Journal of Antimicrobial Chemotherapy (1991) 29, 313-321
Detection of antibiotic-induced platelet dysfunction in whole blood using flow cytometry Connie S. IngaDs and Earl H. Freimer*
Departments of Medicine and Microbiology, Medical College of Ohio. P.O. Box 10008, Toledo, Ohio 43699, USA
Introduction
Normal platelet function depends upon a series of reactions that are localized at the surface of the platelet membrane. These are initiated by the binding of an agonist with its specific glycoprotein receptor. Receptor-ligand interactions have been described for the agonists ADP, epinephrine and collagen (Philips, 1985). Interactions also have been characterized for such platelet-associated proteins as von Willebrand factor (vWf), fibrinogen,fibronectin,thrombin and thrombospondin (Philips, 1985). The identification of specific platelet membrane surface receptors and the concomitant development of monoclonal antibodies to these receptors have served to characterize the pathogenesis of several acquired or congenital platelet associated disorders. Furthermore, the use of these monoclonal antibodies in conjunction with flow cytometry provides a means for quantitative studies of both normal and abnormal platelet function. A number of recent reports suggest that treatment of infections with various betalactam antibiotics may result in bleeding due either to inhibiton of platelet function, or to impaired synthesis of vitamin K-dependent clotting factors, (Bang etal., 1982; 'Correspondence to: Dr Earl H. Freimer, Department of Microbiology, Medical College of Ohio, 3000 Arlington Avenue, Toledo, OH 43699-0008, USA. 313 0305-7453/92/030313+09 $02.00/0
© 1992 The British Society for AntimicrobUl Chemotherapy
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
Using flow cytometry and activation-dependent monoclonal antibodies, we have developed a technique based on forward angle-light scatter (FALS) and immunofluorescence that simultaneously detects human platelet activation, secretion, and aggregation in whole blood. To detect the effects of cefotetan and latamoxef, both of which contain an N-MTT side chain, and of free W-MTT and cefoxitin, which does not contain the A'-MTT side chain, on platelet activation and secretion, platelets were stained by the indirect method using a murine-produced platelet specific activation-dependent monoclonal antibody, SI2, and a goat anti-mouse IgG fluorescein-conjugated antibody. S12 binds to a 140kd alpha granule membrane protein (GMP-140) that is expressed during secretion. Single parameter, 256 channel, log integrated greenfluorescencehistograms were generated, and negative and positive fluorescent populations were defined. Latamoxef and cefotetan reduced the number of platelets expressing S12 by more than 43%. In contrast, cefoxitin reduced the number of platelets expressing S12 by only 13-5%. The inhibition of GMP-140 expression per platelet was calculated by converting the log data to linear fluorescence intensity. Latamoxef and cefotetan inhibited expression of GMP-140 by 88% and 87% respectively. Free tf-MTT inhibited its expression by 68%. In contrast cefoxitin reduced GMP-140 expression per platelet by only 45%.
314
C S. IngaDs and K H. Frdmer
Materials and methods Whole blood preparations Whole blood (nine parts venous blood to one part 3-8% w/v trisodium citrate) was drawn from volunteers who had been free of all medications, including aspirin, for at least ten days. Aliquots of 1 mL of anticoagulated blood were incubated for 20 min at
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
Lipsky, 1984; Andrassy, Bechtold & Ritz, 1985; Agnelli et al., 1986). The specific mechanism(s) by which these antimicrobial agents impair platelet function have not been established. However, they appear to interfere with platelet aggregation induced by agonists such as ADP, epinephrine and collagen. This occurs with platelets obtained either from normal volunteers or from patients who have received multiple doses of the antibiotic, (Hooper, Haney & Stone, 1980; Quintillani, 1983; Uotila & Suttie, 1983; Bach, 1984; Barza et al., 1986; Welage et al., 1987). Interaction of the antibiotics with receptor sites on the surface of the platelet membrane may be responsible for such an effect, although it unlikely that the inhibitory effect on different groups of agonists would be the result of interaction with a single specific receptor site. In a previous study, we have demonstrated the inhibitory effects of both latamoxef, a • /Mactam antibiotic containing an N-methylthiotetrazole-side chain, and free JV-methylthiotetrazole on platelet aggregation in vitro (Ingalls, Somani & Freimer, 1989). Qualitative and quantitative differences in the response patterns of platelet aggregation invoked by the specific agonists were recorded by aggregometry. These studies demonstrated that free N-methylthiotetrazole (N-MTT) as well as latamoxef inhibit platelet aggregation and thus impair platelet function. Although the molecular basis by which these compounds interact with platelets has not been established, the drugs, or intermediate drug metabolites, appear to interfere with the interaction of specific agonists and their respective platelet membrane surface receptor sites. Aggregation is the culmination of a series of activation reactions: expression of the fibrinogen receptor on the platelet glycoprotein Ilb/IIIa complex; binding of fibrinogen; and granular secretion. Platelet activation marks the beginning of this series of events and is thus an essential mediator of aggregation. Therefore, further investigation of the effects of various beta-lactam antibiotics on platelet activation in relation to aggregation seemed warranted. With the use of flow cytometry and an activation-dependent monoclonal antibody, we have developed a technique that simultaneously detects platelet activation and secretion, and aggregation in whole blood based on forward angle light scatter properties and immunofluorescence. To assess the effects of cefotetan and latamoxef both of which contain an W-MTT side chain, of free N-MTT and of the non N-MTT side chain antibiotic cefoxitin on platelet activation and secretion, agonist-induced platelet secretion was measured as the expression of a 140 kd a-granule membrane protein, GMP-140, (Berman et al., 1986; McEver & Martin, 1984; Steinberg et al., 1985). Platelets were stained by the indirect method using a murine produced platelet specific activation-dependent monoclonal antibody, SI2, and commercial goat antimouse IgG-fluorescein-conjugated antibody, GAM-FTTC. S12 binds to a 140 kd alpha-granule membrane proteb, GMP-140, that is expressed at the platelet membrane surface during secretion, (McEver & Martin., 1984).
