Reprinted from
Proc. Nat. Acad. Sei. USA
Vol. 70, No. 9, pp. 2629-2632, September 1973

Leukemia-Specific DNA Sequences in Leukocytes of the Leukemic
Member of Identical Twins

(virogene hypothesis/RNA-directed DNA polymerase)

W. BAXT*, J. W. YATESt, H. J. WALLACE, JR.t, J. F. HOLLAND?, AND S. SPIEGELMAN*

* Institute of Cancer Research and the Department of Human Genetics and Development, College of Physicians and Surgeons,
Columbia University, 99 Fort Washington Ave., New York, N.Y. 10032; and t Department of Medicine A, Roswell Park
Memorial Institute, Buffalo, New York 14203

Contributed by S. Spiegelman, June 12, 197s

ABSTRACT   The discovery in human leukemic cells
of particulate elements encapsulating 70s RNA and RNA-
directed DNA polymerase made possible the synthesis of
a [3H]DNA probe that could detect leukemia-specific
sequences in the DNA of normal and leukemic individuals.
In an earlier study of a series of unrelated leukemic pa-
tients, we established that the nuclear DNA of their leu-
kemic cells contain particle-related sequences that cannot
be detected in leukocytes of normal individuals. This
result is inconsistent with the virogene concept that de-
mands the inclusion of one complete copy of oncogenic
information in the genome of every normal cell.

The present study carries this analysis one step further
by showing, with two sets of identical twins, that the
leukemic member contains particle-related sequences in
the DNA of his leukocytes that cannot be detected in the
leukocytes of his healthy identical sibling. This finding
implies that the additional leukemia-specific information
found in the DNA of the leukemic individuals must have
been inserted subsequent to fertilization. This outcome
argues against the virogene hypothesis or any other etio-
logic concept that invokes vertical transmission throng11
the germ line of the particle-related information found
uniquely in the DNA of leukemic cells.

We have shown by molecular hybridization that human
adenocarcinomas of the breast (l), leukemias (2), sarcomas
(3), and lymphomas (4) contain RN4 molecules possessing a
small but significant homology to RNAs of tumor viruses
that cause the corresponding malignancies in mice. More telling
for a viral involvement was the demonstration with the
simultaneous detection test (5) that the RNA detected in
these human cancers was 70 S in size and eiicapsulated with
RNA-directed DNA polymerase in a particle possessing a
density between 1.16 and 1.19 g/ml (6-9). Furthermore,
DNA synthesized endogenously by these RNA-enzyme
complexes exhibited evidence of complementarity to RNAs of
the analogous murine viruses (6-9).
Taken together, these experiments documented the exis-
tence in human neoplasias of particulate elements possessing
four features diagnostic of animal RNA tumor viruses. The
data did not of course establish that the particles identified
were causative, but the evidence for their involvement was
sufficiently convincing to encourage further exploration of
their significance.
In our studies of human leukemias we developed the meth-
odology req.uired to separate the particles coiitainiiig the
RNA-dependent DNA polymerase and its 70s RNlZ tem-
plate. These particulate elements were then used to generate
3H-labeled DNA probes suitable for detecting corresponding

2629

sequences in genomes of any cells of interest. We found (9)
that RNA of human leukemic particles shared sequences with
the DNA of normal cells, a feature observed (IO) with animal
RNA tumor viruses and the normal DNA of their indigenous
hosts. Sequences common to both normal DNA\ and the [3H 1-
DNA synthesized by leukemic particles were removed by
exhaustive hybridization to a vast excess of normal DNA
followed by hydroxyapatite chromatography to separate
duplexes from unpaired [3H]DNA. The residue was then
used to obtain an answer to the the following question: "Does
the DNA of leukemic cells contain sequences that are not found
in the DNA of normal cells?" The data obtained with eight
leukemic patients were satisfyingly clear cut. In all cases it
was possible to demonstrate (9) that the I>NA of leukemic
cells contained particle-related sequences that could not be
found in DNA of leukocytes of normal individuals. The
sensitivity of the assay was such that 1/40Oth of an equivalent
per normal genome would have been detected. We have shown
(Kufe, I)., Peters, W. P. & Spiegelman, S., in manuscript)
that a similar situation exists in human lymphomas, including
Burkitt's tumors and Hodgkin's disease, and in feline sar-
comas (10).
The fact that specific particle-related sequences are unique
to the DN,4 of leukemic cells possesses obvious im1)lications
for the etiology and possible control of this disease. It is
evident that a specific segment of tlie viral-related sequences
detected earlier by hybridization has in some manner been
inserted into the genome of tlie leukemic cells. Our inability
to detect these specific sequences in the DNA of normal
leukocytes argues against the virogene-oncogene hypothesis
(11, la), which stipulates that a complete copy of tlie iiiforma-
tion required for virus production and malignant transforma-
tion is vertically transmitted arid is therefore preseiit in the
DNA of all cells in all animals prone to cancer.

