From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
Colony-forming Unit-Granulocyte-Macrophage and DNA Synthesis of Human
Bone Marrow Are Circadian Stage-dependent and Show Covariation
By Rune Smaaland, Ole D. Laerum, Robert B. Sothern, Olav Sletvold, Robert Bjerknes, and Knut Lote
Bone marrow samples from sternum and iliac crests were
harvested every 4 hours during 19 24-hour periods from 16
healthy male volunteers, and myeloid progenitor cells were
cultured by the colony-forming unit-granulocyte-macrophage (CFU-GM) assay. A large interindividual variation was
observed in the mean number of colonies during each
24-hour period, with the highest 24-hour mean colony number being about 600% greater than the lowest (range:
16 & 2.3 t o 100.3 & 4.5). For each individual the difference
between the lowest and highest colony number throughout
the day ranged from 47.4% t o 256.3% of the mean colony
number of each series. A circadian stage-dependent variation
in the number of colony-forming units of myeloid progenitor
cells (CFU-GM) of human bone marrow was demonstrated,
with values 150% higher, on the average, during the day as
compared with the night. The overall data (891 CFU-GM
replicates) exhibited a significant 24-hour rhythm (P .001)
with an acrophase at midday (12.09 hours with 95% confidence limits from 10.32 t o 13.49 hours) and a trough at
midnight. This 24-hour variation was found t o covary with
DNA synthesis in the total proliferating bone marrow cell
population. A seasonal effect on CFU-GM numbers was
detected by ANOVA (P = .014) and by the least squares fit of
a 1-year cosine (P = .015), with the highest number found in
summer. The potential relevance of these findings should be
examined in relation t o cytotoxic cancer therapy, use of
hematopoietic growth factors, and bone marrow transplantation.
0 1992by The American Society of Hematology.
B
ONE MARROW suppression is commonly associated
with cytotoxic treatment of cancer, and is generally
seen following combination therapy using different cytotoxic drugs.1.2 Not only may it lead to serious infections and
bleeding tendency, but it may also force dose reductions
and/or 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
limited. Thus, the bone marrow-dependent side effects can
lead to a suboptimal treatment of cancer patient^.^-^
The cytotoxic effect on normal bone marrow cells is
caused by the potentially irreversible damage to stem cells
and early committed progenitor cells, as well as to regulatory stroma cells in the bone marrow mi~roenvironment.~
In
laboratory animals it has been well documented that the
susceptibility to cancer chemotherapy of many normal
tissues, including bone marrow, shows circadian variat i o n ~ . ~In
- ’ ~addition, time-dependent toxicity to the bone
marrow has also been shown in cancer patients. Doxorubicin and tetrahydropyranyl-doxorubicinexert a lower toxicity to the myeloid lineage when administered in the early
morning, as compared with late afternoon,11J2 whereas
carboplatin demonstrates its lowest toxicity toward the
megakaryocyte lineage when administered in the late after-
have investigated a possible circadian variation in circulating human myeloid progenitor cells (CFU-GM). Highest
colony number was found in the morning (at 9:OO when
sampled every 3 hours around-the-clock) in nine healthy
men17 or in the late afternoon at 3:OO (with sampling only at
8:OO AM, 11:OO AM, and 3:OO PM).’~
The circadian rhythmic variations in bone marrow drug
cytotoxicity, as well as bone marrow proliferative parameters, have not been generally recognized or implemented in
practical treatment,ll which may, in part, be attributed to
the fact that the earlier reported bone marrow data have
been too limited to allow extrapolation to the population at
large. There are also no reports on human bone marrow
CFU-GM rhythmicity.
Because of this paucity of data concerning the potentially
important circadian stage-dependence of proliferative activity of human myeloid progenitor cells, especially in relation
to cytostatic cancer therapy and hematopoietic growth
factors, we conducted a study investigating possible circadian variations in myeloid progenitor cell proliferation in a
CFU-GM assay in 16 healthy male volunteers during 19
24-hour periods, with bone marrow samples being obtained
310011.~3
From The Gade Institute, Department of Pathology, Haukeland
Hospital, University of Bergen; the Division of Geriatric Medicine, the
Department of Public Health and Primary Health Care, The Deaconess Hospital; the Department of Pediatrics, Haukeland Hospital; the
Department of Oncology, Haukeland Hospital, University of Bergen,
Bergen, Norway; and the Rhythmomehy Laboratory, the Department
of Surgery, University of Minnesota, Minneapolis.
Submitted November 12,1991; accepted January 3,1992.
Supported by grants from the Norwegian Cancer Society. R.S. is a
Fellow of the Norwegian Cancer Society.
