CA2080255A1

CA2080255A1 - Primitive hematopoietic stem cell preparations

Info

Publication number: CA2080255A1
Application number: CA002080255A
Authority: CA
Inventors: Connie J. Eaves; Allen C. Eaves; Peter M. Lansdorp
Original assignee: Individual
Current assignee: Individual
Priority date: 1992-07-31
Filing date: 1992-10-09
Publication date: 1994-02-01
Also published as: CA2120482A1; AU4693693A; WO1994003587A1

Abstract

ABSTRACT
A cell preparation containing primitive hematopoietic stem cells obtained from the blood of leukemic patients. The stem cells are Philadelphia chromosome positive and produce detectable clonogenic progenitors in long term culture. A cell preparation comprising primitive hematopoietic stem cells obtained from the blood of normal individuals which produce detectable clonogenic progenitors in long term culture. Quantitative assays for the cell preparations and, methods for testing for substances which affect hematopoiesis of the hematopoietic stem cells are described.

Description

208025~

Title: PRIMITIVE HEMATOPOIETIC ST~M CELL PREPARATIONS

FIELD OF THE INVENTION
The invention relates to the isolation and use of primitive hematopoietic stem cells from leuke~ic patients and from blood of normal individuals; quantitative assays for same; and, methods for testing for substances which affect hematopoiesis of the hematopoietic stem cells.
BACgGROVND OF THE INVENTION
The regenerative capacity and life-long maintenance of the hematopoietic system is dependent on a primitive subpopulation of stem cells with extensive self-renewal, proliferative and differentiation potential. Totipotent hematopoietic stem cells with the capacity to repopulate lymphoid and myeloid tissue in myeloablated recipients have been documented (Abramson, S. Miller, R.G. h Philips, R.A., J. Exp. Med., Vol. 145, pp. 1567-1579, 1977; Mintz, B. Anthony, K. & Litwin, S. Proc. Natl. Acad. Sci USA, Vol. 81, pp. 7835-7839, 1984; Turhan et al., N. Engl. J.
Med., Vol. 320, pp. 1655-1661, 1989; Keller et al., Nature, Vol. 318, pp. 149-154, 1985; Dick et al., Cell, Vol. 42, pp. 71-79, 1985; Lemischka et al., Cell, Vol. 45, pp. 917-927, 1986; Snodgrass, R. & Keller, G., EMBO J., Vol. 6, pp. 3955-3960, 1987. Capel et al., Proc. Natl.
Acad. Sci. USA, Vol. 86, pp. 4564-4568, 1989; Capel et al., Blood, Vol. 75, pp. 2267-2270, 1990; Van Zant, et al., Blood, Vol. 77, pp. 756-763, 1991).

- Evidence points to a hierarchy of stem cells with differing potentials for sustaining hematopoiesis when transplanted in vivo. Cells with long term hematopoietic reconstituting ability can be distinguished by a number of physical and biological properties from cells that only generate mature progeny in short-term in vivo or in vitro clonogenic assays (Hodgson, G.S. & Bradley, T.R., Nature, Vol. 281, pp. 381-382; Visser et al., J. Exp. Med., Vol.

, ....

. ~ - .
.

.

20802~S

59, pp. 1576-1590, 1984; Spangrude et al., Science, Vol.
241, pp. 58-62, 1988; Szilvassy et al., Blood, Vol. 74, pp. 930-939, 1989; Ploemacher, R.E. & Brons, R.H.C., Exp.
Hematol., Vol. 17, pp. 263-266, 1989).

There have been substantial efforts to develop methods to detect, isolate and characterize the most primitive hematopoietic stem cells. Such cell populations are expected to be useful in bone marrow transplants for the treatment of a variety of conditions such as AIDS, leukemia and certain anemias, and in gene therapy. The cell populations may also be used to identify growth factors, to screen growth factors and in assays to study the development of hematopoietic cells.

Primitive hematopoietic stem cells capable of initiating long term culture have been identified. The number of clonogenic cells present after 5 to 8 weeks in long term cultures initiated with normal hematopoietic cells allows the detection of a very primitive class of clonogenic cell precursors that exhibit properties characteristic of cells with long-term in vivo reconstituting potential (Sutherland et al., Blood, Vol. 74, p. 1563, 1986 -Udomsakdi et al., Exp. Hematol., Vol. 19, p. 338, 1991.) These normal human "long-term culture-initiating cells"
(LTC-IC~ can be quantitated by limiting dilution analysis, which then allows the proliferative potential of individual LTC-IC to also be determined (Sutherland et al., Proc. Natl. Acad. Sci., Vol. 87, p. 3584, 1990).

Normal marrow LTC-IC are well maintained in LTC
estab.lished from a single input inoculum (Eaves et al., Effects of Therapy on Biology and Kinetics of Residual Tumor, Part A: Pre-Clinical Aspects, p. 223, 1990 - Eaves et al., Ann. N.Y. Acad. Sci. Vol. 628, p. 298, 1991) and similar kinetics are seen when highly purified LTC~IC from normal marrow are seeded onto preestablished feeders ... . . . .

c : ~ -, 20802~5 (Sutherland et al., Proc. Natl. Acad. Sci., Vol. 87, p.
3584, 1990).

Abnormalities of the primitive hematopoietic stem cells - may give rise to a variety of conditions, including leukemia, a family of disorders characterized by the progressive proliferation of abnormal leukocytes. Chronic - myeloid leukemia (CML) is a multi-lineage clonal hematopoietic malignancy characterized by excessive production of granulocytes and the presence in the leukemic cells of a consistent rearrangement of the BCR
and ABL genes, typically manifested in metaphase preparations as the Philadelphia chromosome (Ph1) (Goldman, Balliere's Clinical Haematology, Vol. 1, 1987). The initial cell transformed and hence the origin of the leukemic clone is believed to be a totipotent hematopoietic cell with lymphoid as well as myeloid differentiation potential. (Fialkow et al., Proc. Natl.
Acad. Sci., Vol. 58. p. 1468, 1967). The most primitive hematopoietic cells are difficult to study because they make up such a small proportion of all the nucleated cells in the blood and marrow. Nevertheless, effects of BCR-ABL
eY.pression in their behaviour are of key interest because it is th~se cells that are believed to be responsible for the initial amplification of the leukemic clones in patients with CML.

CML patients with elevated white blood cell (WBC) counts show dramatic increases in the number of Ph1-positive clonogenic progenitors in their circulation (Eaves et al., Bailliere's Clinical Haematology Vol. 4, pg. 931, 1990 -Dowding et al. Int. J. Cell. Cloning Vol. 4, p. 331, 1986). Continued production of Ph1-positive clonogenic cells for many weeks can occur at a high level when peripheral blood cells from such patients are cultured on irradiated marrow cell adherent layers established from normal individuals (Eaves et al., Proc. Natl. Acad. Sci., ': ' :

,. ~ . '.
:
- ~ :

- 4 ~ 208025~
Vol. 83, p. 5306, l986).

Allogenic bone marrow transplantation is a useful therapy for some leukemias, including CML, although the procedure is limited by the availability of compatible donors.
Autologous bone marrow transplantation is a less useful treatment for leukemia in general, and CML in particular, due to the difficulties of o~taining a bone marrow preparation which is devoid of primitive hematopoietic stem cells of leukemic origin and which may proliferate in the donor causing relapse. Studies of primitive leukemic hematopoietic cells have been limited by the lack of a suitable assay for their identification, and of methods for their selective isolation or ablation. In particular, the biological consequences of BCR-ABL kinase expression in very primitive human hematopoietic cells has been difficult to investigate because methods for their selective isolation have not been avAilable.

SUMMARY OF THE INVENTION
The present inventors have isolated the most primitive human hematopoietic stem cells yet defined from patients with chronic myeloid leukemia. It was surprisingly found that a highly enriched population of these cells could he obtained from the blood of CML patients with increased white blood cell counts. The primitive leukemic hematopoietic stem cells were found to be capable of generating clonogenic cells, on average 3-4 clonogenic cells, after at least 5 weeks of culture on competent feeder cells and accordingly are alternatively referred to herein as CML long term culture initiating cells (CML LTC-IC). The cells were also found to be relatively 4-hydroxy-cyclophosphamide sensitive, HLA-DR positive, rhodamine-bright and larger in size (higher FLS) than normal peripheral blood LTC-IC. The present inventors were also able to obtain a cell preparation having a purity of circulating CML LTC-IC of approximately 10~, which is 5 to ,, :: . ' ' ' . ~ . ~:

20802~5 6 fold higher than the purest populations of normal LTC-IC
thus far isolated from normal blood or marrow samples (Lansdorp PN et al, J Exp Med 172:363, 1990; Udomsakdi C
et al, Exp Hematol 19:338,1991; Sutherland HJ et al PNAS
USA 87: 3584, 1990).

The present inventors also have surprisingly found that CML LTC-IC cells can be quantitated by measurement of the number of clonogenic cells produced after at least 5 weeks in culture on competent fibroblasts. A linear relationship was found for the number of clonogenic cells present 5 weeks after seeding light density peripheral cells from CNL patients with high numbers of circulating Ph1 positive clonogenic cells onto irradiated normal marrow adherent layers and the number of peripheral blood cells initially added. Neasurement of the frequency of CML LTC-IC by limiting dilution analysis allowed derivation of their average clonogenic cell output at the 5 week time point.

The present inventors further surprisingly found that CML
LTC-IC are selectively disadvantaged in long term culture compared to normal LTC-IC.

Furthermore, the present inventors have found primitive hematopoietic stem cells i.e. LTC-IC, in normal blood.
- They have significantly shown that the number of clonogenic cells present in long term cultures of T cell-depleted fractions of normal blood after 5 weeks, is linearly related to the input number of peripheral blood cells over a wide range of cell concentrations, thereby permitting the quantitation of circulating LTC-IC by limiting dilution analysis. Using this approach, the concentration of LTC-IC in the circulation of normal adults was found to be 2.9 + 0.5 per ml. This is about 75-fold lower than the concentration of circulating clonogenic cells and represents a frequency of LTC-IC
relative to all nucleated cells that is about 100 fold .,.... . : :

- .. ~ . :

:,, .~ ' . .

2a8~2~5 lower than that measured in normal marrow aspirate samples. Characterization studies revealed most circulating LTC-IC to be small (low forward light scatter and side scatter), CD34+, Rh-123dUll, HLA-DR- and 4-S hydroperoxycyclophosphamide-resistant cells with differentiative and proliferative potentialities similar to LTC-IC in normal marrow. Isolation of the light-density, T-cell-depleted, CD34+, and either HLA-DR low or Rh-123dUll fraction of normal blood yielded a highly enriched population of cells that were 0.5 - 1% LTC-IC, a purity which is comparable to the most enriched populations of human marrow LTC-IC reported to date.

Therefore, the present invention relates to a cell preparation comprising primitive hematopoietic stem cells obtained from leukemic patients which cells are Philadelphia chromosome positive and produce detectable clonogenic progenitors in long term culture. In an embodiment of the invention, the cell preparation is obtained from the blood of CML patients with increased white blood cell counts and the primitive hematopoietic stem cells in the preparation are CD34~and HLA-DR+, FSChi9h, SCC~W, and Rh-123~Ve. Preferably the cell preparation has a purity of circulating primitive leukemic hematopoietic stem cells of approximately 10%.

The invention further provides a cell preparation comprising primitive hematopoietic stem cells obtained from the blood of normal individuals which cells produce detectable clonogenic progenitors in long term culture. In an embodiment of the invention, the primitive hematopoietic stem cells in the preparation are CD34'and HLA-DR-, ~sc.Ow SCCLW, and Rh-123dU~ and 4-hydroperoxycyclophosphamide-resistantwithdifferentiative and proliferative potentialities similar to LTC-IC in normal marrow. Preferably the cell preparation has a purity of circulating primitive hematopoietic stem cells .~. . , . ~ .:' . ' - . :
:, ' ' : -,, -: . : -.:

20~025S

of approximatel~ 0.5 - 1%.

The invention also provides a method for preparing a cell preparation comprising primitive hematopoietic ~tem cells obtained from leukemic patients which cells are Philadelphia chromosome positive and produce detectable clonogenic progenitors in long term culture comprising obtaining a sample from a leukemic patient, preferably a blood sample from a leukemic patient having an elevated white blood cell count, most preferably having a white blood cell count greater than 20 x 109 white blood cells per litre of blood; and isolating a cell preparation from the cell sample which contains cells which are Philadelphia chromosome positive and produces detectable clonogenic progenitors in long term culture. Preferably, the cell preparation obtained substantially comprises primitive leukemic hematopoietic stem cells which are CD34' HLA-DR+, FSChi9h, SCCLW, and Rh-123'Ve. Preferably, a cell preparation having a purity of circulating primitive leukemic hematopoietic stem cells of approximately 10% is obtained.

The invention further provides a method for preparing a cell preparation comprising primitive hematopoietic stem cells obtained from the blood of normal individuals which cells produce detectable clonogenic progenitors in long term culture comprising obtaining a blood sample from a normal individual, and isolating a cell preparation from the cell sample which contains cells which produce detectable clonogenic progenitors in long term culture.
Preferably, the cell preparation obtained substantially comprises primitive hematopoietic stem cells which are CD34~ and HLA-DR-, FSC~Wl SCClW, Rh-123dUL~ and 4-hydroperoxycyclophosphamide-resistantwithdifferentiative and proliferative potentialities similar to LTC-IC in normal marrow. Preferably, a cell preparation having a purity of circulating primitive hematopoietic stem cells ~ ! . , . ' .
: ' ' " ' ; ' , ,' , '` ~ ' ' ' ' ''' ~ '' -' 2~8~2~S

of approximately 0.5 to 1% is obtained.

The invention also relates to a method for quantitating primitive leukemic hematopoietic stem cells in patients with leukemia comprising obtaining a sample which contains primitive leukemic hematopoietic stem cells from a leukemic patient, preferably a blood sample from a leukemic patient having an elevated white blood cell count, preferably greater than 20 ~ 10~ white blood cells per litre of blood; optionally enriching primitive leukemic hematopoietic stem cells in the sample;
coculturing the sample with a feeder cell layer for at least five weeks under conditions which permit the production of clonogenic progenitor cells; harvesting nonadherent cells and adherent cells; subjecting the harvested cells to secondary assay culture under suitable conditions to express clonogenic progenitor cells;
detecting and quantitating the number of clonogenic progenitor cells; and, quantitating the number of primitive leukemic hematopoietic stem cells in the sample on the basis of the linear relationship between the number of clonogenic progenitor cells and the number of primitive leukemic hematopoietic stem cells initially in the sample.

The advantages of the above described method of the present invention include its relative simplicity, ease of standardization, and applicability to quantitation of primitive hematopoietic cells in primary patient samples.
Accordingly, the method may be used in the diagnosis of CML and to characterize the disease state of a patient.
Treatments for leukemia may be evaluated by determining the number and characteristics of primitive hematopoietic leukemic stem cells in the blood and bone marrow of a patient at time periods before and after treatment using the methods of the invention. The method may also provide a useful system for studying the interactions of normal and leukemic primitive hematopoietic stem cells in vivo.

: ~ . - . ., , , . :.: . ~ - . :
.. .. ~ , .

20~0255 g The invention further relates to a method for quantitating primitive hematopoietic stem cells in the blood of normal individuals comprising obtaining a blood sample which contains primitive hematopoietic stem cells from a normal individual, optionally enriching primitive hematopoietic stem cells in the blood sample; coculturing the sample with a feeder cell layer for at least five weeks under conditions which permit the production of clonogenic progenitor cells; harvesting nonadherent cells and adherent cells; subjecting the harvested cells to secondary assay culture under suitable conditions to express clonogenic progenitor cells; detecting and quantitating the number of clonogenic progenitor cells;
and, quantitating the number of primitive hematopoietic stem cells in the blood sample on the basis of the linear relationship between the number of clonogenic progenitor cells and the number of primitive hematopoietic stem cells initially in the sample.

The invention further relates to a method of purging a mammalian bone marrow sample containing primitive leukemic hematopoietic stem cells to prepare a bone marrow cell suspension substantially free of primitive leukemic hematopoietic stem cells comprising: obtaining a sample of bone marrow cells from a leukemic patient; substantially depleting red blood cells in the sample; coculturing the depleted bone marrow cell sample with a feeder cell layer;
and harvesting and suspending the cultured bone marrow cells. The bone marrow cell suspension may be further cultured on a feeder cell layer; and the cultured bone marrow cells may be harvested and cultured.

The invention still further relates to a system for testing for a substance that affects hematopoiesis of primitive hematopoietic stem cells comprising: preparing a cell suspension that is enriched in primitive hematopoietic stem cells; coculturing the cell suspension , . . .

