ANALYTICAL
204,
BIOCHEMISTRY
59-64
(19%)
A Calorimetric Method for Detection of Specific Ligand Binding Hubert
Wang
and Suzanne
Cellular
Biochemistry
Received
October
Research
K. Becknerl and Development,
Life Technologies,
Inc.
The binding of ligands to specific cell surface receptors is rapid, specific, saturable, and reversible. Such interactions have been traditionally examined using a radioreceptor assay which employs an iodinated ligand. Evaluation of such data by Scatchard analysis (1) permits the determination of the number of receptor sites/ cell as well as the affinity of the ligand for the receptor. Radioligand binding is very sensitive, enabling detection of fmoles of bound ligand, but is complicated by the need to use radiolabeled ligand, which may be inactivated during the radiolabeling process. In many instances, where the ligand is slightly hydrophobic, nonspecific binding can be significant and can make data evaluation difficult. Such difficulties are compounded when receptor numbers are low. Recently flow cytometry has been used to quantitate (2,3) and enrich (4-6) growth factor receptor positive cells. While such meth-
’ Present town, MD
address: 208’74.
Maryland
14, 1991
The binding of IL-2 and IL-3 to the factor-dependent cell lines CTB6 and 32D, respectively, was determined using biotinylated ligand detected by the addition of a streptavidinlalkaline phosphatase conjugate and amplified with a phosphatase amplification system. Binding of both ligands was detectable after incubation with as little as 20 fmol of ligand and could be inhibited with a lo-fold molar excess of nonbiotinylated ligand. No binding was observed when biotinylated ligand was incubated with a receptor negative cell line (PC-12) and IL-2 was unable to compete with biotinylated IL-3 binding to 32D cells, further demonstrating specificity. These studies indicate that biotinylated ligands can be used as a nonradioactive method to detect specific, high-affinity cell surface receptors. TV 1992 Academic Press,
Inc.. Gaithersburg,
Cellco,
Inc.,
0003.x97/92 $5.00 Copyright :C 1992 by Academic Press, All rights of reproduction in any form
12321
Middlebrook
Road,
German-
odology is rapid, quantitative, and sensitive, the use of expensive equipment and trained operators is required. Lymphokines are a diverse class of growth factors which mediate growth and differentiation of hematopoietic cells. IL-2* is a 14.6-kDa peptide secreted in vitro by mitogen- or antigen-activated lymphocytes that stimulates the proliferation and differentiation of T-lymphocytes (7), whereas IL-3 is a 2%kDa colony stimulating factor that stimulates the growth and differentiation of pluripotent stem cells to all major lineages (8-10). The biologic responses to both are mediated by interaction with specific cell surface receptors (11-14). Unlike nonhematopoietic cells, which can have up to one million growth factor receptors/cell (15), IL-2 and IL-3 responsive cells generally have less than 50,000 receptors/cells. Although radioligand assays have been utilized to study IL-2 and IL-3 binding, such studies have been difficult due to high nonspecific binding, the availability of high specific activity radioligands and the low number of high-affinity receptors. For example, CTB6 cells have been shown to have 58,000 high-affinity (K, = 130 PM) and 170,000 low-affinity (K, = 28 nM) receptors/cell (16). Several murine IL+dependent lines express even lower numbers of high-affinity (K, = 173 PM) receptors varying between 450 and 2500 sites/cell (17-19). We have developed a rapid and sensitive nonradioactive assay to detect IL-2 and IL-3 receptor binding utilizing biotinylated ligand. The presence of biotin was detected by binding to streptavidin. Conjugation of streptavidin to alkaline phosphatase allows amplification of the signal based on the dephosphorylation of NADPH and subsequent cycling of NADH to NAD through a redox reaction system generating a colorimetric signal (20,21). Utilizing this methodology, we demonstrate that the binding of IL-2 and IL-3 to the receptor
’ Abbreviations used: IL-2, interleukin-2; IL-3, interleukin-3; bovine serum albumin; FBS, fetal bovine serum; ELISA, linked immunosorbent assay; FITC, fluorescien isothiocyanate.
