Data vs. Dogma: HIV-1 Integrations Driving Clonal Selection

HIV type 1 (HIV-1) is a lentivirus that causes a rapid viremia infecting roughly a billion cells per day, followed by a chronic illness characterized by progressive immunodeficiency due to the almost complete destruction of T helper cells (CD4⁺ cells). The replication of HIV-1 involves integration of its proviral DNA into the host genome, a process that occurs in a semirandom pattern with a few known hot spots and a generic bias for active transcription units.¹ With the exception of the untargeted integration, all known viral pathogenicity factors have been eliminated from clinically used lentiviral gene vectors.

One of the rationales for developing gene vectors based on HIV-1 was the lack of evidence that HIV-1 integration is oncogenic, in contrast to the typical findings with many other retroviruses. However, HIV-1 replication is cytopathic and immunogenic, causing cell death or target cell lysis by effector cells of the immune system unless the integrated proviral genome is defective or transcriptionally silenced. Thus, in the case of HIV-1, proto-oncogenic insertions are expectedly hidden by the dominant cytopathic effects. However, with the introduction of highly active combinatorial antiretroviral therapy (cART), there is a constant increase in the numbers of patients in whom viral replication is under pharmacological suppression, thus prolonging the half-life of infected and uninfected T cells. Importantly, in a new study of long-term cART recipients, Maldarelli and colleagues provide strong genetic evidence for the occurrence of clonal T-cell expansion driven by mutations of growth-regulatory genes as a direct result of HIV integration.²

Studying T lymphocytes and mononuclear cells from a relatively small cohort of patients using linker-mediated PCR and subsequent next-generation sequencing, Maldarelli et al. report the preferential expansion of defined clones with recurrent integration patterns in potential proto-oncogenes. The patients developed a heavily oligoclonal composition of the HIV-infected cell pool over time. Strong data sets arguing for positive selection linked to the retroviral integration event were demonstrated for MKL2 and BACH2, candidate proto-oncogenes for which, however, a direct mechanism of their role in clonal expansion remains elusive. Importantly, although genetic evidence acquired on the basis of high-throughput integration analysis is convincing, the authors provide no evidence for transcriptional or posttranscriptional changes in target gene function. Also, they did not use additional, independent methods to assess the exact dynamics and size of clonal expansion. Nevertheless, their report challenges a predominant view that has frequently been used to argue that HIV-based vectors are unlikely to act as insertional mutagens. We now find support for the conclusion that HIV-1 integration is a potential mutagenic event, in addition to its established role as the source of a highly cytopathic virus.

There is no question that this report will trigger the study of larger patient cohorts and the establishment of comprehensive databases with additional functional studies, very similar to the studies performed for many years in animal models with replication-competent murine leukemia virus.³ The need for such studies is further supported by previous reports of insertional mutations caused by HIV-derived gene vectors, revealing that the integration pattern of HIV-1 leaves enough room for potentially problematic integration events.^4,5,6

Established procedures for long-term follow-up of patients who receive lentiviral vector–transduced cells will not need to be revised based on the new findings. The key message of this study is that clonality analyses are relevant not only for recipients of cells transduced with HIV-derived vectors but also for HIV-infected patients treated with cART.² It remains to be seen whether randomly acquired secondary genetic hits trigger malignant cell expansion, on the basis of the initial driver mutation introduced by the semirandom integration event.

Along these lines, more evidence will be required to establish the mechanism by which HIV insertion into genes such as MKL2 or BACH2 drives clonal expansion, as well as to demonstrate what role this plays in the progression of HIV infection. Which additional proto-oncogenes or tumor suppressor genes might be involved in insertional skewing of the HIV-infected cell population? Is the residual viremia observed in association with these clones caused by a subpopulation of cells that escaped epigenetic silencing of the provirus? Indeed, subtle differences in the epigenetic regulation of promoter and enhancer regions may contribute to clonal progression, as shown for gammaretroviral vector insertions.⁷ Another interesting question is whether the virions produced by the progeny of the expanded clone are replication- competent, thus contributing to the clonal evolution of the virus in the course of the disease. Or are such clones undergoing additional selection for defective genomes?

Some may suggest alternative explanations for the expansion of HIV-1-infected clones harboring similar integration events, questioning the apparent hypothesis of a driver mutation introduced by insertional mutagenesis. Alternative mechanisms of clonal expansion could include the escape from negative selection (caused by the cytopathicity and immunogenicity of HIV replication), clonal expansion as the result of altered antigen stimulation, indirect effects of homeostatic proliferation, and the existence of other mutations acquired on top of the viral integration event. In cases of overt clonal dominance, deep sequencing may reveal both changes in the HIV provirus and additional cellular mutations present in the expanded clones.

The emerging databases, coupled with functional investigations, will provide important reference platforms for the long-term follow-up of patients treated with genetically modified cells, as in the case for the retroviral tagged cancer gene database obtained on the basis of experimental systems using oncogenic retroviruses.³ Collectively, these studies will contribute to an undogmatic discussion of vector safety, trigger further refinements of vector design to reduce the risk of target gene dysregulation, support an evidence-based benefit–risk assessment in the field of gene therapy, and help to identify HIV-infected individuals who might be at increased risk to acquire a secondary cancer.