Letters - Clinical Chemistry

Letters The Origin of Circulating Free DNA

To the Editor: Almost every report on circulating DNA identifies apoptosis or necrosis or both as the main source of free circulating DNA in serum and plasma. A hallmark of apoptosis is DNA degradation, in which chromosomal DNA is 1st cleaved into large fragments (50 –300 kb) and subsequently into multiples of nucleosomal units (180 –200 bp) (1 ). This ladder pattern is also visible after electrophoresis of circulating DNA and is frequently considered to be evidence that apoptosis may be the source of the observed DNA fragments in plasma (2, 3 ). It has been shown, however, that the characteristic ladder pattern can also be observed for actively released DNA (3 ). The contents of apoptotic cells are rapidly ingested by professional phagocytes or neighboring cells through mechanisms that are not fully understood (4 ), and the DNA is consequently completely digested by DNase II in lysosomes (1 ). Thus the possibility exists that DNA fragments released by apoptosis are removed before appearing in the circulation. If this engulfment of apoptotic bodies is impaired or cell death is increased enough to produce substantial amounts of circulating DNA, inflammation would definitely be a problem and autoimmunity would occur frequently in cancer and other conditions involving increased circulating DNA (1, 4 ). Radiotherapy, chemotherapy, and other cancer treatments cause cell death by apoptosis, and less circulating DNA is found in cancer patients after treatment than before treatment, possibly because of the inhibitory effect of treatment on the proliferation of cancer cells. Furthermore, in the early stages of cancer, when little cell death seems to occur, circulating DNA may already be present in higher than normal concentrations. As the cancer burden increases, so does the rate of cell death and the amount of proliferating cancer cells, with a concomitant increase in circulating DNA. In addition,

many cancer cells are resistant to apoptosis, and to escape the immune system proliferating cells lose the ability to become apoptotic (2, 3 ), a process that also runs counter to the notion of apoptosis as the main mechanism for generating free DNA. Necrosis, on the other hand, produces large DNA fragments. Even with no cells dying, the DNA concentration in culture medium increases in proportion to the proliferation of cultured cancer cells. Human lymphocytes have also been observed to actively release doublestranded DNA into culture medium to a certain concentration (5 ), irrespective of incubation time. A similar observation was made in experiments with frog auricles: DNA was released to the same concentrations during successive transfer of auricles to fresh medium, purified frog DNA did not inhibit release of DNA, and damaged auricles did not yield more DNA into the medium (5 ). Therefore, quantities of released DNA are similar regardless of the proportion of dying cells (5 ), with the exception applicable for cancer cells, which can release more DNA than normal cells (2 ). Evidence for preferential release of DNA was found by comparing the proportion of Alu repeat sequences to ␤-globin in serum and lymphocytes (5 ). We thus conclude that apoptosis and necrosis are not the main source of circulating DNA, although they may contribute. DNA released by living cells is a viable alternative. The mechanism of DNA clearance from plasma is poorly understood. Increased amounts of circulating DNA in the blood of patients may reflect disturbance of the equilibrium between the release of DNA by living cells and the clearance of DNA. The low concentrations of circulating DNA in healthy individuals may reflect a lower rate of DNA release by cells or a rapid removal of DNA by the optimal functioning of clearance mechanisms, and when this equilibrium is disturbed, the amount of circulating DNA increases. Circulating DNA can be found in a variety of conditions, and although these conditions are unrelated, the presence of circulating nucleic acids

is a common feature, and thus some kind of correlation ought to be found that may point to a common mechanism of origin. Although researchers have been studying the origin of circulating DNA for more than 30 years, the mechanism of release still has to be elucidated. A reasonable possibility is that more than one mechanism is involved; if so, the variables influencing the relative contributions and the interactions between the mechanisms must be understood for optimal utilization of this very valuable, minimally invasive biomarker. Thus more work needs to be done to determine the mechanism(s) of release and clearance as well as the significance of circulating DNA in the body. Grant/funding support: None declared. Financial disclosures: None declared. Acknowledgments: We thank Dr. Phiyani Lebea for proofreading the final manuscript. References 1. Nagata S, Nagase H, Kawane K, Mukae N, Fukuyama H. Degradation of chromosomal DNA during apoptosis. Cell Death Differ 2003;10:108 – 16. 2. Anker P, Mulcahy H, Chen XQ, Stroun M. Detection of circulating tumour DNA in the blood (plasma/serum) of cancer patients. Cancer Metastasis Rev 1999;18:65–73. 3. Stroun M, Lyautey J, Lederrey C, Olson-Sand A, Anker P. About the possible origin and mechanism of circulating DNA: apoptosis and active DNA release. Clin Chim Acta 2001;313:139 – 42. 4. Viorritto IC, Nikolov NP, Siegel RM. Autoimmunity versus tolerance: can dying cells tip the balance? Clin Immunol 2007;122:125–34. 5. Stroun M, Anker P. Circulating DNA in higher organisms cancer detection brings back to life an ignored phenomenon. Cell Mol Biol (Noisy-legrand) 2005;51:767–74.

Maniesh van der Vaart Piet J. Pretorius* School of Biochemistry North-West University Potchefstroom Campus Potchefstroom, South Africa * Address correspondence to this author at: School of Biochemistry, NorthWest University, Potchefstroom Campus, 11 Hoffmann St., Potchefstroom 2531, South Africa. Fax 27(0)18-299-2316; e-mail [email protected]. DOI: 10.1373/clinchem.2007.092734

Clinical Chemistry 53, No. 12, 2007

2215

2216

A Receptor-Mediated Mechanism to Support Clinical Observation of Altered Albumin Variants

