- Email: [email protected]
A semi-automated micro-assay for H2O2 release by human blood monocytes and mouse peritoneal macrophages
Journal of Immunological Methods, 78 (1985) 323- 336
323
Elsevier JIM03459
A Semi-Automated Micro-Assay for H202 Release by Human Blood Monocytes and Mouse Peritoneal Macrophages Jon De la Harpe ] and Carl F. Nathan 2 Laboratory of Cellular Physiology and Immunology, The Rockefeller University, N e w York, U.S.A.
(Received 14 December 1984, accepted 31 December 1984)
H202 secreted by mononuclear phagocytes can be detected by monitoring the horseradish peroxidase-catalyzed oxidation of fluorescent scopoletin. This technique has been adapted to a semi-automated micro-scale with the aid of automatic fluorescence and absorbance micro-culture plate readers to measure H202 and protein, respectively, in the same culture wells. With these adaptations the assay can accurately and precisely detect as little as 0.1 nm01 H202 or 1 ~tg cell protein, permitting the calculation of specific secretion (nmol H 2 0 2 / m g cell protein) from as few as 2 x 104 human blood monocytes or mouse peritoneal macrophages. Cumulative H202 secretion in individual wells may be recorded non-destructively at frequent intervals for time course measurements. Less than 1 rain is required to record the fluorescence in all 96 wells of a micro-culture plate. The assay is highly reproducible, with standard deviations for tripficates typically less than 5-10% of the mean, and gives values in close agreement with those obtained in 10-fold larger samples by previous methods. Using this assay it is feasible to process 1000 samples per day, with order of magnitude savings in labor, cells, sera, media, cytokines, and reagents compared to earlier forms of the assay. The assay is useful in evaluating the cellular effects of cytokines and for assaying their activity in chromatographic fractions and hybridoma cultures. We are currently using the assay to monitor the administration of interferon-gamma to patients with neoplasia. Key words: reactive oxygen intermediates- 1t202 - human monocytes - mouse macrophages - automated microassay
Introduction A n i m p o r t a n t t o o l i n t h e s t u d y o f m o n o n u c l e a r p h a g o c y t e s is q u a n t i t a t i o n o f t h e i r secretion of reactive oxygen intermediates. These incompletely reduced forms of 1 Reprint requests should be addressed to: Dr. Jon De la Harpe, The Rockefeller University Box 280, 1230 York Ave., New York, NY 10021, U.S.A. 2 C.F. Nathan is a research awardee of the Irma T. Hirschl Trust and the Rita Allen Foundation. Abbreviations: PMA, phorbol myristate acetate; HPO, horseradish peroxidase; ROI, reactive oxygen intermediates; KRP, Krebs-Ringer phosphate buffer; KRPG, KRP containing 5.5 mM glucose; PBS, phosphate-buffered saline; DMSO, dimethyl sulfoxide; a-MEM, Eagle's minimum essential medium, alpha variant; FBS, fetal bovine serum; HuS, human serum; BSA, bovine serum albumin; SD, standard deviation. 0022-1759/85/$03.30 © 1985 Elsevier Science Publishers B.V. (Biomedical Division)
324 oxygen, including 02, H202, and OH', can play a major role in the anti-microbial, anti-tumor, and inflammatory functions of monocytes and macrophages. The capacity of mononuclear phagocytes to secrete reactive oxygen intermediates (ROI) holds additional interest because it is closely regulated by cytokines, both positively (interferon-gamma (Nathan et al., 1983)) and negatively (factors from malignant and some non-malignant cells (Szuro-Sudol and Nathan, 1982)). Enhancement of H202 secretory capacity in mononuclear phagocytes from human subjects (C.F. Nathan, C.R. Horowitz, J. De la Harpe, S. Vadhan-Raj and S. Krown, in preparation) and mice (Murray et al., 1985) given recombinant interferon-gamma further heightens interest in the methodology of such measurements. Thus, there is a pressing need for an assay that can quantify the specific activity of ROI release'(nmol product/mg cell protein) on large numbers of samples using low numbers of cells. Such an assay could be employed for monitoring in vivo administration of macrophage activating factors, testing fractions from chromatographic and other steps in the purification of regulatory cytokines, and screening hybridoma supernatants for the production of such cytokines or of antibodies against them. The present report describes a semi-automated, micro-scale modification of the assay based on the horseradish peroxidase (HPO)-catalyzed oxidation of fluorescent scopoletin by H202 as originally introduced by Andreae (1955) and adapted for neutrophils by Root et al. (1975) and for macrophages by Nathan and Root (1977) and Nathan (1981). With the use of automatic fluorescence and absorbance microculture plate readers, the modified assay provides precise, accurate, specific activities for H202 secretion by as few as 2 × 104 monocytes or macrophages per sample, with the ability to detect as little as 0.1 nmol H202 and 1 #g of cell protein. H202 readings are non-destructive and permit repetitive monitoring of the time-course of secretion in individual wells. The fluorescence of 96 samples can be read in less than a minute, so that it is feasible to process 1000 or more samples per day, with an order of magnitude saving in cells, sera, media, cytoldnes, and reagents compared to earlier forms of this assay.
Materials and Methods
Materials
Krebs-Ringer phosphate buffer (KRP) was 129 mM NaC1, 4.86 mM KC1, 0.54 mM CaC12, 1.22 mM MgSO4, 15.8 mM sodium phosphate, pH 7.35, 300-315 mosM. For human cells this was modified to contain 145 mM NaC1 and 5.7 mM sodium phosphate (KRP low phosphate). For the H202 assay solution this KRP buffer was supplemented with 5.5 mM glucose. Phosphate-buffered saline (PBS) was 0.9 mM CaC12, 0.5 mM MgC12, 2.7 mM KC1, 136,9 mM NaC1, 1.5 mM KH2PO4, and 15.2 mM Na2HPO 4, pH 7.35. Fresh KRPG and PBS solutions were prepared weekly and stored at 4°C. Scopoletin (Sigma) was prepared as a 1 mM solution in KRP by dissolution for 24 h at 37°C, sterile filtered, and stored at 4°C in the dark. Casein (Kodak, sodium caseinate, practical grade) was prepared as a 6% solution in 0.95% w / v NaC1,
325 autoclaved, and stored at 40C. Sodium m-periodate (Sigma) was prepared immediately before use as a 5 mM solution in 0.95% NaC1 and filtered through a 0.45 #m Millex filter (Millipore Corp., Bedford, MA). HPO was Sigma P8250 Type II, 150-200 purpurogallin U/mg. It was prepared at 600 purpurogallin U / m l and was stored at -20°C. The triggering agent 4fl-phorbol 12fl-myristate 13a-acetate (Sigma) was prepared as a 300/xg/ml solution in dimethylsulfoxide (DMSO) and stored at 80°C. The working solution was stored at - 20°C and discarded after 1 week, the 80°C stock was discarded after 1 year. Trypan blue was 0.4% in normal saline (Gibco, Grand Island, NY) and was diluted 2-fold in 0.95% NaC1 before use. Ficoll-Paque was from Pharmacia Fine Chemicals, Piscataway, NJ. Bovine serum albumin (BSA) was from Armour Pharmaceutical Co., Phoenix, AZ. The modified Lowry solution A (Lowry et al., 1951) was 4% Na2CO 3 in water. Lowry solution B was 0.1 g CuSO4 • 5H20, 0.2 g tri-sodium citrate. 2H20 in 20 ml H20, and the Folin and Ciocalteau phenol reagent (Accra-Lab., Bridgeport, NJ) was 100 g / l sodium tungstate, 150 g/1 lithium sulfate, 25 g/1 sodium molybdate dihydrate, 100 ml/1 hydrochloric acid, 50 ml/1 phosphoric acid, 2 ml/1 bromine water. Eagle's minimum essential medium, alpha modified (a-MEM) and RPMI were from KC Biologicals, Lenexa, KS. a-MEM was supplemented with 2.2 g/1 NaHCO 3, and brought to pH 7.35 and 300-315 mosM. RPMI was supplemented with 2 mM glutamine. Fetal bovine serum (FBS) was supplied by HyClone, Logan, UT. Human serum (HuS) was prepared as described in Nakagawara et al. (1981). Penicillinstreptomycin solution (5000 U / m l penicillin, 5000/~g/ml streptomycin in normal saline) was from Gibco. Human cells were cultured in 25% HuS, 100 U / m l penicillin, 100/~g/ml streptomycin, 2 mM glutamine in RPMI (RPMI 25% HuS). Mouse peritoneal cells were cultured in 10% FBS, 100 U / m l penicillin, 100 # g / m l streptomycin in a-MEM (a-MEM 10% FBS). -
-
Methods Human peripheral blood monocytes. Human monocytes were obtained from 10 ml of freshly drawn heparinized venous blood. The sample was diluted in an equal volume of sterile saline, layered over 15 ml of Ficoll-Paque in a 50 ml polypropylene conical tube and centrifuged for 20 min at 600 × g at 20°C. The mononuclear cells were aspirated from the liquid/liquid interface (6-7 ml), added to 7 ml RPMI in a 15 ml conical polypropylene tube, and centrifuged for 10 min at 350 × g at 4°C. The pellet was resuspended in 15 ml RPMI, centrifuged for 10 min at 200 × g at 4°C, and then resuspended in 1 ml RPMI 25% HuS. A small volume (10 ~tl) was diluted 10-fold into trypan blue for counting and viability estimation in a hemocytometer. The cell suspension was then diluted in RPMI 25% HuS to achieve the desired final cell density. Peritoneal macrophages. Cells were obtained from female Nelson-Collins strain (NCS) mice (20-30 g) supplied by the Rockefeller University Laboratory Animal Research Center. Mice were sacrified by cervical dislocation and the peritoneal cavity lavaged with 2 washes of 5-10 ml a-MEM. For resident cells, mice received no prior treatment. To obtain activated cells, the mice were injected intraperitoneally
