Isolation and characterization of serum procalcitonin from patients with sepsis

Peptides 22 (2001) 2099 –2103

Isolation and characterization of serum procalcitonin from patients with sepsis Wolfgang Weglo¨hnera, Joachim Struckb,*, Christina Fischer-Schulzb, Nils G. Morgenthalerb, Albrecht Ottoc, Claude Bohuond, Andreas Bergmannb b

a InVivo Diagnostica Entwicklungsgesellschaft mbH, Neuendorfstr. 25, D-16761 Hennigsdorf, Germany Research Laboratories, B.R.A.H.M.S Diagnostica GmbH, Neuendorfstr. 25, D-16761 Hennigsdorf, Germany c Max-Delbru¨ck-Centrum fu¨r Molekulare Medizin, D-13125 Berlin, Germany d Institut Gustave Roussy, F-94805 Villejuif, France

Received 22 January 2001; accepted 25 May 2001

Abstract Procalcitonin (PCT) is one of the precursors in the synthesis of calcitonin in thyroidal C-cells and other neuroendocrine cells. PCT and other calcitonin precursors are elevated in the serum of many conditions leading to systemic inflammatory response syndrome. The measurement of PCT in patients suffering from severe bacterial infections is a useful tool for the diagnosis of sepsis. Furthermore, therapeutic decisions are often based on the increase or decline of serum PCT levels. PCT was reported to have 116 amino acids. The aim of our study was the determination of the primary structure of serum PCT from septic patients. Sera containing high PCT-concentrations (⬎100 ng/ml) were collected from 22 patients with severe sepsis and were pooled for further purification (12.7 ␮g total concentration of PCT). Pooled PCT was purified on a CT 21-immunoaffinity column, further purified by reversed phase HPLC, and the resulting pure PCT was digested with endoproteinase Asp-N. N-terminal Edman sequencing showed that the first two amino acids (Ala-Pro) of the proposed pro-peptide were missing. Further analyses by MALDI-TOF mass spectroscopy resulted in a distinct mass signal of 12640 Da ⫾ 0.1%, which is in concordance with the theoretical molecular weight of the N-terminal truncated form (12628 Da). As opposed to previous suggestions, we could not detect any chemical modifications of PCT. In summary, we could demonstrate that PCT in the serum of septic patients is a peptide of only 114 amino acids, instead of the predicted 116 amino acids, lacking the N-terminal dipeptide Ala-Pro. This information on the primary structure of PCT might help in further studies on the physiological role of PCT during sepsis. © 2001 Elsevier Science Inc. All rights reserved. Keywords: Procalcitonin; PCT; Dipeptidyl peptidase IV; DPIV; Primary structure

1. Introduction Levels of procalcitonin (PCT) and other calcitonin precursors are elevated in many conditions leading to systemic inflammatory response syndrome (SIRS), like bacterial infection [2], pancreatitis [4,14], burns [12] or polytrauma [11]. Since PCT levels correlate with the severity of bacterial infection, PCT has been established over the last years as a useful marker for the diagnosis and therapy monitoring of sepsis, severe sepsis and septic shock of bacterial origin [5,8,15,16]. The pathophysiological background of elevated PCT during severe bacterial infection is still unknown. * Corresponding author. Tel.: ⫹49-3302-883-797; fax: ⫹49-3302-883621. E-mail address: [email protected] (J. Struck).

Unpublished data from our laboratory show, that the rise in PCT concentration is due to an increased production in several non thyroidal tissues during sepsis (e.g. liver). The detection of PCT in the clinical routine is usually performed by a two-site immunometric chemiluminescence assay (ILMA), which has a functional assay sensitivity of 300 pg/ml. In addition to this diagnostic value, experiments using a hamster model showed that increased PCT concentrations exacerbate mortality, whereas immunoneutralization employing an anti-calcitonin antibody increases survival [13]. Therefore it is of interest to characterise the structure of serum PCT from septic patients. This was done so far only by size exclusion chromatography and by binding of antibodies that have been raised against peptides representing the calcitonin-, katacalcin-, and N-terminal moiety of PCT

0196-9781/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved. PII: S 0 1 9 6 - 9 7 8 1 ( 0 1 ) 0 0 5 4 1 - 1

2100

W. Weglo¨hner et al. / Peptides 22 (2001) 2099 –2103

[3]. Deduced from that it was suggested that PCT is encoded by the CALC-1 gene. Its cDNA sequence predicts a primary translational product of 141 amino acids, with potential processing by the signal peptidase to a final size of 116 amino acids. Recently we could demonstrate the in vitro cleavage of the first two N-terminal amino acids of recombinant PCT 1–116 by dipeptidyl peptidase IV (DPIV/CD26) [17]. While it has been speculated that PCT might be glycosylated [9] no data exist on post translational modifications of PCT [3]. It is also unknown, whether PCT contains an internal disulphide bond as does mature calcitonin. It is important to elucidate the structure of PCT as a basis for i) understanding the function of PCT, ii) developing therapeutic intervention strategies and iii) developing novel diagnostic tests for PCT. We describe here the isolation of PCT from sera of septic patients and determination of the primary structure by mass spectrometry and N-terminal sequencing.