Address reprint requests to Rune Smaaland, MD, The Gade
Institute, Department of Pathology, Haukeland Hospital, University of
Bergen, 5021 BeTen, 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 1734 solely to
indicate this fact.
0 1992 by The American Society of Hematology.
0006-497119217909-0012$3.00/0
Circadian stage-dependent variations in proliferative
activity in human bone marrow have, until now, been
largely unknown. We have recently reported a circadian
stage-dependent variation in DNA synthesis of total bone
marrow cells in humans, with bone marrow being harvested
every 4 hours around-the-clock during 19 24-hour peri0ds.1~
Our results were consistent with earlier findings on DNA
synthesis of human bone marrow myeloid cells (by tritiated
thymidine technique) in four volunteers, each providing
four samples (at noon, 6:OO PM, midnight, and 6:OO AM)^^
and two reports of the mitotic activity in six subjects
(sampled at noon, 6:OO PM, midnight, and 6:OO AM)'^ and
one subject (sampled at 1:20 PM, 3:00 PM, 7:OO PM, 10:45 PM,
and 5:45 AM),'^ which indicated lower DNA synthesis and
mitotic activity during late evening hours (near midnight)
and during the night. To our knowledge, only two studies
Blood, Vol 79, No 9 (May 1). 1992: pp 2281-2287
2281
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
2202
SMAALAND ET AL
at 4-hour intervals. In addition, we correlated the number
of myeloid colonies with the DNA synthesis of total
nucleated bone marrow cells in each sample.
MATERIALS AND METHODS
Subjects. Between November 1986 and August 1988 we obtained hone marrow samples from 16 healthy male volunteers
(mean age, 33.7 years; range, 19 to 47) during 19 24-hour periods,
ie, three subjects underwent the sampling procedure twice.14 To
find out if the study was feasible, the investigators started taking
samples from themselves. All volunteers gave their informed
written consent to enter the study, which was approved by and
performed in accordance with the guidelines of the regional
medical ethics committee. All individuals followed their regular
diurnal activity schedule with sleep at night for at least 3 weeks
before the experiment, following a diurnal activity routine with
average time of arising at 7:OO AM and of retiring at 11:OO PM.
Protocol. During each study period, the subjects continued
their usual activities between sampling times. They went to sleep
after the midnight sample was taken, and were awakened once for
the 4:OO AM sample.
To reduce the possibility that the repeated puncture procedure
itself would interfere with the results, the start of each experiment
was randomized to either 8:00 AM, noon, or 4:OO PM,with the first
time of sampling repeated at the end of each study for a total of
seven samples per profile. The sequence of sampling from the three
different anatomical sites was also randomized. To exclude that the
variations should 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 bone marrow (results not
shown).
Venous blood was also obtained from the subjects at the same
time as bone marrow sampling to measure the cortisol level and
determine other hematologic parameters (total number and differential count of white blood cells). The blood was obtained as the
initial procedure or immediately after the anesthesia of periost and
before the bone marrow puncture. In this way, an artificially
increased level of cortisol resulting from the puncture procedure
itself was a ~ 0 i d e d . The
l ~ subjects’ circadian rhythm was validated
by determination of a cortisol level at every timepoint of sampling,
which showed the usual 24-hour pattern for all individuals, ie,
cortisol levels high in the morning and low in the evening.
Bone marrow sampling. The puncture site was infiltrated with a
local anesthetic (Lidocain 20 mg/mL; Astra, Sodertalje, Sweden).
No other premedication was given. Following periost anesthesia,
bone marrow was obtained by puncturing the sternum or one of the
anterior iliac crests every 4 hours during a single 24-hour period.
The marrow was aspirated into each of two syringes, one for bone
marrow culturing in soft agar for estimating growth of CFU-GM,
the other for analysis of fraction of nucleated bone marrow cells in
DNA synthesis by flow cytometry.
CFU-GMassay. Myelopoietic colonies were cultured using the
single layer method described by Burgess et al.*O Human placenta
conditioned medium (HPCM) was used as colony-stimulating
factor according to the method of Schlunk and Schleyer2l Before
plating in agar, bone marrow cells were suspended at a concentration of 1 x lo6 cells per mL in Dulbecco’s modification of Eagle’s
medium with 2% penicillin, streptomycin, and glutamine, and 3.2%
of nonessential amino acids. From a mixture of 2.5 mL medium, 1.0
mL 20% fetal calf serum (Flow, Irvyne, Scotland), 0.5 mL HPCM,
0.5 mL of cell suspension, and 0.5 mL 3% agar (Agar Noble Difco,
Detroit, MI), aliquots of 1mL with 1 x lo5cells were plated per 35
mm dish (Nunclon, Roskilde, Denmark), with 8 dishes per timepoint. The dishes were incubated for 13 days at 37°C in 5.0% COz.
