Development of the VOA

The possibility of a reservoir of latently infected CD4+ T cells harbouring proviral DNA but not producing virus was identified in early clinical studies on peripheral blood from HIV infected persons [1]. When antiretroviral therapy later transformed the prognosis for people living with infection it became clear that this reservoir of latently infected cells would be a major barrier to eradicating the virus. The number of resting CD4+ T cells harbouring replication competent viruses has been considered the most accurate measure of the latent pool as it is thought that these cells are the source of the virus which rebounds after cessation of antiviral therapy [2, 3]. The demonstration that a population of CD4+ HLA-DR- T cells from patients on effective treatment could be made to release virus upon stimulation with phytohaemagglutinin (PHA) led to efforts to quantify the latent reservoir [4–6]. These studies established the principle of quantitation by seeding resting CD4+ T cells in limiting dilution before activating the cells with a combination of PHA and irradiated allogeneic peripheral blood mononuclear cells (PBMCs). Due to the low frequency of latently infected cells it was necessary to obtain a large volume blood sample (approximately 200 ml) from study participants.

The amount of HIV Gag protein released by reactivating a single latently infected cell is below the limit of detection of a standard p24 ELISA. Therefore the released virus is amplified by co-culture with permissive cells. In the original assay the cells were co-cultured with PHA stimulated, CD8-depleted PBMCs from an HIV-negative donor. Maximum likelihood statistics became the standard to estimate the frequency of latently infected resting CD4+ T cells, expressed as infectious units per million cells (IUPM) [7].

The use of allogeneic CD8-depleted PBMCs to amplify outgrowth virus has biological and practical limitations. PBMCs from different donors are not equally permissive for HIV infection, leading to different readouts depending on the PBMC donor. This problem is particularly acute when amplifying R5 tropic virus as expression levels of CCR5 vary widely between donors, CXCR4 on the other hand tends to be highly expressed making detection of X4 tropic virus less problematic. One solution is pre-screening of donor PBMCs for CCR5 expression level and selecting for high CCR5-expressing donors only. Another option is using cells from matched donor-patient pairs that are shown in vitro to exhibit robust viral replication, a method that also requires extensive pre-screening prior to the actual assay [8]. Although donor matching reduces inter assay variability for samples obtained from the same HIV-positive donor, it makes the procedure more labour intensive; and together with the need to perform leukapheresis to obtain sufficient latently infected cells, makes it harder for VOA to scale up to meet the needs of a clinical trial.

Novel VOA

A number of modifications have been introduced to improve the original viral outgrowth assay. Two new methods utilise cell lines engineered to express high levels of CCR5 [9, 10]. These cell line based assays have been shown to give IUPM readings similar to PBMC based assays [9, 10] and to give highly consistent results when sampling the same patient repeatedly [10]. Cell line based systems are also less labour intensive as the cells do not have to be freshly prepared from blood and stimulated with PHA for each assay. One assay uses a custom antibody cocktail to purify resting CD4+ T cells from PBMCs of HIV positive donors in a single step, significantly reducing processing time from approximately 2 h to 45 min [10], a huge advantage when scaling up assay numbers for clinical trials. Another recent modification proposed to reduce the work required for each assay is stimulating the resting CD4+ T cells with CD3/CD28 microbeads rather than PHA and irradiated allogeneic PBMCs. Although the results correlated well with those derived from the standard stimulus, the microbead based method appeared to be less sensitive [11].

Read-outs

Recently a number of strategies have been proposed for optimising the outgrowth assay by improving the sensitivity of the detection system or simplifying the read-out. Besides substituting the PBMC based co-culture with a CD4/CXCR4/CCR5 expressing clonal cell line, we recently reported that the observation of viral cytopathic effect in these cells showed a near-perfect correlation with the detection of p24 from infected cells (Fig. 1) [10], further simplifying the assay. Alternatively, use of a more sensitive PCR based detection method instead of the p24 ELISA can reduce the co-culture time required to detect replication-competent HIV [9]. The TILDA assay [12] utilises the detection of an increase in multiply spliced HIV RNA in CD4+ T cells after stimulation. Similar to the VOA, the cells are seeded in limiting dilution to obtain a quantitative measure. It uses total rather than resting CD4+ T cells, which has the advantage of requiring a smaller blood volume than a VOA and it does not need an extended period of culture for detection, however this means that it measures the inducible reservoir in a population of cells different to that used in the VOA. Another new assay, TZA, utilizes TZM-bl cells [13]. These cells produce β-galactosidase in response to viral Tat, enabling the presence of HIV to be determined by the enzymatic cleavage of a luminescent compound. In contrast to TILDA the TZA assay utilises resting CD4+ T cells but also requires a significantly smaller blood volume than the standard VOA. Neither of these assays correlated closely with IUPM measured by a standard VOA implying that they may not measure the same component of the HIV reservoir. Another caveat is that as neither assay requires viral replication to detect reactivated virus, neither provides unequivocal proof that the detected viral products are derived from replication-competent virus.

Another novel technique that significantly shortens the assay time of the VOA, by use of an enhanced sensitivity of p24 detection at ultra-low concentrations, has recently been applied to HIV latency research [14]. This ultrasensitive assay utilises Simoa single molecule automated array technology [15] which can measure sub-femtomolar concentrations of analytes using a micro-bead based antibody capture system. This promising method allows the detection and quantitation of p24 produced by a single reactivated infected CD4+ T cell [14]. Although this method is very powerful in detecting viral protein production very early after reactivation, it is not yet clear if the measure represents replication-competent virus and how it correlates to the VOA. Therefore, the results of ongoing clinical studies like RIVER (Clinical trial registry identifier: NCT02336074) which employ both a VOA and the ultrasensitive p24 assay, among many other measures of the latent HIV reservoir, will be of tremendous interest in the search for a definitive measure, or combination of measures, of the clinically significant HIV reservoir.

As efforts to cure HIV progress it may become necessary to be able to detect smaller reservoirs with greater confidence. A recent development on the VOA has been the development of the murine outgrowth system [16–18]. It involves the injection of PBMCs or CD4 T cells into immunodeficient NSG mice [17, 18] or humanised mice [16]. These systems offer enhanced sensitivity as they are capable of sampling large number of cells. They can also give a semi-quantitative estimate of the reservoir size and outgrowth from patient cells where a standard VOA had been negative [16]. However, these assays are not easily scalable, limiting their deployment.

Clinical utility