Current HIV/AIDS Reports

, Volume 7, Issue 2, pp 53–59

Systems Biology-Based Approaches to Understand HIV-Exposed Uninfected Women


    • National Laboratory for HIV ImmunologyPublic Health Agency of Canada
  • J. Sainsbury
    • National Laboratory for HIV ImmunologyPublic Health Agency of Canada
  • F. A. Plummer
    • National Laboratory for HIV ImmunologyPublic Health Agency of Canada
  • T. Blake Ball
    • National Laboratory for HIV ImmunologyPublic Health Agency of Canada

DOI: 10.1007/s11904-010-0039-3

Cite this article as:
Burgener, A., Sainsbury, J., Plummer, F.A. et al. Curr HIV/AIDS Rep (2010) 7: 53. doi:10.1007/s11904-010-0039-3


Worldwide HIV infects women more frequently than men, and it is clear that not all exposed to HIV become infected. Several populations of HIV-exposed uninfected (EU) women have been identified, including discordant couples and sex workers. Understanding what provides natural protection in EU women is critical in vaccine or microbicide development. However, correlates of protection in these women are still unclear. Most studies have used classical methods, examining single genes or cellular factors, a mainstay for traditional immunobiology. This reductionist approach may be limited in the information it can provide. Novel technologies are now available that allow us to take a “systems biology” approach, which allows the study of a complex biological system and identifies factors that may provide protection against HIV infection. Herein we report developments in discovery-based systems biology approaches in EU women and how this broadens our understanding of natural protection against HIV-1.


HIV/AIDSProteomicsTranscriptomicsSystems biologyHIV resistanceBiomarkersMucosal immunology


The global HIV/AIDS pandemic currently affects an estimated 33 million individuals [1]. Even after decades of research and implementation of prevention programs there has yet to be a successful vaccine or microbicide developed. However, several studies worldwide have identified individuals, who although highly exposed, remain uninfected by HIV. The study of these individuals may provide clues to what provides protective immunity against HIV-1; this information could be used to design a successful vaccine and/or preventative treatment.

There has been particular interest in studying exposed uninfected (EU) women. Globally, the HIV-1 pandemic disproportionately affects women due to numerous biological and social factors, and this is evident in sub-Saharan Africa and other developing areas [2]. Thus, interest has grown in the development of preventative technologies for women, such as vaginal microbicides [3], and EU women may provide information on critical correlates of protection.

What do we know about EU subjects? Besides known genetic factors that confer immunity to HIV-1, such as the CCR-5 mutation [4], many factors have been correlated with protection in EU individuals, although few have been validated in other cohorts, or follow-up studies. These include specific innate and adaptive factors, including CD8+ T-cell responses [5], upregulation of specific KIRS [6], polymorphisms of the IRF-1 gene [7], polymorphisms in DC-SIGN receptors [8], and others [9].

Additional innate and adaptive mucosal responses have been detected in EU women. This includes the overexpression of anti-HIV-1 factors RANTES [10], cc-chemokines SLPI and MIP-α/β [11], and neutralizing immunoglobulin A (IgA) [12]. However, some of these factors, such as HIV-1–specific IgA, have been shown to be correlated with repeated exposure rather than resistance to HIV-1 infection [13]. This highlights the fact that many of these factors correlate with exposure due to high-risk behavior (commercial sex work for example) rather than a true correlate of protection.

Although these approaches have yielded a greater understanding of potential immunity against HIV-1 in EU women, the reductionist approach taken thus far may under-appreciate the complexity of the immune response. This can be realized when one searches the literature for all previously described HIV-1–inhibitory factors present in the mucosa of the female genital tract. They include numerous factors, including RANTES, SLPI, α/β-defensins, lysozyme, lactoferrin, calprotectin, cystatin, and histone H2A, to name just a few [1419]. This only represents a fraction of detectable protein factors in cervicovaginal fluid, of which at least 500 unique protein species have been characterized (many with unknown biological functions [20]). Therefore, a better global understanding of this complex site is important if we are to have a better understanding of factors that confer protection against HIV-1. Indeed, the dynamic properties of a biological system cannot always be inferred from characteristics of individual isolated components of cells or organisms.

Thus, a “systems biology”-based approach may provide a better understanding with respect to the host immune response to HIV. Systems biology attempts to take into account the dynamic relationships among and between genes and proteins to determine how these interactions come together to form a functional biological system [21]. With the development of “–omics” technologies (such as transcriptomics, proteomics, and metabolomics), systems biology has become a reality and has been a useful tool to study various human diseases, including cancer, diabetes, obesity, and neurological disorders. Here we review the information generated from transcriptomic and proteomic-based studies of EU women and discuss the potential these approaches bring to the field.