326 with 1 ml of 5 m M N a I O 4 in isotonic saline (Tsunawaki and Nathan, 1984) 2 - 5 days prior to harvest (periodate-elicited cells), or with 1 ml 6% sterile sodium caseinate in isotonic saline 4 days before harvest (casein-elicited cells). The cells were harvested sterilely and kept on ice to minimize adherence. They were centrifuged (15 min, 200 x g, 4°C) and, if necessary, the erythrocytes were lysed by resuspending the pellet in 5 ml 0.2% saline for 30 s followed by the addition of 5 ml of 1.9% saline and 5 ml a - M E M 10% FBS. The cells were then pelleted again and resuspended in a few milliliters of a - M E M 10% FBS. Cell density and viability were determined as described above. Cell cultures. Cells were cultured in 96-well flat-bottomed plates (Coming Glass Works, Coming, N Y or Flow Laboratories, McLean, VA), typically at 1 x 105 per well in 100/zl of culture medium. The peripheral rows of wells in the plate were not used for cell culture but received 100 /~l sterile distilled water per well. Cell-free controls received 100/~l of the serum-containing medium used to culture the cells. To ensure adherence, cells were cultured for at least 2 h before washing. This initial culture period could be extended to as long as 5 days. When human cells were cultured more than a few hours the culture medium was changed after 2 h. The medium in the well was aspirated through a 21-gauge needle and replaced with 100 /zl of fresh medium. HeO2 assay. The cultures were washed 3 times with PBS at 37°C as follows: the plate was inverted over a basin and the medium flicked out with a vigorous wrist action. The plate was then held vertical and gently lowered into a 1 liter beaker of PBS at 37°C. The PBS was removed by again flicking the plate over a basin. It is critical that no bubbles are trapped in the culture wells during this procedure, and the plate should be examined through the side of the beaker while immersed to ensure this. Wells which are incompletely washed yield very high values in the protein assay. After filling and emptying the wells with PBS 3 times, the outer surfaces of the plate were blotted dry on a paper towel and the assay mixture (prewarmed to 37°C) was dispensed into the wells (100 F1/well) using an Eppendorf repeater pipette (Brinkmann Instrument Co., Westbury, NY). The assay mixture was prepared immediately before use from stock solutions and consisted of 30 FM scopoletin, 1 m M NAN3, 1 purpurogallin u n i t / m l HPO, 100 n g / m l PMA in K R P G for mouse cells and in K R P G (low phosphate) for human cells. PMA is suspected to be a tumor promoter and care was taken to avoid skin contact with solutions containing PMA. Immediately after the addition of the assay mixture the plate was placed in a filter fluorometer (MicroFluor MR600, Dynatech Laboratories, Alexandria, VA) and the fluorescence, in arbitrary units, recorded for each well. The plate was then covered with its lid and transferred to an incubation chamber maintained at 37°C and continuously flushed with water-saturated air at 37°C. After 60 or 90 min the fluorescence in'each well was again recorded using the MicroFluor. To monitor the rate of release of H202 the fluorescence may be read at intervals as short as 5 rain. The MicroFluor requires about 30 s to record all the values for a 96-well plate. Cell protein assay. A series of protein standards (0-10 Fg BSA in 100 Fl) was prepared using BSA dissolved in the same batch of assay mix used for measuring
327 H202 release by the cells. The peripheral rows A 1-12 and H 1-12 of the micro-culture plate, which were left empty during the H202 assay, were used for the protein standards. The cell protein estimation was performed without removing the H202 assay mix from the wells. To ensure the disruption of the cells and the solubilization of the cell protein, we added 10 /~1 of 1.0 M NaOH to each well, including the BSA standards. The gap between the plate and its lid was then sealed by stretching a strip of Parafilm around the edge, and this assembly was incubated at 37°C with gentle agitation on a rotary shaker for 60 rain. To estimate cell protein, a modification of the method of Lowry et al. (1951) was used. Upon completion of the alkaline digestion, each well received 100 #1 of modified Lowry reagent C (12 ml 4% Na2CO 3 + 0.5 ml Lowry reagent B). After 10 min at room temperature the plate was placed on a vortexer (Dynatech minishaker) and agitated vigorously while 10/tl of Folin's reagent (diluted by adding 0.2 vols. of water) was added to each assay well using an Eppendorf repeater pipette. After a further 30 min at room temperature the absorbance at 690 nm was measured in each well using a filter absorbance reader (BioTek EL307, BioTek Instruments, Burlington, VT). A similar strategy for measuring cell protein in microculture wells has been described by Shopsis and Mackay, 1984. In all the assays described any bubbles in the wells were eliminated using a hypodermic needle before fluorescence or absorbance readings were taken.
Results We shall refer to the assay for H202 release performed in 24-well plates (Nathan, 1981) as the macro-assay, while the assay performed in 96-well plates, as described in this report, will be referred to as the micro-assay. We examined the reliability and sensitivity of measuring scopoletin oxidation in the micro-culture plate system. The Microfluor reader has a preset range of 0-4096 arbitrary fluorescence units. The measurement of scopoletin by the instrument used in this study is linear over the range of 0-3 nmol of scopoletin per well. The addition of 1.5 nmol of H202 to 100 ~1 of assay mix containing 3 nmol of scopoletin reduced the measured fluorescence from 2651 to 1359 arbitrary units. In the absence of cells, the measurement of scopoletin by fluorescence was not affected by the other components of the assay system (horseradish peroxidase, NaN 3 or PMA). We did not find it necessary to use the special non-fluorescent 96-well plates available from Dynatech. Conventional 96-well tissue-culture plates have acceptably low background fluorescence levels - in the region of 50 arbitrary units. We observed a gradual decrease in measured fluorescence in cell-free control wells, in the range of 100-400 arbitrary units, over the 60 or 90 min incubation for the standard assay. In order to determine the appropriate method to correct for this change when evaluating the experimental results, it was important to know whether this decrease was due to oxidation of the scopoletin in the control wells or to a change in the sensitivity of the measuring system. Using a Perkin-Elmer MPF-4A spectrofluorometer, we measured the fluorescence of pooled control samples re-
328
moved from the wells at the start and end of the incubation period, and established that this change is not due to a loss of fluorescence of the assay solution, and so not due to oxidation of the scopoletin. We believe that this change in measured fluorescence is due to 2 factors: (1) A change in the shape of the meniscus in the wells containing assay mix, resulting in a change in the effective light-path and in the refractive properties of the air-liquid interface. We examined the effect of the meniscus shape on the measured fluorescence of a given amount of scopoletin. Wells containing 1.5 nmol of scopoletin in volumes ranging from 50 to 300/~1 of K R P G register the same fluorescence in the MicroFluor (within 10%). However, when the volume is increased from 300 to 340 /~1, the recorded fluorescence increases by 80%. This is the volume range in which the liquid level reaches the rim of the well and the meniscus changes from concave to convex. (2) A change in the sensitivity of the fluorometer, confirmed by a change in the value recorded for the fluorescent glass reference disk in the instrument. The specific fluorescence of scopoletin decreases with increasing temperature. The assay protocol is designed to minimize any temperature differences at the times the first and second fluorescence readings are taken. H202 release by cells. Human blood mononuclear cells were plated at densities
~d
.c_ E 0
3000
!1t
g
~
2000
.D O
g u
~O00 =
I
I
I
2 3 Cells/well (×10-5)
I
I
4
5
Fig. 1. H202 release by human blood monocytes. Human blood mononuclear cells were plated at densities of 0-5 × 105 cells/well. After 2 h incubation the cultures were washed with PBS, triggered with 100 ng/ml PMA, and assayed for H202 release by measuring the loss of fluorescence of scopoletin upon HPO-catalyzed oxidation by H202. Inset: periodate-elicited mouse peritoneal cells were plated at densities of 0-3 × 10 s cells/well and similarly assayed after overnight incubation. (n = 6, vertical bars represent 2 SD.)
