Development of gates to measure the immature platelet fraction in C57BL/6J mice using the Sysmex XN-V series multispecies hematology analyzer

Abstract

The immature platelet fraction (IPF) is a measure of newly released platelets, which has been used as a marker of platelet production in multiple human studies but is not widely available in multispecies analyzers. We developed gates to measure the IPF in diluted and undiluted murine blood samples on the Sysmex XN-1000V multispecies hematology analyzer. IPF gates were created using undiluted and diluted (1/10) blood samples obtained from adult and newborn (postnatal day 10, P10) C57BL/6J wild-type (WT) mice, and from 3 murine models of thrombocytopenia: c-MPL^−/− mice, which lack the thrombopoietin receptor (hyporegenerative); antibody-mediated thrombocytopenia; and acute inflammation-induced thrombocytopenia. P10 mice were chosen because, at their size, we could consistently obtain (by terminal phlebotomy) the blood volume needed to run an undiluted sample. The undiluted blood IPF gate successfully differentiated between mechanisms of thrombocytopenia in both adult and P10 mice. For diluted samples, 2 IPF gates were generated: a thrombocytopenic (T) gate, which performed well in samples with platelet counts (PCs) <800 × 10⁹/L in adult mice and <500 × 10⁹/L in newborn mice, and a non-thrombocytopenic (NT) gate, which performed well in samples with PCs above these thresholds. PCs and IPFs measured in diluted blood using these gates agreed well with those measured in undiluted blood and had good reproducibility. These diluted gates allow for the accurate measurement of PCs and IPFs in small (10 µL) blood volumes, which can be obtained easily from adult and newborn mice as small as P1 to assess platelet production serially.

Thrombocytopenia is a frequent clinical problem in humans of all ages as well as in domestic animals.^3,7,14,15,19 The differential diagnosis of thrombocytopenia is broad, and establishing the underlying etiology is critically important to determine the appropriate therapy. The underlying mechanisms of thrombocytopenia can be broadly categorized as decreased platelet survival (as a result of platelet destruction or increased consumption), decreased platelet production, or a mixture of both. Categorizing thrombocytopenia into one of these general groups narrows the differential diagnosis and can be extraordinarily helpful in reaching the correct diagnosis.

Until recently, assessing platelet production required obtaining a bone marrow sample to evaluate megakaryocyte number and morphology. In addition to being invasive, bone marrow studies are also technically difficult in small human and animal patients. In 1992, reticulated platelets (also known as shift platelets) were described; these are large immature platelets, newly released from the bone marrow, that have a high RNA content, which allows them to be identified and quantified by flow cytometry.^1,6,12,16,19 However, the clinical applicability of the immature platelet measurement was hindered by the lack of standardization and wide variability in research protocols.

Over the past decade, new methods in commercial hematology analyzers allowed for the automated quantification of the immature platelet fraction (IPF), a clinical equivalent to the immature platelet count (PC).¹³ IPF can be used as a marker of platelet production, in a manner similar to the use of the reticulocyte percentage as a marker of red cell production.⁶ Changes in the IPF may indicate physiologic as well as pathologic processes, and multiple human clinical studies have supported the value of the IPF in the evaluation of platelet disorders and as a predictor of PC recovery.^2,4,8,10 Advances in human hematology analyzers were accompanied by progress in multispecies analyzers, and gates were created to measure the IPF in dogs using the Sysmex XT-2000iV.¹²

Animal models are also frequently utilized in research to elucidate physiologic and pathologic processes. For the study of hematologic problems, the mouse remains the preferred in vivo model given its molecular and physiologic similarities with humans. Using both wild-type and mutant strains of mice, investigators are able to model disease states and investigate mechanisms of disease. However, the amount of blood able to be collected for serial sampling in mice is limited, especially when studying newborn mice. Previously, our group created a gate to measure the mouse IPF on the Sysmex XT-2000iV, which was based on the natural separation between mature (older) and immature (younger and larger) platelets visible in the platelet gate, similar to a flow cytometry plot.¹⁷

The new Sysmex XN-V series multispecies hematology analyzer (XN-1000V; Sysmex) was launched in 2017 for animal use only and introduced several advances, including a function that allows for the creation of new gates to evaluate desired parameters. Using this function, we developed a gate to measure the IPF in C57BL/6 mice in this new analyzer. To study very small blood samples (10 µL), which can be obtained from newborn mice as young as P1 (postnatal day 1) or serially from adult mice, we also developed slightly modified IPF gates to accurately measure this parameter in diluted blood samples and examined the agreement between measures obtained in the diluted and non-diluted samples. Finally, we tested the within-run precision of the PC, IPF, and immature PC (IPC; defined as IPF × PC) in diluted samples measured by the Sysmex XN-1000V.

