British Jorrrnal
of Hacmatology, 1975,31, 449.
Microthrombocytosis and Platelet Fragmentation Associated with Idiopathic/Autoimmune Thrombocytopenic Purpura IQBAL KHAN,DOROTHEA ZUCKER-FRANKLIN AND SIMON KARPATKIN
Department $Medicine, New York Univcrzity School of Medicine
(Received
10 April
1975; acceptedfor publication 16 M a y 1975)
SUMMARY. Platelet volume distribution curves were obtained in 20 control subjects and in z I patients with idiopathic/autoimmune thrombocytopenic purpura. A striking increase in microthrombocytes as well as megathrombocytes was noted in 86% of patients on one or more occasions, particularly in the presence of severe thrombocytopenia. The entire spectrum of platelet volume distribution curves noted iii patients could be reproduced experimeiitally in rabbits following intravenous injection of anti-platelet antibody. Differential centrifugation studies with coiitrol subjects revealed that microthrombocytes were light platelets and megathrombocytes were heavy platelets. Electron microscopy in patients with thrombocytopenia revealed that microthrombocytes were composed of intact small platelets as well as platelet fragments. It is concluded that severe peripheral destruction of platelets is associated with an increase in microthrombocytes as well as megathronibocytes. Increased megathrombocytes are associated with increased peripheral destruction or utilization of platelets (Garg et al, 1969, 1971, 1972; Karpatkin et a!, 1971; Chatterji & Lynch, 1971; Murphy et al, 1972; Sahud, 1972; Kraytman, 1973). Little is known about the remainder of the platelet volume distribution. The development of the more sensitive P-64 Channel Analyser with electronic recorder attached to a Coulter Model B Counter has now made the measurement of a total platelet volume distribution more reliable. The platelet volume distribution was measured in 21 patients with idiopathic/autoimmune thrombocytopenic purpura (ATP/ITP) with varying platelet counts, who were studied on several occasions. A striking increase in microthrombocytes was noted in 86% of the patients, particularly in the presence of severe thrombocytopenia. The entire spectrum of platelet volume distribution with increased microthrombocytes as well as megathrombocytes was reproduced in rabbits injected with anti-platelet antibody. The present report provides the methodology and documentation for these observations. METHODS
The patients studied were diagnosed as having chronic ATP/ITP as determined by criteria previously reported (Karpatkin et al, 1971). Correspondence: Dr S. Karpatkin, New York University School of Medicine, 550 First Avenue;New York, N.Y. 10016,U.S.A.
449
450
I. Khqn, D. Zucker-Franklin and S. Karpatkin
New Zealand white rabbits weighing 3-4 kg were employed for the induction of thrombocytopenia with anti-rabbit platelet antibody, obtained from guinea-pigs. The guinea-pig anti-rabbit antiplatelet antibody was prepared and adsorbed, as described previously (Weintraub & Karpatkin, 1974). Blood was drawn into EDTA-anticoagulated Vacutainer tubes (Becton, Dickinson and Co., Rutherford, N.J.). All operations were performed at room temperature unless otherwise noted, Platelet counts were determined manually, by phase optics, employing 3% procaine hydrochloride as diluent (Karpatkin et all 1971). Platelet rich plasma was obtained by centrifugation in plastic z mm diameter tubes at 250 g for 30 s. The specimen was diluted in isoton (Coulter Electronics, Hialeah, Florida) with the aid of 3.3 pl pipette to a platelet count of 5000-10 ooo per 0.1 ml volume in order to reduce coincidence counting to less than I%. A P-64 Channel Analyser with automatic electronic recording device was attached to a ‘transistorized’ Model B Coulter Counter with a 70 pm aperture tube and calibrated with latex particles of 3.35 fl so that windows 4-~oorepresentedz-25 fl volumes. The diluting fluid, time interval between blood removal and volume measurement, and concentration of platelets were held constant at all times. Preliminary studies revealed no change in platelet volume distribution over a z h interval, with a tendency toward a shift of the platelet volume distribution curve to the right at longer time intervals. All samples were measured within 2 h. Platelet volume distribution was measured at the time interval required for the counter to enumerate a fixed number of caunts, employing the ‘count mode’ switch (1024 counts). This time interval in 20 normal subjects as well as patients with normal platelet counts was 7-8 s. Longer time intervals were obtained with low platelet counts or with increased numbers of megathrombocytes (i.e. a wider platelet volume distribution). These conditions were at risk to ‘electronic noise’ at the lower window thresholds. This could be circumvented to negligible interference by either of three procedures. (I) Addition of sdicient platelets to the isoton cuvette to achieve a platelet count of 5000-10 OOO/O:I ml. ( 2 ) Monitoring the same sample on the ‘time mode’, and fixing this interval at LO s. (3) Running a blank consisting of an aliquot of platdet poor plasma in isoton on the ‘time mode’ switch for the interval of time required to count the platelets with the ‘count mode’ switch. Under these conditions, significant ‘electronic noise’ interference did not occur until time intervals of approximately 30-45 s. The introduction ofsignificant noise (i.e. > 10% of a microthrombocytosispeak) was rarely encountered These samples were not included in this study. All measurements were made in duplicate. The aperture tube was kept clear by continual monitoring of the oscilloscope screen for ‘interference’ distribution patterns. Background counts, as determined on the Coulter Model B, were kept at below 50 prior to use of the P-64 Channel Analyser. Twenty healthy laboratory personnel were employed to determine the mean platelet volume distribution scan given in Fig I. Note the skewing to the right. The shaded area represents f2 standard deviations (SD) &om the mean. Microthrombocytes were arbitrarily defined by a mean i-2 SD,Lso value. The Lso represents the 50 percentile point on the first ascending limb (left-side) of the platelet volume distribution curve where it intercepts the horizontal axis (Lso= 8.38 f3.18). Megathrombocytes were similarly arbitrarily defined by a mean f 2 SD, Rlo value. The R, represents the 10th percentile point on the descending limb (right side) of the platelet volume distribution curve where it intercepts the horizontal
