92
Journal of the Royal Society of Medicine Volume 84 February 1991
Changing concepts of oral epithelium: discussion
W J Hume PhD FDSRCPS Leeds LS2 9LU
paper
Department of Dental Surgery, Leeds University Dental School, Clarendon Way,
Keywords: stem cells; proliferation; heterogeneity; bromodeoxyuridine; clonogens
Introduction Advances in our understanding of the proliferative organization in oral epithelium has been a slow process for several reasons. These include a lack of suitable experimental models to test hypotheses, difficulty in extrapolating from other renewing cell systems and the need for new investigative techniques. This paper reviews some of the ways in which our concepts of cell proliferation and differentiation have changed in the last 20 years, then discusses some of the findings from the introduction of new techniques and finally, takes a glimpse into the future.
Concepts of proliferative organization It is hardly surprising that the early concepts of proliferative organization held that all cells in the basal layer of epithelium of skin and mucosa were capable of division, that they were likely to be homogeneous with roughly equal cell cycle times, that their removal suprabasally to differentiation was the result of random processes and that suprabasal migration was the direct result of an increased basal cell 'pressure' induced during cell division. As initial hypotheses, these seemed reasonable. Examples ofthe experimental work underpinning these concepts can be seen in Leblond et al.1 Between 1969 and 1975, a number of workers showed that these concepts were too simple. This began with the finding that mouse skin epithelium, in addition to being organized vertically into recognizable strata, demonstrated a horizontal organization as a series of columns of cells extending from the surface to the basal layer2'3. The overlap between cells of the stratum corneum was minimal, which suggested that a complex pattern of migration from the basal layer had to occur to maintain this highly-ordered and non-random cell stacking. This in turn focused attention on the pattern of cell proliferation in the basal layer which had also to be highly organized and non-random and led to the concept of the existence of a number of proliferative subpopulations, of which the stem cell population seemed to be of particular importance for tissue homeostasis and in disease4'5. The effect of this work was to produce the concepts that not all cells in the basal layer can divide, that a number are instead already differentiating, that cells capable of division comprise a number of subpopulations and finally, that cell migration is not caused by cell proliferation pressure but is an independent biological process6. Our work on the filiform papilla as a basic model showed several novel findings; that basal cells were arranged on their connective tissue cores according
to an age-related proliferative hierarchy, that most proliferation occurred amongst the deepest placed basal cells, that the ones near the tops of the connective tissue cores were incapable of division and that there was an easily observable migration of basal cells along the connective tissue from a stem cell zone. Analysis of the stem cell population at the origin of the cell flow pattern,- showed that they had a dramatic circadian rhythm in proliferation, were able to store tritiated thymidine in a long-lived nucleotide pool, were radiosensitive and might undergo selective segregation of DNA at mitosis7. More recently, we investigated the rete ridge structure of mouse gingiva and sulcus epithelium and found that once again, basal cells are arranged in a strict age-related manner on the connective tissue cores and undergo a gradual decrease in proliferative capacity as they move higher (ie the older they are) in the basal layer8 (Figure 1). We also quantified for. the first time, the rate at which basal cells migrate along undulating basement membranes9, finding that the migration rate of 1.2 cell positions/hour in gingiva and 2.4 cell positions/hour in sulcus epithelium is as fast as the migration rate in intestinal epithelium. These studies show a remarkable degree of proliferative organization within gingival epithelium and give further support for stem cell concepts in oral epithelium. Although it can be said that the use of oral epithelium as a model for the investigation of the behaviour of proliferative subpopulations followed on from the work that had already begun in rodent skin, it is also true that because of its particular morphological structure and marked circadian variation in cell proliferation, it has contributed significantly to the development of an integrated cell proliferation model applicable also to bone marrow, intestine, testis, nematodes, and lately to developmental situations.