The question of whether or not the virogene concept is
applicable to the human disease is central to the development
of a rational approach toward its control and cure. The
availability of two sets of identical twins, each containing a
leukemic individual, provided an extraordinary opliortunity
to pursue this issue further. The data derivable from this
situation could serve to test the validity of our earlier findings,
and, if confirmatory, sharpen the biological implications of
our conclusions concerning the etiology of this disease.
Since identical twins derive their genomes from the same
fertilized egg, any vertically transmitted information must he

2630 Medical Sciences: Baxt et al.

1000-

000


600

E
0
e 400-

200

Proc. Nut. Acad. Sei. USA 70 (1973)

-

-

-

'Oo0l 000

0

8 400

2ool 0

OL

Yo -Twins

Leukemic

n

r

Elution Temperature, "C

FIG. 1. Hydroxyapatite elution profile of a hybridization
reaction of recycled [3H]DNA probe of the leukemic twin to
nuclear DNA from normal leukocytes, normal twin leukocytes,
and leukocytes from the same leukemic twin. The annealing
reaction mixtures contained 20 A260 units of cellular DNA and
4 fmol of [3H]DNA, in a final volume of 0.1 ml. The reaction
was brought to 98" for 60 sec, and 40 pmol of NaCl was added.
The reaction mixture was then incubated at 60" for .50 hr. The
reaction was stopped by addition of 1 ml of 50 mM NaHPOa (pH
6.8). The sample was then passed over a column of hydroxy-
apatite of 20-ml bed volume at 60". The column wab washed with
40 ml of 0.15 M NaHP04 (pH 6.8) at 60", 80", 88", and 95".
Fractions of 4 ml were collected, the A260 of each fraction was
read, and the DNA was precipitated with 2 pg/ml of carrier
yeast RNA and 10% trichloroacetic acid. The precipitate was
collected on Millipore filters, which were dried and counted. In
all cases, more than 80% of the nuclear DNA reannealed.  A
background count of 8 cpm was subtracted in all instances.

present in both. The virogene hypothesis would therefore
predict that if the leukemic member of the pair contains
specific particle-related DNA sequences, these should also be
found in the leukocyte DNA of his healthy sibling. The data
obtained and described here do riot agree with this prediction.
We fourid the same difference between the healthy and leuke-
mic identical siblings as we had earlier found between un-
related leukemic patients and random healthy blood donors.
In the two sets of identical twins examined, the leukemic
member possessed particle-related sequences in his myelo-
blast DXA that could riot be detected in the leukocyte DNA
of his healthy sibling.

MATERIALS AND METHODS

Clinical. The Yo male twins were 24 years old when one of
them (DYo) was diagnosed as having acute myelocytic
leukemia. The patielit was leukophoresed with an I1311 cell
separator when his peripheral leukocyte count was 27,500
(907, myeloblasts). Evidence for the moriozygosity of the Yo
pair came from concordance in fingerprint analysis, HL-h

(1 and 2) typing, and of the erythrocyte major and minor
antigenic groups.
The Ka twins were 26 years old when one of them (PNa)
was diagnosed as having acute myelocytic leukemia.' The
patient was leukophoresed with an IEhl cell separator when
his peripheral leukocyte count was 28,000 arid yielded a
leukocyte population containing 407, myeloblasts. Evidence
for the monozygosity of the Na pair came from concordance

          , HL-A (groups 1 and 2) typing, arid of
the erythrocyte major and minor antigenic groups.

In both cases, leukocytes from the healthy sibling were
obtained from the buffy coats separated from a unit of whole
blood.

