Calcif Tissue Int (1990) 47:276-283
Calcified Tissue International 9 1990Springer-VerlagNew York Inc.
Monolayer Cultures of Normal Human Bone Cells Contain Multiple Subpopulations of Alkaline Phosphatase Positive Cells Toshikatsu Matsuyama, K.-H. William Lau, and Jon E. Wergedal Departments of Medicine and Biochemistry, Loma Linda University, and Mineral Metabolism Unit, Jerry L. Pettis Memorial Veterans' Hospital, 11201 Benton Street, Loma Linda, CA 92357 USA
Summary. Cytochemical staining of normal human bone cells in monolayer cultures for alkaline phosphatase (ALP) indicated that the cultures contained mixed-cell populations. Time course evaluations of the cytochemical staining revealed, in addition to the ALP-negative cell population, at least two subpopulations of ALP-positive human bone cells with different levels of ALP. A cytochemical method has been developed which separates the ALP-positive cells into high and intermediate ALP subpopulations. In this method, human bone cells were stained for ALP using an azo-dye method and incubating at 4~ for 10 and 30 minutes, respectively. We defined the cell population that stained positively for ALP at 10 minutes as strong ALP-positive cells, and both strong and intermediate cells were stained at 30 minutes. The intermediate cells were determined from the difference between the values at the two time points. The intra- and interassay variations of the assay, with the same investigator in blinded investigations, were both less than 10% and the interobserver variation was approximately 25%. Analysis of the distribution of ALP levels in cells with a laser densitometer confirmed the presence of at least t h r e e cell s u b p o p u l a t i o n s . 1,25(OH)2D 3 treatment increased the proportions of both ALP-positive cell populations, whereas TGFbeta treatment increased only the intermediate ALP-positive cell population. On the contrary, fluoride increased the proportion of the strong ALP cells, and IGF-1 had no effect on the proportions of either ALP-positive subpopulation. When the ALPspecific activity was compared with the percentage of each ALP-positive subpopulations for the cells treated with effectors, the ALP-specific activity correlated with the total ALP-positive and with the strong ALP-positive populations but not with the intermediate ALP-positive subpopulation. In summary, this study represents the first evidence that
normal human bone cells in monolayer cultures contained at least two subpopulations of ALPpositive cells, and that bone cell effectors could have differential effects on each cell population. Key words: Human bone cells - - ALP-positive cells Cell populations - - 1,25(OH)2D 3 - - TGF-beta - Osteoblast.
The isolation and culture of human bone cells in monolayer cultures have been shown to be very useful in the investigation of bone cell metabolism and regulation. Several laboratories, including ours, have developed methods for isolating and culturing human bone cells [i-5]. Most isolation methods involved either collagenase digestion [1, 2] or outgrowth from explants of trabecular bone chips [35]. Although most of the resulting preparations exhibited osteoblast-line characteristics, they appeared to contain mixed populations of cells with various degree of differentiation [1, 2, 5]. It is now clear that the effects of many metabolic effectors on the bone cells depended on the differentiation status of the cells [2, 6, 7]. Thus, to better understand the actions of the bone effectors, it is necessary to separate the subpopulations for investigation or to investigate each subpopulation independently in the mixed cell preparations. Alkaline phosphatase (ALP) has been suggested to be a cell surface marker of osteoblastic differentiation [8-10]. Osteoblast-line cells can have various levels of ALP activity, suggesting that osteoblastlines can show different degrees of differentiation. In this study, we developed a cytochemical assay method that distinguished, in addition to the ALPnegative cell population, at least two subpopulations of ALP-positive cells in normal human bone
T. Matsuyama et al.: Human Bone Cell Subpopulations monolayer cell cultures, and tested the responses of these subpopulations to several bone cell effectors.
Materials and Methods
277 Reader at 410 nm against a reference wavelength of 490 rim. One unit of enzyme activity equals the amount of enzyme that is required to hydrolyze 1 p~mole of PNPP/minute at 37~ The enzyme activity was standardized against cellular protein content, which was determined according to Lowry et al. [11].
Materials
Cytochemical Staining of ALP Activity
Collagenase and Dulbecco's modified Eagle's Medium (DMEM) were from Gibco (Grand Island, NY). Calf serum was from Flow Laboratories (McLean, VA). Tissue culture supplies were from Falcon Industries (Oxnard, CA). Paranitrophenyl phosphate (PNPP), naphthol ASTR phosphate (3-hydroxy-2-naphthoic acid 4-chloro-2-methyl-anilide, sodium salt), red violet LB diazonium chloride (5-chloro-4-benzamide-2-methylbenzene-diazonium chloride, hemi-zinc chloride salt), fast blue BB salt (4benzoylamino-2,5-diethoxybenzenediazonium chloride hemi [zinc chloride] salt), and sodium fluoride were obtained from Sigma Chemicals Co. (St. Louis, MO). Transforming growth factor-beta 1 (TGF-bl) was obtained from R and D systems (Minneapolis, MN). Insulin-like growth factor 1 (IGF-1) was a generous gift of CIBA GEIGY (Basel, Switzerland). Other reagents were of reagent grade and were obtained from Sigma Chemical Co.
