Specificity of pseudorabies virus as a retrograde marker of sympathetic preganglionic neurons: implications for transneuronal labeling studies

Brain Research, 617 (1993) 103-112 © 1993 Elsevier Science Publishers B.V. All rights reserved 0006-8993/93/$06.00

103

BRES 19022

Specificity of pseudorabies virus as a retrograde marker of sympathetic preganglionic neurons: implications for transneuronal labeling studies A.S.P. Jansen,

D.G. Farwell

and A.D. Loewy

Department of Anatomy and Neurobiology, Washington UniL,ersity School of Medicine, St. Louis, MO 63110 (USA) (Accepted 16 February 1993)

Key words: Autonomic nervous system; Pseudorabies virus; Sympathetic nervous system; Sympathetic preganglionic neuron; Transneuronal cell body labeling

The purpose of the present study was to examine the specificity of the Bartha strain of pseudorabies virus (PRV) as a CNS retrograde marker. This information is critical in assessing whether this virus has potential value as a specific transneuronal marker. The model system chosen for analysis was the intermediolateral cell column (IML)--the principal site of origin of sympathetic preganglionic neurons (SPNs). Two experiments were performed. The first experiment established the usefulness of this model system and the second examined the properties of PRV as a retrograde cell body marker. In the first experiment, injections of two different conventional retrograde cell body markers (cholera toxin-/3 subunit (CTb) and Fluoro-Gold) were made in two ipsilateral sympathetic structures (viz., stellate ganglion and adrenal gland) in the same rat. This experiment established that (1) heterogenous SPNs originate in the same cell clusters that form the IML at the T4-T 8 levels and 2) SPNs innervate specific sympathetic targets with almost none providing a dual innervation of the stellate ganglion and adrenal gland. This mosaic arrangement of target-specific SPNs makes the IML an excellent CNS site for this type of study. The second experiment followed the same paradigm: PRV was injected into the stellate ganglion and CTb into the adrenal gland (and vice versa). These experiments established that PRV infections of one functional class of SPNs did not produce infections in nearby, functionally unrelated SPNs and did not cause a reduction in the SPN cell population, except under conditions of severe gliosis. These two properties increase the probability that Bartha PRV may be used as a specific retrograde transneuronal marker of central autonomic pathways.

INTRODUCTION

in a non-specific fashion by exocytosis into the extracellular space where they may infect n e a r b y n e u r o n s or

T h e viral t r a n s n e u r o n a l labeling m e t h o d is a n ext r e m e l y powerful n e u r o a n a t o m i c a l tool for the identification of hierarchical chains of c e n t r a l n e u r o n s i n n e r -

glial ceils or by a specific t r a n s f e r process to c o n t i g u o u s second o r d e r n e u r o n s that synapse o n the infected first o r d e r n e u r o n s . A f u r t h e r possibility is spread to the

vating a CNS cell g r o u p or a specific e n d - o r g a n . Application of this t e c h n i q u e is straightforward. A n injection of live h e r p e s viruses is m a d e in a p a r t i c u l a r CNS site or p e r i p h e r a l tissue such as a gland, muscle, or auton o m i c ganglion. Viruses are t a k e n u p a n d t r a n s p o r t e d into the n u c l e u s of the first-order n e u r o n s i n n e r v a t i n g the injected structure, w h e r e they u n d e r g o replication. O n c e viral D N A replication a n d assembly are com-

glial ceils that contact the infected n e u r o n . F r o m the s t a n d p o i n t of using viruses as CNS t r a n s n e u r o n a l markers, it is critical to use a n e u r o t r o p i c virus that is t r a n s m i t t e d by a trans-synaptic m e c h a n i s m . Such a

plete, newly f o r m e d virions exit the nucleus, acquiring a n e n v e l o p e derived from the i n n e r n u c l e a r m e m b r a n e . T h e viruses are t h e n t r a n s p o r t e d in the cytoplasm of the infected n e u r o n via the e n d o p l a s m i c r e t i c u l u m a n d Golgi a p p a r a t u s . S u b s e q u e n t l y , they are released either

virus u n d e r the most favorable c o n d i t i o n s can be used to label f u n c t i o n a l l y c o n n e c t e d sets of n e u r o n s that r e g u l a t e a specific target-site. By selecting the appropriate n e u r o t r o p i c virus a n d t e m p o r a l conditions, it is now possible to localize the first, second, a n d sometimes third o r d e r n e u r o n s involved in the control of a p a r t i c u l a r structure a n d these virally infected n e u r o n s are easily identified by s t a n d a r d i m m u n o h i s t o c h e m i c a l procedures.

Correspondence: A.D. Loewy, Department of Anatomy and Neurobiology, Washington University School of Medicine, 660 South Euclid Avenue, St. Louis, MO 63110, USA. Fax: (1) (314) 362-3446.

