- Email: [email protected]
The retrograde axonal transport of nerve growth factor
Brain Research, 68 (1974) 103-121 © ElsevierScientificPublishingCompany,Amsterdam- Printed in The Netherlands
THE RETROGRADE FACTOR
AXONAL
TRANSPORT
OF
NERVE
103
GROWTH
I. A. HENDRY, K. STt)CKEL, H. THOENEN AND L. L. IVERSEN
Department of Pharmacology, Biocenter of the University, Basel (Switzerland) and (L.L.L) MRC Neurochemical Pharmacology Unit, Department of Pharmacology, University oJ" Cambridge, Cambridge (Great Britain) (Accepted August 30th, 1973)
SUMMARY
A retrograde axonal transport of nerve growth factor (NGF) from the adrenergic nerve terminals in the mouse iris to the cell bodies of postganglionic sympathetic neurones in the superior cervical ganglion has been demonstrated. After injection of iodinated nerve growth factor (125I-NGF) into the anterior eye-chamber there was a relatively rapid accumulation of radioactivity in the superior cervical ganglia on both injected and non-injected sides, as was the case after subcutaneous injection. However, 4 h after intraocular injection a preferential accumulation of radioactivity became apparent in the superior cervical ganglion on the injected side, and this difference between the ganglia on injected and non-injected sides gradually increased to a maximum at 16 h. Transection of the postganglionic adrenergic fibres as well as the prior intraocular injection of colchicine abolished the preferential accumulation of 125I-NGF in the superior cervical ganglion of the injected side, whereas the destruction of adrenergic nerve terminals by 6-hydroxydopamine did not impair the preferential accumulation. It is concluded that the retrograde axonal transport of NGF, which was estimated to take place at a rate of about 2.5 mm/h, depends on a colchicine-sensitive mechanism as does the orthograde rapid axonal transport. However, the uptake of NGF may not only take place from the nerve terminals but also from the preterminal parts, as has been shown in other studies with horseradish peroxidase. Autoradiographic studies strongly supported the existence of a retrograde transport by showing a clear localization of radioactivity in a small number of neurones in the superior cervical ganglion on the injected side, whereas on the non-injected side there was only a diffuse distribution of radioactivity throughout the ganglion.
104
i.A. HENDRY et al.
INTRODUCTION
The mouse submaxillary gland protein nerve growth factor (NGF) has been known for many years to have profound growth-promoting properties on the sympathetic nervous system of neonatal animals 16. It has more recently been shown that the administration o f large doses of N G F to developing mice 10 and rats 22 induces a selective increase in the activity of tyrosine hydroxylase (TH) and dopamine fl-hydroxylase (DBH), key-enzymes in the synthesis of noradrenaline (NA) in adrenergic neurones. These effects and the converse actions of N G F antiseruml6, ~v, suggest that N G F may be essential for the normal development o f adrenergic neurones in the peripheral sympathetic nervous system. Recently it has also been shown that even in adult animals a decrease in the tissue and plasma levels of N G F which occurs after sialectomy or castration, is also followed by a decrease in the activity of T H H and other enzymes involved in NA biosynthesis 23 in mice sympathetic ganglia. This effect could be reversed by administration of N G F z3. Thus, it would appear that N G F is essential for the maintenance of a normal function of the peripheral adrenergic neurones throughout the entire life of the animals. In neonatal mice the unilateral removal of a submaxillary gland produced a reduction in T H activity in the ipsilateral superior cervical ganglion. This effect could be reversed by the local administration of a cellulose-bound slow-release form of N G F at the site of the removal of the gland. It was suggested that a possible mechanism for this action might be a retrograde transport of N G F to the neuronal cell bodies via the adrenergic axons H. The phenomenon o f retrograde axonal transport of proteins has been described for several different neuronal populations: for example horseradish peroxidase from the optic rectum to the ganglion cells in the eyOS; horseradish peroxidase and albumin from the tongue to the hypoglossal nucleus 14. It is possible that such retrograde transport mechanisms may have a special significance in providing the neuronal cell body with a means of sampling the external environment around its axon terminals, thereby enabling the neurone to respond to trophic factors in this environment. N G F may represent such a trophic factor, and it was the purpose of the present study to provide direct evidence for the retrograde axonal transport of this biologically important compound. MATERIALS AND METHODS
N G F was prepared from mouse submaxillary glands by the method of Bocchini and Angeletti 3 and was labelled with 125iodine by a modification of the method of Greenwood et al. s. In order to iodinate all the N G F molecules we used a mixture of sodium 125iodide (supplied by the Radiochemical Centre Amersham, Bucks., England) and sodium a27iodide (supplied by Merck, Darmstadt, Germany). The reaction mixture contained 650 #g (24.5 nmoles) NGF, 2 mCi sodium 125iodide, 11.3 #g (75 nmoles) sodium 127iodide and 28.2 #g (100 nmoles) chloramine-T in a total volume of 400 /~1 of Tris-HC1, pH 8. The reaction took place at room temperature and was stopped after 30 min with 19 /tg (I00 nmoles) sodium metabisulphite (Na2S205) in
RETROGRADE TRANSPORT OF
NGF
105
100 #1 Tris-HC1 buffer, pH 8. The labelled NGF was separated from unreacted iodine by use of a Bio-Gel P2 (supplied by BIO-RAD Laboratories, Richmond, Calif., U.S.A.) column (0.9 cm × 15 cm) equilibrated with 0.1 M acetic acid followed by dialysis against several changes of saline. This fully iodinated NGF had a specific activity of 0.73 #Ci/#g. For the autoradiographic study iodinated NGF with a higher specific activity of 10 #Ci/#g was prepared by replacing the sodium 127iodide by sodium 125iodide. In this experiment the reaction mixture contained 66 /~g NGF (2.5 nmoles), 10 #Ci sodium 125iodide (5.7 nmoles), 8.46 big chloramine-T (30 nmoles) in a total volume of 80 #1 Tris-HCl, pH 8. The reaction was stopped after 30 min with 9.5 big (50 nmoles) Na2S205 in 50 #1 Tris-HCl buffer, pH 8, followed by dialysis against several changes of saline. Subsequently the lz5I-NGF was directly used for injection. Both labelled preparations of N G F retained their full biological activity in the chick dorsal root ganglion bioassay of Fenton 6. For all experiments male albino mice of the Swiss strain were used. Their body weight ranged between 30-35 g. The animals were kept at a constant temperature of 23 ~- 1 °C and they were supplied with NAFAG-pellets (supplied by NAFAG AG, Gossau SG, Switzerland) and water ad libitum. For the injection of labelled NGF into the anterior chamber of the eye, the animals were anaesthetised with ether, and in each case 0.25 #g of the compound was injected in 2 #1 of saline buffered with sodium phosphate at pH 7.4 using a Hamilton glass syringe. Subcutaneous injections were made at a site low on the back, in order to be as far removed from the superior cervical ganglion as possible. Surgical transection of the postganglionic nerve fibres of the superior cervical ganglion was carried out using a dissection microscope under ether anaesthesia. For the determination of 125I-NGF in superior cervical ganglia after intraocular or subcutaneous injection the animals were killed with ether and the ganglia dissected using a stereoscopic microscope. The ganglia were then rinsed in saline and the radioactivity of the intact tissue determined in a Packard gammacounter at a counting efficiency of approximately 50 ~. For autoradiographic studies of the localization of 125I-NGF in superior cervical ganglia the ganglia were fixed in 2.5 ~ glutaraldehyde in 0.1 M sodium phosphate buffer, pH 7.4 (v/v), post-fixed in osmium tetroxide (1 9/o w/v) in phosphate buffer, dehydrated in ethanol and embedded in an epoxy resin. Thin sections of the embedded tissue (1-2/,m) were mounted on glass slides, dipped in Ilford L 4 emulsion and stored in the dark at 4 °C for 28 days prior to development with Kodak D 19b developer. The tissue sections were examined by phase contrast microscopy, and autoradiographic activity was localized by dark field microscopy. Further experiments designed to determine the proportion of radioactivity present in the superior cervical ganglia in the form of antigenically intact NGF were carried out as follows: I~5I-NGF (0.25 /zg) was injected into the right eye. Twelve hours after the injection the animals were killed and the superior cervical ganglia of the injected and non-injected sides removed. In view of the small amounts of radioactivity present in the tissues the ganglia of 5 animals were pooled, homogenized in 2 ml of 0.05 M veronal buffer, pH 8.6 containing 0.9 ~o NaCI, 0.1 9/0sodium azide and
[. A. HENDRY et al.
106
At.,
o =
360 c 300 t~ \tm 2 4 0 '-: E
i80.
7:;1:1::c:2%
4i
d_ d 120 • 60
. . . . . . ,/-"~.=-Iz4~ 4 8 12 16 20 24 36 48
~5
2
injection (hours)
Time offer
cO
420-
B
m 360 L i/[''
300E d.
180 -
.~
120-
~
right side
/I' "% -
240 -
I
~-~left side
12
24
600
Time offer injection
48 (hours)
Fig. 1. A: time course of total accumulation of '2Sl in each superior cervical ganglion after unilateral intraocular injection (for details see Methods) of 0 . 2 5 fig a 2 5 I - N G F . E a c h point represents the mean i S . E . M . (vertical bars) for groups of 5-7 animals. B: time course of accumulation of lzsI in superior cervical ganglia after subcutaneous injection of 0.25 p g o f 1 2 S I - N G F . Each point represents the mean ± S . E . M . (vertical bars) from groups of 6 animals.
