Diagnosing Newborns

ILLUSTRATION: FERRUCCIO SARDELLA

The first few days of a newborn’s life can be critical to his or her future well-being, for it may be the only point at which physicians have a chance to tell if something is wrong and treat it. Metabolic diseases arise from inherited errors in enzymes in any number of pathways, but the symptoms are inevitably due to the buildup of a biosynthetic precursor or product that the body cannot excrete.

In diseases such as phenylketonuria, the error is in amino acid processing. Similarly, with medium-chain acyl-CoA dehydrogenase deficiency (MCAD), the problem is one of incorrect fatty acid metabolism (see table).

Although individual metabolic diseases are rare—phenylketonuria has an incidence of 1 in 23,000 in the United States—their cumulative effect can be devastating. In a recent review, David Millington, a scientist at the Duke University Medical Center (Durham, NC), wrote that up to 1 in 3500 infants born in the United States is affected by an inborn error of metabolism (1). Unfortunately, although the benefits of newborn screening for such diseases have been witnessed time and again, technical challenges have limited their application. For example, until recently each metabolic disorder has required its own test, presenting a financial and technical drain on hospital resources. Diagnostic tests such as that for phenylketonuria (which assays the inhibition of bacterial growth due to the presence of phenylalanine) look for specific markers indirectly and do not quantitatively assay the problem-causing metabolite itself. It would be far better if a single technique could identify several compounds characteristic of the various disease states. This is where mass spectrometry (MS) has come to the fore.

Mass spectrometry
“It used to be the case that we had a particular test for a particular analyte,” recounts Blas Cerda, director of R&D for tandem MS product development at PerkinElmer LifeSciences (http://lifesciences.perkinelmer.com). “So you would develop your extraction or sample prep procedure for that particular instrument. Now, with MS, we can screen for groups of analytes.”

Newborn testing by MS starts with the same steps used in the more traditional methods. A small amount of blood (2) or urine (3) is taken from a newborn within the first 1–2 postnatal days and is absorbed onto a piece of filter paper (Figure 1). This sample is sent to the laboratory, where a small disk is punched out from the spot and loaded into a multiwell plate, where it is extracted with various solvents. For example, a sample might first be immersed for several hours in a methanol solution containing isotopically labeled standards (amino or organic acids). After drying, the sample is derivitized in acidic 1-butanol, washed in hexane, and dried before dissolution in electrospray solvent in preparation for MS.

Figure 2. Blue blood. Phenylketonuria is characterized by a dramatic increase of phenylalanine (Phe) compared with the other amino acids. Normal sample, top; phenylketonuria, bottom. (Image courtesy of PerkinElmer LifeSciences.)

Once the molecules in the sample are injected into a triple quadrupole tandem mass spectrometer, they are divided into selective characteristic masses in the first spectrometer and then fragmented through collision with argon atoms. The daughter ions are analyzed in the second spectrometer. The ion patterns are then analyzed by comparing the sample peaks with those of the internal standards using peak identification and pattern recognition software (Figure 2).

But what does this mean quantitatively?

In testing for MCAD, clinicians look at the levels of medium-chain acyl carnitines, whose normal concentration in newborn blood is 0.05–0.10 µmol/L, according to Cerda. From a 75-µL blood spot, perhaps a 3-mm (~3-µL) disk will be extracted in 100–200 µL of solvent, and of this, 10–20 µL will be injected into the tandem MS. This means that the technician will be looking for femtomole to picomole quantities of each metabolite.

To ensure testing accuracy, the Centers for Disease Control and Prevention (CDC, Atlanta) closely monitors the performance of each testing center. “The CDC has initiated what they call a proficiency testing program for the newborn screening labs,” explains Cerda. “They send out spiked blood spots that they prepare with four amino acids and four acyl carnitines. They then report the mean and standard deviations against which you can check your results.”