329 o .....-a 1~t3
8oc/
o f
o
Time (minul~m)
Fig. 2. Time course of H202 release. Periodate-elicited mouse periotoneal macrophages, adhering after plating 1 x 105/well, were triggered with 100 n g / m l PMA. H202 release was measured by following the
loss of fluorescenceof scopoletin upon HPO-catalyzedoxidation by H202. Each point is from 1 of 3 wells. The same 3 wells were monitored repeatedlyto obtain the time-coursedepicted.
of 0 - 5 x 105 per well and incubated for 2 h. They were then triggered with 100 n g / m l PMA and assayed for H202 release. The recovery of mononuclear leukocytes after the Ficoll-Paque procedure was usually about I × 107 cells, of which about 20~ were monocytes, from 10 ml of heparinized blood. The change in measured fluorescence for the 60 rain assay period was linear with cell density over the range 0.5-3 x 105 cells plated per well (Fig. 1). Similar data were obtained for periodateelicited mouse peritoneal macrophages (Fig. 1, inset). To illustrate the time-course of H202 production by the human cells we measured fluorescence at intervals over 150 rain. The results are shown in Fig. 2, and are consistent with the kinetics observed using the cuvette assay (Nathan, 1981). Cellprotein estimation. To obtain an estimate of the number of cells adhering to the well after the washing procedure, we measured the cell protein in the well. Fig. 3 shows the cell protein measured in wells after the assay for H202 release by human cells described above. In a separate experiment we plated periodate-elicited macrophages at densities up to 1.25 × 106 per well. The measured protein in the wells after washing reached a plateau at 20-25 #g per well (Fig. 3, inset). All cell protein measurements were corrected for residual serum protein from the culture medium adhering to the plastic by subtracting the mean value for protein measured in the cell-free control wells. In a typical experiment the mean and standard deviation (SD) were 0.65 ± 0.10/~g for 100 #1 RPMI 25% HuS; n = 6. The HPO in the assay mix does not contribute to the measured protein because it is present at the same concentration in the protein standard wells. The absorbance values obtained in the protein assay were decreased by the prior addition of excess H202 ( > 3 nmol), but were not affected by the quantities of H202 released by the cells in the standard assay. To compare cell adherence in the macro-assay and the micro-assay we used periodate-elicited mouse periotoneal macrophages. For the macro-assay these were plated at 1 x 106//well on glass coverslips or directly on the plastic of the 16 nun culture well. For the micro-assay they were plated at 2 × 105/well. After 2 h
330
T~
t
I I
,
Y 10 i
/
it t it
i E ®
,'o(
<~ o
Q.
Ce,,s/we,,
lo-5)
./
][/I
J_
I/i
2
I 1
I 5
2
I
I
4
5
Cells/well (Xl0-5) Fig. 3. Measurement of adherent cell protein. After completion of the H202 release assay (Fig. 1) the cell protein in each well was measured in situ by the method of Lowry et al. (1951). To obtain adherent cell protein values the total protein measured was corrected for protein measured in cell-free control wells. Inset: adherent cell protein measured for mouse peritoneal macrophages plated at densities of 0-12.5 x 105 peritoneal cells/well. (n = 6, vertical bars represent 2 SD.)
incubation the cells were washed and the adherent-cell protein measured. For 1 x 106 cells in the macro-assay plated on glass the mean and SD (n = 3) were 31.4 + 1.3 #g protein, and plated on plastic 20.2 __+3.5 /~g. In the micro-assay for 2 x 105 cells we measured 3.6 + 0.2/tg. These figures are in good agreement with the ratio of 5:1 for the surface areas of the 24-well and the 96-well cell culture areas. In a separate experiment we found that the protein measured in the adherent cells 2 h after plating 1 x 105 casein-eficited cells/well was equivalent to 70% of the cell protein measured in 1 x 105 cells in the initial suspension, which is consistent with the percentage of adherent cells in these preparations (Nathan and Root, 1977). Calculation of specific release of 14202. Using the data from the peroxide and protein assays we were able to express the activity of a cell culture in specific terms of nmol H202 produced per mg of cell protein over 60 or 90 min. The specific release of H202 for a particular well over 90 min was calculated using the formula: nmol
H202 released
(E 0-
W)-(E9o-
W)(CO- W)/(Cgo- W)
(C O- W )
× S
where E 0 is the initial fluorescence reading for t h e well; E90 is the fluorescence reading at 90 min; W is the fluorescence recorded in an empty well; CO and C9o are the mean fluorescence readings in the cell-free control wells at 0 and 90 min respectively; and S is the amount of scopoletin, in nanomoles, added to each well at the start of the assay.
331 1000
A C
-t
E 0
~O
E
O~
500
I
0
E c:
I
N
t t, '
o I
I
2
Cells/well
0
I 1
_
3
( x 10 -b)
I 2 Cells/well
I 3
I 4
I 5
(X|O"5)
Fig. 4. Specific release of H202. The specificrelease of H202 by human monocytes was independent of cell density in the range of 0.5-4× 105 mononuclear cells plated per well, of which about 20% were monocytes. Inset: for mouse peritoneal macrophages H202 release was independent of density in the range 0.2-2 x 105 peritoneal cells plated per well. (n = 6, vertical bars represent 2 SD.)
Because the decrease observed in the control wells is due to a change in the efficiency of measurement, and not to a loss or oxidation of scopoletin, we have used a proportional correction in the formula, as opposed to a subtractive correction. An ideal control condition for the micro-assay fluorescence measurements would utilize cells whose ability to undergo a respiratory burst has been blocked in some specific way. Our present understanding of the pathways involved in this response is limited and we do not know of a satisfactory way to set up such a control at this point. Fig. 4 shows that, for human blood monocytes, cell density had little effect on specific H202 production over the range of 1 - 4 × 105 mononuclear cells plated per well (corresponding to approximately 2 - 8 x 104 monocytes). For mouse peritoneal macrophages the specific H202 release was independent of cell density in the range of 0.2-2 x 105 cells plated per well, corresponding to about 1 x 104 to 1 x 105 macrophages (Fig. 4, inset). For both cell types we observed a diminished specific release at higher cell densities. With the semi-automated procedures described in this report it is now possible to assay as m a n y as 1000 macrophage cultures in 1 experiment. The computation involved in analyzing the data from such an experiment is prohibitive if done by hand using a pocket calculator. It is possible to transmit the data obtained by both the MicroFluor and the BioTek EL307 to a laboratory computer by means of an RS232 serial interface. We have developed a data analysis program written in BASIC for an IBM PC laboratory computer. This program takes as input t h e 0 a n d 60 or 90 min fluorescence readings and the 690 nm absorbance readings and, using the formula described above, computes the specific release of H202 for each well,
332 0.5
"~
0,4
c 0 m
0.3
0 0
c
g o.z .Q 0.!
0
I 2
I 4
I 6
I 8
I 10
BSA (pg)
Fig. 5. The absorbance at 690 nm for the standards in the protein assay. The data are from measurements taken with the BioTek EL307 microplate ahsorbance reader. The curve drawn to fit the data is based on the equation y = 0 . 0 6 0 8 x °'86°3 derived by the method of least squares. Data points are from individual wells.
with the appropriate corrections for fluorescence changes and for protein measured in the cell-free control wells. The protein values are obtained by using an equation of the form y = a x b generated to fit the values for the protein standards in rows A and H (Fig. 5). Cell protein values are computed from the 690 nm absorbance values using this equation. Where calculation is done without the aid of a computer the MicroFluor may be set to subtract the background fluorescence recorded in an empty well (A1) from all the readings taken. The formula above may then be simplified by omitting the subtraction of the mean fluorescence recorded in empty wells (W).
400
c
3OO
0
0 0
o
°
~
8~
o
zoo E
loo
O• --f/
I 10
I ~00
O 1000
PMA (ng/ml)
Fig. 6. The effect of PMA concentration on H202 release. Periodate-elicited mouse periotoneal macrophages were assayed for H202 release in the micro-assay using PMA concentrations of 0-1000 ng/ml. The data show a maximal response plateau in the range of 10-100 ng/ml. The ordinate scale is logarithmic. Data are from individual wells.
333
The mean and SD (n = 3) for measured levels of H202 production (in nmol H202/mg cell protein/90 min) by resident peritoneal macrophages (105 + 49) and by casein (332 + 46)- and sodium periodate (630 + 38)-elicited cells were consistent with the published values reported using the macro-assay (Tsunawaki and Nathan, 1984) and the results for human monocytes (845 + 10, n = 4) are consistent with those reported by Nakagawara et al. (1981). We have used both the micro- and the macro-assay to measure H202 production by cells from the same preparation and obtained results which agreed to within 5% (data not shown). We examined the effect of PMA concentration on the amplitude of the H202 release by the cells (Fig. 6) and confirmed (Tsunawald and Nathan, 1984) that there is a broad peak in the response between 10 and 100 ng/ml PMA, while the response is abolished at concentrations above 500 ng/ml.