After treatment, the medium was drawn off, the cells were rinsed with assay buffer, and replaced with a buffer containing 50 mM Tris-HC1 (pH 8.6), 100 mM NaC1, 5 mM KC1, 1 mM CaCI2, and 1 mM MgC12. The staining of ALP activity was accomplished by an azo-dye method as described by Burstone [12]. The substrate was naphthol ASTR phosphate (0.8 mg/ml) and the diazonium salt was red violet LB diazonium salt (0.6 mg/ml) in the same Tris buffer. The staining was carried out either at 4 or 37~ for the indicated time. The cells were then washed with the same Tris buffer, the ALP-positive cells were counted in five representative fields per well, and the percentage of ALP-positive cells were determined. ALP-positive cells were defined as those cells clearly showing red reaction products on their cell surface. For the analysis of ALP staining intensity with the Laser Densitometer (Zeineh Video Densitometer, Biomed Instruments, Inc., Fullerton, CA), the diazonium salt was changed to Fast Blue BB because of the better absorption of the resulting azo dye. For the analysis, only cells that were well spread out were measured and the readings were corrected for background by determining the optical density next to the cell.
Cell Culture Preparation Human bone cells were isolated from the trabecular bone of femoral head samples obtained from hip replacement surgery as previously described [1]. Diced bone samples were incubated with crude collagenase (2 mg/ml) in DMEM at 37~ for 2 hours. The cell suspension was rinsed once with DMEM and plated in DMEM containing 10% calf serum. The culture medium was changed every 2 to 3 days. Cells were grown to confluence and were passaged with 0.1% trypsin in calcium- and magnesium-free Puck's Salin G. Cells were subsequently passaged every 2 weeks using 1:4 dilution ratio. Cells from passages 2-6 were used for the studies, and were plated in multiwell plates at a density of 50--100 cell/ram2 in DMEM containing 10% calf serum. After the cells were established (2 days) the medium was then changed to serum-free DMEM, and the effectors were added 1 hour later. Unless otherwise stated, treatments were for 2 days.
Statistical Analyses All data represented the mean and SD of at least six replicates, and significances were determined either by a two-tailed Student's test or by a x2-test. Correlation analyses were performed with a MICROSTAT computer program. For analysis of the frequency distribution, the mean and SD of the A L P ( + ) and ALP( + + ) subpopulations were calculated from the grouped frequency data by the method of Daniel [13]. Determination of normal distribution was carried out by the Chi-square goodness of fit test.
Results
Preparation of cell extracts After treatment, the medium was drawn off. The cells were then rinsed twice with phosphate-buffered saline, and lysed in 0.5 ml of 0.1% Triton X-100 for I hour at 37~ The cell extract was kept frozen at -20~ until assay.
ALP assay ALP activity was determined in a reaction mixture containing 30 mM PNPP, 150 mM carbonate buffer, and 1 mM MgC12 (pH 10.3) with a final volume of 0.2 ml in 96-well microwell plates. The reaction was initiated with the addition of cell extract (0.05 ml) and the reaction mixture incubated at 37~ for 1 hour. The absorbance was determined with a Dynatech Microwell Plate
Evidence for Multiple Subpopulations of ALP-Positive Cells in Normal Human Bone Cells Initial cytochemical evaluations of the ALPp o s i t i v e c e l l s in o u r h u m a n b o n e c e l l m o n o l a y e r c u l tures revealed that there were approximately 20% of cells stained positively for ALP activity using the m e t h o d a c c o r d i n g t o B u r s t o n e [12]. T h u s , t h e s e findings confirmed that our human bone cell cultures were hetereogenous with respect to ALP content. T h e r e w e r e a l s o v a r i a t i o n s in i n t e n s i t y o f A L P s t a i n i n g in t h e p o s i t i v e c e l l s , i n d i c a t i n g v a r i o u s l e v els o f A L P in t h e c e l l s . T o t e s t w h e t h e r t h e c e l l u l a r
278
T. Matsuyama et al.: Human Bone Cell Subpopulations
50
40 O
:~ ~o
0
I 10
I 20
I 30
I 40
Incubation Time (rain) Fig. 1. Time courses of the cytochemical staining for ALP of normal human bone cells. The culture medium was removed and the ALP substrate and diazonium salt solution was added. The cells were incubated at 4~ (Q) and 37~ ((3). At indicated time points, four replicates were stopped, and the percentage of ALPpositive cells was determined as described in Methods. Cells were counted in five representative fields per well by the same investigator. Results were reported as means and SD.
