- Email: [email protected]
A morphometric analysis of transmitter identified dendrites and nerve terminals
Brain Reseurch
Bulletin, Vol. 9, pp. 53-60,1982.Printed in the U.S.A.
A Morphometric Analysis of Transmitter ,Identified Dendrites and Nerve Terminals L. F. AGNATI,
K. FUXE,’ L. CALZA, T. HijKFELT, 0. JOHANSSON, F. BENFENATI AND M. GOLDSTEIN
DPpartment of Human Physiology, University of Modena, 41100 Modena, Italy Department of Histology, Karolinska Institutet, Box 60400, S-104 01 Stockholm, Sweden and Department of Psychiatql, New York University Medical Center, New York, NY
AGNATI, L. F., K. FUXE, L. CALZA, T. HGKFELT,
F. BENFENATI AND M. GOLDSTEIN. A BRAIN RES. BULL. 9(1-6) 5?-60, 1982.-The present method is exemplified on coronal sections of the medulla oblongata containing phenylethanolamineN-methyltransferase (PNMT) immunoreactive nerve cell bodies and their processes and on coronal sections of the pons containing the locus coeruleus, where PNMT immunoreactive nerve terminals have been demonstrated together with the dopamine+-hydroxylase immunoreactive nerve cell bodies. Morphometric analysis of the processes (both length and branches) and of the nerve terminals involve as a first step the division of the area under study into squares 100 pm wide, which are superimposed on a Cartesian plane. The uniformity of the density distribution of the nerve terminals and the processes (branches or length) can be analyzed by Lorenz curves, which in a quantitative way can measure the degree of unevenness and thus represent a measure of concentration. A concentration index can therefore be calculated. By the use of the densitometric approach it also becomes possible to study the density distribution of the nerve terminals with the highest antigen contents. The present method will make it possible to quantitate morphological changes occurring in processes and nerve terminals of transmitter-identified neurons.
morphometricanalysis
Morphometric
analysis
of transmitter
Transmitter
identified
0. JOHANSSON,
dendrites
identified neurons
IN a large number of studies the distribution and morphology of catecholamine (CA) nerve terminals and dendrites have been described by subjective evaluation in the microscope or
by inspection of microphotographs, using Falck-Hillarp methodology for the cellular demonstration of CA stores or using immunohistochemical methods to evaluate the cellular localization of tyrosine-hydroxylase, dopamine-phydroxylase (DBH) and phenylethanolamine-N-methyltransferase (PNMT) [8, 10, 16, 19, 20, 211. In recent papers [4,9] methods for the morphometric characterization of transmitter-identified nerve cell groups have been introduced together with objective criteria for defining nerve cell groups in subpopulations of nerve cells within one single nerve cell group. In the present paper we introduce a morphometric analysis of transmitter-identified neurons at the dendritic and nerve terminal level. The present method is illustrated on one coronal section of the rostra1 part of the medulla oblongata showing PNMT positive cell bodies and their processes [ 193, and on adjacent coronal sections of the pons in which DBH immunoreactive cell bodies or PNMT ‘Please send reprint requests to Professor Kjell Fuxe, Department Sweden.
Copyright 0 1982 ANKHO
and
nerve
terminals.
Medulla oblongata
immunoreactive locus coeruleus
Locus coeruleus
nerve [ 191.
terminals
are demonstrated
in the
METHOD Male specific pathogen-free Sprague-Dawley rats (b.wt. 150 g) were used. For the analysis of medulla oblongata the rats were pretreated with colchicine (75 Fg) intraventricularly 24 hr before killing. All rats were perfused with 50 ml Tyrode solution followed by 500 ml icecold 4% (W/V) paraformaldehyde in 0.1 M phosphate buffer [23] for 30 minutes. After perfusion the medulla oblongata and the pons were dissected out and immersed in the same fixative for 90 minutes. The brain tissues were then rinsed in 5% (W/V) sucrose in 0.1 M phosphate buffer for at least 24 hr. The medulla oblongata was cut on a freeze microtome. The immunoperoxidase procedure was adopted from Grzanna et al. [14]. Briefly, 50 pm freeze microtome sections of colchicine-treated, Formalin-fixed brains, were incubated with the PNMT antibodies (diluted 1: 160), followed of Histology, Karolinska Institutet,
Box 60400, S-104 01 Stockholm,
International Inc.-0361-9230/82/070053-08$03.00/O
AGNA’I‘I
FIG. 1, Coronal section of the dorsomediai part of the rostrai medulla oblongata where PNMT positive nerve cell bodies and processes of adrenaline cell group C2 [I83 have been visualized by means of PAP technique (see text). The vertical arrow point in the dorsal direction and the horizontal arrow in the medial direction. x300.
t7’ Al
MORPHOMETRIC
ANALYSIS
OF MONOAMINE
PROCESSES taining neurons. The purgation the production of the antiserum ously [ll, 12, 131.
of this enzyme as well as has been described previ-
RESULTS
Morphometric Analysis of PNMT Immunoreactive Nerve Cells of the Adrenaline Cell Group C2 at the Dendritic Level
FIG. 2. Density distribution of length of processes and number of branches of PNMT positive cells in the adrenaline cell group C2. Part of the original microphotograph is shown in Fig. 1. The whole montage is not shown, since the processes are not clearly shown, when the montage is reduced to publication format. For further details, see text.
