Host Defense Mechanisms Against Viruses

Gut Microbiome

Before birth, the intestine is generally sterile. The intestinal tract becomes colonized soon after birth with a variety of ingested environmental and maternal microorganisms. This process is influenced by many factors, including mode of delivery, diet, environment, and use of antibiotics. ^, For example, a breast-fed, term infant normally has an intestinal microbiota in which bifidobacteria predominate over potentially harmful bacteria, whereas in formula-fed infants, enterococci, bacteroides, and clostridia predominate. Experiments in mice demonstrate that the beneficial effects of commensal bacteria are mediated via Toll-like receptors (TLRs). The recognition of commensal bacterial-derived molecules by TLRs represents a critical component of the symbiosis between the host and indigenous microflora and is important for protection against gut injury and associated mortality. ^{, ,}

Gut Immune System Development

The development of the gut immune system is initiated before birth by a genetic program that drives the formation of Peyer patches and mesenteric lymph nodes, but its postnatal maturation depends on the establishment of a balanced indigenous microbiota. Intestinal commensal microorganisms provide signals that foster normal immune system development and influence the ensuing immune responses. Signals delivered by these commensal microorganisms drive the development of isolated lymphoid follicles, stimulate maturation of Peyer patches, and initiate migration of IgA-producing plasma cells, innate lymphoid cells, and mature T lymphocytes into the mucosa.

Models of Study of Fetal and Neonatal Immunity

Due to obvious safety and ethical issues, studies involving tissue and blood sampling of fetuses and neonates are limited. Most immunology studies in adult humans have been developed and performed with large-volume samples, but the volume of blood that can be acquired from neonates is typically on the order of a few milliliters at most. Amniotic fluid collection, used clinically for prenatal genetic diagnosis, can also give information about fetal inflammation and stress. Placentas may be obtained from elective terminations of pregnancy in the first trimester or after delivery in preterm or term infants. Tissue acquisition from fetuses and neonates is limited to fetal tissues obtained after medical termination of pregnancy, neonatal tissues from cases of sudden and unexpected infant death (SUID), or occasionally from small surgical biopsies. Maternal-embryo interactions in the first 8 weeks of gestation following embryo implantation probably are foundational in establishing immune system parameters, and this period of development likely differs significantly from later stages, but there is very little detailed or reliable information on tissues from this period in humans. Preliminary studies performed in the 1980s showed that cells involved in innate immune responses develop as early as the fourth week of gestation in humans, and mature fetal T and B lymphocytes can be detected by the end of the first trimester of gestation.

For these reasons, animal models are frequently used to investigate early immune development, but these models may not reliably recapitulate human immune development. Modeling of in utero development is challenging because of anatomic and physiologic differences between humans and animals. For example, no other species has a placental structure and function identical to that of humans. Postnatal animal models of neonatal immune development and response are also complicated by differences from humans. Neonatal mice are used often because of the ease of breeding and the availability of reagents that facilitate elegant mechanistic studies. The gestational period of mice (about 3 weeks) is quite short, and postnatal development is very rapid compared with humans. In contrast to humans, mouse T lymphocytes only develop after birth. It is difficult to identify days of life in neonatal mice that temporally correspond appropriately to weeks or months of human postnatal life. A variety of molecular characteristics of the immune response are significantly different between young mice and humans. For example, variation in TLR structure between species drives different responses to TLR agonists, which result in species-specific pathogen recognition. Therefore, mouse models often poorly mimic human inflammatory diseases and immune responses. Nonhuman primate models of immune development during pregnancy and the neonatal period are being investigated and have some promise, but they are expensive and require very special facilities and expertise.

B Cells/Antibodies

The antigen-specific receptor of B cells is formed by recombination of variable (V), diversity (D), and joining (J) genes. Recombining V, D, and J genes achieves a diverse repertoire of antigen receptors that can recognize the myriad foreign antigens in a highly regulated process consistent with ability to respond to pathogens without causing autoimmune phenomena. Although neonatal mice use different dominant antibody variable genes compared with adult mice, it appears that the general makeup of human neonatal repertoires is similar to that of adults. In the human fetus, early B cell development is associated with progressive diversification of the antibody gene repertoire. At term, the antibody gene repertoire possesses mature levels of junctional diversity, including nontemplated (N) and palindromic (P) type of additions at the V/D and D/J gene junctions. Furthermore, the average amino acid length of the complementarity-determining regions (antibody-variable loops) appears to be similar to that of adults. In contrast to adult antibody gene sequences, however, neonatal sequences have lower levels of somatic mutations. ^, This finding is presumably due to lack of extensive prior exposure to foreign antigens that drives terminal center formation, which is associated with B cell proliferation and differentiation in concert with the molecular program of somatic hypermutation. Using unmutated recombined antibody gene sequences to encode antiviral antibodies leads to the secretion of antibodies with low affinity for virus antigens, and consequently low antiviral function. Somatic hypermutation is a complex process that requires antigen presentation, CD4 T lymphocyte help via soluble factors and cell surface receptor expression, and interaction with B cells. In a robust germinal center reaction, stimulation of B cells leads to expression of activation-induced cytidine deamidase (AID) and error-prone DNA polymerases. It is possible that there is a lower level of intrinsic response in neonatal B cells in the germinal center milieu. However, human cord blood B cells do up-regulate the transcription of genes involved in somatic hypermutation, including AID and polymerases, following stimulation with CD154 and cytokines (mimicking T helper cell interaction). Human newborns can develop potent antibody responses to T cell–dependent vaccine antigens. Infants exhibit a profound deficiency in antibody response to polysaccharides, which is most pertinent in response to the capsular polysaccharides of some pathogenic bacteria including Streptococcus pneumonia , Neisseria meningitides, and Haemophilus influenzae ; it is not clear whether this deficiency also pertains to the response to glycans present on the surface glycoproteins of many pathogenic viruses.