New investigative techniques Tritiated thymidinel bromodeoxyuridine double labelling Conventional tritiated thymidine (3HTdr) autoradiography has provided considerable information about average values for proliferative parameters of cells but is much less able to indicate the number of subpopulations present and their specific proliferative attributes. In an attempt to overcome this problem we have spent a number of years perfecting a double-labelling technique by combining autoradiography with the immunodetection of bromodeoxyuridine (BrdU) labelling of the same cells in tissues. 3HTdR is used to label
Based on paper read to Section of Odontology, 23 April 1990
0141 0768/91/
02009270602001/0 © 1991
The Royal Society of Medicine
Journal of the Royal Society of Medicine Volume 84 February 1991 93 a. Area 1 .(Sulcus Epithelium)
TOTAL DOSE (Gy)
30
20
0
10 0 I
0
C4)
E
0
IFGM' 20-
i
b. Area 3.(Attached Gingiva)
10-
0' 0
40 60C 20 Cell position in rete ridge
Figure 1. The decreasing proliferative potential with increasing basal cell position on the basement membrane in mouse sulcus and gingival epithelium. The deepest basal cell is at position 1. Data are based on an average of individual labelling index values 40 min after injection of 3HTdR. Observations were made at 3-h intervals and 24-h average values plotted to nullify the effect of the circadian rhythm in proliferation. Both tissues show a gradual fall in Ll with increasing basal cell position, a result that is incompatible with the concept of random processes and a homogeneous basal layer. J=junctional epithelium. FGM=free gingival margin. Reprinted8 with permission from Archives of Oral Biology, vol. 34, Pergamon Press plc, C 1989
specific cohort of cells and -after a period of time BrdU is given and the cells or tissues sampled to determine what proportion of cells -are double-labelled or single labelled with 3HTdR or BrdU. As both compounds mark cells undergoing DNA synthesis, a double-labelling experiment where BrdU is given from 0 to 12 h after 3HTdR will provide an estimate for the duration of the S phase and independent of this, the flux rates for the transit of cells from S into G2 (efflux rate) and from G, into S (influx rate). When applied to mouse tongue epithelium which has a prominent circadian variation in DNA synthesis, three main findings emerge. Firstly, that the duration of DNA synthesis for cells in S at 0900 and 2100 h are markedly different at 5.8 and 9.4 h respectively; secondly, that the influx and efflux values fall into a band of either high or low values; thirdly, that all flux values suddenly change from high to low, or vice versa, between 4 h and 5 h after labelling with 3HTdR10. These findings go some way to explain the nature of the circadian proliferative rhythm and suggest the prominent role played by subpopulations of keratinocytes. The same double-labelling method can be used to assess cell-cycle times by giving BrdU once the 3HTdR-labelled cohort has completed DNA synthesis. These cells will become double-labelled only a
-14
Figure 2. Radiation survival curves for mouse tongue epithelium using an in vitro keratinocyte assay system. The lowest curve represents in vitro irradiaton only and, is the control curve. The middle curve is the effect of a single in vivo dose of 5 Gy and further in vitro irradiation of the cell suspension with doses of0 to 10%Gy. The top curve shows that an interval of 3 days between 'the in vivo' and in vitro irradiation produces a 7-fold increase (vertical arrows) in the number of colonies in vitro.- This is due in part to the regeneration in vivo of clonogenic cells. The restoration ofthe shoulder to the curve indicates that repair ofradiation damage is also involved
when they enter the S phase of their next cell cycle. Applying this method to mouse tongue epithelium confirms that the proliferative circadian rhythm embraces subpopulations with cycle times of 24, 36, 44 and 48 h". There would seem little doubt that this double-labelling method is particularly versatile and, despite its own set of technical problems for the novice, is much easier thanidouble-labelling with 14C and 3HTdR. Development of an in vivo/in vitro clonogenic assay One of the techniques that has been responsible for a considerable increase in our knowledge of keratinocyte physiology is that of Rheinwald and Green'2, who grew colonies of keratinocytes in vitro from cell suspensions on a feeder layer of 3T3 fibroblasts. We have used this method to provide information about the radiosensitivity of oral and skin keratinocytes in vitro and in vivo and have developed a method to try to relate the clonogenicity of in vivo keratinocyte stem cell populations with in vitro clonogens. This is achieved by irradiating keratinocyte cell suspensions with doses of gamma radiation from 0 to 10 Gray. The higher the dose, the greater the proportion of proliferating cells that are killed, ie there are fewer survivors capable of growing in vitro as colonies. Plotting the colonies at each dose, as a percentage of controls, gives a dose survival curve
94
Journal of the Royal Society of Medicine Volume 84 February 1991
(Figure 2). The next stage is to observe how in vivo irradiation changes the shape ofthe in vitro radiation survival curve. A single dose of 5 Gray was given in vivo and animals killed immediately or after 3 days. From the tongues, keratinocyte suspensions were irradiated as already described. Figure 2 shows that the in vitro and the in vivo/in vitro curves are superimposable but that the effect of in vivo irradiation followed by an interval of 3 days before sampling, is to increase the number of clonogens, assessed by the ability to produce colonies in vitro, 7-fold. There are two main components to this increase, in vivo repair of sublethal and potentially lethal radiation damage and in vivo regeneration of clonogenic stem cells. Deciding upon the relative influence of each is not easy, but we estimate that in vivo stem cells regenerate by undergoing between one and two population doublings during the 3 days. This method can be adapted to provide information about the number and behaviour of clonogenic stem cells in vivo following perturbation by a variety of agents, and so increase understanding of stem cell biology.
