A new simple fluorometric assay for phagocytosis

A new simple fluorometric assay for phagocytosis

Journal of Immunological Methods, 88 (1986) 175-183 175 Elsevier JIM 03865 A new simple fluorometric assay for phagocytosis Tatsuya Oda and Hiroshi...

763KB Sizes 1 Downloads 72 Views

Journal of Immunological Methods, 88 (1986) 175-183

175

Elsevier JIM 03865

A new simple fluorometric assay for phagocytosis Tatsuya Oda and Hiroshi Maeda * Department of Microbiology, Kumamoto University School of Medicine, Kumamoto 860, Japan (Received 2 October 1985, accepted 22 November 1985)

A highly sensitive, but simple and quantitative, fluorometric assay method for phagocytosis by cells such as macrophages and polymorphonuclear leukocytes was developed by utilizing fluorescent particles. Escherichia coli, Serratia marcescens, yeast, and latex particles were conjugated with fluorescein isothiocyanate and used as fluorescent particles. The assay procedure requires phagocytic cells, appropriate medium, fluorescent particles, sodium dodecyl sulfate, microtiter culture plate (24 wells), clinical centrifuge, and fluorescence spectrophotometer. One hundred assays can be done within 30 min after the incubation period. A time course analysis with this method showed that the phagocytosis of all these particles was dependent on temperature, and that the number of particles ingested by cells increased rapidly during the initial 30 min of incubation at 37°C. Free fluorescent particles can be removed effectively by aspiration from the well. At 0°C, very few particles were ingested by cells or adsorbed onto the phagocytic cell surface as confirmed by fluorescence microscopy. An inhibitory effect of cepharanthin and sodium azide on phagocytosis was also confirmed by this method. The differential susceptibility of E. coli B and S. marcescens to phagocytosis also could be determined by this method. Key words: Phagocytosis; Fluorometry," Macrophage," Granulocyte

Introduction

It is well known that cells such as macrophages or polymorphonuclear leukocytes (PMN) ingest many kinds of microorganisms and inert particles by phagocytosis (Stossel, 1975). Phagocytosis is an important part of the cellular defense system, and it consists of an integrated series of complex events. In general, the most frequently used assay method to quantify phagocytosis is the direct counting of the actual number of particles per cell under a microscope (Wehinger and Hofacker, 1976). Another method developed to estimate phagocyto* To whom correspondence should be addressed. Abbreviations: PMN, polymorphonuclear leukocyte; KRP, Krebs-Ringer phosphate buffer, pH 7.4; PBS, 0.01 M phosphate-buffered 0.15 M saline, pH 7.4; HBSS, Hanks' balanced salt solution; FITC, fluorescein isothiocyanate; SDS, sodium dodecyl sulfate.

sis has used cell-associated radiolabeled bacteria (Verhoef et al., 1977). These procedures are, however, tedious and time-consuming. Furthermore, it is very difficult to analyze the time course of particle uptake properly during the phagocytic process. The present method will not only solve these problems, but also will provide more reproducible and more accurate data on the phagocytic process. In addition, we found that the present method could be readily applied to analysis of some of the factors involved in phagocytosis, such as opsonin, temperature, and various drugs.

Materials and methods Microorganisms Escherichia coli B and Serratia marcescens kurus

3958 isolated from a patient in Kumamoto Uni-

0022-1759/86/$03.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)

176 versity Hospital were cultured and maintained in tryptosoy medium. After overnight culture at 37°C, the bacteria were harvested by centrifugation, washed 4 times with 0.01 M phosphate-buffered 0.15 M saline (PBS), pH 7.4, and washed once with distilled water followed by lyophilization. Dried powder of baker's yeast (Gibco, Grand Island, NY) was washed extensively with PBS followed by distilled water, and was then lyophilized. All organisms were heat killed (121°C, 15 rain) before lyophilization.

Preparation of fluorescent particles Fluorescent latex particles (2/Lm diameter) were obtained from PolySciences, Warrington, PA. Fluorescent microorganisms were prepared as follows: 20 mg of lyophilized bacteria or yeast was suspended in 5 ml of 0.1 M carbonate-bicarbonate buffer, p H 9.0, containing 0.9% saline at 4°C. To this solution, 1 mg of fluorescein isothiocyanate (FITC) was added. After mixing for 30 min at 4°C, the cells were washed vigorously several times with cold PBS by centrifugation and once with distilled water, and were then lyophilized. The supernatant of the last centrifugation was found uncontaminated by free fluorescent dye.