Although no one but the CDC and the particular laboratory knows how well the laboratory performed, the program has been instrumental in improving testing protocols. In June 2000, the CDC, the Health Resources and Services Administration, and the National Newborn Screening and Genetics Resource Center convened a workshop in San Antonio to discuss the concerns of integrating tandem MS into newborn screening programs generating a series of recommendations (4).

Multiplexing
But as clinicians continue to push the envelope on the number and types of disorders for which they want to test, scientists are being pressed to come up with new MS protocols, and much of this development is occurring at the level of sample preparation.

“The challenge is to develop a sample preparation that would be amenable to very different types of analytes,” says Cerda. “For example, you have very nonpolar fatty acids that you want to extract at the same time as the more polar amino acids.”

In the past, multiplexing has relied on the ability of technicians to develop extraction processes that maximize the number of compounds that can be analyzed in a single MS run, but those days are swiftly coming to an end. As the properties of analytes being tested continue to diversify and their number grows, other methods will have to be added to the tandem MS repertoire. One of these is liquid chromatography.

“At some point, we are going to have to put some HPLC in front of the MS,” agrees Cerda. But this will lengthen the amount of time required to produce test results and, thus, “at that point, we are going to have to start developing rapid ways of doing HPLC-MS-MS for these kinds of studies.”

But regardless of current technical limitations, no one can deny that MS has had a large impact. “Tandem MS has really revolutionized the way we do newborn screening analysis,” continues Cerda. “We can now help many more people in many more ways than we could before. We can now screen for more than 20 disorders that we couldn’t screen for before, so the impact to the community once this technology gets more established will be great.”

Beyond newborn screening
Recently, several research groups have asked whether the technologies used in newborn screening could be applied to the general population. Are there biomarkers that signal the stages of disease progression, especially the presymptomatic stage, of ailments such as rheumatoid arthritis? Or could biomarkers be used to track patients’ responses to drug treatment? And how do biomarkers change with time?

One use of MS is in the search for biomarkers that serve as warning signs against specific drug treatments. Tomika Kuhara and colleagues at Kanazawa Medical University (Ishikawa, Japan) recently developed a highly sensitive GC-MS procedure (5) for diagnosing defects in pyrimidine metabolism. This procedure might save the lives of patients who should not be treated with the anticancer drug 5-fluorouracil (5FU). Because of its pyrimidine base, 5FU can cause severe neurotoxicity and even death in patients with pyrimidine degradation deficiencies.

Pyrimidines are metabolized in four steps that are catalyzed by dihydropyrimidine dehydrogenase (DHPDH), dihydropyrimidinase (DHP), -ureidopropionase, and three aminotransferases. Typically, people with pyrimidine metabolism deficiencies lack sufficient amounts of DHPDH—a condition characterized by excessive amounts of thymine and uracil in urine—and/or DHP—manifested by large amounts of dihydrothymine and dihydrouracil in urine.

In developing the technology, the researchers took urine and treated it with urease to remove or degrade excess urea. They then spiked the treated urine with isotopically labeled uracil, orotate, and creatinine, simulating elevated levels in typical and moderate cases of DHPDH and DHP deficiencies. The extracts were then injected into a benchtop GC-MS instrument. The resulting mass chromatograms showed separate peaks for thymine, uracil, dihydrothymine, and dihydrouracil. The researchers concluded that this method allows simultaneous and quantitative determination of compounds in urine to determine whether a patient has pyrmidine metabolism deficiencies.

Moving beyond genomics
Such studies also interest the corporate world, which is paying a great deal of attention to biomarker discovery from the perspective of their use as prognostic and diagnostic markers as well as their potential as drug targets. Beyond Genomics (Waltham, MA, www.beyondgenomics.com) has developed a series of genomic, proteomic, and metabolomic technologies to identify biomarkers for external clients and for its own models of disease states. The company looks at both control and perturbed (disease-bearing or drug-treated) populations to determine the levels of everything from proteins to small-molecule metabolites.

“Clearly, the problem at the protein level is that of complexity and the variability of the physical properties of various proteins,” says Steve Naylor, chief technical officer of Beyond Genomics. “Any one peak coming off your liquid chromatogram and going directly into the mass spec might have 10, 15, or 20 peptides in there. That’s a tall order to put into the mass spectrometer.”