104 Most of the studies that have been published so far deal with central autonomic pathways 2'7-gA3A9'22'23. However, this technique has been used to study the central connections of other peripheral tissues 11'16-18'24'25 and central neural circuits as well 4-6'12'1a'27. In addition, the feasibility of using a /3-galactosidase expressing form of pseudorabies virus in electron microscopic studies of transneuronally labeled neurons has been demonstrated 1°. Thus, it is clear that this technology offers promise as being an extremely valuable technique for deciphering CNS connections. In evaluating data derived from viral transneuronal labeling studies, it is important to recognize that this labeling method depends on producing a highly limited active viral infection restricted to a small set of neurons. Various types of viruses infect neurons (e.g. herpes-, polio-, rabies-, and reoviruses), but not all of these are useful transneuronal tracers. To date, most investigators have used either herpes-simplex virus type-1 (HSV-1) or pseudorabies virus (PRV) for neuroanatomical studies. However, not all forms of these viruses are suitable. In particular, wildtype forms of HSV-1 and PRV usually produce uncontrolled central viremias 22'24, although some investigators did not find this was a problem in the neuronal system they studied 27. Nevertheless, this type of potential problem can be avoided by using attenuated forms of these viruses, but not all of these weakened strains are useful 6"22. One weakened strain of pseudorabies virus, a swine herpes virus called Bartha PRV which is used in the pig industry as a vaccine to prevent the 'mad itch' or Aujesky's disease, produces specific transneuronal labeling patterns in the CNS (e.g. see refs. 5, 16, 22, and 23). Bartha PRV has been shown to be a specific retrograde transneuronal marker in the sympathetic nervous system 2°. While this finding strongly supports the usefulness of Bartha PRV as a specific transneuronal marker, we felt it was critical to examine the specificity of this particular virus in the CNS under similar conditions as have been employed in our previous neuroanatomical investigations. The CNS model system we selected was the intermediolateral cell column (IML) which is the site of origin of sympathetic preganglionic neurons (SPNs). This cell group was particularly attractive for analysis because different functional types of SPNs often lie next to each other in the same spinal segment of the IML. Thus, our objective was to determine if a viral infection was produced in one type of SPN (e.g. sympathoadrenal SPNs) would the infection spread to adjacent non-related SPNs that innervate a different target (e.g. stellate SPNs). Quantitative data

are presented that indicate this did not occur. This finding fulfills one of the criteria set forth in our earlier study regarding the specificity of the transneuronal labeling method 2°. Thus, the observations presented here support the idea that Bartha PRV may be used a specific retrograde transneuronal marker in the CNS. MATERIALS AND METHODS All experiments were performed on male Sprague-Dawley rats (n = 178, weight = 180-220 g, Sasco Inc., O'Fallon, MO) under general anesthesia with sodium pentobarbital (50 m g / k g ) . Two separate surgical procedures were performed in each rat and the details are described below. During the postoperative recovery periods, none of the animals exhibited signs of systemic or local infections, appeared ill, or displayed abnormal behavior such as piloerection, itching, hyperarousability, or aggressiveness.

Experiment 1. Double retrograde cell body labeling The first experiment was designed to determine whether stellate and adrenal SPNs lie within the same cell clusters of the IML. The right stellate ganglion was injected with 300 nl of a 1% solution of cholera toxin /3-subunit (CTb; List Biological Laboratories, Campbell, CA) dissolved in distilled water. Two days later, the rats were anesthesized and the right adrenal gland was injected with 5 pA of a 4% suspension of Fluoro-Gold (Fluorochrome, Englewood, CO) made in sterile saline. Four days following the second injection, the rats were anesthetized and perfused transcardially with 300 ml of saline followed by 500 ml of 4% paraformaldehyde in 0.t M sodium phosphate buffer (pH = 7.4). The spinal cords were removed, postfixed overnight in the 4% paraformaldehyde solution, transferred to a 30% sucrose in 0.1 M phosphate buffer containing 0.1% sodium azide and stored at 4°C for 2 - 3 days. The T 4 - T s levels of the spinal cord were prepared for histological study. Each spinal segment was marked by a dorsoventral needle puncture that extended through the midline. The rostral entry point of the dorsal roots served as the landmark to define each spinal segment. The spinal cord was sectioned longitudinally on a freezing microtome at 5 0 / z m . The sections incubated overnight at room temperature in a 1:20,000 solution of goat anti-CTb serum (List Biological Laboratories), 10% normal rabbit serum, and 0.3% Triton X in 0.02 M potassium phosphate-buffered saline (KPBS). The following day, the sections were washed twice in KPBS, placed in a solution containing a 1 : 100 dilution of biotinylated rabbit anti-goat (Vector Laboratories, Burlingame, CA) made in KPBS containing 10% normal rabbit serum. The sections were incubated in this solution for 2 h, then, washed twice in KPBS and transferred to a 1 : 100 solution of streptavidin-conjugated rhodamine isothiocyanate (RITC, Jackson I m m u n o R e s e a r c h Laboratories, West Grove, PA) in 10% normal rabbit serum in KPBS. After 2 h, the sections were washed in KPBS (2x), m o u n t e d on gelatin-coated slides, coverslipped in a glycerol-PBS mountant (Citifluor Ltd, London, UK). Labeled fluorescent neurons were mapped using an X - Y plotter attached to an MD2 digitizer (Minnesota Datametrics Corp., St. Paul, MN) connected to the stage of a microscope that was equipped with epifluorescence optics.

