From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
Variations of Fat Tissue Fraction in Abnormal Human Bone Marrow Depend Both
on Size and Number of Adipocytes: A Stereologic Study
B y Ciril Rozman, Joan-Carles Reverter, Evarist Feliu, Lluis Berga, Maria Rozman, and Carme Climent
Studies dealing with the number or size of individual
adipose cells in abnormal human bone marrow are lacking.
To ascertain whether variations in fat tissue fraction
depend on the size of individualadipocytes or their number
or both, a stereologic study of 30 human bone marrow
specimens ( I O with aplasia, 10 with hyperplasia, and 10
with dysplasia)was performed. A total of 23,435 adipocyte
profiles were measured and two stereologic parameters
were obtained in each specimen: mean diameter and
number of cells per mm3 of bone marrow. The fat tissue
fraction correlated positively with the sire ( r = .79;
P < .001) and the number/volume ( r = .77; P < .OOl) of
adipocytes. The significance of both adipocyte size and
adipocyte number/volume was confirmed by stepwise
multiple regression, in which the size alone explained
62.5% of fat tissue fraction and both size and number/
volume explained 95.8% of fat tissue fraction. These
results are discussed from a pathophysiologic point of
view.
0 1990 by The American Society of Hematology.
A
group was represented by 10 biopsy specimens corresponding to
LTHOUGH adipose tissue is a major structural compopatients with myelodysplastic syndromes: 4 with refractory anemia
nent of the bone marrow, it has been relegated during
with excess of blasts (RAEB), 4 with RAEB in transformation, 1
many years to a purely passive role as a space filler in
with
sideroblastic refractory anemia, and 1 with 5q- syndrome.
response to hematopoietic variations.’ It seems clear that fat
Biopsy
specimens were taken from the anterior iliac crest with
cells are actively involved in the hematopoietic process,
Jamshidi needle. Plastic (methyl-methacrylate) embedding of undeserving as a direct metabolic support tissue rather than as a
calcified cores was used.16 Sections 3 pm in thickness were stained
purely mechanical
There is increasing interest in
with Gill’s hematoxilin.” From each specimen, 8 to 10 random
studying the changes of the marrow adipose cells in different
microphotographs were obtained and enlarged to final magnification
pathologic conditions, as far as their m o r p h ~ l o g i c a l , ~ ~ of
~ x 150 for the stereologic analysis. The magnification was confunctional,’ and biochemical”.’ I features are concerned.
trolled by means of a microscopic calibration device, Carl Zeiss
When the bone marrow adipose tissue fraction is studied in
5 + 100/100mm.
The stereologic analysis was performed with a digitizer (91 11A
humans, its variations are expressed in percentage with
Graphic Tablet, Hewlett Packard) connected to a computer (9845B,
respect to the marrow surface or
Investigations
Hewlett Packard). Specific software was developed for this research.
on number or size of individual adipose cells in human bone
The minimum sample size for each group of micrographs was
marrow are scant. In a previous stereologic s t ~ d y , ’we
~
determined by using the progressive mean technique with a confishowed that age-related variations of fat tissue fraction in
dence limit of +5%.’* A total of 23,435 adipocyte profiles were
normal human bone marrow depend both on size and number
measured with a mean of 781.2 per specimen. Because even small
of adipocytes. The purpose of the present study is to analyze,
adipocytes are easily identified in plastic embedded sections, no
by stereologic methods, whether the variations of marrow fat
special stain was used for this purpose. On the other hand, as detailed
tissue fraction in abnormal bone marrow are due to changes
later, a compensation for “optically lost caps” was routinely pereither in size or number, or both, of the fat cells.
formed.
MATERIALS AND METHODS
Thirty bone marrow specimens were studied: 10 with aplasia, 10
with hyperplasia, and 10 with dysplasia. As a source of aplastic bone
marrow, biopsy specimens obtained from 10 patients with aplastic
anemia were used. The hyperplastic group was composed of 10
biopsy specimens obtained from patients with chronic myeloproliferative disorders: 5 with essential thrombocytemia, 3 with chronic
granulocytic leukemia, and 2 with polycythemia vera. The dysplastic
From the Postgraduate School of Haematology “Farreras
Valenti’*>Hospital Clinic. University of Barcelona: Escuela T6cnica
Superior de Ingenieros de Caminos, Canales y Puertos; and Polytechnic University of Catalunya, Barcelona, Spain.