~ '. - . ':

with a feeder cell layer for at least 5 weeks in the presence of a substance which is suspected of affecting the hematopoiesis of primitive hematopoietic stem cells and assessing LTC-IC maintenance, clonogenic cell production, and/or production of nonadherent cells. The cell suspension may comprise primitive hematopoietic stem cells obtained from the blood of leukemic patients, preferably from the blood sample of a leukemic patient having an elevated white blood cell count, or from the blood of normal individuals. The substance may be added to the culture of the cell suspension and feeder cell layer or the feeder cell layer may be genetically engineered to express the substance i.e. the feeder cell layer may serve as an endogenous source of the substance.

DESCRIPTION OF TH~ DRAWINGS
The invention will now be described in relation to the drawings in which:
Figure 1 is a graph showing the relationship between the number of CML peripheral blood cells seeded into long term culture and the number of clonogenic cells detected after 5 weeks;
Figure 2 is a graph showing limiting dilution - analysis of CML peripheral blood cells seeded in culture in decreasing numbers and assayed for clonogenic cells at 5 weeks;
Figure 3 is a graph showing the LTC-IC
concentration in the peripheral blood of CML patients and normal individuals;
Figure 4 is a graph showing the differential kinetics of CML (solid symbols) and normal (open symbols) long term culture initiating cells in vitro;
Figure 5 is a diagram showing the distribution, according to light scatter characteristics, of total cells, clonogenic cells and LTC-IC in the light density fraction of CML blood;
Figure 6 is a diagram showing bivariate contour .
... . . . .

: - .. : , . .

208025~

plots of a representative sample of normal and CML light density blood cells in the low SSC window CD34~DR~W, CD34+DRhi~h CD34+RH-123dUl~andCD34+Rh-123bri9htsubpopulations;
Figure 7 is a bar graph showing the distribution S of clonogenic cells and LTC-IC within the CD34~ fraction of circulating CNL cells subdivided according to their high or low expression of HLA-DR;
Figure 8 is a bar graph showing the distribution of clonogenic cells and LTC-IC within the CD34+ fraction of circulating CML cells subdivided according to their uptake of Rh-123;
Figure 9 is a bar graph showing survival of circulating CML clonogenic cells and LTC-IC after a brief exposure to 4-HC;
Figure 10 is a bar graph showing the number of nonadherent (N~) cells in 5-week old cocultures initiated with equal numbers of sorted cells seeded onto M2-lOB4 feeders producing human growth factors or human MF is compared with NA cells in 5-week old cocultures containing control N2-lOB4;
Figure 11 is a bar graph showing the number of clonogenic cells in 5-week old cocultures containing M2-lOB4 cells producing human growth factors or human NF
compared with control M2-lOB4 cells;
Figure 12 is a bar graph showing the number of LTC-IC as determined by limiting dilution analysis in 5-week old cocultures initiated with equal numbers of sorted cells seeded onto human growth factor producing M2-lOB4 cells or human NK compared with the LTC-IC content of 5-week old cocultures containing control M2-lOB4 cells;
Figure 13 shows the linear relationship between the number of light density ((1.077g/cm3) T cell-depleted peripheral blood cells from a representative normal individual seeded onto pre-established, irradiated normal marrow feeders and the total number of clonogenic cells detected when these LTC were harvested and assayed in methylcellulose 5 weeks later;

~' ' .
: : -20~0255 Figure 14 shows limited dilution analysis of data from a representative experiment in which decreasing numbers of light density T cell-depleted normal peripheral blood cells were seeded onto irradiated marrow feeders and the number of clonogenic cells detectable after 5 weeks was then determined;
Figure 15 shows bivariate contour histograms of light density T cell-depleted normal peripheral blood cells stained with anti-CD34 and anti-HLA-DR;
Figure 16 shows light scatter profiles of T
cell-depleted light density normal blood cells (Panel A) and the mean + SEM of the percentages of nucleated cells (open bar), clonogenic cells (stippled bar), and LTC-IC
(solid bar) in each sorted fraction (Panel B);
Figure 17 show a representative histogram of CD34+, light density T cell-depleted normal blood cells (in the previously described low SSC window shown in Figure 4A) double-stained with PE conjugated anti-HLR-DR;
Figure 18 show a representative histogram of CD34~, light density T cell-depleted normal blood cells double-stained with Rh-123 and sorted into CD34~RH-123 and CD34~RH-123bri9ht fractions (fractions 1 and 2, respectively; Panel A) and the mean + SEN of the percentages of nucleated cells (open bar), clonogenic cells (stippled bar), and LTC-IC (solid bar) in each sorted fraction are shown in (Panel B) (n=3); and Figure 19 shows a comparison of the number of clonogenic cells and LTC-IC surviving a 30 minute exposure to lOO~g/ml of 4-HC at 37C with 7% erythrocytes present.

DETAILED DESCRIPTION OF THE INVENTION

As hereinbefore mentioned the invention relates to a cell preparation comprising primitive hematopoietic stem cells obtained from leukemic patients that are Ph1-positive and capable of initiating long term culture and producing clonogenic precursors. The stem cells may be further . .; , . .
.. . . , . , ~ . . .
.

- - . .: , 20802~5 characterized as CD34~ and DR', FSChi~h, SCCIW, and/or Rh~
123~Ve. Preferably the cell preparation has a purity of circulating primitive leukemic hematopoietic stems cell~
of appro~imately 10~.

The invention further provides a cell preparation comprising primitive hematopoietic stem cells obtained from the blood of normal individuals which cells produce detectable clonogenic progenitors in long term culture.
The primitive hematopoietic stem cells in the preparation may be further characterized as CD34', and HLA-DR-, FSCLW, SCC~W, and Rh-123dUll and 4-hydroperoxycyclophosphamide-resistant with differentiative and proliferative potentialities similar to LTC-IC in normal marrow.
Preferably the cell preparation has a purity of circulating primitive hematopoietic stem cells of approximately 0.5 ~

The primitive hematopoietic stem cells are also referred to herein as LTC-IC and primitive hematopoietic stem cells from a patient with leukemia or CML are referred to as leukemic or CNL LTC-IC. The term LTC-IC used herein refers to a cell that after a minimum period of 5 weeks in culture together with certain marrow adherent cells, but in the absence of exogenous growth factors, produces detectable clonogenic progenitor cells.

As hereinbefore mentioned the invention also relates to a method for preparing a cell preparation comprising primitive hematopoietic stem cells, which are characterized as Ph1-positive and capable of producing clonogenic progenitors detectable in long term culture comprising obtaining a sample from a leukemic patient, preferably a blood sample from a leukemic patient having an elevated white blood cell count, most preferably having a white blood cell count greater than 20 x 109 white blood cells per litre of blood; and isolating a cell preparation ".. , ~ . , :

- - , :
. . . .~:

' ~:~: , ':

20~02~

from the cell sample which is Philadelphia chromosome positive and produces detectable clonogenic progenitors in long term culture. Preferably, the method provides a cell preparation substantially comprising primitive leukemic hematopoietic stem cells which are CD34~ HLA-DR+, FSChigh, SCC~W, and Rh-123~Ve. Preferably, a cell preparation having a purity of circulating primitive leukemic hematopoietic stem cells of approximately 10% is obtained.

The cell preparation may be prepared from blood samples of leukemia patients, preferably from CNL patients having an elevated white blood cell count, most preferably containing greater than 20 x 109 white blood cells per liter. Normal (Ph1-negative) LTC-IC persist in the marrow of many CNL patients (and to a much lesser extent, normal clonogenic cells). Blood samples having elevated white cell counts were found to be a source of cells that are reproducibly, significantly, and preferentially enriched for neoplastic progenitors, eliminating the need for laborious genotyping studies. Thus normal cells, even if present in such samples at normal levels~ remain well below the limit of detectability in the methods of the present invention (as illustrated by the calculations shown in Table 4). The present inventors have also shown that more than 97% of LTC-IC from CML patients with high white blood cell counts and markedly elevated LTC-IC
concentrations exhibit abnormal functional properties (i.e., self-maintenance) in LTC, consistent with a leukemic origin.

The present inventors have shown that the primitive Ph1-positive primitive hematopoietic stem cells differ from their counterparts in normal individuals with respect to a number of functionally related properties. The differences are suggestive of a deregulation in the control of cell proliferation in CML at the level of the cells initially responsible for maintenance and expansion :
: - -, , :
.

: ~ , . .

2~80255 of the Ph~-positive clone without alteration of their commitment to, or early differentiation down, each of the hematopoietic lineages. It is expected that these findings may provide a useful framework for future analysis of the mechanism of BCR-ABL-induced multi-lineage disease.

The majority of both clonogenic cells and LTC-IC in the circulation of CML patients with high white blood cell counts are phenotypically similar to one another with respect to size (FSC), expression of CD34 and HLA-DR, uptake of Rh-123 and sensiti~ity to 4-HC. In both cases, the predominant phenotype is that of proliferating or activated cells (i.e., high FSC, high expression of HLA-DR, high Rh-123 uptake and relative sensitivity to 4-HC).
However, subtle differences between circulating clonogenic cells (more activated) and LTC-IC (less activated) in CML
patients are consistently noted. This predominant, I'abnormal'' phenotype, described herein is essentially the opposite of that previously shown for the majority of clonogenic cells and LTC-IC in the circulation of normal adults, also shared by the majority of LTC-IC in normal marrow ti.e., low FSC, low expression of HLA-DR, low Rh-123 uptake and relative insensitivity to 4-HC).

Separation techniques based on one or more of the distinct phenotypic characteristics of the primitive leukemic hematopoietic stem cells may be used to isolate the cells from a sample. A cell preparation comprising primitive leukemic stem cells may be isolated fro~l a sample based on the light scattering properties of the primitive leukemic stem cells. The stem cells may be found in cell fractions with high FSC and low SSC or fractions with low FSC and low SSC. The cell preparation may also be isolated based on other phenotypic characteristics of the stem cells, for example, their ability to express detectable levels of HLA-DR or CD34 or their ability to retain Rh-123.

''~ : . ,. .' ':

2~8025~

~enerally, the primitive leukemic hematopoietic stem cells are CD34+, ~LA-DRhi9h and Rh-123'Ve.

In one embodiment of the invention, a cell preparation comprisi~g primitive leukemic hematopoietic stem cells may be isolated by first removing erythrocytes, granulocytes and platelets from the sample. In particular, a light density fraction (C1.077 g/cm3) may be isolated by centrifugation of the blood on Ficoll/Hypaque (FH) to eliminate the majority of erythrocytes, granulocytes and platelets. T-cell depletion is generally not required, since it was found that the number of T cells in initial CML blood samples was at or below 2~ of the total.
. .
The light density fraction may be fractionated as generally described below, to obtain a fraction enriched in CD34'and DR~ cells. In particular, cells, in the light density fraction may be washed and resuspended in a physiologic solution, for example Hank's solution with .
fetal calf serum (FCS) containing sodium azide (NaN3)(HFN).
The suspended cells may be stained with an anti-CD34 antibody, for example 8G12, directly conjugated to a detectable marker, preferably a fluorescent label, most preferably phycoerythrin (PE) or fluorescein isothiocyanate (FITC), following the methods described in Lansdorp et al., J. Exp. Med., Vol. 172, p.363, 1990, then double-stained with Rhodamine, preferably Rhodamine-123 (RH-123) (Udomsakdi et al., Exp. Hematol, Vol. 19, p. 338, 1991) or antibody specific for HLA-DR conjugated to a detectable marker, preferably with a fluorescent label, most preferably anti-HLA-DR-PE may be used (Sutherland et al., Blood Vol. 74, p. 1563, 1989).

One skilled in the art will appreciate that the detectable markers conjugated to the HLA-DR and CD34 antibodies should be selected so as to permit measurement of the binding of the two antibodies to their respecti~e .^--. - -~. . ~ ., .
: ~,. ` ' :, . , ' ,~ : '''' ' ' . ' ~ ' ' '~

20802~

antigens; that is, the detectable markers should be easily distinguished from each other to facilitate simultaneous measurement.

The stained, fluorescently labelled cells may be analyzed and sorted using a fluorescence-activated cell sorter (FACS) such as the Becton Dickenqon FAC Star Plus.
Fluorescence of Rh-123, FITC and PE-labelled cells may be measured using 530/30 and 575/26 band pass filters, respectively, after calibration of the FACS prior to each sort, preferably with 10 ~m fluorescent beads. Gates may be set to exclude most of the granulocytes and erythrocytes using forward light scatter (FSC) and side scatter (SSC) characteristics as described in Udomsakdi et al., Exp. Hematol, Vol. 19, p. 338, 1991. Cells appearing ` within this light scatter window (see Figure 5A) constitute on average 15-20% of the total light density fraction of CML blood cells. Sorted cells may be collected in physiological solution, preferably in Hank's solution with 50% FCS and may be maintained at, for example 4C until required.

Sorted cells may be used to initiate long term cultures, by the methods more particularly described below. The clonogenic progenitor cells resulting from long term culture may be assayed for clonogenic erythropoietic (BFU-E), granulopoietic (CFU-GM), and multilineage (CFU-GEMM) progenitors as described below.

The cell preparation comprising primitive leukemic hematopoietic stem cells of the invention may also be used to characterize the differentiative potential of CML LTC-IC. The relative numbers of different types of clonogenic progenitors present in 5 week-old LTC ( see discussion below) provides a consistent average overall measure of the differentiative behaviour of LTC-IC assayed under standard LTC conditions. To assess whether this parameter ...- . . ... ...
... , :

20g~25S

is altered in the LTC-IC present in CML blood, the ratio of BFU-E, CFU-GM and CFIJ-GEMM numbers before and after LTC
of light density CML blood cells may be assessed.

As hereinbefore mentioned the invention further provideR
a method for preparing a cell preparation comprising primitive hematopoietic stem cells obtained from the blood of normal individuals which cells produce detectable clonogenic progenitors in long term culture comprising obtaining a blood sample from a normal individual, and isolating a cell preparation from the cell sample which contains cells which produce detectable clonogenic progenitors in long term culture. Preferably, the cell - preparation obtained substantially comprises primitive hematopoietic stem cells which are CD34' and HLA-DR-, FSC~W, SCC~W, Rh-123dULL and 4-hydroperoxycyclophosphamide-resistant with differentiative and proliferative potentialities similar to LTC-IC in normal marrow.
Preferably, a cell preparation having a purity of circulating primitive hematopoietic stem cells of approximately 0.5 to 1% is obtained.
.~ .
Separation techniques based on one or more of the distinct phenotypic characteristics of the primitive hematopoietic stem cells may be used to isolate the cells from a blood sample. A cell preparation comprising primitive stem cells 25 - may be isolated from a sample based on the light scattering properties of the primitive stem cells. The stem cells may be found in cell fractions with low FSC and low SSC. The cell preparation may also be isolated based on other phenotypic characteristics of the stem cells, for example, their ability to express detectable levels of HLA-DR or CD34 or their ability to retain Rh-123.
Generally, the primitive leukemic hematopoietic stem cells are CD34~, HLA-DR-Ve and Rh-123dU~L. Specific protocols for isolating LTC-IC from the blood of normal individuals are set forth in the examples herein.

.;,., ~ .
-: - .

- ~ ~
~ ' 208025~

As hereinbefore mentioned, the invention provides a method for quantitating primitive leukemic hematopoietic stem cells in patients with leukemia comprising obtaining a cell sample which contains primitive lsukemic hematopoietic stem cells from a leukemic patient;
optionally enriching primitive leukemic hematopoietic stem cells in the sample; coculturing the sample with a feeder cell layer for at least five weeks under conditions which permit the production of clonogenic progenitor cells;
harvesting nonadherent cells and adherent cells;
subjecting the harvested cells to secondary assay culture under suitable conditions to express clonogenic progenitor cells; detecting and quantitating the number of clonogenic progenitor cells; and, quantitating the number of primitive hematopoietic stem cells in the sample on the basis of the linear relationship between the number of clonogenic progenitor cells and the number of primitive leukemic hematopoietic stem cells initially in the sample.
The above method may be used to ~uantitate CML LTC-IC.

The above method of the invention may be carried out on heparinized cell samples obtained from the bone marrow or blood, preferably the peripheral blood, of a mammal, preferably a human. In a preferred embodiment, the cell sample may be obtained from a patient with CML, having an - 25 elevated white blood cell count, preferably of >20 x 104 white blood cells per litre of blood. As discussed above such samples are substantially devoid of normal LTC-IC
hematopoietic stem cells and are enriched in leukemic LTC-IC hematopoietic stem cells. Thus it is not necessary to distinguish normal and leukemic LTC-IC as is required for bone marrow cell samples, as described below.

The cell sample may be pre-treated prior to coculturing with the feeder layer to enrich primitive leukemic hematopoietic stem cells in the sample. Red blood cells may be removed from the cell sample, for example by a ... .
.': ' . .

.

