imaging neutrophil activation: analysis of the translocation and utilization of nad(p)h-associated...

JOUttNAL OF CELLULAR PHYSIOLOGY 152:145F156 (1992)

Imaging Neutrophil Activation: Analysis of the Translocation and Utilization of

NAD( P) H-Associated Autof luorescence During Anti body-Dependent Target Oxidation

BlNC LlANG AND HOWARD R. PETTY* Department of Biological Sciences, Wayne State University, Defroit, Michigan 48202

Fluorescence intensified/enhanced microscopy has been used to study the metabolic activation of living human neutrophils in time-lapse sequences. The autofluorescence associated with NAD(P1H’s emission band was studied within individual quiescent and stimulated cells. Excitation of NAD(P)H-associated autofluorescence was provided by a high-intensity Hg-vapor lamp. The background-subtracted autofluorescence signals were computer enhanced. In some cases the ratio image of NAD(P)H-associated autofluorescence to tetramethylrhodamine methyl ester (TRME) fluorescence, which was found to be uniformly distributed within neutrophils, was calculated to normalize autofluorescence in- tensities for cell thickness. Activation of the NADPH oxidase by phorbol myristate acetate, FF, N-formyl-methionyl-leucyl-phenylalanine (FMLP), or tumor necrosis factor (TNF) dramatically reduced autofluorescence levels. Membrane solubilization with sodium dodecyl sulfate eliminated autofluorescence. Thus, control experiments indicated that most or all of the detectable NAD(P)H-associated autofluorescence was due to NAD(P)H, consistent with previous non-microscopic studies. To understand the metabolic events surrounding the internalization and oxidative destruction of targets, we have imaged the NAD(P)H-associated autofluorescence of neutrophils and the Soret band of antibody coated target erythrocytes during cell-mediated cytotoxicity. Absorption contrast microscopy of the erythrocyte’s Soret band is an especially sensitive indicator of the entry of reactive oxygen metabolites into this target’s cytosol. Thus, it is possible to spectroscopically dissect and image the substrate (NADPH) and product (02-) reactions of the NADPH oxidase in living unlabeled neutrophils. During real-time experiments at 37”C, the level of NAD(P)H-associated autofluorescence surrounding phagosomes greatly increases before the disappearance of the target’s Soret hand. NAD(P)H-associated autofluorescence in the vicinity of phagocytosed erythrocytes is greatly diminished after target oxidation. This suggests that NAD(P)H is translocated to the vicinity of phagosomes prior to the oxidation of targets. The apparent cytosolic redistribution of NAD(P)H was confirmed by ratio imaging microscopy to control for cell thickness. We suggest that NADPH including its sources and/or carriers accumulate near phagosomes prior to target oxidation and that local KADPH molecules are consumed during target oxidation. 0 1992 Wiley-Liss, Inc.

Human neutrophils participate in host defense against bacterial, fungal, viral, and likely neoplastic disease. Although the mechanisms of neutrophil-mediated cytotoxicity are complex and incompletely under- stood, reactive oxygen metabolites are thought to play a central role (Francis et al., 1988; Clark and Klebanoff, 1975; Hafeman and Lucas, 1979; Mandell, 1974; Kle- banoff, 1980). Reactive oxygen metabolites are produced after exposure to certain non-physiological stimuli (e.g., phorbol esters) or physiological stimuli such as chemotactic factors, cytokines, aggregated IgG, and complement-treated zymosan (Nathan, 1987, 1989; Goldstein et al., 1975). During stimulation there are increases in oxygen consumption and hcxose mono- 0 1992 WILEY-LISS, INC.

phosphate shunt activity (Baldridge and Gerard, 1933; Sbarra and Karnovsky, 1959). Oxygen is consumed by a cyanide-insensitive NADPH oxidase during the respiratory burst according to the equation

NADPH + 20,-+NADP+ + 20; + H + . The production of superoxide anions (Babior et al., 1973) leads to the formation of additional reactive oxygen metabolites such as singlet oxygen, hydroxyl radi-

Received November 4,1991; accepted January 31,1992. *To whom reprint requestsicorrespondence should be addressed.

146 LIANG AND PETTY

Fig. 1. Experimental apparatus used to image NAD(P)H-associated autofluorescence in living cells. A 500 W Hg-vapor lamp with a quartz condenser provided the excitation light. The excitation light was shut- tered (1) and on occasion passed through neutral density filters (2). To remove IR radiation a water-filled cell with quartz windows was placed in the light path (3). Omega optical filters were placed at position 4. A KG-5 IR filter protected the microscopic system from a

cals, and HOCl (Iyer et al., 1961; Weiss et al., 1977; Tauber and Babior, 1977; Krinsky, 1974). The cyanide- insensitive NADPH oxidase is assembled at plasma or phagolysosomal membranes during triggering from several components including a flavoprotein, a cy- tochrome b 558, and the p47 and p67 cytosolic factors. Oxidase activity may be influenced by degranulation and protein kinase C activity (Wymann et al., 1987).

