From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
DNA Synthesis in Human Bone Marrow Is Circadian Stage Dependent
By Rune Smaaland, Ole D. Laerum, Knut Lote, Olav Sletvold, Robert 6.Sothern, and Robert Bjesknes
Fraction of human bone marrow (BM) cells in DNA synthesis
has been studied by sampling B M from the sternum or the
iliac crests every 4 hours during one 24-hour period in 16
healthy male volunteers. Three of the subjects underwent the
sampling procedure twice, resulting in 19 24-hour profiles.
The percentage of cells in DNA synthesis measured by flow
cytometry demonstrateda large variation along the circadian
time scale for each 24-hour profile, with a range of variation
from 29% to 339% from lowest to highest value. Seventeen
profiles (89.5%) had the highest DNA synthesis during waking hours between 08:OO hours and 2O:OO hours, and the
lowest percentage of cells in DNA synthesis between 0O:OO
hours and 0 4 : O O hours. TKe mean value of the lowest DNA
synthesis for each 19 2ehour period was 8.7% & 0.6%. while
the mean value of the highest DNA synthesis was 17.6% 2
0.6%, ie, a twofold difference. There was no difference in
DNA synthesis between winter and summer. A significantly
higher DNA synthesis was demonstrated for samples obtained from sternum as compared with the iliac crests, but
the same circadian pattern was demonstrated for both
localizations.By taking circadian stage-dependent variations
in DNA synthesis into account it may be possible to reduce
B M sensitivity to cytotoxic chemotherapy, to increase the
effect of hematopoietic growth factors as well as increase
the fraction of proliferating cells with careful selection of
time of day for harvesting B M cells for auto- or allografting.
o 1991 by The American Society of Hematology.
B
To our knowledge only two studies measuring the DNA
synthesis in human BM according to circadian stage have so
far been r e p ~ r t e d , ' ~in. ~one and four individuals, respectively.
Therefore, there has been an urgent need for a more
extensive study of a possible temporal variation in proliferative activity of the human BM. If large enough, such
temporal variations in BM cell proliferation could be of
clinical importance both relative to optimization of cytotoxic therapy and administration of hematopoietic growth
factors. The selection of time of day for harvesting BM cells
for auto- or allografting could possibly also be optimized.
We have conducted a study investigating the DNA
synthesis in human BM cells sampled several times during
19 24-hour periods in 16 healthy male subjects. BM cells
were aspirated by a standard technique used in the clinic,
and analysis of DNA content has been performed by flow
cytometry.
ONE MARROW (BM) suppression is commonly associated with cytotoxic treatment of cancer, and is
generally seen following combination therapy using different cytotoxic drugs.'.' It represents a major problem in
cancer chemotherapy, because therapeutic response usually requires drug doses inducing BM hypoplasia. The
cytotoxic effect on the BM is due to a potentially irreversible damage of pluripotent stem cells, early committed
progenitor cells, and proliferating cells later in the maturation process, as well as to regulatoly stroma cells in the BM
microenvironment? This sensitivity to cytotoxic therapy is
to a great extent related to the high proliferation rate of BM
cell~,4.~
although other mechanisms may be involved as well.
Acute BM suppression may not only lead to serious
infections, but also to dose reductions and postponement of
treatment courses, as well as reduced duration of useful
treatment. In addition, the possibilities of treatment in the
event of relapse may be reduced.
It is well documented that the susceptibility to cancer
chemotherapy shows circadian variations in laboratory
animals."8 In addition to reduced mortality due to acute
toxicity, it has also been shown that an increase in tumor
effect or cure rate can be ~ b t a i n e d , ~ or
. ~ .that
' ~ it is possible
to eliminate or reduce drug-induced death due to toxicity,
while still using an effective dose.I4 Circadian and circannual variations in proliferative activity in murine BM, both
regarding colony-forming unit granulocyte-macrophage
(CFU-GM), CFU-spleen (CFU-S), and DNA synthesis,
have also been shown."-*'
In addition, clinical studies have demonstrated a circadian dependence of cytotoxic drugs to BM toxicity, showing
less dose reductions, less treatment related complications,
and less postponements of drug courses when drugs have
been administered at certain
There are also a few
clinical studies either demonstrating or suggesting a reduced chance of relapse as well as increased long-term
survival when cytotoxic therapy has been administered at
specific times of the
These time-dependent variations in toxicity and survival
have not been generally recognized in practical-clinical
treat~nent.'~
This lack of recognition may partly be due to
the fact that there are few data on directly measured
biologic rhythms of proliferative parameters in human BM.
Blood, Vol77, No 12 (June 15), 1991: pp 2603-2611
MATERIALS AND METHODS
Subjects. From November 1986 to August 1988 we obtained
BM samples from 16 healthy male volunteers (mean age = 33.7
years; range 19 to 47 years) during 21 24-hour periods, ie, five
subjects underwent the sampling procedure twice. To find out if the
study was feasible, practically and ethically, the investigators
started out sampling on themselves. Therefore, two of the first
volunteers were MDs who had to do night-work. These two
From The Gade Institute, Department of Pathology, Department of
Oncology, and Department of Pediatrics, Haukeland Hospital, University of Bergen, Bergen, Norway; The Geriatric Department, The
Deaconess Hospital, University of Bergen, Bergen, Norway; and The
Rhythmometry Laboratory, University of Minnesota, Minneapolis.
Submitted December 13, 1990; accepted February 13, 1991.
Supported by the Norwegian Cancer Society and Michael Irgens
Flocks Legacy. RS. is a fellow of the Norwegian Cancer Society.
Address reprint requests to Rune Smaaland, MD, The Gade
Institute, Dept. of Pathology, Haukeland Hospital, University of
Bergen, 5021 BeKen, Norway.
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be hereby marked
"advertisement" in accordance with 18 U.S.C. section I734 solely to
indicate thisfact.
0 1991 by The American Society of Hematology.
0006-4971J91/7712-0015$3.