Microthrombocytosis in Thrombocytopenia
451
axis (Rlo = 55.1 & 15.7). The mean peak value, or mode, was similarly obtained + z SD (mode = 17.5+ 5.64). Per cent megathrombocytes were also arbitrarily defined for Table I1 as the per cent total platelets having a volume of 12.5 fl or greater (i.e. all platelets enumerated between windows 50 and IOO divided by all platelets enumerated between windows 4 and 100, multiplied by 100). For ultrastructural studies platelets obtained from three patients with thrombocytopenia (MUN, GAR, JOH) were processed as follows: (I) The anticoagulated whole blood was centrifuged at zoo g for 8 min. The platelet rich plasma and buffy coat cells were removed and placed directly into 6% phosphatebuffered glutaraldehyde, 0.067 M, pH 7.4. (2) The platelet rich plasma without the buffy coat layer was centrifuged at 2800 g for 10 min. The plasma was removed and the sedimented platelets resuspended in 3% phosphatebuffered glutaraldehyde. (3) The plasma of above was centrifuged at 27 ooo g for 3 0 min at 4°C. The resulting sediment was also fixed with 3 % phosphate-buffered glutaraldehyde. Similar, though preliminary studies were carried out on platelets isolated from rabbits which had been treated with antiserum to rabbit platelets. Fixation with glutaraldehyde took place for 1-24 h. The specimens were then processed as detailed elsewhere (Zucker-Franklin, 1969). A Siemens Elmiskop I electron microscope was used to view thin sections. In order to facilitate comparison of platelet size all survey electron photomicrographs were taken at an accelerating voltage of 60 kV, with the same apertures and uniform instrument magnification of 2500. RESULTS The mean platelet volume distribution curve for 20 healthy laboratory personnel f z SD is given in Fig I. Eighty-two platelet volume distribution scans were evaluated in 21 patients with ATP/ITP. Their age, sex, number of determinations, platelet count range and identification symbol are cited in Table I. Besides the increase in megathrombocytes previously reported in hyperdestructive platelet disorders a striking increase in microthrombocytes was also noted. Representative examples of eight different patients are reported in Fig 2. These reveal a shift of the platelet volume distribution curve to the left as well as the right. The shift toward the left with smaller platelets was demonstrable as either a separate peak distinguishable from the remainder of the platelet volume distribution as in Fig 2 0 and Fig 2E-H; or as a shift of the ciitirc curve to the lcft as in Figs ZB and 2C. A marked shift to the left was generally associated with low platelet counts of less than 60x 10~11. The data for all 82 measurements are plotted in Fig 3 , where the Lso and peak window size of these curves are compared to the control curve & 2 SD.86% of the 21 patients demonstrated a pronounced shift to the left (ix. Lsoless than 2 SD from the control curve), on one or more occasions. At least 50% of the total number of 82 measurements on these 21 patients were similarly shifted. As shown in Figs 4(a) and 4(b), platelets obtained by centrifugation of platelet rich plasma at 2800 g from a normal subject were about twice as large as those sedimeiited from patient
I. Khan, D. Zucker-Franklin and S . Kaspatkin
Window size
FIG I. Mean platelet volume distribution curve for 20 normal subjects + z Standard Deviations. An aliquot of platelet-rich plasma was diluted in isoton and then monitored on a Coulter Model B attached to a P64 Channel Analyser and automatic electronic recorder, employing a 70 pm aperture tube. Each window was calibrated to equal 0.25 fl.
Patient MUN SAT GAR RE1 JOH SIM
FIN ME
YAN ROD BRO MIL RIN STR SLA GEL BUR WAR HAS SCH
BEN
Age
Sex
No. of determinations
Platelet count range x 10911.
39
M
25
31
I1 I1
1-1400 10-45 10-4s
39 16
F F M F F F F M
a1
M
60
F F F
75 16 27
53 a2
27
65
5 4 4 3 3 2 2 2
40-90 40-120 70-80 15-90
98-1I 5 3-60 7-200
I I
65
90
81 59
M
I0
M
I I I
a8 6 35 a6
F
I
M M M
I I I
17
F
I
F
0.5-590
45 90 53 45 6 60 I2 20
Symbol+ Open square Closed circle Open circle Closed triangle Open triangle Closed square Closed star Inverted closed triangle Closed diamond Closed diamond Closed diamond Closed diamond Closed diamond Closed diamond Closed b o n d Closed diamond Closed diamond Closed diamond Closed diamond Closed diamond Closed diamond
Microthrombocytosis in Thrombocytopenia
G
00 -
H
-
60 40
453
20 x 10911
-
-
20 I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
)
454
I. Khan, D.Zutker-Franklin and S. Karpatkirt
JOH at 27 000 g following removal of her 2800 g pellet. Her 2800 g fraction contained platelets of the same size as the normal subject’s 2800 g fraction. The 27 000 g fraction comprised less than 10% of the platelets contained in the specimen and was not seen in the sedinient of a similarly prepared control subject’s plasma. It is noteworthy that the centrifugal force and the packing of these small platelets did not cause degradation or other morphological changes associated with aggregation. Microthrombocytes appear to have fewer pseudopods than normal-sized platelets sedimented at 2800 g.
P.SC
L “T ir. +
88
12’1 10.0
2* 0
L50
Peak
FIG 3. Scattergram of ‘Lso’and ‘peak‘ windows of 21 different patients studied on 82 different occasions. Each symbol refers to patients cited in Table I. The area within the two boxes refers to + z SDs.
Because previous work had suggested that smaller platelets were lighter platelets (Karpatkin, 19%; Amorosi et a/, 1971; Charmatz & Karpatkin, 1974), an experiment was designed to
determine whether the increased microthrombocytes seen in these patients could be simulated by differential centrifugation of normal platelet-rich plasma in which presumably largerheavier platelets could be removed preferentially. The data from Fig 5 and Table I1 indicate that large platelets are heavier platelets and that lighter platelets can be isolated by differential centrifugation of normal platelet rich plasma, resulting in platelet volume distribution curves similar to that seen in our patients. Similar data were obtained using EDTA, citrate or oxalate as anticoagulant (I I experiments).
P,
k
FIG 4. (a) Platelets obtained from PRP of a control subject show normal size. (b) Platelets obtained from patient JOH following recentrifugation of her plasma at 27 ooo g. These platelets are obviously smaller. Both at same magnification, x 8000.
