Journal of Pharmaceutical and Biomedical Analysis 111 (2015) 159–162
Contents lists available at ScienceDirect
Journal of Pharmaceutical and Biomedical Analysis journal homepage: www.elsevier.com/locate/jpba
Short communication
Metabolic responses during hemodialysis determined by quantitative 1 H NMR spectroscopy Masako Fujiwara a,∗ , Itiro Ando b , Kazuhisa Takeuchi a,c , Shiro Oguma d , Hiroshi Sato b , Hiroshi Sekino c , Keisuke Sato d , Yutaka Imai a a Department of Planning for Drug Development and Clinical Evaluation, Graduate School of Pharmaceutical Sciences, Tohoku University, 6-3, Aramaki-aza-aoba, Aobaku, Sendai 980-8578, Japan b Laboratory of Clinical Pharmacology and Therapeutics, Graduate School of Pharmaceutical Sciences, Tohoku University, 6-3, Aramaki-aza-aoba, Aobaku, Sendai 980-8578, Japan c Koujinkai Central Clinic, 2-1-6, Tsutsujigaoka, Miyaginoku, Sendai 983-0852, Japan d Koujinkai Nagamachi Clinic, 2-8-2, Happonmatsu, Taihakuku, Sendai 982-0001, Japan
a r t i c l e
i n f o
Article history: Received 26 December 2014 Received in revised form 26 March 2015 Accepted 29 March 2015 Available online 7 April 2015 Keywords: Hemodialysis Lactate Metabolic response 1 H NMR spectroscopy Pyruvate
a b s t r a c t A large proportion of patients with end-stage renal disease have lifelong hemodialysis (HD) treatment. HD rapidly and indiscriminately removes necessary small metabolites together with uremic toxins from plasma into dialysate. To investigate metabolic responses to HD, we determined the levels of metabolites through time-course monitoring of 1 H NMR spectroscopy of dialysate during HD. The dialysate sample is stable for analysis because it contains only small metabolites without proteins. It was collected noninvasively from 9 HD patients with chronic glomerular nephropathy, at 6 time points during 4 h of HD in 5 sessions. Creatinine, alanine, lactate, pyruvate and valine were simultaneously quantified on a onedimensional single-pulse spectrum with a single standard compound. The concentration of creatinine exhibited monotonous decay with time, while that of valine decreased slowly and then maintained its levels throughout an HD. Lactate, alanine and pyruvate increased at 2–3 h after the initiation of HD. They exhibited remarkable responses to HD with production from the body. The time-course of change in the 4 metabolites of lactate, pyruvate, alanine, and valine had reproducible behavior unique to each patient during the HD. This finding may be applied to distinguish metabolic status in HD patients. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Most patients with end-stage renal disease have lifelong hemodialysis (HD) therapy in Japan. In HD, extracorporeal circulation is applied to remove accumulated uremic toxins, electrolyte and fluid in plasma into the dialysate through a membrane during a 4 h HD session 3 times a week. The rate of exclusion in 4 h is approximately 10 times faster than that of previous accumulation for 2 or 3 days. In the therapy, not only toxins but also all the small metabolites of nutrients and physiologically necessary bioactive molecules are filtered indiscriminately [1]. The rapid removal of various metabolites in repeated therapies may cause profound metabolic stress to patients. In our previous study on 1 H NMR-based metabolomics of plasma from HD patients, post-HD plasma levels of lactate
∗ Corresponding author. Tel.: +81 22 795 4528; fax: +81 22 795 4532. E-mail address:
[email protected] (M. Fujiwara). http://dx.doi.org/10.1016/j.jpba.2015.03.035 0731-7085/© 2015 Elsevier B.V. All rights reserved.