ALP varied continuously or exhibited discrete levels, we incubated cells with the ALP substrate (and diazonium salt capture reagent) for varying period of time before determining the percentage of ALPpositive cells. We reasoned that if the variation in ALP content was continuous, a time plot of percentage of ALP-positive cells should be linear, otherwise we should detect discrete plateaus. We found that, staining at 37~ the time plot showed two apparent plateaus at approximately 5-8 minutes and 20-30 minutes of incubation, respectively (Fig. 1). To better characterize the time course, we slowed down the enzymatic reactions by decreasing the staining temperature to 4~ The two welldefined and distinct plateaus were still evident in the time course at approximately 10 and 30 minutes, respectively. It should be noted that the percentages of ALP cells at the plateaus appeared to be very similar, if not identical for the two time curves (i.e., 8-9% and 18-20%, respectively). Together, these data suggest that our monolayer human bone cell cultures contained at least two subpopulations of ALP-positive cells with different levels of ALP. Because the cytochemical staining of ALP-positive cells at 4~ yielded more reproducible and betterdefined plateaus than at 37~ our subsequent cytochemical studies were performed at 4~ We next tested whether the relative proportions of these subpopulations of bone cells would be affected by a bone cell effector. Because 1,25(OH)2D3 has been known to stimulate ALP activity in bone cells of several species [14-16], we tested whether this agent would affect the time course of the cytochemical staining of ALP in our human bone cell monolayer cultures. Figure 2 shows that, again, two w e l l - d e f i n e d p l a t e a u s w e r e o b s e r v e d in the
0 0
10
2O
30
40
Incubation Time (min) Fig. 2. Effects of 1,25(OH)2D 3 on the time course of the cytochemical staining for ALP. The procedure was the same as for Figure l except that 10-SM 1,25(OH)ED 3 ( O ) o r solvent control (i.e., 0.01% ethanol) (O) was added at the beginning of the 2-day incubation, and that the staining procedure was performed at 4~ Data were shown as means and SD of four replicates.
1,25(OH)2D3-treated cultures that occurred at the same times, 10 and 30 minutes, as in the controls. H o w e v e r , 1 , 2 5 ( O H ) E D 3 significantly increased the relative proportions of the cells in each plateau (i.e., from 8 to 27% for the 10 minute plateau form, and from 19 to 40% for the second plateau form). No additional plateau was found in the time course even after a 60 minute staining (data not shown). These findings suggest that 1 , 2 5 ( O H ) E D 3 significantly altered the relative proportions of each subpopulation of human bone cells identified by the cytochemical assay, but it did not create additional ALP-positive subpopulations. To further analyse the cell population, histograms of the distribution of ALP cells as a function of incubation time were prepared (Fig. 3). The distribution in control cultures was significantly different from a single normal distribution (P < 0.05), but was not significantly different from the distribution of two normal populations with mean +- SD of 5.2 -2.3 and 22.7 -+ 6.6 minutes, respectively. This further supports the presence of two ALP-positive subpopulations. Because we consistently observed two plateaus at 10 and 30 minutes in each time course when staining at 4~ we have decided to choose these two time points to quantitate the relative proportions of the two ALP-containing human bone cell subpopulations. The portion of cells stained positively after 10 minutes staining represents the amount of cells that contain strong ALP activity [i.e., A L P ( + +)], and that after a 30 minute staining, indicates the total ALP-positive cells. The cells that were not stained after a 30 minute incubation were termed as ALP-negative cells [i.e., A L P ( - ) ] . The cells that contain intermediate levels of ALP [i.e., ALP(+)] that take longer than 10 minutes but shorter than 30
T. Matsuyama et al.: Human Bone Cell Subpopulations
279
A 10 8
I I
C
I
1
12 r
6 ~
B
i
1,25 Vitamin O Treated
10 ca
4
o Q.
~
o
(o
u
4
o I--
2
N
8 6 4 2
0 0
10
20
30
40
o
Incubation Time (rain)
Fig. 3. Histogram showing the distribution of subpopulations of human bone cells. The data from Figure 2 were replotted to show the percent of total cells that become positive for ALP in each time interval for control and 1,25(OH)2D3-treated cells.
minutes to stain were determined by subtracting the mean value of the A L P ( + +) from the 30 minute value or total positive cells.
40
60
80
100
Intensity of ALP stain
t20
140
160
180
(Arbitrary unit)
Fig. 4. Histogram showing the distribution of control ceils as a function of staining intensity measured with a laser densitometer. Cells were stained for 30 minutes at 4~ The range of optical densities was divided into intervals of five units (arbitrary units) and the percent of cells with densities in each interval was plotted. The distribution was divided into three groups of cells labeled A-C according to the Chi-square goodness of fit test. Density measurements were carried out on a total of 150 cells.
