Antibody Polyreactivity: A Challenger of Immune Paradigms

ABSTRACT

Polyreactivity refers to the ability of antibodies to bind multiple unrelated antigens, encompassing polyspecificity and promiscuity. Here, we discuss the diverse molecular mechanisms underlying polyreactivity, including conformational dynamics, sequence‐ and structural characteristics of antigen‐binding sites. The importance of polyreactive antibodies in immune defence and immune homeostasis is highlighted as well as their potential pathological consequences. Polyreactivity is seen as a continuum rather than a discrete property so that ultimately all antibodies possess some degree of polyreactivity. The challenges in defining antibody specificity are examined, and a shift towards quantitative thinking in antibody research is suggested. This would foster the adoption of novel methodologies to study complex antibody–antigen interactions at a systems level. Finally, the deeper understanding of polyreactivity's potential implications for the current antibody paradigm is critically evaluated.

This review article discusses the definition and functions of polyreactive antibodies, as well as their physiological roles. It also presents the concept of polyreactivity as a challenger of immunological dogmas.

1. Introduction

The conceptual framework of adaptive immunity is based on the clonal‐selection theory, which was first proposed by Frank Macfarlane Burnet in 1957 [1, 2]. This theory postulates that any antigenic epitope is recognised by a corresponding specific clone of B‐ or T‐cell receptor. Notwithstanding, pioneering studies conducted over five decades ago on monoclonal immunoglobulins derived from human or mouse myeloma (subsequently from hybridomas) demonstrated that a considerable proportion of antibodies do not adhere to the principles of the clonal selection theory. Instead, they have the capacity to recognise many unrelated antigenic determinants with analogous binding affinities [3]. Due to their antigen‐binding behaviour, these monoclonal antibodies have been referred to by a number of alternative terms, including multispecific, polyspecific, polyreactive and non‐specific. In the early stages of research, the generation of antibodies with low specificity was regarded as an unwelcome consequence of the immune response. Subsequent investigations, however, revealed the existence of important biological functions performed by antibodies with broad antigen‐binding capabilities. These antibodies were found to be indispensable components of adaptive immune responses, capable of conferring benefits such as directly contributing to the immune defence against pathogens or exerting immune homeostatic activities [4, 5]. Conversely, in specific contexts, these antibodies were also reported to have detrimental functions, including the ability to mediate tissue damage in certain pathological conditions.

A number of review articles have outlined the key characteristics, mechanisms and biological functions of antibodies with broad antigen‐binding potential [5, 6, 7, 8, 9, 10, 11, 12, 13, 14]. This review builds on this existing research by analysing recent findings on the molecular mechanisms of broad antigen binding in antibodies and presenting new insights into their biological functions. First, we define antibody polyreactivity and distinguish between the different types of antigen‐binding reactivity. The next section outlines the functions of polyreactive antibodies in immune defence and homeostasis. The molecular mechanisms leading to polyreactivity are also described. Finally, the paper explores the philosophical aspects of antibody specificity and polyreactivity, demonstrating how the broad reactivity of antibodies challenges established immunological paradigms and identifies opportunities to reconcile conflicting views.

2. Definition

Antibodies (or B‐cell receptors) that manifest binding potential to a broad repertoire of distinct antigens have been referred to by several terms, including polyreactive, multispecific, polyspecific, promiscuous and non‐specific. These terms have frequently been used interchangeably, even though they have different semantic connotations.

To more accurately define the tendency of certain antibodies to bind a broad range of antigens, it is first necessary to examine the meaning of the term ‘specificity’ [15]. The term ‘antibody specificity’ is used to describe the capacity of an immunoglobulin molecule to discriminate in its binding between different antigens. As Van Regenmortel has observed, to illustrate the concept of specificity, it is necessary to have at least three entities: two antigens and an antibody [16]. An essential quantitative parameter associated with specificity is binding affinity, which quantifies the stability of the antigen–antibody complex. If an antibody exhibits a markedly higher affinity for a given antigen in comparison to other antigens, it is considered specific for that antigen [16]. Given the potential for variation in the number of entities employed for probing antibody specificity, the substance of the term ‘antigen specificity’ is inherently relative and susceptible to bias within the context of the experimental setting. Accordingly, a specific antibody can be defined as one that can recognise a given antigen of interest among a panel of unrelated antigens. The categorisation of an antibody as ‘specific’ can be influenced by the breadth of the panel used to assess its antigen‐binding characteristics or the affinity thresholds employed. Consequently, if a limited number of antigens is employed, the antibody may exhibit a high degree of binding affinity solely to the antigen of interest, resulting in its classification as highly specific or ‘monospecific’. Conversely, when a highly diverse panel of potential antigens is utilised, the same antibody may demonstrate comparable binding affinity to multiple structurally unrelated antigens, leading to its designation as ‘polyspecific’. It is important to note that the polyspecific antibody will also possess a high discriminating potential, as it will bind only to a subset of antigens among a large diversity of potential options. The substantial molecular surface area of antigen‐binding sites and their abundance of solvent‐exposed amino acid residues, which are typically located within the interior of other proteins, render immunoglobulins inherently highly interactive molecules [17, 18]. It is therefore unlikely that there exists an antibody molecule with exquisite specificity in virtue of its intrinsic nature. Rather than dichotomising monospecific vs. polyspecific antibodies, it is more accurate to view the recognition of diverse antigens as a continuum of affinities and specificities.