What of the future? To understand fully the importance of stem cells in normal tissue homeostasis and their role in disease, it will be necessary to enrich them from cell suspensions as a prerequisite to determining their molecular biological attributes. We can, at last, commence this task as a result of the basic work undertaken to date. The involvement of subpopulations in proliferative'circadian rhythms in tongue will allow us to enrich S phase'cells at different times of day in the confident hope that their distinct cell cycle and S phase durations are indicators for different subpopulations. Analysis of differences in messenger RNA expression will generate complementary DNA clones for individual subpopulations which, by the technique of in situ hybridization, will allow the tissue localization of each to be determined. Our task must be to generate molecular biological probes for the stem cells and for the oldest, differentiating basal cells and eventually to use triple labelling with 3HTdR/BrdU in combination with a marker for individual subpopulations (eg stem cells or early differentiating cells). Such a regimen would be an extremely powerful way of investigating cell behaviour in vivo and- in vitro, during development, homeostasis and disease.
Acknowledgments: I am grateful to Mrs S Keat,-Miss J Moore and Mr J Thompson for their technical assistance and to the Yorkshire Cancer Research Campaign for funding some of these studies.
References 1 Leblond CP, Greulich RC, Pereira JM. Relationship of cell formation and cell migration in the renewal of stratified squamous epithelium. In: Montagna W, Billingham RE, eds. Advances in biology ofskin, vol 5. New York: Macmillan, 1964:39-67 2 Mackenzie IC. Ordered structure of the stratum corneum of the mammalian skin. Nature 1969;222. 881-3 3 Mackenzie IC. Ordered structure of the epidermis. J Invest Dermatol 1975;65:45-51 4 Potten CS. The epidermal proliferative unit: the possible role of the central basal cell. Cell Tissue Kinet 1974;7:77-88 5 Lajtha LG. Stem cell concepts. In: Potten CS, ed. Stemt cells, their identification and characterisation. Edinburgh: ChurchiU Livingstone, 1983:1-11 6 Potten CS, Hume WJ, Parkinson EK. Migration and mitosis in epidermis. Br J Dermatol 1984;111:695-9 7 Hume WJ. Stem cells in oral epithelia. In: Potten CS, ed. Stem cells, their identification and characterisation. Edinburgh: Churchill- Livingstone, 1983:233-70 8 Kellett M, Hume WJ, Potten CS. A topographical study of the circadian rhythm in labelling index of mouse gingival and floor of mouth epithelium, including changes in labelling activity with individual cell position on the epithelial ridges. Arch Oral Biol 1989;34:321-8 9 Kellett M, Hume WJ, Potten CS. Labelling studies in gingiva provide an estimate of the rate of cell migration along the basement membrane and evidence in support of an hierarchical proliferative organisation. Epithelia 1989;1:245-55 10 Hume WJ, Thompson J. Double labelling of cells with tritiated thymidine and bromodeoxyuridine reveals a circadian-dependent variation in duration of DNA synthesis and S phase flux rates in rodent oral epithelium. Cell Tissue Kinet 1990;23:313-23 11 Hume WJ. DNA synthesising cells in oral epithelium have a range of cell cycle durations: evidence from double-labelling studies using tritiated thymidine and bromodeoxyuridine. Cell Tissue Kinet 1989;22:377-82 12 Rheinwald JG, Green H. Serial cultivation of strains of human epidermal keratinocytes: the formation of keratinising colonies from single cells.. Cell 1975; 6:331-44
(Accepted 22 August 1990)
Journal of the Royal Society of Medicine Volume 84 February 1991
Changing concepts of oral epithelium: discussion
W J Hume PhD FDSRCPS Leeds LS2 9LU
paper
Department of Dental Surgery, Leeds University Dental School, Clarendon Way,
Keywords: stem cells; proliferation; heterogeneity; bromodeoxyuridine; clonogens
Introduction Advances in our understanding of the proliferative organization in oral epithelium has been a slow process for several reasons. These include a lack of suitable experimental models to test hypotheses, difficulty in extrapolating from other renewing cell systems and the need for new investigative techniques. This paper reviews some of the ways in which our concepts of cell proliferation and differentiation have changed in the last 20 years, then discusses some of the findings from the introduction of new techniques and finally, takes a glimpse into the future.