BSA, enzyme-
59 Inc. reserved.
60
WANG
AND BECKNER
positive cell lines CTB6 and 32D, respectively, is specific and reversible. The utility of this methodology for rapid screening of growth factor receptor positive cells is discussed. MATERIALS
AND
METHODS
Materials. All reagents with the exception of BSA were obtained from Life Technologies (GIBCO BRL); BSA was purchased from Sigma Chemical Co. Human II-2 and murine IL-3 were biotinylated as described by Taki et al. (3). The biological activity of the biotinylated IL-2 and IL-3 was determined by measuring the incorporation of thymidine into the murine factor-dependent lines CTB6 and 32D, respectively, as described (22,23). For both ligands, the activity of biotinylated ligand was >85% that of the native ligand, indicating that there was no adverse effect of biotinylation. The cell lines CTB6 and 32D were cultured to a density of 5 X lo5 cell/ml in RPM1 1640 plus 10% fetal bovine serum (FBS) supplemented with 10 U/ml human IL-2 (3100 U/pg) andmurine IL-3 (60,500 U/pug), respectively. Calorimetric assay method. The cells were harvested from culture, washed free of lymphokine and FBS by aseptically bubbling CO, through the media (24), washed twice with RPM1 1640, and resuspended in RPM1 to a density of lo6 cells/ml. One hundred microliters of cell suspension ( lo5 cells) was added to the wells of a 96-well plate and centrifuged for 5 min at 2000g. Following centrifugation, 100 ~1 of a 1.25% glutaraldehyde solution was added to the wells to fix the cells (which normally grow in suspension) and the plates were incubated at room temperature for 30 min. The effect of glutaraldehyde on ligand binding is unknown. Attempts to attach the cells with fibronectin or vitronectin were not successful since adherence is variable when using these attachment factors. The fixative was aspirated and the wells were washed with 200 ~1 of 50 mM Tris-HCl, pH 7.5, and 150 mM NaCl (Buffer A) and then blocked for 30 min with 200 ~1 of 50 mM Tris-HCl, 150 mM NaCl plus 1% BSA and 0.05% NaN, (Buffer B). Dilutions of biotinylated ligand were made in Buffer A containing 0.1% BSA. The blocker was aspirated from the wells and the indicated concentrations of biotinylated/nonbiotinylated ligand added to triplicate wells in a final volume of 100 ~1. The plates were incubated for 1 h at room temperature, the time required for binding to reach equilibrium (14). Post incubation, all wells were washed four times with 200 ~1 Buffer B and incubated with 50 ~1 streptavidin/alkaline phosphatase for 45 min at room temperature. Wells were washed four times with Buffer B and the signal amplified by the ELISA amplification system (Life Technologies), according to the instruction manual. The substrate in this system,
NADPH, is dephosphorylated by bound alkaline phosphatase to NADH. The NADH then activates a secondary enzyme system consisting of a redox cycle driven by diaphorase (which reduces a tetrazolium salt to an intensely colored formazan dye) and alcohol dehydrogenase (which oxidizes ethanol to acetaldehyde). Each molecule of NADH takes part in numerous cycles of the second reaction and color development is linear with time and the concentration of streptavidinlalkaline phosphatase (20,21). Absorbance was measured at 490 nm after 30 min (A,,,). Flow cytometry. Flow cytometry analysis of CTB6 and 32D cells was performed following incubation with biotinylated IL-2 and IL-3, respectively. Cells were counted and brought to a density of lo7 cells/ml, washed as above, and washed twice with phosphate-buffered saline containing 0.5% BSA and 0.5% NaN, (Buffer C). Following wash steps, 100 ~1 of cell suspensions was transferred to 12 x 75 glass test tubes and incubated with the indicated concentrations of factor; control tubes received no factor. After a l-h incubation at 4”C, cells were washed twice with Buffer C and 1 pg of FITCstreptavidin conjugate was added to each tube. Following 30 min at 4”C, cells were again washed twice with Buffer C and then resuspended to 1 ml in the same buffer. The cells were analyzed using a Coulter Epics Profile II flow cytometer. At least lo5 cells were counted from each tube. RESULTS
AND
DISCUSSION
To determine if the binding of biotinylated ligand was concentration dependent, cells were incubated with increasing concentrations of the respective ligand as described. The binding of biotinylated IL-2 or biotinylated IL-3 to the cell lines CTB6 or 32D, respectively, was detectable to the fmole range. A,,, could be measured when lo5 cells were incubated with ligand, over a concentration range of 20-700 fmol/lOO ~1 (Fig. 1). Binding of IL-2 and IL-3 was not saturable at concentrations of 684 and 358 fmol/lOO ~1, respectively. Higher concentrations were not attempted due to the expense of the reagents. With both IL-2 and IL-3, A,,, increased with increasing concentrations of biotinylated ligand. HOWever, the signal generated by IL-3 was consistently lower than that generated by equivalent molar concentrations of IL-2. This could be explained by differences in receptor number or differential biotinylation of the ligands; a greater number of conjugated biotins would allow the binding of more streptavidin/alkaline phosphatase thereby generating a stronger signal. Due to the expense of these reagents, quantitation of the number of biotin molecules/factor was not possible. The binding of biotinylated IL-2 and IL-3 to their respective cell lines was also measured by flow cytometry (Fig. 2). No shift in fluorescence was observed when
A COLORIMETRIC
2&l
300
400
METHOD
6cKr
BiotinyiatedLigand(fmol)
0.4,
B
/
Eliotinylated Ligand (fmd) FIG. 1. Binding of biotinvlated IL-2 and IL-3 to CTB6 and 32D cells. CTB6 (A) and32D (B) cells (105) were fixed and incubated with biotinylated IL-2 (A) or IL-3 (B), as described. Data are expressed as A,,, (signal generation) vs fmol biotinylated ligandllO0 ~1 in the incubation reaction and represent the average f SD of triplicate determinations representative of three separate experiments.
