Clinical Chemistry 55:10
1824–1833 (2009)
Automation and Analytical Techniques
Measurement of Ribosomal RNA Turnover In Vivo by Use
of Deuterium-Labeled Glucose
Julien Defoiche,1,2 Yan Zhang,1 Laurence Lagneaux,3 Ruth Pettengell,4 Andrea Hegedus,1 Luc Willems,2,5
and Derek C. Macallan1*
BACKGROUND: Most methods for estimation of rates of
RNA production are not applicable in human in vivo
clinical studies. We describe here an approach for measuring ribosomal RNA turnover in vivo using [6,62
H2]-glucose as a precursor for de novo RNA synthesis.
Because this method involves neither radioactivity nor
toxic metabolites, it is suitable for human studies.
METHODS:
For method development in vitro, a lymphocyte cell line (PM1) was cultured in the presence of
[6,6-2H2]-glucose. RNA was extracted, hydrolyzed enzymatically to ribonucleosides, and derivatized to either the aldonitrile tetra-acetate or the pentafluoro
triacetate derivative of the pentose before GC-MS.
We identified optimum derivatization and analysis
conditions and demonstrated quantitative incorporation of deuterium from glucose into RNA of dividing cells.
RESULTS:
Pilot clinical studies demonstrated the applicability of this approach to blood leukocytes and solid
tissues. A patient with chronic lymphocytic leukemia
received [6,6-2H2]-glucose (1 g/kg) orally in aliquots
administered every 30 min for a period of 10 h. When
we analyzed CD3– B cells that had been purified by
gradient centrifugation and magnetic-bead adhesion,
we observed deuterium enrichment, a finding consistent with a ribosomal RNA production rate of about
7%/day, despite the slow division rates observed in
concurrent DNA-labeling analysis. Similarly, in 2 patients with malignant infiltration of lymph nodes, administration of [6,6-2H2]-glucose (by intravenous infusion for 24 h) before excision biopsy allowed
estimation of DNA and RNA turnover in lymph node
samples.
CONCLUSIONS: Our study results demonstrate the
proof-of-principle that deuterium-labeled glucose
1
Centre for Infection, St George’s, University of London, London, UK; 2 Department of Molecular & Cellular Biology, FUSAG, Gembloux, Belgium; 3 Department of Hematology Bordet Hospital/ULB, Brussels, Belgium; 4 Department of
Hematology, St George’s, University of London, London, UK; 5 Molecular and
Cellular Epigenetics, Interdisciplinary Cluster for Applied Genoproteomics
(GIGA), University of Liège, Belgium.
* Address correspondence to this author at: Centre for Infection, St George’s,
1824
may be used to analyze RNA turnover, in addition to
DNA production/cell proliferation, in clinical samples.
© 2009 American Association for Clinical Chemistry
Variations in both the type and amount of RNA in the
cell are pivotal for the regulation of intermediary metabolism. Cellular RNA is in a state of rapid turnover,
and various RNA species have widely varying life spans.
Messenger RNAs (mRNA)6 have the shortest survival
times, with half-lives of minutes to hours (1 ), and
modification of mRNA stability is an important regulatory point for control of protein translation. Transfer
RNAs (tRNA) have longer half-lives, measured as
hours to days. Ribosomal RNA (rRNA) persists for several days; estimates for rRNA half-life in vitro range
from ⬍3 days (human fibroblasts) (2 ), through 3.8
days (18S rRNA moiety in H1299 cells) (3 ), to about
7.5 days (cultured rat fibroblasts) (4 ).
RNA turnover represents the aggregate rate of synthesis and degradation of all RNA species. Changes in
mRNA levels are best addressed using molecular approaches or microarrays. Small RNA species such as
tRNA may also be quantified using specific probes. Analyzing rRNA kinetics is more difficult, particularly if
human in vivo measurements are envisaged. In animal
studies U-C14–labeled glucose and cytidylic acid have
been used to estimate the turnover in rat organs (5 )
and, in vitro, 3H-uridine may be used (6 ). Analogous
approaches for clinical studies are, however, limited by
the toxicity of such radioactive isotopes.
rRNA kinetics is intimately related to cellular metabolism. Changes in RNA content precede changes in
protein metabolism in response to cytokine or hormone action or cell division (7 ), and changes in RNA
turnover must de facto precede changes in RNA content. Dysregulation of ribosome biosynthesis is associ-
University of London, Cranmer Terrace, London SW17 0RE, UK. Fax ⫹44-208725-3487; e-mail [email protected].
Received October 22, 2008; accepted July 15, 2009.
Previously published online at DOI: 10.1373/clinchem.2008.119446
6
Nonstandard abbreviations: mRNA, messenger RNA; tRNA, transfer RNA; rRNA,
ribosomal RNA; ATA, aldonitrile tetra-acetate; PFTA, pentafluoro tri-acetate; NCI,
negative chemical ionization mode; PBMC, peripheral blood mononuclear cells.
Measuring RNA Turnover by Use of Deuterium-Labeled Glucose
Fig. 1. Metabolic pathway and analytic strategy for RNA labeling through the de novo nucleotide synthesis
pathway.
