ac research
Anal. Chem. 2008, 80, 3515–3521
Accelerated Articles Subattomole Sensitivity in Biological Accelerator Mass Spectrometry Mehran Salehpour,* Go ¨ ran Possnert, and Helge Bryhni Ion Physics, Ångstro¨m Laboratory, Department of Engineering Sciences, Box 534, Uppsala University, SE-751 21 Uppsala, Sweden The Uppsala University 5 MV Pelletron tandem accelerator has been used to study 14C-labeled biological samples utilizing accelerator mass spectrometry (AMS) technology. We have adapted a sample preparation method for small biological samples down to a few tens of micrograms of carbon, involving among others, miniaturizing of the graphitization reactor. Standard AMS requires about 1 mg of carbon with a limit of quantitation of about 10 amol. Results are presented for a range of small sample sizes with concentrations down to below 1 pM of a pharmaceutical substance in human blood. It is shown that 14Clabeled molecular markers can be routinely measured from the femtomole range down to a few hundred zeptomole (10-21 mol), without the use of any additional separation methods. Measurement techniques with high sensitivities are of general interest to the analytical field. Isotopic labeling methods with consequent radioactive decay measurements such as liquid scintillation counting (LSC) have been very successful in providing valuable data in numerous diverse applications for a number of decades. The most widely used isotope is 14C, utilized in a variety of organic compounds within different disciplines of science including chemistry, biology, and medicine. One limitation of LSC in the biological field and in particular the clinical field is the detection sensitivity. As the total applied radiation exposure to human volunteers may not be arbitrarily high, this sets some boundary conditions on the lowest limit of detection, especially in cases where one is studying low level traces (for a recent review see Lappin and Tempel1). An analogous, yet technologically completely different method is accelerator mass spectrometry * Corresponding author. Fax: +46 18 555 736. E-mail address:
[email protected]. (1) Lappin, G.; Temple, S. Radiotracers in Drug Development; CRC Press, Taylor and Francis Group: Boca Raton, FL, 2006. 10.1021/ac800174j CCC: $40.75 2008 American Chemical Society Published on Web 04/19/2008
(AMS). Whereas LSC detects the β-decay rate, dictating the need for a certain mass of the sample to give a detectable signal, AMS measures the isotopes directly and individually using single particle counting and therefore enhancing the sensitivity by 5-6 orders of magnitude compared to LSC. For 14C, with a half-life of 5760 years, the maximum specific activity achievable for this isotope is 2.3 GBq/mmol (1.38 × 1011 dpm/mmol). With this half-life, one disintegration per minute will occur on average within a population of 4.35 × 109 atoms of 14C. The sensitivity of AMS is therefore explained because unlike LSC, which detects the number of atomic disintegrations, AMS counts individual atoms of 14C in relation to the abundance of 12C. Furthermore, LSC requires more than a gram of carbon to get a reasonable count rate (>10 atomic disintegrations per minute, dpm) whereas AMS requires less than 1 mg of the sample and the data acquisition times are normally about a few minutes to get a 1% accuracy. AMS is a well established method, utilizing medium energy particle accelerators (>1 MeV/nucleon) to perform precise and ultrasensitive isotopic ratio measurements with extremely high sensitivity in the 1015:1 range. The technology has resulted in a number of applications (for a review see Elmore and Phillips2) in diverse disciplines such as archeology using the 12C/14C ratio (Τ1/2 ∼ 5760 years), geochronology with 9Be/10Be (Τ1/2 ∼ 1.5 × 106 years), and environmental sciences with 127I/129I (Τ1/2 ∼ 17 × 106 years). Within the bioanalytical field, Vogel et al. pioneered the implementation of AMS for biochemical samples as a new analytical tool.3 The method is based on using a 14C labeled molecule as the biological marker, for example, administered in the form of a pharmaceutical drug to animals or humans. The samples are taken as blood, urine, breath, tissue, biopsies, etc. and are analyzed with AMS, with a sensitivity demonstrated down (2) Elmore, D.; Phillips, F. M. Science 1987, 236, 543–550. (3) Vogel, J. S.; Turteltaub, K. W. Nucl. Instrum. Methods B 1994, 92, 445– 453.
