Human immunodeficiency virus-like particles with consensus envelopes elicited broader cell-mediated peripheral and mucosal immune responses than polyvalent and monovalent Env vaccines

Abstract

Envelope (Env) sequences from human immunodeficiency virus (HIV) strains can vary by 15–20% within a single clade and as much as 35% between clades. Previous AIDS vaccines based upon a single isolate often could not elicit protective immune responses against heterologous viral challenges. In order to address the vast sequence diversity in Env sequences, consensus sequences were constructed for clade B and clade C envelopes and delivered to the mouse lung mucosa on the surface of virus-like particles (VLP). Consensus sequences decrease the genetic difference between the vaccine strain and any given viral isolate. The elicited immune responses were compared to a mixture of VLPs with Envs from primary viral isolates. This polyvalent vaccine approach contains multiple, diverse Envs to increase the breadth of epitopes recognized by the immune response and thereby increase the potential number of primary isolates recognized. Both consensus and polyvalent clade B Env VLP vaccines elicited cell-mediated immune responses that recognized a broader number of clade B Env peptides than a control monovalent Env VLP vaccine in both the systemic and the mucosal immune compartments. All three clade C Env vaccine strategies elicited similar responses to clade C peptides. However, both the consensus B and C Env VLP vaccines were more effective at eliciting cross-reactive cellular immune responses to epitopes in other clades. This is the first study to directly compare the breadth of cell-mediated immune responses elicited by consensus and polyvalent Env vaccines.

Introduction

The development of an effective AIDS vaccine has been hampered by the intrinsic diversity among circulating populations of HIV-1 in various geographical locations. There is a need to develop vaccines that can elicit enduring protective immunity to variant HIV-1 strains. Based on viral genetic distances and positions in phylogenetic trees, HIV-1 is divided into three separate groups: M, O, and N [1]. Group O infections are limited to central Africa and Group N infections are found in a small number of patients in Cameroon [2]. The vast majority of viral isolates are designated as Group M and they are responsible for the AIDS pandemic continuing worldwide. Due to the diversity in amino acid sequences between isolates within Group M, this group is further subdivided into nine subtypes or clades (A–D, F–H, J and K) [1]. The diversity between isolates in different clades can vary by as much as 35%.

The ultimate goal of an AIDS vaccine is to elicit potent cellular and humoral immune responses that will result in enduring, broadly protective immunity. Over the past 20 years, a number of potential AIDS vaccines have elicited immune responses that decreased viral set-points and maintained CD4⁺ T cells in vaccinated animals, but sterilizing immunity has not been achieved [3], [4], [5], [6]. Often, this limited protection occurs when the challenge virus matches the viral proteins in the vaccine [3], [4], [5], [6]. However, these artificial challenge models are not reflected in human infections and therefore the diversity of HIV-1 isolates requires vaccine designs that elicit broadly reactive immunity.

The greatest diversity is localized to the viral envelope glycoproteins. This may reflect the primary role in eliciting host immune recognition and responses that result in progressive evolution of the envelope proteins during persistent infection. Interestingly, while Env variation is widely assumed to be a major obstacle to AIDS vaccine development, there is very little experimental data in animal or human lentivirus systems addressing this critical issue. One method previously used to address Env variation in AIDS vaccines is a polyvalent vaccine strategy [4], [7], [8], [9], [10], [11]. These vaccines consist of a mixture of divergent isolates of the same antigen administered simultaneously. Polyvalent vaccines function by presenting a wide range of epitopes that cover a majority of individual strains. Vaccines for pneumococcus, poliovirus, and influenza virus have used this strategy [12]. Polyvalent vaccines increase the strength and breadth of humoral immune responses compared to monovalent vaccines [8], [9], [10], [13], [14], [15], [16], [17], [18]. The increase in the number of immunogens, below a certain threshold, does not result in immune interference [19].

More recently, centralized sequences have been developed as an alternative vaccine strategy to address viral sequence diversity. These artificial sequences are designed using computational methods to minimize the distance between a vaccine strain and a primary wild-type isolate. The three main strategies for developing centralized vaccines are center of the tree, ancestral, and consensus (for review see [20], [21]). Each of these designs has advantages for vaccine development. Consensus sequences minimize the degree of sequence dissimilarity between vaccine immunogens and circulating virus strains by creating artificial sequences based upon the most common amino acid in each position in an alignment [22], [23], [24], [25], [26], [27]. Recently, vaccine strategies utilizing consensus or ancestral HIV-1 Gag and Env sequences have proven potent inducers of CTL activity [23], [28], [29], [30], [31]. These consensus Envs form native structures and facilitate infection via the CCR5 coreceptor [32], [33].

In this study, monovalent, polyvalent and consensus Env vaccines were compared for the ability to elicit broadly reactive immune responses following vaccination in mice. Each set of Env vaccines was generated from clade B or clade C envelope sequences. HIV-1 virus-like particles (VLP) were used to deliver these Env sequences to the immune system in their native trimeric structure on the surface of a viral particle and the elicited immune responses were compared for the breadth of Env reactivity.