To the Editor: Dolcini et al. (1 ) recently reported in this journal a novel splice-site mutation (denoted Bartin) that causes deletion of exon 11 in the human serum albumin (HSA) gene, ALB. In spite of their uncertain relevance to pathophysiology of diseases, differences in HSA sequence have interesting correlations with functional properties and stability. Bisalbuminemia (or alloalbuminemia) is a rare inherited or acquired condition characterized by the occurrence of 2 circulating components that are observed, typically, during routine clinical electrophoresis or in genetic surveys. Over the past decades several cases of genetic polymorphisms, described in peer-reviewed papers and listed in the analbuminemia register (2 ), have been characterized as representing site-specific mutations, splice-site mutations, or frame-shift mutations. The deletion reported by Dolcini et al. (1 ) would give rise to a C-terminally truncated HSA variant of 410 amino acids rather than the 585 amino acids of the wild-type protein, and would almost completely delete domain III of HSA. Furthermore, the authors discuss the impact of the C-terminal end of HSA on in vivo stability. Interestingly, several other reported mutations also generate altered C-terminal HSA variants, resulting from truncation or elongation as well as single mutations in the primary sequence. All identified HSA variants with such alteration are present in serum of heterozygous carrier individuals in the range of 2%–30% of total HSA, a fact that underscores the significance of the C-terminal domain III on stability. In light of these observations, it is noteworthy that a widely expressed receptor known as the neonatal Fc receptor (FcRn) was recently found to bind HSA (3 ). This receptor is known to be the major homeostatic regulator of the serum half-life (approximately 20 days) of IgG, as dem-

Letters

onstrated in FcRn-deficient mice that show dramatically reduced circulating concentrations of IgG (3 ). Recently, the same phenomenon was shown to be valid for the long-lived HSA molecule (4 ). An FcRn-deficient strain degraded albumin twice as fast as the wild-type strain. This finding indicates that FcRn rescues as much albumin from degradation as is produced by the liver. Interestingly, individuals with the rare human syndrome familial hypercatabolic hypoproteinemia show marked decreases of both endogenous IgG and HSA (4 ). This clinical observation was an unresolved question for decades, until Wani and colleagues (4 ) elegantly showed that the syndrome is characterized by deficient FcRn expression due to a mutation affecting ␤2-microglobulin, which together with the so-called heavy chain constitutes the heterodimeric receptor. Molecular studies reveal that IgG as well as HSA bind FcRn heavy chain in a strictly pH-dependent fashion, binding at acidic pH and releasing at physiological pH (4, 5 ). The binding sites for the 2 ligands are located distally on the ␣2-domain of the heavy chain, a position that allows simultaneous binding, and the binding of IgG does not affect binding of HSA (5 ). In the proposed regulatory mechanism (3, 4 ), FcRns, predominantly localized intracellularly in endothelial cells that cover the bloodstream, capture circulating IgG and HSA taken up by fluidphase pinocytosis, after the ligands enter acidified endosomal compartments. The lower pH herein facilitates binding to FcRn, and the interactions trigger recycling back to the cell surface and release at pH 7.2–7.4. IgG and HSA that escape receptor binding go to lysosomal degradation. An alternative cooperative pathway may be bidirectional transport between the bloodstream and the extravascular space guided by the same pH-dependent mechanism. In light of the findings of Dolcini et al. (1 ), it is noteworthy that this FcRn-mediated recycling mechanism is completely dependent on binding to domain III of HSA. Unfortunately, no cocrystal structure or site-directed

mutagenesis studies have been performed to investigate the exact interaction site on domain III, but the abnormal HSA variants described lack domain III or have structural alterations in domain III, which is absolutely crucial for the FcRn interaction. These characteristics surely will affect the pH-dependent binding to FcRn and rescue of the protein from degradation. These novel and important findings should be taken into consideration, and they may contribute to reevaluation of existing data and explain clinical observations regarding altered HSA variants.

Grant/funding support: This work was supported by grants from the Steering board for Research in Molecular Biology, Biotechnology and Bioinformatics (EMBIO) at the University of Oslo and The Norwegian Cancer Society. Financial disclosures: None declared.

References 1. Dolcini L, Caridi G, Dagnino M, Sala A, Gokce S, Sokucu S, et al. Analbuminemia produced by a novel splicing mutation. Clin Chem 2007;53: 1549 –52. 2. The Albumin Website. http://www.albumin.org (accessed August 2007). 3. Roopenian DC, Akilesh S. FcRn: the neonatal Fc receptor comes of age. Nat Rev Immunol 2007; 7:715–25. 4. Anderson CL, Chaudhury C, Kim J, Bronson CL, Wani MA, Mohanty S. Perspective—FcRn transports albumin: relevance to immunology and medicine. Trends Immunol 2006;27:343– 8. 5. Andersen JT, Dee Qian J, Sandlie I. The conserved histidine 166 residue of the human neonatal Fc receptor heavy chain is critical for the pH-dependent binding to albumin. Eur J Immunol 2006;36:3044 –51.

Jan Terje Andersen Inger Sandlie Department of Molecular Biosciences, University of Oslo Oslo, Norway Address correspondence to the authors at: Department of Molecular Biosciences, University of Oslo, P.O. Box 1041, 0316 Oslo, Norway. Fax 47 22 85 40 61; e-mail [email protected] or inger.sandlie@imbv. uio.no. DOI: 10.1373/clinchem.2007.097071


2217

Stability of Plasma Homocysteine, S-Adenosylmethionine, and SAdenosylhomocysteine in EDTA, Acidic Citrate, and Primavette™ Collection Tubes

To the Editor: After blood collection, homocysteine (Hcy) is generated in blood cells and is continuously released into the plasma. Stabilizing Hcy in blood samples requires immediate sample centrifugation and plasma separation from blood cells, sample cooling, or the use of special collection tubes (1, 2 ). To investigate the stability of Hcy and its precursors S-adenosyl-Hcy (SAH) and S-adenosyl-methionine (SAM), we collected fasting blood samples from healthy individuals (n ⫽ 8, age 25– 46 years) and renal patients (n ⫽ 9, GFR 11– 48 mL/min, 78 –90 years). Patients were recruited from the Department of Internal Medicine IV-Nephrology and Hypertension of the Saarland University Hospital. Controls were selected among hospital employees. The local ethics committee approved the usage of blood samples from patients, and all participants gave informed consent. Samples were collected into EDTA, acidic citrate, and Primavette™ tubes. The samples were incubated at 4 °C or at room temperature for 0, 2, 6, and 24 h and were centrifuged afterward. A total of 357 samples were analyzed for Hcy by HPLC (Immundiagnostik) and by fluorescence polarization immunoassay (FPIA; Abbott). Plasma SAM and SAH were determined by a modified liquid chromatography–tandem mass spectrometry method according to Gellekink et al. (3 ) in 210 samples from a representative subgroup with 10 individuals. Baseline Hcy values (EDTA) ranged from 5 to 16 ␮mol/L in healthy individuals and from 9 to 65 ␮mol/L in renal patients. Baseline plasma SAM and SAH (EDTA) were determined from 68 to 142 and 9 to 16 nmol/L in healthy individuals and from 148 to 392 and 31 to 361 nmol/L in renal patients. The individual increase of Hcy in EDTA samples at room temperature after 24 h was significantly