A colony was defined as more than 50 cells. All colonies were
counted blindly without knowledge of the actual timepoint of
harvesting.
Flow qtometry. The bone marrow samples for flow cytometric
analyses were handled and analyzed as described earlier.14Briefly,
two parallel bone marrow single cell suspensions were analyzed on
a Cytofluorograf 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
M cells. Thus, the second peak of the DNA
from the G2
histogram contained only the G2 + M cell population. The total
number of cells analyzed for each sample was 3 to 4 x 104.
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 percentages of cells in the G1 GO,
S, and G2 M phases were calculated. The mean CV (coefficient
of variation) of the DNA histograms was 3.3%.
Statistical analysis. Data were analyzed by Student’s t-test
(two-tailed; paired t-tests used for paired analysis of groups) and
one-way analysis of variance (ANOVA), using data both in original
units and normalized to percent of mean for both CFU-GM and
DNA synthesis. In addition, the individual data for CFU-GM and
DNA synthesis were analyzed for circadian rhythm by the fitting of
a 24-hour cosine by the method of least-squares (Cosinor analysis).22The rhythm characteristics estimated by this method include
the mesor (rhythm-adjusted mean), the amplitude (half the difference between the minimum and maximum of the fitted cosine
function), and the acrophase (time of peak value in fitted cosine
function). APvalue for rejection of the zero-amplitude assumption
is determined for each data series. Although the cosinor method
involving a single fitted period may not accurately represent the
true characteristics of the actual time-dependent variations if
asymmetries exist in a time-series,z the procedure is nevertheless
useful for objectively assessing and quantifying periodicities selected a prioriz4 The Spearman correlation test was used for
analyzing the relationship between CFU-GM and DNA synthesis.
+
+
+
RESULTS
Interindividual difference in CFU-GMnumber. Significant
interindividual variations were observed in the mean number of colonies during each 24-hour period, with the highest
24-hour mean colony number being 627% greater than the
lowest (range of mean number of colonies: 16.0 f 2.3 [SEI
in series 5 to 100.3 f 4.5 in series 8) (Table 1).
Circadian variation in CFU-GM number. A large circadian variation in CFU-GM was observed for each individual, with an intra-individual difference between the lowest
and highest colony number ranging from 47.4% (series 8) to
256.3% (series 5) about the mean colony number of each
24-hour series (mean difference = 136.1% f 12.5%) (Table l). Overall, the mean CFU-GM at midnight was 38.3 f
2.5 colonies, as compared with the highest mean value of
59.7 -r- 3.1 at noon, ie, 156% higher (Fig 1). When
expressing the CFU-GM as percent of mean, the corresponding difference was 170% between the observed trough
at midnight and the observed macroscopic peak at 4:OO PM
(Fig 2). Irrespective of hour, the mean value across individual series for the lowest CFU-GM number was 13.1 f 3.1,
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
2283
CIRCADIAN CFU-GM AND DNA SYNTHESIS COVARIATION
Table 1. Circadian Variations in Myeloid Progenitor Cells (CFU-OM) in Human Bone Marrow and Result of Analysis for Time Effect by ANOVA
and Single Cosinor
~~
~
Analysis of Replicates by
Series
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Subject
ID
AA
FLJ