increased compared with pre-HD plasma levels [2]. The increments of lactate suggested production from the body during the HD. To date, it has never been monitored with time how metabolites respond in the intermediate time of HD. In the present study, we investigated the time-course behavior of important metabolites which play a key-role in metabolic pathways related to lactate by 1 H NMR of dialysate through an HD. The 1 H NMR technique for biofluid needs minimum sample preparation [2,3] unlike GC/MS or LC/MS methods. On a spectrum all the protons of various kinds of metabolites are detected with the same sensitivity. Spent hemodialysate consists of only small metabolites filtered proteins in plasma. We demonstrated in our previous study [3] that 1 H NMR of dialysate ensured quantification of the metabolites with enough sensitivity. A dailysate can be preserved stably at −25 ◦ C for a long time because it is free from bacteria and enzymes. In the case of HD patients, blood collection is limited because they abolish kidney functions including hematogenesis and are mostly anemic. Dialysate collection was done non-invasively and frequently. Recently, we have verified dialysate to be an excellent surrogate for plasma in quantifying
160
M. Fujiwara et al. / Journal of Pharmaceutical and Biomedical Analysis 111 (2015) 159–162
small metabolites during HD by NMR analysis [3]. The quantitative 1 H NMR spectroscopy of dialysate offers a novel and simple tool for clinical investigations on HD patients.
decays were processed using the software ‘Aline2 for Windows’ (V. 6.0, JEOL Ltd.).
2.4. Quantification of metabolites
2. Materials and methods
The signals of creatinine (CH3 , 3.05 ppm), lactate (CH3 , 1.33 ppm), pyruvate (CH3 , 2.38 ppm), alanine (CH3 , 1.47 ppm) and valine (2CH3 , 0.98, 1.04 ppm) for dialysate were assigned on the spectra referenced to TSP as a chemical shift standard (0.00 ppm), and quantified by integrations of these signal-areas referenced to TSP (3CH3 ) as a quantitation standard as well. Glucose, acetate, and citrate were not analyzed here because they are composed in the original dialysate. To obtain the plasma levels, the concentrations in dialysate were converted with the ratios derived in our previous study [3]. The levels can be roughly estimated by multiplication by a factor of 3.5, in common with metabolites.
2.1. Patients The study protocol was approved by the Ethical Committee of Koujinkai Hospital (Koujinkai Hemodialysis Clinic, Miyagi, Japan). Nine patients with maintenance HD were recruited and written informed consent was obtained from them. They were selected randomly from chronic glomerular nephropathy (CGN) patients undergoing stable HD without disorders of energy metabolisms such as diabetes mellitus. All patients were virtually anuric and dialyzed with high-performance polysulfone membranes. The flow rates of blood and dialysate were 200, and 500 mL/min, respectively. The total dialysate used in 4 h HD was 120 L. 2.2. Sample collection and preparation From 2010 to 2012, dialysate were collected from the 9 patients during HD every 2 months for 10 months for each patient. Dialysate was collected at 15, 30 min, 1, 2, 3 and 4 (final) h after the initiation of HD in every session. Dialysates were sampled in standard plastic tubes and stored at −25 ◦ C. They were thawed on ice immediately before use. For NMR measurement, 50 l of deuterium oxide to provide a field-frequency lock and sodium 3-(trimethylsilyl) propionate 2, 2, 3, 3-d4 (TSP) as an internal chemical shift and an intensity reference were added to the sample of 550 l. Ten minutes after centrifugation of the sample at 13,000 rpm, we placed the supernatant into a 5 mm NMR tube.
creanine
2.3. Quantitative NMR spectroscopy
15m
lactate
citrate
pyruvate
alanine
4h AcAc
3h
valine
acetate
2h lys+arg
1h
3HB
glu
30m gln
leu
ppm
Single-pulse 1 H NMR spectra were recorded at 25 ◦ C internal probe temperature using a 600 MHz NMR spectrometer (ECA600, JEOL Ltd., Tokyo, Japan). The water signal was suppressed by a presaturation pulse sequence, and the flip angle was set to 90◦ . Sixty-four scans were collected for 64 K data points with a spectrum-width 9 kHz. The repetition times were set to 42.4 s for dialysate, which ensured larger than 5 × T1 [3]. Free induction
Fig. 1. Partial 1 H NMR spectra of dialysate during HD. Time-series of dialysate spectra measured by 1D single-pulse NMR in a typical patient were aligned where TSP peak heights were determined. The dialysate buffer used was citrate, thus citrate peaks were revealed to be constant during the 4-h session. (Conversely, when an acetate-containing buffer was used, acetate peaks were constant in the spectra.) AcAc: acetoacetate, 3HB: 3-hydroxybutyrate, gln: glutamine, glu: glutamate, lys: lysine, arg: arginine, leu: leucine.