To address this complexity, the company adopted the 2-D MS system developed by David Clemmer at the University of Indiana, which combines ion-mobility MS with time-of-flight (TOF) MS. As Naylor describes it, “Ion-mobility mass spec is kind of like doing electrophoresis in the gas phase.” Thus, in the IMS drift tube, peptides are separated on the basis of size on a millisecond time scale and at high resolution. The peptides are then funneled into the TOF-MS, where they are separated on a mass-to-charge basis.

These experiments generate massive data sets that the company analyzes with proprietary bioinformatic software. “The trick is the pattern recognition and clustering software on the back end that allows us to generate these difference components,” continues Naylor. “We can start to tie this together and put pathways or maps into place. What pops out, almost by default, is an interesting series of biomarkers.”

Beyond Genomics is in discussion with several companies about the search for biomarkers involved in oncology and rheumatoid arthritis, and work is also well under way in search of biomarkers related to Alzheimer’s disease.

“We are working with Elan Pharmaceuticals to find a biomarker in human plasma that would go hand-in-glove with the vaccine that they’re working on,” says Naylor. “Their rationale is that if you’ve got a vaccine for Alzheimer’s, how do you know to whom to give it? You need some kind of early-onset marker or predisposition to Alzheimer’s so that they can use that in conjunction with their administration of the vaccine.”

Patient stratification
SurroMed (Mountain View, CA; www.surromed.com) uses an integrated platform of technologies, including cytometry, enzyme-linked immunosorbent assays, and MS to make strides in biomarker discovery.

“From a mass spectrometry perspective, we’re looking at proteins, peptides, and small-molecule metabolites,” says Christopher Becker, director of chemistry at SurroMed. “We try to keep an open mind as to what types of molecules will provide us with the greatest insight and most reliable markers for a disease or drug interaction.”

SurroMed, a developing therapeutics company, uses the biomarkers to stratify patient populations on the basis of the effects and efficacy of specific drugs. “By using biomarkers and understanding how different patients respond to different drugs, we think that there are drugs out there that can be rescued,” continues Becker. “We can use our technologies to do proper patient stratification, a kind of personalized medicine that is now appropriate, and get value out of these therapeutics, probably by in-licensing these drugs.”

The current mass spectrometers are more than sufficiently sensitive and robust to continue locating new low-level markers for several years to come. It is in the arenas of sample preparation and data analysis software, however, where the real progress will have to come before MS is a universal tool in clinics and biopharmaceutical firms. But these technologies are coming. Whether in the form of liquid or gas chromatographs coupled to mass spectrometers or bioinformatic algorithms—biosystematics, as Steve Naylor calls it—identifying changing biomarker expression patterns more finely and quickly, the next generation of diagnostic tools is being developed as we speak.

References

1. Millington, D. S. Am. Scientist 2002, 90, 40–47.
2. Bertholf, R. L. Gas chromatography and mass spectrometry in clinical chemistry. Encyclopedia of Analytical Chemistry; Meyers, A. B., Ed.; John Wiley & Sons: Chichester, U.K., 2000; Vol. 2, pp 1314–1336.
3. Kuhara, T. J. Chromatogr. B: Biomed. Sci. Appl. 2001, 758, 3–25.
4. Using tandem mass spectrometry for metabolic disease screening among newborns. Morbidity and Mortality Weekly Report 2001, 50, RR-3; www.cdc.gov/mmwr/PDF/rr/rr5003.pdf.
5. Kuhara, T.; Ohdoi, C.; Ohse, M. J. Chromatogr. B Biomed. Sci. Appl. 2001, 758, 61–74.

Randall C. Willis is an assistant editor of Modern Drug Discovery. Send your comments or questions regarding this article to mdd@acs.org or the Editorial Office by fax at 202-776-8166 or by post at 1155 16th Street, NW; Washington, DC 20036.