On the benefits of sin

Abstract

Antibodies (Abs) induced during infections with immunodeficiency viruses are subject to a form of original antigenic sin, termed repertoire freeze. This phenomenon encompasses conditions in which antigen (Ag)-specific B-cells and free Ab induced against early viral variants recognize viral escape mutants sufficiently to compete for Ag with naïve B-cells. As previously activated Ag-specific Abs and B-cells are more abundant than their naïve counterparts, they out-compete naïve B-cells and can be selected to undergo repeated rounds of somatic hypermutation and affinity maturation that drive repeated rounds of immune selection and viral escape. This situation prevents or diminishes the ability of B-cells carrying novel Ab-specificities to become activated and produce free Ab, facilitating viral escape. The enactment of repertoire freeze is illustrated in several features of anti-HIV antibodies, including persistently skewed κ/λ light chain ratios, preferential variable region gene usage, and the accumulation of Abs with extensive mutations within their variable regions. Furthermore, several investigators documented the presence of anti-viral Abs carrying a common idiotype, designated 1F7, from early infection onwards. In fact, anti-idiotypic suppression of these Abs in SHIV-infected rhesus macaques allowed the development of Abs that more effectively neutralized autologous contemporaneous viruses. Although most research suggests that repertoire freeze is undesirable for controlling an active infection, recent evidence has demonstrated that potentially protective broadly neutralizing Abs (BnAbs) develop within the freeze susceptible 1F7-idiotypic repertoire. This observation suggests that repeated rounds of selection of 1F7-idiotypic Abs may drive the extensive variable region mutation that characterizes BnAbs. In this review, we address how the demonstrated overlap between 1F7-idiotypic repertoire freeze and potentially protective Ab responses can be unravelled to generate novel vaccine concepts. Furthermore, we address how idiotypic regulation of the humoral immune response could be useful for sustaining protective Ab responses.

Keywords: :

• 1F7
• HIV
• repertoire freeze
• BnAbs

Introduction

Antibody (Ab) responses directed against the human immunodeficiency virus (HIV) envelope (Env) glycoproteins exert strong immunological pressure. Anti-viral humoral immune responses, which are induced during primary infection, are capable of virus neutralization and of mediating cytolysis of virally-infected cells through collaboration with cells of the innate immune system, such as natural killer (NK) cells.^1,2 As a result of this immunological pressure and the frequent introduction of mutations and shifts in glycosylation patterns, viral variants that have escaped these early immune responses emerge.^3,4 At some point in this process, the adaptability of components of the anti-viral Ab response is debilitated by a form of (OAS), known as repertoire freeze.^5,6

The phenomenon of original antigenic sin (OAS) refers to the tendency of the humoral immune response to reutilize clones activated during primary exposure to a pathogen after exposure to an antigenically similar, yet distinct, version of the germ, resulting in anti-viral Abs that are of a lower affinity and less efficient neutralizers of the secondary infectious entity.⁷ The occurrence of OAS has been demonstrated to occur in natural and experimental infections of humans and mice with several pathogens, most notably influenza and dengue viruses.^8-12

Repertoire freeze, with regard to HIV infections, refers to a phenomenon where Abs induced against early viral variants, despite not mediating effector functions that efficiently control viral replication, recognize escape mutant viruses sufficiently to forestall production of Abs with novel antigen (Ag) specificities.^5,6 Instead, ongoing stimulation of the originally induced B-cells results in boosts of the original Ab response. As memory B-cells are present in germinal centers upon reactivation during secondary immune responses,^13,14 the manifestation of repertoire freeze likely creates continuous cycles of somatic hypermutation, affinity maturation, immune pressure, and viral escape. The occurrence of this phenomenon is supported by the accumulation of anti-HIV Abs with high levels of somatic hypermutation during chronic HIV infection.¹⁵