Analytical Chemistry, Vol. 80, No. 10, May 15, 2008
3515
Figure 1. Schematic diagram of a biological accelerator mass spectrometry experiment. The different steps i-iv involved in the measurements of biological samples are shown (see text).
to 10 amol (10-18).4 The method is not limited to 14C labeling but is by far dominated by it. Other isotopes can also be used in organic substances such as 3H, 41Ca, and 129I. The availability of AMS and its ease of use has improved significantly during the past decade,5 and consequently the costs associated with the technique have dropped substantially, reducing the cost of analyzing a sample to a few hundred Euros. This is partly due to the technology being more mature and therefore more reliable and partly due to the fact that new compact and less expensive accelerators have been introduced to the market which has enabled successful commercialization of AMS into the pharmaceutical and biochemical field. All the different variety of experiments that have been performed throughout the years using the LSC counting method can now be realistically performed with AMS, with hundreds of thousands times improved sensitivity. One exciting application of biological AMS in the pharmaceutical research and development sector is the microdosing concept. The major pharmaceutical companies have been struggling with the rising costs of research and development while the number of approved drugs has been decreasing for many years. Pharmacological drugs take many years to develop and are very costly. A major portion of the costs are a result of the high attrition rate, i.e., the drugs that, for miscellaneous reasons, are rejected, often after years of research and many experiments with animals and humans. The microdosing concept is based on using very small doses of drugs (1/100 of the therapeutic dose or a maximum of 100 µg per person) to reduce toxicity and side-effect risk issues associated with therapeutic doses. Consequently this requires a very sensitive method, such as AMS, to determine the drug concentration. Lappin and Garner performed the first AMS (4) Miyashita, M.; Presley, J. M.; Buchholz, B. A.; Lam, K. S.; Lee, Y. M.; Vogel, J. S; Hammock, B. D. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 4403–4408. (5) Suter, M. Nucl. Instrum. Methods B 2004, 22, 3-224; 139-148.
3516
Analytical Chemistry, Vol. 80, No. 10, May 15, 2008
microdosing experiments,6 the usefulness of which, in combination with AMS is at least twofold: it can be utilized to (i) select candidate drugs with the best pharmacokinetic properties and (ii) determine suitable starting dose in first-time-in-man studies. As microdosing was approved by the FDA7 in November 2006 (also earlier by EMEA8) it has increased the demand and the level of interest in this type of measurement. To increase the sensitivity of the method has therefore obvious advantages. Attomole sensitivity in AMS has been demonstrated,4 but although a number of references are made to the possibility of zeptomole sensitivity,1,9 it has not been demonstrated. We are presenting data on zeptomole sensitivity of a pharmaceutical drug in human blood utilizing AMS technology. EXPERIMENTAL SECTION The different steps in a biological AMS are shown in Figure 1 are (i) the sample with the 14C marked molecule is dried and converted to graphite which is suitable for the AMS ion source, (ii) the negative atomic and molecular ions are created from the graphite samples and accelerated, (iii) the high energy collision in a gas cell obliterates any traces of molecular species, and (iv) the 12C/14C ratio is measured. Accelerator Mass Spectrometry. A 5 MV Pelletron tandem accelerator (NEC Inc. Middleton, WI) installed in 2001, is used for the AMS studies. The ion source is a negative ion SIMS (secondary ion mass spectrometry) sputter ion source which uses (6) Lappin, G.; Garner, C. Nat. Rev. Drug Discovery 2003, 2, 223–240. (7) Food and Drug Administration, U.S. Department of Health and Human Services Guidelines for Industry Investigators and Reviewers. Exploratory IND Studies; January 2006. (8) EMEA, Position paper on Non-Clinical Safety Studies to Support Clinical Trails with a Single Microdose. Position Paper CPMP/SWP/2599; June 23, 2004. (9) Vogel, J. S.; Love, A. Methods in Enzymology, Burlingame, A. L., Ed.; Academic Press: NewYork, 2005.