Detection of platelet dysfunction byflowcytometry
315
Reagent preparations Antibodies in this study were S12 (obtained from Dr. Roger McEver, Oklahoma Medical Research Foundation, Oklahoma City, OK, USA) and GAM-FTTC, a fluorescein isothiocyanate (FTTQ conjugated goat antimouse IgG (Ortho Diagnostic Systems Inc., Raritan, NJ, USA). The agonists used in this study were purchased from Bio/Data Corporation, Hatboro, PA, USA. ADP was a lyophilized preparation of adenosine5'-diphosphate. Epinephrine was a lyophilized preparation of adrenaline. Three /Mactam antibiotics were used in this study. Cefoxitin was supplied as the sodium salt (Merck Sharp & Dohme, West Point, PA, USA). Cefotetan was provided also as the sodium salt (Stuart Pharmaceuticals, Wilmington, DE, USA). Latamoxef diammonium was supplied by Eli Lilly and Co. (Indianapolis, IN, USA). The antibiotics were prepared immediately before use. Concentrated (lOx) solutions were prepared in 1 mL of 3.8% trisodium citrate in 10 mL plastic syringes and a 9mL volume of whole blood was then drawn into each syringe.
Flow cytometric analysis Samples were analyzed using an EPICS 753 Flow Cytometer equipped with two argon lasers and one dye laser (Coulter Electronic, Hialeah, FL, USA). The cytometer was set up to measure log forward angle light scatter (LFALS), log 90 light scatter (L90LS) and log green fluorescein isothiocyanate fluoresence. Immunocheck beads of known diameter and fluorescence were used daily to ensure instrument alignment and stability (Coulter Immunology, Hialeah, FL, USA). Light scatter gates were initially determined using partially purified platelets. Single parameter 256 channel, log integrated green fluorescence histograms were generated, and the negative and positive fluorescent populations were defined. An 'IMMUNO' computer program (EASY EPICS, Coulter Electronics. Hialeah, FL, USA) was used to quantitate the percentage of positive platelets. Greater than 95% of the events within these gates were positive when stained with OKM-5 FTTC a monoclonal antibody that binds to platelets (Ortho Diagnostic Systems Inc., Raritan, NJ, USA). Comparable results were obtained when whole blood preparations were used. In specimens in which 'micro' aggregation occurred, the light scatter gates were extended but still kept below the level of the red blood cells.
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
37°C in the presence of cefoxitin, of cefotetan, of latamoxef or of free N-MTT. Phosphate buffered saline (PBS) served as a control. Antibiotics were used at a final concentration of 3mg/mL. The final concentration of free N-M'l'l was 28 mM (3-86 mg/mL). These aliquots of blood were then incubated with S12 antibody (5 /ig/ml) and a 1:1 mixture of ADP. epinephrine (2x 10-3M and 1 x 10"4!* respectively) for 20 min at 37°C. PBS served as a non-activated control. A 50 /zL aliquot of blood was removed from each sample and fixed in 1 mL of 1% paraformaldehyde in PBS for 1 h at 4°C. Fixed whole blood samples were washed by centrifugation at 1600; for 10 min and resuspended in 50/iL PBS. Samples were then incubated with 5/iL of FTTC-conjugated goat anti-mouse IgG (GAM-FTTC, Ortho Diagnostic Systems Inc., Rantan, NJ, USA) for 20 min at room temperature. The preparations were washed twice, resuspended in PBS and analyzed by flow cytometry.
C. S. IngaOs and E. H. Freimer
316
NON-flCTIURTCD CONTROL PLRTKLITS
IB HCTIURTID CONTROL PLflTKLITS
Frtq 12 1 HDP:IP I RCTIURTED HOX-TRERTED PLRTELETS
^ . J
15 Channel Nunbtr
26 1 X18
.
IJ
25
Figure 1. Fluorescence histogram showing the effect of latamoxef (MOX) on ADP: Epinephrine (EPI) induced S12-GAM-FTTC binding. Qtrated whole blood wai incubated for 20 min in the presence of either latamoxef (3 mg/L) or PBS. Aliquot* of each were then incubated for 20 min with S12-GAM-FTTC and a 1:1 mixture of ADP:EPI. PBS also served at a non-activated control. This histogram is from a single experiment, but is representative of seven similar experiments.
Results Figures 1 to 4 are fluorescence histograms that demonstrate the distribution of GMP-140 expression (SI2 binding) on non-activated control platelets, activated control platelets, and activated platelets treated either with an antibiotic or with N-MTT. The fluorescent signals differentiate three distinct populations of platelets. Non-activated platelets have a background autofluorescence intensity that peaks at channel 5-1. Activated control platelets that have not been exposed either to antibiotic or to N-MTT have a fluorescence intensity that peaks at channel 148-3, which is 29-6 times as intense as the non-activated background control. In contrast, the fluorescence intensity of platelets treated with latamoxef, and then activated with ADP:epinephrine, peaks at channel 80-0, which is 15-7 times as intense as the non-activated background control (Figure 1). The relative level of GMP-140 expression in cefotetan-treated activated platelets is essentially the same as that observed with latamoxef treatment (Figure 2). The fluorescence intensity of these platelets peaks at channel 83-2, which is 16-6 times the non-activated background control. The peak of platelets treated with cefoxitin is seen at 128-0, a higher channel than platelets treated with either latamoxef or cefotetan (Figure 3). This fluorescence intensity peak is 25-1 times the background control. Platelets treated with free W-MTT
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
xie
Detection of platelet dysfunction by flow cytometry
26
317
NOH-flCTIURTED "CONTROL PLHTKLETB
f»DP: IPI RCIIUflTED CONTROL PLATELETS
16
HDP:IP I . flCTIUHTID TCT-TRERTED PLflTELITS
15 Channel Nunbtr
2B 1 X1B
25
Flgnc 2. Fluorescence histogram showing the effect of cefotetan (JET) on ADP: Epinephrine (EPI) induced S12-GAM-FTTC binding. Citrated whole blood was incubated for 20 mm in the presence of cefotetan (3 mg/L) or PBS. Aliquots of each were then incubated for 20min with S12-GAM-FITC and a 1:1 mixture of ADP-.EPI. PBS also served as a non-activated control. This histogram is from a tingle experiment, but it representative of (even similar experiments.