Concepts of proliferative organization It is hardly surprising that the early concepts of proliferative organization held that all cells in the basal layer of epithelium of skin and mucosa were capable of division, that they were likely to be homogeneous with roughly equal cell cycle times, that their removal suprabasally to differentiation was the result of random processes and that suprabasal migration was the direct result of an increased basal cell 'pressure' induced during cell division. As initial hypotheses, these seemed reasonable. Examples ofthe experimental work underpinning these concepts can be seen in Leblond et al.1 Between 1969 and 1975, a number of workers showed that these concepts were too simple. This began with the finding that mouse skin epithelium, in addition to being organized vertically into recognizable strata, demonstrated a horizontal organization as a series of columns of cells extending from the surface to the basal layer2'3. The overlap between cells of the stratum corneum was minimal, which suggested that a complex pattern of migration from the basal layer had to occur to maintain this highly-ordered and non-random cell stacking. This in turn focused attention on the pattern of cell proliferation in the basal layer which had also to be highly organized and non-random and led to the concept of the existence of a number of proliferative subpopulations, of which the stem cell population seemed to be of particular importance for tissue homeostasis and in disease4'5. The effect of this work was to produce the concepts that not all cells in the basal layer can divide, that a number are instead already differentiating, that cells capable of division comprise a number of subpopulations and finally, that cell migration is not caused by cell proliferation pressure but is an independent biological process6. Our work on the filiform papilla as a basic model showed several novel findings; that basal cells were arranged on their connective tissue cores according
to an age-related proliferative hierarchy, that most proliferation occurred amongst the deepest placed basal cells, that the ones near the tops of the connective tissue cores were incapable of division and that there was an easily observable migration of basal cells along the connective tissue from a stem cell zone. Analysis of the stem cell population at the origin of the cell flow pattern,- showed that they had a dramatic circadian rhythm in proliferation, were able to store tritiated thymidine in a long-lived nucleotide pool, were radiosensitive and might undergo selective segregation of DNA at mitosis7. More recently, we investigated the rete ridge structure of mouse gingiva and sulcus epithelium and found that once again, basal cells are arranged in a strict age-related manner on the connective tissue cores and undergo a gradual decrease in proliferative capacity as they move higher (ie the older they are) in the basal layer8 (Figure 1). We also quantified for. the first time, the rate at which basal cells migrate along undulating basement membranes9, finding that the migration rate of 1.2 cell positions/hour in gingiva and 2.4 cell positions/hour in sulcus epithelium is as fast as the migration rate in intestinal epithelium. These studies show a remarkable degree of proliferative organization within gingival epithelium and give further support for stem cell concepts in oral epithelium. Although it can be said that the use of oral epithelium as a model for the investigation of the behaviour of proliferative subpopulations followed on from the work that had already begun in rodent skin, it is also true that because of its particular morphological structure and marked circadian variation in cell proliferation, it has contributed significantly to the development of an integrated cell proliferation model applicable also to bone marrow, intestine, testis, nematodes, and lately to developmental situations.