Microthonibocytosis in Thronibocytopenia
0
,
I
I
,
I
10 20 30 40 50 60 70 80 90 100
-
-I Zoo0 rpm
-
-- / I
I
I
2200 rpm
l
l
1
10 20 30 40 50 60 70 80 90 100 1 0 20 30 40 50 60 70 80 90 100 Window size
1
I
1
1
'
1
FIG 5 . Platelet volume distribution curves from EDTA-anticoagulated blood of a normal subject following differential centrifugation at increasing rpm in a Sorvall RC3 cenmfuge at 4°C. See Table I1 for further details.
-n
1
1600 rpm
I. Khan, D. ZuckeriFrunklin
456
atid
S.Kurpatkiri
The next experiment was designed to determine whether similar platelet volume distribution curves could be reproduced in rabbits following the injection of 1.5 ml of antiplatelet antibody. Fig 6 demonstrates a similar increase in microthrombocytes, associated with antibody-induced thrombocytopenia. In this experiment, the increase in microihrombocytes was noted within 15 min. At 30 min, there was an associated increase in megathrombocytes. At 5 h post anti-platelet antibody infusion, the platelet volume distribution appeared to return toward normal. However, at 24 h increased microthrombocytes and megathrombocytes were again noted. Restoration towards normal did not occur until the fifth and sixth day with absolute megathronibocytosis noted on the fourth day. The increase in rhBLB 11. Differential centrifugation of EDTA-anticoagulated blood*
Platelet count RPM
Vol
(MI)
W.B.
5.0
600
1.4 I .6 1.8 1.9
800 1000 I200
1400 I600 I800 2Ooo
2200
2400 2600
2.0 2.0
2.1 2.1 2.1 2.2
2.1
x
10911.
I75 380 350 400 175 130 115 I00
40 60 I0
5
% Megathrombocytes 14.98 12.12
8.68 7.40 5.99 5.81
4.74 5.02
6.14 6.35 7.95 7.96
Megathrombocytes x 10~11. 26.220 46.042 30.380 29.593 10.487 7.552 5.451 5.023 2.457 l.905 0.795 0.398
% PIatelet yield 100.0
60.8 64.0 82.3 38.0 29.7 26.3 24.0 9.6 7.2 2.5
1.2
5 ml of EDTA-anticoagulated blood was repeatedly centrifuged at increasing rpm in a Sorvall
RC3 centrifugue (column I) for 20 nlin at 4"C, in a graduated, calibrated centrifuge tube. The volume of the platelet-rich plasma was recorded following each centrifugation and is given in column 2. The platelet count of the platelet-rich plasma is given in column 3. The megathrombocyte number for the platelet-rich plasma is given in column 5 , and calculated by multiplying the % megathrombocytesby the platelet count. The platelet yield is given in column 6 and determined as per cent of the original whole blood (W.B.)platelet count and volume, given on line I columns I , 2 and 3.
megathrombocytes persisted longer than the increase in microthrombocytes. Similar results were obtained in I 2 such experiments. Preliminary electron microscopic studies performed on platelets obtained fiom rabbits 15 min after treatment with anti-rabbit platelet serum showed marked poikilocytosis and fragmentation of platelets in platelet rich plasm centrifuged at 2800 g. Similar fragmentation was seen in the 2800 g specimen of GAR, one of the three patients who was studied. Two further experiments were performed in splenectomized rabbits (one of which is cited in Fig 7) in order to determine the contribution of the spleen to the platelet volume distribution. Again, both an increase in microthrombocytes and megathrombocytes were noted as early as 5 min post antiplatelet antibody infusion. The increase in microthrombocytes persisted for at least 9 h whereas the increase in megathrombocytes was present at 24 h.
Microthrombocytosis irz Thrombocytopenia
457
\
\
440x109/L.
0 10 20 30 40 50
0
10 20 30 40 50
Window size
FIG 6. Platelet volume distribution curves of an intact rabbit (basal platelet count 324 ooo/min3) following a single intravenous dose of 1.5 ml of guinea-pig anti-rabbit platelet antibody. The time interval and platelet count pre-injection (-) and post-injection (- - -) is given in each panel.
458
I. Khan, D. Zucker-Franklin and S. Karpatkin
I
0
\
10 20 30 4 0 50
Window a i m
FIG7. Platelet volume distributioncurvesin a splenectomizedrabbit (basal platelet count 340 000/nun3) following a single intravenous dose of 1.5 ml of guinea-pig anti-rabbit platelet antibody. The time interval and platelet count pre-injection (-) and post-injection (- - -) are given in each panel.