Discussion
Existing assays for the cellular release of H202 include those based on: (1) catalase-dependent oxidation of formate to CO2 (Iyer et al., 1961), or (2) of H202 to 02 (Zatti et al., 1968), (3) generation of fluorescent products upon the oxidation of leukodiacetyl-2,7-dichlorofluorescein (Keston and Brandt, 1965), or of (4) homovanillic acid (Rossi, 1980), (5) the loss of fluorescence upon the oxidation of scopoletin (Andreae, 1955), (6) change in absorbance upon oxidation of phenol red (Pick and Keisari, 1980), or formation of enzyme-substrate complexes with (7) yeast cytochrome c peroxidase (Boveris et al., 1972) or (8) horseradish peroxidase (Kakinuma et al., 1979). Method (1) is inaccurate and method (2) is insensitive. Method (7) requires purification of yeast cytochrome peroxidase in substrate (rather than catalytic) quantities. Methods (3)-(6) share the disadvantage of dependence on horseradish peroxidase, that is, susceptibility to interference by reducing co-substrates that compete with the indicator. Among these, however, method (3) suffers the further disadvantage that the substrate must be de-esterified before incubation with the cells and method (6) that the reaction must be terminated to develop the color. In addition, the sensitivity of spectrophotometric detection (method 6) is less than that of fluorometric detection. Finally, method (4) involves high rates of auto-oxidation of the substrate. In codtrast, scopoletin is a highly stable, non-auto-oxidizable, non-toxic fluorophore that can be incubated directly with cells for continuous or intermittent determination of H202 concentrations (Nathan, 1981) with considerable accuracy and sensitivity (Boveris et al., 1977). Previous methods for detection of H202 release by mononuclear phagocytes using scopoletin oxidation have, however, been restricted by the relatively small numbers of samples that could be processed by one individual in one day (in the region of a hundred), by the number of cells needed for routine samples (about 106), and by the volume of test media required (0.3-1.0 ml). The modifications of the scopoletin assay described here markedly extend its versatility and utility.
334 Scopoletin (7-hydroxy-6-methoxy-LH-benzopyran-2-one) is a derivative of the highly fluorescent compound umbelliferone which has been used in developing a variety of fluorogenic enzyme substrates having application in ELISA technology. It has excitation and emission maxima at 320-380 nm and 420-520 nm respectively in KRP. The Dynatech MicroFluor MR600 microtest-plate fluorescence reader is equipped with standard excitation (310-390 nm) and emission (420-480 nm) filters suitable for measuring umbelliferone. It has a low-pressure mercury light source with emission bands in the region of 365 nm. The assay described in this report relies on the loss of fluorescence by scopoletin upon HPO-catalyzed oxidation by H202, and so the Microfluor is well suited for use in a miniaturized scopoletin assay. The micro-assay has proved to be at least at sensitive as the assays performed in 24-well plates or in cuvettes, yielding measures of specific release of H202 consistent with reported values in the literature and with values obtained in parallel macro-assays. The primary attractive features of the new assay are: (1) The feasibility of processing large numbers of samples. It is now a realistic undertaking to process 1000 samples in 1 experiment. Using the macro-assay this would be prohibitively demanding in terms of labor and materials. This is of particular value in studies of the dose-response curves and interactions of agents modulating the production of H202; for the chromatographic isolation of such agents; and for screening hybridomas for production of cytokines or monoclonal antibodies. (2) Time saved. We estimate that assaying 50 samples with the macro-assay requires about 10 h. With the micro-assay the same number of samples may be processed in 2 h. (3) Fewer cells required per sample. The micro-assay requires at most 1 / 1 0 t h the number of cells needed for the macro-assay. In animal studies this represents a saving in terms of monetary costs and sacrifice of animals. In clinical studies the availability of cells is frequently severely limited, and the minimal requirements of the new assay present an invaluable advantage. (4) Tenfold smaller volumes of additives needed in experiments designed to examine the effect of such additives on the macrophage response. This is of particular value in working with human sera, purified material or supernatants from hybridoma cultures. (5) In the macro-assay the measurement of H202 release and the measurement of cell protein are done on separate cultures. The disruption and loss of adherence of macrophages to the coverslip after treatment with PMA (Nathan, 1981) precludes using the same coverslip for the protein assay after completion of the H202 assay. In the micro-assay we have developed a system whereby we measure cell protein in the same well after completion of the assay for H202 release without removing the assay solution. This allows for a more direct relation of activity to cell protein. (6) The fluorescence measurement in the micro-assay is non-destructive, and multiple readings may be taken on a single culture. This property was previously unique to the cuvette assay, which in turn was limited by the small number of samples which could be processed. The micro-assay combines the best aspects of both the cuvette assay and the macro-assay.
335
A micro-assay for H202 release by macrophages, based on the HPO-mediated oxidation of phenol red, was described by Pick and Mizel (1981). The scopoletinbased micro-assay offers several advantages over the phenol red assay as described: (1) In the phenol red assay the contents of all the cell culture wells in the plate are pooled for a conventional protein assay after alkaline digestion of the cells. This provides an estimate of the average cell protein content in the wells. Agents which affect macrophage activation may also affect their adherence or mass (Nathan, 1981), making it important to have individual protein determinations for each well as is the case in the scopoletin micro-assay. (2) Measurements in the phenol red assay are destructive, and to record multiple time-points separate cultures are required for each point. The scopoletin measurements are non-destructive. (3) The alkaline solution of phenol red is not stable and it is necessary to construct an H202 standard curve for each assay. Scopoletin and its oxidation product are stable and allow the calculation of results based on stoichiometry. The semi-automated micro-assay described here has already proven useful in monitoring the administration of recombinant interferon-gamma to patients (C.F. Nathan, C.R. Horowiiz, J. De la Harpe, S. Vadhan-Raj and S. Krown, in preparation) and to mice (Murray et al., 1985), in screening column chromatography fractions in the purification of cytokines (De la Harpe and Nathan, unpublished), and in testing for monoclonal antibodies to cytokines (Szuro-Sudol and Nathan, unpublished).
Acknowledgement We wish to thank Dr. Zanvil Cohn for his critical review of this report.
References Andreae, W.A., 1955, Nature (London) 175, 859. Boveris, A., N. Oshino and B. Chance, 1972, Biochem. J. 128, 617. Boveris, A., E. Martino and A.O.M. Stoppani, 1977, Anal. Biochem. 80, 145. Iyer, G.Y.N., M.F. Islam and J.H. Quastel, 1961, Nature (London) 192, 535. Johnston, Jr., R.B., 1978, Fed. Proc. 37, 2759. Kakinuma, K.,T. Yamaguchi, M. Kaneda, K. Shimada, Y. Tomita and B. Chance, 1979, J. Biochem. 86, 87. Keston, A.S. and R. Brandt, 1965, Anal. Biochem. 11, 1. Lowry, O.H., N.J. Rosebrough, A.L. Fan: and R.J. Randall, 1951, J. Biol. Chem. 193, 265. Murray, H.W., G.L. Spitalny and C.F. Nathan, 1985, J. Immunol. in press. Nakagawara, A., C.F. Nathan and Z. Cohn, 1981, J. Clin. Invest. 68, 1243. Nathan, C.F., 1981, in: Methods for Studying Mononuclear Phagocytes, eds. D.O. Adams, P.J. Edelson and H.S. Koren (Academic Press, New York) pp. 499-510. Nathan, C.F. and R.K. Root, 1977, J. Exp. Med. 146, 1648. Nathan, C.F., H.W. Murray, M.E. Wiebe and B.Y. Rubin, 1983, J: Exp. Med. 158, 670. Pick, E. and Y. Keisari, 1980, J. Immunol. Methods 38, 161.
336 Pick, E. and D. Mizel, 1981, J. Immunol. Methods 46, 211. Root, R.K., J. Metcalf, N. Oshino and B. Chance, 1975, J. Clin. Invest. 55, 945. Rossi, F., P. Bellavite, A. Dobrina, P. Dri and G. Zabucchi, 1980, in: Mononuclear Phagocytes Functional Aspects, Part II, ed. R. Van Furth (Martinus Nijhoff, The Hague) pp. 1187-1213. Shopsis C. and G.J. Mackay, 1984, Anal. Biochem. 140, 104. Szuro-Sudol, A. and C.F. Nathan, 1982, J. Exp. Med. 156, 945. Tsunawaki, S. and C.F. Nathan, 1984, J. Biol. Chem. 259, 4305. Zatti, M., F. Rossi and P. Patriarca, 1968, Experientia 24, 669.