Systems Biology Platforms Used to Study EU Women: Transcriptomics, Proteomics, and Genome-Wide Association Studies

There has been significant advancement in the development of technologies to study biological systems of cells and organisms. Although the idea of systems biology is not new, the maturation of technologies that allow generation of complex information is fairly recent. Development of high-throughput DNA/RNA microarrays and serial analysis of gene expression technologies (SAGE) allows the screening of whole genome expression (transcriptomics). Two-dimensional differential in-gel electrophoresis (2D-DIGE) and high-resolution mass spectrometers have made the study of global protein expression possible (proteomics). Mass spectrometry-based platforms are also being used in other areas of systems biology, including lipidomics (study of lipids), metabolomics (study of cellular metabolites), and even degradomics (study of protein degradation products). Furthermore, genome-wide association (GWA) studies are being used to identify genetic variations, or single nucleotide polymorphisms (SNPs), associated with disease phenotypic traits. Employing DNA microarray technology, up to 1 million SNPs can be interrogated, as well as determining copy number variation, and the imminent release of affordable whole-genome sequencing will enable increasingly comprehensive GWA studies.

The power of these technologies is their ability to generate vast amounts of unbiased information of a disease that may be the result of multiple factors. This is likely to be the case with infectious diseases such as HIV, as the interactions between virus and host are notoriously complex. Certainly then, these technologies would bring value to understand what provides protection in EU women. Indeed, many large-scale studies are underway, including a GWA examination of commercial sex worker (SW) cohorts from Kenya and other EU populations (eg, acutely exposed hemophilics), which is being organized by the CHAVI consortium (, but results of these have not yet been published. To date, only two systems biology-based approaches have been reported in EU women—one examining the transcriptome of EU women in an Italian cohort of discordant couples, and three proteomic studies of EU women belonging to commercial SW cohorts in Kenya.

Investigating the Transcriptome of EU Women

The transcriptome is the entire transcript content of a cell and represents a global perspective on whole genome expression. Genomic transcripts, or mRNA, can vary significantly depending on the developmental stage of the cell, or from external influences. Thus, the examination of the transcriptome of a biological sample, whether it is whole tissue, or a single cell, provides a window into its physiological state. Its utility in examining human diseases stems from the presumption that any quantitative or qualitative differences in gene expression in a biological sample can provide information into understanding the genetic basis for its cause.

Technologies used to examine global gene expression include both hybridization-based and sequence-based approaches. Hybridization-based expression array technology involves incubating fluorescently labeled cDNA with high-density oligo microarrays [22]. Hybridization-based approaches are high throughput, cost effective, and have allowed for high-resolution transcriptome mapping. In addition, when microarray probes are designed to interrogate all known exons and/or exon junctions, alternatively spliced isoforms can be detected and quantified [23]. However, the design of microarrays is limited by existing knowledge of the genome. Furthermore, it is prone to technical issues such as high background levels, limited dynamic range of detection, and difficulties normalizing expression levels between experiments. The choice of samples studied, tissue or cell type, can significantly affect the outcome and interpretation of these results as well.

Sequence-based approaches offer alternatives to hybridization-based methods, and include SAGE [24], cap analysis of gene expression [25], and massively parallel signature sequencing [26]. These are all tag-based methods and involve the isolation and Sanger sequencing of 10–21 base-pair sequence tags that can uniquely identify transcripts. While high throughput, the use of short tags means alternative-splicing isoforms cannot always be identified and the genomic position of the transcript be mapped. RNA-Seq is a recently developed transcriptomic protocol and aims to overcome many of the limitations of these approaches [22]. In using deep-sequencing technologies to characterize cDNA transcriptome libraries generated from relevant tissue [27]. RNA-Seq is a cost-effective, high-throughput procedure and can generate comprehensive information with respect to transcript quantity and isoform, and genomic position, without a priori knowledge. Despite the limitations of cDNA microarrays, the cost effectiveness and high throughput of these approaches make them a popular choice.

