Assessment of neutrophil chemotaxis by laser and video densitometry

Journal of hmmmologieal Method.~. 14t~( 19q2) 183- 187

183

,~ P~92ElsevierScience PublishersB.V. All rights reserved(11)22-175t)/~2/$(~5 I)11

JIM 06264

Assessment of neutrophil chemotaxis by laser

and video densitometry N.L.A. Misso ;', T.L. K a n g b. M.J. Phillips h and P.J. T h o m p s o n ;' " Department of Medicine. Unil'erslty c~f Western At~stral~a. am/ i, Department o]" Rc~pirator~"Medicine. Sir ('barb's (htlrdm'r IloffntuL Queen Elizabeth II Medical Cetltrt.. Nctlh#lds. |t~A. 600~). Australia

(Received 2(i May Igt;q,revisedreceived Ifi October Htgl. acceptedt) December 1991)

A laser densitometer and a video densitometer were used to evaluate neutrophil chemotaxis slides which were also assessed by the standard microscopic technique. A linear relationship was observed between cell number per high power field (HPF) and peak height by laser densitometry (r 2 = (I.66) or peak area by video densitometry ( r : = I).81). For wells containing more than 20 c e l l s / H P F the least variability was observed with video densitometry. Quantitative results from neutrophil chemotaxis assays performed in 48-well microchambers can be obtained rapidly and conveniently by the use of commercially available video densitemeters. Key words: Microchc,-notaxischamber: Densitometer;Polymorphenuclearleukocyte:Ouantilication

Introduction The 48-well micro-chemotaxis assembly (Falk et al., 1980) permits assays of neutrophil chemotaxis to be performed with small numbers of cells and small volumes of chemoattractants. The method commonly used to evaluate such assays is to fix and stain the filter and count all the cells in randomly selected high power fields (HPF) under the microscope but quantitation by this method is time-consuming and tedious (Wilkinson, 1982; Bignold, 1988). Recently the use of densitometry to assess chemotactic responses of tumour cells (Taraboletti et al., 1987), smooth muscle cells (Shimura et al., 1989) and granulocytes (Klomp et

Corre.ff)ondence to: N.L.A. Mi',so. University Department of Medicine, Queen Elizabeth II Medical Centre. Nedlands. W.A. 6()()9. Australia (Tel.: /61 t~) 389 34~4: Fax: (t~l 9) 389 2816).

al., 1989) has bccn reported. However, detailed analysis of commercial densitometers to assess neutrophil chcmotaxis in microchambers and the correlation of such methods with microscopic cell counting have not previously been evaluated. We report the use of the LKB 22(12 Ultroscan laser densitomcter and the Bio-Rad 620 video densitometcr to scan stained chemotaxis filters and compare this method with the standard microscopic technique.

Materials and methods N e u t r o p h i l isolation a n d c h e m o t ~ : i s assay

Blood (20 ml) was obtained from healthy volu.ateers who gave informed consent. Neutrophils 1purity > 95%, viability > 98%) wcrc isolated by ccntrifugation of hcparinised whole blood on a discontinuous Percoll density gradient (Roberts et al., 1984). Chemotaxis assays were performed

184 in 48-well micrnchcmotaxis chambers (Neuroprobe, Bethcsda. MD) using polyvinylpyrrolidone-free polycarbonatc filters with 3 # m pores (Nucleporc, Pleasanton, CA). The wells of the lower compartment wcrc filled with 26 #,1 of chcmoattractant diluted in Gey's medium. The chcmoattractants used were platelct-activating factor (PAF, Sigma Chemical Co., St. Louis. MO, 10 ~-10 5 M), Icukotrienc B 4 (a gift from Dr. J. Rokach, Merck-Frosst, Pointe Clairc-Dorval. Quebec, Canada, I0-~-I0 -5 M) and zymosanactivated serum (0, 10 ~, 10--', 10 -3 dilution). Neutrophils (5(} #1, 5 × ll}a neutrophils per well) were pipcltcd into the upper compartment and the chambers were incubated at 37°C for 30 rain in an atmosphere of 5 ~ CO_,. After incubation the cells which had not migrated were scraped from the upper surface of the filters which were thcn fixed in methanol for 10 s and allowed to dry before staining with Diff-Ouick. The filters were rinsed with distilled water, placed on a glass slide and allowed to dry prior to mounting.