Submitted February 22,1990; accepted May I I , 1990.
Supported by Grant No. PB-86.0593 from the Direccibn General
de Investigacibn y Ciencia. Ministerio de Educacibn y Ciencia.
Address reprint requests to Professor C. Rozman. Postgraduate
School of Haematology. Hospital Clinic, 170 Villarroel. 08036
Barcelona, Spain.
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 1990 by The American Society of Hematology.
0006-4971/90/7605-0020$3.00/0
892
In each specimen, the following morphometric parameters were
determined: (1) percent of marrow fat tissue fraction; (2) mean
adipocyte diameter (6);(3) number of adipocytes/volume unit
(NJ, expressed per cubic millimeter.
Percentage of marrow fat tissue fraction was calculated by
fractional area (AA)measurement. The area of each marrow space
was measured by direct digitizing. The fat tissue area within each
marrow space was obtained from the sum of areas corresponding to
adipose cells. The latter were estimated from tr a/2 b/2 where a and
b are diameters of each adipocyte profile.
For the calculation of 6,we used the Giger and Riedwyl method’’
suitable for objects of spherical shape and normal distribution. Mean
profile axial ratio of adipocytes ranged from 1.24 to 1.35, limits
within which the objects can reasonably be treated as spherical.*’
The adequacy of the Giger and Riedwyl method was further
confirmed by the fact that adipocyte diameter did not significantly
depart from gaussian distribution.
Individual adipocyte profile diameters were calculated from
2(ab)’12 and plotted as a histogram of 10 to 20 size classes. To
compensate for “optically lost caps”, the histogram waszxtended to
the left by linear extrapolation. The mean diameter (d) from the
provisionally completed histogram was calculated, and a first estimate of the mean diameter was made: 6, = 4d/tr. From the
nomogram devised by Giger and Riedwyl,” a refined estimate D2
was derived. This was used for a more accurate estimation of the
smaller size classes obtained by linear extrapolation. If N is the total
number of transections measured, then the number N, in the k‘h
Blood, Vol 76,No 5 (September l), 1990:pp 892-895
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
893
STEREOLOGY OF HUMAN BONE MARROW ADIPOCYTES
Table 1. Results in Aplasia, Hyperplasia, and Dysplasia
Parameters
Group
Fat Tissue
Fraction(%)
Mean (SEMI
Mean Adipocyte
Diameter (cm)
Mean (SEM)
Adipocytes/mm3
Mean (SEMI
A. Aplasia (N = 10)
E. Hyperplasia (N = 10)
C. Dysplasia ( N = IO)
71.10 (3.78)
16.39 (1.60)
22.32 (5.05)
62.77 (2.47)
50.04 (1.23)
51.67 (1.98)
5,219.77 (622.99)
2,151.09 (200.81)
2,528.09 (422.01)
All patients
36.61 (5.01)
54.83 (1.52)
3,299.65 (359.01)
AvBP<.001
A v C P C ,001
A v 8 P < ,001
A V C P = .003
8 v C P = NS
BvCP=NS
No. of
A V B P = .001
A V C P = .003
BvCP=NS
Abbreviation: NS, not significant.
where K is a constant depending on the size distribution of the
particles, and B is a shape constant. N, corresponds to number of
profiles per section area, and V, to the volume fraction. Instead of
using the Weibel and Gomez” nomogram, the constant was computed according to Lindberg and Vorwerk.’’
To evaluate the relative contribution of N, and 6 to the variations
of fat tissue fraction, the ratio N,/n was estimated in each case. As
controls,data from our previous work15were used.