by sheep antiserum to rabbit immunoglobulin (diluted 1:50), and horseradish peroxidase-antiperoxidase (PAP) complex (diluted 1:40; the PAP complex was a generous gift from Prof. L. A. Sternberger, Department of Anatomy and Center for Brain Research, University of Rochester, Rochester, NY). Finally, they were treated with diaminobenzidine and hydrogen peroxide, followed by OsO,, dehydrated in ethanol and xylene, and coverslipped in Entellan. The sections were examined in a Leitz Orthoplan light microscope using bright-field optics. Kodak Panatomic-X film was used for photography. For further details on the methodology used, the reader is referred to Johansson and Backman (in preparation). The pons was cut in a cryostat (Dittes, Heidelberg, West-Ge~any) to give sections 10 pm thick; these were processed according to the indirect immunofluorescence technique of Coons [6], essentially as described by Hokfelt et al. [ 171. Briefly, repeated sequences of two adjacent sections containing the locus coeruleus were used. One of them was incubated with DBH antiserum (dilution 1: 100) and the other with PNMT antiserum (dilution 1: 100) ,at +4”C for 24 hr, rinsed in phosphate buffered saline (PBS), tpcubated with fluoresceine-isothiocyanate (FITC) conjugated sheep antirabbit antibodies (Statens Bakteriologiska Laboratorium, SBL, Stockholm, Sweden) for 30 minutes at +37”C, rinsed in PBS mounted in a mixture of glycerol and PBS (3:l) and examined in a Zeiss Standard fluorescence microscope equipped with an oil dark field condenser and a Schott BG 12 (3 or 4 mm) and a Zeiss 56 stop filter. All sera containing 0.3% T&on-X-100 [15] to solubilize membranes and in this way facilitate penetration. Scopix RP 1 black and white film (Gaevert, Belgium) was used for photography. The antiserum against DBH was used as a marker for the noradrenaline (NA) cell bodies of the locus coeruleus. The purification of the enzyme as well as the production of the antiserum has been described in detail [ll, 12, 131. The antiserum produced against the adrenaline synthesizing enzyme PNMT was used as a marker for the adrenaline con-
The morphometric procedure is examplified on a coronal section of the rostra1 medulla oblongata which in its dorsomedial part contain the PNMT positive nerve cell bodies of the adrenaline cell group C2 (see Fig. 1). The morphometric procedure starts with a Cartesian representation of the localization of the various branches which has been obtained by taking the midline as the x-axis and a straight line tangential to the ventral border of the fourth ventricle and perpendicular to the midline as the y-axis. Thereafter, squares with a side of 100 pm are superimposed on this Cartesian plane and an evaluation of the number of branches and of the length of the branches have been carried out for each square (Fig. 2). As seen in Fig. 2, the PNMT positive processes and thus the number of branches are especially rich within the central part of the adrenaline cell group. This is a feature of a nerve cell group of transmitter-identi~ed neurons which can be of interest for, e.g., comparisons between different nerve cell groups or for comparisons within one and the same cell group at different rostra&caudal levels. Therefore, it is important to develop a method which can quantitate the density and uniformity of distribution of the length of processes and number of branches in order to have an indication of the pattern of innervation within a certain area. This is possible by studying the uniformity of the density distribution of a certain type of parameter using standard statistical techniques such as the Lorenz curves to determine the concentration [22]. The concept of concentration as used in the present study has been extensively employed in studies of population density [?I. The measure of concentration gives the degree of unevenness in quantitative terms. The Lorenz curve is a diagram of the con~ent~tion, that is of the unevenness of the dist~bution of a certain parameter such as number of branches or length of processes among N squares, in which the brain area analyzed has been subdivided. The principle according to which a Lorenz curve is constructed is illustrated in Fig. 3 [25]. Thus, let us consider the n quantities x,, xB . . . xk . . . x,, where xL
AGNAT
Sh
ITHELORENZCURVE AND THE INDEXOF CONCENTRATION 1
k,‘TAL.
rm uwmyz CUMS MD M IWXES OF CCNCENTRATIOU FOR TH )ISTRtBUTW Cf PROCESS LENGTHAM BRANCHlWOFo*BER IN TIE C2 AREA @I&IT-m PROFILES ) BRANCHING NUMBER DISTIWTION~ R = 0 5, PROCESS LENGTH USTRlEUTlON R =
FIG. 3. Schematic representation of the Lorenz curve and the index of concentration (=R). On the x-axis the quantities P,=k/n are reported, where k= I,? , n and in the present case are the squares of Fig. 2 ranked according to, e.g., the number of branches. On the y-axis the values qk=VkN,, are given, where V,=x, + x2 + + xk and xk represents the value of the parameter under study (e.g., number of branches) for the Kth square. It should be noted that xII
tween the observed index of concentration
k =
and the maximal concentration. is represented by the ratio
( ;; ) (Pk-q&
( z:
Pk
0
0
O.lO
0.20
0.30
0.40
0 50
0.60
0.70
080
000
Pk E yn
FIG. 4. Lorenz curves and the concentration indexes for the distribution of the number of branches and of the length of processes of the PNMT immunoreactive neurons of adrenaline cell group C? shown in Figs. I and 2.
The
)
This ratio is in the case of an even distribution zero and in the case of a maximal concentration one. The Lorenz curves and the concentration indexes for the distribution of the number of branches and of the length of processes shown in the various squares of Fig. 2 are shown in Fig. 4. Morphonwtric Tmn inm1.s
Ancrlysis
of‘ PNMT lmmunorcac.ti~c
Nrr\lc,
The analysis will be exemplified on PNMT positive nerve terminals within the locus coeruleus. The noradrenaline nerve cell bodies of the locus coeruleus have been demonstrated by means of a DBH antiserum in a section adjacent to the section showing the PNMT immunoreactive nerve terminals within the locus coeruleus. Also in this case the first step in the morphometric procedure is to obtain a Cartesian representation of the localization of the PNMT nerve terminals and of the DBH immunoreactive nerve cell bodies. which probably are innervated by the PNMT immunoreactive nerve terminals. A schematic representation of this procedure is seen in Fig. 5. It now becomes possible to count the
PIG. 5. Schematic diagram of the DBH positive nerve cell bodies of the locus coeruleus (striped area) and of the position of the squarcq ( 100 pm wide) used in the counting procedure.
MORPHOMETRIC
ANALYSIS
OF MONOAMINE
57
PROCESSES
XHEMATIC REPRESENTATIONOF MH POSITIVE CELL 30DV AND PNMT-POSITIVE TERMINAL DlSTRlBUTlON N THE LOCUS COERULEUS AREA (P= 2.6mm)
,,0
THE LORENZ CURVES AND THE INDEXES OF CONCENTRATION FOR THE DISTRISUTION OF DPH-POSITIVE CELLS AND PNMTPOSITIVE TERMINALS IN THE LOCUS COERUEUS AREA (P= 2.6mm)
Ifl H POSITIVE CELL BODIES PNMT POSITIVE TERMINAL! -I) ~
c
060
< >’ 0.50 II I? 0.40
*
LATERAL’
FIG. 7. Lorenz curves for the density distribution of PNMT positive nerve terminals and DBH positive cell bodies in the locus coeruleus. The respective indexes of concentration are also given.
FIG. 6. Density distribution of PNMT positive boutons and DBH positive cell bodies in the locus coeruleus. For further details, see text.