Materials and methods

Study overview

Our study was conducted using a stepwise approach. First, we designed a gate to measure the IPF in undiluted blood (100 µL) from healthy adult C57BL/6J (wild type, WT) mice. Based on prior data, we set this gate to measure IPF values of 5–6%. Next, we used the same gate to measure the IPF in undiluted blood samples obtained from 3 adult models of thrombocytopenia and from P10 WT and thrombocytopenic pups (using the same models of thrombocytopenia). After establishing that this gate performed well in both newborn and adult WT and thrombocytopenic mice when undiluted blood samples were used, we tested its performance in 1/10 diluted blood samples from the same adult and P10 thrombocytopenic and non-thrombocytopenic groups. These studies demonstrated that the original gate overestimated the IPF consistently in diluted non-thrombocytopenic samples, and to a larger extent in thrombocytopenic samples. Thus, we created 2 new diluted IPF gates to improve the accuracy of the measurements in diluted blood (one for thrombocytopenic and one for non-thrombocytopenic samples) and established the PC thresholds below which the thrombocytopenic gate performed better in adult and newborn mice. Finally, we tested the accuracy and within-run precision of these new diluted gates in adult and P10 WT mice and in the 3 models of thrombocytopenia.

Animals

All animals were housed in a pathogen-free environment, and all experiments were approved by the Boston Children’s Hospital Animal Care and Use Committee. Adult (6–12 wk old) and newborn P10 WT C57BL/6J mice (Jackson Labs) and c-MPL^−/− mice were used to model different etiologies of thrombocytopenia. P10 mice were used because they were the smallest and/or youngest mice from which the amount of blood needed to run an undiluted sample plus 3 diluted samples could be obtained consistently by terminal phlebotomy.

Models of thrombocytopenia

c-MPL^−/− mice lack the thrombopoietin receptor, c-MPL, which results in reduced platelet production and a platelet count ~15% of that in WT C57BL/6J mice. These mice were expected to have a low-normal IPF and were used to model thrombocytopenia resulting from decreased platelet production. Immune thrombocytopenia was induced by intraperitoneal (IP) injection of WT mice with a purified rat monoclonal antibody directed against mouse GPIb-alpha, a platelet surface glycoprotein that is part of the receptor for von Willebrand factor (Emfret Analytics), at a dose of 4 μg/g in adults and 2 μg/g in P10 pups. The lower dose was given to P10 mice because they have a lower mean PC than adult mice (Table 1). At these doses, this antibody induces rapid platelet depletion followed by recovery over 5 d.¹¹ Blood samples were obtained 3 d after antibody injection, to measure the IPF during the recovery phase, when an elevated IPF would be expected. Acute inflammation-induced thrombocytopenia was generated in adult and newborn mice by lipopolysaccharide (LPS) injection (O111:B4; Millipore Sigma) at a dose of 10 µg/g body weight IP. We have extensively characterized the murine response to this LPS dose and have found that it induces an acute inflammatory response with rapid platelet and white blood cell (WBC, leukocyte) declines, and acute elevation of inflammatory cytokines, including IL-6. These mice were euthanized 12–24 h after LPS injection, at a time when they exhibited thrombocytopenia as a result of rapid platelet consumption without a significant increase in thrombopoiesis.

Table 1.

Platelet count (PC), immature platelet fraction (IPF), and immature platelet count (IPC) measured in adult and newborn mice with different etiologies of thrombocytopenia.

