Interstitial Deletion Constitutes the Major Mechanism for Loss of
Heterozygosity on Chromosome 20q in Polycythemia Vera
By Fotios A. Asimakopoulos, James G.R. Gilbert, Micheala A. Aldred, Thomas C. Pearson, and Anthony R. Green
An acquired deletion of the long arm of chromosome 20
is a recurrent abnormality in myeloproliferative disorders,
particularlypolycythemia
vera and myelodysplastic syndromes. The association of 20q deletions with myeloid
"stem cell" disorders suggests that thedeletions mark the
plays
site ofone or moregenes, loss or inactivation of which
a role in the regulationof normal hematopoietic progenitors.
We have recently performed a detailed molecular analysis
of 20q deletions in peripheral blood (PB) granulocytes and
defined a commonly deleted region of 16 t o 21 centimorgan
(CM).To further reduce the size of the common deleted region wehave searched for small deletionsor mitoticrecombination events, neither of which would detected
be
by conventional cytogenetics. We have studied 48 patients with
polycythemia vera and four patientswith idiopathic myelofibrosis. In each case, cytogenetic analysis had either failed
or had shown noabnormalities of chromosome 20. Seventeen microsatellite markers that span the common deleted
region were used t o search for loss of heterozygosity in
granulocyte DNA. No instance of microsatellite instability
was observed in a total of 880 comparisons of granulocyte
and T-cell DNA. Granulocyte DNA from four patientsexhib-
ited allele loss on 20q. In each casethe allele loss was caused
by an interstitial deletion because heterozygosity at distal
markers was retained and because quantitative Southern
blotting demonstrated hemizygosity.Loss of heterozygosity
in PB granulocytes would be masked by the presence of
significantnumbersof
normal granulocytes not derived
from the malignant clone. Therefore, the human androgen
receptor assay (HUMARAI was used t o determine granulocyte clonality. In 21 of 27 informative female patients the
majority of the granulocytes were clonally derived. In 5 patients thegranulocytes appeared polyclonal and in l patient
unilateral X inactivation was observed in both granulocytes
and T cells. These results show that, in the vast majority of
patients presented here, the failure t o detect loss of heterozygosity cannot be attributed to the presence of normal
polyclonal granulocytes. Our results therefore show thatallele loss on chromosome 20q in polycythemia vera does not
commonly involve mitotic recombination or chromosome
loss and that microsatellite instability isa rare event in this
disorder.
0 1996 by The American Societyof Hematology.
A
centered on the proximal G(+) band, 20q12. Microsatellite
polymerase chain reaction (PCR) and quantitative Southern
blotanalysisweresubsequentlyused
tocharacterize 20q
deletions in cell lines and peripheralblood (PB) granulocytes. Heterogeneity of both centromeric andtelomeric
breakpoints was shown,thus supporting the existence of one
or more tumor suppressor geneson20q. In addition, we
defined a commondeleted region of 16 to 21 CM which
contains ADA, PLC1,TOPI, SEMG1, and PPGB. Several
candidate tumor suppressor genes were found to lie outside
thecommondeletedregion,
including SRC, HCK, RBLl
(p107), PTPNI, and CEBPP.
Loss of heterozygosity for polymorphic markers is frequently observed in a wide variety of tumors and in several
cases has been shown to mark the site of tumor suppressor
genes."-14 Loss ofheterozygosity can result from several
mechanisms including mitotic recombination, gene conversion, chromosomedeletion, and chromosomenondisjunction
with or withoutreduplication." Different mechanisms accountfor loss of heterozygosity in differenttumors:loss
and reduplicationof chromosome 13 occurs commonly in
retinoblastomas"; whole chromosome loss resulting in monosomy for chromosome 10 occurs frequently in glioblastoma16; whereas mitotic recombination involving chromosomes 17p and 1 I is the major mechanism in glioma and
rhabdomyosarcoma."""
Mutations in genes responsible for fidelity of DNA replication and/or repair represent a recently recognized class of
genetic alterations in cancer. The resultant genomic instability can manifest as alterations in the length of simple repeat
sequences between tumor and normal tissues." Originally
described in hereditary nonpolyposis colon cancer and sporadic colorectal cancer as well as other sporadic epithelial
tumors,*".*' instability of simple repeat sequences has been
shown to occur at variable rates in hematologic malignan-
CQUIRED DELETIONS of the long arm of chromosome 20 area recurring abnormality in myeloid hematologic disorders. They were originally described in polycythemiavera (PV)' and theyrepresentthemost
common
chromosomalabnormality associatedwiththis
However, 20q deletions are also found in othermyeioproliferative d i ~ o r d e r s the
, ~ myelodysplastic syndromes and acute
myeloid leukemia (AML),h-R
but typically at lower rates than
in polycythemia. They arerarely reported in association with
lymphoid malignancies.' This pattern of disease association
implies that20q deletions markthe site of one or more genes,
loss orinactivation of which contributes to thedysregulation
of hematopoietic progenitors.
We recently performed a detailed cytogenetic andmolecular analysisof 20q deletions.'.'" Our data showed twoclasses
of deletions: large deletions encompassing
both Giemsa positive [G( +)] bands on 20q and small deletions encompassing
onlyone G(+) band.Reversechromosome
painting was
used to show that the cytogenetic common
deleted region
From the Department of Haematology, University of Cambridge,
MRC Centre,Cambridge; and the Department of Haematology,
United Medical and Dental Schools, St Thomas's Hospital,London,
UK.
Submitted December 21, 1995; accepted May 28, 1996.
Supported in part by the Leukaemia Research Fund and the Wellcome Trust (Grants No. 33748 and 35981).
Address reprint requests to Anthony R. Green, PhD, MRCPath,
FRCP, University of Cambridge, Department of Haematology, MRC
Centre, Hills Rd, Cambridge CB2 2QH, UK.
The publication costsof this article were defrayed in part by page
chargepayment. This article must therefore be hereby marked
"advertisement" in accordance with 18 U.S.C. section 1734 solely to
indicate this fact.