Fig. 1. Geometric means of plasma Hcy (n ⫽ 17) obtained by HPLC and FPIA and geometric means of plasma SAM and SAH (n ⫽ 10) in dependency on incubation time of the blood samples (0, 2, 6, and 24 h), temperature (room temperature and 4 °C), and collection tubes (F, EDTA; 䉬, acidic citrate; f, Primavette). The concentrations after 2-, 6-, and 24-h incubation were compared with the corresponding baseline values by applying the Wilcoxon test for paired samples with Bonferoni correction. *, P ⬍0.013.

higher (P ⬍0.001) in healthy individuals (8.7–23.6 ␮mol/L) than in renal patients (0.1–9.5 ␮mol/L).

Fig. 1 presents the geometric means of plasma Hcy obtained by HPLC and FPIA and plasma SAM

2218

and SAH after blood sample incubation over a period of 0, 2, 6, and 24 h at room temperature or at 4 °C. In the Primavette, Hcy obtained by HPLC decreased only slightly (⫺4.8%) after 6-h incubation at room temperature and returned to the baseline value after 24 h. At 4 °C a significant decrease of Hcy (6.9%) after 24-h incubation was found (P ⫽ 0.003). Applying FPIA for the determination of Hcy, no significant changes of Hcy compared with the baseline values were found. However, the comparison between baseline Hcy determined by FPIA and HPLC revealed that Hcy values obtained by FPIA were significantly higher (median difference 11%) than values obtained by HPLC (P ⬍0.001). In EDTA and acidic citrate plasma we observed no significant difference between Hcy obtained by FPIA and HPLC. The positive bias seen with Primavette tubes was likely due to interferences with the FPIA method by its proprietary components, which are kept secret by the manufacturer. In acidic citrate samples Hcy increased slowly at room temperature, reaching the level of significance after 6 h (FPIA) and 24 h (HPLC), respectively. At 4 °C Hcy was stable over a 24-h period. In EDTA tubes a strong increase of Hcy was observed that was markedly decreased at 4 °C. In the Primavette, SAM was stable at room temperature and at 4 °C. At room temperature SAM decreased in EDTA and to a smaller extent in acidic citrate tubes. This decrease was clearly decelerated at 4 °C. At room temperature SAH increased in all 3 collection tubes and reached the highest values in the Primavette after 24 h. The increase was attenuated at 4 °C. In Primavette and acidic citrate samples we observed 7- and 3-fold increases of plasma SAH after 24-h incubation at room temperature, whereas Hcy was stable and increased by 20%, respectively (geometric means). In EDTA samples SAH and Hcy increased 1.8-fold. Interestingly, the increase of SAH after 24 h correlated negatively with the increase of Hcy after 24 h in Primavette and acidic citrate samples

Letters

(r ⫽ ⫺0.647; P ⫽ 0.002) but not in EDTA samples (r ⫽ 0.067). Hcy is generated in erythrocytes from its precursor SAH by catalysis of SAH hydrolase. Therefore, the inhibition of SAH hydrolase activity causes an increase of SAH and no or low increase of Hcy. In contrast, in EDTA samples the increase of SAH is low because SAH is metabolized to Hcy. Furthermore, a leakage of SAH from erythrocytes into the plasma might occur because cellular SAH concentrations are approximately 10-fold higher than in plasma (4 ). In blood samples Hcy can be stabilized by the addition of an SAH hydrolase inhibitor (5 ). In conclusion, our results indicate that the stabilizing effect of Primavette and acidic citrate on Hcy is due to the inhibition of SAH hydrolase activity. The inhibition of SAH hydrolase is more efficient in Primavette than in acidic citrate tubes. However, in Primavette samples Hcy obtained by FPIA was approximately 11% higher than Hcy obtained by HPLC.

Grant/funding support: None declared. Financial disclosures: None declared.

References 1. Willems HP, den Heijer M, Lindemans J, Berenschot HW, Gerrits WB, Bos GM, et al. Measurement of total homocysteine concentrations in acidic citrate- and EDTA-containing tubes by different methods. Clin Chem 2004;50:1881–3. 2. Bisse´ E, Epting T, Ko¨gel G, Obeid R, Gempel K, Huzly D, et al. Clinical validation of a new blood collection tube for the accuracy of total homocysteine measurement by different methods. Clin Biochem 2007;40:739 – 43. 3. Gellekink H, van Oppenraaij-Emmerzaal D, van Rooij A, Struys EA, den Heijer M, Blom HJ. Stable-isotope dilution liquid chromatographyelectrospray injection tandem mass spectrometry method for fast, selective measurement of S-adenosylmethionine and S-adenosylhomocysteine in plasma. Clin Chem 2005;51:1487–92. 4. Becker A, Henry RM, Kostense PJ, Jakobs C, Teerlink T, Zweegman S, et al. Plasma homocysteine and S-adenosylmethionine in erythrocytes as determinants of carotid intima-media thickness: different effects in diabetic and non-diabetic individuals. The Hoorn study. Atherosclerosis 2003;169:323–30. 5. Martin I, Gibert MJ, Vila M, Pintos C, Obrador A, Malo O. Stabilization of blood homocysteine in an epidemiological setting. Eur J Cancer Prev 2001;10:473– 6.