SCATTER
DBH immunoreactive
nerve cell bodies and the PNMT immunoreactive nerve terminals and to obtain a density distribution of these structures as shown in Fig. 6. It is evident
that some of the PNMT immunoreactive nerve terminals are present in a periventricular zone, where no DBH positive nerve cell bodies are found. These results indicate a possible innervation of noradrenaline dendrites by PNMT positive nerve terminals. The empty squares which are found in the density distribution of noradrenaline cell bodies will not be considered in the subsequent analysis. The Lorenz curves for both nerve terminal and cell body distribution of the locus coeruleus are reported in Fig. 7. In view of the fact that a concentration is observed a study of the correlation between nerve terminal and cell body distribution has been performed (Fig. 8). In this way it is possible to assess whether regions in which noradrenaline cell bodies are more densely packed also contain a higher number of PNMT immunoreactive nerve terminals. No correlation is found. The mean density of PNMT immunoreactive nerve terminals per noradrenaline cell body has been calculated to be 4.3. This figure represents a morphological evaluation of the possible degree of adrenergic control of the noradrenaline cell bodies of the locus coeruleus. In the morphometric analysis of transmitter-identified nerve terminals it is also possible to include the densitometric approach by Agnati et al. [2] to evaluate fluorescence intensity in the transmitter-identified nerve terminals of the histochemical preparation. There exists a linear relationship between catecholamine contents and fluorescence yield as
DUGRAM
OF
THE
rERMNALSPER SCW4REuVE~ WRVECE
NUMBER OF PtWT-POSlTlVE MRVE TgAMSER OF DWPOSlTlVE
s
,-
,rw-0.05
NS
I-
l
I-
0 a
c
a l
a
l
:
l
I.
t
1
2 MI-R
3 OF WI-WSITIVE
4
5 NERVE CELLS
6
f 7
PER-SQUARE
FIG. 8. Correlation between PNMT positive nerve terminals and DBH positive nerve cell body distribution in the locus coeruleus. For further details, see text.
AGNATI
E’l AL.
.. .* ‘,
. .
:
“6
1
J*
:
. .> .’
.
_ *-
w
.
.
*
*
..
,,-.
..
.
1
._
...a
;
:
.
.I( .. .
.
2. . -
. .
,
.. I
. .
.’
: I
,,)
.
. .
., ,
.
i
.
.
FIG. 9. A-C. Densitometric evaluation of PNMT immunofluorescence intensity (A) in a single microscopic field of the ventral part of locus coeruleus. In B all tones are shown and in C only the highest tones. Magnification 195x. For further details. see text.
FIG. IO. A-C. Densitometric evaluation of DBH immunofluorescence intensity (A) observed in a single microscopic field of the ventral part of locus coeruleus. In B all tones are shown and in C only the highest tones. Section adjacent to section shown in Fig. 9. Magnification 195x. For further details, see text.
far as the central catecholamine nerve terminal systems are concerned [I], while the relationship between antigen con-
centration and fluorescence yield in immunohistochemical preparations has not been as clearly defined. Therefore, in the present analysis of the PNMT immunoreactive nerve terminals only the separation of all tones versus the highest tones has been performed. The densitometric approach has
been carried out for both the DBH positive cell bodies and immunoreactive nerve terminals. The results of this procedure are shown in Fig. 9 and Fig. 10. The fluorescence microphotograph analyzed is shown to the left in Figs. 9 and 10, respectively. The middle photograph of Figs. 9 and 10 shows all the tones and the photographs to the right in Figs. 9 and 10 only show the highest tones of the for the PNMT
MORPHOMETRIC
ANALYSIS
OF MONOAMINE
PROCESSES
fluorescence. In this way it has been possible to evaluate the density distribution of the PNMT immunoreactive nerve terminals and DBH positive cell bodies having the highest fluorescence intensity and a correlation between these two types of structures has been performed. An increase in the correlation coefficient with respect to that calculated for all the tones will indicate that nerve terminals with a high PNMT content may be correlated to DBH positive nerve cells with a high DBH content. However, this is not the case since the Pearson’s correlation coefficient was found to be r=0.09. DISCUSSION
The present morphometric method using Lorenz curves and the index of concentration makes it possible to describe in a quantitative way the degree of unevenness of distribution of transmitter-identified processes and nerve terminals within a certain area of the brain. By using in addition the densitometric approach of Agnati et al. [2] it also becomes possible to study selectively the degree of unevenness of distribution of those processes and nerve terminals which have the highest fluorescence intensity. By this quantitative procedure it becomes possible to establish which area contains the highest density of processes. This knowledge is of importance since in this way it is possible to obtain an indication as to where the main input to the nucleus studied is located. Thus, axodendritic interaction is the most common interaction in the central nervous system [24, 26, 271.
Likewise, by knowing the detailed pattern of distribution of the nerve terminals it may be possible to establish the location of those nerve cells which are the main target for the nerve terminal networks analyzed. An important advantage of the present method is that it makes possible the quantitation of morphologic changes taking place in transmitter-identified dendrites and terminals in relation to various types of treatments and states such as the degenerative changes occurring in the aging brain [28,29]. It becomes possible to quantitatively define the morphological changes existing between control rats and rats with various types of genetic syndromes such as spontaneous hypertension, diabetes mellitus etc. Finally, this procedure makes it possible to define in quantitative terms the pre- and postnatal development of processes and nerve terminal networks of transmitter-identified neurons, which will be of importance for future progress in developmental neurobiology.
ACKNOWLEDGEMENTS This
work
has
been supported by grants (04X-715, 04X-2887) from the Swedish Medical Research Council, by a grant (MH25504) from the National Institute of Mental Health and by an international CNR grant as well as by a grant from Knut & Alice Wallenberg’s Foundation. L. F. Agnati has been a recipient of a visiting scientist fellowship from the Swedish Medical Research Council. Laura Calza is a recipient of a “Legato Dino Ferrari.”
