Introduction

Newly diagnosed immune thrombocytopenia (ITP) patients are usually administered one of the first-line treatments available. Few children will not respond to the first treatment, and only 70–85% will eventually have a durable complete response and cure at 1 year post diagnosis.1,2,3,4,5

None of the clinical factors that have been proposed to predict a cure at 1 year has a sufficiently strong predicting value. These include sudden appearance of the disease, viral disease preceding the presentation, lower platelet count, age <10 years, male sex, and lack of mucosal bleeding at diagnosis.6,7,8,9,10,11,12,13,14,15 Interestingly, intravenous immunoglobulin (IVIG) treatment with or without steroids seems to protect against the onset of chronic disease.7,8

A murine model showed increased platelet apoptosis in response to antibodies against integrin aIIbb3, while this effect was prevented by IVIG treatment.16 Since antibodies are often detected in ITP patients,17 we examined whether platelet apoptosis is involved in ITP pathogenesis. More precisely, we examined whether a difference exists in the pathophysiologic mechanisms, namely, the potential platelet apoptosis presentation at diagnosis between the non-chronic patients versus those who will eventually have a chronic disease.

Several studies have addressed the role of apoptosis in ITP patients. Platelet apoptosis may stem from autoantibodies directed against surface platelets’ proteins.16 In adult patients, platelet apoptosis was pronounced, as measured by increased phosphatidylserine (PS) expression, loss of mitochondrial inner membrane potential (ΔΨm), and higher caspase 3 activation.17,18,19 Winkler et al.19 demonstrated that increased platelet apoptosis at diagnosis in pediatric patients became normalized following IVIG treatment. Other studies have shown that the presence of CD8+ T cells may suppress megakaryocyte apoptosis, leading to impaired platelet formation.20,21 This is the first study to examine differences between non-chronic and chronic pediatric patients’ platelet apoptotic markers resulting from serum taken at the time of diagnosis.

Accurately predicting the course of ITP (namely, the ability to assess, at diagnosis, which patients are more susceptible to becoming chronic) will open the door to trying different first-line treatments up front for those children who are predicted to have a chronic disease. This may change the course of the disease of this subgroup, which is currently very challenging.

Materials and methods

Patients

Blood samples were collected from 66 patients with a clinical diagnosis of primary ITP admitted consecutively to five pediatric departments in Israel between 2010 and 2014. Some patients were admitted at times when the collection of samples was not possible and, therefore, were not recruited to the study. Only 42 samples were analyzed due to inadequate remaining blood samples. Reasons for exclusion were a blood sample in the wrong tube, an insufficient amount of blood, and patient’s arrival at a time when no one was available to accept the sample and separate the serum.

All blood samples were taken prior to initiation of treatment, and this cohort of patients was prospectively followed up. Parents gave informed consent for their child’s participation in the study, which was approved by the respective Ethics Committee of the five major medical centers: Sheba, Barzilai, Laniado, Sha’arei Tzedek, and the Schneider Children’s Hospital. ITP was defined by the report of the International Working Group (IWG) as a disorder characterized by isolated thrombocytopenia (peripheral blood platelet count <100 × 109/l) in the absence of other causes or disorders that may be associated with thrombocytopenia.5 Thirteen healthy participants aged 5–18 years served as a control group.

Clinical data

Clinical data were collected from medical records, and 12 months post diagnosis, each child was categorized into one of the two groups: non-chronic (resolved by 1 year, n = 26) and chronic (defined as persisting after 1 year, n = 16). The clinical presentation and laboratory findings of the two groups were compared and analyzed.

Blood sampling and processing of serum

Serum preparation from patients and controls

Blood samples were taken at diagnosis from ITP patients prior to therapy initiation, as well as from healthy controls, and were centrally processed in the coagulation laboratory at the Sheba Medical Center. All samples underwent heat inactivation at 56 °C for 30 min, destroying any enzymatic activity and leaving the residual immunoglobulin. The sera were used to study their effects on normal platelets ex vivo, as schematically described in Fig. 1 and detailed below. Sera were frozen at −80 °C until the experiments were performed.

Fig. 1: A flow chart of the procedures for measuring apoptosis markers induced by ITP patients’ sera on normal control platelets.
figure 1

The flow chart describes the process the samples underwent including ITP patientsʼ serum as well as controls.

Full size image

Measurement of apoptotic markers

Figure 1 depicts the process that the samples underwent.