2.1. Polyspecificity vs. Cross‐Reactivity

In contrast to polyspecific antibodies, which bind to structurally unrelated antigen determinants (epitopes), cross‐reactive antibodies represent a distinct category of antibodies with the capacity for broad antigen binding. These antibodies are able to bind to similar epitopes that are displayed by different antigens due to molecular mimicry [11, 16, 19]. To exemplify, a cross‐reactive antibody is one that recognises a similar sequence motif in the polypeptide chain, which is present in distinct proteins. Certain antibodies are likewise adept at recognising common chemical adducts that can be displayed on macromolecules as a consequence of post‐translational modifications or oxidative damage [20]. The binding to the chemical moiety displayed on various proteins can endow such antibodies with broad cross‐reactivity to distinct proteins. However, in fact, these antibodies are monospecific in terms of recognition of the same epitope (chemical adduct).

2.2. Polyspecificity vs. Binding Promiscuity

A further category of antibody reactivity can be distinguished within the broader spectrum of antibody reactivities. It has been observed that certain antibodies exhibit binding activity to the majority of the targets present in the panels utilised for the assessment of their antigen‐binding activity. Additionally, these antibodies have been shown to interact with polymer surfaces and chromatographic matrices [12, 13]. Typically, the interactions of these antibodies are characterised by low affinity. In an attempt to propose a nomenclature for these antibodies, they have been recently designated as ‘polyreactive’, ‘sticky’ and ‘non‐specific’ [12, 13]. Therefore, in contrast to polyspecific antibodies, which can bind to unrelated antigens while still exhibiting discriminative capacity, non‐specific antibodies utilise more general and less discriminatory molecular features (see below) to bind to the majority of macromolecules present in the screening panels [13].

The term ‘reactivity’ is used to describe the capacity of an entity to enter into a reaction (or to bind to an antigen in the case of an antibody). It is a more general concept than that of ‘specificity’. Consequently, the polyreactivity of antibodies more accurately represents the general inclination towards interaction with a multitude of antigens (Figure 1 and Table 1). The term is applicable to subcategories of antibodies with broad reactivity, including polyspecific or indiscriminatory binding capacity. In other words, all polyspecific antibodies are polyreactive, but not all polyreactive antibodies are polyspecific (Figure 1). Antibodies with fuzzy indiscriminate antigen binding may be more accurately described as promiscuous, a term used in protein science to denote degenerate binding or catalytic activities of proteins [21, 22]. The two categories of polyreactive antibodies are opposed to cross‐reactive antibodies, which are monoreactive by virtue of recognising identical or similar epitopes displayed by distinct proteins.

FIGURE 1.

Polyreactivity and cross‐reactivity of antibodies. The coloured oval shapes indicate antigens. The geometric forms (triangle, rectangle and star) indicate epitopes.

TABLE 1.

Definitions of terms about the antigen‐binding reactivity of antibodies.

Term	Synonyms	Definition	Notes
Polyspecific antibody	Multispecific antibody	Interaction of an antibody molecule with well‐defined structurally unrelated epitopes.	The binding affinity to unrelated epitopes can be substantial and indistinguishable from that of monospecific antibodies.
Promiscuous antibody	Non‐specific antibody, sticky antibody and degenerated antibody	Interaction of an antibody molecule with fuzzy epitopes. A promiscuous antibody molecule can bind in alternative ways to a single antigen.	The binding affinity of promiscuous antibodies is usually low.
Polyreactive antibody	Multireactive antibody	High‐order term depicting the ability of antibody molecule to bind to many unrelated antigens. It integrates the binding behaviour of polyspecific and promiscuous antibodies.	In certain literature sources, the polyreactivity is opposed to the polyspecificity of antibodies.
Cross‐reactive antibody	Heterophile antibody	An antibody molecule that recognises similar (molecular mimicry) or identical epitopes displayed by unrelated antigens.	Cross‐reactive antibodies are by virtue monospecific antibodies.

Throughout this article, the term ‘polyreactivity’ is used to refer to antibodies that demonstrate binding to multiple antigens. Where necessary, the specific subtype of polyreactivity involved will be specified. Table 1 summarises the terminology used for the description of antigen‐binding activity of antibodies.

2.3. Determination of Polyreactivity of Antibodies

Different methods have been employed to determine the antigen‐binding polyreactivity of antibodies. Classical immune assays, such as ELISA, immunoblot and indirect immunofluorescence, have been widely used [6, 23, 24, 25, 26, 27, 28, 29]. These assays have limitations. They are semi‐quantitative or they use only a limited number of antigens. These assays are also difficult to standardise and may vary between laboratories. The limited number and the nature of the antigens used in ELISA may give rise to some inconsistencies in the estimation of the polyreactivity of antibodies. Another drawback is that they are susceptible to avidity effects because antigens are typically fixed to solid surfaces. More recently, the polyreactivity of monoclonal antibodies has been assessed using protein arrays. These arrays display over 9000 human proteins and hence provide sufficient antigen breadth to comprehensively assess the polyreactive behaviour of antibodies. The protein microarray technology has been used to quantify the polyreactivity of broadly neutralising antibodies against the influenza virus and HIV‐1 [30, 31, 32]. A shortcoming of the method is that it is costly, which limits its use for screening a large number of monoclonal antibodies.