0.5 % bovine albumin. This homogenate was incubated with solid phase N G F antibodies 9 and slowly mixed on a rotary mixer for 24 h at 4 °C. After this the immunoadsorbent paper was washed free of unbound radioactivity and the bound radioactivity was measured. The percentage of the total radioactivity binding to the immunoadsorbent was taken to represent antigenically intact NGF. In those experiments in which the effect of the destruction of adrenergic nerves on the retrograde transport was studied the animals were injected intravenously on 3 consecutive days with 0.32 mmoles/kg 6-hydroxydopamine hydrochloride (supplied by Fluka, Buchs SG, Switzerland). For intraocular injection of colchicine (supplied by Koch-Light Laboratories Ltd., Colnbrook, Bucks., England) a solution of 20 mg/ml was used. Two #1 of this solution were injected into the anterior eye-chamber at various times prior to the injection of t25I-NGF. RESULTS
Accumulation of radioactivity in superior cervical ganglia after intraocular and subcutaneous injection of t~SI-NGF After the intraocular injection of 0.25 fig of t25I-NGF there was a relatively rapid accumulation of radioactivity in superior cervical ganglia on both injected and non-injected sides, which reached a maximum after 6 h (Fig. 1A). However, as early
RETROGRADE u E o
TRANSPORT
oF NGF
107
E ~.
.~_ ~
I~: ~> ~
60 ~4 Time
3~ 48
injection
(hours)
B
g2 "T
g E
o
8;2,~2o2. offer
"6 .---
109+ 8" 7654 3 2 1
j/ ....... 6 1012
Time a f t e r
,..
6
it--,.-/]
.
20 2A 36 /.8
injection
[hours)
Fig. 2. A: difference in 12'~Iaccumulation between superior cervical ganglia from injected (intraocular injection of 0.25/~g of 125I-NGF) and non-injectedside. B: ratio of 1251accumulation between the ganglia from the injected and the non-injected side. as 4 h after the injection there was a statistically significant (P < 0.05) difference between the amount of radioactivity accumulated in the ganglia on the injected and non-injected sides (Figs. 1A, 2A). This difference increased gradually and reached a maximum after 16 h. The differential accumulation becomes even more impressive if it is expressed in terms of the ratio between injected and non-injected side which reaches a value of about 3 after 12 h, 3.5 after 24 h and 9 after 48 h (Fig. 2B). It is likely that the initial rapid increase in radioactivity in both ganglia is due to 125I-NGF which escaped from the anterior eye-chamber into the general circulation and reached the ganglia via the blood stream. This assumption is strongly supported by the observation that the time-course of the initial accumulation of radioactivity in both superior cervical ganglia is very similar after intraocular and subcutaneous (Fig. 1) injections of 125I-NGF.
Effect of transeeting the postganglionie adrenergiefibres and intraocular administration of colehicine Although the initial rapid increase in radioactivity in both superior cervical ganglia after intraocular injection is due to blood-borne lzSI-NGF, the slowly increasing difference between left and right sides is probably due to a retrograde transport of labelled material from the adrenergic nerve terminals in the eye to the cell bodies in the superior cervical ganglia. Since the only known neural connections between the eye and the superior cervical ganglion are postganglionic sympathetic fibres, their transection should abolish the differential accumulation of radioactivity
108
I.A. HENDRY et al.
between the injected and non-injected sides. This indeed was the case. Surgical transection o f the postganglionic nerve fibres emanating f r o m the superior cervical ganglion completely abolished the difference in the accumulation o f 125I-NGF between left and right ganglia normally observed 12 h after intraocular injection (Table 1). The fact that there was no significant (P > 0.05) difference in the accumulation o f radioactivity between the operated and unoperated ganglia 2 h after the subcutaneous injection o f 125I-NGF (Fig. 3) shows that the operation had little direct effect on the ability o f the ganglia to accumulate blood-borne N G F . Hence it follows that intact postganglionic adrenergic nerve fibres seem to be a prerequisite for the preferential accumulation o f 125I-NGF in the superior cervical ganglion on the injected side. The rate o f the retrograde axonal transport o f N G F f r o m the adrenergic nerve terminals to the cell b o d y seems to be rather rapid as far as can be estimated by an approximate calculation. This calculation is based on the assumption that the distance between the eye and the ganglion is about 1 cm and that the first significant difference in radioactivity between the ganglia on the injected and non-injected sides after 4 h represents the first moiety o f 125I-NGF reaching the ganglion o f the injected side by retrograde axonal transport. This calculated rate of 2.5 m m / h is similar to the 3 m m / h reported for the transport o f horseradish peroxidase f r o m the chick optic tectum to the retinal ganglion cells by La Vail and La VaiP 5. This rate o f transport is also in the range o f the fast orthograde transportT, 19 which seems to depend on intact neurotubules and can be blocked by colchicine2, 5. It was, therefore, o f interest to determine the effect o f colchicine on the retrograde axonal transport o f a25I-NGF. Intraocular injection o f
TABLE I EFFECTS OF COLCH1CINE AND TRANSECTION OF THE POSTGANGLIONIC NERVE FIBRES ON THE DIFFERENTIAL ACCUMULATION OF 125I
Groups of 5-6 animals were injected into the anterior chamber of right eye with 0.25/~g of 12~I-NGF in 2/zl of saline at the appropriate time after treatment with colchicine or surgery. Both superior cervical ganglia were removed 12 h after the intraocular injection of le5I-NGF and the total radioactivity of the intact tissues was measured in a Packard gamma-counter. The difference in the amount of 1~5Iaccumulated in the two ganglia was then expressed as the mean ± S.E.M. of the individual differences in counts/min. Treatment
Albumin in saline Transection of postganglionic nerve fibres Colchicine 2/~1 (20 mg/ml)
Time be]bre 1251-NGF injection (h)
Dtfference in ganglionic 1251 (counts~rain)
2 0
217 ± 22 10 ± 34'**
0 2 12
229 ± 42* 70 ± 31 ** 25 -~: 37***
* Differs from 0, P < 0.005, but not from control animals, P :> 0.05. ** Differs from 0 and control animals, P < 0.05. *** Differs from control animals, P < 0.01, but not from 0, P ~ 0.05.
RETROGRADE TRANSPORT OF N G F
109
~m
intact
I
I
transected t"
.9
O. O
oo
I
Ot1' ÷÷
O÷
control 6-0HDA after 2h
control 6-0HDA after 16h
]Fig, 3. Effects of 6-hydroxydopamine and surgical transection of the postganglionic nerve trunk on
the accumulation of 1251in the superior cervical ganglia after subcutaneous injection of 1 ktg of I~sINGF. The animals were killed 2 or 16 h after injection with 125I-NGF. Transection of the nerve fibre was carried out 24 h before the animals were sacrificed. 6-Hydroxydopamine treatment consisted of injecting the animals intravenously on 3 consecutive days with 0.32 mmoles/kg. The last dose was given 24 h before the injection of 125I-NGF. Results are expressed as means ~ S.E.M. (vertical bars) for groups of 5-7 animals. * Left and right ganglia do not differ from each other; P < 0.1. ** Left and right ganglia differ from each other; P < 0.005. + Differs from corresponding ganglia in control animals; P < 0.5. ++ Differs from corresponding ganglia in control animals; P < 0.01.
4 0 / t g o f colchicine 12 h before that o f 0.25/zg 125I-NGF completely abolished the differential accumulation in the superior cervical ganglion (Table I). However, if colchicine was injected at the same time as 125I-NGF then the differential accumulation was maintained, suggesting that colchicine requires some time to exert its action on the axonal transport mechanism, possibly by diffusion to the preterminal parts o f the adrenergic axon since destruction o f the nerve terminals does not seem to impair materially the retrograde transport, as will be shown below in the experiments with 6-hydroxydopamine.
Destruction of adrenergic nerve terminals by 6-hydroxydopamine: effect on retrograde axonal transport of 1251-NGF In order to determine whether intact adrenergic nerve terminals are essential for the retrograde axonal transport o f 125I-NGF a further series o f animals was treated with 6 - h y d r o x y d o p a m i n e which is k n o w n to destroy adrenergic nerve terminals selectively in adult animals 24. The animals were injected intravenously on 3 consecutive days with 0.32 mmoles/kg o f 6-hydroxydopamine. Twenty-four hours after the last dose they received an intraocular injection o f 0.25/~g o f 125I-NGF. In spite o f the destruction o f the adrenergic nerve terminals by 6-hydroxydopamine there was a significant difference in the accumulation o f radioactivity between the ganglia o f the injected and non-injected side, indicating that the destruction o f the nerve terminals
110
i.A. HENDRYet al.