Table 1 Baseline characters of study patients (n = 9). Patient no.
1
2
3
4
5
6
7
8
9
Mean ± SD
Sex Age (yr) Weight (kg)
F 70 46.1 ± 0.6
F 61 52.6 ± 6.9
F 58 38.1 ± 0.1
M 73 44.3 ± 0.2
M 55 58.4 ± 1.1
F 69 51.7 ± 0.3
M 64 69.7 ± 1.1
F 60 55.3 ± 0.2
F 75 36.4 ± 0.7
65 ± 7 50.3 ± 6.6
HD dutration (yr) dialysate Kt/V
31 A/C 1.59 ± 0.0
19 C 1.72 ± 0.1
31 C 1.9 ± 0.1
10 C 1.5 ± 0.1
21 C/A 1.4 ± 0.1
33 C 1.4 ± 0.2
15 C 1.56 ± 0.1
31 A 1.9 ± 0.1
3 A 2.0 ± 0.1
22 ± 11
Creatinine (mg/dL) Glucose (mg/dL) Hemoglobin (g/dL) Hematocrit (%) Albumin (g/dL) Basal disease
9.8 ± 0.2 80.8 ± 1.1 11.3 ± 0.5 33.8 ± 1.6 3.6 ± 0.2
10 ± 0.9 102 ± 8.8 11 ± 0.7 32.7 ± 1.9 3.9 ± 0.1
9.7 ± 0.4 99.8 ± 15 9.6 ± 0.9 28.5 ± 3.0 3.9 ± 0.1 TP
11.7 ± 0.2 134 ± 39.7 11.1 ± 0.3 32.5 ± 1.1 3.9 ± 0.16
12.7 ± 0.5 96.6 ± 8.9 11 ± 0.3 32.4 ± 0.9 4.0 ± 0.1
9.8 ± 0.6 108 ± 11.8 10.2 ± 0.6 30.6 ± 2.0 3.7 ± 0.2
17.2 ± 0.2 132± 18.6 11.1± 0.5 33.6± 1.2 4.01± 0.1 NS
10 ± 0.4 92.6 ± 8.4 11.5 ± 0.3 33.4 ± 1.0 3.9 ± 0.1 IgA
8.5 ± 0.9 86.8 ± 9.4 11.2 ± 0.6 32.9 ± 1.8 3.5 ± 0.1 SLE
11.0 ± 2.5 103.6 ± 17 10.9 ± 0.6 32.3 ± 1.6 3.8 ± 0.2
1.7 ± 0.2
Clinical data are expressed as an average of 5 time measurements in each patient and values are mean ± S.D. In the right hand row, mean and S.D. of values in 9 patients are demonstrated. In HD, heparin was used as the blood anticoagulant. During the HD session, supplemental nutrient was not added. The dialysate buffer contained glucose (150 mg/dL), sodium (140 mEq), potassium (2 mEq), calcium (2.5–3 mEq), and citrate or acetate. A; bicarbonate dialysis buffers (8 mM acetate), C: dialysis buffer (0.7 mM citrate), A/C: 3 times of A and 2 times of C, C/A: 4 times of A and once C. Kt/V: dialysis adequacy values estimated with the single-pool model described by Daugirdas [1]. Nine patients are CGN, which is a kind of syndrome. Some of them had basal diseases which are shown in the bottom line in the table. TP: toxemia of pregnancy, NS: nephrosclerosis, IgA: IgA nephropathy, SLE: systemic lupus erythematosus.
