With products already on the market, biochips are likely to have an increasing
impact on genetic diagnostics, drug discovery, and basic research applications.
The notion of a cheap and reliable computer chip look-alike that performs
thousands of biological reactions is very attractive to drug developers. Because
these chips automate highly repetitive laboratory tasks by replacing cumbersome
equipment with miniaturized, microfluidic assay chemistries, they are able
to provide ultra-sensitive detection methodologies at significantly lower
costs per assay than traditional methods—and in a significantly smaller
amount of space.
At present, applications are primarily focused on the analysis of genetic
material for defects or sequence variations. Corporate interest centers around
the potential of biochips to be used either as point-of-care diagnostics or
as high-throughput screening platforms for drug lead identification. The key
challenge to making this industry as universally applicable as processor chips
in the computer industry is the development of a standardized chip platform
that can be used with a variety of "motherboard" systems to stimulate
widespread application.
Historical perspective It is important to realize that a biochip is not a single product, but
rather a family of products that form a technology platform. Many developments
over the past two decades have contributed to its evolution.
In a sense, the very concept of a biochip was made possible by the work
of Fred Sanger and Walter Gilbert, who were awarded a Nobel Prize in 1980
for their pioneering DNA sequencing approach that is widely used today. DNA
sequencing chemistry in combination with electric current, as well as micropore
agarose gels, laid the foundation for considering miniaturizing molecular
assays. Another Nobel-prize winning discovery, Kary Mullis's polymerase
chain reaction (PCR), first described in 1983, continued down this road by
allowing researchers to amplify minute amounts of DNA to quantities where
it could be detected by standard laboratory methods. A further refinement
was provided by Leroy Hood's 1986 method for fluorescence-based DNA sequencing,
which facilitated the automation of reading DNA sequence.
Further developments, such as sequencing by hybridization, gene marker
identification, and expressed sequence tags, provided the critical technological
mass to prompt corporate efforts to develop miniaturized and automated versions
of DNA sequencing and analysis to increase throughput and decrease costs.
In the early and mid-1990s, companies such as Hyseq and Affymetrix were formed
to develop DNA array technologies (see Table 1).
Current state The availability of genetic sequence information in both public and corporate
databases has gradually shifted genome-based R&D away from pure sequencing
for sequencing's sake and toward gene function−oriented studies.
It soon became apparent to everyone involved in genomics that gene sequence
data alone was of relatively little clinical use unless it was directly linked
to disease relevance. This, in turn, has driven the development of the field
of pharmacogenomics—an approach that seeks to develop drugs tailored
to individual genetic variation (see Pharmacogenomics, pp.
40−42).
In this regard, DNA-based biochips are at present used primarily for two
types of analysis. First, they have been used successfully for the detection
of mutations in specific genes as diagnostic "markers" of the
onset of a particular disease. The patient donates test tissue that is processed
on the array to detect disease-related mutations. The primary example of this
approach is the Affymetrix GeneChip. The p53 GeneChip is designed to detect
single nucleotide polymorphisms of the p53 tumor-suppressor gene; the
HIV GeneChip is designed to detect mutations in the HIV-1 protease and also
the virus's reverse transcriptase genes; and the P450 GeneChip focuses
on mutations of key liver enzymes that metabolize drugs. Affymetrix has additional
GeneChips in development, including biochips for detecting the breast cancer
gene, BRCA1, as well as identifying bacterial pathogens. Other examples
of biochips used to detect gene mutations include the HyGnostics modules made
by Hyseq.
A second application for DNA-based biochips is to detect the differences
in gene expression levels in cells that are diseased versus those that are
healthy. Understanding these differences in gene expression not only serves
as a diagnostic tool, but also provides drug makers with unique targets that
are present only in diseased cells. For example, during the process of cancer
transformation oncogenes and proto-oncogenes are activated, which never occurs
in healthy cells. Targeting these genes may lead to new therapeutic approaches.
Examples of biochips designed for gene expression profile analysis include
Affymetrix's standardized GeneChips for a variety of human, murine, and
yeast genes, as well as several custom designs for particular strategic collaborators;
and Hyseq's HyX Gene Discovery Modules for genes from tissues of the
cardiovascular and central nervous systems, or from tissues exposed to infectious
diseases.
Besides these two immediate array-based applications for this technology,
a number of companies are focusing on creating the equivalent of a wet laboratory
on a chip. One example is Caliper's LabChip, which uses microfluidics
technology to manipulate minute volumes of liquids on chips. Applications
include chip-based PCR as well as high-throughput screening assays based on
the binding of drug leads with known drug targets.
Finally, in addition to DNA and RNA-based chips, protein chips are being
developed with increasing frequency. For example, a recent report describes
the development of a quantitative immunoassay for prostate-specific membrane
antigen (PSMA) based on a protein chip and surface-enhanced laser desorption/ionization
mass spectrometry technology1.