323
Elsevier JIM03459
A Semi-Automated Micro-Assay for H202 Release by Human Blood Monocytes and Mouse Peritoneal Macrophages Jon De la Harpe ] and Carl F. Nathan 2 Laboratory of Cellular Physiology and Immunology, The Rockefeller University, N e w York, U.S.A.
(Received 14 December 1984, accepted 31 December 1984)
H202 secreted by mononuclear phagocytes can be detected by monitoring the horseradish peroxidase-catalyzed oxidation of fluorescent scopoletin. This technique has been adapted to a semi-automated micro-scale with the aid of automatic fluorescence and absorbance micro-culture plate readers to measure H202 and protein, respectively, in the same culture wells. With these adaptations the assay can accurately and precisely detect as little as 0.1 nm01 H202 or 1 ~tg cell protein, permitting the calculation of specific secretion (nmol H 2 0 2 / m g cell protein) from as few as 2 x 104 human blood monocytes or mouse peritoneal macrophages. Cumulative H202 secretion in individual wells may be recorded non-destructively at frequent intervals for time course measurements. Less than 1 rain is required to record the fluorescence in all 96 wells of a micro-culture plate. The assay is highly reproducible, with standard deviations for tripficates typically less than 5-10% of the mean, and gives values in close agreement with those obtained in 10-fold larger samples by previous methods. Using this assay it is feasible to process 1000 samples per day, with order of magnitude savings in labor, cells, sera, media, cytokines, and reagents compared to earlier forms of the assay. The assay is useful in evaluating the cellular effects of cytokines and for assaying their activity in chromatographic fractions and hybridoma cultures. We are currently using the assay to monitor the administration of interferon-gamma to patients with neoplasia. Key words: reactive oxygen intermediates- 1t202 - human monocytes - mouse macrophages - automated microassay
Introduction A n i m p o r t a n t t o o l i n t h e s t u d y o f m o n o n u c l e a r p h a g o c y t e s is q u a n t i t a t i o n o f t h e i r secretion of reactive oxygen intermediates. These incompletely reduced forms of 1 Reprint requests should be addressed to: Dr. Jon De la Harpe, The Rockefeller University Box 280, 1230 York Ave., New York, NY 10021, U.S.A. 2 C.F. Nathan is a research awardee of the Irma T. Hirschl Trust and the Rita Allen Foundation. Abbreviations: PMA, phorbol myristate acetate; HPO, horseradish peroxidase; ROI, reactive oxygen intermediates; KRP, Krebs-Ringer phosphate buffer; KRPG, KRP containing 5.5 mM glucose; PBS, phosphate-buffered saline; DMSO, dimethyl sulfoxide; a-MEM, Eagle's minimum essential medium, alpha variant; FBS, fetal bovine serum; HuS, human serum; BSA, bovine serum albumin; SD, standard deviation. 0022-1759/85/$03.30 © 1985 Elsevier Science Publishers B.V. (Biomedical Division)
324 oxygen, including 02, H202, and OH', can play a major role in the anti-microbial, anti-tumor, and inflammatory functions of monocytes and macrophages. The capacity of mononuclear phagocytes to secrete reactive oxygen intermediates (ROI) holds additional interest because it is closely regulated by cytokines, both positively (interferon-gamma (Nathan et al., 1983)) and negatively (factors from malignant and some non-malignant cells (Szuro-Sudol and Nathan, 1982)). Enhancement of H202 secretory capacity in mononuclear phagocytes from human subjects (C.F. Nathan, C.R. Horowitz, J. De la Harpe, S. Vadhan-Raj and S. Krown, in preparation) and mice (Murray et al., 1985) given recombinant interferon-gamma further heightens interest in the methodology of such measurements. Thus, there is a pressing need for an assay that can quantify the specific activity of ROI release'(nmol product/mg cell protein) on large numbers of samples using low numbers of cells. Such an assay could be employed for monitoring in vivo administration of macrophage activating factors, testing fractions from chromatographic and other steps in the purification of regulatory cytokines, and screening hybridoma supernatants for the production of such cytokines or of antibodies against them. The present report describes a semi-automated, micro-scale modification of the assay based on the horseradish peroxidase (HPO)-catalyzed oxidation of fluorescent scopoletin by H202 as originally introduced by Andreae (1955) and adapted for neutrophils by Root et al. (1975) and for macrophages by Nathan and Root (1977) and Nathan (1981). With the use of automatic fluorescence and absorbance microculture plate readers, the modified assay provides precise, accurate, specific activities for H202 secretion by as few as 2 × 104 monocytes or macrophages per sample, with the ability to detect as little as 0.1 nmol H202 and 1 #g of cell protein. H202 readings are non-destructive and permit repetitive monitoring of the time-course of secretion in individual wells. The fluorescence of 96 samples can be read in less than a minute, so that it is feasible to process 1000 or more samples per day, with an order of magnitude saving in cells, sera, media, cytoldnes, and reagents compared to earlier forms of this assay.
Materials and Methods
Materials
Krebs-Ringer phosphate buffer (KRP) was 129 mM NaC1, 4.86 mM KC1, 0.54 mM CaC12, 1.22 mM MgSO4, 15.8 mM sodium phosphate, pH 7.35, 300-315 mosM. For human cells this was modified to contain 145 mM NaC1 and 5.7 mM sodium phosphate (KRP low phosphate). For the H202 assay solution this KRP buffer was supplemented with 5.5 mM glucose. Phosphate-buffered saline (PBS) was 0.9 mM CaC12, 0.5 mM MgC12, 2.7 mM KC1, 136,9 mM NaC1, 1.5 mM KH2PO4, and 15.2 mM Na2HPO 4, pH 7.35. Fresh KRPG and PBS solutions were prepared weekly and stored at 4°C. Scopoletin (Sigma) was prepared as a 1 mM solution in KRP by dissolution for 24 h at 37°C, sterile filtered, and stored at 4°C in the dark. Casein (Kodak, sodium caseinate, practical grade) was prepared as a 6% solution in 0.95% w / v NaC1,
325 autoclaved, and stored at 40C. Sodium m-periodate (Sigma) was prepared immediately before use as a 5 mM solution in 0.95% NaC1 and filtered through a 0.45 #m Millex filter (Millipore Corp., Bedford, MA). HPO was Sigma P8250 Type II, 150-200 purpurogallin U/mg. It was prepared at 600 purpurogallin U / m l and was stored at -20°C. The triggering agent 4fl-phorbol 12fl-myristate 13a-acetate (Sigma) was prepared as a 300/xg/ml solution in dimethylsulfoxide (DMSO) and stored at 80°C. The working solution was stored at - 20°C and discarded after 1 week, the 80°C stock was discarded after 1 year. Trypan blue was 0.4% in normal saline (Gibco, Grand Island, NY) and was diluted 2-fold in 0.95% NaC1 before use. Ficoll-Paque was from Pharmacia Fine Chemicals, Piscataway, NJ. Bovine serum albumin (BSA) was from Armour Pharmaceutical Co., Phoenix, AZ. The modified Lowry solution A (Lowry et al., 1951) was 4% Na2CO 3 in water. Lowry solution B was 0.1 g CuSO4 • 5H20, 0.2 g tri-sodium citrate. 2H20 in 20 ml H20, and the Folin and Ciocalteau phenol reagent (Accra-Lab., Bridgeport, NJ) was 100 g / l sodium tungstate, 150 g/1 lithium sulfate, 25 g/1 sodium molybdate dihydrate, 100 ml/1 hydrochloric acid, 50 ml/1 phosphoric acid, 2 ml/1 bromine water. Eagle's minimum essential medium, alpha modified (a-MEM) and RPMI were from KC Biologicals, Lenexa, KS. a-MEM was supplemented with 2.2 g/1 NaHCO 3, and brought to pH 7.35 and 300-315 mosM. RPMI was supplemented with 2 mM glutamine. Fetal bovine serum (FBS) was supplied by HyClone, Logan, UT. Human serum (HuS) was prepared as described in Nakagawara et al. (1981). Penicillinstreptomycin solution (5000 U / m l penicillin, 5000/~g/ml streptomycin in normal saline) was from Gibco. Human cells were cultured in 25% HuS, 100 U / m l penicillin, 100/~g/ml streptomycin, 2 mM glutamine in RPMI (RPMI 25% HuS). Mouse peritoneal cells were cultured in 10% FBS, 100 U / m l penicillin, 100 # g / m l streptomycin in a-MEM (a-MEM 10% FBS). -
-
Methods Human peripheral blood monocytes. Human monocytes were obtained from 10 ml of freshly drawn heparinized venous blood. The sample was diluted in an equal volume of sterile saline, layered over 15 ml of Ficoll-Paque in a 50 ml polypropylene conical tube and centrifuged for 20 min at 600 × g at 20°C. The mononuclear cells were aspirated from the liquid/liquid interface (6-7 ml), added to 7 ml RPMI in a 15 ml conical polypropylene tube, and centrifuged for 10 min at 350 × g at 4°C. The pellet was resuspended in 15 ml RPMI, centrifuged for 10 min at 200 × g at 4°C, and then resuspended in 1 ml RPMI 25% HuS. A small volume (10 ~tl) was diluted 10-fold into trypan blue for counting and viability estimation in a hemocytometer. The cell suspension was then diluted in RPMI 25% HuS to achieve the desired final cell density. Peritoneal macrophages. Cells were obtained from female Nelson-Collins strain (NCS) mice (20-30 g) supplied by the Rockefeller University Laboratory Animal Research Center. Mice were sacrified by cervical dislocation and the peritoneal cavity lavaged with 2 washes of 5-10 ml a-MEM. For resident cells, mice received no prior treatment. To obtain activated cells, the mice were injected intraperitoneally
