AU7735201A - Fetal cell recovery method - Google Patents

Description

0@ 0

S

@0

S

.00 0 00 SO 0

S

S.

S

005 0S 0 *0@

AUSTRALIA

PATENTS ACT 1990 PATENT OF ADDITION NAME OF APPLICANT: Genetype A. G.

ADDRESS FOR SERVICE: DAVIES COLLISON CAVE Patent Attorneys 1 Little Collins Street Melbourne, 3000.

INVENTION TITLE: "Fetal cell recovery method" The following statement is a full description of this invention, including the best method of performing it known to us: 0 0 0000 0000 0 @550 .00.

0 0 .555 05

S

P:\OPER\FSpo\f=eW =11s.ps3.dmc-02IIOAfI -2- FIELD OF THE INVENTION The present invention relates to a method for recovering fetal cells from the uterus of a pregnant female using antibodies to distinguish maternal cells having a maternal cell surface antigen encoded by a maternal allele of a polymorphic genetic locus from fetal cells lacking the maternal cell surface antigen.

*o 0@ •The present invention is a patent of addition of Australian Patent No 649027. That "o 10 patent discloses a method for the recovery of fetal cells from the blood sample of a pregnant woman using antibodies to distinguish maternal cells having a maternal cell surface antigen encoded by a maternal allele of a polymorphic genetic locus from fetal cells lacking the maternal cell surface antigen.

0a 0@0000 Oea 15 BACKGROUND OF THE INVENTION 0 The examination of fetal cells for early detection of fetal diseases and genetic S° abnormalities is undertaken in approximately one out of every thirty pregnant women.

The main indication is maternal age (over 35 years). The tests may involve DNA gene 20 typing or, more commonly, the use of live fetal cells for chromosomal karyotyping.

Fetal cells are usually obtained by amniocentesis, the removal of amniotic fluid from the amniotic cavity within the amniotic sac or placenta. The procedure presents a risk of harm to the fetus, particularly after the first trimester of pregnancy. The risk to the fetus together with the high cost of the procedure have prevented the establishment of examination of fetal cells for early detection of abnormalities as a routine procedure in pregnancy.

In the late 1970s and early 1980s, Herzenberg and his colleagues reported that fetal cells were present in maternal blood as early as 15 weeks gestation. The authors separated maternal and fetal cells using fluorescence-activated cell sorting (FACS) by staining PAOPER\FsspccWfta c]S-spS3.docO2/10A) I -3maternal blood for a distinguishing paternal HLA antigen. The authors state that the demonstration that fetal cells enter maternal circulation and can be isolated by FACSenrichment procedures could have practical significance in enabling karyotyping without the need for amniocentesis. The authors state that this would be possible if the frequency of successful isolation of cells at 15 weeks is sufficiently high and the cells could be induced to divide (enter metaphase). Furthermore, extensive HLA typing reagents, or other cell surface reagents would need to be developed to distinguish maternal and fetal cells. To date, the technique has not been successfully adapted for use as a clinical technique for either karyotyping or fetal DNA analysis.

Fetal cells present in maternal blood have been used to perform analysis of genes present in the fetus. In one technique, the maternal and fetal cells were not separated and the DNA from the cell mixture was amplified with Y chromosome-specific primers to determine whether the fetus is male. It has been suggested that DNA amplification 15 techniques can also be performed to detect gene sequences associated with disease in this manner. Of course, the method cannot be used where the mother is a carrier for the trait.

S° Although maternal peripheral blood can be used to source fetal cells for diagnostic o:Oo 20 testing, the technique may not be suitable in all circumstances. For example, for pregnancies after the first, maternal peripheral blood may include some contaminating fetal cells from previous pregnancies, since fetal cells may maintain in maternal circulation for some years after pregnancy.

It is preferable to conduct sampling early to enable the detection of fetal disease and genetic abnormalities as early as possible in the pregnancy. An advantage of uterine fetal cell sampling is that collection can take place in the early stages of gestation. Fetal cells have been detected as early as 6 weeks and up to 17 weeks.

In 1971, Shettles (Nature 230:52 (1971)) proposed that during pregnancy, fetal trophoblasts migrate from the placenta through the uterine veins into the uterine cavity and could reach the endocervical canal. He tested this hypothesis by collecting P:\OPER\Faslspc\fetal clls-spoc3.doc-02/10 I -4transcervical cells and staining with a Y-chromosome fluorescent dye for fetal sex determination. He correctly diagnosed the sex of 10 fetuses. Since then endocervical sampling has been used as a source of fetal cells for diagnostic testing. Detection methods have included cell culture followed by metaphase staining, PCR for Ychromosome sequences (Morris and Williamson 1992), and fluorescence in situ hybridisation ofX and Y chromosome signals (Adinolfi et al. 1993).

DESCRIPTION OF THE PRIOR ART 10 Herzenberg and his colleagues have described methods for separating maternal and fetal cells in maternal blood using fluorescence-activated cell sorting (FACS). In Herzenberg et al, Proc. Natl. Acad. Sci. USA 76:1453-1455 (1979), cells in blood samples from week pregnant HLA A2-negative women were stained for HLA A2 antigen. Stained cells were separated by FACS and collected to enrich the population of fetal cells.

Although the technique was demonstrated to effectively identify male, HLA A2positive cells in maternal blood, to date the technique has not been successfully adapted for general applicability. In Iverson et al, Prenat. Diag. 1:61-73 (1981), peripheral blood lymphocytes (PBLs) from either 15 week or 21 to 25 week pregnant women were examined. If the woman was HLA A2-negative, her cells were stained with anti-HLA 20 A2 reagents, sorted by FACS onto microscope slides (for fetuses that were HLA A2positive), stained with quinacrine and examined microscopically for Y chromatinpositive cells. The authors report that fetal cells enter the maternal circulation as early as weeks gestation.

Detection of trophoblastic cells retrieved from the endocervical canal has been reported as early as 6 weeks (Ishai et al. Prenat Diagn 15:961 (1995)). The success of detection of fetal cells in transcervical cell samples depends largely upon the method of collection. The use of cotton swabs has been found to collect few fetal cells and increases the risk of collecting cells of male partners (Adinolfi et al. Prenat Diag 15:943(1995)). More successful methods have been cytobrush, aspiration and lavage.

Cytobrush samples contain more maternal cellular debris (Kingdom et al. Obst. Gyn PAOPER\Fms~pecrl1 1s-p3.do-02/I0A) I 86:283 (1995)) primarily squamous and endometrial cellular elements, leukocytes and macrophages (Bulmer et al. Prenat. Diag. 15:1143 (1995)).

Trophoblastic cellular elements have also been retrieved from behind the cervical mucus plug by flushing the endocervical canal with saline and collecting the lavage (Rhine et al. Birth Defects 12:231 (1977)). Comparison of endocervical and intrauterine lavage has not shown that trophoblastic cells are more frequent at the Scervical or uterine pole of the endocervical canal (Adinolfi et al. Prenat Diag So15:943(1995)).

Various methods have been utilised in the analysis of transcervical cell samples.

Fluorescence in-situ hybridisation and polymerase chain reaction have been used to detect aneuploidies and Y chromosome derived DNA sequence in male fetuses (Griffith-Jones et al. Br. J. Obstet. Gynaecol. 99:508(1992), Adinolfi et al. Lancet 15 342:403 (1993)). Quantitative fluorescent PCR has also been used to detect chromosome specific DNA sequences (Pertl et al. Lancet 343:1197 (1994)). PCR has been used to amplify Rh(D) positive sequences from Rh(D) negative mothers (Adinolfi et al. Lancet 345:318 (1995)).

20 X-linked and autosomal recessive diseases have been extremely difficult to diagnose using transcervical cell samples due to contamination by maternal cells (Adinolfi et al.

Prenat Diag 17:539 (1997)). The isolation of trophoblastic cells is important for the diagnosis of these diseases and monoclonal antibodies, density gradients and short term in vitro cultures have been proposed to address this problem (Adinolfi and Sherlock, Human Reproduction Update 3:383 (1997)). In vitro culture methods have been used, but suffer from the problem that maternal fibroblasts can overrun the trophoblast cultures. The isolation of clumps of trophoblasts by ficoll separation or manual selection has also been used to extract fetal DNA for diagnostic purposes (Adinolfi et al. (1997)). However, contamination with maternal DNA was common.

Bianchi et al, Cytometry 8:197-202 (1987) report a technique that allows direct hybridization to the DNA of cells which were flow sorted onto nitrocellulose filters P:\OPER\Fasspcc\fctal clls-spe3.doc-02/10/I -6which eliminates the need for a DNA isolation step. The method was performed on human cord blood. The authors state that the technique is useful in situations where there is a limited amount of DNA available for analysis such as for fetal cells recovered from maternal blood.

U.S. Pat. No. 4,675,286 (to Calenoff, issued Jun. 23, 1987) describes a method for •0 obtaining fetal cells for diagnostic examination in which cells derived from the placenta are collected from the uterine cavity and outer surface of the amniotic sac and are incubated with a separation antibody which binds preferentially to maternal cells. The S 10 separation antibody may be to the conserved region of a Class 1 HLA antigen or Pmicroglobulin, on the basis that fetal trophoblast cells derived from the placenta do not ever express HLA antigens.

Practical Flow Cytometry (Second Edition) by Howard M. Shapiro, Alan R. Liss, Inc.

15 1988 is a comprehensive work on flow cytometry. The book describes how to build, purchase and use a flow cytometer and how to design analyses and analyze the data produced thereby.

0•••0 Truneh et al, Cytometry 8:562-567 (1987) describe a method for detection of very low 20 receptor numbers on cells by flow cytometry. The method involves staining the cells with vesicles containing thousands of fluorochrome molecules, which vesicles are conjugated to antibodies with the desired specificity.

Albright et al, Cytometry 7:536-543 (1986) describe the use of centrifugal separation of cells in sputum specimens as an alternative to flow cytometry.

Society for Clinical Cytology 1988 Abstracts p.6 describe in situ hybridization for detection of structural and numerical abnormalities using nonradioactive probes for detection of aneuploidy and translocations (Pinkel et al, No. 13). The immunopotentiality of cancer patients was studied by sorting peripheral blood T cell subsets by a multiparameter analysis using monoclonal antibodies and flow cytometry (Nonura et al, No. 14).

P:\OPER\Fassp=feca1 mlls-sp3.doc-4)IOiAI -7- Each of the above-described references and the references cited therein is incorporated herein by reference in its entirety.

