- Email: [email protected]
Emperipolesis of lymphoid cells in vagal efferent neurons following an intraneural injection of ricin into the vagus nerve in rats
Neuroscience Letters 270 (1999) 153±156
Emperipolesis of lymphoid cells in vagal efferent neurons following an intraneural injection of ricin into the vagus nerve in rats Yee-Kong Ng, Eng-Ang Ling* Department of Anatomy, Faculty of Medicine, National University of Singapore, MD10, 4 Medical Drive, Singapore 117597, Singapore Received 29 April 1999; received in revised form 5 June 1999; accepted 5 June 1999
Abstract Injection of a minute amount of the toxic lectin, Ricinus communis agglutinin-60 (RCA-60) into the vagus nerve resulted in a selective destruction of the vagal efferent neurons in the ipsilateral dorsal motor nucleus (DMN). This has elicited a massive in¯ux of mononuclear leucocytes, notably macrophages and T-lymphocytes, as detected with ED-1 and OX-19 antibodies, respectively. A small number of B-lymphocytes as identi®ed by OX-33 antibody, were also observed in the neuropil of DMN. The in¯ux of mononuclear leucocytes into the neuropil of DMN was by way of diapedesis, peaking in frequency at 4±6 days after the RCA administration. The in®ltrated lymphocytes were closely associated with or penetrated the soma of the vagal neurons, some bearing intact axo-somatic synaptic contacts. The entrapped lymphocytes in neurons underwent degeneration and subsequently disintegrated. Macrophages and plasma cells in the neuropil did not appear to penetrate the neuronal soma. It is concluded that emperipolesis of lymphocytes, presumably cytotoxic T-cells, in RCA-poisoned neurons may represent a form of effector-target cell contact leading to cytotoxicity. In doing so, however, the invading lymphocytes were destroyed by the contents of RCA picked up by the neurons. The absence of macrophages and plasma cells in the RCA-poisoned neurons suggests the cellular speci®city of emperipolesis. q 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Ricin; Vagal neurons; Immunohistochemistry; Lymphocytes; Macrophages
Vagal efferent neurons in the dorsal motor nucleus (DMN) are selectively destroyed following the suicidal retrograde transport of the toxic lectin Ricinus communis agglutinin-60 (RCA-60) injected into the vagus nerve at the cervical region [7±9,13,16]. A remarkable feature of the DMN in RCA-treated rats was the massive in¯ux of mononuclear leucocytes including monocytes and lymphocytes. While it is unequivocal that some of the in®ltrated monocytes would transform into macrophages engaged in the phagocytosis of neuronal debris, the involvement of lymphocytes in the degeneration process of vagal neurons remained unclear; in fact, virtually nothing is known about the fate of the in®ltrated lymphocytes. This study was aimed to investigate some of the cellular events associated with the degeneration of vagal neurons killed by RCA by following the migration of mononuclear cells into the DMN and to examine their interactions with the vagal neurons over a time course sequence. Along with this, the immunopheno* Corresponding author. Tel.: 165-874-3200; fax: 165-7787643. E-mail address: [email protected] (E.-A. Ling)
typic features of the in®ltrated elements were studied. This information would help to better understand the immunological response in neural degeneration induced by a foreign antigen. Male Wistar rats, ranging from 150 to 200 g, were used in this study. Each rat was given, under chloral hydrate anesthesia, a single injection of 3 m1 0.1% Ricinus communis agglutinin-60 (RCA-60, Sigma) in 0.01 M phosphate buffer into the right vagus nerve. The RCA-60 was slowly introduced with a Hamilton syringe into the vagus nerve where it is crossed by the superior belly of the omohyoid muscle. The site of injection was then dried with a cotton bud. Control rats received equal volume of vehicle injection. The rats were allowed to survive for 4, 6, 15, 22, 30 and 60 days. At least two rats were perfused at each of the above time intervals with a mixed aldehyde solution composed of 2% paraformaldehyde and 3% glutaraldehyde in 0.1 M cacodylate buffer adjusted to pH 7.4. After perfusion which lasted for 30 min, the brainstem containing the DMN of vagus was removed. Transverse vibratome sections of 100 mm thickness were prepared. They were post®xed in