TABLE 1I ACCUMULATION OF 12,5] IN SUPERIOR CERVICAL GANGLIA AFTER INJECTION OF
1251-NGFINTO THE ANTE-
RIOR CHAMBER OF THE RIGHT EYE
Animals were injected into the anterior chamber of the right eye with a dose of 0.25/Ag of ~251-NGF. The left and right superior cervical ganglia were removed after 4 or 16 h and the radioactivity was measured in a gamma-counter. The values given represent the mean ± S.E.M. of counts/rain/ganglion for groups of 5 animals. 6-Hydroxydopaminetreatment consisted of injecting the animals intravenously on 3 consecutivedays with 0.32 mmoles/kg. The last dose was given 24 h before the injection of 1251-NGF. Time after injection ( h) 4 16
Treatment
Right ganglion ( counts/min)
Left ganglion ((counts~rain)
Control 6-OHDA Control 6-OHDA
276 i 194 ± 120 , 87 ,
238 ± 189 , 50 -42 ,
25.2 18.1" 10.0 4.2**
9.6 12.8" 8.9*** 11.6"**
* Differs from untreated control animals, P < 0.05. ** Differs from untreated control animals, P < 0.005. *** Differs from the right ganglion, P < 0.001.
did not impair the retrograde transport of 125I-NGF (Table II). However, the accumulation of radioactivityin both ganglia was less than that observed in untreated animals, suggesting that the 6-hydroxydopamine treatment may have had some direct effect on the ganglionic cell bodies. The assumption of a direct damaging effect to the cell bodies by 6-hydroxydopamine is supported by the observation that as early as 4 h after the intraocular injection of 125I-NGF (when a small contribution from the retrograde axonal transport to the radioactivity accumulated in the ganglion was to be expected) the radioactivity measured in superior cervical ganglia of 6-hydroxydopamine treated animals was considerably lower (P < 0.05) than that in untreated controls (Table II). In order to provide further evidence of a direct effect of 6-hydroxydopamine on the accumulation of blood-borne 125I-NGF by the adrenergic cell body, the effect of this drug was examined in animals in which the postganglionic fibres had been cut. A subcutaneous injection of l #g of 12~I-NGF was given and the accumulation of radioactivity in the superior cervical ganglion measured. As in ganglia of untreated animals there was no difference between the radioactivity accumulated 2 h after injection by ganglia with intact and transected postganglionic fibres. However, there was a significantly (P < 0.05) smaller accumulation of radioactivity in transected and intact ganglia of animals treated with 6-hydroxydopamine as compared to untreated animals (Fig. 3). Sixteen hours after the injection of 12SI-NGF there was much less radioactivity accumulated in the ganglia with transected postganglionic fibres both in 6-hydroxydopamine treated animals and untreated controls. However, the accumulation of radioactivity on both transected and intact sides was significantly (P < 0.01) lower in 6-hydroxydopamine treated animals than in untreated controls, whereas the ratio be-
RETROGRADETRANSPORTOF NGF
111
tween transected and non-transected side was virtually the same in controls and 6hydroxydopamine treated animals. Thus, it can be concluded that intact nerve terminals are not a prerequisite for the retrograde axonal transport of 125I-NGF and that the reduced accumulation of radioactivity in the superior cervical ganglion after treatment with 6-hydroxydopamine is due to a direct damaging effect of this amine on the cell body, impairing the accumulation of blood-borne lz~I-NGF. Furthermore, these experiments have also shown that a considerable part of 125I-NGF injected subcutaneously accumulated in the superior cervical ganglion by way of retrograde axonal transport.
Autoradiographic studies of superior cervical ganglia after intraocular injection of 125I-NGF Autoradiographic examination of ganglia taken from the injected and noninjected sides of animals that had received intraocular lzSI-NGF 12 h previously, strongly supported the view that retrograde axonal transport of the labelled material had occurred on the injected side. In superior cervical ganglia from the injected side there was a clear localization of radioactivity in a small number of ganglionic neurones, together with a diffuse background radioactivity throughout the ganglion. In ganglia from the non-injected side only the diffuse labelling was seen, with no clear localization over specific neurones (Figs. 5, 6). The labelled neurones on the injected side were often seen to contain most radioautographic activity over the cytoplasm, with relatively little labelling in the nucleus (Fig. 6). Similar results were observed in ganglia from both sides of 4 animals. No accurate estimate was made of the proportion of neurones that became intensely labelled after injection of 125I-NGF into the eye, but the proportion was almost certainly less than 10 ~ of all ganglionic neurones (Fig. 4).
Determination of antigenically intact 125I-NGF in the superior cervical ganglia after intraocular injection The half-time of circulating exogenously administered 125I-NGF is very short (20-30 min) and a considerable part of the radioactivity present in the plasma after intravenous injection of 125I-NGF represents degradation products and not intact NGF 23. Therefore, it seemed to be important to determine the proportion of the radioactivity present in the ganglia as intact NGF. Because of the extremely low amount of NGF accumulated in the ganglia it was not possible to determine its biological activityL Thus, an immunoadsorption method was used, involving the use of solid phase NGF antibodies previously described for a radioimmunoassay for NGF 9. This procedure estimated the proportion of azSI-NGF which remained bound to at least one antigenic site of the NGF molecule, but it did not allow us to determine the exact amount of biologically active NGF present in the tissues. However, any radioactivity that did not bind to the immunoadsorbent was almost certainly devoid of biological activity. Using this method we found that in homogenates of ganglia from the injected side 60 ± 3 % of the total radioactivity bound to the solid phase antibodies when the ganglia were examined 12 h after the injection. This contrasts with a binding of only 35 ± 2 % of the radioactivity present in the ganglia from the non-injected side of the
112
I. A. HENDRY et a l
Fig. 4. Photomontage of phase contrast micrographs (A) and dark field autoradiograms (B) of mouse superior cervical ganglion from the ioected side of an animal which received an intraocular injection of 125I-NGF 12 h previously. A small number of intensely labelled cells are observed, together with diffuse background labelling of the ganglion. × 10t).
RETROGRADE TRANSPORT oF N G F
113
114
1. A. HENDRY et al.
Fig. 5. Photomontage of phase contrast micrographs (A) and dark field autoradiograms (B) as in Fig. 4, but of a superior cervical ganglion from the side opposite to that used for intraocular injection. × 100.
RETROGRADE TRANSPORT OF N G F
115
116
i . A . HENDRY et al.
Fig. 6. Phase contrast micrographs (A) and dark field autoradiograms (B) of corresponding microscopic fields from mouse superior cervical ganglia on the injected (1) and non-injected (2) sides from an animal that received an intraocular injection of 12H-NGF 12 h previously. Note that the labelled cell in the lower left hand corner of B1 c~rresponds to a neurone (A1), and that labelling is less intense over the nucleus. × 400.
RETROGRADE TRANSPORT OF N G F
117
118
i.A. HENDRY et al.
same animals. This difference in antigenically intact N G F between the two ganglia represents that proportion which is protected during retrograde transport. The circulating N G F is more rapidly metabolized. Indeed, at this time only 8 & 1 To of the radioactivity in the serum is bound to the solid phase antibodies, indicating that the majority of the radioactivity in the serum was not N G F but that there was a specific accumulation of radioactivity in the ganglia representing N G F which contained at least one intact antigenic site. DISCUSSION
After previous experiments had suggested that N G F may exert its physiological control of adrenergic neurones as a 'retrograde trophic factor' acting from the periphery on the perikaryon 11, the present experiments have provided evidence for the existence of such a retrograde axonal transport. This evidence is based on the observation that after intraocular injection of a25I-NGF there was a preferential accumulation of radioactivity in the superior cervical ganglion of the injected side. This preferential accumulation could be abolished by surgical transection of the postganglionic adrenergic fibres which are the only known neural connections between the eye and the superior cervical ganglion. The most direct evidence for retrograde axonal transport is provided by the autoradiographic studies which showed a very intense labelling of a few neuronal cell bodies in the superior cervical ganglion on the injected side. This selective accumulation of radioactivity over a few neuronal ceils is most plausibly explained by a retrograde transport from the few axons leading to the iris; any other form of perineuronal transport would be likely to result in a more diffuse distribution of labelled N G F in the ganglia. The absence of any such cells on the non-injected side eliminates the possibility that this selective labelling was due to the presence of a small proportion of ceils in the ganglion with a specific ability to take up NGF. In addition to this retrograde axonal transport of N G F there seems to exist a mechanism for the accumulation of blood-borne N G F in the sympathetic ganglia which is responsible for the rapid initial accumulation occurring after subcutaneous or intraocular injection. The mechanism of this accumulation of circulating N G F in sympathetic ganglia is not clear and it remains to be established whether the binding sites are localized at the surface of the neuronal membrane or whether circulating N G F is transported into the neuronal ceils. It is noteworthy that even after subcutaneous injection of 125I-NGF a considerable part of the radioactivity accumulating in the sympathetic ganglia seems to reach the adrenergic cell bodies by retrograde axonal transport. This can be deduced from the observation that 16 h after subcutaneous injection of 1251-NGF the radioactivity present in the superior cervical ganglia with transected postganglionic fibres amounted to about 5 0 ~ o f that on the intact side. The fact that the accumulation of radioactivity 2 h after subcutaneous injection did not differ between transected and intact ganglia eliminates the possibility that the reduced accumulation after 16 h results from an impairment of the mechanism responsible for the accumulation of circulating N G F as a consequence of axotomy. In contrast to the intactness of the axons, the integrity of the adrenergic nerve
RETROGRADE TRANSPORT OF N G F
119
terminals does not seem to be a prerequisite for the retrograde axonal transport of NGF. This can be concluded from the observation that the selective destruction of adrenergic nerve terminals by 6-hydroxydopamine does not impair the retrograde transport of I~5I-NGF, i.e. 16 h after intraocular injection of 125I-NGF the ratio between the radioactivity in the ganglia of the injected and non-injected sides was the same in 6-hydroxydopamine treated animals and controls. However, pretreatment with 6-hydroxydopamine reduced the amount of radioactivity present in superior cervical ganglia both 2 and 16 h after subcutaneous injection of I~I-NGF. This implies that the reduced accumulation of N G F most probably results from a direct damaging effect of 6-hydroxydopamine on the cell body impairing both the accumulation of NGF reaching the cell body by retrograde axonal transport and by the blood stream. Thus, the impairment of the accumulation of radioactivity in sympathetic ganglia after systemic administration of 125I-NGF in animals pretreated with 6hydroxydopamine cannot be taken as evidence for retrograde transport of N G F 1. It merely represents a manifestation of a damaging effect of 6-hydroxydopamine on the cell body. Such an effect was not unexpected since it had been shown in previous experiments that in adult rats 6-hydroxydopamine seems to have a direct damaging effect on the cell bodies of adrenergic neurones, since 6-hydroxydopamine treatment blocked - - at least transiently - - the trans-synaptic induction of tyrosine hydroxylase that normally occurred in the superior cervical ganglia after cold exposure or administration of reserpine4, 21. Since the destruction of adrenergic nerve terminals with 6-hydroxydopamine did not interfere with the retrograde transport of 12~I-NGF one has to assume that the uptake of NG F into the adrenergic axon is not limited to the nerve terminals. This is in agreement with the observation that horseradish peroxidase can penetrate into the axons of peripheral nerves through the Schwann cells 13. Furthermore, there is evidence that in the mouse iris regeneration of adrenergic nerve terminals starts very rapidly after destruction by 6-hydroxydopamine18. Thus, it is possible that uptake of N G F takes place at these sites of regeneration, and that the ability for rapid regeneration may be related to this uptake of NGF. Although the present experiments did not allow an accurate determination of the rate of retrograde axonal transport it seems that the values calculated for the mouse sympathetic nervous system (2.5 mm/h) are very close to those determined by La Vail and La VaiP 5 in the chick embryo (3 mm/h) and by Kristensson e t al. 14 in adult rabbits (2.5 mm/h). The latter workers have also shown that the retrograde transport was blocked by ischaemia. The values calculated for the rate of retrograde axonal transport are within the range of the rapid orthograde transport which depends on an intact energy metabolism 2° and on functionally intact neurotubules, i.e. it can be blocked by colchicine and vinblastine2, 5. In view of the many similarities between the orthograde and retrograde axonal transport it is likely that both these fast transport systems involve similar mechanisms. The question arises as to whether there are two independent transport systems for retrograde and orthograde transport, or whether the same transport system performs
120
i.A. HENDRYet al.