DISCUSSION These data clearly reveal the presence of increased microthrombocytes in addition to increased megathrombocytes (previously reported by Garg et a!, 1969, 1971) in hyperdestructive platelet disorders. The increase in microthrombocytes was observed in 86% of the 21 patients, who were measured on 82 different occasions. This presented as a distinct peak or as a shift of the entire platelet volume distribution curve to the left. It was only observed in patients with severe thrombocytopenia, i.e. platelet counts of less than 6 0 x 109/l. These observations have recently been reported in abstract form by Karpatkin & Khan (1974) and Prost-Dvojakovic et a1 (1974). The platelet volume distribution patterns noted in these 21 patients could be simulated in rabbits subjected to intravenous anti-platelet antibody. A striking shift to the left was noted as early as 5 min. This could be demonstrated in intact as well as splenectomized rabbits, suggesting that the spleen was not required and that the mechanism for this response may be intravascular. The time course for the development of increased microthrombocytes and megathrombocytes is of interest. An increase in microthrombocytes was always noted at the earliest time
Microthrombocytosis in Thrombocytopenia
459
point at which a specimen could be obtained, and usually returned to normal prior to the disappearance of increased megathrombocytes. In the splenectomized rabbit, the increase in microthrornbocytes had disappeared within 9 h whereas increased megathrombocytes were present at 24 h. In the intact rabbit there appeared to be two waves of increased microthroml bocytes and megathrombocytes. During the first wave, the increased microthrornbocytes disappeared within 3-5 h along with partial disappearance of the increase in megathrombocytes; at 24 h a resumption of the pattern of increased microthrornbocytes and megathromon bocytes was again noted, with a drop in platelet count from zoo x 1o9/I. to 125 x 10~11. day 2 and further drop in platelet count to 80x 109/l. on day 3. The shift to the left disap peared by the third day, whereas increased megathrombocytes were still present on days 4, 5 and 6, before disappearing on day 7. It is postulated that the second wave of thrombocytopenia with shifts in the platelet volume disp-ibution curve may be due to release of antibody from antibody-coated platelets which were destroyed in the reticuloendothelial system. It should also be noted that megathrombocytosis was noted on days 4, 5 and 6 in the intact rabbit, and at 24 h in the splenectomized rabbit. The increased microthrombocyte pattern could be obtained by differential centrifugation of normal platelet-rich plasma, following removal of the larger-heavier platelets. This suggests that these microthrom%ocytesseen in clinical conditions and simulated in the rabbit are ‘light’ platelets. This is substantiated by the electron microscopic studies, wherein smaller platelets were observed in the patient’s ‘platelet-poor plasma’ after further centrifugation at 27 ooo g for 3 0 min. The mechanism for the production of microthrombocytes as well as platelet fragments remains obscure. Microthrombocytes could be produced by the release of two or more smaller platelets from a larger megathronibocyte. This release of platelets intravascularly has been postulated as a possible physiologic mechanism for the production of platelets during a thrombopoietic stress (Garg et a!, 1971; Karpatkin & Garg, 1974). If this mechanism does exist, it is conceivable that antibody injury might also lead to the conversion of megathrombocytes to microthrombocytes. It is also possible that microthrornbocytes are more resistant to antibody damage and/or removal by the spleen compared to the remainder of the platelet population. Platelet fragmentation may be secondary to antibody-mediated injury to 4 single platelet or group of platelets. Such a mechanism could result from the formation of platelet aggregates. These are known to have irregular shapes with many pseudopods (Zucker-Franklin, 1970). Pseudopods may be broken off following shearing forces. This postulate is supported by the observation that aggregates were noted at 5 min and that many of the platelet fragments appeared to be derived from platelet pseudopods. Regardless of the mechanisms, the observation of increased microthrornbocytes and platelet fragmentation associated with increased megathrombocytes, provides a useful clinical tool for the prediction of severe hyperdestructive immunologic disorders. ACKNOWLEDGMENTS
This work was supported by grant 13336-05 of the National Heart and Lung Institute and Contract DADA 17-68-C-8163 of the U.S. Army Research and Development Command. The authors are indebted to Ira Fink for excellent technical assistance.
I.Khan, D. Zucker-Franklin and S . Karppetkin
460
This work was presented at the 31st annual meeting of the American Federation for Clinical Research, Atlantic City, New Jersey, 5 May 1974(Clinical Research, 1g74,22,3964). REFERENCES AMOROSI,E., GARG,S.K. & KARPATKIN,S. (1971) Heterogeneityof human platelets. N.Identification of a young platelet population with [Se75]selenomethionhe. British Joumal ofHaematology, ax, 227. CHARMATZ, A. & KARPATICIN, S. (1974) Heterogeneity of rabbit platelets. I. Employment of an albumin density gradient for separation of a young platelet population identified with Se75-selenomethionine. Thrombosis et Diathesis Haemorrhagica, 31, 11s. CHAITERJI,A.K. & LYNCH,E.C. (1971) Circulating large platelets. (Letter). New England Joumal of Medicine, 284, 1440. GARG,S.K., AMOROSI, E.L. & KARPATKIN, S. (1969) The large platelet on peripheral smear as an index of thrombopoiesis. Blood, 31, 851. GARG,S.K.,AMOROSI, E.L. & KARPATKIN, S. (1971) Use of the megathrombocyteas an index of megakaryocyte number. New EnglandJournal ofMedicine, 284,
11.
GARG, S.K.,LACKNHB, H. & KARPATKIN, S. (1972) The increased percentage of megathrombocytes in various clinical disorders. Annals oflntemal Medicine, 77.361. KAWATIUN,S. (19%) Heterogeneity of human platelets. I. Metabolic and kinetic evidence.suggestive of young and old plateletr.Journul of Clinical Invest&tion, 48, 1073. KAEPATIUN,S. & GARG,S.K.(1974) The megathrombocyte as an index of platelet production. (Annotation). British Journal of Haematology, 16, 307. KAWATKIN,S., GARG,S.K.& SISKIND, G.W. (1971) Autoimmune thrombocytopenic purpura and the compensated thrombocytolytic state. American Journal OfMedicine, 51, I.
KARPATKIN, S. &KHAN, I. (1974) Intravascular thrombocytolysis in autoimmune thrombocytopenic purpura (ATP). (Abstract). Clinical Research, aa,
KBAnw, M.(1973) Platelet size in thrombocytopenias and thrombocytosis of various origin. Blood. 4 1 s 587.
MURPHY, S., O m ,F.A., NAUIAN, J.L., LUSCH,C.J., GOLDBERG, S. & G m m , F.H.(1972.) Platelet size and kinetics in hereditary and acquired thrombocytopenia. New England Journal .f Medicine, a86, 499. PROST-DVOJAKOVIC, R.J.,LE TOBIC,F., DESNOYERS. P., LABAUME, J. & SAMANA, M.(1974) Critical study of the distribution of platelet volumes in thrombocytopenic purpuras. XV Congress o j the International Society ofHematology,Jerusalem, p 197. SAHUD, M.A. (1972) Platelet size and number in alcoholic thrombocytopenia. New England Journal of Medicine, 186, 355. WEINTRAUB, A.H. & KARPATKIN,S. (1974) Heterogeneity ofrabbit platelets. 11. Use ofthe megathrombocyte to demonstrate a thrombopoietic stimulus. Joumal of Laboratory and Clinical Medicine, 83, 896. ZUCKHR-FRANKLIN, D. (1969) Microfibrils of blood platelets: their relationship to microtubules and contractile protein. Journal 9 f Clinical Investigation, 48,165. ZUCKBB-FRA", D. (1970) The ultrastructure of megakaryocytes and platelets. Regulation of Hematopoiesis (ed. by A. S. Gordon), p 1553. AppletonCentury-Crofts, New York.
of Hacmatology, 1975,31, 449.