OOJO
2603
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
2604
subjects were omitted in the final analysis. After a pilot study of five
individuals, it was found acceptable to include more subjects in the
study, which was approved and performed in accordance with the
guidelines of the regional medical ethics committee. All volunteers
gave their informed written consent to enter the study, and all
individuals included in the analysis followed their regular diurnal
activity schedule with sleep at night for at least 3 weeks before the
experiment. The subjects continued their usual activities during the
study period in between times of sampling. They went to sleep after
the 0O:OO hour sample was taken, and were awakened once for the
04:OO sample. Their diurnal rhythm was validated by determination
of the cortisol level at every sampling point, which showed the
usual circadian pattern for all individuals, ie, high morning levels
and low evening levels.
Protocol. Following periost anesthesia, BM was obtained by
puncturing the sternum or one of the anterior iliac crests every 4
hours during a 24-hour period. To reduce the possibility that the
repeated puncture procedure itself would interfere with the
results, the start of the experiment was randomized to either 08:00,
1200, or 16:00hours, with the first time of sampling repeated at the
end of each study for a total of seven samplesiprofile. The sequence
of sampling from the three different anatomical sites was also
randomized. No premedication was given. To exclude that any
variations found could be attributed to sample dilution caused by
local bleeding at the puncture site, differential counts were
performed on smears from all individual samples. No samples had
to be discarded because of unacceptable large peripheral blood
admixture, ie, all smears were characteristic of BM (results not
shown).
Venous blood was also obtained from the subjects at the same
time as BM sampling to determine hematologic parameters (total
and white differential blood cell counts) and cortisol measurements. The blood was obtained as the initial procedure or
immediately after the anesthesia of periost before the BM puncture. In this way an artificially increased level of cortisol resulting
from the puncture procedure itself was avoided.”
Procedure for BM sampling and sample handling. The puncture
site was infiltrated with a local anesthetic (Lidocain, 20 mgiml;
Astra, Sweden). No other premedication was administered. After
BM (0.2 mL) was aspirated into a 2-cm3 syringe, one part of the
sample was used for routine smears, while one droplet was stained
directly for DNA flow cytometry (direct staining). Another droplet
was placed onto each of two tilted microscope slides, to let the
blood run down, thereby possibly increasing the fraction of marrow
elements, which were immediately removed from the cover slides
by a thin blade (made wet beforehand) of a knife and stained
(indirect staining). Thus, two parallel samples from the same site
were stained at each timepoint. Both samples of BM cells were
added to 2 mL of ice-cold staining solution consisting of ethidium
bromide, detergent, and RNAse according to the method described by Vindel@vv.’*The tubes were sealed and the solution
shaken before being placed in an ice bath for at least 10 minutes.
Flow cytometly. Both single cell suspensions were analyzed on a
Cytofluorograph 50 H (Ortho Diagnostic Systems, Inc, Westwood,
MA), interfaced to a Model 2150 Computer (Ortho). In the
cytogram obtained, both the peak and the area of the red
fluorescence signal were used for region-setting to discriminate the
(G1 + GO) doublets from the real G2 + M cells. Thus, the second
peak of the DNA histogram contained only the G2 + M cell
population. This procedure was performed because the G1 + GO
doublets may “contaminate” the G2 + M peak in the DNA
histogram, leading to errors in the relative distribution of the
different cell cycle phases. The total number of cells analyzed for
each sample was 3 to 4 x lo4. Computerized analyses of the cell
cycle distribution in the histograms were performed using the
constant function of the cell cycle analysis program, by which the
SMAALAND ET AL
+
percentages of cells in the G1 + GO, S, and G2 M phases were
~alculated.”.~‘
The mean coefficient of variation (CV) of the DNA
histograms was 3.3%.
Evaluation of fraction of cells in DNA synthesis (S-phase) was
performed by taking the mean value of the S-phase of the two
differently stained samples at each timepoint. In addition, the
direct, the indirect, and the maximum values at each timepoint
were recorded to more thoroughly evaluate the variation along the
24-hour scale. The maximum value obtained at each timepoint was
included in the analyses because it may possibly represent the BM
sample with highest fraction of proliferative cells.
Statistical analysis. Data were analyzed by Student’s t-test
(two-tailed; paired t-tests used for paired analyses of groups) and
one-way analysis of variance (ANOVA), using data both in original
units and as percentages of the individual mean DNA synthesis. In
addition, the individual data obtained for each way of evaluating
the DNA synthesis phase were analyzed for circadian rhythm by a
computerized inferential statistical method involving the fitting of a
24-hour cosine by the method of least squares (Cosinor analysi~)?~
The rhythm characteristics estimated by this method include the
mesor (rhythm-adjusted mean), the amplitude (half the difference
between minimum and maximum of fitted cosine function), and the
acrophase (time of peak value in fitted cosine function). A Pvalue
for rejection of the zero-circadian amplitude assumption was
determined on each data series. While the cosinor method may not
accurately represent the true characteristics of the actual timedependent variations if assymetries exist in a time-series,” the
procedure is nevertheless useful for assessing the presence of
peri~dicities.~’
Individual rhythm characteristics were summarized
for the group by population mean c ~ s i n o r .Spearman
~~
rank
correlation test was performed for testing the correlation between
the direct and indirect method of analyzing the DNA synthesis.
RESULTS
Circadian and circannual variation of DNA Jynthesis.
The value of fraction of cells in DNA synthesis of BM cells
harvested at each timepoint showed a large variation along
the circadian scale for all 19 24-hour periods (Table 1). This
finding was not explained by a corresponding variation in
distribution of proliferative cells as judged by differential
count of the BM smears at each timepoint, because there
was no direct covariation between these two parameters.