REFERENCES Agnati, L. F., K. Andersson, F. Wiesel and K. Fuxe. A method to determine dopamine levels and turnover rate in discrete dopamine nerve terminal systems by quantitative use of dopamine fluorescence obtained by Falck-Hillarp methodology. J. Nrurosci. Meth. 1: 365-373, 1979. Agnati, L. F., F. Benfenati, P. Cortelli and R. D’Alessandro. A new method to quantify catecholamine stores visualized by means of the Falck-Hillarp technique. Neurosci. Lrtf. 10: 1l-17, 1978. Agnati, L. F., K. Fuxe, K. Andersson, F. Benfenati, P. Cortelli and R. D’AIessandro. The mesolimbic dopamine system: evidence for a high amine turnover and for a heterogeneity of the dopamine neuron population. Neurosci. Left. 18:45-51, 1980. Aenati. L. F.. K. Fuxe. I. Scardovi. P. Monari. L. Calza. F. Benfenati, T. Hokfelt and J. de Mey.‘Criteria for’the identification of transmitter-identified cell groups and their morphometric characterization. J. Neurosci. Me&., in press. Chan-Palay, V. Indoleamine neurons and their processes in the normal rat brain and in chronic diet-induced thiamine deficiency demonstrated by uptake of 3H-serotonin. J. camp. Neural. 176: 467-494, 1977. 6. Coons, A. H. Fluorescent antibody methods. In: General Cyrochemical Methods, edited by J. F. Danielli. New York: Academic Press, 1958, pp. 399-422. 7. Duncan, 0. D. The measurement of population distribution. In: Muthematical Demography, edited by D. Smith and N. Keyfitz. Berlin: Springer Verlag, 1977, pp. 349-363. 8. Fuxe, K. Evidence for the existence of monoamine neurons in the central nervous system. IV. The distribution of monoamine nerve terminals in the central nervous system. Acta physiol. wand. 64: Suppl. 247, 39-85. 1965.
a.
9. Fuxe, K., L. F. Agnati, K. Andersson, V. Locatelli, P. Eneroth, T. Hiikfelt, V. Mutt, T. McDonald, M. F. El Etreby, I. Zini and L. Calza. Concepts in neuroendocrinology with emphasis on neuropeptide-monoamine interactions in neuroendocrine regulation. In: Progress in Psychoneuroendocrinology, edited by F. Brambilla, G. Racagni and D. De Wied. Amsterdam: Elsevier/North-Holland Biomedical Press, 1980, pp. 47-6 1. 10. Fuxe, K., T. Hokfelt, L. Olson and U. Ungerstedt. Central monoaminergic pathways with emphasis on their relation to the so called ‘extrapyramidal motor system.’ Pharmac. Ther., B 3: 169210, 1977. 11. Goldstein, M. Enzymes involved in the catalysis of catecholamine biosynthesis. In: Methods in Neurochemistry, edited by R. N. Ubell. New York: Plenum Press. 1972. DD. 317-340. 12. Goldstein, M., B. Anagnoste, L. S. Freed&r, M. Roffman, R. P. Ebstein, D. H. Park, K. Fuxe and T. Hokfelt. Characterization, localization and regulation of catecholamine synthesizing enzymes. In: Frontiers in Catecholamine Research, edited by E. Usdin and S. Snyder. New York: Pergamon Press, 1973, pp. 69-78. 13. Goldstein, M., K. Fuxe and T. Hokfelt. Characterization and tissue localization of catecholamine synthesizing enzymes. Pharmac. Rev. 24: 293-309, 1972. 14. Grzanna, R., M. E. Molliver and J. T. Coyle. Visualization
of central noradrenergic neurons in thick sections by the unlabeled antibody method: A transmitter-specific Golgi image. Proc.
natn. Acad.
Sri. U.S.A.
75: 2502-2506,
1978.
15. Hartman, B. K., D. Zide and S. Udenfriend. The use of dopamine+-hydroxylase as a marker for the noradrenergic pathways of the central nervous system in the rat. Proc. natn. Acud. Sci. U.S.A. 69: 2722-2726, 1972. 16. Hokfelt, T., K. Fuxe and M. Goldstein.
Applications of immunohistochemistry to studies on monoamine cell systems with special reference to nervous tissues. Ann. N. Y. Acad. Sci. 254:
407-432,
1975.
60
17. H&felt, T., K. Fuxe, M. Goldstein and T. H. Joh. lmmunohistochemical localization of three catecholamine synthesizing enzymes: aspects on methodology. Histochemic 33: 231-254, 1973. 18. Hokfelt, T., K. Fuxe, M. Goldstein and 0. Johansson. lmmunohistochemical evidence for the existence of adrenaline neurons in the rat brain. Bruin Res. 66: I-17, 1974. 19. Hokfelt, T., M. Goldstein, K. Fuxe, 0. Johansson, A. Verhofstad, H. Steinbusch, B. Penke and J. Vargas. Histochemical identification of adrenaline containing cells with special reference to neurons. In: Central Adrenaline Neurons, edited by K. Fuxe, M. Goldstein, B. Hokfelt and T. H&felt. Oxford: Pergamon Press, 1980, pp. 19-47. 20. Hokfelt, T., 0. Johansson, K. Fuxe, M. Goldstein and D. Park. lmmunohistochemical studies on the localization and distribution of monoamine neuron systems in the rat brain. 1. Tyrosine hydroxylase in the mes- and diencephalon. Med. Rio/. 54: 427453, 1976. 21. Lindvall, 0. and A. Bjorklund. Organization of catecholamine neurons in the rat central nervous system. In: Hundbook of Psyc,hopharmaco/oRy, edited by L. L. lversen and S. H. Snyder. New York: Plenum Press, 1978. pp. 13&231.
AGNATl
El’ AL.
22. Mills, F. C. Statistkal Methods. New York: Henry Holt and Company. 1955. 23. Pease, D. C. Buffered formaldehyde as a killing agent and primary fixative for electron microscopy. Anut. Rrc,. 142: 342. 1962. 24. Ramon-Moliner, C. An attempt at classifying nerve cells on the basis of their dendritic patterns. J. camp. Nvrrrol. 119: 2 I l-227. 1962. 25. Scardovi, I. Appunti di Stcrtisrictr-I. Bologna: Patron, 1980. 26. Shall, D. A. Dendritic organization in the neurons of the visual and motor cortices of the cat. .I. Anat. 87: 387-406. 1953. 27. Szentagothai, J. Models of specific neuron arrays in thalamic relay nuclei. Acfa morph. hung. 15: 113-124, 1957. 28. Tennyson. V. M., R. E. Barrett, G. Choen, L. Cote. R. Heikkila and C. Mytilineou. Correlation of anatomical and biochemical development of the rabbit neostriatum. Prop. Rrtrin Rev. 40: 203-217, 1973. 29. Wisniewski, H. M. and R. D. Terry. Morphology of the aging brain. human and animal. Proc. Rrairt Rcs. 40: 167-186. 1973.
