MoniqueRobinson Haemostasis Research Unit, Haematology Department, University College of London, 98 Chenies Mews, London WC1E 6HX, UK. E-mail: monique.robinson@ ucl.ac.uk
Animal in vivo biotinylation studies have demonstrated that thiazole orange (TO) labels the youngest cells in the circulation. TO has since been widely used for the measurement of reticulated platelets. As recent findings suggest that at high concentrations TO also labels platelet dense granules non-specifically, the value of previous work is unclear. Mepacrine also labels platelet dense granules and can detect storage pool defects. In this study, a mouse in vivo biotinylation model was used to determine the specificity of TO and mepacrine staining on platelets recently released into the circulation. The mean life span of biotin/TO (low), biotin/TO (high) and mepacrine/TO dual-positive platelets was 1.4 d (SD 0.5), 2.2 d (SD 0.2) and 2.3 d (SD 0.3) respectively (n = 6) compared with a life span for biotin-positive platelets of 4.9 d (SD 1.6). TO (low), TO (high) and mepacrine labelled 8.0% (SD 3.1), 43.9% (SD 8.3) and 40.0% (SD 9.9) of the total platelet population respectively (results of 30 samples from six mice), which decreased to 6.8% (SD 3.9), 26.6% (SD 6.9) and 25.7% (SD 10.6) after thrombin degranulation. The shorter life span and lack of thrombin sensitivity of TO (low)-positive platelets, suggests that TO (low) measures reticulated platelets specifically. The comparative life spans and thrombin sensitivity of TO (high) and mepacrine-positive platelets suggest that TO (high) labels platelet dense granules. These data also suggest that dense granules are lost during platelet ageing.
Platelets are anucleate cell fragments that are derived from bone marrow megakaryocytes. They retain mRNA derived from their parent cells, but are probably unable to synthesize their own. Ingram & Coopersmith (1969) first described the appearance of platelets containing RNA-like material, which stained positive with methylene blue by supravital staining, after acute blood loss in dogs. Similarly to red cell reticulocytes, these methylene blue-positive platelets were representative of the youngest platelets in the circulation and were termed reticulated platelets. Kienast & Schmitz (1990) have since described a method that identified reticulated platelets by flow cytometry using thiazole orange (TO). TO is a dye specific for nucleic acids that exhibits a several thousand-fold increase in fluorescence emission upon binding to RNA or DNA ( Lee et al, 1986 ) and has been widely used for the measurement of reticulated platelets. The estimation of the percentage of reticulated platelets using TO has been established as a measure of bone marrow platelet production and is sensitive enough to discriminate between consumptive and aplastic causes of thrombocytopenia (for a review, see Harrison et al, 1997 ). The utility of mRNA measurement in patients receiving growth factor therapy has been demonstrated previously ( O'Malley et al, 1996 ) in which a rise in reticulated platelet levels was detected after 3 d of infusion of human subjects with megakaryocyte growth and development factor, a recombinant form of thrombopoietin.
Biotin has previously been used as a means of measuring platelet life span in various animal models ( Ault & Knowles, 1995). After intravenous injection, biotin covalently labels free amino groups on the surface of cells in the circulation. Newly released platelets are thus, by definition, biotin negative, whereas mature platelets remain biotin positive until the end of their life span. In this manner, the entire platelet population will gradually become biotin negative as the platelet population ages. Such in vivo biotinylation studies within animal models have provided the definitive proof that TO-positive platelets represent the youngest platelets in the circulation ( Ault & Knowles, 1995; Dale et al, 1995 ), and, in the case of mice, identifiable as reticulated platelets for about 1.5 d, with a total life span of 4.5 d ( Ault & Knowles, 1995). Despite this, it has been well established that a proportion of the TO signal is RNase insensitive ( Ault et al, 1992 ). Robinson et al (1998) compared TO labelling of thrombin receptor activating peptide (TRAP)-degranulated normal and granule-deficient platelets from patients with storage pool disease and Hermansky Pudlak syndrome and found decreased TO fluorescence in these samples. TO was thus found to label not only platelet RNA but also platelet dense granules non-specifically. It is thus possible that the level of TO labelling observed in various clinical settings by some authors could also be related to platelet granularity rather than to their RNA content.
Mepacrine uptake into platelet dense granules has been successfully utilized to diagnose patients with various types of storage pool disease (for a review, see Corash et al, 1986 ). Mepacrine is a fluorescent antimalarial acridine derivative with a reported high affinity for adenine nucleotides. Da Prada & Pletscher (1975) found that mepacrine accumulated rapidly and selectively into platelet dense granules, so that the number of fluorescent granules per platelet could be measured by microscopy. Flow cytometry was found to improve the sensitivity of the detection of mepacrine-labelled platelets ( Gordon et al, 1995 ; Wall et al, 1995 ).