The reference to any prior art in this specification is not, and should not be taken as, an acknowledgment or any form of suggestion that that prior art forms part of the common general knowledge in Australia.

Throughout this specification and the claims which follow, unless the context requires S 10 otherwise, the word "comprise", and variations such as "comprises" and "comprising", will be understood to imply the inclusion of a stated integer or step or group of integers or steps but not the exclusion of any other integer or step or group of integers or steps.

SUMMARY OF THE INVENTION The present invention provides a method for selectively recovering fetal cells from the uterus of a pregnant female. The method is performed on a uterine sample from a pregnant female having a first cell surface antigen encoded by a first allele of a polymorphic genetic locus and combining the cells with a first antibody specific for the :20 first cell surface antigen. Cells bound to the antibody are separated from any cells which are not bound to the antibody and non-bound, separated fetal cells are recovered.

The method is based on the differential reactivities of the maternal and fetal cells to antibodies specific for polymorphic cell surface antigens, particularly MHC antigens.

The method separates maternal and fetal cells based on the reactivity of the maternal first cell surface antigen with an antibody specific for the maternal first cell surface antigen and recovering the non-bound cells. In particular, the fetal and maternal cells are separated based on the non-reactivity of the fetal cells to an antibody specific for a maternal cell surface antigen which was not passed on to the fetus, being a nontransmitted maternal allele.

P:\OPER\Fsp ctaf I clls-sp3doc-02/IO)I -8- In a preferred embodiment, the cells are also contacted with a second antibody specific for a second maternal cell surface antigen encoded by a second allele of a polymorphic gene locus. The maternal cells react with both antibodies. Except when the fetus inherits both maternal alleles, fetal cells react with, at most, one of the antibodies. In a most preferred embodiment, the alleles are for a single genetic locus, preferably an HLA locus. In this way when the fetus does not inherit the nontransmitted allele from the father, the fetal cells, having inherited only one of the two maternal alleles, react with only one of the two antibodies. In instances where the fetus is the same HLA type as the mother for those alleles where the fetus inherits the non-maternally transmitted S 10 allele from the father), there will be no difference between the reactivity of the fetal and 04 maternal cells with the antibodies. To reduce or avoid this occurrence, testing of both 0• .00 parents is preferred.

When more than one antibody is used, the antibodies can be applied simultaneously or 15 sequentially. Simultaneous application requires that the different antibodies can be S. individually identified, whether by using different fluorochromes or by other means.

"s The differential antibody reactivities of the fetal and maternal cells can be used to separate the cells in two general ways based on affixing the antibodies to a solid phase 20 or based on FACS using fluorochrome-labelled antibodies. For solid phase separations, the cells are separately contacted with solid phase-affixed antibodies for each selected maternal antigen magnetic bead-affixed antibodies). Maternal cells bind to each antibody. Fetal cells fail to react with antibodies for nontransmitted antigens, are not bound to the solid phase, and are recovered. For FACS-based separations, the antibodies are labelled with a fluorochrome, and unlabelled fetal cells are separated from labelled maternal cells. In a preferred embodiment, antibodies for two maternal alleles are labelled with different fluorochromes and double-labelled maternal cells are separated from unlabelled and single-labelled fetal cells.

The separated fetal cells can be used in a variety of procedures. The DNA in the recovered fetal cells can be used to determine a variety of genetic traits, particularly using DNA amplification methods. In a preferred embodiment, the fetal cells are P:\OPER\Fas\spec\fca clls-spe3.doc-02/10/l -9analysed individually by single cell PCR analysis. In a further preferred embodiment, the fetal cells are cultured, preferably from single cell suspensions and the cultures used for karyotyping, genetic analysis or therapeutic purposes. Cytogenetic abnormalities classically revealed by karyotyping and by in situ hybridization may be revealed by identification of PCR amplification products by methods such as gel electrophoresis.

DETAILED DESCRIPTION OF THE INVENTION *0 so0 The present invention provides a method for selectively recovering fetal cells from a 10 uterine sample. The method is based on separating maternal and fetal cells based on •differential reactivities of the cells to antibodies specific for polymorphic surface antigens of the cells, particularly MHC antigens and more particularly HLA antigens.

For HLA loci (or any other polymorphic genetic loci), the fetus inherits one allele for an HLA locus from the mother. The non-reactivity of fetal cells with an antibody specific S 15 for the antigen expressed by the nontransmitted allele can be used to separate fetal and maternal cells.

Specifically, an antibody for the selected maternal antigen can be affixed to a solid phase. Maternal cells become solid-phase affixed by reaction with the antibody. When 20 the antigen is encoded by the nontransmitted allele (and the fetus does not inherit the nontransmitted maternal allele from the father), fetal cells do not bind to the antibody and can be recovered. Alternatively, the antibody can be labelled with a fluorochrome and labelled maternal cells separated from any unlabelled fetal cells by FACS.

Preferably, antibodies specific for two different maternally expressed cell surface antigens are used. When those antigens are encoded by alleles at the same locus, and the mother is heterozygous at that locus (and the fetus does not inherit the nontransmitted maternal allele from the father), the mother and fetus have one allele in common and one different allele for that locus. Therefore, when a maternal uterine cavity or transcervical cell sample is combined with antibodies for those selected alleles, the maternal cells react with both antibodies and fetal cells react with only one of the antibodies, (unless the allele not inherited from the mother is inherited from the P:\OPER\Fas\spccfctaI clls-sp3.doc-O2/IIOA) I father and then the fetal cells will not be able to be separated from the maternal cells on the basis of these alleles. Where this is the case, locus should be selected for differentiation of the maternal and fetal cells).

When the two selected maternal alleles are from different unlinked polymorphic genetic loci, there is a one in four chance that the fetus will inherit both alleles from the mother, where the mother is heterozygous at both loci. In this case, three out of four times, the o• fetus will inherit either one or neither of the two maternal alleles from the mother.

Therefore, when a maternal uterine cavity or transcervical cell sample is combined with S 10 antibodies for each selected allele of those HLA loci (and the fetus does not inherit the nontransmitted maternal allele from the father), the maternal cells react with both antibodies and the fetal cells react with one antibody (two out of four times) or neither antibody (one out of four times). Where the fetus inherits both maternal alleles, whether derived from the mother or father, another antigen should be selected for differentiation 15 of the maternal and fetal cells.

Where the mother is homozygous at one of the two selected loci, the fetus will have a one in two chance of inheriting both alleles from the mother, but where the mother is homozygous at both selected loci, the fetus will inherit both alleles and it will only be So 20 possible to differentiate maternal from fetal cells on the basis of paternal antigens at the selected loci. However, if the fetus has also inherited both alleles from the father, so that the fetus is, like the mother, homozygous at both loci, another locus will need to be used to differentiate fetal and maternal cells.

When the antibodies for two maternal antigens are labelled with different fluorochromes, single-labelled and unlabelled fetal cells are readily separated from double-labelled maternal cells using fluorescence-activated cell sorting (FACS), resulting in separation and recovery of fetal cells. Likewise, when antibodies for more than two maternal antigens are used, each can be labelled with a different fluorochrome and the fetal cells identified and separated on the absence of one or more cell surface antigens encoding a non-transmitted maternal allele.

P:\OPER\Fas\spec\fctaI mls-spe3.dow-02~110) -11- Similarly, when antibodies for a cell-surface antigen encoded by a nontransmitted maternal allele are bound to a solid phase, such as a magnetic bead, only maternal cells bind to the solid phase. The non-bound fetal cells can be readily separated from the bound maternal cells and recovered.

The recovered fetal cells can be used in a variety of assays. The DNA in the recovered fetal cells can be used to determine a variety of genetic traits. In a preferred embodiment, the fetal cells are cultured and the cultures used for karyotyping. It is also envisaged that cytogenetic abnormalities classically revealed by karyotyping and by in S 10 situ hybridization may also be revealed by gel electrophoresis of PCR products.

The present fetal cell recovery method is advantageous because it provides a means of purifying fetal cells where information regarding the paternal contribution is not necessary. Not only is there no need to identify alleles of the natural father that may be .15 distinguishing, but, so long as the fetus does not inherit nontransmitted maternal alleles S o from the father, only a limited number of antigens need be investigated to find a pattern that distinguishes maternal and fetal cells. Thus knowledge of neither parental see contribution is critical, although it may be helpful to carrying out the invention.

20 There is also no need to use a mixture of antibodies specific for all non-maternal alleleencoded antigens for the selected loci. The method does not rely on attempting to identify fetal cells with discriminating paternal alleles that the child may have inherited.

Furthermore, since the method separates fetal and maternal cells based on absence of the maternal antigen in fetal cells, fetal cells can be effectively separated even if the paternal antigen had never been previously encountered. Once antigens present on the maternal cells are identified, any of the nontransmitted antigens can form the basis of the separation. Furthermore, the presence of cells in the sample which do not react with one of the antibodies indicates that the fetal cells have been distinguished based on differential antibody reactivity. No additional evaluation needs to be made.

The method also does not attempt to distinguish the fetal cells based on "fetal" antigens such as fetal hemoglobin which may also be expressed by some maternal cells. In the P:\OPER\Fas.spec\fal calls-spc3.dO4)210A)I 12present method, rather than selecting a fetal antigen that is present on some maternal cells, the method utilizes a maternal antigen which is not present on the fetal cells. Once an antibody that does not react with some cells in the sample is identified, the cells which do not react with the antibody are fetal cells. Furthermore, nucleated fetal cells that express HLA antigens fail to react with the antibody because the particular HLA antigen is not transmitted to that fetus. The lack of reactivity is not related to a particular fetal stage of development or to a particular fetal cell type.

i: ~Selection of Polymorphic Genetic Locus-Encoded Antigens S 10 The fetus inherits half of its chromosomes from its mother and half from its father. For each polymorphic locus, such as an HLA locus, the fetus inherits one allele from its mother and one from its father. Therefore, the mother and fetus always have at least one allele for each polymorphic locus in common. The maternal alleles that the fetus does not inherit are unique to the maternal cells unless the fetus inherits the nontransmitted 15 allele from the father. Of course there will be no alleles unique to the mother at a locus S if the mother is homozygous at a locus. (Although there will be an allele unique to the fetus if it has inherited a non-maternal allele from the father).

ooooo For any polymorphic locus, the fetus will inherit at most one of the maternal alleles.

20 When the mother is heterozygous for the locus, the mother and fetus will have the same alleles only when the fetus inherits the nontransmitted allele from the father. The likelihood of that event depends on the frequency of the nontransmitted allele in the population. Loci are preferably selected to minimize the likelihood that the fetus inherited the allele for the selected antigen from the father as described below.