Bulletin, Vol. 9, pp. 53-60,1982.Printed in the U.S.A.
A Morphometric Analysis of Transmitter ,Identified Dendrites and Nerve Terminals L. F. AGNATI,
K. FUXE,’ L. CALZA, T. HijKFELT, 0. JOHANSSON, F. BENFENATI AND M. GOLDSTEIN
DPpartment of Human Physiology, University of Modena, 41100 Modena, Italy Department of Histology, Karolinska Institutet, Box 60400, S-104 01 Stockholm, Sweden and Department of Psychiatql, New York University Medical Center, New York, NY
AGNATI, L. F., K. FUXE, L. CALZA, T. HGKFELT,
F. BENFENATI AND M. GOLDSTEIN. A BRAIN RES. BULL. 9(1-6) 5?-60, 1982.-The present method is exemplified on coronal sections of the medulla oblongata containing phenylethanolamineN-methyltransferase (PNMT) immunoreactive nerve cell bodies and their processes and on coronal sections of the pons containing the locus coeruleus, where PNMT immunoreactive nerve terminals have been demonstrated together with the dopamine+-hydroxylase immunoreactive nerve cell bodies. Morphometric analysis of the processes (both length and branches) and of the nerve terminals involve as a first step the division of the area under study into squares 100 pm wide, which are superimposed on a Cartesian plane. The uniformity of the density distribution of the nerve terminals and the processes (branches or length) can be analyzed by Lorenz curves, which in a quantitative way can measure the degree of unevenness and thus represent a measure of concentration. A concentration index can therefore be calculated. By the use of the densitometric approach it also becomes possible to study the density distribution of the nerve terminals with the highest antigen contents. The present method will make it possible to quantitate morphological changes occurring in processes and nerve terminals of transmitter-identified neurons.
morphometricanalysis
Morphometric
analysis
of transmitter
Transmitter
identified
0. JOHANSSON,
dendrites
identified neurons
IN a large number of studies the distribution and morphology of catecholamine (CA) nerve terminals and dendrites have been described by subjective evaluation in the microscope or
by inspection of microphotographs, using Falck-Hillarp methodology for the cellular demonstration of CA stores or using immunohistochemical methods to evaluate the cellular localization of tyrosine-hydroxylase, dopamine-phydroxylase (DBH) and phenylethanolamine-N-methyltransferase (PNMT) [8, 10, 16, 19, 20, 211. In recent papers [4,9] methods for the morphometric characterization of transmitter-identified nerve cell groups have been introduced together with objective criteria for defining nerve cell groups in subpopulations of nerve cells within one single nerve cell group. In the present paper we introduce a morphometric analysis of transmitter-identified neurons at the dendritic and nerve terminal level. The present method is illustrated on one coronal section of the rostra1 part of the medulla oblongata showing PNMT positive cell bodies and their processes [ 193, and on adjacent coronal sections of the pons in which DBH immunoreactive cell bodies or PNMT ‘Please send reprint requests to Professor Kjell Fuxe, Department Sweden.
Copyright 0 1982 ANKHO
and
nerve
terminals.
Medulla oblongata
immunoreactive locus coeruleus
Locus coeruleus
nerve [ 191.
terminals
are demonstrated
in the
METHOD Male specific pathogen-free Sprague-Dawley rats (b.wt. 150 g) were used. For the analysis of medulla oblongata the rats were pretreated with colchicine (75 Fg) intraventricularly 24 hr before killing. All rats were perfused with 50 ml Tyrode solution followed by 500 ml icecold 4% (W/V) paraformaldehyde in 0.1 M phosphate buffer [23] for 30 minutes. After perfusion the medulla oblongata and the pons were dissected out and immersed in the same fixative for 90 minutes. The brain tissues were then rinsed in 5% (W/V) sucrose in 0.1 M phosphate buffer for at least 24 hr. The medulla oblongata was cut on a freeze microtome. The immunoperoxidase procedure was adopted from Grzanna et al. [14]. Briefly, 50 pm freeze microtome sections of colchicine-treated, Formalin-fixed brains, were incubated with the PNMT antibodies (diluted 1: 160), followed of Histology, Karolinska Institutet,
Box 60400, S-104 01 Stockholm,
International Inc.-0361-9230/82/070053-08$03.00/O
AGNA’I‘I
FIG. 1, Coronal section of the dorsomediai part of the rostrai medulla oblongata where PNMT positive nerve cell bodies and processes of adrenaline cell group C2 [I83 have been visualized by means of PAP technique (see text). The vertical arrow point in the dorsal direction and the horizontal arrow in the medial direction. x300.
t7’ Al
MORPHOMETRIC
ANALYSIS
OF MONOAMINE
PROCESSES taining neurons. The purgation the production of the antiserum ously [ll, 12, 131.
of this enzyme as well as has been described previ-
RESULTS
Morphometric Analysis of PNMT Immunoreactive Nerve Cells of the Adrenaline Cell Group C2 at the Dendritic Level
FIG. 2. Density distribution of length of processes and number of branches of PNMT positive cells in the adrenaline cell group C2. Part of the original microphotograph is shown in Fig. 1. The whole montage is not shown, since the processes are not clearly shown, when the montage is reduced to publication format. For further details, see text.