Microthrombocytosis and Platelet Fragmentation Associated with Idiopathic/Autoimmune Thrombocytopenic Purpura IQBAL KHAN,DOROTHEA ZUCKER-FRANKLIN AND SIMON KARPATKIN
Department $Medicine, New York Univcrzity School of Medicine
(Received
10 April
1975; acceptedfor publication 16 M a y 1975)
SUMMARY. Platelet volume distribution curves were obtained in 20 control subjects and in z I patients with idiopathic/autoimmune thrombocytopenic purpura. A striking increase in microthrombocytes as well as megathrombocytes was noted in 86% of patients on one or more occasions, particularly in the presence of severe thrombocytopenia. The entire spectrum of platelet volume distribution curves noted iii patients could be reproduced experimeiitally in rabbits following intravenous injection of anti-platelet antibody. Differential centrifugation studies with coiitrol subjects revealed that microthrombocytes were light platelets and megathrombocytes were heavy platelets. Electron microscopy in patients with thrombocytopenia revealed that microthrombocytes were composed of intact small platelets as well as platelet fragments. It is concluded that severe peripheral destruction of platelets is associated with an increase in microthrombocytes as well as megathronibocytes. Increased megathrombocytes are associated with increased peripheral destruction or utilization of platelets (Garg et al, 1969, 1971, 1972; Karpatkin et a!, 1971; Chatterji & Lynch, 1971; Murphy et al, 1972; Sahud, 1972; Kraytman, 1973). Little is known about the remainder of the platelet volume distribution. The development of the more sensitive P-64 Channel Analyser with electronic recorder attached to a Coulter Model B Counter has now made the measurement of a total platelet volume distribution more reliable. The platelet volume distribution was measured in 21 patients with idiopathic/autoimmune thrombocytopenic purpura (ATP/ITP) with varying platelet counts, who were studied on several occasions. A striking increase in microthrombocytes was noted in 86% of the patients, particularly in the presence of severe thrombocytopenia. The entire spectrum of platelet volume distribution with increased microthrombocytes as well as megathrombocytes was reproduced in rabbits injected with anti-platelet antibody. The present report provides the methodology and documentation for these observations. METHODS
The patients studied were diagnosed as having chronic ATP/ITP as determined by criteria previously reported (Karpatkin et al, 1971). Correspondence: Dr S. Karpatkin, New York University School of Medicine, 550 First Avenue;New York, N.Y. 10016,U.S.A.
449
450
I. Khqn, D. Zucker-Franklin and S. Karpatkin
New Zealand white rabbits weighing 3-4 kg were employed for the induction of thrombocytopenia with anti-rabbit platelet antibody, obtained from guinea-pigs. The guinea-pig anti-rabbit antiplatelet antibody was prepared and adsorbed, as described previously (Weintraub & Karpatkin, 1974). Blood was drawn into EDTA-anticoagulated Vacutainer tubes (Becton, Dickinson and Co., Rutherford, N.J.). All operations were performed at room temperature unless otherwise noted, Platelet counts were determined manually, by phase optics, employing 3% procaine hydrochloride as diluent (Karpatkin et all 1971). Platelet rich plasma was obtained by centrifugation in plastic z mm diameter tubes at 250 g for 30 s. The specimen was diluted in isoton (Coulter Electronics, Hialeah, Florida) with the aid of 3.3 pl pipette to a platelet count of 5000-10 ooo per 0.1 ml volume in order to reduce coincidence counting to less than I%. A P-64 Channel Analyser with automatic electronic recording device was attached to a ‘transistorized’ Model B Coulter Counter with a 70 pm aperture tube and calibrated with latex particles of 3.35 fl so that windows 4-~oorepresentedz-25 fl volumes. The diluting fluid, time interval between blood removal and volume measurement, and concentration of platelets were held constant at all times. Preliminary studies revealed no change in platelet volume distribution over a z h interval, with a tendency toward a shift of the platelet volume distribution curve to the right at longer time intervals. All samples were measured within 2 h. Platelet volume distribution was measured at the time interval required for the counter to enumerate a fixed number of caunts, employing the ‘count mode’ switch (1024 counts). This time interval in 20 normal subjects as well as patients with normal platelet counts was 7-8 s. Longer time intervals were obtained with low platelet counts or with increased numbers of megathrombocytes (i.e. a wider platelet volume distribution). These conditions were at risk to ‘electronic noise’ at the lower window thresholds. This could be circumvented to negligible interference by either of three procedures. (I) Addition of sdicient platelets to the isoton cuvette to achieve a platelet count of 5000-10 OOO/O:I ml. ( 2 ) Monitoring the same sample on the ‘time mode’, and fixing this interval at LO s. (3) Running a blank consisting of an aliquot of platdet poor plasma in isoton on the ‘time mode’ switch for the interval of time required to count the platelets with the ‘count mode’ switch. Under these conditions, significant ‘electronic noise’ interference did not occur until time intervals of approximately 30-45 s. The introduction ofsignificant noise (i.e. > 10% of a microthrombocytosispeak) was rarely encountered These samples were not included in this study. All measurements were made in duplicate. The aperture tube was kept clear by continual monitoring of the oscilloscope screen for ‘interference’ distribution patterns. Background counts, as determined on the Coulter Model B, were kept at below 50 prior to use of the P-64 Channel Analyser. Twenty healthy laboratory personnel were employed to determine the mean platelet volume distribution scan given in Fig I. Note the skewing to the right. The shaded area represents f2 standard deviations (SD) &om the mean. Microthrombocytes were arbitrarily defined by a mean i-2 SD,Lso value. The Lso represents the 50 percentile point on the first ascending limb (left-side) of the platelet volume distribution curve where it intercepts the horizontal axis (Lso= 8.38 f3.18). Megathrombocytes were similarly arbitrarily defined by a mean f 2 SD, Rlo value. The R, represents the 10th percentile point on the descending limb (right side) of the platelet volume distribution curve where it intercepts the horizontal