MJ
RK
EK
RBJ
SF
IK
RB
BCS
RSI
RS2
OH
os1
os2
GW
KLI
KL2
ODL
(Totals)
Data Limits
Age
(yr)
No. of
Data
Low
High
19
23
24
25
28
30
30
31
31
33
34
35
35
38
39
39
42
42
46
55
47
56
36
42
30
56
38
48
49
62
42
44
59
54
42
43
39
49
891
24.1 f 1.7
37.9 f 2.2
4.0 f 0.4
10.6 f 1.3
5.0 f 2.0
30.0 f 3.1
3.5 f 0.4
73.4 f 5.2
13.1 f 1.5
19.0 f 1.3
31.9 f 1.6
41.0 f 20.0
1.9 f 0.3
9.9 f 2.1
37.7 f 1.8
30.0 f 1.5
5.8 f 0.8
10.1 f 2.4
5.3 f 1.3
1.9 f 0.3
131.1 f 5.4
159.4 f 6.8
82.9 f 3.8
160.9 f 13.1
46.0 f 12.8
53.2 f 4.6
36.5 f 2.6
120.9 f 3.8
68.6 f 4.9
70.4 f 2.3
86.9 f 9.3
93.0 f 4.5
122.6 f 5.4
51.1 f 4.4
80.4 f 2.7
119.3 f 6.8
108.1 f 2.8
35.3 f 3.5
104.3 f 3.2
160.9 f 13.1
ROC'
1%)
Arithmetic
Mean f SE
162.4 65.9 f 4.8
154.4 78.7 f 7.3
187.0 42.2 f 4.1
158.9 94.6 f 9.9
16.0 f 2.3
256.3
55.6 41.7 f 2.3
162.6 20.3 f 1.8
47.4 100.3 f 4.5
151.6 36.6 f 3.2
122.1 42.1 f 2.7
92.2 59.7 f 3.4
81.5 f 2.8
63.8
190.1 63.5 f 6.4
131.2 31.4 f 2.4
72.9 58.6 f 2.8
99.6 89.7 f 4.9
199.4 51.3 f 6.3
127.3 19.8 f 2.0
153.7 64.4 f 4.9
291.2
54.6 f 1.3
F
46.2
111.8
122.1
92.8
17.3
4.8
66.0
7.7
57.2
33.9
34.9
8.0
202.0
33.5
27.4
19.8
245.9
3.7
83.7
5.8
P
<.001
< .001
< ,001
< ,001
< ,001
,003
< ,001
< .001
< .001
< .001
< ,001
< ,001
< ,001
< .001
< .001
< .001
< .001
.003
< .001
<.001
Parameters of 24-h Cosine Fit
Cosinor
P
Mesor 2 SE
Amplitude 2 SE
Acrophase
<.001
,871
< .001
< .001
.052
648
< .001
.004
,192
<.001
<.001
.060
< .001
< .001
< .001
.002
< .001
.094
< .001
<.001
67.6 f 3.1
78.9 f 7.4
38.6 f 3.2
103.3 f 8.6
18.3 f 2.5
41.2 f 2.5
21.4 f 1.6
105.3 f 4.2
36.3 f 3.3
42.3 f 2.0
62.4 f 3.3
80.7 f 2.6
65.4 f 4.5
30.9 f 2.0
55.9 f 1.6
90.4 f 4.3
45.5 f 4.2
18.7 f 2.1
61.2 f 4.2
53.8 f 1.3
42.0 f 4.6
5.6 f 10.6
26.9 f 4.3
52.5 f 11.5
9.8 f 4.0
3.3 f 3.6
10.4 2 2.2
23.2 f 6.5
8.4 f 4.6
20.5 f 3.2
16.4 f 3.8
10.0 f 4.2
43.1 f 6.2
14.1 f 2.9
23.8 f 2.3
23.4 -t 6.2
41.4 f 5.6
5.9 f 2.7
28.3 f 5.9
8.3 f 1.8
08.25
08.04
13.08
16.47
21.51
03.36
03.01
15.13
04.27
16.28
18.08
03.03
16.34
15.01
10.41
13.23
10.03
19.08
10.46
12.09
'ROC, range of c h a n g e t h e difference from lowest to highest value expressed as percent of the mean. tANOVA, analysis of variance across all
timepoints using all data in original units. Cosinor, individual 24-h rhythm characteristicsfrom the least squares fit of a 24-h cosine. Acrophase
reference, 00.00, with sleep/rest between 00.00-08.00.
T
130-
T
6
/
N of dishes counted
381
34
I
00
04
08
12
16
Time of day (clock hour)
20
Fig 1. Circadian variation in human bone marrow CFU-OM. Timepoint means and standard errors from 19 series in 16 clinically healthy
men.
00
04
08
12
16
Time of day (clock hour)
20
Fig 2. Circadian covariation in human bone marrow CFU-GM and
DNA synthesis. Timepoint means and standard errors from 19 series
of CFU-GM and DNA synthesis in 16 clinically healthy men.
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
SMAALAND ET AL
2284
Table 2. Statistical Evaluation of Circadian Stage-dependent Variation of Myeloid Progenitor Cells ICFU-GM) in Human Bone Marrow
Analysis by
Variable
Units
No.
Arithmetic
Mean -c SE
F
P
P
Both sites
Crest only
Sternum only
Both sites
Crest only
Sternum only
No. of colonies
No. of colonies
No. of colonies
YOof mean
YOof mean
% of mean
891
569
322
891
569
322
54.5 2 1.3
48.7 f 1.5
64.9 2 2.3
100.0 2 1.9
89.6 f 2.3
118.4 f 3.1
5.84
6.73
4.23
13.23
7.42
12.41
<.001
<.001
<.001
<.001
<.001
<.001
<.001
Characteristics of 24-h Cosine Fit
Cosinor
.012
.016
<.001
<.001
,001
~
Mesor
?