Ulrich Hu¨bner1 Heike Schorr1 Rudolf Eckert2 Ju¨rgen Geisel1 Wolfgang Herrmann1* 1

Department of Clinical Chemistry and Laboratory Medicine Faculty of Medicine University of Saarland Homburg, Germany 2

Department of Geriatric Rehabilitation St. Ingbert Hospital St. Ingbert, Germany * Address correspondence to this author at: Department of Clinical Chemistry and Laboratory Medicine, Central Laboratory, University Hospital of the Saarland, Bldg. 57, D-66421 Homburg/Saar, Germany. Fax 49 68411630703; e-mail [email protected]. DOI: 10.1373/clinchem.2007.093930

From Syndrome to Spectrum: What Evolution Suggests about the Status of the Metabolic Syndrome

To the Editor: In 2005 the American Diabetes Association and European Association for the Study of Diabetes questioned the value of the metabolic syndrome as a diagnostic category (1 ). In the same year, Reaven delivered a noted “obituary” for the metabolic syndrome in this journal, while defending the insulin resistance syndrome as a pathophysiologic entity (2 ). But despite these broadsides, the concept of the metabolic syndrome is attracting continued support. PubMed lists 1775 articles with “metabolic syndrome” in the title or abstract published during 2005, 2566 during 2006, and 1689 for the first 6 months of 2007. Clearly, rumors of the syndrome’s demise have been greatly exaggerated. There are several reasons why the metabolic syndrome continues to thrive as a concept. First, there is a widespread sense that its components are sufficiently coupled mech-

2219


anistically to suggest some kind of disease entity. Second, the term and concept retain utility simply for want of a fuller description of the underlying pathology. Third, the concept is self-sustaining, in that the existing body of research attracts further research and comment. Fourth, industry promotes the syndrome in symposia and review articles, alongside related concepts such as the “cardiometabolic syndrome” and “cardiometabolic risk”. But perhaps most interestingly, the metabolic syndrome thrives because the notion of a distinct pathophysiologic entity indulges the intuitive essentialism and predilections for simplification and category formation that inhere in human cognition as evolutionary legacies (3 ). To Homo sapiens, a unified syndrome has gut appeal and even mystique—as exemplified by the allure of the phrase “syndrome X” (4 ). A closely connected nosologic issue that reflects similar inclinations is the tendency to dichotomize continuous variables into crude categories of disease and health. It is possible, however, that no consensus will ever be achieved on a definition for the metabolic syndrome as a diagnostic category, and that no disease entity will be conclusively delimited. This lack of resolution, in a different sense, is also a matter of our evolutionary legacy. The systems that transact the business of cells and regulate human physiology evolved not with rational foresight, but through bricolage, the selective but ramshackle accretion of countless innovations of contingent origin (5 ). Consequently the regulatory systems of higher organisms are highly complex, interconnected, and multifarious; riven with redundancies and lacunae; and resistant to systematization. This form of complexity is the fundamental reason why it is so difficult to define the metabolic syndrome, either as disease entity or diagnostic category. It is also why the metabolic syndrome cannot be adequately described on any single level, such as that of genetics, molecular biology, intercellular signaling, or physiolog-

ical subsystems. And hence too, the number of pathophysiologic factors implicated in the metabolic syndrome, such as hepatic, hemodynamic, endothelial, and inflammatory elements, continues to grow. Human cognition is serial and weighted toward simplicity; cell and organismal physiology is parallel and complex. The error in our conceptualization of the metabolic syndrome lies in overintegration, for in this setting it is to be expected that a spectrum or range of pathology should exist rather than a single cardinal disease pathway. Perhaps at the top level it is therefore preferable to envisage not a metabolic syndrome, but a cardiovascular, renal, and metabolic pathophysiologic matrix (CRM matrix), within which a range of related benign and disease states may develop. The characteristic range of disease states that emerge within this matrix might be described as CRM spectrum disorders. The insulin resistance syndrome of Reaven might prove to be one such disorder; alternatively, several distinct metabolic syndromes, or defined axes of variation within a pathologic continuum, might be characterized. That is a matter for research. But as research proceeds, its correct task might be framed not as one of unification, but of taxonomy, whose goal is to elaborate the nature and relationships of the full range of CRM spectrum disorders and identify appropriate interventions for each. Such an approach might also impact on the assessment of risk. One might, for instance, conceive of a CRM risk profile with an associated scoring system. The concepts of a CRM matrix and CRM spectrum disorders run counter to the instinct to overintegrate, but allow instead for a family of disorders or range of pathology to be comprehended within a single analytic perspective. These concepts also provide a more realistic framework for ongoing research. As population genetics and molecular physiology advance and a more meticulous parsing of disease states becomes possible, and as computer-

based prediction models grow in sophistication, medicine will need to move beyond the era of the allembracing syndrome and adopt a more systematic approach that better reflects the nature of biological systems and their evolutionary history.

Grant/funding support: None declared. Financial disclosures: None declared. Acknowledgments: The author is grateful to an unnamed reviewer for helpful comments on the manuscript. References 1. Kahn R, Buse J, Ferrannini E, Stern M. The metabolic syndrome: time for a critical appraisal: joint statement from the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care 2005;28: 2289 –304. 2. Reaven GM. The metabolic syndrome: requiescat in pace. Clin Chem 2005;51:931– 8. 3. Pinker S. The Blank Slate. London: Penguin, 2002. 4. Reaven GM. Syndrome X. Blood Press Suppl 1992;4:13– 6. 5. Monod, J. Chance and Necessity. London: Penguin, 1997.

Alastair Matheson 5 Highcroft Rd. London N19 3AQ, United Kingdom Fax 44-20-7272-0282 E-mail [email protected]. DOI: 10.1373/clinchem.2007.096321

Effect of Corticosteroid Therapy on Low-Molecular–Weight Protein Markers of Kidney Function

To the Editor: Serum cystatin C, ␤2-microglobulin, and ␤-trace protein are endogenous markers of glomerular filtration rate (GFR). Cystatin C, in particular, is a promising alternative to creatinine for the detection of incipient renal failure. However, corticosteroids affect the extrarenal metabolism of cystatin C, which limits the use of cystatin C as a marker of GFR in a variety of clinical settings. Low-molecular–weight (LMW) ␤-trace protein might be a useful alternative in this respect. The present study set

2220

Letters

Table 1. Factors influencing the serum concentrations of the LMW proteins.a Whole model r2 b inulin clearance, mL 䡠 min⫺1 䡠 (1.73 m2)⫺1 P value b prednisone dose, mg/m2/d P value a