Experiment 2. Viral transport studies The second experiment was designed to determine whether a P R V infection in one functional class of SPNs produced infections in nearby SPNs. A similar two-stage surgical experiment was performed in which 2 separate sympathetic targets were injected: one with a retrograde marker and the other with PRV. In the first set of these experiments CTb was injected in the right stellate ganglion (400 nl of a 1% solution of CTb) and then, after 2 - 5 days postoperative recovery, the right adrenal gland was injected with P R V (100 nl, Bartha's PRV, titer = 2 × 107 p f u / m l ) . The converse experiment was

105 also performed in a second group of rats. Both sets of rats were allowed to survive an additional 3 or 4 days following the PRV injection and were then processed for histochemical studies as described above. Serial sections of the T4-T s spinal cord were cut at 50/zm in the longitudinal plane, placed in a 0.02 M KPBS solution containing both a 1:10,000 dilution of goat anti-CTb antibody and a 1:4,000 dilution of mouse polyclonal anti-PRV (kindly supplied by Dr. T. Ben-Porat), containing 5% normal donkey serum and 0.3% Triton X. The sections were incubated on a shaker overnight at room temperature. The following day, they were washed twice in KPBS, transferred to a solution containing a 1:100 dilution of biotinylated affinity-purified donkey anti-mouse serum (Jackson ImmunoResearch Laboratories, West Grove, PA) in 5% normal donkey serumKPBS, incubated for 4 h on a shaker at room temperature, washed twice in KPBS, and then, transferred to a solution containing both 1:100 streptavidin-RITC (Jackson Lab) and 1:50 FITC-conjugated affinity-purified donkey anti-goat (Jackson Lab) in 5% normal donkey serum in KPBS (4 h, 25°C). The sections were then washed in KPBS (2 x), mounted on gelatin-coated slides, coverslipped, viewed with a microscope equipped for epifluorescence optics and mapped with a X - Y plotting system. The brains were cut in the transverse plane at 50 /~m on a freezing microtome and a l-in-5 series of tissue sections was incubated in a 1:10,000 dilution of a pig polyclonal anti-PRV antibody (kindly supplied by Dr. K.B. Platt, Iowa State Univ., Ames, Iowa) and stained with a 1:250 dilution of FITC-streptavidin-biotin complex (Sigma, St. Louis). The sections were mounted on glass slides, coverslipped with Krystalon (EM Diagnostic Systems, Gibbstown, N J), and examined with epifluorescence optics for PRV-labeled cells.

The a priori conditions for selecting spinal cord material for analysis depended on two factors: unilateral PRV labeling in the IML and transneuronal labeling in the brain. Because there was a need to have high quality histological material for accurate cell counts, some very stringent conditions were placed on this sampling process. Successful staining of both CTb and PRV was obtained in the spinal cord of about 50% of the rats (100 out of 178). Only 26 rats from this population were used in the present study and the exclusion of the other data deserves comment. Within the set of successful experiments, two types of problems were encountered. First, some variations in the quality of the histology forced us to discard some material because the intensity of the staining was too weak for accurate cell counts. Second, the PRV labeling was highly variable ranging in some experiments from only a few cells per segment to the other extreme in which there were cases containing robust bilateral IML labeling accompanied by a severe gliosis, Both extremes were avoided. The rationale for this exclusion was that neither type of experiment would be representative of the type of experimental material we have used in our previous studies 21'22. Thus, all the rats selected for analysis at the 4 day survival period had unilateral PRV labeling in the IML with only a few contralateral labeled SPNs (approx. 8-10 cells) and transneuronal labeling in the brain. The transneuronal labeling correlated well with our previous results 22'23. Namely, labeling was found in the rostroventrolateral medulla, ventromedial medulla, caudal raphe nuclei, A5 cell group, and paraventricular hypothalamic nucleus. Additionally, some variation was noted in the stellate ganglion experiments where only a small number of the rats displayed labeling in the lateral hypothalamic area and zona incerta; these areas showed labeling only if there had been intense labeling in the IML that

® o~

50

Stellate Ganglion total = 2132 ± 180

25 o

"6

m

Gland total = 835 • 37 Adrenal

25 o

C8

®

T1 T2

T3 T4

T5

T12 T13 L1 L2

T6 T7 T8 T9 T10 Tll Spinal Segment

GSte;latoe~/ ~~~*Cholera Toxin

I-~ ¢/)

T5

o

300

ilion

d

250

"10

--~ 200 ¢1 .,a

o~

T6

._1

t50 100

.Q

E

T7

50

Z

0

Gland "

~uoro-Gold

T5 T6 T7 Spinal Segment

Fig. 1. A: segmental distribution of labeled sympathetic preganglionic neurons after Fluoro-Gold injections into the stellate ganglion and adrenal gland taken from Strack et al. 21. Each value is an average of data collected from five rats and the number provided on the right side of each graph indicates the total number of labeled cells in the entire spinal cord + SEM. To obtain the raw data for these experiments, see Table I of ref. 21. B: experimental design in which different retrograde markers were injected into two different sympathetic targets--the adrenal gland and stellate ganglion. C: segmental distribution of labeled sympathetic preganglionic neurons in the T5-T 7 spinal segments in double labeling experiments. Values indicated are the mean +_S.E.M.

