P U L M O N A R Y ARGYROPHZL C E L L S A T H I G H A L T I T U D E W. TAYLOR Department of Pathology, University of Liverpool
PLATESLXIII-LXV ARGYROPHIL cells in human lung were first described by Feyrter as recently as 1954. The probable reason for their late discovery is that they are not easily distinguishable in sections stained by haematoxylin and eosin (fig. 1). Special methods such as the silver nitrate argyrophil technique of Grimelius (1968) are needed for their demonstration. Previous methods for argyrophilia have not produced such consistent results. Sections stained by this method show that human lung may contain large numbers of these cells (fig. 2). The ultrastructure of pulmonary argyrophil cells is characterised by small bodies each about 120 nm in diameter (fig. 3). These bodies are termed " dense-core secretory vesicles " and consist of a dense osmiophilic core separated from an outer limiting membrane by a narrow clear halo. In the rabbit the vesicles have been shown to contain 5-hydroxytryptamine (Lauweryns et al., 1974). Where the argyrophil cells occur in groups in the rabbit, nerve endings have been demonstrated ramifying amongst them (Lauweryns and Cokelaere, 1973). These cells are situated in bronchi and bronchioles, and in alveolar walls, and their free surfaces are in contact with the inspired gases. In these sites, and with a nerve supply, they could conceivably act as chemoreceptors for respiratory gases, perhaps playing a part in the pulmonary vasoconstrictor response to hypoxia (Laros, 1971). The purpose of the present study was to determine whether the lungs of rabbits living under hypoxic conditions at high altitude contain any more argyrophil cells than members of the same species living at sea level. MATERIALS AND METHODS A study was made of tissue from six rabbits, one male and five females, which had been born and spent their entire lives at Cerro de Pasco in the Peruvian Andes at a n altitude of 4300 m above sea level. The study of native rabbits of the Peruvian Andes precludes the use of animals of the same strain as controls. It is not possible to match the animals for weight since the high altitude rabbits were much smaller than any laboratory strain of rabbit of the same age. It was thought more important to match the animals for age since it has been suggested that, at least in the rat, pulmonary argyrophil cells decline in number after the neonatal period (Moosavi, Smith and Heath, 1973). Laboratory strains of rabbit could have been used as test animals in a suitable hypobaric chamber, thus removing the disadvantage of having controls of a different strain. However, animals in hypobaric chambers are low altitude animals exposed to acute hypoxia which is interrupted by daily episodes of repressurisation for feeding and cleaning. The use of a hypobaric chamber would simulate the rapid transportation of a sea-level rabbit to high altitude. When a sea-level-dwelling man Received 2 Sept. 1976; accepted 11 Nov. 1976. J. PATH.-VOL.
122 (1977)
137
138
W. TAYLOR
attempts this acute change in altitude he undergoes initially accommodation and then after some months or years, acquired acclimatisation. He probably never manages to live at high altitude as easily as the natives who are born there, and whose condition may be described as natural acclimatisation (Heath and Williams, 1977). The native Peruvian rabbits employed were indigenous high-altitude rabbits which had been exposed to chronic hypoxia from birth and which should exhibit natural acclimatisation. The rabbits obtained were being bred locally for food. Their breed was indeterminate and they were designated “ Peruvian Test Rabbits ” (Pl-6). At the High Altitude Research Station at Cerro these animals were killed by intraperitoneal injection of pentobarbitone sodium and weighed. The thoracic organs were rapidly removed and each lobe of the lung cut off and placed in Bouin’s solution. After initial fixation of 10 minutes’ duration to assist cutting of the tissues, blocks 3-4 mm thick were cut approximately at right angles to the major airways, commencing at the hilum and progressing towards the periphery of the lung. Each block was left attached to its neighbour by a small area of lung tissue at the pleural surface to aid in orientation of the blocks. The rabbit has an appendage to the lower lobe of the right lung, the cardiac lobe, and this was treated in the same way as the other lobes. After slicing, each lobe was fixed in Bouin’s solution for 48 hr and then transferred back to the United Kingdom. On arrival 4 or 5 wk later the blocks were separated from their last attachment to each other, numbered consecutively from the hilum outwards for each lobe, and processed into paraffin wax. The number of blocks obtained from P1-6 was 25, 25, 33, 35, 35 and 30 respectively. Six New Zealand White rabbits of corresponding ages were studied as controls. Their tissues were treated in the same manner except that the storage time in 70 per cent. alcohol was only 2 days. They were designated “ Liverpool Control Rabbits ” (Ll-6). The number of blocks obtained from L1-6 was 22, 33, 30, 33, 34 and 34 respectively. Two adjacent sections cut from each block of tissue were stained with the Grimelius argyrophil technique and with haematoxylin and eosin. Argyrophil cells occurred singly and in groups in bronchi and bronchioles and in alveolar walls. Those in bronchi and bronchioles were designated “ airway ”, and those in alveolar walls ‘‘ parenchymal ” argyrophil cells and groups of cells. The numbers of individual and groups of argyrophil cells were counted by drawing a grid over the coverslip of each slide and systematically scanning each square. The area of each irregularly shaped section of lung was found using a point-counting technique and a WILD M501 Automatic Sampling Microscope. This device automatically moves its stage one graticule width across the section in the desired direction enabling the operator to point count the area of the section. For the point counting a 42-point Weibel graticule was employed on a projection head. Over lo00 points were counted for most sections and in no case was the number counted less than 380. Despite the number of sections counted the total areas of P1, P2 and P6 were small. This was because these three animals were younger and lighter (table I) and therefore smaller than P3, P4 and P5. Since they were not fixed in distension all the lungs were partly collapsed. Fewer blocks could conveniently be cut from the smaller collapsed lungs of P1, P2 and P6, accounting for the smaller total areas examined in those animals. In each rabbit the number of argyrophil cells and groups of cells present was expressed per cmz of tissue.