Numerous polymorphic loci are well known, as are methods for determining alleletypes for those loci directly or indirectly by determining the cell surface antigens expressed by the loci. The selected loci have at least two alleles each of which comprises a significant percentage of the population. Preferably, the loci have numerous alleles which are present in a significant percentage of the population. Selection of such loci maximizes the likelihood that the father will not provide the nontransmitted maternal allele to the fetus.

P:\OPER\Fa\speofemI =Ulssp3.doc-)2/R)A) 13- In a preferred method the mother is typed and in a more preferred method, both parents are typed. In particular, the results of these tests are utilised to select the antibodies, alleles and loci to be used in the detection and separation of fetal cells. In a preferred embodiment, the selected alleles of the loci are routinely detected using serological methods. In this way, the antibody reagents available to type the loci are then available for use as reagents in the separation method of this invention. However, recent advances in DNA technology have vastly improved the detail to which these antigens can be characterised. Techniques such as restriction fragment length polymorphism S 10 (RFLP), sequence specific primer analysis (SSP), sequence specific oligonucleotide 0 probe analysis (SSOP or just SSO) and direct sequencing can also be used for typing.

Reverse hybridization line probe assay (LiPA) offers many advantages over the more classical techniques used for HLA typing as well as over the newer DNA-based techniques. The main advantage of the LiPA HLA typing system is the requirement of 15 only one amplification step and one hybridisation step to obtain multiple answers. The S presentation of result on a membrane strip allows simple processing and analysis.

Suitable loci encode cell surface antigens present on cells in early through late stages of differentiation. However, any surface antigens expressed by nucleated fetal cells can be 20 used. Exemplary suitable genetic loci encode the blood group antigens, particularly the ABO group, the Kidd group, the Duffy group, the Lewis group, the P group, the I group, and like groups. Polymorphic loci which would provide results are those associated with the minor histocompatibility antigen loci and any other polymorphic loci that encode cell surface antigens present on fetal and maternal cells. Most preferred genetic loci for use in the method are the HLA loci. The antigens are expressed by all nucleated cells and are present on fetal cells early in development. The following tables demonstrate the antigens with the greatest diversity and show the most frequent molecular variants in Australia.

HLA-A (the more frequent antigens only) Antigen Broad Group No. of molecular types* Most common alleles Al 8 A*0101 A2 49 A*0201, A*0202 A3 8 A*0301 PAOPEM~s~spe\feI c]Is-spc3.doc-O2/IOA)I 14 All A23 A24 A26 A29 A31 A32 A33 A34 A36 A66 A68 A69 A74 A80 A9 A9 AlO AlO A19 A19 A19 A19 A19 AlO AlO A28 A28 A19 A*l1Ol A*2301 A*2402 A*2501 A*2601 A*2901, A*2902 A*3001 A*3101 A*3201 A*3301 na na na A*6801 A*6901 na A* 8001 0@ 0 Oe eo

C@

0 0 eeg e CG go S 0 e 00

S

e.g ee 0 eee e gee...

6 ego.

0 OeeO 0@OS

S

e.g.

ego...

o C go e

C

e *e

C

gee HLA-B (the more frequent antigens only) Antigen Broad Group No. of molecular types B7 25 B8 13 B13 7 B14 6 B15 64 B18 13 B27 23 B35 37 B37 5 B38 B16 7 B39 B16 24 35 B41 B42 B44 B46 B47 B49 B51 B52 B53 B54 B56 B57 Most common alleles B*0702 B*0801 B* 1301 B* 140 1( 64 B*1402( 65 B* 1501 B*1801, B*1802 B*2701, B*2702 B*3501, B*3502 B*3701 B*3801 B*3901 B*4001 na na B*4402 B*4501 na B*4701 B*4901 B*5001 B*5 101 B*5201 B*5301 na B*5501, B*5502 B*5601 B *5701 B12 B12 B21 B21 B5 B5 B5 B22 B22 B22 B17 P:\OPER~as\sjxecfwI clls-spc3.doc-02/IOA)I B58 B17 6 B*5801 B59 1 na B40 6 B*4001 B61 B40 8 B*4002 B62 B15 17 B*1501 B63 B15 2 B*1517 B67 2 na B73 1 na B78 5 na B81 2 na B82 2 na B83 1 na HLA-C (the more frequent antigens only) Antigen Broad Group No. of molecular types* Most common alleles Cwl 4 Cw*0101 Cw2 4 Cw*0202 Cw3 12 Cw*0303 Cw9 Cw3 3 Cw*0303 CwlO Cw3 3 Cw*0304 Cw4 8 Cw*0401 4 Cw*0501 Cw6 7 Cw*0602 Cw7 14 Cw*0701, Cw*0702 Cw8 9 Cw*0802 Cwl2 7 Cw*1203 Cwl3 1 na Cwl4 4 na Cwl5 10 Cw*1502 Cwl6 3 Cw*1601 The number of variants is approximate, as there will be more reported regularly.

Furthermore, methods for determining the HLA type of an individual are well known and numerous antibodies for HLA antigens are commercially available. This description uses the HLA antigens as exemplary. Similar considerations apply to use of antigens encoded by other polymorphic genetic loci.

For the present method, if the maternal HLA type is known, any HLA antigen-specific antibodies that are available for any two maternal HLA antigens can be used. Preferably the antibodies are specific for the antigens of any HLA locus at which the mother is heterozygous.

P:\OPER\Fas\spec\felal clls-spe3.doc-o2/1OAi -16- When the maternal HLA type is unknown, the alleles for at least one HLA locus, preferably two or more loci, are determined, preferably by using antibodies specific for antigens of that locus. When antibodies specific for two alleles of a locus react with the maternal cells, those antibodies can be used to identify and separate fetal cells. In this way, it is clear at the outset of the separation procedure that antibody reagents useful in the method are available. Alternatively, the HLA type can be determined by DNA methods and an allele can be selected based on availability of antibodies for the encoded Santigen.

0 Preferably, the loci are any two of the Class I loci B, and C) and the Class II (DR, DQ and DP) loci. Preferably the locus is HLA-A, HLA-B, or HLA-C or a combination of these. An individual heterozygous at each HLA locus must express four major HLA specificities (HLA-A and HLA-B) and 2 minor (HLA-C) ones. More preferably the locus is HLA-A or HLA-B or both, being highly polymorphic. Since those loci are 15 commonly typed by serological methods, HLA antigen-specific antibodies for many alleles of those loci are commercially available. In particular, there are numerous 0o monoclonal antibodies for HLA-A antigens.

00o• In selecting the HLA alleles to use, an important consideration is the availability of 20 antibodies for the antigens expressed by the alleles. When antibodies for the antigens of "a number of alleles present in the mother are available, preferably, the alleles chosen are from a locus that has numerous alleles present in significant numbers in the population, as described above. The selected alleles can be for two or more different loci.

Antibodies specific for the antigen of any of the alleles of maternal HLA loci can be selected for use in separating maternal and fetal cells of the maternal uterine cavity or transcervical cell sample. The antibodies can be polyclonal or monoclonal, preferably monoclonal.

When the alleles for two loci are used, the fetus will only inherit both alleles if recombination has not occurred. In that case, the use of two loci will have the same result as use of a single locus. That is, half of the time the selection results in an P:\OPER\Fasp\speccla clls-spe3.doc-4)2/10I) -17effective separation and half of the time, the fetus has the same alleles as the mother and no separation is achieved. To ensure separation using two alleles, the selected alleles should preferably be from different chromosomes, that is from different maternal haplotypes.

Most DNA methods and serologic methods do not determine the haplotype of the mother. For two genetic loci, four allele types are determined, but no information as to the alleles that comprise a haplotype is determined. For the present method, it is S°preferred that the haplotype is determined to ensure that the selected alleles are not S 10 inherited together. Methods for determining haplotypes include analysis of haploid cells (exemplified by analysis of DNA in sperm) and a method where DNA is diluted to haploid (Ruano et al, Proc. Natl. Acad. Sci USA, 87:6296-6300 (1990)). A preferred method is described in EPO application Serial No. 90.309107.2, filed 8/20/90 entitled Intron Sequence Analysis Method for Detection of Adjacent and Remote Locus Alleles 15 as Haplotypes by Malcolm J. Simons, published Feb. 25, 1991 Preferably, both maternal alleles for a genetic locus are determined, the alleles are necessarily present on different chromosomes and the fetus inherits one and only one of the alleles from the mother. Therefore, whenever appropriate reagents are available, 20 antibodies specific for two antigens encoded by alleles of a single genetic locus are selected.

As stated previously, the antibodies are preferably specific for the two antigens expressed by the alleles of one HLA locus. In another preferred embodiment, the antibodies are specific for both antigens expressed by the alleles of more than one HLA locus. When the fetus inherits the nontransmitted maternal allele(s) from the father, the separation process is repeated selecting antibodies to antigens produced by the alleles of another HLA locus or loci. Similar considerations apply for selection of antigens encoded by other polymorphic genetic loci.

In the absence of determination of parental genotype, a panel of 2 or more antibodies for antigens encoded by a selected locus or loci can be used. Preferably the panel comprises P\OPER\Tpc\f.l cs-sp3.doc-02IO11I 18antibodies to one or more of the alleles of HLA-A, B or C. Population surveys in a variety of human populations have identified over 110, 235 and 65 alleles at the HLA- A, B and C loci respectively. The panel preferably comprises an antibody to each of the alleles significantly represented in the population, in particular those identified in the tables above.

As a result of linkage, HLA-A, B and C alleles may be inherited together as a haplotype. For example, Aland B8 are often inherited together, as are A3 and B7.

Therefore it may not be efficient to use antibodies to more than one HLA locus. More S 10 preferably the panel of antibodies is selected against HLA-A, B or C antigens. Thus in a preferred embodiment, the panel comprises antibodies against some or all of A*0101, A*0201, A*0202, A*0301, A*1101, A*2301, A*2402, A*2501, A*2601, A*2901, A*2902, A*3001, A*3101, A*3201, A*3301, A*6801, A*6901, and A*8001.

Preferably, the panel comprises antibodies to antigens encoded by a HLA locus and a 15 non-HLA locus.

Selection of alleles may be based on ethnicity of the parents. For example, HLA-A34, which is present in 78% of Australian Aborigines, has a frequency of less than 1% in Australian Caucasoids. Alleles may also be selected for association with disease. For o 20 example, B27 is known to be associated with Ankylosing spondylitis, Reiter's disease (syndrome), and Acute anterior uveitis. The use of this allele would thus serve a dual function in cell separation and disease indication.