Peritoneal macrophages Male mice (ddy) were injected intraperitoneally with 1 KE (clinical unit) of OK-432 in 0.1 ml of Hanks' balanced salt solution (HBSS) to elicit peritoneal macrophages. OK-432 is a heat-killed preparation of Streptococcuspyogenes with penicillin (kindly provided by Chugai Pharmaceutical Co., Tokyo). Three days after the injection the resulting peritoneal exudate cells were harvested by peritoneal lavage with 10 ml of cold HBSS containing 10 U heparin/ml, which yielded 1 - 2 × 10 v cells/mouse. After centrifugation at 200 x g, the cell pellets were resuspended in 2 ml of 0.2% saline for 30-60 s to lyse erythrocytes. Isotonicity was restored with 2 ml of 1.6% saline, followed by addition of 6 ml HBSS. Centrifugation was repeated, and the cells were resuspended in cold R P M I 1640 medium supplemented with 10% heat-inactivated fetal calf serum. The cells were placed in 16 m m wells (Falcon, 24-well plastic plate) at a density of 1 × 10 6 cells/well, and were incubated for 3 h at 37°C under 5% CO 2 and 95%

air. Then the medium was removed and the adherent cells were washed vigorously 3 times with HBSS to remove nonadherent cells. More than 95% of these adherent cells exhibited the morphological and staining characteristics of macrophages.

Granulocytes Human PMN were prepared from heparinized venous blood of healthy donors. Fifty ml of heparinized blood was mixed with 10 ml of 6% dextran (molecular weight 200000) in PBS and was kept at room temperature for 1 h while most of the erythrocytes settled to the bottom of the tube. The leukocyte-rich plasma was separated manually and was centrifuged at 160 x g for 5 min. The resulting pellet was resuspended in 4 ml of RPMI 1640 medium, and 0.5 ml of the cell suspension was layered on the top of a centrifuge tube (1.5 x 12 cm) that contained 10 ml of 58% Percoll in RPMI 1640 medium, and was centrifuged at 400 X g for 30 min at 4°C. The sedimented cells, containing granulocytes (PMN) and some contaminating erythrocytes, were washed 3 times with RPMI 1640 medium. The remaining erythrocytes were lysed with distilled water for 30 s, and the resultant PMN were washed twice in the above medium. The final PMN suspension was adjusted to contain 2 x 106 cells/ml of RPMI 1640 medium, and 0.5 ml samples of the cell suspension were placed in 16 mm wells as described and incubated for 30 rain at 37°C under 5% CO 2 and 95% air. The nonadherent cells were then removed by washing 3 times with PBS. More than 95% of the adherent cells were identified as PMN by Wright-Giemsa staining and were viable by trypan blue dye exclusion.

Serum Pooled mouse serum was prepared from blood obtained after clotting; samples from 5 mice were pooled just before use. Pooled human serum was prepared from 5 healthy donors similarly. The serum was stored at - 80°C and was used within 3 weeks.

Opsonization Pooled human and mouse sera were used as a convenient source of opsonins for human PMN

177

and mouse macrophage phagocytosis, respectively. Fluorescent particles were preincubated in 50% pooled serum diluted with KRP (Krebs-Ringer phosphate buffer, pH 7.4) at 37°C for 30 min. After opsonization, particle suspensions were added to the assay mixture to a final concentration of 5% serum.

~c_1C L

io° o

Measurement of phagocytosis

Ingestion of fluorescent particles by phagocytic cells was measured as follows. Fluorescein-labeled particle suspensions in KRP were added to adherent phagocytic cells (1 x 1 0 6 cells/well), which were overlaid with 0.5 ml of KRP. The final particle-to-phagocytic cell ratios were as follows: latex particles 50 : 1, yeast 50 : 1, E. coil 500 : 1, and S. marcescens 500 : 1. These ratios of particles to cells produced maximal rates of phagocytosis in preliminary experiments in which the ratios were varied. After incubation at 0 or 37°C for various times, phagocytic cells were washed 3 times with PBS by aspiration to remove any non-cell-associated particles. Cells were then solubilized by addition of 1 ml of 25 mM Tris-HC1 buffer, pH 8.5, containing 0.2% sodium dodecyl sulfate (SDS). The solubilized solution was then transferred to a cuvette for measurement of the fluorescent intensity by a fluorescence spectrophotometer (Hitachi Model 650-40) with an excitation wavelength of 490 nm and an emission wavelength of 520 nm. The number of phagocytosed particles was determined from the relationship between the fluorescent intensity and number of particles de-

