Computer-assisted image analysis assay of human neutrophil chemotaxis in vitro

Computer-assisted image analysis assay of human neutrophil chemotaxis in vitro

Journal oflmmunological Methods, 144 (1991) 43-48 © 1991 Elsevier Science Publishers B.V. All rights reserved 0022-1759/91/$03.50 43 JIM06099 Compu...

387KB Sizes 14 Downloads 17 Views

Journal oflmmunological Methods, 144 (1991) 43-48 © 1991 Elsevier Science Publishers B.V. All rights reserved 0022-1759/91/$03.50

43

JIM06099

Computer-assisted image analysis assay of human neutrophil chemotaxis in vitro Per Jensen and Arsalan Kharazmi Statens Seruminstitut, Department of Clinical Microbiology, Rigshospitalet, Copenhagen, Denmark (Received 12 November 1990, revised received 29 May 1991, accepted 1 July 1991)

We have developed a computer-based image analysis system to measure in-filter migration of human neutrophils in the Boyden chamber. This method is compared with the conventional manual counting techniques. Neutrophils from healthy individuals and from patients with reduced chemotactic activity were used. The cells migrating into the filter were counted automatically at depths 20 /xm apart commencing from the upper surface on the filter. The major advantages of this method are reproducibility and the counting of many fields which provides better data for statistical analysis of the results. Another advantage of the assay is that it can be used to show the migration pattern of different populations of neutrophils from both healthy individuals and patients. Key words: Neutrophil; Chemotaxis; Image analysis; Computerized method

Introduction

Polymorphonuclear leukocytes (PMN) provide the first line of host defence by virtue of their ability to mobilize and rapidly migrate from the peripheral blood through endothelial spaces and tissue matrices to sites where they eventually engulf and destroy microorganisms. The ability of cells to migrate in vivo towards an inflammatory site is a complex process that necessitates recognition of the chemoattractant and orientation and movement along a concentration gradient of the attractant (Snyderman and Pike, 1984). Defects in PMN chemotaxis have been shown to cause severe recurrent bacterial infections (Gallin, 1981;

Correspondence to: A. Kharazmi, Department of Clinical Microbiology, Rigshospitalet 7806, Tagensvej 20, DK-2200, Copenhagen N, Denmark.

Roberts and Gallin, 1983). The Boyden chamber is the most popular assay which has been developed to measure PMN chemotaxis (Boyden, 1962; Zigmond and Hirsch, 1973; Allan and Wilkinson, 1978; Wilkinson, 1982) and cell migration is measured either at the leading edge or at the lower surface of the filter. The major short-coming of the Boyden chamber technique is the necessity for time-consuming and tiring manual microscopic counting of cells in the filters. Computerbased methods for the quantitation of leukocyte chemotaxis have been described previously (Wine and Kellermeyer, 1976; Turner, 1979; Van Dyke et al., 1979; MacFarlane et al., 1987), in which either the lower surface or the leading front method are used, but not counts of cells at several levels in the membrane. Therefore, although these methods may detect certain abnormalities of cell migration they do not reveal overall changes affecting all cells. In order to detect

44 possible differences between cell populations it is necessary to determine the distribution of all cells through the entire filter. We have developed a computer-assisted image analysis method of PMN chemotaxis which is superior to previously described methods in that it is possible to measure the number of cells that have migrated into the filter at several different depths from the initial point of cell application. Furthermore, this method is able to reveal-the migration patterns of neutrophil populations with differing chemotactic characteristics.

Equipment The image analysis system as shown in Fig. 1 consisted of an optical Ergolux Leitz microscope (Leitz, F.R.G.) with an autofocusing device (Image House, Copenhagen, Denmark), a Philips L D H 600 camera (Philips, The Netherlands), a PC-Vision frame grabber video board (Image House, Copenhagen), an IBM compatible personal computer equiped with a 20 Mb harddisk (Macro-88, AT-3, E R G O system), an E G A monitor (Mitsubishi Electric), and an FX-800 Epson printer.

Software

PMNs were isolated from citrated human peripheral blood by dextran sedimentation and sodium metrizoate-Ficoll (Lymphoprep, Nyegaard, Oslo, Norway) gradient centrifugation (B¢yum, 1968). Erythrocytes were removed by hypotonic lysis and PMNs were resuspended at 1 x 106 cells/ml in Gey's balanced salt solution (GBSS), obtained from Statens Seruminstitut, Copenhagen and containing 1% human serum albumin (Statens Seruminstitut). Informed consent was obtained either from the subjects themselves or from their guardians.

