Pathological Mechanisms Of Heparindependent Antibodies

In our experience, anti-PF4-H antibodies of the IgG isotype are present in at least 85% of patients with clinical HIT. In the remaining cases, IgA, IgM, or both isotypes—but only when present at high concentrations—could be involved. This is based on studies of HIT in which the anti-PF4-H antibodies were fully isotyped (Amiral et al., 1996c). Although the clinical picture and positive platelet aggregation tests supported the diagnosis of immune HIT, only IgM and/or IgA isotypes of anti-PF4-H antibodies were found, and no IgG was detected. This finding is nevertheless controversial, as recent studies tend to demonstrate the preeminent role of the IgG isotype in development of HIT (Lindhoff-Last et al., 2001; Warkentin, 2004, 2005; Warkentin et al., 2005; Greinacher, 2006). However, HIT cases associated with high concentrations of IgM and/or IgA isotypes (Amiral et al., 1996c; Meyer et al., 2006) could be underdiagnosed, depending on the study inclusion criteria. In any event, these intriguing observations require explanation for how IgM and IgA antibodies could trigger thrombocytopenia, with or without thrombosis.

It is well accepted that IgG antibodies to PF4-H can become pathogenic when they interact with platelets, particularly when PF4-H-IgG complexes bind to the platelet Fc receptors (FcyRIIa) (Kelton et al., 1988) (see Chapter 8). Another group proposed that, in addition, an IgG receptor polymorphism on leukocyte FcyRIIIa, different from that of FcyRIIa, could also be involved (Gruel et al., 2004). Our observations indicate that other mechanisms for PF4-H-antibody complexes binding onto blood cells could be involved. These could result not only if the PF4-H complexes bind to the cell surfaces through their heparin-binding sites (Van Rijn et al., 1987; Horne and Alkins, 1996; Horne and Hutchison, 1998) but possibly also through PF4-binding sites (Capitanio et al., 1985; Rybak et al., 1989). Although HIT antibodies recognize PF4-H complexes in the fluid phase (Newman et al., 1998), it is uncertain whether this typically occurs in vivo before interaction of PF4-H-IgG complexes with the platelet surface, or whether HIT antibodies only bind after PF4-H complexes first become attached to the platelet surface. Recent reports have shown that anti-PF4-H antibodies from patients with HIT can activate ECs (especially microvascular ECs) and also monocytes, and thereby induce release of TF (Pouplard et al., 2001; Arepally and Mayer, 2001; Blank et al., 2002).

Regardless, the clinical state of the patient, determining the extent of platelet and EC activation, seems to be a key factor for determining whether clinical HIT results (Boshkov et al., 1993; Reininger et al., 1996). This contribution occurs in several ways: activated platelets release high amounts of PF4 that can complex with heparin, and activated platelets also expose a higher density of heparin-binding sites (Horne and Chao, 1989). Consequently, these platelets may be even more readily activated by heparin-dependent antibodies. This situation occurs in patients with acute or chronic platelet activation associated with CPB, atherosclerosis, inflammation, infections, cancer, diabetes, and orthopedic surgery, among others.

Another factor determining HIT antibody formation is the type of heparin that binds to PF4, which depends on its oligosaccharide composition, polysacchar-ide length, and grade of sulfation (Lindahl et al., 1994; Greinacher et al., 1995). Formation of PF4-H complexes requires a heparin molecule containing at least 12-14 oligosaccharide units and a high sulfation grade (more than three sulfate groups per disaccharide) (Amiral et al., 1995). Furthermore, binding of heparin to blood and ECs also increases with heparin molecule length and sulfation grade (Sobel and Adelman, 1988; Horne and Chao, 1990; Harenberg et al., 1994). Heparin structure thus has a dual effect in HIT: it is required not only to form PF4-H complexes but also to target these complexes onto cell surfaces. These factors could explain the higher frequency of PF4-H antibody development and of HIT in patients receiving UFH, compared with those receiving LMWH (Poncz, 2005; Greinacher, 2006). With UFH, PF4-H complexes are larger and are more easily formed, requiring a lower heparin concentration than with LMWH. For the latter drug, only the subset of molecules containing at least 12-14 oligosaccharide units (MW > 3600 Da) can generate immunoreactive PF4-H complexes. Thus, because LMWH has a lower propensity to form PF4-H complexes and binds less readily to platelets and ECs, LMWH therapy is also less likely to result in thrombocytopenia even when pathologic HIT antibodies are already present.