Although recent technological advances in optical microscopy have allowed the visualization of pH gradients (Bright et al., 1987), Ca2+ gradients (Sawyer et al., 1985), the transmembrane proximity of receptors and microfilaments (Zhou et al., 19911, and target oxidation and cytolysis (Francis et al., 1988; Petty et al., 1989, 1992), i t has not been possible to visualize metabolic changes within cells during triggering. To better understand the complex web of events surrounding neutrophil-mediated target destruction, we have developed a microscopic analog of Chance’s NAD(P)H detection methods (Chance et al., 1962,1965). This experimental approach allows NAD(P)H-associated autof luorescence of living cells to be monitored in real time. To correlate cell metabolic activity with target cell oxidation, we have combined NAD(P)H-associated flourescence imaging with Soret band (absorption-contrast) transmitted light microscopy, which detects the entry of oxidants into target erythrocytes (Francis et al., 1988).

failure of element 3. A 350 DF50 filter selected the excitation light. A 419 nm long-pass dichroic mirror reflected the light t o a quartz objee tive (5). The fluorescence emission was reflected through a 419 nm long-pass filter (6) to a SIT camera (7). To detect changes in the erythrocyte’s Soret band, a 430 nm filter (8) was placed in front of a CCD camera (9). The cells were held at 37°C by a heatingkooling device (10).

This allows us to simultaneously monitor both physiological parameters of interest. In addition to the technological contributions discussed below, we show the formation of spatial gradients and large concentration fluctuations of NAD(P)H-associated autofluorescence during neutrophil activity.

MATERIALS AND METHODS Materials

Phorbol myristate acetate (PMA; Sigma Chemical Co., St. Louis, MO) was stored frozen in dry DMSO a t 4 x lop5 M. Sodium fluoride was also obtained from Sigma. Tumor necrosis factor (TNF; Cetus Corp., Em- eryville, CA) was diluted to lo4 units/ml prior to experiments. Tetramethylrhodamine, methyl ester, perchlo- rate (TRME) was obtained from Molecular Probes (Eugene, OR).

Neutrophil preparation Peripheral blood neutrophils were prepared from

blood obtained from a finger prick or venipuncture using heparinized tubes. Several drops of fresh blood obtained from a finger prick were placed on glass coverslips. Samples were incubated in a humidified incubator a t 37°C for 30 min. The coverslips were then rinsed with Hank’s balanced salt solution (HBSS; Gibco, Grand Island, NY) to remove clots. Adherent

IMAGING NEUI'ROPHIL ACTIVATION

cells remained attached to the coverslips. Clot prepara- tions contained 92-95% neutrophils, <3% eosinophils, and 4-896 monocytes as judged by Giemsa staining. Large quantities of neutrophils were isolated from venous blood using Ficoll-Hypaque solutions (Sigma) and step-density gradient centrifugation. Trypan blue staining indicated that 95-99% of these cells were viable.

Phagocytosis Sheep erythrocytes (RBCs) (Cleveland Scientific,

Cleveland, OH) at 10' cellsiml were opsonized with a rabbit anti-sheep RBC IgG fraction (Cappel Labs., Mal- vern, PA) at a dilution of 112,000 in HBSS for 15 min. at 37°C. Dilutions of less than 111,000 resulted in target aggregation. Opsonized RBCs were washed twice with HBSS. The targets were then added to washed adherent neutrophils in HBSS for 5 to 15 min at 37°C. The mono- layers were washed with HBSS, then placed on a heated microscope stage.

Fluorometry Steady-state fluorescence spectroscopy measure-

ments were performed with a Perkin-Elmer (Norwalk, CT) MPF-66 fluorometer linked to a Perkin-Elmer model 7300 computer utilizing PELCS-I11 software. Typical experimental settings were excitation a t 350 nm and emission at 460 nm with slit widths of 5 nm. Inner filter effects were found to be important at higher NADPH concentrations. To avoid this artifactual decrease in emission intensity, the path length was mini- mized using 2 mm 0.d. Raman scattering capillaries. The capillaries were held in place by 1 cm x 1 cm blocks. Each experiment was comprised of fifteen measurements at various rotational positions of the capil- lary.

OPTICAL MICROSCOPY Transmitted light microscopy

Optical microscopy was performed with an auto- mated Zeiss axiovert microscope (Carl Zeiss, New York, NY) interfaced to a Perceptics Biovision system (Knox- ville, TN). Differential interference contrast (DIC) and bright field microscopy were performed with conven- tional Zeiss optics. Soret band absorption-contrast transmitted light microscopy was performed as previously described (Francis et al., 1988; Petty et al., 1989). A 430110 nm discriminating band pass filter was inserted in the light path in front of a charge-coupled device (CCD) camera (model 72; Dage-MTI, Michigan City, IN). Figure 1 illustrates the insertion of the 430 nm filter. In some cases a 430 nm band-pass filter was inserted between the light source and condenser. A NA = 0.55 condenser was used to allow the insertion of a heatingicooling apparatus. This apparatus held samples at a nominal temperature of 37°C.

NAD(P)H-associated autofluorescence microscopy

NAD(P)H-associated autofluorescence was imaged by epi-fluorescence microscopy. Figure 1 illustrates the important points regarding the microscopic system. EX- citation was provided by a 500 W Oriel Corp. (Stratford, CT) quartz Hg lamp housed in a model 6601 housing.

147

C

Distance

Fig. 2. Fluorescence microscopy of TRME-labeled neutrophils. Cells were labeled with 10 pgiml TRME in HBSS for 15 min at 37°C. A, B: DIC and fluorescence micrographs, respectively, of a representative cell population labeled with TRME. Line profile analyses of cells labeled 1 ,2 , and 3 in B are shown in traces 1,2 , and 3 in c, respectively. The length of each line profile trace in panel C is about 10 pm. These data show that TRME is uniformly distributed within neutrophils. A and B: X650.

