And Dendritic Atrophy by the N-Methyl-D-Aspartate Receptor

Proc. Nati. Acad. Sci. USA Vol. 89, pp. 5197-5200, June 1992 Neurobiology Reduction of posttraumatic transneuronal "early gene" activation and dendritic atrophy by the N-methyl-D-aspartate receptor antagonist MK-801 (transneuronal degeneration/rat hippocampus/excitatory amino acid/neurodegeneratlve disorders/parvalbumin) ROBERT NITSCH* AND MICHAEL FROTSCHER Institute of Anatomy, University of Freiburg, D-7800 Freiburg, Federal Republic of Germany Communicated by J. Szentagothai, February 24, 1992

ABSTRACT The removal ofa major hippocampal afferent In our experimental paradigm, neurons are selectively system, the glutamatergic fibers from the entorhinal cortex, labeled with antibodies against the calcium-binding protein results in transneuronal changes in postsynaptic inhibitory parvalbumin (10, 16). These cells form a subpopulation of neurons using y-aminobutyric acid (GABA) as a neurotrans- hippocampal GABAergic cells and are stained in a Golgi-like mitter. This study shows that these transneuronal alterations manner (17)-i.e., their entire dendritic arbor can be ob- are reduced by the selective N-methyl-D-aspartate (NMDA) served extending into the termination zones ofthe entorhinal receptor antagonist (+)-MK-801. Thus, systemic injection of fibers. This allows for an analysis of distal dendritic changes. (+)-MK-801 prior to and after entorhinal lesion abolishes the Previous studies have demonstrated that all parvalbumin- retraction of distal dendrites from the termination zones of immunoreactive neurons in the hippocampus and fascia degenerating entorhinal fibers and reduces the swelling oftheir dentata contain the inhibitory neurotransmitter GABA (17) distal segments. Also, entorhinal lesion results in the appear- and that the entorhinal fibers, in addition to contacting ance of c-os protein-like immunoreactivit in hippocampal principal neurons (18), establish synapses with parvalbumin- neurons and glial cells, which again is blocked by (+)-MK-801 immunoreactive cells (19). Thus, our in vivo model enables us administration. These data suggest that NMDA receptor- to study transneuronal changes in identified postsynaptic mediated neurotoxicity due to postlesional glutamate elevation elements after entorhinal lesion. results in early gene expression and in transneuronal dendritic changes. Similar processes may play a role in Alzheimer's dsease, since there is a severe degeneration of the glutamater- MATERIALS AND METHODS gic entorhino-hippocampal projection in this neurodegenera- Twenty-two adult male Sprague-Dawley rats (250-g body tive disorder. weight) raised under standard laboratory conditions were used for the first set of experiments. The NMDA receptor Glutamate neurotoxicity has been studied in a variety of blocker (+)-MK-801 was administered i.p. as an aqueous experimental in vitro model systems (1-4), and the involve- solution (1 mg/kg of body weight 30 min prior to and every ment of excitatory amino acids in acute (5, 6) and chronic 12 hr after operation until sacrifice). Under sodium pento- neuronal degeneration (7-9) has been demonstrated. barbital anesthesia (45 mg/kg i.p.), the animals were sub- N-Methyl-D-aspartate (NMDA) receptor-mediated pro- jected to unilateral electrolytic lesion ofthe entorhinal cortex cesses are known to play a key role in these neurotoxic by using a stereotaxic apparatus (19). Coordinates were glutamate effects (5, 6). After entorhinal cortex lesion (10), adjusted according to a stereotaxic atlas. Animals that re- we have demonstrated transneuronal alterations of periph- ceived an electrolytic lesion on the pial surface ofthe cerebral eral dendrites of hippocampal neurons using -aminobutyric cortex and unoperated animals were used as controls. The acid (GABA) as a transmitter. The majority of the perforant animals were perfusion-fixed under deep sodium pentobar- path fibers arising from the entorhinal cortex and terminating bital anesthesia (60 mg/kg i.p.) 48 hr postlesion by using a in the hippocampus and fascia dentata is known to use the fixative containing 4% paraformaldehyde and 15% saturated excitatory amino acid glutamate as a transmitter (11). Ele- picric acid in 0.1 M phosphate buffer (pH 7.4). Forty- vation of extracellular concentration of glutamate in the micrometer-thick Vibratome sections of the midsepto- hippocampus in the course of neuronal degeneration has temporal portion ofthe hippocampus were immunostained. A recently been demonstrated by intracerebral microdialysis monoclonal mouse anti-Parv antiserum (16), diluted 1:5000, (12). To study the involvement of a postlesional pathological was used. The avidin-biotin horseradish peroxidase complex glutamate release leading to NMDA receptor-mediated neu- (ABC) technique was used, and the tissue-bound peroxidase rotoxicity, we have studied the effects of the selective a chro- NMDA receptor antagonist (+)-MK-801 (13) on transneu- was visualized with diaminobenzidine (DAB) as ronal dendritic changes. mogen. Incubation times, dilution ofthe secondary antibody, Recent studies have shown that the expression of "imme- and the ABC compound as well as buffers were the same as diate early genes" such as c-fos is related to NMDA receptor- described in detail elsewhere (19). Sections were osmicated mediated neurodegeneration under pathological conditions and embedded in Eukitt (Merck). Control sections were (14, 15). Thus, we analyzed the effects of MK-801 on the immunostained with all but the primary antiserum. No stain- expression of c-fos protein in the same experimental para- ing was observed under these conditions. Swellings of distal digm. The activation ofthe immediate early gene c-fos can be dendritic tips of immunoreactive neurons (10) were counted demonstrated on the level of protein synthesis by using antibodies to c-fos protein (15). Abbreviations: NMDA, N-methyl-D-aspartate; GABA, '-taminobu- tyric acid; DAB, diaminobenzidine; ABC, avidin-biotin horseradish peroxidase complex. The publication costs of this article were defrayed in part by page charge *To whom reprint requests should be addressed at: Institute of payment. This article must therefore be hereby marked "advertisement" Anatomy, University of Freiburg, Albertstrafle 17, D-7800 in accordance with 18 U.S.C. §1734 solely to indicate this fact. Freiburg, Federal Republic of Germany.