CTB6 cells were incubated with 685 fmol biotinylated IL-2. Significant signal was detected when cells were incubated with 6850 fmol biotinylated IL-2 and fluorescence was clearly significant with 68,500 fmol. These data support previous studies that demonstrated IL-2 receptor positive cells by flow cytometry following incubation with biotinylated (3) or fluoresceinated (2) IL-2. However, IL-3 receptor positive cells could not be detected even when cells were incubated with 35,700 fmol biotinylated IL-3 (Fig. 2). While this may reflect the low number of IL-3 receptors on 32D cells or a lower level of ligand biotinylation, it is clear that, for the detection of both IL-2 and IL-3 receptor positive cells, the colorimetric binding assay is at least two orders of magnitude
FOR
LIGAND
BINDING
61
more sensitive than flow cytometry. Although both the calorimetric method and flow cytometry involve biotinstreptavidin methodology, it is not surprising that the calorimetric method is more sensitive, since there is significant amplification of the initial signal (dephosphorylation). Each molecule of NADH takes part in many cycles of the second reaction, so the final signal is not 1:l as is flow cytometry. The binding of both biotinylated IL-2 and IL-3 was inhibited by nonbiotinylated ligand, indicating specificity of ligand binding (Fig. 3). CTB6 and 32D cells were incubated with nonsaturating concentrations of biotinylated IL-2 (68 fmol) and IL-3 (178 fmol), respectively, in the presence of l- to IO-fold excess of nonbiotinylated ligand (Fig. 3). For both ligands a 2-fold excess of nonbiotinylated ligand (136 and 356 fmol, respectively) resulted in greater than 50% inhibition of absorbance at 490. While a lo-fold excess of nonbiotinylated IL-2 (680 fmol/IOO ~1) completely inhibited IL-2 binding, IL-3 binding was only inhibited by 80%. Higher concentrations of nonbiotinylated IL-3 were not tested. These results are consistent with specific and reversible receptor ligand interactions and indicate that biotinylated IL-2 and biotinylated IL-3 are binding to IL-2 and IL-3 receptors, respectively. While nonbiotinylated ligand specifically inhibited signal generation, attempts to correct for nonspecific binding and to determine saturation of ligand binding were not possible. Although the signal generated with any concentration of biotinylated ligand could be competed with excess nonbiotinylated ligand, higher concentrations of biotinylated ligand required much higher concentrations of nonbiotinylated ligand to fully compete. For example, although lo-fold excess of nonbiotinylated IL-2 completely inhibited signal generation with 68 fmol of biotinylated IL-2, a IO-fold excess of nonbiotinylated IL-2 only inhibited signal generation with 340 fmol of biotinylated IL-2 by 50%. Inhibition of signal generation by higher concentrations of biotinylated ligand by a lo-fold excess of nonbiotinylated ligand was even less. Therefore, the cost of the ligands became prohibitive to determining actual saturation concentrations and nonspecific binding could not be determined over the full range of biotinylated ligand concentrations. Due to the nature of this assay, it is not possible to determine receptor number or binding constants. The data in Fig. 3 suggest that only a small percentage of the biotinylated ligand in the reaction mix actually binds to the cells. For example, incubation with 700 fmol of biotinylated IL-2 is not saturating (Fig. 1) while in Fig. 3, 680 fmol of nonbiotinylated IL-2 completely inhibits signal generation with 68 fmol of biotinylated IL-2. Attempts to determine the amount of unbound biotinylated ligand remaining in the reaction mix after the l-h incubation by using this in a second incubation support
62
WANG
AND
BECKNER
FIG. 2. Flow cytometry analysis of CTB6 and 32D cells following incubation with biotinylated with 0 (A), 685 (B), 6850 (C), 68,500 (D) fmol/lOO ~1 biotinylated IL-2 and 0 (E), 35,700 (F) fmol number of cells counted at a given level of fluorescence. At least lO’/cell were counted/sample.
the conclusion that only a small percentage of the biotinylated ligand actually binds to the cells. This is similar to what is observed in radioligand studies, where only a small percentage of the total counts added binds to cells at saturation. However, unlike radioactive ligands, the specific activity of the biotinylated ligand
ligand. Cells (106) were incubated IL-3, as described. Data are plotted
for 1 h as the
cannot be determined. Attempts to quantitate color generation with biotinylated ligand by comparison to a standard curve of alkaline phosphatase in the amplification reaction were not successful, as the amplification reaction cannot be controlled to that degree of sensitivity. A further complication is the fact that the initial
A COLORIMETRIC
METHOD
signal is amplified, and although the amplification exhibits linear kinetics, it is impossible to determine how many times NADH cycles through the redox cycle. Further evidence that signal generation is the result of a specific receptor ligand interaction was the inability of nonbiotinylated IL-2, at a 19-fold molar excess, to successfully compete and inhibit the signal generation by biotinylated IL-3 (Table 1). Additionally, when the binding of biotinylated IL-2 to PC12 cells (a cell line derived from a transplantable rat adrenal pheochromocytoma not known to express IL-2 receptors) was tested, no significant signal was observed (data not shown). These data together with experiments in Fig. 3
110,
A
FoldCompetitor
B
1
10 0
1
012
FOR
LIGAND
63
BINDING TABLE
1
Effect of Nonbiotinylated IL-3
Binding
IL-2 on Biotinylated to 32D
Cells
Ligand 180 fmol 180 fmol
bIL-3 bIL-3
A 490
+ 3400
fmol
0.150 0.142
IL-2
i 0.013 i 0.011
Note. 32D cells were incubated as described with 180 fmol biotinylated IL-3 (bIL-3) in the absence or presence of 3400 fmol nonbiotinylated IL-2. Data represent the average of triplicate determinations i SD.