2. Material and methods 2.1. Serum samples Sera containing high PCT-concentrations (⬎100 ng/ml) were collected from patients (n ⫽ 22) with severe sepsis according to the American College of Chest Physicians/ Society of Critical Care Medicine Consensus Conference [1] following ethical guidelines, and all serum samples were frozen immediately at -20°C. Prior to analysis samples were pooled (final volume: 67.8 ml), filtered through a 0.45 ␮m filter and PCT immune reactivity (PCT-IR) was measured by using LUMItest PCT BRAHMS Diagnostica GmbH, Hennigsdorf, Germany). The PCT-IR concentration of the serum pool was 187 ng/ml (12.7 ␮g in total). The pool was diluted 1:1 in PBS, pH 7.8 containing 10 mM EDTA, 100 ␮M leupeptin (Sigma, Deisenhofen, Germany) and 50 ␮M amastatin (Sigma, Deisenhofen, Germany) to prevent unspecific degradation from serum proteases.

Unbound antibodies were removed by washing with 50 ml of buffer A. The PCT containing pool (135.6 ml) was passed two times through the CT 21 column at a flow rate of 1 ml/min at 4°C. After washing the column with 10 ml of buffer A, bound peptide was desorbed by elution with buffer B (100 mM acetic acid, 10% methanol) at 1 ml/min. Column efflux was continuously detected for UV absorption at 280 nm. Fractions of 0.5 ml were collected, protein containing fractions were pooled (final volume: 2.5 ml), and further purified by reversed phase HPLC. 2.3. RP-HPLC Affinity purified material was applied directly to ␮-bondapack RP-C18 column (Waters, Eschborn, Germany), eluent A was 5% acetonitrile, 20 mM ammonium acetate, and eluent B was 90% acetonitrile, 20 mM ammonium acetate. The flow rate was 1 ml/min. The column was equilibrated with eluent A. After sample injection (2.5 ml) a linear gradient from 100% A/0% B (v/v) to 55% A/45% B (v/v) within 50 min, followed by a linear gradient from 55% A/45% B (v/v) to 0% A/100% B (v/v) within 5 min was used. Column efflux was continuously detected for absorption at 214 nm. Fractions of 1 ml were collected and dried using a speed vac vacuum dryer, reconstituted in distilled water (200 ␮l), and tested for PCT immune reactivity using LUMItest PCT. 2.4. Desalting samples To remove traces of remaining ammonium acetate, lyophilised samples were dissolved in 50 ␮l of 0.1% trifluoroacetic acid and bound to 5 mg of POROS 20 R1 reversephase material (Perkin Elmer Biosystems, Langen, Germany), washed three times with 200 ␮l 0.1% trifluoroacetic acid and eluted with 50 ␮l of 60% acetonitrile containing 0.1% trifluoroacetic acid, and lyophilised before mass spectometry and N-terminal Edman sequencing. 2.5. Peptide map

2.2. Affinity chromatography 1.0 ml of calcitonin monoclonal antibody (directed against amino acids 12 to 22 of the calcitonin moiety of PCT; CT 21; 1 mg/ml, protein-A purified) in buffer A (20 mmol/l sodium phosphate, pH 7.8) was oxidised by addition of 5 mg sodium periodate (Merck, Darmstadt, Germany), incubated for 45 min at room temperature and desalted using a NAP10 column (Amersham Pharmacia Biotech, Freiburg, Germany). The desalted material (1.5 ml) was mixed with 0.5 ml of pre-washed (buffer A) carbolink gel material (Pierce, Rockford, IL, USA). The mixture was slightly shaken for 15 h at 4°C and filled into a polycarbonate column (diameter 0.5 cm, Biorad, Mu¨ nchen, Germany).