and then activated with ADP:epinephrine give a fluorescence intensity that peaks at channel 111-4 or 22-8 times the background control (Figure 4). These results are recorded in the Table. The percentage of platelets expressing GMP-140 (S12 positive) was calculated using Table. Inhibition of platelet function by /J-lactam antibiotics as determined by flow cytometry
intensity*
Percent of S12 inhibition'
29-6 15-7 16-6 25-1 22-8
—
—
45-9 43-8 13-5 25-0
88-0
Multiples Platelet treatment Control (nonactivated) Control (activated) Latamoxef Cefotetan Cefoxitin FreeN-MTT
Peak channel'
of
Percent of GMP-140 inhibition'
51 148-3 80-0 83-2 128-0 111-4
870 450 68-0
"These numbers rcpmeut the mean peak rhannrif offluorescenceintensity. These numbers repieseut multiples offluorescenceintensity when compared to the nonactivated control. These numbers represent the percentage of platelets inhibited from expressing S12. These numbers represent the percentage of inhibition of GMP-140 expression per platelet
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
Fr«q 1 2 1 X1O
318
C S. IngaQs and E. H. Frdmer
NON-nCTIURTID CONTROL PLRTELETS
nDP:EPI HCIIUHIID CONTROL PLHTILETS
Channel
Nunber
X10
Figure 3. Fhioretcence histogram showing the effect of cefoxitin (FOX) on ADP: epinephrine (EPI) induced S12-GAM-FTTC binding. Citrated whole blood was incubated for 20min in the presence of either cefoxitin (3mg/L) or PBS. Aliquots of each were then incubated for 20min with S12-GAM-FTTC and a 1:1 mixture of ADP:EPI. PBS also served as a non-activated control. This histogram is from a single experiment, but is representative of seven similar experiments.
the 'IMMUNO' program. This program uses a channel by channel subtraction algorithm to estimate the percentage of positive fluorescent events in histograms in which positive events are not clearly resolved. From these values the percentage of platelets inhibited from expressing S12 is calculated as follows: control platelets (activated)-treated platelets xlOO control platelets (activated) Using this formula, the following percentages were calculated. Latamoxef and cefotetan inhibited the number of platelets expressing SI2 by 45-9% and 43-8%, respectively. Free 7V-MTT reduced the number of platelets expressing S12 by 25%. On the other hand, cefoxitin reduced the same parameter by only 13.5%. These results also
are recorded in the Table. To estimate the relative quantities of S12 per platelet, the per cent inhibition of GMP-140 expression per platelet was calculated. The log data was converted to linear fluorescence intensity. The inhibiton of GMP-140 expression per platelet then was calculated using the following formula (Schmid, Schmid & Giorgi, 1988): b + c -b~ c
xlOO
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
HDP:IP I nCTIURTED TOX-TRKflTED PLRTCLCTS
Detection of pbtekt dysfunction by flow cytometry
2B
319
HOM-HCTIUHTrD "CONTnOL PLRTILETS
IB
RDP:IPI flCTIUHTtO CONTROL PLRTELCTS
rraq 12 1
RDP:IP I nCTIUHTED N-HTT-TREflTID PLflTILETg,
15
20 1
2E
C h a n n i l Munber X1B
Flgne 4. Fluorescence histogram showing the effect of free AT-methylthiotetrazole (Af-MTT) on ADP: epinephiine (EPI) induced S12-GAM-FTTC binding. Gtrated whole blood was incubated for 20min in the presence of either N-MTT (3-86 mg/L) or PBS. AliquoU of each were then incubated for 20 min with S12-GAM-FTTC and a 1:1 mixture of AOP-.EPI. PBS also served as a non-activated controL This histogram is from a single experiment, but is representative of seven similar experiments.
b = corrected base (l-02697;5); +c, mean channel of positive (activated) control platelets; exp, mean channel of positive drug-treated platelets; and — c, mean channel of negative (nonactivated) control platelets. Mean channels were derived from the Immuno Analysis Results program (EASY 2, Coulter Electronics, Hialeah, FL, USA). Latamoxef and cefotetan inhibited the amount of GMP-140 expression per platelet by 88% and 87%, respectively. Free N-MTT inhibited GMP-140 expression by 68%, whereas cefoxitin reduced the amount of GMP-140 expression per platelet by only 45%.
Discussion
ADP stimulation of whole blood preparations results in platelet activation, thus initiating the secretory process. During this process, substances found within the a-granules and dense bodies are secreted from the cells. It is well established that platelet activation induced by agonists such as ADP is consistently enhanced by epinephiine, and, that this heightened effect is relatively independent of donor or anticoagulant, (Grant & Scrutton, 1980; O'Brien, 1964; Plow & Marguerie, 1980; Shattil, Budzinski & Scrutton, 1989). The 140 kd a-granule glycoprotein, GMP-140, is
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
xie
320
C S. IngaDs and E. H. Fretmer
Acknowledgement This study was supported by a grant from Merck, Sharp & Dohme, West Point, Pennsylvania, USA. References Agnelli, G., Del Favero, A., Parisc, P., Gucrciolini, R., Pasticol, B., Noncil, G. G. et al. (1986). Cephalosporin-induced hypoprothrombinemia: is the N-methylthiotctrazok side chain the culprit? Antimicrobial Agents and Chemotherapy 29, 1108-9.