5197 Downloaded by guest on September 24, 2021 5198 Neurobiology: Nitsch and Frotscher Proc. Natl. Acad. Sci. USA 89 (1992)

3 DAB for parvalbumin and DAB/Ni for c-fos protein. Control sections were immunostained with all but the primary anti- sera. No staining was observed under these conditions.

RESULTS Forty-eight hours after lesion of the entorhinal cortex, the distal dendrites of parvalbumin-positive neurons have retracted from the termination sites of the degenerating entorhinal fibers close to the hippocampal fissure (Figs. 1 and 2b) (10). Their dendritic tips in stratum lacunosum- :.: moleculare of the hippocampus proper and in the outer .... molecular layer ofthe fascia dentata displayed large varicose swellings (Fig. 2b). Sections from nonoperated animals (Fig. 2a) and from animals that had received a (+)-MK-801 treatment prior to and every 12 hr after entorhinal lesion (Fig. 2c) FIG. 1. Schematic drawing of hippocampus, showing two parv- showed smooth dendrites of parvalbumin-immunoreactive albumin-positive neurons with theircell bodies located in the hilus (h) hippocampal neurons traversing the stratum lacunosum- next to the granule cell layer offascia dentata (g) and CA1 pyramidal cell layer of hippocampus proper (p), respectively. Their dendrites moleculare of the hippocampus proper and dentate neurons run through the molecular layer (ml) and stratum radiatum (r) and traversing the outer molecular layer of the fascia dentata, stratum lacunosum-moleculare (Im), respectively, and approach the meeting each other at the hippocampal fissure. In comparison hippocampal fissure (th). In stratum lacunosum-moleculare and in with lesioned animals, the number of swellings on distal tips the outer molecular layer next to the hippocampal fissure, the of parvalbumin-positive dendrites was reduced by 78% ± 3 perforant path fibers (arrows) from the entorhinal cortex terminate. on the dentate and by 64% ± 4 on the hippocampal side ofthe The small framed area 2 corresponds to the region shown in the fissure when the lesion was accompanied by a (+)-MK-801 micrographs ofFig. 2. Large framed area 3 corresponds to the region treatment (Fig. 2 b and c; n = 4). shown in the micrographs of Fig. 3 a-c. Immunoreactivity for c-fos protein in hippocampal neu- on both sides of the hippocampal fissure by using a drawing rons and glial cells 5 hr after entorhinal lesion (Fig. 3b) can tube. be blocked almost completely by (+)-MK-801 treatment prior were used for the second set to operation (Fig. 3c), thus emulating the absence of immu- Sixteen animals (see above) noreactivity as observed in nonoperated animals (Fig. 3a). In of experiments. Thirty minutes prior to unilateral lesion of lesioned but non-(+)-MK-801-treated animals, neurons in the the entorhinal cortex, (+)-MK-801 was administered (3 hilus showed pronounced c-fos protein-like immunoreactiv- mg/kg i.p.) as an aqueous solution. Five hours after the ity (Fig. 3b), including parvalbumin-containing cells as re- operation, the animals were sacrificed. Anesthesia, lesion, vealed by double labeling studies (Fig. 3d). fixation, and tissue preparation were the same as described above. This time a monoclonal mouse anti-Parv antiserum (16) diluted 1:5000 and a polyclonal goat anti-c-fos protein DISCUSSION antiserum (purchased from Cambridge Research Biochemi- One interpretation of these data is that transneuronal alter- cals, Harston, U.K.-) diluted 1:2000 were used. By using the ation after entorhinal cortex lesion results from postlesional ABC and peroxidase-antiperoxidase (PAP) technique, re- glutamate elevation (12), leading to NMDA receptor- spectively, the tissue-bound peroxidase was visualized with mediated c-fos expression (20, 21). This in turn may activate