01

103

104 105 10 e Concentration {number of particles/ml)

107

Fig. 1. Relationship between fluorescent intensity and particle concentration. Fluorescence-labeled latex particles (e), yeast ( 0 ) , E. coil (A), and S. marcescens (r,) were suspended in 20 m M Tris-HCl buffer at pH 8.5, containing 0.2% SDS at the indicated bacterial number, and the fluorescent intensity of each suspension was measured as described in the text.

termined for standard separately. To quantify the adherent cell number, the protein content in an aliquot of the solubilized solution was determined by the method of Lowry et aL (1951) with bovine serum albumin as a standard. The extent of phagocytosis can thus be expressed in terms of number of particles per cell.

(B)

o

_=, TABLE I

o.

P H A G O C Y T O S I S O F PREOPSON1ZED LATEX PARTICLES BY P H A G O C Y T I C CELLS: C O R R E L A T I O N OF D A T A O B T A I N E D BY MICROSCOPY A N D F L U O R O M E TRY Cells

Macrophages PMN

Average number of phagocytosed particles per cell ~ By microscopy

By fluorometry

8.3 (7.8-8.8) 4.5 (4-5.1)

9.2 (6.3-12.1) 6.1 (4.3-7.9)

a After 120 min of incubation at 37°C, more than 200 phagocytic cells per sample were counted under the microscope. The same samples were measured by fluorometry. Data are the m e a n s of three experiments; the ranges are given in parentheses.

~4

30

60

90

Time ( m i n )

1"~"0

0

30

60

90

120

Time (min)

Fig, 2. Time course of phagocytosis of fluorescent latex particles by phagocytic cells. Phagocytosis of opsonized (e) or nonopsonized ( © ) particles at 37°C and that of opsonized (A) or nonopsonized (zx) particles at 0°C by mouse peritoneal macrophages (A) or h u m a n PMN ( B ) were measured as described in the text. The data were expressed as the number of ingested particles per cell; each point is the mean of duplicate values.

178

J

Fig. 3. Photomicrographs of mouse peritoneal macrophages (A, B) and human PMN (C, D) with ingested fluorescent latex particles under phase-contrast (A,C) and fluorescence ( B , D ) conditions. Preopsonized particles were added to adherent cells in these experiments. After incubation for 120 rain at 37°C, samples were taken for microscopic evaluation (× 600).

179

Fluorescence microscopy All samples in this assay were studied by fluorescence microscopy (inverted fluorescence microscope equipped with phase contrast; Olympus Model IMT-2, Tokyo) at x 600 magnification. In the case of phagocytosis of fluorescent latex particles, more than 200 phagocytic cells/sample were counted visually to determine the number of ingested particles per cell, and the same samples were subjected to measurement by fluorescence spectrophotometer.

concentration. From these standard curves, therefore, the particle number could be determined quantitatively by the measurement of fluorescent intensity.

Kinetics of phagocytosis of fluorescent latex particles The time course of uptake of fluorescent latex particles by mouse peritoneal macrophages or human PMN was examined at 0 and 37°C. Both macrophages and PMN readily ingested preopsonized latex particles in a temperature-dependent fashion (Fig. 2). At 37°C, the rate of inges-

Assay for H_~O2 release Production of H202 by human PMN during phagocytosis was quantitated by the scopoletin method (Nathan and Root, 1977), based on the horseradish peroxidase-dependent oxidation of scopoletin by H202. Briefly, 1 x 10 6 PMN were incubated in 1 ml of KRP, which contained 1 ~M scopoletin and 20 U / m l of horseradish peroxidase, with preopsonized or nonopsonized bacteria (1 x 10~/ml) in 1 × 1 cm light path cuvettes at 37°C. The fluorescence of scopoletin was measured continuously with a fluorescence spectrophotometer (Hitachi Model 650-40) at an excitation wavelength of 350 nm and an emission wavelength of 460 nm. Under assay conditions, the loss of fluorescence was proportional to the concentration of H202.