Microthrombocytosis in Thrombocytopenia
451
axis (Rlo = 55.1 & 15.7). The mean peak value, or mode, was similarly obtained + z SD (mode = 17.5+ 5.64). Per cent megathrombocytes were also arbitrarily defined for Table I1 as the per cent total platelets having a volume of 12.5 fl or greater (i.e. all platelets enumerated between windows 50 and IOO divided by all platelets enumerated between windows 4 and 100, multiplied by 100). For ultrastructural studies platelets obtained from three patients with thrombocytopenia (MUN, GAR, JOH) were processed as follows: (I) The anticoagulated whole blood was centrifuged at zoo g for 8 min. The platelet rich plasma and buffy coat cells were removed and placed directly into 6% phosphatebuffered glutaraldehyde, 0.067 M, pH 7.4. (2) The platelet rich plasma without the buffy coat layer was centrifuged at 2800 g for 10 min. The plasma was removed and the sedimented platelets resuspended in 3% phosphatebuffered glutaraldehyde. (3) The plasma of above was centrifuged at 27 ooo g for 3 0 min at 4°C. The resulting sediment was also fixed with 3 % phosphate-buffered glutaraldehyde. Similar, though preliminary studies were carried out on platelets isolated from rabbits which had been treated with antiserum to rabbit platelets. Fixation with glutaraldehyde took place for 1-24 h. The specimens were then processed as detailed elsewhere (Zucker-Franklin, 1969). A Siemens Elmiskop I electron microscope was used to view thin sections. In order to facilitate comparison of platelet size all survey electron photomicrographs were taken at an accelerating voltage of 60 kV, with the same apertures and uniform instrument magnification of 2500. RESULTS The mean platelet volume distribution curve for 20 healthy laboratory personnel f z SD is given in Fig I. Eighty-two platelet volume distribution scans were evaluated in 21 patients with ATP/ITP. Their age, sex, number of determinations, platelet count range and identification symbol are cited in Table I. Besides the increase in megathrombocytes previously reported in hyperdestructive platelet disorders a striking increase in microthrombocytes was also noted. Representative examples of eight different patients are reported in Fig 2. These reveal a shift of the platelet volume distribution curve to the left as well as the right. The shift toward the left with smaller platelets was demonstrable as either a separate peak distinguishable from the remainder of the platelet volume distribution as in Fig 2 0 and Fig 2E-H; or as a shift of the ciitirc curve to the lcft as in Figs ZB and 2C. A marked shift to the left was generally associated with low platelet counts of less than 60x 10~11. The data for all 82 measurements are plotted in Fig 3 , where the Lso and peak window size of these curves are compared to the control curve & 2 SD.86% of the 21 patients demonstrated a pronounced shift to the left (ix. Lsoless than 2 SD from the control curve), on one or more occasions. At least 50% of the total number of 82 measurements on these 21 patients were similarly shifted. As shown in Figs 4(a) and 4(b), platelets obtained by centrifugation of platelet rich plasma at 2800 g from a normal subject were about twice as large as those sedimeiited from patient
I. Khan, D. Zucker-Franklin and S . Kaspatkin
Window size
FIG I. Mean platelet volume distribution curve for 20 normal subjects + z Standard Deviations. An aliquot of platelet-rich plasma was diluted in isoton and then monitored on a Coulter Model B attached to a P64 Channel Analyser and automatic electronic recorder, employing a 70 pm aperture tube. Each window was calibrated to equal 0.25 fl.
Patient MUN SAT GAR RE1 JOH SIM
FIN ME
YAN ROD BRO MIL RIN STR SLA GEL BUR WAR HAS SCH
BEN
Age
Sex
No. of determinations
Platelet count range x 10911.
39
M
25
31
I1 I1
1-1400 10-45 10-4s
39 16
F F M F F F F M
a1
M
60
F F F
75 16 27
53 a2
27
65
5 4 4 3 3 2 2 2
40-90 40-120 70-80 15-90
98-1I 5 3-60 7-200
I I
65
90
81 59
M
I0
M
I I I
a8 6 35 a6
F
I
M M M
I I I
17
F
I
F
0.5-590
45 90 53 45 6 60 I2 20
Symbol+ Open square Closed circle Open circle Closed triangle Open triangle Closed square Closed star Inverted closed triangle Closed diamond Closed diamond Closed diamond Closed diamond Closed diamond Closed diamond Closed b o n d Closed diamond Closed diamond Closed diamond Closed diamond Closed diamond Closed diamond
Microthrombocytosis in Thrombocytopenia
G
00 -
H
-
60 40
453
20 x 10911
-
-
20 I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
)
454
I. Khan, D.Zutker-Franklin and S. Karpatkirt
JOH at 27 000 g following removal of her 2800 g pellet. Her 2800 g fraction contained platelets of the same size as the normal subject’s 2800 g fraction. The 27 000 g fraction comprised less than 10% of the platelets contained in the specimen and was not seen in the sedinient of a similarly prepared control subject’s plasma. It is noteworthy that the centrifugal force and the packing of these small platelets did not cause degradation or other morphological changes associated with aggregation. Microthrombocytes appear to have fewer pseudopods than normal-sized platelets sedimented at 2800 g.
P.SC
L “T ir. +
88
12’1 10.0
2* 0
L50
Peak
FIG 3. Scattergram of ‘Lso’and ‘peak‘ windows of 21 different patients studied on 82 different occasions. Each symbol refers to patients cited in Table I. The area within the two boxes refers to + z SDs.
Because previous work had suggested that smaller platelets were lighter platelets (Karpatkin, 19%; Amorosi et a/, 1971; Charmatz & Karpatkin, 1974), an experiment was designed to
determine whether the increased microthrombocytes seen in these patients could be simulated by differential centrifugation of normal platelet-rich plasma in which presumably largerheavier platelets could be removed preferentially. The data from Fig 5 and Table I1 indicate that large platelets are heavier platelets and that lighter platelets can be isolated by differential centrifugation of normal platelet rich plasma, resulting in platelet volume distribution curves similar to that seen in our patients. Similar data were obtained using EDTA, citrate or oxalate as anticoagulant (I I experiments).
P,
k
FIG 4. (a) Platelets obtained from PRP of a control subject show normal size. (b) Platelets obtained from patient JOH following recentrifugation of her plasma at 27 ooo g. These platelets are obviously smaller. Both at same magnification, x 8000.