The light was collected by a quartz condenser. The lamp was attached t o an Oriel model 68810 power supply. Ozone produced by the lamp was removed from the laboratory by filtration through activated charcoal. UV safety goggles were worn during all experiments. Infra- red radiation was removed from the excitation light using a water-filled optical filter with quartz windows. A KG-5 filter (Omega Optical, Brattleboro, VT) provided additional protection for the sensitive optical ele- ments. Excitation light was selected using a 350150 nm discriminating band-pass filter. The excitation light was reflected to the sample using a 419 nm long-pass dichroic mirror. Infinity-corrected 4 0 ~ and 100 x Zeiss neoflur objectives and a quartz 100 x (non-infinity-corrected) Zeiss objective were used in these studies. The autofluorescence was selected using a 419 nm long-pass

148

' O 0 1

LIANG AND PETTY

A

Emission Wavelength (nm)

Fig. 3. Fluorescence spectra of NADPH. A: Emission spectra of 0.1 mM NADPH (trace a) and NAl)P+ (trace b) in PBS using a n excitation wavelength of 350 nm and 5 nm slit widths. Although NADPH is highly fluorescent, NADP' displays no significant fluorescence. B: NADPH concentration dependence of the fluorescence emission spectra. Spectra at concentrations of 0.30,0.25,0.20,0.15,0.10,0.05, and 0.01 mhl are illustrated in trace8 a-g, respectively.

filter. Autofluorescence images were acquired by a Dage-MIT model 66 SIT camera.

Typical experiments were conducted using a mini- mum of a 400 W lamp setting. The Oriel Hg lamp was replaced every three to six months. To avoid deliterious effects of UV radiation on cells, the samples were briefly illuminated to collect images. The cumulative

UV exposure of cells during the approximately 60 min observation period of most experiments was <60 sec.

Image acquisition and processing The gain and offset of the CCD and SIT cameras'

video signals were adjusted in analog mode and after digitization by the image processor. Background im-

IMAGING NEUTROPHIL ACTIVATION 149

60

40

20

0 0 . 0 5 0 . 1 0 0 . 1 5 0 . 2 0 0 . 2 5 0.30 0 . 3 5

NADPH (mM) Fig. 4. Quantitative relationship between NADPH fluorescence and concentration. The relative fluorescence intensity (arbitrary units) is shown at the ordinate. The concentration of NADPH (mM) is given at the abscissa. The data points represent the mean ? SEM for 15 measurements of a representative experiment (n = 3). The correlation coefficient for this experiment was 0.999.

ages for both the CCD and SIT cameras were recorded from regions of the slide devoid of cells. The background images were subtracted from the raw digitized images. Due to the low level of autofluorescence, 256 images were averaged during each 8.5 sec observation period. The background-subtracted images were stored on hard disk, streaming tape, or a Sony model 5850 video recorder. Fluorescence images were generally scaled over a broader pixel intensity range. Relative differences in fluorescence intensity were also assessed with the Bio- vision software.

Image ratioing Image ratioing microscopy was performed to pro-

vide additional confirmation that cytoplasmic thickness was not responsible for the spatial variation of NAD(P)H-associated autofluorescence. Neutrophils were labeled with TRME a t 10 pgiml in HBSS for 15 min at 37°C. Figure 2 shows data validating the ability of TRME to uniformly distribute within neutrophils. Panels A and B show representative DIC and fluorescence micrographs of TRME-labeled neutrophils. Line profile analysis of three cells at the center and right- hand sides of panels A and B is given in panel C. Simi- lar results were obtained in three separate experiments. Our data show that during these conditions the TRME label was uniformly distributed within cells; therefore, ratio images formed using TRME correct for cell thickness.

NAD(P)H-associated autofluorescence images were obtained as described above. TRME was chosen because

it uniformly distributes within neutrophils, its spectral properties are distinct from NAD(P)H and independent of cytoplasmic pH and Ca2* levels, and i t is comparatively resistant to 0;-mediated destruction. Images of TRME were collected using a 540 DF 20 nm excitation filter, a 560 nm long-pass dichroic mirror, and a 590 DF 30 nm emission filter. TRME images were acquired within 30 sec of the NAD(P)H-associated autofluorescence image; no change in cell morphology was observed during this period. The ratio image NAD(P)H- associated autofluorescenceiTRME was calculated using the Perceptics floating point processor. The photomicrographs reported were obtained using a Polaroid (Boston, MA) freeze-frame video recorder and Kodak Plus-X pan ASA 125 film.