20802~

brief exposure to ammonium chloride. For blood cell samples, a preparation of light density cells ~for example <1.077 gm/cm3) may be isolated by centrifugation on Ficoll-Mypaque. Preferably, where the initial cell sample is obtained from the blood of a mammal without an elevated white blood cell count, the sample may be depleted of T
cells. T cell depletion may be effected, for example by incubation of the light density cells with 2-amino-ethylisothiouronium bromide-treated sheep erythrocytes for 30 minutes on ice followed by further centrifugation to remove the rosetted T cells as described generally by Marsden et al., J. Immunol. Nethods, Vol. 33, p. 323, 1980. T cell removal prevents the spontaneous activation and outgrowth in vitro of Epstein-virus transformed B
lymphocytes. Since the T cell content of the high white blood cell count peripheral blood specimens obtained from CML patients is already decreased to a few percent, further removal of T cells from these samples may not be necessary.

An enriched blood sample which is CD34' and DR~ may also be used in the method. Such an enriched sample may be obtained by treating the blood sample with antibodies to CD34 labeled with a detectable marker and antibodies to HLA-DR labeled with a detectable marker, wherein the antibodies to CD34 and antibodies to HLA-DR are labeled with different detectable markers, and isolating an enriched sample which is CD34+ and DR~ by means of the labels. Preferably the detectable marker is a fluorescent label, most preferably phycoerythrin (PE) or fluorescein isothiocyanate (FITC) (Lansdorp et al., J. Exp. Med., Vol.
172, p.363, 1990).

The samples or enriched samples are cocultured with a feeder cell layer for at least five weeks under conditions which permit the production of clonogenic progenitor cells. Examples of suitable feeder cells include .~. ..
- ~

~ ', - ~ ' .

208û2~5 irradiat~d normal marrow adherent cell layers, subcultured from primary LTC. Sutherland et al., Blood, Vol. 74, p.
1563, 1989 provide a discussion of suitable feeder layers.
It will be appreciated that other adherent cells, such as murine fibroblasts may be used as the feeder cell layer.

The culture conditions are as generally described in Eaves et al., J. Tissue Culture Methods, Vol. 13, p. 55, 1991.
In particular the culture may be maintained at physiological temperatures, preferably in the range from 30-40C, preferably 32-37C. In a preferred embodiment the cultures are maintained at 37C for the first 3-4 days and then subsequently at 30-35C, most preferably 33C. P.fter the first 7 to 10 days half of the medium and half of the nonadherent cells may be removed and replaced with new LTC
medium. This feeding procedure may then be repeated periodically, preferably after 14 days and at weekly intervals thereafter.
. . .
After at least 5 weeks, nonadherent cells and adherent cells may be harvested. The adherent cells may be harvested using known techniques for example using trypsin.

The harvested cells are subjected to secondary assay culture under suitable conditions to express clonogenic progenitor cells and the number of clonogenic progenitor cells are detected and quantitated. The secondary assay culture may be a semi-solid assay culture, for example a methylcellulose culture, and erythropoietic (BFU-E), granulopoietic (CFU-GN) and multi-lineage (CFU-GENN) progenitors may be detected and quantitated. Preferably, the assay is performed in a standard methylcellulose culture containing human erythropoietin and agar-stimulated human peripheral leukocyte conditioned medium (Terry Fox Laboratory Media Preparation Service, Vancouver, BC~. Harvested cells may be plated at .
'. " . ;' ' ' ' : ~ , ~

appropriate concentrations preferably 5 x 105 cells per 1.1 ml assay, in replicate methylcellulose cultures under standardized conditions optimized for expression of the colony-forming potential of these cells as assessed after 18-21 days incubation at 37C using previously described colony scoring criteria (Cashman et al., Blood, Vol. 66, p. 1002, 1985).

The methodology for colony generation and criteria for colony recognition are generally as described in Coulombel et al., Blood, Vol. 62, p. 291, 1983. Total clonogenic cell numbers refers to the sum of BFU-E, CFU-GM, CFU-GENM
detected in direct assays using these procedures.

Where the original cell sample may be contaminated with normal primitive hematopoietic stem cells, primitive leukemic hematopoietic stem cells may be identified by genotyping the colonies produced in the secondary culture.
Genotyping may be carried out by cytogenetic analysis of individually removed single or small pools of colonies as generally described in Fraser et al., Cancer Genet.
Cytogenet., Vol. 24, p.1, 1987, to allow the proportion of normal and leukemic clonogenic cells to be determined.

Where CML bone marrow cell samples are used in the method of the invention, it is necessary to perform cytogenic analysis on the colonies produced from the clonogenic progenitors to distinguish Phl-positive and Ph1-negative LTC-IC. In contrast to CML blood, Ph1-positive primitive leukemic hematopoietic stem cells would be expected to represent a minority population relative to normal LTC-IC
in CML marrow (Oster et al., Blood (abstr) Vol. 70, p.
266a 1987 - Barnett et al., Bone Narrow Transplant, Vol.
4, p. 345, 1989). The concentration of Ph1-positive LTC-IC

... . . .

.
.- ~

-- :

(relative to other nucleated cells) in CML marrows has been found to be quite variable and, in general, is markedly reduced, both by comparison to LTC-IC values in control marrows and by comparison to normal (Ph1-negative) LTC-IC co-existing in the same CML marrow.

The relative number of primitive hematopoietic stem cells in the sample is quantitated on the basis of the linear relationship between the number of clonogenic progenitor cells and the number of primitive leukemic hematopoietic cells cocultured with the feeder cell layer.

It is possible to quantitate absolute primitive leukemic hematopoietic stem cells in the sample by employing limiting dilution analysis in the method of the invention.
For such methods, samples containing varying concentrations of primitive leukemic hematopoietic stem cells, for example, 100 ~1 aliquot samples containing from 50 to 2 x 105 light density cells may be cocultured with a feeder cell layer and the harvested cells may be plated in methylcellulose assay cultures to enable detection of one or more clonogenic cells in each sample. From the proportion of positive and negative primitive leukemic hematopoietic stem cells defined in this way, the frequency of primitive leukemic hematopoietic stem cells in different input samples may be calculated preferably using Poisson statistics (Taswell, C., J. Immunol. Vol.
126, p. 1614, 1981 - Coller et al., Methods in Enzymology, Vol. 121, p. 412, 1986).

The proliferative potential of the primitive hematopoietic stem cells as indicated by the average 5 week output of clonogenic cells per primitive leukemic hematopoietic stem cells may then be derived in each case. Knowledge of the 5 week clonogenic cell output per primitive leukemic hematopoietic stem cells allows absolute values to be derived from total 5 week clonogenic cell yields measured . .

. , - . .
: , - :' .' .:

208~25~

in cul~ures initiated with non-limiting inocula.
Primitive leukemic hematopoietic stem cells values may thus be obtained for samples, and the concentration of primitive leukemic hematopoietic stem cells per ml of blood then calculated assuming 100~ LTC-IC recovery in the light density fraction assayed (Sutherland et al., Blood, Vol. 74, p. 1563, 1989).
.

The present inventors have demonstrated that, on average, primitive leukemic hematopoietic stem cells circulate at a 10-fold lower frequency than clonogenic cells although these two parameters showed a highly significant association. By comparison the ratio of circulating primitive leukemic hematopoietic stem cells to clonogenic cells in normal blood appears to be much lower.
.

Primitive CML hematopoietic stem cells were found to produce on average, a similar number of clonogenic cell progeny after 5 weeks cocultured with a feeder cell layer as do their normal counterparts in the blood or marrow of normal individuals. However, a number of abnormalities in the primitive CML hematopoietic stem cell population were also revealed. First, their distribution between marrow and blood was shown to be grossly altered, even more dramatically than is the case for Ph1-positive clonogenic cells. Both populations increase exponentially in the blood with linear increases in the white blood cell count, but Ph1-positive primitive hematopoietic stem cells appear to be present at relatively reduced frequencies in CNL
marrow whereas Ph1-positive clonogenic cell frequencies in CNL marrow are relatively normal (Eaves et al., Exp.
Hematol., Vol. 8, p. 235, 1980).

In human long term cultures, primitive normal clonogenic cells in the adherent layer alternate weekly between a quiescent and a dividing state (Cashman et al., Blood, Vol. 66, p. 1002, 1985) and in murine long term cultures, ~.

20~02~

it has been possible to demonstrate that extensive proliferation of some long-term totipotent reconstituting cells does occur. In long term cultures initiated with primitive leukemic hematopoietic stem cells, their derivative primitive clonogenic progeny divide continuously suggesting a defective but unregulated mechanism for inevitable expansion of the Ph~-positive clone. Thus the method of the invention may serve as an important model for further dissection of the mechanisms that regulate normal versus CNL recovery patterns in vivo.

As hereinbefore mentioned, the invention also relates to the use of the above described method of the invention in the diagnosis of CML and in characterizing the disease state of a patient with leukemia by quantitating primitive leukemic hematopoietic stem cells in the blood and bone marrow of a patient. The efficacy of various treatments for leukemia, including chemotherapy, radiation therapy and bone marrow transplants may be assessed using the method of the invention by determining the number and characteristics of primitive hematopoietic leukemic stem cell~ in the blood and bone marrow of the patient at time periods before and after treatment.

Thus, the quantitative assay of the invention for a primitive leukemic hematopoietic stem cell population facilitates a variety of studies to further characterize these cells, obtain a better estimate of the number of leukemic stem cells in individual leukemia patients, and to devise more effective treatment strategies both in and ex vivo.

As hereinbefore mentioned the invention further relates to a method for quantitating primitive hematopoietic stem cells in blood samples of normal individuals comprising obtaining a blood sample which contains primitive hematopoietic stem cells from a normal individual, :: ~

-: ., .
.
: . , :.. : : - :

208025~

optionally enriching primitive hematopoietic stem cells in the blood sample; coculturing the sample with a feeder cell layer for at least five weeks under conditions which permit the production of clonogenic progenitor cells;
harvesting ncnadherent cells and adheren$ cells;
subjecting the harvested cells to secondary assay culture under suitable conditions to express clonogenic proqenitor cells; detecting and quantitating the number of clonogenic progenitor cells; and, quantitating the number of primitive hematopoietic stem cells in the blood sample on the basis of the linear relationship between the number of clonogenic progenitor cells and the number of primitive hematopoietic stem cells initially in the sample.

The blood samples from normal individuals may be blood samples containing peripheral blood mononuclear cells obtained from normal individuals as a byproduct of plateletphereses. Samples may be depleted of T cells by incubation with 2-aminoethylbromide isothiournium-treated sheep red blood cells and subsequent isolation of a light density fraction after centrifugation on Ficoll-Hypaque.
An enriched blood sample from a normal individual which is CD34' may also be used in the method. Such an enriched sample may be obtained by treating the blood sample with antibodies to CD34 labeled with a detectable marker and isolating an enriched sample which is CD34~by means of the labels. Preferably the detectable marker is a fluorescent label, most preferably phycoerythrin (PE) or fluorescein isothiocyanate (FITC) (Lansdorp et al., J. Exp. Med., Vol.
172, p.363, 1990).

The blood sample or enriched blood sample may be cocultured with a feeder cell layer to produce clonogenic progenitor cells and the clonogenic progenitor cells may be detected and quantitated using the procedures outlined above. The number of primitive hematopoietic stem cells in the blood sample is determined on the basis of the linear ' 208025~

relationship between the number of clonogenLc progenitor cells and the number of primitive hematopoietic stem cells initially in the sample as discus-~ed previously.

The invention further relates to a system for testing for a substance that affects hematopoiesis of primitive hematopoietic stem cells comprising: preparing a cell suspension that is enriched in primitive hematopoietic stem cells; coculturing the cell suspension with a feeder cell layer for at least 5 weeks in the presence of a substance which is suspected of affecting the hematopoiesis of primitive hematopoietic stem cells and assessing LTC-IC maintenance, clonogenic cell production, and/or production of nonadherent cells.

The cell suspension may comprise primitive hematopoietic stem cells obtained from the blood of normal individuals.
The cell suspension may also be obtained from blood or bone marrow of leukemic patients, preferably from the blood of a leukemic patient having an elevated white blood cell count and the primitive hematopoietic stem cells may be Philadelphia chromosome positive and produce detectable clonogenic progenitors in long term culture. Preferably the cell suspension is highly enriched in primitive hematopoietic stem cells. Methods for pre-treating samples to obtain enriched preparations of primitive hematopoietic stem cells have been described above.

The substance may be added to the culture of the cell suspension and feeder cell layer, or the feeder cell layer may be genetically engineered to express the substance i.e. the feeder cell layer may serve as an endogenous source of the substance. Nurine marrow-derived stromal cell lines such as N2-lOB4, may be engineered by retroviral-mediated gene transfer to produce specific substances such as specific human factors. Examples of specific human growth factors which can be produced using . .
; ~' , " :
- ~ ' :, ` ~ ::
.: . .
' -~ .

208025~

the engineered feeder cell layer are G-CSF, GM-CSF, and interleukins such as IL-3, IL-4 and IL-6. The engineered cell feeder layer is constructed and maintained such that it releases the substance into the medium for the desired period of time for the long term culture. Through the use of genetically engineered growth factor-producing feeders it may be possible to reproduce the way in which regulator substances may be localized in the adherent layer and/or presented to adjacent hematopoietic cells on the assumption that this might influence the nature and magnitude of their effects. The feeder cell layer may be human marrow stromal cells which may be induced to produce human factors such as G-CSF and GN-CSF.

The system may be used to analyze the affects of substance(s) on different stages of hematopoiesis.
Effects on cells at very early, intermediate, and late stages of hematopoiesis may be evaluated by assessing the number of clonogenic cell precursors, clonogenic cells, and mature granulocyte and macrophage progeny present in the cultures after 5 weeks.

LTC-IC maintenance, clonogenic cell production and production of nonadherent cells may be assessed using the methods described herein.

The present inventors have surprisingly found that CML
LTC-IC are selectively disadvantaged in long term culture compared to normal LTC-IC. In particular, the present inventors have shown that, in spite of a normal output of clonogenic cell progeny by Ph1-positive LTC-IC and the provision of a pre-established feeder layer derived from a normal marrow donor, their initial maintenance in the - ' ' 20~025~

LTC system was highly compromised relative to normal LTC-IC. The beh~viour of normal and leukemic LTC-IC in the LTC may indicate how these cells behave in vivo under analogous conditions of stimulation.

Accordingly, the invention further relates to a method of purging a mammalian bone marrow sample containing primitive leukemic hematopoietic stem cells to prepare a bone marrow cell suspension substantially free of - primitive leukemic hematopoietic stem cells comprising culturing the bone marrow sample in long term culture in vitro. It will be appreciated that such a purging method will be useful for removing leukemic precursor cells from a bone marrow sample to be introduced into a human recipient, for example in the case of autologous bone marrow transplan~ for a leukemia patient. The preparation of samples and the long term culture procedure have been previously described herein.

The following non-limiting examples are illustrative of the present invention.
EXA~PLES
Example 1 Preparation of Cell Samples Heparinized bone marrow (BN) aspirate cells were obtained with informed consent from normal individuals and Ph1-positive CML patients undergoing marrow harvests for transplantation. Heparinized normal blood from additional normal individuals and CNL blood from CML patients undergoing routine hematologic assessment was similarly obtained. All CML patients were Ph1-positive and in stable chronic phase. For the initial experiments with CNL
blood, only samples containing >20 x 109 white blood cells (WBC) per L were used, as this allowed selection of patients whose circulating Ph1-positive progenitors were sufficiently elevated that even after maintenance in LTC
only Ph1-positive cells were detected (see Figure 3A) thus t , ~
- ., ' ~

20~025~

avoiding the necessity for confirmatory progenitor genotyping as is required for similar experiments with CML
BM .

BM cells for initiation of LTC were either used without further processing or after removal of contaminating red blood cells when the nucleated cell count in the original specimen was less than 2 x 107 cells/ml. For clonogenic cell assays, red cells were first lyzed by brief exposure to ammonium chloride generally as described in Turhan et al., N. Engl. J. Med., Vol. 320, p. 1655, 1989, and the cells then washed twice in Iscove's medium plu6 2~ fetal calf serum. For blood cell samples, light density (<1.077 gm/cm3) cells were isolated by centrifugation on Ficoll-Hypaque either with (normal blood) or without ( CML blood) T cell-depletion. This involved incubation of the light density cells with 2-aminoethylisothiouronium bromide-treated sheep erythrocytes for 30 minutes on ice followed by further centrifugation at 4C on Ficoll-Hypaque to remove the rosetted T cells, generally as described in Marsden et al., J. Immunol. Methods, Vol. 33, p. 323, 1980. These were reduced, on average, by this procedure by more than 98~ according to FACScan analysis of CD2-positive cells in a lymphocyte gated population. The primary purpose of T cell removal was to prevent the spontaneous activation and outgrowth in vitro of Epstein-virus transformed B lymphocytes which preliminary experiments showed occurs regularly within 5 weeks when inadequately T cell--depleted normal peripheral blood samples were co-cultured with irradiated marrow adherent layers. Since the T cell content of the high WBC count peripheral blood specimens obtained from CML patients was already decreased to a few percent, further removal of T
cells from these samples was not undertaken and development of spontaneous transformants was not encountered.