the transport in both directions, the direction of transport depending on the properties of the material to be transported. At least for the retrograde transport one could assume that all the substances transported are enclosed in vesicles - - as has been shown for horseradish peroxidasO 2,z5 - - and that the surface of these vesicles might determine the direction of the transport. The assumption of a retrograde transport within vesicles would possibly also explain how the 125I-NGF reaching the cell body by retrograde axonal transport is better protected from metabolic degradation than that reaching the cell body via the blood stream. The specificity for the macromolecules to be taken up by the nerve terminals and to be transported to the perikaryon does not seem to be very high. However, if this retrograde axonal transport should have any physiological significance for the function of the sympathetic neurones it should exhibit some degree of specificity. We believe that this specificity may be conferred by the nature of the molecules transported and not by the transport mechanism itself. Thus, N G F is able to exert its specific effects on the adrenergic neurones once it has been brought to the cell body by retrograde axonal transport. It has been shown that after axotomy there is a change in the enzyme pattern in the adrenergic neurone, with a decrease in the relative amounts of the enzymes involved in NA biosynthesis 4. Moreover, castration which leads to a fall in tissue and plasma NGF, also caused a similar decrease in the enzymes involved in NA biosynthesis 11. If N G F is one of the factors important for the normal function of an adrenergic neurone in contact with its end organ, then the retrograde axonal transport of this substance may provide a mechanism for its control. ACKNOWLEDGEMENTS
This work was supported by the Swiss National Foundation for Scientific Research (Grant No. 3.653.71) and by the U.K. Medical Research Council. I. A. Hendry was the recipient of a fellowship from the Postgraduate Medical Foundation of Sydney University.
REFERENCES 1 ANGELETTI, R. H., ANGELETTI, P. U., AND LEvI-MONTALCINI, R., Selective accumulation of 1251labelled nerve growth factor in sympathetic ganglia, Brain Research, 46 (1972) 421425. 2 BANKS, P., AND MAYOR, D.0 lntra-axonal transport in noradrenergic neurons in the sympathetic nervous system, Biochem. Soc. Symp., 36 (1972) 133-149. 3 BOCCHINI, V., AND ANGELETTI, P. U., The nerve growth factor: purification as a 30,000-molecularweight protein, Proc. nat. Acad. Sci. (Wash.), 64 (1969) 787-794. 4 BRIMIJOIN, S., AND MOLINOFF, P. B., Effects of 6-hydroxydopamine on the activity of tyrosine hydroxylase in sympathetic ganglia of the rat, J. Pharmacol. exp. Ther., 178 (1971) 417424. 5 DAHLSTR6M,A., Axoplasmic transport (with particular respect to adrenergic neurons), Phil. Trans. B, 261 (1971) 325-358. 6 FENTON,E. L., Tissue culture assay of nerve growth factor and of the specificantiserum, Exp. Cell Res., 59 (1970) 383-392. 7 GEFFEN,L. B., AND LIVETT,B. G., Synaptic vesicles in sympathetic neurons, Physiol. Rev., 5l (1971) 98-157.
RETROGRADE TRANSPORT OF N G F
121
8 GREENWOOD,F. C., HUNTER, W. M., AND GLOVER,J. S., The preparation of lalI-labelled human growth hormone of high specific radioactivity, Biochem. J., 89 (1963) 114-123. 9 HENDRY, 1. A., Developmental changes in tissue and plasma concentrations of the biologically active species of nerve growth factor in the mouse, by using a two-site radioimmunoassay, Biochem. J., 128 (1972) 1265-1272. 10 HENDRY, I. A., AND IVERSEN, L. L., Effect of nerve growth factor and its antiserum on tyrosine hydroxylase activity in mouse superior cervical sympathetic ganglion, Brain Research, 29 (1971) 159-162. 11 HENDRY,|. A., AND IVERSEN,L. L., Changes in tissue and plasma concentrations of nerve growth factor following removal of the submaxillary glands in adult mice and their effects on the sympathetic nervous system, Nature (Lond.), 243 (1973) 500-504. 12 HEUSER, E., FREEMAN,A. R., AND KASHNER, A., Stimulation-dependent alteration in peroxidase uptake at lobster neuromuscular junctions, Science, 173 (1971) 733-734. 13 KRISHNAN,N., AND SINGER, M., Penetration of peroxidase into peripheral nerve fibres, Amer. J. Anat., 136 (1973) 1-14. 14 KRISTENSSON,K., OLSSON, Y., AND SJOSTRAND,J., Axonal uptake and retrograde transport of exogenous proteins in the hypoglossal nerve, Brain Research, 32 (1971) 399-406. 15 LA VAIL, J. H., AND LA VAIL, M. M., Retrograde axonal transport in the central nervous system, Science, 176 (1972) 1416-1417. 16 LEvI-MONTALCINI, R., AND ANGELETTI, P. U., Nerve growth factor, Physiol. Rev., 48 (1968) 534-569. 17 LEVI-MONTALCINI,R., AND BOOKER, B., Destruction of sympathetic ganglia in mammals by antiserum to a nerve growth protein, Proc. nat. Acad. Sci. (Wash.), 46 (1960) 384-391. 18 MALMFORS,T., The effect of 6-hydroxydopamine on the adrenergic nerves as revealed by the fluorescence biochemical method. In 6-Hydroxydopamine and Catecholamine Neurons, NorthHolland, Amsterdam, 1971, pp. 47-58. 19 OCHS, S., Characteristics and a model for fast axoplasmic transport in nerve, J. NeurobioL, 2 (1971) 331-345. 20 OCHS, S., AND RANISH, N., Metabolic dependence of fast axoplasmic transport in nerve, Science, 167 (1970) 878-879. 21 THOENEN, H., Biochemical alterations induced by 6-hydroxydopamine in peripheral adrenergic neurons. In 6-Hydroxydopamine and Catecholamine Neurons, North Holland Publ. Co., Amsterdam, 1971, pp. 75-85. 22 THOENEN,H., ANGELETTI,P. U., LEvI-MONTALCINI,R., AND KETTLER, R., Selective induction by nerve growth factor of tyrosine hydroxylase and dopamine fl-hydroxylase in the rat superior cervical ganglia, Proc. nat. Acad. Sci. (Wash.), 68 (1971) 1598-1602. 23 THOENEN,H., HENDRY, I. A., STOCKEL,K., PARAVICINI,U., AND OESCH, F., Regulation of enzyme synthesis by neuronal activity and by nerve growth factor. In Dynamics of Degeneration and Growth in Neurons, Wenner-Gren Center International Symposium, in press. 24 THOENEN,H., AND TRANZER,J. P., Chemical sympathectomy by selective destruction of adrenergic nerve endings with 6-hydroxydopamine, Naunyn-Schmiedeberg's Arch. exp. Path. Pharmak., 261 (1968) 271-288. 25 ZACKS, S. I., AND SAITO, A., Uptake of exogenous horseradish peroxidase by coated vesicles in mouse neuromuscular junctions, J. Histochem. Cytochem., 17 (1969) 161-170.