RP-HPLC-purified serum PCT (fraction 51) was taken to dryness in a speedvac centrifuge and was digested at 37°C for 16 h with 1 ␮g of sequencing-grade endoproteinase Asp-N (Roche Diagnostics, Mannheim, Germany) dissolved in 100 ␮l of 50 mM sodiumphospate buffer, pH 8.0. Asp-N peptide map was obtained by reverse-phase HPLC on a ␮RPC C2/C18 SC 2.1/10 column (Amersham Pharmacia Biotech, Freiburg, Germany) using the Smart system (Amersham Pharmacia Biotech, Freiburg, Germany). The flow rate was 100 ␮l/min at 24°C employing a gradient of acetonitrile in 0.1% trifluoroacetic acid. Peptides of interest were analysed by MALDI-TOF or loaded onto a Biobrenecoated glass fiber filter of a Procise sequencer (Type 494

W. Weglo¨ hner et al. / Peptides 22 (2001) 2099 –2103

Fig. 1. Elution profile of RP-HPLC separation of affinity purified serum PCT. 8.2 ␮g of affinity purified serum PCT were separated on a ␮-bondapack RP-C18 column (see material and methods). UV absorption profile at 214 nm is shown by the upper uninterrupted line (left scale). PCT concentration of each fraction as measured by Lumitest PCT is depicted by the rhombi in the lower line (right scale).

from Perkin Elmer Biosystems, Langen, Germany) and sequenced using standard protocols. 2.6. Mass spectrometry Peptides were analysed by matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry (MALDITOF) using a Micromass ToffSpec instrument (Micromass, UK). A saturated solution of ␣-cyano-4-hydroxycinnamic acid (Sigma, Deisenhofen, Germany) in aqueous 40% acetonitrile/0.1% trifluoracetic (v/v) acid was used as matrix. Samples dissolved in 40% acetonitrile/0.1% trifluoroacetic acid (0.8 ␮l) and matrix (1.2 ␮l) were mixed directly on target, air-dried, and analysed in the linear mode. Data of 20 –50 laser shots were collected and signal-averaged before analysis.

2101

Fig. 2. MALDI-MS spectrum of fraction 51. After RP-HPLC separation of affinity purified serum PCT-IR 20% of fraction 51 were investigated by MALDI-TOF MS. The mass peaks 4213 and 6320 resemble the higher protonated forms (M⫹3H)3⫹ and (M⫹2H)2⫹ of the 12640 Da peptide (M⫹H)⫹.

(19,4% recovery) in total. Fractions 45 to 48 contained 2489 ng in total (19,6% recovery). 3.2. Mass spectrometry and N-terminal sequencing To identify the nature of the PCT-IR the peak fractions 45, 46, 47, 50 and 51 were analysed by MALDI-TOF mass spectroscopy. While the analysis of fraction 51 resulted in a distinct mass signal of 12640 Da ⫾ 0.1% (Fig. 2), the resulting spectra of the other fractions showed broad peaks (Fig. 3) which did not allow an exact mass determination. Furthermore the spectra from the first RP-HPLC peak (fraction 46 and 47) showed also smaller peptides (Fig. 3). The peptides in fraction 47 and 51 were futher analysed by N-terminal Edman sequencing. The N-terminal amino acid sequence of the two peptides (fraction 47: FRSALE and fraction 51: FRSALESSPADPATL) showed that the first

3. Results 3.1. Purification of serum PCT Sera with high procalcitonin immune reactivity (PCTIR) were collected from patients with severe sepsis. After pooling, a concentration of 12.7 ␮g PCT in 67.8 ml was obtained. The pooled sera were diluted, and after addition of protease inhibitors applied onto a CT21-immunoaffinity column. After washing the column, bound antigen was eluted by pH change and the fractions containing PCT-IR were pooled (2.5 ml containing 8.2 ␮g PCT-IR: 65% recovery). The affinity prepared material was further purified by RP-HPLC (Fig. 1). All fractions were lyophilised and their PCT-IR content measured after reconstitution. The chromatogram is shown in Fig. 1. The main PCT-IR concentration was eluted in fraction 50 –52 containing 2465 ng

Fig. 3. MALDI-MS spectrum of fraction 47. After RP-HPLC separation of affinity purified serum PCT-IR 20% of fraction 47 were investigated by MALDI-TOF MS. The mass peaks 5419, 5694 and 6358 resemble the higher protonated forms (M⫹2H)2⫹ of the 10838, 11373 and 12713 Da peptides (M⫹H)⫹.