0304-3940/99/$ - see front matter q 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 9 9) 00 49 0- 5
154
Y.-K. Ng, E.-A. Ling / Neuroscience Letters 270 (1999) 153±156
2% osmium tetroxide in 0.1 M cacodylate buffer. After dehydration in alcohol, the tissue slices were embedded in an Araldite mixture. Ultrathin sections were cut, double stained with uranyl acetate and lead citrate and examined in a JEOL 1200EX eletron microscope. Immunohistochemical study was con®ned to rats killed at 4 and 6 days since massive in®ltration of mononuclear cells was observed at these two time points. The rats were perfused ®rst with Ringer's solution followed by an aldehyde ®xative composed of a mixture of periodate-lysineparaformaldehyde containing 2% paraformaldehyde as described previously [12]. Following removal, the brainstem was post-®xed in the same ®xative for 2±4 h and was kept in 0.1 M phosphate buffer containing 15% sucrose overnight at 408C. Serial coronal frozen sections were cut at 40 mm thick and incubated in one of the following monoclonal antibodies diluted in 1:100 with phosphate buffered saline: ED1 (Harlan SeraLab, MAS-341), OX19 (Harlan SeraLab, MAS-099b), and OX33 (Harlan SeraLab, MAS258b) monoclonal antibodies for detection of macrophages, CD5 equivalent T-cells and CD45RA B-cells, respectively. After incubation, the sections were rinsed in PBS for 15 min and then reacted in Vectastain ABC Kit (PK4002, Vector Laboratories) against mouse IgG for half an hour in a humidi®ed chamber. After this, the sections were rinsed in PBS and treated with a solution of 3-3 0 -diaminobenzidine tetrahydrochloride (DAB, Sigma-5637) and hydrogen peroxide (5 mg DAB, 10 ml PBS, 30 ml 0.3% H2O2). The sections were counterstained with 1% methyl green, dehydrated and mounted in Permount. In agreement with previous studies [7±9], at 4 and 6 days after RCA-60 injection into the right cervical vagus, massive in®ltration of mononuclear leucocytes was observed in the ipsilateral DMN of the vagus. Most of the in®ltrated cells were macrophages and T-lymphocytes, as evidenced by their immunoreaction with OX-19 (Fig. 1) and ED-1 (Fig. 2), respectively. The in®ltrated ED-1 and OX-19 immunoreactive (IR) cells were distributed throughout the ipsilateral DMN, with some of them distributed in the ventral part of the nucleus tractus solitarius (NTS). ED-1IR monocytes were frequently lodged in the vascular walls and their vicinity (Fig. 2). Some OX-19 positive cells were closely associated with the blood vessels (Fig. 1). Few to moderate numbers of OX-33-IR cells were also scattered in the ipsilateral DMN (Fig. 3). In control rats receiving vehicle injection, the above-mentioned immunopositive cells were absent. At 4 and 6 days after RCA injection, some lymphocytes identi®ed by their small nucleus with large chromatin masses and scanty cytoplasm were seen to penetrate the external basal lamina (Fig. 4) to enter the neuropil. In the latter, the neuronal somata were often surrounded by lymphocytes (Fig. 5). A variable number of cytoplasmic processes of lymphocytes appeared to deeply penetrate the neuronal somata. In some section pro®les, as many as up to ®ve lymphocytes were seen in a neuronal soma (Fig. 6). In
what may be an early stage of in®ltration of neuronal soma by lymphocytes, the immediate neuronal cytoplasm surrounding the invaded lymphocytes appeared dissoluted, as indicated by its absence of organelles (Fig. 7). Free degenerating lymphocytes were not uncommon in the neuropil. Sometimes, they were phagocytosed by macrophages which were characterized by an irregular or ¯attened nucleus with margination of chromatin (Fig. 8). Plasma cells in close apposition to neuronal soma were common (Fig. 9). They were characterized by the extensive pro®les rough endoplasmic reticulum. Injection of RCA into the vagus nerve selectively destroys the vagal efferent neurons as a result of the suicide retrograde transport of the toxic ricin. Associated with the neuronal breakdown was the massive in¯ux of mononuclear cells with a preponderance of lymphocytes. The in®ltration peaked at 4±6 days after the toxin injection, and the in®ltrated cells appeared to accumulate in the perivascular space before being dispatched to the neuropil