primary ions in the keV energy range to cause the emission of sputtered secondary ions. Two independent cesium (Cs+) sputter ion sources are used, both with automated multiple sample holders. One ion source is homemade (25 samples) and is used for samples with high 14C contents. The other is a high-throughput commercial ion source (40 samples, MC-SNICS, NEC) providing about 3 times higher ion beam current than the other ion source (∼1.2 µA of 13C- prior to detection). The latter is preferentially utilized for archeological studies or other samples with low 14C content. We enforce this separation so that in case of an accidental contamination, the high sensitivity, daily archeological measurements are not affected. The secondary negative ions from the sample are accelerated to a potential of about 50 kV. Mass-to-charge ratio (m/z) separation occurs through a 90° magnet, and the ions are then steered, focused, and injected into the accelerator toward the positive potential of the accelerator terminal (maximum potential +5 MV). The high energy ions collide with the gas molecules in a lowpressure cell where all molecular entities, including isomers (e.g., 12,13 CH-) undergo nonadiabatic, ion-electron collisions. As a result of this, they are stripped off their electrons and are dissociated by “Coulomb explosion”, resulting in a distribution of atomic positive charge states with no molecular species present. The positively charged ions are now accelerated once again, this time, away from the terminal and are mass-to-charge analyzed through another 90° magnet. The beam goes through a number of focusing elements and a 30° switch-magnet before unambiguous identification in an energy sensitive semiconductor detector. The vacuum throughout the system is on the order of 3 × 10-8 mbar. Isotopic ratios are then measured by single particle counting of 14 C and current measurements for 13,12C. The 14C/12C ratio, R, is presented in units of Modern or as a percentage of Modern (pMC). One Modern corresponds to present day 14C content which is approximately 98 amol 14C/mg 12C, 13.56 dpm/g 12C or 6.1 fCi/mg 12C depending on the unit of preference. It is common procedure to have one 14C-free sample (old carbon) and 2-3 standard reference material samples (oxalic acid II, obtained from NIST, Boulder, CO) in the holder for every AMS experiment. A sample is nominally run for 5 min after which the reference samples are measured, and if there are any significant deviations the software will trigger an alert for inconsistencies. A standard sample is run for three acquisition periods of 5 min each, and the average value is presented. The accuracy is in the range of 0.4-0.8% depending on the absolute current from the ion source which affects the Poisson distribution’s statistical fluctuations and hence the measured error. The accelerator and ion sources operate in an automatic mode for overnight measurements. Finally, to compensate for isotopic fractionation effects, the combusted samples (CO2 gas) are also analyzed in an isotope ratio mass spectrometer where the 13C/12C ratios are measured. Any fractionation effects are compensated for according to the formulation by Stuiver et al.10 and Mook and van der Plicht.11 For biological samples, however, the isotopic fractionation, δ13C is well under 1% and is normally not significant compared to the experimental uncertainties. (10) Stuiver, M.; Polach, H. A. Radiocarbon 1977, 19, 355–363. (11) Mook, W. G.; van der Plicht, J. Radiocarbon 1999, 41, 227–239.