1/Cystatin C, L/mg

1/ß2-microglobulin, L/mg

1/ß-Trace protein, L/mg

0.72 0.0093, 0.0081 to 0.0104 ⬍0.0001 ⫺0.0023, ⫺0.0045 to ⫺0.0001 0.0445

0.67 0.0051, 0.0040 to 0.0061 ⬍0.0001 0.0065, 0.0045 to 0.0085 ⬍0.0001

0.63 0.0110, 0.0087 to 0.0133 ⬍0.0001 0.0114, 0.0069 to 0.0159 ⬍0.0001

Multiple linear regression analysis between the reciprocal of the respective serum concentrations, inulin clearance, and corticosteroid dose.

out to compare the effect of corticosteroid therapy on the serum concentrations of cystatin C, ␤2-microglobulin, and ␤-trace protein. We studied a group of 108 children being treated or followed for malignancy (n ⫽ 41) or renal disease (n ⫽ 67). In the former group 14 patients (34%) were treated with glucocorticoids, in the latter 18 (27%). We compared single-injection inulin clearance studies in 76 patients not receiving steroids with 32 in patients receiving corticosteroid treatment (median dose 33.0 mg prednisoneequivalent per m2 body surface area per day, range 1.2–70.4). Mean (SD) age was 9.7 (5.8) years, mean (SD) GFR 92.8 (34.6) mL 䡠 min⫺1 䡠 (1.73 m2)⫺1. Patients included in the study had to be on corticosteroid therapy for at least 5 days or had corticosteroids discontinued for at least 10 days. In 16 patients, a paired analysis was performed before and during high-dose corticosteroid therapy with a median (range) dose of 36.3 mg/m2 (11.5 to 61.4 mg/m2). This analysis included 10 children in the reinduction protocol of acute lymphoblastic leukemia (ALL-9 protocol of the Dutch Childhood Oncology Group) and 6 with nephrotic syndrome. The study was approved by local and national ethics committees, and written informed consent was obtained from the patients and/or their guardians. The markers were measured by particle-enhanced immunonephelometry on a Behring Nephelometer II (DADE Behring, Marburg, Germany). As expected for endogenous GFR markers, multiple linear regression analysis between the reciprocals of the LMW protein serum concentrations, and inulin clearance showed a

highly significant positive relationship (Table 1). There was a strong positive relationship between prednisone dose and the reciprocals of both ␤2-microglobulin and ␤-trace protein. Cystatin C showed a weak but statistically significant negative relationship with inulin clearance. The paired analysis revealed that corticosteroid therapy significantly decreased mean (SD) ␤2-microglobulin from 1.67 (0.46) mg/L to 1.19 (0.59) mg/L (mean difference ⫺0.48 mg/L, 95% CI ⫺0.76 to ⫺0.21, P ⫽ 0.002 paired t-test) and ␤-trace protein from 0.80 (0.23) mg/L to 0.54 (0.15) mg/L (mean difference ⫺0.26 mg/L, 95% CI ⫺0.37 to ⫺0.15, P ⫽ 0.0001). Cystatin C concentrations, by contrast, did not change significantly: 0.862 (0.144) mg/L without vs 0.840 (0.210) mg/L with corticosteroids; P ⫽ 0.76. Mean (SD) inulin clearance tended to be higher with corticosteroids, 127 (25) mL 䡠 min⫺1 䡠 (1.73 m2)⫺1 vs 118 (25) mL 䡠 min⫺1 䡠 (1.73 m2)⫺1; P ⫽ 0.14. The ratio between a cystatin C-based GFR estimate and measured inulin clearance tended to be lower during corticosteroid treatment, but this difference was not statistically significant [mean difference ⫺0.030 (CI ⫺0.133 to 0.074); P ⫽ 0.54]. Our data confirm a dose-dependent decrease in serum ␤2-microglobulin and a weak but significant increase in cystatin C during corticosteroid treatment. Corticosteroid therapy was associated with a strong dose-dependent decrease in ␤-trace protein in both the cross-sectional and the paired analyses. Indirect evidence to support these findings comes from Po¨ge et al. (1 ), who studied steroid-treated transplant recipients and found a stronger increase in cystatin C than in ␤-trace protein with decreasing GFR, although the

opposite was shown in children who did not receive glucocorticoids (2 ). In another study, however, Risch et al. (3 ) followed 6 patients with subarachnoid hemorrhage treated with high-dose methylprednisolon, and did not find a significant change in serum ␤-trace protein concentration. Their results may have been due to unrecognized changes in GFR or leakage of ␤-trace protein through the blood-brain barrier. We previously found in paired analysis of children with steroid-sensitive nephrotic syndrome that cystatin C concentrations were unchanged on and off corticosteroids (4 ), a finding that may be attributable to a parallel increase in GFR that obscured increased cystatin C production. This theory is supported by the slightly lower ratio between estimated and measured GFR during corticosteroid therapy. Chronic corticosteroid administration increases GFR by causing vasodilation of glomerular resistance vessels. Although this mechanism is maintained in chronic renal failure, it may be blunted in acute renal injury, possibly explaining why this phenomenon was not observed in patients treated for renal allograft rejection (5 ). In conclusion, glucocorticoid therapy leads to a dose-dependent underestimation of GFR calculations based on cystatin C and overestimation of those based on ␤2-microglobulin and ␤-trace protein. In patients receiving corticosteroids, ␤-trace protein offers no advantage over cystatin C as a marker of GFR.