of DMN. The exact signal to attract the in®ltration of mononuclear cells after RCA injection is uncertain. Since in®ltration of mononuclear cells was absent in the control rats receiving vehicle injection, it must be attributed to the injected ricin taken up by the DMN efferent neurons. It is possible that minute amounts of RCA may have been released from dying neurons and this may alter the integrity of blood-brain barrier in the DMN, and serve as the speci®c chemotactic factor for the attraction of mononuclear cells. Leakage of RCA may also have elicited secretion of cytokines or upregulation of adhesive molecules from local glial or vascular elements. This would facilitate the migration of lymphocytes, macrophages and plasma cells into the perivascular space as well as in the neuropil of the DMN and then trigger off a local immunological response. The most unusual feature in the DMN after RCA injection was the penetration of a variable number of lymphocytes into some of the surviving vagal efferent neurons. The internalization of lymphocytes into vagal neurons is reminiscent of `emperipolesis' described in other tissues [2,6,15]. In vitro studies have shown the capability of lymphocytes invading mixed cultures of different tissues especially of mesenchymal origin [5,6]. In biopsies of human melanoma, lymphocytes were seen within the cytoplasm of non-pigmented, nucleated melanoma cell [2]. The exact mechanism of internalization or emperipolesis of lymphocytes into some DMN neurons remains unclear. It is speculated that ricin uptake could have induced the expression of speci®c receptors or adhesive molecules on the neuronal surface that might have attracted the lymphocytes, which were eventually internalized into the neurons. Interestingly, in human laryngeal papilloma, neutrophils in various stages of degeneration and disintegration were seen within the neoplastic cells [1]. In the present study, some of the vagal neurons containing intracellular lymphocytes showed long projecting dendrites; in some, the synaptic contacts associated with the soma remained structurally
Y.-K. Ng, E.-A. Ling / Neuroscience Letters 270 (1999) 153±156
155
Fig. 1. Photomicrograph of the caudal medulla of a rat, 4 days after RCA injection into the right vagus nerve. Note the massive in®ltration of OX-19-immunoreactive (IR) T-cells (arrows) in the ipsilateral dorsal motor nucleus of vagus (DMN) and nucleus tractus solitarius (NTS), but not in the hypoglossal nucleus (HN). Some immunopositive cells (arrowheads) are closely associated with the blood vessels. V, fourth ventricle. Scale bar, 100 mm. Fig. 2. Massive in®ltration of ED1-IR cells in the ipsilateral DMN and part of the NTS, but not in the HN, 4 days after RCA injection. Some immunoreactive monocytes (arrows) appear to be lodged in the vascular walls and their vicinity. V, fourth ventricle. Scale bar, 100 mm. Fig. 3. Photomicrograph showing the caudal medulla, 4 days following RCA injection into the right vagus nerve. OX-33-IR cells (arrows) are scattered in the ipsilateral DMN across in the mid-line. V, fourth ventricle. Scale bar, 100 mm. Fig. 4. A lymphocyte (*) herniates through the external basal lamina (EBL) into the neuropil of DMN, 6 days after RCA injection. Scale bar, 0.5 mm. Fig. 5. Five in®ltrated lymphocytes (*) are in close association with the soma of a DMN neuron (N). The soma is postsynaptic to two axon terminals (circled area), 6 days after RCA injection. Scale bar, 1 mm. Fig. 6. Five lymphocytes (*) are entrapped in the cytoplasm of a DMN neuron (N), 6 days after RCA injection. The intraneuronal lymphocytes are at different stages of lysis. Scale bar, 1 mm. Fig. 7. A lymphocyte is within the cytoplasm of a neuron (N). The neuronal cytoplasm in the immediate vicinity (*) of the lymphocyte is free from any organelles (lysis of cytoplasm). Circled area, synaptic contacts; 6 days after RCA injection. Scale bar, 0.5 mm. Fig. 8. A degenerating cell in the cytoplasm of a macrophage (Ma). A lymphocyte (*) is partially engulfed by the macrophage, 6 days after RCA injection. Scale bar, 0.5 mm. Fig. 9. A plasma cell (P) capping the soma of a DMN neuron (N) containing a degenerating cell presumably lymphocyte (asterisk). Circled area, synaptic contact, 6 days after RCA injection. Scale bar, 1 mm.