M. Fujiwara et al. / Journal of Pharmaceutical and Biomedical Analysis 111 (2015) 159–162
161
3. Results and discussion
3.3. Estimation of metabolites loss
3.1. Characters of 9 patients
Quantifications of metabolite at 6 time points have been performed. The time course of change in dialysate levels of creatinine, lactate, pyruvate, alanine and valine in 9 patients are illustrated in Fig. 2. We calculated the amount of losses of the metabolites during the HD, and summarized in Table 2. The average loss over 9 patients of lactate, pyruvate, alanine and valine were 32 ± 11, 2 ± 1, 7 ± 2, 4 ± 1 mM (mean ± S.D.), respectively. The amounts of the 2 amino acids (alanine: 0.6 g, valine: 0.5 g) were well coincident with the values in the past work [4]. To our knowledge, no data are available on the loss of pyruvate and lactate during HD. An total loss average of 34 ± 12 mM of lactate and pyruvate per 4 h was calculated. They may be mediated by energy losses producing ATP through TCAcycle [5], which we will discuss elsewhere.
Every patient underwent biochemical examination of plasma every 2 weeks before HD under usual conditions, without fasting. Of these data, adjacent ones to each sample collection of 5 times for NMR measurement are summarized in Table 1, together with the patients’ characteristics, HD conditions, and basal disease from the medical records.
3.2. Time-series of 1 H-NMR spectra Examples of 1 H-NMR spectra in the low-frequency (ı 0.75–3.15 ppm) regions of dialysate from one patient during HD are illustrated in Fig. 1. Creatinine peaks diminished monotonously [1]. Other metabolites, amino acids (alanine, valine) and organic acids (lactate, pyruvate) did not decrease with creatinine, lactate and pyruvate increased extensively at around 2 h after the initiation of HD.
3.4. Time course of change in metabolites concentration As shown in Fig. 2, creatinine levels in dialysate decreased with time monotonously, which indicated that creatinine had little
Fig. 2. Dialysate concentrations of metabolites during HD. From top to bottom of the figure, time-courses of changes in metabolite concentrations in dialysate are depicted in patients no. 1–9. In each graph, the X-, Y-axes indicate hours during HD and dialysate concentration (mM), respectively. The left row shows creatinine. The central row shows lactate (solid lines, left axes) and pyruvate (dotted lines, right axes). The right hand row shows alanine (solid lines) and valine (dotted lines). In each figure, dots and vertical bars indicate mean values and standard error (S.E.) of 5 HD sessions, respectively.
162
M. Fujiwara et al. / Journal of Pharmaceutical and Biomedical Analysis 111 (2015) 159–162
Table 2 Average loss of metabolites per HD over 5 HD treatments. Patient no. Lac Pyr Ala Val
1 48 3.4 6.7 3.4
2 43 3.4 7.9 3.4
3 28 2.3 4.5 2.6
4 31 2.3 6.7 3.4
5 29 2.3 6.7 4.3
6 (mM)
7
24 1.1 5.6 5.1
44 3.4 10 6.0
8
9
mean
SD
mean
SD (g)
24 2.3 6.7 5.1
14 1.1 3.4 2.6
32 2.3 6.7 4.3
11 1.1 2.2 0.9
2.9 0.2 0.6 0.5
1.0 0.1 0.2 0.1
For calculation of the total loss amount of above 4 metabolites exclusion into dialysate, the average levels of them in dialysate during 4 h were assumed to be the average of the 2 representative points, i.e., at 1 h and 3 h from the initiation of HD, and then the average levels were multiplied by the total dialysate amount (120 L). In the right hand row, mean and SD of each metabolite are converted to units of (g), using molecular weight lac (90), pyr (88), ala (89), val (117), respectively.