SE
Amplitude
53.8 f 1.3
48.4 2 1.5
63.2f 2.4
98.6 f 1.9
90.2 f 2.3
115.7 f 3.2
?
SE
8.3f 1.8
6.0f 2.0
10.7 f 3.7
18.6 f 2.7
19.3 f 3.3
18.1 f 4.9
0
(95% limits)
12.09
11.22
11.22
13.10
14.04
10.30
(10.32,13.44)
(08.40,14.04)
(09.12,13.32)
(12.08,14.12)
(12.52,15.16)
(08.48.12.12)
Comparison of circadian results for different sites of estimation of CFU-GM in 19 series obtained from 16 men.
Abbreviations: ANOVA, analysis of variance across all timepoints using all data in original units or as percent of mean. Cosinor, 24-hrhythm
charactaristicsfrom analysis by single cosinor using data in both original units and as percent of mean. Acrophase (0)
in clock hour and minute from
reference (00.00 hours), with sleeplrest between 00.00 and 08.00hours.
and for highest CFU-GM the mean was 115.5 f 10.5, ie,
almost 900% (882%) higher.
Each CFU-GM series showed a highly statistically significant effect of circadian stage when analyzed by ANOVA.
Fourteen of 19 series (73.7%) demonstrated a significant
24-hour rhythm at P I .05 in CFU-GM from the leastsquares fit of a 24-hour cosine, while one additional
individual showed borderline significance (P = .06). The
acrophase for 12 of the 14 series was during the day
between 8:30 AM and 500 PM. The overall data (891
CFU-GM replicates) exhibited a significant 24-hour rhythm
(P < .001) with an acrophase at midday (12:09 PM with
95% confidence limits from 10:32 AM to 1:49 PM), and
trough at midnight. Cosinor analysis of CFU-GM as percent of mean also yielded significant results (P < .001),
with similar acrophase and trough (Table 2 ) .
CFU-GM number according to anatomical site. Colony
replicates from the sternum and iliac crests were also
analyzed separately for circadian rhythm. Whereas a significant 24-hour rhythm for both anatomical sites of bone
marrow harvesting could be shown with comparable times
for acrophase and trough, a statistically significant difference in mesor value was found between the sternum (63.2
colonies) and iliac crests (48.4 colonies) (P < ,0001).
Relation between CFU-GM number and DNA synthesis.
We also compared the 24-hour rhythm of CFU-GM with
the rhythm in DNA synthesis measured in the total bone
marrow. Three individual examples of circadian covariation
150
A
between CFU-GM and DNA synthesis with different sampling start times are shown in Fig 3. In Fig 2 the combined
pattern for CFU-GM and DNA synthesis is shown, each
normalized to percent of mean for comparison. Here, the
mean value of CFU-GM (as opposed to all replicates) at
each circadian stage for each individual is used for a more
exact comparison with the single DNA synthesis observation at each timepoint. Both variables show a statistically
significant circadian stage-dependent variation by ANOVA
and cosinor analysis as percent of mean, and they show a
very close circadian covariation, with acrophase (and 95%
limits) at 1:33 PM (11:oo AM to 4:08 PM) and 12:45 PM (1052
AM to 2:40 PM) for CFU-GM and DNA synthesis, respectively (Table 3). A corresponding 24-hour variation was
significant by ANOVA when using original units for both
CFU-GM and DNA synthesis, although cosinor analysis
was not statistically significant for CFU-GM because of the
wide range of nonnormalized mean individual colony numbers (see above). No ultradian (12-hour) component could
be detected for the group in either variable. In addition, by
normalizing the absolute values for CFU-GM and DNA
synthesis to percent of mean, a highly significant correlation
between these two proliferative parameters could be shown
(P < .OOOl).
Circannual variation in CFU-GM number. We found
that the number of colonies of myeloid progenitor cells also
varied according to the time of year, with higher levels in
summer (57 f 7, n = 27) than in winter (32 f 6, n = 22).
:;Bb
v-’a.
4
“0
I110
90
1 yQ-
70:
50
.0--
50
. & CFU-OM
1 -
-
--0
S-phase
I --“
-
g
,’
A
7
1 w
1,:‘
50
O
6
tJ
-
-
m
Fig 3. Individual circadian covariation in human bone marrow CFU-GM and DNA synthesis in healthy adult men. A comparison of sampling
start times. For comparison, values for CFU-GM and DNA synthesis (percent of cells in S-phase) reexpressed as percent of series mean. [A) Start
8:OO AM, July 2 to 3.1987; subject OS, age 39. ( 6 )Start noon, March 22 to 23,1988; subject MJ, age 24. (C) Start 4:W PM, August 26 to 27,1988;
subject AA, age 19.