Microthonibocytosis in Thronibocytopenia
0
,
I
I
,
I
10 20 30 40 50 60 70 80 90 100
-
-I Zoo0 rpm
-
-- / I
I
I
2200 rpm
l
l
1
10 20 30 40 50 60 70 80 90 100 1 0 20 30 40 50 60 70 80 90 100 Window size
1
I
1
1
'
1
FIG 5 . Platelet volume distribution curves from EDTA-anticoagulated blood of a normal subject following differential centrifugation at increasing rpm in a Sorvall RC3 cenmfuge at 4°C. See Table I1 for further details.
-n
1
1600 rpm
I. Khan, D. ZuckeriFrunklin
456
atid
S.Kurpatkiri
The next experiment was designed to determine whether similar platelet volume distribution curves could be reproduced in rabbits following the injection of 1.5 ml of antiplatelet antibody. Fig 6 demonstrates a similar increase in microthrombocytes, associated with antibody-induced thrombocytopenia. In this experiment, the increase in microihrombocytes was noted within 15 min. At 30 min, there was an associated increase in megathrombocytes. At 5 h post anti-platelet antibody infusion, the platelet volume distribution appeared to return toward normal. However, at 24 h increased microthrombocytes and megathrombocytes were again noted. Restoration towards normal did not occur until the fifth and sixth day with absolute megathronibocytosis noted on the fourth day. The increase in rhBLB 11. Differential centrifugation of EDTA-anticoagulated blood*
Platelet count RPM
Vol
(MI)
W.B.
5.0
600
1.4 I .6 1.8 1.9
800 1000 I200
1400 I600 I800 2Ooo
2200
2400 2600
2.0 2.0
2.1 2.1 2.1 2.2
2.1
x
10911.
I75 380 350 400 175 130 115 I00
40 60 I0
5
% Megathrombocytes 14.98 12.12
8.68 7.40 5.99 5.81
4.74 5.02
6.14 6.35 7.95 7.96
Megathrombocytes x 10~11. 26.220 46.042 30.380 29.593 10.487 7.552 5.451 5.023 2.457 l.905 0.795 0.398
% PIatelet yield 100.0
60.8 64.0 82.3 38.0 29.7 26.3 24.0 9.6 7.2 2.5
1.2
5 ml of EDTA-anticoagulated blood was repeatedly centrifuged at increasing rpm in a Sorvall
RC3 centrifugue (column I) for 20 nlin at 4"C, in a graduated, calibrated centrifuge tube. The volume of the platelet-rich plasma was recorded following each centrifugation and is given in column 2. The platelet count of the platelet-rich plasma is given in column 3. The megathrombocyte number for the platelet-rich plasma is given in column 5 , and calculated by multiplying the % megathrombocytesby the platelet count. The platelet yield is given in column 6 and determined as per cent of the original whole blood (W.B.)platelet count and volume, given on line I columns I , 2 and 3.
megathrombocytes persisted longer than the increase in microthrombocytes. Similar results were obtained in I 2 such experiments. Preliminary electron microscopic studies performed on platelets obtained fiom rabbits 15 min after treatment with anti-rabbit platelet serum showed marked poikilocytosis and fragmentation of platelets in platelet rich plasm centrifuged at 2800 g. Similar fragmentation was seen in the 2800 g specimen of GAR, one of the three patients who was studied. Two further experiments were performed in splenectomized rabbits (one of which is cited in Fig 7) in order to determine the contribution of the spleen to the platelet volume distribution. Again, both an increase in microthrombocytes and megathrombocytes were noted as early as 5 min post antiplatelet antibody infusion. The increase in microthrombocytes persisted for at least 9 h whereas the increase in megathrombocytes was present at 24 h.
Microthrombocytosis irz Thrombocytopenia
457
\
\
440x109/L.
0 10 20 30 40 50
0
10 20 30 40 50
Window size
FIG 6. Platelet volume distribution curves of an intact rabbit (basal platelet count 324 ooo/min3) following a single intravenous dose of 1.5 ml of guinea-pig anti-rabbit platelet antibody. The time interval and platelet count pre-injection (-) and post-injection (- - -) is given in each panel.
458
I. Khan, D. Zucker-Franklin and S. Karpatkin
I
0
\
10 20 30 4 0 50
Window a i m
FIG7. Platelet volume distributioncurvesin a splenectomizedrabbit (basal platelet count 340 000/nun3) following a single intravenous dose of 1.5 ml of guinea-pig anti-rabbit platelet antibody. The time interval and platelet count pre-injection (-) and post-injection (- - -) are given in each panel.