The image processing software was a version of the GIPS program developed by Image House (Copenhagen) and modified to run our system including driving the autofocusing of the microscope. The image processing software was a modular system provided with a number of general image processing facilities. The analysis program permitted calculations of segmentation, labelling, perimeter, area and compactness. The system was suitable for measurements of both cell numbers and cell dimensions. The video signal was converted to digital forms which were then received by a PC-Vision frame grabber. The frame grabber was controlled by the personal computer which also controlled the autofocusing microscope processor and field scanner.

Chemotaxis

Computer count

The assay used was a modification of a Boyden chamber technique as previously described (Kharazmi et al., 1983). Briefly, the upper compartment of the Boyden chamber was filled with 0.5 ml of cell suspension separated from the lower compartment by a cellulose filter of 3 /xm pore size (Millipore, Bedford, MA, U.S.A.). The lower compartment was filled with 0.5 ml of either the chemotactic peptide f-Met-Leu-Phe (10 -5 M), obtained from Sigma, St. Louis, MO, U.S.A. or zymosan (Sigma) activated human serum (ZAS). The chambers were incubated at 37 °C for periods of 60-165 min after which the filters were removed, fixed in 96% ethanol, stained with hematoxylin, and mounted on microscope slides with the cell application side down. All experiments were carried out in triplicate.

To calibrate the microscope and the video camera i n / z m units a 1 mm objective micrometer was used. The 10 × objective, corresponding to a pixel area of 0.8990/xm 2 and a total image area

Materials and methods

Polymorphonuclear leukocytes (PMN)

:-'] Fig. 1. Schematic diagram of the computer-assisted image analysis system. 1: microscope;2: video camera, 3: autofocusing device; 4: vision frame grabber; 5: personal computer; 6: printer.

45

of 0.2358 m m 2, was used for cell counting. The cells were counted on a binary image. For each preparation five fields were chosen (one in the centre and four in the four corners). The microscope was focused visually by the operator on the surface of the filter for each field. The autofocusmg device was p r o g r a m m e d to focus down into the m e m b r a n e at 20 ~ m intervals. Five fields, each at four different depths were counted on each filter by the computer. The images were stored on the hard disk and analysed, after which the number of objects per field were printed out. Only objects within the frame or touching the lower or right boundaries were counted. Objects within an area range of 1.8-1000 izm 2 were counted. The lower range was selected to eliminate background noise from smaller objects. The upper range was chosen to make sure that all cells were counted.

Manual count A 25 x magnification objective was used for manual counting and five fields at the lower surface of the filter (equivalent to a distance of 150 I-~m) were counted. The area in each field was 0.1406 m m 2. Filters used in these experiments were incubated for 120 min at 37 ° C after which they were turned upside down and incubated for another 30 min.

Results Chemotaxis towards ZAS The data on the chemotaxis of neutrophils towards zymosan activated serum (ZAS) are shown in Fig. 2. Incubation time is plotted against the number of objects counted per field. The number of objects per field for various depths of cell migration into the filter are given. A minim u m of 60 min incubation was required for the chemotaxis assay. After 60 min there were only a few ceils detectable at a distance of 90 p.m and no cells were seen further up in the filter. The autofocusing device was p r o g r a m m e d to move every 20 /zm so that the same ceils were not counted twice. Therefore, levels of 90, 110, 130, and 150 /xm were chosen. After 75 min incubation, a large number of cells had migrated over a

200 175 150 125

o_

[]

90 pm

[]

110 i.tm

• •

130 am 150 p.m

25-

~][

o .[~- . 60 J

. 75

.

.

.

.

.

.~

.

90 105 120 135 150 165 90 165 ZAS ~ control Time (min.)

Fig. 2. Chemotaxis of neutrophils towards zymosan activated serum (ZAS) after various incubation periods. The results are presented as the number of objects per field at different levels in the filter (90-150 tzm). G e y ' s b a l a n c e d salt solution was used as a buffer control. The bars represent standard errors of the m e a n s of triplicates from one experiment.

distance of 90 # m and only very few were present in the upper levels of the filter. After 90 min incubation, cells were found at various distances beyond 9 0 / z m . The number of cells at any given distance increased with increasing incubation period in the chemotaxis assay. Some of the cells reached the opposite side of the filter (150 Izm) after 90 min of incubation. After 165 min incubation, there were a large number of cells detectable at different depths of the filter including the 150 ~ m level.