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
2285
CIRCADIAN CFU-GM AND DNA SYNTHESIS COVARIATION
Table 3. Statistical Evaluation of Circadian Stage-dependent Covariation of Myeloid Progenitor Cells (CFU-GM) and DNA Synthesis 1% cells in
S-phase) in Human Bone Marrow
Analysis by
Variable
CFU-GM
DNA
CFU-GM
DNA
Units
No. of colonies
YOin S-phase
YOof mean
% o f means-phase
Characteristics of 24-h Cosine Fit
Cosinor
No.
Arithmetic
Mean 2 S E
F
P
P
Mesort SE
121
127
121
127
54.4 f 3.4
13.2 f 0.3
100.0 f 4.7
100.0 f 2.2
0.79
3.70
2.72
4.40
,557
,004
,023
.001
.232
.002
.021
<.001
53.7 & 3.4
13.1 f 0.3
98.8 4.7
99.4 f 2.1
*
Amplitude
2
SE
8.3 f 4.8
1.6 & 0.4
18.9 f 6.7
11.5 f 3.0
0
(95% limits)
11.41
12.28
13.33
12.45
(10.24, 14.32)
(11.00, 16.08)
(10.52, 14.40)
(4
Comparison of circadian results of concomitant estimation of CFU-GM and DNA in 19 series obtained from 16 men.
Abbreviations: ANOVA, analysis of variance across all timepoints using timepoint means in original units or as percent of mean. Cosinor, 24-h
rhythm characteristicsfrom analysis by single cosinor using data in both original units and as percent of mean. Acrophase (0)
in clock hour and
minute from reference (00.00 hours), with sleephest between 00.00-08.00’Kours.
An effect of season was detected by ANOVA (P= .014)
and by the least-squares fit of a 1-year cosine (P= .015).
The time of estimated high values as indicated by the
acrophase was August 12, with 95% confidence limits from
July 8 to September 14.
DISCUSSION
Dose is a critical determinant of the efficacy of cancer
chemotherapy.z Increasing evidence indicates that any
compromise of dosage or delay in the treatment schedule
diminishes the likelihood of cancer control or cure for
drug-sensitive t ~ m o r s . ~ - This
~ , ~ ~may
, ~ 7be explained by the
fact that the effect of most antineoplastic drugs is dose
intensity-dependent, ie, higher doses over a shorter time
span increase the response rates and proportion of cures.
On the other hand, toxic effects to normal tissues limit the
amount of dose that can be administered, with bone
marrow suppression at present being the main doselimiting factor. Therefore, searching for and improving
therapeutic selectivity is critical.
We show in this study a significant circadian stagedependent variation in the colony number of human bone
marrow myeloid progenitor cells, ie, CFU-GM, which has
the same circadian patterning as DNA synthesis of the total
number of bone marrow nucleated cells. The statistically
determined peak (acrophase) of CFU-GM was found
between 8:OO AM and 7:OO PM, ie, during daytime and early
evening, in 14 (73.7%) of 19 profiles from 16 individuals
(Table 1). When pooling all individual data, the differences
become smaller because of interindividual differences in
waveform and phasing. However, the circadian variations
are, to a large extent, predictable, with the highest number
of CFU-GM and highest DNA synthesis occurring during
daytime hours (near noon), while the lowest values occur
around midnight. A huge intraindividual circadian stage
dependent variation in CFU-GM number was found, the
range of change around the mean 24-hour value from the
lowest to highest CFU-GM number of each individual
throughout the day averaging almost 140%. This difference
in the capacity of producing myeloid progenitor cells
throughout the day, as well as the large interindividual
variation observed in the 24-hour mean number of colonies,
might possibly help to predict bone marrow sensitivity for
an individual to, eg, a chemotherapeutic agent.
A possible intrinsic variability in bone marrow proliferation according to localization of sampling site was addressed by analyzing the data from the iliac crests versus the
sternum separately. Earlier it has been shown that core
biopsies obtained simultaneously from both iliac crests give
a very good reproducibility of cytokinetic data,= thereby
indicating a functional homogeneity with regard to proliferation status at different anatomical bone marrow sites. In
accordance with the study by Dosik et a1,= no significant
difference in DNA synthesis was found between the two
iliac crests by us.14However, a significantly higher number
of cells in DNA synthesis was measured in samples obtained from the sternum as compared with the iliac crests.I4
This was also the case for CFU-GM numbers in the present
study. The most likely explanation for this observation is
that there is less blood contamination of the sternum
sample. However, the same circadian pattern of variation
for both variables was observed in both sites and when
individual samples were pooled irrespective of sites. For
each site and variable, highest values were found during the
day and lowest during the night, with a very close relation in
timing of acrophase and trough for the two anatomical sites.
