22Persistence of HIV-1 latent reservoir cells during antiretroviral therapy (ART) is a major 23 obstacle for curing HIV-1. Latency-reversing agents (LRAs) are under intensive development to 24 reactivate and eradicate latently infected cells; however, there are a few useful models for 25 evaluating LRA activity in vitro. Here, we established a chronically HIV-1-infected culture 26 system harboring thousands of different HIV-1-infected cell clones with a wide distribution of 27 HIV-1 provirus similar to that observed in vivo. A combination of an LRA and an anti-HIV-1 28 drug successfully inhibited viral re-emergence after drug discontinuation, demonstrating 29 "experimental cure" in the in vitro model. We demonstrated that the epigenetic environment of 30 the integrated provirus plays a role in determining drug susceptibility. Our widely distributed 31 intact provirus elimination (WIPE) assay will be useful for optimizing therapeutics against
32HIV-1 latency and provides mechanistic insights into the selection of heterogeneous Advances in antiviral therapy have dramatically improved the therapeutic options available for 36 treating human immunodeficiency virus type 1 (HIV-1) infection. However, even with the most 37 potent combined antiretroviral therapy (cART), HIV-1-infected patients remain on medication 38 throughout their lifetime because HIV-1 persists in viral reservoirs in vivo regardless of 39 treatment 1-3 . In this regard, the "shock and kill" approach, which first activates cells latently 40 infected with HIV-1 2,3 using small molecule agents called HIV-1 latency-reversing agents 41 (LRAs), is a possible strategy for curing HIV-1 4-9 . LRAs reverse HIV-1 latency and induce viral 42 production in cells latently infected with the virus. In theory, infected cells that express viral 43 antigens are then killed by the human immune system, such as cytotoxic T lymphocytes, or viral 44 cytopathic effects 10-12 . However, LRAs that appear potent in in vitro assays are not necessarily 45 effective in vivo because the viral reservoir situation is quite different in vitro and in vivo [13][14][15][16] .
46The host factors shaping the HIV-1 reservoir in vivo include the immunological status with 47 respect to the virus, anatomical location, and a variety of host cells. From the viral perspective,
48wide heterogeneity is noted in vivo, such as the viral sequence, presence of defective proviruses,
49integration sites (ISs) in the host cellular genomic DNA, and expansion of some infected 50 clones 17-21 . These factors potentially affect the efficacy of LRA in vivo.
51As the "shock" step may trigger the production of infectious virus and thereby induce 52 de novo infection, it is essential to combine LRAs with the existing anti-HIV-1 drugs. However,
53no suitable in vitro model system to evaluate the efficacy of such combination therapies exists.
54Currently available in vitro models for HIV-1 latency, such as ACH2, J1.1, and U1 cells, carry 55 only one or two integrated proviruses with a specific genetic and epigenetic...