RESULTS Properties of NAD(P)H

In this study we exploit certain physical attributes of reduced pyridine nucleotides and hemoglobin to map biochemical changes accompanying the oxidative destruction of erythrocytes. Previous studies by Chance and co-workers have shown that the autofluorescence of NADH and NADPH can be used to monitor the metabolic state of living tissues from various species (Chance and Jobsis, 1959; Chance and Thorell, 1959; Chance et al., 1962, 1965; Barlow and Chance, 1976). This approach has been recently extended to flow cytometry (Hafeman et al., 1982). Figure 3A shows the excitation and emission spectra of NADPH and NADP+. Although NADPH is highly fluorescent,

150 LIANG AND PETTY

Fig. 5. NAD(P)H-associated autofluorescence of neutrophils. DIC (panels A,C,E,G,I,K) and fluorescence (panels B,D,F,H,J,L) micrographs of neutrophils are shown. A and B show neutrophils incubated in buffer alone for 20 min at 37°C. C and D show neutrophils exposed to 0.1% SDS. The disruption of cell morphology was accompanied by the loss of autofluorescence. When neutrophils are treated with acti-

NADP' displays no detectable fluorescence. This differential was exploited by our imaging methods. Very similar spectroscopic results were obtained using NADH and NAD' (data not shown). Figure 3B shows the fluorescence emission spectra of NADPH at various concentrations from 0.01 to 0.30 mM. This includes the intracellular concentration range associated with resting and phagocytosing neutrophils (Patriarca et al., 1971). At NADPH concentrations greater that 0.2 mM, inner filter effects were found to affect the measured emission intensity. As described in Materials and Methods, this was eliminated using high quality glass capillaries which have very short path lengths. Figure 4 shows the fluorescence intensity at an emission wavelength of 450 nm plotted against NADPH concentration. This indicates that the fluorescence intensity is directly proportional to NADPH concentration.

Most or all of the autofluorescence of neutrophils at 450 nm is due to reduced pyridine nucleotides (Hafe- man et al., 1982). During resting conditions the concen-

vators of the NADPH oxidase (30 nM PMA, panels E, F; 20 m M NaF, panels G, H; 100 nM FMLP, panels I, J; and 400 Uiml TNF, panels K, L) their level of autofluorescence is greatly diminished. The inhibition of autofluorescence was greatest for SDS, PMA, and NaF. All fluorescence images were processed in an identical fashion to facilitate com- parison. ~ 2 7 5 .

tration of cytosolic NADPH is 4.43-fold higher than NADH levels (Patriarca et al., 1971). Although this ratio decreases during phagocytosis, the amount of NADH remains constant (Patriarca et al., 1971). Since earlier experiments have shown that alterations in the level of NAD(P)H during neutrophil phagocytosis represent changes in NADPH concentration, we suggest that most of the differences reported below are due to changes in NADPH concentration.

Microscopy of NAD(P)H-associated fluorescence Figure 5 shows representative DIC and autofluores-

cence micrographs of human neutrophils. The cells were allowed to adhere to coverslips for 20 min at 37°C. As the DIC micrograph of Figure 5A shows, a focal plane near the basal surface was chosen. Panel B shows the corresponding NAD(P)H-associated autofluorescence photomicrograph. As this micrograph shows, non-adherent (spherical) cells are brighter while the adherent cells are dimmer. The fluorescence is uni-

IMAGING NEUTROPHIL ACTIVATION 151

formly distributed throughout the cells’ cytosols. This is not surprising since a small molecule such as NADPH should be able to rapidly diffuse throughout the cytosol.

If the autofluorescence micrograph of Figure 5 represents the concentration of reduced pyridine nucleotides, it should be possible to alter the level of autofluorescence by treating the cells with appropriate reagents, including activators of the NADPH oxidase. Figure 5C,D shows ncutrophils treated with 0.1% SDS for 20 rnin a t room temperature. As expected, the cell morphology is dramatically changed because of membrane solubilization. This treatment should allow small cytoplasmic metabolites such as NAD(P)H to escape. When viewed by fluorescence microscopy, no autofluorescence was observed. This supports the idea that the autofluorescence at this wavelength is due to small dissociable cytoplasmic components. Additional evidence supporting the NAD(P)H origin of the neutro- Phil’s autofluorescence at this wavelength can be found by adding reagents which are known to alter NAD(P)H concentrations. Figure 5E-L shows neutrophils exposed to PMA, F-, FMLP, and TNF, four reagents known to promote oxidation of NADPH to various de- grees. PMA was added to cells at a final concentration of 4 x lop8 M for 10 min a t 37°C. The autofluorescence of the cells was dramatically reduced, although a few fluorescent non-adherent cells could be found (panels E and F). Neutrophils were also treated with fluoride ions which are known to trigger superoxide production (Curnutte et al., 1979). Cells were treated with 20 mM NaF for 30 min a t 37°C followed by optical microscopy. Figure 5G,H shows DIC and NAD(P)H autofluorescence photomicrographs of neutrophils treated with NaF. During these conditions, the neutrophils’ autofluorescence at 450 nm nearly disappears. FMLP was added to cells at a final uniform concentration of lop7 M for 30 min at 37°C. These treatments dramatically decreased the levels of neutrophil autofluorescence. TNF is a moderate activator of superoxide production (Yuo et al., 1989). When incubated with neutrophils at 400 Uiml, the cells readily spread on glass surfaces. A modest reduction in autofluorescence is observed (Fig. 5K,L). These dramatic changes cannot be easily ac- counted for by changes in cell shape. However, to pro- vide definitive evidence we have performed ratio imaging microscopy. Figure 6 shows representative ratio images of NAD(P)H-associated autofluorescence/ TRME of neutrophils. Since TRME distributes uniformly throughout neutrophils under the conditions described above, the NAD(P)H-associated autofluor- escenceiTRME ratio removes any optical effects due to cell thickness. Resting cells displayed uniform intracellular NAD(P)H-associated autofluorescence levels (Fig. 6A,B). Although the total amount of autofluorescence was decreased after NaF treatment, the residual autofluorescence was distributed in a punctate fashion. Identical results were obtained during three separate experiments. Ratio imaging microscopy was especially useful in detecting NAD(P)H-associated autofluorescence in the hyaloplasmic region. The punctate autofluorescence in the perinuclear region and hy- aloplasm frequently correspond to granular structures seen in DIC (Fig. 6C,D). However, since the auto-