Recently, a new method of assessing polyreactivity was introduced to provide a more quantitative and standardised measure of the degree of polyreactivity of monoclonal antibodies. This method is referred to as the polyspecificity reagent assay and relies on the display of antibodies on yeast [33] or their immobilisation on beads [34], followed by the detection of labelled cellular proteins binding in solution using flow cytometry. The advantage of this assay is that it enables the screening of large antibody libraries and can provide quantitative data that has been proven to reliably identify polyreactive antibodies [35, 36, 37]. The polyspecificity reagent assay is now frequently used as an essential method for estimating the polyreactivity of therapeutic antibodies.

In addition to analysing polyreactive monoclonal antibodies, methods exist for assessing the polyreactivity of B‐cell receptors and determining the level of polyreactive antibodies in polyclonal immunoglobulin samples. To assess BCR polyreactivity, flow cytometry analyses using a fluorochrome‐labelled panel of antigens were employed [4]. Reactivity to aromatic nitrophenols (e.g., dinitrophenol) has been proposed and proven to be a reliable indicator of the level of polyreactive antibodies in polyclonal systems [38, 39, 40, 41].

3. Functions of Polyreactive Antibodies

Generally, the physiological functions of polyreactive antibodies can be divided into two categories: functions related to immune defence and functions related to immune regulation and maintenance of homeostasis.

3.1. Role of Polyreactive Antibodies in Immune Defence

3.1.1. Diversification of Immune Repertoires

At the level of immune repertoires, the potential of some B‐cell receptors to bind, albeit with low affinity, to multiple distinct antigens can tremendously amplify the immune diversity. Although the sequence diversity of antigen‐binding sites generated as a result of genetic processes of recombination and mutation during the ontogeny of B cells is vast, at any given moment, the repertoire of antigen‐binding specificities is limited and far insufficient to cover all potential epitopes in practically infinite antigenic space [42, 43]. Negative selection of autoreactive B cells in the bone marrow also introduces holes in the repertoire. Thus, the polyreactivity of a fraction of B cells can fill the gaps by amplifying the breadth of antigen binding at the repertoire level at a tolerable affinity (Figure 2) [9, 44, 45]. This can be especially true for receptors with indiscriminate promiscuous antigen binding, where binding hundreds to thousands of unrelated antigens is feasible, as can be deduced by protein array experiments with certain monoclonal polyreactive antibodies [46].

FIGURE 2.

Polyreactivity broadens the functional diversity of B‐cell repertoires. Upper panel: B‐cell repertoire that consists of cells expressing only highly specific immunoglobulin receptors will have specificity gaps due to the elimination of many B‐cell clones during negative selection. Lower panel: B cells expressing polyreactive immunoglobulin receptors can compensate for these gaps in the repertoire.

In their works, James and Tawfik argued that the primordial promiscuous activities of proteins are the starting material for evolutionary optimisation and the building of proficient activities [21]. Thus, promiscuous B‐cell receptors have higher evolvability, and upon binding of antigen with low affinity, they can be refined and ultimately give rise to a variant with improved affinity and discriminatory potential for a given antigen [9].

3.1.2. Innate‐Like Immune Function

Polyreactive IgM antibodies, which frequently constitute what are known as natural antibodies, constitute a constant component of the immune repertoires of all individuals. They are generated in the absence of specific immune stimulation and exhibit considerably greater stability than repertoires of adaptive antibodies [10, 47]. IgM antibodies frequently perform innate‐like functions within immune repertoires due to their polyreactivity (Figure 3). In particular, they interact with repetitive epitopes displayed on the surface of bacteria or viruses [47, 48]. It has been demonstrated that they are able to opsonise pathogens and trigger the activation of the complement system or phagocytosis by macrophages [48] (Figure 3). Furthermore, it has been demonstrated that natural polyreactive IgM antibodies can increase the immunogenicity of blood‐borne pathogens by up to 1000‐fold by trapping and diverting them to the secondary lymphoid organs [49, 50] (Figure 3). In vivo experiments have demonstrated that polyreactive IgM antibodies are unable to combat bacterial or viral infections on their own. However, they provide a window of opportunity by delaying the infection, allowing the development of a specific adaptive immune response to occur [51, 52]. Accordingly, Baumgarth et al. have shown that IgM generated by B1− (typical natural antibodies) and B2− (primary immune response) have complementary roles in protecting against the influenza virus [53]. A recent study in mice demonstrated that an IgG repertoire devoid of pathogen‐specific antibodies can confer protection against viral infection in a non‐specific manner, akin to the natural IgM repertoire [54]. This protection can be attributed to the presence of polyreactive IgG, which provides an innate‐like defence against the pathogen (Figure 3).

FIGURE 3.

Roles of polyreactive antibodies in immune defence. (A) Innate‐like function of polyreactive antibodies. Polyreactive IgG and IgM can opsonise pathogens and thus direct them for phagocytosis by macrophages (1). The opsonisation can also result in activation of the complement system (2), further enhancing phagocytosis (3) or causing a direct cytotoxic effect, as for example the lysis of bacteria (4). Alternatively, pathogens opsonised by both polyreactive antibodies and complement can be trapped in secondary lymphoid organs and trigger a specific immune response (5). (B) Polyreactive antibodies can also contribute to adaptive immune defence. These antibodies can better tolerate variability in the epitopes of highly evolvable viruses, such as HIV‐1 and influenza, thus broadening the neutralisation capacity. (C) Antigen‐binding polyreactivity can contribute to the enhancement of the avidity of highly specific antibodies, as described for the neutralisation of HIV‐1. Created in https://BioRender.com.