RESULTS Individual argyrophil cells in the airways were often spindle-shaped or conical (fig. 4). When spindle-shaped, the nucleus of the cell, which was distinguishable only as an oval agranular area, tended to be nearer the airway basement membrane than the luminal surface. Conical individual cells had their bases on the airway basement membrane. In both types a thin tapering process extended towards the free surface but did not always reach it within the plane of section. Groups of argyrophil cells in the airways had a corpuscular
PULMONARY ARG YROPHIL CELLS
139
structure (fig. 5). Often ten or more cells could be counted in a group in the plane of section. In the airways argyrophil cells or groups of cells were often present close to bifurcations and sometimes in the mucosa overlying an aggregate of lymphoid tissue. TABLEI Sex, weight and age of test ( P l d ) and control ( L l d ) rabbits Sex
Weight (g)
Age (months)
PI P2 P3 P4 P5 P6
F M F F F F
345 380 800 770 600 375
4
L1 L2 L3
F
3300 2950 2900 2900 3920 5155
F F F F F
IA
L5 L6
4 6 6 6 4 6
4 4 4 6 6
In the parenchyma the shape of the individual cells and groups of cells was more variable and the groups of cells tended to be smaller (figs. 6 and 7). A subjective assessment of the size of airway and parenchymal groups of cells revealed no difference between the test and control groups. TABLEI1 Areas of section examined (cm2) and numbers of individual and groups of argyrophil cells counted Areas of section examined (cm2)
Parenchyma &p Cells
Groups of cells
P1 P2 P3 P4 P5 P6
4.943 5.792 12.158 9.599 10-147 7.299
6 7 26 27 -. 21 47
13 26 48 55
L1 L2 L3
20.975 24664 23.041 21408 26836 27.144
35 15 11 I5 11 12
133 84 35 80
L4 L5 L6
76
73
40
80
Airways -
7
Cells
Groups of cells
4 22 63 48 57 33
35 80 64 83 110 115
116 85
164 101 67 148 143 120
75
143 146 84
Quantitative results The sex, weight and age of the test and control animals is given in table I. Table I1 shows the area of section examined and the number of individual argyrophil cells and groups of argyrophil cells counted in the airways and
W. TAYLOR
140
parenchyma of test and control animals. Individual argyrophil cells and groups of argyrophil cells occurred with greatest frequency in the medium-sized bronchi with cartilage in their walls but were also present in small bronchi without cartilage, in bronchioles, and in the walls of the alveolar ducts and alveoli. They were present in the alveolar walls in smaller numbers than in the bronchi and bronchioles. Wilcoxon’s ranking test (Dixon and Massey, 1957) was employed to assess the statistical significance of differences between test and control groups. This test makes no assumption about the equality of variances or the type of distribution of results. It is carried out by placing the result obtained for each animal TABLE I11
Number of groups of argyrophil cellslcm2
PI P2 P3
P4 P5 P6
L1 L2 L3
LA L5 L6
Means
Parenchyma
Airways
2.630 (3) 4489 (8) 3.948 (7) 5.730 19, 7.490 (ii) 9.864 (12)
7.081 (7) 13.812 (11) 5.264 4) 8.M7 $9,
ioiii (io) 15.756 (12)
6.341 (10) 3.406 (5) 1.519 (2) 3.652 (6j 1.491 (1) 2.947 (4)
7.819 (8) 4095 (2)
Test = 5.69 Control = 3.23
Test = 10.23 Control = 5.22
2.908 ( 1 )
- . ._
6.756 (Sj 5.329 (5) 4.421 (3)
Wilcoxon, p = 0094 Wilcoxon, p
= 0.026
Figures in brackets indicate rank of each value.
in a rank from lowest to highest, regardless of whether the rabbit concerned is a test or control animal. The sum of the ranks for test or control animals is calculated and the corresponding value of p is obtained from Wilcoxon’s table. Table 111 shows the incidence of groups of argyrophil cells in the test and control animals. In both airways and parenchyma there were more groups of cells in the test than in the control rabbits. In the airways the value of p obtained (0.026) indicates that there are 2.6 chances in 100 of this difference between test and control animals having been obtained by chance, i.e., the difference is significant at the 5 per cent. level. No significant difference was present for groups of cells in the parenchyma. In the case of individual cells, a significant difference was found in the parenchyma (table IV). In the parenchyma of the test animals there were more individual cells/cm2 and this difference was significant at the 1 per cent. level. No significant difference was demonstrated between test and control groups in the case of individual cells in the airways.
TAYLOR
PLATE LXIII
PULMONARY ARGYROPHIL
CELLS
FIG. 1 .--Small bronchus in human lung. The bronchus is apparently lined by the familiar ciliated epithelial cells but the smaller, darker nuclei (arrow) are nuclei of argyrophil cells. Haematoxylin and eosin. s 600.