Age/Genotype or condition	Mice (n)	Platelet parameters			MPV (fL)
		PC (×109/L)	IPF	IPC (×109/L)
Adult
Wild-type	13	1,279 ± 215	0.064 ± 0.015	83 ± 33	6.6 ± 0.2
c-MPL−/−	10	145 ± 67	0.051 ± 0.014	7 ± 4	6.9 ± 0.3
Antibody	5	242 ± 266	0.341 ± 0.100	74 ± 64	8.4 ± 0.5
Inflammation	9	489 ± 173	0.076 ± 0.016	38 ± 17	6.7 ± 0.1
p*		<0.0001	<0.0001	<0.0001	<0.0001
Neonate
Wild-type	11	846 ± 192	0.074 ± 0.015	62 ± 19	7.2 ± 0.2
c-MPL−/−	14	83 ± 32	0.072 ± 0.019	6 ± 2	7.8 ± 0.7
Antibody	5	508 ± 187	0.263 ± 0.055	131 ± 48	8.8 ± 0.4
Inflammation	9	500 ± 117	0.091 ± 0.021	45 ± 11	7.4 ± 0.3
p		<0.0001	<0.0001	<0.0001	<0.0001

Values represent means ± SD.

MPV = mean platelet volume.

^*From one-way analysis of variance, testing for equal means across genotype or condition.

Experimental procedures

Blood samples were obtained from adult and newborn mice by terminal retro-orbital blood draw under isoflurane anesthesia. Blood was collected in EDTA microtainer tubes (Becton Dickinson) via heparin-coated capillary tubes. Undiluted blood samples (100 μL of anticoagulated undiluted blood) were run on the XN-V series multispecies hematology analyzer (XN-1000V; Sysmex) immediately following collection. For diluted samples, 10 μL of anticoagulated blood was diluted 1/10 in normal saline (0.9% NaCl) for a final volume of 100 μL. All diluted samples were run in triplicate immediately after dilution. Of note, each sample was visually assessed for the presence of clots and, if any were visualized, the sample was excluded. In addition, the Sysmex XN-1000V has a flag that informs the user if there is platelet clumping. All of our samples had no significant clumping.

Generation of IPF gates

The Sysmex XN series analyzer has a dedicated fluorescence channel for platelets (PLT-F), which extends the counting time 5-fold and thus increases the accuracy and precision of PCs, particularly when the PC is low. The PLT-F channel stains and counts platelets by optical fluorescence using a platelet-specific, patented fluorescent dye (Fluoro-cell PLT), which contains the RNA dye oxazine.⁵ In the Sysmex XN series analyzers, the IPF is calculated in this channel, compared with the previous Sysmex XE/XT series in which counting was performed as part of the reticulocyte analysis using a combination of polymethine and oxazine dyes. In the XE/XT series, as in flow cytometry, immature platelets were visible as a distinct population separated from the mature platelets.¹⁷ In the human and veterinary Sysmex XN analyzers, the platelets with the highest fluorescence and forward scatter (immature platelets) are separated from the mature platelets based on a preset gate, which is used to calculate the IPF. To accommodate different animal species and strains, the manufacturer’s software package allows user-defined gates to be inserted in specific plots. We capitalized on this feature to design a gate to identify the population of immature platelets in healthy adult and newborn C57BL/6J mice. Specifically, we first set a gate around the platelet population to exclude debris, erythrocytes, and reticulocytes. A second gate was then placed within the first gate to separate the majority of the platelet population (~95%, mature platelets) from the smaller percentage of highly fluorescent and larger platelets, representing the IPF (undiluted blood gate, Fig. 1A). In the absence of a clear separation between mature and immature platelets on the screen, the gate was set to separate 5–6% of platelets with the highest fluorescence from the rest of the platelets in adult C57BL/6J mice. This percentage was selected based on IPF values published previously by our group for adult C57BL/6J mice using the Sysmex XT-2000iV analyzer, wherein the immature platelets could be visually identified as a separate population.¹⁷

Figure 1.

Screenshots of the platelet and immature platelet fraction (IPF) gates in the Sysmex XN-1000V. First, a gate was set around the platelet cloud (PLT, white lines), and then a second gate was set to select a small subpopulation of larger and more fluorescent platelets, representing 5–6% of all platelets in adult wild-type (WT) mice (IPF, yellow lines). A. Undiluted blood gate in a WT adult mouse. B. Non-thrombocytopenic diluted blood gate in a WT adult mouse. C. Thrombocytopenic diluted blood gate in a c-MPL^−/− adult mouse, with a very low IPF. D. Thrombocytopenic diluted blood gate in a WT adult mouse with antibody-induced thrombocytopenia and a high IPF.

