Clinical relevance of antiplatelet antibodies and the hepatic clearance of platelets in patients with immune thrombocytopenia

To the editor:

Immune thrombocytopenia (ITP) is an autoimmune disorder whose primary pathogenesis is mediated by Fc receptor (FcR) clearance of antibody-opsonized platelets by spleen macrophages.¹  Recently, however, an Fc-independent mechanism of platelet clearance mediated by antibodies recognizing the platelet membrane glycoprotein Ib molecule (anti-GPIb antibodies) and leading to hepatic clearance of platelets was proposed by Li et al.²  More specifically, they found that both murine and human anti-GPIb antibodies induced platelet GPIb desialylation in vitro; murine in vivo studies suggested that anti–GPlb-opsonized desialylated platelets lacking sialic acid become the ligand for the Ashwell-Morell receptor (AMR) on hepatocytes³  and are removed by endocytosis. According to the hypothesis, this mechanism may predict for failure to respond to splenectomy, a recognized highly effective ITP treatment.⁴  To date, however, this hypothesis has not been tested in ITP patients.

To address this, we retrospectively analyzed 93 adult primary ITP patients who were screened for antiplatelet antibody specificities and underwent platelet survival studies (PSSs) to estimate the site of clearance. This study was approved by the hospital review board. Moreover, patients signed an informed consent to clinical data use at enrollment in a local ITP database.

Our results suggest that the specificities of the antiplatelet autoantibodies do not predict or mediate a skewed hepatic clearance pattern.

Charts of adult ITP patients who had data on both PSSs and autoantibody testing were reviewed by 2 independent reviewers (S.C. and M.C.).

Antiplatelet antibodies are routinely searched for during testing at diagnosis of ITP (or, for patients diagnosed elsewhere, at first referral to our center); PSSs are performed in patients who fail or have a suboptimal response to first-line steroid therapy and are candidates for splenectomy.

Measurements of platelet-bound (direct) and plasma (indirect) immunoglobulin G (IgG) antiplatelet activity were performed using a commercially available solid-phase enzyme-linked immunosorbent assay (PakAutoAssay, Immucor GTI Diagnostic). Blood samples were added to microwells coated with monoclonal-captured glycoprotein IIb-IIIa, Ib-IX, Ia-IIa, allowing antibodies, if present, to bind; an alkaline phosphatase–labeled secondary reagent (anti-IgG/IgM/IgA) was added and incubated for 30 minutes, and optical density was measured at 405 nm.

Patient platelets were labeled with ¹¹¹In-oxide by methods previously described,⁵  and reinfused platelets were used for the PSSs. To determine platelet survival time, platelet-bound radioactivity was assessed from 10-mL venous blood samples taken at 30 minutes, 2 hours, and 4 hours postinfusion and daily thereafter. Sampling was maintained up to 168 hours postinjection or until platelet-bound radioactivity was <10% of the initial value. A multiple-hit method was used to calculate the mean platelet survival according to the International Committee for Standardization in Hematology.⁶  Platelet clearance pattern was determined by images from a γ camera. Scintigraphic indices obtained were spleen/heart, liver/heart, spleen/liver, spleen/spleen at 30 minutes, and liver/liver at 30 minutes according to standard procedures.⁷  Platelet clearance pattern was defined as splenic uptake (if splenic uptake measured on the third day was >1.2 times the splenic uptake measured at 30 minutes from platelet reinfusion), hepatic uptake (if hepatic uptake measured on the third day was >1.2 times the splenic uptake measured at 30 minutes from platelet reinfusion), or mixed uptake (if both splenic and hepatic uptakes measured on the third day were >1.2 times the splenic uptake measured at 30 minutes from platelet reinfusion).⁸ 

Standard descriptive statistical analysis techniques were used for all the collected variables using Stata/SE 14.1 (StataCorp, College Station, TX). The Fisher’s exact test was used to assess the association between categorical variables.

Our study population comprised 93 ITP patients with a median age at diagnosis of 48 years (range, 19-86 years). Of these, 56 out of 93 (60.2%) were splenectomized; in the remaining 37 patients (39.8%), splenectomy had not been performed yet or was subsequently deemed not necessary because of a late response to medical therapy.