Measurement of PS exposure and mitochondrial electrochemical potential using flow cytometry

Washed platelets taken from at least three healthy donors were incubated with serum samples (patients, healthy controls, or ionophore as a positive control) and stained with fluorescein isothiocyanate-conjugated Annexin V (Fischer Scientific, USA) for PS exposure and DIOC63 (Fischer Scientific) for ΔΨm separately as previously described17,18,22,23,24 and then analyzed on a FACScalibur (BD Biosciences—see Supplement for elaboration). The effect was measured as the relative number of platelets showing an increased apoptosis response compared to the controls (mean + 2 SD).

Human Apoptosis Array

We used RayBio® Human apoptosis antibody microarray slides (RayBiotech, Inc.) to measure the expression level of 43 proteins involved in the apoptosis process. First, we incubated sera (50 µl) from ITP patients or healthy controls, with washed platelets taken from healthy donors (350 µl in 3 duplicates from 3 different donors), for 1 h at room temperature. Second, the platelets were lysed (following dilution with 500 µl of platelet buffer [pH 6.5] and centrifuged at 800 rpm for 10 min) with 55 µl of lysis buffer for 30 min on ice. Third, the supernatant of the lysed platelets containing an identical protein concentration (10 ng protein determined by a BCA kit [PierceTM BCA Protein Assay Kit, Thermo Fisher Scientific])25 was incubated overnight in the RayBio-Human Apoptosis Array. The array was performed according to the manufacturer’s instructions: 2-h incubation with biotinylated antibodies, followed by a 2-h incubation with labeled streptavidin. The fluorescence detection of the microarray slides was performed using a laser fluorescence scanner by the RayBiotech company. The signal intensities obtained in each slide were calculated in relation to the positive control (biotinylated proteins) and compared to the negative control (bovine serum albumin as background level). The RayBio® Human Apoptosis Antibody Microarray Kit was validated in comparison to a western blot analysis using Jurkat cells treated with apoptosis inducer (RayBiotech, Inc.).

Statistical analysis

Statistical analyses were performed with the SPSS software (version 24), using the following tests: analysis of variance followed by Bonferroni post hoc comparisons or measurement of Pearson correlation coefficient between several parameters.

Results

Patient population and clinical presentation

The study included 42 children aged 6 months–18 years (average 4.1 years, median 3 years), all diagnosed with ITP according to the IWG guidelines.5 Twenty-one of the patients were female. One patient had a vaccination administered approximately 3 weeks prior to presentation, and another 17/42 had suffered from an upper respiratory tract infection 2–3 weeks prior to presentation. At diagnosis, 38/42 patients presented with petechias on the skin, 7/42 suffered from epistaxis, 6/42 had oral mucosal bleeding, and 1 patient suffered from upper gastrointestinal bleeding.

No difference in platelet count was detected between non-chronic (18,222 ± 13,412) and chronic (18,625 ± 20,858) ITP patients at the time of diagnosis.

Regarding treatment modalities, 26 patients had received first-line therapy: IVIG (23/42) and steroids (3/42). Sixteen out of 42 were observed without treatment. Six were treated additionally with second-line therapy: steroids (3/42), Anti-D (1/42) Mabtera (1/42), and splenectomy (1/42) (see Table 1). Demographic and clinical data between the non-chronic and chronic groups are also compared in Table 1: there were a statistically significant higher proportion of females in the chronic group, as reported above.

Table 1 Clinical data of the ITP patients at diagnosis.
Full size table

Apoptotic markers

Mitochondrial potential and Annexin V surface expression

The effect of sera taken from ITP patients on normal control platelets was measured as increased PS expression and loss of ΔΨm, using platelets taken from at least three different normal control individuals. The relative amount (above the normal = 0) was 0.72 ± 0.28 for decreased ΔΨm and 0.58 ± 0.34 for increased PS expression, suggesting increased platelet apoptotic markers following incubation of normal platelets with sera taken from ITP patients. On the other hand, no difference was found in the surface expression of ΔΨm or PS apoptotic markers between non-chronic and chronic ITP patients (Fig. 2). A significant and strong correlation between the two markers (r = 0.705, p < 0.001) was observed. Interestingly, the correlation was higher in platelets that were incubated with the sera taken from non-chronic ITP patients [r = 0.802, p < 0.001] than with the sera taken from chronic patients [r = 0.544, p = 0.05].

Fig. 2: The phosphatidylserine surface expression (PS) and the mitochondrial inner membrane potential (ΔΨm) in control platelets after incubation with sera taken from non-chronic ITP patients (white) or chronic ITP patients (gray).
figure 2

The figure shows a mean ± SEM above the effects of sera taken from healthy controls.