0 1996 by The American Society of Hematology.
0006-4971/96/8807-0018$3.00/0
2690
Blood, Vol 88,No 7 (October l ) , 1996:pp 2690-2698
MECHANISMS
269 1
FOR LOH ON CHROMOSOME 209
~ i e s . In
~ ~the- ~only
~ study of microsatellite instability in
myeloproliferative disorders, a high incidence was reported
in the accelerated and blastic phase but not the chronic phase
of chronic myeloid le~kemia.'~
Relatively little is known about the role of tumor suppressor genes, including mutator genes, in the pathogenesis of
the myeloproliferative disorders. Cytogenetically visible
chromosomal deletions [eg, del(20q), del( 13q)l are observed
in a minority of cases, but the identity of the critical target
genes remains obscure. No studies of loss of heterozygosity
in myeloproliferative disorders have been reported. Neuman
et alZ6examined 60 patients with primary or secondary myelodysplastic syndrome (MDS) for loss of heterozygosity on
chromosomes 5 and 7. They found no evidence for loss of
heterozygosity other than by major deletion or monosomy.
Similar data were reported by Shepherd et a]'' in a study
investigating loss of heterozygosity of chromosome 5 in
MDS. However, these studies were only able to use a relatively small number of probes, so small regions of loss of
heterozygosity may well have been missed. Moreover, clonality studies were not performed in either of these reports.
Although the precise correlation between genetic and
physical distances is not yet clear, the common deleted region clearly spans a large segment of the chromosomal arm.
Therefore, we have attempted to narrow the region of interest
by searching for small deletions or mitotic recombination
events, neither of which would be detected by conventional
cytogenetics. Patients with a known 20q deletion were excluded. A high-density map of the common deleted region
was constructed and a panel of microsatellite markers mapping to regular intervals across the common deleted region
was used to search for chromosome 20q allelic losses. In
addition we used the HUMARA assay" to show that the
majority of PB granulocytes were clonally derived in most
informative patients. Our data demonstrate that microsatellite instability is a rare event and that the main mechanism
underlying loss of heterozygosity for 20q sequences in polycythemia vera is interstitial deletion rather than mitotic recombination or chromosome loss.
MATERIALS AND METHODS
Patients. PB samples from 48 patients with polycythemia vera
and 4 patients with primary myelofibrosis were studied. All samples
were obtained from patients attending Addenbrookes Hospital (Cambridge, UK) or St Thomas's Hospital. Diagnosis of polycythemia
vera was established according to the Polycythemia Vera Study
Group (PVSG) criteria. The median time from diagnosis of polycythemia vera was 8 years (range, 1 to 28 years). All PV patients
except two male patients had received myelosuppressive therapy
(32P,busulfan, or hydroxyurea). Two patients with idiopathic myelofibrosis were diagnosed in 1993, 1 patient in 1992, and 1 patient in
1991.
Cell separation. Highly purified granulocyte and T-cell fractions
were prepared from PB using magnetic beads conjugated to antiCD15 and anti-CD2 antibodies, respectively (Dynabeads; Dynal,
Oslo, Nonvay). Ten to 20 mL of EDTA-anticoagulated blood was
layered onto Histopaque 1077 cushions (Sigma, St Louis, MO). Cells
from the mononuclear fraction were washed in phosphate-buffered
saline/O.l% bovine serum albumin and incubated with anti-CD2
conjugated beads according to manufacturer's recommendations.
Cells from the granulocyte pellet were similarly processed with anti-
cM
0
-
D20S106
-
D20S174
5
- D2OS107
-
PLCl
-
D20S108
10
D20S55
D20S46
- D20S110
15
CDR
D2OSl69
ADA
- D20S119
-
20
-
25
30
D20S17
-
D20S197
-
D20S176
-
D20S60
-
D20S100
I
D20S43
D20S16
35
40
45
50
- D20S102
Fig 1. Sex-averaged genetic map of the common deleted region.
Loci that could be uniquely ordered are shown against a centimorgan
scale. The most likely locations of loci that could not be uniquely
placedare indicated by vertical bars on the right-handside. The proximal limit of the common deleted region ICDRI is defined by Asimakopoulos et al" and the distal limit by Roulston et al."
CD15 beads. In each case the percentage of Dynabead-bound cells
was determined after each separation procedure using a counting
chamber. This was routinely in the range of 90% to 100%. Cells
were suspended in sterile, autoclaved water at a concentration of IO5
cells/mL, boiled for 5 minutes, and stored at -80°C until required.
Microsatellite PCR. Microsatellite PCR was performed essentially as previously described." Briefly, aliquots of 1,250 cells in
12.5 pL of water were amplified in 25 pL of PCR reaction mixtures
containing buffer (50 "OIL KCl, 10 "OIL Tris HCI pH 8.3,
1.5 "OIL
MgClz, 0.1% gelatine), dNTPs (125 pmoIL each), primers (25 pmol each), and Taq polymerase (0.5 U). In each reaction
mixture 1/10 of one of the primers was end-labeled with 32P-ATP
using polynucleotide kinase. Samples were amplified in an MJ Research programmable thermal controller (MJ Research, Watertown,
MA) for 30 cycles (94°C for 40 seconds; annealing temperature for
30 seconds; no extension step) with an initial denaturation (94"C, 5
minutes) and a final extension step (72"C, 15 minutes). Samples
were analyzed on 8 m o m urea, 6% polyacrylamide gels followed
by autoradiography. Primer sequences are available from the Genome Data Base (GDB; Johns Hopkins University, Baltimore, MD)
except AFMa132xe9 (communicated by J. Weissenbach [GBnBthon,
Paris, France] on a collaborative basis).