With regard to HIV, transcriptomics has been used to understand the mechanism of HIV-1 infection in vitro, determining factors responsible for inhibiting virus replication in nonpermissive cells [28, 29]. Recently, a population-based study examined the transcriptome of highly exposed persistently seronegative (HEPS) individuals [30•]. In this study, transcriptome mRNA expression analysis was performed on peripheral blood T cells from EU/HEPS subjects, their HIV-1–infected sexual partners, and healthy controls. Results from SAGE of activated T lymphocytes revealed two gene products, PRDX2 and IL-22, having enhanced expression in EU women, as compared with controls. The NK-enhancing factor, peroxiredoxin II (PRDX2), has been previously demonstrated to have anti-HIV activity [31], while IL-22 presented a novel candidate for protection against HIV infection. IL-22 was demonstrated to induce upregulation in cervical epithelial cells of β-defensins 2 and 3, both of which have been reported to inhibit HIV replication. IL-22 is also involved in the production of acute-phase proteins such as the acute-phase serum amyloid A (A-SAA). Complementary proteome analysis by Missé et al. [30•] showed the upregulation of A-SAA in the same cohort of HEPS individuals. Further, incubation of immature dendritic cells with A-SAA resulted in increased CCR5 phosphorylation and decreased CCR5 expression, decreasing the susceptibility of these cells to in vitro infection. With the identification and validation of a cascade involving IL-22 as a correlate for host resistance to HIV infection, this study elegantly demonstrated the potential of transcriptomics to elucidate underlying mechanisms of protection without a priori hypothesis. Certainly, this accentuates the strength of systems biology to generate information on candidate genes in a short period of time.

Investigating the Proteome of EU Women

Proteomics, from a systems biology perspective, is the study of all proteins within a cell, tissue, biological fluid, or organism. This can involve the identification and catalogue of proteins present, or an examination of differential expression between biological states. The field of proteomics has generated a wide variety of platforms and approaches, but there is no single method or strategy allowing the routine and complete analysis of the entirety of the proteome. The reason for this lies in the inherent complexity of the proteome. The number of protein products far exceeds the number of genes in the corresponding genome, and combined with the number of post-translation modifications, splice variants, and other modifications, their analysis is a challenge. The dynamic ranges of proteins, which can span ten orders of magnitude in abundance, adds to this complexity. As a result, many different platforms are needed to examine the diverse nature of proteins. However, to date, proteomic studies have yet to exceed 50% coverage of the entire proteome of a given cellular system, or 10% of any living organism.

The benchmark platforms used to examine differential protein expression include gel-based techniques, such as 2D-DIGE, or mass spectrometry-based techniques incorporating liquid chromatography separation. 2D-DIGE involves labeling protein samples with different fluorescent dyes (which does not interfere with the isoelectric point or migration pattern), mixing them together, and co-resolving on a two-dimensional gel. Excitation with specific light wavelengths can then give relative quantitative information of the protein spots on the gel for each sample. Spots of interest are then excised, digested by trypsin, and identified by tandem mass spectrometry.

All mass spectrometry-based platforms generally involve a similar workflow, consisting of pre-fractionation of the sample by various forms of chromatography and digesting the proteins with trypsin to generate peptides that can be quantified and identified by a mass spectrometer. This is commonly followed by differential stable isotope labeling (eg, iTRAQ, ICAT, SILAC), which can distinguish peptides between different sample sets and provides a starting point for quantification. Recently, label-free approaches have been emerging as an alternative to labeling, although with lower accuracy [32]. Peptides are then further fractionated by chromatography and then profiled by a mass analyzer to provide qualitative and quantitative information. Many types of high-resolution mass analyzers are available, such as quadrupole time-of-flight, tandem time-of-flight, ion trap, orbitrap, and ion cyclotron resonance instruments [33]. Currently, the Fourier transform-ion cyclotron resonance and orbitrap are becoming instruments of choice due to their high resolution and mass accuracy.

Some of these technologies have been applied to study differences in the proteome of genital secretions of EU women. Several previous studies have suggested that EU subjects have qualitative and quantitative differences in their mucosal immune system. The first of these studies examined genital secretions of female SWs from the Pumwani SW cohort in Kenya, a well-described EU population that can be epidemiologically defined as relatively resistant to HIV infection [34]. Using 2D-DIGE, protein expression profiles were compared between HIV-1–resistant, HIV-1–infected, and HIV-1–uninfected SWs, as well as HIV-1–uninfected non-SW controls (n = 39) [35•]. Analysis revealed over 15 proteins to be differentially expressed between HIV-1–resistant women and controls, some of which with a greater than eightfold change. Most overexpressed proteins were antiproteases, including those in the serpin B family (B1, 3, 4, and 13), α-2 macroglobulin-like 1, and cystatin A. Underexpressed proteins in HIV-1–resistant women included complement components, apolipoproteins, and Rho dissociation inhibitor (Table 1). A more comprehensive examination of over 500 individuals in this same SW cohort using a complementary platform, 2D LC-FT-MS, confirmed previous findings but also identified other antiproteases to be overexpressed (Burgener et al., Unpublished data [36]). Of the more than 350 unique proteins identified, 29 proteins were found to be differentially expressed (> twofold abundance) between HIV-1–resistant and HIV-1–uninfected, and HIV-1–infected control groups. These included many of the same overexpressed proteins indentified in the 2D-DIGE study, but also included more of the serpin antiprotease family. Underexpressed proteins included different complement components and inflammatory proteases (cathepsins; Table 1). The difference in findings is not surprising given the techniques used, as 2D-DIGE examines intact proteins (to-down approach) while 2D LC-FT-MS examines peptides (bottom-up approach), and thus they often identify different parts of the proteome, a routine observation in the proteomics field [37].
Table 1

