Changing lanes with capillary electrophoresis
The reagents: costs to consider
Choosing among laboratory intruments with price ranges of $50,000
to half a million dollars can be a daunting task. It’s one that clinical
laboratories face as they adopt or expand their use of capillary electrophoresis,
a technique that is swiftly supplanting traditional slab gel electrophoresis
in DNA analysis and other applications.
About 100 laboratories across the country are performing capillary
electrophoresis (CE) analysis now, and a sharp increase in that
number is predicted as molecular pathology continues to expand.
Clinical laboratories increasingly need to weigh the comparative
advantages of the many CE instruments on the market because CE has
become used for a steadily widening range of assays.
"The instruments cost a lot, but they add tremendously to the
generation and evaluation of both clinical and research data," sums
up Karin D. Berg, MD, MS, assistant professor of pathology and oncology
for the Johns Hopkins University School of Medicine and director
of Johns Hopkins’ molecular diagnostics (oncology) laboratory. "The
question is, do you need the Mercedes, or will the Hyundai do?"
Or, she suggests, perhaps the choice should be a compromise like
Dr. Berg recently discussed technical issues of capillary electrophoresis
at a meeting of the Association for Molecular Pathology. Audience
members from academia, hospitals, and industry were looking at new
instruments, with some just getting into CE and others talking about
expanding their throughput.
Her laboratory is in the latter category. "We’re running our single capillary
instrument 24/7, about 98 percent of the time, and our workload is expanding.
We’re actually bringing up two new cancer diagnostic tests we barely have the
capacity to handle." The Johns Hopkins laboratory has gone from one assay on
its machine five years ago to seven assays now—"and we could do more than
that," she adds.
Established and emerging technologies
Electrophoresis involves the migration of charged electrical species
dissolved or suspended in an electrolyte solution through which
an electric current passes. While electrophoresis has been used
for several decades, only in the last 10 years has it become the
key technology for DNA sequencing and fragment analysis. "Clearly,
the development of PCR [polymerase chain reaction] set the stage
for this being a useful tool," Dr. Berg says.
The most common methods of analyzing DNA sequence and structure
have used electrophoresis in gels to separate the fragments of DNA
according to their size. (Each kind of DNA molecule travels through
a thin slab of gel at a definite size-related speed; short molecules
of DNA travel more rapidly than long ones.) "We put the current
on the gel and it attracts the charged DNA, which migrates different
distances based on how big it is," she explains, comparing the slab
gel to a sieve.
The specific use of electrophoresis in a capillary has been more
recent. Capillary electrophoresis has been around only since the
1980s, and since 1995 its use has burgeoned in molecular diagnostics
and DNA analysis.
CE allows for the sizing of fragments of DNA that have been "PCR’d."
Like slab gel electrophoresis, it is based on sieve separation and
may also use a gel, but it involves moving a substance from one
end of a small bore capillary, between 25 and 75 µm internal diameter,
to the other end, separating mixtures into subsets in the process.
CE can be used to analyze proteins and amino acids, inorganic ions,
whole cells, chiral drugs, organic acids, dyes, surfactants, and
viral particles, in addition to nucleic acids.
For DNA analysis, CE has several large advantages over slab gels,
Dr. Berg says. "The first and biggest is that it allows you to resolve
the size of a piece of DNA down to a single base. It has very high
resolution, and that’s really helpful for sequencing. While there
are ways of getting single-base resolution out of slab gels—and
that’s how sequencing was done for a long time—it’s really
difficult to pour those gels, and it’s time-consuming." The longer
a piece of DNA gets, she notes, the less resolution there is. "But
within lengths of 50 to 500 bases you can get good resolution."
The other advantage of CE is the ability to fluorescently label
multiplexed PCR products with four or five different colors, which
eliminates the need for radioactivity. "The compounds are not dangerous,
and the technique overall is better, simpler, and easier," she says.
The high voltages of CE, usually from 10 to 30,000 volts, allow
rapid separations. "You can use a fairly high voltage to apply the
current, and that’s good because you can move things much faster.