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
a marker of platelet secretion. The murine monoclonal antibody, SI2, specifically recognizes the GMP-140 glycoprotein molecule. Thus, the competency of activation and secretion can be assessed by measuring the binding of S12 to the GMP-140 molecule that associates with the platelet plasma membrane during alpha-granule secretion, (McEver & Martin, 1984). Using these S12-GAM-FTTC labeled platelets as a marker of activation and secretion, fluorescence histograms were produced by flow cytometric analysis. This enabled us to quantitatively assess the effects of the antibiotics latamoxef, cefotetan, and cefoxitin, and the effects of free N-MTT on platelet activation and secretion in whole blood. Results of these in-vitro studies clearly demonstrate that the N-MTT containing cephalosporins latamoxef and cefotetan, and free N-MTT significantly inhibit the stimulation of platelets by the agonists ADP and epinephrine. These N-MTT side chain containing antibiotics and the free-N-MTT molecule inhibited normal platelet function by reducing both the total number of platelets activated and the number of GMP-140 molecules measured on the surface of stimulated platelets. Stimulated platelets exposed to either latamoxef or cefotetan expressed less than 14% of the number of GMP-140 molecules measured on the surface of activated control platelets. Stimulated platelets exposed to free 7V-MTT expressed only 32% of the number of GMP-140 molecules present on activated control platelets. In contrast, stimulated platelets exposed to cefoxitin, which does not contain an N-MTT side chain, expressed 55% of the number of molecules of GMP-140 that were measured on the activated control platelets. The fact that a much greater reduction in the number of GMP-140 molecules expressed per platelet occurred with latamoxef, cefotetan or free N-MTT than with cefoxitin suggests that the W-MTT side chain may play a critical role in antagonizing the aggregation of platelets by .W-MTT side chain containing cephalosporins in vivo. Nevertheless, because of the severity of the inhibition of platelet function caused by these cephalosporins in vivo factors other than the .N-MTT side chain itself may well be involved. Whether inhibiton of platelet activation occurs via the direct interference with initial ligand binding to surface receptors or within the secretory process itself has not been determined. We, however, propose to answer this question through the use of other platelet specific monoclonal antibodies, such as PAC-1, (Shattil, Cunningham & Hoxie, 1987). PAC-1, an IgM murine monoclonal antibody specific for the activated complex form of the platelet surface glycoprotein Ilb/IIIa, will be used to detect the initiation of agonist-induced activation. By this approach, antibiotic induced inhibiton of platelet activation as a result of interference with ligand binding and expression of the activated ETb/nia complex can be studied.
Detection of platelet dysfunction by flow cytometry
321
(Received 10 July 1991; accepted 4 October 1991)
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
Andrassy, K., Bechtold, H. & Ritz, E. (1985). Hypoprothrombinemia caused by cephalosporins. Journal of Antimicrobial Chemotherapy 15, 133-6. Bach, M. C. (1984). Prolonged bleeding time associated with 'low-dose' moxalactam therapy. Journal of the American Medical Association 251, 3082 Bang, N. U., Tessler, S. S., Heldcnreich, R. O., Marks, C. A. & Mattler, L. E. (1982). Effects of moxalactam on blood coagulation and platelet function. Reviews of Infectious Diseases 4, Suppl. S546-54. Barza, M., Furie, B., Brown, A. E. & Furie, B. C. (1986). Defects in vitamin K-dependent carboxylation associated with moxalactam treatment. Journal of Infectious Diseases 153, 1166-9. Bennan, C. L., Yeo, E. L., Wencel-Drakc, J. D., Furie, B. C , Ginsberg, M. H. & Furie, B. (1986). A platelet alpha granule membrane protein that is associated with the plasma membrane after activation. Journal of Clinical Investigation 78, 130-7. Grant, J. A. &. Scrutton, M. C. (1980). Positive interaction between agonists in the aggregation response of human platelets: interaction between ADP, adrenaline and vasopressin. British Journal of Haematology 44, 109-25. Hooper, C. A., Haney, B. B. & Stone, H. H. (1980). Gastointestinal bleeding due to vitamin K deficiency in patients on parenteral cefamandole. Lancet i, 39—40. Ingalls, C. S., Somani, P. & Freimer, E. H. (1989). Inhibition of platelet aggregation by moxalactam and free N-mcthylthiotetrazole. Clinical Therapeutics 11, 640-51. Iipsky, J. J. (1984). Mechanism of the inhibition of the gamma-carboxylation of glutamk acid by N-methylthiotetrazole-containing antibiotics. Proceedings of the National Academy of Sciences of the USA 81, 2893-7. McEver, R. P. & Martin, M. N. (1984). A monoclonal antibody to a membrane glycoprotein binds only to activated platelets. Journal of Biological Chemistry 259, 9799-804. O'Brien, J. R. (1964). A comparison of platelet aggregation produced by seven compounds and a comparison of their inhibitors. Journal of Clinical Pathology 17, 275-81. Phillips, D. R. (1985). Receptors for platelet agonists. In Platelet Membrane Glycoproteins (George, J. N., Nurden, A. T. & Phillips, D. R., Eds), pp. Plenum, New York Plow, E. F. & Marguerie, G. A. (1980). Induction of the fibrinogen receptor on human platelets by epinephrine and the combination of epinephrine and ADP. Journal of Biological Chemistry 255, 10971-7. Quintillani, R. (1983). Bleeding disorders associated with newer cephalosporins. Clinical Pharmacy 2, 360-1 Schmid, I., Schmid, P. & Giorgj, J. V. (1988). Conversion of logarithmic channel numbers into relative linear fluorescence intensity. Cytometry 9, 533-8. Shattil, S. J., Budzynski, A. & Scrutton, M. C. (1989). Epinephrine induces platelet fibrinogen receptor expression, fibrinogen binding, and aggregation in whole blood in the absence of other excitatory agonists. Blood 73, 150-8. Shattil, S. J., Cunningham, M. & Hoxie, J. A. (1987). Detection of activated platelets in whole blood using activation-dependent monoclonal antibodies and flow cytometry. Blood 70, 307-15 Steinberg, P. E., McEver, R. P., Shuman, M. A., Jacques, Y. V. & Bainton, D. F. (1958). A platelet alpha-granule membrane protein (GMP-140) is expressed on the plasma membrane after activation. Journal of Cell Biology 101, 880-6. Uotila, L. & Suttie, J. (1983). Inhibition of vitamin K-dependent carboxylase in vitro by cefamandole and its structural analogs. Journal of Infectious Diseases 148, 571-8. Welage, L. S., Wilton, J. H., Adelman, M. H., Grasela, T. H. & Schentag, J. J. (1987). The production and excretion of N-methylthiotetrazole (N-MTT) by normal volunteers given cefoperazone, moxalactam, and cefotetan. In Efficacy and Cost Implications of the New Cephalosporins. Stuart Pharmaceuticals Division of ICI, Wilmington, DE.