Full size image

Although no significant correlation was found between platelet counts and the apoptotic markers, a weak link between low platelet count and loss of ΔΨm in response to chronic ITP patients’ sera was detected; however, it did not reach a statistical significance (r = −0.46, p = 0.112).

Analysis of 43 apoptotic proteins’ expression in normal platelets after incubation with sera taken from ITP patients

We compared each of the 43 apoptotic proteins’ expression to the negative control expression (background level 56.9 ± 35.3) and found 5 proteins whose average expression level was less than mean + 2 SD of the negative control (cut off = 103.5): BID (80.3), caspase 8 (61.9), CIAP2 (72.4), IGFBP4 (96.6), and Survivin (93.1). These proteins were excluded from further analysis due to extremely low expression levels in platelets.

Expression of apoptotic proteins in platelets after incubation with sera of ITP patients or normal controls

The expression level of the remaining 38 apoptotic proteins in normal platelets, which were incubated with sera taken from ITP patients, was compared with their expression after incubation with sera taken from healthy individuals, using analysis of variance followed by Bonferroni post hoc comparisons. Only one protein (sTNFR2) showed an increased level of expression after incubation with sera taken from ITP patients (129.8 ± 73.9) compared to incubation with sera taken from healthy individuals (60.4 ± 32.4), which was similar to the background (56.9 ± 35.2). This suggests that the sera taken from ITP patients triggered a sTNFR2 expression in the platelets (p = 0.002), whereas the basic expression level of sTNFR2 in platelets was negligible.

Expression of apoptotic proteins in normal platelets after incubation with sera of non-chronic versus chronic pediatric ITP patients

Five out of 38 proteins were detected with a significantly increased level of expression after incubation with sera taken from non-chronic ITP patients compared to that of chronic ITP patients (Table 2); BIM, CD40, IGFBP2, P21, and SMAC. No protein was found to have a decreased expression in platelets after incubation with sera taken from non-chronic ITP patients compared to that of chronic ITP patients. Interestingly, in comparison to platelets incubated with sera taken from healthy normal subjects, the expression level of all six apoptotic proteins was higher following incubation with non-chronic ITP patient, while three proteins (IGFBP2, P21 and SMAC) exhibited a decreased expression in platelets after incubation with sera taken from chronic pediatric ITP patients (Fig. 3).

Table 2 The average expression (units) level in platelets after incubation with sera taken from ITP patients and healthy controls.
Full size table
Fig. 3: The experssion level of  proteins associated with apoptosis in platelets.
figure 3

The expression level in platelets after incubation with sera taken from ITP patients, compared to sera taken from healthy controls, shows a mean ± SEM effect of sera from non-chronic ITP patients (white) and chronic ITP patients (gray).

Full size image

A comparison among the six proteins whose expression was affected by sera taken from ITP patients revealed significant positive correlations among five of them as shown in Table 3: only IGFBP2’s level of expression inversely correlated with CD40 and SMAC level of expressions. No significant correlation was found among these six proteins and the platelet counts at the time of diagnosis.

Table 3 Pearson correlation coefficient between the expression levels in platelets after incubation with sera taken from ITP patients and healthy controls.
Full size table

Discussion

In this study, we found an increased expression of five apoptotic proteins on control platelets incubated with sera of non-chronic pediatric ITP patients, compared to the sera of their chronic ITP counterparts, at the time of diagnosis. Although pediatric ITP patients’ sera induced apoptosis in normal platelets compared to sera from healthy donors, no significant difference was found in the apoptotic markers’ PS surface expression and loss of ΔΨm, between the groups.

The proteins with increased expression in non-chronic versus chronic ITP patients are BIM, CD40, IGFBP2, P21 and SMAC. The expression of one protein in particular, sTNFR2, was elevated in platelets that were incubated with sera from non-chronic or chronic ITP patients compared to healthy subjects. The elevated level of sTNFR2 in platelets after incubation with sera taken from ITP patients is in line with its high level in autoimmune diseases.

BIM and SMAC are both pro-apoptotic proteins that act via the mitochondrial intrinsic pathway.26,27 Not surprisingly, we found a significant correlation (r = 0.684) between their expression level in platelets following incubation with the sera. Moreover, a significant correlation (r = 0.421, p = 0.04) was found between the BIM expression level and ΔΨm in platelets following incubation with non-chronic ITP sera, suggesting that non-chronic ITP sera trigger platelet apoptosis via the mitochondrial intrinsic pathway.