secretion from the body during the sessions [1]. Although the levels were different, their time-course changes were similar among all the patients. The error-bars over 5 repetitions for every point were small, which indicated a stable HD. On the other hand, lactate, pyruvate, and alanine exhibited increments in the middle time of HD. The increments indicated enhanced appearances from the body, which overcame continuous filtering from blood. In metabolic pathways, lactate and alanine are coupled directly with pyruvate. The lactate vs. pyruvate (L/P) ratio is used as one of the clinical markers of hypoxia [6]. The normal levels of the ratio is 10, which is coincident with our results. Lactate behaved in consonance with pyruvate in all patients, where the L/P ratio had no particular value throughout the HD. Judging from the value, the patients had no sign of hypoxia when lactate levels increased. Valine concentrations decreased slowly with time and then maintained the levels during the sessions in most patients. The levels of alanine, valine and also alanine/valine ratios were reproducible over 5 time’s sessions and unique to each patient. From Fig. 2, average lactate levels in dialysate were between 0.08 and 0.5 mM, thus lactate plasma levels were estimated to be 0.28–1.75 mM (Section 2.4). The range is within plasma lactate levels in healthy subjects and also HD patients at baseline (0.25–1.8 mM), not reached the alarm levels such as lactate acidosis (∼5 mM) [7]. The other 3 metabolites in plasma levels, alanine, pyruvate and valine, were maintained within normal levels [8]. In these responses, the levels of 4 metabolites were within the physiological regulations. The time course of changes in metabolites and their levels had unique features for each metabolite and each patient. The CGN patient had a unique and the reproducible response in 5 sessions. The responses might be detected only under the stress of HD and hidden under usual conditions untreated. On the lowest column in Table 1, several basal diseases leading to CGN being diagnosed were indicated; these might be one of the factors influencing the responses. Patient no. 9 had a particularly blunt response in metabolite levels; her original disease was systemic lupus eryhtematosus. This is only one case; however, there may
be a potential for diagnosis metabolic status of HD patients by our analysis. 4. Conclusions This is the first observation of the time-course metabolic responses during HD by quantitative 1 H-NMR spectroscopy. We demonstrated that for 4 metabolites, lactate, pyruvate, alanine and valine, a homeostatic reaction with productions from the body occurs during HD, unique to each patient. Using dialysate, noninvasive and quantitative-NMR detection of plasma metabolites was successfully done. The present study will provide future scope for personalized therapy in HD patients with complex backgrounds and etiologies. Acknowledgements This work was in part supported by Grants-in-Aids from the Ministry of Education, Culture, Sports, Science and Technology of Japan (24590042 and 24590178) and from Miyagi Kidney Foundation 2012. We thank the staff of Koujinkai Nagamachi Clinic, who collected samples from patients and gave us information about the patients. References [1] J.T. Daugirdas, Second generation logalithmic estimations of single-pool variable volume Kt/V: an analysis of error, J. Am. Soc. Nephrol. 4 (1993) 1205–1213. [2] M. Fujiwara, T. Kobayashi, T. Jomori, Y. Maruyama, Y. Oka, H. Sekino, Y. Imai, K. Takeuchi, Pattern recognition analysis for 1 H NMR spectra of plasma from hemodialysis patients, Anal. Bioanal. Chem. 394 (2009) 1655–1660. [3] I. Ando, K. Takeuchi, S. Oguma, H. Sato, H. Sekino, Y. Imai, M. Fujiwara, 1 H NMR spectroscopic quantification of plasma metabolites in dialysate during hemodialysis, Magn. Reson. Med. Sci. 12 (2013) 129–135. [4] T.A. Ikizler, P.J. Flakoll, R.A. Parker, R.M. Hakim, Amino acid and albumin losses during hemodialysis, Kidney Int. 46 (1994) 830–837. [5] M. Erecinska, D.F. Wilson, Regulation of cellular energy metabolism, J. Membr. Biol. 70 (1982) 1–14. [6] W.A. Neill, P.E. Jensen, G.B. Rich, J.D. Werschkul, Effect of decreased O2 supply to tissue on the lactate: pyruvate ratio in blood, J. Clin. Invest. 48 (1969) 1862–1869. [7] B. Phypers, J.M.T. Pierce, Lactate physiology in health and disease, Crit. Care Pain 62 (2006) 128–132. [8] L.A. Cynober, Plasma amino acid levels with a note on membrane transport: characteristics, regulation, and metabolic significance, Nutrition 18 (2002) 761–766.