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
Detection of antibiotic-induced platelet dysfunction in whole blood using flow cytometry Connie S. IngaDs and Earl H. Freimer*
Departments of Medicine and Microbiology, Medical College of Ohio. P.O. Box 10008, Toledo, Ohio 43699, USA
Introduction
Normal platelet function depends upon a series of reactions that are localized at the surface of the platelet membrane. These are initiated by the binding of an agonist with its specific glycoprotein receptor. Receptor-ligand interactions have been described for the agonists ADP, epinephrine and collagen (Philips, 1985). Interactions also have been characterized for such platelet-associated proteins as von Willebrand factor (vWf), fibrinogen,fibronectin,thrombin and thrombospondin (Philips, 1985). The identification of specific platelet membrane surface receptors and the concomitant development of monoclonal antibodies to these receptors have served to characterize the pathogenesis of several acquired or congenital platelet associated disorders. Furthermore, the use of these monoclonal antibodies in conjunction with flow cytometry provides a means for quantitative studies of both normal and abnormal platelet function. A number of recent reports suggest that treatment of infections with various betalactam antibiotics may result in bleeding due either to inhibiton of platelet function, or to impaired synthesis of vitamin K-dependent clotting factors, (Bang etal., 1982; 'Correspondence to: Dr Earl H. Freimer, Department of Microbiology, Medical College of Ohio, 3000 Arlington Avenue, Toledo, OH 43699-0008, USA. 313 0305-7453/92/030313+09 $02.00/0
© 1992 The British Society for AntimicrobUl Chemotherapy
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
Using flow cytometry and activation-dependent monoclonal antibodies, we have developed a technique based on forward angle-light scatter (FALS) and immunofluorescence that simultaneously detects human platelet activation, secretion, and aggregation in whole blood. To detect the effects of cefotetan and latamoxef, both of which contain an N-MTT side chain, and of free W-MTT and cefoxitin, which does not contain the A'-MTT side chain, on platelet activation and secretion, platelets were stained by the indirect method using a murine-produced platelet specific activation-dependent monoclonal antibody, SI2, and a goat anti-mouse IgG fluorescein-conjugated antibody. S12 binds to a 140kd alpha granule membrane protein (GMP-140) that is expressed during secretion. Single parameter, 256 channel, log integrated greenfluorescencehistograms were generated, and negative and positive fluorescent populations were defined. Latamoxef and cefotetan reduced the number of platelets expressing S12 by more than 43%. In contrast, cefoxitin reduced the number of platelets expressing S12 by only 13-5%. The inhibition of GMP-140 expression per platelet was calculated by converting the log data to linear fluorescence intensity. Latamoxef and cefotetan inhibited expression of GMP-140 by 88% and 87% respectively. Free tf-MTT inhibited its expression by 68%. In contrast cefoxitin reduced GMP-140 expression per platelet by only 45%.
314
C S. IngaDs and K H. Frdmer
Materials and methods Whole blood preparations Whole blood (nine parts venous blood to one part 3-8% w/v trisodium citrate) was drawn from volunteers who had been free of all medications, including aspirin, for at least ten days. Aliquots of 1 mL of anticoagulated blood were incubated for 20 min at
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
Lipsky, 1984; Andrassy, Bechtold & Ritz, 1985; Agnelli et al., 1986). The specific mechanism(s) by which these antimicrobial agents impair platelet function have not been established. However, they appear to interfere with platelet aggregation induced by agonists such as ADP, epinephrine and collagen. This occurs with platelets obtained either from normal volunteers or from patients who have received multiple doses of the antibiotic, (Hooper, Haney & Stone, 1980; Quintillani, 1983; Uotila & Suttie, 1983; Bach, 1984; Barza et al., 1986; Welage et al., 1987). Interaction of the antibiotics with receptor sites on the surface of the platelet membrane may be responsible for such an effect, although it unlikely that the inhibitory effect on different groups of agonists would be the result of interaction with a single specific receptor site. In a previous study, we have demonstrated the inhibitory effects of both latamoxef, a • /Mactam antibiotic containing an N-methylthiotetrazole-side chain, and free JV-methylthiotetrazole on platelet aggregation in vitro (Ingalls, Somani & Freimer, 1989). Qualitative and quantitative differences in the response patterns of platelet aggregation invoked by the specific agonists were recorded by aggregometry. These studies demonstrated that free N-methylthiotetrazole (N-MTT) as well as latamoxef inhibit platelet aggregation and thus impair platelet function. Although the molecular basis by which these compounds interact with platelets has not been established, the drugs, or intermediate drug metabolites, appear to interfere with the interaction of specific agonists and their respective platelet membrane surface receptor sites. Aggregation is the culmination of a series of activation reactions: expression of the fibrinogen receptor on the platelet glycoprotein Ilb/IIIa complex; binding of fibrinogen; and granular secretion. Platelet activation marks the beginning of this series of events and is thus an essential mediator of aggregation. Therefore, further investigation of the effects of various beta-lactam antibiotics on platelet activation in relation to aggregation seemed warranted. With the use of flow cytometry and an activation-dependent monoclonal antibody, we have developed a technique that simultaneously detects platelet activation and secretion, and aggregation in whole blood based on forward angle light scatter properties and immunofluorescence. To assess the effects of cefotetan and latamoxef both of which contain an W-MTT side chain, of free N-MTT and of the non N-MTT side chain antibiotic cefoxitin on platelet activation and secretion, agonist-induced platelet secretion was measured as the expression of a 140 kd a-granule membrane protein, GMP-140, (Berman et al., 1986; McEver & Martin, 1984; Steinberg et al., 1985). Platelets were stained by the indirect method using a murine produced platelet specific activation-dependent monoclonal antibody, SI2, and commercial goat antimouse IgG-fluorescein-conjugated antibody, GAM-FTTC. S12 binds to a 140 kd alpha-granule membrane proteb, GMP-140, that is expressed at the platelet membrane surface during secretion, (McEver & Martin., 1984).
