta-tubulin, the 150 kDa neurofilament subunit (NF-M), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were examined by quantitative in situ hybridization of adult rat retinal ganglion cells (RGCs) after axotomy in the optic nerve or during regeneration in a peripheral nerve (PN) graft. Soon after optic nerve section beta-tubulin, NF-M, and GAPDH mRNA levels decreased and remained low during the 1 month studied. In these retinas beta-tubulin mRNA fell to approximately 50% of normal controls. However, in the PN-grafted retinas, where approximately 20% of the surviving axotomized RGCs regenerate their axons, there were "hot spots" of beta-tubulin mRNAs where neuronal levels were nearly 300% higher than in controls. By retrograde neuronal labeling these hot spots were shown to correspond to the injured RGCs that regrew their axons into the PN graft; beta-tubulin mRNA levels in nonregenerating RGCs of the same retinas averaged 63% of controls. We suggest that interactions of RBC axons and components of the grafts' non-neuronal environment play a key role in the over fourfold differences in beta-tubulin mRNA levels observed between injured and regenerating RGCs.

Temporal changes in beta-tubulin and neurofilament mRNA levels after transection of adult rat retinal ganglion cell axons in the optic nerve.

Axons of adult mammalian retinal ganglion cells (RGCs) do not regenerate spontaneously after injury in the optic nerve and show a persistent decrease in the rate of transport of tubulin and neurofilament proteins. To investigate further the expression of cytoskeletal proteins in these axotomized CNS neurons, mRNA levels of beta-tubulin and the 150 kDa neurofilament subunit (NF-M) were measured after interrupting the optic nerve 9 mm from the eye. Northern blots of RNA extracted from whole retinas after optic nerve transection showed that the total level of both of these mRNAs fell after injury. To determine if this decrease was a result of the death of axotomized RGCs or reflected changes in individual neurons, RNA probes were hybridized to radial cryostat sections of normal and axotomized retinas from 1 d to 6 months after injury. Grain counts revealed two trends of tubulin expression in RGCs. An early increase in tubulin mRNAs in the axotomized RGCs was followed by a later decrease. Such an increase in tubulin mRNA levels has been correlated with regenerative growth in other neurons. By 1 week after injury, the beta-tubulin mRNA levels decreased to 70% of the control value. Moreover, the time of this fall coincided with the onset of a marked slowing of cytoskeletal transport that follows injury in the optic nerve. In contrast, NF-M mRNA levels dropped immediately after axotomy, and remained at 80% of the control level. It is suggested that the transient increase in tubulin mRNAs may reflect an early regenerative response whose persistence depends on further growth cone interactions with the substrate.

Selective impairment of slow axonal transport after optic nerve injury in adult rats.

To investigate cellular responses of injured mammalian CNS neurons, we examined the slow transport of cytoskeletal proteins in rat retinal ganglion cell (RGC) axons within the ocular stump of optic nerves that were crushed intracranially. RGC proteins were labeled by an intravitreal injection of 35S-methionine, and optic nerves were examined by SDS PAGE at different times after injury. In one group of rats, the RGC proteins were labeled 1 week after crushing. From 14 to 67 d after axotomy, the labeling of tubulin and neurofilaments was reduced in relation to other labeled proteins and to the labeling of tubulin and neurofilaments in the intact optic nerve of controls. To determine whether this reduction in labeling was due to an alteration in axonal transport after axotomy, we prelabeled RGC proteins 1 week before crushing. In such experiments, the rate of slow axonal transport of tubulin and neurofilaments decreased approximately 10-fold from 6 to 60 d after injury. Our results cannot be due only to the retrograde degeneration of RGCs and injured axons caused by axotomy in the optic nerve, because fast axonal protein transport and the fluorescent labeling of many axons were preserved in the ocular stumps of these optic nerves. This selective failure of the slow axonal transport of tubulin and neurofilaments may affect the renewal of the cytoskeleton and contribute to the gradual degeneration of RGCs that is observed after axotomy. The alterations in slow transport we document here differ from the enhanced rates we previously reported when injured RGC axons regenerated along peripheral nerve segments grafted to the ocular stump of transected optic nerves (McKerracher et al., 1990).