.> lesion ';C %t 6esion'.. 1t; t control32 4 1 4 -8 01 \. t 4K j S ; I * > eit f-{ fA 1 t i f

S > jt. *-_i s ', ',i ' Hj,

ft o w eb .;. 4 . +. .s :f

!,,o-A# awit'^ ;.aZ.I

I-

1,I .. 1. ¢l Iml gS>A,''''.'s .. i. : I FIG. 2. (a) Parvalbumin-immunoreactive distal dendrites of hippocampal and dentate neurons in a 40-,m-thick Vibratome section from a control rat (similar region as framed in box 2 in Fig. 1). The distal dendrites of immunoreactive dentate neurons run through the molecular layer (ml) of the fascia dentata and approach the hippocampal fissure (fh). Peripheral dendrites of hippocampal parvalbumin-positive neurons reach the hippocampal fissure from the opposite side via the stratum lacunosum-moleculare of the hippocampus proper (lm). Numerous immunoreactive dendrites of hippocampal and dentate parvalbumin-immunoreactive neurons, respectively, traverse the hippocampal fissure (thin arrows). The layers shown here are the termination zones of afferent fibers from the entorhinal;* j_ Zcortex. (b) Same layers as in a 48 hr after an entorhinal lesion. No (+)-MK-801 administration has been performed. Distal dendrites ofparvalbumin-immunoreactive hippocampal and dentate neurons have retracted from the hippocampal fissure by forming large swellings of their distal tips (small arrows). Note the lack of dendrites traversing the hippocampal fissure. (c) Section from an animal that received (+)-MK-801 injections (1 mg/kg i.p.) 30 min prior to and every 12 hr after the operation until the animal was sacrificed 48 hr postlesion. The majority ofparvalbumin-immunoreactive dendrites has not retracted and traverses the hippocampal fissure from both sides (thin arrows). Swellings of dendritic distal tips are rare (small arrows). (Bar = 50 ,Lm.) Downloaded by guest on September 24, 2021 Neurobiology: Nitsch and Frotscher Proc. Natl. Acad. Sci. USA 89 (1992) 5199