:~

2~80255 Example 2 Long-Term Cultures (LTC) Test cells, prepared as described in Example 1, were cultured in LTC medium generally as described in Eaves et al., J. Tissue Culture Methods, Vol. 13, p. 55, 1991, at varying initial cell concentrations in dishes or wells containing (or not containing) a standardized number of irradiated (15 gy) normal marrow adherent layer cells subcultured from primary LTC (Sutherland et al., Blood, Vol. 74, p. 1563, 1989) according to the particular experimental design. All LTC were maintained at 37C for the first 3-4 days and then subsequently at 33C. After the first 7 to 10 days half of the medium and half of the nonadherent cells were removed and replaced with new LTC
medium. This feeding procedure was then repeated after 14 days and at weekly intervals thereafter. For all LTC-IC assays, cultures were initiated on normal marrow feeder layers, as described above, and then maintained for 5 weeks. At this time all of the nonadherent cells and the trypsinized adherent cells were harvested, washed and plated in methylcellulose cultures for quantitation of clonogenic cells as described below.

For LTC-IC maintenance studies, the entire contents of primary LTC were harvested at the times indicated, aliquots used to initiate secondary LTC on new irradiated feeders and these secondary LTC were then maintained a further 5 weeks prior to harvesting and plating of the cells in methylcellulose assays. The number of clonogenic cells detected at this time provided a relative measure of the LTC-IC in the primary LTC at the time they were harvested. This value was then normalized by the number of LTC-IC detected in primary LTC-IC assays of the original cell suspension, to yield a percent input value.

Example 3 Colony Assays , . .
.-, ~

..
: ; -~ ~ - . .:

~ - . ' - - ~ .

Erythropoietic (BFU-E), granulopoietic (CFU-GN and multi-lineage (CFU-GEMM) progeni~ors were assayed by plating test cell suspensions at appropriate concentrations in replicate methylcellulose cultures under standardized conditions optimized for expression of the colony-forming potential of these cells as assessed after 18-21 days incubation at 37C using previously described colony scoring criteria (Cashman et al., Blood, Vol. 66, p. 1002, 1985). In all assays of fresh or cultured CML marrow, colonies produced in methylcellulose were genotyped by cytogenetic analysis of individually removed single or small pools of colonies (Fraser et al., Cancer Genet.
Cytogenet., Vol. 24, p.1, 1987) to allow the proportion of normal and leukemic clonogenic cells, and hence LTC-IC, to be determined.

E~ample 4 CNL LTC-IC ~SSAYS
The validity of using the clonogenic cell producing property of LTC-IC as an endpoint for their quantitation depends on the existence of a linear relationship between the number of LTC-IC seeded into the cultures and the number of clonogenic cells present 5 weeks later regardless of the input LTC-IC concentration. This was previously demonstrated for assays of LTC-IC in normal marrow cell suspensions (Sutherland et al., Proc. Natl.
Acad. Sci., Vol. 87, p. 3584, 1990).

Previous studies had shown that LTC initiated with peripheral blood cells from CML patients with high WBC
counts (and a marked elevation in circulating Ph1-positive progenitors, (Eaves et al., Bailliere's Clinical Haematology Vol. 4, pg. 931, 1990)) when analyzed 4-8 weeks later contained high numbers of exclusively Ph1-positive clonogenic progenitors (Eaves et al., Proc. Natl.
Acad. Sci., Vol. 83, p. 5306, 1986) in contrast to LTC
initiated with CML BM (Coulombel et al., N. Engl. J. Med., .
, . - . .
. ". ~.
-`, .
.~ .
.

20802~

Vol. 308, p. 1493, 1983). This suggested that it might be possible to use CML peripheral blood from such patients as an enriched source of primitive leukemic cells to investigate the relationship between cell input and S leukemic clonogenic cell output 5 weeks later.

In a series of experiments using cell samples, prepared from the peripheral blood from 8 different CML patients with high WBC counts, in which the number of light density peripheral blood input cells was varied from 5 x 103 to a maximum of 107 cells per 2.5 ml LTC (in 35 mm tissue culture dishes), the slope of the line relating the logarithm of the inoculum size (total nucleated cells) to the logarithm of the number of clonogenic cells detected : after these LTC had been maintained for 5 weeks was l.05 lS + 0.21 (which is not significantly different from l.0, p~0.5).

The results for a representative patient are illustrated in Figure l, and show that the conditions described for the LTC-IC assay in Example 2 are suitable for the detection and quantitation of Ph1-positive LTC-IC. Figure l demonstrates the linear relationship between the number of light density CML peripheral blood cells seeded into individual LTC (containing irradiated pre-established normal marrow feeders) and the number of clonogenic cells detected in secondary methylcellulose assays of cells harvested from these primary LTC 5 weeks after their initiation. Each point represents a single LTC. All points are derived from a single representative experiment using cells from a CML patient with a WBC count of l90 x : 30 109/L. The slope of the regression line fitted to this data set is 0.81+0.13.

., Example S
: Measurements of leukemic LTC-IC in CML blood and marro~.
Because the relative output of clonogenic cells from :: ~ . . : : . . . :
. . . ~ : . .
. . . :
- . - : -- . ,:

20802~

circulatinq leukemic LTC-IC was found to be constant under the assay conditions used even down to limiting numbers of input cells, as shown, for example, in Figure 1, quantitation of absolute leukemic LTC-IC numbers by limiting dilution analysis was possible. For such experiments, blood from CNL patients with elevated WBC
counts was again used as a highly and selectively enriched source of leukemic LTC-IC. Irradiated normal marrow derived feeders were subcultured into 96 well flat bottom Nunclon plates (Sutherland et al., Proc. Natl. Acad. Sci., Vol. 87, p. 3584, 1990) and then from 50 to 2 x 105 light density cells added per well in volumes of 1~0 ~1 with 23 + 1 wells per group. Five weeks later, all of the cells in each well were suspended and plated in methylcellulose assay cultures to enable detection of one or more clonogenic cells per well. From the proportion of positive and negative LTC defined in this way, the frequency of LTC-IC in 6 different input samples was calculated using Poisson statistics (Taswell, C., J.
Immunol. Vol. 126, p. 1614, 1981 - Coller et al., Nethods in Enzymology, Vol. 121, p. 412, 1986).

Results for a representative experiment are shown in Figure 2. Figure 2 shows limiting dilution analysis of data from a representative experiment in which decreasing numbers of light densi~y CML peripheral blood cells (from a patient with a WBC count of 21 x 109/L) were seeded onto irradiated marrow feeders and the cultures then assayed 5 weeks later for the presence (positive cultures) or absence (negative cultures) of 21 clonogenic cell. In this experiment, the frequency of LTC-IC in the suspension assayed (i.e., the reciprocal of the concentration of cells that gave 37~ negative cultures) was 1 per 7.6 x 104 cells (95% confidence limits = 1 per 5.3 x 104 - 1 per 11.0 X 104).

The average 5 week output of clonogenic cells per Ph1-. . ; -.
.

-208025~

positive LTC-IC was then derived in each case. The results of this latter calculation are shown in Table 1 together with results obtained when the same procedure and assay conditions were used to analyze LTC-IC in normal marrow or blood. The proliferative potential o~ all thase types of LTC-IC as assessed by this 5 week clonogenic cell output endpoint can be seen to be both similar and relatively constant, providing further support for the use of the LTC-IC assay to quantitate and charac~erize a very primitive Ph1-positive cell type.

Knowledge of the 5 week clonogenic cell output per leukemic CML allows absolute values to be derived from total 5 week clonogenic cell yields measured in cultures initiated with non-limiting inocula, which are experimentally easier to perform than limiting dilution analyses. LTC-IC values were thus obtained for peripheral blood samples from an additional 20 CML patients, and the concentration of LTC-IC per ml of blood then calculated assuming 100% LTC-IC recovery in the light density fraction assayed (Sutherland et al., Blood, Yol. 74, p.
1563, 1989). The results are shown in Figure 3 together with circulating LTC-IC values obtained from similar measurements of T-depleted, light density peripheral blood samples from a large series of normal individuals.

Figure 3 shows the LTC-IC concentration (per ml) in the peripheral blood of different CML patients (solid circles) as compared to 23 normal individuals (the open circle in each panel shows the mean + SEM of 2.9 + 0.5 LTC-IC per ml measured in these individuals) as a function of the WBC
count (per ml) (Panel A), or the peripheral blood clonogenic progenitor tBFU-E + CFU-GM + CFU-GENM) content per ml (Panel B). Absolute LTC-IC values were obtained either directly by limiting dilution analysis, or indirectly from the total clonogenic cell output measured at week 5 divided by the average number of clonogenic ~,: . . - ~ : -: . .~ : . . -2~802~5 cells produced per LTC-IC; i.e., 3 and 4 for CML and normal LTC-IC, respectively, as described above. A
significant association ~etween the two parameters measured in Panel B is indicated by a Spearmann's rank correlation coefficient rS=0.77 (p~0.000l).

In Figure 3A, LTC-IC concentrations in CML blood are plotted as a function of the W~C count. It can be seen that LTC-IC numbers increase exponentially such that values >105-fold higher than normal circulating LTC~IC
levels are seen in patients with the largest tumor burdens. In Figure 3B, the number of circulating LTC-IC
in individual CML patients is plotted as a function of the number of circulating clonogenic cells ( BFU-E plus CFU-GM
plus CFU-GEMM per ml) in the same patient. On average, leukemic LTC-IC were found to circulate at a l0-fold lower frequency than clonogenic cells although these two parameters showed a highly significant association (Spearman's rank correlation coefficient, rS=0.77, p<0.000l, n=26). By comparison the ratio of circulating LTC-IC to clonogenic cells in normal blood appeared much lower (~1:80).

LTC-IC assays were also performed using CML marrow samples. However, in each of these experiments, cytogenetic analyses were performed on the colonies produced from the clonogenic progenitors present after 5 weeks in LTC to distinguish Ph1-positive and Ph1-negative LTC-IC as, in contrast to CNL blood, Ph1-positive LTC-IC
would be anticipated to frequently represent a minority population relative to normal LTC-IC in CML marrow. The concentration of Ph1-positive LTC-IC (relative to other nucleated cells) in the 12 CML marrows analyzed was quite variable and in general markedly reduced, both by comparison to LTC-IC values in control marrows (i.e. '2.8 + l.4 Ph1-positive LTC-IC per 106 CML marrow cells as compared to 5S + 12 LTC-IC per 106 marrow cells ~rom normal ' 20~0255 individuals, n=13), and by comparison to normal (Ph1-negative) LTC-IC co-existin~ in the same CML marrows tested (for which a Yalue of 5.4 + 1.2 per I06 cells was obtained).

In summary, Ph1-positive LTC-IC were found to produce on average, a similar number of clonogenic cell progeny after 5 weeks in LTC as do their normal counterparts in the blood or marrow of normal individuals. However, a number of abnormalities in the CML LTC-IC population were also revealed. First, their distribution between marrow and blood was shown to be grossly altered, even more dramatically than is the case for Ph~-positive clonogentic cells. Both populations increase exponentially in the blood with linear increases in the WBC count, but Ph1-positive LTC-IC appear to be present at relatively reduced frequencies in CNL marrow whereas Ph1-positive clonogenic cell frequencies in CML marrow are relatively normal (Eaves et al., Exp. Hematol., Vol. 8, p. 235, 1980).
Second, in spite of a normal output of clonogentic cell progeny by Ph1-positive LTC-IC and the provision of a pre-established feeder derived from a normal marrow donor, their initial maintenance in the LTC system was highly comprised relative to normal LTC-IC. Whether this is due to an intrinsic defect in the Ph1-positive LTC-IC that is not subject to extrinsic modulation and/or whether such differences may also prevail in vivo have yet to be determined. ~owever, it is interesting to speculate that the behaviour of normal and leukemic LTC-IC in the LTC may indicate how these cells behave in vivo under analogous conditions of stimulation. One might then expect to see evidence of a growth advantage of the stem cells in the Ph1-positive clone in vivo only when most co-existing normal stem cells were in a quiescent state. The latter might be anticipated to occur in chronic phase CML
patients managed with conventional therapy, but a situation more closely resembling that obtained in LTC

.
,, ~ . .
- : ,, ~ :

might occur in vivo, albeit transiently/ following more intensive treatment. It is interesting to note that clinical experience fits well with these predictions (Goto et al., Blood, Vol. 59, p. 793, 1982; Kantar~ian et al., J. Clin. Oncol., Vol. 3, p. 192, 1~85).

Example 6 Diferential maintenance of normal and leu~emic LTC-IC in culture Previous studies have found that normal marrow LTC-IC are well maintained in LTC established from a single input innoculum (Eaves et al., Effects of Therapy on ~iology and Kinetics of Residual Tumor, Part A: Pre-Clinical Aspects, p. 223, 1990; Eaves et al., Ann. N.Y. Acad. Sci. Vol. 628, p. 298, 1991) and similar kinetics are seen when highly purified LTC-IC from normal marrow are seeded onto preestablished feeders (Sutherland et al., Proc. Natl.
Acad. Sci., Vol. 87, p. 3584, 1990). Figure 4 shows the corresponding results obtained when light density peripheral blood cells from CML patients with high WBC
counts were seeded onto irradiated human marrow feeders and the number of LTC-IC were determined by harvesting these primary LTC and performing secondary LTC IC assays as described above.

In particular, Figure 4 shows the differential kinetics of CML (solid symbols) versus normal (open symbols) LTC-IC in LTC initiated from cells seeded onto irradiated normal marrow feeders. Values shown are means -~ SEM after normalization of data in individual experiments by setting LTC-IC values in the primary inoculum in each experiment to 100%; n=6 for CML (peripheral blood LTC-IC), n=5 for LTC-IC in normal blood (open circles) and n=2 for LTC-IC
in normal marrow topen triangles). Open squares show previously published data for LTC-IC in normal unseparated marrow cultured in the absence of pre-established feeders -(Eaves et al., Ann N.Y. Acad. Sci., Vol. 628, p. 298, : :
.
., 208025~

1991) .

For comparison, analogous experiments were performed for primary LTC established by seeding light density, T-depleted normal peripheral blood or normal marrow buffy coat cells onto pre-established marrow feeders. Normal LTC-IC maintenance in such cultures was the ~ame regardless of the source of LTC-IC with no decrease in overall population size during the first 10 days. In contrast, the leukemic LTC-IC population showed an immediate and rapid rate of decline down to ~3~ of input values within the same initial period during which time the cultures had not been manipulated in any way except to reduce the temperature from 37C to 33C.

The following materials and methods were used in the studies outlined in Examples 7 to 9:

Cells Heparinized blood samples were obtained with informed consent from CNL patients undergoing routine hematologic assessment. All patients were Ph1-positive and in chronic phase. As shown in Table 2, the number of circulating LTC-IC in all patients studied was abnormally elevated (by a factor of from >400-fold to ~ 105-fold above the average normal value of ~ 2.9 + 0.5 per mL. The light density fraction (<1.077 g/cm3) was isolated by centrifugation of the blood on Ficoll/Hypa~ue (FH) to eliminate the ma~ority of erythrocytes, granulocytes and platelets and to obtain a preliminary enrichment of progenitor cells.

Normal blood samples were obtained with informed consent from normal individuals undergoing platelet/leukapheresis and from these a light density T cell-depleted fraction was then isolated by rosetting with sheep erythrocytes and centrifugation on FH as generally described in Marsden et al., J. Immunol. Methods., Vol. 33, p. 323, 1980). The ~ - . ; ~ . . , - . . - : .

~ , , ' ~ ': , ' : ~: ~ :, . ,:. ,:

208~25~

number of remaining CD2-t (T cells) detected by FACScan analysis of this T cell-depleted, light density fraction of normal blood cells represented, on average, <2~ of the total. Since the number of T cells in initial CML blood samples was already at or a below this level, the T-cell depletion step was not performed on CML blood samples.
Heparinized normal marrow aspirate cells were leftovers obtained with informed consent from allogenic donors providing marrow for transplantation.