Results

100

D

(A)

55

I

I

10 210 110o 101 102 103 104 105

0 I001-



(B)

O

8 :B

Relationship between the fluorescent intensity and number of particles Fluorescent particles used in this study were suspended in KRP. The suspension was vigorously agitated by sonication to disperse particles completely, after which the number of particles was counted directly by a hemocytometer under a microscope or by a Coulter counter (Coulter Electronics, U.S.A., Model ZBI). For the determination of fluorescent intensity of the particles by fluorescence spectrophotometry, particles were suspended in 20 mM Tris-HC1 buffer, pH 8.5, containing 0.2% SDS; the excitation wavelength was 490 nm and the emission wavelength was 520 nm. As shown in Fig. 1, the fluorescent intensity of these particles was directly proportional to their

:

0

I

I

10-2 10-1100 101 102 103 104 105 Concentration (uM)

Fig. 4. Effect of sodium azide and cepharanthin on the phagocytosis of preopsonized fluorescent latex particles by mouse peritoneal macrophages (A) or human PMN (B) at 37°C. In the presence of the indicated concentrations of sodium azide (e) or cepharanthin (C)), the number of ingested particles per cell was measured after 60 min of incubation at 37°C, as described in Fig. 2. The data were expressed as percent inhibition.

180

tion of particles by these phagocytic cells was essentially linear with time up to 60 min, and then reached a plateau after 120 min. In contrast, uptake of particles at 0°C was almost negligible, and

the amount of ingested particles after 60 min was less than 5% of that at 37°C. Furthermore, both phagocytic cells required preopsonization with serum to achieve optimal

O

~



w

Fig. 5. Microscopic evaluation of phagoeytosis of FITC-labeled microorganisms by fluorescence quenching. Mouse peritoneal macrophages (A, C, E) or h u m a n P M N (B, D, F ) were incubated with preopsonized FITC-labeled yeast (A, B), E. coli (C, D), or S. marcescens (E, F ) as described in the text. After 120 min of incubation at 37°C, methylene blue was added to each sample to quench the numerous extracellular microorganisms, which are not visible. Then photographs were taken. Only ingested microorganisms retained their intense fluorescence as shown.

181

uptake of particles. The uptake of nonopsonized particles by macrophages or PMN was significantly lower than that of opsonized particles (Fig. 2): only 37 and 39% of maximal uptake was achieved, respectively. Fluorescence photomicrography showed many particles within phagocytic cells, for 37°C incubation, whereas for incubation at 0°C, there was almost no uptake of particles, or very little adherence to phagocytic cells (Fig. 3). Furthermore, the mean number of particles per phagocytic cell was highly correlated when fluorometric and microscopic data were compared (Table I). These findings confirmed that the measurements of uptake by the fluorometric method provided an accurate determination of the internalization of these particles.

Effect of sodium azide and cepharanthin on phagocytosis of fluorescent latex particles We examined the effects of cepharanthin, a known antihemolytic and membrane-stabilizing agent (Utsumi et al., 1976), and of sodium azide on the phagocytosis of preopsonized fluorescent latex particles. Phagocytic cells were exposed to each of these agents at various concentrations together with the latex particles. After incubation at 37°C for 1 h, the amount of ingested particles was determined. As shown in Fig. 4, both sodium azide and cepharanthin inhibited phagocytosis of the particles in a dose-related manner. For macrophages, sodium azide at 100 mM and cepharanthin at 0.1 mM resulted in a decrease in phagocytosis to the level of about 10 and 20% of controls, respectively. Similar results were obtained for human PMN (Fig. 4). Application to phagocytosis of fluorescence-labeled microorganisms As an example of the clinical or research application of our assay method, the phagocytosis of fluorescence-labeled yeast, E. coli B and S. marcescens kums 3958 was examined. In these experiments we employed a differential fluorescence quenching technique with methylene blue (final concentration 0.123%) to distinguish intracellular from extracellular microorganisms as described (Hed, 1977; Hed and Stendahl, 1982). Extracellular or surface-bound organisms ad-