ASIMAKOPOULOS ET AL
2692
Table 1. Analysis of Granulocyte DNA for Loss of Heterozygosity
Locus
D20S107
PLCl
D20S108
D20S110
D20S46
D20S55
D20S169
D20S119
AFMa132xe9
D20S43
D20S16
D20S197
D20S176
D20S60
D20S100
D20S102
PCKl
Two Alleles
One Allele
Uninformative
34
41
32
36
44
44
39
32
42
41
44
42
37
42
44
37
40
3
4
3
2
2
4
1
2
3
2
3
3
2
2
1
0
0
15
7
17
13
6
4
11
18
7
9
4
7
13
8
7
15
12
Microsatellite PCR was performed on DNA from highly purified
granulocytes and T cells. Primer pairs corresponding to 17 loci on
chromosome 20q were used. "Two Alleles" denotes the presence of
two alleles in granulocyte DNA at an informative locus (ie, one for
which two alleles were detected in T-cell DNA). "One Allele" denotes
the presence of a single allele in granulocyte DNA at an informative
locus. Results were assessed by visual inspection of autoradiographs.
We have previously shown that this method can detect allele loss if
Numbers represent the number of patients
present in 240% of
for which each marker was tested. In three patients with interstitial
deletions, a few markers predicted to map within the deletions were
not assayed (see Table 2).
HUMARA donulie assay. Aliquots o f 10,000 Dynabead-separated cells in 100 pL o f water were phenol chloroform extracted
and the D N A precipitated using glycogen as carrier. D N A pellets
were resuspended in 14 pL o f I X Hlin 1 buffer and 40 U o f enzyme
added. Reactions were left to proceed at 37°C overnight. Digested
A
GT
GT
D20S119 D20S102
B
GT
Table 2. Details of Allele Loss in Granulocyte DNA
From Patients With Loss of Heterozygosity
Patients
Locus
HS
D20S107
1
1
1
1
PLC1
D20S108
D20S110
D20S46
D20S55
0205 169
D20S119
AFMa 132xe9
020343
D20S16
D20S197
D20S176
D20S60
D20S100
D20S102
PCKl
U
1
1
1
1
U
1
U
ND
2
2
2
2
PP
HH
IN
1
1
1
1
1
1
1
1
U
U
1
1
1
ND
U
U
U
U
1
1
ND
1
U
1
1
1
1
1
1
2
2
2
U
U
1
1
ND
1
1
1
1
2
1
1
1
U
2
2
2
2
U
Microsatellite PCR was performed as outlined in the legend to Table
1.
Abbreviations: 2, two alleles were amplified from granulocyte DNA
at an informative locus (ie, one for which two alleles were detected
in T cell DNA); 1, a single allele was amplified from granulocyte DNA
at an informative locus; U, uninformative; ND, not done.
D N A was then added to a final volume o f 30 pL o f PCR mixtures
at the same concentrations o f reagents as for other microsatellite PCR
reactions (see above). Samples were amplified i n an MJ Research
programmable thermal controller for 30 cycles (94OC for 40 seconds;
65°C for 30 seconds; 72°C for 30 seconds) with an initial denaturation (94°C. 5 minutes) and a final extension step (72"C, 15 minutes).
Products were analyzed on gels prepared in the same way as for
other microsatellite PCR products. Band densities were measured
on autoradiographic films using NIH Image Analysis Software vs.
GT
D20S176 D20S100
C
GT
GT
PLCl
D20S100
Fig 2. Loss of heterozygosity for loci on chromosome 20q. Microsatellite PCR was performed on PB granulocyte DNA (G) or PB T-cell DNA
(TI. Corresponding loci appear below each panel. (A) Patient MD. Microsatellite PCR using granulocyte DNA showed no loss of heterozygosity
for chromosome 20q markers. (B1 Patient PP. Microsatellite PCR resulted in amplification of two alleles for marker D20S100 but only one allele
for marker D20S176. IC) Patient IN. One allele is consistently less intense in granulocytes with both markers used.
2693
MECHANISMS FOR LOH ON CHROMOSOME 20q
1
ADA
Fig 3. Quantitative Southern analysis of granulocyte DNA from patients exhibiting loss of heterozygosity. Filters were hybridized to t h e control probe
p4.8 and subsequently to two probes from within
the common deleted region (ADA and S€MGI).'o OD
ratios were calculated as described in Materials and
Methods. Lane 1, granulocytes from a normal donor;
lane 2, cytogenetically normal lymphoblastoid cell
line KS10; lane 3, lymphoblastoid cell linewith a 2Oq
deletion, KS7; lane 4, patient HS; lane 5, patient PP;
lane 6, patient HH; lane 7, patient IN.
O.D. ratios:
1.57 (this software is in the public domain and available through
the Internet).
Estimation of the percentage of clonalgranulocytes. Percentages of clonal granulocytes were estimated using granulocyte and
T-cell allele intensity ratios. Granulocyte and T-cell DNA was predigested with Hha I and amplified withHUMARA primers as described above. Allele ratios were derived by dividing the intensity
of one allele (allele 1) by the intensity of the other allele (allele 2).
In each case, allele ratios were corrected for preferential amplification of alleles by comparison with the equivalent ratios obtained
from undigested granulocyte or T-cell DNA, to give granulocyte
and T-cell "corrected ratios," respectively.
A mixed population of clonal and polyclonal granulocytes consists
of a proportion of cells which are clonal (designated GC) and a
proportion of polyclonal cells (designated GP).All clonal granulocytes have the same active allele (eg. allele 1) and the same inactive
allele (allele 2). Only a proportion of the polyclonal granulocytes
has the same pattern (allele 1 active and allele 2 inactive). If this
proportion is designated 'h," the rest of the polyclonal granulocytes
("1 - n") will have the opposite pattern (allele l inactive and allele
2 active). The value of "n" is a measure of the degree of unequal
Lyonization in a polyclonal population of cells from a particular
patient.
In a mainly clonal population of granulocytes amplification of
Hha I predigested DNA will produce a strong and a weak allele.
The weak allele will derive solely from those polyclonal cells that
have an allele activity pattern opposite to thatof clonal cells. Its
intensity can be represented by [(l - n)Gp]. The intensity of the
strong allele will be derived from both the clonal cells present [G,]
together with a contribution from the polyclonal cells that have
identical allele activity pattern [nGp].