106

.g Adrenal = t

11

o

T5

c

o

O!

gland

Stellate 0

ganglion=

~

t

Double- =1~ labeled

o d~

0

©

o°~

T6

( O~

I 250 ~m J

Fig. 2. Distribution of labeled sympathetic preganglionic neurons in the T 5 and T 6 segments of a horizontal section through the intermediolateral cell column (IML) after CTb was injected into the stellate ganglion and Fluoro-Gold was injected into the adrenal gland. Note that sympathetic preganglionic neurons that innervate different sympathetic targets lie near each other in the same cell clusters. In this section, only one cell was double labeled, as indicated by the arrow.

produced gliosis. Finally, one exception in our sampling process needs to be noted: some rats used in the 3-day survival period experiments had excellent labeling in the spinal cord but did not have transneuronal labeling in the brain. The latter result is most likely due to too short of a postinjectiion survival period for optimal viral replication in transneuronally infected neurons. Control histochemical reactions The specificity of the immunohistochemical reagents was examined in a set of control experiments. Tissues containing only CTb or PRV retrogradely labeled neurons were stained by the double immunolabeling procedure to verify that false positives did not occur; none were found. Similarly, omission of primary antibodies and staining with secondary antibodies alone did not result in cell body labeling.

RESULTS

IML as a test site for PRV retrograde labeling method The IML was chosen as a test site for two reasons: (1) some SPNs innervating the stellate ganglion lie within the same cell clusters of the I M L as those innervating the adrenal gland (Expt. 1); and (2) the number of SPNs found in the T s - T 7 levels of the I M L

innervating these two targets was small. As reported earlier, about 50 to 140 adrenal SPNs and 40 to 320 stellate SPNs are found in each of these spinal segments 21. From a practical standpoint, a small sample size such as this is a particularly convenient feature for quantitative studies in which systematic cell counts of all the labeled neurons were performed. Expt. 1. Double retrograde labeling experiments. The stellate ganglion and the adrenal gland are innervated by SPNs originating in nine different segments of the rat spinal cord 21. These two populations of neurons share a common segmental origin in the T a - T 8 spinal segments (Fig. 1A). As shown in Fig. 1B, when two different retrograde cell body markers were injected in these targets in the same animal, a similar overlapping distribution was obtained. In addition, the number of SPNs labeled in the present experiments correlated well with our earlier findings 2~ (The raw data for Fig. 1A can be obtained Table I of ref. 21; these data are highly similar to the cell counts obtained in the present study and presented in Fig. 1C).

Fig. 3. Sympathetic preganglionic neurons innervating different sympathetic structures lie within the same cell clusters. A: stellate sympathetic preganglionic neurons were retrogradely labeled with CTb (orange) and sympathoadrenal preganglionic neurons were retrogradely labeled with Fluoro-Gold (white). B: PRV retrogradely labeled sympathoadrenal SPNs (red) and CTb retrogradely labeled stellate SPNs (yellow) lie within the same cell clusters (3 days PRV postinjection). C: at 4 days postinfection, the number of P R V retrogradely labeled sympathoadrenal SPNs (orange) increased and did not affect the CTb retrogradely labeled stellate sympathetic preganglionic neurons (green). D: in some cases, PRV infections produced a local gliosis in the IML (red). As shown here, there were cases in which the glial reaction did not affect the viability of local stellate sympathetic preganglionic neurons (green-yellow). All of the examples used here also exhibited PRV transneuronal labeling in the brain. Bar = 100 txm.

107 n

108 TABLE I

Doubled labeled SPNs following PRV infections of specific types of SPNs Results are expressed as m e a n +_SEM. There was no statistical difference between individual spinal segments in a given group or between control and experimental groups of double labeled cells as determined by Student's t test. CTb, cholera toxin /3-subnunit; FG, Fluoro-Gold; IML, intermediolateral cell column; SPNs, sympathetic preganglionic neurons.