3.1.3. Function in Adaptive Immune Defence

Additionally, polyreactive antibodies have been demonstrated to play a role in adaptive immune responses. For instance, effective neutralising antibody responses to pathogens with high genetic variability, such as HIV‐1 and the influenza virus, are frequently accompanied by a high proportion of pathogen‐specific polyreactive antibodies with broad virus neutralisation activities [30, 55, 56, 57, 58, 59, 60]. It is noteworthy that these antibodies display an intriguing dichotomous tendency to bind the envelope proteins of viruses with high affinity and specificity, yet simultaneously manifest indiscriminate binding to a vast number of unrelated antigens. The available experimental evidence suggests that the broadly neutralising antibodies do not exhibit cross‐reactivity, but rather the binding promiscuity [29, 30, 31]. This can explain the very scarce generation of these antibodies in immune responses, as B cells expressing them may be subjected more often to negative selection during the early phases of development. This was demonstrated using transgenic mouse models, in which B cells expressing the variable regions of polyreactive, broadly neutralising HIV‐1 antibodies 2F5 and 4E10 were found to undergo negative selection at an early stage of development [61, 62]. These studies may help explain the insurmountable difficulty in the development of an HIV‐1 vaccine that induces a broadly protective immune response against diverse viral strains.

It is worthy of note that it has been demonstrated that there is a direct correlation between the polyreactivity of broadly neutralising antibodies and both their neutralisation breadth and potency [29, 63, 64]. The specific role of polyreactivity in the neutralisation of viruses with high genetic variability remains to be fully elucidated. However, it may be related to the enhancement of avidity by a process referred to as heteroligation [65]. This process may be particularly important for the neutralisation of HIV‐1, where viral envelope proteins are sparsely distributed on the surface of the virus, making bivalency binding through the Fab portions of an antibody practically impossible. The heteroligation model proposes that polyreactivity increases the avidity of broadly neutralising antibodies by enabling the antibody molecule to cross‐link the envelope protein with another, unrelated target on the viral surface by engaging two Fab arms [65]. Another possible consequence of polyreactivity in the neutralisation of highly variable viruses may be an increased tolerance to variability in the epitope. Thus, polyreactive antibodies may tolerate slight variations in the sequence of vulnerability sites on envelope proteins better than highly specific antibodies due to the promiscuity of their antigen‐binding sites [66] (Figure 3). It is worth noting that the antigen‐binding sites of broadly neutralising influenza and HIV‐1 antibodies are characterised by increased flexibility, which may explain their greater adaptability in recognition of epitopes with sequence variability [66].

The clinical implications of the polyreactivity of the broadly neutralising antibodies are unclear. Successful pre‐clinical studies and clinical trials of passive prophylaxis or therapy for HIV‐1 infection have been conducted using broadly neutralising antibodies that do not exhibit polyreactivity, as polyreactive antibodies demonstrate poor pharmacokinetics and are generally excluded from clinical applications [67, 68, 69, 70, 71, 72, 73]. Further investigation is needed to determine whether polyreactive, broadly neutralising antibodies could be used for specific therapeutic purposes. For example, they could be useful in the treatment of rapidly progressing infections (such as influenza), where the shorter half‐life of these antibodies may not be a concern.

3.2. Role of Polyreactive Antibodies in Homeostasis

3.2.1. The Establishment and Maintenance of a Commensal Equilibrium With the Microbiome

Investigations in humans and mice have demonstrated that B cells populating the mucosal surfaces of the intestines often secrete antibodies manifesting polyreactivity [74, 75, 76]. It is therefore evident that while the prevalence of polyreactivity among peripheral IgG‐positive memory B cells in mice is reported to be approximately 20%, a considerably higher percentage, namely 50%–70% of mucosa‐associated IgA+ B cells, secrete antibodies with broad antigen‐binding reactivity [76]. Additionally, studies of B‐cell populations in neonates have demonstrated an increase in polyreactivity and a tendency to bind to diverse commensal bacteria [77, 78]. In light of these findings, it was hypothesised that polyreactive antibodies contribute to the establishment of a mutualistic relationship with the microbiome. Although the mechanisms of this process are not yet fully understood, it is thought that direct binding of polyreactive IgA to diverse bacterial species is of importance. An elegant study by Fransen et al. also demonstrates that intestinal polyreactive antibodies binding to bacteria with low affinity may contribute to the development of highly specific IgA antibodies for given bacteria by transferring the bacteria through the epithelial layer [79]. Once the microbiome is established, polyreactive antibodies can contribute to the maintenance of the mutualistic relationship with the microbiota by surveillance through continuous recognition of bacteria (Figure 4).

FIGURE 4.

Homeostatic functions of polyreactive antibodies. Created in https://BioRender.com.