Because the TO fluorescence of platelets has been shown to increase with the length of incubation and concentration of TO, the aim of this study was to determine the specificity of TO (at low and high concentrations, as used by other investigators) for newly released platelets, as determined by in vivo biotinylation of mice using the method of Ault & Knowles (1995). As a control for platelet dense granule labelling, the impact of mepacrine staining on newly released platelets was also examined. Thrombin degranulation of samples in conjunction with TO and mepacrine staining facilitated the identification of granular (non-specific) staining. The data from this study will provide valuable insight into the conditions required for TO to stain young platelets specifically for the accurate measurement of human reticulated platelet levels.
MATERIALS AND METHODS
N-Hydroxysuccinimido-biotin (NHS-biotin), dimethyl sulphoxide (DMSO), Gly–Pro–Arg–Pro, mouse thrombin and mepacrine were from Sigma-Aldich Chemical Co (St. Louis, MO, USA); streptavidin-RED 670 was from Gibco BRL (Gaithersburg, MD, USA); thiazole orange (Retic-Count solution) was from Becton Dickinson Immunocytometry Systems (San Jose, CA, USA); FITC hamster IgG isotype standard, FITC anti-mouse CD61, FITC mouse IgG1 isotype standard and FITC anti-mouse CD62P were from PharMingen (San Diego, CA, USA); 0.9% sodium chloride was from Baxter Healthcare (Thetford, Norfolk, UK); iloprost was from Schering Healthcare (West Sussex, UK); and Isoton® II was from Beckman-Coulter (Luton, UK).
In vivo biotinylation
CD-1 [Charles-River (UK), Margate, UK] mice were injected intravenously with biotin according to the method of Ault & Knowles (1995). Briefly, 6 mg of NHS-biotin was dissolved in 150 μl of DMSO and diluted 1:10 with 0.9% sodium chloride immediately before injection to give a final concentration of 4 mg/ml. This solution (150 μl) was then injected i.v. into the tail vein. The injection was repeated 1 h later. Blood samples from biotinylated mice were analysed by flow cytometry over a period of 7 d.
Dual TO/mepacrine and streptavidin labelling
Mice were placed in a 37°C chamber for 10 min before blood collection to facilitate location of the tail vein and venepuncture. A 21-gauge needle was inserted into the vein and a drop of blood allowed to form, which was collected. A total of 50 μl of blood was added to 10 μl of 0.105 m citrate and 5 μl of 25 m m Gly–Pro–Arg–Pro. Anticoagulated whole blood (5 μl) was incubated with: (i) TO (low) (1 ml of Retic-Count diluted 1:10 in Isoton® II) for exactly 30 min; (ii) TO (high) (1 ml of undiluted Retic-Count) for exactly 30 min; or (iii) 1 ml of 25 μm mepacrine for exactly 15 min at room temperature (20–22°C) in the dark. Control tubes consisted of 5 μl of blood incubated with 1 ml of Isoton® II alone. Iloprost (100 ng/ml) was added to control and to test tubes (to prevent the platelets from clumping) before centrifugation at 800 g for 5 min. The supernatants were discarded and the pellets gently resuspended. The washed blood cells were then incubated for 15 min with a predetermined saturating concentration of streptavidin-RED 670 (2 μl of stock at 250 μg/ml), and were then resuspended in 0.5 ml of Isoton® for analysis. Platelets were identified by flow cytometry (Coulter XL-MCL) according to their characteristic size (log forward scatter) and granularity (log side scatter). An electronic gate was drawn around the platelet cloud within a Listmode gate to exclude red and white cells and 10 000 platelet events were collected. Verification of the platelet gate was performed by incubating 5 μl of blood with a saturating concentration of anti-CD61-FITC, or its relevant isotype control, for 20 min; 0.5 ml of Isoton® II was then added for flow cytometric analysis.
Anticoagulated whole blood (5 μl) was incubated with 0.5 U/ml mouse thrombin in the presence of 2.5 m m (final concentration) Gly–Pro–Arg–Pro and HEPES-buffered saline (HBS) with Ca2+ (0.145 m NaCl, 5 m m KCl, 1 m m MgSO4, 10 m m HEPES, 2.5 m m Ca2+) ( Chronos et al, 1994 ) for 15 min at room temperature. Dual TO/biotin or mepacrine/biotin staining was then performed as outlined above. Verification of platelet degranulation was performed by incubating degranulated whole blood with a saturating concentration of anti-CD62p, or its relevant isotype control, for 20 min; 0.5 ml Isoton® II was then added before flow cytometric analysis.
Data are expressed as the means plus one standard deviation. Significance was analysed using a paired Student's t-test.