Treatment

Number of animals

Total number labeled cells in T5- T 7 ipsilateral IML

Number of animals double-labeled cells

Percentage of double-labeled cells in total population

Control: FG-Adrenal SPNs + CTb-Stellate SPNs

5

656__+ 126

4_+ 1

0.6%

4 6

1066 __+293 1 144__+146

8 __+6 10_+5

0.8% 0.9%

5

584_+ 38

2_+1

0.3%

5 6

640_+ 61 821-+ 93

5_+3 4_+2

0.8% 0.5%

Experimental: PRV-Adrenal SPNs + CTb-Stellate SPNs 3 day 4 day CTb-Adrenal SPNs + PRV-Stellate SPNs 3 day 4 day

without gliosis with gliosis

Adrenal and stellate SPNs lie near each other in the T s - T 7 levels of the IML, often residing in the same cell clusters (Figs. 2 and 3A,B). This spatial property makes this spinal nucleus an ideal test site for the analysis of the transneuronal viral labeling method. A second important feature that makes the IML an excellent system in this type of experiment is the fact that very few SPNs innervated both the sympathetic targets (Table I). Only 4 out of 656 SPNs found in the T s - T 7 levels project to both sympathetic targets. This clearly indicates this phenomenon is a rare event and supports the findings of Appel and Elde ~ who reported that SPNs project only to target-specific structures. This property was important for the design of Expt. 2 because if PRV infected nearby SPNs it would cause an increase in the number of double-labeled SPNs and thus, would serve as an excellent indicator that PRV infection spread in the IML in a non-specific manner. By using such a system, any change in the number of double-labeled SPNs could be easily detected in a quantitative study.

Expt. 2. Effect of PRV retrograde cell body labeling on nearby neurons in the IML. Prior to beginning this study, we hypothesized that if PRV produced nonspecific infections in the IML, two potential results may occur: (1) the number of double-labeled SPNs would increase or (2) PRV may induce a local glial reaction which would produce secondary cytopathic effects on nearby SPNs resulting in a decrease in the neuronal population in the IML. Of course, both resuits could occur concurrently as well. Our data indi-

cate that an increase in the number of double-labeled SPNs did not occur (Table I); this finding strongly suggests that PRV does not infect nearby neurons in a non-specific manner. Fig. 4 summarizes the effect local PRV infections have on nearby SPNs. PRV infections of sympathoadrenal SPNs had no effect on the number of CTb-labeled stellate SPNs (Fig. 4A), but the results of the converse experiment were more complicated. Under these conditions at 4 days postinjection, two groups could be distinguished: one group showed no reduction in the number of sympathoadrenal SPNs (Fig. 4B) and a second group exhibited a loss of SPNs along with a concurrent gliosis (Fig. 4B). Even within this latter subgroup, the results were not uniform. There were individual cases in which a severe gliosis developed but this cellular reaction did not produce a decrease in the number of CTb retrogradely labeled sympathoadrenal SPNs. Additionally, although a rare occurrence, there were individual cases in which PRV was injected into the adrenal gland that showed a marked gliosis in the area of the IML (Fig. 3D). These rats, however, did not show a decrease in the number of CTb stellate-labeled SPNs. While our experiments do not resolve the reason for this variation, they do, however, indicate local SPNs (i.e non-PRV labeled) may remain viable, even in the presence of a local viral infection affecting the SPNs within the same cell cluster of the IML. Nevertheless, when a severe gliosis develops in the region of the first-order neurons, it is best not to use this type of material for neuroanatomical investigations.

109

Stellate Ganglion~

450f

C,'o,era

,,

Ic~.....I~'~ I

~

Toxin

-~

T5

Effect of Local PRV ] Infection on Stellate SPN Labeling /

350

m

Control, n=6|

T6

"6~ 150 t

5op

T7 Adrenal

:i:::::~

-- PRV

T5

Gland

N

T7

Spinal Segment

Stellate Ganglion~

200

,~~~RV

"~

T5

150

n

,1oo

T6 T7

T6

Effect of Local PRV Infection on Adrenal SPN Labeling m Control,n=6 q" 3 Day,,n=s T~I ~ 4 Days, n=5 I~1 4 Days, n=6 - gliosis ~ , ~

~ 50 ~;

Cholera

Toxin

T5

T6

T7

Spinal Segment

Fig. 4. P R V infections of one functional class of SPNs did not reduce the population of nearby sympathetic preganglionic neurons, except in some cases when a severe glial reaction developed.

DISCUSSION The results presented here demonstrate an important property regarding PRV as a transneuronal marker: when first-order neurons are retrogradely labeled with PRV, PRV does not produce non-specific infections in nearby neurons. This finding rules out one of the conditions that could be a potential source of

false positive transneuronal labeling (Fig. 5). While this observation supports our contention that this virus can be used as a specific retrograde transneuronal marker, it does not exclude the possibility that first order labeled neurons release PRV which becomes disseminated in the neuropil. Under these conditions, PRV could be taken up by synapses impinging on other local neurons and thus, produce false-postive labeling via a

110

Retrograde Transneuronal Labeling Specific

Nonspecific

2ndOrder

1stOrder Extracetlular Space

•i!tl

[

Target~B

Fig. 5. W h e n PRV is injected into a specific target, the virus is retrogradely transported to the nucleus of the neuron innervating that structure. The virus undergoes replication in the infected neuron, and subsequently can produce transneuronal labeling by specific transneuronal labeling of the neural circuits synapsing on the infected neuron (1) and become disseminatied in the extracellular space (2) to cause non-specific transneuronal labeleing by two potential routes: uptake by local, non-related synapses (3) and infections of nearby, non-related neurons (4). The experiments presented here rule out condition 4 as a mechanism of retrograde transneuronal labeling.