The range of change from lowest to highest DNA
synthesis value during the 24-hour spans varied between
29% and 339%, with a mean and median difference of
118.2%
18.4% and 102.9%, respectively. As shown in
Table 1, 17 of 19 series showed the highest DNA synthesis
between 08:OO hours to 20:OO hours according to the cosinor
analysis, ie, during daytime hours or early evening, and
correspondingly, the trough of DNA synthesis during late
evening and night. A complementary analysis showed that
17 of the 19 sampling periods had a lower mean fraction of
cells in DNA synthesis from 0O:OO to 04:OO hours as
compared with the mean DNA synthesis from 08:OO to
20:OO hours (P < .OOOl), and the DNA synthesis in the time
span from 20:OO to 04:OO hours was lower in 16 of 19 periods
as compared with the DNA synthesis from 08:OO to 16:OO
hours (P < .005). The mean value of the lowest and highest
S-phase was 8.7% 2 0.5% and 17.6% 2 0.6%, respectively,
ie, a difference of 102.3% or a twofold variation in DNA
synthesis depending on the time of measurement. Illustrative DNA histograms for two different subjects at two
*
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
2605
BONE MARROW DNA SYNTHESIS RHYTHM
Table 1. Circadian Variation in Fraction (%) of Cells in S-Phase and Result of Single Cosinor Analysis
Data Limits
(S-phase)
Series
Subject ID
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
AA
FLJ
MJ
RK
EK
SF
RB
RBJ #2
IK
BCS
RS #1
RS #2
OH
os #2
GW
KL #1
KL #2
ODL #1
ODL #2
Age
(v)
Parameters of 24-h Cosine Fit:
N of Data
Low
High
ROC (%)
Mesor 2 SE
7
7
7
7
7
6
6
7
6
6
7
7
7
7
7
7
7
5
7
9.3
8.4
5.2
10.6
7.7
6.8
10.4
9.2
9.2
8.1
8.4
6.7
9.3
14.4
5.0
6.9
9.3
7.0
14.0
17.9
19.2
22.8
18.4
18.4
14.4
14.4
22.6
13.7
13.5
14.9
20.5
19.6
18.5
18.7
14.0
17.2
16.2
19.7
91.9
129.9
338.5
73.9
139.0
111.8
39.1
146.4
49.2
66.7
77.4
206.0
111.4
28.9
276.8
102.9
84.4
131.4
40.7
14.1 f 1.1
11.9 t 0.8
13.8 t 2.1
14.4 f 1.3
12.5 t 0.7
11.2 f 1.5
12.9 f 0.5
14.1 f 1.7
12.4 f 0.6
10.9 f 0.9
12.5 t 0.7
12.9 t 1.5
13.7 t 1.8
16.4 t 0.4
12.4 f 2.0
11.9 f 1.0
11.9 f 1.1
12.7 f 2.1
16.2 2 0.9
19
23
24
25
28
30
31
31
31
33
34
35
35
39
39
42
42
46
47
Amplitude
?
SE
3.2 2 1.4
4.1 2 1.1
4.1 2 2.8
0.9 2 1.8
4.2 5 1.0
2.5 5 2.1
1.6 5 0.7
5.5 5 2.6
1.6 5 1.0
2.0 2 1.2
2.2 t 1.0
4.3 f 2.2
0.3 f 2.6
2.3 t 0.5
4.2 2 3.0
1.1 5 1.6
2.5 f 1.4
2.7 2 3.3
0.4 2 1.2
Acrophase'
04:55
08:14
11:21
13:58
13:53
09:06
04:41
14:21
15:43
14:54
13:39
12:23
07:45
11:39
18:30
10:56
17:43
17:24
09:46
Abbreviation: ROC, range of change from lowest to highest value.
*In hours and minutes after local midnight.
different timepoints (daytime and midnight) are shown in
Fig 1.
Although almost all subjects had their highest DNA
synthesis during daytime, differences in phasing along the
24-hour period between the subjects were observed, ie, the
I
1
SubjectRS
Subject BS
5
e
a
e
6
12.00 hours
a
4
c
Relative DNA content
V
00.00hours
-
4
c
Relative DNA content
Fig 1. DNA histograms for t w o different subjects for two timepoints along the 24-hour time scale (day and midnight). The two
peaks (2C and 4C) in each histogram designate the GO G1-phase
and G2
M-phase. The part of the histogram in between is the
S-phase. The height (ie, the area) of the S-phase expresses the
percentage of cells in DNA synthesis.
+
+
time of highest and lowest DNA synthesis differed to some
extent between the individual subjects. Six examples of
individual circadian stage-dependent variations of fraction
of cells in DNA synthesis are shown in Fig 2 to demonstrate
the slightly different phasing and the magnitude of variation
in intraindividual DNA synthesis. The individual mean
S-phase value of the 24-hour sampling period varied from
10.9% and 16.6%, ie, a difference of 52.3%. Due to this
interindividual difference, the data were also normalized
and expressed as percentage of the mean value. When
pooling the data for all subjects both relative to the mean
and highest S-phase values, a consistent pattern was seen,
with a statistically significant lower DNA synthesis around
midnight as compared with the day (Fig 3). The rhythm
characteristics for the different ways of calculating the
DNA synthesis data are depicted in Table 2. Due to
different phasing among the subjects, the difference between the lowest and highest values is smaller as compared
with the individual values. As can be seen from Table 2, the
circadian stage-dependent variation is statistically significant for all methods of evaluating the data, analyzed both
by ANOVA and the Cosinor method.
Because the time of sampling started either at 08:00,
12:00, or 16:OO hours, DNA synthesis for the pooled data
over 32 hours was evaluated (Fig 4). This makes it possible
to observe the DNA synthesis for two consecutive dayperiods, demonstrating highest values during daytime (with
a reproducible dip at midday) and lower late eveninghight
values in between.
No difference in DNA synthesis between winter, ie,
October to March (13.2 rfr 0.4; n = 13), and summer, ie,
April to September (13.3 f 0.6; n = 8) was observed,
including all 21 24-hour profiles.