156
Y.-K. Ng, E.-A. Ling / Neuroscience Letters 270 (1999) 153±156
intact suggesting that the affected cells remain viable. Why some neurons survived after being invaded by lymphocytes is also dubious. One possible explanation may be that the minute amounts of ricin was at too low levels to kill the neurons, but was at toxic levels to kill the internalized lymphocytes. The signi®cance of the penetration of lymphocytes into the neuronal soma remains enigmatic. It could be an attempt for the invaded T cells to lyse the DMN neurons which were recognized as non-self because of their contents of ricin. In other words, emperipolesis may represent an effector-target cell contact leading to cytotoxicity. The dissolution of the cytoplasm immediately surrounding the entrapped lymphocytes indicates the possibility of release of cytotoxic substances. The obligatory role of physical contact between cytotoxic lymphocytes and target cells has been stressed in both the NK cells, as well as the antibody-dependent, in vitro systems [4,6,10]. In the present study, the emperipolesis of lymphocytes into vagal neurons probably represents a more effective mode of destroying the ricin-poisoned cells. The fate of the entrapped lymphocytes seems unequivocal as shown by their destruction within the neuronal cytoplasm, but the mechanism of their destruction remains unclear. It is speculated that they were poisoned by ricin contained in the neuronal cytoplasm. Against this view is the fact that in some host cells, the cytoplasmic organelles, e.g. rough endoplasmic reticulum remained structurally intact. If ricin were to be present at a toxic level, one would expect the disruption of rough endoplasmic reticulum since it is known to inhibit protein synthesis [3,11,14]. Interestingly some lymphocytes which did not penetrate the neurons degenerated in the interstitial spaces and were removed by macrophages. It is likely that some free ricin might have been released by vagal neurons that had earlier picked up the ricin. The occurrence of plasma cells in the neuropil suggests their involvement in the neutralization of the toxic ricin. While the emperipolesis of lymphocytes in vagal neurons was a consistent feature, it is remarkable that neither macrophages nor plasma cells penetrated the neurons suggesting the cellular speci®city of this unique process in the nervous tissue. The non-af®nity of macophages and plasma cells to the affected neurons could be due to absence of some speci®c cell-surface receptor molecules, which may be expressed only on T cells. The work is supported by Research Grants (RP950363
and RP970328) from the National University of Singapore. The authors wish to thank Mr. Tajuddin B. Marican, Mrs. Yong Eng Siang, Mrs. Ng Geok Lan, Mr. Yick Tuck Yong, Miss. Chan Yee Gek and Mr. Gobalakrishnan for their expert technical assistance. [1] Ahmed, M.M., Studies on human laryngeal papilloma. Acta Oto-Laryngol., 92 (1981) 563±567. [2] Burns, E.R., Zucker-Franklin, D. and Valentine, F., Cytotoxicity of natural killer cells. Correlation with emperipolesis and surface enzymes. Lab. Invest., 47 (1982) 99±107. [3] Endo, Y., Mitsui, K., Motizuki, M. and Tsurugi, K., The mechanism of action of ricin and related toxic lectins on eukaryotic ribosomes. J. Biol. Chem., 262 (1987) 5908±5912. [4] Glauert, A.M. and Sanderson, C.J., The mechanism of Kcell mediated cytotoxicity III. The ultrastructure of K-cell projections and their possible role in target cell killing. J. Cell Sci., 35 (1979) 355±366. [5] Hughes, D., Raine, C.S. and Field, E.J., Invasion of neurons in vitro by nonimmune lymphocytes: an electron microscopic study. Br. J. Exp. Pathol., 49 (1968) 356±359. [6] Ioachim, H.L., Emperipolesis of lymphoid cells in mixed cultures. Lab. Invest., 14 (1965) 1784±1794. [7] Ling, E.A. and Leong, S.K., Effect of intraneural injection of ricinus comrnunis agglutinin-60 into the rat vagus nerve. J. Neurocytol., 16 (1987) 373±387. [8] Ling, E.A. and Leong, S.K., In®ltration of carbon-labelled monocytes into the dorsal motor nucleus following an intraneural injection of ricinus communis agglutinin-60 into the vagus nerve in rats. J. Anat., 159 (1988) 207±218. [9] Ling, E.A., Shieh, J.Y., Wen, C.Y., Chan, Y.G. and Wong, W.C., Degenerative changes of neurons in the superior cervical ganglion following an injection of ricinus communis agglutinin-60 into the vagus nerve in hamsters. J. Neurocytol., 19 (1990) 1±9. [10] Ling, N.R., Acton, A.B., Roitt, I.M. and Doniach, D., Interaction of Iymphocytes from immunized hosts with thyroid and other cells in culture. Br. J. Exp. Pathol., 46 (1965) 348±359. [11] MacConnel, W.P., Eurman, D.W. and Kaplan, N.O., The action of ricinA-chain on ribosomes. Biochem. Biophys. Res. Commun., 108 (1982) 809±814. [12] Ng, Y.K. and Ling, E.A., Induction of major histocompatibility class II antigen on microglial cells in postnatal and adult rats following intraperitoneal injections of lipopolysaccharide. Neurosci. Res., 28 (1997) 111±118. [13] Oeltmann, T.N. and Wiley, R.G., Wheat germ agglutininricin. A-chain conjugate is neuronotoxic after vagal injection. Brain Res., 377 (1986) 321±328. [14] Olsnes, S., Refsnes, K. and Pihl, A., Mechanism of action of the toxic lectins abrin and ricin. Nature, 249 (1974) 627±631. [15] Schelton, E. and Dalton, A.J., Electron microscopy of emperipolesis. J. Biophys. Biochem. Cytol., 6 (1959) 513. [16] Wiley, R.G., Blessing, W.W. and Reis, D.J., Suicide transport: destruction of neurons by retrograde transport of ricin, abrin and modeccin. Science, 216 (1982) 889±890.