Analysis of Cell Subpopulations by Laser Densitomitry The method of using incubation time to separate ALP( + + ) and A L P ( + ) subpopulations has an element of subjectivity (as shown below). To test for the presence of subpopulations with a less subjective (but more time-consuming) method, a laser densitometer was used to measure the optical density in cells with and without 1,25(OH)2D 3 treatment and stained for ALP (30 minutes at 4~ The resulting distribution of ALP staining intensities was significantly different from a normal distribution (P < 0.001) and could be divided into 3-4 subpopulations labeled A-D (Figs. 4 and 5) (note that Figs. 4 and 5 plot % cells versus increasing ALP intensity which is the opposite of Figure 3 where increasing time represents decreasing ALP activity). Prior to measuring the optimal density, the cells were identified as ALP positive or ALP negative, i.e., A L P ( - ) . This identification showed that A L P ( - ) corresponded to subpopulation A which had a normal distribution of very low intensities and that ALP-positive cells corresponded to subpopulations B-D. Subpopulations B and C had normal distributions of intensities and probably correspond with A L P + and A L P + + , respectively. In the 1,25(OH)2D3-treated cells, there was an additional group D that may have been part of C, but formed a skewed distribution when combined. This analy-
20
o
~o ~/I
A
B
C
]
D
I
I
I
o CL =
I,-
0 ~
0
20
40
60
80
100
Intensity of ALP stain
120
140
160
180
{Arbitrary unit)
Fig. 5. Histogram showing the distribution of 1,25(OH)2D 3treated cells as a function of staining intensity measured with a laser densitometer. The percent of cells within each five unit (arbitrary units) interval was plotted and subpopulations were calculated, as in Figure 4. The arbitrary units of Figures 4 and 5 are similar but not identical. Similar data were obtained in a second experiment.
sis of cell ALP levels showing evidence of subpopulations of bone cells confirms the results of the more subjective method.
Evaluation of the Method To test the potential application of the timed stain-
280
T. Matsuyama et al.: Human Bone Cell Subpopulations
Table 1. Reproducibility of the cytochemical staining procedure Intraassay Coefficient Experiment
n
Exp. 1
4
Exp. 2
6
Exp. 3
6
Interassay
% ALP-positive cells 8.33, 8.93, 8.11, 9.43 8.01, 7.26, 7.80, 8.26, 8.12, 8.55 7.00, 7.26, 6.78, 6.88, 7.21, 8.07
Mean
SD
of variation
8.70
0.52
5.94
8.00
0.40
5.13
7.20
0.42
5.89
7.97
0.61
7.69
Normal human bone cell cultures were prepared and maintained as described in Methods. ALP-positive cells were stained at 4~ for 10 minutes as described in Methods. Results were presented as percentage of cells that were stained positively for ALP.
ing method for determining subpopulations, we have investigated the reproducibility and precision of this approach. With the same investigator, we found that the intraassay variation was approximately 5% and the interassay variation was approximately 7% (Table 1). With multiple investigators in a blind study (i.e., none of the investigators knew which group was treated and which was the control group), we found that interobserver variation was estimated to be 24.5% (Table 2). It should be noted that each investigator showed that 1,25(OH)ED 3 produced changes in subpopulations despite some differences in the relative proportions of the subpopulations.
With this cytochemical method, we found that 1,25(OH)2D 3 increased the amount of both subpopulations of ALP-positive cells [i.e., A L P ( + ) and A L P ( + +)]; TGF-beta appeared to increase only the intermediate form [i.e., ALP(+)] without affecting the strong positive form [i.e., A L P ( + +)]. In contrast, fluoride increased the proportion of the strong form but not the intermediate form. IGF-1 had no apparent effects on either subpopulation of ALP containing human bone cells. These findings indicate an interesting possibility that different bone cell effectors might have differential effects on different subpopulation human bone cells. In the same experiment, we found that 1,25(OH)ED3 and TGFbeta, but none of the other tested agents, significantly increased the specific activity of cellular ALP in these human bone cells. We have also evaluated the correlations between the specific activity of the cellular ALP with the relative proportions of each of the subpopulations of bone cells (Fig. 6). We found that there was a strong positive correlation (r = 0.927, P < 0.05) between the percent of total ALP cells with the specific activity of cellular ALP (Fig. 6A). A weaker positive correlation (r = 0.880, P < 0.05) was detected between the specific activity of cellular ALP and the percent of strong ALP-positive cells (Fig. 6B). However, we did not find any significant correlation (r = 0.206) between the specific activity and the relative proportion of the intermediate form.
I::
Effects of Bone Cell Effectors on each Subpopulation of Human Bone Cells
O.
E
100
Z)
E
Because we can identify three different subpopulations of human bone cells by this cytochemical method [ A L P ( - ), ALP( + ), ALP( + + )], we have decided to evaluate the effects of the following bone cell effectors: (1) 1,25(OH)ED 3 (this secosteroid has been shown to stimulate the bone cell ALP activity) [14-16]; (2) TGF-beta [this agent has recently been shown to stimulate human bone cell ALP activity (unpublished observation)]; (3) fluoride (it has been shown to directly act on osteoblasts to stimulate their proliferation and increase their cellular ALP content) [17, 18]; and (4) IGF-I [this growth factor stimulates bone cell proliferation without affecting the cellular ALP content (unpublished observation)]. For comparison, we have also determined the effects of these factors on the specific activity of cellular ALP (Table 3).
y
A, TotaJ
9 B. Strong A L P Cells :----
C. Intermediate ALP~q/
50 0. _1 ,< o (3. (0
r -
.942
p < .005
r = ~ p
Calcified Tissue International 9 1990Springer-VerlagNew York Inc.