Once the antibodies for use for a particular sample are selected, the antibodies are either bound to a solid phase or used to label the cells, depending on the type of separation procedure.

Source and Purification of Cells The uterus is divided into the broad-ended fundus, body and thin isthmus that is the uterine cervix. Histologically, the uterus is composed of three layers: 1) an outer perimetrium composed of connective tissue, 2) a thick smooth muscle and elastic tissue, P\OPER\Fassp\fta mell-spe3.doc-OYIOAu! -19myometrium, and 3) a mucosal epithelial lining called the endometrium. The cervix is made mostly of dense connective tissue, about 2.5 cm in length and is covered interiorly by a mucous secreting ciliated epithelium at the upper regions and by stratified squamous epithelium at the vaginal end.

The present method is designed to recover fetal cells present in a uterine sample. As referred to herein "uterine sample" includes a sample from the uterus and sampling may take place from any surface in the uterus including the outer surface of the amniotic S° sac, the placenta and in particular the uterine cavity or endocervical canal. Cervical 10 sites are preferred for sampling, since this is likely to cause least impact on the fetus.

Maternal uterine cavity or transcervical cell samples generally contain two main types of nucleated fetal cells: cytotrophoblasts and syncytiotrophoblasts cells. Preferably, the main types of nucleated fetal cells are maintained in the purified sample. Other cell 15 types may be sorted such as villous trophoblasts and villous mesenchyme, although villous trophobasts do not express HLA molecules. The type and quantity of each cell type depends on the site and method of collection.

•*Vee The cells can be obtained by a variety of relatively safe procedures. Methods of 20 collection include cervical cotton swab, cytobrush, aspiration of cervical mucus, lavage of the endocervical canal and uterine lavage. In particular, samples can be obtained from aspiration of mucus from just above the internal os or the lower uterine cavity.

Lavage is generally conducted with a saline wash, but other isotonic solutions are suitable. Typically, endocervical lavage with 5-10ml or intrauterine lavage with saline provides sufficient fetal cells upon separation from maternal cells. The sample may be collected using a combination of methods.

The sample can be in any solution which maintains cell integrity and minimises cell lysis, damage or clumping of cells, preferably a physiological solution, or more preferably, a saline solution or tissue culture medium with or without the addition of sera. Cell collection and preparation should be gentle enough not to degrade or fragment P:\OPER\Fasspep\fciaJ mIl-sp3.doc-4021ArU fetal cells by harsh treatment. The sample is preferably stored at 0' to 4 C until use to minimize contamination and cell degradation.

As referred to herein "contaminants" refers to dead and damaged cells and non-cellular matter including mucus, but not cell clumps. Aspirate and lavage samples may contain clumps of cells with the characteristics of trophoblastic cells. To aid fetal cell separation, the clumps of cells are preferably treated to obtain a suspension of single .0cells. The clumps may be separated by techniques known to a skilled person, such as 1 enzymatic, chemical or mechanical separation. For example, enzymatic separation may utilise protease or trypsin, chemical separation may utilise acetyl cysteine and mechanical separation may involve gentle teasing, sonication, aspiration or micromanipulation.

The number of fetal cells in the sample varies depending on factors including the age of 15 the fetus, maternal age, previous pregnancies or miscarriage, operator's skill level, Smethod of sampling, number and frequency of samplings, the volume of washing in each lavage and the volume aspirated. Cells derived from the uterine cavity or transcervical cells may have different characteristics, such as size, cell type, cell recovery, level or type of contamination or fluid substrate. The skilled person will take 20 these factors into account in obtaining and preparing the cell sample.

00 Preferably, the female is in her first trimester of pregnancy. More preferably the fetus is between 6 to 17 weeks gestation. In some circumstances it may be appropriate to obtain fetal cells from a female after the birth, or from the placenta, since these cells may have therapeutic applications. Reference to the collection of cells from a pregnant female includes a female who has recently been pregnant.

The female may be a human or another mammal. Preferably the mammal is a domesticated species, such as a cow, sheep, goat, dog, cat or horse.

P:\OPER\Fas\spcc\fetal clls-spe3.doc-02/10i -21 Solid Phase Separation As stated previously, fetal cells can be separated from maternal cells based on the lack of reactivity of the fetal cells with an antibody specific for a nontransmitted maternal cell-surface antigen. Briefly, the maternal antigen-specific antibody can be bound to the solid phase directly or through use of a solid phase-affixed bridge which binds the antigen-specific antibody. Maternal cells bind to the bridge and are affixed to the solid 0 phase. When the antigen-specific antibody is for the antigen of the nontransmitted

C.

allele, fetal cells do not bind to the solid phase and can be separated and recovered.

10 Separation procedures are described below.

Selection and Preparation of Solid Phase for Separation The solid phase for separating fetal cells can be any one of those used conventionally for antibody-based cell separations, such as plastic plates, dishes, flasks, and tubes; 15 beads, polyacrylamide, Sephadex, agarose-polyacrolein, trisacryl and latex; nylon disks and ox red blood cells. The solid phase can be in the form of a dish, flask or tube with which the cells are contacted, present as beads or disks in a container or packed into a column beads and red blood cells) through which the cells are passed. Most Spreferred is use of magnetic beads present in a tube or other container. Affinity 20 chromatography methods for cell isolation are well known and are described in Affinity Chromatography: A Practical Approach, ed. P.D.G. Dean, IRL Press (1985) and the references cited therein, each of which is incorporated herein by reference in its entirety.

The antibody can be bound to the selected solid phase by conventional processes for the solid phase materials including adsorption, ionic bonding, van der Waals adsorption, electrostatic bonding, or other non-covalent or covalent bonding. Procedures for noncovalent bonding are described in U.S. Pat. No. 4,528,267. Procedures for covalently bonding antibody solid supports are described by Ichiro Chibata in IMMOBILIZED ENZYMES. Halsted Press: New York (1978) and A. Cuatrecasas, J. Bio. Chem.

245:3059 (1970), the entire contents of which are hereby incorporated by reference.

PAOPER\Faspspc\fiwl cl-sp6.do.4O2/0A)1 -22- The antibody can be bound to the solid phase either directly or indirectly such as through the use of a surface layer to which the antibody is attached. For example, the surface can be coated with a protein and the protein coupled with the antibody using a coupling agent (for example, glutaraldehyde). The antibody can be bound by applying the antibody in aqueous solution to a surface coated with a layer having free isocyanate groups (such as a polyether isocyanate). Additionally, the antibody can be coupled to a hydroxylated material by means of cyanogen bromide.

:0 Alternatively, an antibody "bridge" can be bound to the solid phase and used to S* 10 indirectly bind the anti-maternal antigen antibody either prior to or during the separation procedure. Bridges include antibodies which are specific for immunoglobulin of the species of the anti-maternal antigen antibody, also referred to as secondary antibodies.

For mouse monoclonal antibodies, exemplary bridge antibodies are goat anti-mouse immunoglobulin, goat anti-mouse IgG and goat anti-mouse IgM. Other bridges include use of staph protein A (SPA) (which binds the Fc region of immunoglobulin) or use of avidin to bind biotinylated antibodies. When applicable, specialized methods for binding antibodies to a particular type of solid phase are described below. Many solid phase-based cell separations are generally performed by reacting the specific antibody with the cells and then reacting the cells with solid phase-affixed second antibody or 20 other bridge.

For solid phase applications, following binding of the antibody, the solid phase can be "blocked" to reduce non-specific binding such as by use of water-soluble, non-immune animal proteins including albumins, casein and non-fat milk.

In a preferred solid phase separation method, the antibody is affixed to a metal or magnetic bead and combined with the cells in a tube or other container. Following binding, solid phase-affixed maternal cells are removed from the cell suspension medium using a magnet. Beads which are attracted by a magnet are commercially available. A preferred bead is sold by Dynal AS under the name DYNABEAD (available from Robbins Scientific of Mountain View, Calif.). Antibodies can be conjugated to the beads according to the manufacturer's directions. In addition, beads P:\OPER\Fas pec~fetI cclls-p3.doc-O 2/IiA)l 23 having goat anti-mouse IgG antibodies conjugated to the surface are available. Any mouse monoclonal IgG antibody can be readily bound to the bead by incubation with the bead in a physiologic solution. A tube holder containing a magnet to which the DYNABEAD beads are attracted is also available from Dynal AS.

For a column separation, the column packing materials must have appropriate spacing so that maternal cells can react with an antibody affixed to the column materials and fetal cells which do not have the antigen can pass through the column. Column

S.

materials which are suitable for use in separating cells include various types of beads 10 andoxredbloodcells.

Similarly, an antibody for the selected maternal antigen can be affixed to a solid phase such as a microtiter plate, a culture dish, a test tube, a centrifuge tube or like suitable containers for cells. Non-bound fetal cells can be separated from maternal cells and removed by gently washing the surface of the solid phase. Solid phase materials which are suitable for affinity separation of cells are well known and are commercially available.

As stated above, for each type of the solid phase the antibodies can either be affixed to 20 the solid phase or bound to the solid phase by reaction with a solid phase-affixed bridge.

Additional alternative embodiments which provide the same result can be readily envisaged by one of ordinary skill.

Preparation of Cells Suspension media and appropriate cell concentrations for solid phase-based separation procedures can be those used conventionally for solid phase-affixed, antibody-based cell separations. The suspension medium can be any physiologic medium compatible with antibody function and with preservation of the fetal and maternal cells; e.g., physiologic saline, phosphate buffered saline and like solutions used for immunoassay procedures. The suspension medium can also be a tissue culture medium which optimizes the ability to culture recovered cells. Suitable tissue culture media are described in the description of suspension media for FACS separation methods. The P:\OPER\Fas\spec~\fclal cclls-sp3.doc4)2/IIOA) -24suspension medium can additionally contain antibiotics, and like additives commonly present in tissue culture media. The suspension medium should be free from reagents which can interfere with subsequent use of the separated cells. For example, preservatives are preferably not used if the cells are to be cultured following separation.

For separations using bead-affixed antibody, the suspension medium is preferably calcium- and magnesium-free to minimize monocyte activity and thus decrease the tendency of the monocytes to engulf the beads during the separation process.

10 Solid phase separation procedures are performed in a relatively short period of time, so SS SO Sthat use of PBS is sufficient to maintain cell viability throughout the separation procedure. However, the cells are preferably resuspended in a tissue culture medium following separation for uses involving culture of the cells.