sorbed methylene blue, which resulted in the quenching of their fluorescence. Such quenched organisms were readily distinguishable from brightly fluorescent intracellular organisms by fluorescence microscopy. As shown in Fig. 5, almost all cell-associated organisms including fluorescent yeast, E. coli B, and S. marcescens were not quenched by methylene blue, which indicates that the number of truly phagocytosed organisms was determined by our assay method. Similar to the fluorescent latex particles, phagocytosis of these microorganisms by macrophages and PMN was also dependent on temperature and opsonin. The number of ingested particles increased linearly with incubation time at 37°C, whereas phagocytosis was almost negligible at 0°C (Fig. 6). Without opsonin treatment, the uptake of these microorganisms was decreased. In addition, there was a significant difference between E. coli and S. marcescens in the uptake

(A)

(B)

6|Yeast

@Yeast

0 E.coli

=E I $.marcescens

z 3

:

Ol E.col i

Eo

z"~ [_S.marcescens

1

0

30 60 90 120 Time (rain)

0

30 60 90 120 Time (rnin)

Fig. 6. Phagocytosis of fluorescence-labeled yeast, E, coli, and S. m a r c e s c e n s by mouse peritoneal macrophages (A) or h u m a n P M N (B). Phagocytosis of opsonized (e) or nonopsonized ( © ) microorganisms at 37°C and that of opsonized (A) or nonopsonized (zx) microorganisms at 0°C by phagocytic cells were measured as described in Fig. 2 and text. Each point is the mean of duplicate values. Particle-to-cell ratios were as follows: yeast 50 : 1, E. coli 500 : 1 and S. m a r c e s c e n s 5 0 0 : 1.

182

~ .0

',B)

%

"~ 0.5

£ 10 Time (min)

20

0

10 Time (min)

20

Fig. 7. H202 release from h u m a n P M N after addition of £ coli (A) or S. marcescens (B). On addition of preopsonized (O) or nonopsonized ( O ) bacteria to the P M N suspension, fluorescent intensity of scopoletin was reduced at 37°C at the indicated time. (a,) represents the control (no addition). Data are plotted as nmol of H202 released per 1 × 106 cells.

by these phagocytic ceils. Although phagocytosis of both bacteria was temperature-dependent, the rate of phagocytosis of E. coli was greater than that of S. marcescens. After 60 min of incubation, the ingested number of preopsonized bacteria by macrophages and PMN was 98/cell and 14/cell for E. coli and 22/cell and 5/cell for S. marcescens, respectively. Relationship between H202 release and phagocytosis o f bacteria

It is well known that macrophages and PMN release H202 into the extracellular medium during the phagocytosis and that these events parallel each other (Root et al., 1975). Therefore, we measured HzO 2 release from a PMN suspension during phagocytosis of E. coli or S. marcescens. Fig. 7 shows the time course of H 202 release from PMN after the addition of opsonized or nonopsonized bacteria (cell-to-bacteria ratio was 1 : 100). When PMN suspensions were mixed with opsonized E. coli, H202 release occurred linearly after a short lag of time; however, in the case of S. marcescens only a small amount of H202 was release. These results are consistent with the differential susceptibility of E. coli B and S. marcescens kurus 3958 to phagocytosis by a factor of 4.4.

Discussion

The method presented here for the assay of phagocytosis by using fluorescence-labeled par-

ticles is simple and rapid, and it permits quantitative measurement of the rate of phagocytosis with considerable sensitivity. Michell et al. (1969) and Stossel et al. (1972) reported that it was preferable to measure the accumulation of particles within cells rather than the disappearance of particles from the incubation medium and that the initial rate of phagocytosis was important. Our method fulfills these criteria. In addition to these criteria, a crucial characteristic of a technique for measuring phagocytosis is its ability to distinguish particles absorbed onto the surfaces of phagocytic cells from particles actually internalized. In this regard, all fluorescencelabeled particles used in this study did not tend to aggregate or bind nonspecifically to cells during assay. The evidence for this contention is that phagocytosis of these particles is temperature-dependent and that there are no cell-associated particles after incubation of cells and particles at 0°C; zero time values were essentially zero (Figs. 2 and 6). In addition, appropriate metabolic inhibitors completely prevent uptake of the particles as expected (Fig. 4). Furthermore, in the case of fluorescence-labeled microorganisms such as yeast or bacteria, extracellular or surface-bound particles were readily distinguishable from intracellular particles by a differential fluorescence quenching technique with methylene blue (Fig. 5). Spectroscopic evidence, which reflected the local pH of the probe molecules, showed that the fluorescent particle was in an acidic environment phagosome (data not shown). Another advantage of the present method is related to the fluorescence-labeled particles such as bacteria and yeast, which can be easily prepared as described here. The fluorescent labeling does not require expensive materials or apparatuses for fluorescence detection and for protection from radiohazards. In addition to these fluorescencelabeled microorganisms, artificial fluorescent latex particles used in this study are commercially available. In experiments designed to test the effect of sodium azide and cepharanthin on phagocytosis, these agents greatly inhibited the phagocytosis of fluorescent latex particles depending on the concentration of the agent. Cepharanthin, a biscoclaurine alkaloid, is known as an antihemolytic