by sheep antiserum to rabbit immunoglobulin (diluted 1:50), and horseradish peroxidase-antiperoxidase (PAP) complex (diluted 1:40; the PAP complex was a generous gift from Prof. L. A. Sternberger, Department of Anatomy and Center for Brain Research, University of Rochester, Rochester, NY). Finally, they were treated with diaminobenzidine and hydrogen peroxide, followed by OsO,, dehydrated in ethanol and xylene, and coverslipped in Entellan. The sections were examined in a Leitz Orthoplan light microscope using bright-field optics. Kodak Panatomic-X film was used for photography. For further details on the methodology used, the reader is referred to Johansson and Backman (in preparation). The pons was cut in a cryostat (Dittes, Heidelberg, West-Ge~any) to give sections 10 pm thick; these were processed according to the indirect immunofluorescence technique of Coons [6], essentially as described by Hokfelt et al. [ 171. Briefly, repeated sequences of two adjacent sections containing the locus coeruleus were used. One of them was incubated with DBH antiserum (dilution 1: 100) and the other with PNMT antiserum (dilution 1: 100) ,at +4”C for 24 hr, rinsed in phosphate buffered saline (PBS), tpcubated with fluoresceine-isothiocyanate (FITC) conjugated sheep antirabbit antibodies (Statens Bakteriologiska Laboratorium, SBL, Stockholm, Sweden) for 30 minutes at +37”C, rinsed in PBS mounted in a mixture of glycerol and PBS (3:l) and examined in a Zeiss Standard fluorescence microscope equipped with an oil dark field condenser and a Schott BG 12 (3 or 4 mm) and a Zeiss 56 stop filter. All sera containing 0.3% T&on-X-100 [15] to solubilize membranes and in this way facilitate penetration. Scopix RP 1 black and white film (Gaevert, Belgium) was used for photography. The antiserum against DBH was used as a marker for the noradrenaline (NA) cell bodies of the locus coeruleus. The purification of the enzyme as well as the production of the antiserum has been described in detail [ll, 12, 131. The antiserum produced against the adrenaline synthesizing enzyme PNMT was used as a marker for the adrenaline con-
The morphometric procedure is examplified on a coronal section of the rostra1 medulla oblongata which in its dorsomedial part contain the PNMT positive nerve cell bodies of the adrenaline cell group C2 (see Fig. 1). The morphometric procedure starts with a Cartesian representation of the localization of the various branches which has been obtained by taking the midline as the x-axis and a straight line tangential to the ventral border of the fourth ventricle and perpendicular to the midline as the y-axis. Thereafter, squares with a side of 100 pm are superimposed on this Cartesian plane and an evaluation of the number of branches and of the length of the branches have been carried out for each square (Fig. 2). As seen in Fig. 2, the PNMT positive processes and thus the number of branches are especially rich within the central part of the adrenaline cell group. This is a feature of a nerve cell group of transmitter-identi~ed neurons which can be of interest for, e.g., comparisons between different nerve cell groups or for comparisons within one and the same cell group at different rostra&caudal levels. Therefore, it is important to develop a method which can quantitate the density and uniformity of distribution of the length of processes and number of branches in order to have an indication of the pattern of innervation within a certain area. This is possible by studying the uniformity of the density distribution of a certain type of parameter using standard statistical techniques such as the Lorenz curves to determine the concentration [22]. The concept of concentration as used in the present study has been extensively employed in studies of population density [?I. The measure of concentration gives the degree of unevenness in quantitative terms. The Lorenz curve is a diagram of the con~ent~tion, that is of the unevenness of the dist~bution of a certain parameter such as number of branches or length of processes among N squares, in which the brain area analyzed has been subdivided. The principle according to which a Lorenz curve is constructed is illustrated in Fig. 3 [25]. Thus, let us consider the n quantities x,, xB . . . xk . . . x,, where xL
AGNAT
Sh
ITHELORENZCURVE AND THE INDEXOF CONCENTRATION 1
k,‘TAL.
rm uwmyz CUMS MD M IWXES OF CCNCENTRATIOU FOR TH )ISTRtBUTW Cf PROCESS LENGTHAM BRANCHlWOFo*BER IN TIE C2 AREA @I&IT-m PROFILES ) BRANCHING NUMBER DISTIWTION~ R = 0 5, PROCESS LENGTH USTRlEUTlON R =
FIG. 3. Schematic representation of the Lorenz curve and the index of concentration (=R). On the x-axis the quantities P,=k/n are reported, where k= I,? , n and in the present case are the squares of Fig. 2 ranked according to, e.g., the number of branches. On the y-axis the values qk=VkN,, are given, where V,=x, + x2 + + xk and xk represents the value of the parameter under study (e.g., number of branches) for the Kth square. It should be noted that xII
tween the observed index of concentration
k =
and the maximal concentration. is represented by the ratio
( ;; ) (Pk-q&
( z:
Pk
0
0
O.lO
0.20
0.30
0.40
0 50
0.60
0.70
080
000
Pk E yn
FIG. 4. Lorenz curves and the concentration indexes for the distribution of the number of branches and of the length of processes of the PNMT immunoreactive neurons of adrenaline cell group C? shown in Figs. I and 2.
The
)
This ratio is in the case of an even distribution zero and in the case of a maximal concentration one. The Lorenz curves and the concentration indexes for the distribution of the number of branches and of the length of processes shown in the various squares of Fig. 2 are shown in Fig. 4. Morphonwtric Tmn inm1.s
Ancrlysis
of‘ PNMT lmmunorcac.ti~c
Nrr\lc,
The analysis will be exemplified on PNMT positive nerve terminals within the locus coeruleus. The noradrenaline nerve cell bodies of the locus coeruleus have been demonstrated by means of a DBH antiserum in a section adjacent to the section showing the PNMT immunoreactive nerve terminals within the locus coeruleus. Also in this case the first step in the morphometric procedure is to obtain a Cartesian representation of the localization of the PNMT nerve terminals and of the DBH immunoreactive nerve cell bodies. which probably are innervated by the PNMT immunoreactive nerve terminals. A schematic representation of this procedure is seen in Fig. 5. It now becomes possible to count the
PIG. 5. Schematic diagram of the DBH positive nerve cell bodies of the locus coeruleus (striped area) and of the position of the squarcq ( 100 pm wide) used in the counting procedure.
MORPHOMETRIC
ANALYSIS
OF MONOAMINE
57
PROCESSES
XHEMATIC REPRESENTATIONOF MH POSITIVE CELL 30DV AND PNMT-POSITIVE TERMINAL DlSTRlBUTlON N THE LOCUS COERULEUS AREA (P= 2.6mm)
,,0
THE LORENZ CURVES AND THE INDEXES OF CONCENTRATION FOR THE DISTRISUTION OF DPH-POSITIVE CELLS AND PNMTPOSITIVE TERMINALS IN THE LOCUS COERUEUS AREA (P= 2.6mm)
Ifl H POSITIVE CELL BODIES PNMT POSITIVE TERMINAL! -I) ~
c
060
< >’ 0.50 II I? 0.40
*
LATERAL’
FIG. 7. Lorenz curves for the density distribution of PNMT positive nerve terminals and DBH positive cell bodies in the locus coeruleus. The respective indexes of concentration are also given.
FIG. 6. Density distribution of PNMT positive boutons and DBH positive cell bodies in the locus coeruleus. For further details, see text.