THE RETROGRADE FACTOR
AXONAL
TRANSPORT
OF
NERVE
103
GROWTH
I. A. HENDRY, K. STt)CKEL, H. THOENEN AND L. L. IVERSEN
Department of Pharmacology, Biocenter of the University, Basel (Switzerland) and (L.L.L) MRC Neurochemical Pharmacology Unit, Department of Pharmacology, University oJ" Cambridge, Cambridge (Great Britain) (Accepted August 30th, 1973)
SUMMARY
A retrograde axonal transport of nerve growth factor (NGF) from the adrenergic nerve terminals in the mouse iris to the cell bodies of postganglionic sympathetic neurones in the superior cervical ganglion has been demonstrated. After injection of iodinated nerve growth factor (125I-NGF) into the anterior eye-chamber there was a relatively rapid accumulation of radioactivity in the superior cervical ganglia on both injected and non-injected sides, as was the case after subcutaneous injection. However, 4 h after intraocular injection a preferential accumulation of radioactivity became apparent in the superior cervical ganglion on the injected side, and this difference between the ganglia on injected and non-injected sides gradually increased to a maximum at 16 h. Transection of the postganglionic adrenergic fibres as well as the prior intraocular injection of colchicine abolished the preferential accumulation of 125I-NGF in the superior cervical ganglion of the injected side, whereas the destruction of adrenergic nerve terminals by 6-hydroxydopamine did not impair the preferential accumulation. It is concluded that the retrograde axonal transport of NGF, which was estimated to take place at a rate of about 2.5 mm/h, depends on a colchicine-sensitive mechanism as does the orthograde rapid axonal transport. However, the uptake of NGF may not only take place from the nerve terminals but also from the preterminal parts, as has been shown in other studies with horseradish peroxidase. Autoradiographic studies strongly supported the existence of a retrograde transport by showing a clear localization of radioactivity in a small number of neurones in the superior cervical ganglion on the injected side, whereas on the non-injected side there was only a diffuse distribution of radioactivity throughout the ganglion.
104
i.A. HENDRY et al.
INTRODUCTION
The mouse submaxillary gland protein nerve growth factor (NGF) has been known for many years to have profound growth-promoting properties on the sympathetic nervous system of neonatal animals 16. It has more recently been shown that the administration o f large doses of N G F to developing mice 10 and rats 22 induces a selective increase in the activity of tyrosine hydroxylase (TH) and dopamine fl-hydroxylase (DBH), key-enzymes in the synthesis of noradrenaline (NA) in adrenergic neurones. These effects and the converse actions of N G F antiseruml6, ~v, suggest that N G F may be essential for the normal development o f adrenergic neurones in the peripheral sympathetic nervous system. Recently it has also been shown that even in adult animals a decrease in the tissue and plasma levels of N G F which occurs after sialectomy or castration, is also followed by a decrease in the activity of T H H and other enzymes involved in NA biosynthesis 23 in mice sympathetic ganglia. This effect could be reversed by administration of N G F z3. Thus, it would appear that N G F is essential for the maintenance of a normal function of the peripheral adrenergic neurones throughout the entire life of the animals. In neonatal mice the unilateral removal of a submaxillary gland produced a reduction in T H activity in the ipsilateral superior cervical ganglion. This effect could be reversed by the local administration of a cellulose-bound slow-release form of N G F at the site of the removal of the gland. It was suggested that a possible mechanism for this action might be a retrograde transport of N G F to the neuronal cell bodies via the adrenergic axons H. The phenomenon o f retrograde axonal transport of proteins has been described for several different neuronal populations: for example horseradish peroxidase from the optic rectum to the ganglion cells in the eyOS; horseradish peroxidase and albumin from the tongue to the hypoglossal nucleus 14. It is possible that such retrograde transport mechanisms may have a special significance in providing the neuronal cell body with a means of sampling the external environment around its axon terminals, thereby enabling the neurone to respond to trophic factors in this environment. N G F may represent such a trophic factor, and it was the purpose of the present study to provide direct evidence for the retrograde axonal transport of this biologically important compound. MATERIALS AND METHODS
N G F was prepared from mouse submaxillary glands by the method of Bocchini and Angeletti 3 and was labelled with 125iodine by a modification of the method of Greenwood et al. s. In order to iodinate all the N G F molecules we used a mixture of sodium 125iodide (supplied by the Radiochemical Centre Amersham, Bucks., England) and sodium a27iodide (supplied by Merck, Darmstadt, Germany). The reaction mixture contained 650 #g (24.5 nmoles) NGF, 2 mCi sodium 125iodide, 11.3 #g (75 nmoles) sodium 127iodide and 28.2 #g (100 nmoles) chloramine-T in a total volume of 400 /~1 of Tris-HC1, pH 8. The reaction took place at room temperature and was stopped after 30 min with 19 /tg (I00 nmoles) sodium metabisulphite (Na2S205) in
RETROGRADE TRANSPORT OF
NGF
105
100 #1 Tris-HC1 buffer, pH 8. The labelled NGF was separated from unreacted iodine by use of a Bio-Gel P2 (supplied by BIO-RAD Laboratories, Richmond, Calif., U.S.A.) column (0.9 cm × 15 cm) equilibrated with 0.1 M acetic acid followed by dialysis against several changes of saline. This fully iodinated NGF had a specific activity of 0.73 #Ci/#g. For the autoradiographic study iodinated NGF with a higher specific activity of 10 #Ci/#g was prepared by replacing the sodium 127iodide by sodium 125iodide. In this experiment the reaction mixture contained 66 /~g NGF (2.5 nmoles), 10 #Ci sodium 125iodide (5.7 nmoles), 8.46 big chloramine-T (30 nmoles) in a total volume of 80 #1 Tris-HCl, pH 8. The reaction was stopped after 30 min with 9.5 big (50 nmoles) Na2S205 in 50 #1 Tris-HCl buffer, pH 8, followed by dialysis against several changes of saline. Subsequently the lz5I-NGF was directly used for injection. Both labelled preparations of N G F retained their full biological activity in the chick dorsal root ganglion bioassay of Fenton 6. For all experiments male albino mice of the Swiss strain were used. Their body weight ranged between 30-35 g. The animals were kept at a constant temperature of 23 ~- 1 °C and they were supplied with NAFAG-pellets (supplied by NAFAG AG, Gossau SG, Switzerland) and water ad libitum. For the injection of labelled NGF into the anterior chamber of the eye, the animals were anaesthetised with ether, and in each case 0.25 #g of the compound was injected in 2 #1 of saline buffered with sodium phosphate at pH 7.4 using a Hamilton glass syringe. Subcutaneous injections were made at a site low on the back, in order to be as far removed from the superior cervical ganglion as possible. Surgical transection of the postganglionic nerve fibres of the superior cervical ganglion was carried out using a dissection microscope under ether anaesthesia. For the determination of 125I-NGF in superior cervical ganglia after intraocular or subcutaneous injection the animals were killed with ether and the ganglia dissected using a stereoscopic microscope. The ganglia were then rinsed in saline and the radioactivity of the intact tissue determined in a Packard gammacounter at a counting efficiency of approximately 50 ~. For autoradiographic studies of the localization of 125I-NGF in superior cervical ganglia the ganglia were fixed in 2.5 ~ glutaraldehyde in 0.1 M sodium phosphate buffer, pH 7.4 (v/v), post-fixed in osmium tetroxide (1 9/o w/v) in phosphate buffer, dehydrated in ethanol and embedded in an epoxy resin. Thin sections of the embedded tissue (1-2/,m) were mounted on glass slides, dipped in Ilford L 4 emulsion and stored in the dark at 4 °C for 28 days prior to development with Kodak D 19b developer. The tissue sections were examined by phase contrast microscopy, and autoradiographic activity was localized by dark field microscopy. Further experiments designed to determine the proportion of radioactivity present in the superior cervical ganglia in the form of antigenically intact NGF were carried out as follows: I~5I-NGF (0.25 /zg) was injected into the right eye. Twelve hours after the injection the animals were killed and the superior cervical ganglia of the injected and non-injected sides removed. In view of the small amounts of radioactivity present in the tissues the ganglia of 5 animals were pooled, homogenized in 2 ml of 0.05 M veronal buffer, pH 8.6 containing 0.9 ~o NaCI, 0.1 9/0sodium azide and
[. A. HENDRY et al.
106
At.,
o =
360 c 300 t~ \tm 2 4 0 '-: E
i80.
7:;1:1::c:2%
4i
d_ d 120 • 60
. . . . . . ,/-"~.=-Iz4~ 4 8 12 16 20 24 36 48
~5
2
injection (hours)
Time offer
cO
420-
B
m 360 L i/[''
300E d.
180 -
.~
120-
~
right side
/I' "% -
240 -
I
~-~left side
12
24
600
Time offer injection
48 (hours)
Fig. 1. A: time course of total accumulation of '2Sl in each superior cervical ganglion after unilateral intraocular injection (for details see Methods) of 0 . 2 5 fig a 2 5 I - N G F . E a c h point represents the mean i S . E . M . (vertical bars) for groups of 5-7 animals. B: time course of accumulation of lzsI in superior cervical ganglia after subcutaneous injection of 0.25 p g o f 1 2 S I - N G F . Each point represents the mean ± S . E . M . (vertical bars) from groups of 6 animals.