Therefore, the corrected granulocyte ratio (&) will be represented
by the following equation: & = ( I - n)Gp/nGp + G, (equation
l). The value of n (degree of unequal Lyonization in a polyclonal
population from a particular patient) can be estimated from the Tcell corrected ratio. This calculation is based on two premises:
3
4
5
6
7
UmW m . 9 w 1.0
1.0
SEMGl
O.D. ratios:
2
0.4 0.4 0.4 0.3 0.3
W""-1.0 1.0 0.4 0.4 0.5 0.4 0.4
Firstly, that T cells are not part of themalignant clone; and secondly,
that the degree of unequal Lyonization is similar in granulocytes
and T cells from the same patient. The T-cell corrected ratio (RT)
will be equal to:RT = (1 - n)Tp/nTp(equation 2). where n is the
proportion of cells that has inactivated allele 2 and Tpis the proportion of polyclonal T cells. If all T cells are polyclonal, TP = 1, and
so n can be calculated from RT.
The value of n derived from equation 2 can now be substituted
into the first equation. If equation 1 is then solved for GCthen the
following equation can be derived: GC = RT - &/ItT (& + I )
(equation 3). This equation was used to calculate the proportion of
clonal granulocytes presented in Table 3.
DNA extraction and quanritativeSouthernblotting
analysis.
High-molecular-weight DNA was extracted from granulocyte pellets
according to standard methods, digested with restriction enzymes,
electrophoresed in 0.8% agarose gels in 1 X TBE, and blotted onto
Hybond N+ nylon membranes (L. Chalfont, Bucks, UK) usingcapillary transfer. The DNA was alkali fixedonthe
membranes and
hybridized to 50 ng of '*P-labeled probe DNA. Probes consisted of
an ADA 1.2-kb genomic fragment containing exon 4 and pHSgdX
clone (SEMGI).'" A probe from the short arm of chromosome 20
(p4.8) was used for control hybridizations." Membranes were
washed at high stringency (0.2X S S C , 0.1 % sodium dodecyl sulfate,
2 X 15 minutes at room temperature, 1 X 15 minutes at 65 to 68°C)
and exposed to autoradiographic film.
Band intensities on the film were measured using the NIH Image
Analysis software vs. 1.57. The optical density ratio for each marker
was calculated by dividing the relevant band density with the band
density for the control probe (p4.8)on the same filter. The resultant
values were then normalized by comparison to the optical density
(OD) ratio obtained after hybridization of the same probes to DNA
from the cytogenetically normal KSlO cell line.
RESULTS
Construction of a genetic map of the common deleted
region. A number of genetic maps of the human genome
2694
ASIMAKOPOULOS ET AL
A
% HhaI digested DNA
100 90
80
60 50 40 30 20
70
10
0
B
n
aa
1.0
CI
w
-.
I
m
0.8
\
w
w
c1
0.6
c1
I
m
Li
aJ
Y
Fig 4. HUMARA PCR conditions result in linear
amplification. (A) Autoradiograph of a mixing experiment. DNA from a lymphoblastoid cell line derived
from a heterozygous female patient was mixed at
........
e
.
I
G
i
!
I
0
I
\
i
I
i
*
40
60
m \. .
\-
m
I
resulting mixes were amplified as described in Materials and Methods using HUMARA primers. (6)
Allele
ratios (fainter alleleldenser allele) were normalized
by comparison t o the ratio obtained using DNA not
DNA
predigested with Hha I. The resultant values were
plotted against percentage of Hha I-digested DNA.
The values represent the mean of four independent
experiments. Standard deviations were less than
0.04 in all instances.
have been published r e ~ e n t l y . ~However,
~ . ~ ~ the GCnCthon
map29consisted only of microsatellite markers isolated at
GCnCthon and other maps did not provide good coverage
across the entire common deleted region on chromosome
20q. Therefore, we constructed a high-density genetic map
of this region to aid analysis of deletions and loss of heterozygosity in this locality (Fig 1). The map was constructed
using genotypic data from the Cooperative Human Linkage
Center as described by Aldred et al (submitted). All pairs of
loci are supported by odds of at least 250: 1, with the exception of D20S17-D20S119 where the odds against inversion
are only 20: 1.
There are no inconsistencies between this and other published genetic
The data are also in agreement with
ordering information derived from previous analysis of patients with 20q deletions and with a physical map constructed
across the common deleted region (Aldred et al, submitted).
Markers AFMa132xe9 and P C K l were not present in the
genotype dataset used and therefore could not be mapped.
However, AFMal32xe9 maps close to D20S 17 by physical
means (Aldred et al, submitted) and P C K l has been reported
to map telomeric to D20S 102 on 20q 13.2 (First International
Workshop on Human Chromosome 20 Ma~ping”~).
Loss of heterozygosity on chromosome 209 in PB granulocytes. We have searched for loss of heterozygosity on chromosome 20q in 48 patients with polycythemia vera and 4
patients with idiopathic myelofibrosis. Thirteen markers
were selected that span the common deleted region at inter-
0
20
% HhaI digested
80
100
MECHANISMS
2695
FOR LOH ON CHROMOSOME 20q
Table 3. Densitometric Analvsis of HUMARA Assay Results
Granulocyte
Corrected
Clonality
Designation
Patient
Ratio (R,)
T-cell
% Clonal
Corrected
Ratio IR,)
Granulocytes
IGc)
~
Clonal
MM
KG
VG
EG
EP
JK
IG
MD
DR
AB
JM
DH
GP
WG
GG
SB
BW
0.000
0.006
0.015
0.014
0.020
0.015
0.023
0.048
0.016
0.052
0.030
0.123
0.253
0.068
0.169
0.081
0.270
0.517
1.007
3.087
1.386
2.040
0.859
0.993
2.073
0.239
0.945
0.311
1.079
4.493
0.274
0.960
0.326
1.855
100
99
98
98
97
97
96
93
92
90
88
79
75
71
70
70
67
Polyclonal
JM
MT
MG
ES
0.923
0.260
0.462
0.506
1.223
0.524
0.420
0.371
13
40
A
A
Granulocyte and T-cell clonality was assessed using the HUMARA
assay. All experiments were performed on at least two occasions.