326 with 1 ml of 5 m M N a I O 4 in isotonic saline (Tsunawaki and Nathan, 1984) 2 - 5 days prior to harvest (periodate-elicited cells), or with 1 ml 6% sterile sodium caseinate in isotonic saline 4 days before harvest (casein-elicited cells). The cells were harvested sterilely and kept on ice to minimize adherence. They were centrifuged (15 min, 200 x g, 4°C) and, if necessary, the erythrocytes were lysed by resuspending the pellet in 5 ml 0.2% saline for 30 s followed by the addition of 5 ml of 1.9% saline and 5 ml a - M E M 10% FBS. The cells were then pelleted again and resuspended in a few milliliters of a - M E M 10% FBS. Cell density and viability were determined as described above. Cell cultures. Cells were cultured in 96-well flat-bottomed plates (Coming Glass Works, Coming, N Y or Flow Laboratories, McLean, VA), typically at 1 x 105 per well in 100/zl of culture medium. The peripheral rows of wells in the plate were not used for cell culture but received 100 /~l sterile distilled water per well. Cell-free controls received 100/~l of the serum-containing medium used to culture the cells. To ensure adherence, cells were cultured for at least 2 h before washing. This initial culture period could be extended to as long as 5 days. When human cells were cultured more than a few hours the culture medium was changed after 2 h. The medium in the well was aspirated through a 21-gauge needle and replaced with 100 /zl of fresh medium. HeO2 assay. The cultures were washed 3 times with PBS at 37°C as follows: the plate was inverted over a basin and the medium flicked out with a vigorous wrist action. The plate was then held vertical and gently lowered into a 1 liter beaker of PBS at 37°C. The PBS was removed by again flicking the plate over a basin. It is critical that no bubbles are trapped in the culture wells during this procedure, and the plate should be examined through the side of the beaker while immersed to ensure this. Wells which are incompletely washed yield very high values in the protein assay. After filling and emptying the wells with PBS 3 times, the outer surfaces of the plate were blotted dry on a paper towel and the assay mixture (prewarmed to 37°C) was dispensed into the wells (100 F1/well) using an Eppendorf repeater pipette (Brinkmann Instrument Co., Westbury, NY). The assay mixture was prepared immediately before use from stock solutions and consisted of 30 FM scopoletin, 1 m M NAN3, 1 purpurogallin u n i t / m l HPO, 100 n g / m l PMA in K R P G for mouse cells and in K R P G (low phosphate) for human cells. PMA is suspected to be a tumor promoter and care was taken to avoid skin contact with solutions containing PMA. Immediately after the addition of the assay mixture the plate was placed in a filter fluorometer (MicroFluor MR600, Dynatech Laboratories, Alexandria, VA) and the fluorescence, in arbitrary units, recorded for each well. The plate was then covered with its lid and transferred to an incubation chamber maintained at 37°C and continuously flushed with water-saturated air at 37°C. After 60 or 90 min the fluorescence in'each well was again recorded using the MicroFluor. To monitor the rate of release of H202 the fluorescence may be read at intervals as short as 5 rain. The MicroFluor requires about 30 s to record all the values for a 96-well plate. Cell protein assay. A series of protein standards (0-10 Fg BSA in 100 Fl) was prepared using BSA dissolved in the same batch of assay mix used for measuring
327 H202 release by the cells. The peripheral rows A 1-12 and H 1-12 of the micro-culture plate, which were left empty during the H202 assay, were used for the protein standards. The cell protein estimation was performed without removing the H202 assay mix from the wells. To ensure the disruption of the cells and the solubilization of the cell protein, we added 10 /~1 of 1.0 M NaOH to each well, including the BSA standards. The gap between the plate and its lid was then sealed by stretching a strip of Parafilm around the edge, and this assembly was incubated at 37°C with gentle agitation on a rotary shaker for 60 rain. To estimate cell protein, a modification of the method of Lowry et al. (1951) was used. Upon completion of the alkaline digestion, each well received 100 #1 of modified Lowry reagent C (12 ml 4% Na2CO 3 + 0.5 ml Lowry reagent B). After 10 min at room temperature the plate was placed on a vortexer (Dynatech minishaker) and agitated vigorously while 10/tl of Folin's reagent (diluted by adding 0.2 vols. of water) was added to each assay well using an Eppendorf repeater pipette. After a further 30 min at room temperature the absorbance at 690 nm was measured in each well using a filter absorbance reader (BioTek EL307, BioTek Instruments, Burlington, VT). A similar strategy for measuring cell protein in microculture wells has been described by Shopsis and Mackay, 1984. In all the assays described any bubbles in the wells were eliminated using a hypodermic needle before fluorescence or absorbance readings were taken.
Results We shall refer to the assay for H202 release performed in 24-well plates (Nathan, 1981) as the macro-assay, while the assay performed in 96-well plates, as described in this report, will be referred to as the micro-assay. We examined the reliability and sensitivity of measuring scopoletin oxidation in the micro-culture plate system. The Microfluor reader has a preset range of 0-4096 arbitrary fluorescence units. The measurement of scopoletin by the instrument used in this study is linear over the range of 0-3 nmol of scopoletin per well. The addition of 1.5 nmol of H202 to 100 ~1 of assay mix containing 3 nmol of scopoletin reduced the measured fluorescence from 2651 to 1359 arbitrary units. In the absence of cells, the measurement of scopoletin by fluorescence was not affected by the other components of the assay system (horseradish peroxidase, NaN 3 or PMA). We did not find it necessary to use the special non-fluorescent 96-well plates available from Dynatech. Conventional 96-well tissue-culture plates have acceptably low background fluorescence levels - in the region of 50 arbitrary units. We observed a gradual decrease in measured fluorescence in cell-free control wells, in the range of 100-400 arbitrary units, over the 60 or 90 min incubation for the standard assay. In order to determine the appropriate method to correct for this change when evaluating the experimental results, it was important to know whether this decrease was due to oxidation of the scopoletin in the control wells or to a change in the sensitivity of the measuring system. Using a Perkin-Elmer MPF-4A spectrofluorometer, we measured the fluorescence of pooled control samples re-
328
moved from the wells at the start and end of the incubation period, and established that this change is not due to a loss of fluorescence of the assay solution, and so not due to oxidation of the scopoletin. We believe that this change in measured fluorescence is due to 2 factors: (1) A change in the shape of the meniscus in the wells containing assay mix, resulting in a change in the effective light-path and in the refractive properties of the air-liquid interface. We examined the effect of the meniscus shape on the measured fluorescence of a given amount of scopoletin. Wells containing 1.5 nmol of scopoletin in volumes ranging from 50 to 300/~1 of K R P G register the same fluorescence in the MicroFluor (within 10%). However, when the volume is increased from 300 to 340 /~1, the recorded fluorescence increases by 80%. This is the volume range in which the liquid level reaches the rim of the well and the meniscus changes from concave to convex. (2) A change in the sensitivity of the fluorometer, confirmed by a change in the value recorded for the fluorescent glass reference disk in the instrument. The specific fluorescence of scopoletin decreases with increasing temperature. The assay protocol is designed to minimize any temperature differences at the times the first and second fluorescence readings are taken. H202 release by cells. Human blood mononuclear cells were plated at densities
~d
.c_ E 0
3000
!1t
g
~
2000
.D O
g u
~O00 =
I
I
I
2 3 Cells/well (×10-5)
I
I
4
5
Fig. 1. H202 release by human blood monocytes. Human blood mononuclear cells were plated at densities of 0-5 × 105 cells/well. After 2 h incubation the cultures were washed with PBS, triggered with 100 ng/ml PMA, and assayed for H202 release by measuring the loss of fluorescence of scopoletin upon HPO-catalyzed oxidation by H202. Inset: periodate-elicited mouse peritoneal cells were plated at densities of 0-3 × 10 s cells/well and similarly assayed after overnight incubation. (n = 6, vertical bars represent 2 SD.)