0.5 % bovine albumin. This homogenate was incubated with solid phase N G F antibodies 9 and slowly mixed on a rotary mixer for 24 h at 4 °C. After this the immunoadsorbent paper was washed free of unbound radioactivity and the bound radioactivity was measured. The percentage of the total radioactivity binding to the immunoadsorbent was taken to represent antigenically intact NGF. In those experiments in which the effect of the destruction of adrenergic nerves on the retrograde transport was studied the animals were injected intravenously on 3 consecutive days with 0.32 mmoles/kg 6-hydroxydopamine hydrochloride (supplied by Fluka, Buchs SG, Switzerland). For intraocular injection of colchicine (supplied by Koch-Light Laboratories Ltd., Colnbrook, Bucks., England) a solution of 20 mg/ml was used. Two #1 of this solution were injected into the anterior eye-chamber at various times prior to the injection of t25I-NGF. RESULTS
Accumulation of radioactivity in superior cervical ganglia after intraocular and subcutaneous injection of t~SI-NGF After the intraocular injection of 0.25 fig of t25I-NGF there was a relatively rapid accumulation of radioactivity in superior cervical ganglia on both injected and non-injected sides, which reached a maximum after 6 h (Fig. 1A). However, as early
RETROGRADE u E o
TRANSPORT
oF NGF
107
E ~.
.~_ ~
I~: ~> ~
60 ~4 Time
3~ 48
injection
(hours)
B
g2 "T
g E
o
8;2,~2o2. offer
"6 .---
109+ 8" 7654 3 2 1
j/ ....... 6 1012
Time a f t e r
,..
6
it--,.-/]
.
20 2A 36 /.8
injection
[hours)
Fig. 2. A: difference in 12'~Iaccumulation between superior cervical ganglia from injected (intraocular injection of 0.25/~g of 125I-NGF) and non-injectedside. B: ratio of 1251accumulation between the ganglia from the injected and the non-injected side. as 4 h after the injection there was a statistically significant (P < 0.05) difference between the amount of radioactivity accumulated in the ganglia on the injected and non-injected sides (Figs. 1A, 2A). This difference increased gradually and reached a maximum after 16 h. The differential accumulation becomes even more impressive if it is expressed in terms of the ratio between injected and non-injected side which reaches a value of about 3 after 12 h, 3.5 after 24 h and 9 after 48 h (Fig. 2B). It is likely that the initial rapid increase in radioactivity in both ganglia is due to 125I-NGF which escaped from the anterior eye-chamber into the general circulation and reached the ganglia via the blood stream. This assumption is strongly supported by the observation that the time-course of the initial accumulation of radioactivity in both superior cervical ganglia is very similar after intraocular and subcutaneous (Fig. 1) injections of 125I-NGF.
Effect of transeeting the postganglionie adrenergiefibres and intraocular administration of colehicine Although the initial rapid increase in radioactivity in both superior cervical ganglia after intraocular injection is due to blood-borne lzSI-NGF, the slowly increasing difference between left and right sides is probably due to a retrograde transport of labelled material from the adrenergic nerve terminals in the eye to the cell bodies in the superior cervical ganglia. Since the only known neural connections between the eye and the superior cervical ganglion are postganglionic sympathetic fibres, their transection should abolish the differential accumulation of radioactivity
108
I.A. HENDRY et al.
between the injected and non-injected sides. This indeed was the case. Surgical transection o f the postganglionic nerve fibres emanating f r o m the superior cervical ganglion completely abolished the difference in the accumulation o f 125I-NGF between left and right ganglia normally observed 12 h after intraocular injection (Table 1). The fact that there was no significant (P > 0.05) difference in the accumulation o f radioactivity between the operated and unoperated ganglia 2 h after the subcutaneous injection o f 125I-NGF (Fig. 3) shows that the operation had little direct effect on the ability o f the ganglia to accumulate blood-borne N G F . Hence it follows that intact postganglionic adrenergic nerve fibres seem to be a prerequisite for the preferential accumulation o f 125I-NGF in the superior cervical ganglion on the injected side. The rate o f the retrograde axonal transport o f N G F f r o m the adrenergic nerve terminals to the cell b o d y seems to be rather rapid as far as can be estimated by an approximate calculation. This calculation is based on the assumption that the distance between the eye and the ganglion is about 1 cm and that the first significant difference in radioactivity between the ganglia on the injected and non-injected sides after 4 h represents the first moiety o f 125I-NGF reaching the ganglion o f the injected side by retrograde axonal transport. This calculated rate of 2.5 m m / h is similar to the 3 m m / h reported for the transport o f horseradish peroxidase f r o m the chick optic tectum to the retinal ganglion cells by La Vail and La VaiP 5. This rate o f transport is also in the range o f the fast orthograde transportT, 19 which seems to depend on intact neurotubules and can be blocked by colchicine2, 5. It was, therefore, o f interest to determine the effect o f colchicine on the retrograde axonal transport o f a25I-NGF. Intraocular injection o f
TABLE I EFFECTS OF COLCH1CINE AND TRANSECTION OF THE POSTGANGLIONIC NERVE FIBRES ON THE DIFFERENTIAL ACCUMULATION OF 125I
Groups of 5-6 animals were injected into the anterior chamber of right eye with 0.25/~g of 12~I-NGF in 2/zl of saline at the appropriate time after treatment with colchicine or surgery. Both superior cervical ganglia were removed 12 h after the intraocular injection of le5I-NGF and the total radioactivity of the intact tissues was measured in a Packard gamma-counter. The difference in the amount of 1~5Iaccumulated in the two ganglia was then expressed as the mean ± S.E.M. of the individual differences in counts/min. Treatment
Albumin in saline Transection of postganglionic nerve fibres Colchicine 2/~1 (20 mg/ml)
Time be]bre 1251-NGF injection (h)
Dtfference in ganglionic 1251 (counts~rain)
2 0
217 ± 22 10 ± 34'**
0 2 12
229 ± 42* 70 ± 31 ** 25 -~: 37***
* Differs from 0, P < 0.005, but not from control animals, P :> 0.05. ** Differs from 0 and control animals, P < 0.05. *** Differs from control animals, P < 0.01, but not from 0, P ~ 0.05.
RETROGRADE TRANSPORT OF N G F
109
~m
intact
I
I
transected t"
.9
O. O
oo
I
Ot1' ÷÷
O÷
control 6-0HDA after 2h
control 6-0HDA after 16h
]Fig, 3. Effects of 6-hydroxydopamine and surgical transection of the postganglionic nerve trunk on
the accumulation of 1251in the superior cervical ganglia after subcutaneous injection of 1 ktg of I~sINGF. The animals were killed 2 or 16 h after injection with 125I-NGF. Transection of the nerve fibre was carried out 24 h before the animals were sacrificed. 6-Hydroxydopamine treatment consisted of injecting the animals intravenously on 3 consecutive days with 0.32 mmoles/kg. The last dose was given 24 h before the injection of 125I-NGF. Results are expressed as means ~ S.E.M. (vertical bars) for groups of 5-7 animals. * Left and right ganglia do not differ from each other; P < 0.1. ** Left and right ganglia differ from each other; P < 0.005. + Differs from corresponding ganglia in control animals; P < 0.5. ++ Differs from corresponding ganglia in control animals; P < 0.01.
4 0 / t g o f colchicine 12 h before that o f 0.25/zg 125I-NGF completely abolished the differential accumulation in the superior cervical ganglion (Table I). However, if colchicine was injected at the same time as 125I-NGF then the differential accumulation was maintained, suggesting that colchicine requires some time to exert its action on the axonal transport mechanism, possibly by diffusion to the preterminal parts o f the adrenergic axon since destruction o f the nerve terminals does not seem to impair materially the retrograde transport, as will be shown below in the experiments with 6-hydroxydopamine.
Destruction of adrenergic nerve terminals by 6-hydroxydopamine: effect on retrograde axonal transport of 1251-NGF In order to determine whether intact adrenergic nerve terminals are essential for the retrograde axonal transport o f 125I-NGF a further series o f animals was treated with 6 - h y d r o x y d o p a m i n e which is k n o w n to destroy adrenergic nerve terminals selectively in adult animals 24. The animals were injected intravenously on 3 consecutive days with 0.32 mmoles/kg o f 6-hydroxydopamine. Twenty-four hours after the last dose they received an intraocular injection o f 0.25/~g o f 125I-NGF. In spite o f the destruction o f the adrenergic nerve terminals by 6-hydroxydopamine there was a significant difference in the accumulation o f radioactivity between the ganglia o f the injected and non-injected side, indicating that the destruction o f the nerve terminals
110
i.A. HENDRYet al.