Antibodies

Respiratory viruses are the most common cause of hospitalization of infants in the developed world, and thus they are interesting models for probing development of antibody-mediated immunity. The most common viral causes of serious lower respiratory tract disease in humans are RSV; parainfluenza virus (PIV) types 1, 2, and 3; and influenza viruses. The mechanisms by which antibodies contribute to resolution of or protection against infection or disease caused by respiratory viruses have been increasingly elucidated over the last decade through in vitro studies of neutralization and in vivo studies of protection. Antibodies neutralize respiratory viruses in vitro using a wide variety of molecular mechanisms. The cell substrate used for neutralization assays matters greatly, because the level of expression of membrane receptors and intracellular restriction factors varies among continuous cell lines. Important antibody characteristics that regulate neutralization potency include the Fab affinity for viral antigen, avidity of multivalent interactions, immunoglobulin (Ig) isotype (IgM, IgG, or IgA) and subclass (IgG1, IgG2, IgG3, or IgG4), concentration, molar ratio of antibody combining sites to epitopes on the surface of viral particles, state of polymerization (monomeric/polymeric IgA or IgM regulated by joining J chain), ability to bind the polyimmunoglobulin receptor (pIgR), ability of certain antibodies to fix complement, and virus protein and epitope specificity. Immunity is also affected by many viral factors, such as replication cycle features including use of diverse host molecules as attachment factors, mode of entry into the cell, pH dependence of fusion (neutral or low pH), and the natural site of replication. Individual characteristics of a specific virus strain can alter the antigenic recognition capacity of antibodies (corresponding to diverse serotypes, clades, antigenic subgroups, and amino acid sequences of viral surface proteins). Some of the major mechanisms of antibody-mediated virus neutralization ( Table 115.2 ) are reviewed next.

Aggregation

Antibodies possess the potential for bivalent or higher valency interactions, depending on isotype, allowing one antibody molecule to bind more than one copy of an epitope. Recognition of quaternary structures through binding of multiple epitope copies on a single oligomer or multiple oligomers can be associated with high potency of virus neutralization. If a multivalent antibody directed against viral surface proteins crosslinks multiple free virions in solution, however, these particles can be aggregated. It is thought that aggregation of multiple infectious virions into a single particle reduces the number of those virions to a single infectious particle, because size constraints suggest that a single particle usually attaches only to one cell. Aggregates caused by IgM, IgG, or IgA to influenza virus have been visualized directly in electron microscopy studies.

Blocking of Attachment

Antibodies can inhibit virus attachment to cell surface receptors or attachment factors; virion particles that cannot attach to cells do not initiate infection. The principal target of neutralizing antibodies for influenza virus is the hemagglutinin (HA) protein, which mediates both virus attachment to its cellular receptor (a glycoconjugate terminating in sialic acid) and fusion with an intracellular membrane at low pH. Some of the most potent anti-HA antibodies bind directly to the recessed receptor-binding site (RBS) for sialic acid on the head domain of HA, by inserting an antibody hypervariable loop (complementarity-determining region) into the site. Interestingly, there are canonical molecular modes of interaction of these antibody loops with the HA RBS that use either a charge-based interaction with an antibody loop aspartate or pi-pi stacking of aromatic residues to mimic the molecular interactions with sialic acid. The stacking of aromatic residues between attachment-blocking antibodies and virus RBS regions has also been observed in other families of viruses, such as neutralizing antibodies to filoviruses, and thus may be a very common and important mechanism of virus receptor blocking. Anti-HA IgA and IgG antibodies inhibit influenza virus attachment to both mammalian cell culture monolayers and tracheal cell epithelia. ^{, ,} The basis for the conventional hemagglutination inhibition test, often used for serologic assays of functional antiinfluenza antibodies, is blockade of virus attachment to sialic acid on the surface of red blood cells. Influenza is not unique in this regard. The HN (HA-neuraminidase [NA]) glycoprotein of PIV mediates attachment to sialic acid-containing host-cell receptors. Antibodies to several of the major antigenic sites on PIV3 HN are neutralizing. RSV does not use sialic acid as receptor; the cellular receptors for RSV are poorly defined, but the RSV G (glycosylated) glycoprotein is considered to be the virus surface glycoprotein that mediates attachment. Interestingly, antibodies to the RSV G protein typically possess a low level of neutralizing activity in vitro, and most monoclonal antibodies (mAbs) directed to the RSV G protein mediate only partial neutralization, even at maximal concentrations. This incomplete neutralization effect likely is due to heterogeneity in glycosylation of the virus protein.