DNA Jynthesisaccording to staining method. The procedure of letting the blood component of the BM run down
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
2606
SMAALAND ET AL
19
8
-
221
Subject FLI
18-
f
s
9 17m
'g
3
16-
.
d
15-
2
n
08
12
16
20
00
04
08
08
Time of t h e day (MET.)
12
16
20
00
04
08
Time of the day (M.E.T.)
Subject RS
0J
OJ
08
12
16
20
00
04
08
1
12
Time of the day (M.E.T.)
l6 1
16 20 00 04 08
Time of the day (M.E.T.)
12
201 Subject EK
Subject KL #I
2
I ,
16
20
00
04
08
12
16
16
Time of the day (M.E.T)
the cover slide and then analyzing the cell components
remaining on the c0ve.r slide (indirect staining method)
represents a method intended to increase the fraction of
proliferating cells of the BM aspiration sample. This is a
simple method to use both for conventional investigation of
. , .
20
Fig 2. DNA synthesis variation along the 24-hour time span
in six different subjects, sampling of BM being performed every 4 hours (N = 7 sampleslsubject). Results are expressed as
the mean of two parallel analyses. The time of starting the experiments was randomized to
08:OO. 12:OO. and 16:OO hours
(M.E.T. = mean European time).
Id
* .
00
04
OS
12
Time of the day (h1.E.T)
16
BM smears and for more specific investigations of a purer
BM sample.
Nearly the same pattern of circadian variation was seen
for the two ways of staining the cells, ie, the direct and the
indirect method (Fig 5). The fraction of cells in DNA
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
2607
BONE MARROW DNA SYNTHESIS RHYTHM
Mean Values
(Percent of Mean)
Mean Values
(Original Units)
15.5
f
2 14.5
-1
T
1201
T
P
13.5
n
Er 12.5
90
9 11.5
-
ANOVA for Time Effect:
F - 4 . 4 0 , ~4.001
0
10.5
12
16
Time of the day (M.E.T.)
08
04
00
20
I
00
20
Highest Values
(Percent of Mean)
Highest Values
(Original Units)
T
Fig 3. Circadian variation in human BM DNA synthesis in 19 =-hour periods from 16 clinically healthy
men (total N = 127). Timepoint means (absolute values and percentage) and standard errors of DNA
synthesis are depicted along the 24-hour time scale.
In addition, the highest DNA synthesis value measured of the two parallel samples is depicted correspondingly.
04
08
12
16
Time of the day (M.E.T.)
T
1204
90 -
-1
ANOVA for Time Effect:
F=5.2S,ppcU.001
F- 4.68, p ~ 0 . 0 0 1
11
7
04
08
12
16
Time of the day (M.E.T.)
00
00
20
08
12
16
Time of the day (M.E.T.)
04
20
using the mean value of the two methods was 13.2%f 0.3%
for the whole material.
DNA synthesis according to anatomical localization. We
found no statistical difference of the S-phase between the
right (n = 46) and left (n = 44) iliac crests, with overall
means being 12.2% f 0.5% and 13.1%& O S % , respectively
(P= .16). A significantly highex S-phase was observed for
samples obtained from the sternum (n = 51) as compared
with the iliac crests (n = 90); 14.5%f 0.5% versus 12.6% f
0.3%,respectively (P = .0015). Comparison of timepoints
by t-test showed a statistically significant difference between the two localizations for the samples obtained at
08:OO hours and 00.00 hours (P < .01 and P < .05, respec-
synthesis was slightly higher for each timepoint when the
indirect staining method was used. The difference was
significant only for two timepoints, at 08:OO hours and 0O:OO
hours; P < .01 and P < .001, respectively. However, when
comparing the paired data available for all timepoints
(n = 120), a highly significant difference was observed
between the two methods, with a larger fraction of cells in
DNA synthesis using the indirect staining method as
compared with the direct staining method, 14.2% f 0.3%
versus 12.7% 5 0.3%, respectively (P < .0001). A highly
significant correlation was found between the two methods
when comparing the two ways of BM sampling (r = .62;
P < .OOOl). The fraction of cells in DNA synthesis when
Table 2. Statistical Evaluation of Circadian Stage-DependentVariation of DNA Synthesis in Human BM
Analysis by:
ANOVA
Variable
Units
N
Mean value
Mean value
Highestvalue
Highestvalue
Direct method
Direct method
Indirect method
Indirect method
Original
% o f mean
Original
% of mean
Original
% of mean
Original
% of mean
127
127
127
127
122
122
113
113
Arithmetic Mean t
13.20 k 0.32
100.0 k 2.2
14.342 0.36
100.0 2 2.2
12.50 0.34
100.0 2.4
14.16 2 0.37
100.0 2 2.3
*
*
SE
Cosinor
Population Mean Cosinor Summary:
F
P
P
Mesor ? SE
Amp (95% limits)
0
3.70
4.40
4.68
5.25
2.70
3.69
2.29
2.45
,004
,001
<.001
<.001
,024
.004
.051
,038
,006
,004
13.06 2 0.35
,003
14.14
1.54 (0.44, 2.67)
11.9 (3.7, 20.4)
1.79 (0.64, 2.96)
12.6 (4.9, 20.5)
1.36 (0.21,2.53)
11.6 (1.6, 21.7)
1.49 (0.03, 3.02)
10.0 (0.3, 20.4)
1300
1304
1312
1312
1212
1212
1340
1344
,002
,019
.022
,045
,042
0.36
12.49 & 0.44
14.05 k 0.43
(95% limits)
(09:24,
(09:32,
(1O:OO.
(10:04,
(07:48,
(07:52,
(07:28,
(07:40,
16:OO)
16:04)
16:04)
16:08)
15:56)
16:OO)
17:56)
17:52)
Comparison of circadian results for different methods of estimation of DNA synthesis in 19 series obtained from 16 men. ANOVA, analysis of
variance across all timepoints using all data in original units or as percentages of mean; Cosinor, summary of indiv+dual 24-hour rhythm
characteristics by population mean cosinor using amplitude (Amp) in both oroginal units and percentages as of Mesor (Mesor, 24-hour
rhythm-adjusted mean); Acrophase (0)
reference, OO:OO, with sleepkest between 00:00-08:OO.The 95% limits for Amp and B giwm i f P I .@5from
cosinor analysis.