(A), Pathway for incorporation of deuterium from glucose into RNA; not all intermediates are shown. (B), Stages in isolation
and analysis of the ATA derivative of adenosine. Labeled hydrogens are shown as D. G6P, glucose 6-phosphate; R5P, ribose
5-phosphate; NDP, ribonucleoside diphosphate; NDKs, nucleoside-diphosphate kinases; NTP, ribonucleoside triphosphate.
ated with cancer; for example, overexpression of rRNA
may lead to malignant transformation (8 ). In chronic
lymphocytic leukemia cells, investigation using
methionine-methyl-tritium revealed defective regulation of new ribosome assembly (9 ). The dynamic nature of rRNA contrasts with genomic DNA, which remains very stable within nondividing cells. The
genomic DNA synthesis rate therefore represents a surrogate for proliferation, a property that has been exploited in recently described methods for measuring
cell proliferation using stable (nonradioactive) isotope
labeling in vivo in humans (10 –14 ). Applications include the measurement of lymphocyte kinetics in HIV
and HTLV-I infection, malignant cell proliferation in
leukemia, and vascular smooth muscle in atheroma
formation (15–20 ).
In this study, we set out to develop an analogous
isotopic approach for measuring rRNA synthesis in
vivo in human clinical studies. We hypothesized that
rRNA could be labeled in the same way as DNA
(11, 12 ), with deuterium-labeled glucose through the
de novo nucleotide synthesis pathway (Fig. 1A). In the
cell, [6,6-2H2]-glucose is phosphorylated and converted via the pentose cycle into [5,5-2H2]phosphoribose pyrophosphate; this process results in
labeled precursor for ribonucleotide synthesis. To test
this hypothesis, we first performed the 4 steps previ-
ously described for DNA labeling: label administration,
nucleoside extraction, derivatization, and GC-MS
analysis (Fig. 1B) (10 –12, 21 ). We validated or modified each step until optimized for RNA/ribonucleoside
analysis using adenosine as a marker of RNA metabolism. We then investigated total RNA production in
rapidly proliferating tissue culture cells in vitro, and
then determined whether this approach may be applied
in vivo in humans to measure rRNA turnover in circulating lymphocytes and tissue samples.
Materials and Methods
METHOD DEVELOPMENT
Isolation of ribonucleosides from RNA. Ribonucleosides
were separated from RNA by sequential digestion with
nuclease S1 (18 h, 37 °C), phosphodiesterase I (2 h,
37 °C), and alkaline phosphatase (1 h, 37 °C) (enzymes
from Sigma), as previously described (22 ), and analogous to the approach for DNA (10 –12 ). Efficiency of
digestion was confirmed by HPLC against ribonucleoside standards. Adenosine was separated from other
ribonucleosides by use of a solid-phase extraction column (LC-18, Supelco), and purity and efficiency were
confirmed by HPLC.
Clinical Chemistry 55:10 (2009) 1825
Derivatization of ribonucleosides and GC-MS analysis.
Derivatization conditions required for GC-MS analysis
are a compromise between the formation of the derivative and its destruction. For DNA analysis, deoxydenosine may be converted to its aldonitrile triacetate
derivative by reaction with hydroxylamine/pyridine
(1% wt/vol, 100 °C, 45 min) followed by acetic anhydride (room temperature). Adenosine was derivatized
with the same reagents but at varying time and temperature combinations, as described below. Analysis of the
aldonitrile tetra-acetate (ATA) derivative was performed by GC-MS (HP-225 column; Agilent 6890/
5973, Agilent Technologies) under positive chemical
ionization with methane. Ions were monitored in selected ion-monitoring mode at m/z 256 and 258 for
adenosine and 198 and 200 for deoxyadenosine; these
represent the M⫹0, M⫹2 mass isotopomers of the dominant ion after ionization.
We also investigated an alternative derivatization
approach generating the pentafluoro triacetate (PFTA)
derivative of adenosine, modifying the protocol previously described for deoxyadenosine (10 ). Samples were
reacted with aqueous O-(2,3,4,5,6-pentafluorobenzyl)hydroxyl-amine hydrochloride solution (1 g/L) and glacial acetic acid at various temperature/time combinations. After the mixtures cooled to room temperature,
acetic anhydride and N-methylimidazole were added
with rapid mixing. The reaction was allowed to proceed
for 15–20 min. After cooling, water was added and the
derivative extracted from the aqueous phase with dichloromethane. GC-MS analysis was performed as
above but in negative chemical ionization mode (NCI).
IN VITRO LABELING OF CELLS IN TISSUE CULTURE
We performed in vitro studies of label incorporation
from [6,6-2H2]-glucose into cellular RNA with PM1
cells (a clone derived from human cutaneous T-cell
lymphoma) grown in suspension (RPMI 1640 media,
10% FCS, Gibco BRL). Media were labeled by adding
[6,6-2H2]-glucose (Cambridge Isotope Laboratories)
and cells were cultured for varying periods up to 96 h
before harvesting. RNA was extracted by guanidine
thiocyanate/phenol extraction in a monophase solution (Trizol Sigma). Media enrichments were confirmed by GC-MS of the ATA derivative of glucose.