Following purification of the sample, if any, the cells are washed and the pellet is 60009* resuspended in the selected suspension medium at an appropriate concentration for solid S:00o phase separations. Generally, those concentrations range from about 105 to about 108 cells/mi, preferably about 106 to about 107 cells/ml.

0000 The cell suspension is contacted with the solid phase-affixed antibody as described hereinafter.

Solid Phase Separation Process To effect the separation, cells of the sample are combined with a solid phase-affixed antibody specific for one of the selected maternal antigens for a time sufficient for antibody binding. The period of time varies depending on the temperature and is well known. Incubations of as little as one minute at room temperature are adequate. Longer periods may be required at 4 C. More important is that there is sufficient time for the cell surface antigen to come into contact with the solid phase-affixed antibody. Usually times of from about 10 to about 60 minutes are sufficient.

P:\OPER\FAsp\l j c-sp.3.dO4)2IOAi)I The cells bound by the maternal antigen-specific antibody (and thus to the solid phase) are separated from cells which are not bound to the solid phase. When there are cells in the sample which are not bound by the selected antibody, those cells are recovered.

When all cells in the sample are bound by the first antibody, the procedure is repeated using an antibody for the second selected maternal cell surface antigen and non-bound, separated cells are recovered. The separations using the first and second antibodies can be performed sequentially. Preferably, the separations are separately performed at substantially the same time using different aliquots of the sample cells.

S 10 For bead separations, maternal cells always bind to the bead. Fetal cells do not bind when the solid phase-affixed antibody is for a nontransmitted maternal antigen. The cells in the sample can be divided and an aliquot of cells can be combined separately Se.e with beads containing antibody specific for each of the selected maternal antigens simultaneously. Fetal cells do not bind to one of the sets of beads, if the nontransmitted allele was not inherited from the father. Solid phase-affixed maternal cells are separated S. from non-bound fetal cells by separating the beads from the suspension medium.

For magnetic beads, the cells are incubated with the beads for about an hour, preferably at 4 C. The incubation is preferably not more than about an hour to avoid clumping by 20 monocytes. The beads are present in the incubation at about 106 to 107 beads/ml. A S. preferred cell to bead ratio is about 2 beads/cell. Following incubation, maternal cells affixed to the beads are moved to the side of the tube in which the incubation is o performed using a magnet and non-bound fetal cells are removed by gentle washing.

The final wash medium preferably contains not more than 1% protein.

For column-based separations, the method is accomplished by passing an aliquot of the sample through the column at a flow rate sufficient to permit antibody binding. Fetal cells pass through the column when the solid phase-affixed antibody is for a nontransmitted maternal antigen. The cells in the sample can be divided and an aliquot of cells can be passed through a column containing antibody specific for each of the selected maternal antigens simultaneously. Fetal cells elute from one of the columns, if the nontransmitted allele was not inherited from the father.

P:\OPER\Fas\pc\fdtI c1h-sp.3do4)2/IOA)I -26- Alternatively, cells can be passed through one column and, when no fetal cells elute, either cells can be removed from the column and passed through another column or a second aliquot of cells can be passed through a column for the second maternal antigen.

Cells can be separated from column materials or other solid phase surfaces using an excess of the maternal antigen or a competitive inhibitor for the antibody binding site, by mechanical disruption pipetting), or, for ox red blood cells, by lysis of the erythrocytes. Additional techniques for removing cells affixed to particular surfaces are S•well known and include cleavage of particular bonds used to attach the antibody to the 0 o 10 solid phase; e.g. cleavage of disulfide bonds with thiol. Preferably, separate aliquots of S0 °O the cells are used so that fetal cells having the selected maternal antigen need not be separated from the solid phase, although in some cases, eg small sample volumes, this may not be possible.

For separations using containers such as plates or tubes, the cells are incubated with the solid phase-affixed antibody for a time sufficient for antibody binding. Following 0 incubation, any non-bound cells are gently washed from the surface and recovered.

000* 0 0 00 FACS Separation S"The fetal cells can also be separated from maternal cells based on the differential labeling of the cells and separation of the labelled cells using FACS. Briefly, as stated previously, the maternal cells can be labelled with an antibody for a selected maternal antigen. Preferably, the maternal cells are labelled with more than one antibody for antigens encoding alleles at the same or different loci. When the two antibodies used are selected for the two maternal antigens encoded by alleles of the same heterozygous genetic locus, fetal cells are single-labelled, unless the allele not transmitted from the mother is provided by the father. The separation procedure is described below using double labelled maternal cells and single labelled fetal cells as an example.

P:\OPER\Fas\spec\feal clls-spc3.doc-02/10/01 -27- Preparation of Cells Suspension media and appropriate cell concentrations for FACS-based separation procedures are those conventionally used. Following purification, if any, the cells are washed and the pellet is resuspended in the suspension medium at an appropriate concentration for sorting. The concentration should not be too dilute. However, the flow cytometer can be adjusted to provide an appropriate cell flow rate using relatively concentrated or relatively diluted samples. Preferably the cell concentration is from about 10 6 to about 5 x 10 8 more preferably from about 5 x 106 to about 1 x 10 7 cells/ml.

The suspension medium is a physiologic solution, such as a physiologic buffer, to maintain cell integrity. Most physiologic buffers, e.g. Tris buffer, phosphate buffer citrate buffer, phosphate buffered saline (PBS) are suitable. Balanced salt solutions such as Earle's balanced salt solution (EBSS), Hank's balanced salt solution (HBSS), and Gey's balanced salt solution (GBSS) are also suitable. Preferably, the suspension medium is a tissue culture medium basal medium Eagle and Dulbecco's modified Eagle's medium), more preferably an enriched tissue culture *0 medium suitable for use with lymphocyte cultures such as RPMI 1640. The use of a 096* tissue culture medium, particularly a medium adapted for the growth of the sample cell 20 type, provides an environment which maximizes cell stability.

The suspension medium can additionally be supplemented with a protein source at a relatively high concentration. The protein source can be albumin such as bovine serum albumin (BSA) or, preferably, human serum albumin (HSA) at a concentration in the range of from about 5 to about 10%. Alternatively, the protein source can be serum such as fetal calf serum or human serum at a concentration of from about 5 to about Since lymphocytes can be stimulated by the presence of foreign antigens, preferably the protein source is human. Most preferred is the use of about 5% autologous plasma which can be harvested from a purified blood sample and is non-immunogenic.

P:\OPERTFas\specffa cclls-spc3.doc-4)2/IOA) -28- Alternatively, the protein source can be added to the flow cytometer collection vessel, rather than to the suspension medium, to cushion the fall of the cell into the vessel, enhancing cell stability.

Labelling the Cells The cells of the sample, preferably purified cells, are labelled with fluorescent antibodies specific for the antigens encoded by at least one maternal polymorphic locus, selected as described previously. The antibodies can be polyclonal or monoclonal, preferably monoclonal. Preparation of polyclonal and monoclonal antibodies for an 10 antigen of interest is well known.

For HLA antigens, as stated previously, HLA antigen-specific antibodies are commercially available. Typically the HLA Class I loci B and C) and the Class II DR and DQ loci are determined by serological methods. Therefore, antibodies specific for those antigens are readily available. Sources of HLA antigen-specific antibodies include Genetic Systems (Seattle, Wash.) and C6 Diagnostics (Mequon, Wis.). Blood group antigens are also determined serologically and the antibodies are commercially available.

20 The antibody is labelled with a dye that facilitates cell sorting, particularly a fluorochrome. Suitable dyes for FACS analysis and/or separation are well known.

Those dyes are described in Practical Flow Cytometry (Second Edition) by Howard M.

Shapiro, supra, at pages 115-198. Preferred dyes are fluorochromes including fluorescein fluorescein isothiocyanate--FITC), rhodamine tetramethylrhodamine isothiocyanate--TRITC), phycoerythrin allophycocyanin (APC) and Texas Red (Molecular Probes, Eugene, Oreg.). The combinations of fluorochromes used for labeling are chosen so that distinguishable wavelengths of light are emitted. A preferred combination is a fluorochrome that emits green light together with one that emits red or orange light, FITC with PE or Texas red.

For those flow cytometers that can perform a four-color sort, the cells can be labelled with antibodies for antigens expressed by four alleles. In that case, preferably, the P:\OPER\Fas\poc\ffcaI cs-spe3.doc/)0I/I0 -29antibodies are specific for both antigens expressed by the alleles of two maternal HLA loci and more preferably the mother is heterozygous at those loci. Maternal cells are labelled with all four fluorochromes. Fetal cells are labelled with two of the four fluorochromes when none of the nontransmitted maternal alleles is inherited from the father. By using four fluorochromes from two loci, the fetal cells remain distinguishable from the maternal cells even when the fetus inherits one of the nontransmitted maternal alleles from the father. A second staining is only necessary when the fetus inherits both nontransmitted maternal alleles from the father or the mother is homozygous for both loci. When the antibodies are for antigens expressed by three or four maternal loci, 10 using the additional dyes increases the likelihood that the fetus did not inherit each of the maternal alleles. Where the machine can perform more than a four-colour sort, the •cells can be labelled with antibodies for antigens expressed by as many alleles as the number of fluorochromes the FACS has the ability to sort.

The antibody can be labelled with the fluorochrome, directly or indirectly, by well S known methods. The conjugation methods for attaching labels to antibodies generally can be used for these purposes. Direct labeling methods for dyes such as FITC are described in Catty et al, in "Antisera in Immunoassays with Special Reference to Monoclonal Antibodies to Human Immunoglobulins", IMMUNOASSAYS FOR THE 20 80's, supra, pp 133-153 and THE HANDBOOK OF EXPERIMENTAL IMMUNOLOGY 4th Edition, Vol. 1: Immunocytochemistry, ed. D. M. Weir, Blackwell Scientific Publications and in the publications cited in those references. The entire contents of each of those references is hereby incorporated by reference.

Preferably the labelled antibody is purified to remove unbound label prior to use.

Preferably, the procedure used also purifies the antibody composition to provide the immunoglobulin fraction, more preferably, to provide the IgG fraction for polyclonal antibodies or, for monoclonal antibodies, the IgG or IgM fraction depending on the isotype of the antibody. Procedures for isolating the antibody fraction of an antibody include the use of recombinant protein G for IgG and immunoprecipitation for IgM.

Procedures that additionally separate unbound fluorochrome are well known and P:\OPERA a\ spcc\fcala ctl1s-spc3.doc-(2/IOAII include use of a DEAE G 35 sephadex (Pharmacia) column. Exemplary purification procedures are described in detail in the examples.