3.2.2. Clearance of Damaged Cells and Macromolecules

One of the most significant homeostatic functions of polyreactive antibodies is their involvement in the clearance of damaged cells, cellular debris, or modified macromolecules [47] (Figure 4). The repertoire of polyreactive antibodies in circulation can be continuously filtered out for reactivities to molecules present on the cellular surface or in plasma under homeostatic physiological conditions. However, alterations in homeostasis and the emergence of novel molecular patterns, such as those resulting from changes in membrane composition during apoptosis or necrosis of cells, oxidation of plasma or cellular macromolecules, or the release of substantial quantities of intracellular antigens (as observed during intravascular haemolysis), will engage the repertoire of polyreactive antibodies that are continuously present in circulation [47, 80, 81, 82, 83]. The homeostatic function is primarily executed by polyreactive IgM due to a significant enhancement in avidity. Consequently, through opsonisation of damaged cells or through binding to modified macromolecules (such as lipoproteins), the polyreactive antibodies, in conjunction with complement, facilitate their removal from circulation [47, 81].

This function of polyreactive natural IgM serves to reduce inflammation [82, 84]. Furthermore, the binding of natural polyreactive IgM to oxidised epitopes on plasma lipoproteins or apoptotic cells may provide protection from the development of atherosclerosis [85, 86, 87, 88, 89].

Notably, a recent study has demonstrated that due to their affinity for binding to necrotic cells, natural polyreactive IgM can assist in the clearance of damaged tissue and even accelerate the regeneration processes in a model of liver injury [90] (Figure 4). In another recent report, it was demonstrated that natural IgM can bind to cellular microvesicles and prevent their pro‐coagulation potential, thus inhibiting pulmonary thrombosis in a mouse model [91].

The capacity of natural polyreactive IgM to ‘sense’ altered self was also shown to contribute to their role in cancer surveillance [92, 93]. Interestingly, a recent study demonstrated that many antibodies targeting the tumour‐specific antigen MET, which were isolated from B cells infiltrating breast and lung cancers, also exhibited binding to unrelated antigens [94]. It has been hypothesised that these polyreactive antibodies may play a role in the anti‐tumour immune response in humans.

3.3. Pathogenic Potential of Polyreactive Antibodies

It is a notable possibility that the sentinel activity of polyreactive antibodies in healthy humans may potentially contribute to the immunopathology of certain diseases by turning against normal tissues. Thus, polyreactive antibodies are inherently capable of recognising minor alterations in the membrane surface. To exemplify, polyreactive IgM antibodies have been observed to deposit on cellular surfaces following ischemia–reperfusion [95, 96, 97]. It was demonstrated that in this condition, as a consequence of cellular stress, non‐muscular myosin is displayed on the cellular surface [96]. This neo‐antigen can be bound by polyreactive IgM antibodies. This results in the activation of the classical complement system, which in turn leads to a further exacerbation of tissue damage. A recent study demonstrated that patients with atypical haemolytic syndrome can produce polyreactive IgM antibodies that can induce the activation of the classical complement system pathway on the surface of a model cell line [98]. Therefore, it can be concluded that pathological activation of the complement system in this condition may not be mediated solely by perturbations in the alternative complement pathway, as was previously thought, but also by the classical pathway of complement, which is triggered by antibodies.

The frequency of B cells expressing polyreactive antigen receptors or soluble polyreactive antibodies has been reported to increase considerably in various autoimmune diseases, such as systemic lupus erythematosus (SLE) [99, 100, 101, 102], rheumatoid arthritis [103, 104], multiple sclerosis [105, 106], type 1 diabetes [107] and Sjögren's syndrome [108]. This increase is believed to result from defective central and peripheral B‐cell tolerance checkpoint mechanisms [109]. Despite evidence of an increased frequency of B cells expressing polyreactive antibodies in autoimmune diseases, direct evidence of the pathogenic potential of these antibodies is scarce. In this regard, one study demonstrated that a human monoclonal polyreactive antibody, isolated from a patient with SLE, exhibits neurotoxic activity in mice [110].

It has been demonstrated that polyreactive antibodies have pathogenic potential in organ transplantation [111, 112, 113]. Thus, polyreactive antibodies in patients with kidney grafts have been suggested to mediate humoral rejection by recruiting the complement system [112]. These antibodies have an increased ability to bind to apoptotic cells, as well as to multiple proteins present in tissue lysates. Infiltrates of B cells expressing elevated levels of polyreactive and apoptotic cell‐reactive antibodies were also found in patients with cardiac allograft vasculopathy [114]. These studies suggest that assessing the level of polyreactive antibodies in patients may be an important biomarker for predicting the outcome of transplantation.

3.4. Undesirable Polyreactivity of Therapeutic Antibodies

In recent years, antibodies with broad antigen‐binding potential have been the subject of considerable interest within the biotechnology industry [12, 13, 35]. The indiscriminate binding to multiple antigens is regarded as a significant drawback of therapeutic antibodies, potentially compromising their pharmacokinetics and increasing the risk of adverse effects in patients [12]. As might be expected, there has been a notable increase in research activity aimed at elucidating the underlying mechanisms and determining the factors that influence the broad antigen‐binding reactivity of monoclonal antibodies. This has been accompanied by the development of experimental and computational tools for the prediction of this antibody behaviour [35, 115, 116, 117].