Fig. 6. Ratio imaging microscopy of neutrophils. DIC (panels A,C,E), NAD(P)H-associated autofluorescence/TRME ratio images (panels B,D,G), and Soret (panel F) photomicrographs are shown. Neutrnphils were labeled with TRME, then incubated with buffer (A,Bj, 20 mM NaF (C,D), or IgG-opsonized sheep erythrocytes (E-GI. Ratio imaging microscopy corrects for the effects of cell thickness. Although resting cells have a uniform distribution of NAD(P)H-associated autofluorescence, exposure to NaF or IgG-opsonized erythrocytes triggers the formation nf autnfluoresrent clusters. Images in B and G are directly comparable; the ratio image in D required greater autoscaling. x 700.

fluorescence level is low, we cannot exclude the possibility that it is due to trace quantities of non-NAD(P)H fluorochromes. Therefore the microscopic NAD(P)H- associated autofluorescence imaging methods, in anal- ogy with its macroscopic and flow cytometric counter- parts (Chance et al., 1962,1965; Hafeman et al., 19821, primarily or only detect the redox state of the neutrophils’ NADH and NADPH pools.

Antibody-dependent phagocytosis triggers large and localized changes in NAD(P)H-associated

fluorescence intensity that correlate with target oxidation

In the preceding paragraphs we have characterized the ability of fluorescence microscopy to detect the metabolic activation of neutrophils by imaging their NAD(P)H-associated autofluorescence. Previous studies in this laboratory (Francis e t al., 1988; Petty et al., 1989) have shown that the entry of oxidative molecules into target erythrocytes can be conveniently monitored by imaging the Soret band at its 430 nm edge using transmitted light microscopy. We have combined these two methods to spatially and temporally resolve the metabolic changes in neutrophils that accompany the oxidative destruction of targets. Figure 7 shows two sets of low (panels A-C) and high (panels D-F) magnification images of neutrophils incubated with antibody- coated RBCs. Each set contains a DIC, a transmitted light Soret absorption-contrast image, and an NAD(P)H-associated autofluorescence image of cells. Neutrophils and targets were incubated at 37°C for 20 min, then transferred to a microscope stage held at 37°C. The neutrophils’ autofluorescence is distributed throughout the cytoplasm in a non-uniform fashion. In general the autofluorescence intensity is enhanced in the vicinity of targets, often appearing punctate (Fig.

152 LIANG AND PETTY

Fig. 7. Autofluorescence of neutrophils after phagocytosis. Neutro- phils were incubated a t 37°C in the presence of IgG-opsonized sheep erythrocytes. A-C: Corresponding DIC, Soret band transmitted light, and autofluorescence micrographs of neutrophils. The autofluorescence is non-uniformly distributed in the neutrophil. It is often found enhanced or clumped near phagocytosed erythrocytes. This is also

seen in the higher magnification DIC, Soret, and fluorescence micrographs of D-F. This example, which demonstrates a moderate level of clumped autofluorescence, shows fluorescence at the perimeter of two phagocytosed erythrocytes and some small clumps near a phagolyso- some (arrows). A,B,C, X275; D,E,F, x700.

7C,Fj. Identical results were obtained on five separate occasions using ratio imaging microscopy (Fig. 6E-GI. Figure 8 shows a time sequence of Soret and NAD(P)H- associated autofluorescence micrographs of neutrophils and antibody-coated erythrocytes during antibody-dependent cellular cytotoxicity a t 37°C. The corresponding Soret and NAD(P)H-associated autofluorescence micrographs are shown in adjacent panels. This series of micrographs was collected at 10 min intervals of time. The time scale of these observations match the 15-90 min required to trigger the oxidase of adherent neutrophils (Nathan, 1987, 1989). As these micrographs show, the level of NAD(P)H-associated autofluorescence varies greatly in the vicinity of phagosornes. Over the time course of these studies, cytosolic fluorescence including the fluorescence surrounding phagosomes increased prior to target RBC oxidation followed by a rapid fall in fluorescence after target oxidation.

The results described above are due to physiological changes of the cells, not photochemical artifacts. Firstly, changes in autofluorescence parallel the changes in NADPH levels brought about by exposure to exogenous substances (Figs. 5,6). Furthermore, if illumination was photochemically changing cells, then the illuminated cells should all behave identically (e.g., they should all photobleach). This cannot be true since we often observe different cells in the same microscopic field brighten and dim simultaneously (Fig. 8). We have observed that prolonged illumination does not alter the Soret band of erythrocytes. To eliminate any potential problem, cells were illuminated intermit- tently for a few seconds during these experiments. Cells were exposed to <60 sec of cumulative UV illumination

in each series of experiments. Therefore, we believe that these changes in cellular autofluorescence represent largely changes in NAD(P)H concentration brought about by cell triggering.