FIG. 2.--Small bronchus in human lung. Large numbers of argyrophil cells are present. Grimelius silver nitrate argyrophil technique. s 600.
TAYLOR
PLATE
PULMONARY ARGYROPHIL
LXIV
CELLS
FIG. 3.-Rabbit lurg. Typical dense core secretory vesicles (arrow) in the cytoplasm of a group of airway argyrophil cells. Electron micrograph. x 18,750.
FIc;. 4.-Control-rat bit lung. A spindle-shaped individual argyrophil cell is present in an airway. The tapering process does not reach the airway lumen within the plane of section. Grimelius stain. x 1500.
TAYLOR
PLATELXV PULMONARY ARGYROPHIL
FIG.
5.-Test-rabbit
CELLS
lung. A group of-argyrophil cells in a small bronchus. corpuscular structure. Grimelius stain. x 1500.
FIG. 6.-Test-rabbit lung. An individual argyrophil cell (arrow: in a n alveolar wall. Such cells are often pokigonal in shape in this site. Grimelius stain. x 1500.
The group has a
FIG.7.-Test-rabbit lung. A group ofargyrophi1 cells in an alveolar wall. Such groups tend to contain fewer cells and to have a less characteristic shape than their airway counterparts. Grimelius stain. x 1500.
PULMONARY ARG YROPHIL CELLS
141
DISCUSSION The results indicate that in animals exposed to hypoxia from birth there a.re more groups of argyrophil cells in the airways and more individual argyrophil cells in the lung parenchyma than in sea-level controls. A nerve supply has been demonstrated to groups of argyrophil cells in the rabbit (Lauweryns and Cokelaere, 197:1), but not as yet to individual cells, so it would not be surprising to find that the groups of cells behave differently from individual cells in hypoxic condit:.ons. What is surprising is the apparent inconsistent behaviour of cells and groups of cells in the two sites, airways and parenchyma. The failure to demonstrate a difference between test and control animals in all TABLEIV Number of individual argyrophil cellslcrnz Parenchyma
Airways
P1 P2 P3 P4 P5 P6
1,214 (7) 1.209 (6) 2.139 (10) 2.813 (11) 2.070 (9) 6.439 (12)
0.809 1) 3.798 15) 5.182 (8) 5.001 (7j 5.617 (11) 4.521 (6)
L1 L2 L3
1.669 (8) 0.608 (4) 0.477 (3) 0.685 (5) 0.410 (1) 0442 (2)
5.530 (10) 3446 (4) 3.255 (3) 6.527 (12) 5.440 (9) 3.095 (2)
L4 L5 L6
Means
Test Control
= =
2.65 0.72
Test Control
Wilcoxon, p
=
0.008 Wilcoxon, p
= 415 = 4.55 = 0.938
Figures in brackets indicate rank of each value.
categories doe:; not necessarily mean that no such difference exists. The possibility that individual argyrophil cells and groups of cells have different functions according to their site should be considered. Other authors have failed to describe individual argyrophil cells in the alveolar walls of the rabbit and the demonstration that such cells are present in larger numbers in chronically hypoxic r;ibbits helps confirm their existence. Nothing is known about whether the number of pulmonary argyrophil cells in the rabbit varies during the lifetime of the animal. These cells have been found in neonatal rabbits but have not been quantified. The presence of more argyrophil cells in high-altitude rabbits than in sea-level controls could be the result of their undergoing hyperplasia in early life, or it could be a genetic characteristic of the native high-altitude rabbit, i.e., a result of natural selection occurring over the many thousands of years which Peruvian rabbits have lived at high altitude. The outstanding environmental difference between the test and control arimals is the hypoxia of the high-altitude environment. It is suggested that the difference in incidence of argyrophil cells in test and control
142
W. TA YLOR
animals, whether it is inherited or acquired, may be an effect of hypoxia. If acquired, any increase in argyrophil cells would probably occur in early extrauterine life, on exposure to hypoxia, rather than in utero. The foetus is probably no more hypoxaemic at high altitude than it is at sea level. Both situations expose the foetus to hypoxaemia but the high-altitude environment is compensated for by an abnormally heavy placenta which probably has a greater density of histological components (Heath and Williams). An increase in the number of argyrophil cells could theoretically occur by division of existing cells. However, division of individual cells would then produce groups of cells. The finding that there were more individual cells in the parenchyma of the test animals is not compatible with this explanation. Similarly, the finding of more groups of cells in the airways of the test animals was not associated with a decrease in the number of individual cells. Also the groups of cells in the test animals were no bigger than those in the control animals, It is suggested that the increased number of individual cells and groups of cells could arise by differentiation of stem cells. The groups of cells are innervated and it is possible that newly arisen groups of argyrophil cells acquire their nerve supply by migration of nerve fibres and differentiation of nerve endings. This phenomenon has been demonstrated in rabbit ear chambers where vasomotor nerves migrate into an ear chamber to supply the smoothmuscle of developing arterioles. Sensory nerves may also grow into healing wounds (Florey and Jennings, 1962). The hypothesis that pulmonary argyrophil cells are chemoreceptors which are sensitive to hypoxia has been put forward by Lauweryns and Cokelaere (1973), who referred to groups of argyrophil cells in airways as