Several lines of evidence support the concept and widespread occurrence of repertoire freeze of the anti-viral humoral immune response during HIV infection. One of the most basic examples that anti-HIV humoral immune responses are susceptible to original antigenic sin was observed following vaccination with HIV-1 SF2 gp120.¹⁶ Individuals who received this vaccine were not protected upon exposure to circulating viruses and those that became infected produced a secondary response against the HIV-1 SF2 Env rather than the infecting strain. Evidence of repertoire freeze during natural HIV infection has also been observed. For example, HIV-infected individuals often exhibit skewed Ab light chain κ/λ ratios, which are maintained for years regardless of the rate of disease progression.¹⁷ Restricted gene usage within the anti-HIV humoral immune response has also been documented.^17-19 Other investigators have reported Abs against earlier viral variants in contemporary plasma samples, despite viral escape many months past.^2,4 Perhaps the best and most easily observed example of repertoire freeze, however, is the appearance of a common idiotype, designated as 1F7, on anti-HIV Abs during primary infection (i.e., within 42 days of infection) and its maintenance and expansion throughout chronic infection.²⁰ 1F7-idiotypic Abs occur in approximately 70% of HIV-infected individuals,²¹ including those infected with diverse HIV-1 clades.²² The 1F7-idiotypic Abs are polyclonal, being selected from several different variable heavy chain genes and being directed against a variety of distinct viral epitopes.^21,23 Furthermore, particular variable heavy chain gene families do not appear to uniformly express the 1F7-idiotype. Indeed, screening of the Ab repertoire from HIV-infected individuals demonstrated the presence of idiotypic and non-idiotypic Abs with common variable heavy chain genes.²³

The 1F7-idiotype is represented by a region spread between the CDRH3 and FR3 and is recognized by a monoclonal murine IgM raised by immunization with anti-HIV Abs.^24,25 This anti-idiotypic Ab represents a novel form of anti-idiotypic Ab. The current classification of anti-idiotypic Abs depends on the relation of the idiotope target to the antigen binding site (Table 1). Anti-idiotypic Ab2 α (Ab2α) binds to an idiotope outside the antigen binding site and thus does not interfere with antigen binding of the antibody.²⁶ Ab2 β (Ab2β) recognizes the antigen binding site, blocks antigen binding and is considered an internal image of antigen.²⁶ Ab2 gamma (Ab2γ) binds near the antigen binding site and can interfere with antigen binding to some degree but not completely.²⁷ These Ab2s are restricted to species that harbor corresponding idiotypic Abs, or Ab1s. The 1F7 anti-idiotypic antibody differs from the classical Ab2α in two important aspects. (1) 1F7 binds to antibodies in different primate species including humans, chimpanzees, and macaques.^21,28,29 (2) 1F7 detects antibodies against different viral antigens of HIV-1 and Hepatitis C.^21,30,31 While the idiotope recognized by 1F7 is outside the antigen binding site of human Ab1,²¹ the absence of these unique properties in the current classification of Ab2αs prompted us to propose a new type of Ab2. In line with current nomenclature we add the term “Ab2 delta” (Ab2δ) to describe anti-idiotypic Abs such as murine 1F7 (Table 1).³²

Table 1. List of different types of anti-idiotypic antibodies

Anti-Idiotype class	Property	Reference
Ab2 Alpha	Non-binding site idiotope	Jerne, 1974
Ab2 Beta	Internal antigen image idiotope	Jerne, 1974
Ab2 Gamma	Near antigen binding site idiotope	Kohler, 1989
Ab2 Delta	Non-binding site, non-disease specific, and outbred shared idiotope	Kohler in Hoffmann, 2012

Involvement of the 1F7-idiotype in repertoire freeze is also reflected in studies with non-human primate species. Indeed, 1F7-idiotypic Abs have been observed in HIV-infected chimpanzees and rhesus macaques infected with the simian immunodeficiency virus (SIV) or the chimeric simian-human immunodeficiency virus (SHIV).^28,29 Elegant studies of the 1F7-idiotypic Ab repertoire in rhesus macaques provide some of the clearest evidence for repertoire freeze. Treatment of macaques with an Ab against the 1F7-idiotype suppressed 1F7-idiotypic Abs and allowed novel Ab specificities to emerge, which proved more efficient at neutralizing autologous contemporaneous virus (ACV).^33,34