Sample Preparation. The samples are normally dried in vacuum and are pumped until a pressure of less than 5 × 10-6 mbar is reached. Samples containing more than a few tens of microliters of liquid (e.g., urine) must be dried separately in an oven or should alternatively be freeze-dried. The sample preparation method of Ognibene et al.12 has been modified and optimized for preparation of small size samples as will be briefly described here. The biological samples are weighed and placed in prebaked quartz tubes with about 100 mg of CuO powder and are then vacuum-sealed using a high temperature hydrogen and oxygen torch. The sealed quartz tubes, which have one end as a breakseal, are placed in an oven and heated to 900 °C for 2.5 h and are allowed to cool slowly. The quartz tubes are connected to a vacuum system with the break-seal end in a metallic bellow which can be bent to puncture the break-seal and release the gas. Once the vacuum has reached the 10-6 mbar region, the vacuum pump valve is closed and the sample gas (mostly CO2) is cryogenically transferred into a septa-sealed borosilicate vial through a hypodermic needle. After the gas transfer, the needle is removed trapping the gas in the graphitization reactor consisting of a septaseal vial with 100 mg of zinc and a smaller vial containing iron powder (1 mg) separated by a few borosilicate balls. The vial containing the sample gas is removed and placed in another oven with the upper part of the vial, including the septa-seal, remaining at room temperature. The vial is heated to 525 °C for 6 h where the graphitization takes place through catalytic deposition of carbon (graphite) aggregates on the iron powder. The samples are finally collected and pressed into an aluminum holder (o.d. ) 1 mm) and sent for AMS analysis. For standard AMS measurements, about 1 mg of carbon is used for each sample corresponding to 10 µL of blood (10% carbon content), 25 µL blood plasma (4.5% carbon content), or 100 µL of urine (about 1% carbon).9 A number of modifications have been implemented in order to optimize the method for small samples. (1) Miniaturization of the graphitization reactor: The standard samples contain about 1 mg of carbon undergoing high temperature graphitization at pressures of a few bars. The small samples that we have studied (down to a few tens of micrograms) have considerably less carbon with correspondingly lower pressure in our standard vials (2 mL), leading to inefficient and irreproducible graphitization. We have therefore reduced the size of the graphitization reactor to a total volume of 0.7 mL. The small reactors facilitate production of graphite with sufficiently high negative ion yields in our sputter ion source as will be discussed in the next section. (2) Gas transfer: The transfer of the sample CO2 from the combustion tube to the graphitization vial takes place through a hypodermic needle which has an inner diameter of 260 µm and a length of 1.8 cm. For large samples, in spite of low gas conductance through the needle, the high pressure gradient enforces the gas transfer. However, for smaller samples the pressure gradient becomes too low to transfer the gas efficiently. Allowing long periods to cryogenically collect the gas is not possible as the septa seal will cool and harden after a few minutes and will lead to a gas leak once the needle is removed. We therefore heat the quartz combustion tube which contains the sample CO2 gas to about 300 °C prior to gas transfer into the graphitization vial. Similarly, in an effort to increase the pressure, all volumes have been (12) Ognibene, T. J.; Bench, G.; Vogel, J. S. Anal. Chem. 2003, 75, 2192–2196.