Fusion Inhibition

Even if viruses attach to cells, some antibodies can neutralize the tethered virion particles at the cell surface before cell entry, especially if they inhibit fusion of virus particles with cell membranes. RSV can be inhibited in vitro by the presence of immune serum during the first 60 minutes after inoculation. This finding suggests that some neutralizing antibodies inhibit events that follow virion attachment to host cells. The role of fusion proteins varies depending on the type of virus (especially enveloped versus nonenveloped virion particles). Fusion proteins of enveloped viruses merge the virus lipid membrane with that of the host cell, either at the cell surface or for some viruses in the endosome. The fusion (F) protein of paramyxoviruses, for example, mediates direct fusion of the viral envelope with the cell membrane after attachment at neutral pH. Many antibodies directed to RSV or PIV3 F glycoproteins are effective at virus neutralization. The three-dimensional structure of paramyxovirus F proteins and the function of diverse regions of the F proteins during fusion of viral and cellular membranes have been defined recently. F proteins require cleavage from an F ₀ precursor to F ₁ and F ₂ subunits for infectivity, and a highly conserved hydrophobic region near the cleavage site functions as a fusogenic peptide. Some paramyxovirus-neutralizing anti-F antibodies inhibit fusion at the cell surface membrane. Fusion inhibition (FI) activity often is measured by inhibition of multinucleate giant cell (syncytium) formation caused by cell-to-cell fusion after productive infection of cell monolayer cultures. Because viral surface proteins traffic to the cell surface during the processes of assembly and budding (which are necessary for egress from the cell), antibodies can inhibit virus egress. Some anti-F antibodies appear to inhibit the release of progeny paramyxoviruses from infected cells. The epitope of binding to the F protein is important to the extent and mechanism of function observed, because some RSV or PIV F antibodies bind to F but do not neutralize virus, some of these antibodies bind to F protein and neutralize virus in vitro, and still others exhibit both in vitro neutralization and FI properties. ^, Major neutralizing sites have been defined for these fusion proteins by competition-binding assays, determination of the nucleotide sequence of the F proteins of mAb escape mutants, crystallography of mAb-virus protein complexes, and other techniques. Such studies have facilitated the first successful “reverse vaccinolog” studies to design structure-based epitope vaccines, using as a starting point the atomic resolution structure of a neutralizing antibody (motavizumab) in complex with the epitope derived from RSV F protein.

Other viruses, including influenza virus as a prototype, do not fuse with host cell membranes at neutral pH at the cell surface; instead, the attached virus particles enter the cell by endocytosis, then fuse with endosomal membranes when the pH of the vesicle drops (typically to a pH of about 5.2). Postattachment inhibition of endocytosis of influenza has been reported. Polymeric IgM appears to be especially effective in this setting, likely due to the large amount of steric hindrance associated with the 5 or 6 copies of IgM in such complexes.

Inhibition of Uncoating

Even after endocytosis, antibodies that remain bound to viral protein targets may mediate neutralization. Some influenza HA-specific neutralizing antibodies prevent uncoating of virus due to low-pH membrane fusion. ^, Some anti-HA IgGs appear to mediate neutralization of influenza virus replication at steps even later than primary uncoating, although the mechanism of such effects is unclear. ^,

Complement-Enhanced Antibody-Mediated Neutralization

The complement system comprises about 30 proteins and protein fragments that are part of the innate immune system. Most complement components are synthesized as inactive precursors in the liver and distributed by systemic circulation. Complement protein levels are low in neonates, especially the terminal elements of the complement cascade. As a result, neonates cannot form the membrane attack complex that is necessary for some antimicrobial responses. Certain isotypes of antibodies fix complement by interaction in the CH2 domain of antibodies, in the Fc region of the Ig molecule. The complement system mediates diverse immune functions, including aggregation of pathogens (as discussed previously for antibodies), chemotaxis-enhancing properties for neutrophils and macrophages, opsonization to enhance phagocytosis of virions by Fc receptor (FcR)–bearing phagocytic cells, and lysis of cells and virions by assembly of a membrane attack complex. Complement fixation (the combining of complement with antigen-antibody complexes) is mediated by certain isotypes of antibodies, and this activity often enhances viral neutralization. Without complement, antibody function is reduced. Complement is also necessary for optimal recruitment of immune factors during local responses, and it plays a role in antigen presentation by enhancing uptake of foreign antigens into antigen-presenting cells (APCs) through various complement receptors.