Re: Genetic engineering costs

From: Robert J. Bradbury (bradbury@aeiveos.com)
Date: Wed Apr 23 2003 - 11:12:14 MDT

  • Next message: Lee Daniel Crocker: "Prey: Am I the only one?"

    On Tue, 22 Apr 2003, Adrian Tymes, responding to my comments
    on the costs of genetic engineering wrote:

    > Hmm. Correct me if I'm wrong, please, but none of the
    > above sounds fundamentally difficult to manufacture.

    Some is and some isn't. The plastic consumables are't
    difficult to manufacture one just tends to need lots of
    them. The enzymes needed to manipulate DNA need to be
    of very high quality so those are more expensive. One
    may pay several hundred dollars for a very small tube
    of enzymes (a few ml).

    > Which means, in theory, if one had a large marget of
    > would-be genetic engineers, one could set up efficient
    > mass production of all of these components, especially
    > the consumables and enzymes. The main reason they
    > cost so much now is that the effort to set up initial
    > manufacturing has to be amortized among a limited set
    > of customers.

    In part I'd agree. There are probably only of the order
    10's of thousands of molecular biologists in the world.
    We might be up into the 100's but certainly not millions.

    > The problem is that it isn't cheap *right now*.
    > Is this correct?

    In large part -- if you had many more working in the field
    you would get more manufacturers and innovations, patent
    expiration & competition would tend to bring the costs
    down.

    You have to keep in mind how many "specialty" items various
    types of research requires. The various reagent manufacturers
    all (perhaps 3-5 significant ones) all have catalogs that
    are an 1-2 inches thick.

    > If so, then it would seem one of the more efficient
    > ways to solve the problem would be to document some
    > simple examples of the type of engineering you'd like
    > to make cheap - basically, recipes that others could
    > follow - and give them away.

    There are several major books one can by on "protocols"
    of various types of molecular biolgy, recombinant DNA
    manipulation, cell biology, etc.

    > Good point. One could, say, engineer or find a host
    > optimized for replication of virii; this would be a
    > significant cost, but one-time.

    The problem is that people are working with perhaps
    half-a-dozen viruses for gene therapies for various
    reasons -- each needs a "specialized" cell for
    production. So its hard to get the demand to develop
    such cells until we start having situations where
    we are doing 10's of thousands of gene therapies.
    That will be hard because of the problems we've
    had getting gene therapies to work and the regulatory
    overhead of setting up such trials. I think we've
    only had in the clinics or under development a
    thousand or so.

    > Typical virii are much smaller than typical cells, no?

    Yes.

    > So, would it be possible to pass the solution,
    > containing both virus and cells, through a filter with
    > small enough pores that only the virus could fit
    > through?

    Actually much of the time the viruses are "absorbed" by
    cellular processes normally involved in the transport of
    various molecules/proteins. That is why viruses tend to
    be tissue specific -- they have adapted to make use of
    naturally occuring receptors.

    You are correct in that the viruses must be filtered out
    for any therapeutic applications.

    > If limiting copies/overwrites would be difficult, then
    > perhaps it might be easier to ensure that overwrites
    > would be harmless? Say, design the virii such that,
    > if one infects an already infected cell, it will
    > merely knock out the previous copy and replace it.
    > (Which is why I use the term "overwrite".)

    Yes, this would require "site-specific integration".
    We know of only 1 virus currently that manages this.

    > True, but then you face the problem of integrating
    > these secondary nuclei into existing cells, putting
    > only one secondary nuclei into each cell instead of
    > multiple in one cell and none in its neighbors, making
    > sure they divide when they rest of the cell divides,
    > et cetera. It seems, on first glance, that by the
    > time you solve those problems you'd wind up with
    > something very much like a non-self-reproducing virus
    > anyway.

    The *difference* is that with secondary nuclei (SN) you can
    carry at least 600 genes, perhaps as many as 4000.
    With viruses you are limited to probably 200 or even
    6-50 probably in most viruses currently (all gene
    theapies I've seen I believe are only attempting
    a single gene).

    You can also make the secondary nuclei (SN) programs
    complex enough that they could detect the other SN
    and regulate their own copy number. Making them
    divide based on the normal division signals would
    not be difficult. Getting them to separate properly
    is what is difficult -- even normal eukaryotic cells
    don't manage to separate the replicated chromosomes
    properly each time.

    But most cells in the body do not replicate -- at least
    not frequently. Really only the stem cells and the
    entire Red blood cell/White blood cell systems do so.
    So one could receive substantial benefits from such
    architectures before one solved the replication
    problem.

    Robert



    This archive was generated by hypermail 2.1.5 : Wed Apr 23 2003 - 11:24:58 MDT