Monolayer Cultures of Normal Human Bone Cells Contain Multiple Subpopulations of Alkaline Phosphatase Positive Cells Toshikatsu Matsuyama, K.-H. William Lau, and Jon E. Wergedal Departments of Medicine and Biochemistry, Loma Linda University, and Mineral Metabolism Unit, Jerry L. Pettis Memorial Veterans' Hospital, 11201 Benton Street, Loma Linda, CA 92357 USA
Summary. Cytochemical staining of normal human bone cells in monolayer cultures for alkaline phosphatase (ALP) indicated that the cultures contained mixed-cell populations. Time course evaluations of the cytochemical staining revealed, in addition to the ALP-negative cell population, at least two subpopulations of ALP-positive human bone cells with different levels of ALP. A cytochemical method has been developed which separates the ALP-positive cells into high and intermediate ALP subpopulations. In this method, human bone cells were stained for ALP using an azo-dye method and incubating at 4~ for 10 and 30 minutes, respectively. We defined the cell population that stained positively for ALP at 10 minutes as strong ALP-positive cells, and both strong and intermediate cells were stained at 30 minutes. The intermediate cells were determined from the difference between the values at the two time points. The intra- and interassay variations of the assay, with the same investigator in blinded investigations, were both less than 10% and the interobserver variation was approximately 25%. Analysis of the distribution of ALP levels in cells with a laser densitometer confirmed the presence of at least t h r e e cell s u b p o p u l a t i o n s . 1,25(OH)2D 3 treatment increased the proportions of both ALP-positive cell populations, whereas TGFbeta treatment increased only the intermediate ALP-positive cell population. On the contrary, fluoride increased the proportion of the strong ALP cells, and IGF-1 had no effect on the proportions of either ALP-positive subpopulation. When the ALPspecific activity was compared with the percentage of each ALP-positive subpopulations for the cells treated with effectors, the ALP-specific activity correlated with the total ALP-positive and with the strong ALP-positive populations but not with the intermediate ALP-positive subpopulation. In summary, this study represents the first evidence that
normal human bone cells in monolayer cultures contained at least two subpopulations of ALPpositive cells, and that bone cell effectors could have differential effects on each cell population. Key words: Human bone cells - - ALP-positive cells Cell populations - - 1,25(OH)2D 3 - - TGF-beta - Osteoblast.
The isolation and culture of human bone cells in monolayer cultures have been shown to be very useful in the investigation of bone cell metabolism and regulation. Several laboratories, including ours, have developed methods for isolating and culturing human bone cells [i-5]. Most isolation methods involved either collagenase digestion [1, 2] or outgrowth from explants of trabecular bone chips [35]. Although most of the resulting preparations exhibited osteoblast-line characteristics, they appeared to contain mixed populations of cells with various degree of differentiation [1, 2, 5]. It is now clear that the effects of many metabolic effectors on the bone cells depended on the differentiation status of the cells [2, 6, 7]. Thus, to better understand the actions of the bone effectors, it is necessary to separate the subpopulations for investigation or to investigate each subpopulation independently in the mixed cell preparations. Alkaline phosphatase (ALP) has been suggested to be a cell surface marker of osteoblastic differentiation [8-10]. Osteoblast-line cells can have various levels of ALP activity, suggesting that osteoblastlines can show different degrees of differentiation. In this study, we developed a cytochemical assay method that distinguished, in addition to the ALPnegative cell population, at least two subpopulations of ALP-positive cells in normal human bone
T. Matsuyama et al.: Human Bone Cell Subpopulations monolayer cell cultures, and tested the responses of these subpopulations to several bone cell effectors.
Materials and Methods
277 Reader at 410 nm against a reference wavelength of 490 rim. One unit of enzyme activity equals the amount of enzyme that is required to hydrolyze 1 p~mole of PNPP/minute at 37~ The enzyme activity was standardized against cellular protein content, which was determined according to Lowry et al. [11].
Materials
Cytochemical Staining of ALP Activity
Collagenase and Dulbecco's modified Eagle's Medium (DMEM) were from Gibco (Grand Island, NY). Calf serum was from Flow Laboratories (McLean, VA). Tissue culture supplies were from Falcon Industries (Oxnard, CA). Paranitrophenyl phosphate (PNPP), naphthol ASTR phosphate (3-hydroxy-2-naphthoic acid 4-chloro-2-methyl-anilide, sodium salt), red violet LB diazonium chloride (5-chloro-4-benzamide-2-methylbenzene-diazonium chloride, hemi-zinc chloride salt), fast blue BB salt (4benzoylamino-2,5-diethoxybenzenediazonium chloride hemi [zinc chloride] salt), and sodium fluoride were obtained from Sigma Chemicals Co. (St. Louis, MO). Transforming growth factor-beta 1 (TGF-bl) was obtained from R and D systems (Minneapolis, MN). Insulin-like growth factor 1 (IGF-1) was a generous gift of CIBA GEIGY (Basel, Switzerland). Other reagents were of reagent grade and were obtained from Sigma Chemical Co.