demonstrate the specificity of the interaction of the biotinylated ligands with their receptor sites. As demonstrated by the current study, the specificity and sensitivity of this assay make it a useful method of screening for the presence of specific receptor sites and this assay has several advantages over possible alternatives. Cell surface receptors for many ligands are altered due to a variety of conditions including differentiation, transformation, development, and carcinogenesis (25,26). Although it is not possible to evaluate this binding data by Scatchard analysis to determine receptor number and binding affinity, this nonradioactive binding assay permits a far more sensitive method for detection of cell surface receptors than does flow cytometry. The use of amplification technology allows the detection of receptors on cells expressing low receptor numbers such as hematopoietic cells which may not be detectable in systems such as flow cytometry, where a 1:l relationship exists between ligand binding and signal intensity (Figs. 1 and 2). Given the extreme sensitivity of the method and the linearity of signal generation, it is possible to monitor relatively small quantitative changes in receptor levels. Additionally, by using a biochemically active ligand rather than an antibody or radiolabeled ligand, only active, physiologically relevant receptor forms are identified. Easy and rapid screening of large sample numbers under a variety of binding conditions is possible due to the microtiter format of the assay. These properties suggest that this calorimetric methodology may have wide application to rapid, quantitative screening of growth factor receptors under a variety of physiologically relevant conditions. ACKNOWLEDGMENTS
3
4
5
6
7
6
970
FoldCompetitor
FIG. 3. Competition of biotinylated ligand binding with nonbiotinylated ligand. CTB6 (A) and 32D (B) cells (105) were incubated with (A) biotinylated IL-2 (68 fmol/lOO ~1) or (B) IL-3 (178 fmol/lOO ~1) as described in the absence or presence of the indicated fold molar excess of nonbiotinylated ligand. Data are expressed as percentage control A,,, (biotinylated ligand alone) and represent the average f SD of triplicate determinations representative of three separate experiments.
The tometry Horuk
authors are grateful to Wes Russ for performing studies and to Drs. Bill Farrar, Vie Rebois, for critical review of this manuscript.
the flow cyand Richard
REFERENCES 1.
Munson,
2.
Harel-Bellen, A., Mishal, Z., Willette-Brown, J., and Farrar, W. L. (1989) J. Zmmunol. Methods 119, 127-133. Taki, S., Shimamura, T., Abe, M., Shirai, T., and Takahara, Y. (1989) J. Zmmunol. Methods 122, 33-41.
3.
P. J., and
Rodbard,
D. (1983)
Science
220,
979-981.
64
WANG
4. Newman, 5. Newman,
W., et al. (1989)
W., Beall, Z. I., and Bertolini, 16,366.
6.
J. Cell Phys.
L. D., Levine, D. R. (1989)
AND
141, 170-178.
M. A., Cone, J. L., Randhawa, J. Biol. Chem. 264, 16,259-
Foxwell, B. M. J., Taylor, D., Greiner, B., Mihatsch, vieri, V., and Ryffel, B. (1988) J. Zmmunol. Methods
M. J., Oli-
9. 10. 11. 12. 13.
Smith,
K. A. (1980)
Zmmunol.
Rev.
Mitjavila, Cantrell,
J. W. (1986)
Annu.
Reu. Zmmunol.
4, 205-230.
M. T., et al. (1989) J. Cell Phys. 138, D. A., and Smith, K. A. (1984) Science
16.
Birchenall-Sparks, M. C., Farrar, W. L., Rennick, D., Killian, P. L., and Ruscetti, F. W. (1986) Science 233, 455-458. Nicola, N. A., and Peterson, L. (1986) J. Biol. Chem. 261, 12,384-12,389.
17.
166,
172.
G., and Cohen,
18.
Park, L. S., Friend, 205-210.
19.
May, mun.
20. 21. 22.
617-624.
W. S., and
Ihle,
S. A. (1979)
Reu. Biochem.
S. (1986)
J. Biol.
Chem.
261,
J. N. (1986)
Biochem.
Biophys.
Res.
Com-
135,870-879.
S., Ferm,
M. W., Ou,
A.,
and
Methods W., and
Self,
C. H.
(1985)
J. Zm-
K. A. (1978)
J. Zm-
76,389-393. Smith,
12,2027-2035.
23.
Ihle, J. N., Keller, R. A., and Morse
Itoh, N., Yonehara, K., Ishii, A., Yahara,
24.
Michiel, D. F., Garcia, G. G., Evans, G. A., and (1991) Cytokine 3, 428-438. Messing, E. M. (1990) Cancer Res. 50, 2530-2537. Weidner, U., Peter, S., Strohmeyer, T., Hussnatter, R., and Sies, H. (1990) Cancer Res. 50, 4505-4509.
S., Scheurs, J., Gorman, I., Arai, K., and Miyajima,
D. M., MaruYama, A. (1990) Science
Robb,
160,
R. J., Greene, 1126-1133.
W. C., and Rusk,
C. M. (1984)
J. Exp.
Med.
25. 26.
48,
D., and Gillis,
Stanley, C. J., Johannsson, munol. Methods 83, 89-95. Self, C. H. (1985) J. Zmmunol. Gillis, munol.
Annu.
Park, L. S., Friend, D., Price, V., Anderson, D., Singer, J., Prickett, K. S., and Urdal, D. L. (1989) J. Biol. Chem. 264,5420-5427.
247,324-327. 14.
Carpenter, 193-216.
51, 337-357.
Ihle, J. N., Keller, J., Oroszlan, S., Henderson, L., Copeland, T., Goldwasser, E., Schrader, J., Palaszynski, E., Dy, M., and Lebel, B. (1983) J. Zmmunol. 131, 282-287. Schrader,
15.
113, 221-
229. I. 8.
BECKNER
J., Greenberger, III, H. C. (1982)
J. S., Henderson, J. Zmmunol. 129,
L., Yetter, 1377-1391.
Farrar,
W.
Ackermann,
L.