Grant/funding support: The study was supported by a grant from Madeleine Schickedanz Kinder-Krebsstiftung. The immunonephelometric assays were a


kind gift from Dade Behring (Marburg, Germany). Financial disclosures: A.B. received honoraria from Dade Behring (Marburg, Germany) and DAKO (Glostrup, Denmark). Acknowledgments: We are indebted to the patients and their parents who agreed to participate in the study. This study would have been impossible without the diligent work of the pediatric nursing staff of VU Medical Center performing the inulin clearance studies. We also wish to thank Mareike Reichelt and Jennifer Roos, who did the immunonephelometric measurements at Bonn University Hospital and the technicians, who did the inulin determinations at VU Medical Center. Ingrid Metgod was involved in the logistics of the study. Lyonne van Rossum is kindly acknowledged for her advice during the introduction of the single injection inulin clearance technique, and Netteke Schouten is recognized for her support and comments on the study protocol. References 1. Po¨ge U, Gerhardt TM, Stoffel-Wagner B, Palmedo H, Klehr HU, Sauerbruch T, et al. beta-Trace protein is an alternative marker for glomerular filtration rate in renal transplantation patients. Clin Chem 2005;51:1531–3. 2. Bokenkamp A, Franke I, Schlieber M, Duker G, Schmitt J, Buderus S, et al. Beta-trace protein: a marker of kidney function in children: “Original research communication-clinical investigation”. Clin Biochem 2007;40:969 –75. 3. Risch L, Saely C, Reist U, Reist K, Hefti M, Huber AR. Course of glomerular filtration rate markers in patients receiving high-dose glucocorticoids following subarachnoidal hemorrhage. Clin Chim Acta 2005;360:205–7. 4. Bokenkamp A, van Wijk JA, Lentze MJ, StoffelWagner B. Effect of corticosteroid therapy on serum cystatin C and beta2-microglobulin concentrations. Clin Chem 2002;48:1123– 6. 5. Risch L, Herklotz R, Blumberg A, Huber AR. Effects of glucocorticoid immunosuppression on serum cystatin C concentrations in renal transplant patients. Clin Chem 2001;47:2055–9.

Arend Bo¨kenkamp1* Ce`leste A.R.C. Laarman1 Katja I. Braam1 Joanna A.E. van Wijk1 Wijnanda A. Kors1 Marijke Kool1 Janneke de Valk1 Anna A. Bouman2 Marieke D. Spreeuwenberg3 Birgit Stoffel-Wagner4 Departments of 1 Pediatrics, 2 Clinical Chemistry, and

3

Clinical Epidemiology and Biostatistics, VU Medical Center Amsterdam, The Netherlands 4

Department of Clinical Biochemistry, Bonn University Medical Center Bonn, Germany * Address correspondence to this author at: Department of Pediatrics, Vrije Universiteit Medical Center, De Boelelaan 1117, NL-1081 HV Amsterdam. Fax 31-20 – 4440849; e-mail Bokenkamp@ VUmc.nl. DOI: 10.1373/clinchem.2007.094946

Ammonium 5-Bromo-7-fluorobenzo-2oxa-1,3-diazole-4-sulphonate: A New Fluorogenic Reagent for the Determination of Aminothiols by HPLC

To the Editor: The measurement of aminothiols such as cysteine (Cys), cysteinyl-glycine (Cys-Gly), homocysteine (Hcy), and glutathione (GSH), as well as the corresponding disulfides, has gained high interest within the biomedical community because such molecules are important biomarkers for a wide range of diseases. In particular, increased total plasma Hcy is now considered a risk factor for cardiovascular disorders (1 ) as well as other degenerative conditions. Although various methods for the measurement of aminothiols are available, HPLC coupled with fluorometric detection is one of the most suitable techniques for determination of minute amounts of thiols (2 ). Most methods require precolumn derivatization, and although different types of fluorogenic reagents for thiols have been proposed, the most commonly used and sensitive reagent is the commercially available ammonium 7-fluorobenzo2-oxa-1,3-diazole-4-sulfonate (SBD-F) (3 ). With this method the completion of the reaction takes a very long time and the conditions required are quite strict, i.e., derivatization must be

2221

carried out at 60 °C for a 1-h period in a moderately alkaline medium (pH ⫽ 9.5) in the presence of excess reagent. Moreover, fast oxidation or degradation of thiols, as well as of other species participating in the reaction, may take place under these conditions. To circumvent these drawbacks a number of new fluorogenic reagents possessing the benzofurazan structure have been proposed (4 ), but all require heating and/or long reaction time. All of the molecules used in aminothiol assays are characterized by the presence of 2 substituents in positions 4 and 7 of the 2,1,3-benzoxadiazole skeleton: the leaving group (halogen) and a 2nd group whose function is to increase both reactivity and solubility in water. In the search for new fluorogenic reagents with benzofurazan structure along with an additional reactivityenhancing substituent to improve reactivity, we synthesized ammonium 5-bromo-7-fluorobenzo-2-oxa-1,3diazole-4-sulfonate (SBD-BF) and characterized this reagent by means of spectroscopic and analytical data. The results of a preliminary kinetic study, carried out with Cys and Hcy nucleophiles by means of the ultraviolet-visible detection technique, indicated that SBD-BF reacts about 3 times faster at 25 °C than SBD-F at 60 °C. SBD-BF was therefore more reactive than SBD-F, and it reacted with thiols under milder conditions and during a time period of about 15 min. This result was supported by timecourse studies, carried out with reversed-phase HPLC, on the derivatization reactions of SBD-BF with the aminothiols used in the present work. Moreover, the spectra of the derivatives of Cys with SBD-F and SBD-BF, recorded with a PerkinElmer MPF 44 A spectrofluorimeter, were practically superimposable, suggesting that our new reagent SBD-BF is a suitable precolumn derivatization reagent for reversed-phase HPLC fluorometric determination of thiols. Calibration curves for the derivatives of Cys, Cys-Gly, Hcy, and GSH with SBD-BF, obtained in concentration ranges from 0.125 to 50 ␮mol/L

2222

Letters

(5 ), may be attributable to oxygen exposure. In conclusion, SBD-BF reacts with thiols under mild conditions, at 25 °C for a time period of approximately 15 min, and seems to be a very promising fluorogenic reagent for aminothiol derivatization.

Grant/funding support: Financial support was provided by Ministero dell’ Istruzione, dell’Universita` e della Ricerca (PRIN2005). Financial disclosures: None declared.