: 10 Staining and Flow Cytometry Cells were washed twice and resuspended in Hank's solution with 2% fetal calf serum (FCS) and 0.1% sodium azide (NaN3)(HFN). Cells were first stained with an anti-CD34 antibody (8G12) directly conjugated to phycoerythrin (PE) 3Or fluorescein isothiocyanate (FITC), following the methods described in Lansdorp et al., J. Exp. Ned., Vol.
172, p.363, 1990, then double-stained with 0.1 ~g/ml Rhodamine-123 (RH-123) (Udomsakdi et al., Exp. Hematol, Vol. 19, p. 338, 1991) or 1-2 ~g of anti-HLA-DR-PE (107 cells/ml) (Sutherland et al., Vol. 74, p. 1563, 1989).
Stained cells were analyzed and sorted using a Becton Dickinson FACStarPlUs fluorescence-activated cell sorter (FACS) equipped with an argon laser emitting at 488 nm.
Fluorescence of Rh-123, FITC and PE-labelled cells was measured using 530/30 and 575/26 band pass filters, respectively, after calibration of the FACS prior to each sort with 10 ~m fluorescent beads. In some experiments, gates were set to exclude most of the granulocytes and erythrocytes using previously described forward light scatter (FSC) and side scatter (SSC) characteristics.
(Udomsakdi et al., Exp. Hematol, Vol. 19, p. 338, 1991) Cells appearing within this light scatter window constituted on average 15-20% of the total light density fraction of CML blood cells, as shown in Figure 5A.
Figure 5 shows distribution according to light scatter characteristics of total cells (representative sample -.:

. ~ . . .
, , . - . , .
:' , . , ' .

Panel A), and clonogenic cells (open bar~) and LTC-IC
(solid bars) (combined results for patients - Panel B) in the light density fraction of CML blood. Results for fractions I, II and III shown in Panel B are as defined in Panel A. Error bars in Panel B indicate the mean + lSEM
of values obtained on each of 5 patients studied individually. Sorted cells were collected in Hank~s ~olution with 50% FCS and were maintained at 4C until plated.

Functional ~ssay~
Cells from primary blood samples or from LTC harvests were assayed for clonogenic erythropoietic (BFU-E), granulopoietic (CFU-GM), and multilineage (CFU-GEMM) progenitors in standard methylcellulose cultures containing 3 units per ml of human erythropoietin and 10%
agar-stimulated human peripheral leukocyte conditioned medium (Terry Fox Laboratory Media Preparation Service, Vancouver, BC). The methodology for colony generation and criteria for colony recognition were generally as descried in Coulombel et al., Blood, Vol. 62, p. 291, 1983. Total clonogenic cell numbers refers to the sum of BFU-E, CFU-GM, CFU-GEMM detected in direct assays using these procedures. LTC-IC assays were performed by seeding an aliquot of the test cell suspension into cultures containing a feeder layer of irradiated (1500 cGy) normal (allogenic) marrow cells (3 X 104 cells per cm3). These were subcultured from the adherent layer of previously established 2-4 week old LTC. (Eaves et al., Proc. Natl.
Acad. Sci. USA, Vol. 83, p. 5306. 1986, Sutherland et al., Proc. Natl. Acad. Sci. USA, Vol. 87, p. 3584, 1990). LTC
were initially maintained for 3-5 days at 37C, then ; switched to 33C thereafter. They were then fed weekly by replacement of half of the growth medium (an enriched ~-medium containing 12.5% horse serum, 12.5% fetal calf serum, 10-4 M 2-mercaptoethanol and 10-6 M hydrocortisone) containing half of the nonadherent cells, with fresh .

.
. .. . .
.

2~802~5 growth medium.

After a total of 5 weeks, or as specified, the nonadherent cells were removed, washed, and combined with cells from those harvested from the trypsinized adherent layer.
These harvested LTC cells were then assayed for clonogenic cells in standard methylcellulose cultures at an appropriate concentration (usually 5 x 104 or 105 cells per l.1 ml assay). The total number of clonogenic cells (i.e.
BFU-E plus CFU-GN plus CFU-&EMM) present in 5 week-old LTC
provides a relative measure of the number of LTC-IC
originally present in the test suspension, as described above. In some cases absolute LTC-IC values were calculated by dividing this number by 3, which is the average output of clonogenic cells per leukemic LTC-IC as shown above by limiting dilution analysis.

~xample 7 Phenotype Analysis of CML LTC-IC
To determine the light scattering properties of circulating clonogenic cells and LTC-IC in patients with CML, light density blood cells were sorted into three fractions (as illustrated in Figure 5A) and the results compared with data for normal BN (Sutherland et al., Vol.
74, p. 1563, 1989) and blood progenitors (Figure 5B).
Most of the nucleated cells (~ 85% in the light density fraction of CML blood had a high SSC (Fraction III) in contrast to the light density cells in normal blood where the proportion of such cells is much lower (<40~, data not shown). The mean number of clonogenic cells and LTC-IC
recovered in each sorted fraction was determined and expressed as a percentage of the total number of progenitors present in the starting (light density) cell suspension of each sample studied. As shown in Figure 5B, it can be seen that the majority of both the clonogenic cells and LTC-IC in CML blood were consistently found in fraction II (i.e. cells with high FSC but low SSC). Cells :

.:
,' : . - :::: .. . .
:: -20~2~

from this fraction also generally produced more nucleated cells (as well as clonogenic cells) after 5 weeks in LTC
(both in the adherent and nonadherent layer) than other fractions on a per cell basis. However, a significant proportion of the circulating CML clonogenic cells (~ 15%) and LTC-IC (~ 30~) were detected in a population characterized by low FSC and low SSC (fraction I). Some circulating C~L clonogenic cells (~ 5%) were found amongst the cells with a high SSC (fraction III). These findings suggest subtle differences between circulating C~L
clonogenic cells and LTC-IC in terms of their overall light scattering properties.

This was reinforced by additional experiments in which fraction II wa~ subdivided further into 2-3 additional fractions. Analysis of these showed that the circulating CML clonogenic cells were more concentrated in fractions containing cells with a slightly higher FSC by comparison to the distribution of LTC-IC in the same fractions. The high FSC of circulating clonogenic cells in CML patients differs markedly from the FSC typical of clonogenic cells in the circulation of normal individuals, but is very similar to the majority of clonogenic cells in normal BM.
(Sutherland et al., Vol. 74, p. 1563, 1989) Since very few progenitors were present in fraction III, only cells ~5 in fractions I and II were analyzed in all subsequent sorts.

Figure 6 shows bivariate contour plots of a single representative sample of normal (Panels A, C & E) and CML
(Panel B, D & F) light density blood cells in the low SSC
window (fractions I and II in Figure lA). CD34+ cells (gated as shown by the vertical lines in Panels C and D, or the horizontal lines in Panels E and F) were subdivided into CD34+DRLW and CD34+DRhi~h subpopulations as shown by the horizontal lines in Panel C, or CD34+DRlW, CD34'DR+ and CD34+DR++ subpopulations as shown by the two horizontal '` `.: ' : ' .
: .

: ~ :

208025~

lines in Panel D, or CD34~h-123dU~l and CD34+Rh-123bri~ht populations as shown by the vertical lines in Panels ~ and F. Unstained cells are shown in Panels A and B.

Figure 6 shows representative distributions of light density normal and CML blood cells gated for low SSC after two colour staining for expression of CD34 and HLA-DR, or expression of CD34 and uptake of Rh-123. A much larger proportion of light density CML blood cells were found to express readily detectable levels of CD34 than is the case for normal blood cells in the same light scatter window (compare Panels D and F with C and E in Figure 6). The CML cells also contained a higher proportion of cells that expressed readily detectable levels of HLA-DR or that retained Rh-123 by comparison to normal blood. Figures 7 and 8 show the results obtained when the CD34+, SSC~W
cells were sorted according to their expression of HLA-DR
(Figure 7) or Rh-123 uptake (Figure 8) and then analyzed functionally for clonogenic cells or LTC-IC content.

Figure 7 shows the distribution of clonogenic cells (Panel A) and LTC-IC (Panel B) within the CD34' fraction of circulating CML cells subdivided (as shown by the lower horizontal line in Figure 6D) according to their high or low expression of HLA-DR ( solid bars). The mean progenitor recovery + lSEN is expressed as a percent of the total number light density progenitors recovered within the low SSC fraction shown in Figure 5A from studies of 4 different pati.ents. For comparison, previously obtained analogous results for normal BM (open bars, n=6) (Sutherland et 21., Vol. 74, p. 1563, 1989) and normal blood (stippled bars, n=3) progenitors are also included in this figure.

Figure 8 shows the distribution of clonogenic cells (Panels A) and LTC-IC (Panel B) within the CD34' fraction o~ circulating CML cells subdivided (as shown in Figure . . ~ .~ , .............. .
: . .:, ~
.
: . : . . .

20802~

6F) according to their uptake of Rh-123 (solid bars). The mean progenitor recovery + lSEM is expressed as a percent of the total number of light density progenitors recovered within the low SSC fraction shown in Figure 5A from Studies of 4 different patients. For comparison previously obtained analogous results for normal AN (open bars, n=6 (~6) and normal blood (stippled bars, n=3) progenitors are also included.

It can be seen that most of ~he clonogenic cells in CML
blood, like most of the clonogenic cells in normal marrow, expressed readily detectable levels of HLA-DR (Figure 7A) and showed positive staining with Rh-123 (Figure 8A). In this respect, however, they both differ markedly from the clonogenic cells found in normal blood, of which very few show a DRhi9h or Rh-123bri9ht phenotype. Further subdivision of the CD34+ DRhi9h fraction of CML blood cells into DR+ and DR++ subpopulations, as defined in Figure 6D, revealed the presence of clonogenic cells in both (Tables 3 and 4).
Interestingly, a proportion (~10%) of the clonogenic progenitors were also found in the DRlW or Rh-123 dUl~
; subpopulations of CD34+ CML blood cells. Although none of these were specifically genotyped, it is unlikely that significant numbers in either of these latter phenotypically defined "normal" subgroups were residual normal progenitors since normal progenitors, even if present at normal levels, would have accounted for <10% of the progenitors in the DRLW (Table 4) or Rh-123 fractions of all patients studied.

When the sorted CML cells were assayed for LTC-IC, the majority (~75%) were also present amongst the CD34+ DRhi9h cells (Figure 7B). This is also in contrast to normal LTC-IC, the majority of which in either blood (~ 100%) or BM ~ 55%) express little or no HLA-DR. Thus isolation of CD34+ DRhi9h populations of cells from the peripheral blood of CNL patients (either DR+ or DR++) yields a highly . ~ .
.
, ' ' , ,~ ,.

enriched LTC-IC population (Table 3). As noted for the circulating clonogenic cells in the same CML blood samples, a proportion of the LTC-IC (in this ca~e, ~ 30%3, was also found in the CD34+DRlW fraction. ~ecause of the marked elevation in total LTC-IC numbers in these samples, the number of CD34~DRlw LTC-IC was also consi~tently greatly ()500x) in excess of values for CD34+DRlW LTC-IC in the normal circulation (see Table 4~. Similarly, most of the LTC-IC in the CML blood samples were Rh-123bri9ht (Figure 8B) in contrast to the LTC-IC in either the blood or marrow of normal individuals. However, on average, ~ 20%
of the circulating LTC-IC in patients with CNL were found to have a Rh-123dUlL phenotype, of which (1% would have been anticipated to be residual normal LTC-IC even if these were still present at normal levels.

It can be seen from Table 2 that the initial frequencies of the clonogenic cells and LTC-IC in the CML blood samples studied, although elevated, were quite variable both on a volume and on a per nucleated cell basis.
Variability was also encountered after these progenitors were separated into various subpopulations as shown in Table 3 for light density, CD34 ', DR~W or DRhi~h cells.
However, on average, the purity of circulating CML LTC-IC
in the CD34+DRLW and CD34+DR+ fractions was approximately 10% (Table 2). Corresponding values for the frequency of clonogenic cells in the CD34+DRLW and CD34+DR+ fractions were 10~ and 20% (Table 2). As for normal blood, recovery of LTC-IC in the light density, SSCLW, CD3~+ fraction of CML blood was high (129~) and of clonogenic cells was lower (73~) suggesting exclusion of some CML clonogenic cells with the gating criteria used.

Example 8 Sensitivity of ~MT. progenitors to 4-hydroperoxycyclo-phosphamide (4-HC) The present inventors show that LTC-IC in normal blood, . .

.

2~2~

- 47 _ like LTC-IC in normal BM, are relatively resistant to 4-HC, as are circulating clonogenic cells, whereas clonogenic cells in normal BM are more 4-HC-sensitive.
Recent clinical findings indicate that reconstitution of hematopoiesis with Ph1-negative cells can be achieved in some CML patients receiving 4-HC-treated autologous BM
transplants. (Carlo-Stella et al., Bone Narrow Transplant, Vol. 8, p. 265, 1991) This suggests that transplantable Ph1-positive stem cells may be more sensitive to 4-HC than normal stem cells. The present inventors have now evaluated the 4-HC sensitivity of circulating CML
clonogenic cells and LTC-IC and compared these to normal clonogenic cells and normal LTC-IC. In this series of experiments, LTC-IC function was assessed in terms of the clonogenic cell content of LTC evaluated after 4 and 8 weeks (rather than after 5 weeks, as in the studies described above), since previous experiments, had revealed differences in the 4-HC sensitivity of normal LTC-IC
measured by these two different endpoints. (Winton et al., Exp. Hematol, Vol. 15, p. 710, 1987) Results for light density CNL blood cells exposed to 100%
~g/ml of 4-HC under standard transplant exposure conditions (i.e., 2 x 107 cells/ml with 7% red cells for 30 minutes at 37C) are shown in Figure 9, together with previous data for normal progenitors tested using the same procedures and reagents.

Figure 9 shows the survival of circulating CML clonogenic cells (Day O) and LTC-IC (4 and 8 week clonogenic cell output endpoints) after a brief exposure to 100 ~g/ml 4-HC
(30 minutes at 37C in the presence of 7~ red blood cells, cells at 2 X 107 cells/ml). Results for circulating CML
progenitors (solid bars showing mean + lSEM for 4 different patients) are shown for comparison together with previously obtained results for normal BM (open bars, n=6) and normal blood (stippled bars, n=3) progenitors treated - '. ' '' ' ~ ' . ~ .

208~2~

using the same conditions.

Circulating CML clonogenic cells and clonogenic cellæ in normal marrow were simply reduced (to ~ 1~% of initial numbers) by this treatment. LTC-IC in the same CNL blood samples appeared only slightly more resistant and were significantly more sensitive (p~0.01) than normal LTC-IC
from any source.

Example 9 Differentiative potential of ~,MT. LTC-IC
Previous studies have shown that the relative numbers of different types of clonogenic progenitors present in 5 week-old LTC provides a consistent average overall measure of the differentiative behaviour of LTC-IC assayed under standard LTC conditions. (Sutherland et al., Blood Vol.
74, p. 1563, 1989) To assess whether this parameter is altered in the LTC-IC present in CML blood, the ratio of BFU-E, CFU-GM and CFU-GEMM numbers before and after LTC of light density CNL blood cells was assessed. As shown in Table 4, after 5 weeks in LTC the proportion of progenitors identified as CFU-GM increased as documented previously for LTC-IC in the blood and marrow of normal individuals, - The materials and methods used in the studies outlined in ~xamples 10 to 15 are detailed below:

Bone marrow cells. Aliquots of normal human marrow cells were obtained from informed and consenting allogeneic bone marrow transplant donors at the time of marrow harvests and with approval of the Clinical Screening Committee for Research Involving Human Subjects of the University of British Columbia (Vancouver, Canada). Percolled low-density cells (~1.068 g/mL) were stained with anti-CD34 antibody directly conjugated to fluorescein isothiocyanate .~,. , ~ : ' - , . . .

"

2~802~5 (8&12-FITC) (Lansdorp PM et al. J. Exp. Med. 172:363, 1990), and HLA-DR directly conjugated to phycoerythrin (HLA-DR-PE; Becton Dickinson, Mountain View, CA), and then ~orted on a FACStarP~Us (BD FACS Sy~tems; Becton Dickinson) (Sutherland HJ, et al. Proc. Natl. Acad. Sci. USA 87:384, 1990). Cells were sorted within low to intermediate forward light scatter and low 90 light scatter gates to include cells with properties similar to small lymphocytes. Cells were additionally sorted for high CD34 expression and very low or negative HLA-DR expression.
This sorting allowed isolation of a subpopulation representing ~0.4% of total bone marrow cells that was enriched ~400-fold in cells that produce clonogenic cells detected after 5 weeks in LTC (Sutherland HJ, et al. Blood 74:1563, 1989; Sutherland HJ et al. Proc. Natl. Acad.
Sci. USA 87:3584, 1990).