Measurement of neutrophil chemotaxi.~ For microscopic assessment, five high power fields ( 1(100 × ) were randomly selected for each well and all the neutrophils counted. The data were expressed as mean ccll number per HPF. 18 slides with a total of 576 chemotaxis wells were counted on three separate occasions by the same observer. Densitometric assessment was performed by scanning the same slides with a LKB 2202 UItroscan laser densitometer (LKB, Bromma, Sweden). The scan speed was 10 r a m / r a i n and the absorbance range was set at 0 - 0 . 1 0 D units. For each slide the instrument was zeroed on a track containing no chemotaxis wells, The maximum optical density of each stained area of filter v, hich represented one chemotaxis well was measured as the peak height determined on an LKB 222(I recording integrator. The slide was scanned longitudinally along four tracks each of which contained up to eight chemotaxis wells and each slide was scanned on .hree separate occasions. The slides were also scanned or. a Bio-Rad 620 video densitometer (Bio-Rad, Richmond, CA) which was set for measurement in the reflectance mode. The slide was placed on a white board and

background absorbancc on a track which contained no chemotaxis wells was subtracted automatically. The slide was scanned along four tracks each of which contained upto eight chemotaxis wells and an absorbance peak was associated with each well. Peak shape and resolution were enhanced using a filter frequency of 2.0 lines/mm, an enhancement frequency of 0.I l i n e s / m m and a boost factor of 5. The four scans representing the data from a single slide were fed in to a personal computer, peaks were identified manually and peak areas (OD × mm) calculated using the Bio-Rad 1-D Analyst programme. Again each of the 18 slides was scanned on three separate occasions.

Data analysis and statistics For each method, three independent measurements were made on each of the 576 wells. Means, standard deviations, standard errors of the mean and coefficients of variation for each well and for groups of wells together with frequencies of distribution of the wells and correlation and linear regression analyses v, ere performed using the SAS program for personal computers. Results and discussion

The distribution of the numbers of wells over the range of cell numbers counted per HPF is shown in Fig. I A. Comparison with the distributions with respect to peak height as measured by laser densitometry (Fig. I B) and peak area by video densitometry (Fig. IC) showed reduced well numbers at the upper ends of the distributions and an increase in the numbers of wells with low peak heights and areas. Thus by microscopy there were 61} wells with less than 5 c e l l s / H P F (6.3% of the maximum number of cells counted), whereas there were 123 wells with peak heights of less than 50 (6.2% of the maximum peak height measured) and 150 wells with peak areas of less than 0.15 (6.4% of the maximum peak area measured). Calculation of the frequencies using arbitrarily defined peak height and area divisions may explain skewing of the densitometric distributions relative to microscopy. However, microscopy is not necessarily more accurate because error is introduced by the selec-

t i o n o f j u s t five H P F w h i c h m a y r e p r e s e n t o n l y 2 % o f t h e s u r f a c e a r e a o f t h e w e l l in w h i c h migrated cells may not be homogeneously diso

TSO

ca. -,

250

"'.

A

ro

. ..;.:.: . .." •

.. - . 7 ~ ; <

"

.

..

. : •.

-..,,.-

-..

..-% ",.,:

• ".-'"'.'...o~:?.: • ..~,~...-~,:'.-?..,:-:.'.".,

;-

. •.,

. ?

•

0 0

l=

10

20

30

40

50

60

70

80

B

o

s Io 15 20 25 30 25 40 4S so 5s eo ss ro

~,.s ¸

75

CelIslHPF

Q

140

2.0

•

~ _-~ ,.0

.- !-. ;.;.:-.

.~ : ' - . - • ~ .~:'.';~" "

0.5

",-

•

~

.

.,,_ e 1.5

120

""" :~'~ "~'"'" '~'~'~" "

0 10

2o o

, o

, s

, lo

, I~

, 2o

, ~

Peak zso

, ~

, 3s

, 4o

height

, 4s

. . . . so

ss

~

6~

7o

x 10

C

14a

w

120

•~ 4o. 2o IS 30 45 60 75 go 105120135150185180195210 225

Peak

area

x 10 -2

Fig. I. The distribution frequencies of the 576 wells on 18 chemotaxis slides grouped in increments of 5 cclls/HPF as assessed by microscopy (A), in increments of 50 peak height units as assessed by laser densitometry (//) and in increments of 0.15 peak area units as assessed by video densitometry (C).

20

30

Microscopy

>,

40

50

60

70

80

(cellslHPF)

Fig. 2. Regression analysis showing the correlation between peak height by laser densitometer and ceUs/HPF by microscopy (• 2 = 11.66. P < ().IX)l) (A) and between peak area by video densitnmeter and cclIs/HPF by microscopy (r" = 0.81. P < 0.0t)1) (B). The means of the triplicate measurements by each method h)r each of the 576 wells were used in the regression analysis.

t r i b u t c d ( K I o m p et al., 1989) a n d t h e d i l i g e n c c a n d f a t i g u e o f t h e o b s e r v e r will g r e a t l y i n f l u e n c e the results. Using densitometry the wells can be scanned across their diameter allowing a greater proportion of the surface area of the well to be assessed and subjective influences should be eliminated. The problems related to the use of polycarbonate membranes such as detachment of unf i x e d c e l l s ( B i g n o l d , 1989) a r e i n h e r e n t in t h e assay itself and are unlikely to cause greater error if a s s e s s m e n t is b y d e n s i t o m e t r y r a t h e r t h a n m i croscopy.