The bulk of evidence indicates avoiding repertoire freeze is desirable in order to develop effective humoral immune responses during HIV infection or in the context of HIV vaccination. Nonetheless, certain characteristics of frozen anti-HIV Ab repertoires are embodied in the same Abs that potentially protect against acquisition of HIV infection. For example, anti-HIV Abs capable of neutralizing a broad spectrum of viral isolates exhibit extensive somatic hypermutation.³⁵ This level of somatic hypermutation is far too high to occur from single exposure to Ag and most likely reflects multiple rounds of selection and affinity maturation. Highlighting the possibility of repertoire freeze contributing to the evolution of protective responses is the remarkable observation that well-characterized broadly neutralizing Abs (BnAbs) uniformly express the 1F7-idiotype.²⁰ In addition, Abs that recognize HIV across diverse clades also express the 1F7-idiotype.²² Therefore, while repertoire freeze of anti-HIV humoral immune responses may be undesirable during HIV infection, a greater understanding of this phenomenon may assist with the design of vaccines that can induce Abs that protect against HIV infection. This review/perspective will discuss how understanding repertoire freeze, and its association with 1F7-idiotypic Abs, may explain several aspects of HIV humoral immune responses, including the potentially beneficial induction of BnAbs and maintenance of vaccine-induced anti-HIV humoral immune responses. We propose Ab repertoire freeze concurrent with 1F7 idiotype selection as a pathway to understanding how BnAbs arise in chronic infection and potentially as a means to select and maintain protective anti-HIV Abs through vaccination.

Repertoire Freeze and Selection of BnAbs and Other Protective Abs

Approximately 20% of HIV-infected individuals develop Abs capable of neutralizing a broad spectrum of HIV-1 isolates.^36,37 These Abs, termed BnAbs, target several distinct regions of the HIV Env glycoprotein heterodimer, which consists of gp41 and gp120. Targeted regions include the glycan shield³⁸ and CD4 binding site of gp120,³⁵ the membrane-proximal external region (MPER) of gp41,³⁹ as well as several epitopes that are contingent upon the successful alignment of homotrimers of gp41/gp120 heterodimers on the viral surface.⁴⁰ These Abs are seen as ideal candidates for an effective HIV vaccine to induce, as upon passive transfer to rhesus macaques and vector-induced expression in humanized mice they protect from infection with SHIV and HIV, respectively.^41-45 Despite the attractiveness of these targets for vaccines and the considerable interest they have invoked, attempts to induce BnAbs through vaccination have failed. The impediment to stimulating BnAbs through vaccination is, at least in part, thought to be related to the extensive somatic hypermutation required to generate these Abs.^35,46

The extensive somatic hypermutation exhibited by BnAbs suggests these Abs undergo repeated selection, which may actually be favored by anti-HIV Ab repertoire freeze. Indeed, the idea that these Abs can drive both viral evolution and their own repeated selection is supported by the observation that while these Abs neutralize a broad spectrum of viral variants, they do not slow progression toward the acquired immune deficiency syndrome (AIDS).⁴⁷ Further corroborating this suggestion is the fact that six well-characterized BnAbs (i.e., b12, 2G12, VRC01, 2F5, 4E10, and Z13e1) all express the 1F7-idiotype.²⁰ 1F7-idiotypic Abs are induced during primary infection and are maintained and expanded throughout chronic infection. As such, it is probable that the repeated selection to which 1F7-idiotypic Abs are subjected underlies the preferential selection of 1F7 idiotypic Abs to become BnAbs.