TABLE 1I ACCUMULATION OF 12,5] IN SUPERIOR CERVICAL GANGLIA AFTER INJECTION OF
1251-NGFINTO THE ANTE-
RIOR CHAMBER OF THE RIGHT EYE
Animals were injected into the anterior chamber of the right eye with a dose of 0.25/Ag of ~251-NGF. The left and right superior cervical ganglia were removed after 4 or 16 h and the radioactivity was measured in a gamma-counter. The values given represent the mean ± S.E.M. of counts/rain/ganglion for groups of 5 animals. 6-Hydroxydopaminetreatment consisted of injecting the animals intravenously on 3 consecutivedays with 0.32 mmoles/kg. The last dose was given 24 h before the injection of 1251-NGF. Time after injection ( h) 4 16
Treatment
Right ganglion ( counts/min)
Left ganglion ((counts~rain)
Control 6-OHDA Control 6-OHDA
276 i 194 ± 120 , 87 ,
238 ± 189 , 50 -42 ,
25.2 18.1" 10.0 4.2**
9.6 12.8" 8.9*** 11.6"**
* Differs from untreated control animals, P < 0.05. ** Differs from untreated control animals, P < 0.005. *** Differs from the right ganglion, P < 0.001.
did not impair the retrograde transport of 125I-NGF (Table II). However, the accumulation of radioactivityin both ganglia was less than that observed in untreated animals, suggesting that the 6-hydroxydopamine treatment may have had some direct effect on the ganglionic cell bodies. The assumption of a direct damaging effect to the cell bodies by 6-hydroxydopamine is supported by the observation that as early as 4 h after the intraocular injection of 125I-NGF (when a small contribution from the retrograde axonal transport to the radioactivity accumulated in the ganglion was to be expected) the radioactivity measured in superior cervical ganglia of 6-hydroxydopamine treated animals was considerably lower (P < 0.05) than that in untreated controls (Table II). In order to provide further evidence of a direct effect of 6-hydroxydopamine on the accumulation of blood-borne 125I-NGF by the adrenergic cell body, the effect of this drug was examined in animals in which the postganglionic fibres had been cut. A subcutaneous injection of l #g of 12~I-NGF was given and the accumulation of radioactivity in the superior cervical ganglion measured. As in ganglia of untreated animals there was no difference between the radioactivity accumulated 2 h after injection by ganglia with intact and transected postganglionic fibres. However, there was a significantly (P < 0.05) smaller accumulation of radioactivity in transected and intact ganglia of animals treated with 6-hydroxydopamine as compared to untreated animals (Fig. 3). Sixteen hours after the injection of 12SI-NGF there was much less radioactivity accumulated in the ganglia with transected postganglionic fibres both in 6-hydroxydopamine treated animals and untreated controls. However, the accumulation of radioactivity on both transected and intact sides was significantly (P < 0.01) lower in 6-hydroxydopamine treated animals than in untreated controls, whereas the ratio be-
RETROGRADETRANSPORTOF NGF
111
tween transected and non-transected side was virtually the same in controls and 6hydroxydopamine treated animals. Thus, it can be concluded that intact nerve terminals are not a prerequisite for the retrograde axonal transport of 125I-NGF and that the reduced accumulation of radioactivity in the superior cervical ganglion after treatment with 6-hydroxydopamine is due to a direct damaging effect of this amine on the cell body, impairing the accumulation of blood-borne lz~I-NGF. Furthermore, these experiments have also shown that a considerable part of 125I-NGF injected subcutaneously accumulated in the superior cervical ganglion by way of retrograde axonal transport.
Autoradiographic studies of superior cervical ganglia after intraocular injection of 125I-NGF Autoradiographic examination of ganglia taken from the injected and noninjected sides of animals that had received intraocular lzSI-NGF 12 h previously, strongly supported the view that retrograde axonal transport of the labelled material had occurred on the injected side. In superior cervical ganglia from the injected side there was a clear localization of radioactivity in a small number of ganglionic neurones, together with a diffuse background radioactivity throughout the ganglion. In ganglia from the non-injected side only the diffuse labelling was seen, with no clear localization over specific neurones (Figs. 5, 6). The labelled neurones on the injected side were often seen to contain most radioautographic activity over the cytoplasm, with relatively little labelling in the nucleus (Fig. 6). Similar results were observed in ganglia from both sides of 4 animals. No accurate estimate was made of the proportion of neurones that became intensely labelled after injection of 125I-NGF into the eye, but the proportion was almost certainly less than 10 ~ of all ganglionic neurones (Fig. 4).
Determination of antigenically intact 125I-NGF in the superior cervical ganglia after intraocular injection The half-time of circulating exogenously administered 125I-NGF is very short (20-30 min) and a considerable part of the radioactivity present in the plasma after intravenous injection of 125I-NGF represents degradation products and not intact NGF 23. Therefore, it seemed to be important to determine the proportion of the radioactivity present in the ganglia as intact NGF. Because of the extremely low amount of NGF accumulated in the ganglia it was not possible to determine its biological activityL Thus, an immunoadsorption method was used, involving the use of solid phase NGF antibodies previously described for a radioimmunoassay for NGF 9. This procedure estimated the proportion of azSI-NGF which remained bound to at least one antigenic site of the NGF molecule, but it did not allow us to determine the exact amount of biologically active NGF present in the tissues. However, any radioactivity that did not bind to the immunoadsorbent was almost certainly devoid of biological activity. Using this method we found that in homogenates of ganglia from the injected side 60 ± 3 % of the total radioactivity bound to the solid phase antibodies when the ganglia were examined 12 h after the injection. This contrasts with a binding of only 35 ± 2 % of the radioactivity present in the ganglia from the non-injected side of the
112
I. A. HENDRY et a l
Fig. 4. Photomontage of phase contrast micrographs (A) and dark field autoradiograms (B) of mouse superior cervical ganglion from the ioected side of an animal which received an intraocular injection of 125I-NGF 12 h previously. A small number of intensely labelled cells are observed, together with diffuse background labelling of the ganglion. × 10t).
RETROGRADE TRANSPORT oF N G F
113
114
1. A. HENDRY et al.
Fig. 5. Photomontage of phase contrast micrographs (A) and dark field autoradiograms (B) as in Fig. 4, but of a superior cervical ganglion from the side opposite to that used for intraocular injection. × 100.
RETROGRADE TRANSPORT OF N G F
115
116
i . A . HENDRY et al.
Fig. 6. Phase contrast micrographs (A) and dark field autoradiograms (B) of corresponding microscopic fields from mouse superior cervical ganglia on the injected (1) and non-injected (2) sides from an animal that received an intraocular injection of 12H-NGF 12 h previously. Note that the labelled cell in the lower left hand corner of B1 c~rresponds to a neurone (A1), and that labelling is less intense over the nucleus. × 400.
RETROGRADE TRANSPORT OF N G F
117
118
i.A. HENDRY et al.
same animals. This difference in antigenically intact N G F between the two ganglia represents that proportion which is protected during retrograde transport. The circulating N G F is more rapidly metabolized. Indeed, at this time only 8 & 1 To of the radioactivity in the serum is bound to the solid phase antibodies, indicating that the majority of the radioactivity in the serum was not N G F but that there was a specific accumulation of radioactivity in the ganglia representing N G F which contained at least one intact antigenic site. DISCUSSION
After previous experiments had suggested that N G F may exert its physiological control of adrenergic neurones as a 'retrograde trophic factor' acting from the periphery on the perikaryon 11, the present experiments have provided evidence for the existence of such a retrograde axonal transport. This evidence is based on the observation that after intraocular injection of a25I-NGF there was a preferential accumulation of radioactivity in the superior cervical ganglion of the injected side. This preferential accumulation could be abolished by surgical transection of the postganglionic adrenergic fibres which are the only known neural connections between the eye and the superior cervical ganglion. The most direct evidence for retrograde axonal transport is provided by the autoradiographic studies which showed a very intense labelling of a few neuronal cell bodies in the superior cervical ganglion on the injected side. This selective accumulation of radioactivity over a few neuronal ceils is most plausibly explained by a retrograde transport from the few axons leading to the iris; any other form of perineuronal transport would be likely to result in a more diffuse distribution of labelled N G F in the ganglia. The absence of any such cells on the non-injected side eliminates the possibility that this selective labelling was due to the presence of a small proportion of ceils in the ganglion with a specific ability to take up NGF. In addition to this retrograde axonal transport of N G F there seems to exist a mechanism for the accumulation of blood-borne N G F in the sympathetic ganglia which is responsible for the rapid initial accumulation occurring after subcutaneous or intraocular injection. The mechanism of this accumulation of circulating N G F in sympathetic ganglia is not clear and it remains to be established whether the binding sites are localized at the surface of the neuronal membrane or whether circulating N G F is transported into the neuronal ceils. It is noteworthy that even after subcutaneous injection of 125I-NGF a considerable part of the radioactivity accumulating in the sympathetic ganglia seems to reach the adrenergic cell bodies by retrograde axonal transport. This can be deduced from the observation that 16 h after subcutaneous injection of 1251-NGF the radioactivity present in the superior cervical ganglia with transected postganglionic fibres amounted to about 5 0 ~ o f that on the intact side. The fact that the accumulation of radioactivity 2 h after subcutaneous injection did not differ between transected and intact ganglia eliminates the possibility that the reduced accumulation after 16 h results from an impairment of the mechanism responsible for the accumulation of circulating N G F as a consequence of axotomy. In contrast to the intactness of the axons, the integrity of the adrenergic nerve
RETROGRADE TRANSPORT OF N G F
119
terminals does not seem to be a prerequisite for the retrograde axonal transport of NGF. This can be concluded from the observation that the selective destruction of adrenergic nerve terminals by 6-hydroxydopamine does not impair the retrograde transport of I~5I-NGF, i.e. 16 h after intraocular injection of 125I-NGF the ratio between the radioactivity in the ganglia of the injected and non-injected sides was the same in 6-hydroxydopamine treated animals and controls. However, pretreatment with 6-hydroxydopamine reduced the amount of radioactivity present in superior cervical ganglia both 2 and 16 h after subcutaneous injection of I~I-NGF. This implies that the reduced accumulation of N G F most probably results from a direct damaging effect of 6-hydroxydopamine on the cell body impairing both the accumulation of NGF reaching the cell body by retrograde axonal transport and by the blood stream. Thus, the impairment of the accumulation of radioactivity in sympathetic ganglia after systemic administration of 125I-NGF in animals pretreated with 6hydroxydopamine cannot be taken as evidence for retrograde transport of N G F 1. It merely represents a manifestation of a damaging effect of 6-hydroxydopamine on the cell body. Such an effect was not unexpected since it had been shown in previous experiments that in adult rats 6-hydroxydopamine seems to have a direct damaging effect on the cell bodies of adrenergic neurones, since 6-hydroxydopamine treatment blocked - - at least transiently - - the trans-synaptic induction of tyrosine hydroxylase that normally occurred in the superior cervical ganglia after cold exposure or administration of reserpine4, 21. Since the destruction of adrenergic nerve terminals with 6-hydroxydopamine did not interfere with the retrograde transport of 12~I-NGF one has to assume that the uptake of NG F into the adrenergic axon is not limited to the nerve terminals. This is in agreement with the observation that horseradish peroxidase can penetrate into the axons of peripheral nerves through the Schwann cells 13. Furthermore, there is evidence that in the mouse iris regeneration of adrenergic nerve terminals starts very rapidly after destruction by 6-hydroxydopamine18. Thus, it is possible that uptake of N G F takes place at these sites of regeneration, and that the ability for rapid regeneration may be related to this uptake of NGF. Although the present experiments did not allow an accurate determination of the rate of retrograde axonal transport it seems that the values calculated for the mouse sympathetic nervous system (2.5 mm/h) are very close to those determined by La Vail and La VaiP 5 in the chick embryo (3 mm/h) and by Kristensson e t al. 14 in adult rabbits (2.5 mm/h). The latter workers have also shown that the retrograde transport was blocked by ischaemia. The values calculated for the rate of retrograde axonal transport are within the range of the rapid orthograde transport which depends on an intact energy metabolism 2° and on functionally intact neurotubules, i.e. it can be blocked by colchicine and vinblastine2, 5. In view of the many similarities between the orthograde and retrograde axonal transport it is likely that both these fast transport systems involve similar mechanisms. The question arises as to whether there are two independent transport systems for retrograde and orthograde transport, or whether the same transport system performs
120
i.A. HENDRYet al.