AdI

.1. 4' h

10 4p - -. 40

FIG. 3. (a-c) Forty-micrometer-thick Vibratome sections of rat hippocampus stained with antibodies against c-fos protein (similar region as framed in box 3 in Fig. 1; for abbreviations see the legend of Fig. 1). (a) Section from a control animal. Note the almost complete absence of staining in neurons and glial cells. (b) c-fos protein immunoreactivity in neurons and glial cells of hippocampus 5 hr after entorhinal lesion. No (+)-MK-801 administration has been performed. Note the pronounced staining of cells deep to the granule cell layer and in the hilus and next to the pyramidal cell layer. (c) Almost complete absence of c-fos protein immunoreactivity in neurons and glial cells 5 hr after entorhinal lesion of an animal that received a (+)-MK-801 injection (3 mg/kg i.p.) 30 min prior to the operation. (d) Double labeling for parvalbumin and c-fos protein of the hippocampus of an animal 5 hr after entorhinal lesion. No (+)-MK-801 treatment has been performed. A parvalbumin- immunoreactive nonpyramidal cell in the hilus (h), visualized by the DAB chromogen (light grey staining of perikaryal and dendritic cytoplasm, thin arrows), contains a nucleus that is heavily immunoreactive for c-fos protein as visualized by the DAB/Ni chromogen (black staining, fat arrow). Granule cells (g) are immunoreactive for c-fos protein but lack parvalbumin immunoreactivity. (Bar in b = 200 Am (for a-c); bar in d = 50 Aum.) genes regulating cytoskeletal remodeling of postsynaptic changes after entorhinal lesion by (+)-MK-801 may be due to dendrites. NMDA receptor activation may thus cause the the blockade of NMDA receptor-induced calcium influx, acute cytopathology occurring after status epilepticus, cere- which otherwise would lead to c-fos expression and to bral ischemia, and trauma (5, 6, 22). The electrolytic lesion of cytoskeletal changes (25). In a previous study we showed that the entorhinal cortex as applied in the present experiments these cytoskeletal changes result in persisting dendritic al- may mimic some of the pathological mechanisms of status terations (10). The present study demonstrates the involve- epilepticus (23, 24) by resulting in an increased glutamate ment of NMDA receptors in initial stages of this process. release from perforant path terminals in the fascia dentata Activation of mRNAs encoding for cytoskeletal proteins and hippocampus. However, the persisting dendritic retrac- such as actin and tubulin has been demonstrated for the rat tion of parvalbumin-containing neurons after entorhinal le- fascia dentata after entorhinal lesion (27). It may be of sion (10) cannot be explained by glutamate neurotoxicity interest that persisting dendritic retraction after entorhinal alone. Terminal degeneration of entorhinal fibers may play a lesion has been shown exclusively for parvalbumin- major role in these postsynaptic dendritic alterations. It containing cells (10) and not for dentate granule cells (28). appears that the maintenance of peripheral dendritic seg- Granule cells lack this calcium-binding protein (17). Our ments of parvalbumin-containing neurons requires the spe- results suggest a pronounced activation of c-fos in parvalbu- cific afferent innervation from the entorhinal cortex (10). min-containing neurons (Fig. 3 b and d). The role of the The increase in intracellular calcium, which is induced by calcium-binding protein parvalbumin in controlling postle- NMDA receptor activation, has been related to long-term sional NMDA receptor-induced calcium influx as well as changes in neuronal excitability, neuronal development, and long-term effects of the MK-801 treatment remain to be cell death, probably mediated by changes in gene expression elucidated. (13, 20, 21, 25). Administration of NMDA receptor antago- The glutamatergic projection from the entorhinal cortex to nists such as (+)-MK-801 is effective in protecting neurons the hippocampus is among the first pathways to be severely from glutamate-induced degeneration (26) and from trau- affected in Alzheimer disease (29). In this neurodegenerative matic degeneration and blocks the activation of the immedi- disorder and after entorhinal