Cell line~. ~2 (Mann R, et al. Cell 33:153, 1983) and NIH-3T3 cell lines were cultured in Dulbecco's modified Eagle'~ medium (DMEM) with high glucose (4.5g/L) and 10%
heat-inactivated calf serum (for ~2 cells) or 10% fetal calf serum (FCS). M2-lOB4 cells, a cloned murine (B6C3F1) marrow fibroblast cell line (Lemoine FM, et al. Exp.
Hematol 16:718, 1988), were maintained in RPMI medium plus 10% FCS. AML-193 cells (Santoli C., et al. ~. Immunol 139:3348, 1987) (American Type Culture Collection [ATCC], Rockville, MD) were growth in Iscove's medium with 20% FCS
and 10% medium conditioned by the 5637 cell line (ATCC).
NFS-60 cells (Weinstein Y. et al. Proc. Natl. Acad. Sci.
USA 83:5010, 1986), obtained from Dr. J. Ihle (National Cancer Institute-Frederick Cancer Research Facility, Frederick, MD), were grown in RPMI with 20~ FCS and 5 pokeweed-stimulated mouse spleen cell conditioned medium.
B9 cells were grown in DMEN with 10% FCS and 100 U/mL IL-6 (Lansdorp PM et al. in Potter M, Melchers F (eds): Current Topics in Microbiology and Immunology. New York, NY, Springer-Veriag, 1986, p. lQ5).

- - . . . .
. , -~' ' " ~ ' , - ~ :

20802~

Retroviral vectors and viral producer cell line~.
The retroviral vectors for human GM-CSF and ~-CSF have been described in Hogge ~E, et al. Blood 77:493, 1991).
Similar principles were used to construct a vector containing a human IL-3 cDNA. Briefly, an IL-3 cDNA ( from Genetics Institute, Boston, MA), was truncated at the 3' end to remove A-T rich sequences thought to be responsible for destabilizing mRNA transcripts (Shaw G., Kamen R. Cell 46:659, 1985). This 632-bp cDNA frag~ent was linked to a 250-bp PvuII-BglII fragment from the pXl vector containing the promoter from the herpes simplex thymidine kinase (tk) gene (Anderson WF. et al. Proc. Natl. Acad. Sci. USA
77:5399, 1980) and the tk-IL-3 cassette inserted into the Xho 1 site in the N2 retroviral vector, which is 3' of the neor gene (Eglitis MA, et al. Science 230:1395, 1985).
Retroviral constructs were transfected into the c2 ecotropic packaging cell line (Mann R., et al. Cell 33:153, 1983). Individual clones of G418 resistant (G418r) transfected cells were isolated, expanded, and assessed both for viral titer by the ability of their growth medium to generate G418r NIH-3T3 cells, and for the production of growth factor bioactivity on growth factor-responsive cell lines. Clones producing viral titers greater than 105 colony-forming units/mL were used to infect N2-lOB4 cells.

Stromal feeders. Irradiated (15 Gy of 250-kV peak x-rays) normal human marrow adherent layer feeders (NF) were prepared as previously described (Sutherland HJ, et al.
Blood 74:1563, 1989; Sutherland JH, et al.Proc. Natl.
Acad. Sci. USA 87:3584, 1990). To generate human growth factor-producing M2-lOB4 feeders, cell-free growth medium was harvested from viral producer cells and, together with 8~g/mL polybrene, added to subconfluent cultures of N210-B4 cells. After 4 hours of incubation at 37C the virus-containing medium was replaced with standard growth medium. Forty-eight hours post-infection the cells were typsinized, replated in growth medium containing 0.4 mg/mL

...... . . . .
., . - ~ :

..

20802~

G418, and the cells grown to confluence, at which time the growth medium was tested for growth factor bioactivity.
Mass cultures of these retrovirally infected, G418r, growth factor-producing M2-lOB4 cells were subsequently mzintained and passaged as continuous cell line~. Using standard techni~ues and hybridization of blots to 3ZP-oligolabeled GM-CSF, G-CSF, or IL-3 cDNA probes, Southern and Northern analysis (Feinberg AP, Vogel~tein B Anal.
Biochem. 132:6, 1983; Sambrook J., et al.: Nolecular cloning: A laboratory manual. Cold Spring Harbor, NY, Cold Spring Harbor Laboratory, 1989) showed grossly intact proviral DNA and the expected full-length and spliced retroviral transcripts in the infected M2-lOB4 cells. N2-lOB4 feeders were prepared before the initiation of cocultures by seeding 3 x 105 M2-lOB4 cells into 35-mm corning tissue culture dishes (Corning Glassworks, Corning NY) or into Nunc 96-well plates (A/S Nunc, Roskilde, Denmark) at 104 cells per well. In cultures containing cells from more than one growth factor-producing cell line (to test the effect of specific combinations of growth factors), equal numbers of each of the types of cells were used, keeping the total cells plated constant at the values given above. All M2-lOB4 feeders were irradiated with 80 Gy of x-rays.

Coculture~. In a total of 16 experiments, 800 to 11,000 sorted human bone marrow cells were placed in cultureR
with or without feeders (as indicated) in LTC medium.
Cultures were then maintained at 33C for 5 weeks with weekly half-medium changes as previously described (Sutherland HJ, et al.: Blood 74:1563, 1989; Sutherland JH, et al. Proc. Natl. Acad. Sci. USA 87:3584, 1990). At the end of 5 weeks, all nonadherent cells were removed and counted, and the adherent cells were then suspended by trypsinization. Alquots equal to 1/3 to 1/2 of total adherent and nonadherent cells were plated in standard methylcellulose cultures for assessment of total , erythropoietic (BFU-E), granulopoietic (CFU-GM), and multilineage (CFU-GEMM) progenitors detected 20 days after initia~ion (Cashman JD., e~ al. Blood 75:~6, 1990).
Aliquots were also reseeded on top of new irradiated normal human NF in 96-well plates for assessment of LTC-IC
content by limiting dilution analysis and measurement of total clonogenic cell content after an additional 5 weeks (Sutherland JH, et al. Proc. Natl. Acad. Sci. USA
87:3584, 1990).

Growth factor bioactivity. Growth factor bioactivity was measured in growth media collected 2 days after a complete change of the medium in confluent cultures of viral producer cells or feeders, and in media removed from cocultures at weekly intervals. Bioactivity was measured by comparing the stimulation of 3H-thymidine incorporation into appropriate growth factor-responsive cell lines to that obtained with recombinant growth factor standards.
Recombinant GN-CSF and IL-3 were gifts from Biogen (Geneva, Switzerland) and Behring (Frankfurt, Germany), and recombinant IL-6 was purchased from R & D Systems, Inc. (Minneapolis, MN). Recombinant G-CSF was purchased from Amersham (Oakville, Canada). GM-CSF and IL-3 levels were measured on human ANL 193 cells, G-CSF on NFS 60 cells, and IL-6 on B9 cells.

Example lO
Growth factor production by engineered M2-lOB4 cells.
Human growth factor-producing N2-lOB4 cells were generated by infection of the cells with ecotropic retrovirus capable of the transfer and expression of both the neor gene, which renders eukaryotic cells resistant to the neomycin analogue G418, and the cDNAs for either human GM-CSF, G-CSF, or IL-3. When retrovirally infected M2-lOB4 cells had grown to confluence under G418 selection, samples of their growth medium were tested for growth factor bioactivity (Table 13. Bioactivity was detected ':' : . - , ' ~ : ' ' . , , ' 20802~5 only from cells infected with the appropriate virus and the levels measured ranged from ~l to 20 ng/mL.
BioactiYity from cultures containing two or three types of growth factor secreting M2-lOB4 cells was twofold to threefold lower, consistent with the lower number of each type of cell in these cultures. These remained stable for at least 2 months in the absence of G418 selection, even after the cells were irradiated. Bioactivity at levels approximately e~ual to the levels from the feeders alone ; 10 was also detected in media removed from cocultures of M2-lOB4 cells with purified human marrow cells, and levels remained unchanged throughout the period of the experiments. In cocultures with uninfected M2-lOB4 feeders no bioactivity could be detected. Assays for IL-6 were also performed on media conditioned by M2-lOB4 cells and media removed weekly from cocultures. These assays showed IL-6 levels to be consistently less than 0.01 ng/mL. In previous experiments the concentration of purified recombinant growth factor required to stimulate half-maximal hematopoietic colony growth from nonadherent marrow cells placed in short-term methylcellulose assays has been shown to be 0.01 ng/mL for GM-CSF, 10 ng/mL for G-CSF, and l ng/mL for IL-3 (Hogge DE., et al. Blood 77:493, l991), suggesting that growth factor production by the retrovirally infected M2-lOB4 cells was sufficient to warrant testing these cells as feeders in LTC.

xample 11 Capacity of N2-lOB4 cells to support human hematopoiesis.

Total numbers of nonadherent cells, clonogenic cells, and LTC-IC in cocultures 5 weeks old were measured to assess the ability of control (uninfected and/or N2-infected) M2-lOB4 cells to support hematopoiesis at these three levels of hematopoietic cell development (Table 2). Results obtained in each case were compared with those obtained from cultures containing normal human MF or no feeders 2~8~255 (i.e., hematopoietic cells seeded directly onto plastic).
Despite the lack of detectable G-CSF, GM-CSF, IL-3 or IL-6 in cultures containing control M2-lOB4 cells, significantly support for all levels of hematopoiesis was evident by comparison to results for cultures without feeders. For LTC-IC maintenance and production of clonogenic cells, M2-lOB4 cells were almost as effective as normal human MF. However, human MF did offer a significant improvement over M2-lOB4 cells when effects on terminal cell numbers (nonadherent cell production) were assessed.

Example 12 Specific growth factor effects on terminal hematopoiesis.

Nonadherent cell numbers in cultures 5 weeks old containing growth factor-producing M2-lOB4 cells were compared in a paired t-test to cultures with control M2-lOB4 cells (Fig. 1) and to cultures with human NF. IL-3-producing M2-lOB4 cells alone were not different than control N2-lOB4 cells (P = .4~. However, all other types of growth factor-producing feeders, either alone or in combination, increased nonadherent cell output above that seen with control M2-lOB4 feeders (P ~ .005). GM-CSF-producing M2-lOB4 cells with or without other growth factor-producing M2-lOB4 cells were most effective in this regard. Alone, they supported the production of ~20 times more nonadherent cells than cultures containing control M2-lOB4 cells, and N4 times more nonadherent cells even than cultures containing human MF (P ~ .005). Although the combination of IL-3- and G-CSF-producing M2-lOB4 cells was less effective for the promotion of terminal cell amplification than any feeder producing GM-CSF, it was equivalent to human MF in this regard.

The maintenance of LTC-IC was found to be supported by control murine stromal cells as effectively as by standard ' , "' .~: . . :
.. . . ..
- ~ .- :~ . , :-: : : --,. .. .

20802~

human marrow adherent layers. The presence of G-CSF and interleukin-3-producing M2-lOB4 cells in combination was found to further enhance the maintenance and early differentiation of these cells without a decline in their proliferative potential as measured by the clonogenic output per LTC-IC. However, this effect was lost if GM-CSF-producing feeders were also present. On the other hand, in the presence of GM-CSF-producing feeders, the output of mature granulocytes and macrophages increased 20-fold.

Example 13 Specific growth factor effects on clonogenic cell output.

Production of clonosenic cells was analyzed in the same experiments by comparison of numbers of clonogenic cells in 5-week-old cocultures to contxols using a paired t-test (Fig. 2). G-CSF feeders alone and G-CSF plus IL-3-producing feeders provided more support than control M2-lOB4 feeders (P ~ .05) and G-CSF plus IL-3 feeders were twice as supportive as human NF, although this later differ~nce did not quite reach statistical significance (P
= .12). GN-CSF feeders either alone or in combination with G-CSF or IL-3 feeders resulted in clonogenic cell output values at 5 weeks that were close to those obtained in cultures with control M2-lOB4 cells. To distinguish whether the IL-3-plus G-CSF-producing feeder combination increases clonogenic cells by increasing the number of LTC-IC recruited to differentiate, or by increasing the proliferative ability displayed by the LTC-IC originally present, an additional series of experiments was undertaken. In these experiments, sorted normal bone marrow cells were seeded at a limiting dilution onto either IL-3-plus G.-CSF- producing M2-lOB4 cells or human MF, and the frequency and avexage clonogenic cell output by individual LTC-IC was then determined from a knowledge of the clonogenic content of wells measured 5 weeks later.

, '~ ' 208025~

The results from five such experiments suggest a slight but not statistically significant advantage for the IL-3-plus G-CSF-secreting M2-lOB4 cell~ for both parameters assessed, i.e., the proportion of initially seeded cells detected as LTC-IC was 1.4% and 1.1% and the averag~
number of clonogenic cells produced per LTC-IC detected was 5.5 and 4.6 for the IL-3 + G-CSF feeders and human NF, respectively.

Example 14 Specific growth factor effects on LTC-IC maintenance. By plating cells at limiting dilution on human NF, the absolute number of LTC-IC in a population can be quantitated (Sutherland JH, et al. Proc. Natl. Acad. Sci.
USA 87:3584, 1990). This analysis can be performed on cells removed at various time points from a culture to provide a measure of the ability of the conditions prevailing in the cultures to promote the maintenance and/or self-renewal of LTC-IC. Approximately 25% of the ` number of input LTC-IC were detected after 5 weeks in LTC
initiated by seeding sorted marrow onto human NF. (Mean + SEM LTC-IC per 1,000 sorted cells originally plated =
16.7 + 4.0 on day O, and = 4.3 + 1.0 at 5 weeks, in six experiments). The ability of control and growth factor-producing N2-lOB4 cells to maintain LTC-IC was similarly assessed by quantitating the number of LTC-IC remaining after 5 weeks in primary cultures containing various feeders. As shown in Fig. 3, the combination of IL-3 plus G-CSF feeders in the primary cultures allowed better maintenance of LTC-IC than control M2-lOB4 feeders (P <
.05) and was even somewhat better than human NF.GM-CSF-producing feeders alone, or together with G-CSF-producing feeders, provided less LTC-IC maintenance than human MF (P
.05) and any culture that contained GM-CSF-producing feeders appeared worse than control N2-lOB4 cells for LTC-IC maintenance (P = .14 to .18). The other feeder combinations tested provided support of LTC-IC maintenance '., ' '`, . :' :: ' 208025~

that did not differ significantly from that obtained with human MF or M2-lOB4 cells.

Example l~
Lack of any growth factor effect on the proliferative potential displayed by LTC-IC present after 5 weeks in culture.
In addition to determining the number of LTC-IC maintained under various coculture conditions, the proliferative potential of these cells, as indicated by the average number of clonogenic progenitors produced per LTC-IC
(CFU/LTC-IC) before and after culture, was measured by limiting dilution analysis. CFU/LTC-IC was the same for LTC-IC maintained on human MF for 5 weeks as for the LTC-IC in the original purified marrow sample (4.0 + 0.7 v 4.3 + 0.4). Moreover, despite the fact that the number of LTC-IC maintained in primary cocultures with various types of growth-factor producing M2-lOB4 cells varied from twofold higher to threefold lower than the number of LTC-IC maintained in cultures containing human MF, the proliferative potential of the Lrrc-Ic present after 5 weeks was not influenced by the type of feeder used in the primary culture (analysis of variance, P = .46) (Table 3~.

The materials and methods used in the studies outlined in Examples 16 to 19 are detailed below:

CELLS peripheral blood mononuclear cells were obtained with informed consent from normal volunteer donors as a byproduct of plateletphereses performed at the Vancouver General Hospital, Canada. Cells were further depleted of T cells by incubation with 2-aminoethylbromide isothiouronium-treated sheep blood cells for 30 minutes at 4C and subsequent isolation of the light density (1.077gm/cm3) fraction after centrifugation on Ficoll-.:
::

. ~, :

208025~

hypaque (FH) as described previously. (Marsden M. et al, J. Immunol Methods, 33:323, 1990.) Random checking of this procedure showed that less than 2~ of the recovered cells were CD2 positive (T cells) by FACScan analysis. Normal ~one marrow (BM) aspirate cells ere obtained with informed consent from normal donors of allogenic morrow for transplantation. BM cells were either used directly, or after lysis of contaminating red blood cells by brief exposure to ammonium chloride, (Turhan AG et al, N. Engl.
- 10 J. Med. 320:1655, 1989) or after centrifugation on FH as indicated.