An inverse correlation was found between the anti-apoptotic protein IGFBP2 and the pro-apoptotic SMAC protein, or the pro-activation CD40 protein, in accordance with their different effects on platelets. No significant correlations were found between IGFBP2 expression and the other proteins: BIM, p21, and IGF1sR (Table 3).

Interestingly, p21, which is connected to cell cycle nucleus proteins,28 is highly expressed in platelets, although they have no nucleus and do not divide (Table 2). Sera taken from chronic ITP patients decreased its expression in the platelets (Fig. 3) and displayed a positive correlation with pro-apoptotic SMAC and the pro-activation CD40,29 which is enigmatic for its function in the platelets.

Two pathways of apoptosis in platelets have been described, despite their lacking a nucleus. The first, a cell-intrinsic mitochondria-dependent pathway,22 is the basis for the results presented above. Platelets, which contain functional mitochondria and alpha granules, are capable of protein synthesis and RNA splicing. In response to stimulation, they are able to adhere and spread on an extracellular matrix, expose P-selectin and PS, and undergo apoptosis.16

The second, an extrinsic pathway, may be initiated by the interaction between death ligands belonging to the TNF superfamily and the cell surface TNF receptors. This explains the higher level of the TNF receptor super-family proteins CD40 and sTNFR2 in platelets following incubation with sera taken from non-chronic ITP patients compared to normal sera, as seen in the pro-apoptotic mitochondrial proteins (BIM and SMAC).

Notably, although increased platelet expression following incubation with sera taken from non-chronic ITP versus chronic ITP patients was found in six proteins, only four—the two pro-apoptotic mitochondrial proteins (BIM and SMAC) and the two TNF receptor super-family proteins (CD40 and sTNFR2)—were higher, compared to normal sera. The levels of the two other proteins, p21 and IGFBP2, were equal in platelets exposed to normal sera and chronic ITP patients’ sera.

The main process leading to thrombocytopenia might differ in non-chronic versus chronic pediatric ITP. Antigenic mimicry with antibodies leading to apoptosis is believed to be the leading cause in non-chronic disease. While decreased maturation of megakaryocytes may be the main initiator in chronic pediatric ITP, it appeared as a decreased number of apoptotic megakaryocytes.30 This may explain why non-chronic patients’ sera induced higher apoptotic markers compared to chronic ones.

This study has some limitations; first, the study tested the effects of ITP sera on normal platelets and not on the ITP patients’ platelets. Second, we examined the sera at one time point, upon initial diagnosis; ideally, this should be done at several different time points and correlated with the response to treatment. Third, the apoptotic protein panel checks the quantitative levels of the markers; however, this does not tell us in what configuration (e.g., phosphorylation) they are, whether or not they are active, or precisely where in the cell (mitochondria or cytosol) they are found. This latter fact may explain why some of the markers that are anti-apoptotic increased in response to non-chronic patients’ sera, perhaps being in a non-active configuration.

Altogether, our results suggest that the pathophysiology of non-chronic versus chronic ITP is somewhat different, although they share a nearly identical clinical presentation. Future studies are needed to expand our knowledge of these markers.

This data is one more step in the attempt to categorize ITP patients at diagnosis into non-chronic and chronic ones. This, in turn, may enable us to tailor a more specific and exact treatment plan to these seemingly different clinical entities, currently being treated in the same way, as has been suggested by other researchers.31

Drugs that target the BCL2 pathway’s proteins (BIM being one of them) are used in cancer therapy and have the effect of increasing apoptosis.28 Thrombocytopenia is the predominant dose-limiting toxicity. We speculate that, if we were able to block this pathway by targeted therapy, we would decrease apoptosis and increase the platelet count. This could be an alternative second-line therapy for patients.

Heitinik-Polle et al.7 and others1,2,3,4,5,6,7,8,9,10,32 suggest certain criteria as predictors of chronic ITP in children at the time of diagnosis. These include sex, age, infection or vaccination, insidious onset, platelet count, and antinuclear antibody titer. Our data can add some knowledge of apoptosis markers that may aid in prognosis prediction in combination with previous predictors. Since our assays are indirect and require normal platelets for the evaluation, the development of an apoptosis test using ITP patients’ platelets is preferred. Future studies are needed to evaluate additional options, such as devising a kit that can measure apoptotic markers in a simple blood test, or flow cytometric assays that are feasible in samples with a low platelet count.33,34,35