neuro-epithelial bodies. Although Lauweryns’ and Cokelaere’s work involved subjecting neonatal rabbits to acute hypoxia and the present study is concerned with chronic hypoxia, the finding that there are more groups of argyrophil cells in the airways and more individual cells in the parenchyma in the chronically hypoxic test animals provides support for Lauweryns’ and Cokelaere’s hypothesis. Lauweryns and Cokelaere (1 973) have pointed out that pulmonary argyrophil cells appear ideally situated to play a part in the pulmonary vasoconstrictor response to hypoxia. They propose that the cells detect the hypoxia in the bronchial tree and alveoli and a locally acting vasoactive substance released by them produces vasoconstriction, the entire process being subject to the influence of the central nervous system. The mechanism of hypoxic pulmonary vasoconstriction has long been in doubt and this interesting speculation deserves further investigation. In this context Lauweryns and Cokelaere have demonstrated exocytosis of dense-core vesicles from groups of argyrophil cells across the cell membrane adjacent to the bronchial basement membrane in experimental hypoxia. These vesicles have the structure of typical polypeptide secretion vesicles. However, the implication that a locally-acting polypeptide released from pulmonary argyrophil cells is the mediator of the pulmonary vasoconstrictor response to hypoxia must be interpreted with some caution. The only vasoactive substance as yet demonstrated in these cells is 5-hydroxytryptamine,
PULMONARY ARG YROPHIL CELLS
143
and it is now suggested that recent work has eliminated this substance as the chemical mediator in this response (Fishman, 1976). However, other vasoactive substances, in addition to 5-hydroxytryptamine, may be elaborated by the cells. Any vasoactive substance released by bronchial argyrophil cells would have access to the bronchial capillaries, which drain into the pulmonary veins. The main site of the vasoconstrictor response to hypoxia is the pulmonary arterioles, but any substance released into the bronchial capillaries does not come into contact with those vessels. However, although pulmonary arteriolar constriction dominates the hypoxic pressor response, there may well be a venous element also (Fishman), and a polypeptide mediator elaborated by hypoxia sensitive argyrophil cells might possibly participate in pulmonary venoconstriction. Numerous other functions in a variety of pathological conditions have been proposed for the pulmonary argyrophil cell (Lauweryns and Cokelaere) but as yet there is no evidence to support these suggestions. It seems likely, though, that the pulmonary argyrophil cell is the cell of origin of the bronchial carcinoid tumour and oat-cell carcinoma, both of which may produce endocrine effects.
SUMMARY The numbers of individual argyrophil cells and groups of argyrophil cells were compared in rabbits which had been born and had spent their entire lives at a height of 4300 m above sea level and in sea-level controls. In the bronchi and bronchioles there were more groups of argyrophil cells in the high-altitude rabbits (mean 10*23/cm2)than in the sea-level controls (mean 5*22/cm2). In the alveolar walls there were more individual cells in the highaltitude rabbits (mean 2-65/cm2)than in the sea-level controls (mean 0*72/cm2). These differences may be either inherited or acquired and in either case it is suggested that the likely explanation for the differences is hypoxia. If acquired, the differences may be due to differentiation of argyrophil cells from a stem cell. The results provide evidence that individual argyrophil cells exist in the alveolar walls. They suggest that individual argyrophil cells and groups of argyrophil cells may have different functions according to their site. I am grateful to Professor D. Heath for helpful advice and criticism and for providing the material from the high-altitude rabbits obtained during field work with Professor P. Harris supported by a grant from the Nuffield Foundation. REFERENCES
DIXON, W. J., AND MASSEY, F. J. 1957. Introduction to statistical analysis, 2nd ed., McGraw Hill, New York, p. 289.
FEYRTER, F. 1954. Zur pathologie des argyrophilen helle-zellen-organes im bronchialbaum des menschen. Virchows Arch. path. Anat. Physiol., 325, 723. FISHMAN, A. P. 1976. Hypoxia on the pulmonary circulation. How and where it acts. Circulation Res., 38, 221. FLOREY, H. W., AND JENNINGS, M. A. 1962. In General pathology, 3rd ed., Lloyd-Luke, London, pp. 459, 474. GRIMELIUS, L. 1968. A silver nitrate stain for a 2 cells in human pancreatic islets. Acta SOC.Med. upsal., 73, 243.
144
W. TA YLOR
HEATH, D., AND WILLIAMS, D. R. 1977. Man at high altitude, 1st ed., Churchill Livingstone, Edinburgh, pp. 220-221, 227-228. LAROS,C. D. 1971. Local chemical regulation of the flow resistance in the bronchial tree and pulmonary circulation. Respiration, 28, 120. LAUWERYNS, J. M., AND COKELAERE, M. 1973. Hypoxia-sensitive neuroepithelial bodies: intrapulmonary secretory neuroreceptors modulated by the CNS. Z. Zellforsch. mikrosk. Anat., 145, 521. LAUWERYNS, J. M., COKELAERE, M., THEUNYNCK, P., AND DELEERSNYDER, M. 1974. Neuroepithelial bodies in mammalian respiratory mucosa : light optical, histochemical and ultrastructural studies. Chest, 65, 22 (supplement). MOOSAW, HOMEIFCA, SMITH,P., AND HEATH,D. 1973. The Feyrter cell in hypoxia. Thorax, 28, 729.