In diluted samples from healthy adult and newborn mice, applying the original gate overestimated the IPF values consistently (compared to the undiluted samples from the same animal), and thus the IPF gate for diluted samples was modified slightly to achieve the best agreement with the corresponding undiluted samples. However, diluted samples from thrombocytopenic mice run on the newly created diluted gate showed suboptimal agreement with the corresponding undiluted samples. To address this, 2 slightly different XN-1000V gates were developed to measure the IPF in diluted samples: one that performed best in non-thrombocytopenic samples (NT gate, Fig. 1B), and one that performed best in thrombocytopenic samples (T gate, Fig. 1C, and 1D), defined as a PC <800 × 10⁹/L in adult mice and <500 × 10⁹/L in newborn mice. These values were chosen because they were ~2 SD below the mean PC for adult and newborn WT mice (Table 1), and because studies directly comparing the 2 diluted gates identified those PCs as the thresholds below which the T gate consistently provided more accurate IPF values (i.e., in better agreement with the corresponding undiluted IPF).

Examination of other hematologic parameters

Commonly used hematologic parameters to assess platelets, red blood cells (RBCs, erythrocytes), and WBCs were measured in both undiluted and diluted samples, and the agreement between those samples was evaluated.

Statistical analysis

We used one-way analysis of variance to compare PCs, IPF, and IPC levels in whole blood among groups of adult and neonatal mice. For each hematologic parameter, we illustrated the agreement between the level determined in undiluted blood and the average of triplicate diluted samples from the same animal with plots of diluted versus undiluted blood values. To measure the bias for each parameter (degree of systematic disagreement), we calculated 100% × (mean difference between diluted and undiluted) ÷ mean undiluted. To measure the reproducibility of each parameter in diluted samples, we estimated the variance σ² among replicates within animal and its standard error (SE). For this analysis, we pooled data from 10 to 14 animals from each group (Table 2) to yield a set of 30–42 data points per group, which was subjected to variance component analysis using a random-effects regression model to identify the SD among replicates of a given sample. We log-transformed the data for analysis and expressed the reproducibility as a coefficient of variation, CV = 100% × (exp(σ) – 1) (equivalent to 100% × standard deviation ÷ mean of a normal distribution, the conventional definition of CV for untransformed data). This analysis was also conducted separately on a set of historical samples obtained with the previous XT-2000iV analyzer in 3 categories of animals: adult WT, adult c-MPL^−/−, and newborn WT. We then constructed a z-statistic comparing the variance in the 2 analyzers (XN-1000V and XT-2000iV) for each animal group (difference in σ² divided by SE of the difference), with a small p value indicating a significant difference in reproducibility between the analyzers. We used SAS v.9.4 (SAS Institute) for all computations.

Table 2.

Coefficient of variation for platelet parameters measured in triplicate diluted blood samples using the Sysmex XN-1000V. Blood for these measurements was obtained from adult and newborn wild-type (WT) and c-MPL^−/− mice.

Parameter	Adult WT	Adult c-MPL−/−	Neonate WT	Neonate c-MPL−/−
Platelet count	3.4	6.7	3.6	10.9
Immature platelet fraction	3.5	11.5	5.1	12.5
Immature platelet count	5.3	12.7	6.4	14.4
Mean platelet volume	1.2	5.1	2.0	11.7

Values are CV (%), estimated by random-effects analysis of variance of log-transformed measures.

Results

Undiluted blood IFP gate

The PC, IPF, and IPC were measured in undiluted blood samples from 37 adult mice: 13 healthy WT mice, 10 c-MPL^−/− mice, 5 with antibody-mediated thrombocytopenia, and 9 with inflammation-induced thrombocytopenia. c-MPL^−/− mice had the most severe thrombocytopenia and also the lowest IPC (with no IPF increase), consistent with a production defect (Table 1). Mice with antibody-induced thrombocytopenia also had significantly reduced PCs but a very high IPF (34.1 ± 10%, mean ± SD), indicating accelerated platelet production following antibody-induced platelet depletion (increased peripheral destruction). LPS-treated mice had less severe thrombocytopenia without a significant elevation in the IPF. P10 mice (n = 39) with the same etiologies of thrombocytopenia exhibited similar results (Table 1). Taken together, these results indicated that the undiluted IPF gate allows for a mechanistic evaluation of thrombocytopenia in adult and newborn mice.