IN VIVO LABELING OF HUMAN LYMPHOCYTES AND TISSUE
Peripheral blood cells. A patient with chronic lymphocytic leukemia was given 1g/kg 6,6-2H2-glucose (Cambridge Isotope Laboratories) orally for a period of 10 h,
as half-hourly aliquots administerd after a priming
dose, as previously described (16, 21 ). The patient was
a 59-year-old man with Binet stage A disease, who had
a peripheral blood B-cell count of 8.9 ⫻ 109 cells/L and
1826 Clinical Chemistry 55:10 (2009)
an unmutated immunoglobulin variable region heavychain phenotype. The study was granted ethics approval by Bordet Hospital in Brussels, Belgium. The
patients gave written informed consent, and all procedures were performed in conformity with the Declaration of Helsinki.
Blood samples were taken for estimation of deuterium enrichment at 7 time points from 3–30 days after
the day of ingestion of labeled glucose. Peripheral
blood mononuclear cells (PBMC) were isolated by Ficoll density gradient centrifugation and CD3– B cells
purified by negative selection using CD14/CD56 and
CD3-coupled magnetic beads (MACS, Miltenyi Biotec
GmbH). After cryopreservation (RNA-later, Ambion),
both DNA and total RNA were extracted; mRNA was
isolated from total RNA by use of MicroPoly(A) Purist
kit (Ambion). Samples were analyzed for deuterium
enrichment as described above. Plasma glucose enrichment was measured using the ATA derivative (monitoring ions m/z 328 and 330) as described (12 ).
Analysis of RNA synthesis in tissue samples. In vivo RNA
turnover in lymph nodes was investigated in lymph
node excision biopsy samples from 2 patients. Patient
A had non-Hodgkin lymphoma (NHL01). Patient B
was undergoing investigational procedures for HIVassociated lymphadenopathy; histology showed Kaposi sarcoma. Labeling consisted of a primed 24-h intravenous infusion of 1.3 g/kg [6,6-2H2]-glucose
(Cambridge Isotope Laboratories) initiated 24 h before
the time of biopsy (12 ). Blood samples for measurement of plasma glucose deuterium enrichment were
taken during infusion. Tissues samples were homogenized (Polytron) in Trizol reagent (Sigma), and RNA
and DNA were isolated. Deoxyadenosine and adenosine were purified and analyzed for deuterium content
by GC-MS as described above.
MODELING AND CALCULATION
We analyzed the kinetics of label incorporation and
loss in the in vivo studies by using a mathematical
model analogous to that previously described for measurement of DNA/cell turnover (23, 24 ), as shown in
Fig. 1 in the Data Supplement that accompanies the
online version of this article at http://www.clinchem.
org/content/vol55/issue10. Experimental results were
expressed as fractional enrichment (F), defined as the
proportion of RNA labeled relative to the labeling rate
in its precursor, given by: F ⫽ E/b, where E is deuterium
enrichment in RNA and b is the deuterium enrichment
in plasma glucose, corrected as previously described
(12 ). Fractional enrichment/time curves were modeled
and solved using nonlinear least squares regression
(Sigmaplot v8.02; SPSS) to yield best-fit values for the
fractional synthesis rate of RNA (s), and the fractional
Measuring RNA Turnover by Use of Deuterium-Labeled Glucose
loss rate of labeled RNA (l*) according to the
relationships:
F共t兲 ⫽
s
共1 ⫺ e ⫺l*t 兲
l*
t ⱕ ␶,
during labeling
(1)
F共t兲 ⫽
s
共1 ⫺ e ⫺l* ␶ 兲e ⫺l* 共 t⫺␶ 兲
l*
after labeling
t ⬎ ␶,
(2)
where t is time from the beginning of labeling and ␶ is
the duration of labeling. The derivation of these equations is given in online Supplemental Fig. 1. Doubling
times (t2) and half-lives (t1/2), where given, were calculated as ln2/s and ln2/l*, respectively.
Results
RNA DIGESTION TO RIBONUCLEOSIDES AND DERIVATIZATION
OF ADENOSINE
HPLC confirmed high retrievals of adenosine, with
⬎95% purity, from RNA. When optimal conditions
for derivatization of adenosine were investigated, we
observed, first, that no appreciable derivatization of
adenosine (6 mmol/L) occurred at or below 80 °C (Fig.
2A). However, at the higher temperatures required for
production, the ATA derivative was unstable (Fig. 2B).
An optimal compromise between production and destruction was achieved at 110 °C for 2 h, conditions
adopted in subsequent experiments.
When deoxyribonucleosides and ribonucleosides
were derivatized together at 110 °C, the derivative of
deoxyadenosine formed rapidly and then degraded,
whereas the derivative of adenosine formed more
slowly (Fig. 2C). Although there was a crossover point
(about 75 min), neither substrate yielded more than
50% of maximum yield under such conditions. We
therefore concluded that unless cellular material is
available in abundance, it is better to separate DNA and
RNA first, then derivatize separately under different
conditions. Estimates of the deuterium content of standards, containing known amounts of 5,5-D2-ribose
(Cambridge Isotope Laboratories), based on the parent
and M⫹2 ion abundances (m/z 256, 258; Fig. 2, D and
E) yielded excellent correlation between measured and
predicted values (Fig. 2F); subsequently standards were
run with all samples for calibration.