New investigative techniques Tritiated thymidinel bromodeoxyuridine double labelling Conventional tritiated thymidine (3HTdr) autoradiography has provided considerable information about average values for proliferative parameters of cells but is much less able to indicate the number of subpopulations present and their specific proliferative attributes. In an attempt to overcome this problem we have spent a number of years perfecting a double-labelling technique by combining autoradiography with the immunodetection of bromodeoxyuridine (BrdU) labelling of the same cells in tissues. 3HTdR is used to label
Based on paper read to Section of Odontology, 23 April 1990
0141 0768/91/
02009270602001/0 © 1991
The Royal Society of Medicine
Journal of the Royal Society of Medicine Volume 84 February 1991 93 a. Area 1 .(Sulcus Epithelium)
TOTAL DOSE (Gy)
30
20
0
10 0 I
0
C4)
E
0
IFGM' 20-
i
b. Area 3.(Attached Gingiva)
10-
0' 0
40 60C 20 Cell position in rete ridge
Figure 1. The decreasing proliferative potential with increasing basal cell position on the basement membrane in mouse sulcus and gingival epithelium. The deepest basal cell is at position 1. Data are based on an average of individual labelling index values 40 min after injection of 3HTdR. Observations were made at 3-h intervals and 24-h average values plotted to nullify the effect of the circadian rhythm in proliferation. Both tissues show a gradual fall in Ll with increasing basal cell position, a result that is incompatible with the concept of random processes and a homogeneous basal layer. J=junctional epithelium. FGM=free gingival margin. Reprinted8 with permission from Archives of Oral Biology, vol. 34, Pergamon Press plc, C 1989
specific cohort of cells and -after a period of time BrdU is given and the cells or tissues sampled to determine what proportion of cells -are double-labelled or single labelled with 3HTdR or BrdU. As both compounds mark cells undergoing DNA synthesis, a double-labelling experiment where BrdU is given from 0 to 12 h after 3HTdR will provide an estimate for the duration of the S phase and independent of this, the flux rates for the transit of cells from S into G2 (efflux rate) and from G, into S (influx rate). When applied to mouse tongue epithelium which has a prominent circadian variation in DNA synthesis, three main findings emerge. Firstly, that the duration of DNA synthesis for cells in S at 0900 and 2100 h are markedly different at 5.8 and 9.4 h respectively; secondly, that the influx and efflux values fall into a band of either high or low values; thirdly, that all flux values suddenly change from high to low, or vice versa, between 4 h and 5 h after labelling with 3HTdR10. These findings go some way to explain the nature of the circadian proliferative rhythm and suggest the prominent role played by subpopulations of keratinocytes. The same double-labelling method can be used to assess cell-cycle times by giving BrdU once the 3HTdR-labelled cohort has completed DNA synthesis. These cells will become double-labelled only a
-14
Figure 2. Radiation survival curves for mouse tongue epithelium using an in vitro keratinocyte assay system. The lowest curve represents in vitro irradiaton only and, is the control curve. The middle curve is the effect of a single in vivo dose of 5 Gy and further in vitro irradiation of the cell suspension with doses of0 to 10%Gy. The top curve shows that an interval of 3 days between 'the in vivo' and in vitro irradiation produces a 7-fold increase (vertical arrows) in the number of colonies in vitro.- This is due in part to the regeneration in vivo of clonogenic cells. The restoration ofthe shoulder to the curve indicates that repair ofradiation damage is also involved
when they enter the S phase of their next cell cycle. Applying this method to mouse tongue epithelium confirms that the proliferative circadian rhythm embraces subpopulations with cycle times of 24, 36, 44 and 48 h". There would seem little doubt that this double-labelling method is particularly versatile and, despite its own set of technical problems for the novice, is much easier thanidouble-labelling with 14C and 3HTdR. Development of an in vivo/in vitro clonogenic assay One of the techniques that has been responsible for a considerable increase in our knowledge of keratinocyte physiology is that of Rheinwald and Green'2, who grew colonies of keratinocytes in vitro from cell suspensions on a feeder layer of 3T3 fibroblasts. We have used this method to provide information about the radiosensitivity of oral and skin keratinocytes in vitro and in vivo and have developed a method to try to relate the clonogenicity of in vivo keratinocyte stem cell populations with in vitro clonogens. This is achieved by irradiating keratinocyte cell suspensions with doses of gamma radiation from 0 to 10 Gray. The higher the dose, the greater the proportion of proliferating cells that are killed, ie there are fewer survivors capable of growing in vitro as colonies. Plotting the colonies at each dose, as a percentage of controls, gives a dose survival curve
94
Journal of the Royal Society of Medicine Volume 84 February 1991
(Figure 2). The next stage is to observe how in vivo irradiation changes the shape ofthe in vitro radiation survival curve. A single dose of 5 Gray was given in vivo and animals killed immediately or after 3 days. From the tongues, keratinocyte suspensions were irradiated as already described. Figure 2 shows that the in vitro and the in vivo/in vitro curves are superimposable but that the effect of in vivo irradiation followed by an interval of 3 days before sampling, is to increase the number of clonogens, assessed by the ability to produce colonies in vitro, 7-fold. There are two main components to this increase, in vivo repair of sublethal and potentially lethal radiation damage and in vivo regeneration of clonogenic stem cells. Deciding upon the relative influence of each is not easy, but we estimate that in vivo stem cells regenerate by undergoing between one and two population doublings during the 3 days. This method can be adapted to provide information about the number and behaviour of clonogenic stem cells in vivo following perturbation by a variety of agents, and so increase understanding of stem cell biology.