Biological function of selected genes and proteins found differentially expressed in EU women as determined by systems biology techniques


Protein biological function

Biological sample

Study group



Overexpressed in EU women

Serpin family

Protease inhibitors

Cervical mucosa

Pumwani SWC


Burgener et al. [35•, 36]

α2 macroglobulin-like 1 protein

Elastase inhibitor, broad-spectrum endopeptidase inhibitor

Cervical mucosa

Pumwani SWC


Cystatin A + B

Cysteine protease inhibitors

Cervical mucosa

Pumwani SWC


S100A7 protein

Innate immune response, epidermis development

Cervical mucosa

Pumwani SWC



Cytokine, innate immune response

Blood T lymphocytes

DC, Italy


Misse et al. [30•]

Peroxiredoxin II


Blood T lymphocytes

DC, Italy


Acute-phase serum amyloid A

Apolipoprotein, inflammation

Blood serum

DC, Italy



Serine protease inhibitor

Cervical mucosa

Kibera and Pumwani SWC, Kenya


Iqbal et al. [38•]

Underexpressed in EU women

Complement components

Immune response

Cervical mucosa

Pumwani SWC, Kenya


Burgener et al. [35•, 36]



Cervical mucosa

Pumwani SWC, Kenya


Rho GDP dissociation inhibitor

Anti-apoptosis, cell motility, cell adhesion

Cervical mucosa

Pumwani SWC, Kenya


2D-DIGE two-dimensional differential in-gel electrophoresis, DC discordant couples, EU exposed uninfected, SAGE serial analysis of gene expression, SWC sex worker cohort

This observation that specific antiproteases are overabundant in the mucosa of EU women was confirmed by another cross-sectional study on two cohorts, including the Pumwani SW cohort described, and another SW population in Kenya, the Kibera SW cohort. Genital secretions from 124 HIV-resistant, 30 HIV-uninfected, and 161 HIV-infected SWs were collected for the study. Equal amounts of cervical lavage protein (1 μg) were spotted onto CM10 ProteinChip arrays (cationic surface) and profiled by SELD-TOF mass spectrometry. From this, a 6-kDa protein peak was discovered to be overexpressed significantly in HIV-1–resistant women, later identified by tandem mass spectrometry to be trappin-2/elafin [38•], a serine antiprotease. The increase was over twofold in HIV-resistant women over that of HIV-1–uninfected (P = 0.0012) and HIV-1–infected women (P = 0.0022). In a stratified case-control format, elafin/trappin-2 levels were measured in 112 individuals from the Kibera SW cohort, in whom protein levels were correlated to HIV infection outcome. In this blinded investigation, 24 of 112 women who became infected had twofold lower levels of elafin/trappin-2 than EU women. Women who had elafin/trappin-2 levels 1 standard deviation over the nonresistant population mean were significantly less likely to get infected (OR = 5.9; 95% CI = 1.9–24.4; P < 0.006).

It is interesting to note that most of the overexpressed proteins in EU women were protease inhibitors, most of which are produced by epithelial cells. Protease inhibitors play a variety of physiological roles, including regulating proteolytic activity of proteases involved in complement activation, coagulation, inflammation, and extracellular matrix remodeling. Serpins, which encompass a large family of antiproteases, are known to be important regulators of inflammation, the immune response, and aid in wound repair by inhibiting extracellular proteases [39]. A2ML1 is produced by epidermal granular keratinocytes and has been implicated to be a regulator of desquamation and protection of the epithelium against protease-mediated damage [40]. Trappin-2, a serine antiprotease, inhibits neutrophil-secreted proteases, such as elastase, and is important in wound repair, maintaining the integrity of the epithelial barrier, and controlling inflammation [41], demonstrating overlapping biological activities with serpins and A2ML1.