Emperipolesis of lymphoid cells in vagal efferent neurons following an intraneural injection of ricin into the vagus nerve in rats Yee-Kong Ng, Eng-Ang Ling* Department of Anatomy, Faculty of Medicine, National University of Singapore, MD10, 4 Medical Drive, Singapore 117597, Singapore Received 29 April 1999; received in revised form 5 June 1999; accepted 5 June 1999
Abstract Injection of a minute amount of the toxic lectin, Ricinus communis agglutinin-60 (RCA-60) into the vagus nerve resulted in a selective destruction of the vagal efferent neurons in the ipsilateral dorsal motor nucleus (DMN). This has elicited a massive in¯ux of mononuclear leucocytes, notably macrophages and T-lymphocytes, as detected with ED-1 and OX-19 antibodies, respectively. A small number of B-lymphocytes as identi®ed by OX-33 antibody, were also observed in the neuropil of DMN. The in¯ux of mononuclear leucocytes into the neuropil of DMN was by way of diapedesis, peaking in frequency at 4±6 days after the RCA administration. The in®ltrated lymphocytes were closely associated with or penetrated the soma of the vagal neurons, some bearing intact axo-somatic synaptic contacts. The entrapped lymphocytes in neurons underwent degeneration and subsequently disintegrated. Macrophages and plasma cells in the neuropil did not appear to penetrate the neuronal soma. It is concluded that emperipolesis of lymphocytes, presumably cytotoxic T-cells, in RCA-poisoned neurons may represent a form of effector-target cell contact leading to cytotoxicity. In doing so, however, the invading lymphocytes were destroyed by the contents of RCA picked up by the neurons. The absence of macrophages and plasma cells in the RCA-poisoned neurons suggests the cellular speci®city of emperipolesis. q 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Ricin; Vagal neurons; Immunohistochemistry; Lymphocytes; Macrophages
Vagal efferent neurons in the dorsal motor nucleus (DMN) are selectively destroyed following the suicidal retrograde transport of the toxic lectin Ricinus communis agglutinin-60 (RCA-60) injected into the vagus nerve at the cervical region [7±9,13,16]. A remarkable feature of the DMN in RCA-treated rats was the massive in¯ux of mononuclear leucocytes including monocytes and lymphocytes. While it is unequivocal that some of the in®ltrated monocytes would transform into macrophages engaged in the phagocytosis of neuronal debris, the involvement of lymphocytes in the degeneration process of vagal neurons remained unclear; in fact, virtually nothing is known about the fate of the in®ltrated lymphocytes. This study was aimed to investigate some of the cellular events associated with the degeneration of vagal neurons killed by RCA by following the migration of mononuclear cells into the DMN and to examine their interactions with the vagal neurons over a time course sequence. Along with this, the immunopheno* Corresponding author. Tel.: 165-874-3200; fax: 165-7787643. E-mail address: [email protected] (E.-A. Ling)
typic features of the in®ltrated elements were studied. This information would help to better understand the immunological response in neural degeneration induced by a foreign antigen. Male Wistar rats, ranging from 150 to 200 g, were used in this study. Each rat was given, under chloral hydrate anesthesia, a single injection of 3 m1 0.1% Ricinus communis agglutinin-60 (RCA-60, Sigma) in 0.01 M phosphate buffer into the right vagus nerve. The RCA-60 was slowly introduced with a Hamilton syringe into the vagus nerve where it is crossed by the superior belly of the omohyoid muscle. The site of injection was then dried with a cotton bud. Control rats received equal volume of vehicle injection. The rats were allowed to survive for 4, 6, 15, 22, 30 and 60 days. At least two rats were perfused at each of the above time intervals with a mixed aldehyde solution composed of 2% paraformaldehyde and 3% glutaraldehyde in 0.1 M cacodylate buffer adjusted to pH 7.4. After perfusion which lasted for 30 min, the brainstem containing the DMN of vagus was removed. Transverse vibratome sections of 100 mm thickness were prepared. They were post®xed in
0304-3940/99/$ - see front matter q 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 9 9) 00 49 0- 5