Detection of platelet dysfunction byflowcytometry
315
Reagent preparations Antibodies in this study were S12 (obtained from Dr. Roger McEver, Oklahoma Medical Research Foundation, Oklahoma City, OK, USA) and GAM-FTTC, a fluorescein isothiocyanate (FTTQ conjugated goat antimouse IgG (Ortho Diagnostic Systems Inc., Raritan, NJ, USA). The agonists used in this study were purchased from Bio/Data Corporation, Hatboro, PA, USA. ADP was a lyophilized preparation of adenosine5'-diphosphate. Epinephrine was a lyophilized preparation of adrenaline. Three /Mactam antibiotics were used in this study. Cefoxitin was supplied as the sodium salt (Merck Sharp & Dohme, West Point, PA, USA). Cefotetan was provided also as the sodium salt (Stuart Pharmaceuticals, Wilmington, DE, USA). Latamoxef diammonium was supplied by Eli Lilly and Co. (Indianapolis, IN, USA). The antibiotics were prepared immediately before use. Concentrated (lOx) solutions were prepared in 1 mL of 3.8% trisodium citrate in 10 mL plastic syringes and a 9mL volume of whole blood was then drawn into each syringe.
Flow cytometric analysis Samples were analyzed using an EPICS 753 Flow Cytometer equipped with two argon lasers and one dye laser (Coulter Electronic, Hialeah, FL, USA). The cytometer was set up to measure log forward angle light scatter (LFALS), log 90 light scatter (L90LS) and log green fluorescein isothiocyanate fluoresence. Immunocheck beads of known diameter and fluorescence were used daily to ensure instrument alignment and stability (Coulter Immunology, Hialeah, FL, USA). Light scatter gates were initially determined using partially purified platelets. Single parameter 256 channel, log integrated green fluorescence histograms were generated, and the negative and positive fluorescent populations were defined. An 'IMMUNO' computer program (EASY EPICS, Coulter Electronics. Hialeah, FL, USA) was used to quantitate the percentage of positive platelets. Greater than 95% of the events within these gates were positive when stained with OKM-5 FTTC a monoclonal antibody that binds to platelets (Ortho Diagnostic Systems Inc., Raritan, NJ, USA). Comparable results were obtained when whole blood preparations were used. In specimens in which 'micro' aggregation occurred, the light scatter gates were extended but still kept below the level of the red blood cells.
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
37°C in the presence of cefoxitin, of cefotetan, of latamoxef or of free N-MTT. Phosphate buffered saline (PBS) served as a control. Antibiotics were used at a final concentration of 3mg/mL. The final concentration of free N-M'l'l was 28 mM (3-86 mg/mL). These aliquots of blood were then incubated with S12 antibody (5 /ig/ml) and a 1:1 mixture of ADP. epinephrine (2x 10-3M and 1 x 10"4!* respectively) for 20 min at 37°C. PBS served as a non-activated control. A 50 /zL aliquot of blood was removed from each sample and fixed in 1 mL of 1% paraformaldehyde in PBS for 1 h at 4°C. Fixed whole blood samples were washed by centrifugation at 1600; for 10 min and resuspended in 50/iL PBS. Samples were then incubated with 5/iL of FTTC-conjugated goat anti-mouse IgG (GAM-FTTC, Ortho Diagnostic Systems Inc., Rantan, NJ, USA) for 20 min at room temperature. The preparations were washed twice, resuspended in PBS and analyzed by flow cytometry.
C. S. IngaOs and E. H. Freimer
316
NON-flCTIURTCD CONTROL PLRTKLITS
IB HCTIURTID CONTROL PLflTKLITS
Frtq 12 1 HDP:IP I RCTIURTED HOX-TRERTED PLRTELETS
^ . J
15 Channel Nunbtr
26 1 X18
.
IJ
25
Figure 1. Fluorescence histogram showing the effect of latamoxef (MOX) on ADP: Epinephrine (EPI) induced S12-GAM-FTTC binding. Qtrated whole blood wai incubated for 20 min in the presence of either latamoxef (3 mg/L) or PBS. Aliquot* of each were then incubated for 20 min with S12-GAM-FTTC and a 1:1 mixture of ADP:EPI. PBS also served at a non-activated control. This histogram is from a single experiment, but is representative of seven similar experiments.
Results Figures 1 to 4 are fluorescence histograms that demonstrate the distribution of GMP-140 expression (SI2 binding) on non-activated control platelets, activated control platelets, and activated platelets treated either with an antibiotic or with N-MTT. The fluorescent signals differentiate three distinct populations of platelets. Non-activated platelets have a background autofluorescence intensity that peaks at channel 5-1. Activated control platelets that have not been exposed either to antibiotic or to N-MTT have a fluorescence intensity that peaks at channel 148-3, which is 29-6 times as intense as the non-activated background control. In contrast, the fluorescence intensity of platelets treated with latamoxef, and then activated with ADP:epinephrine, peaks at channel 80-0, which is 15-7 times as intense as the non-activated background control (Figure 1). The relative level of GMP-140 expression in cefotetan-treated activated platelets is essentially the same as that observed with latamoxef treatment (Figure 2). The fluorescence intensity of these platelets peaks at channel 83-2, which is 16-6 times the non-activated background control. The peak of platelets treated with cefoxitin is seen at 128-0, a higher channel than platelets treated with either latamoxef or cefotetan (Figure 3). This fluorescence intensity peak is 25-1 times the background control. Platelets treated with free W-MTT
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
xie
Detection of platelet dysfunction by flow cytometry
26
317
NOH-flCTIURTED "CONTROL PLHTKLETB
f»DP: IPI RCIIUflTED CONTROL PLATELETS
16
HDP:IP I . flCTIUHTID TCT-TRERTED PLflTELITS
15 Channel Nunbtr
2B 1 X1B
25
Flgnc 2. Fluorescence histogram showing the effect of cefotetan (JET) on ADP: Epinephrine (EPI) induced S12-GAM-FTTC binding. Citrated whole blood was incubated for 20 mm in the presence of cefotetan (3 mg/L) or PBS. Aliquots of each were then incubated for 20min with S12-GAM-FITC and a 1:1 mixture of ADP-.EPI. PBS also served as a non-activated control. This histogram is from a tingle experiment, but it representative of (even similar experiments.