Diluted blood IPF gates

PCs and IPF percentages obtained using the T or NT gate in diluted blood (as indicated based on the PC) agreed well with those measured in the corresponding undiluted blood (bias of −0.1 to −1.2% for PCs, 4.2–7.5% for IPFs; Fig. 2).

Figure 2.

Agreement between platelet counts (PCs; top) and immature platelet fractions (IPFs; bottom) measured in undiluted and diluted blood samples obtained from adult and neonatal mice with normal PCs and various degrees of thrombocytopenia from different etiologies. Samples with low PCs that fell within the dark gray box in both adults and neonates were analyzed using the thrombocytopenic diluted blood IPF gate. Samples with normal or near-normal PCs that fell within the light gray area were analyzed using the non-thrombocytopenic diluted blood IPF gate. Dots represent values from individual mice, and the lines represent perfect agreement. Disagreement is measured by bias = 100% × (mean difference between diluted and undiluted PC) ÷ mean undiluted PC.

Measurement of other CBC parameters in diluted samples

There was also good agreement between measurements obtained in diluted and undiluted samples of the same animal, consistent with prior studies (Suppl. Fig. 1).¹⁸

Reproducibility of platelet parameters measured in diluted samples

The within-run precision of the PC, IPF, IPC, and mean platelet volume measured in diluted samples, expressed as the CV among triplicates, was <7% for all 4 parameters in adult and newborn WT mice, and <15% in severely thrombocytopenic adult and newborn c-MPL^−/− mice (Table 2). Given that we had similar historical data obtained from adult WT and c-MPL^−/− mice and from newborn (P2) WT mice measured using the Sysmex XT-2000iV analyzer, we then compared the reproducibility of diluted measurements for these and other hematologic parameters in these populations between the 2 analyzers. The CV for most parameters was significantly lower when using the Sysmex XN-1000V compared to the XT-2000iV, indicating better within-run precision with the Sysmex XN-1000V (Suppl. Table 1).

Discussion

The IPF and IPC values obtained with the new gate on the XN-1000V analyzer closely matched our previously published values for WT and c-MPL^−/− adult and newborn mice measured using the XT-2000iV,^9,17 and provided different profiles depending on the mechanism of the thrombocytopenia. As in our prior study, the IPF was similar in WT and c-MPL^−/− P10 and adult mice, but the IPC was strikingly lower in the transgenic mice.⁹

Comparison of T and NT diluted gates and the mean IPF measured in 3 diluted samples compared to the corresponding undiluted samples confirmed both the accuracy and precision of results from diluted samples obtained from animals with normal as well as low PCs. Importantly, the IPF and IPC measured in diluted samples had better within-run precision in the XN-1000V analyzer than in the XT-2000iV, likely as a result of the newer technology that extended the platelet counting time 5 times and thus improved the precision of the counts, particularly when the PC is low.

Our study has some limitations. First, because of the inability to visually establish the separation between mature and immature platelets on the Sysmex XN-1000V, we used historical normative IPF values to determine the placement of the IPF gate in undiluted samples from adult WT mice. Second, it is unclear why an adjustment of the original undiluted gate is required to compensate for both sample dilution and the presence of thrombocytopenia in diluted samples. We hypothesize that the low cell count in diluted samples (and particularly in thrombocytopenic, diluted samples) results in overestimation of the IPF when run in the undiluted gate. Third, we also recognize that our model of immune thrombocytopenia, which was used to generate thrombocytopenic mice with a high IPF during platelet recovery, does not mimic the ongoing platelet destruction seen in naturally occurring immune thrombocytopenias in human or veterinary patients. Finally, we observed that, in mice with severe microcytosis, small RBCs could reach the platelet gate and artificially elevate the PC and/or the IPF. Thus, in samples with microcytosis, it is critical that investigators examine the actual plots to ensure that the PC and IPF are not affected by the presence of microcytic RBCs. As a safeguard for such circumstances, user-defined instrument flags may be enabled to indicate the need to review the results further. Despite these limitations, we believe that the application of these new IPF gates in the Sysmex XN-1000V will provide a valuable new tool to study platelet production in newborn and adult mice. A similar approach can be applied to develop IPF gates to study other animal species.