Our data are in good agreement with those published by
L6vi et a1 for murine bone marrow.29 The highest and
lowest number of CFU-GM were found during the activity
and rest spans, respectively, corresponding to the times of
highest and lowest DNA synthesis in mice. Further, these
times corresponded to the times of highest (midactivity)
and lowest (midrest) susceptibility of the anticancer agent,
4’-O-tetrahydropyranyl doxorubicin, which has its main
effect on DNA synthesis.
If healthy tissues exhibit regular circadian variations of
cell proliferation and cancer cells do not, or have a different
circadian pha~ing,”-~O
simple timing of cytotoxic drug administration should improve the efficacy and reduce the toxicity
of cancer chemotherapy, since these drugs are nonselective
in damage done to both healthy and cancerous tissues. It
may then be possible to reduce bone marrow suppression of
cytotoxic drugs exerting their main effect on proliferating
cells by administering the drugs or the major dose of a
continuous drug infusion during the time span of lowest
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
SMAALAND ET AL
2286
proliferative activity of healthy tissue, such as bone marrow
cells, ie, in late evening or during sleep by night (for
diurnally active individuals). Cells in the DNA synthesis
phase will then be less susceptible and cells in the G1 phase
will have more time for repairing damage before entering
into the DNA synthesis phase. The possibility of treating
patients at an optimal circadian time, interval and schedule
is today feasible and cost-effective through programmable
drug delivery s y s t e m ~ . * The
~ J ~ circadian stage-dependent
synchrony between S-phase and numbers of functional
myeloid progenitor cells (CFU-GM), ie, their clonability,
strengthens the possibility of optimizing therapy because
this finding suggests a common circadian phasing of both
myeloid progenitor cells and more mature proliferating
cells.
Another potentially important aspect of these findings is
that it may be possible to increase the effect of biological
response modifiers like granulocyte-macrophage colonystimulating factor, granulocyte colony-stimulating factor,
and interleukin-3. Administration of these substances at
the time of greatest responsiveness of the bone marrow
relative to proliferative circadian rhythms may more effectively accelerate the regeneration of granulocytelthrombocyte numbers after conventional or high-dose cytotoxic
therapy, with or without autologouslallogeneic bone marrow transplantation. As hematopoietic growth factors may
be potentially beneficial for patients with acquired immunodeficiency syndrome, myelodysplastic syndrome, or aplastic anemia:* an optimal use of these factors relative to
biological rhythms may also increase their usefulness in
these disease entities. Yet another point to be made is that
the time of harvesting bone marrow cells for bone marrow
transplantation could possibly be optimized by taking these
proliferative rhythms into consideration.
Therefore, we suggest that the circadian stage-dependent variations of myeloid progenitor cells (CFU-GM), as
well as the circadian variation in cell cycle distribution of
bone marrow cells, be further exploited in an attempt to
reduce bone marrow suppression of cytotoxic drugs by their
proper timing, thereby increasing the dose intensity and
efficacy of these agents.
ACKNOWLEDGMENT
We are indebted to the volunteer subjects in the study. The
authors gratefully acknowledge the skillful technical assistance of
J. Solsvik, G. Olderfiy, and D.A. Sandnes, and we thank Professor
E. Haus for critical review of the manuscript.
REFERENCES
1. Lohrman H-P, Schreml W Cytotoxic Drugs and the Granulopoietic System. New York, NY,Springer-Verlag, 1982
2. Pollak MN, Brennan LV, Antman K, Elias A, Cannistra SA,
Socinsky MA, Schnipper LE, Frei E, Griffin JD: Recombinant
GM-CSF in myelosuppression of chemotherapy. N Engl J Med
320:253, 1989 (letter)
3. Frei E 111, Canellos GP: Dose: A critical factor in cancer
chemotherapy. JAMA 69:585,1980
4. DeVita V Jr: Dose-response is alive and well (editorial). J
Clin Oncol4:1157,1986
5. Hryniuk WM, Figueredo A, Goodyear M: Applications of
dose intensity to problems in chemotherapy of breast and colorectal cancer. Semin Oncol14:3,1987 (suppl4)
6. Simon R, Korn E L Selecting drug combinations based on
total equivalent dose (dose intensity). J Natl Cancer Inst 821469,
1990
7. Haus E, Halberg F, Scheving LE, Pauly JE, Cardoso S, Kuhl
JFW, Sothern RB, Shiotsuka RN, Hwang DS: Increased tolerance
of leukemic mice to arabinosyl cytosine with schedule adjusted to
circadian system. Science 177230,1972
8. Scheving LE, Haus E, Kiihl JFW, Pauly JE, Halberg F,
Cardoso SS: Close reproduction by different laboratories of characteristics of circadian rhythms in 1-p-D-arabinofuranosylcytosine