329 o .....-a 1~t3
8oc/
o f
o
Time (minul~m)
Fig. 2. Time course of H202 release. Periodate-elicited mouse periotoneal macrophages, adhering after plating 1 x 105/well, were triggered with 100 n g / m l PMA. H202 release was measured by following the
loss of fluorescenceof scopoletin upon HPO-catalyzedoxidation by H202. Each point is from 1 of 3 wells. The same 3 wells were monitored repeatedlyto obtain the time-coursedepicted.
of 0 - 5 x 105 per well and incubated for 2 h. They were then triggered with 100 n g / m l PMA and assayed for H202 release. The recovery of mononuclear leukocytes after the Ficoll-Paque procedure was usually about I × 107 cells, of which about 20~ were monocytes, from 10 ml of heparinized blood. The change in measured fluorescence for the 60 rain assay period was linear with cell density over the range 0.5-3 x 105 cells plated per well (Fig. 1). Similar data were obtained for periodateelicited mouse peritoneal macrophages (Fig. 1, inset). To illustrate the time-course of H202 production by the human cells we measured fluorescence at intervals over 150 rain. The results are shown in Fig. 2, and are consistent with the kinetics observed using the cuvette assay (Nathan, 1981). Cellprotein estimation. To obtain an estimate of the number of cells adhering to the well after the washing procedure, we measured the cell protein in the well. Fig. 3 shows the cell protein measured in wells after the assay for H202 release by human cells described above. In a separate experiment we plated periodate-elicited macrophages at densities up to 1.25 × 106 per well. The measured protein in the wells after washing reached a plateau at 20-25 #g per well (Fig. 3, inset). All cell protein measurements were corrected for residual serum protein from the culture medium adhering to the plastic by subtracting the mean value for protein measured in the cell-free control wells. In a typical experiment the mean and standard deviation (SD) were 0.65 ± 0.10/~g for 100 #1 RPMI 25% HuS; n = 6. The HPO in the assay mix does not contribute to the measured protein because it is present at the same concentration in the protein standard wells. The absorbance values obtained in the protein assay were decreased by the prior addition of excess H202 ( > 3 nmol), but were not affected by the quantities of H202 released by the cells in the standard assay. To compare cell adherence in the macro-assay and the micro-assay we used periodate-elicited mouse periotoneal macrophages. For the macro-assay these were plated at 1 x 106//well on glass coverslips or directly on the plastic of the 16 nun culture well. For the micro-assay they were plated at 2 × 105/well. After 2 h
330
T~
t
I I
,
Y 10 i
/
it t it
i E ®
,'o(
<~ o
Q.
Ce,,s/we,,
lo-5)
./
][/I
J_
I/i
2
I 1
I 5
2
I
I
4
5
Cells/well (Xl0-5) Fig. 3. Measurement of adherent cell protein. After completion of the H202 release assay (Fig. 1) the cell protein in each well was measured in situ by the method of Lowry et al. (1951). To obtain adherent cell protein values the total protein measured was corrected for protein measured in cell-free control wells. Inset: adherent cell protein measured for mouse peritoneal macrophages plated at densities of 0-12.5 x 105 peritoneal cells/well. (n = 6, vertical bars represent 2 SD.)
incubation the cells were washed and the adherent-cell protein measured. For 1 x 106 cells in the macro-assay plated on glass the mean and SD (n = 3) were 31.4 + 1.3 #g protein, and plated on plastic 20.2 __+3.5 /~g. In the micro-assay for 2 x 105 cells we measured 3.6 + 0.2/tg. These figures are in good agreement with the ratio of 5:1 for the surface areas of the 24-well and the 96-well cell culture areas. In a separate experiment we found that the protein measured in the adherent cells 2 h after plating 1 x 105 casein-eficited cells/well was equivalent to 70% of the cell protein measured in 1 x 105 cells in the initial suspension, which is consistent with the percentage of adherent cells in these preparations (Nathan and Root, 1977). Calculation of specific release of 14202. Using the data from the peroxide and protein assays we were able to express the activity of a cell culture in specific terms of nmol H202 produced per mg of cell protein over 60 or 90 min. The specific release of H202 for a particular well over 90 min was calculated using the formula: nmol
H202 released
(E 0-
W)-(E9o-
W)(CO- W)/(Cgo- W)
(C O- W )
× S
where E 0 is the initial fluorescence reading for t h e well; E90 is the fluorescence reading at 90 min; W is the fluorescence recorded in an empty well; CO and C9o are the mean fluorescence readings in the cell-free control wells at 0 and 90 min respectively; and S is the amount of scopoletin, in nanomoles, added to each well at the start of the assay.
331 1000
A C
-t
E 0
~O
E
O~
500
I
0
E c:
I
N
t t, '
o I
I
2
Cells/well
0
I 1
_
3
( x 10 -b)
I 2 Cells/well
I 3
I 4
I 5
(X|O"5)
Fig. 4. Specific release of H202. The specificrelease of H202 by human monocytes was independent of cell density in the range of 0.5-4× 105 mononuclear cells plated per well, of which about 20% were monocytes. Inset: for mouse peritoneal macrophages H202 release was independent of density in the range 0.2-2 x 105 peritoneal cells plated per well. (n = 6, vertical bars represent 2 SD.)
Because the decrease observed in the control wells is due to a change in the efficiency of measurement, and not to a loss or oxidation of scopoletin, we have used a proportional correction in the formula, as opposed to a subtractive correction. An ideal control condition for the micro-assay fluorescence measurements would utilize cells whose ability to undergo a respiratory burst has been blocked in some specific way. Our present understanding of the pathways involved in this response is limited and we do not know of a satisfactory way to set up such a control at this point. Fig. 4 shows that, for human blood monocytes, cell density had little effect on specific H202 production over the range of 1 - 4 × 105 mononuclear cells plated per well (corresponding to approximately 2 - 8 x 104 monocytes). For mouse peritoneal macrophages the specific H202 release was independent of cell density in the range of 0.2-2 x 105 cells plated per well, corresponding to about 1 x 104 to 1 x 105 macrophages (Fig. 4, inset). For both cell types we observed a diminished specific release at higher cell densities. With the semi-automated procedures described in this report it is now possible to assay as m a n y as 1000 macrophage cultures in 1 experiment. The computation involved in analyzing the data from such an experiment is prohibitive if done by hand using a pocket calculator. It is possible to transmit the data obtained by both the MicroFluor and the BioTek EL307 to a laboratory computer by means of an RS232 serial interface. We have developed a data analysis program written in BASIC for an IBM PC laboratory computer. This program takes as input t h e 0 a n d 60 or 90 min fluorescence readings and the 690 nm absorbance readings and, using the formula described above, computes the specific release of H202 for each well,
332 0.5
"~
0,4
c 0 m
0.3
0 0
c
g o.z .Q 0.!
0
I 2
I 4
I 6
I 8
I 10
BSA (pg)
Fig. 5. The absorbance at 690 nm for the standards in the protein assay. The data are from measurements taken with the BioTek EL307 microplate ahsorbance reader. The curve drawn to fit the data is based on the equation y = 0 . 0 6 0 8 x °'86°3 derived by the method of least squares. Data points are from individual wells.
with the appropriate corrections for fluorescence changes and for protein measured in the cell-free control wells. The protein values are obtained by using an equation of the form y = a x b generated to fit the values for the protein standards in rows A and H (Fig. 5). Cell protein values are computed from the 690 nm absorbance values using this equation. Where calculation is done without the aid of a computer the MicroFluor may be set to subtract the background fluorescence recorded in an empty well (A1) from all the readings taken. The formula above may then be simplified by omitting the subtraction of the mean fluorescence recorded in empty wells (W).
400
c
3OO
0
0 0
o
°
~
8~
o
zoo E
loo
O• --f/
I 10
I ~00
O 1000
PMA (ng/ml)
Fig. 6. The effect of PMA concentration on H202 release. Periodate-elicited mouse periotoneal macrophages were assayed for H202 release in the micro-assay using PMA concentrations of 0-1000 ng/ml. The data show a maximal response plateau in the range of 10-100 ng/ml. The ordinate scale is logarithmic. Data are from individual wells.
333
The mean and SD (n = 3) for measured levels of H202 production (in nmol H202/mg cell protein/90 min) by resident peritoneal macrophages (105 + 49) and by casein (332 + 46)- and sodium periodate (630 + 38)-elicited cells were consistent with the published values reported using the macro-assay (Tsunawaki and Nathan, 1984) and the results for human monocytes (845 + 10, n = 4) are consistent with those reported by Nakagawara et al. (1981). We have used both the micro- and the macro-assay to measure H202 production by cells from the same preparation and obtained results which agreed to within 5% (data not shown). We examined the effect of PMA concentration on the amplitude of the H202 release by the cells (Fig. 6) and confirmed (Tsunawald and Nathan, 1984) that there is a broad peak in the response between 10 and 100 ng/ml PMA, while the response is abolished at concentrations above 500 ng/ml.