Statistical methods. Significanceof correlation coefficients and
difference between means were evaluated by t statistic^.'^ Contribution of different variables to the percentage of fat tissue fraction was
analyzed by means of stepwise multiple regressionz4using the
software 15010 on a 9845B computer (Hewlett Packard). Adequacy
of fitting model was controlled by graphic representationof standardized residuals.
higher fat tissue fraction group, both mean adipocyte diameter and number/volume were significantly lower in the
former as compared with the latter group (Table 2). The fat
tissue fraction correlated positively with the mean diameter
(r = .79; P < .001) and the number/volume (r = .77;
P < .001) of adipocytes. Thus, with increasing fat tissue
fraction, there is an increase in both adipocyte size and
number, whereas the reverse is also true. A stepwise multiple
regression was performed using the fat tissue fraction as a
dependent variable and the mean size of adipocytes @) and
their number/volume (N,) as independent variables. The
first parameter to enter the regression a t a significant level
was the mean size of adipocytes (D)and their number/
volume (N,) as independent variables. The first parameter to
enter the regression at a significant level was the mean size of
adipocytes (D),
which explained (R’)62.5% of the dependent variable. The number/volume (N,) also contributed to
the regression significantly. The fitted equation: fat tissue
fraction (percent) = - 104.0114 + 2.61293 (D)+ 0.00838
(N,) explained (R’)95.8%of the dependent variable.
The study of relative contribution of adipocyte number
and size to the variations of fat tissue fraction by means of
Nv/D ratio showed (Table 3) that this was significantly
increased in aplasia with respect to normal controls, whereas
no significant difference was observed in cases of hyperplasia
and dysplasia.
RESULTS
DISCUSSION
The results are presented in Table 1 . The values of all three
parameters (fat tissue fraction, mean adipocyte diameter,
and number of adipocytes/mm3) were significantly higher in
aplasia as compared with hyperplasia and dysplasia, whereas
the differences between the latter two groups were not
significant. When the cases were classified in lower and
An accurate histomorphometry of bone marrow relies on
two factors: (1) plastic embedding of undecalcified specimens, and (2) use of stereologic techniques. It has been
shown that paraffin embedding of bone marrow can lead to
considerable shrinkage, amounting to approximately 1 5%,25
producing an under-estimation of the hematopoietic tissue
small size class may be calculated from
Nk = dk aN/(&)’,
where a = width of a histogram block, and d, = mean diameter of
transections in the kth block. From this completed and rEfined
histogram the analysis was repeated and fresh estimates of N, d, 6,,
and 6, were obtained. This second n2was considered as the final
mean adipocytediameter (6).
For the determination of parameter N, we used the method of
Weibel and Gomez”:
Table 2. Results. Comparison Between Lower and Higher Fat Tissue Fraction
Fat tissue fraction <50% (N = 19)
Fat tissue fraction 250% (N = 1 1)
Mean Adipocyte
Diameter (pm)
Mean (SEM)
Adipocytes/mm3
Mean (SEMI
50.3 (1.06)
62.65 (2.24)
P < .001
2,238.14 (229.50)
5,124.08 (571.42)
P < .001
No. of
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
894
ROZMAN ET AL
Table 3. Relative Contribution of Adipocyte Number and Size t o
Variations of Fat Tissue Fraction
Group
Ratio NJD
Mean ISEM)
A. Aplasia (N = 10)
8. Hyperplasia (N = 10)
C. Dysplasia (N = 10)
D. Control (N = 20)
87.20 (13.71)
43.57 (4.5)
48.94 (8.47)
47.88 (2.06)
A vD, P = .02; B vD, NS; C vD. NS (NS, not significant).
fractionz6 and a disruption of the limits between bone
trabeculae and marrow spaces. From the lack of such
disruption in plastic-embedded undecalcified specimens, it
can be assumed that the shrinkage is of no major importance
in these conditions.
To obtain reliable histomorphometric measurements, the
use of stereologic methods is mandatory. If these are not
used, an important underestimation or overestimation of
different parameters may result. For example, calculation of
the mean diameter of adipocytes directly from transected
profiles would lead to a striking underestimation of this
parameter. A plane cut through the volume produces a
section containing circular profiles whose diameters range
from the maximum for equatorial sections through the
largest adipocyte to near zero for tangential cuts. Thus, the
corresponding correction by means of sterologic method is
mandatory. In a previous study2’ we demonstrated that real
mean adipocyte diameter is more than 16% higher than mean
diameter of transected profiles. On the other hand, it is
well-known18that direct counting of particle profile sections
tends to produce an incorrect estimation of number/volume
parameter. Stereologic methods accounting for size-frequency distribution and shape factor yield much more
accurate results.