SCATTER
DBH immunoreactive
nerve cell bodies and the PNMT immunoreactive nerve terminals and to obtain a density distribution of these structures as shown in Fig. 6. It is evident
that some of the PNMT immunoreactive nerve terminals are present in a periventricular zone, where no DBH positive nerve cell bodies are found. These results indicate a possible innervation of noradrenaline dendrites by PNMT positive nerve terminals. The empty squares which are found in the density distribution of noradrenaline cell bodies will not be considered in the subsequent analysis. The Lorenz curves for both nerve terminal and cell body distribution of the locus coeruleus are reported in Fig. 7. In view of the fact that a concentration is observed a study of the correlation between nerve terminal and cell body distribution has been performed (Fig. 8). In this way it is possible to assess whether regions in which noradrenaline cell bodies are more densely packed also contain a higher number of PNMT immunoreactive nerve terminals. No correlation is found. The mean density of PNMT immunoreactive nerve terminals per noradrenaline cell body has been calculated to be 4.3. This figure represents a morphological evaluation of the possible degree of adrenergic control of the noradrenaline cell bodies of the locus coeruleus. In the morphometric analysis of transmitter-identified nerve terminals it is also possible to include the densitometric approach by Agnati et al. [2] to evaluate fluorescence intensity in the transmitter-identified nerve terminals of the histochemical preparation. There exists a linear relationship between catecholamine contents and fluorescence yield as
DUGRAM
OF
THE
rERMNALSPER SCW4REuVE~ WRVECE
NUMBER OF PtWT-POSlTlVE MRVE TgAMSER OF DWPOSlTlVE
s
,-
,rw-0.05
NS
I-
l
I-
0 a
c
a l
a
l
:
l
I.
t
1
2 MI-R
3 OF WI-WSITIVE
4
5 NERVE CELLS
6
f 7
PER-SQUARE
FIG. 8. Correlation between PNMT positive nerve terminals and DBH positive nerve cell body distribution in the locus coeruleus. For further details, see text.
AGNATI
E’l AL.
.. .* ‘,
. .
:
“6
1
J*
:
. .> .’
.
_ *-
w
.
.
*
*
..
,,-.
..
.
1
._
...a
;
:
.
.I( .. .
.
2. . -
. .
,
.. I
. .
.’
: I
,,)
.
. .
., ,
.
i
.
.
FIG. 9. A-C. Densitometric evaluation of PNMT immunofluorescence intensity (A) in a single microscopic field of the ventral part of locus coeruleus. In B all tones are shown and in C only the highest tones. Magnification 195x. For further details. see text.
FIG. IO. A-C. Densitometric evaluation of DBH immunofluorescence intensity (A) observed in a single microscopic field of the ventral part of locus coeruleus. In B all tones are shown and in C only the highest tones. Section adjacent to section shown in Fig. 9. Magnification 195x. For further details, see text.
far as the central catecholamine nerve terminal systems are concerned [I], while the relationship between antigen con-
centration and fluorescence yield in immunohistochemical preparations has not been as clearly defined. Therefore, in the present analysis of the PNMT immunoreactive nerve terminals only the separation of all tones versus the highest tones has been performed. The densitometric approach has
been carried out for both the DBH positive cell bodies and immunoreactive nerve terminals. The results of this procedure are shown in Fig. 9 and Fig. 10. The fluorescence microphotograph analyzed is shown to the left in Figs. 9 and 10, respectively. The middle photograph of Figs. 9 and 10 shows all the tones and the photographs to the right in Figs. 9 and 10 only show the highest tones of the for the PNMT
MORPHOMETRIC
ANALYSIS
OF MONOAMINE
PROCESSES
fluorescence. In this way it has been possible to evaluate the density distribution of the PNMT immunoreactive nerve terminals and DBH positive cell bodies having the highest fluorescence intensity and a correlation between these two types of structures has been performed. An increase in the correlation coefficient with respect to that calculated for all the tones will indicate that nerve terminals with a high PNMT content may be correlated to DBH positive nerve cells with a high DBH content. However, this is not the case since the Pearson’s correlation coefficient was found to be r=0.09. DISCUSSION
The present morphometric method using Lorenz curves and the index of concentration makes it possible to describe in a quantitative way the degree of unevenness of distribution of transmitter-identified processes and nerve terminals within a certain area of the brain. By using in addition the densitometric approach of Agnati et al. [2] it also becomes possible to study selectively the degree of unevenness of distribution of those processes and nerve terminals which have the highest fluorescence intensity. By this quantitative procedure it becomes possible to establish which area contains the highest density of processes. This knowledge is of importance since in this way it is possible to obtain an indication as to where the main input to the nucleus studied is located. Thus, axodendritic interaction is the most common interaction in the central nervous system [24, 26, 271.
Likewise, by knowing the detailed pattern of distribution of the nerve terminals it may be possible to establish the location of those nerve cells which are the main target for the nerve terminal networks analyzed. An important advantage of the present method is that it makes possible the quantitation of morphologic changes taking place in transmitter-identified dendrites and terminals in relation to various types of treatments and states such as the degenerative changes occurring in the aging brain [28,29]. It becomes possible to quantitatively define the morphological changes existing between control rats and rats with various types of genetic syndromes such as spontaneous hypertension, diabetes mellitus etc. Finally, this procedure makes it possible to define in quantitative terms the pre- and postnatal development of processes and nerve terminal networks of transmitter-identified neurons, which will be of importance for future progress in developmental neurobiology.
ACKNOWLEDGEMENTS This
work
has
been supported by grants (04X-715, 04X-2887) from the Swedish Medical Research Council, by a grant (MH25504) from the National Institute of Mental Health and by an international CNR grant as well as by a grant from Knut & Alice Wallenberg’s Foundation. L. F. Agnati has been a recipient of a visiting scientist fellowship from the Swedish Medical Research Council. Laura Calza is a recipient of a “Legato Dino Ferrari.”