4. Mechanisms of Antibody Polyreactivity

A considerable number of studies have been conducted with the objective of delineating the sequence or molecular features of antibodies that can explain their polyreactivity. However, these studies yielded a plethora of conflicting results. For instance, some studies have indicated that polyreactivity is more prevalent in antibodies with a germ‐line configuration of variable regions [36, 44, 118, 119, 120, 121, 122], whereas other research has suggested that the incidence of polyreactive antibodies increases during affinity maturation [28, 123, 124]. Similar discrepancies have been observed in the length of CDR H3, with reports indicating that polyreactivity correlates with longer CDR H3 loops, and others showing no significant differences in the length of this region [60, 76, 123, 124, 125] (for a representative structure of the antibody Fab fragment site and antigen‐binding site, see Figure 5).

FIGURE 5.

Structural organisation of Fab fragment of antibody. As a representative example, the structure of Fab fragment of highly polyreactive HIV‐1 neutralisation antibody m66.6 is shown. The heavy chain is coloured in blue the light chain is coloured in orange. The immunoglobulin domains are indicated. The antigen‐binding site is highlighted with read circle. The antibody has long and protruding CDR H3 loop. The structure of m66.6 (PDB: 4NRZ), was visualised by UCSF Chimera software.

A number of studies have indicated that the broad antigen‐binding activity of antibodies is associated with higher structural dynamics of the antigen‐binding site [29, 120, 126, 127, 128, 129, 130]. Nevertheless, recent structural and bioinformatics analyses comparing polyreactive and monoreactive antibodies have demonstrated that polyreactive antibodies may also possess antigen‐binding sites with greater rigidity than monoreactive ones [125, 131].

It is our view that these discrepancies stem primarily from the utilisation of disparate methodologies for the evaluation of polyreactivity and the application of varying criteria for the classification of an antibody as polyreactive. As proposed above, polyreactivity can be defined as the phenomenon of antibodies exhibiting different types of broad antigen‐binding activity, including polyspecificity and promiscuity. These modes of binding may have completely different molecular requirements for the recognition of multiple antigens. Furthermore, the antibody repertoires assessed in different studies vary in size or are selected or not for recognition of specific antigens. These differences can also introduce biases in the results. In recent years, more standardised procedures for the assessment of antibody polyreactivity have been implemented (see above), and larger antibody libraries have been assessed [35, 36, 37]. Notwithstanding the potential for methodological differences to introduce bias, a key message emerging from the existing literature is that there are many potential pathways for achieving broad antigen binding (Figure 6). This reflects the considerable sequence and molecular diversity observed in antibody repertoires. Therefore, it is important to recognise that generalisations about the mechanism of antibody polyreactivity may not be entirely accurate. Instead, it is necessary to accept that there are multiple pathways (potentially contradictory) leading to polyreactivity. The following section delineates the significant sequence and molecular correlates that have been linked to distinct forms of polyreactivity.

FIGURE 6.

Alternative mechanisms of antibody polyreactivity. Polyreactivity has been associated with high flexibility of antigen‐binding sites. This feature allows adaptability to distinct epitopes. Some antibodies can manifest polyreactivity without elevated molecular flexibility. Large molecular surfaces of the antigen‐binding sites allow docking of different epitopes using alternative parts of the molecular surface. Another mechanism resulting in polyreactivity is the presence of patches of similar charges or hydrophobicity on the antigen‐binding sites. These patches drive complementary association with patches on other macromolecules.

4.1. Conformational Dynamics of Antigen‐Binding Site

The conformational dynamics of antigen‐binding sites on antibodies can vary significantly between different antibodies. This conformational flexibility is typically concentrated in the CDR H3 loop [129, 132, 133] (Figure 5). It is noteworthy that polyreactivity has been previously linked to this region of antibodies, as well as to the entire heavy immunoglobulin variable region [37, 134, 135, 136]. A substantial body of research has established a correlation between the polyreactivity of antibodies and the structural dynamics of their antigen‐binding site [11, 21, 29, 44, 60, 120, 121, 126, 127, 128, 137, 138]. A greater degree of flexibility may facilitate enhanced adaptability to the diverse molecular characteristics of disparate antigens (Figure 6). The extent of conformational changes can range from minor movements of CDR loops in the binding site, which is referred to as ‘induced fit’, to significant reconfigurations of the binding site, which is referred to as ‘conformational isomerism’ [120, 126, 127, 130, 137, 138]. The aforementioned mechanisms were demonstrated by means of structural, kinetic and molecular dynamics simulation analyses, which provided evidence that they play a role in the recognition of unrelated antigens. A substantial reconfiguration of the binding site as a result of conformational isomerism allows a single antibody to bind completely unrelated antigens with relatively high affinity and specificity, thereby exhibiting a form of polyspecificity [126, 137]. The recognition of different epitopes may involve different contact residues from the antibody. The induced fit can compensate for slight differences in the epitopes of different antigens [127, 139].

A recent study utilising structural and molecular dynamics analyses demonstrated that individual polyreactive antibodies may possess antigen‐binding sites with greater rigidity compared to those of monoreactive antibodies [131]. Although the study involved only five polyreactive antibodies, it demonstrated that they are characterised by a more interconnected binding surface, which in turn restricts conformational dynamics. The findings of this study indicate that the flexibility of the antigen‐binding site is not a prerequisite for broad antigen‐binding activity. The rigid binding surfaces may facilitate another type of broad reactivity, known as differential antigen positioning [140] (Figure 6). The antigen‐binding sites of antibodies possess a considerable surface area, which allows for the formation of multiple alternative combinations of contacting residues for the binding of distinct epitopes. It has been demonstrated that antibodies can present with a polyspecific binding mode, engaging with different proteins or peptides using different parts of the antigen‐binding site [140]. In this regard, a flatter and more rigid binding surface would present with lower energetic penalties (unfavourable changes in entropy) in comparison to a more flexible surface.

