Anal. Chem. 2004, 76, 328-334
An Interface for Direct Analysis of 14C in Nonvolatile Samples by Accelerator Mass Spectrometry Rosa G. Liberman,† Steven R. Tannenbaum,† Barbara J. Hughey,‡,§ Ruth E. Shefer,‡ Robert E. Klinkowstein,‡ Chandra Prakash,| Shawn P. Harriman,| and Paul L. Skipper*,†
Biological Engineering Division, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, Newton Scientific, Inc., 245 Bent Street, Cambridge, Massachusetts 02141, and Pfizer Global Research and Development, Groton, Connecticut 06340
We describe here apparatus and methods for direct analysis of 14C in biological specimens by accelerator mass spectrometry (AMS). Liquid samples, including plasma and urine, are deposited by pipet into a bed of CuO powder that fills a space within a rigid, refractory support. Volatile components are removed under reduced pressure prior to analysis. The CuO matrix is locally heated with an infrared laser while it is contained within a sealed chamber that is swept with He carrier gas. Heating induces combustion of the applied sample, and the carrier gas transports the CO2 that is formed to the AMS instrument’s ion source, which is appropriately modified for use with CO2. A rodent study of drug clearance with [14C]-acetaminophen was performed to provide plasma and urine specimens, which were analyzed with this overall approach and by liquid scintillation counting for comparison. Results presented here confirm the potential utility of laser-induced sample combustion as an alternative to graphite production for AMS analysis of 14C. Anticipated benefits of the present approach include reduced risk of sample cross-contamination, decreased analysis time, and greater compatibility with robotics. Accelerator mass spectrometry (AMS) is a well-established analytical technique that has become a fundamental tool for detection of low-abundance stable and long-lived radioactive isotopes in various research fields such as archeology, oceanography, nuclear physics, astrophysics, and the geological sciences.1 Since its invention in the early 1970s, this powerful technique has revolutionized the field of radiocarbon dating, allowing the detection of 14C in small organic samples with unprecedented accuracy and sensitivity.2 The potential of AMS for detection of 14C in biomedical studies was also recognized at about the same * Corresponding author: (fax) 617 252 1787; (e-mail)
[email protected]. † Massachusetts Institute of Technology. ‡ Newton Scientific, Inc. § Current address: Department of Mechanical Engineering, MIT, Cambridge, MA 02139. | Pfizer Global Research and Development. (1) Elmore, D.; Phillips, F. M. Science 1987, 236, 543-550. (2) Litherland, A. E. Annu. Rev. Nucl. Part. Sci. 1980, 30, 437-473.
328 Analytical Chemistry, Vol. 76, No. 2, January 15, 2004
time.3 Beginning in the early 1990s, biomedical AMS has become an active area of research.4-6 AMS as currently practiced in the analysis of 14C generally involves conversion of sample carbon to graphite for introduction into the ion source. Graphite production includes the sequential steps of oxidation to CO2 and reduction to elemental carbon. Various approaches to executing these steps are taken by different laboratories, but all involve multiple operations that are difficult to automate. Graphite samples are well-suited for AMS at very low isotope ratios ( 0.99) for isotope ratio as a function of actual amount of 14C, but the CV for the four runs was substantially greater (32%). The LLOQ for this type of sample appeared to be higher than for plasma samples, since the lowest concentration gave a signal only marginally greater than the blank. At 0.0093 dpm, though, the signal was severalfold greater than the blank, indicating that 0.01 dpm is a reasonable estimate of the LLOQ. Since the urine samples were undiluted, this corresponds to 10 dpm/mL. Two runs were made with fecal homogenate samples. The difference between the two slopes was 11% of the average value, and the regression lines had r 2 ) 0.998 and 0.981. In these analyses, the sample with 0.0031 dpm was omitted, but it was apparent that the LLOQ was substantially above this level: the (13) Skipper, P. L.; Hughey, B. J.; Liberman, R. G.; Choi, M. H.; Wishnok, J. S.; Klinkowstein, R. E.; Shefer, R. E.; Tannenbaum, S. R. Nucl. Instrum. Methods B, in press.
sample with 0.0093 dpm was 3-4-fold above the blank, so 0.01 dpm is a reasonable estimate of the limit of quantitation. In terms of sample concentration, this corresponds to a value of ∼1000 dpm/g. The higher value in comparison to the other two types of samples reflects the fact that plasma and urine comprise no more than ∼3% carbon by weight, while feces have a much higher carbon content. The dynamic range reflected by these results is 2 orders of magnitude. The current limitations on extending the range by decreasing the LLOQ is insufficient discrimination against species other than 14C2+ by the AMS instrument, as well as low transmission through the accelerator and high-energy analyzer. These are characteristics of the AMS instrument used in this work, which is itself a prototype, and cannot be regarded as a deficiency of the interface or of the real-time sample combustion approach to AMS analysis. With improvements to the AMS instrument, it is expected that the LLOQ will decrease by greater than 1 order of magnitude. As a practical matter, though, the range of isotope concentration in a set of samples run sequentially may continue to be restricted to more or less within the present limit by the response time of the ion source. Again, this is not a characteristic of the interface; a different type of ion source might permit a wider dynamic range. Pharmacokinetic Study Samples. Results of analysis of the male rat plasma samples are presented in Figure 5. Sufficient radioactivity for liquid scintillation counting was present only in samples taken at the first three time points. At these time points, the results of AMS and LSC analysis agreed within the margins of error. Data from the female rat samples (not shown) were very similar. Plasma samples were also analyzed after precipitation of protein with 2 volumes of acetonitrile, and the results are included in the figure. At the early time points, 8 h and earlier, isotope concentrations in whole plasma and the protein-precipitated plasma samples were indistinguishable. At 24 and 48 h, however, there were notable differences, consistent with the isotope being mostly in protein-bound form. Apparent protein binding appeared to be Analytical Chemistry, Vol. 76, No. 2, January 15, 2004
333
Table 2. Isotope Levels in Rat Urine Samples Following Administration of [14C]-Acetaminophena time (h)
AMS
LSCb
0-8 8-24 24-48
Male 6540c ( 1850 243, 265 26, 32
4280 371 44
0-8 8-24 24-48
Female 1940, 2020 209, 228 21, 22
2115 204 16
a Values are given in units of dpm/µL ( SD. b Results of single measurements with counting statistics of