Densitometric analyses of one set of results are presented here.
Where duplicate densitometric analysis was possible, individual results lay within 10% of the mean. Percentages of clonal granulocytes
were estimated using the formula described in Materials and Methods. Allele intensity ratios obtained using Hha I-digested DNA were
corrected by comparison with the equivalent ratio using undigested
DNA. Clonality designations were as follows: "Clonal," the proportion
of clonal granulocytes exceeds 50%; "Polyclonal," the proportion of
clonal granulocytes is less than 50% but the T-cell corrected ratio is
greater than 1:3; "Skewed," the proportion of clonal granulocytes is
<50% and the T-cell corrected ratio is <1:3. None of the cases amenable to densitometry produced a "skewed" pattern.
Abbreviations: RG, granulocyte corrected ratio; RT, T-cell corrected
ratio; Gc, percentage of clonal granulocytes. A, a calculation of the
percentage of clonal cells gave values of <O in patients MG and ES.
vals of 1 to 3 cM. The unpublished GCnCthon microsatellite
AFMa132xe9 has been included to provide a marker between D20S119 and D20S17. In addition, we included four
loci (D20S60, D20Sl00, D20S102, PCKI) that map to
20q13.2 in the vicinity of a unique deletion in a case of
AML M4Eo reported by Roulston et al.35
Our results are summarized in Table 1 and Fig 2. The vast
majority of patients did not display loss of heterozygosity for
any of the 2 4 markers used (Fig 2A). However, allele loss
over large regions of chromosome 2Oq was found in 4 patients,
all with polycythemia vera. Complete or near complete loss of
one allele was seen in 3 patients (Fig 2B) whereas in 1 patient
the intensity of one allele was consistently reduced (Fig 2C).
The results in the latter patient suggest the presence of a minor
subpopulation of granulocytes with no loss of heterozygosity
on 2Oq. In each of the 4 patients, the region involved was
extensive (Table 2). The distal breakpoint was determined in
all 4 patients. Precise centromeric breakpoints were not defined
but were proximal to PLCl in all 4 cases. Retention of heterozygosity for at least one distal marker in each patient showed that
the mechanism responsible for loss of heterozygosity in all 4
patients did not involve whole chromosome loss or simple
mitotic recombination.
Our data also provide insight into the prevalence of microsatellite instability in polycythemia vera. Seventeen loci
were examined in both granulocyte and T-cell samples from
52 patients (including the 4 patients with primary myelofibrosis) with the exceptions shown in Table 2. Thus, a total
of 880 separate comparisons between granulocyte and T-cell
DNA were performed. No instance of microsatellite instability was observed, thus demonstrating the rarity of the process
in polycythemia vera.
Loss of heterozygosity resulted from simple interstitial
deletion. The results obtained in the 4 patients who exhibited loss of heterozygosity were consistent with either interstitial deletion or complex mitotic recombination events.
Therefore, quantitative Southern blotting was used to distinguish between these possibilities. The results of these experiments appear in Fig 3. In all four patients the OD ratios
obtained using two probes (ADA and SEMGI) from the region of allele loss" were approximately half that of the
normal controls. The same results were obtained for both
probes in a second independent experiment. These data show
that granulocyte DNA from all 4 patients was hemizygous
for ADA and SEMGl, a finding that indicates the presence
of an interstitial deletion and not mitotic recombination.
Analysis of clonality using the HUMARA assay. Loss of
heterozygosity in PB granulocytes would be masked by the
presence of significant numbers of normal granulocytes not
derived from the malignant clone. Moreover, a minority of
patients with myeloproliferative disorders have been reported to have polyclonal granulocytes.'6 To ascertain to
what extent this phenomenon may have impaired our ability
to detect loss of heterozygosity, we assessed the clonality of
granulocytes from our female patients. It was decided to use
the HUMARA assay because of its high rate of heterozygosity and consistent patterns of
We developed a method for calculating the percentage
of clonal cells present in a population of granulocytes (see
Materials and Methods). The percentage of clonal granulocytes is calculated from the granulocyte corrected ratio by
using the T-cell corrected ratio to indicate the degree of
unequal Lyonization in that patient. To use the allelic ratios
to estimate the proportion of clonal cells, it was essential to
prove that the conditions used for HUMARA PCR resulted
in amplification within the linear range. To test the linearity
of the PCR conditions, DNA from a cell line derived from
a heterozygous female was digested with Rsa I, which leaves
the region between HUMARA primers intact. PCR using
Rsa I-digested DNA should result in amplification of both
alleles. A second aliquot was digested with Rsa I and Hha
I, and resulted in amplification of only one allele. DNA from
the Rsa I-digested sample was mixed at 10% increments
with DNA from the Rsa VHha I-digested aliquot (Fig 4A).
Fifty-four nanograms of resultant mixes (corresponding to
the amount of DNA from 10,000 diploid cells3*)was amplified using the conditions described in Materials and Methods.
ASIMAKOPOULOS ET AL
2696
A
B
G
A
-
T
- + - +
D
C
G
T
-A
- + - +
G
-A
T
G
- + - +
T
-A
- + - +
Fig 5. HUMARA analysis of patient samples. DNA was extracted from highly purified PB granulocytes and T cells and amplified using
HUMARA primers with I+) or without (-1 predigestion with the methylation sensitive enzyme Hha 1. (A) Patient MD. Monoclonalii of
granulocytes but not T cells in this patient is inferred by the disappearance of one allele after digestion of granulocyte but not T-cell DNA
with Hha 1. (B) Patient JM. Neither granulocyte nor T-cell DNA displayed allele-specific digestion with Hha 1, suggesting that both cell types
are polyclonal. (C) Patient PS. Allele-specific digestion of granulocyte but not T-cell DNA from a patient with two HUMARA alleles separated
by one trinucleotide repeat. For an explanation of similar patterns, see the study by Hughes." (D) Patient SS. Allele-specific digestion of both
granulocyte and T-cell DNA from a patient with two HUMARA alleles separated by one trinucleotide repeat.