References

Fig. 1. Chromatograms of (A): derivatized mixture of Cys, Cys-Gly, Hcy, and GSH, each 5 ␮mol/L; and (B): derivatized human plasma. Conditions: Hypersil ODS RP C18 200 ⫻ 2.1 mm (i.d.) column (5 ␮m), thermostated at 23 °C; gradient elution running from 98% eluent A (0.1 mL/L aqueous trifluoroacetic acid) and 2% eluent B (acetonitrile) to 4% B (5 min), 6% B (10 min), flow rate 0.45 mL/min; total analysis time 16 min. Fluorescence intensities measured with excitation at 385 nm and emission at 515 nm.

in borate buffer, exhibited excellent linearity between peak areas and concentrations over the entire range, with r2 values of 0.9992, 0.9993, 0.9989, and 0.9991, respectively. The lower limit of detection (signal-tonoise ratio of 3:1) for the derivative of Hcy was 3 nmol/L, corresponding to 60 fmol (the upper limit was not tested). Comparative chromatographic runs of standard solution of aminothiols in borate buffer (Fig. 1A) and human plasma (Fig. 1B) after derivatization with SBD-BF demonstrate that this new fluorogenic reagent is suitable for thiol determination in biological samples. Briefly, 100 ␮L of human plasma was mixed with 10 ␮L of a solution of Tris (2-carboxyethyl)phosphine (75 mmol/L in borate buffer, pH 7.4) and allowed to react at room temperature for 15 min. Then 90 ␮L of a solution of trichloroacetic acid (100 mL/L in wa-

ter with EDTA 1 mmol/L) was then added and the sample was centrifuged for 6 min at 14.5g. An aliquot of the resulting solution (100 ␮L) was mixed with 380 ␮L of a solution of SBD-BF (25 mmol/L in 0.125 mol/L borate buffer with EDTA 5 mmol/L, pH 9.5) and 20 ␮L of NaOH 1 mol/L, incubated at room temperature for 30 min, and then acidified with 50 ␮L of HCl 1 mol/L. In a final step, 20 ␮L of the sample was injected into a Agilent 1100 HPLC system equipped with fluorescence and ultraviolet-visible detectors. The samples, as well as the starting solutions, were carefully flushed with nitrogen and protected from air until injection onto the column. Because oxygen is indeed a strong florescence quencher, some previously reported unexpected results, i.e., decrease in the fluorescence intensities

1. Mangoni AA, Jackson SHD. Homocysteine and cardiovascular disease: current evidence and future prospects. Am J Med 2002;112: 555– 64. 2. Ducros V, Demuth K, Sauvant MP, Quillard M, Causse´ E, Candito M, et al. Methods for homocysteine analysis and biological relevance of the results. J Chromatogr B 2002; 781:207–26. 3. Imai K, Toyo’oka T, Watanabe Y. A novel fluorogenic reagent for thiols: ammonium 7-fluorobenzo-2-oxa-1,3-diazole-4-sulfonate. Anal Biochem 1983;128:471–3. 4. Okabe K, Wada R, Ohno K, Uchiyama S, Santa T, Imai K. Development of hydrophilic fluorogenic derivatization reagents for thiols: 4-(N-acetyl-aminosulfonyl)-7-fluoro-2,1,3-benzoxadiazole and 4-(Ntrichloroacetylaminosulfonyl)-7-fluoro-2,1,3benzoxadiazole. J. Chromatogr A 2002;982: 111– 8. 5. Krijt J, Vackova M, Kozˇich V. Measurement of homocysteine and other aminothiols in plasma: advantages of using tris(2-carboxyethyl)phosphine as reductant compared with trin-butylphosphine. Clin Chem 2001;47:1821– 8.

Giorgio Cevasco* Anna Maria Mumot Carlo Scapolla Sergio Thea Dipartimento di Chimica e Chimica Industriale Universita` di Genova and Consiglio Nazionale delle Ricerche Genova, Italia * Address correspondence to this author at: Dipartimento di Chimica e Chimica Industriale, Consiglio Nazionale delle Ricerche, Via Dodecaneso, 31 I-16146 Genova, Italy. Fax (⫹39)010 353 6107; e-mail [email protected]. DOI: 10.1373/clinchem.2007.095406

2223


44 Single-Nucleotide Polymorphisms Expressed by Placental RNA: Assessment for Use in Noninvasive Prenatal Diagnosis of Trisomy 21

To the Editor: For noninvasive prenatal diagnosis, markers that directly reflect changes in chromosome dosage are preferred over indirect markers that are associated with epiphenomena (1, 2 ). The RNA:single-nucleotide polymorphism (SNP) allelic ratio strategy was described recently as a means to directly assess fetal chromosome dosage in maternal plasma (2 ). Quantitive comparison of the allelic expression ratios of a placentally expressed, chromosome 21– encoded gene, placentaspecific 4 (PLAC4), enabled detection in maternal plasma of the differences between 2 (normal) or 3 copies of chromosome 21 (2 ). The RNA:SNP ratio strategy is currently limited to a subset of the population with heterozygosity of the SNP used. Theoretically, an increase in population coverage can be obtained by inclusion of additional SNPs within PLAC4 or other chromosome 21– encoded transcripts with placental expression and detectability in maternal plasma (2 ). We therefore tested 44 SNPs expressed by 7 chromosome 21– encoded, placentally expressed genes (2 ), PLAC4, collagen, type VI, alpha 2 (COL6A2), collagen, type VI, alpha 1 (COL6A1), BTG family, member 3 (BTG3), ADAM metallopeptidase with thrombospondin type 1 motif, 1 (ADAMTS1), chromosome 21 open reading frame 105 (C21orf105), and amyloid beta (A4) precursor protein (peptidase nexin-II, Alzheimer disease) (APP), for their potential use in noninvasive prenatal diagnosis. All

SNP markers were tested for their presence in 1st-trimester plasma and their absence in nonpregnant women. Peripheral blood samples were collected from pregnant women attending the Prenatal Diagnostic Centre of the VU University Medical Center. All participants gave informed consent before study inclusion. The study was approved by the ethics committee at our institution. We collected EDTA blood samples between weeks 9 and 14 of pregnancy, before invasive diagnostic procedures were performed. Samples were processed and RNA extracted as described previously, with automated isolation (BioRobot MDX) (1 ). RNA extraction from PAX gene tubes was performed using the BioRobot MDX with a standardized protocol (Qiagen). For selected genes, allele frequencies were determined by cycle sequencing with a Big Dye terminator, followed by capillary electrophoresis (ABI 3100XL). Within the transcripts of the 7 genes of interest, 44 SNPs were identified (www.hapmap.org) (Table 1). Primers flanking these SNPs were designed with similar thermodynamic characteristics to permit RTPCR analysis in single runs. All primers were intron spanning, except for the primers of PLAC4. Using a sensitive, 2-step, 1-tube RT-PCR assay (Superscript II RT-PCR, Invitrogen) supplemented with 1 mol/L betaine to increase reverse transcriptase efficiency and enzyme stability (1 ), the marker set was tested in placental tissue (positive control), plasma from nonpregnant women (negative control), and pregnant women. During the initial screen, we