The inverse relationship between the prevalence of fat cells
in red bone marrow and the activity of hematopoiesis is an
established fact.28On the other hand, reciprocal age-related
variations of hematopoietic cellularity and fat tissue fraction
are ~ e l l - k n o w n . ’ Whereas
~ * ’ ~ ~ ~during
~
the first decade of life
there is a striking predominance of hematopoietic cellularity
and relatively scanty fat tissue, the reverse is true in older
age. In a previous stereologic study’j we showed that such
age-related variations of fat tissue fraction in normal human
bone marrow depend both on size and number of adipocytes.
Variations in fat tissue fraction may be due to changes in
the size of individual adipocytes, or their number, or both.
Very few data are available on the relative importance of
these two factors in experimental animals, whereas before
our studies pertinent data in humans were lacking. Bathija et
al’ demonstrated, in rabbits, that the stimulation of erythropoiesis leads to a striking reduction in size of red marrow fat
cells, whereas the volume of systemic fat cells remains
unchanged. These findings suggest that increased hematopoiesis stimulates lipolysis of red marrow fat cells, and that both
populations, hematopoietic cells and marrow fat cells, may
be functionally related, probably the latter metabolically
supporting the former. In a study of rat marrow, the
stimulation of erythropoiesis over 7 days of hypoxia led to a
great reduction in both size and number of marrow adipocytes, with normalization of these parameters after the
cessation of experimental condition^.'^ The reverse phenomenon, ie, an increase of marrow adipocyte size and/or number
in states of depressed hematopoiesis, might be presumed;
however, it has not yet been demonstrated.
In our previous study” we showed that physiologic agerelated variations of fat tissue fraction depend both on size
and number of adipocytes. The present study is the first to
demonstrate that this physiologic phenomenon is maintained
during different physiopathologic situations such as bone
marrow aplasia, hyperplasia, and dysplasia. An important
degree of variation in size and number can be observed in
these abnormal conditions. As expected, the capacity of
adipocytes to expand and increase their size is not unlimited.
In hyperplasia and dysplasia the relative contribution of
adipocyte size and number to the variations of fat tissue
fraction is not different from controls, but the ratio number/
size increased significantly in aplastic patients, suggesting
that after fat cells have expanded to their maximum, a
disproportional increase in number is required to fill void
marrow spaces. Additional stereologic studies on adipocytes
may yield further insight on the relationship between hematopoiesis and its supporting tissue.
REFERENCES
1. Erslev AJ: Medullary and extramedullary blood formation.
Clin Orthop 52:25, 1967
2. Gordon MY, King JA, Gordon-Smith EC: Bone marrow
fibroblasts, fat cells and colony-stimulating activity. Br J Haematol
46:151, 1980
3. Dexter TM, Allen TD: The regulation of growth and development of normal and leukaemic cells. J Pathol 141:415,1983
4. Gordon MY, Hibbin JA, Dowding Ch, Gordon-Smith EC,
Goldman JM: Separation of human blast progenitors from granulocytic, erythroid, megakaryocytic, and mixed colony-forming cells by
“panning” on cultured marrow-derived stromal layers. Exp Hematol
13:937, 1985
5. Bathija A, Davis S, Trubowitz S: Marrow adipose tissue:
Response to erythropoiesis. Am J Hematol5:315, 1983
6. Bathija A, Ohanian M, Davis S, Trubowitz S : The marrow fat
cell: Response to x-ray induced aplasia. Life Sci 25:921, 1979
7. Brookoff D, Weiss L: Adipocyte development and the loss of
erythropoietic capacity in the bone marrow of mice after sustained
hyper-transfusion. Blood 60:1337, 1982