The measurement of the production of tRNAMet1 in the Friend erythroleukemia cell.

We have used the gene for tRNAMet1 as a hybridization probe to measure the production of tRNAMet1 in the Friend erythroleukemia cell. In this cell, the relative concentration of tRNAMet1 (i.e., the percentage of total steady-state tRNA representing tRNAMet1) is 1.60 +/- 0.18. To study the relative synthesis of tRNAMet1, cells were labeled in vivo with [3H]uridine for periods ranging from 4 to 24 h, and the tRNA was isolated. The fraction of newly-synthesized tRNA representing tRNAMet1 (1.72% +/- 0.11) does not change when different in vivo labeling times are used. This value is similar to the relative concentration of tRNAMet1 in the older steady-state tRNA (1.61% +/- 0.18). The similar relative synthesis values using different labeling times, plus evidence presented that the total tRNA population decays homogeneously (t 1/2 = 110 h) indicate that tRNAMet1 has a cytoplasmic stability similar to the general tRNA population, and that its concentration relative to the tRNA population is established within the nucleus or soon after exiting the nucleus. Measurements of the synthesis of tRNAMet1 in isolated nuclei, relative to the synthesis of total RNA polymerase III transcripts, showed that this relative synthesis (0.291% +/- 0.017) is only 17% of the relative concentration of tRNAMet1 in the cytoplasm, which may reflect the presence of sequences other than tRNA in total nuclear polymerase III transcripts.

The use of cloned tRNA genes for the purification and measurement of specific tRNAs.

We have previously reported the ability of a cloned tRNAMeti gene (pt145) to bind tRNAMeti specifically [5]. In this paper, we show that a pBR322 plasmid containing the tRNAAsn gene of Xenopus (pt38 - donated by Stuart Clarkson) will specifically bind to mouse tRNAAsn when total mouse tRNA, extracted from uninduced Friend erythroleukemia cells, is hybridized to the gene probe. One-dimensional electrophoresis of the hybridizing tRNA in 20% polyacrylamide reveals one major band and several small-molecular-weight minor bands. The hybridizing tRNA has been identified as tRNAAsn by partial RNA sequencing and the detection of both the Q base and t6A. The steady-state concentration of tRNAAsn in the uninduced Friend cell was determined by hybridizing tRNA labeled in vitro to pt38. 1% of the total tRNA hybridized, representing 0.017 pg tRNAAsn/cell. The fraction of newly synthesized tRNA representing tRNAAsn or tRNAMeti was also determined by hybridizing tRNA labeled in vivo to either pt38 or pt145, respectively. 0.96% and 0.85% of the tRNA hybridized to pt38 and pt145, respectively.

The isolation and measurement of tRNAmeti using RNA/DNA hybridization.

A pBR322 plasmid containing the initiator tRNAmet gene of Xenopus (pt145 - donated by Stuart Clarkson) will specifically bind to mouse initiator tRNAmet (tRNAmeti) when total mouse tRNA, extracted from uninduced Friend erythroleukemia cells, is hybridized to the gene probe. One dimensional electrophoresis of the hybridizing tRNA in 20% polyacrylamide reveals one major band (95%) and a minor band. The hybridizing tRNA has been identified as initiator tRNAmet by RNA sequencing. Hybridization of tRNAtotal to another plasmid containing the Xenopus gene for tRNAasn results in two bound species with different electrophoretic mobilities than the tRNA bound to the initiator tRNAmet gene. pt145 has been used to measure the steady state concentration of initiator tRNAmet in the uninduced and erythroid Friend cell, and in the unfertilized egg and 21 h blastula of the sea urchin. Initiator tRNAmet represents 0.91% and 0.52% of the tRNA populations extracted from uninduced and erythroid Friend cells, respectively. Based upon the total tRNA content per cell, there is a 3.8 fold decrease in initiator tRNAmet per cell during erythroid differentiation. tRNA extracted from unfertilized eggs and 21 h blastula of the sea urchin both have 0.5% of total tRNA as initiator tRNAmet (approximately 1.5 pg).