While repertoire freeze provides a potential mechanism for selection of BnAbs during natural HIV infection, it offers little indication of the exact means required for these Abs to be successfully selected during vaccination. Recent publications profiling the evolution of autologous viruses in individuals that have BnAb-like Abs could be expanded to include viral samples from time points prior to the development of BnAbs and provide clues regarding what Env immunogens are required to select BnAbs.⁴⁸ Furthermore, we recently demonstrated that 1F7-idiotypic Abs generated during an infection with one HIV-1 subtype react with Env glycoproteins from diverse HIV-1 subtypes²² suggesting that 1F7-idiotypic Abs target conserved regions within the HIV Env and that repeated selection, somatic hypermutation, and affinity maturation of these Abs may ultimately endow them with the ability to neutralize a broad spectrum of viral variants. Longitudinal studies of the B-cell population producing these Abs or sequencing of the genes utilized to produce these Abs will be necessary to determine if these Abs are indeed those that develop into BnAbs. If so, a greater understanding of both the initial Ab responses and viral Env evolution that is required for selecting BnAbs could facilitate the production of a BnAb-inducing vaccine.

Although the majority of research on Ab-mediated HIV vaccines has focused on neutralizing Abs, recent research suggests that neutralization may not be necessary for Ab-mediated protection from HIV infection. The RV144 vaccine trial in Thailand demonstrated partial protection, even though only virus-binding Abs in the absence of either strong neutralizing Ab or robust cytotoxic T-lymphocyte responses were induced.⁴⁹ Furthermore, the vaccine regimen used stimulates Abs capable of inducing NK cell-mediated antibody dependent cellular cytotoxicity (ADCC).⁵⁰ This led some investigators to suggest that ADCC was the mechanism of protection, and that ADCC-competent anti-viral Abs may be sufficient for protection.⁵¹ These hypotheses are supported by the observation that the protective capacity of BnAb b12 is diminished when a version of the Ab that mediates no ADCC is passively transferred to rhesus macaques prior to exposure to SHIV.⁵² Additional support for this possibility comes from a recent study in the rhesus macaque SHIV model of HIV-infection. Burton et al.⁴³ demonstrated that a non-neutralizing Ab clone directed against the gp41 component of the HIV Env can mediate limited protection against SHIV challenge in rhesus macaques.

Little is currently known about the role of Ab repertoire freeze in the selection of non-neutralizing Abs. However, previous research demonstrated the 1F7-idiotype on non-neutralizing Abs,²¹ and maintenance of non-neutralizing responses has been observed in the face of viral escape of non-neutralizing Ab effector functions, such as ADCC.⁴ As such, it appears non-neutralizing Abs are also subject to Ab repertoire freeze. Although repertoire freeze impedes the ability of Abs to mediate neutralization of a constantly evolving virus during natural infection,^33,34 the sharing of a common idiotype and understanding the mechanism of Ab clonal maintenance of both neutralizing and non-neutralizing Abs could be beneficial for designing long lasting HIV vaccines.

Utilizing Idiotypic Repertoire Freeze to Maintain Vaccine Induced Abs

Several mechanisms have been hypothesized and/or demonstrated to play a role in maintaining the presence of Ab clones during repertoire freeze of the anti-viral humoral immune response.^5,6 Most important among these mechanisms is the preservation of sufficient interaction between escape mutant viral variants and Ab clones directed against earlier viral variants to forestall stimulation of novel B-cells with fresh Ag-specificities. Although there has been little discussion of it to date, the notion that repertoire freeze involves maintenance of Abs commonly expressing the 1F7-idiotype raises the possibility that regulatory idiotypic interactions could play a role in maintaining repertoire freeze of anti-viral humoral responses.

Interactions between complementary Abs/B-cell receptors play a fundamental role in regulating humoral immune responses and can have both stimulatory and suppressive effects.^26,53-59 This is consistent with idiotypic/anti-idiotypic interactions playing a role in maintaining anti-viral humoral repertoire freeze during HIV infection. Previous research demonstrated anti-1F7-idiotypic Abs in HIV-infected individuals.²⁴ As the 1F7-idiotope falls outside of the Ag-binding region of anti-HIV Abs,²¹ it is conceivable that the interaction between 1F7-idiotypic Abs and anti-1F7-idiotypic Abs provides a non-interfering, activating signal to 1F7 expressing anti-viral Abs that provides them a selective advantage and contributes to the their maintenance. Beyond being simply a theoretically interesting possibility, a greater understanding of interactions of 1F7-idiotypic Abs and anti-1F7-idiotypic Abs could be essential for understanding how to maintain vaccine-induced anti-HIV Abs and ultimately elicit BnAb.

