HIV/AIDS and the Perth Group. Was:"AIDS Treatment News"...

Todd Miller todd33 at
Sat Oct 3 08:29:45 EST 1998

fm.raaphorst at wrote:
> In article <3614289A.5A26 at>,
>   todd33 at wrote:

> > How could he claim that the antibodies from the rabbit were
> > specific for "HIV" given that the antigen could not possibly
> > have been pure "HIV" and that apparently, at that time, no one
> > had really investigated methods of trying to further purfiy
> > "HIV"?
> >
> It doesn't matter at all whether the vaccination happens with a pure
> isolate - it all depends on the selection method that's used to
> characterize the resulting antibodies.

There was no information about this in the Gallo papers, but I
agree this is one way to deal with the problem.

> Immunization with something that 'is at best half "HIV"', and the
> rest consisting of cellular material from a human cell line, will
> work just fine. Of course you'll generate a lot of B-cell lines
> producing antibodies specific for common, cellular, proteins. But
> in addition there will be B-cell lines that produce an antibody
> against HIV-proteins. Even if these proteins haven't been
> characterized by molecular cloning, you can identify the antibodies
> that recognize them. For instance, you screen the collection of
> antibodies you generated against uninfected cells and infected
> cells. Immunofluorescence will do the trick. Antibodies specific
> for common cellular antigens will bind to both infected and
> uninfected cell lines. Antibodies specific for an HIV-encoded
> protein, for instance the envelope proteins, will only bind to
> the infected cell lines.

This could be done, although some experiments would have to be
performed to show that antigens unique to the infected cell line
were truly viral, and not some endogenous protein induced by the
experiment/infection.  I guess there are also methods to use such
Western blots preparatively, so that the antibodies binding to
specifically induced antigens could be purified.  The following
link shows the proteins that are purified from 2 different lines
of infected cells (lanes B and C), compared to uninfected cells
(lane A).  I wouldn't want the task of trying to purify polyclonal
antibodies from a rabbit that were specific for "infected" vs
"uninfected" cells in this situation.  Gallo does show the
blots for uninfected vs infected, but selects only serum that
was shown to be "positive" on infected at a 1:100 dilution,
and then instead uses a 1:500 dilution for the blots with
infected and uninfected extracts.  Frankly, even in the original
_Science_ paper, it is extremely difficult to see anything,
probably because of the highly diluted antibody.

Thanks for your response Frank.

Todd Miller

More information about the Immuno mailing list