When the PFTA derivative was analyzed by NCI
(Fig. 3A), it yielded 2 chromatographic peaks (Fig. 3B)
corresponding to cis and trans isomers. Analyses performed at varying temperature and time conditions
demonstrated progressive formation of the derivative
at 100 °C or 110 °C. At 100 °C the derivative was rela-
tively stable, as shown by the ribose derivative, which
forms rapidly (Fig. 3C). Maximal yields were obtained
from adenosine at ⱖ16 h (Fig. 3C). Analysis of enriched standards demonstrated that either of the dominant ions, m/z 513 and 433, could be used for quantification (Fig. 3D); the former corresponds to the parent
ion whereas the latter is consistent with loss of [acetate ⫹ HF]. ATA and PFTA derivatives both gave excellent results for enrichment analysis (Fig. 2F and 3D).
The PFTA derivative appeared about 2–3 times more
sensitive than ATA in terms of abundance vs amount
injected (Fig. 3E) and in terms of the starting number
of cells required to obtain abundances from which isotope ratios could be derived, roughly 1 million for PFTA
(Fig. 3F) and 3 million for ATA. The nonlinearity of Fig.
3F compared to Fig. 3E suggests that some stages of sample processing may be saturable at high cell numbers. The
PFTA derivative would thus appear preferable, but if NCI
analysis is not available, ATA derivatization is an alternative. The in vitro and in vivo experiments described below
were all performed using the ATA derivative with positive
chemical ionization analysis.
LABELING OF DIVIDING LYMPHOCYTES IN VITRO
When PM1 cells were grown in labeled media, deuterium enrichment increased progressively in both adenosine from RNA and deoxyadenosine from DNA.
Initially adenosine enriched more quickly than deoxyadenosine (Fig. 4); in a typical experiment, the fractions of new RNA and DNA at 2 h were estimated to be
4.1% and 0.4%, respectively, and at 6 h 10.9% and
4.1%. The excess of RNA synthesis over DNA synthesis
at these earlier time points is likely to result from rapid
turnover (and thus labeling) of pools of RNA, such as
mRNA and tRNA. Beyond about 6 h the enrichment
curves for RNA and DNA become parallel (Fig. 4), suggesting that new RNA and DNA are being synthesized
at similar, very rapid rates. The lack of divergence
would be consistent with saturation of rapid turnover
RNA subspecies. RNA synthesis beyond this point may
correlate so closely with new cell synthesis because of
the need to generate a new complement of RNA for
each new cell, which dominates RNA metabolism in
such rapidly dividing cells. After prolonged incubation,
deuterium enrichments converged toward a plateau at
about 50% of that for the glucose in the media (Fig. 4),
giving an estimate of the level of labeling of intracellular RNA precursors.
IN VIVO LABELING
Turnover of RNA in peripheral blood leukocytes. To investigate in vivo production of RNA, we used an approach that was analogous to that used in vitro, but
several major differences should be noted. First, labelClinical Chemistry 55:10 (2009) 1827
Fig. 2. Production of the ATA derivative of adenosine with hydroxylamine/pyridine and acetic anhydride.
Abundance represents the amount of aldonitrile tetra-acetate derivative of adenosine (ATA) formed and is expressed as the
GC-MS signal in arbitrary units for the relevant ion (m/z 256). (A), Temperature dependence of formation of ATA derivative from
adenosine: 60 °C (⽧), 80 °C (f), 100 °C (Œ). (B), Stability of ATA derivative from adenosine at 90 °C (f ) and 110 °C (⽧).
(C), Effects of simultaneous derivatization of isomolar solutions of adenosine (f ) and deoxyadenosine (diamond) at 110 °C. (D),
Mass spectrum of ATA derivative showing the dominant ion with positive chemical ionization at m/z 256. (E), Chromatographic
curve of the derivative of unlabeled (m/z 256) and labeled (m/z 258) adenosine. (F), Observed (Ob) vs predicted (Th) deuterium
enrichment for ATA derivative of ribose standards of known enrichment with 5,5-D2-ribose (0%–1%); enrichment is expressed
as atom percent excess (APE), the proportional abundance of 2H2-labeled moieties within the total amount of derivative. The
RNA standard curve was characterized by a slope y ⫽ 1.0988x and correlation coefficient of R2 ⫽ 0.9981.
ing in vivo was transient rather than continuous, constituting a “pulse-chase” design. Second, measurements were made during unlabeling rather than
labeling. Third, whereas most RNA production in very
rapidly dividing tissue-culture cells is related to the
1828 Clinical Chemistry 55:10 (2009)
production of new RNA for new cells, RNA turnover
rates in vivo, in which most cells divide far less frequently, are more likely to reflect intracellular turnover
of RNA rather than RNA production related to cell
division. Fourth, we considered that labeling curves re-
Measuring RNA Turnover by Use of Deuterium-Labeled Glucose
Fig. 3. Production of the PFTA derivative of adenosine.