A major problem with potential Ab-based HIV vaccines, including RV144, is the rate that anti-viral Ab titers wane between immunizations.^60,61 Reinforcing this is the observation that the fading protection of the RV144 immunization is correlated with and is possibly a direct effect of the gradual loss of anti-viral Abs after completing the immunization regimen.⁶² This pattern of gradually losing anti-viral humoral responses after vaccination suggests that constant or recurrent presence of an immunogen is required to maintain anti-HIV Abs. Although some investigators have studied the possibility of using live-attenuated HIV to induce and maintain anti-HIV Abs,^63,64 this is considered too dangerous for implementation. Studies in primate models have demonstrated that attenuated viruses have the potential to revert to pathogenic variants and cause disease.⁶⁵

However, a safe potential mechanism to mimic the persistence of an immunogen in a vaccinated individual would be to induce anti-idiotypic Abs mediating stimulatory interactions with vaccine-induced idiotypic anti-HIV Abs to sustain their presence. Previous research has demonstrated that vaccine-induced anti-HIV Abs display the 1F7 idiotype.²⁹ As such, it may be possible to induce anti-1F7-idiotypic Abs by vaccination, following, prior to, or simultaneous with induction of anti-HIV Abs. These anti-1F7-idiotypic Abs would be able to interact with the 1F7-idiotypic anti-viral B-cells and sustain their presence via activating receptor cross-linking. Furthermore, the continued occurrence of receptor-based idiotypic/anti-idiotypic interactions should not interfere with the ability of circulating anti-HIV Abs to recognize virus or virus-infected cells, as the 1F7-idiotope falls outside of the antigen binding region.²¹

One major concern regarding a vaccine concept that involves inducing anti-anti-HIV Abs is that such Abs have the potential to present as autoimmune Abs. Indeed, the gp120 component of the HIV Env has the potential to mimic major histocompatibility complex class II.⁶⁶ As such, it is foreseeable that the idiotypic cascade induced after the initiation of the anti-HIV humoral immune response includes anti-HIV Abs that mimic CD4, as well as anti-anti-HIV Abs that mimic MHC II and bind CD4 mimicking anti-HIV Abs or CD4 itself. Several investigators have independently demonstrated the presence of anti-CD4 Abs in HIV-infected individuals, and some reports have linked these Abs with progression toward AIDS.^67,68 One report has even demonstrated that CD4-binding Abs from HIV-infected individuals can be anti-anti-HIV Abs.⁶⁹ The suggested vaccine strategy, however, should avoid selection of these CD4-binding Abs, as the anti-anti-HIV Abs being induced will be directed against the 1F7-idiotypic region of the anti-HIV Abs that falls outside of the viral-binding paratope.

A major caveat to this strategy for sustaining anti-HIV Abs is that it is based on the prediction that anti-HIV Ab titers wane after vaccination because of insufficient stimulation. An alternative possibility is that these Abs are eliminated by suppressive anti-idiotypic interactions. Indeed, it may be possible that vaccinated individuals develop anti-1F7-idiotypic Abs, as well as the expected 1F7-idiotypic anti-HIV Abs. These anti-1F7 idiotypic Abs may be able to induce complement-mediated destruction or ADCC of anti-HIV Ab-producing B-cells, reducing the number of cells capable of producing anti-HIV Abs. Assessing involvement of these phenomena in the waning of anti-HIV humoral immune responses after vaccination would require elimination of anti-1F7-idiotypic Abs for a period long-enough to allow anti-HIV Abs to be established as the dominant partner in the idiotypic/anti-idiotypic interaction.