What of the future? To understand fully the importance of stem cells in normal tissue homeostasis and their role in disease, it will be necessary to enrich them from cell suspensions as a prerequisite to determining their molecular biological attributes. We can, at last, commence this task as a result of the basic work undertaken to date. The involvement of subpopulations in proliferative'circadian rhythms in tongue will allow us to enrich S phase'cells at different times of day in the confident hope that their distinct cell cycle and S phase durations are indicators for different subpopulations. Analysis of differences in messenger RNA expression will generate complementary DNA clones for individual subpopulations which, by the technique of in situ hybridization, will allow the tissue localization of each to be determined. Our task must be to generate molecular biological probes for the stem cells and for the oldest, differentiating basal cells and eventually to use triple labelling with 3HTdR/BrdU in combination with a marker for individual subpopulations (eg stem cells or early differentiating cells). Such a regimen would be an extremely powerful way of investigating cell behaviour in vivo and- in vitro, during development, homeostasis and disease.
Acknowledgments: I am grateful to Mrs S Keat,-Miss J Moore and Mr J Thompson for their technical assistance and to the Yorkshire Cancer Research Campaign for funding some of these studies.
References 1 Leblond CP, Greulich RC, Pereira JM. Relationship of cell formation and cell migration in the renewal of stratified squamous epithelium. In: Montagna W, Billingham RE, eds. Advances in biology ofskin, vol 5. New York: Macmillan, 1964:39-67 2 Mackenzie IC. Ordered structure of the stratum corneum of the mammalian skin. Nature 1969;222. 881-3 3 Mackenzie IC. Ordered structure of the epidermis. J Invest Dermatol 1975;65:45-51 4 Potten CS. The epidermal proliferative unit: the possible role of the central basal cell. Cell Tissue Kinet 1974;7:77-88 5 Lajtha LG. Stem cell concepts. In: Potten CS, ed. Stemt cells, their identification and characterisation. Edinburgh: ChurchiU Livingstone, 1983:1-11 6 Potten CS, Hume WJ, Parkinson EK. Migration and mitosis in epidermis. Br J Dermatol 1984;111:695-9 7 Hume WJ. Stem cells in oral epithelia. In: Potten CS, ed. Stem cells, their identification and characterisation. Edinburgh: Churchill- Livingstone, 1983:233-70 8 Kellett M, Hume WJ, Potten CS. A topographical study of the circadian rhythm in labelling index of mouse gingival and floor of mouth epithelium, including changes in labelling activity with individual cell position on the epithelial ridges. Arch Oral Biol 1989;34:321-8 9 Kellett M, Hume WJ, Potten CS. Labelling studies in gingiva provide an estimate of the rate of cell migration along the basement membrane and evidence in support of an hierarchical proliferative organisation. Epithelia 1989;1:245-55 10 Hume WJ, Thompson J. Double labelling of cells with tritiated thymidine and bromodeoxyuridine reveals a circadian-dependent variation in duration of DNA synthesis and S phase flux rates in rodent oral epithelium. Cell Tissue Kinet 1990;23:313-23 11 Hume WJ. DNA synthesising cells in oral epithelium have a range of cell cycle durations: evidence from double-labelling studies using tritiated thymidine and bromodeoxyuridine. Cell Tissue Kinet 1989;22:377-82 12 Rheinwald JG, Green H. Serial cultivation of strains of human epidermal keratinocytes: the formation of keratinising colonies from single cells.. Cell 1975; 6:331-44
(Accepted 22 August 1990)