Densitometric analysis of the results showed that the intensity of the upper allele exhibited a linear relationship with
the percentage of Hha I-digested DNA in the sample (Fig
4B).
All female patients (n = 31) were included in the HUMARA clonality analysis (the remainder 21 patients were
male). Three of these patients had idiopathic myelofibrosis
(patients HP, DR, DH) and the rest had polycythemia vera.
Four of the 3 1 patients were homozygous and therefore uninformative. Of the remaining 27 patients, 21 displayed wellseparated alleles, the intensity of which could be analyzed
by densitometry (see Table 3). An allele intensity ratio of
1:3 has previously been used as a practical (albeit arbitrary)
threshold for assessing the clonality of tumor cell populat i o n ~ . ' ~If~ there
'
is random X-chromosome inactivation (Tcell allele ratio = I). a granulocyte allele ratio of 1:3 would
correspond to 50% clonal granulocytes. In 17 of 2 1 informative patients that were assessed densitometrically, the proportion of clonal granulocytes exceeded 50% (Fig 5A). In 4
patients most or all of the granulocytes appeared polyclonal
(Fig 5B).
Six patients exhibited alleles too similar in size to allow
accurate densitometric analysis, and so the autoradiographs
were assessed visually. In 4 of these patients (including the 2
female patients with loss of heterozygosity for 20q markers)
preferential digestion of one allele was clearly observed in
granulocytes but not T cells (Fig 5C). In one patient no
preferential allele loss was seen in either granulocytes or
T cells (polyclonal pattern) and in the remaining patient
preferential allele loss was seen in both granulocytes and T
cells ("skewed" pattern) (Fig 5D). This result would be
consistent with either skewed Lyonization or the production
of both granulocytes and T cells by the malignant progenitors
in this patient.
Taken together, our results show that the majority of granulocytes were clonally derived in 21 of 27 informative patients and were mostly polyclonal in 5 patients. Therefore,
our data suggest that in most patients the failure to detect
loss of heterozygosity for loci on chromosome 20q did not
reflect masking by normal granulocytes.
DISCUSSION
In this report we have searched for loss of heterozygosity
within the common deleted region on chromosome 20q in
samples from 48 patients with polycythemia vera and 4 patients with primary myelofibrosis. Seventeen microsatellite
markers spaced at approximately 1- to 3-cM intervals were
used. Only four samples displayed loss of heterozygosity
and in each case the underlying mechanism was a previously
undetected interstitial deletion. These results show that in
polycythemia vera, as in MDS, mitotic recombination is rare.
Chromosome 20 is a small chromosome and the detection
of small deletions using conventional cytogenetics is technically demanding. Therefore, it has been proposed that small
submicroscopic deletions may exist but escape detection.
Our data suggest that such deletions are, in fact, rare. One
possible explanation would be the presence of widely spaced
hotspots for chromosome breakage. However, this seems
unlikely because there is considerable heterogeneity of both
centromeric and telomeric breakpoints.'" Moreover, it should
be emphasized that our data do not exclude the existence of
very small deletions ( < I to 3 cM) which would be beyond
the resolution of this study.
One potential reason for our failure to detect loss of heterozygosity in most patients might have been the presence
of significant numbers of polyclonal normal granulocytes.
Gilliland et a13' reported PB granulocytes that were polyclonal in one of three patients with polycythemia vera. More-
2697
MECHANISMS FOR LOH ON CHROMOSOME 20q
over, in further studies, total PB leukocytes or bone marrow
nucleated cells gave nonclonal patterns in 2 of 26 patients
with polycythemia.42-" The presence of significant numbers
of polyclonal granulocytes would mask loss of heterozygosity occumng in a subpopulation of clonally derived granulocytes. Previous mixing experiments have assessed the sensitivity of microsatellite PCR for the detection of a deletion
carried by a subpopulation of cells. These studies showed
that visual inspection of autoradiographs could detect reduced intensity of one allele if the deletion were present in
240% of these cells." Therefore, it was important to assess
the clonality of PB granulocytes in our patients.
To address this issue we have used the HUMARA assay
to estimate the proportion of clonally derived granulocytes
in all female patients included in this study. Five of 27
informative patients exhibited polyclonal granulocytes. In 21
of these informative patients, at least 50% of the granulocytes were clonally derived. These results show that, in the
majority of patients presented here, the failure to detect loss
of heterozygosity cannot be attributed to the presence of
normal polyclonal granulocytes. One patient, in whom densitometric analysis of results was not possible, displayed unilateral patterns of X inactivation in both granulocytes and T
cells. Unilateral X inactivation of T-cell DNA may reflect
skewed Lyonization or involvement of T cells in the malignant clone. Previous reports have shown monoclonal derivation of PB leukocytes in 24 of 26 patients with polycythemia
Vera.36,42-44A6 However, in view of the heterogeneity of X
inactivation patterns in different
it has been shown
that it is important to use T cells as a control to exclude
skewed Lyonization?* Before this report, purified granulocytes and T cells have been studied in only a few pat i e n t ~ . ~ ~ "Granulocytes
'.~
were clonal and T cells polyclonal
in 10 patients, both granulocytes and T cells were polyclonal
in 1 patient, and a skewed pattern was seen in both granulocytes and T cells in 1 patient. Two additional patients reported by Gilliland et a136demonstrated unilateral X inactivation of purified granulocytes but the clonality of lymphoid
cells was not studied. In none of these reports was an attempt
made to estimate the proportion of clonal cells.