204,
BIOCHEMISTRY
59-64
(19%)
A Calorimetric Method for Detection of Specific Ligand Binding Hubert
Wang
and Suzanne
Cellular
Biochemistry
Received
October
Research
K. Becknerl and Development,
Life Technologies,
Inc.
The binding of ligands to specific cell surface receptors is rapid, specific, saturable, and reversible. Such interactions have been traditionally examined using a radioreceptor assay which employs an iodinated ligand. Evaluation of such data by Scatchard analysis (1) permits the determination of the number of receptor sites/ cell as well as the affinity of the ligand for the receptor. Radioligand binding is very sensitive, enabling detection of fmoles of bound ligand, but is complicated by the need to use radiolabeled ligand, which may be inactivated during the radiolabeling process. In many instances, where the ligand is slightly hydrophobic, nonspecific binding can be significant and can make data evaluation difficult. Such difficulties are compounded when receptor numbers are low. Recently flow cytometry has been used to quantitate (2,3) and enrich (4-6) growth factor receptor positive cells. While such meth-
’ Present town, MD
address: 208’74.
Maryland
14, 1991
The binding of IL-2 and IL-3 to the factor-dependent cell lines CTB6 and 32D, respectively, was determined using biotinylated ligand detected by the addition of a streptavidinlalkaline phosphatase conjugate and amplified with a phosphatase amplification system. Binding of both ligands was detectable after incubation with as little as 20 fmol of ligand and could be inhibited with a lo-fold molar excess of nonbiotinylated ligand. No binding was observed when biotinylated ligand was incubated with a receptor negative cell line (PC-12) and IL-2 was unable to compete with biotinylated IL-3 binding to 32D cells, further demonstrating specificity. These studies indicate that biotinylated ligands can be used as a nonradioactive method to detect specific, high-affinity cell surface receptors. TV 1992 Academic Press,
Inc.. Gaithersburg,
Cellco,
Inc.,
0003.x97/92 $5.00 Copyright :C 1992 by Academic Press, All rights of reproduction in any form
12321
Middlebrook
Road,
German-
odology is rapid, quantitative, and sensitive, the use of expensive equipment and trained operators is required. Lymphokines are a diverse class of growth factors which mediate growth and differentiation of hematopoietic cells. IL-2* is a 14.6-kDa peptide secreted in vitro by mitogen- or antigen-activated lymphocytes that stimulates the proliferation and differentiation of T-lymphocytes (7), whereas IL-3 is a 2%kDa colony stimulating factor that stimulates the growth and differentiation of pluripotent stem cells to all major lineages (8-10). The biologic responses to both are mediated by interaction with specific cell surface receptors (11-14). Unlike nonhematopoietic cells, which can have up to one million growth factor receptors/cell (15), IL-2 and IL-3 responsive cells generally have less than 50,000 receptors/cells. Although radioligand assays have been utilized to study IL-2 and IL-3 binding, such studies have been difficult due to high nonspecific binding, the availability of high specific activity radioligands and the low number of high-affinity receptors. For example, CTB6 cells have been shown to have 58,000 high-affinity (K, = 130 PM) and 170,000 low-affinity (K, = 28 nM) receptors/cell (16). Several murine IL+dependent lines express even lower numbers of high-affinity (K, = 173 PM) receptors varying between 450 and 2500 sites/cell (17-19). We have developed a rapid and sensitive nonradioactive assay to detect IL-2 and IL-3 receptor binding utilizing biotinylated ligand. The presence of biotin was detected by binding to streptavidin. Conjugation of streptavidin to alkaline phosphatase allows amplification of the signal based on the dephosphorylation of NADPH and subsequent cycling of NADH to NAD through a redox reaction system generating a colorimetric signal (20,21). Utilizing this methodology, we demonstrate that the binding of IL-2 and IL-3 to the receptor
’ Abbreviations used: IL-2, interleukin-2; IL-3, interleukin-3; bovine serum albumin; FBS, fetal bovine serum; ELISA, linked immunosorbent assay; FITC, fluorescien isothiocyanate.