154
Y.-K. Ng, E.-A. Ling / Neuroscience Letters 270 (1999) 153±156
2% osmium tetroxide in 0.1 M cacodylate buffer. After dehydration in alcohol, the tissue slices were embedded in an Araldite mixture. Ultrathin sections were cut, double stained with uranyl acetate and lead citrate and examined in a JEOL 1200EX eletron microscope. Immunohistochemical study was con®ned to rats killed at 4 and 6 days since massive in®ltration of mononuclear cells was observed at these two time points. The rats were perfused ®rst with Ringer's solution followed by an aldehyde ®xative composed of a mixture of periodate-lysineparaformaldehyde containing 2% paraformaldehyde as described previously [12]. Following removal, the brainstem was post-®xed in the same ®xative for 2±4 h and was kept in 0.1 M phosphate buffer containing 15% sucrose overnight at 408C. Serial coronal frozen sections were cut at 40 mm thick and incubated in one of the following monoclonal antibodies diluted in 1:100 with phosphate buffered saline: ED1 (Harlan SeraLab, MAS-341), OX19 (Harlan SeraLab, MAS-099b), and OX33 (Harlan SeraLab, MAS258b) monoclonal antibodies for detection of macrophages, CD5 equivalent T-cells and CD45RA B-cells, respectively. After incubation, the sections were rinsed in PBS for 15 min and then reacted in Vectastain ABC Kit (PK4002, Vector Laboratories) against mouse IgG for half an hour in a humidi®ed chamber. After this, the sections were rinsed in PBS and treated with a solution of 3-3 0 -diaminobenzidine tetrahydrochloride (DAB, Sigma-5637) and hydrogen peroxide (5 mg DAB, 10 ml PBS, 30 ml 0.3% H2O2). The sections were counterstained with 1% methyl green, dehydrated and mounted in Permount. In agreement with previous studies [7±9], at 4 and 6 days after RCA-60 injection into the right cervical vagus, massive in®ltration of mononuclear leucocytes was observed in the ipsilateral DMN of the vagus. Most of the in®ltrated cells were macrophages and T-lymphocytes, as evidenced by their immunoreaction with OX-19 (Fig. 1) and ED-1 (Fig. 2), respectively. The in®ltrated ED-1 and OX-19 immunoreactive (IR) cells were distributed throughout the ipsilateral DMN, with some of them distributed in the ventral part of the nucleus tractus solitarius (NTS). ED-1IR monocytes were frequently lodged in the vascular walls and their vicinity (Fig. 2). Some OX-19 positive cells were closely associated with the blood vessels (Fig. 1). Few to moderate numbers of OX-33-IR cells were also scattered in the ipsilateral DMN (Fig. 3). In control rats receiving vehicle injection, the above-mentioned immunopositive cells were absent. At 4 and 6 days after RCA injection, some lymphocytes identi®ed by their small nucleus with large chromatin masses and scanty cytoplasm were seen to penetrate the external basal lamina (Fig. 4) to enter the neuropil. In the latter, the neuronal somata were often surrounded by lymphocytes (Fig. 5). A variable number of cytoplasmic processes of lymphocytes appeared to deeply penetrate the neuronal somata. In some section pro®les, as many as up to ®ve lymphocytes were seen in a neuronal soma (Fig. 6). In
what may be an early stage of in®ltration of neuronal soma by lymphocytes, the immediate neuronal cytoplasm surrounding the invaded lymphocytes appeared dissoluted, as indicated by its absence of organelles (Fig. 7). Free degenerating lymphocytes were not uncommon in the neuropil. Sometimes, they were phagocytosed by macrophages which were characterized by an irregular or ¯attened nucleus with margination of chromatin (Fig. 8). Plasma cells in close apposition to neuronal soma were common (Fig. 9). They were characterized by the extensive pro®les rough endoplasmic reticulum. Injection of RCA into the vagus nerve selectively destroys the vagal efferent neurons as a result of the suicide retrograde transport of the toxic ricin. Associated with the neuronal breakdown was the massive in¯ux of mononuclear cells with a preponderance of lymphocytes. The in®ltration peaked at 4±6 days after the toxin injection, and the in®ltrated cells appeared to accumulate in the perivascular space before being dispatched to the neuropil of DMN. The exact signal to attract the in®ltration of mononuclear