CULTURES. Cells from primary blood or BM samples or from LTC were assayed for erythroid (BFU-E), granulopoietic (CFU-~M~, and multilineage (CFU-GEMM) colony-forming cells in standard methylcellulose cultures containing 3 units per ml of human erythropoietin and 10% agar-stimulated human peripheral leukocyte conditioned medium. This methodology and the criteria used for colony recognition have been described in detail in Cashman J. et al, Blood 66:1002, 1985. LTC-IC assays were initiated by seeding an aliquot of the test cell suspension into cultures containing irradiated (1500 c~y) allogenic marrow cells (3 X 104 per cm2) that had been subcultured from the adherent layer of previously established 2-4 week old LTC.
(Sutherland HJ et al, Blood 74:1563, 1989; Eaves CJ et al, J. Tissue Culture Methods 13:55, 1991.) LTC-IC assay cultures were then fed weekly by replacement of half of the growth medium containing half of the nonadherent cells with fresh growth medium (-medium supplemented with inositol, folic acid, glutamine, 10-4 M 2-mercaptoethanol, 10-6 M hydrocortisone sodium hemisuccinate, 12.5% horse serum and 12.5 fetal calf serum (FCS). In most experiments LTC-IC assays were performed in cultures set up in 2.5 ml. volumes in 35 mm tissue cultures dishes, although for the limiting dilution assays, smaller, appropriately scaled down (0.1 ml) cultures were used as ~:
-20802~S

described previously. (Sutherland HJ, Proc. Natl. Acad.
Sci. USA 87:3584, 1990.) After a total of 5 weeks (unless specified otherwise), the nonadherent cells were removed, washed and combined with cells harvested from the adherent fraction by trypsiniazation. (Lansdorp PM et al, J. Exp.
Med. 172:363, l990.) These cells were then adjusted to a concentration suitable for plating in methylcellulose assays (to yield <200 colonies per 1.1 ml assay culture.
For a detailed description of the LTC-IC assay procedure, see Eaves CJ et al, J. Tissue Culture Nethods 13:55, 1991.
In the experiments reported there (unless specified otherwise), the number of clonogenic cells present in LTC
harvested after 5 weeks (i.e. the number of BFU-E plus CFU-GM plus CFU-GENM present in both the nonadherent and adherent fractions at this time) was used to provide a quantitative, albeit relative, measure of the number of LTC-IC originally seeded into the LTC. However, this number of clonogenic cells can be directly converted to an absolute number of LTC-IC simply by dividing by 4, since this is the average number of clonogenic cells calculated to be present in 5 week-old cultures per initial LTC-IC
seeded.

STAINING AND FLOW CYTONETRY. Cells were prepared for staining by resuspension in Hank's solution containing 2~
FCS and 0.01~ sodium azide (HFN). They were then incubated with 1-2 ~g/ml of anti-HLA-DR-phycoerythrin (PE) (107 cells/ml) or RH-123 at a final concentration of 0.1 ~g/ml as described in Sutherland HJ et al, Blood 74:1563, 1989, and Udomsakdi C. et al, Exp. Hematol 19:338, 1991.
In some cases cells were stained with an anti-CD34 antibody (8Gl2)36 directed conjugated to PE or fluorescein isothiocyanate (FITC). Stained cells were scored using a Becton Dickinson FACStar Plus (FACS) equipped with an argon laser emitting at 488 nm. Fluorescence of RH-123, FITC-, and PE-labelled cells was measured using 530/30 and 575/?6 band pass filters, respecti~ely, after calibration : .:

:

2~802~5 - 6~ -of the FACS prior to each sort using 10 ~m fluorescent beads. In some experiments, cells were gated according to their forward light sca~ter characteristics (FSC) and side scatter characteristics (SSC) to exclude most erythrocytes and granulocytes, as described in Sutherland HJ et al, Blood 74:1563, 1989, and Udomsakdi C. et al, Exp. Hematol 19:338, 1991. Cells appearing in this light scatter window (see Figure 4A, fractions I and II) constituted >60~ of the total density fraction T cell depleted blood cells. Cells were collected after sorting in Hank's solution containing 50~ FCS and were maintained at 4C
until plated.

~xample 16 QUANTITATION OF LTC--IC IN NORMAL BLOOD.
In an series of experiments, the number of clonogenic cells present after 5 weeks in LTC initiated with T cell-depleted suspensions of normal peripheral blood mononuclear cells seeded onto pre-established, irradiated marrow adherent layers was found to be a linear function of the number of cells initially added over a 1000-fold range of input cell numbers. Results for a representative experiment are shown in Figure 13.

Figure 13 shows a linear relationship between the number of light density (~1.077g/cm3) T cell-depleted peripheral blood cells from a representative normal individual seeded onto pre-established, irradiated normal marrow feeders and the total number of clonogenic cells detected when these LTC were harvested and assayed in methylcellulose 5 weeks later. The slope of the regression line fitted to this data set is 0.92 + 0.09.

Three such dose response experiments also included a series of assay cultures (20-25 per point) which were seeded with limiting numbers of LTC-IC (i.e. ~1 LTC-IC per assay culture). From the proportion of positive and .. . ~ ; . ~ . . : .
- ... ' . . ~ . .

.
- ~ ' -:
, . . . ..... - ~ : . ...

- : -2~80255 negative assay cultures (containing >1 clonogenic cell each, or none, respectively,) absolute frequencies of LTC-IC in the original test cell suspension were calculated using Poisson statistics (Porter EH et al, Br. J. Cancer 17:583, 1963, and Taswell C., J. Immunol 126:1614, 1981) (Figure 14). Figure 14 shows the limited dilution analysis o data from a representative experiment in which decreasing numbers of light density T cell-depl~ted normal peripheral blood cells were seeded onto irradiated marrow feeders and the number of clonogenic cells detectable after 5 weeks was then determined. For this experiment, the frequency of LTC-IC in the starting cell suspension ~i.e., the reciprocal of the concentration of test cells that gave 37% negative cultures) was 1 per 1.5 x 105 cells (95% confidence limits = 1 per 9.9 x 104 - 1 per 2.2 x 105 cells). From this value and a knowledge of the total number of clonogenic cells produced by a large number of cells of the same input suspension, the avera~e 5 week output of clonogenic cells per LTC-IC in normal blood was calculated. This value was found to be ~.7 + 1.2 (see Example 5) which is similar to the value of 4.3 + 0.4 that we reported for LTC-IC in normal BM. (Sutherland HJ, Proc.
Natl. Acad. Sci. USA 87:3584, 1990.) Bulk measurements of the 5 week clonogenic cell content of assay cultures initiated with T cell- depleted blood samples from other normal adults could then also be used to derive absolute LTC-IC per ml values using this average clonogenic output per LTC-IC conversion factor. Table 9 shows the average concentration of LTC-IC in the peripheral blood calculated from values measured on 23 normal adults, together with the average concentration of circulating clonogenic cells (BFU-E plus CFU-GM plus CFU-GEMM) obtained for the same 23 samples. The derived value of ~3 LTC-IC per ml is ~75-fold lower than the concentration of circulating clonogenic cells both measured here (Table 9) and reported previously. (Sutherland HJ. et al, A Practical Guide. Boca Raton, CRC Press Inc., 1991, pp 155, and Ogawa M. et al, ~ . . .
.

20802~5 Blood S0:1081, 1977.) Hence the frequency of LTC-IC
relative to other nucleated cells in the blood (~l per 2 X 106) is ~100-fold lower than the frequency of LTC-IC
relative to other cells in the BM. (Sutherland HJ, Proc.
Natl. Acad. Sci. USA 87:3584, 1990.) Example 17 PHENOTYPE OF CIRCULATING LTC-IC.
The distribution of LTC-IC and clonogenic cells in various phenotypically-defined subpopulations of the T cell-depleted, light density fraction of normal peripheral blood were then assessed. ~hese were obtained using the FACS to separate cells on the basis of their light scattering properties, expression of CD34, HLA-DR, and RH-123 uptake.

Figure 15 shows the bivariate contour histograms of light density T cell-depleted normal peripheral blood cells stained with anti-CD34 and anti-HLA-DR. Panel B shows the distribution of these cells in the low side scatter window (fraction I ~ II in Figure 4A). Panel C shows the distribution of HLA-DRhi9h and HLA-D~W cells after also gating for CD34~ cells as indicated in Panel B. The light scattering properties of CD34'HLA-DRLW and CD34~ELA-DRhi9h cells are demonstrated in Panels E and F, respectively.
Unsorted, unstained control and irrelevant (lD3) antibody-stained cells are shown in Panels D and A, respectively.

As illustrated in Figure 15, even after removal of the R
cells from the light density fraction of leukapheresis samples, the frequency of cells expressing readily detectable levels of CD34 (as defined by the vertical gate shown in Figure 15B) was still very low (0.1-0.5%) as compared to the non-T cell-depleted light density fraction of normal marrow, where values of 1-4% are typically obtained, (Civin CI. et al in Knapp W. et al (eds):
Leucocyte Typing IV. White Cell Differentiation Antigens, Oxford, Oxford University Press, 1989, pp 818) even using - ~ . , ........... . ~ . . .

.

20802~5 ~imilarly stringent gating criteria. (Sutherland HJ et al, Blood 74:1563, 1989.~ Most of the cells in the fraction defined as CD34~ expressed no or low levels of HLA-DR
(Figure 15C), and had low SSC properties (Figure 15E and F). Cells with a CD34~ and HLA-DRlW phenotype (defined by the horizontal gate shown in Figure 3C) were found almost exclusively amongst the smallest light density cells (low FSC, Figure 15E).

Figure 16 shows the light scatter profiles of T cell-depleted light density normal blood cells (Panel A). The mean ~ SEM of the percentages of nucleated cells (open bar), clonogenic cells (stippled bar), and LTC-IC (solid bar) in each sorted fraction are shown in Panel B (n=4).

Figure 16 shows the distribution of LTC-IC and clonogenic cells observed when the total light density T cell-depleted fraction of peripheral blood cells was subdivided into 3 populations defined by their light scattering properties: l-low FSC, intermediate SSC; II - intermediate to high FSC, low SSC; and III - all remaining cells (i.e.
open FSC, intermediate SSC). Although each gated population contained approximately equal number of cells, virtually all LTC-IC and most of the clonogenic cells were consistently found in the fraction containing the smallest cells (I). No LTC-IC and less than 5% of all clonogenic cells were found in fraction III. Therefore subsequent sorts, only cells in the low SSC fractions (I and II) were analyzed.

Figure 17 shows a representative histogram of CD34~, light density T cell-depleted normal blood cells (in the previously described low SSC window shown in Figure 16A) double-stained with PE conjugated anti-HLR-DR. CD34+HLA-DRL~ cells were further subdivided into CD34~HLA-DR
(fraction 1) and CD34+HLA-DR~ (fraction 2) cells as shown in Panel A. The remaining cells are CD34'HLA-DRhi9h, : .
~ , .

, 208025~

indicated as fraction 3 in Panel A and referred to as CD34'HLA-DR~ cells in Panel B. The dark histogram in Panel A shows the profile for unstained cells. This mean + SEM
of the percentages of nucleated cells (open bar), clonogenic cells (stippled bar), and LTC-IC (solid bar) in each sorted fraction are shown in Panel B (n=3).

Figure 17 shows the results of functional assays performed on cells sorted both according to their expression of CD34 and HLA-DR. In this case, only CD34' cells were assayed.
These were then divided into HLA-DRhi~h and HLA-DR~W cells were further subdivided into an HLA-DR- and an HLA-DR~
population. No LTC-IC and very few directly clonogenic cells were detected in the CD34~HLA-DRhi3h fraction.
However, further subdivision of the remaining CD34~, HLA-DRLWcells did allow some differential separation of LTC-IC
and directly clonogenic cells, moxe of the latter (~40% vs ~10% LTC-IC) being found in the HLA-DR~ fraction. Table 10 shows the enrichment and recovery values obtained for LTC-IC and clonogenic cells in various HLA-DR
subpopulations of the light scatter gated, CD34' fraction, as compared to the unstained, light density, T cell-depleted starting population in these experiments.
Recovery of LTC-IC in the CD34', HLA-DRLW fraction was >100% in all 5 experiments performed suggesting that all circulating LTC-IC express readily detectable levels of CD34, as do those in normal BM. (Sutherland HJ et al, Blood 74:1563, 1989.) Isolation of a rare subpopulation of circulating cells defined by the same properties as previously used to purify BM LTC-IC (i.e. low density, low forward light scatter, high expression of CD34 and low expression of HLA-DR), allowed a much greater enrichment (>l,000-fold beyond the light density, T cell depletion step) of circulating LTC-IC to be routinely obtained.
Thus, even though the initial frequency of LTC-IC in normal peripheral blood as much lower (on a per cell basis), the final purity of LTC-IC achievable from normal - - : . , 208025~

peripheral blood using these parameter~ (~0.5-1%, Table 10~ was approximately the same as the best yet described for normal BM. (Sutherland HJ, Proc. Natl. Acad. Sci. USA
87:3584, 1990.) Recovery of clonogenic cells in these same experiments appear to be somewhat lower (Table 10), suggesting that some circulating clonogenic cells may have been excluded from the CD34~ population gated for in these studies, or that subopitimal plating efficiency of clonogenic cells may have been achieved when highly purified populations were assay. Failure to detect additional clonogenic cells in the higher FSC/SSC
fractions (II and III, Figure 16) due to potential inhibition of their colony-forming ability by the presence of increased numbers of monocytes were ruled out by mixing experiments (i.e. no reduction of clonogenic cells detected when cells from Fraction I were mixed with cells from Fraction III in a 1:2 ratio).

The results of combined staining for CD34 expression and RH-123 uptake are shown in Figure 18. In particular, Figure 18A is a representative histogram of CD34', light density T cell-deplet0d normal blood cells double-stained with Rh-123 and sorted into CD34+RH-123dULL and CD34'RH-123bri9ht fractions (fractions 1 and 2, respectivel~; Panel A). The dark histogram in Panel A shows the profile for unstained cells. The mean + SEM of the percentages of nucleated cells (open bar), clonogenic cells (stippled bar), and LTC-IC (solid bar) in each sorted fraction are shown in Panel B (n-3).

In this case, no difference was noted between circulating LTC-IC and clonogenic cells in terms of their distribution between the CD34' RH-123bri9ht fractions with mor than 80% of both being found in the Rh-123dUlL but most of the clonogenic cells are Rh-123br;9ht, thus allowing their differential isolation by sorting according to this parameter. (Udomsakdi C. et al, Exp. Hematol 19:338, ., ~
. .
,.

.

1991.) Nevertheless, the final purity of LTC-IC in the light density, T cell-depleted, CD34', Rh-123dUl~ fraction of normal blood was similar to that obtained by ~electing for HLA-DRIw cells (Table 10) or by application of the same criteria to BM. (Udomsakdi C. et al, Exp. Hematol 19:338, 1991.) This reflects a similarly greater overall enrichment achieved with blood versus mArrow using either HLA-DE expression or retention of Rh-123 as the final separation parameter.
Example 18 4-HC SENSITIVITIES OF CIRCULATING PROGENITORS. Because normal circulating clonogenic cells were known to be a ~uiescent population (Eaves CJ et al in Goldman JM (ed):
Bailliere's Clinical Haematology. Vol. 1, #4. Chronic 1~ Myeloid Leukaemia, London, sailliere Tindall, 1987, pp 931) and appeared phenotypically to be more similar to LTC-IC in either blood or BM than to the clonogenic cells found in the BM, it was of interest to compare the sensitivities of circulating clonogenic cells and LTC-IC
to 4-HC, using the same type of treatment protocol that is in widespread clinical use for treating autologous marrow transplant. In this set of experiments, LTC-IC function (~efore or after exposure to 4-HC) was assessed in terms of the clonogenic cell content of assay cultures evaluated after 4 and 8 weeks (rather than after 5 weeks as in the experiments described above), since previous experiments and revealed differences in LTC assays of BM samples for autologous transplant when these two time points were compared. (Winton EF et al, Exp. Hematol 15:710, 1987, and 3Q Eaves CJ et al, Blood Cells (in press).) Results for LTC-IC and clonogenic cells in normal peripheral blood and BM
are shown in Figure 19.

Figure 19 shows a comparison of the number of clonogenic cells and LTC-IC surviving a 30 minute exposure to lOO~g/ml of 4-HC at 37C with 7% erythrocytes present.
Values shown are the mean + SEN of the percentages of .
- ' -~
:~ ' ' , clonogenic cells and LTC-IC from normal BM (open bars) and normal blood (solid bars) as a percent of values for control cells (n=3 or BM and n=4 for blood cells).

A dramatic difference in the effect of a 30 minute exposure at 37C to lOO~g/ml of 4-HC on the viability of clonogenic cells and LTC~IC appear to be similar to BM
LTC-IC in their relative resistance to 4-HC. For LTC-IC
from both sources, a slight increase in 4-HC resistance was noted for LTC-IC defined by the longer clonogenic cell output endpoint (i.e., 8 weeks).
' Example 1~
DIFFERENTIATIV~ POTENTIAL EXPRESSED BY CIRCULATING LTC-IC
IN LTC. Table 11 shows the relative proportions of BFU-E, CFU-GM and CFU-GEMM in the total clonogenic population of 5 week-old LTC initiated with circulating LTC-IC of varying purities,and compares these to the relative numbers of these same types of clonogenic cells in the original blood samples. Data for unseparated and LTC-IC
enriched cell populations from normal BM obtained in previous studies (Sutherland HJ et al, Blood 74:1563, 1989) are also shown in Table 11 for comparison. It can be seen that the differentiative behaviour exhibited by LTC-IC in normal blood and BM is similar and is also not affected by the purity of the ~TC-IC in the starting population. In both cases, a significant skewing towards the generation of CFU-GM by comparison to the number of CFU-GM and BFU-E actually found in normal blood or BM was observed. To some extent this might be expected because all stages of granulopoietic cell differentiation are supported in the LTC system whereas erythropoiesis appears to be blocked at the stage of mature BFU-E production.
(Coulombel L. et al, Blood 62:291, 1983.) As a result, this latter contribution to total BFU-E numbers in vivo is absent from LTC-derived populations.