A form of antigen‐binding polyspecificity that is mediated by a rigid antigen‐binding surface is the use of polyspecific cofactor molecules, such as heme [141, 142]. A fraction of antibodies in the human antibody repertoire are capable of binding heme, resulting in the acquisition of polyreactivity. Heme itself is a promiscuous hydrophobic molecule that interacts with numerous proteins [143]. This allows certain antibodies to utilise the molecular imprint of heme as a molecular bridge for binding of unrelated proteins [141, 142, 144].

4.2. Sequence Correlates Associated With Polyreactivity

A multitude of studies have sought to identify distinguishing characteristics between polyreactive and monoreactive antibodies by comparing their respective sequences. Polyreactivity has been observed in both germline antibodies [36, 44, 118, 119, 120, 121, 122] and antibodies with mutated variable regions [29, 36, 60, 123, 125]. There is a lack of consensus regarding the prevalence of polyreactivity among distinct subpopulations of B cells. Consequently, some studies have indicated that approximately 5% of human germline antibodies exhibit polyreactivity [28], while more recent research has demonstrated that 25% of unmutated human antibodies display polyreactivity [36]. Furthermore, polyreactivity has been linked to the preferential utilisation of specific gene families for the variable region, as evidenced by more frequent usage of the human VH1‐69 gene [36, 125].

Analyses of extensive antibody repertoires have demonstrated that polyreactivity is linked to elevated frequencies of specific amino acid residues in complementarity‐determining region (CDR) loops. For instance, polyreactivity was repeatedly associated with elevated levels of positively charged amino acids (arginine and lysine), aromatic amino acids (tyrosine and tryptophan) and hydrophobic residues such as valine [145, 146, 147, 148]. Moreover, it was demonstrated that the contribution of a given residue can be context‐dependent and that sequence motifs can also play a role [37, 146]. Another study demonstrated that the sequence characteristics of variable regions linked to polyreactivity may differ depending on the B‐cell type from which the antibody repertoires were derived [149].

In conclusion, the sequence correlation studies have not elucidated a definitive, robust correlation that can explain the polyreactivity of all antibodies. This finding reiterates the notion that numerous alternative strategies for achieving broad antigen‐binding activity are feasible.

4.3. Structural Correlates of Antibody Polyreactivity

The emergence of certain molecular properties of the antigen‐binding site was found to be associated with antibody polyreactivity. Consequently, a multitude of empirical findings point to the conclusion that polyreactive antibodies possess markedly elevated net charges within their variable region, or CDR H3 loops [36, 148, 150]. Other studies of the molecular surface of antigen‐binding sites have revealed that polyreactive antibodies possess surface patches of positively charged residues, which result in a strong accumulation of positive electrostatic potential [13, 148, 151, 152, 153, 154]. Furthermore, studies have identified the presence of hydrophobic patches on the surface of antigen‐binding sites of polyreactive antibodies [13, 152]. Interestingly, polyreactive binding has also been observed in antibodies with disproportionally negatively charged antigen‐binding sites, albeit less frequently [155]. The emerging properties of the binding site, when considered as an overall charge, can explain the polyreactivity of antibodies with promiscuous or indiscriminate antigen binding (Figure 6). For example, these antibodies can be directed and bind to oppositely charged surfaces on a multitude of proteins or other macromolecules. As would reasonably be expected, polyreactive antibodies frequently exhibit binding to the surface of apoptotic cells, where the negatively charged phosphatidylserine is displayed [47, 81].

The binding promiscuity that is mediated by hydrophobic or excessively charged patches will result in indiscriminate binding to all biomolecules that display hydrophobic or oppositely charged surfaces with low affinity. This form of binding may lack exquisite molecular precision and exhibit low affinity. Nevertheless, it can play a significant role in immune surveillance. It is noteworthy that the binding of such antibodies can be significantly enhanced when they present their binding sites in a multivalent manner (as in the case of IgM) and target repetitive epitopes on the surfaces of pathogens, damaged cells, or aggregated proteins. It is important to note that other proteins identified as ‘charge pattern recognition molecules’ [156], such as C1q, use electrostatics as a strategy of molecular mimicry and also display multivalency to promiscuously bind multiple targets with high‐binding avidities.

The detailed account of the types and properties of polyreactive antibodies shows such a diversity and ubiquity of this ‘exceptional’ property that the key concepts of antibody specificity may need reassessment.

5. What Is an Antibody?

What is an antibody if not specific? It is easier to discuss the well‐defined immunoglobulin molecules than the older concept of antibodies. Initially, the term antibody was defined operationally [157]. However, even today, the majority of details from immunoglobulin molecular and structural biology are still used to support the same old notion. The knowledge of antibodies has informed extremely successful prophylaxis and spectacular immunotherapy [158]. This knowledge also ensured the remarkable designability of the monoclonals as tools [159]. Although only part of the vast immunological ontology, the concept of antibody behaves as a typical Kuhnian paradigm. A very useful and resilient one, but not without its pitfalls, brilliantly discussed by Van Regenmortel [160]. For instance, antibody polyreactivity seems counterintuitive, a defect rather than a feature.