non-specific uptake of PRV at nearby synaptic sites. To test of this hypothesis, two different types of transneuronal markers of equal efficiency would be required: PRV and some type of non-viral transneuronal marker. To date, this second type of neuroanatomical tracer remains unavailable and so this particular problem cannot be analyzed. The subject of viral specificity has been addressed in an earlier study in which experiments were performed indicating that Bartha PRV is a highly specific transneuronal marker in the sympathetic nervous system 2°. The most compelling argument came from the observation that PRV injections into specific targets of the head (viz., eye and ear) produced specific neuronal labeling patterns in the IML that correlated remarkably well with earlier physiological data regarding the spinal origins of sympathetic outflows to these target

tissues. Using a similar argument, Zemanick and coworkers 2v demonstrated that two forms of HSV-1 produced specific transneuronal labeling patterns in the primate motor system. Both of these observations, therefore, strongly support the idea that viruses may be used as specific transneuronal markers. Li and coworkers ~ examined the specificity of HSV-1 in the IML using a similar experimental paradigm as presented here. The fluorescent dye DiI was injected in one sympathetic target and HSV-1 in another. They found that 5 - 1 2 % of the SPNs were double-labeled and this number remained relatively constant during the survival period of 5 to 8 days. It is impossible to evaluate the significance of this finding because these investigators did not present control experiments using two different conventional retrograde markers. Since almost all sympathetic outflows arise from targetspecific sets of SPNs t, it is conceivable that their finding of 5-12% double-labeled SPNs is really evidence that HSV-I is a non-specific marker. Moreover, our results indicate that the occurrence of double labeled SPNs was a rare event in normal and well as PRV infected material. Thus, the issue of whether HSV-1 is a specific transneuronal marker needs further investigation. In the experiments presented here, it is clear that PRV does not spread from neuron-to-neuron in an indiscriminate manner. However, there are several observations that indicate that PRV may not be a perfect transneuronal marker and two examples will be presented. First, when PRV is injected in the adrenal gland (or any sympathetic ganglion) it predominantly causes retrograde labeling in the IML, but for unexplained reason it frequently produces cell body labeling in the ventral horn 22. This bugbear has cropped up regularly in our studies of central sympathetic pathways and has forced us, in some studies, to be highly discriminatory in our selection of histological material. A similar problem has been encountered by Cabot and his coworkers 3 in their studies using the C fragment of tetanus toxin as a transneuronal marker. They suggested this unexpected ventral horn labeling may be due to spillage or leakage of the marker during the injection procedure. However, this explanation does not totally exclude the possibility that some other type of transneuronal transfer may occur. Second, one unusual labeling pattern occurred in our PRV transneuronal study of the CNS cell groups that innervate the motor outflow to skeletal muscle 16. Overall, the results correlated well with earlier anatomical and physiological studies, except for one finding: when a-motoneurons of one specific motor column

111 were retrogradely labeled with PRV, apart from the expected transneuronal labeling in the intermediate spinal gray matter, there was a group of putative amotoneurons ventral to the motor column under study that was consistently labeled. This finding was interpreted to be due to dendrodendritic viral spread. Since dendrites of a-motoneurons often radiate beyond the boundaries of their motor column, it was envisioned that PRV was shed from dendrites of infected motoneurons that extend into the extracolumnar neuropil. These extracolumnar viral infections were confined solely to the motoneurons innervating appendicular muscles and did not involve motoneurons innervating axial muscles. For this reason, we suggested that this may not be a spurious finding. Regardless of the cellular mechanism producing the results described above, it is clear that careful analysis of the labeling pattern of the first-order neurons is crucial for the successful interpretation of data generated by this technique. Variations in severity of infection of the first-order neurons greatly affect the specificity of the transneuronal labeling pattern seen in the CNS. This suggests that for each neural system there is a narrow range in which specific transneuronal labeling occurs; deviations from this critical range result in either no labeling or non-specific 'over-infections'. The optimal amount of virus needed to produce a specific infection for a particular system needs to be established in an empirical fashion. Once this is determined and a series of experiments is underway, it is important to bear in mind that some extreme variations will still occur. This variability probably relates to the fact outbred rats are used in these experiments and just like humans who show individual differences in their susceptibility to viral infections, the rats also react differently. Since it is impossible to know a priori which cases will result in specific transneuronal labeling, a rigorous comparison of the first-order labeling pattern obtained with PRV and a conventional retrograde marker (e.g. Fluoro-Gold, CTb) is critical for defining the individual experiments that should be accepted for any given neuroanatomical study. We have stressed the need for this approach in our previous studies 16'19-23. For example, in our study of the CNS inputs to the pterygopalatine parasympathetic preganglionic neurons, PRV cases were selected on the basis of the PRV labeling pattern in the parasympathetic preganglionic nucleus that matched the results obtained in the control material with Fluoro-Gold 19. Thus, just like in any other neuroanatomical study, a judicious selection of cases is necessary. Finally, it should be recognized that when new results are found with the viral transneuronal labeling method, the data should still be consid-