Platelet degranulation was monitored with CD62p expression. The concentration of thrombin that achieved maximum degranulation was established to be 0.5 U/ml. Unstimulated blood was also monitored to check that samples were not activated during sample preparation. Thrombin degranulation of platelets resulted in a decrease in forward scatter signal, but did not significantly affect CD61 labelling. Thrombin degranulation did, however, alter biotin intensity, so that dual TO/biotin or mepacrine/biotin analysis was not possible. The results of thrombin degranulation on TO (low), TO (high) and mepacrine are shown in Table I, and can be visualized in Fig 5.
NHS-biotin has been used previously to measure the life span of circulating platelets ( Ault & Knowles, 1995; Dale et al, 1995 ). Several investigators have demonstrated that in vivo biotinylation of mouse platelets and erythrocytes does not affect either the total platelet count or platelet life span. Ault & Knowles (1995) also reported that there was no loss of platelet function in human blood when it was exposed in vitro to similar levels of biotin. The extrapolated data for life span obtained in this study, using NHS-biotin as an in vivo label (4.9 d), were similar to those estimated by Ault & Knowles (1995) (4.5 d). In this study, we observed similar signal–noise ratios as Dale et al (1995) and Ault & Knowles (1995), with the majority of platelets becoming biotin positive after the two injections. Over the course of a week, the entire platelet population became progressively biotin negative. Dual TO (low)/biotin labelling clearly shows that it is the TO-positive platelets that become biotin negative first, confirming the observation made by Ault & Knowles (1995) and Dale et al (1995) that TO labels the youngest or reticulated platelets within the circulation. Dual staining of biotin with a higher concentration of TO, however, resulted in non-specific staining of platelets. Although some of the TO-positive platelets were biotin negative (similarly to platelets stained with a low concentration of TO), the remainder of the TO-positive platelets did not become biotin negative within 48 h. These platelets, despite staining positive for TO, were therefore not newly released platelets. Whereas TO (low) thus labelled young platelets specifically (presumably detecting their RNA), TO (high) appeared to label mature platelets non-specifically. This observation was consistent with the fact that thrombin degranulation significantly reduced the levels of TO staining observed with TO (high) but not TO (low) ( Fig 5).
If the estimates of platelet life span (4.9 d) and reticulated platelet life span (1.4 d) as measured using TO (low) are correct, then we would expect ≈ 30% of the platelets to be reticulated platelets as opposed to the 8.0% seen in this study. There are two possible reasons for this difference. First, it is possible that the youngest platelets are sequestered to the spleen where they mature before being released into the circulation. Indeed, it has been reported that the spleen sequesters the younger and larger platelets ( Freedman & Karpatkin, 1975; Shulman & Watkins, 1999). Second, as reticulated platelets are larger ( Karpatkin, 1969a, 1978b Amorosi et al, 1971 ) and haemostatically more reactive ( Karpatkin, 1969b, 1978a), they might be removed from the circulation preferentially to mature platelets.
Mepacrine stains platelet dense granules rapidly and selectively, thus providing a good control for platelet granular labelling. The fact that mepacrine and TO (high) showed some specificity for the youngest platelets with a life span of 2.3 and 2.2 d as opposed to 4.9 d could indicate either that the youngest platelets tend to be larger and contain more dense granules than mature platelets or that old platelets are inefficient at mepacrine uptake and/or that mepacrine is non-specifically labelling platelet mRNA in addition to platelet dense granules.
Circulating platelets are heterogeneous in terms of size, with a log normal distribution. It remains unclear, however, exactly how this heterogeneity arises. There are three main possibilities: first, that platelet volume heterogeneity is created by changes in platelet volume as the platelets age in the circulation; second, that platelet heterogeneity is created primarily during thrombopoiesis; or third, a combination of these. The resulting circulating population of platelets from either model would thus be heterogeneous in terms of size. Our results indicate that mepacrine shows some specificity for young platelets, which would support the first school of thought. The fact that thrombin degranulation results in a significant yet not complete loss of signal, however, would also suggest that it is possible that mepacrine labelled other platelet components, such as RNA ( Wall et al, 1995 ).
In conclusion, low concentrations of TO were found to label specifically young platelets within the circulation, with a life span of 1.4 d. Mepacrine-positive platelets also exhibited a shorter life span (2.3 d) than that of the entire circulating population (4.9 d). Higher concentrations of TO were also found to label not only young platelets but also more mature platelets non-specifically, resulting in a longer life span of 2.2 d. The similar life spans observed with mepacrine and TO (high) suggest that high concentrations of TO label platelet dense granules. There are three possible explanations for the shortened life span observed with mepacrine-positive platelets. First, it is possible that young platelets have a larger volume with increased numbers of dense granules which decrease as the platelet ages; second, that old platelets are inefficient at taking up mepacrine into the platelet dense granules; and third, that mepacrine is non-specifically staining platelet mRNA. Indeed, preliminary in vitro RNase and degranulation experiments in human blood have suggested that mepacrine also labels mRNA (unpublished data). It is highly possible, therefore, that TO and mepacrine could stain platelet RNA at low concentrations and platelet dense granules at higher concentrations.