cells after RCA injection is uncertain. Since in®ltration of mononuclear cells was absent in the control rats receiving vehicle injection, it must be attributed to the injected ricin taken up by the DMN efferent neurons. It is possible that minute amounts of RCA may have been released from dying neurons and this may alter the integrity of blood-brain barrier in the DMN, and serve as the speci®c chemotactic factor for the attraction of mononuclear cells. Leakage of RCA may also have elicited secretion of cytokines or upregulation of adhesive molecules from local glial or vascular elements. This would facilitate the migration of lymphocytes, macrophages and plasma cells into the perivascular space as well as in the neuropil of the DMN and then trigger off a local immunological response. The most unusual feature in the DMN after RCA injection was the penetration of a variable number of lymphocytes into some of the surviving vagal efferent neurons. The internalization of lymphocytes into vagal neurons is reminiscent of `emperipolesis' described in other tissues [2,6,15]. In vitro studies have shown the capability of lymphocytes invading mixed cultures of different tissues especially of mesenchymal origin [5,6]. In biopsies of human melanoma, lymphocytes were seen within the cytoplasm of non-pigmented, nucleated melanoma cell [2]. The exact mechanism of internalization or emperipolesis of lymphocytes into some DMN neurons remains unclear. It is speculated that ricin uptake could have induced the expression of speci®c receptors or adhesive molecules on the neuronal surface that might have attracted the lymphocytes, which were eventually internalized into the neurons. Interestingly, in human laryngeal papilloma, neutrophils in various stages of degeneration and disintegration were seen within the neoplastic cells [1]. In the present study, some of the vagal neurons containing intracellular lymphocytes showed long projecting dendrites; in some, the synaptic contacts associated with the soma remained structurally
Y.-K. Ng, E.-A. Ling / Neuroscience Letters 270 (1999) 153±156
155
Fig. 1. Photomicrograph of the caudal medulla of a rat, 4 days after RCA injection into the right vagus nerve. Note the massive in®ltration of OX-19-immunoreactive (IR) T-cells (arrows) in the ipsilateral dorsal motor nucleus of vagus (DMN) and nucleus tractus solitarius (NTS), but not in the hypoglossal nucleus (HN). Some immunopositive cells (arrowheads) are closely associated with the blood vessels. V, fourth ventricle. Scale bar, 100 mm. Fig. 2. Massive in®ltration of ED1-IR cells in the ipsilateral DMN and part of the NTS, but not in the HN, 4 days after RCA injection. Some immunoreactive monocytes (arrows) appear to be lodged in the vascular walls and their vicinity. V, fourth ventricle. Scale bar, 100 mm. Fig. 3. Photomicrograph showing the caudal medulla, 4 days following RCA injection into the right vagus nerve. OX-33-IR cells (arrows) are scattered in the ipsilateral DMN across in the mid-line. V, fourth ventricle. Scale bar, 100 mm. Fig. 4. A lymphocyte (*) herniates through the external basal lamina (EBL) into the neuropil of DMN, 6 days after RCA injection. Scale bar, 0.5 mm. Fig. 5. Five in®ltrated lymphocytes (*) are in close association with the soma of a DMN neuron (N). The soma is postsynaptic to two axon terminals (circled area), 6 days after RCA injection. Scale bar, 1 mm. Fig. 6. Five lymphocytes (*) are entrapped in the cytoplasm of a DMN neuron (N), 6 days after RCA injection. The intraneuronal lymphocytes are at different stages of lysis. Scale bar, 1 mm. Fig. 7. A lymphocyte is within the cytoplasm of a neuron (N). The neuronal cytoplasm in the immediate vicinity (*) of the lymphocyte is free from any organelles (lysis of cytoplasm). Circled area, synaptic contacts; 6 days after RCA injection. Scale bar, 0.5 mm. Fig. 8. A degenerating cell in the cytoplasm of a macrophage (Ma). A lymphocyte (*) is partially engulfed by the macrophage, 6 days after RCA injection. Scale bar, 0.5 mm. Fig. 9. A plasma cell (P) capping the soma of a DMN neuron (N) containing a degenerating cell presumably lymphocyte (asterisk). Circled area, synaptic contact, 6 days after RCA injection. Scale bar, 1 mm.