BSA, enzyme-
59 Inc. reserved.
60
WANG
AND BECKNER
positive cell lines CTB6 and 32D, respectively, is specific and reversible. The utility of this methodology for rapid screening of growth factor receptor positive cells is discussed. MATERIALS
AND
METHODS
Materials. All reagents with the exception of BSA were obtained from Life Technologies (GIBCO BRL); BSA was purchased from Sigma Chemical Co. Human II-2 and murine IL-3 were biotinylated as described by Taki et al. (3). The biological activity of the biotinylated IL-2 and IL-3 was determined by measuring the incorporation of thymidine into the murine factor-dependent lines CTB6 and 32D, respectively, as described (22,23). For both ligands, the activity of biotinylated ligand was >85% that of the native ligand, indicating that there was no adverse effect of biotinylation. The cell lines CTB6 and 32D were cultured to a density of 5 X lo5 cell/ml in RPM1 1640 plus 10% fetal bovine serum (FBS) supplemented with 10 U/ml human IL-2 (3100 U/pg) andmurine IL-3 (60,500 U/pug), respectively. Calorimetric assay method. The cells were harvested from culture, washed free of lymphokine and FBS by aseptically bubbling CO, through the media (24), washed twice with RPM1 1640, and resuspended in RPM1 to a density of lo6 cells/ml. One hundred microliters of cell suspension ( lo5 cells) was added to the wells of a 96-well plate and centrifuged for 5 min at 2000g. Following centrifugation, 100 ~1 of a 1.25% glutaraldehyde solution was added to the wells to fix the cells (which normally grow in suspension) and the plates were incubated at room temperature for 30 min. The effect of glutaraldehyde on ligand binding is unknown. Attempts to attach the cells with fibronectin or vitronectin were not successful since adherence is variable when using these attachment factors. The fixative was aspirated and the wells were washed with 200 ~1 of 50 mM Tris-HCl, pH 7.5, and 150 mM NaCl (Buffer A) and then blocked for 30 min with 200 ~1 of 50 mM Tris-HCl, 150 mM NaCl plus 1% BSA and 0.05% NaN, (Buffer B). Dilutions of biotinylated ligand were made in Buffer A containing 0.1% BSA. The blocker was aspirated from the wells and the indicated concentrations of biotinylated/nonbiotinylated ligand added to triplicate wells in a final volume of 100 ~1. The plates were incubated for 1 h at room temperature, the time required for binding to reach equilibrium (14). Post incubation, all wells were washed four times with 200 ~1 Buffer B and incubated with 50 ~1 streptavidin/alkaline phosphatase for 45 min at room temperature. Wells were washed four times with Buffer B and the signal amplified by the ELISA amplification system (Life Technologies), according to the instruction manual. The substrate in this system,
NADPH, is dephosphorylated by bound alkaline phosphatase to NADH. The NADH then activates a secondary enzyme system consisting of a redox cycle driven by diaphorase (which reduces a tetrazolium salt to an intensely colored formazan dye) and alcohol dehydrogenase (which oxidizes ethanol to acetaldehyde). Each molecule of NADH takes part in numerous cycles of the second reaction and color development is linear with time and the concentration of streptavidinlalkaline phosphatase (20,21). Absorbance was measured at 490 nm after 30 min (A,,,). Flow cytometry. Flow cytometry analysis of CTB6 and 32D cells was performed following incubation with biotinylated IL-2 and IL-3, respectively. Cells were counted and brought to a density of lo7 cells/ml, washed as above, and washed twice with phosphate-buffered saline containing 0.5% BSA and 0.5% NaN, (Buffer C). Following wash steps, 100 ~1 of cell suspensions was transferred to 12 x 75 glass test tubes and incubated with the indicated concentrations of factor; control tubes received no factor. After a l-h incubation at 4”C, cells were washed twice with Buffer C and 1 pg of FITCstreptavidin conjugate was added to each tube. Following 30 min at 4”C, cells were again washed twice with Buffer C and then resuspended to 1 ml in the same buffer. The cells were analyzed using a Coulter Epics Profile II flow cytometer. At least lo5 cells were counted from each tube. RESULTS
AND
DISCUSSION
To determine if the binding of biotinylated ligand was concentration dependent, cells were incubated with increasing concentrations of the respective ligand as described. The binding of biotinylated IL-2 or biotinylated IL-3 to the cell lines CTB6 or 32D, respectively, was detectable to the fmole range. A,,, could be measured when lo5 cells were incubated with ligand, over a concentration range of 20-700 fmol/lOO ~1 (Fig. 1). Binding of IL-2 and IL-3 was not saturable at concentrations of 684 and 358 fmol/lOO ~1, respectively. Higher concentrations were not attempted due to the expense of the reagents. With both IL-2 and IL-3, A,,, increased with increasing concentrations of biotinylated ligand. HOWever, the signal generated by IL-3 was consistently lower than that generated by equivalent molar concentrations of IL-2. This could be explained by differences in receptor number or differential biotinylation of the ligands; a greater number of conjugated biotins would allow the binding of more streptavidin/alkaline phosphatase thereby generating a stronger signal. Due to the expense of these reagents, quantitation of the number of biotin molecules/factor was not possible. The binding of biotinylated IL-2 and IL-3 to their respective cell lines was also measured by flow cytometry (Fig. 2). No shift in fluorescence was observed when
A COLORIMETRIC
2&l
300
400
METHOD
6cKr
BiotinyiatedLigand(fmol)
0.4,
B
/
Eliotinylated Ligand (fmd) FIG. 1. Binding of biotinvlated IL-2 and IL-3 to CTB6 and 32D cells. CTB6 (A) and32D (B) cells (105) were fixed and incubated with biotinylated IL-2 (A) or IL-3 (B), as described. Data are expressed as A,,, (signal generation) vs fmol biotinylated ligandllO0 ~1 in the incubation reaction and represent the average f SD of triplicate determinations representative of three separate experiments.