For dyes such as PE which are more difficult to attach while maintaining antibody activity, the antibody can be labelled with biotin. The dye can be attached by incubation of the biotinylated antibody with PE-avidin or PE-strepavidin (which are commercially available) either concurrently with or, preferably, following incubation of the antibody with the cells. A procedure for conjugating biotin to an antibody is described in Edward SA. Bayer et al, "The Avidin-Biotin Complex in Affinity Cytochemistry", in METHODS 0 10 IN ENZYMOLOGY Vol. 62 (1979) That article is incorporated herein by reference in its entirety.

An exemplary preferred direct labeling method for FITC is described in detail in the examples. A preferred indirect labeling method for biotinylation of an antibody and attachment of PE-strepavidin is also described.

To label the cells with the labelled, HLA antigen-specific antibodies, the cells are 0***0 incubated with the antibodies for a time sufficient for substantially complete antibody binding under the conditions used. An excess amount of antibody is preferably used. A 20 concentration of about 2 pg for 106 cells is desirable. However, use of 50 p1 of purified antibody at a concentration of about 40 gig/il is sufficient.

The cells are preferably incubated at about 40 C to maintain cell integrity. Incubation for about 30 minutes at 4°C is usually sufficient for substantially complete antibody binding. The sample is preferably mixed, as by using a hematology blood rocking device, during the incubation to ensure contact of the antibodies with the cells.

Preferably, the incubation is performed in the dark when using a fluorochrome label.

Secondary reactions incubation of fluorochrome-labelled avidin with biotinlabelled cells) are performed in the same manner.

Exemplary preferred cell labeling procedures for direct and indirect labels are described in detail in the examples.

PA0PER\FmspccW\=tI md1s-spc3.dm-4JlO1A)I -31 Sorting Labelled Cells Flow cytometry is a process in which the measurement of physical and/or chemical characteristics is made while the cells or particles pass, usually individually, through a measuring apparatus in a fluid stream. Biological particles, usually cells, have been subjected to flow cytometric analysis using acoustic, nuclear radiation, electronic and optical sensors. Optical measurements are used for the widest range of applications.

a Most of the present applications of flow cytometers derive from the ability of the 10 apparatus to define and quantify heterogeneous cell populations. Physical and chemical characteristics or parameters of cells which can be measured by flow cytometry include cell size, cell shape, pigment content, protein content, DNA content and DNA base 0. ratio. In addition, cells can be labelled with one or more fluorochrome(s) and characterized based on color differences.

The most demanding applications of flow cytometry require identification and subsequent characterization of subpopulations of cells. Both flow sorting and multiparameter analysis are used for this purpose.

20 Flow sorting employs electrical and/or mechanical means to divert cells with preselected characteristics from the main stream, and can be used to isolate populations of viable cells with more homogeneous characteristics than could be obtained by any other means. Flow sorting is particularly useful in circumstances in which further characterization of the selected cells requires short- or long-term maintenance in culture or analytical procedures which cannot be accomplished by flow cytometry.

The parameters which can be used to sort cells include measurements of light scattered by cells at two different angles (less than 2 commonly called forward scatter, and about 90 commonly called side scatter) from an incident laser beam. The fluorescence measurement capability of a multiparameter flow cytometer incorporating the light scattering measurements can then be used to determine other characteristics of the individual cell subpopulations.

P:\OPER\Fas\spcc\ftal clls-spe3.doc4)-I2/10 -32- For samples containing substantially larger numbers of fetal cells, the cells can be selected by a one or two step separation procedure by binding the cells to solid phaseaffixed anti-maternal HLA antigen antibodies. For example, maternal cells and fetal cells bind to a column or other solid phase (such as a tissue culture dish surface or magnetic beads) with the shared maternal antigen antibody. Only maternal cells bind to a column (or other solid phase) having antibody to the antigen the fetus did not inherit.

In this way, the fetal cells are separated from the maternal cells when incubated with the antibody to the antigen the fetus did not inherit.

For purposes of the present method, the cells are sorted using the four color patterns 1. produced by differential staining with two fluorochromes as selection criteria. When using red and green fluorochromes, the four patterns are unstained, red only, (3) green only and red and green.

Following staining or labeling of the cells with a red and a green fluorochrome for each of the two maternal alleles, maternal cells are red and green. Fetal cells are either red or green, depending on the allele the fetus inherited. By selecting cells exhibiting either red S or green fluorescence, but not both, the fetal cells can be purified.

Since most flow cytometers can sort on three parameters, an additional selection criterion can be cell size. This criterion eliminates dead cells, debris and contamination, such as sperm. Cell clumps can also be removed by this method, however, cell clumps may constitute clumps of fetal cells and can be sorted for further separation using the methods described above for separating cells.

In a preferred embodiment, an additional selection criterion is DNA content. Fetal cells having greater than 2C DNA content can be determined using a number of vital-staining fluorochromes such as the Hoechst dyes, DAPI (4'-6-diamidino-2-phenylindole), hydroethidine and 7-aminoactinomycin D (7AMD). The fluorochrome used depends on the labels used to select the fetal cells. A second laser capable of emitting UV light is P:\OPER\Fsspecfeal els-spe3.doc-02/10A/ I OS 0

S.

OS

S

S

S@ 0

S

S

S

S

S..

*005 -33 required to excite Hoechst and DAPI dyes. Each of the above-described dyes can be used with FITC and PE.

A flow cytometer can process about 10' to 108 cells per hour, usually about 4 x 107 cells per hour. Typically, the sample is prepared from at least about 2x108 cells which can be sorted in about 5 hours to provide sufficient fetal cells for analysis. However, substantially fewer cells may be required, depending on the analysis method to be used.

The ability of the cell sorter to separate maternal and fetal cells ultimately depends on 10 the percentage of fetal cells in the sample. To obtain a fetal cell sample that is at least about 60% pure (60% of the sorted cells are fetal cells), the fetal cells should constitute about 0.001% of the maternal cells or greater. Preferably, the sample contains more preferably 90% fetal cells post-sorting.

When 100% purity is desired, the sorted cells can be micromanipulated. For example, cell suspensions containing an individual cell per a preselected volume of suspension medium can be prepared by limiting dilution or by using a FACS with single cell sorting capabilities. Drops containing individual cells can be placed in suitable containers 96 well plates) and examined visually with a fluorescent microscope to 20 identify single-labelled (or unlabelled) cells. Wells containing those cells can be marked and the cells pooled.

For PCR analysis, analysis can be performed using a single, fetal cell. For karyotyping, the analysis can be performed using as few as five cells in metaphase. The number of cells necessary to obtain 5 metaphase cells will vary depending on the method used to induce metaphase and the length of culture required. For cells selected to have 2 C or greater DNA, a substantially shortened culture period may be used.

Alternatively, ways can be envisaged of identifying monozygosity (indicative of the presence of a monogenic disease) in a mixed cell population containing minimal fetal material including as few as one fetal cell in ten cells.

P:\OPER\Fas\spcc\felal clls-spc3.doc-042/l I -34- Following sorting, the separated cells can be washed twice in a physiologic buffer and resuspended in an appropriate medium for any subsequent analysis to be performed on the cells.

In circumstances where a second separation is necessary, cells can be incubated at 37 °C for about 2 hours, then washed to remove previously used antibodies, thus additional samples may not be required.

Post-Recovery Processing Following the present recovery method, whether based on solid phase or FACS separation, the fetal cells can be used in the same manner as fetal cells obtained by other methods such as amniocentesis and chorionic villus biopsy. The cells can be used as a source of DNA for analysis of the fetal alleles, as by polymerase chain amplification.

15 PCR analysis methods have been used to detect, for example, fetal sex, P-thalassemia, phenylketonuria (PKU), and Duchenne's muscular dystrophy.

Alternatively, the cells can be cultured in the same manner as biopsy materials for karyotyping analyses. However, the incubation period may be significantly shortened if 20 a DNA content of greater than or equal to 2C is used as a selection criterion.

Single cell suspensions can be prepared for analysis or further culturing. These single cells can be subjected to single PCR and subsequent genetic analysis or cultured to amplify the cell number and genetic material. The advantage of using single cell suspensions is that where the above described cell sorting on the basis of cell surface antigens does not provide 100% purity, but enrichment for fetal cells, the single cell suspensions allow analysis of individual cells, such that fetal cells can be more easily identified.

The purified fetal cells can also be used for therapeutic purposes, having potentiality lacking in fully differentiated cells. The cells may be used for transplantation and/or genetic manipulation.

PAOPERFasrspccctI Is-spe3.o-4)2/IOA)I This invention is further illustrated by the following specific but non-limiting examples.

Temperatures are given in degrees Centigrade and concentrations as weight percent unless otherwise specified.

EXAMPLE 1 Preparation of Labelled Antibodies Two monoclonal antibodies specific for two HLA A locus alleles were labelled. The 10 antibody designated GSP 16.1 (Anti-A 1, 9, 10 and 11; IgM monoclonal antibody from Genetic Systems, Seattle, Wash.; 10th International Workshop No. 2004) was labelled with FITC. The antibody designated GSP 20.1 (Anti-A2, w69; IgG monoclonal antibody from Genetic Systems, Seattle, Wash.; 10th International Workshop No. 2021) was labelled with biotin. Cells which reacted with the biotinylated antibody were 15 subsequently reacted with PE-labelled strepavidin. The reagents used to purify and label the antibodies are described below.

0 0 For the IgG antibody (GSP 20.1), the antibody was first purified by column chromatography using a MABTrap G Kit (Pharmacia LKB) to remove other proteins, 20 particularly albumin. This kit utilizes Protein G Sepharose 4 Fast Flow Chromatography media. (Use of Protein G Sepharose is equivalent.) The antibody-containing eluate ml) was concentrated in an Amicon B125 concentrator until the final volume was approximately 2.0 ml (approximately four times the original volume). The antibody was removed from the concentrator to a 12 x 75 mm test tube. The volume was adjusted to exactly 2.5 ml with biotinylation buffer.

Biotinylation Buffer 0.42 gm Na 2

CO

3 8.06 gm NaHCO 3 QS to 1 liter, pH to 8.4 P:\OPER\Fas\pec\fcal cclls-spc3.doc-2/10A)I -36- The IgM antibody was not pre-purified. A 0.5 ml antibody sample was diluted to 2.5 ml with FITC conjugation buffer.