tolerance by mice. Cancer Res 36:1133,1976
9. E v i F, Hrushesky W, 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
10. Boughattas NA, Levi F, Hecquet B, Lemaigre G, Roulon A,
Fournier C, Reinberg A Circadian time dependence of murine
tolerance for carboplatin. Toxic01 Appl Pharmacol96:233,1988
11. Hrushesky W J M Circadian timing of cancer chemotherapy.
Science 228:73,1985
12. E v i F, Benavides M, Chevelle C, Le Saunier F, Bailleul F,
Misset J-L, Regensberg C, Vannetzel J-M, Reinberg A, Math6 G:
Chemotherapy of advanced ovarian cancer with 4’-tetrahydropyra-
nyl doxorubicin and cisplatin: A randomized phase I1 trial with an
evaluation of circadian timing and dose-intensity. J Clin Oncol
8:705,1990
13. 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
14. Smaaland R, Laerum OD, Lote K, Sletvold 0, Sothern RB,
Bjerknes R: DNA synthesis in human bone marrow is circadian
stage dependent. Blood 77:2603,1991
15. Mauer AM: Diurnal variation of proliferative activity in the
human bone marrow. Blood 26:1,1965
16. 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
17. Ross DD, Pollack A, Akman SA, Bachur NR: Diurnal
variation of circulating human myeloid progenitor cells. Exp
Hematol8:954,1980
18. Verma DS, Fisher R, Spitzer G, Zander AR, McCredie KB,
Dicke KA: Diurnal changes in circulating myeloid progenitor cells
in man. Am J Hematol9185,1980
19. Ginsberg L, Ludman PF, Anderson JV, Burrin JM, Joplin
G F Does stressful venepuncture explain increased midnight serum cortisol concentration? Lancet 2:1257,1988
20. Burgess AW, Wilson EMA, Metcalf D: Stimulation by
human placental conditioned medium of hemopoietic colony
formation by human marrow cells. Blood 49:573,1977
21. Schlunk T, Schleyer M The influence of culture conditions
on the production of colony-stimulating activity by human placenta. Exp Hematol8:179,1980
22. Nelson W, Tong Y, Lee JK, Halberg F Methods for cosinor
rhythmometry. Chronobiologia 6:305,1979
23. Van Cauter E: Endocrine rhythms, in Arendt J, Minors DS,
Waterhouse JM (eds): Biological Rhythms in Clinical Practice.
London, UK, Wright, 1989, p 23
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
CIRCADIAN CFU-GM AND DNA SYNTHESIS COVARIATION
24. Klevecz RB, Shymko RM, Blumenfeld D, Braly PS: Circadian gating of S phase in human ovarian cancer. Cancer Res
47:6267,1987
25. Cheson BD, Lacerna L, Leyland-Jones B, Sarosy G, Wittes
RE: Autologous bone marrow transplantation. Ann Intern Med
110:51,1989
26. Hryniuk W, Bush H The importance of dose intensity in
chemotherapy of metastatic breast cancer. J Clin Oncol 21281,
1984
27. Evans WE: Clinical pharmacodynamicsof anticancer drugs:
A basis for extending the concept of dose-intensity. Blut 56241,
1988
2287
28. 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
29. E v i F, Blazcek I, FerlB-Vidovic A. Circadian and seasonal
changes in murine bone marrow colony forming cells affect
tolerance for 4’tetrahydropyranyladriamycin.Exp Hematol16:696,
1988
30. Smaaland R, Lote K, Laerum OD, Vokac Z A circadian
study of cell cycle distribution in non-Hodgkin’slymphomas. Annu
Rev Chronopharmacol5:383,1987
31. NienhuisW Hematopoietic growth factors. Biologic complexity and clinical promise (editorial). N Engl J Med 318:916, 1988
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
1992 79: 2281-2287
Colony-forming unit-granulocyte-macrophage and DNA synthesis of
human bone marrow are circadian stage-dependent and show
covariation
R Smaaland, OD Laerum, RB Sothern, O Sletvold, R Bjerknes and K Lote
Updated information and services can be found at:
http://www.bloodjournal.org/content/79/9/2281.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.