Chemotaxis towards f-Met-Leu-Phe Fig. 3 shows the results of neutrophil chemotaxis towards the chemotactic peptide f-Met-Leu-

400 [] [] • •

350

~ 300

90 IJ-m 110~m 130pro 150pro

~ 25o ,~ 20o "5

~ 150 .c o z

100

~ 50 60 i

75

90 105 120 135 150 165 90 165 fMLP i control Time (rain.)

Fig. 3. Chemotaxis of neutrophils towards f-Met-Leu-Phe (fMLP). Conditions were the same as those in Fig. 2.

46 3OO

300 [] I~

250

90 /am 110 p.m 130 lam

• •

~o~ 200

~el 250

150 p.m

200

15o

~ 100 Z

~

5o fMLP

"6

I

E

ZAS

fMLP

Patient I

15o

-~ z

1oo

~

so

ZAS

fMLP ZAS Patient 2

control

fMLP ZAS control

Figs. 4 and 5. Comparison of chemotaxis of neutrophils from two patients with reduced chemotaxis with chemotaxis of healthy subjects. The bars represent standard errors of the means of triplicates.

Phe. A pattern of cell migration similar to that towards ZAS was observed with f-Met-Leu-Phe. The front line of the cells detectable at 90 /zm was observed after 75 min incubation. The maximum number of cells reaching the 90 /zm distance was achieved after 150 min of incubation. After 105 min incubation cells were detectable at the opposite side of the filter (150 tzm).

Chemotaxis of cells from patients Experiments were carried out to compare PMN chemotaxis of normal controls with that of two patients with reduced chemotaxis. Patient 1 was a 2-year-old girl with a sustained intra-abdominal abscess and patient 2, a 1-year-old girl with bacterial pneumonia. The chemotaxis results of these patients are shown in Figs. 4 and 5. With patient

~" 20O

1, there were no detectable cells at any level of the filter in the ZAS assay whereas in the f-MetLeu-Phe assay there were only a few cells at the 150 /zm level and increasing number of cells at the upper levels of 130, 110 and 90 /zm in the filter (Fig. 4). With patient 2 the counts at 150 /zm were reduced in the f-Met-Leu-Phe assay as compared to other levels (Fig. 5).

Comparison of manual and computer counts A comparison of manual counts with computer counts are shown in Fig. 6. Using linear regression analysis it was shown that there was a good correlation between manual counting and computer counting at the 150/zm level with a regression coefficient of r = 0.97.

Discussion m

~o~ 150

100

m

[] []

= 50



i



i

,

i

100 200 300 Computer counted chemotaxis (objects/field)

400

Fig. 6. Comparison by linear regression coefficient analysis of manual with computer counting (number of objects per field) of neutrophils on the lower surface of filter membranes. (r = 0.97).

This paper describes the use of a computer-assisted image analysis system for examination of the chemotaxis of human peripheral blood polymorphonuclear leukocyte chemotaxis. Using this technique, it was possible to obtain a more refined measurement of neutrophil chemotaxis by counting the cells at various levels in the filter. Furthermore, the assay could be performed over a shorter period of time and was both less tiring and less laborious than the conventional manual methods. It is known that there are great day-today and operator variations in the analysis of