ered to be provisional until they are confirmed with other neuroanatomical tracing methods. Before the viral transneuronal labeling technique gains widespread use, some additional problems need to be recognized. These include the fact that neurotropic viruses have tropisms for particular classes of neurons and therefore, any one virus may not be a universal transneuronal tracer. An example of this is the apparent refractory nature of certain classes of neurons such as the mesencephalic trigeminal sensory neurons to PRV infection II. Similarly, we observed that Bartha PRV was transported from skeletal muscle to dorsal root sensory neurons, but these neurons failed to showed infections after 4 days survival time 16. Similar limitations have been discovered for HSV-1; some strains of HSV-1 are selective anterograde transneuronal markers and others behave as selective retrograde transneuronal markers 27. The reason for these differences is not understood but probably relates to several factors, including the types of viral receptors present on a neuron, the ability of the virus to enter the neuron, and the axoplasmic transport properties of the virus. A final point deserves comment: glial reactions sometimes occur during PRV infections. In most of the cases used in this and previous studies from our laboratory, we have avoided using histological material with marked gliosis. However, there were some instances in which a profound glial reaction developed without any apparent effect on the nearby, non-PRV infected neurons. In other cases, an intense gliosis occurred in regions of the IML that were formerly occupied by neurons. The glial cells appeared to invade this region and partition it from the rest of the IML. Thus, glial responses may function to restrict the extracellular spread of virus by isolating it and phagocytosing infected neurons 15. There is no evidence to suggest that PRV infections are spread via glial cells 22. Finally, Rinaman et al. 15 suggested that glial cells were incapable of producing mature forms of the virus necessary for productive infections and cogently argued that two factors - - isolation of an infected zone and inability to synthesize mature forms of virus - - greatly favors the likelihood that PRV infections spread from one neuron to another in synaptically related sets of neurons. Since herpesviruses bind preferentially to synaptosomes as opposed to cell membranes from neuronal perikarya and astrocytes 26, this strengthens the argument that viral transneuronal labeling occurs mainly at synapses. If all of these factors are taken into consideration including selection of the appropriate attenuated virus, choosing the minimal viral concentration necessary to produce controlled infections of first order neurons,

112 and critical evaluation of labeling patterns produced in the first-order neurons, the viral transneuronal labeling method undoubtedly will continue to emerge as an extremely useful technique for deciphering CNS connections. Acknowledgements. This work was supported by the National Institute of Heart, Lung, and Blood of the National Institutes of Health (HL 25449). We thank Dr. K. Platt for supplying us with pseudorabies virus, Dr. T. Ben-Porat for the mouse anti-PRV antiserum, and Mr. Xay Van Nguyen for excellent technical assistance.

REFERENCES 1 Appel, N.M. and Elde, R.P., The intermediolateral cell column of the thoracic spinal cord is comprised of target-specific subnuclei; evidence from retrograde transport studies and immunohistochemistry, J. Neurosci. 8 (1988) 1767-1775. 2 Blessing, W.W., Li, Y.W. and Wesselingh, S.L., Transneuronal transport of herpes simplex virus from the central vagus to brain neurons with axonal inputs to central vagal sensory nuclei in the rat, Neuroscience, 42 (1991) 261-274. 3 Cabot, J.B., Mennone, A., Bogan, N., Carroll, J., Evinger, C. and Erichsen, J.T., Retrograde, trans-synaptic and transneuronal transport of fragment C of tetanus toxin by sympathetic preganglionic neurons, Neuroscience, 40 (1991) 805-823. 4 Card, J.P., Rinaman, L., Schwaber, J.S., Miselis, R.R., Whealy, M.E., Robbins, A.K. and Enquist, L.W., Neurotropic properties of pseudorabies virus: uptake and transneuronal passage in the rat central nervous system, J. Neurosci., 10 (1990) 1974-1994. 5 Card, J.P., Whealy, M.E., Robbins, A.K., Moore, R.Y. and Enquist, L.W., Two a-herpesvirus strains are transported differentially in the rodent visual system, Neuron, 6 (1991) 957-969. 6 Card, J.P., Whealy, M.E., Robbins, A.K., Moore, R.Y. and Enquist, L.W., Pseudorabies virus envelope glycoprotein gI influences both neurotropism and virulence during infection of the rat visual system, Z ViroL, 66 (1992) 3032-3041. 7 Gieroba, Z.J., Li, Y.W., Wessenlingh, S.L. and Blessing, W.W., Transneuronal labeling of neurons in rabbit brain after injection of herpes simplex virus type 1 into the aortic depressor nerve, Brain Res., 558 (1991) 264-272. 8 Jansen, A.S.P., Ter Horst, G.J., Mettenleiter, T.C. and Loewy, A.D., CNS cell groups projecting to the submandibular parasympathetic preganglionic neurons in the rat: a retrograde transneuronal viral cell body labeling study, Brain Res., 572 (1991) 253-260. 9 Li, Y.W., Ding, Z.Q., Wesselingh, S.L. and Blessing, W.W., Renal and adrenal sympathetic preganglionic neurons in rabbit spinal cord: tracing with herpes simplex virus, Brain Res., 573 (1992) 147-152. 10 I_x3ewy, A.D., Bridgman, P.C. and Mettenleiter, T.C., /3-Galactosidase expressing recombinant pseudorabies virus for light and electron microscopic study of transneuronally labeled CNS neurons, Brain Res., 555 (1991) 346-352. 11 Martin, X. and Dolivo, M., Neuronal and transneuronal tracing in the trigeminal system of the rat using herpes virus suis, Brain Res., 273 (1983) 253-276.