Preparation of [ 311]DNA. Leukemic leukocytes were gently
opened with a Dounce homogenizer. The extract was cen-
trifuged at 3000 rpm for 15 miii at 2" to remove nuclei and
then at 10,000 rpm for 15 min at 2" to remove mitochondria.
The resultant supernatant was layered on 25%  sucrose-
10 mNI Tris.HC1 (pH 8.3)-0.15 h#I NaC1-10 m1#l EDTA and
centrifuged in a Beckman SW-41 rotor at 40,000 rpm for 1
hr at 2". The resulting pellet was suspended in 1 ml of 10 mM
Tris.HC1 (pH 8.3) and used in a typical endogenous RNA-
directed DNA polymerase reaction to generate [3H]DN.Z (6).
The product was brought to 0.3 M NaOH aiid incubated at
37" for 2 hr to destroy RKA. To provide acceptable back-
grounds for such probes, it is imperative that all self-com-
plementary material be removed before use. To this end, the
alkali-treated material was 1ieut)ralized with 40 my1 HC1 and
1 M Tris. HC1 (pH 7.4) and then self-annealed, corresponding
to (lot values of 0.01-0.25 as defined by 13ritten aiid Kohne
(13), at 60" for 48 lir. It was then passed over an HTP Hiorad
hydroxyapatite coluniri with a 4-in1 bed volume at 60". Single-
stranded [3H]I>NAi was eluted with 4 ml of 0.15 21 NaHP04
(pI1 6.8), wliicli recorers 90% of the iiiput. This was then
layered upoii :L Sephades G-50 column with a bed volume of
100 ml (coarse). The DNA region of the column was pooled,
ad the [31-I]L)NAi was precipitated with 2 pg/ml of yeast
carrier RNA aiid two volumes of 95% et,hanol. The pellet
was dissolved iii 3 mM EDTA\ and stored at 2". [3FI]I>?jA
baiided iii Cs2S01 pradierits :it deiisities between 1.42 and 1.45
g/ml.

Preparation of Nuclear DNA. The nuclear pellet (3000 rpni
for 15 min) was sus1)eiided in 10 in31 Tris.HC1 (pK 83-0.15
;VI NaC1-10 mM EDT.4 aiid lysed by addition of Na dodecyl
sulfate toa concentration of 1% in the presence of 1 i\l ?;aC104.
lifter repeated extraction with chloroform-isoamyl alcohol
(4%), the 1IN-l was precipitated with alcohol and spooled
out of the solution. The DNA was dissolved in 10 nil1 Tris.
HCl (pH 8.2)-3 mh1 EDTA and sheared to 6-8 S, as measured
in a 10-30y0 glycerol gradient. Shearing was accomplished
with a I%roiiwill I3iosoiiic IV sonifier with the microtip at
maximum power for two 30-sec intervals. The sample was
then incubated with 0.3 ;\I NaOH for 16 hr at 37". After
neutralization, DNA\ was precipitated with alcohol and
dissolved in 10 mA1 Tris.HC1 (pII 8.2)- 3 in11 EDT-4. The
DN.l preparations had an AZBo/Azso rat,io greater than 1.8.
More than 90% renatured :is measured by hydrosyapati te
chromatography.

Annealing Conditions. r\niiealiiig reactions coritaiiied >60
AZFO units of nuclear LIXA, 0.1-1.0 pmol of [3H]DNAi, in a
final volume of 0.1 ml. The reaction w:is brought to 100" for

Proc. Nat. Acad. Sci. USA 70 (1373)

Leukemia-Specific DNA Sequences 2631

60 sec, and 40 pmol of NaCl was added. The reaction mixture
was then incubated at 60". The reaction was stopped by
addition of 2 ml of 0.15 31 NaHP04 (pH 6.8). The sample
was then passed over a column of HTP hydroxyapatite with a
3O-ml bed volume (15, 16) at 60". The column was washed
with 20 ml of 0.15 RI NaHP04 (pfI 6.8) at 60", 80°, 88", and
95". Fractions of 4 ml were collected. AzGo of each fraction
was read, and the DNA was precipitated with 2 pg/ml of
yeast carrier RNA and 10% trichloroacetic acid. The precipi-
tate was collected on Millipore filters, which were dried and
counted as described (6).
Tlie method identifies unpaired strands that elute at 60"
and poorly paired duplexes that disassociate at 80". Only the
duplexes disassociating and eluting at 88"-95' are counted here
as hybridized.