156
Y.-K. Ng, E.-A. Ling / Neuroscience Letters 270 (1999) 153±156
intact suggesting that the affected cells remain viable. Why some neurons survived after being invaded by lymphocytes is also dubious. One possible explanation may be that the minute amounts of ricin was at too low levels to kill the neurons, but was at toxic levels to kill the internalized lymphocytes. The signi®cance of the penetration of lymphocytes into the neuronal soma remains enigmatic. It could be an attempt for the invaded T cells to lyse the DMN neurons which were recognized as non-self because of their contents of ricin. In other words, emperipolesis may represent an effector-target cell contact leading to cytotoxicity. The dissolution of the cytoplasm immediately surrounding the entrapped lymphocytes indicates the possibility of release of cytotoxic substances. The obligatory role of physical contact between cytotoxic lymphocytes and target cells has been stressed in both the NK cells, as well as the antibody-dependent, in vitro systems [4,6,10]. In the present study, the emperipolesis of lymphocytes into vagal neurons probably represents a more effective mode of destroying the ricin-poisoned cells. The fate of the entrapped lymphocytes seems unequivocal as shown by their destruction within the neuronal cytoplasm, but the mechanism of their destruction remains unclear. It is speculated that they were poisoned by ricin contained in the neuronal cytoplasm. Against this view is the fact that in some host cells, the cytoplasmic organelles, e.g. rough endoplasmic reticulum remained structurally intact. If ricin were to be present at a toxic level, one would expect the disruption of rough endoplasmic reticulum since it is known to inhibit protein synthesis [3,11,14]. Interestingly some lymphocytes which did not penetrate the neurons degenerated in the interstitial spaces and were removed by macrophages. It is likely that some free ricin might have been released by vagal neurons that had earlier picked up the ricin. The occurrence of plasma cells in the neuropil suggests their involvement in the neutralization of the toxic ricin. While the emperipolesis of lymphocytes in vagal neurons was a consistent feature, it is remarkable that neither macrophages nor plasma cells penetrated the neurons suggesting the cellular speci®city of this unique process in the nervous tissue. The non-af®nity of macophages and plasma cells to the affected neurons could be due to absence of some speci®c cell-surface receptor molecules, which may be expressed only on T cells. The work is supported by Research Grants (RP950363
and RP970328) from the National University of Singapore. The authors wish to thank Mr. Tajuddin B. Marican, Mrs. Yong Eng Siang, Mrs. Ng Geok Lan, Mr. Yick Tuck Yong, Miss. Chan Yee Gek and Mr. Gobalakrishnan for their expert technical assistance. [1] Ahmed, M.M., Studies on human laryngeal papilloma. Acta Oto-Laryngol., 92 (1981) 563±567. [2] Burns, E.R., Zucker-Franklin, D. and Valentine, F., Cytotoxicity of natural killer cells. Correlation with emperipolesis and surface enzymes. Lab. Invest., 47 (1982) 99±107. [3] Endo, Y., Mitsui, K., Motizuki, M. and Tsurugi, K., The mechanism of action of ricin and related toxic lectins on eukaryotic ribosomes. J. Biol. Chem., 262 (1987) 5908±5912. [4] Glauert, A.M. and Sanderson, C.J., The mechanism of Kcell mediated cytotoxicity III. The ultrastructure of K-cell projections and their possible role in target cell killing. J. Cell Sci., 35 (1979) 355±366. [5] Hughes, D., Raine, C.S. and Field, E.J., Invasion of neurons in vitro by nonimmune lymphocytes: an electron microscopic study. Br. J. Exp. Pathol., 49 (1968) 356±359. [6] Ioachim, H.L., Emperipolesis of lymphoid cells in mixed cultures. Lab. Invest., 14 (1965) 1784±1794. [7] Ling, E.A. and Leong, S.K., Effect of intraneural injection of ricinus comrnunis agglutinin-60 into the rat vagus nerve. J. Neurocytol., 16 (1987) 373±387. [8] Ling, E.A. and Leong, S.K., In®ltration of carbon-labelled monocytes into the dorsal motor nucleus following an intraneural injection of ricinus communis agglutinin-60 into the vagus nerve in rats. J. Anat., 159 (1988) 207±218. [9] Ling, E.A., Shieh, J.Y., Wen, C.Y., Chan, Y.G. and Wong, W.C., Degenerative changes of neurons in the superior cervical ganglion following an injection of ricinus communis agglutinin-60 into the vagus nerve in hamsters. J. Neurocytol., 19 (1990) 1±9. [10] Ling, N.R., Acton, A.B., Roitt, I.M. and Doniach, D., Interaction of Iymphocytes from immunized hosts with thyroid and other cells in culture. Br. J. Exp. Pathol., 46 (1965) 348±359. [11] MacConnel, W.P., Eurman, D.W. and Kaplan, N.O., The action of ricinA-chain on ribosomes. Biochem. Biophys. Res. Commun., 108 (1982) 809±814. [12] Ng, Y.K. and Ling, E.A., Induction of major histocompatibility class II antigen on microglial cells in postnatal and adult rats following intraperitoneal injections of lipopolysaccharide. Neurosci. Res., 28 (1997) 111±118. [13] Oeltmann, T.N. and Wiley, R.G., Wheat germ agglutininricin. A-chain conjugate is neuronotoxic after vagal injection. Brain Res., 377 (1986) 321±328. [14] Olsnes, S., Refsnes, K. and Pihl, A., Mechanism of action of the toxic lectins abrin and ricin. Nature, 249 (1974) 627±631. [15] Schelton, E. and Dalton, A.J., Electron microscopy of emperipolesis. J. Biophys. Biochem. Cytol., 6 (1959) 513. [16] Wiley, R.G., Blessing, W.W. and Reis, D.J., Suicide transport: destruction of neurons by retrograde transport of ricin, abrin and modeccin. Science, 216 (1982) 889±890.