5.1. Alternative Views of Immunoglobulin Specificity

High‐throughput binding assays were used in several fascinating studies examining reactivity on a repertoire level [161, 162, 163, 164, 165]. Still, these studies remain a peripheral research niche, like the related idiotypic network theory. It proposed several system‐level hypotheses, but failed to prove them [166]. Probably, the key problem with all potential paradigm‐breaking antibody‐related concepts is the lack of a suitable novel methodology.

Antibody specificity, like affinity, is a quantitative characteristic. Using cut‐offs and dilutions, antibody reactivity is reduced to a discrete variable, masking most of the complexity. On the other hand, any monoclonal, irrespective of its specificity, will select thousands of short peptide mimotopes from a phage display library [167]. Phage display panning on monoclonals thus provides a comprehensive image of antibody specificity. Furthermore, the affinities of a monoclonal antibody to large sets of structures span orders of magnitude, with all those above micromolar affinity being biologically relevant [168]. Probably, the statistical characteristics of the distribution of affinities in general would be better‐suited descriptors of antibody specificity than a sketchy assay on whatever 10 or so proteins one happens to have in one's fridge.

5.2. More on the Statistical Image of Specificity—Quantitative Tolerance

Apart from being a simplification by discretisation, the common notion of specificity is also relative. The effective specificity in the internal environment depends on the antigenic landscape [169]. A series of transgenic animal studies showed the dependence of tolerance on antigen availability and BCR affinity [170, 171]. The human proteome contains only about 10⁷ potential linear epitopes and, probably, the same range of conformational ones. Proteins provide the highest diversity of structures, so non‐protein epitopes represent a small fraction. Under normal conditions, the epitopes accessible to the BCR for negative selection are certainly even fewer [169]. B‐cell tolerance to intracellular proteins and epitopes hidden in the native conformation of proteins is almost non‐existent [172]. It seems not to be enforced, as demonstrated in the cases of B‐cell and T‐cell epitope crypticity [173, 174, 175, 176, 177]. Also, antibodies to low‐abundance soluble proteins like cytokines are commonly detected [178]. Since antibody specificity is an adaptation to avoid autoreactivity [179], it needs to differentiate only against the self‐antigen landscape—a small accessible part of the enormous epitope space.

These considerations also draw attention to the notion of a quantitative tolerance. It is not a new idea—both B‐cell and T‐cell receptor selection depend on affinity thresholds [180]. Extended further, this concept would suggest that the tolerance tunes the physiological expression of self‐reactive antibodies to a threshold defined by the product of their affinity to and the typical levels of expression of the respective proteins/antigens [181]. In the case of antitumor immunity, this would allow targeting immune functions to self‐antigens, even if they are just overexpressed, for example, as tumour‐associated antigens. The current antibody paradigm appears to be mostly agnostic to this type of quantitative, tolerance‐dependent anti‐tumour specificity. Similar ideas have recently been further developed for B cells [182, 183] and T cells [184].

As speculative as it may sound, quantitative tolerance is not a new idea [185, 186, 187]. It can provide valuable insights, for example, into anti‐idiotypic phenomena, the interplay between antigenic immune receptor repertoires and the microbiome and so on. Let's consider that all antibodies have a degree of polyreactivity at low affinities (although some are characterised by a much higher degree than average). Their self‐reactivity would depend not on a rare accidentally cross‐reactive self‐antigen but on the combined concentration (mean field) of all weakly cross‐reactive self‐epitopes. Conversely, autoreactivity to an epitope would be formed by the combined activity of multiple weakly cross‐reactive antibodies. In this respect, there is no reason to exclude the very antigen receptor repertoires from the antigenic landscape. If the available repertoire participates indeed in the selection of the emerging repertoire [188, 189], including in the context of quantitative tolerance, the result is a network of idiotypic relations without the need for constant anti‐idiotypic immune responses.

5.3. Systems Immunology Methodology

The study of these complex interactions necessitates novel techniques. For its mathematical models and network‐based concepts, the systems immunology methodology derives data from various high‐throughput omics technologies, including single‐cell RNA‐seq, proteomics and metabolomics [190]. The antigen‐specific immune receptor repertoires are currently studied primarily through repertoire sequencing, which utilises sophisticated data processing methods [191, 192]. RepSeq experiments provide valuable insights into the clonal structure, diversity and organisation of the repertoire. For instance, Miho et al. established high‐performance computing platforms to analyse large‐scale antibody repertoire networks, revealing three fundamental principles: reproducibility, robustness and redundancy [192]. These network‐based analyses demonstrate that antibody repertoires maintain a functional architecture through quantitative principles, with networks being robust to the removal of up to 50%–90% of randomly selected clones, while being fragile to the removal of public clones. For similar techniques to have a substantial impact, they probably need to be linked and become interchangeable with high‐throughput repertoire reactivity studies [167, 193, 194]. Some efforts have also been made to train AI models to predict specificity from sequences [195].

Yet, all these approaches are still at various stages of initial development. Is this a new paradigm? Not yet. However, these ideas push the boundaries of the current one and offer a glimpse of what lies beyond it.

Author Contributions

A.D.P. and J.D.D. conceived the review and wrote the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Pashov A. D. and Dimitrov J. D., “Antibody Polyreactivity: A Challenger of Immune Paradigms,” Immunology 176, no. 4 (2025): 421–437, 10.1111/imm.70048.

Data Availability Statement

Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.