PLATESLXIII-LXV ARGYROPHIL cells in human lung were first described by Feyrter as recently as 1954. The probable reason for their late discovery is that they are not easily distinguishable in sections stained by haematoxylin and eosin (fig. 1). Special methods such as the silver nitrate argyrophil technique of Grimelius (1968) are needed for their demonstration. Previous methods for argyrophilia have not produced such consistent results. Sections stained by this method show that human lung may contain large numbers of these cells (fig. 2). The ultrastructure of pulmonary argyrophil cells is characterised by small bodies each about 120 nm in diameter (fig. 3). These bodies are termed " dense-core secretory vesicles " and consist of a dense osmiophilic core separated from an outer limiting membrane by a narrow clear halo. In the rabbit the vesicles have been shown to contain 5-hydroxytryptamine (Lauweryns et al., 1974). Where the argyrophil cells occur in groups in the rabbit, nerve endings have been demonstrated ramifying amongst them (Lauweryns and Cokelaere, 1973). These cells are situated in bronchi and bronchioles, and in alveolar walls, and their free surfaces are in contact with the inspired gases. In these sites, and with a nerve supply, they could conceivably act as chemoreceptors for respiratory gases, perhaps playing a part in the pulmonary vasoconstrictor response to hypoxia (Laros, 1971). The purpose of the present study was to determine whether the lungs of rabbits living under hypoxic conditions at high altitude contain any more argyrophil cells than members of the same species living at sea level. MATERIALS AND METHODS A study was made of tissue from six rabbits, one male and five females, which had been born and spent their entire lives at Cerro de Pasco in the Peruvian Andes at a n altitude of 4300 m above sea level. The study of native rabbits of the Peruvian Andes precludes the use of animals of the same strain as controls. It is not possible to match the animals for weight since the high altitude rabbits were much smaller than any laboratory strain of rabbit of the same age. It was thought more important to match the animals for age since it has been suggested that, at least in the rat, pulmonary argyrophil cells decline in number after the neonatal period (Moosavi, Smith and Heath, 1973). Laboratory strains of rabbit could have been used as test animals in a suitable hypobaric chamber, thus removing the disadvantage of having controls of a different strain. However, animals in hypobaric chambers are low altitude animals exposed to acute hypoxia which is interrupted by daily episodes of repressurisation for feeding and cleaning. The use of a hypobaric chamber would simulate the rapid transportation of a sea-level rabbit to high altitude. When a sea-level-dwelling man Received 2 Sept. 1976; accepted 11 Nov. 1976. J. PATH.-VOL.
122 (1977)
137
138
W. TAYLOR
attempts this acute change in altitude he undergoes initially accommodation and then after some months or years, acquired acclimatisation. He probably never manages to live at high altitude as easily as the natives who are born there, and whose condition may be described as natural acclimatisation (Heath and Williams, 1977). The native Peruvian rabbits employed were indigenous high-altitude rabbits which had been exposed to chronic hypoxia from birth and which should exhibit natural acclimatisation. The rabbits obtained were being bred locally for food. Their breed was indeterminate and they were designated “ Peruvian Test Rabbits ” (Pl-6). At the High Altitude Research Station at Cerro these animals were killed by intraperitoneal injection of pentobarbitone sodium and weighed. The thoracic organs were rapidly removed and each lobe of the lung cut off and placed in Bouin’s solution. After initial fixation of 10 minutes’ duration to assist cutting of the tissues, blocks 3-4 mm thick were cut approximately at right angles to the major airways, commencing at the hilum and progressing towards the periphery of the lung. Each block was left attached to its neighbour by a small area of lung tissue at the pleural surface to aid in orientation of the blocks. The rabbit has an appendage to the lower lobe of the right lung, the cardiac lobe, and this was treated in the same way as the other lobes. After slicing, each lobe was fixed in Bouin’s solution for 48 hr and then transferred back to the United Kingdom. On arrival 4 or 5 wk later the blocks were separated from their last attachment to each other, numbered consecutively from the hilum outwards for each lobe, and processed into paraffin wax. The number of blocks obtained from P1-6 was 25, 25, 33, 35, 35 and 30 respectively. Six New Zealand White rabbits of corresponding ages were studied as controls. Their tissues were treated in the same manner except that the storage time in 70 per cent. alcohol was only 2 days. They were designated “ Liverpool Control Rabbits ” (Ll-6). The number of blocks obtained from L1-6 was 22, 33, 30, 33, 34 and 34 respectively. Two adjacent sections cut from each block of tissue were stained with the Grimelius argyrophil technique and with haematoxylin and eosin. Argyrophil cells occurred singly and in groups in bronchi and bronchioles and in alveolar walls. Those in bronchi and bronchioles were designated “ airway ”, and those in alveolar walls ‘‘ parenchymal ” argyrophil cells and groups of cells. The numbers of individual and groups of argyrophil cells were counted by drawing a grid over the coverslip of each slide and systematically scanning each square. The area of each irregularly shaped section of lung was found using a point-counting technique and a WILD M501 Automatic Sampling Microscope. This device automatically moves its stage one graticule width across the section in the desired direction enabling the operator to point count the area of the section. For the point counting a 42-point Weibel graticule was employed on a projection head. Over lo00 points were counted for most sections and in no case was the number counted less than 380. Despite the number of sections counted the total areas of P1, P2 and P6 were small. This was because these three animals were younger and lighter (table I) and therefore smaller than P3, P4 and P5. Since they were not fixed in distension all the lungs were partly collapsed. Fewer blocks could conveniently be cut from the smaller collapsed lungs of P1, P2 and P6, accounting for the smaller total areas examined in those animals. In each rabbit the number of argyrophil cells and groups of cells present was expressed per cmz of tissue.