Discussion
Existing assays for the cellular release of H202 include those based on: (1) catalase-dependent oxidation of formate to CO2 (Iyer et al., 1961), or (2) of H202 to 02 (Zatti et al., 1968), (3) generation of fluorescent products upon the oxidation of leukodiacetyl-2,7-dichlorofluorescein (Keston and Brandt, 1965), or of (4) homovanillic acid (Rossi, 1980), (5) the loss of fluorescence upon the oxidation of scopoletin (Andreae, 1955), (6) change in absorbance upon oxidation of phenol red (Pick and Keisari, 1980), or formation of enzyme-substrate complexes with (7) yeast cytochrome c peroxidase (Boveris et al., 1972) or (8) horseradish peroxidase (Kakinuma et al., 1979). Method (1) is inaccurate and method (2) is insensitive. Method (7) requires purification of yeast cytochrome peroxidase in substrate (rather than catalytic) quantities. Methods (3)-(6) share the disadvantage of dependence on horseradish peroxidase, that is, susceptibility to interference by reducing co-substrates that compete with the indicator. Among these, however, method (3) suffers the further disadvantage that the substrate must be de-esterified before incubation with the cells and method (6) that the reaction must be terminated to develop the color. In addition, the sensitivity of spectrophotometric detection (method 6) is less than that of fluorometric detection. Finally, method (4) involves high rates of auto-oxidation of the substrate. In codtrast, scopoletin is a highly stable, non-auto-oxidizable, non-toxic fluorophore that can be incubated directly with cells for continuous or intermittent determination of H202 concentrations (Nathan, 1981) with considerable accuracy and sensitivity (Boveris et al., 1977). Previous methods for detection of H202 release by mononuclear phagocytes using scopoletin oxidation have, however, been restricted by the relatively small numbers of samples that could be processed by one individual in one day (in the region of a hundred), by the number of cells needed for routine samples (about 106), and by the volume of test media required (0.3-1.0 ml). The modifications of the scopoletin assay described here markedly extend its versatility and utility.
334 Scopoletin (7-hydroxy-6-methoxy-LH-benzopyran-2-one) is a derivative of the highly fluorescent compound umbelliferone which has been used in developing a variety of fluorogenic enzyme substrates having application in ELISA technology. It has excitation and emission maxima at 320-380 nm and 420-520 nm respectively in KRP. The Dynatech MicroFluor MR600 microtest-plate fluorescence reader is equipped with standard excitation (310-390 nm) and emission (420-480 nm) filters suitable for measuring umbelliferone. It has a low-pressure mercury light source with emission bands in the region of 365 nm. The assay described in this report relies on the loss of fluorescence by scopoletin upon HPO-catalyzed oxidation by H202, and so the Microfluor is well suited for use in a miniaturized scopoletin assay. The micro-assay has proved to be at least at sensitive as the assays performed in 24-well plates or in cuvettes, yielding measures of specific release of H202 consistent with reported values in the literature and with values obtained in parallel macro-assays. The primary attractive features of the new assay are: (1) The feasibility of processing large numbers of samples. It is now a realistic undertaking to process 1000 samples in 1 experiment. Using the macro-assay this would be prohibitively demanding in terms of labor and materials. This is of particular value in studies of the dose-response curves and interactions of agents modulating the production of H202; for the chromatographic isolation of such agents; and for screening hybridomas for production of cytokines or monoclonal antibodies. (2) Time saved. We estimate that assaying 50 samples with the macro-assay requires about 10 h. With the micro-assay the same number of samples may be processed in 2 h. (3) Fewer cells required per sample. The micro-assay requires at most 1 / 1 0 t h the number of cells needed for the macro-assay. In animal studies this represents a saving in terms of monetary costs and sacrifice of animals. In clinical studies the availability of cells is frequently severely limited, and the minimal requirements of the new assay present an invaluable advantage. (4) Tenfold smaller volumes of additives needed in experiments designed to examine the effect of such additives on the macrophage response. This is of particular value in working with human sera, purified material or supernatants from hybridoma cultures. (5) In the macro-assay the measurement of H202 release and the measurement of cell protein are done on separate cultures. The disruption and loss of adherence of macrophages to the coverslip after treatment with PMA (Nathan, 1981) precludes using the same coverslip for the protein assay after completion of the H202 assay. In the micro-assay we have developed a system whereby we measure cell protein in the same well after completion of the assay for H202 release without removing the assay solution. This allows for a more direct relation of activity to cell protein. (6) The fluorescence measurement in the micro-assay is non-destructive, and multiple readings may be taken on a single culture. This property was previously unique to the cuvette assay, which in turn was limited by the small number of samples which could be processed. The micro-assay combines the best aspects of both the cuvette assay and the macro-assay.
335
A micro-assay for H202 release by macrophages, based on the HPO-mediated oxidation of phenol red, was described by Pick and Mizel (1981). The scopoletinbased micro-assay offers several advantages over the phenol red assay as described: (1) In the phenol red assay the contents of all the cell culture wells in the plate are pooled for a conventional protein assay after alkaline digestion of the cells. This provides an estimate of the average cell protein content in the wells. Agents which affect macrophage activation may also affect their adherence or mass (Nathan, 1981), making it important to have individual protein determinations for each well as is the case in the scopoletin micro-assay. (2) Measurements in the phenol red assay are destructive, and to record multiple time-points separate cultures are required for each point. The scopoletin measurements are non-destructive. (3) The alkaline solution of phenol red is not stable and it is necessary to construct an H202 standard curve for each assay. Scopoletin and its oxidation product are stable and allow the calculation of results based on stoichiometry. The semi-automated micro-assay described here has already proven useful in monitoring the administration of recombinant interferon-gamma to patients (C.F. Nathan, C.R. Horowiiz, J. De la Harpe, S. Vadhan-Raj and S. Krown, in preparation) and to mice (Murray et al., 1985), in screening column chromatography fractions in the purification of cytokines (De la Harpe and Nathan, unpublished), and in testing for monoclonal antibodies to cytokines (Szuro-Sudol and Nathan, unpublished).
Acknowledgement We wish to thank Dr. Zanvil Cohn for his critical review of this report.
References Andreae, W.A., 1955, Nature (London) 175, 859. Boveris, A., N. Oshino and B. Chance, 1972, Biochem. J. 128, 617. Boveris, A., E. Martino and A.O.M. Stoppani, 1977, Anal. Biochem. 80, 145. Iyer, G.Y.N., M.F. Islam and J.H. Quastel, 1961, Nature (London) 192, 535. Johnston, Jr., R.B., 1978, Fed. Proc. 37, 2759. Kakinuma, K.,T. Yamaguchi, M. Kaneda, K. Shimada, Y. Tomita and B. Chance, 1979, J. Biochem. 86, 87. Keston, A.S. and R. Brandt, 1965, Anal. Biochem. 11, 1. Lowry, O.H., N.J. Rosebrough, A.L. Fan: and R.J. Randall, 1951, J. Biol. Chem. 193, 265. Murray, H.W., G.L. Spitalny and C.F. Nathan, 1985, J. Immunol. in press. Nakagawara, A., C.F. Nathan and Z. Cohn, 1981, J. Clin. Invest. 68, 1243. Nathan, C.F., 1981, in: Methods for Studying Mononuclear Phagocytes, eds. D.O. Adams, P.J. Edelson and H.S. Koren (Academic Press, New York) pp. 499-510. Nathan, C.F. and R.K. Root, 1977, J. Exp. Med. 146, 1648. Nathan, C.F., H.W. Murray, M.E. Wiebe and B.Y. Rubin, 1983, J: Exp. Med. 158, 670. Pick, E. and Y. Keisari, 1980, J. Immunol. Methods 38, 161.
336 Pick, E. and D. Mizel, 1981, J. Immunol. Methods 46, 211. Root, R.K., J. Metcalf, N. Oshino and B. Chance, 1975, J. Clin. Invest. 55, 945. Rossi, F., P. Bellavite, A. Dobrina, P. Dri and G. Zabucchi, 1980, in: Mononuclear Phagocytes Functional Aspects, Part II, ed. R. Van Furth (Martinus Nijhoff, The Hague) pp. 1187-1213. Shopsis C. and G.J. Mackay, 1984, Anal. Biochem. 140, 104. Szuro-Sudol, A. and C.F. Nathan, 1982, J. Exp. Med. 156, 945. Tsunawaki, S. and C.F. Nathan, 1984, J. Biol. Chem. 259, 4305. Zatti, M., F. Rossi and P. Patriarca, 1968, Experientia 24, 669.