A

t20 I I

~

so

~ 6o ~ 4o 2

0

12o.

5 10 15 20 25 30 35 40 45 50 55 6o 65 7 0 > 7 5

B

100 ao

o

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

5 10 15 20 25 30 35 40 45 SO 55 60 65 70 >75

14o

c

120 I. too ~ 8o ~ 6o ~ 4o 20

,

0

,

,

,

,

,

,

,

,

,

,

,

,

,

10 15 20 25 30 35 40 45 50 55 60 65

, 70

~

>75

CelIs/HPF

Fig. 3. The mean (+ SEM) coefficients of variation for the wells grouped by cell number/HPF plotted against the same increments of cells/llPF (A) and the mean (+ SEM) coefficients of variation for the same groups of wells as assessed by laser densit~metry(B) and video densitometry(C).

By microscopy 50 of the 576 wells were determined to be 'zero wells' in that neutrophils were not detected in these wells. By laser densitometry 63 wclls had a peak height of 0 and this included

48 of the 'zero wells' by microscopy. By video densitometry 50 wells had a peak area of 0 and 46 of these were 'zero wells' by microscopy. If microscopy is regarded as accurate, laser densitometry correctly identified 96% of the 'zero wells' while the video densitometer identified 92%. The laser densitometer gave a higher false zero rate than the video densitometer (2.6% vs 0.7%7. Therefore the thresholds of both densitometric techniques for detecting low cell numbers are comparable to microscopic counting. Regression analysis of the mean peak heights determined by laser densitometry compared v~ith the mean cell number per HPF by microscopy showed a linear relationship (r z = 0.66, P < 0.001) (Fig. 2A). The mean peak areas by video densitometry also showed a linear correlation with microscopy ( r 2 = 0.81, P < 0.001) (Fig. 2B). The variability in assessment by the three methods was investigated by calculating mean coefficients of variation (CV) for the wells grouped by cell n u m b e r / H P F . The mean CVs were plotted against those increments of cell n u m b e r / H P F for microscopy (Fig. 3A) and the mean CVs for the same groups of wells as assessed by laser and video densitometry were also determined (Figs. 3B and 3C). A similar pattern of variation was seen for the three methods with a high variability in those wells with low cell numbers and a rapid decline to low variability (mean CV < 20%) as cell numbers increased. In wells containing more than 20 c e l l s / H P F the least variation was observed with video densitometry. One previous study (Klomp et al., 1989) has reported the use of densitometry to assess granuIocy.te chemotaxis in multi-well chambers. However a specially adapted transmission densitometer was used and correlation with microscopy and the variability associated with each method were not examined. The present study has shown that the use of generally available commercial laser and video densitometers gives good correlation with microscopy while variability in densitometric assessment is no greater than for microscopy. Although it does not provide an absolute measure of chemotaxis, densitometry can be used to obtain an index of relative chemotactic responses with considerable savings in time and effort.

V i d e o d e n s i t o m e t r y w o u l d a p p e a r to b c a p a r t i c u larly u s e f u l t e c h n i q u e w i t h a d v a n t a g e s o v e r cell c o u n t i n g for t h e r a p i d , c o n v e n i e n t m e a s u r e m e n t of relative chemotactic responses.

References Bignold. L.P. (10881 Measurement of chemotaxis of polymorphonuclear leukocytes in vitro. J. Immunol. Methods 1(18. I. Bignold. L.P. (1989) Use of sparse-pore polycarhonate (Nuclepore) membrane for the measurement of chemotaxis of polymorphonuelear leukocytes. J. Immunol. Methods 118, 217. Falk. W., Goodwin, R.II. and Leonard. E.J. 09811) A 48-well microchemotaxis assembly fi)r rapid and accurate measurement of leukocyte migration. J. Immunol. Methods 33, 239.

Klomp. J.P.(i.. Te Velde. A.A. and Figdor. C.G. (It;go}) Rapid densitometric delermination of cell migration and cell adhesion in a microchemotaxis chambcr. J. Immunol. Methods 118. 47. Roberts. R.L. Mounessa, N.L and Gallin. J.I. (19841 Increasing exlraccllular potassium causes calcium-dependen! shape change and facilitates concanavalin A capping in human neutrophds. J. ImmunoL 132. 2(~). Shimura. J.. Suzuki. K.. Mitsuka. M. and Umczu. K. (19~91 An automated analysis of chemotaxis in vitro using a computer assisted scanning densitometer. Anal. Biochem.

177. 72. Taraholetti. G.. Roberls. D.D. and Liotta. L.A. (lg871 Thromhospondin-induecd tumor cell migration: haptotaxis and chemotaxis are mediated by different molecular domains. J. Cell Biol. 105. 241~). Wilkinson. P.C. (1'~82)The measurement of icuk(~-'yte chemotaxis. J. Immunol. Methods 51. 133.