RESULTS Individual argyrophil cells in the airways were often spindle-shaped or conical (fig. 4). When spindle-shaped, the nucleus of the cell, which was distinguishable only as an oval agranular area, tended to be nearer the airway basement membrane than the luminal surface. Conical individual cells had their bases on the airway basement membrane. In both types a thin tapering process extended towards the free surface but did not always reach it within the plane of section. Groups of argyrophil cells in the airways had a corpuscular
PULMONARY ARG YROPHIL CELLS
139
structure (fig. 5). Often ten or more cells could be counted in a group in the plane of section. In the airways argyrophil cells or groups of cells were often present close to bifurcations and sometimes in the mucosa overlying an aggregate of lymphoid tissue. TABLEI Sex, weight and age of test ( P l d ) and control ( L l d ) rabbits Sex
Weight (g)
Age (months)
PI P2 P3 P4 P5 P6
F M F F F F
345 380 800 770 600 375
4
L1 L2 L3
F
3300 2950 2900 2900 3920 5155
F F F F F
IA
L5 L6
4 6 6 6 4 6
4 4 4 6 6
In the parenchyma the shape of the individual cells and groups of cells was more variable and the groups of cells tended to be smaller (figs. 6 and 7). A subjective assessment of the size of airway and parenchymal groups of cells revealed no difference between the test and control groups. TABLEI1 Areas of section examined (cm2) and numbers of individual and groups of argyrophil cells counted Areas of section examined (cm2)
Parenchyma &p Cells
Groups of cells
P1 P2 P3 P4 P5 P6
4.943 5.792 12.158 9.599 10-147 7.299
6 7 26 27 -. 21 47
13 26 48 55
L1 L2 L3
20.975 24664 23.041 21408 26836 27.144
35 15 11 I5 11 12
133 84 35 80
L4 L5 L6
76
73
40
80
Airways -
7
Cells
Groups of cells
4 22 63 48 57 33
35 80 64 83 110 115
116 85
164 101 67 148 143 120
75
143 146 84
Quantitative results The sex, weight and age of the test and control animals is given in table I. Table I1 shows the area of section examined and the number of individual argyrophil cells and groups of argyrophil cells counted in the airways and
W. TAYLOR
140
parenchyma of test and control animals. Individual argyrophil cells and groups of argyrophil cells occurred with greatest frequency in the medium-sized bronchi with cartilage in their walls but were also present in small bronchi without cartilage, in bronchioles, and in the walls of the alveolar ducts and alveoli. They were present in the alveolar walls in smaller numbers than in the bronchi and bronchioles. Wilcoxon’s ranking test (Dixon and Massey, 1957) was employed to assess the statistical significance of differences between test and control groups. This test makes no assumption about the equality of variances or the type of distribution of results. It is carried out by placing the result obtained for each animal TABLE I11
Number of groups of argyrophil cellslcm2
PI P2 P3
P4 P5 P6
L1 L2 L3
LA L5 L6
Means
Parenchyma
Airways
2.630 (3) 4489 (8) 3.948 (7) 5.730 19, 7.490 (ii) 9.864 (12)
7.081 (7) 13.812 (11) 5.264 4) 8.M7 $9,
ioiii (io) 15.756 (12)
6.341 (10) 3.406 (5) 1.519 (2) 3.652 (6j 1.491 (1) 2.947 (4)
7.819 (8) 4095 (2)
Test = 5.69 Control = 3.23
Test = 10.23 Control = 5.22
2.908 ( 1 )
- . ._
6.756 (Sj 5.329 (5) 4.421 (3)
Wilcoxon, p = 0094 Wilcoxon, p
= 0.026
Figures in brackets indicate rank of each value.
in a rank from lowest to highest, regardless of whether the rabbit concerned is a test or control animal. The sum of the ranks for test or control animals is calculated and the corresponding value of p is obtained from Wilcoxon’s table. Table 111 shows the incidence of groups of argyrophil cells in the test and control animals. In both airways and parenchyma there were more groups of cells in the test than in the control rabbits. In the airways the value of p obtained (0.026) indicates that there are 2.6 chances in 100 of this difference between test and control animals having been obtained by chance, i.e., the difference is significant at the 5 per cent. level. No significant difference was present for groups of cells in the parenchyma. In the case of individual cells, a significant difference was found in the parenchyma (table IV). In the parenchyma of the test animals there were more individual cells/cm2 and this difference was significant at the 1 per cent. level. No significant difference was demonstrated between test and control groups in the case of individual cells in the airways.
TAYLOR
PLATE LXIII
PULMONARY ARGYROPHIL
CELLS
FIG. 1 .--Small bronchus in human lung. The bronchus is apparently lined by the familiar ciliated epithelial cells but the smaller, darker nuclei (arrow) are nuclei of argyrophil cells. Haematoxylin and eosin. s 600.