In the work reported here, no difference in the size of the
microsatellite alleles amplified from granulocytes and T cells
was detected in a total of 880 separate comparisons. Therefore, our data show that genetic instability is extremely rare
in PV. There is relatively little information available about
microsatellite instability in hematologic malignancies. In a
study of 21 patients with lymphomas and acute leukemias,
both lymphoblastic and myeloid, using 10 microsatellite
markers?' microsatellite instability was only demonstrated
in one case of AML and two patients with B-cell lymphoid
malignancies (Burkitt's lymphoma and B-cell acute lymphoblastic leukemia). Kaneko et alx reported 6 patients with
MDS, two of whom exhibited microsatellite instability at 4
of 4 loci, whereas this was not observed in the other 4
patients, even using samples obtained after leukemic evolution. Furthermore, in chronic myeloid leukemia, a high incidence of genetic instability has been reported in the accelerated and blastic phase but not the chronic phase of the
disease.'*
In summary, our data show that allele loss on chromosome
2% in patients with PV does not commonly involve mitotic
recombination or chromosome loss and that microsatellite
instability is extremely rare in this disorder. In addition, the
HUMARA assay was used to show that most PB granulocytes were clonal in the majority of female patients and so
the failure to detect loss of heterozygosity cannot be attributed to the presence of normal polyclonal granulocytes.
ACKNOWLEDGMENT
We are grateful to Dr Maria Messinezy (Department of Haematology, St Thomas's Hospital, London, UK) for providing clinical samples.
REFERENCES
1. Kay HEM, Lawler SD, Millard RE: The chromosomes in polycythaemia vera. Br J Haematol 12507, 1966
2. Rege-Cambrin G, Mecucci C, Tricot G, Michaux JL, Louwagie
A, Van Hove W, Francart H, van den Berghe H: A chromosomal
profile of polycythaemia. Cancer Genet Cytogenet 25:235, 1987
3. Swolin B, Weinfeld A, Westin J: A prospective long-term
cytogenetic study in polycythaemia vera in relation to treatment and
clinical course. Blood 72:386, 1988
4. Diez-Martin JL, Graham DL, Pettit RM, Dewald GW: Chromosome studies in 104 patients with polycythaemia vera. Mayo Clin
Proc 66:287, 1991
5. Mertens F, Johanssen B, Heim S, Kristoffersson U, Mitelman
F Karyotypic patterns in chronic myeloproliferative disorders: Report on 74 cases and review of the literature. Leukemia 5:214, 1991
6. Knapp RH,Dewald GW, Pierre RV: Cytogenetic studies in
174 consecutive patients with preleukemia or myelodysplastic syndromes. Mayo Clin Proc 60507, 1985
7. Mufti GJ: Chromosomal deletions in the myelodysplastic syndrome. Leuk Res 16:35, 1992
8. Mittelman F: Catalog of Chromosome Aberrations in Cancer.
New York, NY, Wiley-Liss, 1991
9. Nacheva E, Holloway T, White N, Carter N, Grace C, Green
A R Characterization of chromosome 20q deletions in patients with
myeloproliferative disorders or myelodysplastic syndromes. Cancer
Genet Cytogenet 80:87, 1995
10. Asimakopoulos FA, White NJ, Nacheva E, Green AR: Molecular analysis of chromosome 20q deletions associated with myeloproliferative disorders and myelodysplastic syndromes. Blood
84:3086, 1994
11. Ponder B: Gene losses in human tumours. Nature 335:400,
1988
12. Patterson H: Approaches to proto-oncogene and tumour suppressor gene identification. Eur J Cancer 28A:278, 1992
13. Lasko D, Cavenee W, Nordenskjold M: Loss of constitutional
heterozygosity in human cancer. Annu Rev Genet 25:281, 1991
14. Levine A: The tumor suppressor genes. Annu Rev Biochem
62:623, 1993
15. Hansen MF,Cavenee WK: Genetics of cancer predisposition.
Cancer Res 475518, 1987
16. James CD, Carlbom E, Dumanski JP, Hansen M, Nordenskjold M, Collins VP, Cavenee WK: Clonal genomic alterations in
glioma malignancy stages. Cancer Res 485546, 1988
17. James CD, Carlbom E, Collins VP, Cavenee WK Mitotic
recombination of chromosome 17 in astrocytomas. Proc Natl Acad
Sci USA 86:2858, 1989
18. Mulligan LM, Matlashewski GJ, Scrable HJ, Cavenee WK:
Mechanisms of p53 loss in human sarcomas. Proc Natl Acad Sci
USA 875863, 1990
2698
19. Scrable HJ, Witte DP, Lampkin BC, Cavenee W K Chromosomal localisation of the human rhabdomyosarcoma locus by mitotic
recombination mapping. Nature 329:645, 1987
20. Wooster R, Cleton-Jansen AM, Collins N, Mangion J, Cornelis RS, Coper CS, Gusterson BA, Ponder BAJ, von Deimling A,
Wiestler OD, Cornelisse CJ, Devilee P, Stratton MR: Instability of
short tandem repeats (microsatellites) in human cancers. Nature
Genet 6:152, 1994
21. Aaltonen LA, Peltomaki P, Leach FS, Sistonen P, Pylkkanen
L, Mecklin JP, Jarvinen H, Powell SM, Jen J, Hamilton R, Petersen
GM, Kinzler KW, Vogelstein B, de la Chapelle A: Clues to the
pathogenesis of familial colorectal cancer. Science 260:8 12, 1993
22. Robledo M, Martinez B, Arranz E, Trujillo MJ, Gonzalez
Ageitos A, Rivas C, Benitez J: Genetic instability of microsatellites
in hematological neoplasms. Leukemia 9:960, 1995
23. Hubner G, Battmer K, Paaz U, Link H: Monitoring of relapse
and remission in acute leukaemias by DNA fingerprint analysis. Br
J Haematol 85:320, 1993
24. Kaneko H, Horiike S, Inazawa J, Nakai H, Misawa S: Microsatellite instability is an early event in the myelodysplastic syndrome
(letter). Blood 86: 1236, 1995
25. Wada C, Shionoya S, Fujino Y, Tokuhiro H, Akahoshi T,
Uchida T, Ohtani H: Genomic instability of microsatellite repeats
and its association with the evolution of chronic myelogenous leukemia. Blood 83:3449, 1994
26. Neuman WL, Rubin CM, Rios RB, Larson RA, LeBeau MM,
Rowley JD, Vardiman JW, Schwartz JL, Farber RA: Chromosomal
loss and deletion are the most common mechanisms for loss of
heterozygosity from chromosomes 5 and 7 in malignant myeloid
disorders. Blood 79:1501, 1992
27. Shepherd L, Cameron C, Galbraith P, Windsor S, Lillicrap
D: Absence of allelic loss on chromosome 5q by RFLP analysis in
preleukaemia. Leuk Res 15:297, 1991
28. Allen RC, Zogbi Hy, Moseley AB, Rosenblatt HM, Belmont
JW: Methylation of Hpa I1 and Hha I sites near the polymorphic
CAG repeat in the human androgen receptor gene correlates with X
chromosome inactivation. Am J Hum Genet 51:1229, 1992
29. Gyapay G, Morissette J, Vignal A, Dib C, Fizames C, Millasseau P, Marc S, Bemardi G, Lathrop M, Weissenbach J: The 199394 GCnCthon human genetic linkage map. Nature Genet 7:246, 1994
30. Matise TC, Prlin M, Chakravarti A: Automated construction
of genetic linkage maps using an expert system (Multimap): A human genome linkage map. Nature Genet 6:384, 1994
3 1, Buetow KH, Weber JL, Ludwigsen S, Scherpbier-Heddema
T, Duyk GM, Sheffield VC, Wang Z, Murray JC: Integrated human
genome-wide maps constructed using the CEPH reference panel.