8. Cohen GI, Greenberger JS, Canellos GP:Effect of chemotherapy and irradiation on interactions between stromal and hemopoietic
cells in vitro. Scanning Electron Microsc 1:359, 1982
9. Touw I, Lowenberg B: No stimulative effect on adipocytes on
hematopoiesis in long-term human bone marrow cultures. Blood
61:770, 1983
10. Tavassoli M, Houchin DN, Jacobs P Fatty acid composition
of adipose cells in red and yellow marrow: A possible determinant of
haemopoietic potential. Scand J Hematol 18:47, 1977
1 1 . Malik F, Gordon MY, Goldman JM, Gordon-Smith EC:
Comparison of the composition of fat cells obtained from the marrow
of normal individuals or of subjects with aplastic anemia and from
bone marrow cultures. Exp Hematol 12:191, 1984
12. Burkhardt R, Frisch B, Bart1 R: Bone marrow in haematological disorders. J Clin Pathol35:257, 1982
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
STEREOLOGY OF HUMAN BONE MARROW ADIPOCYTES
13. Bartl R, Frisch B, Burkhardt R: Bone marrow biopsies
revisited. A new dimension for haematologic malignancies. Basel,
Switzerland, Karger, 1982, p 8
14. Frisch B, Lewis SM, Burkhardt R, Bartl R: Biopsy pathology
of bone and bone marrow. Biopsy Pathology Series. London, UK,
Chapman and Hall, 1985, p 9
15. Rozman C, Feliu E, Berga LI, Reverter JC, Climent C,
Ferrin MJ: Age-related variations of fat tissue fraction in normal
human bone marrow depend both on size and number of adipocytes:
A stereological study. Exp Hematol 17:34, 1989
16. Burkhardt R: Bone marrow and bone tissue: Colour Atlas of
Clinical Histopathology. Berlin, Germany, Springer, 1972, p 8
17. Gill GW, Frost JK, Miller KA: A new formula for a half
oxidized hematoxylin solution that neither overstains nor requires
differentiation. Acta Cytochem 18:300, 1974
18. Aherne WA, Dunnil MS: Morphometry. London, UK, Edward Arnold, 1982, p 27
19. Giger H, Riedwyl H: Bestimmung der Grossenverteilung von
Kugeln aus Schnittkreisradien. Biometr Zschr 12:156, 1953
20. Williams MA: Quantitative Methods in Biology. Amsterdam,
The Netherlands, North Holland, 1977, p 53
21. Weibel ER, Gomez DM: A principle for counting tissue
structures on random sections. J Appl Physiol 17:343, 1962
22. Lindberg LG, Vorwerk P On calculating volumes of trans-
895
sected bodies from two-dimensional micrographs. Lab Invest 23:315,
1970
23. Sokal RR, Rohlf FJ: Biometry. The principles and practice of
statistics in biological research. San Francisco, CA, Freeman, 1969,
p 571
24. Draper NR, Smith H: Applied regression analysis. New
York, NY, Wiley, 1966, p 86, 171
25. Cullen WC, McDonald T P Comparison of stereologic techniques for the quantification of megakaryocyte size and number. Exp
Hematol 14:782, 1966
26. Kerndrup G, Pallesen G, Melsen F, Mosekilde L: Histomorphometrical determination of bone marrow cellularity in iliac crest
biopsies. Scand J Hematol23:110, 1980
27. Rozman C, Feliu E, Berga LI, Reverter JC, Climent C,
Ferrin MJ: Estudio estereol6gico de las cClulas grasas en la medula
6sea humana. Sangre 33:206, 1988
28. Tavassoli M, Maniatis A, Crosby WH: Induction of sustained
hemopoiesis in fatty marrow. Blood 43:33, 1974
29. Hartsock RJ, Smith EB, Petty CS: Normal variations with
aging of the amount of hematopoietic tissue in the bone marrow from
the anterior iliac crest. Am J Clin Pathol43:326, 1965
30. Slavin BG, Yoffey JM, Yaffe P: Response of Marrow Adipocytes to Hypoxia and Rebound. Advances in the Morphology of Cells
and Tissues. New York, NY, Liss, 1981, p 231
From www.bloodjournal.org by guest on June 15, 2017. For personal use only.
1990 76: 892-895
Variations of fat tissue fraction in abnormal human bone marrow
depend both on size and number of adipocytes: a stereologic study
C Rozman, JC Reverter, E Feliu, L Berga, M Rozman and C Climent
Updated information and services can be found at:
http://www.bloodjournal.org/content/76/5/892.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.