2102

W. Weglo¨ hner et al. / Peptides 22 (2001) 2099 –2103

Table 1 Analysis of the Asp-N digested serum PCT Peptide

PCT aa

theor. MW

determ. MW

amino acid sequence of PCT

1 2 3 4 5 6 7 8 9

3–12 13–19 20–31 32–52 53–73 74–95 96–99 100–103 104–116

1065.1 732.8 1312.6 2457.6 2319.6 2417.8 439.5 532.5 1503.6

Nd Nd 1312.2 2456.7 2315.9 2417.3 Nd Nd 1503.2

FRSALESSPA DPATLSE DEARLLLAAVQ DYVQMKASELEQEQEREGSSL DSPRSKRCGNLSTCMLGTYTQ DFNKFHTFPQTAIGVGAPGKKR DMSS DLER DHRPHVSMPQNAN

Giving theoretical peptide masses as [M ⫹ H]⫹, all cysteines in reduced form and methionines not oxidised. Amino acid sequence as deduced from the cDNA–sequence of human procalcitonin (Genbank: J00109). Amino acids determined by N-terminal sequencing of the peptides are shown in bold. nd: not determined

two amino acids (Ala-Pro) of the proposed propeptide were missing. No further amino acids, especially alanine or proline could be detected in the first two steps showing that the truncated peptide was not contaminated by the longer form. The theoretical molecular weight of the truncated form (12628 Da) is in good correlation with the experimental observed 12640 Da (⫾0.1%). 3.3. Peptide map To prove that the serum PCT consists of only 114 amino acids with no further post translational modifications we digested the remaining material from fraction 51 with the endoprotease Asp-N and analysed the resulting peptides by mass spectroscopy and/or N-terminal sequencing. The results are shown in Table 1. Five of the resulting fragments were completely sequenced, including the N- and C-terminal peptides. The remaining four peptides were only partially sequenced and therefore additionally analysed by mass spectroscopy. The largest difference between theoretical and observed molecular weight was found for peptide 5. This difference could partially be explained by the existence of a disulphide bond between the two cysteins. This disulphide bond and the resulting ring structure is a well known feature of the mature calcitonin (see discussion).

4. Discussion Our knowledge on the structure of circulating procalcitonin in septic patients is based so far mainly on hypotheses. For developing solid strategies to therapeutically neutralise endogenous PCT and to quantitate PCT, the structure of PCT must be known. Therefore we isolated PCT from a pool of sera from septic patients and determined its primary structure.

4.1. Chemical modifications of PCT It was claimed that procalcitonin is a glycoprotein [9], based on the analysis of in vitro translated rat PCT-mRNA in the presence of microsomal membranes of canine pancreas. In another study a proportion of procalcitonin, produced in a rat medullary thyroid carcinoma cell line, was found to be glycosylated [7]. More recently it was discussed, that PCT might be phosphorylated [10]. In contrast, our analyses show that human serum PCT derived from septic patients is not chemically modified whatsoever. Obviously, there are numerous differences in the experimental settings that could account for the different observations. In mature calcitonin the two highly conserved cysteine residues present in the molecule are intramolecularly linked by a disulphide bond. Our analyses indicate that such a disulphide bond may also exist in PCT. While the resolution in the mass spectroscopy was not high enough to prove this without doubt, we could demonstrate the presence of such a disulphide bond in recombinant PCT from E. coli (data not shown). This indication may argue against the speculation, that in the maturation cascade of calcitonin the disulphide bond formation occurs only after procalcitonin has been cleaved by prohormone convertases [10]. 4.2. PCT moieties and their amino acid sequence Our purification scheme involved affinity chromatography employing an antibody against the calcitonin moiety of PCT. Consequently, we can draw conclusions only on PCT species containing this moiety, and not on smaller N-terminal fragments which may as well occur to a small extent [3]. Among the PCT immunoreactive peptides it was the largest PCT species which was most abundant. The smaller peptides— due to their heterogeneity and low relative concentrations—are likely to result from proteolytic degradation of the largest species rather than from alternative biosynthetic pathways. The amino acid sequence of the PCT species corresponded perfectly to the calcitonin-I mRNA derived from