183

agent against hemolysis by snake venom because of its membrane-stabilizing activity and its ability to decrease the fluidity of the biological membrane (Utsumi et al., 1976). Therefore, these results indicate the importance of membrane fluidity in phagocytosis, in addition to that of energy supply as already reported (Karnovsky, 1962). In general, opsonization is required for effective phagocytosis. Two major opsonic substances in serum, complement and immunoglobulin, have been well documented (Verhoef et al, 1977). Results of our present studies of the opsonic effect on the uptake of fluorescent latex particles and microorganisms by macrophages or PMN agreed with previous findings. When the particles were opsonized with 50% pooled serum, phagocytosis of these particles was enhanced. When E. coli B or S. marcescens was subjected to phagocytosis, a difference in the rate of phagocytosis was found. In both PMN and macrophages, the rate of phagocytosis of E. coli B was greater than that of S. marcescens. These differences in phagocytic susceptibility may explain the clinical course of the bacterial infection. Namely, S. marcescens kums 3958 is a pathogenic isolate whereas E. coli B is not. Thus, the clinical vulnerability of the host to the bacterial infection may in part be determined at the level of phagocytosis. These results were confirmed by another experiment o n H 2 0 2 release from PMN mixed with E. coli B or S. marcescens (Fig. 7). When PMN were mixed with E. coli B at 37°C, H202 release was clearly demonstrated. However, when S. marcescens was used, only a small amount of H 2 0 2 w a s released. These results may imply that E. coli B is more easily phagocytosed by PMN than S. marcescens. Surface charge and the degree of hydrophobicity (Ohman et al., 1982) have been

considered important in determining the acceptability of particles for ingestion. Therefore, a differential susceptibility of E. eoli B and S. marcescens to phagocytic cells might be due in part to differences in their surface structure or properties that influence phagocytic cell-particle interaction. In conclusion, the present data warrant that this method is sensitive, reproducible, rapid, and accurate for the measurement of phagocytosis, and appears to be of wide applicability.

Acknowledgement We gratefully acknowledge the assistance of Masataka Nagao for the preparation of fluorescent bacteria.

References Hed, J., 1977, FEMS Microbiol. Lett. 1, 357. Hed, J. and O. Stendahl, 1982, Immunology 45, 722. Karnovsky, M.L., 1962, Physiol. Rev. 42, 143. Lowry, O.H., N.J. Rosebrough, A.L. Farr and R.J. Randall, 1951, [l. Biol. Chem. 193, 265. Michell, R.H., S.J. Pancake, J. Noseworthy and M.L. Karnovsky, 1969, J. Cell Biol. 40, 216. Nathan, C.F. and R.K. Root, 1977, J. Exp. Med. 146, 1648. Ohman, L., J. Hed and O. Stendahl, 1982, J. Infect. Dis. 146, 751. Root, R.K., J. Metcalf, N. Oshino and B. Chance, 1975, J. Clin. Invest. 55, 945. Stossel, T.P., 1975, Semin. Haematol. 12, 83. Stossel, T.P., R.J. Mason, J. Hartwig and M. Vaughan, 1972, J. Clin. Invest. 51,615. Utsumi, K., M. Miyahara, K. Sugiyama and J. Sasaki, 1976, Acta Histochem. Cytochem. 9, 59. Verhoef, J., P.K. Peterson and P.G. Quie, 1977, J. Immunol. Methods 14, 303. Wehinger, H. and M. Hofacker, 1976, Eur. J. Pediatr. 123, 125.