Recycling of Probe on Normal DNA to Separate Leukemic
from Xormal Sequences. Reactions were done with normal
DNA and the eluates were collected in I-ml fractions at 60".
The peak fractioiis were combined and the resulting pools
were passed over a Sephadex G-50 column with a 20-ml bed
volume. Regions containing DNA were collected, and the
DXAI was precipitated with 2 pg/nil of yeast carrier RNA and
2 volumes of 9575 ethanol. The material eluted at 60" was
used as the recycled product unable to hybridize to normal
DNA.

RESULTS

The strategy of the experiments to be described may be out-
lined as follows: (1) Isolate from the leukemic cells of each
patient the particles encapsulating the RNA-directed DXLI
polymerase, and its 70s RNA template (8). (2) Use this frac-
tion to geuerate [ 3H]DNA eiidogeiiously synthesized in the
presence of high conceritratioiis (400 pg/ml) of actinomycin
1) to inhibit host aiid viral 1)N:Z-directed DNA$ synthesis
(14-16). (3) Purify the [3H]DXA by hydroxyapatite and
Sephadex chromatography with care being exercised to
remove all self-complementary material. (4) Sequences shared
with normal DSA are removed by exhaustive hybridization
in the presence of vast excess of normal DNA followed by
hydroxyapatite chromatography to separate paired from
unpaired [aH]LISA. In this step the normal DXA used came
from leukocytes of healthy unrelated blood donors aiid not
from the normal twin. To have used the latter would have
obviously prejudged the issue. (5) The residue of [3H]I)N,Z
probe that does riot pair with normal DNA is then used to
test for the presence of sequences in the leukocyte IIN-I of the
patient arid in that of his healthy identical sibling. In such
experiments it is imperative to adjust DNA coiicentratioris
and durations of hybridizations so that sequences present
only once per genome equivalent will be readily detected.
In the experiments to be described, this was done by running
the annealing reactions to Cot (concentration of DNA nucleo-
tides in mol/liter times sec) values of 10,000 and above.
111 carrying out the fourth step for removal of normal
sequences from the [3H]I)NAi, annealing reactions were set
ulj contaiiiiiig 60 ityaO units of riormal cellular IIXA, 0.1
pmol of [31-T]IlNAZ (3000 cpm), and 15 pmol of NaHP04
(pH 7.2) in a final volume of 0.1 ml. The reaction was brought
to 98" for 60 see and 40 pniol of NaCl was added.

hnne:iling was done at 60" for 50 hr. The reaction mixture
was then added to 1 ml of 0.15 11 NalIP04, passed over a
30-ml hydroxyapatite column at 60", and washed with 0.1 5

M NaHP04; eluates were collected in I-ml fractions and
yielded 25% of the input [3H ]DNA. The peak fractions were
combined, and the resulting pools were passed over a Sephadex
G-50 column of 20-ml bed volume. The DNA regions were
collected, and the DNAi was precipitated with 2 pg/ml of
carrier yeast RNA and 2 volumes of 95% ethanol.
This material, which cannot hybridize to normal DNA, was
then used to examine the nuclear DNA from leukocytes of
the normal twin and from that of his leukemic sibling. The
results obtained with the two sets of twins are shown in Fig. 1
and reveal a response pattern indistinguishable from our
previous findings with eight unrelated leukemic patients. In
that study we showed that after recycling on normal DNA
the particle-derived [3H]DNA showed no ability to form well-
paired duplexes (requiring 88' and above to melt) with normal
DNA but did form stable duplexes with the DNA of the
original patient. In the present instances, we see that the
same situation holds between the members of each twin pair.
After removal of normal DNA sequences with unrelated
normal DNA, the remaining [3H]DNA can still form stable
duplexes with the DNA from the leukemic individual but
not with the DNA from his healthy identical sibling.