You can’t move things as fast through solid gels because they will
melt with high voltage." Slab gels are actually liquids when heated
and solid when cooled. They cannot tolerate high temperatures without
melting (or burning). Because capillaries have a very high surface
area, the heat generated inside the capillary by the high voltages
used is effectively dissipated through its walls.
CE results are very reproducible because the method uses an internal size
standard. "On a usual gel, like a slab gel, you have to run a size standard
in a different lane than the lane containing your analyte. Which means you have
to have a straight edge across the gel and match up bands. It’s much less accurate."
Capillaries, by contrast, much more readily allow size comparisons between lanes
because the size standard is run in the capillary with the sample being sized.
Merits and drawbacks
Academic institutions are major users of CE. "A lot of universities
have core facilities that do sequencing. One facility will purchase
a machine, then perform sequencing for a lot of investigators on
campus," Dr. Berg notes. But industry also widely employs CE. Celera
Genomics and other large sequencing companies use this technology
on an enormous scale. Some of these institutions have multiple capillary
electrophoresis sequencers that can actually sequence as many as
384 samples simultaneously.
Their needs are different from those of pathologists, who indeed
do some sequencing. "But genetics, or looking at sequences, oftentimes
when it’s done in medical labs is quite expensive. There are some
institutions or laboratories that do sequence analysis. At my lab,
we can do sequencing—but we usually don’t."
Pathologists are more likely to perform fragment analysis, she
notes, which in some respects is easier to do. "There are some things
you can’t analyze by sequence analysis," Dr. Berg says. She offers
T-cell receptor gamma gene rearrangements as an example of fragment
analysis: "If in some piece of tissue there are lymphocytes that
the surgical pathologist says look atypical, but is not really sure
if they are neoplastic-reactive, we can do TCR gamma gene rearrangement
analysis and probably add some information."
"Cancer cells are derived from a single cell, and the original
cancer cell gives rise to the rest of the neoplastic cells," Dr.
Berg continues. "T lymphocytes rearrange the TCR gene during normal
development; if all the T-cells in a population have the same rearrangement
of that gene, the cells are monoclonal and most likely arose from
the same cell. These populations are more likely to be neoplastic.
If there are many different arrangements, then it’s a polyclonal
population and less likely to be neoplastic."
"We can’t call something definitely malignant or not—but
we can tell the surgical pathologist that it’s clonal or polyclonal,
and that can help in conjunction with the rest of the clinical and
pathologic data. It’s not a diagnosis; it’s not enough to stand
on its own. It’s a piece of the whole interpretation. But in the
proper clinical and pathologic context, it can be very helpful in
solidifying an opinion one way or another."
CE gives a highly visual depiction of the distinction, she adds.
"We actually PCR across a place where the gene rearranges, so if
it’s polyclonal on the capillary, there’s a normal distribution
or bell curve. If it’s a clonal arrangement, then we see a spike,"
Dr. Berg described an actual case of a patient with a hepatic
transplant, for whom surgeons found an approximately 4-cm mass in
the portal region of the transplant. "The question was whether this
was a lymphoproliferative disorder—a neoplastic disorder which
can occur after transplant and is often related to immunosuppression—or
a reactive process. The first would usually require aggressive intervention
with chemotherapy, while the other would be treated more conservatively.
"We did a PCR assay in our lab, and with the polyacrylamide gel,
if there is a nice crisp band you’d call it clonal and that would
support a diagnosis that it was neoplastic. And there was a nice
crisp band. But we ran the exact same PCR product on the capillary,
and we were able to resolve the sizes into four or five separate
peaks, which is consistent with oligoclonality, rather than a neoplastic
process. Oligoclonal populations almost always are reactive processes."
"There is no way using the gel slab that I would have seen this,"
Dr. Berg says, citing this case as a "poster child" for CE’s advantages.
"With the capillary, it was so clear what was going on. I’d say
looking at a picture of the polyacrylamide slab gel was sort of
like playing Russian roulette."
