Letters

Retraction

Science  26 Mar 2004:
Vol. 303, Issue 5666, pp. 1974
DOI: 10.1126/science.303.5666.1974b

In our Report “Reelin promotes peripheral synapse elimination and maturation” (1), we presented data indicating that Reelin, a protein previously implicated in brain development (2), promotes synapse elimination and maturation at the neuromuscular junction. These conclusions were based mainly on our observations that Reelin was expressed and secreted by spinal motor neurons at the peripheral synapse; that mutant mice lacking Reelin (reeler) exhibited connectivity defects that included multiple fiber and multiple synapse innervation; and that intramuscular injection of Reelin stimulated axonal withdrawal. Several groups recently failed to reproduce our findings regarding the connectivity defects in reeler mice [see the Technical Comment in this issue (3)], prompting us to reexamine all the main points of the study. Two laboratories (G.D. and F.K.) attempted to replicate the data using reagents and mice similar to those used for our original Report, including two reeler strains obtained from the Jackson Laboratory and expanded by inbreeding for several generations.

The expression data for Reelin (Fig. 1 in our Report) and Reelin signaling proteins (Fig. 2) at the neuromuscular junction could not be reproduced. Recent immunofluorescence experiments in our laboratories also indicate that spinal motor neurons do not, in fact, express Reelin, consistent with the findings of others (4).

Upon reexamination, the number of end plates in reeler mice was not increased in whole mount preparations of the diaphragm as compared with wild-type mice, unlike the results shown in Fig. 3 and Table 1. Single fiber analysis confirmed that reeler muscles bear only one end plate per fiber, not multiple end plates as indicated in Fig. 4, disproving one important aspect of the reported connectivity phenotype. We were able to confirm the abnormal morphology and reduced size of reeler end plates. However, because the reeler mice that we examined in the Report were severely impaired due to genetic background and inbreeding, we now recognize that these abnormalities may have resulted from profound ataxia or central defects, rather than from absence of Reelin in the muscle. The same caveat applies to the electron microscopy data (Fig. 3 and Table 2).

The second aspect of the reeler connectivity phenotype, the presence of multiple axons innervating the same end plate (Fig. 4 and Table 1), has been clearly disproved by others and could not be replicated in our laboratories [see the Technical Comment and Response in this issue (3, 5)]. Our physiological recordings (Fig. 4I) may have been contaminated by an artifact resulting from incomplete block of action potentials in atrophic preparations of reeler muscles. Because no polyneuronal innervation exists in adult reeler synapses, the reported effect of injected Reelin in promoting axonal withdrawal (Fig. 4K) clearly must be wrong. Recent experiments in G.D.'s laboratory involving Reelin injection in the muscle of young wild-type mice, similar to those described in Fig. 4L, produced unreliable results. Because the activity of Reelin itself is in question, we also seriously doubt that the serine protease inhibitor Pefabloc SC prevents Reelin from promoting multiple axonal withdrawal as reported in Fig. 4L. Thus, the conclusion that Reelin promotes synapse elimination through its protease activity is not warranted. We did not attempt to replicate the data concerning the effect of Pefabloc SC alone (Fig. 4M), which may be correct, consistent with other studies implicating endogenous serine proteases in synapse elimination (6, 7).

With the exception of the in situ hybridization (produced by C.H. in M.S.'s laboratory), electron microscopy (J.C. Jr.'s laboratory), and electrophysiology data (D.R.M.'s laboratory), all data published in the Report were produced by C.C.Q. when he was in G.D.'s laboratory. We cannot presently explain the lack of reproducibility of the data. All authors agree that the study is critically flawed to an extent that we must retract it.

References and Notes

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.

Navigate This Article