Quantitative data from combined Soret and NAD(P)H- associated autofluorescence experiments are given in Figure 9. This figure shows ten separate experiments on the oxidation of target erythrocytes (Fig. 9A-J). The percentage of cells with a dark (non-oxidized) Soret band is plotted at 5 time points spaced 10 min apart. The percentage of cells with bright autofluorescence near targets is also plotted in each panel. The entire figure represents kinetic data from 250 target cells. The targets’ Soret bands disappear as cytoplasmic NAD(P)H-associated autofluorescence levels decrease (e.g., NADPH oxidation). The best target responses were obtained when the cytosolic fluorescence levels began a t higher levels. Lower initial fluorescence levels gave smaller changes in Soret absorption. Some heterogeneity in these data was expected since we previously mentioned that the autofluorescence level often rises prior to target oxidation and the oxidative events are asynchronous. However, as these quantitative data show, the autofluorescence levels fall as the targets are oxidized, thus correlating these two phenomena. Similar results were obtained in twelve additional experiments performed at various time scales. We suggest that the level of NADPH near the perimeter of phagosomes rises prior to target oxidation, then falls after oxidation.

DISCUSSION The studies described in this paper quantitate the

temporal and spatial relationships between NAD(P)H-

IMAGlNG NEUTROPHIL ACTIVATION

Fig. 8. Temporal analysis o f target oxidation and effector autofluorescence changes. Combined Soret band transmitted light microscopy (A,C,E,G) and NAD(P)H-associated autofluorescence (B,D,F,H) images are shown. Adjacent panels represent the same point in time. This series o f micrographs was collected at 10 min intervals of time. As this sequence o f pictures shows, autofluorescence increases prior to target oxidation and falls after target oxidation (n = 22) X800.

153

154

40

20

L U N G AND PETTY

100 - 80 - 60 - 40 -

100

80

60

40

20

0

100

80

60

40

’ 20

80 4

20 :l!?A+ 0

100

40

20

0 1

E loo I 8 0 ] 60 \

0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6

100 1 F 100

80

60

40

20

0 1 0 1 2 3 4 5 6 0 1 2 3 4 5 6

Fig. 9. Quantitative summary of combined Soret band transmitted light and NAD(P)H-associated autofluorescence studies. A J : Ten different experiments each consisting of 11-27 target erythrocytes. The total number of targets shown in this figure is 250. For each panel

associated fluorescence intensity and target cell oxidation. To our knowledge these data represent the first photomicrographs of NAD(P)H-associated autofluorescence in cells. The reorganization of cytoplasmic NAD(P)H-associated autofluorescence observed in this study likely corresponds to the neutrophil’s metabolic “activation.” Previous studies by Chance and col- leagues (Chance and Jobsis, 1959; Chance and Thorell, 1959; Chance et al., 1962, 1965; Barlow and Chance, 1976) and others (Thorell, 1980; Hafeman et al., 1982) have defined the specificity and applicability of NAD(P)H fluorescence in cell and tissue characterization. By exploiting recent developments in optical materials, image detection, and computer-based processing, we have been able to image the comparatively low-level autofluorescence associated with this spectral region. To ensure that the fluorescence emission following excitation at 320 to 380 nm was due to NAD(P)H, we performed several types of control studies. When the neutrophil’s cytosol was extracted by detergent treatment, no autofluorescence could be detected. This supports the idea that diffusible or extractable pro- toplasmic components were responsible for the autofluorescence emission. Certain reagents such as PMA, FMLP, and F- are known to trigger the neutrophil’s respiratory burst and NADPH oxidase activity. Conse-

loo 1 I

80

40

20

0 L

J

( 1 0 1 2 3 4 5 6 0 1 2 3 4 5 6

the percentage of cells showing a dark (non-oxidized) Soret band ( 0 ) and bright (NAD(P)H-associated) autofluorescence ( 0 ) are given at five time points spaced 10 min apart. In general, targets are oxidized as NAD(P)H-associated autofluorescence disappears.

quently, the NADPH level falls in response to stimulation. When cells are examined during treatment with PMA, FMLP, or F-, their autofluorescence is dramatically reduced. These results were confirmed by ratio imaging microscopy, which controls for cell thickness. Taken together, these and earlier studies indicate that the autofluorescence of neutrophils due to excitation a t 320-380 nm is largely or exclusively due to NAD(P)H.

The accumulation of NAD(P)H-associated autofluorescence about the perimeter of phagosomes represents a previously unrecognized step in neutrophil-mediated cytotoxicity. Both NAD(P)H-associated autofluorescence microscopy and ratio imaging microscopy showed the punctate accumulation of autofluorescence in the vicinity of phagosomes (Figs. 6-8). This stimulus-dependent accumulation of NAD(P)H-associated autof luorescence has been unambiguously linked to the func- tional activity of the NADPH oxidase, the deposition of reactive oxygen metabolites in targets, using Soret band microscopy. The Soret band of hemoglobin disappears upon exposure to superoxide anions (Francis et al., 1988). Therefore, i t is possible to temporally and spatially correlate cytosolic NAD(P)H-associated autofluorescence levels with the entry of oxidants into targets. As Figure 8 shows, the NAD(P)H-associated autofluorescence levels in the vicinity of phagosomes

155 IMAGING NEUTROPHIL ACTIVATION

rises prior to target oxidation followed by a dramatic decrease. Since the fall in NAD(P)H-associated autofluorescence concentration correlates with target oxidation, we propose that local NADPH molecules are consumed by the NADPH oxidase in the production of superoxide anions. The superoxide anions, which are released at the trans or internal face of the phagosome membrane, enter the target erythrocyte where they ox- idize hemoglobin.