Abundance represents the amount of PFTA derivative formed and is expressed as the GC-MS signal in arbitrary units for the
relevant ion (m/z 433 or 513). (A), mass spectrum of ATA derivative showing dominant ions with NCI at m/z 433, 493 and 513.
(B), Chromatographic curve of the PFTA derivative of adenosine showing unlabeled (m/z 513) and labeled (m/z 515) isotopomers
of cis and trans isomers. (C), Formation of the PFTA derivative of adenosine (f ) at 100 °C for 3, 5, 16, and 24 h. Also shown
is the PFTA derivative of ribose (e ), which forms rapidly and demonstrates stability at this temperature. (D), Observed versus
predicted deuterium enrichment of ribose standards enriched with known amounts of 5,5-D2-ribose (0%–1%) for PFTA
derivative using m/z 515:m/z 513 ratio; APE is defined in Fig. 2. Line shows linear regression, slope 1.74, R2 ⫽ 0.99. (An
identical curve was obtained using the 435:433 ratio, data not shown). (E), Comparison of the abundance vs amount injected
of the PFTA (〫) and ATA (f ) derivatives of adenosine. (F), Curve of abundances according to starting cell number for RNA
analysis by PFTA. The horizontal line represents the approximate minimum abundance for reliable isotopic ratio estimation (⬃4
⫻ 106); this abundance corresponds to analysis of about 1 million cells.
flected almost exclusively the dynamics of rRNA because measurements were delayed until ⱖ3 days postlabeling to allow equilibration between vascular and
extravascular lymphoid compartments, a procedure
also used for DNA analysis (24 ). This selectivity was
confirmed by models of RNA dynamics (Fig. 5A),
Clinical Chemistry 55:10 (2009) 1829
suggests that these cells are metabolically active and
that, in contrast with PM1 cells in tissue culture, most
RNA production is not directly attributable to cell
division.
TURNOVER OF RNA IN LYMPH NODES
Fig. 4. Progressive labeling of RNA and DNA in PM1
cells grown in media labeled with deuterated glucose.
Results are expressed as fractional enrichments (F) of adenosine (RNA, ⽧) and deoxyadenosine (DNA, E), relative to
glucose enrichment, in cells growing in media labeled with
approximately 30% [6,6-2H2]-glucose. Modeling a monoexponential rise to plateau with variable time-delay gave a
mean (SEM) asymptote of 46.0% (1.5%) for RNA and
47.1% (2.2%) for DNA.
which demonstrated that mRNA and tRNA would be
unlikely to contribute to the total labeling curve beyond 3 days postlabeling.
In peripheral blood from 1 patient with chronic
lymphocytic leukemia, CD3– lymphocytes consisting
predominantly of B cells showed significant incorporation of deuterium into both RNA and DNA after labeling. Labeling curves for RNA and DNA followed similar patterns (Fig. 5B); RNA labeling reached higher
initial levels but then fell more rapidly, consistent with
faster turnover than DNA, and became undetectable by
day 21. Peak RNA labeling was consistent with a proportional synthesis rate (s) of about 7%/day, equivalent
to a doubling time (t2) of approximately 10 days (Fig.
5B). Labeled total RNA was lost at a rate (l*) of about
15%/day, equivalent to a mean half-life (t1/2) of about 5
days. The disparity between rates of synthesis and loss
derives from the fact that synthesis relates to the whole
RNA pool, whereas loss reflects only labeled RNA,
which may be lost at a rate more rapid than average if
the RNA pool is heterogeneous. When mRNA was removed from RNA fractions before analysis, similar, albeit slightly decreased, rates were obtained, about 6%
and 12% per day for s and l*, respectively (Fig. 5C).
This finding confirmed that mRNA makes only a small
contribution to labeling at these time points. The substantial difference between RNA- and DNA-labeling
rates (7.0% vs 2.7%/day, respectively; Fig. 5, B and D),
1830 Clinical Chemistry 55:10 (2009)
In diseased lymph nodes from 2 patients, substantial
rates of RNA turnover were measured. In a node infiltrated with lymphoma (patient A), total RNA production rate was estimated to be 9.6%/day, equivalent to a
doubling time (t2) of 7 days (Fig. 6A). In a node affected
by HIV-associated Kaposi sarcoma (patient B), higher
rates of RNA turnover were estimated; production rate
was 36.9%/day (Fig. 6B), equivalent to a doubling time of
about 2 days. In samples from both study patients RNA
turnover greatly exceeded DNA turnover (by 4- and
5-fold), estimated cell proliferation rates being 2.4%/day
(t2 29 days) and 7.7%/day (t2 9 days) respectively.
Discussion
This study demonstrates the proof-of-principle that in
vivo production and disappearance rates for RNA may be
estimated from rates of incorporation of deuterium from
labeled glucose in clinical studies. The use of a stable isotope avoids radiation exposure, in contrast to some previous approaches, and is safe for human studies. In
PBMC, a pulse-chase approach with delayed sampling
can be used to selectively target ribosomal RNA turnover.
