Abstract Body

An accidental contagion during the production of theoretically non-infectious HIV-1 laboratory-recombinant viruses, used only for research purpose, is here described.

HIV-1 infection was diagnosed by ELISA assay, confirmed by Western Blot, and HIV-RNA/DNA test. Baseline plasma samples were used for routine pol and V3 sequences, PBMC for the whole HIV-1 genome sequencing and for in-vitro isolation. Phylogenetic analyses using Neighbor Joining (NJ) and Maximum Likelihood (ML) methods revealed the source of infection. Ethics committee approval and signed informed consent were obtained.

A lab worker resulted HIV-1 positive during a routine screening HIV test. In the 6 months preceding HIV diagnosis, he/she exclusively worked on the production of a recombinant nef-defective HIV, starting from a nef/env-defective NL4.3 vector and JRFL env-encoding plasmid. In the same laboratory other HIV-1 constructs were handled by other researchers at the same time. A thorough investigation did not evidence any laboratory accident during the whole period. At diagnosis, CD4 were 392 cells/μL, HIV-RNA 3.30 log cps/mL, and HIV-DNA 157 cps/106 PBMC. Virus was R5-tropic. After 25 months of stable HIV-1 RNA and CD4-cells, a sudden 1-log HIV-RNA increase occurred; TDF/FTC/RPV cART was started, with full viremia control.

NJ trees revealed a B-subtype virus, totally unrelated with other 629 HIV-1 clinical strains collected at the same hospital, but clustering exclusively with NL4.3 for pol sequence, and with JRFL for V3 (bootstrap: 96% and 100%, respectively). Genetic distance confirmed the homology of pol and V3 with NL4.3 and JRFL, respectively (0.002±0.002 and 0.000±0.000). Full-length viral sequence, performed in a different laboratory, confirmed the results and revealed a NL4.3/JRFL recombinant strain surprisingly expressing nef. A primary isolate obtained from his/her PBMC culture confirmed the nef-expressing NL4.3/JRFL recombinant strain. By inferring ML trees, the entire HIV-1 genome clustered again with NL4.3 (bootstrap>99%), with the exception of env (strongly linked with JRFL, bootstrap=99.7%). To date, how nef gene (absent in both vectors) entered into the recombinant virus, and mode of contagion, remain both unclear.

This clinical case highlights that in-vitro recombination procedures with per se non-infectious vectors, in a laboratory handling multiple HIV constructs, may still represent a risk of HIV contagion notwithstanding increasing biosafety efforts.