W. Weglo¨ hner et al. / Peptides 22 (2001) 2099 –2103

the CALC-1 gene. We deduced no sequence corresponding to the calcitonin-II mRNA, which could be derived from the CALC-1 gene by alternative splicing. Most notably, the largest PCT species detected contained only 114 amino acids. So far the entire literature on PCT had assumed a size of 116 amino acids, PCT 1–116” [10], a molecule with two additional amino acids at the N-terminus. This would be the expected size after the cleavage of the signal peptide. In another study we produced PCT 1–116 as a recombinant protein in E.coli, and we could show that dipeptidyl peptidase IV (DP IV/CD26) can cleave off the first two N-terminal amino acids Ala-Pro and yield PCT 3–116 [17]. This is exactly the molecule we detected in sepsis sera, suggesting that DP IV/CD26 — be it soluble or membrane bound—acts on PCT 1–116 to yield PCT 3–116 also in vivo. Recently, DP IV/CD26 was shown to process a number of chemokines including RANTES (regulated on activation normal T cell expressed and secreted), SDF-1 (stromal cell-derived factor-1), GCP-2 (granulocyte chemotactic protein-2), MDC (macrophage-derived chemokine) and others, generating naturally occurring truncated molecules with a significantly altered biological activity. Due to the small amount of PCT peptides it is not possible to investigate each serum individually. Therefore our findings do not allow the conclusion that PCT 3–116 is present in all sera. However, no higher molecular weight forms of PCT could be found, which is in favour of only one major circulating form. Furthermore the constant cooling of the used serum samples and the low concentration of soluble DPIV in sera indicates that significant cleavage of AlaPro post-collection is unlikely. Once the function of PCT in sepsis is clear, it will be interesting to study the influence of DP IV/CD26 on the physiological role of PCT. In summary, we could demonstrate that PCT in patients with sepsis is a peptide of only 114 amino acids, instead of the predicted 116 amino acids. References [1] American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med 1992;20:864 –74.

2103

[2] Assicot M, Gendrel D, Carsin H, Raymond J, Guilbaud J, Bohuon C. High serum procalcitonin concentrations in patients with sepsis and infection. Lancet 1993;341:515– 8. [3] Becker KL, Nylen ES, Cohen R, Snider RH. Calcitonin: structure, molecular biology, and actions. In: Bilezkian JP, Raisz LG, Rodan GA, editors. Principles of bone biology. San Diego: Academic Press Inc., 1996. p. 471–94. [4] Brunkhorst FM, Eberhard OK, Brunkhorst R. Early identification of biliary pancreatitis with procalcitonin. Am J Gastroenterol 1998;93: 1191–2. [5] Dandona P, Nix D, Wilson MF, Aljada A, Love J, Assicot M, Bohuon C. Procalcitonin increase after endotoxin injection in normal subjects. J Clin Endocrinol Metabol 1994;79:1605– 8. [6] De Meester I, Korom S, Van Damme J, Scharpe S. CD26, let it cut or cut it down. Immunol Today 1999;20:367–75. [7] Flanders KC, Birnbaum RS, Mahoney WC, Roos BA. Biosynthesis of an asparagine-linked oligosaccharide-containing calcitonin by a rat medullary thyroid carcinoma cell line. Biochemistry 1986;25:2995– 3000. [8] Gendrel D, Assicot M, Raymond J, Moulin F, Francoual C, Badoual J, Bohuon C. Procalcitonin as a marker for the early diagnisis of neonatal infection. J Pediatr 1996;128:570 –3. [9] Jacobs JW, Lund PK, Potts JT, Bell NH, Habener JF. Procalcitonin is a glycoprotein. J Biol Chem 1981;256:2803–7. [10] Meisner M. Biochemistry. In: Meisner M, editor. Procalcitonin. Stuttgart, New York: Thieme, 2000. p. 15– 46. [11] Mimoz O, Benoist JF, Edouard AR, Assicot M, Bohuon C, Samii K. Procalcitonin and C-reactive protein during the early posttraumatic systemic inflammatory response syndrome. Intensive Care Med 1998; 24:185– 8. [12] Nylen ES, O’Neill W, Jordan MH, Snider RH, Moore CF, Lewis M, Silva OL, Becker KL. Serum procalcitonin as an index of inhalation injury in burns. Horm Metab Res 1992;24:439 – 43. [13] Nylen E, Whang KT, Snider RH, Steinwald PM, White JC, Becker KL. Mortality is increased by procalcitonin and decreased by an antiserum reactive to procalcitonin in experimental sepsis. Crit Care Med 1998;26:1001– 6. [14] Rau B, Steinbach G, Gansauge F, Mayer JM, Grunert A, Beger HG. The potential role of procalcitonin and interleukin 8 in the prediction of infected necrosis in acute pancreatitis. Gut 1997;41:832– 40. [15] Snider RH, Nylen ES, Becker KL. Procalcitonin and its component peptides in systemic inflammation: immunochemical characterization. J Investig Med 1997;45:552– 60. [16] Whang KT, Steinwald PM, White JC, Nylen ES, Snider RH, Simon GL, Goldberg RL, Becker KL. Serum calcitonin precursors in sepsis and systemic inflammation. J Clin Endocrinol Metabol 1998;83: 3296 –301. [17] Wrenger S, Ka¨ hne T, Bohuon C, Weglo¨ hner W, Ansorge S, Reinhold D. Amino-terminal truncation of procalcitonin, a marker for systemic bacterial infections by dipeptidyl peptidase IV (DPIV). FEBS Lett 2000;466:155–9.