CTB6 cells were incubated with 685 fmol biotinylated IL-2. Significant signal was detected when cells were incubated with 6850 fmol biotinylated IL-2 and fluorescence was clearly significant with 68,500 fmol. These data support previous studies that demonstrated IL-2 receptor positive cells by flow cytometry following incubation with biotinylated (3) or fluoresceinated (2) IL-2. However, IL-3 receptor positive cells could not be detected even when cells were incubated with 35,700 fmol biotinylated IL-3 (Fig. 2). While this may reflect the low number of IL-3 receptors on 32D cells or a lower level of ligand biotinylation, it is clear that, for the detection of both IL-2 and IL-3 receptor positive cells, the colorimetric binding assay is at least two orders of magnitude
FOR
LIGAND
BINDING
61
more sensitive than flow cytometry. Although both the calorimetric method and flow cytometry involve biotinstreptavidin methodology, it is not surprising that the calorimetric method is more sensitive, since there is significant amplification of the initial signal (dephosphorylation). Each molecule of NADH takes part in many cycles of the second reaction, so the final signal is not 1:l as is flow cytometry. The binding of both biotinylated IL-2 and IL-3 was inhibited by nonbiotinylated ligand, indicating specificity of ligand binding (Fig. 3). CTB6 and 32D cells were incubated with nonsaturating concentrations of biotinylated IL-2 (68 fmol) and IL-3 (178 fmol), respectively, in the presence of l- to IO-fold excess of nonbiotinylated ligand (Fig. 3). For both ligands a 2-fold excess of nonbiotinylated ligand (136 and 356 fmol, respectively) resulted in greater than 50% inhibition of absorbance at 490. While a lo-fold excess of nonbiotinylated IL-2 (680 fmol/IOO ~1) completely inhibited IL-2 binding, IL-3 binding was only inhibited by 80%. Higher concentrations of nonbiotinylated IL-3 were not tested. These results are consistent with specific and reversible receptor ligand interactions and indicate that biotinylated IL-2 and biotinylated IL-3 are binding to IL-2 and IL-3 receptors, respectively. While nonbiotinylated ligand specifically inhibited signal generation, attempts to correct for nonspecific binding and to determine saturation of ligand binding were not possible. Although the signal generated with any concentration of biotinylated ligand could be competed with excess nonbiotinylated ligand, higher concentrations of biotinylated ligand required much higher concentrations of nonbiotinylated ligand to fully compete. For example, although lo-fold excess of nonbiotinylated IL-2 completely inhibited signal generation with 68 fmol of biotinylated IL-2, a IO-fold excess of nonbiotinylated IL-2 only inhibited signal generation with 340 fmol of biotinylated IL-2 by 50%. Inhibition of signal generation by higher concentrations of biotinylated ligand by a lo-fold excess of nonbiotinylated ligand was even less. Therefore, the cost of the ligands became prohibitive to determining actual saturation concentrations and nonspecific binding could not be determined over the full range of biotinylated ligand concentrations. Due to the nature of this assay, it is not possible to determine receptor number or binding constants. The data in Fig. 3 suggest that only a small percentage of the biotinylated ligand in the reaction mix actually binds to the cells. For example, incubation with 700 fmol of biotinylated IL-2 is not saturating (Fig. 1) while in Fig. 3, 680 fmol of nonbiotinylated IL-2 completely inhibits signal generation with 68 fmol of biotinylated IL-2. Attempts to determine the amount of unbound biotinylated ligand remaining in the reaction mix after the l-h incubation by using this in a second incubation support
62
WANG
AND
BECKNER
FIG. 2. Flow cytometry analysis of CTB6 and 32D cells following incubation with biotinylated with 0 (A), 685 (B), 6850 (C), 68,500 (D) fmol/lOO ~1 biotinylated IL-2 and 0 (E), 35,700 (F) fmol number of cells counted at a given level of fluorescence. At least lO’/cell were counted/sample.
the conclusion that only a small percentage of the biotinylated ligand actually binds to the cells. This is similar to what is observed in radioligand studies, where only a small percentage of the total counts added binds to cells at saturation. However, unlike radioactive ligands, the specific activity of the biotinylated ligand
ligand. Cells (106) were incubated IL-3, as described. Data are plotted
for 1 h as the
cannot be determined. Attempts to quantitate color generation with biotinylated ligand by comparison to a standard curve of alkaline phosphatase in the amplification reaction were not successful, as the amplification reaction cannot be controlled to that degree of sensitivity. A further complication is the fact that the initial
A COLORIMETRIC
METHOD
signal is amplified, and although the amplification exhibits linear kinetics, it is impossible to determine how many times NADH cycles through the redox cycle. Further evidence that signal generation is the result of a specific receptor ligand interaction was the inability of nonbiotinylated IL-2, at a 19-fold molar excess, to successfully compete and inhibit the signal generation by biotinylated IL-3 (Table 1). Additionally, when the binding of biotinylated IL-2 to PC12 cells (a cell line derived from a transplantable rat adrenal pheochromocytoma not known to express IL-2 receptors) was tested, no significant signal was observed (data not shown). These data together with experiments in Fig. 3
110,
A
FoldCompetitor
B
1
10 0
1
012
FOR
LIGAND
63
BINDING TABLE
1
Effect of Nonbiotinylated IL-3
Binding
IL-2 on Biotinylated to 32D
Cells
Ligand 180 fmol 180 fmol
bIL-3 bIL-3
A 490
+ 3400
fmol
0.150 0.142
IL-2
i 0.013 i 0.011
Note. 32D cells were incubated as described with 180 fmol biotinylated IL-3 (bIL-3) in the absence or presence of 3400 fmol nonbiotinylated IL-2. Data represent the average of triplicate determinations i SD.