Industry challenges A key challenge to the biochip industry is standardization. Both the assays
and the ancillary instrumentation need to be interfaced so that the data can
be easily integrated into existing equipment. This is particularly important
when genetic diagnostic applications are at stake, because important clinical
decisions are to be based on the interpretation of gene chip readouts, and
these results need to be independent of the manufacturer of the biochip.
An example of an effort to address this issue is the formation of the Genetic
Analysis Technology Consortium (GATC) by Affymetrix and Molecular Dynamics2. The aim of this group is to establish an industry standard for
the reading and analysis of many types of chips. In debating whether or not
to join this consortium, companies are forced to decide whether their market
niche will be broad use across the industry or highly customized applications
in niche areas. When the decision is for the latter, it is unlikely that they
will spend the time or money to standardize their product.
There are also important technical challenges for this industry that are
fueling a highly competitive R&D race in order to establish market dominance.
This is especially true in the "reader" technology to detect and
decipher biochip readouts. Despite efforts to standardize this technology,
novel platforms are being developed that promise higher throughput. One technology
is that appears to have particular promise is the "optical mapping"
of DNA. This method involves elongating and fixing DNA molecules onto derivatized
glass slides in order to preserve their biochemical accessibility. It has
the added feature of being able to maintain sequence order after enzymatic
digestion. This system has shown promise for high throughput and accurate
sequence analysis when integrated with appropriate detection and interpretation
software3. Whether it will emerge as the system of choice, however,
remains to be determined.
Finally, it is sometimes asked whether mass spectrometry can be part of
next-wave biochip technology. As currently conceived biochips are essentially
immobilized arrays of biomolecules, whereas mass spectrometry can determine
molecular structure from ionized samples of material. Therefore, it is difficult
to envisage a direct connection between the two, but perhaps in the future
certain aspects of biochip analysis might be performed by mass spectrometry
approaches.
Future directions Biochip development will benefit increasingly from applications developed
for other industries. For example, flame hydrolysis deposition (FHD) of glasses
has many applications in the telecommunications industry, and is now also
being applied toward the development of new biochips. A recent report describes
how FHD was used to deposit silica with different refractive indices, resulting
in microstructures that can be readily incorporated onto a chip and that integrate
both optical and fluidic circuitry on the same device4.
Biochips are also continuing to evolve as a collection of assays that provide
a technology platform. One interesting development in this regard is the recent
effort to couple so-called representational difference analysis (RDA) with
high-throughput DNA array analysis. The RDA technology allows the comparison
of cDNA from two separate tissue samples simultaneously. One application is
to compare tissue samples obtained from a metastatic form of cancer versus
a non-metastatic one in successive rounds. A "subtracted cDNA library"
is produced from this comparison which consists of the cDNA from one tissue
minus that from the other. If, for example, one wants to see which genes are
unique to the metastatic cancer cells, a high density DNA array can be built
from this subtractive library to which fluorescently labeled probes are added
to automate the detection process of the differentially expressed genes. One
study using this method compared a localized versus a metastatic form of Ewing's
sarcoma and demonstrated that 90% of the genes examined had expression levels
that differed between the two cancers by more than twofold5.
Another area of interest for future development is protein-based biochips.
These biochips could be used to array protein substrates that could then be
used for drug-lead screening or diagnostic tests. If a biosensor apparatus
is built into these biochips a further application might be to measure the
catalytic activity of various enzymes6. The ability to apply
proteins and peptides on a wide variety of chip substrates is currently an
area of intense research. The goal is to be able to control the three-dimensional
patterning of these proteins on the chips through either nano-patterning on
single layers or protein self assembly7.
The future will also see novel practical extensions of biochip applications
that enable significant advances to occur without major new technology engineering.
For example, a recent study described a novel practical system that allowed
high-throughput genotyping of single nucleotide polymorphisms (SNPs) and detection
of mutations by allele-specific extension on standard primer arrays. The assay
is simple and robust enough to enable an increase in throughput of SNP typing
in non-clinical as well as in clinical labs, with significant implications
for areas such as pharmacogenomics8.
Finally, another development of protein biochips involves the use of powerful
detection methodologies such as surface plasmon resonance (SPR). A recent
study describes the use of SPR to detect the interaction between autoantibodies
and 2-glycoprotein I (a2GPI) immobilized on protein sensor chips,
this interaction being correlated with lupus. SPR enabled the interaction
to be detected at a very low density of protein immobilization on the chip,
and this approach therefore has significant potential for the future9.
Conclusions As this fast-maturing field already boasts sales of products, biochips
are likely to have a significant business future. We can expect that advances
in microfluidic biochip technology will enable the miniaturization of devices
that will allow highly sensitive analysis of complex biological interactions
in real time. These advances promise to transform genetic diagnostics and
drug screening because of their reproducibility, low cost, and speed.
Reprinted from Nature Biotechnology 16, 981−983
(1998).
2-glycoprotein I (