Analytical Chemistry, Vol. 80, No. 10, May 15, 2008
3517
minimized, including the combustion quartz tube and the vacuum transfer tubes. (3) Vacuum: For standard biological samples (milligrams) a few minutes of pumping with the high vacuum pump is sufficient from a vacuum contamination point-of-view. However, small samples are more susceptible to contaminations affecting the measurements, e.g., adsorbed atmospheric CO2 during handling or contaminations originating from the degassing of the vacuum transfer system. Consequently, we have pumped the samples for at least an hour prior to gas transfer, reaching a pressure of below 5 × 10-6 mbar measured in the region between the pump and the graphitization vial. (4) Extended graphitization periods: For standard samples, high temperature graphite formation is completed within a few hours. However, for small samples, longer graphitization times of up to 18 h are required. Special care must be taken to avoid contamination by any carbon-containing substances or tools in the laboratory which could affect the isotopic ratio. All vials are prebaked, 2 h at 900 °C for quartz vials and 6 h at 550° for borosilicate vials and parts. RESULTS AND DISCUSSIONS The antipsychotic drug (Remoxiprid, C16H23BrN2O3 · HCl) has been used as the labeled molecule for our experiments. The drug has been 14C-marked with a specific activity of 2.035 GBq/mmol, dissolved in water and then in human blood. The sample was diluted in human blood in a number of steps spanning 3 orders of magnitude, starting from a concentration of about 170 pM (activity of 0.36 Bq/mL). The samples were prepared according to the method described in the earlier section, using amounts which varied in volume from 0.1-10 µL. The total amount of drug per sample thus varied in the range from about 20 ag to 3 pg. The samples were weighed, prepared, and measured by AMS. The concentration, K, of the labeled compound in fg of sample per mg of blood, plasma, etc. is given by K ) (RM - RN)Ψ(W ⁄ L)
(1)
where L is the specific molar activity of the 14C-labeled compound in fCi/fmol, Ψ is the carbon mass fraction of the sample, W is the molecular weight of the compound in fg/fmol, RM is the measured 14C amount of the samples per unit mass of 12C, and RN is the corresponding value for an unlabeled sample.13,14 R values are given in units of Modern (1 Modern ) 6.1 fCi/mg C). Similarly, the concentration of the compound, χ in fmol per mg of blood is given by χ ) δRΨ ⁄ L
(2)
where, δR ) RM - RN. Thus, for every experiment requiring qualitative measurement of K or χ, an unlabeled sample needs to be measured. Figure 2 shows AMS measurement results from seven different concentrations of 14C-Remoxiprid in human blood, performed for standard sample sizes (>5 µL of blood or 500 µg of carbon content) as well as sample sizes of 1, 2, and 3 µL corresponding to 100, 200, and 300 µg of carbon in the samples, respectively. The small variations observed is a result of the experimental (13) Salehpour, M. S., Possnert, G., Bryhni, H., Palminger-Halle´n, I., Ståhle, L. Appl. Radiati. Isot., accepted for publication. (14) Vogel, J. S. Nucl. Instrum. Methods B 2000, 172, 884–891.
3518
Analytical Chemistry, Vol. 80, No. 10, May 15, 2008
Figure 2. The measured isotopic ratio, RM in percent Modern (pMC) as a function of seven different molar concentrations of 14CRemoxiprid in human blood. For each concentration, four different sample sizes have been used: a standard sample containing >0.5 mg of carbon (>5 µL of blood, O), 300 µg of carbon (3 µL of blood, [), 200 µg of carbon (2 µL of blood, ]), and 100 µg of carbon (1 µL of blood, ×).
uncertainty contribution of about 1-2% arising from sample-tosample fluctuation including the statistical error. The data in Figure 2 shows that much smaller samples can be used to perform concentration measurements with a few percent accuracy as in the case of standard samples, requiring 10 times less material. This performance is considerably better than what has been reported using the sample preparation method of Ognibene et al.12 and is attributed to the following modifications. We have miniaturized the gas handling assembly, including the vacuum tubes and the graphitization reactor. The latter is important from a graphitization point-of-view. For small samples, the CO2 partial pressure is lower by a factor of up to 10 for the data in Figure 2 compared to a standard sample. Thus, the graphitization becomes inefficient leading to low secondary negative ion yields in the ion source. Consequently, reducing the reactor volume facilitates graphitization of smaller samples. We reduced the reactor volume by a factor of 3 down to 0.7 mL. Another feature of the sample preparation geometry is the low pressure conductance through the hypodermic needle between the space where the sample gas is released and the graphitization reactor. This setup works well for large samples where the pressure gradient facilitates gas transfer. For small samples we have found that it is necessary to heat up the sample gas in order to be able to transfer the gas within a short enough time while cryogenically cooling the graphitization reactor. Longer cooling time will lead to hardening of the septa-seal material with subsequent leakage of the sample gas after the removal of the needle. The vacuum properties of the system can also affect the measurements.13 We have ensured that good vacuum conditions are maintained during the whole process, pumping over an hour for each sample (