demonstrate the specificity of the interaction of the biotinylated ligands with their receptor sites. As demonstrated by the current study, the specificity and sensitivity of this assay make it a useful method of screening for the presence of specific receptor sites and this assay has several advantages over possible alternatives. Cell surface receptors for many ligands are altered due to a variety of conditions including differentiation, transformation, development, and carcinogenesis (25,26). Although it is not possible to evaluate this binding data by Scatchard analysis to determine receptor number and binding affinity, this nonradioactive binding assay permits a far more sensitive method for detection of cell surface receptors than does flow cytometry. The use of amplification technology allows the detection of receptors on cells expressing low receptor numbers such as hematopoietic cells which may not be detectable in systems such as flow cytometry, where a 1:l relationship exists between ligand binding and signal intensity (Figs. 1 and 2). Given the extreme sensitivity of the method and the linearity of signal generation, it is possible to monitor relatively small quantitative changes in receptor levels. Additionally, by using a biochemically active ligand rather than an antibody or radiolabeled ligand, only active, physiologically relevant receptor forms are identified. Easy and rapid screening of large sample numbers under a variety of binding conditions is possible due to the microtiter format of the assay. These properties suggest that this calorimetric methodology may have wide application to rapid, quantitative screening of growth factor receptors under a variety of physiologically relevant conditions. ACKNOWLEDGMENTS
3
4
5
6
7
6
970
FoldCompetitor
FIG. 3. Competition of biotinylated ligand binding with nonbiotinylated ligand. CTB6 (A) and 32D (B) cells (105) were incubated with (A) biotinylated IL-2 (68 fmol/lOO ~1) or (B) IL-3 (178 fmol/lOO ~1) as described in the absence or presence of the indicated fold molar excess of nonbiotinylated ligand. Data are expressed as percentage control A,,, (biotinylated ligand alone) and represent the average f SD of triplicate determinations representative of three separate experiments.
The tometry Horuk
authors are grateful to Wes Russ for performing studies and to Drs. Bill Farrar, Vie Rebois, for critical review of this manuscript.
the flow cyand Richard
REFERENCES 1.
Munson,
2.
Harel-Bellen, A., Mishal, Z., Willette-Brown, J., and Farrar, W. L. (1989) J. Zmmunol. Methods 119, 127-133. Taki, S., Shimamura, T., Abe, M., Shirai, T., and Takahara, Y. (1989) J. Zmmunol. Methods 122, 33-41.
3.
P. J., and
Rodbard,
D. (1983)
Science
220,
979-981.
64
WANG
4. Newman, 5. Newman,
W., et al. (1989)
W., Beall, Z. I., and Bertolini, 16,366.
6.
J. Cell Phys.
L. D., Levine, D. R. (1989)
AND
141, 170-178.
M. A., Cone, J. L., Randhawa, J. Biol. Chem. 264, 16,259-
Foxwell, B. M. J., Taylor, D., Greiner, B., Mihatsch, vieri, V., and Ryffel, B. (1988) J. Zmmunol. Methods
M. J., Oli-
9. 10. 11. 12. 13.
Smith,
K. A. (1980)
Zmmunol.
Rev.
Mitjavila, Cantrell,
J. W. (1986)
Annu.
Reu. Zmmunol.
4, 205-230.
M. T., et al. (1989) J. Cell Phys. 138, D. A., and Smith, K. A. (1984) Science
16.
Birchenall-Sparks, M. C., Farrar, W. L., Rennick, D., Killian, P. L., and Ruscetti, F. W. (1986) Science 233, 455-458. Nicola, N. A., and Peterson, L. (1986) J. Biol. Chem. 261, 12,384-12,389.
17.
166,
172.
G., and Cohen,
18.
Park, L. S., Friend, 205-210.
19.
May, mun.
20. 21. 22.
617-624.
W. S., and
Ihle,
S. A. (1979)
Reu. Biochem.
S. (1986)
J. Biol.
Chem.
261,
J. N. (1986)
Biochem.
Biophys.
Res.
Com-
135,870-879.
S., Ferm,
M. W., Ou,
A.,
and
Methods W., and
Self,
C. H.
(1985)
J. Zm-
K. A. (1978)
J. Zm-
76,389-393. Smith,
12,2027-2035.
23.
Ihle, J. N., Keller, R. A., and Morse
Itoh, N., Yonehara, K., Ishii, A., Yahara,
24.
Michiel, D. F., Garcia, G. G., Evans, G. A., and (1991) Cytokine 3, 428-438. Messing, E. M. (1990) Cancer Res. 50, 2530-2537. Weidner, U., Peter, S., Strohmeyer, T., Hussnatter, R., and Sies, H. (1990) Cancer Res. 50, 4505-4509.
S., Scheurs, J., Gorman, I., Arai, K., and Miyajima,
D. M., MaruYama, A. (1990) Science
Robb,
160,
R. J., Greene, 1126-1133.
W. C., and Rusk,
C. M. (1984)
J. Exp.
Med.
25. 26.
48,
D., and Gillis,
Stanley, C. J., Johannsson, munol. Methods 83, 89-95. Self, C. H. (1985) J. Zmmunol. Gillis, munol.
Annu.
Park, L. S., Friend, D., Price, V., Anderson, D., Singer, J., Prickett, K. S., and Urdal, D. L. (1989) J. Biol. Chem. 264,5420-5427.
247,324-327. 14.
Carpenter, 193-216.
51, 337-357.
Ihle, J. N., Keller, J., Oroszlan, S., Henderson, L., Copeland, T., Goldwasser, E., Schrader, J., Palaszynski, E., Dy, M., and Lebel, B. (1983) J. Zmmunol. 131, 282-287. Schrader,
15.
113, 221-
229. I. 8.
BECKNER
J., Greenberger, III, H. C. (1982)
J. S., Henderson, J. Zmmunol. 129,
L., Yetter, 1377-1391.
Farrar,
W.
Ackermann,
L.