the transport in both directions, the direction of transport depending on the properties of the material to be transported. At least for the retrograde transport one could assume that all the substances transported are enclosed in vesicles - - as has been shown for horseradish peroxidasO 2,z5 - - and that the surface of these vesicles might determine the direction of the transport. The assumption of a retrograde transport within vesicles would possibly also explain how the 125I-NGF reaching the cell body by retrograde axonal transport is better protected from metabolic degradation than that reaching the cell body via the blood stream. The specificity for the macromolecules to be taken up by the nerve terminals and to be transported to the perikaryon does not seem to be very high. However, if this retrograde axonal transport should have any physiological significance for the function of the sympathetic neurones it should exhibit some degree of specificity. We believe that this specificity may be conferred by the nature of the molecules transported and not by the transport mechanism itself. Thus, N G F is able to exert its specific effects on the adrenergic neurones once it has been brought to the cell body by retrograde axonal transport. It has been shown that after axotomy there is a change in the enzyme pattern in the adrenergic neurone, with a decrease in the relative amounts of the enzymes involved in NA biosynthesis 4. Moreover, castration which leads to a fall in tissue and plasma NGF, also caused a similar decrease in the enzymes involved in NA biosynthesis 11. If N G F is one of the factors important for the normal function of an adrenergic neurone in contact with its end organ, then the retrograde axonal transport of this substance may provide a mechanism for its control. ACKNOWLEDGEMENTS
This work was supported by the Swiss National Foundation for Scientific Research (Grant No. 3.653.71) and by the U.K. Medical Research Council. I. A. Hendry was the recipient of a fellowship from the Postgraduate Medical Foundation of Sydney University.
REFERENCES 1 ANGELETTI, R. H., ANGELETTI, P. U., AND LEvI-MONTALCINI, R., Selective accumulation of 1251labelled nerve growth factor in sympathetic ganglia, Brain Research, 46 (1972) 421425. 2 BANKS, P., AND MAYOR, D.0 lntra-axonal transport in noradrenergic neurons in the sympathetic nervous system, Biochem. Soc. Symp., 36 (1972) 133-149. 3 BOCCHINI, V., AND ANGELETTI, P. U., The nerve growth factor: purification as a 30,000-molecularweight protein, Proc. nat. Acad. Sci. (Wash.), 64 (1969) 787-794. 4 BRIMIJOIN, S., AND MOLINOFF, P. B., Effects of 6-hydroxydopamine on the activity of tyrosine hydroxylase in sympathetic ganglia of the rat, J. Pharmacol. exp. Ther., 178 (1971) 417424. 5 DAHLSTR6M,A., Axoplasmic transport (with particular respect to adrenergic neurons), Phil. Trans. B, 261 (1971) 325-358. 6 FENTON,E. L., Tissue culture assay of nerve growth factor and of the specificantiserum, Exp. Cell Res., 59 (1970) 383-392. 7 GEFFEN,L. B., AND LIVETT,B. G., Synaptic vesicles in sympathetic neurons, Physiol. Rev., 5l (1971) 98-157.
RETROGRADE TRANSPORT OF N G F
121
8 GREENWOOD,F. C., HUNTER, W. M., AND GLOVER,J. S., The preparation of lalI-labelled human growth hormone of high specific radioactivity, Biochem. J., 89 (1963) 114-123. 9 HENDRY, 1. A., Developmental changes in tissue and plasma concentrations of the biologically active species of nerve growth factor in the mouse, by using a two-site radioimmunoassay, Biochem. J., 128 (1972) 1265-1272. 10 HENDRY, I. A., AND IVERSEN, L. L., Effect of nerve growth factor and its antiserum on tyrosine hydroxylase activity in mouse superior cervical sympathetic ganglion, Brain Research, 29 (1971) 159-162. 11 HENDRY,|. A., AND IVERSEN,L. L., Changes in tissue and plasma concentrations of nerve growth factor following removal of the submaxillary glands in adult mice and their effects on the sympathetic nervous system, Nature (Lond.), 243 (1973) 500-504. 12 HEUSER, E., FREEMAN,A. R., AND KASHNER, A., Stimulation-dependent alteration in peroxidase uptake at lobster neuromuscular junctions, Science, 173 (1971) 733-734. 13 KRISHNAN,N., AND SINGER, M., Penetration of peroxidase into peripheral nerve fibres, Amer. J. Anat., 136 (1973) 1-14. 14 KRISTENSSON,K., OLSSON, Y., AND SJOSTRAND,J., Axonal uptake and retrograde transport of exogenous proteins in the hypoglossal nerve, Brain Research, 32 (1971) 399-406. 15 LA VAIL, J. H., AND LA VAIL, M. M., Retrograde axonal transport in the central nervous system, Science, 176 (1972) 1416-1417. 16 LEvI-MONTALCINI, R., AND ANGELETTI, P. U., Nerve growth factor, Physiol. Rev., 48 (1968) 534-569. 17 LEVI-MONTALCINI,R., AND BOOKER, B., Destruction of sympathetic ganglia in mammals by antiserum to a nerve growth protein, Proc. nat. Acad. Sci. (Wash.), 46 (1960) 384-391. 18 MALMFORS,T., The effect of 6-hydroxydopamine on the adrenergic nerves as revealed by the fluorescence biochemical method. In 6-Hydroxydopamine and Catecholamine Neurons, NorthHolland, Amsterdam, 1971, pp. 47-58. 19 OCHS, S., Characteristics and a model for fast axoplasmic transport in nerve, J. NeurobioL, 2 (1971) 331-345. 20 OCHS, S., AND RANISH, N., Metabolic dependence of fast axoplasmic transport in nerve, Science, 167 (1970) 878-879. 21 THOENEN, H., Biochemical alterations induced by 6-hydroxydopamine in peripheral adrenergic neurons. In 6-Hydroxydopamine and Catecholamine Neurons, North Holland Publ. Co., Amsterdam, 1971, pp. 75-85. 22 THOENEN,H., ANGELETTI,P. U., LEvI-MONTALCINI,R., AND KETTLER, R., Selective induction by nerve growth factor of tyrosine hydroxylase and dopamine fl-hydroxylase in the rat superior cervical ganglia, Proc. nat. Acad. Sci. (Wash.), 68 (1971) 1598-1602. 23 THOENEN,H., HENDRY, I. A., STOCKEL,K., PARAVICINI,U., AND OESCH, F., Regulation of enzyme synthesis by neuronal activity and by nerve growth factor. In Dynamics of Degeneration and Growth in Neurons, Wenner-Gren Center International Symposium, in press. 24 THOENEN,H., AND TRANZER,J. P., Chemical sympathectomy by selective destruction of adrenergic nerve endings with 6-hydroxydopamine, Naunyn-Schmiedeberg's Arch. exp. Path. Pharmak., 261 (1968) 271-288. 25 ZACKS, S. I., AND SAITO, A., Uptake of exogenous horseradish peroxidase by coated vesicles in mouse neuromuscular junctions, J. Histochem. Cytochem., 17 (1969) 161-170.