Date Published: September 12, 2017
Publisher: BioMed Central
Author(s): Amy E. Baxter, Julia Niessl, Antigoni Morou, Daniel E. Kaufmann.
Despite the tremendous success of anti-retroviral therapy (ART) no current treatment can eradicate latent HIV reservoirs from HIV-infected individuals or generate, effective HIV-specific immunity. Technological limitations have hampered the identification and characterization of both HIV-infected cells and HIV-specific responses in clinical samples directly ex vivo. RNA-flow cytometric fluorescence in situ hybridisation (RNA Flow-FISH) is a powerful technique, which enables detection of mRNAs in conjunction with proteins at a single-cell level. Here, we describe how we are using this technology to address some of the major questions remaining in the HIV field in the era of ART. We discuss how CD4 T cell responses to HIV antigens, both following vaccination and HIV infection, can be characterized by measurement of cytokine mRNAs. We describe how our development of a dual HIV mRNA/protein assay (HIVRNA/Gag assay) enables high sensitivity detection of very rare HIV-infected cells and aids investigations into the translation-competent latent reservoir in the context of HIV cure.
In spite of remarkable progress over the last 30 years, key challenges remain for the HIV/AIDS field. Persons infected with HIV require life-long anti-retroviral therapy (ART) to suppress viral replication, as the virus survives long-term in the form of latent HIV viral reservoirs  (VR). A functional cure is needed to treat these individuals and remove the financial and social burdens of long-term ART. In addition, the major hope to limit the spread of the epidemic, an effective HIV vaccine, remains elusive. We propose that the in-depth study of HIV infection in HIV-positive subjects is required to reach these two goals. Firstly, although the phenomenon of latency is well accepted, there remains a limited understanding of cells that support active HIV replication in ART-naïve subjects and serve as long-lived latent VR from which the virus can rebound in treated persons. Secondly, HIV-infected individuals can raise adaptive and humoral responses to the virus [2–4], which are highly inter-related [5, 6]. Studying how these HIV-specific responses vary between individuals, particularly in cases where HIV replication is naturally controlled, and how they compare to responses elicited by other pathogens or vaccines, provides valuable information for prophylactic HIV vaccine design. Lastly, bringing these two themes together, understanding the interactions between the VR and the individual’s immune system is critical for the development of therapeutic vaccines capable of eliminating HIV-infected (HIV+) cells in the cure context.
HIV-infected individuals mount detectable T cell-driven responses to the virus that may be important components of viral control . Yet, in the only trial showing protection to date (RV144), vaccine-induced T cell responses were highly variable and did not correlate with protection . However, the HIV-specific responses detected in these studies are defined and limited by the assays available to monitor them. Population-level techniques, such as Luminex bead arrays, provide only an overview of the response, masking crucial single-cell information. Activation-induced marker (AIM)-style assays [11, 12] detect a heterologous antigen-responsive population at the single-cell level, but alone provide limited functional information. Intracellular cytokine staining (ICS) enables single-cell analysis but requires the addition of protein trafficking inhibitors that may affect concurrent cell phenotyping; and furthermore a number of important CD4 T cell (CD4) cytokines stain poorly in ICS. Thus, cell populations of exceptional interest to the HIV field may be missed by such assays, with antigen-specific T follicular helper cells (Tfh) as a prime example.
The major source of HIV during chronic infection has long been recognized as CD4 T cells. The precise characterization of CD4 subsets harbouring replication-competent virus (i.e. provirus contributing to spreading infection) is crucial for the development of therapeutic vaccine or cure strategies targeting these VR; however studies have been limited by the strategies used to detect HIV+ cells. For example, in vitro infection studies are limited by the requirement for cellular activation, which alters cell characteristics, preventing phenotyping. Viral DNA measurement by PCR for total or integrated viral genes [17–19] and other assays performed on bulk cell subsets provide the VR size at a population, not single-cell, level; and limiting dilution strategies including the Quantitative Viral Outgrowth Assay (QVOA) [20, 21] and Tat/Rev Limiting Dilution assay (TILDA)  provide VR cell frequencies but do not allow phenotypic characterization of the infected cells. Therefore the relative contribution of distinct subsets can only be inferred by pre-sorting cell populations of interest prior to running the assay. This complexity is further confounded by the high prevalence of defective HIV genomes ; thus, PCR-based techniques represent a maximal estimate for VR size  while QVOA represents a minimal estimate . Therefore, a technique was required that enabled single-cell phenotyping of the HIV+ cells that were likely to contain non-defective, functional proviruses.
Despite its tremendous success, ART does not represent a cure. HIV is able to persist in the form of a latent reservoir from which the virus can rebound when a patient discontinues therapy. The identification, quantification and monitoring of the size of this reservoir presents a key challenge for HIV cure research. As mentioned above, multiple techniques are currently used, but caveats remain. We have used the HIVRNA/Gag assay to quantify the translation-competent latent reservoir in virally-suppressed, ART-treated individuals at a single-cell level, identifying rare HIVRNA+/Gag+ CD4 following maximal stimulation in vitro (median ~4 per million, ). Interestingly, we also identified very rare HIVRNA+/Gag+ in the absence of stimulation in a subset of samples from these individuals with suppressed plasma viremia (median ~1 per million in samples where events were detected). Whether this finding corresponds to limited viral reactivation or isolated HIV Gag protein/GagPol mRNA production in the absence of complete viral replication remains to be determined. Indeed, the implication that ongoing infection and spread could occur in the presence of suppressive ART is exceptionally controversial [28–30].
We have described here our use of RNA Flow-FISH to investigate interactions of HIV with peripheral CD4 T cells (both as a HIV reservoir and as an immune cell responding to viral antigen). However, the technique can easily be adapted to investigate tissues, alternative cell types including macrophages, for detection of multiple viral mRNAs and additional host factors. In summary, RNA Flow-FISH represents a powerful, highly versatile technique that has multiple applications for the broader HIV field.