FIG. 2.--Small bronchus in human lung. Large numbers of argyrophil cells are present. Grimelius silver nitrate argyrophil technique. s 600.
TAYLOR
PLATE
PULMONARY ARGYROPHIL
LXIV
CELLS
FIG. 3.-Rabbit lurg. Typical dense core secretory vesicles (arrow) in the cytoplasm of a group of airway argyrophil cells. Electron micrograph. x 18,750.
FIc;. 4.-Control-rat bit lung. A spindle-shaped individual argyrophil cell is present in an airway. The tapering process does not reach the airway lumen within the plane of section. Grimelius stain. x 1500.
TAYLOR
PLATELXV PULMONARY ARGYROPHIL
FIG.
5.-Test-rabbit
CELLS
lung. A group of-argyrophil cells in a small bronchus. corpuscular structure. Grimelius stain. x 1500.
FIG. 6.-Test-rabbit lung. An individual argyrophil cell (arrow: in a n alveolar wall. Such cells are often pokigonal in shape in this site. Grimelius stain. x 1500.
The group has a
FIG.7.-Test-rabbit lung. A group ofargyrophi1 cells in an alveolar wall. Such groups tend to contain fewer cells and to have a less characteristic shape than their airway counterparts. Grimelius stain. x 1500.
PULMONARY ARG YROPHIL CELLS
141
DISCUSSION The results indicate that in animals exposed to hypoxia from birth there a.re more groups of argyrophil cells in the airways and more individual argyrophil cells in the lung parenchyma than in sea-level controls. A nerve supply has been demonstrated to groups of argyrophil cells in the rabbit (Lauweryns and Cokelaere, 197:1), but not as yet to individual cells, so it would not be surprising to find that the groups of cells behave differently from individual cells in hypoxic condit:.ons. What is surprising is the apparent inconsistent behaviour of cells and groups of cells in the two sites, airways and parenchyma. The failure to demonstrate a difference between test and control animals in all TABLEIV Number of individual argyrophil cellslcrnz Parenchyma
Airways
P1 P2 P3 P4 P5 P6
1,214 (7) 1.209 (6) 2.139 (10) 2.813 (11) 2.070 (9) 6.439 (12)
0.809 1) 3.798 15) 5.182 (8) 5.001 (7j 5.617 (11) 4.521 (6)
L1 L2 L3
1.669 (8) 0.608 (4) 0.477 (3) 0.685 (5) 0.410 (1) 0442 (2)
5.530 (10) 3446 (4) 3.255 (3) 6.527 (12) 5.440 (9) 3.095 (2)
L4 L5 L6
Means
Test Control
= =
2.65 0.72
Test Control
Wilcoxon, p
=
0.008 Wilcoxon, p
= 415 = 4.55 = 0.938
Figures in brackets indicate rank of each value.
categories doe:; not necessarily mean that no such difference exists. The possibility that individual argyrophil cells and groups of cells have different functions according to their site should be considered. Other authors have failed to describe individual argyrophil cells in the alveolar walls of the rabbit and the demonstration that such cells are present in larger numbers in chronically hypoxic r;ibbits helps confirm their existence. Nothing is known about whether the number of pulmonary argyrophil cells in the rabbit varies during the lifetime of the animal. These cells have been found in neonatal rabbits but have not been quantified. The presence of more argyrophil cells in high-altitude rabbits than in sea-level controls could be the result of their undergoing hyperplasia in early life, or it could be a genetic characteristic of the native high-altitude rabbit, i.e., a result of natural selection occurring over the many thousands of years which Peruvian rabbits have lived at high altitude. The outstanding environmental difference between the test and control arimals is the hypoxia of the high-altitude environment. It is suggested that the difference in incidence of argyrophil cells in test and control
142
W. TA YLOR
animals, whether it is inherited or acquired, may be an effect of hypoxia. If acquired, any increase in argyrophil cells would probably occur in early extrauterine life, on exposure to hypoxia, rather than in utero. The foetus is probably no more hypoxaemic at high altitude than it is at sea level. Both situations expose the foetus to hypoxaemia but the high-altitude environment is compensated for by an abnormally heavy placenta which probably has a greater density of histological components (Heath and Williams). An increase in the number of argyrophil cells could theoretically occur by division of existing cells. However, division of individual cells would then produce groups of cells. The finding that there were more individual cells in the parenchyma of the test animals is not compatible with this explanation. Similarly, the finding of more groups of cells in the airways of the test animals was not associated with a decrease in the number of individual cells. Also the groups of cells in the test animals were no bigger than those in the control animals, It is suggested that the increased number of individual cells and groups of cells could arise by differentiation of stem cells. The groups of cells are innervated and it is possible that newly arisen groups of argyrophil cells acquire their nerve supply by migration of nerve fibres and differentiation of nerve endings. This phenomenon has been demonstrated in rabbit ear chambers where vasomotor nerves migrate into an ear chamber to supply the smoothmuscle of developing arterioles. Sensory nerves may also grow into healing wounds (Florey and Jennings, 1962). The hypothesis that pulmonary argyrophil cells are chemoreceptors which are sensitive to hypoxia has been put forward by Lauweryns and Cokelaere (1973), who referred to groups of argyrophil cells in airways as