and then activated with ADP:epinephrine give a fluorescence intensity that peaks at channel 111-4 or 22-8 times the background control (Figure 4). These results are recorded in the Table. The percentage of platelets expressing GMP-140 (S12 positive) was calculated using Table. Inhibition of platelet function by /J-lactam antibiotics as determined by flow cytometry
intensity*
Percent of S12 inhibition'
29-6 15-7 16-6 25-1 22-8
—
—
45-9 43-8 13-5 25-0
88-0
Multiples Platelet treatment Control (nonactivated) Control (activated) Latamoxef Cefotetan Cefoxitin FreeN-MTT
Peak channel'
of
Percent of GMP-140 inhibition'
51 148-3 80-0 83-2 128-0 111-4
870 450 68-0
"These numbers rcpmeut the mean peak rhannrif offluorescenceintensity. These numbers repieseut multiples offluorescenceintensity when compared to the nonactivated control. These numbers represent the percentage of platelets inhibited from expressing S12. These numbers represent the percentage of inhibition of GMP-140 expression per platelet
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
Fr«q 1 2 1 X1O
318
C S. IngaQs and E. H. Frdmer
NON-nCTIURTID CONTROL PLRTELETS
nDP:EPI HCIIUHIID CONTROL PLHTILETS
Channel
Nunber
X10
Figure 3. Fhioretcence histogram showing the effect of cefoxitin (FOX) on ADP: epinephrine (EPI) induced S12-GAM-FTTC binding. Citrated whole blood was incubated for 20min in the presence of either cefoxitin (3mg/L) or PBS. Aliquots of each were then incubated for 20min with S12-GAM-FTTC and a 1:1 mixture of ADP:EPI. PBS also served as a non-activated control. This histogram is from a single experiment, but is representative of seven similar experiments.
the 'IMMUNO' program. This program uses a channel by channel subtraction algorithm to estimate the percentage of positive fluorescent events in histograms in which positive events are not clearly resolved. From these values the percentage of platelets inhibited from expressing S12 is calculated as follows: control platelets (activated)-treated platelets xlOO control platelets (activated) Using this formula, the following percentages were calculated. Latamoxef and cefotetan inhibited the number of platelets expressing SI2 by 45-9% and 43-8%, respectively. Free 7V-MTT reduced the number of platelets expressing S12 by 25%. On the other hand, cefoxitin reduced the same parameter by only 13.5%. These results also
are recorded in the Table. To estimate the relative quantities of S12 per platelet, the per cent inhibition of GMP-140 expression per platelet was calculated. The log data was converted to linear fluorescence intensity. The inhibiton of GMP-140 expression per platelet then was calculated using the following formula (Schmid, Schmid & Giorgi, 1988): b + c -b~ c
xlOO
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
HDP:IP I nCTIURTED TOX-TRKflTED PLRTCLCTS
Detection of pbtekt dysfunction by flow cytometry
2B
319
HOM-HCTIUHTrD "CONTnOL PLRTILETS
IB
RDP:IPI flCTIUHTtO CONTROL PLRTELCTS
rraq 12 1
RDP:IP I nCTIUHTED N-HTT-TREflTID PLflTILETg,
15
20 1
2E
C h a n n i l Munber X1B
Flgne 4. Fluorescence histogram showing the effect of free AT-methylthiotetrazole (Af-MTT) on ADP: epinephiine (EPI) induced S12-GAM-FTTC binding. Gtrated whole blood was incubated for 20min in the presence of either N-MTT (3-86 mg/L) or PBS. AliquoU of each were then incubated for 20 min with S12-GAM-FTTC and a 1:1 mixture of AOP-.EPI. PBS also served as a non-activated controL This histogram is from a single experiment, but is representative of seven similar experiments.
b = corrected base (l-02697;5); +c, mean channel of positive (activated) control platelets; exp, mean channel of positive drug-treated platelets; and — c, mean channel of negative (nonactivated) control platelets. Mean channels were derived from the Immuno Analysis Results program (EASY 2, Coulter Electronics, Hialeah, FL, USA). Latamoxef and cefotetan inhibited the amount of GMP-140 expression per platelet by 88% and 87%, respectively. Free N-MTT inhibited GMP-140 expression by 68%, whereas cefoxitin reduced the amount of GMP-140 expression per platelet by only 45%.
Discussion
ADP stimulation of whole blood preparations results in platelet activation, thus initiating the secretory process. During this process, substances found within the a-granules and dense bodies are secreted from the cells. It is well established that platelet activation induced by agonists such as ADP is consistently enhanced by epinephiine, and, that this heightened effect is relatively independent of donor or anticoagulant, (Grant & Scrutton, 1980; O'Brien, 1964; Plow & Marguerie, 1980; Shattil, Budzinski & Scrutton, 1989). The 140 kd a-granule glycoprotein, GMP-140, is
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
xie
320
C S. IngaDs and E. H. Fretmer
Acknowledgement This study was supported by a grant from Merck, Sharp & Dohme, West Point, Pennsylvania, USA. References Agnelli, G., Del Favero, A., Parisc, P., Gucrciolini, R., Pasticol, B., Noncil, G. G. et al. (1986). Cephalosporin-induced hypoprothrombinemia: is the N-methylthiotctrazok side chain the culprit? Antimicrobial Agents and Chemotherapy 29, 1108-9.