Using Soret band microscopy and antibody-coated sickle cells as targets, we have recently shown that reactive oxygen metabolites can enter targets in an asymmetric (or unidirectional) fashion (Petty et al., 1992). As previously mentioned, the NAD(P)H-associated fluorescence is frequently found as aggregates near phagosomes. The asymmetric distribution of NAD(P)H about phagosomes may thus account for the asymmetric entry of oxidants into targets.

The biochemical mechanism of NAD(P)H-associated autofluorescence accumulation near phagosomes has not been addressed in this study. The increase in NAD(P)H-associated autofluorescence intensity could be due to the recruitment of NAD(P)H sources, NAD(P)H, and/or NAD(P)H binding proteins. Since the diffusion coefficient of small molecules such as NAD(P)H in the focal plane is roughly lop5 cm’isec, the “non-specific” accumulation of NAD(P)H and the col- lection of NAD(P)H sources near the phagosome do not seem physically compatible with our observations. There are no apparent diffusion gradients during target oxidation; these autofluorescence patterns are sharp and punctate and therefore do not show the gradients expected for diffusion by line profile analysis. Electron microscopic studies have not shown the recruitment of specific organelles about the perimeter of phagosomes (Oliver, 1978). The simplest self-consistent picture emerges by proposing the NADPI-I carrying proteins are recruited to the vicinity of phagosomes. Such proteins may be detected by studying the pro- tease-sensitive solvent effects on the NADPH emission spectrum of cytosolic extracts. Recently, Umei et al. 11991) have reported an NADPH-binding protein par- ticipating in the respiratory burst. It is a speculative possibility that we have imaged the trafficking of this protein in living neutrophils using methods described above.

The intracellular signaling pathways that lead to recruitment of NAD(P)H-associated autofluorescence to phagosomes are not known. Ca2+ fluctuations may be required for both degranulation and the assembly of the NADPH oxidase (Jaconi et al., 1990; Lew et al., 1984). Furthermore, Ca2+ is found about the perimeter of target particles during and after phagocytosis (Sawyer et al., 1985). These correlations suggest that Ca2+ may play a role in recruiting NAD(P)H-associated autofluorescence to phagosomes.

Although this study has focused upon imaging metabolic changes within neutrophils during cell triggering, it has numerous potential applications. For example, i t could be used to follow the metabolic program of individual cells during development. Moreover, the entry of toxic oxygen metabolites into tumor cells may play a central role in their destruction andlor mutagenesis (Clark and Klebanoff, 1975; Hafeman and Lucas, 1979;

Weitzman and Stossel, 1981). The redox states of both the effector’s and target’s cytosols could be simultaneously monitored with this approach.

ACKNOWLEDGMENTS This work has been supported by NIH grant A K A

27409. LITERATURE CITED

Babior, B.M., Kipnis, R.S., and Curnutte, J.T. (1973) The production by leukocytes of superoxide, a potential bactericidal agent. J. Clin. Invest., 52t741-744.

Baldridge, C.W.: and Gerard; R.W. (1933) The extra respiration of phagocytosis. Am. J. Physiol., 103:235-236.

Barlow, C.H., and Chance, B. (1976) Ischemic areas in prefused rat hearts: Measurement by NADH fluorescence photography. Science, 193r909-910.

Bright, G.R., Fisher, G.W., Regowska, J., and Taylor, D.L. (1987) Fluorescence ratio imaging microscopy: Temporal and spatial measurements of cytoplasmic pH. J. Cell Biol., 104t1019-1033.

Chance, B., and Jobsis, F. (1959) Changes in fluorescence in a frog sartorius muscle following a twitch. Nature, 184t195-196.

Chance, B., and Thorell, B. (1959) Localization and kinetics ofreduced pyridine nucleotide in living cells by microfluorometry. J. Biol. Chem., 234t3044-3050.

Chance, B., Cohen, P., Jobsis, F., and Schoener, B. (1962) Intracellular oxidation-reduction states in vivo. The microfluorometry of pyridine nucleotide gives a continuous measurement of the oxidation state. Science, 137:499-508.

Chance, B., Williamson, J.R., Jamieson, D., and Schoener, B. (1965) Properties and kinetics of reduced pyridine nucleotide fluorescence of the isolated and in vivo rat heart. Biochem. Z., 341:358377.

Clark, R.A., and Klebanoff, S.J. (1975) Neutrophil-mediated tumor cell cytotoxicity: Role of the peroxidase system. J . Exp. Med., 141.1442-1447.

Curnutte, J.T., Babior, B.M., and Karnovsky, M.L. (1979) Fluoride- mediated activaton of the respiratory hurst in human neutrophils. J. Clin. Invest., 63:637-647.

Francis, J.W., Boxer, L.A., and Petty, H.R. (1988) Optical microscopy of antibody-dependent phagocytosis and lysis of erythrocytes hy living normal and chronic granulomatous disease neutrophils: A role of superoxide anions in extra- and intra-cellular lysis. J. Cell. Phys- iol,, I35:l-12.

Goldstein, I.M., Roos, D., Kaplan, H.B., and Weissmann, G. (1975) Complement and immunoglobulins stimulate superoxide production by human leukocytes independently of phagocytosis. J. Clin. Invest., 56t1155-1163.