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
SMAALAND ET AL
.Tamplini Day 2
08
12
16 20 00 04 08 12
Time of the day(M.E.T.)
16
Flg 4. C i d k n varktion in human BM DNA svnthosb for 19
=hour p.riod.(total N = 127). Timepoint mwmand standard OWON
am given. The time r u l e b extended for 32 hours due to the different
times of starting each indMdualstudy 1text).
tivcly) with marginal statistical significance at 1 6 0
(P= .07).
To scc if thc higher proliferative activity in the samples
obtaincd from thc stcrnum could in part explain the
observed circadian rhythm, and to see if the rhythm at the
two localizations was thc samc, we pooled all thc samplcs
taken from stcrnum and iliac crests at cach timc point and
analyzed cach sitc scparatcly. Duc to the intcrindividual
diffcrcnccs in the mean S-phase, the data wcrc also
normalizcd. A statistically significant effcct of time was
found for the samplcs obtaincd from stcrnum whcn analyzcd as original valucs and as pcrccntage of the mcan, P =
. 0 3 and P = .02 (ANOVA), rcspcctivcly. When analyzed
.Sunphi Day 2
r
17
-
;
;
IS - ;
16-
I
14-
;
I
13-
;
I
12 - ;
II
-
I
I
;
I
IO-
;
.knqIiinRDUy2
I
OR
12
16 20
00 0.1
OR
I
'
12
16
Time of the day(M.E.T.)
Flg 5. Comprrhon oftmtiOn ofdh in DNA-
Kcordlng
to method of a ~ m n the
g BM m p k . (direct W-1or i n d l m t [-&I),
whh tho time rule extended for 32 hours.
by cosinor we found P = .02 both for original values and as
perccntagc of thc mean, with acrophasc at 1216 hours and
12:48 hours, respectively. The samc circadian pattcm was
scen for the DNA synthesis in the samples from thc iliac
crcst as for thc stcmum, although thc rhythm was lcss
pronounccd. The P valuc was .Mwhen analyzing thc crcst
data as percentage of the mean. Thc acrophasc was found
at 1247 hours (Fig 6).
The possibility that multiple punctures into the bonc
could havc an impact on DNA synthesis by virtue of stress
was considered. We therefore compared the level of the
stress rclatcd hormone cortisol at thc start and at the end of
thc sampling proccdure for cach individual, ie, 24 hours
apart. No statistically significant diffcrcncc was observed
(P = .97). Ncithcr was thcrc a statisticallysignificant diffcrcnce in perccntagc of cclls in DNA synthcsis samplcd 24
hours apart (P = .22).
In addition, any possible stress rclatcd effcct on overall
pattern of the poolcd data was minimized in the study
protocol by starting sampling at threc diffcrcnt clock hours.
DISCUSSION
There is now increasing evidence that any compromise of
dosage or dclays in treatment schedule diminish thc likelihood of canccr control or cure?'.ww On thc other hand, it
sccms likely that treatment of some cancers would improve
if doses of cytotoxic drugs could bc incrcascd.'" This is due
to thc fact that the efficacy of most antineoplastic drugs is
dosc intensity-dcpendcnt, ie, higher doscs ovcr a shorter
timc span increase thc responsc rates and proportion of
curcs." Thercforc, it hccomcs important to reducc the toxic
effccts to normal sensitive tissues, especially the BM.
Scvcral ways of circumventing this problem arc bcing
cxplorcd, such as BM transplantation (auto- or allografting),"' use of recombinant human hematopoietic growth
factors (granulocytc-macrophagccolony-stimulating factor
(GM-CSF]/granulocytc-CSF [ G-CSF]):"." and rctroviral
transfer of thc MDRl gcnc to primary hcmatopoictic
progcnitor cclls (murinc studies)." In this study we suggest
another approach, as we havc demonstrated large circadian
variations in fraction of BM cells in DNA synthesis in each
of 19 24-hour periods, which can be taken advantagc of
whcn trying to optimize trcatment of canccr paticnts.
A total of sevcn 4-hourly samples were obtained from
each individual, with a median differcnce bctween the
lowest and highest S-phasc valuc of 103%. Although there
wcrc intcrindividual diffcrenccs in circadian phasing, a
well-known phcnomenon for most rhythmic physiologic
parameters, a consistent and statisticallysignificant pattcrn
was manifested for the group as a whole whcn pooling the
data from all subjccts. The pcriod of lowest DNA synthesis
was found around midnight, and the highest DNA synthcsis
was found during the day. This pattcrn was statistically
validatcd for cach of four diffcrcnt ways of dctermining thc
DNA synthesis at each timcpoint, as prcsentcd in Tablc 2.
Bccausc the sampling period cxtended for 32 hours for the
poolcd data, it was possiblc to graph the DNA synthesis
over two consecutive daytimc periods. Higher DNA synthesis during daytimc on thc sccond day aftcr having bccn low
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
2609
BONE MARROW DNA SYNTHESIS RHYTHM
T
T
T
T
Fig 6. Circadian variation in
DNAsynthesis according to ssmpling site {sternum [-.I
or iliac
crests [-O-]) along the 24-hour
time scale.
01
Iliac Crests
I
-
-
00
04
08
12
16
Time of the day(M.E.T.)
during night corroborates the finding of a higher DNA
synthesis during the daytime of the first day. However, due
to different phasing between the subjects, the difference
between the lowest and highest value was smaller than that
found for the individual subjects. The average time of
lowest fraction of cells in DNA synthesis was computed to
be at around 01.00 hours (trough) by cosinor analysis, and
correspondingly, the highest fraction of cells in DNA
synthesis was found to be at around 13.00 hours (acrophase).