Nature Genet 6391, 1994
32. Wunderle V, Dib C, Fizames C, Morissette J, Hazan J, Hansmann I, Whitehouse D, Vergnaud G, Weissenbach J: The EUROGEM map of human chromosome 20. Eur J Hum Genet 2:242, 1994
33. Rothchild CB, Akots G, Hayworth R, Pettenati MJ, Rao PN,
Wood P, Stolz F-M, Hansmann I, Serino K, Keith TP, Fajans SS,
Bowden DW: A genetic map of chromosome 2Oq12-q13.1: Multiple
highly polymorphic microsatellite and RFLP markers linked to the
ASIMAKOPOULOS ET AL
maturity-onset diabetes of the young (MODY) locus. Am J Hum
Genet 52: 110, 1993
34. Smith CL, Keith T, Hansmann 1, Weissenbach 3 , Asimakopoulos FA, Bowden DW, Deleuze JF, Dutton ER, Fasman KH,
Green T, Hadchouel M, Hazan J, Hilgartner S, Kingsbury DK, Loder
B, Malafosse A, Meunier-Rotival M, Pearson PL, Siracusa LD,
Steinlein 0, White N, Williamson CM: Report of the first international workshop on human chromosome 20 mapping 1993. Cytogenet Cell Genet 66:78, 1994
35. Roulston D, Espinosa R, Stoffel M, Bell GI, LeBeau MM:
Molecular genetics of myeloid leukaemia: Identification of the commonly deleted segment of chromosome 20. Blood 82:3424, 1993
36. Gilliland DG, Blanchard KL, Levy J, Perrin S, Bunn HF:
Clonality in myeloproliferative disorders: Analysis by means of the
polymerase chain reaction. Proc Natl Acad Sci USA 88:6848, 1991
37. Busque L, Zhu J, DeHart D, Griffith B, Willman C, Carroll
R, McBlack P, Gilliland DG: An expression based clonality assay
at the human androgen receptor locus (HUMARA) on chromosome
X. Nucleic Acids Res 22:697, 1994
38. A d a m RLP, Knowler JT, Leader DP: The Biochemistry of
the Nucleic Acids. London, UK, Chapman and Hall, 1986
39. Gale RE, Wheedon H, Linch DC: X chromosome inactivation
patterns using HPRT and PGK polymorphisms in haematologically
normal and post-chemotherapy females. Br J Haematol 79: 193, 1991
40. Gale RE, Wainscoat JS: Clonal analysis using X-linked DNA
polymorphisms. Br J Haematol 85:2, 1993
41. Busque L, Gilliland DG: Clonal evolution in acute myeloid
leukemia. Blood 82:337, 1993
42. Anger B, Janssen JWG, Schrezenmeier H, Hehlmann R,
Heimpel H, Bartram CR: Clonal analysis in chronic myeloproliferative disorders using X-linked DNA polymorphisms Leukemia 4:258,
1990
43. Lucas GS, Padua RA, Masters GS, Oscier DG, Jacobs A:
The application of X chromosome gene probes to the diagnosis of
myeloproliferative disease. Br J Haematol 72:530, 1989
44. Taylor KM, Shetta M, Talpaz M, Kantarjian HM, Hardikar
S, Chinault AC, McCredie KB, Spitzer G: Myeloproliferative disorders: Usefulness of X-linked probes in diagnosis. Leukemia 3:419,
1989
45. Asimakopoulos FA, Holloway TL, Nacheva EP, Scott MA,
Fenaux P, Green AR: Detection of chromosome 20q deletions in
bone marrow metaphases but not peripheral blood granulocytes in
patients with myeloproliferative disorders or myelodysplastic syndromes. Blood 87:1561, 1996
46. Tsukamoto N, Morita K, Maehara T, Okamoto K, Sakai H,
Karasawa M, Naruse T, Omine M: Clonality in chronic myeloproliferative disorders defined by X chromosome linked probes: Demonstration of heterogeneity in lineage involvement. Br J Haematol
86:253, 1994
47. Tan SS, Williams EA, Tam PPL: X-chromosome inactivation
occurs at different times in different tissues of the postimplantation
mouse embryo. Nature Genet 3:170, 1993
48. Gale RE, Wheadon H, Boulos B, Linch DC: Tissue specificity
of X chromosome inactivation patterns. Blood 83:2899, 1994
49. Hughes A: Optimization of microsatellite analysis for genetic
mapping. Genomics 15:433, 1993