FITC Conjugation Buffer 3.18 gm Na 2

CO

3 5.86 gm NaHCO 3 QS to 1 liter, pH to 9.6 A PD-10 column from Pharmacia LKB was washed with 25 ml of the appropriate 10 conjugation buffer (FITC conjugation buffer for the FITC-labelled antibody and biotin conjugation buffer for the biotinylated antibody). The 2.5 ml sample of each antibody o was applied to the appropriate column. When the column stopped dripping, a 3 ml aliquot of the appropriate conjugation buffer was applied. Each antibody-containing eluate was collected.

Each eluate was concentrated in an Amicon B125 concentrator until the final volume was 1 ml. The antibody was removed from the concentrator and placed in a 1.5 ml microcentrifuge tube. The amount of protein present was assessed using a kit (Protein **00 Assay Kit BioRad, Richmond, Calif.) that utilizes the Bradford method of protein quantitation.

Once the quantity of antibody (the amount of protein) was calculated, the FITC (Sigma Chemical Co., cat. #F-7250) and biotin (Molecular Probes, cat. #S-1582) were added as follows. A 100 p. aliquot of the working solution of FITC per mg of protein (Stock Solution FITC: Dissolve 10 mg in 1 ml dimethylsulfoxide (DMSO) and Working Solution FITC: Dilute stock to 1 mg/ml in FITC conjugation buffer) was added to the IgM antibody. A 120 kl of working solution of biotin (Stock Solution Biotin: Dissolve mg in 1 ml DMSO and Working Solution Biotin: Dilute stock to 1 mg/ml in DMSO )/mg protein was added to the IgG antibody. Each solution was mixed, and each tube was covered with aluminum foil and rocked on a hematology blood rocker overnight at room temperature.

PAOPERXFas'pcc\fernI =cHis-spe3.dom-4)2JR) I -37- One PD-10 column for each antibody was prepared by running 25 ml of Phosphate Buffered Saline (PBS), pH 7.4 (0.01M phosphate, 0.15M NaC1) through it. Following the incubation, each antibody preparation was diluted to 2.5 ml with PBS and layered over the column. When the column stopped dripping, a 3 ml aliquot of PBS was run through and the eluate was collected.

Each antibody was again concentrated in the Amicon B125 to a final volume of 1 ml.

The protein content was again quantitated, and the final concentration was adjusted to 40 gg/ml. In the case of the IgM antibody, the amount of antibody present was 10 estimated to be one-half of the total protein present and was diluted to 40 tg/ml. Bovine c Sserum albumin (BSA) was added to a final concentration of 10 mg/ml and sodium azide was added to a final concentration of 0.1%.

EXAMPLE 2 15 Sample Collection and Preparation of Cells for Labelling A sample was collected from a woman at 15 weeks of pregnancy. The sample was

*BO°

s obtained by lavage of the endocervical canal with 5ml of sterile saline and the sample was then washed twice with RPMI 1640 tissue culture media (Mediatech, Inc., 20 Herndon, Va.) with 5% fetal calf serum. The sample was incubated in acetyl cysteine to dissolve mucus and obtain a single cell suspension. The samples were then counted on a hemocytometer and adjusted to 10 7 /ml in RPMI.

EXAMPLE 3 Labelling Cells in a Sample Labelled antibody is usually used at a concentration of about 2 ug/10 6 cells. In this case, il of the antibody per 10 6 cells was used for staining. The cells were stained and sorted effectively using that amount of antibody. Increased antibody can be used to reduce false negatives, although specificity and cell recovery are reduced.

P:\OPER\Fas\pec\fc a cclls-spc3.doc-)2/10/01 -38- The cells (106 cells in 100 pl of RPMI 1640 tissue culture media plus 5% plasma), prepared as described in Example 2, were stained as described below to assess the conjugation of the label to the antibodies.

Tube #1-No antibody added (negative) Tube #2-GSP 16.1 (Anti-All, IgM monoclonal antibody) FITC conjugated; 50 l1 Tube #3-GSP 20.1 (Anti-A2, IgG monoclonal antibody) Biotin conjugated; 50 l1 Tube #4-GSP 16.1+GSP 20.1; 50 ul of each

S

0 8 10 The tubes were mixed and incubated in the dark at 4°C for 30 minutes. After this time, 10 ptl of TAGO (Burlingame, Calif.) PE-strepavidin was added to tubes #3 and This reagent reacts with biotin leading to a biotin/avidin-PE complex which allows the biotin-labelled cells to be fluorescent. The cells were incubated an additional 30 minutes after which the cells were washed twice with ice cold PBS and resuspended in 1 ml of

PBS.

*S*

EXAMPLE 4 Sorting Labelled Cells 20 Cells which were labelled as described in Example 3 were subjected to flow cytometric analysis using log green fluorescence (FITC) versus log red fluorescence (The cells were from a donor having the HLA type A2, All.) This results in a plot that is termed a two-dimensional (2D) plot, contour plot (if contours are drawn), dot plot (if dots are used to represent cells present) or cytogram. When the X-axis representing FITC and the Y-axis representing PE fluorescence are plotted against one another, the resulting graph can be divided into four quadrants. Lower left is where the negative or unstained cells fall. Lower right is where the FITC- labelled (green) cells fall. Upper left is where the PE- labelled (red) cells fall. Upper right is where the FITC- and PE-labelled (green and red) cells fall.

P:\OPER\Fas\spcIeal ccs-sp3.doco-02/lOl) -39- This property of cell identification based upon their labeling characteristics as they fall into the various quadrants of the graph was used to locate and subsequently sort out those located cells. In addition forward and side scatter measurements were used to gate cells and reduce cell debris and contaminants.

EXAMPLE Dilution Studies to Evaluate Rare Cell Sorting 0o In these studies, cells from a sample from donor A (sample A) are mixed with cells 10 from a sample from donor B (sample B) to mimic rare event cell analysis/sorting using a flow cytometer. This series of experiments is performed to evaluate the ability of flow cytometry to not only define but also purify these rare cells. These studies are performed by diluting samples from two individuals who have one A locus allele in common. In this way, a predetermined number of cells from each individual are known to be present in a sample.

0* o Antigens for GSP 16.1 and GSP 20.1 are both found on the surface of sample A cells.

These cells, when reacted with these monoclonal antibodies, have both red and green fluorescence on their surface. Antigen GSP 20.1 is also found on the surface of sample B cells. These cells, when reacted with this monoclonal antibody, have only red fluorescence on their surface. Thus, the labeling patterns mimic the patterns of a pregnant woman and the fetus.

Sample A is mixed into sample B at ratios of 0.1% and 0.001%. In the second study, sample A is mixed into sample B at ratios of 0.001% and 0.0001%.

The dilutions of A into B are calculated so that at least 1000 cells of A are present in the test tube.

The cells are labelled according to the procedure in Example 3. Monoclonal antibodies GSP 16.1 and GSP 20.1 are both added to each of the tubes of mixed cells using 50 t1d of antibody prepared as described in Example 1 per 106 cells. The cells are vortexed, P:\OPER\Fasspec\fcIal clls-spc3.doc-4)2/10A)I and the tube was covered with aluminum foil and placed on an ice pack which is mounted on a hematology blood rocking device.

After the antibody incubation, strepavidin-phycoerythrin (TAGO) is added to each tube of biotin-labelled cells using 10 pl of undiluted PE-strepavidin per 106 cells. The cells are again vortexed and incubated as previously for another 30 minutes.

Following the incubation, the samples are centrifuged at 1000 g for 3 minutes, washed twice with ice cold PBS, resuspended in RPMI to approximately the original volume and placed on ice.

A separate set of three control tubes containing 106 sample A cells only is reacted in the same manner as above with GSP 16.1, with GSP 20.1 and with GSP 16.1 in combination with GSP 20.1. A fourth aliquot of cells is left untreated with monoclonal antibodies. This group of four tubes is used to assess the antibody labeling of the cells in the study, and to set the flow cytometry instrument for color compensation, log amplification offset, signal gains, etc. The fluorescence parameters of the instrument are adjusted to produce four quadrant separation of the cells (as discussed above).

20 The lowest dilution tubes are analyzed/sorted first. Only one region (gate) is defined for S. sorting. This is an area where the single-labelled sample B cells would fall on log green versus log red fluorescence parameter display. By defining only one region, the computer has to make one decision in the evaluation of which cells to sort. If the cell fell into the defined gate region of the display, the cell is sorted.

The instrument is set to sort in one drop deflection/purity mode. This means a cell is sorted only if it can confidently be sorted without an unwanted cell being sorted with it.

The inclusion of unwanted cells can occur if the cells are located in close proximity to one another resulting in unwanted cells being included in a drop of fluid including wanted cells. This one drop deflection/purity mode lends an added discrimination to the process resulting in a higher percentage of purity in the sorted cells of choice.

P:\OPER\TFasspec\fel cllsspe3.doo4)2/ol)I -41 The average speed of the sort is approximately 10,000 cells/second. As the cells passed through the instrument, statistics are computed to evaluate the number of cells falling in the sort gate region as well as the number of cells which the instrument sorted.

A collection apparatus for the sorted cells consisted of a 10 ml beaker fitted into a solid support located directly under the sort streams. These beakers are layered with 100 to 200 dtl of approximately 50% autologous plasma in RPMI. This layer helps to cushion 10 the cells when they enter the collection vessel.

After at least 3,000 of the cells which are located in the gated region are sorted, the contents of the beaker is poured into another test tube and resubmitted to the flow cytometer for statistical analysis. This process of evaluating the efficacy of the sorting 0o procedure by analyzing the sorted cells using the instrument is referred to as "playback", "sort playback", or simply, "second pass". The statistics obtained from this evaluation of the sorted cells provides information on the alteration in the percentage of the two cell o populations from the original tube due to the sort process. It is from this analysis of the sorted cells that the percentage of purity or enrichment of the sorted population is evaluated. It is unlikely that the process will produce a population of 100% purity from low frequency, but the method may be repeated to further purify the fetal cell •00•00 population.

•Using this process, the study demonstrates that in tubes where a sample is diluted at or greater than 0.01%, the flow cytometer may not be able to discriminate the signal of the labelled cells as clearly as with the higher concentrations. However, the reliability of detecting only the rare cells starts to diminish. This is attributed to statistical problems related to the instrument whereby stray signals caused by inherent noise in the system due to cell clumps, debris, etc., may be evaluated in the region designated for our rare cells. To overcome this limitation, additional parameters, such as a third fluorescent color, cell size, etc. can be used to locate the rare cells.