lesions in rats, an elevation of ate early gene c-fos (1-3, 13). The reduction of transneuronal a pathological glycoprotein in the hippocampus has been Downloaded by guest on September 24, 2021 5200 Neurobiology: Nitsch and Frotscher Proc. Natl. Acad Sci. USA 89 (1992) reported (30). It is ofinterest that Alzheimer-type cytopatho- 13. for review see Lodge, D. & Johnson, K. M. (1991) Trends logical changes have been induced in cultured human neurons Pharmacol. Sci. Special Report, 13-18. by glutamate exposure (31, 32) and have been observed in a 14. Cole, A. J., Saffen, D. W., Baraban, J. M. & Worley, P. F. chronic neurological syndrome linked to dietary consump- (1989) Nature (London) 340, 474-476. tion of excitatory amino acid toxins (Guam disease) (7). The 15. Uemura, Y., Kowall, N. W. & Beal, M. F. (1991) Brain Res. 542, 343-347. data of the present study suggest that primary neuropatho- 16. Celio, M. R., Baier, W., Scharer, P. A., De Viragh, P. A. & logical degeneration of entorhinal projection neurons in Alz- Gerday, C. (1988) Cell Calcium 9, 81-86. heimer disease (29) may result in secondary NMDA receptor- 17. Nitsch, R., Soriano, E. & Frotscher, M. (1990) Anat. Embryol. mediated dendritic alterations in GABAergic hippocampal 181, 413-425. neurons. Such subtle dendritic changes have to be taken into 18. Matthews, D. A., Cotman, C. W. & Lynch, G. (1976) Brain consideration, since parvalbumin-containing neurons are re- Res. 115, 1-21. ported to survive in this disease (33). 19. Zipp, F., Nitsch, R., Soriano, E. & Frotscher, M. (1989) Brain Res. 495, 161-166. The authors thank Dr. D. K. Meyer for critical discussions in the 20. Kleinschmidt, A., Bear, M. & Singer, W. (1987) Science 238, course of the experiments, Drs. G. M. Peterson and H. D. Hofmann 355-358. for remarks on the manuscript, Dr. M. R. Celio for providing the 21. Sonnenberg, J. L., Mitchelmore, C., Macgregor, P. F., Hem- anti-parvalbumin antiserum, and G. Schmidt and M. Winter for stead, J., Morgan, J. I. & Curran, T. (1989) J. Neurosci. Res. excellent technical assistance. This study was supported by the 24, 72-80. Deutsche Forschungsgemeinschaft (Fr 620/1-5; Ni 344/1-1). 22. Isokawa, M. & Levesque, M. F. (1991) Neurosci. Lett. 132, 212-216. 1. Garthwaite, G. & Garthwaite, J. (1989) Neusosci. Lett. 97, 23. Sloviter, R. S. (1983) Brain Res. Bull. 10, 675-697. 316-322. 24. White, J. D., Gall, C. M. & McKelvy, J. F. (1987) J. Neurosci. 2. Rothman, S. M. (1987) Neuroscience 22, 471-480. 7, 753-759. 3. Choi, D. W. (1987) J. Neurosci. 7, 369-379. 25. Morgan, J. I. & Curran, T. (1986) Nature (London) 322, 552- 4. Foster, A. C., Gill, R. & Woodruff, G. N. (1988) J. Neurosci. 555. 8, 4745-4754. 26. Choi, D. W., Jae-Young, K. & Peters, S. (1988) J. Neurosci. 8, 5. Simon, R. P., Griffiths, T., Evans, M. C., Swan, J. H. & 185-1%. Meldrum, B. S. (1984)J. Cereb. BloodFlow Metab. 4, 350-361. 27. Phillips, L. L. & Steward, 0. (1990) Mol. Brain Res. 8, 249- 6. Auer, R. N., Kalimo, H., Olsson, Y. & Wieloch, T. (1985) Acta 257. Neuropathol. 67, 279-288. 28. Caceres, A. & Steward, 0. (1983) J. Comp. Neurol. 214, 7. Spencer, P. S., Nunn, P. B., Hugon, J., Ludolph, A. C., Ross, 387-403. S. M., Roy, D. N. & Robertson, R. C. (1987) Science 237, 29. Hyman, B. T., VanHoesen, G. W., Damasio, A. R. & Barnes, 517-522. C. L. (1984) Science 225, 1168-1170. 8. Weiss, J. H. & Choi, D. W. (1988) Science 241, 973-975. 30. May, P. C., Lampert-Etchells, M., Johnson, S. A., Poirier, J., 9. Seawright, A. A., Brown, A. W., Nolan, C. C. & Cavanagh, Masters, J. N. & Finch, C. E. (1990) Neuron 5, 832-839. J. B. (1990) Neuropathol. Appl. Neurobiol. 16, 153-169. 31. DeBoni, U. & McLachlan, D. R. (1985) J. Neurol. Sci. 68, 10. Nitsch, R. & Frotscher, M. (1991) Brain Res. 554, 304-307. 105-118. 11. Storm-Mathisen, J. & Ottersen, 0. P. (1984) in Cortical Inte- 32. Mattson, M. P. (1990) Neuron 4, 105-117. gration (Raven, New York), pp. 105-130. 33. Hof, P. R., Kox, K., Young, W. G., Celio, M. R., Rogers, J. 12. Beneviste, H., Drejer, J., Schousboe, A. & Diemer, N. H. & Morrison, J. H. (1991) J. Neuropathol. Exp. Neurol. 50, (1989) J. Neurochem. 43, 1369-1374. 451-462. Downloaded by guest on September 24, 2021