REFERENCES Agnati, L. F., K. Andersson, F. Wiesel and K. Fuxe. A method to determine dopamine levels and turnover rate in discrete dopamine nerve terminal systems by quantitative use of dopamine fluorescence obtained by Falck-Hillarp methodology. J. Nrurosci. Meth. 1: 365-373, 1979. Agnati, L. F., F. Benfenati, P. Cortelli and R. D’Alessandro. A new method to quantify catecholamine stores visualized by means of the Falck-Hillarp technique. Neurosci. Lrtf. 10: 1l-17, 1978. Agnati, L. F., K. Fuxe, K. Andersson, F. Benfenati, P. Cortelli and R. D’AIessandro. The mesolimbic dopamine system: evidence for a high amine turnover and for a heterogeneity of the dopamine neuron population. Neurosci. Left. 18:45-51, 1980. Aenati. L. F.. K. Fuxe. I. Scardovi. P. Monari. L. Calza. F. Benfenati, T. Hokfelt and J. de Mey.‘Criteria for’the identification of transmitter-identified cell groups and their morphometric characterization. J. Neurosci. Me&., in press. Chan-Palay, V. Indoleamine neurons and their processes in the normal rat brain and in chronic diet-induced thiamine deficiency demonstrated by uptake of 3H-serotonin. J. camp. Neural. 176: 467-494, 1977. 6. Coons, A. H. Fluorescent antibody methods. In: General Cyrochemical Methods, edited by J. F. Danielli. New York: Academic Press, 1958, pp. 399-422. 7. Duncan, 0. D. The measurement of population distribution. In: Muthematical Demography, edited by D. Smith and N. Keyfitz. Berlin: Springer Verlag, 1977, pp. 349-363. 8. Fuxe, K. Evidence for the existence of monoamine neurons in the central nervous system. IV. The distribution of monoamine nerve terminals in the central nervous system. Acta physiol. wand. 64: Suppl. 247, 39-85. 1965.
a.
9. Fuxe, K., L. F. Agnati, K. Andersson, V. Locatelli, P. Eneroth, T. Hiikfelt, V. Mutt, T. McDonald, M. F. El Etreby, I. Zini and L. Calza. Concepts in neuroendocrinology with emphasis on neuropeptide-monoamine interactions in neuroendocrine regulation. In: Progress in Psychoneuroendocrinology, edited by F. Brambilla, G. Racagni and D. De Wied. Amsterdam: Elsevier/North-Holland Biomedical Press, 1980, pp. 47-6 1. 10. Fuxe, K., T. Hokfelt, L. Olson and U. Ungerstedt. Central monoaminergic pathways with emphasis on their relation to the so called ‘extrapyramidal motor system.’ Pharmac. Ther., B 3: 169210, 1977. 11. Goldstein, M. Enzymes involved in the catalysis of catecholamine biosynthesis. In: Methods in Neurochemistry, edited by R. N. Ubell. New York: Plenum Press. 1972. DD. 317-340. 12. Goldstein, M., B. Anagnoste, L. S. Freed&r, M. Roffman, R. P. Ebstein, D. H. Park, K. Fuxe and T. Hokfelt. Characterization, localization and regulation of catecholamine synthesizing enzymes. In: Frontiers in Catecholamine Research, edited by E. Usdin and S. Snyder. New York: Pergamon Press, 1973, pp. 69-78. 13. Goldstein, M., K. Fuxe and T. Hokfelt. Characterization and tissue localization of catecholamine synthesizing enzymes. Pharmac. Rev. 24: 293-309, 1972. 14. Grzanna, R., M. E. Molliver and J. T. Coyle. Visualization
of central noradrenergic neurons in thick sections by the unlabeled antibody method: A transmitter-specific Golgi image. Proc.
natn. Acad.
Sri. U.S.A.
75: 2502-2506,
1978.
15. Hartman, B. K., D. Zide and S. Udenfriend. The use of dopamine+-hydroxylase as a marker for the noradrenergic pathways of the central nervous system in the rat. Proc. natn. Acud. Sci. U.S.A. 69: 2722-2726, 1972. 16. Hokfelt, T., K. Fuxe and M. Goldstein.
Applications of immunohistochemistry to studies on monoamine cell systems with special reference to nervous tissues. Ann. N. Y. Acad. Sci. 254:
407-432,
1975.
60
17. H&felt, T., K. Fuxe, M. Goldstein and T. H. Joh. lmmunohistochemical localization of three catecholamine synthesizing enzymes: aspects on methodology. Histochemic 33: 231-254, 1973. 18. Hokfelt, T., K. Fuxe, M. Goldstein and 0. Johansson. lmmunohistochemical evidence for the existence of adrenaline neurons in the rat brain. Bruin Res. 66: I-17, 1974. 19. Hokfelt, T., M. Goldstein, K. Fuxe, 0. Johansson, A. Verhofstad, H. Steinbusch, B. Penke and J. Vargas. Histochemical identification of adrenaline containing cells with special reference to neurons. In: Central Adrenaline Neurons, edited by K. Fuxe, M. Goldstein, B. Hokfelt and T. H&felt. Oxford: Pergamon Press, 1980, pp. 19-47. 20. Hokfelt, T., 0. Johansson, K. Fuxe, M. Goldstein and D. Park. lmmunohistochemical studies on the localization and distribution of monoamine neuron systems in the rat brain. 1. Tyrosine hydroxylase in the mes- and diencephalon. Med. Rio/. 54: 427453, 1976. 21. Lindvall, 0. and A. Bjorklund. Organization of catecholamine neurons in the rat central nervous system. In: Hundbook of Psyc,hopharmaco/oRy, edited by L. L. lversen and S. H. Snyder. New York: Plenum Press, 1978. pp. 13&231.
AGNATl
El’ AL.
22. Mills, F. C. Statistkal Methods. New York: Henry Holt and Company. 1955. 23. Pease, D. C. Buffered formaldehyde as a killing agent and primary fixative for electron microscopy. Anut. Rrc,. 142: 342. 1962. 24. Ramon-Moliner, C. An attempt at classifying nerve cells on the basis of their dendritic patterns. J. camp. Nvrrrol. 119: 2 I l-227. 1962. 25. Scardovi, I. Appunti di Stcrtisrictr-I. Bologna: Patron, 1980. 26. Shall, D. A. Dendritic organization in the neurons of the visual and motor cortices of the cat. .I. Anat. 87: 387-406. 1953. 27. Szentagothai, J. Models of specific neuron arrays in thalamic relay nuclei. Acfa morph. hung. 15: 113-124, 1957. 28. Tennyson. V. M., R. E. Barrett, G. Choen, L. Cote. R. Heikkila and C. Mytilineou. Correlation of anatomical and biochemical development of the rabbit neostriatum. Prop. Rrtrin Rev. 40: 203-217, 1973. 29. Wisniewski, H. M. and R. D. Terry. Morphology of the aging brain. human and animal. Proc. Rrairt Rcs. 40: 167-186. 1973.