DISCUSSION

Using lymphocyte toxicity tests, Levine and his collaborators
(17) examined 10 sets of identical twins and found antigenic
differences in seven instances between the leukemic and
healthy members of each pair. A41though consistent with the
results reported here, such phenotypic dissimilarities do not
necessarily imply a difference in genome content since they
could reflect rather a disparity in gene expression. It is worth
explicitly ernphasizing that experiments that focus on pheno-
type or search for specific RNAs in tumor cells are limiting
their attention to active genes in the form of transcripts or
their translated products. In contrast, the experiments de-
scribed here and in our previous study of leukemias (9) center
on DNA sequences and, as such, include all genes, active or
silent.
The fact that we could establish a sequence difference be-
tween identical twins implies that the additional information
found in the DNA of the leukemic members was inserted
after the zygote was formed. This outcome agrees with the
results and conclusions of our previous study (9) of unrelated
leukemic patients arid is consistent with the original provirus
concept (18). The data argue against the validity of the
virogene hypothesis (ll), which demands that if leukemia-
specific sequences are found in the DNA of the individual
with the disease they must surely also exist in the genome of
his identical healthy sibling. We have shown (Uaxt, W., &
Spiegelman, S., manuscript in preparation) that human
leukemia-specific sequences are partially homologous to
RNA of Rauscher leukemia virus. This latter fact, along with
the outcome from the monozygotic twins, permits a biologi-
cally more informative conclusion with respect to the virogene
hypothesis and the etiology of human leukemia. One could
have argued from the study of the unrelated leukemia patients
that they had the disease because, in fact, each inherited the
necessary m:ilignant information in his genome. Tlie results
with the moiiozygotic twins contradict this possibility. The
data imply that even those who come down with this malig-
iiancy do iiot inherit the required information in their germ
lines.

2632 Medical Sciences: Baxt et al.

Proc. Nat. Acad. Sei. USA 70 (1973)

This research was supported by Contract no. NO1-CP-3-3258
within the Virus Cancer Program of the National Cancer ln-
stitute and by Research Grants CA-02332, CA-10044, and CA-
,5834, National lnstitutes of Health, Bethesda, Md. and the Irma
T. Hirschel Scholar Award.

1.  Axel, R., Schlom, J. & Spiegelman, S. (1972) Nature 235,
  32-36.
2. Hehlmann, R., Kufe, D. & Spiegelman, 8. (1972) Proc.
  Nat. Acad. Sci. USA 69, 43.5-439.
3.  Hehlmann, R., Kufe, D. & Spiegelman, S. (1972) Proc. Nat.

4. Kufe, D., Hehlmann, R. & Spiegelman, S. (1972) Science

Acad. Sei. USA 69, 1727-1731.

.-

175, i82-185.

5.
6. Baxt, W. G., Hehlmann, R. & Spiegelman, S. (1972)

Schlom, J. & Spiegelman, S. (1971) Science 174, 840-843.

--

Nature Neu, Bid. 240, 72-79.
Gulati, S. C., Axel, R. & Spiegelman, S. (1972) Proc. Nut.
Acad. Sei. USA 69, 2020-2024.
Axel, R., Gidati, S. C. & Spiegelman, S. (1972) Proc. Nut.

7.


8.

Acad. Sei. USA 69, 3133-3137.

9.


10.


11.

12.

13.
14.

15.

16.

17.

18.

Baxt, W. G. & Spiegelman, S. (1972) Proc. Nut. Acad. Sci.
USA 69,3737-3741.
Ruprecht, R. M., Goodman, N. C. & Spiegelman, 8. (1973)
Proc. Nnt. Acad. Sci. USA 70, 1437-1441.
Huebner, R. & Todaro, G. (1969) Proc. Nut. Acad. Sci.
USA 64, 1087-1094.
Todaro, G. J. & Huehner, R. J. (1972) Proc. Nut. Acad. Sci.
USA 69, 1009-1015.
Britten, R. J. & Kohne, 1). E. (1968) Science 161,529-540.
Manly, K. F., Smoler, D. F., Bromfield, E. & Baltimore, 11.
(1971)J. Virol. 7, 106-111.
Varmus, H. E., Levinson, W. E. & Bishop, J. M. (1971)
Nature New Biol. 223, 19-21.
Ruprecht, R. M., Goodman, N. C. & Spiegelman, S. (1973)
Riochim. Biophys. Acta 294, 192-203.
Levine, P. H., Herberman, It. B., Rosenberg, E. B., Mc-
Clure, P. D., Roland, A,, Pienta, R. J. & Ting, R. C. Y.
(1972) J. Nut. Cancer Inst. 49, 943-952.
Temin, H. M. (1964) Proc. Nut. Acad. Sci. USA 52,323-329.