DISCUSSION These data clearly reveal the presence of increased microthrombocytes in addition to increased megathrombocytes (previously reported by Garg et a!, 1969, 1971) in hyperdestructive platelet disorders. The increase in microthrombocytes was observed in 86% of the 21 patients, who were measured on 82 different occasions. This presented as a distinct peak or as a shift of the entire platelet volume distribution curve to the left. It was only observed in patients with severe thrombocytopenia, i.e. platelet counts of less than 6 0 x 109/l. These observations have recently been reported in abstract form by Karpatkin & Khan (1974) and Prost-Dvojakovic et a1 (1974). The platelet volume distribution patterns noted in these 21 patients could be simulated in rabbits subjected to intravenous anti-platelet antibody. A striking shift to the left was noted as early as 5 min. This could be demonstrated in intact as well as splenectomized rabbits, suggesting that the spleen was not required and that the mechanism for this response may be intravascular. The time course for the development of increased microthrombocytes and megathrombocytes is of interest. An increase in microthrombocytes was always noted at the earliest time
Microthrombocytosis in Thrombocytopenia
459
point at which a specimen could be obtained, and usually returned to normal prior to the disappearance of increased megathrombocytes. In the splenectomized rabbit, the increase in microthrornbocytes had disappeared within 9 h whereas increased megathrombocytes were present at 24 h. In the intact rabbit there appeared to be two waves of increased microthroml bocytes and megathrombocytes. During the first wave, the increased microthrornbocytes disappeared within 3-5 h along with partial disappearance of the increase in megathrombocytes; at 24 h a resumption of the pattern of increased microthrornbocytes and megathromon bocytes was again noted, with a drop in platelet count from zoo x 1o9/I. to 125 x 10~11. day 2 and further drop in platelet count to 80x 109/l. on day 3. The shift to the left disap peared by the third day, whereas increased megathrombocytes were still present on days 4, 5 and 6, before disappearing on day 7. It is postulated that the second wave of thrombocytopenia with shifts in the platelet volume disp-ibution curve may be due to release of antibody from antibody-coated platelets which were destroyed in the reticuloendothelial system. It should also be noted that megathrombocytosis was noted on days 4, 5 and 6 in the intact rabbit, and at 24 h in the splenectomized rabbit. The increased microthrombocyte pattern could be obtained by differential centrifugation of normal platelet-rich plasma, following removal of the larger-heavier platelets. This suggests that these microthrom%ocytesseen in clinical conditions and simulated in the rabbit are ‘light’ platelets. This is substantiated by the electron microscopic studies, wherein smaller platelets were observed in the patient’s ‘platelet-poor plasma’ after further centrifugation at 27 ooo g for 3 0 min. The mechanism for the production of microthrombocytes as well as platelet fragments remains obscure. Microthrombocytes could be produced by the release of two or more smaller platelets from a larger megathronibocyte. This release of platelets intravascularly has been postulated as a possible physiologic mechanism for the production of platelets during a thrombopoietic stress (Garg et a!, 1971; Karpatkin & Garg, 1974). If this mechanism does exist, it is conceivable that antibody injury might also lead to the conversion of megathrombocytes to microthrombocytes. It is also possible that microthrornbocytes are more resistant to antibody damage and/or removal by the spleen compared to the remainder of the platelet population. Platelet fragmentation may be secondary to antibody-mediated injury to 4 single platelet or group of platelets. Such a mechanism could result from the formation of platelet aggregates. These are known to have irregular shapes with many pseudopods (Zucker-Franklin, 1970). Pseudopods may be broken off following shearing forces. This postulate is supported by the observation that aggregates were noted at 5 min and that many of the platelet fragments appeared to be derived from platelet pseudopods. Regardless of the mechanisms, the observation of increased microthrornbocytes and platelet fragmentation associated with increased megathrombocytes, provides a useful clinical tool for the prediction of severe hyperdestructive immunologic disorders. ACKNOWLEDGMENTS
This work was supported by grant 13336-05 of the National Heart and Lung Institute and Contract DADA 17-68-C-8163 of the U.S. Army Research and Development Command. The authors are indebted to Ira Fink for excellent technical assistance.
I.Khan, D. Zucker-Franklin and S . Karppetkin
460
This work was presented at the 31st annual meeting of the American Federation for Clinical Research, Atlantic City, New Jersey, 5 May 1974(Clinical Research, 1g74,22,3964). REFERENCES AMOROSI,E., GARG,S.K. & KARPATKIN,S. (1971) Heterogeneityof human platelets. N.Identification of a young platelet population with [Se75]selenomethionhe. British Joumal ofHaematology, ax, 227. CHARMATZ, A. & KARPATICIN, S. (1974) Heterogeneity of rabbit platelets. I. Employment of an albumin density gradient for separation of a young platelet population identified with Se75-selenomethionine. Thrombosis et Diathesis Haemorrhagica, 31, 11s. CHAITERJI,A.K. & LYNCH,E.C. (1971) Circulating large platelets. (Letter). New England Joumal of Medicine, 284, 1440. GARG,S.K., AMOROSI, E.L. & KARPATKIN, S. (1969) The large platelet on peripheral smear as an index of thrombopoiesis. Blood, 31, 851. GARG,S.K.,AMOROSI, E.L. & KARPATKIN, S. (1971) Use of the megathrombocyteas an index of megakaryocyte number. New EnglandJournal ofMedicine, 284,
11.
GARG, S.K.,LACKNHB, H. & KARPATKIN, S. (1972) The increased percentage of megathrombocytes in various clinical disorders. Annals oflntemal Medicine, 77.361. KAWATIUN,S. (19%) Heterogeneity of human platelets. I. Metabolic and kinetic evidence.suggestive of young and old plateletr.Journul of Clinical Invest&tion, 48, 1073. KAEPATIUN,S. & GARG,S.K.(1974) The megathrombocyte as an index of platelet production. (Annotation). British Journal of Haematology, 16, 307. KAWATKIN,S., GARG,S.K.& SISKIND, G.W. (1971) Autoimmune thrombocytopenic purpura and the compensated thrombocytolytic state. American Journal OfMedicine, 51, I.
KARPATKIN, S. &KHAN, I. (1974) Intravascular thrombocytolysis in autoimmune thrombocytopenic purpura (ATP). (Abstract). Clinical Research, aa,
KBAnw, M.(1973) Platelet size in thrombocytopenias and thrombocytosis of various origin. Blood. 4 1 s 587.
MURPHY, S., O m ,F.A., NAUIAN, J.L., LUSCH,C.J., GOLDBERG, S. & G m m , F.H.(1972.) Platelet size and kinetics in hereditary and acquired thrombocytopenia. New England Journal .f Medicine, a86, 499. PROST-DVOJAKOVIC, R.J.,LE TOBIC,F., DESNOYERS. P., LABAUME, J. & SAMANA, M.(1974) Critical study of the distribution of platelet volumes in thrombocytopenic purpuras. XV Congress o j the International Society ofHematology,Jerusalem, p 197. SAHUD, M.A. (1972) Platelet size and number in alcoholic thrombocytopenia. New England Journal of Medicine, 186, 355. WEINTRAUB, A.H. & KARPATKIN,S. (1974) Heterogeneity ofrabbit platelets. 11. Use ofthe megathrombocyte to demonstrate a thrombopoietic stimulus. Joumal of Laboratory and Clinical Medicine, 83, 896. ZUCKHR-FRANKLIN, D. (1969) Microfibrils of blood platelets: their relationship to microtubules and contractile protein. Journal 9 f Clinical Investigation, 48,165. ZUCKBB-FRA", D. (1970) The ultrastructure of megakaryocytes and platelets. Regulation of Hematopoiesis (ed. by A. S. Gordon), p 1553. AppletonCentury-Crofts, New York.