A potential problem is that DNA synthesis may vary as a
function of the site from which the BM sample is taken,
both with regard to anatomical site and within the BM of
the actual site.46Conceptually, one would regard the total
red BM as one organ, being affected by the same endogenous physiologic and hormonal factors, and thus making
site-dependent variations less important. This concept is
supported by earlier reported data by Dosik et a1: who
showed a very close correlation between DNA synthesis in
BM samples obtained simultaneously from right and left
iliac crests by biopsies. A good reproducibility, although
with larger individual variations, was also demonstrated by
bilateral simultaneous aspirations. In agreement with the
results of Dosik et al;’we found no significant difference in
S-phase between the right and left iliac crests. In the
present study sampling was also performed from sternum,
the BM from which showing a slightly higher S-phase. For
obvious reasons it was not feasible to take samples from
different localizations at the same time, and in our study
protocol we tried to minimize this potential site-related
problem by harvesting BM from different sites (ie, sternum
and iliac crests) at the same timepoints for different
individuals.
The circadian stage-dependent rhythm was most pronounced for the sternum samples, but the same pattern of
circadian variation was demonstrated for the samples from
the iliac crests. For both localizations, the time of acrophase and trough was at about the same time, and
accordingly corresponded to the acrophase and trough of
the combined data. This finding rules out the possibility of
/> d
00
04
08
12
16
20
20
Time of the day M.E.T.)
different sampling sites contributing to the observed circadian stage dependence of DNA synthesis.
Thus, the demonstration of nearly the same value of
DNA synthesis in the left and right iliac crests indicates
strongly that the total red BM must be looked on as a
functional entity, and the same circadian variation in the
BM of the sternum as of the iliac crests further corroborates
this functional homogeneity. The higher DNA synthesis
found in the sternum samples is most likely due to less
blood contamination of these samples. This finding implies
that the circadian stage-dependent variations demonstrated in this study may be even larger. It should be noticed
that the DNA synthesis values found during daytime in the
present study (14.3 f 0.4 [mean] and 15.6 f 0.5 [highest])
are not much different from DNA values obtained by
trephine and from filtered BM fragments, 15.3% and
16.5%, respectively, reported by Zbroja et aLa This strongly
suggests that the BM samples analyzed in our study are
representative of BM nucleated cells. The major reason for
this relatively high DNA synthesis in the aspirate may be
the small amount of BM harvested for &w cytometry
analysis.
The finding that there was no statistically significant
difference in cortisol and DNA synthesis measured 24
hours apart for the 19 24-hour periods contradicts the
possibility of a stress-induced circadian rhythm of the DNA
synthesis. In addition, a possible stress effect related to
sampling was minimized by starting sampling at three
different clock hours. These three factors, therefore, strongly
negate the notion of a stress-induced circadian rhythm of
the DNA synthesis.
Likewise, it was found that neither the differential count
of the BM smears nor a dilution effect of mature granulocytes leaving the BM compartment could explain the
circadian variation in DNA synthesis observed (results not
shown).
Our data are in good agreement with those of Mauer,
who found that [3H]TdR-labeled cells of the myeloid
lineage were clearly higher during the day as compared with
midnight in three of four individuals, and with a trend
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
2610
SMAALAND ET AL
towards lower DNA synthesis at midnight in the fourth
individual. In that study, BM was sampled four times at
different clock hours during a 42-hour time span, with 12 to
18 hours between time points.M Although the study of
Mauer was an in vitro study because exposing individuals to
the isotope was not feasible, the short exposure for 1 hour
after the sample was obtained should express the DNA
synthesis at the time of sampling. Killmann et al made
corresponding observations in one human volunteer in an
earlier study with regard to mitotic indices, demonstrating
an increase in this proliferative parameter from early
morning to late evening.29
The study reported here is thus a major expansion of
these two earlier studies, and is in addition an in vivo study
analyzing the proliferative activity at the exact time of
sampling, demonstrating a statistically significant reproducible circadian stage-dependent variation in DNA synthesis
of the human BM.
The potential importance of the data presented here is
underlined by the results recently published by U v i et al for
murine BMT demonstrating a corresponding circadian
variation in CFU-GM and DNA synthesis, with highest and
lowest values of these parameters in the activity and rest
span, respectively. The benefit of taking such rhythms into
consideration was demonstrated in the same study by a
circadian toxicity rhythm of the anticancer agent 4‘-0tetrahydropyranyl doxorubicin, demonstrating the lowest
toxicity when the CFU-GM and DNA synthesis were
lowest.
Therefore, by taking circadian stage-dependent variations in DNA synthesis into account it may be possible to
reduce BM toxicity of S-phase-specific drugs or drugs
having a major effect on DNA synthesis by administering
the drugs or the major dose of a continuous drug infusion
during the time of lowest proliferative activity, ie, late
evening or at night in diurnally active individuals. Cells in
the S-phase will then be less susceptible, and cells in the
GO/G1 phase will have more time for repairing damage
before entering into the S-phase. However, in this context
pharmacokinetic and pharmacodynamic properties of drugs
must also be considered.
An additional important aspect of these findings is that it
may be possible to increase the effect of biologic response
modifiers like GM-CSF and G-CSF by administering the
optimal dose at the time of greatest responsiveness of the
BM. This timed administration may increase their usefulness and effect, and possibly also reduce their side effects.
The data further suggest that it may be possible to increase
the fraction of proliferating cells with careful selection of
time of day for harvesting BM cells for auto- or allografting.
However, the existence of interindividual differences in
circadian time structure, as well as the possibility that
patients may have a different inherent time structure due to
their illness, suggests that BM sampling should be performed at the two timepoints of anticipated highest and
lowest DNA synthesis to verify this circadian pattern. This
verification is easily performed by the sampling procedure
described here, which is a routine clinical procedure. It is
important that a small amount of BM aspirate is harvested
to reduce the “contamination” of peripheral blood. Another way to individualize chronotherapy will be to monitor
marker parameters in peripheral blood related to BM cell
proliferative activity. Such studies are underway.