After treatment, the medium was drawn off, the cells were rinsed with assay buffer, and replaced with a buffer containing 50 mM Tris-HC1 (pH 8.6), 100 mM NaC1, 5 mM KC1, 1 mM CaCI2, and 1 mM MgC12. The staining of ALP activity was accomplished by an azo-dye method as described by Burstone [12]. The substrate was naphthol ASTR phosphate (0.8 mg/ml) and the diazonium salt was red violet LB diazonium salt (0.6 mg/ml) in the same Tris buffer. The staining was carried out either at 4 or 37~ for the indicated time. The cells were then washed with the same Tris buffer, the ALP-positive cells were counted in five representative fields per well, and the percentage of ALP-positive cells were determined. ALP-positive cells were defined as those cells clearly showing red reaction products on their cell surface. For the analysis of ALP staining intensity with the Laser Densitometer (Zeineh Video Densitometer, Biomed Instruments, Inc., Fullerton, CA), the diazonium salt was changed to Fast Blue BB because of the better absorption of the resulting azo dye. For the analysis, only cells that were well spread out were measured and the readings were corrected for background by determining the optical density next to the cell.
Cell Culture Preparation Human bone cells were isolated from the trabecular bone of femoral head samples obtained from hip replacement surgery as previously described [1]. Diced bone samples were incubated with crude collagenase (2 mg/ml) in DMEM at 37~ for 2 hours. The cell suspension was rinsed once with DMEM and plated in DMEM containing 10% calf serum. The culture medium was changed every 2 to 3 days. Cells were grown to confluence and were passaged with 0.1% trypsin in calcium- and magnesium-free Puck's Salin G. Cells were subsequently passaged every 2 weeks using 1:4 dilution ratio. Cells from passages 2-6 were used for the studies, and were plated in multiwell plates at a density of 50--100 cell/ram2 in DMEM containing 10% calf serum. After the cells were established (2 days) the medium was then changed to serum-free DMEM, and the effectors were added 1 hour later. Unless otherwise stated, treatments were for 2 days.
Statistical Analyses All data represented the mean and SD of at least six replicates, and significances were determined either by a two-tailed Student's test or by a x2-test. Correlation analyses were performed with a MICROSTAT computer program. For analysis of the frequency distribution, the mean and SD of the A L P ( + ) and ALP( + + ) subpopulations were calculated from the grouped frequency data by the method of Daniel [13]. Determination of normal distribution was carried out by the Chi-square goodness of fit test.
Results
Preparation of cell extracts After treatment, the medium was drawn off. The cells were then rinsed twice with phosphate-buffered saline, and lysed in 0.5 ml of 0.1% Triton X-100 for I hour at 37~ The cell extract was kept frozen at -20~ until assay.
ALP assay ALP activity was determined in a reaction mixture containing 30 mM PNPP, 150 mM carbonate buffer, and 1 mM MgC12 (pH 10.3) with a final volume of 0.2 ml in 96-well microwell plates. The reaction was initiated with the addition of cell extract (0.05 ml) and the reaction mixture incubated at 37~ for 1 hour. The absorbance was determined with a Dynatech Microwell Plate
Evidence for Multiple Subpopulations of ALP-Positive Cells in Normal Human Bone Cells Initial cytochemical evaluations of the ALPp o s i t i v e c e l l s in o u r h u m a n b o n e c e l l m o n o l a y e r c u l tures revealed that there were approximately 20% of cells stained positively for ALP activity using the m e t h o d a c c o r d i n g t o B u r s t o n e [12]. T h u s , t h e s e findings confirmed that our human bone cell cultures were hetereogenous with respect to ALP content. T h e r e w e r e a l s o v a r i a t i o n s in i n t e n s i t y o f A L P s t a i n i n g in t h e p o s i t i v e c e l l s , i n d i c a t i n g v a r i o u s l e v els o f A L P in t h e c e l l s . T o t e s t w h e t h e r t h e c e l l u l a r
278
T. Matsuyama et al.: Human Bone Cell Subpopulations
50
40 O
:~ ~o
0
I 10
I 20
I 30
I 40
Incubation Time (rain) Fig. 1. Time courses of the cytochemical staining for ALP of normal human bone cells. The culture medium was removed and the ALP substrate and diazonium salt solution was added. The cells were incubated at 4~ (Q) and 37~ ((3). At indicated time points, four replicates were stopped, and the percentage of ALPpositive cells was determined as described in Methods. Cells were counted in five representative fields per well by the same investigator. Results were reported as means and SD.