neuro-epithelial bodies. Although Lauweryns’ and Cokelaere’s work involved subjecting neonatal rabbits to acute hypoxia and the present study is concerned with chronic hypoxia, the finding that there are more groups of argyrophil cells in the airways and more individual cells in the parenchyma in the chronically hypoxic test animals provides support for Lauweryns’ and Cokelaere’s hypothesis. Lauweryns and Cokelaere (1 973) have pointed out that pulmonary argyrophil cells appear ideally situated to play a part in the pulmonary vasoconstrictor response to hypoxia. They propose that the cells detect the hypoxia in the bronchial tree and alveoli and a locally acting vasoactive substance released by them produces vasoconstriction, the entire process being subject to the influence of the central nervous system. The mechanism of hypoxic pulmonary vasoconstriction has long been in doubt and this interesting speculation deserves further investigation. In this context Lauweryns and Cokelaere have demonstrated exocytosis of dense-core vesicles from groups of argyrophil cells across the cell membrane adjacent to the bronchial basement membrane in experimental hypoxia. These vesicles have the structure of typical polypeptide secretion vesicles. However, the implication that a locally-acting polypeptide released from pulmonary argyrophil cells is the mediator of the pulmonary vasoconstrictor response to hypoxia must be interpreted with some caution. The only vasoactive substance as yet demonstrated in these cells is 5-hydroxytryptamine,
PULMONARY ARG YROPHIL CELLS
143
and it is now suggested that recent work has eliminated this substance as the chemical mediator in this response (Fishman, 1976). However, other vasoactive substances, in addition to 5-hydroxytryptamine, may be elaborated by the cells. Any vasoactive substance released by bronchial argyrophil cells would have access to the bronchial capillaries, which drain into the pulmonary veins. The main site of the vasoconstrictor response to hypoxia is the pulmonary arterioles, but any substance released into the bronchial capillaries does not come into contact with those vessels. However, although pulmonary arteriolar constriction dominates the hypoxic pressor response, there may well be a venous element also (Fishman), and a polypeptide mediator elaborated by hypoxia sensitive argyrophil cells might possibly participate in pulmonary venoconstriction. Numerous other functions in a variety of pathological conditions have been proposed for the pulmonary argyrophil cell (Lauweryns and Cokelaere) but as yet there is no evidence to support these suggestions. It seems likely, though, that the pulmonary argyrophil cell is the cell of origin of the bronchial carcinoid tumour and oat-cell carcinoma, both of which may produce endocrine effects.
SUMMARY The numbers of individual argyrophil cells and groups of argyrophil cells were compared in rabbits which had been born and had spent their entire lives at a height of 4300 m above sea level and in sea-level controls. In the bronchi and bronchioles there were more groups of argyrophil cells in the high-altitude rabbits (mean 10*23/cm2)than in the sea-level controls (mean 5*22/cm2). In the alveolar walls there were more individual cells in the highaltitude rabbits (mean 2-65/cm2)than in the sea-level controls (mean 0*72/cm2). These differences may be either inherited or acquired and in either case it is suggested that the likely explanation for the differences is hypoxia. If acquired, the differences may be due to differentiation of argyrophil cells from a stem cell. The results provide evidence that individual argyrophil cells exist in the alveolar walls. They suggest that individual argyrophil cells and groups of argyrophil cells may have different functions according to their site. I am grateful to Professor D. Heath for helpful advice and criticism and for providing the material from the high-altitude rabbits obtained during field work with Professor P. Harris supported by a grant from the Nuffield Foundation. REFERENCES
DIXON, W. J., AND MASSEY, F. J. 1957. Introduction to statistical analysis, 2nd ed., McGraw Hill, New York, p. 289.
FEYRTER, F. 1954. Zur pathologie des argyrophilen helle-zellen-organes im bronchialbaum des menschen. Virchows Arch. path. Anat. Physiol., 325, 723. FISHMAN, A. P. 1976. Hypoxia on the pulmonary circulation. How and where it acts. Circulation Res., 38, 221. FLOREY, H. W., AND JENNINGS, M. A. 1962. In General pathology, 3rd ed., Lloyd-Luke, London, pp. 459, 474. GRIMELIUS, L. 1968. A silver nitrate stain for a 2 cells in human pancreatic islets. Acta SOC.Med. upsal., 73, 243.
144
W. TA YLOR
HEATH, D., AND WILLIAMS, D. R. 1977. Man at high altitude, 1st ed., Churchill Livingstone, Edinburgh, pp. 220-221, 227-228. LAROS,C. D. 1971. Local chemical regulation of the flow resistance in the bronchial tree and pulmonary circulation. Respiration, 28, 120. LAUWERYNS, J. M., AND COKELAERE, M. 1973. Hypoxia-sensitive neuroepithelial bodies: intrapulmonary secretory neuroreceptors modulated by the CNS. Z. Zellforsch. mikrosk. Anat., 145, 521. LAUWERYNS, J. M., COKELAERE, M., THEUNYNCK, P., AND DELEERSNYDER, M. 1974. Neuroepithelial bodies in mammalian respiratory mucosa : light optical, histochemical and ultrastructural studies. Chest, 65, 22 (supplement). MOOSAW, HOMEIFCA, SMITH,P., AND HEATH,D. 1973. The Feyrter cell in hypoxia. Thorax, 28, 729.
