DNA synthesis in human bone marrow is circadian ...

From bloodjournal.hematologylibrary.org by guest on July 10, 2011. 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

Information about reproducing this article in parts or in its entirety may be found online at: http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#repub_requests Information about ordering reprints may be found online at: http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#reprints Information about subscriptions and ASH membership may be found online at: http://bloodjournal.hematologylibrary.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.


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 bloodjournal.hematologylibrary.org by guest on July 10, 2011. 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

*


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


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


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