ALP varied continuously or exhibited discrete levels, we incubated cells with the ALP substrate (and diazonium salt capture reagent) for varying period of time before determining the percentage of ALPpositive cells. We reasoned that if the variation in ALP content was continuous, a time plot of percentage of ALP-positive cells should be linear, otherwise we should detect discrete plateaus. We found that, staining at 37~ the time plot showed two apparent plateaus at approximately 5-8 minutes and 20-30 minutes of incubation, respectively (Fig. 1). To better characterize the time course, we slowed down the enzymatic reactions by decreasing the staining temperature to 4~ The two welldefined and distinct plateaus were still evident in the time course at approximately 10 and 30 minutes, respectively. It should be noted that the percentages of ALP cells at the plateaus appeared to be very similar, if not identical for the two time curves (i.e., 8-9% and 18-20%, respectively). Together, these data suggest that our monolayer human bone cell cultures contained at least two subpopulations of ALP-positive cells with different levels of ALP. Because the cytochemical staining of ALP-positive cells at 4~ yielded more reproducible and betterdefined plateaus than at 37~ our subsequent cytochemical studies were performed at 4~ We next tested whether the relative proportions of these subpopulations of bone cells would be affected by a bone cell effector. Because 1,25(OH)2D3 has been known to stimulate ALP activity in bone cells of several species [14-16], we tested whether this agent would affect the time course of the cytochemical staining of ALP in our human bone cell monolayer cultures. Figure 2 shows that, again, two w e l l - d e f i n e d p l a t e a u s w e r e o b s e r v e d in the
0 0
10
2O
30
40
Incubation Time (min) Fig. 2. Effects of 1,25(OH)2D 3 on the time course of the cytochemical staining for ALP. The procedure was the same as for Figure l except that 10-SM 1,25(OH)ED 3 ( O ) o r solvent control (i.e., 0.01% ethanol) (O) was added at the beginning of the 2-day incubation, and that the staining procedure was performed at 4~ Data were shown as means and SD of four replicates.
1,25(OH)2D3-treated cultures that occurred at the same times, 10 and 30 minutes, as in the controls. H o w e v e r , 1 , 2 5 ( O H ) E D 3 significantly increased the relative proportions of the cells in each plateau (i.e., from 8 to 27% for the 10 minute plateau form, and from 19 to 40% for the second plateau form). No additional plateau was found in the time course even after a 60 minute staining (data not shown). These findings suggest that 1 , 2 5 ( O H ) E D 3 significantly altered the relative proportions of each subpopulation of human bone cells identified by the cytochemical assay, but it did not create additional ALP-positive subpopulations. To further analyse the cell population, histograms of the distribution of ALP cells as a function of incubation time were prepared (Fig. 3). The distribution in control cultures was significantly different from a single normal distribution (P < 0.05), but was not significantly different from the distribution of two normal populations with mean +- SD of 5.2 -2.3 and 22.7 -+ 6.6 minutes, respectively. This further supports the presence of two ALP-positive subpopulations. Because we consistently observed two plateaus at 10 and 30 minutes in each time course when staining at 4~ we have decided to choose these two time points to quantitate the relative proportions of the two ALP-containing human bone cell subpopulations. The portion of cells stained positively after 10 minutes staining represents the amount of cells that contain strong ALP activity [i.e., A L P ( + +)], and that after a 30 minute staining, indicates the total ALP-positive cells. The cells that were not stained after a 30 minute incubation were termed as ALP-negative cells [i.e., A L P ( - ) ] . The cells that contain intermediate levels of ALP [i.e., ALP(+)] that take longer than 10 minutes but shorter than 30
T. Matsuyama et al.: Human Bone Cell Subpopulations
279
A 10 8
I I
C
I
1
12 r
6 ~
B
i
1,25 Vitamin O Treated
10 ca
4
o Q.
~
o
(o
u
4
o I--
2
N
8 6 4 2
0 0
10
20
30
40
o
Incubation Time (rain)
Fig. 3. Histogram showing the distribution of subpopulations of human bone cells. The data from Figure 2 were replotted to show the percent of total cells that become positive for ALP in each time interval for control and 1,25(OH)2D3-treated cells.
minutes to stain were determined by subtracting the mean value of the A L P ( + +) from the 30 minute value or total positive cells.
40
60
80
100
Intensity of ALP stain
t20
140
160
180
(Arbitrary unit)
Fig. 4. Histogram showing the distribution of control ceils as a function of staining intensity measured with a laser densitometer. Cells were stained for 30 minutes at 4~ The range of optical densities was divided into intervals of five units (arbitrary units) and the percent of cells with densities in each interval was plotted. The distribution was divided into three groups of cells labeled A-C according to the Chi-square goodness of fit test. Density measurements were carried out on a total of 150 cells.