Ribosomal RNA turnover is of interest for several
reasons. In lymphocytes, for example, it distinguishes
metabolic from proliferative activity; thus terminally
differentiated cells, which lack the ability to divide, may
appear quiescent on proliferative assays but may have
considerable metabolic activity, manifested as accelerated RNA turnover. Tumor cells may have different
metabolic activity from nonmalignant cells, and treatment may modify RNA metabolism (8, 9 ). Applying
this approach to leukemic cells, we demonstrate that
although such cells are relatively quiescent in proliferative terms, with a lifespan of months (16, 17 ), they
have significant metabolic activity (Fig. 5). Further interpretation of the RNA turnover rate of such leukemic
cells would be enhanced by comparison with other cell
types and lymphocytes in other settings.
The approach we describe may also be used in tissue samples to derive production rates for total RNA.
In diseased lymph nodes, rates of RNA labeling suggest
high metabolic activity. Comparison with DNA labeling suggests that most RNA production is not simply
due to the accumulation of newly divided cells. Because
the lag between labeling and measurement (⬍24 h) was
shorter in these experiments than in those with PBMC,
the measured results include short-lived RNA species
(mRNA, tRNA) as well as rRNA. Interpretation is lim-
Measuring RNA Turnover by Use of Deuterium-Labeled Glucose
Fig. 5. In vivo labeling of RNA and DNA in CD3ⴚ lymphocytes from a patient with chronic lymphocytic leukemia.
(A), Theoretical model for contribution of mRNA (lowest thin line), tRNA (next lowest thin line), and rRNA (bold dashed) to total RNA
labeling (bold), the sum of the 3 components. Calculations assume that RNA comprises 4% mRNA, 8% tRNA, and 88% rRNA, with
half-lives of 0.3, 0.6, and 7 days respectively, and that each compartment behaves as a single homogenous compartment with equal
synthesis and loss rate constants, k ⫽ s ⫽ l such that F(t) ⫽ (1 ⫺ e⫺kt ) during labeling (t ⱕ ␶), and F(t) ⫽ (1 ⫺ e⫺k␶) 䡠 e⫺k(t⫺␶)
postlabeling (t ⬎ ␶), with a labeling period (␶) of 10 h. (B–D), Experimental data showing the fraction of labeled total RNA (B), fraction
of labeled total RNA fraction remaining after removal of mRNA (C), and the fraction of labeled DNA (D), after orally administered
labeling with [6,6-2H2]-glucose. Solid lines show curve fits generated using the model described. Inset data show modeled synthesis
rates (s) for RNA and proliferation rates (p) for DNA together with loss rates (l*) for labeled RNA and disappearance rates for labeled
DNA (d*), as described. Error bars represent SDs of triplicate measurements by GC-MS.
ited by lack of comparative information from “normal” lymphoid tissues in these investigations and
thus they can be considered only pilot studies; such
information could be elucidated in further experiments. Despite this limitation, the feasibility of this
approach for clinical studies of tissue samples is
clearly demonstrated.
The analysis of deuterium-labeled RNA parallels
that used for DNA (10 –12 ), but different derivatization conditions are required, possibly because of
greater stability of the pentose ring of adenosine. Analysis of the ATA derivative of adenosine has previously
been described by Boros et al., who used [1,2-13C2]glucose labeling to investigate the intermediary metabolism of RNA in thiamine-responsive megaloblastic
anemia (25 ). This label is not appropriate for turnover
measurements because the 1-carbon is lost during
nonoxidative pentose production; the preserved
6-position is thus a preferable labeling site. Because
RNA labels through the de novo pathway via 5,5-2H2–
phosphoribose pyrophosphate, and because this pathway is more active for purines than pyrimidines, selective analysis of purines is advantageous. An allowance
for the contribution of nucleoside salvage or other
routes by which nonlabeled pentose moieties might enter the precursor pathway is made by correcting the
value used for precursor enrichment (12 ). An estimate
of this correction factor may be made by labeling cells
to saturation, as in the experiment shown in Fig. 4. The
value obtained in the experiment shown is the same for
both RNA and DNA, consistent with shared precursor
pathways. The reason why the asymptote does not
reach the previously documented value for DNA precursor dilution [0.6 – 0.65 (12 )] is unclear, and further
experiments to validate this value in both lymphoid
and nonlymphoid cells are indicated. Validation in
various cell types under differing conditions would enable wider application of this approach; for example,
Clinical Chemistry 55:10 (2009) 1831
Fig. 6. Estimated rates of production of DNA and
RNA in lymph node samples from patients with lymph
node disease.
Results show fraction of labeled total DNA and RNA in
biopsy samples after intravenous labeling with [6,6-2H2]glucose for 24 h in 2 patients with malignant infiltration of
lymph nodes, 1 with lymphoma (A) and one with HIVassociated Kaposi sarcoma (B). Error bars represent SD of
duplicate measurements by GC-MS; for DNA, replicates
were so close that error bars are too small to be visible.
“resting” lymphocytes may make greater use of the salvage pathway than activated cells (26, 27 ). However, it
should be noted that because DNA and RNA data are
corrected by the same dilution factor, any error would
not affect the comparisons of RNA and DNA turnover
rates shown in Figs. 5 and 6.