The next higher dilution tube (0.001%) is sorted directly onto a microscope slide coated with autologous plasma (dry). Evaluation of the wet mount of this sorted population P:\OPER\Farspcc~fctaI clls.spo3.d4)2/JOA)I 42using a fluorescence microscope reveals quality of sorting. Visual inspection may reveal a background of dye crystals, debris, etc. which, if played back, would have been picked up by the flow cytometer and counted resulting in what would appear as a lower overall recovery of dual-labelled cells. Less diluted samples are expected to have higher purity than more diluted samples. Therefore, the 0.01% dilution would be expected to have greater purity.

The evaluation of the dilution-sort studies demonstrated that the present method makes o• S°it possible to locate, sort and enrich the diluted cell population appreciably from its original low percentage.

S'. 0 000 EXAMPLE 6 Dilution Studies Using Single-Labelled Rare Cells Another study is performed as described in Example 5 in which the single-labelled cells are diluted to be the rare cells. (The antibody labeling procedure is not repeated.) Sample A is treated as the maternal, dual-labelled cells and sample B as the fetal, S single-labelled cells at dilutions of 0.1, 0.01 and 0.001%. This demonstrates that the 6669 fetal cells can be separated from the maternal cells based upon a single fluorescent label. This more closely mimics the inventive method since the mother's antigens can be determined, allowing her cells to be dually labelled. As demonstrated by this study, any 00 !combination of the four quadrants where cells can be placed (unstained, red only, green only and red plus green) can be used to separate/sort the cells of interest.

EXAMPLE 7 FACS Separation of Rare Cells A sample is taken from the cervical canal of a woman in her ninth week of pregnancy by endocervical lavage. As in the studies described in Examples 5-6, the mother and father shared a common antigen for the A locus. The mother's other A antigen is A2, PAOPERFa~qxcWe dI =e-spc3.doo-02/I0AJI 43 which reacts with the GSP 20.1 antibodies. The father's discriminating A antigen reacts with GSP 16.1 antibodies;.

For this study, the maternal GSP 20.1 antibodies are labelled with biotin/PE and the paternal GSP 16.1 antibodies are labelled with FITC as described in Example 1. The sample is purified and stained as described in Examples 2 and 3. For this study, maternal cells are stained red; paternal cells green.

S• Cells in the maternal sample are predominantly red with rare (fetal) cells labelled with red and green. Out of about 5.5 x 106 sorted maternal cells, cells retrieved by the red and green sort criteria comprise about 0.1% of the cell population. Of those sorted cells, about 20% are red and green labelled.

This experiment demonstrates that fetal cells are present in the cervical canal of a woman in her first trimester of pregnancy, and can be retrieved using a fluorescent staining pattern that distinguishes maternal and fetal HLA antigens encoded by a single SHLA locus.

0goO o0oo EXAMPLE 8 Separation of Rare Cells Using Magnetic Beads 0 00 S. Additional samples of the cells used in Example 7 are separated using DYNABEAD magnetic beads (Robbins Scientific, Mountain View, Calif.). The GSP 16.1 antibody (IgM), the fetal HLA A antigen not present on maternal cells, is affixed to the beads by conjugation of the antibody according the manufacturer's directions. This antibody separation functions in the same manner as separations based on binding maternal cells, except that, in this case, the rare, fetal cells are solid phase-affixed and maternal cells are non-bound (rather than maternal cells for the nontransmitted antigen being solid phase-affixed and fetal cells being non-bound as in the inventive method).

The maternal sample is purified by ammonium chloride lysis and cells are suspended in PBS at 4°C. The manufacturer recommends using 1 x 106 to 1 x 107 beads per ml and a P:\OPER\Fas\spcc\felal cllsspe3.doc-02/IOA) -44ratio of about five beads per cell. However, upon consultation, a ratio of about 1 or fewer beads per cell is suggested for applications where cells of choice are removed from a mixture of cells.

An experiment using a ratio of 0.05 and 0.5 beads/cell is performed. The GSP 16.1 antibody is conjugated to the beads ("16.1 beads") and a negative control antibody (GSP 56.1, an antibody for HLA B 27.1, an antigen not present in either parent) is conjugated to a second set of beads ("56.1 beads") to assess non-specific binding. The purified cells are incubated with the beads at 4 C for 60 minutes.

The 16.1 beads using a 0.05 to 1 ratio had too few beads to evaluate. The 16.1 beads using a 0.5 to 1 ratio had readily visible rosettes of cells. The 56.1 beads had a few cells per bead and no prominent rosettes. Cytocentrifugation of the control bead preparation demonstrated that the rosetted cells are mainly monocytes.

To verify the fetal origin of the rosetted cells, the study is repeated using the bead to cell *0 ratio of 0.5. Since the beads are sufficiently large to permit binding of FITC-labelled cells, the cells are counter-labelled with FITC-labelled GSP 16.1 antibodies prior to incubation with the beads so that fetal cells are green. The cells are incubated with the 20 beads as described above, except that whenever fluorescent labels are used, the incubation is performed in the dark. Following incubation, the solid phase affixed fetal cells are separated from maternal cells using a magnet and removed from the magnetic beads. The separated fetal cells are placed in RPMI 1640 medium for further analysis.

Analysis of the solid phase-affixed cells showed the cells are green, confirming that the cells are of fetal origin.

Claims (24)

1. A method for recovering fetal cells from a uterine sample obtained from a pregnant female having a first cell surface antigen encoded by a first allele of a polymorphic genetic locus and a second, different cell surface antigen encoded by a second allele of a polymorphic genetic locus, said method comprising: a. contacting cells of said sample with a first antibody specific for said first cell surface antigen for a period of time sufficient for antibody binding; o b. separating maternal cells bound to said first antibody from fetal cells 0oo which are non-bound; and 0g0 c. recovering said separated fetal cells.

2. The method of claim 1 wherein said first antibody is labelled with a 00*0 fluorochrome and cells bound to said first antibody are separated from cells which are not bound to said first antibody using fluorescence-activated cell sorting. S.

3. The method of claim 2 additionally comprising the steps of a. contacting said cells with a second antibody specific for a second, different cell surface antigen with a second different fluorochrome label, for a period of time sufficient for antibody binding to produce double labelled maternal cells; and b. separating double labelled maternal cells from single bound or non- bound fetal cells. P:\OPER\Fs\pc\fIaI mIIs-sp3.dx4-()2IIJ) -46-

4. The method of claim 1 wherein said first antibody is affixed to a solid phase and cells bound to said first antibody are separated from cells which are not bound to said first antibody.

The method of claim 4 additionally comprising the steps of a. contacting the cells of said sample with a second solid phase-affixed second antibody specific for said second antigen for a period of time sufficient for antibody binding; and B. 0• S° b. separating cells bound to each of said solid phases by said first antibody 0e S or by said second antibody from any non-bound cells.

6. The method of claim 4 or 5 wherein said solid phase is magnetic beads and solid phase-affixed cells are separated from non-bound cells using a magnet. o 0 ro

7. The method of any one of claims 3 or 5 wherein the cells are separately contacted with said first antibody and said second antibody.

8. The method of any one of claims 3 or 5 wherein said cells are contacted sequentially or simultaneously with said first antibody and said second antibody. S..

9. The method of any one of claims 3 or 5 to 8 wherein said first and said second antibodies are specific for cell surface antigens expressed by different alleles of the same polymorphic genetic locus.

The method of any one of claims 3 or 5 to 8 wherein said first and said second antibodies are specific for cell surface antigens expressed by said first and second alleles of two different polymorphic genetic loci.

11. The method of any one of claims 1 to 10 wherein said first allele is a MHC allele P:\OPER\Fas\spcc\flci =dIs-spc3.dc4J2/IOA)I -47-

12. The method of claim 11 wherein said MHC allele is an HLA Class I allele.

13. The method of any one of claims 9 or 10 wherein said first and said second alleles are MHC alleles.

14. The method of claim 13 wherein said first and said second alleles are HLA Class I alleles.

15. The method of any one of claims 12 or 14 wherein said HLA allele is selected from the group consisting of HLA-A, HLA- B, and HLA-C. 0*

16. The method of any one of claims 1 to 15 additionally comprising the step of determining a genotype of said pregnant female for said loci.

17. The method of any one of claims 1 to 16 additionally comprising the step of determining a genotype of a father of the fetus for said loci. 0000

18. The method of any one of claims 1 to 17 wherein said sample is treated to reduce cell clumping. O

19. The method of claim 18 wherein said sample is treated with acetyl cysteine.

The method of any one of claims 1 to 19 wherein said sample is treated to reduce contaminants prior to contact with said first antibody.

21. The method of any one of claims 1 to 20 comprising the additional step of culturing said recovered fetal cells.

22. The method of any one of claims 1 to 21 comprising the additional step of obtaining DNA from one or more recovered fetal cells. P:\OPER\Fas\pcc\fclal clls-spc3.doc-02/IO I -48-

23. A method for obtaining DNA from fetal cells recovered from a uterine sample obtained from a pregnant female having different first and second antigens expressed by a first and a second allele of an HLA locus, said method comprising: a. contacting the cells of said sample with a first antibody specific for said first antigen and labelled with a first fluorochrome and with a second antibody specific for said second antigen and labelled with a second, different fluorochrome for a period of time sufficient for antibody binding to produce a S° sample containing labelled cells; 6 6 b. separating double-labelled cells of said sample from single-labelled cells using fluorescence-activated cell sorting to produce separated, single-labelled cells; c. recovering said separated, single-labelled cells; and d. obtaining DNA from said recovered fetal cells. b 0 *:ego:

24. A method for recovering and culturing fetal cells from a uterine sample obtained from a pregnant female having different first and second antigens expressed by a first and a second allele of an HLA locus, said method comprising: a. contacting the cells of said sample with a first antibody specific for said first antigen and labelled with a first fluorochrome and with a second antibody specific for said second antigen and labelled with a second, different fluorochrome for a period of time sufficient for antibody binding to produce a sample containing labelled cells; b. separating double-labelled cells of said sample from single-labelled cells using fluorescence-activated cell sorting to produce separated, single-labelled cells; and PAOPERT\Fspoc6feal .s-p0.d-02/1RiA I 49 C. recovering said separated, single-labelled cells; and d. culturing said recovered fetal cells. A method according to any one of claims 1 to 24 substantially as described herein before with reference to the examples. DATED this 2 nd day of October 2001 Genetype A. G. by DAVIES COLLISON CAVE Patent Attorneys for the Applicants a me C 00 00 0 0 0 *a *0 0 0 6 Oe 0 *00 as 4 S.. **oo a* 0 D**4 *0 so*# 6000 6 S** 0