12 McLean, J.H., Shipley, M.T. and Bernstein, D.I., Golgi-like, transneuronal retrograde labelling with CNS injections of herpes simplex virus type-l, Brain Res. Bull., 22 (1989) 867-881. 13 Nadelhaft, I., Vera, P.L., Card, J.P. and Miselis, R.R., Central nervous system neurons labelled following the injection of pseudorabies virus into the rat urinary bladder, Neurosci. Lett., 143 (1992) 271-274. 14 Norgren, R.B. and Lehman, M.N., Retrograde transneuronal transport of herpes simplex virus in the retina after injection in the superior colliculus, hypothalamus and optic chiasm, Brain Res., 479 (1989) 374-378. 15 Rinaman, L., Card, J.P. and Enquist, L.W., Spatio-temporal responses of astrocytes, ramified microglia and brain macrophages to central neuronal infection with pseudorabies virus, J. Neurosci., 13 (1993) 685-702. 16 Rotto-Percelay, D.M., Wheeler, J.G., Osorio, F.A., Platt, K.B. and Loewy, A.D., Transneuronal labeling of spinal interneurons and sympathetic preganglionic neurons after pseudorabies virus injections in the rat medial gastrocnemius muscle, Brain Res., 574 (1992) 291-306. 17 Rouiller, E.M., Capt, M., Dolivo, M. and De Ribaupierre, F., Tensor tympani reflex pathways studied with retrograde horseradish peroxidase and transneuronal viral tracing techniques. Neurosci. Lett., 72 (1986) 247-252. 18 Rouiller, E.M., Capt, M., Dolivo, M. and De Ribaupierre, F., Neuronal organization of the stapedius reflex pathways in the rat: a retrograde HRP and viral transneuronal tracing study. Brain Res., 476 (1989) 21-28. 19 Spencer, S.E., Sawyer, W.B., Wada, H., Plan, K.B. and Loewy, A.D., CNS projections to the pterygopalatine parasympathetic preganglionic neurons in the rat: a retrograde transneuronal viral cell body labeling study, Brain Res., 534 (1990) 149-169. 20 Strack, A.M., and Loewy, A.D., Pseudorabies virus: a highly specific transneuronal cell body marker in the sympathetic nervous system, J. Neurosci., 10 (1990) 2139-2147. 21 Strack, A.M., Sawyer, W.B., Marubio, L.M. and Loewy, A.D., Spinal origin of sympathetic preganglionic neurons in the rat, Brain Res., 455 (1988) 187-191. 22 Strack, A.M., Sawyer, W.B., Platt, K.B. and Loewy, A.D., CNS cell groups regulating the sympathetic nervous outflow to adrenal gland as revealed by transneuronal cell body labeling with pseudorabies virus, Brain Bes., 491 (1989) 274-296. 23 Strack, A.M., Sawyer, W.B., Hughes, J.H., Platt, K.B. and Loewy, A.D., A general pattern of CNS innervation of the sympathetic outflow demonstrated by transneuronal pseudorabies viral infections, Brain Res., 491 (1989) 156-162. 24 Ugolini, G., Kuypers, H.G.J.M. and Simmons, A., Retrograde transneuronal transfer of herpes simplex virus type 1 (HSV1) from motoneurons, Brain Res., 422 (1987) 242-256. 25 Ugolini, G., Kuypers, H.G.J.M. and Strick, P.L., Transneuronal transfer of herpes virus from peripheral nerves to cortex and brainstem, Science, 243 (1989) 89-91. 26 Vahlne, A., Svennerholm, B., Sandberg, M., Hamberger, A. and Lycke, E., Differences in attachment between herpes simplex type 1 and type 2 viruses to neurons and glial cells, Infect. Immunol., 28 (1980) 675-680. 27 Zemanick, M.C., Strick, P.L. and Dix, R.D., Direction of transneuronal transport of herpes simplex virus 1 in the primate motor system is strain dependent, Proc. Natl. Acad. Sci. USA, 88 (1991) 8048-8051.