47

human neutrophil chemotaxis (Valerius, 1977). The ease in counting large areas in many fields and at different levels in the filter provides better data for statistical analysis and better reproducibility. Cell counting at different levels in the filter as suggested by Maderazo and Woronick (1978) and Howe et al. (1980) seems to be necessary in order to obtain an overall measure of cell migration. The present method fulfills this requirement. In order to establish the pattern of cell migration in the filter the chemotaxis assay was performed at different time periods ranging from 60 to 165 min. It was shown that a minimum of 75 min was required to observe the leading front at the 90 /xm level of the filter. A comparison of results using f-Met-Leu-Phe with those obtained using zymosan activated serum as chemoattractants showed that the peak chemotactic response to f-Met-Leu-Phe was earlier than that to ZAS. This may have been due to a faster equilibration of chemoattractant gradient on both sides of the filter by f-Met-Leu-Phe due to its small. In the studies with cells from patients cells it was observed that when using only the lower surface cell counting method in patient 1 the chemotactic response to ZAS was absent and that to f-Met-Leu-Phe was markedly reduced. Similarly, in patient 2 the response to both chemoattractants was reduced. In contrast with the present method a large number of cells were observed at various levels in the filter, particularly at the 90 Izm distance. This may indicate that in a given patient only certain subpopulation of neutrophils, i.e., the fastest moving cells, exhibit reduced chemotaxis whereas the others are normal. Based on these findings it may be concluded that the present method gives a more precise picture of neutrophil chemotactic activity in patients with reduced chemotaxis. The major advantage of this method is that it can be used to show the migration pattern of different populations of neutrophils with different chemotactic abilities.

Acknowledgements The expert technical assistance of Hanne Tamstorf and Anne Asanovski is appreciated.

This work was partially supported by the Danish Biotechnology Programme, Medical Immunology Centre.

References Allan, R.B. and Wilkinson, P.C. (1978) A visual analysis of chemotactic and chemokinetic locomotion of human neutrophil leukocytes. Exp. Cell Res. 111, 191. Boyden, S. (1962) The chemotactic effect of mixtures of antibody and antigen on polymorphonuclear leukocytes. J. Exp. Med. 115, 453. Boyum, A. (1968) Isolation of mononuclear cells and granulocytes from human blood. Scand. J. Clin. Lab. Invest. 21, 77. Gallin, J.l. (1981) Abnormal phagocyte chemotaxis: pathophysiology, clinical manifestations, and management of patients. Rev. Infect. Dis. 3, 1196. Howe, G.B., Swettenham, K.V. and Currey, H.J.F. (1980) Polymorphonuclear motility: measurement by computerlinked image analysis. Blood 56, 696. Kharazmi, A., Hoiby, N. and Valerius, N.H. (1983) Effect of antimalarial drugs on human neutrophil chemotaxis in vitro. Acta Pathol. Microbiol. Immunol. Scand., Sect. C. 91, 293. MacFarlane, G.D., Herzberg, M.C. and Nelson, R.D. (1987) Analysis of polarization and orientation of human polymorphonuclear leukocytes by computer-interfaced video microscopy. J. Leukocyte Biol. 41,307. Maderazo, E.D. and Woronick, C.L. (1978) Micropore filter assay of human granulocyte locomotion: problems and solutions. Clin. Immunol. Immunopathol. 11, 196. Nelson, R.D., Quie, P.D. and Simmons, R.L. (1975) Chemotaxis under agarose: a new and simple methods for measuring chemotaxis and spontaneous migration of human polymorphonuclear leukocytes and monocytes. J. Immunol. 115, 1650. Roberts, R. and Gallin, J.I. (1983) The phagocytic cell and its disorders. Ann. Allergy 51,330. Snyderman, R. and Pike, M.C. (1984) Transductional mechanisms of chemoattractant receptors on leukocytes. In: R. Snyderman (Ed.), Regulation of Leukocyte Function. Plenum, New York, pp. 1-28. Turner, S.R. (1979) ACDAS: An automated chemotaxis data acquisition system. J. Immunol. Methods 28, 355. Valerius, N.H. (1977) Neutrophil granulocyte chemotaxis in vitro. Comparison of the response to casein and bacterial chemotactic factor, and evaluation of an automatic method for counting cells on a membrane filter surface. Acta Pathol. Microbiol. Scand. Sect. C 85, 289. Van Dyke, T.E., Reilly, A.A., Horoszewics, H., Gagliardi, N. and Genco, R.J. (1979) Computerized image analysis for measuring chemotaxis in Boyden chambers. J. Immunol. Methods 31,271. Wilkinson, P.C. (1982) The measurement of leukocyte chemotaxis. J. Immunol. Methods 51, 133.

48 Wine, A.C. and Kellermeyer, R. (1976) A computer-based method for the quantitation of leukocyte chemotaxis. J. Lab. Clin. Med. 88, 487. Zigmond, S.H. and Hirsch, J.G. (1973) Leukocyte locomotion

and chemotaxis: new methods for evaluation and demonstration of a cell-derived chemotactic factor. J. Exp. Med. 137, 387.