Choice of label is critical. For fast-turnover moieties such as RNA, the precursor must be rapidly incorporated then rapidly disappear. Glucose, with its short
in vivo half-life, is an ideal candidate and may be given
intravenously or orally (12, 21 ). A universal issue in
labeling studies is recirculation of label, which is likely
to be minimal for deuterated glucose because glucose is
rapidly metabolized and deuterium is primarily lost
into the very large body water pool after oxidation. Although heavy water (2H2O) is useful for labeling DNA
in slowly dividing cells, in which incorporation of deuterium takes place over weeks (14, 28 ), 2H2O would
not be suited to RNA analysis because of the relatively
high rates of RNA turnover; RNA pools would become
saturated with label and useful information obscured.
Reducing RNA turnover to a single parameter is,
of course, an oversimplification. Each RNA species,
having its own rate of synthesis and degradation, will
contribute to the overall pool labeling in proportion to
its abundance and labeling rate. The measured value is
therefore a composite, as illustrated in Fig. 5A. (Such a
model of 3 separate homogenous populations is still
excessively reductionist because considerable heterogeneity exists within RNA species.) Furthermore, heterogeneity within sampled cell populations must also
be considered; for example lymphoid samples may
1832 Clinical Chemistry 55:10 (2009)
contain nonmalignant lymphoid and myeloid cells, as
well as malignant cells, each with their own proliferation/turnover rate and metabolic profile. Nonetheless,
when samples are well defined and carefully analyzed,
biologically meaningful comparisons should still be
possible.
Further method development could focus on dissecting the relative contribution of different RNA species. We separated mRNA from total RNA from
chronic lymphocytic leukemia cells for analysis but did
not obtain a sufficient signal for reliable quantification
of mRNA alone, although we were able to measure the
turnover rate of the total RNA fraction remaining after
removal of mRNA. With more cells and/or greater analytic sensitivity, together with earlier sampling time
points, experiments with other RNA species would be
possible. In these studies, we used samples of 106–107
cells; the minimum required for reliable analysis will
depend on the RNA content of the cell-type of interest
but appears to be around 106 cells with the protocol
above. This sensitivity range would make PBMC subpopulations in individuals with normal peripheral
blood lymphocyte counts amenable to analysis.
These studies describe a generic approach that
may be tailored toward specific applications. We envisage this approach being used as a research investigative
tool in observational and interventional studies of disease mechanisms, rather than as a diagnostic test. Further protocol development would be determined by
the RNA species under investigation and the tissue/cell
of interest; in this study we have shown how analyzing
rRNA turnover in lymphoid tissues can yield novel insights into disease processes such as leukemia and
lymphoma.
Author Contributions: All authors confirmed they have contributed to
the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design,
acquisition of data, or analysis and interpretation of data; (b) drafting
or revising the article for intellectual content; and (c) final approval of
the published article.
Authors’ Disclosures of Potential Conflicts of Interest: Upon
manuscript submission, all authors completed the Disclosures of Potential Conflict of Interest form. Potential conflicts of interest:
Employment or Leadership: J. Defoiche, Fonds National de la Recherche Scientifique (FNRS); L. Lagneaux, FNRS; L. Willems, FNRS.
Consultant or Advisory Role: None declared.
Stock Ownership: None declared.
Honoraria: None declared.
Research Funding: FNRS.
Expert Testimony: None declared.
Role of Sponsor: The funding organizations played no role in the
design of study, choice of enrolled patients, review and interpretation
of data, or preparation or approval of manuscript.
Measuring RNA Turnover by Use of Deuterium-Labeled Glucose
Acknowledgments: We are grateful to Dr. L. Pickering and Dr. G.
Panayiotakopoulos for help with clinical studies in patients
with lymph node disease. We would like thank also Dr. B. Asquith for
advice on modeling and to P. Akanji for contributing to analysis.
References
1. Ross J. mRNA stability in mammalian cells. Microbiol Rev 1995;59:423–50.
2. Gillery P, Georges N, Wegrowski J, Randoux A,
Borel JP. Protein synthesis in collagen latticecultured fibroblasts is controlled at the ribosomal
level. FEBS Lett 1995;357:287–9.
3. Yi X, Tesmer VM, Savre-Train I, Shay JW, Wright
WE. Both transcriptional and posttranscriptional
mechanisms regulate human telomerase template RNA levels. Mol Cell Biol 1999;19:3989 –97.
4. Halle JP, Muller S, Simm A, Adam G. Copy number, epigenetic state and expression of the rRNA
genes in young and senescent rat embryo fibroblasts. Eur J Cell Biol 1997;74:281– 8.
5. Gerber G, Gerber G, Altman KI. Studies on the
metabolism of tissue proteins, I: turnover of collagen labeled with proline-U-C14 in young rats.
J Biol Chem 1960;235:2653– 6.
6. Cooper HL. Studies on RNA metabolism during
lymphocyte activation.Transplant Rev 1972;11:3–
38.