Indeed, one may speculate that in the years to come, the
optimization of doses, intervals and scheduling of doses,
and treatment courses of both existing and upcoming drugs
and biologic substances will become just as important as the
institution of novel drugs and biologic response modifiers.
The possibility of treating patients at an optimal circadian
time, interval, and schedule is today feasible and costeffective through programmable drug delivery systems.
ACKNOWLEDGMENT
We are indebted to the volunteer subjects in the study. The
authors gratefully acknowledge the skilful technical assistance of
Jan Solsvik, Gro OlderQy, and Dagny Ann Sandnes.
REFERENCES
1. Gale RP: Myelosuppressive effects of antineoplastic chemotherapy, in Testa NG, Gale RP (eds): Hematopoiesis. Long-Term
Effects of Chemotherapy and Radiation. New York, NY, Marcell
Dekker, 1988, p 63
2. Evans WE: Clinical pharmacodynamics of anticancer drugs:
A basis for extending the concept of dose-intensity. Blut 56241,
1988
3. Hryniuk W, Figueredo A, Goodyear M: Applications of dose
intensity to problems in chemotherapy of breast and colorectal
cancer. Semin Oncol14:3, 1987 (suppl4)
4. Lohrman H-P, Schreml W: Cytotoxic Drugs and the Granulopoietic System. New York, NY,Springer-Verlag, 1982
5. Pollak MN: Recombinant GM-CSF in myelosuppression of
chemotherapy. N Engl J Med 320:253,1989 (letter)
6. Haus E, Halberg F, Scheving LE, Pauly JE, Cardoso S, Kiihl
JFW, Sothern RB, Shiotsuka RN, Hwang DS: Increased tolerance
of leukemic mice to arabinosyl cytosine with schedule adjusted to
circadian system. Science 177230,1972
7. Scheving LE, Haus E, Kiihl JFW, Pauly JE, Halberg F,
Cardoso SS: Close reproduction by different laboratories of characteristics of circadian rhythm in 1-P-D-arabinofuranosykytosine
tolerance by mice. Cancer Res 36:1133,1976
8. U v i F, Hrushesky WJM, Blomquist C, Lakatua D, Haus E,
Halberg F, Kennedy BJ: Reduction of cisdiamminedichloroplatinum nephrotoxicity in rats by optimal circadian drug timing.
Cancer Res 42:950,1982
9. Kiihl JFW, Haus E, Halberg F, Scheving LE, Pauly JE,
Cardoso SS, Rosene G: Experimental chronotherapy with ara-C.
Comparison of murine ara-C tolerance on differently timed treatment schedules. Chronobiologia 1:316,1974
10. Scheving LE, Bums ER, Halberg F, Pauly JE: Combined
chronochemotherapy of L1210 leukemic mice using 1-P-D-arabinofuranosylcytosine, cyclophosphamide, vincristine, methylprednisolone, and cis-platinum. Chronobiologia 17:33,1980
11. Scheving LE, Burns ER, Pauly JE, Halberg F: Circadian
bioperiodic response of mice bearing advanced L1210 leukemia to
combination therapy with adriamycin and cyclophosphamide. Cancer Res 40:1511,1980
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
BONE MARROW DNA SYNTHESIS RHYTHM
12. Sothern RB, E v i F, Haus E, Halberg F, Hrushesky WJM:
Control of a murine plasmacytoma with doxorubicin-cisplatin:
Dependence on circadian stage of treatment. J Natl Cancer Inst
81:135, 1989
13. von Roemeling R, Hrushesky WJM: Determination of the
therapeutic index of floxuridine by its circadian infusion pattern. J
Natl Cancer Inst 82386,1990
14. Scheving LE, Bums ER, Pauly JE, Halberg F, Haus E:
Survival and cure of leukemic mice after circadian optimization of
treatment with cyclophosphamide and 1-p-D-arabinofuranosykytosine. Cancer Res 373648,1977
15. Stoney PJ, Halberg F, Simpson HW: Circadian variation in
colony-forming ability of presumably intact murine bone marrow
cells. Chronobiologia 2:319,1975
16. Scheving LE, Burns ER, Pauly JE, Tsai TH: Circadian
variation in cell division of the mouse alimentary tract, bone
marrow, and corneal epithelium, and its possible implication in cell
kinetics and cancer chemotherapy. Anat Rec 191:479,1978
17. Bartlett P, Haus E, Tuason T, Sackett-Lundeen L, Lakatua
D: Circadian rhythm in number of erythroid and granulocytic
colony forming units in culture (ECFU-C and GCFU-C) in bone
marrow of BDFl male-mice, in Haus E, Kabat H F (eds): Proceedings of the XV International Conference of the International