- , ~ ~ .

' :" -: . -- -208025~

The present invention has been described in detail and with particulsr reference to the preferred embodiments;
however, it will be understood hy one having ordinary skill in the art that changes can be made thereto without departing from the spirit and scope thereof.

J~. . ~ ' ' , . ' ' ' . ' ,~'',, , ~',.' ' ~' ' '' ` ' . ~ . ' . '' , ' ' ' ' . ' ' ' . ' .

.' ' .' ' ' ' :- ' , " ' :
-:: . ', ' :
' .; ~ ' ' - 69 - 20802~5 TAE~LE 1 No. of dono~e~c cdl~ per LTC-IC
~5wk~

Normal BM 4.3 t 0.4 (5~
Nomlal Blood 3.7 ~ 1.2 (3) CML Blood 3.1 + 0.416) Mean +SEM of lralue~ from ~3~) experlmc~t~ calculatcd by multlplylng the frequcncy of LTC-IC In cach acperiment (dcterm~ed by llmltlng dllutlon assays) by the total number o~ celb plated In dl LTC to determlne the total number of LTC-IC for that exper~erlt. The total content of clonogenlc progenlto~s In all LTC for ~n ~Idual ~ent was obtained dlrectly grom clonogen1c progenltor ass~rs.

- . , ~ 70 - 20802~5 Frequency of Pr~mlUvc Progenltors In the CML PaUcnts 8tudled Patlcnt No. WBC/ALClono8cnlc Ccll~/ml, LTC-lC/mLa Ix 10~
.
2g.000 17,000 2 156 704,0G0 266,000 3 137 .72.000 1~,000 4 62 82,000 7.200 262 1.060,000145,000 6 104 161,u~10 8.300 7 1 10 86,000 1.300 8 142 344,000 12,500 9 ~36 1,090,000 10,000 Mean + SEM162 ~ 40403,000 + 145,00052.000 + 31,000 a These numbers rcprcsent absolute LTC-IC valucs calcubted ~s descrlbed h the Mdhods .. . .. .
. - .

- : ~:

:. .' 20802~5 ~ ~ 5 e 8 ~ ~Y~. ;;~

c y ; ~ 5 ~ ¦ ~ J ~ 5 :~ ~ . ~ ~ ~ i ~; 5 ~

5 ~ a ~ ~ 9 ~ e t I i 3 ~ 5 ` e i ~ t~ 1 ~ ~ ~ ~ C

3 ~ ~ ~ i i l ~ ia~ ~ ~
a a a ,,,,," " " v_ , , " ., ; -:

2080~55 a 5 "

~Z ~ 0~0~ j COCO~O~ CO ~0~ j . R' ~
R5 ~ 888 88 O 80 ~ ~ ~ 3 ~ 5, R~ ~ ¦ ~b--~ - 0 j ~IR a3 ~U ~ ~ ~3~5 ~ R~5 ~ f 5 R R ~ ~ ~ -O N ~o ~N~ R !~ ~ ~

~ 8 ~ N ~ ~ ~ ;N r~ 5 ~ f ~ D D

~ ~ ~ ~ e ~ g ~ z .

20802~

Relattve Numbers IExpressed as a Pcrcent of the Total) of Dlfrerent l~pcs o~
Clonogcnlc CeUs Prescnt In CML E~lood and Produccd by LTC-IC In CML Blood a~er 5Weelcs In LTC

O~lgln o~ No. d Clonogenlc CcL~ l,TC-IC
SamplesSamplesBFU-E CFU4M CFU-GEMM BFU-E CFU-GM CFU-GEMM

CMLPB 17 65+3 3u~+3 1.3+0.215+3 B3~3 1.2+0.4 Normal PEla 23 74 ~ 3 24 + 2 2.2 1 0.3 11 + 2 89 + 2 0.5 + 0.2 NormalBMb 20 36+3 62~4 1.2+0.29+2 91 +2 0.8+0.3 a From 23 b From l~i .

.
- ~ - ;

.

~able 6 Growth F~ctor Product~on by Retrovirally Infected M2-lOB4 Cell~

Bioactivity (ng/mL) in Medium ~.
Retroviral Ve~tor GN-CSF G-CS~ 3 ~-.
~ninfectsd or N2 <O.Ol ~O.l <O.1 10 N2-tkG~-CSF 1.0 ~O.l <O.1 N2-tkG-CSF ~0.03 20 ~O.l N2-tkIL-3 <O.Ol <O.1 6 :: .

.

' Table 7 Comp~ri30n of the Content of Five-Week-Old Cocultures Containing Differ0nt Type3 of or No Feeder (plastic) ~2-1084 ~uman MF Plastic Nonndherent cells per 13 + 6 28 + 6* 1.2 + O.4*
cell pl~ted (no. (15) (13) (12) experiments) 10 Clonogsnic cells per 4.8 + 1.0 6.2 ~ 1.1 0.10 + 0.03 100 cells plated (no. sxperi~ent ) (15) (14) (15) LTC-IC per 1,000 c~ll~ 2.3 ~ 0.4 3.4 + 0.8 0 plated (no. (12) (10) (4) experim~nt~) * Using a paired t-test a significant difference from ~2-lOB4 was observed.

r~ :
'' ' ' : ~
.
.- . ' `. : ' ~ : ' .' ' ' "' ~' ', . '~ . ' ''' ''' ' ' ' ~ ~ ; .

~ ' ' ' .

2~0~55 _ 76 -Table 8 Number of Clonogenic Progenitor~ per L~C-IC Harvested From Five-Week-Old Cocultur~ Containing Different Types of Feeder ~
No. of Clonogenic Feeder Progenitor~ Per LTC-IC

Human MF 4.0 + 0.7 N2-lOB4 3.3 + 0.7 M2-lOR4 + G 5.5 ~ 1.0 N2-lOB4 ~ GM 6.4 + 2.1 M2-lOB4 + I~-3 2.8 + 0.8 M2-lOB4 + G + GN 3.7 ~ 1.1 N2-lOB4 + I~-3 + G 4.7 + 1.1 M2-lOB4 + Ih-3 + GM 4.3 + 1.0 M2-lOB4 ~ IS-3 + G + GM 6.0 + 2.4 _ Abbrevi~tions: G,G-CSF; GN,G~-CSF.

~' '': - ' : : -2~025~

Tablc 1. Quantl~tlon of LTC-IC and Clono~ lc Cells In Nonnal PB

Cell l~pe Concentrat~on IPer ml) BFU-E 170 ~ 20 CFU-GM 51 ~ 5 CFU-GEMM 4.6 + 0.6 LTC-IC 2.9 + 0.5 .

Values for Indhrldual patlents were calculated by multlplying the progenltor frequency per 105 cells by the total nucleatcd cell recove~y aft both the T cell-depletlon and FH denslt~r cent~ugatlon steps and then agaln ~y the WBC pcr ml. Values shown are the mean SEM of data obtatned from 23 dlaerent normal Indhrlduals.

-: ~

.

- 78 - 2~8025~

Table 2. Frequency. Enrlc~cnt and Recovy of LTC-IC and ClonogeDlc Cells In Va~ous SubpopulaUon~ of the CD34+. T Cen-depleted Llght D~ty ~ractlon of Nomlal Perlpheral 13boli Cells De~lned Acc~dlng to Uleir Exp~lon d HLA-DR

.
Cell ~pe Souroe F requencyb Er~mentC % Recove~d No. of Evaluated ~ Exp .
LTC-lCa ur~orted~ 0.0022 + 0.0004 - - 5 cells DRhlgh f Og 5 DRlowf 3.7 ~ 1.11930 + 470 300 + 40 2 DR~h 1.0 _ 0.554~ ~ 270 48 + 36 3 DR I 2.8 + 0.51470 + 340 280 + 17 3 Clonogenlc unsorted~0. 1 1+ 0.02 - . 5 cells celh DRhlgh t 7.6 + 0.4 65 + 22 2.7 + 1.2 4 DRIwf N.D,~

DR+h 21 ~ 2 240 ~ 40 21 + 6 4 DR I 15 + 4 160 + 40 38 + 5 4 , a Measured as the total number of clonogenlc cells present after 5 weeks (I.e.. -~dC the absolute LTC-IC number) b Frequency of the cell typc evalualed ILTC-IC or total clonogenic ceDs) rclatlve to all nucleated cells In the populatlon analyzed. fTo co~vert the LTC-IC frequencies shown to absolute frequenclcs. dlvide by 4.1 c Calcula5ted by dhrldlng the progenltor frequency pcr 105 sorted ceL~ls by the progcnitor frcquency pcr 10 unsorted, T cell-depleted. IUht-den~ty ceL~ In each lndhrldual acper~ment. and then derivlng thc mcan _ SEM of these values for the number of c~cper~nent~ performed.
d Calcubted by mulUpb1ng thc percentage of cell~ rctrlcvcd In the fractlon Indlcated by the correspondhg cakulated progenitor er~chmcnt for cach Indlvidual exper~nent ~dellned in ~ootnote c). and then derl~rtng the mcsn t SEM of thesc values for the number of e~cperlments perfonned.
e Llght denslty. T cell-depleted cell~.

TABLE 10 cont ' d f AS deflned In F'~gure 15C .
g Nonc dctccted: Le., d~.l5. Enrlcbment and mt~v~value~ therefore t cakulatçd.
h Refer~ to a ~ubpopulatlon o~ DR~ ceDs dcDned z~ fractlon 2 In F~gurc 17A.
Rcfers to a ~ubpopulauon of CD34+HI~-DE~1W cdb deflned as ~actlon 1 In ~gure 17A.
Not donc as a ~eparate measu~ment. DRlW ~ecovy ~ralues can bc ~nfased b~r addlng togethcr values for DE~ and DP:.

.. , ?

, , ~ .: - . ~ , : ... . . .

- 80- 208025~

Relative Proportion~ of Dif ferent Type~ of Clonogenic Cell~
Detected Before and After S weeks in LTC ( 9~ of total ) .

Orlglnal No. of Clonogenlccellsa LTC-lCb Progenltor Sampl~
Source BFU-E C~V-Gbl CFU-CEMM BF'U-E CF~-GM CFI~-GEMM
Llght dcnslty fractlon of nonnal bloodC 23 74 ~ 3 24 2 2.2 t 0.3 11 ~ 2 89 1 2 0.5 ~ 0.2 LTC-IC
enrlchcd fractlon of norrnal bloodd 6 72 + 5 28~ 5 0.8*0.3 11~ 1 89~ 2 1.0 1 0.6 Nonnal BMe 10 36~3 62~4 1.2 l 0.2 9~2 91 1 2 0.8~0.3 L rC -IC
er~ched fractlon of normal E3Me 10 24 ~ 5 75 ~ S 0.4 ~ 0.3 9 + 3 90 + 4 1.4 + 0.8 a Data shovm are the mcan ~ SEM of proportlonal values for spec1flc clonogenlc ccll types expressed as a percent of aU clonogenlc cclls ILe.. E~FU-E plu~ CFU-GM plus CFU-GEMM~
measured ~n standard short-term mcthylce~ulosc assays.
b Data shown arc the mean * 5EM of proportlonal valucs for spcc~c clonogenlc ccD typcs exprcsscd as a pc~lt of all clonogen~c ccDs ILe.~ ElFU E plus CFU-GM plus CFU-GEMM~
measurcd h me~hylcdlulose a~ay~ Or ccDs harvested rrom 5 weck-old LTC.
c Sarne samplcs as In Table 9 .
d Data rrom LTC IC In rractlon I ICD34~. DR ) ~n F4~urc l7 Ind~ a~d fractlon 1 ICD34~. Rh-123dUII~ In F~gurt 18 (n=3) .
C Data ror nonnal BM rr~TI prr~ously publlsbed studlcs.20

Claims

1. A cell preparation comprising primitive hematopoietic stem cells, wherein the hematopoietic stem cells are obtained from the blood of leukemic patients, are Philadelphia chromosome positive and produce detectable clonogenic progenitors in long term culture.

2. A cell preparation comprising long term culture initiating cells obtained from the blood of a patient with chronic myeloid leukemia having elevated white blood cell count.

3. A cell preparation as claimed in claim 1 or 2 wherein the hematopoietic stem cells are CD34+ and HLA-DR+.

4. A cell preparation as claimed in claim 3 wherein the hematopoietic stem cells are FSChigh, SCClow, and Rh-123+ve.

5. A cell preparation as claimed in claim 1 or 2 having a purity of about 10%.

6. A method for preparing a cell preparation as claimed in claim 1 comprising obtaining a sample from a leukemic patient, preferably a blood sample from a leukemic patient having an elevated white blood cell count, most preferably having a white blood cell count greater than 20 x 109 white blood cells per litre of blood;
and isolating a cell preparation from the cell sample which comprises cells which are Philadelphia chromosome positive and produce detectable clonogenic progenitors in long term culture.

7. A method for quantitating primitive leukemic hematopoietic stem cells in patients with leukemia comprising obtaining a sample which contains primitive leukemic hematopoietic stem cells from a leukemic patient;
optionally enriching primitive leukemic hematopoietic stem cells in the sample; coculturing the sample with a feeder cell layer for at least five weeks under conditions which permit the production of clonogenic progenitor cells;
harvesting nonadherent cells and adherent cells;
subjecting the harvested cells to secondary assay culture under suitable conditions to express clonogenic progenitor cells; detecting and quantitating the number of clonogenic progenitor cells; and, quantitating the number of primitive leukemic hematopoietic stem cells in the sample on the basis of the linear relationship between the number of clonogenic progenitor cells and the number of primitive leukemic hematopoietic stem cells initially in the sample.

8. A method as claimed in claim 7 wherein the sample is enriched by combining the sample with antibodies to CD34 labeled with a detectable marker and antibodies to HLA-DR labeled with a detectable marker wherein the antibodies to CD34 and antibodies to HLA-DR are labeled with different detectable markers, and isolating a cell fraction which is CD34+ and DR+ detectable markers.

9. A method as claimed in claim 8 wherein the antibodies are labeled with different fluorescent labels.

10. A method as claimed in claim 7 wherein the sample is a blood sample from a leukemic patient having an elevated white blood cell count, preferably greater than 20 X 109 white blood cells per litre of blood.

11. A method of purging a mammalian bone marrow sample containing primitive leukemic hematopoietic stem cells to prepare a bone marrow cell suspension substantially free of primitive leukemic hematopoietic stem cells comprising: obtaining a sample of bone marrow cells from a leukemic patient; substantially depleting red blood cells in the sample; coculturing the depleted bone marrow cell sample with a feeder cell layer; and harvesting and suspending the cultured bone marrow cells.

12. A system for testing for a substance that affects hematopoiesis of primitive hematopoietic stem cells comprising: preparing a cell suspension that is enriched in primitive hematopoietic stem cells;
coculturing the cell suspension with a feeder cell layer for at least 5 weeks in the presence of a substance which is suspected of affecting the hematopoiesis of primitive hematopoietic stem cells and assessing LTC-IC maintenance, clonogenic cell production, and/or production of nonadherent cells.

13. A cell preparation comprising primitive hematopoietic stem cells obtained from the blood of normal individuals and which produce detectable clonogenic progenitors in long term culture.

14. The cell preparation as claimed in claim 13, wherein the primitive hematopoietic stem cells in the preparation are CD34+ and HLA-DR-, FSClow, SCClow, and Rh-123dull and 4-hydroperoxycyclophosphamide-resistant.

15. The cell preparation as cliamed in claim 13 which has a purity of circulating primitive hematopoietic stem cells of approximately 0.5 - 1%.

16. A method for quantitating primitive hematopoietic stem cells in blood of normal individuals comprising obtaining a blood sample which contains primitive hematopoietic stem cells from a normal individual, optionally enriching primitive hematopoietic stem cells in the blood sample; coculturing the sample with a feeder cell layer for at least five weeks under conditions which permit the production of clonogenic progenitor cells; harvesting nonadherent cells and adherent cells; subjecting the harvested cells to secondary assay culture under suitable conditions to express clonogenic progenitor cells; detecting and quantitating the number of clonogenic progenitor cells;
and, quantitating the number of primitive hematopoietic stem cells in the blood sample on the basis of the linear relationship between the number of clonogenic progenitor cells and the number of primitive hematopoietic stem cells initially in the sample.