Admittedly, she says, there are analyses in which gels may produce
superior results. For example, when performing microsatellite instability
analysis using a radioactive assay, Dr. Berg says it was easy to
tell on the gel that something was amiss, maybe easier than using
"When you look at the capillary result, it’s more subtle. So it’s
not quite as good for this. But the advantage is you don’t have
to use radioactivity, and obviously in the clinical laboratory we
much prefer technologists not have to use radioactivity."
CE typically works with only one type of DNA at a time, she says.
"So if you’re trying to sequence somebody’s genetic makeup that
may or may not have a polymorphism in it, that’s easy, because all
the cells in the sample should have essentially the same sequence.
With mixed cell populations, sequencing is more difficult: It’s
really hard to sequence complex mixtures. However, you can do fragment
analysis with CE on complex mixtures using some assays."
Dr. Berg described residual disease testing as an application
of CE in engraftment studies of bone marrow transplant patients.
"We use CE to see if the graft is stable; if not, then the disease
may be recurring. We do identity testing of the patient and donor
for genetic patterns before the transplant. You might see one green
peak at a specific locus in the patient pattern and two in the donor’s
pattern. Then post-transplant, if you see two green peaks, it shows
the patient has all donor marrow."
"In this particular assay, there were actually 10 loci being PCR’d
simultaneously"—a big improvement over gel slab technology.
"With gel slab technology, "you’d actually have to run 10 separate
lanes on a gel for each patient to see the data, so CE is a lot
Other clinical applications pertain to decisions on cancer treatment.
An assay created by Kathleen Murphy, PhD, assistant director of
Dr. Berg’s laboratory, tests for the FLT3 mutation. "The mutation
of this receptor is important in acute myologenous leukemia patients,
because patients with this mutation generally do worse, and they
may also be responsive to new investigational chemotherapy agents."
Glioma subtyping is another application that has been approved for clinical
use and will be the subject of study this year at Johns Hopkins. Kimmo Hatanpaa,
MD, PhD, a neuropathology fellow who spent time doing collaborative work with
the molecular diagnostics laboratory in the last year, has created an assay
that helps to discriminate between astrocytomas and oligodendrogliomas. These
two types of tumors can be difficult to distinguish histologically, but the
distinction is important prognostically, helping clinicians decide who will
and will not benefit from aggressive treatment.
Price, throughput, and other factors
How can laboratories decide which CE instrument is for them? Of
course, a primary criterion is price, Dr. Berg says. "That’s essentially
throughput-related. But it’s also partly governed by the company
and the reputability of the company. I think for a lot of people
it’s important to think realistically about what they need in their
laboratory, and to purchase accordingly."
For example, she says, a 96-capillary instrument from one company
costs more than $300,000, but it’s really best for sequencing and
may not be needed for looking at DNA fragments. Some machines permit
an easier transition between fragment analysis and sequencing, while
others typically require capillaries of different lengths, or different
types of matrices for the capillaries, or different electrophoresis
parameters such as current or temperature.
Several CE instruments are on the market now, and Dr. Berg reviewed
the major ones, recounting her own experience with them and the
factors Johns Hopkins considers in acquiring new instruments. Enormous
instruments like the MegaBACE 4000 (Molecular Dynamics) are capable
of running 384 capillaries at once. "But that instrument is mainly
useful for big sequencing groups that are trying to just get genome
sequences," Dr. Berg says. "Those tend to be more biotechnology
applications. They may become medically important, but haven’t gotten
For the clinical laboratory market, one of the largest providers
of CE is Applied Biosystems, which provided Celera Genomics with
instruments for the work it did in sequencing the human genome,
and is actually part of the same company. "They’ve done a good job
of marketing their stuff," she says. "We’ve had the ABI 310, their
single-capillary unit, for about five years. We purchased it and
I got to do a lot of hands-on work with it. We brought four or five
assays up on that machine in our lab, so the machine has become
my friend." The instrument has limited capacity, though; because
Dr. Berg’s lab is using the instrument at 98 percent of capacity,
she is seeking an instrument with higher throughput. The 310 is
priced at about $55,000.