Table 1 summarizes antibody testing and PSSs results. The time interval between antibody testing and PSSs was ≤12 months in 71% of patients, with only 17.2% of patients tested at ≥24 months.

Among our patients, 53 out of 93 (57%) tested positive for antiplatelet antibodies by the direct test: 19 out of 53 (20.4%) tested positive for only one auto-antibody, whereas 8 out of 53 (8.6%) tested positive for 2 autoantibodies and 26 out of 53 (28%) tested positive for 3 autoantibodies.

Among the 40 out of 93 (43%) autoantibody-negative patients, the spleen was the sole site of platelet clearance in 24 out of 40 (60%), the liver was solely involved in 4 out of 40 (10%), and 12 out of 40 (30%) had mixed clearance patterns. Among the 53 out of 93 autoantibody-positive patients, splenic clearance occurred in 31 out of 53 (58.5%), hepatic clearance was observed in 7 out of 53 (13.2%), and mixed clearance was seen in 15 out of 53 (28.3%).

Upon analysis, the site of platelet clearance was not influenced by either antibody positivity or negativity (P = .910) or number of different antibody specificities (eg, 1 vs 2 or 3; P = .802). Moreover, positivity for anti–GPIb-lX antibody (32 out of 53 patients; isolated, 4 out of 32 patients; combined with other specificities, 28 out of 32 patients) was not associated with site of platelet clearance (P = .270).

Table 2 summarizes outcomes of the 56 out of 93 splenectomized patients; results were matched with PSSs and autoantibody results. Response to splenectomy was classified according to International Working Party definitions,⁹  and overall, our results were consistent with published data⁴  (complete response [CR], 47 out of 56 patients [84%]; response [R], 4 out of 56 patients [7.1%]; and no response [NR], 5 out of 56 patients [8.9%]). No difference in remission rate was observed between the 52 patients with splenic (94.3%) and mixed (94.1%) clearance (P > .999), and this was confirmed when analysis was restricted to triple-positive (12 out of 52) patients (P > .999); CR/R was less frequent in cases of hepatic clearance compared with both splenic and mixed clearance patterns (P = .040). However, no association was found between outcome and antibody negativity or positivity (P = .358), number of antibody specificities (P = .245), or GPIb-IX positivity, whether present alone or combined with other specificities (P = .324)

Recent murine studies suggested that antiplatelet antibody specificity may determine the platelet’s fate by activating either FcR- or non–FcR-mediated mechanisms, resulting in splenic or hepatic uptake, respectively.¹ 

However, in contrast to the experimental results in mice, our in vivo data in ITP patients suggest that antiplatelet antibody specificity, including anti-GPlb, is not associated with either a specific clearance pattern or failure to respond to splenectomy.

In patients with multiple antibody specificities (including GPIb), mixed or hepatic PSS patterns may result from platelet uptake by both hepatocytes (GPIb-mediated platelet desialylation) and macrophages (FcR-mediated, independent of antibody specificity), which are present in the reticuloendothelial system of both spleen and liver. In our case series, 1 patient with isolated positivity to anti–GPIIb-IIIa and 5 out of 26 triple-positive patients had limited hepatic clearance, probably resulting from hepatic macrophage FcR-mediated platelet clearance, similarly to splenic clearance mechanisms. Conversely, 13 out of 26 triple-positive patients had limited splenic uptake. Moreover, of the 4 patients with isolated GPIb positivity, 2 had splenic uptake and 1 had mixed uptake, and the liver was solely involved in only 1 patient.

AMR-mediated hepatic platelet clearance in vivo may represent a physiological mechanism involved in platelet homeostasis. Platelets desialylate as they circulate, thereby becoming the primary ligand for the AMR,¹⁰  and this interaction regulates hepatocyte thrombopoietin production.¹¹  Desialylation also occurs when platelets are activated by several physiological stimuli, and AMR clearance may be relevant in attenuating the coagulopathy of sepsis.^12-14  Our results support the indications of international ITP guidelines,¹⁵  which suggest that both PSSs and glycoprotein-specific antibody testing are not mandatory in ITP workup or management. However, if available on a single-center basis, these tests may help to gain insight into the prevalent mechanism underlying thrombocytopenia (increased clearance vs deficient production) in a specific patient who fails first-line therapy with glucocorticoids.