7. Grimble GK, Malik SB, Boza JJ. Methods for measuring tissue RNA turnover. Curr Opin Clin Nutr
Metab Care 2000;3:399 – 408.
8. Ruggero D, Pandolfi PP. Does the ribosome translate cancer? Nat Rev Cancer 2003;3:179 –92.
9. Rubin AD. Defective control of ribosomal RNA
processing in stimulated leukemic lymphocytes.
J Clin Invest 1971;50:2485–97.
10. Busch R, Neese RA, Awada M, Hayes GM, Hellerstein MK. Measurement of cell proliferation by
heavy water labeling. Nat Protoc 2007;2:3045–
57.
11. Macallan DC, Fullerton CA, Neese RA, Haddock
K, Park S, Hellerstein MK. Measurement of cell
proliferation by labeling of DNA with stable
isotope-labeled glucose: studies in vitro, in animals and in humans. Proc Natl Acad Sci U S A
1998;95:708 –13.
12. Macallan DC, Asquith B, Zhang Y, de Lara CM,
Ghattas H, Defoiche J, Beverley PC. Measurement
of proliferation and disappearance of rapid turnover cell populations in human studies using
deuterium-labelled glucose. Nat Protoc 2009;4:
1313–27.
13. Neese RA, Siler SQ, Cesar D, Antelo F, Lee D,
Misell L, et al. Advances in the stable isotopemass spectrometric measurement of DNA synthesis and cell proliferation. Anal Biochem 2001;298:
189 –95.
14. Neese RA, Misell LM, Turner S, Chu A, Kim J,
Cesar D, et al. Measurement in vivo of proliferation rates of slow turnover cells by 2H2O labeling
of the deoxyribose moiety of DNA. Proc Natl Acad
Sci U S A 2002;99:15345–50.
15. Chu A, Ordonez ET, Hellerstein MK. Measurement
of mouse vascular smooth muscle and atheroma
cell proliferation by 2H2O incorporation into
DNA. Am J Physiol Cell Physiol 2006;291:C1014 –
C1021.
16. Defoiche J, Debacq C, Asquith B, Zhang Y, Burny
A, Bron D, et al. Reduction of B cell turnover in
chronic lymphocytic leukaemia. Br J Haematol
2008;143:240 –7.
17. Messmer BT, Messmer D, Allen SL, Kolitz JE,
Kudalkar P, Cesar D, et al. In vivo measurements
document the dynamic cellular kinetics of chronic
lymphocytic leukemia B cells. J Clin Invest 2005;
115:755– 64.
18. McCune JM, Hanley MB, Cesar D, Halvorsen R,
Hoh R, Schmidt D, et al. Factors influencing T-cell
turnover in HIV-1-seropositive patients. J Clin Invest 2000;105:R1– 8.
19. Hellerstein M, Hanley MB, Cesar D, Siler S, Papageorgopoulos C, Wieder E, et al. Directly measured kinetics of circulating T lymphocytes in
normal and HIV-1-infected humans. Nat Med
1999;5:83–9.
20. Asquith B, Zhang Y, Mosley AJ, de Lara CM,
Wallace DL, Worth A, et al. In vivo T lymphocyte
dynamics in humans and the impact of human
T-lymphotropic virus 1 infection. Proc Natl Acad
Sci U S A 2007;104:8035– 40.
21. Ghattas H, Darboe BM, Wallace DL, Griffin GE,
Prentice AM, Macallan DC. Measuring lymphocyte kinetics in tropical field settings. Trans R Soc
Trop Med Hyg 2005;99:675– 85.
22. Crain PF. Preparation and enzymatic hydrolysis of
DNA and RNA for mass spectrometry. Methods
Enzymol 1990;193:782–90.
23. Asquith B, Debacq C, Macallan DC, Willems L,
Bangham C. Lymphocyte kinetics: the interpretation of labelling data. Trends Immunol 2002;23:
596 – 601.
24. Macallan DC, Asquith B, Irvine A, Wallace D,
Worth A, Ghattas H, et al. Measurement and
modeling of human T cell kinetics. Eur J Immunol
2003;33:2316 –26.
25. Boros LG, Steinkamp MP, Fleming JC, Lee WN,
Cascante M, Neufeld EJ. Defective RNA ribose
synthesis in fibroblasts from patients with
thiamine-responsive
megaloblastic
anemia
(TRMA). Blood 2003;102:3556 – 61.
26. Cohen A, Barankiewicz J, Lederman HM, Gelfand
EW. Purine and pyrimidine metabolism in human
T lymphocytes: regulation of deoxyribonucleotide
metabolism. J Biol Chem 1983;258:12334 – 40.
27. Reichard P. Interactions between deoxyribonucleotide and DNA synthesis. Annu Rev Biochem
1988;57:349 –74.
28. Voogt JN, Awada M, Murphy EJ, Hayes GM,
Busch R, Hellerstein MK. Measurement of very
low rates of cell proliferation by heavy water
labeling of DNA and gas chromatography/pyrolysis/isotope ratio-mass spectrometric analysis.
Nat Protoc 2007;2:3058 – 62.
Clinical Chemistry 55:10 (2009) 1833