used 3 chorionic villus samples (weeks 8, 11, and 12) to test markers for placental expression. Screening of pregnant and control plasma was done in triplicate. To minimize the effect of biological variation of marker levels in plasma, each of the 3 screens in plasma was performed on pooled RNA fractions isolated from individual females in series of 44. In practice, this process was performed by downstream pooling of the concentrated, individual RNA fractions isolated from plasma (10 ␮L each) after automated extraction of 44 different plasma samples. For the SNPs of most use, the final screen was done by individual analysis of 6 pregnant plasma and 6 control plasma samples. With the use of RNA isolated from EDTA plasma, 5 of 44 SNP markers were detectable in maternal plasma and absent in nonpregnant plasma: rs8130833 (PLAC4), rs9977003 (PLAC4), rs11554667 (COL6A2), rs9637170 (COL6A1), and rs2187247 (C21orf105) (Table 1). In contrast, in RNA isolated from whole blood collected in Paxgene tubes, no SNP markers fulfilled the criterion of absence in nonpregnant blood. Identical analysis of hPL RNA (3 ) excluded false positivity, because in RNA recovered from whole blood in Paxgene tubes, this marker was clearly present and absent, respectively, in samples obtained from pregnant and nonpregnant females (data not shown). We conclude the following: (a) Although the PAX gene tube reagent that stabilizes RNA may be beneficial for RNA isolation from whole blood, the large contribution of intracellular RNA from maternal

Table 1. SNPs expressed by placental RNA and present in maternal plasma but not control plasma. No.

Gene

Exon

SNPa

HET FREQb

A1c

A2c

Forward primer

Reverse primer

1 2 3 4 5

PLAC4 PLAC4 COL6A2 COL6A1 C21orf105

1 1 28 23 2

rs8130833 rs9977003 rs11554667 rs9637170 rs2187247

0.448 0.339 ⬍0.1% ⬍0.1% 0.5

A A C A A

G G G C C

GGGACTCGCCGCTAGGGTGTCT AACCGTGGGACCAGTGTAGAAGAATG CACAGCAGGTGCGCAACATG CCTATCGGACCTAAAGGCTAC GCGCGCTCTCCGGGTTCCAACC

GGTGGGGATCCCTTATGCATGG GGGCAAGTGGAAAACACGCAGT AAGCGCCGGGCCTTGTG TCCAAAATCTCGCATTCGTC GGGGCCTGTCCACTTCGGTGGTAG

Selected from among 44 tested SNPs of the 7 gene selected genes. The additional SNPs tested (n ⫽ 39), and their primer sequences are available on request. Heterozygote frequencies (HET FREQ) are given for white individuals only. c A1 and A2, variant alleles. a b

2224

peripheral blood cells prevents widespread prenatal use. Prenatal PAX gene tube use appears to be limited to genes with high relative expression differences between placental tissue and maternal blood cells, such as hPL. (b) Our data confirm the utility and high expression of PLAC4 (2 ). (c) The use of SNP markers is restricted to specific exons for genes with complex transcriptional organization, such as COL6A1 and COL6A2. (d) For the transcripts of COL6A2 and COLA1 with placental specificity (for example encompassing exon 23 in COL6A1), SNPs remain to be identified for use in RNA-SNP assays. The heterozygote frequencies of rs11554667 (COL6A2) and rs9637170 (COL6A1) are ⬍0.1% in the white population we tested. (e) Alternatively, for COL6A2 and COL6A1, the combined detection of exons with specificity (exons 28 and 23 for COL6A2 and -6A1, respectively) with additional exons carrying SNPS with high heterozygosity (rs2839114, rs1053312) might yield useful combinations. (f) The predictive power of C21orf105 (1, 4 ) for

Letters

prenatal diagnosis should be retested with the RNA:SNP allelic ratio strategy by use of rs2187247. (g) Our data permit an evidence-based selection of target genes and markers to increase the population coverage of the allelic ratio strategy for noninvasive prenatal diagnosis of trisomy 21.

Grant/funding support: Part of this work was supported by the SAFE network (Project Number LSHB-CT2004-503243). Financial disclosures: None declared. Acknowledgments: We greatly appreciate the continuous support from the Department of Obstetrics and Gynaecology. References 1. Oudejans CBM, Go ATJJ, Visser A, Mulders M, Westerman BA, Blankenstein MA, et al. Detection of chromosome 21-encoded mRNA of placental origin in maternal plasma. Clin Chem 2003;49:1445–9. 2. Lo YM, Tsui NB, Chiu RW, Lau TK, Leung TN, Henung MM, et al. Plasma placental RNA allelic ratio permits noninvasive prenatal chromosomal aneuploidy detection. Nat Med 2007;13:218 –23. 3. Ng EK, Tsui NB, Lau TK, Leung TN, Chiu RW, Panesar NS, et al. mRNA of placental origin is readily detectable in maternal plasma. Proc Natl Acad Sci U S A 2003;100:4748 –53.

4. Go AT, Visser A, Mulders MA, Twisk JW, Blankenstein MA, van Vugt JM, et al. C21orf105, a chromosome 21-encoded mRNA, is not a discriminative marker gene for prediction of Down syndrome in maternal plasma. Prenat Diagn 2007;27:146 –9.

Attie T.J.I. Go1 Allerdien Visser2 Monique A.M. Mulders2 Marinus A. Blankenstein2 John M.G. van Vugt1 Cees B.M. Oudejans2* 1

Departments of Obstetrics/ Gynecology, and 2 Clinical Chemistry, VU University Medical Center Amsterdam, The Netherlands * Address correspondence to this author at: Department of Clinical Chemistry, VU University Medical Center, De Boelelaan 1117, 1081 HV Amsterdam, The Netherlands. Fax 31-20-444 3895; email [email protected]. DOI: 10.1373/clinchem.2007.093146