Mechanistically, how may these factors be important in HIV-1 infection? As inflammation of the female genital tract is known to increase susceptibility to HIV-1 infection [3], serpins/trappin-2 could have a plausible role in protection. Indeed, studies have shown that serpin B1 is crucial in protecting against protease-mediated inflammatory damage during host infection by Pseudomonas aeruginosa [42]. In addition, serpin B4 has been proposed to protect the epithelial barrier against chymase-induced inflammation and tissue degradation [43]. Many serpins inhibit cathepsin G, an inflammatory protease that acts as a chemoattractant for monocytes/neutrophils that can also enhance HIV infectivity in vitro, an effect of which serpins can directly abrogate [44]. Furthermore, some serpins directly inhibit HIV-1 replication, such as serpin A1 and C1, along with other antiproteases such as cystatins A and B [16, 45, 46]. Recently, trappin-2/elafin has been shown to have HIV-1–specific inhibitory activity, potentially as a binding inhibitor, supporting a role in mediating HIV-1 resistance [47]. Interestingly, IL-22, a cytokine upregulated in EU women of the Italian cohort, can stimulate innate defense molecules such as β-defensins in epithelial cells, and can also induce expression of antiproteases, including serpin A3 [48]. This suggests it may be a possible regulator of these factors, although it is unknown if this would extend to the mucosal compartment.

Many downregulated proteins in EU women have known roles in promoting HIV-1 infection. Complement components serve as chemoattractants that contribute to inflammatory responses. They are activated upon HIV-1 infection and can enhance infection by allowing opsonized virus to bind to complement receptor–positive cells and infect them more efficiently [49]. It also allows for fusion to nonpermissive cells, such as follicular dendritic cells. The complement cascade is also regulated by many antiproteases, including serpins, so lowered levels may be a result of this relationship [50]. Rho dissociation inhibitor has been implicated in mediating HIV-1–infected cell migration though tight junctions [51]. The observation that several factors that facilitate HIV-1 infection are differentially expressed in mucosal fluid of EU women supports a hypothesis that these factors could be working synergistically to provide a resistant phenotype.


Systems biology approaches offer novel insight into the biology of EU women and how they may be protected from infection by HIV-1. The results to date suggest that novel epithelial cell-secreted factors may play a role in altered susceptibility to HIV infection. Certainly the biological activity of many of these factors makes this a plausible hypothesis, as A-SAA has shown anti-HIV activity in vitro, as have many of the elevated mucosal antiproteases (such as trappin-2, cystatins, and serpins). In addition, these mucosal factors may provide a general anti-inflammatory effect in the female genital tract and enhance the maintenance of the epithelial barrier. A common regulator of these processes remains to be elucidated, but it is possible IL-22 could be playing a role, as it stimulates the production of several innate factors. This, in combination with the underexpression of HIV-1 activators, such as complement components, suggests that a combination of an overabundance of antiviral/anti-inflammatory and fewer activating factors may contribute to a protective environment, or at least one that is less likely to facilitate efficient HIV-1 infection at mucosal surfaces. Taken together, the likelihood of the protective mechanisms in EU women being the result of a sole factor seems to be less plausible, and a protective phenotype might lie in the synergy of these factors. Indeed, this hypothesis is supported by the observation that cationic antimicrobial factors at mucosal surfaces act in concert, while others cancel each other out, suggesting that this complex interplay is important [15, 52].

The utility of systems biology approaches is underscored in these studies. It demonstrates the complexity of human immunobiology and identifies new players that might not have been uncovered using a hypothesis-driven approach. However, in the absence of validation, the weakness in these techniques lies in the complexity of information that is generated, which makes it challenging to tease out which factors are important. As is the case with many systems biology experiments, reduction of this complexity is achieved by sorting genes/proteins by degree of differential abundance, which is based on an assumption that factors that are the most different are the most important, and this may not be the case. Clearly then, going forward, being mindful of the caveats of these studies will be important for hypothesis generation and to understand the biological significance of these findings.

Future studies should include quantitation of all of these factors longitudinally and correlated to disease outcome in high-risk individuals, in multiple cohorts, to validate a role in protection against HIV-infection. Furthermore, expansion of proteomic/transcriptomic studies of other biological sample types, especially those of the female genital tract (such as ectocervical tissue samples), would be useful to better understand host immunity at this understudied site. However, alternate technologies are available that could reveal novel insights into EU women. This includes, but is not limited to, epigenetic studies, which examine nonheritable genetic modifications that affect gene expression. It is entirely possible that these gene/protein expression perturbations observed in EU women are regulated by alterations not encoded genetically and should be explored. At the very least, the data from systems biology approaches have provided new avenues of investigation into what comprises protective immunity in EU women and warrant the expansion of these types of studies.


No potential conflicts of interest relevant to this article were reported.

Copyright information

© Springer Science+Business Media, LLC 2010