The ABI 3100, with 16 capillaries, is much more expensive but
has better linear range, which is important when laboratories are
trying to quantitate PCR products. While the 3100 has similar capabilities,
"its chemistries are not quite as flexible. It’s not quite the same
as the 310, so you have to be sure you can transfer your assays
over when you buy one."
The ABI 3700, priced at about $330,000, has a different optical
detection system and is typically most useful for high-volume sequencing
rather than fragment analysis. "For sizing PCR products, it’s not
quite as clear how good the 3700 is," Dr. Berg says. Most important,
it is a high-end instrument even for the clinical laboratory at
Johns Hopkins. "We will not consider purchasing it. It’s way above
what we need right now."
Beckman Coulter has two CE instruments. The P/ACE MDQ is a single-capillary
instrument used mostly for non-DNA applications. "The one we’re
actually looking at is called the CEQ 2000XL. It lists close to
$100,000 and has eight capillaries instead of 16 or one. Its sizing
is good, and it’s interesting to us because Beckman makes so many
clinical instruments. Seventy-five percent of their business is
for clinical labs, so they’re trying very hard to make them clinical-lab
friendly, with disposable modules that are prepackaged, that you
can just plug into the machine without having to do a lot of manipulations."
On the other hand, because Beckman Coulter uses a different chemistry
than ABI or MegaBACE instruments, "there can be problems for labs
that are already running assays because you have to re-label your
Molecular Dynamics manufactures three CE instruments. In the past,
"they have really targeted the big, high-throughput gene analysis
sequencing labs, and now they’re trying to focus more on the middle-throughput
labs—for example, the core facility sequencing units at smaller
academic institutions, as well as clinical laboratories."
Molecular Dynamics’ MegaBACE 500 has 48 capillaries and lists
at $175,000, but boasts the capability of running 16 capillary subunits
independent of one another. "That’s a tremendous advantage, because
you’re able to get more throughput but you don’t have to use all
those expensive capillaries if you don’t need them all."
Like Beckman Coulter, Agilent offers a one-capillary CE system
that separates a lot of different substances, not just nucleic acids.
Dr. Berg’s lab has tried out Agilent’s 2100 Bioanalyzer, which uses
a microfluidics platform with 16 capillaries on a chip. "It has
a really small footprint, and you can run DNA analysis fast," she
says. "You can do 12 samples in 30 minutes. But you don’t actually
fluorescently label the DNA, so you can really only separate one
thing at a time."
She believes the 2100 may have applications in some clinical laboratories,
but it doesn’t have the resolution that the larger capillary instruments
have. "It’s maybe between five and 10 bases. It’s not going to tell
the difference between fragments only one base different in size,
so it’s not quite enough resolution for us." The instrument is probably
more useful for industrial companies that use the "lab on a chip"
format to quantitate RNA, she adds.
Chip technology will probably influence CE’s place in the technology
picture, Dr. Berg believes. "At least right now, CE looks to be
a growing platform for DNA analysis. That could change, and there
are other approaches like real-time PCR that are very viable for
some applications. It’s always hard to predict because the field
is expanding so quickly. Some people argue that CE instruments will
continue to become smaller and be done on microfluidics platforms."
"At least for clinical applications, my understanding is that
people, especially in the larger academic laboratories, are moving
from single-capillary to multiple-capillary instruments." In biotechnology,
companies can benefit from CE depending on their mission. "If you’re
actually doing the science to identify mutations or polymorphisms
by sequence analysis, then CE is useful."
"There are a lot of different technologies out there and it’s
not clear what will sink or swim," Dr. Berg concludes. But she believes
that "thinner is better," and that capillary electrophoresis has
many advantages over other nucleic acid sizing techniques. So far,
she says, "It seems that capillaries have ’swum.’"
Anne Paxton is a writer in Seattle.