Hafeman, D.G., and Lucas, Z.L. (1979) Polymorphonuclear leukocyte- mediated, antibody-dependent, cellular cytotoxicity against tumor cells: Dependence on oxygen and the respiratory burst. J. Immunol., 123.5542.

Hafeman, D.G., McConnell, H.M., Gray, J.W., and Dean, P.N. (1982) Neutrophil activation monitored by flow cytometry: Stimulation by phorbol diester is an all-or-none event. Science, 215:673-675.

lyer, G.Y.N., Islam, M.F., and Quastel, J.H. (1961) Biochemical as- pects of phagocytosis. Nature, 192r535-541.

Jaconi, M.E.E., Lew, D.P., Carpentier, J.-L, Magnusson, K.E., Sjogren, M, and Stendahl, 0. (1990) Cytosolic free calcium elevation mediates the phagosome-lysosome fusion during phagocytosis in human neutrophils. J . Cell Biol., 110;1555-1564.

Klebanoff, S.J. (1980) Myeloperoxidase-mediated cytotoxic systems. In: The Reticuloendothelial System: A Comprehensive Treatise. A.J. Sbarra and R.R. Strauss, eds. Plenum Press, New York, pp. 279-308.

Krinsky, N.I. (1974) Singlet excited 0, as a mediator of the antibacte- rial action of leukocytes. Science, 186:363-365.

Lew, D.P., Wollheim, C.B., Waldrogel, F.A., and Pozzan, T. (1984) Modulation of cytosolic free calcium transients by changes in intracellular calcium-buffering capacity: Correlation with exocytosis and 0, production in human neutrophils. J. Cell Biol., 99:1212-1220.

Mandell, G.L. (1974) Bactericidal activity of aerobic and anaerobic polymorphonuclear neutrophils. Infect. Immun., 9.337341.

Nathan, C.F. (1987) Neutrophil activation on biological surfaces. Mas- sive secretion of hydrogen peroxide in response to products of mac- rophages and lymphocytes. J . Clin. Invest., 80:1550-1560.

Nathan, C.F. (1989) Respiratory hurst in adherent neutrophils: Trig- gering by colony-stimulating factors CSF-GM and CSF-G, Blood, 73t301306.

156 LIANC AND PETTY

Oliver, J.M. (1978) Cell biology of leukocyte abnormalities-membrane and cytoskeletal function in normal and defective cells. Am. J. Pathol., 93 ~221-260.

Patriarca, P., Cramer, R., Moncalvo, S., Rossi, F., and Romeo, D. (1971) Enzymatic basis of metabolic stimulation in leucocytes during phagocytosis: The role of activated NADPH oxidase. Arch. Bio- chem. Biophys., 145255-262.

Petty, H.R., Francis, J.W., and Anderson, C.L. (1989) Cell surface distribution of Fc receptors I1 and 111 on living human neutrophils before and during antibody dependent cellular cytotoxicity. J. Cell. Physiol., I 4 1 ~598-605.

Petty, H.R., Liang, B., and Maher, R.J. (1992) Mapping the entry of reactive oxygen metabolites into target erythrocytes during neutrophil-mediated antibody-dependent cellular cytotoxicity. J. Cell. Physiol., 150t447450.

Sawyer, D.W., Sullivan, J.A., and Mandell, G.L. (1985) Intracellular free calcium localized in neutrophils during phagocytosis. Science, 230r663-666.

Sbarra, A.J., and Karnovsky, M.L. (1959) The biochemical basis of phagocytosis I. Metabolic changes during the ingestion of particles by polymorphonuclear leukocytes. J. Biol. Chem., 234t1355-1362.

Tauber, A.I., and Babior, B.M. (1977) Evidence for hyroxyl radical production by human neutrophils. J. Clin. Invest., 60r374-379.

Thorell, B. (1980) Intracellular red-ox steady states as basis for cell characterization by flow cytofluorometry. Blood Cells, 6r745-751.

Umei, T., Babior, B.M., Curnutte, J.T., and Smith, R.M. (1991) Identi- fication of the NADPH-binding subunit of the respiratory burst oxidase. J. Biol. Chem., 266r6019-6022.

Weiss, S.J., King, G.W., and LoBuglio, A.F. (1977) Evidence for hydroxyl radical generation by human monocytes. J. Clin. Invest. 60t370-373.

Weitzman, S.A., and Stossel, T.P. (1981) Mutation caused by human phagocytes. Science, 212t546-547.

Wymann, M.P., von Tscharner, V., Derauleau, D.A., and Baggiolini, M. (1987) The onset of the respiratory burst in human neutrophils: Real-time studies of H,O,. formation reveal a rapid agonist-induced transduction process. J . €3101. Chem., 262r12048-12053.

Yuo, A., Kitagawa, S., Suzuki, I., Urabe, A., Okabe, T., Saito, M., and Takaku, F. (1989) Tumor necrosis factor as an activator of human granulocytes. J. Immunol., 142:1678-1684.

Zhou, M.J., Todd, R.F., and Petty, H.R. (1991) Detection oftransmem- brane linkages between immunoglobulin or complement receptors and the neutrophil's cortical microfilaments by resonance energy transfer microscopy. J. Mol. Biol., 218:263-268.

imaging neutrophil activation: analysis of the translocation and utilization of nad(p)h-associated...

Documents