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
a marker of platelet secretion. The murine monoclonal antibody, SI2, specifically recognizes the GMP-140 glycoprotein molecule. Thus, the competency of activation and secretion can be assessed by measuring the binding of S12 to the GMP-140 molecule that associates with the platelet plasma membrane during alpha-granule secretion, (McEver & Martin, 1984). Using these S12-GAM-FTTC labeled platelets as a marker of activation and secretion, fluorescence histograms were produced by flow cytometric analysis. This enabled us to quantitatively assess the effects of the antibiotics latamoxef, cefotetan, and cefoxitin, and the effects of free N-MTT on platelet activation and secretion in whole blood. Results of these in-vitro studies clearly demonstrate that the N-MTT containing cephalosporins latamoxef and cefotetan, and free N-MTT significantly inhibit the stimulation of platelets by the agonists ADP and epinephrine. These N-MTT side chain containing antibiotics and the free-N-MTT molecule inhibited normal platelet function by reducing both the total number of platelets activated and the number of GMP-140 molecules measured on the surface of stimulated platelets. Stimulated platelets exposed to either latamoxef or cefotetan expressed less than 14% of the number of GMP-140 molecules measured on the surface of activated control platelets. Stimulated platelets exposed to free 7V-MTT expressed only 32% of the number of GMP-140 molecules present on activated control platelets. In contrast, stimulated platelets exposed to cefoxitin, which does not contain an N-MTT side chain, expressed 55% of the number of molecules of GMP-140 that were measured on the activated control platelets. The fact that a much greater reduction in the number of GMP-140 molecules expressed per platelet occurred with latamoxef, cefotetan or free N-MTT than with cefoxitin suggests that the W-MTT side chain may play a critical role in antagonizing the aggregation of platelets by .W-MTT side chain containing cephalosporins in vivo. Nevertheless, because of the severity of the inhibition of platelet function caused by these cephalosporins in vivo factors other than the .N-MTT side chain itself may well be involved. Whether inhibiton of platelet activation occurs via the direct interference with initial ligand binding to surface receptors or within the secretory process itself has not been determined. We, however, propose to answer this question through the use of other platelet specific monoclonal antibodies, such as PAC-1, (Shattil, Cunningham & Hoxie, 1987). PAC-1, an IgM murine monoclonal antibody specific for the activated complex form of the platelet surface glycoprotein Ilb/IIIa, will be used to detect the initiation of agonist-induced activation. By this approach, antibiotic induced inhibiton of platelet activation as a result of interference with ligand binding and expression of the activated ETb/nia complex can be studied.
Detection of platelet dysfunction by flow cytometry
321
(Received 10 July 1991; accepted 4 October 1991)
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
Andrassy, K., Bechtold, H. & Ritz, E. (1985). Hypoprothrombinemia caused by cephalosporins. Journal of Antimicrobial Chemotherapy 15, 133-6. Bach, M. C. (1984). Prolonged bleeding time associated with 'low-dose' moxalactam therapy. Journal of the American Medical Association 251, 3082 Bang, N. U., Tessler, S. S., Heldcnreich, R. O., Marks, C. A. & Mattler, L. E. (1982). Effects of moxalactam on blood coagulation and platelet function. Reviews of Infectious Diseases 4, Suppl. S546-54. Barza, M., Furie, B., Brown, A. E. & Furie, B. C. (1986). Defects in vitamin K-dependent carboxylation associated with moxalactam treatment. Journal of Infectious Diseases 153, 1166-9. Bennan, C. L., Yeo, E. L., Wencel-Drakc, J. D., Furie, B. C , Ginsberg, M. H. & Furie, B. (1986). A platelet alpha granule membrane protein that is associated with the plasma membrane after activation. Journal of Clinical Investigation 78, 130-7. Grant, J. A. &. Scrutton, M. C. (1980). Positive interaction between agonists in the aggregation response of human platelets: interaction between ADP, adrenaline and vasopressin. British Journal of Haematology 44, 109-25. Hooper, C. A., Haney, B. B. & Stone, H. H. (1980). Gastointestinal bleeding due to vitamin K deficiency in patients on parenteral cefamandole. Lancet i, 39—40. Ingalls, C. S., Somani, P. & Freimer, E. H. (1989). Inhibition of platelet aggregation by moxalactam and free N-mcthylthiotetrazole. Clinical Therapeutics 11, 640-51. Iipsky, J. J. (1984). Mechanism of the inhibition of the gamma-carboxylation of glutamk acid by N-methylthiotetrazole-containing antibiotics. Proceedings of the National Academy of Sciences of the USA 81, 2893-7. McEver, R. P. & Martin, M. N. (1984). A monoclonal antibody to a membrane glycoprotein binds only to activated platelets. Journal of Biological Chemistry 259, 9799-804. O'Brien, J. R. (1964). A comparison of platelet aggregation produced by seven compounds and a comparison of their inhibitors. Journal of Clinical Pathology 17, 275-81. Phillips, D. R. (1985). Receptors for platelet agonists. In Platelet Membrane Glycoproteins (George, J. N., Nurden, A. T. & Phillips, D. R., Eds), pp. Plenum, New York Plow, E. F. & Marguerie, G. A. (1980). Induction of the fibrinogen receptor on human platelets by epinephrine and the combination of epinephrine and ADP. Journal of Biological Chemistry 255, 10971-7. Quintillani, R. (1983). Bleeding disorders associated with newer cephalosporins. Clinical Pharmacy 2, 360-1 Schmid, I., Schmid, P. & Giorgj, J. V. (1988). Conversion of logarithmic channel numbers into relative linear fluorescence intensity. Cytometry 9, 533-8. Shattil, S. J., Budzynski, A. & Scrutton, M. C. (1989). Epinephrine induces platelet fibrinogen receptor expression, fibrinogen binding, and aggregation in whole blood in the absence of other excitatory agonists. Blood 73, 150-8. Shattil, S. J., Cunningham, M. & Hoxie, J. A. (1987). Detection of activated platelets in whole blood using activation-dependent monoclonal antibodies and flow cytometry. Blood 70, 307-15 Steinberg, P. E., McEver, R. P., Shuman, M. A., Jacques, Y. V. & Bainton, D. F. (1958). A platelet alpha-granule membrane protein (GMP-140) is expressed on the plasma membrane after activation. Journal of Cell Biology 101, 880-6. Uotila, L. & Suttie, J. (1983). Inhibition of vitamin K-dependent carboxylase in vitro by cefamandole and its structural analogs. Journal of Infectious Diseases 148, 571-8. Welage, L. S., Wilton, J. H., Adelman, M. H., Grasela, T. H. & Schentag, J. J. (1987). The production and excretion of N-methylthiotetrazole (N-MTT) by normal volunteers given cefoperazone, moxalactam, and cefotetan. In Efficacy and Cost Implications of the New Cephalosporins. Stuart Pharmaceuticals Division of ICI, Wilmington, DE.
Downloaded from http://jac.oxfordjournals.org/ at Frankfurt Univesity Library on May 28, 2015