Society for Chronobiology. Basel, Switzerland, Karger, 1982, p 160
18. Aardal NP, Laerum OD: Circadian variations in mouse bone
marrow. Exp Hematol 11:792,1983
19. Haus E, Lakatua DJ, Swoyer J, Sackett-Lundeen L Chronobiology in hematology and immunology. Am J Anat 168:467,1983
20. Aardal NP: Circannual variations of circadian periodicity in
murine colony-forming cells (CFU-C). Exp Hematol12:61, 1984
21. Sletvold 0, Laerum OD: Multipotent stem cell (CFU-S)
numbers and circadian variations in aging mice. Eur J Hematol
41:230,1988
22. Sletvold 0, Laerum OD, Riise T: Age-related differences
and circadian and seasonal variations in myelopoietic progenitor
cell (CFU-GM) numbers in mice. Eur J Haematol40:42,1988
23. Hrushesky WJM: Circadian timing of cancer chemotherapy.
Science 228:73,1985
24. Kerr DJ, Lewis C, O’Neill B, Lawson N, Blackie RG, Newell
DR, Boxall F, Cox J, Rankin EM, Kaye SB: The myelotoxicity of
carboplatin is influenced by the time of its administration. Hematol
Oncol8:59,1990
25. E v i F, Benavides M, Chevelle C, Le Saunier F, Bailleul F,
Misset J-L, Regensberg C, Vannetzel J-M, Reinberg A, Mathe G:
Chemotherapy of advanced ovarian cancer with 4’-O-tetrahydropyranyl doxorubicin and cisplatin: A randomized phase I1 trial with
an evaluation of circadian timing and dose-intensity. J Clin Oncol
23705,1990
26. Rivard GE, Infante-Rivard C, Hoyoux C, Champagne J:
Maintenance chemotherapy for childhood acute lymphoblastic
leukemia: Better in the evening. Lancet 2:1264, 1985
27. Hrushesky WJM, von Roemeling R, Sothem RB: Circadian
chronotherapy: From animal experiments to human cancer chemotherapy, in Lemmer B (ed): Chronopharmacology. Cellular and
Biochemical Interactions. New York, NY, Marcel Dekker, 1989, p
439
28. Hrushesky WJM, von Roemeling R, Sothern RB: Preclinical
and clinical cancer chemotherapy, in Arendt J, Minors DS,
Waterhouse JM (eds): Biological Rhythms in Clinical Practice.
London, UK, Wright, 1989, p 225
29. Killmann S-A, Cronkite EP, Fliedner TM, Bond VP: Mitotic
indices of human bone marrow cells. I. Number and cytologic
distribution of mitosis. Blood 19:743, 1962
2611
30. Mauer AM: Diumal variation of proliferative activity in the
human bone marrow. Blood 26:1,1965
31. Ginsberg L, Ludman PF, Anderson JV, Burrin JM, Joplin
GF: Does stressful venepuncture explain increased midnight serum cortisol concentration? Lancet 2:1257, 1988
32. VindelBv LL: Flow microfluorometric analysis of nuclear
DNA in cells from solid tumors and cell suspensions. Virchows
Arch [B] 24:227,1977
33. Dean PN, Jett JH: Mathematical analysis of DNA distributions derived from flow microfluorometry. J Cell Biol60523, 1974
34. Gray JW, Dean PN: Display and analysis of flow cytometric
data. Annu Rev Biophys Bioeng 9509, 1980
35. Nelson W, Tong Y,Lee JK, Halberg F Methods for cosinor
rhythmometry. Chronobiologia 6:305,1979
36. Van Cauter E: Endocrine rhythms, in Arendt J, Minors DS,
Waterhouse JM (eds): Biological Rhythms in Clinical Practice.
London, UK, Wright, 1989, p 23
37. Kleveu RB, Shymko RM, Blumenfeld D, Braly PS: Circadian gating of S phase in human ovarian cancer. Cancer Res
47:6267,1987
38. Hryniuk W, Bush H: The importance of dose intensity in
chemotherapy of metastatic breast cancer. J Clin Oncol 2:1281,
1984
39. DeVita V Jr: Dose-response is alive and well (editorial). J
Clin Oncol4:1157,6986
40. Simon R, Korn E L Selecting drug combinations based on
total equivalent dose (dose intensity). J Natl Cancer Inst 82:1469,
1990
41. Moormeier JA, Williams SF, Kaminer LS, Garner M, Bitran
JD: High-dose tri-alkylator chemotherapy with autologous stem
cell rescue in patients with refractory malignancies. J Natl Cancer
Inst 82:29, 1990
42. Cheson BD, Lacerna L, Leyland-Jones B, Sarosy G, Wittes
RE: Autologous bone marrow transplantation. Ann Intern Med
11051, 1989
43. Bronchud MH, Scarfe JH, Thatcher N, Crowther D, Souza
LM, Alton NK, Testa NG, Dexter TM: Phase 1/11 study of
recombinant human granulocyte colony-stimulating factor in patients receiving intensive chemotherapy for small cell lung cancer.
Br J Cancer 56:809,1988
44. Antman KS, Griffin JD, Elias A, Socinski MA, Yan L,
Cannistra SA, Oette D, Whitley M, Frei E 111, Schnipper LE:
Effect of recombinant human granulocyte-macrophage colony
stimulating factor on chemotherapy-induced myelosuppression. N
Engl J Med 319593,1988
45. McLachlin JR, Eglitis MA, Ueda K, Kantoff PW, Pastan IH,
Anderson WF, Gottesman MM: Expression of a human complementary DNA for the multidrug resistance gene in murine hematopoietic precursor cells with the use of retroviral gene transfer.
J Natl Cancer Inst 821260,1990
46. Lord BI: The architecture of bone marrow cell populations.
Int J Cell Cloning 8:317,1990
47. Dosik GD, Barlogie B, Gohde W, Johnston D, Tekell JL,
Drewinko B: Flow cytometry of DNA content in human bone
marrow: A critical reappraisal. Blood 55:734,1980
48. Zbroja RA, Wass J, Vincent PC, Young GAR: Fragment
filtration: A method for the accurate determination of flow
cytometric kinetic data from bone marrow aspirates. Exp Hematol
14:85, 1986
49. Ltvi F, Blazsek I, Ferle-Vidovic A Circadian and seasonal
rhythms in murine bone marrow colony-forming cells affect tolerance for the anticancer agent 4’-O-tetrahydropyranyladriamycin
(THP). Exp Hematol16:696,1988
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
1991 77: 2603-2611
DNA synthesis in human bone marrow is circadian stage dependent
R Smaaland, OD Laerum, K Lote, O Sletvold, RB Sothern and R Bjerknes
Updated information and services can be found at:
http://www.bloodjournal.org/content/77/12/2603.full.html
Articles on similar topics can be found in the following Blood collections
Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests
Information about ordering reprints may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#reprints
Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.org/site/subscriptions/index.xhtml
Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American
Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.
Copyright 2011 by The American Society of Hematology; all rights reserved.