It has been known for many years that T-cells that are more
activated, meaning that they display markers on their surface like CD38 that
show how reactive they are to infections, and more differentiated, meaning that
they display markers such as CD45RA indicating that they have developed specialised
cell-killing ability, are much more likely to be infected with HIV than cells
which are quiescent and undifferentiated.

Dr Asier Sáez-Cirión from the Institut Pasteur found that
cells’ vulnerability to HIV infection is also dependent on cells’ metabolic (energy)
requirements; in particular, how fast they burn glucose. His team did a series
of experiments that showed that, by and large, cells that were more activated and
differentiated also burn glucose faster, which is to be expected, as they are doing
more work.

They took T-cells and managed to infect 12% of them in the
lab dish. The most differentiated cells, T-effector-memory cells, were the most
liable to infection, with 20% infected.

But, importantly, 0.9% of the least active and
differentiated cells, the T-naïve cells, were also infected. The one thing that
distinguished these low-activation but infected cells was that they had unusually
high energy demands for cells of their type.

One interesting aspect of the role of glucose metabolism in
HIV infection is that it seems to sustain ongoing
replication, rather than first infection. For HIV infection to happen at all,
cells need to express on their surfaces the receptor molecules CD4 and CCR5, which
are present in greater density on more differentiated cells. But for ongoing,
productive infection to be sustained the cell also needs to keep burning glucose
at a high rate, and indeed viral production peaked in these cells some three to
five days after initial infection rather than immediately. So drugs that
interfered with glucose metabolism may act at a different stage of the viral
timeline than entry inhibitors.

The scientists therefore cultivated infected cells with
several drugs that inhibited glucose metabolism. Such drugs could pose a
significant threat of toxicity, as they interfere with one of the most basic
and universal biological processes.

However they found that an altered version of glucose,
2-deoxyglucose or 2-DG, “decreased HIV infection of CD4 T-cells with minimal
cell toxicity.” This molecule looks like glucose to cellular receptors but can’t
be ‘cracked’ to release energy in the way regular glucose can. Its specific effect
was to reduce the ability of the cellular machinery to produce viral
components. Interestingly, 2-DG worked retrospectively – it shut down viral
replication in cells even if they had been infected eight hours previously, and
did so so efficiently that it effectively reversed potential new infections.

If HIV does manage to infect cells with low activity and
differentiation, they are more likely to turn into quiescent ‘reservoir’ cells
that are potential sources of future waves of virus rather than cells currently
producing lots of viral particles. Importantly, 2-DG reduced the number of both
reservoir cells and actively productive ones.

2-DG was considerably less toxic than other metabolic suppressors
tried by the researchers. This seemed to be because although it caused a ‘glucose famine’
that killed T-cells, it was more likely to kill HIV-infected cells than
non-infected cells, probably because their energy requirements are greater.

Dr Sáez-Cirión’s team did not just experiment on cells
infected in the lab dish; they also took T-cells from six people with HIV on antiretroviral
therapy, added a cellular activating chemical, and then added 2-DG. The glucose
analogue potently stopped the T-cells from reactivating and producing new
virus.

These experiments are early-stage preclinical
investigations, and many steps will be needed to find out if using metabolic
inhibitors is safe. But they have found
a new vulnerability of HIV to a class of molecules that has
already been investigated in cancer therapy and are simple, easy-to-make. The fact that they work on cellular rather than viral machinery
suggests that viral resistance might not be a problem, and they may hold the
potential to be used as agents that could preferentially kill HIV-infected reservoir cells.