PF4-H-reactive antibodies targeted at platelets induce platelet activation, resulting in thrombocytopenia and, often, thrombosis. Occasionally, heparin-induced thrombosis occurs in the absence of thrombocytopenia (Hach-Wunderle et al., 1994; Bux-Gewehr et al., 1996). Platelet activation by the IgG isotype antibodies is mediated by interaction with the platelet FcyRIIa receptors (Kelton et al., 1988; Denomme et al., 1997). Some studies suggest an important role for an FcyRIIa polymorphism (Brandt et al., 1995; Burgess et al., 1995). However, the role of the FcyRIIa receptor polymorphism is controversial (Arepally et al., 1997; Denomme et al., 1997; Suh et al., 1997; Bachelot-Loza et al., 1998) (see Chapter 8).

Platelet activation might also occur through other mechanisms such as direct antibody binding to exposed cell antigens (Rubinstein et al., 1995), a phenomenon that is dependent on the antigen electric charge (Schattner et al., 1993). Heparin is highly electronegative. Evidence for direct activation through antigen binding is supported by the positive platelet aggregation response produced by some patient plasma samples containing only anti-PF4-H antibodies of the IgM and/or IgA isotypes. Formation of heparin-containing immune complexes on cell surfaces can initiate blood and EC interactions, and this can enhance their activating effects. Cell-cell interactions may occur and be amplified through release products that chemoattract and activate cells or through transcellular metabolism (Nash, 1994; Marcus et al., 1995). Platelet products (e.g., PF4) and platelet-derived microparti-cles (Warkentin et al., 1994) can induce activation of leukocytes (Aziz et al., 1995; Jy et al., 1995; Petersen et al., 1996). Leukocyte-release products, such as cathepsin G, can directly activate platelets and cleave b-thromboglobulin to the active chemokine, NAP-2, thus establishing an amplification loop. Platelet-leukocyte aggregates can form in vivo, contributing to vascular occlusion, especially in limb vessels (Fig. 3). In a recent study, antibodies to PF4-H from patients with HIT were shown to induce synthesis of TF by monocytes in the presence of PF4 and heparin (Pouplard et al., 2001) or by microvascular ECs (Blank et al., 2002). This could be a complementary pathway for inducing thrombosis.

Variability in certain biologic characteristics of anti-PF4-H antibodies influences their potential for inducing HIT. Platelet activation caused by anti-PF4-H antibodies is usually weak and is only pathogenic when amplification mechanisms are involved. This is demonstrated by the variable lag phase observed in platelet aggregation studies performed with different plasmas or sera from HIT patients. Antibody concentration is another important factor for determining the extent of platelet activation. Antibody affinity is also crucial: the higher the affinity, the lower the concentration of antibodies required for activating platelets. Recently, a subset of antibodies to PF4-H complexes that had platelet-activating properties was isolated in three patients with HIT (Amiral et al., 2000). These platelet-activating antibodies had the highest avidity for PF4-H. In contrast, the bulk of antibodies against PF4-H in these patients had no effect on platelet activation. Also, when IgM or IgA isotypes are present, affinity for PF4-H complexes is usually lower than that of IgG isotypes and, consequently, high concentrations are necessary for pathogenicity. Lastly, HIT antibodies do not all bind to the same epitopes on PF4-H complexes, and this specificity could be an additional important factor (Horsewood et al., 1996; Pouplard et al., 1997; Suh et al., 1998). At least two neoepitopes have been identified on PF4 that are distinct from the "region of positive charge" to which heparin binds (Ziporen et al., 1998; Li et al., 2002) (see also Chapter 6). Thus, anti-PF4-H antibodies are not equivalent, and those with the strongest affinity are most pathogenic.

Platelet activation in HIT involves amplification through ADP receptors (Polgâr et al., 1998) and involves GPIIb/IIIa (Hérault et al., 1997; Jeske et al., 1997). These findings further emphasize the importance of platelet activation amplification loops for producing the clinical manifestations of HIT.

0 0

Post a comment