Analysis of Cell Subpopulations by Laser Densitomitry The method of using incubation time to separate ALP( + + ) and A L P ( + ) subpopulations has an element of subjectivity (as shown below). To test for the presence of subpopulations with a less subjective (but more time-consuming) method, a laser densitometer was used to measure the optical density in cells with and without 1,25(OH)2D 3 treatment and stained for ALP (30 minutes at 4~ The resulting distribution of ALP staining intensities was significantly different from a normal distribution (P < 0.001) and could be divided into 3-4 subpopulations labeled A-D (Figs. 4 and 5) (note that Figs. 4 and 5 plot % cells versus increasing ALP intensity which is the opposite of Figure 3 where increasing time represents decreasing ALP activity). Prior to measuring the optimal density, the cells were identified as ALP positive or ALP negative, i.e., A L P ( - ) . This identification showed that A L P ( - ) corresponded to subpopulation A which had a normal distribution of very low intensities and that ALP-positive cells corresponded to subpopulations B-D. Subpopulations B and C had normal distributions of intensities and probably correspond with A L P + and A L P + + , respectively. In the 1,25(OH)2D3-treated cells, there was an additional group D that may have been part of C, but formed a skewed distribution when combined. This analy-
20
o
~o ~/I
A
B
C
]
D
I
I
I
o CL =
I,-
0 ~
0
20
40
60
80
100
Intensity of ALP stain
120
140
160
180
{Arbitrary unit)
Fig. 5. Histogram showing the distribution of 1,25(OH)2D 3treated cells as a function of staining intensity measured with a laser densitometer. The percent of cells within each five unit (arbitrary units) interval was plotted and subpopulations were calculated, as in Figure 4. The arbitrary units of Figures 4 and 5 are similar but not identical. Similar data were obtained in a second experiment.
sis of cell ALP levels showing evidence of subpopulations of bone cells confirms the results of the more subjective method.
Evaluation of the Method To test the potential application of the timed stain-
280
T. Matsuyama et al.: Human Bone Cell Subpopulations
Table 1. Reproducibility of the cytochemical staining procedure Intraassay Coefficient Experiment
n
Exp. 1
4
Exp. 2
6
Exp. 3
6
Interassay
% ALP-positive cells 8.33, 8.93, 8.11, 9.43 8.01, 7.26, 7.80, 8.26, 8.12, 8.55 7.00, 7.26, 6.78, 6.88, 7.21, 8.07
Mean
SD
of variation
8.70
0.52
5.94
8.00
0.40
5.13
7.20
0.42
5.89
7.97
0.61
7.69
Normal human bone cell cultures were prepared and maintained as described in Methods. ALP-positive cells were stained at 4~ for 10 minutes as described in Methods. Results were presented as percentage of cells that were stained positively for ALP.
ing method for determining subpopulations, we have investigated the reproducibility and precision of this approach. With the same investigator, we found that the intraassay variation was approximately 5% and the interassay variation was approximately 7% (Table 1). With multiple investigators in a blind study (i.e., none of the investigators knew which group was treated and which was the control group), we found that interobserver variation was estimated to be 24.5% (Table 2). It should be noted that each investigator showed that 1,25(OH)ED 3 produced changes in subpopulations despite some differences in the relative proportions of the subpopulations.
With this cytochemical method, we found that 1,25(OH)2D 3 increased the amount of both subpopulations of ALP-positive cells [i.e., A L P ( + ) and A L P ( + +)]; TGF-beta appeared to increase only the intermediate form [i.e., ALP(+)] without affecting the strong positive form [i.e., A L P ( + +)]. In contrast, fluoride increased the proportion of the strong form but not the intermediate form. IGF-1 had no apparent effects on either subpopulation of ALP containing human bone cells. These findings indicate an interesting possibility that different bone cell effectors might have differential effects on different subpopulation human bone cells. In the same experiment, we found that 1,25(OH)ED3 and TGFbeta, but none of the other tested agents, significantly increased the specific activity of cellular ALP in these human bone cells. We have also evaluated the correlations between the specific activity of the cellular ALP with the relative proportions of each of the subpopulations of bone cells (Fig. 6). We found that there was a strong positive correlation (r = 0.927, P < 0.05) between the percent of total ALP cells with the specific activity of cellular ALP (Fig. 6A). A weaker positive correlation (r = 0.880, P < 0.05) was detected between the specific activity of cellular ALP and the percent of strong ALP-positive cells (Fig. 6B). However, we did not find any significant correlation (r = 0.206) between the specific activity and the relative proportion of the intermediate form.
I::
Effects of Bone Cell Effectors on each Subpopulation of Human Bone Cells
O.
E
100
Z)
E
Because we can identify three different subpopulations of human bone cells by this cytochemical method [ A L P ( - ), ALP( + ), ALP( + + )], we have decided to evaluate the effects of the following bone cell effectors: (1) 1,25(OH)ED 3 (this secosteroid has been shown to stimulate the bone cell ALP activity) [14-16]; (2) TGF-beta [this agent has recently been shown to stimulate human bone cell ALP activity (unpublished observation)]; (3) fluoride (it has been shown to directly act on osteoblasts to stimulate their proliferation and increase their cellular ALP content) [17, 18]; and (4) IGF-I [this growth factor stimulates bone cell proliferation without affecting the cellular ALP content (unpublished observation)]. For comparison, we have also determined the effects of these factors on the specific activity of cellular ALP (Table 3).
y
A, TotaJ
9 B. Strong A L P Cells :----
C. Intermediate ALP~q/
50 0. _1 ,< o (3. (0
r -
.942
p < .005
r = ~ p