Both of these possibilities could be tested in either the rhesus macaque SHIV or SIV infection models. Previous research has demonstrated that macaques infected with SHIV or SIV exhibit anti-viral Abs that bear the 1F7-idiotype.²⁸ Furthermore, a peptide that reflects at least a portion of the 1F7-idiotype that is present on anti-viral Abs has been identified and demonstrated to bind to both the murine anti-1F7-idiotypic Ab and anti-1F7-idiotypic Abs present in HIV-infected individuals.²⁴ This peptide could be used to screen vaccinated macaques for the presence of anti-1F7-idiotypic Abs. If such Abs are present, the peptide could be converted into a cytotoxic peptide, which could be utilized in vivo to eliminate anti-1F7-idiotypic B-cells, allowing an assessment of their effect on the anti-viral 1F7-idiotypic Ab population. If no anti-1F7-idiotypic Abs are present in the vaccinated macaques, the peptide could be utilized to induce anti-1F7-idiotypic Abs. This would enable assessment of the potential for utilizing anti-1F7-idiotypic Abs to sustain 1F7-idiotypic anti-viral humoral immune responses.

There is much precedent from outside of the HIV literature that idiotypic/anti-idiotypic interactions can be utilized to alter the Ab repertoire, and that some idiotypic-based interventions can provide immunotherapeutic benefit. Indeed, several groups have demonstrated that anti-idiotypic Abs can enhance or suppress idiotypic Abs.^56-59 Several factors could potentially contribute to the direction of the effect of anti-idiotypic Abs. One important factor appears to be the isotype of the anti-idiotypic Ab.⁵⁶ The value of these previous observations is reflected in the contemporary development of idiotypic-based understandings of autoimmunity and immunotherapies for cancers and autoimmune diseases.^70-72

Although it is currently unknown why anti-HIV Abs wane with duration since last immunization, it is possible that suppressive idiotypic/anti-idiotypic interactions play a role. If such interactions are not mediating a suppressive effect on anti-HIV Ab titers, it may be feasible to establish a concentration of anti-idiotypic Abs via vaccination sufficient to stimulate and sustain idiotypic anti-HIV Abs. Future exploration of the humoral immune responses of vaccinated individuals should investigate the presence of anti-1F7-idiotypic Abs. If these Abs are present, their role in suppressing anti-HIV immune responses should be elucidated. Alternatively, the potential to use stimulatory idiotypic/anti-idiotypic interactions can be investigated in primate models that produce 1F7-idiotypic anti-viral Abs.

Conclusion

As was recently reviewed by van Regenmortal,⁷³ most research regarding the induction of anti-HIV BnAbs via vaccination is hampered by a reductionist approach. Many attempts have been made to induce BnAbs simply by utilization of ligands that have been observed to bind to the Abs in in vitro assays, with little consideration of the vastly complex array of potential paratopes that constitute the Ab variable region or of developmental pathways and immunoregulatory mechanisms required for eliciting such highly mutated Ab. BnAbs are frequently viewed as stand alone biochemical structures, and, with a few exceptions in the literature,^74,75 rarely seen as structures that develop within the context of an organism with numerous immunoregulatory mechanisms. The view presented in this review/perspective is an attempt to develop a balanced approach to the BnAb problem, which takes into consideration the potential developmental and immunoregulatory mechanisms that encourage development of BnAbs within an organism. The goal of this manuscript was to present repertoire freeze as a pathway for selecting vaccine-induced anti-HIV Ab responses that can be sustained through the multiple rounds of affinity maturation required to generate HIV BnAbs, thus drawing attention to the pragmatic value of repertoire freeze. By understanding the anti-HIV humoral immune response through the framework of the repertoire freeze hypothesis, it is possible to broaden the approach to generating HIV BnAbs with novel vaccine constructs that would otherwise be unanticipated.

Conflict of Interest

The authors declare a conflict of interest. MSP, SM, HK, and MDG are members of the scientific advisory board for Network Immunology Inc.

Funding

This work was supported by a grant from the Canadian Institutes for Health Research (CIHR) #HVI 79515 and the Fonds de la Recherche en Santé du Québec (FRSQ) AIDS and Infectious Diseases Network. M.S.P is supported by a CIHR Vanier Scholarship.

Acknowledgments

We wish to acknowledge Geoffrey W Hoffmann, chief scientist of Network Immunology Inc., for his assistance in preparing this manuscript.